Biothiols, Part B: Glutathione and Thioredoxin: Thiols in Signal Transduction and Gene Regulation, Volume 252 (Methods in Enzymology)

Preface Biothiols participate in numerous cellular functions, such as biosynthetic pathways, detoxification by conjugation, and cell division. Studies on oxidative stress in recent years have amply documented the key role of thiols-more specifcally the thiol-disulfide status of the cell--in a wide array of biochemical and biological responses. Awareness of the great importance of biothiols in cellular oxidative injury has grown along with the recognition of free radicals in biological processes. The reactions of thiols with free radicals are not only of interest in free radical chemistry: the most abundant nonprotein thiol in the cell, glutathione, is essential for the detoxification of peroxides as cofactors of various selenium-dependent peroxidases. The high concentration of glutathione in cells clearly indicates its general importance in metabolic and oxidative detoxification processes. In many ways, glutathione may be considered the master antioxidant molecule, a phrase which the late Alton Meister, one of the pioneers in glutathione research and a contributor to this volume, used. Bolstering of glutathione by other thiols, both natural (such as a-lipoic acid) and synthetic (such as Ebselen and several other drugs), has been investigated as a therapeutic approach to the oxidative component of various pathologies. Moreover, the redox changes of several thiol-containing proteins may be involved in key regulatory steps of the enzyme as well as in cell proliferation. The contributions to Volumes 251 and 252 of Methods in Enzymology (Biothiols, Parts A and B) provide a comprehensive and detailed account of the methodology relating to the molecular mechanisms underlying the multiple functions of biothiols, with emphasis on their interaction at the biochemical and molecular biological levels in cellular reactions, with oxidants and other biological and clinical implications of thiols. The contributions in this volume (Part B) include methods on glutathione: distribution, biosynthesis, metabolism, and transport; signal transduction and gene expression; thioredoxin and glutaredoxin; and synthetic mimics of biological thiols and thiols inhibitors. In Part A (Volume 251) methods are included relating to thiyl radicals; chemical basis of thiol/disulfide measurements: monothiols: measurement in organs, cells, organelles, and body fluids; dithiols: oMipoic acid; and protein thiols and sulfides. Credit must be given to the experts in various specialized areas selected to provide state-of-the-art methodology. The topics and methods included in these volumes were chosen on the excellent advice of the volume advisors, Bob B. Buchanan, Enrique Cadenas, Carlos Gitler, Arne Holmgren, Alton Meister, and Helmut Sies, to whom I extend my thanks and most grateful appreciation. LESTER PACKER

xiii

C o n t r i b u t o r s to V o l u m e 2 5 2 Article numbers are in parentheses following the names of contributors. Affiliations listed are current,

MARIE-ODILECHRISTEN(32, 33, 34), Laboratoire de ThOrapeutiqueModerne, SOL VA Y Pharma, 95151 Suresnes Cedex, France

MONIQUE ADOLPHE (34), Laboratoire de Pharmacologie Cellulaire, Ecole Pratique des Hautes Etudes, 75006 Paris, France MICHAEL T. ANDERSON(17), Stanford Medical School, Beckman Center, Stanford, California 94305 FREDRIK ~SLUND (29), Medical Nobel Institute for Biochemistry, MBB, Karolinska Institutet, S-171 77 Stockholm, Sweden K. D. AUMANN(5), Gesellschaftfar Biotechnologische Forschung (GBF), Braunsehweig, Germany E L A I N E S. G . B A R D E S (16), Department of Molecular Cancer Biology, Duke University Medical Center, Durham, North Carolina 27710 ROBERTM. BELL (16), Department of Molecular Cancer Biology, Duke University Medical Center, Durham, North Carolina27710 ERNESTBEUTLER(8), Department of Molecular and Experimental Medicine, Scripps Research Institute, La Jolla, California 92037 MIKAEL BJORNSTEDT(21, 22), Department of Medical Biochemistry and Biophysics, The Medical Nobel Institute for Biochemistry, Karolinska lnstitutet, S-171 77 Stockholm, Sweden RoY A. BORCHARDT(16), Department of Molecular Cancer Biology, Duke University Medical Center, Durham, North Carolina 27710 R. BRIGELIUS-FLOHE(5), Deutsches Institut far Erniihrungsforschung, Postdam-Rehbriicke, Germany BoB B. BUCHANAN(24), Department of Plant Biology, University of California, Berkeley, Berkeley, California 94720 ZHEN-HAI CHEN (14), Department of Pharmacology and Nutrition, School of Medicine, University of Southern California, Los Angeles, California 90033

GEORGE L. DALE (8), Department of Medicine, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma 73190 PAULETTE DECO'Iq'IGNIES (25), Physiologie V~gOtale Mol~culaire, Universit~ de ParisSud, 91405 Orsay-Cedex, France WULF DROGE (26), Division of Immunochemistry, Deutsches Krebsforschungszentrum, D-69120 Heidelberg 1, Germany H. JANE DYSON (30), Department of Molecular Biology, The Scripps Research Institute, La Jolla, California 92037 L. FACKIR (33), Laboratoire de Chimie-Physique, Universitd R~ne Descartes, 75270 Paris Cedex 06, France L. FLOHE (5), Gesellschafi far Biotechnologische Forschung (GBF), Braunschweig, Germany

Department of Molecular Pharmacologyand Toxicology, University of Southern California, Los Angeles, California 90033 JAMESA. FUCHS(9), Department of Biochemistry, University of_Minnesota, St. Paul, Minnesota 55108 RICHARD S. GLASS (31), Department of" Chemistry, University of Arizona, Tucson, Arizona 85721 RAYUDU GOPALAKRISHNA(14), Department of Pharmacology and Nutrition, School of Medicine, University of Southern California, Los Angeles, California 90033 USHA GUNDIMEDA(14), Department of Pharmacology and Nutrition, School of Medicine, University of Southern California, Los Angeles, California 90033 HENRY JAY FORMAN (7),

ix

X

CONTRIBUTORS TO VOLUME 252

LEONARD A. HERZENBERO (17), Stanford Medical School, Beckman Center, Stanford, California 94305 TOD P. HOLLER (12), Parke-Davis Pharmaceutical, Division of Warner-Lambert Company, Ann Arbor, Michigan 48105 ARNE HOLMGREN (21, 22, 29), Department of Medical Biochemistry and Biophysics, The Medical Nobel Institute for Biochemistry, Karolinska lnstitutet, S-171 77 Stockholm, Sweden PAUL B. HOPKINS (12), Department of Chemistry, University of Washington, Seattle, Washington 98195 TOSHIYUKIHORI (36), Department of Biological Responses, Institute for Virus Research, Kyoto University, Kyoto, Japan STEVAN R. HUBBARD (13), Department of Biochemistry and Molecular Biophysics, Columbia University, New York, New York 10032 RONALD S. HUTCHISON (23), Department of Plant Biology, The Ohio State University, Columbus, Ohio 43210 EMMAMUELLE ISSAKIDIS (25), Physiologie V~g~tale Mol~culaire, Universitd de ParisSud, 91405 Orsay-Cedex, France ROLE D. ISSELS(11), Medizinische Klinik lII, Universitat MEinchen, D-81366 Munich, Germany; and Institut far Klinische Hi~matologie, GSF MEinchen, D-81377 Munich, Germany TAKEO IWAMOTO (7), Department of Biochemistry, Kansas State University, Manhattan, Kansas 66506 SATOSHI IWATA (36), Department of Biological Responses, Institute for Virus Research, Kyoto University, Kyoto, Japan JEAN-PIERRE JACQUOT (25), Physiologie V~g~tale Mol~culaire, Universit~ de Paris-Sud, 91405 Orsay-Cedex, France DEAN P. JONES (1), Department of Biochemistry, Emory University School of Medicine, Atlanta, Georgia 30322 DANIEL JORE (33), Laboratoire de ChimiePhysique, Universit~R~ne Descartes, 75270 Paris Cedex 06, France

BRIAN KETTERER (6), Department of Biochemistry, CRC Molecular Toxicology Research Group, University College and Middlesex School of Medicine, London W1P 6DB, United Kingdom KAROLY KOBREHEL (24), Laboratoire de Technologie des Cdrdales, INRA, 34060 Montpellier Cedex 01, France TAKAHITO KONDO (8), Department of Pathological Biochemistry, Atomic Disease Institute, Nagasaki Univeristy School of Medicine, Nagasaki, Japan SUSHIL KUMAR (22), Department of Medical Biochemistry and Biophysics, The Medical Nobel Institute for Biochemistry, Karolinska Institutet, S-171 77 Stockholm, Sweden LAWRENCE H. LASH (2), Department of Pharmacology, Wayne State University School of Medicine, Detroit, Michigan 48201 MARTINE LEMAIRE (25), Physiologie V~g~tale Mol~culaire, Universit~de Paris-Sud, 91405 Orsay-Cedex, France RuI-MING LIu (7), Departments of Molecular Pharmacology, Toxicology, Pathology, and Pediatrics, University of Southern California, Los Angeles, California 90033 TE-CHUNG LIU (20), Department of Nutrition, Case Western Reserve University, Cleveland, Ohio 44106 M. MAIORINO (5), Department of Biological Chemistry, University of Padua, Padua, Italy THOMAS MEIER (11), Institute far Klinische Hamatologie, GSF Mtinchen, D-81377 Munich, Germany ALTON MEISTER* (3), Department of Biochemistry, Cornell University Medical College, New York, New York 10021 DAVID J. MEYER (6), Department of Biochemistry, CRC Molecular Toxicology Research Group, University College and Middlesex School of Medicine, London WIP 6DB, United Kingdom MYROSLAWAMIGINIAC-MASLOW (25), Physiologie V~g~tale Mol~culaire, Universit~ de Paris-Sud, 91405 Orsay-Cedex, France *Deceased

CONTRIBUTORS TO VOLUME 2 5 2

xi

MASASHI MIZUNO (19), Department of Molec-

HEIKE SCHENK (26), Division of Immuno-

ular and Cell Biology, Universityof California, Berkeley, Berkeley, California 94720 DONALD R. ORT (23), Photosynthesis Research Unit, USDA/ARS, Department of Plant Biology, University of Illinois, Urbana, Illinois 61801

chemistry, Deutsches Krebsforschungszentrum, D-69120 Heidelberg, Germany D. SCHOMBUR6 (5), Gesellschaftfar Biotechnologische Forschung (GBF), Braunschweig, Germany KLAUS SCHULZE-OSTHOFF (26), Division of Immunochemistry, Deutsches Krebsforschungszentrum, D-69120 Heidelberg, Germany PETER SCHURMANN (28), Laboratoire de Biochimie Vegetale, Universit~ de Neuchdtel, CH-2000 Neuch~tel, Switzerland MICHAEL MING SHI (7), Departments of Molecular Pharmacology, Toxicology, Pathology, and Pediatrics, University of Southern California, Los Angeles, California 90033 HELMUT SIES (35), Heinrich-Heine-Universitilt, Instimt far Physiologische Chemie 1. D-40001 Diisseldorf Germany FRANK J. T. STAAL (17), Department of Immunology, Netherlands Cancer Institute, 1066 CX, Amsterdam, Netherlands' THRESSA C. STADTMAN (31), Laboratory of Biochemistry, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland 20892 YUICHIRO J. SUZUKI (18), Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, California94720 MARC THEVENIN (34), Laboratoire de Toxicologie, Facult( de Pharmacie, 75270 Paris Cedex 06, France F. URSINI (5), Department of Chemistry, University of Udine, 1-33100 Udine, Italy

LESTER PACKER (18, 19), Department of Mo-

lecular and Cell Biology, University of California, Berkeley, Berkeley, California 94720 MULCHAND V. PATEL (20), Department of

Biochemistry, State University of New York at Buffalo, Buffalo, New York 14214 ANDREW F. G. QUEST (16), Department of Molecular Cancer Biology, Duke University Medical Center, Durham, North Carolina 27710 PATR1CE RAT (34), Laboratoire de Pharma-

cologie Cellulaire, Ecole Pratique des Hautes Etudes, 75006 Paris, France DONALD J. REED (10), Department of Bio-

chemistry and Biophysics, Environmental Health Sciences Center, Oregon State University, Corvallis, Oregon 97331 T1MOTHY W. ROBINSON (7), Departments of Molecular Pharmacology, Toxicology, Pathology, and Pediatrics, Universityof Southern California, Los Angeles, California 90033 JOAQUfN ROMA (15), Department of Physiol-

ogy, Experimental Toxicology and Neurotoxicology Unit, School of Medicine, University of Valencia, 46010 Valencia, Spain FRANCISCO J. ROMERO (15), Department of

Physiology, Experimental Toxicology and Neurotoxicity Unit, School of Medicine, University of Valencia, 46010 Valencia, Spain A. ROVERI (5), Department of Biological

Chemistry, University of Padua, Padua, Italy MARJORIE RUSSEL (27), Laboratory of Genet-

ics, The Rockefeller University, New York, New York 10021 NORIHITO SATO (36), Department of Biologi-

cal Responses, Institute for Virus Research, Kyoto University, Kyoto, Japan

NATARAJ N. VETTAKKORUMAKANKAV (20),

Department of Biochemistry, State University of New York at Buffalo, Buffalo, New York 14214 JEAN-MICHEL WARNET (34), Laboratoire de Toxicologie, Facult~ de Pharmacie, 75270 Paris Cedex 06, France MICHAEL P. WASHBURN (4), Department of Biochemistry, Michigan State University, East Lansing, Michigan 48824 WILLIAM W. WELLS (4), Department of Biochemistry, Michigan State University, East Lansing, Michigan 48824

xii

C O N T R I B U T O R S TO V O L U M E

252

ELIZABETHWILLOTT(16), Department of Mo-

DIAN PENG XU (4), Department of Biochemis-

lecular Cancer Biology, Duke University Medical Center, Durham, North Carolina 27710 JOSHUAH. WONG (24), Department of Plant Biology, University of California, Berkeley, Berkeley, California 94720 WEN QIN XIE (16), Department of Molecular Cancer Biology, Duke University Medical Center, Durham, North Carolina 27710

try, Michigan State University, East Lansing, Michigan 48824 AKIRA YAMAUCHI(36), Department of Bio-

logical Responses, Institute for Virus Research, Kyoto University, Kyoto, Japan JUNJI YODOI (36), Department of Biological

Responses, Institute for Virus Research, Kyoto University, Kyoto 606-01, Japan

[ 1]

GLUTATHIONE

DISTRIBUTION

IN NATURAL

PRODUCTS

3

[I] G l u t a t h i o n e D i s t r i b u t i o n i n N a t u r a l P r o d u c t s : Absorption and Tissue Distribution

By DEAN P. JONES Introduction G l u t a t h i o n e ( G S H ) is p r e s e n t in m o s t p l a n t a n d a n i m a l tissues f r o m w h i c h t h e h u m a n d i e t is d e r i v e d . M u c h o f this G S H is b i o a v a i l a b l e b e c a u s e cells o f t h e g a s t r o i n t e s t i n a l t r a c t h a v e N a + - d e p e n d e n t a n d N a + - i n d e p e n d e n t u p t a k e m e c h a n i s m s , 1-3 a n d n e t a b s o r p t i o n o f G S H f r o m t h e l u m e n into t h e v a s c u l a r c i r c u l a t i o n o c c u r s in t h e s m a l l intestine. 4 In a d d i t i o n , G S H can f u n c t i o n as an a n t i o x i d a n t in t h e ingesta, c a n m a i n t a i n a s c o r b a t e in a red u c e d a n d f u n c t i o n a l f o r m , 5 c a n d i r e c t l y r e a c t with a n d i n a c t i v a t e toxic e l e c t r o p h i l e s in t h e diet, 6 a n d c a n b e b r o k e n d o w n to y i e l d cysteine. 7 S t u d i e s have shown that dietary GSH enhances metabolic clearance of dietary p e r o x i d i z e d lipids a n d d e c r e a s e s t h e i r n e t a b s o r p t i o n 8'9 a n d t h a t c o n s u m p tion o f f o o d s high in G S H c o n t e n t is a s s o c i a t e d w i t h a b o u t a 50% r e d u c tion in risk o f o r a l a n d p h a r y n g e a l cancer, w T h e r e f o r e , a n i m p r o v e d u n d e r s t a n d i n g o f t h e d i s t r i b u t i o n o f G S H in f o o d s a n d t h e f a c t o r s affecting G S H a b s o r p t i o n a n d d i s t r i b u t i o n c a n b e e x p e c t e d to p r o v i d e t h e basis for i m p r o v e d f o o d s e l e c t i o n a n d p r e p a r a t i o n to r e d u c e risk o f c h r o n i c diseases. 1M. K. Hunjan and D. F. Evered, Biochim. Biophys. Acta 815, 184 (1985). 2 L. J. Dahm, P. S. Samiec, J. W. Eley, E. W. Flagg, R. J. Coates, and D. P. Jones, in "Free Radicals: From Basic Science to Medicine" (G. Poli, E. Albano, and M. U. Dianzani, eds.), p. 506. Birkhaeuser, Basel, 1993. 3 M. T. Vincenzini, F. Favilli, and T. Iantomasi, Biochim Biophys, Acta 1113, 13 (1992). 4 T. M. Hagen and D. P. Jones, Am. J. Physiol. 252, G607 (1987). 5 B. S. Winkler, Biochim. Biophys. Acta 925, 258 (1987). P. S. Samiec, L. J. Dahm, E. W. Flagg, R. J. Coates, J. W. Eley, and D. P. Jones, in "Free Radicals and Antioxidants in Nutrition" (F. Corongiu, S. Banni, M. A. Dessi, and C. RiceEvans, eds.), p. 269. Richelieu Press, London, 1993. 7 N. Tateishi and Y. Sakamoto, in "Glutathione: Transport and Turnover in Mammals" (Y. Sakamoto, T. Higashi, and N. Tateishi, eds.), p. 13. Japan Scientific Societies Press, Tokyo, 1983. 8 D. P. Kowalski, R. M. Feeley, and D. P. Jones, J. Nutr. 120, 1115 (1990). 9 T. Y. Aw and M. W. Williams, Am. J. Physiol. 263, G665 (1992). m E. W. Flagg, R. J. Coates, D. P. Jones, T. E. Byers, R. S. Greenberg, G. Gridley, J. K. McLaughlin, W. J. Blot, M. Haber, S. Preston-Martin, J. B. Schoenberg, D. F. Austin, and J. F. Fraumeni, Am. J. Epidemiol. 139, 453 (1994).

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

4

GLUTATHIONE

[11

Glutathione Contents of Foods General Considerations

Glutathione is readily oxidized by food processing and preparation, and many foods contain reactive electrophilic compounds that can remove GSH even after food is ingested. 6 Thus, analytically precise measures of GSH in foods only provide an index of GSH delivered to gastrointestinal epithelia. In addition, oxidized forms of GSH, such as oxidized glutathione (GSSG), are also present in foods and are at least partially available as GSH because the intestine contains a GSSG reductase mechanism, tl Consequently, direct measures of GSH probably underestimate GSH availability; whereas, measures of total GSH ( G S H t ) following reduction of food extracts with dithiothreitol probably overestimates GSH that is bioavailable. Many factors affect food contents of nutrients, and established food analysis protocols, such as those of the Association of Official Analytical Chemists, 12 should be used whenever possible. As water content is highly variable in stored foods, dry weight or total nitrogen contents should be determined for all samples. Conditions of growth, seasonal and geographic conditions, plant variety, duration and conditions of storage and processing, and preparation are all expected to affect GSH content of foods, but no systematic studies of the effects of these factors on GSH content are available. The methods that we use to assay GSH contents of foods were developed to provide a database for epidemiological studies of dietary intake of GSH and risk of diseases in humans. 13 We based our food items on the National Cancer Institute's Health Habits and History Questionnaire ( H H H Q ) , 14'15 which includes 98 food items and groups that account for over 90% of calories, fiber, and 18 major nutrients in diets in the United S t a t e s . 14,16 For each individual food item or group, a specific food and method of preparation is selected based on those most commonly reported in the second National Health and Nutrition Survey, a population-based survey conducted in the United States between 1976 and 1980. 11 T. M. Hagen, G. T. Wierzbicka, B. B. Bowman, T. Y. Aw, and D. P. Jones, Am. Z PhysioL 259, G530 (1990). 12 S. Williams (ed.), "Official Methods of Analysis." (1141 pages). Association of Official Analytical Chemists, Arlington, Virginia, 1984. 13 D. P. Jones, R. J. Coates, E. W. Flagg, J. W. Eley, G. Block, R. S. Greenberg, E. W. Gunter, and B. Jackson, Nutr. Cancer 17, 57 (1992). 14 G. Block, C. M. Dresser, A. M. Hartman, and M. D. Carroll, Am. J. Epidemiol. 122,13 (1985). 15 G. Block, A. M. Hartman, C. M. Dresser, M. D. Carroll, J. Gannon, and L. Gardner, Am. J. Epidemiol. 124, 453 (1986). 16 G. Block, C. M. Dresser, A. M. Hartman, and M. D. Carroll, Am. J. EpidemioL 122, 27 (1985).

[ 1]

GLUTATHIONE DISTRIBUTION IN NATURAL PRODUCTS

5

Standard Protocol

For each food item, at least three different brand name products or three different samples (different sources) are purchased from major chain supermarkets, cafeterias, or restaurants. When samples are stored prior to assay, this is done under usual storage conditions, that is, in a refrigerator for perishable foods and in a freezer for frozen foods. Prior to sampling, foods are prepared according to common habits in the United States. Nonedible portions are removed, and foods that are normally cooked are prepared accordingly. In general, steamed or boiled foods are cooked in a small amount of water until tender and then drained. Pan-fried foods, such as pork chops, are fried at medium heat for 5 min or until done (loss of pink or red color in meats). Because GSH can be rapidly lost on homogenization, we developed a standard protocol that provides a correction for GSH loss based on simultaneously assayed samples with known amounts of GSH added. Five 1-g portions are taken from one sample of each food item and treated as follows. One portion is placed in an aluminum tray and dried for at least 24 hr at 105° to obtain a dry weight. The other four portions are added to beakers containing 10 ml of ice-cold homogenization solution A, B, C, or D. Solution A is 0.9% (w/v) NaC1 containing 5 mM EDTA. Solutions B, C, and D are made from solution A by adding 10 ~M GSH, 20/zM GSH, or 5 mM dithiothreitol, respectively. Foods that are solutions or homogenates are mixed on a vortex mixer prior to removal of 1-ml aliquots for addition to test tubes containing 0.5 ml of ice-cold 30% (w/v) trichloroacetic acid. Emulsions (e.g., milk) are extracted 17 to remove lipids before acidification. For other foods, portions are homogenized using a Polytron for a minimum duration to obtain a homogeneous suspension, and 1-ml of aliquots are immediately transferred to ice-cold 30% trichloroacetic acid. Following centrifugation to remove insoluble material, 0.5-ml aliquots are transferred to microcentrifuge tubes containing 0.1 ml of 40 mM iodoacetic acid. 18Each solution is adjusted to pH 8 by addition of i M KHCO3 in I M KOH and incubated for 1 hr at room temperature. One-half milliliter of 1-fluoro-2,4-dinitrobenzene (1.5% in ethanol; v/v) is then added, and samples are incubated overnight at room temperatures in the dark. a8 Samples are then stored at 0o-4 ° for up to 2 weeks prior to assay by highperformance liquid chromatography (HPLC). The HPLC procedure is essentially that described by Reed et al. ~8 using 17 j. Folch, M. Lees, and G. H. Sloane Stanley, J. BioL Chem. 226, 497 (1957). 18 D. J. Reed, J. R. Babson, P. W. Beatty, A. E. Brodie, W. W. Ellis, and D. W. Potter, Anal. Bioehem. 106, 55 (1980).

6

GLUTATHIONE

[11

100

80

60

$ 20

I

0

i

I

10

I

I

20

I

i

30

I

40

i

I

50

Time (rain) FIG. 1. Analysis of GSH in potato by HPLC. The sample (1 g) was homogenized in 10 ml of solution A and derivatized as described. Twenty-five microliters were injected on the HPLC system. The peak at 40.0 min is the S-carboxymethyl-N-dinitrophenyl derivative of GSH. The GSSG derivative is the peak at 44.3 min.

an Ultrasil NH2 column from Beckman (Norcross, GA) or an aminopropyl column from Custom LC (Houston, TX). Initial conditions are 1 ml/min flow with 80% solvent A (80% methanol) plus 20%o solvent B (5 M sodium acetate, pH 4.6, in 80% methanol). Initial conditions are maintained for 10 min followed by a linear gradient to 95% B over 40 min and reequilibration with initial conditions for 10 min prior to analysis of the next sample. Derivatives are detected at 365 nm and quantified relative to standards by integration (Fig. 1). Integral values for portions taken in solutions A, B, and C are plotted as a function of added GSH content (0, 10, and 20/zM), and the negative of the x axis intercept is taken as the GSH concentration (Fig. 2). Values are multiplied by 11 to convert back to sample concentrations and appropriately recalculated to be expressed in terms of millimolar or as mg/g dry weight. Values are converted to milligrams per usual portion size by use of the H H H Q medium portion sizes from the H H H Q database. Total GSH values are estimated from the same graphical representation as the negative of the x axis intercept of a line through the point (0,D) with the same slope as the line through the data from portions A - C (see Fig. 2).

[ 1]

GLUTATHIONE DISTRIBUTION IN NATURAL PRODUCTS

7

600

•~.

~

.=

~2oo

400

IO,BJ _/J

-10

0

10

20

G S H Added (~tNO

F1~. 2. Graphical analysis of the GSH content of bell pepper from multiple samples with a method of additions. Integral values from replicate samples homogenized in solutions A, B, and C were graphed as a function of the respective amounts of GSH in each solution, namely, 0, 10, and 20 ~M. These are identified as points (0,A), (10,B), and (20,C) (filled circles). Extrapolation of the linear least-squares best-fit line back to the x axis provides an estimate of the negative of the GSH content. For most foods, the slope of this line approximated that of the GSH contents of solutions A-C without added food sample (open circles). For some foods, GSH-reactive materials result in less than complete recovery of added GSH (filled squares). Extrapolation of the linear least-squares best-fit line for these points thus allows a correction for incomplete recovery. For estimation of GSHt, the integral value for the sample taken in solution D is plotted as (0,D), and a line is drawn through this point with a slope equivalent to the best fit for the corresponding solutions A-C. The negative of the x intercept is taken as the GSHt concentration. All values are corrected for dilution prior to expression per wet weight, dry weight, or total nitrogen.

Glutathione Content in Food Groups A l t h o u g h k n o w l e d g e of G S H c o n t e n t s of foods r e m a i n s i n c o m p l e t e , studies have p r o v i d e d insight i n t o the r a n g e of values a n d p r o v i d e s o m e useful g e n e r a l i z a t i o n s . O f p a r t i c u l a r i m p o r t a n c e , c o n s i d e r a b l e v a r i a t i o n in G S H c o n t e n t occurs in different foods, a n d e s t i m a t e d daily i n t a k e s of G S H vary c o n s i d e r a b l y b e t w e e n individuals. T h e G S H c o n t e n t s of r e p r e s e n t a t i v e foods are g i v e n in T a b l e I. D a i r y products, eggs, fats, a n d oils, m o s t b e v e r a g e s , a n d m o s t cereal g r a i n p r o d u c t s have very little G S H . I n contrast, fresh fruits a n d v e g e t a b l e s a n d freshly c o o k e d m e a t s are high in G S H . F r o z e n foods g e n e r a l l y have G S H c o n t e n t s c o m p a r a b l e to fresh foods, b u t o t h e r f o r m s of p r e s e r v a t i o n u s u a l l y result

8

GLUTATHIONE

[1]

TABLE I REPRESENTATIVE TOTAL GLUTATHIONE CONTENTS IN COMMONLY CONSUMED FOODS a

Food Vegetables Asparagus (fresh, cooked) Carrots (raw) Carrots (fresh, cooked) Carrots (canned, heated) Potatoes (fresh, cooked) Fruits Oranges (fresh) O r a n g e juice (frozen, reconstituted)

Apples (fresh) Apple juice (bottled) Meats Chicken (fried) Ham (boiled) Beef (ground, fried) Bacon (fried) Cereal, grains, and legumes Pinto beans (canned, heated) Peanut butter Bread, whole wheat Rice Dairy products Milk (whole)

Cheese, American

GSHt (mg/100 g)

(rag/medium portion)

28.3 - 5.4 7.9 -4- 1.5 5.8 - 1.5 0.0 13.6 - 2.5

26.3 5.4 4.3 0.0 12.7

7.3 - 0.8 4.2 _+ 0.1 3.3 _+ 0.5 0.0

10.6 7.9 4.6 0.0

13.1 23.3 17.5 5.0 0.6 2.4 1.2 1.6

-+ 2.8 -+ 6.5 _+ 0.8 -+ 0.7 +-+ -

0.0 0.0

0.1 0.5 0.4 0.1

GSHt

13.4 13.0 14.9 0.8 0.9 0.4 0.6 2.1 0.0 0.0

a Values of GSHt represent

total GSH recovered after reduction by dithiothreitol, expressed as mean _+ S E M for 100 g wet weight.

in substantial or complete loss of GSH. 13 Normal cooking results in variable loss, typically with different consequences for plant and animal products. Plant products lose GSH to forms that are not converted back to GSH with dithiothreitol, whereas animal products have forms that are reducible back to GSH. In addition to the loss with cooking, it is noteworthy that some foods contain substantial amounts of GSH-reactive materials. This is revealed by analyzing the percentage of added GSH that is recovered in the method of additions described above (see Fig. 2, dashed line). 6 Ground beef has a substantial loss of added GSH, whereas beef steak does not. This suggests that disruption of tissue by grinding allows lipid peroxidation to occur and may contribute to an increased load of reactive electrophiles and oxidants in the diet. Additional studies are needed to establish the identity of the

[ 1]

GLUTATHIONE DISTRIBUTIONIN NATURALPRODUCTS

9

reactive species and whether they contribute to diet-related disease processes.

D i s p o s i t i o n of Orally S u p p l i e d R e d u c e d a n d Oxidized G l u t a t h i o n e U p t a k e of G S H by epithelial cells of the gastrointestinal tract can be studied in freshly isolated cells, 19 cultured cells, 2° and in intact preparations where the mucosa is sampled by scraping, a Loss of G S H from the lumen can also be directly measured as a function of time or location along gastrointestinal tract, u However, m e a s u r e m e n t of absorption, which occurs along with degradation and resynthesis, is technically m o r e difficult and must be addressed in terms of the b r o a d e r context of thiol-disulfide balance which includes cysteine and cystine. Cysteine and cystine are also present in the diet and are actively transported in the small intestine. In addition, cysteine is critical for iron absorption. 21 and is transported from the epithelium into the lumen to regulate thiol-disulfide balance. 2t Glutathione and related products are also released into the intestinal lumen in bile. Thus, a variety of intact biological preparations, including cannulated bile duct, t3 lymph fistula, a4 and in v i v o blood sampling, 25 are needed to assess fully the disposition of orally supplied G S H and GSSG. A m e t h o d that we have found to be particularly useful for short-term (up to 1 hr) intestinal absorption studies involves direct m e a s u r e m e n t of lumen to vasculature transit of G S H in an in situ, vascularly perfused small intestinal preparation. 4,26 For this preparation, we use male, SpragueDawley rats (VAF/1) (Sasco, O m a h a , NE) weighing 200-320 g. They are maintained under controlled conditions of t e m p e r a t u r e and humidity on a 12 hr light/dark cycle. Rats are typically given food and water a d l i b i t u m , and experiments are p e r f o r m e d during the late morning or early afternoon when rats are not feeding. Rats are anesthetized with ketamine/xylazine, and heparin is injected into the left femoral vein. The jejunum is isolated and cleared of its contents by flushing with 40 ml of 37 ° saline (0.9%, w/v NaC1). The bile duct, celiac artery, renal arteries, and aorta (proximal to the celiac artery) are ligated 19L. H. Lash, T. M. Hagen, and D. P. Jones, Proc. Natl. Acad. Sci. U.S.A. 83, 4641 (1986). 2oK. Sundqvist, Y. Liu, K. Arvidson, K. Ormstad, L. Nilsson, R. Toftgard, and R. C. Grafstrom. In Vitro Cell Dev. Biol. 27A, 562 (1991). 21M. Layrisse, C. Martinez-Torres, I. Leets, P. Taylor, and J. Ramirez, J. Nutr. 114,217 (1984). 22L. J. Dahm and D. P. Jones, Am. J. Physiol. 267, G292 (1994). 23D. Eberle, R. Clarke, and N. Kaplowitz, J. Biol. Chem. 256, 2215 (1981). 24T. Y. Aw, M. W. Williams, and L. Gray, Am. J. Physiol. 262~G99 (1992). 25B. H. Lauterburg, J. A. Adams, and J. R. Mitchell, Hepatology (Baltimore) 4, 586 (1984). 26T. M. Hagen, C. Bai, and D. P. Jones, FASEB J. 5, 2721 (1991).

10

GLUTATHIONE

[ 11

and the rat is killed by cutting through the diaphragm and vena cava. The aorta is then cannulated distal to the renal arteries, and retrograde perfusion is begun with Krebs-Henseleit buffer (pH 7.4, 37 °) containing 12.5 mM HEPES and equilibrated with 95% 02 and 5% CO2 (v/v). This allows perfusion of the small intestine by flow through the superior mesenteric artery. Perfusate is collected by cannulating the portal vein. The flow rate is usually set at 1 ml/min because this optimizes the ability to measure low transport rates. Under these conditions, the 02 consumption rate is about 60% of maximal, but there is no loss of viability (as measured by lactate dehydrogenase release) or loss of vascular integrity (as measured by polyethylene glycol permeability). Body temperature is maintained by a heated operating tray or a heat lamp. After preperfusion for 5 min to clear the vasculature of blood, the jejunum is instilled with Krebs-Henseleit medium (in mM: 120 NaC1, 4.8 KCI, 2.6 CaCI2, 1.2 MgSO4, i KH2PO4, 2.4 NaHCO3) and other additions as appropriate. In a typical experiment measuring GSH uptake (Fig. 3), 4,26,27 [35S]GSH is placed in the lumen, and samples are taken from the portal vein every minute. Aliquots are derivatized and analyzed by HPLC as described above. Use of this method allows direct measures of the rate of GSH transport as well as the rate of GSH degradation and cysteine absorption. 4 Substitution of choline chloride for NaCI in the Krebs-Henseleit medium allows study of Na+-independent transport. 4 Transport of polyethylene glycol from the lumen to the vasculature provides a measure of loss of vascular integrity. The contribution of GSH synthesis can be virtually eliminated by inclusion of 1 mM buthionine sulfoximine (BSO) in the luminal and perfusate solution. 4 Inhibition of GSH degradation poses more of a technical problem in that 250 /xM acivicin, a concentration that gives over 99% inhibition of y-glutamyltransferase in mucosal homogenates, only gives about 70% inhibition in the in situ vascularly perfused intestine. Inhibition can be increased to more than 95% by increasing the acivicin concentration to 1 mM, but this has the risk of inhibiting GSH transport and other processes because acivicin becomes a nonspecific alkylating reagent at high concentrations. An alternative means to inhibit y-glutamyltransferase is to use serineborate, which gives greater than 99% inhibition with 20 mM each of serine and borate at pH 7.5. When [glycine-3H]GSH is used for tracer studies, we have found that inclusion of 5 or 10 mM glycine in the solutions effectively eliminates only the contribution of GSH synthesis to the measured vascular [3H]GSH. 27 T. M. Hagen, G. T. Wierzbicka, A. H. Sillau, B. B. Bowman, and D. P. Jones, Am. J. Physiol. 259, G524 (1990).

[ 1]

GLUTATHIONE DISTRIBUTION IN NATURAL PRODUCTS

11

;9 0

10 Time (min)

Fie. 3. Uptake of GSH in vascularly perfused small intestine of the rat. The 35S-labeled GSH was placed in the lumen at zero time, and the portal vein effluate was collected and counted by scintillation counting (filled circles). In these experiments, performed in the laboratory by C. Bai, the intestine was pretreated with acivicin to inhibit GSH degradation. Under such conditions, greater than 90% of the total radiolabel is GSH as analyzed by HPLC. Without acivicin treatment, 30% or more of the radiolabel is recovered as cysteine or cystine (4). A comparable concentration of [14C]polyethylene glycol shows little radiolabel appearance in the perfusate (open circles). Additional details are given in Refs. 4, 26, and 27.

Hormonal regulation of GSH transport 26 and intestinal thiol-disulfide regulation22 have been successfully studied using the in situ vascularly perfused intestinal preparation. Additional studies are needed to determine whether GSH absorption or thiol-disulfide regulation are affected by the status of other antioxidants or by exposure to oxidants or reactive electrophiles in the diet. Modulation of glutathione availability and absorption can potentially provide an effective means to reduce morbidity from intestinal distress during chemo- and radiotherapies. Systemic Availability and Tissue Distribution of Absorbed Glutathione Dietary GSH is typically only a small fraction of the total sulfur amino acids in the diet. In addition, the amount of GSH consumed in a day is only a small fraction of the total in the body. Thus, it appears unlikely that a measurable effect of dietary GSH on systemic GSH would be detectable. However, blood plasma concentrations are typically low, and increases in blood plasma concentrations have been detected following oral loads in

12

[ 1]

GLUTATHIONE Control

{Z~ +GSH

6' A

E -to9 O

2,

Liver

Kidney

3.

Heart

Lung

Brain

intestine

Skin

Iml BSO control

•

E

2'

+Amino adds

i--1 +GSH

o

•

0

,

Liver

,

Kidney Heart

,

Lung

Brain Intestine Skin

FIG. 4. Effect of oral supplementation with GSH on tissue GSH concentrations in mice. Top: Mice were given tap water which did not contain BSO but were otherwise treated identically to the mice used for studies shown in the bottom graph. Controls received 0.3 ml saline by gavage i hr prior to sacrifice and tissue sampling. The GSH-treated animals received 100 mg/kg body weight in the saline gavage. Bottom: Mice were treated with BSO in the drinking water for 5 days prior to gavage with either 0.3 ml saline, saline with GSH (100 mg/ kg body weight), or saline with an equivalent amount of constitutive amino acids. Asterisks represent significant differences between GSH-treated and respective controls at p < 0.05.

rats, 4'27 mice, 28 and humans. 29 Although increases in blood plasma G S H are transient following oral consumption, such increases may be important in detoxication of reactive species in the blood 3° or in protection of cell surfaces from oxidative damage. 31 A particularly useful approach to study systemic availability of G S H is to use animals treated with the G S H synthesis inhibitor, BSO. 28 The com28 T. Y. Aw, G. Wierzbicka, and D. P. Jones, Chem.-Biol. Interact. 8 0 , 89 (1991). 29 T. M. Hagen and D. P. Jones, in "Glutathione Centenial: Molecular and Clinical Implications" (Y. Sakamoto, T. Higashi, N. Taniguchi, and A. Meister, eds.), p. 423. Academic Press, San Diego, 1989. 30 C. D. Thomson, M. F. Robinson, J. A. Butler, and P. D. Whanger, Br. Z Nutr. 69, 577 (1993). 31 p. C. Davidson, P. Sternberg, D. P. Jones, and R. L. Reed, Invest. Ophthamol. Vis. Sci. 35, 2843 (1994).

[ 1]

GLUTATHIONE DISTRIBUTION IN NATURAL PRODUCTS

13

pound is an irreversible inhibitor of 7-glutamylcysteine synthetase (glutamate-cysteine ligase), the rate-limiting enzyme in GSH synthesis.3eWith this treatment, mice can be maintained for several days and have substantially decreased tissue GSH concentrations. Substantial ultrastructural effects have been noted, 33,34 but there is no overt pathology. We use outbred albino ICR mice (30-40 g, Sasco) that are maintained on mouse food and water ad libitum. 28 After a 1-week preconditioning period to allow animals to adjust to the animal care facility, 20 mM BSO is included in the drinking water. From the volumes consumed, this amounts to about 80 tzmol BSO per mouse per day. After 5 days food is removed, and, on the sixth day, mice are given an oral load of GSH or amino acids in 0.3 ml of 0.9% (w/v) NaCI by gavage. Control animals receive saline only. One hour after gavage, animals are sacrificed, and 10 ml of ice-cold saline is perfused through the aorta to wash blood from the tissues. Tissue samples are immediately taken, weighed, and placed in ice-cold 30% trichloroacetic acid. After centrifugation to remove protein, aliquots are derivatized and analyzed for GSH content by HPLC as described above. Results from the experiments show that tissue concentrations of GSH can be increased by orally supplied GSH even though GSH synthesis is inhibited, and an equivalent amount of constituent amino acids does not increase tissue GSH concentrations (Fig. 4). Comments Numerous studies have established the importance of GSH in protection against a variety of toxicological and pathological processes. Glutathione is a compound that occurs naturally in the diet and can be readily supplemented or administered therapeutically. Although there are several exciting reports on the possibility that nutritional variations in GSH consumption and oral administrations of GSH can have beneficial effects against specific disease processes, relatively few scientific studies of this nature are available. The described methods provide approaches to more detailed investigation of the nutritional and physiological control of GSH consumption and utilization and should provide an important resource for investigations of the role of dietary GSH in human health.

32 O. W. Griffith and A. Meister, J. Biol. Chem. 254, 7558 (1979). 33 j. Martensson, A. Jain, W. Frayer, and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 87, 5296 (1990). 34j. Martensson, A. Jain, and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 87, 1715 (1990).

14

GLUTATHIONE

[2]

[2] I n t r a c e l l u l a r D i s t r i b u t i o n o f T h i o l s a n d D i s u l f i d e s : Assay of Mitochondrial Glutathione Transport

By LAWRENCEH. LASH Introduction Metabolism of glutathione (GSH) in mammalian cells exhibits many unique characteristics: GSH is synthesized by two ATP-dependent enzymes, rather than by the normal protein synthetic mechanism, and is degraded by a specific enzyme, 7-glutamyltransfcrase (EC 2.3.2.2), which is found selectively on the luminal plasma membranes of epithelial cells, including those of the renal proximal tubule and the small intestinal epithelium. Besides synthesis and degradation, another component of the turnover of cellular GSH is transport across plasma membranes and intracellular membranes. Nutritional and biochemical studies on GSH turnover have shown that the GSH pool within many mammalian cell types, such as liver 1'2 and renal proximal tubule, 3 is divided into two distinct subpools, one comprising the majority of total intracellular GSH (70-85% of total) that turns over relatively rapidly (h/2 of 2 hr) and one that comprises a smaller portion of total intracellular GSH (15-30% of total) that turns over relatively slowly (tl/2 of 30 hr). Investigations employing digitonin fractionation and marker enzymes showed that the larger pool of intracellular GSH is cytosolic in origin, whereas the smaller, slowly turning over pool is predominantly mitochondrial in origin. Much less is known about the properties and regulation of the mitochondrial pool of intracellular GSH as compared with the cytosolic pool because the mitochondrial pool comprises a relatively minor fraction of total cellular GSH and because the cytosolic pool is more easily accessible to analysis and manipulation. However, numerous studies have provided evidence that the mitochondrial pool of GSH is critical for the maintenance of both mitochondrial and cellular function and that several types of chemically induced injury are associated with depletion or oxidation of the mitochondrial, rather than the cytosolic, pool of G S H . 4-9 1 M. J. Meredith and D. J. Reed, J. BioL Chem. 257~ 3747 (1982). 2 F. J. R o m e r o and H. Sies, Biochem. Biophys. Res. Commun. 123, 1116 (1984). 3 R. G. Schnellmann, S. M. Gilchrist, and L. J. Mandel, Kidney Int. 34, 229 (1988). 4 M. C. Beatrice, D. L. Stiers, and D. R. Pfeiffer, J. Biol. Chem. 259, 1279 (1984). s K. L~-Qu6c and D. LS-Qu6c, Arch. Biochem. Biophys. 216, 639 (1982). 6 K. LS-Qu6c and D. L~-Qu6c, J. Biol. Chem. 260, 7422 (1985).

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

[2]

MITOCHONDRIAL GLUTATHIONE TRANSPORT

15

The inability to detect de n o v o synthesis of GSH within hepatic mitochondria 1° suggested that transport of cytosolic GSH as the intact tripeptide across the mitochondrial inner membrane into mitochondrial matrix must occur to provide the organelle with GSH. Because the fraction of total cellular GSH found in mitochondria approximates the fractional volume of mitochondria within cells such as hepatocytes or renal proximal tubular cells, there is probably little if any gradient of GSH across the mitochondrial inner membrane between cytosol and mitochondrial matrix. The subcellular distribution of GSH cannot be due to simple diffusion, however, but must be actively regulated because: (i) the mitochondrial pool of GSH turns over much more slowly than the cytosolic pool; (ii) the GSH molecule has a net negative charge at physiological pH ranges; and (iii) the mitochondrial matrix is negatively charged relative to the cytosol (i.e., there is a membrane potential across the mitochondrial inner membrane) and there is a pH gradient across the mitochondrial inner membrane (i.e., matrix pH is higher than the cytosolic pH). Indeed, Kurosawa et al.ll and Martensson and Meister 12described transport systems for GSH uptake into rat liver mitochondria. The hepatic mitochondrial system is dependent on membrane potential and the presence of a proton gradient. Two transport systems were distinguished based on kinetics, a high-affinity, low-capacity system and a low-affinity, high-capacity system. 12 Kaplowitz and colleagues 13 have found that activity of the hepatic system is altered by chronic ethanol feeding of rats. Schnellmann 14 found evidence for a GSH transport system in renal cortical mitochondria. Lash and colleagues15 determined that this transport process is electroneutral and appears to involve exchange with dicarboxylic acids, such as malate and succinate. This chapter describes methods to quantitate transport of GSH as the intact tripeptide into rat renal mitochondria. Procedures are described to analyze uptake of GSH into mitochondria using both isolated renal cells and isolated renal cortical mitochondria, although the methods are applicable 7 K. L~-Qu6c and D. L~-Qu6c, Arch. Biochem. Biophys. 273, 466 (1989). 8 H. Sies and K. M. Moss, Eur. J. Biochem. 84, 377 (1978). 9 T. Yagi and Y. Hatefi, Biochemistry 23, 2449 (1984). 10 O. W. Griffith and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 82, 4668 (1985). ~t K. Kurosawa, N. Hayashi, N. Sato, T. Karnada, and K. Tagawa, Biochem. Biophys. Res. Commun. 67, 367 (1990). le j. Martensson and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 86, 471 (1989). 13 j. C. Fernandex-Checa, C. Garcia-Ruiz, M. Ookhtens, and N. Kaplowitz, J. Clin. Invest. 87, 397 (1991). 14 R. G. Schneltmann, Life Sci. 49, 393 (1991). 15 T. B. McKernan, E. B. Woods, and L. H. Lash, Arch. Biochem. Biophys. 288, 653 (1991).

16

GLUTATHIONE

[2]

to measurement of transport into mitochondria from other tissues with appropriate, minor modifications.

Biological Preparations

Preparation of Isolated Renal Cortical Mitochondria Although description of biological preparations is not the primary focus of this chapter, a few points will be made that are relevant to the analysis of GSH transport. Suspensions of isolated renal cortical mitochondria that are suitable for GSH transport studies are obtained from renal cortical homogenates by differential centrifugation methods (e.g., see Johnson and Lardy16 and Lash and Sa1117). These procedures employ homogenization in buffers containing a high concentration of an impermeant solute, such as 225 mM sucrose, and potassium phosphate, magnesium chloride, and an organic buffer, such as Tris-HC1 or triethanolamine hydrochloride. Additionally, a divalent cation chelator such as ethylene glycol bis(/3-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA) should be included in the preparation of the mitochondria, but not in the final suspension medium. This removes much of the calcium ions that are released from intracellular stores during homogenization and isolation of the mitochondria. Without chelation of the calcium ions, damage to mitochondrial membranes can occur. The buffer composition is important because it enables isolation of mitochondria that maintain a high degree of structural integrity and selective permeability, thereby making the preparation suitable for transport studies.

Analysis of Mitochondrial Uptake in Intact Renal Cells Two methods can be employed to analyze uptake of GSH into mitochondria with intact renal cells as the biological system. In the first method, the mitochondrial content of GSH is determined after incubations with intact cells by fractionation of cells with digitonin. The digitonin fractionation method (see Lash and Sail 17 for methodological details) involves rapid mixing of cells in a microcentrifuge tube with the top layer that contains buffer and 0.15% (w/v) digitonin in a microcentrifuge tube and centrifugation through a silicone oil-mineral oil layer (6: 1; v/v) into a 40% (v/v) 16D. Johnson and H. Lardy,this series, Vol. 10, p. 94. 17L. H. Lash and J. M. Sail, in "MitochondrialDysfunction"(L. H. Lash and D. P. Jones, eds.), Methodsin Toxicology,Vol. 2, AcademicPress, San Diego, p. 8. 1993.

[2]

MITOCHONDRIAL GLUTATHIONE TRANSPORT

17

glycerol layer. This procedure enables rapid separation of cytosolic and mitochondrial compartments with minimal postisolation artifacts. The second method (see Lash and Sall av for methodological details) enables preparation of mitochondria from suspensions of intact cells, thereby providing isolated mitochondria for direct measurement of transport. A distinct advantage of this method is that cell populations derived from specific regions of a tissue, such as a particular nephron segment in mammalian kidney, can be used as the source of mitochondria. This method is the only one available that gives mitochondria from suspensions of intact cells that exhibit a high degree of morphological and functional integrity. Isolated cells are mixed with digitonin and are centrifuged at a low speed. Further steps beyond simple digitonin treatment are necessary to isolate the mitochondria because digitonin permeabilizes the plasma membranes, thereby releasing soluble components of the cytosol, but it does not alter the intracellular cytoskeleton. To separate mitochondria from intracellular cytoskeletal elements, pellets from the digitonin treatment step are mixed with the protease nagarse, and the suspension is filtered through polycarbonate membranes. The nagarse releases intracellular organelles from the cytoskeletal network, and the polycarbonate membranes are inert toward biological materials. Use of the polycarbonate membranes provides a means to recover intracellular organelles. As with the mitochondria isolated from homogenates, mitochondria isolated by the digitonin-nagarse and filtration method are resuspended in a buffer that should optimize the structural integrity of the mitochondrial inner membrane. This buffer contains an impermeant molecule (i.e., sucrose), potassium chloride, and an organic buffer. This method was originally developed for isolated rat hepatocytes, 18 but has also been applied to the isolation of mitochondria from renal proximal tubular and distal tubular cells from rat. 19 Mitochondria isolated from intact cells can be incubated, processed, and analyzed identically to mitochondria obtained from tissue homogenates (see below).

Assessments of Functional Integrity When any biological preparation is employed for membrane transport studies, it is critical that the functional integrity of the model is validated. Use of an in vitro model that is compromised either morphologically or functionally may alter transport behavior so that nonspecific permeability or collapse of ion gradients across membranes interfere with measurement of carrier-mediated transport. 18 C. Bai, C. W. Slife, T. Y. Aw, and D. P. Jones, Anal. Biochem. 179, 114 (1989). 19 L. H. Lash, Pharmacologist 35, 148 (1993).

18

GLUTATHIONE

[21

Numerous methods are available to assess the functional integrity of suspensions of isolated renal mitochondria (see Lash and Salll7). These methods include spectrophotometric monitoring of matrix swelling with polyethylene glycol or leakage of matrix enzymes and assessment of respiratory control by measurement of rates of state 3 and state 4 oxygen consumption in an oxygraph. Determination of respiratory control allows an assessment of how well integrated, energy-linked processes are maintained. For the two methods employing intact, isolated cells, mitochondrial functional integrity can also be assessed by convenient methods. Because the objective with digitonin fractionation is to measure distribution of GSH into the mitochondrial compartment, it is critical that the viability of the isolated cell suspension is maintained and that the fractionation method produces an adequate separation of mitochondrial and cytosolic compartments. Procedures to assess the quality of the separation use marker enzyme activity distributions. 17 For the isolation of mitochondria from cell suspensions by the digitonin-nagarse treatment and filtration method, the functional integrity of the filtrates, which are essentially suspensions of isolated mitochondria, is evaluated by the same assays as are used with the isolated mitochondria prepared from homogenates.

Measurement of Mitochondrial Glutathione Transport General Principles

The accurate measurement of transport of GSH into or from renal mitochondria requires that GSH degradation and oxidation be minimized and that the mitochondrial compartment be efficiently separated from the extramitochondrial compartment. In experiments with intact renal cells, GSH oxidation can occur by nonenzymatic as well as enzymatic mechanisms.2° Enzymatic oxidation can be catalyzed by the copper-dependent thiol oxidase that is present on the basolateral plasma membrane, z° Oxidation can be prevented or minimized by addition of a metal ion chelator, such as bathophenanthroline disulfonate (BPDS) or ethylenediaminetetraacetic acid (EDTA). BPDS is preferred because it is readily water-soluble and EDTA interferes with the high-performance liquid chromatography (HPLC) assay for GSH (see below). Degradation of GSH is almost completely inhibited by preincubation of cells or mitochondrial fractions with the potent, irreversible inhibitor of y-glutamyltransferase acivicin [L(aS,5S)-a-amino-3-chloro-4,5-dihydro-5-isoxazoleacetic acid], zl In isolated 20L. H. Lash, D. P. Jones, and M. W. Anders, Rev. Biochem. Toxicol. 9, 29 (1988). 21D. J. Reed,W. W. Ellis,and R. A. Meck,Biochem. Biophys. Res. Commun. 94, 1273(1980).

[2]

MITOCHONDRIAL GLUTATHIONE TRANSPORT

19

renal plasma membrane vesicles, 50% inhibition of y-glutamyltransferase activity is obtained by preincubation with an acivicin concentration of 6 /xM, with essentially complete inhibition (>98%) occurring with an acivicin concentration of 0.25 mM. 22 In contrast, use of acivicin concentrations above 1 mM causes nonspecific alkylation of cellular nucleophiles, often producing cytotoxicity. Reagents and Materials: Biological Materials and Processing of Samples Biological Materials Renal cortical mitochondria (3-6 protein/ml) Isolated renal cells (2-5 × 106 cells/ml) General Materials Plastic Erlenmeyer flasks, 25 ml Acivicin, 2.5 mM in saline (10× solution) Plastic microcentrifuge tubes, 1.5 ml Microcentrifuge, 13,000 g maximum Materials for Isolated Mitochondria Dibutyl phthalate (d of 1.046 g/ml) Mitochondrial buffer: 225 mM sucrose, 10 mM potassium phosphate, pH 7.4, 5 mM MgC12, 20 mM KC1, 20 mM triethanolamine hydrochloride, pH 7.4, 0.1 mM phenylmethylsulfonyl fluoride (PMSF), 2 mM EGTA (included in all but final suspension of mitochondria) Materials for Isolated Renal Cells: Digitonin Fractionation Digitonin (for aqueous solutions; Sigma, St. Louis, MO), recrystallized two times from hot, absolute ethanol; digitonin stock solution (prepared fresh) contains 1.5 mg digitonin/ml N-morpholinoethanesulfonic acid (MES) buffer Silicone oil (d of 1.05 kg/liter) (Aldrich, Milwaukee, WI) White light paraffin oil, viscosity 125/135 (Fisher Scientific, Pittsburgh, PA) Krebs-Henseleit buffer (for suspension of isolated cells): 118 mM NaC1, 25 mM NaHCO3, 4.8 mM KC1, 25 mM HEPES, 1.2 mM MgSO4, 2.6 mM CaCI2, 1.0 mM KH2PO4; adjust to pH 7.2 MES buffer: 19.8 mM EDTA, 19.8 mM EGTA, 250 mM D-mannitol, 19.8 mM MES, pH 7.4 2: L. H. Lash and D. P. Jones, J. BioL Chem. 259, 14508 (1984).

20

OLUTATHIONE

[21

Materials for Isolated Renal Cells: Digitonin/Nagarse Treatment and Filtration Buffer A: 70 mM Tris base, 0.25 M sucrose, pH 7.0 Buffer B: 50 mM HEPES, 0.1 M sucrose, 0.1 M KC1, 1 mM EDTA, pH 7.2 Nagarse (bacterial protease, type XXVII from Sigma); stock solution, prepared fresh, contains 50 mg nagarse/ml buffer B MES buffer (see above) Digitonin (see above); stock solution contains 2.0 mg digitonin/ml MES buffer Bovine serum albumin (BSA; fraction V, Sigma); stock solution, stored at 0°, contains 0.2 g BSA/ml deionized water Polycarbonate membranes (25-mm diameter, 2.0-/zm pore size) with Swinnex filter holders (Millipore, Bedford, MA); filters are presoaked in distilled water for 30 min before being placed in the filter holder and are washed with 2 ml of buffer A before sample filtration

Reagents and Materials: Transport Assays Materials for Transport Method 1: Filtration and Enzymatic Assay Mixed cellulose ester membrane filters, 0.45-/xm pore size, 25-mm diameter Borosilicate glass test tubes, 16 x 125 mm Sulfosalicylic acid, 10% (w/v) Spectrophotometer Triethanolamine base, 25% (v/v) GSSG reductase [EC 1.6.4.2, glutathione reductase (NADPH)], 50, U/ml NADPH, 0.3 mM 5,5'-Dithiobis(2-nitrobenzoic acid) (DTNB), 6 mM

Materials for Transport Method 2: Centrifugation and Analysis by Chromatography Perchloric acid, 70% (v/v) 7-Glutamylglutamate, 15 mM, in 0.3% (v/v) perchloric acid; used as an internal standard for HPLC assay of GSH (see below) Bathophenanthroline disulfonic acid (BPDS), 15 mM, in deionized water Iodoacetic acid, 100 mM (prepared fresh daily) KOH (2 M)-KHCO3 (2.4 M) 1-Fluoro-2,4-dinitrobenzene, 1% (v/v), in absolute ethanol (prepared fresh daily)

[2]

MITOCHONDRIAL GLUTATHIONE TRANSPORT

21

Incubation and Sampling Procedures All incubations are performed at 25 ° to 28 ° for two reasons: biochemical processes in mitochondria often occur too rapidly at physiological temperatures (e.g., 37 °) to permit measurement, and the functional stability of mitochondrial preparations often declines rapidly at higher temperatures. Incubations are performed in 25-ml Erlenmeyer flasks made of polypropylene and are conducted in a Dubnoff metabolic shaking incubator (60 cycles/ min; Precision Scientific, Chicago, IL). In incubations of either whole renal cells or isolated renal mitochondria with GSH or oxidized glutathione (GSSG), it is critical to pretreat the biological material with acivicin for 15 min prior to addition of GSH or GSSG. 7-Glutamyltransferase is present at extremely high activities in brush border membranes of renal cells. Although the plasma membrane contamination of renal mitochondria may be fairly low, the extremely high activities of 7-glutamyltransferase indicate that a small contamination (e.g., 5% of total protein content) may produce a significant amount of GSH degradation. Mitochondria or isolated cells (9 volumes) are mixed with 2.5 mM acivicin (1 volume of 10x solution) and incubated for 15 min at 25 ° or 28 °. The GSH or GSSG transport in mitochondria or cells is measured by one of two methods (see below). In both cases, with either mitochondria or cells, incubations are performed identically in 25-ml Erlenmeyer flasks: Five volumes of acivicin-pretreated mitochondria or cells are mixed with 1 volume of 6x uptake solution (GSH dissolved in mitochondrial or KrebsHenseleit buffer). Total incubation volumes are between 2 and 3 ml. Transport Method 1: Filtration and Enzymatic Assay. Aliquots (0.5 ml) of mitochondrial suspensions from cortical homogenates or from isolated cells are rapidly filtered under vacuum on 0.45-/xm pore size, 25-mm diameter filters. After washing filters twice with 2 volumes of ice-cold incubation buffer, filters are placed in 16 x 125 mm borosilicate glass test tubes with 0.6 ml of distilled water for 30 min on a shaking water bath to release intramitochondrial contents retained on the filters. Samples (0.5 ml) are mixed with 10% (v/v) sulfosalicylic acid (0.25 ml) in 1.5-ml plastic microcentrifuge tubes and are centrifuged in a microcentrifuge at 13,000 g for 2 min. Acid-soluble supernatants are neutralized with triethanolamine base (25%, v/v) and are assayed for total GSH by an enzymatic recycling assay using GSSG reductase and Ellman's reagent (i.e., DTNB). ~3 Nonspecific binding of GSH to filters is assessed by addition of a known amount of GSH to filters without mitochondria, and is typically less than 5% of initial GSH content of mitochondrial samples. 15 Nonspecific binding of GSH to mito23 O. W. Griflith, Anal. Biochem. 106, 207 (1980).

22

GLUTATHIONE

121

chondria is assessed by addition of a known amount of GSH to filters after filtration of mitochondria, and is also typically only 5% of the endogenous amount of GSH normally present in renal mitochondrial samples. 15 Recovery of GSH is assessed by addition of a known amount of GSH to mitochondrial samples just prior to filtration, and is greater than 85%.

Transport Method 2: Centrifugation and Analysis by Chromatography. Aliquots (0.5 ml) of mitochondrial suspensions from cortical homogenates or from isolated cells are layered on 0.45 ml of dibutyl phthalate over 0.55 ml of perchloric acid (10%, v/v) containing 1% (w/v) BPDS and 75 nmol 7-glutamylglutamate (internal standard for HPLC method). Tubes are centrifuged at 13,000 g for 1 min. The supernatant (i.e., top layer) can be sampled to analyze extramitochondrial GSH. The oil layer is gently washed twice with incubation buffer. This step is designed to minimize any carryover of extracellular material into the bottom layer. As much as possible of the oil layer is removed, and a 0.5-ml aliquot of the perchloric acid layer is removed with a syringe. The use of dibutyl phthalate, based on the method described by Fariss and Reed, 24 is done to separate extramitochondrial medium from intramitochondrial medium. The acid-soluble material in the bottom layer of the microcentrifuge tube is analyzed for GSH, GSSG, and related metabolites (e.g., cysteine, glutamate) by ion-exchange HPLC. 24 The method involves derivatization of thiols with iodoacetic acid to form the S-carboxymethyl derivative, and derivatization of amino acids, disulfides, and S-carboxymethyl derivatives with 1-fluoro-2,4-dinitrobenzene to form the corresponding N-dinitrophenyl, N,N-bisdinitrophenyl, and S-carboxymethyl-N-dinitrophenyl derivatives, respectively. The derivatives are then separated on an amine column with a mobile phase of 80% (v/v) aqueous methanol and a linear gradient of acetate ions. Recovery of GSH or GSSG, assessed as described above, is greater than 90%. Studies on Uptake of Glutathione by Renal Cortical Mitochondria from Rats In studies with renal cortical mitochondrial suspensions, Lash and colleagues 15 demonstrated the time- and concentration-dependent uptake of intact GSH across the mitochondrial inner membrane by a cartier-mediated mechanism. A major advantage of using the above-mentioned methods for analysis of GSH uptake into mitochondria is that the methods are specific for measurement of the intact tripeptide rather than for potential degradation products of GSH (i.e., constituent amino acids). This is critical, even 24M. J. W. Fariss and D. J. Reed, this series, Vol. 143, p. 101.

