METHODS
IN
M O L E C U L A R B I O L O G Y TM
Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK
For further volumes: http://www.springer.com/series/7651
.
Single Molecule Analysis Methods and Protocols
Edited by
Erwin J.G. Peterman and Gijs J.L. Wuite Department of Physics and Astronomy, VU University Amsterdam, Amsterdam, The Netherlands
Editors Erwin J.G. Peterman Department of Physics and Astronomy VU University Amsterdam Amsterdam, The Netherlands
[email protected] Gijs J.L. Wuite Department of Physics and Astronomy VU University Amsterdam Amsterdam, The Netherlands
[email protected] ISSN 1064-3745 e-ISSN 1940-6029 ISBN 978-1-61779-281-6 e-ISBN 978-1-61779-282-3 DOI 10.1007/978-1-61779-282-3 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2011935030 ª Springer ScienceþBusiness Media, LLC 2011 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer ScienceþBusiness Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper Humana press is a part of Springer Science+Business Media (www.springer.com)
Preface Life scientists have been brought up for ages with the idea that life is driven, directed, and shaped by biomolecules, working on their own or in concert. Only since a decade of two to three, it has been possible to study the properties of molecules in ultimate isolation: individual molecules. Technical breakthroughs in the field of sensitive fluorescence microscopy have made it possible to observe single fluorescent molecules and measure their properties. Other researchers have developed optical tweezers into a method to measure the mechanic properties of single molecules. Around the same time, atomic force microscopy has been developed, with a spatial resolution good enough to resolve single biomolecules. Together, these techniques (and several other ones) have been applied more and more to the study of biologically relevant molecules, such as DNA, DNA-binding proteins, and motor proteins. These single-molecule approaches have led both not only to new views into how biomolecules bring about biology, but also to novel insights in the way physical and statistical principles underlie the behavior and mechanism of biomolecules. By now, single-molecule tools are slowly becoming commonplace in molecular biophysics, biochemistry, and molecular and cell biology. Thanks not only to their success, but also to their accessibility: in the beginning, these tools were solely developed and custom-built by (bio)physicists; now, commercial tools are becoming available. We foresee that this trend will prevail, and single-molecule tools will play an even more prominent role in molecular biology. The aim of Single-Molecule Techniques: Methods and Protocols is to provide a broad overview of single-molecule approaches applied to biomolecules, on the basis of clear and concise protocols. In addition, we provide a solid introduction to the most widely used single-molecule techniques. The idea is that these introductions, together with the protocols provide enough basis for nonspecialists to make the step to single-molecule experiments. The protocols contain a “Notes” section, in which the authors provide tips and tricks, rooted in experience, that are often decisive between failure and success. Our selection of topics for Single-Molecule Techniques: Methods and Protocols cannot be all-inclusive, but we hope that we have covered the most important ones, which can serve as a starting point for further exploration of single-molecule methods. The volume opens with four chapters on optical tweezers. In Chapter 1, a general overview of the method is provided. In the next chapters, protocols of applications of optical tweezers to studies of DNA/RNA (Chapters 2 and 3) and motor proteins (Chapter 4) are presented. The second part of the volume (Chapters 5–10) deals with single-molecule fluorescence tools. First, a general overview of these techniques is provided (Chapter 5), followed by protocols for fluorescent labeling of proteins (Chapter 6). In the following chapters, applications to motor proteins (Chapter 7), structural proteins in living bacteria (Chapter 8) and DNA (Chapter 9) are explained. In the last chapter of this part, the method fluorescence correlation spectroscopy is presented in detail (Chapter 10). The next part of the volume deals with atomic force microscopy (Chapters 11–14). Also this part opens with a general overview of the approach (Chapter 11), followed by protocol chapters describing applications to DNA and DNA-binding proteins (Chapter 12), protein folding and unfolding (Chapter 13), and viruses (Chapter 14). In the remaining chapters of the book methods
v
vi
Preface
that fall outside these three categories are discussed, including magnetic tweezers (Chapter 15) and tethered particle motion (Chapter 16). We have taken great care to provide a broad and thorough overview of the exciting and still emerging field of single-molecule biology in this volume Single-Molecule Techniques: Methods and Protocols. It is unavoidable that there is some overlap between the chapters. Furthermore, it might well be that within a few years new techniques have emerged and become important that are not discussed here. Nevertheless, we hope that the presented protocols will be useful to many researchers, inspire them and help them to go single molecule! Amsterdam, The Netherlands
Erwin J.G. Peterman Gijs J.L. Wuite
Contents Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction to Optical Tweezers: Background, System Designs, and Commercial Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Joost van Mameren, Gijs J.L. Wuite, and Iddo Heller 2 Optical Trapping and Unfolding of RNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Katherine H. White and Koen Visscher
v ix
1
3
DNA Unzipping and Force Measurements with a Dual Optical Trap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ismaı¨l Cisse´, Pierre Mangeol, and Ulrich Bockelmann
Probing the Force Generation and Stepping Behavior of Cytoplasmic Dynein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Arne Gennerich and Samara L. Reck-Peterson 5 A Brief Introduction to Single-Molecule Fluorescence Methods . . . . . . . . . . . . . . . . Siet M.J.L. van den Wildenberg, Bram Prevo, and Erwin J.G. Peterman
1 21
45
4
6 7
8
63 81
Fluorescent Labeling of Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Mauro Modesti Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 Till Korten, Bert Nitzsche, Chris Gell, Felix Ruhnow, Ce´cile Leduc, and Stefan Diez Exploring Protein Superstructures and Dynamics in Live Bacterial Cells Using Single-Molecule and Superresolution Imaging. . . . . . . . . . . . . . . . . . . . . 139 Julie S. Biteen, Lucy Shapiro, and W.E. Moerner
9
Fluorescence Microscopy of Nanochannel-Confined DNA . . . . . . . . . . . . . . . . . . . . . 159 Fredrik Persson, Fredrik Westerlund, and Jonas O. Tegenfeldt 10 Fluorescence Correlation Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 Patrick Ferrand, Je´roˆme Wenger, and Herve´ Rigneault 11
Introduction to Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 Pedro J. de Pablo
12
Sample Preparation for SFM Imaging of DNA, Proteins, and DNA–Protein Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 Dejan Ristic, Humberto Sanchez, and Claire Wyman Single-Molecule Protein Unfolding and Refolding Using Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 Thomas Bornschlo¨gl and Matthias Rief
13
14
How to Perform a Nanoindentation Experiment on a Virus. . . . . . . . . . . . . . . . . . . . 251 Wouter H. Roos
vii
viii
Contents
15
Magnetic Tweezers for Single-Molecule Manipulation . . . . . . . . . . . . . . . . . . . . . . . . 265 Yeonee Seol and Keir C. Neuman
16
Probing DNA Topology Using Tethered Particle Motion. . . . . . . . . . . . . . . . . . . . . . 295 David Dunlap, Chiara Zurla, Carlo Manzo, and Laura Finzi
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
Contributors JULIE S. BITEEN • Department of Chemistry, University of Michigan, Ann Arbor, MI, USA ULRICH BOCKELMANN • Laboratoire de Nanobiophysique, UMR Gulliver CNRS – ESPCI, Paris, France THOMAS BORNSCHLO¨GL • Department of Physics, TU Munich, Garching, Germany ISMAI¨L CISSE´ • Laboratoire de Nanobiophysique, UMR Gulliver CNRS – ESPCI, Paris, France STEFAN DIEZ • Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany DAVID DUNLAP • Department of Cell Biology, Emory University, Atlanta, GA, USA PATRICK FERRAND • Mosaic Group, Institut Fresnel, CNRS, Aix-Marseille Universite´, Ecole Central Marseille, Marseille, France LAURA FINZI • Department of Physics, Emory University, Atlanta, GA, USA CHRIS GELL • Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany ARNE GENNERICH • Department of Anatomy and Structural Biology, Gruss-Lipper Biophotonics Center, Albert Einstein College of Medicine, Bronx, NY, USA IDDO HELLER • Department of Physics and Astronomy, VU University, Amsterdam, The Netherlands TILL KORTEN • Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany CE´CILE LEDUC • Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany JOOST VAN MAMEREN • JPK Instruments AG, Berlin, Germany PIERRE MANGEOL • Laboratoire de Nanobiophysique, UMR Gulliver CNRS – ESPCI, Paris, France CARLO MANZO • Department of Physics, Emory University, Atlanta, GA, USA MAURO MODESTI • Laboratory of Genome Instability and Carcinogenesis, CNRS, Marseille, France W.E. MOERNER • Department of Chemistry, Stanford University, Stanford, CA, USA KEIR C. NEUMAN • Laboratory of Molecular Biophysics, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, USA BERT NITZSCHE • Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany PEDRO J. DE PABLO • Departamento de Fı´sica de la Materia Condensada Universidad Auto´noma de Madrid, Madrid, Spain FREDRIK PERSSON • Department of Physics, University of Gothenburg, Gothenburg, Sweden ERWIN J.G. PETERMAN • Department of Physics and Astronomy, VU University Amsterdam, Amsterdam, The Netherlands BRAM PREVO • Department of Physics and Astronomy, VU University Amsterdam, Amsterdam, The Netherlands ix
x
Contributors
SAMARA L. RECK-PETERSON • Department of Cell Biology, Harvard Medical School, Boston, MA, USA MATTHIAS RIEF • Department of Physics, Technical University Munich, Garching, Germany HERVE´ RIGNEAULT • Mosaic Group, Institut Fresnel, CNRS, Aix-Marseille Universite´, Ecole Central Marseille, Marseille, France DEJAN RISTIC • Department of Cell Biology and Genetics, Erasmus Medical Center, Rotterdam, The Netherlands WOUTER H. ROOS • Department of Physics and Astronomy, VU University Amsterdam, Amsterdam, The Netherlands FELIX RUHNOW • Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany HUMBERTO SANCHEZ • Department of Cell Biology and Genetics, Erasmus Medical Center, Rotterdam, The Netherlands YEONEE SEOL • Laboratory of Molecular Biophysics, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, USA LUCY SHAPIRO • Department of Developmental Biology, Stanford University, Stanford, CA, USA JONAS O. TEGENFELDT • Department of Physics, Lund University, Lund, Sweden Department of Physics, University of Gothenburg, Gothenburg, Sweden KOEN VISSCHER • Department of Physics, University of Arizona, Tucson, AZ, USA JE´ROˆME WENGER • Mosaic Group Institut Fresnel, CNRS, Aix-Marseille Universite´, Ecole Central Marseille, Marseille, France FREDRIK WESTERLUND • Department of Physics, University of Gothenburg, Gothenburg, Sweden KATHERINE H. WHITE • Department of Chemistry and Biochemistry, University of Arizona, Tucson, AZ, USA SIET M.J.L. VAN DEN WILDENBERG • Department of Physics and Astronomy VU University Amsterdam, Amsterdam, The Netherlands GIJS J.L. WUITE • Department of Physics and Astronomy, VU University Amsterdam, Amsterdam, The Netherlands CLAIRE WYMAN • Department of Cell Biology and Genetics, Erasmus Medical Center, Rotterdam, The Netherlands Department Radiation Oncology, Erasmus Medical Center, Rotterdam, The Netherlands CHIARA ZURLA • Department of Physics, Emory University, Atlanta, GA, USA
Chapter 1 Introduction to Optical Tweezers: Background, System Designs, and Commercial Solutions Joost van Mameren, Gijs J.L. Wuite, and Iddo Heller Abstract Optical tweezers are a means to manipulate objects with light. With the technique, microscopically small objects can be held and steered while forces on the trapped objects can be accurately measured and exerted. Optical tweezers can typically obtain a nanometer spatial resolution, a piconewton force resolution, and a millisecond time resolution, which make them excellently suited to study biological processes from the single-cell down to the single-molecule level. In this chapter, we provide an introduction on the use of optical tweezers in single-molecule approaches. We introduce the basic principles and methodology involved in optical trapping, force calibration, and force measurements. Next, we describe the components of an optical tweezers setup and their experimental relevance in single-molecule approaches. Finally, we provide a concise overview of commercial optical tweezers systems. Commercial systems are becoming increasingly available and provide access to single-molecule optical tweezers experiments without the need for a thorough background in physics. Key words: Optical tweezers, Optical trap, Radiation pressure, Single molecule, Trap stiffness calibration, Force spectroscopy, Instrument design, Commercial optical tweezers, Molecular motors, DNA–protein interactions
1. Introduction 1.1. History of Optical Tweezers
At the heart of optical tweezers techniques is the interaction between light and matter. The minute forces that are generated in this interaction can be used to displace and trap microscopic objects. In 1970, Ashkin laid the foundations for present-day optical tweezers techniques. At Bell labs, Ashkin observed that micron-sized latex spheres (beads) were attracted toward the center of an argon laser beam of a few mW power (1). It is this attractive force that makes optical trapping possible. Ashkin also observed, however, that the laser light scattered and propelled the beads forward. By using two counterpropagating beams, he managed to avoid forward propulsion, and thus created the first stable
Erwin J.G. Peterman and Gijs J.L. Wuite (eds.), Single Molecule Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 783, DOI 10.1007/978-1-61779-282-3_1, # Springer Science+Business Media, LLC 2011
1
2
J. van Mameren et al. forward scattered light is deflected
external force on trapped particle
trapped particle
forcused laser
restoring force
forcused laser
Fig. 1. Schematic of optical trapping. Left: A tightly focused laser beam (cone ) attracts refractive objects (dark sphere ), such as glass beads, nanoparticles, or even whole cells to its focus. Right : External forces pushing or pulling on the particle slightly displace it from the center of the focus, leading to a slight deflection of the forward-scattered laser light. This deflection forms the basis for quantitatively detecting the forces and displacements experienced by the trapped object.
optical trap for beads suspended in water. It was not until 1986 that Ashkin together with Chu and others demonstrated the present form of optical tweezers that uses a single, tightly focused laser beam to stably trap particles – of diameters between 25 nm and 10 mm – in three dimensions (see Fig. 1, left) (2). Later on, Chu and others used techniques inspired by optical tweezers to trap and cool atoms, which brought him the 1997 Nobel Prize in physics (3, 4). 1.2. Optical Tweezers in Biology
Currently, optical tweezers have found widespread applications in biology (5–7). One of the important reasons for the success of optical tweezers in biology is that it provides biological scientists with “microscopic hands” to manipulate biological objects and feel or exert forces, yet with the same low level of invasiveness as light microscopy techniques. Furthermore, the length scales, time scales, and force scales accessible to optical tweezers are biologically relevant from the single-cell down to the single-molecule level. In 1987, Ashkin presented the first applications of optical tweezers in biology by manipulating individual viruses and living bacteria (8). By a correct choice of laser power and wavelength, photodamage to biological samples could be minimized, which allowed trapping and manipulation of single living cells (9). Since
1 Introduction to Optical Tweezers
3
Fig. 2. Prototypical single-molecule optical tweezers assays. Top : A single kinesin motor protein bound with its two heads to an optically trapped bead moves along a surface-immobilized microtubule track. Its 8-nm steps, the forces exerted, and the mechanics of the stepping have been elucidated in such assays. Bottom: DNA suspended between two optically trapped beads.
the late 1980s, optical tweezers approaches have been extended down to the single-biomolecule level (10–20). In these singlemolecule studies, the biomolecules of interest are not themselves trapped directly, but are manipulated through optically trapped microbeads that act as handles and force transducers. A large fraction of this single-molecule work includes the study of the activity of individual motor proteins (10, 14, 20). With optical tweezers, the motion and forces generated by these motor proteins have been studied and controlled to reveal their dynamics and energetics (Fig. 2, top). Another important area of research includes the study of biopolymers, such as DNA (7, 13, 16, 21). In these experiments, the DNA molecule is attached to one or more optically trapped beads which allows stretching the molecule and studying its mechanical properties through force spectroscopy (Fig. 2, bottom). In addition, this layout has been used to studyDNA–protein interactions (12, 15, 19, 22). A wide range of DNA–protein interactions affect the structure of DNA, and thus
4
J. van Mameren et al.
the (force-dependent) length of the DNA molecules. In optical tweezers, these length changes can be observed by measuring the displacements of the microbeads. Examples include the study of DNA-binding proteins and the activity of DNA and RNA polymerases. With the advent of commercial optical tweezers systems in recent years, this powerful single-molecule technique is approaching maturation and is becoming more and more accessible to a wide range of biological scientists. As with the development of commercial fluorescence and AFM techniques, it is to be expected that commercial optical tweezers greatly contribute to our knowledge of biology on the single-molecule level. As a final motivation to read more about optical tweezers: in an interview with Physics Today, Nobel Prize winner Steven Chu said that he would not be surprised if in the coming decennium another Nobel Prize would be attributed to groundbreaking discoveries in molecular biology facilitated by optical tweezers or other single-molecule techniques (23).
2. Principles of Optical Tweezers Techniques
2.1. Forces in an Optical Trap
The basic physical principle underlying optical tweezers is the radiation pressure exerted by light when colliding with matter. For macroscopic objects, the radiation pressure exerted by common light sources is orders of magnitude too small to have any measurable effect: we do not feel the light power of the sun pushing us away. However, for objects of microscopic dimensions (> l, one speaks of the “ray-optics” regime while the regime where d fc, however, the power spectrum falls off like 1/f 2, which is characteristic of free diffusion. The inverse of the corner frequency represents the time response of the optical trap, which is typically in the order of 1–0.1 ms. For shorter timescales, the particle does not “feel” the confinement of the trap which means that behavior of biological systems at timescales more rapid than this response time cannot be detected by the optical tweezers. The two power spectra of Fig. 4, acquired at two different trap stiffnesses, illustrate that when higher trap stiffnesses are used (i.e., by increasing the laser power), the bead fluctuations at low frequencies are reduced and the time response of the optical trap increases. On the other hand, a higher trap stiffness implies smaller bead displacement at a given force which means that more accurate bead displacement detection is required to obtain the same force sensitivity. It is important to note that, in practice, the detector used to determine the bead position reads uncalibrated displacement fluctuations u(t) (i.e., as some voltage rather than as a displacement
8
J. van Mameren et al.
in nanometers). The response of the detector R, which has units m/V, relates the displacement to u(t) as x(t) ¼ Ru(t). To fully calibrate an optical trap, the power spectrum of the uncalibrated displacement fluctuations Su(f) is fitted with a Lorentzian: Su ðf Þ ¼
Su; 0 fc 2 : ðfc 2 þ f 2 Þ
Once the parameters Su;0 and fc are obtained, the trap stiffness can be calculated using k¼
2kB T pSu;0 fc
or k ¼ 2pgfc ;
and, providing the bead diameter and solvent viscosity are known, the detector response can be calculated using "
kB T R¼ 2 p gSu;0 fc 2
"
#1=2 25 C
¼
#1=2 5:0 1020 m3 =s : Su;0 fc 2 d
Finally, to convert uncalibrated displacement data to forces, the displacement signal should be multiplied by R and the trap stiffness such that F ¼ kx ¼ kRu.
3. Optical Tweezers Systems An optical tweezers setup consists of various dedicated components. In this section, we discuss the role of these components in optical tweezers function and performance. Below, we divide and discuss the components in five groups: the trap, the environment of the trap, trap steering, position and force detection, and the environment of the setup. In addition, we also discuss the ability to combine optical tweezers with other techniques and the different optical trapping assays that are typically used in biological experimentation. For illustration, Fig. 5 shows the schematic layout of an optical tweezers setup that combines two steerable optical traps with fluorescence microscopy. 3.1. The Optical Trap
At the heart of every single-beam optical tweezers instrument is the microscope objective, which creates a tight focus to form a stable optical trap. Tight focusing implies that a significant fraction of the incident light comes in at large angles such that the scattering force is overcome by the gradient force. The maximum incidence angle of the light Ymax is determined by the numerical aperture (NA) of the objective used to focus the laser beam. This is a measure for the solid angle over which the objective lens can
1 Introduction to Optical Tweezers
9
InGaAs InGaAs
InGaAs QPD (x,y) signal attenuation
laser
TRAP 2
TRAP 1
isolator main shutter
shutters
moving
OPTICAL MICROSCOPE
TRAP STEERING
power control
LASER
Fig. 5. Optics layout of a typical optical tweezers instrument. The various parts of the system as described in the text are bounded by dashed boxes. The depicted layout was adopted from that of the “NanoTracker” commercial optical tweezers platform of JPK Instruments (see Subheading 5).
gather light and is defined as: NA ¼ n sin Ymax, where n is the refractive index of the immersion medium (i.e., the medium between the objective lens and the sample) and Ymax is one-half the angular aperture. The value of n varies between 1.0 for air and 1.5 for most immersion oils. For typical oil-immersion objectives, an NA of 1.4 corresponds to a total acceptance angle of the objective of about 130 . To obtain a stable three-dimensional optical trap, the laser beam entering the objective has to be wide enough to fill or overfill the back aperture of the objective. This way, one provides sufficient convergent, high-angle rays that contribute to counteracting the scattering force. The maximum forces that can be exerted by an optical trap can be enhanced by either increasing the laser power or optimizing the quality of the focal spot of the laser and the refraction of the laser by the trapped particle (i.e., choosing the proper optics and materials with proper optical densities; see below). The laser power can be increased only up to a certain limit, above which more laser light would lead to heating or photodamage of the examined system (often delicate biomaterials) or even of the optics in the instrument (27). Therefore, care must be taken to optimize the quality of the focal spot of the trapping laser. The difference in refractive index of the trapped object n2 compared to that of the
10
J. van Mameren et al.
surrounding medium n1 determines how strongly the incident rays are refracted and, consequently, how strong the trapping force is. The required balance between gradient and scattering forces yields an optimal refractive index of n2 ¼ 1.69 (5). One often uses silica (glass) particles (n2 ¼ 1.37–1.47) or polystyrene particles (n2 ¼ 1.57). The trapping forces that can be obtained for polystyrene particles are, thus, higher. When using an oilimmersion objective, the refractive index of the immersion oil matches both that of the objective lens and that of the glass of the sample. Therefore, the maximum NA can be achieved with oil-immersion objectives. However, due to the refractive index mismatch between the sample glass and the buffer, spherical aberrations deteriorate the quality of the laser focus when the distance of the focus to the sample surface increases (optical trapping in water with oil-immersion objectives is typically performed within several tens of micrometers of the glass surface) (28). To allow equally stable trapping at any distance to the surface, waterimmersion objectives are often used. Despite the fact that the NA is somewhat compromised (typically, NA ¼ 1.2 for waterimmersion objectives), the ability to move away from the sample surface without lowering the trap quality can be a good reason to use water-immersion objectives. For optical trapping, single-mode continuous wave lasers with a Gaussian beam profile (i.e., operated in the lowest, TEM00, mode) are commonly used. The laser power typically ranges from a few hundred milliwatts up to several watts. Important properties of the laser for stable trapping are low intensity fluctuations and high pointing stability (little angular and transverse wandering of the beam). Of particular interest for biological experiments is the wavelength. Since the light intensity at the focus is very high, heating and damage through light absorption by the often delicate biological material need to be considered. Near-infrared lasers (800–1,200 nm, most often the 1,064-nm line from diode-pumped solid-state lasers) are typically used because of the low absorption of biological material as well as water in this spectral range (5). Although in current biological applications optical traps are most commonly formed by tightly focusing a single laser beam (2), the first stable optical trap was accomplished in 1970 by using two counterpropagating focused laser beams (1). The counterpropagating layout does not require tight focusing such that low NA (ms). Certain fluorophores show different on–off blinking behavior, for example, fluorescent proteins are known to undergo cis–trans isomerization and intramolecular proton transfer, both resulting in long-lived dark states (18). Another example is the photoblinking of QDs, which is caused by ejection of electrons from the semiconductor core (19). Considering all the fluorophore characteristics mentioned above the ideal single-molecule fluorophore (1) has high fluorescence quantum yield and molecular extinction coefficient, (2) has welldefined excitation and emission wavelengths, (3) shows steady emission intensity (20), (4) can be followed for a long time using high illumination intensity, (5) does not affect the natural behavior of the single molecule, (6) shows no blinking behavior, (7) can be easily attached to the molecule of interest, (8) is soluble in buffers used (21), and (9) its characteristics are well described. 2.3. Fluorophores Used for Single-Molecule Research
Next, we take a closer look at the different classes of fluorescent labels and compare and discuss some of their characteristics important for single-molecule research (Fig. 2). (1) Synthetic
Fig. 2. Structure, size, and spectra of different fluorescent probes used in single-molecule fluorescence microscopy. (a) Chemical structure and absorption/emission spectrum of the synthetic dye rhodamine 6G (R6G). (b) Structure (Protein Databank entry 1S6Z (58)) and absorption/emission spectrum of enhanced Green Fluorescent Protein (eGFP). (c) Schematic representation of a functionalized QD consisting of a core and different shells and corresponding absorption/emission spectrum.
5 A Brief Introduction to Single-Molecule Fluorescence Methods
87
dyes have been around for decades, are commercially available and constructed to suit the means of use. They are constructed in a way that they contain different reactive groups, such as maleimides or succynimidyl esters, which can be used for attachment of the label to a protein or biomolecule of interest. Succynimidyl esters react with free amino groups, which are available in large quantities on the surface of most proteins. Maleimides or other sulfhydryl reactive probes can be used for more specific labeling of cystein residues, which are generally less abundant. Because synthetic dyes have been around for a long time, their characteristics have been optimized and labeling protocols are widely available. Their small size (~0.5 nm) minimizes the chance of causing steric hindrance to the labeled molecule. Cyanine and rhodamine dyes (Fig. 2a) are most often used for in vitro single-molecule research, in particular Cy3, Cy5, Alexa555 of the cyanine family and R6G and Texas Red of the rhodamine family (20). (2) QDs are very bright fluorophores with a very wide range of absorption wavelengths, narrow (about 10 nm) and symmetric emission bands and quantum yields close to 90% (Fig. 2c) (22, 23). QDs in general consist of a CdSe, CdTe, InP, or InAs core and a ZnS shell. Their size, shape, and structure can be controlled precisely, in order to tune the emission from visible to infrared wavelengths (24, 25). For biological applications, QDs are normally coated to make them hydrophylic or prepare them for specific attachment to the biomolecule of interest (26). Compared to organic dyes, QDs are brighter (molecular extinction coefficients between 105 and 106 L/mol/cm). They are also more photostable and therefore can be followed longer and their position can be determined with higher accuracy (21). However, their size varies between 6 and 60 nm (with coating), which is relatively large in comparison to organic dyes (~0.5 nm). In addition, as mentioned above, they can suffer from on/off blinking on a wide range of time scales (19). (3) Green fluorescent protein (GFP, Fig. 2b) is an autofluorescent protein from the jellyfish Aequorea Victoria, which is very well suited for (but not restricted to) in vivo applications. A key advantage of GFP is that the protein is genetically encoded, does not require a cofactor and that every protein copy can be labeled by fusing its gene with that of GFP (27). Another advantage is that it is less sensitive to its surroundings than many synthetic dyes (28). Disadvantages of GFP are that it is rather large (27 kDa, ~4 nm in diameter), its emission shows blinking behavior and its photostability is substantially less than good synthetic dyes (18). By now, many different variants with different colors, optimized for different organisms have been developed on the basis of the A. Victoria protein and related proteins in other organisms (29). A very exciting recent development is the generation of photoactivatable and photoswitchable fluorescent proteins that are very well suited for superresolution methods like PALM (30). (4) Biological
88
S.M.J.L. van den Wildenberg et al.
materials also contain naturally occurring fluorescent molecules. Many of these (such as tryptophan and NADH), which are widely used in bulk fluorescence measurements, are not photostable enough and absorb and emit too far in the ultraviolet to allow single-molecule detection. Protein cofactors, such as chlorophyll and flavin, can be very fluorescent and have been used for in vitro applications (31, 32). Their occurrence is however limited to a small subset of proteins, which hinders general application.
3. Methods 3.1. Microscopic Detection of Single Fluorophores
Single, fluorescently labeled particles can be detected using a fluorescence microscope. The main components of such a microscope are an illumination source for excitation of the fluorophores, filters to extract light of wanted colors and suppress unwanted light, an objective to direct the excitation light and efficiently collect the emission light, and a detector. Since much of the emitted light is lost in the detection pathway and the total number of photons that a fluorophore can emit is restricted due to photobleaching, it is imperative in single-molecule fluorescence studies to use optimal components in each part of the instrument. In the next section, we discuss the different parts of a fluorescence microscope in more detail.
3.1.1. Light Sources
To maximize the signal obtained from a single fluorophore, it is essential to excite it with a wavelength close to its absorption maximum. In bulk fluorescence microscopy often broad-band sources, such as metal halide or mercury arc lamps, are used (33). A key advantage of these sources is that they are broad-band and contain several intense spectral lines. This allows them to be used to excite spectrally distinct fluorophores, with the proper excitation filters to suppress unwanted lamp light. Relatively new on the market are LED sources, which are more monochromatic than lamps (spectra with a width of several tens of nanometers). LEDs are much more energy efficient and thus produce less heat. In single-molecule applications, in most cases, lasers are used for fluorescence excitation since they emit monochromatic and collimated light, allowing better separation of excitation from fluorescence light and a more straightforward construction of complex optical paths using mirrors, lenses, filters, polarizers, etc. In addition, collimated laser beams can be focused to diffraction limited spots (see below), which is indispensable for confocal fluorescence microscopy. Apart from cost, there are two key disadvantages of lasers. (1) Laser beams are Gaussian and this can result in uneven excitation intensity profiles. (2) Lasers are in many cases monochromatic meaning that for
5 A Brief Introduction to Single-Molecule Fluorescence Methods
89
each spectrally distinct fluorophore an additional laser needs to be purchased and installed. 3.1.2. Filters and Dichroic Mirrors
Optical filters are used to separate fluorescence light from scattered excitation light and other background signals. On the basis of their transmission spectra two classes of filters can be discerned: edge filters and band-pass filters. Edge filters transmit light above (long-pass) or below (short-pass) a specific wavelength, and block the other light. Band-pass filters transmit only a narrow range of wavelengths and block wavelengths on either side of this range. The performance of the filters depends on three aspects: the percentage of transmission of the desired light, the optical density in the blocked region of the spectrum, and the steepness of the edges between the transmitted and blocked regions (14). In the past, filters were based on stained glass, which often suffered from considerable autofluorescence. Later, thin film interference filters, consisting of repetitive, thin layers evaporated on a surface were developed with substantially better performance and flexibility. Recently, new technologies to make precise thin layers based on ion-beam sputtering have further improved filter performance. Nowadays, filters that are designed for simultaneous excitation and/or detection of several, spectrally distinct fluorophores are available. A typical fluorescence microscope consists of three optical filters. (1) The excitation filter that selects one line or band from excitation source to illuminate the sample. In many cases, the use of an excitation filter is not required when using lasers. (2) A dichroic beam-splitting mirror that reflects excitation light in the direction of the objective, but transmits fluorescence light collected by the objective. (3) The emission filter that is used to select for an emission band and block any residual excitation light. In principle, for each light source and dye combination a separate combination of these filters needs to be used.
3.1.3. Detectors in SingleMolecule Fluorescence Microscopy
In single-molecule fluorescence methods, the number of photons emitted is very limited, putting forward strong demands on the quantum efficiency and noise characteristics of detectors used. Broadly speaking, two distinct classes of detectors can be used, depending on the imaging modality (see below) (14, 34). (1) Point detectors, such as avalanche photodiodes (APD) and photomultipliers, which do not provide position information but are capable of counting single photons with high time resolution. APDs have high quantum efficiency, but a small active area, which makes alignment tedious. Photomultipliers usually have a lower quantum efficiency and worse dark noise characteristics, but a larger active area. For single-molecule applications, photomultipliers are usually only used in the blue part of the spectrum, where APDs perform poorly. In general, point detectors are used in
90
S.M.J.L. van den Wildenberg et al.
confocal instruments or when time resolution is important. Key applications are in burst analysis of diffusing molecules, for example, in fluorescence correlation spectroscopy (FCS) (35, 36) or in FCS-like experiments (37). (2) Array detectors, such as charged coupled devices (CCDs) are the most widely used detectors in wide-field fluorescence microscopy. CCD detectors are twodimensional array detectors that can be read out in one of three ways, full-frame, frame-transfer, or interline-transfer. The latter two are fast and allow for continuous detection. Limitations of CCD detectors are read-out noise due to analog-to-digital conversion (a problem solved in modern electron multiplying CCDs) and the relatively slow speed: an entire frame is integrated and read out. For single-molecule detection, frame rates up to ~100 per second can be reached. Key advantages are the almost unity quantum efficiency (in the visible) and the very low dark currents. These detectors are, therefore, optimally suited in conditions when acquisition times of down to ~10 ms are sufficient (38). 3.1.4. Microscope Objectives
The key optical element in a fluorescence microscope is the objective. The objective concentrates the excitation light in the sample and collects the emitted light. In principle, an objective is nothing more than a strong lens, but in order to fulfill its tasks it needs to be highly corrected for optical aberrations, which can only be achieved by complex designs involving multiple optical elements. Important properties with respect to single-molecule experiments are the magnification, which determines the size of the field of view and the number of pixels over which a single fluorescent particle is imaged and the numerical aperture (NA), which is a measure of the angle over which photons can be detected, defined as (33): NA ¼ n sin ymax :
(4)
With n is the refractive index of the medium between the sample and the objective, and ymax is the maximum half-angle of the collection cone of the objective. A fluorophore emits light in all directions and to maximize the collected light the NA of the objective should be as high as possible. This may be achieved by using an objective designed to be immersed in a high-n medium (14, 39). Furthermore, the NA is not only important for collection efficiency, but it also determines the resolution of the optical system, as is discussed in the next section. 3.1.5. The Resolution of a Fluorescence Microscope
A fluorescent molecule is much smaller than the wavelength of visible light and after excitation the molecule emits photons in random direction. Using a single microscope objective, it is impossible to collect light emitted in all directions and consequently only a fraction of the emitted photons can be collected. The circular apertures of the microscope optics (in particular the
5 A Brief Introduction to Single-Molecule Fluorescence Methods
91
objective) result in diffraction of the transmitted light, which causes the fluorescent particle not being imaged as an infinitely small point, but as an Airy disk, with a finite width and side lobes. A three-dimensional representation of this diffraction pattern is referred to as the point spread function (PSF). The diameter (d) of the PSF only depends on the NA of the objective and the wavelength of the light (l) (33): d¼
1:22l : 2NA
(5)
The width of the PSF is also a measure of the resolution of the optical system: when the distance between two closely spaced point sources is less than d, the images of the sources overlap and their peaks cannot be resolved. This definition of the resolution is called the Rayleigh criterion (33). For a typical fluorescence microscope (NA ¼ 1.4, l ¼ 505 nm), the resolution is ~220 nm. 3.2. Key Imaging Modalities in Fluorescence Microscopy 3.2.1. Confocal Fluorescence Microscopy
In confocal fluorescence microscopy, the sample is illuminated with a diffraction limited spot and an image is acquired by moving this spot over the sample (Fig. 3a). To this end, a collimated laser beam is coupled in the microscope objective, resulting in a tightly focused spot with a diameter of typically 200–300 nm, ruled by the same effects of diffraction as discussed above. The fluorescence resulting in the focus is collected by the objective, separated from the excitation beam by a dichroic mirror and further spectrally filtered by an emission filter. In addition, the fluorescence light is
Fig. 3. Schematic representation of the three most common microscope designs used in single-molecule fluorescence microscopy. (a) Confocal setup. (b) Epi-wide field setup. (c) TIRF-wide field setup. A zoom of the excitation path, the laser is coupled off-axis into the objective. The laser is reflected on the glass-water interface and creates an exponentially decaying evanescent wave in the sample. Ex is Excitation source, L lens, O objective, E emission filter, D Dichroic mirror, and A confocal aperture.
92
S.M.J.L. van den Wildenberg et al.
spatially filtered with a pinhole before being detected with a point detector (APD or photomultiplier). The pinhole increases spatial resolution, but its key purpose is to suppress out-of-focus background signal, allowing optical sectioning. To create an image, scanning of the beam (using galvanic mirrors) or the sample (using a piezo stage) is required. In single-molecule experiments, excitation intensities need to be reduced to avoid saturation or limit photobleaching, resulting in rather long image acquisition times. Consequently, the most important single-molecule application of confocal microscopy is to study molecules freely diffusing in and out the confocal spot and FCS. 3.2.2. Wide-Field Epifluorescence Microscopy
In wide-field epifluorescence microscopy, the excitation beam is not tightly focused to a diffraction-limited spot in the sample plane, but illuminates a substantially larger area (Fig. 3b). The fluorescence arising from this illuminated area can be detected with an array detector, such as a CCD camera. Uniform illumination throughout the sample can be obtained by focusing the excitation beam in the back-focal plane of the objective. In case of a laser, this results in a collimated, Gaussian-shaped beam illuminating the sample. Filters and dichroic mirrors are applied in exactly the same way as confocal fluorescence microscopy. Single-molecule wide-field fluorescence microscopy is particularly useful when molecules are moving in the sample or when a time resolution of ~100 Hz is sufficient. As discussed above, modern CCD cameras are superior with respect to dark noise and quantum efficiency, compared to point detectors used in confocal microscopy. In addition, wide-field fluorescence microscopes are optically simpler and often require less frequent and precise alignment (14).
3.2.3. Total Internal Reflection Fluorescence Microscopy
As discussed above, wide-field approaches have, in certain cases, advantages over confocal approaches, at the cost of time resolution and worse suppression of out-of-focus background fluorescence than in confocal microscopy. This latter problem is addressed in total internal reflection fluorescence microscopy (TIRF), in which the evanescent wave resulting from a totally internally reflected laser beam is used for fluorescent excitation. The evanescent wave, generated by reflection on a glass water interface, penetrates into the medium with lower refractive index (the water) and its intensity I drops exponentially with distance (z) into the low-index medium (40): z I ðzÞ ¼ I0 exp ð Þ: d
(6)
With I0 the intensity at the surface and d the decay constant. The decay constant of the intensity of the evanescent wave is on the order of 100 nm, depending on the angle of incidence and the
5 A Brief Introduction to Single-Molecule Fluorescence Methods
93
refractive indices of glass and medium. To obtain total reflection and generate an evanescent wave, the angle of incidence of the laser beam impinging on the glass/water interface needs to be larger than the critical angle (yc): yc ¼ arcsinðn1 =n2 Þ;
(7)
where n1 and n2 are the refractive indexes of water and glass, respectively. At the glass-water interface, the internal reflection is achieved at an yc of 61 . These high angles of incidence are usually obtained in one of two ways: (1) by using a prism, coupled to the coverslip on the side of the sample opposite to the objective. (2) By using an objective with NA higher than 1.45 and coupling in the laser beam off axis, on the edge of the entrance pupil (41) (Fig. 3c). The key advantage of TIRF microscopy is that only fluorophores close to the interface will be excited and fluorophores deeper in the sample will not, resulting in a reduced background. Disadvantages of TIRF microscopy are that the excitation intensity strongly depends on the depth in the sample, making comparison of fluorescence intensities difficult and that the polarization of the evanescent wave is complex.
4. Measurables 4.1. Counting the Number of Fluorescent Molecules Within a Diffraction-Limited Spot
So when an instrument is built and a sample prepared, how does one know one is looking at a single molecule? Observation of a single diffraction-limited spot is not enough: the Rayleigh criterion tells us that when particles are too close, they cannot be resolved and are imaged as a single spot. The most straightforward signature of fluorescence arising from a single fluorophore is stepwise photobleaching: the intensity is for a while rather constant, until photobleaching occurs and the signal abruptly drops to the background level. Another signature is the intensity, which should be constant from molecule to molecule; however, orientation or polarization effects can substantially modulate fluorescence intensity (see below). If the intensity of a fluorescent spot is due to more than one fluorophore, the number of fluorophores can be determined in two ways. (1) By comparing the fluorescence intensity of the spot to the average value of single fluorophores (Fig. 4a). To correct for possible bleaching within the first frame, the intensity at t ¼ 0 can be extrapolated by fitting an exponential decay to the fluorescence intensity profile over time (42). (2) One can also determine the number of fluorophores by counting the number of bleaching steps. This approach has, for example, been used to determine the number of Ase1p dimers incorporated in multimers bound to microtubules (43). On the other hand, at conditions when photobleaching is negligible the
94
S.M.J.L. van den Wildenberg et al.
