- Email: [email protected]
Imaging of Single Fluorescent Molecules Using Video-Rate Confocal Microscopy
Biochemical and Biophysical Research Communications 287, 323–327 (2001) doi:10.1006/bbrc.2001.5574, available online at http://www.idealibrary.com on
Imaging of Single Fluorescent Molecules Using Video-Rate Confocal Microscopy Hisashi Tadakuma,* Junichi Yamaguchi,* Yo Ishihama,* and Takashi Funatsu* ,† ,1 *Department of Physics, School of Science and Engineering, Waseda University, 3-4-1 Okubo, Tokyo 169-8555, Japan; and †Precursory Research for Embryonic Science and Technology, Japan Science and Technology Corporation, Kawaguchi Center Building 1-8, Honcho 4-Chome, Kawaguchi-city, Saitama 332-0012, Japan
Received August 15, 2001
Single fluorescent molecules in aqueous solution were imaged for the first time at video-rate using Nipkow disk-type confocal microscopy. Performance of this method was evaluated by imaging single kinesin molecules labeled with fluorescent dyes of tetramethylrhodamine (TMR) or IC5. Photodecomposition lifetimes of the fluorophores were ⬃10 s for TMR and ⬃2 s for IC5 under the incident laser power of 0.5 W/mm 2. Both the fluorescence intensity and the photobleaching rate were proportional to the laser power from 0.65 to 3 W/mm 2. 2D sliding movement of single kinesin molecules along microtubules on glass surface and 3D Brownian motion of individual kinesin molecules in viscous solution could be observed using this microscopy. These results indicated that this method could be applicable to the study of single molecular events in living cells at real time. © 2001 Academic Press Key Words: single-molecule imaging; confocal microscopy; fluorescence microscopy; Brownian motion.
Single-molecule biophysics has recently advanced and opened up a new era in the field of life sciences (1, 2). Imaging and manipulating single biomolecules enabled us to study dynamics of elementary process, and to obtain invaluable information that was not extractable through the conventional ensemble-averaged experiments. Following the success in visualizing a single fluorophore on an air-dried surface (3), a single fluorophore attached to a protein molecule in aqueous solution is first observed by using total internal reflection fluorescence microscopy (TIRFM) (4). Application of single-molecule imaging in vitro includes translational (5) and rotational (6) motion of single motor proteins, Abbreviations used: TIRFM, total internal reflection fluorescence microscopy; TMR, tetramethylrhodamine; NA, numerical aperture; MSD, mean square displacement. 1 To whom correspondence and reprint requests should be addressed. Fax: ⫹81-3-5286-3863. E-mail: [email protected].
chemical reaction in a single enzyme molecule (4, 7), and simultaneous measurement of chemical and mechanical events in a motor protein (8). Single-molecule imaging has recently been applicable to the study of living cells. Ligand-receptor interaction and oligomerization of adhesion proteins are visualized in living cells (9, 10). TIRFM is widely used for single-molecule imaging to reduce the background fluorescence by means of the local illumination of the evanescent field. For living cells, however, this method is not generally applicable since illumination using TIRFM is confined to extremely thin interfacial layers around cell membrane. To overcome these issues, we used a video-rate confocal microscopy (CSU10, Yokogawa Electric Co., Japan) to visualize single fluorescent molecules. CSU10 has two Nipkow disks, and light incident on 20,000 microlenses in the first disk is focused on corresponding pinholes on the second disk (11). This configuration increases transmission rate of the incident light to 50% while the conventional single Nipkow-disk scanner transmits only 2–3%. These disks rotate at 1800 rpm and a frame of image can be obtained every 2 ms. Owing to this video-rate confocal microscopy, 2D sliding movement of single TMR-kinesin molecules along microtubules and 3D Brownian motion of TMR-kinesin molecules in viscous solution could be observed. The photodecomposition lifetime of the single fluorophore up to 10 s was long enough for observing some biological phenomena. Thus, this study provided a basis for the single-molecule imaging in living cells. MATERIALS AND METHODS Proteins. Kinesin was prepared from bovine brain (12) and labeled with tetramethylrhodamine (TMR)-5-maleimide (Molecular Probes, U.S.A.) or IC5-maleimide (Dojindo, Japan) with dye to protein ratio of 0.6 or 1.0, respectively. Tubulin was prepared from porcine brain and labeled with Cy5 (Amersham–Pharmacia, Japan) according to the method of Hyman (13). Cy5 labeled tubulin was polymerized in the assembly buffer (80 mM Pipes, pH 6.8, 5 mM
323
0006-291X/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
Vol. 287, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