[2]

MITOCHONDRIAL GLUTATHIONE TRANSPORT

23

with pretreatment of cells or mitochondria with acivicin, because 7-glutamyltransferase activity is so prominent in renal subcellular fractions. A time course of intramitochondrial GSH content in renal cortical mitochondria incubated with either buffer or 5 mM (Fig. 1) shows that the organelles rapidly accumulate GSH. In the absence of exogenous GSH, intramitochondrial GSH content decreases by one-third from 4.77 to 3.13 nmol/mg protein during 10 min of incubation at 28°. In the presence of 5 mM exogenous GSH, in contrast, intramitochondrial GSH content increases nearly 3-fold to 13.3 nmol/mg protein after 2 min of incubation, and then slowly declines over the next several minutes. The energetics of mitochondrial GSH transport is complex, but the transport mechanism does not appear to be coupled to the membrane potential, as illustrated in Fig. 2. Uncoupling of mitochondria with the protonophore CCCP (carbonyl cyanide m-chlorophenyl hydrazone) or with 2,4-DNP does not significantly affect GSH uptake, suggesting that the negatively charged GSH molecule is either cotransported with a cation or undergoes countertransport (i.e., exchange) with another anion. In con-

15

12

9

6

0

0

I

I

5

tO

Time (min)

FIG. 1. T i m e course of G S H uptake by isolated rat renal cortical mitochondria. Isolated mitochondria from rat kidney cortex ( 3 - 6 m g protein/ml) were incubated with either buffer (filled circles) or 5 m M G S H (open circles) at 28 ° for the indicated times. The G S H content in mitochondria was m e a s u r e d by HPLC. Results are the m e a n s of experiments from seven mitochondrial preparations.

24

GLUTATHIONE

[2]

100

75 o

50

25

0

FI~. 2. Studies on the energetics of GSH uptake by isolated rat renal cortical mitochondria. Isolated mitochondria from rat kidney cortex (3-6 mg protein/ml) were preincubated for 15 min with the indicated compound and were then incubated for 5 min with 5 mM GSH at 28°. The GSH content in mitochondria was measured by HPLC. Results are the means of experiments from three to six mitochondrial preparations. Asterisks signify results significantly different (p < 0.05) from the corresponding control incubation.

trast, ATP, KCN, and calcium chloride significantly inhibit G S H uptake, suggesting an indirect coupling to mitochondrial energetics. Investigation of the interaction of G S H with mitochondrial anions, by looking at effects on the same side of the mitochondrial membrane (cis effects) and on the opposite side of the mitochondrial membrane (trans effects) (Fig. 3), shows that divalent anions, such as succinate and malate, undergo exchange transport with G S H whereas monovalent anions, such as pyruvate, do not interact with G S H transport. The interaction with L-glutamate is more complex, involving both competitive inhibition and trans-stimulation. Comments Although we have applied these methods to the study of G S H transport in renal mitochondria, they should be equally applicable to mitochondria from other tissues. Analysis of G S H transport with three biological models

[2]

25

MITOCHONDRIAL GLUTATHIONE TRANSPORT 200. []

Cis uptake (%)

[]

Trans uptake (%)

150-

o

~

100

eL

50-

W Pyruvate

Glutamate

Malate

Succinate

Flc. 3. Interactions of mitochondrial GSH transport in rat kidney with mono- and dicarboxylates. Isolated mitochondria from rat kidney cortex (3-6 mg protein/ml) were incubated for 5 min with 5 mM GSH and incubated simultaneously (cis) or preincubated for 15 min to preload the mitochondria (trans) with the indicated mono- or dicarboxylic acid (10 mM). The GSH content in mitochondria was measured by HPLC. Results are the means of experiments from three to five mitochondrial preparations. Asterisks signify results significantly different (p < 0.05) from the corresponding control incubation.

was described. The first method involves studies with suspensions of isolated mitochondria from renal cortical homogenates. The second method involves incubation of suspensions of intact renal cells with GSH and isolation of cytosolic and mitochondrial fractions with digitonin, thus enabling analysis of GSH compartmentation into mitochondria. The final method is an extension of the digitonin method, which involves subsequent treatment with the protease nagarse and filtration on polycarbonate membranes to isolate mitochondria. The unique advantage of this biological model is that intact, viable mitochondria can be isolated from purified preparations of specific cell populations. Critical features of the methods involve the efficient separation of extra- and intramitochondrial space and the ability of the analytical methods to quantitate precisely and unambiguously intact GSH rather than degradation products.

26

GLUTATHIONE

[31

Acknowledgments Research from the author's laboratory was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant 2 R01-DK40725. Lawrence H. Lash is a recipient of a Research Career Development Award from the National Institute of Diabetes and Digestive and Kidney Diseases (Grant 1 K04-DK02090).

[31 Glutathione Biosynthesis and Its Inhibition By A L T O N

MEISTER

Introduction Glutathione (GSH) is synthesized in two steps that are catalyzed, respectively, by 7-glutamylcysteine synthetase (glutamate-cysteine ligase) and GSH synthetase1: u-Glutamate + l-cysteine + ATP ,-~ l-T-glutamyl-L-cysteine + ADP + Pi

(1) L-T-Glutamyl-L-cysteine + glycine + ATP ~-- GSH + ADP + Pi

(2) The reaction catalyzed by T-glutamylcysteine synthetase is feedback inhibited by GSH, and thus the enzyme functions normally at less than its maximal rate. Inhibition by GSH is not allosteric, but is competitive with respect to glutamate. Inhibition is associated with binding of the glutamyl moiety of GSH to the glutamate binding site of the enzyme, and there is evidence that the thiol moiety of GSH binds to another site on the enzyme, possibly to the cysteine binding site. 2'3 Earlier studies on the isolation, properties, catalytic mechanisms, and inhibition of the enzyme and GSH synthetase have been reviewed in this seriesJ ~/-Glutamylcysteine synthetase isolated from rat kidney has a molecular weight of about I00,000 and can be dissociated into subunits that have molecular weights of 73,000 and 27,700. The heavy subunit was isolated after dissociation of the enzyme under nondenaturing conditions in the presence of 50 mM dithiothreitol. The heavy subunit exhibits all of the catalytic activity of the isolated enzyme as well as feedback inhibition by GSH. Although the possibility that the light subunit is not an integral part of the enzyme was considered, studies in which both subunits were cloned i G. F. Selig and A. Meister, this series, Vol. 113, p. 379. 2 p. Richman and A. Meister, J. Biol. Chem. 250, 1422 (1975). 3 C.-S. Huang, L.-S. Chang, M. E. Anderson, and A. Meister, J. Biol. Chem. 268, 19675 (1993).

METHODSIN ENZYMOLOGY,VOL.252

Copyright© 1995by AcademicPress,Inc. All fightsof reproductionin any formreserved.

[3]

GLUTATHIONE BIOSYNTHESISAND ITS INHIBITION

27

and sequenced have indicated that both are essential. Thus, the cDNA for the heavy subunit was isolated from rat kidney and found to code for a protein of 637 amino acid residues (Mr 72,614)-4 The cDNA was expressed in Escherichia coli, and the recombinant heavy subunit was purified.3 The latter is readily distinguishable from E. coli ~/-glutamylcysteine synthetase by making use of the fact that the kidney enzyme is markedly inhibited by cystamine. In contrast, the E. coli enzyme is not inhibited by this reagent. 5 The purified recombinant enzyme as well as the purified heavy subunit obtained from the isolated heterodimeric enzyme were found to exhibit much lower affinity for glutamate than the native enzyme.3 In addition, both the recombinant enzyme and the isolated heavy subunit were found to exhibit a much higher sensitivity to GSH inhibition than the holoenzyme. These findings indicate that the heavy subunit alone would not be significantly active under in vivo conditions. The GSH analog ~/-glutamyl-aaminobutyrylglycine (ophthalmic acid) was found to inhibit the isolated enzyme only slightly, but it inhibited much more effectively after treatment of the isolated enzyme with dithiothreitol. In contrast, ophthalmic acid inhibited the recombinant enzyme and the isolated heavy subunit substantially without dithiothreitol treatment. This supports the conclusions that the light subunit has a regulatory function that affects the affinity of the enzyme for glutamate and GSH, and that feedback inhibition by GSH involves reduction of the enzyme as well as competition between GSH and glutamate for the glutamate binding site of the enzyme. The cDNA for the light subunit of rat kidney ~/-glutamylcysteinesynthetase was isolated, sequenced, and expressed in E. coli. 6 The cDNA was found to code for a protein of 274 amino acid residues (Mr 30,548). Coexpression of the heavy and light subunit cDNAs in E. coli or separate expression of the cDNAs followed by mixing of the expressed proteins led to recombinant holoenzyme preparations that exhibit catalytic and GSH feedback inhibitory properties that are virtually identical to those shown by the holoenzyme isolated from rat kidney. These studies establish that the light subunit is an integral part of the enzyme and also that the light and heavy subunits are separately coded for. Rat kidney 3,-glutamylcysteine synthetase thus differs significantly in some of its properties from E. coli "/-glutamylcysteine synthetase. The E. coli enzyme is composed of a single polypeptide chain. Furthermore, as noted above, it is not inhibited by cystamine. Computer-aided comparison of the amino acid sequences of the heavy subunit and the E. coli enzyme showed a low degree of similarity,4 4N. Yan and A. Meister,J. Biol. Chem. 265, 1588 (1990). 5C.-S. Huang, W. Moore, and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 85, 2464 (1988). 6C.-S. Huang, M. E. Anderson, and A. Meister,J. Biol. Chem. 268, 20578 (1993).

28

GLUTATHIONE

[3]

but greater similarity was found between the light subunit of the kidney enzyme and the E. coli enzyme. 6 It is notable that the rat kidney enzyme and the E. coli enzyme have similar apparent Km values, turnover numbers, and specificity; both enzymes are inhibited by G S H and to about the extent. Neither enzyme is significantly inhibited by the G S H analog ophthalmic acid, but both enzymes are inhibited by ophthalmic acid when the enzymes are treated with dithiothreitol.

Inhibition of 3~-Glutamylcysteine S y n t h e t a s e b y B u t h i o n i n e Sulfoximine Buthionine sulfoximine (BSO) the most widely used inhibitor of G S H biosynthesis. The use of BSO for O

II CH3CH2CH2CH2SCH2CH2CHNH2COOH

il

NH Buthionine sulfoximine such inhibition developed from earlier studies on the inhibition of glutamine synthetase (glutamate-ammonia ligase) by methionine sulfoximine. Methionine sulfoximine, previously known as a convulsant agent, was found to bind to the glutamate binding site of glutamine synthetase and to undergo phosphorylation on the sulfoximine nitrogen atom by ATP. 7'8 Phosphorylation of the sulfoximine nitrogen atom is apparently analogous to the formation of ~/-glutamyl phosphate in the normal catalytic reaction. Binding of methionine sulfoximine phosphate to the active site inhibits the enzyme irreversibly. O f the four diastereomers of methionine sulfoximine, only the (2S, Ss) form is phosphorylated on the enzyme. 9'I° It is only this isomer that produces convulsions. Evidence that T-glutamyl phosphate is also an intermediate in the reaction catalyzed by 3~-glutamylcysteine synthetase led to examination of methionine sulfoximine as a possible inhibitor of the latter enzyme; it was found that "y-glutamylcysteine synthetase is irreversibly inhibited by the same isomer of methionine sulfoximine that inhibits gluta7 R. Ronzio and A. Meister, Proc. Natl. Acad. ScL U.S.A. $9, 164 (1968). s For a review, see A. Meister, Chemtracts: Biochem. Mol. Biol. 3, 75 (1992). 9 A. Meister, in "Glutathione Centennial: Molecular Perspectivesand Clinical Implications," (N. Taniguchi, T. Higashi, Y. Sakamoto, and A. Meister, eds.), Chap. 1, p. 3. Academic Press, New York, 1989. 10j. M. Manning, S. Moore, W. B. Rowe, and A. Meister, Biochemistry 8, 2681 (1969).

[3]

GLUTATHIONE BIOSYNTHESIS AND ITS INHIBITION

29

mine synthetase, n Administration of methionine sulfoximine to experimental animals leads to a decrease in the synthesis of glutamine and also that of GSH. Although such inhibition of GSH synthesis is effective, the usefulness of methionine sulfoximine in experimental studies is limited because it produces convulsions and has lethal effects that appear to be related to the inactivation of brain glutamine synthetase. Studies on the mapping of the active site of glutamine synthetase led to the conclusion that the S-methyl group of methionine sulfoximine binds to the ammonia binding site of the enzyme. 8 Sulfoximines with larger S-alkyl moieties were found not to inhibit glutamine synthetase significantly because they are hindered from binding to the ammonia binding site of this enzyme. However, such compounds effectively bind to and inactivate 7-glutamylcysteine synthetase. Thus, BSO does not inhibit glutamine synthetase significantly and thus decreases GSH levels without affecting the synthesis of glutamine (and does not produce convulsions). In addition to BSO, several other S-substituted homocysteine sulfoximine compounds inactivate 7-glutamylcysteine synthetase without affecting glutamine synthetase. These include S-(n-pentyl)homocysteine sulfoximine, which is similar in action to BSO. 12 The n-hexyl and S-(n-heptyl) compounds have been reported to be toxic. Other inhibitory sulfoximines were also found including S-2-methyl-n-butylhomocysteine sulfoximine and the corresponding 3-methyl, 3,3'-dimethyl, 2-ethyl, and cyclohexyl derivatives. 9 As first shown for methionine sulfoximine, only the (2 S, Ss) diastereomers of these compounds inactivate 7-glutamylcysteine synthetase. L-Buthionine (S,R)-sulfoximine is commercially available; it exhibits about 50% of the effectiveness of the pure L-(S)-isomer. BSO is itself not chemically reactive. L-Buthionine (R)-sulfoximine is useful as a control compound for possible detection of effects of BSO that are unrelated to its effect on GSH synthesis. For example, L-buthionine (R)-sulfoximine does not inhibit proliferation of normal human T lymphocytes, whereas L-buthionine (S,R)-sulfoximine d o e s ) 3 Other aspects of inhibition by BSO and its isomers have been discussed. 14,1sBSO is effective when administered to experimental animals orally or parentally. It is not readily transported across the blood-brain barrier, whereas BSO ethyl ester is more effectively 11 p. G. Richman, M. Orlowski, and A. Meister, J. Biol. Chem. 248, 6684 (1973). 12 O. W. Griffith, this series, Vol. 143, p. 286. ~3M. Suthanthiran, M. E. Anderson, V. K. Sharma, and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 87, 7185 (1990). 14 A. Meister, Pharmacol. Ther. 51, 155 (1991). 15 O. W. Griffith, in "Glutathione Centennial: Molecular Perspectives and Clinical Implications" (N. Taniguchi, T. Higashi, Y. Sakamoto, and A. Meister, eds.), Chap. 20, p. 285. Academic Press, New York, 1989.

30

6LUTATHIONE

[41

transported. 16 Inhibition of GSH synthesis by administration of BSO is currently under study for treatment of drug- and radiation-resistant tumors (see Meister14). Several other compounds that inhibit y-glutamylcysteine synthetase in vitro have been studied including cystamine, L- and D-3-amino-l-chloro-2pentanone, and L-2-amino-4-oxo-5-chloropentanoate (see Seelig and Meisterl), y-Methylene D-glutamate inactivates the enzyme by binding to an active-site thiol. 17 S-Sulfocysteine and S-sulfohomocysteine inhibit the enzyme by serving as transition-state analogs. 18 16 R. Steinherz, J. M~rtensson, D. Wellner, and A. Meister, Brain Res. 518, 115 (1990). 17 R. P. Simondsen and A. Meister, J. Biol. Chem. 261, 17134 (1986). 18 W. Moore, H. L. Wiener, and A. Meister, J. Biol. Chem. 262, 16771 (1987).

[4] G l u t a t h i o n e By

: Dehydroascorbate

WILLIAM W. WELLS, DIAN PENG Xu,

Oxidoreductases

and MICHAELP.

WASHBURN

Introduction The glutathione:dehydroascorbate oxidoreductases (EC 1.8.5.1) catalyze the following reaction: 2GSH + D H A ~ GSSG + AAH2 Dehydro-L-ascorbic acid (DHA), the first chemically stable oxidation product of L-ascorbic acid (AAH2), is generated in plant or animal cells from L-ascorbic acid by one- or two-electron oxidation reactions in which semidehydroascorbic acid ( A A H . ) may or may not be a required intermediate. The oxidation may proceed through interaction with superoxide free radical (02 ~) or hydroxyl radical (. OH), generated by H202 and 02 ~ interaction in the presence of Fe 3÷. Superoxide is produced by respiration and by a number of enzymes, for example, xanthine oxidase, aldehyde oxidase, and others. Oxidation of many compounds including hydroquinone, flavins, thiols, catecholamines, dialuric acid, herbicides, paraquat, and C C 1 4 produces 02 ~. In addition, D H A may be produced by a number of dioxygenases that use AAH2 as cofactors such as proline 4-hydroxylase.1 1 K. I. Kivirikko and R. Myllyl~i, this series, Vol. 144, p. 96.

METHODSIN ENZYMOLOGY,VOL.252

Copyright© 1995by AcademicPress,Inc. All fightsof reproductionin any formreserved.

[4]

GLUTATHIONE: DEHYDROASCORBATEOXIDOREDUCTASES

31

Regardless of whether D H A is derived by disproportionation of A A H . 2AAH. ~ DHA + AAH2 or whether A A H 2 reacts with H202 chemically or as a result of ascorbate peroxidase (plant) activity, A A H 2 + H202 -'+ 2 H 2 0 + D H A for normal cellular functions, it is essential to regenerate ascorbic acid from the D H A produced and to maintain a highly reduced A A H 2 / D H A set point. The reduction of D H A by glutathione (GSH), especially at an alkaline pH, is a well-recognized chemical reaction. 2 Szent-Gy6rgyi first demonstrated that the oxidized f o r m of vitamin C, D H A , could be regenerated to ascorbic acid by plant and animal tissue extracts that contained GSH. 3 In 1941, C r o o k characterized plant ascorbate oxidase activity and showed that it could be separated from the G S H - d e p e n d e n t D H A reductase catalyzed reaction. 4 The D H A reductase enzyme has been purified and characterized from spinach leaves. 5 The properties of glutathione : dehydroascorbate oxidoreductase, also known as glutathione dehydrogenase (ascorbate), f r o m spinach leaves was reported in this series by Stahl et al. 6 Following the early observations of Szent-Gy6rgyi, further evidence for animal D H A reductases was obtained in various tissues of the guinea pig by Hughes 7 and in h u m a n leukocytes by Christine et aL 8 During a search for alternative functions for thioltransferase, also known as glutaredoxin, this small protein, with thiol/disulfide transhydrogenase ( E C 1.8.4.1, glutat h i o n e - h o m o c y s t i n e transhydrogenase) and electron carrier properties in the ribonucleotide reductase system, 9 was found to have intrinsic D H A reductase activity. 1° In addition, bovine protein disulfide-isomerase (EC 5.3.4.1), but not m a m m a l i a n thioredoxin nor thioredoxin reductase, had D H A reductase activity as measured by the direct spectrophotometric procedure of Stahl e t al. 11 2 H. Borsook, H. W. Davenport, C. E. P. Jeffreys, and R. C. Warner, J. Biol. Chem. 117, 237 (1937). 3 A. Szent-GyOrgyi,Biochem. J. 22, 1387 (1928). 4 E. M. Crook, Biochem. J. 35, 226 (1941). 5 M. A. Hossain and K. Asada, Plant Cell Physiol. 25, 85 (1984). 6 R. L. Stahl, L. F. Liebes, and R. Silber, this series, Vol. 122, p. 10. 7 R. E. Hughes, Nature ( L o n d o n ) 203, 1068 (1964). s L. Christine, G. Thomson, B. Iggo, A. C. Brownie, and C. P. Stewart, Clin. Chim. Acta 1, 557 (1956). 9 A. Holmgren, Proc. Natl. Acad. Sci. U.S.A. 73, 2275 (1976). mW. W. Wells, D. P. Xu, Y. Yang, and P. A. Rocque, J. Biol. Chem. 265, 15361 (1990). it R. L. Stahl, L. F. Liebes, C. M. Farber, and R. Silber, Anal, Biochem. 131, 341 (1983).

32

GLUTATHIONE

I41

Assay Method

Principle A direct 11 and an indirect 12 method of measurement of D H A reductase activity may be chosen depending on the purity of the enzyme preparation. For relatively pure samples of dehydroascorbate reductase of animal cell origin (either thioltransferase or protein disulfide-isomerase), the basic direct spectrophotometric method described by Stahl et al. is preferred. 11 The activity is measured by following the change in absorbance at 265 nm as dehydroascorbic acid is reduced to ascorbic acid in a glutathionedependent manner according to the reaction 2GSH + DHA ~ GSSG + AAH2 In crude preparations of cell extracts, absorbance interference at 265 nm from nucleotides, etc., can be expected. Accordingly, the above reaction can be coupled with regeneration of GSH from oxidized glutathione (GSSG) by measuring the oxidation of NADPH in the presence of excess glutathione disulfide reductase as monitored at 340 nm spectrophotometrically.12

Dehydroascorbate Reductase Assay Reagents 20 mM GSH freshly prepared in Chelex-treated triple-distilled water 20 mM D H A freshly prepared in 10 mM sodium phosphate, 0.5 mM EDTA, pH 5.5 400 mM Sodium phosphate, 2 mM EDTA, pH 6.85 Preparation of Dehydroascorbic Acid. Prepare 20 mM DHA by bromine oxidation of a 20 mM solution of L-ascorbic acid dissolved in 10 mM sodium phosphate, 0.5 mM EDTA, pH 5.5. To 2 ml of the ascorbic acid solution in a graduated glass centrifuge tube, add 10/zl bromine (Aldrich, Milwaukee, WI, 99.99+%), stored over copper wire in a glass bottle sealed with Parafilm. After vortexing at room temperature for 30 sec, the excess bromine is sparged with nitrogen in a ventilation hood until the solution becomes colorless. The nitrogen sparging is continued for 10 min. Chelextreated triple-distilled water is added to adjust the volume to 2.0 ml. Solutions of DHA, prepared according to this procedure, stored on ice, and protected from light, are stable for at least 4 hr. Procedure. The standard reaction mixture contains 200 mM sodium phosphate buffer, pH 6.85, 1.0 mM EDTA, 3.0 mM GSH, and 1.5 mM 12E. I. Anderson and A. Spector, Invest. Ophthalmol.10, 41 (1971).

[41

GLUTATHIONE : DEHYDROASCORBATE OXIDOREDUCTASES

33

DHA with or without enzyme in a total of 500/zl. The increase in absorbance is recorded at 265 nm and 30° after initiating the reaction by the addition of DHA. The reaction, which is linear for up to 1 min is run simultaneously with a cuvette containing all reagents without the enzyme. A similar volume of buffer, used to store the enzyme, is substituted for the enzyme, and the resulting absorbance change is subtracted from that obtained with enzyme present. The millimolar extinction coefficient of ascorbic acid, reported by Daglish, is 14.7.13 Units. A unit of enzyme activity is defined as the amount of enzyme that reduces 1/xmol of L-dehydroascorbic acid in 1 min at 30°. Protein is determined by the bicinchoninic acid protein assay protocol according to the manufacturer's directions (Pierce, Rockford, IL) with bovine serum albumin as standard. Purification Procedures Two mammalian D H A reductases have been identified, m The cytoplasmic reductase, also known as thioltransferase or glutaredoxin, has been purified from numerous sources, including rat liver] 4 pig liver, 15 calf thymus, 16 rabbit bone marrow, 17 human placenta] 8 and human erythrocytes.19 A glutaredoxin homolog encoded by vaccinia virus was reported to be a virion-associated enzyme with thioltransferase and dehydroascorbate reductase activities.2°

Purification of Pig Liver Thioltransferase (Dehydroascorbate Reductase) Liver (2 kg) from 4- to 6-month-old pigs is used as starting material. 15 Except where indicated, all purification procedures are carried out at 4°. The liver is washed of as much blood as possible with buffer containing 10 mM HEPES, pH 7.5, 25 mM KC1, 5 mM MgC12, and 75/~g/ml phenylmethylsulfonyl fluoride (PMSF) and minced into small pieces. The same buffer is used to give a volume of 3200 ml, and the mixture is homogenized in portions with a Tekmar Tissumiser (Tekmar Co., Cincinnati, OH) at 70 V 13 C. Daglish, Biochem. J. 49, 635 (1951). 14 K. Axelsson, S. Ericksson, and B. Mannervik, Biochemistry 17, 2978 (1978). 15 Z.-R. Gan and W. W. Wells, Anal Biochem. 162, 265 (1987). 16 M. Luthman, S. Ericksson, A. Holmgren, and L. Thelander, Proc. Natl. Acad. Sci. U.S.A. 76, 2158 (1979). 17S. Hopper, R. S. Johnson, J. E. Vath, and K. Biemann, J. BioL Chem. 264, 20438 (1989). is K. Larson, V. Ericksson, and B. Mannervik, this series, Vol. 113, p. 520. 19j. j. Mieyal, D. W. Starke, S. A. Gravina, C. Dothey, and J. S. Chung, Biochemisry 30, 6088 (1991). 20 B.-Y. Ahn and B. Moss, Proc. Natl. Acad. Sci. U.S.A. 89, 7060 (1992).

34

GLUTATHIONE

[41

for 3-5 min. The homogenate is passed through two layers of cheesecloth, and a 2.4 M sucrose solution is added to give a final sucrose concentration of 0.24 M. The homogenate is centrifuged at 25,000 g for 40 min. The supernatant is adjusted to pH 5.4 with 5% acetic acid and centrifuged at 20,000 g for 40 min. The resulting supernatant fraction is adjusted to pH 7.0 with 6 N NaOH and is referred to as the high-speed supernatant. Step 1: Heating. The high-speed supernatant is heated by placing in a water bath, maintained at 70 °, with continuous stirring. When the supernatant reaches 62°, the suspension is immediately placed on ice and then centrifuged at 15,000 g for 30 min at 4 °. Step 2: Fractionation with Ammonium Sulfate. Ammonium sulfate is slowly added to the resulting supernatant fraction to give a final concentration of 40% (w/v). After standing on ice for 10 min, the suspension is centrifuged at 15,000 g for 30 min. Ammonium sulfate is added to the supernatant solution to give a final concentration of 65% (w/v), and the resulting suspension is centrifuged at 15,000 g for 30 min after standing on ice for 15 min. Step 3: Sephadex G-75 Gel Filtration. The pellet from the ammonium sulfate precipitation is diluted to a protein concentration of about 150 mg/ ml with water containing 2 mM dithiothreitol and loaded onto a Sephadex G-75 column (5 x 100 cm) which is equilibrated with 10 mM sodium phosphate buffer, pH 6.7. The flow rate is 1.5 ml/min, and the peak fractions are pooled. Thioltransferase activity appears after most of the protein in the Sephadex G-75 column. The salt fraction appears after the thioltransferase peak. Step 4: Sephadex G-50 Gel Filtration. The thioltransferase fractions from the Sephadex G-75 column are pooled (about 450 ml). The soluble proteins are concentrated by 70% ammonium sulfate precipitation. The concentrated proteins are incubated with 5 mM dithiothreitol (40 ml) for 15 min at room temperature, loaded on a Sephadex G-50 column (3 x 110 cm), and equilibrated and eluted with 10 mM sodium phosphate buffer, pH 6.1. Step 5: First CM-Sepharose Chromatography. Thioltransferase fractions (65 ml in 10 mM sodium phosphate buffer, pH 6.1) from the Sephadex G-50 gel filtration are incubated with 5 mM dithiothreitol for 15 min at room temperature, then loaded on a CM-Sepharose column (2.5 × 25 cm) equilibrated with 10 mM sodium phosphate buffer, pH 6.1. The column is first washed with 50 to 100 ml of 10 mM sodium phosphate buffer, pH 6.1, and the enzyme is eluted with 10 mM sodium phosphate buffer, pH 7.0. The thioltransferase fractions (about 30 ml) are pooled and lyophilized. Step 6: Second CM-Sepharose Chromatography. The lyophilized enzyme from the first CM-Sepharose step is incubated in a small volume of 10 mM dithiothreitol for 15 min at 30°. The incubation mixture is loaded on a Sephadex G-25 column equilibrated and eluted with 10 mM sodium phos-

[4]

35

GLUTATHIONE : DEHYDROASCORBATE OXIDOREDUCTASES TABLE I PURIFICATION OF PIG LIVER THIOLTRANSFERASEa

Purification step High-speed supernatant Heat and ammonium sulfate precipitation Sephadex G-75 Sephadex G-50 First CM-Sepharose Second CM-Sepharose

Protein (mg)

Thioltransferase activity (units)

111,750 35,280

10,060 9090

4000 461 81 24

7450 5650 3950 3220

Specific activity (units/rag) 0.09 0.26 1.9 12 49 134

Yield (%)

Purification (-fold)

100 90

1 3

74 56 39 32

21 133 544 1490

a Two kilograms liver from 4- to 6-month-old pigs was used as starting material. Reproduced from Gan and Wells 15 with permission.

phate buffer, pH 6.4. The thioltransferase fractions (1 ml per tube) are directly collected into test tubes which contain 0.1 ml of 20 mM 2-hydroxyethyl disulfide. Thioltransferase fractions are pooled and loaded on a CMSepharose column (1.5 x 28 cm) equilibrated with 10 mM sodium phosphate buffer, pH 6.4. The bound enzyme is first washed with 50 to 100 ml of 10 mM sodium phosphate, pH 6.4, and then eluted by a linear gradient formed by mixing 80 ml of 10 mM sodium phosphate buffer, pH 7.0, with 80 ml of 10 mM sodium phosphate buffer, pH 7.0, containing 0.1 M NaC1. The peak of thioltransferase activity elutes at a concentration of 45 mM NaC1. The thioltransferase fractions from the CM-Sepharose chromatography are lyophilized and reduced by 10 mM dithiothreitol before desalting on a Sephadex G-25 column equilibrated with 20 mM sodium phosphate, pH 7.5, and 20% (v/v) glycerol. The purified enzyme is shown to be homogeneous by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), isoelectric focusing, and high-performance liquid chromatography. The protein has an Mr of 11,740 and, in the reduced form, a pI of 6.4. A summary of the purification is shown in Table I. The recombinant pig liver enzyme has been purified after transformation and expression of the cDNA in Escherichia coli JM105. 21 Thioltransferase purification requires some specific modifications depending on its source. Expression of recombinant D H A reductase results in an 8.6% yield of the soluble protein after three to four simple steps, including heat and ammonium sulfate precipitation, Sephadex G-100 gel filtration, and a final CM-Sepharose ion-exchange chromatography step. 21 For human thioltransferase ( D H A reductase), the procedure of Mieyal et al. is recom2~ y. Yang and W. W. Wells, J. Biol. Chem. 265, 589 (1990).

36

GLUTATHIONE

141

TABLE II KINETIC PARAMETERS FOR MAMMALIAN DEHYDROASCORBATE REDUCTASES a

Parameter kcat (rain-a)d Km(app) (mM) e DHA GSH

Thioltransferaseb

Protein disulfide-isomerase c

374

16

0.26 3.4

2.80 2.9

2.4 × 104 1.8 × 103

1.0 x 102 0.9 × 102

k e a t / g m ( M -1 sec -1)

DHA GSH

The parameters were determined using either constant GSH (3.0 mM) or DHA (1.5 mM). Reactions were run spectrophotometrically in 0.5 ml at 30°, 200 mM sodium phosphate, pH 6.85, 1 mM EDTA, at 265 nm. All data were corrected for nonenzymatic reactions between DHA and GSH. b Recombinant pig liver thioltransferase,z° c Bovine liver protein disulfide-isomerase isolated according to a modification of the procedure of Yang and Wells.el d kcatvalues were calculated by dividing Vmax(app)by the molar concentrations of the two enzymes. e Km(app) values were calculated by nonlinear least-squares fit of the velocity versus substrate concentration data. a

m e n d e d . 19 I m p r o v e d y i e l d s o f h u m a n t h i o l t r a n s f e r a s e a r e o b t a i n e d w h e n t h e b u f f e r s f o r t h e l a t t e r s t e p s a r e s u p p l e m e n t e d w i t h 20% (v/v) glycerol. T h e e n z y m e is s t a b l e for o v e r 1 y e a r if s t o r e d at c o n c e n t r a t i o n s o f 1 0 - 2 0 m g / m l in 20 m M s o d i u m p h o s p h a t e , p H 7.5, a n d 20% (v/v) g l y c e r o l at - 7 0 °. Protein D i s u l f i d e - I s o m e r a s e T h e m i c r o s o m a l D H A r e d u c t a s e , t h a t is, p r o t e i n d i s u l f i d e - i s o m e r a s e , m a y b e p u r i f i e d f r o m b o v i n e liver to n e a r h o m o g e n e i t y b y a m o d i f i c a t i o n o f t h e m e t h o d o f H i l l s o n et aL 22 T h e e n z y m e is f u r t h e r p u r i f i e d b y 85% a m m o n i u m sulfate p r e c i p i t a t i o n f o l l o w e d b y dialysis a g a i n s t a b u f f e r cont a i n i n g 25 m M p o t a s s i u m p h o s p h a t e , p H 8.0, a n d 1 m M E D T A . T h e dialyz a t e is a p p l i e d to a c o l u m n (2.5 × 25 cm) o f D E A E - S e p h a d e x A - 5 0 equilib r a t e d w i t h t h e s a m e buffer. A f t e r w a s h i n g w i t h s e v e r a l b e d v o l u m e s o f t h e e q u i l i b r a t i o n buffer, a l i n e a r g r a d i e n t o f 0 - 0 . 7 M N a C I in t h e s a m e b u f f e r is a p p l i e d w i t h a t o t a l v o l u m e o f 500 ml. z3 T h e activity o f a p r e p a r a 22D. A. Hillson, N. Lambert, and R. B. Freedman, this series, Vol. 107, p. 281. 23 S. Bjelland, B. Foltmann, and K. Wallevik, Anal. Biochem. 142, 463 (1984).

[4]

GLUTATHIONE

CHzOH I

HO-C-H

.,c- 0 -po

H

Cz~C" OfO* + (DHA)+~X.H S-

E" "SH (OH)

: DEHYDROASCORBATE

CHz0H

HO-C-H " O-

37

OXIDOREDUCTASES

ICHz0H

HO-C-H " O

o

C-C W" O S

.c- T_-o

H C=C --

OH OH (AA)

.~,~I +GS-H E,SH

E"S-SG

(OH)

G S H ~ "~#H) ~.~

GSSG

E-S-

\SH (OH)

FIG. 1. Proposed scheme for the DHA reductase mechanism of thioltransferase. Reaction of the nucleophilicCys-22leads to a thiohemiketal intermediate. The reduced product, ascorbic acid (AA), is displaced by glutathione in the active mutant C25S to form a mixed disulfide that is further displaced by GSH to form GSSG and reduced enzyme. Alternatively, an intramolecular disulfide pathway can displace the reduction product, ascorbic acid, followed by exchange with 2 mol of GSH successively. [From W. W. Wells, Y. Yang, T. L. Deits, and Z.-R. Gan, in "Advances in Enzymology and Related Areas of Molecular Biology" (A. Meister, ed.), Volume 66, p. 149. Copyright © 1993 John Wiley & Sons, Inc. Reprinted by permission of John Wiley & Sons, Inc.] tion may be monitored by the insulin reduction assay of Holmgren, 24 the transhydrogenase activity essentially as described by Ibbetson and Freedm a n y or the reconstitution of scrambled ribonuclease assay of Hillson et a t 22 Bovine liver protein disulfide-isomerase concentrated to 10-20 mg/ml in 20 m M sodium phosphate, p H 7.5, and 20% (v/v) glycerol at - 7 0 ° is stable for several months. Analysis of the enzyme preparation by S D S P A G E show that it is free of material with a molecular weight of about 12,000. The exclusion of thioltransferase from the preparations is further verified by the absence of thioltransferase activity using the standard assay with S-sulfocysteine and GSH. 14,15 Properties The D H A reductase properties of thioltransferase and protein disulfideisomerase have been reported. 1° A kinetic study of both enzymes, using D H A prepared as described here, is shown in Table II. Although L-cysteine or cysteamine reduce D H A to A A H 2 chemically, neither thioltransferase nor protein disulfide-isomerase increase the rate of reduction (data not shown). The D H A reductase activity of thioltransferase is inhibited by 24A. Holmgren, J. Biol. Chem. 254, 9627 (1979). 25A. L. Ibbetson and R. B. Freedman, Biochem. J. 159, 377 (1976).

38

GLUTATHIONE

15]

iodoacetamide and other alkylating agents (data not shown) and by platinum complexes. 26 By using site-directed mutagenesis techniques, the essential amino acids at the catalytic center of porcine thioltransferase (dehydroascorbate reductase) were determinedY Seven mutant cDNAs were constructed, subcloned in plasmid pKK233-2, and transfected into E. coli JM105. The proteins were induced by isopropyl-l-thio-/3-D-galactopyranoside treatment at high yields. Activity comparisons showed that Cys-22 is essential for D H A reductase activity and that Cys-25 may participate in the reaction mechanism for the native enzyme; however, the latter is not essential, as substitution of Cys-25 with serine results in increased activity for D H A reductase. The proposed mechanism for the D H A reductase action of thioltransferase is shown in Fig. 1 and discussed elsewhere. 28 Maellaro et al. 29 have reported the purification and partial characterization of a glutathione-dependent dehydroascorbate reductase from rat liver that lacks thioltransferase activity. Acknowledgments This work was supported in part by Research Grants DK-44456 and CA-51972 from the National Institutes of Health. 26 W. W. Wells, P. A. Rocque, D.-P, Xu, Y. Yang, and T. L. Deits, Biochem. Biophys. Res. Commun. 180, 735 (1991). 27 y . Yang and W. W. Wells, J. Biol. Chem. 266, 12759 (1991). 2s W. W. Wells, Y. Yang, T. L. Deits, and Z.-R. Gan, in "Advances in Enzymology and Related Areas of Molecular Biology" (A. Meister, ed.), p. 149. John Wiley & Sons, Inc., New York, 1993. 29 E. Maellaro, B. Del Bello, L. Sugherini, A. Santucci, M. Comporti, and A. F. Casini, Biochem. J. 301, 471 (1994).

[51 D i v e r s i t y o f G l u t a t h i o n e By F.

Peroxidases

URSINI,M. MAIORINO,R. BRIGELIuS-FLoHI~,K. D. AUMANN, A. ROVERI, D. SCHOMBURG,and L. FLOHE

Introduction Selenium was identified as a toxic factor for grazing animals in the first half of the twentieth century and since then has been considered hazardous. 1

1 Committee on Medical and Biological Effects of Environmental Pollutants, "Selenium." National Academy of Science, Washington, D.C., 1976.

METHODSIN ENZYMOLOGY,VOL.252

Copyright© 1995by AcademicPress,Inc. All rightsof reproductionin any formreserved.

[5]

DIVERSITY OF GLUTATHIONE PEROXIDASES

39

The discovery by Schwartz and Foltz 2 that selenium was also an essential micronutrient in higher animals did not markedly change the prevalent view of selenium being a poison. Only long after the identification of the first selenoenzymes in bacteria 3"4 and mammals 5"6 was a Recommended Dietary Allowance gradually established] In fact, the putative biological roles of the selenoenzymes, in particular those of the glutathione peroxidases (GPX), proved instrumental in the understanding of selenium deficiency syndromes in livestock and humans, although the emerging complexity of selenium enzymology still precludes definitive conclusions (for review, see Refs. 8-10). The selenoproteins analyzed so far contain selenocysteine residues as integral parts of the peptide backbones. The selenocysteine is encoded by TGA, H'12 formerly known only as a terminating codon. The selenol group, owing to its low pK and high nucleophilicity, qualifies such selenoproteins for redox catalysis. Correspondingly, most bacterial and mammalian selenoproteins could be characterized as oxidoreductases, with the function of the heavily selenium-loaded protein P remaining unknown. 9"1° The first protein from higher vertebrates shown to be a true selenoprotein was GPX, 5 an unusual peroxidase discovered by Mills in 195713 and shown by Cohen and Hochstein TM to be of pivotal importance for the protection of red blood cells against oxidative damage. In the 1970s it had become common to ascribe the antioxidant effects, if not all biological roles, of selenium in mammals to GPX. In 1982, however, the search for further antioxidant proteins led to the discovery of a new "peroxidation

2 K. Schwarz and C. M. Foltz, J. Amer. Chem. Soc. 79, 3292 (1957). 3 D. C. Turner and T. C. Stadtman, Arch. Biochem. Biophys. 154, 366 (1973). 4 j. R. Andreesen and L. G. Ljungdahl, J. Bacteriol. 116, 867 (1973). 5 L. Floh6, W. A. Gtinzler, and H. H. Schock, F E B S Lett. 32, 132 (1973). 6 j. T. Rotruck, A. L. Pope, H. E. Ganther, A. B. Swanson, D. G. Hafeman, and W. G. Hoekstra, Science 179, 558 (1973). 7 O. A. Levander, in "Selenium in Biology and Medicine" (A. Wendel, ed.), p. 205. SpringerVerlag, Berlin, 1989. 8 L. Floh6, in "Glutathione: Chemical, Biochemical, and Medical Aspects" (D. Dolphin, R. Poulson, and O. Avramovic, eds.), p. 643. Wiley, New York, 1989. 9 T. C. Stadtman, Annu. Rev. Biochem. 49, 93 (1980). lo R. F. Burk and K. E. Hill, Annu. Rev. Nutr. 13, 65 (1993). 11 I. Chambers, J. Frampton, P. Goldfarb, N. Affara, W. McBain, and P. P. Harrison, E M B O J. 5, 1221 (1986). ~2F. Zinoni, A. Birkmann, T. C. Stadtman, and A. B6ck, Proc. Natl. Acad. Sci. U.S.A. 83, 4650 (1986). 13 G. C. Mills, J. Biol. Chem. 229, 189 (1957). 14 G. Cohen and P. Hochstein, Biochemistry 2, 1420 (1963).

40

GLUTATHIONE

[51

inhibiting protein," a5 now known as phospholipid hydroperoxide glutathione peroxidase (PHGPX). In contrast to the classic cytosolic GPX (cGPX), P H G P X is a monomeric protein of 19 kDa and is intriguing in its ability to reduce hydroperoxidases of esterified phospholipids and cholesterol in biomembranes. In 1987 a glycosylated extracellular GPX was isolated from plasma (pGPX) 16 and more recently a new type of GPX apparently prevalent in the gastrointestinal tract (giGPX) 17 was detected; the latter two, like cGPX, are tetrameric proteins. All the glutathione (GSH) peroxidases proved to be selenoproteins, all proved to be distinct gene products, and all proved to be homologous with one another but structurally and phytogenetically unrelated to the selenium-containing bacterial oxidoreductases and the other two sequenced mammalian selenoproteins, the type I thyronine deiodinase TM and protein p.19 In this chapter we compile the structural knowledge on various GPX types to provide a basis for discussion of the possible biological meaning of the structural diversification within this family of selenoenzymes. Molecular Evolution of the Glutathione Peroxidase Superfamily The selenium-dependent peroxidases have long been considered a late achievement of evolution, as they were only detected in vertebrates. This view now has to be revised. Not only is the structural relatedness between the various mammalian GPX types so low that the time point of evolutionary divergence has to be dated to a period long before the origin of mammals, 2°'21 but an obviously typical GPX gene has also been discovered in an invertebrate species, namely, S c h i s o t o s o m a m a n s o n i , z2 Further proteins not containing selenocysteine but being homologous with the mammalian GPX types have been identified in various vertebrate and invertebrate animals, in plants, and even in prokaryotes. Thus, the selenium-dependent 15F. Ursini, M. Maiorino, M. Valente, L. Ferri, and C. Gregolin, Biochim. Biophys. Acta 710, 197 (1982). 16 K. T. Takahashi, N. Avissar, J. Within, and H. Cohen, Arch. Biochem. Biophys. 256, 677 (1987). 17 F. F. Chu, J. H. Doroshow, and S. Esworthy, J. Biol. Chem. 268, 2571 (1993). 18 M. J. Berry, L. Banu, and P. R. Larsen, Nature (London) 349, 438 (1991). 19 K. E. Hill, R. S. Lloyd, J.-G. Yang, R. Read, and R. F. Burk, J. Biol. Chem. 266,10050 (1991). 20 R. Schuckelt, R. Brigelius-Floh6, M. Maiorino, A. Roveri, J. Reumkens, W. Strassburger, F. Ursini, B. Wolf, and L. Floh6, Free Radical Res. Commun. 14, 343 (1991). 21 R. Brigelius-Floh6, K. D. Aumann, H. B16cker, G. Gross, M. Kiess, K. D. KlOppel, M. Maiorino, A. Roveri, R. Schuckelt, F. Ursini, E. Wingender, and L. Floh6, J. Biol. Chem. 269, 7342 (1994). 22 D. L. Williams, R. J. Pierce, E. Cookson, and A. Capron, Mol. Biochem. Parasitol. 52, 127 (1992).

[5]

DIVERSITY OF GLUTATHIONE PEROXIDASES

41

oxidoreductases belong to a superfamily of proteins which in part can hardly or not at all be rated as peroxidases. Unfortunately, it is not straightforward to deduce precisely the molecular evolution of the superfamily from the percentage of misfits between aligned sequences, because the molecular diversity has in part been generated by intron/exon shuffling. The PHGPX enzyme, for instance, contains two partial sequences that are very similar to the corresponding ones in cGPXfl 1 They are interrupted by sequences unrelated or poorly related to cGPX. The highly conserved parts of the sequence are represented in the PHGPX gene by two of the seven exons which again are homologous to parts of the two exons of the cGPX gene. A dendrogram based on mutation statistics, if supposed to have any relevance to molecular evolution, can, however, only be constructed by comparing partial sequences that are unequivocally homologous. For the dendrogram of the GPX superfamily shown in Fig. 1,21-41 we chose the sequences corresponding to the third exon of the PHGPX gene, which meets this prerequisite. Obviously, the GPX superfamily branches early into several molecular clades. Only within 23 D. Lincoln, H. BlOcker, U. Leuner, W. Lehnberg, and G. Graf, "GENMON Version 4.1 Manual." Gesellschaft fiir Biotechnologische Forschung, Braunschweig, Germany, 1989. 24 S. Yoshimura, S. Takekoshi, K. Watanabe, and Y. Fuji-Kuriyama, Biochem. Biophys. Res. Commun. 154, 1024 (1988). 25 y. Ho, A. J. Howard, and J. D. Crapo, Nucleic Acids Res. 16, 5207 (1988). 26 W. A. Gtinzler, G. J. Steffens, A. Grossmann, S.-M. A. Kim, F. Otting, A. Wendel, and L. Floh6, Hoppe-Seyler's Z. Physiol. Chem. 365, 195 (1984). 27 K. Ishida, T. Morino, K. Takagi, and Y. Sukenaga, Nucleic Acids Res. 15~ 10051 (1987). 28 M. Akasaka, J. Mizoguchi, S. Yishimura, and K. Watanabe, Nucleic Acids Res. 17, 2136 (1989). 29 R. A. Sunde, J. A. Dyer, T. Moran, J. K. Evenson, and M. Sugimoto, Biochem. Biophys. Res. Commun. 193, 905 (1993). 30 R. A. Sunde, J. A. Dyer, T. V. Moran, and J. K. Evenson, unpublished; EMBL Acc. No. L 24896, 1994. 31 F. F. Chu and S. R. Esworthy, unpublished, EMBL Acc. No. X 71973, 1994. 32 M. C. Criqui, E. Jamet, Y. Parmentier, J. Marbach, A. Durr, and J. Fleck, Plant Mol. Biol. 18, 623 (1992). 33 D. Holland, G. Ben-Hayyim, Z. Falin, L. Camoin, A. D. Strosberg, and Y. Eshdat, Plant Mol. Biol. 21, 923 (1993). 34 K. Takahashi, M. Akasaka, Y. Yamamoto, C. Kobayashi, J. Mizoguchi, and J. Koyama, J. Biochem. (Tokyo) 108, 145 (1990). 35 j. M. Martin-Alonso, S. Gosh, and M. Coca-Prados, J. Biochem. (Tokyo) 114, 284 (1993). 36 H. Hatta and T. Moriuchi, Biochem. J. 109, 918 (1991). 37 A. C. Perry, R. Jones, L. S. Niang, R. M. Jackson, and L. Hall, Biochem. J. 285, 863 (1992). 38 N. B. Ghyselinck and J.-P. Dufaure, Nucleic Acids Res. 18, 7144 (1990). 39 T. N. Dear, K. Campbell, and T. H. Rabbitts, Biochemistry 30, 10376 (1991). 40 M. J. Friedrich, L. C. DeVeaux, and R. J. Kadner, J. Bacteriol. 167, 928 (1986). 41 E. Cookson, M. L. Baxter, and M. E. Selkirk, Proc. Natl. Acad. Sci. U.S.A. 89, 5837 (1992).

r

Lr] II .IJ

tO E

4..I

oJ F-q

x

C. .iJ

L

E

x" X

m h

X El. L9

X n O

X El. t9

X n (D

X rl t.O

X rl (-9

X n tD

X n C.9

X 13_ ~

X n CD

0J

~

-I O

IT

4-J

0J

C

01

17

CL

13.

.~

.O .O ro

.,4 >

I= ~

C ~I

C~ .r4

-IJ ~0

C ~

O 0'~

o

T

e

O_

rr

~

c

>

IT

03

~

D

tO

T X n

X n (.~

X 13_ CD

X 13_ (.~

X n (-9

tO C

f0 I=

C" .,4

tO n"

~

C

T

0

tO

ffl

~ ~ ~

~

7

~

T X 12.

"r" X 12.

T x 13_

~

~

0J

~ to

4a ro

U to m

~

rr

T X 12.

T X 13_

ZC X 12.

I0 r/-

,-4 O u

10 o

~--

~"

QJ Z

FIG. 1. Dendrogram of the glutathione peroxidase superfamily calculated from conserved amino acid residues of aligned protein sequences according to Ref. 23 applying a gap penalty of 5. The pertinent sequences are as follows: mouse GPX, u rat cGPX, 24'25 bovine GPX, 26 human cGPX, 27 rabbit cGPX, 2s human giGPX, 17 porcine P H G P X , 21'29 rat P H G P X , 3° human P H G P X , 31 Schistosoma m a n s o n i GPX, 22 G P X homolog ( G P X H ) of Nicotiana sylvestris, 32 G P X H from Citrus sinensis, 33 human p G P X , 34 bovine p G P X , 3s rat p G P X , 36 the epididymal G P X homolog ( e G P X H ) of mouse, 37 rat, 38 and Macaca fascicularis, 38 another rat G P X H , 39 the cobalamine-binding protein of Escherichia coli, 4° and a nematode G P X H . 41

[5]

DIVERSITY OF GLUTATHIONE PEROXIDASES

43

a distinct type of the protein does the molecular similarity reasonably correspond to the relatedness of the species. Three major clades can be clearly delineated from the dendrogram (Fig. 1) and can be even more precisely defined when the common molecular characteristics obvious from the sequence alignment (see below and Fig. 3) are considered: (i) the cGPX family with its side branch giGPX; (ii) pGPX together with GPX homologs in which the selenocysteine residue is replaced by cysteine, both characterized as secreted proteins by typical signal peptides, and (iii) the PHGPX family. Whether the common ancester of the GPX superfamily was a selenoprotein or a cysteine-containing homolog cannot be deduced from the available sequences. The only prokaryotic member of the superfamily detected so far, a cobalamine-binding protein of Escherichia coli, 4° does not contain selenocysteine, and despite ongoing efforts, functionally active glutathione peroxidases have not yet been found in prokaryotes. This might indicate that, within the GPX superfamily, the use of T G A as a codon for selenocysteine was not achieved except during eukaryotic evolution. This view, however, conflicts with the observation that selenocysteine in bacterial selenoproteins is also encoded by TGA, 42 although the mechanism for the differential use of T G A as codon for selenocysteine or stop differs in proand eukaryotes. In bacteria selenocysteine incorporation requires a defined stem-loop in the immediate neighborhood of the U G A triplet in the mRNA coding region, 43,44whereas in eukaroytes proper selenocysteine incorporation is guaranteed by a selenocysteinyl insertion sequence (SECIS) in the 3' untranslated tail of the mRNA, its distance from the U G A codon being largely irrelevant. 45 In each case, the formation of a selenoprotein requires the fit of the m R N A structure comprising the U G A codon for selenocysteine and a SECIS and special decoding factors plus the availability of a selenocysteinyl-tRNA with an appropriate anticodon. Despite basic similarities, however, the mechanisms of selenoprotein biosynthesis of pro- and eukaryotes diverged to an extent that a eukaryotic gene encoding a selenoprotein, if expressed in bacteria, inevitably yields a truncated peptide resuiting from chain termination at the U G A codon. The puzzle of how the coevolution of pertinent genes sustained the ability of organisms to build selenoproteins remains to be solved. In view of the multiple ways of being evolutionary deleted and the limited chance of being reinvented, seleno42 A. BOck, K. Forchhammer, J. Heider, and C. Baron, Trends Biochem. Sci. 16, 463 (1991). 43 F. Zinoni, J. Heider, and A. B/)ck, Proc. Natl. Acad. Sci. U.S.A. 87, 4660 (1990). 44 G. E. Garcia and T. C. Stadtman, J. Bacteriol. 173, 2093 (1991). 45M. J. Berry, L. Banu, J. W. Harnex, and P. R. Larsen, E M B O J. 12, 3315 (1993).

44

GLUTATHIONE

[5]

proteins maintained up to mammals should deserve more than curiosity, as their mere existence stresses their biological importance.

S t r u c t u r e a n d F u n c t i o n of G l u t a t h i o n e P e r o x i d a s e s C o m m o n Features

Bovine c G P X has been extensively investigated functionally. 46-49 With the identification of c G P X as a selenoprotein 5° and the determination of its amino acid sequence 26 and three-dimensional (3D) structure, 51 a rather detailed mechanism of action of the enzyme could be proposed 52 which, however, needs to be revised (see below). As far as p G P X and P H G P X are concerned, we still have to rely on functional and structural homology with cGPX. In functional terms the three types of G P X appear very similar. In all cases the reaction catalyzed is identical in principle [Eq. (1)]. The kinetic 2 R S H + R O O H ~ RSSR + H 2 0 + R O H

(1)

pattern is characterized as a Ping-Pong mechanism with indefinite Michaelis constants, indefinite maximum velocities, 46'47'53-55 and negligible, if any, product inhibition. It can be described by the simple Dalziel equation [Eq. (2)]. This kinetic mechanism is routinely explained by a sequence of [Eo! _ ~1 ~2 + - v [ROOH] [RSH]

(2)

bimolecular reactions of the substrates with the enzyme (E) or derivatives (F, G) thereof [Eqs. (3)-(5)]. Comparison of Eqs. (3)-(5) with the Dalziel 46 L. Floh4, O. Loschen, W. A. Giinzler, and E. Eichele, Hoppe-Seyler's Z. Physiol. Chem.

353, 987 (1972). 47W. A. Giinzler, H. Vergin, I. MUller, and L. Floh6, Hoppe-Seyler's Z. Physiol. Chem. 353, 1001 (1972). 48L. Floh6, W. A. Giinzler, G. Jung, E. Schaich, and F. Schneider, Hoppe-Seyler's Z. Physiol. Chem. 352, 159 (1971). 49L. Floh6 and W. A. Giinzler, in "Glutathione" (L. Floh& H. C. BenOhr, H. Sies, H. D. Waller, and A. Wender, eds.), p. 132. Thieme, Stuttgart, 1974. 5oL. Floh6, W. A. Gtinzler, and H. H. Schock, FEBS Lett. 32, 132 (1973). 51O. Epp, R. Ladenstein, and A. Wendel, Eur. J. Biochem. 133, 51 (1983). 52R. Ladenstein, O. Epp, W. A. Giinzler, and L. Floh4, Life Chem. Rep. 4, 37 (1986). 53F. Ursini, M. Maiorino, and C. Gregolin, Bioehim. Biophys. Acta 839, 62 (1985). 54M. Maiorino, A. Roved, C. Oregolin,and F. Ursini, Arch. Biochem. Biophys. 251,600 (1986). 5s R. S. Esworthy, F.-F. Chu, P. Geiger, A. W. Girotti, and J. H. Doroshow, Arch. Bioehem. Biophys. 3ff7, 29 (1993).

[5]

DIVERSITY OF GLUTATHIONE PEROXIDASES

45

E + ROOH

k G + H20 k+~

(4)

G + RSH

k:~* E + RSSR

(5)

equation [Eq. (2)] obtained empirically yield the definitions of the Dalziel coefficients: qbl -

%

] k+l

1 -

k+3 + k+5

(6)

(7)

With the questionable exception of rabbit cGPX 56 for which a tendency toward substrate saturation kinetics at extreme GSH levels has been described, Eqs. (3)-(5) would be satisfactory to describe the kinetic mechanisms of all selenoperoxidases analyzed so far. These minimum assumptions, however, are incompatible with substrate-specificity studies. Although distinct differences in the generally low specificity for the hydroperoxide substrates do exist between the different GPX types, they may be explained by a modified accessibility of the catalytic site and, in view of the lacking saturation kinetics, do not justify implication of the intermediate formation of a specific complex between E and ROOH. Despite the lack of saturation kinetics, the pronounced specificity of cGPX for GSH, however, indicates a complex formation between the donor substrate and the oxidized enzyme forms F and G [Eqs. (8) and (10)], before reactions (9) and (11), respectively, can proceed: F + RSH

(F. RSH)

(8)

(F-RSH)

k:~ G + H20

(9)

G + RSH ~ (G.RSH)

k*4

(G-RSH)

~,:5 E + RSSR

(10) (11)

The rate constants k+2 and k+4 are obviously much smaller than k+3 and k+5, respectively, which means that the rate constants for complex formation, k+2 a n d k+4, become rate limiting and approximate k+ 3 and k+5 in the overall reactions [Eqs. (4) and (5)]. The lack of any detectable substrate saturation further implies that regeneration of the ground state E via Eqs. (10) plus 56 p. Morrissey and P. J. O'Brian, Can. J. Biochem. 58, 1012 (1980).

46

GLUTATHIONE

[51

(11) must proceed faster than the formation of the intermediate G according to Eqs. (8) and (9). Deviations of the ratios of rate constants k+2/k'+3, k+4/ k~5, o r k÷2/k+4may be envisaged in certain GPX species. If substantial, they might affect the kinetic pattern without, however, invalidating the basic assumptions explaining the catalytic cycle. Common to all types of glutathione peroxidases to be discussed here is the involvement of selenium in the catalysis. [The so-called nonselenium G P X 57 is a structurally unrelated GSH thioltransferase which mimicks the overall reaction (1) by an entirely different mechanism.] Their biosynthesis strictly depends on the availability of selenium.They become inactivated if their selenocysteines are carboxymethylated by i o d o a c e t a t e 49,58,59 o r if the selenium is eliminated. 6° Further, no relevant peroxidase activity could so far be attributed to any of the sulfur-containing homologs of the superfamily.

Putative Molecular Mechanism of Bovine Cytosolic Glutathione Peroxidase Although bovine cGPX may be considered to be one of the best investigated proteins, some uncertainties with respect to its molecular mechanism remain. These are predominantly due to the kinetics of the enzyme, which make it theoretically impossible to freeze any of the postulated complexes or intermediates (F. GSH), G, and (G. GSH). The groundstate E and the first intermediate F are not easily investigated either because of stability problems. X-ray studies, for instance, aimed at the analysis of the ground state yielded the 3D structure of an enzyme form in which the selenocysteine was oxidized by air to a seleninic acid derivative. 51 This oxidation state most probably can be attributed to a sluggish enzyme derivative which does not become fully active until after prolonged incubation with thiols. 8,61 Removal of the two oxygens from the selenium atom in a molecular model did not, however, essentially alter the 3D structure after energy minimization and molecular dynamics calculations (K. D. Aumann, unpublished). We may thus consider the X-ray structure of Epp et aI.,51 though performed on an overoxidized enzyme, as a reliable basis for our attempts to understand the essential molecular properties of the various enzyme forms implicated in the catalytic cycle. Clearly, the ground state of the enzyme (E) must be characterized by 57 R. F. Burk and R. A. Lawrence, in "Functions of Glutathione in Liver and Kidney" (H. Sies and A. Wendel, eds.), p. 114. Springer-Verlag, Berlin, 1978. 58 R. J. Kraus, J. R. Prohaska, and H. E. Ganther, Biochim. Biophys. Acta 615, 19 (1980). 59 M. Maiorino, C. Gregolin, and F. Ursini, this series, Vol. 186, p. 448. 60 H. E. Ganther and R. J. Kraus, this series, Vol. 107, p. 593. 61 C. Little, R. Olinescu, K. G. Reid, and P. J. O'Brian, J. Biol. Chem. 245, 3632 (1970).

[5]

DIVERSITY OF GLUTATHIONE PEROXIDASES

47

having its selenocysteine residue in the reduced form, that is, as a dissociated selenol. This is obvious (i) from the selective susceptibility of the donor substrate-reduced enzyme by iodoacetate 49 and (ii) from the speed of its reaction with hydroperoxides 46'47which could hardly be expected from any other functional groups of this protein. The selenium is located in the center of a flat impression composed of hydrophobic residues. 51 It is fixed in this position from the interior of the protein core by a glutamine and a tryptophan residue, each forming hydrogen bridges with the selenium (see Figs. 2a and 5a). 51 In the catalytic triad of selenocysteine, glutamine, and tryptophan, a complete dissociation of the selenol is certainly enforced and its nucleophilic attack on the peroxide is facilitated. Because no other redox active residue is available in the immediate neighborhood of the selenol and a free radical mechanism has been ruled out, 6a we have to assume that the selenium has to change its oxidation state from - 2 to 0 on collision with the peroxide. A simple bimolecular reaction between the selenol and the peroxide, mediated by a transition state rather than by a specific enzyme-substrate complex, is also suggested by the speed of the reaction, the broad range of hydroperoxides accepted at comparable rates, 8 and the lack of any structural pecularities reminiscent of a substrate binding pocket. 5~ The next step, the formation of the complex (F. GSH) according to Eq. (8), has to comply with the strict specificity of cGPX for GSH. Further, in this complex the sulfur of GSH has to be directed into a position to enforce a fast, non-rate-limiting intramolecular reaction yielding the intermediate G [Eq. (9)]. Calculations 62a show that these requirements may be met in the following way (see Fig. 2). The glycine carboxylic group of GSH binds electrostatically to the positively charged guanidino group of Arg57; Arg-184 forms a hydrogen bridge with the carbonyl of the cysteine residue of GSH; another hydrogen bridge is built between the amino group of the y-glutamyl residue of GSH and the carbonyl of Gly-53, whereas the carboxylic group remains loosely oriented toward a positively charged area composed of Arg-103 and by Lys-91 of the neighboring subunit (Lys-91B). In this complex the sulfhydryl group of GSH is in an ideal position to react with the oxidized selenium of the active site. In the resulting intermediate G (see Fig. 2b), the glutathione molecule remains essentially bound as in the preceding complex. It should be mentioned that this proposal for G obtained by molecular modeling 63 differs from that originally discussed by Epp e t al. 51 in which the carboxylic groups 62 L. S. Harman, D. K. Carver, J. Schreiber, and R. P. Mason, J. Biol. Chem. 261, 1642 (1986). 62a K.-D. Aumann, D. Schomburg, and L. Flohd, in preparation. 63 L. Flohr, K.-D. Aumann, R. Brigelius-Flohr, D. Schomburg, W. Strassburger, and F. Ursini, in "Active Oxygens, Lipid Peroxides and Antioxidants" (K. Yagi, ed.), p. 299. Japan Scientific Societies Press, Tokyo/CRC Press, Boca Raton, Florida, 1993.

