Drug-Nucleic Acid Interactions

Preface Progress in molecular biology and studies of small molecule binding to nucleic acids have been inextricably lin...

Author: Jonathan B. Chaires | Michael J. Waring

62 downloads 1886 Views 12MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Preface

Progress in molecular biology and studies of small molecule binding to nucleic acids have been inextricably linked. A testament to that fact is the inclusion of eight papers directly concerned with drug-DNA interactions among the recently published list of the 100 most cited articles in the Journal of Molecular Biology. Few other scientific areas are as well represented on that list. Small molecules have perhaps taught us more about DNA than DNA has taught us about small molecules. Watson, for example, notes in the Molecular Biology of the Gene that the "fact that intercalation occurs so readily indicates that it is energetically favored... [and] is additional evidence for the metastability of the double-helical structure--its ability to assume many inherently unstable configurations that normally revert quickly back to the standard B conformation." From that point of view, intercalation provided one of the very first indications of the plasticity of DNA, an area that has blossomed to reveal an incredible diversity of structural forms. Perhaps the most widespread interest in small molecules that bind to nucleic acids stems from their potential as useful pharmaceutical agents. Indeed, some of the very best anticancer drugs are well-documented DNA binders. While interest in drug-DNA interactions has at times waned, recent advances in chemical synthesis, analytical instrumentation to measure binding, and structural biology have greatly enhanced the potential for rational design of new therapeutic compounds. Accordingly, studies on the interaction of small molecules with nucleic acids have taken on new life and have helped spawn several emergent biotechnology companies dedicated to exploiting the promise of making new types of pharmaceuticals targeted at nucleic acids. The aim of this volume is to consolidate key methods for studying ligandnucleic acid interactions, both old and new, into a convenient source. Accordingly, we have solicited from experts in a variety of disciplines articles that concisely but completely describe useful methods and strategies for studying small molecule binding to nucleic acids. Techniques that are useful now range from biophysical and chemical approaches to methods rooted in molecular and cell biology. We hope that this volume will serve as a useful compendium of methods both to newcomers entering the field as well as to scientists already actively engaged in research in this area. JONATHANB. CHAIRES MICHAELJ. WARING

xiii

Contributors to Volume 3 4 0 Article numbers are in parentheses following the names of contributors. Affiliations listed are current.

CHRISTIAN BAILLY (24, 31), INSERM

CARLEEN M. CULL1NANE (23), Pharma-

U-524, and Laboratoire de Pharmacologie Antitumorale du Centre Oscar Lambret IRCL, 59045 Lille, France

cology and Developmental Therapeutics Unit, Peter MacCallum Cancer Institute, Victoria 3002, Australia

ALBERT S. BENIGHT (8), Department

SUZANNE M. CUTrS, (23), Department of

of Chemistry, University of Illinois, Chicago, Illinois 60607 and DNA Codes LLC, Chicago, Illinois 60601

Biochemistry, La Trobe University, Bundoora, Victoria 3083, Australia

LAWRENCE A. BOTTOMLEY(11), School of

Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332 SOPHIA Y. B REUSEGEM(10), Laboratoryfi~r

Fluorescence Dynamics, Department of Physics, University of Illinois, Urbana, Illinois 61801 JONATHAN B. CHAIRES (1, 5, 27), De-

partment of Biochemistry, University of Mississippi Medical Center, Jackson, Mississippi 39216 YEN CHOO (30), Gendaq Limited, London

NW7 lAD, United Kingdom BABUR Z. CHOWDHRY(6), School of Chem-

ical and Life Sciences, University of Greenwich, London SE18 6PF, United Kingdom ROBERT M. CLEGG (10), Laboratory for

Fluorescence Dynamics, Department of Physics, University of Illinois, Urbana, Illinois 61801 DONALD M. CROTHERS (3, 23), Depart-

ment of Chemistry, Yale University, New Haven, Connecticut 06520-8107 MARK S. CUBBERLEY (28), Department of

Chemistry and Biochemistry, University of Texas, Austin, Texas 78712

JAMES C. DABROWIAK(21), Department of

Chemistry, Center for Science and Technology, Syracuse University, Syracuse, New York 13244 TINA M. DAVIS (2), Department of Chem-

istry, Georgia State University, Atlanta, Georgia 30303 PETER B. DERVAN (22), Department of

Chemistry, California Institute of Technology, Pasadena, California 91125 MAGDALENAERIKSSON(4), Department of

Physical Chemistry, Chalmers University of Technology, Gothenburg SE-41296, Sweden, and Department of BiDchemistry, University of Gothenburg, Gothenburg SE-40530, Sweden CHRISTOPHE ESCUDI~ (16), Laboratoire

de Biophysique, INSERM U201, CNRS UMR 8646, Museum National d'Histoire Naturelle, 75231 Paris Cedex 05, France IZABELA FOKT (27), M. D. Anderson Can-

cer Center, University of Texas, Houston, Texas 77030 KEITHR. Fox (20), Division of Biochemistry

and Molecular Biology, School of Biological Sciences, University of Southampton, Southampton S016 7PX, United Kingdom

x

CONTRIBUTORS TO VOLUME 340

THI~RI~SE GARESTIER (16), Laboratoire

BESIK I. KANKIA (7), Department of

de Biophysique, 1NSERM U201, CNRS UMR 8646, Museum National d'Histoire Naturelle, 75231 Paris Cedex 05, France

Pharmaceutical Sciences, University ~f Nebraska Medical Center, Omaha, Nebraska 68198

JERRY GOOD1SMAN (21), Department of

ASMITA KUMAR (33), Department of Bio-

Chemistry, Center for Science and Technology, Syracuse University, Syracuse, New York 13244 DAVID E. GRAVES (18), Department of

Chemistry, University of Mississippi, University, Mississippi 38677 KEITH A. GRIMALDI(17), CRC Drug-DNA

Interactions Research Group, Royal Free and University College Medical School, University College London, London WI P 8BT, United Kingdom VLAD1M1RM. GUELEV (28), Department of

Chemistry and Biochemistry, Universi~" of Texas, Austin, Texas 78712 Krebs Institute for Biomolecular Science, Department of Chemisto, University of Sheffield, Sheffield $3 7HF, United Kingdom

IHTSHAMUL HAG (6),

JOHN A. HARTLEY (17), CRC Drug-DNA

chemistry, University of Mississippi, Jackson, Mississippi 39216 DONALD W. KUPKE (7), Department of

Chemistrry, University of Virginia, Charlottesville, Virginia 22901 ANDREW N. LANE (12), Division of Molecu-

lar Structure, National Institute for Medical Research, London NW7 IAA, United Kingdom GREGORY H. LENO (33), lnfgen Incorpo-

rated, DeForest, Wisconsin 53532 PETER T. LILLEHEI (l l), School of Chem-

istry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332 R. SCOTT LOKEY (28), Department of

Chemistry and Biochemistr); University of Texas, Austin, Texas 78712

Interactions Research Group, Royal Free and University College Medical School, University College London, London W1P 8BT, United Kingdom

FRANK G. LOONTIENS (10), Laboratory

PAUL B. HOPKINS (19), Department of

RYAN A. LUCE (19), Department of Chem-

for Biochemistry, WEVIO, University of Gent, Gent 9000, Belgium

Chemistry, University of Washington, Seattle, Washington 98195

istr); University of Washington, Seattle, Washington 98195

LAURENCE H. HURLEY (29), College of

CHRISTOPHE MARCHAND (32), Laboratory

Pharmacy, University of Arizona, Tucson, Arizona 85721 and Arizona Cancer Center, Tucson, Arizona 85724

of Molecular Pharmacology, Division of Basic Sciences, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892

Gendaq Limited, London NW7 laD, United Kingdom

MARK ISALAN (30),

Chemistry and Biochemistry, University of Texas, Austin, Texas 78712

Department of Pharmaceutical Sciences, University of Nebraska Medical Center, Omaha, Nebraska 68198

TERENCE C. JENKINS (6), Yorkshire Cancer

CLAIRE J. MCGURK (17), CRC Drug-DNA

Research Laboratory of Drug Design, Cancer Research Group, University of Bradford, Bradford BD7 1DP, United Kingdom

Interactions Research Group, Royal Free and University College Medical School, University College London, London WI P 8BT, United Kingdom

BRENT L. IVERSON (28), Department of

LUIS A. MARKY (7),

CONTRIBUTORS TO VOLUME 340 PETER J. MCHUGH (17), CRC Drug-DNA

Interactions Research Group, Royal Free and University College Medical School, University College London, London W1P 8BT, United Kingdom MARK P. MCPIKE (21), Department of

Chemistry, Centerfor Science and Technolog); Syracuse University, Syracuse, New York 13244

xi

R. PHILLIPS (23), Department of Biochemistry, LaTrobe Universit3; Bundoora, Victoria 3083, Australia

DON

YVES POMMIER (32), Laboratory of Molec-

ular Pharmacolog); Division of Basic Sciences, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892 JOSl~ PORTUGAL(25, 27), Departamento de

Chemistry and Biochemistry, Universi~ of Texas, Austin, Texas 78712

Biologia Molecular y Celular, lnstituto de Biologia Molecular de Barcelona, CSIC, Barcelona 08034, Spain

NOURI NEAMATI(32), Laboratory of Molec-

WALDEMAR PRIEBE (27), M. D. Ander-

ular Pharmacology, Division of Basic Sciences, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892

son Cancer Center, University of Texas, Houston, Texas 77030

MEREDITH M. MURR (28), Department of

JAROSLAV NESETI~IL (8), Department of

Applied Mathematics, Faculty of Mathematics and Physics, Charles Universi~, 118 O0 Praha 1, Czech Republic PETER E. NIELSEN (15), Department of

Medical Biochemistry and Genetics, The Panum Institute, University of Copenhagen, Copenhagen DK-2200, Denmark BENGT NORD~N (4), Department of Phys-

ical Chemistry, Chalmers University of Technology, Gothenburg SE-41296, Sweden

TERESA PRZEWLOKA (27), M. D. Ander-

son Cancer Center, University of Texas, Houston, Texas 77030 PETER REGENFUSS (10), Laboratory for

Fluorescence Dynamics, Department of Physics, Universi~' of Illinois, Urbana, Illinois 61801 JINSONG REN (5), Department of Biochem-

istry, University of Mississippi Medical Center, Jackson, Mississippi 39216 RICCELLI (8), Department of Chemistry, University of Illinois, Chicago, Illinois 60607, and DNA Codes LLC, Chicago, Illinois 60601

PETER V.

Department of Chemistry, University of Illinois, Chicago, Illinois 60607, and Integrated DNA Technologies, Coralville, Iowa 52241

RICHARD D. SHEARDY(26), Department of

PETR PAN(~OSKA(8), Department of Chem-

School, Elmwood Park, New Jersey 07407

RICHARD OWCZARZY (8),

istry, University of Illinois, Chicago, Illinois 60607, and Center for Discrete Mathematics, Applied Computer Science and Applications DIMAT1A, Charles University, Prague, Czech Republic, and DNA Codes LLC, Chicago, Illinois 60601 MARY ELIZABETH PEEK (13), School of

Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332

Chemistry and Biochemistry, Seton Hall Universit); South Orange, New Jersey 07079 ANGELA M. SNOW (26), Memorial High CHARLES H. SPINK (9), Department of

Chemistry, State University of New York, Cortland, New York 13045 DAEKYU SUN (29), Institute for Drug De-

velopment, San Antonio, Texas 78245 JIAN-SHENG SUN (16), Laboratoire de

Biophysique, INSERM U201, CNRS UMR 8646, Museum National d'Histoire Naturelle, 75231 Paris Cedex 05, France

xii

CONTRIBUTORS TO VOLUME 340

MICHAEL J. TILBY (17), Cancer Research

Unit, Medical School, University of Newcastle Upon Tyne, Newcastle NE2 4HH, United Kingdom JOHN W. TRAUGER (22), Department of Chemistry, California Institute of Technology, Pasadena, California 91125 JOHN O. TRENT (14, 27), James Graham Brown Cancer Center, Department of Medicine, University of Louisville, Louisville, Kentucky 40202 PETER M. VALLONE (8), Department of Chemistry, University of Illinois, Chicago, Illinois 60607 and National Institute of Standards" and Technology, Biotechnology Division, Gaithersburg, Mao, land 20899 MICHAEL J. WARING (20, 24), Department of Pharmacology, University of Cam-

bridge, Cambridge CB2 IQJ, United Kingdom SUSAN E. WELLMAN (9), Department of Pharmacology and Toxicolog), University of Mississippi Medical Center, Jackson, Mississippi 39216 LOREN DEAN WILLIAMS (13), School of

Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332 W. DAVID WILSON (2), Department of

Chemistry, Georgia State University, Atlanta, Georgia 30303 HONGZH1XU (33), Department of Biochem-

istry, University of Mississippi, Jackson, Mississippi 39216 STEVEN M. ZEMAN (3), Department of

Chemistry, Yale University, New Haven, Connecticut 06520

[1]

ANALYSTSOF LIGAND-DNA BINDINGISOTHERMS

3

[1] Analysis and Interpretation of Ligand-DNA Binding Isotherms By JONATHANB. CHA1RES

Introduction To attain a reasonable understanding of any ligand-receptor interaction, it is necessary to answer the questions posed by Scatchard ~ more than 50 years ago: "How many? How tightly? Where? Why? What of it?" The first two Questions (and in part the third) can be answered by equilibrium binding studies, and are the primary focus of this chapter. The remaining questions concisely express the concerns of structural and functional studies, and may be addressed by X-ray crystallography, nuclear magnetic resonance (NMR) techniques, molecular modeling, and a variety of chemical and molecular biological methods. Macromolecular binding is a phenomenon of general interest, and the underlying general principles are the same for ligand binding to proteins or to nucleic acids. A number of excellent general treatments of macromolecular binding are available that explain the underlying physical chemistry in detail .2-6 What distinguishes the binding of small molecules to DNA from their binding to proteins is the need to account for behavior arising from the lattice properties of linear DNA molecules. Various neighbor exclusion models have evolved to cope with that complexity, and are described. An excellent discussion of the principles of nucleic acid binding interactions is provided by Bloomfield et al. 7 Determination of the binding constant K allows the binding free energy change, AG, to be calculated by the standard Gibbs equation, AG = - R T In K, where R is the gas constant and T is the temperature in degrees Kelvin. From studies of the temperature dependence of the binding constant, or (preferably) by calorimetric studies, the binding enthalpy (AH) may be obtained. The binding free energy may then be partitioned into its enthalpic and entropic components, AG = AH -- TAS, where AS is the entropy change. Knowledge of these thermodynamic parameters

I G. Scatchard, Ann. N.Y. Acad. Sci. 51,660 (1949). 2 j. T. Edsall and J. Wyman, "Biophysical Chemistry." Academic Press, New York, 1958. 3 j. Wyman and S. J. Gill, "Binding and Linkage." University Science Books, Mill Valley, California, 1990. 41. M. Klotz, "Ligand Receptor Energetics." John Wiley & Sons, New York, 1997. 5 E. diCera, "Thermodynamic Theory of Site-Specific Binding Processes in Biological Macromolecules." Cambridge University Press, Cambridge, 1995. 6 G. Weber, "Protein Interactions." Chapman & Hall, New York, 1992. 7 V. A. Bloomfield, D. M. Crothers, and J. Ignacio Tinoco, "Nucleic Acids: Structures, Properties and Functions," 1st Ed. University Science Books, Sausalito, California, 2000.

METHODS IN ENZYMOLOGY, VOL. 340

Copyright © 2001 by Academic Press All rights of reproduction in any form reserved. 0076-6879/00 $35.00

4

BIOPHYSICALAPPROACHES

[ 11

provides a firm foundation for understanding the molecular forces that govern the binding reaction, allowing one to begin to address Scatchard's question "Why?" Details of attempts to parse binding free energies for ligand-DNA interactions in order to understand the contribution of various molecular forces are described in publications from this and other laboratories, s- 12 The aim of this chapter is to offer a concise guide for the analysis and interpretation of ligand-DNA binding isotherms. Methods for experimentally obtaining binding data are not discussed because detailed, practical descriptions of experimental protocols are available. 13 15 In this chapter, examples of binding data are taken from results obtained in the author's laboratory with the anticancer agent daunomycin (daunorubicin). Daunomycin is perhaps the best-characterized DNA intercalator, and its binding to a wide variety of DNA sequences and structures has been thoroughly investigated. ~6,17 Model-Independent Approaches Figure 1 shows the results from two types of binding experiments, each of which addresses one of Scatcbard's queries as directly as possible. The method of continuous variations Is-a1 may be used to construct a so-called Job plot (Fig. 1A). Binding stoichiometries may be determined from such plots without recourse to any assumed binding model. For the data shown in Fig. 1A for the interaction of daunomycin with calf thymus DNA, an inflection near 0.2 mol fraction ligand indicates a binding stoichiometry of one ligand per 3 or 4 base pairs. The exact stoichiometry from the inflection at 0.21 mol fraction is (1.0 - 0.21)/0.21 = 3.76 base pairs. This value represents the predominant binding mode, although an s j. B. Chaires, Anticancer Drug Des. 11,569 (1996). 9 j. B. Chaires, Biopolymers 44, 201 (1997). l01. Haq, J. E. Ladbury, B. Z. Chowdhry, T. C. Jenkins, and J. B. Chaires, J. Mol. Biol. 271,244 (1997). II j. Ren, T. C. Jenkins, and J. B. Chaires, Biochemistry 39, 8439 (2000). 12 S. Mazur, F. A. Tanious, D. Ding, A. Kumar, D. W. Boykin, I. J. Simpson, S. Neidle, and W. D. Wilson, J. Mol. Biol. 300, 321 (2000). 13 X. Qu and J. B. Chaires, Methods Enzymol. 321, 353 (2000). L4T. C. Jenkins, in "Drug-DNA Interaction Protocols" (K. R. Fox, ed.), Vol. 90, pp. 195-218. Humana Press, Totowa, New Jersey, 1997. 15 p. C. Dedon, in "Current Protocols in Nucleic Acid Chemistry" (S. L. Beaucage, D. E. Bergstrom, G. D. Glick, and R. A. Jones, eds.), Vol. 1, pp. 8.2.1-8.2.8. John Wiley & Sons, New York, 2000. 16 j. B. Chaires, in "Advances in DNA Sequence Specific Agents" (L. H. Hurley, ed.), Vol. 2, pp. 141167. JAI Press, Greenwich, Connecticut, 1996. 17 j. B. Chaires, Biophys. Chem. 35, 191 (1990). 18 E Job, Ann. Chim. (Paris) 9, 113 (1928). 19 C. Y. Huang, Methods Enzymol. 87, 509 (1982). 2o A. Waiters, Biomed. Biochim. Acta 44, 132t (1985). 21 E G. Loontiens, E Regenfuss, A. Zechel, L. Dumortier, and R. M. Clegg, Biochemistry 29, 9029 (1990).

[1]

ANALYSIS OF L I G A N D - D N A BINDING ISOTHERMS

I

I

I

200

I

/

I

1.2

,

,

,

,

5 ,

1.0

,

g

t

0.8

-200 0.6 -400

0.4

-600

0.2

-800

o.o

•

B

, I , I

0.0

0.2

0.4

0.6

0.8

-20

-18

Mole Fraction Daunomycin

-16

I

I

-14

-12

-t0

In Cf

FIG. I. Daunomycin binding to calf thymus DNA. (A) Job plot obtained from fluorescence titration studies. A F is the difference in fluorescence emission intensity between solutions of daunomycin alone and in the presence of DNA. The minimum indicates a binding stoichiometry of 3 or 4 base pairs. (B) Binding isotherm for the daunomycin--calf thymus DNA interaction. The fractional saturation was calculated assuming a 3-bp binding site. The abscissa is the natural logarithm of the free daunomycin concentration.

inflection near 0.5-0.6 mol fraction indicates an additional binding mode at higher drug concentrations. The results shown here, based on fluorescence data, agree well with data based on absorbance changes. 2° The Job plot thus answers the question "How many?" directly. In studies of ligand-DNA interactions, this method has been underutilized and its advantages largely unappreciated. In the case of multiple binding modes, the method of continuous variations is particularly valuable, and clearly reveals complexities in the binding process. Published examples for the groove-binder Hoechst 3325821 and for the bisintercalating anthracycline WP63122 illustrate the value of the method in cases of complicated, multimode binding interactions. Figure 1B shows a titration binding isotherm for the daunomycin-calf thymus DNA interaction. In this form, the fractional occupancy of binding sites is shown as a function of the natural logarithm of the free daunomycin concentration (Ct-). The fractional occupancy was calculated from the experimentally determined binding

22 F. Leng, W. Priebe, and J. B. Chaires, Biochemistry 37, 1743 (1998).

6


[ 1]

ratio r (moles daunomycin bound per mole base pair) and the binding stoichiometry was determined from the Job plot shown in Fig. 1A. The form of the plot shown in Fig. 1B is regarded by some 3 as the most fundamental representation of binding data because the logarithm of the free ligand activity is proportional to the chemical potential of the ligand. For simple binding to identical, noninteracting sites, titration binding curves should be symmetric about a midpoint located at a ligand concentration that is the reciprocal of the association binding constant, and should cover a span of 1.8 lOgl0 units (4.14 In units) in going from 0.1 to 0.9 fractional saturation. 3'4'6 The data shown in Fig. 1B cover a span of 5.4 in units (2.4 log~0 units) and represent an essentially complete binding titration curve. The span is greater than expected for simple binding, which indicates negative cooperativity, neighbor exclusion, or heterogeneity of binding sites. Perhaps the main advantage of the data shown in Fig. 1B is that they may be analyzed in a model-independent way by using the Wyman concept of median ligand activity. 3'5 The free energy of ligation (AGx) to go from a state where no ligand is bound to a degree of saturation of k? is given by Eq. (1): P 2 AGx = RT Jo In Cfg~"

(1)

where RT has its usual meaning. The pronounced advantage of Eq. (1) is that it provides a free energy estimate to attain any degree of saturation without recourse to any specific binding model. Numerical integration of the data in Fig. 1B yields an estimate of AGx = - 7 . 8 kcal mol I for the full ligation of a daunomycin binding site. Free energies derived from binding constants obtained by curve fitting to specific models must agree with this model-independent value if the model is reasonable. Neighbor Exclusion Models Figure 2 shows data for the daunomycin-calf thymus DNA interaction in the form of a Scatchard plot, J by far the most common representation of binding data for ligand-DNA interactions. To explain the curvature in such plots, a variety of neighbor exclusion models were proposed, 23'24 and these have become the most commonly used models for the interpretation of binding isotherms. Neighbor exclusion models assume (in their simplest form) that the DNA lattice consists of an array of identical and noninteracting potential binding sites. The base pair is commonly defined as the lattice binding site for duplex DNA. Ligand binding to any one site occludes neighboring sites from binding as defined by the site size n. As the lattice approaches saturation, the probability of finding a stretch 23 D. M. Crothers, Biopolymers6, 575 (1968). 24j. D. McGheeand R H. yon Hippel,J. Mol.Biol. 86, 469 (1974).

[ 1]


8.0x10

5

,

,

,

I

,

,

7

I

6.0xl 0s

"',O • 4.0xl

0s

• o~',~

G~

C3

2.0x10 5

e ~ - . .

0.0

I

I

I

I

I

I

I

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

r FIG. 2. Scatchard plot for the daunomycin-calf thymus DNA interaction. The solid line is the best fit of the neighbor exclusion model [Eq. (2)] to the experimental data yielding the parameters shown in Table I. The dashed line is the best fit with the exclusion parameter constrained to an integral value of 3.

of unoccupied DNA n base pairs long decreases, producing the curvature seen in Fig. 2. The curvature does not result from a decrease in the intrinsic binding affinity, but rather arises from the decreased probability of finding a free site of the appropriate size. McGhee and von Hippe124 derived a closed form equation that embodies the neighbor exclusion model [Eq. (2)]:

r

--

=

K(1

-

nr)

I

1;

1--"r

]°' J

(2)

where K is the association constant for ligand binding to an isolated lattice site, n is

8

BIOPHYSICAL APPROACHES

[ 1]

the neighbor exclusion parameter, and r is the binding ratio. Since its publication in 1974, the McGhee and von Hippel article has been cited more than 1350 times, and is certainly the most commonly used model for the interpretation of ligand binding to DNA. It should be noted that Crothers 23 originally derived a neighbor exclusion model 6 years before McGhee and von Hippel, using the statistical mechanics matrix method, but did not offer a convenient closed form equation for use in fitting experimental data. However, starting with Crother's characteristic equations, 23 it is straightforward to obtain an equation identical to Eq. (2) by simple algebraic rearrangement. The two models are therefore equivalent. Equation (2) is commonly used as a fitting function for nonlinear least-squares analysis of ligand-DNA binding isotherms. 13'25 In fact, Eq. (2) is not appropriate for such purposes, because nonlinear fitting by most methods assumes that the independent variable is error free, and that all of the experimental uncertainty resides in the dependent variable. 26 Because the binding ratio r is experimentally determined, it contains error. Worse, the dependent variable r/Cf is a derived quantity so that error in r is propagated into the dependent variable. Nonetheless, nonlinear fitting methods are inevitably used to extract K and n value from experimental data. Two excellent software packages are routinely used in this laboratory for nonlinear curve fitting, FitAll (MTR Software, Toronto, Canada) and Origin (Microcal, Northampton, MA). FitAll now contains a module for Monte Carlo analysis z7 of the error in parameter estimates. Origin contains a module for a rigorous evaluation of parameter error by determination of their upper and lower bounds at any chosen confidence interval.28 Table I shows the results of fits of the data shown in Fig. 2 to the neighbor exclusion model. The binding constant, K = 6.6 x 105 M 1, may be used to calculate AG = - 7.8 kcal mol -I. That value is in excellent agreement with that obtained for the model-independent approach described above. The exclusion parameter, n = 3.3, agrees well with the estimate of the site size obtained by the method of continuous variations. The nonintegral value of the exclusion parameter (n) poses a problem. For neighbor exclusion models, n should strictly be an integer quantity. A fractional value makes no physical sense for a DNA lattice composed of identical, noninteracting sites] Table I shows, however, that if n is constrained to an integer value (either 3 or 4), the standard deviation of the fit degrades significantly. Figure 2 shows the best fit obtained with n = 3. Systematic deviation between the fit and the data are clear, with the data having more curvature than the calculated function. This results in nonrandom residuals, indicating an inadequate fit of the model to the data. 28 The fractional values of n that are required to obtain a statistically 25 j. j. Correia and J. B. Chaires, Methods Enzymol. 240, 593 (1994). 26 M. L. Johnson, Methods Enzymol. 210, 106 (1992). 27 M. Straume and M. L. Johnson, Methods Enzymol. 210, 117 (1992). 28 M. L. Johnson and L. M. Faunt, Methods Enzymol. 210, 1 (1992).

[]]


9

TABLE I NONLINEARLEAST-SQUARESFITS OF DAUNOMYCINBINDINGDATATO NEIGHBOREXCLUSIONMODELSa

Model McGhee-von Hippel n = 3.0 n = 4.0

Friedman-Manning n = 3.0 n ----4.0

K/IOS(M I) 6.6 ± 5.9± 5.4 i 7.2+ 7.0 ± 7.2 ±

0.2 0.1 0.4 0.2 0.2 0.5

N (bp)

Standard deviation

3.3 + 0. l Fixed Fixed 3.1 + 0.1 Fixed Fixed

51,330 56,250 154,600 53,310 53,280 133,300

a McGhee-von Hippel refers to the neighbor exclusion model specified by Eq. (2). Friedman-Manning refers to the model specified by Eq. (3). K is the association constant and n is the neighbor exclusion

parameter expressed in base pairs. The standard deviation is of the best fit.

acceptable fit indicate that the neighbor exclusion model is not, in fact, an appropriate model for the data. Fundamental assumptions of the model must be violated. As is discussed below, the most likely assumption that is violated is that sites are identical, when in fact they are heterogeneous. An extension of the simple McGhee-von Hippel model incorporates an additional term to account for ligand-ligand cooperativity. In principle, integral values of n coupled with negative ligand-ligand cooperativity might adequately fit data such as are shown in Fig. 2. Unfortunately, we previously have shown that attempts to incorporate the added cooperativity parameter are statistically unwarranted and that attempts to force fits to integral values of n are futile. 25 Friedman and Manning derived a variant of the neighbor exclusion model that incorporates aspects of polyelectrolyte theory.29'3° Counterion condensation around DNA is dictated by the spacing of the charged phosphates along the double helix. 31'32 Counterion release coupled to the binding of a charged ligand to DNA provides an energetically favorable contribution to the binding free energy. For DNA intercalation reactions, a complexity arises that the FriedmanManning model addresses. Intercalation results in a separation of the stacked base pairs in duplex DNA, and increased phosphate spacing. This alters the polyelectrolyte properties of the duplex, and results in additional counterion release. As the DNA lattice is saturated with intercalator, there is an ever-changing increase in the 29 R. A. Friedman and G. S. Manning, Biopolymers 23, 2671 (1984). 3o R. A. G. Friedman, G. S. Manning, and M. A. Shahin, in "Chemistry and Physics of DNA-Ligand Interactions" (N. R. Kallenbach, ed.), pp. 37~64. Adenine Press, Schenectady, New York, 1988. 31 M. T. Record, Jr., C. E Anderson, and T. M. Lohman, Q. Rev. Biophys. 11, 103 (1978). 32 G. S. Manning, Q. Rev. Biophys. 11, 179 (1978).


10

[ 1]

average phosphate spacing, resulting in a systematic decrease in the polyelectrolyte contribution to the binding free energy. A closed form equation was derived 29,3° to embody this model for a univalent intercalator binding to B-form DNA in excess univalent salt solutions:

r K(2+r]-(2+~)lO_[c~i~2o2,1~+~](l_nr)[1-nr ],,-1 C---~=

\2-~-rJ

1 -(n --~l)r

(3)

In Eq. (3) K and n have the same meaning as given above, and the added exponential terms describe the decrease in polyelectrolyte contribution to the binding free energy over the course of lattice saturation. The parameter ff0 is the dimensionless charge density parameter and is a function of the structure of duplex DNA. Specifically, ~'0 = qZ/EkTb, where q is the charge of an electron, e is the bulk dielectric constant of the liquid, k is the Boltzmann constant, T is the temperature in degrees kelvin, and b is the charge spacing on the DNA chain. For standard B-form duplex DNA, b = 1.7 A, and ~'o = 4.2. The results of fits of the data shown in Fig. 2 to the Friedman-Manning model are listed in Table I. In all cases examined, the fits are statistically worse than those obtained with the simpler McGhee-von Hippel model. The standard deviations of the fits shown in Table I are all larger for the Friedman-Manning model than for the McGhee-von Hippel model when both K and n are allowed to vary, even though the former contains an additional parameter. The curvature imposed on the fitting function by the added exponential terms in Eq. (3) evidently makes it more difficult to match the curvature in the data. Statistically, therefore, there is no justification for use of the Friedman-Manning model instead of the simpler McGhee-von Hippel neighbor exclusion model. However appealing the underlying theory, the reality of the experimental data provides the ultimate test of the model. In this case, the inclusion of added complexity of polyelectrolyte effects is statistically unwarranted. Use of the neighbor exclusion model has become the standard practice in studies of ligand-DNA interactions. The preceding discussion, however, poses some serious questions about its use. Fractional values for the neighbor exclusion parameters are inevitably required to accurately describe the curvature in experimental data, yet have no meaning in the context of the model because integral values were assumed in the derivation of the model. Fractional neighbor exclusion values signify that the model is not an appropriate description of the actual data. The specific assumption of neighbor exclusion models that is violated is most likely that lattice sites are homogeneous. Chemical and enzymatic footprinting methods have shown unambiguously that such is not the case, and that most ligands bind to DNA sites with a wide distribution of affinities. 33'34 For the specific case of daunomycin, for examples, footprinting studies revealed a strong preference for triplet binding sites with the sequence 5'-(A/T)GC or 5'-(A/T)CG, where the notation

[1]

ANALYSIS OF L I G A N D - - D N A BINDING ISOTHERMS

11

(A/T) means that either A or T can occupy the position. 35'36 Further, sequences containing runs of AT base pairs were revealed by footprinting not to bind daunomycin with appreciable affinity. 16 All DNA lattice sites are clearly not identical, in which case a central tenet of the neighbor exclusion model is violated, vitiating its use as an appropriate model for the analysis of binding isotherms. This conclusion was strongly supported in the single example available, where both macroscopic binding studies and footprinting studies were carried out on the same homogeneous fragment, the 165-bp tyrT DNA fragment. 37 In that study, the macroscopic binding isotherm was complex in shape, could not be fit to the neighbor exclusion model, and clearly revealed a class of high-affinity binding sites whose number was consistent with the number of high-affinity sites visualized by the companion footprinting titration study. The neighbor exclusion principle has also been questioned on other grounds. Rao and Kollman 38 carried out molecular mechanics and molecular dynamics studies from which they concluded that there was no stereochemical basis for neighbor exclusion, at least for the intercalation of 9-aminoacridine into DNA. NMR studies from the Wilson laboratory 39'4° showed conclusively that actinomycin D can bind to adjacent 5'-GpC dinucleotide sites within an oligonucleotide, a finding that contrasts with the 5- or 6-bp exclusion parameter normally associated with the drug. While these are limited and perhaps specialized cases, they do raise questions about the exact physical basis for the neighbor exclusion phenomenon. Dinucleotide Binding Model If the neighbor exclusion model is excluded, what model can be used for the analysis of ligand-DNA binding data? One possibility is to assume heterogeneity of binding sites at the outset. Crothers in fact introduced an early variant of the neighbor exclusion model that incorporated simple base pair selectivity, with an added parameter to account for the relative affinity of GC versus AT base pairs. 23 That simple "two-site" model has not seen wide application. Another, more radical possibility is to abandon the neighbor exclusion concept entirely and to ascribe 33 j. C. Dabrowiak, A. A. Stankus, and J. Goodisman, in "Nucleic Acid Targeted Drug Design C" (L. Propst and T. J. Perun, eds.). Marcel Dekker, New York, 1992. 34 M. J. Waring and C. Bailly, J. Mol. Recognir 7, 109 (1994). 35 j. B. Chaires, K. R. Fox, J. E. Henera, M. Britt, and M. J. Waring, Biochemist~ 26, 8227 (1987). 36 j. B. Chaires, J. E. Herrera, and M. J. Waring, Biochemistry 29, 6145 (1990). 37 C. Bailly, D. Suh, M. J. Waring, and J. B. Chaires, Biochemistry 37, 1033 (1998). 38 S. N. Rao and P. A. Kollman, Proc. Natl. Acad. Sci. U.S.A. 84, 5735 (1987). 39 W. D. Wilson, R. L. Jones, G. Zon, E. V. Scott, D. L. Banville, and L. G. Marzilli, J. Am. Chem. Soc. 108, 7113 (1986). 40 E. V. Scott, R. L. Jones, D. L. Banville, G. Zon, L. G. Marzilli, and W. D. Wilson, Biochemist~ 27, 915 (1988).

12


[ 11

the curvature in Scatchard plots entirely to heterogeneity. We have explored the simplest case in this scenario, a model in which a dinucleotide binding site is assumed. There are 16 possible dinucleotide combinations, 10 of which are unique. These are (5' -+ 3'): AT, AA (---- TT), TA, AC ( = GT), CA ( = TG), GC, GG ( = CC), CG, GA ( = TC), and AG ( = CT). Because intercalators insert between adjacent base pairs and make contact with both, dinucleotide selectivity is not an unreasonable starting point. For binding to dinucleotides (MN), each of which has a unique affinity (KMN), the binding isotherm is described by rD

MN fMNKMNCf

(4)

1 q'- KMNCf

where the binding ratio ro is now expressed as moles of ligand per mole of dinucleotide. The remaining variables are the dinucleotide frequency (fMN) and the free ligand concentration (C0. Dinucleotide frequencies were experimentally determined and tabulated for a numerous natural DNA samples, 41 and may be fixed as constants for a given DNA. Because there are 10 unique dinucleotide steps, the equation has 10 terms, and 10 binding constants must be obtained by nonlinear least-squares fitting of experimental data. Although at first glance the exercise of resolving 10 parameters may seem hopeless or even ludicrous, we show that it can in fact be done with the return of reasonable results. Figure 3 shows the fit of daunomycin-calf thymus DNA binding data to the dinucleotide model. Note that the binding ratio is now expressed in terms of total dinucleotide concentration rather than the usual base pair concentration. The solid line represents the best fit to the dinucleotide binding model. Estimates of 10 binding constants are obtained, and are summarized in Fig. 4. The data in Fig. 4 are average values of KMN estimates obtained from fitting daunomycin binding data obtained with eight different natural DNA samples with known and widely varying dinucleotide frequencies. These samples ranged from Clostridium perfringens DNA (31% GC content) to Micrococcus lysodeikticus DNA (72% GC content). Unique KMN values are returned that range over two orders of magnitude. Figure 4 shows that AA, AT, and TG steps represent low-affinity sites. TA, GC, and CG have intermediate affinity. High-affinity sites are TC, AC, AG, and GG steps. Does this analysis make sense? Is it valid? Several observations suggest that the answer is "yes" to both questions. First, footprinting studies of the daunomycinDNA interaction showed that sequences protected from DNase I cleavage by the bound drug were enriched in the dinucleotides AC, AG, GG, GC, and CG relative to the tyrT DNA fragment alone that was used for footprinting. 35 These are among the very dinucleotide steps with the highest KMN values. In contrast, analysis of unprotected cleavage sites from the footprinting experiments showed that such 41 G. D. Fasman, "Handbook of Biochemistry and Molecular Biology," 3rd Ed. CRC Press, Cleveland, Ohio, 1976.

[l ]

ANALYSIS OF LIGAND-DNA BINDINGISOTHERMS

13

0,75 ]

0.50

0.00

t _.t.._

1E-9

1E-8

1E-7

1E-6

1E-5

1E-4

Cf, M FIG. 3. Binding isotherm for the daunomycin-calf thymus DNA interaction. The binding ratio on the ordinate is expressed as moles of daunomycin bound per dinucleotide. Free daunomycin concentration is shown on the abscissa. The solid line represents the best fit to the 10-site, dinucleotide binding model described in text.

sites were enriched in the dinucleotides AA, AT, and TA. 16'35 Two of these steps have the lowest measured KMN. Second, KMN values can be used to calculate a total binding free energy for the loading of the calf thymus DNA lattice that is in excellent agreement with the model-independent approach described earlier in this chapter. The total free energy (AGT) is specified by MN

MN

AGT = ~ fMNAGMN = Z fMN(-RT In KMN)

(5)

where A G M N is the free energy for binding to the dinucleotide step MN. Using the known fMN values for calf thymus DNA and the KMN values shown in Fig. 4, a value of AGT = --7.8 kcal mol I is obtained, in excellent agreement with that obtained by model-independent analysis. Finally, the binding constants shown in Fig. 4 may be used in combination with the known dinucleotide frequency of calf thymus DNA to simulate a binding isotherm, as shown in Fig. 5. If these simulated data are then fit to the neighbor exclusion model, values o f K = 6.5 x 105 M - I and n -- 3.3 bp are obtained. These values are in excellent agreement with those obtained by fits to actual experimental data (Fig. 2). The key point of this exercise is

14

BIOPHYSICALAPPROACHES 10 7

I

I

!

108

I

I

I

I

[ 1] I

I

l~,//'~~i,X~ i

10 s

104

t AA

AT

TA

t

I

t

t

T G T C AC AG GG GC C G

Dinucleotide Step FIG. 4. Affinity profile for the interaction of daunomycin with the 10 unique dinucleotide steps. The binding constants (K) are average values obtained by fitting of binding isotherms for the interaction of daunorubicin with eight different DNA samples with widely varying dinucleotide frequencies. The error bars show standard deviations from the mean values.

that dinucleotide heterogeneity can account entirely for the curvature in Scatchard plots at least as well as the neighbor exclusion model does. The model presented in this section is somewhat radical and vitiates the conventional wisdom of the neighbor exclusion model. But because a fundamental assumption of the neighbor exclusion model (the homogeneity of potential binding sites) is demonstrably incorrect, the development of other models is mandatory. Furthermore, a few studies have questioned the physical basis of neighbor exclusion. 38-4° The dinucleotide model is consistent with the results of more than a decade of footprinting studies, which show that ligands inevitably bind with a wide range of affinities along the DNA lattice. In one sense, we have come full circle. Early binding isotherms for proflavin42 were curved and were interpreted in term of site heterogeneity, namely with two classes of sites later attributed to intercalation and "outside" binding. The dinucleotide model assumes a different kind

42 A. R. Peackocke and J. N. H. Skerren, Trans. Faraday Soc. 52, 261 (1956).

[1]

ANALYSIS OF L I G A N D - D N A BINDING ISOTHERMS I

I

I

I

~

15

I

5 x 1 0 5.

4xl 0 5

3x10 5

0 2x 10 5

lx10 5

I

I

I

I

I

0,05

0.10

0.15

0.20

0.25

r bound

0.30 ,-,

FIG. 5. Simulated Scatchard plot for the daunomycin-calf thymus DNA interaction. Dinucleotide binding constants (Fig. 4) were used in combination with the known dinucleotide frequency for calf thymus DNA to simulate binding data (solid circles), which were then cast into the form o f a Scatchard plot. If these simulated data are then fit to the neighbor exclusion model [Eq. (2)], the best fit (solid line) yields K -- 6.5 × 105M -1 and n = 3.3 bp.

of heterogeneity involving a larger number of sites derived from a fundamental property of any DNA, its nearest neighbor dinucleotide frequency.

Coping with Cooperativity In some cases, ligand-DNA binding isotherms exhibit evidence of positive cooperativity. In these cases, data cast into the form of a Scatchard plot show positive slopes at low binding ratios (Figs. 6 and 7). Analysis and interpretation of such isotherms become even more complicated. McGhee and yon Hippe124 (and

16

BIOPHYSICALAPPROACHES 4

I

I

I

[ 11 I

I

!3 v

2

+

%÷÷**~+÷%Q~. •

- -

4~m J.

~ $$ m

I 0

I

I

I

I

0.1

0.2

0.3

0.4

r-bound FIG. 6. Allosteric binding of daunomycin to poly(dA).poly(dT). Binding data were obtained for the interaction of daunomycin with poly(dA).poly(dT) in buffered 0.2 M NaCI solutions (pH 7.0). The solid line was calculated for the parameters listed in Table II for the Crothers allosteric binding model. The -I- + + line shows the best fit to the McGhee-von Hippel Neighbor exclusion model with added ligand-ligand cooperativity [Eq. (6)].

Crothers earlier 23) presented an extended form of the neighbor exclusion model that included an additional term to account for ligand-ligand interactions. The equation for this model is

r C--f = R =

K(l_nr)[(2w-1)(l-nr)+r-R]n-l[l-(n+l)r+R]2 Y(~--- U d - n ~ 2(1 - nr) {[1 - (n + l)r] 2 + 4mr(1 - nr)} 1/2

(6)

where K, n, and r have the same meaning as given above, and co is the cooperativity parameter, co is defined as the equilibrium constant for moving two ligands bound at isolated sites into proximity, such that they occupy contiguous lattice sites. If ~o > 1.0, positive cooperativity results and ligands bind preferentially next to one another. If ~o < 1.0, negative cooperativity results, and ligand binding at adjacent sites is hampered. A distinctive feature of this model is that the DNA lattice remains and inert array of identical sites. All cooperativity results from ligandligand interactions of an unspecified nature. A contrasting model that can account for positive cooperativity is the Crothers allosteric model. 43 The underlying concept is radically different from the McGhee-von Hippel model. The allosteric model is analogous to the classic Monod-Wyman-Changeux model for allostery 44 derived to explain the cooperative 43 N. Dattagupta, M. Hogan, and D. M. Crothers, Biochemistry 19, 5998 (1980). 44 j. Monod, J. Wyman, and J.-P. Changeux, Z Mol. Biol. 12, 88 (1965).

[ l]

ANALYSIS OF LIGAND--DNA BINDING ISOTHERMS I

I

I

I

17

I

1.2x10 s

1.0xlO s

Z Form

8.0x104

"')

•

B Form

6.0xl 0 4

4.0x10 4

2.0xl 0 4

0.0 I

0.0

~

I

0.1

~

I

~

0.2

I

0.3

~

I

0.4

0.5

r

FIG. 7. Allosteric binding of daunomycin to left-handed Z-DNA. Binding data were obtained for the interactions of daunomycin with Z-form [poly(dG-dC)]2in buffered 3.0 M NaCI solutions (pH 7.0). The solid line was calculated for the parameters listed in Table II for the Crothers allosteric binding model. At the start of the binding isotherm, the polymer is in the left-handed Z form. Beyond the maximum near r = 0.3, daunomycin binding has allosterically converted the polymer to the right-hand form.

b i n d i n g of ligands to proteins, such as the b i n d i n g of o x y g e n to hemoglobin. Crothers built on the basic concept of allostery, but included specific details of the m a c r o m o l e c u l a r conformational transition and ligand b i n d i n g that were appropriate for a D N A lattice. The allosteric model assumes that the D N A lattice can exist in one of two conformational forms (1 and 2). The transition of the lattice between these forms proceeds by a nucleation step, followed by propagation steps: O-2S

Nucleation:

...111111...

< ~ ... 112111...

Propagation:...112111...

< ~ ... 112211...

S

The equilibrium constant for nucleation is cr2s and that for propagation is s. The allosteric model then assumes that ligand can bind to each conformational form

18


[1]

with unique neighbor exclusion binding parameters, (K1, n l, o~1) and ( K 2 , t/2,092). Cooperativity arises when the ligand binds selectively or preferentially to one of the conformational forms. Binding then drives an allosteric conformational transition to the form with higher ligand affinity. Eight parameters are needed to compute binding isotherms. There is no closed form, analytic equation for the allosteric model. Crothers et al. wrote a Fortran program that calculates binding isotherms according to the derived statistical mechanical model, 43 a program that has seen modest circulation 45 47 and that has been modified on occasion 4s'49 for easier use. No nonlinear fitting routine has yet been developed that incorporates the allosteric model, and users must optimize binding parameters by successive approximation along with judicious constraint of those selected parameters that can be estimated by independent methods. 5° Nonetheless, use of the allosteric model is unavoidable in some cases, as examples will show. It is important to contrast the key features of the McGhee-von Hippel and allosteric models. In the McGhee-von Hippel model, positive cooperativity arises from ligand-ligand interactions while the DNA lattice remains in a single conformation. For protein binding to DNA, such ligand-ligand interactions might be visualized as protein-protein contacts formed when adjacent lattice sites are occupied. For small molecule ligands, although such interactions could also occur, it is less easy to ascribe a molecular picture to the process. In contrast, positive cooperativity arises in the Crothers allosteric model from an underlying conformational transition in the DNA lattice to a form with higher ligand binding affinity. The allosteric model was perhaps the first clear statement of the possibility of structural selectivity in ligand-DNA interactions. Two examples will illustrate positive cooperativity in ligand-DNA interactions. Figure 6 shows the binding of daunomycin to poly(dA) • poly(dT). Independent studies have shown that poly(dA), poly(dT) undergoes a premelting transition between two helical forms. 51-53 The binding data in Fig. 6 show a positive slope at low r values, and pass through a maximum near r = 0.05. Attempts to fit these data to the extended McGhee-von Hippel model yield unsatisfactory results. The best fit to that model (with K = 1.4 × 104M -j , n = 2.3 bp, ~o = 2.4) is shown by the 45 j. B. Chaires, J. Biol. Chem. 261, 8899 (1986). 46 G. Y. Walker, M. E Stone, and T. R. Krugh, Biochemistry 24, 7462 (1985). 47 G. T. Walker, M. E Stone, and T. R. Krugh, Biochemistry 24, 7471 (1985). 48 G. T. Walker, Ph.D. Thesis, University of Rochester, 1986. 49 D. Snh, Ph.D. Thesis, University of Mississippi Medical Center, 1993. 5o j. B. Chaires, J. Biol. Chem. 261, 8899 (1986). 51 j. E. Herrera and J. B. Chaires, Biochemistry 28, 1993 (1989). 52 S. S. Chan, K. J. Breslauer, M. E. Hogan, D. J. Kessler, R. H. Austin, J. Ojemann, J. M. Passner, and N. C. Wiles, Biochemistry 29, 6161 (1990). 53 S. S. Chan, K. J. Breslauer, R. H. Austin, and M. E. Hogan, Biochemistry 32, 11776 (1993).

[1]


19

TABLE II PARAMETER ESTIMATES FOR ALLOSTERICMODEL "FIT" TO DAUNOMYCINBINDING TO POLY(dA).POLY(dT) AND Z-DNA Parameter o" s

KI(M -I ) K2(M -I ) K2/KI ?11 n2

Description

Poly (dA).Poly(dT) a

Z-DNA ~'

Nucleation parameter Propagation constant Association constant for binding to form 1 Association constant for binding to form 2 Ratio of association constants Neighbor exclusion parameter (form 1) Neighbor exclusion parameter (form 2)

0.01 0.985 4,800 21,000 4.3 2.0 2.0

0.00 l 0.635 8,000 350,000 44 2.0 2.3

a Daunomycin binding to poly(dA).poly(dT) in buffered 0.2 M NaCI solutions (pH 7.0). Parameters result from analysis of the binding data shown in Fig. 7. b Daunomycin binding to Z-form [poly(dG-dC)]2 in buffered 3.0 M NaCI solutions (pH 7.0). Parameters result from analysis of the binding data shown in Fig. 6.

curve with the (+) symbols; no value of ~o can be found that can produce a curve that matches the steep positive slope at the start of the isotherms. The model, with an assumed inert DNA lattice, cannot match the experimental data. In contrast, the allosteric model can match the shape of the binding data well, as shown by the solid curve in Fig. 6. The best estimates of the parameters for the allosteric model that describe the binding data are collected in Table II. The key driving force for the allosteric conversion of poly(dA), poly(dT) is the >4-fold preference for binding to form 2 of the polynucleotide. A variety of additional physical and enzymatic tools was used to demonstrate that daunomycin binding was indeed coupled to a conformational change in poly(dA), poly(dT) 51 the essential feature of the allosteric model. Figure 7 shows daunomycin binding to poly(dG-dC) under solution conditions that initially favor the left-handed Z conformation of the polynucleotide. 45'54 The allosteric model with the parameters listed in Table II was used to obtain the solid curve matching the experimental data. In this case, the underlying conformational transition to which binding is coupled is the Z-to-B conversion. Compared with the first example of binding to poly(dA) • poly(dT), this system is much more cooperative. The reason for that is the greater preference of daunomycin for the right-hand form relative to the left-handed form (with K z / K I -~ 44; Table II), compared with its preference for the two right-handed helical forms of poly(dA) • poly(dT) (with K2/KI = 4; Table II).

54 X. Qu, J. O. Trent, I. Fokt, W. Priebe, and J. B. Chaires, Proc. Natl. Acad. Sci. U.S.A. 97, 12032

(2O0O).

20


[ 1]

L i g a n d B i n d i n g to O l i g o n u c l e o t i d e s Advances in synthetic methods have made it relatively easy to prepare DNA or RNA oligonucleotides of precisely defined length and sequence, and it is now fashionable to use such molecules for ligand binding studies. The significant advantage of oligonucleotides is their homogeneity. There are, however, disadvantages. First, end effects may become a consideration in oligonucleotide studies. Neighbor exclusion models typically assume an "infinite lattice" specifically to avoid end effects, and therefore such models become inapplicable to oligonucleotide systems. Polyelectrolyte theory and experiment have shown that significant end effects exist for oligonucleotides less than about 24 bp in length. 55'56 Second, problems of appropriate representation of all possible sequence elements may arise in oligonucleotide systems. There are 10 unique dinucleotide combinations, as discussed above, and it is rare that a given oligonucleotide is appropriately designed to contain all possible dinucleotide steps. It is always possible that a high-affinity interaction may go undetected because the appropriate site is absent in the oligonucleotide chosen for study. Binding constants may thus be biased by the choice of sequence unless a large number of oligonucleotides is studied. For trinucleotides, the situation becomes even worse, because there are 64 possible triplets, 32 of which are unique. These concerns should not be taken as a call to abrogate oligonucleotide studies, but rather to acknowledge their appropriate role. Complete binding studies should be comprehensive and should systematically move from long natural DNA samples through synthetic polynucleotides of defined simple repeating sequences to oligonucleotides of precisely defined sequence. Oligonucleotide systems are of particular value for the study of the energetics of binding when other experiments have defined the sequence of the preferred binding site for a given ligand. Neighbor exclusion models are generally inapplicable for the analysis of ligand binding to oligonucleotides because they are based on an "infinite lattice" assumption. Binding isotherms in these cases are best analyzed by the classic and simple stoichiometric binding models that are described in any number of texts and monographs. 2-4 Figure 8 shows an example of daunomycin binding to a 24-bp duplex oligonucleotide with the sequence (5'-TGCATGCATGCATGCATGCATGCA)2. This oligonucleotide was designed and synthesized to contain a repetitive motif containing the preferred daunomycin binding site that emerged from footprinting studies [5'-(A/T)GC; where (A/T) means A or T]. Binding data were fit to the simple expression for multiple identical, noninteracting sites: r

~tK Cf -

-

-

-

1 +KCf

(7)

55M. C. Olmsted,C. E Anderson,and M. T. Record,Jr., Proc. Natl. Acad. Sci. U.S.A. 86, 7766(1989), 56M. C. Olmsted,C. E Anderson,and M. T. Record,Jr., Biopolymers 31, 1593(1991).

[ 1]


I

I

!

I

2

4

6

8

21

(D "O

"5

EI1)

cO

O to

E3

0

0

10

Ct~ee,gM FIG. 8. Binding of daunomycin to a 24-bp duplex oligonucleotide. Data (solid circles) are presented as a direct plot, with the best fit to Eq. (7) shown as the solid line.

where r is now expressed as moles of daunomycin bound per mole of oligonucleotide, n is the number of sites, and Kis the association constant. Nonlinear leastsquares fitting of the data yields K = 4.0 (4-0.1) × 106M - 1 and n = 4.3 (-t-0.3). Evidently only four molecules of daunomycin are binding to each oligonucleotide, even though there are six potential sites that contain the preferred triplet sequence. It is possible that the sites near the ends are disfavored, or that there is anticooperativity that disfavors binding to adjacent sites. In the latter case, more complicated models with added parameters would need to be invoked and implemented. An excellent example of such an analysis applied to daunomycin binding to hexanucleotide sequences was provided by Rizzo and co-workers. 57 Summary Binding studies provide information of fundamental and central importance for the complete understanding of ligand-DNA interactions. Studies of ligand binding to long natural DNA samples, to synthetic deoxypolynucleotides of simple repeating sequence, and to oligonucleotides of defined sequence are all needed to 57 V. Rizzo, C. Battistini, A. Vigevani, N. Sacchi, G. Razzano, F. Arcamone, A. Garbesi, E P. Colonna, M. Capobianco, and L. Tondelli, J. Mol. Recognit. 2, 132 (1989).

22


[21

begin to understand the interaction in detail. Binding studies provide entry into the thermodynamics of the DNA interactions, which in turn provides great insight into the molecular forces that drive the binding process. This chapter summarizes both model-dependent and -independent approaches for the analysis and interpretation of binding isotherms, and should serve as a concise guide for handling experimental data. Acknowledgment W o r k in the a u t h o r ' s l a b o r a t o r y w a s f u n d e d b y g r a n t C A 3 5 6 3 5 f r o m the N a t i o n a l C a n c e r Institute.

[2] S u r f a c e P l a s m o n R e s o n a n c e of RNA-Small Molecule

Biosensor Analysis Interactions

By TINA M. DAVIS and W. DAVID WILSON Introduction Because some of the most devastating human diseases are caused by RNA viruses, such as human immunodeficiency virus (HIV) and hemorrhagic fever viruses such as dengue and Ebola, l RNA is an attractive therapeutic target in drug design. 2-4 The unique structural folds present in RNA, but not DNA, offer the possibility of much higher recognition specificity by small molecules. Structured sites in RNA can be targeted with the same specificity with small molecules as structured binding regions of proteins. Despite the potential advantages of targeting RNA, few drugs are known that target RNA, and rational design of drugs that target RNA is in the beginning stages.5-SA summary of some of the RNAs that have been successfully targeted in drug design and interaction studies is shown in Table 1.2,9-26 I M. B. A. Oldstone, "Viurses, Plagues, and History." Oxford University Press, Oxford, 1998. 2 W. D. Wilson and K. Li, Curr. Med. Chem. 7, 73 (2000). 3 K. Michael and Y. Tor, Chem. Eur. J. 4, 2091 (1998). 4 F. Walter, Q. Vicens, and E. Westhof, Curr. Opin. Chem. Biol. 3, 694 (1999). 5 K. Li, M. Fernandez-Saiz, C. T. Rigl, A. Kumar, K. G. Ragunathan, A. W. McConnaughie, D. W. Boykin, H. J. Schneider, and W. D. Wilson, Bioorg. Med. Chem. 5, 1157 (1997). 6 T. Hermann and W. Westhof, Curt Opinion Biotechnol. 8, 278 (1998). 7 C. Chow and F. M. Bogdan, Chem. Rev. 97, 1489 (1997). 8 M. Afshar, C. D. Prescott, and G. Varani, Curr. Opin. BiotechnoL 10, 59 (1999). 9 M. J. Rogers, Y. V. Bukhman, R. E McCutchan, and D. E. Draper, RNA 3, 815 (1997). 10 G. L. Conn, R. R. Gutell, and D. E. Draper, Biochemistry 37, 11980 (1998).


Copyright © 2001 by Academic Press All rights of reproduction in any form reserved. 0076-6879f00 $35.(X)

22


[21

begin to understand the interaction in detail. Binding studies provide entry into the thermodynamics of the DNA interactions, which in turn provides great insight into the molecular forces that drive the binding process. This chapter summarizes both model-dependent and -independent approaches for the analysis and interpretation of binding isotherms, and should serve as a concise guide for handling experimental data. Acknowledgment W o r k in the a u t h o r ' s l a b o r a t o r y w a s f u n d e d b y g r a n t C A 3 5 6 3 5 f r o m the N a t i o n a l C a n c e r Institute.

[2] S u r f a c e P l a s m o n R e s o n a n c e of RNA-Small Molecule

Biosensor Analysis Interactions

By TINA M. DAVIS and W. DAVID WILSON Introduction Because some of the most devastating human diseases are caused by RNA viruses, such as human immunodeficiency virus (HIV) and hemorrhagic fever viruses such as dengue and Ebola, l RNA is an attractive therapeutic target in drug design. 2-4 The unique structural folds present in RNA, but not DNA, offer the possibility of much higher recognition specificity by small molecules. Structured sites in RNA can be targeted with the same specificity with small molecules as structured binding regions of proteins. Despite the potential advantages of targeting RNA, few drugs are known that target RNA, and rational design of drugs that target RNA is in the beginning stages.5-SA summary of some of the RNAs that have been successfully targeted in drug design and interaction studies is shown in Table 1.2,9-26 I M. B. A. Oldstone, "Viurses, Plagues, and History." Oxford University Press, Oxford, 1998. 2 W. D. Wilson and K. Li, Curr. Med. Chem. 7, 73 (2000). 3 K. Michael and Y. Tor, Chem. Eur. J. 4, 2091 (1998). 4 F. Walter, Q. Vicens, and E. Westhof, Curr. Opin. Chem. Biol. 3, 694 (1999). 5 K. Li, M. Fernandez-Saiz, C. T. Rigl, A. Kumar, K. G. Ragunathan, A. W. McConnaughie, D. W. Boykin, H. J. Schneider, and W. D. Wilson, Bioorg. Med. Chem. 5, 1157 (1997). 6 T. Hermann and W. Westhof, Curt Opinion Biotechnol. 8, 278 (1998). 7 C. Chow and F. M. Bogdan, Chem. Rev. 97, 1489 (1997). 8 M. Afshar, C. D. Prescott, and G. Varani, Curr. Opin. BiotechnoL 10, 59 (1999). 9 M. J. Rogers, Y. V. Bukhman, R. E McCutchan, and D. E. Draper, RNA 3, 815 (1997). 10 G. L. Conn, R. R. Gutell, and D. E. Draper, Biochemistry 37, 11980 (1998).


Copyright © 2001 by Academic Press All rights of reproduction in any form reserved. 0076-6879f00 $35.(X)

[2]

R N A INTERACTIONS

23

TABLE I SMALL MOLECULESAND THEIR BIOLOGICALRNA TARGETS Sequence and secondary structure

RNA

Small molecule ligand

Ref.

uAA A rRNA OTPase

AOCAGccAUcAuU AGAuUCGGuuG UA,~,

A# AGcGUA cC U CGAuA

Thiostrepton

9, 10

Aminoglycosides

11-13

A U 5'-GCC GGU-3' rRNA A site

AC CC -CCAC UGC GC-5' -GGUG UCG # A - 3 '

AAa U~

tRNA

tRNA phe

Hoechst 33258 Bleomycin

14, 15, 16

Ribozymes

Hepatitis delta virus ribozyme Hammerhead RNA Group I intron

Aminoglycosides

17, 18, 19

HIV- 1 RRE

AAcuGCGAC GC CG GUA-5' uGACGCUCACGGuACAG-3'

Aminoglycosides Diphenylfuran polycations

20 21, 22

HIV-I TAR

GUC~GA'UCu" ' t tJ 5A GA-5 G G AGCUC --,- UCU-3'

Aminoglycosides Aminoquinozalines Acridine polycations

23 24 25

ACc G C G c c G C c c CC-5' CUCCUGUGG-GGUC_3' G U

Aminoglycosides

26

Thymidylatesynthase mRNA

RNA-ligand interactions can be investigated by the same techniques as DNAligand interactions. Nuclear magnetic resonance (NMR) 27-33 and X-ray crystallography34 can be used to obtain detailed structural information about RNA-ligand complexes, although obtaining the quantity of RNA necessary for such studies is Jl D. Fourmy, M. I. Recht, S. C. Blanchard, and J. D. Puglisi, Science 274, 1367 (1996). 12 D. Fourmy, S. Yoshizawa, and J. D. Puglisi, J. Mol. Biol. 277, 333 (1998). 13 M. I. Recht, D. Fourmy, S. C. Blanchard, K. D. Dahlquist, and J. D. Puglisi, J. Mol. BioL 262, 421 (1996). 14E. g. Bichenkova, S. E. Sadat-Ebrahimi, A. N. Wilton, N. O'Toole, D. S. Marks, and K. T. Douglas, Nucleosides Nucleotides 17, 1651 (1998). 15 S. E. S. Ebrahimi, A. N. Wilton, and K. T. Douglas, Chem. Commun. 4, 385 (1997). 16C. E. Holmes, R. J. Duff, G. A. van der Marel, J. van Boom, and S. M. Hecht, Bioorg. Med. Chem. 5, 1235 0997). 17 U. yon Ahsen, J. Davies, and R. Schroeder, Nature (London) 353, 368 (1991 ). I8 j. Rogers, A. H. Chang, U. yon Ahsen, and R. Schroeder, ,l. Mol. Biol. 259, 916 (1996).

24


[2]

often much more costly than it is for DNA. In addition, making NMR resonance assignments for RNA is often much more difficult than it is for DNA because of decreased spectral dispersion in the NMR spectra of RNA. Chemical and nuclease footprinting can provide information regarding drug-binding site(s) and sequence specificity of binding. 32'33'35-38 Theoretical studies of RNA complexes have reached a high level of sophistication and can now independently provide important structural and dynamic information. 4 Melting experiments by spectroscopic methods such as UV and circular dichroism (CD) can provide information regarding the thermodynamics of drug binding.5,38 43 CD has the added ability to provide information regarding the

19 T. Hermann and E. Westhof, J. Mol. Biol. 276, 903 (1998). 2o M. L. Zapp, S. Stern, and M. R. Green, Cell 74, 969 (1993). 21 K. Li, G. Xiao, T. Rigl, A. Kumar, D. W. Boykin, and W. D. Wilson, in "Structure, Motion, Interaction and Expression of Biological Macromolecules: Proceedings of the 10th Conversation in the Discipline Biomoleculor Stereodynamics," University of Albany, New York (R. H. Sarma and M. H. Sarma, eds.), pp. 137-145. Adenine Press, Schenectady, New York, 1998. 22 L. Ratmeyer, M. L. Zapp, M. R. Green, R. Vinayak, A. Kumar, D. W. Boykin, and W. D. Wilson, Biochemistry 35, 13689 (1996). 23 S. Wang, P. W. Huber, M. Cui, A. W. Czaruik, and H.-Y. Mei, Biochemistry 37, 5549 (1998). 24 H.-Y. Mei, D. P. Mack, A. A. Galan, N. S. Halim, A. Heldsinger, J. A. Loo, D. W. Moreland, K. A. Sannes-Lowery, L. Sharmeen, and A. W. Czarnik, Bioorg. Med. Chem. 5, 1173 (1997). 25 E Hamy, V. Brondani, A. Florsheimer, W. Stark, M. J. J. Blommers, and T. Klimkait, Biochemisttiy 37, 5085 (1998). 26 A. R. Ferre-D'Amare, K. Zhou, and J. A. Doudna, Nature (London) 395, 567 (1998). 27 M. Katahira, S. Kobayashi, A. Matsugami, K. Ouhashi, S. Uesugi, R. Yamamoto, K. Taira, S. Nishikawa, and P. Kumar, Nucleic Acids Symp. Sel: 42, 269 (1999). 28 W. H. Gmeiner, Curt Med. Chem. 5, 115 (1998). 29 S. Yoshizawa, D. Founny, and J. D. Puglisi, EMBO J. 17, 6437 (1998). 30 Q. Chen, R. H. Shafer, and 1. D. Kuntz, Biochemistry 36, 11402 (1997). 31 L. Jiang, A. K. Suri, R. Fiala, and D. J. Patel, Chem. Biol. 4, 35 (1997). 32 E Hamy, V. Brondani, A. Florsheimer, W. Stark, M. J. Blommers, and T. Klimkait, Biochemistry 37, 5086 (1998). 33 M. I. Recht, D. Fourmy, S. C. Blanchard, K. D. Dahlquist, and J. D. Puglisi, J. Mol. Biol. 262, 421 (1996). 34 E Takusagawa, K. T. Takusagawa, R. G. Carlson, and R. F. Weaver, Bioorg. Med. Chem. 5, 1197 (1997). 35 N. Gelus, C. Bailly, E Hamy, T. Klimkait, W. D. Wilson, and D. W. Boykin, Biool2~. Med. Chem. 7, 1089 (1999). 36 N. Gelus, E Hamy, and C. Bailly, Bioorg. Med. Chem. 7, 1075 (1999). 37 G. Rosendahl and S. Douthwaite, Nucleic Acids Res. 22, 357 (1994). 38 L. Dassonneville, E Hamy, P. Colson, C. Houssier, and C. Bailly, Nucleic Acids Res. 25, 4487 (1997). 39 j. E. Draper and T. C. Gluick, Methods Enzymol. 259, 281 (1995). 4o D. S. Pilch, M. A. Kirolos, X. Liu, G. E. Plum, and K. J. Breslauer, Biochemistry 34, 9962 (1995). 41 U. Sehlstedt, P. Aich, J. Bergman, H. Vallberg, B. Norden, and A. Graslund, J. Mol. Biol. 278, 31 (1998).

[21

RNA INTERACTIONS

25

mode of drug binding to RNA (groove binding versus intercalation). Fluorescence experiments are useful for obtaining binding information provided the small molecule has fluorescent properties that are altered on binding to RNA. 23 Calorimetric techniques such as differential scanning calorimetry (DSC) and isothermal calorimetry (ITC) provide direct quantitative information regarding the thermodynamics of drug binding and are perhaps the most reliable methods for obtaining such information. 4°-42 Calorimetric methods, however, usually require substantially more material than spectrophotometric methods, and they have difficulty measuring large binding constants. Gel band shift has been the most widely used method for studying nucleic acid-ligand interactions. 5,35 Providing that the RNA-small molecule complex can be separated from free RNA (i.e., there is a detectable band shift from drug binding) and that the binding kinetics are relatively slow, the technique can provide accurate binding constants and stoichiometry with only minimal quantities of RNA and drug. The technique is particularly useful for screening focused combinatorial libraries that meet the above-described criteria. Many RNA-small molecule complexes are not easily separated on gels, however, or have kinetics that are too fast for resolution on gels. It is clear that additional methods are required for such complexes. Several newer techniques for studying nucleic acid-small molecule interactions have been described. Ibis Therapeutics (Carlsbad, CA) a division of Isis Pharmaceuticals, has developed a sophisticated mass spectrometry-based assay for screening mixtures of small molecules for binding to RNA. 44-46 The technique permits multiple ligands and multiple RNAs to be screened simultaneously. In addition, the technique can provide binding affinities and binding sites on RNA. This is an excellent method for analysis of multiple compounds when the appropriate equipment is available. Luedtke and Tor 47 have developed a novel solid-phase assay for investigating RNA-small molecule interactions. The assay was used to identify small molecules that bind to the Rev-responsive element (RRE) fragment of the HIV virus and was based on the competition between potential RNA binders and a fluorescent Rev peptide. A biotinylated RRE fragment was immobilized on insoluble agarose beads that were covalently modified with streptavidin. A ternary complex was formed 42 Z. Xu, D. S. Pilch, A. R. Srinivasan, W. K. Olson, N. E. Geacintov, and K. J. Breslauer, Bioorg. Med. Chem. 5, 1137 (1997). 43 G. Luck, C. Zimmer, and B. C. Baguley,Biochim. Biophys. Acta 782, 41 (1984).

44 S. A. Hofstadler, K. A. Sannes-Lowery,S. T. Crooke, D. J. Ecker, H. Sasmor, S. Manalili, and R. H. Griffey, Anal. Chem. 71, 3436 (1999). 45 R. H. Griffey, S. A. Hofstadler, K. A. Sannes-Lowery,D. J. Ecker, and S. T. Crooke, Proc. Natl. Acad. Sci. U.S.A. 96, 10129 (1999). 46 K. Hamasaki and R. R. Rando, Anal. Biochem. 261, 183 (1998). 47 N. W. Luedtke and Y. Tor, Angew. Chem. 39, 1788 (2000).

26


[2]

on addition of fluorescein-labeled Rev peptide. Ligand binding was indirectly detected by monitoring the quantity of fluorescent peptide either released into solution or remaining on the solid support as ligand bound to the RNA. The assay offers a simple, rapid, and quantitative method that facilitates the discovery and characterization of RNA binders directed at a specific complex. Surface plasmon resonance (SPR) biosensors are another relatively new technique for characterizing nucleic acid-small molecule interactions. Although SPR biosensors have been commercially available for more than a decade, their application to nucleic acid-small molecule systems is just developing. Only a handful of SPR articles describing macromolecule-small molecule interactions have been published to date, and only a few of these focus on nucleic acid-small molecule systems.48 51 As with the methods described above, SPR offers a rapid and powerful tool for screening small molecule libraries. 49'52 An advantage of SPR is that the method simultaneously can provide kinetic and equilibrium characterization of the interactions of active compounds with macromolecules. 53 Such information gives important insight for understanding whether inhibitory activity is governed by the kinetic or equilibrium (or both) properties of the interaction. For example, Hendrix e t al. found that whereas two similar aminoglycoside antibiotics have significantly different inhibitory potency, their binding constants are rather similar9 The dissociation rates for the two molecules, however, are substantially different, accounting for the more than 100-fold difference in inhibitory activity. When one of the interacting species can be captured on a biosensor surface, the advantages of SPR over conventional interaction analyses can be numerous. SPR monitors molecular interactions in real time 53 and is a significant improvement over optical methods for systems involving strong binding and/or low absorbance or fluorescence. In addition, material requirements are minimal when compared with spectrophotometric techniques, generally requiring only picomole to nanomole quantities, while simultaneously providing accurate and reproducible data. Although excellent articles on nucleic acid-small molecules have begun to appear, to our knowledge, a comprehensive guide for investigating nucleic acid-small molecule systems through SPR has not been published to date. For this reason and because of the commercial availability of high-sensitivity SPR instruments that are amenable for studying these systems, we have focused this review on how to study RNA-small molecule interactions with SPR biosensors. 48 C.-H. Wong,M. Hendrix, W. S. Priestley, and W. A. Greenberg, Chem. Biol. 5, 397 (1998). 49C.-H. Wong,M. Hendrix, D. D. Manning, C. Rosenbohm,and W. A. Greenberg,J. Am. Chem. Soc. 120, 8319 (1998). 50M. Hendrix, E. S. Priestley, G. E Joyce, and C.-H. Wong,J. Am. Chem. Soc. 119, 3641 (1997). 51 L. Wang, C. Bailly, A. Kumar, D. Ding, M. Bajic, D. W. Boykin, and W. D. Wilson, Proc. Natl. Acad. Sci. U.S.A. 97, 12 (2000). 52 P.-O. Markgren, M. Hamalained, and U. H. Danielson,Anal. Biochem. 265, 340 (1998). 53 Biacore, "BIACoreBIAapplicationsHandbook."Biacore, Uppsala, Sweden, June 1994.

[2]

RNA INTERACTIONS

Description of Surface Plasmon Resonance

27 Biosensors

SPR biosensors employ surface plasmon resonance to qualitatively and/or quantitatively describe molecular interactions. Although several companies now offer SPR biosensors, 54 currently, instruments from BIAcore (Uppsala, Sweden) are the most widely used. The BIAcore 2000 biosensor is used in our laboratory, and the experimental guidelines and technical details presented in this review are specific to this biosensor and the BIAcore 3000 biosensor. However, the basic principles of using SPR to study molecular interactions are universal. In biosensor experiments, the interaction between molecules in solution (the analyte in BIAcore literature) and molecules immobilized on a sensor chip surface (the ligand in BIAcore literature) is monitored in real time. ~3 Labeling is not required and there are many immobilization chemistries available depending on experimental designY Biosensor experiments involve immobilizing one of the interacting species on a sensor chip surface either covalently or through affinity capture (such as biotinstreptavidin or antibody-antigen) to produce a biospecific surface. 55 Binding of analyte to this biospecific surface is monitored through SPR in a thin gold film that is the base of all sensor chips. The change in SPR response is directly related to changes in the refractive index at the biospecific surface. The refractive index at the surface is directly related to the concentration of molecules at the surface. Binding of analyte to the immobilized ligand, for example, causes a change in the refractive index at the surface, which in turn leads to a change in the reflected light angle at which SPR is observed (the SPR angle). Binding data are presented in the form of a sensorgram, which is a plot of the SPR angle, converted to resonance units (RU), versus time 53 (Fig. 1). For the most commonly used sensor chips, a change of 1000 RU is equivalent to binding of about 1 ng of protein and 0.8 ng of nucleic acid surface concentration per millimeter squared. 53,56,57 The greater response per nanogram of nucleic acid is a result of their generally higher refractive index increment with respect to proteins. Small molecules can have quite different refractive index increments (RIIs) than proteins and nucleic acids and the importance of this difference is discussed below. Experimental

Design and Protocols

Experimental Design~Considerations

There are several important factors to consider when designing an SPR experiment. At the forefront of experimental design are the questions of which flow cell surface to use, which species to immobilize, what chemistry to use to immobilize it, 54R. L. Rich and D. G. Myszka,Curl: Opin. Biotechnol. 11, 54 (2000). 55 Biacore, "BIACoreBIAtechnologyHandbook." Biacore, Uppsala, Sweden, June 1994. 56M. Buckle, R. M. Williams, M. Negroni, and H. Buc, Proc. Natl. Acad. Sci. U.S.A. 93, 889 (1996). 57R. J. Fisher, M. Fivash, J. Casasfinet, J. W. Erickson, A. Kondoh, S. V. Bladen, C. Fisher, D. K. Watson, and T. Papas, Protein Sci. 3, 257 (1994).

28


[2]

3 n-

steadyassociation state

dissociation

.1 ~

-50

0

50

100 Time

150

200

250

300

350

(second)

FIG. 1. Example sensorgramcollected on a BIAcore 2000 biosensor.The association, steady state, and dissociation regions are indicated. The data result from binding of a small molecule (5 nM) to immobilizednucleic acid. and how much of it to immobilize. In an ideal experiment, both interacting species would be immobilized in reverse experiments to ensure that any data collected is not affected by the immobilization and other experimental factors. If only one can be immobilized then it is advantageous to immobilize the lower molecular weight species because the response signal on analyte binding is dependent on the molecular weight (the greater the molecular weight, the greater the signal). 53 In RNA (or D N A ) - s m a l l molecule systems, it is often impractical or impossible to immobilize the small molecule, especially when the purpose of the experiment is to screen a small molecule library. Subtle changes in a small molecule can drastically affect binding, making perturbation-free immobilization and data interpretation difficult. In contrast, immobilizing RNA through a 5' or 3' tether is unlikely to affect specific binding at a site in the interior of the molecule. Fortunately, we have found

[21

RNA INTERACTIONS

29

excellent signal-to-noise ratios (S/N) with small molecule binding, and where it has been possible to determine the equilibrium constant through both solution and SPR methods for small molecule-nucleic acid complexes, the constants are in excellent agreement. 58 Measuring binding of the low molecular weight compound does, however, require considerable attention to baseline stability. Methods that we have found to give the necessary S/N ratios are presented throughout. For immobilizing nucleic acids, the biotin-streptavidin affinity complex has been the immobilization method used in the vast majority of BIAcore SPR experiments,48-51.59-63 although other immobilization methods are appearing. 64 In the biotin-streptavidin immobilization method, biotinylated RNA (see Preparation of RNA, below) is immobilized on a sensor chip that has been derivatized with streptavidin (see Preparation of Sensor Chip, below). The rather large affinity constant for the biotin-streptavidin complex creates a quite stable binding surface over the time course of most experiment. However, in our experience, streptavidinnucleic acid sensor chips gradually lose binding activity through repeat use. We have been able to immobilize additional nucleic acid onto the sensor chip to achieve the original binding level but the activity of the surface continues to decline. Alternative immobilization strategies, including covalent attachment on activated carboxymethylated dextran through both a primary amine and a thiol group tethered to the 5' end of the DNA, offer the potential for longer chip stability. However, in our hands, neither strategy yielded enough immobilized nucleic acid for small molecule studies. It appears that the negative charge on the carboxymethyl dextran inhibits sufficient immobilization of the negative nucleic acid even at extremes of salt and pH in reasonable time periods. We were, however, able to quantitatively immobilize a 5'-thiolated nucleic acid directly onto a gold surface, using the method outlined in Heine and Tarlov,65 but the sensor chip lost significant activity rapidly, probably due to noncovalent adsorption of the nucleic acid on the surface. The interaction between small molecule compounds and the gold sensor chip surface is a point that needs investigation as research on this 58 S. Mazur, F. A. Tanious, D. Ding, A. Kumar, D. W. Boykin, I. J. Simpson, S. Neidle, and W. D. Wilson, J. Mol. Biol. 300, 321 (2000). 59 p. j. Bates, H. S. Dosanjh, S. Kumar, T. C. Jenkins, C. A. Laughton, and S. Neidle, Nucleic Acids Res. 23, 3627 (1995). 6o G. Bischoff, R. Bischoff, E. Birch-Hirschfeld, U. Gromann, S. Lindau, W.-V. Meister, S. Bambirra, C. Bohley, and S. Hoffman, J. Biomol. Struct. Dynam. 16, 187 (1998). 61 B. Persson, K. Stenhag, E Nilsson, A. Larsson, M. Uhlen, and E-A. Nygren, Anal. Biochem. 246, 34 (1997). 62 C. Rutigliano, N. Bianchi, M. Tomassetti, L. Pippo, C. Mischiati, G. Feriotto, and R. Gambari, Int. J. Oncol. 12, 337 (1998). 63 T. M. Nair, D. G. Myszka, and D. R. Davis, Nucleic Acids Res. 28, 1935 (2000). 64 C. I. Webster, M. A. Cooper, L. C. Packman, D. H. Williams, and J. C. Gray, Nucleic Acids Res. 28, 1618 (2000~[. 65 T. M. Heme and M. J. Tarlov, J. Am. Chem. Soc. 119, 8916 (1997).

30


[2]

type of immobilization proceeds. We are currently working on a new method for covalently immobilizing nucleic acid onto a modified sensor chip that will allow for regeneration of the original surface after the chip has lost activity or when it is desirable to immobilize a different nucleic acid sequence. In the meantime, however, immobilization through the biotin-streptavidin affinity complex is the recommended strategy to efficiently and reproducibly immobilize nucleic acids for SPR experiments. The next question that must be addressed in sensor chip surface preparation is how much nucleic acid to immobilize. For kinetic experiments it is usually best to immobilize the smallest amount of nucleic acid (or any ligand) possible, while maintaining the desired signal-to-noise ratio, in order to minimize mass transport effects. Mass transport of analyte to the surface will alter kinetic data when the rate of mass transport is slower than or on the same time scale as the interaction kinetics. 53 Because a high concentration of surface binding sites depletes the analyte at the surface, the more nucleic acid immobilized the greater the contribution from mass transport. Crouch et al. suggest an immobilization level less than 100 RU for kinetic analysis of RNA-protein interactions. 66 When the analyte is a small molecule, however, it becomes necessary to increase the ligand surface density because the instrument response from small molecule binding will be much less than the response from protein binding. We typically immobilize about 250 RU of hairpin nucleic acid ('v50 bases in length) for small molecule assays. In our experience, the increase in surface density does not affect kinetic data because small molecules diffuse more rapidly than macromolecules and are not as limited by mass transport as macromolecules. Our small molecule nucleic acid experiments to date have not been limited by mass transport. It is important, however, to account for mass transport when it is present, to correctly analyze kinetic data. See Refs. 67-69 for a detailed description of processing kinetic data influenced by mass transport and additional suggestions for minimizing its effects. Because mass transport does not affect steady state data, it is advantageous for signal-to-noise optimization to immobilize a higher level of nucleic acid for equilibrium analyses, especially when working with small molecules. Experimental Protocols Preparation o f Sensor Chip. For immobilizing biotin-RNA (or any biotinnucleic acid) (see Preparation of RNA, below) on a sensor chip, the sensor chip

66R. J. Crouch, M. Wakasa, and M. Haruki, in "RNA-ProteinInteractionProtocols"(S. Haynes, ed.), pp. 143-160. Totowa, New Jersey, 1999. 67B. Goldstein, D. Coombs,X. He, A. R. Pineda, and C. Wofsy,J. Mol. Recognit. 12, 293 (1999). 68D. G. Myszka, Curt: Opin. Bioteehnol. 8, 50 (1997). 69T. A. Morton and D. G. Myszka,Methods Enzymol. 295, 268 (1998).

[2]

RNA INTERACTIONS

31

must be modified to a streptavidin surface. BIAcore offers high-quality premade streptavidin sensor chips (SA sensor chip) that are ready for immediate use. However, it is possible and in some cases worthwhile to prepare streptavidin sensor chips (Refs. 50 and 53, and unpublished results, 2000) using standard (CM5) dextran surfaces or pioneer chips (chips that are under development but are available from BIAcore) with features such as a low-density carboxyl surface. The low-density carboxyl surfaces uses dextran but has less negative charge and so may be advantageous when investigating the interactions between RNA and molecules with high positive charge. We use the procedure outlined below for immobilizing streptavidin on CM5 sensor chips. The procedure is straightforward, and in our experience, generates streptavidin chips of similar quality to the premade chips available from BIAcore but with lower cost and additional flexibility. REQUIRED MATERIALS AND SOLUTIONS.The following is adapted from Ref. 55. CM5 or Pioneer sensor chip HEPES-buffered saline (HBS) buffer: 10 mM HEPES (pH 7.0), 150 mM NaC1, 3 mM EDTA, 0.005% (v/v) polysorbate 20 (running buffer) N-Hydroxysuccinimide (NHS): 100 mM in water N-Ethyl-N'-(dimethylaminopropyl)carbodiimide (EDC): 400 mM in water Acetate buffer (pH --~ 5): 10 mM (immobilization buffer) Streptavidin, 200-400 #g/ml [Pierce (Rockford, IL); Molecular Probes (Engine, OR); or Sigma (St. Louis, MO)] in immobilization buffer Ethanolamine hydrochloride: 1 M in water (pH 8.5) (deactivation solution) Sodium dodecyl sulfate (SDS): 0.05% (w/v) IMPORTANTNOTES 1. The sensor chip and solutions should equilibrate to room temperature for at least 30 min prior to use. 2. Because of the small diameter of the tubes in the microfluidics, all buffers should be filtered (0.22-#m pore size filter) and degassed daily prior to use: 3. The "amine coupling kit" available from BIAcore contains the NHS, EDC, and ethanolamine hydrochloride. The kit reagents do not need to be filtered or degassed. These reagents are also available from other suppliers but must be of sufficient quality so as not to clog the microfluidics, and they must be filtered and degassed. 4. For efficient immobilization of streptavidin (or any protein), it is necessary to use an immobilization buffer with a pH below the pl of the protein. If the macromolecule has a negative charge during immobilization, the amount of

32


[2]

material that can be linked to the surface will be limited. Different sources and batches of streptavidin can vary in pl (pl 5-6), resulting in the necessity to optimize the immobilization procedure for each new batch (see Preconcentration of Streptavidin, below). BIAcore reports the use of 200 #g/ml in 10 mM acetate, pH 4.5, for streptavidin from Sigma, and 400 #g/ml in 10 mM acetate, pH 5.0, for streptavidin from Molecular Probes. We currently use streptavidin at 200 #g/ml in 10 mM acetate, pH 4.5, for streptavidin from Pierce. 5. BIAcore Software requires specification of volumes rather than times when performing injections. Therefore, with a flow rate of 5 #l/min, a 2-min injection must be specified as 10 #1. PRECONCENTRAT1ON OF STREPTAVID1N. Efficient immobilization of proteins through amine coupling on a carboxymethylated dextran surface involves electrostatic attraction between the negative charges on the surface matrix (carboxymethyl dextran) and positive charges on protein when it is below its pl (termed preconcentration). Ideally, the solution should have low ionic strength (50 mM or less) to maximize preconcentration. 55 It is important to note that if the pH of the buffer is too low (pH < 3-3.5), the dextran matrix becomes protonated and preconcentration is less efficient. 55 In addition, amine coupling requires uncharged amino groups and is therefore favored by higher pH. Clearly, the optimum pH is a compromise between efficient preconcentration and efficient coupling. To optimize the immobilization of streptavidin through amine coupling, injections of streptavidin in buffers differing in pH should be done as outlined below. Three buffers at pH 4.5, 5.0, and 5.5 are generally sufficient. We currently use 10 mM acetate buffer although any buffer lacking primary amine groups should suffice (note that a Tris buffer is not suitable for amine coupling). PRECONCENTRATIONPROTOCOL. The following is adapted from Ref. 55.

1. Dock a CM5 or Pioneer sensor chip into the instrument. 2. Prime the instrument three times with HBS running buffer. 3. Start a sensorgram at 5 #l/min. 4. Perform a 2-min "QUICK INJECT" of each streptavidin-pH combination from'high to low pH. Begin with a streptavidin concentration of 200 #g/ml. 5. Perform a l-min "QUICK INJECT" with "Extra Clean" consisting of 0.05% (w/v) SDS to remove any residual streptavidin from the fluidics. 6. Perform a 1-min "QUICK INJECT" with "Extra Clean" consisting of HBS buffer to remove any residual SDS from the fluidics. 7. Using the crosshair, note the preconcentration level in RU. BIAcore suggests that a streptavidin immobilization level of ~5000 RU is suitable for most applications. 55 We currently immobilize 2500-3500 RU of streptavidin. We have found this immobilization level suitable for our applications.

[2]

RNA INTERACTIONS

33

Typically, 60-80% of the protein preconcentration level is immobilized. Therefore, aim for an electrostatic preconcentration level 150% of the desired immobilization yield. 55 For example, to immobilize 3000 RU of streptavidin, use conditions that favor an electrostatic preconcentration level of ~4500 RU. Using the data from the preconcentration experiment above, determine which pH is best for immobilizing the desired amount of streptavidin, keeping in mind that a lower pH may also reduce coupling efficiency. If none of the solutions produces the desired response, it may be necessary to increase the concentration of streptavidin [see Important Notes (4) above]. IMMOBILIZATIONOF STREPTAVIDIN. After choosing the appropriate pH to immobilize the desired amount of streptavidin, follow the procedure outlined below for immobilizing streptavidin. 1. Run a sensorgram at 5 #l/min. 2. With NHS in one vial and EDC in other, use the DILUTE command to make a 1 : 1 mixture of NHS-EDC (note that the NHS-EDC solution must be prepared fresh immediately prior to use). 3. Inject NHS-EDC for 7 min (35 #1) to activate the carboxymethyl surface to reactive esters. Higher or lower activation may be desired for preparation of nonstandard surfaces. 4. Inject streptavidin in the appropriate buffer for 7 min (35 #1) to immobilize streptavidin. 5. Inject ethanolamine hydrochloride for 7 min (35 /~1) to deactivate any remaining reactive esters. If immobilization fails after achieving appropriate levels of preconcentration then it may be necessary to increase the time of activation (the volume of NHS-EDS injected) or the streptavidin concentration if the pH is relatively low. Alternatively, EDC is quite labile and if the EDC solution is not freshly made or has not been kept properly frozen it could contain extensive inactive material. We keep EDC as a dry powder in a desiceant bottle at 4 ° and make fresh solutions as needed rather than aliquot all the EDC as suggested in the BIAcore amine coupling kit. We have found that streptavidin immobilization is straightforward as long as fresh solutions are used. Preparation of RNA: Derivatization to Incorporate Biotin. The biotin-streptavidin capture method requires that the RNA be modified with biotin. The ideal level of biotinylation for BIAcore nucleic acid applications is one biotin per nucleic acid. Small RNAs can be synthetically prepared with biotin on either the 5' or 3' terminus and are commercially available. Larger RNAs must be prepared by in vitro transcription. An excellent source for the synthesis of RNA by in vitro transcription is Wyatt et al. 7° The use of guanosine 5'-monophosphorothioate (GMPS) to prime a transcription reaction is a convenient way to site-specifically

34


[21

incorporate a single 5' terminal thiol group into RNA. 50'70-72 The thiol can then be modified to biotin with biotin-maleimide or biotin-iodoacetamide conjugates (Molecular Probes). A concise procedure for synthesizing GMPS from guanosine can be found in Burgin and Pace. 72 Alternatively, larger RNAs can be transcribed so as to incorporate 4-thiouridine internally, 73 and then modified with biotin as described above. Drawbacks to this method are that (1) additional experiments must be performed to ensure the internal biotin moiety does not disrupt native structure or function, and (2) it is difficult to achieve a low level of biotinylation unless only a couple of uridines are present in the RNA. RNA can also be sitespecifically modified to contain a thiol group on the 5' terminus through a kinase reaction. 7t Briefly, RNA is prepared by in vitro transcription, using GMP or GTP as the priming nucleotide instead of GMPS, or obtained commercially unmodified. The Y-phosphate(s) are then removed with phosphatase followed by transfer of the thiophosphate of ATPvS to the RNA by T4 polynucleotide kinase. Other methods to incorporate biotin into RNA include oxidation of the 3' terminal cis-diol to dialdehyde by periodate followed by reaction with biotin-hydrazine conjugates (Molecular Probes). Alternatively, the Y-phosphate of RNA (or DNA) can be converted to a 5'-phosphorimidazolide with imidazole and a water-soluble carbodiimide, and subsequently reacted with biotin-amine derivatives (Molecular Probes). Detailed protocols as well as limitations and necessary considerations for all of the above-described methods can be found in Qin and Pyle 71 (that article focuses on modification of RNA with fluorophores, but biotin derivatives can be used instead). It should be noted that regardless of the method used to modify RNA, excess biotin must be efficiently removed after modification to avoid competition with biotin-RNA for streptavidin binding sites and to enable reproducible levels of immobilization. Gel filtration on a desalting column and dialysis are simple methods to separate excess biotin from biotin-RNA. 55 Immobilization o f RNA. Once derivatized with biotin, the RNA is ready to be immobilized on a streptavidin-coated sensor chip. Generally, immobilization methods that rely on rapid kinetics and high-affinity binding of the ligand to the surface, such as the streptavidin-biotin interaction, do not require preconcentration experiments to assist in immobilization. 55 To date, we have worked with relatively short oligonucleotide hairpins (< 50 bases). For this size nucleic acid, we routinely use HBS as the immobilization and running buffer. Larger RNAs may require either higher salt to minimize electrostatic repulsion or a pH lower than the pl of streptavidin to favor electrostatic interactions with the negatively charged RNA, We routinely use a concentration of ~25 nM oligonucleotide (1 : 1 biotin : RNA 7oj. R. Wyatt, M. Chastain, and J. D. Puglisi, Biotechniques 11, 764 (1991). 71 p. Z. Qin and A. M. Pyle, Methods 18, 60 (1999). 72A. B. Burgin and N. R. Pace, EMBO,I. 9, 4111 (1990). 73j. E Milligan and O. C. Uhlenbeck, Methods Enzymol, 180, 51 (1989).

[21

RNA INTERACTIONS

35

ratio) in HBS buffer when immobilizing nucleic acids less than 50 bases in length. It may be necessary to increase the concentration when using larger nucleic acids. A concentration that is too high, however, will make control over the amount of ligand immobilized difficult. In addition, too high a concentration may cause the outer dextran layer to bind RNA rapidly, potentially blocking inner layers and limiting the amount of RNA that can be immobilized. As discussed in the introduction, we typically immobilize about 250 RU of nucleic acid for small molecule studies. REQUIRED MATERIALSAND SOLUTIONS Streptavidin-coated sensor chip (SA chip or prepared as outlined above). HBS buffer: 10 mM HEPES (pH 7.0), 150 mM NaC1, 3 mM EDTA, 0.005% (v/v) polysorbate 20 (running buffer) Activation buffer: 1 M NaC1, 50 mM NaOH Biotin-modified RNA: ~25 nM in appropriate running buffer Biotin (optional) IMPORTANTNOTES 1. RNA is extremely sensitive to nucleases and base hydrolysis. Many researchers strictly use diethyl pyrocarbonate (DEPC)-treated water when making solutions for RNA research. We have found water from commercial systems such as NANO (Barnstead, Inc., Dubuque, Iowa) or Milli-Q (Pharmacia, Peapack, NJ) of sufficient quality for RNA work. 2. Because of the sensitivity of RNA, it is critical to maintain a "nucleasefree" working environment. It is always necessary to run the BIAcore Desorb method just prior to working with RNA, especially if there are multiple users on the instrument. In addition, any external instrument components such as the needle and the tubing placed in the running buffer should be routinely treated with RNase Zap (Ambion, Austin, TX) or another chemical to inactivate/remove any ribonucleases. 3. Researchers at the National Institutes of Health (NIH, Bethesda, MD) routinely cleanse the fluidics of the BIAcore instrument by injecting 20 #1 of RNase Zap solution through each flow cell at 20 #l/min followed by ten 20-#1 injections of DEPC-water (Ref. 66 and Inna Gorshkova, personal communication, 2000). It may be advisable to run this procedure when there are multiple instrument users, if a ribonuclease has ever been used in an experiment, or if problems with RNA degradation during an experiment are encountered. IMMOBILIZATIONPROTOCOL. The following is adapted from Ref. 55. 1. Dock the streptavidin-coated sensor chip into the instrument. 2. Prime the instrument three times with HBS buffer (note that priming does not need to be done if the streptavidin chip was just made as outlined above,

36


[2]

has not been removed from the instrument and HBS has been the running buffer). 3. Start a sensorgram with a 20-/zl/min flow rate. 4. Inject activation buffer for 1 min (20 ~1) three times to remove any unbound streptavidin from the sensor chip. 5. Allow buffer to flow for at least 5 min before immobilizing the RNA. Note that NaOH will facilitate hydrolysis of the RNA, so it is critical that buffer flows long enough to remove any trace of NaOH. 6. Select the desired flow cell on which to immobilize the RNA and start a new sensorgram for only that flow cell. Take care not to immobilize RNA on the flow cell chosen as the control flow cell. Generally flow cell 1 ("fcl") is used as a control and RNA is not immobilized on it. It is often desirable to immobilize a different RNA (perhaps wild-type and mutants) on different flow cells. In this situation it is necessary to immobilize ligands one at a time. Alternatively, it may be desirable, depending on specific needs, to immobilize different amounts of the same ligand on each flow cell. 7. Using M A N U A L INJECT with a flow rate of 2 #l/rain, load the loop with ~ 1 0 0 #1 of RNA and inject over the desired flow cell(s). Using the crosshair, track the number of resonance units immobilized and stop the injection after the desired level is reached. As a rule of thumb, do not inject more than 80-90% of the loaded volume, to minimize sample dispersion effects at the end of the injected sample. If necessary, reload the loop and inject more RNA. It may be necessary to alter immobilization conditions such as ionic strength, pH, and ligand concentration to obtain the desired immobilization level. 8. At the end of the injection and after the baseline has stabilized, use the crosshair to determine the resonance units of RNA immobilized and record this amount. The amount of RNA immobilized is required to determine the theoretical number of small molecule binding sites for the flow cell. 9. Repeat steps 7 and 8 for each flow celI-RNA combination according to experimental design. i0. Some researchers block the remaining biotin binding sites and the biotin binding sites on the streptavidin in the control flow cell with free biotin. This will depend on any nonspecific interactions between components in the flow solution with streptavidin on the sensor chip surface. STORAGEOF RNA SENSOR CHIP. Researchers at the NIH store oligonucleotide sensor chips in Tris-buffered saline (TBS) buffer to maintain binding activity (Inna Gorshkova, personal communication, 2000). To do so, remove the sensor chip from the white cartridge, being careful not to touch the flow cells. Immerse and store the chip in RNase-free TBS buffer. To subsequently use the sensor chip, remove it from the TBS. Shake the sensor chip a couple of times to remove the excess buffer, and then leave the chip to air dry in a sterile environment until it is completely

[2]

RNA INTERACTIONS

37

dry. Do not touch or wipe the flow cells in any way. Insert the sensor chip into the white cassette and dock. If storing sensor chips dry, it may be advisable to run a solution of RNase inhibitors over the sensor chip prior to undocking, Data Collection and Analysis We routinely obtain high-quality binding data on a BIAcore 2000 instrument, even with only 10% site saturation of small molecules (MW 300-600) with only a few hundred resonance units of immobilized nucleic acid (Fig. 1). For binding constants of 107-108 M -l, as observed with many RNA complexes (our unpublished results, 2000), small molecule concentrations from 1 nM to 1 # M in the flow solution allow accurate determination of binding constants. Methods to extract binding constants from experimental data have been described in detail. TM In addition to instrument performance, data collection and data processing influence the quality of the data and the information that can be extracted from it. Myszka published a comprehensive review outlining the steps necessary to obtain high-quality biosensor data. 75 The article discusses many of the potential pitfalls of biosensors and provides suggestions for avoiding them. The discussion below focuses on application of those principles to nucleic acid-small molecule complexes. Data Collection

Some of the issues affecting the quality and accuracy of the data obtained through SPR were addressed above. For example, the advantages of working with a homogeneous surface with as low a surface density as possible have been discussed. In addition, one flow cell should be left blank (no RNA immobilized) to provide a control flow cell. Some researchers immobilize biotin or an unstructured RNA on the control flow cell to match the control and RNA flow cells as closely as possible. However, with most compounds in our laboratory, we have not found significant adsorption on the dextran or streptavidin surface. In addition, adding a different RNA to the control surface could introduce unknown sources of nonspecific binding in the blank subtraction. Additional issues such as the mode and length of sample injection, sample concentration range and order of injection, regeneration and buffer injections, and the integrity of the surface are addressed when setting up an experiment. Surface interactions can be a serious problem and must be evaluated for each compound series to be investigated. There are several injection modes available in the BIAcore 2000 and 3000 instruments for delivering sample to the sensor chip. The injection modes differ in the amount of volume that can be injected. More importantly, they differ in

74j. j. Correiaand J. B. Chaires,Methods Enzymol. 240, 593 (1994). 75D. G. Myszka,J. Mol. Recognit. 12, 279 (1999).

38


[2]

the amount of sample dispersion that occurs during transport and injection. It is generally best to use the injection mode "KINJECT" for sample injections. This mode offers the maximum number of air segments surrounding the flowing sample, minimizing the amount of sample dispersion and dilution with buffer. Myszka points out that high flow rates are advantageous since faster switching between running buffer and sample helps to deliver a consistent sample plug. 75 In addition, high flow rates help to minimize rebinding of analyte to ligand during the dissociation phase when the association rate is fast. Higher flow rates are not necessary in steady state experiments, and may not be feasible if the kinetics are slow and sample is limited because faster flow rates require more sample. Sample concentrations should vary over a wide range (at least 100-fold). In general, the middle of the sample concentration range should be near 1/Ka if the Ka can be approximated. If the Ka is unknown, a broad concentration range should be used in a preliminary experiment to obtain an estimate of the K a. Ideally, the order of sample injection should be randomized. Many of the small molecules we have worked with, however, adsorb nonspecifically to the tubing of the microfluidics and are slowly released over the course of the experiment. In this situation, we have found that injecting samples from low to high concentration is necessary for eliminating artifacts in the data from adsorption carryover. It is also important to replicate each injection at least once. At the end of the injection, a regeneration step should be performed to remove any analyte still bound to the surface. Many of our compounds dissociate rapidly enough to regenerate the surface with buffer flow alone and do not require a separate regeneration step. When working with RNA, each solution and injection increases the potential for degradation so the least number of steps the better, as long as the quality of the data is not compromised. When a regeneration step is required, we typically use 1 M NaCI. We have used 1 M NaCI under slightly acidic conditions but found that salt alone was sufficient to dissociate most cationic compounds from RNA. To subsequently remove any remaining NaCI, we perform two 1-min injections of running buffer prior to the end of the cycle followed by a 5-min wait with running buffer flowing. After the next cycle has begun, we have the instrument wait another 5 min with running buffer flowing to ensure baseline has stabilized before the next sample injection. When working with small molecules that often give changes of less than 2 RU at steady state, it is essential that the baseline not drift significantly during the injection. We typically achieve a drift less than 0.5 RU/min by putting in the 5-rain wait (instrument specification is 1 RU/min), and such attention to the baseline is essential in small molecule binding experiments. For our experiments, we identify a well-behaved and well-characterized analyte as a standard to periodically evaluate the integrity of the surface when working with both RNA and DNA. In some of our experiments the RRE hairpin was the immobilized RNA. Because the Rev-RRE complex is a well-behaved 1 : 1 complex under appropriate conditions and has previously been studied by SPR, 5° we used

[2]

RNA INTERACTIONS

39

the Rev peptide to periodically check the amount and integrity of the RRE immobilized. If a well-described analyte is not available, any well-behaved molecule that binds to the immobilized RNA can be injected after RNA immobilization and calibrated for use as a standard for that surface. The same concentration of the standard is analyzed periodically and the resonance unit response compared with that of the initial injection. By using a calibrator molecule, the amount of RNA on the chip can be routinely checked and any loss can be accounted for during data processing. We have found a calibrator essential when working with the same RNA sensor chip for extended periods of time, especially if the chip has been stored for any length of time. If there are questions about the chip, a more thorough evaluation of the surface, including a complete concentration gradient experiment with a standard, should be performed to compare kinetic rates and immobilization levels with those obtained after the RNA was first immobilized. Below is a method that we routinely use to collect small molecule data on nucleic acid surfaces. This method is set for a flow rate of 10 #l/min (FLOW 10) over flow cells 1,2, and 3 (FLOWPATH 1,2,3). The samples are injected as written (STEP) from low to high concentration and then the injection sequence is repeated (TIMES 2). Note that before any drug is injected, 10- and 5-min buffer injections are done to enable double referencing (see Data Analysis, below). 75 In addition, the volume of drug injected is set as a variable so that the least amount of volume required to reach a steady state is used for each concentration. Much less time is required for the association reactions at high concentration of analyte.

Typical Method MAIN RACK 1 thermo_c RACK 2 thermo_a FLOWCELL 1,2,3,4 LOOP DB340 STEP APROG drug %sample2%position2%volume2%conc2 ENDLOOP APPEND Continue END DEFINE APROG drug PARAM %sample2%position2%volume2 %conc2 KEYWORD Concentration %Conc2 CAPTION %conc2%sample2 over RRE (gradient surface) F L O W 10 FLOWPATI-I 1,2,3

40


[2]

WAIT 5:00 KINJECT %position2 %volume2 300 QUICKINJECT r2f4 20 ! 1M NaCI EXTRACLEAN QUICKINJECT r2f3 10 !Buffer EXTRACLEAN QUICKINJECT r2f3 10 !Buffer EXTRACLEAN WAIT 5:00 END DEFINE LOOP DB340 LPARAM %sample2%position2%volume2 %conc2 TIMES 2 Buffer Buffer DB340 DB340 DB 340 DB340 DB340 DB340 DB340 DB340 DB340 DB340 DB340 DB340 DB340 DB340 DB340

r2f3 r2f3 r2al r2a2 r2a3 r2a4 r2a5 r2a6 r2a7 r2a8 r2a9 r2al0 r2bl r2b2 r2b3 r2b4 r2b5

100 50 100 100 100 100 50 50 50 50 50 50 50 50 50 50 50

0.000u 0.000u 0.025u 0.075u 0.125u 0.250u 0.375u 0.500u 0.625u 0.750u 1.000u 1.750u 2.500u 3.250u 5.000u 6.750u 10.00u

END

Data Processing After the data have been collected, there are several processing steps that must be carried out before any quantitative information can be extracted. We use BIAevaluation 3.0 software for data processing. Zeroing on the y axis (RU) and then the x axis (time) are the first steps in data processing. Because the flow cell surfaces are not identical to each other, the refractive index of each surface is different,

[2]

RNA INTERACTIONS

41

causing the flow cells to register at different positions on the y axis. Zeroing the data on the y axis is necessary to allow the responses of each flow cell to be compared. Generally the average of a stable time region of the sensorgram, prior to sample injection, should be selected and set to zero. Because the flow cells are aligned in series, sample is not injected across the flow cells simultaneously. Zeroing on the x axis aligns the beginnings of the injections with respect to each other. Myszka has shown that artifacts, such as spikes at the beginning and end of injections and baseline fluctuations, are usually present in all flow cells. 75 The two data processing steps outlined next serve to minimize these artifacts and also to correct for the bulk shift that results when injection buffer and running buffer are not identical. In the first step, the control flow cell sensorgram is subtracted from the reaction flow cell sensorgrams. This removes the bulk shift contribution to the change in resonance units. The last step in data processing is required to remove systematic deviations that are frequently seen in the sensorgrams. Referred to as "double referencing," this process removes the systematic drifts and shifts in baseline that are frequently observed even in control cell-subtracted sensorgrams. In the data collection method shown above, two buffer injections are performed, one for each volume amount used for sample injection. In double referencing, plots are made for each flow cell separately, overlaying the control cell-corrected sensorgrams from buffer and all sample injections. The buffer sensorgram is then subtracted from the sample sensorgrams. If multiple buffer injections were done, averaging these sensorgrams before subtracting from the sample sensorgram can improve the data further. At this point, the data should be of optimum quality and are ready to be analyzed.

Data Analysis In most systems, both equilibrium and kinetic rate constants can be extracted from SPR data. The equilibrium constant can be obtained from fitting steady state data directly, or from kinetic data. The association equilibrium constant (K) is the ratio of the association (ka) and dissociation rate (kj) constants. Comparing the K value obtained both ways is a good way to evaluate the models used to fit the data. Most cationic molecules that bind RNA also interact nonspecifically with the RNA, but frequently both K, ka, and ka for the strong specific site can be determined accurately because the equilibrium constants for the two types of binding differ markedly. Having both kinetic and equilibrium information can yield important insight into understanding whether any biological activity of the small molecule is governed by the kinetic or equilibrium (or both) constants of binding to RNA. Knowledge of the stoichiometry of the system is essential for obtaining correct kinetic and binding constants as well as for obtaining a complete description of the system being studied. While at first glance determining stoichiometry seems straightforward, this is often not the case when working with nucleic acid-cation

42


[2]

TABLE II REFRACTIVE INDEX INCREMENTS AND REFRACTIVE INDEX INCREMENT RATIOS WITH B S A AND d A d T

RI incrementa Compound BSA Neomycin Netropsin Quinacrine Berenil Pentamidine DB404 dGdC dAdT

RI increment ratio vs BSAh

RI increment ratio vs dGdCh

590.5nm 633nm 679.5nm 590.5nm 633nm 679.5nm 590.5nm 633nm 0.18 0.15 0.22 0.24 0.35 0.22 0.28 0.22 0.21

0.17 0.15 0.21 0.25 0.34 0.22 0.28 0.22 0.21

0.17 0.15 0.21 0.24 0.33 0.22 0.27 0.22 0.21

-0.9 1.2 1.4 2.0 1.3 1.6 1.2 1.2

-0.9 1.2 1.4 2.0 1.3 1.6 1.3 1.2

-0.9 1.2 1.4 1.9 1.3 1.6 1.3 1.2

0.8 0.7 1.0 1.2 1.7 1.1 1.3 1.0 --

0.8 0.7 1.0 1.2 1.6 1.1 1.3 1.1 --

679.5 nm 0.8 0.7 1.0 1.2 1.6 I. 1 1.3 1.1 --

RI increment (RII) = A(An)/~xC; determined at 590.5,633, and 679.5 nm. RI increment ratio = RllcompoundI/Rllcompound2 = (~n/~C)compound l/(6n/~C)compound 2. interactions because nonspecific b i n d i n g will always be present. The refractive index increments (RIIs) of small molecules can vary considerably more than those of proteins because small molecules can contain a m u c h larger variety of potential constituent units. While it is recognized that the a m o u n t of analyte b o u n d and the molecular weight of the analyte contribute to the observed b i n d i n g response, the RII is often an overlooked contributor to the b i n d i n g response. W h e n studying m a c r o m o l e c u l a r interactions, such as p r o t e i n - p r o t e i n and p r o t e i n - n u c l e i c acid systems (the m a i n focus of biosensor studies until recently), the RIIs of the interacting species are in fact generally irrelevant because the RIIs of most proteins and nucleic acids are similar. We have found, however, that the RIIs of small molecules can vary by a factor of two, and can be smaller or significantly larger than the RIIs of proteins and nucleic acids (Table II). 76 Highly conjugated molecules have significantly larger RIIs than those of unconjugated molecules. Because the RIIs of small molecules can be different from those of proteins and nucleic acids, it is essential that such a difference be accounted for during data interpretation to correctly determine stoichiometry, and subsequently kinetic and equilibrium constants. The m a x i m u m BIAcore instrument response for a 1 : 1 b i n d i n g interaction can be predicted with the following equation: (RUpred)max = RUE \ M ~ L J

L~ - ]

where (RUpred)max is the predicted m a x i m u m instrument response in resonance units for b i n d i n g at a single site, RUE is the experimental a m o u n t of R N A 76 T. M. Davis and W. D. Wilson, Anal. Biochem. 284, 348 (2000).

[2]

RNA INTERACTIONS

43

immobilized on the chip in resonance units, MWA is the molecular weight of the analyte, MWL is the molecular weight of the RNA, (On/OC)Ais the RII of the analyte, and (On/OC)Lis the RII of the RNA. The maximum observed instrument response [(RUobs)max] will approach this predicted value as the concentration of analyte bound at the surface increases and saturation of the single binding site occurs. For systems with analyte-to-ligand stoichiometry greater than 1, RU(pred)rnax must be multiplied by the number of binding sites per macromolecule to obtain the maximum response: (RUpred)sat = (RUpred)maxP

(2)

where (RUpred)sat is the predicted instrument response on saturation of p binding sites in resonance units, (RUpred)maxis from Eq. (1), and p is the number of binding sites per macromolecule. The maximum instrument response on saturation of all binding sites, (RUobs)sat, is obtained from fitting SPR results, but a value for (RUpred)max is essential in order to correctly determine stoichiometry from (RUobs)sat. Small molecules generally have more than a single binding site in nucleic acid complexes. Because nonspecific binding can occur with cationic molecules and RNA, the RII ratio is critical for accurate determination of stoichiometry. Consider, for example, typical parameters for the cationic small molecule DB404 binding to DNA: RUL of 250, MWA of 467, MWL of 6465, and (On/OC)A/(On/OC)Lratio (RII ratio) of 1.3 (Table II). Without correcting for the RII difference, these parameters would predict an RUma× per binding site of 18 [Eq. (1)]. However, accounting for an RII ratio of 1.3, the actual predicted RUmax per binding site is 24. For a dimer binding mode, which we have encountered with similar compounds, 51 an experimental (RUobs)sat near 50 could incorrectly be characterized as a trimer binding mode [predicted at 54, Eq. (2)] if the RII ratio is not considered. Likewise, for a trimer binding mode, (RUobs)sat would register around 72, which could be mistaken for four binding sites per RNA [predicted at 72, Eq. (2)] if not corrected for the difference in RIIs. Clearly, the RII ratio is a critical factor to consider when interpreting SPR data involving small molecules. It should be noted that because stoichiometry is an integral number [p in Eq. (2)], it should only be necessary to have an approximate value of the RII to determine p. Available data may serve as a guide for estimating the RIIs of other small molecules if such data cannot be experimentally determined. 76

Equilibrium Analysis Although our laboratory uses the SPR biosensor to obtain kinetic constants and screen small combinatorial libraries, in many applications the binding constants of small molecule-nucleic acid systems are the primary requirement. Many of our compounds bind specifically to multiple sites on the nucleic acid, especially our

44


[21

small m o l e c u l e - R N A systems, and also display various degrees of nonspecific binding. To analyze equilibrium data with BIAevaluation software, the data should be treated in the following manner. After processing the data as outlined above, overlay the sensorgrams of all the analyte concentrations for the flow cell(s) with immobilized RNA (Fig. 2A). Using the software, calculate an average of the blankcorrected data in the steady state region of each sensorgram and then plot the average as a function of analyte concentration (Fig. 2B). The result is a recognizable binding isotherm. The simplest model system for determination of a binding constant from such results is one with a stoichiometric ratio of one binding site per

50

....

, ....

, ....

, ....

, ....

, ....

, ....

, ....

A

4o

3O

mr 20

10

0

-10

.... -50

' .... 0

~ .... 50 Time

' .... 100

' .... 150

' . . . . . . . . 200 250

' .... 300

350

(second)

FIG. 2. (A) Sensorgramsat different concentrationsof the small molecule DB244 binding to DNA. The data have been processed as described in text. (B) The average response at steady state for the curves shown in (A), plotted as a function of concentration.The result is a typical binding isotherm.

[2]

RNA INTERACTIONS 30

.

.

.

.

I

.

.

.

.

I

.

.

.

.

I

45 .

.

.

.

I

.

.

.

. d

25

20 ¢n

O:: 15

10

0

.

0

.

.

.

I

.

.

1x10" 6

.

.

I

.

.

.

.

2xlO" 6

I

.

.

.

.

3xlO" 6

Concentration

I

.

.

4xlO" 8

.

.

5xlO" 6

(M)

FIG.2. (Continued) RNA that can be fit with Eq. (3), using either the BIAevalution software or any suitable program, such as KaleidaGraph, for this purpose: KA[AI ] RUss = RUmax (1 ~K~A[AI)J

(3)

where RUss is the RU response at steady state, KA is the dissociation constant, RUmax is the fitted value for the maximum binding capacity of the surface in resonance units, and [A] is the concentration of analyte in the flow solution. Note that the free analyte concentration in equilibrium with bound analyte {[A] in Eq. (3)} is constant in the flow solution. The dissociation constant, Ko, is 1]KA. Note that in the ideal case, RUmax equals (RUpred)maxand (RUss)max. In some cases

46


[2]

it may not be possible to go to a high enough analyte concentration to reach the (RUss)max plateau. Also, if the surface has significantly degraded, RUmax will be less than Rg(pred)max. As mentioned above, most small cation-nucleic acid complexes have stoichiometry greater than 1. Many of our systems involve specific binding at one or two sites followed by additional nonspecific binding at higher concentration. For these systems, we typically visualize the data in the form of both direct and Scatchard plots while fitting by nonlinear least squares, using Eq. (3) or a more complex model. After obtaining the average of the data in the steady state region as described above, we plot the data as either R U J [ A ] as a function of RU~s or r/[A] as a function of r, where r is RWss/RW(pred)max.Plotting the data in Scatchard form can visually reveal considerable information about the binding constants, stoichiometry of specific and nonspecific binding, and cooperativity.51 Figure 3A provides an excellent example of complex binding information obtained on a BIAcore 2000 and presented in a Scatchard plot. In this example, the differences in binding constants, stoichiometry, and cooperativity for binding of two low molecular weight aromatic cations, DB293 and DB270, to two different DNA hairpins, oligo 1 and oligo 2-1, illustrate the power of SPR biosensors. The cooperative binding of two molecules of DB293 to oligo 2-1 is clear. This type of information is difficult to obtain by other methods. Equilibrium constants were obtained by fitting the Scatchard plots with either a single-site model [Eq. (4) with K2 = 0] or with the two-site model in Eq. (4): r =

KI[A] + 2K1K2[A] 2 1 + KI[A] + 2K1K2[A] 2

(4)

where r represents the moles of bound compound per mole of DNA hairpin and is calculated as described above, Kl and Kz are macroscopic association binding constants, and [A] is the concentration of analyte in the flow solution. Figure 3B provides another example of SPR data presented in the form of a Scatchard plot. The data are for binding of a small aromatic tetracation to an RNA hairpin. The data indicate a stoichiometry of two specific binding sites per RNA hairpin, but additional nonspecific binding occurs as the concentration increases. The binding constant for nonspecific binding is substantially lower that the binding constants for the two specific binding events. Higher salt concentrations help to reduce the amount of nonspecific binding.

Kinetic Analysis Because binding data is collected in real time with SPR biosensors, the technique is particularly well suited for kinetic analyses. Currently, association constants in the range of 103-106 M-1 sec-I and dissociation constants in the range of 10-5-10 -1 sec -1 can be accurately obtained on BIAcore instruments. These

[2]

RNA INTERACTIONS

47

ranges can be extended, however, with proper data handling, favorable S/N, and other conditions. One of the most important parameters to consider during kinetic analysis of SPR data is mass transport, which is transport of the analyte to the solid phase covalently bound with ligand. Although it is best to optimize data collection parameters so as to remove any contribution to the kinetics from mass transport, accurate kinetic constants can still be obtained when mass transport contributions

2 10 7

A

1.5 10 7

oo

~-.~"1 10 7

5 10 6

O

O

o

0

0.5

1

1.5

2

r FIG. 3. Examples of SPR data visualized by Scatchard plots. (A) Scatchard plot for binding of the small molecules DB293 and DB270 to two different DNA olignucleotides along with best fit binding curves are shown. (A, A) DB293 and DB270 binding to oligo 1; (O, O) DB293 and DB270 binding to oligo 2-1. [These data are reprinted with permission from L. Wang, C. Bailly, A. Kumar, D. Ding, M. Bajic, D. W. Boykin, and W. D. Wilson, Proc. Natl. Acad. Sci. U.S.A. 97, 12 (2000).] (B) Scatchard plot for binding of the small molecule DB340 to RRE with the best fit for the two strong binding sites (our unpublished data, 2000).

48


[21

8 x l 06

I

'

'

'

B 7 x l 06

6 x l 06

5 x l 06

0 4 x l 06 •

B

•

•

•

3 x l 06

2 x l 06

lx10 6

i° ' i

0

0.5

1

1.5

2

.

.

.

.

.

.

.

I

*

.

2.5

3

r

FIG,3. (Continued) cannot be entirely removed. 67-69 With small molecules, however, mass transport effects often do not exist. Indeed, we have never experienced mass transport limitations with any of the small molecules we have investigated to date. It is essential, however, to test for mass transport effects, by methods such as varying the flow rate, before proceeding with data collection and kinetic analysis. The best way to process kinetic data obtained with SPR biosensors is through global analysis, especially for complex systems. 77 Global parameters are constrained to the same value for each curve in the data set, in contrast to local 77T. A. Morton, D. G. Myszka, and I. M. Chaiken,Anal. Biochem. 227, 176 (1995).

[2]

RNA INTERACTIONS

49

parameters, which can have different values for each curve. Global analysis offers a more stringent test of reaction models and a statistically more reliable calculation of the parameters. 77 Currently, global analysis offers the most reliable fitting method for obtaining optimum fitted kinetic constants from SPR data. BIAevalution software versions 3.0 and above have the option to fit data globally. CLAMP is an excellent program for fitting SPR biosensor data that combines numerical integration and nonlinear global curve fitting routines. 69'78 The software is user-friendly and can be downloaded at http:/]www.hci.utah.edu]groups/interaction]clamp.htm. As with equilibrium analysis, the simplest system for fitting kinetic data is one with a stoichiometric ratio of one binding site per RNA. The classic 1 : 1 rate equation [Eq. (5)] is converted to SPR terms in Eq. (6). d[AB] dt dRU

dt

- ka[A][B] - kd[AB]

(5)

-- kaC(RUmax - RU) - kdRU

(6)

where [AB] is the concentration of complex and is directly related to the instrument response, resonance units in Eq. (6); [A] is the concentration of free analyte, which is the concentration of analyte in the flow, and is equal to the constant C in Eq. (6); [B] is the concentration of immobilized ligand that is not bound by analyte, and is equal to the maximum amount of B immobilized on the surface, RUmax, minus the amount of B bound with analyte, RU; ka is the association constant; and kd is the dissociation constant. For a 1 : 1 system with no mass transport, Eq. (6) should be used during fitting analyses once the data has been processed as described earlier. As mentioned above, most of our RNA-small molecule systems have stoichiometry greater than 1 and require more complex models for fitting. Complex models, however, usually have too many parameters for unique fits. Because SPR biosensors permit low analyte concentrations, it is often possible to isolate and fit only the strongest binding site. SPR biosensors can also be used to identify and rank small molecules that bind to an immobilized ligand, using data in the association and dissociation regions of a sensorgram, without actually determining the kinetic constants, s2 Once identified, a detailed kinetic analysis of the compounds that bind within the desired rate region can be determined in subsequent experiments. BIAsimulation software (BIAcore) is an excellent tool for generating simulated sensorgrams to evaluate the effect kinetic and experimental parameters have on the shape of a binding curve. For example, the program can be used to determine what amount of ligand to immobilize given a set of kinetic constants so as to minimize mass transport effects. The reader is urged to explore the applications of BIAsimulation during experimental design and data analysis. 78 D. G. Myszka and T. A. Morton, Trends Biochem. Sci. 23, 149 (1998).