MgCl 2, 1 mM EGTA, 1 mM GTP, and 33% glycerol) at 37°C, and stabilized by adding 4 M Taxol. Optics. Nipkow disk-type confocal unit (CSU10, Yokogawa Electric Co., Japan) was installed to an inverted microscope (IX-70, Olympus Optical Co., Ltd., Japan). TMR was illuminated using a green solid-state laser (50 mW, 532 nm, -Green Model 4601, Uniphase, U.S.A.) and IC5 with a He–Ne laser (20 mW, 633 nm, GLG5601, NEC, Japan). A laser beam was depolarized by passage through a 1/4 plate and introduced into CSU10 without an optical fiber. It was focused at the inlet of CSU10 with numerical aperture (NA) of 0.01. The laser beam illuminated a specimen with diameter of up to 40 m by using oil-immersion objective (PlanApo 100⫻ NA ⫽ 1.4, Olympus). Dichroic mirrors and emission filters used for TMR and IC5 imaging were as follows; a dichroic mirror of DM560 (Asahi Spectrum Co., Japan) and emission filters of EM560 (Asahi Spectrum Co.) and 605DF80 (Omega Optical Inc., U.S.A.) for TMR imaging, and a dichroic mirror of DM650 (Asahi Spectrum Co.) and an emission filer of 680DF55 (Omega Optical Inc.) for IC5 imaging. Fluorescent image was taken by an ICCD camera (ICCD-350F, Video Scope International, U.S.A.) coupled to an image intensifier (VS41845, Video Scope International). Video signal was enhanced with an image processor (Argus20, Hamamatsu Photonics, Japan), and recorded on S-VHS format videotapes. Images were analyzed using a program written specifically for this analysis on a Halcon image processor (MVTec Software GmbH, Germany) or Scion image (Scion Corp., U.S.A.). Fluorescence intensities were calculated by integrating signals from 9 ⫻ 9 pixels and subtracting background. Single-molecule imaging in 2D plane. Coverslips were washed with 0.1 M KOH, ethanol and then dried in a clean bench as previously described (4). TMR-kinesin or IC5-kinesin in Buffer-A (80 mM Pipes, pH 6.9, 2 mM MgCl 2, 1 mM EGTA) was immobilized on a coverslip surface in the presence of oxygen scavenger system (25 mM glucose, 2.5 M glucose oxidase, 10 nM catalase, 10 mM dithiothreitol) (14). The laser power density at the specimen plane was estimated by measuring the power of laser that passed the objective lens and the illumination area of the specimen. Fluorescence intensity was measured as described above. Movement of single kinesin molecules along microtubule was observed as follows. A flow cell was made from a glass slide and coverslip with two slivers of film 50 m thick. Cy5-microtubules in Buffer-A containing 40 M Taxol were infused into the flow cell to be adsorbed on the coverslip. The glass surface was then blocked with 1 mg/ml casein (No. 073-19, Nacalai Tesque, Inc., Japan) to prevent the nonspecific adsorption of kinesin. TMR-kinesin in Buffer-A containing 40 M Taxol, 1 mM ATP, 1 mg/ml casein, 10 mM dithiothreitol, 0.2% methyl cellulose and oxygen scavenger system, was infused into the flow cell. The final concentration of TMR-kinesin in the motility assay was 1 nM. The slivers of film in flow cell were removed, and then the flow cell was sealed with the nail enamel. Microtubule and kinesin were observed using confocal microscopy at 23°C. The position of the fluorophore was determined using the homemade centroid calculating macro on Scion Image. Velocity was determined by dividing the run length by run time (the duration while kinesin moved on microtubule). Single-molecule imaging in 3D space. Solution containing 5 nM kinesin, 15% acrylamide, 5% bisacrylamide and 10% ammonium persulfate was mixed with N,N,N⬘,N⬘-tetramethylethylenediamide, infused into a flow cell, and acrylamide gels were polymerized for 1 h. Confocal images at various depths from the glass interface were obtained, and fluorescence intensities of the individual molecules were analyzed. To evaluate the aberration of objective, fluorescent microbeads in 0.1 m diameter (Molecular Probes, U.S.A.), instead of TMR-kinesin, were added to polyacrylamide gels. Series of optical sections were obtained at every 0.1-m interval, and fluorescence intensities of individual microbeads were analyzed. For single-molecule imaging of 3D Brownian motion of TMRkinesin, Buffer-A containing 10 nM TMR-kinesin, 0.1 mg/ml ␣-casein and 60% (w/w) sucrose was infused into the flow cell. The viscosity
FIG. 1. Single-molecule imaging of fluorescent kinesin molecules using Nipkow disk-type confocal microscopy. (A) Fluorescence micrograph image of TMR-kinesin molecules attached to the glass surface. The image was integrated over eight video frames. Scale bar, 2 m. (B) Quantized photobleaching of the TMR molecule indicated by the arrow in A. The TMR molecule was photobleached at the time indicated by the arrowhead. (C) Distribution of the fluorescence intensities of TMR-kinesin molecules that were bleached with single (open columns) and two (closed columns) steps. The power of the incident laser was 1.9 W/mm 2. Solid lines indicated Gaussian functions fitting to the data. Arrowheads indicated average fluorescence intensities. (D) Fluorescent intensity of the TMR-kinesin molecules as a function of laser light intensity. The bars indicated standard deviation. The data were fitted by a linear line function. (E) Time course of the decrease in number of TMR molecules due to photobleaching. TMRkinesin molecules were illuminated by a green laser at the light intensity of 1.9 W/mm 2. The data were fitted by a single exponential function with photodecomposition lifetime of 3.1 s. (F) Rate constants of the photobleaching of TMR-kinesin (closed circles) and IC5kinesin (open circles) molecules as a function of laser light intensity.
coefficient of 60% sucrose solution was 53.2 mPa 䡠 s. Diffusional motion of kinesin was observed at a 2-m distance from the glass surface. First, diffusion constant in a horizontal direction (D x ⫺y ) was determined based on Einstein’s equation:
4D x⫺y ⌬t ⫽ 具⌬x 2 ⫹ ⌬y 2 典
[1]
where, ⌬t was time interval and 具⌬x 2 ⫹ ⌬y 2 典 was two-dimensional mean square displacement (MSD). Next, diffusion constant in vertical direction (D z ) was determined using the following equation (15):
324
Vol. 287, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Sequential images of single TMR-kinesin molecule sliding on a microtubule at 23°C. (A) A fluorescence micrograph image of IC5-labeled microtubule. (B) A series of fluorescent micrograph images of TMR-kinesin taken every 1 s without frame averaging. Arrowheads indicated kinesin moving along microtubule. Numerals represented the time in seconds. Scale bar, 2 m.