Fig. 4. Measurable parameters in single-molecule fluorescence microscopy. (a) The number of fluorophores within a fluorescent spot, even though they are not resolvable, can be deduced from the intensity of that spot. (b) Localization of single fluorophores. The fluorophores are imaged as Airy disks, which can be approximated with a 2D Gaussian. The limit in which two neighboring fluorophores can still be resolved is described by Rayleigh criterion. (c) Colocalization is achieved by using labels with different colors. By imaging both color channels simultaneously on a CCD camera, the precise localization of both fluorophores can be determined. (d) FRET reports on the distance between two spectrally distinct dyes, and can be used to study for example intramolecular conformational changes. When the distance between the dyes is in the order of tens of A˚ (20 A˚ < r < 90 A˚), the energy can get transferred from the donor (D) to the acceptor (A) causing an increase in acceptor signal. When the distance between the dyes is large (r > 90 A˚), no energy transfer occurs and the acceptor signal will be low. (e) Fluorescence polarization reports on orientation or orientational dynamics. Circular polarized light can be used to excite dyes in all orientations. Subsequently, the emitted light is filtered for a specific polarization.
changes in numbers of molecules due to association or dissociation can be measured. This approach has been used to determine the number of Rad51 monomers disassembling from DNA in a single burst (44). 4.2. Localization of Single Molecule
We have seen above that the resolution of a fluorescence microscope is limited by diffraction, to about half the wavelength of the emitted light. The resolution is a measure for how close two point sources can be to be still resolvable (Fig. 4b), it does not restrict the accuracy with which the location of a single point source can
5 A Brief Introduction to Single-Molecule Fluorescence Methods
95
be determined. By fitting the resulting image with the PSF (often an approximation with a Gaussian is sufficient), the location of the maximum of the image can be determined with far greater accuracy than the width of the PSF. This method is frequently used in single particle tracking (SPT) (38, 45). Given the noise encountered in most single-molecule experiments a PSF of the microscope with a full width at half maximum of ~1.5–2 pixel yields best results for the accuracy (46). The uncertainty in the localization of a point source (Dx) depends on the size of the pixels (a), the number of photons (N), the background noise (b), the standard deviation of the PSF (s) (46, 47): D E s 2 ða 2 =12Þ 8ps 4 b 2 ðDx Þ2 ¼ þ þ 2 2 : (8) N N a N The first term represents the photon counting noise (s2/N), the second term represents pixilation noise arising from the uncertainty of where in the pixel the photon arrived ((a2/12)/N). The final term is due to background noise. Under typical single-molecule fluorescence conditions, position accuracies down to about 2 nm can be achieved (47). 4.3. Detection of Motion of Single Molecules
Given this high localization accuracy, the positions of an emitting fluorophore can be determined in each image from a time stack of images and subsequently a trajectory can be reconstructed by connecting the positions. Using this approach, the motion of single-molecules can be accurately determined. Care has to be taken that the motion of the molecules are not too large within the acquisition time of an image, since this can smear out the Gaussian intensity profile, complicating fitting. This problem can be avoided by using short acquisition times and increasing the excitation intensity, at the cost of enhanced photobleaching. It is important in single-molecule tracking to find the proper balance between movement of a particle within the acquisition time, and the total number of time points (frames) over which the particle is observed (48). Motion of biomolecules can be directional (for example driven by motor proteins) or diffusive (like membrane proteins). To analyze the precise nature of mobility, often the mean square displacement (MSD) is calculated as a function of time. Motion with constant speed (and direction) leads to an MSD that increases with the square of time while diffusive motion results in a linear increase of the MSD with time. The localization uncertainty leads to a constant offset in the MSD, due to its timeindependence (38, 49). The MSD analysis was, for example, used to show that, depending on the exact conditions, the motor protein kinesin-5 can switch between different mode of motility, diffusion, and directed motion (50).
96
S.M.J.L. van den Wildenberg et al.
4.4. Colocalization of Fluorescent Molecules
One of the key interests in (cell) biology is to resolve which proteins interact and how. To this end, proteins of interest can be labeled with differently colored fluorophores (51–53). Subsequently, the different dyes can be excited by the appropriate lasers and the fluorescence signal can be separated in two or more wavelength channels and detected independently using different cameras or side-by-side on one. In this way, different biomolecules can be tracked simultaneously and their motion can be correlated to resolve whether they move independently or interact (part of the time) (Fig. 4c). High-resolution colocalization was applied to show myosin V’s alternating heads while it walked hand-over-hand along an actin filament (53).
4.5. Fo¨rster Resonance Energy Transfer
Positions and distances of single fluorophores can be determined with an accuracy that is substantially smaller than the diffraction limit using PSF fitting (see above). This approach is very powerful, but has its limitations, in particular in its poor time resolution and its inability to resolve multiple molecules that are closer than the optical resolution, without photobleaching them. An excellent method to measure relative distances and changes on a length scale of ~2–9 nm is Fo¨rster resonance energy transfer (FRET) (Fig. 4d). In FRET, two spectrally distinct fluorophores are used. One, with the highest energy excited state, is excited and serves as donor, the other as acceptor. When the two fluorophores are close and their dipoles oriented favorably, dipole–dipole coupling can occur and excitations can be transferred from donor to acceptor. The distance dependency of the FRET efficiency (E ) is: E¼
1 1 þ ðR=R0 Þ6
(9)
with R the distance between donor and acceptor and R0 the Fo¨rster distance. The Fo¨rster distance is defined as the distance at which half the fluorescence of the donor is transferred to the acceptor. The Fo¨rster distance depends on the overlap of the emission spectrum of the donor with the absorption spectrum of the acceptor, the relative orientation of donor and acceptor dipole moments and the fluorescence quantum yield of the donor. It has a typical value of about 5 nm (5). FRET has proven to be a valuable tool to study conformational dynamics in nucleic acids and proteins. Examples are the folding of ribozymes (54) and the observation of conformational dynamics in kinesin-1 (55). 4.6. Fluorescence Polarization
Another way to measure conformational dynamics of single biomolecules is to use the polarization of the fluorescence signal. Absorption and emission are governed by the interaction of the absorption and emission transition dipole moments of the chromophore, which are vectors, with the electric component of the electromagnetic light field, also a vector. Using polarized light for excitation and/or a polarizer in the emission path allows
5 A Brief Introduction to Single-Molecule Fluorescence Methods
97
obtaining the orientation and dynamics of the transition dipole moment. Care needs to be taken that the fluorophore is not free to rotate with respect to the biomolecule of interest, but that its orientation is tightly linked to that of the biomolecule. This can be achieved by using bisfunctional fluorophores that are connected with two chemical links to the protein or nucleic acid of interest (56). One way of determining dipole orientations on the single-molecule level is to excite with circular polarized light and to split the resulting fluorescence in two perpendicular linear polarized signals, detected with two APDs or side-by-side on a CCD chip (Fig. 4e). Another way is to detect without polarizers, but to use alternating (linear) polarization of the excitation light. If one combines polarized excitation with polarized detection, a separation can be obtained of the depolarization due to rapid fluorophore orientation (on the nanosecond scale) and much slower conformational changes. Polarization methods have, for example, been applied to study the conformational changes occurring during stepping of the kinesin-1 motor protein (57).
5. Concluding Remarks Here, we have provided a short and broad overview of the wealth of single-molecule fluorescence approaches and their backgrounds. These tools have become indispensible in the study of diverse processes, such as the active and diffusive motion of biomolecules, their conformational changes, and their assembly and disassembly. In the following Chapters 6 to 10, several approaches are discussed in more depth, including detailed protocols. References 1. Stokes, G. G. (1852) On the Change of Refrangibility of Light, Philosophical Transactions of the Royal Society of London 142, 463–562. 2. Herschel, J. F. W. (1845) No. I. On a Case of Superficial Colour Presented by a Homogeneous Liquid Internally Colourless, Philosophical Transactions of the Royal Society of London 135, 143–145. 3. Herschel, J. F. W. (1845) No. II. On the Epipolic Dispersion of Light, Being a Supplement to a Paper Entitled, “On a Case of Superficial Colour Presented by a Homogeneous Liquid Internally Colourless”, Philosophical Transactions of the Royal Society of London 135, 147–153. 4. Brewster, D. (1846) On the decomposition and dispersion of light within solid and fluid
bodies, Transitions of the Royal Society of Edinburgh 16, 11. 5. Lakowicz, J. R. (2006) Principles of Fluorescence Microscopy, 3 ed., Springer, New York. 6. Hirschfeld, T. (1976) Optical Microscopic Observation of Single Small Molecules, Journal of the Optical Society of America 66, 1124–1124. 7. Nguyen, D. C., Keller, R. A., Jett, J. H., and Martin, J. C. (1987) Detection of Single Molecules of Phycoerythrin in Hydrodynamically Focused Flows by Laser-Induced Fluorescence, Analytical Chemistry 59, 2158–2161. 8. Peck, K., Stryer, L., Glazer, A. N., and Mathies, R. A. (1989) Single-Molecule Fluorescence Detection - Auto-Correlation Criterion and Experimental Realization with
98
S.M.J.L. van den Wildenberg et al.
Phycoerythrin, Proceedings of the National Academy of Sciences of the United States of America 86, 4087–4091. 9. Moerner, W. E., and Kador, L. (1989) Finding a Single Molecule in a Haystack - OpticalDetection and Spectroscopy of Single Absorbers in Solids, Analytical Chemistry 61, A1217–A1223. 10. Orrit, M., and Bernard, J. (1990) Single Pentacene Molecules Detected by Fluorescence Excitation in a Para-Terphenyl Crystal, Physical Review Letters 65, 2716–2719. 11. Shera, E. B., Seitzinger, N. K., Davis, L. M., Keller, R. A., and Soper, S. A. (1990) Detection of Single Fluorescent Molecules, Chemical Physics Letters 174, 553–557. 12. Tinnefeld, P., and Sauer, M. (2005) Branching out of single-molecule fluorescence spectroscopy: Challenges for chemistry and influence on biology, Angewandte ChemieInternational Edition 44, 2642–2671. 13. Weiss, S. (1999) Fluorescence spectroscopy of single biomolecules, Science 283, 1676–1683. 14. Moerner, W. E., and Fromm, D. P. (2003) Methods of single-molecule fluorescence spectroscopy and microscopy, Review of Scientific Instruments 74, 3597–3619. 15. Soper, S. A., Nutter, H. L., Keller, R. A., Davis, L. M., and Shera, E. B. (1993) The Photophysical Constants of Several Fluorescent Dyes Pertaining to Ultrasensitive Fluorescence Spectroscopy, Photochemistry and Photobiology 57, 972–977. 16. Wieser, S., and Schutz, G. J. (2008) Tracking single molecules in the live cell plasma membrane-Do’s and Don’t’s, Methods 46, 131–140. 17. Kapanidis, A. N., and Weiss, S. (2002) Fluorescent probes and bioconjugation chemistries for single-molecule fluorescence analysis of biomolecules, Journal of Chemical Physics 117, 10953–10964. 18. Peterman, E. J. G., Brasselet, S., and Moerner, W. E. (1999) The fluorescence dynamics of single molecules of green fluorescent protein, Journal of Physical Chemistry A 103, 10553–10560. 19. Kuno, M., Fromm, D. P., Hamann, H. F., Gallagher, A., and Nesbitt, D. J. (2000) Nonexponential “blinking” kinetics of single CdSe quantum dots: A universal power law behavior, Journal of Chemical Physics 112, 3117–3120. 20. Joo, C., Balci, H., Ishitsuka, Y., Buranachai, C., and Ha, T. (2008) Advances in single-molecule fluorescence methods for molecular biology, Annual Review of Biochemistry 77, 51–76. 21. Resch-Genger, U., Grabolle, M., CavaliereJaricot, S., Nitschke, R., and Nann, T. (2008)
Quantum dots versus organic dyes as fluorescent labels, Nature Methods 5, 763–775. 22. Alivisatos, A. P. (1996) Semiconductor clusters, nanocrystals, and quantum dots, Science 271, 933–937. 23. Kaji, N., Tokeshi, M., and Baba, Y. (2007) Single-molecule measurements with a single quantum dot, Chemical Record 7, 295–304. 24. Alivisatos, P. (2004) The use of nanocrystals in biological detection, Nature Biotechnology 22, 47–52. 25. Gao, X. H., Yang, L. L., Petros, J. A., Marshal, F. F., Simons, J. W., and Nie, S. M. (2005) In vivo molecular and cellular imaging with quantum dots, Current Opinion in Biotechnology 16, 63–72. 26. Michalet, X., Pinaud, F. F., Bentolila, L. A., Tsay, J. M., Doose, S., Li, J. J., Sundaresan, G., Wu, A. M., Gambhir, S. S., and Weiss, S. (2005) Quantum dots for live cells, in vivo imaging, and diagnostics, Science 307, 538–544. 27. Chalfie, M., Tu, Y., Euskirchen, G., Ward, W. W., and Prasher, D. C. (1994) Green Fluorescent Protein as a Marker for GeneExpression, Science 263, 802–805. 28. Giepmans, B. N. G., Adams, S. R., Ellisman, M. H., and Tsien, R. Y. (2006) Review - The fluorescent toolbox for assessing protein location and function, Science 312, 217–224. 29. Shaner, N. C., Steinbach, P. A., and Tsien, R. Y. (2005) A guide to choosing fluorescent proteins, Nature Methods 2, 905–909. 30. Patterson, G. H., and Lippincott-Schwartz, J. (2002) A photoactivatable GFP for selective photolabeling of proteins and cells, Science 297, 1873–1877. 31. Lu, H. P., Xun, L. Y., and Xie, X. S. (1998) Single-molecule enzymatic dynamics, Science 282, 1877–1882. 32. Rutkauskas, D., Novoderezhkin, V., Cogdell, R. J., and van Grondelle, R. (2005) Fluorescence spectroscopy of conformational changes of single LH2 complexes, Biophysical Journal 88, 422–435. 33. Murphy, D. B. (2001) Fundamentals of light microscopy and electronic imaging, Wiley-Liss, Inc. 34. Michalet, X., Siegmund, O. H. W., Vallerga, J. V., Jelinsky, P., Millaud, J. E., and Weiss, S. (2007) Detectors for single-molecule fluorescence imaging and spectroscopy, Journal of Modern Optics 54, 239–281. 35. Magde, D., Elson, E. L., and Webb, W. W. (1974) Fluorescence Correlation Spectroscopy .2. Experimental Realization, Biopolymers 13, 29–61. 36. Eigen, M., and Rigler, R. (1994) Sorting Single Molecules - Application to Diagnostics
5 A Brief Introduction to Single-Molecule Fluorescence Methods and Evolutionary Biotechnology, Proceedings of the National Academy of Sciences of the United States of America 91, 5740–5747. 37. Verbrugge, S., Kapitein, L. C., and Peterman, E. J. G. (2007) Kinesin moving through the spotlight: Single-motor fluorescence microscopy with submillisecond time resolution, Biophysical Journal 92, 2536–2545. 38. Anderson, C. M., Georgiou, G. N., Morrison, I. E. G., Stevenson, G. V. W., and Cherry, R. J. (1992) Tracking of Cell-Surface Receptors by Fluorescence Digital Imaging Microscopy Using a Charge-Coupled Device Camera Low-Density-Lipoprotein and InfluenzaVirus Receptor Mobility at 4-Degrees-C, Journal of Cell Science 101, 415–425. 39. Hecht, E. (1998) Optics, third edition. 40. Dickson, R. M., Norris, D. J., Tzeng, Y. L., and Moerner, W. E. (1996) Three-dimensional imaging of single molecules solvated in pores of poly(acrylamide) gels, Science 274, 966–969. 41. Axelrod, D. (2001) Total internal reflection fluorescence microscopy in cell biology, Traffic 2, 764–774. 42. Leake, M. C., Greene, N. P., Godun, R. M., Granjon, T., Buchanan, G., Chen, S., Berry, R. M., Palmer, T., and Berks, B. C. (2008) Variable stoichiometry of the TatA component of the twin-arginine protein transport system observed by in vivo single-molecule imaging, Proceedings of the National Academy of Sciences of the United States of America 105, 15376–15381. 43. Kapitein, L. C., Janson, M. E., van den Wildenberg, S., Hoogenraad, C. C., Schmidt, C. F., and Peterman, E. J. G. (2008) Microtubule-Driven Multimerization Recruits ase1p onto Overlapping Microtubules, Current Biology 18, 1713–1717. 44. van Mameren, J., Modesti, M., Kanaar, R., Wyman, C., Peterman, E. J. G., and Wuite, G. J. L. (2009) Counting RAD51 proteins disassembling from nucleoprotein filaments under tension, Nature 457, 745–748. 45. Schmidt, T., Schutz, G. J., Baumgartner, W., Gruber, H. J., and Schindler, H. (1996) Imaging of single molecule diffusion, Proceedings of the National Academy of Sciences of the United States of America 93, 2926–2929. 46. Thompson, R. E., Larson, D. R., and Webb, W. W. (2002) Precise nanometer localization analysis for individual fluorescent probes, Biophysical Journal 82, 2775–2783. 47. Yildiz, A., and Selvin, P. R. (2005) Fluorescence imaging with one manometer accuracy:
99
Application to molecular motors, Accounts of Chemical Research 38, 574–582. 48. Saxton, M. J. (1997) Single-particle tracking: The distribution of diffusion coefficients, Biophysical Journal 72, 1744–1753. 49. Gross, D., and Webb, W. W. (1986) Molecular Counting of Low-Density-Lipoprotein Particles as Individuals and Small Clusters on CellSurfaces, Biophysical Journal 49, 901–911. 50. Kwok, B. H., Kapitein, L. C., Kim, J. H., Peterman, E. J. G., Schmidt, C. F., and Kapoor, T. M. (2006) Allosteric inhibition of kinesin-5 modulates its processive directional motility, Nature Chemical Biology 2, 480–485. 51. Lacoste, T. D., Michalet, X., Pinaud, F., Chemla, D. S., Alivisatos, A. P., and Weiss, S. (2000) Ultrahigh-resolution multicolor colocalization of single fluorescent probes, Proceedings of the National Academy of Sciences of the United States of America 97, 9461–9466. 52. Agrawal, A., Deo, R., Wang, G. D., Wang, M. D., and Nie, S. M. (2008) Nanometer-scale mapping and single-molecule detection with color-coded nanoparticle probes, Proceedings of the National Academy of Sciences of the United States of America 105, 3298–3303. 53. Churchman, L. S., Okten, Z., Rock, R. S., Dawson, J. F., and Spudich, J. A. (2005) Single molecule high-resolution colocalization of Cy3 and Cy5 attached to macromolecules measures intramolecular distances through time, Proceedings of the National Academy of Sciences of the United States of America 102, 1419–1423. 54. Zhuang, X. W., Bartley, L. E., Babcock, H. P., Russell, R., Ha, T. J., Herschlag, D., and Chu, S. (2000) A single-molecule study of RNA catalysis and folding, Science 288, 2048–2051. 55. Mori, T., Vale, R. D., and Tomishige, M. (2007) How kinesin waits between steps, Nature 450, 750–715. 56. Corrie, J. E. T., Craik, J. S., and Munasinghe, V. R. N. (1998) A homobifunctional rhodamine for labeling proteins with defined orientations of a fluorophore, Bioconjugate Chemistry 9, 160–167. 57. Asenjo, A. B., and Sosa, H. (2009) A mobile kinesin-head intermediate during the ATPwaiting state, Proceedings of the National Academy of Sciences of the United States of America 106, 5657–5662. 58. Rosenow, M. A., Huffman, H. A., Phail, M. E., and Wachter, R. M. (2004) The crystal structure of the Y66L variant of green fluorescent protein supports a cyclization-oxidationdehydration mechanism for chromophore maturation, Biochemistry 43, 4464–4472.
.
Chapter 6 Fluorescent Labeling of Proteins Mauro Modesti Abstract Many single-molecule experimental techniques exploit fluorescence as a tool to investigate conformational dynamics and molecular interactions or track the movement of proteins in order to gain insight into their biological functions. A prerequisite to these experimental approaches is to graft one or more fluorophores on the protein of interest with the desired photophysical properties. Here, we present detailed procedures for the current most efficient methods used to covalently attach fluorophores to proteins. Alternative direct and indirect labeling strategies are also described. Key words: Fluorescence, Fluorophore, Maleimide, Cys light, NHS ester, Lys light, RAD51, RPA, HOP2/MND1, RecA, hSSB
1. Introduction The design of fluorescence-based single-molecule experiments requires choosing an optimal fluorophore as marker to monitor a particular protein activity. Tryptophan intrinsic fluorescence emission can be very useful to study the folding, conformational dynamics, and interactions of a protein. However, since quartz optics are required, complicated photophysical properties and many proteins lack tryptophan residues; the choice of tryptophan as a fluorophore is often unpractical for single-molecule experimental techniques. Thus, attaching an “extrinsic” fluorophore moiety to the protein of interest is the most frequent route used to make the protein glow. One approach commonly used for in vivo experiments is to generate a chimer of the protein of interest by fusion to an intrinsically fluorescent protein that fluoresces in the visible/near infrared range of the electromagnetic spectrum. Collections of intrinsically fluorescent proteins exist, such as the jellyfish green fluorescent protein (GFP) and its variants that can be attached by genetic engineering to either the Erwin J.G. Peterman and Gijs J.L. Wuite (eds.), Single Molecule Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 783, DOI 10.1007/978-1-61779-282-3_6, # Springer Science+Business Media, LLC 2011
101
102
M. Modesti
amino or the carboxyl terminus of a protein or even inserted in frame internally to the protein (1). These protein-based fluorophores are very useful for in vitro experiments as well, but they are relatively bulky and can perturb the original protein function. In addition, their photophysical properties are complex and influenced by pH, ionic strength, or any factors affecting protein folding. Instead of fusing such an intrinsic fluorescent protein, small synthetic fluorescent organic compounds can covalently be attached to the protein of interest. A large number of small fluorophores, including ATTO dyes, CF dyes, cyanine dyes, HiLyte Fluors, Alexa Fluors, or DyLight Fluors, have been developed and are commercially available. These small, photostable, and bright fluorophores are generally much less sensitive to buffer conditions as compared to intrinsically fluorescent proteins. Moreover, the chemical and photophysical properties of these small fluorophores are usually well-defined, which allows selection of the dye with the most optimal properties for a given single-molecule application. In a fluorescence correlation spectroscopy experiment, for example, a dye with high quantum yield, weak nonspecific binding to the protein of interest, and resistant to irreversible photobleaching and to triplet state excitation can be optimally selected. Or, if the application requires detection and visualization of a single fluorophore, one would want to select a dye with reduced blinking behavior. Most of these chemical dyes can be purchased in a “functionalized” form with a reactive group for specific covalent attachment to proteins on, for instance, the –SH group of cysteine residues or the –NH2 group of lysine residues. In this chapter, we describe procedures for labeling proteins with such functionalized fluorophores, as well as alternative procedures. The presentation of these protocols is intended to guide researchers in biophysics that do not have much experience with protein handling. Each method is illustrated with an example from our laboratory not only to show to the reader how results should look like, but also to highlight commonly encountered problems and drawbacks when using these procedures.
2. Materials 2.1. Buffers and Solutions
1. 0.5 M MES–NaOH, pH 6.2. 2. 0.5 M MOPS–NaOH, pH 7.0. 3. 1 M Tris–HCl, pH 7.5; 1 M Tris–HCl, pH 8.0. 4. 1 M HEPES–NaOH, pH 8.2. 5. 2 M imidazole–HCl, pH 7.5, stored at dark at +4 C.
6 Fluorescent Labeling of Proteins
103
6. 0.5 M EDTA–NaOH, pH 8.0. 7. IPTG dissolved in water at 1 M and stored at 20 C. 8. 1 M DTT freshly prepared as a solution in water. 9. 5 M NaCl. 10. 3 M KCl. 11. Anhydrous DMSO stored at +4 C in a desiccator. 12. Glycerol. 13. SDS 20% (w/v). 14. ß-mercaptoethanol. 15. Coomassie Brilliant Blue R-250. 16. Ampicillin sodium salt, solution at 100 mg/ml in water, store at 20 C. 17. Chloramphenicol, solution made at 34 mg/ml in ethanol, store at 20 C. 18. Storage buffer A: 0.3 M KCl, 20 mM Tris–HCl, pH 8, 1 mM DTT, 0.5 mM EDTA, and 10% glycerol. 19. Labeling buffer A: 0.5 M NaCl, 50 mM MOPS–NaOH, pH 7, 0.5 mM EDTA, and 10% glycerol. 20. Storage buffer B: 0.5 M KCl, 20 mM Tris–HCl, pH 8, 1 mM DTT, 0.5 mM EDTA, and 10% glycerol. 21. Labeling buffer B: 0.5 M NaCl, 50 mM HEPES–NaOH, pH 8.2, 0.5 mM EDTA, and 10% glycerol. 22. Labeling buffer C: 1 M NaCl, 50 mM MOPS–NaOH, pH 7, 0.5 mM EDTA, and 10% glycerol. 23. PBS: 137 mM NaCl, 2.7 mM KCl, 4.3 mM Na2HPO4, and 1.47 mM KH2PO4. Adjust to pH 7.4 with NaOH. 24. 2 lysis buffer: 1 M NaCl, 40 mM Tris–HCl, pH 7.5, 4 mM bmercaptoethanol, 10 mM imidazole, pH 8, and 20% glycerol. 25. Buffer R: 50 mM KCl, 20 mM Tris–HCl, pH 7.5, 1 mM DTT, 0.5 mM EDTA, and 10% glycerol. 26. Protein sample buffer: 2% w/v SDS, 62.5 mM Tris–HCl, pH 6.8, 25% glycerol, 0.01% w/v bromophenol blue, and 0.72 M ß-mercaptoethanol. 27. Staining solution: 10% ethanol, 7% acetic acid, and 1 g/L Coomassie Brilliant Blue R-250. 28. Destaining solution: 10% ethanol and 7% acetic acid. 2.2. Dyes
Alexa Fluor dyes (Invitrogen) and ATTO dyes (Sigma-Aldrich) are dissolved in anhydrous DMSO and used immediately.
104
M. Modesti
2.3. SDS-PAGE
1. Precast NuPAGE Bis-Tris Gels with MOPS running buffer (Invitrogen). 2. Prestained Precision Plus Protein Standards (BIO-RAD).
2.4. Columns
1. Econo-Pac 10DG columns (BIO-RAD). 2. PD SpinTrap G-25, HisTrap FF, HiTrap Q HP, and HiTrap Heparin HP columns (GE Healthcare).
2.5. Media and Bacterial Expression
Rosetta/pLysS cells (Novagen) were used as host for inducible expression of hRPA-eGFP. Cells were grown in LB broth (bactotryptone 10 g/L, yeast extract 5 g/L, NaCl 10 g/L) supplemented with antibiotics as indicated.
3. Methods The “Cys light” and “Lys light” methods for covalent attachment of small organic fluorophores to proteins are described. Alternative labeling procedures are presented that can be used in case the latter two methods fail to yield suitable reagents. Each method is illustrated with an example from our laboratory, highlighting commonly encountered problems when using these methods. We describe means to analyze the extent and the specificity of the labeling reaction. Importantly, whatever the labeling method used, it is essential to verify that the original activity of the protein has not been altered by the labeling procedure. 3.1. The Cys Light Method: Labeling of the hHOP2–MND1 Protein Complex
The maleimide chemical group reacts with the –SH group (thiol) of cysteine residues of proteins to form a covalent thioether bond (see Note 1). Because of the high specificity and efficiency of this reaction, the functionalization of organic fluorescent compounds with a maleimide group has developed as the method of choice over other –SH reactive groups, such as iodoacetamide. A large selection of organic fluorophores functionalized with a maleimide group is available commercially. Below, we present a modification of the labeling method recommended by Molecular Probes but adapted to the labeling of the hHOP2–hMND1 protein complex on cysteine residues. The hHOP2 and hMND1 proteins form a heterodimeric complex (2). Each protein contains three cysteine residues. The lack of three-dimensional structural information does not allow prediction of the surface-exposed cysteine residues of the complex. Since the focus of this chapter is on protein labeling procedures and not purification methods, we assume that the researcher has access to a source of purified protein. For all procedures, work on ice or in the cold as much as possible and avoid exposure to light when handling fluorophores.
6 Fluorescent Labeling of Proteins
105
1. To ensure that surface-exposed cysteines are in a reduced form and reactive toward the maleimide group, freshly prepared DTT (see Note 2) is added to a final concentration of 20 mM to 3 ml of the protein solution at a concentration of 2 mg/ml in storage buffer A and incubated on ice for 30 min. 2. After reduction, the protein sample is buffer exchanged (see Note 3) into deoxygenated labeling buffer A (see Notes 4 and 5) and finally recovered into 4 ml. 3. Alexa Fluor 488 C5 maleimide (or Alexa Fluor 546 C5 maleimide) is dissolved in anhydrous DMSO (see Note 6) and immediately added to the protein solution at fivefold molar excess of dye over the protein (see Note 7). From this step on, protect the sample from exposure to light as much as possible using aluminum foil. The dye solution should be added rapidly drop by drop while stirring the solution with a small magnetic bar to avoid local concentration effects. The reaction mixture is left to incubate at +4 C for 2 h with stirring. 4. At the end of the reaction, DTT is added to 10 mM final concentration and further incubated for 30 min to quench the excess reactive dye. 5. The volume is brought up to 6 ml with labeling buffer A and centrifuged at 20,000 g in a Sorval SS34 rotor for 30 min to remove aggregates (see Note 8). The sample is divided into two aliquots of 3 ml, and each aliquot is buffer exchanged into the desired buffer using Econo-Pac 10DG columns to remove excess dye (see Note 3). In this case, we buffer exchange into storage buffer A. The two aliquots are pooled giving 8 ml of labeled protein solution that should be at around 0.7 mg/ml, if no loss by aggregation has occurred. At this stage, the sample can be concentrated if desired (see Note 9), aliquoted, and stored at 80 C after flash freezing in liquid nitrogen. Even after passage on the Econo-Pac 10DG column, presence of free dye is often observed. To remove the residual free dye, we dialyze the protein samples (8 ml) prior to concentration for 4 h against 2 L of storage buffer A (see Note 10). 6. The extent of the labeling reaction was assessed by denaturing and reducing SDS-PAGE analysis using 20 ml of the preparation (see Note 11). As shown in Fig. 1, the labeling is apparently complete since the mobility of both proteins after labeling is retarded compared to the unlabeled control (see Note 12). The dye has been covalently attached to both hHOP2 and hMND1 subunits. The degree of labeling (DOL or dye-to-monomer ratio) is estimated spectrophotometrically (see Note 13), giving in this case DOL of 4.5 and 4.3 for the Alexa Fluor 488 C5 maleimide and the Alexa Fluor 546 C5 maleimide-labeling reaction, respectively. This suggests that four of the possible
M. Modesti
U
A-4 88 A-5 46
106
M 50 kDa 37
hHOP2 hMDN1
hHOP2 hMDN1
hHOP2 hMDN1
Coomassie staining
25 20 15 10
before Coomassie staining 473 nm Excitation LP 510 nm
before Coomassie staining 532 nm Excitation LP 575 nm
Fig. 1. Fluorescent labeling of the hHOP2–hMND1 complex by the Cys light method. The hHOP2–hMND1 complex was labeled with Alexa Fluor 488 maleimide (A-488) or Alexa 546 Fluor maleimide (A-546) and compared to the unlabeled preparation (U) by denaturing and reducing SDS-PAGE analysis. M ¼ protein size standards. After electrophoresis, the gel was first scanned with a Fluor Imager FLA-5100 (Fujifilm) to detect emission of Alexa Fluor 488 (middle panel ) or Alexa Fluor 546 (bottom panel ). Next, the gel was stained with Coomassie Brilliant Blue R-250 to reveal all proteins and visualized by bright-field illumination (top panel ).
six Cys residues in the heterodimeric complex are exposed to solvent. It can also be noticed that the fluorescent signal after labeling is more intense for hMND1 than for hHOP2 (see Fig. 1; bottom panel obtained by fluor imager scanning of the gel before Coomassie staining). This difference in intensity suggests that hMND1 has more cysteine residues exposed to solvent than hHOP2, but further analysis by mass spectrometry is required to prove this point. 7. This protocol is relatively large scale, but the reactions can be scaled down to 100 ml when optimizing reaction conditions keeping the protein concentration at around 1–2 mg/ml. First, treat the protein sample with DTT as described in step 1. Then, buffer exchange the protein into labeling buffer A
6 Fluorescent Labeling of Proteins
107
using a PD SpinTrap G-25 column (see Note 14). Perform reaction series to optimize labeling conditions (label-toprotein ratio, time of incubation, temperature of incubation, pH, buffer conditions, and ionic strength and we typically use NaCl or KCl). The reaction is quenched by addition of DTT as in step 4. Ten microliters of each reaction mixture can directly be analyzed by SDS-PAGE analysis to verify covalent attachment of the dye. Excess free dye is visible in the running front of the gel. The reaction that is most optimal can further be buffer exchanged into a desired storage buffer using a PD SpinTrap G-25 column as described above. In general, this small-scale preparation gives enough reagent to perform pilot experiments in single-molecule setups keeping in mind that free dye might still be present in the preparation. 8. When using recombinant protein that contains a polyhistidine tag, the labeling reaction can be done on a 1 ml HisTrap FF. Load 1–5 mg of protein on the column in a labeling buffer free of DTT and EDTA and pH not lower than 7. As a general practice, keep all fractions for analysis by SDS-PAGE analysis if required. Dilute the DMSO dye solution in 2 ml of labeling buffer at the desired dye-to-protein ratio. Gently flush the dye solution in the column and let incubate at the desired temperature for the desired time. After the reaction, flush 10 ml of labeling buffer to remove free dye. Elute the labeled protein with labeling buffer A containing 250 mM imidazole, pH 8.0 (adding 1 mM EDTA to the elution buffer is fine but may strip the nickel). First, flush 1 ml of elution buffer on the column and let sit for 15 min. Recover the protein by injecting 2 ml of elution buffer. Discard the column, as it is best to avoid reusing it to prevent contamination with different dyes. Dialyze into the desired storage buffer containing 10% glycerol. Aliquot, flash freeze in liquid nitrogen, and store at 80 C. 3.2. Site-Specific Labeling by the Cys Light Method: Labeling of hRAD51
For many fluorescent single-molecule applications, it is advantageous to obtain a preparation of a fluorescent variant of the protein of interest in which every monomer is specifically labeled at a selected surface position. The development of such a reagent is of course more time consuming, but in the end, it greatly helps interpretation and analysis of results. As an example, we describe here how the Cys light method was used to specifically label a protein at a specific surface position (3). The human RAD51 recombinase monomer contains five cysteine residues of which Cys31 and Cys319 are exposed to solvent according to structural predictions. To obtain a homogeneous population of monomers having a single and discrete label, each on the same position, namely, residue Cys31, we mutated Cys319 to Ser by site-directed mutagenesis of the plasmid expression construct, thereby removing the –SH group
108
M. Modesti
at position 319. Labeling of the C319S hRAD51 variant with Alexa Fluor 488 C5 maleimide or Alexa Fluor 555 C2 maleimide was performed as described in Subheading 3.1 (steps 1–6). DOL of 0.8 and 1.1 was measured, respectively, and has shown in Fig. 2a; both labeling reactions went to completion as judged by the retarded mobility of the labeled samples in the gels. Mass spectrometry of the full-length labeled protein samples (see Note 15) shows that the Alexa Fluor 488 C5 maleimide-labeled sample is homogeneous, containing mostly monomers with one single dye covalently attached (Fig. 2b, left panel). In contrast, analysis of the Alexa Fluor 555 C2 maleimide-labeled sample shows that a substantial fraction of monomers have incorporated two dye molecules (Fig. 2b, right panel). Thus, different dyes with different carbon arm linkers may have different specificities. The identification of the residue(s) giving nonspecific labeling requires further detailed mass spectrometry analysis after protease digestion. 3.3. The Lys Light Method: Labeling of hSSB1 and hSSB2
The other popular chemistry used to covalently attach a fluorophore to a protein is succinimidyl ester or N-hydroxysuccinimide (NHS) ester conjugates, which are reactive toward amine groups, such as e-amino groups of lysines or the amine terminus of proteins, forming a chemically stable amide bond (see Note 1). Below, we describe the labeling of hSSB1 and hSSB2 proteins with modifications of the procedure recommended by ATTO-TEC (see Note 16) (4). Lysine residues are typically abundant in proteins and the procedure can give very high DOL. 1. The protein sample (3 ml) is supplied in storage buffer B at a concentration of 2 mg/ml and buffer exchanged in labeling buffer B using an Econo-Pac 10DG column and recovered in 4 ml. Maintaining the pH just above 8 ensures that the exposed e-amino groups are sufficiently deprotonated and, thus, reactive toward the NHS ester and that the competing reaction with hydroxyl ions is minimized. 2. Atto 488-NHS ester is dissolved in anhydrous DMSO and added at twofold molar excess of dye over the protein (see Notes 6 and 7). The dye solution should be added rapidly drop by drop while stirring the solution with a small magnetic bar to avoid local concentration effects. The reaction mixture is left to incubate at +4 C for 1 h or at room temperature for 30 min with stirring. We find that the reaction is quite fast and that with some samples a 5-min incubation at room temperature is sufficient to label the protein to completion. 3. The reaction is quenched by addition of Tris–HCl, pH 7.5, to 0.1 M and further incubated for 5 min.
M
Before
U
M
After
U
Before
M
109
A-555
U
A-555
M
A-488
a
A-488
6 Fluorescent Labeling of Proteins
U
After
Coomassie staining
b 8000 37,210
37,210
37,907
8000
Unlabeled
Intensity [a.u.]
Unlabeled
A-488
6000
A-555 ∆m1 = 951 ∆m2 = 2072
6000
∆m = 697 4000
4000
M2+
38,161
M2+ 39,282
2000 2000
0 0 20000 25000 30000 35000 40000 45000
m/z
20000
25000
30000
35000
40000
45000
m/z
Fig. 2. Site-specific fluorescent labeling of hRAD51 by the Cys light method. (a) Preparations of the hRAD51 variant C319S were labeled with Alexa Fluor 488 maleimide (A-488) or Alexa Fluor 555 maleimide (A-555) and compared to the unlabeled preparation (U) by denaturing and reducing SDS-PAGE analysis. M ¼ protein size standards. After electrophoresis, dye attachment was directly visualized by bright-field illumination of the gels (before panels) or after Coomassie Brilliant Blue R-250 staining (after panels). (b) The degree and homogeneity of labeling were analyzed by MALDI-TOF of the full-length protein preparations (lower panels ).
110
M. Modesti
4. The volume is brought up to 6 ml with labeling buffer and centrifuged at 20,000 g in a Sorval SS34 rotor to remove aggregates. The sample is divided into two aliquots of 3 ml, and each aliquot is buffer exchanged into the desired buffer using Econo-Pac 10DG columns to remove excess dye (see Note 3). In this case, we buffer exchange into storage buffer B. The two aliquots are pooled giving 8 ml of labeled protein solution that should be at around 0.7 mg/ml if no loss by aggregation has occurred. At this stage, the sample can be concentrated if desired (see Note 9), aliquoted, and stored at 80 C after flash freezing in liquid nitrogen. Even after passage on the Econo-Pac 10DG column, some low level of free dye is often observed. To remove the residual free dye, we dialyze the protein samples (8 ml) for 4 h against 2 L of storage buffer B (see Note 10). 5. Figure 3 shows an assessment of the extent of the labeling procedure by SDS-PAGE analysis. DOL of six and five dyes per monomer was measured for hSSB1 and hSSB2, respectively (13 and 11 possible lysine residues for hSSB1 and hSSB2, respectively). However, although we obtained very high DOL with this method, the labeled proteins were not able to bind DNA as efficiently as the unlabeled proteins. Further optimization (lower dye-to-protein ratio, shorter time, and lower
Fig. 3. Fluorescent labeling of hSSB1 and hSSB2 by the Lys light method. Preparations of hSSB1 or hSSB2 were labeled with ATTO 488 NHS (A-488) and compared to the unlabeled preparations (U) by denaturing and reducing SDS-PAGE analysis. M ¼ protein size standards. After electrophoresis, dye attachment was visualized by UV light transillumination of the gel (bottom panels) or after Coomassie Brilliant Blue R-250 staining (top panels).
6 Fluorescent Labeling of Proteins
111
temperature of incubation) is required to find labeling conditions that preserve the original activity of the proteins. Alternatively, the Cys light method could be tried. 3.4. N-Terminus Labeling by the Lys Light Method: RecA Labeling
Because the N-terminal amine group of proteins has a pKa value that is lower than 9, performing a labeling with an NHS ester dye conjugate at pH 7 or lower can in principle target the attachment of the dye to the N-terminal amine group of proteins. At this pH, e-amino groups of solvent-exposed lysines are expected to be fully protonated and nonreactive toward NHS esters. Thus, applying the Lys light method at low pH could in principle be a convenient method to generate a preparation of a fluorescent variant of the protein of interest in which every monomer is specifically labeled at a specific surface position. We tried to apply this principle to the labeling of the RecA protein. First, we produced a variant of RecA with a cleavable N-terminal polyhistidine tag that can be cut by incubation with the TEV protease (see Fig. 4a). After TEV protease cleavage, we can therefore assess whether the incorporated label has been specifically attached to the N-terminus of the protein. 1. The protein sample (3 ml) is supplied in storage buffer A at a concentration of 2 mg/ml and buffer exchanged in labeling buffer C using Econo-Pac 10DG column and recovered in 4 ml. 2. ATTO 488-NHS ester or ATTO 633-NHS ester is dissolved in anhydrous DMSO and added rapidly drop by drop while stirring at twofold molar excess of dye over the protein (see Notes 6 and 7). The reaction mixture is incubated at room temperature for 30 min with stirring. 3. The reaction is quenched by addition of Tris–HCl (pH 7.5) to 0.1 M and further incubated for 5 min. 4. The reaction is processed as in Subheading 3.3, step 4, using storage buffer A as recovery buffer. 5. Figure 4b shows the analysis of the extent of labeling and the label incorporation after TEV cleavage. Clearly, it can be seen that the label has been attached to the N-terminal fragment but also internally. DOL of 2.2 and 2.3 for the ATTO 488 and ATTO 633 labeling was measured spectrophotometrically. Thus, under these conditions of pH 7, specificity of the reaction toward the N-terminus was not observed. Mass spectrometry analysis of the labeled full-length proteins shows that we obtained a very complex and heterogeneous mixture with species that have acquired even four dyes per monomer (see Fig. 4c). Perhaps, the e-amino groups of solvent-exposed lysine residues of RecA have a pKa value lower than expected due to their local environment. It can be noticed that the
112
M. Modesti
Fig. 4. Fluorescent labeling of RecA by the Lys light method at low pH. (a) The RecA protein was produced with an N-terminal His tag cleavable by the TEV protease. (b) The His-RecA preparation was labeled with either ATTO 488 NHS (A-488) or ATTO 633 NHS (A-633) and compared to the unlabeled preparations (U) before and after TEV protease cleavage (+TEV) by denaturing and reducing SDS-PAGE analysis. M ¼ protein size standards. After electrophoresis, the gel was first scanned with a Fluor Imager FLA-5100 (Fujifilm) to detect emission of ATTO 488 (middle panel ) or ATTO 633 (bottom panel ). Next, the gel was stained with Coomassie Brilliant Blue R-250 to reveal all proteins and visualized by bright-field illumination (top panel ). (c) The degree and homogeneity of labeling were analyzed by MALDI-TOF of the full-length protein preparations.