50


[2]

Thermodynamic Analysis The BIAcore 2000 biosensor has the ability to collect data at temperatures from 4 to 40 °. Dissociation and kinetic constants obtained at different temperatures can be further analyzed to obtain thermodynamic information. In a review, 79 Myszka and Morton present examples of calculating the Gibbs free energy change of binding, the enthalpy and entropy of binding through van't Hoff analysis, and the enthalpic and entropic contributions to the free energy of activation through Eyring analysis for a model protein-protein system from data collected on a BIAcore 2000 biosensor. We refer the reader to this review as a guide to performing such analyses. Chaires 8° has presented experimental considerations required to obtain accurate thermodynamic parameters through van't Hoff analyses. A powerful approach to obtaining complete thermodynamic profiles for cation-nucleic acid interactions is to determine the equilibrium constant (Gibbs energy) as a function of temperature by SPR methods, and to directly determine the reaction enthalpy as a function of temperature by calorimetery. 58 Conclusion Surface plasmon resonance biosensors are firmly established for monitoring macromolecular interactions and are becoming a popular technique for characterizing small molecule complexes. Only more recently, however, has the technology advanced to the point that small molecule systems can be studied without the need for a high molecular weight tag. SPR biosensors are amenable to the study of kinetics and equilibrium properties of many systems as well as screening of combinatorial libraries, and the potential applications of biosensors are rapidly expanding. SPR biosensors offer an excellent method for investigating small molecule-macromolecule systems because many small molecules are either not fluorescent or do not yield notable gel band shifts on binding to macromolecules. In addition, such small quantities are required that precious or sparse samples can be investigated with minimal material requirements. This is in direct contrast to UV or CD spectroscopies, where a 10- to 100-fold increase in sample quantity may be necessary to obtain a satisfactory signal-to-noise ratio. This review is intended to serve as a guide for beginning users of SPR biosensors. Other excellent reviews and articles have been noted and the reader is encouraged to explore them also. It is worth reemphasizing that the refractive index increments of small molecules must be considered for accurate analysis of SPR data for systems with stoichiometry greater than 1. The application of biosensors 79D. G. Myszka,MethodsEnzymol. 323, 325 (2000). 80j. B. Chaires,Biophys. Chem. 64, 15 (1997).

[3]

MEASUREMENT OF BINDING AND UNWINDING

51

to small molecule systems is so new that advances, pitfalls, and other important considerations in this area are likely to develop rapidly.

Acknowledgments Work from our laboratory described in this review has been supported by the NIH. We also acknowledge Farial Tanious for helpful suggestions and data for some figures and Lei Wang for data for some figures. We appreciate preprints from Dr. David Myszka.

[3] Simultaneous Measurement of Binding Constants

and Unwinding Angles by Gel Electrophoresis By STEVEN M. ZEMAN and DONALD M. CROTHERS

The perturbation of DNA topology due to ligand binding is a phenomenon of profound biological importance. Protein-induced DNA bending is an important, in many cases indispensable, factor in the regulation of gene expression, and several drugs utilized in the treatment of cancer derive their therapeutic efficacy from an ability to intercalate DNA, unwinding the double helix in the process. The methodology we describe pertains to the latter class of compounds, namely small organic molecules. It offers the advantage over related methods that D N A topological changes and drug binding constants are measured simultaneously and, perhaps most importantly, under identical reaction conditions. The original intercalation model 1describes a lengthening of the D N A helix and an associated unwinding of the phosphodiester backbone DNA on drug binding in order to accommodate a molecule of intercalator. As this unwinding varies with intercalator, the unwinding angle (¢) is a valuable biophysical parameter when evaluating the characteristics of a compound interacting with DNA. To date, most experimental approaches for measuring ¢ have exploited the special conformational characteristics of closed circular D N A (ccDNA). As isolated from prokaryotes and viruses, ccDNA is negatively supercoiled, that is to say, its writhe (Wr) and linking difference ( A L k ) are negative. *'2'3 The unwinding angle ~b is the change in I L. S. Lerman, J. Mol. Biol. 3, 18-30. * Lk = Tw + Wr, where Tw is the twist about the helical axis and Wr is the writhe or superhelicityof the helical axis through three-dimensionalspace. Whereas Tw and Wr and each geometric properties that vary with environmentalconditions such as temperature and ionic strength, the linking number Lk is a topologicalproperty that is invariantin the absenceof single-strandednicks. ALk is definedas the difference in linking number or superhelicalturns between an idealized fully relaxed, nonligated closed circular DNA for which Wr = 0, and a topoisomer j o f nonzero Wr. Whereas ALk can



[3]


51

to small molecule systems is so new that advances, pitfalls, and other important considerations in this area are likely to develop rapidly.

Acknowledgments Work from our laboratory described in this review has been supported by the NIH. We also acknowledge Farial Tanious for helpful suggestions and data for some figures and Lei Wang for data for some figures. We appreciate preprints from Dr. David Myszka.

[3] Simultaneous Measurement of Binding Constants

and Unwinding Angles by Gel Electrophoresis By STEVEN M. ZEMAN and DONALD M. CROTHERS

The perturbation of DNA topology due to ligand binding is a phenomenon of profound biological importance. Protein-induced DNA bending is an important, in many cases indispensable, factor in the regulation of gene expression, and several drugs utilized in the treatment of cancer derive their therapeutic efficacy from an ability to intercalate DNA, unwinding the double helix in the process. The methodology we describe pertains to the latter class of compounds, namely small organic molecules. It offers the advantage over related methods that D N A topological changes and drug binding constants are measured simultaneously and, perhaps most importantly, under identical reaction conditions. The original intercalation model 1describes a lengthening of the D N A helix and an associated unwinding of the phosphodiester backbone DNA on drug binding in order to accommodate a molecule of intercalator. As this unwinding varies with intercalator, the unwinding angle (¢) is a valuable biophysical parameter when evaluating the characteristics of a compound interacting with DNA. To date, most experimental approaches for measuring ¢ have exploited the special conformational characteristics of closed circular D N A (ccDNA). As isolated from prokaryotes and viruses, ccDNA is negatively supercoiled, that is to say, its writhe (Wr) and linking difference ( A L k ) are negative. *'2'3 The unwinding angle ~b is the change in I L. S. Lerman, J. Mol. Biol. 3, 18-30. * Lk = Tw + Wr, where Tw is the twist about the helical axis and Wr is the writhe or superhelicityof the helical axis through three-dimensionalspace. Whereas Tw and Wr and each geometric properties that vary with environmentalconditions such as temperature and ionic strength, the linking number Lk is a topologicalproperty that is invariantin the absenceof single-strandednicks. ALk is definedas the difference in linking number or superhelicalturns between an idealized fully relaxed, nonligated closed circular DNA for which Wr = 0, and a topoisomer j o f nonzero Wr. Whereas ALk can



52


[3]

twist plus writhe in a linear molecule on ligand binding; it is usually assumed that there is no net change in writhe, in which case q~ is equal to the number of degrees by which the DNA is unwound about its helical axis per ligand molecule bound. The topological properties of DNA can be altered by the enzyme topoisomerase I (Topo I), 4'5 whose discovery opened new experimental avenues for the determination of Lk and ~b.Topo I facilitates thermodynamic equilibration between supercoiled and relaxed forms of ccDNA by creating a transient single-strand nick at which superhelical strain can be relieved, as shown schematically in Fig. IA. Methods for the determination of ~b that employ Topo I and intercalating ligand are based on the processes shown in Fig. 1B-D. On intercalation by a small, usually aromatic molecule (Fig. 1, step a), naturally supercoiled ccDNA is unwound (ATw < 0), so that the writhe becomes less negative at constant Lk (Fig. 1B). For a certain critical amount of bound intercalator (Fig. 1C), all of the negative supercoiling is taken up by negative twist produced by unwinding at drug binding sites, writhe disappears, and the ccDNA exists as a relaxed circle in which the helical axis lies fiat in a plane. Further titration with intercalator past this topological equivalence point (Fig. 1D) continues to unwind the helix, introducing supercoils in an opposite sense, so that Wr is positive. When ccDNA that has been intercalated to some extent is acted on by Topo I (Fig. 1, step b), a fully relaxed species is generated whose ATw reflects the presence of bound intercalator. In this fully relaxed species, Wr = 0. When the intercalator is removed, usually by organic extraction, the resulting change in Tw is offset by compensatory Wr such that Lk remains unchanged. (Lk, the algebraic sum of Tw and Wr, is the number of times the two DNA strands wind about each other, and is an invariant property intrinsic to any nonnicked topoisomer of ccDNA.) The change in Wr on ligand removal manifests itself conformationally as a reappearance of superhelical turns, the magnitude of which is determined by the original amount of intercalator bound at the time of enzymatic religation. Specifically, the linking number deficit - A L k in the final state is equal to the twist change - A T w produced by intercalator binding in step a (Fig. 1).

assume a fractional value because of the idealized reference state, Lk is integral by definition. Excellent treatments of ccDNA topology exist in the literature. 2"3 2 A. D. Bates and A. Maxwell, in "DNA Topology" (D. Rickwood and D. Male, eds.), pp. 17-61. Oxford University Press, Oxford, 1993. 3 N. R. Cozzarelli, T. C. Boles, and J. H. White, in "DNA Topology and Its Biological Effects" (N. R. Cozzarelli and W. C. Wang, eds.), pp. 139-184. Cold Spring Harbor Laboratory Press, Plainview, New York, 1990. 4 j. C. Wang and L. E Liu, (1979) in "Molecular Genetics, Part liP' (J. H. Taylor, ed.), pp. 65-88. Academic Press, New York, 1979. 5 j. C. Wang, in "Nucleic Acid Research" (K. Mizobuchi, I. Watanabe, and J. D. Watson, eds.), pp. 549-566. Academic Press, New York, 1983.

[3]

53


A

B

Wr=-5

Wr=.-5

ATw=0

ATw=0 i ~ _ ~

I

C Wr=-~

D

Ia

ATw=0 i

_

:

Wr=+2

l

c ~Lk=0

~LI(=-2

&l.k=-5

,~J..k=-7

Wr=0 ATw = 0

Wr=-2 ATw = 0

Wr=-5 ATw = 0

Wr=-7 ATw = 0

FIG. I. Schematic topo 1 relaxation reactions of ccDNA at no, low ([I]L), medium ([I]M), and high ([I]H) ligand binding, a, Drug binding; b, topo I relaxation; c, ligand removal by organic extraction. Ligand molecules are represented as nodules along the helical axis of the ccDNA shown, (A) Reaction without ligand, so that the ccDNA product is fully relaxed with ALkc = 0. (B) Reaction containing an amount of ligand that, when bound to the DNA, is sufficient to titrate out two superhelical turns (sht) so that the ccDNA prior to enzymatic relaxation contains - 3 sht. After steps b and c these two titrated sht are manifested in a final ALkc of - 2 . (C) Reaction containing just enough bound ligand to fully unwind all five sht, so that product and starting ccDNAs are indistinguishable. (D) Binding of ligand in high concentration to +2 sht beyond the equivalence point of (C). Here, the ccDNA product bears this additional superhelicity in its ALkc of - 7 . (In reality, some of the superhelical stress is taken up in Tw, so the number of Wr turns is less than the linking number deficit.)

54


[3]

In reality, the resulting ceDNA is not a single species, but rather exists as a family of topoisomers (substantially identical DNA molecules differing relative to one another only in their topology), which differ by integral increments of the linking number. ALkj is defined for topoisomer j as the difference in superhelical turns (Tw + Wr) between an idealized, fully relaxed species for which Wr = 0, and a species j of writhe Wrj. Thermal fluctuations at the time of religation give rise to a Boltzmann distribution of topoisomers around an average ALk. 6-9 This central ALk (ALkc) is taken as the average most abundant topoisomer in the Boltzmann population, and is used in the following method as a comparative value between populations arising from Topo I reaction on ccDNA with varying intercalator concentrations. The simplest method to quantitate such topoisomeric reaction products is resolution by agarose gel electrophoresis. Separation results because the speed of topoisomer migration through an agarose gel is directly proportional to the absolute value of ALk. U n w i n d i n g a s M e a s u r e of E x t e n t of L i g a n d B i n d i n g The present method to calculate ~b and the binding constant Ka relies on resolving topoisomeric reaction products in two electrophoretic dimensions 10.~1 and analyzing the degrees of unwinding in a way formally analogous to the classic treatment of optical binding data.12 The incorporation of a second gel dimension, run perpendicular to the first, resolves topoisomers according to the handedness of their superhelicities, thus enabling resolution of negative and positive ccDNA topoisomers, which otherwise comigrate in just one electrophoretic dimension. Mathematical analysis of reaction products employs the Scatchard binding isotherm adjusted for nearest neighbor exclusion by bound ligand) 3~14expressed in closed formlS: I__/__"= Ka( 1 _ nr)F CF

1 - nr

],,-1

k 1 -- (-nT-1)r J

(1)

6 W. Keller, Proc. Natl. Acad. Sci. U.S.A. 72, 4876 (1975). 7 M. Shure and J. Vinograd, Cell 8, 215 (1976). 8 D. E. Depew and J. C. Wang, Proc. Natl. Acad. Sci. U.S.A. 72, 4275 (1975). 9 D. E. Pulleyblank, M. Shure, D. Tang, J. Vinograd, and H.-P. Vosberg, Proc. Natl. Acad. Sci. U.S.A. 72, 4280 (1975). 10 R. Bowater, E Aboul-Ela, and D. M. Lilley, Methods Enzymol. 212, 105 (1992). I I C.-H. Lee, H. Mizusawa, and T. Kakefuda, Proc. Natl. Acad. Sci. U.S.A. 78, 2838 (1981). 12 G. Scatchard, Ann. N.Y. Acad. Sci. 51, 660 (1949). 13 D. M. Crothers, Biopolymers 6, 575 (1968). 14 W. Mfiller and D. M. Crothers, J. Mol. Biol. 35, 251 (1968). 15 j. D. McGhee and P. H. Von Hippel, J. Mol. Biol. 86, 469 (1974).

[3]


55

where CF is the free ligand concentration, 1"is the ratio of bound ligand to base pairs, Ka is the intrinsic association constant under the chosen conditions, and n is the number of sites (normally base pairs) covered by one molecule of bound ligand. Equation (1) simplifies to linear form at low binding when r 0 (sin 45 ° > 0). If the arrows instead are antiparallel, that is, forming an angle of 135 °, we have an attractive interaction (sin 135 ° < 0). Repulsive orientation corresponds to the exciton couplet seen at the high energy end of the spectrum, that is, the band at the short-wavelength side of the absorption maximum of the monomer drug absorption peak. For right-handed DNA this component should display a positive CD and a negative LD, which may be deduced as follows. 2. LD is determined by the resultant vector of the two vectors (~A + ~B) in Eq. (III.3), in this case a vector bisecting the angle between the two drug units, that is, symmetrically in between 90 and 45 ° to the helix axis. We thus obtain the angle 67.5 °, and according to Eq. (5) we thus conclude a negative LD for the short-wavelength band. For the case of oppositely directed arrows--the attractive coupling, that is, the exciton band at longer wavelength than the monomer--one instead finds the vector (]..La -- ~ B ) oriented at 157.5 ° relative to the helix axis, and thus displaying a positive LD signal. 3. To deduce the sign of the CD, we again consider the resultants of the two /ZA and/~B vectros, and start with the in-phase interaction observed at high energy, lt~A -]- jt~B. CD is positive if any component of the magnetic moment created by circulation of charge around this direction is parallel, and negative if the magnetic moment is antiparallel to this direction. With the anticipated arrangement in the groove of a right-handed DNA molecule, we find that/~A and/~a will create an anticlockwise rotation of charge around the vector ~A +/ZB. Using the right hand and pointing with the index finger in the direction of the current, the thumb will point in the direction of the magnetic moment. Because/ZA + ~B and m are thus antiparallel, the CD signal will be negative. Correspondingly, for the out-of-phase component, (/~A --/~B) and m will be parallel, so the CD at the long-wavelength side will show positive intensity. So with the dimer bound to right-handed DNA in the fashion assumed, a bisignate exciton CD spectrum should appear having a negative peak at short wavelength and a positive peak at long wavelength.

[5]

RAPID SCREENING OF LIGAND-DNA BINDING

99

[5] Rapid Screening of Structurally Selective Ligand Binding to Nucleic Acids B y JINSONG REN and JONATHAN B. CHAIRES

Introduction DNA is polymorphic, and exists in a variety of secondary and tertiary structures that are dictated by both sequence and environmental conditions.1 Unique DNA structures represent potential targets for small molecules, and provide a promising new avenue for drug development. 2,3 Small molecules that selectively bind to a particular structure could interfere with whatever biological function involves that structure. For example, multistranded triplex and tetraplex nucleic acid structures have attracted considerable attention as targets for small molecule. 2 6 Attempts to rationally design small molecules that bind selectively to a particular structure are hampered by the lack of a rapid and convenient assay for structural selectivity. Typically, thermal denaturation studies monitored by UV absorbance are used, but such an assay is both time consuming and incapable of screening a large number of different structures simultaneously. In addition, melting curves in the presence of bound ligand are often complex and multiphasic, 7-9 making the analysis and interpretation of data difficult. Alternatively, binding constants for the interaction of ligands with the target structure can be measured and compared with standard duplex DNA, but such measurements are even more time consuming and tedious than are those involved in obtaining melting curves. To circumvent these difficulties, we have devised a novel competition dialysis assay for rapidly screening ligand binding to a variety of nucleic acid structures) °- 12 Practical details of the assay are described here, along with examples of experimental data.

I S. Neidle, "Oxford Handbook of Nucleic Acid Structure." Oxford University Press, New York, 1999. 2 j. L. Mergny and C. Helene, Nat. Med. 4, 1366 (1998). 3 T. C. Jenkins, Curl: Med. Chem. 7, 99 (2000). 4 j. L. Mergny,G. Duval-Valentin,C. H. Nguyen, L. Perrouault,B. Faucon, M. Rougee, T. MontenayGarestier, and E. Bisagni, Science 256, 1681 (1992). 5 L. H. Hurley and E L. Boyd, Trends Pharmacol. Sci. 9, 402 (1988). 6 L. H. Hurley, J. Med. Chem. 32, 2027 (1989). 7 L. Cai, L. Chen, S. Raghavan, R. Ratliff, R. Moyzis, and A. Rich, Nucleic Acids Res. 26, 4696 (1998). 8 D. M. Crothers, Biopolymers 10, 2147 (1971). 9 j. D. McGhee, Biopolymers 15, 1345 (1976). l0 j. Ren and J. B. Chaires, Biochemistry 38, 16067 (1999). 11j. Ren and J. B. Chaires, J. Am. Chem. Soc. 122, 424 (2000). 12j. Ren, C. Bailly, and J. B. Chaires, FEBS Lett. 470, 355 (2000).

METHODS 1N ENZYMOLOGY, VOL. 340


100


[5]

Principle of Method and Historical Background The competition dialysis assay is based on the fundamental thermodynamic principle of equilibrium dialysis.13 The approach is easily summarized. A macromolecule solution is placed inside a semipermeable dialysis membrane with pore sizes that allow small molecules to pass through, but which retain the large macromolecules. The dialysis unit is then suspended in a solution containing small molecule ligand. At equilibrium, the chemical potential of the free ligand must be equal inside and outside the dialysis unit, and any excess ligand on the macromolecule side of the membrane may be attributed to binding to the macromolecule. Equilibrium dialysis has been widely used to measure the binding of small molecules or ions to macromolecules to provide the primary data for the determination of accurate binding constants.13.14 Muller and Crothers first introduced an important variant of the equilibrium dialysis experiment, the competition dialysis procedure.15 That procedure was first used to evaluate the base specificity of drug-DNA interactions. The procedure utilized a three-part dialysis chamber, in which two dialysis membranes separated the central chamber from two outside chambers. Two DNA samples of identical concentration but of differing base composition were placed in each of the two outer chambers. A ligand solution was placed in the central chamber, and allowed to equilibrate among the three chambers. If the ligand bound selectively to AT or GC base pairs, more ligand would accumulate in the chamber containing the DNA with the base composition containing the highest percentage of the preferred base. Rather detailed quantitative inferences concerning the nature of the preferred binding site are possible by this method using simple probability concepts.15'16 This version of the competition dialysis method unfortunately saw little widespread application, and was supplanted by footprinting methods that provide a higher resolution glimpse into sequence selective binding. Competition dialysis was, however, adapted to provide a tool for probing structural selective ligand binding to nucleic acids. Becker and Dervan used the approach to show selective binding of bis(methidum)spermine to a D N A - R N A hybrid. 17 Chaires and coworkers used the method as a probe for ligand binding to left- and right-handed DNAJ s'19 More recently, the method was used to probe ligand binding to duplex and triplex DNA. 2° The assay described here is a simple and straightforward extension of the Crothers competition dialysis method. Instead of a three-chambered dialysis 13K. E. vanHolde, W. C. Johnson, and R S. Ho, "Principles of Physical Biochemistry."Prentice Hall, Upper Saddle River, New Jersey, 1998. 14I. M. Klotz, "Ligand Receptor Energetics." John Wiley & Sons, New York, 1997. 15W. Muller and D. M. Crothers, Eur. ,l. Biochem. 54, 267 (1975). 16j. B. Chaires, in "Advances in DNA Sequence Specific Agents," Vol. l, pp. 3-23. (L. H. Hurley, ed.). JAI Press, Greenwich, Connecticut, 1992. 17M. M. Becker and E B. Dervan, J. Am. Chem. Soc. 101, 3664 (1979).

[5]

RAPID SCREENING OF L I G A N D - - D N A BINDING

10 1

lyzer g DNA

I#M Free tigand

Stir bar FIG. 1. Schematic diagram of the experimental arrangement for the competition dialysis assay. Disposable dialysis units (either Spectrum DispoDialyzers or Pierce Slide-A-Lyzers) are used to contain the desired nucleic acid structures. Structures are dialyzed against a common test ligand solution to reach equilibrium. The amount of ligand bound to each structure is then measured by absorbance spectroscopy or fluorescence.

apparatus, disposable dialysis units are used to contain a wide variety of nucleic acid structures at identical concentrations. These units are simply placed into a beaker containing ligand solution, as illustrated in Fig. 1. At equilibrium, the free ligand concentration is identical throughout the system, and preferential binding by a particular structure leads to a greater accumulation of total ligand within that dialysis unit. Structural selectivity can be easily measured by measuring total ligand concentration within each dialysis unit. The challenge to be met is to find a suitable buffer in which the nucleic acid structures of interest are all stable. In the first-generation assay, l° 13 structures were found to be stable in a simple phosphate buffer containing 200 mM NaCI. The assay was expanded to include 19 structures in the second-generation assay l~ using the same buffer conditions. Description of these structures and a detailed protocol for the assay follow.

18 j. B. Chaires, Biochemistry 24, 7479 (1985). 19 S. Satyanarayana, J. C. Dabrowiak, and J. B. Chaires, Biochemistry 32, 2573 (1993). 2(I I. Haq, J. E. Ladbury, B. Z. Chowdhry, and T. C. Jenkins, J. Ant. Chem. Soc. 118, 10693 (1996).

102


[5]

Experimental Method

Overview of Assay The competition dialysis assay is straightforward and simple. First, stock solutions of the structures of interest are prepared and characterized. Second, the assay is set up for the study of a ligand of interest by filling disposable dialysis units with solutions of the various structures at identical concentrations, placing these in a common ligand solution, and allowing the system to reach dialysis equilibrium. Third, concentrations of total ligand within each dialysis unit are determined by absorbance or fluorescence spectroscopy (or any other convenient analytical tool). Finally, results are graphed for analysis and interpretation.

Materials and Sample Preparation Dialysis Units. The first-generation assay used 1.0- or 0.5-ml DispoDialyzer units (Spectrum Medical Industries, Rancho Domingneg, CA; www.spectrumlabs. com) of the desired molecular weight cutoff (MWCO), usually 3500 MWCO. Subsequently, a less expensive and overall more suitable alternative was found, Pierce Slide-A-Lyzer MINI dialyzer units (Pierce, Rockford, IL, www.piercenet.com). Pierce proved to be helpful and cooperative in providing technical information about their product, and is the recommended source for dialysis units for the assay. The quality of each minidialysis unit is evaluated before experiments by dialysis against water to ensure against leaks. In that test, 200 #1 of water is added to each dialysis unit, which are then dialyzed against water for 24 hr. The volume of solution within each dialysis unit is then measured. Units with excess volumes of water that exceed the original 200/zl should be discarded. Dialysis units may be reused several times after careful rinsing with buffer solution. Nucleic Acids. The nucleic acid structures used in the current version of the competition dialysis assay are shown schematically in Fig. 2 and listed in Table I. All samples are dissolved in BPES buffer, consisting of 6 mM NazHPO4, 2 mM NaHzPO4, 1 mM NazEDTA, 0.185 M NaC1 (pH 7.0). Samples of DNA from ClostHdium perfringens, calf. thymus, and Micrococcus lysodeikticus are purchased from Sigma (St. Louis, MO), and are sonicated, phenol extracted, and purified as previously described. 21 Poly(dA), poly(dT), poly(U), poly(dC), poly(dA)-poly(dT), poly(dA-dT), and poly(dG-dC) are purchased from Pharmacia Biotech (Piscataway, N J). Poly(rA) and poly(A)-poly(U) are purchased from Sigma. Deoxyoligonucleotides 5'-T2G20T2, 5'-G10TaGI0, and 5'-AG3(TTAG3)3 are purchased from Research Genetics (Huntsville, AL). Synthetic single-stranded and duplex polynucleotides are used without further purification. The poly(rA)-poly(dT) DNA-RNA hybrid is prepared by mixing 21j. B. Chaires, N. Dattagupta,and D. M. Crothers,Biochemistry21, 3933 (1982).

[5]

RAPID SCREENING OF L I G A N D - D N A BINDING

1

103

Form

Samples

Single-strand

poly dA, poly dT poly A, poly U

Duplex

natural DNA, Z-DNA RNA, DNA:RNA hybrid, polynucleotides,

Triplex

poly dA:[poly dTL (DNA) polyA:[poly U]2 (RNA)

Tetraplex

[T2G2oT]. A(G3I-IA)3G3 [G,oT4G,o]. [Poly dC].

(Parallel) (Folded) (G-wire) (i-motif)

FIG. 2. Schematic of structures used in the current generation of the competition dialysis assay.

poly(rA) and poly(dT) in a 1 : 1 molar ratio, heating to 90 °, and slowly cooling to room temperature. The DNA and RNA triplex forms are prepared by mixing poly(dA)-poly(dT) with poly(dT) [or poly(A)-poly(U) with poly(U)] in a 1:1 molar ratio, heating to 90 °, and slowly cooling to room temperature. Tetraplex DNA and i-motif DNA are prepared by heating the oligonucleotide 5'-T2G20T2, 5'-Gl0TaGl0, 5'-A(G3TTA)3G3, or poly(dC) to 90 ° for 2 min, slowly cooling to room temperature and then equilibrating for 48 hr at 4 ° before use. Left-handed Z-DNA is prepared by bromination of poly(dG-dC) as previously describedY Concentration Determinations. Concentrations of nucleic acid samples are determined by UV absorbance measurements, using the extinction coefficients and absorption maxima listed in Table I. Stock solutions of nucleic acids are prepared and maintained at 75/zM concentration, using the monomeric unit of each polynucleotide as the concentration standard: nucleotides (nt) for singlestranded forms, base pairs (bp) for duplex forms, triplets for triplex forms, and quartets for tetraplex forms. The extinction coefficients listed in Table I all refer to these concentration standards. Sample Quality Control. All nucleic acid preparations are characterized by circular dichroism, using a JASCO (Tokyo, Japan) J500A instrument, and by thermal 22 A. Moiler, A. Nordheim, S. A. Kozlowski, D. J. Patel, and A. Rich, Biochemistry 23, 54 (1984).

104


[5]

TABLE I NUCLEIC ACID CONFORMATION AND SAMPLES USED IN COMPETITION DIALYSIS EXPERIMENTSa

Conformation

DNA/oligonucleotide

Single-strandpurine

Poly (dA) Poly(A) Single-strandpyrimidine Poly(dT) Poly(U) Duplex DNA CIostridium po~[Hngens (31% GC) Calf thymus (42% GC)

DNA-RNAh y b r i d Duplex RNA Z-DNAb Triplex DNA Triplex RNA Tetraplex DNA 1 Tetraplex DNA 2 Tetraplex DNA 3 i-motif

X(nm) E (M Icm I)

T~,~('C)

%H

---

257 258 264 260 260

8,600 9,800 8,520 9,350 12,476

-82.5

1.24 1.29 --1.28

Miclvcoccus lysodeikticus

260 260

12,824 13,846

85.5 > 100

1.36 1.16

(72% GC) Poly(dA)-poly(dT) Poly(dA-dT) Poly(dG-dC) Poly(rA)-poly(dT) Poly(A)-poly(U) Br-poly(dG-dC) Poly(dA)-[poly(dT)]2 Poly(A)[poly(U)]2 (5'-T2G20T2)4 5'-A(G3TTA)3G~ 5'-(Gl0T4GI0)x [Poly(dC)]4

260 262 254 260 260 254 260 260 260 260 260 274

12,000 13,200 16,800 12,460 14,280 16,060 17,200 17,840 39,267 73,000 39,400 27,200

75.2 67.8 > 100 70.9 62.5 > 100 42.5 62.5 89.0 66.8 89.0 46.9; 55.0

1.58 1.53 1.49 1.54 1.59 1.67 1.06 1.04 1.02 1.38

'~x, Wavelength;e, molar extinctioncoefficient;Tin,meltingtemperature;%H, hyperchromicity. ~'A.Moiler,A. Nordheim, S. J. Kozloski,and A. Rich, Biochemistry23, 56 (1984).

denaturation monitored by UV absorbance. Denaturation studies are done with a Varian (Palo Alto, CA) Cary 3E spectrophotometer equipped with a Peltier temperature control unit. Charateristic melting temperatures and hyperchromicity values are listed in Table I. Spectra and melting curves for all samples are available as supplementary materials to published articles.l°' 11

Competition Dialysis Protocol For each competition dialysis assay, 400 ml of the dialysate solution containing 1 /zM test ligand concentration is placed into a beaker. A volume of 180 #1 (at 75 # M monomeric unit) of each of the DNA samples listed in Table I is pipetted into a separate Slide-A-Lyzer MINI dialysis unit with 7000 M W C O membrane. All 19 dialysis units are placed in a MINI dialysis flotation device (Pierce) and then the whole unit is placed in the beaker containing the dialysate solution. The beaker is covered with Parafilm, wrapped in foil, and allowed to equilibrate with continuous stirring for 24 hr at room temperature (20-22°). At the end of the

[5]

RAPID SCREENINGOF LIGAND--DNA BINDING

105

equilibration period, DNA samples are carefully collected in the comer of the MINI dialysis unit and transferred to microcentrifuge tubes. The DNA samples are then made to a final concentration of 1% (w/v) sodium dodecyl sulfate (SDS) by the addition of appropriate volumes of a 10% (w/v) stock solution. The total concentration of drug (Ct) with each MINI dialysis unit is then determined spectrophotometrically, using wavelengths and extinction appropriate for each ligand. An appropriate correction for the slight dilution of the same resulting from the addition of the stock SDS solution is made. The free ligand concentration (C0 is determined spectrophotometrically with an aliquot of the dialysate solution, although its concentration usually does not vary appreciably from the initial 1 #M concentration. The amount of bound drug is determined as Cb = Ct - C f . Data are plotted as a bar graph, using Origin software (version 5.1 ; Microcal, Northampton, MA) for plotting and analysis. In the initial experiments, and on occasion thereafter, the competition dialysis method is checked by a detailed mass-balance accounting of total ligand, using the procedures described in detail by Klotz.]4

Results and Interpretation Figure 3 shows results obtained with the standard intercalator ethidium. Resuits from both the first-generation (Fig. 3A) and second-generation (Fig. 3B) assays are shown. Additional structures were included in the second-generation assay. These results compare binding to 19 different structures. No binding is observed to any single-stranded DNA or RNA forms [poly(A), poly(U), poly(dA), and poly(dT)]. Duplex DNA, both from natural sources or synthetic polynucleotides,

M tyso DNA CT DNA

M lyso DNA CT DNA C perf DNA polydT polydA polyU polyA

C perl DNA poly dA poly dT 0

3

6

[Bound]~M

9

o

3

6

[Bound]pM

FIG.3. Resultsobtainedforthe simpleintercalatorethidium,usingthe first-generationassay(A) that included 13differentnucleicacid structures,or the expandedsecond-generationassay(B) that included the 19 structures listed in TableI.

106


-3

0

C ¢- Cr0NA b'--b

S

4

[5]

•

3

4J

Cb.

FIG. 4. Alternate graphical representations of experimental data. Data obtained by the secondgeneration assay for ethidium (Fig. 3B) are replotted in alternate forms. (A) The difference in the amount bound relative to the amount bound to duplex calf thymus (CT) DNA is shown. (B) The difference relative to the arithmetic mean bound to all structural forms is shown.

binds ethidium well. Poly(dA)-poly(dT), however, binds ethidium only weakly, a result that arises from the unusual structure adopted by the polymer in solution. 23 25 RNA [poly(A)-poly(U)] and the DNA-RNA hybrid (poly(rA)-poly(dT) show the most ethidium binding, a finding consistent with more laborious titration binding studies. 26 Left-handed Z-DNA shows only weak binding, although ethidium can aliosterically convert the polynucleotide to an intercalated right-handed form at higher-ligand concentrations, z7 DNA and RNA triplex forms bind ethidium about as well as duplex forms, consistent with equilibrium binding studies. 28 Finally, ethidium binds weakly to G-tetraplex structures, and not at all to the unusual i-motif. The results of Fig. 3 provide a rapid comparison of ethidium binding to 19 different structures, and provide a sound basis for selecting particular structures for more detailed biophysical or biochemical studies. Figure 4 shows two of many possible alternate graphical representations of competition dialysis data, using the results obtained with ethidium in the secondgeneration assay (Fig. 3B). In Fig. 4A, the difference in the amount of ethidium bound to a particular structure relative to that bound to calf thymus DNA is shown. 23 S. M. He, Y. C. Sun, S. L. Wei, C. B. Wu, E U. Wang, S. D. Wang, and J. X. Xie, Yao Hsueh Hsueh Pao 28, 859 (1993). 24 j. E. Herrera and J. B. Chaires, Biochemistry 28, 1993 (1989). 25 S. S. Chan, K. J. Breslauer, M. E. Hogan, O. J. Kessler, R. H. Austin, J. Ojemann, J. M. Passner, and N. C. Wiles, Biochemistry 29, 6161 (1990). 26 j. L. Bresloff and D. M. Crothers, Biochemistry 20, 3547 (1981). 27 G. T. Walker, M. P. Stone, and T. R. Krugh, Biochemistry 24, 7462 (1985). 28 p. V. Scaria and R. H. Shafer, J. Biol. Chem. 266, 5417 (1991).

[5]

RAPID SCREENING OF LIGAND--DNA BINDING

107

Calf thymus DNA represents a standard duplex DNA, so this plot emphasizes differences relative to that standard. Positive values of the difference in the amount bound clearly indicate preferential binding relative to calf thymus DNA, while negative values indicate poorer binding. Figure 4B shows data plotted as the difference from the mean value of binding to all structural forms. In that plot, the difference is relative to the simple arithmetic mean, with positive values indicative of preferential binding. Other manipulations of the data are possible and may be desirable on occasion. Perhaps most useful is calculation of the apparent binding constant, Kapp, which is simply given by the equation Cb Kapp -- Cf x [NA]total

where Cb and Cf are the bound and free ligand concentrations, respectively, and [NA]total is the total nucleic acid concentration. Typically, in our experimental setup, Cf = 1 # M and [NA]total = 75/.zM. Cb is the experimentally measured amount of ligand bound to each form. This treatment assumes that potential binding sites are in excess, so possible neighbor exclusion effects can be neglected. The units of Kapp would refer to the monomeric unit (nucleotide, base pair, etc.) of each polymer. Values of Kapp determined from competition dialysis were found to be in excellent agreement with binding constants obtained from more complete titration studies.12 The primary utility and the power of the competition dialysis assay are illustrated in Fig. 5, in which the structural selectivity of four different compounds is compared. The compounds are 9-aminoacridine, quinacrine, methylene blue, and telomerase inhibitor V (Fig. 5D). At a glance, the distinct structural preferences of these compounds are evident. 9-Aminoacridine (Fig. 5A) is a well-known intercalator, and shows the expected binding to duplex DNA structures. It has a strong preference, however, for triplex DNA and binds well to the parallel-stranded, "G-wire" tetraplex structure (tetraplex 3). Quinacrine (Fig. 5B) shows an even stronger preference for triplex DNA. Methylene blue (Fig. 5C) shows a strong preference for multistranded DNA structures, and binds well to all tetraplex forms and to triplex DNA. Finally, telomerase inhibitor V shows a similar preference for multistranded structures, but in contrast shows appreciable binding to certain duplex forms, poly(dA-dT) and the DNA-RNA hybrid. Telomerase inhibitor V is one of the few compounds assayed that binds appreciably to single-stranded forms, in this case poly(A). The few examples shown here are the tip of a rather large iceberg of experimental results. We have efficiently and rapidly screened more than 150 compounds to date by the competition dialysis assay. The results obtained with these compounds provide a comprehensive view of structural selectivity from which principles will emerge that should serve to guide the rational design process for

108


[5]

~ CI~

2

4

[Bound] g M

s

e

v

CH~ NHCI~(CH2)3N(C2HJz OCH3 -x~

o

3

s

O

3

IS [Bound]

v

9

9

12

12

g M

FIG. 5. Competition dialysis results obtained with (A) 9-aminoacridine, (B) quinacrine, (C) methylene blue, and (D) telomerase inhibitor V (2,6-bis[3-(N-piperidino)propionamido]anthracene9,10-dione; Calbiochem, La Jolla, CA).

new small molecules that can target specific structures. So far, detailed results on several compounds have been published,l°-12 and descriptions of many more are in preparation. We hope that the description of the protocol presented here will allow other laboratories to implement the method as a routine part of their studies of drug-nucleic acid interactions. Acknowledgment This work was supported by grant CA35635 from the National Cancer Institute.

[7]

VOLUME EFFECTS OF DRUG-NUCLEICACID INTERACTIONS

149

toward high-order nucleic acid system. The purpose of this chapter has been (1) to illustrate the importance and power of calorimetry as a biophysical tool, and (2) to outline certain practical considerations in the use of ITC and DSC methods for DNA triplex/tetraplex stability and binding studies. It must be stressed that, if used properly and within limits, calorimetry represents a powerful and versatile array of techniques that can yield great insight into the energetics of biomolecular stability and binding processes. The genuine strength of calorimetry becomes apparent when it is used in conjunction with parallel techniques such as UV-visible/CD/fluorescence spectrophotometry, NMR spectroscopy, X-ray crystallography, and stopped-flow kinetic methods. Such a multitechnique approach to studies of DNA-drug interactions (or indeed any binding interaction) can provide a comprehensive and cohesive picture for the underlying biomolecular events. Acknowledgments We thank Ms. Maria Halla (Universityof Greenwich) for conducting all the DSC experiments with the G2 tetraplex. Financial support for the authors' laboratories is generouslyprovidedby the Universityof Sheffield(to I.H.)and YorkshireCancerResearch(to T.C.J.)and Universityof Greenwich (to BZC). I.H. is an EPSRC AdvancedResearchFellow.

[7] Volume Changes Accompanying Interaction of Ligands with Nucleic Acids By LUIS A. MARKY, DONALD W. KUPKE, and BESIK I. KANKIA Introduction The biological function of nucleic acids is carried out through interactions with other molecules, specifically with proteins. To completely understand the expression of genes, it is imperative to obtain a complete description of the structure, thermodynamics, and kinetics of these interacting systems. In the specific case of the energetics of protein-nucleic acid complex formation, few thermodynamic investigations have been carried out because of the limited solubility of proteins. Alternatively, the energetic contributions of the specific molecular interactions observed in the structure of the protein-nucleic acid complex can be obtained by measuring complete thermodynamic profiles of the interaction of small molecules with nucleic acids. At the same time, some of these small molecules, called ligands or "drugs," interact with nucleic acids to prevent replication and transcription. The extent of their biological activity depends on several factors, such as binding affinity, specificity, and binding mode. The literature is full of reports on the molecular



150


[7]

forces involved in these types of association reactions, but little is known of the hydration effects that accompany these reactions or the association of two smaller solutes. We are finding answers to the following questions: how does water hydrate a small or large solute? What are the physical properties of water around charged, polar, or nonpolar groups? The measurement of the volume change, AV, for association reactions with a proper interpretation usually yields information about these hydration effects. Water plays a fundamental role in the stability of the secondary and tertiary structures of proteins and nucleic acids. 1.2 In particular, the nucleic acid molecule is well hydrated and its hydration depends on the nucleic acid conformation and composition.l'3 7 The hydration properties provide a rationale for the conformational plasticity (A, B, or Z) of DNA; for example, slight changes of the water activity can shift the equilibrium to favor one duplex conformation over others.I At the present time, the measurement of hydration changes has been achieved with the development of high-sensitivity density techniques, which make it possible to determine the change in volume of samples containing 1 mg or less of solute. 8'9 However, the interpretation of results remains sometimes difficult because of the existence of two types of water that are hydrating a particular macromolecule or complex: water of electrostriction around charged and polar groups and hydrophobic water around nonpolar or organic groups. For example, a release of water molecules is expected to be measured in the formation of a simple complex AB from the reaction of its component A and B solutes and to a first approximation, which is due to the overlap of hydration shells of the solutes in the complex. In biochemical systems at constant temperature and pressure, the volume expansions and contractions attending reactive chemical changes reflect the average change in the molar volume of water in the vicinity of the participating solutes. The solutes themselves ordinarily do not possess the propensity to change the solution volume to the same order of magnitude (a relatively large net difference in the amount of void spaces within globular macromolecules perhaps being an exception). Liquid water by virtue of its quadrupolar nature [and possessing a large dipole moment of 1.85 Debye (D) units, or more in the liquid phase] has the capacity to change its molar volume dramatically. The open, quasi-tetrahedral structure of bulk water I W. Saenger, "Principles of Nucleic Acid Structure." Springer-Verlag, New York, 1984. 2 C. Tanford, Adv. Protein Chem. 23, 121 (1968). 3 M.-J. B. Tunis and J. E. Hearst, Biopolymers 6, 1325 (1968). 4 L. A. Marky and D. W. Kupke, Biochemistry 28, 9982 (1989). 5 E. Westhof, Annu. Rev. Phys. Chem. 17, 125 (1988). 6 V. A. Buckin, B. I. Kankiya, N. V. Bulichov, A. V. Lebedev, V. P. Gukovsky, A. P. Sarvazyan, and A. R. Williams, Nature (London) 340, 321 (1989). 7 W. A. Buckin, B. I. Kankiya, A. P. Sarvazyan, and H. Uedaira, Nucleic Acids Res. 17, 4189 (1989). 8 G. T. Gillies and D. W. Kupke, Rev. Sci. lnstrum. 59, 307 (1988). 9 D. W. Kupke and J. W. Beams, Methods Enzymol. 26, 74 (1972).

[7]

VOLUME EFFECTS OF DRUG--NUCLEICACID INTERACTIONS

151

is some 50% larger in molar volume than if the individual molecules were packed in a noninteracting hexagonal array.l° Accordingly, ion-water dipole interactions are likely to be the major sources of biochemical volume changes (at constant temperature and pressure). The intense electric field close to an ion induces a local collapse of the bulk water structure by compressing the adjacent water dipoles at pressures corresponding to several thousand atmospheres, conditions that may even reduce the molar volume of water to "-~12 ml/mol, lO This compression, which depends on the charge and radius of an ion and on the local dielectric gradient, is known colloquially as electrostriction, j°'l] Hence, neutralization reactions, formation of ion pairs (salt bridges), and the coordination of metal ions result in water decompressions that can give rise to relatively large expansions at millimolar reactant concentrations. 12 The changes in molar volume of water molecules interacting over much shorter effective distances (dipole~lipole, hydrophobic-dipole, partial charge-dipole, etc.) are less well understood. According to electrostriction theory, we would expect these compressions to be comparatively small. Also, it is not clear whether the interaction of organic groups with water can in some cases cause a small expansion relative to bulk water. ]3 The volume effect of hydrophobic moieties on charged solutes is known to augment the contractions 14'15; however, unless these moieties undergo chemical changes during mixing of aqueous phases, the difference in volume from this source should contribute only a marginal effect. The electrostriction effect for hydrogen bonding in nucleic acid base pairing, between the partially charged nitrogen and oxygen donor/acceptor atoms, is predicted to be small because the contraction is approximately proportional to the square of the charge density. 16 Moreover, the volume change attending the loss of hydrogen-bonding water to these sites prior to the formation of a duplex would tend to compensate for the putative contraction accompanying hydrogen bond formation between the bases. It appears, therefore, that an uptake of water molecules is observed in the formation of a nucleic acid duplex, from the mixing of its complementary strands. 17,18This is because the resulting volume change is interpreted to reflect the compensating effects that take place between changes in electrostriction of the duplex water quadrupoles and changes of hydrophobically bound water of the single strands. 10B. E. Conway,"Ionic Hydration in Chemistry and Biophysics."Elsevier, New York, 1981. ] I E J. Millero, in "Waterand AqueousSolutions"(R. A. Home, ed.), p. 519. Wiley-lnterscience,New York, 1972. 12D. W. Kupke and B. S. Shanck,,I. Phys. Chem. 93, 2101 (1989). 13E H. Stillinger, Science 209, 451 (1980). 14E. J. King,J. Phys. Chem. 73, 1220 (1969). ]5 j. E. Desnoyers,Phys. Chem. Liq. 7, 63 (1977). 1rE J. Millero, Chem. Rev. 71,147 (1971). 17K. Zieba, T. M. Chu, D. W. Kupke, and L. A. Marky, Biochemistly 30, 8018 (1991). 18L. A. Marky and D. W. Kupke, Methods Enzymol. 323, 419 (2000).

152


[71

In this chapter, we initially provide a brief theoretical and experimental description of current techniques for the measurement of volume changes; we then illustrate the applicability of this approach for the interaction of ligands with nucleic acids. We provide specific examples from data obtained in our laboratories; these include ligands with different physical binding modes, the contribution of nucleic acid conformation, bent deoxyoligonucleotides, and the formation of oligomer duplexes with chemically attached ligands. M e a s u r i n g V o l u m e C h a n g e for A s s o c i a t i o n R e a c t i o n There are several techniques that are currently in use for the measurement of the volume change, 2xV, for an association reaction. Model-dependent techniques are based on the dependence of the equilibrium constant on hydrostatic pressure; model-independent techniques include direct measurements by dilatometry and indirect measurements of the mass and density of reactants and products of a particular reaction.

Dependence of K on Pressure The volume change is determined by the following relationship: A V = (O A G /

Op)T, substitution of AG(---- -RTlnK) yields AV = -RT(O In K/Op)T, where R is the gas constant and T is the temperature. Thus, the volume change is obtained from the dependence of the equilibrium constant, K, on hydrostatic pressure and may be monitored by optical techniques. The measurement involves the placement of a sample solution in a quartz cell contained in an optical high-pressure cell. The concentration of reactants and products is measured optically, using standard absorption or fluorescence techniques, as a function of hydrostatic pressure in the range of 1 to 3000 atm ( t 0 . 1 - 3 0 0 MPa). In this method, it is assumed that the AV is independent of pressure, that is, (OAV/Op)T equal to zero. This is equivalent to an optical measurement of model-dependent enthalpies, A HuH, using the van't Hoff equation AH~n = RT2(OIn K/OT)p, in which it is assumed that AH~n is independent of temperature or ACp is zero.

Dilatometry Dilatometry is a straightforward method for A V measurements. In dilatometry, ~ 6 ml of solution of each reactant is placed in each arm of an inverted "Y" glass tube with arms widely separated; a water-immiscible liquid, for example, heptane, proved to be inert, is placed above the reactants to separate them initially and fill up the entire tube. This tube is closed with a capillary tube and the level of the organic liquid is within the linear scale of the capillary in millimeters. The whole tube is placed in a water bath with a temperature control of 4- 0.001 ° and the height of the capillary is read after reaching temperature equilibrium. Then, the reactants

[7]

VOLUME EFFECTS OF DRUG-NUCLEIC ACID INTERACTIONS

153

are mixed by moving the dilatometer sideways without removing it from the bath. The height of the liquid in the capillary is read again and the difference in this height, negative or positive, is proportional to the extent of the volume contraction or expansion of the particular reaction. The main disadvantage of this technique is the large amount of materials needed for each independent measurement.

Indirect Methods The AV value is obtained from the measurements of the mass and the equilibrium density of solutions before and after mixing. The masses are obtained with a microbalance, and the densities are obtained with a magnetic suspension densimeter or with an Anton Paar (Graz Austria) densimeter. The observed change in volume, Av, on adding reactant A to reactant B is given by A1) = Pproducts -- 1)reactants = mmix/Pmix -- (mA/PA -}- mB/PB)

(1)

where m is the mass in grams and p is the density of the solution in grams per milliliter. The sum of the two terms within parentheses gives the initial volume before mixing. The value of Av in milliliters is then normalized per mole of the limiting reagent to give AV. Our laboratories use this procedure extensively; therefore, we describe in the following sections the techniques in use for the density measurements.

Magnetic Suspension Densimetry The instrument used in this study is an improved version of the earlier design for biochemical purposes by Senterl9and has been previously described, s This instrument requires only 100 #1 per measurement, and its sensitivity is such that it can deliver volume differences with a precision of less than half a nanoliter. This densimeter is calibrated with aqueous KC1 solutions of known density. The density of each sample is obtained by relating the measured voltage to the straightline calibration equation of voltage versus density. This is because the electrical current required to stabilize the tiny permanent magnet jacketed at a present height below the meniscus is directly proportional to the density of the surrounding fluid. 9 According to Eq. (1), small weighing errors have no appreciable effect on A v and the three density values, while independent, need not be of high absolute accuracy because it is their differences that are required for A v. The density is measured with a precision of 5 × 10 - 6 g/ml. The temperature is always kept within 4-0.001 °. In these experiments, the concentration of nucleic acid samples (in phosphate) ranges from 2 to 3 mM; thus, the effect of solute-solute interactions on the volume property is deemed to be negligible. 2° Usually, equal volumes of solutions of each reactant 19j. p. Senter, Rev.Sci. lnstrum. 40, 334 (1969). 20R. S. Berry, S. A. Rice, and J. Ross, "Physical Chemistry." Oxford University Press, New York, 2000.

154


[7]

are mixed to give about 300 #1 of the final solution in 0.4-ml polyethylene tubes to prevent evaporation and to allow for rinsings, and the mean value of three readings is taken for A v. Three of these solutions are normally studied. The molar volume change is obtained by dividing A v by the number of moles of the limiting reagent. Oscillating M e t h o d

In the oscillating method the density is measured with an Anton Paar densimeter (DMA 60) with two microcells (DMA 602). Each glass cell is a U-shaped tube with a volume of 0.2 ml, and is electromagnetically forced into oscillation. To reduce the influence of temperature fluctuations, a differential method is used in which two identical cells are connected to a common thermostat. The sample cell is filled with the sample solution while the reference cell is filled with solvent of equal composition. The density is obtained by measuring the period of oscillation, T, of the solution (or gas), using the relationship p = A ( T 2 - B ) , where A and B are constants determined from a calibration curve. This instrument is calibrated by using the densities and periods of water and air, or solutions of known densities as in the case of the magnetic suspension densimeter. The precision in the density is 4- 5 × 10 6 g/ml. The investigator should be cautious in using this instrument for density determinations of viscous macromolecule solutions; it is best to reduce the viscosity of these solutions by sonication. M o l e c u l a r I n t e r p r e t a t i o n of V o l u m e C h a n g e To understand the physical significance of the volume change of an association reaction, it is best to discuss initially the apparent molar volume, qbV, of a solute molecule. This is defined as the volume difference of a solution relative to that of the solvent 11,21: • V = ( V - n l V l ) / n 2 = Vm -]-

AVh

(2)

where V is the volume of a solution that contains nl mole of the solvent (water) and n2 moles of solute; VI is the molar volume of pure solvent and is in turn equal to the sum of two terms, that is, the fight-hand side of Eq. (2), which includes Vm (the intrinsic molar volume of the solute that is inaccessible to the surrounding solvent) and AVh (the hydration contribution), which is the volume change of water around the solute molecule as a result of solute-water interactions, and the void volume between the solute molecule and the surrounding water (the so-called thermal volume because it is due to thermally induced molecular vibrations). 22'23 21 D. W. Kupke, in "Physical Principles and Techniques of Protein Chemistry" (S. J. Leach, ed.), p. 1. Academic Press, New York, 1973. 22 S. Terasawa, H. Itsuki, and S. Arakawa, J. Phys. Chem. 79, 2345 (1975). 23 D. E Kharakoz, J. Sol. Chem. 21, 569 (1992).

[7]


155

In general, the majority of solutes have positive Vmvalues and negative AVh values. For most solutes, the sum of the magnitudes of the intrinsic and thermal volumes is greater than the magnitude of AVh, which leads to positive values of the apparent volume.11,24,25 The few exceptions are some metal ions with negative ~ V values because of their small size and strong hydration contribution.1 l The volume changes accompanying the binding of a ligand molecule to a DNA molecule can be expressed in the following way: AV = AVm + AAVh, where AVm is the change in the intrinsic volume of both DNA and ligand (normally negligible) and A A Vh is the hydration contributions of the interacting molecules, DNA and ligand molecules. The AVm of association reactions involving nucleic acids without conformational changes is normally considered equal to zero, that is, the intrinsic volume of a nucleic acid molecule remains the same. Therefore, A V is simply equal to the volume changes due to the rearrangements of water molecules in the hydration shells of the interacting molecules. Because hydrating water has a lower molar volume than bulk water at room temperature, a positive value of AV indicates an expansion of the system or release of water molecules from hydration shells while a negative value of A V is indicative of a system contraction or uptake of water molecules. Another parameter that can be estimated quantitatively from A V measurements is the actual number of water molecules involved in a reaction, but only if the type of the participating water is known. In most cases, this analysis is complicated because during DNA-ligand interaction different types of atomic groups are involved, including charged, polar, and nonpolar groups. These groups have characteristic hydration states involving two types of water molecules that have different molar volumes. However, a lower estimate of the number of water molecules, N, participating in a particular association reaction can be obtained from the following relationship: N = A A V h / ( V E - Vw); where VE and Vw are the molar volumes of electrostricted and bulk water, respectively. In using this equation, the apparent molar volume of electrostricted water is considered equal to 15.5 ml/mo116,18 and the assumption is made that the apparent molar volume of structural water is higher than this value. Volume Effects on Interaction of Ligands with DNA a s F u n c t i o n of B i n d i n g M o d e A small ligand interacts with a nucleic acid double helix in different physical binding modes, and by physicochemical attachment. Some of these binding 24 H. Hoiland, in "Thermodynamic Data for Biochemistry and Biotechnology" (H.-J. Hinz, ed.), p. 17. Springer-Verlag, Berlin, 1986. 25 H. Durchschlag, in "Thermodynamic Data for Biochemistry and Biotechnology" (H.-J. Hinz, ed.), p. 45. Springer-Verlag, Berlin, 1986.

156


[7]

modes includes the following: outside the helix, where the complex is stabilized by electrostatic interactions between the positive charges of the ligand and the negatively charged phosphate groups of the nucleic acid; by intercalation of planar aromatic rings between adjacent base pairs, which are stabilized mainly by van der Waals and electrostatic interactions; and by geometric complementarity with the grooves, the complex is stabilized by a combination of electrostatic, van der Waals, and hydrogen-bonding interactions. Typical examples include the electrostatic binding of Mg 2÷ ions, the intercalators ethidium and propidium, and the minor groove ligand netropsin. In this section, we discuss the volume effects that take place in the binding of these ligands to calf thymus DNA (CTDNA). In addition, we show the volume effects for the interaction of actinomycin D with CT-DNA, which is an example of both intercalative and minor groove binder. Electrostatic Interaction o f m g 2+ Ions

Mg 2+ plays essential roles in biology by stabilizing the secondary structure of DNA, the tertiary folding of RNA molecules, and protein-nucleic acid complexes. 1,26 In general, metal ions are well hydrated, particularly small cations such as Mg2+. 11'27 DNA molecules are also strongly hydrated I and it has been shown that among the DNA conformations the B form has the strongest hydration. 1,28 Therefore, the interaction of Mg 2+ with DNA should result in strong hydration effects if their hydration layers overlap. Consistent with this expectation, we obtained a AV of 22 ml/mol of bound Mg 2+ to CT-DNA (see Table I). The positive value indicates an overall release of water molecules, which is consistent with earlier measurements of 27 ml/mol at lower ionic strengths. 29'3° This comparison at different ionic strengths also confirms the electrostatic nature of Mg 2+ binding. Because in the binding of 1 mol of Mg 2+ there is a release of 2 mol of Na + ions, the resulting AV value also includes the hydration contribution for this exchange of Na + with Mg 2+ ions in the ionic atmosphere of DNA. 31 However, DNA-condensed sodium ions keep their hydration layers intact and the hydration contribution from the release of Na + may be considered negligible. X-Ray structural data of MgZ+-oligonucleotide crystals have indicated that Mg 2+ ion binds in the major groove of DNA, keeping intact its first hydration layer. 32 Previous results from our laboratory have shown that Mg 2+ binding is somewhat specific to 26V. K. Misra and D. E. Draper, Biopolymers 48, 113 (1999). 27E J. Millero, G. K. Ward, E K. Lepple, and E. V. Hoff,J. Phys. Chem. 78, 1636 (1974). 28D. Rentzeperis, D. W. Kupke, and L. A. Marky,Biopolymers 33, 117 (1993). 29R. M. Clement, J. Sturm, and M. P. Daune, Biopo(ymers 12, 405 (1973). 3oB. I. Kankia, Biophys. Chem. 84, 227 (2000). 31 V. A. Buckin, H. Tran, V. Morozov,and L. A. Marky,J. Am. Chem. Soc. 118, 7033 (1996). 32L. McFail-Isom,X. Shui, and L. D. Williams, Biochemistry 37, 17105 (1998).

[7]

VOLUMEEFFECTSOF DRUG-NUCLEICACIDINTERACTIONS

157

DNA by forming inner-sphere complexes with dG. dC base pairs, 33 justifying the bending of GC sequences induced by Mg 2+. We arrive at the conclusion that the measured dehydration effects correspond to a partial removal of water molecules from the hydration shells of both Mg 2+ and DNA atomic groups. Furthermore, we estimate a removal of approximately nine water molecules in this interaction by dividing AV ( = 22 ml/mol) by the overall molar compression of electrostricted water of 2.5 ml/mol obtained earlier. 18,27 Intemalative Binding o f Ethidium and Propidium

Ethidium is a classic intercalator that has been investigated extensively and is commonly used in biotechnology as a fluorescent dye to characterize nucleic acid molecules in gel electrophoresis experiments. This charged molecule has a planar phenanthroline ring that intercalates into DNA base pairs with a neighbor exclusion parameter of 0.5 base pair, while propidium, a similar molecule, possesses an extra charged side chain (see structures in Fig. 1). Table I lists the resulting AVvalues for the interaction of these ligands with DNA at two different [ligand]/[DNA] ratios in the intercalative mode. These ratios correspond to two different slopes of the heat (and ultrasonic velocity) versus ligand-DNA titrations curves prior to saturation of DNA by intercalation (data not shown). All A V values are negative, corresponding to an uptake of water molecules; the values for ethidium are consistent with earlier dilatometric measurements. 34 However, two features should be noted about our volume effects: at the smaller degree of binding, the uptake of water molecules for each drug is larger; and propidium shows a larger uptake of water molecules. It is difficult to interpret unambiguously the resulting volume effects of intercalation. Several hydration contributions are involved in the formation of the intercalative complex. Some of these contributions include (1) the release of water due to electrostatic interactions of the positive charges of the ligands and the negative phosphates of DNA, (2) the hydrophobic release of water from the ligands prior to intercalation, and (3) the hydrophobic uptake of water after the intercalation of the phenanthroline ring due to the unwinding and lengthening of DNA, which also gives a higher exposure of organic groups to the solvent. The net effect is that the uptake of water of the last contribution overrides the release of water of the first two contributions. This is consistent with the higher uptake of water molecules at the smaller degree of binding, in which larger changes of the DNA structure takes place, and with the presence of heat capacity effects that have been measured in the binding of these ligands to CT-DNA. 35'36 33W.A. Buckin,B. I. Kankiya,D. Rentzeperis,and L. A. Marky,J. Am. Chem. Soc. 116,9423 (1994). 34E Delben,E Quadrifoglio,V. Giancotti, and V. Crescenzi,Biopolymers 21, 331 (1982). 35j. Ren, T. C. Jenkins, and J. B. Chaires, Biochemistry 39, 8439 (2000). 36A. M. Soto, C. Dziowgo,and L. A. Marky, in "12th Annual Gibbs Conferenceon Biothermodynamics," Book of Abstracts, p. 98. Carbondale, IL, 1998.

158


H2N~~~

?=.:,

(~

NH2

H

C2H5

MeVal I Sar

MeVal "-7 I Sar

o

Pro

Pro

o

D-Val

D-Val

I

I Thr , NH I CO

I

I'--Thr ,

NH I CO [r~N'~~

'

NH2 N~'N('CH2)3N+(C2H5)H3 2C

Propidium

I

I

~

~

Ethidium

I

2

[71

-,~'NH2 H2N NH HN

NH2

CH3

,Y'N CH3

CH3

Actinomycin D

H2N

O

"~'O

CH3

Netropsin

FIG. 1. Chemical structures of ethidium, propidium, actinomycin D, and netropsin.

The overall uptake of water molecules is higher for propidium and this may be due to its extra quaternary amino charged group. If it is assumed that both ligands form similar structures with DNA, the differential AV is --9 ml/mol, which corresponds to an uptake of about four water molecules. Minor Groove Binding A classic groove binder is the dicationic oligopeptide netropsin (see Fig. 1). Netropsin binds with high affinity and high specificity to the minor groove of 4-5 dA- dT base pairs of DNA (for references see Marky and Kupke4); it actually penetrates down to the floor of this groove. This complex is stabilized by electrostatic, van der Waals interactions of the ligand with the sugar-phosphate backbone, and by hydrogen bonding between the amide protons of netropsin with N-3 of adenine and 0-2 of thymine facing the floor of the minor groove. Binding affinities drop

[7]


159

TABLE I VOLUME EFFECTS ACCOMPANYINGINTERACTIONOF LIGANDS WITHCALF THYMUSDNAa Ligand Mg 2+ Ethidium Propidium Netropsin Actinomycin D

AV (ml/mol) 22 - 3l - 12I' -40 -21 h 1 16

Binding mode Electrostatic Intercalation (0.20) Intercalation (0.50) Intercalation (0.18) Intercalation (0.38) Minor groove Intercalation and grooves

"All values determined at 20 ° in 10 mM sodium HEPES, pH 7.5. Values in parentheses are the final [ligand]/[DNA] ratios used in measurements. t, These AV values have been measured at the indicated [ligand]/[DNA] ratios but corrected for the initial binding of ligands.

quite significantly with the increase in GC content of the DNA 37 because the NH2 of guanine blocks by steric hindrance the deep penetration of the ligand in the minor groove. In the case of CT-DNA, which has "-~40% GC content, netropsin binds to this polymer with a binding affinity of ~105. The observed volume effect is close to zero (see Table I) and consistent with netropsin lying in or partially penetrating the minor groove. The release of water molecules, from the formation of ion pairs, cancels the immobilization of water by the partial exposure of the netropsin and/or van der Waals interactions (see below for the volume effects of the binding of netropsin to oligomers containing AT sequences). Intercalation and Minor Groove Binding Actinomycin D is an antiobiotic commonly used clinically in the treatment of rhabdomyosarcoma and Wilms' tumor in children (for references see JaresErijman e t a/.38). It consists of a planar phenoxazone chromophore and two pentapeptide cyclic rings of similar sequence (see Fig. 1). The pharmacological action of actinomycin D is generally attributed to its tight and specific binding to DNA, which results in the inhibition of transcription elongation by blocking the RNA polymerase. Binding to DNA takes place by intercalation of the phenoxazone chromophore and by the formation of hydrogen bonds between specific groups 37 D. Rentzeperis, L. A. Marky, and D. W. Kupke, Biopolymers 32, 1065 (1992). 38 E. A. Jares-Erijman, R. Klement, R. Machinek, R. M. Wadkins, L. A. Marky, B. I. Kankia, and T. M. Jovin, Nucleosides Nucleotides 16, 661 (1997).

160


[7]

of the pentapeptide cyclic rings and the nucleic bases facing the minor groove of the DNA duplex. 39 This ligand is not charged; therefore, the hydration effects would be attributed to the uptake of water from the dissociation of actimomycin D aggregates prior to binding and from the intercalation event, and to the release of water from hydrogen bonding and from the physical placement of the hydrophobic peptide chains of the ligand on the minor groove of DNA. Inspection of Table I shows a positive AV, or release of water. To a first approximation, this can be explained as a cancellation of the opposing contributions from the dissociation of aggregates and the intercalation event with the physical placement of the peptide groups. What remains is the hydrogen-bonding term, which allows for the formation of tighter complexes and the release of more water molecules. The volume effect for the formation of one hydrogen bond between polar atomic groups has been estimated previously to be ~ 4 ml/mol. 23 Dividing our AV of 16 ml/mol by this number presents the formation of four hydrogen bonds between actinomycin D molecule and CT-DNA. I n t e r c a l a t i o n of E t h i d i u m a n d P r o p i d i u m w i t h RNA P o l y n u c l e o t i d e s To include the role of nucleic acid conformation, we reported the volume effects for the interaction of ethidium and propidium with poly(rA), poly(rU) and poly(rA), and the results are compared with those presented for CT-DNA. A DNA helix adopts the B conformation and poly(rA)-poly(rU) adopts the A conformation, whereas poly(rA) forms a single-stranded helix (due to the stacking of adenines) in the A-like conformation. Table II lists the resulting AV values for the intercalation of these ligands with poly(rA) • poly(rU) and at two different [ligand]/[DNA] ratios. Similar to the results with DNA, the AV values are negative and indicate an uptake of water molecules. Also, at the smaller degree of binding, the uptake of water molecules for each drug is higher and propidium shows the most uptake of water molecules. All the hydration contributions mentioned above for the intercalative binding to DNA hold for RNA. Relative to CT-DNA and at the lower degree of binding, the water uptake of ethidium and propidium binding with RNA is somewhat smaller, whereas it is marginally larger at the higher degree of binding. The lower uptake of water may be explained in terms of the larger helical radius for RNA, which induces a better shielding of the phenanthroline ring from the solvent. This explanation does not take into account the smaller hydration of the free RNA polynucleotide and any contributions from differences in the nucleotide sequence. The AV values for ethidium and propidium binding to poly(rA) are both positive (see Table II). This means that the hemiintercalation of the ligands, between 39 T. R. Krugh, Proc. Natl. Acad. Sci. U.S.A. 69, 1911 (1972).

[7]


161

TABLE I1 VOLUME EFFECTS ACCOMPANYINGINTERACTIONOF LIGANDSWITHPOLYRIBONUCLEOTIDESa Ligand

AV (ml/mol)

Binding mode

Poly(rA)opoly(rU) Ethidium Propidium

-24 - 15~ -31 - 2 6 t'

Intercalation Intercalation Intercalation Intercalation

(0.25) (0.52) (0.20) (0.45)

Poly(rA) Ethidium Propidium

13 17

Hemintercalation Hemintercalation

a All values determined at 20 ° in 10 mM sodium HEPES, pH 7.5. Values in parentheses are the final [ligand]/[DNA] ratios used in the measurements. b These AV values have been measured at the indicated [ligand]/[DNA] ratios but corrected for the initial binding of ligands.

two adjacent adenines, is accompanied by a release of water molecules. The singlestranded poly(rA) helix does not have grooves (which are normally well hydrated in the duplex state) and the phenanthroline ring is partially buried; therefore, the volume effect reflects primarily the water release of electrostatic interactions between negatively charged phosphates of the nucleic acid and the positively charged atomic groups of the ligands. B i n d i n g of N e t r o p s i n to M o d e l B e n t S e q u e n c e s We studied the interaction of netropsin with a set of isomeric oligonucleotide duplexes, [d(GAaTaC)]2 and [d(GTaA4C)]2; each oligomer binds two netropsin molecules and end-to-end ligation of the first oligomer forms curved polymers, which run in gel electrophoresis experiments with a slower mobility. The comparison of the resulting volume effects yielded the type of water involved in bent sequences. We obtained volume changes by mixing netropsin and oligomer duplex in two different ways: to form 1 : 1 ligand-oligomer complexes followed by the addition of a second equivalent of ligand, and to form directly the 2 : 1 complexes; both methods yielded similar results shown in Table III. Binding of netropsin to the first site of each oligomer generated a volume contraction whereas a volume expansion is observed in the binding to the second site. The magnitude of these volume effects was higher with the GAaT4C oligomer than with its GT4AaC isomer. Thus, at this salt concentration of 0.1 M, we infer that a net hydration accompanies the

162


[71

TABLE III VOLUME EFFECTS ACCOMPANYINGINTERACTIONOF NETROPSIN WITH DNA OLIGOMERSa Duplex

AV (ml/mol)

[d(GA4T4C)12

- 100(first site) 37(second site) -63(saturated complex) -83(first site) 15(second site) -68(saturated complex)

[d(GT4A4C)]2

All values determined at 20" in 10 mM sodium phosphate buffer, 100 mM NaC1, 0.1 mM Na2EDTA, pH 7.

total binding to these oligomer duplexes as reflected by the apparent compression of water molecules. The [d(GA4T4C)]2 is slightly more hydrated by 5 ml/mol of duplex and indicates that this net hydration depends on the DNA sequence and the type of site of the oligomer duplex. To obtain a better understanding of the observed differences, we have dissected the overall volume profiles into the individual contributions for netropsin binding to the first and second site, respectively, and have taken the differences for each site, using the formation of the netropsin[d(GT4AaC)]2 as a reference state. The resulting AAV parameters are 17 ml/mol for the first site and 22 ml/mol for the second site. A similar procedure for the dissection of the AAG ° terms obtained from melting experiments (data not shown) yielded a marginal AAG ° of 0.4 kcal/mol for each binding site of netropsin. ~8 These results indicate that in the sequential binding of netropsin, there is an initial increase in hydration due to a hydrophobic or structural effect (opposite signs of AAV and AAG°), followed by a dehydration event due to electrostriction effects (similar signs of AAV and AAG°). ~8 Analysis of the electrophoretic mobility of the free and netropsin-bound oligomer duplexes shows that the mobility of the free [d(GAaT4C)]2 is lower than that of the free [d(GT4A4C)]2; the 1 : 1 netropsinoligomer complexes ran with similar mobilities whereas for the 2 : 1 complexes the trend is reversed. These results strongly indicate that binding of the first netropsin straightens the curved [d(GA4T4C)]2 duplex, yielding an important correlation with our volumetric studies in that the straightening of this oligomer invokes the participation of structural water. F o r m a t i o n of O l i g o m e r D u p l e x e s w i t h C o v a l e n t l y A t t a c h e d L i g a n d s One way to study the hydration properties of covalently attached moieties to DNA is by comparing the AV values for the formation of oligomer duplexes,

[7]

VOLUME EFFECTS OF DRUG-NUCLEICACID INTERACTIONS

163

by mixing its component single strands with and without the covalently attached moieties. In this section, we present results for two sets of oligonucleotides: the inclusion of two enantiomers of benzo[a]pyrene diol epoxide, (BPDE), (+)-BPDE and (-)-BPDE, with d(CCATCGCTACC)/d(GGTAGCGATGG) at the central Gforming DNA adducts; and the covalent attachment of cisplatin, cis-[Pt(NH3)2Cl2], to the central G G step of d(CTCTGGTCTC)/d(GAGACCAGAG). The volume effects accompanying the formation of a nucleic acid, from mixing its complementary strands, are interpreted to reflect changes in the electrostriction and/or hydrophobicity of water dipoles, which are immobilized by each of the participating species. For instance, the single strands have the bases more exposed to the solvent and immobilize more structural water. The base pairing and base pair stacks of the duplex bring the bases away from the solvent, yielding a duplex with a higher charge density that interacts with counterions more effectively and immobilizes more electrostricted water. The net result will depend on these two effects, which tend to compensate for each other. Furthermore, in the comparison of volume changes for each of these two sets we are assuming that the modified strands make a similar contribution to AV; this may be true because the complementary strands for a particular set have similar chemical compositions. Benzo[a]pyrene Diol Epoxide Adducts

BPDE molecules are metabolized in vivo and bind covalently to DNA. Their biological activities are significantly dependent on their stereochemical configurations. For instance, (+)-BPDE is strongly tumorigenic, whereas ( - ) - B P D E is not. 4°'41 All AV values were measured with a magnetic suspension densimeter. The formation of each duplex is accompanied by an uptake of water molecules as seen in the negative AV values of Table IV. The main observation is that the uptake of water for the unmodified duplex and the ( - ) - B P D E duplex is the same, whereas the water uptake is about 50% larger for the (+)-BPDE duplex. 42 Relative to the unmodified duplex, we obtained A AV values of 8 ml/mol of duplex for the ( - ) - B P D E duplex and - 6 5 ml/mol of duplex for the (+)-BPDE duplex. The modified duplexes are less stable, A AG ° of "-~5kcal/mol, but the overall counterion uptake is similar for all three duplexes, 40.8 mol of Na + per mole of duplex. The opposite signs of these thermodynamic parameters for the (+)-BPDE duplex reveal a high ordering of structural water whereas the similarity of the signs for the ( - ) - B P D E duplex indicate a marginal uptake of electrostricted water. The nuclear magnetic resonance (NMR) structures have indicated that these adducts lie in the minor groove of this duplex with different orientations: the (+)-anti-BPDE toward 4oA. H. Conney,Cancer Res. 42, 4875 (1982). 41p. Brookesand M. R. Osborne,Carcinogenesis 3, 1223 (1982). 42L. A. Marky,D. Rentzeperis,N. E Luneva,M. Cosman,N. E. Geacintov,and D. W. Kupke,J. Am. Chem. Soc. 118, 3810 (1996).

164


[7]

TABLE IV VOLUME EFFECTS ACCOMPANYINGFORMATION OF DUPLEXES WITH CHEMICAL ALTERATIONS a Duplex

AV(ml/mol)

Inclusion of BPDE moieties h Unmodified With (-)-BPDE With (+)-BPDE

- 144 - 136 -209

Inclusion of cisplatin' Unmodified With cisplatin

19 - 35

a All measurements determined at 20 °. ~' Inclusion of BPDE moieties in 20 mM sodium phosphate buffer, 100 mM NaC1, pH 7. c Inclusion of cisplatin in 10 mM Na-HEPES, pH 7.5.

the 5' end of the unmodified strand and the ( - ) - B P D E toward the 3' end of the unmodified strand. 43 The sequence of this duplex is symmetric with respect to the modified base pair at the center, and therefore the (+)-BPDE is exposing more organic groups to the solvent. In addition, end-to-end ligation of (+)-BPDE duplexes yields bent structures 44 similar to the [d(GA4T4C)]2 duplex, and accordingly we conclude that bent duplex structures have a higher ordering of hydrophobic water.

Cisplatin The antitumor activity of cisplatin, a widely used drug, is attributed to its covalent interaction with DNA. The major cisplatin adducts are with GG or AG intrastrand cross-links. We used the Anton Paar densimeter to measure the volume changes accompanying the formation of duplexes. Unmodified duplex formation is accompanied by a marginal release of water molecules, whereas the cisplatin duplex reveals a significant uptake of water molecules (see Table IV); X-ray 45 and NMR 46 studies of d(CCTCTGGTCTCC)]d(GGAGACCAGAGG) with a cisplatin attached to the G G position demonstrated that the adduct causes the guanine bases to roll toward one another by 49 °, leading to an overall helix bend angle of 78 ° . We 43 C. De los Santos, M. Cosman, B. E. Hingerty, V. Ibanez, L. A. Margulis, N. E. Geacintov, S. Broyde, and D. J. Patel, Biochemistry 31, 5245 (1992). 44 B. Mar, Ph.D. Dissertation, New York University, 1994. 45 p. M. Takahara, A. C. Rosenzweig, C. A. Frederick, and S. J. Lippard, Nature (London) 377, 649 (1995). 46 A. Gelasco and S. J. Lippard, Biothemistry 37, 9230 (1998).

[8]

D N A SEQUENCE-DEPENDENT STABILITY AND GRAPHS

165

investigated here a shorter version of this duplex, without one dC- dG base pair at either end, and it is expected to have a similar bend at the modified base pair. We have done an analysis similar to that of the previous set of duplexes and obtained, relative to the unmodified duplex, a AAV value of --54 ml/mol of duplex for the cisplatin duplex and a AAG ° of "-~3kcal/mol. 46 The opposite signs of AAV and AAG ° reveal that the bent cisplatin duplex is immobilizing structural water. Acknowledgment This work was supported by grant GM42223 (L.A.M.) from the National Institutes of Health.

47 B. 1. Kankia and L. A. Marky, unpublished data (2000).

[8] Calculating Sequence-Dependent Melting Stability of Duplex DNA Oligomers and Multiplex Sequence Analysis by Graphs By ALBERT

S. BENIGHT, PETR PANt~OSKA, RICHARD OWCZARZY,

PETER M . VALLONE, JAROSLAV NESETI},IL, a n d PETER V. RICCELLI

Introduction DNA duplex formation by annealing or hybridization of two complementary single strands is a central reaction in many important cellular events and such fundamental biological processes as replication, transcription, and recombination. In vitro binding and melting studies have shown, for short duplex DNA oligomers of 12 to 22 base pairs, that thermodynamic stability of flanking sequences is inversely related to the binding strength of several DNA restriction enzymes and site-specific binding ligands.l-3 Because DNA hybridization is sequence dependent, this basic reaction also underlies many research and commercial diagnostic applications. 4-6

1 A. S. Benight, E J. Gallo, T. M. Paner, K. D. Bishop, B. D. Faldesz, and M. J. Lane, Adv. Biophys. Chem. 55, 1 (1995). 2 p. V. Riccelli, P. M. Vallone, I. Kashin, B. D. Faldasz, M. J. Lane, and A. S. Benight, Biochemistry 38, 11197 (1999). 3 p. M. Vallone and A. S. Benight, Biochemiso'y 39, 7835 (2000). 4 E. A. Wagar, J. Clin. Lab. Anal. 10, 312 (1996). 5 j. G. Wetmur, Crit. Rev. Biochem. Mol. Biol. 26, 227 (1991). 6 K. B. Mullis, Ann. Biol. Clin. 48, 579 (1990).

METHODSIN ENZYMOLOGY,VOL.340

Copyright© 200l by AcademicPress All rightsof reproductionin any formreserved.

0076-6879/00$35.00

[8]


165

investigated here a shorter version of this duplex, without one dC- dG base pair at either end, and it is expected to have a similar bend at the modified base pair. We have done an analysis similar to that of the previous set of duplexes and obtained, relative to the unmodified duplex, a AAV value of --54 ml/mol of duplex for the cisplatin duplex and a AAG ° of "-~3kcal/mol. 46 The opposite signs of AAV and AAG ° reveal that the bent cisplatin duplex is immobilizing structural water. Acknowledgment This work was supported by grant GM42223 (L.A.M.) from the National Institutes of Health.

47 B. 1. Kankia and L. A. Marky, unpublished data (2000).

[8] Calculating Sequence-Dependent Melting Stability of Duplex DNA Oligomers and Multiplex Sequence Analysis by Graphs By ALBERT

S. BENIGHT, PETR PANt~OSKA, RICHARD OWCZARZY,

PETER M . VALLONE, JAROSLAV NESETI},IL, a n d PETER V. RICCELLI

Introduction DNA duplex formation by annealing or hybridization of two complementary single strands is a central reaction in many important cellular events and such fundamental biological processes as replication, transcription, and recombination. In vitro binding and melting studies have shown, for short duplex DNA oligomers of 12 to 22 base pairs, that thermodynamic stability of flanking sequences is inversely related to the binding strength of several DNA restriction enzymes and site-specific binding ligands.l-3 Because DNA hybridization is sequence dependent, this basic reaction also underlies many research and commercial diagnostic applications. 4-6

1 A. S. Benight, E J. Gallo, T. M. Paner, K. D. Bishop, B. D. Faldesz, and M. J. Lane, Adv. Biophys. Chem. 55, 1 (1995). 2 p. V. Riccelli, P. M. Vallone, I. Kashin, B. D. Faldasz, M. J. Lane, and A. S. Benight, Biochemistry 38, 11197 (1999). 3 p. M. Vallone and A. S. Benight, Biochemiso'y 39, 7835 (2000). 4 E. A. Wagar, J. Clin. Lab. Anal. 10, 312 (1996). 5 j. G. Wetmur, Crit. Rev. Biochem. Mol. Biol. 26, 227 (1991). 6 K. B. Mullis, Ann. Biol. Clin. 48, 579 (1990).


Copyright© 200l by AcademicPress All rightsof reproductionin any formreserved.

0076-6879/00$35.00

166


[8]

Many assays utilize sequence-specific hybridization as the basis for recognition of nucleic acids. The functional utility of these assays relies heavily on the fidelity of DNA sequence-dependent hybridization and stability of the hybridized products. A number of practical applications employ hybridization between relatively short duplex regions. Stability of the hybridized products is determined by the sequence dependence of the transition temperatures (Tin values) of the duplex complexes. To optimize probe and primer sequence design 5'7 it is of particular necessity to be able to accurately predict the Tm values of DNA oligomer duplexes under a given set of conditions (ionic strength and strand concentration) from their base pair sequencesf In addition, an understanding of solution hybridization and DNA sequence-dependent thermodynamic stability has direct practical applications in the design of multiplex reactions in which many desired hybridizations are required to occur with high fidelity over a narrow temperature range, with all the required strands present in the same reaction mixture. These sequences are said to be "isothermal," in that they have the same Tm (or at least differ over a narrow range of T,,~ values). DNA microarrays are being designed or contemplated that will have thousands to millions of different DNA sequences on them. 9 For these microarrays to be useful all sequences will be required to hybridize precisely to their complement strand, and to no other. For these reasons there is widespread interest in analytical calculations of sequence-dependent stability of duplex DNA. A significant challenge in multiplex reaction design is the selection of isothermal sequences. An obvious approach to solving this problem might involve brute force calculation of the stabilities of all sequences, and then grouping them according to their calculated thermodynamic stability. We present a graph theory approach that can be employed to analyze and classify sequence families. This representation of sequences provides an efficient means for grouping them according to their graph topologies. Using these graph methods it is not necessary to calculate the stability of every individual sequence, group sequences in isothermal sets according to their calculated stabilities, and then count the sequences in each group. Rather, it is possible to classify sequences according to their graphical representation in such a way that all sequences within a class, represented by the same graph, can readily be counted, and by definition have the same sequence composition (base pairs and nearest neighbors) and thus must have identical calculated thermostabilities. The presentation is in two parts. In the first part, the nearest-neighbor (n-n) model for characterizing DNA sequence-dependent stability is reviewed. A set of n-n parameters is presented and the use of these parameters to calculate sequencedependent stability of short linear DNA oligomers is demonstrated. In the second 7 W. Rychlik, W. J. Spencer, and R. E. Rhoads, Nucleic Acids Res. 18, 6409 (1990). s R. Owczarzy, P. M. Vallone, E J. Gallo, T. M. Paner, M. J. Lane, and A. S. Benight, Biopolymers (Nucleic AcidSci.) 44, 217 (1997). R. J. Sapolsky, L. Hsie, A. Berno, G. Ghandour, M. Mittmann, and J. B. Fan, Genet. Anal. 14, 187 (1999).

[8]


167

part, application of graph theory to sequence analysis is described. Essential elements of the theory are presented and necessary mathematical relationships are recalled. Finally, several illustrative examples are presented. Review of A n a l y t i c a l M e t h o d s a n d R e s u l t s In this section the basic model of melting short duplex DNA is reviewed. Key parameters are identified. The models that consider DNA sequence dependence are also reviewed, and a set of n-n parameters evaluated from melting studies of DNA dumbbells is presented. Use of the n-n stability values to calculate the thermodynamic parameters and Tm values of two short DNA duplex oligomers is demonstrated.

Two-State Melting Theory Melting of short duplex DNAs ( 3'

10 g Kab

Sequence ~

(M-I)

AATT TAAT ATAT TATA TTAA

10 -7 (1/C50) "

10 8kond

koff d

(M-1 sec-I)

(sec t)

5.2 ± 0.2 2.78 ± 0.85 2.47 ± 0.03 (1) (1) 0.271 ± 0.003 0.051 ± 0.(121 1.57 ± 0.03 (0.0518) (0.0183) 0.240 ± 0.002 0.022 ± 0,014 1.26 ± 0.002 (0.0459) (0.0079) 0.026 ± 0.001 0.008 ± (I.004 0.68 ± 0.01 (O.0O5O) (0.0028) 0.030 ± 0.001 0.007 ± 0.005 1.93 ± 0.07 (0.0058) (0.0024)

10 g Kkine

(M-~)

0.419 ± 0.005

5.9 ± 0.1

7.0 ± 0.4

0.22 ± 0.02

10.5 ± 0.3

0.12 ± 0.04

66 ± 3

0.010 ± 0.001

96 ± 4

0.020 ± 0.003

a The 5' --->3' sequences are constructed in DNA hairpin structures as in Fig, lB. h Ka values were determined by fluorescence equilibrium titrations of 1.6 nM p-OH Hoechst with excess DNA (our unpublished data, 2000). The errors are for the fit to the single titration curve. The relative values of Ka (relative to the AATT sequence) are given in parentheses, c Data from Abu-Daya etal. [footprinting experiments; A. Abu-Daya, P. M. Brown, and K. R. Fox, Nucleic Acids Res. 23, 3385 ( 1995)1. d The kon and koff values are from the plots in Fig. 7, except for the koff for AATT (taken from Fig. 5). The errors are for the linear fit to the data. e Kki,1 = kon/koff for a simple bimolecular association.

constants s u m m a r i z e d in Table III shows that the kinetic traces b e c o m e slower as the binding to the different sequences b e c o m e s stronger. This difference in the reaction rate is almost exclusively due to the dissociation rate parameter koff. Again, the stopped-flow reactions were fitted as m o n o e x p o n e n t i a l s and the plots o f kobs as a function of concentration are linear for the five sequences (Fig. 7). B e c a u s e the koff for p - O H H o e c h s t and the TAAT, ATAT, TATA, and T T A A sequences are larger than for AATT, they could be determined f r o m the intercept in the concentration plots. The koff ['or the A A T T site was obtained by displacement kinetics with p o l y [ d ( A - 5 B r U ) ] as described above. The ko, values are similar (0.68 x 108 to 2.47 x 108 M 1 s e c - I ) , and are also similar to the kon values for the derivatives in Table II. In contrast, the koft" values for the sequences differ by a factor of almost 200 and define the affinity of p - O H Hoechst for the different sequences. Thus, as o p p o s e d to binding of p - O H H o e c h s t to p o l y m e r i c D N A s (Table I), binding of a H o e c h s t dye to the A A T T site (Table II), as w e l l as to the four additional (A/T)4 sites (Table III), is characterized by a ko, value that is independent of the dye derivative or the (A/T)4 sequence. The affinity is determined exclusively by the koff or the lifetime of the c o m p l e x that is formed. This important conclusion m i g h t apply generally to other peripheral binding processes, including regulatory

[ 1O]

HOECHSTDYEBINDINGKINETICS

225

40 ,-, 30

A

20 ,_~o

A

~'~ 10 / D ~ ~1~"

~ J

300

• AATT TAAT ' ATAT I

1 10 -7 2 10 -7 [(A/T)4 site] (M) ' /~

B

zoo

100 I

0

5 10.7 [(A/T)4 site] (M)

I

1 10.6

FIG. 7. Determination of konand koffforp-OH Hoechst binding to five (A/T)4 sites in a DNA hairpin. The kobsvalues correspond to the time constants for monoexponential fitting of the stopped-flow traces using the DNA hairpins in excess over 2 or 10 nMp-OH Hoechst (as in Fig. 6). (A) Strongest binding sequences AATT, TAAT, and ATAT;(B) weaker binding sequences TATA and TTAA.

proteins and enzymes. The specificitycomes from the dissociation rate parameter and the shape of the ligand. Comparison o f Kinetics with Equilibrium Data As for the derivatives, the kinetic association constant Kki, = kon/koff is identical or close to the corresponding K~ value obtained by equilibrium fluorescence titrations.19 T h e s e stopped-flow and equilibrium data confirm and refine the footprinting data o f A b u - D a y a et al. 23

23 A. Abu-Daya, E M. Brown, and K. R. Fox, NucleicAcids Res. 23, 3385 (1995).

226


[ 1 0]

Effect o f Base-Pair Roll and Minor-Groove Width on p - O H Hoechst Affinity

The highest affinity of p-OH Hoechst for the AATT sequence seems to correlate with the zero base pair roll and reduced minor groove width (as small as 3.1 ]k) at the central AATT sequence in the d(CGCGAATTCGCG)2 duplex, probably due to the binding of monovalent cations in the bottom of the groove.24 This contrasts with TATA as the poorest affinity site: in solution, the [d(CGCTATACGCG)]2 sequence shows an increased roll (17.2 °) at the two TpA steps that widens the minor groove to an overall width of 7.2 A. 25 Likewise, the single and central TpA step in CGATTAATCG has a local roll of 9.4 ~' and corresponds to a minor groove width of about 8 ]~26 Although narrowing of the minor groove at the AATT sequence correlates with the affinity, it is puzzling that sequencedependent tightening of the minor groove in the AATT sequence does not reduce its value of kon that is the largest of the five sequences investigated (see the next section).

More C o m p l e x Kinetics: T w o - S t e p B i n d i n g M o d e l s We have seen above that the progress curves of p-OH Hoechst binding to polymeric DNAs and to defined oligonucleotides can be represented well by a single exponential. Analysis and interpretation of the concentration-dependent rate data are then straightforward, corresponding to a single-step bimolecular reaction mechanism, with the binding step being almost diffusion controlled and the dissociation rate determining the affinity of the dye for the DNA target. In addition, the reaction rate parameters obtained from this simple kinetic analysis agree well with the equilibrium titration data. However, it is reasonable to wonder how a molecule that fits so snugly into the minor groove can bind with a kinetic association constant that is close to diffusion controlled, especially considering the extended nature of the ligand. In addition, we know that there are specific attachment sites between the dye and the DNA that require a precise positioning of the dye in the groove. For every collision with the DNA the dye cannot be oriented correctly for insertion into the helix groove, and we might suspect that the geometry and size of the minor groove would slow down the association step. Indeed, this is the case, as we show, and it is likely that this conclusion applies to other dye molecules that fit precisely into relatively deep sites in the DNA molecule.

24X. Shui, C. C. Sines, L. McFail-Isom,D. VanDerveer,and L. D. Williams, Biochemistry 37, 16877 (1998). 25J.-W. Cheng,S.-H. Chou,M. Salazar, and B. R. Reid, J. Mol. Biol. 228, 118 (1992). 26j. R. Quintana, K. Grzeskowiak,K. Yanagi,and R. E. Dickerson,J. Mol. Biol. 225, 379 (1992).

[10]

HOECHST DYE BINDING KINETICS I

I

227

i

100 8O

~= 60

g 40 !

2o

A 1

I

I

+

I

o

'

I

I

I

d

,,,

eO

~a

2F

0

+ It~

0.1

] '

i

0.2 T i m e (s)

I

0.3

i

i

0.4

FIG. 8. Association kinetics of p-OH Hoechst and the AATT hairpin with improved signal-tonoise ratio. (A) Average of 20 traces (2000 points per trace) for binding of 8 nM p-OH Hoechst to 52 nM AATT hairpin and its single-exponential fit (corresponding to r = 98.6 msec). (B), (C), and (D) compare the residuals for different fits: (B) monoexponential (R = 0.99869); (C) biexponential (rfast = 45.7 msec and rslow = 113.3 msec, R = 0.99952); (D) triexponential (r~ = 11.3 msec, t2 67.7 msec, and r3 = 2.0 msec, R = 0.99958). =

ExperimentalApproach To detect any deviations from single-exponential behavior we have increased the signal-to-noise levels of the kinetic data acquisition by using higher dye and D N A concentrations. We have done this deliberately with the well-defined D N A oligomers so as not to confuse any deviations with polymeric effects. We have also m a i n t a i n e d the pseudo-first-order conditions with the D N A hairpin in excess over the dye. Figure 8A shows an improved kinetic trace for p - O H Hoechst b i n d i n g to the A A T T hairpin and its single-exponential fit; the residuals of the fit are shown in Fig. 8B. The systematic deviation in the residuals seen in Fig. 8B is found with most

228

BIOPHYSICAL APPROACHES 0

i

[ 101

i

7°' 1 obs,i

60

[]

obs,ii [

•

,-, 50 40 "~° 30 20

plateau value

10

~

~ I

0

~

~

0 4

I

5 10 -8 1 10 -7 1.5 10 7 [AATT D N A hairpin] (M)

2 10 -7

FIG. 9. Concentration plot for binding of p-OH Hoechst to the AATT hairpin, For each DNA concentration about 20 stopped-flow traces (500 points per trace) were averaged and fitted by a double exponential. The linear fit through the inverse fast time constants has a slope of 2.7 × 108 M - I sec -1 and an intercept equal to 13 sec 1.

of the DNA concentrations and with all the DNA hairpin sites in this study. The deviation is small, as we would expect because the previous one-exponential curves already represented the data well. However, this systematic deviation indicates that a single exponential is not an adequate representation of the progress curves.

Biexponential Association Kinetics Fitting the data with two exponentials results in a significant improvement in the residuals for most of the concentrations the residuals now approach a random deviation around zero (Fig. 8C). The fitting parameters obtained show relatively large errors, as we expected considering the previous satisfactory one-exponential fits, and the plots of the inverse time constants ( 1/rfast = kobs,i and 1/ rslow = kobs,ii) versus the DNA concentration, for p-OH Hoechst binding to AATT, show relatively large scatter (Fig. 9). However, the concentration plot in Fig. 9 suggests that the kobs, i values increase approximately linearly with DNA concentration while the kobs,ii values increase slowly and eventually reach a plateau value.

Model 1: Bimolecular Association followed by lsomerization Because we observe two exponential processes there are at least two welldefined kinetic steps for the reaction. Several two-step reactions 27'28 are consistent with the behavior seen in Fig. 9. We can, for example, consider a classic binding

[ 10]

HOECHST DYE BINDING KINETICS

229

model consisting of a bimolecular association reaction followed by a conformational change of the complex: k+l

k+2

DNA+dye ~C~-D k I

k-2

(i)

with Ki = k+l/k-l and K2 ----k+2/k 2. Under first-order conditions, with the DNA sites in large excess over the dye, and assuming that the bimolecular reaction is rapid compared with the rate of the second reaction, kobs.i and ko6s.ii for the above-described binding mechanism are described by 28 kobs,i = k I + k+I(DNA0)

kobs,ii =

k+zKl (DNA0) KI(DNA0) + 1 + k-2

(2) (3)

with (DNA0) the total DNA concentration. A linear fit of the kobs,i in Fig. 9 gives a slope k+l -- 2.7 × 108M -I sec -I, close to the association rate parameter obtained from the single-exponential analysis (2.47 × 108M-Isec-1). The intercept k_l = 13 sec -I is not so well defined and is about 10 times larger than the value from the single-exponential analysis. These kinetic parameters define Ki ----k+l/k_j = 2.1 × 107M -l for the first bimolecular reaction step, significantly smaller than Kkin obtained in the single-exponential analysis or Ka obtained from the equilibrium titrations (see Table III, Ka = 25K1). From the plot of kobs,ii v e r s u s the total DNA concentration (Fig. 9) we can deduce that we are already in the concentration region where kobsd i increases only slowly and has almost reached its plateau value of k+2 + k 2 estimated as 15 sec- J. Relatively high DNA concentrations were used, much higher than the dye concentration (8 nM), in order to be under pseudo-first-order conditions. These concentrations are too high for estimating the reverse rate parameter k_2 from the intercept. However, we can calculate the effective equilibrium constant of this reaction model to be Ka --- KI(1 + K2), assuming that we are observing both bound states as one state in the equilibrium experiments. Because Ka ----25KI (see above), we expect that K2 = k + z / k - 2 = 24. Because we estimated that k+2 q-- k - 2 = 15 s e c -1 w e can solve for k+2 and k-2. The result is k+2 = 14.4sec -I and k 2 = 0.6sec -1. The latter value of k_2 compares well with the overall dissociation rate of p-OH Hoechst from the AATT site measured with poly[d(A-5BrU)] (0.419 sec -1) and is also close to the intercept of the linear concentration plot for the single-exponential fit (1.6 sec-1). The analysis of our kinetic data according to the above-described two-step mechanism [Eq. (1)] contributes considerably to our understanding of the 27 C. F. Bernasconi, "Relaxation Kinetics," Chapter 3, Table 3.1. Academic Press, New York, 1976. 28 A. Van Landschoot, E G. Loontiens, R. M. Clegg, and T. P. Jovin, Era: J. Biochem. 103, 313 (1980).

230


[ 10]

mechanism of the reaction. Specifically, it shows why we can measure an association rate that is close to diffusion controlled, assuming a single-step binding mechanism. The p-OH Hoechst molecule collides with the DNA molecule (in a diffusion-controlled step), and there is at least one additional step following this collision in which the dye molecule snuggles into the minor groove and makes its specific contacts with the DNA. The dye molecule cannot dissociate from the DNA until this second monomolecular step is reversed, requiring the dye to find its way out of the enclosed and restrictive environment of the deep, narrow minor groove. This is a slow process that involves the specific interactions between the dye and molecular groups in the DNA minor groove, and, as we discussed above, its rate determines the affinity of the dye for the particular DNA site. The process may also involve dynamics of the DNA molecule itself, an intriguing possibility that must await further investigations. Note that the kinetic constants k-i = 13 sec -I and k+2 = 14.4sec I are approximately equal, so that when we calculate the ratio of k+l and the dissociation rate measured competitively with poly[d(A-5BrU)]) we obtain a value that agrees well with the equilibrium association constant found by accurate fluorescence titration experiments. This means that the dye molecule in the initial, "outside" bound state has about equal probability to dissociate completely from the DNA molecule into the solution or to enter the enclosed state of the minor groove. Model 2: Two Single-Step Associations

The binding-plus-isomerization model described above is in agreement with high-resolution (1.5-A) X-ray diffraction data that suggest that a Hoechst dye, bound in a particular orientation, can "slide" over the AATT binding site over a distance of 3 ~.29 However, in addition, the X-ray data suggest that the dye can bind in two opposite orientations that partially overlap and mutually exclude each other. This requires the dye to dissociate from the site before reentering, a process that is not included in the above-described two-step binding-plus-isomerization mechanism with only one d y e - D N A complex species. Therefore we consider an alternative model that is kinetically equivalent but mechanistically different, consisting of two single-step associations forming two complexes that mutually exclude each other: k+~

DNA + dye ~- C k ~

k+b

and

DNA + dye ~ D

(4)

k-t,

with KA = k+a/k-a and KB = k+b/k b. Both alternative mechanisms are difficult to distinguish. 3° For the two competing single-step associations the inverse fast and slow time constants (1/rf~t = kobs,i 29C. J. Squire, L. J. Baker,G. R. Clark, R. E Martin, and J. White,Nucleic Acids Res. 28, 1252(2000). 30R. D. Viale, J. Theor. Biol. 31,501 (1971).

[ 1 O]

HOECHSTDYE BINDINGKINETICS

231

and 1/Vslow = kobs.ii) are described by kobs.i = k_a + k+a(DNA0)

kobs.ii

=

k+b(DNA0) KA(DNA0) + 1

Jr- k_ b

(5) (6)

The faster reaction has the same properties as in the first mechanism [Eq. (1)], giving KA = 2.1 x 107M -l . Assuming that both C and D contribute in our equilibrium fluorescence titrations, the overall association constant Ka (5.2 x 108 M - l ) equals KA + KB, which characterizes the slower reaction as forming the tightest complex with KB = 5.0 x 10SM -l. The plateau value in the kobs,ii plot (Fig. 9) is n o w k - b q- (k+b/Ka) = 15 sec 1. Neither the slope k+b n o r the intercept k_ b can be determined from our plot of kobs,ii v e r s u s the DNA concentration (Fig. 9); however, for the complexes C and D, k - a and k - b are expected to be different. We therefore also reexamined the dissociation experiments with poly[d(A5BrU)] under conditions of improved signal-to-noise ratio, and data for the p-OH Hoechst-AATT complex are shown in Fig. 10. The dissociation curves are fitted significantly better with two (Fig. 10C) rather than with one exponential (Fig. 10B). However, the fitted rates are close to each other and differ by a factor of 2.6 only, possibly reflecting a small difference between the values of k-a and k-b. This is an indication of multiple bound states of the dye molecule in the DNA minor groove, each state having its distinctive overall dissociation rate.

Triexponential Association Kinetics As a further note, we mention that in some cases a two-exponential fit of the association kinetic traces still shows systematic deviations that are reduced by fitting with three exponentials (Fig. 8D). The deviation from biexponential behavior is small and we cannot reliably deconvolute another process from the data. Perhaps by varying the experimental conditions (different temperatures, pressures, ionic strength) we will be able to decide whether this is a real effect. At this level of accuracy care is needed not to be confused by small artifacts, even if they are systematic. The instrumentation is stable, but there could be small nonlinearities that are difficult to expose unless one can reproduce the exact conditions of the progress curve. Conclusions In this chapter we have given the reader a working example of a kinetic analysis involving a ligand that binds strongly to specific sequences of DNA. The Hoechst dye-DNA interaction is an informative model system for the study of the highly A/T-specific minor groove binding to DNA. Titrations at equilibrium allow us to determine the high association constants. However, the determination of binding

232


[ 10]

A 80 ca

60 o

40

20

~

I

,a 2LI/

'

'

=21

.

.

.

.

B

~

Time (s) FIG. 10. Dissociation kinetics of the p-OH Hoechst-AATT complex determined by displacement with poly[d(A-5BrU)]. (A) Average of 15 kinetic traces (500 points per trace) for a mixture of 10 nM Hoechst 33258 and 10 nM AATT hairpin, forming 6.5 nM complex, rapidly mixed with 210 nM poly[d(A-5BrU)] using stopped flow. For the sake of clarity no simulation is superposed on the data. (B) Residuals for a monoexponential fit (1/kobs = 2.1 sec). (C) Residuals for a biexponential fit (rl = 1.5 sec, r2 = 3.9 sec).

kinetics considerably augments our understanding of the factors responsible for the high affinity. Unless the data are accurate the stopped-flow curves are well described by single exponentials and the kinetic analysis suggests a simple, single-step binding mechanism. The association rate is almost diffusion controlled and the specificity of the dye comes solely from the reverse rate parameter, possibly determined by the number of dye-DNA interactions in the minor groove that must be broken for the dye molecule to escape. When more interactions become involved, as with the strongest ligand in this study (bis-m-OH), the time for the ligand to dissociate becomes longer, and its affinity increases. Similar principles might well apply to other DNA-binding drugs as well as to the docking of enzymes and regulatory proteins on to DNA.

[ 10]

HOECHST DYE BINDING KINETICS

233

We have discussed details of a two-exponential analysis in order to demonstrate the complexities that can be expected when dye molecules interact strongly with specific DNA sequences. Such expanded kinetic models [Eqs. (1) and (4)] became necessary after increasing the signal-to-noise levels of the kinetic experiments, because the one-exponential analysis was no longer adequate to fit the data. The first of the two-step models [Eq. (1)] has allowed us to understand how a molecule that fits so snugly in the minor groove can still behave as though the association step is approximately diffusion controlled. In addition, we have shown that the affinities of p-OH Hoechst for different sequences are reflected in the dissociation rates--not in the association steps. The two-step mechanisms we have discussed are commonly applied to simple chemical reactions, as well as to complex biomacromolecular reactions. Multistep mechanisms for macromolecular reactions reveal the concept of conformational changes initiated by binding events that are so prevalent in biology. In this case, the step(s) following the initial binding event is probably related to the nestling of the loosely bound dye molecule into the enclosed, confined binding pocket of the DNA minor groove. This reaction step probably also requires small dynamic conformational adjustments of the DNA molecule. The second of the two-step models [Eq. (4)] is kinetically equivalent to the first, but requires that the dye completely dissociate from the site to occupy it again, possibly in a different orientation. Importantly, high-resolution X-ray diffraction data 29 corroborate our two models. It is necessary to remember that there are certainly other mechanisms that can account for the data. Kinetic experiments can never prove the exclusive validity of one particular model, and binding of p-OH Hoechst could be more complex than a two-step mechanism. However, combining equilibrium experiments with a variety of kinetic experiments can provide the rationale for understanding and quantifying fundamental steps of the binding mechanism, and sheds light on the general physical principles that produce highly selective and extremely strong interactions involving complex biological macromolecules, such as these dyeDNA interactions. Acknowledgments This work was initiated with the support of (N)FWO and (B)OZF (to EG.L.). We thank Kenneth T. Douglas and Seyed E. Ebrahimi for their gift of the Hoechst derivatives and Etienne Wouters and Eddy Van Outryve for their help in constructing and optimizing the equipment.

234


[1 11

[ 1 1 ] Scanning Force Microscopy of Nucleic Acid Complexes By PETER T. LILLEHEIand LAWRENCE A.

BOTTOMLEY

Introduction Small ligands bind to DNA by (1) intercalation between adjacent base pairs, (2) binding within the major or minor grooves, (3) nonclassic modes] or (4) a combination of these modes. Definitive assays for mode of binding are threedimensional (3D) structure determination by X-ray diffraction or nuclear magnetic resonance (NMR) spectroscopy. These labor-intensive techniques are applicable only to binding studies of short DNA fragments because of difficulties in obtaining diffraction-quality crystals or interpretation of complicated chemical shift data. Their use is often precluded by lack of site specificity, rapid exchange, or multiple binding modes. Assays suitable for long DNA fragments involve viscometry, sedimentation, and linear and circular dichroism. These methods are reliable when ligands bind by conventional intercalative or minor groove modes, 2 but can be confounded by mixed and nonclassic modes. All traditional assays require large quantities of material. Whereas traditional techniques average over population and time, microscopic imaging techniques enable structural determinations of complexes from single molecules that are effectively frozen in time and space. Evidence of the mode and site of drug binding is, in principle, directly attainable from image analysis of a small number of nucleic acid molecules. Optical microscopy has insufficient spatial resolution for visualization of most drug-DNA complexes. The elaborate sample preparation techniques required for electron microscopy have precluded its use as an assay for drug binding. Scanning force microscopy (SFM, also known as atomic force microscopy) enables direct visualization of single DNA strands. A tip, fabricated onto the end of a readily deformable, rectangular or triangular-shaped cantilever, is positioned near or on the surface of the sample to be imaged. The tip and sample are moved with respect to one another and the cantilever bends in response to surface topography. Laser light is reflected off the cantilever onto a position-sensitive photodetector to track cantilever deflection. An image is produced by plotting cantilever deflection as a function of the position of the tip on the sample. SFM has advantages over optical and electron microscopy for imaging nucleic acids. Images can be acquired in vacuum, under liquid, or in air at ambient, 1 L. A. Lipscomb, K X. Zhou, S. R. Presnell, R. J. Woo, M. E. Peek, R. R. Plaskon, and L. D. Williams, Biothemistry 35, 2818 (1996). 2 D. Suh and J. B. Chaires, Bioorg. MeN. Chem. 3, 723 (1995).


Copyright@ 2001 by AcademicPress All rightsof reproduclion in any lbrm reserved. 0076-6879/00 $35.00

[1 1]

SFM OF NUCLEICACID COMPLEXES

235

cryogenic, or elevated temperatures. Sample preparation is minimal. Addition of a stain or contrast agent, replica formation, or application of a conductive coating is not required. As a result, SFM is inherently a higher resolution technique. SFM has limitations. Images of nucleic acids can be obtained only when the molecule is physisorbed or chemisorbed to an atomically flat substrate. Image features are dependent on tip geometry and structural details are limited only to the portions of the molecule in direct interaction with the tip. Analysis and interpretation of images of nucleic acids require (1) assessment of tip shape and diameter as well as how these parameters may have changed during the course of imaging, (2) consideration of elastic and inelastic sample deformation, and (3) knowledge of the sources and treatment of image artifacts. Even with these limitations, the capability of acquiring subnanometer scale images of single molecules in vitro renders SFM superior to other forms of microscopy for determining nucleic acid structures.

Instrumental Considerations Imaging Modes Three modes are used to image nucleic acids: contact, noncontact, and intermittent contact imaging. Schematics of each are given in Fig. 1. Each has advantages and limitations. Contact Mode. As the name implies, with the contact imaging mode the tip is kept in contact with the sample. Contact mode images are obtained by plotting cantilever deflection as a function of the xy coordinates of the tip on the sample. Distortion of soft surfaces (or compliant molecules adsorbed on them) is commonplace with this imaging mode because of variation in the vertical force exerted by the tip on the sample. To image under fixed vertical load, a feedback mechanism is required. As the cantilever bends in response to surface topography, the scanner is retracted or extended to return cantilever deflection to its original value. Images are obtained by plotting the z piezovoltage as a function of the xy coordinates of the tip on the sample. Because the movements of the scanner are calibrated to subangstrom dimensions, accurate topographic images are obtained with properly tuned scanner feedback control. High-quality images of DNA can be acquired in just a few minutes with this imaging mode. Minimal vertical loading must be used to eliminate sample deformation or damage and tip wear. In the absence of strong adsorbate-substrate interaction, molecules will be readily displaced with this imaging method. When samples are imaged in air under ambient humidity, a thin film of water forms on the sample and tip surface. Contact between these water layers produces increased vertical and lateral forces that cause severe artifacts in the image. 3 Whenever possible, contact mode imaging should either be performed under a dry atmosphere or under fluid.

236


a

I

11 11

%

FIG. 1. Schematic of the tip-surface interaction during (a) contact mode, (b) noncontact mode, and, (c) intermittent contact mode imaging.

Noncontact Mode. The noncontact imaging mode maintains the sample and probe tip in close proximity while scanning. The cantilever oscillates at its resonant frequency and any change in the forces acting on the cantilever (magnetic, electrostatic, and/or van der Waals forces) will be manifested as a shift in resonant frequency. Topographical information is obtained by monitoring resonant frequency shifts as a function of the xy coordinates of the tip above the sample. Tip wear and sample deformation or damage are eliminated in noncontact mode imaging. Because the forces are often small, tip-sample separation must be minimized. This mode is best suited for imaging under vacuum, in which van der Waals interactions are not dampened by intervening liquid or by adsorbed gases on the surface of the tip and sample. 3 T. Thundat, R. J. Warmack, D. E Allison, L. A. Bottomley, A. Lourenco, and T. L. Ferrell, J. Vac. Sci. Technol. A. 10, 630 (1992).

[1 1]

SFM OF NUCLEICACID COMPLEXES

237

Intermittent Contact Mode. Presently, most biological samples are imaged us-

ing intermittent contact. In this mode, the cantilever is oscillated at its natural resonance or driven at some selected frequency. Separation of the sample and tip is reduced until the oscillation of the cantilever is dampened. Images are acquired by tracking the shift in either amplitude or phase of oscillation as a function of the xy coordinates of the tip above the sample. Maps of sample topography are produced by recording shifts in amplitude whereas maps of sample compliance are produced by recording shifts in phase. Resolution is enhanced with intermittent contact imaging compared with contact imaging because the reduced tip wear allows for imaging with sharper and more delicate probe tips. Sample and tip damage is reduced because the tip is not shearing the surface; rather, its motion is always normal to the sample. Sticking of the tip to the surface is minimized because tip momentum is sufficient to overcome surface adhesion forces. The method by which the cantilever is oscillated varies among microscope manufacturers. Digital Instruments (Santa Barbara, CA) uses TappingMode whereas Molecular Imaging (Phoenix, AZ) uses MAG mode. In TappingMode, a piezoelectric oscillator is placed in contact with the cantilever holder and the entire cantilever assembly is oscillated at a desired frequency.4 The MAG mode oscillates a cantilever, coated with a magnetic material, with an induction coil placed beneath the sample. The oscillating magnetic field emanating from the coil drives the cantilever.5 Cantilever oscillation amplitudes are reduced compared with TappingMode. The relative advantages and limitations of each method are the subject of an ongoing debate in which we chose not to engage. In the context of this chapter, we shall treat both modes as equivalent. Tip and Scanner Calibration Cantilevers, tips, and the scanner are critical components of the microscope. Commercially available rectangular and V-shaped cantilevers are appropriate for nucleic acid imaging. The latter provides lower mechanical resistance to vertical deflection and high resistance to lateral torsion. 6 The size, shape, and material composition of an SFM tip have important consequences on the imaging of nucleic acids. Because the resultant scanning probe microscopy (SPM) image is always some convolution between the tip shape and image topography, tip artifacts will always be intrinsic to the technique.

4 D. A. Waiters, J. P. Cleveland, N. H. Thomson, P. K. Hansma, M. A. Wendman, G. Gurley, and V. Elings, Rev. Sci. lnstru. 67, 3583 (1996). 5 W. H. Han, S. M. Lindsay, and T. W. Jing, Appl. Phys. Lett. 69, 4111 (1996). 6 R. Howland and L. Benatar, in "A Practical Guide to Scanning Probe Microscopy." Park Scientific Instruments, Sunnyvale, California, 1996.

238


[ 1 1]

Even though only the end of an SFM probe tip interacts with the molecule, the size of the tip dramatically affects the quality of images of nucleic acids. Tip irregularities away from the end are of importance only when imaging larger structures such as nucleosomes of chromatin. 7 For simplicity, the shape of the end of the tip is routinely assumed to be spherical when estimating actual feature widths from the apparent widths seen in micrographs. 8 The relationship between the observed diameter, Dobserved, and tip radius, Rtip, is given in Eq. (1). Dobserved =

4

-I/2 (Rtiprfeature) -

(1)

Equation (1) is applicable only when the radius of the feature, rfeature, is less than or equal to the radius of tip. When Rti p 1 M (as the water concentration is about 55 M, this means >98% occupancy at Ka = 1 M). Typically, a relatively small number of water-DNA contacts is observed, some in the major groove and others in the minor groove. This is largely a technological problem related to resolution and the fact that there are comparatively few protons available to report interactions with water. It is not possible to measure an NOE between water and ring nitrogen or phosphate atoms because of substantial CSA relaxation. Equations (14)-(16) imply that the cross-relaxation rate constant is measured in both the laboratory and rotating frames. Usually, however, a single mixing times or perhaps two mixing times are used for each experiment, and the ratio of the peak intensities is taken as a measure of r. However, this requires that the spectra are also accurately scaled, which is difficult with independent experiments of different types. It is therefore convenient to use an internal reference. A convenient one is the cross-relaxation between Cyt H5 and Cyt H6, for which the distance is known and fixed. The ratio of the cross-peak intensity (actually area in cross-section through the water frequency) to that of the standard scales the desired ratio in Eq. (16) by exactly 2, because for the cytosine the correlation time is in the slow tumbling limit, and J(0) >> J(2o)). Taking the ratio also reduces some of the error associated with longitudinal relaxation, allowing longer mixing times to be used than would 65 G. Otting, E. Liepinsh, and K. Wtithrich, Science 254, 974 (1991).

268


[ 12]

otherwise be the case. Because cross-relaxation in the rotating frame is up to twice as fast as in the laboratory frame, it is usual to use mixing times in that ratio (which would then give approximately equal intensities if other relaxation processes are the same). For moderate-sized oligonucleotides at 5 to 1@, where the rotational correlation time is < 10 nsec, mixing times of 50 msec (NOESY) and 25 msec (ROESY) seem to give good results. 64 Another major difficulty, and one that can seriously affect quantitative aspects of this kind of experiment, is that effective T1 relaxation is much faster than expected because of radiation damping, which becomes increasingly efficient as the magnetic field strength is increased. To avoid this problem in NOESY, a flip-back kind of experiment should be u s e d . 66 Once the cross-peaks have been identified and quantified, it is necessary to account for other artifacts that can influence the analysis. One is that there may be protons resonating at the same frequency as the water, and give rise to an ordinary NOE (e.g., H3' to H8). Although this is unlikely to be significant for C2'-endo sugars and short mixing times, the effect can often be ruled out by changing the temperature, as the water frequency is rather strongly temperature dependent compared with solute protons. A much more severe problem arises from exchangeable protons. A proton that exchanges with solvent during the mixing time gives rise to a cross-peak. In NOESY, this will be the same sign as the NOEs, although such peaks are usually also much broader. In ROESY, the exchange cross-peak is of opposite sign to direct ROE peaks (but in ROESY, spin diffusion changes the sign of the cross-peak). Thus exchanging protons can be readily identified. However, if a proton exchanges from, say, an NH2 group to water, and the NH2 is also close to a nonexchangeable proton, then there will be an indirect interaction between water and the target proton as far as magnetization transfer is concerned. Unfortunately, this exchangemediated transfer will have a normal ROE sign, and cannot be distinguished from a direct ROE except by careful mixing of time dependence studies and calculations based on known structures. Often potential water-solute cross-peaks are ignored if there is an exchangeable proton within 5 • of the target proton. This may be unnecessarily restrictive; if the exchange rate constant can be estimated from the exchange peak with the water, then the geometry and simple simulation of time courses can help discriminate between genuine direct NOE (ROE) and exchangemediated effects. These latter effects tend to become important in drug-DNA c o m p l e x e s . 66,67

Spectrometer Requirements For titrations, typically one needs about 0.5 ml at 0.25-0.5 mM DNA (and therefore > 5 - 1 0 mM ligand solution) in a buffer that does not contain protonated 66 G. Otting, Prog. Nucl. Magn. Reson. Spectrosc. 31, 259 (1998). 67 H. E. L. Williams and M. S. Searle, J. Mol. Biol. 290, 699 (1999).

[ 12]

NMR STUDIESOF DRUG-DNA COMPLEXES

269

salts. Sodium phosphate plus KC1 is a good general salt system that buffers close to pH 7, and is essentially independent of temperature. At lower pH values, deuterated acetate can be used, but should not be lyophilized as acetic acid escapes. Titrations can be done in both H20 and D20 buffers. For He0, it is necessary to suppress the strong solvent signal; the pulsed gradient-based techniques such as Watergate 68 work well, and avoid the problems of saturation transfer that occurs with presaturation, as well as giving much better suppression. Titrations need to be carried out to greater than the stoichiometric point. For weakly soluble drugs, one can prepare a series of solutions of drug-DNA complexes by adding the DNA to the dilute drug (possibly in a nonaqueous solvent). The aromatic drugs are not always soluble in water without extensive aggregation. This creates a problem for binding and stoichiometry at NMR concentrations and also for determining the reference ligand shifts in the absence of DNA. For structural analysis, a 1 : 1 complex can usually be prepared by adding excess ligand, and resolving on a small desalting column. 4° This works only for tightly binding ligands. Solvent suppression techniques, and proton-detected X experiments require pulsed-field gradient capability (now common with modern spectrometers). The gradients are used for dephasing signals and coherence selection. HSQC experiments can easily be run at natural abundance (13C) at millimolar DNA-drug concentrations. Large z gradients (> 50 G/cm) can also be used for measuring the translation diffusion coefficient with the pulsed field gradient spin-echo method, 53'54 which can be valuable for investigating the hydrodynamic properties of the DNA and the DNA--drug complex (cf. bending). For this application, the gradient must be carefully calibrated using, for example, doped HOD and also against systems having known translation diffusion coefficients. The gradient linearity also needs to be checked; it may be necessary to restrict the length of the sample, using Shigemi tubes (we routinely use 16 to 18 mm for this purpose). In general, an inverse 5-mm probe can be used as the standard for modern spectrometers. If aggregation is a problem, more dilute solutions can be studied with minimal sensitivity cost, using 8-ram-diameter tubes (also available as Shigemi tubes, using about 1 ml). With a dilution of 3-fold, there is a small reduction in sensitivity, but may be sufficient to reduce dimerization to a manageable level. 69 The sample requirements are also reduced by using the increased sensitivity associated with 18.8-T spectrometers, which allow titrations to be carried out at 100 # M or less, and good quality two-dimensional spectra can be acquired at 25 # M with 8-mm tubes or a cryoprobe. For quantifying peak areas or volumes, it is essential to produce spectra that have fiat baselines or base planes, are free of ridges and spikes, and are adequately digitized. For measuring coupling constants from double quantum filtered correlation spectroscopy (DQF-COSY) or primitive 68 M. Piotto, V. Saudek, and V. Sklenar, J. Biomol. Struct. 2, 661 (1992). 69 J.-L. Asensio, T. Brown, and A. N. Lane, Struct,re 7, 1 (1999).