12D z ⌬t ⫽ 具 ␦ z 2 典,
(Fig. 1B). Figure 1C showed the distributions of the fluorescence intensity of the individual spots that were photobleached in a single step (open columns) or two steps (close columns). The distribution was well fitted by two Gaussian distributions. The mean intensity of the fluorescent spots that were bleached in two steps was as twice strong as the value of those bleached in single step. These results indicated that single fluorescent molecules could be seen using confocal microscopy. A well-known problem of the fluorescence imaging with the confocal microscopy is a saturation of the fluorescence intensity of the fluorophore. We thus examined the dependence of the fluorescence intensity of the TMR-kinesin on the excitation laser intensity. We found a linear relationship between the laser power and the fluorescence intensity in the power range from 0.65 to 3 W/mm 2 (Fig. 1D). In our optical configuration,
[2]
where, ␦ t was the resident time of TMR-kinesin in a confocal cross section and ␦ z was the depth of the confocal optical section.
RESULTS AND DISCUSSION 2D Imaging of Single Fluorescent Molecules CSU10, a video-rate confocal unit using Nipkow disk, was installed to an inverted microscope (see details described under Materials and Methods). Laser beam was directly introduced into CSU10 without using a single mode optical fiber with NA 0.11, because high laser power density required for single-molecule imaging could not be obtained in default configuration. Inhomogeneity of illumination due to interference was less than 10% in a specimen plane. To evaluate the performance of single-molecule imaging using this confocal microscopy, kinesin was labeled with fluorescent dye of TMR and attached to the coverslip, then illuminated with green laser of 1.9 W/mm 2 at the specimen plane (Fig. 1). Through several approaches, it was confirmed that individual fluorescent molecules could be seen using this method. First, the number of fluorescent spots depended on kinesin concentration. The number of fluorescence spots was proportional to the applied kinesin concentration except when kinesin was extremely diluted and adsorption during preparation was not negligible. We found a very small number of light spots (less than 0.001 spots per square micrometer), presumably dusts, in the absence of kinesin. Effects of these dusts were negligible in the analysis. The time course of the fluorescence intensity of the individual spots showed quantized photobleaching
FIG. 3. Changes in fluorescence intensities with increasing penetration into solution. (A) Relative fluorescence intensities of single TMR-kinesin molecules embedded in a polyacrylamide gel were measured at 0, 2, 5, and 10 m away from the glass-solution interface. The bars indicated standard deviations. N ⫽ 76 ⬃ 116. The data was fitted by a linear line function. (B) Relative fluorescence intensities of microbeads as a function of the distance from the glass solution. Fluorescent microbeads with diameter of 0.1 m were embedded in a polyacrylamide gel, and fluorescence micrograph images were obtained every 0.1 m. Open circles indicated the average of relative fluorescence intensities of microbeads (N ⫽ 10) located at 0, 2, 5, and 9 m from the glass solution interface as shown by arrows.
325
Vol. 287, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Diffusion of TMR-kinesin molecules in 60% sucrose solution. (A) Sequential images of a single TMR-kinesin molecule taken at video-rate. Numerals represented the time in milliseconds. Scale bar, 2 m. Measurement was performed at a temperature of 23°C. (B) 2D mean square displacement as a function of time. Bars indicated standard errors of the data. (C) Duration of TMR-kinesin observed during passing through the confocal plane. Arrowhead indicates the average duration (0.13 s).
CSU10 scanned the illumination area with 256 pinholes, and we observed a single fluorophore at an illumination intensity of 12 W per pinhole when laser power density was 1.9 W/mm 2 at the specimen plane. Using a single-beam laser, the fluorescence intensity of the fluorophore is reported to saturate at about 1 mW (16). At our setting, we observed a single molecule at about 1/100 laser intensity of the saturation power. This observation might account for the linear dependency of the fluorescence intensity on the excitation laser intensity in our study. Figure 1E showed the photodecomposition of TMRkinesin under the illumination power of 1.9 W/mm 2. Photobleaching rates of TMR and IC5 were measured under the various laser power conditions. We found a linear relationship between laser power and photobleaching rate (Fig. 1F). As a demonstration of this microscopy for visualizing protein–protein interactions, sliding movement of TMR-kinesin molecules along microtubule was visualized. Microtubules labeled with Cy5 were attached to a coverslip of the chamber (Fig. 2A), and 1 nM of TMRkinesin was infused in the presence of ATP. The movement of fluorescent spots of kinesin was observed (Fig. 2B). The sliding velocity was 0.29 ⫾ 0.17 m/s (mean ⫾ SD, n ⫽ 7), which was in good agreement with the reported value for native kinesin movement on axoneme (17). Individual fluorescent spots of kinesin could be hardly observed when using conventional epifluorescence microscopy because of the background fluorescence of kinesin freely diffusing in solution.