6 Fluorescent Labeling of Proteins
113
DOL determined spectrophotometrically only give average and approximate values. 6. To further optimize the labeling reaction, we performed additional labeling reactions with ATTO 633 NHS as described in steps 1–5 using labeling buffer C in which the MOPS–NaOH, pH 7.0, is replaced with MES–NaOH, pH 6.2, or HEPES–NaOH, pH 8.2. Unfortunately, as judged by fluorescence quantification after TEV cleavage, the labeling in the three different pH conditions behaved similarly giving both N-terminal and internal attachment of the dye (data not shown). We conclude that at least one internal Lys residue is reactive toward the NHS ester group even when the pH is buffered at 6.2. Perhaps, the local environment is influencing the pKa of this putative surface-exposed lysine residue. However, another group has been successful in specifically labeling RecA at the N-terminus using a 5(6)-carboxyfluorescein, succinimidyl ester (5). Thus, the chemical properties of the dye might also influence the specificity of the reaction. Alternatively, we have to optimize the ratio of dye to protein and/ or the reaction kinetics when working with ATTO NHS ester conjugates. 3.5. Fusion to an Intrinsically Fluorescent Protein: Production and Purification of hRPAeGFP
Sometimes, fluorescent labeling by chemical modification is not efficient or destroys the original activity of the protein. An alternative to make the protein glow is to generate a fusion of the protein to an intrinsically fluorescent protein, such as eGFP. Below, we describe this approach for the labeling of hRPA, a single-stranded DNA-binding heterotrimeric protein complex (6). We generated a bacterial polycistronic expression construct phRPA-eGFP (or phRPA-mRFP1 expressing a red fluorescent construct) that produces the large subunit of hRPA tagged at its C-terminus with a polyhistidine-tagged variant of eGFP to facilitate purification (details of plasmid constructs are available upon request). Below, we describe the expression and purification procedures as well as the DNA-binding activity analysis. 1. The expression plasmid is transformed in Rosetta/pLysS cells (Novagen), and cells containing the plasmid are selected on LB plates supplemented with ampicillin (100 mg/ml) and chloramphenicol (34 mg/ml). A single colony is used to inoculate 50 ml of LB+ amp + cm and incubated overnight at 37 C with agitation. The next morning, 3 L of medium in a 6-L flask are inoculated with 30 ml of overnight preculture and incubated at 37 C with vigorous shaking until the OD at 600 nm reaches 0.5. At that stage, IPTG is added to a final concentration of 1 mM and the temperature of the incubator is turned down to +15 C. Incubation at +15 C is continued for at least 16 h. Cells are collected by centrifugation at
114
M. Modesti
3,500 g and resuspended in 10 ml of PBS. The cell paste is flash frozen in liquid nitrogen and stored at 80 C. 2. To extract the protein, the frozen cell paste is quickly thawed in lukewarm water and immediately chilled on ice. Keep working in the cold from now on. The suspension should become very viscous due to cell lysis and release of genomic DNA. Add 1 volume of 2 lysis buffer and resuspend by mixing with a pipette (see Note 17). The viscosity of the sample is reduced by sonication (see Note 18). The mixture is clarified by centrifugation at 20,000 g for 1 h at +4 C (Sorval SS42 rotor). The supernatant containing RPA-eGFP is flushed through a filter device with 0.45-mm pores (Millipore). Cleaner preparations are usually obtained when using a chromatography system, such as an ¨ KTAFPLC (GE Healthcare). However, since most biophysics A laboratories are not equipped with such equipment, we present a procedure that can be performed manually with a syringe and gives very pure preparations as shown in Fig. 5a. The supernatant is loaded (slowly, drop by drop) on a 1-ml HisTrap FF column preequilibrated with 1 lysis buffer. The column is washed with 10 volumes of 1 lysis buffer. The column is further washed in 5 ml steps by increasing the imidazole concentration starting from 5 to 10, 20, 50, and finally 250 mM imidazole in 1 lysis buffer. The bulk of hRPA-eGFP should elute in the last 250 mM imidazole step. The protein sample is dialyzed against 2 L of buffer R for at least 2 h at +4 C, but overnight is also fine (see Note 10). Slowly load the protein sample on a 1-ml HiTrap Heparin HP column equilibrated with buffer R and wash with at least 10 ml of buffer R. Step wash the column (5 ml) by increasing the concentration of KCl in buffer R by 50 mM increments starting from 50 to 500 mM. The bulk of hRPA-eGFP elutes around 200–250 mM. Aliquot the fraction, flash freeze in liquid nitrogen, and store at 80 C. Presence of nucleases should be tested. If desired, the protein can be further purified by chromatography through a 1-ml Hitrap Q HP column by performing exacting the same protocol as for the HiTrap Heparin HP. Elution of hRPA-eGFP occurs around 250–300 mM KCl. We advise to keep every fraction for analysis by SDS-PAGE. 3. Figure 5a shows the denaturing SDS-PAGE analysis of a hRPA-eGFP preparation obtained by the Histrap/Hitrap heparin protocol described above. Figure 5b presents the analysis of the single-stranded DNA-binding activity by electrophoretic mobility shift assay. Reassuringly, the presence of the eGFP tag fused at the C-terminus of the large hRPA subunit does not affect its DNA-binding activity in this assay.
a KDa 150
M RPA70-eGFP
100 75 50 37
RPA32 25 20 15 10
b
hRPA
hRPA-mRFP1
hRPA-eGFP
+
+
+ + + + + +
RPA14
+ + + + +
+ + + + +
Protein ssDNA
U.V. light excitation
mRFP1 emission
eGFP emission
Ethidium bromide and mRFP1 emission
Fig. 5. Purification and single-stranded DNA binding of hRPA-eGFP. (a) SDS-PAGE/ Coomassie staining of an aliquot of the hRPA-eGFP preparation (2 mg) purified on HisTrap and Heparin columns. M ¼ molecular weight standard. (b) Electophoretic mobility shift assay showing that the fusion of eGFP or mRFP1 to C-terminus of the large subunit of hRPA does not interfere with its single-stranded DNA-binding activity. For each preparation, the protein (1 mM for the highest final concentration, and diluted in twofold increments) was incubated with or without 200 ng of PhiX174 single-stranded DNA in a 10 ml reaction volume containing 50 mM KCl, 20 mM Tris–HCl, pH 7.5, 5% glycerol, 1 mM EDTA, 1 mM DTT, and 100 mg/ml BSA. After a 20-min incubation, the binding reactions were directly analyzed by electrophoresis in a 0.6% agarose/ Tris–Borate–EDTA. Before staining, the gel was scanned with a Fluor Imager for detection of mRFP1 or eGFP emission (middle two panels ). Afterward, the gel was stained with ethidium bromide to detect DNA and visualized on a UV light transilluminator (top panel ) or by scanning the gel with a Fluor Imager (bottom panel , notice that with these settings both mRPF1 and ethidium bromide emissions contribute to the signal).
116
M. Modesti
3.6. Indirect Strategies for Protein Fluorescent Labeling
It is also possible to label a protein by adding an affinity tag by genetic engineering, allowing indirectly labeling with secondary fluorescent reagents specific for the tag. Here, we enumerate and briefly describe a few of the possible tags that can be used and cite companies from which reagents can be obtained (see Note 19). 1. The AviTag (Avidity) consists in a 15 amino acid peptide tag (GLNDIFEAQKIEWHE) that can specifically be biotinylated in vivo or in vitro on the lysine residue using the biotin ligase (BirA) from E. coli. Commercially available fluorescent avidin or streptavidin can subsequently be attached “nearly covalently” to the AviTagged protein. 2. The strep tag (IBA-biotagnology) is a very small peptide that can be fused to a protein of interest by genetic engineering. The strep tag mimics biotin and can bind streptavidin or avidin with high affinity. Addition of fluorescently labeled streptavidin or avidin, thus, indirectly labels the biotinylated protein. 3. The FlAsH/ReAsH system (Invitrogen): The small 6-amino acid 1 kDa tetracysteine tag (CCPGCC) that can be fused to a protein of interest by genetic engineering coordinates the FlAsH-EDT2 or ReAsH-EDT2 compounds with high affinity. These biarsenical compounds fluoresce upon coordination to the tetracysteine tag. 4. Fluorescently labeled antibodies specific to the protein of interest could also be used for indirect labeling as long as they do not interfere with its activities. Instead, a number of fluorescently labeled monoclonal antibodies specific for various epitopes that can be fused to the protein of interest by genetic engineering are commercially available. The FLAG, HA, and myc epitopes are among the most popular ones. Finally, since many recombinant proteins are produced as fusions to the glutathione S-transferase or the maltose-binding protein, these moieties can also be used as anchor point for a fluorescent antibody. 5. Fluorescent nanocrystals (quantum dots, Evident Technologies, Invitrogen) functionalized with NHS ester or maleimide groups, or coupled to streptavidin or antibodies, might be used for direct or indirect protein labeling (7, 8).
4. Notes 1. See Molecular Probes Handbook at http://www.invitrogen. com/site/us/en/home/References/Molecular-Probes-TheHandbook.html for a description of the chemistry.
6 Fluorescent Labeling of Proteins
117
2. The 1 M DTT stock solution should be prepared freshly. 3. The Econo-Pac 10DG column is first equilibrated by gravity flow with 20 ml of labeling buffer (or recovery buffer). The 3 ml sample is next loaded on the column and after it has fully entered the gel, add 4 ml of labeling buffer to elute the protein. When performing a buffer exchange after labeling, most of the dye should stay in the column bed. 4. Buffers are deoxygenated by gently bubbling argon gas for 30 min. 5. Most protocols do not recommend high-salt concentrations during labeling. However, we find that many proteins that we work with require at least 250 mM salt (NaCl or KCl) to be maintained in solution. The high-salt concentration does not affect the reactivity of the maleimide toward the –SH groups. Performing the labeling in the presence of 0.5 M NaCl (and sometimes, even up to 1 M NaCl) helps in obtaining more homogeneous labeling by avoiding protein aggregate formation. It is important to keep the pH below 7.5; otherwise, the maleimide group can also react with unprotonated primary amines. 6. We purchase dyes in small quantities, typically 1 mg, and use them promptly. Routinely, we dissolved the dye in anhydrous DMSO (stock kept in a desiccator) because many dyes have a poor solubility in water. Once dissolved in DMSO by vortexing, spin the solution in a microfuge to remove insoluble matters. It is best to immediately use the dye solution in a labeling reaction. However, we find that the dye solution in anhydrous DMSO can be kept for several weeks when stored at 20 C in an O-ring screw cap tube. Reactivity drops with time due to maleimide (or NHS ester) hydrolysis. It is, thus, important to avoid contact with humid environment and test the labeling efficiency of the dye after prolonged storage in anhydrous DMSO. 7. To obtain complete labeling, the choice of the molar ratio of dye to protein during labeling can be critical as well as the incubation time and temperature. We typically use fivefold and one- to twofold molar excess of dye for maleimide and NHS ester reaction, respectively. However, it is wise to first perform small-scale pilot reactions to optimize conditions (see Subheading 3.1, step 7). 8. It is common to find that some or all of the protein forms aggregates after labeling. This is not a surprise since the surface properties of the protein can be changed after labeling. Another dye can be tried or alternatively use a variant of the dye with a different charge, for example. 9. Protein concentration can conveniently be achieved using Amicon Ultra centrifugal filter devices from Millipore. They
118
M. Modesti
come in different sizes and with different MWCOs. Follow the supplier recommendation for the choice of centrifugal speed. We first spin an aliquot of buffer through the filter before adding the protein sample. Be careful not to concentrate too much as the protein might aggregate at high concentration. Proceed first in 5-min time intervals and gentle resuspension with a Pasteur pipette without damaging the filter. 10. We use SnakeSkin Pleated Dialysis Tubing from Pierce (10,000 MWCO). Protein aggregation may occur after dialysis. 11. Before loading the protein sample on the gel, 1 volume of protein sample buffer is added to the protein aliquot and the mixture is heated at 95 C for 5 min. After electrophoresis, the gel is placed in destaining solution and can directly be scanned with a Fluor Imager to visualize and quantify fluorescent signals. Free dye runs with the electrophoresis front and diffuses away after prolonged incubation. Note that this fluorescence imaging procedure is very useful to determine if the dye exhibits nonspecific binding to the protein of interest. For Coomassie staining, the gel is placed in staining solution and incubated for 1 h to overnight on a rotating table. The gel is next destained by several incubations in destaining solution (note that Coomassie quenches the fluorescence). 12. The mobility shift after labeling is often, but not always, detectable. The mobility shift depends on the mass-to-charge ratio, which can be affected by attachment of the dye to the protein and the resolving power of the gel chosen. 13. Measurement and calculation of DOL by spectrophotometry are described in detail at (http://www.atto-tec.com/index. php?id¼62&L¼1). 14. First, equilibrate the PD SpinTrap G-25 spin column by five consecutive 400 ml washes in labeling buffer and then proceed as recommended by the supplier. Do not exceed 100 ml for the sample. 15. Mass spectrometry procedures are not described here since most universities and research centers have nowadays access to a mass spectrometry service. For full-length protein mass measurements by MALDI-TOF, it is important to work at protein concentrations of at least 5 mM and to avoid salt, detergents, and glycerol. Reversed-Phase ZipTips from Millipore can be very convenient. Sinapinic acid often works well as matrix for ionization of full-length proteins. Finer mass spectrometry analysis to identify modified residues after labeling can also easily be performed by in-gel trypsin digestion (or using other proteases). 16. http://www.atto-tec.com/index.php?id¼62&L¼1.
6 Fluorescent Labeling of Proteins
119
17. Protease inhibitor cocktails can be added during resuspension. We typically use 1 mM PMSF in this protocol. Avoid EDTA in your sample because it chelates the Ni2+. Do not go higher than 20 mM ß-mercaptoethanol or add DTT to avoid reducing Ni2+. 18. Sonicate with a microtip in short intervals of max 1 min with interruption on ice to avoid heating of the sample. Alternatively, benzonase nuclease (Novagen) can be added to the sample, but Mg2+ should then be added. 19. There is an important danger in using antibody, quantum dots, or fluorescent streptavidin to label proteins because these reagents may have multiple binding sites and can thus, under the wrong conditions, multimerize the protein of interest.
Acknowledgments Work in our laboratory is supported by Laserlab-Europe, the Association for International Cancer Research (AICR), the Agence pour la Recherche contre le Cancer (ARC), The city of Marseille, The Region of Provence-Alpes-Coˆtes d’Azur, and the Mediterranean Institute of Microbiology (IMM/IFR88). We would like to thank Sabrina Lignon from the mass spectrometry platform, Marielle Bauzan from the fermentation unit, and Yann Denis from the transcriptome platform of the IMM/IFR88 for advice and help with instrumentation and services. We thank Marc Wold (University of Iowa) for the gift of the p11d-tRPA polycistronic expression construct. References 1. Nathan C. Shaner, Paul A. Steinbach, & Roger Y. Tsien. (2005). A guide to choosing fluorescent proteins. Nature Methods. 2, 905–909. 2. Petukhova GV, Pezza RJ, Vanevski F, Ploquin M, Masson JY, Camerini-Otero RD. (2005). The Hop2 and Mnd1 proteins act in concert with Rad51 and Dmc1 in meiotic recombination. Nat. Struct. Mol. Biol. 12, 449–453. 3. Modesti M, Ristic D, van der Heijden T, Dekker C, van Mameren J, Peterman EJ, Wuite GJ, Kanaar R, Wyman C. (2007). Fluorescent human RAD51 reveals multiple nucleation sites and filament segments tightly associated along a single DNA molecule. Structure. 15, 599–609. 4. Richard DJ, Bolderson E, Cubeddu L, Wadsworth RI, Savage K, Sharma GG, Nicolette ML, Tsvetanov S, McIlwraith MJ, Pandita RK,
Takeda S, Hay RT, Gautier J, West SC, Paull TT, Pandita TK, White MF, Khanna KK. (2008). Single-stranded DNA-binding protein hSSB1 is critical for genomic stability. Nature. 453, 677–681. 5. Galletto R., Amitani I., Baskin R.J., Kowalczykowski S.C. (2006). Direct observation of individual RecA filaments assembling on single DNA molecules. Nature. 443, 875–878. 6. Henricksen LA, Umbricht CB, Wold MS. (1994) Recombinant replication protein A: expression, complex formation, and functional characterization. J. Biol. Chem. 269, 11121–11132. 7. Visnapuu M.-L., Duzdevich, D. and Greene, E. C. (2007). Using Total Internal Reflection Fluorescence Microscopy, DNA Curtains, and Quantum Dots to Investigate Protein-DNA
120
M. Modesti
Interactions at the Single-molecule Level. Modern Research and Educational Topics in Microscopy. Me´ndez-Vilas, A. and Diaz, J. (Eds). Microscopy series N 3 Vol. 1, pp 297–308.
8. Visnapuu ML, Greene EC. (2009). Singlemolecule imaging of DNA curtains reveals intrinsic energy landscapes for nucleosome deposition. Nat. Struct. Mol. Biol. 16, 1056–1062.
Chapter 7 Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules Till Korten, Bert Nitzsche, Chris Gell, Felix Ruhnow, Ce´cile Leduc, and Stefan Diez Abstract Recent developments in optical microscopy and nanometer tracking have greatly improved our understanding of cytoskeletal motor proteins. Using fluorescence microscopy, dynamic interactions are now routinely observed in vitro on the level of single molecules mainly using a geometry, where fluorescently labeled motors move on surface-immobilized filaments. In this chapter, we review recent methods related to single-molecule kinesin motility assays. In particular, we aim to provide practical advice on: how to set up the assays, how to acquire high-precision data from fluorescently labeled kinesin motors and attached quantum dots, and how to analyze data by nanometer tracking. Key words: Kinesin, Single molecule, Imaging, TIRF microscopy, Motor protein, Microtubule
1. Introduction In vitro motility assays, where the movement of molecular motors along surface-bound cytoskeletal filaments is imaged by optical methods, have greatly improved our understanding of the mechanochemical properties of motor proteins. First, such assays were performed by binding the motors to large, micron-sized beads, which could be followed by video microscopy (1–3) or held in an optical trap (4–6). Later on, Funatsu et al. pushed the sensitivity of the fluorescence microscope to the limit of being able to visualize individual motor molecules labeled with cyanine-based fluorophores (7). Using total internal reflection fluorescence (TIRF) microscopy, the processive movement of individual kinesin-1 molecules along microtubules was visualized by labeling the motors with Cy3 (8) or with green fluorescent protein (GFP) (9). Since then, the number of single-molecule fluorescence measurements of kinesin and dynein motors interacting with Erwin J.G. Peterman and Gijs J.L. Wuite (eds.), Single Molecule Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 783, DOI 10.1007/978-1-61779-282-3_7, # Springer Science+Business Media, LLC 2011
121
122
T. Korten et al.
microtubules has vastly expanded. Nowadays, not only the stepping of motors is observed in these assays, but also diffusion, (de)polymerizing activities of motors (and other microtubuleassociated proteins), as well as interactions between motors and other proteins on the microtubule lattice (10–19). In this chapter, we describe protocols for the imaging of single kinesin-1 molecules stepping along surface-immobilized microtubules using TIRF microscopy. These procedures should be understood as examples that need to be adapted in order to work for other motors or filaments (see also Notes 17 and 24). As the name indicates, TIRF microscopy excites fluorophores using light that is totally internally reflected at the interface of two materials with different refractive indices (e.g., glass and buffer). This reflected light produces an electromagnetic near field (termed “evanescent field”) that decays exponentially with a decay length of 50–200 nm into the material with lower refractive index (see Fig. 1, left panel; for reviews, see refs. 20–23). A low penetration depth ensures that only fluorophores are excited which are very close to the surface. This “optical sectioning” reduces the background noise caused by molecules diffusing freely in solution, thus allowing the imaging of single fluorescent molecules. Due to the sensitivity of this fluorescence technique, it is essential to use clean glass surfaces that can be passivated efficiently (Subheading 3.2). Microtubules that self-assemble in vitro (Subheading 3.1) can be attached to these surfaces via antibodies
Fig. 1. Imaging single GFP-labeled kinesin-1 molecules moving along microtubules. Left : Schematic drawing of the assay. Microtubules are bound to a passivated glass surface with anti-tubulin antibodies. Single GFP-labeled kinesin-1 motors are imaged using the evanescent field generated by a totally internally reflected laser beam. Middle : Three frames of an image sequence showing a single kinesin-1 molecule (indicated with arrows) moving along a microtubule. Right : Kymograph of the same image sequence (microtubule not shown). Time progresses downward while the horizontal axis represents the distance along the microtubule. To record this data, we used an inverted fluorescence microscope (Zeiss Axiovert 200 M, Zeiss, Jena, Germany) with a commercially available TIRF-slider and an alpha PlanApochromat 100 oil 1.46 NA DIC objective (both Zeiss). Fluorescence signals were recorded using an EMCCD camera (Ixon DV 897, Andor, Belfast, N. Ireland). Excitation was achieved using either a mixed gas argon-krypton laser (Innova 70C Spectra; Coherent, Santa Clara, CA) or a liquid-waveguide coupled Lumen 200 metal arc lamp (Prior Scientific Instruments Ltd., Rockland, MA). Image acquisition and basic data processing were done using MetaMorph (Universal Imaging, Downingtown, PA).
Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules
Streptavidin
Qdot
Antibody GFP
Kinesin-1
Pluronic F127
Microtubule
123
20 y-position (nm)
7
m
8n
10
0
Glass 0
10
20 30 x-position (nm)
40
50
60
number of steps
40
8 nm
walked distance (nm)
60
20
40
20
0 0 2
4
6 8 time (s)
10
12
0
10 20 step size (nm)
30
Fig. 2. Resolving 8 nm steps of a single kinesin-1 molecule using TIRF microscopy. Top-left: Schematic drawing of the assay. In addition to the assay explained in Fig. 1, a streptavidin-coated Qdot is attached to the kinesin-1 molecule via a biotinylated anti-GFP antibody. Top-right: Tracked x–y data of the Qdot positions. The dashed line indicates the average path of the Qdot. Bottom-left: Walked distance of the Qdot projected onto its average path plotted against time. Steps identified by a step-finding algorithm (47) are overlaid. Bottom-right: Histogram of step sizes identified by the step-finding algorithm (data from 12 molecules). The imaging setup used to acquire this data was the same as that was used for Fig. 1.
(Subheading 3.3). Single GFP- or quantum dot (Qdot)-labeled, truncated kinesin-1 molecules can then be observed moving along these microtubules. Using Gaussian model-based tracking algorithms, the position of these molecules can be localized with higher precision than the resolution of an optical microscope (24–28). With bright fluorescent labels, it is possible to resolve 8 nm steps of kinesin-1 (Fig. 2) and even smaller steps produced by two motors working collectively (29).
2. Materials 2.1. In Vitro Polymerization of Microtubules
1. BRB80: 80 mM Piperazine-N,N0 -bis(2-ethanesulfonic acid) (PIPES), pH 6.9, with KOH, 1 mM ethylene glycol tetraacetic acid (EGTA), 1 mM MgCl2. 2. Microtubule polymerization mix: BRB80 supplemented with 24% ml DMSO, 20 mM MgCl2, and 5 mM GTP.
124
T. Korten et al.
3. Tubulin, bovine (Cytoskeleton Inc, Denver, CO; unlabeled #5.11.2146; rhodamine labeled #5.11.2148). Store at 80 C. Stable for ~6 months. 4. Taxol (Paclitaxel; Sigma-Aldrich, Steinheim, Germany, #T7191): Dissolve to a final concentration of 100 mM in DMSO. Store 10-ml aliquots at 20 C. Stable for at least 1 year. 5. BRB80T: Dilute Taxol to a final concentration of 10 ml in BRB80. 2.2. Preparation of the Sample Chambers
1. Coverglasses 18 18 mm and 22 22 mm (Menzel, Braunschweig, Germany, #BB018024A1 and #BB022022A1, ask for thickness 1.5!). 2. Soap bath: 2% Mucasol in water. 3. Acetone. 4. Ethanol. 5. Nanopure water (>18 MO/m). 6. Piranha solution: Carefully mix 30% H2O2 with 97% H2SO4 at a ratio of 1:2 in a suitable container. 7. 0.1 M KOH in nanopure water. 8. Trichloroethylene (TCE)/dichlorodimethylsilane (DDS) solution: 0.05% DDS in TCE. Add TCE first, and stir while adding DDS to TCE (see Note 9). Prepare freshly in the clean container used for cover glass treatment (see Subheading 3.2.4). 9. Methanol. 10. Nescofilm (Roth, Karlsruhe, Germany, #2569.1).
2.3. Surface Passivation and Microtubule Binding
1. Tetraspeck fluorescent microspheres (0.2 mm, Invitrogen, Carlsbad, CA, #T7280): 1:100 dilution in BRB80 (see Subheading 2.1). 2. Anti-beta-tubulin antibody, SAP4G5 (Sigma-Aldrich, Steinheim, Germany, #T7816): Typically, a 1:200 to 1:1,000 dilution in BRB80 is used (see Note 16). The undiluted stock is stored at 4 C for several weeks. 3. F127 solution: 1% Pluronic F127 (Sigma-Aldrich, Steinheim, Germany, #P2443) is dissolved in BRB80 overnight, filtered (0.22-mm syringe filter), and stored at 4 C. Stable for ~6 months. 4. BRB20: 20 mM PIPES, pH 6.9, with KOH, 1 mM EGTA, 1 mM MgCl2. 5. BRB20T: 10 mM Taxol (see Subheading 2.1) in BRB20.
7
Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules
2.4. Imaging GFP-Labeled Kinesin-1 Motors
125
1. Casein (Sigma-Aldrich, Steinheim, Germany, #C7078): Add 35 ml BRB80 (see Subheading 2.1) to 1 g casein in a 50-ml Falcon tube. Place the Falcon tube in a rotating wheel at 4 C and allow the casein to dissolve overnight. Let the large precipitates sediment in the Falcon tube placed upright. Centrifuge the supernatant at 10,000 g at 4 C for 5 min. Filter the supernatant using a 0.22-mm syringe filter. Determine the concentration by measuring the absorbance at 280 nm. Dilute to the desired concentration (e.g., 10 mg/ml), snap freeze 40-ml aliquots in liquid nitrogen, and store at 20 C. Stable for at least 1 year. 2. GFP kinesin-1: We use truncated, GFP-labeled kinesin-1 constructs (rkin430GFP), which consist of the first 430 amino acids of kinesin-1 fused to a GFP and a his-tag at the tail domain. Cloning and purification have been described (30). 3. Antifade solution (see Notes 20–24): Supplement BRB20T (see Subheading 2.3) with final concentrations of 10 mM Taxol (see Subheading 2.1) and antifade reagents: 40 mM D-glucose, 160 mg/ml glucose oxidase (SERVA Electrophoresis GmbH, Heidelberg, Germany, #22778), 20 mg/ml catalase (Sigma-Aldrich, Steinheim, Germany, #C40), and 10 mM 1,4-dithio-DL-threitol (DTT). 4. Motor solution: Add final concentrations of 100 mg/ml casein, 1 mM ATP, and 50–500 pM GFP kinesin-1 to the antifade solution (see Notes 24 and 25).
2.5. Imaging Quantum Dot-Labeled Kinesin Motors
1. Biotinylated anti-GFP antibodies (produced by our antibody facility in-house): Biotinylated using EZ-Link NHS-PEO Solid Phase Biotinylation Kit (Pierce, Rockford, IL, #21450). 2. Qdots 655, streptavidin-coated quantum dots, emission peak at 655 nm (Invitrogen, Carlsbad, CA, #Q10121MP). Store at 4 C. Stable for ~1 year.
3. Methods 3.1. In Vitro Polymerization of Microtubules
Microtubules are commonly reconstituted from purified tubulin. Their physical structure, including the number of protofilaments and the stiffness, strongly depends on the assembly conditions (31–33). Below, we describe the production of stabilized (i.e., nondynamic) microtubules using GTP and the microtubulestabilizing drug Taxol (see Note 1). 1. Add 1.25 ml of microtubule polymerization mix to 5 ml of 4 mg/ml tubulin in BRB80 (containing 3–25% rhodaminelabeled tubulin, see Note 2).
126
T. Korten et al.
2. Incubate at 37 C for 30 min to allow microtubule assembly. 3. Dilute assembled microtubules by adding 493.75 ml of BRB80 supplemented by 10 mM Taxol and store them for up to 1 week at room temperature (see Notes 1 and 3). 4. Immediately before using the microtubules, centrifuge 50 ml of the microtubule solution at 100,000 g for 5 min and resuspend (see Note 4) in 200 ml of BRB80T (final tubulin concentration ~100 nM). 3.2. Preparation of the Sample Chambers
3.2.1. Cover Glass Precleaning
Clean, low-fluorescence glass surfaces are essential for singlemolecule TIRF microscopy of kinesin motors. Formed into simple flow cells, these glass surfaces allow convenient perfusion of solutions for immobilization of microtubules and introduction of kinesin. Unwanted nonspecific adsorption of fluorescent motors can be effectively prevented by surface passivation. 1. Place an equal number of 18 18-mm and 22 22-mm cover glasses into Teflon racks, ensuring that adjacent cover glasses cannot contact. 2. Sonicate for 15 min in a soap bath; rinse in deionized water for 1 min. 3. Bathe, sequentially, in acetone for 10 min, ethanol for 10 min, and nanopure water for 1 min.
3.2.2. “Piranha Solution” Cleaning
1. Transfer the cover glasses directly from the water bath to Piranha solution. Bathe for 1 h, heating the solution to 60 C. Extreme caution should be used, as Piranha solution is highly corrosive and potentially explosive (see Note 5). 2. Transfer racks directly from the Piranha solution sequentially to three nanopure water baths, bathing for 1 min each (see Notes 6 and 7).
3.2.3. Activating OH Groups on the Glass Surface for Silanization Using KOH
1. Transfer racks from the third nanopure water bath to a 0.1 M KOH bath for 15 min. 2. Transfer racks sequentially through two nanopure water baths, bathing in each for 1 min. 3. Remove from water bath and dry cover glasses completely with clean nitrogen gas (see Note 8).
3.2.4. Glass Silanization
1. Gently place the cover glasses into a container filled with TCE/DDS solution so that the glasses are completely immersed; bathe for 1 h. 2. Transfer the silanized glass, sequentially, through three methanol baths, placed in an ultrasonic bath for times of 5, 15, and 30 min.
7
Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules
127
3. Remove from the final methanol bath and dry cover glasses completely with clean nitrogen gas. Store in clean, sealed glass containers (see Notes 10 and 11). 3.2.5. Sample Chamber Assembly and Use
1. Cut out several small (30 2 mm) Nescofilm strips using a (razor) blade (see Notes 12 and 13). 2. Arrange the Nescofilm strips on one side of a silanized 22 22-mm cover glass, forming a series of parallel channels (up to 4), and cut off (using a razor blade) those parts of the strips that protrude over the edges of the cover glass. 3. Place an 18 18-mm silanized cover glass on top, sandwiching the Nescoflm strips. 4. Transfer the assembly (22 22-mm cover glass facing down) to a hot plate at a temperature of 150 C. 5. The two cover glasses are firmly joined by the melted Nescofilm. This process can be monitored by a change in the optical properties of the Nescofilm. Gently applying pressure to the top cover glass ensures a tight seal. 6. Without delay, place the flow chamber onto the surface of a metal (e.g., aluminum) block, at room temperature, for fast cooling. 7. Mount the assembly in an appropriate holder. The process results in channels with a volume of 5 ml, and solutions can then be repeatedly perfused into the channels independently.
3.3. Surface Passivation and Microtubule Binding
We found that a hydrophobic surface can be very efficiently blocked using F127, a tri-block copolymer consisting of a hydrophobic poly(propylene oxide) (PPO) block in between two hydrophilic poly(ethylene glycol) (PEG) blocks. The PPO block strongly adsorbs onto the hydrophobic-rendered glass surface. The outer PEG parts form a polymer brush that is very effective in blocking protein adsorption. Stabilized microtubules are bound to the surface via anti-tubulin antibodies (see Fig. 1). 1. Perfuse Tetraspeck fluorescent microspheres into the hydrophobic-rendered glass flow cell with the assistance of a vacuum (see Notes 14, 15, and 29). 2. After a 5-min incubation, flush out excess microspheres with BRB80 (see Note 15). 3. Perfuse anti-beta-tubulin antibodies into the flow cell (see Notes 16 and 17). 4. After a 5-min incubation, flush out excess antibodies with BRB80.
128
T. Korten et al.
5. Using F127 solution, passivate the remaining exposed surface in order to block any further nonspecific surface binding of proteins. 6. After a 10–30 min incubation, remove excess F127 with 10 channel volumes of BRB80 (see Note 18). 7. Perfuse the solution containing microtubules (see Subheading 3.1). The incubation time depends on the desired microtubule density in the channel. Typical times range from 5 to 15 min. 8. Wash out unbound microtubules with BRB20T (see Note 19). 3.4. Imaging GFPLabeled Kinesin-1 Motors
Due to the nature of TIRF microscopy, only fluorophores that are close to the surface are excited. Therefore, it is possible to have free GFP-labeled motors in solution and still observe single kinesin-1 molecules moving along surface-immobilized microtubules. Using this technique, it is possible to measure biophysical parameters of the motors, such as: speed, run length, and dwell time. However, the limited resolution of an optical microscope in conjunction with fast bleaching (see Notes 20–23) and low fluorescence intensity of GFP does not allow the observation of individual steps of the motor protein. This limitation can be overcome using highly efficient fluorescence probes, like Qdots, together with advanced image analysis (see Subheadings 3.5 and 3.6). 1. Perfuse 2 channel volumes of motor solution into a channel containing surface-bound microtubules. Place the sample onto the microscope stage (see Notes 26–29). 2. Using arc lamp illumination and a rhodamine filter set, focus on the microtubules (see Note 30). Switch to laser illumination (488 nm wavelength) and a GFP filter set. In order to find the optimal TIRF angle, adjust the angle of the laser until the image becomes dark, and then turn back just until single molecules become visible in the whole field of view (see Notes 31 and 32). 3. Switch back to arc lamp illumination and rhodamine filters, move to a different field of view, and record a snapshot of the microtubule positions. Switching to laser illumination and the GFP filter set, acquire a continuous sequence of images, each with 100 ms exposure time (see Notes 33–35). An example microscope setup used to observe a kinesin-1 molecule moving along microtubules is given in the caption of Fig. 1.
7
Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules
3.5. Imaging Quantum Dot-Labeled Kinesin Motors
129
1. Mix 10 nM streptavidin-coated Qdots with 1 nM biotinylated anti-GFP antibodies. Incubate for 5 min. 2. Add 0.1–1 nM GFP-kinesin-1 to the above solution and incubate for 5 min (see Note 36). 3. Prepare the motor solution by supplementing the antifade solution from Subheading 2.4 with 4% (v/v) Qdot–kinesin solution, 100 nM ATP (if individual steps are to be resolved at 10 frames/s), and 100 mg/ml casein. 4. Continue with step 1 from Subheading 3.4 (see Notes 37 and 38).
3.6. Image Analysis: Tracking of Motors
GFP molecules with a diameter of 2 nm or Qdots with a diameter of ~20 nm appear as point sources of light; therefore, their intensity profile can be regarded identical to the point spread function (PSF) of the imaging system. The PSF is well-approximated by a 2D Gaussian (34, 35). Therefore, fluorescent particles can be localized by fitting their camera-captured intensity profiles to a 2D Gaussian. One possibility of implementing this approach is fitting by a radially symmetric 2D Gaussian with a fixed width corresponding to the dimension of the PSF. However, we found it important to also allow radially symmetric 2D Gaussians with nonfixed widths (Eq. 1) as well as elliptical 2D Gaussians (Eq. 2). These additional degrees of freedom accommodate slightly defocused images as well as particles of larger sizes (such as fluorescent Tetraspeck beads in the 200-nm range). " # h ðx x^Þ2 þ ðy ^y Þ2 I ðx; yÞ ¼ exp (1) 2ps2 2s2 I ðx;yÞ ¼
h pffiffiffiffiffiffiffiffiffiffiffiffiffi 2psx sy 1 r2 " !# 1 ðx x^Þ2 ðx x^Þ ðy ^y Þ ðy ^y Þ2 exp r þ ð1 r2 Þ sx sy 2s2x 2s2y (2)
Here, I(x, y) denotes the intensity value at pixel position x, y; h denotes the height of the Gaussian; x^, ^y denote the position of the center of the fitting function; s, sx, sy denote the widths; and r denotes the correlation coefficient determining the orientation of the Gaussian. 3.6.1. Implementation of the Tracking Routines
A tracking algorithm was developed in MATLAB (The MathWorks, Natick, MA). First, every frame is processed individually to determine the positions of all objects: (1) Rough scanning. Gray-scale images are converted to binary images using
130
T. Korten et al.
predefined threshold values. Objects with intensity values above this threshold are classified as particles. (2) Fine scanning. The regions around the particles are fitted by Eq. 1 or 2 using estimated starting parameters. The algorithm employs a nonlinear least squares fitting routine (36, 37) and estimates the error for each parameter. Second, the fitted particle positions from individual frames are connected into tracks using an adapted feature point-tracking algorithm (38). Thereby, a cost function for four subsequent frames is calculated and includes parameters, like position, velocity, direction, or intensity. Occlusions are handled with a post-processing step, where corresponding tracks are merged automatically. The final tracking results can be displayed in an overlay with the original images offering manual verification of the fitting. The described tracking approach was implemented in a MATLAB software package named Fluorescence Image Evaluation Software for Tracking and Analysis (FIESTA) which is freely available (39). Using FIESTA and microtubule-attached Qdots, we were able to resolve the 8 nm steps that kinesin-1 takes on the microtubule (see Fig. 2) and even smaller steps produced by two motors working collectively (29).
4. Notes 4.1. In Vitro Polymerization of Microtubules
1. As an alternative to the taxol stabilization of microtubules described in Subheading 3.1, stabilized microtubules can also be obtained by tubulin assembly in the presence of guanosine 50 -[a,b-methylene] triphosphate (GMP-CPP). For this, supplement 100 ml BRB80 with 2 mM tubulin, 4 mM MgCl2, and 1 mM GMP-CPP. Allow the microtubules to assemble for 2 h at 37 C. Immediately before use, centrifuge the microtubule solution at 100,000 g for 5 min and resuspend in BRB80 (final tubulin concentrations 0.4–4 mM). If particularly stable microtubules (lasting for several weeks at room temperature) are desired, assembled microtubules should be centrifuged immediately after polymerization and resuspended in BRB80 containing 10 mM Taxol. 2. When using chemically labeled tubulin, keep the labeling ratio as low as possible. Controls should be performed to check that labeling does not generate artifacts, such as decreased or increased run length of motor proteins on the microtubules. 3. Low temperatures cause microtubule depolymerization. Therefore, do not keep microtubules on ice or in the refrigerator.
7
Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules
131
4. Resuspending and pipetting microtubules may break them by shearing. This effect can be reduced by slow, controlled pipetting and by cutting off the pipette tip to enlarge the opening. 4.2. Preparation of the Sample Chambers
5. Piranha solution is extremely corrosive. Work with proper protective equipment and under a fume hood. Dispose of Piranha properly. Used solutions must not be stored in closed containers since gas formation continues and closed containers bare the risk of explosions (40, 41). 6. Clean glass surfaces are wetted by water. A test for cleanliness is to check whether a droplet of water spreads immediately over the complete cover glass. If a drop with a contact angle not close to 0 forms, the cover glasses are dirty! 7. Clean glass surfaces are an attractant for dirt. Therefore, it is best to use clean surfaces right away. 8. Use a clean nitrogen source for drying; use a filter on the nitrogen gas line, and clean the nitrogen gun in acetone to remove oil. 9. DDS forms insoluble crystals during storage; store upright in a sealed container at 4 C; avoid agitating the DDS in its bottle. Use a long needle and syringe to pierce the Teflon seal on the DDS bottle and draw off the required amount from the top. Using the syringe, add DDS below the TCE surface to avoid air contact. DDS vigorously reacts with water, resulting in cross-linking and crystal formation. Crystals, if present on cover glass surfaces, greatly reduce the surfaces’ ability to be passivated. Store DDS under nitrogen or in a desiccator. 10. The quality of the silanization can be checked by putting a water droplet onto the glass surface and measuring its contact angle. The water contact angle should ideally be larger than 100 . Over time, the quality of silanization deteriorates. Use DDS-coated glasses within 1 week after preparation. 11. It is essential to prepare the cover glasses at all stages in as clean an environment as is possible; dry glasses in an environment free from any sources of dust; store the glass in doublesealed holders and bags; open the containers only briefly to remove glasses. 12. In place of Nescofilm, double-sided sticky tape can be used to form channels. 13. Using multiple layers of Nescofilm or tape can increase the channel volume; this decreases the surface area-to-volume ratio of the channel, mitigating the effects of any undesirable surface binding.