270


[ 1 21

exclusive correlation spectroscopy (PE-COSY) spectra, digitization and resolution are important, as also is sensitivity. Poorly digitized, noisy spectra do not generally give reliable coupling constants or peak volumes. 25 For these purposes, high fields and long acquisition times are essential. In NOESY spectral peak volumes are also sensitive to saturation effects from rapid recycle times. Either long relaxation delays must be used (based on measured T1 values), or careful normalization of the peak volumes is required if fitting to the peak volumes is to be attempted. 59 Calculation Strategies The analysis of the three bond coupling constants will show whether the DNA and drug are predominantly to be found in a single conformer, or whether multiple state averaging has to be considered. If the former approach is chosen (and justified by the experimental data to hand), then the calculation strategy becomes very simple, and the same as for unliganded DNA. 26 Simulated annealing protocols and restrained molecular dynamics (rMD) with and without time averaged restraints (TARs) 26'7°'71 will be effective in refining the structure, from which many important conclusions can be derived about the nature of the interacting groups. A non-TAR structure can give misleading results about the extent of the fluctuations and how they are sampled compared with the free DNA, and therefore cause an important component of the overall affinity, and possibly specificity, to be missed. It is also clear that the free and ligand-bound structures and dynamics must be compared for any conclusions about affinity or stability to have any justification whatsoever. It is usually obvious where the binding site is from a cursory analysis of patterns of chemical shift perturbations and intermolecular NOEs (see below). The latter usually also indicate the preferred orientation of the drug in the binding pocket. However, other information may indicate that there are two possible orientations, or multiple overlapping binding sites, in which case the structure analysis becomes complicated; some way is needed to assign the NOEs to different possible states. Under these conditions, forcing a unique structure on the data could be misleading when relating it to specificity and affinity. There is no alternative to doing higher level structure calculations, using one of the ensemble methods or, at the very least, MD-TAR. The dynamics simulations also provide valuable information about flexibility of the system on the subnanosecond time scale, which can be compared with the experimental relaxation data. Order parameters extracted from 13C relaxation data can be compared with those calculated from the autocorrelation functions constructed from the dynamics trajectory. For preference, the system should be simulated in H20, as this generally gives superior results, and is in any case essential for calculating nanosecond dynamics of 70T. E. Torda, R. M. Scheek, and W. E van Gunsteren,Chem. Phys. Lett. 157, 2890 (1989). 71 U. Schmitz,N. B. Ulyanov,A. Kumar,and T. L. James,J. Mol. Biol. 234, 373 (1993).

[ 1 2]


271

/sNH 2 X H2N

X: -N=N-NH-

berenii

-O-(CH2)3-O-

propamidine

-O-(CH2)s-O"

pentamidine

FIG. 1. Chemical structure of bisamidine compounds.

the systems. Furthermore, it also allows one to analyze the positions and persistence of individual hydration sites, which can be compared with the (probably limited) information from the NMR data. Taken together, the structure, conformational flexibility, dynamics, and hydration data can be used for more detailed calculations of drug binding (such as SASA; see [6] in this volume59a), and for parsing the free energy relationships. Specific E x a m p l e s

Minor Groove Binding by Bisamidinium Compounds Members of the bisamidinium class of compounds (Fig. 1) are simple and cytotoxic. They bind to the minor groove of DNA with moderate affinity (l-10 # M range) and with a preference for AT-rich sites (Table I). 72 Extensive correlations of affinity and potency are available, and there are several X-ray and NMR structures of these drugs bound to DNA. 73-79 Titrations of the DNA with ligand always showed fast or fast intermediate exchange, implying moderate affinity (in agreement with footprinting and optical methods). Figure 2 shows NOESY spectra of the d(CGCAAATTTGCG)2 : propamidine complex in H20, and titration shifts for selected DNA resonances. The titration shifts (Fig. 2B) show that the binding site is in the AT-rich central region. Furthermore, the symmetry and extent of the shift profile suggest an extended, asymmetric

72 C. Zimmer and U. Wahnert, Ppvg. Biophys. Mol. Biol. 47, 31 (1986). 73 K. J. Edwards, T. C. Jenkins, and S. Neidle, Biochemistry 31, 7104 (1992). 74 Y. C+ Jenkins, D. G. Brown, S. Neidle, and A. N. Lane, Eur J. Biochem. 213, 1175 (1993). 75 T. C. Jenkins and A. N, Lane, Biochim. Biophys. Acta 1350, 189 (1997). 76 M. R. Conte, T. C. Jenkins, and A. N. Lane, Eur. J. Biochem. 229, 433 (1995). 77 S. Hu, K. Weisz, T. L. James, and R. H. Shafer, Eu~: J. Biochem. 204+ 31 (1992). 78 K. R. FOX,C. E. Sansom, and M. E G. Stevens, FEBSLett. 266, 150 (1990). 79 C. A. Laughton, T. C. Jenkins, K. R. Fox, and S. Neidle, Nucleic Acids Res. 18, 4479 (1990).

272


[ 12]

TABLE I DRUG-DNAINTERACTIONS:NMR STUDIES Ligand

Footprint (bp)

Kd (298 K) (/zM)

Site (NMR)

Berenil

3.5

Propamidine Pentamidine Hoechst 33258

4 4.5 4-5

1-2 > I0 4 ~5 >5 0.01

AAT ATG TAA AAT AATT AAATT

4

2

YpG

11

2

Minor groove

lntercalator

Nogalamycin Major groove

Triplex 1l-mer

AGAAGAAAGGA

binding site such that the drug can bind in both orientations with essentially equal affinity, and that exchange between the binding modes is fast on the NMR shift time scale. The shift profile indicates an effective binding site of about 4 bp, in reasonable agreement with the DNase I footprint. As Table ! shows, the shift profile generally is a good indicator of the binding site size for this class of ligands. NOEs between the ligand protons and DNA protons (Fig. 2A) were exclusively to the minor groove (AdeC2H and H I ' plus some H4' and H5'/H5"), and to the bases expected from the shift profile. This confirmed the location of the binding site in the minor groove. The chemical shifts of the ligand also change on binding to the minor groove (Table II). In general, the downfield shifts increase with increasing site size, except for pentamidine, which can effectively bind only to the longest AT tract. Docking and energy minimization showed that the binding site was 3 base-pairs for berenil, and not symmetrically displaced with respect to the pseudodyad in AATT and AAATTT tracts. Similar results were obtained for berenil, propamidine, and pentamidine acting on AATT and AAATTT. The main differences were that berenil can occupy more than one site in the extended AT tracts. For propamidine, essentially all resonances were in fast exchange. The largest shift difference was about 0.5 ppm (250 Hz at 11.5 T), implying an exchange rate constant of at least 2000 sec -1. However, berenil binding showed intermediate exchange behavior. The line width is smallest in the free DNA, increased during the titration, and finally narrowed again on reaching saturation. The variation of the line width with drug concentration was analyzed according to the fast exchange equations: LW - LW ° = 47rpa(1 - p a ) A v 2 / k

(17)

[12]

N M R STUDIES OF DRUG-DNA COMPLEXES

273

A PNH 2

P3,5

P 2,6

4,5" =

G

- r 7 H l ' / 91b • °

5.5-

~

e

Q

O

6.5 Q

7.5

o

1

,0*0

8.5 ¸ I , i 9 1 8.7

,

i 8.3

' I7'9I ' ' ' '7'5

7.1

6.7

6.3

chemical shift/ppm

B 0.4

B 0.2 E

-0.2

[]

[] []

-0.4

-0.6

C

G

HI' H2 H4'

C

A

A

A

T

T

T

G

C

G

FIG. 2. Interaction of d(CGCAAATTTGCG)2 with propamidine. (A) NOESY spectrum of the DNA-propamidine complex recorded in HeO at 277 K. P, ligand propamidine. (B) Titration shifts; &(complex) - S(DNA) for selected protons. [Reproduced with permission of Elsevier from A. N. Lane, T. C. Jenkins, and T. A. Frenkiel, Biochim. Biophys. Acta 1350, 205 (1997).1

where Pa is the m o l e fraction o f the unliganded D N A , A v is the frequency difference in hertz, determined from the free and fully saturated shift values, and k is the rate constant = kl[L] ÷ kl. As Kd was known, the on and off rate constants could be determined, koff = 500 sec -1, and kon = 5 × 108 M - I sec - l ,

274


[ 121

TABLE II SHIFTS OF BISAMIDINECOMPOUNDS BINDING TO DIFFERENT TARGETS IH chemical shift (ppm)

Ligand Berenil Propamidine

Pentamidine

Proton

Free

A2Ta

A2T2h

A3T3'

2, 6 3, 5 2, 6 3, 5

7.67 7.84 7.16 7.78 4.34 2.33 8.63 8.30 7.14 7.77 4.21 1.90 1.66

7.68 7.87 7.29 7.87 4.38 2.36 --------

7.68 7.90 7.42 8.00 4.48 2.41 8.78 8.63 7.31 8.05 4.47 1.93 1.70

7.97 8.18 7.52 8.10 4,56 2.48 8,95 8,85 7.21 7.95 4.33 1.87 1.61

NH2a NH2b 2, 6 3, 5 c~ fi y

a d(GCAATGAGCG),d(CGCTCATTGC). t, d(CGCGAATTCGCG)2. c d(CGCAAATTTGCG)2.

which is close to that expected for a diffusion limited reaction, indicating minimal reorganization of the initial d r u g - D N A complex (i.e., a preorganized site). For other ligands that bind less tightly, the dissociation rate constant was substantially larger, suggesting that the affinity for these ligands is determined by the off rate constant. The ligands showed only small effects on coupling constants, glycosyl torsion angles determined from NOE build-up curves, and 3tp N M R spectra, which is consistent with minimal perturbation of the D N A structure on forming the complexes, 74,75 and agrees with the binding kinetics (see above). This was confirmed by analyzing the structures of the DNA with and without ligand (Fig. 3); the structures were essentially identical within the experimental error and the limitations of the calculations. In the asymmetric D N A fragment, it was shown that there is some tolerance of a GC base pair by both berenil and propamidine, 76 which agrees with the observed low sequence selectivity (about 10-fold). 78,79 The structures all showed insignificant DNA distortion, and that the ligands make numerous good van der Waals contacts with the walls of the minor groove. The amidinium groups are able to make electrostatic interaction with the phosphodiester backbone. The larger ligands may bind their optimal target more weakly than berenil because the latter is preformed in the correct conformation, whereas the flexible

[12]


275

FIG.3. Solution structure of propamidine:d(CGCAAA'lTI'GCG)2.The structure was determined with rMD (Laneet al.64). The boundligand is knownin black, viewedend-onfrom the minorgroove.

linker in propamidine and pentamidine gives an unfavorable entropic contribution to binding. d(CGCAAATTTGCG)2 and d(CGCGAATTCGCG)2 both showed the expected water of hydration in the minor groove adjacent to the Ade residues; all AdeC2H showed negative NOEs to water. On binding propamidine, some water molecules were displaced, but others remain associated with the DNA, and probably also with the drug molecules. Thus, the amount of water displaced by drug binding is less than might be estimated simply by displacement of the minor groove hydration. Potential energy calculations give the correct position for the binding

276


[ 12]

OH (HaC)2HN+~,.,.~

"r-o

o

Vy% # OH

O

OH

.~

0-1"-1-o4 .

H 3 C O ~

OCH3 C

G

A

1"

FIG. 4. Structureof nogalamycinand its complexwith DNA. Top:Nogalamycinstructure. Bottom: Structure of nogalamycinintercalated into the TpG/ApCsite. 67

site, and rank order the affinities for alternative sites. 75 A detailed parsing of the free energy has been carried out by Jenkins and co-workers (see [6] in this volume59a).

Intercalation by Nogalamycin The anthracycline antibiotic nogalamycin (Fig. 4) binds preferentially to YpG steps of DNA by intercalation, with the sugar in the minor groove. In contrast to the minor groove binders, the intercalators induce major local rearrangements and distortion of the DNA, which may be paid for by intrinsic binding energy; intercalators generally have rather low affinity for DNA. The structure of DNAnogalamycin complexes has been solved by both X-ray diffraction and NMR. The NMR spectrum is characterized by large changes due to the intercalation, in both 1H and 31E Indeed, the large (1-ppm) downfield shift of the YpG is characteristic of a B[-to-Bii transition. This is a way for the DNA to unwind, as intercalation

[12]


277

requires separation of the two base pairs. The CSA of the affected phosphate increases, as one might expect as the symmetry and therefore bonding about the phosphorus atom are changed. However, the new conformation appears to be quite rigid by 31p relaxation.4 The structure of phosphodiester backbone necessitates a large change in the backbone torsion angles, which is primarily achieved by the B~-to-Bli transition. However, some other local backbone torsions alter to accommodate the intercalator. Thus, in the 1 : 1 complex with d(ATGCAT) • d(TACGTA), nogalamycin intercalated at the TpG/ApC step. The T moved primarily into the N domain of sugar conformations, whereas the other sugars remained mainly in the S domain. The CpA step, however, showed the transition to the BH conformation (e, ~" = - 6 0 , 180), with a concomitant change in G(A) to 180 ° (trans) and a compensating change in el(A) from g - to g+ (crankshaft motion). Other residues showed minimal changes in local conformations. These data show that the binding is notably asymmetric across the two affected base pairs, 67 reflecting the nonsymmetric structure of the drug, and the lifting of the degeneracy of the self-complementary DNA duplex. To accommodate the aromatic ring, the helical twist is now only about 20 °, compared with the usual 36 ° in B-DNA, and the helical rise is now 7 A, reflecting the sandwich of the aromatic intercalator adding about 3.4 ]k. In addition, other helical parameters for the 2 base pairs forming the intercalation site are altered, and even in restrained MD, showed substantial fluctuations about the average values. Larger fluctuations would be expected in a free dynamics trajectory or using the MD-TAR protocol. 26'7° Thus the local distortion of the DNA by the intercalator is severe, and must cost substantial potential energy. However, the destacking is presumably compensated for by making new van der Waals and electrostatic interactions with the intercalating ring system. Despite the severe local distortion imposed by the intercalator, the remainder of the molecule is normal B-DNA, and the effects do not propagate far along the double helix. This is in accord with many other observations of the inherent flexibility of the DNA backbone, that major distortions are rapidly damped out by the large number of degrees of freedom in the DNA duplex. 7'34 The hydration of the nogalamycin-DNA complex was also examined by NMR methods and simulation, 67 which could be compared with information from the X-ray structures. A remarkable feature is the incomplete agreement among the methods. Some of this is likely to be due to the inherent limitations of each method; NMR measures a time-dependent water residence and can detect only waters that persist for longer than about 100 psec close to a solute proton (i.e., no water associated with the phosphates can be detected by NMR), whereas X-ray and simulation detect mainly population averages. X-Ray crystallographic refinement methods and resolution have been shown to affect the estimate of water, and there are certainly limitations in the treatment of water potential in MD simulations. Nevertheless, significantly bound water molecules (residence times >0.5 nsec at 15°) were detected in the major groove, and indications of hydration of the bound

278


[ 121

drug were obtained. The simulations showed some interesting correlations with the experimental data, especially regarding the hydration of the bound drug, and suggested that there may be water-mediated interactions between drug side chains and the DNA.

Major Groove Binding: Triplex-FormingOligonucleotides In addition to minor groove recognition and intercalation, it is possible to recognize the major groove with oligonucleotides, making use of Hoogsteen hydrogen bonding to a polypurine stretch. The most promising avenue in this area of antigene appears to be the parallel motif, despite the need to protonate the cytosine at N3. High affinity and specificity have been achieved with modified bases, such as including a positive charge on the base, and adding propanyl chains at the 5 position. To date the only full and detailed conformational information about DNA triplexes has come from NMR spectroscopy. There are now several structures of the parallel motif, and the basic observations are the same; the minor groove of the underlying duplex is compressed, and the helix axis is displaced about 2 A. The sugars in the Hoogsteen strand, especially the protonated cytosines, appear to be less S state than usual, consistent with the notion that the most stable triplex is one in which the Hoogsteen oligonucleotide has C3'-endo sugars. 5'8°'81 This is shown by the relative stability of triplexes containing a DNA third strand and an RNA third strand. 8° 82 The presence of the positive charge on the protonated cytosine residues was verified by 13C NMR spectroscopy, which showed that both the chemical shifts and the one-bond C H coupling constant (C5 and C6) reflect the different electron density of CH + and C (Fig. 5). 17'69 The charge also appears to alter the local conformation of the purine residues, and was associated with a transition between y(g+) and y(t) of the purine nucleotide 5' to the CH + residue. The value of y was determined from a combination of coupling constants at H4' and NOEs and unusual shifts for H3' and sequential NOE. In the usual g+ rotamer, the H4' as a small coupling to H3' (S sugar) and two equal, small coupling to H5' and H5". The sum of the coupling is less than 12 Hz. However, for the t conformer, one of the couplings is large (12 Hz) and the other is small, leading to a splitting of the H4' resonance. This can be seen readily in cross-section of an NOESY through the H3'-H4', with high digital resolution along F2. In addition, in the DQF-COSY, the appearance of a strong H4'-H5'/H5" cross-peak and of NOEs between H4', and H5'/H5" also proves the presence of g - or t (Fig. 6). The NOEs in turn allow a stereospecific assignment to be made. In all cases that we have examined, the presence of y (t) is accompained by an unusual up field shift of H3', which shows an NOE to so R. F. Macaya, P. Schultze, and J. Feigon, J. Ant, Chem. Soc. 114, 781 (1992). Sl J.-L. Asensio, R. Carr, T. Brown, and A. N. Lane, J. Am. Chem. Soc. 121, 11063 (1999). 82 H. Han and E B. Dervan, Proc. Natl. Acad. Sci, U.S.A. 90, 3806 (1993).

[12]

NMR STUDIES OF DRUG-DNA COMPLEXES

279

H6/H8 E

136_

G3 o..4-qp~o Gqo ~ o

Q.

140_ = 0

•~

144 _

61 0

148-

"¢: 152 O (,3 e~ 156 -

C8

~ C9+C10

C6

~ C15+C16

o

otO~

od AC2

83 7:8

C2

7;5 7.~ 6~9

1H Chemical Shift (ppm)

FIG. 5. HSQC spectrum CH +. 13C-IH HSQC spectrum of d(AAGGAA) - d(TTCCTT), d(TTC + C+TT) recorded at 14.1 T and 30 °, showing the C8, C6 and C2 regions. The protonated carbons (15 and 16) have significantly different shifts and nJcH from C9 and C10. ]Reproduced with permission from J.-L. Asensio, T. Brown, and A. N. Lane, Nuclei(' Acids Res. 26, 3677 (1998). Oxford University Press.]

H4' HS' HS"

~-81

_

~

= r

o

H4' HS' HS"

381

~J

oo

& .

~

HS' HS"

46t

48:

H3'

" r~ 5 . 8 5.96.0" G.1-

• 4.5S

. GS H4' O.Z4' O.051F(hkl)lmax, O.051F(hkl)lmax > IF(hkl)l > O.lOIF(hkl)lmax, O. lOIF(hkl)lmax > IF(hkl)l > O.151F(hkl)[max, etc.

286


b

[ 13]

"

1016

\ 4~

- elongat~

S

FIG. 2. Precession photograph of a crystal of the DNA-porphyrin complex CGATCG-CuTMPyP4 [Cu(II)meso-(4-N-tetramethylpyridyl)porphyrin,NDB entry DDF060].I6 This 6 ° screened precession photograph, taken with Ni-filtered Cu radiation, recorded the hOI layer. In this crystal (a = 39.49 ~, c = 56.15 ~, c~ = 90 °, ~ = 120, space group P 6122) the pseudo-helical axis is nearly parallel to the crystallographic c axis, causing extremely large values of IF(1 0 16)1. Anisotropy in crystalline order causes the diffraction pattern to fade out at a lower angle along the a* axis than along the c* axis. The elongated cross is caused by rotational disorder about the helical axis. The 6-fold screw axis along the c axis is indicated by systematic absences along the c* axis.

weak in the skewed helical axis DNA, and ~80% are weak in the pseudo-infinite helical axis DNA. The effects of a pseudo-infinite helical axis of DNA on the diffraction pattern can been observed directly in the precession photograph shown in Fig. 2. In this case (a DNA-porphyrin complex), the pseudo-infinite helical axis is directed nearly

[13]

X-RAY CRYSTALLOGRAPHYOF DRUG-DNA COMPLEXES

li

"

0

d,

I

.........

0.0 0.5

287

X

7.2 Al [3.6A t -0.5

FIG. 3. Patterson map from a crystal of duplex [d(CGTACG)]2 bound to the bisintercalator D232 (a = 28.24 A, b = 28.24 ,&, c = 72.74 ~, e~ = 90.0,13 = 90.0 ~, 3' = 120.0~) .14 The pseudo-infinite helical axis is parallel to the crystallographic c axis.

along the crystallographic c axis. Extremely intense reflections with Miller indices 1 0 16, 1 0 -16, - 1 0 16, and - 1 0 - 1 6 are observable in the photograph. The origin of differences in dynamic ranges of IF(hkl)l from globular proteins and DNA crystals can be understood in part from a Patterson analysis. In a globular protein crystal, interatomic vectors generally have random lengths, directions, and origins. By contrast, a B-DNA or intercalated B-DNA crystal will yield a large number of"stacking vectors" with conserved lengths ('--3.4 A, 6.8 ,~, 10.2 A, etc.), directions (along the helical axis), and semiconserved origins (from the planes of base pairs). When the B-DNA is organized in an end-to-end fashion in the crystal (with a pseudo-infinite helical axis), then all the stacking vectors are nearly aligned. Therefore a DNA Patterson map can contain intense peaks spaced by 43.4 A. A Patterson map from a bisintercalated DNA crystal (NDB entry DD0018)14 with a pseudo-infinite helical axis is shown in Fig. 3. A Fourier decomposition of the electron density within a B-DNA or intercalated DNA crystal gives large amplitude waves with wavelengths of ~3.4 ~, ~6.8 ]k, ~ 10.2 ~, etc. For example, F(hkl)m.,~× in the data set collected by Hunter and co-workers 1~ on TGTACA-4'-epiadriamycin (Table I) has the Miller index 0 0 16. Dividing the length of the crystallographic c axis (52.39 ~) by l = 16 gives a d spacing for this reflection of 3.27 A. The Miller plane normal is parallel to t4 X. Shui, M. E. Peek, L. A. Lipscomb, Q. Gao, C. Ogata, B. R Roques, C. Garbay-Jaureguiberry, and L. D. Williams, Curl: Med. Chem. 7, 59 (2000).

288


[ 131

the c* axis. The spacing and direction of this intense reflection are consistent with base-base and base-intercalator stacking, and alignment of the helical axis along the crystallographic c axis. The second and third ranked F(hkl) in this data set have Miller indices l 1 16 and 0 2 16 (3.25-A d spacing with Miller plane normals within 15° of the c axis). Thus the Fourier decomposition of the electron density in this crystal contains three stacking waves of similar wavelength and direction. Those stacking waves have amplitudes nearly 40-fold greater than the median intensity in the data set and dominate the diffraction pattern. Their positions are tightly clustered in reciprocal space.

Data Collection Strategy One of the goals of most diffraction experiments is to collect data to the highest resolution possible. The highest resolution data have the lowest intensities, requiring long collection times, intense X-ray beams, and sensitive detectors. It is often impossible to collect weak high-resolution data simultaneously with strong stacking reflections. One part of the strategy employed in our laboratory is to collect stacking intensities independently from high-resolution data by collecting multiple data sets on the same crystal. The stacking intensities are collected with short exposures, low amperage on a rotating anode, or an attenuated the beam at a synchrotron. The short- and long-exposure data sets are scaled and merged. It is best to exclude the stacking reflections from the scaling, for example, by merging data from 20 to 2.6 A (short exposure) with data from 3.1 to 1.4 ]k (long exposure). In this case the overlap, used for scaling, is 3.1 to 2.6/~ and would not contain the ~3.4-2k stacking reflections. A second part of our strategy is to identify the stacking reflections prior to initiating data collection, and to set the collection parameters to ensure accurate measurement of their intensities. Detector overload and peak overlap must be avoided by empirical adjustment of exposure time and crystalto-detector distance. The stacking reflections are in close proximity in reciprocal space, and are generally broader than other reflections.

Crystallographic Anisotropy Elongated flexible molecules that lack lateral structural hooks exhibit characteristic types of crystalline disorder. In the precession photograph in Fig. 2, the diffraction pattern of a DNA-porphyrin complex is seen to extend out to the edge of the photograph along c*, but to fade out at lower resolution along a*. Thus the diffraction pattern is anisotropic, indicating that the disorder within the crystal is anisotropic. The DNA is more highly ordered along the helical axis than along the perpendicular to the helical axis. These systematic errors in the data can be attenuated, once the refinement is near completion, by anisotropic or local scaling

[13]

X-RAY CRYSTALLOGRAPHYOF DRUG-DNA COMPLEXES

289

of observed-to-calculated data. One cost of these corrections is that they naturally introduce additional parameters to the refinement. Additional information about disorder within the DNA-porphyrin crystal is provided by the faint elongated cross pattern, stretching between the intense stacking reflections (Fig. 2, "X"). This elongated cross is reminiscent of a DNA fiber diffraction pattern. Our interpretation of the origin of the elongated cross in Fig. 2 is that a fraction of the DNA within the crystal is disordered by random rotation about the helical axis, just as in a fiber. Structure Determination

Molecular Replacement Determination of helical axis orientation by Patterson analysis can be combined with symmetry information to simplify structure solution by molecular replacement. For example, the Patterson map and strong 0 1 16 reflection indicate that the helical axis of the CGTACG-porphyrin complex is nearly parallel to the c axis. Simple volume calculations and spectroscopic analysis of a dissolved crystal indicate that the asymmetric unit contains one strand of DNA plus one porphyrin molecule. If the DNA forms a duplex, then the duplex must be centered on a crystallographic 2-fold axis. The DNA-porphyrin complex must be centered on one of the two crystallographic 2-fold axes (in space group p61(5)22). Two possible orientations, which differ by 180 °, are possible on each 2-fold axis. Thus the molecular replacement search is limited to four one-dimensional translations. However, the success of molecular replacement always depends on an accurate search model. The CGTACG-porphyrin structure contains an unanticipated flipped-out base. Therefore molecular replacement failed in this case, even though location and orientation of the complex were correctly anticipated.

Multiple Isomorphous Replacement and Multiple Wavelength Anomalous Diffraction Structure determination by MIR or MAD requires substitution with heavy atoms. Derivatives can be generated by de novo crystal growth of modified DNA. Cytosine and uracil can be substituted with bromine or iodine at the C5 position.14 16 A search of the NDB indicates 30 bromine-substituted DNA fragments have been crystallized. Guanines are effective targets for soaking, for example, by accepting platinum at the N7 position. ~6 J5C. L. Kielkopf,K. E. Erkkila, B. E Hudson,J. K. Barton, and D. C. Rees,Nat. Struct.Biol. 7, 117 (20OO). 16L. A. Lipscomb,E X. Zhou,S. R. Presnell, R. J. Woo,M. E. Peek,R. R. Plaskon,and L. D. Williams, Biochemistry35, 2818 (1996).

290


[ 14]

Summary Here we have stressed important differences between protein and DNA crystallography. Crystal growth and data collection methodologies are not directly transferable between the two subfields. In addition, we note that analysis of symmetry and packing of DNA crystals can be useful and a uniquely aesthetic exercise. Acknowledgments This work was supported by the National Science Foundation (Grant M C B - 9 9 7 6 4 9 8 ) and the A m e r i c a n C a n c e r Society (Grant R P G - 9 5 - 1 1 6 - 0 3 - G M C ) .

[14] Molecular Modeling of Drug-DNA Complexes: An Update By J O H N

O. TRENT

Introduction This is an exciting time in the field of molecular modeling of DNA and DNA :ligand structures. There have been several advances in the field, such as second-generation force fields and the implementation of the particle mesh Ewald (PME) methods that more accurately reproduce experimental structures. When combined with the ever-increasing speed of computational resources (supercomputers and, in particular, parallel processor workstations), this places molecular simulations of biologically relevant systems at the fingertips of researchers. We now are able to investigate systems that previously were not practical, either because of less reliable force fields or computational expense. This is not to say that the discipline of molecular modeling has fully matured, as we are only at the second generation of force fields. However, with the increasing reliability, physical size of simulated DNA and proteins, and length of simulations, more researchers are turning toward simulations to aid them in the rationalization and prediction of experimental data, and in the design of specific experiments. The application of molecular modeling (and structural biology as a whole) to nucleic acids is, and will continue to be, increasingly important in future, particularly in light of the complete mapping of the human genome. The ability to target a particular gene or control region of genetic DNA in a sequence-specific fashion is one of the biggest challenges in drug design today. The strategy of inhibiting the transcription or translation of a specific gene has several advantages. These include low drug concentration needed (in theory, only a few molecules of a gene-targeting

METHODS IN ENZYMOLOGY,VOL. 340

Copyright© 2001 by AcademicPress All rights of reproductionin any form reserved. 0076-6879/00$35.00

290


[ 14]

Summary Here we have stressed important differences between protein and DNA crystallography. Crystal growth and data collection methodologies are not directly transferable between the two subfields. In addition, we note that analysis of symmetry and packing of DNA crystals can be useful and a uniquely aesthetic exercise. Acknowledgments This work was supported by the National Science Foundation (Grant M C B - 9 9 7 6 4 9 8 ) and the A m e r i c a n C a n c e r Society (Grant R P G - 9 5 - 1 1 6 - 0 3 - G M C ) .

[14] Molecular Modeling of Drug-DNA Complexes: An Update By J O H N

O. TRENT

Introduction This is an exciting time in the field of molecular modeling of DNA and DNA :ligand structures. There have been several advances in the field, such as second-generation force fields and the implementation of the particle mesh Ewald (PME) methods that more accurately reproduce experimental structures. When combined with the ever-increasing speed of computational resources (supercomputers and, in particular, parallel processor workstations), this places molecular simulations of biologically relevant systems at the fingertips of researchers. We now are able to investigate systems that previously were not practical, either because of less reliable force fields or computational expense. This is not to say that the discipline of molecular modeling has fully matured, as we are only at the second generation of force fields. However, with the increasing reliability, physical size of simulated DNA and proteins, and length of simulations, more researchers are turning toward simulations to aid them in the rationalization and prediction of experimental data, and in the design of specific experiments. The application of molecular modeling (and structural biology as a whole) to nucleic acids is, and will continue to be, increasingly important in future, particularly in light of the complete mapping of the human genome. The ability to target a particular gene or control region of genetic DNA in a sequence-specific fashion is one of the biggest challenges in drug design today. The strategy of inhibiting the transcription or translation of a specific gene has several advantages. These include low drug concentration needed (in theory, only a few molecules of a gene-targeting


Copyright© 2001 by AcademicPress All rights of reproductionin any form reserved. 0076-6879/00$35.00

[ 1 4]

MOLECULAR MODELING OF DRUG-DNA COMPLEXES

291

drug w o u l d be required per cell), the r e m o v a l of nonspecific side effects (observed for m o s t current c h e m o t h e r a p e u t i c drugs), and the ability to investigate the effect o f "turning o f f " one or m o r e genes. There are several D N A - b a s e d approaches that are currently under investigation to do this, such as the antigene l-5 (targeting transcription by triplex formation), antisense 4'6'7 (targeting translation), and aptamer s-~3 (posttranslational, targeting particular proteins) approaches. The antisense and antigene approaches use the natural base c o m p l e m e n t a r i t y of base pairs and base triplets, respectively, to target relatively long regions (typically 2 0 - 3 0 bases) o f R N A or D N A . In theory it is possible to use synthetic organic ligands as an alternative to targeting specific gene sequences, and several researchers ~4-19 are investigating this approach. This o f course relies on the ability to design sequencespecific ultratight D N A - b i n d i n g ligands. Our understanding o f the c o n s e q u e n c e s o f specific genetic manipulation is s o m e w h a t limited at this stage, but is likely to expand with the advent of new technologies such as G e n e Chip arrays 2° and proteomics.2

J K. R. Fox, Curr. Med. Chem. 7, 17 (2000). 2 K. M. Vasquez and J. H. Wilson, Trends Biochem. Sci. 23, 4 (1998). 3 H. G. Kim, J. F. Reddoch, C. Mayfield, S. Ebbinghaus, N. Vigneswaran, S. Thomas, D. E. Jones, and D. M. Miller, Bioehemistry 37, 2299 (1998). 4 p. C. vander Vliet and J. D, Gralla, eds., Biochim. Biophys. Acta Gene Struet. Express. 1489(1) (1999). This issue is dedicated to oligonucleotide therapies and has several excellent reviews. 5 p. p. Chan and P. M. Glazer, J. Mol. Med. 75, 267 (1997). 6 S. Agrawal and E. R. Kandimalla, Mol. Med. Today 6, 72 (2000). 7 S. T. Crooke, Methods Enzymol. 313, 3 (2000). s S. A. Robertson, K. Harada, A. D. Frankel, and D. E. Wemmer, Biochemistry 39, 946 (2000). 9 I, Smimov and R. H. Shafer, Biochemistry 39, 1462 (2000). I0 D. E. Huizenga and J. W. Szostak, Biochemistry 34, 656 (1995). 11 L. Gold, J. BioI.Chem. 270, 13581 (1995). 12 p. j. Bates, J. B. Kahlon, S. D. Thomas, J. O. Trent, and D. M. Miller, J. Biol. Chem. 274, 26369 (1999). 13 M. Koizumi and R. R. Breaker, Biochemistry 39, 8983 (2000). 14N. R. Wurtz and P. B. Dervan, Chem. Biol. 7, 153 (2000). ~5p. B. Dervan and R. W. Burli, Curr. Opin. Chem. Biol. 3, 688 (1999). 16C. A. Hawkins, R. P. d. Clairac, R. N. Dominey, E. E. Baird, S. White, P. B. Dervan, and D. E. Wemmer, J. Am. Chem. Soc. 122, 5235 (2000). 17 R. L. d. Clairac, B. H. Geierstanger, M. Mrksich, P. B. Dervan, and D. E. Wemmer, J. Am. Chem. Soc. 119, 7909 (1997). 18j. B. Chaires, E Leng, T. Przewloka, I. Fokt, Y. H. Ling, R. Perez-Soler, and W. Priebe, J. Med. Chem. 40, 261 (1997). 19B. Martin, A. Vaquero, W. Priebe, and J. Portugal, Nucleic Acids Res. 27, 3402 (1999). 20 U. Scherf, D. T. Ross, M. Waltham, L. H. Smith, J. K. Lee, L. Tanabe, K. W. Kohn, W. C. Reinhold, T. G. Myers, D. T. Andrews, D. A. Scudiero, M. B. Eisen, E. A. Sausville, Y. Pommier, D. Botstein, P. O. Brown, and J. N. Weinstein, Nat. Genet. 24, 236 (2000). Comment in Nat. Genet. 24, 208 (2000). 21 A. A. Alaiya, B. Franzen, G. Auer, and S. Linder, Electropholesis 21, 1210 (2000).

292


[ 14]

It is apparent for most biologically relevant systems that a multidisciplinary approach, including biochemistry, bioinformatics, biophysics, chemical synthesis, molecular biology, and structural biology [nuclear magnetic resonance (NMR), X-ray crystallography, molecular modeling], is needed to fully elucidate the system of interest. This review covers the advances since 1995 in DNA and DNA : ligand molecular modeling. There are several excellent reviews 22-25 that cover the development and theoretical aspects of molecular modeling, and this review briefly summarizes these advancements and focuses on the applications of these to biologically relevant systems. Molecular Modeling

Introduction Molecular modeling is a broad term that encompasses ab initio quantum mechanical calculations, semiempirical calculations, and empirical calculations (charge-dependent molecular mechanics force fields). These techniques can be used to study the three-dimensional structure, dynamics, and properties of a molecule of interest. The force field approach has been traditionally used for DNA calculations of any size (more than a few bases) because of the computational expense of the ab initio or semiempirical approaches. However, the increasing CPU power of workstations and the advent of hybrid methods such as quantum mechanics/molecular mechanics (QM/MM) 26'27 now make it possible to treat a region of interest with a high level of theory while considering the surrounding shell of the system at a less computationally intensive level. Molecular modeling, particularly molecular mechanics and dynamics, are highly complementary to macromolecular NMR and X-ray crystallography. A combined approach is desirable and, where the parent experimental structure is available, the simulation can be partially validated 28 by reproducing the structure (assuming the structure is correct). The system of interest can then be investigated with more confidence.

Advances in Molecular Modeling of DNA Issues of molecular motion and flexibility are expected to be a crucial component in developing a fundamental understanding of the relationship between DNA structure and function in molecular biophysics and genetics. Diverse

22 D. Beveridge and K. Campbell, Curl: Opin. Struct. Biol. 10, 182 (2000). 23 T. E. Cheatham and P. Kollman, Annu. Rev. Phys. Chem. 51,435 (2000). 24 T. E. Cheatham, B. R. Brooks, and P. A. Kollman, in "Current Protocols in Nucleic Acid Chemistry," p. 7.5.1. John Wiley & Sons, New York, 1999. 25 T. E. Cheatham and B. R. Brooks, Theor. Chem. Acc. 99, 279 (1998).

[14]

MOLECULAR MODELING OF D R U G - D N A COMPLEXES

293

biophysical methods can be applied, but no single experimental technique is capable of full elucidation of the dynamical structure of DNA in a solvent environment. Molecular dynamics simulation can, in principle, provide a complete theoretical description of DNA structure and motions, and is thus a valuable independent means of developing models and interpreting experimental data. Furthermore, molecular dynamics force fields now play a key supplementary role in the determination of structures by iterative refinement procedures in crystallography and NMR spectroscopy. Both X-ray crystallography and NMR rely on fitting experimental data [electron density, or nuclear Overhauser effect restraints (NOEs), respectively] to an empirical force field. It should be noted that if the experimental data has low-quality regions, then the structures are dependent on the inherent force field to a greater extent. It is becoming increasingly popular for NMR investigators to use "complete system" protocols with full periodic box solvation and the particle mesh Ewald (PME) methods to improve the structure refinement (see Duplex DNA : Ligand Complexes, below). This seems a logical approach as the NOEs are determined in the solution phase, and so the best approximation of the solution phase should be used for the structure refinement. Molecular dynamics simulations have proved to be an excellent adjunct to experimental techniques for the determination of protein structures. 29 However, it is only the advances in force fields and the accurate representation of long-range effects that make molecular dynamics extremely powerful for the study of DNA structures. The two most significant advances in modeling of DNA have been the development of second-generation force fields and the effective treatments of longrange electrostatic interactions. Three second-generation all-atom force fields 3°-32 specifically designed for simulations including explicit solvent were developed in 1995. Two, AMBER 31 and CHARMM, 32'33 have been updated and extensively applied to DNA simulations. Discussions of the relative merits of these force fields are available 22,23,28 and in some instances similar effects are being reported. In both cases the force fields are now based on ab initio calculations with derived parameters.

26 Q. Cui and M. Karplus, J. Phys. Chem. B 104, 3721 (2000). 27 B. Hernandez, E J. Luque, and M. Orozco, J. Comput. Aided Mol. Des. 14, 329 (2000). 28 W. E van Gunsteren and A. E. Mark, J. Chem. Phys. 108, 6109 (1998). 29 p. A. Kollman, Acc. Chem. Res. 29, 461 (1996). 3o M. Levitt, M. Hirshberg, R. Sharon, and V. Daggett, Comput. Phys. Commun. 91, 215 (1995). 3JW. D. Cornell, E Cieplak, C. I. Bayly, I. R. Gould, K. M. Merz, D. M. Ferguson, D. C. Spellmeyer, T. Fox, J. W. Caldwell, and R A. Kollman, J. Am. Chem. Soc. 117, 5179 (1995). http://www.amber.ucsf.edu, and references therein. 32 A. D. MacKerell, J. Wiorkiewicz-Kuczera, and M. Karplus, J. Am. Chem, Soc. 117, 11946 (1995). 33 B. R. Brooks, R. E. Bruccoleri, B. D. Olafson, D. J. States, S. Swaminathan, and M. Karplus, J. Cornp. Chem. 4, 187 (1983). http://yuri.harvard, edu/.

294


[ 14]

Use of the Ewald methods in the treatment of long-range electrostatic interactions has become popular with efficient implementations of the particleparticle mesh Ewald, 34'35 and the particle mesh Ewald 36-38 methods (see Sagui and Darden 39 for a review). Molecular dynamics simulations using the PME method show significant improvement in the dynamic stability of simulated DNA, and stable multiple nanosecond trajectories of explicitly solvated [using periodic boundary conditions (PBC)] DNA are now routine using PME methods. However, possible artifacts 4°,4j from this treatment may occur if the system is not neutrally charged or exhibits large effective charge separation, although a net correction term has been reported 42 to remove artifacts of energy for systems with a net charge. Also the "flying ice cube" 43 is a potential problem with periodic velocity rescaling, which can be alleviated by periodically removing the translational and rotational motion or reassigning the velocities. Finally, the treatment of low dielectric solvents has been investigated4°'41 with potential disturbing aspects. The popular choice for simulations includes a second-generation force field with a PME implementation and PBC, and these have been shown to reproduce experimental observations (see below). There has been a resurgence in the use of implicit solvation techniques, as these are computationally less intensive and greater simulation times are available. Originally implemented by Still et a1.,44 the generalized Born/solvent accessibility (GB/SA) method, with appropriate parameters, can reproduce hydration free energies, 45'46 and the development of GB parameters compatible with AMBER is included in AMBER 6.0. However, the implicit solvation methods do not include specific ion and water association interactions. DNA Duph, x Simulation

Some of the first solution-phase simulations47-51 showed that nanosecond simulations were stable and reproduced a B-form DNA, instead of the intermediate A 34R. W. Hockney and J. W. Eastwood, "'Computer Simulation Using Particles." McGraw-Hill, New York, 1981. 35 B. A. Luty, 1. G. Tironi,and W. F. van Gunsteren.,/. Chem. Phys. 103, 3014 (1995). 3f~T. A. Darden, D. M. York, and L. G. Pedersen,.1. Chem. Phys. 98, 10089(1993). 3vU. Essmann,L. Perera, M. L. Berkowitz,T. A. Darden, H. Lee, and L. G. Pedersen,.1. Chem. Phys. 103, 8577 (1995). :~sH. G. Petersen,J. Chem. Phys. 103, 3668 (1995). 3'~C. Sagui and T. A. Darden,Aimu. Rev. Biophys. Biomol. StI'I~('L 28, 155 (1999). 4OR H. Hunenbergerand J. A. McCammon,Biophys. Chem. 78, 69 (1999). 41 R H. Hunenbergerand J. A. McCammon,J. Chem. Phys. 110, 1856(1999). 42 S. Bogusz,T. E. Chealham,and B. Brooks,.I. Chem. Phys. 109, 7070 (1998). 4:~S. C. Harvey,R. K. Z. Tan,and T. E. Cheathmn,.I. Comput. Chem. 19, 726 (1998). 44 W. C. Still, A. Tempczyk,R. C. Hawley,and T. Hendrickson,.I. Am. Chem. Soc. 112, 6127 (1990). 45 B. Jayaram, D. Sprous, and D. L. Beveridge,.1. Phys. Chem. B 102, 9571 (1998). 46 B. N. Dominyand C. L. Brooks,J. Phys. Chem. B 103, 3765 (1999).

[14]


295

and B forms from previous simulations. This was extended 52'53 with the AMBER second generation force field by Comell et al. 31 on a larger time scale (>5 nsec) with varied Na + ion distributions, using PME methods. These simulations suggested that the force field reproduced B-type DNA extremely well. The fact that the simulation was closer to canonical B-DNA than the crystal structure does not make the simulation less accurate; rather, the crystal structure had contributing crystal packing effects. It is more appropriate to compare NMR solution structures with solvated simulations and in this case there was good agreement with the NMR structure 54 from Lane's group. However, few helical parameters were available for comparison. Improvements in the deficiencies such as unwinding of the DNA, average sugar pucker, and )~ angles in the Cornell force field have been made. 55 C o n f o r m a t i o n a l Transitions

The stable dynamics simulations enabled Cheatham et al. to examine conformational transitions, such as the A- to B-form transition. Under low-salt conditions, using the Cornell et al. 31 force field, the A- to B-form transition occurred in 500 psec. 56 It also was shown 57 that models of structures that experimentally stabilize in the A form, such as RNA : RNA and RNA : DNA duplexes, were stabilized in the A form by the Cornell et al. 31 force field. In addition, the phosphoramidatemodified DNA, which has a known preference for the A form, was observed to undergo a B- to A-form transition in the simulation. 58 Cheatham et al. 59 and Beveridge's group 6° have extended these simulations by modeling experimental conditions known to stabilize A-from DNA, such as high ethanol content, and found stabilization of the A-form DNA. Langley has developed the BMS force field for nucleic acids that shows the B- to A-form DNA transitions in 75% ethanol, and also reproduces this under high-salt conditions (4 M NaC1). Under ionic conditions 47 Y. E. Cheatham, J. L. Miller, T. Fox, T. A. Darden, and P. A. Kollman, J. Am. Chem. Soc. 117, 4193 (1995). 48 D. A. Zichi, ,l. Am. Chem. Soc. 117, 2957 (1995). 49 S. Weerasinghe, P. E. Smith, V. Mohan, Y.-K. Cheng, and B. M. Pettitt, J. Am. Chem. Soc. 117, 2147 (1995). 5o S. Weerasinghe, P. E. Smith, and B. M. Pettitt, Biochemistry 34, 16269 (1995). 51L. Yang, S. Weerasinghe, E E. Smith, and B. M. Pettitt, Biophys. J. 69, 1519 (1995). 52 M. A. Young, B. Jayaram, and D. L. Beveridge, .l. Am. Chem. Soc. 119, 59 (1997). 53 M. A. Young, G. Ravishanker, and D. L. Beveridge, Biophys. J. 73, 2313 (1997). 54 A. Lane, T. C. Jenkins, T. Brown, and S. Neidle, Biochemiso 3, 30, 1372 (1991). 55 T. E. Cheatham, E Cieplak, and E A. Kollman, J. Biomol. Struct. Dyn. 16, 845 (1999). 56 T. E. Cheatham and E A. Kollman, J. Mol. Biol. 259, 434 (1996). 57 Y. E. Cheatham and P. A. Kollman, J. Am. Chem. Soc. 119, 4805 (1997). 58 p. Cieplak, I. T. E. Cheatham, and E A. Kollman, J. Am. Chem. Soc. 119, 6722 (1997). 59 T. E. Cheatham, M. E Crowley, T. Fox, and E A. Kollman, Proc. Natl. Acad. Sci. U.S.A. 94, 9626 (1997). 6o D. Sprous, M. A. Young, and D. L. Beveridge, J. Phys. Chem. B 102, 4658 (1998).

296


[ 141

known to stabilize A-form DNA the Cornell e t a l . 31 force field also showed spontaneous transition of B-form to A-form DNA. 6j The A- to B-form transition also has been observed for triplex simulations 62 using A M B E R (PBC and PME conditions). The conformation of duplexes involving peptide nucleic acid (PNA) also has been investigated using A M B E R and C H A R M M . Molecular dynamics simulations of the PNA : DNA and PNA : RNA duplexes, using A M B E R (PBC and PME), by Soliva et al. 63 demonstrated the reproduction of PNA : DNA and PNA : RNA NMR structures. They found that the radius of convergence was wide, although it was noted that convergence can fail if the starting model is severely in error. Sen and Nilsson 64 used C H A R M M (PBC) to investigate the PNA : PNA duplex, PNA : DNA antiparallel duplex, PNA : DNA parallel duplex, and DNA : DNA duplex and reported excellent agreement for the PNA : PNA duplex with the crystallographic structure and good agreement of the PNA : DNA antiparallel duplex with the N M R structure. Overall this is encouraging, as the second-generation force fields are beginning to reproduce experimental structures with experimental conditions. Further improvements are likely to extend the experimental conditions that can be reproduced, for example, the B- to Z-form and Z- to B-form transitions, which are currently under investigation. 65"66 DNA Hydration and Ion Interactions

There has been great interest in the hydration and ion interactions of DNA stemming largely from the Williams 67 high-resolution (1.4-]k) crystal structure, which clearly shows the detailed spine of hydration of the minor groove extending past the first solvation shell. Several groups 52'53'57-59'61,68-75 have shown that molecular dynamics simulations can accurately reproduce several features of DNA hydration 61 T. E. Cheatham and E A. Kollman, Structure 5, 1297 (1997). 62 G. C. Shields, C. A. Laughton, and M. Orozco, J. Am. Chem. Soc. 119, 7463 (1997). 63 R. Soliva, E. Sherer, E J. Luque, C. A. Laughton, and M. Orozco, J. Am. Chem. Soc. 122, 5997 (2000). 64 S. Sen and L. Nilsson, ,l. Am. Chem. Soc. 120, 619 (1998). 65 j. O. Trent, unpublished data, Gridrun program (available from the author). 66 T. E. Cheatham, personal communication (2000). 67 L. D. Williams, Biochemistry 37, 8341 (1998). 68T. E. Cheatham, J. Srinivasan, D. A. Case, and R A. Kollman, J. Biomol. Struct. Dyn. 16, 265 (1998). 69 y. Duan, R Wilkosz, M. Crowley, and J. M. Rosenberg,J. Mol. Biol. 272, 553 (1997). 7o M. Feig and B. M. Pettitt, J. Mol. Biol. 286, 1075 (1999). 7t M. A. Young and D. L. Beveridge,J. Mol. Biol. 281, 675 (1998). 72 D. R. Langley, J. Biomol. Struct. Dyn. 3, 487 (1998). 73 M. Feig and B. M. Pettitt, J. Phys. Chem. B 101, 7361 (1997). 74 M. Feig and B. M. Pettitt, Biophys. J. 75, 134 (1998). 75 D. R. Langley,J. Biomol. Struct. Dyn. 16, 1366 (1999).

[14]


297

including the spine of hydration in A tracts, hydration in the major groove, the cones of hydration around phosphates, and the differences in the hydration patterns for A-form DNA and RNA. Specific ionic interactions also are being reproduced by simulations, 52'53'57"59' 61,71,72,75,76 such as the ion association in the minor groove of B-form DNA and in the major groove of A-form DNA. Moreover, molecular dynamics reproduced specific effects on the stabilization of the structure that are consistent with experimental data, such as enhanced bending of the A tracts, 71'77the B- to A-form transitions, 61'72 and differences in the behavior of specific ions. 76'78 Again, the ability to reproduce structures consistent with experimental conditions is highly encouraging. Poisson-Boltzmann calculations have been used to look at multivalent ions and the B- to Z-form transitions. 79 A review by Pack et aL so details Monte Carlo and Poisson-Boltzmann approaches to examine divalent cations and the electrostatic potential surrounding DNA. Modeling of DNA : Ligand Complexes Several examples of modeling DNA:ligand complexes and multistranded DNA have been chosen, with particular emphasis on the utilization of the advances mentioned above. Duplex DNA : Ligand Complexes. Several groups have been applying explicit water and PME methods to aid in the refinement of NMR data. Williams and Searle 81 have studied the structure, dynamics, and hydration of nogalamycin with d(ATGCAT)2 by using NMR and molecular dynamics and show that many sequence-dependent structural features observed in X-ray crystal structures lie within the range of dynamic fluctuations observed in the molecular dynamics simulations. The Searle and Laughton groups also have examined s2 the DNA minor groove recognition by a tetrahydropyrimidinium analog of Hoechst 33258 by NMR and molecular dynamics studies of the complex with d(GGTAATTACC)2, and the molecular recognition and dynamic equilibrium between a pentacyclic acridinium salt and DNA sequences in a combined approach of optical spectroscopic techniques, NMR, and molecular dynamics 83 (AMBER, PBC, and PME). The ability to calculate thermodynamic parameters for DNA : ligand complexes allows direct comparison with experiments. The absolute free energy of association 76 A. D. MacKerell, J. Phys. Chem. B 101,646 (1997). 77 C. E. Bostock-Smith, C. A. Laughton, and M. S. Searle, Biochem. J. 342, 125 (1999). 78 A. P. Lyubartsev and A. Laaksonen, J. Biomol. Struct. Dyn. 16, 579 (1998). 79 M. Gueron, J. Demaret, and M. Filoche, Biophys. J. 78, 1070 (2000). 80 G. R. Pack, L. Wong, and G. Lamm, Biopolymelw 49, 575 (1999). 81 H. E. Williams and M. S. Searle, J. Mol. Biol. 290, 699 (1999). 82 C. E. Bostock-Smith, C. A. Laughton, and M. S. Searle, Nucleic Acids Res. 26, 1660 (1998). 83 C. E. Bostock-Smith, E. Gimenez-Arnau, S. Missailidis, C. A. Laughton, M. E Stevens, and M. S. Searle, Biochemisoy 38, 6723 (1999).

298


[ 14]

of netropsin with the Dickerson dodecamer has been determined with the AMBER suite of programs by Singh and Kollman 84 and a fully detailed atomic model for both solute and solvent. The calculated free energy of - 10.3 kcal/mol is in good agreement with the experimentally determined value of - 1 1.5 kcal/mol, and the number of water molecules found to diffuse into the uncomplexed DNA was equal to the number of water molecules in the binding site of the uncomplexed crystal structure. The free energy difference between thymine and difluorotoluene in DNA duplex has been determined by molecular dynamics and thermodynamic integration and was found 85 to be in good agreement with experimental values. This was extended to make predictions about the effect of the difluorotoluene substitution in triplex DNA. Triplex DNA Complexes. The simulation of triplex DNA is well established with reports from Laughton and co-workers62'86 and Pettitt,49'5°'87 among others. It has been shown 62 for the TAT parallel triplex structure, by using molecular dynamics (AMBER with PME), that the A-, B-, and P-form (PNA-like) starting models converge to a single B-form structure that is consistent with experimental evidence. We have observed similar convergence effects for the RNA third strand: DNA triplexes (see Triplex Simulation of Antigene Approach, below). Molecular dynamics simulations with AMBER have been used 88 to investigate the PNA. DNA : PNA system and convergence of starting models also has been demonstrated. It is well known that certain ligands have a preference for triplex DNA over other forms of DNA, and a new dialysis assay (see [5] in this volume TM) enables the direct comparison of binding preferences of ligands for different forms of duplex, triplex, and quadruplex DNA. Qualitative modeling 89 of intercalating mono- and disubstituted amidoanthraquinone regioisomers has rationalized the relative affinities determined by DNase I footprinting. Mithramycin and distamycin also have been investigated for triplex stability by gel mobility shifts, nuclease and chemical hypersensitivity, two-dimensional gel topological analyses, triplex-specific antibody-binding studies, and qualitative modeling. 9° These minor groove binders have preferences for GC (A-type DNA) and AT (B-type DNA), respectively. Modeling reproduced the experimental observation that mithramycin 84 S. B. Singh and E A. Kollman, J. Am. Chem. Soc. 121, 3267 (1999). 85 E. Cubero, C. A. Laughton, E J. Luque, and M. Orozco, J. Am. Chem. Sot., in press (2000). 86 R. Soliva, C. A. Laughton, F. J. Luque, and M. Orozco, J. Am~ Chem. Soc. 120, 11226 (1998). 87 Y.-K. Cheng and B. M. Pettitt, Biopolymers 35, 457 (1995). 88 G. C. Shields, C. A. Laughton, and M. Orozco, J. Am. Chem. Soc. 120, 5895 (1998). 88a j. Ren and J. B. Chaires, Methods Enzymol. 340, [5] 2001 (this volume). 89 M. D. Keppler, M. A. Read, E J. Perry, J. O. Trent, T. C. Jenkins, A. E Reszka, S. Neidle, and K. R. Fox, Era: J. Biochem. 263, 817 (1999). 9o N. Vigneswaran, J, Thayaparan, J. Knops, J. O. Trent, D. M. Miller, and W. Zacharias, Biol. Chem., in press (2001).

[14]

MOLECULAR MODELINGOF DRUG-DNA COMPLEXES

299

destabilizes triplex and also inhibits triplex formation, whereas distamycin stabilizes triplex and facilitates triplex formation. Q u a d r u p l e x D N A Complexes. Molecular dynamics simulations have reproduced the high-resolution crystallographic structure and NMR structures of several quadruplex species. The parallel d(G4)4 quadruplex simulation 91 including the central core Na + ions reproduces the crystal structure extremely well [ < 1-]k root mean square deviation (RMSd)], and it was observed that the Na + ions are important for stability. The antiparallel quadruplex d(G4T4G4)2 simulation 91 also compares will with the NMR structure. A comparison 92 of the in vacuo and fully solvated (AMBER, PME) NMR refinement of the antiparallel d(G3T4G3)2 quadruplex indicated that explicit solvent treatment led to better R factors. An NMR investigation 93 into the determinants of DNA quadruplex structures and potassium binding, using C H A R M M (implicit solvation), showed that potassium is needed for the chair-type but not for the basket-type quadruplex. Ligands that bind to quadruplexes have been of interest 94 as potential telomerase inhibitors. The amidoanthraquinone series has been reported 95 to have inhibitory activity and qualitative modeling shows that "end pasting" is one possible binding mode for the human telomere quadruplex. More detailed simulations 96 using explicit water and the consistent force field (CFF) supported this, although other binding sites were not examined. The porphyrin class of telomerase inhibitors has been examined 97 by isothermal titration calorimetry (ITC), spectrophotometry, and fully solvated molecular dynamics (AMBER and PME), and this clearly showed intercalation in a number of quadruplexes. Protocols for Molecular Modeling of DNA: Ligand Complexes When contemplating a molecular modeling simulation the first priority is to determine the fundamental question to be answered and to decide which (if any) method is the most appropriate to achieve this. There is little point in performing a simulation that cannot, by the experimental design and limitations of the algorithms, provide the answer or insight into the fundamental question. 91 N. Spackova,I. Berger, and J. Sponer,J. Am. Chem. Soc. 121, 5519 (1999). 92G. D. Strahan, M. A. Keniry, and R. H. Shafter, Biophys. J. 75, 868 (1998). 93W.M. Marathias and P. H. Bolton,Biochemistry 38, 4355 (1999). 94D. Sun, B. Thompson, B. E. Cathers, M. Salazar, S. M. Kerwin, J. O. Trent. T. C. Jenkins, and S. Neidle, J. Med. Chem. 40, 2113 (1997). 95p. j. Perry, A. P. Reszka, A. A. Wood,M. A. Read, S. M. Gowan, H. S. Dosanjh, J. O. Trent, T. C. Jenkins, L. R. Kelland, and S. Neidle,J. Med. Chem. 41, 4873 (1998). 96M. A. Read, A. A. Wood, J. R. Harrison, S. M. Gowan, L. R. Kelland, H. S. Dosanjh, and S. Neidle, J. Med. Chem. 42, 4538 (1999). 97I. Haq, J. O. Trent, B. Z. Chowdhry,and T. C. Jenkins, J. Am. Chem. Soc. 121, 1768 (1999).

300


[ 14]

M o l e c u l a r D y n a m i c s Simulations

It should be emphasized that no two systems are alike and that most protocols must be adapted to every situation. Simulations should be constantly monitored to check progress. Five components are required to performing molecular dynamics simulations, namely, model building, charge generation, mechanics and dynamics, visualization, and analysis. Various software packages are available that perform one or more of these functions. The increasing abundance of more user-friendly modeling programs is enabling a greater number of researchers to utilize molecular modeling. However, extreme care should be taken and a "black box" approach should never be used. Investigators should understand in detail the "default" parameters and should tailor the simulation to meet their needs in accordance with the literature on simulations. The software used in this chapter is Macromodel, 9s GAMESS, 99 the AMBER suite, 3j GRASP, 1°° Insight II, and Curves. I°1 There are three basic steps in the overall simulation: the generation of starting models, the actual simulation, and the evaluation of the production run. Generation o f Starting Models D N A Duplex Generation. Generation of the DNA starting model (or models) can be done by a variety of software programs, such as NUCGEN, 3j NAB, 1°2 and Macromodel, 98 which can generate different forms (A-form, B-form, Z-form) of duplex DNA. The Protein Database 1°3 (PDB) and the Nucleic Acid Database TM (NDB) are invaluable resources for obtaining experimentally determined structures. Care should be taken with the addition of hydrogen atoms to experimentally derived crystal and software-generated DNA structures, as some software can place them in nonoptimal positions. Also, atom names and the format of the PDB file often need to be changed when generating structures in one program and using another for the simulation because, although there is a standard PDB format, historical variations are still common. The file conversion program Babel 1°5 is

98E Mohamadi, N. G. Richards, W. C. Guida, R. Liskamp, M. Lipton, C. Caufield, G, Chang, T. Hendricksen,and W. C. Still,J. Comp. Chem. 11,440 (1990). http://www.schrodinger.com. 99M. W. Schmidt, K. K. Baldridge,J. A. Boatz, S. T. Elbert, M. S. Gordon,J. J. Jensen, S. Koseki, N. Matsunaga,K. A. Nguyen,S. Su, T. L. Windus,M. Dupuis,and J. A. Montgomery,.I. Comp. Chem. 14, 1347 (1993). http://www.msg.ameslab.gov/gamess/gamess.html. lo0 A. Nicholls, K. Sharp, and B. Honig,Proteins Struct. Funct. Genet. 11, 281 (1991). http://trantor. bioc.columbia.edu/grasp/. 101R. Lavery and H. Sklenar, J. Biomol. Struct. Dyn. 6, 63 (1998). http://www.ibpc.fr/ UPR9080/curindex.html. 1olT. Macke and D. A. Case, in "Molecular Modelingof Nucleic Acids" (N. B. Leontes and J. S. Lucia, eds.), p. 379. AmericanChemical Society, Washington,D.C., 1998. 1o3H. M. Berman,J. Westbrook,Z. Feng,G. Gilliland,T. N. Bhat, H. Weissig,I. N. Shindyalov,and R E. Bourne,Nucleic Acids Res. 28, 235 (2000). 1o4H. M. Berman,W. K. Olson,D. L. Beveridge,J. Westbrook, A. Gelbin,T. Demeny,S. H. Hsieh, A. R. Srinivasan,and B. Schneider,Biophys. J. 63, 751 (1992).

[ 14]


301

useful in the manipulation of different file formats, although most modeling software will now export PDB files. A program that converts nonstandard PDB files to AMBER-style PDB files with atom renaming (based only on identifying the element of each individual atom) is available from the author. Generation of Triplex DNA. There are a number of NMR triplex structures in the PDB and NDB that can be modified for the generation of the particular triplex sequence and motif (parallel or antiparallel) of interest. The most notable X-ray crystal structure is that of Rhee et al., 1°6 which has an overlap region of three consecutive base triplets consisting of C+-GC, BU-ABU (BU, 5-bromouracil), and C+-GC. Alternatively, the fiber diffraction structures of parallel and antiparallel triplexes are available 1°7 109 and a simple algorithm can be used to generate the triplex of interest. However, care should be taken as the triplex structure generated by this method has enforced symmetry. Additional equilibrium periods should be utilized to relax the symmetry prior to using these structures as starting models. Generation of Quadruplex DNA. The generation of parallel and antiparallel quadruplex is based on crystallographic and NMR structures, particularly the high-resolution (0.75-~) parallel d(TGGGGT)4 structure of Phillips et al.ll° and the antiparallel human telomere d[AG3(T2AG3)3] NMR structure of Wang and Patel.I I1 There are a number of different motifs available in the PDB and NDB and, as the structures of interest are usually short (2-4 base quartets), these structures can easily be modified for use. DNA Ligand Generation. Generation of the starting model for each DNA ligand is variable. Many different programs can be used to generate the initial threedimensional model of the ligand. The three-dimensional coordinates may be available from the Cambridge Structure Database, j J2 or the structure of the DNA : ligand complex may already have been reported. More commonly, when dealing with a new ligand, it will have to be optimized and parameterized in some way. The geometry can be optimized at a level suited to the overall calculation. Many of the van der Waals, bond, angle, torsion, and improper torsion parameters can be derived by analogy. A particular force field may already be parameterized for the ligand, such Jo5 p. Waiters and M. Stahl, ftp://ccl.osc.edu/pub/chemistry/software/UNIXfoabel/. 106 S. Rhee, Z. Han, K. Liu, H. T. Miles, and D. R. Davies, Biochemistry 38, 16810 (1999). 1o7 G. Raghunathan, H. T. Miles, and V. Sasisekharan, Biochemistry 32, 455 (1993). 108 G. Raghunathan, H. T. Miles, and V. Sasisekharan, Biopolymers 36, 333 (1995). 109 E B. Howard, H. T. Miles, K. Liu, J. Frazier, G. Raghunathan, and V. Sasisekharan, Biochemistry 31, 10671 (1992). lm K. Phillips, M. Orozco, E Q. Luque, S. J. Teat, W. Clegg, and B. Luisi, to be published (PDB code 1DJ4). IlL y. Wang and D. J. Patel, Structure 1,263 (1993). 112 E H. Allen, S. Bellard, M. D. Brice, B. A. Cartright, A. Doubleday, H. Higgs, T. Hummelink, B. G. Hummerlink-Peters, O. Kennard, W. D. S. Motherwell, J. R. Rodgers, and D. G. Watson, Acta Crystallogr B 35, 2331 (1979),

302


[ 14]

as MM2, and can be used to provide a reasonable starting model. If conformational flexibility is present, a grid search by torsion driving (one to three torsion angles) or Monte Carlo searches (greater than three torsion angles) can be used to fully explore the conformational space of the ligand. Other conformational techniques can be employed such as pure methods JBW, ADF, TMC, and SD or mixed methods jl3 ADF/JBW, SD/TMC (original MCSD method), SD/TMC, SD/JBW, and SD/ADF/JBW. It is important to get close to the "global" minima as the ligand may undergo a conformational change on binding to DNA and this energy penalty is included in the overall energy of the complex. The starting models can be used for ab initio optimization and analogous parameters may be incorporated into the force field, although great care must be taken as the format of force fields may be different in the terms they use. The worst case scenario is that full parameterization using ab initio calculations is required, although this is faster than in the past. The level of effort that should be expended is related to the scientific question being asked. (For a tutorial on parameter development see Kollman's tutorial, http://www.amber.ucsf.edu/amber/newparams.html. The calculation of accurate partial charges uses an ab initio program, such as Gaussian 98 or GAMESS 99 (this can be done with parallel processors). The accurate representation of electrostatic interactions is crucial for a force field intended for application to biological molecules, particularly as DNA is a highly charged polyanionic molecule. The charges also should be consistent with the force field. When using the A M B E R force field at least the 6-31G* basis set should be used, and restricted electrostatic potential (RESP) j 14,~15 charges are advisable. DNA : Ligand Complex Generation. The starting position(s) of the ligand in the D N A : l i g a n d complex is a crucial element. Ideally a similar structure has been solved and coordinates can be superimposed to produce a reasonable starting model. However, obviously not all systems of interest have been solved, and although force fields have improved and can find stable and experimentally derived DNA conformations from starting models that are reasonably distant from the final structure, the DNA : ligand starting complex must be chosen with care. There are two pitfalls here; the first is that the conformations become locked into a local minima so that it cannot move over local energy barriers, and second, user bias is incorporated so that the final structure derived is that which was considered most likely by the investigator. Ideally the ligand should be placed near the DNA and allowed to "find" the most stable bound complex. However, even if this were possible, it would require simulations that are on a much greater time scale than are currently accessible. There are a number of ways to overcome this situation. l J3 SD, stochastic dynamics; ADF, all degrees of freedom Metropolis Monte Carlo; TMC, torsional Monte Carlo; JBW,jumping between wells importance-samplingMonte Carlo. 114C. I. Bayly, P. Cieplak, W. D. Comell, and P. A. Mollman,J. Phys. Chem. 97, 10269 (1993). 115W. D. Cornell, E Cieplak, C. I. Bayly, and E A. Kollman,J. Am. Chem. Soc. 115, 9620 (1993).

[14]


303

Simulated annealing, Monte Carlo searches, and docking calculations can be used to obtain starting structures. Often the ligand is in a known class of DNA binder, such as an intercalator or a groove binder, which reduces the uncertainty of binding location. Also, additional experimental evidence, such as DNase I footprinting and sequence preference techniques such as REPSA, 116 can provide information that is useful in generation of the complex. Generation o f DNA : lntercalator Starting Models. The intercalation site can usually be obtained and modified from similar structures from the PDB or NDB for duplex DNA. A known intercalator step can be incorporated into the sequence of interest. Alternatively, an intercalation step can be built by placing the intercalation site bases at a similar rise and helical twist as known intercalation sites. The phosphodiester (or other) backbone can be inserted and optimized while constraining the rest of the DNA. The optimization should follow a protocol similar to that used for the overall simulation and include minimization and dynamics combined with solvent treatment to arrive at a reasonable structure. Constraints can be relaxed for portions of the DNA; however, the rise should not be, as the intercalation site will collapse without the ligand present. We have developed a simple approach (Gridrun 65) to generate starting structures for new intercalator complexes, as we have found that standard docking programs tend not to reproduce the intercalated ligand location. The only way to sample all possible binding sites is to do a grid search of all possible positions of the ligand in the intercalation site. The grid search in Gridrun translates the ligand in an 8 × 8 ~ (any size can be used) grid on the xz plane (with the helical axis of the DNA aligned to the y axis). In combination with this, full rotation in 15 ° increments (any increment can be used) rotates the ligand around the y axis. This typically generates 2025 structures, although this can be reduced, as the translation and rotation increments can be modified to a user-specified range if additional information is known and combined with chemical or structural intuition. Molecular mechanics minimization using the implicit GB/SA solvation approximation and AMBER94 force field within Macromodel 7.098 (modified to include AMBER charges from the input file) is run on all structures (this can be done with parallel processors) and an energy window of 50 kcal is used to keep possible starting structures. Incorporation of a heavy atom position comparison during the minimization reduces the actual number of models fully minimized, thus decreasing the actual computation time. The region surrounding the intercalation site can be excluded from the calculation by using a frozen shell that further reduces the computational expense. Additional conformational freedom can be included in this approach, such as torsion driving of any substituents on the planer aromatic intercalation component. Monte Carlo searches also can be included for a large

116E Hardenbol,J. C. Wang,and M. W. Van Dyke,Bioconiug. Chem. 8, 617 (1997).

304


[ 14]

number of torsion angles of the substituent, although this rapidly becomes computationally expensive. The grid search can be enhanced by molecular dynamics for relaxation of the intercalation complex if a reduced number of starting models is used (for computation expense). In fact, any conformational space search techniques available in Macromode198 can be used. The results of this type of search depend on the quality of the intercalation site. There are a number of other, more sophisticated approaches available for this, but the author has found that this is relatively efficient and removes any doubt about sampling conformational space. Obviously, at this time an explicit solvation method would be time consuming and computationally expensive for a grid search. A selection of the lowest energy models is then used for explicit solvation simulations, using AMBER. 3t As with all starting models, care should be taken in the selection, as this is an approximate method and the user does not want to exclude reasonable models. We have found that explicit solvation is an essential factor as water is an integral structural component of DNA. This method has reproduced the daunomycin intercalation position and has been applied successfully to Z-DNA intercalation, triplex intercalation, and quadruplex intercalation. Generation of Minor Groove-Binding Ligand Starting Models. In general, groove-binding ligand starting models can be generated on the basis of known structures as they typically do not greatly distort DNA and are AT specific, although some prefer GC regions. Some classes of ligands cause a conformation change in the DNA, such as mithramycin-induced B- to A-form transition. A method for the generation of minor groove-binding ligand starting models, similar to that used for the generation of intercalation site starting models, has been implemented by the author. To find the optimal binding site for the ligand, a long sequence containing multiple binding specificities is used and the ligand is simply translated and rotated around the helical axis in the minor groove, and evaluated by molecular mechanics minimization using the implicit GB/SA solvation approximation and AMBER94 force field within Macromodel. 98 This method has reproduced the binding sites of berenil, pentamidine, distamycin, netropsin, and Hoechst 33258, and has been successfully applied to both duplex and triplex minor groove-binding ligands. 9°

Molecular Dynamics Simulation Molecular dynamics simulation has two components: the equilibrium phase and the production phase. There are many different opinions about how to carry out the equilibrium and production phases, but only our chosen method is presented here. In fully solvated systems the equilibrium period is essential for the reorganization and relaxation of the solvent and counterions before the production phase to reduce potentially destabilizing van der Waals and electrostatic interactions. In some cases, if the starting model is not optimal, the DNA : ligand complex should be equilibrated in vacuo prior to addition of solvent, to remove any potentially

[14]


305

bad contacts. Equilibrating the solvent with the DNA:ligand complex initially restrained reduces artifacts from the interactions of the nonoptimized solvent with the complex. The gradual removal of restraints allows the system to slowly interact fully. The equilibrium period is dependent on each simulation and is typically on the order of 30-100 psec. However, some systems may require longer equilibrium periods; the section Triplex Simulation of Antigene Approach (below) shows an equilibrium period for triplex DNA that required 155 psec, and ionic interactions may require even much longer periods. The systems should be monitored by individual terms of the force field such as bond, angle, dihedral, van der Waals, and electrostatics for anomalies, as well as density, pressure, potential energy, and temperature to establish whether equilibrium has been reached. The production phase is run for the time period required, typically 5001000 psec, and can be done in a variety of ways. This protocol uses the Sander program of AMBER 31 at constant pressure with the PME summation and a time step of 1.5 fsec. The time step is usually varied between 1 and 2 fsec. A shorter time step may be required if coordinate resetting is too large and causes the simulation to stop. Typically the bond lengths or only the hydrogen atom bond lengths are fixed, using the SHAKE method. The equilibrium and production phases should be constantly monitored and not just checked at completion, as every system is different and a number of things can happen that require intervention. Often, molecular simulations are iterative and the process may require modification from the generation of starting models to the production phase.

Evaluation of Simulation Evaluation of the molecular dynamics trajectory will depend on what fundamental question is being asked. In simple cases, if a stable structure for comparison is required, simple time averaging over snapshots near the end of the production run may be sufficient. To compare two structures use of the root mean square deviations (RMSd) of coordinates is common. It should be noted that the average structure is precisely that, an average, and is analogous to crystal structures and NMR average structures. Detailed dynamics may be concealed in average structures. If the dynamics of the system are of interest, monitoring a particular region or specific interactions over the whole production phase is informative, and a variety of parameters may be calculated, such as order parameters from correlation functions, to compare with experimentally derived NMR parameters. There is a wealth of information in the trajectory regarding DNA-ligand interactions and dynamics (nanosecond and subnanosecond range) of the DNA, ligand, DNA : ligand complex, solvent, and counterion interactions with the DNA and ligand. Any three-dimensional quantity can be measured such as hydrogen bonding, van der Waals interactions, and solvent density. There are a number of parameters specific for DNA structure that can be used for comparison. These can be divided into five groups: the global property (helical twist); the backbone torsion angles

306


[ 14]

or, r , y, 6, e; and (; the furanose sugar conformation; the glycosidic torsion angle X; and the base helical parameters. The backbone torsion angles are indicative of the form of DNA. The pseudorotation of the furanose ring defines the sugar pucker amplitude and the pseudorotation phase angles also are indicative of the form of DNA, and a population of conformations is not uncommon. The glycosidic angle )~ defines the orientation of the sugar to the base. The helical parameters are defined for the axis (x displacement, y displacement, and inclination), base pair (slide, shift, rise, twist, roll, and tilt), and bases (stagger, stretch, shear, opening, buckle, and propeller twist). The two common programs used for helicoidal parameters are Curves l°l and NEWHELIX, 117 although more recent programs are more mathematically consistent I Js (see Lu et al)19 for a review of nucleic acid analysis programs). Care should be taken when comparing helicoidal parameters that are calculated by different programs, as different definitions of the reference frame may be used. Typically three simulations are needed: the DNA : ligand complex, DNA only, and the ligand only. The energy of binding can be calculated in two ways. If the same number of water molecules was added (manual manipulation of the topology file) to different conformations of the DNA : ligand complexes, a direct comparison of energies is possible. Or more commonly, the systems are stripped of water and minimized, and the simple equation of Ebinding = Ecomplex--EDNA -Eligand is used, although this does not directly take into consideration the hydration energetic component. Average structures distort the water positions as they are moving rapidly compared with the complex. Movies of the trajectories can be made using Moil-view 12° or VMD TM and some imaging may be necessary (see http://www.amber.ucsf.edu/amber/tutorial/ polyA-polyT/analysis.html for guidelines on this). Hydration may be investigated by looking at the radial distribution of water around a region of interest, using Carnal. 31 Also, the water density can be calculated by Watden 122 in order to identify regions with high-density or long-lived water molecules. General Molecular Simulation Protocol

GENERATION OF STARTING MODELS 1. Creation of a three-dimensional model of the ligand: A b initio partial charge calculation (at least at the 6-31 G* basis set level) of the ligand with restricted 117R. E. Dickerson, Nucleic Acids Res. 17, 1797 (1989). 118M. S. Babcock,E. E Pednault, and W. K. Olson,J. Mol. Biol. 237, 125 (1994). 1L9X. J. Lu, M, S. Babcock,and W. K. Olson, J. Biomol. Struct. Dyn. 16, 833 (1999). 120C. Simmerling,R. Elber, and J. Zhang, in "Modellingof BiomolecularStructure and Mechanisms" (A. Pullman, ed.), p. 241. Kluwer, Amsterdam,The Netherlands, 1995. 121W. Humphrey,A. Dalke, and K. Schulten,J. Mol. Graphics 14, 33 (1996), 122C. A. Laughton,Watden,a programto calculate waterdensities from AMBERmoleculardynamics trajectories.

[14]


307

electrostatic potential charge fit (RESP) and parameterization of the force field (if required) 2. Generation of starting models of DNA complex (Insight II, Macromodel,98 PDB, 103 NDB, 104 NUCGEN ,31 NAB 102) 3. Minimization of starting model in v a c u o with GB/SA solvation 4. Counterion addition: This can be done in at least three ways: (1) counterions added at the most electronegative positions, calculated by solving the PoissonBoltzmann equation (DELPHI123); (2) counterions added as per the rules in the Edit or LEAP program of AMBER; or (3) counterions randomly placed inside the periodic box 5. Periodic box of TIP3P water molecules added, typically 10-A box around the complex (~3500 water molecules) MOLECULARSIMULATION:EQUILIBRIUMPHASE 6. Initial equilibrium of the solvent: a. Minimization, holding DNA complex fixed [100 kcal (mol ./~)-1 restraints, 10,000-step steepest descents] b. Molecular dynamics at 100 K, holding DNA complex fixed [100 kcal (mol- A)-1 restraints, 1.5-fsec time step, 25-psec dynamics] c. Minimization, holding DNA complex fixed [100 kcal (mol./~)-1 restraints, 10,000-step steepest descents] d. Minimization of total system (10,000-step steepest descents) 7. Equilibrium of system: a. Molecular dynamics at 100 K, holding DNA complex fixed [100 kcal (mol. ~)-J restraints, 1.5-fsec time step, 25-psec dynamics] b. Molecular dynamics at 300 K, holding DNA complex fixed [100 kcal (mol. ]k)-I restraints, 1.5-fsec time step, 25-psec dynamics] c. Molecular dynamics at 300 K, gradually reducing the restraints on DNA complex over 100 psec [starting at 100 kcal (mol. ~)-1 restraints going to 0 kcal (mol. A) -I, 1.5-fsec time step, 80-psec dynamics) MOLECULARSIMULATION:PRODUCTIONPHASE 8. Production run: Unrestrained molecular dynamics at 300 K for 1 nsec, using the AMBER98 force field in the isothermal isobaric ensemble (P = 1 atm, T = 300 K), using periodic boundary conditions and the PME algorithm. A 1.5-fsec time step was used with all bond distances frozen using SHAKE for the 1-nsec production run 123B. Honigand A. Nicholls,Science 268, 1144(1995). http://trantor.bioc.columbia.edu/delphi/.

308


[ 141

EVALUATION 9. Sampling of production run and analysis: Time-averaged structures (50 snapshots sampled in the last 50 psec of the production trajectory) with further minimization (10,000 steps steepest descents), helical parameter structural analysis using Curves lm (helical parameters), trajectory analysis for energetics, hydration, and dynamic parameters 10. Calculation of binding energies from the DNA, ligand, and DNA : ligand simulations: Direct comparison of two or more conformations of the ligand in the DNA : ligand complexes E x a m p l e s of M o l e c u l a r M o d e l i n g of Biologically R e l e v a n t D u p l e x , Triplex, and Quadruplex Systems Four examples of modeling biologically relevant DNA : ligand complexes follow, representing duplex, triplex, and quadruplex DNA. The duplex example shows a new DNA ligand that allosterically converts B-form DNA to Z-form DNA, which could lead to a new class of DNA ligands with a novel mechanism. The first triplex example shows the structure of the RNA and 2'-modified RNA : DNA triplex system (and an A-form to B-form conformational transition) and is involved in the structure-based drug design of antigene agents. The second triplex example deals with triplex stabilization and shows how stereochemically and energetically reasonable models can be generated when structural data are unavailable. The quadruplex example examines the stoichiometry of ligands binding to quadruplex systems, as the binding site of these ligands has not been fully elucidated, and is part of a structure-based drug design effort for the development of telomerase inihibitors. Duplex Intercalation in Z-DNA

The binding interactions of (-)-daunorubicin (WP900) (Fig. 1) a newly synthesized enantiomer of the anticancer drug (+)-daunorubicin, with right- and lefthanded DNA have been studied TM quantitatively by equilibrium dialysis, fluorescence spectroscopy, and circular dichroism. (+)-Daunorubicin binds selectively to right-handed DNA, whereas the enantiomeric WP900 ligand binds selectively to left-handed DNA. Further, binding of the enantiomeric pair to DNA is clearly chirally selective and each of the enantiomers was found to act as an aliosteric effector of DNA conformation.124 Molecular dynamics studies using the AMBER suite of programs 3~ resulted in a stereochemically feasible model for the intercalation of WP900 into left-handed Z-DNA. 124X. Qu, J. O. Trent,I. Fokt, W. Priebe, and J. B. Chaires, Proc. Natl. Acad. Sci. U.S.A. 9, 12032 (2000).

[14]


A

309

B

FIG. 1. Structures of (+)-daunorubicin (A) and (-)-daunorubicin, WP900 (B).

The strong preference of (+)-daunorubicin for right-handed, B-form DNA is thought to arise from the precise fit of the daunosamine moiety into the minor groove. 125 High-resolution crystal structures of daunorubicin complexed to DNA shows that the angle of the daunosamine relative to the intercalated chromophore is such that it aligns with the minor groove, allowing a favorable stereochemical fit. 126 Because there are no crystal structures of WP900 bound to Z-DNA, computer simulations using molecular dynamics were applied to build a reasonable model of the complex. Figure 2 shows WP900 intercalated into left-handed DNA. The model is energetically and stereochemically feasible, and provides a detailed view of a left-handed DNA intercalation complex. Simulations of all possible combinations of daunorubicin enantiomers and DNA forms yielded only two stable complexes, one with daunorubicin bound to B-form DNA and another with WP900 bound to Z-form DNA. The B-form DNA : WP900 and Z-form DNA : daunorubicin complexes were not stable and resulted in disruption of duplex stability and loss of intercalation. Such ligand-induced disruptions of the disfavored B- and Z-form duplexes with WP900 and daunorubicin, respectively, may contribute to the allosteric effect by facilitating the backbone conformation change. The simulated B-form DNA: daunorubicin complex (not shown) resembles the published crystal structure.126 The Z-form DNA : WP900 complex (Fig. 2) shows that the AMBER95 force field is able to reproduce the left-handed DNA, as the structure is conserved under the salt conditions used, and the intercalation site is stable. The models reproduce the experimental observations the WP900 preferentially binds to Z-form DNA and daunorubicin preferentially binds to B-form DNA. The preference of WP900 for Z-form DNA arises from the orientation and shape complementarity of the D-daunosamine and the minor J25 j. B. Chaires, J. Biol. Chem. 261, 8899 (1986). 126 A. H. Wang, G. Ughetto, G. J. Quigley, and A. Rich, Biochemistry 26, 1152 (1987).

310


[ 14]

FIG.2. Energy-minimizedaveragemodelof the left-handedZ-DNA-WP900intercalationcomplex formed with the d(CGCGCGCG)2duplex. groove. In contrast, the L-daunosamine group of daunorubicin has steric clashes with the minor groove and backbone of Z-DNA, which causes a positional change of the intercalated chromophore that decreases the overall integrity of the binding site. Analogous effects are observed for WP900 and the disfavored B-form DNA. Intercalation of WP900 into Z-DNA is the likely binding mode, a supposition that is supported by our computer simulations using molecular dynamics (Fig. 2), where a stereochemically reasonable WtX)00 : Z-DNA intercalation complex can be formed. Because the same methods accurately reproduce the structure of the

[14]


31 1

B-DNA : daunorubicin intercalation complex, for which high-resolution X-ray data are available, the model derived from molecular dynamics is plausible. We note that a variety of computer simulations have been performed previously in attempts to obtain models for a Z-DNA intercalation complex.127' 12s The methods employed in our studies represent significant improvements over these earlier attempts for several reasons. First, a greatly improved AMBER force field with the PME summation method has been used. Second, our simulations include both solvent and ions, whereas earlier studies modeled structures in vacuo. Finally, reliable highresolution structural data for our system were available as starting models for at least the B family of structures used in our simulations. The average structure obtained from the molecular dynamics simulation for the daunorubicin : B-DNA intercalation complex is consistent with, and essentially reproduces, existing structural data. J26 The Z-DNA simulations are in ~imilar close agreement with existing structural data ~°4 for Z-DNA for the nonintercalated region of the complex. No structural data exist for a Z-DNA intercalation complex. Modeling Protocol. The starting model for daunorubicin was taken from the literature j26 and was inverted to create the WP900 enantiomer. Parameterization and partial charges were obtained from ab initio calculations at the 6-31 G* level with the GAMESS 99 package, using the restrained electrostatic potential fit routine. 114,115 The geometry of the resulting daunorubicin closely resembled the published crystal structure. 126Starting models for the right-handed B and left-handed Z forms of d(CGCGCGCG)2 were created with the standard conformations provided in the Macromodel program. 98 The intercalation site (highlighted in bold face) for the B-form DNA was built by inserting a standard B-form phosphate backbone intercalation site conformation. 126 The Z-form intercalation site (highlighted in boldface) was built by separating the bases of the Z-form of d(CGCGCGCG)2 by 6.8 A with reduction of the helical twist and inserting phosphate backbones, which were subsequently optimized before the intercalator was added. Daunorubicin and wPg00 ligand molecules were then inserted into the generated intercalation site by superimposing the DNA : daunorubicin crystal structure, and four models were used for simulations: B-forms of DNA : daunorubicin and DNA : WP900, and the corresponding Z-forms of DNA : daunorubicin and DNA : WP900. The intercalation sites were optimized by molecular mechanics and dynamics with restraints of 100 kcal (tool - A) i applied to the nonintercalation site bases using the implicit solvation within Macromodel. The optimal positioning of each intercalator was determined by a grid search with translation of 8 × 8 A and rotation ~ 9 0 (using 1 ]~ and 15~ resolution steps). The resulting structures were used as starting structures for the explicit solvation calculations. [27 M. Prabhakaran and S. C. Harvey, Biopolymers 27, 1239 (1988). 128 G. Gupta, M. M. Dhingra, and R. H. Sarma, .I. Biomol. Struct. Dyn. l, 97 (1983).

312


[ 14]

Model structures were hydrated using standard A M B E R 5.031 rules in a 10-~ box of TIP3P waters. Sodium cations for the B-form complexes were added, using the Edit placement routine, and chloride anions were added randomly for charge neutrality. The Na + and C1- counterions for the Z-form complexes were added randomly to simulate a neutral 3 M aqueous NaC1 solution. The systems were heated slowly to 300 K and equilibrated carefully during 50 psec with gradual decrease of restraints on the DNA and ligand from 100 to 10 kcal (mol. ~ ) - I . Further 400- and 600-psec periods of restrained molecular dynamics were used for the B-form and Z-form complexes, respectively, to ensure complete solvent equilibrium. Molecular dynamics simulations using the AMBER95 force field were performed in the isothermal isobaric ensemble (P = 1 atm, T = 300 K) with the A M B E R 5.0 program, 31 using periodic boundary conditions and the PME algorithm. A 2-fsec time step was used with all bond distances frozen by using SHAKE. After heating and equilibration, molecular dynamics production runs of 800 psec were used to derive average structures for the complexes taken from 50 snapshots accumulated in the last 50 psec with subsequent minimization. Triplex Simulation o f Antigene Approach

The c-myc gene plays an important role in the regulation of cellular proliferation and differentiation. The c-myc protooncogene encodes a DNA-binding phosphoprotein129 131 that plays an important regulatory role in cell proliferation and differentiation, and is overexpressed in many types of cancer. The c-myc gene product appears to play an important role in the final common pathway regulating cellular proliferation, which makes it an ideal target for transcriptional inhibition. The Miller group has shown that DNA-binding drugs and triplex-forming oligonucleotides inhibit transcription by preventing the binding of regulatory proteins that are necessary for transcriptional activity.132'133 They have developed the molecular, cellular, and whole animal model systems with which to effectively screen for synthetic compounds that inhibit expression of target genes in a sequence-specific manner. They have paid particular attention to factors that influence the intracellular stability of triplex-forming oligonucleotides and that allow formation of triplex DNA by a broader variety of sequences. 134 129E O. Donner, I. G. Wilke, and K. Moelling,Nature (London) 296, 262 (1982). 130M. D. Cole, Cell GrowthDiffer. 65, 715 (1991). J31 K. Kelly and U. Siebenlist, J. Biol. Chem. 263, 4828 (1998). 132V. W. Campbell, D. Davin, S. Thomas, D. Jones, J. Roesel, R. Tran-PaUerson, C. A. Mayfield, B. Rodu, D. M. Miller, and R. A. Hiramoto,Am. J. Med. Sci. 307, 167 (1994). 133D. Chaudharyand D. M. Miller, Biochemistry 34, 3438 (1995). 134N. Vigneswaran, C. A. Mayfield, B. Rodu, R. James, H. G. Kim, and D. M. Miller, Biochemistry 35, 1106 (1996).

[14]


313

There has been much interest in triplex approaches to antigene therapy. This is primarily due to the potential for durable sequence-specific inhibition of gene expression, as demonstrated by reports showing effective in vitro and cell culture inhibition of transcription of several genes by triplex-forming oligonucleotides. A contributing factor is the Food and Drug Administration (FDA) approval of an antisense oligonucleotide for use in acquired immunodeficiency syndrome (AIDS) patients with cytomegalovirus (CMV), demonstrating that oligonucleotides can be used safely as therapeutic agents. Several approaches to triplex formation and stabilization are being used, including modified oligonucleotides, 135 137 DNA binders, and oligonucleotide-DNA binder conjugates ~38,139that stabilize triplex formation. Use of the pyrimidine motif (pyrimidine third-strand hydrogen bonded via Hoogsteen bonds parallel to the purine strand of the duplex) has not been examined in detail as the cytosine in the third strand must be protonated. One article 14° has shown that the pH requirement of the pyrimidine motif may not be as problematic as previously thought. The NMR experiments unequivocally showed that the pKa of the protonated cytosine residue is higher than physiologic pH for internal positions. The advantage of the pyrimidine motif over the purine motif is that the purine triplex-forming oligonucleotide is always G-rich and, under physiologic concentrations of monovalent cations, especially K + and Na +, the third strand can self-associate and form dimers and tetramers in the presence of these ions, thereby reducing the effective concentration of the third strand available for triplex formation. There are problems associated with all oligonucleotide-based therapies, such as reagent delivery, cellular uptake, and stability (both thermodynamic and enzymatic). Because of these issues, and in-depth understanding of triplex binding is critical in optimizing triplex-forming oligonucleotide design. Nuclease sensitivity of phosphodiester oligonucleotides is a major problem hindering their use as gene-specific compounds in vivo. RNA oligonucleotides with modifications at the T-position of the ribose sugar are more nuclease resistant than unmodified DNA oligonucleotides and are able to form stable triplexes with duplex DNA.141 It has been shown that the most stable RNA :DNA triplexes are

135 j. Michel, G. Gueguen, J. Vercauteren, and S. Moreau, Tetrahedron 53, 8457 (1997). 136 E J. Bates, C. A. Laughton, T. C. Jenkins, D. C. Capaldi, P. D. Roselt, C. B. Reese, and S. Neidle, Nucleic Acids Res. 24, 4176 (1996). 137 S. A. Cassidy, E Slickers, J. O. Trent, D. C. Capaldi, E D. Roselt, C. B. Reese, S. Neidle, and K. R. Fox, Nucleic Acids Res. 25, 4891 (1997). 138 j. Robles and L. W. McLaughlin, J. Am. Chem. Soc. II9, 6014 (1997). 139 A. E Faruqi, S. H. Krawczyk, M. D. Matteucci, and P. M. Glazer, Nucleic Acids Res. 25, 633 (1997). 140 j. L. Asensio, A. N. Lane, J. Dhesi, S. Bergqvist, and T. Brown, J. Mol. Biol. 275, 8 l 1 (1998). 141 A. M. Iribarren, B. S. Sproat, E Neuner, I. Sulston, U. Ryder, and A. L. Lamond, Ptw'. Natl. Acad. Sci. U.S.A. 87, 7747 (1990).

314


[ 141

2'-O-methyl-RNA third strand bound to duplex DNA. 142 The unmodified RNA third strand bound to duplex DNA is less stable that the T-modified RNA but is more stable than the parent DNA triplex. An NMR structure of an intramolecular RNA third strand [r(UCUCUCUU)] : DNA duplex linked by hexakis(ethylene glycol) has been reported by Feigon's group. 143 Another relevant structure has been reported by Lane's group/44 This is of an intramolecular RNA third strand [r(UCUUC)] : DNA duplex linked by hexakis(ethylene glycol), and an intramolecular 2'-O-methyl-RNA third strand [r(UCUUC)] : DNA duplex linked by hexakis(ethylene glycol). However, the structures from the two groups are not in complete agreement with respect to the sugar conformations, base inclination, and backbone parameters. As these are the only two NMR structures deposited, there is an obvious need for more structural information to define the RNA and modified RNA : DNA triplex system. A molecular modeling report 62 used the AMBER force field to calculate 1-nsec molecular dynamics trajectories for the TAT pyrimidine triplex, starting from the A- and B-forms (analogous to A- and B-DNA conformation) and the P-form (analogous to the peptide nucleic acid structure, which can be loosely described as an exaggerated A-form). The structures converged to a single structure that was consistent with the B-form proposed by NMR-derived structures from Patel's group.145'146 This study further demonstrated that high-level simulations can model the A- to B-form transition, as the starting models were in the A, B, and P forms. We have applied molecular dynamics simulations to determine the structure of the RNA and 2'-modifed RNA triplex-forming third strand targeted to the P2 promoter of the protooncogene c-myc. As there was no structure available when this work commenced, we implemented a similar protocol as Shields et al. 62 That is, we started with three models for the sequence r(UUUCUUC) • d(CGAAAGAAGTTTTCTTCTTTCG), the A-form triplex (all three strands in the A form), the B-form triplex (all three strands in the B form), and the AB form (duplex DNA in the B form and the triplex third strand in the A form) (Fig. 3). The target triplex was selected for several reasons: the duplex DNA is part of the murine c-myc P2 promoter 5'-d(AAAGAAG); the target duplex has no contiguous guanine bases, which would require destabilizing contiguous protonated cytosine bases in the third strand; a hairpin duplex decreases the possible number of species in solution

142E J. Bates, J. E Reddock, S. D. Thomas, V. Guarcello, A. Arrow, R. Dale, and D. M. Miller, unpublished results (2000). 143 C. H. Goffredgen, E Shultze, and J. Feigon, J. Am. Chem. Soc. 120, 4281 (1998). 144 j. L. Asensio, R. Carr, T. Brown, and A. N. Lane, J. Am. Chem. Soc. 121, 11063 (1999). 145 I. Radhakrishnan, X. Gao, C. de los Santos, D. Live, and D. J. Patel, Biochemistry 30, 9022 (1991). 146 I. Radbakrishnan and D. J. Patel, Structure 2, 395 (1994).

[14]


A-form

AB-form

315

B-form

FIG.3. Startingmodels for the moleculardynamicsstudies of the RNA : DNA triplex systems. and aids in the assignment of the resonance signals of the duplex. The T4 loop was designed from the literature, 147 and dCG was added to the 5' site to give enhanced stability by avoiding AT fraying and to give information on the triplex-duplex boundary. The nanosecond trajectories, after 155 psec of equilibrium, for the three starting models converge to a single AB-hybrid form (distinct from the starting AB form) (Fig. 4, Tables I and II). This shows again that the modeling protocol has allowed transition of the A-form DNA duplex to the more stable B-form DNA duplex (the AB-form triplex has an underlying B-form duplex). The Feigon structure t43 generally compares well with the modeled structure, with subtle differences in the sugar conformations, base inclination, and backbone parameters. The converged model is in much closer agreement with the Lane structure. 144 It should be noted that the two most recent RNA third-strand structures are not the same in the detailed analysis and there may be some sequence dependence. It has been shown that the most stable RNA : DNA triplexes are T-O-methylmodified RNA third strand to duplex DNA. 142 A surface plasmon resonance technique (BIAcore, Uppsala, Sweden) also has been used 142 to examine the real-time association and dissociation between a 24-mer triplex-forming oligonucleotide and a 30-bp hairpin duplex representing the murine c - m y c P2 promoter. These experiments showed that both the association and dissociation of the triplex were slower for a 2'-O-methyl-RNA third strand compared with its RNA analog. We undertook a similar modeling approach as with the RNA : DNA triplex for the T-O-methylRNA : DNA triplex. 147Z. Kuklenyik,S. Yao,and L. G. Marzilli,Eur. J. Biochem. 236, 960 (1996).

316


[ 141

FIG.4. The final converged average AB-form structure for the RNA : DNA triplex systems. Starting from the A, AB, and B forms, the A and AB forms converged to a single AB-form structure. However, the B form did not converge and the base triplets were disrupted because of the unfavorable steric clash of the T-O-methyl group with the backbone hindering the B-to-A transition. It should be noted that the ends of the triplex regions frayed, which is consistent with DNase I footprinting data 142 of the longer RNA triplex formed at the c-myc P2 promoter sequence. The calculated order of stability for the modeled triplexes is DNA triplex < RNA : DNA triplex double duplex invasion >> triplex > duplex invasion. However, whereas the invasion complexes have a slow rate of formation, which is further decreased by increasing ionic strength, major groove triplexes form much faster, but virtually nothing is known about their properties in general at this stage. Therefore, if possible, the best advice is to target a homopurine site of 8-10 bases in length, using a bis-PNA (Fig. 6). The complex loses stability below about 8 bases and sequence specificity above about 10-12 bases. It is also advisable to include three or four positive charges, for example, in the form of lysines, in order to increase the binding rate. Most probably the increased cationic character of the PNA ensures a high local concentration of the PNA in close proximity to the 50 V. V. Demidov, M. V. Yavnilovich, and M. D. Frank-Kamenetskii, Biophys. J. 72, 2763 (1997). 51 j. Lohse, C. Hui, S. H. SOnnichsen, and P. E. Nielsen, in "DNA and RNA Cleavers and Chemotherapy of Cancer and Viral Diseases" (B. Meunier, ed.), pp. 133-141. NATO ASI Series. Kluwer Academic Publishers, Dordrecht, The Netherlands, 1996. .52 p. Bigey, S. H. S0nnichsen, P. E. Nielsen, and B. Meunier, Bioconjug. Chem. 3, 267 (1997). 53 M. Footer, M. Egholm, S. Kron, J. M. Coull, and P. Matsudaira, Biochemistry 35, 10673 (1996). 54 B. Armitage, T. Koch, H. Frydenlund, H. Orum, H. G. Batz, and G. B. Schuster, Nucleic Acids Res. 25, 4674 (1997).

[15]

PNA TARGETINGOF DOUBLE-STRANDEDDNA

339

H-Lys-Lys-TTTTJJTJTJ ~ . 5- AAAAGGAGAG ~-- 3 ' H2N- Lys- TTTTCCTCTC ~ FIG. 6. Schematic drawing of a bis-PNA binding to its purine target. The bis-PNA is constructed with G-recognizing pseudoisocytosine (J) in the parallel Hoogsteen strand, and the Watson-Crick strand is binding in an antiparallel orientation. The linker is conveniently constructed from three 8amino-3,6-dioxaoctanoic acid units [-NH-(CH2)2-O-(CH2)2-O-(CH2)-CO-]. H-, The amino terminal of the PNA; H2N-, the amidated carboxyl terminal.

DNA, thereby increasing the probability of invading the dynamically breathing and therefore transiently open DNA helix. 27'37'38 Finally, pseudoisocytosine should be used in the Hoogsteen strand to avoid pH sensitivity of the binding. On the other hand, if binding is possible at acidic pH, the presence of cytosines in the Hoogsteen strand will increase both binding rate and sequence discrimination. 2° If homopurine targets are not available, the double duplex invasion principle is an obvious option, although not yet thoroughly explored and characterized. In this case the target (about 10 bp) should contain at least 50% AT base pairs. In principle, it should also be possible to target nonhomopurine targets by triplex-forming bis-PNAs if novel bases are employed that recognize cytosine or thymine in the "Hoogsteen mode" [such as the E-base 55 (Fig. 7)]. Many attempts to construct such bases have been reported in a DNA as well as a PNA context, but the results have been rather disappointing56; no efficient candidates have yet been identified. Finally, under certain conditions not yet fully understood simple duplex invasion complexes may form and be sufficiently stable to yield biological effects, but this area needs much further research before it can be recommended as a reliable method.

° ...... ....... H

.,.,..'1~

" - - _ _ - ~ N''H''" T

l

N~'~,,

N'%"~N

A

0

FIG. 7. Chemical structure of a putative triplet formed by binding of the E-base to the T-A base pair.

55 A. B. Eldrup, O. Dahl, and E E. Nielsen, J. Am. Chem. Soc. 119, 11116 (1997). 56 K. R. Fox, Curr. Med. Chem. 7, 17 (2000).

340

CHEMICAL AND MOLECULAR BIOLOGICAL APPROACHES

[ 1 6]

Concluding Remarks The results obtained so far clearly show that PNA is an interesting and useful reagent for sequence-specific targeting of double-stranded DNA. It is also clear that we are far from understanding all the relevant aspects, both as regards binding modes and mechanism of this methodology. This is particularly true when considering effects and possible (medical) applications in live cells and animals. Continued studies and development of peptide nucleic acids should yield both a deeper understanding of bimolecular recognition principles and also contribute to various areas of molecular biology, gene therapeutic drug development, and genetic diagnostics. Acknowledgment This work was supported by the Lundbeck Foundation.

[16]

Drug

By

Interaction

with

Triple-Helical

CHRISTOPHE ESCUD[~, THI~RESE GARESTIER,

and

Nucleic

Acids

JIAN-SHENG SUN

Introduction Triple-helical DNA was first discovered more than 40 years ago. Renewed interest in this structure developed at the end of the 1980s because of two discoveries. First, sequence-specific recognition of double-helical DNA can be achieved by short oligonucleotides forming a short intermolecular triple helix.l'2 This recognition can be used to design tools for manipulating double-stranded DNA in vitro as well as for interfering with DNA-related processes in vivo. Numerous studies have now shown that the so-called triplex-forming oligonucleotides (TFOs) can be used to cleave or chemically modify a DNA target, 3 and that TFOs can interfere with binding of transcription factors, transcription elongation, or DNA repair. 4 Moreover, it has been shown that DNA can adopt an intramolecular triple-helical structure, also called H-DNA, under certain conditions. 5 The formation of this structure has been shown in bacteria and may play a role in the 1 T. Le Doan, L. Perrouault, D. Praseuth, N. Habhoub, J.-L. Decout, N. T. Thuong, J. Lhomme, and C. H61~ne, Nucleic Acids Res. 15, 7749 (1987). 2 H. E. Moser and P. B. Dervan, Science 238, 645 (1987). 3 N. T. Thuong and C. H61~ne, Angew. Chem. Int. Ed. 32, 666 (1993). 4 C. Giovannangeli and C. H6l~ne, Antisense Nucleic Acid Drug Dev. 7, 413 (1997). 5 R. D. Wells, D. A. Collier, J. C. Hanvey, M. Shimizu, and F. Wohlrab, FASEB J. 2, 2939 (1988).


Copyright© 2001by AcademicPress All rightsof reproductionin any formreserved. 0076-6879/00 $35.1,)0

340


[ 1 6]

Concluding Remarks The results obtained so far clearly show that PNA is an interesting and useful reagent for sequence-specific targeting of double-stranded DNA. It is also clear that we are far from understanding all the relevant aspects, both as regards binding modes and mechanism of this methodology. This is particularly true when considering effects and possible (medical) applications in live cells and animals. Continued studies and development of peptide nucleic acids should yield both a deeper understanding of bimolecular recognition principles and also contribute to various areas of molecular biology, gene therapeutic drug development, and genetic diagnostics. Acknowledgment This work was supported by the Lundbeck Foundation.

[16]

Drug

By

Interaction

with

Triple-Helical

CHRISTOPHE ESCUD[~, THI~RESE GARESTIER,

and

Nucleic

Acids

JIAN-SHENG SUN

Introduction Triple-helical DNA was first discovered more than 40 years ago. Renewed interest in this structure developed at the end of the 1980s because of two discoveries. First, sequence-specific recognition of double-helical DNA can be achieved by short oligonucleotides forming a short intermolecular triple helix.l'2 This recognition can be used to design tools for manipulating double-stranded DNA in vitro as well as for interfering with DNA-related processes in vivo. Numerous studies have now shown that the so-called triplex-forming oligonucleotides (TFOs) can be used to cleave or chemically modify a DNA target, 3 and that TFOs can interfere with binding of transcription factors, transcription elongation, or DNA repair. 4 Moreover, it has been shown that DNA can adopt an intramolecular triple-helical structure, also called H-DNA, under certain conditions. 5 The formation of this structure has been shown in bacteria and may play a role in the 1 T. Le Doan, L. Perrouault, D. Praseuth, N. Habhoub, J.-L. Decout, N. T. Thuong, J. Lhomme, and C. H61~ne, Nucleic Acids Res. 15, 7749 (1987). 2 H. E. Moser and P. B. Dervan, Science 238, 645 (1987). 3 N. T. Thuong and C. H61~ne, Angew. Chem. Int. Ed. 32, 666 (1993). 4 C. Giovannangeli and C. H6l~ne, Antisense Nucleic Acid Drug Dev. 7, 413 (1997). 5 R. D. Wells, D. A. Collier, J. C. Hanvey, M. Shimizu, and F. Wohlrab, FASEB J. 2, 2939 (1988).


Copyright© 2001by AcademicPress All rightsof reproductionin any formreserved. 0076-6879/00 $35.1,)0

[16]

DRUG INTERACTIONWITHTRIPLE-HELICALNAs

341

regulation of some eukaryotic genes (see Mirkin and Frank-Kamenetskii 6 for a review). One limitation to the possible use of TFOs as therapeutic agents or the possible biological role of H-DNA is the low stability of some commonly studied triplexes under physiological conditions. This problem may be overcome by the action of proteins that can specifically bind triple-helical DNA, and thus would be able to stabilize them. 7-9 Alternatively, low molecular weight compounds that can bind triplex DNA more tightly than duplex DNA would also have the ability to enhance triplex stability. If their target is an intramolecular triplex, whose formation may interfere with the regulation of gene expression, these molecules may directly affect the gene expression profile. Therefore, such molecules constitute potential drugs, either in conjunction with TFOs or by themselves as structure-specific DNA-binding compounds. In this chapter, after a brief description of the characteristics of triple-helical DNA, we describe the physicochemical aspects of drug binding to triplex DNA, with special emphasis on the different methods used to study the binding of low molecular weight compounds to triplex DNA, through a review of examples published in the literature. We then discuss some methodological and general considerations about targeting triplehelical nucleic acids with drugs, and present the biological applications of such drugs. C h a r a c t e r i s t i c s of T r i p l e - H e l i c a l DNA The structure of triple-helical DNA has been extensively studied (see Radhakrishnan and Patel l° for a review). A triplex-forming oligonucleotide binds into the major groove of double-stranded DNA (dsDNA) at an oligopyrimidine • oligopurine sequence. Specific recognition is achieved by formation of hydrogen bonds between bases in the third strand and purine bases in the target sequence (Fig. 1) (see Sun et al. 11 for more details). Pyrimidine-containing oligonucleotides bind in a parallel orientation with respect to the oligopurine sequence by formation of Hoogsteen hydrogen bonds between a T or a protonated C in the third strand and a T. A or a C. G base pair in the duplex target, respectively. Oligonucleotides containing purines can also bind double-helical DNA. G can form two reverse Hoogsteen hydrogen bonds with a C. G base pair and A or T can do so with a T. A base pair. In this configuration, binding of the o!igonucleotide occurs in an antiparallel orientation. Some (G, T)-containing oligonucleotides may also bind parallel

6 S. M. Mirkin and M. D. Frank-Kamenetskii, Annu. Rev. Biophys. Biomol. Struct. 23, 541 (1994). 7 A. L. Guieysse, D. Praseuth, and C. H61~ne, J. MoL Biol. 267, 289 (1997). 8 M. Musso, L. D. Nelson, and M. W. Van Dyke, Biochemistry 37, 3086 (1998). 9 L. D. Nelson, M. Musso, and M. W. Van Dyke, J. Biol. Chem. 275, 5573 (2000). l0 I. Radhakrishnan and D. J. Patel, Biochemistry 33, 11405 (1994). 11 j. S. Sun, T. Garestier, and C. H61~ne, Curr. Opin. Struct. Biol. 6, 327 (1996).

342


Pyrimidine 3'

[1 g]

motif

51

~'-W-C W...~

-~--C-H

C..->

109 M-1 for duplex DNA. 27 This technique also allows some structural investigation. The CD spectrum of a ligand bound to double-stranded DNA can differ from that of the same ligand bound to triple-stranded DNA. Such a phenomenon, which has been observed with Hoechst 29 or berenil, 3° but not with netropsin, suggests that the conformations of the ligand are slightly different when it is bound to the duplex or the triplex. Linear Dichroism

Linear dichroism (LD) can been used to study the binding of polyaromatic compounds to homopolymeric helical structures. For BePI, BgPI, the benzo[f]pyrido[4,3-b]quinoxaline (BfPQ) derivative (9), and BQQ (6), the similar magnitude of reduced LD of the electronic transitions of the polynucleotide bases and of the bound ligands suggests an orientation of the plane of the fused-ring systems perpendicular to the helix axis, which is indicative of intercalation. 33 This technique

25 D. S. Pilch, M. J. Waring, J. S. Sun, M. RougEe, C. H. Nguyen, E. Bisagni, T. Garestier, and C. H61bne, J. Mol. Biol. 232, 926 (1993). 26 y. W. Park and K. J. Breslauer, Proc. Natl. Acad. Sci. U.S.A. 89, 6653 (1992). 27 M. Durand, N. T. Thuong, and J. C. Maurizot, J. Biol. Chem. 267, 24394 (1992). 28 M. Durand and J. C. Maurizot, Biochemistry 35, 9133 (1996). 29 M. Durand, N. T. Thuong, and J. C. Maurizot, Biochimie 76, 181 (1994). 30 M. Durand, N. T. Thuong, and J. C. Maurizot, J. BiomoL Struct. Dyn. 11, 1191 (1994). 31 D. S. Pilch, M. A. Kirolos, and K. J. Breslauer, Biochemistry 34, 16107 (1995). 32 D. S. Pilch and K. J. Breslauer, Proc. Natl. Acad. Sci. U.S.A. 91, 9332 (1994). 33 S. K. Kim, J. S. Sun, T. Garestier, C. Hfl~ne, E. Bisagni, A. Rodger, and B. Norden, Biopolymers 42, 101 (1997).

[1 B]


347

has also shown that various complexes of ruthenium can intercalate into triplehelical DNA and stabilize it. 34

Nuclear Magnetic Resonance Spectroscopy Only two studies have employed nuclear magnetic resonance (NMR) to explore ligand binding to triplex DNA. In the first, NMR was used to monitor the solvent exchange properties of the exchangeable protons of the ligand. This technique allowed measurement of self-association constants for both the BePI (1) and BgPI (2) derivatives, 23 but did not provide high-resolution structural information. The binding of coralyne was also investigated by using short intramolecular triplex DNA with different sequences. 35 The intercalative binding mode was confirmed. However, at NMR concentrations, and even in the presence of an excess of DNA, two different binding sites were observed: a primary site between two T-A*T triplets, and a neighboring site with a lower affinity. These complexes led to two sets of overlapping spectra, which made the precise assignment of proton resonances impossible.

Hydrodynamic Methods Viscometric titration of closed circular duplex DNA is commonly used to determine whether a compound intercalates into double-helical DNA. Intercalating agents usually induce the unwinding of supercoiled plasmids and the lengthening of linear helical structures, which are reflected by an increase in the viscosity of the samples. Several polyaromatic molecules that stabilize triple helices bind duplex DNA by intercalation, as evidenced by unwinding of supercoiled plasmid DNA. 36 In the case of BePI, an increase in the relative contour lengths of both double- and triple-stranded homopolymer structures [i.e., poly(dA) • poly(dT) and poly(dA) • [poly(dT)]2] has been demonstrated. 25

Competition Dialysis Competition dialysis techniques provide a direct method to obtain the DNAbinding preferences of a candidate ligand. For example, this method has been used to show that the 1,4-disubstituted anthraquinone 11 binds preferentially to homopolymeric duplexes whereas the 2,6-derivative 12 favors binding to triplexes, 37 An improved competition dialysis procedure has been developed that allows quick 34 S.-D. Choi, M.-S. Kim, S. K. Kim, E Lincoln, E. Tuite, and B. Norden, Biochemistry 36, 214 (1997). 35 A. A. Moraru-Allen, S. Cassidy, J. L. Asensio Alvarez, K. R. Fox, T. Brown, and A. N. Lane, Nucleic Acids Res. 25, 1890 (1997). 36 C. Marchand, C. Bailly, C. H. Nguyen, E, Bisagni, T. Garestier, C. H61~ne, and M. J. Waring, Biochemistry 35, 5022 (1996). 37 I. Haq, J. E. Ladbury, B. Z. Chowdhry, and T. C. Jenkins, J. Am. Chem. Soc. 118, 10693 (t996).

348

C H E M I C AAND L MOLECULARBIOLOGICALAPPROACHES

[16]

and efficient comparison of the relative affinity of DNA-binding agents for different structural motifs, such as triplex DNA, i-DNA, tetraplex DNA, or Z-DNA and various duplex DNA samples with different base contents. 38 In this method, equal volumes of solutions of different DNA structures at the same concentration are dialyzed against a common solution containing the ligand being studied. After dialysis, the amount of ligand bound to a particular DNA structure is measured by spectrophotometry. The amount of bound ligand is proportional to the association constant for binding of the ligand to this DNA structure. This study has confirmed the structural preference for triplex DNA (represented by a poly(dA) • [poly(dT)]2) of previously studied DNA-binding agents, such as BePI or coralyne, and made possible the discovery of the strong triplex selectivity of some less well-characterized ligands, such as berberine (13) or a novel 3,3'-diethyloxadicarbocyanine (15). 39 Further experiments are required to apply this technique to triplexes with different sequences. Calorimetric Methods

Isothermal titration calorimetry (ITC) provides the only means to determine directly the molar calorimetric binding enthalpy, which can be used to determine the equilibrium binding constant for a bimolecular DNA-ligand interaction. This method has shown that preferential binding of the disubstituted amidoanthraquinone derivative 12 to triplex DNA is enthalpically driven whereas binding of 11 to duplex DNA is entropically driven. 37 This technique has also suggested that the stoichiometry of binding of netropsin to a triplex would be 2 : 1.4° One molecule of netropsin would bind in the Watson-Crick groove, and another would bind in the Watson-Hoogsteen groove. The assignment of this new binding site was suggested by the fact that the AH for binding of netropsin to this second site was exothermic. This is expected for binding of a ligand in a hydrophobic groove by a mechanism involving mostly van der Waals interactions. Gel-Shift Assays

Triplex formation can be detected by gel-shift methods. Binding of the third strand to a radiolabeled duplex results in the formation of a complex with a slower migration rate. Low molecular weight drugs are too small to induce a significant mobility shift when binding to triplex DNA. However, the formation of a triple helix that does not form in the absence of ligand can be detected, and an apparent dissociation constant can be estimated by the concentration of third strand required 38j. Ren and J. B. Chaires, Biochemiso'y 38, 16067 (1999). 39j. Ren and J. B. Chaires, J. Am. Chem. Soc. 122, 424 (2000). 40 D. Rentzeperis and L. A. Marky,J. Am. Chem. Soc. 117, 5423 (1995).

[16]

DRUG INTERACTIONWITH TRIPLE-HELICALNAs

349

for 50% triplex formation. Such a method has shown that BeP141 and various polyamine analogs 42 can stabilize triple helices with a GT antiparallel third strand. In the former case, triplex formation could be detected at 4 °, but not at room temperature, although the melting temperature (measured by thermal denaturation experiments) was as high as 51 °. This is probably due to too short a lifetime of the triplex-drug complex at room temperature, which results in dissociation of the complex during gel migration.

Footprinting Methods Footprinting methods provide a powerful tool for investigating the sequence specificity of DNA-binding agents. DNase I, dimethyl sulfate (DMS), diethyl pyrocarbonate (DEPC), osmium tetroxide, EDTA-Fe II, and methidiumpropyl-EDTA (MPE) have been used to probe triplex DNA alone or complexed with triplexbinding agents. Cleavage by DNase I, DMS, or MPE is inhibited by triplex formation itself. In the presence of a specific triplex-binding agent, the concentration of third strand required to produce a footprint is lower than that required in the absence of ligand. For example, DNase I and MPE-Fe I1 footprinting reveal that the binding constant for a 27-mer oligonucleotide is enhanced by a factor of 10 in the presence of the BfPQ derivative 9. 36 As another example, the concentration of a 10-mer oligonucleotide required for 50% inhibition of cleavage by DNase I can be reduced by at least a factor of 100 (i.e., from more than 50 to 0.5 #M) in the presence of a 10 # M concentration of the naphthylquinoline derivative 7. 43 DMS footprinting has been used to show that basic oligopeptides can stabilize triplehelical DNA. 44 It should be mentioned that footprinting methods can also detect structural changes induced by the ligands. For example, an enhanced cleavage of the oligopyrimidine target strand at a specific site, surrounded by footprints on both sides, has been reported when the binding of a GT antiparallel third strand was investigated by DNase I footprinting in the presence of BePI. 41 Enhanced cleavage has also been described at the 5'-thymine of 5'-TTC sequences, when probing a triple helix with a pyrimidine third strand in the presence of BPQ and using MPE as cleaving agent. 36

Molecular Modeling In the absence of high-resolution structural information, quantitative structureactivity relationship (QSAR) studies and/or molecular modeling can be used to 41 C. Escudr, J. S. Sun, C. H. Nguyen,E. Bisagni, T. Garestier, and C. Hrl~ne, Biochemistry35, 5735 (1996). 42 M. Musso, T. Thomas, A. Shirahata, L. H. Sigal, M. W. VanDyke,and T. J. Thomas, Biochemistry 36, 1441 (1997). 43 S. A. Cassidy, L. Strekowski, W. D. Wilson, and K. R. Fox, Biochemistry33, 15338 (1994). 44W.N. Potaman and R. R. Sinden, Biochemistry34, 14885 (1995).

350

CHEMICAL AND MOLECULARBIOLOGICALAPPROACHES

[16]

understand the critical chemical moieties required for preferential binding to triplex DNA. This has been done for related compounds in the benzopyridoindole family, where molecular modeling has provided some insights into possible mechanisms responsible for the different efficiency of the compounds by proposing a model for intercalation of these compounds into triplex DNA. 45 According to this model, the side chain of BePI (1) lies in the Watson-Hoogsteen groove whereas the side chain of BgPI (2) lies in the Watson-Crick groove. A similar model was built for the benzo[f]pyrido[3,4-b]quinoxaline derivative 3. This model suggested that stacking interactions with base triplets could be improved by adding a new ring to the aromatic system. The resulting rational design of a new compound based on this model gave a very efficient triplex-binding agent, BQQ (6). ~9 The same strategy applied to BePI (1) and BgPI (2) led to the synthesis of the new benzoindoloquinoline derivatives 4 and 5. 46 Another study provided a rationale for the different effects of the disubstituted anthraquinone derivatives 11 and 12 on triple helices.47 Molecular modeling suggested that 12 can intercalate with one side chain located in the Watson-Crick groove whereas the other lies in the WatsonHoogsteen groove. The threading intercalation provides good stacking interactions between the chromophore and base triplets. For 11, however, the relative position of the two side chains does not accommodate both intercalation and binding of the third strand. As a result, the triple helix is destabilized. Smface Plasmon Resonance

Surface plasmon resonance has been used to study the kinetics of a pyrimidinemotif triple-helix formation in the presence and in the absence of BPI ligands. Under given conditions, the presence of 6 # M BePI (1) or BgPI (2) accelerates the association rate constant by 27- and 23-fold, respectively, as compared with that of the TFO alone. The dissociation rate constant was decreased by a factor of 1.2 and 4.5, respectively. Therefore, these triplex-specific ligands enhance the binding constant of TFO by 32- and 104-fold, respectively (Sun et al., unpublished data, 1999). The main contribution of BPI ligands to triplex formation is evidently to the association phase. This implies that BPI ligands likely act on the nucleation step by stabilization of the initiation fragment. BgPI derivatives exhibited a larger effect on the dissociation (and thus increased the lifetime of the triplex) than BePI. This can be explained by a better stacking interaction of BgPI with base triplets than BePI. It should be pointed out that the lifetime of the triplex is relevant when the triple-stranded structure must interfere with transcription or other biological 45C. Escud6, C. H. Nguyen,J. L. Mergny,J. S. Sun, E. Bisagni, T. Garestier,and C. H61bne,J. Am. Chem. Soc. 117, 10212 (1995). 46C. H. Nguyen,C. Marchand, S. Delage,J. S. Sun, T. Garestier.C. HEl~ne,and E. Bisagni,J. Am. Chem. Soc. 120, 2501 (1998). 47K. R. Fox, E Polucci, T. C. Jenkins, and S. Neidle, Piw'. Natl. Acad. Sci. U.S.A. 92, 7887 (1995).