48

GLUTATHIONE

[51

of GSH formed salt bridges with the Arg-57 and Arg-184 residues. By this mode of binding, however, the GSH molecule would be forced into a less relaxed Y-shaped conformation; correspondingly, the energy minimization calculations refused such a solution. Also, in the original proposal the active site selenodisulfide bridge of G was shielded by the glutathionyl residue in a way which would probably prevent any attack of the second GSH molecule. 8 In contrast, the model presented here leaves the selenodisulfide bridge freely accessible. As outlined above, the regeneration of E from G by GSH proceeds faster than the formation of G. This is again achieved by directing the second GSH molecule into an adequate position. Molecular modeling combined with calculation of energy minimization and molecular dynamics yields a proposal for (G. GSH), which again is essentially electrostatic (Fig. 2c). Apart from some hydrogen bridges the second GSH molecule forms with the first one and with some residues of the enzyme including those of the neighboring subunits, the predominant binding forces are the positive charges of Arg-184 now binding to the glycine carboxyl of the second GSH and of Arg-103 plus Lys-91B attracting the T-glutamyl carboxyl of the substrate. In an alternative model (not shown) Arg-185 turns around to compete with Arg-184 for the cosubstrate. In both models the S- of GSH can split the selenodisulfide bridge to generate the product oxidized glutathione (GSSG) and to close the catalytic cycle. Lack of product inhibition finally may be explained by the conformation of GSSG assumed after dissociation of the enzyme-product complex. In an aqueous environment, the four carboxylic groups of GSSG tend to repel one another electrostatically. Thus, a double Y-shaped conformation twisted by 90° around the disulfide bond is favored. Consequently, the product no longer fits into the frame of positive charges surrounding the catalytic site of the enzyme and its affinity is substantially lowered. The mechanism proposed here is consistent with all experimental data available such as X-ray analysis, sl kinetic patterns, 46'4~ inhibition, 49 and substrate specificities.48 It should also be stressed that the residues invoked to play a major role in the catalytic cycle are conserved in all cGPX species sequenced so far (see Fig. 3). It may therefore be stated that the ideas summarized here provide a reasonably realistic proposal for the first example of an enzymatic selenium catalysis. Structural Peculiarities of Individual Glutathione Peroxidases Despite a low incidence of identities in the aligned sequences, molecular models of pGPX and P H G P X could be developed, 63 As the core of the modeled proteins was not changed substantially on simulated energy expo-

blue bails, respectively. For sake of clarity the sizes of these balls are much smaller than the van der Waals radii (30%). Residues belonging to the second subunit of the dimer model are shown in beige. (a) "Empty" ground state of the enzyme (E) essentially corresponding to the X-ray structure. 51 (b) Model for the catalytic intermediate G, in which GSH by its sulfur (yellow) is covalently bound to the selenium. The negatively charged carboxylic groups of GSH noncovalently bound to the arginine and lysine residues are shown as red balls. (c) The complex of G with the second GSH molecule (pink). Note how precisely the sulfur of the second GSH is directed toward the selenadisulfide bridge of G. All computer graphics were obtained by means of the BRAGI program (GBF, Braunschweig, Germany).

FIG. 3. Alignment of available sequences of the GPX superfamily. Pertinent references are given in the legend to Fig. 1. The residues forming the catalytic triad in GPX, selenocysteine (position 52 in the bovine cGPX sequence), glutamine-87, and tryptophan-165 are marked in red and other highly conserved residues in beige. The arginine residues 57, 103, 184, and 185 and lysine-91 (position numbers of bovine cGPX) involved in GSH binding by cGPX are shown in blue. Green defines the signal peptides of pGPX and secreted GPX homologs. Signal peptide processing sites are either taken from the corresponding references or estimated according to the algorithm proposed by von Heijne (Nucl. Acids Res. 14, 4683, 1986). HHH and 13~il3mean a-helices and pleated sheets, respectively, in the bovine cGPX structure.

O

S~

iqil

g.

0

•

Q~IZt

:S:I= :1:

O0 r~lZ

m ~

~

~m~0J :l: I! Q

. f : I~l

,-

i/

Fro. 4. Comparison of the Ccz carbon backbone of a bovine cGPX subunit51 with other members of the cGPX superfamily. (a) Residues of bovine cGPX (red) which are conserved in homologous positions in all members of the superfamily. Note that these residues accumulate around the active-site selenium (orange ball) in the pleated sheets and turns. (b) Overlay of the bovine cGPX structure and a model of human pGPX. Here the pleated sheets are marked in green, and the cz helices in red. Deviations of the pGPX model from the cGPX structure are shown in yellow. (c) Overlay of the PHGPX model with cGPX. Note the substantial deletion of the random loop and of the small c~ helix left of the active center.

FIG. 5. Comparison of the active site structure of bovine cGPX 51 with models of human pGPX and porcine PHGPX. (a) Active site of cGPX with Arg-57 (lower left comer), Arg-103 (lower right

comer) Arg-184 (top), and Arg-185 (upper left comer) in red, Gln-87 in lilac, Trp-165 and Phe-167 in yellow, and selenocysteine in orange. The size of the selenium (big ball) corresponds to its van der Waals radius. Residues supposed to be functionally important are stressed with bigger sticks. (b) Overlay of cGPX in the same colors and human pGPX in pink. Note that Arg residues 57 and 184 are replaced by Gin and His, respectively. (c) Overlay of cGPX with the model of porcine PHGPX (pink). Note that all arginines invoked in substrate binding are missing.

[5]

DIVERSITY OF GLUTATHIONE PEROXIDASES

49

sure in molecular dynamics calculations, the models may be considered fairly realistic. According to the molecular models, the homology of the sequences finds its counterpart in a conservation of the basic structure of the core of the proteins. Figure 4a shows the Ca backbone of a bovine cGPX subunit, 51 which demonstrates where the residues conserved in the entire superfamily are located. Interestingly, conserved residues are particularly frequent in the intimate environment of the active site selenium (or sulfur, respectively). This highly conserved part of the 3D structure is built up by three distant protein loops of the primary sequence, namely, NVA..Sec (or C)..T (positions 47-59 in the bovine cGPX sequence), L.FPCNQF..Q (positions 7787), and WNF (positions 165-167). The conserved clusters comprise the residues Gin, (seleno)Cys, and Trp which appear to represent the catalytic triad, qualifying the enzymes for a more or less efficient redox catalysis depending on the presence of selenium or sulfur (see Fig. 5a for more details). Further highly conserved residues, for example, those in the pleated sheets and Gly-176 forming a turn, are considered responsible for the conservation of the global structure of the molecules, which is obviously indispensible for the formation of the particular catalytic triad. The structural similarity of the two families of cytosolic and plasma GPX is demonstrated in Fig. 4b, showing an overlay of the experimentally determined Ca backbone of bovine cGPX and the model of human pGPX. There is virtually no difference whatsoever in the arrangement of the wellordered secondary structural elements, namely, the four pleated sheets and the four helices. However, a small shortcut in the large loop (right-hand side of Fig. 4) is seen in the pGPX structure. Because this loop forms the subunit interaction site of cGPX, the deviation might slightly affect the affinity of the subunits. Calculations of the subunit interaction forces of pGPX, 63 however, confirm a tetramerization, which is in agreement with the apparent Mr of pGPX. The differences between P H G P X and cGPX are more pronounced (Fig. 4c). Again, the shapes of the core of these enzyme families comprising the pleated sheets and three of the helices appear largely identical. However, the deletions in the P H G P X sequences cause a substantial shortcut of the subunit interaction loop plus an elimination of a small helix (left site) also involved in subunit interaction in cGPX. Thus, both the interfaces responsible for dimerization (loop) and tetramerization (helix) in cGPX are destroyed, which is consistent with the monomeric nature of PHGPX. In the P H G P X model, because of the deletions and the resulting monomeric nature, the active site appears more freely accessible. This could provide a reasonable explanation for the ability of P H G P X to attack more complex hydroperoxy lipids including those integrated into lipid layers.

50

GLUTATHIONE

[51

Comparing the active sites of cGPX with more remote relatives poses rather than answers questions. Figure 5a shows a closeup of the active site of bovine cGPX as derived from the X-ray study of Epp et al. 5a The four arginines in positions 57, 103, 184, and 185 surrounding the catalytic triad and postulated to contribute to the correct orientation of the donor substrates are conserved in all cGPX species sequenced so far (Fig. 3), but one, Arg-185, is replaced by Thr in the most closely related offspring of that family, giGPX. Also, Lys-91B contributing to substrate binding is exchanged by an uncharged amino acid residue (Gin). In the pGPX family, of the five basic residues implicated only Arg-103 and Arg-185 are conserved (Fig. 3), and any isosteric charged substitute for Arg-57 and Arg184 facilitating binding of the first GSH molecule in cGPX is not detectable in the pGPX model, as evident from the overlay shown in Fig. 5b. In the P H G P X family, finally, none of the charged residues is retained in homologous positions (Fig. 3), nor does the P H G P X model suggest any isosteric and similarly charged substitute which could be envisaged to bind GSH (Fig. 5c). For the GPX from Schistosoma mansoni, however, which is the only nonvertebrate GPX sequence known so far, one could envisage a functional substitution of Arg-57 and Arg-184 by lysine residues (Fig. 3). From the alignment and the molecular models one might deduce that the donor substrate specificity declines in the order cGPX > giGPX > pGPX > PHGPX. Also, any attempt to generate a plausible model of the complex (F. GSH), as described for cGPX above, has so far failed in the case of pGPX and PHGPX. Nevertheless, these enzymes have been discovered as functionally active glutathione peroxidases, and despite considerable efforts no physiological donor substrate acting faster than GSH has been found so far. For pHGPX, however, considerable activity has been detected with dithiols like lipoic acid and dithiothreitol (DTT), 64 which are poor substrates for cGPX. Also, the limiting rate constant of the reducing part of the catalytic cycle, that is, probably k+2 of Eq. (8) (see above), has consistently been determined to be one order of magnitude smaller for both pGPX 55 and P H G P X 59 than for cGPX. 47'48'55 This finding is consistent with the view that, in the pGPX and P H G P X family, GSH as donor substrate is less precisely forced into a suitable position to react with the active site selenium. In both cases, it may be reasonably questioned whether the real physiological donor substrates have yet been identified. 64 M. Maiorino, in "Biological Oxidants and Antioxidants; New Development in Research and Health Effects" (L. Packer and E. Cadenas, eds.), p. 69. Hyppocrates Verlag, Stuttgart, 1994.

[5]

DIVERSITY OF GLUTATHIONE PEROXIDASES

51

Apart from the questions related to the physiological donor substrates of pGPX and PHGPX, further peculiarities reveal a functional diversification within the GPX superfamily. The substitution of selenocysteine by cysteine in the catalytic triad of many of the members is incompatible with the assumption of all enzymes being peroxidases in the sense of efficient peroxide removers. Unfortunately, however, few functional studies on the cysteine-containing GPX homologs are available. For the homolog in E. coli, a role as a cobalamine-binding protein has been demonstrated, n° in citrus plants the homolog is induced by salt stress, 33 and the epidydimal homologs respond to androgens. 38

Functional Diversification of Selenoperoxidases Ample evidence reveals that cGPX, in concerted action with the pentose phosphate cycle and glutathione reductase, efficiently metabolizes intracellular H202, particularly in ceils or cellular compartments low in or devoid of catalase/As evident from the early studies of Cohen and Hochstein 14and the k+a values in the range of 108 M -1 s e c 1,46,47 cGPX can also successfully compete with catalase as long as the rate of GSH regeneration can cope with the generation or influx of H202. 8 The broader range of accepted hydroperoxides renders cGPX even more competitive in the defense against oxidative damage. In agreement with early in vitro experiments with isolated membranes, 65 it could also be envisaged that cGPX contributes to the prevention of lipid peroxidation in biomembranes if the latter is initiated by soluble hydroperoxides. To this end, the cGPX system can be rated as a highly efficient and indispensible constituent of cellular devices protecting against oxidative stress. Gastrointestinal GPX, being structurally very similar and functionally indistinguishable from cGPX, has to be seen in the same context. Whether the scope of biological functions of these most typical glutathione peroxidases is definitively described by "defense against soluble hydroperoxides" should, however, be left unanswered. Historically, P H G P X was detected and exclusively discussed as being an indispensible device to prevent unspecific lipid peroxidation within biomembranes. 15 In this respect, it is indeed the substitute for cGPX, to which this function was erroneously attributed. 65 In a concerted action with vitamin E, P H G P X breaks the chain reactions leading to oxidative lipid destruction in the following way: peroxy radicals generated by addition of molecular oxygen on hydrogen abstraction from unsaturated phospholipids are 65 L. Floh6 and R. Zimmerman, in "Glutathione" (L. Floh6, H. C. Bent~hr, H. Sies, H. D. Waller, and A. Wendel, eds.), p. 245. Thieme, Stuttgart, 1974.

52

GLUTATHIONE

[5]

first reduced by vitamin E to the corresponding hydroperoxides; these are reduced by P H G P X with the consumption of GSH; thereby, the decomposition of the hydroperoxides into alkoxy radicals and reinitiation of free radical chains are prevented. 66,67 Again, the scope of P H G P X function appears inadequately described by this narrow model. In vitro, cGPX plus GSH was shown to block the activity of the key enzymes of prostaglandin and leukotriene biosynthesis, the cyclooxygenase6s and the 5-1ipoxygenase. 69 However, in carefully designed experiments taking advantage of the differential response of cGPX and P H G P X to alimentary selenium supply, P H G P X was identified as the more likely candidate to regulate leukotriene biosynthesis in v i v o . 70 Beyond this observation, a preferentially regulatory role of P H G P X is further suggested by its overall low tissue levels, its extremely unusual tissue distribution with prevalence in reproductive and endocrine tissues, 71 susceptibility to posttranslational modifications, 2°,21 change of subcellular location during spermatogenesis, strong gonadotropin dependency of its biosynthesis, 72 and multiple putative hormone-responsive elements in the P H G P X gene. 2a Also, the complete elimination of any GSH-binding residues in the active site of P H G P X and the low specificity for GSH allows one to question its nature as a real glutathione peroxidase. The flat hydrophobic surface surrounding the catalytic site of P H G P X would allow an intimate interaction not only with lipid layers but also with proteins. Thus, oxidation equivalents from some peroxides could be transferred via oxidized selenium to protein thiol groups of functional relevance. It is even more tempting to speculate along these lines in case of pGPX. Although the active site of pGPX with two of the typically four arginine residues being conserved is still reminiscent of a real GPX, an antioxidant function has to be ruled out for this enzyme in view of the extremely low extraceUular GSH levels, the lack of any efficient extracellular GSH regenerating system, and the comparatively low rate constants for GSH. 17 Extracellular GSH levels in the micromolar range could just serve to keep pGPX in a reduced state under undisturbed physiological conditions. An oxidative

66 M. Maiorino, M. Coassin, A. Roveri, and F. Ursini, Lipids 24, 721 (1989). 67 F. Ursini, M. Maiorino, and A. Sevanian, in "Oxidative Stress: Oxidants and Antioxidants" (H. Sies, ed.), p. 319. Academic Press, London, 1991. 68 M. A. Warso and W. E. Lands, J. Clin. Invest. 25, 667 (1985). 69 M. Haurand and L. Floh6, Biol. Chem. Hoppe-Seyler 369, 133 (1988). 70 F. Weitzel and A. Wendel, J. Biol. Chem. 268, 6288 (1993). 71 A. Roveri, M. Maiorino, and F. Ursini, this series, Vol. 233, p. 202. 72 A. Roveri, A. Casasco, M. Maiorino, P. Dalan, A. Calligaro, and F. Ursini, J. Biol. Chem. 267, 6142 (1992).

[6]

GLUTATHIONE TRANSFERASE PURIFICATION

53

burst of polymorphonuclear leukocytes or macrophages could keep the active site selenium oxidized long enough to allow a transfer of the oxidant signal onto thiol groups on distant cellular surfaces.

Conclusion In conclusion, a comparison of the structures of pGPX and PHGPX with that of cGPX should open our eyes to new possible roles of these more recently discovered GPX congeners far beyond the concept of antioxidant defense. The "atypical" members of the GPX family at least will find their biological role more likely in the realm of redox regulation, for example, for the release of lipid mediators 7°,73 and cytokines, TM signal transduction, 73 and NF-~:B-mediated transcriptional control. 75'76 73 S. Orrenius, M. J. Burkitt, G. E. Kass, J. M. Dypbukt, and P. Nicotera, Ann. Neurol. 32, 33 (1992). 74 Fo U. Schade, R. Engel, S. H~irtling, J. Holler, and D. Jakobs, lmmunobiology 187, 283 (1993). 75 T. Okamoto, H. Ogivara, T. Hayashi, A. Mitsui, T. Kawabe, and J. Yodoi, Int. ImmunoL 4, 811 (1992). 76 N. Israel, M. A. Gougerot-Pocilado, F. Aillet, and J. L. Virelizier, Z lmmunol. 149, 3386 (1992).

[6] P u r i f i c a t i o n

By D A V I D

of Soluble Human S-Transferases

Glutathione

J. M E Y E R a n d BRIAN KETTERER

Introduction Soluble glutathione transferases (GSTs) (EC 2.5.1.18) catalyze glutathione-dependent nucleophilic substitution and addition reactions with endogenous and xenobiotic electrophiles as well as the reduction of hydroperoxides, quinones, and quinoneimines. 1They are intracellular dimeric enzymes (molecular weight approximately 50,000), representing four related gene families named alpha, mu, pi, and theta. 2 Both homodimers and heterodimers may occur within but not between gene families. 1 B. Ketterer and L. G. Christodoulides, Adv. Pharrnacol. 27, 36 (1994). 2 B. Mannervik, Y. C. Awasthi, P. G. Board, J. D. Hayes, C. Di Ilio, B. Ketterer, I. Listowsky, R. Morgenstern, M. Murarnatsu, W. R. Pearson, C. B. Pickett, K. Sato, C.-P. D. Tu, M. Widersten, and C. R. Wolf, Biochem. J. 281, 305 (1992).

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

54

GLUTATHIONE

[61

General Considerations Glutathione transferases of classes alpha, mu, and pi but not theta have a high affinity for glutathione (GSH) conjugatesJ As a result affinity chromatography with immobilized GSH conjugates, such as amino-linked S-hexylglutathione-agarose 4 or sulfur-linked glutathione-agarose, 5 enables the isolation in one step of a fraction containing alpha, mu, and pi class enzymes. Whereas the former also binds glyoxalase II and minor amounts of uncharacterized proteins of approximately Mr 35,000, the latter retains only GSTs. Subsequent hydroxylapatite and anion-exchange chromatography steps enable the purification of the major human GSTs (AI-1, A1-2, A2-2, Mla-la, Mla-lb, M3-3, and PI-1) as described below. The published purifications of human 0 class GSTs, which do not bind to GSH affinity matrices, namely, GSTTI-1 and GSTT2-2, have been described, 6'7 but because they yield small amounts of unstable enzyme, they are not given in this chapter. The same problems arise with respect to rat theta class GST5-5. In this case the problem has been overcome by the use of a heterologous bacterial expression system. This procedure should be applicable to the human homolog GST-I?I-1 and is described below. All the procedures described give a high yield (50-90%); therefore, detailed analyses of purification are not included. Our laboratory has not isolated GSTM2-2 or the heterodimers GSTM1-2 or GSTM2-3, which are relatively abundant in skeletal muscle. A pr6cis of the method used by Hussey et al. 8 to isolate GSTM2-2 is given below. Readers are referred to the literature for less studied enzymes, such as GSTM4-4 from brain 9 and GSTs omega (co) 1 and 2 from liverJ ° Purification of Glutathione Transferases A I - I , AI-2, A2-2, and M1 from Liver Normal human liver contains the three a class GSTs A1-I, A1-2, and A2-2; GST M1 has two expressing alleles G S T M I * A and G S T M I * B and 3 D. J. Meyer, Xenobiotica 23, 823 (1993). 4 C. Guthenberg and B. Mannervik, Biochem. Biophys. Res. Commun. 86, 1304 (1979). 5 p. C. Simons and D. L. Vander Jagt, this series, Vol. 77, p. 235. 6 D. J. Meyer, B. Coles, S. E. Pemble, K. S. Gilmore, G. M. Fraser, and B. Ketterer, Biochem. J. 274, 409 (1991). 7 A. J. Hussey and J. D. Hayes, Biochem. J. 286, 929 (1992). 8 A. J. Hussey, L. A. Kerr, A. D. Cronshaw, D. J. Harrison, and J. D. Hayes, Biochem. J. 273, 323 (1991). 9 T. Suzuki, D. C. Shaw, and P. G. Board, Biochem. J. 274, 405 (1991). 10H. Ahmad, S. S. Singhal, M. Saxena, and Y. C. Awasthi, Biochim. Biophys. Acta 1161, 333 (1993).

[61

GLUTATHIONE TRANSFERASE PURIFICATION

55

the null allele GSTMI*O. GSTM1 expression occurs in approximately 60% of the population. Of these a small proportion are GSTM1 A,B heterozygotes. GST subunits Mla and Mlb differ in only one amino acid residue conferring a single charge difference, resulting in distinct chromatographic properties. The presence of GSTMI-1 can be most easily tested by measuring soluble GSH transferase activity toward trans-4-phenyl-3-buten-2-one. 11 Materials Equipment. The following equipment is needed: Waring blendor (or similar tissue homogenizer/food processor); ultracentrifuge; low-pressure chromatography column (15-30 cm × 5 cm i.d.); recording spectrophotometer; Pharmacia (Piscataway, N J) FPLC (fast protein liquid chromatography) system or equivalent; Bio-Rad (Richmond, CA) hydroxylapatite (HPHT) column and guard column or low-pressure chromatography column (25-35 cm × 1 cm i.d.); Pharmacia Mono Q anion-exchange column or equivalent; Sephadex G-25 columns (e.g., PD-10 from Pharmacia); and vacuum filtration apparatus to remove particulates from FPLC buffers. Chemicals. Required reagents are as follows: phenylmethylsulfonyl fluoride (PMSF); dithiothreitol (DTT); reduced glutathione (acid); S-hexylglutathione (acid); S-linked glutathione-agarose; polyethylene glycol (molecular weight 15,000-20,000); dialysis tubing; 1-chloro-2,4-dinitrobenzene (recrystallized from ethanol); Tris base (high purity), piperazine hydrate: glycerol (low UV absorbing); and hydroxylapatite (HTP grade) for lowpressure chromatography from Bio-Rad should an HPHT column be unavailable. Water for all solutions should be glass-distilled and deionized. Solutions

Buffer 1 is required for tissue homogenization and affinity chromatography. It contains 0.2 M NaC1, 2 mM dithiothreitol, 2 mM EDTA, 20 mM sodium phosphate, pH 7.0, 25 /~M phenylmethylsulfonyi fluoride. The last is added to the stirred solution from a syringe fitted with a 27-gauge needle containing a 25 mM stock solution in 2-propanol just prior to use Buffer 2, used for elution of GSTs from the affinity matrix, is also prepared just prior to use as follows: to 5 mM S-hexylglutathione in 0.1 mM Tris base is added 5 mM glutathione and the solution adjusted to pH 9.0 at 25° (i.e., pH 9.6 at 4°) Buffer 3, required for hydroxylapatite chromatography, contains 5% (w/v) glycerol, 0.5 mM dithiothreitol, 10 mM sodium phosphate, pH 6.8 11 W. H. Habig and W. B. Jakoby, this series, Vol. 77, p. 398.

56

GLUTATHIONE

[61

Buffer 4 is for the gradient elution from hydroxylapatite and contains 5% (w/v) glycerol, 0.5 mM dithiothreitol, 250 mM sodium phosphate, pH 6.8 Buffer 5, required for anion-exchange FPLC of alpha class GSTs, is composed of 10% (w/v) glycerol, 1 mM dithiothreitol, 30 mM piperazine hydrochloride, pH 9.55. The dithiothreitol is added after adjusting the pH. The buffer is made 0.2 M NaC1 for the gradient elution Buffer 6, for anion-exchange FPLC of mu class GSTs, contains 10% (w/v) glycerol, 1 mM dithiothreitol, 20 mM Tris-HC1, pH 7.8. The dithiothreitol is added after adjusting the pH. The buffer is made 0.3 M NaCI for gradient elution Buffer 7, used for the storage of GSTs, is 10% (w/v) glycerol, 50 mM NaC1, 0.2 mM dithiothreitol, 5 mM sodium phosphate, pH 7.0. All buffers used for FPLC are vacuum-filtered through glass micro fiber filters and 0.2-/.~m pore nitrocellulose.

Tissue Extraction and Affinity Chromatography Safety Note. Care should be taken in the handling and disposal of human tissue extracts in order to avoid exposure to pathogens. Vaccination against hepatitis B is advised. Starting Material. The routine preparation of 2-8 mg amounts of hepatic GSTs is as follows. Normal liver, obtained from either an organ donor unit or mortuary as soon as possible postmortem, is cut into approximately 1.5cm cubes and either used immediately or frozen in liquid N2 and stored at - 7 0 °. The sample chosen for illustration here was HL135, from Tennessee Donor Services (Nashville, TN), which in addition to GSTs A1-1, A1-2, and A2-2 contained the products of both the GSTMI*A and GSTMI*B alleles, namely, GSTMla-la, GSTMlb-lb, and GSTMla-lb. Day 1. Twenty grams of frozen liver is added to sufficient (~200 ml) buffer 1 to cover the blades of a Waring blendor. When partially thawed, the tissue is homogenized for 15 sec at low speed and 10 sec at high speed. The homogenate is centrifuged at 105,000 g for 1 hr and the supernatant poured through precooled glass wool to remove lipid. Meanwhile, 40 ml S-linked glutathione-agarose in a 5-cm diameter column is cooled and equilibrated by passing buffer 1 at 00-4 ° through it at approximately 1.5 ml/min under gravity. The soluble supernatant fraction is applied followed by 100-150 ml of buffer 1 and the effluent disposed of. Thirty milliliters of buffer 2 is applied to the column followed by more buffer I and fractions of approximately 6 ml collected. The elution of GSTs is followed by determining the optical absorption at 280 nm using buffer 2 as a blank (Fig. 1). It may be misleading to measure GST activity at this stage because the

[6]

GLUTATHIONE

TRANSFERASE

PURIFICATION

57

A28o 1,0

A2eo BUFFER 2 0

I

0

I

I

I

t

I

5o

VOL BUFFER 2

I (ml)

FIG. 1. Glutathione-agaroseaffinitychromatography.After application of the soluble fraction from 18 g liverto a 40-mlcolumnof glutathione-agaroseand washing,the GST pool (~25 mg) was eluted with buffer2.

S-hexylglutathione present is a potent inhibitor. The GST pool (-25 ml) is dialyzed twice against 1.5 liters unfiltered buffer 3 at 00-4 °. The affinity column is cleaned with 20 ml of 6 M guanidinium hydrochloride pH 7, and stored in 3 M NaCI; it may occasionally need repacking. The next step usually requires hydroxylapatite FPLC; however, if instead a low-pressure hydroxylapatite column is to be used, it is freshly prepared and equilibrated using approximately 20 g hydroxylapatite HTP in a column 1.5-2 cm in diameter. The column outlet is adjusted to maintain pressure at less than 25 cm of liquid. Day 2. The dialyzed GST pool is concentrated to 4-5 ml by laying the dialysis bag on a bed of cold dry polyethylene glycol, and buffers 3 and 4 for hydroxylapatite FPLC are made ready. Following equilibration with buffer 3, the HPHT column is washed with a 0-100% gradient of buffer 4 (15 ml) at 0.3 ml/min. Pressure should be applied to the HPHT column gradually, to avoid matrix compression and consequent low flow rates; the guard column is particularly sensitive in this respect. The FPLC steps are carried out at room temperature. The concentrated GST pool is applied to the HPHT column (2.5-ml portions using a 3-ml sample loop of PTFE tubing) at 0.3 ml/min, and the flow through, which contains a small amount of GSTPI-1, is collected. The mu and alpha class GST fractions are then eluted with a gradient of 0100% buffer 4 in about 30 ml (Fig. 2). Fractions (~0.7 ml) are collected in polystyrene tubes and kept on ice.

58

[61

GLUTATHIONE

ALPHA CLASS

A2ao

t MU CLASS I

/

/'/

I

/

GSTMla-la I

//

A/

0

0

/ !

10 20 GRADIENT VOL (ml)

|

--

. 10

FIG. 2. Separation of liver GSTs using hydroxylapatite. The GST pool (Fig. 1) was separated into alpha and mu classes using a Bio-Rad HPHT column fitted to an FPLC system at a flow rate of 0.35 ml/min. The flow-through fraction contains small amounts of GSTPI-1 and M3-3.

If low-pressure hydroxylapatite chromatography is used instead of FPLC, the concentration step is omitted and the column is run at 0°-4 ° with a 300 ml gradient at about 0.4 ml/min (day and night of day 2). The pooled GST fractions are then concentrated on day 3. Day 3. The alpha class GSTs are separated as follows. The FPLC Mono Q anion-exchange column is cleaned with 1 column volume of 70% acetic acid and equilibrated with 1 column volume of buffer 5 followed by 1.5 column volumes of filtered 2 M NaC1 and finally 5 column volumes of buffer 5. An effluent pH monitor may be used to ascertain equilibration. The alpha class GST pool is transferred into buffer 5 by applying 2 ml to each of three PD-10 (Sephadex G-25) columns, equilibrated in buffer 5, followed by 0.5 ml buffer 5; 3 ml buffer 5 is then added and the 3 eluates collected into a polypropylene syringe for application to the Mono Q column. Note that this procedure yields a more stringent removal of phosphate ions than would be achieved using the manufacturer's guidelines. The flow through should not contain GST. The AI-1, A1-2, and A2-2 enzymes are eluted with a linear gradient of buffer 5 containing NaC1. To obtain satisfactory separation when using the Mono Q HR5/5 column, less than 6 mg protein should be applied per run. Larger amounts may be separated using the HR10/10 column (Fig. 3). Because buffer 5 has a relatively high pH, after separation the fractions are transferred to buffer 7 using PD-10 columns

I6]

59

GLUTATHIONE TRANSFERASE PURIFICATION GSTA1-2

A28o

GSTA2-2

1.0

GSTA1-1

/

Y 0.1

/ / / / /

z / 0

--

I. /

,z 0

I 5O

I

--O

100

GRADIENT VOL (ml}

FIG. 3. Separation of major c~ class GSTs, The liver c~ class GST fraction (Fig. 2) was separated into GSTs AI-I, A1-2, and A2-2 using anion-exchange FPLC (Mono Q HR10/10) in buffer 5 at flow rate of 1.2 ml/min.

prepared in advance. Immediately after this exchange, the GST fractions are placed on ice. The GSTMI-1 fraction from hydroxylapatite chromatography is usually almost pure. If separation of Mla-la, Mla-lb, and Mlb-lb is required, the mu class GST pool is transferred to buffer 6 using PD-10 columns and the fraction applied to the Mono Q HR5/5 column. The flow-through fraction contains GSTMla-la. Elution with buffer 6 containing 0.3 M NaC1 (Fig. 4) yields first Mla-Mlb followed by Mlb-Mlb. Fractions are collected in plastic tubes on ice and transferred immediately into buffer 7 for storage. The purity and concentration of the GSTs are determined by reversedphase high-performance liquid chromatography (HPLC) and the activity determined (see below). If the GSTs are to be used within 3 weeks, they are kept on ice. For longer storage, small samples are transferred to polypropylene tubes and snap-frozen in liquid nitrogen. However, up to 30% loss of enzyme activity may occur on freezing and thawing. Yield. Liver GST content varies considerably. 12'13A GST-rich, GSTM1positive liver may yield 4-8 mg GSTA1 and A2 and 1-2 mg GSTM1 per 10 g tissue. 12B. van Ommen, J. J. P. Bogaards, W. H. M. Peters, B. Blaauboer, and P. J. van Bladeren, Biochem. Z 269, 609 (1990). 13 B. Ketterer, D. J. Meyer, E. Lalor, P. Johnson, F. P. Guengerich, L. M. Distlerath, P. E. B. Reilly, F. F. Kadlubar, T. J. Flammang, Y. Yamazoe, and P. H. Beaune, Biochem. Pharmacol. 41, 635 (1991).

60

GLUTATHIONE

[6]

[NaCI] (M)

Aa80

1.0

--

0.15

/ /

/ill// J

t~

m

~

0

,

,,/

t

I

10 Gradient vol (ml)

0

FIG. 4. Separation of GSTs Mla-la, Mla-lb, and Mlb-lb. The/z class GST fraction from the hydroxylapatite step (Fig. 2) was separated into allelic forms by anion-exchange FPLC (Mono Q HR5/5) in buffer 6 at a flow rate of 0.3 ml/min.

Purification of Glutathione Transferase P I- 1 from Kidney The human kidney is a good source of GSTPI-I but also contains alpha class GSTs together with small amounts of mu class GSTs, including M33. The procedure is very similar to that for the liver GSTs (see above) with the following modifications. The soluble fraction from 40 g of kidney is applied to the 40-ml affinity column. If the t~ class GSTs are not required, the hydroxylapatite step is omitted and the total GST pool brought to pH 7.4 by dropwise addition of 0.5 M sodium acetate, pH 6.3, and concentrated. If a hydroxylapatite column is used, the flow-throdgh fraction is concentrated. The GSTPI-I is then purified, either from the total GST pool or the hydroxylapatite flow-through fraction, by anion-exchange FPLC in buffer 6 exactly as described for the separation of GSTM-I (Fig. 5). Yield. Ten grams kidney should yield 2 - 4 mg GSTPI-I. Purification of Glutathione Transferase M3-3 from Testis The human testis contains high levels of alpha class GSTs, GSTM3-3, and GSTPI-I, together with smaller amounts of other mu class GSTs. The

[6]

GLUTATHIONE TRANSFERASE PURIFICATION

61

GST P1-1

A28o

/

I

z

/ /

0.2

0,5 z ¢

/

/

,"

GSTM3-3

Z

I

0

I •

l

l

10 GRADIENT VOL (ml)

20

FIG, 5. Purification of GSTP-1, The kidney GST pool was applied to a hydroxylapatite column and GSTPI-1 separated from the flow-through fraction by anion-exchange FPLC exactly as described for liver GSTMI-1 (Fig. 4). Small amounts of pure GSTM3-3 are also obtained.

GSTM3-3 may be purified by minor modification of the method for the purification of hepatic GSTs. The total GST pool, obtained by affinity chromatography, is dialyzed against buffer 3 modified to contain 50 mM rather than 10 mM sodium phosphate. The resulting flow-through fraction then contains GSTPI-1 and GSTM3-3 as major components, together with lesser amounts of other mu class GSTs. After concentration, this fraction undergoes anion-exchange FPLC in buffer 6, exactly as described above. If pure alpha class GSTs are no t required, the hydroxylapatite step can be omitted. In this case, the total GST pool is simply neutralized with 0.5 M sodium acetate (see above), concentrated, transferred into buffer 6, and applied to the Mono Q column (Fig. 6). Yield. Approximately 2 mg GSTM3-3 should be obtained per 10 g testis.

Purification of Glutathione Transferase M2-2 from Skeletal Muscle The concentration of GSTs in skeletal muscle is in the region of 4% that of the liver. However, because skeletal muscle accounts for 40% of body mass, the total hepatic and skeletal muscle GST activities are similar. Skeletal muscle is a very important component of extrahepatic GST activity. Glutathione transferases in skeletal muscle mainly comprise GSTs PI-1

62

[61

OLUTATHIONE GSTM3-3 CO

A2eo /

~

GSTPI-1

0

10

i i"

2O

z

0.15

0

GRADIENT VOL (ml) FIG. 6. Purification of GSTM3-3. The whole testis GST pool was fractionated by anionexchange FPLC (Mono Q HR5/5) under the conditions of Figs. 4 and 5. The major peak between GSTPI-1 and M3-3 is a heterodimer (M2-M3). Without prior hydroxylapatite separation, the GSTPI-1 peak is not pure.

and subunits M1 (in subjects with a G S T M I * G S T M 1 genotype), M2, and M3 occurring in both homo- and heterodimeric forms. Hussey et al. 18 have purified GSTM2-2 (referred to by them as N2N2) from skeletal muscle using similar techniques to those used above. A recommended procedure based on their published work is described below. Solutions

Buffer a is used for homogenization and affinity chromatography on S-hexylglutathione-agarose. The composition is 0.2 M NaC1, 0.5 mM dithiothreitol in 20 mM Tris-HC1 at pH 7.8. Elution from the affinity column is brought about with buffer a containing 5 mM S-hexylglutathione (the pH must be readjusted after addition of S-hexylglutathione) Buffer b is used for gradient elution by FPLC anion exchange using Mono Q columns. It contains 20 mM Tris-HC1 at pH 8.4, 0.5 mM dithiothreitol. Elution is in two stages using a linear gradient from 0 to 0.15 M NaC1 Buffer c for gradient elution from hydroxylapatitite chromatography is 10 mM sodium phosphate, pH 7.0, containing 0.5 mM dithiothreitol. Elution is brought about by a 10-350 mM phosphate gradient Procedure. To purify GSTM2-2 it is advisable to use muscle from subjects with a GSTM1 null phenotype, that is, muscle with activity toward trans-4-phenyl-3-buten-2-one 11 of less than about 0.04 nmol/min/mg protein. This avoids complications arising from the presence of M l a and M l b homodimers and their heterodimers with one another or with GSTM2

[6]

GLUTATHIONE TRANSFERASE PURIFICATION

63

and GSTM3. The purification procedure is carried out at 00-4 ° unless otherwise stated. For a final yield of approximately 1 mg M2-2, 300 g skeletal muscle is homogenized in 900 ml buffer a and clarified soluble supernatant obtained as described above. After dialysis against three changes of 5 liters of buffer a over 24 hr, the GSTs are separated by application to 60 ml S-hexylglutathione-agarose in a 5-cm-diameter column equilibrated in buffer a. After washing with 200 ml buffer a, the GSTs are eluted with buffer a containing 5 m M S-hexylglutathione and monitored at A280. The G S T pool (~40 ml) is dialyzed twice against 2 liters buffer b over 24 hr and concentrated to approximately 5 ml using polyethylene glycol. The sample is applied at 0.3 ml/min to an FPLC Mono Q column equilibrated in buffer b at room temperature and GSTs eluted with a gradient of 0-0.15 M NaC1 in buffer b. Two major peaks absorbing at 280 nm are obtained. The first to elute is a mixture of GSTs PI-1 and M2-2, whereas the second peak contains the GSTM2-3 heterodimer. Fractions from the first peak containing GSTM2-2 and GSTPI-1 are pooled and transferred into buffer c at room temperature and applied to a column of hydroxylapatite (Bio-Rad H P H T or low-pressure column of 1 x 10 cm; see above). The GSTPI-1 is not retained, and GSTM2-2 is eluted with a gradient of sodium phosphate.

Purification of T h e t a Class G l u t a t h i o n e T r a n s f e r a s e s E x p r e s s e d in E s c h e r i c h i a coli

The theta class enzymes have proved difficult to purify from rat and human tissue fully active and in high yield. 6a In the case of rat GST5-5 this has been overcome by its heterologous expression in Escherichia coli, 14 which has facilitated a simplified, high yield procedure. The methods used for purifying theta class GSTs from both rat and human tissues are very similar; therefore, it is anticipated that the simplified procedure will also be successful for the purification of rat GST12-12 and human GSTs TI-1 and T2-2 when bacterially expressed material becomes available. (The c D N A sequence of GSTT1 has already been obtained, is E q u i p m e n t . N e e d e d equipment includes an ultracentrifuge, low-pressure chromatography column 50 x 2.5 cm (i.d.), Pharmacia F P L C system or equivalent, and a Mono Q anion-exchange column. 14R. Thier, J. B. Taylor, S. E. Pemble, W. Griffith Humphreys, M. Permark, B. Ketterer, and F. P. Guengerich, Proc. Natl. Acad. Sci. U.S.A. 90, 8576 (1993). 15S. E. Pemble, K. R. Schroeder, S. R. Spencer, D. J. Meyer, E. Hallier, H. M. Bolt, B. Ketterer, and J. B. Taylor, Biochem. J. 300, 271 (1994).

64

GLUTATHIONE

[6]

Chemicals. The following reagents are required: Matrex gel Orange A (Amicon, Danvers, MA); DNase I; phenylmethylsulfonyl fluoride; dithiothreitol; polyethylene glycol (molecular weight 15,000-20,000); dialysis tubing; piperazine hydrate; and glycerol (low UV absorbing). Solutions

Buffer A used for homogenization contains 25/xM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 10 mM sodium phosphate, pH 7.0 Buffer B for dialysis and Matrex gel Orange A chromatography contains 1 mM dithiothreitol, 1 mM EDTA, 5% (w/v) glycerol, 10 mM sodium phosphate, pH 7.0 Buffer C used for gradient elution from Matrex gel Orange A is buffer B made 1 M with respect to KC1 Buffer D for cleaning the Matrex gel Orange A is 6 M guanidinium hydrochloride adjusted to pH 7.0 with Tris base Buffer E used for guanidinium ion removal is 0.4 M sodium phosphate pH 7.0 Buffer F used for anion-exchange FPLC contains 5% (w/v) glycerol, 30 mM piperazine-HC1, pH 9.55, with 1 mM dithiothreitol added after pH adjustment. NaCI is added to 0.2 M for gradient elution. The FPLC buffers are filtered through glass microfiber and 0.2-/zm cellulose nitrate filters TABLE I SPECIFIC ACTIVITIESOF PURIFIED GLUTATHIONETRANSFERASES GST enzyme

Specific activity" (tzmol/min/mg protein)

A1-1 A2-2 M1-1 M2-2 M3-3 P1-1 R a t 5-5

130 130 250 171 b 15 140 220 c

a M e a s u r e d at 37 ° with 1 m M 1-chloro-2,4-dinitrobenzene and 1 m M glutathione at p H 6.5 b Results of H u s s e y et al. 8 c M e a s u r e d at 37 ° with 0.5 m M 1,2-epoxy-3 (pnitrophenoxy)propane and 5 m M glutathione at p H 6.5.

[6]

GLUTATHIONE TRANSFERASE PURIFICATION

65

Procedure. All operations except FPLC are carried out at 0o-4 °. Bacteria are disrupted by sonication in buffer A. DNase I (1000 Kunitz units per 3 liters original culture) is added, and the suspension is centrifuged for 1 hr at 105,000 g. The supernatant is dialyzed against three changes of 1 liter buffer B over 48 hr. The Matrex gel Orange A column is poured (2.5 × 25 cm) and washed with 15 ml buffer D to remove noncovalently bound dye, 30 ml buffer E to remove guanidinium ions, and equilibrated with 2 column volumes of buffer B. The dialyzed soluble fraction is applied to the column at a flow rate of 1-2 ml/min and the flow through discarded. The retained GST is eluted at a similar flow rate using a linear gradient formed from 200 ml buffer B and an equal weight of buffer C. The major peak absorbing at A280 is GST 5-5, which is approximately 85% pure. This fraction is concentrated by placing in dialysis tubing on a bed of dry polyethylene glycol. The Matrex gel Orange A column is cleaned with 15 ml buffer D and stored in 0.05% (w/v) sodium azide. The GST 5-5 is finally purified by anion-exchange FPLC (Mono Q) exactly as described above for the human ot class GSTs using buffer 5 (Fig. 3). The GST 5-5 elutes at about 25 mM chloride ion.

Analysis of Glutathione Transferase Purity and Activity Enzyme purity may be assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), isoelectrofocusing, and reversedphase HPLC. The HPLC method gives the best resolution and quantitative results. The column packing must be of the "wide-pore" type (>300/~) for high yield and around 5-/zm particle size for optimal resolution. Good separations have been described using Dynamax 16,17 and Vydac 12,1s columns. Elution is by a gradient of 0.05% (v/v) trifluoroacetic acid in acetonitrile. The GST subunits elute in the order PI, Mla, Mlb, AI, M3, and A2. Glutathione transferase activity is assayed using 1-chloro-2,4-dinitrobenzene as electrophilic substrate for the alpha, mu, and pi class enzymes and with 1,2-epoxy-3-(p-nitrophenoxy)propane for GST 5-5.11 To obtain accurate initial rates of reaction, it is necessary to dilute the GSTs with a suitable solution such as 15% (w/v) glycerol, 0.5 mM dithiothreitol, 50 mM sodium phosphate, pH 7.0. The specific activities of GSTs purified as described here are given in Table I. 16 D. J. Meyer, E. Lalor, B. Coles, A. Kispert, P. ~lin, B. Mannervik, and B. Ketterer, Biochem. J. 260, 785 (1989). 17 D. J. Meyer, J. M. Harris, K. S. Gilmore, B. Coles, T. W. Kensler, and B. Ketterer, Carcinogenesis 14, 567 (1993). 18j. A. Johnson, K. A. Finn, and F. L. Siegel, Biochem. J. 282, 279 (1992).

66

GLUTATHIONE

and

[71

[7] M e a s u r e m e n t of 7-Glutamyl Transpeptidase 7-Glutamylcysteine Synthetase Activities in Cells

B y HENRY JAY FORMAN, MICHAEL MING SHI, TAKEO IWAMOTO,

RuI-MINO LIU, and TIMOTHY W. ROBISON Introduction There has been an increasing interest in glutathione ( G S H ) as it is the major nonprotein thiol in cells and is a preferred, if not specific, thiol substrate for several enzymes in xenobiotic metabolism and antioxidant defense. Under such stress conditions, cells must maintain glutathione and may even increase glutathione content above the steady state for maximum protection. The intraorgan transport and d e n o v o synthesis of glutathione are two of the possible mechanisms for maintaining or increasing glutathione. I'2 As there is little circulating cysteine and only some epithelial cells have the capacity for uptake of intact glutathione, 3 circulating glutathione, produced principally by the liver, is itself the major source of the required cysteine for d e n o v o glutathione synthesis in other organs. 4'5 In intraorgan transport of glutathionc, "y-glutamyl transpcptidase [yGT; T-glutamyltransfcrasc, {5-glutamyl)-peptide:amino acid 5-glutamyltransfcrase, E C 2.3.2.2], an enzyme located on the outer surface of the cell membrane, catalyzes the first step in the breakdown of extracellular glutathione [Eq. (i)]. The physiological donor of the glutamyl moiety can be L-T-Glutamyl-L-cysteinylglycine + amino acid ---> L-T-glutamylamino acid + L-cysteinylglycine

(I)

glutathione or its thiyl conjugates, such as leukotriene C4 (LTC4). 6 The enzyme can also transfer the "y-glutamyl moiety from other compounds, such as 7-glutamyl-4-methoxy-2-naphthylamide, which has been used in histochemical staining for the enzyme, 7 and can act as a glutaminase. 8 The acceptor can be an amino acid, a dipeptide, another molecule of glutathione, 1M. E. Anderson and A. Meister, this series, Vol, 143, p. 313. 2S. M. Deneke and B. L. Fanburg, Am. J. Physiol. 257, L163 (1989). 3T. M. Hagen, L. A. Brown, and D. P. Jones, Biochem. Pharmacol. 35, 4537 (1986). 40. W. Griffith and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 76, 5606 (1979). 5j. M. Williamson,B. Boettcher, and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 79, 6246 (1982). 6L. Orning, S. Hammarstrom, and B. Samuelsson,Proc. Natl. Acad. Sci. U.S.A. 77, 2014(1980). 7 A. M. Rutenburg, H. Kim, J. W. Fischbein, J. S. Hanker, H. I. Wasserkrub, and A. M. Seligman, J. Histochem. Cytochem. 17, 517 (1969). s G. A. Thompson and A. Meister, J. Biol. Chem. 254, 2956 (1993). METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc. All rights of reproduction in any form reseiwed.

[7]

'y-GLUTAMYLTRANSPEPTIDASEAND TGCS

67

hippurate analogs, or water. 9 Once formed, both cysteinylglycine and Tglutamyl-amino acids can be further metabolized to supply amino acids to the cell that may then be used for synthesis, including that of glutathioneJ The rate-limiting step in glutathione biosynthesis from the constitutive amino acids is catalyzed by T-glutamylcysteine synthetase (TGCS; glutamate-cysteine ligase; EC 6.3.2.2) [reaction (2)]. Once formed, L-Glutamate + ATP + L-cysteine--> L-T-glutamyl-L-cysteine + ADP + Pi

(2)

T-glutamylcysteine can be linked to glycine in a second ATP-requiring reaction, catalyzed by glutathione synthetase, to produce glutathione. The enzyme also can use other amino acids as the acceptor of the T-glutamyl linkJ ° Principles

y-Glutamyl Transpeptidase The specificity of TGT toward y-glutamyl compounds is quite broad and has allowed the development of several different assays. The most common assay uses L-T-glutamyl-p-nitroanilide with the product detected by absorbance spectrophotometry, n A more sensitive assay was developed by using the fluorescent substrate L-y-glutamyl-7-amino-4-methylcoumarin (y-glutamyl-AMC) [Eq. (3)]. 12 The product, AMC, has quite well-sepaL-y-Glutamyl-AMC + glycylglycine--+ L-y-glutamylglycylglycine+ AMC

(3)

rated excitation-emission spectra from the substrate (Fig. 1). This assay has been modified (as described below) for use with homogenates of cells having quite low TGT activity. Another assay for TGT that has been used for measurement of TGT activity in cells depends on the conversion of LTC4 to leukotriene D4 (LTD4) and separation by high-performance liquid chromatography) 3 This assay was particularly useful in assaying the activity of a T-glutamyl cleaving enzyme related to TGT, which does not recognize the synthetic substrates T-glutamyl-AMC and y-glutamyl-p-nitroanilide.13 As other enzymatic or nonenzymatic reactions may cleave y-glutamyl com9 G. A. Thompson and A. Meister, J. Biol. Chem. 255, 2109 (1993). 10 G. F. Seelig and A. Meister, this series, Vol. 113, p. 379. 11 S. B. Rosalki, Adv. Clin. Chem. 17, 53 (1975). ~z G. D. Smith, J. L. Ding, and T. J. Peters, Anal. Biochem. 100, 136 (1979). 13N. Heisterkamp, E. Rajpert-De Meyts, L. Uribe, H. J. Forman, and J. Groffen, Proc. Natl. Acad. Sci. U.S.A. 88, 6303 (1991).

68

[7]

GLUTATHIONE

ooot

B 1000 -I-glu-AMC ex at 370

-...

T-glu-AMC em at 440 800.

8001

600.

600-

8

iO

m O

E_ 400-

A 400.

200

200. ...... ///

'\ A M C em at 440',,

O.

320

"

~o

............................... ii:::.................................. I

380

3~o

Excitation Wavelength (nm)

400

0 .

.

.

.

~

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

360' 3;0' 480' 4~0',~0' 4;0' 4~0 s;o Emission Wavelength (nm)

Fic. 1. Fluorescence spectra of 7-glutamyl-AMC and AMC. Solutions of 20/~M 7-glutamylAMC or 20/zM AMC in the assay buffer (see text) were used. Excitation spectra (A) were measured for y-glutamyl-AMC with emission at 400 nm (maximum emission wavelength for y-glutamyl-AMC) and 440 nm (maximum emission wavelength for AMC) and for AMC with emission at 440 nm. Emission spectra (B) were measured for 7-glutamyl-AMC with excitation at 349 nm (maximum excitation wavelength for y-glutamyl-AMC) and 370 nm (maximum excitation wavelength for AMC) and for AMC with excitation at 370 nm.

pounds, the inhibitor acivicin [AT-125; L-(aS,5S)-a-amino-3-chloro-4,5-dihydro-5-isoxazoleacetic acid] is used to assess specificity of the enzymatic reaction. 14

"y-Glutamylcysteine Synthetase In the most commonly used assay for ~/GCS, A D P generated in reaction (2) is measured by coupling to the pyruvate kinase or lactate dehydrogenase reactions. 1° Other methods determine the Pi produced in reaction (2). 15'16 Unfortunately, both methods compete with other ADP- or Pi-forming 14 L. Allen, R. Meek, and A. Yunis, Res. Commun. Chem. Pathol. Pharmacol. 2?, 175 (1980). z5 A. N. Lestas and J. M. White, Ann. Clin. Biochem. 20, 241 (1983). 16 E. Beufler and T. Gelbart, Clin. Chim. Acta 158, 115 (1986),

[7]

"y-GLUTAMYLTRANSPEPTIDASEAND yGCS

69

systems, particularly ATPases, that are abundant in cells. Development of new assays with greater specificity has been required to enable measurement of activities in cells, where the rate of enzymatic catalysis is sufficient for physiological requirements but where the commonly used assays are too insensitive. One such method is the measurement of the product, 7-glutamylcysteine,iv To avoid the potential loss of -/-glutamylcysteine, which easily oxidizes, another assay using modifications of the method of Huang et al. TM substitutes L-a-aminobutyrate for L-cysteine [Eq. (4)]. 1° Buthionine L-Glutamate + ATP + L-c~-aminobutyrate L-3~-glutamyl-L-a-aminobutyrate + ADP + Pi

(4)

sulfoximine, an inhibitor of yGCS, is used to be certain that formation of product is through catalysis by this enzyme. 19

Methods y-Glutamyl Transpeptidase

The following assays are modifications of the method of Smith et al.12 Procedure 1: E n d Point Assay. A 0.1 ml enzyme solution (cell extract) is added to 0.25 ml of a reaction mixture containing 0.1 M 2-amino-2-methyl1,3-propanediol (ammediol)-HC1 buffer, pH 8.6, 20 mM glycylglycine, 20 /~M 7-glutamyl-AMC, and 0.1% (v/v) Triton X-100. y-Glutamyl-AMC can be prepared as a 1 mM stock solution in methoxyethanol by sonication. The reaction is carried out at 37° for 30 min and terminated by adding 1.5 ml cold 50 mM glycine buffer. The fluorescence is measured at 440 nm with an excitation wavelength of 370 nm. Product formation is calculated by comparison with a standard curve of AMC. The specific activity is confirmed by using acivicin (10/zM) to inhibit the enzymatic reaction. The advantage of the end point assay is that multiple samples can be assayed simultaneously. This is important when activity is low and requires a prolonged incubation to produce a significant amount of product. Fortunately, under such conditions product formation can be linear for hours. Procedure 2: Kinetic Assay o f E n z y m e Activity. When sufficient activity is present, the kinetic assay can be convenient. The reaction is started by adding 0.1 ml enzyme solution to a fluorescence cuvette containing 0.75 ml of the reaction mixture (as in procedure 1) and 0.15 ml of 0.2 M TrisHC1 buffer, pH 7.9, at 37 °. The fluorescence is monitored continuously at J7K. Ohno and M. Hirata, Biochem. Biophys. Res. Commun. 168, 551 (1990). 18C.-S. Huang, W. R. Moore, and A. Meister, Proc. Natl, Acad. Sci. U.S.A. 85, 2464 (1988). 19O. W. Griffith,J. Biol. Chem. 257, 13704 (1982).

70

GLUTATHIONE

[71

440 nm with excitation at 370 rim. The activity of the enzyme is calculated as in procedure 1.

y-Glutamylcysteine Synthetase The activity of yGCS is determined by measuring the formation of yglutamyl-a-aminobutyrate. Each reaction mixture (final volume 100 /zl) contains 100 mM Tris-HCl, pH 8.2, 50 mM KCI, 20 mM MgCI2, 2 mM EDTA, 10 mM ATP, 5 mM glutamic acid, 5 mM a-aminobutyric acid, and cell homogenate. To remove competing acceptors and inhibitors, cell homogenates are concentrated by centrifuging at 5000 g with a Centricon10 column (Amicon, Danvers, MA) at 4°. The mixture is incubated for 0, 15, or 30 rain at 37 ° to assess linearity. After the incubation, each reaction is terminated by adding 20/xl of 50% (w/v) trichloroacetic acid containing 0.5 mM norleucine as an internal standard. Samples are centrifuged at 15,000 g. The supernatants are analyzed using an amino acid analyzer system. A 2-/zl aliquot of each sample is exposed to phenylisothiocyanate in the presence of a base (diisopropylethylamine). The reaction is quantitative, yielding phenylthiocarbamoylamino compounds. Separation of the phenylthiocarbamoyl derivatives is optimized using a gradient buffer system. The buffers are filtered and degased. Solvent A consists of 140 mM sodium

from Tris buffer

0.6-

ID

E c

,;

0.4

o

0

0.2

0.00

1'0

2 1'5

20

"rime (rain) FIG. 2, Amino acid analysis for 7-glutamyl-a-aminobutyrate formation catalyzed by "yglutamylcysteine synthetase. Assays were performed as described in the text using an Applied Biosystems (Foster City, CA) Model 420A/130A amino acid analyzer.

[7]

~/-GLUTAMYLTRANSPEPTIDASEAND 7GCS

71

acetate buffer, pH 6.5, and solvent B is 70% acetonitrile and 30% (by volume) 32 mM sodium acetate buffer, pH 6.1. The derivatives are injected into a PTC C~8 column (Brownlee Scientific, 5 /zm, 220 × 2.2 mm i.d. obtained from Cole Scientific, Calabasas, CA) equilibrated in 95% solvent A and 5% solvent B at 38°. A linear gradient with a flow rate of 0.3 ml/ min is ramped to 32% solvent B in 10 rain, 60% solvent B in another 10 min, and finally 100% solvent B in 5 min. Elution is then maintained with 100% solvent B for 5 min. The elution of derivatives is monitored at 254 nm. The amount of ~/-glutamyl-~-aminobutyrate produced is quantified by comparison to authentic standards derivatized and analyzed as above. The retention times of the standard amino acids glutamate, 3,-glutamyl-a-aminobutyrate, a-aminobutyrate, and norleucine are 5.4, 8.7, 12.6, and 16.6 min, respectively (Fig. 2). The specificity of the method is confirmed using buthionine sulfoximine to inhibit 3~GCS. One unit of enzyme activity is defined as the amount that catalyzes formation of 1/zmol of product per hour at 37°. The method can detect differences of product formation in the nanomole range. Discussion Several studies have shown that yGT-mediated utilization of extracellular glutathione contributes to protection against oxidative injury. 2°-23 The importance of 3,GCS in maintaining glutathione in the face of the challenge of even endogenous generation of H202 has also been demonstrated. 24 Thus, development of assays sensitive enough for detection of low but physiologically significant activities of -/GT and yGCS will become increasingly important in evaluating the role of the enzymes in cellular resistance to toxic stresses.

2o G. L. Jensen and A. Meister, Proc. NatL Acad. Sci. U.S.A. 80, 4714 (1983). 21 M. F. Tsan, J. E. White, and C. L. Rosano, J. Appl. PhysioL 66, 1029 (1989). 22 H. J. Forman and D. C. Skelton, Am. J. Physiol. 259, L102 (1990). 23 M. Chang, M. Shi, and H. J. Forman, Am. J. Physiol. 6~ L637 (1992). 24 j. Martensson, A. Jain, W. Frayer, and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 86, 5296 (1989).

72

GLUTATHIONE

[8]

Thiol Transport

from Human

[81

Red Blood Cells

By TAKAHITO KONDO, GEORGE L. D A L E , a n d ERNEST BEUTLER

Introduction Glutathione (GSH) is present in human red cells at a concentration of approximately 2.5 mM. The antioxidant functions of GSH are well described and have been substantiated by numerous genetic disorders reflecting increased oxidation as a result of abnormal GSH metabolismJ '2 Another role for GSH in the red cell is the detoxification of xenobiotics by covalent conjugation. Both of these roles for GSH result in products that the red cell can either metabolize or extrude into the extracellular space. Specifcally, excess oxidized glutathione (GSSG) is either enzymatically reduced or rapidly transported out of the cell. Similarly, GSH conjugates of xenobiotics are exported out of the erythrocyte. Both transport systems utilize ATP as an energy source for extrusion, and it is these transport processes that are the focus of this chapter. The transport of GSSG was first observed in oxidatively stressed red cells, both normal and glucose-6-phosphate dehydrogenase deficient.3 Transport was shown to be both rapid and energy-dependent. Subsequently, Lunn et aL 4 demonstrated that a modest basal level of GSSG transport occurred in native red cells at a rate sufficient to account for the previously unexplained turnover of intracellular GSH. In 1981, Board 5 demonstrated that red cells treated with 1-chloro-2,4-dinitrobenzene consumed nearly all of their intracellular GSH in conjugating this xenobiotic but, more importantly, the cells also extruded the GSH-2,4-dinitrophenyl adduct in an energy-dependent process. Further progress in delineating the nature of the GSSG transport process was made by utilizing inside-out vesicles (IOV) prepared from red cells to decipher the molecular requirements for the transport process. 6 This work not only revealed many of the parameters associated with GSSG transport but also complicated the issue by demonstrating that two transport pro1 E. Beutler and G. L. Dale, in "Coenzymes and Cofactors: Glutathione--Chemical, Biochemical and Medical Aspects, Part B" (D. Dolphin, R. Poulson, and O. Avramovic, eds.), p. 291. Wiley, New York, 1989. 2 X. Shan, T. Y. Aw, and D. P. Jones, PharmacoL Ther. 47, 61 (1990). 3 S. K. Srivastava and E. Beutler, J. Biol. Chem. 244, 9 (1969). 4 G. Lunn, G. L. Dale, and E. Beutler, Blood 54, 238 (1979). 5 p. G. Board, FEBS Lett. 124, 163 (1981). 6 T. Kondo, G. L. Dale, and E. Beutler, Proc. Natl. Acad. Sci. U.S.A. 77, 6359 (1980).

METHODSIN ENZYMOLOGY.VOL.252

Copyright© 1995by AcademicPress.Inc. All rightsof reproductionin any formreserved.