132
T. Korten et al.
4.3. Surface Passivation and Microtubule Binding
14. The cover glasses used here are hydrophobic. Therefore, a vacuum line is needed to help draw the first solution into the channels. After that, filter paper can be used. Caution is needed not to introduce air bubbles into the channel because air immediately denatures any proteins on the surface. 15. The flow profile in the channel is parabolic; thus, there is slower movement of the solution near the channel walls. In order to exchange solutions in the channels thoroughly, it is necessary to flow several channel volumes of solution. Unless noted otherwise, use 5 channel volumes. 16. Antibody concentrations should be as low as possible to not compromise surface blocking; places where antibodies are attached to the surface are not blocked by F127 in the subsequent surface passivation step. 17. As an alternative to anti-tubulin antibodies, in some cases it can be advantageous to use anti-fluorophore antibodies (42) or neutravidin (18) binding to tubulin labeled fluorescently or with biotin, respectively. 18. The channels can dry out. This damages proteins in the channel (see above). To avoid this, keep the sample chamber in a moist environment by covering it (e.g., with the lid of a pipette tip box) and adding a moist tissue. Alternatively, a drop of buffer can be added to each end of the channel, but in this case evaporation increases the salt concentration. If no additional solutions are to be perfused, channels can be sealed for long-term experiments using immersion oil or vacuum grease. 19. After the microtubules are immobilized on the surface, the low-salt buffer BRB20 is used because it increases the run length of kinesin-1.
4.4. Imaging GFP-Labeled Kinesin-1 Motors
20. In order to reduce photobleaching of fluorescent labels and to provide a reducing environment for the proteins, we add an oxygen-sequestering antifade cocktail to all our motor solutions. The antifade cocktail consists of D-glucose, glucose oxidase, catalase, and DTT. When adding these components, add the glucose oxidase last and just before actually using the motor solution, for this starts the antifade reaction. Once the antifade is complete, do not use the respective solution longer than 15–30 min. Otherwise the pH might drop significantly due to acid produced by the glucose oxidase. 21. High concentrations of glucose oxidase and catalase protein can interfere with the assay. At the same time, it is important to have high enzymatic activities in order to keep the oxygen concentration low. Therefore, it is important to always use
7
Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules
133
glucose oxidase and catalase of the highest available relative activities (i.e., units per mg). 22. For convenience and reproducibility, store all antifade components separately at 20 C in aliquots of 10 ml at a 100 concentration. Catalase and glucose oxidase are dissolved in BRB80 and snap frozen in liquid nitrogen. The frozen aliquots are stable for 6 months; when thawed and stored on ice, aliquots maintain activity for several hours. DTT is dissolved in nanopure water, stored at 20 C, and stable for ~3 months. 23. DTT can be substituted with 2-mercaptoethanol (BME). Using 10 mM BME can prolong the time until bleaching of GFP by ~50%. However, BME is less stable than DTT. Therefore, antifade solution containing BME should not be used longer than 30 min. Also, ensure that the BME stock is fresh and not oxidized by air. 24. When using motor proteins other than kinesin-1, in most cases the buffer conditions of the motor solution have to be adjusted. While BRB buffers are well-suited for microtubules and kinesin-1, our experience is that other motor proteins tend to aggregate under these conditions. In those cases, it can help to use other buffering reagents, vary the salt concentration, and add low amounts of detergents, like 0.1% Tween 20. Never add calcium and always add at least 1 mM EGTA because free calcium ions depolymerize microtubules. Also, phosphate buffers are not very suitable because phosphate is a product of the motor hydrolysis cycle. 25. The concentration of motor protein needed for singlemolecule imaging greatly depends on the quality of surface blocking and the salt concentration. The flow channels have a high surface-to-volume ratio. Therefore, motor protein adsorption to the surface greatly influences the amount of free motor in solution. With increasing salt concentration, the affinity of the motor protein for the microtubule is reduced. Thus, a higher salt concentration requires more freely diffusing motor in solution. 26. To reduce thermal drift during imaging, allow the flow chamber to thermally equilibrate for about 2 min before image acquisition (sample already in contact with the immersion fluid, if non-air microscope objectives are used). 27. Another source of stage drift is airflow (e.g., in airconditioned rooms). Shielding the microscope by a closed box (e.g., made of transparent plastic) provides a simple and effective solution. 28. The epi-fluorescence arc lamps on many microscopes generate considerable heat. To avoid unwanted sample drift, and also heating the sample itself, we do not mount our arc lamp on
134
T. Korten et al.
the microscope and couple the light to the microscope through a multimode optical fiber. 29. To compensate for stage drift, fluorescent Tetraspeck beads can be used as drift control in the sample. 30. In order to minimize photobleaching of GFP-labeled kinesin-1, it is best to always focus on the rhodamine-labeled microtubules and then switch to the GFP filter set (start the recording immediately without refocusing). For this to work, it may be necessary to compensate for chromatic aberrations by measuring the focus offset between the GFP and the rhodamine channels and compensate for this difference before switching channels. Alternatively, a highly corrected objective, such as a Zeiss alpha Plan-Apochromat 100 oil 1.46 NA DIC (Zeiss, Jena, Germany), can be used. 31. Due to the nature of TIRF imaging, the laser comes out of the objective as a collimated beam (as opposed to a confocal microscope). Therefore, extreme precautions must be taken when the laser is on. (1) Make sure that no reflective material is above the microscope. (2) Never add or remove the sample while the laser is on. (3) Place a sheet of paper over the sample while adjusting the TIRF angle. This diffuses the laser beam while helping to see whether the laser still comes out of the objective. (4) Adjust the TIRF angle such that the laser points away from you. 32. Accurate focusing on the back focal plane can be achieved by observing the transmitted laser beam emerging from the objective some meters away on the ceiling, and adjustment of the position of the focusing lens in the TIRF condenser can then be performed. 33. For single-molecule imaging, we typically use a laser beam power of 1–3 mW, measured as it exits the microscope objective. A higher laser power reduces the time until photo bleaching of the motors while a lower power decreases tracking accuracy. 34. A good compromise between collecting enough photons per pixel while maintaining sufficient temporal resolution is critical. We found that for many applications exposure times of 30–100 ms per frame are optimal. 35. Keep your microscope clean, and use good antireflectioncoated filters and lenses. This reduces undesirable interference with reflections of the laser light from the optical components and diffraction from dust on the surfaces of optical components, which otherwise may cause a spatially nonuniform illumination due to interference fringes.
7
Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules
4.5. Imaging Quantum Dot-Labeled Kinesin Motors
135
36. When binding Qdot–streptavidin conjugates to the biotinylated antibody, make sure that there is no other source of biotin that might compete for free streptavidin-binding sites. 37. For imaging of Qdot-labeled kinesin motors, the same considerations as above apply, except that a higher laser power can be used in order to increase spatial and/or temporal resolution. We typically use a laser beam power of 5–10 mW, measured as it exits the microscope objective. 38. Qdots are bright and photostable fluorescent emitters. Therefore they are ideally suited for high-precision nanometer tracking. One drawback of Qdots is that they blink on various timescales (ranging from submilliseconds to many seconds). Although there is no way to eliminate blinking completely, it has been reported that the use of the reducing agents BME and DTT reduces blinking dramatically (43). There are also recent reports about the synthesis of nonblinking Qdots (44–46). The localization uncertainty for GFP molecules is a magnitude higher (10–40 nm) than for Qdots (1–4 nm).
Acknowledgments The authors would like to thank all current and former members of the Howard and Diez labs for the tremendous efforts in helping to develop the described protocols. References 1. Sheetz, M. P., and Spudich, J. A. (1983) Movement of Myosin-Coated Fluorescent Beads on Actin Cables Invitro, Nature 303, 31–35. 2. Yanagida, T., Nakase, M., Nishiyama, K., and Oosawa, F. (1984) Direct Observation of Motion of Single F-Actin Filaments in the Presence of Myosin, Nature 307, 58–60. 3. Spudich, J. A., Kron, S. J., and Sheetz, M. P. (1985) Movement of Myosin-Coated Beads on Oriented Filaments Reconstituted from Purified Actin, Nature 315, 584–586. 4. Svoboda, K., Schmidt, C. F., Schnapp, B. J., and Block, S. M. (1993) Direct observation of kinesin stepping by optical trapping interferometry, Nature 365, 721–727. 5. Rief, M., Rock, R. S., Mehta, A. D., Mooseker, M. S., Cheney, R. E., and Spudich, J. A. (2000) Myosin-V stepping kinetics: A molecular
model for processivity, Proc. Natl. Acad. Sci. U. S. A. 97, 9482–9486. 6. Schnitzer, M. J., Visscher, K., and Block, S. M. (2000) Force production by single kinesin motors, Nat. Cell. Biol. 2, 718–723. 7. Funatsu, T., Harada, Y., Tokunaga, M., Saito, K., and Yanagida, T. (1995) Imaging of Single Fluorescent Molecules and Individual ATP Turnovers by Single Myosin Molecules in Aqueous-Solution, Nature 374, 555–559. 8. Vale, R. D., Funatsu, T., Pierce, D. W., Romberg, L., Harada, Y., and Yanagida, T. (1996) Direct observation of single kinesin molecules moving along microtubules, Nature 380, 451–453. 9. Pierce, D. W., HomBooher, N., and Vale, R. D. (1997) Imaging individual green fluorescent proteins, Nature 388, 338–338.
136
T. Korten et al.
10. Helenius, J., Brouhard, G., Kalaidzidis, Y., Diez, S., and Howard, J. (2006) The depolymerizing kinesin MCAK uses lattice diffusion to rapidly target microtubule ends, Nature 441, 115–119. 11. Varga, V., Helenius, J., Tanaka, K., Hyman, A. A., Tanaka, T. U., and Howard, J. (2006) Yeast kinesin-8 depolymerizes microtubules in a length-dependent manner, Nat. Cell. Biol. 8, 957–960. 12. Bieling, P., Laan, L., Schek, H., Munteanu, E. L., Sandblad, L., Dogterom, M., Brunner, D., and Surrey, T. (2007) Reconstitution of a microtubule plus-end tracking system in vitro, Nature 450, 1100–1105. 13. Brouhard, G. J., Stear, J. H., Noetzel, T. L., Al-Bassam, J., Kinoshita, K., Harrison, S. C., Howard, J., and Hyman, A. A. (2008) XMAP215 is a processive microtubule polymerase, Cell 132, 79–88. 14. Fink, G., Hajdo, L., Skowronek, K. J., Reuther, C., Kasprzak, A. A., and Diez, S. (2009) The mitotic kinesin-14 Ncd drives directional microtubule-microtubule sliding, Nat. Cell. Biol. 11, 717–723. 15. Varga, V., Leduc, C., Bormuth, V., Diez, S., and Howard, J. (2009) Kinesin-8 motors act cooperatively to mediate length-dependent microtubule depolymerization, Cell 138, 1174–1183. 16. Seitz, A., Kojima, H., Oiwa, K., Mandelkow, E. M., Song, Y. H., and Mandelkow, E. (2002) Single-molecule investigation of the interference between kinesin, tau and MAP2c, EMBO J. 21, 4896–4905. 17. Dixit, R., Ross, J. L., Goldman, Y. E., and Holzbaur, E. L. (2008) Differential regulation of dynein and kinesin motor proteins by tau, Science 319, 1086–1089. 18. Telley, I. A., Bieling, P., and Surrey, T. (2009) Obstacles on the microtubule reduce the processivity of Kinesin-1 in a minimal in vitro system and in cell extract, Biophys. J. 96, 3341–3353. 19. Korten, T., and Diez, S. (2008) Setting up roadblocks for kinesin-1: mechanism for the selective speed control of cargo carrying microtubules, Lab Chip 8, 1441–1447. 20. Axelrod, D., Burghardt, T. P., and Thompson, N. L. (1984) Total internal reflection fluorescence, Annu. Rev. Biophys. Bioeng. 13, 247–268. 21. Thompson, N. L., and Steele, B. L. (2007) Total internal reflection with fluorescence correlation spectroscopy, Nat. Protoc. 2, 878–890. 22. Gell, C., Berndt, M., Enderlein, J., and Diez, S. (2009) TIRF microscopy evanescent
field calibration using tilted fluorescent microtubules, J. Microsc. 234, 38–46. 23. Gell, C., Brockwell, D. J., and Smith, D. A. M. (2006) Handbook of Single Molecule Fluorescence Spectroscopy, Oxford University Press, Oxford, UK. 24. Gelles, J., Schnapp, B. J., and Sheetz, M. P. (1988) Tracking kinesin-driven movements with nanometre-scale precision, Nature 331, 450–453. 25. Anderson, C. M., Georgiou, G. N., Morrison, I. E., Stevenson, G. V., and Cherry, R. J. (1992) Tracking of cell surface receptors by fluorescence digital imaging microscopy using a charge-coupled device camera. Lowdensity lipoprotein and influenza virus receptor mobility at 4 degrees C, J. Cell Sci. 101 (Pt 2), 415–425. 26. Hua, W., Young, E. C., Fleming, M. L., and Gelles, J. (1997) Coupling of kinesin steps to ATP hydrolysis, Nature 388, 390–393. 27. Yildiz, A., Forkey, J. N., McKinney, S. A., Ha, T., Goldman, Y. E., and Selvin, P. R. (2003) Myosin V walks hand-over-hand: single fluorophore imaging with 1.5-nm localization, Science 300, 2061–2065. 28. Nitzsche, B., Ruhnow, F., and Diez, S. (2008) Quantum-dot-assisted characterization of microtubule rotations during cargo transport, Nature Nanotechnology 3, 552–556. 29. Leduc, C., Ruhnow, F., Howard, J., and Diez, S. (2007) Detection of fractional steps in cargo movement by the collective operation of kinesin-1 motors, Proc. Natl. Acad. Sci. U. S. A. 104, 10847–10852. 30. Rogers, K. R., Weiss, S., Crevel, I., Brophy, P. J., Geeves, M., and Cross, R. (2001) KIF1D is a fast non-processive kinesin that demonstrates novel K-loop-dependent mechanochemistry, EMBO J. 20, 5101–5113. 31. Meurer-Grob, P., Kasparian, J., and Wade, R. H. (2001) Microtubule structure at improved resolution, Biochemistry (Mosc). 40, 8000–8008. 32. Pierson, G. B., Burton, P. R., and Himes, R. H. (1978) Alterations in number of protofilaments in microtubules assembled in vitro, J. Cell Biol. 76, 223–228. 33. Ray, S., Meyhofer, E., Milligan, R. A., and Howard, J. (1993) Kinesin follows the microtubule’s protofilament axis, J. Cell Biol. 121, 1083–1093. 34. Cheezum, M. K., Walker, W. F., and Guilford, W. H. (2001) Quantitative comparison of algorithms for tracking single fluorescent particles, Biophys. J. 81, 2378–2388.
7
Fluorescence Imaging of Single Kinesin Motors on Immobilized Microtubules
35. Thompson, R. E., Larson, D. R., and Webb, W. W. (2002) Precise nanometer localization analysis for individual fluorescent probes, Biophys. J. 82, 2775–2783. 36. More´, J. J. (1977) The Levenberg-Marquardt Algorithm: Implementation and Theory, in Numer. Anal. (Watson, G. A., Ed.), pp 105–116, Springer Verlag. 37. Coleman, T. F., and Li, Y. (1994) On the Convergence of Reflective Newton Methods for Large-Scale Nonlinear Minimization Subject to Bounds, Math. Program. 67, 189–224. 38. Chetverikov, D., and Veresto´y, J. (1999) Feature Point Tracking for Incomplete Trajectories, Computing 62, 321–338. 39. Ruhnow, F., Zwicker, D., and Diez, S. (2011) Tracking single particles and elongated filaments with nanometer precision, Biophys. J. 100, 2820–2828. 40. Dobbs, D. A., Bergman, R. G., and Theopold, K. H. (1990) Piranha Solution Explosion, Chem. Eng. News 68, 2–2. 41. Erickson, C. V. (1990) Piranha Solution Explosions, Chem. Eng. News 68, 2–2. 42. Reuther, C., Hajdo, L., Tucker, R., Kasprzak, A. A., and Diez, S. (2006) Biotemplated
137
nanopatterning of planar surfaces with molecular motors, Nano Lett 6, 2177–2183. 43. Hohng, S., and Ha, T. (2004) Near-complete suppression of quantum dot blinking in ambient conditions, J. Am. Chem. Soc. 126, 1324–1325. 44. Chen, Y., Vela, J., Htoon, H., Casson, J. L., Werder, D. J., Bussian, D. A., Klimov, V. I., and Hollingsworth, J. A. (2008) “Giant” multishell CdSe nanocrystal quantum dots with suppressed blinking, J. Am. Chem. Soc. 130, 5026–5027. 45. Mahler, B., Spinicelli, P., Buil, S., Quelin, X., Hermier, J. P., and Dubertret, B. (2008) Towards non-blinking colloidal quantum dots, Nat. Mat. 7, 659–664. 46. Wang, X. Y., Ren, X. F., Kahen, K., Hahn, M. A., Rajeswaran, M., Maccagnano-Zacher, S., Silcox, J., Cragg, G. E., Efros, A. L., and Krauss, T. D. (2009) Non-blinking semiconductor nanocrystals, Nature 459, 686–689. 47. Kerssemakers, J. W., Munteanu, E. L., Laan, L., Noetzel, T. L., Janson, M. E., and Dogterom, M. (2006) Assembly dynamics of microtubules at molecular resolution, Nature 442, 709–712.
.
Chapter 8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells Using Single-Molecule and Superresolution Imaging Julie S. Biteen, Lucy Shapiro, and W.E. Moerner Abstract Single-molecule imaging enables biophysical measurements devoid of ensemble averaging, gives enhanced spatial resolution beyond the optical diffraction limit, and enables superresolution reconstruction of structures beyond the diffraction limit. This work summarizes how single-molecule and superresolution imaging can be applied to the study of protein dynamics and superstructures in live Caulobacter crescentus cells to illustrate the power of these methods in bacterial imaging. Based on these techniques, the diffusion coefficient and dynamics of the histidine protein kinase PleC, the localization behavior of the polar protein PopZ, and the treadmilling behavior and protein superstructure of the structural protein MreB are investigated with sub-40-nm spatial resolution, all in live cells. Key words: Single-molecule imaging, Live-cell imaging, Live-cell PALM, Superresolution imaging, Superlocalization, Caulobacter crescentus
1. Introduction Since its advent 20 years ago, single-molecule fluorescence imaging has given rise to a host of exciting experiments (1). Beyond enabling fundamental investigations of the physics of emissive molecules, one main advantage of this technique is its utility in biologically relevant, live-cell experiments. The optical fluorescence microscope is an important instrument for cell biology, as light can be used to noninvasively probe a sample with relatively small perturbation of the specimen, enabling dynamical observation of the motions of internal structures in living cells. Single-molecule epifluorescence microscopy extends these capabilities by achieving nanometer-scale resolution, taking advantage of the fact that one can precisely characterize the
Erwin J.G. Peterman and Gijs J.L. Wuite (eds.), Single Molecule Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 783, DOI 10.1007/978-1-61779-282-3_8, # Springer Science+Business Media, LLC 2011
139
140
J.S. Biteen et al.
point spread function (PSF) of a microscope, allowing the center of a distribution, and thus the exact position of an emitter, to be localized with accuracy much better than the diffraction limit itself. This localization accuracy improves beyond the diffraction limit roughly as one over the square root of the number of detected photons (2). Single-molecule imaging has shown utility in the investigation of live-cell images from a number of different cellular systems (3–9). More recently, single-molecule epifluorescence microscopy has been used to probe the inner workings of live bacteria. The small size of bacterial cells makes the optical diffraction limit particularly restrictive, which has stimulated the push toward superlocalization and superresolution to overcome this obstacle. As a result, the nascent field of bacterial structural biology has benefited greatly from single-molecule investigations of proteins in live cells. The overall shapes of such cells can be seen in a standard light microscope, but those interested in probing subcellular details, such as protein structure and localization, have typically had to resort to in vitro characterization combined with extrapolation to the cellular environment, as well as to indirect biochemical assays. While cryo-electron microscopy (EM) can provide extremely high spatial resolution, fixation or plunge freezing is essential, and in contrast to labeling by fluorescent protein (FP) fusions, methods for identifying specific proteins by cryo-EM are still lacking. As a consequence, bacterial cell biology is an area of study ripe for investigation with direct, noninvasive optical methods of probing position, coupling, and structure, with resolution below the standard diffraction limit. Several groups have extended single-molecule imaging techniques to live bacterial samples (10–15). In this chapter, we focus on the application of single-molecule imaging and single-molecule-based superresolution studies of live cells to investigate the localization, movement, and structure of three proteins, PleC, PopZ, and MreB, that play important functions in Caulobacter crescentus cell cycle progression (10, 11, 14, 15). The C. crescentus cell cycle is characterized by dynamic changes in polar morphology and asymmetric cell division, and thus is a key model organism for study of bacterial developmental processes. Each round of dimorphic C. crescentus cell division gives rise to two distinct daughter cells: a motile, nonreplicating swarmer cell with a polar flagellum and a replicating stalked cell with an adhesive stalk (Fig. 1). This process of asymmetric division is partly driven by patterns of protein localization, which explains why study of specific proteins in this system is of great interest.
8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells. . .
141
Fig. 1. Schematic of the dynamic localization of the proteins PleC (stars ), PopZ (circles ), and MreB (lines ) over the course of the Caulobacter crescentus cell cycle. PleC is localized to the flagellar pole of the nonreplicating swarmer cell. As the swarmer cell differentiates into a stalked cell, PleC becomes delocalized. As the stalked cell progresses into a predivisional cell, a new flagellum is formed opposite the stalk and PleC is localized to the new flagellar pole. PopZ is also positioned at the flagellar end of swarmer cells. Upon differentiation into a stalked cell, a second concentration of PopZ begins to accumulate at the opposite pole (distal from the stalk) with an intensity that increases as the cell cycle progresses. Prior to the onset of the division process, MreB is arranged in a helix along the longitudinal axis of the cell. As the cell begins cytokinesis, MreB is assembled into a ring at the division plane. The helical structure is then reassembled in the two daughter cells.
2. Materials 2.1. Growth Media
1. PYE: 2 g bactopeptone, 1 g yeast extract, 0.3 g MgSO4, and 0.0735 g CaCl2 per L PYE. The mixture is autoclaved for 20 min and vented for 10 min to sterilize. 2. PYE solid media: Add 1.5% Bacto Agar (15 g/L); no CaCl2 needed. 3. M2G minimal media: 930 mL water, 50 mL 20 M2 salts (17.4 g Na2HPO4, 10.6 g KH2PO4, and 5.0 g NH4Cl per L), 1 mL 0.5 M MgSO4, 10 mL 20% glucose, 1 mL 10 mM FeSO4 (Sigma #F-0518), and 5 mL 0.1 M CaCl2.
3. Methods 3.1. Sample Preparation
Single-molecule fluorescence imaging of proteins in living cells can be accomplished via a genetically encoded fusion of the protein of interest to a photostable fluorescent protein emitting at long enough wavelengths to avoid cellular autofluorescence. However, single-molecule imaging requires a low copy number of fluorophores in the cell at any one time so that the ~250-nm
142
J.S. Biteen et al.
diameter image spots from single molecules do not overlap. This low concentration can be achieved in several ways: 1. The gene encoding the protein–FP fusion is integrated into the chromosome in lieu of the gene encoding the native protein. In such strains, the initial FP concentration is generally too high for single-molecule imaging, so the emitting FP population must be reduced via photobleaching before measurements are made (10). 2. The gene encoding the protein–FP fusion is integrated into the chromosome under the control of an inducible promoter while maintaining the gene encoding the untagged wild-type protein at its native chromosomal locus. This has the advantage of allowing the experimenter to control the level of tagged protein while maintaining a significant background of the native protein in order to minimize any physiological effects of the tagging (11, 14). 3. A large number of protein–FP fusions are expressed in the cell, and the number of emitting FPs in any imaging frame can be subject to optical control, for instance by using a photoactivatable FP or by bleaching all of the FPs and then photoreactivating a sparse subpopulation (15). In order to resolve the emission from single FPs above background, cell and sample autofluorescence must be minimized. In C. crescentus, this is accomplished by the following: 1. After initial growth in rich PYE medium, cells are grown in minimal M2G medium for >24 h. 2. Blue and green FP mutants are avoided in favor of yellow and red FPs. At these longer wavelengths, cell autofluorescence is greatly diminished. The experiments discussed below all use the enhanced yellow fluorescent protein EYFP (S65G, V68L, S72A, T203Y) which can be excited at 514 nm. The high resolution provided by single-molecule imaging also requires the cells to lie still on a substrate. This is accomplished by preparing an agarose pad onto which the cells are deposited for imaging. 1. Agarose is added to M2G to a final concentration of 1.5–2% by mass. 2. The agarose/M2G mixture is heated in the microwave for 1–2 min until boiling, shaking every 30 s. 3. 1 mL warm agarose solution is pipetted onto a clean slide and then covered by a clean coverslip to ensure a flat surface.
8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells. . .
143
4. Just prior to imaging, the coverslip is removed, 5 mL cells are deposited on the agarose pad, and the cells are then covered by a clean coverslip. 5. The sample is sealed with wax (melted on a hot plate) to prevent drying (see Note 1). The sample of cells on agarose must also contain fiduciary markers that can quantify stage drift over the imaging time. Fiducials can be any bright, nonbleaching emitters, for example quantum dots (11) or fluorescent beads (15); these are added to the cells in nM concentration (see Note 2). 3.2. Imaging
Single-molecule imaging of bacterial cells is accomplished in a standard wide-field epifluorescence microscopy configuration, and the general imaging considerations have been described in detail (16). Confocal or total internal reflection methods are often not necessary because the entire cell itself is mostly within the depth of focus of a high numerical aperture (NA) microscope (~500–1,000 nm). Because the photon emission rate from a single molecule is typically ten orders of magnitude smaller than the number of scattered photons per second irradiating one pixel of the detector, appropriate filtering is necessary to reject scattered laser light. To optimize the number of detected photons, the single-molecule fluorescence microscope should incorporate a high-NA oil-immersion objective, sharp and carefully chosen filters, and a sensitive detector capable of detecting single photons. In modern experiments, this detector is generally an electronmultiplying charge-coupled detector (EMCCD) (11, 14, 15), though single-molecule imaging has also been accomplished in live cells with an intensified CCD (10). For imaging of EYFP–protein fusions (see Note 3), 1. The sample is excited with a 514-nm Ar+ ion or solid-state laser with similar wavelength. 2. The excitation beam is converted from linear to circular polarization with a l/4 wave plate in order to avoid the preferential detection of certain chromophore orientations (which often cannot be directly correlated with the orientation of the protein under investigation due to the FP fusion). 3. In the detection pathway, a 525-nm longpass dichroic filter and a 530-nm longpass filter are used to reject scattered laser light with an optical density of OD 8–10. Bandpass filters may also be used when necessary to reject other sources of background. 4. In experiments where photoswitching is useful, photoreactivation of EYFP to control the emitting fluorophore density is performed with a 407-nm laser. This light is also
144
J.S. Biteen et al.
circularly polarized and is coupled into the excitation pathway with a dichroic mirror. Superresolution imaging based on single-molecule fluorescence and photoreactivation is accomplished by capturing different sparse subsets of emitters in each imaging frame. Upon photobleaching of a sparse subset, a new subset is generated by photoreactivation and the imaging cycle continues. The final image is generated based on the sum of localizations from many single-molecule imaging frames (15, 17–20). EMCCD cameras can operate at rates as fast as ~50 frames/s for a fairly large field and even faster for subimages. The integration time is chosen based on the experimental system. 1. A longer integration time (100 ms) increases the signal-tonoise ratio when detecting quasi-stationary or slow-moving FPs. These long integration times are applied to the slowly diffusing mobile protein PleC (10) and to MreB monomers incorporated into filaments (11, 15). The long-integrationtime imaging cannot easily observe fast-moving emitters. 2. On the other hand, a shorter integration time is employed to resolve freely or quickly diffusing FPs. Typical fast integration times used in C. crescentus are 15.4–32.2 ms for the quickly diffusing unpolymerized proteins, PopZ and MreB (11, 14). 3.3. Data Processing
Single molecules of the protein–FP fusion are identified by several criteria, including: 1. Digital (one-frame) photobleaching 2. Appropriate number of detected photons 3. A diffraction-limited emission PSF The position of a nanometer-sized isolated emitter is at the center of the PSF; this center can be identified by eye as the brightest pixel in the emission spot (10, 14) or the molecule can be more accurately localized by fitting the PSF to a 2D, symmetric Gaussian function (11, 15). High-density superresolution images based on multiple cycles of single-molecule imaging frames require the fitting software to be automated (15), and it is essential to carefully analyze the data in terms of numbers of photons detected and localization precision in order to represent the position determinations in an unbiased manner (20). Absolute protein positions and direction of movement are determined relative to fiducials and to bright- or dark-field images of the C. crescentus cells. The midpoint and endpoints of each cell are observable, the stalked end is recognized by its stalk, and the cell cycle stage can be identified from the cell morphology and the presence or absence of a stalk. It would also be possible to use DIC images, but the changes in microscope optics compared to those
8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells. . .
145
for single-molecule fluorescence imaging would have to be carefully calibrated. The techniques described in this chapter measure images in only two dimensions, but 2D measurements of diffusion on the surface of the bacterial cell membrane can be corrected appropriately by simulating 3D diffusional movement along an appropriately curved membrane (10, 11). 3.4. Dynamic Localization of the PleC Histidine Protein Kinase
A short region of the PleC histidine kinase, near its N-terminus, is inserted in the inner (cytoplasmic) membrane while the major portion of the protein resides in the cytoplasm. In C. crescentus, the dynamic localization of PleC regulates polar organelle biogenesis, motility, and asymmetric cell division by localizing to the flagellar pole at specific points of the cell cycle (Fig. 1). The swarmer cell must differentiate into a stalked cell, shedding its flagellum and building a stalk, in order to replicate and divide, and during this differentiation process, PleC becomes delocalized. As the stalked cell progresses into a predivisional cell, a new flagellum is formed opposite the stalk and PleC is localized to the new flagellar pole (21). Three different mechanisms have been postulated for the crucial process of PleC localization: free diffusion of PleC through the inner membrane followed by capture and immobilization at a binding site (22, 23), directed insertion of PleC into the inner membrane at the flagellar pole at the time of translation (24, 25), and active transport of PleC to the pole by a motor protein. Deich et al. used single-molecule microscopy to distinguish among these putative mechanisms by measuring the diffusion coefficients and movement directionality of PleC molecules in C. crescentus cells as a function of cell type and position within the cell (10). 1. Strain EJ148 of C. crescentus was created with PleC–EYFP fusions replacing the wild-type PleC on the chromosome and expressed under control of the PleC promoter. 2. The EYFP tags are exposed to 514-nm illumination until all but a sparse subset remained in the emissive state. 3. These isolated single-molecule emitters are imaged and their positions recorded as a function of time in order to measure the movement of PleC molecules in the cells. Figure 2 shows representative fluorescence images of this system. The first image in each of the series in Fig. 2a–c is a dark-field image illustrating the position, size, and orientation of the cell under investigation. The other images in each series represent successive 100-ms imaging frames of the same cells in which each punctate spot is the fluorescence image of a single PleC-EYFP molecule. Figure 2a shows images acquired uninterruptedly (every 100 ms). Due to the relatively slow dynamics of the membrane-bound PleC, such continuous imaging bleached
146
J.S. Biteen et al.
Fig. 2. Dynamics of single PleC histidine kinase molecules. (a–c) Leftmost image: Dark-field image of the cell under investigation. Columns 2–7: Consecutive fluorescence images of single PleC-EYFP molecules spaced by different time intervals: (a) images acquired every 100 ms, (b–c) images acquired every 1 s. (d) Measured mean square displacement (triangles ) and geometry-corrected mean square displacement (circles ) vs. time lag. The corrected data slope of 0.049 mm2/s corresponds to a diffusion coefficient, D ¼ 12 2 mm2/s. (e) Direction of motion in stalked cells without localized PleC (black bars ) and swarmer and predivisional cells with localized PleC (gray bars ), where s indicates motion toward the flagellar pole. Each bin represents one-tenth of the cell length, with the polar bins excluded. Reprinted with permission from ref. 10.
the molecules before significant motion could be observed. Indeed, the lower molecule in Fig. 2a is bleached during the last frame, after only 600 ms of imaging. Time-lapse (TL) imaging was, therefore, utilized to increase the information attained before photobleaching. Figure 2b, c shows two different cells in which a 100-ms imaging frame is acquired every 1 s. In this way, more movement is observed before photobleaching. The single-molecule investigation of PleC in C. crescentus demonstrated several key aspects of PleC dynamics. Bulk microscopy indicates the presence of a bright, fixed PleC locus at the flagellar pole of swarmer and predivisional cells, but by resolving PleC-EYFP on a single-molecule level, a population of mobile PleC molecules could be identified in these cells. Mobile PleC molecules were also seen in the stalked cells, which show no evidence of localization in bulk microscopy. The trajectories of
8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells. . .
147
400 such mobile molecules were further investigated by tracking the PleC-EYFP molecules, and the measured mean square displacement (MSD) of the molecules at different time lags is shown in the triangles in Fig. 2d. Because the observed motion is a 2D projection of 3D motion along a membrane surface, the actual distance moved by the molecule is larger than the observed movement. By simulating diffusion on a cell surface, modeled as a cylinder with spherical ends, the measured MSD is converted to a geometry-corrected MSD (circles in Fig. 2d). The geometrycorrected MSD is linear with the time lag, Dt, consistent with Brownian diffusion with a diffusion coefficient, D ¼ MSD/(4Dt), of 12 2 mm2/s. A normalized coordinate system was created in which the variable s (0 s 1) indicates the position of a molecule along the long axis of a C. crescentus cell. The position s ¼ 0 represents the flagellar pole, and s ¼ 1 the stalked pole. The geometrycorrected coefficients D for the mobile PleC-EYFP molecules were classified based on the cell type in which the molecules were imaged (stalked, swarmer, or predivisional) and their position in the cell along the s axis. No measurable difference in D was observed for the different cell types. The geometry-corrected coefficient D was also not found to vary significantly as a function of s, except for a small reduction in D at the cell flagellar pole of swarmer and predivisional cells, perhaps attributable to immobilization at the binding site during the course of observation. Finally, Fig. 2e shows the direction of PleC-EYFP motion along the s axis, where s indicates motion toward the flagellar pole, as a function of position and cell type. Here, stalked cells, which do not have localized PleC, are in black, and swarmer and predivisional cells, which have localized PleC in addition to the mobile PleC under consideration, are in gray, and each cell is divided into ten equal bins, so length bin 2 includes data for molecules starting with 0.1 < s < 0.2. All molecules were equally likely to move in the +s or s direction, regardless of cell type or PleC position. The motion of PleC in the membrane appears to be governed by Brownian diffusion and is the same at all times in the cell cycle and all positions in the cell. In addition, the absence of directed motion argues against the existence of an active transport mechanism. Overall, the results of single-molecule investigations of PleC in C. crescentus were consistent with the diffusion and capture model of PleC localization at the flagellar pole. 3.5. Diffusion Dynamics of the Polar Protein PopZ
The ability of single-molecule microscopy to track individual intracellular proteins was subsequently applied to the prolinerich protein PopZ in live C. crescentus cells. PopZ is required to anchor the C. crescentus chromosome origins to the cell poles via the DNA-binding protein ParB, an anchoring that is critical for the control of its temporal and spatial localization in the cell cycle.
148
J.S. Biteen et al.
PopZ is positioned at the flagellar end of nonreplicating swarmer cells (Fig. 1). Upon differentiation into a stalked cell, a second concentration of PopZ begins to accumulate at the opposite pole (distal from the stalk) with an intensity that increases as the cell cycle progresses. This second polar accumulation of PopZ is, thus, in place to tether the newly replicated chromosomal origin when it completes its transit across the cell. The mechanism of PopZ localization was investigated on a single-molecule level (14). 1. The merodiploid strain GB175 of C. crescentus was created with an unperturbed popZ locus and popZ-eyfp under control of the vanillate promoter at the vanA chromosomal locus. 2. Leakage of the vanillate promoter ensures the expression of 1–10 copies of PopZ-EYFP per cell in a background of untagged PopZ molecules. 3. These isolated single-molecule emitters are imaged and their positions recorded as a function of time. Two subpopulations of PopZ were, thereby, identified: polelocalized molecules that remained fixed within the measurement localization accuracy (60 nm) and mobile molecules that were resolved with fast (32.2 ms) imaging frames. Figure 3a, d shows the motion of two molecules tracked in one cell under investigation. In Fig. 3a, the tracks are overlaid on a transmitted white-light image of the cell with the cell outline shown (dotted line) while in Fig. 3d, the longitudinal positions of the same two molecules are recorded as a function of imaging time. In Fig. 3a, d, the gray traces (indicated by the white arrow in Fig. 3a) show the position of a polar-localized molecule, and the black traces show the trajectory of a mobile molecule that explores a large portion of the cell interior. Figure 3b–f shows the trajectories of 12 fixed (gray) and mobile (black) PopZ-EYFP molecules in five cells. The mobile PopZ-EYFP molecules exhibited random, diffusive movement in the cytoplasm during most of the time they were tracked, though on some occasions, they remained fixed at a pole (e.g., Fig. 3d at 3–4 s). This movement is consistent with a diffusion and capture mechanism for PopZ localization at the cell pole. The spatial dynamics of single PopZ molecules would not have been observable without single-molecule imaging. 3.6. Movement of the Structural Protein MreB
The structural protein MreB mediates polarity, chromosome segregation, and cell shape in C. crescentus (26–28). Bulk-level imaging and biochemical studies of the cytoplasmic MreB protein are consistent with the formation of a dynamic superstructure made up of actin-like filaments. This structure provides the machinery for key cellular processes, and the dynamics of MreB molecules were investigated on a single-molecule level (11).
8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells. . .
149
Fig. 3. Visualization of single PopZ-YFP molecules in live cells. (a) Positions of two PopZEYFP molecules within a cell, with heavy lines tracking the distance moved between 32.2-ms frames. One molecule (black ; indicated by arrow at top right ) remains localized to the pole, and the other (dark gray ) has increased mobility. The tracks are overlaid on a white-light image of the cell, which is outlined in black. (b–f) Timedependent behavior of 12 single molecules in five cells. Stationary and mobile molecule positions are both tracked with heavy lines. The heavy lines are dotted during dark (blinking-off) periods. Thin horizontal dotted lines in (b) and (d) mark the positions of the poles opposite to the pole at which stationary molecules are localized. The trajectories in (d) correspond to the longitudinal position of the molecules tracked in (a) as a function of time. Reprinted with permission from ref. 14.
1. The merodiploid strain LS3813 of C. crescentus was generated by integrating a single Pxyl::eyfp-mreB construct into the xylX locus in a C. crescentus chromosome already containing a wild-type, unlabeled copy of MreB under its endogenous promoter (26). 2. Cells with only a small number of isolated EYFP-MreB molecules are prepared by incubating the cells in the M2G minimal medium with 0.0006–0.003% xylose for 4 h. 3. The cells are imaged with 15.4-ms imaging frames in order to resolve the tagged MreB molecules. Figure 4a1 shows one such frame, in which three single molecules were captured. The image is smoothed by applying
150
J.S. Biteen et al.
Fig. 4. Analysis of motion of EYFP-MreB. (a) 15.4-ms integration time fluorescence images of single EYFP-MreB molecules in a C. crescentus cell. White line shows the cell outline. (a1) 15.4-ms epifluorescence image showing three single EYFP-MreB molecules. The top and bottom molecules (arrowheads ) are stationary on this timescale, and the middle molecule (arrow ) is mobile. (a2) Smoothed image of (a1) obtained by applying a low-pass filter. (a3) Recorded trajectory of the mobile (middle ) molecule in (a1). (a4) Summed image of 450 sequential imaging frames. The fluorescence from the two stationary molecules is still evident, whereas the middle molecule does not appear on this 6.93-s integration timescale (scale bar, 1 mm). (b) Measured (open circles ) and geometry-corrected (filled circles ) MSD of fast-moving MreB molecules. The solid line is a linear fit of the corrected data. (c) Velocity autocorrelation of fastmoving molecules; this quantity drops to zero at the very first time lag. (d) Inverted-contrast, 100-ms integration time fluorescence images of single slow-moving EYFP-MreB molecules in a C. crescentus cell at 30-s intervals. The arrows illustrate the directional movement of the molecules and the black line shows the cell outline. The superlocalized track of a single molecule leads to a superresolution image of the filament through which the molecule is treadmilling. (e) Measured MSD versus time lag for slow-moving MreB molecules. The solid line is a quadratic fit to the data, indicative of directional motion. (f) Velocity autocorrelation of slow-moving MreB; this quantity remains positive over at least 80 s. Reprinted with permission from ref. 11.
a low-pass filter (3 3 kernel of 0.0625, 0.125, 0.0625, 0.125, 0.25, 0.125, 0.0625, 0.125, and 0.0625) in Fig. 4a2 for enhanced visibility. In this cell, two subpopulations of MreB were distinguished from one another: slow-moving molecules and fast-moving molecules (arrowheads and arrow, respectively, in Fig. 4a1). The trajectory of the fast-moving molecule in Fig. 4a1, obtained by fitting the fluorescence image in every frame to a 2D Gaussian function, is plotted in Fig. 4a3. This molecule diffused rapidly and explored a large portion of the cell. Four hundred and fifty successive frames were summed to form the fluorescence image in Fig. 4a4. After this 7-s integration
8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells. . .