[161


351

process. Kinetic knowledge of the effect of triplex-specific ligands on triple-helix formation provides insight into their mechanisms of action, and helps to design even better ligands. Targeting Triple-Helical Nucleic Acids with Drugs

Choice of Experimental System Physicochemical methods are generally applicable to homopolymeric nucleic acid structures, mostly poly(dA) • [poly(dT)]2 or poly(dAG) • [poly(dTC)]2 in the case of triple helices. Such experiments are useful for determining the binding mode, observing the specific features of intercalative versus groove-binding compounds, and for assessing cooperative binding. Although they can reveal triplexstabilizing properties, they are poorly predictive of the capability of these compounds to stabilize short triplex sequences containing C. G*C + base triplets at low concentration. For example, coralyne was suggested to stabilize the poly(dAG) • [poly(dTC)]2 triple helix efficiently, but further experiments using short oligonucleotides showed that this effect was weak. 35 The choice of the sequence is also important. Ethidium bromide was shown to stabilize triple helices made o f T . A*T base triplets, but destabilized triple helices containing C • G*C + base triplets. 15This was attributed to electrostatic repulsion between ethidum bromide and the protonated cytosines. Except for hydrodynamic methods and linear dichroism, most physicochemical techniques can be applied to short oligonucleotides, which opens the possibility of exploring various sequence contexts. The use of intramolecular triple helices artificially enhances the stability of triple-helical DNA and ensures the formation of a correctly folded and stoichiometric structure, which can be an interesting advantage, especially in the design of NMR experiments.

Choice of Experimental Conditions Triple helix formation is known to be strongly dependent on pH and ionic conditions. Experimental results have shown that solution conditions also have an important influence on the ability of drugs to bind triplex DNA. Indeed, electrostatic interactions are likely to provide an important contribution to DNA binding. Aliphatic nitrogen atoms, often found on the side chains of intercalating compounds, are generally protonated below pH 8.0. The pKa of aromatic nitrogen atoms can be much lower. For example, the pKa of the aromatic N-10 atom of BePI (1) was measured by recording the absorption spectrum at different pH. A pKa of 6 was measured for the free ligand, which was increased to 8 when the ligand was bound to triple-helical DNA. 25 Binding to DNA can therefore stabilize the form with the highest affinity for DNA, that is, the protonated form. The naphthylquinoline derivative 7 was shown to stabilize a triple helix with a GA third strand only at acidic pH. This was likely due to a requirement for protonation of

352


[ 16]

the chromophore. It has also been reported that berenil, a minor groove-binding drug, can stabilize triple-helical DNA at low ionic strength, but destabilizes it at higher ionic strength. 3°

Binding Mode Study The binding mode of various triplex binding agents has been studied using techniques such as those described above. As these are all indirect methods, the binding mode is generally deduced by using several different techniques. Both intercalative and groove-binding mode have been observed for drug binding to triple-helical DNA. The binding mode is usually the same for duplex and triplex DNA, although small differences in the structure of the nucleic acid~trug complex are suggested by circular dichroism for some groove-binding agents, and by linear dichroism for intercalators. DAPI (14) is the only agent that may bind in the minor groove of duplex DNA and intercalate into triplex DNA. 48 NMR has not provided high-resolution structural information about drug-triplex DNA complexes. This may be due to the difficulty of obtaining a triple-helical structure with only one binding site for the ligand. The concentrations used for NMR are usually much higher than the binding affinity for secondary binding sites. Crystals of triplex DNA diffract only at low resolution. Attempts to crystallize triple helical DNA in the presence of triplex-stabilizing agents have not been successful so far.

Structural Specificity Compounds with high selectivity for triplex DNA versus duplex DNA have now been described. However, an in-depth investigation using, for example, the competition dialysis method described above may reveal that such compounds also bind to other types of multistranded nucleic acid structures, for example, tetraplex DNA. The search for compounds specific for triplex DNA is therefore a complicated task. Moreover, different types of triple helices (with a Hoogsteen or a reverse Hoogsteen configuration) can be targeted by a particular drug. The same ligand can stabilize both types of structures (as, e.g., BePI), but the more efficient ligands are not the same for both types of structure. It should be mentioned that triple-stranded structures containing both RNA and DNA strands may also be targeted by drugs. Specific stabilization of such structures has been reported. For example, in the presence of the minor groove binders berenil or DAPI, the poly(rA). [poly(dT)]2 and poly(dT). [poly(rA)]2 triple helices could be formed whereas they did not form in the absence of ligand. 32 Specific and efficient stabilization of RNA-containing triple-helical structures has also been observed with 48 Z. Xu, D. Pilch, A. Srinivasan, W. Olson, N. Geacintov, and K. Breslauer, Bioorg. Med. Chem. 5, 1137 (1997).

[16]


353

BePI (Escud6 et al., unpublished results, 1994). These results suggest that unpredicted triple-stranded structures involving both RNA and DNA may form within living cells in the presence of some drugs.

Binding Site Study Some qualitative information about the sequence preference of a drug may be gleaned by comparing the effects of the drug on different sequences, using thermal denaturation or footprinting experiments. For example, it has been shown that stabilization by BePI was reduced when T- A*T triplets were substituted by C. G*C + triplets. 16 Also, the naphthylquinoline derivative stabilizes the binding of a T6C 2 oligonucleotide, but not a Y2C 6 oligonucleotide.49 These results have been used to infer that the likely intercalation site for the pyrimidine motif is between T- A*T triplets. Indeed, intercalation does occur between T- A*T triplets, but some experiments have suggested that binding between T. A*T and C. G*C + triplets may be possible, especially at high drug concentration. 35 Few studies have addressed the binding of ligands to triple helices with purine-containing third strands. BePI has been shown to considerably enhance triple helix formation with a GT antiparallel third strand, leading to high stability under physiological conditions. 41 In another study, coralyne was shown to stabilize triple helices made of T. A*T triplets with the third strand in an antiparallel orientation. 35 All these results suggest that different sites within different structural contexts of triple-helical DNA may be targeted by drugs. These sites are still to be characterized. Footprinting methods allow the observation of specific cleavage patterns in the presence of triplex-binding agents (see Footprinting Methods, above). However, the results do not allow accurate mapping of the binding sites for these ligands. Thus, further investigations using established high-resolution structural methods are required to provide accurate information about binding sites.

Designing New Ligands for Triple-Helical DNA Most of the compounds that have been shown to stabilize triple-helical DNA efficiently are positively charged intercalators. The efficiency of triplex stabilization is the result of higher binding affinity for the triplex than for the duplex. This structural specificity can be achieved by the shape of a polyaromatic ring system, but also by electrostatic interactions and steric constraints. In particular, the nature and the position of side chains are likely to have an important influence on triplex-stabilizing properties. Rational design based on experimental studies has led to the synthesis of new compounds that bind to triplex DNA with high 49S. A. Cassidy,L. Strekowski,and K. R, Fox,NucleicAcqdsRes. 24, 4133 (1996).

354


[16]

affinity and specificity. 19 New strategies are emerging. For example, a dimer of the naphthylquinoline derivative 7 has been designed, synthesized, and studied9 Although the bisintercalation of this compound in triple helices was not conclusively demonstrated, it was shown to stabilize DNA triplexes much more effectively and at much lower concentrations than the monomer. T r i p l e x - B i n d i n g A g e n t s in M o l e c u l a r a n d C e l l u l a r Biology

Improving Antigene Strategy Initial efforts to find triplex-binding agents were motivated by the need to form stable triple helices under physiological conditions. Indeed, several triplex-binding agents have been shown to improve the stability of triple helices. Triplex-forming oligonucleotides can be used to compete with the binding of transcription factors, or to inhibit transcription initiation or elongation by RNA polymerase. Most studies of triplex-stabilizing agents were carried out with pyrimidine-rich third strand. Although efficient stabilization can be obtained, the stabilized triple helices are still pH dependent. Nevertheless, BePI has been shown to improve the TFO-mediated inhibition of in vitro transcription elongation in two different systems.16'51 Another study has suggested that the use of a triplex-stabilizing polyamine enhanced the TFO-mediated inhibition of the c-myc oncogene expression in MCF-7 breast cancer cells, but the evidence of a triplex-based mechanism was lacking in this case. 52 Further in vivo studies are required to establish the potential of triplex-binding agents as antigene enhancers. These studies will also have to deal with the potential toxicity of the compounds. Another interest in triplex-binding drugs stems from their potential to enhance triplex formation on sequences that have interruptions in the target oligopurine • oligopyrimidine sequence. The relative binding strength of different oligonucleotide substitutions could be affected, although the least destabilizing triplets have been shown to be the same in the absence and presence of ligand in two different studies) TM The combined use of a triplex-binding agent and a modified base analog, for example, 3-nitropyrrole, can promote discrimination for a specific base pair inversion, s4 It should be kept in mind that although the use of triplex-stabilizing agents will broaden the range of potential triplextargeted sites, they will also promote the formation of complexes at nontargeted sites. 50 M. Keppler, O. Zegrocka, L. Strekowski, and K. R. Fox, FEBS Lett. 447, 223 (1999). 51 C. Giovannangeli, L. Perrouault, C. Escud6, N. Thuong, and C. H61~ne, Biochemistry 35, 10539 (1996). 52 T. J. Thomas, C. A. Faaland, M. A. Gallo, and T. Thomas, Nucleic Acids Res. 23, 3594 (1995). 53 S. P. Chandler, L. Strekowski, W. D. Wilson, and K. R. Fox, Biochemistry 34, 7234 (1995). 54 S. Kukreti, J. S. Sun, D. Loakes, D. Brown, C. Nguyen, E. Bisagni, T. Garestier, and C. H61~ne, Nucleic Acids Res. 26, 2179 (1998).

[ 1 6]

DRUG INTERACTIONWITH TRIPLE-HELICALNAs

355

Triplex-Binding Agents and H-DNA It is known that H-DNA can form naturally within supercoiled double-stranded DNA, and formation of this structure has been demonstrated in living bacteria (see Mirkin and Frank-Kamenetskii6). Only one study has addressed the question of the effect of triplex-binding agents on intramolecular triple helices. 55 A mirror repeat sequence was inserted into a plasmid between the Escherichia coli/~-lactamase gene, which confers resistance to the antibiotic ampicillin, and its promoter. This plasmid also carries the resistance gene for the antibiotic tetracycline. It was shown that bacteria transformed with the plasmid containing the mirror repeat sequence grew more slowly than those transformed with the unmodified plasmid in the presence of ampicillin. Both types of cells could grow at the same rate in the presence of tetracycline. Thus, transcription of the/%lactamase gene was apparently inhibited by the mirror repeat sequence, whereas the plasmid could replicate normally. The addition of the triplex-stabilizing agent BePI was shown to further reduce the growth rate in the presence of ampicillin selectively. Chemical probes (chloroacetaldehyde) showed that BePI stabilized the H-DNA structure induced by supercoiling in the plasmid. These experiments suggest that mirror sequences that are able to form H-DNA structures could act as gene regulators, and that regulation of these genes could be influenced by triplex-stabilizing agents. Many sequences that are likely to be able to adopt an H-DNA structure exist in eukaryotic genomes, and many of them are located in gene promoter regions. If these structures do play a functional role, drugs binding to them may serve as tools for investigating this role and may even constitute potential drugs able to interfere with specific biological processes. This has also been postulated in the case of drugs binding to tetraplex DNA. ~6 Some of the triplex-stabilizing agents described above, for example, BePI and coralyne, have cytotoxic properties. Stabilization of intramolecular triple-helical DNA might represent a new mechanism of action for these drugs. Other substances, for example, mithramycin, have been shown to destabilize intermolecular triplexes by binding strongly to duplex DNA. It has been suggested that this antibiotic may inhibit intramolecular triplex formation within a sequence in the c-myc gene. 57 Irreversible Modifications: Triplex-Cleaving Agents Agents that could inflict irreversible damage on triplex DNA would be helpful first as a structural probe for the complexes formed by these agents and the nucleic

55G. Duval-Valentin,T. Debizemont,M. Takasugi,J. L. Mergny,E. Bisagni, and C. H61~ne,J. Mol. Biol. 247, 847 (1995). 56j. L. Mergnyand C. H61bne,Nat. Med. 4, 1366 (1998). 57N. Vigneswaran,C. A. Mayfield,B. Rodu, R. James, H. G. Kim, and D, M. Miller, Biochemistry 35, 1106 (1996).

356


[16]

acids, and second as a probe for detecting the presence and studying the role of triplex DNA within a biological context. In the presence of iron, a reducing agent, and molecular oxygen, EDTA-Fe H can generate hydroxyl radicals through the Fenton reaction. The covalent attachment of EDTA to a TFO has been used to induce specific cleavage of the targeted double-stranded DNA. 2 EDTA has been covalently attached to various triplex intercalators in our laboratory, namely BePI (1), BgPI (2), and BQQ (6). 58,59 Experiments have been conducted with different double-stranded DNAs containing a putative target for triplex formation, in the presence and absence of third strand. It was shown that strong DNA cleavage was promoted by BgPI-EDTA on both strands within the target site, whereas with BePI-EDTA a slightly enhanced cleavage of the oligopyrimidine strand and a footprint in the oligopurine strand were observed. DNA cleavage was observed all along the triplex sequence, with peaks corresponding to more intense cleavage centered around T-A*T stretches. This suggests that there are various sites for intercalation within the triplex region and that these compounds may exhibit some preferential binding to T. A'T-rich sequences. BQQ-EDTA exhibited the same behavior as BgPI-EDTA, except that it was much more specific than BgPI-EDTA, that is, nonspecific cleavage of double-stranded DNA was reduced. These results are in agreement with results of melting temperature experiments. 19 The observed difference in the pattern and efficiency of cleavage can be explained by the structure of the intercalation complex. Efficient cleavage of both strands is observed when the cleaving moiety is located in the Watson-Crick groove, as with BgPI-EDTA and BQQ-EDTA, whereas with BePI-EDTA hydroxyl radicals must diffuse from the Watson-Hoogsteen groove, an inefficient process that results in some weak cleavage of the more accessible strand, the oligopyrimidine Watson strand (see Fig. 3). These results therefore furnish experimental verification for the model that was proposed on the basis of QSAR and molecular modeling studies. 45 The selectivity of the BQQ-EDTA conjugate for a triplex structure was sufficiently high to induce oligonucleotide-directed DNA cleavage at a single site on a 2718-bp plasmid DNA. Experiments aimed at using this compound to cleave intramolecular triplex DNA in a supercoiled plasmid are under way. More recently, a dibenzophenanthroline derivative, 10, which had been shown to stabilize triple helices,t8 was shown to induce cleavage close to the site of binding of a TFO on irradiation at 360 nm. This cleavage occurred only in the presence of the TFO under conditions favorable for triplex formation. The mechanism of this triplex-mediated photoinduced cleavage is under investigation. 6°

58C. Marchand,C. H. Nguyen,B. Ward,J.-S. Sun,E. Bisagni,T. Garestier, and C. H~16ne,Chem. Eur J. 6, 1559(2000). 59R. Zain, C. Marchand,J.-S. Sun,C. H. Nguyen,E. Bisagni,T. Garestier, and C. H616ne,Chem. Biol. 6, 771 (1999). 6oD. Perrin, et al., in preparation (2001).

[16]

357

DRUG INTERACTION WITH TRIPLE-HELICAL N A s

k

k

FIG. 3. A scheme for the cleavage of triple-helical DNA by EDTA-ligand conjugates. The intercalating ligand is shown in gray, with the EDTA-Fe 11moiety represented by a star shape. It generates hydroxyl radicals, which can diffuse over short distances (dashed arrows). Small dark arrows point to the CI' and C4' atoms of deoxyribose. The arrows are located on the side from which the hydrogen atoms (HI' and H4 I) are accessible. Left: The EDTA-Fe II moiety is located in the Watson-Hoogsteen groove, and the HI' and H4' atoms are not readily accessible to hydroxyl radicals, especially that of the Crick strand. Right: The EDTA-Fe n moiety is located in the Watson-Crick groove, and the HI' and H4 f atoms of both the Watson and Crick strands are readily accessible to hydroxyl radicals.

Conclusion Ligands that are selective for triplex DNA can be used to promote triple helix formation, but their interest is much broader as they represent a unique class of structure-specific ligands that could be used as tools to study the biological relevance of triple-stranded structures, or potentially as drugs acting on structurerelated gene regulation. Both high-resolution structural information and thermodynamic as well as kinetic studies still need to be pursued to yield a better understanding of how drugs bind specifically to triplex DNA. Nevertheless, the progress that has been made in characterizing triplex-ligand complexes has led to the design of some specific and efficient triplex-targeted drugs.

Acknowledgments The authors thank all the scientists of the Laboratoire de Biophysique who contributed to the study of triplex-stabilizing agents, and gratefully acknowledge the chemists at the Institut Curie and the Coll6ge de France who synthesized some of the compounds described in this chapter.

358

C H E M I C A L AND M O L E C U L A R B I O L O G I C A L A P P R O A C H E S

[ 17]

[ 17] M e a s u r e m e n t of Covalent D r u g - D N A I n t e r a c t i o n s at the Nucleotide Level in Cells at P h a r m a c o l o g i c a l l y Relevant Doses By CLAIREJ. MCGURK, PETERJ. MCHUGH, MICHAELJ. TILBY, KEITH A. GRIMALDI,and JOHN A. HARTLEY Introduction Many clinically important anticancer drugs, such as those from the alkylating agent and platinum classes, act by forming covalent adducts with DNA. The majority of these adducts are formed with a base specificity and often show limited sequence specificity. For example, cisplatin [cis-diamminedichloroplatinum II] binds mainly the N7 positions of guanine nucleotides in the major groove of DNA and forms mostly intrastrand cross-links at GG and AG. More recently, novel drugs that show a high degree of sequence-selective binding and alkylation in the minor groove of DNA have emerged based on natural product structures. Attempts to rationally design agents that will target DNA damage to unique sequences are ongoing. Sequence specificity has mainly been determined by examining the lesions produced on isolated DNA. Although important, such analyses are limited. Nucleotidelevel mapping of adducts in cells indicates whether binding to a particular sequence is conserved, because sequence context, methylation, and cellular factors such as nucleosomes and DNA-binding proteins influence drug-DNA interactions. Ligation-mediated polymerase chain reaction (LM-PCR) is the sensitive technique that was first used to map nucleotide-level DNA damage in single-copy genes in mammalian cells after exposure to UV. 1'2 Lesions are cleaved with enzymes or chemical reagents to create strand breaks to which a double-stranded oligonucleotide linker is ligated. The molecules, of varying lengths determined by the location of the lesion, are exponentially amplified, separated on sequencing gels, and detected by hybridization with a fragment-specific probe. For many drug-DNA adducts, however, there are no effective cleavage agents available. Single-strand ligation PCR (sslig-PCR) 3-6 was developed, which relies on the blockage of Taq DNA polymerase at covalent lesions and therefore can be I R R. Mueller and B. Wold, Science 246, 780 (1989). 2 G. P. Pfeifer and A. D. Riggs, Genes Dev. 5, 1102 (1991). 3 K. A. Grimaldi, S. R. McAdam, R. L. Souhami, and J. A. Hartley, Nucleic Acids Res. 22, 2311 (1994). 4 K. A. Grimaldi, S. R. McAdam, and J. A. Hartley, Single-strand ligation PCR for detection of DNA adducts, in "Technologies for Detection of DNA Damage and Mutations" (G. R Pfeifer, ed.). Plenum Press, New York, 1996.

METHODS IN ENZYMOLOGY.VOL. 340

Copyright,~) 2001 by AcademicPress All rightsof reproductionin any form reser,,red. 0076 6879/(X1$35.00

[ 1 7]

MEASUREMENT OF COVALENTDRUG-DNA INTERACTIONS

359

used to detect most d r u g - D N A adducts. An initial linear PCR step samples each target molecule many times. Ligation of a single-stranded oligonucleotide with T4 RNA ligase facilitates exponential PCR, followed by a third round of linear PCR with a labeled primer. Terminal transferase-dependent PCR (TD-PCR), developed later, is a related technique that has been used to detect cyclobutane pyrimidine dimers 7 and mycotoxin aflatoxin B l addu cts.8 Like sslig-PCR, it involves initial linear PCR and relies on a DNA polymerase blockage at the sites of lesions. However, an additional step utilizes terminal deoxynucleotidyltransferase (TdT) to add ribo-guanines onto the 3' termini. In combination with a double-stranded linker possessing a cytosine overhang, this provides an efficient substrate for T4 DNA ligase, a more efficient enzyme than T4 RNA ligase in this context. The ligated products are amplified and detected in the same way as for LM-PCR. Currently the major drawback to studying nucleotide level damage is the relatively high doses of damaging agent required to generate detectable levels of lesions. The study of drug-DNA interactions at these doses, although generating important comparative information, may not reflect the pharmacological situation. F o r example, the distribution of lesions may be distorted, many biological responses may be impaired or saturated, or the cells may initiate apoptosis. In particular, studies of the repair of lesions at the nucleotide level are difficult at such high drug doses. Antibodies to O6-ethylguanine, 9'1° thymine glycols, I1 and benzo[a]pyrenediol epoxide (BPDE)-guanine adductsl2,13 have been used successfully to purify fragments of DNA for use in PCR, and have provided the sensitivity to detect adducts at the gene level at relatively low doses. In this chapter we describe how the specificity of DNA damage antibodies can be combined with the sensitivity of genomic sequencing methods to study the lesions produced by DNA-damaging drugs at pharmacologically relevant doses at the nucleotide level in single-copy genes in mammalian cells. A modified method adapted from both the TD-PCR and sslig-PCR procedures is described (see Fig. 1), which can be used to study any covalent DNA lesions that

5 K. A. Grimaldi and J. A. Hartley, Methods Mol. Biol. 90, 157 (1997). 6 K. A. Grimaldi, S. R. McAdam,and J. A. Hartley,Methods Mol. Biol. 113, 241 (1999). 7 j. Komuraand A. D. Riggs, Nucleic Acids Res. 26, 1807 (1998). 8 M. F. Denissenko, T. B. Koudriakova, L. Smith, T. R. O'Connor, A. D. Riggs. and G. P. Pfeifer, Oncogene 17, 3007 (1998). 9 K. Hochleitner,J. Thomale, A. Y. Nikitin, and M. E Rajewsky,Nucleic Acids Res. 19, 4467 ( 1991). l0 j. Thomale, K. Hochleitner, and M. E Rajewsky,J. Biol. Chem. 269, 1681 (1994). I l S. A. Leadon and D. A. Lawrence,J. Biol. Chem. 267, 23175 (1992). 12M. E Denissenko, S. Venkatachalam,E. E Yamasaki, and A. A. Wani,Nucleic Acids Res. 22, 2351 (1994). J3 M. E Denissenko, S. Venkatachalam,Y. Ma, and A. A. Wani, Biotechniques 21, 187 (1996).

360


[ ] 7]

Damaged Genomic DNA, Cut with Appropriate Restriction Enzyme(s) /

Linear PCR using [ Biotinylated Primer I.B

Capture on Streptavidin Dynabeads

......

~O

Ribonucleotide Tailing by Terminal DeoxynueleotidylTransferase

rGrGrG Ligation of Double-Stranded Linker [ by T4 DNA Ligase [

C C C[ [

cccl

IrGrGrG Exponential PCR with Upper Oligo Primer [ and Primer 2

CC(~ . . . . . . . . . . . .

[rGrGrG

~ 2 PCR with Radiolabeled Primer 3

I

cccl

............

Run DNA on Sequencing Gel

C C C[

[ 17]

MEASUREMENTOF COVALENTDRUG-DNA INTERACTIONS

361

prevent the progression of a DNA polymerase. An antibody-dependent purification method is also described that can be used in combination with the modified PCR method to study any lesions to which an antibody is available or can be raised. This combined method significantly improves the sensitivity and quality of the results. As an example of the method, the sequence specificity of binding of cisplatin to a single-copy gene sequence in intact cells is illustrated. Method

DNA-Damaging Agent Treatments PCRs are carried out on damaged DNA, undamaged control DNA, and also Maxam and Gilbert-treated DNA to provide a sequence reference. During MaxamGilbert sequencing, 143/zg of unlabeled DNA is chemically cleaved at base-specific sites and subsequently processed by the PCR technique to produce accurate sequencing lanes that will directly correspond to the lanes containing DNA that has been damaged. The method is applicable to any type of cell, including freshly isolated lymphocytes or cell preparations from solid tissue. For comparison with DNA extracted from drug-exposed intact cells, isolated genomic DNA is treated with the drug and subjected to the same analyses. Treatment of Isolated DNA. Genomic DNA is isolated as described below. The drug incubation time with naked DNA will depend on the reactivity of the drug. For many drugs, 1 hr is sufficient. However, when using cisplatin (David Bull Laboratories, Mulgrave, Australia) we have found that increasing incubation times up to 12 hr generates increasing amounts of damage to template DNA. 1. Incubate 3 # g of DNA in 1.5-ml tubes with DNA-damaging agent at 37 ° for appropriate times in I x Teoa [10x Teoa is 250 mM triethanolamine (pH 7.2), 10 mM EDTA] in a total volume of 50 #1. Teoa should be stored at 4 °. 14 M. Maxam and W. Gilbert, Methods"Enzymol. 65, 499 (1980).

FIG. 1. Outline of the PCR procedure. An intrastrand cisplatin adduct in the starting DNA is indicated by a small triangle. Ten cycles of linear amplification of damaged DNA with a 5'-biotinylated primer produce a family of single-stranded molecules, the Y ends of which are defined by the positions at which Taq polymerase has stopped at DNA lesions. The products are captured on streptavidin-coated magnetic Dynabeads and washed to remove any excess genomic DNA. Terminal deoxynucleotidyltransferase adds three ribo-G molecules to the 3' ends of the amplified DNA fragments. A double-stranded linker with a complementary Y overhang of three C's is ligated to the ribo-G tail, facilitating exponential amplification. A final round of PCR with a radiolabeled nested primer enables the products to be visualized when run on a sequencing gel.

362


[ 17]

2. Add 50/zl of 0.6 M sodium acetate "drug stop" solution (pH 5.2) and 3 volumes of 95% (v/v) ethanol (-20°). Vortex. 3. Place the samples in a dry ice bath for 10 min and centrifuge for 15 min at 13,000g at 4 °. 4. Wash the DNA pellet with 1 ml of 70% (v/v) ethanol (room temperature) by gently inverting the tube, centrifuging for 5 min at 11,600 g at room temperature, and removing the supernatant. Repeat the wash and dry the DNA pellet under vacuum. 5. Resuspend the DNA in 10/xl of distilled water.

Treatment of Cells. The time of incubation of drug with cells will again depend on the reactivity of the drug. For example, with cisplatin lesions can be detected within a few hours, but with carboplatin 24 hr or more is needed because of its slower reactivity. Some drugs, including cisplatin, are able to bind the proteins present in fetal calf serum. Therefore, when using cells in culture it is advisable at low doses to carry out the drug incubations in serum-free medium to achieve optimal drug binding to DNA. For long incubation times, however, it may be necessary to include fetal calf serum in the drug incubation to ensure cell survival. Treatment of Suspension Culture Cells 1. Count the cells and resuspend at a density of 2 x 106/ml in serum-free tissue culture medium. 2. Add the required amount of drug (diluted in serum-free tissue culture medium) to the wells of six-well fiat-bottomed tissue culture plates. 3. Add tissue culture medium to make the volume up to 1.5 ml. 4. Add 1.5 ml of cell suspension (3 x 1 0 6 cells), mix well, and incubate at 37 ° for the appropriate time. 5. Transfer the cells to 2 x 1.5-ml tubes and centrifuge at 3000g for 5 min at 4 °. 6. Wash out the wells with 1 ml of serum-free tissue culture medium. 7. Remove the supernatant from 2 × 1.5-ml tubes. Resuspend and combine both cell pellets in 1 ml of medium from the wash (step 6) into one 1.5-ml tube. Discard the other 1.5-ml tube. 8. Centrifuge at 3000g for 5 min at 4 °. 9. Remove the supernatant. The cell pellet may be stored at - 2 0 ° at this point until DNA isolation.

Treatment of Adherent Cells 1. Grow cells almost to confluence in the wells of six-well flat-bottomed tissue culture plates.

[17]


363

2. Dilute the required amount of drug in 2 ml of serum-free tissue culture medium and add to the cells. 3. Incubate at 37 ° for the appropriate time. 4. Remove the drug medium and gently wash the cells three times with 1 ml of fresh tissue culture medium. [Some cells may become detached from the wells. If so, place into 1.5-ml tubes and centrifuge at 3000g for 5 min at 4 °, remove the supernatant, and add to cells harvested from wells by trypsinization (step 5).] 5. Harvest the cells by trypsinization, centrifuge at 3000g for 5 min at 4 °, and remove the supernatant. The cell pellet may be stored at - 2 0 ° at this point until DNA isolation.

Preparation of DNA 1. Isolate DNA by conventional procedures. The isolated DNA must be of a sufficiently high quality for use in PCR. Contaminating proteins may inhibit the progression of Taq polymerase and generate excess background; therefore additional phenol-chloroform and proteinase K steps may be implemented to increase the quality of the DNA) 5 We have used a kit (Wizard Genomic DNA Purification Kit; Promega, Southampton, UK) that generates a relatively high yield of DNA, 15-30/.zg from 3 × 106 cells. We have obtained equivalent results from DNA by using this kit and DNA that has been further purified with phenol--chloroform and proteinase K. However, these extra protein purification steps may be necessary with some DNA isolation procedures and particular cell lines. An RNase incubation step should be included in the protocol. RNase is necessary, especially if DNA is to be measured spectrophotometrically, as excess RNA could contribute to the total measurement of DNA in the sample and thus could lead to an overestimate of the DNA, creating inconsistency between samples. This is particularly important when quantitating lanes and making comparisons, such as in repair experiments. 2. Digest the DNA with a restriction enzyme that cuts at a site between 100 and a few hundred bases downstream of the binding site of the biotinylated primer that is to be used to create a natural stop site in the first PCR step. 3. Check the DNA concentration fluorimetrically. The measurement of DNA by fluorimeter is extremely accurate, especially when measuring small quantities of DNA. 16 For analyzing single-copy genes we have found that 3 #g of DNA produces the optimal signal. If insufficient material is available (e.g., from clinical samples), 15j. Sambrook,E. E Fritsch, and T. Maniatis, "MolecularCloning: A LaboratoryManual," 2nd Ed. Cold Spring HarborLaboratory,Cold Spring Harbor,New York, 1989. t6 D. Kowalski,Anal.Biochem.93, 346 (1979).

364


[1 7]

then it is possible to use less DNA and add cycles during the second-round exponential PCR step to compensate. However, nanogram quantities of DNA can be used if studying multiple-copy sequences, such as mitochondrial DNA or DNA from organisms with smaller genomes (e.g., yeast), and still achieve equivalent results.

Polymerase Chain Reaction The following PCR procedure, outlined in Fig. 1, is adapted from both the terminal transferase-dependent PCR (TD-PCR) procedure 7 and the single-strand ligation PCR (sslig-PCR) technique. 3-6 Biotinylated primers are used in the linear PCR step to facilitate purification of amplified DNA fragments from genomic DNA; streptavidin-Dynabeads (Dynal Biotech, Oslo, Norway) prevent nonspecific amplification in further PCR steps. Lengthy ethanol precipitation steps are replaced by simple washes of Dynabeads, thus reducing time required and the possibility of intersample variation. Time-consuming electroblotting and hybridization are replaced with a third round of PCR using a labeled, nested primer. We have found that Taq polymerase produces more discrete bands compared with Vent (exo-) (New England BioLabs, Hitchin, UK). Vent produces no amplification product because of interference with biotin. MgCI2 concentrations for linear and exponential PCR steps must be determined for each set of primers individually; optimization experiments are usually carried out on drug-treated isolated DNA. Gelatin is included in the first-round "linear" PCR to improve specificity, which is important when using a single primer to target a single-copy gene. The conditions shown here are for analyzing the nontranscribed strand of a coding region of the human c-jun gene using primers (JT1, JT2, and JT3) as previously describedJ 7 Oligonucleotides. Oligonucleotides are obtained from MWG (Ebersberg, Germany). All oligonucleotides are stored in small aliquots at - 2 0 °. The double-stranded oligonucleotide linker used is linker y as previously described. 7 Lower Oligo: 5'-AATTCAGATCTCCCGGGTCACCGC This oligonucleotide should be gel or high-performance liquid chromatography (HPLC) purified. It must also be 5' phosphorylated and possess a T-terminal amine group to block self-ligation. The amine should be incorporated at synthesis, but phosphorylation is more complete if carried out after purification. Upper Oligo: 5'-GCGGTGACCCGGGAGATCTGAATTCCC 17 y. Tu, S. Tornaletti, and G. E Pfeifer, E M B O J. 15, 675 (1996).

[ 17]


365

This oligonucleotide is complementary to Lower Oligo except for an additional three C's at the 3' terminus. It should also be gel or HPLC purified. Three nested oligonucleotide primers, each 16 to 28 nucleotides long, should be used per reaction. The sequences of the oligonucleotides will depend on the region of interest to be studied. The first oligonucleotide primer should have biotin incorporated at the 5' end (this modification should be incorporated at synthesis). The Tm of the oligonucleotide primers used is calculated according to the manufacturer equation (MWG): Tm

=

69.3 + (0.41 × %GC) - (650/n)

where %GC is the percent GC content and n is the length of the primer. The Tm of the first oligonucleotide primer should usually be between 50 and 60 °. The Tm of oligonucleotide primers 2 and 3 usually falls between 60 and 70 ° and should ideally be as close to the Tm of Upper Oligo as possible (71 °). The sequences of the primers used in the following example are outlined below and are as previously described 17: JT-1.B: 5'-biotin-ACCGGTGCGAGCGAAG

JT-2: 5'-CGTCCTTCTTCTCTTGCGTGGCTCT JT-3: 5'-GGCTCTCCGCCGCCTTCTGGTCTTT

5' Phosphorylation of Lower Oligonucleotide of Linker. Oligonucleotides are phosphorylated with T4 polynucleotide kinase, using kits (GIBCO-BRL, Paisley, UK) with forward reaction buffer. 1. Prepare 200-pmol/#l (200/zM) stocks of upper and lower oligonucleotides with distilled water. 2. To phosphorylate the 5' terminus of the lower oligonucleotide, incubate in a 1.5-ml tube: Lower oligonucleotide (200 pmol]/zl) 11.1 #1 1x Forward reaction buffer [as provided with enzyme; 20/zl 5x is 350 mM Tris-HC1 (pH 7.6), 50 mM MgC12, 500 mM KC1, 5 mM 2-mercaptoethanol] Distilled water 62.9 #1 ATP (100 raM) 1 #1 T4 polynucleotide kinase (10 units/#l) 5 #1 3. Incubate at 37 ° for 2 hr and then heat inactivate the kinase at 65 ° for 20 min. 4. To the phosphorylated lower primers add 11.1 #1 of the upper oligonucleotide (200 pmol/#l). 5. Heat this mixture to 95 ° for 3 rain and gradually cool to 4 ° by placing the sample in a 70 ° removable hot block for 1 min and then placing the hot block into

366


[17]

a refrigerator. When cool, the linker is stored at - 2 0 ° until ready for use in the ligation reaction.

First-Round Linear Polymerase Chain Reaction. The single biotinylated primer used in the first round of PCR determines in which region and on what strand damage is measured (i.e., a primer complementary to the transcribed strand detects damage on the transcribed strand). All the reagents for the PCR steps are stored at - 2 0 ° (except MgCI2 and gelatin, which are stored at 4°), defrosted if necessary, and kept on ice until needed except Taq polymerase, which is kept at - 2 0 ° until needed. Amplifications are carried out in a thermal cycler with a heated lid (MJ PTC -100: MJ Research, Watertown, MA) so that it is not necessary to use mineral oil. 1. Place 3 ~g of template DNA (or beads-antibody-DNA mixture if using immunoprecipitated DNA) in a 0.5-ml tube. Bring the volume to 10 #1 with distilled water. 2. Prepare a PCR cocktail mix in a 1.5-ml tube, using the volumes given below after multiplying by the number of samples. 10x Reaction buffer [as provided with enzyme; 10x is 4.0/~1 100 mM Tris-HCl (pH 9.0 at 25°), 500 mM KC1, and 1% (v]v) Triton X- 100] MgCI2 (25 mM) 1.6 #1 Gelatin (0.2%, w/v) 4.0 #1 Distilled water 10.4/zl Mixture containing a 2.5 mM concentration of each dNTP 4.0 #1 (Amersham Pharmacia Biotech, Little Chalfont, UK) Oligonucleotide primer 1 (0.6 pmol/#l) 1.0/~1 Place 25.0 #1 of the PCR cocktail mix into each of the sample tubes; mix well. 3. In a separate 1.5-ml tube, dilute Taq DNA polymerase (5 units/ttl; Promega) to 0.2 units//~l in a total volume of 5/~1 per sample, using distilled water. Add 5 #1 (1 unit) of diluted Taq DNA polymerase to each 0.5-ml tube; tap gently to mix. 4. Denature at 94 ° for 5 min, and then cycle 10 times at 94 ° for 1 min, 56 ° for 3 min, and 74 ~ for 2 min. For complete extension of all products, incubate at 74 ° for 5 rain. Denature at 95 ° for 2 min and then cool to 4 °.

Capture of Biotinylated Polymerase Chain Reaction Products 1. Transfer streptavidin-coated magnetic beads (Dynabeads M-280; Dynal Biotech) to 1.5-ml tubes. Use 5/zl (10/zg//zi) per sample plus 5 #l extra. 2. Sediment the Dynabeads in a magnetic rack (e.g., Dynal MPC-E) for 30 sec and remove the supematant. 3. Wash the Dynabeads twice by removing the tubes from the magnetic rack, resuspending the Dynabeads in 200 #1 of 1 x WBB [1 x washing and binding

[17]

MEASUREMENT OF COVALENTDRUG--DNA INTERACTIONS

367

buffer; 10 mM Tris-HC1 (pH 7.7), 1 mM EDTA, 2 M NaC1], replacing the tubes in the magnetic rack to sediment the beads, removing the supernatant, and repeating. 4. Resuspend the Dynabeads in 40 #1 of 1 x WBB per sample. 5. Transfer 40-#1 aliquots to fresh 1.5-ml tubes. Sediment the Dynabeads and remove the supernatant from each tube. Remove the 1.5-ml tubes from the magnetic rack. 6. To each 0.5-ml tube containing 40 #1 of products from the first-round PCR add 10 #1 of 5 x WBB [5 x washing and binding buffer; 50 mM Tris-HC1 (pH 7.7), 5 mM EDTA, 10 M NaC1]. Transfer this 50-#1 mixture to each of the 1.5-ml tubes containing prepared Dynabeads. 7. Incubate at 37 ° for 30 rain, agitating occasionally to resuspend the Dynabeads. 8. Wash the Dynabeads three times with 200 #1 ofTE [ 10 mM Tris-HC1 (pH 7.6), 1 mM EDTA] (see step 3). 9. ResuspendtheDynabeadsin50/zlofdistilledwaterandcentrifugeat 11,600g for 10 sec.

TerminalDeoxynucleotidyltransferaseRibo-Tailing. Ribo-G's are added to the 3' ends of the single-stranded products from linear PCR. The majority of the ends will possess three ribo-G's. 7 All the reagents for ribo-tailing and ligation are stored at - 2 0 ° [except polyethylene glycol 8000 (PEG), which is stored at 4°], defrosted if necessary, and kept on ice until needed except for TdT and DNA ligase, which are kept at - 2 0 ° until needed. 1. Place 1.5-ml tubes containing Dynabeads bound to biotinylated PCR products into a magnetic rack and remove the supernatant. 2. Resuspend the Dynabeads in 10/~1 of 0.1 x TE [1 mM Tris-HC1 (pH 7.6), 0.1 mM EDTA]. 3. Prepare a ribo-tail mix in a 1.5-ml tube, using the volumes given below after multiplying by the number of samples. 5 × TdT buffer [as provided with enzyme; 0.5 M potassium 4.00/zl cacodylate (pH 7.2), 10 mM COC12, and 1 mM dithiothreito! (DTT)] rGTP (10 mM; Promega) 4.00 #1 Distilled water 1.33 #1 TdT (15 units/#l; GIBCO-BRL) 0.67 #1 4. Place 10.0 #1 of the ribo-tail mix into the tubes with the Dynabeads and pipette to mix. 5. Incubate at 37 ° for 15 min with occasional agitation. 6. Add 180 #1 of TE [ 10 mM Tris-HC1 (pH 7.6), 1 mM EDTA], place the tubes in a magnetic rack, and remove the supernatant. Wash the Dynabeads again with 200 #I of TE (twice).

368


[17]

7. Resuspend the Dynabeads in 15/zl of 0.1 × TE [ 1 mM Tris-HCl (pH 7.6), 0.1 mM EDTA].

LigationofDouble-StrandedLinker. The families of single-stranded molecules that possess rGTP tails are ligated to the double-stranded linker. The doublestranded linker has a 3-C single-stranded overhang that is complementary to the three ribo-G tails. The ligation reaction contains PEG, which concentrates the reagents and keeps the Dynabeads partially suspended, as is important when incubating for long periods. 1. Prepare a ligation mix in a 1.5-ml tube, using the volumes given below after multiplying by the number of samples. Ligation buffer [622 mM Tris-HCl (pH 7.6), 126 mM MgC12, 2.7 #1 126 mM DTT, and 12.6 mM ATP] Prepared double-stranded linker (20/zM; see above) 3.0/zl PEG (50%, w/v) 9.1 #1 T4 DNA ligase (20 units//zl; Promega) 0.2/zl 2. Place 15.0 #1 of the ligation mix into the tubes with the Dynabeads from the ribo-tailing reaction (using a cut tip, as the mixture is extremely viscous) and pipette to mix. 3. Incubate at 17° overnight. 4. Add 170/zl of TE [ 10 mM Tris-HC1 (pH 7.6), 1 mM EDTA], place the tubes in a magnetic rack, and remove the supernatant. Wash the Dynabeads again with 200/zl of TE (twice). 5. Resuspend the Dynabeads in 40 #1 of distilled water for PCR. Transfer to 0.5-ml tubes.

Second-RoundExponentialPolymerase Chain Reaction. The procedure is carried out with template DNA bound to Dynabeads. Ten picomoles of second-round nested region-specific primer 2 and Upper Oligo primers are required per sample. For each new primer set, the number of cycles necessary during the exponential step will need to be determined empirically and almost always falls between 16 and 22 cycles. It is important that the number of cycles falls in the exponential range, especially when quantifying bands. 1. Prepare a PCR cocktail mix in a 1.5-ml tube, using the volumes given below after multiplying by the number of samples. 10x Reaction buffer [as provided with enzyme; 10x is 10.0/zl 100 mM Tris-HCl (pH 9.0 at 25°), 500 mM KCI, and 1% (v/v) Triton X-100] MgC12 (25 mM) 8.0/zl Distilled water 20.0 #1

[ 1 7]


369

Mixture containing 2.5 mM each dNTP (Amersham 10.0/zl Pharmacia Biotech) Oligonucleotide primer 2 (10.0 pmol/#l) 1.0/~1 Upper Oligo linker primer (10.0 pmol/#l) 1.0 #1 2. Place 50.0 #1 of the PCR cocktail mix into the 0.5-ml tubes with the Dynabeads from the ligation reaction; mix well. 3. In a separate 1.5-ml tube, dilute Taq DNA polymerase (5 units/#l; Promega) to 0.2 units/#l in a total volume of 10/zl per sample, using distilled water. Add 10 #1 (2 units) of diluted Taq DNA polymerase to each 0.5-ml tube; tap gently to mix. 4. Denature at 94 ° for 5 min, and then cycle 22 times at 94 ° for 1 min, 68 ° for 2 min, and 74 ° for 1 min plus a 1-sec extension per cycle. For complete extension of all products, incubate at 74 ° for 5 min.

5' End Labeling of Primer 3for Third-Round Polymerase Chain Reaction. Five picomoles of oligonucleotides is required per sample for third-round PCR. Primer 3 is phosphorylated with T4 polynucleotide kinase, using kits (GIBCO-BRL) with forward reaction buffer. Once labeled the primer is stored at - 2 0 ° until ready for use in the third-round PCR. 1. To label the 5' terminus of the third-round primer 3, incubate in a 1.5-ml tube: Oligonucleotide primer 3 (10 pmol//zl), per sample (plus one 0.5 #1 extra) 5 x Forward reaction buffer [as provided with enzyme; 5 x is 6 #1 350 mM Tris-HC1 (pH 7.6), 50 mM MgCI~, 500 mM KC1, 5 mM 2-mercaptoethanol] [F-32p]ATP (10 #Ci/#l; Amersham Pharmacia Biotech) 1 #1 T4 polynucleotide kinase (10 units//zl) 1 #1 Bring the volume to 30 #1 with distilled water. 2. Incubate at 37 ° for 30 min. 3. Add distilled water to bring the total volume to 5 #1 per sample (plus one extra). 4. Purify labeled primer from unincorporated nucleotide by centrifuging through a Bio-Spin-6 spin column (Bio-Rad, Hemel Hempstead, UK).

Third-Round Labeling Polymerase Chain Reaction. This is carried out immediately after the second-round exponential PCR, using the labeled nested primer 3. 1. Prepare a PCR cocktail mix in a 1.5-ml tube, using the volumes given below after multiplying by the number of samples.

370


[ 17]

10x Reaction buffer [as provided with enzyme; 10x is 1.0 #l 100 mM Tris-HC1 (pH 9.0 at 25°), 500 mM KC1, and 1% (v/v) Triton X-100] MgC12 (25 mM) 1.0 #1 Distilled water 1.8 #1 Mixture containing 2.5 mM each dNTP (Amersham 1.0 #1 Pharmacia Biotech) Labeled oligonucleotide primer 3 (1.0 pmol//zl) 5.0 #1 Taq DNA polymerase (5 units//zl; Promega) 0.2 #1 Tap gently to mix. 2. Place 10/zl of the PCR cocktail mix into the 0.5-ml tubes from the secondround exponential PCR. 3. Cycle five times at 94 ° for 1 min, 68 ° for 2 min, and 74 ° for 1 min plus a 1-sec extension per cycle. For complete extension of all products incubate at 74 ° for 5 min. 4. Centrifuge the 0.5-ml tubes for 10 sec at 11,600g and transfer 100 #1 of supernatant (not Dynabeads) to fresh 1.5-ml tubes. Add 100/zl of distilled water to 0.5-ml tubes to wash the Dynabeads, respin, and transfer 100 #1 of supernatant (not Dynabeads) to the 1,5-ml tubes. 5. Add 20 #1 of 3 M sodium acetate (pH 5.2) and precipitate the DNA with 3 volumes of 95% (v/v) ethanol (-20°). Vortex, and then place the samples in a dry ice bath for 10 min and centrifuge for t5 min at 13,000g at 4 °. 6. Wash the DNA pellet with 200/xl of 70% (v/v) ethanol (room temperature) by gently inverting the tube, centrifuging for 5 min at 11,600g at room temperature, and removing the supernatant. Repeat the wash and dry the DNA pellet under vacuum. 7. Resuspend the DNA in 5 /xl of sequencing gel-loading buffer [96% (v/v) formamide (deionized), 20 mM EDTA, 0.03% (w/v) xylene cyanol, 0.03% (w/v) bromphenol blue]. It is important that the samples be completely and uniformly resuspended. To achieve this, vortex each sample vigorously for at least 10 sec and centrifuge at 11,600g for 10 sec at 4 ° to collect the liquid. Store at - 2 0 °.

Gel Electrophoresis 1. Separate the radiolabeled amplified fragments in a 50 cm x 21 cm x 0.4 mm, 6% (w/v) denaturing polyacrylamide sequencing gel [National Diagnostics, Hull, UK; 5.7% (w/v) acrylamide, 0.3% (w/v) bisacrylamide, 8.3 M urea, 0.1 M Trisborate (pH 8.3), 2 mM EDTA]. 2. Pre-electrophorese the gel for 30 min prior to loading. The gels are run in l x TBE buffer [89 mM Tris-borate, 2 mM EDTA (pH 8.3)] at 1700-2000 V for 2 hr or until the bromphenol blue dye reaches the bottom of the gel.

[ 17]


371

3. Denature the samples at 95 ° for 2-3 rain prior to loading, cool on ice, and load samples into individual wells with an elongated, flat tip (Anachem, Luton, UK). 4. After the run, cover the gel with Saran Wrap and place the gel onto Whatman (Maidstone, UK) 3MM paper, supported by a layer of Whatman DE 81 paper, which binds the shorter fragments that pass through the 3MM paper. Dry the gel on a gel dryer (Bio-Rad) under vacuum at 80 ° for at least 1 hr. 5. Expose the gel to X-ray film (X-Omat LS; Eastman Kodak, Rochester, NY) or phosphoimaging screen for the appropriate time. If an intensifying screen is to be used, the film is exposed at - 7 0 ° . Exposure to X-ray film is usually for about 18 hr without the aid of an intensifying screen, and exposure to phosphoimager screens is usually 2 hr to visualize the bands using these conditions. An example of the method is shown in Fig. 2. Cisplatin adducts are clearly seen in the human c-jun gene when naked genomic DNA is treated with 5 #M drug. A 10-fold higher concentration of drug is required to treat cells in order to detect lesions sufficiently above background. However, the pattern of cisplatin binding is preserved in this sequence in intact cells.

Antibody Purification of DNA A problem with existing genomic sequencing procedures such as sslig-PCR or TD-PCR is that at low doses it is difficult to distinguish between bands representing lesions and background. Antibodies to specific lesions bind only fragments of DNA that contain one or more of these lesions. Purification of these damaged fragments provides an enriched template DNA for PCR. Therefore, an antibody purification procedure (Fig. 3) produces a high signal-to-noise ratio that allows bands representing positions of adducts generated at low doses to be determined by these techniques. Intrastrand cisplatin adduct-specific monoclonal antibodies (ICR4, clone CP9/19) were developed in hooded rats as described. 18 These antibodies can also bind the chemically identical intrastrand adducts produced when carboplatin binds DNA. Antibodies to 6-4 photoproducts, thymidine dimers, and antibodies to the 8oxoguanine adduct to study oxidative DNA damage are available commercially (Kamiya Biomedical, Seattle, WA). Antibodies to melphalan adducts are also available. 19 To visualize bands at low doses, the number of cycles necessary during the second-round exponential PCR step may need to be increased relative to unpurified 18 M. J. Tilby, C. Johnson, R. J. Knox, J. Cordell, J. J. Roberts, and C. J. Dean, CancerRes. 51, 123 (1991). 19 M. J. Tilby, J. M. Styles, and C. J. Dean, Cancer Res. 47, 1452 (1987).

372


Cells I

Cisplatin (,uM)

Naked DNA II

0

50 100

[ 1 7]

I

0

5

'Full-Length'

FIG. 2. Detection of cisplatin adducts on the nontranscribed strand of a coding region of the human

c-jun gene by the PCR procedure. The bands on the autoradiograph indicated by arrows represent fragments of DNA of varying length, which correspond to the positions of cisplatin intrastrand adducts along the sequence. The intensity of each adduct-specific band directly relates to the frequency of adducts at that particular position. Full-length represents the position of the downstream restriction enzyme cut. K562 cells were treated with the indicated concentrations ofcisplatin for 18 hr, and isolated DNA was treated for 1 hr. The sequences of the oligonucleotide primers used (JT 1, JT2, and JT3) were as previously described. 17

[ 17]


373

Damaged GenomicDNA Denature DNA

Bind Adduct Specific Antibody /J~

2, Bind IgG /J~

l

Bind Protein A SepharoseBeads Q

? Spin Beads Down and Wash Supematant Containing Unbound DNA Away LinearPCR using BiotinylatedPrimer 1.B (See Figure 1) FIG. 3. Antibodypurificationof adductedDNA. An intrastrandcisplatinadductin the startingDNA is indicated by a small triangle.

374


[17]

DNA. However, the high signal-to-noise (S/N) ratio obtained indicates that this can be achieved without a significant increase in background. Although not purified by antibody, the chemically cleaved products produced when DNA is treated for Maxam and Gilbert sequencing can be amplified by using the same number of cycles without affecting the ability to read the sequence. Preparation ofDNA. Both naked treated DNA and cells treated with drug and controls are purified by this procedure. It is important to completely digest the DNA with restriction enzymes to ensure maximum sensitivity. Large uncut fragments may be more difficult to isolate by immunoprecipitation. Restriction enzymes that cut as close to the biotinylated primer binding site as possible should be selected to minimize isolation of fragments of DNA that possess an adduct 3' of the primer-binding site. We have obtained the best results with primer and restriction enzyme combinations that produce a cut immediately upstream of the first biotinylated primer-binding site. If possible, this could be carried out as a one-step double-digest reaction when cutting at the site downstream of the primer-binding site. Alternatively, it may be necessary to perform two separate reactions. Binding of Antibodies to DNA. Both rabbit anti-mouse IgG and rabbit anti-rat IgG can be used to bind the cisplatin-specific antibodies. Antibodies are stored in small aliquots at - 8 0 ° and TN-milk is kept in aliquots at - 2 0 °. 1. Place 3/zg of damaged DNA sample in a 0.5-ml tube. Bring the volume to 10/zl with distilled water. 2. Denature damaged DNA at 100 ° for 3 min, cool on ice for 3 min, and centrifuge at 10,500g for 5 sec. 3. Add 10 #1 of cisplatin monoclonal antibody culture supernatant (10 #g/ml) to the 0.5-ml tube. Incubate for 1 hr at 37 °. 4. Dilute IgG (1.8-2.5 mg/ml; Sigma, Poole, UK) 10-fold with TN-milk [10 mM Tris-HCl (pH 7.5), 140 mM NaC1, 5% (w/v) nonfat dry milk powder, and yeast tRNA (0.1 g/liter; GIBCO-BRL)]. 5. Add 10/zl of diluted IgG to the 0.5-ml tube. Incubate at 37 ° for 1 hr.

Treatment of Protein A-Sepharose. Before incubation with the antibody-DNA mix, the protein A-Sepharose beads must be treated to reduce nonspecific binding. 1. Swell and wash the protein A-Sepharose beads (Sigma) according to the manufacturer instructions and store at 4 ° . 2. Transfer the protein A-Sepharose beads (250 mg/ml) to a 1.5-ml tube. Use 5/~1 per sample plus an extra 10/zl of beads and resuspend in 5/zl of TN-milk per sample plus 10/zl extra of TN-milk (i.e., for 10 samples resuspend 60 ~1 of beads in 110/zl of TN-milk).

[ 17]


375

3. Incubate at room temperature for 1 hr with occasional agitation. 4. Centrifuge at 2600g for 30 sec at 4 °, remove the TN-milk supernatant, and resuspend in 5 #1 of fresh TN-milk per sample plus 20 #1 extra of TN-milk.

Capture of Antibody Complex on Beads 1. Spin down and transfer antibody-DNA mixture to flesh 0.5-ml tubes. 2. Add 10 #1 of pretreated beads mixture to each 0.5-ml tube (with a cut tip, as bead mixture is fairly viscous). Incubate at 37 ° for 1 hr with occasional agitation. 3. Centrifuge the 0.5-ml tubes at 2600g for 30 sec and remove 200 #1 of supernatant carefully, leaving behind the white bead pellet. 4. Resuspend the bead mix in 200 #1 of TN [ 10 mM Tris-HC1 (pH 7.5), 140 mM NaCI] to wash the beads, centrifuge at 2600g for 30 sec at 4 °, and remove 200/~1 of supernatant carefully, again leaving behind the white bead pellet. 5. Repeat step 4, first with 200 #1 of TN and again with 200/~1 of TE [10 mM Tris-HC1 (pH 7.6), 1 mM EDTA] and finally with 200 #1 water. 6. Dry the beads-antibody-DNA mixture under vacuum. 7. Resuspend in I 0/~1 of distilled water for use in PCR and store at 4 ° until ready for TD-PCR. The first-round linear PCR is then carried out with the Sepharose beads and antibodies bound to template DNA. Figure 42o shows the results obtained when antibody-purified DNA is processed by the PCR procedure. Adducts are visualized at 5 #M, approximately 10-fold lower than possible with existing procedures (e.g., see Fig. 2). This increased sensitivity demonstrates that there is minimal nonspecific binding of antibody. The evident dose-dependent increase in adduct formation indicates that this technique can be used for quantitative measurement of lesions. Conclusions This modified/shortened protocol PCR protocol based on single-strand ligation PCR (sslig-PCR) 3-6 and terminal transferase-dependent PCR (TD-PCR) v using Taq polymerase works effectively in combination with antibody-purified template DNA to produce a highly sensitive method for detecting covalent lesions at the nucleotide level in intact cells. The sensitivity is such that lesions can be measured at pharmacologically relevant doses and studies of the sequence specificity and repair of individual lesions can be performed. The antibody-PCR combined procedure is extremely sensitive and highly reproducible.

2oD. Rozekand G. P. Pfeifer,Mol. Cell.Biol. 13, 5490 (1993).

376


Cisplatin (~uM) 0

[ 17]

5 10 20 40 80 'Full-Length'

FIG. 4. Detection of cisplatin intrastrand adducts by the PCR procedure, using immunoprecipitated DNA as template. The nontranscribed strand of the promoter region of tile human c-jun gene is shown. K562 cells were treated with the indicated concentrations of the cisplatin for 18 hr. Full-length represents the position of the downstream restriction enzyme cut. The sequences of the oligonucleotide primers used (JD 1, JD2, and JD3) were as previously described. 2°

Acknowledgment This work was funded as part of a Program Grant from the Cancer Research Campaign (SP2000/ 0402) and a Cancer Research Campaign studentship to C. J. McGurk.

[ 18]

COVALENT CROSS-LINKING OF LIGAND--DNA

INTERACTIONS

377

[18] Targeting DNA through Covalent Interactions of Reversible Binding Drugs B y DAVID E. GRAVES

Introduction The interactions of small molecules with nucleic acids have provoked considerable interest in the field of drug design over the past three decades. This interest is motivated by the unique functional role of DNA in living organisms and the structural features of DNA afforded by the planar stacking of aromatic base pairs perpendicular to the long axis of the double helix. J'2 Perturbations to the sequential readout of the D N A bases and/or the ability of this information to be identified and read by proteins and enzymes result in marked changes in the normal functioning of basic cellular processes such as replication, transcription, and DNA repair processes, making DNA an interesting and unique target for potential chemotherapeutic agents. 3-5 Studies as early as 1947 by Michaelis 6 noted the utility of DNA-binding agents as effective chemotherapeutic agents and later the studies of Kirk, the Kerstens, and Goldberg examining the structural and functional features of the interactions of actinomycin D with nucleic acids predicted the true potential of D N A as a powerful target for disrupting cellular metabolism and terminating replication and/or transcription. 7-9 One of the earliest classes of DNA-binding compounds to be associated with targeting DNA was the class of acridines. In the 1940s, simple acridine-based compounds such as 9-aminoacridine (structure shown in Fig. 1) and proflavin were used as antibacterial agents l°'ll and were subsequently used as probes to examine mutagenesis. 12,13 In 1956, correlations between biological activities and DNA binding were implicated in studies conducted by Peacocke and Skerrett, who I X. Qu and J. B. Chaires, Methods Enzymol. 321,353 (2000). 2 j. B. Chaires, Curr. Opin. Struct. Biol. 8, 314 (1998). 3 H. M. Berman and E R. Young, Annu. Rev. Biophys. Bioeng. 10, 87 (1981). 4 W. D. Wilson and R. L. Jones, Adv. Pharmacol. Chemother 18, 177 (1981). 5 S. Neidle and Z. Abraham, CRC Crit. Rev. Biochem. 17, 73 (1984). 6 L. Michaelis, Cold Spring Harb. Syrup. Quant. Biol. 12, 131 (1947). 7 j. M. Kirk, Biochim. Biophys. Acta 42, 167 (1960). 8 H. M. Rauen, H. Kersten, and W. Kersten, Z. Physiol. Chem. 321, 139 (1960). 9 E. Reich and I. H. Goldberg, Prog. Nucleic Acid Res. Mol. Biol. 3, 182 (1964). l0 C. N. Hinshelwood."The Kinetics of Bacterial Change." Clarendon Press, Oxford, UK, 1946. II A. C. R. Dean and C. N. Hinshelwood. "Growth, Function, and Regulation in Bacterial Cells" Clarendon Press, Oxford, UK, 1966. 12S. Brenner, L. Barnett, E H. C. Crick, and A. Orgel, J. Mol. Biol. 3, 121 (1961). 13K H. C. Crick, L. Barnett, S. Brenner, and R. J. Wattg-Tobin,Nature (London) 192, 1227 (1961).


Copyright ~3 2001 by Academic Press All rights of reproduction in any form reserved. 0076-6879/00 $35.00

378


[18]

Sara. L-N-MeVal~

2N~NH2

O

Sar

L-Pro L - P r o

\L-N-MeVal

D-Val D-Val

T.rJ

\Thr

I

J

OC

CO

OCH3

.O

"]~

0

CH3

FlG. 1. Chemical structures of selected DNA-binding agents used to study nucleic acid structure and function. Note the similar heterocyclic chromophores of ethidium (left), 9-aminoacridine (center), and actinomycin D (right).

demonstrated that the antibacterial activity of proflavin was directly related to the interaction of acridine with DNA.14 The intercalative mode of binding of acridines to DNA, originally proposed by Lerman in 1961, is now well characterized and a common binding motif for many anticancer agents. 15 The cytotoxic properties observed for DNA-binding agents have led to the extensive use of these compounds as chemotherapeutic agents in the treatment of numerous diseases including cancer16; hence, the mechanism through which small molecules such as chemotherapeutic agents, mutagens, and carcinogens interact with nucleic acids is of central biological significance. Biophysical studies correlate the physicochemical properties associated with formation of these complexes with various biological effects that are induced by these DNA-binding agents, such as inhibition of DNA replication and transcription, topoisomerase II inhibition, mutagenesis, and carcinogenesis. 17 Significant progress has been made toward unraveling the structural, thermodynamic, and kinetic properties associated with the interactions of ligand-DNA complexes. These studies provide pivotal insight into the design and development of more effective second- and third-generation chemotherapeutic agents that are being used in the successful treatment of many types of diseases; however, despite extensive research efforts, many molecular mechanisms responsible for eliciting the biological actions remain unknown. A key problem associated with examining the interaction of a drug with its DNA-binding site centers on the reversible nature of the ligand-DNA interaction; 14 A. R. Peacocke and N. J. H. Skerrett, Trans. Faraday Soc. 52, 261 (1956). 15 L. S. Lerman, J. Mol. Biol. 3, 18 (1961). 16 M. J. Waring, Annu. Rev. Biochem. 50, 159 (1981). 17 A. H.-J. Wang, in "Nucleic Acids and Molecular Biology" (F. Eckstein and D. M, Lilley, eds.), Vol. 1, pp. 32-54. Springer-Verlag, New York, 1987.

[ 1 8]

COVALENT CROSS-LINKING OF LIGAND--DNA INTERACTIONS

379

hence limiting the microscopic characterization of the complex. 18,19 One strategy used to circumvent this problem is photoaffinity labeling. 2° The usefulness of photoreactive agents as tools to probe structural and/or functional properties of the active sites of proteins has been well documented for more than three decades; however, relatively few applications of photoaffinity labeling have been directed toward the study of ligand-DNA interactions. 21 In the 1970s, Yielding and coworkers realized the usefulness of this method and successfully developed photoreactive analogs of ethidium and acridines as tools for examining ligand-DNA interactions. 22'23 Their studies demonstrated the utility of this method in characterizing ligand-DNA interactions and were extended toward the examination of mitochondrial mutagenesis, bacterial frameshift mutagenesis, and eukaryote DNA repair processes. 24-27 Studies by Graves and Yielding, using a variety of thermodynamic and kinetic measurements, determined that placement of the azido moiety on the phenanthridine ring (ethidium monoazide) did not significantly alter the intercalative DNAbinding properties of the ethidium photoreactive analogs from those of the parent ethidium.28'29 The photoreactive ethidium monoazide was shown to exhibit DNAbinding affinity, binding site size, and base-sequence specificity (in the absence of light) equivalent to those of the parent ethidium bromide; however, on photolysis by visible light, the azide moiety was converted to a highly reactive nitrene resulting in formation of a covalent bond in situ. 3° Over the past two decades, this laboratory and others have exploited this facility for covalent attachment to DNA of ethidium and other photoaffinity analogs of DNA-binding agents in an effort to more closely examine DNA-binding specificities at markedly lower ligand concentrations than had been possible with studies involving the reversibly binding parent molecules. Through covalent attachment of the ligand, much greater sensitivity in characterizing the sequence-selective binding of ligands to DNA is achieved. 31,32 We have extended these studies to include photoaffinity analogs of actinomycin 18 D. E. Graves, in "Advances in DNA Sequence Specific Agents" (L. Hurley and J. Chaires, eds.), Vol. 2, pp. 169-186. Jai Press, Greenwich, Connecticut, 1996. 19j. R. Knowles, Acc. Chem. Res. 5, 155 (1972). 20 E. Neilson and O. Buchardt, Photochem, Photobiol. 35, 317 (1982). 21 W. Lwowski, Ann. N.Y. Acad. Sci. 346, 491 (1980). 22 D. E. Graves, L. W. Yielding, C. L. Watkins, and K. L. Yielding, Biochim. Biophys. Acta 479, 98 (1977). 23 S. C. Hixon, W. E. White, Jr., and K. L. Yielding, Biochem. Biophys. Res. 66, 31 (1975). 24 W. E. White, Jr., and K. L. Yielding, Methods Enzymol. 46, 644 (1977). 25 M. Fukunaga and K. L. Yielding, Mutat. Res. 80, 91 (1981). 26 S. C. Hixon, W, E. White, Jr., and K. L. Yielding, J. Mol. Biol. 92, 319 (1975). 27 L. W. Yielding, W. E. White, Jr., and K. L. Yielding, Mutat. Res. 34, 351 (1976). 28 L. W. Yielding, D. E. Graves, and B. R. Brown, Biochem. Biophys. Res. Commun. 87~ 424 (1979). 29 O. E. Graves, C. L. Watkins, and L. W. Yielding, Biochemistry 20, 1887 (1981). 30 N. Turro, Ann. N.E Acad. Sci. 346, 1 (1980),

380


[ 18]

D (7-azidoactinomycin D) and 4'-(9-acridinylamino)methanesulfon-m-anisidide (m-AMSA)(3-azido-m-AMSA) to probe ligand-DNA interactions. 33-35 DNA I n t e r c a l a t o r s Intercalating agents constitute an important class of drugs that have been used for decades to probe DNA structure and function and in the treatment of various diseases including cancer. 36 The intercalative mode of binding involves the insertion of the planar polycyclic aromatic chromophore of the ligand between adjacent base pairs of the DNA duplex. Besides the chromophore involvement in binding, additional chemical substituents residing on the ligand are highly influential in determining the chemical properties such as charge, hydrophobicity, and aqueous solubility that in turn direct the thermodynamic binding mechanism(s), the geometry of the ligand-DNA complex, and the sequence selectivities observed for many of these agents. 37-39 There are numerous intercalators classified according to the structure of the chromophore, as shown in Fig. 1. Classic intercalators such as ethidium bromide and proflavin were among the first compounds to be studied with regard to their abilities to interact with DNA. Pioneering studies of Lerman4° revealed that addition of acridine, proflavin, or acridine orange to DNA resulted in marked changes in the viscosity and sedimentation coefficients. These observations led Lerman to predict that the planar acridine chromophore inserted between adjacent base pairs of the DNA helix, resulting in perturbations to the helical structure of the DNA. In 1965, the first biophysical characterization of the ethidium-DNA complex was presented by Waring, 41 who described the intercalative interaction of ethidium with DNA by using a variety of spectrophotometric methods. In 1967, these studies were extended by LePecq and Paoletti to include the characterization of the ethidium-DNA complex by fluorescence spectroscopy.42 Over the next several decades, numerous investigators have probed the interactions of small 3t j. M. Hardwick, R. S. von Sprecken, K. L. Yielding, and L. W. Yielding, J. Biol. Chem. 259, 11090 (1984). 32 R. L. Rill, G. A. Marsch, and D. E. Graves, J. Biomol. Struct. Dynam. 7, 591 (1989), 33 R. M. Wadkins and D. E. Graves, Biochemistry 30, 4277 (1991). 34 G. W. Rewcastle, W. A. Denny, W. R. Wilson, and B. C. Baugley, Anti-Cancer Drug Des. 1,215 (1986). 35 T.-L. Shieh, P. Hoyos, E. Kolodziej, J. G. Stowell, W. M. Baird, and S. R. Byrn, J. Med. Chem. 33, 1225 (1990). 36 M. J. Waring, Annu. Rev. Biochem. 50, 159 (1981). 37 j. B. Chaires, Anti-Cancer Drug Des. 11, 569 (1996). 38 R. M. Wadkins and D. E. Graves, Biochemiso 3' 30, 4277 (1991). 39 j. M. Crenshaw, D. E. Graves, and W. A. Denny, Biochemistry 34, 13682 (1995). 4o L. S. Lerman, J. Mol. Biol. 3~ 18 (1961). 41 M. J. Waring, J. Mol. Biol. 13, 269 (1965). 42 j. B. LePecq and C. Paoletti, J. Mol. BioL 27, 87 (1967).

[ 18]

COVALENT CROSS-LINKING OF L I G A N D - - D N A INTERACTIONS

38 1

molecules with nucleic acids by a variety of spectroscopic and crystallographic methods and have provided pivotal information as to the mechanisms of action of numerous biologically active compounds used in the treatment of a variety of diseases, including cancer. From these studies key insights have been gained regarding sequence-selective interactions of small molecules and proteins with DNA, thermodynamic and kinetic properties associated with the interactions of these compounds with DNA, and the abilities of these agents to alter DNA structures and/or enzymatic activities. The vast majority of these studies have dealt with reversibly binding drugs; hence, precluding a detailed characterization of the structural features and perturbations induced by formation of the ligand-DNA complex. Hence, in an effort to circumvent many of the problems associated with characterization of the ligand-DNA complex formed by reversible binding compounds, novel approaches such as photoaffinity labeling were applied. The remainder of this chapter describes the use of photoreactive analogs of several DNA-binding drugs that interact with DNA in a reversible manner. P h o t o r e a c t i v e A n a l o g s of E t h i d i u m Ethidium bromide and 9-aminoacridine are paradigms of DNA-intercalating agents. Biological activities of these compound have been known and exploited for more than half a century because of their propensities for inhibiting nucleic acid synthesis in a variety of parasitic organisms. 43-47 The trypanocidal activity of ethidium was linked to its inhibition of DNA polymerase as early as 1963 and postulated to arise from its interaction with nucleic acids. 48'49 Acridine, first isolated from coal tar more than a century ago, has been used extensively in the development of biologically active analogs, such as 9-aminoacridine and proflavin, that came into broad clinical use as antibacterial agents during World War II. 5°'51 Antimalarial qualities of quinacrine were recognized in the 1930s and the drug continues to be used in the treatment of this disease. 52 Both ethidium and the acridine analogs have been extensively used as DNA-intercalating agents to probe

43 B. A. Newton, Annu. Rev. Microbiol. 19, 209 (1965). 44 D. Kerridge, J. Gen. Microbiol. 42, 71 (1966). 45 j. E Henderson, Prog. Exp. Tumor Res. 6, 84 (1965). 46 R. Tomchick and 3H. G. Mandel, J. Gen. Microbiol. 36, 225 (1964). 47 B. A. Newton, in "Metabolic Inhibitors" (R. M. Hochster and J. H. Quastel, eds.). Academic Press, New York, 1963. 48 M. J. Waring, Biochim. Biophys. Acta 87, 358 (1964). 49 W. n. Elliott, Biochem. J. 86, 562 (1963). 50 A. C. R. Dean, in "Acridines" (R. M. Acheson, ed.), pp. 789-814. John Wiley & Sons, New York, 1975. 51 W. Schulemann, Proc. R. Soc. Med. 25, 897 (1931). 52R. M. Pinder, in "Medicinal Chemistry" (A. Burger, ed.), pp. 508-621. Wiley-lnterscience, New York, 1970.

382


[18]

and characterize DNA structure and function; however, investigation using these agents is limited because of the reversible nature of the ligand-DNA complexes. Photoaffinity labeling, a technique that involves the design and synthesis of analogs with the ability to covalently attach at their binding sites when photoactirated, provides a means of overcoming the limitation of reversible binding. Several photoreactive analogs of ethidium have been synthesized by replacing either or both of its amino groups at the 3- or 8-positions of the phenanthidine ring with an azido substituent (structures shown in Fig. 2A and B ) . 22'53 A

3~:H:anal o og H 2 N ~ (~

C2H5

NH2

N 3 ~

NH2

~H~_..~ N 8-~C a2zid~analog

etparent hidium C2H5 "~.~ 3,8diazidanal o og FIG. 2. (A) Conversion of the parent ethidium (left) to photoreactive analogs (right). Chemical synthesis involves conversion of one or both of the amino groups to the diazonium salt followed by substitution to the azido moiety. The three photoreactive analogs generated are 3-azido8-aminophenanthridinium chloride (top right), 8-azido-3-aminophenanthridinium chloride (middle right), and 3,8-diazidophenanthridinium chloride (bottom right). For synthesis details, see Graves et al. 10°. (B) Conversion of the parent acridine (left) to photoreactive analogs (Hght) including 9-azidoacridine (top right), 3-azido-6-aminoacridine (middle right), and 3-azido-6-amino-10methylacridine (bottom right). For synthesis details, see Graves et al. i00

[18]


383

B ]3

H 9-azd io

I

H2N~N3 H]

H

3-azido-6-amino

acridine

H2N~N [

CH3

3

3-azido-6-amino-

10-methyl

FIG. 2. (continued)

The value of ethidium as a photoaffinity analog is reflected by its ability to mimic the noncovalent DNA interaction of the parent ethidium bromide in forming reversible complexes (in the absence of light) identical to that of the parent compound. Biophysical data comparing the properties associated with these interactions are provided in Table I. Comparative studies of the interactions of the parent ethidium and the 8-azido analog (ethidium monoazide) with DNA reveal that the interactions of ethidium monoazide closely mimic those of the parent compound. 29 In contrast, the structural conformer, 3-azido-8-amino-5-ethyl-6phenyl-phenanthridinium chloride, was reported by LePecq and co-workers to bind DNA with an orientation that was inverted compared with that of the parent ethidium, with the phenyl and ethyl moieties residing in the helical major groove, in contrast to the minor groove interaction of the parent ethidium and ethidium monoazide. 54 53 W. J. Firth, C. L. Watkins, D. E. Graves, and L. W. Yielding, J. Heterocyclic Chem. 20, 759 ( 1981 ). 54 R Laugaa, A. Delbarre, J. B. Le Pecq, and B. P. Roques, Eur J. Biochem. 134, 163 (1983).

384


[18]

TABLE 1 BIOPHYSICAL PROPERTIES OF PHOTOREACTIVE ANALOGS OF ETHIDIUM BROMIDE

Characteristic

Parent ethidium

Ethidium monoazide (8-azido-3-amino)

Ethidium diazide

Visible properties •~max )~max + DNA

Efree ~°bound Fluorescence properties )-max em ~-max em + DNA

Relative intensity Relative intensity+DNA Pbotolysis rate

480 nm 518 nm 5680M 1 cm-I 2500 M - 1 c m - I

458 nm 495 nm 5 2 2 0 M - t cm 1 3560M tcm-l

432 nm 445 nm 5850M-1 cm-i 3880M -lcm t

6 0 0 nm 590 n m 1 21

590 nm 580 nm 1 21 K = 5.7 x 1 0 - 6 s e c - I (at80J I M-2sec-l

500 nm 507 nm 1 0.1 K = 6.1 x 10 5sec I (at80J-I M-2sec I

Yielding and co-workers revealed that the interactions of the ethidium monoazide with DNA resulted in relatively large increases in the fluorescence emission signal on complex formation, suggesting that the monoazide intercalates in a manner similar to that of the parent compoundfl 9 Stopped-flow studies revealed that the parent ethidium and photoreactive ethidium azide exhibit identical apparent rate constants for the dissociation of these drugs from D N A Y Similarly, the intrinsic binding constants and site exclusion sizes for the two compounds are identical, demonstrating the usefulness of the 8-azido-3-aminoethidium analog to characterize ethidium-DNA interactions as described in Table I. The utility of ethidium monoazide to probe biological functions such as mutagenesis and DNA repair processes has been well documented by Yielding and co-workers. Their studies demonstrate that covalent attachment of the ethidium analog to DNA resulted in induction of petite mutants for yeast mitochondria, frameshift mutagenesis in Salmonella, and repairable DNA lesions in human lymphocytes 56-6° Control experiments with the parent ethidium bromide and/or ethidium monoazide under nonphotolyzed conditions did not elicit these biological effects, indicative of a requirement for covalent attachment for their induction.

55 E Garland, D. E. Graves, L. W. Yielding, and H. C. Cheung, Biochemistry 19, 3221 (1980). 56 S. C. Hixon, W. E. White, Jr., and K. L. Yielding, J. Mol. Biol. 92, 319 (1975). 57 L. W. Yielding, W. E. White, Jr., and K. L. Yielding, Murat. Res. 34, 351 (1976). 58 C. E. Cantrell, K. M. Pruitt, and K. L. Yielding, Mol. Pharmacol. 15, 322 (1979). 59 M. Morita and K. L. Yielding, Mutat. Res. 54, 27 (1978). 60 M. Fukanaga and K. L. Yielding, Biochem. Biophys. Res. Commun. 84, 501 (1978).

[ 18]


385

In 1985, Kulkarni and Yielding demonstrated that DNA lesions produced by photolysis of ethidium in intact human lympocytes were alkali labile. 61 Alkali treatment of the ethidium-DNA adduct resulted in fragmentation of the DNA, presumably at the site of covalent modification. This discovery was coupled with emerging advances in DNA sequencing methodology and allowed the exact location of DNA lesions resulting from covalent attachment of the drug (alkylation) to be mapped. Application of this method was further refined by Rill and co-workers, using piperidine treatment of the ethidium-modified DNA adducts to generate DNA strand breaks at the site of covalent modification. 32'62Covalently attached ethidium was initially probed by this method; however, the method was rapidly extended to include photoreative analogs of 9-aminoacridine, quinacrine, actinomycin D, and m-AMSA. Because the ligand is irreversibly bonded to the DNA, the total ligand concentration could be decreased by 100- to 1000-fold compared with the reversible-binding parent compound. Hence, footprinting studies used to discern DNA sequence preferences were markedly more sensitive, allowing observations of novel high-affinity sites. Footprinting studies with the parent ethidium reported this compound to bind DNA in a random manner with no explicit preferences for DNA base sequence. However, studies by Rill and co-workers, employing covalently attached ethidium at low ligand-DNA binding densities, indicated a nonrandom distribution of binding sites, with the order of preference for covalent attachment being G > C, T > A. Analyses of neighboring base effects showed that guanine adducts were most reactive when flanked on the 5' side by another guanine and on the 3' side by either guanine or thymine. Cytosine adducts were most reactive when flanked on the 5' side by another cytosine and on the 3' side by guanine. Interestingly, the reactivities of ethidium with the d(CpG) and d(GpG) sequences were predicted from optical and nuclear magnetic resonance (NMR) studies of Reinhardt and Krugh. 63-65 Overall analyses of all possible intercalation sites revealed a binding order of d(CpG) ~ d(GpG) > d(GpC) ~ d(TpG) d(GpT) > d(GpA) ~ d(ApG) "-- d(TpA) > d(ApT) > d(ApA). 32'62 Ethidium Monoazide as Probe for Topoisomerase H-Mediated DNA Strand Cleavage Many compounds used in the treatment of cancer have the enzyme topoisomerase II as their molecular target. 66 This enzyme is responsible for controlling

61 M. S. Kulkarni and K. L. Yielding, Chem. Biol. Interact. 56, 89 (1985). 62 G. A. Marsch, D. E. Graves, and R. L. Rill, Nucleic Acids Res. 23, 1252 (1995). 63 C. G. Reinhardt and T. R. Krugh, Biochemistry 17, 4845 (1978). 64 R. V. Kastrup, M. A. Young, and T. R. Krugh, Biochemistry 17, 4855 (1978). 65 C. G. Reinhardt and T. R. Krugh, Bioehemistry 16, 2890 (1977). 66 S. J. Froelich-Ammon and N. Osheroff, J. Biol. Chem. 270, 21429 (1995).

386


[1 8]

the topology of the DNA in cells through enzyme-mediated double-strand breakage concomitant with strand passage of an intact DNA duplex through the break and followed by religation of the double-strand DNA break. Through this catalytic cycle, the enzyme controls over- and underwinding of the DNA duplex and can alleviate torsional stress that accumulates ahead of the replication forks. 67'68 The mechanism(s) through which topoisomerase II-targeting agents such as m-AMSA, daunomycin, Adriamycin (doxorubicin), mitoxantrone, actinomycin D, and ellipticine influence the activity of topoisomerase II remain unknown despite more than two decades of intense investigation. 69 Interestingly, these compounds are all classified as reversible DNA-binding agents, with their presumed mode of DNA binding via intercalation. However, not all DNA-binding agents are capable of altering the catalytic activity of topoisomerase II. While m-AMSA is a potent topoisomerase II inhibitor and its presence results in the enhanced stimulation of topoisomerase II-mediated single- and double-strand breaks, the structural conformer o-AMSA is ineffective in eliciting such a response. 7° In this case, both m-AMSA and o-AMSA bind to DNA through intercalation. In addition, o-AMSA h~s been shown to have a binding affinity that is four times greater than that of m-AMSA. 38 A second example of a DNA-intercalating drug that does not influence the catalytic activity of topoisomerase II is the classic intercalator, ethidium bromide. Early studies of Ross and co-workers demonstrated ethidium to be ineffective in eliciting topoisomerase II-mediated DNA cleavage at drug concentrations up to 100/zM. Indeed, the stimulation of topoisomerase II-mediated DNA cleavage by m-AMSA could be reversed with increasing concentrations of ethidium. 71 Hence, intercalation does not appear to be an absolute requirement for the modulation of topoisomerase II activity. In an effort to probe the influence of intercalative DNA binding on modulating topoisomerase II activity, photoaffinity labeling with ethidium monoazide was applied. Ethidium monoazide was chosen for these experiments because of its similarities in DNA binding to that of the parent ethidium bromide. 2~'22 Ethidium monoazide forms a complex with DNA (in the absence of light) that is identical to that of the parent ethidium. However, on photolysis with visible light, the azido moiety is activated to the reactive nitrene and covalent attachment of the ethidium in s i t u is achieved. These data (shown in Fig. 3) demonstrated that covalent attachment of ethidium to DNA resulted in stimulation of the topoisomerase II-mediated cleavage reaction, markedly enhancing both double- and single-strand breaks in the DNA at low covalently attached drug densities. A comparable level of 67j. C. Wang,Annu. Rev. Biochem. 65, 635 (1996). 68N. R. Cozzarelliand J. C. Wang,"DNATopologyand Its BiologicalEffects."Cold Spring Harbor LaboratoryPress, Cold Spring Harbor,New York, 1990. 69A. Y. Chen and L. F. Liu, Annu. Rev. Pharmacol. Toxicol. 34, 191 (1994). 70E. M. Nelson, K. M. Tewey,and L. E Liu, Proc. Natl. Acad. Sci. U.S.A. 81, 1361 (1984). 71W. E. ROSS,W.C. Rowe, B. Glisson, J. Yalowich,and L. R. Liu, CancerRes. 44, 5857 (1984).

[ 1 8]

387

COVALENT CROSS-LINKING OF L I G A N D - D N A INTERACTIONS

35 Single-Strand DNA Breaks

30 t~

~* 25 o < z

a

20

"0

=

10

0

o

~-

5 0 0

20

40

60

80

Covalently Attached Drugs Per Plasmid FIG. 3. Influence of covalently attached ethidium on the enhancement of topoisomerase If-mediated single-strand (©) and double-strand (ff]) DNA breaks. Control experiments using nonphotolyzed ethidium monoazide ( 0 ) and parent ethidium bromide (11) show no enhancement. The cleavage reaction was carried out as described in Marx et al. 74

single- and double-strand cleavage is observed at bound drug concentrations that are 150 times lower than required for equivalent cleavage stimulation by the known topoisomerase II poisons, m-AMSA and etoposide. 72'73 The capacity for covalent attachment of the drug to the DNA provides novel insight into the mechanism(s) through which DNA-binding drugs may influence topoisomerase II activity and provides details of the structural and functional properties of the ternary complex. The study revealed that covalent attachment of ethidium to the DNA resulted in the conversion of an intercalative binding ligand that was ineffective in modulating topoisomerase II activity to one that was highly active in enhancing topoisomerase II-mediated DNA single- and double-strand breaks. Covalent attachment of the ligand to the DNA prior to addition of the enzyme allowed control of the assembly of the ternary complex. In addition, insight into the recognition of DNA structural and/or stability properties by the enzyme was explored through studies comparing the binding affinities of the enzyme to native and ethidium-modified DNAs. These experiments revealed that the enzyme recognizes and binds to modified 72 M. J. Robinson and N. Osheroff, Biochemistry 29, 2511 (1990). 73 N. Osheroff, Biochemistry 28, 6157 (1989). 74 G. Marx, H. Zhou, D. E. Graves, and N. Osheroff, Biochemistry 36, 15844 (1997).

388

CHEMICAL AND MOLECULAR BIOLOGICAL APPROACHES NHSO2CH3

NHSO2CH 3

NHSO2CH 3

NH

NH

NH

H

H

H

m-AMSA

o-AMSA

[1 8 ]

3-azido-m-AMSA

FIG. 4. Structures of the acridine-based topoisomerase II poison, m-AMSA (left); the inactive conformer, o-AMSA (center); and the photoreactive analog, 3-azido-m-AMSA (right).

DNA (1 drug covalently bonded per 60 base pairs) with a binding affinity 5-fold greater than that of native DNA, indicative of the importance of DNA structural perturbations in ternary complex formation.74 P h o t o r e a c t i v e A n a l o g s of m-AMSA (Amsacrine) In the early 1970s, Cain and co-workers at the Cancer Research Laboratory in Auckland reported the synthesis of an anilinoacridine analog [4'-(9-acridinylamino)methanesulfon-m-anisidide] along with its enhanced activity against a variety of experimental tumors and leukemias (structure shown in Fig. 4). 75-77 Subsequent studies by Kohn and co-workers revealed m-AMSA to induce the formation of protein-associated single- and double-strand breaks in nuclear DNA through a ternary complex formed between the m-AMSA, DNA, and topoisomerase II. 78-8° The exact nature of the spatial orientation and role of m-AMSA within the ternary complex remained unknown until photoaffinity labeling using the photoreactive analog 4'-(3-azido-9-acridinylamino)methanesulfon-m-anisidide (3-azidom-AMSA) was applied by Freudenreich and Kreuzer. sl Prior biophysical studies 75 B. E Cain, G. J. Atwell, and W. A. Denny, J. Med. Chem. 18, 1110 (1975). 76 S. W. Hall, J. Friedman, and S. S. Legha, Cancer Res. 43, 3422 (1983). 77 W. Marsoni and R. Wittes, Cancer Treat. Rep. 68, 77 (1984). 78 y. Pommier, M. R. Mattem, R. E. Schwartz, L. A. Zwelling, and K. W. Kohn, Biochemistry 23, 2927 (1984). 79 j. Minford, Y. Pommier, J. Filipski, K. W. Kohn, D. Kerrigan, M. R. Mattem, S. Michaels, R. Schwartz, and L. A. Zwelling, Biochemistry 25, 9 (1986). so y. Pommier, J. Covey, D. Kerrigan, W. Mattes, J. Markovits, and K. W. Kohn, Biochem. Pharmacol. 36, 3477 (1987). st C. H. Freudenreich and K. N. Kreuzer, Proc. Natl. Acad. Sci. U.S.A. 91, 11007 (1994).

[18]

COVALENTCROSS-LINKINGOF LIGAND-DNA INTERACTIONS

389

by Byrn and co-workers demonstrated that DNA-binding properties of 3-azido-mAMSA were comparable to those of the parent m-AMSA. 82 In 1994, Freudenreich and Kreuzer extended the utility of photoaffinity labeling using 3-azido-m-AMSA to provide the first biophysical characterization of a ternary complex formed between the drug, DNA, and topoisomerase II. Prior to photolysis, the 3-azido-m-AMSA demonstrated stimulation of topoisomerase II-mediated DNA strand breaks comparable that of the parent m-AMSA. However, on photolysis, the photoreactive ligand covalently bonded to the substrate DNA. Mapping of the DNA-binding site of the covalently attached ligand revealed the drug to be covalently bonded to the DNA bases immediately adjacent to the two phosphodiester bonds that were cleaved by the enzyme. From these studies, Freudenreich and Kreuzer demonstrated that topoisomerase II created and/or stabilized preferential DNA-binding sites for the inhibitor precisely at the two sites of enzyme-mediated DNA cleavage. 81 P h o t o r e a c t i v e A n a l o g s of A c t i n o m y c i n D Actinomycin D (structure shown in Figs. 1 and 5) has long been a paradigm for sequence-selective binding to DNA. Early studies using native DNAs of heterogeneous base sequence by Gellert et al., Wells and Larson, and Mtiller and Crothers investigated the hydrodynamic, kinetic, and thermodynamic properties of actinomycin D-DNA interactions and demonstrated actinomycin D to bind strongly to double-strand DNAs (typically Kint of 1-5 x 106 M -1) with binding strength directly correlated with GC content of the DNAs. 83-85 Comparisons of the reversible DNA-binding properties of the parent actinomycin D and 7-azidoactinomycin D are provided in Table II. The structure of actinomycin D complexed with DNA was first determined by Sobell and co-workers, who proposed an intercalative binding model involving insertion of the phenoxazone chromophore between the d(GpC) step and placement of the cyclic pentapeptide side chains within the minor groove of the DNA. This orientation allows the drug to form hydrogen bond contacts between the carbonyl oxygen of a threonine residue with the 2-amino group of guanine; hence, dictating the d(GpC) sequence preference. 86'87 Further evidence of the preferred binding to the d(GpC) step was subsequently presented, using a variety of biophysical methods and DNase I footprinting studies. 82T.-L. Shieh, P. Hoyos,E. Kolodziej,J. G. Stowell, W. M. Baird, and S. R. Byrn,J. Med. 1225 (1990). 83M. Gellert,C. E. Smith, D. Neville,and G. Felsenfeld,J. Mol. Biol. 11, 445 (1965). 84W. MUllerand D. M. Crothers,J. Mol. Biol. 35, 251 (1968). 85R. D. Wells and J. E. Larson,J. Mol. Biol. 49, 319 (1970). 86S. C. Jain and H. M. Sobell,J. Mol. Biol. 68, 1 (1972). 87H. M. Sobell,Sci. Am. 231, 82 (1974).

Chem.

33,

390


pentapeptideside

[1 8]

chain~

phenoxazone chromophore

pentapeptsididchai ee n Front View

Side View

FIG. 5. Three-dimensional structure of the sequence-selective DNA-binding agent, actinomycin D. These structures reveal the spatial relationships between the planar phenoxazone chromophore flanked above and below b y the cyclic pentapeptide side chains. The structure was obtained from the Nucleic A c i d Database file (ddh048) ( T a k u s a g a w a

et al? 9) a n d visualized with Insight lI (MSI).

In 1992, Takusagawa and co-workers reported the structure of actinomycin D complexed with the self-complementary deoxyoctanucleotide d(GAAGCTTC)2 by X-ray crystallographic methods. This study revealed that on binding actinomycin D, the DNA becomes severely distorted (i.e., kinked) and, interestingly, displayed an asymmetry of the pentapeptide side chains of actinomycin D residing in the minor groove. Several contacts between the DNA and the ligand were observed, including hydrophobic interactions between the surface of the minor groove and the proline, sarcosine, and methylvaline residues of the cyclic pentapeptides as shown in Fig. 6. 88,89 Early studies by White and Phillips 9° and Fox and Waring 91 revealed complex association and dissociation kinetics needed to describe the actinomycin T A B L E 11 BIOPHYSICAL PROPERTIES OF PARENT ACTINOMYCIN D AND ITS PHOTOREACTIVE ANALOG, 7-AZIDOACTINOMYCIN D

Characteristic

Parent actinomycin D

7-Azidoactinomycin D

Visible properties

)'-max )~max + DNA ~ee 8bound D N A - b i n d i n g properties Kinl Binding site size Sequence specificity Photolysis rate

4 0 0 nm 4 5 0 nm 2 2 , 5 0 0 M l cm I 12,000 M - I cm - l

462 nm 472 nm 29,800 M - 1 c m - I 13,560 M ] c m 1

6 x 105M I(bp) 6 (bp) GC

5 × 105 M -1 (bp) 6 (bp) GC K = 6.2 x 10 3 s e c I (at 80 J I M 2 sec 1

[ 18]

COVALENTCROSS-LINKINGOF LIGAND-DNA INTERACTIONS

391

duplex DNA sequence (5'-CTATTGCA TAC-3')

phenoxazone ring

pentapeptide side chains

FIG.6. Energy-minimizedstructure showing the complexformed between actinomycinD and the 5'-CTATTGCATAC-3'/5'-GTATGCAATAG-3' duplex. The viewis fromthe minorgrooveof the DNA, illustrating the intercalativegeometryof the phenoxazonering at the GpC step and the pentapeptide side chains residing tightly within the minor grooveof the DNA.The structure was generated by Discover and Insight II (MSI). D - D N A interaction. Using homogeneous DNA systems (i.e., synthetic oligo- and polynucleotides), the association kinetics of the ligand-DNA interactions were characterized by several slow, unimolecular processes with qualitatively little or no sequence or length dependence in the binding kinetics. By contrast, utilizing heterogeneous DNAs (i.e., calf thymus), the association kinetics were shown to be highly complex. In 1986, Fox and Waring described a binding model to explain the time-dependent DNase I footprinting patterns, coining the term "shuffling hypothesis." In this model, initial interactions of actinomycin D with the DNA lattice are nonsequence specific. The ligands are presumed to "shuffle" along the lattice sampling binding sites until preferred sites are found. On interaction with the preferred sites, tight binding with slow dissociation is observed. 91 In collaboration with Waring and Bailly, we provided a direct test of the shuffling hypothesis by using the photoreactive analog of actinomycin D, 7-azidoactinomycin D. 92 The 7-azidoactinomycin D analog provides a unique probe for examining the time-dependent binding of actinomycin D to DNA. Similarly, photolysis of the azido moiety to the reactive nitrene converts this analog into a DNAalkylating agent that retains the intercalative DNA-binding characteristics of the s8 S. Kamitori and E Takusagawa,J. Mol. Biol. 225, 445 (1992). 89F. Takusagawa, L. Wen, W. Chu, Q. Li, R. G. Carlson, K. T. Takusagawa, and R. E Weaver, Biochemistry 35, 13240 (1996). 9oR. J. White and D. R. Phillips, Biochemistry 28, 6259 (1989). 91 K. R. Fox and M. J. Waring,Biochemistry 25, 4349 (1986). 92C. Bailly, D. E. Graves, G. Ridge, and M. J. Waring,Biochemistry 33, 8736 (1994).

392


[18]

parent actinomycin D molecule. In this study, the photoreactive actinomycin D analog was allowed to interact with a 32p-labeled restriction fragment for specified lengths of time ranging from 20 sec up to 45 min. After incubation, the drug-DNA complex was photolyzed, rendering the drug irreversibly bonded to the DNA. The modified DNA was then treated with hot piperidine and analyzed on a DNA sequencing gel. Results were quantitated by densitometric analysis and revealed time-dependent movements from selected sites on the DNA lattice; with classically low-affinity sites decreasing in reactivity (i.e., actinomycin D binding) and high-affinity sites increasing in reactivity as shown in Fig. 7. These studies provide conclusive support for the shuffling hypothesis and demonstrate that the locations of actinomycin D binding to the DNA lattice change significantly with time. 92'93 Of interest was the observation that even at the shortest times, the DNA sequencing patterns revealed actinomycin to be absent from DNA sequences having high AT content, indicating that the initial event in the sequence recognition process requires recognition of a global structural feature such as minor groove geometry. The minor groove of the DNA duplex within localized sites having high AT content is narrow in comparison with other regions of DNA with high GC or mixed sequences. The inability of actinomycin D to bind the AT-rich sites even at early equilibration times indicates an initial recognition of a sequence-dependent structural feature necessary for ligand interaction. 92 By the same analogy, DNA sequences with high GC or mixed content seemed to accommodate actinomycin D fairly readily, so that early binding of the ligand was observed at many such sequences (i.e., 5'-GpC-3' and 5'-GGG-3'), as shown in Fig. 3. Photolysis of the 7-azidoactinomycin D-DNA complex results in photoactivation of the azide to the reactive nitrene and adduct formation. The ability to covalently bond the actinomycin D analog gave us a highly sensitive probe to examine sequence selectivity of actinomycin D. Covalent attachment of the drug allows sequence specificity studies to be performed at much lower drug concentrations than in comparable studies using reversible binding compounds; thus, much greater specificity (i.e., high-affinity sites) can be observed. As shown by Kulkarni and Yielding 61 and H61bne and co-workers, 94'95 DNA photoadducts of ethidium monoazide and 3-azidoproflavin linked to DNA are alkali labile, thus circumventing the reliance on DNA cleavage agents such as DNase I or chemical cleavage, both of which are hampered by the need for single-hit kinetics and nonrandom DNA cleavage. Because the ligand-DNA adduct is selectively cut at the site of covalent attachment by treating the adduct with piperidine as for a sequencing reaction, the resolution of the sequence specificity is markedly enhanced. 93G. S. Ridge, C. Bailly, D. E. Graves,and M. J. Waring,Nucleic: Acids Res. 22, 5241 (1994). 94p. Cieplak,S. N. Rao, C H616ne,T. Montenay-Garestier,and P. A. Kollman,J. Biomol. Struct. 5, 361 (1987). 95C. H61~neand N. T. Thuong, Genome 31,413 (1989).

Dyn.

[181


393

40 o o o

5'-A-T-C-C-G-C-T-C-A-C-T-A-G-G-C-G-A-G-T-G-5'

35

x,.E. 30 position 47

~ 25 c-

-a 20 C

m 15 "0

[]

5'-A-T-C-C-G-C-T-C-A-C-

(1)

__-N-- 10

E o

Z

5

position 43

0

i

i

I

i

10

20

30

40

50

Incubation Time Prior to Photolysis (min) FIG. 7. Influence of DNA base sequence on binding of 7-azidoactinomycin D within the DNA sequence 5t-GTGAGCGGAT-3 ' as a function of equilibration time. The photoreactive analog of actinomycin D was added to the DNA and equilibrated for specific times prior to photolysis. At times less than 1 min, actinomycin D is shown to interact with DNA in a relatively nonspecific manner. However, as the equilibration time is increased, preferred binding sites show an increase in binding density, whereas less preferred sites show decreased binding densities. Experiments were performed as described in Bailly et al. 92 and Ridge et al. 93

Using this method, Rill and co-workers examined the sequence-specific interactions of 7-azidoactinomycin D and compared these results with those obtained by DNase I footprinting of the parent compound. 97 Results from these studies are consistent with a predominant preference for intercalation of the 7-azidoactinomycin D phenoxazone ring into dGpdC steps. The intercalation geometry is such that both the flanking G and C residues readily react to yield a piperidine-labile adduct. In addition, these studies revealed that the DNA bases that flank the intercalation site strongly influence the binding specificity of actinomycin D. 97 Using this method, 7-azidoactinomycin D was observed to exhibit a strong binding preference for a novel nontraditional sequence, the d[T(G)nT] motif, where n equals 2 to 4. Careful examination of this motif revaled actinomycin D to bind with the strongest affinity to the d(TGGGT) sequence. 9s,99

96 j. Goodisman, R. Rehfuss, B. Ward, and J. C. Dabrowiak, Biochemistry 31, 1046 (1992). 97 R. L. Rill, G. A. Marsch, and D. E. Graves, J. Biomol. Struct. Dyn. 7, 591 (1989). 98 S. A. Bailey, D. E. Graves, R. Rill, and G. Marsch, Biochemistry 32, 5881 (1993). 99 S. A. Bailey, D. E. Graves, R. Rill, and G. Marsch, Biochemistry 33, 11493 (1994).

394


[18]

Methodologies Using Photoaffinity Analogs as DNA-Binding Agents Photoaffinity labeling is a valuable yet underutilized tool for studying nucleic acid structure and function. This method is easily adaptable to both in v i t r o and in v i v o conditions and can provide key insight into the mechanism(s) of action of many biologically active DNA-binding agents. Conditions for the preparation and purification of the photoreactive analogs of DNA-binding agents are as varied as the agents themselves. The synthetic procedures and chemical properties of all the photoaffinity analogs described in this chapter are published and referenced accordingly.100 105 Utilization of the azido analogs as DNA-binding probes requires that the work be carried out under photographic "darkroom" conditions. In all synthesis, storage, and subsequent experiments, the extent of exposure to light must be rigorously restricted. Exposure to room light for even a few minutes will result in degradation of the photoreactive agent. Red safety lights have been used with no observed degradation of the photoreactive analog. Covalent attachment of the photoaffinity analog to DNA is accomplished by the following general procedure. Photolysis conditions may be adapted to fit the necessary conditions dictated by the nature of the DNA sample being modified. In general, DNA samples in 10 mM buffer (pH 7.0) containing 1 mM disodium EDTA, and 0.01 to 0.1 M sodium chloride, are equilibrated with known concentrations of the photoreactive ligand. The temperature of the ligand-DNA solution may vary, dependent on the length and topology of the target DNA (i.e., native DNAs with higher melting temperatures may be modified at room temperature whereas deoxyoligonucleotides may require lower temperatures to ensure the duplex structure of the DNA target). Care must be taken to prevent premature exposure to light; hence, the ligand is prepared and the concentration is determined under photographic safelight conditions. After an appropriate equilibration time, the ligand-DNA complex is photolyzed for a brief period via exposure to visible light. The method of choice in this laboratory has been to use two Haake Buchler (Paramus, NJ) light boxes, each containing two General Electric daylight No. F15T8-D fluorescent light bulbs delivering a total of ~80 J. m -2 sec-l. A photolysis kinetic profile in buffer is advisable to gauge the amount of time and energy required for photolysis of the ligand. For the ethidium monoazide adducts, the photolysis is carried out at 5 ° to minimize heating of the sample by the close proximity of the light boxes. The nominal time and efficiency of photolytic attachment vary with the particular drug analog. For example, covalent attachment of the ethidium monoazide to calf thymus DNA is around 10oD. E. Graves,L. W. Yielding, C. L. Watkins,and K. L. Yielding, Biochim. Biophys. Acta 479, 98 (1977). Jo~D. E. Gravesand R. M. Wadkins,J. Biol. Chem. 264, 7262 (1989). Jo2W. A. Denny, B. E Cain, G. J. Atwell, C. Hansch, A. Panthananickal, and A. Leo,J. Med. Chem. 25, 276 (1982). Io3W. J. Firth, S. G. Rock, B. R. Brown, and L. W. Yielding,Mutat. Res. 81, 295 (1981). io4 K. L. Yieldingand L. W. Yielding,Ann. N. 1( Acad. Sci. 346, 368 (1980). Io5W. E. White, Jr., and K. L. Yielding,Methods Enzymol. 44, 644 (1977).

[18]

COVALENTCROSS-LINKINGOF LIGAND--DNAINTERACTIONS

395

30% with extended photolysis times. In contrast, the photolytic efficiency of 7azidoactinomycin D is >95% under identical conditions. Prior to photolysis, the ligand-DNA complex is reversible; hence, during photolysis, ligand is undergoing association/dissociation and can react with water to yield the reversible binding hydroxylamine photoproduct. After photolysis, all ligand that has not bonded to the DNA must be removed. In the case of charged ligands such as ethidium monoazide or m-AMSA, this is readily accomplished by small spin columns filled with Chelex resin (Bio-Rad, Hercules, CA). 106'107 For 7-azidoactinomycin D, which is neutral, size-exclusion chromatography is used. The ligand-DNA adducts made in this laboratory have been relatively stable when maintained at 5 ° and kept in the dark. Ethidium adducts formed with supercoited plasmid DNA have maintained the supercoiled topology for up to 6 months with no appreciable degradation of the supercoiled topology when compared with nonmodified supercoiled plasmid stored at 5 ° and wrapped in aluminum foil. However, exposure to room light (prior to and during gel electrophoresis) will result in substantial single-strand cleavage, comparable to that observed for the parent ethidium-supercoiled DNA complex. Conclusions The use of photoaffinity labeling to study ligand-DNA interactions provides a powerful tool for the overall assessment of ligand targeting, complex structure, and biological and/or molecular mechanism(s) resulting from these interactions. Many of the reversible binding ligands that interact with DNA do so with high affinities, have chromophores that absorb in the visible range, and hence are amenable as potential candidates or parent compounds for the development of photoreactive analogs. The utility of the photoaffinity analog depends on the following criteria: (1) the ligand should be photoreactive at visible wavelengths such that cellular and/or DNA damage is not incurred; (2) the lifetime of the reactive product (nitrene) should be sufficiently short to assure quenching by solvent interactions and exclusion of any secondary perturbations; (3) the reaction time of the reactive nitrene should be sufficiently fast to ensure that the covalent adduct formed is representative of the bound ligand in its reversible state; and (4) the resulting ligand-DNA adduct must be stable such that isolation and characterization of the adduct is possible. This chapter has dealt primarily with ethidium, acridine, and actinomycin derivatives, providing key studies utilizing the photoreactive analogs to examine a number of biophysical and structural problems; however, the potential for applying this methology to other DNA-binding agents offers exciting avenues for future research. Io6E L. Gilbert,D. E. Graves,M. Britt, and J. B. Chaires,Biochemistry 30, 10931 (1991). 107p. L. Gilbert,D. E. Graves,and J. B. Chaires,Biochemistry 30, 10925(1991).

396


[19]

[19] Chemical Cross-Linking of Drugs to DNA By RYAN A. LUCK and PAUL B. HOPKINS Introduction It is well known that a number of antitumor agents and carcinogens garner their activities through "chemically cross-linking" DNA. These are compounds that become covalently linked to the DNA duplex. "Chemical cross-linking," however, is an imprecise term in that there are a number of categories of DNA lesions that can be considered the result of these modifications. Four of these categories are considered here. The first of these is interstrand cross-links, which are characterized by covalent linkages formed between the two different strands that form the DNA duplex. Second, there are intrastrand cross-links, covalent linkages formed between two different sites on one strand of DNA. Third, there are monoadducts, which are characterized by the covalent bonding of the compound to one site on one strand of the DNA duplex. Finally, there are a small number of agents that are able to covalently bind DNA via a second endogenous bridging compound. All four categories of chemical cross-links are important in understanding the activity of antitumor agents and carcinogens. As a result, it is important to understand the methods used in their analysis. Initially, this chapter discusses methods for analyzing DNA interstrand crosslinks and includes methods for their detection, location, and determination of their covalent structure. After this, we discuss methods used to analyze intrastrand cross-links and monoadducts. The fourth category of cross-link, wherein agents are covalently linked through endogenous bridging compounds, is not as prevalent as the other categories. As a result, there are fewer published methods for analysis; however, we have developed methodologies in our laboratory for analyzing the formaldehyde-mediated coupling of doxorubicin (Adriamycin) to DNA. This new methodology is explained in the final part of the chapter.

D e t e c t i o n of I n t e r s t r a n d C r o s s - L i n k s

Ethidium Bromide Fluorescence Assay The first challenge for analysis of interstrand cross-links is detection. A number of methods have been developed for detecting the presence of interstrand crosslinks. One of the earliest techniques was the ethidium bromide fluorescence assay. The ethidium bromide fluorescence assay takes advantage of the increase in fluorescence that results when ethidium bromide binds to double-stranded DNA. This assay was initially developed by Morgan and Paetkau in 1972.1 In broad terms,



[19]

CHEMICAL CROSS-LINKING OF DRUGS TO D N A

397

a DNA sample is exposed to an interstrand cross-linking agent. The sample is then mixed with an ethidium bromide solution and the fluorescence of the resuiting mixture is measured. This is followed by heat denaturation, a fairly short cooling period, and another measurement of the fluorescence. If there is no interstrand cross-linking, the fluorescence of the sample after the cooling period will be greatly diminished. The short cooling period does not allow enough time for the single-stranded DNA sequences to find their complements and reform the duplex. However, if an interstrand cross-link is present, then the strands were never completely separated, resulting in the relatively quick reforming of the duplex. Lusthof and co-workers examined the interstrand cross-linking activity of 2,5bis(1-aziridinyl)-3,5-bis(carboethoxyamino)-1,4-benzoquinone (AZQ, I) and several related compounds, using the ethidium bromide fluorescence assay. 2 The specific procedure is outlined below.

(I) 2,5-bis(1-aziridinyl)-3,5-bis(carboethoxyamino)- 1,4-benzoquinone (AZQ) A solution of the reduced form of AZQ (I) (40/~1, 0.05 mM) is combined with 360 #1 of calf thymus DNA solution (6.7 x 10-1 mg/ml) and allowed to react at 37 ° for 1 hr. Excess cross-linking agent is removed with Sephadex G-50 and ethanol precipitation. The DNA precipitate is dissolved in 100 #1 of water overnight at 4 °. Thirty microliters of this solution is combined with 2 ml of aqueous ethidium bromide solution and the fluorescence is measured. The samples are then heat denatured at 95 ° for 3 min, quickly followed by cooling at 20 ° for 7 min. Immediately after this, the fluorescence of the sample is measured. The ratio of the fluorescence after and before denaturation [(FJFu)s] is compared with the same ratio for a blank sample [(FJFu)u]. Using these ratios, an interstrand cross-link percentage is calculated, using Eq. (1): Cross-link percentage = L

l~(~/~)~b)b

J

100%

(1)

Lusthof and co-workers are careful to note that the cross-link percentage is not directly correlated to the percentage of base pairs covalently linked. Instead, it is

I A. R. Morgan and V. Paetkau, Can. J. Biochem. 50, 210 (1972). 2 K. J. Lusthof, N. J. De Mol, L. H. Janssen, W. Verboom, and D. N. Reinhoudt, Chem. Biol. Interact. 70, 249 (1989).

398

CHEMICAL AND MOLECULARBIOLOGICALAPPROACHES I

I

r

[

m

m

/

native duplex DNA

[1 9]

~

Interstrand Cross-link

m

DNA w/ cross-linking agent

FIG. 1. Schematicof DPAGEassay used for detection of interstrand cross-linking. correlated to the percentage of DNA that is in duplex form after the heat denaturation and quick cooling process. Denaturing Polyacrylamide Gel Electrophoresis

Gel electrophoresis is a widely used technique for manipulation of a diversity of biomolecules, including DNA, RNA, and proteins. With the increased availability of relatively short, custom-made oligonucleotides, denaturing polyacrylamide gel electrophoresis (DPAGE) has become a commonly performed assay for manipulating DNA and assessing its interactions with a variety of other compounds. DPAGE is a special type of gel electrophoresis in which the matrix of the gel has been loaded with some type of chemical denaturant. (This denaturant, often urea, is usually a component of the sample loading buffer as well.) The denaturant hinders the formation of the normal DNA duplex by keeping the single-stranded oligonucleotides from base pairing with their complement. This technique is used in a number of ways, including sequencing and purification. DPAGE is also a useful method for the detection of interstrand cross-links. When a double-stranded oligonucleotide is denatured, the single strands move with approximately twice the mobility (compared with the original duplex) through the gel matrix. If an interstrand cross-link has been formed, however, the strands cannot separate. As a result, they retain their retarded mobility and form a discrete band that does not progress nearly as far through the gel (Fig. 1). This slower mobility band can be detected and quantified either through UV shadowing of the gel or by prior radiolabeling of the oligonucleotide. Huang et al. investigated the interstrand cross-linking activity of formaldehyde in 1992. 3 Utilizing a library of self-complementary 17-mer oligonucleotides, they determined which sequences yielded interstrand cross-links, using the DPAGE assay. Solutions of radiolabeled DNA duplexes that have been incubated with 3 H. Huang, M. S. Solomon,and R B. Hopkins,J. Am. Chem. Soc. 114, 9240 (1992).

[19]

CHEMICALCROSS-LINKINGOF DRUGSTO DNA

399

formaldehyde are ethanol precipitated. The precipitate is dissolved in 10 ttl of water followed by the addition of 10 #1 of sample loading buffer [90% (v/v) formamide, 10 mM Tris [pH 7.5], 0.1% (w/v) xylene cyanol, and 0.1 mM EDTA]. This mixture is heated at 90 ° for 2 min, iced for 2-3 min, and then loaded onto the gel. DPAGE is conducted on a Hoefer (San Francisco, CA) thermojacketed Poker Face gel stand, using a 25% (w/v) gel (19 : 1 acrylamide : bisacrylamide, 8 M urea, 0.35 mm thick, 33 x 41 cm, 20-tooth comb). The gel is run at 65-70 W at approximately 55 ° until the xylene cyanol dye had run 14-16 cm. The gel is then transferred onto Whatman (Clifton, N J) 3M filter paper, covered with Saran Wrap, and dried with a Bio-Rad (Hercules, CA) model 583 gel drier. Autoradiography is then used to visualize the single-stranded and cross-linked duplexes. Our laboratory has found that DPAGE is an extremely robust method for detection of interstrand cross-linked DNA. This method has been used for detecting interstrand cross-linking by a variety of compounds, including cisplatin, mitomycin C, nitrogen mustard, formaldehyde, and others. Variations in size and thickness of the gel, urea concentrations, acrylamide : bisacrylamide ratio, temperature, loading buffer composition, and power usually can affect only the aesthetic appearance of the resulting data, not the conclusions. With regard to the loading buffer, it should be noted that as long as the sample is heat denatured, the buffer does not necessarily need to contain a chemical denaturant. Furthermore, cross-linking reactions have been directly loaded onto the gel with good results. If the sample is ethanol precipitated, however, it is important first to dissolve the sample in water prior to adding a urea-based loading buffer. A high concentration of urea often makes it difficult for the DNA precipitate to dissolve, resulting in a DNA sample that never migrates into the gel matrix. D e t e r m i n i n g S e q u e n c e L o c a t i o n of I n t e r s t r a n d C r o s s - L i n k s

Use of Fenton Chemistry The DPAGE assay described in the previous section provides a method for determining which DNA base sequences yield interstrand cross-linking. Using this method, it is possible to began making inferences about which specific DNA bases are involved in formation of an interstrand cross-link. Further investigation is needed, however, in order to make definitive conclusions about the bases involved. One method for determining the sequence location of interstrand cross-links combines the DPAGE methodology with Fenton chemistry. In 1985, Tullius and Dombroski demonstrated that a mixture of iron(II), hydrogen peroxide, EDTA, and ascorbate resulted in cleavage of the DNA backbone. 4 In this mixture, iron(II) reduces hydrogen peroxide to give a hydroxyl radical. The 4 T. D. Tulliusand B. A. Dombroski,Science230, 679 (1985).

400


5'

* 3'

Fe(EDTA)2H202 ascorbate

[ 19]

* 3' *3 ' - -

*3'--

+ cross-linked fragments

*3'--

Radiolabeled products FIG. 2. Hydroxyl radical cleavage of interstrand cross-linked DNA.

hydroxyl radical abstracts a hydrogen atom from the deoxyribose sugar that composes the DNA backbone, resulting in nearly sequence-random cleavage. Ascorbate is present in order to reduce the resulting iron(III) back to iron(II), thereby continuing the process. This procedure, known as hydroxyl radical footprinting, has been used to examine interactions between DNA and proteins as well as global structural changes associated with unusual DNA sequences. Hydroxyl radical footprinting has been used in our laboratory to pinpoint the sequence location of interstrand cross-links. 5 In a DNA duplex that has only one interstrand cross-link, single-hit cleavage generated by Fenton chemistry results in a mixture of products. If the 3' end of one strand is radiolabeled, there are only two categories of products observed when the mixture is analyzed by DPAGE (Fig. 2). The first category is composed of the radiolabeled products that still contain the interstrand cross-link. The second category consists of the short fragments that correspond to cleavage on the 3' side of the cross-link on the radiolabeled strand. When the mixture is run on the gel, the two categories have different mobilities; the fragments that still contain the cross-link are much larger and move much more slowly through the gel. The location of the interstrand cross-link can then be determined by simply analyzing the length of the short fragments generated. This technique also works if the 5' terminus of one of the single strands is selectively radiolabeled. Radiolabeling of either both 5' termini or both 3' termini can also work, as long as the sequence is self-complementary.

[2+ l H3N, ,NH3 Pt cl" "cI

(II) - cis-diamminedichloroplatinum(II) (cisplatin) This technique has been used in our laboratory to determine the exact sequence location of the interstrand cross-link formed by cisplatin (II). 6 Radiolabeled DNA 5 M. E Weidner, J. T. Millard, and E B. Hopkins, J. Am. Chem. Soc. 111, 9270 (1989). 6 E B. Hopkins, J. T. Millard, J. Woo, M. E Weidner, J. J. Kirchner, S. T. Sigurdsson, and S. Rancher, Tetrahedron 47, 2475 (1991).

[19]

CHEMICALCROSS-LINKINGOF DRUGS TO DNA

401

that has been incubated with cisplatin is initially analyzed by DPAGE. The interstrand cross-linked product, which demonstrates approximately half the mobility of the single strands, is visualized by autoradiography and the band is removed from the gel matrix with a razor blade. The gel slice is crushed with a glass rod and incubated at 37 ° in aqueous solution (0.5 M NH4OOCCH3, 1 mM EDTA) for 16 hr. The eluate is extracted with a glass pipette and loaded onto a Waters (Milford, MA) Sep-Pak C~8 cartridge in order to remove salts and gel particulate. On this cartridge, the sample is sequentially washed with 10 ml of aqueous NHaOOCCH3 (10 mM) and 10 ml of water. The sample is then recovered by washing the cartridge with 4 ml of 25% CH3CN-water. This mixture is evaporated in vacuo and the sample is redissolved in water. The 20-#1 reaction mixture for the hydroxyl radical cleavage is then formulated using DNA (100,000 cpm by Geiger counter), (NH4)2Fe(SO4)2 (50/zM), EDTA (100/zM), sodium ascorbate (1 mM), H202 (10 mM), and Tris (pH 7.5, 5 mM). After 1 min at room temperature, the reactions are stopped by addition of 2 #1 of aqueous thiourea (50 mM). The mixture is then lyophilized, dissolved in 5 #1 of sample loading buffer, heated to 90 ° for 3 min, cooled to 0 °, and analyzed by DPAGE. DPAGE in this instance is a 25% (w/v) acrylamide gel cross-linked with 5% (w/v) bisacrylamide, 48% (w/v) urea, 0.35 mm thick, 41 x 37 cm. It is run at 2000 V on a Hoefer thermojacketed Poker Face gel stand at 64-68 °. It has been the experience in our laboratory that the components of the hydroxyl radical cleavage reaction need to the made immediately prior to use, particularly the Fe(II) and sodium ascorbate solutions. Once again, however, the specific composition of the gel used for the DPAGE is not critical. The relatively high temperature of the gel portion of the assay is used in order to make the separation between the various short fragments and the cross-link as well defined as possible. This high temperature is not necessary, however, for the experiment to work. Piperidine Cleavage at Interso'and Cross-Linked Deoxyguanosine Residues

Heating DNA with aqueous piperidine leads to cleavage at deoxyguanosine (dG) residues that are alkylated at N7. There are two steps to this process: (1) depurination and (2) cleavage at the resulting abasic site. This reactivity can be used to demonstrate the presence of interstrand cross-links at N7. Rink et al. and Millard et al. used this assay to determine the sequence preference of the interstrand cross-link formed by mechlorethamine (III).7'gA single-stranded oligonucleotide containing the sequence 5'-d(GJG2G3C) is radiolabeled and annealed to its complement. The resulting duplex is treated with mechlorethamine and the interstrand cross-linked material is isolated by DPAGE and "crush and soak" methodology

c,Ha

cI~N~cI (III) - mechlorethamine

402


[ 19]

described previously. This DNA is then heated in 50 #l of 1 M piperidine in a sealed microcentrifuge tube for 1 hr. The mixture is dried in vacuo, lyophilized twice from 25 #1 of water, dissolved in sample loading buffer, and analyzed by DPAGE. DPAGE analysis shows that cleavage on the radiolabeled strand occurs primarily at G 2. This result, combined with experiments involving the complement strand and other duplexes, helped to determine that the sequence selectivity of the mechlorethamine interstrand cross-link targeted 5'-d (GNC). C o m p o s i t i o n of I n t e r s t r a n d C r o s s - L i n k s

Acid Hydrolysis of Cross-Linked DNA Generally, the next step in the analysis of interstrand cross-links is to determine the exact chemical composition and structure of the covalent linkage. This includes not only information about the cross-link itself, but also its points of attachment to the bases. The quickest way to gather information about the structure of the cross-link is to cleave the oligonucleotides into their single-base components for analysis. This section addresses various methods for this process; it does not detail, however, the various methods for analyzing the resulting components. The methods for identifying these products are similar to those for identifying natural products, including nuclear magnetic spectrometry (NMR), mass spectrometry, authentic sample comparisons, and others. These methods are too broad to be fully explained here and are addressed elsewhere. The initial methodology for breaking DNA into its single base components was acid hydrolysis. In 1961, Brookes and Lawley used acid hydrolysis to identify the products of a variety of alkylating agents and DNA. 9 From these experiments came the first evidence that mechlorethamine (III) was a DNA cross-linking agent, although it was unclear to the authors whether this was an intrastrand or interstrand cross-link. Unfortunately, the details of the acid hydrolysis procedure were not published. Two different procedures for acid hydrolysis of cross-linked DNA were published, however, in regard to investigations of the cis-syn-thymine photodimer (IV). 1°,11 The cis-syn-thymine photodimer is formed at adjacent thymines when DNA is exposed to ultraviolet light. This cross-link is an intrastrand cross-link and thus does not involve any exogenous alkylating agents, but the procedures should be applicable to interstrand cross-links as well (provided they are acid stable). 7 S. M. Rink, M. S. Solomon, M. J. Taylor, S. B. Rajur, L. W. McLaughlin, and E B. Hopkins, J. Am. Chem. Soe. 115, 2551 (1995). 8 j. T. Millard, S. Raucher, and E B. Hopkins, J. Am. Chem. Soc. 112, 2459 (1990). 9 E Brookes and E D. Lawley, Biochem. J. 80, 496 (1961). l0 A. J. Varghese and S. Y. Wang, Nature (London) 213, 909 (1967). 11 D. Weinblum, Biochem. Biophys. Res. Commun. 27, 384 (1967).