3D Imaging of Immobilized Fluorescent Molecules Next we observed fluorescent molecules in 3D space. To evaluate the depth dependency of the fluorescence intensity, TMR-kinesin molecules were immobilized in polyacrylamide gels, and fluorescence images at various depths from the glass solution interface were obtained. Fluorescence intensities of kinesin molecules decreased due to aberration as the distance from the coverslip increased (Fig. 3A), and it became 50% at 10 m from the glass surface. To quantify the intensity and spread of the fluorescence, fluorescent microbeads in 0.1 m diameter were used instead of kinesin. Figure 3B showed the depth dependency of the fluorescence intensity relative to that of beads immobilized on coverslip. The half width of the fluorescence intensity was 0.53 ⫾ 0.1 m (mean ⫾ SD) at the glass surface and 1.05 ⫾ 0.1 m at 10 m from the cover glass surface. 3D Imaging of Brownian Motion of Single Fluorescent Molecules We observed single TMR-kinesin molecules undergoing Brownian motion in solution, and followed their trajectory while they were in confocal cross section at appropriate z position. To slow down the movement of kinesin, 60% sucrose was included in the solution. Figure 4A showed typical sequential video images of a single TMR-kinesin molecule moving at 2 m from the glass surface. The 2D MSD was calculated and plotted
326
Vol. 287, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
against time (Fig. 4B). By fitting a straight line to the data, diffusion constant in a horizontal direction (D x ⫺y ) of 0.75 m 2/s at ⫽ 53.2 mPa 䡠 s was obtained. The diffusion constant in a vertical direction (D z) was determined using the resident time of TMR-kinesin in a confocal cross section ( ␦ t) and the depth of the confocal plane ( ␦ z ⫽ 1.0 m) in Eq. [2]. The depth of the confocal plane ␦z was assumed to be 1.0 m, since the fluorescence intensity decreased to 20% of the maximum intensity when confocal plane was displaced ⫾ 0.5 m from the molecules, and individual fluorescent molecules could not be clearly observed (Fig. 3). The average resident time was 0.13 s for ⫽ 53.2 mPa 䡠 s. D z was calculated to be 0.64 m 2/s for ⫽ 53.2 mPa 䡠 s. The diffusion constants of both lateral and vertical diffusion were quite similar, and they were estimated to be 34 – 40 m 2/s in the absence of sucrose, which was in good agreement with the values measured in ensemble biochemical studies (18, 19). In conclusion, 2D and 3D movement of single molecules was successfully visualized using video-rate confocal microscopy. This method could be applied to the examination of single molecular events inside living cells. ACKNOWLEDGMENTS We thank Messrs. T. Tanaami, A. Ichihara (Yokogawa Electric Co.), and K. Abe (Olympus Co., Ltd.) for valuable comments and Drs. T. Tojo and T. Zako (Waseda University) for critical reading of the manuscript. This work was partly supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture of Japan and by Uehara Memorial Foundation.
REFERENCES 1. Harada, Y., Funatsu, T., Tokunaga, M., Saito, K., Higuchi, H., Ishii, Y., and Yanagida, T. (1998) Single molecule imaging and nanomanipulation of biomolecules. Methods Cell Biol. 55, 117– 128. 2. Weiss, S. (1999) Fluorescence spectroscopy of single biomolecules. Science 283, 1676 –1683. 3. Betzig, E., and Chichester, R. J. (1993) Single molecules observed by near-field scanning optical microscopy. Science 262, 1422–1425. 4. 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.
5. 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. 6. Adachi, K., Yasuda, R., Noji, H., Itoh, H., Harada, Y., Yoshida, M., and Kinosita, K., Jr. Stepping rotation of F1-ATPase visualized through angle-resolved single-fluorophore imaging. (2000) Proc. Natl. Acad. Sci. USA 97, 7243–7247. 7. Lu, H. P., Xun, L., and Xie, X. S. (1998) Single-molecule enzymatic dynamics. Science 282, 1877–1882. 8. Ishijima, A., Kojima, H., Funatsu, T., Tokunaga, M., Higuchi, H., Tanaka, H., and Yanagida, T. (1998) Simultaneous observation of individual ATPase and mechanical events by a single myosin molecule during interaction with actin. Cell 92, 161–171. 9. Sako, Y., Minoguchi, S., and Yanagida, T. (2000) Single-molecule imaging of EGFR signaling on the surface of living cells. Nat. Cell Biol. 2, 168 –172. 10. Iino, R., Koyama, I., and Kusumi, A. (2001) Single molecule imaging of green fluorescent proteins in living cells: E-cadherin forms oligomers on the free cell surface. Biophys. J. 80, 2667– 2677. 11. Ichihara, A., Tanaami, T., Isozaki, K., Sugiyama, Y., Kosugi, Y., Mikuriya, K., Abe, M., and Uemura, I. (1996) High-speed confocal fluorescence microscopy using a Nipkow scanner with microlenses for 3D imaging of single fluorescent molecule in real time. Bioimages 4, 57– 62. 12. Kojima, H., Muto, E., Higuchi, H., and Yanagida, T. (1997) Mechanics of single kinesin molecules measured by optical trapping nanometry. Biophys. J. 73, 2012–2022. 13. Hyman, A., Drechsel, D., Kellogg, D., Salser, S., Sawin, K., Steffen, P., Wordeman, L., and Mitchison, T. (1991) Preparation of modified tubulins. Methods Enzymol. 196, 478 – 485. 14. Harada, Y., Sakurada, K., Aoki, T., Thomas, D. D., and Yanagida, T. (1990) Mechanochemical coupling in actomyosin energy transduction studied by in vitro movement assay. J. Mol. Biol. 216, 49 – 68. 15. Berg, H. C. (1993) Diffusion to capture. In Random Walks in Biology, new, expanded edition, pp. 37– 47, Princeton Univ. Press. 16. Sandison, D. R., Williams, R. M., Wells, K. S., Sticker, J., and Webb, W. W. (1995) Quantitative fluorescence confocal laser scanning microscopy (CLSM). In Handbook of Biological Confocal Microscopy (Pawley, J. B., Ed.), 2nd ed., pp. 39 –53, Plenum Publishing Corp. 17. Friedman, D. S., and Vale, R. D. (1999) Single-molecule analysis of kinesin motility reveals regulation by the cargo-binding tail domain. Nat. Cell Biol. 1, 293–297. 18. Bloom, G. S., Wagner, M. C., Pfister, K. K., and Brady, S. T. (1988) Native structure and physical properties of bovine brain kinesin and identification of the ATP-binding subunit polypeptide. Biochemistry 27, 3409 –3416. 19. Hackney, D. D., Levitt, J. D., and Wagner, D. D. (1991) Characterization of alpha 2 beta2 and alpha 2 forms of kinesin. Biochem. Biophys. Res. Commun. 174, 810 – 815.