[8]

THIOL TRANSPORT FROM HUMAN RED BLOOD CELLS

73

cesses were apparently present in the red cell. O n e system was a lowaffinity, high-Vm~x transport system, and the other was a high-affinity, 1OWVm~x transport. Subsequent work has d e m o n s t r a t e d that the transport of G S H adducts also occurs through two separate transport systems. W h e t h e r the two systems are the same as for G S S G is still an unresolved question. The high-affinity, low-Vma~ transport system for G S H adducts is competitively inhibited by GSSG, whereas the low-affinity, high-Vmax transport is not. 7 A further advance in this field came with purification of a G S S G d e p e n d e n t A T P a s e f r o m h u m a n red cell m e m b r a n e s . 8 Additionally, a G S H conjugate-dependent A T P a s e has b e e n purified 9 which does not utilize GSSG. Characterization of these transport proteins will o p e n the way for cloning of the relevant genes. This is particularly important to resolve a long-standing discussion as to the relationship between the G S S G and G S H - a d d u c t transport systems and the P-glycoprotein found in multiple drug-resistant t u m o r cells lines. 1°

Thiol T r a n s p o r t f r o m W h o l e Cells O x i d i z e d Glutathione Transport f r o m Oxidatively Stressed R e d Ceils

The demonstration that oxidatively stressed red cells extruded G S S G in an energy- and t e m p e r a t u r e - d e p e n d e n t process 3 served as the foundation for this field and has resulted in the subsequent elucidation of the transport processes for G S S G as well as G S H adducts that are detailed below. Red cells are purified f r o m h u m a n blood by cellulose filtration u and washed four times with 0.9% (w/v) saline. The washed red cells are susp e n d e d at a 25% hematocrit in phosphate-buffered saline (PBS; 0.15 M NaC1, 10 m M NaH2PO4, p H 7.4) containing 14 m M glucose and 1.3 m M Na2CrO4. The sodium chromate is added to inhibit glutathione reductase ~2 and thereby allow accumulation of G S S G in an energy-replete cell. The mixture is placed in a W a r b u r g flask along with 200/.d of 30% (w/v) H202 in the center well of the flask. T h e system is then sealed, and the reaction 7 T. P. M. Akerboom, G. Bartosz, and H. Sies, Biochim. Biophys. Acta 1103, 115 (1992). T. Kondo, K. Miyamoto, S. Gasa, N. Taniguchi, and Y. Kawakami, Biochern. Biophys. Res. Commun. 162, 1 (1989). 9 R. Sharma, S. Gupta, S. V. Singh, R. D. Medh, H. Ahmad, E. F. LaBelle, and Y. C. Awasthi, Biochem. Biophys. Res. Cornrnun. 171, 155 (1990). 10I. C. West, Trends Biochern. Sci. 15, 42 (1990). u E. Beutler, in "Red Cell Metabolism: A Manual of Biochemical Methods." Grune & Stratton, New York, 1975. 12S. K. Srivastava and E. Beutler, Biochem. J. 114, 833 (1969).

74

GLUTATHIONE

[8]

proceeds with shaking for 2 hr. Samples are collected at intervals over a 4-hr period; red cells are pelleted, and the supernatants are saved for analysis. The GSSG extruded into the extracellular medium is quantitated enzymatically,I1 and corrections are made for hemolysis during the incubation. 3,~2 The rate of GSSG transport from human red cells under these conditions is approximately 65 nmol/hr/ml cells, a2

Oxidized Glutathione Transport from Native Red Cells The sensitivity required to measure transport from red cells that have not been oxidatively stressed so as to increase the level of GSSG is achieved by incubating the cells with [3H]glycine. This labels preexisting GSH through an exchange reaction catalyzed by glutathione synthetase. In a series of experiments utilizing energy-replete, unstressed red cells, the transport rate out of red cells was found to be 6.7 nmol/hr/ml of red cells. No transport of GSH was observed. The transport rate for GSSG also suggested that the half-life for intracellular GSH would be 4.7 days if the turnover of the peptide were strictly a function of transport. The value of 4.7 days agrees closely with previously measured turnover of total glutathione of 4 days. Human red cells are filtered through cellulose 11 and washed with 0.9% saline. The cells are suspended at a 20% concentration (v/v) in 1% (w/v) bovine serum albumin (BSA), 8 mM glucose, 62 mM NaC1, 40 mM NaH2PO4, 35 mM TES, pH 7.4 (buffer A), with 15/.,Ci of [3H]glycine/ml erythrocytes. The mixture is incubated at 37° for 4 hr, at which time the red cells are washed four times with 0.9% saline. To demonstrate transport the labeled red cells are resuspended at a 20% hematocrit in buffer A and incubated in a shaking water bath at 37°. The incubation is allowed to proceed for 150 min during which 3-ml aliquots are taken for analysis. Each aliquot is centrifuged and the supernatant saved for determination of [3H]GSSG. If hemolysis is noted in the supernatants, correction for release of intracellular metabolites must be taken into account. 4 The [3H]GSSG transported at each time point is quantitated by mixing 2 ml of the reaction supernatant with 14 ml of 0.5 mM glycine, 0.5 mM GSSG, pH 7, and each sample is applied to a 10 × 95 mm column of Dowex-l-X8 (formate form) preequilibrated with water. The column is then eluted with a 100 ml linear gradient of 0 to 1 M ammonium formate; 3.5-ml fractions are collected. A typical chromatographic profile is shown in Fig. 1. The relevant fractions are pooled, and total GSSG radioactivity associated with each time point is determined. To calculate the actual rate of GSSG transport from red cells, the specific

[8]

THIOL TRANSPORT FROM HUMAN RED BLOOD CELLS

75

o-o-o-o

18i

Q)~

14"

~"

10-

o

a~ 6-

/'/\

-loo

to

°~O~o "~o. o

o-o ._._~=Q=Q=Q=Q=Q=

.~ Q= Q=O~

E

\

-o

\O~o

6o

"8

2o

~

-o

FRACTION

FIG. 1. Chromatography of [3H]glutathione. Erythrocytes were incubated with [3H]glycine to label the glutathione pool. A portion of the cells were lysed, incubated with tert-butyl hydroperoxide to oxidize all of the GSH to GSSG, and then chromatographed on Dowex-1 formate. The column was eluted with a 0 to 1 M ammonium formate gradient. Radioactivity is represented by circles (©); the filled circles (Q) are GSSG (nmol/ml). The radioactivity that appeared before fraction 1 is from residual [3H]glycine. (Reprinted from Lunn et al. 4 with permission.) activity of the G S S G after the [3H]glycine incubation must be determined. A n aliquot of the labeled cells is incubated with an excess of tert-butyl hydroperoxide to oxidize intracellular G S H to GSSG. 13 The cells are then precipitated with an equal volume of 4% (w/v) perchloric acid at 0 °. The precipitated protein is pelleted by centrifugation at 1500 g for 15 min, and the supernatant is neutralized with 1 M K2CO3.11 A n aliquot is diluted 4fold with water and then applied and eluted from a Dowex-1 column as described above but without carrier GSSG. The quantitation of G S S G in each fraction is p e r f o r m e d enzymatically 11 and the specific activity calculated as counts per minute (cpm) per micromole of GSSG.

Glutathione-Adduct Transport from Red Cells W h e n red cells are exposed to a xenobiotic such as 1-chloro-2,4-dinitrobenzene (CDNB), enzymatic conjugation of glutathione with C D N B is a3 S. K. Srivastava, Y. C. Awasthi, and E. Beutler, Biochem. J. 139, 289 (1974).

76

GLUTATHIONE

[81

catalyzed by glutathione S-transferase with a resulting product of glutathione-S-2,4-dinitrobenzene (GS-DNP). The transport of the GSH adduct out of the red cell can be readily demonstrated. 5 Red cells are filtered through a-cellulose/microcellulose columns and washed as detailed above. To derivatize erythrocyte GSH, the washed cells are suspended in 4 volumes of 0.9% saline with 1.2 mM CDNB for 15 min at 37°. The cells are then washed twice with buffer A. Under these conditions, approximately 70% of the red cell GSH is conjugated. Transport experiments are performed in buffer A at approximately a 25% hematocrit. The cells are incubated at 37 ° with shaking and samples taken over a 2-hr period. Supernatants are collected by centrifugation, and GSH conjugate present in the supernatants is quantitated spectrophotometrically; the extinction coefficient at 340 nm for the GSH-CDNB conjugate is 9.6 mM -1. Again, corrections for lysis occurring during the transport experiments must be made. The maximum transport rate for GSH adducts is approximately 450 nmol/hr/ml red cells, 5 which is considerably higher than that for GSSG. 12 Alternatively, GS-DNP transport is studied after incubating erythrocytes with 45/zCi of [3H]glycine (25/zCi/nmol) for 2 hr at 37°, followed by washing with saline to remove most of the residual [3H]glycine. Before the transport experiment, cells containing labeled GSH are incubated with 0.5 mM CDNB at 37 ° for 1 hr. After the incubation, the cells are washed three times in 20 volumes of the incubation medium to remove excess CDNB and further incubated for 1 hr at 37°. After incubation, aliquots of the supernatant fraction of the incubation mixture are diluted with 10 volumes of 0.5 mmol/liter GS-DNP and 0.5 mmol/liter glycine, pH 7.0, and applied to a 10 × 95 mm column of Dowex-1 (formate form). Elution and fractionation are the same as those for the GSSG-transport experiment. To calculate the actual rate of GS-DNP transport from red cells, the specific activity of the GS-DNP after the incubation with [3H]glycine must be determined as described above for GSSG transport. Transport Studies with Inside-Out Vesicles Inside-out vesicles (IOV) serve as a useful tool for characterization of systems that transport substances from the inside to the outside of erythrocytes. The concentration of activators, inhibitors, as well as substrates and the temperature and pH of the medium are easily modified. IOV from human erythrocytes were first prepared by Steck and Kant TM using ultracentrifugation methods in a dextran density gradient. Alterna14T. L. Steck and J. A. Kant, this series, Vol. 31, p. 172.

[8]

THIOL TRANSPORT FROM HUMAN RED BLOOD CELLS

77

Km2=7"m lM

~o

y

Krn1=0.1mM

¢

0

I 10

I 20

I 30

I 40

I 50

(G~G,mM) "1 FIG. 2. L i n e w e a v e r - B u r k plot of GSSG transport. Inside-out vesicles were incubated with various concentrations of GSSG, 2 m M A T P , 10 m M MgCI2 in PBS for 2 hr at 37 °. (Reproduced from K o n d o et al. 6)

tively, IOV can be prepared from human erythrocytes using an affinity column of cellulose coupled to concanavalin A. 15 The purity of IOV prepared by the two methods is similar, although with either method there is some contamination with resealed right-side-out vesicles and unsealed vesicles. The existence of two different transport systems for GSSG was first observed using IOW.6 Both systems are dependent on the concentrations of ATP and GSSG (Fig. 2). As mentioned above, the rate of GSSG transport under physiological conditions is sufficient to account for the turnover rate of glutathione in erythrocytes, and larger amounts of GSSG are effluxed when the cells are exposed to oxidative stress. The presence of two transport processes for GSSG provides a reasonable explanation for these previous findings: the high-affinity, IoW-Vma~transport may be important to maintain the steady-state levels of glutathione, and the low-affinity, high-Vmax transport may serve as an emergency pathway for the elimination of GSSG during oxidative stress. Furthermore, pyrimidine nucleotide triphosphates such as CTP and UTP have been found to inhibit competitively the highaffinity, low-Vmax GSSG transport. 6 This inhibition may explain the high levels of intracellular GSH observed in hereditary hemolytic anemia associ15 T. Kondo, this series, Vol. 171, p. 217.

78

GLUTATHIONE

[8]

TABLE I TRANSPORTOF GLUTATHIONES-CoNJUGATESIN HUMANERYTHROCYTESa Erythrocyte Parameter

Reticulocyte-rich

Transport of GS-DNP (nmol/ml RBC/hr) GSH (nmol/ml RBC) ATP (nmol/ml RBC) -r-Glutamylcysteine synthetase (units/g Hb) Glutathione S-transferase (units/g Hb)

468 _+ 38 2.3 ± 0.1 1.9 ± 0.1 1.16 ± 0.11 6.8 -+ 1.7

Reticulocyte-poor 282 2.2 ± 1.8 ± 0.89 ± 6.7 -

42 0.1 0.2 0.16 1.9

a Erythrocytes were separated into two groups by centrifugation on the basis of specific gravity. Values are expressed as means ± SD of four different experiments. Extracellular GS-DNP was corrected by hemolysis. RBC, Red blood cell; Hb, hemoglobin. Data from T. Kondo, K. Yoshida, Y. Urata, S. Goto, S. Gasa, and N. Taniguchi, J. Biol. Chem. 268, 20366 (1993). ated with e r y t h r o c y t e pyrimidine 5 ' - n u c l e o t i d a s e deficiency, which is invariably associated with increased a m o u n t s o f pyrimidine nucleotides. 16 Active transport of G S - D N P can also be m e a s u r e d using this technology. T h e a p p a r e n t g m for G S - D N P is 940 ~ M at an A T P c o n c e n t r a t i o n of 2 m M , and that for A T P is 7 6 0 / z M at a G S - D N P c o n c e n t r a t i o n of 1 m M . 17 W h e n red cells are separated into two g r o u p s on the basis of specific gravity using P e r c o l l - H y p a q u e centrifugation, 18 the transport rate is higher in the reticulocyte-rich fraction than in the r e t i c u l o c y t e - p o o r fraction b o t h in native red cells and I O V f r o m the fractions, suggesting that the transport system for G S - D N P is inactivated or lost during reticulocyte m a t u r a t i o n (Table I). Oxidized Glutathione-Stimulated ATPase from Human R e d Cell M e m b r a n e s T h e existence o f A T P - d e p e n d e n t transport systems suggests the presence of A T P a s e s which are stimulated b y G S S G . E s t i m a t i o n o f the activity o f G S S G - s t i m u l a t e d A T P a s e o f h u m a n red cell m e m b r a n e s is difficult b e c a u s e erythrocytes are very rich in various types o f A T P a s e activities. A G S S G - s t i m u l a t e d A T P a s e was first d e m o n s t r a t e d in the p l a s m a m e m b r a n e fraction of rat h e p a t o c y t e s attached to polycationic b e a d s J 9 Utilization of 16T. Kondo, Y. Ohtsuka, M. Shimada, Y. Kawakami, Y. Hiyoshi, Y. Tsuji, H. Fujii, and S. Miwa, A m . I. Hematol. 26, 37 (1987). 17T. Kondo, M. Murao, and N. Taniguchi, Eur. J. Biochem. 125, 551 (1982). 18L. Vettore, M. C. DeMatteis, and P. Zampini, A m . J. Hematol. 8, 291 (1980). 19p. Nicotera, M. Moore, G. Bellomo, F. Mirabelli, and S. Orrenius, J. Biol. Chem. 260, 1999 (1985).

[8]

THIOL TRANSPORT FROM HUMAN RED BLOOD CELLS

79

such polycationic beads also allowed the demonstration of enzyme activity in red cell membranes, and the activity can also be demonstrated in erythrocyte membranes solubilized in detergents.

Assay Attachment of Red Cell Membranes to Polycationic Beads. Erythrocyte membranes are ionically attached to polyethyleneimine-coated beads (AffiGel 731, Bio-Rad, Richmond, CA) as described by Jacobson and Branton. 2° Two grams of beads are washed three times with 0.2 M NaC1, twice with 150 mM Tris-HCl buffer, pH 7.4, and once with buffer A (7 volumes of 310 mM sucrose and 3 volumes of 49 mM sodium acetate, pH 5.0). Eight milliliters of a 50% (v/v) suspension of beads is added slowly to 4 ml of washed erythrocytes suspended in buffer A. After washing with buffer A twice, hemoglobin is removed from the erythrocytes attached to the beads by washing three times in 30 volumes of 10 mM Tris, pH 7.4, at 0°. The cell membranes are further disrupted by sonication for 5 sec. The membrane/bead suspension, washed with Tris-HC1, pH 7.4, is resuspended in the same buffer and kept at 0 ° until use. Oxidized Glutathione-Stimulated ATPase Activity Using Membrane/ Bead Suspension. One hundred microliters of membrane/bead suspension (20-30 /xg protein) is incubated in a 250-/~1 system containing 100 mM Tris-HC1, pH 7.4, 0.5 mM EDTA, 10 mM MgCl2, 1 mM ATP. The GSSG is added at the concentration indicated. No GSSG is present in the blank system. After preincubation for 5 min at 37°, the reaction is initiated by the addition of 1 mM ATP. The reaction is terminated after 30 rain by the addition of 250/~l of ice-cold 8% (w/v) trichloroacetic acid. The tubes are centrifuged in the cold, and inorganic phosphate (Pi) is extracted from the supernatant into benzene/isobutanol (1:1, v/v) as a molybdate complex 2~ and assayed colorimetrically. The activity is quantitated by Pi released by the addition of GSSG and is expressed as nanomoles of Pi released per minute per milligram protein. Membrane protein is removed from the polycationic beads as described by Cohen et a1.,22 and protein concentration is determined according to the method of Redinbaugh and Turley. 23 Figure 3 depicts the GSSG-stimulated ATPase present in red cell membranes attached to polycationic beads. The activity of the enzyme is esti20 B. S. Jacobson and D. Branton, Science 195, 302 (1977). 21 E. Beutler and W. Kuhl, Biochim. Biophys. Acta 601, 372 (1980). ~2 C. M. Cohen, D. I. Kalishi, B. S. Jacobson, and D. Branton, J. Cell Biol. 75, 119 (1977). 23 M. G. Redinbaugh and R. B. Turley, Anal. Biochem. 153, 267 (1986).

80

GLUTATHIONE

[8]

'7,

E ~. 0.8

~ 0.6 e

~ OA

> 0.2 ~

f o

Kin2=2.0raM ,'o

'

s'o

,;o

(GSSG concentration)l,(mM) 1 FIG. 3. Lineweaver-Burk plot of GSSG-stimulated ATPase activity. The nanomoles Pi released per minute per milligram protein of erythrocyte membranes attached to polycationic beads is expressed as a function of GSSG concentration. [From T. Kondo, Y. Kawakami, N. Taniguchi, and E. Beutler, Proc. Natl. Acad. Sci. U.S.A. 84, 7373 (1987).]

mated as a function of GSSG concentration ranging from 25 /zM to 10 mM. Lineweaver-Burk analysis of the data gives a biphasic plot with two apparent Km values of 0.13 and 2.0 m M GSSG, having apparent Vma~values of 6.4 and 38.5 nmol of Pi released per minute per milligram protein, respectively. Oxidized Glutathione-Stimulated A TPase Using Solubilized Membranes. When the activity of the GSSG-stimulated ATPase is estimated using solubilized red cell membranes, the assay mixture consists of 150/zl of enzyme solution in a 250-/xl system containing 100 m M Tris-HCl, pH 7.4, 0.5 mM EDTA, 10 mM MgCI2, 1 m M ATP, and GSSG. To estimate activities of the distinct GSSG-stimulated ATPases, 50/zM and 5 mM concentrations of GSSG are used to detect the low-Kin and high-Kin enzymes, respectively. After 5 min of preincubation followed by incubation for 30 min at 37 °, the reaction is terminated by the addition of 250/zl of 8% (w/v) trichloroacetic acid as described above. When the protein concentration in the assay system is less than 20/zg, 100/zg bovine serum albumin is added after termination of the reaction. In some studies, [3,-32p]ATP with a specific activity of 1.5 pmol per disintegration per minute (dpm) is used to quantitate [32P]Pi release. Glutathione S-Conjugate-Stimulated A TPase of Human Red Cell Membranes. S-(2,4-Dinitrophenyl)glutathione (GS-DNP) is prepared enzymati-

[8]

THIOL TRANSPORT FROM HUMAN RED BLOOD CELLS

81

2.0

200

2o01/

o

,00 o ._E 6_o100

d~ [] 0

10

20

30 Fraction No.

40

FIG. 4. Affinity chromatography of two GSSG-stimulated ATPases. Erythrocyte membranes solubilized in 0.25% Triton X-100, 10 mM imidazole hydrochloride, pH 7.4, 0.2 mM EDTA, 50/xM 2-mercaptoethanol, 0.5 m M phenylmethylsulfonyl fluoride were chromatographed on S-hexylgluthathione-Sepharose 6B. Elution was with a 0-2 mM S-hexylglutathione gradient. Protein concentrations are represented by filled circles, activities of GSSG-stimulated ATPases by open bars, and GS-DNP-stimulated ATPase by hatched bars.

cally using glutathione S-transferase as described by Wahllander and Sies. 24 The [3H]GS-DNP is prepared using [3H]GSH and 1-chloro-2,4-dinitrobenzene according to the same method used for the unlabeled compound. The activity of the GS-DNP-stimulated ATPase is measured at 37° in a 250-/xl system containing 100 mM Tris-HCl, pH 7.4, 0.5 mM EDTA, 10 mM MgCI2, l m M [y-32p]ATP, 0.1 mM GS-DNP, and 150/zl enzyme solution. As a blank system, no GS-DNP is added to the assay mixture. Released [3aP]Pi is measured in a liquid scintillation counter after extraction with benzene/isobutanol. The activity is expressed as the amount of P~ released by the addition of GS-DNP.

Purification of Oxidized Glutathione-Stimulated A TPase from Human Red Cells Unless otherwise indicated, all procedures are performed at 4°. Hemoglobin-free red cell membranes are prepared by lysing 100 ml of washed red cells in 10 volumes of 10 mM Tris-HCl, pH 7.4, at 0°, and washing five times in the same buffer. White ghosts are suspended in 20 volumes of 0.5 mM Tris-HC1, pH 8.0, for 12 hr to remove most of the extrinsic proteins such as spectrin. Spectrin-depleted ghosts are collected by centrifugation 24 A. Wahllander and H. Sies, Eur. J. Biochem. 96, 441 (1979).

82

GLUTATHIONE

[81

at 15,000 g for 30 min and are solubilized in 0.5% (w/v) Triton X-100 in 10 m M imidazole hydrochloride, pH 7.4, containing 0.2 mM EDTA, 50 ~ M 2-mercaptoethanol, 0.5 mM phenylmethylsulfonyl fluoride (buffer B), for 20 min. Insoluble proteins are removed by centrifugation at 10,000 g for 90 rain. The extract is diluted to adjust the concentration of Triton X100 to 0.25% and then applied to a column of S-hexylglutathione-Sepharose 6B (1 × 6 cm). The column is washed with 50 ml of 0.5 NaC1 in buffer B containing 0.25% Triton X-100, and bound material is eluted with a 60-ml linear gradient of 0-2 m M S-hexylglutathione at a flow rate of 8 ml/hr. Fractions of 1 ml are collected, and the enzyme activities and protein concentration of fractions are determined. As shown in Fig. 3, the presence of of two distinct GSSG-stimulated ATPases is observed by elution from S-hexylglutathione-Sepharose 6B. The specific activity of the low-Km enzyme is 1200 nmol of Pi released/ min/mg protein, and the protein purified from 100 ml erythrocytes is 0.09 mg. The specific activity of the high-Kin enzyme is 980 nmol of Pi released/ min/mg protein, and the protein purified is 0.1 mg. The low-affinity, but not the high-affinity transport system is activated by treatment with dithiothreitol and inhibited by the addition of p-chloromercuribenzoate, iodoacetamide, or N-ethylmaleimide. Mixed disulfides other than GSSG such as Glu-Cys-SSG and Cys-SSG partially stimulate the enzyme, but GSH has no effect. GS-DNP-stimulated ATPase is purified from red cell membranes using the method as described above for GSSG-stimulated ATPase. 25'26 The elution of GS-DNP-stimulated ATPase is identical to that for the highKr~ form of GSSG-stimulated ATPase (Fig. 4). The enzyme has an apparent Km value of 0.14 m M for ATP and 0.038 m M for GS-DNP. These results show the presence of two different GSSG-stimulated ATPases in human red cell membranes with properties resembling those of the GSSG-transport systems. It seems reasonable to conclude that the enzymes play a role in the outward transport of thiols in human red blood cells.

25 T. Kondo, K. Yoshida, Y. Urata, S. Goto, S. Gasa, and N. Taniguchi, Z BioL Chem, 268, 20366 (1993). 26 T. Kondo, Y. Kawakami, N. Taniguchi, and E. Beutler, Pro~ N a ~ Aca& ScL U.~A. 84, 7373 (1987).

[91

GLUTATHIONEMUTANTS

[9]

Glutathione

83

Mutants

By JAMES A. FUCHS

Introduction This chapter concentrates on Escherichia coli mutants altered in the synthesis, reduction, or utilization of glutathione as a model system but briefly discusses some mutants in yeast and humans. Although some contradictory results have been reported for mutants altered in glutathione metabolism, these differences can probably be explained by a combination of the redundancy of functions in E. coli that reduce thiols and a strong selection pressure to compensate for mutations that decrease the cellular concentration of reduced glutathione. As an example, we find that a null gshA mutant strain made by P1 transduction is unable to grow at 42 ° unless supplied with glutathione but the same strain after being maintained for some time under laboratory conditions grows at 42 ° (J. A. Fuchs, unpublished). The first observation that there may be selection pressure to alter glutathione metabolism under laboratory conditions came from a study by Owens and Hartman ~ who found that laboratory strains of E. coli and Salmonella typhimurium characterized excreted up to 30% of their glutathione into the medium, whereas fresh isolates excreted at least 19-fold less. Because the excretion could be eliminated by azide, it was energy dependent. Although 50% of the glutathione in the medium was oxidized, reconstruction experiments 1 indicated that most of the oxidation must have occurred in the aerobic medium after excretion and in sample preparation. As another example, Russel and Holmgren 2 isolated a grx mutant that appeared to have a very different phenotype than that isolated by Kren et aL 3 However, when a double mutant 4 was constructed that was defective in glutaredoxin and thioredoxin, the strain obtained was found to have mutated in several other genes in addition to grx and trxA, presumably to compensate for the primary mutations. Thus, one must be careful in maintenance of mutant strains as well as in selections to prevent accumulation of secondary mutations. This chapter discusses how E. coli mutants defective in the various i R. A. Owens and P. E. Hartman, J. Bacteriol. 168, 109 (1986). 2 M. Russel and A. Holmgren, Proc. Natl. Acad. Sci. U.S.A. 85, 990 (1988). 3 B. Kren, D. Parsell, and J. A. Fuchs, J. Bacteriol. 170, 308 (1988). 4M. Russel, P. Model, and A. Holmgren, J. Bacteriol. 172, 1923 (1990).

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

84

GLUTATHIONE

[91

steps of glutathione synthesis, reduction, and utilization are isolated and characterized and then briefly discusses similar mutants in yeast as well as in humans.

gshB M u t a n t s The first Escherichia coli mutants found to be defective in the synthesis of glutathione were simultaneously isolated and characterized by two groups. 5'6 In an attempt to isolate a mutant in thioredoxin or thioredoxin reductase, Fuchs and Warner 5 isolated a mutant in gshB (gene encoding glutathione synthase) by selecting for ceils with a decreased capacity to reduce ribonucleotide diphosphates to deoxyribonucleotide diphosphates. In the gshB mutant, the activity of one of the subunits of ribonucleotide reductase, the enzyme that carries out the reaction, was significantly decreased. The defect was corrected by restoring glutathione synthase, indicating that gshB was responsible for the decreased activity. The gshB mutant contained no detectable glutathione but contained an increased pool of yglutamylcysteine, approximately the same amount as the glutathione pool in the parent. The gshB mutant had an increased sensitivity to the oxidant diamide. 7 Diamide led to a rapid conversion of small molecular weight thiols to disulfides in the mutant strain but not the parental strain, suggesting that the mutant cell is able to keep y-glutamylcysteine reduced by a system other than glutathione reductase (y-glutamylcysteine is not a substrate for glutathione reductase). However, that system is insufficient to keep small molecular weight thiols reduced when challenged by an external oxidant. The gene encoding gshB was cloned by complementation of a gshB, trxB double mutant which requires glutathione for growth. 8 The cloned gene, inactivated by insertion of a kanamycin cassette, was integrated into the E. coli chromosome by homologous recombination. The gene was then used to map the locus to 63.4 rain on the genetic map and to a region between 3100 and 3120 kilobase pairs on the physical map. s Apontoweil and Berends 6 also isolated a gshB mutation by selection of diamide-resistant ceils followed by detection of colonies that failed to react with sodium nitroprusside. Their studies found that the gshB mutant grew slightly slower than its parent and was devoid of detectable glutathione; however, the mutant was not extensively characterized.

5j. A. Fuchs and H. R. Warner, J. Bacteriol. 124, 140 (1975). 6p. Apontoweil and W. Berends, Biochem. Biophys. Acta 399, 10 (1975). 7 K. A. Hibberd, P. B. Berget, H. R. Warner, and J. A. Fuchs, J. BacterioL 133, 1150 (1978). 8T. Daws, C.-J. Lim, and J. A. Fuchs, J. Bacteriol. 171, 518 (1989).

[9]

GLUTATHIONE MUTANTS

85

gshA M u t a n t s In the same selection used to isolate a $shB mutant, Apontoweil and Berends 6 obtained three gshA mutants. The gshA mutants were found to have an increased sensitivity to a large number of chemical agents, suggesting that the mutation may have altered membrane permeability. However, Apontoweil and Berends 6 did not find that the gshA mutants had increased sensitivity to X rays. The gshA mutation was mapped by P1 transduction to the pheA-naIB region" and shown to be closely linked to srl: :Tnl0. ~°

O t h e r gshA a n d gshB M u t a n t s In an attempt to isolate a mutant consitutive for an adaptive response to alkylating agents, Sedgwick and Robins ix isolated mutants with decreased glutathione content. Because a thiol compound is required for activation of N-methyl-N'-nitro-N-nitrosoguanidine and glutathione is a major source of thiols, this phenotype was not unexpected. In fact, Apontoweil and Berends 6 had earlier observed the phenotype in gshA mutants. Greenberg and Demple ~2 took advantage of the phenotype to isolate an insertional inactivated allele of the gshA gene that allows gshA to be easily moved between strains. The gshA : :TnlOkan mutant was used to investigate the role of glutathione in protection against oxygen during X-irradiation in E. coll. Although Apontoweil and Berends 6 found no protection by glutathione, Morse and DahP 3 using a different gsh allele and nonisogenic strains found that glutathione was the key to the oxygen effect. The study by Greenberg and Demple 12 was consistent with those of Apontoweil and Berends 6 and found no effect of glutathione on the oxygen effect present during X-irradiation. Using 6°C0 as a source of y-irradiation, H a r r o p et al.a4 measured the oxygen enhancement of killing in gshA and gshB mutants and found that the mutants did affect the oxygen enhancement. The fast chemical repair of O2-dependent damage in the strains was measured using a fast mixing and irradiation method, namely, the gas explosion technique. The chemical repair rates in the various E. coil strains varied approximately in proportion to the 02 concentration that gives half-maximal 02 effects (K value), and both the rates of chemical repair and the K values correlated 9p. Apontoweil and W. Berends, Mol. Gen. Genet. 141, 91 (1975). ~oB. L. Hailer and J. A. Fuchs, J. Bacteriol. 159, 1060 (1984). ~1B. Sedgwick and P. Robins, Mol. Gen. Genet. 180, 85 (1980), 12j. T. Greenberg and B. Demple, J. Bacteriol. 168, 1026 (1986). 13M. L. Morse and R. H. Dahl, Nature (London) 271, 661 (1978). 14H. A. Harrop, K. D. Held, and B. D. Michael, Int. J. Radiar Biol. 59, 1237 (1991).

86

GLUTATHIONE

[91

approximately with the levels of nonprotein sulfhydrals in the various strains. The use of E. coli gsh mutants as a model for the role of glutathione in oxidant stress in mammalian hepatocytes was evaluated by Romero and Canada.15 They concluded that the lack of a correlation of GSH content with either oxidant-induced growth inhibitory or lethality effects, the presence of catalase in the cytoplasm in E. coli rather than in peroxisomes as in mammals, and the absence of glutathione peroxidase, made E. coli a poor model for mammalian oxidant stress. Accumulation of K ÷ and glutamate plays a primary role in maintaining osmotic balance in E. coli, as illustrated by the high concentrations of the ions present in cells growing in medium of high osmolarity. McLaggan et al. 16 found that two 3,-glutamyl peptides and glutamine also accumulated during growth at high osmolarity. A mutant deficient in glutathione owing to an insertion in the gshA gene was unable to grow above 1.4 osM, grew more slowly at intermediate osmolarities, and took longer to start growing following osmotic upshock. The involvement of glutathione in osmoregulation was independent of the effect of glutathione on K ÷ retention. Selenite toxicity in Salmonella (a species closely related to E. coli) involves the ability of selenite to act as an oxidizing agent. 17The gsh mutants were found to exhibit a biphasic killing curve as a function of selenite concentrations. At low concentrations, the mutants were more sensitive to selenite than the parent strain, presumably owing to the lack of sulfhydrals, but at higher concentrations the mutants were more resistant than the parent strain. Under these conditions the lack of sulfhydrals appeared to protect as sulfhydrals are needed to form more reactive oxidizing species. Gardner and FridovichTM investigated the effect of glutathione on the superoxide-sensitive [4Fe-4S]-containing aconitase of Escherichia coli. A gshA mutant grew slower than did the parental strain in a defined medium, and this effect was more pronounced when succinate, as opposed to glucose, was supplied as the carbon source. This suggested that the citric acid cycle was compromised in the gshA strain. Aconitase activity was approximately 25% lower in the gshA cells compared to the parental strain growing on either glucose or succinate. Addition of glutathione to the medium stimulated growth of the gshA strain on succinate and also elevated the aconitase activity even in the presence of chloramphenicol, which was added to block protein synthesis. In the presence of 4.2 atm of 02, however, aconitase 15 M. J. Romero and A. T. Canada, Toxicol. Appl. PharmacoL U l , 485 (1991). 16D. McLaggan, T. M. Logan, D. G. Lynn, and W. Epstein, J. Bacteriol. 172, 3631 (1990). 17 G. F. Kramer and B. N. Ames, Mutat. Res. 201, 169 (1988). 18p. R. Gardner and I. Fridovich, Arch. Biochern. Biophys. 301, 98 (1993).

[91

GLUTATHIONEMUTANTS

87

activity decreased in b o t h m u t a n t and parent, with the recovery of the mutant strain being much slower than that of the parent strain. Because both aconitase and the R1 subunit of ribonucleotide reductase have decreased activity in the absence of glutathione, 5 it is reasonable to conclude that other enzymes are likewise affected by the absence of glutathione in the cell.

g s h M u t a n t s in E s c h e r i c h i a coli B Although most experiments with glutathione mutants, like most other experiments in E. coli, have b e e n conduced with K strains, some work has also been conducted with B strains. T h e B strains were isolated independently from K strains and have a different modification and restriction system, making transfer of genes between t h e m difficult and thus allowing t h e m to diverge during evolution. Murata et al. 19 isolated an E. coil B mutant that was resistant to methylglyoxal and found that the mutant excreted 50-fold the normal level of glutathione when given methylglyoxal. The mutant had high levels of the glutathione synthesizing enzymes as well as glyoxal I and II. Murata and K i m u r a 2° isolated a gsh mutant in E. coli B by starting with the strain that excreted glutathione. The strain was mutagenized and then plated on a plate seeded with 107 cells of a strain that was deficient in the synthesis of cysteine but was able to use glutathione as a source of cysteine. Mutants that failed to form halos of the seeded ceils were collected and analyzed for glutathione synthase (gshII) and T-glutamylcysteine synthetase activity (gshI). The mutants had properties similar to gsh mutants in E. coli K12 strains. The gshII gene was cloned by selecting for resistance to tetramethylthiuram disulfide in a gshII mutant 21 and then was sequenced. 22

gor (gshC) M u t a n t s Mutants defective in glutathione reductase (gor) were isolated by different methods. Tuggle and Fuchs 23 employed the same selection used to obtain gshA mutants, assuming that these mutants would be sensitive to diamide. Davis et al. 24 isolated gor mutants without selection using transpo19K. Murata, K. Tani, J. Kato, and I. Chibata, J. Gen. Microbiol. 120, 545 (1980). 20K. Murata and A. Kimura, J. Gen. Microbiol. 18, 10047 (1992). 21K. Murata, T. Miya, H. Gushima, and A. Kimura, Agric. Biol. Chem. 45, 2131 (1981). 22H. Gushima, S. Yasda, E. Soeda, M. Yokoa, M. Kondo, and A. Kimura, Nucleic Acids Res. 12, 9299 (1984). 23C. K. Tuggle and J. A. Fuchs, J. Bacteriol. 162, 448 (1985). 24N. K. Davis, S. Greer, M. C. Jones-Mortimer, and R. N. Perham, J. Gen. Microbiol. 12, 631 (1982).

88

GLUTATHIONE

[91

son mutagenesis followed by direct assays of glutathione reductase activity in colonies on plates. The gor mutation was mapped between 77 and 78 min on the E. coli map and was found to be 32% cotransducible with xyl with T4GT7 phage but not with P1 phage. 24 The gor gene was linked t o pit by P1 transduction, and a clone with an insert of 25.5 kb was found to contain the gor gene approximately 10 kb away from the pit gene. 25 Greer and Perham 26 also isolated a clone that contained the gor gene and determined the DNA sequence of gor. In a selection for hydrogen peroxide-sensitive E. coli cells, Barbado et al.27 found that approximately 20% of the mutants were gor, but the mutants were not further characterized. In accordance with the observation that y-glutamylcysteine in a gshB mutant is reduced although it is not a substrate for glutathione reductase, the characterization of a gor mutant 23 indicated that glutathione reductase was not required in vivo to keep glutathione reduced. Escherichia coli has alternate systems that can keep the redox state of the cell almost fully reduced without glutathione reductase. The addition of an amount of diamide to the gor mutant strain that failed to alter significantly the ratio of reduced glutathione to total glutathione in the parent oxidized most of the cellular glutathione in the gor mutant. Thus, glutathione reductase is not essential for E. coli and does not seem necessary to keep glutathione reduced, but it is important in conditions that are more oxidizing than in an aerobically growing culture of E. coli. 23 Kunert et aL 2s using a cloned gor gene in a gor mutant to investigate the sensitivity of E. coli to methyl viologen found that sensitivity increased with glutathione reductase content. They further reported that the activity of glutathione reductase is essential to maintain a high glutathione content. This result is surprising as glutathione content appears to be controlled by the level of cysteine in the cell. Perhaps the amount of glutathione excreted is affected by reduction. grx M u t a n t s

Glutaredoxin in E. coli can couple the oxidation of glutathione to the reduction of disulfide bonds. Mutants deficient in glutaredoxin were ob25 C. M. Elvin, N. E. Dixon, and H. Rosenberg, Mol. Gen. Genet. 204, 477 (1986). 26 S. Greer and R. N. Perham, Biochemistry 25, 2736 (1986). 27 C. Barbado, M. Ramirez, A. Blanco, J. Lopez-Barea, and C. Pueyo, Curr. Microbiol. 8, 251 (1983). 28 K. J. Kunert, C. F. Cresswell, A. Schmidt, P. M. Mullineaux, and C. H. Foyer, Arch. Biochem. Biophys. 282, 33 (1990).

[91

GLUTATHIONEMUTANTS

89

tained in a selection by Kren et al. 3 based on the assumption that a double mutant defective in the genes encoding thioredoxin reductase and glutaredoxin would have the same phenotype as double mutants defective in other steps of the glutaredoxin-dependent and the thioredoxin-dependent pathways of reduction of protein thiols. These types of double mutants were all unable to reduce SO]- and thus had a requirement for a reduced sulfur source. Because the double mutants were viable at least at temperatures of 37° or less, these double mutants had sufficient ribonucleotide reductase activity for DNA synthesis? Two mutants defective in glutaredoxin were isolated and characterized. Both failed to grow without a reduced sulfur source. Introduction of a trxB gene on a plasmid failed to restore growth with sulfate, indicating that the grx gene product was required for the reduction of sulfate. Strains containing a grx mutant and a nonmutant trxB gene on a plasmid formed filaments when grown at 42°, indicating they were defective in reduction of ribonucleotides. Introduction of a plasmid-borne grx gene into the strains restored all phenotypes to the parental type. Russel and Holmgren z produced a grx null mutant by gene replacement of the chromosomal gene with a cloned grx gene that had been interrupted by the gene encoding kanamycin. The mutant had a phenotype that did not differ from its parent. Because a null mutant would be expected to have a more severe phenotype than a point mutant, this mutant strain appeared to differ from those isolated by Kren et al. 3 owing to either secondary mutations or strain differences. The null grx mutation was mapped at 18.5 min of the E. coli map. 2 A derivative strain that also contained a null trxA mutation was constructed.4 This double mutant had a cystine requirement as well as having a cell separation defect and formed colonies that were mucoid. The latter two phenotypes were found to be independent of the trxA and grx mutations. In addition, Russel et al. 4 found that the double mutant had accumulated an additional mutation in cysA. They observed that in the absence of cystine a trxA, grx strain is inviable owing to accumulation of 3'-phosphoadenosine 5'-phosphosulfate (PAPS). Cystine suppresses the synthesis of PAPS. Growth of a double mutant without the cystine added to the medium leads to a strong selection for a cysA mutation. It was proposed that there is an additional system to reduce disulfides required for the synthesis of deoxyribonucleotides. The difference in phenotypes observed by Kren et al. 3 and Russel et al. 4 appears to be due to accumulation of secondary mutations. The large overlap of functions of enzyme systems involved in reduction of disulfides may allow secondary mutations in regulation to compensate for various primary mutations.

90

GLUTATHIONE

[91

gsh M u t a n t s in Yeast Kistler et al.29 isolated glutathione-deficient (gsh) mutants of the yeast Saccharomyces cerevisiae after U V treatment using N-methyl-N'-nitro-Nnitrosoguanidine (MNNG) as was used in E. coli as a selective agent. Mutant isolates showed a 2 : 2 segregation of the G S H - and G S H ÷ phenotypes, indicating that the mutant was recessive. Complementation analysis indicated that all gsh mutants belong to one complementation group. Ohtake and Yabuuchi 3° cloned the gshl gene by complementation and used gene disruption to show that the resulting mutant had the same properties as the original gsh mutant. The cloned gene was sequenced and found to have 46% similarity with the rat kidney gshl gene and 26% with the E. coli gshl gene. Southern analysis was used to show that gshl is located on the X chromosome. Mutoh et aL 31 cloned the gene for the large subunit of glutathione synthase (EC 6.3.2.3) of Schizosaccharornyces pombe from a S. pombe genomic D N A library by complementation of cadmium hypersensitivity of a glutathione synthase-deficient mutant of S. pombe. The strain harboring the cloned gene expressed more glutathione synthase in spite of having extra copies of only one subunit of the enzyme. Lisowsky 32 isolated a new temperature-sensitive nuclear mutant affecting the biogenesis of functional mitochondria and characterized the mutant as having a complete block of mitochondrial translation at the restrictive temperature. Transformation of the mutant with plasmids from gene banks identified a chromosomal D N A fragment which can restore growth at the restrictive temperature. Sequence analysis identified the gene as the yglutamylcysteine synthetase of yeast. Disruption of the yeast y-glutamylcysteine synthetase caused a drastic reduction of growth on glucose medium. The insertion mutants were not able to grow on plates with glycerol as the sole carbon source, indicating the dependence of mitochondria on glycerol. Crosses between the temperature-sensitive mutant and the insertion mutant indicated that the mutations are in different genes and that the cloned yeast gene acts as an extragenic suppresser when present on a high copy number plasmid.

29M. Kistler, K.-H. Summer, and F. Eakardt, Mutat. Res. 173, 117 (1986). 3oy. Ohtake and S. Yabuuchi, Yeast7, 953 (1991). 31N. Mutoh, C. W. Nakagaawa, S. Ando, K. Tanabe, and Y. Hayashi, Biochem. Biophys. Res. Commun. 181, 430 (1991). 32Z. Lisowsky, Curt. Genet. 23, 408 (1993).

19]

GLUTATHIONEMUTANTS

91

B e r t h e - C o r t i et al. 33 d e t e r m i n e d glutathione ( G S H ) c o n t e n t in S a c c h a r o m y c e s c e r e v i s i a e as a function o f g r o w t h conditions and used the genotoxicity of M N N G , which is activated by G S H , as a quantitation of G S H . Levels of G S H varied greatly with g r o w t h conditions in a m a n n e r similar to that f o u n d for E. coli. gsh Mutants in Humans

T h e first h u m a n m u t a n t s deficient in the synthesis of glutathione were d e t e c t e d as patients with hemolytic anemia. ~,35 T h e patients had a defect in the gene e n c o d i n g a specific e r y t h r o c y t e glutathione synthase. This defect has also b e e n r e p o r t e d for an additional three families. 36'37 A s e c o n d type of defect that causes a generalized defect in glutathione synthase was reported. 38 T h e individual excreted large a m o u n t s o f 5-0xoproline. Sixteen patients with the defect have b e e n reported. 39 In all cases, patients were mentally r e t a r d e d and h a d n u m e r o u s neurological symptoms. T h e sympt o m s could be detected in the n e o n a t a l state. A defect in the gene e n c o d i n g ~/-glutamylcysteine synthetase was rep o r t e d for two siblings with hemolytic anemia. 4°m T h e patients h a d significant excretion of dibasic and m o n o b a s i c m o n o c a r b o x y l i c a m i n o acids and exhibited n e u r o m u s c u l a r degeneration. A family with a generalized defect in glutathione reductase has b e e n reported. 42 T h e individuals h a d p r o b l e m s with cataracts and nonspecified o p h t h a l m o l o g i c a l problems. O t h e r r e p o r t e d cases of low glutathione reductase activity a p p e a r to be due to insufficient riboflavin in the diet. 43 33L. Berthe-Corti, R. Hulsch, U. Nevries, and F. Eckardt-Schupp, Mutagenesis 7, 25 (1992). 34M. Oort, J. A. Loos, and H. K. Prins, Vox Sang. 6, 370 (1961). 35H. K. Prins, M. Oort, J. A. Loos, C. Ziarcheer, and T. Becker, Blood 27, 145 (1966). 36D. N. Mohler, P. W. Majerus, V. Minnich, C. E. Hess, and M. D. Garrick, N. Engl. J. Med. 283, 1253 (1970). 37E. Buetler, T. Gelbart, and C. J. Pegelow, Clin. Invest. 77, 38 (1986). 38E. Jellum, T. Kluge, H. C. Borresen, O. Stokke, and L. Eldjarn, Stand. J. Clin. Lab. lnvest. 26, 327 (1970). 39A. Larsson, in "Glutathione: Metabolism and Physiological Function" (J. Vino, ed.), p. 360. CRC Press, Boca Raton, Florida, 1990. 40p. N. Konrad, F. H. Richards, W. N. Valentine, and D. E. Paglia, N. Engl. J. Med. 286, 557 (1972). 41 F. H. Richards, M. R. Cooper, L. A. Pearce, R. J. Cowan, and C. L. Spurr. Arch. Int. Med. 14, 535 (1974). 42j. A. Loos, H. D. Roos, R. Weening, and J. Houwerzijl, Blood 48, 53 (1976). 43G. W. Lrhr, K. G. Blume, H. W. Rfidiger, and H. Arnold, in "Genetic Variability in the Enzymatic Reduction of Oxidized Glutathione" (L. Flohe, H. BenOhr, H. Sies, H. O. Waller, and A. Wendel, eds.), p. 165. Thieme, Stuttgart, 1974.

92

GLUTATHIONE

[ 101

Conclusion Although E. coli serves as an excellent model for mutants defective in glutathione metabolism in general, additional phenotypes are observed in simple eukaryotes owing to the increased complexity of the cell structure and the fact that glutathione appears to serve many functions including its role in keeping cellular proteins in a reduced state. Additional functions for glutaredoxin appear in multicellular organisms that are more important than the functions present in E. coli.

[ 10] C y s t a t h i o n i n e By DONALD J. REED Introduction Cystathionine is a thioether-containing amino acid without a known function other than serving as an intermediate in the only known transsulfuration pathway in mammals for the biosynthesis of a nonessential amino acid, cysteine. Transsulfuration has been known for more than 50 years, first as a nutritional phenomenon and more recently as a potentially important factor in vascular disease. Methionine is capable of providing cysteine equivalents by the utilization of the sulfur atom of methionine and the carbon skeleton and amino group of serine. The capacity for transsulfuration was initially observed as a decreased requirement for dietary methionine when cystine was present in the diet. Principally as a result of work by Binkley and duVigneaud, 1 the role of the cystathionine pathway in the transsulfuration process was elucidated. Methionine is converted via Sadenosylmethionine by demethylation to homocysteine (Fig. 1).2 Homocysteine condenses with serine in a reaction catalyzed by cystathionine synthase to form the thioeter cystathionine. Cystathionine is subsequently cleaved by y-cystathionase to cysteine plus ammonia and ot-ketobutyrate. This pathway is a major source of cysteine in the liver during dietary limitation of cyst(e)ine or when a high rate of glutathione biosynthesis is required by t h e l i v e r 3 and a limited number of other cell t y p e s . 4 In mammals, such as rats, the liver is the main site of cysteine biosynthe1 F. Binkley and V. du Vigneaud, J. Biol. Chem. 144, 507 (1942). 2 j. D. Finkelstein and S. H. Mudd, J. Biol. Chem. 242, 873 (1967). 3 D. J. Reed and S. Orrenius, Biochem. Biophys. Res. Commun. 77, 1257 (1977). 4 A. E. Brodie, J. Potter, and D. J. Reed, Eur. J. Biochem. 123, 159 (1982).

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc. All tights of reproduction in any form reserved.

[ 1 0]

CYSTATHIONINE

93

CQOH I NH2-C-H I

c.~ ~

coo. I CH2

METHIONINE

I

c~-~ C~Hz

A.de

H 5-METHYLTETRA_~,.~[

H HO

t '%'"-"~C"3)3"C"2C00H

OH /~ACCEPTOR

\\

coo. c

CHz

HOMOCYSTEINE

Ade

.

HO

NH2--IC-H CIHz S I

CH2

OH

$-ADENOSYLHOMOCYSTEINE

"

I

H-C-- NH2 I

COOH CYSTATHIONINE

~-

COOH

COOH I

I

NH2-- IC-H CH2 I

SH CYSTEINE

C=O I

-'1"

CH2 I

-I-

NH3

CH3 CI-KETOBUTYRATE

FIG. 1. Cystathionine pathway and associated enzymes for the biosynthesis of L-cysteine. Enzymes catalyzing the steps are as follows: (1) methionine adenosyltransferase (EC 2.5.1.6), (2) S-Adenosylmethionine methyltransferases (EC 2.1.1.?), (3) adenosylhomocysteinase (EC 3.3.1.1.), (4) betaine-homocysteine S-methyltransferase (EC 2.1.1.5), (5) 5-methyltetrahydrofolate-homocysteine S-methyltransferase (methionine synthase) (EC 2.1.1.13), (6) cystathionine/3-synthase (EC 4.2.1.22), and (7) cystathionine 7-1yase (EC 4.4.1.1).

94

GLUTATHIONE

[101

sis which occurs via the cystathionine pathway. Normally, maintenance of a high glutathione content in the liver is associated with a very dynamic rate of secretion into plasma that is followed by an extensive and rapid extracellular degradation of both reduced glutathione (GSH) and oxidized glutathione (GSSG). These events support the concept that liver glutathione is a physiological reservoir of cysteine. This concept, originally proposed by Tateishi et al. 5 and Higashi et al., 6 involves two glutathione pools that are physiologically separate, one with a fast turnover in the cytosol and another, representing about 10% of the total liver glutathione, with a slow turnover in the mitochondria. 7 Evidence has led to the conclusion that cysteine may be an essential amino acid during premature growth s and even during surgical shock owing to the absence or a loss of function of the cystathionine pathway and transsulfuration of methionine sulfur. 9

Methods Sample Preparation Principle. Accurate measurement of cystathionine and related thiols and disulfides in biological specimens relies in part on sample preparation, namely, the rapid termination of metabolic processes in fluids, cells, and tissues and the prevention of thiol oxidation and thiol-disulfide interchange during the assay procedure. To satisfy these criteria, biological specimens are treated promptly with perchloric acid (PCA; 1 M or 10% v/v) containing a metal chelator such as 1 m M bathophenanthrolinedisulfonic acid (BPDS). Reagents

Perchloric acid (PCA), 70% (double distilled, G. F. Smith Chemicals) Perchloric acid (PCA), 10% (v/v) in metal-free water (Milli-Q Reagent Water System, Millipore, Bedford, MA) containing 1 m M bathophenanthrolinedisulfonic acid (BPDS) (G. F. Smith Chemicals) y-Glutamylglutamate (7-Glu-Glu; Vega Biochemicals) 15 mM, in 0.3% PCA, stored at - 1 0 °, stable at room temperature for several days, used as an internal standard 5 N. Tateishi,T. Higashi, A. Naruse, K. Nakashima,H. Shiozaki,and Y. Sakamoto,J. Nutr. 107, 51 (1977). 6T. Higashi, N. Tateishi, A. Naruse, and Y. Sakamoto,J. Biochem. 82, 117 (1977). 7M. J. Meredith and D. J. Reed, J. Biol. Chem. 257, 3747 (1982). 8 F. V. Pallardo, J. Sastre, M. Areseni, F. Rodrigo, J. M. Estrela, and J. Vina. Biochem. J. 274, 891 (1991). 9j. Vina, A, Girnenez, I. R. Puertes, E. Gasco, and J. R. Vian, Br. J. Nutr. 68, 421 (1992).

[101

CYSTATHIONINE

95

Bathophenanthrolinedisulfonic acid (BPDS, 15 mM), in metal-free water Standard Procedure. Cells, subcellular fractions in suspension, or cell culture media are treated by adding 50/~l y-Glu-Glu (15 raM), 50/~l BPDS (15 raM), and 100/~l 70% PCA. Following mixing with a vortex mixer, freezing, thawing, and centrifugation, a portion of the 10% PCA extract (0.5 ml) is derivatized and analyzed by high-performance liquid chromatography (HPLC) by the method of Reed et al. ~° as modified previously. T M O t h e r Methods. Stabler et al. t3 have examined the elevation of serum cystathionine levels in patients by a new capillary gas chromatographicmass spectrometric assay and measured cystathionine in the serum of normal subjects and patients with clinically confirmed deficiencies of cobalamin and folate. Ohmori et al.24 have described a sensitive determination of cystathionine and assays for cystathionine /3- and y-lyase, as well as cystathionine flsynthase, using HPLC. Cystathionine was cleaved to 2-ketobutyric acid, cysteine, and ammonia by cystathionase. 2-Ketobutyric acid was converted to 2-ethyl-2-hydroxy-6,7-dimethoxyquinoxaline (EHDQ) by reaction with 1,2-diamino-4,5-dimethoxybenzene. When E H D Q was measured in a mobile phase of pH 2.1 using HPLC with UV detection, 250 pmol of L-cystathionine in 250 ~l of the reaction mixture could be determined. Because E H D Q has a strong fluorescence in a mobile phase of pH 6.5 at 447 nm, on excitation at 365 nm, as little as 2.5 pmol of cystathionine in 250 /~l of the reaction mixture could be determined by HPLC with fluorimetric detection. Cystathionase activity was assayed on the basis of the same principle by determining cystathionine in as little as 63 ng of rat liver by fluorimetric detection. Cystathionine/3-synthase activity was measured by the same method by determining cystathionine formed in only 113 ng wet weight of rat liver. Using these methods, both cystathionine fl- and y-lyase activities in S a c c h a r o m y c e s cerevisiae were determined, because quinoxaline derivatives from pyruvate and 2-ketobutyrate could be measured simultaneously by HPLC. Watanabe et al. 15 have developed qualitative analyses of sulfur amino 10D. J. Reed, J. R. Babson, P. W. Beatty, A. E. Brodie, W. W. Ellis, and D. W. Potter, Anal. Bioehem. 106, 55 (1980). 11 M. W. Fariss and D. J. Reed, this series, Vol. 143, p. 101. 12 D. J. Reed, Methods Toxieol. 1B, 421 (1994). 13 S. P. Stabler, J. Lindenbaum, D. G. Savage, and R. H. Allen, Blood 81, 3404 (1993). 14 S. Ohmori, K. Nakata, K. Nishihara, S. Yamamoto, M. Kawase, and S. Tsuboi, J. Chromatogr. 574, 35 (1992). 15 H. Watanabe, Y. Fujita, K. Sugahara, H. Kodama, and S. Ohmori, Biol. Mass Spectrom. 20, 602 (1991).

96

GLUTATHIONE

[ 1O]

acids in the urine of a patient with cystathioninuria using liquid chromatography/atmospheric pressure ionization mass spectrometry. Standard sulfur amino acids and various cystathionine metabolites in the urine of a patient with cystathioninuria were analyzed using liquid chromatography/ mass spectrometry with an atmospheric pressure ionization interface system. Very intense quasi-molecular ions ([M + HI ÷) of synthetic cystathionine, N-monoacetylcystathionine, perhydro-l,4-thiazepine-3,5-dicarboxylic acid, S-(3-hydroxy-3-carboxy-n-propyl)cysteine, S-(2-carboxyethyl)cysteine, S-2-hydroxy-2-carboxyethyl)homocysteine, S-(carboxymethyl)homocysteine, N-acetyl-S-(3-hydroxy-3-carboxy-n-propyl)cysteine, and N-acetylS-(2-carboxyethyl)cysteine were observed by this method. Quasi-molecular ions ([M + H] +) of the sulfur amino acids were observed in the urine sample of the patient with cystathioninuria, and N-acetyl-S-(3-hydroxy-3carboxy-n-propyl)- and N-acetyl-S-(2-carboxyethyl)cysteine as N-substituted sulfur amino acids were also identified in the urine of the same patient. Sugahara et al. 16 have determined cystathionine and perhydro-l,4-thiazepine-3,5-dicarboxylic acid in the urine of a patient with cystathioninuria using column liquid chromatography-mass spectrometry. The concentrations found for cystathionine and perhydro-l,4-thiazepine-3,5-dicarboxylic acid were 1.289 + 0.099 and 0.310 _+ 0.067 mg/ml, respectively. Nutritional Biochemistry of Cystathionine and Related Amino Acids Methionine is an essential amino acid for mammals and must be derived entirely from the diet. In contrast, cysteine is synthesized by mammals, with the cysteine sulfur being derived from methionine and the carbon and nitrogen of cysteine from serine via the cystathionine pathway that functions principally in the liver. The dependence on methionine sulfur implies that cysteine remains a nonessential amino acid only if methionine intake is sufficient to meet the total sulfur amino acid requirement. Conversely, the dietary requirement for methionine is reduced, but not eliminated, when cyst(e)ine intake is adequate. 17 Another sulfur compound, the tripeptide glutathione, is present in many food and feed products and is capable of delivering 100% of its cysteine content to the animal as bioavailable cysteine. Rats, mice, dogs, and pigs have relatively low dietary requirements for sulfur amino acids compared to chicks and catsJ 7 In all the species, cyst(e)ine has been found capable of supplying 50% of the total requirement (w/w) for growth. Although transsulfuration is a biologically efficient process, unless interfered with by inhibition, a higher concentration of methio16K. Sugahara, J. Ohta, M. Takemura, and H. Kodama, I. Chromatogr. 579, 318 (1992). 17 D. H. Baker, this series, Vol. 143, p. 297.

[ 10]

CYSTATHIONINE

97

nine alone is needed to furnish the same sulfur content as a 1 : 1 mixture (w/w) of a methionine-cyst(e)ine combination owing to the molecular weight difference between methionine and cysteine. Excess dietary methionine is generally considered to be the most toxic of the (dietary) essential amino acids. A level twice the requirement is generally well tolerated, but at 3-fold or above levels toxicity results) 7 Fasting or nitrogen-free diet feeding to cause depletion of the sulfur amino acids from tissue pools leads to a greater depletion of sulfur amino acids compared to other essential amino acids. Excess dietary cyst(e)ine is well tolerated even though intracellular concentrations of cysteine are of the order of 30-200 ~M, values that are among the lowest observed for any of the protein amino acids.5 Higher concentrations of cysteine are potentially dangerous to cells because cysteine forms a thiazolidine derivative with pyridoxal phosphate and thereby may deplete cells of that coenzymeJ s Multiple pathways of cysteine catabolism function efficiently to prevent toxic accumulation of cysteine. Glutathione, a physiological reservoir of cysteine, is present in liver cells and other cells in a nontoxic, metabolically active form that participates in an interorgan process to supply cysteine rapidly to all cells in the body. Methionine metabolism is initiated by the formation of S-adenosylmethionine (AdoMet), a sulfonium compound in which each of the carbons attached to sulfur is activated toward nucleophilic attack. In most cases, AdoMet reacts by transfer of the S-methyl group to one of several possible acceptors, with methyl transfer reactions as a group thought to consume about 95% of the AdoMet f o r m e d . 19 Pallardo and co-workers 8 have examined physiological changes in glutathione metabolism in fetal and newborn rat liver. Isolated hepatocytes from fetal, newborn, and adult rats were utilized to examine glutathione metabolism. The GSH/GSSG ratio decreased 15- to 20-fold through the fetal-neonatal-adult transition. The change in the GSH/GSSG ratio was mainly due to an increase in GSSG. It was observed that all enzyme activities involved in the glutathione redox cycle increased during that transition but the relative increases in glutathione peroxidase and glutathione Stransferase were 3-5 times those of glutathione reductase or glucose-6phosphate dehydrogenase. Synthesis of GSH from methionine as a sulfur source was 6 times lower in fetal than in adult hepatocytes. When Nacetylcysteine was used as a sulfur donor to bypass the cystathionine pathway, the rates of GSH synthesis were similar in fetal and adult cells. It was concluded that this was a result of cystathionase activity being very low in is O. W. Grfffith, this series, Vol. 143, p. 366.

19S. H. Mudd and J. R. Poole, Metab. Clin. Exp. 24, 721 (1975).

98

GLUTATHIONE

[10]

fetal cells. The low activity is reflected in the blood amino acid pattern, where the concentration of cysteine rises from 8 to 52/zM from fetuses to adult rats. It was concluded that cysteine may be an essential amino acid for the premature animal. Stabler e t a l J 3 have observed elevation of serum cystathionine levels in patients with cobalamin and folate deficiency. Homocysteine can be methylated to form methionine by the cobalamin and folate-dependent enzyme methionine synthase (5-methyltetrahydrofolate-homocysteine Smethyltransferase). Serum levels of total homocysteine are elevated in greater than 95% of patients with either cobalamin or folate deficiency. Homocysteine can condense with serine to form cystathionine in a pyridoxal phosphate-dependent reaction catalyzed by cystathionine fl-synthase. Cystathionine is subsequently cleaved to cysteine and ot-ketobutyrate by the pyridoxal phosphate-dependent enzyme T-cystathionase. The normal range for serum cystathionine was 65 to 301 nmol/liter (median of 126 nmol/ liter) for 50 normal blood donors. In 30 patients with clinically confirmed cobalamin deficiency, values for cystathionine ranged from 208 to 2920 nmol/liter (median of 816 nmol/liter) and 26 (87%) had levels above the normal range. In 20 patients with clinically confirmed folate deficiency, cystathionine concentrations ranged from 138 to 4150 nmol/liter (median of 1560 nmol/liter) and 19 (95%) had values above the normal range. Five homozygotes for cystathionine/3-synthase deficiency had high values for serum total homocysteine and low or low-normal values for serum cystathionine that ranged from 30 to 114 nmol/liter even though they were on treatment with pyridoxine and had partially responded. It was concluded that levels of cystathionine are elevated in the serum of most patients with cobalamin and folate deficiency and that they are useful in the differential diagnosis of an elevated serum total homocysteine level. Oxidative Stress Effects Because elevated circulating homocyst(e)ine is a risk factor for occlusive ascular disease, Dudman et al. 2° have measured circulating lipid hydroperoxide levels in human hyperhomocysteinemia. They explored whether elevated plasma homocyst(e)ine is associated with increased plasma lipid hydroperoxides that might trigger vascular disease. Plasma containing high levels of homocyst(e)ine was obtained from four patients with a homozygous deficiency of cystathionine /3-synthase activity and also from four heterozygotes with a deficiency of the enzyme after an oral methionine 20 N. P. Dudman, D. E. Wilcken, J. Wang, J. F. Lynch, D. Macey, and P. Lundberg, Arterioscler. Thromb. 13, 1253 (1993).