151
time, the fluorescence from the two slow-moving molecules was still evident, but emission from the fast-moving molecule was smeared out over many pixels and did not appear. The white line in Fig. 4a shows the outline of the C. crescentus cell. Because the dynamics of the moving molecules were unchanged in the presence of the MreB-depolymerizing drug A22, the fast-moving EYFP-MreB molecules were ascribed to a free, unpolymerized population. The behavior of the unpolymerized single molecules was further characterized based on 111 trajectories. The observed MSD of the fast-moving molecules is plotted as a function of time lag (Dt) in the open circles of Fig. 4b. Based on the first four points, a diffusion coefficient of D ¼ MSD/(4Dt) of 1.11 0.18 mm2/s was extracted. This value of D is smaller than expected for a free cytoplasmic protein, but is consistent with the motion of a membrane-associated protein. Since the observed motion represents a 2D projection, simulations were again performed to correct for the three-dimensionality of the cell, and the geometry-corrected MSD of the molecules is plotted in the filled circles of Fig. 4b. The geometry-corrected MSD is linear with the time lag, Dt, consistent with diffusion along the cell membrane with D ¼ MSD/(4Dt) of 1.75 0.17 mm2/s. The velocity autocorrelation function, CV(t), was also calculated for these molecules. As shown in Fig. 4c, for the fastmoving molecules, CV(t) dropped to zero at the very first time step, indicating a nondirected random walk. Further experiments addressed the behavior of the slowmoving MreB. Because slow-moving EYFP-MreB molecules were not observed in the presence of the MreB-depolymerizing drug A22, the slow-moving molecules represent polymerized MreB. The dynamics of 120 single members of this subpopulation were carefully examined. Since the molecules were stationary within the 30-nm resolution of the measurement over the course of the 7-s integration time shown in Fig. 4a4, time-lapse imaging was used. Specifically, 100-ms imaging frames are separated by 9.9 s. In this way, the slow movement of polymerized MreB molecules could be followed over a longer time period before photobleaching. Figure 4d shows eight such fluorescence images in reverse contrast, where a single molecule is tracked for 220 s. The molecule moves from left to right, and then turns and moves right to left at a downward angle, shown by the arrow. The observed MSD of the slow-moving MreB molecules is plotted as a function of Dt in Fig. 4e. Here, the MSD was characterized by a quadratic dependence on time lag, as is typical of directed motion. Also consistent with directed motion is the computed velocity autocorrelation function, CV(t), shown in Fig. 4f, in which CV(t) remains positive until t ¼ 80 s. The slow, directed motion of polymerized MreB was further explored by analyzing the trajectories of each individual molecule
152
J.S. Biteen et al.
in the context of two putative models for the motion of an MreB molecule in an MreB filament: (1) the filaments into which the monomers are incorporated were moving or (2) the monomers themselves were treadmilling through largely stationary filaments, analogous to the motion of actin. If the motion of polymerized MreB is due to whole filament movement, then the observation time for the slow-moving single molecules should be limited by the photobleaching time of the EYFP, independent of experiment timescale. The total emission time of EYFP-MreB before photobleaching under continuous emission was measured, with an average on-time of 4.6 s. However, the true irradiation time before photobleaching of the molecules imaged with 9.9-s time lapses was found to be only 0.8 s. Most of the fluorescence from the polymerized MreB molecules, therefore, disappeared before photobleaching occurred likely as a result of dissociation from the end of a filament and the onset of fast diffusive motion not resolvable with the 100-ms imaging frames. This result indicates that MreB molecules treadmill through filaments with fixed ends. Given that the polymerized MreB molecules exhibited directed motion, a speed value was extracted from each singlemolecule trajectory. The average speed was 6.0 0.2 nm/s. From its crystal structure, the length of an MreB monomer is 5.4 nm (29); the average speed, therefore, corresponded to 1.2 monomer additions per second in steady-state fixed MreB filaments. The average length traveled by a polymerized MreB molecule before dissociation, which corresponds to the filament length, was also measured by considering only MreB molecules that became polymerized after the start of imaging and that dissociated before the end of imaging. The average length of 128 MreB filaments was 392 23 nm – quite small relative to the average cell length of 3.5 mm. Most of the trajectories moved perpendicular to the long axis of the cell. A smaller number of trajectories were oblique. These nonperpendicular trajectories were characterized as moving toward the swarmer pole or toward the stalked pole, but no preferred orientation was observed. 3.7. Cell CycleDependent Superresolution Structure of MreB
By fitting the PSF of isolated nanoscale emitters with 30-nm localization accuracy and following these superlocalized molecules over time, the previous experiment attained superresolution via single-molecule tracking, as displayed in Fig. 4d. That is, as a single labeled MreB treadmills through a filament, the track shows the shape of the filament. This approach was, however, limited by the isolated fluorophore requirement to ~1–3 tagged MreB molecules, and therefore ~1–3 trajectories per cell. To address this upper bound and more fully observe the full superstructure in one cell, photoactivated localization microscopy (PALM) was used to extend the investigation of EYFP-MreB to cells expressing high concentrations of these molecules, and thus to visualize the
8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells. . .
153
superstructure formed by a collection of MreB filaments in these cells (15). In particular, bulk visualization of MreB in C. crescentus indicates that MreB forms a cell cycle-dependent structure (26, 28, 30). Prior to the onset of the division process, MreB is arranged in a helix-like shape along the longitudinal axis of the cell. As the cell begins cytokinesis, MreB is assembled into a ring at the division plane. The helical structure is then reassembled in the two daughter cells (Fig. 1). Single-molecule-based superresolution imaging was applied to measure these structures. 1. C. crescentus cells expressing 100–1,000 copies of EYFP-MreB are created by incubating the strain LS3813, described above, in the M2G minimal medium with 0.006–0.075% xylose for 4 h. 2. The cells are imaged with 100-ms imaging frames in order to resolve only the slow-moving, polymerized MreB molecules. The initial density of fluorescent tags in these cells was initially too great for isolated, single-molecule imaging, but this could be addressed based on the fact that apparently bleached EYFP molecules can be reactivated with violet light (31). 3. After the bleaching of all emissive EYFP-MreB molecules, a sparse subset is reactivated with 407-nm irradiation, where the reactivation intensity of 103–104 W/cm2 is chosen such that at most one EYFP molecule is reactivated in each diffraction-limited region. 4. The positions of these isolated molecules are determined by PSF fitting. 5. The reactivation and imaging process was repeated 20 times until a superresolution image could be reconstructed from the sum of the localizations. To illustrate the photoreactivation process, Fig. 5a–f shows the fluorescence emission from EYFP-MreB molecules in a single cell over time. Each punctate white spot is the emission from a single molecule, and the fluorescence images (acquired with 100-ms integration times) are superimposed on a reversedcontrast white-light image of the cell. After the initial bleaching step, two single EYFP-MreB molecules were observed in the cell (Fig. 5a). Additional imaging with 514-nm light bleached all the fluorophores (Fig. 5b). A 2-s dose of 407-nm laser illumination returned some EYFP-MreB to an emissive state (Fig. 5c, e), though all fluorophores are apparently bleached prior to the 407-nm pulse (Fig. 5d, f). 1. The positions of these molecules are calculated from a fit of the fluorescence images to a 2D Gaussian function and are recorded for every such frame.
154
J.S. Biteen et al.
Fig. 5. Superresolution imaging of MreB superstructure. (a–f) Reactivation of single EYFP-MreB fusions in the same live C. crescentus cell using 407-nm light. Fluorescence images of single EYFP-MreB molecules from 514-nm excitation are overlaid on a reversed-contrast, white-light image of the cell. (a) Initial image showing two isolated emissive EYFP-MreB molecules. (b, d, f) After photobleaching by 514-nm irradiation, the cell contains no emissive EYFP-MreB. (c, e) Reactivated EYFP-MreB molecules are observed following a short 407-nm reactivation pulse (scale bar, 1 mm). (g–j) TL-PALM images of EYFP-MreB in C. crescentus cells created by fitting molecule positions in imaging frames, such as those in (a), (c), and (e). (g, h) Quasi-helical structure in a stalked cell. (i, j) Midplane ring in a predivisional cell. Fluorescence PALM images are shown in (g) and (i). The PALM images in (h) and ( j ) are from the same cells as in (g) and (i), respectively, overlaid on a reversed-contrast, white-light image of the cell (scale bars, 300 nm). Reprinted with permission from ref. 15.
2. At the end of a 4-min acquisition period, these localization events are summed. 3. To form a final superresolution image, each localized molecule is depicted in the reconstruction as a unit area Gaussian with fixed width equal to the average statistical localization accuracy of 30–40 nm. This technique, termed “Live-Cell PALM,” enabled the imaging of the EYFP-MreB superstructure in C. crescentus. In particular, it resolved several MreB bands spanning the length of stalked cells and a tightly focused midplane ring in predivisional cells. Importantly, understanding superresolution features derived from many image acquisitions requires careful consideration of the emitter photophysics and the dynamics, if any, of the underlying structure. Since, as described above, polymerized MreB molecules treadmill slowly along MreB filaments, time-lapse imaging, as described above, was incorporated into live-cell PALM to increase the number of localization events given the fact that the maximum labeling concentration in live-cell experiments is limited due to changes in the cell morphology at high concentrations of fusion
8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells. . .
155
protein. Specifically, a 900-ms delay was introduced between each 100-ms imaging frame, and the sample was only illuminated with 514-nm light during the short acquisition time. 407-nm reactivation pulses were again applied after photobleaching to generate novel sparse subsets of emissive EYFP-MreB. Figure 5g–j presents the results from TL-PALM imaging of EYFP-MreB in C. crescentus. Two distinct MreB superstructures were identified in the cells: a quasi-helical arrangement in a stalked cell (Fig. 5g, h) and a midplane ring in the predivisional cell (Fig. 5i, j). In Fig. 5h, j, the TL-PALM reconstruction is superimposed on a reversed-contrast, white-light image of the cell in which the EYFP-MreB was visualized. The images obtained by TL-PALM were more continuous than those acquired without the introduction of dark periods. Furthermore, the use of TL increased the number of localization events (487 in Fig. 5g and 330 in Fig. 5j) to the point, where the Nyquist criterion for 40-nm resolution was satisfied (32), giving rise to a true superresolution reconstruction of the MreB superstructure in C. crescentus.
4. Summary and Outlook Single-molecule imaging permits high-resolution measurements of fluorescent emitters in time and space. This noninvasive, nonperturbative technique is quite useful for the study of single protein molecules in live cells. By applying the techniques of single-molecule microscopy to EYFP-tagged proteins in live C. crescentus cells, the diffusion dynamics, localization patterns, and structure of three key proteins, PleC, PopZ, and MreB, could be explored. By providing a noninvasive tool for high-resolution imaging that is beginning to approach the spatial resolution of cryo-electron microscopy, single-molecule, superresolution imaging can enable investigations of many more intracellular processes and has become an important biophysical tool. Because the single-molecule-based imaging techniques described here are based on wide-field microscopy, they can further provide the basis for more sophisticated experiments. Specific challenges that are being addressed include improving the axial resolution of single-molecule microscopes (33–35), incorporating multiple excitation and emission wavelengths (36, 37), and imaging thick samples (38).
156
J.S. Biteen et al.
5. Notes Axial drift of the sample, as caused, for instance, by microscope objective settling or by sample drying, is particularly detrimental to superresolution experiments. These notes outline several strategies used by the experimenters to address this concern. 1. Sealing the cells-on-agarose-pad with wax prevents excessive drying of the agarose pad over time. This is important in order to maintain constant cell sample thickness. 2. Despite sealing in wax, samples settle over the course of the experiment. It is, therefore, important to ensure that the coverslip is in close contact with the microscope stage and not hanging through an opening in the microscope stage. 3. The microscope objective settles over time, changing the instrument focus and blurring the acquisition. We have, therefore, found it necessary to let the microscope settle after initial focusing, and then use a PIFOC P721 piezo scanner/objective positioner for any needed further focus adjustments.
Acknowledgments The authors warmly thank members of the Shapiro laboratory, specifically Ellen Judd, Harley McAdams, Grant Bowman, and Zemer Gitai, as well as members of the Moerner laboratory, specifically, Jason Deich, Marcelle Koenig, So Yeon Kim, Anika Kinkhabwala, Michael Thompson, and Nicole Tselentis, for their contributions to this work. This research was supported in part by DARPA Grant MDA-972-00-1-0032, NSF Grant MCB0212503, Department of Energy Grants DE-FG02-04ER63777 and DE_FG02-05ER64136 9 (to LS), and NIH Grants P20-HG003638, R01-GM086196, R01-GM085437, R01GM51426 (to LS), and R01-GM32506 (to LS). References 1. Ambrose W. P., Moerner W. E. (1991) Fluorescence spectroscopy and spectral diffusion of single impurity molecules in a crystal Nature 349, 225–7. 2. Thompson R. E., Larson D. R., Webb W. W. (2002) Precise nanometer localization analysis for individual fluorescent probes Biophys J 82, 2775–83. 3. Sch€ utz G. J., Kada G., Pastushenko V. P., Schindler H. (2000) Properties of lipid micro-
domains in a muscle cell membrane visualized by single molecule microscopy EMBO J 19, 892–901. 4. Sako Y., Minoghchi S., Yanagida T. (2000) Single-molecule imaging of EGFR signalling on the surface of living cells Nat Cell Biol 2, 168–72. 5. Moerner W. E. (2003) Optical measurements of single molecules in cells Trends Anal Chem 22, 544–8.
8 Exploring Protein Superstructures and Dynamics in Live Bacterial Cells. . . 6. Vrljic M., Nishimura S. Y., Moerner W. E., McConnell H. M. (2005) Cholesterol Depletion Suppresses the Translational Diffusion of Class II Major Histocompatibility Complex Proteins in the Plasma Membrane Biophys J 88, 334–47. 7. Vrljic M., Nishimura S. Y., Brasselet S., Moerner W. E., McConnell H. M. (2002) Translational Diffusion of Individual Class II MHC Membrane Proteins in Cells Biophys J 83, 2681–92. 8. Lee H. D., Dubikovskaya E. A., Hwang H., Semyonov A. N., Wang H., Jones L. R., Twieg R. J., Moerner W. E., Wender P. A. (2008) Single-Molecule Motions of Oligoarginine Transporter Conjugates on the Plasma Membrane of Chinese Hamster Ovary Cells J Am Chem Soc 130, 9364–70. 9. Harms G. S., Cognet L., Lommerse P. H. M., Blab G. A., Schmidt T. (2001) Autofluorescent Proteins in Single-Molecule Research: Applications to Live Cell Imaging Microscopy Biophys J 80, 2396–408. 10. Deich J., Judd E. M., McAdams H. H., Moerner W. E. (2004) Visualization of the movement of single histidine kinase molecules in live Caulobacter cells Proc Nat Acad Sci USA 101, 15921–6. 11. Kim S. Y., Gitai Z., Kinkhabwala A., Shapiro L., Moerner W. E. (2006) Single molecules of the bacterial actin MreB undergo directed treadmilling motion in Caulobacter crescentus Proc Nat Acad Sci USA 103, 10929–34. 12. Xie X. S., Choi P. J., Li G. W., Lee N. K., Lia G. (2008) Single-Molecule Approach to Molecular Biology in Living Bacterial Cells Annu Rev Biophys 37, 417–44. 13. Conley N. R., Biteen J. S., Moerner W. E. (2008) Cy3-Cy5 Covalent Heterodimers for Single-Molecule Photoswitching J Phys Chem B 112, 11878–80. 14. Bowman G. R., Comolli L. R., Zhu J., Eckart M., Koenig M., Downing K. H., Moerner W. E., Earnest T., Shapiro L. (2008) A Polymeric Protein Anchors the Chromosomal Origin/ParB Complex at a Bacterial Cell Pole Cell 134, 945–55. 15. Biteen J. S., Thompson M. A., Tselentis N. K., Bowman G. R., Shapiro L., Moerner W. E. (2008) Super-resolution imaging in live Caulobacter crescentus cells using photoswitchable EYFP Nat Methods 5, 947–9. 16. Moerner W. E., Fromm D. P. (2003) Methods of Single-Molecule Fluorescence Spectroscopy and Microscopy Rev Sci Instrum 74, 3597–619.
157
17. Betzig E., Patterson G. H., Sougrat R., Lindwasser O. W., Olenych S., Bonifacino J. S., Davidson M. W., Lippincott-Schwartz J., Hess H. F. (2006) Imaging intracellular fluorescent proteins at nanometer resolution Science 313, 1642–5. 18. Rust M. J., Bates M., Zhuang X. (2006) Subdiffraction-limit imaging by stochastic optical reconstruction microscopy (STORM) Nat Methods 3, 793–5. 19. Hess S. T., Girirajan T. P. K., Mason M. D. (2006) Ultra-high resolution imaging by fluorescence photoactivation localization microscopy Biophys J 91, 4258–72. 20. Thompson M. A., Biteen J. S., Lord S. J., Conley N. R., Moerner W. E. (2010) Molecules and Methods for Super-resolution Imaging Methods Enzymol 475, 27–59. 21. Wheeler R. T., Shapiro L. (1999) Differential localization of two histidine kinases controlling bacterial cell differentiation Mol Cell 4, 683–94. 22. Shapiro L., McAdams H., Losick R. (2002) Generating and exploiting polarity in bacteria Science 298, 1942–6. 23. Rudner D. Z., Pan Q., Losick R. M. (2002) Evidence that subcellular localization of a bacterial membrane protein is achieved by diffusion and capture Proc Nat Acad Sci USA 99, 8701–6. 24. Steinhauer J., Agha R., Pham T., Varga A. W., Goldberg M. B. (1999) The unipolar Shigella surface protein IcsA is targeted directly to the bacterial old pole: IcsP cleavage of IcsA occurs over the entire bacterial surface Mol Microbiol 32, 367–77. 25. Robbins J. R., Monack D., McCallum S. J., Vegas A., Pham E., Goldberg M. B., Theriot J. A. (2001) The making of a gradient: IcsA (VirG) polarity in Shigella flexneri Mol Microbiol 41, 861–72. 26. Gitai Z., Dye N., Shapiro L. (2004) An actin-like gene can determine cell polarity in bacteria Proc Nat Acad Sci USA 101, 8643–8. 27. Gitai Z., Dye N. A., Reisenauer A., Wachi M., Shapiro L. (2005) MreB actin-mediated segregation of a specific region of a bacterial chromosome Cell 120, 329–41. 28. Dye N. A., Pincus Z., Theriot J. A., Shapiro L., Gitai Z. (2005) Two independent spiral structures control cell shape in Caulobacter Proc Nat Acad Sci USA 102, 18608–13. 29. Van Den Ent E., Amos L. A., Lowe J. (2001) Prokaryotic origin of the actin cytoskelecton Nature 413, 39–44.
158
J.S. Biteen et al.
30. Figge R. M., Divakaruni A. V., Gober J. W. (2004) MreB, the cell shape-determining bacterial actin homologue, co-ordinates cell wall morphogenesis in Caulobacter crescentus Mol Microbiol 51, 1321–32. 31. Dickson R. M., Cubitt A. B., Tsien R. Y., Moerner W. E. (1997) On/Off Blinking and Switching Behavior of Single Green Fluorescent Protein Molecules Nature 388, 355–8. 32. Shroff H., Galbraith C. G., Galbraith J. A., Betzig E. (2008) Live-cell photoactivated localization microscopy of nanoscale adhesion dynamics Nat Methods 5, 417–23. 33. Juette M. F., Gould T. J., Lessard M. D., Mlodzianoski M. J., Nagpure B. S., Bennett B. T., Hess S. T., Bewersdorf J. (2008) Threedimensional sub-100 nm resolution fluorescence microscopy of thick samples. Nat Meth 5, 527–9. 34. Huang B., Wang W., Bates M., Zhuang X. (2008) Three-Dimensional Super-Resolution Imaging by Stochastic Optical Reconstruction Microscopy Science 319, 810–3.
35. Pavani S. R. P., Thompson M. A., Biteen J. S., Lord S. J., Liu N., Twieg R. J., Piestun R., Moerner W. E. (2009) Three-dimensional, single-molecule fluorescence imaging beyond the diffraction limit by using a double-helix point spread function Proc Nat Acad Sci USA 106, 2995–9. 36. Shroff H., Galbraith C. G., Galbraith J. A., White H., Gillette J., Olenych S., Davidson M. W., Betzig E. (2007) Dual-color superresolution imaging of genetically expressed probes within individual adhesion complexes Proc Nat Acad Sci USA 104, 20308–13. 37. Bates M., Huang B., Dempsey G. T., Zhuang X. (2007) Multicolor super-resolution imaging with photo-switchable fluorescent probes Science 317, 1749–53. 38. Vaziri A., Tang J., Shroff H., Shank C. V. (2008) Multilayer three-dimensional super resolution imaging of thick biological samples Proc Nat Acad Sci USA 105, 20221–6.
Chapter 9 Fluorescence Microscopy of Nanochannel-Confined DNA Fredrik Persson, Fredrik Westerlund, and Jonas O. Tegenfeldt Abstract Stretching of DNA in nanoscale confinement allows for direct visualization of the genetic contents of the DNA on the single DNA molecule level. DNA stretched in nanoscale confinement also allows for studies of DNA–protein interactions and DNA polymer physics in confined environments. This chapter describes the basic steps to fabricate the nanostructures, to perform the experiments, and to analyze the data. Key words: DNA, Nanochannels, Single molecule, Fluorescence
1. Introduction Stretching of single DNA molecules in nanochannels occurs spontaneously due to the confinement and does thus not require any active application of force, for example through chemically attached anchor groups. Visualization of stretched DNA using fluorescence microscopy provides a diffraction-limited resolution corresponding to approximately 1 kbp. Furthermore, the length of the DNA in the channels scales linearly with the total length of the DNA (1). Confining DNA in nanochannels has previously shown applicability to both single-molecule studies of DNA–protein interactions (2, 3), and studies of DNA conformation and dynamics (1, 4–6). Possible future applications may include fingerprinting and investigations of finer structural rearrangements in the genome, such as copy number variations (CNV), microdeletions and insertions, as well as translocations (7). There are a few important parameters to consider when dealing with confined DNA: 1. Contour length – total length of the DNA backbone, here denoted L.
Erwin J.G. Peterman and Gijs J.L. Wuite (eds.), Single Molecule Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 783, DOI 10.1007/978-1-61779-282-3_9, # Springer Science+Business Media, LLC 2011
159
160
F. Persson et al.
2. Persistence length – length-scale over which the DNA can be considered a rigid rod, here denoted P. 3. Effective width – a measure of the width of the DNA, composed of the physical width of the DNA (~2 nm) and an electrostatic contribution (8). The effective width is here denoted weff. When a long DNA molecule is free in solution it forms a coil, often characterized by its radius of gyration (RG). When confined in a tube-like confinement with a diameter Dav smaller than RG, the DNA stretches out in the tube. As long as the diameter of the tube is larger than P, the DNA can backfold and adopt an elongated coiled up conformation. In this regime, commonly denoted the deGennes regime, the DNA is modeled as a series of noninteracting blobs, where the DNA inside each blob behaves as it would in free solution. This leads to an extension, r, of the DNA along the channel of (9): r weff P 1=3 / ; 2 L Dav 2 , the parameter relating to the tube dimensions, can be replaced Dav by the geometric average of the height, D1, and the width, D2, of a 2 rectangular channel, Dav ¼ D1 D2 (10). What is convenient from an experimental point of view is that r scales linearly with L. This means that a position along the stretched DNA can be directly related to a position along the contour of the DNA, i.e., the sequence, with a resolution primarily determined by the degree of stretching and the optics of the microscope. The authors recommend refs. 1, 4, 6, 11–13 for further insight into the polymer physics of confined DNA. There is also a vast literature on general polymer theory, notably the books by deGennes (14), Doi and Edwards (15), and Rubinstein and Colby (16).
2. Materials 2.1. Fabrication of Chips
There are a multitude of ways to fabricate nanostructured chips depending on the facilities and equipment available (see Note 1) (17). We here present a fabrication scheme based on fused silica wafers, for which the following are needed: 1. Fused silica wafers (Available from Hoya). 2. 110-mm thick fused silica coverslips for sealing of the chips. The thickness is optimized for compatibility with oilimmersion objectives. (Available from Valley Design). 3. Access to clean-room equipment for photo (UV) and electron-beam (e-beam) lithography and reactive ion etching
9
Fluorescence Microscopy of Nanochannel-Confined DNA
161
(RIE), as well as standard resists (e.g., AZ (photo lithography) and ZEP (e-beam lithography) resists) and chemicals from any large supplier. 2.2. Chemicals
Two important additives are b-mercaptoethanol (BME) (see Note 2) for suppression of photobleaching and polyvinylpyrrolidone (PVP) (see Note 3) for suppression of electroosmosis when using electrophoresis. Note that genomic length DNA should be handled with wide-orifice pipettes to avoid shear-induced breakage. (Available from e.g., Molecular Bioproducts)
2.2.1. Buffer for DNA Experiments
Tris–Borate–EDTA (TBE) is a standard buffer for DNA studies, especially for electrophoresis due to its low conductivity that ensures a low degree of Joule heating. TBE buffer consists of the following: 1. Tris (tris(hydroxymethyl)aminomethane): Buffering agent for slightly basic conditions (pH range 7– 9). 2. Boric acid: Weak acid that improves the buffering capabilities of Tris. 3. EDTA (Ethylenediaminetetraacetic acid): Chelating agent that scavenges multivalent metal ions, in particular magnesium (Mg2+). Since multivalent metal ions are common cofactors for many enzymes, such as nucleases that digest DNA, the removal of these ions will prevent enzymatic degradation of DNA.
2.2.2. Protocol for Preparing 1 L 5 TBE Buffer
1. Prepare 0.5 L of 0.5 M EDTA solution by weighing out 93.06 g of disodium EDTA (372.24 g/mol) and adding it to 350 mL of water (see Note 4). EDTA will not go into solution until pH is adjusted to 8.0. Add NaOH pellets to the solution, one by one, while stirring vigorously on a magnetic stir plate. Monitor solution pH using a digital pH meter so as not to overshoot. Bring the final solution volume to 0.5 L with water. 2. Prepare a 5 TBE solution by adding 20 mL of 0.5 M EDTA solution from item 1, 54 g of Tris (121.1 g/mol) and 27.5 g of boric acid (61.8 g/mol) to 800 mL of water. Then adjust pH to 8.0–8.5 by adding HCl while monitoring pH. Bring final solution volume to 1 L with water. 3. Autoclave the buffer. This stock buffer can be diluted to any arbitrary ionic strength. For the following steps, 0.5 TBE buffer is used as an example. 4. Dilute the 5 TBE buffer ten times. Then use a syringe with a 0.2-mm filter to aliquot approximately 1.2 mL of buffer into a large number of 1.5 mL microcentrifuge tubes. Store these
162
F. Persson et al.
tubes in the refrigerator for future use. Degas the buffer prior to use. The final 0.5 TBE solution contains 44.6 mM Tris, 44.5 mM boric acid, and 1 mM EDTA and has an ionic strength of approximately 15 mM at pH 8.5. 2.2.3. Protocol for Staining 1 mL of 10 mg/mL DNA at a Dye Ratio of 10:1 in 0.5 TBE Buffer
Always use wide-orifice pipette tips to handle DNA solutions. 1. Create 250 mL of 50 mg/mL solution of DNA in 0.5 TBE (see Note 5). 2. Pipette 769 mL of 0.5 TBE buffer into a separate 1.5-mL microcentrifuge tube. 3. Pipette 47.5 mL of 0.5 TBE buffer into a 0.65-mL tube. Add 2.5 mL of YOYO-1 from the stock solution (1 mM). This creates a 50 mM dye solution. Work in low light from now on to avoid bleaching of the dye. 4. Pipette 31 mL of the 50 mM dye solution from item 3 into the buffer-filled tube from item 2. This creates a solution with a dye concentration of 1.55 mM. Vortex and centrifuge the solution to evenly distribute the dye. 5. Pipette 200 mL of DNA from item 1 into the buffer filled tube from item 4. Do not ever vortex or centrifuge solutions containing DNA – that will fragmentize the DNA. In order to mix the DNA, use a wide-orifice tip and gently pipette a part of the solution a minimum of three times while evenly distributing the ejected solution throughout the tube. 6. To evenly distribute the dye throughout the population of molecules, wrap the tube in aluminum foil and heat the solution to 50 C for 3 h and then store at 4 C.
2.2.4. Protocol for Preparing 500 mL Loading Buffer and 100 mL DNA in Loading Buffer
1. Mix 485 mL of degassed buffer with 15 mL of BME (see Note 6) in a 0.65-mL microcentrifuge tube. BME will collect at the bottom of the tube, so mixing by pipetting and/or vortexing is essential at this stage. Note that vortexing should not be performed after the DNA has been added. Below, we refer to this as the loading buffer. 2. Pipette 95 mL of the loading buffer into a 0.65-mL microcentrifuge tube. 3. Add 5 mL of the 10-mg/mL solution of stained DNA using wide-orifice pipettes. Mix the solution very gently with the pipette. Work in low light to protect the DNA, and wrap the tube in aluminum foil once the DNA-solution is made. Below, we refer to this as the DNA loading sample.
2.3. DNA Samples
In order to characterize the experimental techniques, it is necessary to use monodisperse DNA. There are a few different purified and monodisperse DNA solutions commercially available, and using
9
Fluorescence Microscopy of Nanochannel-Confined DNA
163
Table 1 Selection of commercially available DNA molecules Name
Length (kbp)
Supplier
l-DNA
48.5
New England Biolabs
l-DNA concatamers (in gel)
48.5 n (n 1–50)
New England Biolabs
Yeast DNA (in gel plug)
3500 – 5700
Bio-Rad Laboratories
T4GT7-DNA
166
Nippon Gene
Charomid 9 (circular)
19.7–42.2
Nippon Gene
T7
39.9
Yorkshire Bioscience
restriction enzymes different size distributions can be obtained. Table 1 above lists a few common examples of commercially available, purified and monodisperse DNA. 2.4. Fluorescence Microscopy
For a thorough introduction to microscopy, the authors recommend the MicroscopyU Web site from Nikon (http://www. microscopyu.com), especially the tutorial section on fluorescence microscopy (http://www.microscopyu.com/articles/fluorescence/ index.html), as well as ref. 18. Owing to the low light levels and the risk of photodamaging the DNA, the optical system must be designed to maximize photon detection probability and signal-to-noise ratio. Key considerations are as follows: 1. High-quality filters with high transmission (~90%) in the wavelength region relevant for the dye used and low transmission in the rest of the spectrum, corresponding to a high optical density (OD > 5). 2. Objectives (40–100) with high numerical apertures (NA). Oil-immersion objectives readily achieve NA of 1.4 but have a working distance (WD) of only 100–200 mm. To combine a good NA and a large WD, a 60 water immersion objective with an NA of 1.0 and a WD of 2 mm, originally designed for electrophysiology by Nikon, can be used. This provides sufficient clearance to accommodate thick and easy-to-use cover glasses in the chip-fabrication process. 3. A detector with high quantum yield (QY) and low noise. Electron-multiplying CCDs (EMCCD) have an integrated, virtually noiseless, amplification on the CCD chip. As opposed to intensification technologies based on multichannel plates, the EMCCD is not easily damaged by excessive light levels.
164
F. Persson et al.
The EMCCD can be back-thinned to allow for a QY approaching 90–95% over the visible spectrum. Less expensive EMCCDs are not back-thinned and thus suffer from QYs that are roughly a factor of two lower. To minimize thermal noise, the CCD is normally cooled from 50 to 100 C in modern cameras. EMCCDs are available from, for example, Andor, Photometrics, and Hamamatsu. See Note 7 for examples of additional optical tools and technologies. 2.5. Addressing the Chip
The two most common ways of manipulating DNA in fluidic systems are by electrophoresis or pressure-driven fluid flow. In order to have both these capabilities, a chip holder with both electrical and air pressure connections to internal reservoirs can be used (Fig. 1) (19). The holder can be fabricated in Lucite® (PMMA), allowing the sample to be illuminated from the top and making it easy to detect bubbles trapped in the reservoirs in the chip holder. However, Lucite® has poor resistance to solvents, as it swells and dissolves easily. For experiments involving more aggressive solvents, a holder made in PEEK (Polyetheretherketone) is more suitable, but then the holder is opaque. The holder in Fig. 1 has eight independently accessible reservoirs that are connected to the fabricated chips by pressing the chip against o-rings mounted in the bottom of the reservoirs using a stainless steel frame that is screwed into the chip holder. The other end of the reservoirs is sealed with o-rings and screws with integrated electrodes, supplying the electrical connection to the buffer for electrokinetic transport. Pumps are needed for controlling DNA using pressure-driven flow. Standard diaphragm pumps capable of producing pressures of up to 5 Bar are sufficient in most cases (available from VWR).
Fig. 1. Example of chuck for mounting samples. Eight reservoirs linked by o-rings to the fluidic access holes on the chip. Each reservoir is individually addressable by pressure and electrical connections. Reproduced with permission from ref. 19.
9
Fluorescence Microscopy of Nanochannel-Confined DNA
165
When using pressure-driven flow the pressure is routed through a network of valves giving the possibility of applying pressures to selected reservoirs while others are kept at ambient pressure. To control the pressure, a needle valve can be used as a leak valve, which enables the pressure to be controlled with an accuracy of down to 2 mBar. Accessories such as manifolds, needle valves, and tubing to direct and control the pressure are available from Cole Parmer. When using electrophoresis to control the DNA, a power supply and electrodes are needed. Platinum wires dipped into the DNA solution in the reservoir are often sufficient as electrodes. The electrophoretic mobility of DNA is on the order of 1 mm/s per V/cm. 2.6. Data Analysis
Commonly used software packages for data analysis are as follows: 1. ImageJ – A Java™ based freeware image processing and analysis software developed at the National Institutes of Health, USA (http://rsbweb.nih.gov/ij/). The software benefits from the extensive use of open-source plug-ins developed by users. The MBF plug-in set from the Biophotonics Facility at McMaster University is recommended (http://www. macbiophotonics.ca/imagej/). 2. MatLab – A common high-level technical computing language from The Mathworks™. 3. FreeMat – Open-source freeware available at http://freemat. sourceforge.net/. 4. GNU Octave – Freeware available at http://www.gnu.org/ software/octave/.
3. Methods 3.1. Design and Fabrication of Chips 3.1.1. Design
A careful design of the DNA-visualization device is crucial for its user-friendliness. Typically, two or four U-shaped inlet channels (50 mm 1 mm, each connected to two reservoirs) for efficient fluid transport are combined with nanoscale channels for stretching of the DNA. In this way, the sample can be transported quickly through the large channels to the entrance of the nanochannels by applying a driving force across the microchannel, enabling rapid exchange of buffer. Using nanochannels of dimensions 100 nm 100 nm ensures both a relatively high degree of DNA stretching (~60%) without encountering many of the problems that appear when the channel size approaches the persistence length of DNA (~50 nm). The degree of stretching can be tuned by altering the buffer conditions (20), such as the ionic strength.
166
F. Persson et al.
See Notes 8–10 for examples on how to include extra functionalities on the chip. An important concern is the introduction of the DNA into the nanochannels. There is a significant entropic barrier between the large micron-scale channels and the nanochannels (21). To facilitate the entry of the DNA, a gradual change in entropy can be used. This can be implemented by a gradually evolving degree of confinement combined with obstacles, such as pillar arrays, for prestretching the DNA (1). 3.1.2. Fabrication
When fabricating nanofluidic channels for optical observation of stretched DNA, there are some requirements to consider: 1. The channels should be sealed. 2. At least one side (substrate or lid) must be optically transparent. 3. The surface of the channels should be negatively charged with a minimal roughness to prevent sticking and entanglement of the DNA. 4. The material used should be hydrophilic to allow for easy wetting of the channels. In the following section, a commonly used fabrication process based on fused silica is outlined. A full-scale clean room, with spinners for resist deposition, mask aligners for exposure of micron scale patterns and an electron-beam writer for definition of nanoscale structures, is required (see Note 11). Reactive-ion etchers are used for etching channels with straight walls. In order to align the nanostructures and the microchannels, it is useful to first define alignment marks in the wafer periphery. This can be done by either etching or depositing metals on the wafer (see Note 12), the latter described below. It is assumed that the clean room used has its own standard processes for the following steps.
3.1.3. Definition of Alignment Marks
1. Treat the fused silica wafers with HMDS (hexamethyldisilazane) to increase resist adhesion. 2. Spin-coat and bake a combination of resists used for liftoff, e.g., a LOR/AZ or other similar sandwich constructs, to enable a pattern with an undercut. 3. Expose and develop the resist to create the undercut structure. 4. Run a low-power oxygen descum plasma to remove remaining resist residues. 5. Evaporate a 5-nm Cr (or Ti) adhesion layer and subsequently a 50–80-nm thick Au layer.
9
Fluorescence Microscopy of Nanochannel-Confined DNA
167
6. Strip the resist using a chemical stripper, e.g., Microposit Remover 1165 or acetone (see Note 13). 3.1.4. Definition of Nanochannels
1. Treat the fused silica wafers with HMDS to increase resist adhesion. 2. Spin-coat and bake a 150–250 nm thick layer of ZEP520A e-beam resist. ZEP is chosen because of its good dry-etch resistance. Other resists can be used, but they often require deposition of an extra metallic etch mask. 3. Thermally evaporate 15 nm Al on top as a discharge layer. (This is only needed when working with isolating substrates such as fused silica.) 4. Expose the resist (exposure dose approximately 280 mC/cm2 at 100 kV). 5. Remove the Al layer using, for example, MF322 developer. 6. Develop the resist using, for example, ZED N50 developer. 7. Run a low-power oxygen descum plasma to remove remaining resist residues. 8. Etch the nanochannels into the fused silica using RIE with CHF3/CF4 chemistry. 9. Strip the resist using a chemical stripper, e.g., Microposit Remover 1165.
3.1.5. Definition of Microchannels
1. Treat the fused silica wafers with HMDS to increase resist adhesion. 2. Spin-coat and bake a 2–5 mm thick layer of photoresist, e.g., an AZ resist, that has relatively high etch resistance. 3. Expose and develop the resist. 4. Run a low-power oxygen descum plasma to remove remaining resist residues. 5. Etch the microchannels (approximately 1 mm deep) using RIE with CHF3/CF4 chemistry. 6. Strip the resist using a chemical stripper, e.g., Microposit Remover 1165 or acetone.
3.1.6. Drilling of Access Holes
There is a multitude of ways of producing access holes through a wafer (see Note 14). Here, we describe a setup based on powder blasting. 1. Spin-coat at least 5 mm photoresist on both sides of the wafer. 2. Cover the backside (i.e., the nonstructured side) with an adhesive plastic film, e.g., 70 mm thick Nitto SWT 20 film (see Note 15). 3. Make holes through the film over the reservoir structures using a scalpel or for example laser ablation.
168
F. Persson et al.
4. Powder-blast using 50–110 mm sized Al2O3 particles from the backside of the wafer (i.e., the nonstructured side). A small powder-blasting tool and the powder can be obtained from Danville Materials. 5. Remove the film, strip the resist in a chemical stripper and/or acetone, and carefully clean the wafers in an ultrasonic bath. 3.1.7. Sealing of the Chips
The last step in the production of the chips is sealing. This can be done in several different ways depending on the material of the chips (see Note 16). Fused silica can be bonded covalently via condensation of hydroxyl groups when two surfaces are brought together. Table 2 summarizes two standard ways of creating a high density of the necessary hydroxyl groups, involving thorough cleaning to remove organic residues and subsequent surface activation (see Note 17). For the RCA-based method, the hydrogen peroxide should be added after the mixture has reached the correct temperature to avoid disintegration of the hydrogen peroxide.
3.2. Chemicals
When using techniques for studying single DNA molecules, it is very important that the molecules are kept in a controlled environment and not subjected to reactive contaminants such as radicals or enzymes that damage or digest DNA. Some of these enzymes, such as endonucleases, are present on our skin to break down foreign DNA that we come in contact with. It is, therefore, crucial that gloves are worn at all times when handling DNA and that all tools and pipette tips that come in contact with either the
Table 2 Two fusion-bonding protocols for fused silica (see Note 17) RCA-based
Piranha-based
Chemical
Time
Chemical
RCA2 at 80 C (1:1:5 HCl–H2O2–H2O)
10 min Piranha (1:3 H2O2–H2SO4)
Time 20 min
Rinse carefully with DI water for 5 min RCA1 at 80 C (1:1:5 NH4OH–H2O2–H2O) 10 min Ammonium hydroxide (NH4OH) 40 min Rinse carefully with DI water for 5 min Blow dry in N2 Press together by hand to form a prebond Anneal at 1,050 C for at least 3 h (ramp temperature at approximately 300 C/h for both heating and cooling)
9
Fluorescence Microscopy of Nanochannel-Confined DNA
169
buffer or the DNA samples have been autoclaved or sterilized in another way, e.g., by wiping them with ethanol. 3.2.1. Fluorescent Labeling of DNA
Dimeric cyanine dyes such as YOYO®-1 (YOYO) (Invitrogen, Carlsbad, CA, USA) are extensively used for imaging single DNA molecules (4, 22) due to their high binding affinity (KA ¼ 1010–1012 M1) and a fluorescence enhancement upon binding to DNA of over 1,000, which ensures a low fluorescence background from unbound dye molecules (23). Figure 2 shows the absorption and emission spectra as well as the chemical structure of YOYO (24). At moderate binding densities, YOYO binds to DNA by bis-intercalation and increases the contour length of DNA by approximately 0.68 nm per YOYO molecule (one base pair (0.34 nm) per intercalation event) (25, 26). Note that the highly charged YOYO affects the physical properties of DNA (11, 27). When staining the DNA, it is important to know the resulting dye ratio ([base pair]–[dye molecule]). The easiest way of calculating this is to use the molar concentration of the dye and DNA, respectively. If the concentration of DNA is known in mg/mL, it is easily converted using the molar mass of one DNA base pair (bp), Mbp ¼ 618 g/mol. The DNA concentration can also be determined by absorbance measurements, using the molar absorption coefficient for DNA at l ¼ 260 nm, e260 ¼ 13,200 cm1 M1 (base pair) or 50 mg/mL for OD 1 (1 cm optical path length).