327
Imaging of Single Fluorescent Molecules Using Video-Rate Confocal Microscopy Hisashi Tadakuma,* Junichi Yamaguchi,* Yo Ishihama,* and Takashi Funatsu* ,† ,1 *Department of Physics, School of Science and Engineering, Waseda University, 3-4-1 Okubo, Tokyo 169-8555, Japan; and †Precursory Research for Embryonic Science and Technology, Japan Science and Technology Corporation, Kawaguchi Center Building 1-8, Honcho 4-Chome, Kawaguchi-city, Saitama 332-0012, Japan
Received August 15, 2001
Single fluorescent molecules in aqueous solution were imaged for the first time at video-rate using Nipkow disk-type confocal microscopy. Performance of this method was evaluated by imaging single kinesin molecules labeled with fluorescent dyes of tetramethylrhodamine (TMR) or IC5. Photodecomposition lifetimes of the fluorophores were ⬃10 s for TMR and ⬃2 s for IC5 under the incident laser power of 0.5 W/mm 2. Both the fluorescence intensity and the photobleaching rate were proportional to the laser power from 0.65 to 3 W/mm 2. 2D sliding movement of single kinesin molecules along microtubules on glass surface and 3D Brownian motion of individual kinesin molecules in viscous solution could be observed using this microscopy. These results indicated that this method could be applicable to the study of single molecular events in living cells at real time. © 2001 Academic Press Key Words: single-molecule imaging; confocal microscopy; fluorescence microscopy; Brownian motion.
Single-molecule biophysics has recently advanced and opened up a new era in the field of life sciences (1, 2). Imaging and manipulating single biomolecules enabled us to study dynamics of elementary process, and to obtain invaluable information that was not extractable through the conventional ensemble-averaged experiments. Following the success in visualizing a single fluorophore on an air-dried surface (3), a single fluorophore attached to a protein molecule in aqueous solution is first observed by using total internal reflection fluorescence microscopy (TIRFM) (4). Application of single-molecule imaging in vitro includes translational (5) and rotational (6) motion of single motor proteins, Abbreviations used: TIRFM, total internal reflection fluorescence microscopy; TMR, tetramethylrhodamine; NA, numerical aperture; MSD, mean square displacement. 1 To whom correspondence and reprint requests should be addressed. Fax: ⫹81-3-5286-3863. E-mail: [email protected].
chemical reaction in a single enzyme molecule (4, 7), and simultaneous measurement of chemical and mechanical events in a motor protein (8). Single-molecule imaging has recently been applicable to the study of living cells. Ligand-receptor interaction and oligomerization of adhesion proteins are visualized in living cells (9, 10). TIRFM is widely used for single-molecule imaging to reduce the background fluorescence by means of the local illumination of the evanescent field. For living cells, however, this method is not generally applicable since illumination using TIRFM is confined to extremely thin interfacial layers around cell membrane. To overcome these issues, we used a video-rate confocal microscopy (CSU10, Yokogawa Electric Co., Japan) to visualize single fluorescent molecules. CSU10 has two Nipkow disks, and light incident on 20,000 microlenses in the first disk is focused on corresponding pinholes on the second disk (11). This configuration increases transmission rate of the incident light to 50% while the conventional single Nipkow-disk scanner transmits only 2–3%. These disks rotate at 1800 rpm and a frame of image can be obtained every 2 ms. Owing to this video-rate confocal microscopy, 2D sliding movement of single TMR-kinesin molecules along microtubules and 3D Brownian motion of TMR-kinesin molecules in viscous solution could be observed. The photodecomposition lifetime of the single fluorophore up to 10 s was long enough for observing some biological phenomena. Thus, this study provided a basis for the single-molecule imaging in living cells. MATERIALS AND METHODS Proteins. Kinesin was prepared from bovine brain (12) and labeled with tetramethylrhodamine (TMR)-5-maleimide (Molecular Probes, U.S.A.) or IC5-maleimide (Dojindo, Japan) with dye to protein ratio of 0.6 or 1.0, respectively. Tubulin was prepared from porcine brain and labeled with Cy5 (Amersham–Pharmacia, Japan) according to the method of Hyman (13). Cy5 labeled tubulin was polymerized in the assembly buffer (80 mM Pipes, pH 6.8, 5 mM
323
0006-291X/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
Vol. 287, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
MgCl 2, 1 mM EGTA, 1 mM GTP, and 33% glycerol) at 37°C, and stabilized by adding 4 M Taxol. Optics. Nipkow disk-type confocal unit (CSU10, Yokogawa Electric Co., Japan) was installed to an inverted microscope (IX-70, Olympus Optical Co., Ltd., Japan). TMR was illuminated using a green solid-state laser (50 mW, 532 nm, -Green Model 4601, Uniphase, U.S.A.) and IC5 with a He–Ne laser (20 mW, 633 nm, GLG5601, NEC, Japan). A laser beam was depolarized by passage through a 1/4 plate and introduced into CSU10 without an optical fiber. It was focused at the inlet of CSU10 with numerical aperture (NA) of 0.01. The laser beam illuminated a specimen with diameter of up to 40 m by using oil-immersion objective (PlanApo 100⫻ NA ⫽ 1.4, Olympus). Dichroic mirrors and emission filters used for TMR and IC5 imaging were as follows; a dichroic mirror of DM560 (Asahi Spectrum Co., Japan) and emission filters of EM560 (Asahi Spectrum Co.) and 605DF80 (Omega Optical Inc., U.S.A.) for TMR imaging, and a dichroic mirror of DM650 (Asahi Spectrum Co.) and an emission filer of 680DF55 (Omega Optical Inc.) for IC5 imaging. Fluorescent image was taken by an ICCD camera (ICCD-350F, Video Scope International, U.S.A.) coupled to an image intensifier (VS41845, Video Scope International). Video signal was enhanced with an image processor (Argus20, Hamamatsu Photonics, Japan), and recorded on S-VHS format videotapes. Images were analyzed using a program written specifically for this analysis on a Halcon image processor (MVTec Software GmbH, Germany) or Scion image (Scion Corp., U.S.A.). Fluorescence intensities were calculated by integrating signals from 9 ⫻ 9 pixels and subtracting background. Single-molecule imaging in 2D plane. Coverslips were washed with 0.1 M KOH, ethanol and then dried in a clean bench as previously described (4). TMR-kinesin or IC5-kinesin in Buffer-A (80 mM Pipes, pH 6.9, 2 mM MgCl 2, 1 mM EGTA) was immobilized on a coverslip surface in the presence of oxygen scavenger system (25 mM glucose, 2.5 M glucose oxidase, 10 nM catalase, 10 mM dithiothreitol) (14). The laser power density at the specimen plane was estimated by measuring the power of laser that passed the objective lens and the illumination area of the specimen. Fluorescence intensity was measured as described above. Movement of single kinesin molecules along microtubule was observed as follows. A flow cell was made from a glass slide and coverslip with two slivers of film 50 m thick. Cy5-microtubules in Buffer-A containing 40 M Taxol were infused into the flow cell to be adsorbed on the coverslip. The glass surface was then blocked with 1 mg/ml casein (No. 073-19, Nacalai Tesque, Inc., Japan) to prevent the nonspecific adsorption of kinesin. TMR-kinesin in Buffer-A containing 40 M Taxol, 1 mM ATP, 1 mg/ml casein, 10 mM dithiothreitol, 0.2% methyl cellulose and oxygen scavenger system, was infused into the flow cell. The final concentration of TMR-kinesin in the motility assay was 1 nM. The slivers of film in flow cell were removed, and then the flow cell was sealed with the nail enamel. Microtubule and kinesin were observed using confocal microscopy at 23°C. The position of the fluorophore was determined using the homemade centroid calculating macro on Scion Image. Velocity was determined by dividing the run length by run time (the duration while kinesin moved on microtubule). Single-molecule imaging in 3D space. Solution containing 5 nM kinesin, 15% acrylamide, 5% bisacrylamide and 10% ammonium persulfate was mixed with N,N,N⬘,N⬘-tetramethylethylenediamide, infused into a flow cell, and acrylamide gels were polymerized for 1 h. Confocal images at various depths from the glass interface were obtained, and fluorescence intensities of the individual molecules were analyzed. To evaluate the aberration of objective, fluorescent microbeads in 0.1 m diameter (Molecular Probes, U.S.A.), instead of TMR-kinesin, were added to polyacrylamide gels. Series of optical sections were obtained at every 0.1-m interval, and fluorescence intensities of individual microbeads were analyzed. For single-molecule imaging of 3D Brownian motion of TMRkinesin, Buffer-A containing 10 nM TMR-kinesin, 0.1 mg/ml ␣-casein and 60% (w/w) sucrose was infused into the flow cell. The viscosity
FIG. 1. Single-molecule imaging of fluorescent kinesin molecules using Nipkow disk-type confocal microscopy. (A) Fluorescence micrograph image of TMR-kinesin molecules attached to the glass surface. The image was integrated over eight video frames. Scale bar, 2 m. (B) Quantized photobleaching of the TMR molecule indicated by the arrow in A. The TMR molecule was photobleached at the time indicated by the arrowhead. (C) Distribution of the fluorescence intensities of TMR-kinesin molecules that were bleached with single (open columns) and two (closed columns) steps. The power of the incident laser was 1.9 W/mm 2. Solid lines indicated Gaussian functions fitting to the data. Arrowheads indicated average fluorescence intensities. (D) Fluorescent intensity of the TMR-kinesin molecules as a function of laser light intensity. The bars indicated standard deviation. The data were fitted by a linear line function. (E) Time course of the decrease in number of TMR molecules due to photobleaching. TMRkinesin molecules were illuminated by a green laser at the light intensity of 1.9 W/mm 2. The data were fitted by a single exponential function with photodecomposition lifetime of 3.1 s. (F) Rate constants of the photobleaching of TMR-kinesin (closed circles) and IC5kinesin (open circles) molecules as a function of laser light intensity.
coefficient of 60% sucrose solution was 53.2 mPa 䡠 s. Diffusional motion of kinesin was observed at a 2-m distance from the glass surface. First, diffusion constant in a horizontal direction (D x ⫺y ) was determined based on Einstein’s equation:
4D x⫺y ⌬t ⫽ 具⌬x 2 ⫹ ⌬y 2 典
[1]
where, ⌬t was time interval and 具⌬x 2 ⫹ ⌬y 2 典 was two-dimensional mean square displacement (MSD). Next, diffusion constant in vertical direction (D z ) was determined using the following equation (15):
324
Vol. 287, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Sequential images of single TMR-kinesin molecule sliding on a microtubule at 23°C. (A) A fluorescence micrograph image of IC5-labeled microtubule. (B) A series of fluorescent micrograph images of TMR-kinesin taken every 1 s without frame averaging. Arrowheads indicated kinesin moving along microtubule. Numerals represented the time in seconds. Scale bar, 2 m.