[ 101

CYSTATHIONINE

99

load. The mean plasma nonprotein-bound homocyst(e)ine level in all subjects was more than 11-fold higher than the mean normal fasting value. Levels of high-density lipoprotein (HDL) cholesteryl ester hydroperoxides (CEOOH), normalized against the concentration of free cholesterol in HDL, were not elevated compared with values for fasting healthy donors. An inverse dependency was observed between plasma total homocyst(e)ine and HDL CEOOH. The ubiquinol-10/ubiquinone-10 ratio in HDL, which is expected to fall during oxidative stress, increased with plasma homocyst(e)ine. Because HDL contains the majority of detectable plasma lipid hydroperoxides, of which CEOOHs are the most abundant, Dudman et al. z° suggest that an elevated plasma homocyst(e)ine level does not enhance oxidative stress, increase the levels of lipid hydroperoxides in plasma, or generate vascular damage by this mechanism. Rusakow et al. 21 examined Oz-induced changes thiols in mice. Because glutathione is an important antioxidant, they hypothesized that changes in lung and systemic availability of GSH and its precursor amino acid, cysteine, are induced by exposure to hyperoxia and that the changes could be modulated by toxic Oz metabolites. In relatively Oz-resistant Swiss-Webster mice, lung GSH increased during O2 exposure, whereas liver GSH and liver and plasma cysteine all decreased. Pair-feeding studies suggested that nutritional deprivation alone did not cause the decrease in plasma cysteine. In lung, sulfur amino acids were not decreased by O2 exposure. Cystathionine increased 6-fold, and y-cystathionase was not inhibited. These findings suggest that hyperoxia increases transsulfuration pathway activity and that cystathionase rate-limits this process in lung. In comparative studies, lung GSH increased in Oz-resistant high-Cu,Zn-superoxide dismutase (SOD) transgenic mice but not in genetically similar, nontransgenic controls (CBYB/6 × B6D/2) during hyperoxic exposure. In addition, liver GSH and plasma cysteine decreased in nontransgenic control but not in highSOD mice, whereas lung cystathionine increased similarly in both groups. Thus, superoxide or its secondary products can modulate, at least in part, the changes in cysteine and GSH. Nonetheless, regardless of strain or SOD status, hyperoxic exposure consistently caused thiol and sulfur amino acid changes, including increased lung cystathionine and GSSG, demonstrating a strong association between these dynamic changes and oxidant stress. Vina and co-workers9 have examined the impairment of cysteine synthesis from methionine in rats exposed to surgical stress. The activity of liver cystathionase (EC 4.4.1.1) was decreased after 3 days of stress induced by surgery. The rate of L-cysteine synthesis from L-methionine was significantly higher in isolated hepatocytes from controls than in hepatocytes from rats 21 L. S. Rusakow, C. W. White, and S. P. Stabler, J.

Appl. Physiol. 74, 989 (1993).

100

OLUTATHIONE

[ 1 O]

suffering from surgical stress. The half-life of L-[2(n)-3H]methionine was significantly higher in rats subjected to surgical stress than in controls. The plasma L-methionine/L-cystine ratio was higher in stressed rats than in controls. L-Cystine was increased significantly in the surgically stressed rats compared with the controls. From these observations, it was concluded that the data were consistent with the hypothesis that the observed inhibition of cystathionase is physiologically important and that L-cysteine might be considered as an essential amino acid in cases of surgical stress. Takeuchi and co-workers 22have examined glutathione levels and related the enzyme activities in vitamin B6-deficient rats fed a high methionine and low cystine diet. Vitamin B6-deficient rats were fed a vitamin B6-deficient diet containing 0.56% (w/w) methionine and 0.075% cystine for 8 weeks. Controls were fed an identical diet supplemented with 10 mg/kg pyridoxine hydrochloride. Glutathione concentrations in each organ examined were similar in control and vitamin B6-deficient rats, and the values were comparably lower after intraperitoneal injection of diethyl maleate, a glutathione S-transferase substrate that causes depletion of glutathione. However, buthionine sulfoximine, an inhibitor of glutathione biosynthesis, caused a significantly greater decrease in glutathione levels in the liver and lungs of vitamin B6-deficient rats relative to controls. Glutathione peroxidases activity in the liver of vitamin B6-deficient rats was higher than in control animals; however, glutathione transferase activity in tissues other than liver of vitamin B6-deficient rats was higher than in the controls. The holoenzyme activities of cystathionine/3-synthase and cystathionine ~/-lyase in the liver of vitamin B6-deficient rats were markedly reduced. These findings indicated that although the activities of enzymes that synthesize cysteine from methionine were decreased by vitamin B6 deficiency, the level of synthesis and supply of cysteine in vitamin B6-deficient rats were sufficient to maintain the same glutathione level as in controls, and that glutathione utilization in the liver was accelerated by vitamin B6 deficiency. Effects of Propargylglycine in Laboratory Animals Abeles and Walsh 23 synthesized and tested a number of acetylenic enzyme substrates and observed the irreversible inactivation of the pyridoxal phosphate-dependent rat liver enzyme y-cystathionase. The irreversible inactivation is accompanied by covalent labeling of the enzyme to the extent of I m01 of labeled inactivator per 80,000 daltons, equivalent to two subunits. Subsequently, an inactivator-enzyme adduct was isolated. 24 The 22 F. Takeuchi, S. Izuta, R. Tsubouchi, and Y. Shibata, J. Nutr. 121, 1366 (1991). 23 R. H. Abeles and C. T. Walsh, J. Am. Chem. Soc. 95, 6124 (1973). 24 W. Washtien and R. H. Abeles, Biochemistry 16, 2485 (1977).

[ lOl

CYSTATHIONINE

101

inactivator, DL-2-amino-4-pentynoic acid (propargylglycine, PPG), has been shown to rapidly inactivate the rat liver cystathionase in vivo when administered to mice and leads to a rapid decrease in the hepatic activity of the enzyme. In this manner, experimental cystathionuria has been induced by a single intraperitoneal administration of propargylglycine to mice. An extensive decrease in the hepatic cystathionase activity occurs within 1 hr and is maintained at a low level of enzyme activity for I day. It is accompanied by a considerable increase in urinary excretion of cystathionine, glycine, and alanine. The strict specificity and high efficiency of propargylglycine to inactivate y-cystathionase appears to ensure quantitative information concerning the actual metabolic significance of the target enzyme. This concept is supported by the fact that the concentration of cystathionine in liver, kidney, and brain increased rapidly on the administration of L-propargylglycine. Glutathione in liver diminished in parallel with the increase in cystathionine, reaching 40% of the original level when the cystathionine level was at a maximum. Following the restoration of the hepatic cystathionase activity, the cystathionine level declined gradually with concomitant increase in the glutathione level. These findings indicate that, in liver, the cystathionine pathway supplied more than half of the cysteine store for the biosynthesis of glutathione. Unlike the case with liver, this metabolic correlation between cystathionine and glutathione is not explicit in kidney and brain, suggesting that the supply of cysteine for the maintenance of glutathione in these tissues is from a source other than the cystathionine pathway. New studies continue to support the concept of liver glutathione being the physiological reservoir for cysteine as it is applied to the other cell types in mammals. Acetylenic substrate analogs are useful in the inactivation of enzymes which abstract carbon-bond hydrogens as protons at a position adjacent to the acetylenic linkage. Specificity of the interactions is a very attractive feature, as it is not based solely on structural similarity to substrate, but also on the mechanism of the catalytic process. It is important to point out that the acetylenic linkage is made chemically reactive as a result of specific enzymatic action on the inhibitor molecule. Inactivators are very useful as active site-directed reagents for enzymes and serve as very specific agents for in vivo inactivation of these specific target enzymes. Propargylglycine is such a potent irreversible inhibitor that only 6.6 × 10 -5 M inactivator is necessary to cause the loss of enzymatic activity of y-cystathionase with a half-life of 2 min at 25 °. In vivo, a single intraperitoneal administration of L-propargylglycine (5 txmol/mouse) leads to almost complete inactivation of y-cystathionase activity within 1 hr. e5 25S. Shinozuka, S. Tanase, and Y. Morino, Eur. J. Biochem. 124, 377 (1982).

1 02

GLUTATHIONE

[ 1 0]

Houghton and Cherian 26 have examined the effects of inhibition of cystathionase activity on glutathione and metallothionein levels in the adult rat. Rats were fed a diet containing methionine (0.66% (w/w)) and cystine (0.20%) for 1 week before receiving three consecutive daily intraperitoneal injections of propargylglycine, at various doses (2.5-375/xmol/kg). They observed that when hepatic cystathionase was inhibited greater than 90% (->50/~mol propargylglycine/kg), renal and hepatic metallothionein and hepatic glutathione were unaltered except at the highest dose. In contrast, renal glutathione was increased 2-fold. The authors concluded that renal and hepatic glutathione pools appear to be altered differently and that renal glutathione may be preferentially maintained even when hepatic glutathione is decreased. Cho and co-workers 27 have examined propargylglycine infusion effects on tissue glutathione levels, plasma amino acid concentrations, and tissue morphology in parenterally fed, growing rats. Amino acid solutions currently used for total parenteral nutrition (TPN) contain little cysteine or cystine. Some premature human infants have low liver activities of y-cystathionase and presumably require preformed cysteine or cystine. Growing animals tend to have higher liver y-cystathionase activity, which makes them unsuitable as models to study effects of cystathionine precursors. Because propargylglycine inhibits y-cystathionase specifically, rats infused with PPG as part of a TPN regimen were evaluated as a potential model. Two groups of rats (120-160 g) were infused for 15 days with TPN regimens, one without and one with PPG (40 ~mol/day). A third group received the TPN control regimen, with methionine added at toxic levels. Propargylglycine treatment significantly decreased plasma cysteine and taurine concentrations and significantly increased plasma cystathionine concentration without affecting methionine concentration. Propargylglycine treatment significantly decreased brain, muscle, liver, intestine, and stomach glutathione concentration without affecting erythrocyte or heart glutathione concentrations. Electron microscopic examination showed no abnormalities in heart and kidney of PPG-treated rats. Hepatocyte glycogen was lower in TPNfed controls than in orally fed rats and was further reduced in TPN-PPGfed animals. Growing rats infused with low doses of PPG show promise as an animal model to study a numbers of important issues concerning human sulfur amino acid metabolism. Acknowledgment The author generally acknowledges the support of NIEHS Grants ES-01978 and ES-00210. 26 C. B. Houghton and M. G. Cherian, I. Biochem. Toxicol. 6, 221 (1991). 27 E. S. Cho, J. Brown-Hovanec, Tomanek, and L. D. Stegink, I. Nutr. 121, 785 (1991).

[11]

PROMOTION OF CYST(E)INEUPTAKE

[ 1 II P r o m o t i o n By

of Cyst(e)ine

103

Uptake

THOMAS MEIER and ROLF D. ISSELS

Introduction Besides its involvement in protein synthesis, the amino acid cysteine has a central function in the cellular metabolism of glutathione. The tripeptide glutathione, which can be defined as the intracellular "package form" of cysteine, is limited in regard to its biosynthesis by the availability of cysteine. Therefore, under conditions where the supply of cellular cysteine is not sufficient, a promotion of cysteine uptake by other chemical compounds would be expected to increase glutathione levels. The importance of glutathione for cellular metabolism is well known. 1 Glutathione can protect cells against oxidative stress, heavy metal ions, and various toxic compounds. Many studies show that depletion of cellular glutathione levels leads to sensitization of cells to radiation and antineoplastic agents 2 and vice versa. The increase of cellular glutathione often corresponds to an increased resistance to these therapeutical modalities. In most eukaryotic cell types cysteine has to be taken up from the extraceUular medium, but some cells (e.g., hepatocytes) are able to synthesize cysteine from methionine via the cystathionine pathway. For promotion studies of cysteine uptake different approaches can be taken. Direct supplementation of cells with cysteine is critical because cysteine exhibits a high degree of toxicity. 3 Several nontoxic cysteine derivatives have been proposed as delivery systems for cysteine. Administration of thiazolidines, ring form cysteine derivatives, can promote glutathione biosynthesis. 4 After entering the cells thiazolidines become intracellularly cleaved and release cysteine. Several cysteine esters can act as cysteine delivery systems and elevate intracellular cysteine and glutathione. -~In addition, administration of extracellular monoethyl or diethyl esters of glutathione rather than glutathione itself can also enhance intraceUular glutathione levels. 6 There is no evidence that intact glutathione can be taken up by mammalian cells] 1A. Meister and M. E. Anderson, Annu. Rev. Biochem, 52, 711(1983). 2 A. Meister, PharrnacoL Ther. 51, 155 (1991), 3y. Nishiuch, M. Sasaki, M. Nakayasu, and A. Oikawa, In Vitro 12, 635 (1976). 4j. M. Williamson,B. Boettcher, and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 79, 6246 (1982). _sM. Butterworth, D. G. Upshall, M. Hobbs, and G. M. Cohen, Biochem. Pharmacol. 45, 1769 (1993). 6E. J. Levy, M. E. Anderson, and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 90, 9171 (1993). 7M. E. Anderson and A. Meister, this series, Vol. 143, p. 313.

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

104

GLUTATHIONE

[ 11 ]

However, extracellular or membranous metabolism of glutathione may form products such as ~-glutamylcysteine, cysteinylglycine, or cysteine which may be transported into cells and utilized for glutathione biosynthesis. 7 For a promotion of cysteine uptake the cellular transport systems are of critical importance. In aqueous solutions rapid autoxidation of most of the free thiol cysteine to cystine occurs. Therefore under physiological conditions only 10-20% of the total of free cystine is present as the reduced form. 8 After entry into a cell a prompt subsequent reduction of cystine to cysteine may be expected. Cystine and cysteine appear to be taken up by different transport systems, and no direct interaction between the systems has yet been found. In general, the transport systems vary considerably among tissues. Three transport systems are known and are explained in more detail. The uptake of cystine can be mediated by the transport system xc- in exchange with glutamate. 9 The uptake system is Na÷-independent and further shared by homocysteate and a-aminoadipate. Some cells express high activity for this anion-exchanging transport (e.g., fibroblasts), in other cells it is strongly inducible (e.g., macrophages), whereas other cell types lack this transport (e.g., lymphocytes). 1°-12 Cysteine is transported into cells by a transport system totally different from xc-. The system ASC (alanine, serine, cysteine) transports neutral amino acids and is Na÷-dependent. 13 Alanine, serine, and cysteine share this transport system, which is ubiquitously distributed among cell types. Exracellular cysteine can also spontaneously react with other thiol compounds to form mixed disulfides under certain conditions. The disulfides can be taken up by the transport system L, which was demonstrated for the disulfide of cysteamine and cysteine to be Na÷-independentJ 4 More detailed descriptions of the amino acid transport systems for cyst(e)ine uptake are reviewed elsewhere. 8'~5 To study the promotion of cyst(e)ine uptake of mammalian cells in vitro we use Chinese hamster ovary (CHO) cells as a model system. The cells have a limited capability to utilize extracellular cystine. 16 The uptake of cystine can be investigated when extracellular cystine is replaced by the 8 S. Bannai, Biochim, Biophys. Acta 779, 289 (1984). 9 M. Makowske and H. N. Christensen, J. Biol. Chem. 257, 5663 (1982). 10T. Ishii, H. Sato, K. Miura, J. Sagara, and S. Bannai, Ann. N. Y. Acad. Sci. 663, 497 (1992). 11 H. Watanabe and S. Bannai, J. Exp. Med. 165, 628 (1987). 12 H. Gmunder, H.-P. Eck, and W. Droge, Eur. J. Biochem. 201, 113 (1991). 13H. N. Christensen, M. Liang, and E. G. Archer, J. Biol. Chem. 242, 5237 (1967). 14R. L. Pisoni, J. G. Thoene, and H. N. Christensen, J. Biol. Chem. 260, 4791 (1985). 15 H. N. Christensen, Physiol. Rev. 70, 43 (1990). 16 R. D. Issels, A. Nagele, K. G. Eckert, and W. Wilmanns, Biochem. Pharmacol. 37, 881 (1988).

[1 ll

PROMOTION OF CYST(E)INE UPTAKE

105

70

r

60

~'

5o 40

30 ~,

20

ffl

oi

0

10

20

30

40

50

60

Time (rain)

FIG. 1. Time course (0-60 min) of the uptake of 0.1 mM [3SS]cystine by CHO cells in McCoy's 5A medium at 37°. The medium contained the following supplements: (Q) 0.4 mM cysteamine; (11) 0.8 mM N-acetylcysteine; ( ~ ) no additional thiol (control). Data are means _+ SD of three independent experiments. [From R. D. Issels, A. Nagele, K. G. Eckert, and W. Wilmanns, Biochem. PharmacoL 37, 881 (1988).]

radioactively labeled compound. Addition of several thiols to the medium of CHO cells can promote uptake of cystine as shown in Fig. 1. A pronounced increase in [35S]cystine uptake compared to untreated control cells is observed when the thiols cysteamine and N-acetylcysteine are added to the medium. The promotion of cystine uptake has also been studied for other thiols. A selection of thiols that have been investigated are summarized in Table I. Dihydrolipoic acid, WR-1065, and Mesna promote cystine uptake; in contrast, a-lipoic acid, N,N-diacetylcystine, and WR-2721 show no effect. The common molecular structure shared by such compounds that promote cystine uptake in CHO cells is the free SH group. Modification of the free thiol group by chemical reaction to form a disulfide bond or the addition of a phosphate moiety to the thiol group abolishes the potential of the compound to promote cystine uptake in CHO cells. Utilization of Cystine for Glutathione Biosynthesis Addition of extracellular thiols to the medium has been demonstrated to influence the intracellular glutathione levels of CHO cells. 17As shown for cysteamine and N-acetylcysteine, intracellular glutathione content increases during thiol exposure (see Table II). Total cellular glutathione of C H O cells 17 R. D. lssels, S. Bourier, J. E. Biaglow, L. E. Gerweck, and W. Wilmanns, Cancer Res. 45, 6219 (1985).

106

[ 11]

GLUTATHIONE TABLE 1 CYSTINE UPTAKE IN CHO CELLS

Thiol"

Promoting cystine uptake N-Acetylcysteine Mesna WR-1065

Cysteamine Dihydrolipoic acid Not promoting cystine uptake WR-272t

a-Lipoic acid

Formula

HOOCCHCHzSH

I

CH3OCNH SO3CH2CH2SH H2NCH2CH2CHzNHCH2CH2SH H2NCHzCHzSH SH SH HOOCCH2CH2CH2CH2CHCHeCH2 H2NCH2CH2CH2NHCH2CH2SPO3 HOOCCH2CH2CH2CH2CH-- CH2----~CH 2

~s__s j N,N-Diacetylcystine

Ref.

d

c

HOOCCHCH2SSCH2CHCOOH CH3OCNH

NHCOCH3

a Mesna, 2-Mercaptoethane sulfonate; WR-1065, N-(2-mercaptoethyl)-l,3-diaminopropane; WR-2721, S-2-(3-aminopropylamino) ethylphosphorothioic acid. b R. D. Issels, A. Nagele, K. G. Eckert, and W. Wilmanns, Biochem. Pharmacol. 37, 881 (1988). c T. Meier and R. D. Issels, Biochem. Pharmacol., submitted for publication. a R. D. Issels and A. Nagele, Cancer Res. 49, 2082 (1989).

increases about 2-fold after exposure to the thiols. Buthionine sulfoximine (BSO), an irreversible inhibitor of glutathione biosynthesis, when added to the medium immediately prior to the addition of cysteamine or Nacetylcysteine could completely block the increase of glutathione. This demonstrates that the increase in glutathione content mediated by cysteamine and N-acetylcysteine is dependent on cellular glutathione biosynthesis. The utilization of radioactively labeled cystine from the medium for glutathione biosynthesis is shown in Fig. 2. More than 90% of the 35S radioactivity in the soluble fraction of the cell lysates could be recovered in the peak of glutathione. These data show that the cysteine moiety of the newly synthesized glutathione is derived from the extracellular pool of cystine. The promoted uptake of cystine in the presence of cysteamine or Nacetylcysteine is dependent on amino acid transport systems of CHO cells. Addition of L-alanine, L-serine, and L-leucine reduced the uptake rate of

[11]

PROMOTION OF CYST(E)INE UPTAKE

107

TABLE II GLUTATHIONE CONTENT OF C H O CELLS AFTER EXPOSURE TO CYSTEAMINE AND N - A C E T Y L C Y S T E I N E a

Treatment

GSH + GSSG

Controls +BSO 0.4 mM Cysteamine +BSO" 0.8 mM N-Acetylcysteine +BSO ~

26.3 _+ 5.6b 17.5 _+ 0.9 61.0 +_ 14.0 16.7 _+ 0.4 41.4 _+ 18.9 16.0 ± 1.4

"After exposure to thiols for 120 rain at 37°. b Means ± SD of three or more independent experiments. c Addition of 1 mM DL-buthionine (S,R)-sulfoximine (BSO) occurred 10 min prior to the addition of cysteamine or N-acetylcysteine. [Modified from R. D. lssels, A. Nagele, K. G. Eckert, and W. Wilmanns, Biochem. Pharmacol. 37, 881 (1988).] cystine, whereas L-glutamate or DL-homocysteate showed no effect. Therefore, part of the uptake p r o m o t e d by cysteamine or N-acetylcysteine occurs most likely in the form of cysteine, which is mainly transported by the ASC system in most cell types. Besides uptake via the A S C system, uptake of mixed disulfides (e.g., cysteine-cysteamine, cysteine-N-acetylcysteine) via system L is also likely. In conclusion, the p r o m o t i o n of cystine uptake induced by extracellular thiols in our model system needs two different steps: first, the formation of cysteine and cysteine mixed disulfides by reaction of the added thiol with extracellular cystine and, second, the uptake of cysteine and/or the mixed disulfide via cellular transport systems. I n Vitro A s s a y of T h i o l - P r o m o t e d C y s t i n e U p t a k e

T w o different experimental protocols are described that can be used to study adherent or n o n a d h e r e n t cell types. A d h e r e n t Cells

The protocol is optimized for C H O cells and might need slight modifications when other cell types are used. A r o u n d 5 × 105 exponentially growing C H O cells are seeded in polystyrene tissue culture petri dishes (35 m m

108

111]

GLUTATHIONE

A

S-12,/,- dinitrophenyl) glutnthione (GS-DNP) 0.01 E3/.0nm

~_a-"

L

I- chloro-2,4-dinitrobenzene

I

B

,

~.

A

.

50000

5 E

@

M

r

i

0

2

i

i

[

1

4

6

8

10

i

12

i

14

i

16

Time {rain}

FIG. 2. Recovery of [3SS]cystine incorporated into S-(2,4-dinitrophenyl)glutathione (GSDNP). (A) Typical elution profile of GS-DNP at 340 nm after HPLC on a/zBondapak C18 column. After cysteamine exposure of CHO cells for 60 min at 37 ° in McCoy's 5A medium containing 0.1 mM [35S]cystine, total GSH of the acid-soluble fraction was transferred to 1chloro-2,4-dinitrobenzene by glutathione S-transferase, forming GS-DNP as described. (B) Measurement of 35S radioactivity in fractions of the eluate after HPLC separation. More than 90% of the total 35S radioactivity applied to the column could be recovered in the peak of GS-DNP. [From R. D. Issels, A. Nagele, K. G. Eckert, and W. Wilmanns, Biochem. Pharmacol. 37, 881 (1988).]

diameter) containing 2 ml of complete McCoy's 5A medium. Twenty-four hours after plating the cell number is 1.5-2 × 106 per dish with a confluency of about 50%. The cells are washed twice with warm phosphate-buffered saline (PBS), and 0.5 ml of the uptake medium is added. The uptake medium consists of McCoy's 5A medium without cysteine, cystine, methionine, and glutathione and is supplemented with 0.1 mM L-[35S]cystine (50-120 mCi/ mmol, Amersham, Braunschweig, Germany). Concentrated stock solutions of thiols are freshly prepared in ice-cold (PBS) which has been previously

[1 I I

PROMOTION OF CYST(E)INE UPTAKE

109

gassed with nitrogen for about 30 min. The thiol stock solution is added in small volumes (5-20/xl) to reach the final concentration. The dishes are immediately placed on a temperature-controlled water bath, the bottom of the dishes being submersed in the water. After the incubation period, the radioactive medium is removed and the cell layer washed three times with ice-cold PBS. The cells are trypsinized (0.25% v/v trypsin) for 2 min at room temperature, and the cell number of an aliquot is determined. The rest of the suspension is centrifuged (500 g, 5 min, at 4 °) and the cell pellet is carefully resuspended with 0.5 ml of 1 N perchloric acid. After incubation for 10 min on ice, the solution is centrifuged at high speed. The acidinsoluble pellet, which is solubilized in 0.5 ml of 0.1 N NaOH for 12 hr, and the acid-soluble supernatant are mixed with a appropriate scintillation cocktail and separately counted in a/3-counter. An aliquot of the NaOH solution can be used for protein determination by the method of Lowry. N o n a d h e r e n t Cells

The cystine uptake in nonadherent cells is determined by using a modification of the method described by Novogrodsky et al.18 Mitogen-activated peripheral blood lymphocytes (PBL) (107 cells/ml) are suspended in the same radioactive uptake medium as described for the adherent cells. The suspension is incubated with shaking in a temperature-controlled water bath, and the thiol stock solution is added in a small volume. Aliquots (100 ixl) are removed at given time points and layered on 0.1 ml of a mixture of mineral oil and dibutyl phthalate (1:4, v/v). After centrifugation in a Beckman Microfuge for 15 sec the Beckman reaction vial is frozen in liquid nitrogen. Then the tip of the vial containing the cell pellet is cut off with a razor. The pellet is solubilized with 1 ml Biolute-S (Zinsser, Frankfurt, Germany) overnight, then mixed with an appropriate scintillation cocktail and counted in a/3-counter. Determination of Intracellular Glutathione Total gluathione including the reduced (GSH) and oxidized (GSSG) form is quantified using the high-performance liquid chromatography (HPLC) method of Reed et aL 19 Briefly, the cell pellet is extracted with 0.5 ml of 1 N perchloric acid, and y-glutamyl glutamate is added as an internal standard. The extract is incubated for 120 rain with iodoacetic i8 A. Novogrodsky, R. E. J. Nehring, and A. Meister, Proc. Natl. Acad. Sci. U.S.A. 76, 4932 (1979). 19 D. J. Reed, J. R. Babson, P. W. Beatty, A. E. Brodie, W. W. Ellis, and D. W. Potter, Anal. Biochern. 106, 55 (1980).

110

GLUTATHIONE

[ 1 1]

acid at pH 8.5 to form S-carboxymethyl derivatives of free thiol groups. Subsequently, the amino groups are derivatized with 1-fluoro-2,4-dinitrobenzene (Sanger's reagent). Aliquots of the reaction mixture are injected onto a/xBondapak amine column (4 × 250 mm, Waters, Eschborn, Germany) and eluted with a sodium acetate gradient (flow rate of 2 ml/min) in a water-methanol-acetic acid solvent at pH 4.5. The dinitrophenyl derivatives are detected at 365 nm. The GSH and GSSG derivatives are quantified in relation to the internal standard. M e a s u r e m e n t of [a~S]Cystine Incorporated into Glutathione Recovery of [3SS]cystine incorporated into GSH is determined by HPLC and scintillation counting as described. 2° The incubation of cells in [35S]cystine-containing McCoy's 5A medium is essentially the same as described for the uptake method (see above). The cell pellet is extracted with icecold I N perchloric acid containing i mM EDTA (50/zl). The supernatant is neutralized with 22 /xl of 1.57 M K3PO4, vortexed, and the resulting precipitate of KCIO4 pelleted after 4 min on ice. The GSSG is reduced by adding 6/zl of a mixture of 20/zl of 100 mM glucose 6-phosphate (in water), 20 /zl of i mM N A D P H (in I% NaHCO3), 20 /zl of 5 mg/ml glucose6-phosphate dehydrogenase (from yeast, 140 U/rag, in water), 20/zl of glutathione reductase (from yeast, suspension of 120 U/ml), and finally 40 /xl of (i mg/ml) glutathione S-transferase. After i0 min at 37°, 5/zl of lchloro-2,4-dinitrobenzene (i0 mM in methanol) is added and the reaction mixture further incubated for 30 min at 37 °. Then the mixture is acidified with 5 /zl of i M H3PO4, centrifuged, and injected onto a /xBondapak C18 column (4 x 250 mm, Waters). The S-(2,4-dinitrophenyl)glutathione conjugate (GS-DNP) is eluted with 0.1 mM H3PO4/methanol (60:40, v/v, flow rate of 1.5 ml/min, Rt (retention time) of 4.65 rain, detection at 340 nm) and the radioactivity in the collected GS-DNP fractions counted as described. It should be noted that the possible 2,4-dinitrophenyl derivatives of cysteine and cysteamine are eluted at a n R t of 3.65 rain. Methological Considerations Investigations of reduced thiols in vitro require some precautions. Autoxidation reactions of reduced thiols during storage and dilution procedures occur spontaneously but can be avoided by simple techniques. During thiol exposure of ceils the autoxidation reaction consumes oxygen and produces reactive oxygen species (e.g., hydrogen peroxide) which may exert cytotoxic effects on the cells. 21 An enhanced toxicity of thiols in CHO cells 20K. G. Eckert and P. Eyer, Biochem. PharmacoL 35, 325 (1986). 21R. D. Issels, J. E. Biaglow,L. Epstein, and L. E. Gerweck, Cancer Res. 44, 3911 (1984).

[1 1]

PROMOTION OF CYST(E)INEUPTAKE

11 1

at higher t e m p e r a t u r e s was d e m o n s t r a t e d to be due to the generation of activated oxygen species. 22 Therefore, depending on the sensitivity of the cell system used, the addition of protection factors (e.g., catalase, heavy metal chelators) may be necessary. 23Besides the oxidation of the thiol group, other oxidation reactions of the thiol molecule m a y also occur. As demonstrated for WR-1065, which p r o m o t e s cystine uptake in C H O and Chinese hamster ovarian carcinoma (OvCa) cells, a increase in cellular glutathione was observed only for O v C a cells. 24In C H O cells the incubation with WR-1065 leads to a significant depletion of intracellular glutathione. Exposure of C H O cells to WR-1065 together with aminoguanidine, an inhibitor of copper-dependent amine oxidases, increased cellular glutathione as observed for other thiolsY Aminoguanidine prevents the enzymatic oxidation of the polyamine moiety of WR-1065 and the production of b r e a k d o w n products which are detoxified by the conjugation with glutathione. 26 Biological I m p l i c a t i o n s Exposure of ceils to several cytostatic agents leads to a depletion of intracellular glutathione. We have d e m o n s t r a t e d that activated ifosfamide (4-hydroperoxyifosfamide), an alkylating anticancer drug, depletes intracellular glutathione of P B L in v i t r o . 27 Also ifosfamide had differential effects on the capacity of lymphocyte subsets to lyse their target ceils which correlate with intracellular glutathione levels. 28 Reconstitution of glutathione levels and lymphocyte functions was achieved by addition of extracellular thiol compounds, (e.g., 2-mercaptoethanesulfonate (Mesna)). 28,29 Exposure of lymphocytes to Mesna in v i t r o increased the uptake of radioactive labelled cystine f r o m the m e d i u m Y Studies investigating the effects of Mesna in v i v o demonstrate p r o m o t i o n of cysteine uptake in PBL. 3° These findings suggest that the mechanisms of cyst(e)ine uptake and elevation of glutathione as demonstrated for C H O cells may also be relevant for other cell systems. z2 R. D. Issels and A. Nagele, this series, Vol. 186, p. 696. 2_'~j. E. Biaglow, R. W. Issels, L. E. Gerweck, M. E. Varnes, B. Jacobson, J. B. Mitchell, and A. Russo, Radiat. Res. 100, 298 (1984). 24R. D. Issels and A. Nagele, Cancer Res. 49, 2082 (1989). 25T. Meier and R. D. Issels, Biochem. PharmacoL, submitted for publication. 26A. Nagele, T. Meier, and R. D. Issels, Prog. YharmacoL Clin. Pharmacol. 8, 309 (1991). 27T. Meier, A. Allenbacher, E. Mueller, G. Multhoff, C. Botzler, M. Wiesnet, and R. D. Issels, Anti-Cancer Drugs 5, 403 (1994). 2s G. Multhoff, T. Meier, C. Botzler, M. Wiesnet, A. Allenbacher, W. Wilmanns, and R. D. lssels, Blood 4, (1995). 29R. D. Issels, T. Meier, E. MUller, G. Multhoff, and W. Wilmanns, MoL Aspects Med. 14, 281 (1993). 30B. Stofer-Vogel, T. Cerny, A. Kiipfer, E. Junker, and B. H. Lauterburg, Br. J. Cancer 68, 590 (1993).

112

GLUTATHIONE

[ I I]

Conclusions Exposure of CHO cells to thiols in vitro leads to enhanced intracellular glutathione levels via an increased uptake of cyst(e)ine. The use of such model systems where effects on intracellular glutathione play a major role may lead to a beter understanding of the fundamental processes which are influenced by the availability of cyst(e)ine and intracellular thiol metabolism.

[ 121

OVOTHIOLS

115

[ 12] Ovothiols By T O D

P. HOLLER

and

PAUL

B. HOPKINS

Discovery a n d D i s t r i b u t i o n Investigation of the molecular mechanism by which hydrogen peroxide is formed during the cyanide-insensitive respiratory burst that immediately follows sea urchin egg fertilization led T u r n e r et al.l to the observation that a cytosolic cofactor was able to confer cyanide-insensitive N A D ( P ) H oxidase activity on the enzyme ovoperoxidase. The isolated cofactor was identified by spectroscopic techniques 2 as the highly methylated 4-mercaptohistidine (3, in Fig. 1), which was eventually named ovothiol C. The positions of the methyl groups were determined by desulfurization of the cofactor with Raney nickel and comparison of the product with authentic samples of N-methylated histidines. These assignments were further demonstrated by nuclear Overhauser enhancement studies and total synthesis. 3 A survey conducted by T u r n e r el al. 4 of the eggs or ovaries of other marine invertebrates revealed two additional methylated mercaptohistidines that were given the names ovothiol A (1 in Fig. 1) and ovothiol B (2 in Fig. 1). A group of researchers in Naples, Italy, independently reported the isolation of a series of mercaptohistidines from marine invertebrates, s which were formulated as isomeric to the ovothiols with respect to the position of ring methylation. Comparison of ovothiol C with a sample of the purportedly isomeric compound isolated in Italy showed them to be identical, 3'4 strongly suggesting that the other mercaptohistidines reported by the Italian group are also members of the ovothiol family. Ovothiols A and B have also been identified in the eggs of trout and salmon, 6 and ovothiol A has been isolated from a marine a l g a l i E. Turner, C. Somers, and B. M. Shapiro, J. Biol. Chem. 260, 13163 (1985). E. Turner, R. Klevit, P. B. Hopkins, and B. M. Shapiro, J. Biol. Chem. 261, 13056 (1986). 3T. P. Holler, A. Spaltenstein, E. Turner, R. E. Klevit, B. M. Shapiro, and P. B. Hopkins, J. Org. Chem. 52, 4420 (1987). 4 E. Turner, R. Klevit, L. J. Hager, and B. M. Shapiro, Biochemistry 26, 4028 (1987). 5 A. Palumbo, G. Misuraca, M. d'Ischia, F. Donaudy, and G. Prota, Comp. Biochem. Physiol. 78B, 81 (1984). E. Turner, L. J. Hager, and B. M. Shapiro, Science 242, 939 (1988). 7 S. Selman-Reimer, R. J. Duhe, B. J. Stockman, and B. R. Selman, J. Biol. Chem. 266, 182 (1991).

METHODS IN ENZYMOLOGY,VOL. 252

Copyright © 1995 by AcademicPress, Inc. All rights of reproduction in any form reserved 7

116

[12]

SIGNAL TRANSDUCTIONAND GENE REGULATION CH3NH2 • HCI ++ CH20(aq) NaCN

a ~

O CH~N,,~H - ["~rr~

O CH3N,,~ H ~,,,NH 2

~ b

i 1 4

S

CH3 N C

5

CH3 N-~./CH2 OH

N~,SCH2C6H4CH 3 6

d

N,~.SCIq2C6H4CH3 7

CH3 N -CH2C1

CHACO2

N f~.SCHeC6HzCH3

N

H

SCHeC6HaCH3

8

9

CH3 . N~--~$CH2C6H4CH3

h

II N

N(CH3)2 SCHzC6H4CH3

10

11 CH3 N ~

CH3 ~q~CO2H ~

~ ~RIR2 N~SCH2C6H4CH 3

RI=H, R2=CH3 10: RI=R2=H 11:RI=R2=CH3

9:

~

CO2H

' ' - ' ~ -I~ SH I~R1R2 N

i 1: RI=R2=H 2: R~=H, R2----CH3 3:RI=R2---CH3

Fro. 1. Chemical synthesis of racemic ovothiols. Reagents: a, (i) H20, (ii) HC(O) OC(O)CH3. b, H2S, (CH3CH2)3N; c, (i) (CH3)3SiCI, (CH3CH2)3N, (ii) CHsC6H4CHzCI; d, CH20, sodium acetate/acetic acid; e, SOC12; f, (i) CHsCONHCH(CO2CH2CH3)2, Nail, (ii) HC1 (aqueous); g, (i) CHaCON(CH3)CH(CO2CH2CH3)z, Nail, (ii) HC1 (aq); h, CH20, NaBH3CN; i, (i) Hg(OCOCF3)2, CF3CO2H, (ii) H2S.

[ 121

OVOTHIOLS

117

Ovothiols from Natural Sources Detection in Tissue

A high-performance liquid chromatography (HPLC) method that allows detection of 0.1-1.0 nmol of ovothiol from tissue has been developed by Turner et al. 4 Tissue samples are homogenized in 4 volumes of water. A 1-ml sample of the homogenate is adjusted to pH 8, mixed with 1/zmol of dithiothreitol (DTT), incubated 10 rain, then admixed with 2.5 tzmol of tritiated [2 x 106 counts/rain (cpm)//xmol], iodoacetic acid for 30 min. Proteins are removed from the sample by precipitation with 10% trichloroacetic acid, followed by centrifugation in a microcentrifuge for 3 rain. The supernatant is loaded onto a 1-ml column of Dowex 50-X8, 100-200 mesh, ammonium form, previously equilibrated with 0.5 M formic acid. The column is washed with 10 ml of 0.5 M formic acid and then eluted with 10 ml of 0.5 M ammonium formate, pH 4.5 (0.5 M ammonium hydroxide adjusted to pH 4.5 with 88% aqueous formic acid). The concentrated eluate is lyophilized twice from water, resuspended in water, and loaded onto a 15-cm Altex ultrasphere ODS C18 column equipped with a standard precolumn. The column is washed for 10 min with a buffer containing 20 mM phosphoric acid and 5 mM sodium heptanesulfonate adjusted to pH 4.5 with aqueous ammonia, and then eluted with a gradient of 0 to 50% methanol in the same buffer. Ovothiols are detected by UV absorbance at 210 nm and scintillation counting. Identification is made by comparison to authentic samples. Isolation

Underivatized ovothiol C has been isolated in 100-mg quantities as described below. Ovothiol A may be obtained in about 10-mg quantities by isolation from cultured marine algae] but milligram quantities of ovothiol B have not been obtained from natural sources. Sea urchins (Strongylocentrotus purpuratus) are gathered intertidally and induced to spawn by intracoelomic injection of 0.50 M KC1. The eggs are collected in 0.45-/zm filtered seawater. The jelly coat is removed from 300 ml of settled eggs by resuspension in 3 liters of artificial seawater (0.5 M NaC1, 10 mM KCI, 2.5 mM NaHCO3, 25 mM EGTA, pH 8.0). The supernatant containing the jelly coat is decanted, and the eggs are resuspended in 600 ml of artificial seawater. Homogenation is accomplished with 10 strokes in a Teflon pestle tissue homogenizer. The homogenate is spun at 15,000 g for 60 rain, and the supernatant is boiled for 10 rain. Precipitated material is removed by centrifugation at 15,000 g for 10 min, and this supernatant is lyophilized to dryness. The residue is redissolved in 80 ml of water and brought to 800 ml with

118

SIGNAL T R A N S D U C T I O N AND GENE REGULATION

[ 12]

absolute ethanol. The solution is evaporated to dryness and extracted a second time, with 80 ml of 90% (v/v) ethanol. Lyophilized material from the second extraction is dissolved in 10 ml of water and applied to a 30-ml column of Dowex 1-X8 (200-400 mesh, hydroxide form) anion-exchange resin. The column is developed with a 300-ml linear gradient of 0 to 0.4 M ammonium formate, pH 7.0 (prepared by neutralization of 0.4 M formic acid with concentrated ammonia). Ovothiol C is identified in the fractions by its UV spectrum (see Fig. 2). This crude material is further purified by HPLC. Samples are adjusted to pH 2.5 with HC1 and loaded onto a Waters (Milford, MA) SP-5PW sulfopropyl cation-exchange column. After washing the column for 10 min with 40 mM formic acid, pH 2.5 (adjusted with concentrated ammonia), the ovothiol C is eluted with a gradient of 40 mM to 2 M formic acid, pH 2.5. Two peaks containing ovothiols are observed. The reduced form elutes at approximately 1.2 M, and the oxidized material, which normally constitutes the majority of the material, elutes at approximately 1.4 M formic acid, pH 2.5. As an alternative to the HPLC separation, acidified material can be applied to a column of Dowex 50-X8, 100-200 mesh, equilibrated with 0.5 M formic acid, and eluted with a gradient of 0.5 M formic acid to 0.5 M ammonium formate, pH 4.5. This material retains some organic impurities, detected by nuclear magnetic resonance (NMR) spectroscopy. 4

¢J

o.2

b

U L 0 0 •~

0.1

250

3~

350

Wovelength (nm) F~c. 2. Ultraviolet spectra of ovothiol C. (a) The UV spectrum of approximately 0.2 mM oxidized ovothiol C. (b) Spectrum taken after addition of a slight molar excess of dithiothreitol. (Reprinted with permission, from Turner et al. 2)

[ 121

OVOTHIOLS

i 19

Chemical Synthesis Chemical syntheses of the ovothiols, in racemic or optically active form, have been described. 8 A synthesis of racemic ovothiols amenable to multigram scale is presented in Fig. 1. The initial phase involves the synthesis of a protected mercaptoimidazole moiety. This general procedure can be used to synthesize a variety of imidazole-4-thiols which might serve as ovothiol analogs for a variety of purposes. 9 In the initial step (Fig. 1) formaldehyde, methylamine, and sodium cyanide are condensed to form an unstable intermediate which is promptly formylated with formic acetic anhydride. The nitrile amide 4 is quite stable and can be purified by distillation at reduced pressure, Addition of gaseous H2S to the nitrile is performed in a mixture of ethanol and triethylamine at room temperature. The crystalline thioamide 5 is cyclized to the imidazole-4-thiol (6) using trimethylsilyl chloride and triethylamine in methylene chloride at - 7 8 °, and the thiol group is protected in situ as the p-methylbenzyl thioether. Elaboration of the amino acid side chain begins with hydroxymethylation using aqueous, acetate-buffered formaldehyde. This particular hydroxymethyl compound (7) can be crystallized from ethyl acetate, avoiding chromatography at this early stage of the synthesis. Addition of 7 in small portions to ice-cold, freshly distilled thionyl chloride, followed by stirring at room temperature and evaporation in vacuo, provides the chloromethyl derivative as the hydrochloride (8). This material is used directly in the next step. Reaction of 8 with 2 equivalents of the sodium salt of diethyl acetamidomalonate in tetrahydrofuran (THF), keeping the reaction temperature below 30 °, followed by hydrolysis for 12 hr in boiling 6 N HC1 and concentration at low pressure affords crude Sprotected ovothiol A (9). This material is diluted with water, brought to pH 7 with aqueous ammonia, and applied to a column of Amberlite XAD4. The column is washed exhaustively with water to remove the glycine, then eluted with batches of 10, 20, 35, and 50% aqueous methanol. Fractions containing the desired product are pooled and evaporated to a yellow solid, which becomes white on trituration with ethanol. Substitution of the sodium salt of diethyl N-methylacetamidomalonate in this procedure afforded Sprotected ovothiol B (10). Reductive methylation of S-protected ovothiol A with aqueous formaldehyde and sodium cyanoborohydride afforded Sprotected ovothiol C (11). The p-methylbenzyl protecting group is removed from each of the Sprotected ovothiols by reaction with mercuric trifluoroacetate in ice-cold T. P. Holler, F. Ruan, A. Spaltenstein, and P. B. Hopkins,J. Org. Chem. 54, 4570 (1989). 9 A. Spaltenstein, T. P. Holler, and P. B. Hopkins, J. Org. Chem. 52, 2977 (1987).

120

SIGNAL TRANSDUCTION AND GENE REGULATION

1121

trifluoroacetic acid containing anisole. The final product (1-3) is most easily desalted and stored as the disulfide, which is obtained by adding a trace of CuC12 to an aqueous solution of the thiol and bubbling oxygen through the mixture. The oxidized material is desalted by loading the aqueous solution onto a column of Dowex 50W-X8, hydrogen form, washing with water, and eluting with 1 M ammonium hydroxide. Ninhydrin-reactive fractions are pooled and lyophilized. Salt-free ovothiol A can then be obtained by reduction of the disulfide in a minimum of water treated at 25 ° with 2 to 3 equivalents of dithiothreitol. For applications compatible with the presence of both reduced and oxidized dithiothreitol, the resulting solution may be used. These contaminants may be removed by chromatography on Dowex 50-X8 ion-exchange resin, ammonium form, containing oxygen-free (N2 purged) aqueous ammonium formate, pH 2.5, then eluted with oxygen-free 1 M ammonium formate, pH 4.0. The eluted ovothiol may be detected by ultraviolet spectroscopy or Ellman's assay. Samples of racemic ovothiol A disulfide synthesized by this method can be obtained, while supplies last, from one of the authors (P.B.H.).

Physical Properties

Spectroscopic Properties The aH NMR spectra of the ovothiols vary somewhat with reduction state and pH of the solution; however, the method remains a useful one for determining purity. The spectra of ovothiol disulfides in approximately neutral DzO include a singlet at approximately ~ 7.8 (imidazole C-2 H), a singlet at approximately 8 3.7 (imidazole N-CH3), and three multiplets at approximately 8 3.3 (C-o~ H), 2.8 (C-/3 H), and 2.3 (C-/3' H). The spectra of ovothiols B and C contain, in addition to the resonances listed above, a singlet at approximately 8 2.8 which arises from the aliphatic N-methyl groups. It has been our experience that XH NMR spectroscopy in aqueous solution is not useful for the determination of the thiol/disulfide ratio of ovothiols, because only a single set of resonances is observed, even at 500 mHz. This is presumably due to fast chemical exchange of the two species. As noted below, ultraviolet spectroscopy can be used for this analysis. The ultraviolet spectra of the ovothiols provide both a useful handle during their detection and isolation and a means of determining their redox state. The UV spectra of reduced and oxidized ovothiols, as seen in Fig. 2, have an isosbestic point at 288 nm, an absorbance ratio of 1.63 at 260 nm, and an absorbance ratio of 2.8 at 236 nm. The oxidized material has

[ ] 2]

OVOTHIOLS

121

significant absorption above 320 nm, imparting a yellow color to concentrated solutions; the reduced form lacks this absorption and is colorless. Ultraviolet spectroscopy provides a particularly simple method for the analysis of the ovothiol content and thiol/disulfide ratio in unknown sampies. The absorbances at 288 and 240 nm of a solution (35 tzg sample/ml) in phosphate buffer, pH 7.0, are measured in a 1-cm cuvette. The weight percent ovothiols in the original sample are Weight percent ovothiol = 100(A240 - A288) Weight percent ovothiol disulfide --- 100(3A28s - A240) Weight percent (ovothiol + disulfide) = 200A288 For example, a sample of ovothiol in phosphate buffer, pH 7, giving an A288 of 0.24 and an Ae40 of 0.55 is 31% ovothiol [100(0.55 - 0.24)]. The same sample is 17% disulfide, or a total of 48% total ovothiol/disulfide. Alternatively, as mentioned below, Ellman's reagent can be used for the analysis of the thiol content of ovothiols. Thiol pKa

Knowledge of the pK, of the thiol group is important in understanding the reactivity of the ovothiols. To avoid the complications of making this determination in the presence of the a-amine and carboxylic acid of the natural substances, Holler and Hopkins ~°titrated 1,5-dimethyl-4-mercaptoimidazole (DMI), a model compound which lacks them. Two observed pKa values at 2.3 and 10.3 were combined with the observation that the Smethyl derivative of DMI has a single pKa of 6.0 to produce the diagram in Fig. 3. At neutral pH, DMI, and by inference the ovothiols, exists almost exclusively as the zwitterion in which the imidazole nitrogen is protonated and the thiol is present as the thiolate. Interestingly, the histidine ring nitrogen is sufficiently basified by the neighboring thiolate to persist in the protonated form to about pH 10, some 4 units higher than the histidine itself. Redox Properties

The midpoint redox potential for the conversion of ovothiol C disulfide to the thiol form has been estimated to be 84 mV positive compared to glutathione.11 The thiol form is rapidly oxidized by hydrogen peroxide. Turner et al. 6 measured a second-order rate constant of 2.0 M -1 s -1 at 30° and pH 7.2 for the reaction of ovothiol C with hydrogen peroxide. Ovothiol disulfides are reduced to the thiol forms by glutathione and, more rapidly, to T. P. Holler and P. B. Hopkins, J. Am. Chem. Soc. U0, 4837 (1988). 11 B. M. Shapiro and E. Turner, Biofactors J. 1, 85 (1988).

122

SIGNAL TRANSDUCTION AND GENE REGULATION

[ 1 2]

CH 3 N~jCH3

(Im - SH)

pKim =

~ K s r

t = 6.7

CH3

CH3

N~CH3

N ~..~CH3 pKzw = 3"7 N+

(~NLs.

SH

(ImH + - SH)

/¢~

X

~ p K

pK 1 = 2 . 3 ~

(Ira - S') 2 = 10.3

OH3

(ImH ÷ - S') FIG. 3. Acid-base equilibria of an ovothiol analog. The pKa values depicted were calculated from the two macroscopic pKa values measured on titration of 1,5-dimethyl-4-mercaptoimidazole (DMI) and the assumption that the pKa at upper left [(ImH+-SH) to (Im-SH)] is identical to that measured on titration of S-methyl-DMI. (Reprinted with permission from Holler and Hopkins. 1° Copyright 1988 American Chemical Society.)

by dithiothreitol. Reduced ovothiols can be determined, as can many other thiols, 12 using Ellman's reagent [5,5'-dithiobis(2-nitrobenzoic acid), DTNB]. The reactivity of the ovothiols in one-electron reduction reactions has been investigated by Holler and Hopkins. 1°'13'14 The first demonstration that the mercaptoimidazole moiety is an effective one-electron donor used ferricytochrome c as a generic one-electron acceptor. The ovothiol analog DMI (see Fig. 3) reacts at physiological pH with a second-order rate constant of 30 M -1 sec -~. The pH-rate profile revealed that the active species is the DMI anion, which reacts with a rate constant of roughly 104 M -I sec-k The ability of ovothiol A to scavenge photochemically generated 12p. C. Jocelyn, this series, Vol. 143, p. 44. 13T. P. Holler and P. B. Hopkins, AnaL Biochem. 180, 326 (1989). 14T. P, Holler and P. B. Hopkins, Biochemistry 29, 1953 (1990).

[131

XAS

STUDIES OF PROTEIN

KINASE

C

123

tyrosyl radicals 12 far surpasses that of glutathione, more closely resembling the reactivity of ascorbic acid and the soluble tocopherol derivative Trolox. This reactivity difference, which is also observed when using Fremy's salt or Banfield's radical, 13 is presumably due to the more favorable free energy of conversion of an aromatic thiol to thiyl radical relative to its aliphatic analog. The presumed thiyl radical intermediate in these reactions seems, in model compounds at least, to be unreactive with oxygen. Biological Significance Ovothiols are present at millimolar concentrations in the eggs of marine invertebrates, persist in the first 2 weeks of embryo development, but are absent from the somatic tissues of adults. This suggests some special role in early development that is no longer necessary in later life. The fact that high concentrations of hydrogen peroxide are generated and used just outside the egg in early fertilization events, and the differences in redox chemistry of the ovothiols compared to glutathione, have led Turner et al. 6 to propose that the ovothiols replace glutathione peroxidase as a hydrogen peroxide scavenger. The ready participation of ovothiols and their analogs, but not glutathione, in one-electron redox chemistry has led Holler and Hopkins 14 to suggest that there may also be a role for the ovothiols in protection of the embryo from free radical-induced damage as well. When the low thiol pKa and relative abundance of the ovothiols are taken into consideration along with the redox properties mentioned above, it becomes clear that the ovothiols have the capability of protecting the egg and the early embryo from damage that might result from a wide range of electrophilic species.

[13] C h a r a c t e r i z a t i o n of the Cysteine-Rich Zinc-Binding Domains of Protein Kinase C by X-Ray Absorption Spectroscopy B y STEVAN R. HUBBARD

Introduction Protein kinase C (PKC) plays a key role in the regulation of cellular signaling events in eukaryotic organisms from yeast to humans, catalyzing the transfer of the terminal phosphate of adenosine triphosphate (ATP) to specific serine or threonine residues of substrate proteins (for review, METHODS IN ENZYMOLOGY. VOL. 252

Copyright © 1995 by Academic Press, lnc, All rights of reproduction in any form reserved,

124

SIGNAL TRANSDUCT1ON AND GENE REGULATION

[ 13]

see Ref. 1). Protein kinase C forms a family of closely related isozymes which are activated at the inner surface of the plasma membrane by diacylglycerol, a product of receptor-mediated phosphatidylinositol hydrolysis. Phosphatidylserine is an essential cofactor for activation, and calcium is required for a subset of isozymes. The single polypeptide chain of PKC consists of an NH2-terminal regulatory domain, in which the divergence among isozymes is primarily found, and a COOH-terminal catalytic domain, which shows sequence similarity to the catalytic domains of other protein kinases. Within the regulatory domain are two adjacent, highly similar sequences of approximately 50 amino acids containing strictly conserved cysteines (six) and histidines (two) in the pattern H-XI2-C-X2-C-Xlo_14-C-X2-C-X4-H-X2-C-X7-C (C6H2) , where the intervening X residues are, in general, more variant. The C6H2 regions of PKC have been shown to be essential for the binding of phorbol esters, which can substitute for diacylglycerol.2 The C6H2 motif is also found in Raf, Vav, chimaerin, diacylglycerol kinase, and the Caenorhabditis elegans unc-13 gene product. Because a cytosolic protein such as PKC is not expected to contain disulfide bonds, the conserved cysteines in the C6H2 regions are likely involved in metal ion coordination. My colleagues and I investigated this proposition using X-ray absorption spectroscopy (XAS), and a report of our findings has been published. 3 Here, I present in greater detail the experimental methods used to show that PKC is a metalloprotein which contains four zinc ions coordinated primarily by cysteine thiolates. Description of X-Ray Absorption Spectroscopy Techniques In XAS, a metal-containing sample is illuminated by a monochromatic beam of X-rays of sufficient energy to stimulate core electron transitions. Absorption is measured by monitoring either the diminution of the X-ray beam as it traverses the sample or the fluorescence emanating from the sample arising from electron relaxation. Two types of XAS techniques were employed in the PKC study. The first, called X-ray fluorescence (XF), was used to identify which if any metal atoms were present in the protein sample and to quantitate the amount. Here, the sample is irradiated by monochromatic X-rays with a fixed energy higher than the absorption edges of the metals of interest, and the energy spectrum of fluorescent photons i S. Stabel and P. J. Parker, Pharmacol. Ther. 51, 71 (1991). 2 y. Ono, T. Fujii, K. Igarashi, T. Kuno, C. Tanaka, U. Kikkawa, and Y. Nishizuka, Proc. Natl. Acad. Sci. U.S.A. 86, 4868 (1989). 3 S. R. Hubbard, W. R. Bishop, P. Kirschmeier, S. J, George, S. P. Cramer, and W. A. Hendrickson, Science 254, 1776 (1991).

[ 13]

XAS STUDIESOF PROTEINKINASEC

125

is recorded. Each atomic element has fluorescence emission lines of characteristic energy which can be used to identify its presence. The second technique, called extended X-ray absorption fine structures (EXAFS), was used to obtain information regarding the chemical environment of the metal atom, namely, the type and number of ligating atoms and metal-ligand distances. In this case, absorption is measured as a function of incident X-ray energy over a range that typically extends from several 100 eV below an absorption edge to 500-1000 eV above. Above the edge, backscattering of the ejected photoelectron by atoms in the near vicinity of the metal atom leads to a modulation of the absorption cross section, which appears as oscillations in the absorption spectrum. 4 The XAS experiments are optimally performed at synchrotron X-ray sources, owing to their high flux and tunability. In a preliminary XF experiment done on beam line X-19A at the National Synchrotron Light Source (NSLS), Brookhaven National Laboratory, we identified stoichiometric amounts of zinc in a PKC sample. At a later date, we collected EXAFS spectra and repeated the XF measurements.

X-Ray Fluorescence Experiment

Sample Preparation The preparation of the PKC sample used for both the XF and EXAFS experiment has been described) Briefly, Sf9 insect cells (Spodoptera frugiperda, fall armyworm ovary) were infected with a recombinant baculovirus containing the cDNA for the rat PKC/3I isozyme. The protein was purified to near homogeneity in three column chromatography steps. To remove any metal ions not tightly bound to the protein, i mM EGTA was included in all column buffers. The sample was fully active as determined by an in vitro kinase assay. The PKC/3I from three preparations was pooled and concentrated in a Centricon-30 (Amicon, Beverly, MA) to approximately 25 mg/ml. The sample used for a control in the XAS experiments was a concentrated pool of fractions that elutes just prior to PKC /3I on the last column.

Data Collection and Processing The detector used for the XF and EXAFS experiments was a liquid nitrogen-cooled array of 13 germanium elements with an energy resolution 4S. P. Cramer and K. O. Hodgson,Prog. Inorg. Chem. 25, 1 (1979).

126

SIGNAL TRANSDUCTION AND GENE REGULATION

[ ] 3]

of approximately 200 eV. 5 For the XF experiment, the signal from one germanium element, which is proportional to the energy of the incident photon, was fed into a Canberra multichannel analyzer (MCA). A pulse height analysis is performed by the MCA, and a histogram of detected photons versus channel number (energy) is displayed and stored. Fluorescence was monitored in the direction normal to the incident X-rays in the horizontal plane. For our experiment, the silicon (220) double-crystal monochromator was set to pass X-rays of 11.2 keV. We were able to detect fluorescent photons with energies from approximately 3.5 to 10.0 keV, which includes Ks emission lines for calcium (atomic number Z = 20) through germanium (Z = 32) and L~ emission lines for tin (Z = 50) through iridium (Z = 77). Zinc standards of 0, 0.5, 1.0, and 2.0 mM were prepared from a zinc atomic absorption standard solution (Sigma, St. Louis, MO); volumes were measured gravimetrically. A Lucite EXAFS cell (outer dimensions of 25.4 by 3.7 by 2.4 mm) with a Mylar window and a sample volume of approximately 70 t~l was used for all samples. The cell was rinsed extensively with an EDTA solution and deionized water between samples. The XF data at room temperature were accumulated for 5 rain for each sample. Raw data in the form of fluorescence counts versus MCA channel number (1024 points per spectrum) were converted to counts versus energy with a two-point linear calibration using the chromium K~ peak (see below) and the incident beam scatter peak (11.2 keV). Each of the energy-calibrated spectra was interpolated onto the same uniformly spaced energy grid and normalized by the integrated counts in the scatter peak to correct for differences in total incident flux. Difference spectra were obtained by subtracting the control spectrum from the PKC/~I spectrum and the 0 mM zinc spectrum from the other zinc standards spectra. The counts in the zinc K~ peaks of the difference spectra were summed. Results

The XF data for the PKC/3I and control samples are shown in Fig. 1. The chromium and copper K~ peaks (5.41 and 8.04 keV) are present in the spectrum of an empty sample cell and derive from the Mylar window; Kapton was later found to be a more suitable window material. The two peaks unique to the PKC ~I spectrum are consistent with zinc K~ and K B emissions (8.63 and 9.57 keV) and are highlighted in the difference spectrum (Fig. 1, inset), The small peak in the difference spectrum at copper K~ 5 S. P. Cramer, O. Tench, M. Yocum, and G. N. George, Nucl. Instrurn. Methods Phys. Res. Sect. A 266, 586 (1988).

[13]

X A S STUDIES OF PROTEIN KINASE C 70

. . . .

i . . . .

i . . . .

I . . . .

L , q l l l l l l

i

I J L~l

.............................. Zn K~

60×

a . . . .

127 Ii

,li

Scatter peak

50E

40O

"~ 30>, -

(

I1)

---~ 100

Cr KcL

-

ZnKcz ~

C

/

/

~ . . . .

3.0

i

,

4.0

,

,

,

I

. . . .

5.0

i

. . . .

6.0

i

. . . .

7,0

i

. . . .

8.0

t

. . . .

9.0

i

. . . .

i

,

,

10.0 11.0 '12.0

Energy (keY) FIG. 1. Energy-calibrated XF spectra of the PKC/3[ (solid fine) and control (dotted line) samples prior to normalization by incident X-ray flux. Inset: Normalized difference spectrum. (Reprinted with permission from Hubbard et al. 3 Copyright © 1991 AAAS.)

is probably not significant owing to the sizable copper Ks peak in the control spectrum. The concentration of zinc in the PKC/3I sample was determined by a comparison of the summed counts in the zinc Ks peaks of the protein and standards spectra. Based on a linear regression analysis of the zinc standards (fit correlation of 0.991), the zinc concentration in the PKC/3I sample was 1.22 + 0.03 raM. The standard deviation was estimated by examining the effects of reasonable changes in energy calibration, spectra normalization, and peak integration on the concentration calculation. The concentration of PKC BI in the sample was obtained by measuring the total protein concentration by amino acid analysis and the PKC 13I purity by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE) and densitometry. Duplicate measurements using a Beckman 6300A amino acid analyzer were made for each of two separate aliquots taken from the XAS sample. The value obtained for the total protein concentration was 29.2 + 1.2 mg/ml. An SDS-polyacrylamide gel (12.5%) stained with Coomassie blue was scanned in two dimensions on a Molecular Dynamics 300A computing densitometer. Several lanes with differing amounts of loaded protein were analyzed; a lane containing just sample buffer served for baseline measurements. According to the densitometry, PKC/3I comprised 76 + 4% of the total stained protein (the most significant contaminant comprised 4%). As a positive control, a homogeneous preparation of ribonuclease H was also run on the gel, for which a purity value of 98% was obtained. Because Coomassie blue, like other protein dyes, shows

128

SIGNAL TRANSDUCTION AND GENE REGULATION

[ 13]

some amino acid composition dependence in protein staining, the systematic error in the purity measurement decreases with increasing protein purity. The combined measurements gave a PKC/3I concentration of 22.2 + 1.5 mg/ml or 0.289 mM (Me 76,781). The zinc ion to PKC/3I stoichiometry is therefore 4.2 _+ 0.3. The error estimate was obtained by applying the standard formula for propagating errors assumed to be independent. Thus, the XF data indicate that four zinc ions are tightly bound to PKC/3I.