Fig. 2. (Left ) Excitation (solid )/emission (dashed ) spectra of YOYO®-1. Adapted from data from Invitrogen. Inset: TOTO®-1 intercalating in DNA, visualized using the open-source viewer JMol. The structure (PDB ID: 108D) was determined by nuclear magnetic resonance (24). (RIGHT ) Chemical structure of YOYO®-1.
170
F. Persson et al.
Knowing the molar concentrations of the DNA and the dye, the dye ratio is readily obtained. In the case of dimeric cyanine dyes, such as YOYO, the DNA should not be stained at dye ratios higher than 10:1 to avoid crowding of dye on the DNA. 3.3. Running Experiments (Loading of DNA)
Prior to mixing the loading buffer, the TBE buffer should be degassed in a vacuum chamber for 2 h to reduce the amount of dissolved air in the system and to avoid bubble formation in the channels. The degassing process can be shortened to about 10–20 min by ultrasonic agitation. Fresh loading buffer should be prepared in conjunction with every experiment, since the BME degrades with time (see Note 18). Use the loading buffer to wet the chip. Placing droplets over the fluidic access holes is normally sufficient to wet the chip by capillary forces (see Note 19). Remaining air bubbles can be removed by applying a pressure across the channels or by submerging the chip in buffer and placing it in a vacuum desiccator overnight for degassing. Using degassed buffer solution during experiments ensures that bubbles formed during the capillary wetting (28) are absorbed into the liquid and also prevents the DNA from degrading. When the chip is properly wetted, it is mounted in the chip holder. Thereafter, the DNA loading sample is added to the desired reservoirs and loading buffer to the remaining reservoirs. The DNA can be moved through the chip by electrokinetic transport or pressure-driven flow. A pressure difference of approximately 100 mBar results in reasonable sample velocities when transporting the DNA in the micron-sized channels, from the inlet reservoirs to the nanostructures. Once the DNA molecules are in close proximity to the inlets of the nanochannels, the driving force is shifted so that it is applied across the nanochannel array instead. The most convenient way to introduce DNA into the nanochannels is to pulse the pressure, switching rapidly between a low pressure and 1–3 Bar. When the DNA molecule of interest is in the nanochannel and in the field of view of the CCD, stacks of images of at least 200 frames are recorded (see Note 20). During the measurements, the coordinates of at least two alignment marks on the chip as well as the stage coordinates for all the recorded stacks should be recorded. This allows for both rotational correction as well as accurate localization of the molecule within the fluidic network in the case of a more intricate design of the chip. There are a few things to be aware of during image/data acquisition: 1. If one of the ends of the DNA molecule appears much brighter, it might have been folded while entering the nanochannel. Given time (usually minutes), the end will unfold;
9
Fluorescence Microscopy of Nanochannel-Confined DNA
171
else, the molecule can be pushed out into the microchannels and reinjected into the nanochannels. 2. A small pressure offset when using pressure-driven flow can cause the molecule to not be in its equilibrium state while imaging. 3. Photonicking may cause the DNA to be cut into smaller pieces while imaging. However, the ordering of the fragments will not change, since two pieces cannot diffuse past each other while confined in a nanochannel. 4. During long imaging periods, the molecules will most likely fade significantly in fluorescence intensity due to photobleaching, especially in the absence of BME. 5. The DNA present close to the nanochannels can suffer from some degree of photobleaching and photonicking during imaging of the DNA in nanochannels. Therefore, an important consideration is to make sure that the illuminated area does not extend beyond the region of interest, if necessary using the field aperture. 3.4. Data Analysis
To extract essential parameters from the movies (or rather stacks of images) of DNA molecules confined in a nanofluidic structure, a simple pattern recognition and fitting script can be used (1). The key steps of the analysis are listed and explained below. 1. The position of the DNA molecule is detected. A region of interest (ROI) is created around the molecule and the rest of the image is discarded to reduce the amount of data to store (Fig. 3a) (19). 2. The pixels are summed over the width of the extended DNA yielding a one-dimensional intensity profile of the molecule. 3. Step 2 is repeated for each frame in the movie, except that the molecule is identified based on its position in the previous frame. Stacking these intensity profiles next to each other yields a timetrace (also known as a kymograph) (Fig. 3b). In this way, a whole movie can be reduced to one single composite image. 4. The intensity profiles are fitted to a model profile, described below, by a least-square algorithm (Fig. 3c). This fitting provides the center position, intensity (with subtracted background) and length of the DNA for each frame in the original movie. The model intensity profile, I(x), consists of a convolution of a modified box function (height I0, length Lx) with a Gaussian point-spread function (PSF), with a full width at half maximum l (FWHM) of 2:35 s0 ¼ 0:61 NA (l is the wavelength of the light
172
F. Persson et al.
Fig. 3. (a) The first fluorescence image in an image stack. A box is drawn enclosing the extended DNA molecule. The scale bar corresponds to 20 mm. (b) Time trace obtained by averaging over the molecule shown in (a) in the direction transverse to the DNA extension for every frame. Each column of pixels corresponds to the averaged intensity profile of one frame. The scale bars correspond to 20 mm and 10 s, respectively. (c) The intensity profile I (x ) and the corresponding fit for one column of the time trace. (d) A histogram over all the fitted lengths from one movie containing 400 frames. The data is well described by a Gaussian distribution (solid line). Reproduced with permission from ref. 19.
and NA is the numerical aperture of the objective) (see Note 21) (1). The model is represented by the following equation: I0 I ðxÞ ¼ Ibg þ 2 x x0 x ðx0 þ L0 Þ pffiffiffi ð1 þ BÞ Erf pffiffiffi ð1 BÞ Erf ; s0 2 s0 2 where Ibg is the background intensity value, B is a numerical factor introduced to allow for a nonconstant background, Erf is the error function, and x0 is the center position of the box function (see Note 22). It is important to realize that the biologically relevant gauge of resolution in these experiments is base pairs and not nanometers
9
Fluorescence Microscopy of Nanochannel-Confined DNA
173
or pixels. Standard fluorescence microscopy has a resolution limited by diffraction to roughly half the wavelength of the detected light. The resolution in base pairs is determined by the degree of DNA stretching, the DNA fluctuations and the total photon budget. Thus, maximum resolution is obtained in channels with the smallest possible cross section yielding fully stretched DNA with a minimum of thermal fluctuations. In practice, there is an optimum resolution for each experiment and it is important to design the experiments accordingly. A relatively low resolution is appropriate when targeting the overall structure of chromosomes and can be done in a fast way without extensive efforts. On the contrary, studies of a selected region of interest along the genome may require the highest possible resolving power, requiring extensive efforts in terms of data acquisition, storage, and handling. In some applications, the resolving power per se is not relevant; instead, the positioning of one or more specific labels contains the important information. In this case, the measurement uncerpffiffiffiffiffi tainty scales roughly as 1= N where N is the number of detected photons, giving accuracies approaching a single nanometer (29).
4. Notes 1. One common fabrication method uses nanoimprint lithography (NIL). This has the benefit that it is possible to order finished master stamps commercially (available from NIL Technology, Denmark), thus eliminating the need for an electronbeam lithography system. A common mass production technique, capable of defining nanostructures, is injection molding. With suitable choice of low-fluorescence polymer matrix, it may prove useful for large series of devices. Although focused ion beam (FIB) milling is a slow linear technique, it may find use for creating complicated three-dimensional structures with resolution comparable to that of electron-beam lithography. Direct laser writing systems (available from Nanoscribe GmbH, Germany) are now also capable of creating complex three-dimensional structures with feature sizes below 100 nm. A multitude of more exotic alternative fabrication techniques are described in the literature. 2. 0.1 M dithiothreitol (DTT) can replace BME as reducing agent. 3. Performance Optimized Polymer 6 (POP6) from Applied Biosystems can be used as an alternative to PVP. 4. Whenever water is mentioned in the context of buffer composition, we refer to ultrapure water with resistivity 18.2 MO cm
174
F. Persson et al.
(at 25 C) (referred to as Milli-Q water when using water purification equipment from the Millipore Corporation). 5. DNA in stock solutions at concentrations of 100–500 mg/mL is very viscous and hard to pipette accurately. Tip the tube sideways and suck in the solution very slowly to ensure that the correct amount of DNA is withdrawn. (For l-DNA from New England Biolabs the stock solution is 500 mg/mL, so 25 mL of the solution is added to 225 mL of 0.5 TBE in a 1.5-mL microcentrifuge tube.) When working with lambda phage DNA, it may be advisable to heat it to 50 C for 10 min in a microcentrifuge tube heater and quench in icy water to avoid concatamers due to the hybridization of the single-strand overhangs. 6. BME, a highly toxic chemical that serves as a biological antioxidant by scavenging oxygen and hydroxyl radicals in the buffer, thereby preventing photobleaching and photoinduced damage (photonicking) of the DNA. An enzymatic oxygen scavenger system may constitute a useful alternative when reducing agents cannot be used. It consists of 0.2 mg/mL glucose oxidase, 0.04 mg/mL catalase and 4 mg/mL b-D-glucose (available from Sigma-Aldrich). The oxygen scavenger system can be combined with BME but typically does not provide any additional benefit for the experiments listed. 7. One useful option in a fluorescence microscope is a unit that sends a selected field of view through two different optical paths and projects the resulting images on two separate areas of the CCD. This allows the user to acquire two (or more) colors or two polarization directions simultaneously. Existing systems include DV2™ from Photometrics and OptoSplit™ from Cairn Research. To improve imaging resolution, a wide range of novel techniques have been developed, each one capable of reaching a resolution of below 100 nm (30). They essentially fall into three categories: local quenching of the fluorescence emission (STED), repeated photoactivation and subsequent imaging of a subset of the fluorophores in the sample (STORM, FPALM), and structured illumination (SIM). Most imaging is carried out with B/W cameras giving information on the intensity in each pixel. For additional contrast information, multicolor (spectroscopic) or fluorescence life-time imaging (FLIM) may be utilized. 8. In order to expose the DNA to a gradual change in confinement in one single chip, it is possible to use funnel-like channels (6). This is an analogue to force spectroscopy techniques using optical tweezers that allows for probing of low DNA
9
Fluorescence Microscopy of Nanochannel-Confined DNA
175
Fig. 4. (Left ) Chip design with nanochannels spanning between two U-shaped microchannels. This design enables fast buffer exchanges since all of the liquid does not have to pass through the nanochannels. The size of each chip as seen in the lower image is 1 1 in. In order to visualize the nanochannel region, the central region is exaggerated. (Right ) Schematics of a similar design as the left one with an added nanoslit oriented perpendicular to the nanochannel array. In the nanoslit, the nanochannels become nanogrooves in the bottom, working as entropic traps for the DNA. This design also allows for enrichment of DNA in the nanogrooves by applying positive pressures at both ends of the nanochannels. In order to visualize the nanochannel region, the central region is exaggerated. The nanoslit and nanochannels are 50 mm and 100 nm wide respectively. Reproduced with permission of The Royal Society of Chemistry from ref. 36.
extensions, corresponding to forces in the femtonewton regime, without anchoring the molecules. Another feature that can be added to the chip is demonstrated in Fig. 4 (right). In this chip design, a nanoslit is etched orthogonal to the nanochannels. This allows for enrichment of DNA in the nanogrooves by applying moderate positive pressures at both ends of the nanochannels. If the slit is sufficiently shallow, entropy keeps the DNA in the grooves while buffer flows through the slit. This design also allows for changing the chemical environment of the DNA, by flushing the desired solution in the slit, while monitoring the DNA in real time. 9. When analyzing long, genomic DNA extra care has to be taken due to the relatively large size of the molecules. Considering human DNA, it can be noted that the largest single DNA molecule has a fully extended length of over 8 cm (chromosome 1). If the DNA molecule is too long for the whole molecule to be easily extended in a single nanochannel, one possibility is to stretch it by shear flow (31) in a device made by conventional photolithography. With this approach, there is no need for nanofabrication. However, it is a dynamic
176
F. Persson et al.
system in which the DNA conformation will not be at equilibrium as compared to DNA confined in nanochannels. 10. The layout of a DNA analysis chip may also contain features for sample preparation such as cell-sorting, cell lysis, DNA extraction and purification, and DNA staining in addition to the actual analysis of the DNA molecules. This kind of integration would enable the analysis of DNA from a single cell, which would be especially useful in the study of e.g., rare circulating tumor cells. A potentially very useful development with this regard is the use of deterministic lateral displacement to continuously guide cells and chromosomes through different chemical environments for cell lysis and DNA staining (32). 11. Even without an e-beam writer, slit-like channels (depth in the nanometer range and widths larger than 0.5 mm) can readily be defined with UV lithography and carefully tuned RIE etching. 12. Alignment marks can alternatively be formed by anisotropic RIE etching, and in the case of silicon also through anisotropic wet etching using, for example, KOH. If etching is used to define the alignment marks, it is important that they provide a sufficient contrast for the alignment in the mask aligner. An etch depth of at least 200 nm is recommended. For metal alignment marks it is also possible to first deposit a layer of metal and subsequently spin on and pattern a photoresist and in a last step etch away the exposed metal. Al is commonly etched using either a wet etch using phosphoric acid or a dry etch containing chlorine chemistry. Au is commonly etched by using wet etches of either potassium iodine or aqua regia (1:3 HNO3–HCl). 13. Instead of using chemicals, the resist can be stripped by an oxygen plasma treatment. However, this is not recommended since it can burn the resist, making it very hard to remove, and also induces roughness on the sample surface. 14. Examples include micromilling, deep reactive ion etching (DRIE) or ultrasonic drilling. However, these techniques often demand some specialized equipment, which is very expensive compared to that needed for powder blasting. 15. Instead of using a soft film to mask the wafers/chips during powder blasting, a metal mask defined in a thick brass plate can be used. The chip is then attached to the metal mask using reversible thermal glue. It should be noted that since the metal mask is hard, it will also be degraded by the powder blasting, which attacks hard surfaces. 16. Polymer-based devices are generally sealed using polymer fusion bonding. The device is bonded to a lid with a polymer film by heating until the polymer layers on the chip and lid intermix. The combination of polymer compositions and
9
Fluorescence Microscopy of Nanochannel-Confined DNA
177
temperatures must be carefully chosen to create a sufficiently strong bond while maintaining the structural integrity of the micro- and nanochannels. Anodic bonding is the standard technique to bond borosilicate glass to silicon, also for silicon with a hydrophilic oxide layer, but it might cause wide nanochannels (nanoslits) to collapse. 17. The Piranha-based protocol can be used to bond silicon with a thin layer of oxide ( hI(t)ihI(t)i and therefore
10
Fluorescence Correlation Spectroscopy
183
G(t) > 1. On the contrary, if t is much larger than the typical residence time, then signals I(t) and I(t + t) have no similarity and hI(t) I(t + t)i can be written as hI(t)ihI(t + t)i, thus G(t) ¼ 1. It is intuitive at this stage that the intermediate values of the ACF will carry the information on the duration of the fluctuation, here the residence time of the molecule in the observation volume. A typical ACF is plotted in Fig. 1c. Although sensitive to single molecule fluctuation, FCS is intrinsically a statistic method, thanks to the huge number of individual events that are analyzed. Although the above explanation is hand waving and limited to the simple case of 2D diffusion, for tutorial purposes, FCS is supported by a rigorous theory as described below. It permits to quantify a wide range of phenomena such as photophysical, photochemical, interaction, diffusion, and transport properties of molecules (3). The scope of this chapter is very practical, “what shall I do to build and run my own FCS setup” has been our guideline when writing this chapter. 1.1. Theoretical Assumptions
The most important assumption behind FCS measurements deals with the concept of stationarity. The system under study is assumed to be stationary, i.e., from a mathematical point of view, both quantities hI(t)i and hI(t) I(t + t)i of Eq. 1 need to be independent of time t. In other words, the total measurement duration Ttot has to be much larger than the fluctuation duration (4). A practical figure of merit of the stationarity is the value of G(t), which should be unity for the largest values of t. As we will see later, a signal that is not stationary is not be suitable for a FCS analysis. The second assumption deals with the shape of the observation volume, which is described by the point spread function (PSF) of the confocal microscope. In the case of a high numerical aperture microscope objective, the PSF is in principle a complicated distribution that depends both on the excitation and collection optics. However, for FCS studies, it is generally assumed that it can be reasonably described by a 3D gaussian distribution (5) " !# x 2 þ y 2 z2 (2) PSFðx; y; zÞ ¼ I 0 exp 2 þ 2 ; o2xy oz where wxy and wz denote the transverse and axial waist, respectively. With this assumption, an effective observation volume is usually defined as V eff ¼ p3=2 w 2xy w z : Although this Gaussian assumption does not constitute a sine qua non condition for performing FCS, it allows us to derive relatively simple analytical expressions for the ACF G(t), as described in Subheading 1.5.
184
P. Ferrand et al.
1.2. Amplitude of the ACF
By writing the fluctuating intensity as I ðtÞ ¼ hI i þ dI ðtÞ, Eq. 1 becomes GðtÞ ¼ 1 þ
hdI ðtÞdI ðt þ tÞi hI i2
:
In case of diffusing molecules, the fluctuations of intensity recorded by an ideal setup are only due to the fluctuation number of molecule in the observation volume. We shall name dN as fluctuation number. For a given count rate per molecule (CRM), hI i ¼ hN i CRM and dI(t) can be simply written as dI(t) ¼ dN(t) CRM, and then, for short delays, the amplitude of the ACF is Gðt ! 0Þ ¼ 1 þ
hdN 2 i 2
hN i
¼1þ
1 ; hN i
(3)
where we have used the property that the quantity N is governed by a Poisson statistics, and as such, hdN2i ¼ hN i. The important result of Eq. 3 emphasizes that FCS is intrinsically designed to analyze small number of molecules. 1.3. Relevant Time Scales
Although FCS can address a large range of time scales, a priori knowledge of the fluctuation time, that we denote tf, is necessary to set the optimum conditions for measurement, that we summarize here as Dt tf T tot ;
(4)
where Dt is the correlator sampling time with which the intensity I(t) is recorded. Indeed, in order to be sensitive to the fluctuations, Dt needs to be much smaller than tf. Note that Dt will be the first channel of the computed ACF, usually named the correlogram. In addition, as discussed in Subheading 1.1, the condition of stationarity requires that the total measurement duration Ttot is much larger than correlogram. 1.4. Signal-to-Noise Ratio in FCS
Noise and statistical accuracy in FCS deserve specific consideration. First, understanding the origin of the signal-to-noise ratio (SNR) enables optimization of the experimental apparatus. Second, proper weighting procedures based on statistical estimates for the noise enable more accurate analysis. Beyond technical noise, two physical phenomena bring fundamental contributions to the noise in FCS: the quantum nature of light that induces shot noise and the stochastic nature of the fluorescence fluctuation process itself. In the most common experimental conditions for FCS: Gaussian molecular detection efficiency, 3D Brownian diffusion, negligible background, and small sample time (Dt ! 0), but without any assumption on the average number of molecules hN i,
10
Fluorescence Correlation Spectroscopy
185
it can be demonstrated that the SNR of the ACF is simply given by (6) pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi CRM T tot Dt SNR t!0 ’ ; (5) ð1 þ 1=hN iÞ1=2 where CRM denotes as previously the CRM. This shows that the SNR in FCS does not depend on the total detected fluorescence, but on the fluorescence CRM, on the square root of the experiment duration Ttot, and on the correlator channel minimum width Dt. This further emphasizes the single molecule nature of FCS and provides guidelines to improve the statistical accuracy in FCS. For a fixed experimental apparatus, one can either increase the excitation power to raise the CRM or wait for longer integration times Ttot. However, both of these strategies have significant practical limitations. First, saturation and photobleaching limit the CRM increase to a certain extent. Second, due to the square root dependence of the SNR on the acquisition time, increasing Ttot to a few hundred seconds has only a minor influence on the SNR. 1.5. Analytical Expression of the ACF
With the assumptions described in Subheading 1.1, analytical expressions of the ACF can be derived for a large variety of physical processes. For the sake of generality, we shall write the ACF as GðtÞ ¼ 1 þ
1 X ðtÞ G D ðtÞ BðtÞ; hN i
(6)
where X(t), GD(t), and B(t) denote the intra- or intermolecular reactions, diffusion, and background contributions, respectively, to the ACF. Note that this separation of different dynamical contribution is only possible when molecular reactions takes place at much shorter time scale than diffusion, which is usually a reasonable assumption. 1.5.1. Inter- and Intramolecular Dynamics
The first factor in Eq. 6 accounts for changes in fluorescence yield as molecules undergo inter- or intramolecular dynamics, such as triplet blinking. The general expression for X(t) is T t ; (7) exp X ðtÞ ¼ 1 þ 1T tT where T is the fraction of molecule in a dark state and tT is the corresponding duration of the triplet fluorescence fluctuation (7).
1.5.2. Translational Diffusion
The factor GD(t) is derived from the diffusion equation that governs the evolution of the concentration C (r, t) of the fluorescent molecule, given by @Cðr; tÞ ¼ DDCðr; tÞ; @t
186
P. Ferrand et al.
where D is the diffusion coefficient of the molecule. For a 3D translational diffusion, G 3D D ðtÞ ¼
1 1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 þ ðt=tD Þ 1 þ s 2 ðt=tD Þ
(8)
with tD ¼ w 2xy =4D is the average residence time in the observation volume and s ¼ wxy =w z is a purely geometrical factor. In case of 2D diffusion, the limit s ! 0 gives G 2D D ðtÞ ¼
1.5.3. Background Signal
1 1 þ ðt=tD Þ
(9)
Last but not least, the factor B(t) accounts for the contribution of background signal in the ACF. It has been demonstrated that (6) hBi 2 ; (10) BðtÞ ¼ 1 hI i where hB i is practically the intensity measured by replacing the fluorescent sample by a nonfluorescent one.
2. Materials 2.1. FCS System
Several companies offer turn-key FCS systems. Most of them are based on confocal laser scanning microscope (CLSM) systems (see Note 1), while some others rely on a more simple optical systems, require no optical alignment but suit therefore to more diluted systems. These systems include excitation laser, appropriate filter sets, photon counting systems, and a dedicated data analysis software. Alternatively, most commercial CLSMs can be customized in order to perform FCS measurements, according to the optical scheme presented in Fig. 2. It requires, however, a solid background in photonics instrumentation, electronics, data acquisition, and programming. The following points need to be considered 1. It is important to ensure that the scanning system can be turned off in one way or an other, in order to perform static measurements. Sometimes, this can be done by setting the scanning range to 0. 2. A system with various available pinhole sizes will be useful in the case of signal concerns (the larger the pinhole, the stronger the signal). 3. A high numerical aperture objective (typically NA ¼ 1.2) is preferable, because it ensures the largest collection efficiency
10
Fluorescence Correlation Spectroscopy
187
Fig. 2. A basic FCS setup as well as a typical correlogram as computed by the correlator. The inset shows a schematic view of the 3D PSF of the system. APD stands for avalanche photodiode.
and therefore the largest CRM, in addition to the highest spatial resolution. 4. Laser excitation power will need to be measured in order to provide controlled and producible operating conditions. Photodiode power meters are usually the best choice for low power measurement, i.e., below a few milliwatts. 5. Because CLSM detection usually relies on photomultiplier tubes in analog mode, the detector must be replaced by one (or two, see Note 2) photon counting module(s) such as avalanche photodiodes (APDs) (e.g., SPCM-AQR series by PerkinElmer Optoelectronics, PDM series by Micro-PhotonDevices, or id100 series by id Quantique). The key parameter to be considered here is quantum efficiency, because this directly affects the CRM. 6. The correlation function can be built up either by using a digital hardware correlator (by correlator.com or ALVGmbH) or by software correlation. Although it means an additional cost, the first approach is the most easy way to implement and also the most secure concerning numerical error and/or artifacts. In this case, the signal form the detectors is correlated on-line, and the ACF can be monitored in real time. Such correlator can work either in linear mode
188
P. Ferrand et al.
or in multitau mode, the later offering a wider range of accessible fluctuation times, thanks to channels of increasing width in a quasi-logarithmic progression. Alternatively, correlation can be performed afterwards on the intensity series, usually in multitau or equivalent computation scheme. The major drawback of this approach is the huge volume of data that need to be stored and the often prohibitive calculation time. For this reason, it is only relevant for slow processes (fluctuation time larger than 1 ms). Even if the correlation is built up on-line, we recommend to always record (with a very low time sampling, typically 100 ms), the time series, for further analysis of stationarity. 7. Last but no least, FCS data analysis strongly relies on advanced fitting procedures. The work flow will be much more effective by using data analysis software offering powerful fitting algorithms, such as Levenberg-Marquardt, as well as a complete macroprogramming language, such as Origin, Igor Pro, Matlab, etc. 2.2. Labeling
When choosing the fluorescent label to be used on the biological system you want to investigate, remember that the golden rule is “to maximize the CRM.” Therefore, 1. Select robust dyes, labels, fluorescent proteins with the highest quantum efficiency. 2. Check carefully the filter set in order to match the spectral features of the label. 3. Always prefer labels that have been extensively documented in the literature. Even if they are not the subject of your investigations, photophysical properties, such as conformal changes, triplet blinking, may be visible on the ACF, and an a priori knowledge will by a real plus when fitting the data. Diffusion coefficient of the molecule in solution is required if the label is used to calibrate the measurement volume of your FCS system.
2.3. Calibration Solution
The calibration solution will be used in order to determine the measurement volume of the system. 1. Choose a solution of a robust dye those fluorescence can be measured with the same laser/filter configuration as the one that labels your biological system. 2. Choose a dye those diffusion coefficient Dcal is documented (8). 3. Feel free to use the same label as for your biological system, if you think it is stable and bright enough.
10
Fluorescence Correlation Spectroscopy
189
4. First prepare a 1-mL aliquot of high concentration in ultrapure water, typically 1 mM. This will constitute the stock solution for future dilutions. Keep it stored at 18 ∘ C. 5. Prepare from the stock solution four successive dilutions in 1-mL aliquots with typical concentrations 10 mM, 1 mM, 100 nM, 10 nM. Store at + 4 ∘ C in the dark.
3. Methods 3.1. Coarse Optical Alignment
Note that although CLSM can produce relative clear images even in case of a moderate misalignment, FCS is optically much more demanding and requires l
A perfect centering of the excitation laser on the back aperture of the microscope objective.
l
A perfect conjugation of the pinhole with the excitation spot.
In case of a commercial FCS system or of a FCS-upgraded CLSM, some of the alignment settings may not be accessible to the user, and therefore require maintenance by the company. Note that this procedure needs to be followed only if the system has not been used for several weeks or if an important modification has been made on the optical system (laser replacement, filter change, etc.). 1. Switch the system on, with the correct excitation sources, microscope objective and filter set. 2. Select the largest pinhole diameter. 3. Stop the beam with an appropriate beam stopper. Set laser power to a moderate level. This should be a few hundreds mW for a one-photon fluorescence process and a few tens of milliwatts for a two-photon fluorescence process. 4. Put a droplet of the appropriate immersion liquid on the objective. 5. Place a standard coverslip on the sample stage. 6. Put about 50 mL of the most concentrated calibration solution (10 mM). 7. Release the laser beam stopper. 8. Move the focus until the laser spot is located inside the solution. You should see the fluorescence spot when looking at the sample, your eyes been protected from laser radiation by goggles. Alternatively, remove the sample holder, and place the coverslip directly on the objective, the capillary effect of the immersion should maintain it horizontally. This ensures that the observation volume is into the solution.
190
P. Ferrand et al.
9. Switch on the detector. Run a static intensity measurement. Reduce laser excitation power if the detector is saturating. 10. Optimize the signal by adjusting the pinhole alignment. Reduce laser excitation power if the detector is saturating. 11. Repeat optimization with the next dilution (1 mM). 12. Repeat optimization with a smaller pinhole diameter. 3.2. Calibration of the Measurement Volume
This procedure has to be followed before any series of measurements, in order to prevent any slight misalignment. In addition, it allows to determine the values of wxy and wz for the system. steps 12–13 allow us to set the best measurement conditions for a given solution on the system. It does not need to be repeated daily. 1. Switch the system on, with the appropriate excitation sources, microscope objective and filter set. 2. Select the largest pinhole diameter. 3. Set laser power to a moderate level. Stop the beam with an appropriate beam stopper. 4. Put a droplet of the appropriate immersion liquid on the objective. 5. Place a standard coverslip on the sample stage. 6. Put about 50 mL of the 100-nM calibration solution. 7. Release the laser beam stopper. 8. Move the focus until the laser spot is located inside the solution. 9. Record a ACF for Ttot ¼ 30 s, and measure simultaneously the average value of the fluorescence intensity hI i. If ACF is too noisy, use a larger measurement duration. 10. Fit the obtained ACF with analytical expression of Eq. 6, in order to obtain values hN i, tD, as well as T and tT, if necessary. Usually, s can be fixed to s ¼ 0. 2 for the first fitting iterations. 11. From hI i and hN i, calculate the count rate per molecule CRM ¼ hIi=hNi. 12. Repeat steps 9–11 for various sizes of pinhole (see Notes 3 and 4) and various excitation powers Pexc. 13. Plot variations of CRM and tD versus Pexc. The range of excitation power for which CRM has a linear dependence upon Pexc and tD does not depends on Pexc is the one which suits to this calibration solution on your system. In practice, use a power 20% below the maximum suitable excitation power.
x10–3
10
Fluorescence Correlation Spectroscopy
40 0 –40 55 x103
1.6 1.5
50 45 0
2
4
6
8
10
t (s)
1.4 G(tau)
191
1.3 1.2 1.1 1.0 10–6
10–5
10–4 tau (s)
10–3
10–2
Fig. 3. Example of ACF recorded in a calibration solution of Rhodamine 6G. Measurement duration was 10 s. Circle are experimental points, while solid line is the fit according to Eqs. 6–8. Inset is the intensity series, top graph is the residuals. Fitting values: hN i ¼ 1.93, T ¼ 0. 133, tT ¼ 4.1 ms, tD ¼ 29 ms, and s ¼ 0.2.
14. The lateral width of the measurement volume is given by pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi wxy ¼ 4D ref tD , where tD is measured in the correct range of excitation power. An example of FCS performed in a 30 nM solution of Rhodamine 6G in water (9) under a typical excitation power of 200 mW is plotted in Fig. 3, as well as the corresponding fit, according to Eqs. 6–8. Note that the background signal was negligible and therefore B(t) was assumed to be unity. Given a value Dref ¼ 280 mm2/s, measurement volume here is characterized by wxy ¼ 179 nm and wxy ¼ 895 nm. 3.3. FCS Measurements
1. Place your sample on the microscope. 2. Set the excitation power to a moderate level. 3. In the case of a cell or other specific sample, use the LSCM capabilities of the system in order to set the correct focus and set the FCS measurement area (by moving the sample or by moving the measurement spot, depending on the system in use). 4. Set the excitation power to a low level, compatible with static recording of intensity (see Note 5).
192
P. Ferrand et al.
5. Record a set of short correlogram as well as the corresponding intensity series. 6. Repeat the measurement on equivalent locations, on another part of the sample. 3.4. FCS Data Analysis
1. Browse all correlograms together with the corresponding intensity series that have been recorded at one location. 2. Discard all correlograms whose intensity series are not stationary (see Note 6). This often happens in complex samples such as cells because of membrane fluctuations, dye aggregation, etc. 3. Average the remaining correlograms. 4. Fit the obtained correlogram with the most appropriate model. If the obtained physical parameters do not make physical sense, reject the fit. 5. Repeat this procedure for all sets of correlograms.
4. Notes 1. FCS is not limited to one-photon absorption fluorescence, but has been successfully extended to nonlinear optical contrasts, such as two-photon absorption fluorescence, etc. From an instrumental point of view, because nonlinear microscopy systems are free of pinhole, detection relies usually on large area photomultiplier tubes. It is therefore subject to parasitic light, in addition to a lower quantum efficiency. In practice, two-photon FCS is subject to strong photobleaching and requires significantly higher incident power (tens of milliwatts) than one-photon FCS (hundreds of microwatts). Nevertheless it has several advantages like being able to excite several fluorophores with the same wavelength; it does not require pinhole, it has a potential deeper penetration depth in scattering samples. 2. Detectors can be subject to measurement artifacts (dead time, after pulse, etc.) that compromise the ACF for short delays (typically for t < 1 ms). In case values at short delays are necessary for the process under study (photophysics, for instance), this random noise can be efficiently rejecting by splitting the signal on two detectors and by cross-correlating the signals obtained by the two detectors. The system illustrated in Fig. 2 is set up in this configuration. 3. Pinhole diameter is directly related to the axial extent of the PSF. Although this feature is crucial for CLSM imaging, because it determines the axial sectioning capability of the
10
Fluorescence Correlation Spectroscopy
193
system, it can often be reasonably sacrificed while doing FCS measurement, a larger pinhole diameter providing a significantly larger CRM, and therefore a larger SNR. Indeed, enlarging the pinhole diameter produces only limited modifications of the diffusion time tD. This is because wxy (transverse width of observation volume) is defined by the excitation beam rather than the collection optics (that defined wz). 4. When repeating measurements with new samples, you may experience a sudden degradation of CRM and/or diffusion time, these two figures being of course much more clearly visible when measuring solutions. This is often due to a defect of the coverslip that is sometimes outside the nominal range indicated by the manufacturer. In this case, it is usually more effective to discard the coverslip and to start a new measurement than to try to adjust the correction collar of the microscope objective. Another common problem is the drying of water between the coverslip and a water-objective lens. 5. Although FCS can address fluctuation processes at all timescales (within the limitations mentioned in Subheading 1.3, of course), slow diffusion cannot be measured in practice, because slow species are usually bleached before leaving the measurement volume. This produces a typical drop of the recorded intensity, as illustrated in Fig. 4d. In this case, scanning FCS approaches such as image correlation spectroscopy and its variants may be more relevant (see Note 7). 6. Figure 4 illustrates some examples of odd-looking correlograms that constitute the usual first taste of FCS for new users and that may be quite discouraging. Possible explanations and suggestions are given for each case. Measurements like cases C and D must be definitely discarded, because they do not satisfy to the condition of stationarity required by the FCS analysis. Although these correlograms may in some cases be successfully fitted, they would produce erroneous values. More generally, these examples emphasize that FCS analysis must be operated by experienced users only, especially for measurements carried out in complex systems such as cells, where the diffusion is mostly non-Brownian and the signal subjected to other spurious fluctuations. 7. Latest development of FCS focus both on hardware and software. From the hardware side the current trends are (a) dual-spot measurement to measure FCS simultaneously at various locations (9) and/or to access to absolute diffusion coefficient (8) and (b) multispot FCS using EMCCD cameras as detector (). From the data analysis side, recent development focus on the improvement of spatio-temporal information by means of scanning FCS (11) and image correlation
194
P. Ferrand et al.
a Problem - No visible correlation
1.4
I(t)
1.5
G(tau)
1.3 0.0
1.2
0.1
0.2
0.3 0.4 t (s)
0.5
0.6
1.1
Diagnosis - Concentration is too high - Sample is out of focus - Sample is static To do - Use a sample with less fluorophores - Check focus
1.0 0.9 0.8 10–6
10–2
10–1 Problem - Low SNR
1.004 1.003
G(tau)
10–4 10–3 tau (s)
I(t)
b
10–5
0
5
1.002
10 t (s)
15
20
To do - Improve CRM (increase excitation power, use a larger pinhole, increase excitation power) - Change fluorophore for a brighter one - Change the microscope objective for one with a higher NA
1.001 1.000 10–6 10–5 10–4 10–3 10–2 10–1 tau(s) 1.008 1.006 G(tau)
100
101 Problem - Intensity is not stationary - Shape of correlogram changed suddently
I(t)
c
0
1
2
3
4
5
t (s)
1.004
1.000 10–5
10–4
d
G(tau)
1.03
10–3 10–2 tau (s)
10–1
100 Problem - Intensity is dropping continuously - Correlogram "flies away"
I(t)
1.04
Diagnosis - sample drift - intercept of a large aggregate To do - Try to purify your sample - Replace long duration measurement by several short duration measurement, discard odd correlograms and average
1.002
10–6
Diagnosis - CRM is too low - Measurement duration is too low
0
2
4
6
8
t (s) 1.02 1.01 1.00 10–6 10–5 10–4 10–3 10–2 10–1 tau (s)
100
10
Diagnosis - photobleaching of the static or slow fraction of the sample
To do - Use a lower excitation power - If you are only interested in the fast moving fraction, also consider the possibility to i) bleach out all the slow fraction, ii) run your FCS measurement on the remaining fast fraction. 101
Fig. 4. A nonexhaustive panel of unsuitable correlograms, with the corresponding intensity series plotted in the inset. For each case, the problem is detailed, diagnosis and possible solutions are given.
10
Fluorescence Correlation Spectroscopy
195
spectroscopy techniques (12) that give access to both temporal and spatial correlations. Another powerful variant of FCS is fluorescence cross- correlation spectroscopy (FCCS), which extends the capabilities of standard FCS by introducing two different fluorescent probes with distinct excitation and/or emission properties, which can be detected in the same confocal volume. It allows to probe interactions between different molecular species (13).
References 1. Magde, D., Elson, E. L., and Webb, W. W. (1972) Thermodynamic fluctuations in a reacting system – measurement by fluorescence correlation spectroscopy. Phys. Rev. Lett.29, 705 2. Berne, B. J., and Pecora, R. Dynamic light scattering: with applications to chemistry, biology, and physics; Dover, 2000 3. Webb, W. W. (2001) Fluorescence correlation spectroscopy: inception, biophysical experimentations, and prospectus. Appl. Opt.40, 3969 4. Tcherniak, A., Reznik, C., Link, S., and Landes, C. F. (2008) Fluorescence correlation spectroscopy: criteria for analysis in complex systems. Anal. Chem.81, 746 5. Rigler, R., Mets, U., Widengren, J., and Kask, P. (1993) Fluorescence correlation spectroscopy with high count rate and low-background – analysis of translational diffusion. Eur. Biophys. J.22, 169 6. Wenger, J., Ge´rard, D., Aouani, H., Rigneault, H., Lowder, B., Blair, S., Devaux, E., and Ebbessen, T. W. (2009) Nanoaperture-enhanced signal-to-noise ratio in fluorescence correlation spectroscopy. Anal. Chem.81, 834–9 7. Widengren, J., Mets, U., and Rigler, R. (1995) Fluorescence correlation spectroscopy of triplet-states in solution – a Theoretical and experimental-study. J. Phys. Chem.99, 13368
8. M€ uller, C. B., Loman, A., Pacheco, V., Koberling, F., Willbold, D., Richtering, W., and Enderlein, J. (2008) Precise measurement of diffusion by multi-color dual-focus fluorescence correlation spectroscopy. Europhys. Lett.83, 46001 9. Ferrand, P., Pianta, M., Kress, A., Aillaud, A., Rigneault, H., and Marguet, D. (2009) A versatile dual spot laser scanning confocal microscopy system for advanced fluorescence correlation spectroscopy analysis in living cell. Rev. Sci. Instrum.80, 083702 10. Burkhardt, M., and Schwille, P. (2006) Electron multiplying CCD based detection for spatially resolved fluorescence correlation spectroscopy. Opt. Express14, 5013 11. Petra´sˇek, Z., Hoege, C., Mashaghi, A., Ohrt, T., Hyman, A., and Schwille, P. (2008) Characterization of protein dynamics in asymmetric cell division by scanning fluorescence correlation spectroscopy. Biophys. J.95, 5476 12. Kolin, D. L., and Wiseman, P. W. (2007) Advances in image correlation spectroscopy: Measuring number densities, aggregation states, and dynamics of fluorescently labeled macromolecules in cells. Cell Biochem. Biophys.49, 141 13. Bacia, K., Kim, S. A., and Schwille, P. (2006) Fluorescence cross-correlation spectroscopy in living cells. Nature Meth.3, 83
.