[ 19]


403

Varghese and Wang exposed approximately 5 g of native calf thymus DNA to ultraviolet conditions. Afterward, the DNA sample is evaporated to dryness, and the material is dissolved in 200 ml of trifluoroacetic acid (TFA). l° This sample is transferred to 20 Pyrex glass tubes, which are then sealed and heated at 170 ° for 90 min. To remove insoluble material, the sample is then filtered. The tubes and the insoluble material are washed twice with 0.1 N HCI. The filtrate is evaporated to dryness. Varghese and Wang than dissolve the residue in TFA again and analyze the components by thin-layer chromatography and NMR.

O~CH3CH3J? HN")x""~-~ ''" %NH

(IV) - cis-syn thymine photodimer Weinblum uses a different method for hydrolyzing DNA containing the cissyn-thymine photodimer, li Approximately 1.5 g of salmon sperm DNA is exposed to the cross-linking conditions. The dried, cross-linked DNA is then dissolved in 3 ml of 70% (w/v) perchloric acid. This is heated to I00 ° for 45 min. Water 150 ml is added and the solution is neutralized by 10 N KOH. The solution is left overnight; the KC104 precipitate is then filtered off. The filtrate is adjusted to pH 10, using ammonium hydroxide, and the hydrolysis products are separated by Dowex chromatography. Enzymatic Digestion of lnterstrand Cross-Links

There are relatively few examples of the utilization of acid hydrolysis to isolate cross-linked components of DNA. This is because the harsh conditions of the procedure can potentially alter the structure of the cross-links, or even reverse their formation. In 1977, Dubelman and Shapiro addressed this limitation by developing methodology for digesting cross-linked DNA enzymatically.12,13 They accomplished this by performing a 98-hr digest of cross-linked DNA, using a combination of deoxyribonuclease I and snake venom phosphodiesterase. A variety of procedures for enzymatic digestion of DNA have since been developed and are outlined here (Table I; structures V and VI). As can easily be seen, a number of enzyme recipes have been used for digestion of cross-linked oligonucleotides. It should be noted that in the case of the nitrous acid cross-link, only the minimum enzymes are used--alkaline phosphatase and 12S. Dubelmanand R. Shapiro,NucleicAcids Res, 4, 1815 (1977). ~3R. Shapiro, S. Dubelman,A. M. Feinberg, P. E Crain, and J. A. McCloskey,J. Am. Chem. Soc. 99, 302 (1977).

TABLE I SAMPLING OF ENZYMATICDIGESTION METHODS FOR CROSS-LINKEDDNA Cross-linking agent CH20

cis-DEP (V)

Mitomycin C (VI)

AZQ (I)

BCNU (VII)

Nitrous acid

Enzymatic digestion procedure initial sample of cross-linked calf thymus DNA (ca. 60 mg) is brought up to 50 ml with digestion buffer (5 mM Tris-HC1, I mM NaCI, 1 mM MgCI2, pH 8.8). Deoxyribonuclease I (2 mg, 4000 units) is added at - 3 0 ° for 16 hr. Venom phosphodiesterase (1.5 rag, 30 units) and spleen phosphodiesterase are added at - 3 0 ° for another 6 hr. Alkaline phosphatase (150/zl, 250 units) is added at - 3 0 ° for 6 hr. Reaction mixtures are analyzed by HPLC Calf thymus DNA ( 100 #g) that has been incubated with cis-DEP (V) is ethanol precipitated and dried. It is then resuspended in digestion buffer (20 mM sodium acetate, 50 mM NaCI, 1 mM ZnSO4, 10 mM MgCI2, pH 4.6). Bovine pancreas deoxyribonuclease I (200 Kunitz units) is added at - 3 0 ° for 4 hr. SI nuclease (1000 units) is added at - 3 0 ° for 16 hr. One-tenth volume of solution of Tris (1.0 M, pH 9) and 5 units of alkaline phosphatase are added at - 3 0 ° for another 4 hr. Reaction mixtures are analyzed by HPLC Calf thymus DNA that has been cross-linked is brought to 3 A260 units/ml in digestion buffer (0.005 M Tris-HC1, 0.001 M MgC12, pH 7.0). Digestion is accomplished at 37 ° with this sequence of enzymes: DNase 1 (16 units/A260 unit) at 0 and 1 hr; snake venom phosphodiesterase (1.25 units/A260 unit, pH increased to 5.5) at 2 and 5 hr; alkaline phosphatase (0.5 unit/A260 unit) at 7 hr, incubation going until 24 hr. Reaction mixtures are analyzed by HPLC Cross-linked DNA (1.0 A260 unit) is dissolved in 100/zl of digestion buffer (50 mM Tris, 10 mM MgCI2, pH 7.5) and digested by DNase I (50 units), DNase 11 (21 units), snake venom phosphodiesterase (5 units), and calf intestinal alkaline phosphatase (15 units) at 25 "~for 18 hr. Reaction mixtures are analyzed by HPLC Cross-linked DNA sequence [5t-d(GC4G4C)]2 (0.4 A260 unit) is digested in a 30-#1 total volume of digestion buffer (50 mM Tris, 10 mM MgC12, pH 8.5) by a combination of DNase I (10 units), calf intestinal phosphatase (10 units), DNase II (4 units), phophodiesterase I (0.5 unit), and Sl nuclease (180 units). Reactions are allowed to go for 3 hr at 3T ~ and are analyzed by HPLC Cross-linked DNA sequence [5'-d(CG)612 (0.35 A260 unit) is digested in a 75-/zl total volume of digestion buffer (50 mM Tris, 10 mM MgCI2, pH 8.9) by a combination of calf intestinal phosphatase (2/~1, 20 units) and phophodiesterase I (6/~1, 4 units). Reactions are allowed to go 8 hr at 37 ° and are analyzed by HPLC

Ref.

h,,

~t

e

Y

a y. E Chaw, L. E. Crane, E Lange, and R. Shapiro, Biochemistry 19, 5525 (1980). t~A. Eastman, Biochemistry 24, 5027 (1985). c A. Eastman, Biochemistry 22, 3927 (1983). d M. Tomasz, D. Chowdary, R. Lipman, S. Shimotakahara, D. Veiro, V. Walker, and G. L. Verdine, Proc. Natl. Acad. Sci. U.S.A. 83, 6702 (1986). e S. C. Alley, K. A. Brameld, and E B. Hopkins, J. Am. Chem. Soc. 116, 2734 (1994). fP. L. Fischhaber, A. S. Gall, J. A. Duncan, and E B. Hopkins, CancerRes. $9, 4363 (1999). g J. J. Kirchner, S. T. Sigurdsson, and P. B. Hopkins, J. Am. Chem. Soc. 114, 4021 (1992).

[19]


405

phosphodiesterase. These are the two primary enzymes needed. As the amount of DNA or the structural perturbations introduced by the cross-link increase, other digestive enzymes are called on to help. It should also be mentioned that with small amounts of DNA, digestion can be accomplished at room temperature in less than 1 hr. [

I

H2N, ,NH2 Pt CI" "CI (V) - cis-dichloro(ethylenediamine)platinum(II)

(cis-DEP) O

H2N~O~NH2 o (VI) - mitomycin C Generally, it is the practice in our laboratory to analyze the results of the enzymatic digestion ofinterstrand cross-linked DNA by using reversed-phase highperformance liquid chromatography (HPLC). Analytical HPLC runs are generally performed on an Alltech (Deerfield, IL) 5-#m, Cls, 250 × 4.6 mm Econosphere column. A typical solvent gradient is run at 1 ml/min. A good starting place for gradient development is as follows: solvent A, 10 mM ammonium acetate; solvent B, 100% acetonitrile; isocratic 92% solvent A for 7 min, a 13-min linear gradient to 70% solvent A, a 10-min linear gradient to 60% solvent A, and a 10-min linear gradient to initial conditions. Conditions are then adjusted to increase separation of the nucleosides and cross-linked material. On occasion, the peak percentage of acetonitrile over the gradient must be increased in order to elute the cross-linked material from the column. In this manner, the cross-linked bases can be isolated and identified by other analytical techniques.

Controlled Depurination of Cross-Links Involving N7 Previously, methodology for piperidine-mediated cleavage of DNA at sites of N7 alkylation has been described. Similarly, N7-alkylated sites are vulnerable to controlled depurination by exposure to heat. This characteristic is advantageous when it is thought that an interstrand cross-link is made between N7 of two purines. Tong and Ludlum have used this method to isolate the interstrand cross-link from DNA that had been treated with N,N'-bis(2-chloroethyl)-N-nitrosourea (BCNU,

406


[1 9]

VII). 14 The procedure they used for controlled depurination is fairly straightforward. Approximately 16 mg of calf thymus DNA that had been exposed to BCNU is ethanol precipitated, washed with ethanol, and dried. The precipitate is then dissolved in 2 ml of sodium cacodylate buffer (25 mM, pH 7.0). This is followed by heating at 100° in a water bath for 15 rain and then cooling in an ice bath. The depurinated oligonucleotides are precipitated by the addition of 0.2 ml of 1 N HCi; this precipitate is then removed by centrifugation. The supernatant is evaporated in v a c u o and redissolved in 200 #l of water. The pH is adjusted to pH 6.0 with about 50 #l of 1 N NH4OH so that the sample is suitable for HPLC. Using this methodology, Tong and Ludlum were able to isolate 1,2-(diguan-7-yl) ethane (VIII), the major interstrand cross-link formed by BCNU. N"0 H I

I

CIJ~N'~N~cI 0 (VII)

- N,

N'-bis(2-chloroethyl)-N-nitrosourea O

O

(VIII) - 1,2-(diguan-7-yl) ethane It should be mentioned that enzymatic digestion can be used to isolate nearly every cross-link that can be isolated by controlled depurination. There are potential scenarios, however, in which controlled depurination would be the preferred method. One possible situation would be if the cross-linking were performed under conditions that prohibited the activity of the enzymes. Another possible situation would be if the researcher wanted only N7-alkylated purines isolated, because of yield or sensitivity issues. If controlled depurination is not necessary, however, enzymatic digestion is the preferred method in our laboratory. G l o b a l S t r u c t u r e of I n t e r s t r a n d C r o s s - L i n k s The final challenge for complete analysis of interstrand cross-links is the determination of the global structure of the entire oligonucleotide. Although there are numerous techniques for obtaining limited information about the global structure of oligonucleotides, the two most powerful methods are X-ray crystallography

J4W. E Tongand D. B. Ludlum,Cancer

Res.

41, 380 (1981).

[19]


407

and nuclear magnetic resonance spectrometry. The details involved in performing either of these types of analysis are too broad to be covered here and are explained elsewhere in this volume.

Analysis of Intrastrand Cross-Links and Monoadducts Most of the techniques described for the analysis of interstrand cross-links can also be used, with little or no modification, for the analysis of intrastrand cross-links or monoadducts. With respect to detection of cross-linking, DPAGE is an excellent technique for observing intrastrand cross-links or monoadducts. (For obvious reasons, the ethidium bromide fluorescence assay is not useful for this goal.) Intrastrand cross-links and monoadducts generally do not hold the duplex together under the denaturing conditions. The presence of a lesion, however, retards the mobility of single-stranded DNA through the gel matrix. As a result, both monoadducts and intrastrand cross-links can be observed by DPAGE, although the separation is not as dramatic as it is with interstrand cross-links. The most proficient method for determining the sequence location involved in interstrand cross-links, the use of Fenton chemistry, is not nearly as applicable with regard to intrastrand cross-links and monoadducts. This technique can be used only if it is possible to detect which hydroxyl radical-generated fragments have the intrastrand cross-link or monoadduct present. Because the change in mobility is less dramatic, this feat is more difficult than it is with an interstrand cross-link. Determination of the location of monoadducts is further complicated by the greater probability that there is more than one lesion. Analysis of results of piperidinemediated cleavage is potentially a useful technique, but only in those cases that involve purine alkylation at N7. After detection of intrastrand cross-links or monoadducts, the most common route in our laboratory is to break the DNA into its components and determine which bases are involved. As it was with interstrand cross-links, the methods of acid hydrolysis, enzymatic digestion, and controlled depurination are all applicable for this process. Also as it was with interstrand cross-links, however, enzymatic digestion is by far the most common method. The same enzyme recipes and analytical techniques used for determining the composition of intersti'and cross-links are the starting point for determining the composition of intrastrand cross-links and monoadducts.

Formaldehyde-Mediated

C o u p l i n g of DNA to A d r i a m y c i n

BackgroundInformation Until now, this chapter has not addressed the fourth category of "chemical cross-links" that was outlined in the introduction. These are the cross-links that result from the coupling of an agent to DNA via a third compound. As an example

408


O

OH

O OH30

OH O HO NH 2

IX

OH

O

OH O

O

O DNA + CH20

O

o

[1 9]

H3c'~O

N-~LL- N -H

3

- - ~

.~--o

~"~CH2~''~"

X

FIG. 3. Formaldehyde-mediated coupling of Adriamycin to DNA.

of this category, we outline the methodology we developed for investigating the formaldehyde-mediated coupling of doxorubicin (Adriamycin, IX; Fig. 3) to DNA. The formaldehyde-mediated coupling of Adriamycin (and the structurally similar daunorubicin) to DNA was discovered by Wang and co-workers using X-ray crystallography. 15'16 In this structure, Adriamycin is linked via its amino group to N2 of a deoxyguanosine residue through a methylene group (X, Fig. 3). The methylene bridge presumably results from successive nucleophilic attacks on formaldehyde by the amino groups of Adriamycin and dG. This structure became of interest to our laboratory when Koch and co-workers related it to the putative interstrand cross-links formed by Adriamycin, as seen by Phillips and co-workers. When DNA was exposed to reductively activated Adriamycin, the Phillips laboratory used DPAGE methodology and observed a band that was consistent with the presence of an interstrand cross-link. 17.=8The Koch laboratory used electrospray mass spectrometry to demonstrate that lesion X was formed under the same reductive conditions.19,2° Koch and co-workers suggested that the formaldehyde resulted either from H202-promoted oxidation of buffer components or from BaeyerVilliger oxidation of Adriamycin itself. (The H202 is generated by reduction of O2 by the reducing agents in the reaction mixture.) Koch and co-workers proposed, but did not definitively prove, that the formation of lesion X was responsible for observation of the putative interstrand cross-links. (For a more complete history of the contributions of various laboratories in this research area, please see Luce et al. 2j ) Our laboratory was able to provide supporting evidence for Koch's suggestion by 15 A. H. Wang, Y. G. Gao, Y. C. Liaw, and Y. K. Li, Biochemistry 30, 3812 (1991). 16 j. Wang, M. Chao, and A. H. Wang, Adducts of DNA and anthracycline antibiotics. In "Anthracycline Antibiotics," Vol. 574, pp. 168-182. American Chemical Society, Washington, D.C., ] 99 l. 17 C. Cullinane, A. van Rosmalen, and D. R. Phillips, Biochemistry 33, 4632 (1994). 18 S. M. Cutts and D. R. Phillips, Nucleic Acids Res. 23, 2450 (1995). 19 D. J. Taatjes, G. Gaudiano, K. Resing, and T. H. Koch, J. Med. Chem. 39, 4135 (1996). 20 D. J. Taatjes, G. Gaudiano, K. Resing, and T. H. Koch, J. Med. Chem. 40, 1276 (1997). 21 R. A. Luce, S. T. Sigurdsson, and P. B. Hopkins, Biochemistry 38, 8682 (1999).

[19]

CHEMICALCROSS-LINKINGOF DRUGSTO DNA

409

confirming that either H202 or CH20 mixed with Adriamycin and DNA generated the same double-stranded band when analyzed by DPAGE. The problem with DPAGE, however, is that it is not a consistent detection method for lesion X. Adriamycin is covalently attached to only one of the two strands of the DNA duplex. Noncovalent interactions between Adriamycin and the nonalkylated strand, as well as noncovalent interactions between the two DNA strands themselves, are essential for keeping the duplex together under the denaturing conditions of DPAGE. This makes the results of the DPAGE experiment difficult to predict for various oligonucleotides, depending on their G-C base pair content and length. If the length of the DNA was too short or if there were not enough G-C base pairs, no double-stranded band would be seen, even if it contained the [5'-d(GC)]2 sequence that was essential for its formation. Another problem is that small variations in experimental conditions lead to drastically different results. The temperature, time length, percentage of urea, and composition of loading buffer all affect the outcome of the DPAGE experiments when detecting the lesion. One could imagine that these different variables might somehow be combined and computed to ascertain how "denaturing" a particular DPAGE experiment was. An experiment must be denaturing enough to pull apart the two strands of a DNA duplex when Adriamycin is simply noncovalently bound. Experimental conditions cannot be so denaturing, however, that the lesion is not strong enough to hold the duplex together. This level of denaturing would obviously vary according to the oligonucletide sequence. Given this level of complexity, we decided to attempt to develop another assay for detection of lesion X that would be quantitative and flexible for a variety of oligonucleotides.

Synthesis of N-Methyl Adriamycin The simplest way to detect the presence of lesion X would be to isolate the formaldehyde-mediated conjugate of Adriamycin with deoxyguanosine by employing the enzymatic digestion methods previously described in this chapter. Attempts to isolate this conjugate, however, returned only unmodified nucleosides and Adriamycin. It is presumed that when lesion X is free from duplex DNA, the aminal in X is too vulnerable to hydrolysis. (For similar reasons, there is no detectable monoadduct when DNA containing lesion X is analyzed by DPAGE.) Therefore, the strategy that we chose was to capture the Adriamycin iminium ion intermediate that was generated when this aminal decomposed. Prior to implementing this strategy, however, it was necessary to synthesize a sample ofN-methyladriamyxin (XI). To synthesize N-methyl Adriamycin, a reaction mixture composed of Adriamycin (100 #M), NaCNBH3 (67 #M), and CH20 (2.5 raM) is assembled in a solution of 2 : 1 CH3CN : H20. After 1.5 hr at room temperature in the dark, the reaction mixture is analyzed by analytical HPLC [Alltech, 5 #m, Cls, 250 x 4.6 mm Econosphere column, 1 ml/min; solvent A, 50 mM ammonium acetate (pH 4.0);

410


[ 19]

solvent B, acetonitrile; isocratic 50% solvent A]. HPLC gives three broad peaks, the first of which is determined to be Adriamycin by comparison of retention times and coinjection. The second and third peaks are determined by electrospray mass spectrometry to be N-methyl Adriamycin (XI) (MH + = 558.5), and N,N-dimethyl Adriamycin (MH + = 572.5), respectively.

CH30

0

OH

0

OH 0

0

H3C'~ HO NH I CH3 (XI) N-methyl adriamycin

Gas Chromatography-Mass Spectrometry Analysis The procedure for generation of samples for gas chromatography-mass spectrometry analysis (GC-MS) is based on that developed by Blakeney and co-workers for separating and analyzing mixtures of amino sugars. 22 The overall strategy for capturing the iminium ion intermediate is outlined here (Fig. 4, structures X I I XV). 2t Reactions of DNA with Adriamycin activated by a variety of different agents are ethanol precipitated after completion of the reaction. After the supernatant is removed and the DNA pellet dried down, 5/zl of glucosamine solution (10 mM) is added as an internal standard. Immediately, 50/zl of aqueous sodium borohydride (5 M) is added, the microcentrifuge tube used as the reaction vessel is pierced with a needle in order to prevent pressure from building up, and the reduction is allowed to progress at 65 ° for 1.5 hr. After the reduction, 50/zl of glacial acetic acid is added to consume the excess sodium borohydride. To acetylate the amino sugars, 100/zl of N-methylimidazole and 1 ml of acetic anhydride are added. After 30 min at room temperature, the acetic anhydride is quenched by addition of 2 ml of H20 and the acetylated amino sugars are extracted with 1 ml of dichloromethane. The dichloromethane is then dried down in vacuo. The samples are kept dry until immediately prior to GC-MS analysis, at which point such sample is dissolved in 100/zl of acetonitrile. GC-MS analysis is performed with a Hewlett-Packard (Palo Alto, CA) 5890 gas chromatograph and a Micromass Trio2000 quadrupole mass spectrometer. GC-MS is conducted with a DB-5 column, a source temperature of 200 °, and an 22 A. B. Blakeney, E J. Harris, R. J. Henry, and B. A. Stone, Carbohydr. Res. 113, 291 (1983).

[ 19]

CHEMICALCROSS-LINKINGOF DRUGS TO DNA 0

OH

O CH30 O

DNA

1OH

0

OH O

H+j OH2 O

k.)"

O

loss of chromophore

S

=

5 M NaBH4

CH30 O

OH O

411

.... t.;NZUN

acetic anhydride,

HH~NHH--Rl-methylimidazole H--t--OH

.o--i--H

OH3 XII: R = H XIII: R = Me

CH2OAc

H'Ji~N~RH+H-Ac H'-I--OAc

AoO--I--H

OH3 XIV: R = H XV: R = Me

HN+

IX + CH20 FIG. 4. Strategy for detection of lesion X.

oven temperature gradient as follows: 140 ° for 1 min, increase at 25°/min to 215 °, increase at 6°/min to 240 °, increase at 70°/min to 280 °, hold at 280 ° for 3 min. Autosampling is performed with a Hewlett-Packard 7674a autosampler, and data analysis is performed with Windows-based Micromass MassLynx software. Along with the samples generated by exposure of DNA to Adriamycin and various activating agents, known concentrations of Adriamycin and N-methyl Adriamycin isolated by HPLC are subjected to the reduction and acetylation procedure. These are the first samples analyzed by GC-MS; they are monitored with total ion current. This is done in order to determine the approximate retention times of the Adriamycin derivatives and glucosamine standard. Once these retention times have been determined, single-ion monitoring (SIM) is used to more precisely measure the most intense signals for XIV (98.1), XV (112.1), and the glucosamine standard (84.1). Quantification of [XV] in the DNA reaction mixtures is then performed in all GC-MS experiments by using the least-squares regression line determined for the N-methyl Adriarnycin standards versus its GC-MS response. In this manner, we have been able to find a correlation between the amount of XV observed by GC-MS and the amount of double-stranded band seen by the DPAGE

412


[20]

methodology. More importantly, however, we have been able to detect significant amounts of XV in DNA duplexes that contained the sequence [5'-d(GC)]2 but contained no other G-C base pairs in the flanking sequence. These are sequences that generally do not show the double-stranded DNA band when analyzed by DPAGE (data not shown). The GC-MS assay described herein opens avenues for further investigation of the sequence specificity of lesion X formation, as well as its possible generation in biological systems.

Caveats about Denaturing Polyacrylamide Gel Electrophoresis Methodology The experience of our laboratory with Adriamycin underscores a serious and potentially general failure of DPAGE analysis with regard to interstrand crosslinking. The literature regarding Adriamycin, as well as a number of its derivatives, presumes that the thermal lability of this putative interstrand cross-link is the obstacle to isolation of the cross-link. It turned out that in this system, noncovalent interactions were evidently strong enough to leave the DNA duplex intact under the conditions of DPAGE. It seems unlikely that the Adriamycin family of DNAbinding agents is the only family in which this is the case.

[20] High-Resolution Complexes

Using

Footprinting Chemical

and

By K E I T H R. Fox and M I C H A E L

Studies

of Drug-DNA

Enzymatic

Probes

J. WARING

Introduction Footprinting is a method that has been developed for determining the sequencespecific binding of small molecules, oligonucleotides, or proteins to DNA. The technique, which was originally designed for analyzing protein-DNA interactions,1 is based on the ability of ligands to protect DNA from enzymatic or chemical cleavage at their binding sites. The footprinting probe cuts or modifies free DNA, while regions to which a sequence-selective agent are bound are protected from attack. The DNA fragment is labeled at the 3' or 5' end of one strand, usually with 32p, but 33p or fluorescent labels can also be used, and the regions of protection are detected by autoradiography or scanning after separating the products of digestion on denaturing polyacrylamide gels. Several enzymatic and chemical agents have been developed as footprinting probes, and each has its own characteristic advantages and disadvantages. I D, J. G a l a s and A. Schmitz, Nucleic Acids Res. 5, 3157 (1978).

METHODS IN ENZYMOLOGY,VOL.340

Copyright@ 2001 by AcademicPress All rights of reproductionin any form reserved. 0076-6879/00$35.[10

412


[20]

methodology. More importantly, however, we have been able to detect significant amounts of XV in DNA duplexes that contained the sequence [5'-d(GC)]2 but contained no other G-C base pairs in the flanking sequence. These are sequences that generally do not show the double-stranded DNA band when analyzed by DPAGE (data not shown). The GC-MS assay described herein opens avenues for further investigation of the sequence specificity of lesion X formation, as well as its possible generation in biological systems.

Caveats about Denaturing Polyacrylamide Gel Electrophoresis Methodology The experience of our laboratory with Adriamycin underscores a serious and potentially general failure of DPAGE analysis with regard to interstrand crosslinking. The literature regarding Adriamycin, as well as a number of its derivatives, presumes that the thermal lability of this putative interstrand cross-link is the obstacle to isolation of the cross-link. It turned out that in this system, noncovalent interactions were evidently strong enough to leave the DNA duplex intact under the conditions of DPAGE. It seems unlikely that the Adriamycin family of DNAbinding agents is the only family in which this is the case.

[20] High-Resolution Complexes

Using

Footprinting Chemical

and

By K E I T H R. Fox and M I C H A E L

Studies

of Drug-DNA

Enzymatic

Probes

J. WARING

Introduction Footprinting is a method that has been developed for determining the sequencespecific binding of small molecules, oligonucleotides, or proteins to DNA. The technique, which was originally designed for analyzing protein-DNA interactions,1 is based on the ability of ligands to protect DNA from enzymatic or chemical cleavage at their binding sites. The footprinting probe cuts or modifies free DNA, while regions to which a sequence-selective agent are bound are protected from attack. The DNA fragment is labeled at the 3' or 5' end of one strand, usually with 32p, but 33p or fluorescent labels can also be used, and the regions of protection are detected by autoradiography or scanning after separating the products of digestion on denaturing polyacrylamide gels. Several enzymatic and chemical agents have been developed as footprinting probes, and each has its own characteristic advantages and disadvantages. I D, J. G a l a s and A. Schmitz, Nucleic Acids Res. 5, 3157 (1978).

METHODS IN ENZYMOLOGY,VOL.340

Copyright@ 2001 by AcademicPress All rights of reproductionin any form reserved. 0076-6879/00$35.[10

[20]

FOOTPRINTINGSTUDIES

413

Enzymes, such as DNase I, are often easy to use but typically overestimate the size of the footprint (on account of their size) and generate uneven cleavage patterns in the absence of added ligand. Chemical agents, such as dimethyl sulfate, osmium tetroxide, or diethyl pyrocarbonate (DEPC), are of limited usefulness because they react only with specific DNA bases. Other chemical probes such as methidiumpropyl-EDTA-Fe(II) [MPE-Fe(II)] act by intercalation and so perturb the DNA structure. Arguably the best cleavage agent for accurately mapping small molecule binding sites on DNA is the hydroxyl radical. P r i n c i p l e s of F o o t p r i n t i n g The DNA fragment of interest is first labeled at a 3 t or 5 f end, to allow detection of the cleavage products. After formation of the ligand-DNA complex, using appropriate buffers and equilibration conditions, the complex is subjected to cleavage. The concentration of the cleavage agent and the time of incubation are adjusted so that on average there is only one cleavage event per DNA strand (i.e., "single-hit" kinetics). To achieve this low level of cutting, conditions are chosen so that at least 50% (preferably >90%) of the DNA remains uncut. C h o i c e of DNA F r a g m e n t Footprinting fragments are usually obtained by restriction enzyme digestion of appropriate plasmids. These restriction fragments are usually between 50 and 200 base pairs in length. Longer fragments can be used, but it is more difficult to limit the digestion to conform with single-hit kinetics and the longer products are not properly resolved. Different laboratories have adopted various natural fragments as standard footprinting substrates, although both our laboratories have extensively used the 160-base pair t y r T DNA fragment 2'3 (Fig. 1). Other laboratories have used various fragments cut out of pBR322 or pBS. In many ways it would be convenient if a few fragments became recognized standards, as this would facilitate direct comparison of the affinities and specificities of ligands examined in different laboratories. It should also be noted that a limitation of the footprinting technique is that the (unknown) preferred binding site(s) for a ligand must be present within the DNA substrate. This may not be a problem for small ligands, which recognize two or three base pairs, but it can become a limitation for compounds with greater sequence selectivity unless that selectivity is known or predicted, in which case "designer" sequences or plasmid inserts can be employed. Because there are 10 possible dinucleotide steps, 64 different trinucleotides, 136 tetranucleotides, and 256 pentanucleotides, it is clear that as the specificity of the ligand increases there 2 H. R. Drew and A.A. Travers,Cell 37, 491 (1984). 3 K. R. Fox and M. J. Waring,Nucleic' Acids Res. 12, 9271 (1984).

414


[[20]

TyrTDNA 5'-AATTCCGGTTACCTTTAATCCGTTACGGATGAAAATTACGCAACCAGTTCATTTT 3'-AAGGCCAATGGAAATTAGGCAATGCCTACTTTTAATGCGTTGGTCAAGTAAAA TCTCAACGTAACACTTTACAGCGGCGCGTCATTTGATATGATGCGCCCCGCTTCCCGA AGAGTTGCATTGTGAAATGTCGCCGCGCAGTAAACTATACTACGCGGGGCGAAGGGCT TAAGGGAGCAGGCCAGTAAAAAGCATTACCCCGTGGTGGGGGTTCCC ATTCCCTCGTCCGGTCATTTTTCGTAATGGGGCACCACCCCCAAGGGCT-5'

Multisite DNAinsert 5'-GGATCCATATGCGGCAATACACATGGCCGATTTCCAACTGCACTAGTCGTAGCGCGA 3'-CCTAGGTATACGCCGTTATGTGTACCGGCTAAAGGTTGACGTGATCAGCATCGCGCT TCAAGGTTAAGCTCCCGTTCTATCCTGGTATAGCAATTAGGGCGTGAAGAGTTATGTA AGTTCCAATTCGAGGGCAAGATAGGACCATATCGTTAATCCCGCACTTCTCAATACAT AAGTACGTCCGGTGGGGTCTGTTTTGTCATCTCAGCCTCGAATGCGGATCCTTCATGCAGGCCACCCCAGACAAAACAGTAGAGTCGGAGCTTACGCCTAGG-

3 ' 5 '

FIG. 1. Sequence of tyrT and multisite DNA fragments. The lower strand of TyrT DNA is usually labeled by filling in at the 3' end of the EcoRI site with [~-32p]dATP while the upper strand is labeled at the 3' end of the Aval site with [ot-32p]dCTP. Two clones of the multisite fragment are used in which the sequence is inserted into the BamHI site of pUC 18 in opposite orientations. HindlII-Sacl restriction fragments are then labeled at the Y end of the HindlIl site, visualizing the upper strand of MSI and the lower strand of MS2.

is a reduced chance of finding the correct site within any given restriction fragment. Even if the preferred binding site is present, a proper analysis of the specificity should examine the binding to related sequences, which may also not be present. In addition, although many ligands specifically recognize one or two bases, their binding affinity may be influenced by the nature of the surrounding bases. For these reasons one of us has prepared a multisite DNA footprinting substrate that contains all 136 tetranucleotide sequences. 4 It is freely available on request as a footprinting standard. The sequences of tyrT and the multisite fragment are presented in Fig. 1. This fragment is used to illustrate the various footprinting reactions described below. Plasmids harboring suitable DNA fragments are often prepared by cesium chloride/ethidium bromide density gradient centrifugation. However, this is expensive and time consuming and the ethidium must be removed at a later stage. Because the final stage in the preparation of a radiolabeled restriction fragment usually involves purification from a polyacrylamide gel the initial plasmid preparation need not be of the highest quality, but needs only to be of sufficient purity 4 M. Lavesa and K. R. Fox, Anal. Biochem. 293, in press (2001).

[20]

FOOTPRINTING STUDIES

415

to enable restriction enzyme digestion and polymerase activity. We routinely use Qiagen (Valencia, CA) or Promega (Madison, WI) Wizard minipreparations for extraction of pUC plasmids from 5-ml cultures and find that this generates between 100 and 200 #1 of radiolabeled DNA, which is suitable for use in footprinting reactions. DNA fragments can also be prepared by polymerase chain reaction (PCR) amplification. These products can be cut with suitable restriction enzymes and labeled at the 3' end. Alternatively, if the 5' end of one primer is blocked, then the PCR products can be directly labeled at the 5' end of the other strand. Another approach is to use a radiolabeled primer directly in the PCR mixtures. R a d i o l a b e l i n g DNA DNA fragments can be labeled at either the 3' or 5' ends. 3' Labeling is generally preferable for DNase I and hydroxyl radical footprinting as the products of digestion comigrate with Maxam-Gilbert sequence markers (see below), whereas 5' end labeling is most convenient for digestion with micrococcal nuclease and DNase II. Chemical cleavage agents, such as DEPC and OsO4, can be used with either. 5' End labeling is achieved with polynucleotide kinase and [y-32p]ATP after removal of the native Y-phosphate group with alkaline phosphatase (not necessary with synthetic oligonucleotides). 3' End labeling is achieved by filling in 3' ends left by digestion with restriction enzymes, using enzymes such as DNA polymerase (Klenow fragment) or reverse transcriptase. Enzymes DNase 1

DNase I was the cleavage agent used in the original footprinting experiments. J This enzyme is a double strand-specific endonuclease that produces single-strand nicks, cutting the O3'-P bond. 5 The enzyme requires divalent cations and has optimal activity in the presence of magnesium and calcium. 6 Manganese has been shown to stimulate the action of DNase I without altering the cleavage pattern and it is often added to buffers containing magnesium. 2 Nonetheless, the enzyme is also active in the presence of calcium as the sole divalent cation. 7 It cuts all phosphodiester bonds, and does not display any simple sequence selectivity. However, DNase I cleavage patterns are often uneven, reflecting sequence-dependent variations in local DNA structure. 2 The enzyme binds by inserting an exposed loop into

5 M. Laskowski. in "The Enzymes" (P. D. Boyer, ed.), Vol. 4, p. 289. Academic Press, London, 1971. 6 p. A. Price, J. Biol. Chem. 250, 1981 (1975). 7 K. R. Fox, Biochem. Biophys. Res. Commun. 155, 779 (1988).

416


[9,0]

the DNA minor groove, bending the DNA toward the major groove. 8,9 As a result An-T~ tracts are poor substrates for the enzyme because they possess a narrow minor groove. GC-rich regions are also poor substrates, as they have a more rigid structure that impedes the bending that is thought to be an essential feature of the enzymatic reaction.l° The DNA-binding surface of DNase I covers about 10 base pairs, explaining why the enzyme leads to overestimations of the size of the drugbinding site. Closest phosphates positioned across the DNA minor groove are not attached to a single base pair, but are staggered by two or three bases. As a result, DNase I footprints on opposing strands are staggered by two or three bases in the 3' direction relative to each other. DNase I cuts the O3'-P bond, generating products having a phosphate at the 5' terminus and hydroxyl at the 3' end. In contrast, Maxam-Gilbert cleavage generates fragments that terminate with phosphate groups at both 3' and 5' ends.lJ For 3' end-labeled DNA substrates Maxam-Gilbert markers comigrate with the products of DNase I cleavage, because both types of labeled fragments possess 5'-phosphates. However, for 5' end-labeled DNA the radiolabeled products terminate with a 3'-hydroxyl group for DNase I but a T-phosphate for Maxam-Gilbert markers. As a result Maxam-Gilbert marker lanes, with the extra phosphate, migrate faster than corresponding DNase I cleavage products. This difference is often overlooked when footprinting with 5' endqabeled fragments, but is significant for short fragments, where the difference in mobility may be as great as two bands. In contrast, enzymes that cut the O5'-P bond, such as micrococcal nuclease or DNase II, generate products that comigrate with Maxam-Gilbert markers when labeled at the 5' (but not the 3') end. Protocol. DNase I (purchased from Sigma, St. Louis, MO) is dissolved in 0.15 M NaC1 containing 1 mM MgC12, and stored (at a concentration of about 7200 Kunitz units/ml) at - 2 0 °. This stock solution is stable to frequent freezing and thawing and is diluted to working concentrations immediately before use. Mix 2 #1 of radiolabeled DNA (dissolved in 10 mM Tris-HC1, pH 7.5, containing 0.1 mM EDTA at a concentration of > 10 cps as determined on a hand-held Geiger counter) with 2 #1 of ligand (dissolved in a suitable buffer such as 10 mM Tris-HCl, pH 7.5, containing 10 mM NaC1). It is usual to examine a range of ligand concentrations both above and below the minimum concentration required to generate a footprint. The ligand concentration is usually taken as that in the mixture before adding the enzyme. Strictly speaking, for ligands that are in rapid exchange with DNA, the concentration should refer to that after adding the enzyme, whereas for those in slow exchange (e.g., triplexes) the concentration should 8 A. Lahm and D. Suck, J. Mol. Biol. 221,645 (1991). 9 S. A. Weston, A. Lahm, and D. Suck, J. Mol. Biol. 226, 1237 (1992). ") M. E. Hogan, M. W. Robertson, and R. H. Austin, Proc. Natl. Acad. Sci. U.S.A. 86, 9273 (1989). It A. M. Maxam and W. Gilbert, Methods Enzymol. 65, 499 (1980).

[20]

FOOTPRINTINGSTUDIES

417

refer to that before adding the enzyme. It all depends on how much reequilibration of the ligand-DNA complex may be expected to occur during the time after the small dilution caused by adding enzyme. The mixture is then left to equilibrate for an appropriate length of time (usually less than 30 rain for small molecules, although it may need to be much longer for triplexes). Digestion is initiated by adding 2 #1 of DNase I (dissolved in 2 mM MgCI2, 2 mM MnC12, 20 mM NaC1). The enzyme concentration is usually determined empirically, and will depend on the reaction conditions (pH, ionic strength, temperature, etc.) as well as the DNA concentration. However, at 20 ° with 10 mM NaC1 and pH 7.5 a suitable enzyme concentration is about 0.03 Kunitz units/ml. The reaction is terminated after 1 rain by adding 3.5 ~1 of 80% (v/v) formamide containing 10 mM EDTA, 2 mM NaOH, and 0.1% (w/v) bromphenol blue. EDTA is the essential component in this stop solution, as it chelates the divalent metal ions that are necessary for DNase I activity. After digestion samples are denatured by heating to a temperature near boiling, followed by rapid cooling on ice. Samples can be loaded directly from the near boiling conditions, although heating for excessive periods can lead to depurination. The products of digestion are then resolved on a denaturing polyacrylamide gel (typically 6-12%, depending on the length of the DNA fragment). R e a c t i o n Conditions. DNase I is active over a wide range of experimental conditions and can tolerate temperatures between 4 and 70 °, ionic conditions up to 1 M, and pH values between 4.5 and 9.0. DNase I is also active in the presence of various organic solvents, which are often necessary for dissolution of insoluble ligands. A 2% (v/v) concentration of methanol or dimethyl sulfoxide (DMSO) can be included in the reaction mixture without significantly affecting enzyme activity. DMSO concentrations as high as 50% (v/v) can be tolerated, although under these conditions the cleavage pattern becomes more even as a result of changes in DNA structure. Some ligands with low sequence selectivity do not produce DNase I footprints because they are in rapid exchange with DNA. In some instances footprints can be induced by lowering the temperature, thereby decreasing the dissociation rate constant(s). 12 Although enzyme concentrations under different conditions are usually empirically determined, approximate values for relative enzyme concentrations are shown in Table I. Similarly, magnesium can be replaced with calcium, although about 10-fold more enzyme is required in the absence of magnesium. By contrast, the enzyme is inhibited by millimolar concentrations of cobalt or zinc. Some workers include a known concentration of unlabeled carrier (calf thymus) DNA in the reaction mixture. This enables accurate determination of the total DNA concentration, because the concentration of radiolabeled DNA is usually vanishingly low, and therefore produces consistent extents of cleavage with different preparations of radiolabeled DNA. However,

12K. R. Fox and M. J. Waring,Nucleic Acids Res. 15, 492 (1987).

418


[20]

TABLE 1 RELATIVE CONCENTRATIONSOF DNase I THAT ARE REQUIREDTO PRODUCE EQUIVALENTAMOUNTSOF CLEAVAGEUNDERVARIOUSREACTIONCONDITIONSa Temperature (°C) 4 20 37 50 65

Relative concentration

Ionic strength


10 1 0.5 1 2

0.01 0.1 1.0

1 5 20

pH


5.0 6.0 7.0 8.0

50 3 1 1

a The values are relative to the concentration required to produce single-hit kinetics in pH 7.5, 10 mM NaCI at 20 °.

this can complicate the quantitative analysis of footprinting, and can usually be cmitted. Limitations and Advantages. DNase I is the most commonly used footprinting agent because it is inexpensive and simple to use. Digestion products can be loaded directly onto polyacrylamide gels without further purification, precipitation, or extraction. The major problem with using DNase I as a footprinting probe is that it produces an uneven ladder of cleavage products, so that the boundaries of the footprint may be difficult to define. In addition, because the enzyme is much larger than most DNA-binding agents it overestimates the ligand-binding site size. A further complication is that enhanced cleavage is often observed around the ligand-binding site, especially in regions such as An. Tn, which are normally poor substrates for the enzyme.13 This is also evident in triplex footprints, where cleavage of the bond at the 3' end of the target purine strand is usually (but not always) enhanced. 14'15 These enhancements have often been interpreted as arising from ligand-induced changes in the local structure of the DNA helix that render it more susceptible to enzyme cleavage. However, it should alsobe noted that apparent intensification of cleavage can also arise from changes in the ratio of free DNA to enzyme. ~6 Because DNase I cuts from the minor groove it might be expected that it would detect only the binding of ligands that interact with the minor groove. However, it is one of the most successful footprinting probes for detecting triplex formation in the DNA major groove. 15'17 This may arise from oligonucleotideinduced changes in either DNA structure or flexibility. Surprisingly, DNase I can

13 C. M. L. Low, H. R. Drew, and M. J. Waring, Nucleic Acids Res. 12, 4865 (1984). 14 T. J. Stonehouse and K. R. Fox, Biochim. Biophys. Acta 1218, 322 (1994). 15 M. D. Keppler and K. R. Fox, Nucleic Acids Res. 25, 4464 (1997). 16 B. Ward, R. J. Rehfuss, J. D. Goodisman, and J. C. Dabrowiak, Nucleic Acids Res. 16, 1359 (1988). 17 K. R. Fox, E. Flashman, and D. Gowers, Biochemistry 39, 6714 (2000).

[20]


419

be used to determine accurately the 3' end of triplex-binding sites on the purinerich target strand, and the enzyme has been successfully employed for comparing the binding sites of closely related oligonucleotides. ~7 Example. Figure 2 shows representative DNase I footprints for echinomycin (a bifunctional intercalator that is known to bind to the dinucleotide step CpG ~3'~8,t9) and Hoechst 33258 [a minor groove-binding ligand that is selective for (A/T)4 tracts2°'21]. The fragment used in this study is the multisite fragment (Fig. 1), which contains all possible tetranucleotide sequences. This fragment has been cloned into the BamHI site of pUC18 in both orientations; labeling the 3' end of the HindlII site visualizes the top strand of MS 1, while for MS2 the bottom strand is visualized. The use of these two clones together enables clear data to be obtained for both ends of this relatively long footprinting substrate. DNase I cleavage of the drug-free controls reveals an uneven cleavage ladder in which some bonds are cut more efficiently than others. The unevenness is less pronounced with this fragment than with tyrT DNA because it was designed to avoid the presence of oligopurine - oligopyrimidine tracts. Inspection of the cleavage patterns reveals that, at these concentrations, echinomycin induces footprints around each of the CpG steps. At several locations this is accompanied by enhanced cleavage, presumably reflecting drug-induced changes in the local DNA structure caused by bis-intercalation of this ligand. Hoechst 33258 produces footprints at the (A/T)4 tracts, although as previously noted TTAA, TATA, and TTAT do not present good binding sites for this ligand. 21 It should also be noted that Hoechst 33258 does not produce any regions of enhanced DNase I cleavage, presumably because it does not distort the DNA on binding. Differential cleavage plots derived from these data are presented in Fig. 6 (see below). Other Enzymes

DNase H DNase II is also a double strand-specific nuclease that produces single-strand breaks in DNA. 22 It cuts the O5'-P bond in a reaction that does not require the presence of divalent metal ions. As a result, cleavage with this enzyme cannot be terminated by addition of EDTA. A further disadvantage of DNase II is that it requires conditions of low pH. Much less is known about the structural preferences 18 M. W. Van Dyke and P. B. Dervan, Science 225, 1122 (1984). 19 G. Ughetto, A. H. J. Wang, G. J. Quigley, G. van der Marel, J. H. van Boom, and A. Rich, Nucleic Acids Res. 13, 2305 (1985). 20 M. A. Carrondo, M. Coil, J. Aymami, A. H.-J. Wang, G. A. van der Marel, J. H. Van Boom, and A. Rich, Biochemistry 28, 7849 (1989). 2t A. Abu-Daya, P. M. Brown, and K. R. Fox, Nucleic Acids Res. 23, 3385 (1995). 22 G. Bernardi, in "The Enzymes" (P. D. Boyer, ed.), Vol. 4, p. 271. Academic Press, London, 1971.

420


MS1 Echy

GC~

[20]

MS2 Hoechst

Echy

Hoechst

CCGG ACGT ACGC

ACGG

TCGCGC ACGA

TCGG

TO

CCGC

FIG. 2. DNase I digestion patterns of fragments MS l and MS2 in the presence of echinomycin or Hoechst 33258. The DNA fragments were each labeled at the 3' end of the HindIIl site, so that the sequence runs 3'---~5' from bottom to top of the gel. Fragment MS I corresponds to the top strand of the sequence shown in Fig. 1, while MS2 corresponds to the bottom strand. Ligand concentrations (micromolar) are shown at the top of each gel lane. The bars indicate the location of the potential ligand-binding sites: CpG for echinomycin and (A/T)4 for Hoechst 33258. Tracks labeled GA are Maxam-Gilbert sequence markers specific for purine bases.

[20]


421

of this enzyme, although it produces a less even pattern of cutting than DNase I. 23 Regions of good cleavage do not correlate across the two strands and it cuts best in oligopurine tracts that contain both A and G. This enzyme has been used only rarely as a footprinting probe. 13 Protocol. DNase II (purchased from Sigma) is dissolved in 20 mM ammonium acetate, pH 5.4, containing 1 mM EDTA and stored (at a concentration of about 1000 units/ml) at 20 °. Mix 2 #1 of radiolabeled DNA (dissolved in 10 mM TrisHC1, pH 7.5, containing 0.1 mM EDTA) with 2 #1 of ligand (dissolved in DNase II digestion buffer: 20 mM ammonium acetate, pH 5.4, containing 1 mM EDTA). Digestion is initiated by adding 2 #1 of DNase II diluted in 20 mM ammonium acetate, pH 5.4, containing 1 mM EDTA. The enzyme concentration is usually determined empirically but will typically be about 1 unit/ml. The reaction is stopped after 3 min by adding 4 #1 of formamide containing 10 mM EDTA and 0.1% (w/v) bromphenol blue. Because DNase II does not require a divalent metal ion, this is not sufficient to inactivate the enzyme and further cleavage is prevented either by placing the reaction on dry ice or by immediately heating to 100 °. Micrococcal Nuclease

Micrococcal nuclease, which is widely used for digesting chromatin to prepare nucleosome core particles, cleaves the O5'-P bond. 24 Cleavage is almost exclusively located at pA and pT bondsY '26 It is thought to bind short regions of single-stranded DNA in a cleft on the enzyme, thereby explaining its preference for cleaving at AT residues. 27 Because of its AT sequence selectivity it is not widely used as a footprinting probe, but it can be used for detecting ligand-induced changes in DNA structure and dynamics. 2~ Protocol. Micrococcal nuclease (purchased from Sigma) is dissolved in 50 mM Tris-HCl, pH 7.6, containing 2 mM CaCI2 and stored (at a concentration of about 2000 units/ml) at - 2 0 °. Mix 2 #1 of radiolabeled DNA (dissolved in 10 mM Tris-HCl, pH 7.5, containing 0.1 mM EDTA) with 2 #1 of ligand (dissolved in a suitable buffer such as 10 mM Tris-HC1, pH 7.5, containing 10 mM NaC1). Digestion is initiated by adding 2 #1 of micrococcal nuclease diluted in 50 mM Tris-HC1, pH 7.6, containing 2 mM CaC12. The enzyme concentration is usually determined empirically and will typically be about l unit/ml. Because micrococcal nuclease possesses exonuclease as well as endonuclease activity it is especially important not to overdigest 23 H. R. Drew, J. Mol. Biol. 176, 535 (1984). 24 A. Deiters, J. C. Galluci, and R. R. Holmes, J. Am. Chem. Soc. 104, 5457 (1982). 25 W. Horz and W. Altenburger, Nucleic Acids Res. 9, 2643 ( 1981 ). 26 C. Dingwall, G. E Lomonossoff, and R. A. Laskey, Nucleic Acids Res. 9, 2659 ( 1981 ). 27 F. A. Cotton, E. E. Hazen, and M. J. Legg, Proc. Natl. Acad. Sci. U.S.A. 76, 2551 (1979). 28 K. R. Fox and M. J. Waring, Biochim. Biophys. Acta 909, 145 (1987).

422


[9~0]

the samples. The reaction can be stopped after 3 min by adding 4 ~tl of formamide containing 10 mM EDTA and 0.1% (w/v) bromphenol blue. Chemical Cleavage Agents

Hydroxyl Radicals The hydroxyl radical, a small, highly reactive and freely diffusible species, generates an even ladder of cleavage products in the absence of added ligand and can be used to assess accurately the size and location of a DNA-binding site as well as to glean some information on local DNA structural variations. 29,3° The hydroxyl radical probe is generated by the Fenton reaction: Fe 2+ + H202 --+ Fe 3+ + OH- + .OH The Fe z+ is usually complexed with EDTA, generating a negatively charged species, thereby preventing direct interaction of Fe 2+ with the DNA. Although the exact mechanism of hydroxyl radical cleavage is still not known with certainty, it is generally accepted that the free radicals attack the DNA backbone at the C4' or C 1' positions, leading to subsequent base removal and strand scission. 3k32 Most of the cleavage products contain 5'- and 3'-phosphate ends. However, a small amount of 3'-phosphoglycolate is formed as well; this product migrates slightly faster than the T-phosphate and is therefore apparent as a weaker additional band when examining short fragments that are labeled at the 5' end. As a consequence hydroxyl radical cleavage patterns are often cleaner for 3' endlabeled DNA fragments. Because hydroxyl radicals are thought to attack DNA from positions that are accessible from the minor groove, this cleavage agent is most useful for detecting the interaction of proteins and ligands that bind in the minor groove. It is a great advantage that hydroxyl radicals produce even DNA cleavage ladders, with little sequence dependence. Single- and double-stranded nucleic acids are cut with similar efficiency.33'34 Several studies have employed hydroxyl radical footprinting to examine the interaction of AT-selective minor groove-binding ligands with both natural 35 and synthetic 21'36 DNA fragments. These substances produce footprints about three base pairs long, which are staggered by about two bases in the 3' direction across

29 T, D. Tullius, B. A. Dombroski, M. E. A. Churchill, and L. Kam, Methods Enzymol. 155, 537 (1987). 3o M. A. Price and T. D. Tullius, Methods Enzymol. 212, 194 (1992). 3t W. K. Pogozelski, J. McNeese, and q\ D. Tullius, J. Am. Chem. Soc. 117, 6428 (1995). 32 W. K. Pogozelski, T. J. McNeese, and T. D. Tullius, J. Biomol. Struct. Dyn. 8, 167 (1991). 33 M. J. Jezewska, W, Bujalowski, and T. M. Lohman, Biochemistry 28, 6161 (1989) [see correction in Biochemistry 29, 5220 (1990)]. 34 D. W. Celander and I". R. Cech, Biochemistry 29, 1355 (1990). 35 j. Portugal and M. J. Waring, FEBS Lett. 225, 195 (1987). 36 A. Abu-Daya and K. R. Fox, Nucleic Acids Res. 25, 4962 (1997).

[20]

FOOTPRINTINGSTUDIES

423

the two DNA strands, typical of agents that cut from the DNA minor groove. It has been suggested that at some sites cleavage of one strand is affected more than its complement, possibly providing information about the strand against which the ligand molecules are most closely stacked. 35 The interaction of mithramycin with GC-rich sequences has also been probed by hydroxyl radical footprintingS and has been used to provide information about variations in the location of the antibiotic when bound to different arrangements of GC residues. 38 In some instances hydroxyl radical footprinting was able to detect multiple ligand-binding sites in regions where DNase I produced a single footprint. Protocol. Mix 3/zl of radiolabeled DNA (dissolved in 10 mM Tris-HCl, pH 7.5, containing 0.1 mM EDTA at a concentration of > 10 cps as determined on a hand-held Geiger counter) with 10 #1 of ligand (dissolved in a suitable buffer such as 10 mM Tris-HC1, pH 7.5, containing 10 mM NaC1). The mixture is left to equilibrate for an appropriate time. Digestion is initiated by adding 8 #1 of freshly prepared "hydroxyl radical mix," which is prepared as follows. Stock solutions of the following are prepared in the best quality water available: 100 mM ferrous ammonium sulfate (stored frozen at - 2 0 °), 0.5 M EDTA (pH 8.0), 100 mM ascorbic acid, 30%(v/v) hydrogen peroxide (Sigma). From these the following dilutions are made: Solution Solution Solution Solution

A: Add 2 #1 of 100 mM ferrous ammonium sulfate to 1 ml of water. B: Add 4 #1 of 0.5 mM EDTA pH 8.0, to 1 ml of water. C: Add 100 ttl of 100 mM ascorbic acid to 1 ml of water. D: Add 10 #1 of 30% (v/v) hydrogen peroxide to 1 ml of water.

Prepare the hydroxyl radical mix by mixing solutions A, B, C, and D in the ratio 1 : 1 : 2 : 2 and use immediately. Hydroxyl radical digestion is conducted for 10 min and stopped by adding 2 #1 of 3 M sodium acetate and 65 #1 of ethanol. The samples are left for 10 min on dry ice, before spinning at 13,000 rpm for 10 min in an Eppendorf centrifuge. The supernant is removed and the pellet is washed with 100 #1 of 70% (v/v) ethanol before drying. The radiolabeled DNA pellet is then redissolved in 8 #1 of gel loading buffer [formamide containing l0 mM EDTA, 1 mM NaOH, and 0.1% (w/v) bromphenol blue]. Samples are boiled for 3 min, before loading onto a gel as described for DNase I. Limitations. Hydroxyl radical footprinting offers a powerful means of identifying ligand- and protein-binding sites on DNA and for assessing variations in local helical structure. However, it has some limitations that need to be taken into consideration. First, hydroxyl radicals can only be used to assess the interaction with agents that bind in the DNA minor groove. Compounds that do not occlude the minor groove do not produce hydroxyl radical footprints. As a result, hydroxyl radicals cannot be used for detecting the formation of DNA triple helices. Second, 37B. M. G. Cons and K. R. Fox, NucteicAcids Res. 17, 5447 (1989). 38M. L. Carpenter,J. N. Marks, and K. R. Fox, Eur J. Biochem, 215, 561 (1993).

424

C H E M I C AAND L MOLECULARBIOLOGICALAPPROACHES

[20]

some sequence-selective ligands including actinomycin, echinomycin, and nogalamycin do not generally produce hydroxyl radical footprints. 39'4° The reasons for this are not clear, because these agents produce footprints with other footprinting probes. A further limitation in using hydroxyl radicals is that the reaction is sensitive to the presence of free radical scavengers. Agents such as glycerol (used as a stabilizer in many protein preparations) and dimethyl sulfoxide or methanol (used to dissolve insoluble drugs) inhibit the cleavage reaction and therefore in their presence higher concentrations of EDTA-Fe (iI) are required. E x a m p l e . Representative hydroxyl radical footprints for the interaction of Hoechst 33258 with the multisite DNA fragment are shown in Fig. 3. It can be seen that hydroxyl radicals produce an even cleavage in the drug-free controls, Although hydroxyl radicals produce a more precisely defined assessment of the ligand-binding site than does DNase I, the footprints themselves are less clear and appear only as attenuations in the local cleavage pattern. As a result the footprints are best identified from densitometric analysis of the data; examples are shown in Fig. 4. Inspection of these plots reveals that the ligand protects four bands from cleavage around each (A/T)4 site, and that two bonds at the center of each footprint are more strongly affected. The footprints are staggered toward the 3' end of the binding site, as expected for a cleavage probe that attacks from the minor groove. Rather surprisingly, this footprinting tool is able to detect binding to some sites (such as TTAA and TTAT) that were not evident with DNase I (Fig. 2). Methidiumpropyl-EDTA-Fe( ll )

The chemical cleavage agent methidiumpropyl-EDTA-Fe(II) was devised by Dervan and colleagues. 4j'42 It consists of an intercalating moiety (methidium, originally known as dimidium) tethered to a moiety of EDTA. In the presence of Fe 2+ and under reducing conditions, free radicals are generated that cleave the DNA. MPE produces a fairly even ladder of cleavage products although, because its action is based on intercalation of its phenanthridinium chromophore, An" Tn tracts, which are poor intercalation sites, show attenuated cleavage. Although M P E Fe(II) is widely used as a footprinting probe for studying the binding of proteins and small molecules to DNA there have been surprisingly few studies in which it has been used for examining triplex formation, 17,43,44 where it produces less clear footprints than does DNase I. It is likely that this footprinting tool does not reveal clear triplex sites because it binds to triplex as well as duplex DNA and perturbs the triplex-duplex equilibrium, as does the parent molecule ethidium. 39M. E. A. Churchill, J. J. Hayes, and T. D. Tullius, Biochemistry 29, 6043 (1990). 4oK. R. Fox, Anticancer Drug Des. 3, 157 (1988). 41 M. W. Van Dyke and R B. Dervan, Nucleic Acids Res. 10, 5555 (1983), 4z M. W. Van Dyke, R. R Hertzberg, and R B. Dervan, Proc. Natl. Acad. Sci. U.S.A. 79, 5470 (1982). 43 1~A. Beal and E B. Dervan, J. Am. Chem. Soc. 114, 4976 (1992). 44C. Marchand, C. Bailly, C. H. Nguyen, E. Bisagni, T. Garestier, and C. H61~ne,Biochemistry 35, 5022 (1996).

[20]

FOOTPRINTING STUDIES MS1

TTAA TATA A ATTA

425

MS2

TTTA, ATAA TAATT TATA

TTAA TTAT TAAA

TTTT

AAAT

TATT

ATAT

FIG. 3. Hydroxyl radical digestion of the multisite fragment in the presence of Hoechst 33258. The fragments were labeled at the 3' end of the HindlIl site, revealing the top strand of Fig. 1 for MS1 and the bottom strand for MS2. Ligand concentrations (micromolar) are shown at the top of each gel lane. The bars indicate the location of the (A/T)4 sites. Tracks labeled GA are Maxam-Gilbert sequence markers specific for purine bases.

Protocol. Radiolabeled D N A (2 #1) is m i x e d with ligand (4 #1) diluted in an appropriate buffer and left to equilibrate. Five microliters of a solution containing 1 # M M P E (Sigma) and 1 /zM ferrous a m m o n i u m sulfate is then added and the mixture is left to equilibrate for a further 5 min. C l e a v a g e is initiated by adding 3 #1 o f 10 m M dithiothreitol ( D T T ) and stopped after 60 min by ethanol precipitation, in the presence o f 0.3 M sodium acetate.

426


[20]

MS1 TTTT

TATT

ATTAAATAT 2,~ ¢v ~..1~• ~ ,'

vv"~j V

MS2

Hoechst

TATA

i

TTAT

TAAA I . . . . .

AATT ,A. 1,1~,£~i.

T_TAAT~ ATAT • IIJ[ , A. A )

.A~AA~,A~AI~AA~I~lt~/~, ,~,v,'w~,~" ~ V

JControl FIG. 4. Densitometric analysis of the hydroxyl radical cleavage plots shown in Fig. 3 in the presence of 3 # M Hoechst 33258. Note that sequences are written 3'---*5' so that the left-hand side corresponds to the bottom of the gel.

Potassium Permanganate and Osmium Tetroxide The reagents potassium permanganate and osmium tetraoxide oxidize thymines 45,46 (and to a lesser extent cytosines) by reaction across the C5-C6 double bond. Double-stranded DNA is resistant to this oxidation because the stacked bases do not permit the out-of-plane attack by the oxidizing agent. The reagents are therefore much more reactive toward single-stranded DNA, or regions in which the C5-C6 bond of thymine is exposed. Osmium tetroxide has been used to detect echinomycin-induced changes in DNA structureY whereas permanganate oxidizes certain thymines in the presence of bleomycin. 4s The exact molecular basis for these changes in reactivity is not known, but it is thought to involve ligandinduced helix perturbations that may well include extension and unwinding. The permanganate reaction is performed by mixing 3 #1 of DNA with 3 #1 of ligand in an appropriate buffer (10 mM Tris-HCl, pH 7.5), followed by addition of 3/zl of 50 mM KMnO4. The reaction is stopped after 3 min by adding 2 #1 of 2-mercaptoethanol, and the DNA is precipitated by adding 10 #1 of 3 M sodium acetate, 75 #1 of water, and 300 #1 of ethanol. After washing and drying the pellet, 45 E. Palecek, Methods Enzymol. 212, 139 (1992). 46 C. M. Rubin and C. W. Schmid, Nucleic Acids Res. 8, 4613 (1980). 47 M. J. McLean, E Seela, and M. J. Waring, Proc. Natl. Acad. Sci. U.S.A. 86, 9687 (1989). 48 K. R. Fox and G. W. Grigg, Nucleic Acids Res. 16, 2063 (1988).

[20]

FOOTPRINTINGSTUDIES

427

strand cleavage is achieved by adding 100/zl of 10% (v/v) piperidine and boiling for 30 min. Diethyl Pyrocarbonate

Diethyl pyrocarbonate (DEPC) reacts predominantly with N7 of purines (located in the DNA major groove of B-DNA) so that subsequent treatment with piperidine leads to strand scission. Native duplex DNA is relatively unreactive to this reagent, so it provides a means of examining DNA structures in which the N7 atom may be more exposed than usual. It has therefore been used to detect the local unwinding adjacent to drug intercalation sites. 49 Echinomycin induces DEPC-hyperreactive sites at adenines 3' to the drug-binding site (i.e, at CGA), 5° whereas bleomycin enhances DEPC reactivity at adenines in the sequence GYA. 45 The reaction is performed by mixing 3 #1 of DNA with 3 #1 of ligand in an appropriate buffer (10 mM Tris-HC1, pH 7.5). The reaction is started by adding 5/zl of DEPC. Because this reagent is immiscible with water it is necessary to agitate the tube at regular intervals. The reaction is stopped after 20 min by precipitating the DNA with 15/zl of 0.5 M sodium acetate and 60 #1 of ethanol. After washing and drying the pellet, strand cleavage is achieved by adding 100/zl of 10% (v/v) piperidine and boiling for 30 min. Example. The effects of echinomycin on DEPC and permanganate modification of the multisite fragment are shown in Fig. 5. In the drug-free controls DEPC produces weak cleavage at purine residues. In the presence of echinomycin a few bands are rendered more susceptible to reaction with DEPC. These all correspond to adenines and the strongest are located on the 3' side of echinomycin-binding sites at TCGA and ACGA. Other adenines that are not close to echinomycinbinding sites are also affected, although to a lesser extent. These may correspond to long-range changes in DNA structure that are transmitted into regions that are several nucleotides removed from the ligand-binding site. Alternatively, they may indicate the location of secondary ligand-binding sites. It is possible that the interaction of echinomycin with these secondary sites is too transient to produce a footprint, but DEPC may be able to detect the unwound DNA structure associated with these short-lived species. It should also be noted that echinomycin does not affect the reaction of DEPC with CCGA and GCGA, suggesting that the ligand has less effect on the structure of sites that are flanked by G and C residues. Figure 5 also shows that echinomycin affects the modification of several thymines by permanganate. Although one of the strongest sites of enhancements occurs at TCGT, most of the other enhancements are not close to CG sites. In

49C. Jeppesenand E E. Nielsen,FEBS Lett. 231, 172 (1988). 5oj. Portugal, K. R. Fox, M. J. McLean,J. L. Richenberg,and M. J. Waring, Nucleic Acids Res. 16, 3655 (1988).

428


[9,0]

MS2

MS1 DEPC KMBO4

DEPC

KMBO,

.

.

.

.

-~ 0 2

~" 005

"-.

'm.

~01

0 O0 10

20 [paromomyein],

I

I

I

O0

30

0

10

~tM

I

20

30

[paromomy¢in], ~tM

I

0

4

~

"~ 021

00

. - ' " " 0

10

:i1'

" 20

30

[paromomycin], I~M

40

0

"-:':-] 10

20

30

40

[paromomy¢in], ~tM

FIG. 7. Selected footprinting plots (positions 301-358) from cleavage of the 176-met with RNase 1. Refer to Fig. 8 for the locations of the sites on the 176-mer.

analyzing RNA secondary structure2j was applied to the 176-mer. This produced the lowest energy RNA structure, shown in Fig. 8. A more detailed structural analysis of the 176-mer in the absence of drug is under way.

Paromomycin Paromomycin is a member of the aminoglycoside class of antibiotics. 22 The drug contains a central 2-deoxystreptamine ring to which are appended monoand disaccharide units (Fig. 2). The five cationic protonated amine groups of paromomycin give the drug solubility in water and provide it with a high electrostatic attraction for RNA. Relatively free rotation about the glycosidic bonds allows the drug to alter its shape and this, along with numerous possibilities for 21 D. H. Mathews, J. Sabina, M. Zuker, and D. H. Turner, J. Mol. Biol. 288, 911 (1991). 22 R. Schroeder, C. Waldsich, and H. Wank, EMBO J. 19, 1 (2000).

[21 ]

D R U G - R N A FOOTPRINT1NG

445

TABLE I EFFECT OF PAROMOMYCINON FOOTPRINTING PLOTSa RNase TI

RNase T 1 Site number G223 G224 G226 A235 G237 C238 A239 G241 A242 C243 U244 C245 C248 U249 G251 G254 A255 G257 G259 G261 G266 C267 A268 A269 C274 G275 C284 G285 A286 C287 G289 G290 G292 G294 G298

RNase I

10% gel

RD RD,W M M M M M M M, W N M, W

6% gel

Site number

RNase 1

SD SD SD

C300 A301 A302 U303 U308 U309 G310 G318 G321 G325 A327 G328 G329 A330 G331 A332 G333 A334 G335 A336 G338 G339 G340 G342 G348 G350 U351 A353 U355 U358 A359 G361/' G367 G369

M M M M M M N M

N

SD SD

SD SD

I I I

1 I 1

N, W

RD, W RD, W RD, W RD, W RD, W N N, W M M M, W

N, W

SD M M M

SD M M M

M M M M M

10% gel

6% gel

N

N

1 I

1

N,W

N

SD SD

SD SD

SD, W

SD, W

SD

SD

SD

SD

SD SD SD SD

SD SD SD SD N

M M M

I 1

M M M M 1

M

SD SD SD

N SD SD SD

M, W

a N, Neutral (no change in cleavage with increased drug concentration); I, increasing (increased cleavage with increased drug); RD, rapid decrease (cleavage decreases rapidly with increased drug); SD, slow decrease (cleavage decreases with increased drug, but slowly); M, maximum (cleavage increases initially with increased drug, then goes through a maximum and decreases); W, weak (low intensity). l, These three sites are representative of the region from position 361 through position 369, all of which are G, except positions 362 and 368. It is impossible to resolve individual sites in this region.

446


[21]

hydrogen bonding (donation and acceptance), makes it difficult to generalize a paromomycin-RNA binding model. However, because of its large size the drug is not likely to bind in the major groove of a normal A-form helix such as would occur in stem regions of the 176-mer (Fig. 8). Puglisi and co-workers 23 used nuclear magnetic resonance (NMR) to uncover details of the binding of paromomycin to the A-site in 16S ribosomal RNA. They found that the drug, which spans about five nucleotides of RNA, adopts an "L-shape" and binds to an asymmetric internal loop caused by noncanonical base pairs in the A-site.

Footprinting Analysis of Paromomycin Binding to 176-mer Inspection of the footprinting plots for RNase I and RNase T1 cleavage of the 176-mer reveals that the former, because of its low base/sequence specificity,16 yields more information about the location of paromomycin-binding sites. By varying the electrophoresis time and the composition of the gel it is possible to resolve most of the sites (single nucleotide resolution), but those farthest from the label, especially in the experiments involving RNase I, remain unresolved. The analysis of the footprinting plots for RNase I and RNase T1 is summarized in Table I. In general, information is included only for sites that have significant intensity and for which cleavage changes noticeably with drug concentration. A footprinting gel for RNase I is given in Fig. 3. The kinetics of cutting in the case of RNase I appear to be zero order in enzyme. In this case the rate of cleavage of phosphodiester linkages of RNA is independent of enzyme concentration, so that apparent increases in the rate on addition of drug must be structural in origin. The zero-order kinetic regimen was used in DNase I footprinting experiments involving the anticancer drug actinomycin D. 24 A footprinting gel involving RNase TI is shown in Fig. 4. The kinetic regimen used in the footprinting experiments involving RNase T1 seems to be first order in enzyme. In this case, the rate of cleavage of a phosphodiester linkage, (rate)i, is given by ki [RNase T 1 ] i , where ki iS the cleavage rate constant at the site and [RNase T 1]i is the concentration of enzyme at site i. This rate law implies that drug binding at certain sites will shift enzyme to other sites, increasing (rate)i through a mass action effect and giving rise to enhancements in the footprinting data. Similar mass action effects were earlier observed in DNase I footprinting experiments involving actinomycin D and a 139-mer DNA restriction fragment. 3'25 Some footprinting plots, all for RNase I, are illustrated in Figs. 6 and 7. As can be seen from the "noise" in the plots, the error in any one optical density measurement is about 0.02 OD unit. The optical densities plotted in Figs. 6 and 7 have been corrected by subtracting the optical density in the control lane from that 23 M. 1. Recht, D. Fourmy, S. C. Blanchard, K. D. Dahlquist, and J, D. Puglisi, J. Mol. Biol. 262, 421 ( ! 996)~ 24 K. D. Bishop, P. N. Borer, Y. Q. Huang, and M. J. Lane, Nucleic Acids Res. 19, 871 ( 1991 ). 25 B. Ward, R. Rehfuss, J. Goodisman, and J. C. Dabrowiak, Nucleic Acids Res. 16, 1359 (1988).

[21]

DRUG-RNA FOOTPRINTING 260

~ A

DIS

447

SD ~G u {ZI'.AG

.%

c G

UG'¢ A ~ t ~

~u"G c

SLI

280

SL2

C.o_~

' ~'c~ .c~.G^ ~GGCG.c-300

G GCu'~ A G

24o.G _'Zc

AAA A A

.uu u G

,,AGA',.,, U A

SL3

AOG.cGAAU "~ ~G ~U A G" "G-/ U" G AAGA',;~CG IjO'GG :560 A U .~.gag '~C GG AG

uc~G C "A~ G G A

A.~U A

220,,G.~U AG

A

,

G"~ A~.G

340

320

~.~-.G

SL4 ~ . c G AGA

cA.u G A A C "G A "G A G 380 G G.~C A A

176-mer H I V - ~ - R N A FIG. 8. Energy-minimized structure of the 176-met from the packaging region ~ of HIV- 1 (LAI). The main stem includes positions 213-238 and 361-388; stem-loop 1 (SLI) includes the dimer initiation site (DIS), having the palindromic sequence 5'-GCGCGC-3'; stem-loop 2 (SL2) includes the 5' splice donor site (SD); stem-loop 3 (SL3); stem-loop 4 (SL4) includes the start codon 5'-AUG-3' for the gag gene (circled).

in the drug lane. The control lane shows fragmentation of the 176-mer prior to the addition of enzyme. Because the amount of degradation of the 176-met is minor, there is some uncertainty in locating the position in the control lane corresponding to the position of a given spot in a drug lane. This could lead to an incorrect value of the optical density to be subtracted, and is responsible for some of the optical densities in Figs. 6 and 7 becoming negative by several hundredths of a unit. No correction for loading error, which can be as large as 10% in a given lane, has been made. Because of this, and because of the "noise," it is best to be wary of any interpretation based on a single point in a footprinting plot. The plots for sites 235, 237, and 266-269 show the simple decrease in intensity with drug concentration associated with a drug-binding site. The last four are associated with adjacent cleavage sites, indicating the same binding event. Because optical density drops to zero when drug concentration exceeds 15 #M, the

448


[2 1]

drug-binding constant is large for this site. For other footprinting plots, such as those associated with sites 300-309, optical density is still decreasing after 30/~M, indicating a much smaller drug-binding constant. The plots for sites 348 and 350 (Table I) show steady increases, or enhancements, which, because of the kinetic regimen of the cutting reaction, are most likely structural. Almost all the other plots of Fig. 6 and 7 show a rapid rise to a maximum, followed by a decrease. The rapid rise must be a structural effect; if it were due to mass action, all sites would show a similar rise, and they certainly do not. Note that the rise occurs over the same concentration range as the drug binding at the strongest binding sites, which suggests an association with those drug binding events. Both RNase I and RNase T 1 reveal paromomycin binding via decreased cleavage with added drug. The decreases occur at lower drug concentrations with RNase I than with RNase T1. Because competitive binding between drug and enzyme is being dealt with, the difference may be due to RNase T1 binding more strongly than RNase I to the RNA, so that higher amounts of drug are required to saturate the site. Without more detailed analysis it is not possible to determine whether one or more than one drug molecule is binding in any region. Considering the RNase I results (Table I and Figs. 6 and 7), it can be seen that the highest-affinity binding sites occur near positions 235,237,266--269, and 274. The first two are at a bulge in the main stem of the 176-met, and the others are near a bulge in the stem of SL1. The footprinting plots for many sites in the experiments with RNase I pass through a maximum, indicated by "M" in Table I. This occurs for regions 238-249 (which includes the linker region between the main stem and SLI as well as part of SL 1), 298-318 (the linker region between SL2 and SL3 and part of the stem of SL3), 325-336 (part of SL3 and SL4 as well as the linker region between them), and 351-361 (stem of SL4 plus the linker between SL4 and the main stem). The maxima appear at ~ 5 - 1 0 #M, the concentration range over which the primary binding sites are accepting drug. Thus, binding to the strongest sites appears to cause a change in the structure of the 176-mer. The decrease in cutting at M sites at higher drug concentration, after the maximum in the footprinting plot, shows that they are secondary sites, binding ligand at higher drug concentration. It is also possible that a decrease in cutting at high drug concentration is not due to drug binding but that a second structural change is occurring. This could be caused by a binding event in another part of the 176-mer. Interestingly, non-contiguous nucleotides sometimes appear to be involved in the same binding process. For example, compare the RNase I footprinting plots (Fig. 6) for nucleotides 235 (A) and 237 (G) with those for sites 266-269 (GCAA). Sites 266-269 flank the bulge in SL1, whereas nucleotides 235 and 237 are in a bulge in the main stem. It is possible that these are separate drug-binding events. However, because loading is taking place in the same concentration range for all of the sites (including nucleotide 274), they may be forming a single site that accepts one drug molecule. Site-directed mutagenesis of the 176-mer to selectively

[21]

DRUG-RNA FOOTPRINTING

449

modify sites and complete quantitative analysis of the footprinting plots should help determine which situation is correct. The footprinting plots for sites 298-318 strongly resemble those for sites 325336 (see Table I and Fig. 7). These nucleotides are not contiguous. All the footprinting plots show maxima at about 10 #M, indicating structural enhancements connected to the primary binding events. It is not unreasonable that all the sites are reacting to the same structural change, but it is unusual that the binding (decrease in cleavage after the maximum) is also the same at all sites. Again, it may be that the sites form a pocket that accepts one drug molecule, or it may be that more than one drug site, having by coincidence the same properties, is involved. The RNase T1 footprinting plots show drug binding at sites 223,224, and 226 (a bulge in the main stem of the 176-mer), 251 and 254 (in the stem of SL 1), 289294 (in the stem and loop region of SL2), 328-342 (near bulges in the stems of SL3 and SL4, plus the linker region), and the region 361-369 (near a bulge in the main stem). The footprinting plots for these sites show a decrease in intensity with increasing drug concentration, sometimes after an initial increase. As is evident from the RNase TI experiment shown in Fig. 4 and the data collected in Table I, enhancements are observed in the palindromic loop of SL1, sites 257, 259, and 261. The enhancements could be due to altered structure, or simply a mass action effect, because the kinetics of cutting are believed to be first order for this enzyme. Summary and Conclusions RNase I and RNase T1 can be used to obtain high-quality footprinting information for paromomycin binding to a 176-mer RNA from the packaging region of HIV-1 (LAI). Controls and scanning procedures are necessary for quantitation of autoradiographic data, so that footprinting plots showing cutting behavior as a function of drug concentration can be used to identify binding sites and regions of altered structure on the 176-mer. From the RNase I footprinting results the primary paromomycin binding sites on the 176-mer are on the main stem and on the stem of SL1, but noncontiguous sequences may be involved in the same binding event. Strong enhancements in cleavage with added drug are also observed, indicating drug-induced structural changes. Drug binding may cause linker regions between stem-loops of the 176-mer to change structure, possibly providing a site or sites for additional drug binding. Because drug binding changes the structure of the packaging region, which may alter its function, paromomycin analogs with enhanced specificity for HIV ~RNA have potential as a new class of agent for treating AIDS. Acknowledgments We thank P. N. Borer for the energy-minimized structure of the 176-mer and for helpful discussions pertaining to the research. This work was in part supported by a grant, GM32691, to P. N. Borer.

450


[22]

[22] Footprinting Methods for Analysis of Pyrrole-lmidazole Polyamide/DNA Complexes B y JOHN W. TRAUGER a n d PETER B. DERVAN

Introduction The design of synthetic ligands that read the information stored in the DNA double helix has been a long-standing goal at the interface of chemistry and biology. Cell-permeable small molecules that target predetermined DNA sequences enable a potential approach for the regulation of gene expression. The development of "pairing rules" for minor groove-binding polyamides offers a code to control sequence specificity.l'2 Crescent-shaped polyamides containing three aromatic amino acids, N-methylimidazole (Im), N-methylpyrrole (Py), and N-methyl3-hydroxypyrrole (Hp), bind double-helical DNA with affinities and specificities comparable to those of natural DNA-binding proteins. 3-6 Sequence specificity is controlled by side-by-side pairs of rings that afford unique set(s) of sequencespecific hydrogen bonds with the minor groove edges of the intact Watson-Crick base pairs. 7'8 The pair Im/Py specifies a G- C base pair while Py/Im targets C. G. The Py/Py pair binds both A. T and T. A. Hp/Py and Py/Hp pairs discriminate T. A from A. T.l-8 Although the generality of these pairing rules has been demonstrated by synthesizing a large number of polyamides that bind many different sequences, we find that there are limitations with regard to some sequence contexts, which are probably due to the sequence-dependent microstructure of the DNA helix. Remarkably, eight-ring polyamides have been shown to permeate cells and, in a few encouraging examples, inhibit the expression of target genes in cell culture experiments .9,10

I p. B. Dervan and R. W. Biirli, Curr. Opin. Chem. Biol. 3, 688 (t999). 2 D. E. Wernmer, Annu. Rev. Biophys. Biomol. Struct. 29, 439 (2000). 3 j. W. Trauger, E. E. Baird, and E B. Dervan, Nature (London) 382, 559 (1996). 4 S. White, E. E. Baird, and P. B. Dervan, Chem. Biol. 4, 569 (t997). 5 j. M. Turner, S. E. Swaltey, E, E. Baird, and E B. Dervan, J. Am. Chem. Soc. 120, 6219 (1998). 6 S. White, J. W. Szewczyk, J. M. Turner, E. E. Baird, and E B. Dervan, Nature (London) 391, 468

(1998). 7 C. L. Kielkopf, E. E. Baird, P. B. Dervan, and D. C. Rees, Nat. Struct. Biol. 5, 104 (1998). 8 C. L. Kielkopf, S. White, J. W. Szewczyk~ J. M. Turner, E. E. Baird, P. B. Dervan, and D. C. Rees, Science 288, 111 (1998). 9 j. M. Gottesfeld, L. Neely, J. W. Trauger, E. E. Baird, and P. B. Dervan, Nature (London) 387, 202

(1997). 10 L. A. Dickinson, IL J. Gulizia, J. W. Tranger, E. E. Baird, D. E. Mosier, J. M, Gottesfeid, and P. B. Dervan, Proc. Natl. Aead. Sci. U.S.A. 95, 12890 (1998).

METHODS1NENZYMOLOGY,VOL.340

Copyright© 2001by AcademicPress ~1 rightsof reproduction~nany.formreserved. 0076-6879/00$35.00

[22]

FOOTPRINTINGTO ANALYZEPy-Im/DNA

451

A pivotal step in the discovery--evaluation process of new polyamide motifs is the characterization of the affinity and specificity of next-generation molecules following the design-sythesis phase (i.e., structure-function relationships). In the design phase, it is often the case that the sequence preference, as well as the energetics of any new molecule binding at each potential site, are not perfectly understood and it is crucial to scan "libraries" of many potential DNA-binding sites in order to identify true high-affinity binding sites. Characterization of equilibriumassociation constants must guide the choice of sequence context for DNA : ligand complexes selected for subsequent structure elucidation by nuclear magnetic resonance (NMR) or X-ray methods. A DNA restriction fragment, typically a few hundred base pairs in length, provides a library of contiguous small molecule-binding sites on a string addressed by a 32p end label. The high-resolution separation of nucleic fragments by gel electrophoresis underpinning "footprinting" methods has revolutionized the field of small molecule-DNA discovery. This chapter describes three complementary footprinting methods and the protocols used for analysis of polyamide : DNA complexes: (1) MPE-Fe(II) footprinting,1 l- 14 (2) affinity cleavage,15-J7 and (3) quantitative DNase I footprint titration 18-21 (Fig. 1). Footprinting with methidiumpropylEDTA-Fe(II) [MPE-Fe(II)] is used to identify high-affinity polyamide-binding sites to near nucleotide resolution (Fig. 1A). Affinity cleavage is used to determine the orientation of the bound polyamide in the minor groove of DNA (Fig. 1A). Quantitative DNase I footprinting is used to determine equilibrium association constants (K~) for polyamide-DNA complexes at previously identified match and mismatch sites (Fig. 1B).

Description of Methods Footprinting was first described for analysis of protein-DNA complexes using the enzyme DNase I. 22 DNase I has been used to footprint small molecules as

l! M. W. Van Dyke, R. P. Hertzberg, and P. B. Dervan, Proc. Natl. Acad. Sci. U.S.A. 79, 5470 (1982). J2 M. W. Van Dyke and P. B. Dervan, Cold Spring Harbor Syrup. Quant. Biol. 47, 347 (1982). 13 M. W. Van Dyke and E B. Dervan, Nucleic Acids Res. 11, 5555 (1983). 14 M. W. Van Dyke and E B. Dervan, Science 225, 1122 (1984). 15 j. S. Taylor, E G. Schultz, and E B. Dervan, Tetrahedron 40, 457 (1984). 16 p. G. Sehultz and E B. Dervan, J. Biomol. Struct. Dyn. 1, 1133 (1984). 17 p. B. Dervan, Science 232, 464 (1986). 18 M. Brenowitz, D. E Senear, M. A. Shea, and G. K. Ackers, Methods Enzymol. 130, 132 (1986). 19 M. Brenowitz, D. E Senear, M. A. Shea, and G. K. Ackers, Proc. Natl. Acad. Sci. U.S.A. 83, 8462 (1986). 20 D. F. Senear, M. Brenowitz, M. A. Shea, and G. K. Ackers, Biochemistry 25, 7344 (1986)~ 21 M. Brenowitz, D. E Senear, E. Jamison, and D. Dalma-Weisha~sz, in "Footprinting of Nucleic Acid-Protein Complexes" (A. Revzin, ed.), p. 1. Academic Press, San Diego, California, 1993. 22 D. Galas and A. Schmitz, Nucleic Acids Res. 5, 3157 (1978).

452


[22]

(A)

5'

3'

5'

3'

(B) C

Increasing [Ligancl] B==

1 0.8 Eo.6 a:~ 0 . 4

Site I

Ref[

0.2-

o lO-ll

10.1o

.... 1~0.9 '

l!O.a ',

1~0.7

Increasing [Ligand]

FIG. 1. (A) Schematic of (left) MPE-Fe(II) footprinting and (right) affinity-cleaving techniques. Cleavage products obtained on a denaturing gel with DNA end labeled either on the 5' or 31 strand are shown. (B) LeJ?: Cleavage pattern generated by quantitative DNase I footprint titration, on a 31 endlabeled DNA fragment in the presence of increasing ligand concentrations. Right: Langmuir binding titration isotherm obtained from DNase I data.

well as proteins. 13,23 However, DNase I does not cleave all sequences equally and footprints can appear significantly larger than the actual base pairs occupied by the D N A - b i n d i n g ligand. 13 Both natural product D N A - b i n d i n g ligands, as well as designed polyamides, typically cover only five or six base pairs and accurate assignment of the b i n d i n g site sequence becomes important. The chemical footprinting reagent, M P E - F e ( I I ) , based on the nonspecific intercalator m e t h i d i u m with E D T A - F e attached, cleaves D N A 24,25 with m i n i m a l sequence specificity in 23 K. R. Fox and M. J. Waring, Nucleic Acids Res. 12, 9271 (1984).

10-s

[22]


453

the presence of a reducing agent and 02, and allows a higher resolution footprint. By directly attaching the EDTA-Fe moiety to the DNA-binding ligand, the orientation of a ligand in its complex with DNA can be determined in a technique called affinity cleavage. 15-17 In addition, affinity cleavage experiments are useful for determining the groove location of a ligand, because distinct 5'- or T-shifted cleavage patterns are observed for cleavage in the major versus the minor groove. Minor groove-binding ligands produce 3'-shifted cleavage patterns j7 (Fig. 1A). Quantitative DNase I footprinting procedures were developed by Ackers and co-workers, and are the foundation of the protocol presented here. 18-21 For quantitative footprinting experiments, equilibrium mixtures of 32p end-labeled DNA and a range of polyamide concentrations are subjected to partial digestion with DNase I. The resulting cleavage products are separated by gel electrophoresis and visualized by autoradiography. DNA sites bound by a polyamide are protected from cleavage, resulting in a gap in the ladder of bands on the gel. The tractional protection of a binding site as a function of polyamide concentration can be determined by densitometric analysis of the gel, and these data can be fitted to a theoretical binding isotherm to determine the equilibrium association constant (Ka) (Fig. 1B). The data analysis for this assay assumes that the polyamide concentration is much higher than the DNA concentration, allowing the approximation [L]free = [L]tot, where [L]fre°is the concentration of polyamide free in solution and [L]tot is the total polyamide concentration. The total equilibration volume specified in this protocol is 400 #1 and the DNA concentration is ~5 pM. As a consequence, apparent association constants greater than 4 2 x 10 ~° M ~ should be regarded as lower limits. In the protocol described here, polyamide and radiolabeled DNA are equilibrated in the absence of any unlabeled carrier DNA (such as calf thymus DNA) to avoid underestimation of association constants due to polyamide binding to sites on the carrier DNA. Materials

DNase I [fast protein liquid chromatography (FPLC) pure], calf thymus DNA (sonicated, deproteinized), and NICK columns are purchased from Pharmacia (Piscataway, NJ) Restriction enzymes, polynucleotide kinase, Taq DNA polymerase, Klenow DNA polymerase, and glycogen are obtained from Boehringer Mannheim (Indianapolis, IN). Sequenase (version 2.0) is purchased from United States Biochemicals (Cleveland, OH). [ot-32p]Thymidine 5'-triphosphate (>3000 Ci/mmol) and [ot-32p]deoxyadenosine 5'-triphosphate (>6000 Ci/mmol) are purchased from Du Pont-NEN (Boston, MA). [F-32p]adenosine triphosphate (endlabeling grade) is from ICN (Costa Mesa, CA). Denaturing polyacrylamide gel 24R. E Hertzbergand P. B. Dervan,J. Am. Chem. Soc. 104, 313 (I 982). 25R. R Hertzbergand R B. Dervan,Biochemistry 23, 3934 (1984).

454


[22]

mix, Tris-borate-EDTA (TBE) packets, and gel plates, spacers, and stands are obtained from GIBCO-BRL (Gaithersburg, MD). Presiliconized 1.5-ml tubes are obtained from Sorenson Biosciences, Inc.

P r e p a r a t i o n of 32p E n d - L a b e l e d DNA End-labeled DNA is prepared from plasmid DNA containing the binding site(s) of interest. The desired DNA fragment is generally prepared by restriction digest and 3' fill-in, using a DNA polymerase and [a-32p]dNTPs (Procedures 1 and 2, below), but can also be prepared by polymerase chain reaction (PCR), using a 5/ 32p-labeled primer (Procedure 3, below). When labeling by 3' fill-in, it is best to cut the end to be labeled with an enzyme whose 51 overhang contains only two different nucleotides because the overhang can then be completely filled in with two labeled dNTPs. The DNA-binding sites should ideally be located 40-100 base pairs from the labeled end of the DNA fragment, although sites of up to ~200 base pairs from the label can usually be resolved. The size of the DNA fragment can range from 100 base pairs to more than 500 base pairs.

Procedure l Procedure 1 is the most convenient labeling procedure, and can be used if (l) both enzymes cut efficiently in the same buffer, and (2) the enzyme that cuts at the unlabeled end leaves either a blunt end or a 3' overhang. Enzyme combinations that work well are EcoRI*/PvuII (this example) and AflII*]FspI (the asterisk denotes the labeled end). 1. Combine plasmid DNA (4 #g) and water in a total volume of 55/zl. Add 7 #1 of H buffer (Boehringer Mannheim), 4/~1 of EcoRI (20 units/#l), and 4 #1 of PvuII (20 units/#l). Incubate for 2-4 hr at 37 °. Add 8 ttl of water, 5 #1 of H buffer, 10 #1 of [~-32P]dATR 10/zl of [a-32p]dTTP, and 10 #1 of Klenow DNA polymerase (2 units/#l). Allow the reaction to proceed for 25 rain at room temperature. Add 5 ttl of dATP/dTTP mix (each at 5 mM) and allow the reaction to proceed for 5 rain. 2. Add 30 #1 of 5x Ficoll loading buffer, mix, and load on a nondenaturing 7% (w/v) polyacrylamide gel (two lanes, each ~75 #1). Run the gel and isolate the DNA as described below.

Plvcedure 2 In this example (EcoRI*/HindIII), the second cutter, HindIII, produces a 5' overhang and requires a buffer different from that required by EcoRI. For these reasons, the second cut is done after the first cut and fill-in.

[221


455

1. Combine plasmid DNA (4 #g) and water in a final volume of 46/zl. Add 8/zl of H buffer (Boehringer Mannheim), 10 #1 of [a-32p]dATP, 10/xl of [c~-32p]dTTP, 4 #1 of EcoRI (20 units/ml), and 2 #1 of Sequenase (8-25 units/ml). Incubate at 37 ° for 4 hr. Add 5/zl of dATP/dTTP mix (each at 5 mM) and 0.5 #1 of Sequenase, and incubate at 37 ° for 20 min. Pass through a NICK column equilibrated with water to remove unincorporated nucleotides. To the 400/zl of eluate add 25/zl of 4 M NaC1, 1/xl of glycogen (20 mg/ml), and 860 #1 of absolute ethanol. Centrifuge at 14,000 rpm for 30 min in a cold room and decant the ethanol solution. Wash the pellet with 70% (v/v) ethanol (100 #1), decant, and dry briefly in a Speed-Vac (Savant Instruments, Hicksville, NY). 2. Dissolve the DNA pellet in 42/~1 of water, 5 #1 of buffer B (BoehringerMannheim), and add 3 #1 of HindlII. Incubate at 37 ° for 2-4 hr. Add 15 #1 of 5x Ficoll loading buffer and load on a 7% (w/v) nondenaturing gel. Run the gel and isolate the DNA as described below. Procedure 3 When suitable restriction sites are not present, a 5' 32p end-labeled DNA fragment can be prepared by PCR. 1. To 60 pmol of primer A (the primer to be labeled) add 80 #1 of water, 10 #1 of 10× kinase buffer, 4 #1 of [y-32p]ATP (~250 #Ci), and 6 #1 of polynucleotide kinase. Incubate at 37 ° for 30 min. Add 5 #1 of 0.5 M EDTA and extract with 25 : 24 : 1 (v/v/v) phenol-chloroform-isoamyl alcohol (four times, 100 #1 each). Purify the DNA with a NICK column equilibrated with water to remove unincorporated ATP. To the 400/zl from the NICK column add 24 #1 of 4 M NaC1, 1 #1 of glycogen (20 mg/ml), and 860 #1 of absolute ethanol, mix, and centrifuge (14,000 rpm) for 30 min in a cold room. Decant, wash with 70% (v/v) ethanol (100 #1), decant, and dry briefly in a Speed-Vac. 2. Dissolve 60 pmol of primer B in 60 #1 of water. To the pellet from step 1, add 50/zl of the primer B solution, 33 #1 of water, 10 #1 of PCR buffer (Boehringer Mannheim), 3.7 #1 of plasmid DNA (0.003/zg/ml), 2 #1 of dNTP mix (each at 10mM), and 1 #1 of 100x bovine serum albumin (New England BioLabs, Beverly, MA). Transfer the solution to a PCR tube. 3. Heat the tube at 70 ° in a thermocycler for 5 min (hot start PCR). Carefully add 4 units (0.8 #1) of Taq DNA polymerase, mix, and cover the solution with mineral oil. 4. Thermocycle as follows: 30 cycles of (1) 94 ° for 1 min, (2) 54 ° for 1 min, (3) 72 ° for 1.5 min, and then 72 ° for 10 min. 5. Carefully transfer the aqueous phase to a new tube. For each 100 #1 of solution, add 30/A of 5 x Ficoll loading buffer. Load onto a 7% (w/v) nondenaturing gel and purify as described below.

456


[22]

Gel Purification of End-Labeled DNA 1. Prepare 7% (w/v) nondenaturing gel mix as follows: to a 125-ml Erlenmeyer flask add 8.1 g of acrylamide (caution: acrylamide is toxic!), 290 mg of bisacrylamide, and 24 ml of 5x TBE, and dilute to 120 ml with water and stir until homogeneous. Initiate polymerization by adding 700 ttl of 10% (w]v) ammonium persulfate (APS) and 42 #1 of N,N,N',N'-tetramethyl ethylene diamine (TEMED), and pour the gel. Prepare 1 liter of I x TBE gel running buffer. Run the gel at ~200 V for ~ 2 hr. 2. Carefully take the gel down and cover it with Saran Wrap. Expose the gel to X-Omat film (Kodak, Rochester, NY) for 30-60 sec. Place the developed film under the gel and cut out the desired band, using a razor blade. Transfer the gel slice to a 1.5-ml tube. 3. Crush the gel slice with a pipette tip and add 700 #1 of elution buffer (20 mM Tris-HC1, 250 mM NaCI, pH 8). Place the tube in a shielded container and soak overnight (8-16 hr) in a 37 ° shaker. Filter the DNA, using a plastic centrifugal filter (Quiksep; Isolab, Singapore) into a 15-ml Falcon 2059 tube, and transfer the eluate to a 1.5-ml tube. Add 1.5 volumes of 2-propanol, mix by inversion, and precipitate by spinning in a microcentrifuge at 14,000 rpm for 30 rain in a cold room. Decant, and then add 100 #1 of 75% (v/v) ethanol, spin briefly (30-90 sec), decant, and dry the DNA pellet briefly in a Speed-Vac. 4. Dissolve the DNA in 100 #1 of water and pass it through a NICK column equilibrated with water. Divide the 400-#1 eluate into aliquots, which are then counted in a scintillation counter and stored at - 8 0 ° until use. P r e p a r a t i o n of P o l y a m i d e S e r i a l D i l u t i o n s Solid-phase methods for the synthesis of polyamide have been described, z6 Polyamide stock solutions are conveniently prepared from dry 25-nmol aliquots. Concentrations are determined by UV absorption, using an experimentally determined extinction coefficient. Extinction coefficients can be estimated on the basis of the number of aromatic rings, using the relation 8690 M- 1 cm- 1/aromatic ring for the absorption maximum between 290 and 315 nm (e.g., for an eightring polyamide, ~ ~ 69,500 M ~ cm -1 at the maximum between 290 and 315 nM). Polyamides having at least one positive charge per eight rings are generally freely soluble up to a concentration of at least 500 ttM. The presence of multiple f%alanines can reduce solubility. It is good practice, especially for polyamides expected to have relatively low solubility, to spin the initial stock solution in 26 E. E. Baird and P. B. Dervan, J. Am. Chem. Soc. 118, 6141 (1996).

[22]

FOOTPRINTINGTO ANALYZEPy-lm/DNA

457

a microcentrifuge for several minutes at high speed (14,000 rpm) to pellet any insoluble material and transfer the supernatant to a new tube before measuring the concentration. For quantitative footprinting experiments, the goal is to prepare binding reactions over a range of polyamide concentrations that produce from 0 to 100% fractional saturation of the DNA-binding site(s) of interest. This generally requires at a minimum a three order of magnitude concentration range. In a typical titration, 10-13 data reactions, 1 control reaction (with no polyamide), and l undigested DNA control should be prepared. Prepare polyamide stock solutions as follows: Dissolve 25 nmol of polyamide in water to give an approximately 10 #M solution. Determine the concentration of this solution by UV absorption, and then prepare 10× polyamide stock solutions with nearly equal spacing on a log scale (e.g., 1, 2, 5, and 10 nM). Quantitative DNase I Footprinting Procedure 1. Prepare a 1.14× DNA solution: Combine 2.06 ml of 5x TKMC, pH 7.0 (50 mM Tris-HC1, 50 mM KC1, 50 mM MgC12, 25 mM CaC12, pH 7.0}, 7.0 ml of water, and 32p-labeled DNA (400,000 cpm). The amount of DNA used is calculated to give a final loading of --~15,000 cpm per lane. 2. Prepare equilibrium binding mixtures: To a 1.5-ml presiliconized Eppendoff tube add 40 #1 of 10x polyamide solution (or water for the control lanes) and 350/zl of 1.14× DNA solution. Allow the mixtures to equilibrate at room temperature for 4-24 hr. Polyamides with association constants over 109 M -l should be allowed to equilibrate for at least 12 hr. 3. Prepare DNase I stop buffer: Combine 40 #1 of glycogen (20 mg/ml), 40/zl of 1 mM bp calf thymus DNA, 107 ~1 of water, 788 #1 of 4 M NaC1, and 425/zl of 0.5 M EDTA, pH 8.0. The EDTA in this buffer quenches DNase I activity by chelating the essential Mg 2+ and Ca 2+ ions. The other ingredients provide for good ethanol precipitation and subsequent resuspension. 4. Prepare a DNase I stock solution: Combine 975/zl of water and 20 #1 of 50 mM dithiothreitol (DTT), and chill the solution on ice. Add 5/zl of DNase I (Pharmacia FPLCPure, 7500 units/ml) and mix the resulting 38 units/ml-solution by inverting the tube several times. Prepare the final stock solution by adding 15-70 #1 of DNase I (38 units/ml) to a prechilled mixture of 20/zl of 50 mM DTT and 930-965 #1 of water (the total final volume should be 1000/zl). Keep the DNase I solution on ice throughout the experiment (prepare this solution freshly from the 7500-units/ml stock and use within l hr). The exact amount of DNase I to use depends on the batch of DNase I and the restriction fragment being used. The goal is to achieve ~50% digestion of the restriction fragment. Although the exact amount of DNase I to use can be determined by running a titration, it is

458


[22]

usually possible to estimate on the basis of the length of the restriction fragment. For a 220-bp fragment, dilute ~ 5 0 / z l of the 38-units/ml stock solution to 1 ml. Shorter restriction fragments require more DNase I, longer fragments less (e.g., for a 440-bp fragment, use ~25 #1 of the 38-units/ml solution). 5. Digest the DNA: To each tube (except the intact DNA control) add 10 #1 of the final DNase I solution and mix by vortexing. Allow the reaction to proceed for 7 rain at room temperature, and then add 50/~1 of "stop buffer" and mix by vortexing. Add 975 #1 of room temperature absolute ethanol and mix by inversion. Carry out this step while using a stopwatch: First, open the tops of all tubes, then start the stopwatch. Every 15 sec, add DNase I to a tube, close the top, mix by vortexing, and place the tube in a microcentrifuge. When all the tubes are in the microcentrifuge, spin them briefly (~2 sec), remove from the centrifuge, and open all the tops. When 7 rain has elapsed, reset the stopwatch. Add stop buffer every 15 sec and vortex. Finally, add ethanol and mix. 6. Precipitate the DNA: Spin the tubes in a microcentrifuge in a cold room at 14,000 rpm for 25-30 min (start the samples spinning right away, and do not spin for longer than 30 min: spinning longer may make resuspension of the DNA more difficult). After precipitation the pellets should be visible. Carefully decant the supernatant, and then add 300 #1 of 70% (v/v) ethanol. Vortex the tube briefly to thoroughly wash the pellet. Spin the tubes briefly (10-20 sec) in a microcentrifuge, and then carefully decant the supernatant. 7. Resuspend the DNA: To each tube add 15 /zl of water and vortex (5-10 sec) to resuspend the DNA (dissolving the samples in water and reconcentrating them increases the chances that the DNA will resuspend properly and not hang in the wells). Freeze the samples by placing them in liquid nitrogen, or in a - 8 0 ° freezer. The samples may be stored overnight at this point at - 8 0 °. Concentrate the solution to dryness in a Speed-Vac (to reduce the chances of hanging lanes, do not overdry; it is best to take the samples out as soon as they are dry). Next, add 7 #1 of 80% (v/v) formamide-1 x TBE loading buffer (prepared by mixing 8 ml of formamide with 2 ml of 5 x TBE). The loading buffer should be stored at 4 ° and discarded after 1-2 months (use of old loading buffer can result in smearing of bands on the gel). Thoroughly vortex each tube (~30 sec) to resuspend the DNA. 8. Denature and load the gel: Denature the DNA by heating for 10 min at 85-90 °, and then immediately place the samples on ice. Load 5/~1 per lane on a prerun 8% (w/v) [ 19 : 1 (w/w) acrylamide-bisacrylamide] denaturing polyacrylamide gel (prerun the gel for 1 5 4 0 min until the temperature of the front gel plate reaches 50-55°). Load one or more chemical sequencing reaction lanes? 7,2s Run the gel at 27A. M. Maxamand W. S. Gilbert,Methods Enzymol. 65, 499 (1980). 28B. L. Iversonand P. B. Dervan,Nucleic' Acids Res. 15, 7823 (1987).

[22]


459

'-~2000 V (running bromphenol blue to the bottom of the gel is good for resolving DNA sites 40-80 base pairs from the labeled end). To help ensure that the gel sticks to only one glass plate when the plates are separated, one of the gel plates may be treated with a silanizing reagent (e.g., SigmaCote; Sigma, St. Louis, MO) occasionally (once every 10 runs or so). 9. Transfer the gel to drying paper, cover with plastic wrap, and dry on a gel dryer for 60 min at 80 °. Expose the gel to a storage phosphor screen (Molecular Dynamics, Sunnyvale, CA) for 8-16 hr.

Data Analysis Image the gel with a phosphorImager such as the Molecular Dynamics 400S PhosphorImager. Intensity data are obtained from the gel image by quantitation using ImageQuant software (Molecular Dynamics). Background-corrected (i.e., quantitate the intact DNA control lane and subtract this background value) volume integration of rectangles encompassing the footprint sites and a reference site at which DNase I reactivity is invariant across the titration provides values for the site intensities (/site) and the reference intensities (Iref), respectively. The apparent fractional occupancy (0app) of the sites is then calculated using Eq. (1): 0app = 1 --

(lsite/lref)/(lOsite/l°ref)

(1)

w h e r e I0site a n d / O r e f are the site and reference intensities, respectively, from a control lane to which no polyamide was added. The ([L], 0app) data points are then fitted to a general Hill equation [(Eq. (2)] by minimizing the difference between 0ap p and 0fit: 0fit = 0mi n + (0ma x --

Omin)(Kna[L]n/1 + K"a[L]n)

(2)

where [L] is the total polyamide concentration, Ka is the equilibrium association constant, and 0mi n and 0max are the experimentally determined site saturation values when the site is unoccupied or saturated, respectively. The data are fitted by a nonlinear least-squares fitting procedure (such as KaleidaGraph software; Synergy, Reading, PA) with Ka, 0min, and 0max as the adjustable parameters, and with n fixed at either 1 or 2. A good fit to a Langmuir isotherm [Eq. (2), n = 1] is consistent with formation of a 1 : 1 polyamide-DNA complex, whereas a good fit to a cooperative isotherm (Eq. (2), n = 2] is consistent with cooperative dimeric binding. 29 The data are normalized by Eq. (3): 0norm = (0app -- 0min)/(0max -- 0min)

(3)

29 C. R. Cantor and E R. Schimmel, "Biophysical Chemistry, Part III: The Behavior of Biological Macromolecules," p. 863. W. H. Freeman, New York, 1980.

460


[22]

Reported association constants are the average values obtained from three independent footprinting experiments.

M e t h i d i u m p r o p y l - E D T A - F e ( I I ) F o o t p r i n t i n g P r o t o c o l a n d Affinity Cleavage Protocols The protocol reported here avoids the use of carrier DNA and gives apparent affinities that are the same as those obtained using the DNase I footprinting assay described above. The final solution conditions are 20 mM HEPES, 200 mM NaCI glycogen (50 tig/ml), 5 mM DTT, pH 7.3, and either (1) 0.5 tiM MPE-Fe(II) for footprinting experiments or (2) 1 tiM Fe(II) for affinity cleavage experiments. 1. Prepare a 1.43× DNA solution: To a 15-ml Falcon 2059 tube add 1286 til of 5× cleavage buffer (100 mM HEPES, 300 mM NaCI, pH 7.3), 225 #1 of 4 M NaC1, 3 ml of water, and labeled DNA (250,000 cpm). 2. Prepare 10× polyamide (or EDTA-polyamide) stock solutions. 3. Set up equilibrations: To a 1.5-ml Eppendorf tube add 280 #1 of 1.43 × DNA solution, 40 #1 of 10x polyamide, and 40 #1 of glycogen (0.5 mg/ml). Allow the solution to equilibrate for 1-18 hr. 4. Prepare a precipitation buffer by combining 35 #1 of glycogen (20 mg/ml), 35 #1 of calf thymus DNA (1 mM bp), and 180 #1 of water. 5a. For MPE footprinting experiments: Prepare 5 #M MPE-Fe(II) by combining equal volumes of 10 tiM MPE and freshly prepared 10 #M Fe(NH4)2(SO4)2. To the polyamide-DNA solution add 40 til of 5 tiM MPE-Fe(II); allow to equilibrate for 10 min. Add 10 #1 of 200 mM DTT (use a freshly thawed DTT aliquot). Allow cleavage to proceed for 30 min. Add 1 ml of ethanol, mix the tubes by inversion, and spin briefly. 5b. For affinity cleavage experiments: To the EDTA-polyamide/DNA solution add 20 til of freshly prepared Fe(NH4)2(SO4)2 and equilibrate for 10-30 min. Add 40 til of DTT (use a freshly thawed aliquot) and allow cleavage to proceed for 30 min. Add 1 ml of ethanol, mix the tubes by inversion, and spin briefly. 6. Precipitate the DNA and run the gel: Add 10 #1 of precipitation buffer to each tube and mix by inversion. Spin the tubes at 14,000 rpm in a cold room for 30 min. Decant, wash with 75% (v/v) ethanol (350 #1), and decant. Resuspend the DNA in 16 til of water, freeze, and Speed-Vac dry. Resuspend in 7 til of 80% (v/v) formamide-1 x TBE loading buffer, heat denature (10 min at 85-90 °, then place on ice), and load 5 til per lane on a denaturing 8% (w/v) gel and subject to electrophoresis. 7. Transfer the gel to paper, cover with plastic wrap, and dry on a gel dryer (45-60 min) at 80 °. Expose the gel to a storage phosphor screen for 8-16 hr and image with a Phosphorlmager.

[22]

FOOTPRINTINGTO ANALYZEPy-Im/DNA o, HN

461

I N

~/, .

,

o o

I

~

. N2N'JL'.~"N~:

I

1 (R=H) 1-E (R=RE)

".

".

O

O

.

.~ O I

.

RE = I ~ . - ' N ' - . P ' N ~, '

~,. N'~

I

N O

o

c

,%0 °

H H H+ "~j...~N....~-.,~N~.~/N-R '~' ~ ~ "

FIG. 2. Structures of polyamide ImPy-fl-lmPy-y-lmPy-/3-ImPy-/3-Dp (1) and EDTApolyamide ImPy-/3-ImPy-y-lmPy-/%ImPy-/3-Dp (l-E) (/3,/3-alanine; y, y-aminobutyric acid; Dp, dimethylaminopropyl).

Data Analysis The extent of cleavage protection [for MPE-Fe(II) footprinting] and cleavage intensities (for affinity cleavage experiments) are determined by quantitation of the gel, using ImageQuant software. For MPE-Fe(II) footprinting experiments, 0app values are calculated for each band, using Eq. (1). For affinity cleavage experiments, quantitation of cleavage products provides the cleavage efficiency directly. C h a r a c t e r i z a t i o n of I m P y - ~ - I m P y - y - l m P y - ~ - I m P y - ~ - D p We reported previously that the eight-ring polyamide ImPy-j6-1mPy-y-lmPy]%ImPy-fl-Dp (1) (Fig. 2) binds to 5'-(A,T)GC(A,T)GC(A,T)-3' target sequences within the human immunodeficiency virus type I (HIV- i) promoter region that are adjacent to binding sites for the cellular transcription factors TBP and Ets- l, blocks HIV-I transcription in vitro, and inhibits HIV-I replication in cell culture, l° As an example application of the quantitative DNase I footprinting protocol described here, we present results of experiments that (1) assessed the specificity of I for a match site versus several mismatch sites, and (2) compared equilibrium association constants determined with TKMC buffer (10 mM Tris-HCl, 10 mM KCI, I0 mM MgCI2, 5 mM CaCl2, pH 7.0), which provides optimal DNase I activity, with those obtained with solution conditions approximating those expected to prevail within a living cell. Complexes of 1 with binding sites on the 282-bp EcoRI/PvuII restriction fragment from pJT2B2 were characterized by quantitative DNase I footprinting, MPEFe(II) footprinting, and affinity cleavage experiments. Plasmid pJT2B2 was prepared by hybridizing the complementary oligonucleotides 5'-CCGGCTTAAGTTC GTGGGCCATGCTGCATTCGTGGGCCATGGTGGATTCGTGGGCCATGTTA CATTCG-3' and 5'-TCGACGAATGTAACATGGCCCACGAATCCACCATGGC

462


[22]

5I

I

A A C A G C ]-T

I TC G A C

] T C

T A A C A

C A T

C A

I

3' FIG. 3. Storage phosphor autoradiogram of a denaturing 8% (w/v) polyacrylamide gel used to separate the products generated by DNase 1 digestion in a quantitative footprinting experiment with polyamide L Lane 1, A-specific sequencing lane (see Dervan17); lane 2, DNase I digestion products obtained in the absence of polymide; lanes 3-12, DNase I digestion products obtained in the presence of 0.002, 0.005, 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, 1, and 2 nM polyamide 1, respectively; lane 13, intact DNA. Polyamide-binding sites and the reference site (shaded bracket) are indicated along the right side of the autoradiogram. All reactions contained 15 kcpm of 32p end-labeled restriction fragment, 10 mM Tris-HCl, 10 mM KC1, 10 mM MgCI2 and 5 mM CaCI2 (pH 7.0, 24°).

464


[22]

TABLE 1 EQUILIBRIUMASSOCIATIONCONSTANTSOF POLYAMIDE 1 FOR ITS MATCH SITE 5'-TGCTGCA-3 'a Buffer t'

Temperature(°C)

Ka (M -I)

TKMC

24 37

1.5 x 10 t° (0.3) 8.4 × 109 (2.3)

lntracellular

24 37

1.9 × 10 I° (0.6) 1.1 x 101°(0.2)

a Values reported are the mean values from at least three DNase I footprint titration experiments. The standard deviation for each value is indicated in parentheses. ~' TKMC buffer: 10 mM Tris-HCl, 10 mM KC1, 10 mM MgCI2, and 5 mM CaCI2, pH 7.0 at 24 °. Intracellular buffer: 10 mM HEPES-HCI, 140mMKCI, 10 mM NaCI, 1 mMMgCI2, I mM spermine, pH 7.2.

5'-AGCTGTT-3' and the "reverse orientation" site 26 5'-TCGTCGA-3'. The (0norm, [L]) data points for 1 binding to these sites were well fit by Langmuir binding isotherms [Eq. (2), n = 1], consistent with formation of the expected 1 : 1 "hairpin" polyamide-DNA complexes. Polyamide 1 is >50-fold specific for its match site 5'-TGCTGCA-3' relative to the double-base pair mismatch sites 5'-TC_C_CAC_CA-3' and 5'-TGTAACA-3'. Additional footprinting experiments indicate that increasing the equilibration temperature from 24 to 37 °, and changing the solution conditions from standard polyamide assay conditions (i.e., TKMC buffer: 10 mM Tris-HC1, 10 mM KC1, 10 mM MgCI2, 5 mM CaC12, pH 7.0 at 24 ° to conditions modeling those encountered within a typical mammalian cell 3° (140 mM KCI, 10 mM NaC1, 1 mM MgC12, I mM spermine, pH 7.2) has little effect on the affinity of polyamide 1 for its match site 5'-TGCTGCA-3' (Table I). We note that DNase I activity is nearly 100-fold lower in the model intracellular buffer compared with TKMC. MPE-Fe(II) footprinting and affinity cleavage experiments confirm that 1 binds to its match site 5'-TGCTGCA-3' (Fig. 6). The observed affinity cleavage pattern, which consists of roughly equal cleavage patterns at both ends of the binding site, is expected because the ligand can bind its pseudosymmetric match site in two orientations. Cleavage patterns are shifted toward the 3' end of each strand at the binding site, consistent with binding of the ligand in the minor groove of DNA. 30 R. J. Jones, K. Y. Lin, J. F. Milligan, S. Wadwani, and M. D. Matteucci, J. Org. Chem. 58, 2983 (1993).

[22]

FOOTPRINTINGTO ANALYZEP y - I m / D N A

A

B

q u_ t,- t~.

1 2 3 4 5

465

t-

Z

I nM1

6 7 8 9

5' - T G G C C C A C G A A T ( ~ C A ( ~ . A T G G C C C 3' - ACCGGGTGCTTACGTCGTACCGGG-

1 nM1-E

,11,

,111,

5 ' - TGGCCCACGAATGCAGCATGGCCC3' -ACCGGGTGCTTACGTCGTACCGGG-

'II'

3' 5'

'II'

3 ' 5 '

5' ~ ~

T C G

C ~

~ ~,

~

~

m

S

T

"]

A

,-I

n

T ....

C A

. . . . . . .

C

.,1 C

A

..... ....

3j

FIG. 6. (A) Storage phosphor autoradiogram of a denaturing 8% (w/v) polyacrylamide gel used to separate the products generated by an MPE-Fe(II) footprinting experiment with polyamide 1 and by an affinity cleavage experiment with polyamide | - E . Lanes 1, 6, and 8, A sequencing lanes; lanes 2 and 4, MPE-Fe(II) cleavage products obtained in the absence of 1; lane 3, MPE-Fe (II) cleavage products obtained in the presence of 1 nM 1; lane 5, intact DNA [no MPE-Fe(II)]; lane 7, cleavage products obtained in the presence of 1 nM l - E ; lane 9, intact DNA (no I-E). The reactions contained 15 kcpm of restriction fragment, 20 mM HEPES, 200 mM NaCI, glycogen (50 #g/ml), and 5 mM DTT at pH 7.0, and 0.5/zM MPE-Fe(II) [for MPE-Fe(II) footprinting, lanes 2 4 ] or 1 # M Fe(II) (for affinity cleavage, lanes 7 and 9). (B) Results of MPE-Fe(II) footprinting (top) and affinity cleavage experiments (bottom) with polyamides I and I-E, respectively. Horizontal and vertical bar heights are proportional to the amount of cleavage protection and cleavage, respectively, at the indicated base.

466


[23]

Acknowledgments We are grateful to the National Institutes of Health (General Medical) and the National Foundation for Cancer Research for research support, and to the National Science Foundation and the Ralph M. Parsons Foundation for predoctoral fellowships to J.W.T.

[23] High-Resolution Transcription Assay for Probing Drug-DNA Interactions at Individual Drug Sites By DON R. PHILLIPS,SUZANNEM. CUTTS, CARLEENM. CULLINANE,and DONALD M. CROTHERS Introduction Approximately half the anticancer agents in routine current clinical use are known to interact with DNA by one or more of the following mechanisms: (1) intercalation [e.g., doxorubicin (Adriamycin), mitoxantrone], (2) groove binding (e.g., distamycin), (3) formation of covalent adducts and/or cross-links (e.g., cisplatin, melphalan, mitomycin C), and (4) incorporation of modified bases (e.g., 5fluorouracil, 6-thioguanine). Although the exact mechanisms of action of these agents remain unresolved in some cases, the apparent critical role of DNA has prompted the initiation of a wide range of studies of these drug-DNA interactions. A detailed understanding of the chemical/biochemical aspects of such interactions with DNA (particularly categories 1-3 above) would be expected to provide the necessary insight to design new generations of more active derivatives. New therapeutic approaches have indeed become apparent as a result of improved understanding of the molecular detail of drug-DNA interactions. Some of the fundamental properties of drug-DNA complexes include (1) the DNA sequence involved in the interaction, (2) the kinetics of the interaction, and (3) the affinity of the interaction. There have been two distinct phases in the development of experimental approaches to determine these parameters. Early procedures relied on a variety of physicochemical methods such as detergent sequestration, equilibrium dialysis, and spectrophotometric/spectrofluorimetic binding studies. Although these procedures all yield useful and valuable information concerning the overall drug-DNA interaction, a general limitation is that they usually yield only average binding parameters resulting from the multiple equilibria occurring at a multitude of binding sites on heterogeneous DNA and do not provide details of the drug-DNA interaction at individual sites on the DNA. To overcome this limitation several new approaches were subsequently developed that relied on the use of identical sequences of DNA (usually from plasmids), rather


Copyright(c) 2(101by AcademicPress All rightsof reproductionin any form reserved. 0076-6879/00$35.00

466


[23]

Acknowledgments We are grateful to the National Institutes of Health (General Medical) and the National Foundation for Cancer Research for research support, and to the National Science Foundation and the Ralph M. Parsons Foundation for predoctoral fellowships to J.W.T.