Extended X-Ray Absorption Fine Structure Experiment

Data Collection and Processing To prevent X-ray-induced protein damage, the PKC/3I sample was cooled to approximately 15 K using a liquid helium cryostat. Twenty-one scans through the zinc K-absorption edge (9.66 keV), from 9.25 to 10.50 keV, were recorded with step sizes of 1.0 eV in the edge region and 2.0 or 3.0 eV in the EXAFS region. The duration of each scan was 35 min. Two scans for baseline subtraction were taken of a cell containing sample buffer. A zinc foil placed downstream of the sample position was used to calibrate the energy for each scan. Transmission EXAFS data were taken at room temperature for a zinc model compound, (1,10-phenanthroline)bis(4-toluenethiolato)Zn2+, in which zinc is coordinated by two sulfur and two nitrogen atoms. 6 Data from the seven active germanium elements of the detector array were weighted by the size of the edge jump and averaged together. No time-dependent changes in the scans, indicative of sample deterioration, were observed. A small edge-like feature in the far EXAFS region of the PKC/3I spectrum, which was attributed to a tungsten contaminant in the cryostat, limited the usable data to below 10.2 keV. In addition, the baseline spectrum showed a zinc K-edge feature with an edge jump 14% of that in the PKC/3I spectrum. This was also due to a contaminant in the cryostat, which was not present in the preliminary setup of the experiment. The baseline spectrum was smoothed, with the edge feature retained, and subtracted from the PKC/3I spectrum. Tests showed that the EXAFS results were not overly sensitive to this subtraction. A cubic spline was fit to the region above the edge to extract the EXAFS oscillations. Conversion from energy E (keV) to photoelectron wave number k (/~-~) was done with an estimate for the threshold energy, E0, of 9.670 keV {k = [0.262(E E0)]}v2. EXAFS spectra were smoothed with a Gaussian function of width 0.1/~-i. 6 T. L. Cremers, D. R. Bloomquist, R. D. Willet, and G. A. Crosby, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. 36, 3097 (1980).

[13]

X A S STUDIES OF PROTEIN KINASE C

129

Curve fitting analyses were done with the nonlinear least-squares program OPT, 7 using curved wave-derived theoretical backscattering amplitudes and phases computed with the program FEFF. 8 The single-scattering expression used to obtain the calculated EXAFS spectrum x¢(k) is given by

e_2 2abk2e_2Rab/X(k) sin[2kRab + aab(k)] xc(k) = S ~ NbAab(k,Rab) ~ where Nb is the number of b-type backscattering atoms at a distance Rab from the absorber atom a, k is the photoelectron wave number, Aab(k, Rab) is the total backscattering amplitude including intrinsic losses, aab(k) is the total phase shift function, Crab is the root mean square deviation of Rab (Debye-Waller-like factor), h(k) is the electron mean free path, and S is a k-independent scale factor relating experimental and calculated spectra, s

Results The zinc K-edge absorption spectrum for the PKC flI sample is shown in Fig. 2a, and the extracted EXAFS oscillations are shown in Fig. 2b. The observed absorption spectrum is a superposition of the absorption features for each of the four zinc sites in the protein. The initial curve fitting analyses clearly indicated that the EXAFS oscillations were primarily due to backscattering from sulfur atoms with a zinc-sulfur distance of approximately 2.3/~ (Fig. 2c). It was more difficult to establish: (1) the average number of coordinating sulfur atoms, (2) the presence of a lower Z ligand (e.g., nitrogen), and (3) the strength of evidence for thiolate-bridged zinc sites. Fitting results for several plausible coordination models are given in Table I. The 3.5S/0.5N model represents a coordination scheme in which two zinc ions are bridged by a single thiolate, one zinc ion having a total of four sulfur ligands and the other having one nitrogen and three sulfur ligands. The 2S/2S model was included to test a four-sulfur model with the same number of independent parameters as for the 3S/1N model (a 3S/1S model was also tested). The model consisting of three sulfur atoms and one nitrogen atom (3S/1N) gave the best fit, and the derived zinc-sulfur and zincnitrogen distances, 2.32 and 2.08 ~ , and mean-square distance deviations are chemically reasonable. Distinguishing between a nitrogen and oxygen backscatterer is problematic, especially when its contribution to the EXAFS is minor. A 3S/10 model would have given an equally good fit, but a nitrogen was favored because of the conserved C6H2 histidines. 7 Z. K. Eccles, Thesis, Stanford University, Stanford, California (1977); modifications to the program were made by G. N. George and S. P. Cramer. 8 j. j. Rehr, J. Mustre de Leon, S. I. Zabinsky, and R. C. Albers, J. Am. Chem. Soc. 113, 5135 (1991).

1 3 0

TRANSDUCTION

S I G N A L

(a)

A N D

G E N E

(:l) ~ "

0.25 ~. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

(J)

i

.~..,

[13]

R E G U L A T I O N

,

.

~: x~ k5 ¢~ ~'-~"

9.6

9.7

9.8 9.9 10.0 Energy (keY)

10.1

(b)

10.2

(c) 8.0

~

4.0

2.o

X 0.0 ~ 3) or, for less than three determinations, the standard deviation of the measurements averaged (n = 4). Highest PDBu binding values determined for individual fusion proteins are shown in parentheses. Data for fusion proteins with no adjacent values in parentheses are single determinations. g ND, Not determined.

162

SIGNAL TRANSDUCTION AND GENE REGULATION

[ 161

50

Cys1 )

n +Glycerol • -Glycerol

40 30 20

10 ND

0

_~ so u

B

ND

(GST-Cys2)

40 3o ,c

° o

-7

2o

-

10 0

~E

0 so ] C

(GST-CyslCys2)

40 30, 20,

10. 20

33 Temperature

37

(*C)

FIG. 2. Optimization of expression conditions for three GST-PKC3, fusion proteins in XL1B E. coli. The GST fusion proteins GST-Cysl (A), GST-Cys2 (B), and GST-CyslCys2 (C) were expressed in XLtB cells following the procedure that eventually became established (see text), except that in some cases (white bars) glycerol was present at a final concentration of 10% (v/v) in the expression medium and the temperature at which expression was induced by addition of IPTG was varied as indicated. Recovery of soluble fusion protein after cell lysis and centrifugation was quantitated by measuring GST activity. Note that in all three cases reduction of the expression temperature to 20° greatly improved the recovery of soluble fusion proteins. In contrast, the presence of glycerol had no beneficial effects in this respect.

p r o b a b l y n o t n e c e s s a r y b e c a u s e r e g u l a r m e d i u m c o n t a i n s a b o u t 2 0 / z M zinc (22 _+ 5 / z M ) , as was l a t e r d e t e r m i n e d in c o n j u n c t i o n with e x p e r i m e n t s using BL21 cells (see Fig. 4). Thus, t r a n s f o r m e d E. coli w e r e initially g r o w n at 37 ° b u t s u b s e q u e n t l y t r a n s f e r r e d to 20 ° p r i o r to t h e i n d u c t i o n o f p r o t e i n synthesis b y the a d d i t i o n o f I P T G to a final c o n c e n t r a t i o n o f 0.2 m M in t h e p r e s e n c e o f l / z M ZnCI2

[16]

PKC-GLUTATHIONE S-TRANSFERASE FUSION PROTEINS

163

once cultures reached an absorbance of 0.7 (see Experimental Procedures). Under such conditions, fusion proteins were present as soluble proteins in E. coli homogenates and could be readily purified in a single step to homogeneity by affinity chromatography. Following this procedure roughly 1 mg of pure GST-PKC'y fusion proteins containing short segments of the regulatory domain (Cysl, Cys2, CyslCys2) was obtained from each 25-30 ml of induced E. coli culture. Conditions established with fusion proteins containing the cysteine-rich regions essentially allowed expression of fusion proteins containing all segments of the PKC~/regulatory domain in XLIB cells (Fig. 3A). However, in some cases, in particular when the V1 or CaLB region was present, affinity-purified proteins migrated with apparent A

1

2

3

4

5

6

7

8

9

10

11

Mr (x10-3) -

106

"

80

-

49.5

-

32.5

-

27.5/GST

-

18.5

- D y e front

B -

106

"

80

-

49.5

-

32.5

-

27.5/GST

-

18.5

- D y e front

FIG. 3. Analysis by SDS-PAGE of fusion proteins expressed in either XLIB or BL21 cells. PKCy (lane 1) or GST-PKC3~ fusion proteins (1/zg) were resolved by SDS-PAGE on 12.5% gels and subsequently stained with Coomassie blue. Fusion proteins isolated from both XL1B (A) and BL21 (B) cells are shown: GST (lane 2), GST-V1 (lane 3), GST-V1Cysl (lane 4), GST-V1CyslCys2 (lane 5), GST-V1CyslCys2CaLB (lane 6), GST-Cysl (lane 7), GST-Cys2 (lane 8), GST-CyslCys2 (lane 9), GST-Cys2CaLB (lane 10), and GST-CaLB (lane 11). Expression in XL1B cells yielded mainly intact fusion proteins in several cases (lanes 7, 8, 9, and I0), but others were severely degraded (lanes 3, 4, 5, 6, and 11). In the latter cases expression in BL21 improved the yield of nondegraded fusion protein. The Mr values of prestained marker proteins (Bio-Rad, Richmond, CA) are indicated at right together with the dye front.

164

SIGNAL T R A N S D U C T I O N A N D GENE R E G U L A T I O N

[ 16]

molecular weights lower than anticipated, suggesting that both the V1 and CaLB sequences were susceptible to proteolysis. To circumvent these problems, an alternative host strain, with reduced protease activity, was sought for protein expression. BL21 cells, an E. coli strain deficient in ornp T and Ion proteases, were transformed with plasmids isolated from the XL1B cells. Protein expression was induced and cells were harvested as described above. A major difference from previous experiments with XL1B was that maximum levels of protein expression were obtained after only 5-8 hr, as will be discussed (Fig. 4). Figure 3B shows an analysis of the same fusion proteins as discussed above purified from BL21 cells. Clearly, alteration of the host cell dramatically improved the recovery of intact fusion proteins, especially in cases where either V1 or CaLB were adjacent to GST (Fig. 3, lanes 3, 5, 6, and 11). However, the yield of larger fusion proteins like GST-V1CyslCys2CaLB was about 4-fold lower (1 mg per 100 ml E. coli culture; see Fig. 4B) than that for GST-Cys2. A particularly difficult fusion protein to obtain intact was GST-V1CyslCys2CaLB. Thus, the time dependence of expression in BL21 cells was optimized using this fusion protein as an example. Expression was monitored as a function of several parameters, namely, by measuring GST activity in E. coli lysates (Fig. 4A), fusion protein and PDBu binding recovery after affinity purification (Fig. 4B), the specific PDBu binding activity (Fig. 4C) and zinc stoichiometry (Fig. 4E). Cell growth (doubling times at 37°: BL21, 24 min; XL1B, 54 min) and protein expression were more rapid in the BL21 cells than in XL1B E. coli. After 5-8 hr both GST activity (Fig. 4A) and specific PDBu binding activity (Fig. 4C) reached constant levels, whereas this occurred to a lesser extent for protein recovery from the affinity column (Fig. 4B) and total PDBu binding in lysates (Fig. 4A). This could have indicated that at longer time intervals more degradation was occurring. However, gel analysis of GST-V1CyslCys2CaLB purified at different time points suggested that there was no difference in this respect. Surprisingly, an inverse relationship was noted with the zinc stoichiometry of purified protein, suggesting that, as time progressed, either zinc became limiting or it was not effectively incorporated into the protein. Our previous studies indicate that although 4 zinc atoms are coordinated within the regulatory domain of P K C , 17 n o t all appear equally important for PDBu binding. 18,19 This may explain why an increase in specific PDBu binding activity was observed (Fig. 4C), despite the decrease in zinc stoichiometry (Fig. 4E). In any case, these studies revealed that zinc stoichiometries for fusion proteins isolated from BL21 cells may vary considerably depending on the time point postinduction at which cells are harvested. Fusion proteins were routinely expressed in BL21 cells for 5-8 hr postinduction, and representative samples of such preparations are shown in Fig.

[16]

PKC-GLUTATHIONE S-TRANSFERASE FUSION PROTEINS

lOO A

165 1000

lOOO "_ @ Protein B o POSu binding

80

80o -~

800

E=

40

~ 600 .E ~ 400

6OO

20

ZOO

200

60 p-

0

~ 0

5 Time

,

'

4o "C

o.

0

10 1'5 2'0 post-induction (hour)

5'

0 Time

2.5

..-..-e

A

==

~5 4 0 0 i~

1 'o 15' 0 post-induction (hours)

"E

~2,0 30

i

25 20

'~1 .S

,~ ~t.o

15

¢

N

10

~o.s

5 0

I

I

I

I

5 10 15 20 Time post-induction (hours)

|

0

Time

I

I

|

5 10 15 20 post-induction (hours)

Flo. 4. Expression of GST-V1CyslCys2CaLB in BL21 cells. Fusion protein expression was induced at 20° as described in the text and followed by measuring several parameters as a function of time after induction with IPTG. The GST activity after centrifugation of BL21 cell lysates is shown in (A), protein (@) and PDBu binding activity (O) recovered by affinity purification from lysates of 100 ml BL21 cell culture in (B), specific PDBu binding activity of affinity-purified fusion protein in (C), and zinc stoichiometry of affinity-purified fusion protein in (E). Fusion protein (1/xg), purified at time points of 2, 5, 8, or 20 hr postinduction, is shown after S D S - P A G E and Coomassie blue staining in (D) (lanes 2, 3, 4, and 5, respectively). Prestained molecular weight marker proteins are shown in lane 1, corresponding, from top to bottom, to Mr values of 106,000, 80,000, 49,500, 32,500, 27,500 and 18,500. Note that although the yield of affinity-purified fusion protein and total PDBu binding activity as well as the specific PDBu binding activity increase with time postinduction, a decrease in zinc stoichiometry was seen.

166

SIGNAL TRANSDUCTION AND GENE REGULATION

[161

3. Preparations of fusion proteins obtained both from XLIB and BL21 cells were additionally characterized with regard to PDBu binding and zinc stoichiometry. These results are summarized in Table I. Most notably, zinc levels were below the anticipated values for all fusion proteins except those containing Cysl, Cys2, and CyslCys2 (XLB1 strain), but even there considerable variation was observed. The PDBu binding values were also highest in these proteins. Thus, although proteins produced in strain XLB1 were less degraded (Fig. 3), both zinc stoichiometries and PDBu binding values were lower in all cases except for PDBu binding to GST-Cys2 (Table I), possibly owing to the more rapid production of GST fusion proteins in BL21 cells, in comparison to XL1B cells. Zinc stoichiometries were highest for the GST-V1CyslCys2CaLB fusion protein at early time points during expression (1-2 hr postinduction) and decreased subsequently (Fig. 4E), being paralleled by a decline in the rate of protein production as fusion protein accumulated within the BL21 cells (Fig. 4A,B). This may indicate that intracellular levels of free (i.e., not protein-bound) zinc, rather than the rate of protein expression, became more critical as time progressed. In general, we found that values for zinc stoichiometries were much lower than expected mainly in fusion proteins with predicted stoichiometries greater than 2. The expression of 1 mg of fusion protein of Mr 35,000-40,000 consumed roughly 25 nmol of zinc per coordinated zinc atom. At a total concentration of 20/zM zinc, roughly 500 nmol of zinc was present in 25 ml expression medium, the volume from which the equivalent of 1 mg of fusion protein was obtained. Even the coordination of 4 atoms of zinc per fusion protein molecule should not have depleted the medium of zinc. Thus, zinc incorporation into fusion proteins appeared to be limited by the rate at which E. coli took zinc up from the medium. In BL21 cells, which grow faster and express fusion proteins more rapidly, such a limitation would be expected to affect zinc stoichiometries more seriously, as was indeed the case (Table I). Concluding Remarks We have described here how conditions that enable the expression and isolation of large quantities of segments of the PKC'y regulatory domain as soluble GST fusion proteins were initially established using XL1B as an E. coli host. Affinity-purified fusion proteins, containing the cysteine-rich region of PKC referred to as Cys2 and deletions thereof, have provided excellent molecular tools to define the PKC regulatory domain elements involved in zinc coordination and PDBu (or diacylglycerol) binding, as well as the relationship between these two properties. 18,19 Furthermore, using PDBu binding as a sensitive method of quantitation, GST-PKC fusion

[161

PKC-GLUTATHIONE S-TRANSFERASE FUSION PROTEINS

167

proteins have enabled us to analyze contributions of several cofactor binding sites in the regulatory domain toward phosphatidylserine-, diacylglycerol-, and calcium-dependent regulation of PKC3, activity.2° For the latter studies, fusion proteins containing almost the entire regulatory domain of PKC~/were employed, which were difficult to obtain in large quantities owing to protein degradation (Fig. 3A). These problems were overcome by using protease deficient BL21 E. coli as host cells for protein expression (Fig. 3B). Unfortunately, however, some proteins expressed in these cells had lower PDBu binding activities than when expressed in XL1B (GST-Cysl, GST-CyslCys2), and zinc stoichiometries were generally lower, too (see Table I). This may have been a result of the more rapid expression kinetics in BL21 cells, as the zinc binding stoichiometry was observed to decline as a function of time after induction of protein expression with IPTG (Fig. 4). Thus, although more rapid expression in conjunction with a reduction of proteolytic activity served well to increase the yield of intact fusion proteins in BL21 cells, the proteins obtained were not necessarily improved in terms of functional (PDBu binding) and structural (zinc coordination) criteria. Our results suggest that expression conditions need to be optimized individually for different proteins and that protein yield and SDS-PAGE analysis are not sufficient to assess the quality of the protein product. Depending on the final objectives, additional criteria are essential, and particular attention should be paid to the rate at which proteins are being expressed. Despite these limitations the procedures described here should yield the quantities of fusion proteins necessary for future nuclear magnetic resonance (NMR) and crystallographic studies of PKC structure. Furthermore, these procedures have already been employed successfully in our laboratory to express raf-1 kinase, another protein kinase with a cysteinerich domain that coordinates zinc, as well as different segments thereof, as soluble GST fusion proteins. 33 Thus, these methods should also provide helpful guidance to the successful expression of other metal ion-coordinating thiol proteins. Acknowledgments Research was supported by National Institutes of Health Grant GM 38737.

33 S. Ghosh, W. Q. Xie, A. F. G. Quest, G. M. Mabrouk, J. C. Strum, and R. M. Bell, Jr. Biol. Chem. 269, 10000 (1994).

168

SIGNAL TRANSDUCTION AND GENE REGULATION

[l 7]

[ 17] R e d o x R e g u l a t i o n o f A c t i v a t i o n o f NF-KB Transcription Factor Complex: Effects of N-Acetylcysteine

By FRANK J. T.

S T A A L , MICHAEL T. ANDERSON, and LEONARD A. HERZENBERG

Introduction NF-KB is a heterodimeric transcription factor complex composed, in its classic form, of two DNA-binding subunits: p50 and p65 (reviewed in Ref. 1). NF-KB was first identified as a DNA-binding activity specific for the KB motif in the immunoglobin kappa (K) enhancer. Cloning of the NF-KB p50 and p65 genes revealed a family of NF-KB/rel proteins that participate in a variety of transcriptionally regulated processes, such as lymphocyte differentiation, responsiveness to cytokines, and embryonic development in Drosophila. NF-KB is present in many cell types in an inactive form in the cytoplasm, bound to a cytoplasmic retention molecule called I-KB (for inhibitor of NF-KB). 2 Nuclear translocation of NF-KB is achieved after stimulation of the cells by many different inducers (cytokines, phorbol esters, viral proteins, oxidants) and presumably involves phosphorylation of I-KB which disrupts its interaction with the p65 NF-KB subunit. 3 The NF-KB complex is then free to migrate to the nucleus and transactivate its various target genes. Studies from several laboratories have revealed that the signal transduction pathways leading to activation of NF-KB are redox regulated. 4-7 That is, antioxidants such as N-acetyl-L-cysteine (NAC) can inhibit activation, whereas oxidants such as hydrogen peroxide (1-[202) can directly activate NF-KB. The N A C inhibition of NF-KB activation first implicated a redoxsensitive step in the NF-KB signal transduction pathway. 4 Further studies have found that a wide variety of compounds which share the ability to modulate intracellular oxidant levels also inhibit NF-KB activation. These 1 G. P. Nolan and D. Baltimore, Curr. Opin. Gen. Dev. 2, 211 (1992). 2 R. Sen and D. Baltimore, Cell (Cambridge, Mass.) 47, 921 (1986). 3 S. Ghosh and D. Baltimore, Nature (London) 344, 678 (1990). 4 F. J. T. Staal, M. Roederer, L. A. Herzenberg, and L. A. Herzenberg, Proc. Natl. Acad. Sci. U.S.A. 87, 9943 (1990). 5 T. Kalebic, A. Kinter, G. Poli, M. E. Anderson, A. Meister, and A. S. Famci, et al., Proc. Natl. Acad. Sci. U.S.A. 88, 986 (1991). 6 S. Mihm, J. Ennen, U. Passara, R. Kurth, and W. Droge, A I D S 5, 497 (1991). 7 R. Schreck, P. Bieber, and P. Baeuerle, E M B O J. 10~ 2247 (1991).

METHODS IN ENZYMOLOGY,VOL 252

Copyright © 1995 by AcademicPress, Inc. All rights of reproduction in any form reserved.

[171

REOOX REGULATIONOF NF-KB

169

results, coupled with the observations that oxidants such as H202 are by themselves sufficient to activate NF-KB, 7 suggests oxidants play key roles within NF-KB signaling pathways. Here we describe detailed methods on the appropriate cell culture conditions to study the influence of oxidants and antioxidants, the preparation of nuclear protein extracts, and the use of gel retardation assays to detect the presence of activated NF-KB in the nucleus. In addition, we briefly describe a reporter gene-based system to assay whether the DNAbound NF-KB is transcriptionally active. Methods and Procedures Principle

The electrophoretic mobility shift assay (EMSA) or gel retardation assay is a powerful method for quantitative and qualitative analysis of p r o t e i n - D N A interactions. The technique is based on the principle that the electrophoretic mobility of a nucleic acid (e.g., a radiolabeled oligonucleotide) through a polyacrylamide gel is impeded when a protein is bound to it. Thus, free (labeled) oligonucleotide and oligonucleotides that have bound protein can be discerned based on their relative migration in the gel. This principle was originally described by Garner and Rezvin s and by Fried and Crothers 9 and extended to the identification of mammalian transcription factors by Singh et a l l ° With the EMSA small amounts of a D N A (or RNA)-binding protein can be detected in a heterogeneous mixture of proteins such as a nuclear extract. Cells and Stimulation Conditions

A variety of cells can be used for measurement of NF-KB. Either constitutive or inducible EMSA-shifted NF-KB complexes are detectable in lymphoid, epithelial, and fibroblast cell lines as well as in primary cells. It is important that the cells are viable, proliferating in logarithmic phase, and grown at the right density. For instance, Jurkat T cells grown at high density (2 x 10 6 cells/ml) can have about 25% lower glutathione (GSH) levels than cells grown at low density. This density-induced change in GSH can confound redox effects of the compounds actually studied. It is therefore good experimental practice to perform the experiments under standardized culture conditions. M. M. Garner and A. Revzin, Nucleic Acids Res. 9, 3047 (1981). 9 M. G. Fried and D, M. Crothers, Nucleic Acids Res. 9, 6505 (1981). ~0H. Singh, R. Sen, D. Baltimore, and P. A. Sharp, Nature (London) 319, 9 (1986).

170

SIGNAL TRANSDUCTION AND GENE REGULATION

[ 17]

Cells are incubated with appropriate stimuli (e.g., cytokines, phorbol esters) that activate NF-KB and the redox-regulating compounds (antioxidants, iron chelators, or oxidants) that may inhibit or increase activation. Both the stimuli and inhibitors should be added from stock solutions that are at least 100x concentrated, pH-neutralized, and sterilized before use. If organic solvents [such as dimethyl sulfoxide (DMSO)] are used, it is important to include a control adding the vehicle only. As a first estimate, it is useful to explore the kinetics of activation and inhibition of NF-KB over a time range from 0.5 to 8 hr; subsequently, one can establish a rough titration curve of the redox-active compound.

Preparation of Nuclear Protein Extract Several methods to prepare nuclear extracts are available, most of which involve salt extraction of whole cell or nuclear homogenates. The original procedure was described by Dignam, Lebovitz, and Roeder and has been modified by others. 11 The nuclear extract method described here works well with 1 x l06 to 2 x 107 cells and is an adaptation from the method described by Schaffner and co-workers. 12 We have compared this method to the original large-scale method and obtained equivalent results for the NF-xB, Oct-l, AP-1, and NF-AT transcription factors (although total protein yield is, of course, lower). 1. After stimulation, the cells are harvested, washed once, and then resuspended in 1.0 ml of cold Tris-buffered saline (TBS). To harvest adherent cells, do not use trypsin but use phosphate-buffered saline (PBS)E D T A (2.5 mM) instead. Subsequent steps are done in the cold room at 4 °. 2. Transfer sample to a 1.5-ml Eppendorf tube. 3. Centrifuge 15 sec at 14,000 rpm. 4. Remove supernatant using a drawn out Pasteur pipette and suction. 5. Resuspend (cellular) pellet in 400/zl of buffer A [10 m M HEPES, pH 7.8/10 mM KC1/2 mM MgC12/0.1 mM EDTA/0.5 mM dithiothreitol (DTT), supplemented with the protease inhibitors phenylmethylsulfonyl fluoride (PMSF, 0.5 mM), antipain (1/zg/ml), leupeptin (0.3/xg/ml), and pepstatin (0.5/zg/ml); DTT and the protease inhibitors are added freshly from stock solutions.]. 6. Incubate cells on ice for 15 min. 7. Add 25 tzl of 10% Nonidet P-40 (NP-40) solution. This solution should be added to the inner part of the Eppendorf cap. Cap the tubes in 11H. Ohlsson and T. Edlund, Cell (Cambridge, Mass.) 45, 35 (1986). 12 E. Schreiber, P. Matthias, M. M. Mueller, and W. Schaffner, Nucleic Acids Res, 17, 6419 (1989).

[17]

REDOX REGULATIONOF NF-rB

171

pairs, invert the tubes, and immediately vortex 10 times for 1 sec. This procedure allows reproducible extraction of nuclear proteins, whereas direct addition of NP-40 can give problems. 8. Centrifuge extract 30 sec at 14,000 rpm. Remove as much as possible of the supernatant (if necessary, centrifuge a second time to remove all liquid). A small white nuclear pellet should be visible. 9. Resuspend pellet in 40/~l of buffer C (50 m M HEPES, pH 7.8/50 mM KC1/300 mM NaC1/0.1 mM EDTA/0.5 mM DTT/10% (v/v) glycerol; add protease inhibitors as for buffer A). Resuspension is usually somewhat difficult. It is important to assure that clumps are at least somewhat dispersed. 10. Salt-extract nuclear proteins by rotating on a rotator for 30 min. 11. Centrifuge 10 min at full speed (14,000 rpm). 12. Transfer supernatant (nuclear extract) to a new Eppendorf tube. Discard pellet (cell debris). The extracts can be stored at - 7 0 ° or used immediately for determination of protein concentration and gel retardation.

Electrophoretic Mobility Shift Assay The amount of protein obtained can be determined using the Bradford assay (Bio-Rad, Richmond, CA). We routinely use 200/.d of reagent with 2.0 ~l of extract and a standard curve of different known concentrations of bovine serum albumin (BSA). We then determine the 560 nm absorbance on a 96-well enzyme-linked immunosorbent assay (ELISA) plate reader, from which the protein concentration in the extract can be calculated. Two micrograms of nuclear extract per condition should be sufficient for detection of NF-KB and other transcription factors by EMSA. Optimal binding of NF-KB is dependent on the final salt concentration in the binding reaction, the concentration of DTY, and the amount of nonspecific competitor DNA. In our hands, optimal binding occurs under conditions of 70 mM salt, 5 mM D ] T , and 2.0/~g poly(dI-dC) as competitor. Poly(dI-dC) is dissolved in Tris-EDTA buffer (TE)/100 mM NaCl, heated to 90°, and slowly cooled (over 30-45 min) to room temperature before use. It may be necessary to vary both the amount of extract and amount of nonspecific competitor to optimize the sensitivity of the assay for detection of NF-KB within different cell types. The radiolabeled probe (see below) is added last; 10,000 counts/ min (cpm) should be sufficient for detection after overnight exposure on film. Binding reactions are done at room temperature for 45 min in a final volume of 20-50 tzl. Subsequently, samples are electrophoresed on a 4.5% polyacrylamide gel with 0.25× TBE (Tris-borate-EDTA buffer, pH 8.3) as running buffer

172

SIGNAL TRANSDUCTION

AND

OENE

[17]

REGULATION

for 1.5-2 hr at 100 V. After electrophoresis, the gel is disassembled, transferred onto two sheets of Whatman (Clifton, N J) filter paper, and dried. A typical 1.5-mm gel requires about 1 hr of drying under vacuum at 80°. The dried gel is exposed to film overnight at - 7 0 ° and developed. For most purposes this exposure time should yield a good signal. If necessary, longer exposures should be used.

Labeling of Probe Many different KB probes can be used. We use a double-stranded probe made by annealing two single-stranded synthetic oligonucleotides with the following sequences. T C GAGT C AGAGGGGAC T T T C C GAG C AGT C T C C C CT G A A A G G C T C A G C The unlabeled probe is stored at 100 ng/tA at - 2 0 °, and 0.5 IA is used to end-label the probe with Klenow D N A polymerase. Mix 0.5/zl probe, 10 /zl GTC 2 mM dNTP mixture (2 mM dATP, 2 mM dGTP, 2 mM dq-TP),

Slirnulafion 20 n t M ~ NAC

~o

.,,~, --

I

I

~

+

-I

.%~

+

_ I~

+

_ I

~,o~

+ I

J

NFr.B

Free Probe

F~G. 1. Inhibition of TNF-a- and PMA-induced NF-KB binding activity by NAC. 293 cells were stimulated with TNF-a, PMA, or both in the presence or absence of 20 mM NAC. Nuclear protein extracts were made and EMSAs performed as described in the text. First and last lanes contain free probe without nuclear protein added. The top arrow indicates the inducible NF-KB band; the bottom arrow marks unbound probe. (Reproduced from Staal e t al. 4 with permission from the publishers.)

[17]

REDOX REGULATIONOF NF-KB

173

30/zl doubly distilled water, 5/zl of 10× Klenow buffer (0.5 mM Tris-C1, pH 7.5, 0.1 M MgC12, 10 mM DTT, 0.5 mg/ml BSA), 1/xl Klenow DNA polymerase (100 U/ml), and 4/zl of [a-32p]dCTP (1 mCi/ml). Incubation is for 30 min at room temperature, after which the reaction is stopped by addition of 50/zl TE. Labeled probe and unincorporated [a-32p]dCTP are separated using a spin column packed with Bio-Rad P6DG or Sephadex G-25. The columns are made using a 1-ml insulin syringe (centrifuge 2 rain at 1600 rpm (500 g), packed volume should equal about 0.9 ml). Two microliters of the labeled probe is counted using a scintillation counter. A good probe should have at least 25,000-50,000 counts/sec (cps).

Analysis of Results Typical results of a gel retardation assay are shown in Fig. 1. The presence of NF-KB under stimulated [tumor necrosis factor (TNF), phorbol 12-myristate 13-acetate (PMA)] conditions is visible as a set of two sharp bands, which presumably correspond to the (p50)2 homodimer (lower band) and the p65-p50 heterodimer (upper band). Clearly, if stimulation is done in the presence of 20 mM NAC, no activated NF-~B can be found. Similar

20 mM ....~ NAC

I

I

--

I

+

I

-

I

-

I

+

I

+

I

-

I

+

1 Oct-1

Free Probe

FIG. 2. N-Acetylcysteine does not inhibit the Oct-1 transcription factor. T h e same nuclear protein extracts from Fig. 1 were probed with a Oct-1 probe, T h e top arrow indicates the constitutive Oct-I band, the bottom arrow the u n b o u n d probe. ( R e p r o d u c e d from Staal et aL 4 with permission.)

174

SIGNAL TRANSDUCTION AND GENE REGULATION

[ 17]

results can be obtained with many different cell lines and stimulation conditions. Quantitative results can be obtained by scanning the gel with a phosphorimager (radioanalytic imaging system). It is important to ascertain the specificity of the NF-KB inhibition by antioxidants such as NAC. Therefore, we usually use the same nuclear protein extracts to probe another transcription factor that is not redoxsensitive. The ubiquitous Oct-1 transcription factor can be useful for this purpose. As shown in Fig. 2, conditions that effectively inhibit activation of NF-KB do not affect Oct-1 expression, demonstrating that NAC selectively inhibits NF-xB. Finally, the presence of activated NF-KB in a nuclear extract does not necessarily mean that NF-KB is transcriptionally active (binding does not imply function). Functional NF-KB can be readily assayed using reporter genes, the transcription of which is directed by an NF-KB-dependent promoter. We have used Jurkat T cells stably transfected with a plasmid containing three KB motifs fused to the lacZ (/3-galactosidase) gene (Jurkat kB5.2). 4 Cells can be stimulated under the appropriate conditions and assayed for/3-galactosidase activity by biochemical means in lysates ~3 or by the fluorescence-activated cell sorting (FACS)-Gal assay using a flow cytometer. TMIt is beyond the scope of this chapter to describe these methods in detail, but readers are referred to protocols given elsewhereJ 5 Acknowledgment Supported in part by NIH CA-42509.

13 M. Roederer, F. J. T. Staal, P. A. Raju, L. A. Herzenberg, and L. A. Herzenberg, Proc. Natl. Acad. Sci. U.S.A. 87, 4884 (1990). 14 G. P. Nolan, S. N. Fiering, J. F. Nicolas, and L. A. Herzenberg, Proc. Natl. Acad. Sci. U.S.A. 85, 2603 (1988). 15 M. Roederer, S. N. Fiering, and L. A. Herzenberg, Methods (San Diego) 2, 248 (1991).

[ 18]

D N A - P R O T E I N INTERACTIONS

[18] R e d o x R e g u l a t i o n o f D N A - P r o t e i n by Biothiols

175

Interactions

By YUICmRO J. SUZUKIand LESTZRPACKER Introduction Redox regulation of DNA-protein interactions plays an important role in mechanisms of actions of many mammalian transcription factors for controlling gene transcription. NF-KB DNA-binding activity is stimulated by reducing agents (dithiothreitol, 2-mercaptoethanol, thioredoxin) and abolished by exposure to sulfhydryl-modifying agents (diamide, N-ethylmaleimide). 1-~ Important sulfhydryl groups such as Cys-62 in the p50 subunit of NF-~B for DNA-binding activity have been identified.6 Similarly, the DNA-binding activity of the Fos-Jun heterodimeric complex (AP-1) has been shown to be regulated by a redox mechanism.7 A ubiquitous protein, redox factor-1 (Ref-1), appears to participate in regulation of AP-1. 8"9 Steroid receptors such as glucocorticoid receptoP ° and progesterone receptoP 1 also require reduced sulfhydryl groups for DNA-binding activity. Hence, the identification of reductants which are capable of influencing the D N A protein interactions has physiological and pharmacological importance as they can enhance the activities of transcription factors, and in turn either positively or negatively modulate gene expression. Here we describe a simple and rapid methodology to screen effects of reductants on DNA-protein interactions. This methodology utilizes the knowledge that DNA-binding proteins usually require the presence of reducing agents such as dithiothreitol (DTI') to exhibit DNA-binding activity. Thus, creating a nonreducing environment by omitting DTT from the I M. B, Toledano and W. J. Leonard, Proc. Natl. Acad. Sci. U.S.A. 88, 4328 (1991). 2 j. A. Molitor, D. W. Ballard, and W. C. Greene, New Biol. 3, 987 (1991). 3 T. Okamoto, H. Ogiwara, T. Hayashi, A. Mitsui, T. Kawabe, and J. Yodoi, Int. Immunol. 4, 811 (1992). 4 j. R. Matthews, N. Wakasugi, J.-L. Virelizier, J. Yodoi, and R. T. Hay, Nucleic Acids Res. 20, 3821 (1992). 5 T. Hayashi, Y. Ueno, and T. Okamoto, J. Biol. Chem. 268, 11380 (1993). 6 j. R. Matthews, W. Kaszubska, G. Turcatti, T. N. C. Wells, and R. T. Hay, Nucleic Acids Res. 21, 1727 (1993). 7 C. Abate, L. Patel, F. J. Rauscher III, and T. Curran, Science 2,49, 1157 (1990). 8 S. Xanthoudakis and T. Curran, EMBO J. 11, 653 (1992). 9 S. Xanthoudakis, G. Miao, F. Wang, Y.-C. E. Pan, and T. Curran, EMBO J. 11, 3323 (1992). 10 C. M. Silva and J. A. Cidlowski, J. Biol. Chem. 264, 6638 (1989). H S. Peleg, W. T. Schrader, and B. W. O'Malley, Biochemistry 28, 7373 (1989).

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

176

SIGNAL TRANSDUCTION AND GENE REGULATION

[ 18]

general protocol allows the isolation of DNA-binding proteins whose activity cannot be detected, yet a potential for DNA-binding activity is retained in the presence of appropriate reducing factors. Briefly, nuclear extracts are isolated by a miniextraction procedure without the use of DTT, 12,13 followed by binding reactions with interesting reductants for electrophoretic mobility shift assays. The results can be obtained within 24 hr.

Preparation of Nuclear Miniextracts 1. Pellet 1 × 106 cells by centrifuging for 10 rain at 1200 rpm in microcentrifuge. 2. Remove supernatant. Wash pellet with 1 ml of cold phosphatebuffered saline (PBS). 3. Pellet ceils by centrifuging for 15 sec at 14,000 rpm in microcentrifuge. 4. Remove supernatant and resuspend pellet in 0.4 ml of buffer A [10 mM HEPES, pH 7.8, 10 mM KC1,2 mM MgCI2, 0.1 mM EDTA, 0.1 mM phenylmethylsulfonylfluoride (PMSF), 5/zg/ml antipain, 5/zg/ml leupeptin, 1 mM DTT*]. 5. Incubate on ice for 15 min to swell ceils. 6. Add 25/M of 10% Nonidet P-40 and vortex vigorously for 15 sec. 7. Centrifuge homogenate for 30 sec at 14,000 rpm in microcentrifuge. 8. Remove supernatant and resuspend nuclear pellet in 50/~1 of buffer C [50 mM HEPES, pH 7.8, 50 mM KC1, 300 mM NaC1, 0.1 mM EDTA, 0.1 mM PSMF, 10% (v/v) glycerol, 1 mM DTT*]. 9. Mix vigorously for 20 rain at 4°. 10. Centrifuge for 5 rain at 14,000 rpm in microcentrifuge at 4 °. 11. Harvest supernatant containing the nuclear proteins. 12. Determine protein concentrations. 13. Samples can be stored in aliquots at - 8 0 ° until use.

Electrophoretic Mobility Shift Assay The simplest and the most user-friendly method to study DNA-protein interactions is the electrophoretic mobility shift assay (EMSA; band-shift assay, gel-shift assay, gel retardation assay). This technique was first devel12E. Schreiber, P. Matthias, M. M. Mueller,and W. Schaffner,Nucleic Acids Res, 17, 6419 (1989). 13F. J. T. Staal, M. Roederer, L. A. Herzenberg, and L. A. Herzenberg, Proc. Natl. Acad. Sci. U.S.A. 87, 9943 (1990). * The DTT is omittedfrom buffersA and C in order to create a nonredueingenvironment.

[ 18]

DNA-PROTEIN INTERACTIONS

177

oped simultaneously by two groups in 1981.14'15 There already exist excellent reviews which describe the theory and step-by-step instructions for the EMSA. 16,17 Thus, users should refer to those sources for detailed instructions. The following is one example of a protocol for performing the EMSA. Binding reaction mixtures (20 ~1) containing the following components are incubated for 20 min at 25°: 10 m M Tris-HC1 (pH 7.5), 50 m M NaCI, 0.2 m M E D T A , 2% (v/v) glycerol, nuclear protein extract, poly(dI-dC) • poly(dI-dC) (nonspecific competitor), 32p-labeled D N A fragment, and DTT. Titrations of nuclear protein extract, nonspecific competitor, and 32p-labeled D N A fragment may be necessary to determine the optimal conditions for the assay. D T T is omitted to create a nonreducing environment. Titration with D T T may also provide useful information in determining the redox characteristics of D N A - p r o t e i n interactions. Reducing factors of interest are also included in the binding reaction. Proteins are separated by electrophoresis through a native polyacrylamide gel, and detection can be made by autoradiography following standard procedures.

E x a m p l e : E n h a n c e m e n t of NF-KB DNA Binding b y Biothiols NF-KB is a mammalian transcriptional activator that is involved in the transmission of signals from the cytoplasm to the nucleus, and it participates in activation of genes involved in the immune, inflammatory, or acute phase responses, such as various cytokines and surface receptors as well as viruses including human immunodeficiency virus (HIV). is Studies also point to the importance of this transcription factor in neurological disorders 19 and atherosclerosis. 2° The DNA-binding activity of NF-~cB is regulated through redox mechanisms. Treatment of the activated form of NF-KB molecule with N-ethylmaleimide, diamide] ,2 or hydrogen peroxide has been shown to eliminate D N A binding. Inversely to the inhibition by oxidants, thiol reductants such as D T T and reduced thioredoxin enhance the DNA-binding activity. 4 The present study demonstrates that other biothiols including dihydrolipoate, N-acetylcysteine, and L-cysteine also enhance the D N A NF-KB interactions. ~ M. M. Garner and A. Revzin, Nucleic Acids Res. 9~ 3047 (1981). t5 M. Fried and D. M. Crothers, Nucleic Acids Res. 9, 6505 (1981). 16G. H. Goodwin, in "Gel Electrophoresis of Nucleic Acids: A Practical Approach" (D. Rickwood and B. D. Haines, eds.), p. 225. IRL Press, Oxford, 1990. ~7j. Carey, this series, Vol. 208, p. 103. 18p. Baeuerle, Biochim. Biophys. A c t s 1072, 63 (1991). 19B. Kaltschmidt, P. A. Baeuerle, and C. Kaltschmidt, Mol. Aspects Med. 14, 171 (1993). 20 F. Liao, A. Andalibi, F. C. deBeer, A. M. Fogelman, and A. J. Lusis, J. Clin. Invest. 91, 2572 (1993).

178

SIGNAL

TRANSDUCTION

AND

GENE

REGULATION

[ 18]

Materials and Methods Jurkat T (acute human leukemia) cells are grown in RPMI 1640 medium supplemented with 10% fetal calf serum, 1% penicillin/streptomycin, and 1% glutamine. Cells are plated at a density of 1 × 106 cells per well in 1.0 ml. Recombinant Escherichia coli-derived human tumor necrosis factor-a (TNF-a; 25 ng/ml; Genentech Inc., South San Francisco, CA) is added, and cells are incubated for 1 hr in an atmosphere of 5% (v/v) .CO2 in air humidified at 37 °. Nuclear extracts are prepared from 1 × 106 cells as described above without the use of DT]?. The binding reaction mixtures (20 /zl) for E M S A containing 1.5 ~g protein of nuclear extract, 0.5 /~g poly(dI-dC)-poly(dI-dC), 32p-labeled oligonucleotide probe for NF-KB (see sequences below), 50 m M NaC1, 0.2 m M EDTA, 2% (v/v) glycerol, and 10 mM Tris-HCl (pH 7.5) are incubated for 20 min at 25 °. In some experiments, binding reaction mixtures also contain DTT (Sigma, St. Louis, MO), dihydrolipoate (Asta Medica, Frankfurt, Germany), Noacetylcysteine, L-cysteine, ascorbate, a-tocopherol, NADH, N A D P H , or pyruvate (Sigma). Proteins are separated by electrophoresis through a native 6% polyacryl-

4-

4-

4-

4-

4-

4-

4-

4-

4-

4-

NF-rB

,~-Freepr~ FIG. 1. Effects of thiol and nonthiol reductants on NF-rB DNA-bindingactivity.Nuclear extracts (NE) from nonstimulatedor TNF-c~-stimulatedcells were prepared without the use of dithiothreitol (DTT). Various reductants were added in binding reaction mixtures for the electrophoretic mobility shift assay before the addition of 3~P-labeled NF-KB-specific oligonucleotide probe. DHLA, Dihydrolipoate;NAC, N-acetylcysteine.

[181

DNA-PROTEIN INTERACTIONS

179

amide gel in a running buffer of 12.5 mM Tris-borate (pH 8.0) containing 0.25 mM EDTA, followed by autoradiography. 5' G A T C C G A G G G G A C T T T C C G C T G G G G A C T T T C C A G G 3' 3' GCT C C C C T G A A A G G C G A C C C C T G A A A G G T CCCTAG 5'

Results and Discussion

In EMSA experiments which were performed without the use of DTT, no observable NF-KB bands were detected in nuclear extracts from either nonstimulated or TNF-o~-stimulated Jurkat T cells (Fig. 1, lanes 2 and 5, respectively). Although no changes were observed in nonstimulated samples (Fig. 1, lanes 3 and 4), the addition of thiol reductants, DTT or dihydrolipoate (10 ~M-1 raM), to the binding reaction mixtures caused concentration-dependent restorations of NF-KB DNA-binding activity in nuclear extracts from TNF-o~-stimulated cells (Fig. 1, lanes 6-11). Other thiols, namely, N-acetylcysteine (Fig.l, lane 12) and e-cysteine (Fig. 1, lane 13), also enhanced the NF-KB DNA-binding activity. In contrast, nonthiol reductants such as ascorbate (Fig. 1, lane 14), ot-tocopherol (Fig. 1, lane 15), NADH, NADPH, and pyruvate (data not shown) did not enhance the DNA-binding activity. The present example demonstrates that reduced thiols including DTT, dihydrolipoate, N-acetylcysteine, and L-cysteine enhance the DNA-binding activity of NG-KB. The order of potencies of the thiols in enhancing DNAbinding activity generally agrees with the order of redox potentials (see Table I). In contrast, nonthiol reductants, including those with very negative redox potentials such as pyruvate, NADH, and NADPH (see Table I), do

TABLE I REDOX POTENTIALSFOR THIOLS AND NONTHIOLSa System

Eo (V)

Pyruvate/acetate + CO2 Dithiothreitol, ox/red NAD(P)+/NAD(P)H + H + Lipoate, ox/red L-Cystine/cysteine

-0.70 -0.33 -0.32 -0.29 -0.22

Midpoint potentials at pH 7.0; ox/red, oxidized form/reduced form.

" E o,

180

SIGNAL TRANSDUCTION

AND GENE

REGULATION

[19]

not enhance DNA-NF-KB interactions. Therefore, biothiols may regulate gene transcription by enhancing the DNA-binding activity of NF-KB transcription factor. Acknowledgments Research was supported by the National Institutes of Health (CA47597) and ASTA Medica. This work was done during the tenure of a research fellowship from the American Heart Association, California Affiliate, to Y. J. S.

[ 191 S u p p r e s s i o n of Protooncogene c-los Expression Antioxidant Dihydrolipoic Acid

By

by

MASASHI M I Z U N O a n d LESTER PACKER

Introduction Cells in a growth-arrested quiescent state require a set of growth factors for initiation of cell proliferation. 1 Competence factors include serum, tumor promoters such as phorbol 12-myristate 13-acetate (PMA), plateletderived growth factor (PDGF), and so on. When quiescent cells are stimulated by these factors, certain genes called immediate early genes are immediately transcribed without de novo protein synthesis. The protooncogene c-fos is one of the immediate early genes, and the product derived from the gene is thought to participate in signal transduction after stimuli. It is induced very rapidly and transiently in response to extracellular stimuli. 2 For example, in the case of PMA stimulation, the expression of cofos is rapidly increased, with m R N A levels reaching a maximum within the first hour after stimulation and decreasing to constitutive levels thereafter. A substantial amount of evidence suggests that the generation of free radicals, such as superoxide anion and hydroxyl radicals, may be involved in tumor promotion. A number of studies indicate that various antioxidants inhibit tumor promotion. Additionally, several lines of evidence indicate that generation of reactive oxygen species (ROS) is one of the earliest events involved in stimulation of cell growth in response to growth factors or in growth-promoting conditions. It has demonstrated that hydrogen peroxide and superoxide anion radical induce expression of c-fos m R N A A. B. Pardee, R. Dubrow, J. L. Ham/in, and R. F. Kletzien, Annu. Rev. Biochem. 47, 715 (1978). z M. E. Greenberg and E. B. Ziff, Nature (London) 311, 433 (1984).

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc. All rights of reproduction in any form reserved.

[191

SUPPRESSION o r c-los BY DIHYDROLIPOIC ACID

181

and D N A synthesis in quiescent BALB/3T3 cells,3 and it has been reported that low concentrations of oxidants act as a mitogenic signal. 4 It has also been reported that ROS are produced by cultured cells when they are stimulated by PMA. 5 Furthermore, the antioxidant retinoic acid suppresses expression of the protooncogene c-los in cells stimulated by PMA and serum. 6 This evidence indicates the possibility that ROS are mitogenic and mediate growth-associated gene expression. oL-Lipoic acid, one of the disulfide compounds, is an essential substance synthesized in the mammalian liver and other organs which serves in physiological systems as a cofactor for ot-ketoacid dehydrogenase reactions (e.g., in the pyruvate dehydrogenase complex), o~-Lipoic acid exerts antioxidant actions in vivo and in vitro] Moreover, the reduced form of o~-lipoic acid, dihydrolipoic acid (DHLA), acts as a potent antioxidant and scavenges superoxide, hydroxyl, peroxyl, and other radical species, s Because the formation of ROS plays an important role in the expression of m R N A after extracellular stimuli, antioxidants might suppress expression of protooncogenes. A method is presented here that utilizes D H L A to investigate the suppressive effect of antioxidants on the expression of the protooncogene c-fos induced by PMA. RNA Isolation and Northern Analysis The difficulty in R N A isolation is that most ribonucleases (RNases) are very stable and active enzymes. It is important that cells are lysed under chemical conditions that promote RNase denaturation. The following steps should be taken to avoid contamination by RNase: (1) all water and salt solutions should be treated with diethyl pyrocarbonate (DEPC); (2) glassware should be heated to 250 ° for 4 hr, and plasticware should be disposable; (3) gloves should be worn at all times during the extraction procedure and when handling materials and equipment. Caution: DEPC is a suspected carcinogen and should be handled carefully in a hood. Cell Culture

Jurkat T cells, a human T-cell leukemia cell line, are grown in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS), 1% glutamine, 3 M. Shibanurna, T. Kuroki, and K. Nose, Oncogene 3, 17 (1988). 4 D. Crawford, I. Zbinden, P. Amstad, and P, Cerutti, Oncogene 3, 27 (1988). 5 M. Shibanurna, T. Kuroki, and K, Nose, Biochern. Biophys. Res. Cornrnun. 144, 1317 (1987). 6 K. J. Busam, A. B. Roberts, and M. B. Sporn, J, Biol. Chem. 267, 19971 (1992). 7 j. Peinado, H. Sies, and T. P. Akerboom, Arch. Biochem. Biophys. 273, 389 (1989). Y. J. Suzuki, M. Tsuchiya, and L. Packer, Free Radical Res, Cornmuns. 15, 255 (1991).

182

SIGNALTRANSDUCTIONAND GENE REGULATION

[191

1% sodium pyruvate, and antibiotics including 100 units/ml of penicillin and streptomycin in 5% (v/v) CO2 in air at 37°. At approximately 70% confluency (7 x l0 s to 9 x 105 cells/ml), cells are suspended at a density of 1 x 106 cells/ml in medium in which the concentration of FCS is reduced from 10 to 1%, and the mixtures are incubated for 16-18 hr in the presence or absence of 0.2 or 0.02 mM D H L A in dimethyl sulfoxide (DMSO) solution. Then PMA (Sigma, St. Louis, MO) is added (0.5/xM) to the medium to stimulate the protooncogene c-los, and the cell suspensions are incubated.

Extraction o f Total R N A Total RNA is prepared essentially according to the acidic guanidiumphenol-chloroform method of Chomczynski and Sacchi9 with slight modification.

Reagents Denaturing solution (solution D): 4 M guanidium thiocyanate, 25 mM sodium citrate (pH 7.0), 0.5% Sarkosyl, 0.1 M 2-mercaptoethanol DEPC-water: Add 0.2 ml DEPC to 100 ml water, shake vigorously, let stand for at least 30 rain, and then autoclave to inactive the remaining DEPC. Water-saturated phenol: Dissolve phenol cyrstals at 68° and saturate with DEPC-water

Procedure 1. Harvest cell suspensions at indicated times and centrifuge for 10 min at 500 g. Remove the supernatant under vacuum. 2. Lyse the cell pellet in 1 ml of denaturating solution per 107 cells. 3. Add 0.1 ml of 2 M sodium acetate (pH 4.0), 1 ml of phenol (water saturated), and 0.2 ml of chloroform-isoamyl alcohol mixture (49:1, v/v) to the solution D homogenate, mixing by inversion after the addition of each reagent. Shake vigorously for 10 sec and cool on ice for 15 min. 4. Centrifuge at 10,000 g for 20 rain at 4 °. Transfer the upper aqueous phase to a new centrifuge tube that has been prechilled on ice. 5. Mix with an equal volume of 2-propanol and place at -20 ° for at least 1 hr to precipitate RNA. 6. Centrifuge at 10,000 g for 20 rain at 4° and discard the supernatant. Dissolve the pelleted RNA in 0.3 ml of solution D and transfer to a 1.5ml Eppendorf tube. 9p. Chomczynskiand N. Sacchi,Anal. Biochem. 162, 156 (1987).

[19]

SUPPRESSION OF

c-los

BY DIHYDROLIPOIC ACID

183

7, Precipitate the R N A with 0.3 ml of 2-propanol at - 2 0 ° for 1 hr. Centrifuge at 10,000 g for 20 rain at ambient temperature and remove the supernatant from the R N A pellet. 8. Wash the R N A pellet with 70% (v/v) ethanol (prepared with D E P C water). Centrifuge 10 rain at 10,000 g and discard ethanol solution, then dry under vacuum for 15 min. Note: R N A is protected from RNase by guanidium thiocyanate prior to Step 8. The R N A is no longer protected now. Take special care that RNase contamination does not occur under any circumstance. 9. Dissolve the R N A in 200/~1 of DEPC-water. Dilute 10/xl of R N A solution to 1 ml and measure the optical density at 260 nm to determine its concentration: 1 OD260 = 40 ~g/ml RNA.

Analysis of RNA Purified R N A is subjected to Northern blot analysis to investigate the effect of D H L A on the expression of the protooncogene c-los when ceils are stimulated with PMA.

Reagents 20× MOPS: 0.4 M 3-(N-morpholino)propanesulfonic acid (MOPS), pH 7.0, 0.I M sodium acetate, 20 mM ethylenediaminetetraacetic acid (EDTA) 20x SSPE: 0.2 M sodium phosphate (pH 7.4), 2.98 M sodium chloride, 20 mM E D T A disodium salt Hybridization buffer: 1 M sodium chloride, 50 mM Tris-HC1 buffer (pH 7.5), 10% dextran sulfate, 1% sodium dodecyl sulfate (SDS), 100 ~g/ml denatured salmon sperm D N A R N A loading buffer: 720/~1 formamide, 80/xl of 20x MOPS, 260/.d formaldehyde, 150/~1 of 50% glycerol, 10 tzl bromphenol blue dye (saturated with DEPC-water), and 280/xl DEPC-water

Procedure 1. Adjust the volume of R N A sample (20 ~g) to 5 ~1 with D E P C water; add 16 ~1RNA loading buffer and I ~1 of 1 mg/ml ethidium bromide. 2. Electrophorese 20 ~g of total R N A in a 1.2% agarose gel containing 0.66 M formaldehyde using 1 × MOPS as running buffer. 3. Transfer the R N A from the gel to a nylon membrane (Gene Screen Plus, N E N - D u Pont, Boston, MA) by upward capillary transfer in 10× SSPE. 4. Place RNA-side down on a UV transilluminator (254 nm wavelength) for 3 rain.

184

[19]

SIGNAL TRANSDUCTION AND GENE REGULATION

5. Label the 40-mer D N A probe for c-fos and glyceraldehyde-3-phosphate dehydrogenase (GAPDH, Oncogene Science, NY) at 5' DNA fragment termini with [y-32p]dATP (ICN Biomedicals, Costa Mesa, CA) using T4 polynucleotide kinase (Amersham, Arlington Heights, IL). 6. Purify radiolabeled D N A probe from unincorporated probe through Sephadex G-25 column. 7. Prehybridize the membrane for at least 2 hr at 65° in hybridization buffer and hybridize in the same solution with 5 x 10 6 t o 10 7 counts/min (cpm) 32p-labeled D N A probe at 65° for 20 to 24 hr. 8. Briefly rinse the membrane four times with 2x SSPE, 0.1% SDS at room temperature, and wash with the same solution prewarmed to 65 ° for 30 min at 65°. 9. Carry out a wash using 2x SSPE, 0.1% SDS for 10 min at room temperature followed by a final incubation in 0.2x SSPE for 5 min at room temperature. 10. Blot excess liquid from the membrane, cover it with plastic wrap, and then expose to X-ray film (Amersham).

Caution: Formamide is a teratogen. Formaldehyde is toxic. Ethidium bromide is a mutagen and potential carcinogen. All operations involving these reagents should be carried out with care. Exposure to UV light is harmful to the eyes and skin. Eyes and skin need protection when using a UV light source.

28S c-fos 18S 28S

18S GADPH 0

15

30

45

60

75

90

120 (min)

FI6. 1. Kinetics of c-fos mRNA expression. Cells were incubated in RPMI 1640 supplemented with 1% FCS for 18 hr with DMSO at a concentration of 2/~l/ml. Thereafter, cells were exposed to 0.5/~M PMA and analyzed at the times indicated. Total RNA was extracted and Northern analyses performed hybridizing with a 32P-labeled c-fos or GAPDH DNA probe.

[19l

SUPPRESSION OF

c-los

185

BY DIHYDROLIPOIC ACID

J 28S

c-los

18S

28S

•~

18S

GAPDH

F]G. 2. Dose-dependent inhibition of c-fos m R N A expression by DHLA. Cells were incubated in the indicated concentration of D H L A in 1% FCS/RPMI 1640 for 18 hr prior to stimulation by 0.5 /zM PMA for another 1 hr. As a control, cells were incubated with the same volume of DMSO solution. Total R N A was extracted and Northern analyses performed hybridizing with a 32p-labeled c-los or G A P D H D N A probe.

Application The kinetics of c-fos accumulation after treatment of cells with PMA was studied as shown in Fig. 1. A rapid and transient increase in c-fos mRNA was observed after PMA stimulation in the absence of DHLA. The level of c-los mRNA became detectable 30 min after the addition of PMA, peaked at 60 min, and returned rapidly to constitutive levels by 2 hr. G A P D H expression, which served as a positive control for a constitutively expressed gene, showed no changes under the different experimental conditions. To investigate whether the onset of c-los mRNA induction by PMA is altered by DHLA, cells were cultured for 18 hr in the presence or absence of D H L A (0.02 or 0.2 raM) before 0.5 /xM PMA was added for another 1 hr (Fig. 2). Suppression of PMA-induced c-fos mRNA levels was already seen at 0.02 mM DHLA. At 0.2 mM DHLA, the level of c-fos mRNA was,

186

SIGNAL T R A N S D U C T I O N A N D G E N E R E G U L A T I O N

[20]

further decreased. This experiment suggests that effects of antioxidants can be estimated at the molecular biological level by using the detection of immediate early genes such as c-los mRNA after PMA stimulation. Acknowledgments Research was supported by National Institutes of Health (CA 47597) and ASTA Medica (Germany).

[201 D i h y d r o l i p o a m i d e

By M U L C H A N D

Dehydrogenase:

Activity Assays

N. V E T T A K K O R U M A K A N K A V , and TE-CHUNG LIU

S. P A T E L , N A T A R A J

Introduction Dihydrolipoamide dehydrogenase (DHLipDH; EC 1.8.1.4) catalyzes the reoxidation of the protein-bound dihydrolipoamide moiety as per the reaction shown. Protein-dihydrolipoamide + NAD ÷ .~ protein-lipoamide + N A D H + H ÷ The enzyme has been extensively characterized from a variety of eubacteria, archaebacteria, and eukaryotes, and the properties of the enzyme are similar among members of these groupsJ -3 This flavoprotein disulfide oxidoreductase catalyzes the above reaction as part of the pyruvate, a-ketoglutarate, and branched-chain a-keto acid dehydrogenase multienzyme complexes as well as in the glycine cleavage multienzyme complex.4'5 The dihydrolipoamide acyltransferase component of these a-keto acid dehydrogenase complexes and hydrogen carrier protein (H-protein) component of the glycine cleavage system are the physiological substrates for this enzyme. The enzyme has a redox disulfide and a tightly but noncovalently bound FAD cofactor, both of which participate in the transfer of electrons from dihydrolipoamide to NAD ÷ (reviewed by Williams1'2). The DHLipDH func1 C. H. Williams, Jr., in "Chemistry and Biochemistry of Flavoenzymes" (F. Muller, ed.), Vol. 3, p. 121. CRC Press, Boca Raton, Florida, 1992. 2 C. H. Williams, Jr., in "The Enzymes" (P. D. Boyer, ed.), Vol. 13, 3rd ed., p. 87. Academic Press, New York, 1976. 3 D. J. Carothers, G. Ports, and M. S. Patel, Arch. Biochern. Biophys. 268, 409 (1989). 4 S. J. Yeaman, Biochem. J. 257, 625 (1989). 5 M. S. Patel and T. E. Roche, FASEB J. 4, 3224 (1990).

METHODS IN ENZYMOLOGY,VOL. 252

Copyright © 1995 by AcademicPress, Inc. All fights of reproductionin any form reserved.