Chapter 11 Introduction to Atomic Force Microscopy Pedro J. de Pablo Abstract Atomic force microscopy (AFM) is an invaluable tool not only to obtain high-resolution topographical images, but also to determine certain physical properties of specimens, such as their mechanical properties and composition. In addition to the wide range of applications, from materials science to biology, this technique can be operated in a number of environments as long as the specimen is attached to a surface, including ambient air, ultra high vacuum (UHV), and most importantly for biology, in liquids. The versatility of this technique is also reflected by the wide range of sizes of the sample that can dealt with, such as atoms, molecules, molecular aggregates, and cells. Indeed, this technique enables biological problems to be tackled from the single-molecule point of view and it allows not only to see but also to touch the material under study (i.e., mechanical manipulation at the nanoscale), a fundamental source of information for its characterization. In particular, the study of the mechanical properties at the nanoscale of biomolecular aggregates constitute an important source of data to elaborate mechano-chemical structure/function models of single-particle biomachines, expanding and complementing the information obtained from bulk experiments. Key words: Scanning, Surface, Cantilever, Liquids, Force curve
1. Introduction The first thing that comes to mind on hearing the word “microscopy” is an optical device that manipulates light to obtain a magnified image of a sample. As a consequence, the first question that is usually posed when seeing an atomic force microscopy (AFM) for the first time is: where do I have to look to see the specimen? A microscope is generally considered a machine in which a source emits particles such as photons or electrons that are used as probes thrown onto the specimen to provoke an interaction. This probe–specimen interaction can be registered by a detector and this result is communicated of to an analyzer, which processes the information received to make it comprehensible. There are two kinds of microscopes that fit readily into this Erwin J.G. Peterman and Gijs J.L. Wuite (eds.), Single Molecule Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 783, DOI 10.1007/978-1-61779-282-3_11, # Springer Science+Business Media, LLC 2011
197
198
P.J. de Pablo
source–specimen–detector–analyzer scheme. The first is the optical microscope, where the photons coming from the incandescent lamp are conveniently manipulated by a system of lenses and mirrors both before and after the interaction with the specimen arriving to the eyepiece where the detector (i.e., the eyes) collects the information. A typical optical microscope can reach a resolution of l/2 ~200 nm (l being the wavelength of the light). Anton van Leeuwenhoek (1632–1723) is credited with bringing the optical microscope to the attention of biologists, even though simple magnifying lenses were already being produced in the 1500s. The electron microscope (EM) is a little more complicated. In this case, the particles that act as probes are not photons but electrons produced by an incandescent wire. Here, electromagnetic lenses are used to manipulate and focus the electron beam in order to provoke the right interaction with the specimen. The electrons are then collected by a screen, which is conveniently monitored. The first prototype electron microscope was built in 1931 by the German engineers Ernst Ruska and Max Knoll and 2 years later, Ruska constructed an electron microscope that exceeded the resolution possible with an optical microscope, reaching about 1 nm. In scanning probe microscopy, a sharp tip of a few nanometers in diameter, which can be considered as a probe, approaches the surface of the sample. The first member of this family of microscopes was the scanning tunneling microscope (STM) invented by Binning and Rhorer (1), who received the Nobel Prize for Physics along with Ruska in 1986. This system is based upon a quantum effect (tunneling) that occurs when a sharp metallic tip is brought to a distance (z) of less than 1 nm from a conductive surface. This effect involves the flow of an electronic current (I) between pffiffiffiffi the surface and the tip according to the formula I / expð ’zÞ, where f is the work-function of the metallic surface (2). The strong dependence on the tip–surface distance can be used to obtain topographic and electronic maps of the sample by moving (i.e., scanning the tip on the surface) while keeping the tip–sample distance constant through a feedback algorithm. Although this tool provides true atomic resolution in UHV conditions, a prerequisite is that both the tip and sample should be conductive. Therefore, it follows that STM is not suitable for biological samples since these are mainly insulators, which need to be covered with a metallic layer (3). In 1986, Binnig et al. (4) invented the AFM combining the principles of both the STM and the so-called stylus profilometer (5). In an AFM, a sharp stylus (approximately tenths of a nanometer) attached to the end of a cantilever is approached to the surface. As a consequence, a force appears between the tip and surface that can be attractive or repulsive (see below) causing
11
Introduction to Atomic Force Microscopy
199
the cantilever to bend. When this bending is controlled with a feedback algorithm, it is possible to obtain a topographic map by scanning the surface in a plane perpendicular to the tip. In the original paper (4), the topographic profiles of a ceramic sample, an insulator, were shown. This is one of the main advantages of AFM: the fact that both tip and sample may be insulators, opening the door to a whole range of possibilities such as the study of biological samples (biomolecules like proteins, membranes, whole cells, etc.).
2. AFM Implementation: Technical Issues
Although there are a variety of ways to control the deflection of the cantilever in AFM we will focus on the beam deflection method (6) since it is commonly employed when working with biological samples. The beam deflection system involves focusing a laser beam on the end of the cantilever and collecting the reflected light with a photodiode. As a consequence, any bending of the cantilever will affect the position of the reflected laser spot on the photodiode. A normal bending will originate a so-called normal force signal Fn on the photodiode sectors, whereas a lateral torsion will result in the so-called lateral force F l. The core of an AFM is the head (Fig. 1a), where the beam
Fig. 1. Principle of an AFM. The software running on the computer (a) controls the electronics, which give and receive information to and from the AFM head (inside the oval ). Panels (b) and (c) show the AFM sensor: the cantilever. Panel (b) shows a rectangular cantilever attached to a chip and panel (c) presents a typical tip at the end of the evident pyramid.
200
P.J. de Pablo
deflection system is integrated along with the piezoelectric-tube that moves the sample in all three directions (x, y, and z). In Fig. 1a, an AFM head configuration is shown where the tip is fixed and the sample is moved by the piezotube. Another configuration known as “stand alone” fixes the sample and makes the scanning movement with the tip. The electronic components receive the signals coming from the photodiode, mainly Fn and Fl, and provide high voltages (~240 V) to move the piezotube to which the sample is attached. The computer is in charge for managing the data and calculating all the parameters required to move the piezotube in a convenient way. Integrated tip and cantilever assemblies can be fabricated from silicon or silicon nitride using photolithographic techniques. More than 1,000 tip and cantilever assemblies can be produced on a single silicon wafer. The cantilevers can be rectangular (see Fig. 1b) or V-shaped and they typically range from 60 to 200 mm in length, 10 to 40 mm in width, and 0.3 to 2 mm in thickness. The typical tip radius is about 30 nm, although sometimes lower diameters can be obtained (Fig. 1c). The cantilever spring constant k ranges between 0.03 and 40 N/m, and it strongly depends on the cantilever’s dimensions. For example, for a rectangular cantilever k ¼ ðEW =4ÞðT =LÞ3, where E is the Young Modulus of the cantilever, and W, T, and L are the cantilever width, thickness, and length, respectively (Fig. 1b). While W and L can be fairly precisely known, T is always difficult to measure. As a consequence, the manufacturers normally provide the cantilever spring constant with an error of 10–30% and therefore the user should calibrate each cantilever (7). The spring constant k is used to calculate the Hookean force applied on the cantilever as a function of bending Dz, i.e. F ¼ k Dz. The vertical resolution of the cantilever (Dz) strongly depends on the noise of the photodiode, since it defines the minimum significant displacement of the laser beam on the detector (see Fig. 1a). A typical cantilever of 100 mm length has about 0.1 A˚ resolution assuming a signal-to-noise ratio of around 1 (6). Another important parameter that can have influence on the vertical resolution is the thermal noise (8): the cantileverp oscillates ffiffiffiffiffiffiffiffiffiffiffiffiffiffi at the resonance frequency with an amplitude Dz ¼ kB T =k, where kB is the Boltzmann constant, T is the absolute temperature, and k the cantilever spring constant. For example, at room ˚ for a cantilever of spring temperature, there is a noise of about 5 A constant 0.02 N/m. 2.1. Interaction Between the AFM-Tip and the Surface
In order to understand the interaction between the tip and sample, we shall refer to potentials rather than forces. Physicists prefer this as potentials are scalar and therefore easier to deal with
11
Introduction to Atomic Force Microscopy
201
Fig. 2. Force curves, feedback, and lateral resolution. Panel (a) shows the force of the tip–sample interaction (represented in the inset ). The numbers inside the circles indicate the cantilever instabilities (see text). Panel (b) shows force versus z-piezo displacement curve in air and (c) in liquid. Panel (d) presents the contact mode showing the variation of Fn (gray ), as a function of a step (dark ). In panel (e), the geometric features of the tip–surface contact are depicted (see text).
than vectors such as forces. For this purpose, let us consider a Lennard-Jones potential. U ðrÞ ¼
A B þ 12 ; 6 r r
where r is the tip–sample distance, A ¼ 1077 Jm6 and B ¼ 10134 Jm12. This describes the interaction of all the atoms in a particular solid (9). A rough approximation to the AFM tip–sample interaction is to consider the approach of two such atoms where, as depicted in the inset of Fig. 2a, the lower part is a solid surface and the above it is the apex of a sharp tip that moves down. We can find an interaction force such that F ðrÞ ¼ ðdU =drÞ and this force is attractive (F < 0) when r > r0 or repulsive (F > 0)
202
P.J. de Pablo
when r < r0, (Fig. 2a). These regions define the attractive and the repulsive regimes of operation, respectively. Now let us consider the role of these regimes in a force versus distance (F–z) AFM experiment that involves the tip approaching to the surface (Fig. 2b). The experiment starts with the tip situated far from the surface in the attractive regime. As the tip is approaching and as soon as the gradient (i.e., the slope) of the force equals the cantilever spring constant, the tip jumps to the surface from dark point 1 to the dark point 2 (both connected by the slope). This is seen in the F–z of Fig. 2b at point A like a sudden jump of the cantilever deflection (vertical scale). Thus, the tip establishes mechanical contact with the surface and it rapidly enters the repulsive regime (FN > 0). The z-piezo (horizontal scale) is elongated until a given FN or deflection value is reached and then stops (point B of Fig. 2b). The external loading force Fext can be calculated as the difference between the zero deflection position (i.e., before the jump to contact in A) and the deflection at point B. Since the vertical scale is about 10 nm per division and the cantilever spring constant k ¼ 0.1 N/m, we have Fext ~ 0.5 div 10 nm/div 0.1 nN/nm ¼ 0.5 nN. Subsequently, the z-piezo retrace cycle starts, and the tip is released from the surface at C (i.e., once more where the derivative of the tip– surface force equals the cantilever spring constant) jumping from the gray point 1 to 2 following the gray arrow. The cantilever deflection jumps-off to zero with a dampened oscillation. This jump-off is known as the adhesion force Fadh (in this case Fadh ~ 2div 10 nm/div 0.1 nN/nm ¼ 2 nN). The total force at point B is the sum of Fext and Fadh (i.e., 2.5 nN). It is interesting to note that no matter how small Fext is the total force applied to the surface will always be at least Fadh. 2.2. Contact Mode
The contact mode is the simplest operational method used for AFM and it was the first to be developed, in the mid-1980s (4). Here, the tip is brought into contact with the surface until a given deflection in the cantilever (Fn) is reached and the tip then scans a square area of the surface to obtain a topographic map. By elongating or retracting the z-piezo, the feedback algorithm tries to maintain the cantilever deflection constant by comparing the Fn signal with a set point reference value established by the user. The topographic data are obtained by recording the z-piezo voltage that the feedback algorithm is applied to correct the cantilever deflection at each position on the surface. Since the z-piezo is calibrated, the voltages are transformed into heights and a topography map is obtained. Let us consider a simple example where the tip is scanning a step (dark line in Fig. 2d) with Fn ¼ k Dz. When the cantilever moves to the upper part of the step, it undergoes a deflection greater than Dz. Therefore, the feedback algorithm retracts the z-piezo in order to achieve the same deflection
11
Introduction to Atomic Force Microscopy
203
Dz as when the tip was in the lower part of the step and as a consequence, a topographic profile of the step is obtained. On the other hand, Fn varies at the step and is corrected by the feedback, which can be observed as a peak in the deflection signal (gray line in Fig. 2d). The latter is known as the “constant deflection” mode, otherwise named as the constant height mode, since the z-piezo is not modified and a map of the changes in Fn is obtained. The reader is encouraged to reproduce Fig. 2d when the tip goes down the step. Let us consider now the lateral resolution that can be achieved in contact mode. We can make an estimate of this parameter by applying the Hertz theory (10), which accounts for the deformation of solids in contact. Once the tip is in contact with the sample, the radius r of the tip–surface contact area is given by (Fig. 2e): 3FR 1=3 r¼ ; 4E where F is the applied force, i.e., Fn, E* is the effective Young modulus expressed by 1 1 vt2 1 vs2 ¼ þ ; E Et Es with Et, nt and Es, ns being the Young Modulus and Poisson ratio for the tip and sample, respectively. R is the effective radius expressed as a combination of the tip Rt and sample radius Rs. 1 1 1 ¼ þ R Rtip Rsample In the case of metals, E ¼ 100 GPa, n ~ 0.5, and the mechanical thermal noise of the cantilever is10 pN, with Rt ~ 20 nm, the radius of contact r is about 0.15 nm, which implies atomic resolution. However, the adhesion force in air (see below) is about 5 nN, increasing r to about 1 nm. Atomic resolution can be achieved by either working in liquids (11) or in UHV conditions (12), where even individual atoms can be chemically identified (13). 2.3. Geometrical Dilation
When the surface asperities are comparable to the tip radius (which is very common in AFM experiments), the size of the tip plays an important role. The tip distorts the image due to the dilation of certain features of the image by the finite tip size (14). Such dilation effects occur in all AFM operational modes. An example is shown in Fig. 3a, b, in which we image singlewalled carbon nanotubes that are essentially rolled-up graphene in the form of a cylinder of a few nanometers in diameter. In Fig. 3c, two profiles of similar carbon nanotubes are compared showing that the nanotubes of Fig. 3b (dark line of Fig. 3c)
204
P.J. de Pablo
Fig. 3. Geometrical dilation and sample preparation. Panels (a) and (b) are AFM images of the same sample of carbon nanotubes adsorbed on silicon oxide. The topographic profiles obtained along the lines in (a) show different widths due to dilation in (c) (see text). Panel (d) shows the geometric parameters in the dilation process. The line depicts the dilated section of the carbon nanotube section (circle).
are wider than those in Fig. 3a (gray line of Fig. 3c) due to the larger tip used for imaging Fig. 3b. The geometric dilation effect is seen in Fig. 3c whereby the scanning of the carbon nanotube by the tip results in the Black profile since the tip cannot get closer to the tube than the tip radius rt. Therefore, by geometrical
11
Introduction to Atomic Force Microscopy
205
considerations of Fig. 3d, the tip radius rt can be calculated as rt ¼ b 2 =2h . The topographies of Fig. 3a, b have been obtained using tips with a radius of 15 nm and 70 nm, respectively. We can conclude that the sharper is the tip the better it is for imaging. It is quite instructive to exanimate the dilation effects in biological specimens, such as single viral particles. A very popular and impressive experiment used to teach AFM is to image graphite (in particular, highly oriented pyrolitic graphite or HOPG) in air conditions (15), where the elastic deformation of the tip–sample contact plus the dilation effects result in atomic corrugation, although atomic defects are not visualized. 2.4. Dynamic Modes
Dynamic modes (DMs) (16) are those in which thepcantilever is ffiffiffiffi oscillated near to or at its resonance frequency (o0 / E T =L 2 for a rectangular cantilever). As the tip approaches the sample, the oscillating amplitude decreases until it establishes contact with the sample, following a similar cycle to that in Fig. 2b but now with oscillation. Therefore, the feedback loop involves the amplitude rather than the Fn and by keeping the oscillation amplitude constant, a topographical map can be obtained. The amplitude is reduced because the resonance frequency (Do) changes with the tip–sample distance z and with the tip–sample interaction force Fts according to Do=o0 / ð1=2kÞðdFts =drÞ, thereby decreasing with attractive forces. The new resonance frequency o is positioned to the left of o0 and since the cantilever is still oscillating at o0, the cantilever amplitude decreases. The very high lateral forces that are applied to the surface (17), which can damage the sample, present an important problem in the contact mode. This is especially important for single biomolecules adsorbed onto a surface, these being delicate samples from which to obtain images. However, when operated in noncontact mode (18), DM does not apply huge dragging forces and as such it is commonly used to image weakly attached molecules on surfaces in air. Maps other than topographical maps can also be obtained in DM, such as the phase (the time difference between the excitation and the response of the cantilever) map, which in air carries information on the composition of the sample (see specific details in ref. 18). When DM is used in liquid, the landscape completely changes as the viscosity of water reduces the resonance frequency about fourfold and the quality factor (a measure of the cantilever damping) of the oscillation is reduced from ~100 to 10. When oscillating the cantilever in liquid, mechanical contact between tip and sample is established, thereby applying lateral and normal forces that may damage the specimen (19).
206
P.J. de Pablo
2.5. Jumping or Pulse Force Mode
3. Biological Applications of Imaging AFM
Jumping or pulse force mode (JM) (20, 21) is a contact mode where lateral tip displacement occurs when the tip and sample are not in mechanical contact, thereby avoiding shear forces and the corresponding damage to the tip–sample system. A F–z curve (Fig. 2b) is obtained at every point of the image, moving the tip to the next point at the end of each cycle when the tip and sample lose their contact. Feedback is engaged at point B of Fig. 2b, moving the z-piezo in a convenient way to maintain a constant deflection or loading force Fext. Adhesion force maps can be routinely obtained by using JM in air, which provides compositional or geometrical information of the surface (22). The adhesion force between the tip and the surface can be described as Fadh ¼ 4pRgL cos y þ 1:5pDgR, where gL is the water surface tension and y is the angle of the water meniscus present between the tip and surface; R is the effective radius (described above), and Dg is the tip–surface energy difference. Although the effective radius R is present in both terms, the first one mainly informs about the hydrophobicity of the sample and as such a rough estimate in air at room temperature results in ~7 nN for a Rt of about 20 nm. In air, the first term is the main contribution to Fadh while the second one depends mainly on the tip–surface geometry. The importance of the first term can be appreciated by comparing the F–z curves taken from glass in both air (Fig. 2b) and liquid (Fig. 2c): in liquid the adhesion force is almost absent, since there is no water meniscus between the tip and sample, although some hysteresis appears in the F–z curve due to the dragging of the water on the cantilever.
Let us look first at the applications of AFM for imaging in biology when operated in air. For example, DM is readily used for imaging DNA on mica. As DNA and mica are both negatively charged, MgCl2 is added to the DNA solution so that the Mg2+ ions are sandwiched between DNA and mica, resulting in the adsorption of DNA molecules on the surface. Subsequently, the sample is dried out and DM is used to image the DNA molecules on the surface. Although DNA itself has been the focus of much research using AFM (23), the obvious application of AFM is to investigate the binding of proteins to DNA, for there is now an important body of literature (24–26). In this kind of single-molecule experiments, the researcher is not interested in the average result of the bulk reaction but rather in the action of single proteins on DNA. Hence, once the protein is pipetted into the DNA solution, the reaction starts and the DNA–protein complex is adsorbed onto the mica at the appropriate time and air dried. Therefore, on the
11
Introduction to Atomic Force Microscopy
207
surface, there is a snapshot of the process taking place between proteins and DNA, and the topography provides single-molecule information of the protein–DNA complex. To pick up just one example of many, this technique has been used to study how the H-NS enzyme stops the activity of DNA-polymerase (27). While the DNA–protein complexes can be seen before copying commences, the DNA-polymerase (the big round blob on the DNA filament) becomes trapped between two pieces of DNA that are bridged by the H-NS enzyme, stopping the copying process. However, in biology, we are particularly interested in AFM imaging in a liquid milieu. This is an important issue since biomolecules are normally fully functional in aqueous solutions. First, good attachment of the biomolecule to the supporting surface is required to achieve good resolution during the imaging process. Although biomolecules can be covalently linked to a chemically modified surface (28), covalent modifications could potentially damage the biomolecules, and therefore this tends to be avoided. Fortunately, physisorption is usually sufficient and thus, the specimens can be directly adsorbed in a physiological buffer. The relevant forces that drive the physisorption process are the van der Waals force, the electrostatic double-layer force (EDL force), and the hydrophobic effect (29). Unlike the van der Waals interaction, the EDL force depends strongly on the concentration and valence of charged solutes, as well as the surface charge density of both surface and specimen. The EDL force between two equally charged surfaces is repulsive and hence opposite to the van der Waals attraction (30). To this end, one of the most extended AFM applications is to image two-dimensional protein crystals in liquid, such as membranes, by using contact mode. Biological macromolecules become attached to the surface (mica, silicon, gold, glass, etc.) when there is a net attractive force between them and the surface pulling their surfaces into contact. This force can be estimated to the light of the DLVO theory (9) like the sum of the electrostatic force between surface and molecule Fel, and the van der Waals interaction FvdW. FDLVO ¼ Fel ðzÞ þ FvdW ðzÞ ¼
2ssurf ssample z=lD Ha e ee e0 6pz 3
where z is the distance between the surface and specimen; ssurf and ssample are the charge densities of surface and specimen, respectively; ee and e0 are the dielectric constants of the electrolyte and the vacuum, respectively; lD is the Debye length that depends on the electrolyte valence (29); and Ha the Hamaker constant. The adsorption of a sample onto freshly cleaved mica (atomically flat) can be manipulated by adjusting both the ion content and the pH of the buffer solution. An estimate of the FDLVO
208
P.J. de Pablo
Fig. 4. Imaging the purple membrane and single viruses. Establishing the buffer conditions for imaging: in (a) the interaction force between the purple membrane and the mica surface is depicted as a function of the distance and the electrolyte concentrations. Panel (b) shows a typical AFM image of a purple membrane in liquids with bacteriorhodopsin, a light absorbing membrane protein (adapted from refs. 29 and 31). Images (c)–(e) are three MVM particles showing three-, two-, and fivefold symmetries, respectively (after ref. 38).
between a purple membrane (a 2D crystal lattice formed by bacteriorhodopsin) and mica is shown in Fig. 4a, highlighting the strong influence of the electrolyte concentration (after ref. 29). Figure 4b shows a region of purple membrane adsorbed on mica (after ref. 31). Interestingly, in this kind of set-up, the cantilever can be used to apply forces that trigger conformational changes in single proteins such us GroEL (32). AFM can also be used to visualize single proteins at work and generally AFM is used in dynamic mode and in liquid for this purpose since it is faster (33). Maximum peak forces of a few nN are applied (19), in relation to the stiffness of the sample (34) and, although these forces could in principle be damaging to the sample, they are applied for very short periods of time such as 10% of an oscillating period (i.e., for cantilever with a resonance frequency of 10 kHz in liquid, the forces are applied during 10 ms for every 0.1 ms). Using this method, it has been possible to visualize, for example, the activity of RNA polymerase on DNA,
11
Introduction to Atomic Force Microscopy
209
for instance (35), or the conformational changes in a DNA–repair complex upon binding DNA (36). Imaging viruses by AFM is a good way to illustrate this methodology. Structural and chemico-physical characterization has been critical to understand the biology of viruses. X-ray crystallography and EM (cryo-EM/IR) techniques have traditionally been used and they provide direct three-dimensional structural information as well as allowing both the interior and the surface of the virus to be visualized. However, a limitation of these techniques is that they are both averaging (“bulk”) techniques and thus they present an average time and space model of the entire population of particles found in the crystal or on the EM grid. As such, these techniques provide rather limited information (although some) on the characteristics of the individual elements in the population that distinguishes them from the average. For this reason, the beautifully symmetrical and apparently perfect models of larger viruses derived from these techniques may be somewhat deceptive as they may not be entirely representative of the individual viruses. Therefore, the imaging and interpretation of individual virus particles should always be a part of the initial virus characterization (37). Since viruses are individual particles that adsorb weakly to typical surfaces, they are prone to destruction by lateral forces in contact mode, as they are not held by a surrounding neighborhood (alike a protein in a two-dimensional crystal). In this case, the jumping mode is used (instead of the contact mode) since the loading forces can be accurately controlled thereby avoiding the application of lateral forces. Single viral particles of the minute virus of mice (MVM) adsorbed in three-, two-, and fivefold symmetries, respectively, are shown in Fig. 4c–e. By making nanoindentations with F–z curves on single viruses, it has been shown that single-stranded DNA within this virus does contribute to the overall mechanical stiffness of the virus particles (38), something that could be important for viral stability during the extracellular cycle. Moreover, it is possible to selectively disrupt these DNA–protein interactions and thereby engineer a virus with altered mechanical properties (39).
4. Perspectives and Future Directions
Some of the current developments in imaging with AFM are aimed at increasing the speed (a typical AFM image takes minutes to be acquired) in order to directly visualize the dynamics of biological processes in real time. For example, conformational changes of single myosin V proteins on mica have been shown with high video rates (80 ms/frame) using very soft cantilevers
210
P.J. de Pablo
Fig. 5. Dynamic mode and stiffness. Panel (a) presents the topography of a mature f29 virion and (b) the simultaneously acquired phase lag related with the stiffness of the sample. They are both compared at a longitudinal profile in panel (c). Panel (d) shows the three-dimensional EM reconstruction of the virion (dark gray ) with a superposed image of the expected AFM topography including the geometric tip dilation (adapted from ref. 42).
with a high frequency of resonance in liquids (40). This was achieved by decreasing the cantilever thickness and reducing the cantilever width and length proportionally. Beyond the indentation experiments (41), dynamic mode operating in liquids can also extract information about mechanical properties (e.g., stiffness), by mapping the phase lag (Fig. 5) (42). Finally, the invention of noninvasive imaging techniques where sample destruction is minimized is also important. For example, frequency modulation AFM (43) is a dynamic technique where forces of about tens of pN can be applied to the surface. This promising technique is based on the use of three simultaneous feedbacks. A phase lock loop ensures that the cantilever is always in resonance while a second feedback (working over the phase lock loop) changes the tip–sample gap to keep a set point frequency, such that its output gives the topography. Finally, a
11
Introduction to Atomic Force Microscopy
211
third feedback is used to maintain the oscillation amplitude constant by changing the amplitude of the cantilever driving signal, which results in more stable operation. References 1. Binnig G & Rohrer H (1982) Scanning Tunneling Microscopy. Helvetica Physica Acta 55, 726–735. 2. Chen CJ (1993) Introduction to Scanning Tunneling Microscopy (Oxford University Press, Oxford). 3. Baro AM, Miranda R, Alaman J, Garcia N, Binnig G, Rohrer H, Gerber C, & Carrascosa JL (1985) Determination of Surface-Topography of Biological Specimens at High-Resolution by Scanning Tunnelling Microscopy. Nature 315, 253–254. 4. Binnig G, Quate CF, & Gerber C (1986) Atomic Force Microscope. Phys Rev Lett 56, 930–933. 5. Schmalz G (1929) Uber Glatte und Ebenheit als physikalisches und physiologishes Problem. Verein Deutscher Ingenieure. 6. Meyer G & Amer NM (1988) Novel Optical Approach to Atomic Force Microscopy. Applied Physics Letters 53, 1045–1047. 7. Sader JE, Chon JWM, & Mulvaney P (1999) Calibration of rectangular atomic force microscope cantilevers. Review of Scientific Instruments 70, 3967–3969. 8. Butt HJ & Jaschke M (1995) Calculation of Thermal Noise in Atomic-Force Microscopy. Nanotechnology 6, 1–7. 9. Israelachvili J (2002) Intermolecular and surface forces (Academic Press, London). 10. Johnson KL (1985) Contact mechanics (Cambridge University Press, Cambridge). 11. Ohnesorge F & Binnig G (1993) True Atomic-Resolution by Atomic Force Microscopy through Repulsive and Attractive Forces. Science 260, 1451–1456. 12. Giessibl FJ (1995) Atomic-Resolution of the Silicon (111)-(77) Surface by Atomic-Force Microscopy. Science 267, 68–71. 13. Sugimoto Y, Pou P, Abe M, Jelinek P, Perez R, Morita S, & Custance O (2007) Chemical identification of individual surface atoms by atomic force microscopy. Nature 446, 64–67. 14. Villarrubia JS (1997) Algorithms for scanned probe microscope image simulation, surface reconstruction, and tip estimation. J Res Natl Inst Stan 102, 425–454. 15. Marti O, Drake B, Gould S, & Hansma PK (1988) Atomic Resolution Atomic Force
Microscopy of Graphite and the Native Oxide on Silicon. Journal of Vacuum Science & Technology a-Vacuum Surfaces and Films 6, 287–290. 16. Martin Y, Williams CC, & Wickramasinghe HK (1987) Atomic Force Microscope Force Mapping and Profiling on a Sub 100-a Scale. Journal of Applied Physics 61, 4723–4729. 17. Carpick RW, Ogletree DF, & Salmeron M (1997) Lateral stiffness: A new nanomechanical measurement for the determination of shear strengths with friction force microscopy. Applied Physics Letters 70, 1548–1550. 18. Garcia R & Perez R (2002) Dynamic atomic force microscopy methods. Surface Science Reports 47, 197–301. 19. Legleiter J, Park M, Cusick B, & Kowalewski T (2006) Scanning probe acceleration microscopy (SPAM) in fluids: Mapping mechanical properties of surfaces at the nanoscale. P Natl Acad Sci USA 103, 4813–4818. 20. Miyatani T, Horii M, Rosa A, Fujihira M, & Marti O (1997) Mapping of electrical doublelayer force between tip and sample surfaces in water with pulsed-force-mode atomic force microscopy. Applied Physics Letters 71, 2632–2634. 21. de Pablo PJ, Colchero J, Gomez-Herrero J, & Baro AM (1998) Jumping mode scanning force microscopy. Applied Physics Letters 73, 3300–3302. 22. De Pablo PJ, Colchero J, Gomez-Herrero J, Baro AM, Schaefer DM, Howell S, Walsh B, & Reifenberger R (1999) Adhesion maps using scanning force microscopy techniques. Journal of Adhesion 71, 339–356. 23. Hansma HG, Sinsheimer RL, Li MQ, & Hansma PK (1992) Atomic Force Microscopy of Single-Stranded and Double-Stranded DNA. Nucleic Acids Research 20, 3585–3590. 24. Lyubchenko YL, Jacobs BL, Lindsay SM, & Stasiak A (1995) Atomic-Force Microscopy of Nucleoprotein Complexes. Scanning Microscopy 9, 705–727. 25. Dame RT, Wyman C, & Goosen N (2003) Insights into the regulation of transcription by scanning force microscopy. Journal of Microscopy-Oxford 212, 244–253. 26. Janicijevic A, Ristic D, & Wyman C (2003) The molecular machines of DNA repair:
212
P.J. de Pablo
scanning force microscopy analysis of their architecture. Journal of Microscopy-Oxford 212, 264–272. 27. Dame RT, Wyman C, Wurm R, Wagner R, & Goosen N (2002) Structural basis for HNS-mediated trapping of RNA polymerase in the open initiation complex at the rrnB P1. Journal of Biological Chemistry 277, 2146–2150. 28. Wagner P, Hegner M, Guntherodt HJ, & Semenza G (1995) Formation and in-Situ Modification of Monolayers Chemisorbed on Ultraflat Template-Stripped Gold Surfaces. Langmuir 11, 3867–3875. 29. Muller DJ, Amrein M, & Engel A (1997) Adsorption of biological molecules to a solid support for scanning probe microscopy. Journal of Structural Biology 119, 172–188. 30. Muller DJ, Janovjak H, Lehto T, Kuerschner L, & Anderson K (2002) Observing structure, function and assembly of single proteins by AFM. Progress in Biophysics & Molecular Biology 79, 1–43. 31. Muller DJ, Schabert FA, Buldt G, & Engel A (1995) Imaging Purple Membranes in Aqueous-Solutions at Subnanometer Resolution by Atomic-Force Microscopy. Biophysical Journal 68, 1681–1686. 32. Viani MB, Pietrasanta LI, Thompson JB, Chand A, Gebeshuber IC, Kindt JH, Richter M, Hansma HG, & Hansma PK (2000) Probing protein-protein interactions in real time. Nature Structural Biology 7, 644–647. 33. Moreno-Herrero F, Colchero J, GomezHerrero J, & Baro AM (2004) Atomic force microscopy contact, tapping, and jumping modes for imaging biological samples in liquids. Physical Review E 69. 34. Xu X, Carrasco C, de Pablo PJ, GomezHerrero J, & Raman A (2008) Unmasking imaging forces on soft biological samples in liquids when using dynamic atomic force microscopy: A case study on viral capsids. Biophysical Journal 95, 2520–2528. 35. Kasas S, Thomson NH, Smith BL, Hansma HG, Zhu XS, Guthold M, Bustamante C,
Kool ET, Kashlev M, & Hansma PK (1997) Escherichia coli RNA polymerase activity observed using atomic force microscopy. Biochemistry 36, 461–468. 36. Moreno-Herrero F, de Jager M, Dekker NH, Kanaar R, Wyman C, & Dekker C (2005) Mesoscale conformational changes in the DNA-repair complex Rad50/Mre11/Nbs1 upon binding DNA. Nature 437, 440–443. 37. Plomp M, Rice MK, Wagner EK, McPherson A, & Malkin AJ (2002) Rapid visualization at high resolution of pathogens by atomic force microscopy – Structural studies of herpes simplex virus-1. American Journal of Pathology 160, 1959–1966. 38. Carrasco C, Carreira A, Schaap IAT, Serena PA, Gomez-Herrero J, Mateu MG, & Pablo PJ (2006) DNA-mediated anisotropic mechanical reinforcement of a virus. P Natl Acad Sci USA 103, 13706–13711. 39. Carrasco C, Castellanos M, de Pablo PJ, & Mateu MG (2008) Manipulation of the mechanical properties of a virus by protein engineering. P Natl Acad Sci USA 105, 4150–4155. 40. Ando T, Kodera N, Takai E, Maruyama D, Saito K, & Toda A (2001) A high-speed atomic force microscope for studying biological macromolecules. P Natl Acad Sci USA 98, 12468–12472. 41. Ivanovska IL, de Pablo PJ, Ibarra B, Sgalari G, MacKintosh FC, Carrascosa JL, Schmidt CF, & Wuite GJL (2004) Bacteriophage capsids: Tough nanoshells with complex elastic properties. Proc. Natl. Acad. Sci. U. S. A. 101, 7600–7605. 42. Melcher J, Carrasco C, Xu X, Carrascosa JL, Gomez-Herrero J, de Pablo PJ, Raman A (2009) Origins of phase constrat in the atomic forces microscopy in liquids. P Natl Acad Sci USA 106, 13655–13660. 43. Hoogenboom BW, Hug HJ, Pellmont Y, Martin S, Frederix PLTM, Fotiadis D, & Engel A (2006) Quantitative dynamic-mode scanning force microscopy in liquid. Applied Physics Letters 88, 193109.
Chapter 12 Sample Preparation for SFM Imaging of DNA, Proteins, and DNA–Protein Complexes Dejan Ristic, Humberto Sanchez, and Claire Wyman Abstract Direct imaging is invaluable for understanding the mechanism of complex genome transactions where proteins work together to organize, transcribe, replicate, and repair DNA. Scanning (or atomic) force microscopy is an ideal tool for this, providing 3D information on molecular structure at nanometer resolution from defined components. This is a convenient and practical addition to in vitro studies as readily obtainable amounts of purified proteins and DNA are required. The images reveal structural details on the size and location of DNA-bound proteins as well as protein-induced arrangement of the DNA, which are directly correlated in the same complexes. In addition, even from static images, the different forms observed and their relative distributions can be used to deduce the variety and stability of different complexes that are necessarily involved in dynamic processes. Recently available instruments that combine fluorescence with topographic imaging allow the identification of specific molecular components in complex assemblies, which broadens the applications and increases the information obtained from direct imaging of molecular complexes. We describe here basic methods for preparing samples of proteins, DNA, and complexes of the two for topographic imaging and quantitative analysis. We also describe special considerations for combined fluorescence and topographic imaging of molecular complexes. Key words: Scanning force microscopy, Atomic force microscopy, DNA–protein complexes, Single-molecule imaging, Combining fluorescence and topography
1. Introduction Proper expression and maintenance of genomic DNA are executed with precision and control by the coordinated action of proteins arranged in specific assemblies on DNA. Understanding how these proteins work together, to package, transcribe, replicate, and repair DNA, requires knowing how they are arranged into functional assemblies. Direct images of protein–DNA complexes are an essential tool to achieve this understanding. They reveal a wealth of inherently correlated information on structures, their variation, and distributions. Scanning force microscopy (SFM), also commonly known as atomic force microscopy (AFM), Erwin J.G. Peterman and Gijs J.L. Wuite (eds.), Single Molecule Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 783, DOI 10.1007/978-1-61779-282-3_12, # Springer Science+Business Media, LLC 2011
213
214
D. Ristic et al.
is an excellent and practical method for direct imaging of proteins, DNA, and complexes of the two at the single molecule/complex level. Molecules and complexes are individually analyzed, providing information on the variety of arrangements possible and their frequency in a mixture. Importantly, this type of single-molecule structural analysis allows coherent description of features that would otherwise be lost in the averaging of bulk analysis. In addition, direct observation allows correlation of multiple structural features of individual molecular complexes. Sample preparation is relatively simple and requires component biomolecules in easily obtainable amounts and purity. SFM imaging has literally provided a new view of the molecular machinery responsible for DNA processing and by this new insight into molecular mechanisms of vital processes such as DNA packaging, repair, replication, and transcription (1, 2). Mechanistic information is obtained by quantitative analysis of image data. Typically, it is necessary to devise an appropriate scheme to divide complexes or structures observed into relevant categories and determine the distribution of these categories in different conditions. For instance, the percentage of DNA bound by a protein at a specific binding site versus at nonspecific sites would be determined as a function of conditions such as the addition of a nucleotide cofactor or another protein. SFM images are also ideal for revealing mechanistically important proteininduced distortions in DNA such as changes in DNA bending, contour length, and flexibility. Protein-induced distortions of DNA can be determined at specific binding sites and at nonspecific sites for comparison, usually from the same sample (3). DNA substrates are constructed with specific sequence or structural features at defined locations, such as a recognition sequence, a single modified or damaged base, nicks, gaps, various lengths of single- and double-stranded DNA, and complex DNA junctions such as those recognized by replication or recombination proteins. In all cases, the DNA strands not including the specific feature are by definition nonspecific binding sites and serve as unavoidable internal control DNA. Proteins and their functional assemblies often involve multiple DNA sites and strands. These functional assemblies, for example, DNA looped between protein bound at two sites or proteins associating to join or connect multiple DNA molecules, are sometimes hard to define by indirect means but obvious by simply looking at images. Biomolecules are typically deposited onto an atomically flat mica surface and imaged in air. The samples are dried of bulk water but not desiccated and likely retain their native structure (4). The volume of the particles observed can be used to estimate size and multimeric state. DNA can be deposited on mica by equilibration on the surface so that it is not kinetically trapped. In this way, the arrangement of the DNA on the surface accurately reflects its
12
Sample Preparation for SFM Imaging of DNA, Proteins. . .
215
properties in solution, such as contour length, flexibility, and presence of bent segments. Protein–DNA complexes prepared in appropriate biochemical conditions are deposited onto mica in a similar fashion. Changes in the DNA induced by bound proteins can be accurately measured. Some proteins have features that are distinct in SFM images, such a long coiled-coil regions or multiple globular lobes, but most proteins appear as similar globular objects. Here, we describe basic methods for obtaining topographic images of DNA, proteins, and their complexes that can be used for a variety of quantitative structural analyses. Many functionally important protein–DNA complexes include multiple component proteins. The identity of components in multi-protein complexes can be estimated based on their volume and known molecular weights. However, it is not always possible to determine molecular composition and stoichiometry unambiguously by volume alone. It is, therefore, necessary to label or tag specific proteins to identify different, possibly similarly sized, components in complex assemblies. Combining fluorescence and topographic imaging allows specific identification of fluorescently labeled proteins and greatly expands the application of SFM to analysis of increasingly complex molecular assemblies. Several instruments that combine SFM topographic imaging with fluorescent imaging are currently commercially available. There are specific challenges for using these instruments for the analysis of protein–DNA complexes that we address here. It is necessary to deposit the molecules of interest onto an atomically flat surface such as mica for topography, but this substrate must also be sufficiently optically transparent to allow fluorescence. In addition, appropriate marker objects are essential to achieve nanometer accuracy in aligning optical and topographic images when the objects of interest are smaller than optical resolution. We briefly describe methods for sample preparation and image alignment that allow properly correlated SFM and fluorescence imaging.
2. Materials 2.1. Instrumentation
1. Scanning probe microscope: These instructions are guided toward eventual imaging by intermittent contact mode in air. We have a Digital Instruments MultiMode scanning probe microscopes. The methods for sample preparation and guidelines for data acquisition are applicable to any similar instrument and imaging mode (see Note 1). 2. Computers (PC, Mac) that meet the requirements of image analysis software. Software such as ImageJ (http://rsbweb. nih.gov/ij/), Image SXM (extended version of NIH image
216
D. Ristic et al.
by Steve Barrett, Surface Science Research Centre, Univ. of Liverpool, Liverpool, U.K.), WSxM (Nanotec Electronica S.L.), or a similar image analysis software is required for quantifications of various features of visualized molecules. 2.2. General Supplies
1. Glass pasture pipettes. 2. Standard facial tissues. 3. Lens cleaning tissues (Whatman 105). 4. Forceps (DZM, Italy). 5. A source of filtrated air or N2. 6. For preparation of all solutions, Mili-Q filtered deionized water (resistivity 16 MO cm, TOC 1–5 ppb). 7. Standard protein deposition buffer: 20 mM HEPES-KOH, pH 7–8, 50–100 mM KCl (or NaCl), and 1 mM DTT (chemical supplied by SigmaAldrich). 8. Standard DNA deposition buffer: 5–10 mM HEPES-KOH pH 7.5, 5–10 mM MgCl2 (chemical supplied by Sigma–Aldrich).
2.3. Sample Substrates
1. Metal disks such as those usually supplied with a scanning probe microscope. 2. Mica sheets (Muscovite Mica V-5 quality, thickness ¼ 0.15–0.21 mm, Electron Microscopy Sciences). 3. Punch & die set (Precision brand) for cutting mica into disks. The diameter of mica disks is 1–2 mm smaller than the diameter of metal disks. 4. Superglue to attach mica disks to metal disks. 5. Invisible tape (19 mm width, Magic from 3M) to cleave mica.
2.4. DNA Preparation
General molecular biology reagents and instruments for preparation of DNA (purification from bacteria or PCR amplification) are needed. General knowledge on methods for DNA preparation, such as those found in various editions of Molecular Cloning: A Laboratory Manual from CSHL Press, is assumed.
2.5. Combining Fluorescence and SFM
For identifying specific molecules and nano-objects in SFM topography images, we use a combined SF-fluorescent microscope setup which consists of the following: An inverted fluorescence microscope equipped with high numerical aperture (1.45) objectives with a minimum magnification of 60 (Nikon TE2000); signal detections with a Cascade II:512B EMCCD camera (Princeton Instruments); running MetaMorph software (Molecular Devices) or custom made Labview (National Instruments) software for microscope operation and image acquisition;
2.5.1. Instrumentation
12
Sample Preparation for SFM Imaging of DNA, Proteins. . .
217
and a coupled NanoWizard®II scanner (JPK instruments). Similar instruments and equivalent components would perform as well. 2.5.2. Sample Substrates for Combined Fluorescence and SFM
1. Mica sheets (Muscovite V-1 quality, from Electron Microscopy Science). 2. Glass coverslips, round, 24-mm diameter, thickness 00 (from Menzel-Glazer). 3. Optical adhesive NOA88 (Norland products). 4. Handheld UV lamp. 5. Sodium tetrahydridoborate 0.25% w/v solution in water (see Note 2). 6. Invisible tape (19 mm width, Magic from 3M).
2.5.3. Fluorescent Markers and Labels
1. FluoSpheres® carboxylate-modified microspheres (0.04 mm diameter, yellow–green fluorescent (505/515), orange fluorescent (540/560), and red fluorescent (580/605) from Invitrogen).
3. Methods 3.1. Considerations for Preparing Proteins Used in SFM Imaging
Purified proteins should be stored and used without addition of stabilizing proteins such as BSA. The protein purity requirements differ with application and the nature of possible contaminants. For example, if the protein of interest binds to DNA, contaminating proteins not bound to DNA can be ignored. Also, DNAbinding proteins should be without any trace of DNA. In general, 80% protein purity, estimated by Coomassie blue staining of the purified protein displayed by gel electrophoresis, is sufficient for SFM analysis.
3.2. Considerations for Preparing DNA Used in SFM Imaging
The length of DNA to be used depends on the eventual data desired and the specific experimental question. In general, DNA should be at least 500 bp or longer so that it is obviously DNA by appearance based on relative width, height, and length. We commonly use DNA in the range of 500–3,000 bp. Linear DNA is generally more useful. Circular DNA tends to fold over itself on the surface, making it harder to analyze. In addition, linear DNA allows determining the location of a specific sequence or feature of interest by its relative position from an end. DNA should be clean and free of proteins or other material that will deposit onto mica and complicate the imaging. Kits/columns for DNA purification (Quiagen, GE Healthcare, Sigma–Aldrich) usually produce DNA of sufficient purity for SFM, though some problems with residual column material or buffer components occasionally occur.
218
D. Ristic et al.
The cleanest DNA is obtained by the following methods depending on the source: from solutions such as PCRs or by enzyme digestion purify DNA by phenol:chloroform extraction followed by chloroform extraction; from gel slices, purify DNA by electroelution; purify DNA from plasmid or phage preparations by cesium gradient purification. These DNA purification and isolation methods often reduce final yield. Therefore, the tradeoff between purity and yield has to be considered when choosing purification methods for individual applications. Ethanol precipitation tends to result in contaminating material when imaging, often assumed to be excess salt. For this reason, ethanol precipitation is avoided or preformed with care to wash excess salt from pellets before resuspension. 3.3. Preparation of Mica Substrates
1. Attach mica to metal by applying a very small drop of supergule onto the metal disks, placing the mica disk over the glue and gently pressing. Avoid glue spreading beyond the mica as this will interfere with cleaving the mica. 2. Freshly cleaved mica is prepared by applying scotch tape (Magic, 3M) to the mica glued to metal and peeling off the top mica layers. The peeled off layer stuck to the tape is inspected to see if it is a complete circle. If a complete layer is not cleaved off, the procedure is repeated until a complete layer, smooth unbroken circle, is removed. The mica is usually cleaved for only a few minutes to a half-hour before use to assure a clean surface.
3.4. Immobilizing Molecules on Mica: General Considerations
In order to visualize molecules in SFM, they have to be immobilized on a surface. However, molecules have to be free from the surface to enable their dynamic interactions and to prevent steric hindrance that might affect molecular interactions. Thus, the immobilization of molecules on the surface has to be carefully controlled to enable imaging while minimizing disturbing the relevant molecular interactions. Molecules can be deposited on a surface through specific or nonspecific interaction. We most often take advantage of relatively nonspecific electrostatic adsorption, which depends on the charge of the surface and molecules, and is sufficient to provide controlled attachment of DNA, proteins, and their complexes. Specific interactions such as streptavidin and biotin or digoxigenin and anti-digoxigenin provide much stronger attachment of molecules to the surface with defined molecule orientation. However, surface modification also increases roughness and interferes with imaging. The most commonly used surface for deposition of biomolecules is Muscovite mica. Mica can be cleaved at crystal planes that produce large atomically flat surfaces. This uniform flat surface allows detection of biomolecules that are only a few
12
Sample Preparation for SFM Imaging of DNA, Proteins. . .
219
nanometers high. The mica surface is negatively charged and the heterogeneous charged domains on most proteins result in sufficient deposition without additional treatment of either surface or protein. HEPES buffers, in biologically relevant pH range (pH 7–8), are preferred for protein deposition. Tris–HCl buffers tend to deposit on mica and interfere with SFM imaging. Other common protein storage or reaction buffer components, such as KCl, MgCl2, DTT, ATP, low concentration of detergents such as NP40, and glycerol, do not interfere with protein deposition or eventual imaging in our experience. The composition of buffers used for deposition of proteins is less strict than that used for deposition of DNA because proteins deposit effectively onto mica in a wider range of salt and pH conditions. Thus, it is a good first step to deposit proteins in buffers that are optimal for maintaining protein structure and/or activity. If the protein of interest is already biochemically characterized, the optimal buffer for deposition for SFM would be the same (or very similar) as the optimal buffer for protein activity. In case of an uncharacterized protein, SFM deposition can be done in a standard deposition buffer consisting of 20 mM HEPES, pH 7–8, 50–100 mM KCl (or NaCl), and 1 mM DTT. If the protein appears aggregated upon deposition, increasing salt concentration and/or including some NP40 (0.05%) often helps to prevent undesired protein–protein interaction. Adsorption of negatively charged DNA on negatively charged mica surface requires the presence of divalent cations. Those interested in the effects of different divalent cations on DNA deposition on mica are referred to published studies on this topic (5–8). The mica surface can be modified for more efficient adsorption of DNA. This has advantages and disadvantages. The treatment of mica with either 3-aminopropyltriethoxysilane (APTES) or poly-L-lysine results in a positively charged surface. DNA binding to such surfaces does not require the addition of divalent cations. The resulting DNA is strongly attached to the modified surface and can be imaged by SFM both in air and in buffer. However, DNA does not equilibrate on these modified surfaces but become kinetically trapped (9). Deposition by kinetic trapping results in DNA conformations on mica that are strongly influenced by interaction with the surface. This complicates and in some cases precludes determining experimentally important changes in DNA conformation, such as measuring DNA bends or distortions induced by protein binding, or changes in DNA conformation induced by binding small molecules prior to deposition. In order to measure and analyze protein-induced changes in DNA, we do not use treated mica (9). In addition, any mica treatment results in a rougher surface that can complicate imaging.
220
D. Ristic et al.
The persistence length of DNA can be determined from SFM images and is used as a test of deposition by equilibration on the surface. Persistence length reflects intrinsic flexibility of a polymer and is a well-characterized feature of DNA. Persistence length is obtained by measurements of the contour length and the end-toend distance of deposited DNA as described in (9). A persistence length of ~50 nm is characteristic of DNA and thus indicates deposition by equilibration. We have described that proteins adhere to mica in a wider variety of buffer conditions than DNA. Because of this, protein-bound DNA can also adhere to mica due to the behavior of the proteins. Thus, a protein-bound DNA will often adhere to mica in deposition conditions that are effective for the protein alone, without the need to add Mg2+ or reduce monovalent cation concentration. It should be noted, however, that free DNA, not bound by proteins, will behave differently and may not adhere to mica in conditions effective for adherence of protein–DNA complexes. Thus, deposition conditions should be chosen carefully depending on the goal of the experiment. For instance, if it is necessary to quantify the amount of protein bound and protein-free DNA, controls that show equal deposition of these need to be included. The best conditions for protein–DNA complex formation should be determined by standard quantitative biochemical assays prior to initiating SFM imaging experiments. This will form the starting point for determining conditions to be used in SFM imaging. Binding reactions for SFM typically require more concentrated DNA than many biochemical assays and there are limitations to the excess of protein that can be used. As a general rule, the solution deposited onto mica should have DNA at 1–10 ng/ml and proteins should not be more than 50-fold molar excess to their binding sites. The upper limit to excess protein is due to problems in observing DNA if the surface becomes covered by protein, or even more dramatically, if protein saturates the surface and prevents DNA binding. Thus, protein to DNA ratios and protein concentrations optimal for biochemical assays may not be optimal for SFM imaging. 3.5. Deposition of Protein for SFM Imaging
1. Prepare a solution of protein to be deposited, concentration of about 0.5 mM in an appropriate buffer is a good starting point. Optimal protein concentration for deposition will differ for each protein depending, for instance, on purity, size, and oligomeric state. Usually, protein concentrations of less than 0.5 mM (final in deposition buffer) provide reasonable coverage of the mica surface for further analysis. 2. Place a drop, 5–30 ml depending on the size of the mica surface, of protein solution onto the freshly cleaved mica surface and allow it to deposit for ~30 s.
12
Sample Preparation for SFM Imaging of DNA, Proteins. . .
221
Fig. 1. Washing mica surfaces with water. The sample substrate is held by forceps on the edge of the metal disk, not over the mica. Water or buffer is washed over the surface using a pasture pipette.
3. Rinse the mica surface with Mili-Q filtered deionized water, about one pasture pipette full, as shown in Fig. 1. 4. Blot excess water by touching a piece of facial tissue to the edge of the mica. 5. Dry the mica surface in a stream of filtrated air (or N2). 6. Observe the sample with the SFM; typically scanning fields of 1 1 mm with a Z scale of 5 nm or less will give a good impression of the protein coverage. 7. Assess the protein coverage and modify deposition if needed. If deposition is too crowded (Fig. 2a), additional dilution of five- to tenfold usually results in surface coverage where proteins are nicely separated on the surface and can be analyzed (Fig. 2b). 3.6. Deposition of DNA for SFM Imaging
1. Prepare a solution of DNA to deposit; DNA 0.5–10 ng/ml in deposition buffer of 5–10 mM HEPES, pH 7.5,and 5–10 mM MgCl2. Since the presence of monovalent cations dramatically reduces Mg2+ promoted adsorption of DNA to mica, the concentration of monovalent salt in the solution being deposited should be less than or equal to the concentration of Mg2+. 2. Place a drop, 5–30 ml depending on the size of the mica surface, of the solution containing DNA onto the freshly cleaved mica surface and let it sit to deposit for 30 s to 1 min.
222
D. Ristic et al.
Fig. 2. Examples of excess or appropriate protein coverage. The 37-kD human RAD51 protein was deposited as described from a buffer containing 25 mM HEPES-KOH, pH 8.0, and 100 mM KCl and imaged in tapping mode in air with a Digital Instruments multimode NanoScope. (a) Too much protein covering the mica surface prevents identification and analysis of individual molecules. Here, a 2 mM solution was used for deposition onto mica. (b) Adequate protein deposition showing many single molecules well separated and few overlapping unresolved molecule pairs. Here, a 0.2 mM solution was used for deposition onto mica. The image dimensions are shown; height is indicated by color according the scale shown on the right.
3. Wash the mica surface with Mili-Q filtered deionized water, about one pasture pipette full, as shown in Fig. 1. 4. Blot excess water by touching a piece of facial tissue to the edge of the mica. 5. Dry the mica surface in a stream of filtrated air (or N2). 6. Observe the sample with the SFM; typically scanning fields of 2 2 mm with a Z scale of 2 nm will give a good impression of the DNA coverage. 7. Assess the DNA coverage and modify deposition if needed. Excess DNA coverage is shown in Fig. 3a and appropriate coverage of DNA on mica is shown in Fig. 3b. If the DNA on the surface is too crowded, then try two- to tenfold dilution of the solution for deposition. If there is not enough DNA on the surface (only one, a few, or no molecules in a 2 2 mm scan), one of the several variations in the deposition will usually help: increase the time for the DNA solution sitting on mica to 1–2 min, increase the DNA concentration, increase the Mg2+ concentration, or decrease the monovalent cation concentration. 3.7. Deposition of Protein–DNA Complexes for SFM Imaging
1. Prepare a protein–DNA binding reaction for deposition. This typically should include DNA at 1–10 ng/ml (see Note 3).
12
Sample Preparation for SFM Imaging of DNA, Proteins. . .
223
Fig. 3. Examples of excess and appropriate DNA coverage. DNA was deposited onto mica as described and imaged in tapping mode in air with a Digital Instruments multimode NanoScope. (a) Too much DNA covering mica prevents identification of individual molecules and precludes any meaningful analysis. Here, a solution of 800-bp linear doublestranded DNA at 10 ng/ml in 10 mM HEPES and 10 mM MgCl was used for deposition onto mica. (b) Appropriate DNA density with clearly separated and nonoverlapping DNA molecules. Depending on the length of the DNA, there should be 5–20 molecules in a field. Here, a solution of 0.5 ng/ml DNA in 10 mM MgCl was used for deposition onto mica. This resulted in about 15 isolated, identifiable, and analyzable linear DNA molecules. This linear DNA consists of a 500-bp double-stranded and 300-nt single-stranded segment. The single-stranded DNA appears as a small knob at one end of the linear molecule. The image dimensions are shown; height is indicated by color according the scale shown on the right.
2. Place a drop of the protein–DNA complex sample onto mica, as described above for protein or DNA alone, and allow it to sit for 30 s to 1 min. 3. Wash off unbound material with a small volume of binding buffer, as shown in Fig. 1. 4. Remove excess buffer by touching a tissue to the edge of the mica surface, but the surface is not dried. 5. Cover the surface with 30–40 ml of buffer containing 10 mM HEPES-KOH (pH 7.5) and 10 mM MgCl2 in order for DNA to attach to mica. 6. Almost immediately or after about 5 s, wash the mica with water, about 1 pasture pipette full, as shown in Fig. 1. 7. Blot excess liquid by touching a tissue to the edge of the mica and dry the surface in a stream of filtered air, as described above. 3.8. Deposition of Protein–DNA Complexes and Free DNA in Binding Reactions for SFM Imaging
1. Prepare a binding reaction and dilute with deposition buffer (10 mM HEPES and 10 mM MgCl2) so that the concentration of DNA is 1–10 ng/ml and the concentration of monovalent cations is equal to or less than 10 mM (see Note 4). This will allow free DNA and protein-bound DNA to adhere to mica. The initial binding reaction may need to be adjusted
224
D. Ristic et al.
so that after dilution the solution that will be deposited still contains DNA at 1–10 ng/ml (see Note 5). 2. Place a drop of the protein–DNA complex sample onto mica, as described above for protein or DNA alone, and allow it to sit for 30 s to 1 min. 3. Wash off the unbound material with water, about 1 pasture pipette full, as shown in Fig. 1. 4. Blot excess water by touching a piece of facial tissue to the edge of the mica. 5. Dry the mica surface in a stream of filtrated air (or N2). 6. Observe the sample with the SFM; typically scanning fields of 2 2 mm or 4 4 mm with a Z scale of 2–5 nm will give a good impression of the sample. Density of DNA on the surface, either with or without bound protein, similar to that shown in Fig. 3b is sufficient for most analysis (see Note 6). 3.9. Guidelines for Collecting Image Sets for Analysis
Once the stoichiometry of DNA and protein is optimized for imaging and good coverage of mica is achieved, for proteins alone, DNA alone, or protein–DNA complexes, a collection of images needs to be obtained for eventual analysis. 1. Issues of actual microscope operation and image acquisition are not discussed or described here (refer to specific instrument operating manual). The operation of a scanning force microscope and data acquisition differ depending on the instrument and are beyond the scope of this article. The samples we have described are usually imaged in our laboratory using intermittent contact or “tapping” mode in air. We use standard silicon tapping tips for a variety of suppliers with equivalent success. It is important that the tips have a confirmed end radius of curvature of about 10 nm or less (see Note 7). 2. For most applications, scan sizes ranging from 1 1 mm up to 4 4 mm are most useful. For instance, when analyzing proteins, scans of 1 1 mm usually provide sufficient resolution and sufficient data per image. For DNA–protein complexes that individually cover more surface, scans of 2 2 mm or 4 4 mm are better (see Note 8). In any case, all images to be used in the same analysis need to be of the same size and resolution. 3. Images should be collected from nonoverlapping fields without selection for areas of interest. If scanning is stable and interference free, images can be collected without changing the scanning parameters; the autoscan function of the microscope software can be used to collect an unbiased series.
12
Sample Preparation for SFM Imaging of DNA, Proteins. . .
225
4. Most analyses require a significant number of molecules or complexes to measure. For example, if 100 or more DNA–protein complexes have to be analyzed and the reaction results in 1/5 of the DNA molecules being bound by protein and deposition results 10 DNA molecules of about 1 kpb long in a 2 2 mm field, then a minimum of 50 such images need to be collected. It will take considerable time to collect these 50 images. In the optimistic case that everything works well, this is a half-day of collecting images. Taking into account that images are needed for control samples, planning is very important to perform efficient imaging experiments. 3.10. Guidelines for Standard Image Analysis
A number of features of protein, DNA, and protein–DNA complexes can be measured from SFM images (1). The size of proteins alone or bound to DNA can be accurately determined from SFM images. Also, protein-induced changes in DNA, such as wrapping, elongating, and bending, can be quantitatively described. DNA length can be measured from SFM images imported into specialized software such as IMAGE SXM. This is a customized version of NIH Image modified to run on a number of operating systems and to import image data automatically from a variety of commercial scanning probe microscopes. 1. We determine DNA contours by manually tracing in an appropriate image analysis program. In the case of DNA–protein complexes, contour length is traced as the shortest possible DNA path through the bound protein. Custom software that can automate DNA length measurements has been developed in several laboratories but is not currently commercially available. The length of DNA protein will indicate whether bound protein alters DNA by wrapping or stretching (see, for example, ref. 10). 2. The volume of proteins (not bound to DNA) can be determined using a semi-automated method developed by Glenn C. Ratcliff and Dorothy A. Erie (11). Using this approach, large protein populations can be analyzed in a rapid and accurate manner. 3. The volume of DNA-bound protein complexes has to be determined by manual tracing. Area and average height of a complex are measured, and a background volume of the same traced area at an adjacent position including DNA is subtracted (12, 13). 4. DNA bending is a feature of many DNA-binding proteins (see, for example, refs. 3, 14). Using SFM, DNA bending can be directly evaluated. A comparison of different methods to determine DNA bend angles is presented in (15).
226
D. Ristic et al.
3.11. Preparation of Mica Substrates for Combined SFMFluorescence
1. Cut the mica disk slightly smaller than the size of the coverslips. 2. Cleave the mica disk with tape until almost transparent. The color at this stage should still be slightly brownish. Be sure that one side of the mica surface is clearly flat (by eye) with no irregularities (as described in Subheading 3.3). The flat side of the mica is to be glued to the coverslip. 3. Place the coverslip to be used onto a lens cleaning tissue (Whatman 105). 4. Put a small drop of the optical glue in the end of a yellow-tip pipette. Spread slightly on the center of a round coverslip (the amount of glue could be enough for 10 coverslips). 5. Place the mica disk over the glue with the flat side facing the glue. 6. Tape the mica and the coverslip and attach to the lens cleaning tissue as shown in Fig. 4. Press down on the middle of the mica with your thumb for a homogeneous distribution of the glue under the surface. 7. Cure the UV glue by placing the UV light around 4 cm over the coverslips and switch on the 350-nm light for 3 min (see Note 9). It is most convenient to keep the glass-glued mica taped to the lens tissue (Fig. 4). A piece of lens tissue including the taped down coverslip is cut to about the size of a microscope slide. Accidentally breaking the coverslip is
Fig. 4. Illustration of preparing mica glued to glass coverslips. The mica is glued to a glass coverslip as described and then attached with tape to a lens cleaning tissue as shown. The tissue is used to pick up the fragile glass coverslip and the sample surfaces are also stored in this way.
12
Sample Preparation for SFM Imaging of DNA, Proteins. . .
227
avoided by handling the tissue. The mica–coverslips can be stored for future use for as long as necessary. 8. Before using the mica–coverslip, the mica surface should be made as thin as possible, by cleaving as many layers as possible using the scotch tape. The transparency of the mica now should be close to 100%, otherwise the focusing step with the short-distance working objectives would be simply impossible (see Note 10). 9. Once the thickness of the mica–coverslip has been assessed (see Note 10), this layer can be stripped off and a new one treated by placing a 200 ml drop of 0.25% w/v sodium tetrahydridoborate on the mica for 20 min to reduce auto-fluorescence. Wash with water as shown in Fig. 1. The surface is now ready for depositing the solution containing the DNA and protein complex to be analyzed. 3.12. Fluorescent DNA and Proteins for Combined SFMFluorescence: General Considerations
3.13. Sample Preparation for Combined SFM-Fluorescence
General considerations for DNA and proteins: Fluorescence labels can be attached to DNA and/or protein. DNA is typically labeled (1) by PCR using fluorescence nucleotide analogs for uniform labeling if the fluorophore does not interfere with protein interactions; (2) with PCR primers including a 50 fluorophore; and (3) by conjugating biotin to DNA, also introduced by PCR with 50 biotin-modified primers, bound to any of a variety of streptavidincoupled fluorophores. Protein can be labeled with fluorophores by a variety of methods such as those described in this issue in the chapter by Modesti. 1. Prepare a DNA, Protein, or DNA–protein complex binding reaction in appropriate functional conditions. 2. Dilute the binding reaction into 20 ml of deposition buffer (10 mM HEPES and 10 mM MgCl2, including 3 pM FluoSpheres® markers) according to the guidelines given above in Subheadings 3.4–3.8. 3. Place a 20–40 ml drop of the diluted sample onto the freshly cleaved mica surface and let it sit for 1 min. 4. Wash off the unbound material with water, about 1 pasture pipette full, as shown in Fig. 1. 5. Blot off excess liquid by touching a tissue to the edge of the mica. 6. Dry the sample in a stream of filtered air.
228
D. Ristic et al.
3.14. General Considerations for Imaging with Combined SFM-Fluorescence
3.15. Using FluoSpheres® to Align Topographic and Optical Images (Fig. 5)
Familiarity with the available fluorescence microscope setup is assumed. Specific operating instructions will vary depending on the setup and are in any case beyond the scope of this article. With respect to excitation light source, we note that a mercury lamp is often sufficient for visualization of quantum dots and FluoSpheres®, and to a limited extent for single fluorophores such as Alexa 633. Laser excitation specific to the dyes used is preferable for most single fluorophore visualization applications. In our setup, the optical microscope is coupled to a NanoWizard®II scanner (JPK instruments). In this instrument, correlation of fluorescence and topographic images is accomplished first by DirectOverlay™ software. Optical images are obtained at the highest magnification possible, usually as 60-mm fields. Topographic images are obtained as fields of 2–10 mm depending on the size of the objects deposited and the eventual analysis needed. 1. Obtain an optical image with at least 3 FluoSpheres®. The FluoSpheres® attach to mica together with the biomolecules during standard deposition and fall at random on the surface (see Note 11).
Fig. 5. Aligning optical and topographic images using FluoSpheres®. A mixture of polystyrene spheres with three different colors, were deposited for combined SFM and fluorescence imaging, as described. The samples were imaged in tapping mode in air with a JPK NanoWizard scanner mounted on a Nikon TE2000 microscope. The density of FluoSpheres® shown is appropriate for both optical and topographic imaging. (a) Optical image, 60 60 mm, of a mixture of polystyrene spheres with three different colors: green and red channels were overlaid. Polystyrene beads are recognized by the co-localization of both signals. The indicated area (5 5 mM, square) was chosen for scanning force microscopy. (b) Overlay of the fluorescence signal with the height image. (c) Topography image of the selected area, Z scale 0–30 nm. The overlay in b shows that the optical and topographic images can be aligned by centering the height image of the FluoSpheres® in the optical signals from the same objects. Panels (b) and (c) present the area scanned rotated about 135 clockwise relative to its position in panel (a).
12
Sample Preparation for SFM Imaging of DNA, Proteins. . .
229
2. Obtain a topographic scan of the same area at a resolution that will allow clear identification of the 20–40-nm diameter FluoSpheres®. 3. Use the microscope software to overlay the two images so that the topographic image of the spheres (always smaller than the diffraction-limited fluorescent spot) is aligned with the center of the optical spots. This overlay now defines the register of the fluorescent and topographic images to nanometer accuracy and can be used to identify fluorescent signals as belonging to specific topographic features (such as specific proteins or fluorescently labeled sites on DNA).
4. Notes 1. Here, we describe the use of SFM imaging of air-dried samples with the following standard settings: scanned in intermittent contact mode (air), using Silicon Tapping/Non-Contact Mode tips 125 mm in length with a spring constant of 25–75 N/m from Applied Nanostructures, drive frequency of the cantilevers on average 300 kHz, and a line rate of 2 Hz to acquire images. 2. Tetrahydridoborate is a flammable solid and should be handled according to the precautions indicated by the supplier. The solution should be made in a fume hood. 3. Because deposition is not always perfect but easy to repeat, it is a good idea to make several depositions of the same sample. When working with protein–DNA binding reactions, depositions should be done at the same time to keep the binding conditions, such as incubation time, constant. It is also a good idea to deposit different dilutions, differing by a factor of 2–5, at once to assure that one will be appropriate for analysis without requiring repetitive binding reactions and additional material. 4. After dilution, the sample should be transferred to mica as fast as possible to avoid changes in binding behavior due to changed salt conditions. Typically, adding a small volume of binding reaction to a premeasured volume of deposition buffer for dilution and immediately pipetting onto mica take less than 20 s. 5. The optimal buffer conditions, stoichiometry of DNA and protein, and the expected number of complexes are best extracted from previous biochemical characterizations. As mentioned above, these will indicate starting conditions as the amounts and concentration of DNA and proteins that
230
D. Ristic et al.
are optimal for SFM imaging may be different. Based on the initial SFM results, it may be necessary to vary the stoichiometry of DNA and protein to obtain sufficient protein–DNA complexes for analysis or to minimize background of unbound proteins. This is best done in small steps, such as twofold changes in concentration of proteins or DNA, in order to improve the image data that can be obtained. 6. If enough material is not deposited on mica, small variations in the dilution step that may effect cation concentrations by twofold or less can also make a big difference. 7. The sharpness of the tips determines the resolution and detail of the images obtained. Although specialized tips with end radius of curvature considerably less than the usual 10 nm (some as small as 2 nm) are available, they are still rather expensive. In our experience, for most applications, the added resolution and detail are minimal and do not justify the extra cost. We use uncoated tips. Even though coatings that increase reflectivity of the back of the cantilevers should not influence the size of the end of the tips, in our experience coated tips produced poorer resolution. 8. Many commercial SFMs produce images with maximally 512 512 pixels, independent of the scan size. Thus, larger scan sizes produce data with lower resolution. Clearly, higher resolution is achieved with smaller scan size (scan of 1 1 mm, 1 pixel ~2 nm vs. scan of 4 4 mm, 1 pixel ~8 nm). Due to the size of the scanning tips, we typically use standard silicon noncontact tips with a radius of curvature of about 10 nm, resolution does not improve much by decreasing scan size below 1 1 mm. 9. When using a UV light, wear UV safety glasses and avoid skin exposure. 10. It is wise, before going further in the sample preparation, to check if the mica glued to glass thickness is appropriate. Deposit the fluorescent test object such as the FluoSpheres® (3 pM solution in deposition buffer) and check if it is possible to focus on the surface in the fluorescent microscope. If it is not possible, more mica layers need to be cleaved off. 11. We have also tested quantum dots as markers for aligning fluorescence and topographic images. However, due to the relatively large percentage of dark quantum dots in the preparations we have used, it is not easy to align the patterns from the topographic and fluorescence image unambiguously. Thus, in our experience, quantum dots are not a robust marker for alignment.
12
Sample Preparation for SFM Imaging of DNA, Proteins. . .
231
References 1. Janicijevic, A., Ristic, D. and Wyman, C. (2003) The molecular machines of DNA repair: scanning force microscopy analysis of their architecture. J Microsc. 212, 264–72. 2. Dame, R.T., Wyman, C. and Goosen, N. (2003) Insights into the regulation of transcription by scanning force microscopy. J Microsc. 212, 244–53. 3. Erie, D.A., Yang, G., Schultz, H. C. and Bustamante, C. (1994) DNA bending by Cro protein in specific and nonspecific complexes: implications for protein site recognition and specificity. Science 266, 1562–6. 4. Ristic, D., Modesti, M., van der Heijden, T., van Noort, J., Dekker, C., Kanaar, R. and Wyman, C. (2005) Human Rad51 filaments on doubleand single-stranded DNA: correlating regular and irregular forms with recombination function. Nucleic Acids Res. 33, 3292–302. 5. Bustamante, C., Vesenka, J., Tang, C. L., Rees, W., Guthold, M. and Keller, R. (1992) Circular DNA molecules imaged in air by scanning force microscopy. Biochemistry 31, 22–6. 6. Vesenka, J., Guthold, M., Tang, C. L., Keller, D., Delaine, E. and Bustamante, C. (1992) Substrate preparation for reliable imaging of DNA molecules with the scanning force microscope. Ultramicroscopy 42–44, 1243–9. 7. Hansma, H.G. and Laney, D.E. (1996) DNA binding to mica correlates with cationic radius: assay by atomic force microscopy. Biophys J. 70, 1933–9. 8. Han, W., Lindsay, S. M., Dlakic, M. and Harrington, R. E. (1997) Kinked DNA. Nature 386, 563.
9. Rivetti, C., Guthold, M. and Bustamante, C. (1996) Scanning force microscopy of DNA deposited onto mica: equilibration versus kinetic trapping studied by statistical polymer chain analysis. J Mol Biol. 264, 919–32. 10. Beerens, N., Hoeijmakers, J. H., Kanaar, R., Vermeulen, W. and Wyman, C. (2005) The CSB protein actively wraps DNA. J Biol Chem. 280, 4722–9. 11. Ratcliff, G.C. and Erie, D.A. (2001) A novel single-molecule study to determine protein– protein association constants. J Am Chem Soc. 123, 5632–5. 12. Wyman, C., Rombel, I., North, A. K., Bustamante, C. and Kustu, S. (1997) Unusual oligomerization required for activity of NtrC, a bacterial enhancer-binding protein. Science 275, 1658–61. 13. van der Linden, E., Sanchez, H., Kinoshita, E., Kanaar, R. and Wyman, C. (2009) RAD50 and NBS1 form a stable complex functional in DNA binding and tethering. Nucleic Acids Res. 37, 1580–8. 14. Janicijevic, A., Sugasawa, K., Shimizu, Y., Hanaoka, F., Wijgers, N., Djurica, M., Hoeijmakers, J. H. and Wyman, C. (2003) DNA bending by the human damage recognition complex XPC-HR23B. DNA Repair (Amst). 2, 325–36. 15. Dame, R.T., van Mameren, J., Luijsterburg, M. S., Mysiak, M. E., Janicijevic, A., Pazdzior, G., van der Vliet, P. C., Wyman, C. and Wuite, G. J. (2005) Analysis of scanning force microscopy images of protein-induced DNA bending using simulations. Nucleic Acids Res. 33, e68.
.
Chapter 13 Single-Molecule Protein Unfolding and Refolding Using Atomic Force Microscopy Thomas Bornschlo¨gl and Matthias Rief Abstract Over the past few years, atomic force microscopy (AFM) became a prominent tool to study the mechanical properties of proteins and protein interactions on a single-molecule level. AFM together with other mechanical, single-molecule manipulating techniques (Bustamante et al., Nat Rev Mol Cell Biol 1:130–136, 2000) made it possible to probe biological molecules in a way that is complementary to single-molecule methods using chemicals or temperature as a denaturant (Borgia et al., Annu Rev Biochem 77:101–125, 2008). For example, AFM offered new insights into the process of protein folding and unfolding by probing single proteins with mechanical forces. Since many proteins fulfill mechanical function or are exerted to mechanical forces in their natural environment, AFM allows to target physiologically relevant questions. Although the number of proteins unfolded with AFM continually increases (Linke and Grutzner, Pflugers Arch 456:101–115, 2008; Zhuang and Rief, Curr Opin Struct Biol 13:88–97, 2003; Clausen-Schaumann et al., Curr Opin Chem Biol 4:524–530, 2000; Rounsevell et al., Methods 34:100–111, 2004), the total number of proteins studied so far is still relatively small (Oberhauser and Carrion-Vazquez, J Biol Chem 283:6617–6621, 2008). This chapter aims at giving protocol-like instructions for people who are actually getting started using AFM to study mechanical protein unfolding or refolding. The instruction includes different approaches to produce polyproteins or modular protein chains which are commonly used to screen for true singlemolecule AFM data traces. Also, the basic principles for data collection with AFM and the basic data analysis methods are explained. For people who want to study proteins that unfold at small forces or for people who want to study protein folding which also occurs typically at small forces (25 mM. Thus, polymerization will already occur during the steps for expression and purification so that no further incubation is needed. If you observe sedimentation in your protein solution due to aggregation, spin down aggregates at 10,000 g for >10 min and use the supernatant for the AFM experiments. 3.1.3. Construct Polyproteins Using “Cysteine Engineering”
1. Use your T7 promoter-based E. coli expression vector containing a His6 tag sequence. 2. Insert your protein of interest into the multiple cloning site of the vector (See Fig. 1g). 3. Choose two residues within the amino acid sequence of your protein of interest and change them into cysteines using the Phusion site-directed mutagenesis kit. These cysteines will later on define the linkage geometry. Therefore, it is critical that they are solvent accessible. Use the protein structure if it is available to determine these linkage points; if not available, try the N- and C-termini as linkage points because they are often solvent accessible. Choosing residues that are located at opposite sites of the protein structure might decrease the hindrance for polymerization due to sterical restrictions. Also, check if your protein already contains solvent-accessible cysteines, and either exchange them with, e.g., alanine or serine or use them as linkage points. For troubleshooting and further details, see the protocol (3). 4. Continue with protein expression and purification (steps 7–14 of Subheading 3.1.1). 5. Concentrate the purified protein in the lysis buffer solution to more than 0.2 mM using, e.g., Centricon centrifugal filter units. 6. For polymerization, incubate the protein solution for ca. 80 h at 37 C if this temperature does not affect your protein. Else, incubate at lower temperatures for longer times. Check the progress of the polymerization using, e.g., SDS-Page with nonreducing SDS buffer. When the wanted degree of polymerization is reached (average of octamers is convenient for AFM measurements), dilute the protein solution to decrease further polymerization (e.g., dilute to 0.02 mM).
240
T. Bornschlo¨gl and M. Rief
7. Continue with the AFM experiments. If you observe sedimentation in your protein solution due to aggregation, spin down aggregates at 10,000 g for >10 min and use the supernatant for the AFM experiments. 3.2. Principal Data Collection and Analysis Using AFM
3.2.1. Experimental Setup and Data Collection
The development and availability of relatively easy-to-use commercial atomic force microscopes allow the application of this technique by researchers who are interested in single-molecule protein folding and unfolding, and who do not necessarily have to be AFM specialists. Most of the commercially available instruments also include automated procedures, e.g., in order to calibrate the force constants of the used cantilevers. For details on how to use your specific AFM, the relevant manual will provide much more useful information than this chapter is able to. Therefore, we will only shortly explain the basic principles of AFM measurements and provide relevant references for further reading. Due to the unspecific binding of proteins between cantilever tip and surface, the collected data will consist of a huge fraction of non-interpretable traces as, for example, traces where many molecules have bound in parallel. The second part of this chapter, therefore, provides a protocol on how to select relevant and interpretable single-molecule data when the modular polyprotein chains from Subheading 3.1 are used for the measurements. This section also explains the principal steps to determine the length increase due to protein unfolding. 1. Apply ca. 10 ml of the protein suspension to a clean glass coverslip or a freshly gold-evaporated glass surface and incubate for 20 min. To avoid air bubbles when you apply a drop (ca. 30 ml) of PBS onto the cantilever (Bio Lever Type A or B, Olympus), use degassed PBS. Join both drops and align the AFM (as shown in Fig. 2a), where a beam of light that is focused on the tip of the cantilever’s back side is reflected onto the middle of a 2-segment photodiode. To determine the bending of the cantilever, read out the deflection D(t) by reading the intensities I measured on the two photodiodes A and B. (D(t) ¼ (IA IB)/(IA + IB)). 2. If not already available, install a piezo control (Physik Instrumente) that allows, e.g., to move the surface back and forth over distances of several micrometer at constant velocity. Applicable velocities for protein unfolding experiments range between 1 nm/s and 10,000 nm/s (some of the limiting factors are instrument drift for slower pulling speeds and hydrodynamic effects for the upper border). 3. Gently approach the cantilver tip toward the surface. You can do so by observing the hysteresis in the deflection signal that originates from the hydrodynamic drag due to a repetitive,
13
Single-Molecule Protein Unfolding and Refolding Using Atomic Force Microscopy
241
Fig. 2. Experimental setup. (a) Schematic drawing of the AFM setup. (b) By applying a triangular voltage signal to the piezo-actuator, the surface should move toward the cantilever and then away from the cantilever at constant velocity. (c) Deflection signal when no protein has been attached between cantilever tip and surface after contact.
forward, and backward movement at constant velocities (~5 mm/s) of the surface. This hysteresis will increase with decreasing cantilever–surface distance. When close to the surface, apply a triangular voltage signal to the piezo-actuator, leading to surface movement as shown in Fig. 2b. If you are close enough to the surface, you should gain a deflection trace as the one shown in Fig. 2c, which defines the position of the surface (slightly tilted vertical line that also defines the proportionality constant p) and the zero line for the force acting on the cantilever (horizontal line). 4. Check your instrument drift. The deflection should stay constant when the surface is not moved. Particles swimming through the optical path of the light beam might lead to slow variations in the signal, while cantilever drift caused, e.g., by the bimetal effect might lead to a constantly creeping drift. Check also if the absolute distance of cantilever to surface stays constant over time. Without drift the curve shown in Fig. 2c should be exactly reproducible after some time, and forward and backward traces at slow velocities 50 pN), while they refold at forces in the lower pN regime where the resolution of commercially available cantilevers can become the limiting factor. It is demanding but possible to observe the refolding of such proteins directly (20, 21). Another approach that has been already used to deduce the folding properties of a protein is to retract the cantilever toward the surface after the protein has been successfully unfolded. After a certain time during which the protein is allowed to refold, one can check if folding had occurred by stretching the whole protein again (22) (17). However, AFM is perfectly suited to study protein unfolding and refolding even at very small forces if this process occurs close to thermodynamic equilibrium. In this case, either many folding/unfolding transitions can be observed during one stretching cycle of one molecule (23) or many stretching and relaxing cycles with highly reproducible force extension pattern can be obtained on the same molecule (5, 24). This allows to separate the signal easily from unspecific background such as drift. It should be noted that the use of an optical trap comprising of a higher force resolution is an alternative for the study of proteins in this force regime, especially if constant force experiments are wanted (25). But the AFM with its higher force constants can provide insights into this low force regime that might be complementary to those measured with optical trapping. The following subchapter provides a protocol for how the force
13
Single-Molecule Protein Unfolding and Refolding Using Atomic Force Microscopy
245
resolution of an AFM in the low force regime can be enhanced by recording many folding/unfolding cycles and by averaging them. 3.3.1. Protein Design and First Steps
1. Use a naturally occurring modular protein or an artificially designed modular protein as matrix protein and insert only once your protein of interest (see Subheading 3.1.1). The used matrix protein should have been already probed with AFM and should be known to be stable enough to not interfere with the unfolding of your protein of interest. Subheading 3.1.1 facilitates the screening for true single-molecule events compared to a polyprotein approach (Subheadings 3.1.2 and 3.1.3) because you expect that the unfolding forces of your protein of interest are small. Moreover, the unfolding/folding behavior of one single protein might be already rather complex so that using the polyprotein approach might complicate the data analysis. 2. Start with the standard measurement and data analysis (Subheadings 3.2.1 and 3.2.2) at slow velocities (~100 nm/s). This will prove successful protein design and purification, and it may give you a first idea of the unfolding behavior of your protein. Since your protein will unfold in the beginning of the force curve where often unspecific interactions occur (e.g., gray effect in Fig. 3a before ~75 nm), data collection using Subheading 3.2.2 might be demanding.
Fig. 3. High resolution experiments for proteins that unfold and refold close to thermodynamic equilibrium. (a) Example curve to study coiled-coil unzipping (black at forces