[23] High-Resolution Transcription Assay for Probing Drug-DNA Interactions at Individual Drug Sites By DON R. PHILLIPS,SUZANNEM. CUTTS, CARLEENM. CULLINANE,and DONALD M. CROTHERS Introduction Approximately half the anticancer agents in routine current clinical use are known to interact with DNA by one or more of the following mechanisms: (1) intercalation [e.g., doxorubicin (Adriamycin), mitoxantrone], (2) groove binding (e.g., distamycin), (3) formation of covalent adducts and/or cross-links (e.g., cisplatin, melphalan, mitomycin C), and (4) incorporation of modified bases (e.g., 5fluorouracil, 6-thioguanine). Although the exact mechanisms of action of these agents remain unresolved in some cases, the apparent critical role of DNA has prompted the initiation of a wide range of studies of these drug-DNA interactions. A detailed understanding of the chemical/biochemical aspects of such interactions with DNA (particularly categories 1-3 above) would be expected to provide the necessary insight to design new generations of more active derivatives. New therapeutic approaches have indeed become apparent as a result of improved understanding of the molecular detail of drug-DNA interactions. Some of the fundamental properties of drug-DNA complexes include (1) the DNA sequence involved in the interaction, (2) the kinetics of the interaction, and (3) the affinity of the interaction. There have been two distinct phases in the development of experimental approaches to determine these parameters. Early procedures relied on a variety of physicochemical methods such as detergent sequestration, equilibrium dialysis, and spectrophotometric/spectrofluorimetic binding studies. Although these procedures all yield useful and valuable information concerning the overall drug-DNA interaction, a general limitation is that they usually yield only average binding parameters resulting from the multiple equilibria occurring at a multitude of binding sites on heterogeneous DNA and do not provide details of the drug-DNA interaction at individual sites on the DNA. To overcome this limitation several new approaches were subsequently developed that relied on the use of identical sequences of DNA (usually from plasmids), rather


Copyright(c) 2(101by AcademicPress All rightsof reproductionin any form reserved. 0076-6879/00$35.00

[23]

In Vitro TRANSCRIPTIONASSAY

467

than the heterogeneous sequences previously used with bacterial or mammalian DNA. These approaches include DNA footprinting (utilizing DNase I or hydroxyl radicals) to yield preferred drug-binding regions, and inhibition of the processivity of DNA-dependent enzymes such as DNA polymerase, DNA exonucleases, and RNA polymerase. The RNA polymerase inhibition assay results in the accumulation of truncated transcripts arising from blockage of the movement of RNA polymerase along the DNA at specific drug-binding sites. An overview of the procedure is shown in Fig. 1. The length of the truncated RNAs therefore reveals the location of these drug-binding sites. In vitro procedures that are based on blocking the progression of RNA polymerase along double-stranded DNA (dsDNA) are now referred to as "in vitro transcription assays" (and are analogous to DNA polymerase inhibition assays), whereas the term "bidirectional transcription footprinting" denotes the use of two counterdirected promoters, such that truncated transcripts arising from transcriptional blockages to both sides of a drug site define the physical "footprint" of the drug-occupied region. In addition to defining the sequence specificity of drug-binding sites on DNA, the transcription assay can also be exploited to yield the relative drug occupancy at each site as well as the complex lifetime at each individual site. The transcription assay is therefore a powerful technique that is able to yield several molecular parameters simultaneously to define how and where a drug interacts with DNA, and this information can be determined at individual drug sites on the DNA. In this chapter we outline the experimental approach to the in vitro transcriptional analysis of drug-DNA interactions. The approach requires a synchronized population of initiated transcription complexes containing RNA of a common length, which is achieved with promoters with a range of specific characteristics: (1) they should not require additional activating elements such as catabolite activator protein (CAP) and cAMP (these would add an unnecessary degree of complexity to the assay); (2) they should be "strong" promoters (i.e., the RNA polymerase should have a high affinity for the promoter region); (3) "slippage" of the start site of transcription should be minimal (the fidelity of the start site of transcription should be >99%, or capable of being "forced" to that level of fidelity); (4) the sequence of the nascent RNA must be such that a stable initiated complex can be formed in the absence of one or more of the four nucleotide triphosphates (5) the rate of formation of the initiated transcripts should be rapid (for experimental convenience); and (6) the half-life of the initiated transcription complex should be at least several hours, A range of promoters appears to satisfy all these criteria (Table I). Because the UV5 promoter had been extensively characterized it was used for the development of this assay, l and has continued to be the major promoter used in our laboratories. I D. R. Phillips and D. M. Crothers,Biochem&try25, 7355 (1986).

468


[23]

~[ RNAPOLYMERASE

HEPARIN

I

INITIATIONNUCLEOTIDES

DRUG

I

-

-

I

ELONGATIONNUCLEOTIDES

~ |

FIG. 1. Overviewof the transcriptionassay. The majorsteps are bindingof RNA polymeraseselectivelyto the lacUV5 promoter, formationof a synchronizedinitiatedtranscriptioncomplex(see Fig. 2 for details), reactionof the initiatedtranscriptioncomplex with the drug of interest, and elongationof the transcriptioncomplex to yield drug-inducedblocked transcripts.

I s o l a t i o n of 5 1 2 - B a s e P a i r lacUV5 DNA F r a g m e n t The lacUV5 promoter is obtained from the plasmid pHWO1, which contains a single copy of the LSUV5 double mutant, 203-bp lac promoter cloned into the EcoRI site of pHW1.2 Only the UV5 mutation at - 9 (numbering with respect to the mRNA start site at + 1) is significant for this work because this confers strong "up"

2 H. Wu and D. M. Crothers, Nature (London) 308, 509 (1984).

[23]

In Vitro TRANSCRIPTION ASSAY

469

TABLE I PROMOTERS THAT YIELD SYNCHRONIZEDINITIATEDTRANSCRIPTIONCOMPLEXES Promoter

Initiation dinucleotide

Absent initiation nucleotide

Initiated transcript length

Ref. a

UV5 N25 TetR SP6 T3 T7 ~-PL

GA AU AG AG AG AG AU

CTP CTP GTP GTP CTP UTP UTP

10 29 1I 9 12 13 15

1 2 3 4 4 4 3

"Key to references: (1) R. J. White and D. R. Phillips, Biochemistry 27, 9122 (1988); (2) R. J. White and D. R. Phillips, Biochemistry 28, 6259 (1989); (3) H. Trist and D, R. Phillips, Nucleic" Acids Res. 17, 3673 (1989); (4) R. J. White and D. R. Phillips, Biochemistry 28, 4277 (1989).

p r o m o t e r c h a r a c t e r i s t i c s to the p r o m o t e r a n d i n i t i a t i o n o f t r a n s c r i p t i o n d o e s not r e q u i r e a c t i v a t i o n b y CAP. 3 A g o o d s u m m a r y o f the s e q u e n c e c h a r a c t e r i s t i c s o f the 2 0 3 - b p f r a g m e n t is available. 4 T h e 2 0 3 - b p f r a g m e n t was ligated into the E c o R I site o f p B R 3 2 2 to yield pRW1.5 T h e U V 5 p r o m o t e r c a n b e e x c i s e d as a 4 9 7 - b p +1

I

5 ' - GGAATTGTGAGCGGATAACAA'I-I'TCACACAGG 3 ' - CC-I-I-AACACTCGCCTATTGTTAAAGTGTGTCC

GpA E.

coli RNA polymerase

ATP, GTP, [32P]UTP +1 I v 5 ' -GGAATTGTGAGCGGATAACAATTTCACACAGG 3 ' - CC-FFAACACTCGCCTATTGTTAAAGTGTGTCC

GAAUUGUGAG ~",xx

T e m p ' l a t e DNA s t r a n d

Radiolabeled

nascent RNA

FIG. 2. Synchronized initiated transcription complex. Initiation of the lacUV5 promoter with E. coli RNA polymerase, GpA, ATP, GTP, and [c~-32p]UTPresults in a stable transcription complex containing a nascent RNA mainly 10 nucleotides in length, because CTP is absent (denoted by the arrowhead). Minor amounts of 17-mer and 23-mer transcripts are also formed, presumably resulting from CTP contaminants of other components (or from misincorporation of other nucleotides). The nascent RNA begins at the - 1 position with G of GpA present in the initiation mixture. The first nucleotide of the transcript formed under normal conditions is denoted as + 1. Radiolabel (32p) is incorporated into the nascent RNA at three sites, denoted with asterisks.

470


[23]

PvulI/SalI fragment. The 497-bp fragment from pRW1 (containing the UV5 promoter) has also been cloned directionally into pSP64 to yield a much higher copy number plasmid, pCC 1. The plasmid yield from this vector is significantly greater than from pBR322-derived vectors. The UV5 promoter is then excised as a 512-bp PvulI/HindlII fragment. 6 The lacUV5 promoter is also available commercially in the vector pGH6 [American Type Culture Collection (ATCC), Manassas, VA]. Procedure

1. Restriction digest pCC 1 with PvulI and HindlII [or use appropriate restriction enzymes/polymerase chain reaction (PCR) primers to isolate the lacUV5 promoter from other sources]. 2. To separate the resulting two DNA fragments, subject the restriction digest to electrophoresis, using a 1.5% (w/v) minisubmarine agarose gel in Tris-borateEDTA (TBE) buffer (lacking ethidium bromide) for 2 hr at 10 V/cm. If ethidium bromide is included in the agarose gel, then single-strand nicks may be induced in the DNA, and this will result in a high background of truncated transcripts during the elongation phase of the transcription assay. 3. Cut a thin slice lengthwise from the gel and stain with ethidium bromide (0.5 #g/ml). 4. Visualize the location of the 512-bp fragment in the gel slice with a transilluminator. 5. Align the gel slice with the main gel and excise the 512-bp fragment from the main agarose gel. 6. Place the agarose gel slice in a Biotrap chamber (Schleicher & Schuell, Keene, NH) and electroelute the DNA fragment. The Biotrap electroelution procedure has a high efficiency of recovery of DNA from agarose (typically greater than 95%). 7. Purify the DNA further by extracting with an equal volume of phenolchloroform and then subject to ethanol precipitation. 8. Redissolve the DNA in TE buffer to a concentration of approximately 100 #g/ml. F o r m a t i o n of S y n c h r o n i z e d I n i t i a t e d T r a n s c r i p t i o n C o m p l e x e s Formation of synchronized complexes initially requires an interaction between Escherichia coil RNA polymerase and DNA containing the [acUV5 promoter.

The use of heparin displaces bacterial RNA polymerase from nonspecific binding 3 A. E. Silverstone,R. R. Arditti, and B. Magasnik,PJvc. Natl. Acad. Sci. U.S.A. 66, 773 (1970). 4 F. Schaeffer,A. Kolb,and H. Buc,EMBOJ. 1, 99 (1982). 5R. J. White and D. R. Phillips,Biochemistry 27, 9122 (1988). 6C. Cullinaneand D. R. Phillips,Nucleic Acids Res. 21, 1857(1993).

[23]

In Vitro TRANSCRIPTIONASSAY

471

sites on the DNA, including the ends of the linear DNA, which have modest affinity for the polymerase. This procedure ensures that only single-copy transcripts result from the subsequent elongation step because the RNA polymerase will be unable to rebind to the promoter because of competition with heparin. Initiation is accomplished by the addition of a nucleotide mix containing a radiolabeled nucleotide (usually UTP) but lacking CTP. The resulting initiated transcription complex comprises a nascent RNA predominantly 10 nucleotides long, which is paused at the first dGTP of the template strand (see Fig. 2), and is sufficiently stable for most drug-DNA studies, with a half-life of 23 hr at 37 °.7 Transcription from the UV5 promoter does not begin exclusively from the + 1 site. When all four nucleotides are present only 59% of transcripts begin at the + 1 site, with 29, 7, and 9% beginning from the - l , +2, and -t-5 positions, respectively.8 To ensure that transcription begins from one site only, all nucleotides are maintained at O--H-", --

~..N

"N " ' ~ I~N_ H._.._..N./~C~--N ~ 0

FIG. 1. Structures of natural and unnatural base pairs. Arrows point to the hydrogen bond (d) donor and (a) acceptor groups. The positions of the minor and major grooves are indicated. Dashed lines refer to the hydrogen-bonding interactions between the base pairs. I, D, U, and M represent inosine, diaminopurine, uracil, and 5-methylcytosine, respectively.

488


[24]

DNA Bases: A, T, G, C, and All the Others The common idea that DNA is composed of only four bases, A, T, G, and C, is an oversimplification because most types of DNA, including human DNA, contain naturally modified forms of the canonical bases. 5-Methylcytosine (5-MeC or M; Fig. 1) was the first modified nucleobase to be discovered more than half a century ago. 3 5-MeC completely replaces C in the DNA from Xanthomonas phage XP-12. 4 Nowadays, it is well established that this unusual but naturally occurring base participates in the control of gene expression in higher organisms. 5 Over the last 50 years many other modified bases have been discovered, particularly in DNA from microorganisms. DNA from the bacteriophage SP-15, which infects Bacillus subtilis, contains a modified pyrimidine [5-(4',5'-dihydroxypentyl)uracil] that replaces 62% of the thymines, and other types of B. subtilis phages have T residues replaced by uracil (U) or 5-bydroxymethyluracil (hm5U). 6'7 Just as T residues are found in tRNA, U residues can be a normal constituent of DNA. It is quite common among diverse types of bacteriophages that one of the normal pyrimidines of DNA is completely or partially replaced by a modified derivative. 6 But the modified base can also be a purine. One of the most interesting is 2-aminoadenine (abbreviated DAP or D for 2,6-diaminopurine; Fig. 1) that replaces adenine in the cyanophage S-2L, which infects the blue-green algae Synechococcus elongatus. 8 Thus, DAP in DNA is compatible with normal DNA function although the DAP. T base pair possesses an extra hydrogen bond compared with A- T because of the additional -NH2 group that hydrogen bonds to the 2-keto of thymine in the minor groove of the double helix (Fig. 1). By contrast, the inosine nucleoside (I) lacks the -NH2 group of guanosine and therefore I. C base pairs can form only two hydrogen bonds instead of the three typical of G. C base pairs. Consequently D and I have become favorite tools with which to examine the critical role of the exocyclic 2-amino group of guanine in DNA recognition by various antibiotics. DNA C o n t a i n i n g I n o s i n e a n d / o r D i a m i n o p u r i n e R e s i d u e s Synthesis The phosphoramidite forms of the I and DAP nucleosides are available by total synthesis and can readily be incorporated into synthetic oligonucleotide 3 R. D. Hotchkiss, J. Biol. Chem. 168, 315 (1948). 4 M. Ehrlich, K. Ehrlich, and J. A. Mayo, Biochim. Biophys. Acta 395, 109 (1975). 5 X. Cheng, Annu. Rev. Biophys. Biomol. Struct. 24, 293 (1995). 6 M. Ehrlich and X.-Y. Zhang, in "Chromatography and Modification of Nucleosides" (C. W. Gehrke and K. C. T. Kuo, eds.). Elsevier, Amsterdam, 1990. 7 M. Ehrlich and K. C. Ehrlich, J. Biol. Chem. 256, 9966 (1981). s I. Y. Khudyakov, M. D. Kimos, N. I. Alexandrushkina, and B. E Vanyushin, Virology 88, 8 ( 1978).

[24]

DNA CONTAININGMODIFIEDBASES

489

sequences. 9 There are numerous instances in which these bases have been used to investigate nucleic acid stability, structure, or interaction with ligands. DAP is frequently introduced into DNA to increase its melting temperature ~° and/or into RNA duplexes It as well as for related applications such as making primers for sequencing and PCR 12'13 or fingerprinting.14 Inosine is also commonly used to probe unusual DNA structures such as parallel-stranded DNA 15 or intrinsically curved sequences, 16 and to evidence minor groove binding to DNA of certain proteins such as high mobility group (HMG) proteins (see below) or the tyrosine kinase c-Abl.17 The potential uses of I and DAP are wide ranging.18 Alternatively, I and DAP can be incorporated into DNA by enzymatic methods via the use of the corresponding triphosphate and polymerases. The triphosphate of 2-amino-adenosine acts as a true analog of ATP in transcription. 19 dITP has long been commercially available. By contrast, dDTP was not manufactured until more recently. It is currently sold by TriLink Biotechnologies (San Diego, CA). Methods have been described to convert 2,6-diaminopurine-2'-deoxyribonucleoside into dDTE The synthesis involves converting the nucleoside to its 5'-monophosphate derivative dDMP, followed by pyrophosphorylation. 2°m Other chemical routes have been reported. 22'23 Incorporation of nucleoside triphosphate analogs by polymerases is a useful method of examining miscoding by different DNA polymerases, dDTP and to a lesser extent dITP are good substrates for a number of polymerases, including heat-stable enzymes such as Taq polymerase. We have established a protocol to incorporate inosine and]or diaminopurine into both strands of a 160-mer fragment containing the Escherichia coli tyrT promoter, 24 which offers a great variety of 9 B. S. Sproat, A. M. Iribarren, R. G. Garcia, and B. Beijer, Nucleic Acids Res. 19, 733 (1991). l0 j. D. Hoheisel and H. Lehrach, FEBS Lett. 274, 103 (1990). I1 G. M. Lamm, B. J. Blencowe, B. S. Sproat, A. M. Iribarren, U. Ryder, and A. I. Lamond, Nucleic AcidsRes. 19, 3193 (1991). 12 T. L. Azhykina, S. I. Veselovskaya, V. A. Myasnikov, V. K. Potapov, and E. D. Sverdlov, Proc. Natl. Acad. Sci. Russia 330, 624 (1993). 13 y. Lebedev, N. Akopyants, T. Azhikina, Y. Shevchenko, V. Potapov, D. Stecenko, D. Berg, and E. Sverdlov, Genet. Anal. 13, 15 (1996). 14 M. 1. Prosnyak, S. 1. Veselovskaya, V. A. Myasnikov, E. J. Efremova, V. K. Potapov, S. A. Limborska, and E. D. Sverdlov, Genomics 21, 490 (1994). 15 S. Mohammadi, R. Klement, A. K. Shchyolkina, J. Liquier, T. M. Jovin, and E. Taillandier, Biochemistry 37, 16529 (1998). 16 N. E. M~llegaard, C. Bailly, M. J. Waring, and E E. Nielsen, Nucleic Acids Res. 25, 3497 (1997). 17 M. H. Dav id-Cordonnier, M. Hamdane, C. B ailly, and J. C. D'Halluin, Biochemistry 37, 6065 (1998). 18 C. Bailly and M. J. Waring, NuHeic Acids Res. 26, 4309 (1998). 19 H. R. Rackwitz and K. H. Scheit, Eulz J. Biochem. 72, 191 (1977). 20 N. N. Kahn, G. E. Wright, L. W. Dudycz, and N. C. Brown, Nucleic Acids Res. 13, 6331 (1985). 21 A. Chollet and E. Kawashima, Nucleic Acids Res. 16, 305 (1988). 22 C. A. Brennan and R. I. Gumport, Nucleic Acids Res. 13, 8665 (1985). 23 L. W. McLaughlin, T. Leong, E Benseler, and N. Piel, Nucleic Acids Res. 16, 5631 (1988).

490


[24]

binding sites and has been extensively employed over the last 20 years in footprinting experiments with numerous drugs and proteins. 25 The four DNA molecules containing either natural bases or inosine residues in place of guanosines (G ~ I substitution), or 2,6-diaminopurine residues in place of adenines (A ~ DAP substitution), or both I and DAP residues, can be synthesized by PCR amplification, using the parameters described in Table I. The sequence of the tyrT fragment and the two primers (Watson primer 1 and Crick primer 1) used for the PCR are specified in Fig. 2. For the sake of convenience, the two strands of the duplex are named the Watson (Fig. 2, top, coding) and Crick (Fig. 2, bottom, noncoding) strand. The use of primers in which the 5' terminal nucleotide bears a 5'-OH or a 5'-NH2 terminus is convenient as it enables selective labeling of one or the other strand in the PCR product. The most straightforward procedure is to initiate the PCR amplification with primers already radiolabeled, but the use of unlabeled primers has the advantage that quantities of the modified DNA may be stored for subsequent use. DAP-containing DNA is routinely obtained in good yield, but making the inosinesubstituted DNA can be more tricky. We found that commercial stock solutions of dITP are not stable and should be replaced frequently (every month). Moreover, because of the limited thermal stability of DNA containing inosine residues in both strands the yield of the reaction can vary substantially from one experiment to another, mostly depending on the freshness (purity) of the dITP preparation. This DNA is by far the least stable of the four species, and is best used promptly. The protocol can be adapted to engineer DNA molecules containing the modified base(s) on one strand only of the duplex. Molecules asymmetrically substituted with inosine and/or 2,6-diaminopurine are best made with different PCR primers that produce complementary strands of different length. 26 For example, to synthesize a duplex DNA containing DAP residues on the radiolabeled Watson strand only, a PCR is set up with the tyrT template, the Watson primer 1 and Crick primer 2, and dDTP plus dTTP, dCTP, and dGTP (Fig. 3). This PCR produces a 162-base pair duplex with DAP residues on both strands. In parallel, a PCR is set up with the four natural nucleotides, using Watson primer 2 and Crick primer 1. It produces a normal duplex of 155 base pairs. Both PCR products are then "polished" by treatment with reverse transcriptase to eliminate any incomplete strands (which could subsequently become labeled) by converting them to DNA molecules of homogeneous length. The normal and DAP-containing duplexes are mixed, denatured by heating, and slowly cooled to allow reannealing of the complementary single strands. The products comprise the parental PCR species plus hybrid molecules containing strands originating from the two different PCRs, but only one duplex can be radiolabeled in the presence of [o~-32p]dCTPand avian myetoblastosis virus 24 C. Bailly and M. J. Waring, Nucleic Acids Res. 23, 885 (1995). 25 M. J. Waring and C. Bailly, J. Mol. Recognit. 7, 109 (1994). 26 S. Jennewein and M. J. Waring, Nucleic Acids Res. 25, 1502 (1997).

[241

DNA CONTAININGMODIFIEDBASES

491

TABLE I POLYMERASE CHAINREACTIONCONDITIONSUSED TO INCORPORATEINOSINE AND/OR 2,6-DIAMINOPURINERESIDUESINTOtyrT DNA" Amplification cycles (20--30) Denaturation Primer annealing Polymerization

Normal and DAP DNA

Inosine and I + DAP DNA

1 min at 94 ° 2 min at 45 ° 10 min at 72 °

1 min at 84 ° 2 min at 45 ° 10 min at 62 °

a In all cases, PCR mixtures contained l0 ng of the DNA template, a 1 #M concentration of each primer, a 250/~M concentration of each appropriate dNTP (dTTP, dCTP, plus dATP or dDTP, and dGTP or dlTP according to the desired DNA), and 5 units of Taq polymerase in a volume of 50 #1 containing 50 mM KC1, 10 mM TrisHCI (pH 8.3), 0.1% (v/v) Triton X-100, and 1.5 mM MgC12. To prevent unwanted primer-template annealing before the cycles began, the reactions were heated to 60 ° before adding the Taq polymerase. After the last cycle, the extension segment was continued for an additional 10 min at 72 ° followed by a 5-min segment at 55 ° and a 5-min segment at 37 °. The purpose of these final segments was to maximize annealing of full-length product and to minimize annealing of unused primer to full-length product. 24

(AMV) reverse transcriptase. This is the hybrid duplex bearing a free 3'-hydroxyl group with a single-stranded template 5' overhang that contains a natural Crick strand and an artificial Watson strand, as schematized in Fig. 3. Finally, purification is easy because of the difference in length between the hybrid and either of the

5"5"5'3'-

GTTACCT AATT CCGGTTACCT AATT CCGGTTACCT 0 i0 GGCCAATGGA

TTAATCCGTT TTAATC TTAATCCGTT

ACG

(Watson primer 2) (Watson primer 1)

ACGGATGAAA

20 AATTAGGCAA

ATTACGCAAC

30 TGCCTACTTT

40 TAATGCGTTG

CAGTTCATTT 50 GTCAAGTAAA

TTCTCAACGT

AACACTTTAC

AAGAGTTGCA

TTGTGAAATG

TCGCCGCGCA

CATTTGATAT 90 GTAAACTATA

GATGCGCCCC i00 CTACGCGGGG

GCTTCCCGAT ii0 CGAAGGGCTA

AAGGGAGCAG

60

GCCAGTAAAA 130 CGGTCATTTT

AGCGGCGCGT

70

AGCATTACCC CGTGGTGGGG 140 150 TCGTAATGGG GCACCACCCC (Crick primer 1) 3 ' - G G G C A C C A C C C C

FIG. 2. Sequences of the

80

120 TTCCCTCGTC

GTTC-3'

158 CAAGGGCT -5' CAAGGGCT -5'

tyrT (A93) DNA and the PCR primers used to prepare the modified

DNA.

492

CHEMICAL AND MOLECULAR

[24]

BIOLOGICAL APPROACHES

•r ~

¢5g

< Z Q

!

0 Z

©

°I

¢.),o 0.)

¢~

¢~..c

.~'r-

"5
Y-OH 14-OH --+ 14-H 9-OH --+ 9-H 3'-NH2 ~ Y-OH; 4'-OH ~ 4~-NH2; 14-OH ~ 14-H Sugar removed Sugar removed; 14-OH ---> 14-H Sugar in/3 orientation

A AGt (kcal/mol) 0.0 0.7 0.9 1. I 1.6 2.0 2.5 3.2

a The quantity AAGt refers to the difference in the thermodynamic binding free energy relative to the reference compound DOX. In all cases, AAGt is positive, indicating a less favorable binding interaction resulting from the structural alteration.

daunosamine "nonelectrostatic" free energy contribution was approximately 2 kcal. This substantial contribution to the binding free energy emphasizes the role of the sugar moiety as a tiny DNA groove binder. These studies allowed the molecular determinants of DNA binding of the anthracyclines to be specified in considerable detail and complemented the structural studies of anthracycline-DNA complexes that emerged from X-ray crystallographic studies. 18,20 M o d u l a r D e s i g n of DNA-Binding A g e n t s Using A n t h r a c y c l i n e - B a s e d B u i l d i n g Blocks The major advantages of our modular design approach are (1) its general applicability to different classes of DNA binders, (2) simplification of a structurebased design process, and (3) its compatibility with combinatorial chemistry methods, enabling facile creation of focused libraries of high-affinity sequence-specific small-molecule DNA binders. The potential of small-molecule binders has long been appreciated and there have been attempts to create anthracycline-based bisintercalating molecules2~-23 with expectations that DNA binding can be dramatically increased to the value of 20 A. H.-J. Wang, in "Nucleic Acids and Molecular Biology" (F. Eckstein and D. M. J. Lilley, eds.). Springer-Verlag, Berlin, 1987. 21 D. W. Henry and G. L. Tong, U.S. Patent 4,112,217, September 5, 1978. 22 D. R. Phillips, R. T. C. Brownlee, J. A, Reiss, and P. A. Scourides, Invest. New Drugs 10, 79 (1992). 23 A. Skorobogaty, R. T. C. Brownlee, C. J. Chandler, I. Kyratzis, D. R. Phillips, J. A. Reiss, and H. Trist, AnticancerDrug Des. 3, 41 (1988).

536


[27]

FIG.4. Structure of daunorubicin with daunomycinone, an intercalating aglycone, and daunosamine, a minorgroove-bindingsugar, represented by Legoblocks. the square of the binding constant of the monointercalator and that the binding site size of bisintercalator as well as its selectivity will be increased relative to the monointercalator. However, from the chemist's point of view, such attempts were not optimized for success because they focused on exploring chemically accessible linking positions (C- 13 and C- 14) on the side chain of the aglycon, the problem being that the side chain participates in specific interactions with DNA and so any modification at those positions would ultimately reduce the binding affinity of each monomer. In exploring the use of anthracycline scaffolds in the design of high-affinity agents, we have utilized our results of how anthracyclines interact with DNA. In our modular approach, a molecule of DNR can be presented as two blocks-an intercalating aglycon (daunomycinone) and a minor groove-binding sugar (daunosamine)--connected via a glycosidic bond (Fig. 4). Simple combination of an aglycon with a sugar substituent gives 3-bp binding modules, for example, DNR, which preferentially binds to 5'-(AFF)GC or 5'-(A/T)CG sequences. Even though the quantity and categories of blocks and modules can be easily increased and utilized by modular design, we limit our discussion to anthracycline-based blocks and selected linkers and demonstrate their application in two different approaches to high-affinity agents: (1) a rational structure-based design approach and (2) a combinatorial approach. These modular design approaches are exemplified by two compounds: (1) WP631 (structure-based design) and (2) WP760 (combinatorial approach). Rational Structure-Based

Design

Further analysis of the DNR molecule dissected into its two blocks clearly indicated that DNR block constituents can be combined in different ways. Indeed, both blocks can be individually combined with other nonanthracycline types of DNA-binding molecules. Our first attempt demonstrated the possibility of using our structure-based design approach to design agents with a binding specificity of at

[27]

NOVELHIGH-AFFINITYDNA-BINDINGAGENTS

537

least 6 bp. We had already established that one DNR molecule is a 3 bp-recognizing module; therefore the proper linking of two DNR molecules should afford a 6 bpbinding agent (Fig. 5). Information critical to our design was obtained by analysis of the crystallographic structure of a DNR-hexanucleotide (5'-CGTACG)2 complex and from our own studies of how monointercalating anthracyclines interact with DNA. In the complex of (5'-CGTACG)2 and DNR, the two DNR molecules are arranged such that they intercalate at the two CG steps and their groove-binding sugar moieties face each other at the center of the complex (Fig. 2). Clearly, designing a proper linker that would facilitate aglycon intercalation and sugar minor groove binding and its attachment point was of the utmost importance (Fig. 5). Further analysis of the DNR-DNA complex indicated that the DNA template places the T-amines at a distance of 6-7 ,~. Therefore, the simplest way to obtain a molecule that would bind tightly to the 6-bp sequence (5'-CGTACG)2 was to design a linker that would cover the required 6- to 7-,~ distance and fit into the groove. One such potential linker we considered was a p-xylyl molecule. Molecular modeling studies could easily verify whether such a linker can fit into the minor groove while holding both DNRs in their original binding sites. Below we present methods used to model our target molecule. Molecular Modeling

Modeling studies successfully reproduced the NMR and X-ray crystallographic structures of DNR : DNA and DOX : DNA. This partially validates the modeling protocols, which is the first crucial step in the rationalization, prediction, and design of DNA-interactive ligands. Simulations were performed in the isothermal isobaric ensemble (300 K, 1 atm) with the AMBER 5.0 program and AMBER95 force field, using periodic boundary conditions and the particle mesh-Ewald algorithm. Molecular dynamics (MD) simulations used the Sander routine (1.5-fsec time step), with SHAKE to freeze all bonds involving hydrogen. Detailed methods for the simulation of DNA intercalators and groove binders can be found in [14] in this volume. TM The optimization of the linker is a two-step process: the generation of starting models and initial screening, and the fully solvated molecular dynamics simulations. The implicit solvation (GB/SA algorithm) can be used to coarsely screen many linker candidates with several different sequence lengths. Some models will be too strained to be stable. The most stable structures are then used for fully solvated molecular dynamics to determine the relative stabilities. It is important to include explicit solvation effects and counterions, as DNA is a polyanionic molecule, and to more accurately represent the hydrophobic and hydrophilic nature of the ligand. Also, the minor groove hydration may be disrupted (which can be reproduced by the fully solvated molecular dynamics protocol), which should be taken into account in the overall stabilization of the 23aj. O. Trent,Methods Enzymol. 340, [14] 2001 (this volume).

538

[27]


©

i

©

X..-

;:m

t'.,,i ~Z e-

•~ ._=

,< Z

+

o

'z,

~9 .,.., o e',

.o

*"

o

o

~D

.-'-~,

~9 •

._

--

~-01

Z

O

b __2E 2

[27]

NOVEL HIGH-AFFINITYDNA-BINDING AGENTS

539

DNA : ligand complex. The same protocols can be used to rationalize experimental results and design new modules and linkers. Combinatorial Approach While there is a great appeal of the rational structure-based design and clearly there has been significant progress in molecular modeling methodology, there are experimental conditions that dictate DNA conformation that are not reproduced at present by molecular modeling. Combinatorial approaches can effectively complement the rational approach. Our modular design approach is perfectly suited for both rational design and combinatorial methodology. In our initial work we chose particular building blocks, which included two intercalating blocks (adriamycinone and daunomycinone), two minor groovebinding sugars (daunosamine and 4'-amino-3'-hydroxy-ol-L-lyxo-hexopyranose), and a range of linkers (Fig. 6). Adriamycinone was selected because the 14-hydroxy group plays an important biological function by distinguishing doxorubicin from daunorubicin. In addition, our studies indicated (Tables I and II) that 14-OH contributes approximately I kcal/mol to the free energy of binding to DNA. Our studies (Tables I and II) also indicated that 4'-amino-3'-hydroxy-ol-L-lyxo-hexopyranose, a sugar used in compound WP608, does not reduce DNA-binding affinity, and thus its contribution to the free energy is similar to that of daunosamine. Molecular modeling indicates that substitution at the 4'-amino group should not obstruct DNA binding of the monomer. Therefore, in addition to the 3'-amine, the 4'-amine can also be used as a convenient attachment point for a linker. Our initial efforts resulted in the parallel synthesis of approximately 65 compounds, and their studies continue. Compounds linked at C-4' appeared to have interesting DNA-binding and biological properties, which are still under evaluation. The first compound from the 4'-linked bisanthracycline family was bisintercalator WP652. Its DNA-binding mode was analyzed by IH NMR, and the structure of its complex with d(TGTACA)2 is shown in Fig. 7. 24 Its interaction with DNA differs from that of WP631, which is able to recognize 6-bp sequences (data for WP631 are shown below), whereas WP652 prefers tetranucleotide sequences. The DNA-binding affinity of WP652 is significantly higher than that of DNR. The energetics of WP652 interactions with DNA are now being studied in detail. One of the possible and immediately available tests to assess the biological potential and selectivity of our newly assembled potential DNA-binding agents is the National Cancer Institute (NCI, Rockville, MD) in vitro disease-oriented primary antitumor screen, which consists of 60 human tumor cell lines. The majority of the compounds we have prepared have been tested in this screen, and the most

24H. Robinson,W. Priebe,J. B. Chaires, and A. H.-J. Wang,Biochemisto'36, 8663 (1997).

540


~ CH30

O

CHs

~

C

H

=

O

H

[27]

0•/1•-3•

OH

OH OH

CH~O

Daunomycinone

OH OH

Adriamycinone

HO~~0

O==~H

O

HaN OH Daunosamine

4-Amino-L-lyxo-hexopyrans~

H%O

Ct 2

() HO

0

Br

.o~o Linkers

FIG. 6. Building blocks used for preparation of the first library: daunomycinone, adriamycinone, daunosamine, 4'-amino-L-lyxo-hexopyranose, and linkers (only selected linkers are shown).

WP652 FIG. 7. Chemical structure of WP652 (left) and its complex with d(TGTACA)2 DNA oligonucleotide

(right). 24

[27]

NOVEL HIGH-AFFINITY DNA-BINDING AGENTS

541

WP760

FIG. 8. Chemical structure of WP760 (left) and molecular modeling-derived structure of WP760DNA complex (right).

interesting pattern of antitumor activity we noticed was for a close analog of WP652, which for a change was a doxorubicin-based DNA-binding agent (WP760, NSC 688130) (Fig. 8) connected like WP652 at C-4' rather than at C-3', like WP631. The NCI screen revealed that WP760 is selectively cytotoxic against melanoma cell lines. Results presented in Figs. 9 and 10 (Fig. 9 containing dose-response curves and Fig. 10 containing mean graphs) clearly indicate the selective cytotoxicity of WP760. In the mean graphs (Fig. 10), bars extending to the fight represent sensitivity of the cell line to WP760 in excess of the average sensitivity of all tested lines. Because the bar is logarithmic, a bar 2 units to the right indicates that the compound achieved a response parameter [i.e., median lethal concentration (LCs0)] at 1/100th the mean concentration required over all cell lines, and thus the cell line is unusually sensitive to that compound. Thus, WP760 appeared to be 10-fold to more than 100-fold more cytotoxic against melanoma cell lines than against other cell lines in the NCI panel. These results are even more striking because for most cell lines, the LCs0 values could not determined even at the highest tested concentration (those are proceeded by ">"). This fact is well illustrated by

542

CHEMICAL

AND

MOLECULAR

BIOLOGICAL

APPROACHES

[9.7]

(A) CNS Cancer tO0

I

[

I

I -7

I -6

I -5

50

~o

-50

-lOe

I -8

-9

L o g 10 o f S a m p l e SF-268 SNB-75

SF-295 U251

Concentration

__ ~ _ _ . _ 4,--- -

SF-539

-4

(Molar) _ _ _ A- - -

SNB-19

. . . . . [] . . . . .

(B) Melanoma 1oo

,~..

~

I

I

~,, ,%~,

50

"-.

~o +), \ \

.50

-9

-8

-7 Log l0 of Sample

LOXIMVI SK-MEL-28

0 _ _• __

FIG. 9. Dose-response

MALME-3M SK-MEL-5 curves

__~..._ _. _ ~ _ _

-6

(B).

-5

Concentration MI4 UACC-257

for a typical carcinoma,

and the selective activity in melanoma

...,,

CNS

-4

(Molar) .__.~___ _ .. ~ .. _

cancer,

SK-MEL-2 UACC-62

. . . . . [] . . . . .

with no specific activity (A),

Panel/Cell Line Leukemia CCRF-CEM HL-60(TB) K-562 MOLT-4 RPMI-8226 SR Non-Small Cell Lung Cancer A549/ATCC EKVX HOP-62 HOP-92 NCI-H226 NCI-H23 NCI-H322M NCI-H460 NCI-H522 Colon Cancer COLO 205 HCT-116 HCT-15 HT29 KM 12 SW-620 CNS Cancer SF-268 SF-295 SF-539 SNB-19 SNB-75 U25 ) Melanoma LOX IMVI MALME-3M MI4 SK-MEL-2 SK-MEL-28 SK-MEL-5 UACC-257 UACC-62 Ov~'ian Cancer IGROVI OVCAR-3 OVCAR-4 OVCAR-5 OVCAR-8 SK-OV-3 Renal Cancer 786 0 ACHN CAKI-I RXF 393 SNI2C TK-IO UO-31 Prostate Cancer PC-3 DU-145 Breast Cancer MCF7 NCl/ADR-RES MDA-MB-23 I/ATCC HS 578T MDA-MB-435 MDA-N MG_MID Delta Range

LC50

LOglo LC50 >

m

-4.60

> -4.60 > -4.60 > -4.60 -4.84

m

-460 -460 -4.60 -4.60 -4.60 -6.54 > -4.60 > -4.60 -7.12

m

> -4.60 > -4.60 > -460 > -4.60 > -4.60

m

> > > > >

mm l• /

7. m m

> > > > > >

-4 60 -4.60 -4.60 -4.60 -4.60 -4.60

>

-460 -6.47 -5.95 -686 -6.18 -6.79 -4.71 -6,94

I

-4,60 -4.60 -4.60 -4.60 -4,60 -4.60

ll

> > > > > >

7. I m

/ m I

m

m m

> -4.60 > -4.60 > -4.60 > >

II

-460 -4.60 -4.98

> -4.60 > -4.60 > > > >

-4.60 -460 -4,60 -460

>

-4.60 -4.92 220 2.52 I

I

I

+3

+2

+1

0

I

I

-1

-2

FIG. 10. Excerpt from the NCI Developmental Therapeutics Program mean graphs lbr the 60-eel line screen. LogloLC5o values are shown.

544


[27]

the dose-response curves, and it further indicates that in practice the difference in cell line sensitivity can be even higher. In summary, identified compound WP760 displays high activity against melanoma tumors and it is hoped that its use in treating melanomas may not be limited by myelosuppression typically caused by anthracyclines. These promising in vitro data warrant further studies of WP760, aimed at assessing its clinical potential on the one hand and its DNA-binding properties and sequence specificity on the other hand. Our studies clearly demonstrated that the process of randomly assembling even a limited number of DNA-binding molecular blocks can lead to novel DNA-binding agents with unique biological activity and, consequently, validate a modular design of high-affinity DNA ligands by the combinatorial approach.

S y n t h e s i s a n d E v a l u a t i o n of B i s i n t e r c a l a t o r W P 6 3 1

Synthesis and Chemical Characterization of WP631 The DNR molecule contains a primary amino group at C-Y, and this amine can be selectively reacted (e.g., alkylated or acylated) in the presence of other functional groups present on DNR. The most direct approach to creating a p-xylyl link was direct alkylation using commercially available a,a-dibromo-p-xylyl (Fig. 11 ). Reaction gave one main product, which could be purified by on a conventional silica gel column and thin-layer chromatography (TLC). The exact experimental procedures are given below. Experimental Procedure. Daunorubicin hydrochloride (282 rag, 0.5 mmol) is dissolved in a mixture of N,N-dimethylformamide (DMF) and CH2C12 (1 : 1, v/v; 6 ml). Na2CO3 (100 rag) and a,cd-dibromo-p-xylene (68.7 mg, 0.26 mmol) are then added. The reaction mixture is stirred at room temperature for 2028 hr. The progress of the reaction is monitored by TLC [chloroform-methanolNH4OH(aqueous), 86 : 13 : 1, v/v/v). After the reaction is complete, the reaction mixture is diluted with CH2C12 (100 ml) and poured into water (100 ml). The organic layer is separated and washed with water until neutral pH is attained, dried over anhydrous Na2SO4, and evaporated under diminished pressure. The crude product is purified by column chromatography (Silicage160, 230-400 mesh; Merck, Rahway, NJ), eluted with CHCI3, and then eluted with CHC13-methanol at ratios of 98 : 2 and 95 : 5 (v/v). The final product is isolated as the free amine (WP630) and precipitated from CH2C12-hexane to give a red solid (178 mg, 0.153 mmol), yield 61.5%. Elemental analysis for (C62H64N2020 x H20) C, H, N: IH NMR (CDC13) 6:13.95 (s, 2H, 6-OH, ll-OH), 8.01 (d, 2H, JI,2 ----7.7 Hz, HI), 7.77 (dd, 2H, J2.3 = 7.7 Hz, H-2), 7.38 (d, 2H, J3,2 = 8.5 Hz, H-3), 7.18 (s, 4H, p-xylene), 5.50 (d, 2H, Jl'.2' ----3.4 Hz, H-I'), 5.28 (bs, 2H, H-7), 4.66 (bs, 2H, 9-OH), 4.07 (s, 6H, OCH3), 4.05 (q, 2H, J5',6' = 6.25 Hz, H-6'), 3.77 (d,

[27]


545

CH30 O O~ CH3

I 1 CH30 OII OO~

+

~

r

CH2 Br

HD

HCI.HI~

IH 2

Daunorubicin

o

OH WP631

FIG. 11. Synthesis of WP631. 2H, J = 12.78 Hz, CH2 from p-xylene), 3.63 (m, 4H, CH2 from p-xylene and H-4'), 3.22 (d, 2H, Jl0a,10e = 18.8 Hz, H-10e), 2.96 (d, 2H, Jt0a,10e = 18.8 Hz, H-10a), 2.98-2.93 (m, 2H, H-3'), 2.41 (s, 6H, 14-CH3), 2.37 (d, 2H, Jga,8e = 14.7 Hz, H-8e), 2.09 (dd, 2H, J8a,8e = 14.9 Hz, Jv,8a = 4.1 Hz, H-8a), 1.76 (td, 2H, J2'a,2'e ~---Jz'a,3' = 13.0 Hz, J2'a,l' = 3.8 Hz, H-2'a), 1.68 (dd, 2H, J3',2'e = 4.7 Hz, Jz'a,Z'e = 13.0 Hz, H-2'e), 1.37 (d, 6H, J5',6' = 6.65 Hz, H-6'). The free amine WP630 is suspended in methanol (2 ml). Then dry 1 N HC1 in methanol is added (to pH 4), followed by an excess of diethyl ether to precipitate the hydrochloride of WP630. The red solid is washed with ether until neutral pH and then dried to give analytically pure WP631 (170 mg, 0.138 mmol), 55.3% yield calculated from daunorubicin, rap: 160-170 ° dec. [ot]Dz5 : 184.7 ° (c 0,05, CHC13-CH3OH, 1 : 1). Analysis (C62H65N202o x 2HC1 x 4H20) C, H, C1, N. 25j. E Brandts and L.-N. Lin, Biochemistry 29, 6927 (1990).

546


[27]

DNA-Binding Properties of WP631 Viscosity Studies: Verification that WP631 Binds to DNA by Bisintercalation. In the absence of high-resolution structural data, hydrodynamic studies, especially viscosity, provide the most reliable means of inferring the binding mode of agents that interact with DNA. 26 The theory of Cohen and Eisenberg 27 predicts that, for monointercalation, plots of the cubed root of the relative viscosity [(rl/rl0) 1/3] versus the binding ratio (bound drug per DNA base pair) ought to have a slope of 1.0. The proved monointercalators ethidium and daunorubicin run as controls in the experiment, and both show viscosity changes in excellent agreement with the theory. For bisintercalators, the slope is expected to be twice that observed for monointercalators, an expectation that has been verified for a variety of bisintercalating compounds. The slope observed for WP631 is nearly double that observed for daunorubicin, consistent with a bisintercalative binding mode (Fig. 12). Ultratight Binding of WP631 to DNA. Traditional spectrophotometric methods for measuring affinity fail for ultratight binding expected for WP631. Reliable alternatives for the accurate determination of binding constants are optical melting studies or differential scanning calorimetry. 25'2s 3o UV melting studies were used to determine the binding constant for the interaction of WP631 with herring sperm DNA in BPE buffer (16 mM total Na÷). In the absence of WP631, the Tm of herring sperm DNA was found to be 67.2 °. In the presence of 10 #M WP631, a concentration sufficient to saturate the DNA lattice, the Tm was elevated to 93.5 °. McGhee 29 has shown that the shift in Tm is a function of the ligand-binding constant and site size. Assuming no interaction of ligand with single-stranded DNA, McGhee derived the equation (1/Tin ° - 1/Tin) = (AHm/R) ln(1 + KL) 1/" where Tm° is the melting temperature of the DNA alone, Tm is the melting temperature in the presence of saturating amounts of ligand, AHm is the enthalpy of DNA melting (per base pair), R is the gas constant, K is the ligand-binding constant at Tin, L is the free ligand concentration (approximated at the Tm by the total ligand concentration), and n is the ligand site size. A value of AHm = 7.0 4- 0.3 kcal mol-1 for herring sperm DNA, determined by separate differential scanning calorimetry experiments, was used. From the experimentally determined increase in Tm observed for WP631, a value of K = 8.8 4- 106M -1 (at 93.5 °) was computed, assuming n = 6 bp. Correction of this value to lower temperatures requires knowledge of the binding enthalpy, which was determined by differential scanning calorimetry. 26D. Suh and J. B. Chaires, Bioorg. Med. Chem.3, 723 (1995). 27G. Cohen and H. Eisenberg,Biopolymers8, 45 (1969). 28D. M, Crothers, Biopolymers10, 2147 (1971). 29j. D. McGhee,Biopolymers15, 1345 (1976). 3oG. Hu, X. Shui, E Leng,W.Priebe,J. B. Chaires,and L. D. Williams,Biochemist~36, 5940 (1997).

[27]


547

1.2 pj OJ

j~ °~4

o~ "t"' A

C:,

1.1

.-.......

1.0

0.0

0.1

0.2

Bound Druglb.p FIG. 12. Viscosity studies of WP631-DNA interaction. The cubed root of the relative viscosity

(q/To) is shown as a function of the ratio of bound drug per DNA base pair: solid circles, ethidium bromide; open diamonds, daunorubicin; solid diamonds, WP63 I. Linear least-squares fits to these data yielded the following slopes: ethidium, 0.84 ± 0.02; daunorubicin, 0.89 ± 0.04; WP631, 1.58 ± 0.08.

Experimental for UV Melting Studies. Ultraviolet DNA melting curves are determined with a Cary 3E UV/visible spectrophotometer (Varian, Palo Alto, CA), equipped with a thermoelectric temperature controller. Sonicated herring sperm DNA at a concentration near 20/zM bp in BPE buffer (2 mM NaH2PO4, 6 mM Na2HPO4, 1 mM NazEDTA, pH 7.0) is used for melting studies. Samples are heated at a rate of 1° min -1, while the absorbance at 260 nm is continuously monitored. Primary data are transferred to the graphics program Origin (Microcal, Northampton, MA) for plotting and analysis. Differential Scanning Calorimetry Determination of Binding Enthalpy. Figure 13 shows the results of differential scanning calorimetry experiments using herring sperm DNA in the presence and absence of saturating amounts of WP631. The area under the curves shown in Fig. 13 provides a model-free estimate of the enthalpy of melting of the DNA alone and of the DNA-WP631 complex. By

548

CHEMICAL AND MOLECULARBIOLOGICALAPPROACHES ,

1200

[27]

1

13

._UZ2"

W

W

o I

40

,

50

fl0

,

TO

1

I

I

80

90

100

110

Temperature, oC FIG. 13. Results of differential scanning calorimetry studies of the melting of herring sperm DNA alone (A) or in the presence of near-saturating amounts of WP631 (B). Excess heat capacity (cal mo1-1 °C -I) is plotted as a function of temperature.

Hess's law, these data may be used to determine the enthalpy of WP631 binding to DNA. The equilibria to be considered, along with the experimentally determined enthalpy values, are as follows: Duplex ~- 2(single strand), AHI = 7.0 i 0.3 kcal mol -I WP63 l-Duplex ~ 2(single strand) + WP631, AH2 = 11.44- 1.0 kcal mol -I By combining these two reactions, the binding reaction and its enthalpy may be obtained: WP631-duplex ~ duplex + WP631, AH3 = AH2 - AHI

[27]

NOVEL HIGH-AFFINITYDNA-BINDING AGENTS

549

The enthalpy A//3 needs to be corrected for the amount of WP631 bound to the DNA, and the sign changed, to obtain the binding enthalpy, A Hb = -- A H3/(mol of WP631/mol of bp) From five determinations, a value of AHb = - 3 0 . 2 4-2.6 kcal mol -l was obtained for the association of WP631 with DNA. Complete Thermodynamic Profile for Binding of WP631 to DNA. AHb may be used, assuming that it is constant with temperature, to calculate the binding constant at 20 ° by application of the standard van't Hoff equation, yielding a value of 2.7 × 10 lj M -I . In BPE buffer at 20 °, the binding constant for the interaction of daunorubicin with herring sperm DNA is 1.6 × 107 M-J (E Leng and J. B. Chaires, unpublished data, 2000), so the binding constant of WP631 indeed approaches the value expected for an ideal bisintercalator. The magnitude of the WP631 binding constant approaches that observed for the binding of many regulatory proteins for their specific DNA-binding sites. Knowledge of the binding constant and the binding enthalpy allows us to construct the complete thermodynamic profile for the binding of WP631 to DNA. The free energy is obtained from the standard relation A G ° = - R TIn K, yielding a value of -15.3 kcal mol-] at 20 °. The entropy may be evaluated from the equation - T A S = AG - A H , yielding AS = - 5 1 cal mol-l K-T at 20 °. The thermodynamic profile indicates that the large, favorable binding free energy o f - 15.3 kcal mol ~is derived from the large negative enthalpic contribution of - 3 0 . 2 kcal mo1-1. Binding is opposed by an unfavorable entropic contribution of TAS = - 14.9 kcal tool -I at 20 °. Experimental for Differential Scanning Calorimetry, Differential scanning calorimetry (DSC) experiments utilize a Microcal MC2 instrument along with its DA2 software (July 1986 version) for data acquisition and analysis. Sonicated herring sperm DNA at a concentration of 1 mM bp in BPE buffer is used for all experiments. A scan rate of 1° min J is used. Primary data are corrected by subtraction of a buffer-buffer baseline, normalized to the concentration of DNA base pairs, and further baseline corrected using the Cp(0) software option. Baseline-corrected, normalized data are transferred to Origin graphics software for integration and plotting. Samples for DSC of DNA plus WP631 are prepared by weighing appropriate amounts of solid WP631 and dissolving the solid directly into 2 ml of 1 mM DNA solution. Any undissolved drug is removed by low-speed centrifugation. The exact amount of WP631 bound to the DNA is determined spectrophotometrically.

X-Ray Diffraction and Nuclear Magnetic Resonance Studies of WP631 Complex with DNA Hexanucleotide d( CGATCG)2 The most important evidence that WP631 binds as intended was obtained by X-ray diffraction 3° and NMR 24 analysis of the structure of the WP63 l complex with DNA hexanucleotide d(CGATCG)e (see Fig. 14). Both the X-ray crystal and

550


[27]

FIG. 14. Stereoscopic view of WP631 :DNA hexanucleotide d(CGATCG)2 complex. 24

NMR structures confirm stable binding to d(CGATCG)2 with little distortion of the optimal DNR intercalation sites. The p-xylyl linker is in a different orientation in the two structures; the crystal structure has it parallel to the groove walls and the NMR structure has it perpendicular to the groove walls. This does not affect the linker distance for the two DNR molecules. Fully solvated molecular dynamics simulations show that both orientations are possible and that the p-xylyl position is dynamic in nature. It should be noted that crystallographic and NMR structures both fit data to an empirical force field and, depending on the resolution and nuclear Overhauser effects (NOEs) defining the ligand interactions, respectively, rely to a greater or lesser degree on the particular force field. B i o c h e m i c a l a n d Biological E f f e c t s of W P 6 3 1 WP631 as Potent Inhibitor of Basal and Spl-Activated Transcription in Vitro

The use of transcription assays in vitro to analyze DNA-drug interactions is considered in detail by Phillips et al. ([23] in this volume). 3] Here, we describe a :~l D, R. Phillips, S. M. Cutts, C. M. Cullinane, and D. M. Crothers, Methods Enzymol. 340, [23] 2001 (this volume).

[27]


551

transcription assay system that allows a direct and quantitative disclosure of the effects of bisanthracycline WP631, as well as any other antitumor drug, on in vitro transcription by RNA polymerase I1.5 The templates used are plasmids containing the adenovirus major late promoter linked to a synthetic DNA fragment that lacks cytidine residues on the transcribed strand (a G-less cassette), which generates transcripts without guanosines. Omission of GTP from the reaction leads to elongation arrest at the end of the cassette. In vitro transcriptions are performed in the absence of GTP; in the presence of RNase T1, which acts as a guanylate-specific ribonuclease; and 3'-O-methyl-GTP, which is added to prevent readthrough transcription of the G-less cassette. Using this assay, the unique RNA that accumulates is the RNase Tl-resistant transcript. 32 We sought to determine the influence of WP631 on an Spl-trans-activated promoter in vitro by comparing its effects with those of an Sp i-lacking promoter. The selective targeting of a transcription factor-binding site might provide a way to interfere with transcription-regulatory processes in vivo. Because the Spl site contains (G + C)-rich tracts, including CpG steps, we foresaw WP631 binding to the same regions, although with different binding affinity. In fact, G-less cassettes are suitable for analyzing almost any DNA-binding d r u g . 4'33'34 WP631 was highly efficient at inhibiting transcription initiation when a plasmid containing an Spl-binding site was used under experimental conditions in which Spl acts as a gene activator. 5 The two plasmids used in the analysis of the effects of WP631 on transcription contained a G-less template. 5 This means that we have been able to distinguish unambiguously between effects that are due to the decrease in transcription initiation and those due to elongation. Presented here is a general protocol that, by taking advantage of the characteristics of a G-less cassette, can be used to analyze any DNA-drug complexes. The utility of the protocol is discussed in the context of WP631 and its powerful activity as an inhibitor of Sp 1-activated transcription.5

DNA Vectors pAdML. The pAdML plasmid, 35 also known as pAdML50[180], contains the strong adenovirus major late promoter linked to an ~ 190-bp G-less cassette. pAdSP01. The pAdSP01 plasmid 5 contains an Spl-binding site. It was constructed by inserting the oligonucleotide 5'-GATTCCGGGGCGGGGCGAATC3', which contains a consensus 5'-(GFF)GGGCGG(G/A)(G/A)(C/T)-3' sequence

32 M. Sawadogo and R. G. Roeder, Proc. Natl. Acad. Sci. U.S.A. 82, 4394 (1985). 33 j. Portugal, B. Martfn, A. Vaquero, N. Ferrer, S. Villamarfn, and W. Priebe, Curt. Med. Chem. 8, 1 (2001). 34 C. Cullinane, S. J. Mazur, J. M. Essigmann, D. R. Phillips, and V. A. Bohr, Biochemist~ 38, 6204 (1999). 35 j. Bernu6s, K. A. Simmen, J. D. Lewis, S. 1. Gunderson, M. Polycarpou-Schwarz, V. Moncollin, J. M. Egly, and I. W. Mattaj, EMBOJ. 12, 3573 (1993).

552


[27]

for Sp I binding, 36 in the unique EcoRI site located immediately upstream of the AdML promoter in pAdML. Cell Extracts and Spl Protein. One of the better characterized cell extracts for studying transcription of eukaryotic genes in vitro is derived from HeLa nuclei. Because HeLa cells can be produced in large quantities and are relatively free of nucleases they are a good choice for producing extracts that work well in transcription assays. 37 HeLa cell extracts contain some Spl protein, which is difficult to remove (e.g., it cannot be properly immunoprecipitated from nuclear extracts). The presence of Spl should therefore be considered whenever activated transcription is analyzed quantitatively. The problem of the presence of endogenous Spl protein in HeLa cells can be avoided by using Drosophila embryo nuclear cell extracts, which are considered to lack Spl protein. 38 Nevertheless, these extracts are difficult to prepare for routine use in transcription studies in vitro. Another alternative is other cell nuclear extracts which have been used in the analysis of the effects of the drug on transcription. 4 Transcription in Presence of WP631. Basal and activated transcription assays are carried out in disposable microcentrifuge tubes. A control tube (containing no added drug) is used together with others containing different amounts of WP631 (or any other drug to be analyzed). It is important to avoid introducing RNase activity and all buffers must be prepared with nuclease-free water. Moreover, potassium salts should be used instead of sodium salts to optimize the RNA polymerase II performance. The experiments should be done at 30 ° . General Protocol for Basal Transcription Assays. The protocol to be used for assaying basal transcription is as follows. 1. Prepare, at 4 °, in a final volume of 20 #1, a reaction mixture containing 30 mM HEPES-KOH (pH 7.9), 7 mM MgC12, 5 # M ZnSO4, 1 mM dithiothreitol (DTT), 0.2 mM EDTA, 2% (w/v) polyethylene glycol 8000 (Sigma, St. Louis, MO), 400 # M each of ATP and CTP, 1 # M UTP, 1 mM 3'-O-methyl-GTP (Amersham Pharmacia Biotech, Piscataway, NJ) 15 U ofRNase TI (Calbiochem, La Jolla, CA), and 10/zCi of [~-32p]UTP (800 Ci/mmol; Amersham, Arlington Heights, IL). 2. Briefly warm the reaction mixture at 30 ° . 3. Add 1 /zl of a 0.2-/zg//zl solution of supercoiled pAdML plasmid. 4. Add the drug (WP631) solutions to the different reaction tubes to obtain the desired final concentrations. 5. Add 50 ktg (1-2 #1) of HeLa nuclear extract (the exact amount will depend on the activity of each extract; it is advisable to check beforehand) and nucleasefree water such that the final volume is 25 #1. 36 M. R. Briggs, J. T. Kadonaga, S. P. Bell, and R. Tjian, Science 234, 47 (1986). 37 j. D. Dignam, P. L. Martin, B. S. Shastry, and R. G. Roeder, Methods Enzymol. 101,582 (1983). 38 A. J. Courey, D. A. Holtzman, S. R Jackson, and R. Tjian, Cell 59, 827 (1989).

[27]

NOVEL HIGH-AFFINITYDNA-BINDINGAGENTS

553

6. Incubate at 30 ° for 60 min. 7. Stop the reaction by adding 100 #l of TESS [1 mM Tris-HC1 (pH 8.0), 2 mM EDTA, 300 mM sodium acetate (pH 7.5), and 0.5% (w/v) sodium dodecyl sulfate (SDS)]. In all the transcription experiments, an internal control for recovery and gel loading [a radioactive (--,300 cpm) unrelated transcript, larger than 250 bp] is also added at this stage. 8. Immediately digest the samples with proteinase K (0.2 mg/ml) at 60 ° for 15 min and extract the entire reaction with 200 #1 of phenol-chloroform. 9. Transfer the upper aqueous phase to a new tube and precipitate it with 100% ethanol. Wash the pellet with 75% (v/v) ethanol. 10. Resuspend the dried pellet carefully in 8-10 #1 of a loading solution consisting of 95% (v/v) formamide, 10 mM EDTA (pH 7.0), 0.02% (w/v) xylene cyanol, and 0.02% (w/v) bromophenol blue. 1 I. Analyze the transcripts by high-voltage electrophoresis in 90 mM Trisborate and 2 mM EDTA (pH 8.3), using 30-cm-long 8% (w/v) polyacrylamide gels containing 7 M urea. 12. After electrophoresis, soak the gel in distilled water, dry under vacuum, and subject it to autoradiography.

Protocolfor Assaying Spl-Activated Transcription. The protocol to be used for assaying Sp 1-activated transcription is practically the same as described above for the basal transcription assays. However, the pAdSP01 plasmid, which contains an Spl-binding site, is used (see above). In step 5, about 20 ng of pure recombinant Spl protein is also added. Pure protein can be prepared in the laboratory36 or purchased from Promega (Madison, WI). Quantitative analyses of transcripts can be carried out with any densitometer. The relative amounts of transcripts observed should be normalized to the total amount of radioactivity loaded, using the internal control for recovery and gel loading. The C50 values (the drug concentrations that reduce electrophoretic band intensity by 50%) can be derived from plots of percentages of transcription versus drug concentration. These C50 values are then used to compare the effects of WP6315 (or any other drug) on basal and activated transcription and to compare the effects of WP631 (or any other drug) with those of other DNA-binding drugs. The inhibition of transcription from either promoter by increasing concentrations of WP631 is apparently due to the effect on transcription initiation by RNA polymerase II. Because there are no CpG or GpC steps in a G-less cassette, there should be no preferred intercalating sites in the transcribed region that would have an effect during the elongation step. WP631 differentially influences promoter-initiated transcription in vitro (Fig. 15). When transcription was analyzed with the AdML promoter, the C50 for WP631 was 0.48 #M. In contrast, when transcription was analyzed using the

554


AdML promoter

AdSP01 promoter

WP631

Control 0.44

1,76

2.2

[27]

WP631

2.64 3.52 !aM Control ÷ Spl 0.02

0.07 0,20

0.50 0.90 tdv

,~_Q, I00"~ "6 z~ O

E

2 s ~ ff

o

%% WP631 [~tM]

0 , ,~

,

~, ' ,7

~

xooOo, % % WP631 [laM]

FIG. 15. Inhibitionof RNApolymeraseII transcriptionin vitroby WP631.Left: Effectof increasing concentrationsof WP631 on basal transcriptionfrom the AdML promoter. Right: Effect of WP631 on Sp 1-activatedtranscriptionfrom the AdSP01 promoter. In both panels, the upper part displaysthe results of electrophoretic gel analysisof the RNA transcripts, while the bottom part corresponds to the results of quantitativeanalysisof the relativeamountof transcription.Note that in the experiments shown, an internal standard for recovery and gel loading was used (see text), but is not shown. The results presented clearly agree with those we have described previously5 and emphasize the strong effect of WP631 on activatedtranscriptionin vitro. AdSP01 promoter, the ability of WP631 to inhibit Sp 1-activated transcription was outstanding and the C50 was in the low nanomolar range (only 60 nM). 5 WP631 was also more efficient at inhibiting activated transcription than basal transcription; at a concentration as low as 0.5 #M, it totally inhibited Sp l-activated transcription in vitro. In comparative experiments performed with daunorubicin, the concentrations of this anthracycline that inhibited the transcription of either promoter were similar. 5'33 The strong effect of WP631 on AdSP01 was fairly distinct (Fig. 15) and consequently rather specific.

Using Gel Retardation and DNase I Footprinting Assays to Complement Information Provided by Transcription Assays in Vitro. More direct evidence of the inhibitory effects of WP631 on Sp 1 binding can be obtained by gel retardation (band shift) and DNase I footprinting.5'33 General protocols of the gel retardation technique are widespread, and the characteristics of the techniques have been discussed in detail elsewhere. 39

[27]


555

Gel retardation experiments are carried out with a labeled Sp 1 oligonucleotide and pure Spl protein. Gel retardation (band-shift) assays using WP631 have been performed in 10 mM Tris-HC1 (pH 7.4) containing 50 mM KC1, 1 mM MgCI2, 0.2 mM EDTA, 0.5 mM DTT, bovine serum albumin (30 #g/ml), and 5% (v/v) glycerol. A typical reaction contains about 20 ng of pure Spl protein and about 2 nmol (in base pairs) of the end-labeled double-stranded oligonucleotide 5'GAATTCGGGGCGGGGCGAATTC-3' in the presence of 1 #g of poly[d(I-C)] (Boehringer Mannheim, Indianapolis, IN). Thereafter, the samples are analyzed on nondenaturing 4.5% (w/v) polyacrylamide gels containing 45 mM Tris-borate and 1 mM EDTA (pH 8.3). The gels are run at low voltage (12 V/cm), dried under vacuum, and subjected to autoradiography. The results obtained with WP631 attest that this bisanthracycline might block the binding of Spl to its putative binding site with high efficiency.33 Drug-mediated inhibition of the interaction between Spl and the oligonucleotide shows the destabilizing effect of WP631 on Spl-DNA complexes (it is also possible to use longer DNA fragments containing an Sp 1 site). In the presence of decreasing concentrations of WP631, the formation and stability of the S p l DNA complexes was more apparent, indicating that the DNA-protein complex was sensitive to the bisanthracycline in a concentration-dependent way. DNase I footprinting assays of the binding of WP631 and the Sp I protein to the same sequence can be performed with the pAdSP01 plasmid, which contains a putative Spl-binding site. A 352-bp DNA fragment purified from pAdSp01 can be labeled at the 5' end of the upper strand with [y-32p]ATP and T4 polynucleotide kinase. 5 General aspects of a footprinting assay and details of its use in DNA-drug interactions are considered in [20] in this volume. 4° In a standard experiment, samples containing 3000 cpm of an end-labeled DNA fragment (about 10 pmol in base pairs), along with different concentrations of WP631 and/or of pure Spl protein, are supplemented with a buffer consisting of 10 mM Tris-HC! (pH 7.4), 50 mM KC1, 2 mM MgC12, 2 mM MnC12, 0.5 mM DTT, and 5% (v/v) glycerol to a final volume of 20/zl and then digested with DNase I (Boehringer Mannheim) at a final concentration of 0.01 U/ml for 2 min at 30 °. The reaction mixture is phenol extracted and ethanol precipitated, and the resulting pellet is dissolved in 85% (v/v) formamide, 10 mM EDTA, and 0.02% (w/v) bromphenol blue. Samples are heated at 95 ° for 2 min prior to electrophoresis. Footprints are resolved by high-voltage electrophoresis in 90 mM Tris-borate and 2 mM EDTA (pH 8.3), using 6% (w/v) polyacrylamide gels containing 8 M urea. Our footprinting studies performed in this way have revealed that a ternary complex is formed with some concentrations of Spl and WP631 on the Splbinding site of the AdSP01 promoter. 33 39 j. B. Taylor, A. J. Ackroyd, and S. E. Halford, Methods Mol. Biol. 30, 263 (1994). 4o K. R. Fox and M. J. Waring, Methods Enzymol. 340, [20] 2001 (this volume).

556


[28] D e s i g n , S y n t h e s i s , and of Polyintercalating

[28]

Characterization Ligands

By VLADIMIR M. GUELEV, M A R K S. CUBBERLEY, MEREDITH M. M U R R , R . SCOTT LOKEY,

and

BRENT L. IVERSON

Introduction Functionalized DNA polyintercalators that bind to DNA in a sequence-selective manner 1-3 represent, at least in principle, an attractive strategy for targeting cellular DNA. As such, polyintercalation could be considered an alternative to approaches such as triple helix-forming oligonucleotides,4,5 minor groove-binding molecules,6 peptide nucleic acids,7 and modified zinc fingers/Indeed, the idea of connecting several intercalating groups via appropriate linkages to create DNAbinding agents with high affinity is certainly not new.9-13 Synthetic trisintercalators have been reported, most notably a polyamine-linked trisacridine compound synthesized by Laug~a et al., 13 that binds to DNA with an affinity n e a r 1014 M -1. Despite noteworthy successes such as this, attempts to extend interactions beyond bisintercalation have often failed to significantly improve binding affinities. In addition, the key issue of sequence specificity has hardly been addressed with polyintercalating systems. Design Considerations There are potential difficulties associated with the design of polyintercalators. As noted above, linking multiple intercalating moieties together does not necessarily result in a corresponding increase in binding constant or even in full 1 j. Chaires, Curt Opin. Strucr Biol. 8, 314 (1998). 2 L. Hurley and J. Chaires (eds.), "Advances in DNA-Sequence Specific Agents," Vol. 2, JAI Press, Greenwich, Connecticut, 1996. 3 E Nielsen, Chem. Eur. J. 3, 505 (1997). 4 D. Gowers and K. Fox, Nucleic Acids Res. 27, 1569 (1999). 5 j. Franqois, J. C. Lacoste, L. Lacroix, and J. L. Mergny, Methods Enzymol. 313, 74 (2000). 6 S. White, J. Szewczyk, J. Turner, E. Baird, and E Dervan, Nature (London) 391, 468 (1998). 7 M. Egholm, O. Buchardt, L. Christensen, C. Behrens, S. Freier, D. Driver, R. Berg, S. Kim, B. Norden, and P. Nielsen, Nature (Londan) 365, 566 (1993). 8 A. Klug, J. Mol. Biol. 293, 215 (1999). 9 L. Wakelin, Med. Res. Rev. 6, 275 (1986). 10 G. Atwell, W. Leupin, S. Twigden, and W. Denny, J. Am. Chem. Soc. 105, 2913 (1983). II j. Hansen, T. Thomsen, and O. Buchardt, J. Chem. Soc. Chem. Commun, p. 1015 (1983). 12 W. Denny, G. Atwell, G. Willmott, and L. Wakelin, Biophys. Chem. 22, 17 (1985). 13 p. Laugfia, J. Markovits, A. Delbarre, J. Le Pecq, and B. Roques, Biochemistry 24, 5567 (1985).

METHODS1NENZYMOLOGY,VOL.340

Copyright~ 2001 by AcademicPress All rightsof reproductionin any formreserved. 0076-6879/(X)$35.00

[28]

POLYINTERCALATING LIGANDS

557

intercalation. Likely explanations include the entropic cost associated with immobilizing a long flexible molecule on the DNA and/or self-stacking of the adjacent aromatic units in a polyintercalator in the absence of DNA. In addition, there are many examples of naturally occurring and synthetic DNA mono- and bisintercalators 1-3,9 that are well studied, but there are no known examples of trisor tetra-, etc., polyintercalators found in nature on which to base synthetic designs. The pursuit of predictable sequence specificity in a high-affinity polyintercalating system will require a detailed structural understanding, but unfortunately the characterization of DNA-bound structure is complicated in the case of polyintercalators. For example, a presumed polyintercalator can potentially have multiple binding modes; accordingly threading versus nonthreading intercalation (see below) and even groove binding should be considered.14 In addition, if polyintercalation is confirmed, then the linker could be placed in the major or minor grooves and there could be various numbers of base pairs between consecutive intercalated aromatic units. Finally, and perhaps most challenging, polyintercalation involves considerable distortion of the DNA helix, the structural and energetic consequences of which must be taken into account in the design of sequence-specific agents. ~5,~6 On the basis of the above considerations, a viable DNA-readout polyintercalator system should be modular, so that the design of the longer polyintercalator analogs can be based on detailed studies of the shorter modules, that is, dimers. Ideally, the linkers will be held deep in one or both grooves of DNA, facilitating sequence-specific interactions. A viable polyintercalating system also needs to be easily synthesized, including the ability to create numerous derivatives to provide for various sequence specificities. Finally, the system needs to be readily characterized structurally when bound to DNA, presumably via multidimensional nuclear magnetic resonance (NMR). Although initial polyintercalator designs were based on attaching several chromophores as side chains of a linear linker, 9-13 Takenaka et aL constructed a trisintercalator with chromophores attached in a head-to-tail fashion.~7 On the basis of the ability of the central intercalator to "thread through" DNA and access both DNA grooves simultaneously, a threading binding mode was proposed for this molecule, in which the linker alternates between the major and minor grooves of the DNA helix.IV' 18 Aside from purely aesthetic appeal, such a binding mode may have several advantages pertaining to biological activity, such as simultaneous blockade of both DNA grooves, improved sequence specificity, and slower dissociation rates from cellular DNA.

14 D. Suh and J. Chaires, Bioorg. Med. Chem. 3, 723 (1995). 15 E Gago, Methods 14, 277 (1998). J6 j. Chaires, Biopolymers 44, 201 (1997). 17 S. Takenaka, S. Nishira, K. Tahara, H. Kondo, and M. Takagi, Supramol. Chem. 2, 41 (1993). J8 C. Lamberson, Ph.D. Thesis, pp. 23-26. University of Illinois, Chicago, Illinois, 1991.

558

Ht~3N


N"A'~-"

,~.~N

2H

4 H

~

O

[9,8]

H

O

n=0-8

NDI amino acid FiG. 1. General structure of the polyintercalators, synthesized by standard SPPS techniques. When specifying polyintercalator composition in text, the 1,4,5,8-naphthalenetetracarboxylic diimide amino acid is referred to as NDI.

1,4,5,8-Naphthatenetetracarboxylic diimide (NDI) derivatives have been shown to monointercalate DNA in a threading manner. 19 They are chemically robust and are easily derivatized into amino acids in which the amino and carboxyl termini are on opposite sides of the molecule (Fig. 1). These features make the NDI unit particularly well suited to be a building block for the construction, via solid-phase peptide synthesis (SPPS) methodology, of modular polyintercalating molecules with a putative threading mode of DNA binding. With the above considerations in mind, we devised a series of polyintercalators in which a number of intercalating NDI units are strung together in a head-to-tail fashion (Fig. 1). 20 The compounds are synthesized by standard 9-fluorenylmethoxycarbonyl (Fmoc)-based solid-phase methods, 21,22 which readily allows for tremendous variation in the linker composition as well as the facile incorporation of even large numbers of intercalating aromatic units. Initially, we reported the first DNA tetraintercalator based on chains of NDI units linked via Gly-Gly-Gly-Lys peptides, 2° and more recently we have expanded the number of intercalating cbromophores to eight units, 23 which appear to fully 19 E Tanious, S. Yen, and W. D. Wilson, Biochemisto' 30, 1813 (1991). 2o R. S. Lokey, Y. Kwok, V. Guelev, C. Pursell, L. Hurley, and B. Iverson, J. Am. Chem. Soc. 119, 7202 (1997). 1 E. Atherton and R. C. Sheppard, "Solid Phase Peptide Synthesis--A Practical Approach." IRL Press, Oxford, (1989). 22 NovaBiochem, "NovaBiochem Catalog & Peptide Synthesis Handbook." NovaBiochem, La Jolla, California, 1999. 23 M. Murr, M. A. Thesis. University of Texas, Austin, Texas, 1999.

[28]

POLYINTERCALATING L|GANDS H2N~

NBOC

TsO HaN O ~ "O" "O

559

Statistical Mixture: RI= R2 = CH2CH2CO2Bz RI= CH2CH2CO2Bz R2 = CH2CH2NBOC RI= R2 = CH2CH2NBOC

OBz

DIEA, i PrOH 94%

O" "N" "O

1 H2, Pd/C 36%

1.) TFA/CH2CI 2

/\

2.) FMOC-AA-OPfp, .OBT, O" "N" "O

/

O

NH

2,6-1utidine, NMP 90 %

O~ "N" "O

/

/

NBOC

NFMOC

2

2a

R~[~NFMOC

3

SCHEME I. Synthetic scheme for the generation of the NDI-amino acid 3, used in the solid-phase synthesis. Although amino acid 2a would be a more convenient building block in the synthesis, we have been unable to isolate this derivative in high yield (see text).

intercalate on DNA binding according to viscometry and UV measurements. We have also used combinatorial methods to isolate a linker that causes a dramatic change in DNA sequence specificity.24 Below we describe a general protocol for the synthesis of NDI-based polyintercalators of up to eight chromophores, as well as the synthesis of mixtures of bisintercalators for library studies. S y n t h e s i s of 1 , 4 , 5 , 8 - N a p h t h a l e n e t e t r a c a r b o x y l i c Diimide C h r o m o p h o r e - A m i n o Acid (3) Scheme 1 (1-3) describes the synthesis of the diimide amino acid 3. Ideally, it would be preferable to eliminate any in-solution couplings, by starting with a protected amino acid such as 2a. Unfortunately, this derivative was found to be highly unstable, having a propensity to spontaneously lose the Fmoc-protecting 24 g. Guelev, M. Harting, R. S. Lokey, and B. Iverson, Chem. Biol. 7, l (2000).

560


[28[

group. The mechanism of this decomposition reaction has not yet been investigated. Therefore, the solid-phase synthesis starts with a diimide-amino acid adduct 3, prepared in solution as follows.

N-( 2-te rt-Buto xycarbonylaminoethyl )-N'-( 2-carbo xyethyl )1,4,5,8-naphthalenetetracarboxylic Diimide (2) 1,4,5,8-Naphthalenetetracarboxylic dianhydride (1) (8.96 g, 33.4 retool) is combined with mono-tert-butoxycarbonylaminoethylamine (5.34 g, 33.4 mmol) and fl-alanine benzyl ester p-toluene sulfonate salt (11.72 g, 33.4 mmol) in 2propanol (450 ml). N,N-Diisopropylethylamine (6.4 ml, 37 mmol) is added to the suspension and the mixture is heated at reflux for 19 hr under an inert atmosphere. The mixture is allowed to cool to room temperature and the suspended solid is collected by filtration, washed thoroughly with diethyl ether, and dried under high vacuum. The statistical mixture of diimide products (17.9 g, 94%) is used directly in the next step without further purification. The above mixture of diimide products is suspended in absolute ethanol (250 ml) and purged with argon. Pd/C (9 g, 10%) is added and the mixture is charged with H2. After 48 hr, the mixture is concentrated to 50 ml (bath temperature 30 °, 20 torr) and diluted with 15% (v/v) triethylamine (TEA)-CHzCI2 (250 ml). The suspension is filtered through Celite, which is rinsed with 10% (v/v) TEA-CHzCI2 until the filtrate is colorless. The filtrate is concentrated to dryness and the resulting solid is loaded onto a large column of silica equilibrated with 10% (v/v) TEA-CH2CI2. Flash chromatography using a gradient of 0 --+ 5% (v/v) methanol in 10% (v/v) TEA-CH2CI2 yields the TEA salt of the title compound as the second yellow band. The product fractions are concentrated to dryness and the resulting solid is dissolved in a minimal amount of 10% (v/v) methanol-CHzCl2. Acetic acid (20 ml) is added followed by the slow addition of hexanes (300 ml). The resulting solid is collected by filtration, triturated several times in methanol to remove residual acetic acid, and dried in vaeuo to yield the title compound 2 as a white powder (5.5 g, 73% based on statistical mixture).

N-2-(N ~-9-Fluorenylmethoxycarbonyl-N~-tert-butoxycarbonyl)lysylaminoethyl N'-( 2-carboxyethyl )- l, 4,5,8-naphthalenetetracarboxylic Diimide (3) The following procedure works for the preparation of most Fmoc-protected amino acids, except for Fmoc-glycine, for which a different workup is necessary because of poor solubility (see below). The t-butyloxycarbonyl (Boc)-protected NDI amino acid 2 (1.5 g, 3.1 mmol) is suspended in CH2C12 (10 ml) and trifluoroacetic acid (TFA, 10 ml) is slowly added. After standing for 15 min, the solvent is evaporated and residual TFA is removed by azeotropic evaporation from heptane (twice). The resulting residue is dried in vacuo for 4 hr. The TFA salt is then

[28]


561

combined with N-9-fluorenylmethoxycarbonyl-(N~-tert-butoxycarbonyl)lysine (Fmoc-Lys-OH, 2.2 g, 3.2 mmol) and 1-hydroxybenzotriazole (HOBO (0.48 g, 3.1 mmol) in N-methylpyrolidone (NMP, 10 ml). 2,6-Lutidine (1 ml, 6 mmol) is added and the suspension is stirred until clear (approximately 2 hr). Adduct Workup of Lysine and All Other Amino Acids The reaction mixture is diluted with ethyl acetate ( 1000 ml) and quickly washed with 0.1 N HC1 (3 x 100 ml). The organic solution is reduced in volume to "-~500ml and 100 ml ofhexane is added. The resulting solid is collected by filtration and dried in vacuo over KOH to yield the title compound 3 as a yellow powder (2.35 g, 90%). Glycine Adduct Workup The reaction mixture is poured into 200 ml of citrate buffer (0.2 M, pH 4) with stirring and filtered. The filtered material is washed with H20 and dried in vacuo. The dry solid is ground to a fine powder, triturated with diethyl ether and with ethyl acetate several times, and dried in vacuo. Solid-Phase Synthesis Coupling/Deprotection Cycle Reagents Needed Fmoc-amino acid-NovaSyn TGA resin N-methylpyrrolidinone (NMP) 2-propanol dimethyl sulfoxide (DMSO) piperidine (20%, v/v) in NMP N-methylmorpholine (NMM) (Benzotriazol-l-yloxy)tris(pyrrolidino)phosphonium hexafluorophosphate (PyBOP) Fmoc-amino acid-NDI adduct 3 Fmoc-protected amino acids Equipment Needed Solid-phase peptide synthesis (SPPS) reaction vessel with coarse frit filter (freshly silanized) Rotary shaker The Fmoc-amino acid-NovaSyn TGA resin (1 equivalent) is added to the solidphase reactor followed by enough NMP to cover the resin. The resin is allowed to swell for 15-20 min.

562


[28]

Deprotection and Wash. The NMP is drained from the resin and piperidineNMP (10 ml/g; 20%, v/v) is added. The mixture is shaken for 15 min before the solvent mixture is drained from the resin and a second portion of piperidine-NMP is added. After 15 min, the solvent mixture is removed and the resin is washed with NMP (three times), 2-propanol (three times), and NMP (three times) in preparation for acylation. Coupling. The Fmoc-amino acid-NDI adduct 3 (3 equivalents) and PyBOP (3 equivalents) are completely dissolved in NMP (sonicate if necessary) to make a 0.2-0.4 M solution. The solution and NMM (6 equivalents) are added to the resin and the mixture is shaken for 45 min. The solvent is drained and the resin is washed with NMP, after which the coupling is repeated. The resin is washed twice with DMSO (to remove all unreacted NDI amino acid) and followed by the usual NMP-2-propanol-NMP wash sequence. Capping. The unreacted amino termini are capped with 5% (v/v) acetic anhydride-5% (v/v) 2,6-1utidine in NMP (30 min). The wash/deprotection/wash/acylation cycle is continued until the desired sequence is achieved. Double couplings and additional DMSO washes are necessary only for the NDI amino acid. Cleavage from Resin and Side Chain Deprotection Reagents and Equipment Needed Glacial acetic acid Dichloromethane (CH2C12) Methanol Diethyl ether Trifluoroacetic acid (TFA) : water mixture (95 : 5, v/v) If Trp, Met, Tyr, Cys, and Phe are used in the synthesis, also needed are phenol, anisole, and 1,2-ethanedithiol (EDT) as cation scavengers.

Equipment Needed Conical flask (25 ml) Plastic centrifuge tube (50 ml) Centrifuge

Procedure. When the desired sequence is achieved, the terminal Fmoc group is removed as described above. The resin is then washed with NMP (three times), acetic acid and CH2C12 (three times), and methanol (three times) before the resin is dried over KOH under high vacuum for several hours. The TFA cleavage mixture (20 mug) is added and the solution is mixed gently. After 1 hr the TFA

[9.8]


563

solution is drained into the 25-ml conical flask and the resin is washed with undiluted TFA (two times). The TFA solution is concentrated and added dropwise to cold diethyl ether (50 ml/g) in a 50-ml centrifuge tube. The solution is allowed to precipitate for 12 hr at - 2 0 °. The precipitate is collected by centrifugation at 2000 rpm at 0 ° for 10 min. The ether is removed by decantation and additional ether is added. The suspension is triturated for several minutes and the precipitate collected as before. The trituration is repeated and the collected precipitate is partially dried under reduced pressure to remove any residual ether. Caution: Drying too long may make the compounds difficult to dissolve. The precipitate is dissolved in ~ 5 - 8 ml of water, filtered, and purified as described below. Alternatively, the solid is extracted with diethyl ether and the residual ether is removed from the aqueous layer by rotary evaporation (at room temperature) prior to purification. No~s 1. The Kaiser test for free amino termini does not give reliable results once

the first NDI residue is introduced. 2. Our experience shows that the NDI-amino acid coupling is by far the most difficult. The problem may be due to aggregation of the diimide chromophores. a. TentaGel-type resins (Rapp Polymere, Tiibingen, Germany) work somewhat better than regular polystyrene/divinylbenzene (DVB) resins. b. Double couplings of--~45 min (with PyBOP as the coupling reagent) are required to achieve coupling efficiency >90%. Capping after the first coupling with acetic anhydride ensures lack of nondiimide by-products that can go undetected. c. N-Methylpyrrolidinone (NMP) gives better results as a solvent than N,Ndimethylformamide (DMF) or N,N-dimethylacetamide (DMA). An additional DMSO wash after the couplings of the NDI amino acid removes any residual excess reagent. 3. On the basis of analysis of library mixtures, we have not detected any other problematic coupling steps or specific "difficult sequences. ''22 However, coupling efficiency and the resulting product purity deteriorate rapidly with increasing number of NDI units. Whereas product purity is "~90% for the dimer analogs (Fig. 2), it was only "-~10% for the octamer. 4. N-Fmoc-protected amino acids (or oligopeptides), when not commercially available, can be readily prepared from the unprotected material, using 9-fluorenylmethyloxycarbonyl-N-hydroxysuccinimide(Fmoc-OSu), according to the procedure of Lapatsanis et al. 25 25L. Lapatsanis,G. Milias, K. Froussios,and M. Kolovos,Synthesis, p. 671 (1983).

564

CHEMICAL AND MOLECULARBIOLOGICALAPPROACHES a

[28]

1,2 3,4

.

.

.

.

.

.

250

.

.

.

.

,

.

,

.

.

300 350 Wavelength (nm)

.

.

,

400 3

5

10 15 Time (min)

20

25

FIG.2. HPLC analysisof the crude productfrom the synthesisof the bisintercalatorLys-NDI-GlyGly-Gly-Lys-NDI-Gly.Gradient:85% water [0.1% (v/v)TFA] ~ 60% CH3CN[0.1% (v/v)TFA]over 60 min at 0.75 ml/min.Usingthe UV spectra (a) the four majorproducts seen in the chromatogram(b) can be differentiatedas containingone NDI (peaks 1 and 2) and two NDI (peaks 3 and 4) units. Mass spectral analysisidentifiedpeak 3 as the desired product. Peaks 1 and 2 are due to monomericNDI by-products and peak 4 (mass 14 units greater than the major product) is thought to be an N-oxide, formed by oxidationof a lysine side chain.

5. If desired, the number of coupling steps can be reduced by using short peptide segments, for example, Fmoc-N-Gly-Gly-Gly-OH (see Note 1).

P u r i f i c a t i o n b y P r e p a r a t i v e C18 H i g h - P e r f o r m a n c e Liquid Chromatography

Reagents Needed High-performance liquid chromatography (HPLC)-grade acetonitrile Distilled water Trifluoroacetic acid (TFA)

Solvent Systems Solvent A: 0.1% TFA-CH3CN (v/v) Solvent B: 0.1% TFA-water (v/v)

Procedure Typically, a gradient of 100% solvent B -+ 0% solvent B over 240 min at 0.75 ml/min yields good results. Holding the gradient (e.g., at 85% solvent B for dimeric derivatives) is sometimes necessary to resolve especially impure mixtures.

[28]

POLYINTERCALATING LIGANDS NH3

HaN

+NH 3

N~ 0

565

N~

X Y Zd

"~-E "0

H 0

N---if.O (~ ~'~--~' ~

H 0

x,z-{

Skip

.

o

y =

H 0

0

H 0

H 0

FIG. 3. General structure of a 360-member library of bisintercalators synthesized to study the effect of amino acid linker on DNA sequence specificity} 4

Library Synthesis Figure 3 shows the composition of a 360-member library synthesized as described below} 4 Equipment. Synthesis is performed on a parallel synthesizer that consists of 30 open sintered glass filter tubes attached to a Teflon block, equipped on the bottom with a vacuum outlet/argon inlet. The reactor is mounted on a rotary shaker. Note: Coupling conditions are the same as described above for the single compounds.

Procedure The NDI-amino acid coupling is performed in one vessel; prior to deprotection the resin is divided by suspending it in NMP in a graduated cylinder and, while stirring, pipetting out equal amounts into the separate reaction vessels. Deprotection is performed in parallel, with a different amino acid added to each vessel. The resin is then recombined and redivided, and deprotection/coupling is repeated the desired number of times. To increase the number of compounds in a mixture, we have added two amino acids per coupling reaction. To minimize differences in

566


[28]

reactivity, we group the amino acids, separating primary from secondary amino acids. Half an equivalent of amino acid is used to allow for shorter peptide linkage ("skip" residue). The mixtures are cleaved with approximately 0.5 ml of TFA-H20 (95 : 5, v/v) per 50 mg of resin (adding more TFA would prevent precipitation). The solution is collected in 15-ml centrifuge tubes and the resin is washed with a minimal amount of TFA (~0.5 ml). The product is precipitated with 10 ml of diethyl ether and centrifuged, and the ether is decanted. This step is repeated two more times, the product is then extracted between water and diethyl ether, and the aqueous layer is loaded on a Waters (Milford, MA) Sep-Pak cartridge, washed with several volumes of 0.1% (v/v) aqueous TFA, eluted with a 70 : 30 mixture of water-acetonitrile [0.1% (v/v) TFA each], and lyophilized. Note: If a given mixture does not precipitate on treatment with ether, the procedure described above for the single compounds should be used. Alternatively, the product is extracted with water, neutralized with triethylamine and lyophilized, and then loaded on a Sep-Pak column and purified as described above. Characterization Because of the acetic anhydride capping of the resin after the first NDI coupling, all impurities in the crude product contain at least one NDI moiety that can be used as a diagnostic chromophore to facilitate characterization.

High-Performance Liquid Chromatography Efficient separation of the crude product or of library mixtures of derivatives is achieved on C 18columns with a gradient of 100% water [TFA, 0.1% (v/v)] --+ 100% CH3CN [TFA, 0.1% (v/v)] over 180 min at 0.75 ml/min. Generally, the compounds with higher numbers of NDI units have longer retention times. Depending on the compounds analyzed, the conditions can be optimized, for example, in the case of mixtures of bisintercalators, a gradient of 85% water [TFA, 0.1% (v/v)] 60% CH3CN [TFA, 0.1% (v/v)] over 60 min at 0.75 ml/min is sufficient and can resolve two bis-NDIs with diastereomeric linkers. Monitoring is by UV (see below) at ~. = 384 nm against a baseline at )~ = 450 nm, or, if available, by liquid chromatography-electrospray ionization (LC-ESI; see below).

Ultraviolet-Visible Spectroscopy The naphthalene diimide chromophore exhibits a characteristic UV spectrum. Generally, because of intramolecular stacking of the chromophores in the compounds with more than one intercalating unit, a hypochromic effect leads to a distinct UV profile for those compounds as compared with the NDI monomer (Fig. 2a). This often allows the unambiguous assignment of the desired multimeric

[28]


567

TABLE I EXTINCTIONCOEFFICIENTSAT 386 nm FOR NDI OLIGOMERS Compound

Oligomer

na

~F b

Gly-NDI-Lys Gly-NDI-Lys-GIy-Gly-GIy-NDI-Lys GIy-(NDI-Lys-Gly-GIy-GIy)2-NDI-Lys GIy-(NDI-Lys-GIy-GIy-GIy)3-NDI-Lys GIy-(NDI-Lys-GIy-Gly-Gly)7-NDI-Lys

Monomer Dimer Trimer Tetramer Octamer

0 I 2 3 4

20,000 27,400 39,200 51,300 104,500

~'SDSc 23,400 44,000 74,600 96,000 190,000

a Refer to Fig. 1. b Measured in 10 mM Tris buffer, 1 mM EDTA, 50 mM NaCI, pH 7.4. "Measured in 2% SDS.

product from the commonly present monomer impurities by simple HPLC/UV (Fig. 2). For crude concentration estimates, the extinction coefficients shown in Table I can be used for any derivative with a given number of diimide chromophores. However, the observed hypochromic effect is somewhat sensitive to linker variations; therefore, independent measurement of the extinction coefficient for each derivative is important for accurate DNA binding studies such as quantitative footprinting.

Electrospray Ionization-Mass Spectrometry Compounds with basic amino acid side chains are conveniently analyzed by ESI-mass spectrometry (MS) in the positive ionization mode. All possible ionization states are usually observed, with the relative intensities of the different species somewhat dependent on the conditions. If the basicities of the components of a mixture are similar, as is the case in a library mixture of compounds with one lysine residue per NDI unit, the relative amounts of compounds are roughly proportional to the sums of all ionized species observed. Figure 4 and Table II show the ESI-MS data for a mixture of bisintercalators, 24 synthesized as described above (without double coupling for the second NDI chromophore).

Future Directions We have demonstrated that a modular DNA polyintercalator system, amenable to a great number of structural variations, is accessible via facile solid-phase synthetic procedures. 2°,23,24 Therefore, a number of the technical challenges associated with producing polyintercalating agents with antigene activity have been successfully addressed. The next step in the design of the sequence-specific polyintercalators requires structural studies of bisintercalating modules bound to their

568


[9,8]

1356.6

1oo-~

°%° !I

90 85 60

Ob 75

7"

890~8 i~08

65 60

1285,7 b13m99"6

28,rnI Im114006

679.2

o~ 45

did

>_ 40

(D tY

25 20~ 3646 422.5

',',

' .....

400

,,,,,J

66 567.6 d I I ~]

....

i ........

600

i.,,,i,

8OO

1000

12O0

1400

..... , ..... ;',',,,,

m/z FIG. 4. ESl-mass spectrum of a mixture of bisintercalators synthesized as described in text. The data are tabulated in Table I1. All 11 nonisobaric compounds were identified as the monoprotonated (m) and/or diprotonated (d) species. Most impurities were identified as monomeric NDI derivatives, resulting from incomplete second NDI coupling (i).

preferred DNA sequences. In this respect, there have been several encouraging developments in bisintercalator design.

Binding Affinity Chaires and co-workers have designed an ultratight binding bisdaunorubicin, WP631,26-28 based on the crystal structure of two daunorubicin monomers complexed to DNA, demonstrating the degree to which structural data can be used to fine-tune the energetics of bisintercalation. Interestingly, the daunorubicin monomers, as well as the dimer with optimal linker, span not two, but four DNA base pairs between the two intercalating aromatic units. 27 We have carried out initial NMR studies of a bis-NDI module, and the resulting structure indicates four base pairs might be the optimal inter-intercalator distance for our NDI systems as well, at least for certain DNA sequences.29 Another 26 j. Chaires, E Leng, T. Przewloka, 1. Fokt, Y. Ling, R. Perez-Soler, and W. Priebe, J. Med. Chem. 40, 261 (1997). 27 G. H u , X. Shui, E Leng, W. Priebe, J. Chaires, and L. Williams, Biochemistry 36, 5940 (1997). 28 E Leng, W. Priebe, and J. Chaires, Biochemistry 37, 1743 (1998). 29 g. Guelev, S. Sorey, J. Lee, J. Ward, D. Hoffman, and B. Iverson, in preparation (2001).

[28]

POLYINTERCALATING

LIGANDS

569

Z ¢g ,.A

E b

E z

© [...

e..,

,-r, [... ©

.7 © [-.

< =~
Y - e x o n u c l e a s e activity cannot p r o c e e d i f primers contain one or m o r e 3' m i s m a t c h e s against their template-binding sites. 10 y. Choo and A. Klug, Proc. Natl. Acad. Sci. U.S.A. 91, 11163 (1994). 11 y. Choo and A. Klug, Curr. Opin. Struct. Biol. 7, 117 (1997).

recombination breakpoint in the two libraries. Recombination is carded out by digestion and religation of both Libl2 and Lib23. The composition of each helical position is indicated in the column of amino acids. Each library contains an invariant wild-type Zif268 region that acts as an anchor to the system, binding a fixed, G-rich, DNA sequence. 12 After selection, the two complementary variable regions of Libl2 and Lib 23 are recombined to make a full-length protein that potentially binds 10 bp of DNA: bases 2x-n'x. of a DdeI restriction enzyme site (CTGAG) that encodes Leu-Ser in positions 4 and 5 of F2 ~

598

ENZYMOLOGY AND BIOLOGICAL APPROACHES

A

~

c~-helix -1 1 2 3 4

i

I

Minicassettel

B

~

x

lrl

[30]

linker

4 5 6 I

Minicassette2

x

x

I

Minicassette3

Minicasseffe 1 5'-GCGGAAGAGAGGCCCTACGCATGCCCTGTCGAGTCCTGCGATCGC CGCCTTCTCTCCGGGATGCGTACGGGACAGCTCAGGACGCTAGCGGCGA-5

MinicasseRe 2 F1 - 1 +1 +2 +3 CGCTTTTCTxxxTCGxxxxx AAAGAxxxAGCxxxxxGGAAT

-TOP -BOTTOM Code for:

TOPCGCTTTTCTcrsTCGgmcca CGCTTTTCTcrsTCGgmcav CGCTTTTCTcrsTCGgmcgh CGCTTTTCTcrsTCGmrsca CGCTTTTCTcrsTCGmrsav CGCTTTTCTcrsTCGmrsgh CGCTTTTCTrmcTCGgmcca CGCTTTTCTrmcTCGgmcav CGCTTTTCTrmcTCGgmcgh CGCTTTTCTrmcTCGmrsca CGCTTTTCTrmcTCGmrsav CGCTTTTCTrmcTCGmrsgh

BOTTOMTAAGGtggkcCGAsygAGAAA TAAGGbtgkcCGAsygAGAAA TAAGGdcgkcCGAsygAGAAA TAAGGtgsykCGAsygAGAAA TAAGGbtsykCGAsygAGAAA TAAGGdcsykCGAsygAGAAA TAAGGtggkcCGAgkyAGAA TAAGGbtgkcCGAgkyAGAAA TAAGGdcgkcCGAgkyAGAAA TAAGGtgsykCGAgkyAGAAA TAAGGbtsykCGAgkyAGAAA TAAGGdcsykCGAgkyAGAAA

-i R,Q,H R,Q,H R,Q,H R,Q,H R,Q,H R,Q,H N,D,A,T N,D,A,T N,D,A,T N,D,A,T N,D,A,T N,D,A,T

+2 D,A D,A D,A RQHKSN RQHKSN RQHKSN D,A D,A D,A RQHKSN RQHKSN RQHKSN

+3 H N,S,T V,A,D H N,S,T V,A,D H N,S,T V,A,D H N,S,T V,A,D

Minicassette 3 F1 +4 +5 +6 5'- CCTTAxxxxxCATATCCGCATCCACA xXXXxGTATAGGCGTAGGTGTGGCC TOPCCTTayccrgCATATCCGCATCCACA CCTTaycghaCATATCCGCATCCACA CCTTaycamsCATATCCGCATCCACA

-TOP -BOTTOM

Code for:

BOTTOMCCGGTGTGGATGCGGATATGcyggr CCGGTGTGGATGCGGATATGtdcgr CCGGTGTGGATGCGGATATGsktgr

+5 I,T I,T I,T

+6 R,Q V,A,E K,N,T

Key to Randomised Nucleotides m=A/C r=A/G w=A/T

s=G/C y=T/C k=T/G

b=C/G/T

v=A/C/G

n=A/C/G/T

h=A/C/T d=A/G/T

FIG. 2. Construction of a gene cassette encoding a zinc finger phage display library with "smart" randomizations. (A) The scheme used to generate selective randomization throughout the u helix of a zinc finger. A set of complementary oligonucleotides is used to construct a series of"minicassettes" that

[30]

SEQUENCE-SPECIFIC ZINC FINGERS

599

The two complementary libraries are therefore designed with unique sequences at their 5' and 3' termini, and the corresponding primers are used to amplify any recombinants of the two libraries. Finally, it should be noted that the selection procedure is amenable to a microtiter plate format, so that selections and most subsequent manipulations can be carried out by liquid-handling robots. This allows substantial automation of the zinc finger engineering procedure. The ability to carry out selective PCR implies that the protocol can even be adapted to selecting complementary library portions in the same tube or well. For example, both universal libraries can be coscreened in a single well, thereby increasing the efficiency of high-throughput applications. The output of such combined selections can be monitored either by selective PCR, or by ELISA of samples of isolated clones, as described below. Library Construction Gene inserts for phage libraries may be constructed by end-to-end ligation of selectively randomized dsDNA "minicassettes," made individually by annealing complementary template oligonucleotides (Fig. 2). The resulting genes are amplified by PCR and encode zinc fingers in a suitable reading frame for cloning and expressing as fusions to the phage minor coat protein, pIII. Our work uses the DNA-binding domain of the transcription factor Zif268 as a scaffold, because it contains three Cysz-His2 zinc fingers whose mode of binding is well understood.12,13 To selectively randomize the ot helix of a zinc finger, the coding region is synthesized using DNA minicassettes, such that helical positions - 1 through 4 are encoded by one cassette (minicassette 2; Fig. 2), whereas positions 4 through 6 are encoded by another cassette (minicassette 3; Fig. 2). These double-stranded cassettes are synthesized with complementary overhangs that anneal through the codon for the fourth or-helical residue, which is invariant. Each cassette actually comprises a library of oligonucleotides synthesized with appropriate codon randomizations so as to encode a given subset of amino acids. Figure 2A shows that a "smart library" of zinc finger genes can be created with three minicassettes: the first cassette is a single sequence and encodes the invariant/3-sheet region, 12N. R Pavletich and C. O. Pabo, Science 252, 809 (1991). 13M. Elrod-Erickson, M. A. Rould, L. Nekludova,and C. O. Pabo, Structure 4, 1171 (1996).

can be annealed and ligated togetherto constructthe randomizedportionof the gene. Note that several additional minicassettes, encoding other fingers, need to be constructed to achieve the randomization scheme outlined in Fig. 1. After ligation of all the minicassettes, the full-lengthconstructis recovered by PCR, using primers that containSfiI/Notl restriction enzymesites for cloninginto a phage vector.(B) Examples of the oligonucleotidesused to achieve the selective randomizationof a single zinc finger.

600

ENZYMOLOGYAND BIOLOGICALAPPROACHES

[30]

whereas the second and third cassettes contain randomizations of the ct helix. Figure 2B shows that each of the library minicassettes comprises numerous oligonucleotides created through a limited number of solid-phase syntheses: minicassette 2 requires oligonucleotides from 12 pairs of syntheses, whereas minicassette 3 requires oligonucleotides from 3 pairs of syntheses. Each oligonucleotide synthesis is designed to introduce limited variability into each cassette; the library complexity is increased by the use of oligonucleotides from multiple syntheses and by the combination of the two minicassettes. Library Construction Protocol. Single-stranded template oligonucleotides are phosphorylated in a kinase reaction prior to assembly (100 pmol of each oligonucleotide in 10/zl of I x T4 kinase buffer, containing 1 mM dATP and 10 U of T4 polynucleotide kinase, 37 °, 1 hr). Complementary single-stranded template oligonucleotides are annealed pairwise to form double-stranded minicassettes (Fig. 2): 100 pmol of each oligonucleotide (or, for smart randomization, 100 pmol of each strand mixture) is mixed in 1 x T4 ligase or kinase buffer, to a final DNA concentration of 10 pmol//zl. Annealing is achieved by heating to 94 ° and then cooling slowly (~ 1 hr) to room temperature. The resulting double-stranded DNA (dsDNA) minicassettes are combined and ligated by adding an equal volume of l x T4 ligase buffer and 8 / t l (3200 U) o f T 4 ligase per 100/zl (16°, 20 hr). Full-length genes are amplified by PCR from the ligation mixture with primers that introduce NotI and Sill restriction sites for cloning into phage vector Fd-TETSN.l° Thorough digestion with these endonucleases is essential for high-efficiency ligation into similarly prepared phage vector (200 U of enzyme per 40/zg of DNA, with 8 hr of incubation in appropriate temperatures and buffers, adding enzymes in stages at 2-hr intervals). Typically, 1/zg of pure phage vector is ligated with a 5-fold excess of gene cassette insert (1 x T4 ligase buffer, 3/zl of T4 ligase, 30-~1 total volume, 16°, 20 hr). Ligation reactions are prepared for electroporation by washing twice with an equal volume of chloroform and precipitating by adding a 1/10 volume of sodium acetate (pH 5.5) and 3 volumes of ethanol. 14 DNA pellets are washed with 70% (v/v) ethanol and resuspended in sterile water to a final concentration of 200 ng//zl. The phage library is cloned by electroporation of recombinant vector into a suitable strain of Escherichia coli, such as TG1. Typically, 0.5/zg of recombinant phage vector can be used with 100/zl of electrocompetent cells, 15 yielding up to --d06 library transformants (2-ram path cuvette, 2.5 kV, 25/zF, 200 f2). After pulsing, cells are immediately resuspended in 1 ml of SOC medium (see Appendix) and incubated without shaking (37 °, 1 hr). Fd-TET-SN confers tetracycline resistance, allowing positive selection of bacterial transformants by plating on 2x YT-agar plates, containing tetracycline (15/zg/ml) (37 °, 16 hr). 14j. Sambrook,E. F. Fritsch,and T. Maniatis, in "MolecularCloning: A LaboratoryManual."Cold Spring HarborLaboratoryPress,Cold SpringHarbor,New York, 1989. 15W. J. Dower,J. E Miller,and C. W. Ragsdale,NucleicAcids Res. 16, 6127 0988).

I30]

SEQUENCE-SPECIFIC ZINCFINGERS

601

Selection from Libraries Selection Protocol. Phage are prepared for selections by scraping library transformant colonies into 2× YT liquid medium, containing tetracycline (15 #g/ml) and 50/xM zinc chloride, and incubating in an orbital shaker (30 °, 220 rpm, 16 hr). The culture supernatant containing phage particles is collected by centrifugation (3700g, 15 min). For each selection, the supernatant is diluted 1 : 10 in 1 ml of phosphate-buffered saline (PBS) containing 1% (w/v) Marvel, 1% (v/v) Tween 20, and sonicated salmon sperm DNA (20 #g/ml). Although these conditions generally work well, using specific competitor DNAs and varying binding conditions can have a significant impact on the outcome of individual selections. The phage mixtures are added to streptavidin-coated tubes or wells (Roche, Rahway, NJ) that have been precoated with biotinylated target DNA (made by annealing two complementary oligonucleotides, one of which is biotinylated). The selection procedure described here is for use with streptavidin-coated tubes and a total reaction volume of 1 ml (however, by scaling down to a 200-/zl volume the process can easily be adapted to a 96-well microtiter plate format). Typically, 1 pmol of target DNA is coated on each tube, in 50 #1 of PBS-Zn (--~20°, 15 min). The addition of 1 ml of 4% (w/v) Marvel blocking agent helps to reduce nonspecific binding. After blocking (-,~20°, 1 hr) tubes are emptied, refilled with 1 ml of phage binding mixture, and left to equilibrate (~20 °, 1 hr). Washing steps are then carried out to remove all unbound phage [20 washes with 1 ml of PBS-Zn containing 2% (w/v) Marvel, 1% (v/v) Tween 20, followed by one wash with PBS-Zn alone]. Retained phage are eluted in 100 #1 of 0.1 M triethylamine, removed to a separate container, and immediately neutralized with an equal volume of 1 M Tris-HCl, pH 7.4. Eluted phage can be stored at - 2 0 °. Fifty microliters of the eluted phage are used to infect 0.3 ml of a logarithmic phase culture ofE. coli TG1. The bacteria are derived from colonies grown freshly on M9 minimal agar as this ensures expression of the F' pilus, which is required for phage infection. Bacteria are infected by the addition of phage and incubating without shaking (37 °, 1 hr). Bacteria are then transferred to 2-5 ml of 2× YT containing tetracycline (15 #g/ml) and 50 #M zinc chloride and grown, as before, to prepare phage supematant (30 °, 220 rpm, 16 hr). Subsequent rounds of selection are carried out as described above, although the amount of competitor DNA may be increased in later rounds to increase the stringency of selection. Three to five rounds of selection are usually sufficient to enrich target-binding clones; however, the progress of individual selections may be monitored by plating out infections to estimate phage yield after each round of selection (Fig. 3; see Related Procedures in the Appendix). It should be noted, however, that phage yield does not by itself give an indication of the quality of the selected zinc fingers because fast-growing nonspecific binders can dominate selections. After selections, pools of complementary phage can be recombined directly without preliminary analysis; however, we recommend that individual clones first

602


I30]

3500 3000 -

=zl. 2500

~

Libl211

2000

Lib23 [ ]

1500 1000 500 0 RI

R2

R3

R4

R5

Round FIG. 3. Phage titers of selections from Libl2 and Lib23, using two DNA target sites, 5'-GCG GCT CTC-31 and 5'-GGA GCG GCG-3r (selected sequences shown in boldface). After each round of selection (R1-5), neutralized phage eluate was used to infect an excess of logarithmic phase E. coli TG1 cells. Serial dilutions of these infections were plated out on TYE-TETplates. Colonies that grew after incubation (15 hr, 37°) were counted to estimate phage yield. Phage yields are shown plotted vertically as number of colony-formingunits per microliter of selection eluate. undergo testing for binding to their respective 5-bp target sites, using an ELISA (see below).

Recombination of Complementary Libraries After three to five rounds of selection, PCR is used to amplify zinc finger genes from Libl2 and Lib23, and the selected portions are recombined by digestion and religation at a unique DdeI restriction site. The DNA template for the initial PCR can be derived from the entire pool of selected phage, or preferably from one or several individual clones previously validated for DNA binding. Individual clones are chosen on the basis of sequence-specific DNA binding, verified by phage ELISA (see below). Recombination begins with a PCR step to amplify all the zinc finger genes present, using primers that are external to the SfiI/NotI gene cassette. One microliter of phage supernatant(s) provides sufficient DNA template on lysing the virions (94 °, 9 min). The PCR itself comprises 30 cycles, using Taq polymerase. After PCR, 50/zl of the reaction is phenol/chloroform extracted, diluted in 2 volumes of DdeI buffer (New England BioLabs, Beverlay, MA), and digested with 60 U of DdeI per 150 #l. The enzyme is heat inactivated (65 °, 20 min) and the reactions are precipitated with a 1/10 volume of sodium acetate (pH 5.5) and 3 volumes of ethanol. Each DNA pellet is resuspended in 10 ttl of T4 DNA ligase buffer (New England BioLabs) and 400 U o f T 4 ligase is added per reaction (20 °, 15 hr). Recombinants, comprising the selected portions of Libl2 and Lib23, are amplified selectively by PCR from 1/zl of the ligation mixture (Fig. 4). For optimal selective amplification, only 20 cycles of PCR are carried out with

[30]

SEQUENCE-SPECIFIC ZINC FINGERS ,,~

W

Ub12

i

I

"

\

603

Ub23

I

W

I

Dde I

Dde I

x.

M

I PCR, Dde I cut, ligation W

I

W

M

W

I

I

M ,~

I W

\

M

SelectivePeR (M/M primers)

Recombinedproduct FIG.4. SelectivePCR to recovercorrectlyrecombinedportionsof the two complementarylibraries. PCR primers that bind the 3t and 5~ends of zinc finger genes are designedto bind either the wildtype Zif268 codingsequence (W) or a mutantDNA sequencecontainingalternativecodons (M). The mutations affect priming events in the PCR but do not alter the protein sequence when translated. The two selected portions of the bipartite libraries(shown in black and gray) containmutant primer recognitionsites.Therefore,two mutant(M/M)primersare requiredto recoverthe correctrecombinants from the mixedpool of ligationproducts. a nonproofreading DNA polymerase such as Taq. It is worth noting that enzymes such as Pfu, which contain 5'--->3' proofreading activity, erode the DNA ends, thus eliminating selective priming. PCR products are cloned back into phage vector as described above (see Library Construction, above). The resulting products of the recombination step, which often comprise pools of clones (defined by the various permutations in which portions from the two libraries can recombine), may be reselected by phage display so that the optimal zinc finger phage is isolated. Alternatively, a number of different clones may be screened for binding to the target DNA sequence.

Coselection of Two Complementary Libraries In devising the present strategy we initially envisaged the simultaneous selection of complementary zinc finger domains from a mixture of L i b l 2 and Lib23

604


[30]

in a single tube. This should be possible as different DNA sequences are used to select zinc fingers from the two libraries, meaning that these would not compete with each other for the DNA used in their selection. We further imagined that the mixture of selected zinc fingers could then be recombined directly to yield product zinc fingers (without previously carrying out ELISAs to verify the presence of binders from each library). Although this approach is feasible in theory, and indeed in practice we have used it to engineer a few DNA-binding domains, we have found that physically separating the selections from the two libraries increases the probability that the engineering will be successful. It often emerges that after coselection of the two libraries, the large majority of selected phage are from just one of the two libraries (this can be checked either by phage sequencing or by selective PCR) and therefore recombinant DNA-binding domains cannot be produced. The reason for this is that certain phage clones have a distinct growth advantage that contributes to the selection process (alongside DNA binding), allowing a few clones to dominate the selection. Thus, separating the selections from the two libraries increases the chance of obtaining phage from both libraries.

Phage Enzyme-Linked Irnmunosorbent Assay Determination of D N A - B i n d i n g S p e c i f i c i t y a n d Affinity Phage ELISA may be used to assess the DNA-binding affinity and specificity of a zinc finger phage clone at any stage of the engineering procedure, but particularly prior to recombination of clones from the two libraries, and at the end point, at which time it is important to validate DNA recognition by the product. By using closely related variants of the DNA binding site, or a variety of binding-site concentrations, it is possible to obtain either a profile of the sequence discrimination and/or the apparent Kd values of interaction.

Enzyme-Linked Immunosorbent Assay Protocol Single bacterial colonies, infected with phage clones derived from library selections, are picked from agar plates. Colonies are transferred to wells in sterile, round-bottom, 96-well plates containing 150/zl of 2 x YT, tetracycline (15/zg/ml), and 50/zM zinc chloride. Plates are incubated with orbital mixing (30 °, 225 rpm, 16 hr). Phage supematant is prepared from the bacterial cultures by centrifuging the 96-well plates (3700g, 15 min), in a suitable swinging-bucket centrifuge. A phage binding mixture is prepared by diluting supernatant 1 : 10 in 1 ml of PBSZn containing 2% (w/v) Marvel, 1% (v/v) Tween 20, and sonicated salmon sperm DNA (20/zg/ml). DNA target sites for ELISA can be made by synthesizing and annealing two complementary oligonucleotides, one of which is biotinylated. These targets are

[30]


605

bound to streptavidin-coated wells in 96-well microtiter plates (Roche). Typically, 0-5 pmol of DNA site is bound in 50/zl of PBS-Zn ('--20°, 15 min). Wells are then blocked (20 °, I hr), using 150 #1 of PBS-Zn containing 4% (w/v) Marvel. Blocking solutions are discarded and 50 #1 of the phage binding mixture is applied to each well. Binding mixtures are left to equilibrate (~20 °, 1 hr) and are then discarded. Wells are washed seven times with 200/~1 of PBS-Zn containing 1% (v/v) Tween 20. A further three washes are carried out with 200 #1 of PBS-Zn alone. Fifty microliters of PBS containing 2% (w/v) Marvel and a 1 : 5000 dilution of horseradish peroxidase (HRP)-conjugated anti-M 13 IgG antibody (Pharmacia Biotech, Piscataway, N J) is added to each well. After incubating (~20 °, 1 hr), antibody binding reactions are discarded and the wells are washed three times with 200 #1 of PBSZn containing 0.05% (v/v) Tween 20. A further three washes are carried out with 200/zl of PBS-Zn alone. ELISAs are developed with 100 #1 of HRP substrate, such as the 3',3',5',5'-tetramethylbenzidine (TMB)-based, ELISA developer solution described below. Colorimetric reactions are stopped after ~5 min by adding 100 #1 of 1 M H2SO4. ELISA signals are quantitated with a spectrophotometer fitted with a microtiter plate reader. Signals should be compared with negative control wells that either lack target DNA or have been coated with unrelated DNA sites.

Scanning Mutagenesis Binding Assays In this phage ELISA variation, each position in the ~10-bp binding site is sequentially mutated in a series of oligonucleotide targets in order to help assess the DNA sequence specificity of the engineered zinc finger protein. The single transition mutations used in the control DNA sites are more stringent controls of DNA-binding specificity than their counterpart transversion mutations. This is primarily because of the size and charge similarities of purines and pyrimidines (the purines A and G are both larger and partially negatively charged, whereas the pyrimidines C and T are smaller and partially positively charged). Transition mutations therefore represent a good test of the ability of zinc fingers to differentiate between closely related DNA sites. Typically, binding is carried out at a DNA concentration of 5-10 riM, which is close to the expected apparent Kd of zinc fingers. In the example shown in Fig. 5, a clone engineered from the bipartite-complementary library against the target site GGA-GCT-CTC is tested against a series of transition mutant sites. Although the construct binds well to its own target site, it appears that discrimination is not flawless throughout the binding site. In particular, the assay reveals that a C-+T transition (Fig. 5; column 8) is tolerated by the protein. Otherwise, however, the overall binding specificity is shown to be fairly good. This type of assay provides an indication of the DNA-binding specificity for a given protein but obviously does not allow a comprehensive study. To achieve

606


[30]

Selection target GGA F3

GCT F2

CTC F1

I Phage ELISA Well

1

2

3

4

5

6

7

8

9

10

G G A G C T C T C

a

G A G C T C T C

G a A G C T C T C

G G g G C T C T C

G G A a C T C T C

G G A G t T C T C

G G A G C e C T C

G G A G C T t T C

G G A G C T C e C

G G A G C T C T t

ii@@@@@@ii@@

I Plot ELISA Well 0.6

~

1

2

3

4

5

6

7

8

9

10

T

G

G

A

G

C

T

C

T

C

E

liiliilll

0.3

0.0

T

a

a

g

a

t

c

t

c

t

FIG. 5. Scanning mutagenesis ELISA binding assay. A zinc finger phage clone was tested for binding both to its full target site (GGA-GCT-CTC) and to sites containing single transition mutations. Each well of tfie ELISA microliter plate is coated with a DNA sequence that contains a different transition mutation (shown as a lower-case nucleotide). DNA binding is revealed by the ELISA absorbance reading (A45o-A650),which is plotted vertically in the graph. All binding reactions were carried out at a DNA concentration of 8 nM.

[301


607

this we have developed microarrays of double-stranded DNAs printed on glass slides. 16 These microarrays can be used to perform phage ELISA in a procedure similar to that described above, but the high density of DNA spotting allows parallel sampling of many thousands of DNA sequences. A u t o m a t i o n of Zinc F i n g e r E n g i n e e r i n g P r o c e s s Most of the steps of the engineering process--bacterial growth, phage selection, colony picking, phage ELISA, PCR, and cloning--can be automated by using commercially available instruments. At Gendaq (London, UK) we use 96or 384-well microtiter plates to carry out phage selections, ELISA reactions, and PCR preparation, on a liquid-handling robotic platform. A robotic arm shuttles the microtiter plates between a pipetting station, a plate hotel, a plate washer, a spectrophotometer, and a PCR block. A colony-picking robot is used to inoculate microcultures of bacteria in microtiter plates in order to provide monoclonal phage for ELISA. A robot that interfaces with the spectrophotometer is capable of returning to the liquid culture archive in order to "cherry-pick" particular clones that are suitable for recombination, or that should be archived. A bar-coding system is used to keep track of the various plates used for phage selections, phage ELISAs, or for archiving interesting clones. Concluding Remarks Previous strategies for engineering zinc fingers by phage display were labor intensive and/or limited in their utility.l If zinc fingers were selected in parallel and subsequently assembled into a multifinger domain, the DNA specificity of the domain was limited to binding only a guanine-rich binding site. 6 If, on the other hand, the selection of zinc fingers was carried out serially, the protein engineering procedure became laborious. 7 The "bipartite library" protocol described here solves the problem of engineering sequence-specific zinc fingers to bind diverse DNA sequences, and at the same time increases the speed and throughput of the engineering process. Appendix: Solutions and Related Procedures

Solutions Enzymes and buffers: T4 DNA kinase, T4 DNA ligase, Taq DNA polymerase, and SfiI, Nod, and DdeI endonucleases are all used under conditions specified by manufacturers (e.g., New England BioLabs, Promega) 16M. Bulyk,G. Church, and Y. Choo,in preparation(2001).

608


[30]

Escherichia coli TG117: supE hsdA thiA(lac-proAB) F' [traD36 proAB + laqI q laqZAM 15] 2 x YT: Bacto-tryptone, 16.0 g/liter; Bacto-yeast extract, 10.0 g/liter; NaCI, 5.0 g/liter. For growth of zinc finger phage add tetracycline (15/zg/ml) and 50/zM zinc chloride 2x YT-agar: Bacto-tryptone, 16.0 g/liter; Bacto-yeast extract, 10.0 g/liter; NaC1, 5.0 g/liter; agar, 15.0 g/liter SOC medium: Bacto-tryptone, 20.0 g/liter; Bacto-yeast extract, 5.0 g/liter; NaC1, 10 raM; KCI, 2.5 mM; MgC12, 10 raM; MgSO4, 10 mM; glucose, 20 mM Minimal agar medium: Na2HPO4, 6.0 g/liter; KH2PO4, 3.0 g/liter; NH4C1, 1.0 g/liter; NaC1, 0.5 g/liter. After autoclaving, add 1 M MgSO4, 1.0 ml/liter; glucose, 2.0 g/liter; 1 M CaCI2, 0.1 ml/liter; 1 M thiamine hydrochloride 1.0 ml/liter; agar, 15.0 g/liter PBS-Zn: PBS (10x stock solution: NaC1, 80 g/liter; KC1, 2 g/liter; Na2HPO4. 7H20, 11.5 g/liter; KHzPO4, 2 g/liter); make up to 1 x and add 50 #M zinc chloride immediately before use TMB-based ELISA developer solution: 0.1 M sodium acetate pH 5.5; 3',3', 5',5'-tetramethylbenzidine (TMB; Sigma, St. Louis, MO ), 0.5 mg/ml; dimethyl sulfoxide (DMSO), 1% (v/v); H202, 0.05% (v/v) Related Procedures Cesium Chloride Gradient Purification of Replicative Form Phage Vector. If Fd-TET-SN is required to clone large phage display libraries, vector DNA is prepared from a l-liter bacterial culture by using a large-scale plasmid preparation kit [e.g., Qiagen (Valencia, CA) Maxiprep], followed by additional purification on a cesium chloride gradient. 14Only double-stranded phage DNA purified on a cesium chloride gradient is suitable for library construction; the library size that may be cloned is severely reduced by lower quality vector. Ten to 100 #g of phage DNA is dissolved in 4 ml of Tris-EDTA (TE) buffer containing 4 g of cesium chloride and 0.1 ml of ethidium bromide solution (10 mg/ml). Samples are well mixed to dissolve and dispensed into 5-ml ultracentrifuge tubes (Beckman, Fullerton, CA). Tubes are centrifuged with a Beckman Vti 65.2 rotor (20°, 370,000 g, 20 hr). Two bands of DNA are visible under UV light. The lower (supercoiled plasmid) band is collected with a syringe and washed four times with water-saturated butanol to remove all traces of ethidium. The resulting solution is diluted 3-fold in TE buffer and then ethanol precipitated to recover pure phage vector. Preparation of Electrocompetent Cells. Various precautions must be taken to achieve the highest level of electrocompetence.15 Escherichia coli colonies are initially selected from freshly inoculated minimal medium plates, which promote 17T. J. Gibson,CambridgeUniversity,Cambridge,UK, 1984.

[30]


609

the growth of bacteria containing the F' plasmid. Cells are then transferred to the rich 2x YT medium to encourage rapid growth and division. Also, all inoculations are carried out in prewarmed medium to prevent the interruption of cell growth. At the cell-harvesting stage, all centrifugations are carried out with precooled (4°), sterile centrifuge pots and buffers to minimize cell damage. To further maintain cell viability, centrifugations are performed at the minimum speed and for the minimum time required to pellet cells, so that resuspensions may be carried out gently. Escherichia coli TG1 cells are streaked out onto minimal medium plates and incubated (37 °, ~15-24 hr). A single colony is transferred into 5 ml of 2x YT medium and incubated until turbid (37 °, ~3 hr). The 5-ml culture is then added to 250 ml of 2x YT medium and incubated until the culture reaches an optical density of 0.6 at 600 nm (37 °, -,~3-4 hr), equivalent to approximately 108 cells/ml. At this point the culture is transferred to 4 × 50 ml sterile tubes and harvested by centrifugation (4 °, 3500 g, 5 min). The cell pellets are resuspended in 4 x 50 ml of cold, sterile Milli-Q water, and then centrifuged as before. The pellets are resuspended in 2 x 50 ml of cold, sterile Milli-Q water, and centrifuged again. These pellets are resuspended in 20 ml of 10% (v/v) glycerol in cold, sterile Milli-Q water and centrifuged once more. The final pellets are resuspended in 0.7 ml of 10% (v/v) glycerol in cold, sterile Milli-Q water. At this stage, the cell suspension is subdivided into 50- to 70-#1 aliquots, and either used directly for electroporation or frozen on dry ice and stored at -80". Cells prepared by this method yield approximately 109 transformants per microgram of pUC-19 supercoiled plasmid DNA, as measured by a standard electroporation assay using 15 pg of DNA. For best results, transformations are carried out by electroporation of freshly prepared (not previously frozen) electrocompetent cells. Estimating Phage Yield. Phage titer from selection eluates and culture supernatants can be estimated by using 1 #1 of phage sample to infect 1 ml of a logarithmic phase culture of E. coli (37 °, 1 hr, no shaking). Infections are serially diluted 10-fold with individual dilutions being spread on 2× YT-agar plates containing tetracycline (15/~g/ml). After incubation (37 °, 16 hr), individual colonies are counted to give an indication of the colony-forming units (phage titer) in the original sample (see Fig. 3). Acknowledgments Y. C. and M. 1. were staff scientist anti graduate student, respectively, at the Medical Research Council Laboratory of Molecular Biology (Hills Road, Cambridge CB2 2QH, UK). Both authors currently work at Gendaq, Ltd. (1-3 Burtonhole Lane, London NW7 lAD, UK).

610


[31]

[31] DNA Relaxation and Cleavage Assays to Study Topoisomerase I Inhibitors By C H R I S T I A N

BAILLY

Introduction The history of the topoisomerase I enzyme is intimately interwoven with that of the plant alkaloid camptothecin (CPT).I This antitumor drug contributed importantly to the elucidation of the DNA-cleaving properties of the enzyme and much of its biochemistry. 2-4 Topoisomerase I is a monomeric t r a n s - e s t e r a s e . The enzyme catalyzes a reversible DNA cleavage reaction by transferring a phosphodiester bond to the active site tyrosine residue. 5'6 This reaction, which does not require an energy cofactor, produces a covalent intermediate in which the enzyme is covalently attached to the 3' terminus of a DNA single-strand break (Fig. 1). This transient enzyme-DNA covalent complex (often called the cleavage complex) is the key element of the reaction. The identification of specific inhibitors is essentially based on their capacity to stabilize this covalent complex. In the absence of inhibitor, the covalent complex is highly labile and freely dissociates to regenerate the active enzyme and the religated DNA molecule with intact double strands. This catalytic cycle is essential for relieving DNA torsional stress during all manipulations of DNA in cells, including replication and transcription. 7 But in the presence of a poison, such as camptothecin, the religation step of the topoisomerase I catalytic reaction can be inhibited, thus producing single-stranded DNA breakages that are subsequently transformed into double-strand breaks lethal for the growing cells. Camptothecin converts topoisomerase I into a cell poison by blocking the religation step, thereby enhancing the formation of persistent DNA breaks responsible for cell death. This is also the case for various synthetic analogs endowed with potent antitumor activities such as the clinically used drugs topotecan and SN38 (the active metabotite of irinotecan) as well as derivatives 9-aminocamptothecin, 9-nitrocamptothecin, DX-8951 f, GG-211, CKD602, and the homocamptothecin derivative BN 80915. But in addition to the camptothecins, an important 1 M. E. Wall and M. C. Wani, CancerRes. 55, 753 (1995). 2 M. Gupta, A. Fujimori, and Y. Pommier, Biochim. Biophys. Acta 1262, I (1995). 3 y. Pommier, E Pourquier, Y. Fan, and D. Strumberg, Biochim. Biophys. Acta 1400, 83 (1998). 4 j. j. Champoux, Plvg. Nucleic Acid Res. Mol. Biol. 60, 111 (1998). 5 M. R. Redinbo, L. Stewart, P. Kuhn, J. J. Champoux, and W. G. J. Hol, Science 279, 1504 (1998). 6 L. Stewart, M. R. Redinbo, X. Qiu, W. G. Hol, and J. J. Champoux, Science 279, 1534 (1998). 7 j. C. Wang, Annu. Rev. Biochem. 65, 635 (1996).

METHODS IN ENZYMOLOOY.VOL 34(1

Copyright~¢;~200J by AcademicPress All righlsof reproductionit] any tbml reserved. 0076-6879/00$35.00

[3 1]

TOPOISOMERASE I INHIBITORS

dissociation

Topo I

binding and recognition

DNA cleavage

61 1

) covalent

re~a~n

complex

FIG. 1. Schematicrepresentation of the topoisomerasecatalyticcycle. The topoisomerizationreaction involves binding of topoisomerase I to DNA, followedby DNA cutting and then religation. group of topoisomerase I inhibitors with disparate structures and origins has been identified. 8- l0 Benzophenanthridine alkaloids, certain bis- and terbenzimidazole derivatives, and several glycosylated indolocarbazole derivatives hold great promise as topoisomerase I-targeted anticancer agents. ~ Several experimental approaches have been employed to identify these inhibitors and to characterize their mechanism of action. In most cases, in vitro assays using purified topoisomerase I and a radioactively labeled DNA substrate have been employed to evidence the poisoning effects of these drugs. This chapter focuses on two frequently used experimental approaches to characterize topoisomerase I inhibitors. These procedures are illustrated mainly with examples from our own work on topoisomerase I-targeted drugs. Further detailed discussion of the effect of topoisomerase I inhibitors at the cellular level is beyond the scope of this chapter. The experiments presented here can be performed with commercial topoisomerase I from calf thymus (Life Technologies, Rockville, MD) or the human enzyme (TopoGen, Columbus, OH). Methods to overexpress and purify topoisomerase I from different organisms, including bacteria, yeast, human placenta, and baculoviruses have been described. 12

8 y. Pommier, Biochimie 80, 255 (1998). 9 y. Pommier, E Pourquier, Y. Urasaki, J. Wu, and G. S. Laco, Drug Resistance Updates 2, 307 (1999). 10C. Bailly, Curt Med. Chem. 7, 39 (2000). 11B. H. Long and B. N. Balasubramanian, Exp. Opin. Ther. Patents 10, 635 (2000). 12M.-A. Bjornstiand N. Osheroff(eds.), DNAtopoisomeraseprotocols.I. DNAtopologyand enzymes. In "Methods in Molecular Biology,"Vol. 94, Humana Press, Totowa,New Jersey, 1999.

612


(-) ethidium

(+) ethidium

z o0-C0 al-~ I-'1F'IF"lF-I

topoisomer~¢

[3 1 ]

z

+

i

+'+' i

40- i ~

-40

411

+o-+

. = -30

t

FIG. 5. (A) Cleavage of a 116-bp DNA restriction fragment from plasmid pTayD by human topoisomerase I in the presence of graded concentrations of the homocamptothecin derivative BN80915 and the conventional camptothecin SN38. The 3' end-labeled fragment (lane DNA) was incubated in the absence (lane Topoi) or presence of the test drug at the indicated concentration (micromolar). Topoisomerase I cleavage reactions were analyzed on a 8% denaturing polyacrylamide gel. Numbers at the right-hand side of the gel show the nucleotide positions, determined with reference to the guanine tracks labeled G. The positions of I 0 major cleavage sites are indicated. (B) Cleavage of a 131-bp DNA restriction fragment from plasmid pTayB by human topoisomerase I in the presence of CPT or the two glycosylated indolocarbazole derivatives 1 and 3 (structures indicated in Fig. 3A).

active metabolite of the clinically used antitumor drug irinotecan. For this purpose, a 116-bp EcoRI-HindlII restriction fragment from plasmid pTayD, uniquely 3' end labeled at the EcoRI site, was used as a substrate for the topoisomerase I cleavage reaction. The cleavage products were analyzed on a sequencing polyacrylamide gel as shown in Fig. 5A. Both BN80915 and SN38 promote DNA cleavage by topoisomerase I, but the effect is more pronounced with the homocamptothecin

622


[3 1]

derivative compared with the conventional camptothecin. Moreover, the patterns of cleavage sites are different. For example, sites 2, 3, 4, 7, 8, and 10 are cleaved more efficiently with BN80915 than with SN38. The converse effect occurs at sites 1 and 5 whereas the cleavage intensities at sites 6 and 9 are comparable for the two drugs. The results confirm that the replacement of the camptothecin lactone E-ring with a homologous seven-membered lactone ring changes the sequencespecificity of drug-induced DNA cleavage by topoisomerase I. In particular, several sites encompassing the sequence 5'-AAC'I'G were found to be specifically cleaved in the presence of homocamptothecin (hCPT) but not with camptothecin. 51 Such cleavage studies with restriction fragments have provided essential information about the role of the o~/fl-hydroxylactone ring of CPT/hCPT in the poisoning of topoisomerase I. 2. Indolocarbazoles derived from the antibiotic rebeccamycin are promising antitumor agents targeting topoisomerase I. l°'j~ Unlike the camptothecins, these drugs bind to DNA even in the absence of topoisomerase I. The planar indolocarbazole chromophore of the tumor active drug NB-506 (compound 1 in Fig. 3A) intercalates between two consecutive base pairs in the DNA double helix, thus placing the appended glucose residue into one of the helical grooves, most likely the minor groove. 19 We used a series of restriction fragments to compare the patterns of topoisomerase I-mediated DNA cleavage obtained in the presence of 1 or its regioisomeric analog 3, which contains two hydroxyl groups at positions 2 and 10 instead of at positions 1 and 11. A typical sequencing gel obtained with a 131-bp EcoRI-HindIII restriction fragment from plasmid pTayB, uniquely 3' end labeled at the EcoRI site, is presented in Fig. 5B. The two drug isomers are equally potent at maintaining the integrity of the topoisomerase I-DNA covalent complexes but stimulate cleavage at different sites on DNA. Compound 1 stabilizes topoisomerase I preferentially at sites having a pyrimidine (T or C) and a G on the 5' and 3' sides of the cleaved bond, respectively, whereas compound 2 induces topoisomerase I-mediated cleavage principally at TG sites only. The two drugs preferentially induce cleavage by the enzyme at different nucleotide sequences. Therefore, we concluded that the positioning of the two hydroxyl groups on the drug chromophore exerts a significant effect on the sequence specificity of topoisomerase I-induced DNA cleavage. 23 Cleavage assays can also be performed with a linear plasmid DNA 3' end labeled, such as the EcoRI-linearized pBR322 DNA. In this case, the cleavage products of the DNA fragment are denatured with alkali (0.45 M NaOH) prior to electrophoresis on agarose gels containing a detergent [0.1% (w/v) SDS]. This procedure does not permit identification of the cleavage site at nucleotide resolution, 51C. Bailly, A. Lansiaux, L. Dassonneville,D. Demarquay,H. Coulomb,M. Huchet, O. Lavergne, and D. C. H. Bigg,Biochemistry 38, 15556 (1999).

[31 ]

TOPOISOMERASE I INHIBITORS

623

but it is relatively quick and the radioactive substrate is easy to prepare. This is a convenient assay for comparing the inhibitory potencies of series of drugs. 52 Conclusion The relaxation and cleavage assays presented here are the most frequently used in vitro methods to study the effects of topoisomerase I inhibitors. However, other

technical approaches can be used. For example, a filter binding assay can also be employed to evidence the stabilizing action of drugs on DNA-topoisomerase I covalent complex formation. 53 Neither topoisomerase I alone nor DNA is retained on nitrocellulose filters whereas the large covalent complexes bind selectively to nitrocellulose in the presence of SDS. The addition of a topoisomerase I poison such as CPT, or certain flavones, 14 increases significantly the amount of material retained on the filter, thus reflecting the drug-induced formation of additional covalent complexes. This assay may be used as a rapid screening procedure for screening libraries but it is not sensitive and requires relatively large amounts of enzyme. In this chapter, the two most widely used in vitro approaches to characterize topoisomerase I inhibitors have been briefly outlined. Not only do these assays help to identify novel topoisomerase I poisons; they can provide important structural data. They can easily be performed by any laboratory that is experienced in methods of molecular biology. We hope that this description will encourage researchers to apply these approaches to discover new anticancer or antiparasitic agents directed against topoisomerase I. Acknowledgment Support from the Association pour la Recherche sur le Cancer is acknowledged.

52 j. F. Riou, P. FossE, C. H. Nguyen, A. K. Larsen, M. C. Bissery, L. Grondard, J. M. Saucier, E. Bisagni, and E Lavelle, Cancer Res. 53, 5987 (1993). 53 S. M. Hecht, D. E. Berry, L. MacKenzie, R. W. Busby, and C. A. Nasuti, J. Nat. Prod. 55, 401 (1992).

624


[32]

[32] In Vitro Human Immunodeficiency Virus Type 1 Integrase Assays B y CHRISTOPHE M A R C H A N D , NOURI NEAMATI, a n d YVES POMMIER

Introduction Integration of the human immunodeficiency virus type 1 (HIV- 1) genome into host cell DNA is an essential step for viral replication. Soon after cell penetration, the single-stranded viral RNA is reverse transcribed into a double-stranded proviral DNA. The next major event is integration, whereby the viral DNA is inserted into the host chromosome to create the proviral state. This event is catalyzed by a viral enzyme called integrase, encoded in the po! gene and generated after proteolysis of the Gag-Pol fusion protein precursor by the HIV-1 protease (Pig. 1). Integrase is responsible for the removal in the cytoplasm of a GT dinucleotide immediately 3' from a conserved CA dinucleotide at the Y-end of both extremities of the viral genome [U3 and U5 long terminal repeats (LTRs)]. After this first step, called Yprocessing, integrase remains bound to the LTRs and this preintegration complex migrates to the nucleus, where the second step of the integration reaction (3'-end joining) occurs. The T-end joining reaction consists of the direct nucleophilic attack of the Y-recessed viral ends on the host chromosome. Both termini of the viral DNA, kept in close proximity, integrate with a 5-bp stagger toward the 51-ends of the target chromosomal DNA. Completion of the integration process requires removal of the two unpaired nucleotides at the 5'-ends of the viral DNA and gap filling, probably accomplished by cellular enzymes {for review see Refs. 1-3). HIV-1 integrase is a 32-kDa protein of 288 amino acids. It can be divided into three functional domains {Pig. 2). The central domain represents the enzyme catalytic core and is required for all enzymatic activities. The N- and C-terminal domains are only required for full enzymatic activity and oligomerization. The catalytic core of HIV-I integrase {residues 50-212) contains three acidic amino acids (Asp-64, Asp-116, and Glu-152), which are highly conserved among integrases from all retroviruses, retrotransposons, and some bacterial transposases. These conserved residues are absolutely required for all integrase activities and are referred to as the D,D(35)E motif. In HIV-l-infected patients, at present, only reverse transcriptase and protease inhibitors are used as a combination in highly active antiretroviral therapy I y. Pommier, A. A. Pilon, K. Bajaj, A. Mazumder, and N. Neamati, Antivira[ Chem. Chemother 8, 463 (1997). 2 p. O. Brown, in "Retroviruses" (J. M. Coffin, S. H. Hughes, and H. E. Varmus, eds.), pp. 161-203. Cold Spring Harbor Press, Cold Spring Harbor, New York, 1998. 3 A. M. Skalka, "Advances in Virus Research." Academic Press, San Diego, California, 1999.

METHODSIN ENZYMOLOGY.VOL.340

In Vitro H I V - 1 INTEGRASE ASSAYS

[32]

U3

R U5

s,I_~.nc'z"z'ccc~

625

U3 R Gag

Pol

Env

U5

3'

ATCGTCA 1

5'

3'-Processing I

C 3'-end joining

1

Hosl

5'-processing gap filling

FIG. 1. Schematic representation of the integration reaction steps. HIV-1 integrase catalyzes 3rprocessing in the cytoplasm and 3t-end joining in the nucleus. The conserved CA dinucleotide is outlined and the 3t-processed GT dinucleotide is underlined. The mechanisms tbr gap filling and 5'-end joining are unknown.

626

[32]


PR

RT

IN

s

s

s ~

s~ p s

~

s s

s~ I~

/" ts ~4~J I I

,

H 1 2 H 1 6 C 4 0 C43

zinc binding

DD(35)E

'

~................................................................... I

D64

D116

K...K...K 211

catalytic core

E152

.... R...R..R

.... K...K...K 244

DNA binding

FIG. 2. The three domains of HIV-1 integrase. The central core domain contains three conserved acidic residues [D,D(35)E] common to all retroviral integrases.

(HAART). Because HIV-1 integrase is absolutely required for viral replication and because it has no cellular equivalent, this enzyme is an excellent target for a chemotherapeutic approach. 4'5 In vitro integration assays are used to elucidate the enzyme mechanisms and to discover, design, and develop integrase inhibitors. Other assays using subcellular fractions from infected cells have been reported (preintegration complex assays, PICs). 6 They are not described in the present chapter. Human Immunodeficiency Virus Type 1 Integrase Preparation The HIV- 1 integrase protein commonly used is a soluble double mutant bearing the mutations F185K and C280S. This mutant is as catalytically active as the wildtype protein. Its higher solubility provides better purification yields. 7 Transformation

The HIV-1 IN/F185K/C280S gene has been cloned into a pET-15b ampicillinresistant vector (Novagen, Milwaukee, WI). Twenty microliters ofEscherichia coli 4 y. Pommier and N. Neamati, Adv. Virus Res. 52, 427 (1999). 5 N. Neamati, C. Marchand, and Y. Pommier, Adv. Pharmacol. 49, 147 (2000). 6 M. S. T. Hansen, G. J. I. Smith, T. Kafri, V. Molteni, J. S. Siegel, and F. D. Bushman, Nat. BiotechnoL 17, 578 (1999).

[32]

In Vitro HIV-1 INTEGRASEASSAYS

627

BL21 (DE3) competent cells (Novagen) are placed into a 1.5 ml prechilled tube (Eppendorf Scientific, Westbury, NY) and kept on ice. Five nanograms of the pET15b/HIV- 1 IN/F 185K/C280S plasmid is then added to the cells and the suspension is incubated for 5 min on ice. The competent bacteria are heat shocked by placing the tube in a water bath at 42 ° for 30 sec and then back on ice for 2 min. Bacteria are diluted by adding 250 #1 of SOC medium (Novagen) and placed for 1 hr at 37 ° in a rotary shaker. The entire tube volume is spread on an agar plate containing ampicillin (100 #g/ml; K. D Medical, Columbia, MD) and grown overnight at 37 °. At least three colonies are picked up and amplified for 12 hr in separate tubes containing 6 ml of Luria-Bertani (LB) broth medium (Digene, Beltsville, MD) and ampicillin (100 #g/ml; Boehringer Mannheim, Indianapolis, IN). A QIAprep Spin minipreparation (Qiagen, Valencia, CA) is performed with 5 ml of culture and the remaining 1 ml is diluted to 40% (v/v) glycerol (GIBCO-BRL/Life Technologies, Rockville, MD), properly labeled and stored at - 8 0 °. The plasmid resulting from the minipreparation is checked on a 1% (w/v) agarose gel and its concentration is determined by UV absorption. The presence of the integrase gene insert can be checked by sequencing of the plasmid. Amplification

BL21 (DE3) competent cells containing the pET- 15b/HIV- 1 IN/F185K/C280S vector are spread on an agar plate containing ampicillin (100/zg/ml) and grown overnight. The colonies are then harvested, expanded in 1 liter of fresh LB broth medium containing ampicillin (100 #g/ml), and incubated in a rotary shaker at 37 °. When the cells reach an optical density of 0.5 at 600 nm, overexpression is induced by adding fresh isopropyl-fi-D-thiogalactopyranoside (IPTG; GIBCO-BRL/Life Technologies) to a final concentration of 0.4 mM. Cells are then incubated for an additional 3 hr. Purification

Cells are harvested by centrifugation (10 min, 3000 rpm) and the pellet is resuspended in ice-cold lysis buffer containing 20 mM HEPES, pH 7.5 (Boehringer Mannheim), 1 M NaC1 (Digene), 5 mM imidazole (Sigma-Aldrich, Milwaukee, WI), 2 mM 2-mercaptoethanol (Sigma-Aldrich), and fresh lysozyme (0.2 mg/ml; Boehringer Mannheim) (15 ml of lysis buffer per 500 ml of culture). The lysate is incubated on ice for 30 min with occasional stirring, sonicated on ice for approximately 1 min and centrifuged (20 min, 30,000g, 4°). The supernatant is then applied to a Sepharose column (diameter 1 cm, volume 15 ml; Bio-Rad, Hercules, CA) containing 1 ml of chelating Sepharose Fast Flow (Amersham Pharmacia Biotech, Piscataway, N J) saturated with a solution of 0.4 M NiSO4 (Sigma-Aldrich). Excess NiSO4 has been previously removed by 1 column 7T. M. Jenkins, A. Engelman,R. Ghirlando,and R. Craigie,J. Biol. Chem.271, 7712 (1996).

628


[32]

volume of elution buffer containing 25 mM HEPES (pH 7.5), 0.5 M NaC1, 2 mM 2-mercaptoethanol, and 5 mMimidazole. The supernatant is washed with 3 column volumes of the same elution buffer but with increasing concentration of imidazole (20, 60, and 250 mM). HIV-1 integrase protein is eluted and fractionated with 5 ml of elution buffer at 750 mM imidazole. Protein fractions are checked for integrase on a 12-20% gradient (w/v) Tricine-SDS-protein gel (Novex, San Diego, CA) revealed by Gel Code Blue (Pierce, Rockford, IL). HIV-1 integrase-containing fractions are pooled and dialyzed twice against 1 liter of buffer containing 50 mM HEPES (pH 7.5), 1 M NaC1, 4 mM EDTA (Quality Biological, Gaithersburg, MD), 2 mM dithiotbreitol (DTT; Boehringer Mannheim), and 20% (v/v) glycerol. Integrase concentration is checked by Micro Protein assay (Bio-Rad) and the protein is diluted 2-fold with a 60% (v/v) glycerol solution. The HIV-I integrase is stored "as is" at - 2 0 ° in 25 mM HEPES (pH 7.5), 0.5 M NaC1, 2 mM EDTA, 1 mM DTT, and 40% (v/v) glycerol. Assays

5'-End Labeling of DNA Oligonucleotides In all the following assays (except choice of nucleophile), double-stranded oligonucleotides are 5'-end labeled by T4 polynucleotide kinase (GIBCO-BRL/ Life Technologies). The single-stranded oligonucleotide (10 pmol) is incubated at 37 ° for 30 min in 50 #1 of kinase buffer containing 10/zCi of [y-32p]ATP (Amersham Pharmacia Biotech) and 10 units of kinase. The labeling solution is then applied to the top of a Sephadex G-25 Quick Spin column (Boehringer Mannheim) and the filtrate is annealed with 20 pmol of the complementary strand for 5 min at 95 ° and for 30 min at 37 °.

3'-End Labeling of DNA Oligonucleotides In one of the following assays (choice of nucleophile), the 21-mer singlestranded oligonucleotide A (Table I) is T-end labeled as follows. Ten picomoles of oligonucleotide A is incubated at 37 ° for 60 min in 50 #1 of terminal transferase buffer containing 10 #Ci of ~-cordycepin TP (NEN Life Science Products, Boston, MA) and 15 units of terminal transferase (GIBCO-BRL/Life Technologies). The labeling reaction results in the addition of a single 32p-labeled nucleotide. The labeling solution is then applied to the top of a Sephadex G-25 Quick Spin column and the filtrate is annealed with 20 pmol of oligonucleotide G (complementary strand) for 5 min at 95 ° and for 30 min at 37 °.

Standard Reaction Conditions Unless otherwise indicated, the integrase assays are performed in "integration buffer" containing 25 mM morpholinepropanesulfonic acid (MOPS, pH 7.2),

In Vitro HIV- 1 INTEGRASEASSAYS

[32]

629

TABLE I OLIGONUCLEOTIDESUSED IN ASSAYS

Oligonucleotide

Size

Sequence

A B C D E F G Au

21-mer 21-mer 19-mer !5-mer 34-mer 30-mer 22-mer 21-mer

5'-GTG TGG AAA ATC TCT AGC AGT-3' 5'-ACT GCT AGA GAT TTT CCA CAC-3' 5'-GTG TGG AAA ATC TCT AGC A-3' 5'-GAA AGC GAC CGC GCC-3' 5'-GTG TGG AAA ATC TCT AGC AGG GGC TAT GGC GTC C-3' 5'-GGA CGC CAT AGC CCC GGC GCG GTC GCT TTC-3' 5'-ACT GCT AGA GAT TTT CCA CAC T-3' 5'-GTG TGG AAA ATC TCT AGC UGT-3 ~

7.5 mM MnC12, 14.3 mM 2-mercaptoethanol, and bovine serum albumin (BSA; final concentration, 0.1 mg/ml).

Integration Assay Measuring 3'-Processing and Strand Transfer In the integration assay (Fig. 3A), a 21-mer double-stranded DNA oligonucleotide (oligonucleotide *A, 5'-end labeled and annealed to oligonucleotide B), corresponding to the last 21 bases of the U5 viral LTR (Table I), is used to follow both T-processing and T-end joining (strand transfer) reactions. In the 3'processing reaction, integrase liberates a GT dinucleotide at the 3'-end of the labeled strand, resulting in the generation of a 19-mer labeled product. The strand transfer (3'-end joining) reaction consists of the insertion of a T-processed oligonucleotide into another DNA target. This strand transfer leads to higher and lower molecular weight species migrating slower and faster, respectively, than the original 21-mer substrate (Fig. 3A). The higher molecular weight species (strand transfer products, STP) are generally used to evaluate strand transfer (integration). For inhibitor testing, the compound to be tested is incubated in integration buffer for 30 min at 37 ° with 400 nM HIV- 1 integrase and 5 nM 5'-labeled doublestranded DNA template (*A/B) in a total volume of 10 #1. The reaction is stopped by adding the same volume of electrophoresis denaturing dye containing 99% (v/v) formamide (Sigma-Aldrich, Milwaukee, WI), 1% (w/v) sodium dodecyl sulfate (SDS), bromphenol blue (0.2 mg/ml; Sigma-Aldrich), and xylene cyanol blue (0.2 mg/ml; Sigma-Aldrich). A preincubation for 15 min at 37 ° with either integrase or DNA can be adopted in order to provide advantage to the drug-protein or drug-DNA interaction, respectively. Samples are then heated for 5 min at 95 ° and loaded on 20% (w/v) acrylamide : bisacrylamide (19 : 1) denaturing gel (Accugel; National Diagnostics, Atlanta, GA) containing 7 M urea (GIBCO-BRL/Life Technologies) in TBE (GIBCO-BRL/Life Technologies).

630


[321

B

A

IN

A~

21 3AGT 21

B

3'-P ~

+

IN

C* 8

t9 2A 21

GT 19

19

S.T, ,'I

3A

21

3A

21

STP

21 3AGT

~*i~L

19

+

E D* F

- +

D

C

|5~110~ 15

IN

IN

~AGT*A T

22

BQi '~ qb|

3O

I/Glycerol or

~ Disintegration[ ~ ~ 4

~|

Water,

I~IA-3'OH

Glycerol-GT*A G TWA

o,~,l

IG) lCl

5'*OH-GT*A

FIG. 3. Catalytic activities in vitro of HIV-l integrase. The asterisk marks the location of the 32p radiolabel. Oligonucleotides are denoted by capital letters with reference to Table I. In the autoradiograms to the right of each panel, the left lane represents DNA alone and the right lane represents DNA plus integrase (1N). T-E Y-Processing; S.T., strand transfer (3' end joining); STP, strand transfer products. (A) Dual assay for 3'-processing and strand transfer, using a blunt-ended 21-mer; (B) strand transfer assay using a "precleaved" 19-mer; (C) disintegration assay using a Y-shaped oligonucleotidic structure; (D) assay to determine the choice of nucleophile, using a Y-labeled double-stranded 22-met oligonucleotide.

Strand Transfer Assay In the strand t r a n s f e r a s s a y (Fig. 3B), a " p r e c l e a v e d " 1 9 - m e r Y - e n d l a b e l e d o l i g o n u c l e o t i d e C is a n n e a l e d to its 2 1 - m e r c o m p l e m e n t a r y s t r a n d B. T h i s substrate, c o n t a i n i n g a r e c e s s e d Y - e n d , is u s e d to d e t e r m i n e w h e t h e r the 3 ' - e n d j o i n i n g

[32]

In Vitro HIV- 1 INTEGRASEASSAYS

63 t

reaction is truly inhibited by a compound or whether the decrease in strand transfer products observed in the integration assay is primarily due to an inhibition of the T-processing step. The conditions of this assay are the same as those used for the integration assay measuring T-processing and strand transfer (see above).

Disintegration Reaction The integration reaction (Fig. 3C), catalyzed in vitro by integrase, is in equilibrium with the reverse reaction called disintegration. Disintegration is the only reaction that the core enzyme (without the N- and C-terminal domains) can catalyze. Therefore the disintegration reaction is commonly used to test whether inhibitors act on the integrase core domain. Disintegration can be monitored by using an oligonucleotide that mimicks an integrated product (Fig. 3C). In this reaction, 10 pmol of a 15-mer 5'-end labeled single-stranded oligonucleotide *D is annealed with 20 pmol of oligonucleotides E, B, and F to form a branched Yshaped double-stranded structure (see 5'-End Labeling of DNA Oligonucleotides, above). During catalysis, HIV-I integrase removes the branch and liberates a 30-met 5'-end labeled double-stranded oligonucleotide. Conditions used in this assay are identical to those used for the integration assay.

Choice of Nucleophile in 3~-Processing Reaction The choice of nucleophile in the T-processing reaction (phosphodiester cleavage) can be analyzed with a 22-mer double-stranded DNA oligonucleotide (oligonucleotide A* T-end labeled and annealed to oligonucleotide G) (see 3'-End Labeling of DNA Oligonucleotides, above). Water, glycerol, or T-OH DNA can be used by integrase as nucleophiles. Products resulting from hydrolysis (L), glycerolysis (G), and nucleotide circularization (C) can be monitored by gel electrophoresis (Fig. 3D). If all these products are inhibited to the same extent by a compound, this compound can be described as exerting a global inhibition on the 3'-processing reaction. The reaction conditions are similar to those used in the integration assay.

DNA-Binding Assays Two binding assays are used to monitor integrase-DNA interactions. The first is a photocross-linking assay in which 254 nm ultraviolet light is used to generate free radicals responsible for cross-linking HIV- 1 integrase to the DNA (Fig. 4A). In this assay, the compound to be tested is incubated in integration buffer containing no BSA for 15 min at 37 ° with 1.5 # M HIV- 1 integrase and 5 nM 5'-labeled doublestranded DNA template (A*/B) in a total volume of 10 #1. BSA is removed from the integration buffer to prevent formation of nonspecific photocross-linking products between BSA and the target DNA template. Samples are then irradiated for 15 min under a 254 nm portable UV lamp (UVP, Upland, CA) at room temperature and at

632

ENZYMOLOGY AND BIOLOGICALAPPROACHES

[32]

+

IN

A CAGT

HIV-1 IN

U.V. 254 nm

,4 DPC

4 21

B

CUGT

UDG

*"~"~

+

.,~,~

HIV-1 IN

IN (I DPC

/0

OT

(121

..l~o NaBH4~

~'HoT/o

.,,,,~,,,~ H .,/O

GT

[32]

In Vitro HIV-1 INTEGRASEASSAYS

633

a distance of 1 cm from the light source. One volume of 2x SDS-Tricine loading buffer (Novex) is added to the samples, which are then heated 5 min at 95 ° and resolved in a 12-20% gradient (w/v) Tricine-SDS-polyacrylamide precast gel (Novex). The second method depends on covalent chemical trapping, via reduction with sodium borohydride, of a Schiff base formed between an appropriately positioned e-amino group of a lysine present on the enzyme, and an aldehydic abasic site, enzymatically introduced into the DNA (Fig. 4B). 8 In this Schiff base assay, an abasic site is generated in the DNA template (A*/B) by using a 21-mer single-stranded oligonucleotide that contains a uracil base incorporated in the oligonucleotide (oligonucleotide Au). The Y-end labeled Au oligonucleotide is annealed to its 21mer complementary strand (oligonucleotide B), and treated in the kinase buffer with 1 unit of uracil DNA glycosylase (GIBCO-BRL/Life Technologies) for 60 min at 37 °. The labeling solution is then applied to the top of a Sephadex G-25 Quick Spin column. In this assay, the compound to be tested is incubated in integration buffer containing no BSA for 15 min at 37 ° with 1.5 # M HIV- 1 integrase and 5 nM Y-labeled abasic site-containing double-stranded DNA template (*Au/B) in a total volume of 10/zl. The covalent trapping is initiated by reduction of the Schiff base, using I/zl of fresh 1 M sodium borohydride. After 5 rain at room temperature, samples are treated with I volume of 2x SDS-Tricine loading buffer, heated for 5 min at 95 °, and loaded on 12-20% gradient (w/v) Tricine-SDS-polyacrylamide gels. Conclusion Success obtained with highly active antiretroviral therapy (HAART) has changed the prognosis of acquired immunodeficiency syndrome (AIDS). However, these treatments are limited by side effects and viral resistance. Discovering inhibitors for new targets such as HIV-1 integrase will help to decrease viral load and emergence of drug resistance. The in vitro HIV-1 integrase assays described here can be modified for high-throughput screening 9 and are routinely used for the discovery of HIV-1 integrase inhibitors for the treatment of AIDS.

s A. Mazumder, N. Neamati, A. Pilon, S. Sunder, and Y. Pommier,J. Biol. Chem. 271, 27330 (1996). 9 D. J. Hazuda, J. C. Hastings, A. L. Wolfe, and E. A. Emini, Nucleic Acids Res. 22, 1121 (1994).

FIG.4. DNA-binding activities of HIV-1 integrase. The asterisk marks the location of the 32p radiolabel. Oligonucleotidesare denotedby capital letters with referenceto TableI. In the autoradiogram to the right of each panel, the left lane represents DNA alone and the right lane represents DNA plus integrase. DPC, DNA-proteincross-links.(A) Photocross-linkingassayfor DNAbinding; (B) chemical cross-linking assay for DNA binding.

634


[33]

[33] Use of X e n o p u s Egg Extracts to Study Effects of DNA-Binding Drugs on Chromatin Assembly, Nuclear Assembly, and DNA Replication B y ASMITA KUMAR, HONGZHI XU, a n d GREGORY H. LENO

Introduction

Xenopus egg extract has been extremely valuable for the analysis of numerous integral processes in cell biology including chromatin assembly] ,2 nuclear assembly and nuclear protein import, 3'4 DNA replication, 5 and cell cycle control. 6 One crucial feature of the Xenopus egg, and egg extract, that makes it so useful for studying these processes is the stockpile of components that are required to support the rapid cell division cycles following fertilization. The female germ cell or oocyte is arrested in prophase of meiosis I within the ovary of the adult frog. During this stage of meiotic arrest oocyte growth or oogenesis occurs, resulting in the production of a stockpile of macromolecules and organelles that are required to support the rapid cell cycles in the early embryo. Fully grown oocytes are then induced to complete meiosis I and enter a second stage of arrest in metaphase of meiosis II. The mature oocyte then passes down the oviduct and is released by the frog as an "unfertilized egg." Extracts are prepared from these unfertilized eggs. On fertilization, the egg is released from metaphase arrest and enters interphase, with the first mitotic cell cycle lasting approximately 90 min and the next 11 cycles lasting only 30 min each. The stockpile of components present within the egg not only supports these remarkably rapid embryonic cell cycles in vivo, but also supports the decondensation and remodeling of sperm chromatin, the assembly of remodeled chromatin into functional nuclei, and the replication of nuclear DNA in egg extracts, mimicking the events within the intact egg. Thus, this cell-free system has remarkable potential for determining the effects of DNA-binding drugs on replication-related processes, with the goal of identifying mechanism(s) leading to cytotoxicity and the possible rational design of new anticancer drugs. In this chapter, we describe how to establish and maintain a Xenopus frog colony, how to obtain unfertilized eggs and prepare egg extracts, and how to isolate sperm chromatin as a source of template DNA. We then provide a general t A. R Wolffe and C. Schild, Methods Cell Biol. 36, 541 (1991). 2 G. H. Leno, Methods Cell Biol. 53, 497 (1998). 3 D. D. Newmeyer and K. L. Wilson, Methods Cell Biol. 36, 607 ( 1991 ). 4 M. J. Lohka, Methods" Cell Biol. 53, 367 (1998). 5 G. H. Leno and R. A. Laskey, Methods Cell Biol. 36, 561 (1991 ). 6 A. W. Murray, Methods CellBiol. 36, 581 (1991).

METHODS IN ENZYMOLOGY~VOL.34(1

Copyright~) 2001 by AcademicPress All rightsof reproductionin any formreserved. (X176-6879/00$35.0(I

[33]

USE OF Xenopus EGG EXTRACTS

635

approach for the analysis of drug effects on the remodeling of sperm chromatin, nuclear assembly, and DNA replication, and provide detailed methodology for analysis of each of these processes. Finally, we provide examples of drug effects on nuclear assembly and DNA replication, using this cell-free system. Establishing and Maintaining Frog Colony

Xenopus laevis females may be purchased from one of several supply companies (e.g., Nasco, Fort Atkinson, WI; Xenopus I, Ann Arbor, MI). Frogs may be purchased immature or as sexually mature females. However, given that it may be several weeks before high-quality eggs can be obtained from newly purchased sexually mature frogs, obtaining immature females and housing them until maturity may be the most cost-effective approach. This is especially true if mature females are present within an established colony and new frogs are obtained to replace older animals or to increase colony size. Sperm cells may be obtained immediately after purchase of sexually mature male frogs. Frogs are maintained in covered polyethylene tanks, 66 cm long ×45 cm wide x 30 cm high. The animals are housed at or below a density of one frog per 4 liters of water. Tank water is supplied as well water filtered through activated charcoal and water quality (volatile organic chemicals) is analyzed quarterly. Charcoal filters are changed as needed but no less than twice per year. Ambient room temperature is maintained at approximately 20 °, with the optimal temperature range between 17 ° and 22 °. Illumination is maintained according to a 12-hr light and 12-hr dark cycle. Frogs are fed daily, Monday-Friday. Approximately 1.5 g of Purina Trout Chow is provided for each female frog whereas the smaller, male flogs are provided approximately 1 g of chow each. One hour after adding food pellets to the tanks, any remaining food is removed with a fine mesh fish net. Tank water is exchanged on Monday, Wednesday, and Friday of each week after feeding. Water is removed from each tank with a pump and the tank walls are scrubbed with a brush to remove any debris. The tank is then filled with fresh charcoal-filtered water. The flogs remain within the tank during this procedure and exhibit no adverse effects. Every 2 weeks the tanks are sanitized by thoroughly scrubbing all surfaces with 2 M NaC1. All debris is removed and the tanks are rinsed with charcoal-filtered water. Water pumps and tubing are sanitized in a like manner. Each frog is removed and placed into a freshly cleaned tank prior to the sanitation process. Obtaining Unfertilized Eggs from Sexually Mature Frogs Mature oocytes are arrested at metaphase of meiosis II within the ovary, awaiting ovulation and fertilization. Unfertilized eggs, from which extracts are prepared, are obtained in the following way. Each frog is primed for ovulation by injection

636

ENZYMOLOGY

AND

BIOLOGICAL

APPROACHES

[33]

,=E

P

®I

c~ o~11o

g~

E

,11

~E

~e~J

®o 121

F, .8

.$

.$

& .c:,-

o~

~=

z~

L~

Z 0

\

[33]


637

of 100 IU of pregnant mare serum gonadotropin (PMSG) 4 days before egg collection. This results in follicle cell stimulation and oocyte growth. Twelve to 15 hr before egg collection, each primed frog is injected with 500 IU of human chorionic gonadotropin (HCG) to induce ovulation. Hormone is prepared in deionized water and administered by subcutaneous injection into the dorsal lymph sac. The frogs are then placed in individual tanks containing high-salt Barth's solution [110 mM NaCI, 15 mM Tris-HC1 (pH 7.4), 2 mM KCI, 2 mM NaHCO3, 1 mM MgSO4, 0.5 mM Na2HPO4],7 which prevents spontaneous activation of the eggs. After egg laying slows or stops, the frogs are returned to filtered water and allowed to recover for a minimum of 3 months before they are reinjected with hormone.

Preparing Extracts from Unfertilized Eggs Eggs from each frog are transferred in high-salt Barth's solution to a l-liter glass beaker for processing, keeping the eggs from each frog separate. The external jelly coat of the egg is removed by incubation in 2% (w/v) cysteine hydrochloride, pH 7.8-7.9, for approximately 10 min at room temperature. The beaker should be rocked during incubation to facilitate removal of the jelly coat, which results in the eggs "settling" together, with no visible space between adjacent eggs. Dejellied eggs are then rinsed with Barth's solution [88 mM NaC1, 15 mM Tris-HC1 (pH 7.6), 2 mM KC1, 1 mM MgC12, 0.5 mM CaC12](7) until debris is no longer visible in the rinse solution. Extract quality will be improved if discolored or misshapen eggs are removed before further processing. After eggs are dejellied, they are artificially "activated." Activation is the entry of the cell into interphase and results in contraction of the pigment in the animal hemisphere. This is achieved by incubation of the eggs with the calcium ionophore A23187 for approximately 5 min (0.2-0.5 #g/ml in Barth's solution). After activation, eggs are rinsed extensively in ice-cold extraction buffer [50 mM HEPES-KOH (pH 7.4), 50 mM KC1, 5 mM MgC12, 2 mM 2-mercaptoethanol] (7). After the final rinse in extraction buffer, eggs are gently poured into ice-cold centrifuge tubes. Transferring in a larger volume of buffer will avoid flattening of the eggs against the glass beaker, which can cause them to rupture. After the eggs have settled to the bottom of each tube, much of the excess buffer can be removed with a pipette. The preparation and activities of egg "low-speed" supernatant (LSS) and "high-speed" supematant (HSS) are illustrated in Fig. 1. The eggs are packed by FIG. 1. The preparation of Xenopus egg extracts and their activities. Stage 1 decondensation (adapted from Lu et al?8). Chromatin remodeling: The core histones H2A, H2B, H3, and H4, the spermspecific basic proteins X and Y, the embryonic linker histone B4 (arrowheads), and the high mobility group protein 1 (HMG 1, arrow), are indicated in these two-dimensional Triton-acid-urea (TAU)-SDSPolyacrylamide gels stained with Coomassie blue (from Philpott and Leno 13). Replication time course (from Leng and Lenol°).

638


[33]

centrifugation at 134g for 1 min at 4 ° in an SW50. 1Ti rotor (Beckman, Fullerton, CA). Buffer should be removed from the surface to prevent excessive dilution of the extract. To separate the various egg components, packed eggs should be crushed by centrifugation at 21,000g for 10 min at 4 ° in an SW50.1Ti rotor (Beckman). Egg components separate into three distinct fractions: a lipid cap, a golden cytoplasm layer, and a plug of yolk platelets and pigment. The cytoplasmic layer is collected by breaking through the lipid cap with a cooled Pasteur pipette. Cytochalasin B, aprotinin, pepstatin A, and leupeptin are all added to the cytoplasmic "extract" to a final concentration of 10 #g/ml. The extract obtained after crushing of the eggs is contaminated with yolk, pigment, and lipid material. This material is "cleared" from the extract by centrifugation at 84,000g for 15 min at 4 c'. The cleared extract, or LSS, may be used to study stage I decondensation and remodeling of sperm chromatin, nuclear assembly and stage II decondensation, and DNA replication (Fig. 1). If LSS is the extract of choice, it may be used fresh or frozen as beads in liquid nitrogen for later use. To prevent damage, frozen extracts should be supplemented with glycerol to 1% (v/v). Separation of the soluble extract components from membrane vesicles is achieved by high-speed centrifugation of the LSS. s Approximately 1.5 ml of LSS is placed in a precooled ultraclear (Beckman) centrifuge tube and overlaid with chilled liquid paraffin. The extract is centrifuged at 114,000g for 1 hr at 4 ° in an SW50.1Ti rotor (Beckman). Three major fractions are produced: a transparent soluble fraction or HSS; a membrane vesicle fraction; and a golden, hard ribosomal pellet (Fig. 1). If fractionation is incomplete after 1 hr, the extract should be centrifuged for an additional 30 rain. If the initial volume of LSS is > 1.5 ml, further centrifugation will nearly always be required. The HSS and membrane fractions are collected by side puncture of the tube with a syringe. As with the LSS, the HSS can be used fresh or frozen for later use after addition of glycerol to 7% (v/v). The HSS is able to decondense and remodel sperm chromatin (Fig. 1) but is unable to assemble this chromatin into a functional nucleus and replicate DNA. Under optimal conditions, nuclear assembly and DNA replication can be reconstituted by combining the HSS with the membrane fraction. It is important to note that not all frogs injected with hormone will produce eggs. Furthermore, even if eggs are produced, their quality or quantity may make extract preparation difficult, if not impossible. We find that in a typical group of animals, one-half to three-quarters of the frogs lay eggs that can be used for extract preparation. It is also important to note that the quality of extract is reduced when eggs are allowed to remain in high-salt Barth's solution for long periods of time. Thus, once a sufficient number of high-quality eggs have been produced,

7 j. j. Blowand R. A. Laskey,Cell 47, 577 (1986). M. A. Sheehan,A. D. Mills, A. M. Sleeman,R. A. Laskey,andJ. J. Blow,J. CellBiol. 106, l (1988).

[33]

USE OF

Xenopus EGG

EXTRACTS

639

they should be processed into extract immediately. All these factors should be considered when determining the number of frogs needed for an injection series. The number of eggs needed to make an extract depends on the chosen method of preparation. If large-scale preparation procedures are employed, as described here, a minimum of "-~1000 eggs is required. This can be a disadvantage, especially if smaller frogs, which often produce fewer eggs than larger animals, are used. Smallscale preparations, using benchtop centrifugation procedures, 9 require fewer eggs and may be the method of choice under these conditions. However, preparation of extract in bulk offers many advantages including the thorough characterization of a single extract and its subsequent use in many experimental settings. "Pooling" eggs from different frogs to increase quantity is not recommended even if the eggs appear to be of similar quality. One bad batch of eggs can adversely affect the activity of the whole pooled extract. P r e p a r a t i o n of Xenoptrs S p e r m Nuclei To prepare Xenopus sperm nuclei, 1° six male X. laevis frogs are injected with 50 IU of HCG one week prior to the removal of the testes. Frogs are killed by injection of a lethal dose (120 mg/frog) of tricaine methane sulfonate (MS-222) into the dorsal lymph sac. Administration of MS-222 is considered an acceptable method of euthanasia for amphibian species by the American Veterinary Medical Association. 11 Testes are then removed from the frogs by dissection and placed in 25 ml of modified Barth's saline [MBS; 20 mM HEPES-KOH (pH 7.5), 110 mM NaC1, 1 mM KC1, 0.75 mM CaC12, 0.82 mM MgSO4, 2.4 mM NaHCO3] in a petri dish on ice. Testes are homogenized by chopping them with a razor followed by repeated passage through a 10-ml syringe without a needle. This material is then transferred to two 50-ml conical tubes and the petri dish is rinsed with an additional 20 ml of MBS. The supernatant is then split into two aliquots and each is layered over 10 ml of 70% (v/v) Percoll, prepared in modified Barth's saline, in a 50-ml conical tube. Mature sperm cells are then sedimented by centrifugation at 1510g for 30 min at 4 °. A single centrifugation step is normally not adequate for the recovery of all mature sperm from the testicular homogenate. Therefore, this material should be removed, gently shaken to loosen clumps of debris, layered on fresh Percoll, and recentrifuged as described above. The sperm pellets are then resuspended in buffer NIM [10 mM HEPES (pH 7.5), 2.5 mM MgC12, and 0.2 M sucrose], combined, and resedimented by centrifugation at 1730g for 15 min at 4 °. The pellet is then resuspended in 10 ml of buffer NIM containing lysophosphatidylcholine (LPC, 1.0 mg/ml) with 100 mM phenylmethylsulfonyl fluoride, 9 j. Newport, Cell 48, 205 (1987). ~o F. Leng and G. H. Leno, J. Cell. Biochem. 64, 476 (1997). It E. J. Andrews, J. Ant. Vet. Med. Assoc. 202, 229 (1993).

640


[33]

and kept on ice for 15 min. During this period the sample should be mixed by inverting the tube every minute. LPC disrupts the plasma and nuclear membranes, allowing free movement of macromolecules into the cell. This membrane "permeabilization" is stopped by adding 5 ml of ice-cold buffer NIM containing 3% (w/v) bovine serum albumin (BSA). Permeabilized sperm nuclei are then pelleted by centrifugation at 430g for 15 min at 4 °, washed once in buffer NIM containing 0.4% (w/v) BSA, and finally resuspended in buffer XN-50% (v/v) glycerol [buffer XN: 50 mM HEPES-KOH (pH 7.0), 250 mM sucrose, 75 mM NaC1, 0.5 mM spermidine, 0.15 mM spermine]. The DNA concentration can be determined by counting nuclei with a hemocytometer and assuming a DNA mass of 3.15 pg per haploid genome. 12 On average, >100 #g of DNA is obtained from the testes of a single frog. The nuclei are stored at - 8 0 ° in buffer XN-50% (v/v) glycerol. S t r a t e g i e s for A n a l y s i s of D r u g E f f e c t s o n C h r o m a t i n R e m o d e l i n g , N u c l e a r A s s e m b l y , a n d DNA R e p l i c a t i o n in Xenopus E g g E x t r a c t There are three general strategies for analysis of drug effects on the decondensation and remodeling of sperm chromatin, the assembly of remodeled chromatin into functional nuclei, and DNA replication (Fig. 2). In the first approach (Fig. 2A), drug, Xenopus sperm nuclei (XSN), egg extract (LSS or HSS), and a reaction mixture (Rx Mix) are combined and incubated at 23 °. The type of extract (LSS vs. HSS), and the time of incubation, will depend on the process to be analyzed (see Fig. 1). In the second approach (Fig. 2B), drug and XSN are preincubated in buffer for 15-30 rain to promote drug-DNA interaction in the absence of extract. The XSN are then sedimented by centrifugation at 770g for 10 min at 4 ° and the buffer containing the drug is removed. Egg extract (LSS or HSS), with or without drug, and Rx Mix are then added and the total mixture is incubated as described above. In the third approach (Fig. 2C), drug is preincubated with extract (LSS or HSS) for 15-30 min to facilitate interaction of drug with extract components in the absence of DNA. The XSN and Rx Mix are then added and the sample is incubated as described above. Chromatin Remodeling In general, the remodeling of sperm chromatin that accompanies stage I decondensation involves the replacement of sperm-specific basic proteins (e.g., X and Y) with histones H2A and H2B from the egg, resulting in the formation of nucleosome cores. It is now clear that in Xenopuseggs, this remodeling is mediated by specific factors that participate in both assembly and disassembly processes. 13 In addition 12 I. B. Dawid, J. Mol. Biol. 12, 581 (1965).


[33]

A

SimultaneousAddition:XSN-LSS-Drug ~LSS U

B

641

_ChromatinRemodeling axMix ~

"%'X&DNA Replicat,on

Preincubation:XSN-Drug

U C

Preincubation'-LSS-Drug

I

I .xMix,

N FIG. 2. Strategies for analysis of drug-induced effects on chromatin remodeling, nuclear assembly, and DNA replication.

to the core histones, the embryonic linker histone B4, and high mobility group protein I (HMG1), are assembled on sperm chromatin by the extract. T M These changes in chromatin composition and structure occur within the first 3 to 10 rain of incubation and can be analyzed by one- or two-dimensional polyacrylamide gel electrophoresis and micrococcal nuclease digestion (Fig. 3).

Analysis of Chromatin Remodeling by Egg Extract Fresh or freshly thawed Xenopus egg extract (LSS or HSS) is supplemented with (1) a reaction mixture (Rx Mix) containing an energy-regenerating system [creatine phosphokinase (150 lzg/ml), 60 mM creatine phosphate],7 cycloheximide to a final concentration of 100/zg/ml, ATP to a final concentration of 2 mM, and deoxynucleoside triphosphates (dNTPs) to readjust pool sizes to 50 # M 7' 15 after J3 A. Philpott and G. H. Leno, Cell69, 759 (1992). 14S. Dimitrov, G. Almouzni,M. Dasso, and A. E Wolffe,Dev.Biol. 160, 214 (1993). 15C. J. Hutchison, R. Cox, R. S. Drepaul, M. Gomperts, and C. C. Ford, EMBOJ. 6, 2003 (1987).

642


[33]

Chromatin Remodeling XSN LSS or HSS Rx Mix +/- Drug 3-10'

Electrophoretic Analysis of Chromattn P r o t e ~ r o c o c c a l

Nuclease Digestion of Chromatin

F Dilute extrac~withbuffer A r

Dilute extract 10-foldwith bufferA

Centrifuge 770xg 10" 40C p~reI Rinse chrornatin let with bufferA

C4mthtuge 770xg 10'cushion 40C 90% Parcoll ,.L

~r Centrifuge 770xg 1G'40C

Incubate 5ragchromatin 200 m! MN dtgesti~-~buffer O.05 U mtcrococcalnuck~se

Extract basic proteinswith 0.5M HCt 4oc

Add 200 ml MN terminationbuffer 0,5 mg/ml proteineseK / ~k ISh37oC PhenoJ-chloroturmVextraction Ethanoi p~fecipitation

Centrifuge 16,000xg 10' 4oC ~f Supernatentfrozenin liquid N2

tyoph,~zed

J SDS PAGE 17% PAG - La,~11mllbuffer system Coomassteb~ueo~SYNerstain XSN

XSN LSS

DNA dissolvedin TE buffer 1 rag/m! RNeseA ; lh37oC

~ TAU/SDS 2-D pAGE Samples dissolved: BM Urea 6% Acetic Acid 4% ~-mercaptoethanol 0,01% pyronioY

Resolutionof DNA fragments 1.8% Metaphoragarosegel TBE buffer 6Vlcm XSN

XSN+LSS 3' I0~

1st Dimension:Triton-Acid-UreaPAGE

1.5 mm tube gels 15% acrylamide 8M urea 6raM Triton X-100 16h@70 V

1

Equilibrate tube gels Laemmli stackinggel buffer 2rid Dimension:SDS PAGE 17% PAG - Laernmlibuffer system Coomassleblue or sliver $teJn XSN + XSN LSS

Ethldivmtxomldes~n

Cc~e

~ue ~La~

[33]


643

dilution of the extract16; (2) drug or an equal volume of diluent alone; and (3) XSN at 77 ng of DNA per microliter of extract. The extent of extract dilution is always kept constant at 30% of the initial volume for all incubations. Samples are incubated at 23 ° for 3 to 10 min. Electrophoretic Analysis of Chromatin Proteins To isolate chromatin-associated proteins after incubation of sperm nuclei in Xenopus egg extract, the mixture should be diluted with buffer A [15 mM TrisHC1 (pH 7.4), 60 mM KC1, 15 mM NaC1, 1 mM 2-mercaptoethanol, 0.5 mM spermidine, 0.15 mM spermine] and the nuclei sedimented by centrifugation at 770g for 10 min at 4 °. The nuclei are then resuspended in buffer A, centrifuged under the same conditions, and finally resuspended in a small volume of buffer H [20 mM H E P E S - K O H (pH 7.4), 50 mM KC1]. Basic proteins are extracted from the sperm chromatin by addition of HC1 to a final concentration of 0.5 M. The sample is immediately centrifuged at 16,000g for 10 rain at 4 ° and the resultant supematant is frozen in liquid nitrogen and lyophilized. For one-dimensional analysis, lyophilized proteins are dissolved in water, combined with 2× sodium dodecyl sulfate (SDS) sample buffer and separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) on a 17% (w/v) gel, using the Laemmli buffer system, 17 and visualized with Coomassie blue stain or by silver stain (Fig. 3). Is For two-dimensional analysis by Triton-acid-urea (TAU) (first dimension, left to right)/SDS-PAGE (second dimension, top to bottom), lyophilized proteins are dissolved in 8 M urea, 5% (v/v) acetic acid, 4% (v/v) 2-mercaptoethanol, and 0.01% (w/v) pyronin Y, loaded onto a 1.5-mm tube gel containing 15% (w/v) acrylamide, 8 M urea, and 6 mM Surfact-Amps Triton X-100 (Pierce, Rockford, IL), and separated by running for 16 hr at 70 V. Tube gels are removed and equilibrated for 20 to 30 min with buffer [80 mM Tris-HC1 (pH 6.8), 50 mM dithiothreitol (DTT), 2% (w/v) SDS, 0.01% (w/v)bromphenol blue] and loaded onto 17% (w/v) SDS-polyacrylamide slab gels run using the Laemmli buffer system. 17 After electrophoresis, slab gels are fixed and stained with silver stain or with 0. 1% (w/v) Coomassie Blue R-250 dissolved in 50% (v/v) J6L. S. Cox and G. H. Leno, J. Cell Sci. 97, 177 (1990). 17U. K. Laemmli,Nature (London) 227, 680 (1970). 18Z. H. Lu, D. B. Sittman, D. T. Brown, R. Munshi, and G. H. Leno, J. Cell Sci. 110, 2745 (1997).

FIG. 3. Analysis of chromatin remodeling in Xenopus egg extracts. SDS-PAGE(adapted from Lu et al?8). TAU-SDS two-dimensional PAGE: The core histones H2A (2A), H2B (2B), H3 (3), and H4 (4), and the sperm-specific basic proteins X and Y, are indicated (from Philpott and Lenol3). Micrococcalnuclease digestion of chromatin (adapted from Lenoe).

644


[33]

methanol-7% (v/v) acetic acid. Gels stained with Coomassie blue are destained in 5% (v/v) methanol-7% (v/v) acetic acid 13 (Fig. 3). M i c r o c o c c a l N u c l e a s e D i g e s t i o n of C h r o m a t i n The enzyme micrococcal nuclease has been used to investigate the subunit structure of sperm chromatin before and after decondensation in egg extracts. 13,J9 Specifically, sperm nuclei are used without extract incubation (XSN) or after incubation in egg extract (LSS or HSS) for 3 to 10 min at a DNA concentration of 77 ng/#l extract. Samples are diluted with 10 volumes of buffer A and overlaid onto a 90% (v/v) Percoll cushion in buffer A. Sperm nuclei are sedimented onto the Percoll cushion by centrifugation at 770g for 10 min at 4 °. Chromatin (5 ~g of DNA) is resuspended in 200 #1 of MN digestion buffer [ 15 mM Tris-HCl (pH 8.5), 60 mM KC1, 15 mM NaC1, 0.34 M sucrose, 15 mM 2-mercaptoethanol, 0.5 mM spermidine, 0.15 mM spermine, 2 mM CaC12] and incubated with 0.05 U of micrococcal nuclease (Sigma, St. Louis, MO) at 23 ° for 3 min. Two hundred microliters of MN termination buffer [20 mM Tris-HCl (pH 8.0), 20 mM EDTA, 0.5% (w/v) SDS] is added along with proteinase K (0.5 mg/ml; Sigma) followed by incubation for 16 hr at 37 °. DNA is then purified by phenol-chloroform extraction followed by ethanol precipitation. Dried samples are dissolved in TE buffer [10 mM Tris-HCl (pH 7.4), 1 mM EGTA] containing RNase A (1 Izg/#l) and incubated at 37 ° for 1 hr. DNA fragments are resolved on a 1.8% (w/v) Metaphor agarose gel (FMC, Rockland, ME) with TBE buffer in an 8-V/cm electric field. The gel is stained with ethidium bromide (Fig. 3). The remodeling of XSN by egg extract results in the formation of nucleosomes (in Fig. 3, compare XSN with XSN + LSS). Nucleosome repeat length is determined by estimating the size of DNA fragments on the basis of electrophoretic mobility. DNA molecular weight markers at 100-base pair (bp) intervals are used to plot the standard electrophoretic mobility versus DNA fragment size interpolation curve.

Selecting Appropriate System Both low-speed and high-speed supernatants (LSS and HSS) promote stage I decondensation and remodeling of sperm chromatin.13'20 However, chromatin remodeling in LSS (Fig. 3) is more efficient than in HSS (Fig. 1). The basis for this difference is not clear, but could involve the loss of specific remodeling factors from the LSS after high-speed centrifugation or, alternatively, a reduction in remodeling factor efficiency in the HSS. In spite of this disadvantage, the HSS may be the extract of choice when drug-membrane interactions need to be avoided or if long exposures of DNA to drug are necessary without the potential complications

19 S. Dimitrov, M. C. Dasso, and A. E Wolffe, J. Cell Biol. 126, 591 (1994). 20 A. Philpott, G. H. Leno, and R. A. Laskey, Cell 65, 569 (1991).

[33]


645

associated with the formation of higher order chromatin structure, nuclear assembly, and DNA replication Nuclear Assembly A functional nucleus is assembled from remodeled sperm chromatin after approximately 30 min in egg LSS. This involves the formation of higher order chromatin structure, the assembly of an intact nuclear membrane, the import of nuclear protein, the formation of a nuclear lamina, and chromatin swelling (stage II decondensation). Three features of nuclear structure and function, namely, the assembly of the nuclear membrane and the nuclear lamina, and the import of nuclear protein (Fig. 4), are routinely used to assess nuclear assembly in egg extract.

Assembly of Nuclear Membrane Determining the presence of a nuclear membrane surrounding decondensed sperm chromatin can be achieved with the lipid dye, Nile red (Sigma). Nile red fluoresces primarily within a hydrophobic environment, thereby reducing background fluorescence in the assay. Nuclear membranes are visualized by placing an aliquot of unfixed sample (e.g., 2 #1) on a slide and mixing it with an aliquot of Nile red (0.15/zg/ml) and Hoechst 33258 (5 #g/ml) to label total DNA. A continuous, peripheral Nile red fluorescence (Figs. 1 and 4) indicates complete nuclear membrane formation around sperm chromatin (Fig. 1) in the extract. 16 The anthracycline antibiotic daunomycin has been shown to disrupt nuclear membrane formation in the extract in a concentration-dependent manner. ~0At high drug concentrations (>20 #M) membrane vesicles do not associate with sperm chromatin (Fig. 5), whereas at a lower drug concentration (5/~M) membrane assembly is complete within most nuclei (see Fig. 3 in Leng and Lenol°).

Assembly of Nuclear Lamina The formation of a complete nuclear lamina by egg extract is a late event in the process of nuclear assembly. 21'22 After formation of a transport-competent nuclear membrane (see below), lamin LjII, the most abundant lamin species in egg extract, accumulates within the nucleus, leading to the formation of a complete lamina structure. 23 If the majority of lamin Lln is removed from egg extract by immunodepletion, the resultant nuclei are unable to initiate replication, 24-26

21 C. J. Hutchison, R. Cox, and C. C. Ford, Development 103, 553 (1988). 22 C. J. Hutchison and I. Kill, J. Cell Sci. 93, 605 (1989). 23 C. J. Hutchison, J. M. Bridger, L. S. Cox, and I. R. KillJ. Cell Sci. 107, 3259 (1994). 24 j. W. Newport, K. L. Wilson, and W. G. Dunphy, J. Cell Biol. 111, 2247 (1990). 25 j. Meier, K. H. S. Campbell, C. C. Ford, R. Stick, and C. J. Hutchison, J. Cell Sci. 98, 271 ( 1991). 26 H. Jenkins, T. Holman, C. Lyon, B. Lane. R. Stick, and C. Hutchison, J. Cell Sei. 106, 275 (1993).

646

[33]


Nut:lear Assembly XSN LSS Rx Mix +/- Drug

Jna Ni'°arLam' 30'

Structural Analysis: Nuclear Membrane &

Nuclear Membrane

Functional Analysis: Nuclear Protein Import

Nucle~arProtsth Imoort Transport probe added T7 RNA Polymarase+/- NLS

Nuclear Lamina

Unfixed sample on slide

~0'-1

h

Samples diluted 70-fold with buffer A Overleyed on 30% Sucrose in buffer A

Lipid dye: Ni~eRed (0.15pg/ml) DNA stain: Hoechst 33258 (5~g/rnl)

~

Nuclear Membrane

- - Scin~lletion ~al

Diluted sample

/I

300 Sucrose

Polylysine.coated coverslip Centdfuge 770xg 5' 230C

Nuclei pelleted on coverslips Rinsed with D-PSS Fixed in 4% Paratarmaldehyde/D-PBS10' 230C Rinsed with D-PBS Membranes solubtlizedwith Methanol.20oC t0'

/

Rinsed with D-PBS Blocked twice with TTBS-SSA

Probe with primely antibody Mouse anti-lamin LIII monoolonalantibody Lo46F7(t :250) 1h 23OC

4

Rinsed with TTBS-BSA Sheep anti-mouse Texas Red-conjugated antibody (1:25) r l h 23oc Rinsed with TTBS-BSA DNA stainedwith 5p,g/mlHoechst33258 Nuclear Lamina

Probe with primely antibody Rabbit anti-T7 RNA polyrnaraseantiserum(1:10,000) ~ 1h23~ Rinsed with TTBS-BSA Donkey anti-rabbitfluorescein-conjugatedantibody (1:25) ~ lh23~C Rinsed with "I-FBS-BSA DNA stained with 5pg/ml Hoechst 33258 Nuclear Transport TT-NLS T7+NLS

FIG. 4. Analysis of nuclear assembly in Xenopus egg extracts. Nuclear lamina (from Lu et al. 18).

[33]


647

Nuclear Membrane

Daunomycin (50 I.tM) FIG. 5. A high concentration of daunomycin inhibits assembly of the nuclear membrane in Xenopus egg extract. Xenopus sperm nuclei and daunomycin (50 #M) were incubated simultaneously (Fig. 2A) in egg extract for 90 min. Unfixed nuclei were stained with Nile Red and viewed by confocal microscopy. These nuclei lack the continuous peripheral fluorescence surrounding control nuclei (Fig. 4, Nuclear Membrane) indicating that assembly of the nuclear membrane is incomplete. The fluorescence associated with the sperm chromatin is due to autofluorescence of daunomycin that is visible at high drug concentrations (from Leng and Lenol°).

demonstrating that initiation of DNA replication in egg extract requires the formation of a complete nuclear lamina. To assay nuclear lamina assembly around sperm chromatin, each sample is diluted 70-fold with buffer A, then layered on 30% (w/v) sucrose in buffer A, and the nuclei are sedimented onto a polylysine-coated coverslip by centrifugation at 770g for 5 min at 23 °. The coverslip containing sedimented nuclei is then rinsed in Dulbecco's phosphate-buffered saline (D-PBS) and the nuclei are fixed by incubation in 4% (w/v) paraformaldehyde-D-PBS for 10 min at 23 °. The coverslip is then rinsed in D-PBS and treated with methanol at - 2 0 ° for 10 min. The use of both paraformaldehyde and methanol may not be necessary or even beneficial in all cases. The coverslip is then rinsed in D-PBS, blocked twice with TTBS-BSA [0.1% (v/v) Tween 20, 25 mM Tris-HC1, (pH 8.0), 137 mM NaC1, 2.7 mM KC1, 1.5% (w/v) BSA], and probed for 1 hr or longer at 23 ° with the primary antibody (e.g., mouse anti-lamin Llu monoclonal antibody, L046F72v diluted with T T B S BSA (e.g., 1 : 250). The coverslip is then rinsed with TTBS-BSA and incubated for 1 hr at 23 ° with a fluorochrome-conjugated secondary antibody (e.g., sheep anti-mouse Texas Red antibody) diluted with TTBS-BSA (e.g., 1 : 25). The coverslip is then rinsed with TTBS-BSA and stained with Hoechst 33258 to label DNA. An example of nuclear lamina staining is shown in Fig. 4.18

27 D. Lourim, A. Kempf, and G. Krohne, J. Cell Sci. 109, 1775 (1996).

648

ENZYMOLOGY AND BIOLOGICALAPPROACHES

[33]

Nuclear Protein Import

Nuclear transport is essential for DNA replication in Xenopus egg extract. 28 To determine whether extract-assembled nuclei are capable of nuclear protein import, one can use a transport probe such as bacteriophage T7 RNA polymerase. Natural T7 RNA polymerase is too large to diffuse through the nuclear pore complex, and as such, is excluded from an intact nucleus. However, T7 RNA polymerase, constructed with the nuclear localization signal (NLS) from the simian virus 40 (SV40) large-T antigen, 29 is actively transported through the nuclear pore complex and accumulates within an intact nucleus. 3° T7 RNA polymerase, with the NLS (T7 + NLS) or without it ( T 7 - N L S ) , is expressed in Escherichia coli BL2131 and purified according to the procedure described by Zawadzki and Gross. 32 Nuclear protein import is assayed in the following way. Sperm chromatin, at 1-3 ng of DNA per microliter of LSS, is incubated for 30 min to allow for nuclear assembly. T7 RNA polymerase is then added to the extract and incubated an additional 30 min to 1 hr, giving a total time in extract of 60 to 90 min. Samples are processed as described for the nuclear lamina, and as illustrated in Fig. 4. After fixation, the coverslip is rinsed in D-PBS, blocked twice with TTBS-BSA, probed for 1 hr or longer at 23 ° with the primary antibody (e.g., rabbit anti-T7 RNA polymerase antiserum), and diluted (e.g., 1 : 10,000), with TTBS-BSA. The coverslip is then rinsed with T T B S - B S A and incubated for 1 hr at 23 ° with a fluorochrome-conjugated secondary antibody (e.g., donkey anti-rabbit fluorescein-conjugated antibody), and diluted (e.g., 1 : 25) with TTBS-BSA. The coverslip is then rinsed with T T B S - B S A and stained with Hoechst 33258 to label DNA. An example o f T 7 + NLS accumulation within the nucleus, and of T 7 - N L S exclusion from the nucleus, is shown in Fig. 4. No fluorescence is observed surrounding the nucleus incubated with T 7 - N L S because the nuclei are sedimented through 30% (w/v) sucrose onto a coverslip and away from excluded protein. DNA Replication Low-speed supematants from Xenopus eggs can assemble sperm chromatin into functional nuclei that are able to initiate and complete semiconservative DNA replication under complete cell cycle control] We routinely use three assays for DNA replication as outlined in Fig. 6. 28L. S. Cox, J. CellSci. 101, 43 (1993). 29B. M. Benton, W. Eng, J. J. Dunn, F. W. Studier, R. Sternglanz, and P. A. Fisher, Mol. Cell. Biol. 10, 353 (1990). 30j. j. Dunn, B. Krippl, K. Bernstein, H. Westphal, and E W. Studier, Gene 68, 259 (1988), 31j. Grodbergand J. J. Dunn, J. Bacteriol. 170, 1245 (1988). 32g. Zawadzki and H. J. Gross, Nucleic Acids Res. 19, 1948 (1991).

DNA Replication XSN LSS P,x Mix

+/- Drug

Single Round of Mass of DNA Synthesized

DNA Replication w/In Individual Nuclei ~

Semiconssrvatlve DNA Replication

Terminate reaclion w/ SDS/EDTA buffer

Samples diluted 70-toid wl buffer A Ovedayed on 30% sucrose in buffer A

Terminate reaction wt SDS/EDTA buffer

D~ke~~n'~,e

Degrade protein w l protelnase K (0.5mg/ml) Jr ~ lh

Degrade woteln w/ pmteinase K (0,5mghnl) 37~Cth

~

~

Extract DNA w/ phen ol:chloroform:isoamyl alcohol (25:24:1)

phenei->phenol-chlomform->chlomform

Centrifuge 770xg 5' 23oC

Spot DNA on G F- C flit ers

Nuclei pelleted on coverslips

N p=

TE buffer to 0.Sml 2.5 m1111%+CsCI/TE buffer Refractive Index 1,41

;

Membranes solubittzed w/

1 Precipitate DNA on Washed filters w/ 10% TCA-2% sodium pyrophosphate

Wash filters 3x wl 5% TCA -> 2x w/95% ETOH

0f

1~ eachMap Dry washed filters & count all filters m liquid scin~llant

|

;

Rinsed w/buffer A

~

] -

ndedayaredw/~,

stock CsCI/'rE solution

I~ Texas Red streptevidtn + Hoechst 33258 or Ruorescein sfreptavidin + pmpidium iodide + RNasa A

I I

,00;0, Llqu~@c~a~a~

Time(rain) 0

0

Fraction

FIG. 6. Analysis of DNA replication in Xenopus egg extracts. Density substitution (adapted from Lu et al.36).

650

ENZYMOLOGYAND BIOLOG|CALAPPROACHES

[33]

Incorporation and Detection of Radiolabeled Deoxyribonucleotides The rate and extent of DNA replication in egg extract can be determined by the incorporation of radiolabeled deoxyribonucleotides. The radiolabeled deoxynucleoside triphosphate (dNTP) [ot-32p]dATR is routinely used at 100 #Ci/ml (800 Ci/mmol) to determine the mass of DNA synthesized from a known mass of DNA added to the extract. This requires knowledge of the endogenous dNTP pools in the extract. Using isotope dilution analysis, the dATP 7 and dCTP 15 pools were shown to be approximately 50 #M. All four dNTPs, that is, dATE dCTP, dTTP, and dGTP, are added to the reaction mixture in replication assays to a final concentration of 50 #M to readjust pool sizes after dilution. This not only ensures sufficient dNTP precursors for replication but also allows direct comparison of radiolabeled precursor incorporation between reactions. Sperm nuclei replicate most efficiently when added between 2 and 10 ng of DNA per microliter of extract. 7 Thus, there appears to be an upper and lower threshold of template concentration tolerated by egg extracts. After incubation of the radiolabeled percursor in extract, the reactions are stopped by adding termination buffer [0.5% (w/v) SDS, 20 mM EDTA, 20 mM Tris-HC1 (pH 8.0)]. Proteinase K (0.5 mg/ml) is added and the sample is incubated for 1 hr at 37 °. The DNA is then extracted with phenol, phenol-chloroform, and chloroform and applied to glass fiber filters (GF-C; Whatman, Clifton, N J). A set of two filters is labeled for each sample. Sixty microliters of sample is spotted onto one filter ("washed") and 15/~1 is applied to the other filter ("unwashed"), representing a 4-fold difference in sample volume. The filters are then allowed to dry. Each filter containing the larger volume of sample is then treated with 10% (w/v) trichloroacetic acid (TCA)-2% (w/v) sodium pyrophosphate for 30 min to precipitate the DNA. These filters are then washed three times with 5% (w/v) TCA and then twice in 95% (v/v) ethanol. The washed filters are dried with a heat lamp and then counted, by liquid scintillation, along with the unwashed filters. An average from duplicate filters for each sample is determined. The percentage of [32p]dATP incorporated multiplied by 0.654 yields the mass of DNA synthesized as nanograms per microliter of extract (Fig. 6).33 Titration experiments have shown that daunomycin effectively inhibits replication in the extract, with 50% inhibition occurring at a total drug concentration of 2.7 # M (see Fig. 2 in Leng and Lenol°).

Incorporation and Detection of Deoxynucleotide Analogs Two deoxynucleotide analogs of thymidine, 5-bromodeoxyuridine triphosphate (BrdUTP) and biotinylated deoxyuridine triphosphate (biotin-dUTP), can be

33A. D. Mills, J. J. Blow,J. G. White, W. B. Amos,D. Wilcock,and R. A. Laskey,J. CellSci. 94, 471 (1989).

[33]

USE OF X e n o p u s EGG EXTRACTS

651

used for analysis of DNA replication in the extract. BrdUTP is nearly entirely substituted for thymidine in nascent DNA 34 and biotin-dUTP incorporation has been shown to increase linearly with DNA content in X e n o p u s sperm nuclei replicating in egg extract. 35 Thus, the incorporation of these analogs accurately represents the extent of DNA replication in this system. Two specific features of these analogs, namely density and specificity of detection, can be exploited for the analysis of DNA replication. The dense precursor BrdUTP is combined with [~-32p]dATP, used as a tracer, in density substitution experiments. This procedure allows semiconservative DNA replication to be distinguished from repair synthesis or incomplete strand synthesis and demonstrates whether replication is occurring under cell cycle control, that is, one round per S p h a s e ] DNA (3 ng/#l) is incubated in LSS containing [c¢-32p]dATP at 100 #Ci/ml (800 Ci/mmol) and 0.25 mM BrdUTP for 3 hr. After incubation, the reaction is stopped with termination buffer (see above), proteinase K (0.5 mg/ml) is added, and the sample is incubated for 1 hr at 37 °. The DNA is then extracted with phenol-chloroform-isoamyl alcohol (25 : 24 : 1, v/v/v), brought to a volume of 0.5 ml with TE buffer, mixed with 2.5 ml of 111% CsC1-TE (refractive index, 1.41), and underlaid with 3 ml of stock CsC1-TE solution. Substituted DNA is then separated from unsubstituted DNA by centrifugation to equilibrium, using a 70.1 Ti rotor (Beckman) run at 84,000g for 60 hr at 20 c'. The resultant gradient is fractionated (0.23 ml/min, 200/zl/fraction) and the refractive index of every fifth fraction is determined. Fractions are spotted onto GF/C (Whatman) filters containing 50 # g of herring sperm DNA as carrier and the filters are washed with TCA and counted by liquid scintillation. A typical density substitution profile is shown in Fig. 6. 36 A single peak of radioactivity is detected at a density of 1.75 g/ml (heavy/light DNA, HL), indicating a single round of semiconservative DNA replication and ruling out extensive DNA repair or partial strand synthesis, which would resolve at densities between hemisubstituted (HL) and unsubstituted (light/light, LL) DNA. 37'38 The presence of completely substituted (heavy/heavy, HH) DNA would indicate rereplication within a single incubation, that is, a single cell cycle. Care must be taken when centrifuging high concentrations of cesium chloride as excessive rotor speed, excessive CsCI concentration, overfilled tubes, or low temperature can all cause crystallization and rotor failure. Nascent DNA can be visualized in replicating nuclei through the use of another TTP analog, biotinylated dUTP. Biotin-dUTP incorporated into nascent DNA can be visualized by staining with fluorescent streptavidin. Using flow cytometry 34 R. M. Harland and R. A. Laskey, Cell 21, 761 (1980). 35j. j. Blow and J. V. Watson, EMBO J. 6, 1997 (1987). 36Z. H. Lu, D. B. Sittman, R Romanowski, and G. H. Leno, Mol. Biol. Cell 9, 1163 (1998). 37 H. M. Mahbubani, J. R J. Chong, S. Chevalier, R Thommes, and J. J. Blow, J. Cell Biol. 136, 125 (1997). 38T. Krude, M. Jackman, J. Pines, and R. A. Laskey, Cell 88, 109 (1997).

652

ENZYMOLOGY AND BIOLOGICAL APPROACHES DNA

[33]

Biotin

Daunomyein

(60 rain.) FIG. 7. Daunomycin disrupts the coordinate initiation of DNA replication in Xenopus egg extract. Xenopus sperm nuclei and daunomycin (5 #M) were incubated simultaneously (Fig. 2A) for 60 rain in egg extract containing biotin-dUTP. Nuclei were sedimented onto coverslips, fixed, and stained with Hoechst 33258 (DNA) and Texas Red streptavidin (biotin) to label nascent DNA.

to detect the fluorescent streptavidin-biotin-dUTP complex, Blow and Watson 35 demonstrated that biotin-dUTP incorporation is proportional to DNA synthesis. The biotin-streptavidin system reveals DNA replication within individual nuclei, an approach that has proved extremely useful for analysis of DNA replication in Xenopus extracts. From 20 to 40 # M 5-biotin-16-dUTP is added to the sample, in the Rx Mix, and incubated at 23 ° for 30-120 min. The sample is diluted 70-fold with buffer A, then layered on 30% (w/v) sucrose in buffer A, and the nuclei are sedimented onto a polylysine-coated coverslip by centrifugation at 770g for 5 min at 23 °. The coverslip containing sedimented nuclei is then rinsed in Dulbecco's phosphatebuffered saline (D-PBS) and the nuclei are treated with methanol at - 2 0 ° for 10 min. Nuclei are rinsed with buffer A and stained with fluorescent streptavidin and a total DNA stain such as Hoechst 33258 or propidium iodide. To eliminate any possible nonspecific signal from labeled RNA, RNase A (50 k~g/ml) may be included in the stain mixture when propidium iodide is used. Biotin-labeled nuclei, stained with Texas Red-streptavidin, are shown in Fig. 6. Initiation events occur synchronously or nearly synchronously at the beginning of S phase within individual sperm nuclei replicating in egg extract. 35 Daunomycin has been shown to disrupt the coordination of initiation events, resulting in asynchronous origin firing in egg extract (Fig. 7).1°

Concluding Remarks The interaction of specific drugs with DNA within the cell can affect essential cellular processes, resulting in cytotoxicity. In many cases, the nature of the

[33]


653

interaction is known; however, the mechanisms by which drugs induce cytotoxicity are not. Xenopus egg extract provides an excellent cell-free system for the identification of these mechanisms that may provide the basis for the design of novel anticancer drugs. Acknowledgments We thank Drs. Paul Fisher and Miguel Berrios for the gift of bacteriophage T7 RNA polymerase and antiserum, and Dr. Georg Krohne for the gift of monoclonal antibody L046F7. We also appreciate help and advice from Zhi Hong Lu and Jennifer Johns. Our work is supported by the National Science Foundation.

A u t h o r Index

A

Amirikyan, B. R., 170 Amos, W. B., 650 An, H., 433 Anantha, N. V., 138, 141(91), 589 Andersen, A. H., 612, 616, 619, 619(26), 623(14) Anderson, C. E, 9, 20, 147, 195,519 Anderson, J. R., 242, 247(35; 36) Andrew, W. H., 574 Andrews, D. T., 291 Andrews, E. J., 639 Andrews, W. H., 574 Anin, M.-F., 485 Apiou, F., 591 Arakawa, H., 614 Arakawa, S., 154 Arcamone, E, 21,531 Archer, T. K., 504 Arditti, R. R., 469(3), 470 Armitage, B., 338 Armstrong, R. W., 77 Arnott, S., 319 Arrow, A., 314, 315(142), 316(142) Arslan, T., 617,619 Arthanari, H., 138 Asensio, J.-L., 118, 131(33), 256, 260(17), 269, 278,278(17; 69), 279, 280(69; 81), 313, 314, 315(144), 317 Asensio Alvarez, J. L., 347, 351(35), 353(35) Asseline, U., 343 Atherton, E., 558 Atwell, G. J., 388, 394, 556, 557(10; 12) Aubertin, A.-M., 433 Auer, G., 291 Ausi6, J., 503,504, 507(14), 508, 508(6), 509(14), 516(6; 14) Austin, R. H., 18, 106, 416 Avilion, A. A., 573,574, 574(6) Aymami, J., 212,419 Azam, M., 138, 141(91), 589 Azhikina, T., 489 Azhykina, T. L., 489 Azizkhan, J. C., 529

Abildgaard, E, 257 Aboul-Ela, E, 54, 61(10), 612 Abraham, A. T., 619 Abraham, Z., 377 Abu-Daya, A., 224, 225,419, 422, 422(21) Ackers, G. K., 451,453(18-20) Ackroyd, A. J., 554(39), 555 Adams, R. R., 574 Adhya, S., 249 Afshar, M., 22 Agbandje, M., 133, 134(73), 136(73), 586 Aggarwal, A. K., 521 Agrawal, S., 291 Ahsen, U. V., 22(17), 23 Aich, E, 24, 25(41) Aida, M. J., 170 Aisner, D. L., 576, 581 Aivasashvilli, V. A., 483 ,~kerman, B., 73 Aki, T., 249 Akopyants, N., 489 Alaiya, A. A., 291 Alessi, K., 111, 114(18), 115(18) Alexandrushkina, N. I., 486 Allan, S., 576 Allawi, H. T., 170, 171(35), 172(35), 174(35), 175(35), 176, 176(35), 178(35), 329 Allen, E H., 301 Allersma, M. W., 240 Alley, S. C., 404 Allfrey, V. G., 336, 337 Allison, D. P., 235(3), 236, 238,249 Allshire, R. C., 573(12), 574 Allsopp, R. C., 573(13), 574, 576(14) Almouzni, G., 641 Almqvist, N., 250 Altenburger, W., 421 Altona, C., 257 Amaratunga, M., 167, 168(11), 169(11) Am6, J. C., 591 655

656

AUTHOR INDEX

B Baas, P., 589 Babcock, M. S., 306 Babiss, L. E., 335,580 Bacchetti, S., 573,574(6), 576, 577,578, 581(33), 582(33) Baer, L. M., 281 Baguley, B. C., 24(43), 25 Bailey, S. A., 393 Bailly, C., 10(34), 11, 24, 25(35), 26, 29(51), 99, 107(12), 108(12), 347,391,392, 392(92), 393(92; 93), 424, 433,485,486, 489, 489(24), 490, 493,494, 495,495(24), 496, 496(24; 39), 499, 499(41), 500, 500(30), 504, 509(13), 516(13), 517(13), 518(13), 610, 611,612, 614, 620, 620(19), 622, 622(10; 19; 23) Baird, E. E., 291,450, 456, 464(26), 529, 529(7), 530, 556 Baird, W. M., 380, 389 Bajaj, K., 624 Bajic, M., 26, 29(51) Baker, L. J., 230, 233(29) Balagurumoorthy, P., 143 Balasubramanian, B. N., 611,622(11) Baldeyrou, B., 612 Baldridge, K. K., 300, 302(99), 311(99), 318(99), 324(99) Bambirra, S., 29 Bamdad, C., 240, 241(26) Banaigs, B., 612 Bandyopadhyay, B., 578,580(52) Bankman, I. N., 242 Bannwarth, W., 212 Bansal, M., 143 Banville, D. L., 11, 14(40) Baran, N., 590 Barany, G., 581 Barbas, C. F., 594 Barber, J., 212 Bare, L. A., 579, 582(60) Barfield, M., 281 Barker, D. L., 479 Barlow, S., 337 Barner, J., 254 Barnett, L., 377 Barton, J. K., 289 Bass, M. B., 574 Basu, S., 138

Bates, A. D., 51(2), 52 Bates, D. L., 503,507(5) Bates, P. J., 29, 118, 131(31), 291,313,314, 315(142), 316(142) Batey, R. T., 281 Baty, D., 88 Batz, H. G., 338 Baudoin, O., 344, 356(18) Bauer, C. J., 258,270(25) Bauer, G. B., 518 Baugley, B. C., 380 Baur, J., 581 Bax, A., 254 Bayly, C. I., 293,296(31), 302, 304(31), 305(31), 307(31), 308(31), 311(114; 115), 312(31), 318(114; 115), 319(31), 324(114; 115) Beabealashvilli, R. S., 483 Beach, D., 576 Beal, P. A., 132, 424 Beams, J. W., 150 Bear, D. G., 248 Beau, J.-M., 281 Becker, M. M., 100 Beebe, T. P., 240, 241(25) Beekman, J. M., 131 Beerli, R. R., 594 Behrens, C., 329, 333(5), 556 Beijer, B., 489 Beijersbergen, R. L., 576, 577,578 Bell, S. P., 552, 553(36) Bellard, S., 301 Bellorini, M., 529 Belotserkovskii, B. P., 332, 334 Benatar, L., 237 Benchimol, S., 576, 577(36) Benight, A. S., 165, 166, 167, 167(8), 168, 168(8; 11-13), 169(11-13), 170, 170(8; 15), 171,171(1; 8; 15; 26), 172(1; 8; 15; 26; 31), 173(8), 175(8), 176(8), 178(8) Bennett, R. J., 590, 592(110) Benseler, F., 489 Bentin, T., 332, 334, 336(27) Benton, B. M., 648 Beoge, F., 612, 623(14) Berding, C., 329 Berg, D., 489 Berg, H., 531 Berg, R. H., 90, 329, 330(2), 331(1), 332, 333(5), 338(1), 556

AUTHOR INDEX Berger, I., 283, 299 Bergman, J., 24, 25(41) Bergqvist, S., 313 Berkowitz, M. L., 294 Berman, H. M., 111,283,300, 307(104; 105), 311(105), 324(104), 377 Bemardi, G., 419 Bemardou, J., 212 Bernasconi, C. E, 228(27), 229 Berno, A., 166 Bernstein, K., 648 Bernu6s, J., 551 Berova, N., 69 Berrman, T. A., 529 Berry, D. E., 623 Berry, R. S., 153 Betts, L., 580 Beveridge, D. L., 292, 294, 295,296, 296(52; 53), 297(52; 53; 7l), 300, 307(105), 311(105) Bezailla, M., 247 Bhat, T. N., 300, 307(104), 324(104) Bhattacharyya, A., 579 Bhattacharyya, N. P., 578,580(53) Bianchi, A., 584, 585(87) Bianchi, N., 29 Bibby, M. C., 215 Bichenkova, E. V., 22(14), 23 Bigam, C. G., 257 Bigey, P., 338 Bigg, D. C. H., 620, 622 Binaschi, M., 620 Birchall, A. J., 263 Birch-Hirschfeld, E., 29 Bisagni, E., 99, 110, 131,131(7), 343, 345,346, 347, 347(23; 25), 349, 350, 351(25), 354, 354(19), 355,356, 356(19; 45), 424, 504, 509(13), 516(13), 517(13), 518(13), 623 Bischoff, G., 29 Bischoff, R., 29 Bishop, K. D., 165, 171(1), 172(1), 446 Bishop, M., 576 Bisi, J. E., 335,580 Bissery, M. C., 623 Bjornsti, M.-A., 611 Blaber, M., 130 Blackburn, E. H., 137,573, 574, 574(2), 575, 575(15), 576, 578,579 Bladen, S. V., 27 Blake, R. D., 170

657

Blakeney, A. B., 410 Blanchard, S. C., 22(1 l; 13; 33), 23, 24, 446 Blasco, M. A., 574, 576, 577(29) Blencowe, B. J., 489 Bl6cker, H., 170, 174(24) Blommers, M. J. J., 22(25), 23(25; 32), 24, 433 Bloomfield, V. A., 3, 6(7), 55,168,208,282 Blow, J. J., 637(7), 638,641 (7), 648(7), 650, 650(7), 651, 651(7), 652(35) Blume, S., 131 Boatz, J. A., 300, 302(99), 311 (99), 318(99), 324(99) Bocker, W., 582 Bodnar, A. G., 576 Boelens, R., 259 Boesenberg, C., 612, 623(14) Boffa, L. C., 336, 337 Bogdan, F. M., 22 Bogusz, S., 294 Bohley, C., 29 Bohr, V. A., 485,551 Boles, T. C., 51(3), 52 Bolton, P. H., 138, 144, 280, 299, 324 Bonaventura, F. L., 216 Bond, J. P., 195 Bond, P. J., 319 Bonnard, I., 612 Bontemps, N., 612 Bonven, B. J., 616 Bonvin, A. M. J. J., 259 Borer, P. N., 119, 262, 446 Borgias, B. A., 253 Borochov, N., 503,508(6), 516(6) Bosshard, H. R., 114, 115(25) Bostock-Smith, C. E., 216, 297 Boston, R. C., 479 Botstein, D., 291 Bottomley, L. A., 234, 235(3), 236, 242, 243(34), 247, 247(34-36) Bourdouxhe, C., 504, 509(13), 516(13), 517(13), 518(13) Bourne, P. E., 300, 307(104), 324(104) Bowater, R., 54, 61(10), 612 Boyd, F. L., 99, 109 Boyd, J., 263 Boykin, D. W., 4, 22, 22(21; 22), 23(21; 22), 24, 24(5), 25(5; 35), 26, 29, 29(51), 50(58) Brabec, V., 519 Bradbury, E. M., 248 Brahmachari, S. K., 143

658

AUTHORINDEX

Brameld, K. A., 404 Brafia, M., 495,499(41) Branddn, L. J., 337 Brandts, J. E, 115, 121(26), 128,545 Brandts, J. M., 128 Braunlin, W. H., 519 Breaker, R. R., 291 Bremer, R. E., 529(7), 530 Brennan, C. A., 489 Brenner, S., 377 Brenowitz, M., 451,453(18-21) Breslauer, K. J., 18, 24, 24(42), 25, 25(40), 106, 114, 115(23), 117, 127(23), 128, 142(29), 143, 143(29), 146(29), 168, 170, 171, 172(38), 174(24), 175(23; 38), 177(17), 178(38), 346, 352, 352(32) Breslin, D. T., 242,247(36) Bresloff, J. L., 106 Breusegem, S. Y., 212,219, 225(19) Brice, M. D., 301 Bridget, J. M., 645 Briem, H., 590 Briggs, M. R., 552, 553(36) Britt, M., 11, 12(35), 395 Bronckart, Y., 620 Brondani, V., 22(25), 23(25; 32), 24, 433 Brookes, P., 163,402 Brooks, B. R., 292,293,294 Brooks, C. L., 294 Brooks, M. W., 577 Broude, N. E., 337 Brown, B. R., 394 Brown, D. G., 271,274(74), 354 Brown, D. T., 643,646(18), 647(18) Brown, N. C., 489 Brown, P. M., 224, 225,419, 422(21), 504, 510(11), 511(11), 513(11), 514, 515(11) Brown, P. O., 291,624 Brown, T., 256, 259, 260(17), 261,262(37), 264(40), 265(38; 40), 269, 269(40), 278, 278(17; 69), 279, 280(69; 81), 284, 285(11), 295,313,314, 315(144), 317, 347, 351(35), 353(35) Browne, K. A., 247 Brownlee, R. T. C., 535 Broyde, S., 164 Bruccoleri, R. E., 293 Bruce, J., 579 Bruhn, S. L., 519 Bruice, T. C., 247

BriJnger, A. T., 259 Bryan, T. M., 577 Buc, H., 27,469(4), 470 Buchardt, O., 90, 329, 330, 330(2), 331,331(1), 332, 332(14), 333(5; 1 l), 335,336, 336(29), 338(1; 14), 379, 556, 557(11) Buchner, J., 581 Buckin, V. A., 150, 156, 157 Buckle, M., 27 Bujalowski, W., 422 Bukanov, N. O., 336, 337, 337(33) Bukhman, Y. V., 22, 23(9) Bulichov, N. V., 150 Bulyk, M., 607 Burger, H., 582 Burgin, A. B., 34 Burke, T. G., 53l, 532(17) Burli, R. W., 291,450, 529(6), 530 Burn, R, 581 Busby, R. W., 623 Bushman, E D., 626 Bustamante, C., 240, 249, 250 Butler, P. J., 503, 507(5) Buttinelli, M., 502 Byrn, S. R., 380, 389

C Cabrol-Bass, D., 433 Caddle, S. D., 576, 578 Cai, L., 99, 283 Cain, B. F., 388,394 Calderone, D. M., 519 Caldwell, J. W., 293,296(31), 304(31 ), 305(31), 307(31), 308(31), 312(31), 319(31) Calladine, C. R., 504, 506(17) Callo, E J., 170, 171(26), 172(26) Camara, J., 620 Campbel, I. D., 263 Campbell, K., 292 Campbell, K. H. S., 645 Campbell, V. W., 312 Campisi, J., 576 Cantor, C. R., 119, 130, 459 Cantrell, C. E., 384 Capaldi, D. C., 313 Capobianco, M., 21 Capp, M. W., 168 Capranico, G., 616, 620, 620(29)

AUTHOR INDEX Carlson, J. W., 239 Carlson, R. G., 23(34), 24, 390(89), 391 Carpaneto, E. M., 336, 337 Carpenter, M. L., 423 Carpousis, A. J., 471 Carr, R., 278,280(81), 314, 315(144), 317 Carrasco, C., 614 Carrascosa, J. L., 249 Carrondo, M. A., 212,419 Carter, C. W., Jr., 282 Cartright, B. A., 301 Cary, R. B., 248 Casasfinet, J., 27 Cascio, D., 83 Case, D. A., 254, 296, 300, 307(103) Cassidy, S. A., 132, 313,347,349, 351(35), 353, 353(35) Cathers, B. E., 299 Caufield, C., 300, 303(98), 304(98), 307(98), 311 (98), 319(98), 320(98) Cech, T. R., 117, 138,422,574, 590 Celander, D. W., 422 Cera, C., 509, 516(26), 517(27) Chaboteaux, C., 620 Chabra, A., 620 Chaiken, I. M., 48, 49(77) Chaires, J. B., 3, 4, 5, 8, 8(13), 9(25), 11, 11(16), 12(16; 35), 18, 19, 19(45), 37, 50, 65, 99, 100, 100(18; 19), 101,101(10; 11), 102, 106, 107(12), 108(10-12), 109, 111, 114(20), 115(20), 117(20), 123(20), 124(20), 126, 126(22), 135, 135(20; 39), 157,203,204, 204(14), 205,207,207(14), 216, 223(15), 234, 245,246, 291,298, 308, 309, 319, 348,377,380, 395,493,496, 500, 500(30), 504, 506(8), 508, 508(8), 509, 509(8; 23), 516(8), 518(8), 529, 531,532, 532(17), 539,540(24), 546, 549(24; 30), 550(24), 556, 557,557(1; 2), 568,592,614 Chalikian, T. V., 168, 177(17) Champagne, M., 509 Champoux, J. J., 610, 620 Chan, P., 291,336 Chan, S. S., 18, 106 Chan, V., 169 Chandler, C. J., 535 Chandler, S. E, 354 Chang, A. H., 22(18), 23 Chang, E., 574

659

Chang, G., 300, 303(98), 304(98), 307(98), 311(98), 319(98), 320(98) Changeux, J.-P., 16 Chanteloup, L., 281 Chao, M., 408 Chapman, A., 131 Chapman, K. B., 574 Chartier, M., 88 Chastain, M., 33(70), 34, 435 Chattopadhyaya, R., 167, 168(11), 169(11) Chaudhary, D., 312 Chaw, Y. E, 404 Cheatham, T. E., 292, 294, 294(47), 295,296, 296(57-59), 297(57; 59; 61) Chemeris, V. V., 138, 323,590 Chen, A. Y., 386 Chen, D. J., 248 Chen, L., 99, 283 Chen, Q., 23(30), 24 Chen, S.-E, 578 Cheng, J.-W., 226, 253, 277(7) Cheng, X., 486 Cheng, Y.-K., 294(49), 295,298,298(49) Chemy, D. I., 248,249 Cherny, D. Y., 332 Cheung, C. L., 238 Cheung, H. C., 384 Chevalier, S., 651 Chiang, S.-Y., 529 Chiu, C. A., 574, 576 Choi, S.-D., 347 Chollet, A., 489 Chong, J. E J., 651 Choo, Y., 593,594, 595(2; 9), 597,607, 607(1) Chou, S.-H., 226 Chow, C., 22 Chowdary, D., 404 Chowdhry, B. Z., 4, 100(20), 101,109, 111, 114(20), 115(20), 117(20), 123(20), 124(20), 126(22), 128, 130(47), 134, 135(20), 139, 141(94), 142(94), 216, 223(15), 266, 271(59a), 276(59a), 299, 320(97), 324(97), 347,348(37), 586, 589 Chrisey, L. A., 240, 241(28) Christensen, L., 329, 330, 333(5; 11), 334, 334(8), 338(8), 556 Christiansen, K., 612, 616, 619,619(26), 623(14) Chu, T. M., 151

660

AUTHOR INDEX

Chu, W., 390(89), 391 Chui, C.-P., 577 Church, G., 607 Churchill, M. E. A., 422,424, 513 Cieplak, P., 293,295,296(31; 58), 302, 304(31), 305(31), 307(31), 308(31), 311(114; 115), 312(31), 318(114; 115), 319(31), 324(114; 115), 392 Clairac, R. L. d., 291 Clairac, R. P. d., 291 Clark, G. R., 222, 230, 233(29) Clarke, P. A., 579 Clegg, R. M., 4, 5(21), 212, 213(11), 216(10), 217,219, 225(19), 228(28), 229 Clegg, W., 301 Cleland, A. N., 251 Clement, R. M., 156 Cleveland, J. P., 237, 239, 240 Clever, J. L., 433,443(8) Cliff, C. L., 590 Clore, G. M., 262 Coffin, J., 433 Cohen, G., 546 Cole, M. D., 312 Coil, M., 212, 419 Collier, D. A., 131,133,340, 343, 351(15) Colonna, E P., 21 Colson, P., 24, 433,504, 509(13), 516(13), 517(13), 518(13), 612, 614, 620(19), 622(19; 23) Combaut, G., 612 Comeau, L., 584, 585(87) Condom, R., 433 Conn, G. L., 22, 23(10) Connelly, P. R., 128 Conney, A. H., 163 Cons, B. M. G., 423,515 Conte, M. R., 258,270(25), 271, 274(76) Conway, B. E., 151 Cook, H. J., 573(13), 574, 576(13) Cook, P. D., 433 Coombs, D., 30, 48(67) Cooper, M. A., 29 Corbo, M., 579 Corda, Y., 485 Cordell, J., 371 Corey, D. R., 334, 336(26), 578,580, 580(54; 56)

Cornell, W. D., 293,296(31), 302, 304(31), 305(31), 307(31), 308(31), 311(114; 115), 312(31), 318(114; 115), 319(31), 324(114; 115) Comilescu, G., 254 Correia, J. J., 8, 9(25), 37,245 Cosman, M., 163, 164 Cotter, E, 579 Cotton, E A., 421 Coull, J. M., 330, 333(11), 338 Coulomb, H., 622 Counter, C. M., 573,574, 574(6), 576, 577, 578 Courey, A. J., 552 Coury, J. E., 242, 243(34), 247,247(34-36) Covey, J., 388 Coviello, G. M., 137,576, 581(32), 582(32) Cowell, J. K., 578,580(52) Cox, L. S., 643,645,645(16), 648 Cox, R., 645 Cozzarelli, N. R., 51(3), 52, 386 Craigie, R., 626(7), 627 Crain, P. F., 403 Cramer, H., 578, 580(52) Crane, L. E., 404 Crane-Robinson, C., 168 Crenshaw, J. M., 380 Crescenzi, V., 157 Crick, E H. C., 377 Crooke, S. T., 25,291 Crothers, D. M., 3, 6, 6(7), 8(23), 11(23), 16, 18(43), 51, 54, 55, 62(13), 64(14), 65, 65(14), 99, 100, 102, 106, 194, 262, 389, 466, 467,468,471,479, 494, 504, 506(8), 508(8), 509, 509(8), 515(24), 516(8), 518(8), 546, 550 Crouch, R. J., 30 Crow, S., 499, 500 Crowley, M. E, 295,296, 296(59), 297(59) Cubberley, M. S., 556 Cubero, E., 298 Cui, M., 22(23), 23(23), 24, 25(23), 433 Cui, Q., 292(26), 293 Cullinane, C., 408,466, 470, 471,473,476(7), 477,478,479, 481,483,483(7), 484, 485, 550, 551 Cundliffe, E., 109 Cunningham, D., 579 Curto, E. B., 264 Cutts, S. M., 408,466, 473,479, 481,550

AUTHOR INDEX Czamik, A. W., 22(23; 24), 23(23; 24), 24, 25(23), 433

D Dabrowiak, J. C., 10(33), 11,100(19), 101,418, 429, 431,432, 436, 446, 446(3), 473 da Costa, P. M., 620 Daggett, V., 293 Dahl, O., 334, 335(28), 339 Dahlquist, K. D., 22(13), 23, 23(33), 24, 446 Dale, R., 314, 315(142), 316(142) D'Alessio, G., 435 Dalke, A., 306 Dalla-Pozza, L., 577 Dalma-Weishausz, D., 451,453(21) Danielson, U. H., 26, 49(52) Danyluk, S. S., 260 Daragan, V. A., 262 Darden, T. A., 294, 294(47), 295 Das, A., 500 Das, G. C., 509 Dasso, M. C., 641,644 Dassonneville, L., 24, 433,612, 614, 620, 622,622(23) Dattagupta, N., 16, 18(43), 65, 102,504, 506(8), 508(8), 509, 509(8), 515(24), 516(8), 518(8) Daune, M. P., 156, 509 D'Aurora, V., 519 Dauter, Z., 139, 324 David-Cordonnier, M. H., 489 Davidoff, M. J., 578 Davidson, K. K., 138,589 Davies, B. A., 580 Davies, D. B., 260 Davies, D. R., 130, 301 Davies, J., 22(17), 23 Davies, R. T., 589 Davin, D., 312 Davis, D. R., 29, 255 Davis, J., 130 Davis, P. W., 118 Davis, T. M., 22, 42, 43(76) Dawid, I. B., 640 D'Costa, N. P., 242 Dean, A. C. R., 377,381 Dean, C. J., 371 Dean, D. A., 336, 479

661

Debbs, R., 203 Debizemont, T., 355 Decaestecker, C., 620 Decker, P., 591 Decout, J.-L., 130, 340 Decoville, M., 485 Dedon, P. C., 4, 518 Degrooth, B. G., 239 Deiters, A., 421 de Jong, R., 258,260(27) Delage, S., 350 Delaglio, F., 254 Delaine, E., 240, 248 de Lange, T., 573,576, 581(31), 584, 585(87), 591 Delbarre, A., 383,556, 557(I 3) Delben, F., 157 Delcourt, S. G., 170 De los Santos, C., 164 delos Santos, C., 315 Demaret, J., 297 Demarquay, D., 620, 622 Demeny, T., 300, 307(105), 311(105) Demidov, V. V., 332, 333,334, 336, 337, 337(33), 338,339(20) Demin, V. V., 249 De Mol, N. J., 397 Dempster, M., 573(12), 574 de Murcia, G., 509, 591 Denissenko, M. F., 359 Denny, W. A., 380, 388,394, 556, 557(10; 12) DePinho, R. A., 576, 577(29) Deprew, D. E., 54 DeRose, J. A., 240 Dervan, P. B., 100, 130, 132, 278,291,340, 356(2), 419, 424, 429, 450, 451,452(13; 15-17; 24; 25), 453,456, 458, 462(17), 464(26), 519, 529, 529(6; 7), 530, 556 Descounts, P., 241 Desnoyers, J. E., 151 Devlin, J. J., 579, 582(60) de Vroom, E., 258,260(27) D'Halluin, J. C., 489 Dhesi, J., 313 Dhingra, M. M., 311 Diallo, R., 582 Diaz, M. O., 577 diCera, E., 3, 6(5) Dickerson, R. E., 83, 212, 226, 306, 496 Dickinson, L. A., 450, 529

662

AUTHOR INDEX

Diekmann, S., 212, 213( 11), 494, 499(33) Dignam, J. D., 552 Dimitrov, S., 641,644 D'Incalci, M., 529 Ding, D., 4, 26, 29, 29(51), 50(58) Dingley, A. J., 281 Dingwall, C., 421 di Stefano, F., 579 Dobrowiak, J. C., 393 Dobrynin, V. N., 131 Dockhorn-Dwomiczak, B., 582 Doktycz, M. J., 170, 171(26), 172(26), 249 Dombroski, B. A., 399, 422,519 Dominey, R. N., 291 Dominy, B. N., 294 Donahue, B. A., 485 Dongchul, S., 493,500(30) Donner, P. O., 312 Domer, L. E, 521 Dorr, A., 579 Dosanjh, H. S., 29, 118, 131(31; 33), 141,299 Doubleday, A., 301 Doudna, J. A., 22(26), 23(26), 24 Douglas, K. T., 22(14; 15), 23,212,214, 215, 222(12), 254 Douthwaite, S., 24 Dower, W. J., 600, 608(15) Dragan, A. I., 168 Dragowska, W., 577 Draper, D. E., 22, 23(9; 10), 156, 435 Draper, J. E., 24 Dreier, B., 594 Dremaldi, K. A., 252 Drenth, J., 282,284(2) Drepaul, R. S., 641 Drew, H. R., 413,415(2), 418,419(13), 421, 421(13), 429(13), 503,504, 506(17), 510(7), 511(7; 9), 514(9), 515(9) Dritschilo, A., 248 Driver, D. A., 329, 333(5), 556 Drobny, G. P., 258 Drohat, A. C., 284, 285(13) Droz, E., 241 Drukier, A. K., 337 Drushlyak, K. N., 131 Duan, Y., 296 Dubelman, S., 403 Dudycz, L. W., 489 Dueholm, K., 330, 333(11) Duff, R. J., 22(16), 23

Dumortier, L., 4, 5(21), 212, 216(10), 217 Dunham, M. A., 577 Dunlop, M. G., 573(12), 574 Dunn, J. J., 648 Dunphy, W. G., 645 Dupuis, M., 300, 302(99), 311 (99), 318(99), 324(99) Durand, M., 346 Durchschlag, H., 155 Durland, R. H., 130 Durselen, R., 239 Duval-Valentin, G., 99, 110, 131(7), 355 Duvic, M., 130 Dy, M., 581 Dynan, W. S., 248 Dyson, H. D., 257 Dziewanowska, Z., 579 Dziowgo, C., 157

E Eastman, A., 404 Eastwood, J. W., 294 Eaton, E. N., 574, 577,578 Ebbinghaus, S., 291 Eblings, V., 237 Ebrahimi, S. E. S., 22(15), 23,214, 215,222(12) Ecker, D. J., 25 Eckhardt, G., 579 Eckstein, D. A., 433,443(8) Edsall, J. T., 3 Edwards, K. J., 271 Efremova, E. J., 489 Egan, W. M., 257 Egholm, M., 90, 329, 330, 330(2), 331, 331(1), 332, 332(14), 333(5; 11), 335, 336, 336(29), 338,338(1; 14), 556 Egly, J. M., 551 Ehrlich, K., 486 Ehrlich, M., 486 Eisen, M. B., 291 Eisenberg, H., 503,508(6), 516(6), 546 Elber, R., 306 Elbert, S. T., 300, 302(99), 311(99), 318(99), 324(99) Eldrup, A. B., 339 Elgin, S. C. R., 495 Elliott, W. H., 381 Ellis, T. M., 577

AUTHORINDEX Ellisen, L. W., 578 Elrod-Erickson, M., 599 Elvingson, C., 73, 88 Emini, E. A., 633 Eng, W., 648 Engelke, J., 262 Engelman, A., 626(7), 627 Englezou, A., 577 Erard, M., 509 Erickson, J. W., 27 Erickson, L. C., 485 Eriksson, M., 68, 90, 92,516 Erkkila, K. E., 289 Escud6, C., 340, 345, 349, 350, 354, 354(19), 356(19; 45) Essigmann, J. M., 485, 551 Essmann, U., 294

F Faaland, C. A., 354 Fairall, L., 585 Faldesz, B. D., 165,171(1), 172(1) Fan, J. B., 166 Fan, P., 219 Fan, Y., 610 Fang, G., 590 Fang, Y., 242 Faruqi, A. F., 313, 336 Fasman, G. D., 12, 508,516, 517(39) Faucon, B., 99, 110, 131(7) Faunt, L. M., 8 Fawthrop, S. A., 266 Fede, A., 212 Fedoroff, O. Y., 138, 323,580, 586, 590 Feig, M., 296 Feigon, J., 110, 139, 142(14), 144, 253,278, 278(5), 281,314, 315(143), 317 Feinberg, A. M., 403 Feigner, P. L., 203,337 Felsenfeld, G., 130, 389 Feng, J., 574 Feng, Z., 300, 307(104), 324(104) Ferguson, D. M., 293,296(31), 304(31), 305(31), 307(31), 308(31), 312(31), 319(31) Feriotto, G., 29 Fernandez, C., 281 Fernandez-Saiz, M., 22, 24(5), 25(5)

663

Ferrara, J. D., 144 Ferre-D'Amare, A. R., 22(26), 23(26), 24 Ferrell, T. L., 235(3), 236, 238 Ferrer, N., 551,554(33), 555(33) Fiala, R., 23(31), 24 Filippov, S. A., 131 Filipski, J., 388 Filoche, M., 297 Finch, J. T., 216, 503,508(4), 516(4) Finkel, S. E., 494 Firth, W. J., 382(53), 383,394 Fisher, C., 27 Fisher, J., 266 Fisher, P. A., 648 Fisher, R. J., 27 Fivash, M., 27 Flanagan, J. M., 280 Flashman, E., 418,419(17), 424(17) Florsheimer, A., 22(25), 23(25; 32), 24, 433 Fokt, I., 19,291,308,529,532, 568 Footer, M., 338 Ford, C. C., 641,645 Ford, K. G., 324 Forrow, S. A., 485 Forsyth, G., 345 Foss6, P., 623 Fourcade, A., 248 Fourmy, D., 22(11-13), 23, 23(29; 33), 24, 446 Fox, K. R., 11, 12(35), 110, 117(12), 125, 130(12), 131,131(12), 132, 132(12; 67), 215,224, 225,271,274(78; 79), 291,298, 339, 347,349, 350, 35 I(35), 353,353(35), 354, 391,391(91), 412, 413,414, 415,417, 418,419, 419(17), 421,422, 422(21), 423, 424, 424(17), 426, 427,429(3), 432,452, 504, 510, 510(11), 511(11), 513(11), 514, 515,515(11), 555,556 Fox, T., 293,294(47), 295,296(31; 59), 297(59), 304(31), 305(31), 307(31), 308(31), 312(31), 319(31) Francisco, C., 612 Francois, J. C., 343,556 Frank, R., 170, 174(24) Frankel, A. D., 291 Frank-Kamenetskii, M. D., 110, 117(9), 130, 131,131(9), 170, 176, 332, 333,334, 336, 337,337(33), 338,339(20), 341 Franzen, B., 291 Frau, S., 212 Frazier, J., 301

664

AUTHOR INDEX

Frederick, C. A., 86, 164, 212 Frederick, N. A., 238 Freeman, R., 255 Freier, S. M., 118, 329, 333,333(5), 336, 556 Freire, E., 114, 115(23), 127(23) Frenkiel, T. A., 266, 273,275(64) Fresco, J. R., 170 Freudenreich, C. H., 388 Friedman, J., 388 Friedman, R. A., 9, 10(29; 30) Fritsch, E. E, 363,435,600, 608(14) Fritz, M., 240 Fritzsche, H., 509, 531 Fritzsche, W., 238 Froelich-Ammon, S. J., 385 Frolkis, M., 576 Froussios, K., 563 Frydenlund, H., 338 Fuchs, R. E, 485 Fujimori, A., 610, 620 Fukanaga, M., 384 Fukunaga, M., 379 Fukusawa, K., 614, 622(23) Fulton, T. B., 579 Funk, W. D., 574 Furlong, C., 578 Futcher, A. B., 573,573(13), 574, 576(5; 14) Futscher, B. W., 485

G Gabbay, E. J., 504, 508(12), 517(12) Gaffney, B. L., 117, 142(29), 143, 143(29), 146(29), 256 Gago, E, 557 Galan, A. A., 22(24), 23(24), 24 Galas, D. J., 412, 451 Gale, E. E, 109 Gall, A. A., 240, 241,241(24) Gallego, J., 570 Gallie, D. R., 250 Gallo, E J., 165, 166, 167(8), 168(8), 170(8), 171(1; 8), 172(1; 8), 173(8), 175(8), 176(8), 178(8) Gallo, M. A., 354 Galluci, J. C., 421 Gambari, R., 29 Gamper, H., 485 Gandhi, V., 131

Gao, Q., 287 Gao, X., 203,315 Gao, Y.-G., 283,408 Garbay-Jaureguiberry, C., 287 Garbesi, A., 21 Garcia, R. G., 489 Garcia, R. J., 249 Garcia de la Torre, J., 263 Garestier, T., 131,340, 341,344, 345,346, 347, 347(23; 25), 349,350, 351(25), 354, 354(19), 356, 356(18; 19; 45), 424 Garland, E, 384 Gaub, H. E., 251 Gaudiano, G., 408 Gaudin, E, 281 Gavazzo, P., 516 Gazdar, A., 581 Geacintov, N. E., 24(42), 25, 163, 164, 352 Gee, J. E., 131 Geierstanger, B. H., 291 Geiger, A., 329 Gelasco, A., 164 Gelbin, A., 111,300, 307(105), 311(105) Gellert, M., 389 Gelus, N., 24, 25(35), 433 Gessner, R. V., 519 Ghandour, G., 166 Ghanouni, E, 485 Ghirlando, R., 626(7), 627 Giancotti, V., 157 Gibbs, S. J., 263,269(53) Giberson, R., 239 Gibson, I., 485 Gibson, Q. H., 214 Gibson, T. J., 608 Gieselmann, R., 240 Giesen, U., 329 Giessner-Prette, C., 254 Gilbert, D. E., 110, 142(14) Gilbert, E L., 395 Gilbert, W., 361,416, 458 Gill, S. J., 3, 6(3) Gilley, D., 579 Gillies, G. T., 150, 153(8) Gilliland, G. L., 284, 285(13), 300, 307(104), 324(104) Gimenez-Arnau, E., 297 Giovannangeli, C., 131,340, 343,354 Giriat, I., 591 Gisself~ilt, K., 73

AUTHOR INDEX Glazer, P. M., 291,313,336 Glisson, B., 386 Gluick, T. C., 24, 435 Glukhov, A. I., 578, 580(55) Gmeiner, W. H., 23(28), 24 Gochin, M., 253 Gocke, E., 616 Godfrey, B. J., 239 Gold, L., 291 Goldberg, I. H., 377,518 Goldstein, B., 30, 48(67) Goldstein, R. E, 168, 170, 170(15), 171, 171(15; 26), 172(15; 26) Goljer, I., 280 Gollahon, L., 581 Gomperts, M., 641 Gonzalez, C., 259 Goodisman, J., 10(33), 11,393, 418, 429, 431, 432, 436, 446, 446(3) Goodman, S. R., 336 Goodsell, D. S., 83,496 Gopalakrishnan, S., 500 Gordeev, S. A., 578,580(55) Gordon, M. S., 300, 302(99), 311 (99), 318(99), 324(99) Gordover, L., 620 Gorenstein, D. G., 253,264 Gorenstein, J., 337 Gotfredgen, C. H., 314, 315(143), 317 Gotoh, O., 170 Gottesfeld, J. M., 450, 529 Gottlieb, G. J., 576 Gould, I. R., 293,296(31), 304(31), 305(31), 307(31), 308(31), 312(31), 319(31) Gowan, S. M., 141,299, 589 Gowers, D., 418, 419(17), 424(17), 556 Graham, D. R., 519 Gralla, J. D., 291, 471 Graslund, A., 24, 25(41) Graves, D. E., 377, 379, 379(32), 380, 382(22; 53; 100), 383,383(29), 384, 384(29), 385, 385(32; 33), 386(22; 38), 387,388(74), 391,392, 392(92), 393,393(92; 93), 394, 395,531,532(17), 614 Graves, D. J., 169 Gray, E. J., 222 Gray, J. C., 29 Gray, P. J., 473 Green, D. K., 573(12; 13), 574, 576(13) Green, M. R., 22(20; 22), 23(20; 22), 24, 433

665

Greenberg, W. A., 26, 29(48; 49), 36(48) Greider, C. W., 137, 573,573(13), 574, 574(6), 575(15), 576, 576(5; 14), 577(29) Greig, M. J., 333 Greisman, H. A., 594, 607(7) Grek, S., 130 Greve, J., 239 Griffey, R. H., 25,333 Griffieth, M. C., 336 Griffith, J. D., 584, 585(87) Griffith, M. C., 333 Grigg, G. W., 426 Grigoriev, M., 331 Grimaldi, K. A., 358,358(5; 6), 359, 364(3-6), 375(3~5) Grodberg, J., 648 Gromann, U., 29 Gromova, I. I., 619 Grondard, L., 623 Gronenborn, A. M., 262 Gross, H. J., 648 Grunewald, U., 239 Gryaznov, S., 493 Grzesiek, S. J., 281 Grzeskowiak, K., 212, 226 Grzeskowial, K., 83 Gu, M., 238 Guarcello, V., 314, 315(142), 316(142) Guedj, R., 433 Gueguen, G., 313 Guelev, V. M., 556, 558,559, 565(24), 567(20; 24), 568,570(20; 29) Gueron, M., 263,297 Guida, W. C., 300, 303(98), 304(98), 307(98), 311(98), 319(98), 320(98) Guieysse, A. L., 110, 331,341 Gukovsky, V. P., 150 Gulizia, R. J., 450, 529 Gumport, R. I., 489 Gunderson, S. I., 551 Gunnell, S., 130 Guntherodt, H. J., 238,240 Guo, Q., 138, 143 Gupta, G., 311 Gupta, J., 577 Gupta, M., 610, 619 Gurley, G., 237 Gurskii, G. V., 194 Gutell, R. R., 22, 23(10) Guthold, M., 250