[201

DIHYDROLIPOAMIDE DEHYDROGENASE

187

tions as a dimer and is inactive as a monomer because the active site is composed of amino acid residues from both subunits. 1'2 The reaction catalyzed by DHLipDH is a Ping-Pong Bi-Bi mechanism that can be divided into two half-reactions: (i) the reduction of the oxidized enzyme [E] to the reduced form [EH2] and (ii) the transfer of electrons from the [EH2] form to the nicotinamide cofactor.6'7 Crucial to the characterization of this flavoenzyme has been the ease with which the enzyme can be assayed in the forward and backward reactions as well as by electron transfer to electron-accepting dyes. The simplest assay procedure is monitoring the NAD+-dependent oxidation of free OLdihydrolipoamide. Dihydrolipoamide + NAD ÷ ,~ lipoamide + NADH + H + The aim of this chapter is to describe the commonly used procedures to monitor the enzyme activity, both from prokaryotic cells as well as from eukaryotic cells. The forward, or physiological, reaction is widely used to monitor the activity of DHLipDH and is described in detail. We first describe the preparation of crude cell exracts to assay DHLipDH activities from prokaryotic and eucaryotic cells, followed by the actual assay procedures.

Preparation of Crude Cell Extracts

Bacterial Cell Extracts Bacterial cell extracts can be prepared for detection of DHLipDH activity by any of the routine cell lysis procedures. These include lysis by osmotic shock, which is done by suspending frozen cells in 50 mM potassium phosphate (pH 7.4) and subsequent treatment with DNase I to reduce the viscosity of the preparation (Haloferax volcanii), s lysis by sonication (6 to 10 30-sec cycles in an ice-cooled sonication cell at 50% pulse efficiency),9 and lysis by French press treatment (Escherichia coli). After the bacterial cells have been lysed, the preparation is clarified by centrifugation at 6000 g for 10 min to remove cell debris and cells that 6 V. Massey, Q. H. Gibson, and C. Veeger, Biochem. J. 77, 341 (1960). 7 V. Massey and C. Veeger, Biochim. Biophys. Acta 48, 33 (1961). s N. N. Vettakkorumakankav, M. J, Danson, D. W. Hough, J. A. Young, M. Davison, and K. J. Stevenson, Biochem. Cell Biol. 70, 70 (1992). 9 S. R. Adamson and K. J. Stevenson, Biochemistry 20, 3418 (1981).

188

SIGNAL TRANSDUCT1ON AND GENE REGULATION

[20]

escaped lysis. The crude extract at this stage can be used for the assay of D H L i p D H activity.

Eukaryotic Cell Preparations If membranes or eukaryotic cell organelles are to be assayed for D H L i p D H activity, the enzyme must be solubilized by treatment with a detergent or sonication before activity can be assayed. Cells (cultured) are washed three times in phosphate-buffered saline (PBS; 130 mM NaC1, 10 mM Na2HPO4, 10 mM NaH2PO4, pH 7.4) and resuspended in hypotonic extraction buffer (20 mM potassium phosphate, pH 7.4; 1% (w/v) Triton X-100) containing 2 mM EDTA, 2mM EGTA, and protease inhibitors [0.2 mM phenylmethylsulfonyl fluoride (PMSF), 0.5 tzg/ml leupeptin, 0.7/zg/ ml pepstatin]. The cell suspensions are incubated on ice for 30 min, subjected to three cycles of freezing and thawing, and finally placed on ice for a further 30 rain to ensure complete solubilization of the enzyme. The whole cell extracts can be used for the determination of D H L i p D H activity. Mitochondrial extracts for the assay of D H L i p D H activity are prepared as follows. The tissues are minced and homogenized (10%, w/v) in a buffer containing 70 mM sucrose, 220 mM mannitol, 2 mM HEPES, and 1 mM EDTA, pH 7.4 (for preparation of mitochondria from liver cells), or 250 mM sucrose, 10 mM Tris, 1 mM 2-mercaptoethanol, pH 7.4, for the preparation of mitochondria from kidney or heart cells. 1° The homogenization of the tissue is performed by six or seven passes through a motor-driven Fisher Dynamix homogenizer. Unbroken cells and nuclei are removed by centrifugation at 650 g for 10 min and the pellet discarded. The floating lipid on the surface of the supernatant may be removed by pipetting. The supernatant is then centrifuged at 12,000 g at 4° for 10 rain to sediment the mitochondria. The mitochondrial pellet is resuspended in half the original volume of homogenization buffer. The suspension is passed once through a hand-held homogenizer and the preparation centrifuged at 12,000 g for 10 rain. The pellet is resuspended in one-fourth the original volume of the homogenization buffer and centrifuged again. The pellet is finally resuspended in phosphate-buffered saline containing the protease inhibitors described above (the amount of solution used to resuspend the pellet is 1 ml/g of tissue used). Aliquots (1 ml) can be stored frozen at - 7 0 ° until required. When required for enzyme assays, the aliquots are thawed and extracted once in the hypotonic extraction buffer with protease inhibitors as described for cell extracts or sonicated to ensure complete lysis. The extract is clarified by centrifugation at 20,000 g for 30 min at 4°. The extract at this stage may be used for activity assays. 1° 10 D. Carothers, Ph.D. Thesis, Case Western Reserve University, Cleveland, Ohio (1987).

[20]

DIHYDROLIPOAMIDE DEHYDROGENASE

189

Preparation o f Cultured Skin Fibroblasts f r o m Patients Deficient in E n z y m e Activity

Control and patient skin fibroblasts are normally cultured in minimum essential medium supplemented with 20% (v/v) fetal calf serum, 100 Ixg/ ml streptomycin, 60 tzg/ml penicillin, and 1.5/zg/ml amphotericin B Y The cells are harvested and lysed by resuspension in phosphate-buffered saline containing 0.5% Triton X-100 and sonicationJ 2

Purification of D i h y d r o l i p o a m i d e D e h y d r o g e n a s e The purification of D H L i p D H from several sources has been reported. General purification strategies include ammonium sulfate fractionation of crude extracts, heat treatment, and chromatography on a hydroxyapatite matrix. 13 The aforementioned procedure has been modified and used to purify the enzyme from several other sources, 14'15including some extremophiles, s Because D H L i p D H is extremely heat stable, a heat treatment step may be routinely included in the purification strategy. The yield of purified enzyme from most sources is approximately 30%. The specific activity of the purified human enzyme 14 is 407 units/rag protein.

S y n t h e s i s of D i h y d r o l i p o a m i d e T o a chilled suspension of DL-lipoamide (200 mg in 4 ml of methanol and 1 ml of water), 1 ml of a cold solution of sodium borohydride (200 mg/ml) is added and stirred on ice for 1 hr. The reaction is acidified to p H 1-2 with 1 M HCI and extracted three times with chloroform. The extracts are pooled and evaporated to dryness on a rotary evaporator. The resulting white residue is resuspended in 10 ml benzene, and 4 ml n-hexane is added. The solution is set on ice until white precipitates are formed. The precipitates are filtered on a Whatman 3MM paper and dried under vacuum. The melting point of DL-dihydrolipoamide is between 50° and 55 °. The reduction of lipoamide to dihydrolipoamide is not complete, and therefore the reagent may be separated from the oxidized form by high performance liquid chromatography (HPLC) on a 0.39 × 15 cm Waters (Milford, MA) C18, Pico-Tag column. The solvent system used is 10 m M u L. Ho, C.-W. C. Hu, S. Packman, and M. S. Patel, J. Clin. Invest 78, 844 (1986). LaT,-C. Liu, H. Kim, C. Arizmendi, A. Kitano, and M. S. Patel, Proc. Natl. Acad. Sci. U.S.A. 90, 5186 (1993). 13C. H. Williams, Jr., J. Biol. Chem. 240, 4793 (1965). 14H. Kim, T.-C. Liu, and M. S. Patel, J. BioL Chem. 2.66, 9367 (1991) 15H. Kim and M. S. Patel, J. Biol. Chem. 267, 5128 (1992).

190

SIGNAL TRANSDUCTION AND GENE REGULATION

[201

sodium acetate, 30/zM EDTA, pH 6.3 (buffer A), and acetonitrile containing 30/zM E D T A (buffer B). The dihydrolipoamide is separated from the lipoamide by a linear gradient of 100% buffer A to 50% buffer B, and the elution of the compounds is monitored at 214 rim. The flow rate is 1 ml/min, and the gradient is established over 50 min (L. David Arscott and C. H. Williams, Jr., personal communication, 1993). The purified, dried dihydrolipoamide may be stored at - 2 0 ° until neededJ 6 Determination of Enzymatic Activity The forward reaction catalyzed by D H L i p D H is the oxidation of dihydrolipoamide and the reduction of NAD +. This can be monitored spectrophotometrically at 340 nm. Different concentrations of dihydrolipoamide have been used by various laboratories for routine assays of the forward reaction. The following final concentrations of various components produce satisfactory activities in our laboratory: 100 mM potassium phosphate, pH 8.0, 1.5 mM EDTA, 3.0 mM DL-dihydrolipoamide, 3.0 mM NAD ÷. Dihydrolipoamide solution is prepared in 95% (v/v) ethanol and appropriately diluted in the final reaction mixture to achieve the final concentration of 3 raM. Typically, 120 mM solutions of both dihydrolipoamide and N A D ÷ are prepared, and 25/zl of each is used in the assay (total volume of I ml). It is noted that in crude cell extracts, the presence of alcohol dehydrogenase can interfere with D H L i p D H activity assays. To avoid such an interference, a solution of dihydrolipoamide dissolved in acetone can be usd. 17 Several laboratories have used lower concentrations of dihydrolipoamideY 8 usually around 0.4 raM. We have found, however, that maximum enzyme activity can be realized on using higher dihydrolipoamide concentrations specified aboveJ 9 The reaction is initiated by the addition of the enzyme preparation. The progress of the reaction is monitored by following the increase in absorbance at 340 nm. An extinction coefficient for N A D H of 6.22 × 103 M -I cm -~ is used in the calculation of enzyme activity. The product of the forward reaction, NADH, is known to inhibit D H L i p D H J '2° Therefore, it is important that only the initial rates are used for determining enzyme activity, as product inhibition is apparent about 1.5 rain after initiation of the reaction. It is important to determine experimentally the appropriate 16 L. J. Reed, M. Koike, M. E. Levitch, and F. R. Leach, J. Biol. Chem. 232, 143 (1958). 17j. R. Sokatch, V. McCuUy, and C. M. Roberts, I. Bacteriol. 148, 647 (1981). 18M. J. Danson and R. N. Perham, Bioehern. J. 159, 677 (1976). 19 H. Kim, Ph.D. Thesis, Case Western Reserve University, Cleveland, Ohio (1992). 20 L. P. Hager and I. C. Gunsalus, J. A m . Chem. Soc. 75, 5767 (1953).

[201

DIHYDROLIPOAMIDE DEHYDROGENASE

191

concentrations of enzyme in the assay mixture, where the change in absorbance at 340 nm is linear over a period of 2-3 min and within 0.050.15 absorbance units. T o minimize the inaccuracies resulting from the inhibition of D H L i p D H activity by N A D H , an N A D + analog, acetylpyridine adenine dinucleotide, can be used to monitor the enzyme activity by following increase in absorbance at 363 nm. 13 The original assay using this compound, as reported by Williams 13 is described below. In a total reaction volume of 1 ml are mixed 50 m M sodium phosphate, p H 7.6, 1 mg bovine serum albumin, 1 m M E D T A , 0.9 m M acetylpyridine adenine dinucleotide, and 0.1 m M dihydrolipoamide. The assay is initiated with the addition of an enzyme preparation. An extinction coefficient of 9.1 × 103 M -1 cm -~ for reduced acetylpyridine adenine dinucleotide is used for calculations of enzyme activity. 21 A n o t h e r concern that has been reiterated is the fact that the dihydrolipoamide used in the enzymatic assays is a mixture of the D and L forms. Only the L- or the (R)-enantiomer is the natural substrate for the enzyme, and biphasic kinetics are observed using the mixture of substrates. 22 Therefore it is necessary to interpret kinetic parameters cautiously when using this mixture. One unit of enzyme activity is defined as the amount of enzyme required for the production of 1/zmol of N A D H or reduced acetylpyridine adenine dinucleotide in 1 min at 25 °. The specific acivity is expressed as units per milligram protein or per mole ravin cofactor (there is 1 mol F A D bound per mole of the subunit enzyme). The amount of enzyme-bound F A D can be quantitated from the absorption at 455 nm using an extinction coefficient of 11,300 M -a cm -~ for enzyme-bound F A D J 3 The D H L i p D H activity (expressed as units/mg protein) from various tissues using the dihydrolipoamide/NAD + assay are illustrated in Table I to provide the reader with a ready source of comparison.

Reverse R e a c t i o n The reverse reaction catalyzed by D H L i p D H is the reduction of lipoamide with N A D H as the reducing agent. The progress of the reaction can be monitored by the disappearance of N A D H at 340 nm. This assay has been previously described and is less common than the assay in the physiological direction. 6 The problem with this assay, especially with the E. coli D H L i p D H , is that N A D H inhibits D H L i p D H at high concentrations, as 2t j. M. Siegel, G. A. Montgomery, and R. M. Bock, Arch. Biochern. Biophys. 82, 288 (1959). 22L. D. Arscott and C. H. Williams, Jr., in "Flavins and Flavoproteins" (K. Yagi, ed.), p. 527. W. deGruyter, New York, 1994.

192

SIGNAL TRANSDUCTION AND GENE REGULATION

[20]

TABLE I SPECIFIC ACTIVITIES OF DIHYDROLIPOAMIDE DEHYDROGENASE IN CELLULAR PREPARATIONS FROM DIFFERENT SPECIES AS MONITORED BY FORWARD REACTION

Species Rat

Haloferax volcanii Escherichia coli

Pig heart Bovine liver Human liver Human enzyme (recombinant)

Preparation

Specific activity (units/mg protein)

Heart (homogenate) Kidney (homogenate) Skeletal (homogenate) Brain (homogenate) Crude extract Purified enzyme Purified enzyme Crude extract Crude extract Purified enzyme Crude extract Purified enzyme Purified enzyme

0.723 0.589 0.161 0.287 0.1 16.1 120.0 1.33 0.066 70.4 0.096 76.0 407.0

Ref. a

8 13 b c d 14

a N. N. Vettakkorumakankav and M. S. Patel, unpublished observations, 1993. b T. Hayakawa, M. Hirashima, S. Ide, M. Hamada, K. Okabe, and M. Koike, J. Biol. Chem. 241, 4694 (1966). c C. J. Lusty, J. Biol. Chem. 238, 3443 (1963). d S. Ide, T, Hayakawa, K. Okabe, and M. Koike, J. Biol. Chem. 242, 54 (1967). it c a n c o n v e r t t h e e n z y m e to t h e i n a c t i v e E H 4 f o r m w h i c h is a c c o m p a n i e d b y a p o s s i b l e d i m e r to m o n o m e r t r a n s i t i o n . 23,24

Diaphorase Activity T h e N A D H - 2 , 6 - d i c h l o r o p h e n o l i n d o p h e n o l d i a p h o r a s e activity can b e m o n i t o r e d b y t h e d i s a p p e a r a n c e of 600 n m a b s o r b a n c e , i n d i c a t i v e of t h e r e d u c t i o n o f t h e d y e , a n d has b e e n p r e v i o u s l y d e s c r i b e d . 13 T h e d e t a i l s o f this a s s a y a r e as follows. T h e final v o l u m e o f t h e r e a c t i o n is 3 ml, a n d t h e a s s a y c o n t a i n s 50 m M s o d i u m p h o s p h a t e , p H 7.2, 2 m g b o v i n e s e r u m a l b u m i n , i m M E D T A , 4 0 / x M 2 , 6 - d i c h l o r o p h e n o l i n d o p h e n o l , a n d 0.2 m M N A D H . T h e a s s a y is i n i t i a t e d b y t h e a d d i t i o n o f t h e e n z y m e p r e p a r a t i o n a n d t h e c h a n g e in a b s o r b a n c e at 600 n m f o l l o w e d at 25 °. A n e x t i n c t i o n coefficient for t h e o x i d i z e d d y e o f 21 x 103 M -I c m -I is u s e d for t h e calculations of enzyme activityY 23j. F. Kalse and C. Veeger, Biochim. Biophys. Acta 159, 244 (1968). 24j. Visser and C. Veeger, Biochim. Biophys. Acta 159, 265 (1968). 25E. P. Steyn-Parve and H. Beinert, Z Biol. Chem. 233, 843 (1958).

[201

DIHYDROLIPOAMIDE DEHYDROGENASE

193

P r o p e r t i e s of Purified E n z y m e The purified D H L i p D H has been extensively characterized from several sources. 1 The properties of all the enzymes are somewhat similar, with an enzyme subunit molecular mass of approximately 50,000 daltons. The oxidized enzyme shows the characteristic absorption spectrum of free F A D with pronounced shoulders at 480 and 430 nm because F A D is associated with the protein. The ratio of the peak at 273 nm to that at 455 nm, which is an excellent indicator of the purity of the preparation, is 5.3.~ The catalytic mechanism of all the enzymes are similar in that they follow a Ping-Pong mechanism. The reaction, as mentioned before, can be divided into two half-reactions (i) the reductive half-reaction where the oxidized enzyme is reduced to the two-electron form and (ii) the oxidative half-reaction where the two electrons are passed to the nicotinamide cofactor, N A D +. The rates of the individual half-reactions can be measured in a stopped-flow spectrophotometer by following absorbance changes at 530 nm (the characteristic absorbance of the two-electron form with a charge transfer between the thiolate anion and the F A D cofactor). 6,7 Dihydrolipoamidedehydrogenase is extremely stable to a wide variety of denaturants including urea. It is, however, very sensitive to high concentrations of guanidinium hydrochloride, as the presence of that denaturant leads to the dissociation of the F A D cofactor from the enzyme. It is also remarkably resistant to the action of several proteases owing to the compactly folded globular structure. 26 The enzyme is extremely heat stable (in fact, heat treatment has been used as a purification step), being stable for at least 5 min at 75°. 15 The optimum p H for D H L i p D H activity is p H 8 for the physiological reaction. 13 The enzyme is sensitive to arsenite, copper, and cadmium (after reduction with N A D H or dihydrolipoamide) 1,27 and to copper without prior reduction. 2s Trivalent arsenicals have been used extensively as probes for the presence of vicinal thiol groups (formed as a result of reduction of the enzyme with N A D H or dihydrolipoamide). ~,9 The enzyme is also sensitive to other reagents that modify cysteines in proteins (e.g., iodoacetamide) because of the indispensible role played by the redox disulfide in catalysis. 1 Finally, D H L i p D H has been characterized extensively, and the steady-state kinetic parameters have been determined (Table II). The primary structure of D H L i p D H has been deduced from the nucleotide sequence of the cloned genes from a variety of different sources includ-

26R. G, Mathews, L. D. Arscott, and C. H. Williams,Jr., Biochim. Biophys. Acta 370,26 (1974). 2~S. Ide, T. Hayakawa, K. Okabe, and M. Koike, J. Biol. Chem. 242, 54 (1967). 28C. Thorpe and C. H, Williams, Jr., Biochemistry 14, 2419 (1975).

194

SIGNAL TRANSDUCTION AND GENE REGULATION

[20]

TABLE II STEADY-STATEKINE37CPARAMETERSOF DIHYDROL1POAMIDEDEHYDROGENASEFROM DIFFERENT SOURCES

Source

Parameter

Value

Ref.

Azotobacter vinelandii

Km (dihydrolipoamide) Km (NAD +) Turnover number~

b

Rat liver

Km (dihydrolipoamide)

390/xM 325/xM 420 sec-1 490/xM 520 heM 345 sec-1 300/zM 200/.~M 550 sec-~ 690/xM 320 gM 421 sec-1

Pig heart

Human (recombinant)

Km (NAD +) Turnover number Km (dihydrolipoamide) Krn (NAD ÷) Turnover number K~ (dihydrolipoamide) Km (NAD ÷) Turnover number

1

1

15

a The turnover numbers were obtained under conditions of maximum velocity. b j. Benen, W. V. Berkel, N. Dieteren, D. Arscott, C. H. Williams, Jr., C. Veeger, and A. DeKok, Eur. Z Biochem. 207, 487 (1992).

ing t h e e u b a c t e r i a E. coli 29 a n d A z o t o b a c t e r vinelandii3°; the a r c h a e b a c t e r i u m Haloferax volcanii31; a n d f r o m m a m m a l s (e.g., p i g h e a r t , 3z human33). T h e r e a r e e x t e n s i v e s e q u e n c e similarities a m o n g t h e e n z y m e s , 1'3 e s p e c i a l l y in t h e n u c l e o t i d e c o f a c t o r b i n d i n g r e g i o n s a n d in t h e active site region. T h e h i g h - r e s o l u t i o n t h r e e - d i m e n s i o n a l s t r u c t u r e s of t h e e n z y m e f r o m A z o t o b a c ter vinelandii 34 a n d P s e u d o m o n a s putida 35 a r e a v a i l a b l e , a n d , again, t h e r e a r e e x t e n s i v e s i m i l a r i t i e s in t h e o v e r a l l t e r t i a r y s t r u c t u r e of t h e e n z y m e f r o m t h e s e sources.

Summary We have described the most commonly used assay procedures determin a t i o n o f t h e D H L i p D H activities f r o m p r o k a r y o t i c a n d e u k a r y o t i c cells. W e h a v e also d e s c r i b e d t h e p r o c e d u r e s for t h e p r e p a r a t i o n o f tissue e x t r a c t s 29p. E. Stephens, H. M. Lewis, M. G. Darlison, and J. R. Guest, Eur. J. Biochem. 135, 519 (1983). 3oA. H. Westphal and A. DeKok, Eur. J. Biochem. 172, 299 (1988). 31 N. N. Vettakkorumakankav and K. J. Stevenson, Biochem, Cell Biol. 70, 656 (1992). 32 G. Otulakowski and B, H. Robinson, J. Biol. Chem. 262, 17313 (1987). 33G. Pons, C. Raefsky-Estrin, D. J. Carothers, R. A. Pepin, A. A. Javed, B. W. Jesse, M. K. Ganapathi, D. Samols, and M. S. Patel, Proc. Natl. Acad. Sci. U.S.A. 85, 1422 (1988). 34A. Mattevi, A. J. Schierbeek, and W. G. J. Hol, J. Mol. Biol. 220, 975 (1991). 3s A. Mattevi, G. Obmolova, J. R. Sokatch, C. Betzel, and W. G. J. Hol, Proteins; Struct. Funct. Genet. 13, 336 (1992).

[201

DIHYDROLIPOAMIDE DEHYDROGENASE

195

to determine the enzymatic activity. We have briefly alluded to the problems inherent in these assay procedures and suggested some precautions for using the assays to determine D H L i p D H activity. W e have also included a statement about the purification procedures and described some of the properties of this enzyme, which might help the reader in planning studies. We have, by no means, a t t e m p t e d to be comprehensive in this regard; therefore, the reader should consult the appropriate references (and references therein). Acknowledgments The authors express gratitude to Prof. Charles H. Williams, Jr., University of Michigan, and Dr. Gail Willsky of the Department of Biochemistry, State University of New York at Buffalo, for critical reading of the manuscript and for helpful suggestions. Work described in this chapter is supported in part by U.S. Public Health Service Grant DK42885.

[2 1]

THIOREDOXIN

AND THIOREDOXIN

REDUCTASE

199

[21] T h i o r e d o x i n a n d T h i o r e d o x i n R e d u c t a s e By ARNE HOLMGREN and MIKAEL BJORNSTEDT Introduction Thioredoxin (Trx) is a small (Mr 12,000) multifunctional and ubiquitous protein characterized by having a redox-active disulfide/dithiol within the conserved active site sequence: -Trp-Cys-Gly-Pro-Cys-. 1-4 Oxidized thioredoxin (Trx-S2) has a disulfide, and reduced thioredoxin [Trx-(SH)2] has a dithiol. This chapter summarizes current methods to determine Trx and thioredoxin reductase (TR). Elsewhere in this volume the genetics of Trx is summarized by Russel 5 and structural studies of Trx in solution by nuclear magnetic resonance ( N M R ) are described by Dyson. 6 A previous volume in this series contains a chapter by H o l m g r e n 7 on the use of N A D P H , TR, and Trx for specific reduction of disulfide bonds in proteins. Buchanan and co-workers s review aspects of the control of photosynthetic enzymes by specific chloroplast thioredoxins and the enzyme ferredoxin-thioredoxin reductase. Thioredoxin reductase specifically reduces Trx-S2 to Trx-(SH)2 using N A D P H 1-3 [reaction (1)]. The Trx-(SH)2 f o r m is a powerful protein disulfide reductase [reaction (2)]. Thus, Trx, TR, and N A D P H , collectively called the thioredoxin system, operate as a powerful N A D P H - d e p e n d e n t protein disulfide reductase system. Note that the redox potential (E~ value) TR

Trx-S2 + N A D P H + H + .

-' Trx-(SH)2 + N A D P *

(1)

spontaneous

Trx-(SH)2 + protein-S2 ~

Trx-S2 + protein-(SH)2

(2)

of H + + N A D P H / N A D P + at p H 7.0 and 25 ° is - 3 1 5 m V and that the E{~ of E s c h e r i c h i a coli Trx-(SH)z/Trx-S2 is - 2 7 0 inV. 1-3 Reaction (1), which I A. Holmgren, Annu. Rev. Biochem. 54, 237 (1985). 2 A. Holmgren, J. Biol. Chem. 264, 13963 (1989). 3 F. K. Gleason and A. Holmgren, FEMS Microbiol. Rev. 54, 271 (1988). 4 H. Eklung, F. K. Gleason, and A. Holmgren, Proteins 2, 13 (1991). 5 M. Russel, this volume [27]. 6 H. J. Dyson, this volume [30]. 7 A. Holmgren, this series, Vol. 107, p. 295. s j. H. Wong, K. Kobrehel, and B. B. Buchanan, this volume [24].

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, Inc All rights of reproduction in any form reserved.

200

THIOREDOXINAND GLUTAREDOXIN

[2 1]

is fully reversible, may thus be used to determine the E6 value of a particular Trx. For example, in Escherichia coli Trx the mutation of Pro-34 to His changed the E~ value of Trx from -270 to -235 i n V . 9 Thioredoxin has been isolated and characterized from a wide variety of prokaryotic and eukaryotic species. 1-4 Mammalian thioredoxins show about 25% sequence identity to the well-characterized E. coli protein with 108 residues. GS-SeH + GSH + NADP +

(2) GS-Se- + H20 ---> Se ° + GSH + OH-

(3)

GS-SeH + NADPH + H20

T~or6R H2Se + GSH + NADP + + OH(a) GS-Se- + GSH ---> HSe- + GSSG (5) HSe- + (O) ---> Se ° + OH(6) HSe- + 2GSH + (O) ---> HSe- + GSSG + H20 (7) SH S

/

HSe- + R

\SH

+ (O) --~ HSe- +

J

+ H20

(8)

kS

R is the active site of thioredoxin or mammalian thioredoxin reductase. Preparation a n d Purification of Selenodiglutathione Mix a 10-/~l aliquot of 0.i M selenite (dissolve Na2SeO3, SeO2, or H2SeO3 in water) with 60/zl of 0.1 M HCI and add 40/zl of freshly prepared 0.I M GSH. The order in which the different components are mixed is 12 S. Kumar, M. Bj0rnstedt, and A. Holmgren, Eur. J. Biochem. 207, 435 (1992). 13X. Ren, M. Bj6rnstedt, B. Shen, M. L. Ericson, and A. Holmgren, Biochemistry 32, 9701 (1993).

[22]

211

SELENITE AND SELENODIGLUTATHIONE

E c

c'q

GS-Se-SG

GSSG

8 1

c t~ zc/

o ,


E "-t

E

"~

E

100

80 60 40 20

B ~) ¢.)

100

I

,

I

,

I

,

I

--

(

'

I

'

¢-

¢1

o

m,7.9 = - 3 4 6 m v

80

t~

0 -,f

60

E E

40

E

20 0 -450

I

•400

,

I

-350

J

I

-300

,

-250

E h (mv) FIG. 1. Equilibrium redox titrations of thiol/disulfide regulatory groups on chloroplast FBPase. Purified FBPase was incubated for 1.5 hr (A) or 3 hr (B) at pH 7.9 in the presence of thioredoxin f a n d 10 mM DTT in various thiol/disulfide ratios. (A) Dependence of FBPase activity, measured as phosphate release from fructose 1,6-bisphosphate, on the ambient redox potential (Eh). (B) Level of fluorescence emission at 485 nm after reacting FBPase for 10 rain with 10 mM monobromobimane. The solid lines are the theoretical curves determined from the nonlinear least-squares fit of the data to the Nernst equation with the value of n fixed at 2. The calculated midpoint potential of the thiol/disulfide regulatory group of FPBase (Era,7.9) for the two titration methods agree closely.

[231

DETERMINING SULFHYDRYL MIDPOINT POTENTIALS

225

Stability of Ambient Redox Potential. It is important that the thiol/ disulfide ratio of the redox buffer not change over the course of the experiment. Oxygen is the major threat to the stability of the ambient redox poise in the relatively reducing potential range in which the titrations are performed even though DTT red is not very susceptible to direct oxidation by 02.1° The FBPase (7 /zg) is incubated at 25° in 0.1 ml of solution containing 100 mM Tricine-NaOH (pH 7.9), 0.2 tzg thioredoxin f, and 10 mM Dq-T ~1 in various thiol/disulfide ratios. The incubation solution has been purged with argon, and samples are kept under a constant stream of humidified argon throughout the incubations. The dipolar ionic buffer Tricine is an effective chelator of trace metals that might otherwise catalyze the oxidation of thiols by oxygen. The stability of the redox poise under these conditions for incubations of 5 hr is confirmed using Ellman's reagent 12 to monitor the concentration of DTT tea. However, it should be noted that in titrations of other enzymes, such as phosphoribulokinase, in which the thiol/disulfide midpoint potential is more oxidizing than FBPase, we find that 10 mM DTT is not adequate to maintain a satisfactorily stable poise throughout the full range of the titration. For those values far below the maximum buffering range of the DTT redox couple, where the concentration of the DTT thiol is very low, we use 80 mM total DTT to ensure a stable redox poise at the most oxidizing potentialsJ 3 Establishing that Equilibrium Has Been Achieved. Although the thioldisulfide exchange rates between thioredoxin and the target enzymes must be relatively rapid under physiological conditions in order to explain observed rates of light activation, this situation cannot be assumed for in vitro titrations. Although the pH sensitivity of thiol/disulfide exchange is well known, more cryptic factors associated with enzyme substrates and cofactors may play important roles in determining the equilibration kinetics. Changes in the protein which affect the local electrostatic environment around the regulatory thiol/disulfide or changes in the physical accessibility due to rearrangements within the protein could have large and unforeseen effects on thiol/disulfide exchange rate. In our experiments, we monitor the time course of enzyme activation in order to ensure that redox equilibrium has been attained. Under the conditions cited for the FBPase titration reported in Fig. 1A, maximum activity is achieved within 1.5 hr, and we observe no loss of activity after 3 hr of incubation at 25°.13 Without thioredoxin f, the rate of activation m W. W. Cleland, Biochemistry 3, 480 (1964). 1~Ultrapure DTT red is purchased from Boehringer Mannheim (Indianapolis, IN) and DTT °x from Sigma (St. Louis, MO), Both are used without further purification. ~2G. L. Ellman, Arch. Biochem. Biophys. 82, 70 (1959). 13 R. S. Hutchison, Ph.D. Thesis, University of Illinois, Urbana/Champaign, 1994.

226

THIOREDOXIN AND GLUTAREDOXIN

[231

of FBPase by direct thiol exchange with DTT is exceedingly slow, and equilibrium is not attained during the lifetime of the protein under these incubation conditions. We find13 that the effect of thioredoxin f on the rate of FBPase activation is saturated at about 10/zg/ml thioredoxin f, which differs from the unsaturable, first-order dependence of FBPase reduction rate on thioredoxin f concentration observed by Clancey and Gilbert. 14 In the presence of thioredoxin f, we observe no dependence on the rate of FBPase activation on the total concentration of DTT present during the incubation.

Redox Titration of Fructose-l,6-Bisphosphatase by Monobromobimane Labeling of Protein Thiols The thioredoxin f-dependent reduction of the regulatory disulfide on FBPase can be measured using the thiol-labeling reagent monobromobimane (mBBr) that reacts rapidly and exclusively with accessible thiol groups, producing a fluorescent, chemically modified protein. 15 Figure 1B shows an equilibrium redox titration of FBPase using mBBr to quantitate the amount of thiol present at different ambient redox potentials. Two hundred micrograms FBPase, 0.4 txg of thioredoxin f, and 10 mM DTT in various thiol/disulfide ratios are incubated for 3 hr in 0.1 ml of 100 mM Tricine-NaOH buffer (pH 7.9) as described above. Monobromobimane is added to 10 mM (from a 250 mM stock solution in acetonitrile). After 10 rain, unreacted mBBr is extracted with 0.1 ml of methylene chloride, leaving the protein-mBBr complex in the aqueous phase. The fluorescence emission intensity of each sample is measured in a SLM Aminco SPF-500 spectrofluorometer (Urbana, IL) at 485 nm with 395 nm excitation light. Fluorescence emission originating from mBBr labeling of protein thiols not involved in the thioredoxin-mediated disulfide/thiol interconversion can be estimated from a sample incubated as above with 10 mM DTT °x. This background fluorescence level is less than 10% of the fluorescence emission from samples incubated in 10 mM DTT red. It should be noted that the fluorescence originating from mBBr labeling of nonregulatory protein thiols may not be as low for all other thioredoxin-regulated chloroplast enzymes as we observed for FBPase. Were an enzyme to have accessible nonregulatory thiol groups of similar redox potential to the regulatory thiols, significant background fluorescence would be anticipated. Samples incubated with 10 mM D T r red were considered to be fully reduced, and the fluorescence intensity of such samples was assigned the value of 100% in Fig. lB. 14 C. J. Clancey and H. R. Gilbert, J. Biol. Chem. 262, 13545 (1987). x5 N. S. Kosower and E. M. Kosower, this series, Vol. 143, p. 76.

[23]

DETERMINING SULFHYDRYL MIDPOINT POTENTIALS

227

There is one additional concern that should be addressed for techniques such as this, in which the measurement technique perturbs the equilibrium. By reacting exclusively and irreversibly with the thiol, mBBr could "pull" the reaction in the direction of thiol formation and thereby produce erroneous results. This is not likely to be a problem in the in vitro titration of chloroplast enzymes as performed here, however, because conditions can be set to ensure that the reaction of mBBr with protein thiols is much more rapid than the thioredoxin-mediated reduction of the protein disulfide. Under the conditions used here for FBPase, the half-time for mBBr labeling is less than 1 min, whereas the half-time for the reductive activation for FBPase is more than 30 min. 13

Analysis of Redox Data Curve Fitting. The data in Fig. 1 were fit to the Nernst equation in the form: 59 [disulfide] Eh --- Em,x + ~'~ log [dithiol] where Eh denotes that the redox potential values are referenced to the standard hydrogen half-cell. Em~ is the symbol for the midpoint potential under conditions when the components of the redox reaction are present at less than unit activities and the ambient pH value (x) for the determination is greater than zero16; in this case the pH was 7.9. There are 2 electrons involved in the thioredoxin-mediated thiol/disulfide interconversion for FBPase, so we have evaluated the equation for n = 2. In Fig. 1A the [disulfide] was calculated as 100% minus the percent maximum FBPase activity at any given Eh, and the [dithiol] was then the percent maximum FBPase activity at that Eh. In Fig. 1B, the [disulfide]/[dithiol] ratio was calculated similarly with the percent maximum fluorescence intensity for mBBr-labeled FBPase replacing FPBase activity in the calculation. The data points were computer fit to the Nernst equation using an iterative nonlinear least-squares analysis. We used Fig P, which is a versatile commercial data analysis package (Biosoft, Cambridge, UK) replete with convenient plotting and graph drawing capabilities. The solid lines in Fig. 1 are the theoretical curves determined from the nonlinear least-squares fit to the Nernst equation with the value of n fixed at 2.

16p. L. Dutton, this series, Vol. 54, p. 411.

228

[241

THIOREDOXIN AND GLUTAREDOXIN

Determination o f Eh. The absolute values determined for the midpoint potential of the regulatory sulfhydryls on FBPase in Fig. 1 are dependent on the Eh values assigned to various ratios of DTTrea/DTT°x at pH 7.9. The initial value for the midpoint potential of DTT at pH 7 reported by Cleland 1° of -332 mV is well supported by measurements employing diverse techniques. 17,18Because the redox reaction of DTT involves protons, the Em of the redox buffer will have a substantial pH dependence which must be taken into account in Eh calculations. The reduced form of DTT is a very weak acid, with the pKa values of the thiols reported to be 9.2 for the first ionization and 10.1 for the secondJ 9 Because DTT °x is a very strong acid, the redox chemistry of DTT will involve one proton per electron at pH values below approximately 8.2 (at higher pH values ionization of the first thiol o n D T T red will lower the proton to electron ratio below 1). Thus for pH values below 8.2, the Em of DTT can be calculated according to Em~ = Em~7 - 0.059 (pH x - pH 7)

We have used the value of Lees and Whitesides is for the Era, 7 for DTI" of -327 mV at 25 °, giving a value for Era,7. 9 of --380 mV. It should be noted that Cleland's original value for Em,8.1 of -366 mV 1° (implying a - 3 0 mV per unit pH dependence) has not been substantiated by subsequent measurements and should not be used for calculations of Eh.

17 D. M. R o t h w a r f and H. A. Scheraga, Proc. Natl. Acad. Sci. U.S.A. 89, 7944 (1992). is W. J. Lees and G. M. Whitesides, J. Org. Chem. 58, 642 (1993). 19 G. M. Whitesides, J. E. Liburn, and R. P. Szajewski, J. Org. Chem. 42, 332 (1977).

[24] By

Thioredoxin

and Seed Proteins

JOSHUA H. WONG, KAROLY KOBREHEL,

and BOB B.

BUCHANAN

Introduction Storage proteins, which are formed and stored within seeds, serve as the main source of nitrogen and as a primary source of carbon for the germination and growth of seedlings. Storage proteins are synthesized after pollination and accumulate during seed maturation. The proteins are mobilized, proteolytically degraded, and utilized when conditions are favorable METHODS1NENZYMOLOGY,VOL. 2.52

Copyright© 1995by AcademicPress,Inc. All rightsof reproductionin any formreserved.

[24]

THIOREDOXIN AND SEED PROTEINS

229

for germinationJ '2 In the case of cereals, the bulk of the storage proteins are water insoluble after the grains mature. In most other types of seeds, the proteins remain soluble and enclosed within a membrane, thereby creating a structure known as a protein body. 3 In some cereals, the protein body m e m b r a n e is disrupted during maturation and drying of the grain, thereby exposing the storage proteins to endogenous proteases of the endosperm. 3,n However, in typically nonglutenous cereals such as corn, rice, sorghum, and millet, storage proteins remain in protein bodies in the mature grains (see references in Ref. 5). Disulfide bonds are one of the main stabilizing forces in storage proteins. The formation of disulfide bonds appears to function in cereals to provide increased structural stability on the one hand and decreased solubility on the other. Both features provide protection against proteolysis. The group includes typically large, insoluble storage proteins (60-100 kDa) as well as generally less abundant low molecular weight soluble proteins (5-30 kDa),4 rich in disulfide bonds, with four or more cystines per molecule. Examples include a wide variety of exogenous a-amylase and protease inhibitors. Except for the ot-amylase/subtilisin inhibitor of barley and other cereals, which inhibits endogenous a-amylase, 6,7 the physiological function of these small disulfide proteins within the seed remains a mystery. It has long been considered that proteins of this type play a bioprotection role by blocking digestive enzymes of invading pests. 8 Several investigators have shown that small disulfide proteins of seeds, like the larger insoluble storage proteins, are degraded during germination? a° Our laboratory has uncovered evidence that both protein groups undergo reduction in conjunction with the degradationJ ~ The NADP+/ thioredoxin system (NTS) appears to play a role in the reduction of critical disulfide groups of seed proteins (S-S ~ 2SH) and, thereby, as seen below, trigger germination. The thioredoxin is reduced by N A D P H via NADP+/ J A. M. Mayer and A. Poljakoff-Mayber, "The Germination of Seeds." 4th Ed,, p. 160. Pergamon, New York, 1989. 2 F. M. Ashton, Annu. Rev. Plant Physiol. 27, 95 (1976). 3 E. Weber and D. Neumann, Biochem. Physiol. Pflanz. 175, 279 (1980). 4 p. R. Shewry and B. J. Miflin, Adv. Cereal Sci. Technol. 7, 1 (1985). s D. B. Bechtel, R. L. Gaines, and Y. Pomeranz, Cereal Chem. 59, 336 (1982). 6j. Hejgaard, I. B. Svendsen, and J. Mundy, Carlsberg Res. Commun. 48, 91 (1983). 7 R. J. Weselake, A. W. MacGregor, R. D. Hill, and H. W. Duckworth, Plant Physiol. 73, 1008 (1983). s C. A. Ryan, Biochem. Plants 6, 351 (1981). 9 p. Hwang and W. Bushuk, Cereal Chem. 50, 147 (1973). 10j. E. Kruger and B. A. Marchylo, Cereal Chem. 62, 1 (1985). ~K. Kobrehel, J. H. Wong, A. Balogh, F. Kiss, B. C. Yee, and B. B. Buchanan, Plant Physiol. 99, 919 (1992).

230

THIOREDOXINAND GLUTAREDOXIN

[241

thioredoxin reductase (NTR), a flavoprotein [Eq. (1)]. The reduced thioredoxin, in turn, reduces disulfide groups on target proteins [Eq. (2)]. The evidence suggests that thioredoxin preferentially reduces intramolecular disulfide bonds relative to intermolecular counterpartsJ 2 Thioredoxin hox + N A D P H + H + -NTR thioredoxin hred + N A D P + (1) (--S--S--) (--SH HS--) Thioredoxin hred + proteinox --~ thioredoxin hox + proteinr~d (--SH HS--) (--S--S--) (--S--S--) (--SH HS--)

(2)

Thioredoxin h, a 12-kDa protein with a catalytically active disulfide group, seems to be ubiquitous in plant cells and to be localized in the cytosol, endoplasmic reticulum, and mitochondriaJ 3-15 At the onset of germination, thioredoxin h appears to be reduced by N A D P H generated by the oxidative pentose phosphate pathway of the cereal endospermJ 1 Once reduced, the thioredoxin, in turn, reduces the large storage proteins (leading to their solubilization), as well as the small disulfide inhibitor proteins (leading to loss of their ability to inhibit target enzymes). On reduction, both the large storage proteins 11A6,17 and the small disulfide enzyme inhibitor proteins ~8,19 show increased susceptibility to proteolysis. There is some evidence t h a t glutathione maintained in reduced form (GSH) by N A D P H and glutathione reductase may also function in the reduction of these proteins [Eqs. (3) and (4)], particularly after thioredoxin has initiated the reduction process. 11 GSSG + N A D P H + H +~ 2GSH + NADP ÷ (--S--S--) (--SH)

(3)

2 G S H + proteinox --~ GSSG + proteinred (--SH) (--S--S--) (--S--S--) (--SHHS--)

(4)

Much of the evidence for the role of thiols in seed germination has come from studies in our laboratory. For these studies, we have adapted a relatively simple, yet specific method for measurement of the thiol redox 12S. Shin, J. H. Wong, K. Kobrehel, and B. B. Buchanan, Planta 189, 557 (1993). 13T. C. Johnson, K. Wada, B. B. Buchanan, and A. Holmgren, Plant Physiol. 85, 446 (1987). 14j. Bodenstein-Lang, A. Buch, and H. Follmann, FEBS Lett. 258, 22 (1989). 15F. Marcus, S. H. Chamberlain, C. Chu, F. R. Masiarz, S. Shin, B. C. Yee, and B. B. Buchanan, Arch. Biochem. Biophys. 287, 195 (1991). 16j. H. Wong, K. Kobrehel, C. Nimbona, B. C. Yee, A. Balogh, F. Kiss, and B. B. Buchanan, Cereal Chem. 70, 113 (1993). 17I. Besse, unpublished findings from our laboratory, 1995. ~8j. Jiao, B. C. Yee, K. Kobrehel, and B. B. Buchanan, J. Agric. Food Chem. 40, 2333 (1992). 19j. Jiao, B. C. Yee, J. H. Wong, K. Kobrehel, and B. B. Buchanan, Plant Physiol. Biochem. 31, 799 (1993).

[24]

THIOREDOXIN AND SEED PROTEINS

231

status of proteins. The technique is a modification of an earlier procedure developed for analysis of chloroplast enzymes regulated by thioredoxin. 2° In that study, the chloroplast proteins of interest, because of their low abundance, were immunoprecipitated to permit quantitation of the reduction of the disulfide groups by light. An uncharged probe, m o n o b r o m o b i m a n e (mBBr), was used to bind sulfhydryl groups specifically and covalently, thereby rendering the p r o t e i n m B B r derivative fluorescent. 2°'21 The fluorescent m B B r - p r o t e i n products can be separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( S D S - P A G E ) and then visualized and quantified. The method can be applied equally well in vitro (both with purified proteins ls'19`2L22 and with extracts composed of many proteins 11,16)or in vivo (with intact seedsU). In either case, the redox status of the seed proteins can be assessed. A description of the procedure is presented below, using wheat as an example. Also, as seen below, the procedure can be applied equally well to disulfide proteins purified from a variety of sources.

S o u r c e s of Materials a n d C h e m i c a l s Materials. Either seeds or semolina of durum wheat (Triticum durum) or seeds or flour of bread wheat (Triticum aestivum) may be used. In our case, the respective cultivars are Desf. cv. M o n r o e or cv. Scout 66. Escherichia coli thioredoxin and N T R are either purchased from American Diagnostica (Greenwich, CT) or isolated from cells overexpressing each protein. 23'24 Wheat thioredoxin h and N T R are isolated from germ, following the procedures developed for spinach leaves. 25 Glutaredoxin (a gift of Prof. A. Holmgren, Karolinska Institutet, Stockholm) is isolated from E. coli cells genetically modified to overexpress the protein. 26 Chemicals. Reagents for S D S - P A G E are purchased from Bio-Rad Laboratories (Richmond, CA). M o n o b r o m o b i m a n e (mBBr) or Thiolyte is obtained from Calbiochem (San Diego, CA). Aluminum lactate and methyl green are products of Fluka Chemicals (Buchs, Switzerland).

20N. A. Crawford, M. Droux, N. S. Kosower, and B. B. Buchanan, Arch. Biochem. Biophys. 271, 223 (1989). 21N. S. Kosower and E. M. Kosower, this series, Vol. 143, p. 76. 22K. Kobrehel, B. C. Yee, and B. B. Buchanan, J. Biol. Chem. 266, 16135 (1991). ~3F. de la Motte-Guery, M. Mijiulac-Maslow, P. Decottiguies, M. Stein, P. Minard, and J. P. Jacquot, Eur. J. Biochem. 196, 287 (1991). 24M. Russel and P. Model, J. Biol. Chem. 263, 9015 (1988). 25F. J. Florencio, B. C. Yee, T. C. Johnson, and B. B. Buchanan, Arch. Biochem. Biophys. 266, 496 (1988). z6A. Holmgren, J. Biol. Chem. 264, 13963 (1989).

232

THIOREDOXINAND GLUTAREDOXIN

[241

Isolation of Protein Fractions Chemicals and Materials

Durum wheat semolina or bread wheat flour, 5 g Tris-HC1 buffer, pH 7.5, 50 mM Ethanol, 70% (v/v) Acetic acid, 0.1 M Procedure. For isolation of the soluble and insoluble proteins, semolina or flour (0.2 g) is extracted sequentially with 1 ml of the following solutions for the indicated times at 25°: (1) 50 mM Tris-HC1 buffer, pH 7.5 (20 min), for the albumins and globulins; (2) 70% ethanol (2 hr) for the gliadins; and (3) 0.1 M acetic acid (2 hr) for the glutenins. 1° Alternatively, the glutenins can be isolated by extraction with sodium myristate. In that case, 80 mg myristate is added to 1 g of flour together with 8 ml distilled water. 27 Similar conditions can be used to extract total proteins in one step. 28During extraction, samples are placed on an electrical rotator and, in addition, occasionally agitated with a vortex mixer. After extraction with each solvent, samples are centrifuged (12,000 rpm for 5 min at 4°) in an Eppendorf microcentrifuge, and supernatant fractions are saved for analysis. In between each extraction, pellets are washed with 1 ml of water and collected by centrifugation as before; the supernatant wash fractions are discarded. By convention, z9 the fractions are designated as follows: (1) albumin/globulin; (2) gliadin; and (3) glutenin. If desired, the albumins and globulins can be separated by dialysis against distilled water; the albumins remain soluble and the globulins precipitate. Following centrifugation (10 min, 10,000 g at 4°), the supernatant fraction containing the albumins is decanted and the globulins are redissolved in 50 mM Tris-HCl buffer, pH 7.5.

Reduction and in Vitro Labeling of Proteins In Vitro Labeling o f Proteins with Monobromobimane Reagents

Tris-HC1 buffer, pH 7.9 at 20 °, 100 mM Thioredoxin, E. coli, 0.1 mg/ml in 30 mM Tris-HC1, pH 7.9 NTR, E. coli, 0.1 mg/ml in 30 mM Tris-HCl, pH 7.9 NADPH, 20 mM in 30 mM Tris-HCl, pH 7.9 27K. Kobrehel and R. Alary, J. ScL Food Agric. 48, 441 (1989). 28K. Kobrehel and B. Matignon, Cereal Chem. 57, 73 (1980). 29T. B. Osborne and C. G. Voorhees,Am. Chem. J. 15, 392 (1893).

[241

THIOREDOXIN AND SEED PROTEINS

233

Dithiothreitol (DTT), 100 mM Glutathione, reduced, 20 mM in 30 mM Tris-HC1, pH 7.9 Glutathione reductase (purified from spinach leaves25), 0.1 mg/ml in 30 mM Tris-HC1, pH 7.9 Glutaredoxin, E. coli, 0.1 mg/ml in 30 mM Tris-HC1, pH 7.9 Monobromobimane (mBBr) or Thiolyte, 20 mM dissolved in acetonitrile and stored protected from light SDS, 10% (w/v) 2-Mercaptoethanol (2-ME), 100 mM Target proteins, 5-30/zg, in extraction solvent; alternatively, as seen below, 5-10 tzg pure disulfide proteins may be used Procedure. Both steps of the reaction are carried out at room temperature. As indicated, 7 txl NTR (0.7/~g) and 10/~1 thioredoxin (1 tzg) (both from E. coli unless specified otherwise) along with 10 t~l N A D P H are added to 70 tzl of 100 mM Tris-HC1 buffer, pH 7.9, and 30 txg of target protein. When thioredoxin is reduced by DTT, the N A D P H and NTR are omitted and 5 txl DTT is added to a final concentration of 1 mM. Reduction by reduced glutathione is performed similarly, but at a final concentration of 2 mM. After incubation for 20 min, 5/zl of mBBr (80 nmol) is added, and the reaction is continued for another 15 min. To stop the reaction and derivatize excess mBBr, 10/zl of 10% SDS and 10 tzl of 100 mM 2-ME are added, and the samples are then applied to the gels. For reduction by glutaredoxin, the thioredoxin and NTR are replaced by 10 Ixl E. coli glutaredoxin (1/xg), 10/xl glutathione reductase (1.4/xg), and 10 txl NADPH.

Germination and in Vivo Labeling of Proteins Germination of Wheat Seeds Materials

Seeds of durum wheat Plastic petri dishes Whatman (Clifton, N J) No. 1 filter paper Growth chamber Procedure. Duplicate sets of 20 to 30 seeds are placed in a plastic petri dish on three layers of Whatman No. 1 filter paper moistened with 5 ml of deionized water. Petri dishes are then transferred to a darkened growth chamber. Germination is carried out for up to 4 days at room temperature, with 5 ml of deionized water being added daily to each petri dish. On the fifth day, the germinated seedlings are harvested.

234

THIOREDOXIN AND GLUTAREDOXIN

[24]

In Vivo Labeling of Proteins with Monobromobimane Chemicals and Materials Endosperm of germinated seedlings Liquid N2 mBBr, 2.0 mM, freshly diluted in 100 mM Tris-HC1 buffer, pH 7.9 Procedure. At the indicated times, the dry seeds or germinated seedlings (selected on the basis of similar radicle or shoot length) are taken from the petri dish and the embryos or germinated axes are removed with a razor blade. Five endosperm from each lot are weighed and then ground in liquid N2 with a mortar and pestle. One milliliter of 2.0 mM mBBr in 100 mM Tris-HCl buffer, pH 7.9, is added just as the last trace of liquid N2 disappears. The thawed mixture is ground for another minute and transferred to a microcentrifuge tube of 1.5 ml capacity. The volume of the suspension is adjusted to 1 ml with the appropriate mBBr or buffer solution. Protein fractions of albumin/globulin, gliadin, and glutenin from endosperm of germinated seedlings are extracted as described above for semolina and flour. The extracted protein fractions are stored at - 2 0 ° until use for electrophoresis. A buffer control (minus mBBr) is included for each time point. Analysis of Labeled Samples by Polyacrylamide Gel Electrophoresis

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis of Monobromobimane-Labeled Samples Analysis by S D S - P A G E of the mBBr-derivatized samples of each protein fraction is performed in 15% polyacrylamide gels at pH 8.5 as described by Laemmli. 3° Before sample application, an aliquot (20/xl) of 50% (v/v) glycerol containing 0.01% (w/v) bromphenol blue is added to each mBBr-labeled sample. Gels of 1.5 mm thickness are developed for 16 hr at a constant current of 9 mA.

Native Gel Electrophoresis for Separating Different Gliadins Reagents and Equipment A gel solution in 100 ml final volume contains the following: 6.0 g acrylamide, 0.3 g bisacrylamide, 0.024 g ascorbic acid, 0.2 mg ferrous sulfate heptahydrate, 0.25 g aluminum lactate, and 85% lactic acid (certified ACS); the gel solution is adjusted to pH 3.1 with lactic acid 30 U. K. Laemrnli, Nature (London) 227, 680 (1970).

[24]

THIOREDOXINAND SEEDPROTEINS

235

Catalyst solution, hydrogen peroxide, 3% Tank buffer, aluminum lactate buffer, 0.5 g/liter, is adjusted to pH 3.1 with lactic acid Tracking dye, methyl green, 0.6% (w/v) in 80% (w/v) sucrose in tank buffer Electrophoresis equipment, Hoefer Model SE600 (Hoefer Scientific Instrument, San Francisco, CA) Procedure. To resolve the different types of gliadins, native polyacrylamide gel electrophoresis is performed in 6% gels of 3 mm thickness (a procedure designed to separate gliadins into the four types, a,/3, y, and ~0) as described by Bushuk and ZiUman 31 and modified for vertical slab gels by Sapirstein and Bushuk. 32 A gel stock solution can be prepared beforehand by mixing acrylamide, bisacrylamide, and aluminum lactate, adjusting to pH 3.1 with lactic acid, and then storing at 4 °. Before use the ascorbic acid (0.024 g) and ferrous sulfate (0.2 rag) are added to 100 ml of the gel stock solution, which is then degassed for 2 hr at 4 °. Because polymerization occurs essentially instantaneously after adding the hydrogen peroxide catalyst, it is best to set up and precool the gel plate (16 × 18 cm) assembly, with the 3-mm spacers, in a cold room. It is also best to pour the gel solution as quickly as possible through a funnel, mounted on a stand, that fits into the space between the plates with the comb already in place. Before sample application an aliquot (20 /xl) of the tracking dye is added to each sample. The duration of electrophoresis is approximately 4 hr, with a constant current of 50 mA. The temperature of the tank buffer ( - 2 - 3 liters) is controlled by circulating water at 21 ° during electrophoresis. (Caution: As gliadins are positively charged at this pH, it is imperative to reverse the polarity of the two electrodes, i.e., by attaching the negative electrode of the gel box to the positive outlet of the power supply and vice versa for the positive electrode.) Electrophoresis is terminated when the lower mobility band of the methyl green tracking dye migrates to about ! cm from the end of the gel. Monobromobimane Removal and Fluorescence Photography Reagents and Equipment

12% (w/v) Trichloroacetic acid 40% Methanol/10% acetic acid (v/v)

31W. Bushuk and R. R. Zillman, Can. J. Plant. Sci. 58, 505 (1978). 32H. D. Sapirstein and W. Bushuk, Cereal Chem. 62, 372 (1985).

236

THIOREDOXIN AND GLUTAREDOXIN

[24]

Spectroline transilluminator, (Spectronics Corp., Westbury, NY), Model TS-365, 365 nm ultraviolet Polaroid Land camera Polaroid Positive/Negative Landfilm, type 55 Wratten gelatin filter No. 8 (cutoff of 460 nm) Procedure. Following electrophoresis, gels are placed in 200 ml trichloroacetic acid and soaked for 4 to 6 hr with one change of solution to fix the proteins; gels are then transferred to 200 ml of 40% methanol/10% acetic acid for 8 to 10 hr (with two to three changes of solution) to remove excess mBBr. The fluorescence of mBBr, both free and protein bound, is visualized by placing gels on a light box fitted with an ultraviolet light source (365 nm). Gels are photographed with Polaroid Positive/Negative Landfilm, type 55, through a yellow Wratten gelatin filter No. 8 (cutoff of 460 nm) (exposure time ranges from 25 to 60 sec at f4.5). The negatives are used for densitometric scanning after being processed and dried (see below).

Protein Staining/Destaining/Photography Reagents Coomassie Brilliant Blue R-250, 0.25% (w/v) 40% methanol/10% acetic acid (v/v) Coomassie Brilliant Blue R-250 (0.1 g dissolved in 10 ml of 95%, v/v, ethanol) Trichloroacetic acid, 12% Procedure. The SDS gels are stained with 100 ml of 0.25% Coomassie Brilliant Blue R-250 in 40% methanol/10% acetic acid for 1 to 2 hr and destained overnight with several changes (200 ml) of the above solvent without the dye. 2° Aluminum lacate native gels are stained overnight in a filtered solution containing 0.1 g Coomassie Brilliant Blue R-250 in 240 ml of 12% trichloroacetic acid. Gels are destained overnight in 200 ml of 12% trichloroacetic acid. 1l Protein-stained gels are photographed with Polaroid type 55 film to produce prints and negatives. Prints are used to determine band migration distances and loading efficiency as well as to quantitate protein SH. Quantitation of Fluorescence [Reduction)

Scanning of Gels The Polaroid negatives of fluorescent gels and prints of wet proteinstained gels are scanned with a laser densitometer (Pharmacia-LKB, Piscataway, N J, UltroScan XL). Intensities of fluorescent bands are quantified by evaluating peak areas after integration with GelScan XL software. The

[24]

237

THIOREDOXIN AND SEED PROTEINS Proteln

Fluorescence 2

3

I 5

4

~ 1

I 5

Protein

Fluorescence I 6

7

8

9

10

I

I

6

10

kDa 97.4 66.2 w

45.0

31.0

21.5

14.4

Gliadins:

Glutenins:

1. 2. 3. 4. 5,

6. Control 7. GSH 8. GSH/GR/NADPH 9. NGS 10. NTS

Control GSH GSH/GR/NADPH NGS NTS

FIG. 1. Relative effectiveness of the NADP+/thioredoxin (NTS), NADP+/glutaredoxin (NGS), and the glutathione/glutathione reductase system (GSH/GR/NADPH) in the reduction of seed proteins (SDS, polyacrylamide gel electrophoresis).

intensity of fluorescent bands, when totally reduced by 2 mM DTT at 100° for 3 min and derivatized with mBBr, is proportional to the amount of added protein (up to 30/zg). It is assumed that the intensity of fluorescence of the proteins under analysis is also a direct measure of percent reduction of sulfhydryl groups. Specific Applications The techniques described above have been used successfully in determining the reduction of seed storage proteins and purified disulfide proteins from seeds as well as other sources. The following examples serve to illustrate use of the technique.

238

[241

THIOREDOXIN AND GLUTAREDOXIN

Effectiveness of Thiols in Reduction of Seed Proteins. By employing the in vitro labeling technique described above, our laboratory has obtained evidence that the NADP+/thioredoxin system reduces wheat storage proteins--gliadins and glutenins--much more effectively than the glutaredoxin or glutathione systems (Fig. 1). T M Similar results have been obtained in an unpublished survey of extracted prolamines and glutelins from several nonglutenous cereals. The system works equally well for pure disulfide proteins from seeds (Fig. 2) as well as other s o u r c e s . 12'19,2°,22 Change in Redox Status of Proteins during Germination. Using the in vivo labeling technique described above, our laboratory has obtained evidence for a change in redox state of wheat proteins during germination.11 We observed a progressive increase of 2- to 3-fold in reduction of the albumin/globulin, gliadin, and glutenin fractions up to the second day of germination; thereafter, reduction declined (Fig. 3). This pattern of redox change was coincident with the proteolytic degradation (or mobilization) of storage proteins. A similar series of events has been observed with barley DSG-1

BBTI

•

,g"

DTT GSH

~ + , ÷1i t

DTT ÷

_

GSH II +

_

C ~l +

DTT II

.

GSH ,

+

C ,

+

I

+ and - refer to thioredoxin (E.coli) C = Control (no reductant) •~ = Thioredoxin

Fie. 2. Dithiothreitol-reduced thioredoxin as a reductant for seed a-amylase and trypsin inhibitor proteins. DSG-1, Wheat a-amylase inhibitor; BBTI, soybean Bowman-Birk trypsin inhibitor (SDS, polyacrylamide gel electrophoresis).

[241

THIOREDOXIN AND SEED PROTEINS

5

I

I

,

o

.~a

i

a

•

,

,

I

239 I

t

o,/

i

i

~ ~ 3

~

o

0

0

1

2

3

4

Day

FIG. 3. In vivo reduction status of protein fractions during germination. [Data obtained by analysis of SDS-polyacrylamide gels.]

(C. Marx, J. H. Wong, and B. B. Buchanan, unpublished results); also in unpublished experiments, the reduction of disulfide bonds of gliadins and glutenins by the NTS was found to facilitate their degradation by proteases (I. Besse, and B. B. Buchanan, unpublished findings). Change in Redox Status of Thiol-Containing Proteins during Seed Maturation. Samples (wheat ears) are harvested at different stages of seed maturation up to complete maturity. Immediately after harvest, samples are put into liquid N2, and grains are dehuUed manually in a mortar in the presence of liquid N2. In vivo labeling with mBBr is performed under the conditions described for the in vivo labeling of germinated seeds. Results to date provide evidence that the storage proteins are synthesized in the reduced form and become oxidized during maturation and drying. Redox Changes during Technological Transformations. An extension of these experiments has revealed that the NADP+/thioredoxin system reduces selected flour proteins and strengthens the dough network formed

240

[251

THIOREDOXIN AND GLUTAREDOXIN

with flour of poor cooking quality, as measured by farinograph 16and baking tests. 33 These results suggest that addition of the NADP+/thioredoxin system improves products prepared from wheat and other cereals. Monobromobimane is currently being used as a probe to identify the major sulfhydryl/disulfide redox changes taking place in proteins during dough strengthening. Samplings are taken at different stages of the dough making process. Dough samples are put into liquid N2, and protein is fractionated and labeled with mBBr as described above.