12D z ⌬t ⫽ 具 ␦ z 2 典,
(Fig. 1B). Figure 1C showed the distributions of the fluorescence intensity of the individual spots that were photobleached in a single step (open columns) or two steps (close columns). The distribution was well fitted by two Gaussian distributions. The mean intensity of the fluorescent spots that were bleached in two steps was as twice strong as the value of those bleached in single step. These results indicated that single fluorescent molecules could be seen using confocal microscopy. A well-known problem of the fluorescence imaging with the confocal microscopy is a saturation of the fluorescence intensity of the fluorophore. We thus examined the dependence of the fluorescence intensity of the TMR-kinesin on the excitation laser intensity. We found a linear relationship between the laser power and the fluorescence intensity in the power range from 0.65 to 3 W/mm 2 (Fig. 1D). In our optical configuration,
[2]
where, ␦ t was the resident time of TMR-kinesin in a confocal cross section and ␦ z was the depth of the confocal optical section.
RESULTS AND DISCUSSION 2D Imaging of Single Fluorescent Molecules CSU10, a video-rate confocal unit using Nipkow disk, was installed to an inverted microscope (see details described under Materials and Methods). Laser beam was directly introduced into CSU10 without using a single mode optical fiber with NA 0.11, because high laser power density required for single-molecule imaging could not be obtained in default configuration. Inhomogeneity of illumination due to interference was less than 10% in a specimen plane. To evaluate the performance of single-molecule imaging using this confocal microscopy, kinesin was labeled with fluorescent dye of TMR and attached to the coverslip, then illuminated with green laser of 1.9 W/mm 2 at the specimen plane (Fig. 1). Through several approaches, it was confirmed that individual fluorescent molecules could be seen using this method. First, the number of fluorescent spots depended on kinesin concentration. The number of fluorescence spots was proportional to the applied kinesin concentration except when kinesin was extremely diluted and adsorption during preparation was not negligible. We found a very small number of light spots (less than 0.001 spots per square micrometer), presumably dusts, in the absence of kinesin. Effects of these dusts were negligible in the analysis. The time course of the fluorescence intensity of the individual spots showed quantized photobleaching
FIG. 3. Changes in fluorescence intensities with increasing penetration into solution. (A) Relative fluorescence intensities of single TMR-kinesin molecules embedded in a polyacrylamide gel were measured at 0, 2, 5, and 10 m away from the glass-solution interface. The bars indicated standard deviations. N ⫽ 76 ⬃ 116. The data was fitted by a linear line function. (B) Relative fluorescence intensities of microbeads as a function of the distance from the glass solution. Fluorescent microbeads with diameter of 0.1 m were embedded in a polyacrylamide gel, and fluorescence micrograph images were obtained every 0.1 m. Open circles indicated the average of relative fluorescence intensities of microbeads (N ⫽ 10) located at 0, 2, 5, and 9 m from the glass solution interface as shown by arrows.
325
Vol. 287, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Diffusion of TMR-kinesin molecules in 60% sucrose solution. (A) Sequential images of a single TMR-kinesin molecule taken at video-rate. Numerals represented the time in milliseconds. Scale bar, 2 m. Measurement was performed at a temperature of 23°C. (B) 2D mean square displacement as a function of time. Bars indicated standard errors of the data. (C) Duration of TMR-kinesin observed during passing through the confocal plane. Arrowhead indicates the average duration (0.13 s).
CSU10 scanned the illumination area with 256 pinholes, and we observed a single fluorophore at an illumination intensity of 12 W per pinhole when laser power density was 1.9 W/mm 2 at the specimen plane. Using a single-beam laser, the fluorescence intensity of the fluorophore is reported to saturate at about 1 mW (16). At our setting, we observed a single molecule at about 1/100 laser intensity of the saturation power. This observation might account for the linear dependency of the fluorescence intensity on the excitation laser intensity in our study. Figure 1E showed the photodecomposition of TMRkinesin under the illumination power of 1.9 W/mm 2. Photobleaching rates of TMR and IC5 were measured under the various laser power conditions. We found a linear relationship between laser power and photobleaching rate (Fig. 1F). As a demonstration of this microscopy for visualizing protein–protein interactions, sliding movement of TMR-kinesin molecules along microtubule was visualized. Microtubules labeled with Cy5 were attached to a coverslip of the chamber (Fig. 2A), and 1 nM of TMRkinesin was infused in the presence of ATP. The movement of fluorescent spots of kinesin was observed (Fig. 2B). The sliding velocity was 0.29 ⫾ 0.17 m/s (mean ⫾ SD, n ⫽ 7), which was in good agreement with the reported value for native kinesin movement on axoneme (17). Individual fluorescent spots of kinesin could be hardly observed when using conventional epifluorescence microscopy because of the background fluorescence of kinesin freely diffusing in solution.
3D Imaging of Immobilized Fluorescent Molecules Next we observed fluorescent molecules in 3D space. To evaluate the depth dependency of the fluorescence intensity, TMR-kinesin molecules were immobilized in polyacrylamide gels, and fluorescence images at various depths from the glass solution interface were obtained. Fluorescence intensities of kinesin molecules decreased due to aberration as the distance from the coverslip increased (Fig. 3A), and it became 50% at 10 m from the glass surface. To quantify the intensity and spread of the fluorescence, fluorescent microbeads in 0.1 m diameter were used instead of kinesin. Figure 3B showed the depth dependency of the fluorescence intensity relative to that of beads immobilized on coverslip. The half width of the fluorescence intensity was 0.53 ⫾ 0.1 m (mean ⫾ SD) at the glass surface and 1.05 ⫾ 0.1 m at 10 m from the cover glass surface. 3D Imaging of Brownian Motion of Single Fluorescent Molecules We observed single TMR-kinesin molecules undergoing Brownian motion in solution, and followed their trajectory while they were in confocal cross section at appropriate z position. To slow down the movement of kinesin, 60% sucrose was included in the solution. Figure 4A showed typical sequential video images of a single TMR-kinesin molecule moving at 2 m from the glass surface. The 2D MSD was calculated and plotted
326
Vol. 287, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
against time (Fig. 4B). By fitting a straight line to the data, diffusion constant in a horizontal direction (D x ⫺y ) of 0.75 m 2/s at ⫽ 53.2 mPa 䡠 s was obtained. The diffusion constant in a vertical direction (D z) was determined using the resident time of TMR-kinesin in a confocal cross section ( ␦ t) and the depth of the confocal plane ( ␦ z ⫽ 1.0 m) in Eq. [2]. The depth of the confocal plane ␦z was assumed to be 1.0 m, since the fluorescence intensity decreased to 20% of the maximum intensity when confocal plane was displaced ⫾ 0.5 m from the molecules, and individual fluorescent molecules could not be clearly observed (Fig. 3). The average resident time was 0.13 s for ⫽ 53.2 mPa 䡠 s. D z was calculated to be 0.64 m 2/s for ⫽ 53.2 mPa 䡠 s. The diffusion constants of both lateral and vertical diffusion were quite similar, and they were estimated to be 34 – 40 m 2/s in the absence of sucrose, which was in good agreement with the values measured in ensemble biochemical studies (18, 19). In conclusion, 2D and 3D movement of single molecules was successfully visualized using video-rate confocal microscopy. This method could be applied to the examination of single molecular events inside living cells. ACKNOWLEDGMENTS We thank Messrs. T. Tanaami, A. Ichihara (Yokogawa Electric Co.), and K. Abe (Olympus Co., Ltd.) for valuable comments and Drs. T. Tojo and T. Zako (Waseda University) for critical reading of the manuscript. This work was partly supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture of Japan and by Uehara Memorial Foundation.
REFERENCES 1. Harada, Y., Funatsu, T., Tokunaga, M., Saito, K., Higuchi, H., Ishii, Y., and Yanagida, T. (1998) Single molecule imaging and nanomanipulation of biomolecules. Methods Cell Biol. 55, 117– 128. 2. Weiss, S. (1999) Fluorescence spectroscopy of single biomolecules. Science 283, 1676 –1683. 3. Betzig, E., and Chichester, R. J. (1993) Single molecules observed by near-field scanning optical microscopy. Science 262, 1422–1425. 4. 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.
5. 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. 6. Adachi, K., Yasuda, R., Noji, H., Itoh, H., Harada, Y., Yoshida, M., and Kinosita, K., Jr. Stepping rotation of F1-ATPase visualized through angle-resolved single-fluorophore imaging. (2000) Proc. Natl. Acad. Sci. USA 97, 7243–7247. 7. Lu, H. P., Xun, L., and Xie, X. S. (1998) Single-molecule enzymatic dynamics. Science 282, 1877–1882. 8. Ishijima, A., Kojima, H., Funatsu, T., Tokunaga, M., Higuchi, H., Tanaka, H., and Yanagida, T. (1998) Simultaneous observation of individual ATPase and mechanical events by a single myosin molecule during interaction with actin. Cell 92, 161–171. 9. Sako, Y., Minoguchi, S., and Yanagida, T. (2000) Single-molecule imaging of EGFR signaling on the surface of living cells. Nat. Cell Biol. 2, 168 –172. 10. Iino, R., Koyama, I., and Kusumi, A. (2001) Single molecule imaging of green fluorescent proteins in living cells: E-cadherin forms oligomers on the free cell surface. Biophys. J. 80, 2667– 2677. 11. Ichihara, A., Tanaami, T., Isozaki, K., Sugiyama, Y., Kosugi, Y., Mikuriya, K., Abe, M., and Uemura, I. (1996) High-speed confocal fluorescence microscopy using a Nipkow scanner with microlenses for 3D imaging of single fluorescent molecule in real time. Bioimages 4, 57– 62. 12. Kojima, H., Muto, E., Higuchi, H., and Yanagida, T. (1997) Mechanics of single kinesin molecules measured by optical trapping nanometry. Biophys. J. 73, 2012–2022. 13. Hyman, A., Drechsel, D., Kellogg, D., Salser, S., Sawin, K., Steffen, P., Wordeman, L., and Mitchison, T. (1991) Preparation of modified tubulins. Methods Enzymol. 196, 478 – 485. 14. Harada, Y., Sakurada, K., Aoki, T., Thomas, D. D., and Yanagida, T. (1990) Mechanochemical coupling in actomyosin energy transduction studied by in vitro movement assay. J. Mol. Biol. 216, 49 – 68. 15. Berg, H. C. (1993) Diffusion to capture. In Random Walks in Biology, new, expanded edition, pp. 37– 47, Princeton Univ. Press. 16. Sandison, D. R., Williams, R. M., Wells, K. S., Sticker, J., and Webb, W. W. (1995) Quantitative fluorescence confocal laser scanning microscopy (CLSM). In Handbook of Biological Confocal Microscopy (Pawley, J. B., Ed.), 2nd ed., pp. 39 –53, Plenum Publishing Corp. 17. Friedman, D. S., and Vale, R. D. (1999) Single-molecule analysis of kinesin motility reveals regulation by the cargo-binding tail domain. Nat. Cell Biol. 1, 293–297. 18. Bloom, G. S., Wagner, M. C., Pfister, K. K., and Brady, S. T. (1988) Native structure and physical properties of bovine brain kinesin and identification of the ATP-binding subunit polypeptide. Biochemistry 27, 3409 –3416. 19. Hackney, D. D., Levitt, J. D., and Wagner, D. D. (1991) Characterization of alpha 2 beta2 and alpha 2 forms of kinesin. Biochem. Biophys. Res. Commun. 174, 810 – 815.
327