666

AUTHOR INDEX

H Haaima, G., 334 Haber, D. A., 578 Habhoub, N., 130, 340 Haefke, H., 238 Hagerman, P. J., 499 Hagmar, E, 88 Hahn, W. C., 576, 577 Hakkinen, A., 262 Halford, S. E., 554(39), 555 Halim, N. S., 22(24), 23(24), 24 Hall, S. W., 388 Haly, B. D., 433 Hamalained, M., 26, 49(52) Hamana, K., 495 Hamasaki, K., 25 Hamdane, M., 489 Hamilton, S. E., 578, 580, 580(56) Hampel, K. J., 131,345 Hamy, E, 22(25), 23(25; 32), 24, 25(35), 433,614 Han, E X., 138, 141(87), 586, 589 Han, H., 138, 141(87), 278,323,586, 589, 590, 592, 592(110) Han, W. H., 237 Han, Z., 301 Hanawalt, P. C., 485 Hande, M. E, 576, 577(29) Hanlon, S., 167, 168(11), 169( 11) Hannon, G. J., 576 Hansch, C., 394 Hansen, H. E, 334 Hansen, J., 556, 557(11) Hansen, M. S. T., 626 Hansma, H. G., 240, 247,248, 251 Hansma, P. K., 237,238,239, 240, 250, 251 Hanvey, J. C., 131,335,340, 580 Haq, 1,4, 100(20), 101,109, 111,114(19; 20), 115(19; 20), 117(20), 123(20), 124(20), 126(22), 134, 135(20), 139, 141(94), 142(94), 216, 223(15), 266, 271(59a), 276(59a), 299, 320(97), 324(97), 347, 348(37), 586, 589 Harada, K., 291 H~ird, T., 76, 79, 81(13), 219 Hardenbol, E, 303 Hardin, C. C., 117

Hardwick, J. M., 379(31), 380 Harel-Bellan, A., 331 Harland, R. M., 651 Harley, C. B., 137,573,573(13), 574, 574(6), 576, 576(5; 14), 577,581,581(32), 582(32) Harnett, J., 620 Harper, L., 620 Harrington, L., 574 Harrington, R. E., 240, 241 (24) Harris, P. J., 410 Harrison, J. R., 299 Hartl, F. U., 581 Hartley, J. A., 252, 358,358(5; 6), 359, 364(3-6), 375(3-6), 485 Hanson, S. D., 581 Haruki, M., 30 Harvey, S. C., 294, 311 Hassman, C. E, 335,580 Hastie, N. D., 573(12), 574 Hastings, J. C., 633 Hawkins, C. A., 291 Hawley, R. C., 294 Hawn, D. D., 24 l Hayakawa, K., 203 Hayes, J. J., 424, 513 Haynie, D. T., 130 Hazen, E. E., 421 Hazuda, D. J., 633 He, S. M., 106 He, X., 30, 48(67) Hearst, J. E., 150, 485 Hecht, S. M., 22(16), 23,617, 619, 619(30; 31; 35; 36), 623 Hegner, M., 240, 241(27) Heim, G., 248 Heldsinger, A., 22(24), 23(24), 24, 433 H61bne, C., 99, 110, 130, 131,131(7), 137,331, 340, 341,343,344, 345,346, 347,347(23; 25), 349, 350, 351(15; 25), 354, 354(19), 355, 356, 356(18; 19; 45), 392, 424, 504, 509(13), 516(13), 517(13), 518(13), 578 Henderson, E., 117, 238, 239 Henderson, J. E, 167, 168(11), 169(11 ), 381 Hendricksen, T., 294, 300, 303(98), 304(98), 307(98), 311(98), 319(98), 320(98) Hendrix, M., 26, 29(48; 49; 50), 31(50), 36(48), 38(50), 433 H6nichart, J. E, 504, 509(13), 516(13), 517(13), 518(13)

AUTHOR INDEX Henningfeld, K. A., 617,619(35; 36) Henry, D. W., 535 Henry, R. J., 410 Hermann, T., 22, 22(19), 23(19), 24 Hernandez, B., 292(27), 293 Herne, T. M., 29 Herrera, J. E., 11, 12(35), 18, 106, 500, 531 Hertzberg, R. P., 424, 452(24; 25), 453, 519 Heus, H. A., 258,260(27) Hicks, M., 519,520, 521(12), 527(12), 528(12) Higgs, H., 301 Hi|bers, C. W., 257,258,260(27) Hill, H. D. W., 255 Hilsenbeck, S. G., 586 Hingerty, B. E., 164 Hinshelwood, C. N., 377 Hiramoto, R. A., 312 Hirose, Y., 138 Hirshberg, M., 293 Hixon, S. C., 379, 384 Hiyama, E., 581 Hiyama, K., 581 Ho, E, 500 Ho, R L. C., 137,576, 581(32), 582(32) Ho, E S., 100 Hochleitner, K., 359 Hockney, R. W., 294 Hoff, E. V., 156, 157(26) Hoffman, D., 568,570(29) Hoffman, E. E, 337 Hoffman, S., 29 Hofstadler, S. A., 25 Hogan, M. E., 16, 18, 18(43), 65, 106, 130, 131, 416, 509, 515(24) Hoger, T., 591 Hoh, J. H., 242 Hohiesel, J. D., 489 Hoiland, H., 155 Hol, W. G. J., 610 Holbrook, J. A., 168 Holman, T., 645 Holmes, C. E., 22(16), 23 Holmes, R. R., 421 Holt, S. E., 575,576, 581 Holtzman, D. A., 552 Hommel, U., 263

667

Honda, K.-J., 170, 175(36), 176(36), 178(36) Honig, B., 300, 307 Hopkins, P. B., 396, 398,400, 401(7; 8), 402, 404, 408, 519 Homer, J. W., 576 Horz, W., 421 Hotchkiss, R. D., 486 Houssier, C., 24, 433,504, 509(13), 516(13), 517(13), 518(13), 612, 614, 620(19), 622(19; 23) Howard, E B., 301,493 Howell, D. K., 167, 168(12), 169(12) Howland, R., 237 Hoyos, P., 380, 389 Hoyt, P. R., 249 Hsie, L., 166 Hsieh, S. H., 300, 307(105), 311(105) Hu, G., 546, 549(30), 568 Hu, G. G., 284, 285(12) Hu, J., 238 Hu, S., 271 Huang, C. Y., 4 Huang, H., 398 Huang, L., 203,337 Huang, P., 131 Huang, R. C., 508,509(19) Huang, W., 581 Huang, Y. Q., 446 Huang, Z., 238 Huber, P. W., 22(23), 23(23), 24, 25(23) Huchet, M., 622 Huchital, D. H., 519, 520, 521(12), 527(12), 528(12) Huckriede, B. D., 257 Hudson, B. P., 289 Hughes, T. R., 574 Hui, C., 338 Huizenga, D. E., 291 Hummelink, T., 301 Hummerlink-Peters, B. G., 301 Humphrey, W., 306 Hunenberger, P. H., 294 Hunter, W. N., 284, 285(11) Hurley, L. H., 99, 109, 138, 141(87), 323,556, 557(2), 558,567(20), 570(20), 573,582, 583,586, 589, 589(85), 590, 592, 592(110) Hurst, R., 581 Hutchison, C. J., 641,645 Hyrup, B., 329

668

AUTHOR INDEX

lbanez, V., 164 Ikeda, N., 138 Ikuta, S., 167, 168(11), 169(11) Inada, M., 281 lng, N. H., 131 lngacio Tinoco, J., 3, 6(7) Inman, E. M., 592 Ippel, J. H., 257,258, 260(27) Iribarren, A. M., 313,489 lsalan, M. D., 593,594, 595(2; 9), 607(1) Ishihara, T., 334, 336(26) Ishikawa, E, 574 lskandar, M., 504, 507(14), 509(14), 516(14) Itsuki, H., 154 lverson, B. L., 458, 556, 558,567(20), 568, 570(20; 29) Izbicka, E., 138,586, 589 lzvolsky, K. I., 336, 337,337(33)

,J Jackman, M., 65 l Jackson, S. E, 552 Jacobs, B. L., 240, 241(24) Jacobson, E. L., 585,591(88) Jacobson, M. K., 585,591(88) Jacobson, S., 518 Jagadeesh, J., 284, 285(13) Jain, S. C., 389 Jakob, F., 612,623(14) Jakob, U., 581 James, R., 312, 355 James, T. L., 253,257, 258, 259, 259(26), 262, 264(26), 265(26), 270, 270(26), 271, 277(26) Jamison, E., 451,453(21) Janin, Y., 345,354(19), 356(19) Janmey, E, 240 Janota, V., 181, 185 Janssen, L. H., 397 Jares-Erijman, E. A., 159 Jaxel, C., 616, 620, 620(28; 29) Jayaram, B., 294, 295,296(52), 297(52) Jayaraman, K., 131 Jelesarov, 1., 114, 115(25), 168 Jen-Jacobson, L., 223 Jenkins, H., 645

Jenkins, T. C., 4, 29, 99, 100(20), 101,109, 110, 111,113(21), 114(20), 115(20), 117(13; 20), 118, 118(21), 123(20), 124(20; 21), 126, 126(22), 130(13), 131,131(13; 31; 33; 40), 132(13; 67), 133, 133(13), 134, 134(73), 135, 135(20; 39), 136(73-75), 137(13; 16), 139, 141,141(94), 142(13; 94), 157,216, 223(15), 261, 262(37), 264(40), 265(38; 40), 266, 269(40), 271,271 (59a), 273,274(74-76; 79), 275(64), 276(59a; 75), 295,298, 299, 313,320(97), 324(97), 347, 348(37), 350, 485,583,586, 589,589(85) Jenkins, T. M., 626(7), 627 Jennewein, S., 490, 493(26) Jensen, J. J., 300, 302(99), 311 (99), 318(99), 324(99) Jensen, K. K., 329 Jeppesen, C., 427 Jezewska, M. J., 422 Jia, X., 580 Jiang, L., 23(31 ), 24 Jin, R., 117, 142(29), 143, 143(29), 146(29) Jing, T. W., 237 Job, D., 485 Job, E, 4 Johnson, C., 371 Johnson C. S., Jr., 263,269(53) Johnson D. K., 249 Johnson M. L., 8 Johnson R. C., 494 Johnson R. K., 617,619(30; 31) Johnson W. C., 100 Johnston, R. E, 479 Jones, D. E., 291, 312 Jones, R. A., 117, 142(29), 143, 143(29), 146(29), 256 Jones, R. J., 464 Jones, R. L., 11, 14(40), 377 Jonsson, M., 73 Jordan, S. R., 580 Joselevich, E., 238 Josey, J. A., 580 Jovin, T. M., 159, 239,248,489 Jovin, T. E, 228(28), 229 Joy, D. C., 238 Joyce, G. E, 26, 29(50), 31 (50), 38(50), 433 Jucker, F. M., 281 Jujawinski, E., 281 Jung, M., 248

AUTHOR INDEX

K Kadonaga, J. T., 552, 553(36) Kafri, T., 626 Kahlon, J. B., 291 Kahn, N. N., 489 Kainosho, M., 262,281 Kakefuda, T., 54 Kallansrud, G., 119 Kallenbach, K. R., 138 Kallenbach, N. R., 143 Kam, L., 422 Kamitori, S., 390(88), 391,495 Kandimalla, E. R., 291 Kang, C. H., 283 Kankia, B. 1., 149, 156, 159, 165 Kankiya, B. I., 150, 157 Kaptein, R., 259 Karplus, M., 292(26), 293 Kasas, S., 250 Kashin, I., 165 Kashlev, M., 250 Kasprzyk, G., 620 Kasprzyk, P. G., 620 Kastrup, R. V., 385 Katahira, M., 23(27), 24 Kataoka, T., 581 Kathiriya, I. S., 578,580(56) Katipally, R. R., 580 Kawano, T, L., 138 Kawashima, E., 489 Kay, L. E., 262, 281(44) Keams, D. R., 219 Kedersha, N. L., 592 Kehr, A., 612, 623(14) Keiderling, T. A., 181 Kelland, L. R., 141,299, 578,586, 589, 589(92) Kelland, S. M., 299 Keller, D. J., 240 Keller, W., 54 Kelly, K., 312 Kempf, A., 647 Keniry, M. A., 299 Kennard, O., 301 Kennedy, A. W., 578,580(52) Kennedy, M. A., 258, 280 Keppler, M. D., 131,132(67), 298, 354, 418 Kemen, P., 240, 241(27) Kerper, P. S., 249

669

Kerridge, D., 381 Kerrigan, D., 388 Kersten, H., 377 Kersten, W., 377 Kerwin, S. M., 138,299, 323,583,586, 589(85), 590 Kerwood, D. J., 253 Kessler, D. J., 18, 130, 131 Kessler, O. J., 106 Kharakoz, D. R, 154 Khil, P. P., 249 Khudyakov, I. Y., 486 Kickhoefer, V. A., 592 Kielkopf, C. L., 289,450, 529(7), 530 Kiely, J. S., 333 Kill, I., 645 Kilmkait, T., 433 Kim, H. G., 291,312, 355 Kim, J.-H., 332 Kim, K., 248 Kim, K.-H., 332 Kim, M.-S., 347 Kim, N. W., 137,576, 577,581,581(32), 582, 582(32) Kim, S.-G., 265,270(59) Kim, S. K., 83, 90, 329,333(5), 346, 347, 556 Kimura, K., 483 Kingsbury, W. D., 617,619(30; 31) Kingston, R. E., 503,518(2) Kirchner, J. J., 400, 404, 519 Kirk, L M., 377 Kirnos, M. D., 486 Kirolos, M. A., 24, 25(40), 346 Kiselyov, A. S., 132, 345 Kiss, R., 620 Kleider, W., 329 Kleiman, L., 508,509(19) Klement, R., 159, 489 Klimkait, T., 22(25), 23(25; 32), 24, 25(35) Klinov, D. V., 249 Klotz, I. M., 3, 6(4), 100, 106(14) Klug, A., 216, 503,508(4), 516(4), 556, 593, 594, 595(2; 9), 597 Knoll, J. H., 577 Knops, J., 298,304(90) Knowles, J. R., 379 Knox, R. J., 371 Koberle, B., 252 Koch, T. H., 338,408

670

AUTHOR INDEX

Koga, S., 578 Kohn, K. W., 291,388,616, 617,619, 620, 620(28; 29) K6hrle, J., 612, 623(14) Koizumi, M., 291 Kojima, C., 262 Kojiri, K., 614 Kolb, A., 469(4), 470 Kollhagen, G., 617, 619,620 Kollman, E A., 11, 14(38), 292, 293, 294(47), 295,296, 296(31; 57-59), 297(57; 59; 61), 298,302, 304(31), 305(31), 307(31), 308(31), 311(115), 312(31), 318(115), 319(31), 324(115), 392 Kolodziej, E., 380, 389 Kolovos, M., 563 Komata, T., 578 Komura, J., 359, 364(7), 375(7) Kondo, H., 557 Kondo, S., 578 Kondo, Y., 578 Kondoh, A., 27 K6nig, E, 585 Koo, H.-S., 332,494 Kool, E. T., 250 Kopka, M. L., 496 Kornberg, R. D., 503,504, 506(18), 518(3) Kosaak, J. C. T., 203 Koseki, S., 300, 302(99), 311 (99), 318(99), 324(99) Koshlap, K. M., 253,278(5) Koudriakova, T. B., 359 Kowlski, D., 363 Kozloski, S. J., 104 Kozlowski, S. A., 103 Krawczyk, S. H., 313 Kreuzer, K. N., 388 Krippl, B., 648 Krohne, G., 647 Kron, S., 338 Krude, T., 651 Krugh, T. R., 18, 106, 160, 385 Kubinec, M. G., 266 Kubista, M., 69, 88 Kubo, Y., 262 Kuhn, H., 332, 333,337, 339(20) Kuhn, P., 610 Kuklenyik, Z., 315 Kukreti, S., 345, 354, 354(19), 356(19)

Kulkarni, M. S., 385,392(61) Kumar, A., 4, 22, 22(21; 22), 23(21; 22), 24, 24(5), 25(5), 26, 29, 29(51), 50(58), 259, 262, 270, 634 Kumar, P., 23(27), 24 Kumar, S., 29, 118, 131(31), 144, 324 Kung, H. C., 280 Kung, P. E, 256 Kuntz, I. D., 23(30), 24, 590 Kupke, D. W., 149, 150, 151, 153(8), 154, 155(18), 156, 157(18), 158(4), 159, 162(18), 163 Kurachi, A., 577 Kurakin, A., 336, 339(38) Kuroi, K., 581 Kurpiewski, M. R., 223 Kurucsev, T., 69, 77 Kushner, D. M., 578, 580(52) Kwok, Y., 558, 567(20), 570(20) Kypr, J., 493 Kyratzis, I., 535

L Laaksonen, A., 297 Labhardt, A., 212 Laco, G. S., 611 Lacoste, J. C., 343,556 Lacroix, L., 343,556 Ladbury, J. E., 4, 100(20), 101, 111, 114, 115(24), 126(22), 134, 216, 223(15), 347, 348(37), 581,586 Laemmli, U. K., 643 Lagutina, I. V., 249 Lahm, A., 416 Lahmy, S., 612 Lamberson, C., 557 Lamm, G. M., 297,489 Lamond, A. L., 313,489 Lancelot, G., 281 Landsorp, P. M., 576, 577(29) Lane, A. N., 118, 126, 131(33; 40), 252,253, 256, 258,259,260(17), 261,262(37), 263, 264(40), 265(38; 40), 266, 269, 269(40), 270(25), 271,273,274(74-76), 275(64), 276(75), 277(7; 34), 278,278(17; 69), 279, 280(69; 81), 295,313,314, 315(144), 317, 347,351(35), 353(35) Lane, B., 645

AUTHORINDEX Lane, M. J., 165, 166, 167(8), 168(8), 170(8), 171(1; 8), 172(1; 8), 173(8), 175(8), 176(8), 178(8), 446 Laney, D. E., 240, 248 Lange, P., 404 Langley, D. R., 296, 297(72; 75) Langmore, J. P., 138 Langowski, J., 248 Lansdorp, P. M., 576, 577 Lansiaux, A., 612, 622 Laoui, A., 137,578 Lapatsanis, L., 563 Laplante, S. R., 262 LaPointe, J. W., 504, 506(18) Latimer, E W., 249 Larsen, A. K., 623 Larsen, H. J., 332,336, 337, 339(37; 38) Larsen, T. A., 83 Larson, J. E., 202, 389 Larsson, A., 29 Laskey, R. A., 421,510, 634, 637(7; 20), 638, 641(7), 644, 648(7), 650, 650(7), 651, 651 (7) Laskowski, M., 415 Latimer, L. J. P., 131,345 Latt, S. A., 194 Laug'~a, P., 383,556, 557(13) Laughlan, G., 324 Laughton, C. A., 29, 118, 131(31), 271,274(79), 280, 296, 297,298,298(62), 306, 313, 314(62), 317 Lavelle, F., 137,578,623 Lavergne, O., 620, 622 Lavery, R., 300, 308(102), 317, 319(102) Lavesa, M., 414 Lawley, P. D., 402 Lawrence, D. A., 359 Lawrence, R. A., 138,589 Le, H., 250 Le, S., 574 Leadon, S. A., 359 Leaman, D. W., 578, 580(52) Leavitt, A. J., 240, 241(25) Lebedev, A. V., 150 Lebedev, Y., 249, 489 Le Doan, T., 130, 340, 343 Lee, C.-H., 54 Lee, G. U., 240, 241(28) Lee, H., 294 Lee, H. W., 576, 577(29)

671

Lee, J., 568,570(29) Lee, J. K., 291 Lee, J. S., 131,345 Lee, M. S., 579 LeFeuvre, C. E., 573,574(6) Legault, P., 253(9), 254, 281 Legg, M. J., 421 Legha, S. S., 388 Le Grimellec, C., 250 Leharne, S. A., 128, 130(47) Lehn, J. M., 344, 356(18) Lehrach, H., 489 Lemaire, M. A., 485 Lemrow, S. M., 433 Leng, E, 5, 291,546, 549(30), 568,639, 645(10), 647(10), 650(10), 652(10) Leng, M., 485,519 Leno, G. H., 634, 637(13; 20), 639,641,643, 643(2; 13), 644, 644(13), 645(10; 16), 646(18), 647(10; 18), 650(10), 651,652(10) Leo, A., 394 Leonard, G. A., 284, 285(1 l) Leong, T., 489 LePecq, J. B., 380 Le Pecq, J. B., 383,556, 557(13) Lepple, F. K., 156, 157(26) Lerman, L. S., 51,378,380 Leroy, J.-L., 263 Leseur-Ginot, L., 620 Lesnik, E. A., 333 Lesser, D. R., 223 Lesuer-Ginot, L., 620 Leteurtre, E, 617,620 Leupin, W., 212, 222, 266, 556, 557(10) Levenson, C., 132 Levitt, M., 293 Levy, G. C., 262 Levy, M. Z., 573(13), 574, 576(14) Lewis, J. D., 551 Lhomme, J., 130, 340 Li, K., 22, 22(21), 23(21), 24, 24(5), 25(5) Li, M., 238 Li, Q., 390(89), 391 Li, Y. K., 408 Liang, K. W., 337 Liang, X., 337 Liaw, Y. C., 408 Lichtsteiner, S., 576 Lieber, C. M., 238 Liepinsh, E., 266, 267

672

AUTHOR INDEX

Lifson, S., 193, 196(3) Ligget, D., 203 Lillehei, P. T., 234 Lilley, D. M., 54, 61(10), 139, 324, 612 Limborska, S. A., 489 Lin, K. Y., 464 Lin, L.-N., 115, 121(26), 128,545 Lincoln, P., 73,347,570 Lindau, S., 29 Linder, S., 291 Lindsay, S. M., 237, 240, 241(24) Lindsey, J., 573(13), 574, 576(13) Lindsey, L. A., 573(13), 574, 576(13) Ling, Y. H., 291,568 Lingner, J., 574 Lipanov, A. A., 212 Lipari, G., 261 Lipman, R., 404 Lippard, S. J., 86, 164, 519 Lipscomb, L. A., 234, 286(16), 287,289 Lipton, M., 300, 303(98), 304(98), 307(98), 311(98), 319(98), 320(98) Liquier, J., 489 Liskamp, R., 300, 303(98), 304(98), 307(98), 311(98), 319(98), 320(98) Liu, K., 301 Liu, L. E, 52, 386 Liu, L. R., 386 Liu, Q., 578 Liu, X., 24, 25(40) Liu, Y., 203 Live, D., 315 Loakes, D., 354 Locker, D., 485 Lockshin, C., 283 Lohka, M. J., 634 Lohman, T. M., 9, 147,422 Lobse, J., 334, 335(28), 338 Lokey, R. S., 556, 558,567(20), 570(20) Lomonossoff, G. P., 421 London, R. E., 255 Long, B. H., 611,622(11) Loo, J. A., 22(24), 23(24), 24, 433 Loontiens, E G., 4, 5(21), 212, 213(11), 216(I0), 217,219, 225(19), 228(28), 229 Lopez-Guajardo, C. C., 583,589(85) Lorch, Y., 503,504, 506(18), 518(3) Louis, J. M., 262 Louissaint, M., 336, 337 Lourenco, A., 235(3), 236

Lourim, D., 647 Loutfy, R. O., 345 Low, C. M. L., 418,419(13), 421(13), 429(13), 504, 511(9), 514(9), 515(9) Lown, J. W., 531 Lu, M., 138, 143 Lu, X. J., 306 Lu, Z. H., 643,646(18), 647(18), 651 Luce, R. A., 396, 408 Luck, G., 24(43), 25 Ludlum, D. B., 406 Luedtke, N. W., 25 Luger, K., 503 Luisi, B., 139, 301,324 Lund, K., 616 Lundberg, A. S., 577 Lundblad, V., 574 Luneva, N. P., 163 Luque, E J., 292(27), 293,296, 298 Luque, F. Q., 301 Lusthof, K. J., 397 Lutter, L. C., 504, 511(16) Luty, B. A., 294 Lwowski, W., 379, 386(21) Lyamichev, V. I., 130, 131 Lyng, R., 79, 81, 81(13), 92 Lyon, C., 645 Lyubartsev, A. P., 297 Lyubchenko, Y. L., 170, 240, 241,241(24), 249

M Ma, Y., 359 Macaya, R. F., 139, 144, 278 MacCross, M., 260 Machinek, R., 159 Mack, D. P., 22(24), 23(24), 24 Macke, T., 300, 307(103) MacKenzie, L., 623 MacKerell, A. D., 293,297 Mader, A. W., 503 Magasnik, B., 469(3), 470 Magno, S. M., 509, 517(27) Mahbubani, H. M., 651 Maher, L. H., I 18 Mahieu, C., 612 Mailliet, P., 137, 578 Makarov, V. L., 138 Malonne, H., 620

AUTHOR INDEX Malvy, C., 248 Mamoon, N. M., 209 Manalili, S., 25 Mandel, H. G., 381 Maniatis, T., 363,435,600, 608(14) Mann, D. B., 485 Mann, M., 574 Manne, S., 239 Manning, D. D., 26, 29(49) Manning, G. S., 9, 10(29; 30) Manor, H., 590 Mantilla, E. J., 519 Mantovani, R., 529 Mar, V., 574 Marathias, V. M., 144, 299, 324 Marcell, V., 589 Marchand, C., 344, 347,350, 356, 356(18), 424, 486, 624, 626 Marco-Haviv, Y., 590 Margeat, E., 250 Margulis, L. A., 164 Mark, A. E., 292(28), 293 Markgren, P.-O., 26, 49(52) Markley, J. L., 257 Markovits, J., 388, 556, 557(13) Marks, D. S., 22(14), 23 Marks, J. N., 423 Marky, L. A., 111, 114(18), 115(18), 138, 143, 149, 150, 151, 155(18), 156, 157, 157(18), 158(4), 159, 162(18), 163, 165, 170, 174(24), 175(23), 348 Marsch, G. A., 379(32), 380, 385,385(32), 393 Marsh, T., 239 Marsoni, W., 388 Martin, A., 262 Martfn, B., 291,529, 551,551(5), 554(5; 33), 555(33) Martin, M. T., 131,345,347(23) Martin, P. L., 552 Martin, R. E, 230, 233(29) Martin, R. G., 262 Marx, G., 387,388(74) Marzilli, L. G., 11, 14(40), 315 Mascetti, G., 516 Masse, J. E., 281 Masters, J. R., 252 Mathews, D. H., 444 Matousek, J., 181,184(43), 186(43), 187(43) Matsudaira, P., 338 Matsugami, A., 23(27), 24

673

Matsumoto, M., 614 Matsunaga, N., 300, 302(99), 311(99), 318(99), 324(99) Matsuoka, Y., 79 Matsuura, A., 574 Mattaj, I. W., 551 Mattern, M. R., 388 Mattes, W., 388 Matteucci, M. D., 313,464 Matts, R. L., 581 Maurizot, J. C., 346 Maxam, A. M., 416, 458 Maxam, M., 361 Maxwell, A., 51(2), 52 Mayfield, C. A., 291,312, 355 Mayo, J. A., 486 Mayo, K. H., 262 Mazen, A., 509 Mazumder, A., 619, 620, 624, 633 Mazur, S. J., 4, 29, 50(58), 485, 55l Mazzarelli, J. M., 494, 499(33) McAdam, S. R., 358,358(6), 359,364(3; 4; 6), 375(3; 4; 6) McCammon, J. A., 294 McCampbell, C. R., 170 McCloskey, J. A., 403 McConnaughie, A. W., 22, 24(5), 25(5) McCurdy, S., 144 McCutchan, R. E, 22, 23(9) McDark, C. J., 252 McFail-Isom, L., 156, 226, 242, 243(34), 247, 247(34-36), 284, 285(12) McGhee, J. D., 6, 7(24), 15(24), 54, 64(15), 65(15), 77, 99, 193, 194, 196(1), 199(l), 201(1), 202(1), 204(9), 209(9), 245, 246(41), 247(41), 546 McGill, N. J., 573(13), 574, 576(13) McGurk, C. J., 358 McHugh, P. J., 252,358 McKenna, R., 133,134(73), 136(73), 586 McKenzie, S. E., 169 McKie, J. H., 214, 222(12) McLaughlin, L. W., 212, 213(11), 313,401(7), 402, 489, 494, 499(33) McLean, M. J., 426, 427, 432,486 McMahon, N. J., 581 McMurray, C. T., 508,509, 509(20), 515(20), 516(20), 518(28) McNeese, J., 422 McPhail, T., 574

674

AUTHOR INDEX

McPhee, J. D., 200, 203 McPherson, A., 282 McPike, M. P., 43 l, 436 McRee, D. E., 282 Meadows, R. E, 264 Mei, H.-Y., 22(23; 24), 23(23; 24), 24, 25(23), 433 Meier, J., 645 Meister, W.-V., 29 Mel'nikov, S. M., 203,206(15) Menissier de Murcia, J., 591 Mergny, J.-L., 99, 110, 131(7), 137,343,350, 351(15), 355,356(45), 556, 578 Merz, K. M., 293,296(31), 304(31), 305(31), 307(31), 308(31), 312(31), 319(31) Meunier, B., 212, 338 Meyerson, M., 574, 577, 578 Michaelis, L., 377 Michaels, S., 388 Michel, J., 313 Michl, J., 69(4), 70 Miles, H. T., 301,319(108; 109), 493 Milias, G., 563 Millard, J. T., 400, 401(8), 402,519 Miller, D. M., 131,291,298, 304(90), 312, 314, 315(142), 316(142), 355 Miller, J. E, 600, 608(15) Miller, J. L., 294(47), 295 Millero, F. J., 151, 154(11), 155(11; 16), 156, 156(11), 157(26) Milligan, J. F., 34, 464 Mills, A. D., 638,650 Minford, J., 388 Minnock, A., 499, 500, 502 Mirkin, S. M., 110, 117(9), 130, 131,131(9), 341 Mischiati, C., 29 Misra, V. K., 156 Missailidis, S., 297 Mittmann, M., 166 Mizan, S., 132, 345 Mizusawa, H., 54 Moate, E J., 479 Modrich, E, 249 Moelling, K., 312 Mohamadi, E, 300, 303(98), 304(98), 307(98), 311(98), 319(98), 320(98) Mohamed, A. J., 337 Mohammadi, S., 489 Mohanty, D., 143 Mc,llegaard, N. E., 332, 335,489, 494

Moiler, A., 103, 104 Mollman, E A., 302, 311(114), 318(114), 324(114) Molteni, V., 626 Moncollin, V., 529, 551 Mongelli, N., 529 Monod, J., 16 Montenay-Garestier, T., 99, 110, 131(7), 343, 351(15), 392, 504, 509(13), 516(13), 517(13), 518(13) Montgomery, J. A., 300, 302(99), 311(99), 318(99), 324(99) Moody, M. H., 324 Moody, E C., 324 Moore, M. H., 324 Mooren, M. M. W., 257 Moraru-Allen, A. A., 347,351(35), 353(35) Moreau, S., 313 Moreland, D. W., 22(24), 23(24), 24 Morgan, A. R., 396(1), 397 Morin, G. B., 574, 575(16), 576, 581,582(16), 583(16) Morita, M., 384 Moroney, S. E., 486 Morozov, V., 156 Morris, E L., 336 Morse, D. E., 238 Mortensen, K., 88 Morton, T. A., 30, 48, 48(69), 49, 49(69; 77) Moseley, H. N. B., 264 Moser, H. E., 130, 340, 356(2) Mosier, D. E., 450, 529 Moss, H., 584,585(87) Mote, J., 485 Motherwell, W. D. S., 301 Moyzis, R., 99, 283 Mrksich, M., 291 Mucenski, M. L., 249 Mueller, P. R., 358 Muller, S., 591 Miiller, W., 54, 64(14), 65(14), 100, 389 Mullis, K. B., 165 Multani, A. S., 578 Munshi, R., 643,646(18), 647(18) Muraoka, M., 493 Murchie, A. I. H., 139, 324 Murphy, M., 131 Murphy, W. R., Jr., 519, 520, 521(12), 527(12), 528(12) Murr, M. M., 556, 558,567(23), 570(23)

675

AUTHOR INDEX Murray, A. W., 634 Musso, M., 341,349 Myasnikov, V. A., 489 Myers, T. G., 291 Mymryk, J. S., 504 Myszka, D. G., 27, 29, 30, 37, 38(75), 39(75), 41(75), 48, 48(68; 69), 49, 49(69; 77), 50

N Nagahara, L. A., 240 Naghibi, H., 126, 135(41) Nair, T. M., 29 Nakamura, H., 247,574 Nakamura, T. M., 574 Nakanishi, K., 69 Nakano, H., 620 Nakano, S.-I., 170, 175(36), 176(36), 178(36) Nakayama, J., 574 Nasuti, C. A., 623 Neamati, N., 624, 626, 633 Neely, L., 450 Negroni, M., 27 Neidle, S., 4, 29, 50(58), 99, 118, 131, 131(31), 132(67), 133, 134(73), 136(73-75), 141, 222, 261,262(37), 264(40), 265(38; 40), 269(40), 271,274(74; 79), 295,298,299, 313,324, 350, 377,486, 578, 583,586, 589, 589(85) Neilsen, E E., 331 Neilson, E., 379 Nekludova, L., 599 Nelson, E. M., 386 Nelson, H. C. M., 216 Nelson, L. D., 341 Nemunaitis, J., 579 Neretina, T., 249 Nesetril, J., 165, 17l(1), 172(1), 181, 184(43), 185,186(43), 187(43) Nettikadan, S., 249 Neuner, E, 313 Neville, D., 389 Newman, M., 521 Newmeyer, D. D., 634 Newport, J. W,, 639, 645 Newton, B. A., 381 Nguyen, C. H., 99, 110, 131,131(7), 337,345, 346,347,347(23; 25), 349, 350, 351 (25), 354, 354(19), 356, 356(19; 45), 424, 623

Nguyen, K. A., 300, 302(99), 311(99), 318(99), 324(99) Ni, E, 264 Nicholls, A., 300, 307 Nicolini, C., 516 Niedergang, C., 591 Nielsen, P. E., 90, 92,248, 329,330, 330(2), 331, 331(1), 332,332(14), 333, 333(5; 11), 334, 334(8), 335,335(28), 336, 336(27; 29), 337,338,338(1; 8; 14), 339, 339(20; 37; 38), 427,489,494, 556, 557(3) Nieves-Neira, W., 617 Nikitin, A. Y., 359 Nikonowicz, E. P., 281 Nilsson, L., 296 Nilsson, E, 29 Nishikawa, S., 23(27), 24 Nishimura, S., 614, 622(23) Nishioka, D., 589 Nishira, S., 557 Nitiss, J. L., 617 Noble, S. A., 335,580 Nod6n, B., 69 Noftz, S. L., 485 Noguchi, K., 614 Nomura, K., 614 Nord6n, B., 24, 25(41), 68, 69(3), 70, 73, 76, 79, 81,81(13), 83, 86, 88, 90, 92, 94, 329, 331, 332, 333(5), 346, 347,516, 517,556, 570 Nordheim, A., 103, 104 Norman, D. G., 324 Norwood, T. J., 258 Nygren, P.-A., 29

O O'Brien, R., 581 O'Connor, T. R., 359 Oden, E I., 240, 241(24) O'Ferrall, C. E., 240, 241(28) Ogata, C., 287 Ogonskie, N., 579 Ohkubo, M., 614 Ohnesorge, E, 239 Ohyahiki, J. H., 582 Ohyashiki, K., 582 Ojemann, J., 18, 106 Okura, A., 614 Olafson, B. D., 293

676

AUTHOR INDEX

Oldfield, E,, 257 Oldstone, M. B. A., 22 Olivas, W. M., 118 Olmstead, M. C., 20 Olovnikov, A. M., 573,573(11), 574 Olson, W. K., 24(42), 25,300, 306, 307(105), 311(105), 352 O'Malley, B. W., 131 Onfelt, B,, 570 Ono, A., 262, 281 Orgel, A., 377 Omstein, R. L., 170 Orozco, M., 280, 292(27), 293,296, 298, 298(62), 301,314(62), 317 Orum, H., 329, 338 Osborne, M. R,, 163 Osheroff, N., 385, 387,388(74), 611 O'Toole, N., 22(14), 23 Otterbach, E, 582 Otting, G., 266, 267, 268 Ottinger, E. A., 581 Otto, P., 170, 171(34), 172(34) Ouellette, M. M., 576, 581 Ouhashi, K., 23(27), 24 Oulton, R., 574 Owczarzy, R., 165, 166, 167(8), 168, 168(8), 170(8; 15), 171(1; 8; 15), 172(1; 8; 15), 173(8), 175(8), 176(8), 178(8)

P Pabo, C. O., 594, 597(12), 599, 607(7) Pace, B., 336 Pace, N, R,, 34 Pack, G, R., 297 Packman, L, C., 29 Padmanabhan, K. P., 144 Paetkau, V., 396(1), 397 Palecek, E., 426 Paloczi, G, T., 251 Palu, G., 509, 517(27) Palumbo, M., 509, 516(26), 517(27) Pan, C., 577 Pancoska, E, 165, 171(1L 172(1), 181,185 Pandit, B., 578,580(53) Paner, T. M., 165, 166, 167(8), 168(8), 170, 170(8), 171(1; 8; 26), 172(1; 8; 26), 173(8), 175(8), 176(8), 178(8) Panetta, G., 502

Pang, D., 248 Panthananickal, A., 394 Paoletti, C., 380 Papas, T., 27 Paquet, E, 281 Paranjape, J. M., 578,580(52) Pardi, A., 253(9), 254, 281 Park, Y. W., 346 Parkinson, J. A., 212,214, 222(12), 254 Parslow, T. G., 433,443(8) Parsons, P. G., 473 Passner, J. M., 18, 106 Patel, D. J., 23(31), 24, 103, 139, 164, 280, 301, 315,341 Pathak, S., 578 Patino, N., 433 Pavletich, N. P., 581,597(12), 599 Payet, D., 495,496(39) Payton, N., 345 Peace, D. J., 577 Peacocke, A. R., 14,378 Pearl, L. H., 324, 581 Pearson, E. C., 503,507(5) Pedersen, H. G., 294 Pedersen, L. G., 294 Pednault, E. P., 306 Peek, M. E., 234, 282, 286(16), 287, 289 Peffer, N. J., 335,580 Perera, L., 294 Pdrez, J. J., 512,515 Perez-Soler, R., 291, 531,568 Perlman, M. E., 255 Perouault, L., 130 Perrin, D., 356 Perrin, L. C., 483 Perrouault, L., 99, 110, 131(7), 340, 343, 354 Perry, P. J., 110, 131,132(67), 137(16), 141, 298,299, 586, 589, 589(92) Persson, B., 29 Perun, T. J., 109 Pervushin, K., 281 Peterson, R. D., 281 Peterson, S. R., 248 Pettitt, B. M., 130, 294(49-51), 295,296, 298, 298(49; 50) Peytou, V., 433 Pfannschmidt, C., 248 Pfeifer, G. P., 358, 359, 364, 365(17), 372(17), 375,376(20) Pham, T~ Q., 144, 324

AUTHOR INDEX Philippart, E, 620 Phillips, D. R., 390(90), 391,408,466, 467,469, 469(5), 470, 471,473,473(5), 476(7), 477, 478,479, 481,482(10), 483,483(7; 10; 13), 484, 485,535,550, 551 Phillips, K., 139, 301,324 Philpott, A., 637(13; 20), 641,643(13), 644, 644(13) Piatyszek, M. A., 137,576, 581,581(32), 582(32) Pickett, S. C., 479 Piel, N., 489 Pieper, R. O., 485 Pietrasanta, L. I., 251 Pilch, D. S., 24, 24(42), 25, 25(40), 131,132, 143,345,346, 347(23; 25), 351(25), 352, 352(32) Pillus, L., 518 Pilon, A. A., 619, 624, 633 Pinder, R. M., 381 Pineda, A. R., 30, 48(67) Pines, J., 651 Piotto, M., 269 Piper, P. W., 581 Pippo, L., 29 Pitts, A. E., 580 Pitts, E. E., 578,580(54) Pjura, E E., 212 Pla Rodas, E, 620 Plaskon, R. R., 170, 234, 286(16), 289 Plaxco, K. W., 251 Plotnikov, V. V., 128 Plukett, W., 131 Plum, G. E., 24, 25(40), 110, 117(10), 118(10), 128, 131(10), 143, 168, 177(17), 208 Pogozelski, W. K., 422 Poland, D., 167, 168(10), 169(10), 193 Polucci, E, 131,350 Polycarpou-Schwarz, M., 551 Pommier, Y., 291,388,610, 611,616, 617, 619, 620, 620(28; 29), 624, 626, 633 Pons, D., 620 Poremba, C., 582 Porter, S. E., 620 Portugal, J., 291,422, 427,432, 503,504, 512, 514, 514(10), 515, 515(10), 529, 551, 551(5), 554(5; 33), 555(33) Post, C. B., 264 Potaman, V. N., 110, 117(15), 118(15), 349 Potapov, V. K., 489

677

Potekhin, S. A., 127, 128 Pourquier, E, 610, 611,617,619 Pouyet, J., 509 Povirk, L. E, 518 Prabhakaran, M, 311 Praseuth, D., 110, 130, 331,340, 341 Prescott, C. D., 22 Prescott, D. M., 138 Presnell, S. R., 234, 247, 286(16), 289 Prestegard, J. H., 280 Preuss, W., 239 Prevost, G., 620 Pribble, J., 579 Price, M. A., 422 Price, E A., 415 Priebe, W., 5, 19, 291,308,529, 531,532, 532(17), 539, 540(24), 546, 549(24; 30), 550(24), 551, 551(5), 554(5; 33), 555(33), 568 Priestley, E. S., 433 Priestley, W. S., 26, 29(48; 50), 31(50), 36(48), 38(50) Pritchard, L. L., 331 Privalov, P. L., 127, 168 Prodromou, C., 581 Prokhorov, V. V., 249 Propst, C. L., 109 Prosnyak, M. I., 489 Prosser, J. K., 117 Prowse, K. R., 137,576, 581(32), 582(32) Prudhomme, M., 614, 620(19), 622(19) Pruitt, K. M., 384 Prunell, A., 510 Przewloka, T., 291,529, 532,568 Puglisi, J. D., 22(11-13), 23, 23(29; 30), 24, 33(70), 34, 281,435,446 Pulleyblank, D. E., 54 Pullman, B., 254, 531 Pulyaeva, H., 337 Pursell, C., 558, 567(20), 570(20) Putman, C. A., 239 Puvvada, M. S., 485 Pyle, A. M., 34

Q Qin, P. Z., 34 Qiu, X., 610 Qu, X., 4, 8(13), 19, 308,377. 614

678

AUTHORINDEX

Quada, J., 583,589(85) Quadrifoglio, E, 157 Quigley, G. J., 283,309, 311(126), 320(126), 419, 519, 531,533(18), 535(18) Quintana, J. R., 212,226

R Rabbani, A., 504, 507(14), 509(14), 516(14) Rackwitz, H. R., 489 Radhakrishnan, I., 280, 315,341 Radmacher, M., 240 Raghavan, S., 99 Raghunathan, G., 301,319(108; 109) Raghuraman, M. K., 117 Ragsdale, C. W., 600, 608(15) Ragunathan, K. G., 22, 24(5), 25(5) Rahmouni, A. R., 485 Rajewsky, M. E, 359 Rajur, S. B., 401(7), 402 Rama Krishna, N., 264 Ramasamy, K., 336 Rando, R. R., 25 Rao, S. N., 11, 14(38), 392 Ratliff, R., 99, 283 Ratmeyer, L., 22(22), 23(22), 24 Raucher, S., 400, 401(8), 402,519 Rauen, H. M., 377 Ravishanker, G., 295,296(53), 297(53) Ray, R., 131 Raymond, E., 138,578,586, 589 Rayner, D. M., 345 Razzano, G., 21 Read, M. A., 13l, 132(67), 141,298,299, 589 Rebar, E. J., 594 Recht, M. I., 22(11; 13), 23, 23(33), 24, 446 Record, M. T., Jr., 9, 20, 147, 168, 195,519 Reddel, R. R., 577 Reddoch, J. E, 291,314, 315(142), 316(142) Redinbo, M. R., 610 Reedjil, J., 519 Rees, D. C., 289, 450, 529(7), 530 Reese, C. B., 313 Reeves, R. H., 242 Regenfuss, P., 4, 5(21), 212,216(10), 217 Rehfuss, R., 393,418,432, 446, 446(3) Reich, E., 377 Reich, N. O., 249

Reid, B. R., 226, 253,258,265,270(59), 277(7), 570, 580 Reid, C. M., 168 Reimann, P., 238 Reines, D., 485 Reinhardt, C. G., 385 Reinhold, W. C., 291 Reinhoudt, D. N., 397 Reiss, J. A., 535 Ren, J., 4, 99, 101(10; 11), 107(12), 108(10-12), 111, 114(20), 115(20), 117(20), 123(20), 124(20), 126, 135(20; 39), 157,298, 348, 592 Rentzeperis, D., 111, 114(18), 115(18), 156, 157, 159, 163,348 Resing, K., 408 Reszka, A. E, 131, 132(67), 133, 134(73), 136(73), 141,298,299, 586, 589 Revenko, |., 248 Rewcastle, G. W., 380 Reynolds, M. A., 259 Reynolds, P. E., 109 Rhee, S., 301 Rhoads, R. E., 166 Rhodes, D., 510, 585 Riccelli, P. V., 165, 171(1), 172(1) Rice, S. A., 153 Rich, A., 99, 103, 104, 130, 212, 283,309, 311(126), 320(126), 419, 519,531, 533(18), 535(18) Rich, N., 473 Rich, R. L., 27 Richards, N. G., 300, 303(98), 304(98), 307(98), 311(98), 319(98), 320(98) Richenberg, J. L., 427, 432 Richmond, M. H., 109 Richmond, R. K., 503 Richmond, T. J., 503 Richter, M., 251 Ridge, G. S., 391,392, 392(92), 393(92; 93) Rief, M., 251 Riggs, A. D., 358, 359, 364(7), 375(7) Rigl, C. T., 22, 24(5), 25(5) Rigl, T., 22(21), 23(21), 24 Rill, R. L., 379(32), 380, 385,385(32), 393,508, 516(22) Rink, S. M., 401(7), 402 Riordan, J. E, 435 Riou, J.-E, 137,578,614, 620(19), 622(19), 623 Rippe, K., 250

679

AUTHOR INDEX Risen, L. M., 333,336 Rivalle, C., 343,504, 509(13), 516(13), 517(13), 518(13) Rivetti, C., 240 Rizzo, V., 21 Roberts, J. J., 371 Robertson, M. W., 416 Robertson, S. A., 291 Robinson, H., 539, 540(24), 549(24), 550(24) Robinson, M. J., 387 Robinson, M. O., 574 Robles, J., 313 Rock, S. G., 394 Rodger, A., 69(3), 70, 81,346 Rodgers, J. R., 30l Rodrigues-Pereira, E., 614, 620(19), 622(19) Rodu, B., 312,355 Roe, J. A., 139 Roe, S. M., 581 Roeder, R. G., 551,552 Roesel, J., 312 Rogers, J., 22(18), 23 Rogers, M. J., 22, 23(9) Rolland, A., 620 Rolli, V., 591 Romanowski, E, 651 Rome, L. H., 592 Roques, B. E, 287,383,556, 557(13) Rosamond, J., 212 Rosamund, J., 254 Roselt, P. D., 313 Rosen, N., 581 Rosenberg, J. M., 296 Rosenbohm, C., 26, 29(49) Rosendahl, G., 24 Rosenfield, S., 584, 585(87) Rosenzweig, A. C., 86, 164 Ross, D. T., 291 Ross, J., 153 Ross, P., 579 Ross, W. E., 386 Roug6e, M., 99, 110, 131,131(7), 343,346, 347(25), 351(15; 25) Rould, M. A., 599 Rouzina, I., 168 Rowe, T. C., 386 Roy, J., 579 Royer, C. A., 250 Rozek, D., 375,376(20) Rubin, C. M., 426

Rubin, E., 614 Ruland, C., 592 Russell, E E., 238 Russo, A. A., 581 Riiterjans, H., 262 Rutigliano, C., 29 Rutter, J. P., 580 Ryberg, H., 88 Ryberg, M., 88 Rychlik, W., 166 Ryder, U., 313,489

S Sabina, J., 444 Sacchi, N., 21 Sadat-Ebrahimi, S. E., 22(14), 23 Sadler, J. E., 144 Saecker, R. M., 168 Saenger, W., 150, 156(1) Sagi, J., 493 Sagui, C., 294 Saito, M., 574 Salazar, M., 138,226, 299, 323,580, 583,586, 589(85), 590, 592 Saleem, A., 614 Sambrook, J., 363,435,600, 608(14) Samper, E., 576, 577(29) Sanaterre, J. E, 203 Sanchez-Garcia, I., 594 Sannes-Lowery, K. A., 22(24), 23(24), 24, 25, 433 Sano, T., 582 Sansom, C. E., 271,274(78) SantaLucia, J., 170, 171(35), 172(35), 174(35), 175(35), 176, 176(35), 178(35), 329 SantaLucia, J., Jr., 168, 170(14), 174(14), 175(14), 176(14), 177(14), 178(14) Sapolsky, R. J., 166 Sargent, D. E, 503 Sarma, R. H., 311 Sarvazyan, A. E, 150 Sasi, R., 508 Sasisekharan, V., 143, 301,319(108; 109) Sasmor, D. J., 25 Sasmor, H., 25 Satyanarayana, S., 100(19), 101,532 Saucier, J. M., 623 Saudek. V., 269

680

AUTHOR INDEX

Sausville, E. A., 291 Savre-Train, I., 576 Sawadogo, M., 551 Scaria, P. V., 106, 143, 343,345(14) Scatchard, G., 54, 62(12), 194, 246 Schaeffer, F., 469(4), 470 Schafer, K. L., 582 Schaffer, T. E., 239, 251 Schaper, A., 248 Scheek, R. M., 270, 277(70) Scheit, K. H., 489 Schellman, J. A., 194 Scheraga, H. A., 167, 168(10), 169(10), 193 Scherf, U., 291 Schiano-Liberatore, A.-M., 620 Schild, C., 634 Schildkraut, I., 521 Schimmel, P. R., 459 Schipper, P. E., 79 Schlessinger, F. B., 509 Schmacher, R. J., 581 Schmid, C. W., 426 Schmidt, C. E, 240 Schmidt, M. W., 300, 302(99), 311(99), 318(99), 324(99) Schmitt, A., 591 Schmitt, L., 250 Schmitz, A., 412, 451 Schmitz, U., 257,258, 259, 259(26), 264(26), 265(26), 270, 270(26), 277(26) Schneider, B., 300, 307(105), 311(105) Schneider, C., 581 Schneider, H. J., 22, 24(5), 25(5) Schreiber, V., 591 Schroeder, R., 22(17; 18), 23,444 Schuitmaker, J. J., 589 Schulemann, W., 381 Schulten, K., 306 Schultz, P. G., 451,452(15; 16) Schultz, R. G., 493 Schultze, P., 139, 144, 278 Schuster, G. B., 242, 247(36), 338 Schwartz, A., 485 Schwartz, R. E., 388 Schwarz, U., 238 Scott, E. V., l 1, 14(40) Scourides, P. A., 535 Scratchard, G., 3, 6(1) Scudiero, D. A., 291

Searle, M. S., 216, 253,254(4), 268, 276(67), 277(4; 67), 297 Sedivy, J. M., 576 Seela, E, 426 Segal, D. J., 594 Seger, D., 503,508(6), 516(6) Sehlstedt, U., 24, 25(41), 83 Sekiguchi, J., 138 Selsing, E., 319 Semenza, G., 240, 241(27) Sen, S., 296 Senear, D. F., 451,453(18-21) Seneviratne, P., 170, 171(35), 172(35), 174(35), 175(35), 176(35), 178(35), 329 Senter, J. P., 153 Sergeyev, V. G., 203, 206(15) Sethson, I., 257 Severin, S. E., 578,580(55) Seymour, C. K., 242 Shafer, R. H., 23(30), 24, 106, 132, 138, 143, 271,291,299, 343,345(14) Shahin, M. A., 9, 10(30) Shanck, B. S., 151 Shao, Z., 337 Shapiro, H., 119 Shapiro, R., 403,404 Sharma, A., 586, 619 Sharma, S., 586 Sharmeen, L., 22(24), 23(24), 24, 433 Sharon, R., 293 Sharp, K., 300 Sharp, S. L., 238 Sharpies, D., 212, 254 Shastry, B. S., 552 Shaw, B. R., 508,516(22) Shay, J. W., 137,575,576, 580, 581, 581(32; 33), 582(32; 33) Shchyolkina, A. K., 489 Shea, M. A., 451,453(18-20) Shea, R. G., 144 Sheardy, R. D., 138, 141(91), 519, 520, 521(12), 527(12), 528(12), 589 Sheehan, M. A., 638 Sheng, S., 337 Sheppard, R. C., 558 Sherer, E., 296 Shevchenko, A., 574 Shevchenko, Y., 489 Shi, Y. B., 485 Shida, T., 138

AUTHORINDEX Shieh, T.-L., 380, 389 Shields, G. C., 280, 296, 298, 298(62), 314(62), 317 Shimizu, M., 131,340 Shimotakahara, S., 404 Shindyalov, I. N., 300, 307(104), 324(104) Shippen-Lentz, D., 574 Shirahama, K., 203 Shirahata, A., 349 Shire, S. J., 143 Shlyaktenko, L. S., 240, 241,241(24), 249 Shokolenko, I., 336 Short, G. F., 617,619(30) Shubsda, M. E, 436 Shui, X., 156, 226, 284, 285(12), 287,546, 549(30), 568 Shultze, P., 314, 315(143), 317 Shuman, S., 619 Shure, M., 54 Siebenlist, U., 312 Siegel, J. S., 626 Sigal, L. H., 349 Sigman, D. A., 519 Sigurdsson, S. T., 400, 404, 408,519 Silverman, R. H., 578, 580(52) Silverstone, A. E., 469(3), 470 Simmen, K. A., 551 Simmerling, C., 306 Simmons, C. G., 578,580(56) Simpson, I. J., 4, 29, 50(58) Sinden, R. R., 349 Sines, C. C., 226 Singh, S. B., 298 Sirr, A., 281 Sittman, D. B., 643,646(18), 647(18), 651 Siva, A. C., 592 Sivolob, A., 510 Sj6berg, B., 88 Skalka, A. M., 624 Skerrett, J. N. H., 14 Skerrett, N. J. H., 378 Sklenar, H., 300, 308(102), 317,319(102) Sklenar, V., 269 Skorobogaty, A., 473,535 Sleeman, A. M., 638 Slickers, E, 313 Sligar, S. G., 239 Smerdon, M. J., 485 Smimov, I., 291 Smith, B. L., 238,250, 251,518

681

Smith C. E., 389 Smith C. I. E., 337 Smith E W., 139, 144, 253,278(5) Smith G. J. I., 626 Smith L. H., 291,359 Smith P. E., 294(49-51), 295,298(49; 50) Smith P. J. C., 319 Smith R., 579 Smith S., 591 Smo orzewska, A., 573 Snow, A. M., 519 Snyder, R. C., 131 Sobell, H. M., 389 Sober, H. A., 194 Soda, H., 586 Soliva, R., 296 Solomon, M. S., 398, 401(7), 402 Song, Y., 209 S6nnichsen, S. H., 338 Sorey, S., 568,570(29) Soto, A. M., 157 Souhami, R. L., 358,364(3), 375(3) Soyfer, V. N., 110, 117(15), 118(15) Spackova, N., 299 Spellmeyer, D. C., 293,296(31), 304(31), 305(31), 307(31), 308(31), 312(31), 319(31) Spencer, W. J., 166 Spink, C. H., 193,203,204, 204(14), 205,207, 207(14) Spisz, T. S., 242 Sponer, J., 299 Sprankle, K. G., 333 Springer, D. L., 485 Sproat, B. S., 313,489 Sprous, D., 294, 295 Squire, C. J., 222,230, 233(29) Srinivasan, A. R., 24(42), 25,300, 307(105), 311 (105), 352 Srinivasan, J., 296 Stahl, M., 301 Stankus, A. A., 10(33), 11,432 Stansel, R. M., 584, 585(87) Star, W. M., 589 Starink, J. P., 239 Stark, W., 22(25), 23(25; 32), 24, 433 States, D. J., 293 Stebbins, C. E., 581 Stec, W., 259 Stecenko, D., 489

682

AUTHOR INDEX

Stefano, J. E., 471 Steiner, R, 578 Stenglanz, R., 648 Stenhag, K., 29 Stephenson, R, 485 Stem, M. A., 519 Stem, S., 22(20), 23(20), 24, 433 Stevens, M. E, 271,274(78), 297 Stevens, R. M. D., 238 Stewart, L., 610 Stewart, N. G., 573, 574(6) Stewart, R. J., 240 Stewart, S. A., 577 Stick, R., 645 Still, W. C., 294, 300, 303(98), 304(98), 307(98), 311 (98), 319(98), 320(98) Stillinger, E H., 151 Stivers, J. T., 284, 285(13) Stone, B. A., 410 Stone, M. E, 18, 106 Stonehouse, T. J., 418 Stowell, J. G., 380, 389 Strahan, G. D., 299 Strahl, C., 578 Straney, D. C., 471 Straub, T., 612,623(14) Straume, M., 8, 114, 115(23), 127(23) Strauss, U. E, 77 Strekowski, L., 132, 345,349, 353,354 Streuli, M., 592 Strumberg, D., 610 Strzelecka, T., 521 Stucky, G. D., 238 Studier, E W., 648 Sturm, J., 156 Sturm, R. A., 473 Sturtevant, J. M., 126, 127, 135(41) Styles, J. M., 371 Su, S., 300, 302(99), 311 (99), 318(99), 324(99) Suck, D., 416, 499 Suda, H., 614 Sugawara, K., 247 Sugimoto, N., 170, 175(36), 176(36), 178(36) Suh, D., 11, 18, 234, 532, 546, 557 Sullivan, W. E, 581 Sulston, I., 313 Sun, D., 138, 141(87), 299, 573,578,582, 583, 586, 589, 589(85), 592

Sun, J. S., 110, 131,340, 341,344, 345,346, 347(23; 25), 349, 350, 351(25), 354, 354(19), 356, 356(18; 19; 45), 504, 509(13), 516(13), 517(13), 518(13) Sun, Y. C., 106 Sunder, S., 633 Suri, A. K., 23(31), 24 Svejstrup, J. Q., 616, 619, 619(26) Sverdlov, E. D., 249, 489 Svinarchuk, F., 248 Swalley, S. E., 450 Swaminathan, S., 144, 293,324 Sweet, R. M., 282 Sykes, B. D., 257 Szabo, A. G., 261,345 Szakonyi, E., 493 Szewczyk, J. W., 450, 529(7), 530, 556 Szoka, E C., 203 Szostak, J. W., 291 Szyperski, T., 281

T Taatjes, D. J., 408 Tabore|li, M., 241 Tagashira, Y., 170 Tahara, K., 557 Taillandier, E., 489 Taira, K., 23(27), 24 Takac, L,, 238 Takagi, M., 557 Takahara, P. M., 86, 164 Takahashi, M., 88 Takashima, K., 203 Takasugi, M., 355 Takenaka, S., 557 Takeyasu, K., 249 Takisawa, N., 203 Takusagawa, E, 23(34), 24, 390(88; 89), 391,495 Takusagawa, K. T., 23(34), 24, 390(89), 391 Tamayo, J., 249 Tamura, A., 126, 135(41) Tan, R. K. Z., 294 Tanabe, L., 291 Tanaka, S., 247 Tanford, C., 150 Tang, C. L., 240 Tang, D., 54

683

AUTHORINDEX Tang, K. S., 500 Tanious, E A., 4, 29, 50(58), 132, 133, 136(74), 345,558,586 Tanious, T. C., 4 Tanizawa, A., 617,620 Tanner, J. E., 263,269(54) Tarlov, M. J., 29 Tate, S., 262 Taylor, J. B., 554(39), 555 Taylor, J.-S., 451,452(15), 485 Taylor, M. J., 401(7), 402 Teat, S. J., 301 Tempczyk, A., 294 Teng, M.-K., 212 Terasawa, S., 154 Terreux, R., 433 Tesmer, V. M., 581 Teulade-Fichou, M.-E, 344, 356(18) Tewey, K. M., 386 Thayaparan, J., 298, 304(90) Thomale, J., 359 Thomas, J. O., 503,507(5) Thomas, S. D., 291,312, 314, 315(142), 316(142) Thomas, T. J., 349, 354, 577 Thommes, E, 651 Thompson, A. M., 573(12), 574 Thompson, B., 299, 583,589(85) Thompson, J. B., 251 Thomsen, T., 556, 557(11) Thomson, N. H., 237,250 Thomson, S. A., 335,580 Thrall, B. D., 485 Thulsprup, E. W., 69(4), 70 Thundat, T., 235(3), 236, 238,240, 249 Thuong, N. T., 130, 131,281,340, 343,346, 354, 392 Thurston, D. E., 485 Tilby, M. J., 358,371 Tilby, N. J., 252 Tinoco, I. T., Jr., 55 Tirado, M. M., 263 Tironi, I. G., 294 Tjerneld, E, 79, 86 Tjian, R., 552, 553(36) Toft, D. O., 581 Tokumasu, F., 249 Tolman, J. R., 280 Tomassetti, M., 29 Tomasz, M., 404, 500

Tomchick, R., 381 Tondelli, L., 21 Toney, J. H., 519 Tong, G. L., 535 Tong, W. E, 406 Tor, Y., 25 Torda, T. E., 270, 277(70) Tordova, M., 284, 285(13) Tomaletti, S, 364, 365(17), 372(17) Toyama, K., 582 Trager, J. B., 581 Tran, H., 156 Tran-Patterson, R., 312 Trant, I. O., 537 Trauger, J. W., 429,450, 529 Travers, A. A., 413,415(2), 494, 495,496(39), 499(33), 502, 503,510(7), 511(7) Trent, J. O., 19, 131,132(67), 139, 141, i41(94), 142(94), 290, 291,296, 298,299, 303(65), 304(90), 308,313, 320(65; 97), 324(97), 529,589 Trinh, L., 579,582(60) Trist, H., 469, 471,535 Tsai, M. J., 131 Tu, Y., 364, 365(17), 372(17) Tuite, E., 347,517 Tulinski, A., 144 Tullius, T. D., 399, 422,424,513,519 Tunis, M.-J. B., 150 Turner, D. H., 444 Turner, J. M., 450, 529(7), 530, 556 Turro, N., 379 Twigden, S., 556, 557(10) Tzfati, Y., 579

U Uedaira, H., 150 Uemura, D., 614 Ueng, L.-M., 619 Uesugi, S., 23(27), 24 Ughetto, G., 283,309, 311(t26), 320(126), 419, 531,533(18), 535(18) Uhlen, M., 29 Uhlenbeck, O. C., 34 Ulibarri, G., 620 Ulyanov, N. B., 259, 270 Urasaki, Y., 611 Usman, N., 212

684

AUTHOR INDEX

V Valenti, M., 617 Vallberg, H., 24, 25(41) Valle, M., 249 Vallone, P. M., 165, 166, 167(8), 168, 168(8), 170(8; 15), 171(8; 15), 172(8; 15), 173(8), 175(8), 176(8), 178(8) Valpuesta, J. M., 249 van Boom, J. H., 22(16), 23,212,258,260(27), 283,419, 519 van der Marel, G. A., 22(16), 23,212, 258, 260(27), 283,419, 519 van der Meulen, R. W., 589 VanDerveer, D., 226 vander Vliet, P. C., 291 Vanderwerf, K. O., 239 Van Dyke, M. W., 303,341 VanDyke, M. W., 349 Van Dyke, M. W., 419, 424, 451,452(13) van Gunsteren, W. E, 270, 277(70), 292(28), 293,294 vanHolde, K. E., 100 van Holde, K. E., 508,509, 509(20), 515(20), 516(20; 22), 518(28) Vanhulst, N. E, 239 Vankatachalam, S., 359 Van Landschoot, A., 228(28), 229 van Leengoed, H. L., 589 van Rosmalen, A., 408,479, 481 van Steensel, B., 573 van Wijk, J., 257 Vanyushin, B. E, 486 Vaquero, A., 291,529, 551,551(5), 554(5; 33), 555(33) Varani, G., 22 Varghese, A. J., 402 Vary, C. P., 473 Vasquez, K. M., 291 Vaziri, H., 576, 577,577(36) Veal, J. M., 580 Veiro, D., 404 Verboom, W., 397 Vercauteren, J., 313 Verdine, G. L., 404 Vergani, L., 516 Veselkov, A. G., 334 Veselovskaya, S. I., 489 Vesenka, J., 239, 240 Vezin, H., 614

Viale, R. D., 230 Viani, M. B., 251 Vicens, Q., 22, 24(4) Vickers, T. A., 336 Vigevani, A., 21 Vigneron, J.-P., 344, 356(18) Vigneswaran, N., 291,298,304(90), 312, 355 Vileponteau, B., 574 Villamarfn, S., 551,554(33), 555(33) Villeponteau, B., 573,574 Vinayak, R., 22(22), 23(22), 24 Vinograd, J., 54 Volenshtein, M. V., 194 V61ker, J., 168, 177(17) Volkovitsky, P., 337 Vologodskii, A. V., 170 von Ahsen, U., 22(18), 23 von Hippel, P. H., 6, 7(24), 15(24), 54, 64(15), 65(15), 77, 194, 204(9), 209(9), 245, 246(41), 247(41), 250 Von Hoff, D. D., 138,578,579, 582, 583,586, 589, 589(85), 592 von Kitzing, E., 494, 499(33) von Sprecken, R. S., 379(31), 380 Vorlickova, M., 493 Vosberg, H.-P., 54 Vournakis, J. N., 473

W Wadkins, R. M., 159, 380, 385(33), 386(38), 394 Wadwani, S., 464 Wagar, E. A., 165 Wagner, C. R., 495 Wagner, P., 240, 241(27) W~ihnert, U., 212, 271,509 Wakasa, M., 30 Wakelin, L., 556, 557(9; 12) Waldsich, C., 444 Walker, G. T., 18, 106 Walker, V., 404 Wall, M. E., 610, 617 Walter, F., 22, 24(4) Waiters, A., 4, 5(20) Waiters, D. A., 237,239 Waiters, P., 301 Waltham, M., 291

AUTHOR INDEX Wang, A. H.-J., 212, 283,378, 408,419, 519, 531,532, 533(18), 535(18), 539, 540(24), 549(24), 550(24) Wang, C., 117, 142(29), 143(29), 146(29), 256 Wang, G., 336 Wang, H., 576 Wang, J. C., 52, 54, 248, 303,386, 408,576, 610 Wang, K. Y., 144, 280, 324 Wang, L. K., 26, 29(51), 617,619(31) Wang, P. U., 106 Wang, S., 22(23), 23(23), 24, 25(23) Wang, S. D., 106 Wang, S. S., 574 Wang, S. Y., 402 Wang, X., 617,619(30; 31; 36) Wang, Y., 139, 167, 168(11), 169(11), 301 Wang, Z., 238 Wani, A. A., 359 Wani, M. C., 610, 617 Wank, H., 444 Ward, B., 119, 356, 393,418, 429, 432, 446, 446(3) Ward, G. K., 156, 157(26) Ward, J., 568,570(29) Waring, M. J., 10(34), 11, 12(35), 109,346, 347, 347(25), 351(25), 378, 380, 381,391, 391(91), 392, 392(92), 393(92; 93), 412, 413,417,418,419(13), 421,421(13), 422, 426, 427,429(3; 13), 432, 452, 485,486, 489,489(24), 490, 493,493(26), 494, 495, 495(24), 496, 496(24; 39), 499, 499(41), 500, 500(30), 502, 504, 509(13), 510, 511(9), 512, 514, 514(9; 10), 515(9; 10), 516(13), 517(13), 518(13), 531,555,612 Warmack, R. J., 235(3), 236, 238,249 Warshaw, M. M., 119 Wartell, R. M., 167, 168(11), 169(11), 170, 172(31), 202 Waterman, M. S., 181 Watkins, C. L., 379, 382(22; 53; 100), 383, 383(29), 384(29), 386(22), 394 Watson, D. G., 301 Watson, D. K., 27 Watson, J. D., 573 Watson, J. V., 651,652(35) Watson, T., 117 Wattez, N., 612 Watts-Tobin, R. J., 377 Weaver, R. E, 23(34), 24, 390(89), 391 Webb, A., 579

685

Weber, G., 3, 6(6) Webster, C. I., 29 Weerasinghe, S., 294(49-51), 295,298(49; 50) Wei, S. L., 106 Wei, W., 576 Weidner, M. E, 400, 519 Weimer, J. J., 241 Weinberg, R. A., 574, 576, 577,578 Weinblum, D., 402, 403(11) Weinrich, S. L., 137, 574, 576, 581(32), 582(32) Weinstein, J. N., 291 Weissig, H., 300, 307(104), 324(104) Weisz, K., 271 Wellman, S. E., 193,209, 210 Wells, R. D., 131,202, 340, 389 Wells, T. N., 241 Wemmer, D. E., 266, 291,450 Wen, L., 390(89), 391 Wendman, M. A., 237 Wenzler, L. A., 240, 241(25) West, M. D., 137, 574, 576, 581(32), 582(32) Westbrook, J., 111,283, 300, 307(104; 105), 311(105), 324(104) Westergaard, O., 616, 619, 619(26) Westhof, E., 22, 22(19), 23(19), 24, 24(4), 150 Westhof, W., 22 Weston, S. A., 416 Westphal, H., 648 Wetmur, J. G., 165,166(5) Wharton, G. III, 520, 521(12), 527(12), 528(12) Wheelhouse, R. T., 138, 141(87), 586, 589 White, J. G., 650 White, J. H., 51(3), 52, 230, 233(29) White, M. A., 581 White, R. J., 390(90), 391,469, 469(5), 470, 471,473(5), 479, 482(10), 483(10; 13) White, S., 291,450, 529(7), 530, 556 White, W. E., Jr., 379, 384, 394 Wiech, H., 581 Wijmenga, S. S., 257,258, 260(27) Wilcock, D., 650 Wiles, N. C., 18, 106 Wilke, 1. G., 312 Wilkosz, P., 296 Willenbring, H., 582 Williams, A. R., 150 Williams, D. H., 29 Williams, H. E. L., 268,276(67), 277(67), 297

686

AUTHORINDEX

Williams, L. D., 156, 226, 234, 242, 243(34), 247, 247(34-36), 282, 284, 285(12), 286(16), 287, 289, 296, 546, 549(30), 568 Williams, R. M., 27 Williamsand, J. M., 240, 241(25) Williamson, J. R., 117, 138, 147,281 Williston, S., 115, 121(26) Willmott, G., 556, 557(12) Wilson, J. H., 291 Wilson, K. L., 634, 645 Wilson, W. D., 4, l 1, 14(39; 40), 22, 22(21; 22), 23(21; 22), 24, 24(5), 25(5; 35), 26, 29, 29(51), 42, 43(76), 50(58), 132, 133, 136(74), 345,349, 354, 377,504, 508(12), 517(12), 558,586 Wilson, W. R., 380 Wilton, A. N., 22(14; 15), 23 Wiltshire, T., 242 Windle, B. E., 138,578,586, 589 Windus, T. L., 300, 302(99), 311(99), 318(99), 324(99) Wiorkiewicz-Kuczera, J., 293 Wiseman, T., 115, 121(26) Wishart, D. S., 257 Wittes, R., 388 Wittung, E, 92, 329, 331,332 Wofsy, C., 30, 48(67) Wohlrab, E, 131,340 Wold, B., 358 Wolfe, A. L., 633 Wolffe, A. E, 634, 641,644 Wong, A. H., 309, 311(126), 320(126) Wong, C.-H., 26, 29(48-50), 31(50), 36(48), 38(50), 433 Wong, L., 297 Wong, S. S., 238 Woo, J., 400, 519 Woo, R. J., 234, 286(16), 289 Wood, A. A., 141,299, 589 Wood, R. D., 252 Woody, R. W., 69 Woolley, A. T., 238 Workman, J. L., 503, 518(2) Wright, G. E., 489 Wright, W. E., 137,576, 580, 581,581(32), 582(32) Wu, C. B., 106 Wu, E, 582 Wu, H. M., 509, 515(24) Wu, H.-W., 65,468

Wu, J., 518, 611 Wu, R. S., 589 Wu, S., 579,582(60) Wurtz, N. R., 291 Wiithrich, K., 266, 267,281 Wyatt, J. R., 33(70), 34, 118,435 Wyman, J., 3, 6(3), 16

X Xiao, G., 22(21), 23(21), 24, 284, 285(13) Xiaoxin, X., 336 Xie, J. X., 106 Xie, L. Y., 576 Ximen, H. Y., 238 Xu, H., 634 Xu, J., 518 Xu, Q., 519 Xu, Y., 203,238 Xu, Z., 24(42), 25,352 Xue, B.-H., 577

Y Yalowich, J., 386 Yamada, A., 614 Yamamoto, R., 23(27), 24 Yamasaki, E. E, 359 Yanagi, K., 226 Yang, D., 262, 281 (44) Yang, G., 250 Yang, J.-C., 266 Yang, L., 294(51), 295 Yao, J., 257 Yao, S., 132, 315,345 Yao, X., 238 Yavnilovich, M. V., 334, 338 Yen, S., 558 Yeung, D. S. K., 574 Yielding, K. L., 379, 379(31), 380, 382(22; 100), 384, 385,386(22), 392(61), 394 Yielding, L. W., 379, 379(31), 380, 382(22; 53; 100), 383,383(29), 384, 384(29), 385,386(22), 394 Yin, S., 485 Yip, R. W., 345 Yokoyama, T., 58 ! Yoneyama, M., 170, 175(36), 176(36), 178(36)

AUTHOR INDEX Yoo, S., 248 York, D. M., 294 York, S. G., 577 Yoshida, E., 614 Yoshikawa, K., 203,206(15) Yoshinari, T., 614, 622(23) Yoshizawa, S., 22(12), 23, 23(29), 24 Young, M. A., 295,296, 296(52; 53), 297(52; 53; 71), 385 Young, P. R., 377 Yu, J., 574 Yu, L., 518 Yui, J., 577

Z Zacharias, W., 298,304(90) Zahler, A. M., 138 Zakian, V. A., 573 Zanatta, N., 262 Zandwijk, N., 589 Zaniewski, E., 504 Zapp, M. L., 22(20; 22), 23(20; 22), 24, 433 Zardecki, C., 283 Zasedatelev, A. S., 194 Zaugg, F., 240, 241(27) Zawadzki, V., 648

Zechel, A., 4, 5(21), 212,216(10), 217 Zegrocka, O., 354 Zelphati, O., 337 Zeman, S. M., 51 Zendegui, J. G , 131 Zhang, J., 306 Zhang, L., 238 Zhang, X., 283 Zhang, X.-Y., 486 Zheng, X.-Y., 238 Zhou, E X., 234, 286(16), 289 Zhou, G., 500 Zhou, H., 387,388(74) Zhou, K., 22(26), 23(26), 24 Zhou, W., 574 Zhu, J., 238,576 Zhu, L., 258 Zhu, N., 203 Zhu, X., 250 Ziaugra, L., 578 Zichi, D. A., 294(48), 295 Zieba, K., 151 Zimmer, C., 24(43), 25,202, 212,271 Zimmer, D. P., 262 Zimmerman, R., 581 Zimnik, O. V., 578,580(55) Zon, G., 11, 14(40), 132, 257,345 Zuker, M., 444 Zwelling, L. A., 388

687

Subject Index

A Acridine DNA intercalation, 379-380 photoreactive azides, 383 4'-(9-Acridinylamino)methanesulfon-manisidide, topoisomerase II interactions azido analog stimulation of DNA strand breaks, 388-389 inhibition, 388 Actinomycin D DNA binding properties, 389-391 kinetics of DNA interactions, 390-391 photoreactive analog, s e e 7-Azidoactinomycin D Actinomycin D DNA binding modes and volume changes, 159-160 nucleosome-drug interactions, 515 Adjacency matrix, s e e DNA duplex stability Adriamycin chemotherapy, 531 energetics of DNA binding, 532-533, 535 formaldehyde-mediated coupling to DNA assay gas chromatography-mass spectrometry, 410--412 N-methyl adriamycin standard synthesis, 409-410 denaturing polyacrylamide gel electrophoresis, 408-409, 412 overview, 407-408 structure of lesion, 408 functional domains, 531 modular drug design using anthracycline-based building blocks, s e e a l s o WP631 combinatorial approach, 539, 541,544 overview, 530, 535-536 rational structure-based design, 536-537, 539

site specificity for DNA, 531 transcription assays of DNA binding, 479, 483 9-Aminoacridine, historical perspective of use, 377-378,381 Antigene, s e e DNA triplex Ascididemin, topoisomerase I inhibition, 613 Association constant equation, 112 noncalorimetric determination for nucleic acid-ligand interactions, 125-126 Atomic force microscopy, s e e Scanning force microscopy 7-Azidoactinomycin D alkali lability of DNA adducts, 392 biophysical properties, 389-390 criteria for ideal photoaffinity analog, 395 DNA binding properties, 389-391 DNA photoreactivity, 391-392 photolysis conditions, 394-395 shuffling hypothesis of binding, 391-392 specificity of binding, 392-393

B Berenil, nuclear magnetic resonance analysis of minor groove binding, 272, 274 BIAcore, s e e Surface plasmon resonance BN80915, topoisomerase I inhibition, 620-622

C Camptothecin, topoisomerase I inhibition, 610, 622 CD, s e e Circular dichroism Cetyltrimethylammonium bromide, McGhee model of DNA melting with ligand, 206 Chromatin~zlrug interactions binding isotherm generation equilibrium dialysis anthracycline antibiotics, 509 dialysis conditions, 508-509 ethidium bromide, 509-510 fluorescence titration, 508

689

690

SUBJECT INDEX

chromatin preparations chicken blood, 504-506 integrity checking, 508 uniform length preparation, 507 circular dichroism analysis, 51 6-517 DNA base distribution in grooves, 503-504 footprinting of reconstituted nucleosomes actinomycin D binding, 515 DNase I footprinting, 511,513-515 gel electrophoresis, 513-514 hydroxyl radical footprinting, 513, 515 netropsin binding, 51 4-515 radiolabeling of DNA, 510 reconstitution with radiolabeled DNA, 510-511 McGhee model of DNA melting with linker histone binding, 209-210 nucleosome purification, 506 prospects for study, 518 sedimentation velocity analysis, 515-516 structure histones, 503 nucleosome, 503 Xenopus sperm chromatin remodeling assays in egg extracts applications, 652-653 DNA replication assays deoxynucleotide incorporation, 650-652 overview, 648 radiolabeled deoxyribonucleotide incorporation, 650 electrophorettic analysis of proteins, 643-644 incubation conditions for drug testing, 641, 643 micrococcal nuclease digestion analysis, 644-645 nuclear assembly assays nuclear lamina assembly, 645,647 nuclear membrane assembly, 645 overview, 645 protein import assay, 648 overview, 640-641 Circular dichroism absorption of polarized light, 69 chromatin~lrug interaction analysis, 516-517 DNA~trug interactions drug dimerization and DNA ineractions, 82 exciton spectra, 97-98

induced signal and structural information, 93, 95-96 modes of binding and spectra interpretation calculations, 82 groove binding, 77, 81 intercalation, 76-77, 79, 81 phosphate binding, 76 signal origin, 69, 75-76 peptide nucleic acid formation, 90, 92 principles, 75-76 RNA-drug interactions, 24-25 triplex DNA~lrug interactions, 346 Cisplatin DNA binding assay, see Polymerase chain reaction, covalent DNA~lrug interaction assay mechanism, 358 volume change studies of DNA adducts, 164-165 Closed circular DNA~trug interactions binding constant and unwinding determination advantages of method, 65~56 calibration of micropipettes, 61 data analysis, 63-64 daunomycin as intercalator, 57-59 ethidium bromide as intercalator, 58, 60, 64 plasmids, 57 principles, 56 reaction conditions, 58-61, 65-66 troubleshooting faint bands, 66-67 resolution of gel, 67 topoisomerase I reaction and incomplete relaxation, 67-68 two-dimensional gel electrophoresis analysis casting, 61-62 image analysis, 62-63 running conditions, 62 staining, 62 linking difference, 51-52, 54, 63-64, 66 nearest neighbor exclusion model, 54-55 topoisomerase I relaxation reactions, 52-53 twist change on ligand removal, 52 unwinding angle, 51-52 ligand binding measurement equations, 54-56 writhe change on ligand removal, 52, 54

691

SUBJECT INDEX Cobalt, DNA binding conformational transition induction, 52O covalent modification, 520 restriction enzyme assay for site determination advantages, 528 agarose gel electrophoresis, 525 AseI linearization, 523-524 BamHI cleavage, 524, 527 binding reaction, 523 DraI leavage, 524, 527 EcoRV cleavage, 524-525,527 endonuclease concentration determination, 525-526 inhibition assessment, 528-528 selection, 522 HindlII cleavage, 524, 527 materials, 522-523 overview, 521 plasmid selection, 522 Pvull cleavage, 524-525,527 site specificity, 520-521 Competition dialysis assay, nucleic acid ligands advantages, 99 applications, 107-108 development, 100-101 dialysis conditions, 104-105 units, 102 interpretation of results binding constant determination, 107 ethidium as intercalator, 105-106 graphical representations, 106-107 nucleic acids concentration determination, 103 preparation, 102-103 quality control, 103-104 principles, 100-102 triplex DNA-drug interactions, 347-348 Cross-linking, DNA interstrand cross-link assays denaturing polyacrylamide gel electrophoresis, 398-399 digestion and chemical composition determination acid hydrolysis, 402-403 controlled depurination, 405-406 enzymatic digestion, 403-405 ethidium bromide assay calculations, 397-398

incubation conditions, 397 principles, 397-397 global structure determination, 406-407 hydroxyl radical footprinting for localization, 399-401 piperidine cleavage at cross-linked residues, 401-402 intrastrand cross-link analysis, 407 monoadduct analysis, 407 types of cross-links, 396 Crothers allosteric model, cooperativity, 16-19, 194

D Daunomycin-DNA interactions chemotherapy, 531 closed circular DNA interactions, 57-59 Crothers allosteric model for cooperativity, 16-19 dinucleotide binding model assumptions, 14-15 binding ratio equation, 12 development, 11-12 free energy equation, 13 unique dinucleotide steps, 12 validation, 12-14 DNA-binding site, 109 energetics of DNA binding, 532-533,535 free energy of ligation, 6 functional domains, 53 I Job plot, 4-5 5-methylcytosine substitution effects on binding, 500 modular drug design using anthracycline-based building blocks, see also WP631 combinatorial approach, 539,541, 544 overview, 530, 535-536 rational structure-based design, 536-537, 539 molecular modeling of (-)-daunorubicin intercalation B-DNA, 309-310 heating and equilibration, 312 hydration, 312 left-handed DNA preference, 308-309 starting model, 311 Z-DNA, 309-312

692

SUBJECT INDEX

neighbor exclusion models assumptions, 6-7 closed form equations, 7-8, 10 cooperativity inclusion, 15-16, 18-19 exclusion parameter, 8-9 fitting software, 8 Friedman-Manning model, 9-10 limitations, 10-11 McGhee-von Hippel model, 6-9 oligonucleotide binding studies, 20--21 Scatchard plot, 6-7 site specificity, 531 titration binding isotherm, 5-6 Diaminopurine, DNA substitution synthesis chemical synthesis, 489 nucleosides, 488-489 polymerase chain reaction, 489-491,502 purification, 491,493 melting temperatures of substituted duplexes, 493-494 protein-binding studies, 494-495,500 minor groove-binding drug studies, 495-496, 499 Diethylpyrocarbonate footprinting, s e e DNA footprinting Differential scanning calorimetry cooperativity analysis, 127-128 high-order DNA melting studies data acqusition, 129-130 data analysis, 130 DNA tetraplex stabilization by cations data acquisition, 144 melting transitions, 146-147 potassium versus sodium, 144-148 sample preparation, 144 thermodynamic parameters, 144-146 thrombin-binding aptamer, 144 sample preparation, 129 instrumentation, 127-129 thermodynamic parameter determination, 127-128 WP631 binding to DNA, 547-549 Dinucleotide binding model assumptions, 14-15 binding ratio equation, 12 development, 11-12 free energy equation, 13 unique dinucleotide steps, 12 validation, 12-14

Dissociation constant equation, 112 noncalorimetric determination for nucleic acid-ligand interactions, 125-126 DNA-binding proteins, s e e Zinc finger engineering DNA cross-linking, s e e Cross-linking, DNA DNA~trug interactions, s e e a l s o s p e c i f i c d r u g s binding modes, 76-77, 234, 466, 529 calorimetry, s e e Differential scanning calorimetry; Isothermal titration calorimetry chromatin, s e e Chromatin-drug interactions closed circular DNA unwinding, s e e Closed circular DNA~lrug interactions competition dialysis, s e e Competition dialysis assay, nucleic acid ligands cross-linking, s e e Cross-linking, DNA drug design criteria, 110-111 McGhee model for melting studies, s e e McGhee model, DNA melting with ligands modified DNA base probing base types, 488,502 groove structure, 486, 500 historical perspective, 485-486 melting temperatures of substituted duplexes, 493-494 minor groove-binding drugs with inosine or diaminopurine substitution, 495-496, 499 protein-binding studies, 494-495,500 synthesis containing inosine or diaminopurine chemical synthesis, 489 nucleosides, 488-489 polymerase chain reaction, 489-491, 502 purification, 491,493 uridine and 5-methylcytosine residues daunomycin binding effects, 500 DNase I sensitivity, 499-500 structural effects, 499-500 synthesis, 499 modular design using anthracycline-based building blocks, s e e a l s o WP631 combinatorial approach, 539, 541,544 overview, 530, 535-536 rational structure-based design, 536--537, 539

SUBJECT INDEX molecular modeling, s e e Molecular modeling, DNA--drug interactions nuclear magnetic resonance, s e e Nuclear magnetic resonance polyintercalators, s e e a l s o 1,4,5,8-Naphthalenetetracarboxylic diimide amino acid design considerations, 556-559 rationale for synthesis, 556 polymerase chain reaction assay, s e e Polymerase chain reaction, covalent DNA~lrug interaction assay scanning force microscopy, s e e Scanning force microscopy spectroscopy, s e e Circular dichroism; Linear dichroism stopped-flow studies, s e e Stopped-flow fluorescence, Hoechst 33258-DNA interactions telomeres, s e e Telomerase; Telomere therapeutic targeting, 109-110, 252, 290-291, 529, 556 transcription assays, s e e Transcription assay, DNA-clrug interactions triplex DNA, s e e DNA triplex volume change, s e e Volume change, nucleic acid-ligand interactions X-ray crystallography, s e e X-ray crystallography, DNA-drug interactions DNA duplex stability graph theory block sequence analysis, 190-192 D N A sequence representation adjacency matrix, 181- 183 Eulerian graph and trails, 184-186 submatrices, 183-184, 186 Eulerian trail finding, 186-187 isothermal sequence number finding decomposition of adjacency matrix, 187-188 different sequence restrictions, 188-189 mathematical nomenclature, 185-186 nearest neighbor model doublet format, 173-175 enthalpy change, 176-180 entropy change, 177 free energy change, 177-180 sequence-dependent end interactions, 171-172 singlet format, 172-175

693

sodium dependence correction, 176-177 stacking interactions, 169-170 transition temperature calculation, 175-176, 178-180 unique interactions, 171-172 rationale for study, 165-166 transition temperature calculation, 170, 175-176, 178-180 isothermal sequences, 166, 181 two-state melting theory dissociation constant, 167 enthalpy of transition, 168 entropy of transition, 168 external degrees of freedom, 168-169 internal degrees of freedom, 167-169 nucleation parameter, 168-169 DNA footprinting diethylpyrocarbonate footprinting echinomycin binding study, 427, 429 incubation conditions, 427 reaction specificity, 427 differential cleavage plots, 429 DNase I footprinting advantages and limitations, 418-4 19 cation dependence, 415 cleavage site, 416 Hoechst 33258 binding study, 419 incubation conditions for drug binding, 416-417 reaction conditions, 417-418 substrate specificity, 415-4 16, 452 WP631 effects on Spl binding, 555 DNase II footprinting, 419, 421 ethidium azide binding sites, 385 fragment selection and preparation, 413-4 15 hydroxyl radical footprinting Fenton chemistry, 399,422 Hoechst 33258 binding study, 424 incubation conditions, 423 interstrand cross-link localization, 399-401 limitations, 423-424 minor groove drug binding studies, 422-423 specificity of cleavage, 422 methidiumpropyl-EDTA-Fe(lI) footprinting cleavage specificity, 452-453 incubation conditions, 425 triplex DNA, 424

694

SUBJECT INDEX

micrococcal nuclease footprinting reaction conditions, 421-422 substrate specificity, 421 nucleosome~trug interactions actinomycin D binding, 515 DNase I footprinting, 511,513-515 gel electrophoresis, 513-514 hydroxyl radical footprinting, 513,515 netropsin binding, 514-515 radiolabeling of DNA, 510 reconstitution with radiolabeled DNA, 510-511 osmium tetroxide footprinting, 426 potassium permanganate footprinting, 425-426 principles, 412-4 13 probe comparison, 413 pyrrole-imidazole polyamide complexes, s e e Pyrrole-imidazole polyamides quantitatie analysis, 429-430 radiolabeling of DNA, 415 triplex DNA, 349, 353 DNA quadruplex differential scanning calorimetry of stabilization by cations data acquisition, 144 melting transitions, 146-147 potassium versus sodium, 144-148 sample preparation, 144 thermodynamic parameters, 144-146 thrombin-binding aptamer, 144 isothermal titration calorimetry sample preparation annealing, 118 buffers, 117 concentration determination, 118-119 melting behavior characterization, 119-120 purity, 117 stabilization by cationic porphyrin buffer conditions, 139 counterion effects, 142 oligonucleotides, 139 stoichiometry of binding, 141-142 thermodynamic parameters, 139-141 molecular modeling ligand specificity, 299 porphyrin intercalation 1:1 complexes, 321-323 2:1 complex, 323

3:1 complex, 322 hydration, 324, 326 starting models, 324 stoichiometry, 320-32 l simulations, 299 starting model generation, 301 spectroscopic studies of cationic stabilization 143 structure, 139, 142-143 telomerase inhibition, 137-138 telomere G-quadruplex targeting anthraquinone analogs, 586, 589 perylene, 590 PIPER, 590 porphyrins, 589 DNase footprinting, s e e DNA footprinting; Pyrrole-imidazole polyamides DNA triplex antigene approach applications, 312-313 improvement with stabilizers, 354 limitations, 313 therapeutic targeting, 130-132 base triples, 341-343 cleaving agent design and uses, 355-356 drug binding studies binding mode studies, 352 buffer conditions, 351-352 calorimetry, 348 circular dichroism, 346 competition dialysis, 347-348 fluorescence spectroscopy, 345-346 footprinting, 349, 353 gel-shift assays, 348-349 linear dichroism, 346-347 nuclear magnetic resonance, 347,352 rational drug design, 353-354 RNA-DNA hybrids, 352-353 sequence selection, 351 surface plasmon resonance, 350-351 thermal denaturation, 343-345,353 types and structures of drugs, 343 viscometric titration, 347 H-DNA stabilization by binding agents, 355 structure, 340-341 isothermal titration calorimetry sample preparation annealing, 118 buffers, 117

SUBJECT INDEX concentration determination, 118-119 melting behavior characterization, 119-120 purity, 117 triplex, 117-118 stabilization by intercalating drugs data acquisition, 134 drug types and specificities, 132-134. 137, 348 thermodynamic parameters. 135-136 titration curves, 134 molecular modeling antigene approach c - m y c P2 promoter targeting, 312, 314-316, 318 hydration, 318-319 RNA as third strand, 313-314 starting model, 318-319 ligand specificity, 298-299 naphtholflavone intercalation, 319-320 peptide nucleic acid complexes, 298 rational drug design, 349-350 starting model generation, 301 nuclear magnetic resonance of inducing oligonucleotides, 278,280 stability with RNA, 313-314 stabilization by proteins and drugs, 341,354 structure, 131-132 Doxorubicin, s e e Adriamycin DSC, s e e Differential scanning calorimetry

E Echinomycin, diethylpyrocarbonate footprinting, 427,429 E c o R I , scanning force microscopy of DNA interactions, 249 Elinafide. inosine or diaminopurine substitution effects on DNA binding, 496. 499 ELISA, s e e Enzyme-linked immunosorbent assay Enthalpy binding enthalpy determination, 3 calorimetric measurement of change, 113-114 nearest neighbor model, 176-180 two-state melting theory. 168 van't Hoffenthalpy, 126-128, 152 Entropy nearest neighbor model, 177 two-state melting theory, 168

695

Enzymeqinked immunosorbent assay, phage assay for zinc finger engineering applications, 595,604 binding conditions and analysis, 605 DNA target preparation, 604-4505 phage preparation, 604 scanning mutagenesis binding assays. 605, 607 Ethidium azides alkali lability of DNA adducts, 385,392 biophysical properties, 383 criteria for ideal photoaffinity analog, 395 DNA binding characteristics of monoazide, 379, 384, 386 footprinting studies and binding sites, 385 photolysis conditions, 394-395 photoreactive analog structures, 382 topoisomerase II-mediated DNA strand cleavage probing mechanism studies, 387-388 stimulation of cleavage reaction, 386-387 Ethidium bromide chromatin~lrug interactions, 509-510 competition dialysis assay. 105-106 DNA cross-linking assay calculations, 397-398 incubation conditions, 397 principles. 397-397 historical perspective of intercalation studies, 380-381 medical use, 38l unwinding assay, s e e Closed circular DNA-drug interactions Eulerian graph, s e e DNA duplex stability

F Factor for inversion stimulation, modified DNA base-binding studies, 494-495 FIS, s e e Factor for inversion stimulation Footprinting, s e e DNA footprinting; RNA footprinting Friedman-Manning model, neighbor exclusion model, 9-10

G Gal repressor, scanning force microscopy of DNA interactions, 249-250

696

SUBJECT INDEX

Gibbs free energy calculation of change, 3, 113 dinucleotide binding model, 13 ligation, 6 nearest neighbor model, 177-180 noncalorimetric determination for nucleic acid-ligand interactions, 125-126

H H-DNA, s e e DNA triplex Heat capacity, calculation of change, 113 Hexaammine cobalt(II1), McGhee model of DNA melting with ligand, 208 Hexadecyltrimethylammonium bromide, McGhee model of DNA melting with ligand, 203-207 High-mobility group proteins, modified DNA base-binding studies, 495 Histone, s e e Chromatin~lrug interactions HIV, s e e Human immunodeficiency virus Hoechst 33258 DNA binding fluorescence titration, 213, 231 mode, 212 stopped-flow studies, s e e Stopped-flow fluorescence, Hoechst 33258-DNA interactions structural studies, 212, 214 footprinting studies DNase I footprinting, 419 hydroxyl radical footprinting, 424 Human immunodeficiency virus integrase, s e e Integrase, HIV-1 packaging region RNA footprinting autoradiogram analysis, 439, 441-442 gel electrophoresis, 437, 439 kinetics of cleavage, 446 paromomycin binding, 433,435,444, 446-449 plot analysis of binding sites, 446-449 RNase I cleavage, 435,437,446, 448 RNase T1 cleavage, 446, 449 RNA synthesis, 435-437 structure, 442-444 therapeutic targeting, 433 pyrrole-imidazole polyamide targeting, 461 structure and derivatives, 212-213

Hydroxyl radical footprinting, footprinting

see

DNA

I Inosine, DNA substitution melting temperatures of substituted duplexes, 493-494 minor groove-binding drug studies, 495-496, 499 protein-binding studies, 494-495,500 synthesis chemical synthesis, 489 nucleosides, 488-489 polymerase chain reaction, 489-491,502 purification, 491,493 Integrase, HIV-1 assays disntegration reaction, 631 DNA-binding assays, 631,633 integration assay, 629 nucleophile selection for phosphodiester cleavage, 63 l oligonucleotide radiolabeling 3~-end, 628 5~-end, 628 reaction conditions, 628-629 strand transfer assay, 630-631 domains, 624 function, 624 recombinant protein purification from Escherichia

coli

amplification, 627 cell harvesting and lysis, 627 nickel affinity chromatography, 627~528 soluble mutant, 626 storage, 628 transformation, 626-627 therapeutic targeting, 624, 626, 633 Isothermal sequence, s e e DNA duplex stability Isothermal titration calorimetry binding constant range limitations, 122 instrumentation, 115-116 nucleic acid--drug interaction analysis advantages, 114 controls, 124 c value, 121 DNA sample preparation annealing, 118 buffers, 117

697

SUBJECT INDEX concentration determination, 118-119 melting behavior characterization, 119-120 purity, 117 tetraplex, 117-118 triplex, 117-118 excess-site method, 124-125 experimental variables, 121-123 interval between injections, 123 ligand concentration determination, 122-123 sample preparation, 120-121 temperature selection, 123 tetraplex DNA stabilization by cationic porphyrin buffer conditions, 139 counterion effects, 142 oligonucleotides, 139 stoichiometry of binding, 141-142 thermodynamic parameters, 139-141 trial experiments, 122 triplex DNA stabilization by intercalating drugs data acquisition, 134 drug types and specificities, 132-134, 137,348 thermodynamic parameters, 135-136 titration curves, 134 ITC, s e e Isothermal titration calorimetry

,J Job plot, daunomycin-DNA interactions, 4-5

K Ku, scanning force microscopy of DNA interactions, 248-249

L Linear dichroism absorption of polarized light, 69 DNA~lrug interactions applications, 68, 82, 92-93 binding rearrangement monitoring, 83, 86 cyanine dye binding, 86, 88 discrimination between intercalation and groove binding, 82-83

drug dimerization and DNA interactions, 82, 97-98 electric transition moment determination, 73-75, 93-95 exciton spectra, 97-98 modes of binding and spectra interpretation, 76-79 orientation of DNA electric field, 73 gel electrophoresis, 73 hydrodynamic flow, 70, 72 polyvinyl alcohol film, 72-73 site-directed linear dichroism of RecA, 88, 90 peptide nucleic acid formation, 90 principles, 70 signal origin, 68-70 triplex DNA-drug interactions, 346--347

M Magnesium, electrostatic interactions with DNA, 156-157 Magnetic suspension densimetry, volume change measurement, 153-154 Mass spectrometry N-methyl adriamycin, gas chromatography-mass spectrometry, 410-412 1,4,5,8-naphthalenetetracarboxylic diimide amino acid, 567 RNA~lrug interaction screening, 25 McGhee model, DNA melting with ligands binding free energy equation, 198 examples cationic lipid ligands applications, 203 cetyltrimethylammonium bromide, 206 hexadecyltrimethylammonium bromide, 203-207 hexaammine cobalt(Ill) as ligand, 208 linker histone binding, 209-210 netropsin and poly[d(AT)]melting, 202-203 helical base pair averages, 197 historical perspective, 193-195 Lifson model foundation, 196-197 ligand binding to both helix and coil, 201 melting temperature binding constant effects, 199 cooperativity effects, 199

698

SUBJECT INDEX

free ligand activity effects, 199 ligand size effects, 198-199 parameter prerequisites and limitations, 201-202, 211 program availability, 211 features, 197-198 titrations with constant total ligand concentration, 199-201 McGhee-von Hippel model cooperativity inclusion, 15-16, 18-19, 194, 247 neighbor exclusion model, 6-9 Melting temperature, s e e DNA duplex stability; McGhee model Methidiumpropyl-EDTA-Fe(ll) footprinting, s e e DNA footprinting; Pyrrole-imidazole polyamides 5-Methylcytosine, D N A substitution daunomycin binding effects, 500 DNase I sensitivity, 499-500 structural effects, 499-500 synthesis, 499 Micrococcal nuclease footprinting, s e e DNA footprinting Molecular modeling, DNA~trug interactions computing power, 290, 292 duplex DNA B-DNA simulation, 294-295 conformational transitions, 295-296 (-)-daunorubicin intercalation B-DNA, 309-310 heating and equilibration, 312 hydration, 312 left-handed DNA preference, 308-309 starting model, 311 Z-DNA, 309-312 hydration, 296-297 ion interactions, 297 peptide nucleic acid complexes, 296 starting model generation, 300-301 thermodynamics of ligand binding, 297-298 evaluation of simulations, 305-306, 308 force field development, 293-294 modular drug design using anthracycline-based building blocks, 537,539 molecular dynamics simulation, 304-305 molecular simulation protocol

equilibrium phase, 307 evaluation, 308 production phase, 307 starting model generation, 306-307 overview, 292 particle mesh Ewald methods, 290, 293-294 quadruplex DNA ligand specificity, 299 porphyrin intercalation 1:1 complexes, 321-323 2:1 complex, 323 3:1 complex, 322 hydration, 324, 326 starting models, 324 stoichiometry, 320-321 simulations, 299 starting model generation, 301 software, 300 starting model generation DNA-intercalator complex, 303-304 DNA-ligand complex, 302-303 ligands, 301-302 minor groove-binding ligands, 304 therapeutic application, 290-292, 312-313 triplex DNA antigene approach c - m y c P2 promoter targeting, 312, 314-316, 318 hydration, 318-319 RNA as third strand, 313-314 starting model, 318-319 ligand specificity, 298-299 naphtholflavone intercalation, 319-320 peptide nucleic acid complexes, 298 starting model generation, 301 validation, 293 visualization, 306 MS, s e e Mass spectrometry

N

1,4,5,8-Naphthalenetetracarboxylic diimide amino acid characterization high-performance liquid chromatography, 566 mass spectrometry, 567 ultraviolet-visible spectroscopy, 566-567 library synthesis, 565-566 polyintercalator design, 558-559

SUBJECT INDEX prospects binding affinity improvement, 568,570 sequence specificity, 570 threading versus nonthreading intercalation, 570 synthesis N-(2-tert-butoxycarbonylaminoethyl)-N'-

(2-carboxyethyl)- 1,4,5,8naphthalenetetracarboxylic diimide, 560 N-2-(Na-9-fluorenylmethoxycarbonyl-N etert-butoxycarbonyl)lysylaminoethylN'-(2-carboxyethyl)- l,4,5,8-

naphthalenetetracarboxylic diimide, 560-561 glycine adduct workup, 561 lysine adduct workup, 561 overview, 559-560 reversed-phase high-performance liquid chromatography, 564 solid-phase synthesis cleavage from resin and side chain deprotection, 562-564 coupling/deprotection cycle, 561-562 NB-506, topoisomerase I inhibition, 614, 622 NDI, see 1,4,5,8-Naphthalenetetracarboxylic diimide amino acid Nearest neighbor model, see DNA duplex stability Neighbor exclusion model assumptions, 6-7 closed circular DNA~lrug interactions, 54-55 closed form equations, 7-8, 10, 54 cooperativity inclusion, 15-16, 18-19 exclusion parameter, 8-9 fitting software, 8 Friedman-Manning model, 9-10 limitations, 10-11 McGhee-von Hippel model, 6-9 oligonucleotide binding studies, 20-21 Netropsin inosine or diaminopurine substitution effects on DNA binding, 496 McGhee model of DNA melting with ligand, 202-203 nucleosome-drug interactions, 514-515 volume changes on binding to model bent sequences, 161-162 NMR, see Nuclear magnetic resonance

699

Nogalamycin, nuclear magnetic resonance of DNA intercalation, 276-278 Nuclear magnetic resonance DNA~lrug interactions bisamidinium compound binding to minor groove berenil, 272,274 hydration analysis, 275-276 pentamidine, 275 propamidine, 272-275 structures, 271 titrations, 271 calculation strategies, 270-271 chemical shifts, 253-256 coupling constants cross-correlation effects, 257-258 Karplus equation, 257 nuclear Overhauser effect error function, 259 sugar conformations, 257-260 field strength, 280 hydration studies nuclear Overhauser effect spectroscopy, 266-268 principles, 265-266 rotating-frame Overhauser enhancement spectroscopy, 266-268 hydrogen bond scalar coupling, 281 instrumentation, 268-270 isotopic labeling, 281 line widths, 255 nogalamycin intercalation, 276-278 nuclear Overhauser effect spectroscopy, 264-265 overview, 252-253 rate constants, 254-256 relaxation times and dynamics 13 value, 260-261 carbon- 13,263 chemical shift anisotropy, 261-263 equations, 260 proton exchange, 263-264 resonance assignment, 253 sensitivity, 269-270 solvent suppression, 269 standards for referencing, 256-257 titrations, 254, 256, 268-269 triplex-forming oligonucleotides, 278,280 tubes, 269 WP631 binding, 549-550

700

SUBJECTINDEX

RNA-drug interactions, 23-24 triplex DNA~lrug interactions, 347,352 Nucleation parameter, two-state melting theory, 168-169 Nucleosome, s e e Chromatin~lrug interactions

O Osmium tetroxide footprinting, footprinting

see

DNA

P PARG, s e e Poly(ADP-ribose)glycohydrolase Paromomycin, s e e Human immunodeficiency virus, packaging region Particle mesh Ewald methods, molecular modeling, 290, 293-294 PCR, s e e Polymerase chain reaction Pentamidine, nuclear magnetic resonance analysis of minor groove binding, 275 Peptide nucleic acid assays for double-stranded DNA binding, 337-338 circular dichroism studies, 90, 92 design for double-stranded DNA binding, 338-339 DNA-binding modes double duplex invasion, 334-335 duplex invasion, 334 overview, 330-331 triplex binding in major groove, 331 triplex invasion, 331-334 invasion complex applications DNA-binding protein inhibition, 336 molecular biology, 337 transcriptional activation, 336 transcriptional arrest, 335-336 linear dichroism studies, 90 molecular modeling duplex DNA complexes, 296 triplex DNA complexes, 298 stability of duplex with DNA, 329-330 structure, 329-330 Perylene, telomere G-quadruplex targeting, 590 Phage display, s e e Zinc finger engineering Photoaffinity labeling of DNA, s e e 4 t -(9-Acridinylamino)methanesulfon-manisidide; 7-Azidoactinomycin D; Ethidium azides

Platinum, DNA modification sites, 519 PNA, s e e Peptide nucleic acid Poly(ADP-ribose)glycohydrolase, inhibition in telomere targeting, 591-592 Polymerase chain reaction modified base DNA synthesis, 489-491,502 radiolabeling of DNA, 4 5 5 Polymerase chain reaction, covalent DNA~lrug interaction assay antibody purification of DNA fragments binding reaction, 374 DNA preparation, 374 overview, 359, 371 protein A-Sepharose beads capture, 375 treatment, 374-375 sensitivity, 371,374 capture of biotinylated products, 366-367 damaging agent treatment cells monolayers, 362-363 suspensions, 362 isolated genomic DNA, 361-362 DNA isolation and digestion, 363-364 first-round amplification, 366 gel electrophoresis of products, 370-37 l ligation of double-stranded linker, 368 ligation-mediated polymerase chain reaction, 358 oligonucleotides end-labeling for third round, 369 melting temperature calculation, 365 phosphorylation, 365-366 sequences, 364-365 optimization, 364 overview of steps, 359-361 polymerase, 364 second-round amplication, 368-369 single-strand ligation polymerase chain reaction, 358-359, 364 terminal deoxynucleotidyltransferase ribo-tailing, 367-368 terminal transferase-dependent polymerase chain reaction, 359, 364 third-round amplication, 369-370 Potassium permanganate footprinting, s e e DNA footprinting Propamidine, nuclear magnetic resonance analysis of minor groove binding, 272-275 Pyrrole-imidazole polyamides

SUBJECT INDEX cell permeability, 450 design, 451 DNA binding specificity, 450 footprinting of DNA complexes association constant determination, 453, 460 DNase I footprinting binding reaction, 457 digestion, 457458 fractional occupancy calculation, 459 gel electrophoresis, 458-459 materials, 453-454 methidiumpropyl-EDTA-Fe(II) footprinting data analysis, 461 ImPy-[3-lmPy--y-ImPy-13-ImPy-[3-Dp characterization, 461,463-464 reaction conditions, 460 overview of probes, 451-453 radiolabeling of DNA gel purification, 456 polymerase chain reaction, 455 restriction digest and fill-in, 454--455 serial dilution of polyamides, 456-457

Q Quadruplex,

see

DNA quadruplex

R RecA, site-directed linear dichroism, 88, 90 Rev-responsive element solid-phase assay for ligand screening, 25-26 surface plasmon resonance conrol, 38-39 RNA~lrug interactions circular dichroism, 24-25 gel shift assays, 25 mass spectrometry screening, 25 nuclear magnetic resonance, 23-24 solid-phase assay for Rev-responsive element ligands, 25-26 surface plasmon resonance, s e e Surface plasmon resonance therapeutic prospects, 22-23 volume change, s e e Volume change, nucleic acid-ligand interactions RNA footprinting binding constant determination, 431432 human immunodeficiency virus packaging region

701

autoradiogram analysis, 439, 441-442 gel electrophoresis, 437,439 kinetics of cleavage, 446 paromomycin binding, 433,435,444, 446-449 plot analysis of binding sites, 446-449 RNase I cleavage, 435,437,446, 448 RNase T1 cleavage, 446, 449 RNA synthesis, 435-437 structure of RNA fragment, 442-444 therapeutic targeting, 433 plot analysis, 431 4 3 3 principles, 431 RNA polymerase DNA~lrug interaction assay, s e e Transcription assay, DNA~lrug interactions scanning force microscopy of DNA interactions, 250-251 RRE, s e e Rev-responsive element

S Scanning force microscopy (SFM) advantages and limitations in nucleic acid imaging, 234-235 components of instrument, 237 drug-DNA complex imaging buffer conditions, 241 groove binding assay, 247-248 immobilization of DNA, 240-242 intercalation assay binding affinity determination, 245 contour length determination, 242-243 exclusion number determination, 244-245 occupied site determination, 245-246 sample preparation, 243-244 Scatchard modeling, 246 site determination, 247 substrates, 240-241 modes for imaging contact mode, 235 intermittent contact mode, 237 noncontact mode, 236 protein-DNA complex imaging E c o R l studies, 249 Gal repressor studies, 249-250 Ku studies, 248-249 prospects, 251

702

SUBJECT INDEX

RNA polymerase studies, 250-251 sample preparation, 248 Trp repressor studies, 250 scanner calibration, 239 tip calibration, 237-239 forces, 239-240 Scatchard plot curvature, s e e Dinucleotide binding model; Neighbor exclusion model daunomycin-DNA interactions, 6-7 DNA ligand limitations, 194 Sedimentation velocity analysis, chromatin~lrug interactions, 515-516 Site-directed mutagenesis, linear dichroism of RecA, 88, 90 SN38, topoisomerase I inhibition, 620-622 Spl, WP631 effects on transcription DNase I footprinting, 555 gel retardation, 554-555 transcription assay, 553-554 SPR, s e e Surface plasmon resonance Stopped-flow fluorescence, Hoechst 33258-DNA interactions association constants, 212, 21 6-217 association rates, 214, 216-217 (A/T)4 sequence binding base-pair roll effects, 226 equilibrium data comparison, 225 minor groove width effects, 226 pseudo-first-order kinetics, 223-225 sequences in study, 223 data analysis, 214-215 derivative binding to isolated AATT site association kinetics, 217, 219 dissociation kinetics, 219-221 effect of m-OH and bis-m-OH phenyl substitutions, 222-223,232 equilibrium data comparison, 221 dissociation rates, 214 DNA types in study, 216 instrumentation, 21 4-215 precautions with dye, 215-216 sample preparation, 215 steady-state titrations, 213, 231 two-step binding models biexponential association kinetics, 228, 233 bimolecular association followed by isomerization, 228-230, 233

data collection, 227-228 rationale, 226 triexponential association kinetics, 231 two single-step associations, 230-231 Surface plasmon resonance advantages over conventional interaction analysis, 26, 50 injection modes, 37-38 principles, 27 RNA~lrug interactions controls, 37-39 data collection program, 39-40 data processing, 40-41 equilibrium analysis, 43-46 experimental design, 27-30 immobilization biotinylated RNA on streptavidin-coated sensor, 34-36 materials, 35 storage of RNA sensor chips, 36--37 strategies, 28-30 kinetic constant determination, 41,45-49 refractive index increment ratios, 42-43 regeneration of surface, 38 resonance units maximum predicted response, 42-43 steady state response, 45-46 RNA biotinylation, 33-34 sample concentrations, 38 sensor chip, streptavidin immobilization coupling reaction, 33 materials, 31-32 overview, 30-31 preconcentration of streptavidin, 32-33 thermodynamic analysis, 50 triplex DNA~lrug interactions, 350-351

T Tankyrase, inhibition in telomere targeting, 591-592 Telomerase antisense oligonucleotide targeting, 579 assays biotinylated primer extension assay, 584 cell extract preparation, 583 primer extension assay, 582-583 telomeric repeat amplification protocol, 582-584

SUBJECT INDEX assembly inhibition, 581 cancer therapy targeting, 577 components, 574 DNA tetraplex inhibition, 137-138 immortalization of cell cultures, 576 knockout mouse phenotype, 575-576 mechanism, 574-575 regulation of expression, 577 reverse transcriptase inhibitors, 578-579 RNA component targeting, 579-580 RNA/DNA hybrid targeting, 580 tumor expression, 137,574-575, 577 Telomere functions, 573 shortening in aging, 573 tandem repeats, 573 targeting cancer therapy rationale, 577 D-loop, 584, 590 t-loop anthraquinone analogs, 586, 589 duplex DNA, 585-586 perylene, 590 PIPER, 590 porphyrins, 589 sequence, 584-585 single-sranded DNA, 586, 589-590 t-loop binding proteins assays, 592 overview, 584-585 poly(ADP-ribose)glycohydrolase, 591-592 tankyrase, 591-592 Tetraplex, s e e DNA quadruplex Topoisomerase I camptothecin inhibition, 610, 622 cleavage reaction cleavage complex and inhibitor targeting, 610 filter-binding assay, 623 inhibitors BN80915, 620-622 indolocarbazoles, 622 SN38, 620-622 religation reaction distinguishing from cleavage ligation reaction, 617,619 oligonucleotide substrates, 617

703

restriction fragment substrates, 619~520 suicide substrates, 619 synthetic oligonucleotides as assay substrates, 6164517 DNA relaxation gel electrophoresis assay, 612-613 inhibitors ascididemin, 613 glycosylated indolocarbazoles, 614 NB-506, 614, 622 reactions, 52-53, 612 inhibitor classes, 61(P4511 sources, 611 unwinding assay, s e e Closed circular DNA~trug interactions Topoisomerase II 4'-(9-acridinylamino)methanesulfon-manisidide azido analog stimulation of DNA strand breaks, 388-389 inhibition of enzyme, 388 cancer role, 385-386 ethidium azide probing mechanism studies, 387-388 stimulation of cleavage reaction, 386-387 intercalator inhibitors, 386 Transcription assay, DNA~lrug interactions Adriamycin binding, 479, 483 bidirectional transcription footprinting applications, 483,485 principles, 482-483 reaction conditions and digestion, 485 dissociation kinetics, 479, 481 drug binding to initiated transcripts, 473, 475 gel elecrophoresis autoradiography, 476-477 phosphorimaging, 479 separation of blocked transcripts, 475-476 principles, 467 promoters ideal characteristics, 467 l a c U V 5 promoter fragment isolation, 468-470 types, 469 sequencing reaction preparation, 472-473 synchronization of transcription initiation complexes, 470472

704

SUBJECT INDEX

WP631 transcription inhibition assay basal transcription assay, 552-553 cell extracts, 552 overview, 550-551 Sp l-activated transcription assay, 553-554 vectors, 551-552 Transition temperature, see DNA duplex stability; McGhee model Triplex, see DNA triplex Trp repressor, scanning force microscopy of DNA interactions, 250

U Uridine, DNA substitution DNase 1 sensitivity, 499-500 structural effects, 499-500 synthesis, 499

V van't Hoff equation, 126 Volume change, nucleic acid-ligand interactions actinomycin D studies, 159-160 apparent molar volume, 154-155 DNA-binding mode effects intercalation and minor groove binding, 159-160 intercalation, 157-158 magnesium electrostatic interactions with DNA, 156-157 minor groove binding, 158-159 overview, 155-156 duplex formation and water uptake, 151 equilibrium constant dependence on pressure, 152 hydration and nucleic acid structure, 150 measurement techniques dilatometry, 152-153 indirect measurement, 153 magnetic suspension densimetry, 153-154 oscillating method, 154 molar volume flexibility of water, 150-151 molecular interpretation, 154-155 netropsin binding to model bent sequences, 161-162 oligomer duplex formation with covalently attached ligands

benzo[a]pyrene diol epoxide adducts, 163-164 cisplatin adducts, 164-165 oligonucleotides for study, 163 RNA intercalation studies, 160-161

W WP631 DNA-binding properties differential scanning calorimetry, 547-549 melting temperature shift, 546 overview, 539, 568 ultraviolet melting studies, 547 viscosity studies, 546 Sp 1 binding effect studies DNase I footprinting, 555 gel retardation, 554-555 structural studies nuclear magnetic resonance, 549-550 X-ray crystallography, 549-550 synthesis, 544-545 transcription inhibition assay basal transcription assay, 552-553 cell extracts, 552 overview, 550-551 Spl-activated transcription assay, 553-554 vectors, 551-552 WP760, melanoma toxicity, 541, 544

X Xenopus egg extract

advantages of system, 634 development of egg, 634 frog colony establishmet and maintenance, 635 preparation activation, 637 centrifugation, 637~638 egg collection, 635,637~638 jelly coat removal, 637 yields, 638 Xenopus sperm chromatin remodeling assays applications, 652-653 DNA replication assays deoxynucleotide incorporation, 6 5 0 ~ 5 2

SUBJECT INDEX overview, 648 radiolabeled deoxyribonucleotide incorporation, 650 electrophorettic analysis of proteins, 643--644 incubation conditions for drug testing, 641, 643 micrococcal nuclease digestion analysis, 644-645 nuclear assembly assays nuclear lamina assembly, 645, 647 nuclear membrane assembly, 645 overview, 645 protein import assay, 648 overview, 640-641 Xenopus sperm chromatin remodeling studies, see Xenopus egg extract nuclei preparation, 639-640 X-ray crystallography, DNA-drug interactions anisotropic diffraction, 288-289 crystal growth precipitants, 282-283 protein crystallization differences, 283 screening parameters, 283-284 toroid analogy, 282 data collection, 284, 288 dynamic range of intensity data, 284-289 overview of steps, 282 structure determination molecular replacement, 289 multiple isomorphous replacement, 289

705

multiple wavelength anomalous diffraction, 289 WP631,549-550

Z Zinc finger engineering bipartite complementary approach advantages, 607 automation, 607 cesium chloride gradient purification of replicative form DNA, 608 cooperating contacts in design, 594-595 coselection of complementary libraries, 603-604 electrocompetent cell preparation, 608-609 library construction, 599-600 materials, 607-608 overview, 595,597 phage enzyme-linked immunosorbent assay applications, 595,604 binding conditions and analysis, 605 DNA target preparation, 604-605 phage preparation, 604 scanning mutagenesis binding assays, 605,607 phage yield estimation, 609 premade libraries, 595 recombination of complementary libraries, 602-603 selection, 601-602 selective polymerase chain reaction, 597, 599 overview, 593 phage display techniques, overview, 594

Drug-Nucleic Acid Interactions

Methods in Enzymology Vol 340: Drug-Nucleic Acid Interactions

Nucleic Acid Metal Ion Interactions (RSC Biomolecular Sciences)

Protein-Nucleic Acid Interactions: Structural Biology (RSC Biomolecular Sciences)

Acid Bath

Acid Bath

Chlorosulfonic Acid

Chlorosulfonic Acid

Acid Rock

Acid Rock

Acid Rock

Acid power!

Acid Row

Acid Rayne

Acid Rock

Acetylsalicylic Acid

Acid Rock

Acid Rock

ACID Power!

Acid Row

Acid Row

Interactions I

Fundamental Interactions

Amino Acid Analysis Protocols

The Boric Acid Murder

The Boric Acid Murder

New Nucleic Acid Techniques

Drug-Nucleic Acid Interactions

Methods in Enzymology Vol 340: Drug-Nucleic Acid Interactions

Nucleic Acid Metal Ion Interactions (RSC Biomolecular Sciences)

Protein-Nucleic Acid Interactions: Structural Biology (RSC Biomolecular Sciences)

Acid Bath

Acid Bath

Chlorosulfonic Acid

Chlorosulfonic Acid

Acid Rock

Acid Rock

Acid Rock

Acid power!

Acid Row

Acid Rayne

Acid Rock

Acetylsalicylic Acid

Acid Rock

Acid Rock

ACID Power!

Acid Row

Acid Row

Interactions I

Fundamental Interactions

Amino Acid Analysis Protocols

The Boric Acid Murder

The Boric Acid Murder

New Nucleic Acid Techniques

Recommend Documents