33 K. Kobrehel, C. Nimbona, B. B. Buchanan, D. Bergmann, J. H. Wong, and B. C. Yee, "Gluten Proteins 1993," p. 381. Assoc. of Cereal Research, Detmold, Germany.

[251 A n a l y s i s a n d M a n i p u l a t i o n o f T a r g e t E n z y m e s Thioredoxin Control

By J E A N - P I E R R E

for

J A C Q U O T , E M M A N U E L L E ISSAKIDIS,

PAULETTE DECOTTIGNIES, MARTINE LEMAIRE,

and

MYROSLAWA MIGINIAC-MAsLOW

Introduction Several systems are able to control the activity, stability, and correct folding of enzymes through dithiol/disulfide isomerization reactions including the enzyme protein disulfide-isomerase, the glutathione-dependent glutaredoxin system, and the thioredoxin systems. 1-3 Plants contain two different thioredoxin systems, one extrachloroplastic in which thioredoxin is reduced through NADPH and the flavoenzyme NADPH-thioredoxin reductase (EC 1.6.4.5), and the other chloroplastic in which the reducing system is photoreduced ferredoxin and the iron-sulfur enzyme ferredoxinthioredoxin reductase. 4'5 To study thioredoxins and their reducing systems, it is necessary to use a target enzyme, the properties of which can be modified after reaction with reduced thioredoxin. Several proteins with this property have been isolated (including fructose-l,6-bisphosphatase, phosphoribulokinase, glyceraldehyde-3-phosphate dehydrogenase, glu1 R. B. Freedman, FEBS Lett. 97, 201 (1979). z A. Holmgren, Annu. Rev. Biochem. 54, 237 (1985). 3 H. Eklund, F. K. Gleason, and A. Holmgren, Proteins: Struct. Funct, Genet. 11, 13 (1991). 4 B. B. Buchanan, Annu. Rev. Plant Physiol. 31, 341 (1980). 5 T. C. Johnson, K. Wada, B. B. Buchanan, and A. Holmgren, Plant Physiol. 85, 446 (1987).

METHODSIN ENZYMOLOGY,VOL.252

Copyright © 1995by AcademicPress,Inc. All rightsof reproductionin any formreserved.

125]

THIOREDOXIN-REGULATED ENZYMES

241

cose-6-phosphate dehydrogenase, and CF1-ATPase). One of the best documented, easy to use, and most versatile enzymes, however, is the NADP +malate dehydrogenase (EC 1.1.1.82). This chapter deals with the purification of the protein from the leaves of C4 plants (in which the enzyme is present at a much higher level relative to C3 plants), the optimal conditions for its reductive activation, and activity measurements. Other procedures have been published for the purification of the protein from C4 or C3 plants. 6'7 A system is described for the expression and purification of the enzyme from Eseherichia coli cells and for its modification to make it even more versatile. Purification and Assay of Corn or Sorghum Leaf NADP+-Malate Dehydrogenase Purification Procedure Materials and Chemicals

Corn or sorghum plants grown on vermiculite, watered twice a week with demineralized water, for 3 weeks in a greenhouse (illumination of 60 W/m 2, 15 hr/day) Tris-HCl, 1 M, pH 8.0, 1 liter stock solution EDTA (Na salt), 1 M, 50 ml stock solution K2-K phosphate buffer, pH 7.2, 0.5 M stock solution (1 liter) 2-Mercaptoethanol Phenylmethylsulfonyl fluoride (PMSF), 50 mM (in 2-propanol) Polyvinylpyrrolidone (PVP 40), 40 g MgCI2, 1 M, 100 ml stock solution NaCI, 150 g Ammonium sulfate (Merck, Darmstadt, Germany), 1 kg Aca 44 (IBF, Villeneuve la Garenne, France), 1.5 liters, swollen DEAE-Sephacel (Pharmacia, Piscataway, N J), 80 ml, swollen Matrex red A, (Amicon, Exernon, France), 25 ml, swollen Dialysis tubing Pair of scissors, Waring blendor (1 gallon capacity), magnetic stirrer, UV monitoring system, peristaltic pump, fraction collector, temperature-controlled recording spectrophotometer Cold room (5 °) Extraction and Purification. Leaves (1 kg) are harvested, deribbed, cut into small pieces, and blended in 2 liters extraction buffer [30 mM Tris6 A. Agostino, P. Jeffrey, and M. D. Hatch, Plant Physiol. 98, 1506 (1992). 7 K. Fickenscher and R. Scheibe, Biochim. Biophys. Acta 749, 249 (1983).

242

THIOREDOXIN AND GLUTAREDOXIN

[25]

HCI, pH 8.0, 1 mM EDTA, 5 mM MgCI2, 14 mM 2-mercaptoethanol, 0.2% (w/v) PVP 40, 100/zM PMSF]. The resulting slurry is filtered through a double layer of nylon net (48 mesh) and centrifuged for 20 min at 10,000 g to eliminate cell debris. Ammonium Sulfate Fractionation. Solid ammonium sulfate (226 g/liter supernatant) is slowly added to the protein solution to bring it to 40% of saturation. After 30 min of stirring at 5 °, precipitating proteins are eliminated by centrifugation (20 min at 10,000 g). Then 120 g of solid ammonium sulfate is added per liter supernatant, which brings it to 60% of saturation. The NADP+-malate dehydrogenase (NADP+-MDH) precipitates at this step. After 30 min of stirring, it is recovered in the protein pellet obtained by centrifuging the extract for 20 min at 10,000 g. The proteins are dissolved in around 100 ml of extraction buffer from which PVP and MgC12 have been omitted and dialyzed overnight against 5 liters of 20 mM potassium phosphate buffer, pH 7.2, containing 1 mM EDTA. The latter buffer (PE) is then used throughout the purification. The sample is cleared after dialysis by spinning for 1 hr at 50,000 g. DEAE-Sephacel Chromatography. The clear supernatant is applied to a DEAE-Sephacel column (2.5 × 15 cm) previously equilibrated in PE buffer at a flow rate of 80 ml/hr. After extensive washing (3 bed volumes), the proteins are eluted with a 500-ml linear gradient of NaC1 (0-600 mM) in PE buffer. Fractions (6 ml) are collected and assayed for activity (see below). The enzyme elutes at a concentration of about 300 mM NaC1. The fractions showing N A D P + - M D H activity are pooled and concentrated by ultrafiltration in an Amicon cell fitted with a YM10 membrane to a final volume of 20 ml. Aca 44 Chromatography. The concentrated sample is applied to an Aca 44 column (5 x 70 cm) previously equilibrated with PE buffer. The filtration is performed at a flow rate of 60 ml/hr, and 7-ml fractions are collected and assayed for activity. The active fractions are pooled and used for the Matrex red A step. Matrex Red A Chromatography. The N A D P + - M D H contained in the solution is then bound to a Matrex red A column (1 x 10 cm) at a flow rate of 25 ml/hr. Unbound material is washed off the column by 5 bed volumes of PE buffer, and the proteins specifically retained (NADP+-dependent enzymes) are eluted using a 400-ml linear gradient of NaC1 (0-2 M) in PE buffer. The plant NADP+-MDH is eluted at a concentration of 1.5 M NaCI. After this step of purification, the enzyme is at least 80% homogeneous and can be used in this form for most purposes. If a fully purified sample is desired a final purification step by hydrophobic interaction chromatography as described below can be added.

[25]

THIOREDOXIN-REGULATED ENZYMES

243

Hydrophobic Interaction Chromatography. The NADP+-MDH can be further purified on a TSK phenyl 5PW column (0.75 x 7.5 cm; LKB). The column is equilibrated with 1.8 M ammonium sulfate in 0.1 M potassium phosphate, pH 7.2, and the proteins eluted with a 30-rain linear decreasing gradient of ammonium sulfate (1.8-0 M) in 0.1 M potassium phosphate, pH 7.2, followed by a 10-min wash with 0.1 M potassium phosphate, pH 7.2, at a flow rate of 1 ml/min. The NADP+-MDH sample is then dialyzed and concentrated using a C30 microconcentrator (Amicon). Dialysis and Storage. The NADP+-MDH sample recovered after Matrex red A chromatography has a very high concentration of salt. Dialyzing it by ultrafiltration at that step can result in denaturation and heavy precipitation. It is therefore recommended to dialyze it overnight against 5 liters of PE buffer, before concentrating it by ultrafiltration as described above. The enzyme can be conveniently stored in aliquots of 200 tzl at a concentration of 1 mg/ml at - 7 0 °. The protein concentration of the sample can be determined from its absorbance at 280 nm (1 A280 unit = 0.8 rag/ ml). The specific activity of the fully activated enzyme is about 400 U/rag protein at 30° (1 U = 1/zmol NADPH oxidized/rain).

Measurement of Activity The NADP+-MDH is totally dependent on activation by reduced thioredoxins for activity. Hence, it requires a preactivation step before the activity can be measured (this is detailed below). Once activated, the activity of the enzyme can be detected as follows.

Reagents Tris-HC1, pH 8.0, 1 M (stock solution 1 liter) Oxaloacetate* (OAA), 75 mM (stock solution 1 ml) NADPH, 15 mM (stock solution 1 ml) Assay Procedure. One cuvette (1 ml) contains 100 tzl Tris-HCl, pH 8, 10/xl stock OAA, 10 txl stock NADPH, and 880 txl water. The reaction is initiated by the addition of activated NADP+-MDH. It is followed spectrophotometrically by the decrease of absorbance at 340 nm (molar extinction coefficient for NADPH of 6200 M -1 cm-X).

Activation Methods for NADP+-Malate Dehydrogenase As explained above, the NADP+-MDH purified from plant leaves requires a preactivation step before its activity can be measured. This can

* Warning: Oxaloacetateis very unstable in solutions above pH 7. It shouldbe dissolvedin water, kept frozen at -20 °, and used within2 weeks.

244

THIOREDOXIN AND GLUTAREDOXIN

[251

be achieved in two ways: either with thioredoxin chemically reduced by dithioreitol (DTT) or with thioredoxin reduced photochemically through light and the ferredoxin-thioredoxin system. Therefore the activation of N A D P * - M D H can be used to detect either thioredoxin (any type when DTT is the reductant) 8 or ferredoxin-thioredoxin reductase (FTR) (reconstituted light activation system). 9

Dithiothreitol- Thioredoxin-Linked Activation Reagents Tris-HC1, pH 8.0, 1 M (stock solution 1 liter) Dithiothreitol (DTT), 100 mM (stock solution 1 ml) N A D P + - M D H , 1 mg/ml Activation Conditions. The reaction mixture contains 10/xl of 1 M TrisHCI, 10 /xl of 100 mM DTT, 5 /xl NADP+-MDH, 5-20 /xl thioredoxin fractions, and water to 100 txl. The activation is started by the addition of thioredoxin. The activation rates are followed by removing 20-/zl samples at 5-rain intervals and injecting them into the spectrophotometer cuvette containing the reaction medium. The rates are proportional to the thioredoxin content of the fractions. Under these conditions, as little as 0.1 txg thioredoxin can be detected per 20-txl sample. Most thioredoxins have a good reactivity with NADP+-MDH. There are, however, at least two cases where the enzyme is poorly activated, namely, with human thioredoxin 1° and with thioredoxins from Dictyosteliurn.ll

Light-Dependent Activation The assay is carried out essentially as described in elsewhere this series, xz with some simplification.

Reagents Tris-HC1, pH 8.0 (stock solution 1 liter) Pea thylakoids, freshly isolated, 2.5 mg chlorophyl/m113 8 p. Schfirmann, K. Maeda, and A. Tsugita, Eur, J. Biochem. 116, 37 (1981). 9 M. Droux, M. Miginiac-Maslow, J. P. Jacquot, P. Gadal, N. A. Crawford, N. S. Kosower, and B. B. Buchanan, Arch. Biochem. Biophys. 256, 372 (1987). 10j. p. Jacquot, F. de Lamotte, M, Fontecave, P. Schfirmann, P. Decottignies, M. MiginiacMaslow, and E. Wollman, Biochem. Biophys. Res. Comrnun. 173, 1375 (1990). ll B. Wetterauer, M. V6ron, M. Miginiac-Maslow, P. Decottignies, and J. P. Jacquot, Eur. J. Biochem. 209, 643 (1992). ~2N. A. Crawford, B. C. Yee, M. Droux, D. E. Carlson, and B. B. Buchanan, this series, Vol. 167, p. 415. 13 M. Miginiac-Maslow, J. P. Jacquot, and M. Droux, Photosynth. Res. 6, 201 (1985).

[25]

THIOREDOXIN-REGULATED ENZYMES

245

Spinach ferredoxin, 2 mg/m114 Spinach thioredoxin m, 1 mg/ml 8 N A D P + - M D H , 1 mg/ml Activation Conditions. In a colorless E p p e n d o r f tube are mixed 10/xl of 1 M Tris-HC1, 10 /~1 pea thylakoids, 5 /zl ferredoxin, 5 /xl N A D P +M D H , 20 txl thioredoxin, 5-20/xl ferredoxin-thioredoxin reductase (FTR) fractions, and water to 100/xl. The activation reaction is initiated by placing the mixture in an illuminated water bath (100 W/m2). Every 5 min a sample of 30 txl is removed and injected into the spectrophotometer cuvette for activity measurement. When using freshly isolated thylakoids, there is no need for anaerobiosis, and water can serve as the electron donor, which makes the system easier to use. (Warning: The thylakoid preparation is not stable and should be kept in the dark on ice and used within 5 hr at most.) With this assay, as little as 0.3 tzg F T R can be detected per 30-/xl sample.

Cloning, O v e r e x p r e s s i o n , a n d Purification of R e c o m b i n a n t NADP +Malate D e h y d r o g e n a s e The purification procedure using plant leaves is time-consuming and requires large amounts of plant material. Thus, it is often more convenient to overproduce the protein in E. coli.

Expression Systems for NADP÷-Malate Dehydrogenase One prerequisite for expressing a eukaryotic protein in prokaryotes is to have a full-length c D N A available. This was realized in our laboratory with a c D N A obtained from a sorghum library. 15 Several plasmids are available for expression of nonfused proteins (including pKK, p P R O K , pTrc, t6 and pET17). All the plasmids have either an NcoI or an NdeI site which allows the introduction of the sequence at the level of the initiator methionine. Depending on the sequence to be introduced, several strategies can be employed for the cloning. The easiest and most versatile way of cloning is indeed the use of the polymerase chain reaction (PCR) as described below. 14S. G. Mayhew, Anal. Biochem. 42, 191 (1971). 1.~C. Cr6tin, P. Luchetta, C. Joly, P. Decottignies, L. Lepiniec, P. Gadal, M. Sallantin, J. C. Huet, and J. C. Pernollet, Eur. J. Biochem. 192, 299 (1990). f6E. Amman and J. Brosius, Gene 40, 183 (1985). 17F. W. Studier, A. H, Rosenberg, J. J. Dunn, and J. W. Dubendorff, this series, Vol. 185, p. 60.

246

THIOREDOXIN AND GLUTAREDOXIN

[251

Reagents 10X Taq polymerase buffer Mix of the four deoxynucleoside triphosphates (dNTPs), 5 mM Upstream and downstream oligonucleotides, 5 t~M stock BioTaq polymerase (5 units/tzl) (Bioprobe, Montreuil-Sous-Bois, France) Oligonucleotide primers For the design of the oligonucleotides it is recommended to use the following rules: both oligonucleotides should hybridize over a 15- to 18base length. Appropriate restriction sites and additional bases (2) can be introduced upstream of the codon for the first amino acid and downstream of the termination codon. A typical example is given below (designed for the use of pET 3d vector). N-terminal protein sequence: S V D A A K C-Terminal protein sequence: L P G E V Ter Upstream oligonucleotide: 5' CCCC A T G GCC GTC GAC GCC GCC A A G 3' NcoI Downstream oligonucleotide: 5' CCGGATC CTA CAC TTC TCC CGG TAG 3' BamHI

This upstream olignonucleotide has been chosen for cloning convenience in the pET 3d vector. The use of the cloning series with the NcoI site sometimes results in the need for changing the first amino acid downstream of the initiator methionine (in this case, Set to Ala) to keep the proper reading frame. This modification does not change the properties of the protein. It can be circumvented by the use of pET12a series in which the upstream cloning site is NdeI. 17 Polymerase Chain Reaction. The reaction mixture contains 10/xl of 10x Taq polymerase buffer, 4 /xl of 5 mM dNTPs, 20 /xl of upstream oligonucleotide stock, 20/xl of downstream oligonucleotide stock, and plasmid containing the nucleotidic sequence (100 ng DNA). The reaction mixture is placed in 1.5-ml Eppendorf tubes tightly closed with a plastic clip (avoid using mineral oil), and DNA is first denatured by heating for 5 min at 95°. The mixture is then cooled on ice, spun briefly, and 0.5/zl BioTaq polymerase added. Twenty cycles are performed in a PCR machine (Biomed Thermocycler 60, Braun, Theres, Germany) (1 min at 94 °, 2 min at 50 °, 3 min at 72°). The resulting product is then purified by agarose gel electrophoresis (1%), extracted with phenol, and concentrated by precipitation with ethanol. Following digestion with the appropriate restriction enzymes (the addition of the two external bases next to the restriction sites is essential for this purpose), the product is ligated to the plasmid previously digested under the same conditions. The resulting construction is then used to transform E. coli cells. DNA minipreparations are realized with random chosen

[251

THIOREDOXIN-REGULATED ENZYMES

247

ampicillin-resistant colonies (a resistance conferred by the plasmid) and checked by digestion with the restriction enzymes used above. All the procedures used for digestion, ligation, transformation, and D N A preparations are quite standard and are described by Sambrook et aL TM The above cloning procedure has proved successful in our laboratory for a number of sequences including NADPH-thioredoxin reductase, t9 thioredoxins, fructose-l,6-bisphosphatase (FBPase) (J.-P. Jacquot, E. Issakidis, P. Decottignies, M. Leymaire, and M. Miginiac-Maslow, unpublished results). Therefore, it seems to be applicable to a broad variety of cases. The ehloroplastic thioredoxin-dependent enzymes are nuclear-encoded. Hence, their full-length cDNA sequence comprises a transit peptide which is processed after protein import into the chloroplast. In the case of N A D P ÷MDH, the transit peptide is cleaved by the bacteria, and hence the sequence can be cloned either as the full-length form or in a short form deprived of the sequence coding for the transit peptide. 2° Expression and Purification o f R e c o m b i n a n t Protein Materials

Incubator shaker, Series 25, New Brunswick Instruments (Edison, NJ) Centrifuge Luria-Bertani (LB) medium, sterile, 5.5 liters containing ampicillin (50 tzg/ml) Stock solutions as for the plant enzyme purification All cultures are performed at 37 °. After transformation of the appropriate E. coli strain (BL21 for the pET plasmid), one colony is used to inoculate 3 ml of LB medium. After 8 hr, the culture is transferred to 400 ml of LB medium and left to grow overnight (15 hr). This culture is in turn used to inoculate 4.8 liters of LB medium distributed in six l-liter Erlenmeyer flasks (each containing 400 ml) and three 2-liter Erlenmeyer flasks (each containing 800 ml). After 3 hr of culture (OD600of 0.6), 100/~M isopropyl-/3-o-thiogalactopyranoside (IPTG) is added. After an additional 5 hr, the cells are collected by centrifugation (5000 g, 10 min). The pellets can then be stored at - 7 0 °. Protein purification is carried out essentially as for the plant enzyme. However, the overproduction of the enzyme by transformed bacteria allows ts j. Sambrook, E. F, Fritsch,and T. Maniatis,in "MolecularCloning:A LaboratoryManual." Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. 1989. f9j. p. Jacquot, R. Rivera-Madrid,P. Marinho, M. Kollarova,P. LeMarrchal, M. MiginiacMaslow, and Y. Meyer, J. MoL Biol. 235, 1357 (1994). 20j. p. Jacquot, E. Keryer, E. Issakidis,P. Decottignies,M. Miginiac-Maslow.J. M. Schmitter. and C. Crrtin, Eur. J. Biochem. 199, 47 (1991).

248

THIOREDOXIN AND GLUTAREDOXIN

[9,5]

some simplification of the procedure. Bacterial extracts are much cleaner than plant extracts, and consequently MgCI2, 2-mercaptoethanol, and PVP can be omitted. The cell pellets are thus resuspended in 50 ml of extraction buffer consisting of 30 mM Tris-HC1 buffer, pH 7.4, 1 mM EDTA, and 100 /zM PMSF. The bacteria are disrupted by three passages in a French pressure cell (60 MPa). The cell debris are removed by centrifugation (30 rain, 20,000 g). Ammonium Sulfate Fractionation. Solid ammonium sulfate (19.4 g per 100 ml supernatant) is slowly added to the protein solution to bring it to 35% of saturation. After 30 min of stirring at 5°, precipitating proteins are eliminated by centrifugation (20 min at 10,000 g). Approximately 15.1 g of solid ammonium sulfate is added per 100 ml supernatant, which brings it to 60% of saturation. The NADP+-malate dehydrogenase precipitates at this point. After 30 min of stirring, it is recovered in the protein pellet obtained by centrifuging the extract for 20 rain at 10,000 g. The proteins are dissolved in approximately 50 ml of extraction buffer and dialyzed overnight against 5 liters of 20 mM potassium phosphate buffer, pH 7.2, containing 1 mM EDTA. The latter buffer (PE) is then used throughout the purification. The sample is cleared after dialysis by spinning for 1 hr at 50,000 g. DEAE-Sephacel Chromatography. Chromatography on DEAE-Sephacel is carried out as for the plant protein. The fractions containing NADP ÷MDH activity (major protein peak) are pooled and directly loaded onto the affinity Matrex red A column. Matrex Red A Chromatography. Matrex red A chromatography is performed as for the plant protein. The enzyme is almost homogeneous after this step (90-95%). Up to 40 mg protein can be obtained and used for most studies without further purification. Dialysis and Storage. Dialysis and storage are as for the plant protein. N-Terminal Sequencing. To check that the recombinant NADP +-MDH is identical to the plant protein, it is necessary to analyze the amino-terminal sequence. The sample is desalted by reversed-phase high-performance liquid chromatography (HPLC) on a Vydac C4 column (4.5 × 150 ram) (Touzart et Matignon, Vitry-sur-Seine, France), eluted with a 20-rain gradient of 35-70% acetonitfile in 0.1% trifluoroacetic acid. This step is also useful to remove the minor contaminants still present in the preparations. The solvents are evaporated in a Speed-Vac concentrator (Savant) and the protein subjected to Edman degradation using an Applied Biosystems (Foster City, CA) Model 476A sequqencer, with on-line HPLC detection of phenylthiohydantoin amino acids. The amino-terminal sequence of the recombinant NADP+-MDH is identical to that of the plant enzyme. The fact that the recombinant protein is produced in the mature form indicates

[25]

THIOREDOXIN-REGULATED ENZYMES

249

that the transit peptide can be cleaved by the E. coli cells. In addition to the major form of the protein, some additional forms, shortened by a few amino acids, are obtained. This microheterogeneity, although less marked, is also observed in the case of the plant protein. Modification of NADP+-Malate Dehydrogenase for Use in NADPH/ NADPH-Thioredoxin Reductase Assay The major drawback for the utilization of NADP+-MDH is that its activation is strongly inhibited by NADP ÷.21 As a consequence, in its native form the wild-type enzyme cannot be used to detect the functioning of the NTR system, as the latter generates oxidized NADP ÷. One way to circumvent this problem is to design an NADP+-MDH molecule, the activation of which is not inhibited by NADP +. This can be achieved by site-directed mutagenesis followed by expression of the recombinant protein in E. coli cells and subsequent purification of the protein. Site-Directed Mutagenesis

The procedures used for site-directed mutagenesis of NADP÷-MDH have been detailed by Issakidis et al. 22 They are based on the use of the method described by Kunkel et al. in this seriesY The technique is based on the use of an uracil-containing single-stranded DNA molecule as a template. Modification is carried out by hybridizing a mutagenic oligonucleotide to the template. Wild-type DNA is then counterselected by transformation of an ung + E. coli strainY First, the cDNA of choice must be cloned into an M13 phagemid which is used to infect E. coli strain RZ1032 (dut-, ung-) in order to generate and purify uracil-substituted single-stranded DNA template. Stock Solutions

10× Kinase buffer: Tris-HCl, 500 mM, pH 7.5; MgCI2, 100 mM; DTT, 20 mM 10× Annealing buffer: Tris-HCl, 200 raM, pH 7.5; EDTA, 20 mM, pH 8.0; NaC1, 500 mM 10× Synthesis buffer: Tris-HCl, 100 raM, pH 7.5; MgCI2,50 raM; DTT, 20 mM ATP 10 mM Mix of the four dNTPs (5 mM each) 21R. Scheibe and J. P. Jacquot, Planta 157, 548 (1983). 22E. Issakidis,M. Miginiac-Maslow,P. Decottignies,J. P. Jacquot, C. Cr6tin, and P. Gadak J. Biol. Chem. 267, 21577 (1992). a3T. A. Kunkel,J. D. Roberts, and R. A. Zakour, this series, Vol. 154, p. 367.

250

THIOREDOXIN AND GLUTAREDOXIN

[25]

T4 kinase (->1 U//zl) Klenow fragment of E. coli D N A polymerase I (->1 U/Izl) T4 ligase (->1 U//zl) Phosphorylation of Mutagenesis Oligonucleotide. Mix the following in an Eppendorf tube: 5/~1 water, 1/zl of 10× kinase buffer, 1/zl oligonucleotide (0.1 /zg//zl), 2 /~I of 10 mM ATP, and ] /zl T4 kinase (->1 U//zl). Incubate for 30 min at 37 °. Inactivate the kinase by heating (10 min, 70°). Spin and store on ice.

Annealing of Phosphorylated Oligonucleotide with Single-Stranded Uracil-Containing Template; Polymerization/Ligation. Mix the following in an Eppendorf tube: 2/zl uracil-containing single-stranded D N A (100 ng//zl), 1/zl phosphorylated oligonucleotide, 1 /zl of 10x annealing buffer, and 6 /zl water. Place the mixture at 70° in a dry bath; then let the temperature decrease to 30 ° on the bench. Keep on ice. Add the following to the annealing mixture: 2/xl water, 2/zl of 10x elongation buffer, 2/zl of a mix containing the four dNTPs (5 mM each), 2/zl of 10 mM ATP, 1/xl T4 ligase (->1 U//zl), and 1 tzl of the Klenow fragment of E. coil D N A polymerase (->1 U//zl). Vortex, spin, place for 5 min on ice, and then 5 min at room temperature. Incubate for 90 min at 37 °. After transformation of an E. coil (ung +, dut ÷) strain (DHSc~), six plaques are picked at random and mutant phagemids are identified by sequencing. This procedure yields from 60 to 100% mutants. The mutant sequences are transferred to one of the expression systems described earlier.

NADP+-Malate Dehydrogenase C405A/C417A as Tool for NADPHThioredoxin Reductase/Thioredoxin h Detection Sorghum leaf NADP+-malate dehydrogenase contains 8 cysteine residues, 4 of which constitute two regulatory disulfide bridges located at the N and C termini of the protein as shown in Fig. 1. It has been found that the mutation of either C405 or C417 to Ala or the combined mutation C405A/C417A yields a protein the activation of which is no longer sensitive

C64 C6g

C405 C417

FIG. 1. Schematic drawing of sorghum NADP+-malate dehydrogenase subunit showing the sequence extensions (in black) which are not present in the NAD+-dependent forms, and the positions of the N-terminal (left) and C-terminal (right) bridges. The cysteine numbering includes the 40 amino acids of the transit peptide, which is cleaved in the mature protein.

[25]

THIOREDOXIN-REGULATED ENZYMES

251

to NADP + inhibition. 24 Hence, the mutant protein can be used in a convenient test for the detection of NADPH-thioredoxin reductase (NTR) or thioredoxin h as described below.

Purification of Mutated NADP+-Malate Dehydrogenase The mutated protein, overproduced in E. coli, is purified essentially by following the same procedure as for the wild-type protein. However, it binds more strongly to the Matrex red A affinity column. As a consequence, a 500-ml linear gradient of NaC1 (0-4 M) in PE buffer must be applied, the protein being eluted at 2.7 M NaC1. This very high salt concentration must be removed by two successive dialysis steps (against 5 liters PE buffer) before concentration, in order to avoid protein precipitation. The resulting mutant protein, where the two C-terminal cysteines have been replaced by alanines, can be activated in the presence of NADP + and hence is suitable for the detection of the NTR/thioredoxin h system.

Use of Mutated NADP+-Malate Dehydrogenase for Detection of NA D PH- Thioredoxin Reductase Reagents Tris-HC1, pH 8 (stock solution 1 liter) Thioredoxin h, 1 mg/m125 N A D P H 10 mM (stock solution 1 ml) Mutated N A D P + - M D H 1 mg/ml Activation Conditions. For activation, 10 /zl of 1 M Tris-HCl, 10 /xl NADPH, 5/xl N A D P + - M D H , 20/xl thioredoxin h, 5-20/xl NTR fractions, and water to 100/xl are mixed in an Eppendorf tube, the reaction being started by the addition of NTR fractions. Aliquots (20/xl) are removed every 5 rain and injected into the spectrophotometer cuvette for the measurement of N A D P + - M D H activity. As little as 0.1 /xg of NTR can be detected per 20-/xl sample with this procedure. In addition, because the overal volume for activation is 100/xl instead of 1 ml for the standard 5,5'dithiobis(2-nitrobenzoic acid) (DTNB) reduction assay, the quantities of thioredoxin to be added are much lower (in a homolgous NTR/thioredoxin h assay, a more or less standard Km value is 2/xM). To reach this concentration, 24/zg should be used per ml versus 2.4/xg in a 100-/xl assay mixture. 24 E. lssakidis, M. Saarinen, P. Decottignies, J. P. Jacquot, C. Crdtin, P. Gadal, and M. MiginiacMaslow, J. BioL Chem. 269, 3511 (1994). 2s p. Decottignies, J. M. Schmitter, S. Dutka, J. P. Jacquot, and M. Miginiac-Maslow, Eur..I. Biochem. 198, 505 (1991).

252

THIOREDOXIN AND GLUTAREDOXIN

[25]

Properties of Mutant C405A/C417A. The mutated protein still requires preactivation by reduced thioredoxin to be active, with activation kinetics similar to those of the wild-type protein (full activation with 10-15 min of preincubation). However, it exhibits a slight spontaneous activity (5% of the activity of the fully activated protein). As a consequence, controls without thioredoxin must be run in all cases. The mutant protein is also more susceptible to denaturation by repeated freeze/thawing and should be stored in small aliquots at - 7 0 °. Use of Site-Directed Mutagenesis for Identification of ThioredoxinControled Disulfide Bridges In our hands, site-directed mutagenesis proved very useful for the identification of the thioredoxin-dependent regulatory disulfides of NADP +MDH. It has also been applied successfully to identify the regulatory disulfide of phosphoribulokinase. 26 It should be applicable to identify other thioredoxin-regulated sites. As a rule, the identification of the disulfide should consist of the replacement of two single cysteines and the combined replacement of both by isosteric amino acids (serines or alanines), giving identical results. In the case where one of the cysteines is involved in substrate binding or catalysis, one of the single mutations (and the double one) might result in an inactive protein which must then be monitored during purification by immunological methods (dot-blots). The thioredoxin sensitivity of the mutant proteins should be checked using different reduction systems (DTT or a specific reductase), as thioredoxin-independent illegitimate disulfides (reducible by DTT, but not by thioredoxin) might be formed during protein folding in E. coli or repeated freeze/thawing of the pure protein. Candidate cysteines for mutagenesis of thioredoxinregulated chloroplastic enzymes can generally be identified from sequence comparisons with equifunctional but thioredoxin-independent enzymes: they appear in sequence extensions or insertions 27 and are strictly conserved among the various plant sources.

26 S. Milanez, R. J. Mural, and F. C. Hartman, J. Biol. Chem. 266, 10694 (1991). 27 R. Scheibe, Bot. Acta 103, 327 (1990).

[26l

THIOREDOXIN AND ACTIVATION OF N F - r B

253

[26] E f f e c t s o f T h i o r e d o x i n o n A c t i v a t i o n o f T r a n s c r i p t i o n Factor NF-KB

By KLAUS

SCHULZE-OsTHOFF, H E I K E SCHENK, a n d W U L F D R O G E

Introduction Studies from different laboratories have shown that the regulation of the intracellular redox state may represent a versatile physiological control mechanism in gene expression. Low molecular weight thiols, such as cysteine or glutathione, control various cellular functions. 1-3 In lymphocytes reducing conditions promote cell proliferation, cytotoxic T-cell activity, and other effector functions. Other processes are facilitated or potentiated under prooxidant conditions or in situations of oxidative stress. 4-1° Reactive oxygen intermediates and lipid peroxides, for instance, have been shown to increase the expression of cytokines,4-6 adhesion molecules, 7,8 and protooncogenes. 9"~° These observations suggest that lymphocyte functions in general and gene expression in particular may be controlled by a delicate balance between prooxidant and antioxidant conditions and that a shift in either of these directions may selectively induce certain sets of genes. The induction of gene expression is in most cases regulated at the transcriptional level by the activation of transcription factors that bind to specific D N A sequences in the promoter region of target genes. The inducible or tissue-specific transcription factors act in conjunction with the basal transcriptional machinery and can dramatically increase the rate of transcription directed by R N A polymerase II. u An important activator of expression of numerous immunoregulatory genes is the transcription factor ~H. Gmtinder, H.-P. Eck, B. Benninghoff, S. Roth, and W. DrOge, Cell ImmunoL 12, 32 (1990). 2 W. Drfge, H.-P. Eck, and S. Mihrn, ImmunoL Today 14, 211 (1992). 3 W. DrOge, H.-P, Eck, and S. Mihrn, Med. lmmunol. 26, 357 (1993). 4 S. Roth and W. Dr6ge, Cell. Immunol. 108, 417 (1987). 5 S. Roth and W. Drfge, Eur. J. lmmunoL 21, 1933 (1991). 6 L. E. DeForge, A. M. Preston, E. Takeuchi, J. Kenney, L. Boxer, and D. R. Rernick, .I. Biol. Chem. 268, 25568 (1993). 7 S. K. Lo, K. Janakidevi, L. Lai, and A. B. Malik, Am. J. Physiol. 264, L406 (1993). s N. Marui, M. K. Offermann, R. Swerlick, C. Kunsch, C. A. Rosen, M. Ahmad, R. W. Alexander and R. M. Medford, J. Clin. Invest. 92, 1866 (1993). 9 D. Crawford, J. Zbinden, P. Amstad, and P. A. Cerutti, Oncogene 3, 27 (1988). 10 K. Nose, M. Shibanuma, K. Kikuchi, H. Kageyama, S. Sakiyama, and T. Kuroki, Eur. J. Biochem. 201, 99 (1991). ~J R. G. Roeder, Trends Biochem. Sci. 16, 402 (I991).

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press, lnc All rights of reproduction in any form reserved.

254

THIOREDOXIN AND GLUTAREDOXIN

[26]

NF-KBJ 2 In noninduced cells, NF-xB resides in the cytoplasm as an inactive complex bound to the inhibitory subunit IKB. On cell stimulation, IKB rapidly dissociates from the inactive complex and allows NF-KB to be translocated into the nucleus and to bind to its D N A sequence. Evidence has been accumulated indicating that this key step is essentially dependent on redox processes. First, NF-KB is rapidly activated in cells treated with hydrogen peroxide or tert-butyl hydroperoxide. 13-15 Second, activation of NF-KB by all stimuli known so far, including cytokines, phorbol esters, and viral transactivator proteins, is commonly inhibited by thiols and other antioxidant compounds. 16-2° However, NF-xB is not directly activated by prooxidants in cell-free systems, implicating the involvement of intermediate redox-sensitive molecules. A signal generated by these molecules may be counteracted by cellular antioxidants. Besides the glutathione system, an important antioxidant is the protein oxidoreductase thioredoxin, zI-23 Here we describe some experimental procedures to evaluate the effects of thioredoxin on NF-KB activation. These methods are suitable to analyze the effects of other thiols as well and may also be used for other transcription factors.

Effects of Recombinant Thioredoxin on NF-t 2 G S H + N A D P +

(1) (2) (3)

N A D P H + H + + XSSX --~ N A D P + + 2XSH

Activity is measured as the decrease of N A D P H absorbance at 340 nm. The reagents in the H E D assay are as follows. A cuvette contains 0.50 ml of 1 m M GSH, 0.4 m M N A D P H , 2 m M E D T A , 0.1 mg/ml BSA, 6/xg/ml glutathione reductase in 0.1 M Tris-C1, p H 8.0. The disulfide substrate H E D is added to each curvette (including the reference curvette) to a final concentration of 0.7 mM. A mixed disulfide between G S H and H E D is formed within 2 min, whereafter Grx is added to the sample cuvettes and buffer to the reference cuvette. Addition of 5 pmol of E. coli G r x l (5/xl of 1.0/zM G r x l ) gives a decrease of absorption at 340 nm of 0.140 min -1. The reaction is linear up to a rate of A340 of -0.200 min -a. The spontaneous background reaction is -0.020 min -1. One unit of activity is defined as the consumption of 1/xmol of N A D P H per minute calculated from the expression (&A340 × V)/(min x 6.2), where Vis the cuvette volume 46 M. Nikkola, F. K. Gleason, M. Saarinen, T. Joelson, O. Bj6rnberg, and H. Eklund, J. BioL Chem. 266, 16105 (1991). 47 y. Yang and W. W. Wells, J. Biol. Chem. 266, 12759 (1991).

1291

GLUTAREDOXIN

289

in milliliters and 6.2 is the millimolar extinction coefficient for NADPH. Pure Grxl, Grx2, and Grx3 from E. coli have the specific activities 220] ~ 287, 34 and 35034 units/rag. Reduction of Dehydroascorbic Acid Reduction of dehydroascorbic acid (DHA) is performed either indirectly, by calculation of the oxidation of N A D P H as described above, or directly as the increase of absorbance at 265 nm (the millimolar absorption coefficient of ascorbic acid is 12.59) by the formation of ascorbic acid. The ingredients in the assay may be as listed above (with the addition of D H A instead of H E D ) or as described by Wells and co-workers48:137 mM sodium phosphate, pH 6.8, i mM EDTA, 2 mM GSH, and 1 mM D H A (without N A D P H and glutathione reductase). Radioirnmunoassay and ELISA of Escherichia coli Grxl The first E. coli strains constructed lacking Grxl were confirmed to lack Grxl protein by a radioimmunoassay, using iodinated purified Grxl. 33 In a subsequent characterization of an analogous set of mutants, MirandaVizuete and co-workers developed a highly sensitive enzyme-linked immunosorbent assay (ELISA) for Grxl determinationsY Purification of Escherichia coli Glutaredoxins Escherichia coli Glutaredoxin 1 The gene coding for the Grxl protein has been cloned 1° and overexpressed in a hpL plasmid system, it The method described below was used for preparation of material used for the determination of the three-dimensional structure of Grxl by NMR. 14'15,17 Escherichia coli N4830/pAHOB1 cells are grown at 30° and induced by heat treatment for 5 hr at 40 °, resulting in expression of Grxl as 20% of the soluble E. coli protein, u Bacterial cells are disrupted in 50 mM Tris, pH 7.5, 1 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride (PMSF). A crude extract is obtained by centrifugation (10,000 g for 15 min). Ammonium sulfate is added to 85% solution, and the precipitated protein is dialyzed against 20 mM potassium phosphate, pH 7.3, 2 mM EDTA. v After dialysis, the material is applied to a column of DEAE-cellulose equilibrated with 20 mM potassium phosphate, pH 7.3, 2 mM E D T A and eluted with 0.2 M potassium phosphate, pH 7.3, 2 mM EDTA. The active fractions are then further purified to homogeneity on 48 W. W. Wells. D. P. Xu, Y. Yang, and P. A. Rocque, J. BioL Chem. 265, 15361 (1990).

290

THIOREDOXIN AND GLUTAREDOXIN

[291

a column of Sephadex G-50. Grx-90 (10% of the material) can be isolated on a small scale by separation on a Mono Q column (Pharmacia, Piscataway, NJ) equilibrated with 50 mM Tris-C1, p H 8.0, 1 mM EDTA, eluting the material with a gradient of 0-350 mM NaC1 in the same buffer. For largescale isolation of Grxl-85, DEAE-cellulose chromatography at pH 6.0 with a gradient of 0.10 to 0.40 M ammonium acetate is employed. 13 To avoid aggregation, the solution is raised to pH 7.5 before ultrafiltration concentration (lyophilization should be avoided), as the solubility of Grxl is very low below pH 5.0. Methods for the preparation of 15N,13C-labeled Grxl and a mixed disulfide of the 15N,13C-labeled Grxl C14S mutant with GSH have been developed) s Escherichia coli Grxl is available from IMCO (Stockholm, Sweden). Escherichia coli Glutaredoxins 2 and 3 The Grx2 and Grx3 forms have been purified from an E. coli AtrxA, grx :: kan strain. 34 A crude extract of the bacteria is prepared as above, whereafter nucleic acids are precipitated with 0.6-0.8% final concentration of slowly added streptomycin solution (10%). The supernatant is acidified to pH 5, and precipitated proteins are discarded by centrifugation. After neutralization, protein is collected with a 40-95% ammonium sulfate precipitation by centrifugation. The precipitate is extensively dialyzed against a buffer of 20 mM Tris, pH 9.5, 1 mM E D T A and then applied to a column of DEAE-cellulose equilibrated with this buffer. Grx2 and Grx3 coelute from the column on addition of 50 mM Tris-Cl, pH 8.0. After concentration and incubation with DTT (10 mM), the 27-kDa E. coli Grx2 separates from the 9-kDa Grx3 on Sephadex G-50 chromatography; alternatively, Grx2 may be removed by heat treatment 75 ° for 5 rain. Activated Thiol-Sepharose Chromatography Activated thiol-Sepharose chromatography is a recently developed technique for affinity chromotography of G r x . 34 This step is for practical reasons usually applied after a gel-filtration step, since it requires the reduced form without DTT. Presumably Grx-(SH)2 binds to the GSH matrix in the column, which is in the form of a GSH-2-pyridyl disulfide, as a mixed disulfide. To stabilize the binding, a buffer of pH 6 is used. For both E. coli Grx148a and pig liver glutaredoxin 47 it has been shown that the Nterminal of the active site cysteine residues has a pKa value that is unusually low, whereas the C-terminal active site residue shows a higher apparent pK~ value (around 9). 49 Thus, at pH 6.0 the SH group of the N-terminal 4sa O. Bj6rnberg and A. Holmgren, to be published. a9 T. E. Creighton, Prog. Biophys. Mol. Biol. 33, 231 (1978).

[291

GLUTAREDOXIN

291

active site cysteine would be in the thiolate form ready to initiate a nucleophilic attack on the GSH-2-pyridyl disulfide in the thiol-Sepharose (Pharmacia). Purification of Mammalian Glutaredoxin The first purification of mammalian glutaredoxin to homogeneity was performed by Luthman and Holmgren,2° which enabled the subsequent determination of its primary structure. 22'22 Partial purifications of presumably the same enzyme had been attempted since 1955 by others. 4'5'23 A breakthrough in the purification of mammalian glutaredoxin came with the pI shift method developed by Gan and Wells,2s which is described in detail in a chapter by Wells et al. elsewhere in this volume.5° Human placenta Grx has been purified to homogeneity, and the amino acid sequence was determined,51 based on the method by Luthman and Holmgren2° and the aforementioned pl shift method. The gene for human Grx was cloned and characterized. 5~ Concluding Remarks Glutaredoxin has been found in all organisms containing GSH, and even in one glutathione-negative species, namely, the archaebacterium Methanobacterium thermoautotrophicum. 52 Its presence in the genome of the fast replicating phage T4 demonstrates its importance for DNA synthesis as a hydrogen donor for RR. 53'54 Except as hydrogen donor for RR, the only well-established function of Grxl in vivo is that for sulfate (or PAPS) reduction. 55'56 Either Grxl or Trx has to be present in order to permit growth in liquid minimal media. 33 Neither Grxl nor Trx is essential for DNA synthesis, owing to the presence in E. coli of the Grx isoenzymes Grx2 and Grx3. 34 However, our data strongly suggest that in wild-type cells Trx and Grxl are both important hydrogen donors for RR, as the levels of ribonucleotide reductase activity in trx A, grx (the gene for Grxl) double mutants are increased 19- to 23-fold and only slightly elevated in single mutantsY In an attempt to investigate if a mutator phenotype was associs0 W. W. Wells, D. P. Xu, and M. P. Washburn, this volume [4]. 5t C. A. Padilla, E. M. Martinez-Galisteo, J. A. B~ircena, G. Spyrou, and A. Holmgren. Eur. J. Biochem. 227, 27 (1995). 52 S. C. McFarlan, C. A. Terrell, and H. P. C. Hogenkamp, J, Biol. Chem. 267, 10561 (1992). 53 O. Berglund, J. Biol. Chem. 244, 6303 (1969). 54 D. M. LeMaster, J. ViroL 59, 759 (1986). 55 M. Russel, P. Model, and A. Holmgren, J, Bacteriol. 172, 1923 (1990). 56 M. Russel, this volume [27].

292

THIOREDOXIN AND GLUTAREDOXIN

[29]

ated with the lack of Trx and/or Grxl, these mutants were investigated for sensitivity to different mutagenic agents. No major differences in sensitivity was recorded between the different mutants, suggesting well-balanced deoxynucleoside-triphosphate (dNTP) pools despite the lack of efficient hydrogen donor systemsY The interconversion of the reducing equivalents between GSH and ascorbic acid is likely to be of great physiological importance 57 and to be mediated by Grx. 48 The general activity of Grx with low molecular weight disulfides or GSHmixed disulfides (see above) has led to its characterization by a number of assays using different disulfide substrates. As it has not been apparent in the earlier studies that the actual substrate in the reaction is a mixed disulfide with GSH, Grx has been named after its activity with its supposed disulfide substrate. Examples of names referring to Grx in the literature include GSH-homocystine reductase 4 and thiol-disulfide transhydrogenaseJ Based on a model which did not include a GSH-mixed disulfide intermediate, Grx was named thioltransferase. 6 To avoid further confusion, we advocate the glutaredoxin name, which is easy to associate with the name of the functionally related disulfide reducing protein thioredoxin. Thus, Grx-(SH)2 acts either as a dithiol reductant with protein disulfides or as a specific GSH-mixed disulfide reductase because of its GSH binding site. The substrate specificity of Grx is often different from that of thioredoxin, which has only low activity (high Km value) with low molecular weight disulfides. For proteins it appears that there is a clear difference in substrate specificity owing to steric and charge factors around the disulfide. This forms the basis for unknown redox regulation effects on proteins with synergistic or opposing effects of Trx and GSH and Grx, Acknowledgments This research was supported by grants from the Swedish Cancer Society (961), the Swedish Medical Research Council (Project 13X-3529), the Inga-Bfitt and Arne Lundbergs Stiftelse, the Knut and Alice Wallenbergs Stiftelse, and the Karolinska Institute. The excellent secretarial assistance of Mrs. Lena Ringd~n is gratefully acknowledged.

57 A. Meister, Biochem. Pharmacol. 44, 1905 (1992).

[30]

NMR

OF THIOREDOXIN

AND GLUTAREDOXIN

[30] N u c l e a r M a g n e t i c R e s o n a n c e and Glutaredoxin

293

of Thioredoxin

By H. JANE DYSON Introduction Thioredoxins and glutaredoxins from a number of sources, with molecular weights of the order of 9000-12,000, are sufficiently small to be amenable to nuclear magnetic resonance (NMR) spectroscopy. NMR spectra of thioredoxin were reported as early as 1976,1 but extensive NMR studies on these proteins were not made until the advent of high-field spectrometers in the 1980s. One of the major advances has been the development of methods to utilize NMR information to derive three-dimensional structures of proteins and other macromolecules (reviewed in Clore and Gronenborn2). This application has provided a major impetus in the study of thioredoxins and glutaredoxins by NMR, as X-ray crystallography cannot be used to derive structures in the majority of cases because crystals are not formed. An exception to this is the well-known X-ray crystal structure of the oxidized form of Escherichia coli thioredoxin,3-5 which was crystallized in the presence of Cu(II).6 Reduced E. coli thioredoxin could not be crystallized under the same conditions owing to the redox reaction between the copper ions and the protein dithiol, and its three-dimensional structure has therefore been solved by NMR spectroscopy;7 high-resolution structures of both forms of the protein under identical conditions have been determined. ~ The structure of reduced human thioredoxin cloned from T lymphocytes and overexpressed in E. coli has also been solved by NMR, 9 and a comparison of the solution structures of the oxidized and reduced forms of a A. Holmgren and G. Roberts, FEBS Letr 71, 261 (1976). 2 G. M. Clore and A. M. Gronenborn, Annu. Rev. Biophys. Biophys. Chem. 20, 29 (1991). 3 B.-O. SOderberg, A. Holmgren, and C.-I. Br~indfn, J. Mol. Biol. 90, 143 (1974). 4 A. Holmgren, B.-O. S6derberg, H. Eklund, and C.-I. Br~indfn, Proc. NatL Acad. Sci. U.S.A. 72, 2305 (1975). S. K. Katti, D. M. LeMaster, and H. Eklund, J. Mol. Biol. 212, 167 (1990). A. Holrngren and B.-O. SOderberg, J. Mol. Biol. 54, 387 (1970). 7 H. J. Dyson, G. P. Gippert, D, A. Case, A. Holmgren, and P. E. Wright, Biochemistry 29, 4129 (1990). 8 M.-F. Jeng, A. P. Campbell, T. Begley, A. Holmgren, D. A. Case, P. E. Wright, and H. J. Dyson, Structure 2, 853 (1994). 9 j. D. Forman-Kay, G. M. Clore, P. T. Wingfield, and A. M. Gronenborn, Biochemistry 30, 2685 (1991).

METHODS IN ENZYMOLOGY, VOL. 252

Copyright © 1995 by Academic Press. Inc. All rights of reproduction in any form reserved,

294

THIOREDOXIN AND GLUTAREDOXIN

1301

structurally similar mutant human protein has been published. 1° The Xray structure of T4 glutaredoxin (thioredoxin) has been determined, ~1 but E. coli glutaredoxin has not been crystallized. The NMR solution structures of both oxidized and reduced glutaredoxin have been determined, m3 As well as providing structural information, the NMR spectrum of a protein can also be used to probe enzyme mechanism and function and the binding of ligands. A number of such studies have been performed for thioredoxins and glutaredoxins and are summarized in this chapter.

Nuclear Magnetic Resonance of Thioredoxin The high thermal stability and small size of thioredoxin (108 amino acids for E. coli and 105 for human) make it a suitable candidate for NMR spectroscopy. Other desirable attributes in a protein to be studied by NMR are a high solubility without aggregation (for maximum sensitivity of measurement), stability at relatively low pH values (for lowered solvent exchange of amide protons), and a high yield of recombinant protein from the bacterial expression system used (for maximum efficiency of incorporation of isotope labels, if these are to be used). Thioredoxins from E. coli and human sources meet many of these criteria. Solution Conditions for Nuclear Magnetic Resonance Measurement Thioredoxin is prepared in the oxidized form using E. coli strain SK3981 containing the thioredoxin gene (trxA) cloned on a derivative of the plasmid pBR325. TM Preparation of labeled thioredoxin utilized strain Cla, which grows well on minimal medium, t5 Purification of protein by column chromatography TM follows a modification of the method of Holmgren and ReichardJ 6 Reduced thioredoxin for NMR is prepared from the oxidized protein by reduction with dithiothreitol (DTT). Interference from the 1H signals of the DTT is avoided either by use of fully deuterated DTT or by passage of the solution containing reduced thioredoxin and unlabeled DTT through 10j. Qin, G. M. Clore, and A. M. Gronenborn, Structure 2, 503 (1994). 11 B.-O. SSderberg, B.-M. Sit)berg, U. Sonnerstam, and C.-I. Brtind6n, Proc. Natl. Acad. Sci. U.S.A. 75, 5827 (1978). i2 p. Sodano, T.-H. Xia, J. H. Bushweller, O. Bjt~rnborg, A. Holmgren, M. Billeter, and K. Wtithrich, J. Mol. Biol. 221, 1311 (1991). 13T.-H. Xia, J. H. Bushweller, P. Sodano, M. Billeter, O. BjSrnberg, A. Holrngren, and K, Wtithrich, Protein Sci. 1, 310 (1992). 14 H. J. Dyson, A. Holmgren, and P. E. Wright, F E B S Lett. 228, 254 (1988). 15 K. Chandrasekhar, G. Krause, A. Holmgren, and H. J. Dyson, F E B S Lett. 284, 178 (1991). 16 A. Holmgren and P. Reichard, Eur. J. Biochem. 2, 187 (1967).

[30]

NMR

OF THIOREDOXIN AND GLUTAREDOXIN

295

i

I

I

I

L

I

I

I

]---

~

i

I

11

10

9

8

7

6

5

4

3

2

1

0

1H Chemical Shift (ppm) FIG. 1. The 500 MHz 1H NMR spectrum of oxidized Escherichia coli thioredoxin in 90% 1H20/10% 2H20, pH 5.7, 30°. The resonance of the 1H20 protons at approximately 4.6 ppm has been deleted for clarity,

a final Sephadex G-25 column under argon to remove the D T T . ~7 Both the oxidized and reduced forms of E. coli thioredoxin are soluble to at least 4 mM at pH 5.7 in 100 mM phosphate buffer1,14 and to 5-8 mM in 150 mM NaCI, 20 mM phosphate, pH 5.7. is This is the lowest pH that can be used before broadening of resonances in the NMR spectrum is observed, indicating the onset of aggregation, t Under these conditions, the 1H NMR spectrum is well-dispersed with narrow resonance lines, typical of a wellfolded, monomeric protein of this size (Fig. 1). The related protein cloned from human T lymphocytes is approximately 25% homologous with the E. coli protein, mainly in the active site region. 19 Only the reduced form of unmodified human thioredoxin has so far been studied by NMR, because of the presence of three cysteine residues in addition to the two in the active site. The additional thiols cause problems with aggregation of the protein on oxidation.2° Reduced human thioredoxin ~7 H. J. Dyson, A. Holmgren, and P. E. Wright, Biochemistry 28, 7074 (1989). is D. M. LeMaster and F. M. Richards, Biochemistry 27, 142 (1988). 19 H. Eklund, F. K. Gleason, and A. Holmgren, Proteins 11, 13 (1991). ~0 j. D. Forman-Kay, G. M. Clore, P. C. Driscoll, P. Wingfield, F. M. Riehards, and A. M. Gronenborn, Biochemistry 28, 7088 (1989).

296

THIOREDOXIN AND GLUTAREDOXIN

[30]

is soluble to at least 2 mM, and solutions prepared in 100-150 mM phosphate buffer containing 0.2 mM DTT under argon appear to be stable at pH 7.0 or 5.5 for at least 2 months, over a temperature range of 40-500. 20 Solubility is reduced below pH 5.6. 21 The recombinant human thioredoxin isolated from an E. coli expression system is heterogeneous at the N terminus, with about 30% containing an additional methionine residue, 2° although the N-terminal methionine is processed more efficiently during bacterial growth in minimal media. 22

Resonance Assignment It is obvious from Fig. 1 that the assignment of all of the proton resonances of thioredoxin to individual protons in the protein molecule is not possible from the one-dimensional (1D) spectrum alone. The development of two-dimensional (2D) and higher dimensional spectroscopy has been an important breakthrough in the resonance assignment problem for macromolecules, e3Higher dimensions are obtained by frequency and phase modulation in a series of 1D spectra, giving spectra with chemical shift (frequency) scales in two or more dimensions that contain cross-peaks off the planar or body diagonal representing the 1D spectrum. Depending on the sequence of the radiofrequency pulses used to obtain the spectrum, different information is conveyed by the presence of cross-peaks at particular positions. For example, in a 2D correlated (COSY) spectrum, a cross-peak between proton resonances at two frequencies indicates that they are coupled, usually by being attached to adjacent carbon or nitrogen atoms, but in a 2D nuclear Overhauser effect (NOESY) spectrum, a cross-peak indicates that the protons are close together, usually within 5 ~ , but not necessarily coupled or close together in the amino acid sequence. It is in fact the information obtained from long-range NOE cross-peaks obtained from NOESY spectra that allows three-dimensional structures to be calculated from NMR data. The two-dimensional spectra of thioredoxin are well-dispersed, as seen, for example, in Fig. 2, which shows a portion of the COSY spectrum of oxidized E. coli thioredoxin. This particular part of the 2D IH spectrum, containing the cross-peaks between the amide protons and C"H, is known as the fingerprint region, since there is only one cross-peak for each amino acid residue (2 for glycine) in this region of the spectrum. Complete 1H resonance assignments have been obtained for the reduced and oxidized 21 j. D. Forman-Kay, G. M. Clore, and A. M. Gronenborn, Biochemistry 31, 3442 (1992). 2z j. D. Forman-Kay, A. M. Gronenborn, L. E. Kay, P. T. Wingfield, and G. M. Clore, Biochemistry 29, 1566 (1990). 23 K. Wiithrich, "NMR of Proteins and Nucleic Acids." Wiley, New York, 1986.

[30}

N M R OF THIOREDOXIN AND GLUTAREDOXIN

297

-3.0 ~Val 86

~' Ala 29

3.5

~Leu 103 ,~

8J

~lle 41

o, Ala 19

4.0 C,~H chemical 4.5 shift (ppm)

Val 55 5.0

Phe 81

Ale 22

~]Aia 56

5.5

~Leu 78 I

10.0

915

I

9.0

I

I

8.5 8.0 NH chemical shift (ppm)

I

7.5

I

7.0

[

6.5

FIG. 2. Portion of a 500 MHz 2QF COSY spectrum of oxidized thioredoxin in 90% IH20/ 10% 2H20, pH 5.7, 30°. This part of the spectrum, known as the fingerprint region, shows the cross-peaks that arise from the coupled N H and C~H protons of the amino acid residues, and each four-lobed cross-peak shows one such connectivity. The inset shows schematically the position of this region in the full 2D spectrum of the protein.

forms of E. coli thioredoxin 17 and for the reduced form of human thioredoxin 2° using IH two-dimensional methods alone.

Structure Determination from 1H Data Once resonance assignments have been obtained, it is possible to calculate a three-dimensional structure of the molecule. This is done by extracting interproton distances from the NMR spectrum via the nuclear Overhauser effect (NOE). An NOE is typically observed between two protons that lie in the folded protein within about 5 A, and the intensity of the NOE is proportional to the inverse sixth power of the internuclear distance. Observation and quantitation of the NOEs between protons, especially those between residues far distant in the sequence, give a set of distance restraints that can be used together with other information such as dihedral angle restraints provided by spin-spin coupling constants, distance geometry, or similar algorithms to give the three-dimensional structure. Such a structure has been obtained for reduced E. coli thioredoxin from 2D 1H

298

THIOREDOXIN AND GLUTAREDOXIN

[30]

FIG. 3. Stereo view showing the best-fit superpositions of the backbone (N, C ", and C) atoms of 12 NMR solution structures of reduced E. coli thioredoxin (residues 3-108) obtained using the distance geometry program DISGEO followed by molecular dynamics refinement with AMBER. The active site (residues 32-35) is at the bottom right of the structure. [Adapted with permission from H. J. Dyson, G, P. Gippert, D. A. Case, A. Holmgren, and P. E. Wright, Biochemistry 29, 4129 (1990). Copyright 1990 American Chemical Society.]

structure has been obtained for reduced E. coli thioredoxin from 2D ~H data, 7 and a superposition of 12 of these structures is shown in Fig. 3. N M R solution structures are usually presented as families. The slight differences in the calculated structures arise because the distances and dihedral angles obtained from N M R data are not known accurately. The data are therefore presented as a set of boundary conditions (upper and lower bounds) rather than as discrete distances or angles. Individual structures calculated using these ranges may have slightly different starting and end points, hence the variability in the structures. The correspondence between the independently calculated structures [usually expressed as a root-mean-square difference (rmsd) between the structures or from a mean structure] gives a measure of how well the distance and dihedral angle restraints determine the structure.

Multidimensional Heteronuclear Nuclear Magnetic Resonance and Determination of High-Resolution Structures Although the homonuclear proton spectrum in this case allowed us to obtain a recognizable and even quite high-resolution structure by some criteria, 24 the problem of resonance overlap in two-dimensional spectra of thioredoxin is still substantial. The introduction of the stable NMR-active 24 M. W. MacArthur and J. M. Thornton, Proteins 17, 232 (1993).

[301

NMR

OF THIOREDOXIN AND GLUTAREDOXIN

299

nuclei 13C and 15N into the protein gives much greater resolution in multidimensional spectra and allows the assignment of a much greater proportion of the available NOEs as distance restraints,2 as well as a more accurate determination of dihedral angle constraints from a number of homonuclear and heteronuclear coupling constants. The additional resolution obtainable from three-dimensional heteronuclear NMR spectra is illustrated in Fig. 4 for oxidized E. coli thioredoxin: the dispersal of resonances into a third dimension according to the z3C chemical shift greatly reduces resonance overlap and allows unambiguous assignment of many more NOE crosspeaks. Assignments of the 15N and ~3C spectra have been made for both oxidized and reduced E. coli thioredoxin,~5"25and high-resolution structures for both forms of the protein using the greater volume of data available from the three-dimensional heteronuclear spectra have been calculated, s Extensive 15N assignments and coupling constant measurements have been made for reduced human thioredoxin,22 and a high-resolution structure has been calculated. 9 The 13C, 1H, and 15N resonance assignments have also been made for the reduced form of a triple mutant of human thioredoxin, in which the three cysteine residues other than those in the active site have been mutated to alanines.26 The mutant has been used in structural studies for the comparison of the oxidized and reduced forms of human thioredoxin. ~° The three-dimensional structures of the E. coli and human thioredoxins are quite similar. The E. coli protein in both the X-ray crystal structure 5 and the NMR solution structuresv,8 has a central five-stranded twisted t3 sheet with two long, regular ot helices and two shorter helices, one irregular helix and one apparently 310 helix. The solution structure of reduced human thioredoxin9,~° shows a similar overall topology, and the central 13 sheet is retained, including the presence of a 13 bulge between the fourth and fifth 13 strands. There are also four helices in human thioredoxin, but they are in general longer and more regular? The active site is similar among all of the thioredoxin structures studied, consistent with the high degree of sequence similarity in this region.

Deuterium Labeling of Thioredoxin Oxidized Escherichia coli thioredoxin has also been extensively used as a paradigm for a number of methodological advances in the labeling of proteins for NMR. A number of papers from the laboratories of LeMaster 25 K. Chandrasekhar, A. P. Campbell, M.-F. Jeng, A. Holmgren, and H. J. Dyson, J. Biomol. N M R 4, 411 (1994). 26 j. D. Forman-Kay, G. M. Clore, S. J. Stahl, and A. M. Gronenborn, J. Biomol. N M R 2~ 431 (1992).

38: