Proc. Nati. Acad. Sci. USA Vol. 75, No. 2, pp. 857-861, February 1978 Cell Biology

Molecular cytochemistry: Incorporation of fluorescently labeled actin into living cells (Physarum/amoeba/fibrils/motile extracts)

D. LANSING TAYLOR AND Yu-Li WANG The Biological Laboratories, Harvard University, Cambridge, Massachusetts 02138

Communicated by Keith R. Porter, November 17,1977

ABSTRACT

Actin labeled with 5-iodoacetamidofluorescein has been incorporated into the functional pool of actin in Chaos carolinensis and Physarum polycephalum by direct microinjection. The functional activity of the labeled actin has been analyzed at three levels of organization as: (a) with the purified actin, (b) in motile extracts of cells, and (c) in living motile cells. The labeled actin exhibited normal polymerization and activated myosin ATPase to a similar extent as unlabeled controls. Labeled actin and endogenous actin were incorporated into contracted pellets to approximately the same extent in motile cell extracts. After labe ed actin had been microinjected into single C. carolinensis cells, the fluorescent actin spread into both the endoplasm and ectoplasm without forming distinct fibrils. In contrast, fluorescent bundles developed in the ectoplasm of P. polycephalum following microinjection of labeled actin. This experimental method in conjunction with fluorescence spectroscopic techniques could become a powerful tool for studying the intracellular distribution and structural changes of components in living cells. Until recently, the field of nonmuscle-cell motility has relied heavily on the techniques of electron microscopy and immunofluorescence microscopy. While valuable information has been obtained concerning the identification and localization of contractile and cytoskeletal proteins (1-7), the application of these techniques is limited by several important factors. First, electron microscopy and immunofluoresence techniques are static tools which are unable to follow the highly dynamic processes involved in cell motility, including the polymerization-depolymerization of proteins; and, second, neither technique yields any information about the microenvironment or conformation of the contractile and cytoskeletal proteins. Thus, the field of cell motility requires some new techniques that permit simultaneous localization and characterization of specific proteins in vivo as well as in vitro. Fluorescent probes have been used as reporter groups of the conformation and activities of biological molecules in vitro. Fluorescent parameters such as lifetime of excited state, quantum yield, excitation spectra, emission spectra, and polarization reflect various properties of the microenvironment of the fluorophores, thus the structure and activities of a fluorescently labeled molecule can be inferred (8, 9). Fluorescence techniques have been applied in vitro to investigate interactions of specific contractile proteins, such as actin polymerization (10, 11), actin-myosin interactions (12), and actin-tropomyosin interactions (13). Because of the great sensitivity of fluorescence techniques, it should be possible to apply fluorescence spectroscopy at the single cell level. During the past few years, the methods and experience gained by various applications of fluorescence The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

spectroscopy in solution have slowly been incorporated into quantitative fluorescence microscopy (microspectrofluorometry) (14-16). For example, analysis of extrinsic fluorescence polarization from single striated muscle fibers has yielded important information about the angular motion of subfragment 1 moieties of myosin during contraction (17). However, to apply such techniques to single living nonmuscle cells, specific proteins must be isolated, labeled extrinsically without destroying the functional activities, and finally reincorporated into target cells. In the present paper we have demonstrated that purified actin can be specifically labeled while maintaining biological activities and can be incorporated into the actin pool in motile models and in selected nonmuscle motile cells. The amoeba Chaos carolinensis and the slime mold Physarum polycephalum have been studied initially because a great deal of biochemical, ultrastructural, and motile model information exists for these cells. MATERIALS AND METHODS Growth of Cells. Microplasmodia of P. polycephalum (strain C H357, a gift of C. N. Holt, Massachusetts Institute of Technology) were grown according to the method of Daniel and Baldwin (18). C. carolinensis were grown as described by Taylor et al. (19), while Dictyostelium discoideum (strain A.3) was grown in axenic culture (20). Physarum microplasmodia and Chaos were starved for 2 days before use by placing them in nutrient-free media. Dictyostelium were harvested during the logarithmic growth phase (21, 22). Preparation of Cells and Models. Motile extracts of D. discoideum were prepared by a differential centrifugation method (22) that was an extension of earlier methods (21, 23). Labeled proteins were added to the motile extract at a final concentration of 0.05-0.15 mg/ml (5.0-15.0% of the total endogenous actin concentration). Contraction of the extract was initiated as described previously (22). Quantitative measurements of labeled proteins remaining in the supernatant following a short clarification step (10,000 X g centrifugation for 10 min) were performed by measuring the absorbance of the dye-actin complex at 495 nm with a Beckman Acta III spec-

trophotometer. Single cell amoeba models were prepared as described earlier (19), as were single cell microinjections (24). The ratio of intracellular labeled actin to unlabeled endogenous actin was adjusted to ca 0.1 by controlling the volume and actin concentration injected into the cells (24). The fluorescein isothiocyanate-labeled bovine serum albumin was injected so the fluorescence intensity was approximately equivalent to that of labeled actin. Abbreviations: 5-IAF, 5-iodoacetamidofluorescein; RITC, rhodamine isothiocynate; FITC, fluorescein isothiocyanate.

857

858

Proc. Natl. Acad. Sci. USA 75 (1978)

Cell Biology: Taylor and Wang

Protein Labeling. Actin was purified to electrophoretic homogeneity from rabbit psoas muscle (25). F-actin (3.0-5.0 mg/ml) was labeled with 5-iodoacetamidofluorescein (5IAF, Molecular Probes, Inc., Roseville, MN) at an initial dye-toprotein molar ratio of 100:1. The dye exhibited a single spot on silica gels with acetone and methanol (1:1, vol/vol) as the solvent. The actin was labeled for 1 hr at 200 in a buffer containing 50 mM Tris-HCl, 100 mM KCl, 0.2 mM CaCl2, 2.0 mM MgCl` at pH 8.0-8.6. The reactants were subsequently saturated with a molar excess of cysteine, followed by the depolymerization of actin, and desalted on a Sephadex G-25 column. The actin was cycled from F-actin to G-actin twice and the final protein concentration was determined by the Lowry procedure with bovine serum albumin as the standard (26). The dye-to-protein ratio was calculated as 0.4-0.6, using an extinction coefficient of 60,000 M-1 cm-' for the bound dye (Y. L. Wang and D. L. Taylor, unpublished data). The emission spectrum was measured for both G-actin and F-actin with a Perkin-Elmer MPF 44 spectrofluorometer. A detailed analysis of the labeled actin will be published elsewhere (Y. L. Wang and D. L. Taylor). Bovine serum albumin (Sigma) was labeled with either fluorescein isothiocyanate (FITC) (Sigma isomer I) or rhodamine isothiocyanate (RITC) (Sigma) by a modification of a previous method (27). The free dye was removed by gel filtration (27) and subsequent dialysis, resulting in a final dye-to-protein ratio of 2.0. The labeled albumin was subsequently treated exactly like the actin. Some control experiments were performed with denatured labeled actin that was prepared by a 3-day dialysis of the labeled actin against 2 mM Tris.HCI buffer, pH 8.0. Optical Methods. The labeled cells, cell models, motile extracts, and actin paracrystals were observed with a Zeiss photomicroscope III equipped with bright-field, polarized light, and epifluorescence optics with 25X Neofluar objectives. A sensitive microscope photometer (Custom Instrumentation, Ravena, NY) connected to a strip chart recorder was used to measure the relative fluorescence intensities. Micrographs were taken with Kodak Tri-X pan film. Actin Paracrystals. Labeled actin (0.5 mg/ml) and unlabeled actin (0.5 mg/ml) were mixed to form different ratios of labeled/unlabeled actin (1:4, 1:16, and 1:32). Actin paracrystals were subsequently induced by the addition of MgCl2 to 25 mM. Actin Assays. Viscosity measurements were made with an Ostwald viscometer (outflow time of buffer, 30 sec) with actin at 0.73 mg/ml and the ATPase activity was measured by a method modified from Martin and Doty (28). RESULTS Properties of Labeled Actin. The labeled actin exhibited an absorption peak wavelength at 495 nm and an emission peak wavelength at 521 nm (Y. L. Wang and D. L. Taylor, unpublished data). The polymerizability and extent of activation of myosin ATPase has been compared between labeled actin and unlabeled controls. The reduced viscosity of unlabeled actin was 5.6 dl/g, while the labeled actin possessed 87% of the control viscosity (4.9 dl/g) after 2 hr of polymerization. Both the unlabeled and labeled actins activated the myosin ATPase up to 13-fold. Motile Extracts. The birefringent contracted pellets exhibited fluorescence only when functional labeled actin was present before the contractions were induced (Fig. 1 a and b). In contrast, both denatured labeled actin and labeled albumin were excluded from the birefringent contracted pellets and the

IM,

FIG. 1. Contracted pellets of motile extracts from D. discoideum demonstrating the birefringence (a) and fluorescence (b) of the same region when labeled actin was added before contractions were induced. (X390.)

pellets exhibited no fluorescence. The labeled actin apparently copolymerized with the endogenous actin in the extract, because the labeled actin was observed in the contracted pellets even when the labeled actin concentration was 0.05 mg/ml.

The labeled actin or BSA was added to the extract before or after induction of contractions. The absorbances of the contraction supernatants at 495 nm were compared to determine the amount of labeled protein incorporated into the contracted pellets. Between 19 and 20% of the functional labeled actin was incorporated into the pellet, while only 1-3% of either labeled albumin or denatured actin was present in the pellet. The contracted pellets also contained 20-23% of the endogenous D. discoideum actin as determined by gel scans using procedures described previously (22). Single-Cell Models. Functional labeled actin was demonstrated to be part of the cellular actin pool by two control experiments with motile model systems of amoeboid movement: (a) "Flare streaming" models (which are single-cell motile models) were prepared by mechanically removing the plasmalemma (19). In the presence of a threshold contraction solution the disrupted cells exhibited fluorescence in streaming loops of cytoplasm when the cells were preloaded with functional labeled actin. Furthermore, the emptied plasmalemmas (23) exhibited strong fluorescence, which indicated that the labeled actin also became associated with the actin in the ec-

Cell Biology:

. *,

Taylor and Wang ..4

:.

Proc. Nati. Acad. Sci. USA 75 (1978)

.... .....:Rt R t

'

*.

ME:.

As,

'm

1:

ii,

:-

f, ...

nMl ..i

.,.

'.Z.

I.."N

-

859

.e

hi:

> ..

.'

:.

.; '.".l,&,. ,,.st.

Is' 4

\

........ s

.:

:.

.:

.

Actin FIG. 3. A region of a single living microplasmodium of P. polycephalum injected with labeled actin. Birefringent bundles (Upper) which are distinctly fluorescent (Lower) can be detected. (X420.)

L

FIG. 2. A cell double-injected with RITC-labeled albumin and 5-IAF-labeled actin and then contracted by microinjecting a contraction solution. The scan shows the distribution of the relative fluorescence of the actin and albumin in the contracted cell. The contracted cell was photographed with Nomarski optics. (X640.)

toplasm-membrane complex (23). In contrast, flare models prepared from cells preloaded with denatured labeled actin or labeled albumin exhibited no fluorescence in the flare loops. In fact, these nonfunctional labeled proteins diffused rapidly into the surrounding medium after rupturing the plasmalemma. (b) Single specimens of Chaos were injected with equal volumes of functional actin labeled with 5-IAF and albumin labeled with RITC. These double-labeled cells were subsequently microinjected with a contraction solution that induced the formation of a transient contracted "knot" (24). These contracted cells were scanned with the photometer system, first measuring the relative fluorescence of labeled actin and then RITC-labeled albumin. Fig. 2 indicates that the contracted region was enriched in labeled actin while the labeled albumin was more randomly distributed. Intact Living Cells. Several fundamental observations were

made on single Chaos microinjected with labeled proteins and observed during normal amoeboid movement over a 2-hr period. The injected cells recovered from the injections within a few minutes (24) when the relaxation solution was used as the carrier. However, the cells recovered slowly when other solutions were used (19, 23). Fluorescent vesicles formed in the cytoplasm following microinjection, particularly if the calcium ion concentration of the carrier solution was above ca. 1 tiM. All observations were terminated if more than 10 fluorescent vesicles formed. Small fluorescent knots formed in the cytoplasm at the site of microinjection only when functional actin was injected. These fluorescent regions usually dispersed within 10-15 min after injection. The fluorescence intensity of functional labeled actin appeared to be relatively uniform in the endoplasm and ectoplasm of the amoeba C. carolinensis. No discrete birefringent and/or fluorescent bundles were detected (29). Interestingly, several amoebae exhibited fluorescent "rods" forming the core of microspikes in the tails or uroids. No fluorescence micrographs of amoebae were possible because the cells moved too fast for the required time of exposure. Physarum exhibited differences from Chaos both in response to microinjection and in patterns of fluorescence of labeled actin. Physarum was much more susceptible to irreversible damage resulting from the injection than was Chaos, and the relaxation solution was required as a carrier for successful injections. Furthermore, no fluorescent vesicles were observed over a 2-hr period. The fluorescence intensity of functional labeled actin was uniform in the endoplasm but exhibited discrete fluorescent bundles in the ectoplasm 30 min following microinjection (Fig. 3). The fluorescent bundles were most obvious along the long axis of the microplasmodia in the same regions where birefringent actin bundles have been detected (30-32). In contrast, both the denatured labeled actin and la-

860

Cell Biology: Taylor and Wang

Magnesium paracrystals of actin at a ratio of 1:16 labeled photographed with polarized light (a) and fluorescence optics (b) under exactly the same conditions used for PhyFIG. 4.

to unlabeled actin sarum in

Fig.

3.

(X420.)

beled albumin showed uniform fluorescence in microplasmodia. Micrographs without excessive blurring were possible only in a few instances when gross movements were very small and the rate of streaming was temporarily low. In order to determine the detectability of fluorescent actin bundles, actin paracrystals were prepared at various ratios of labeled to unlabeled actin. Actin paracrystals the same size or smaller than the actin bundles observed in living Physarum were readily detected by fluorescence even at ratios of labeled to unlabeled actin as low as 1:32. The birefringent, fluorescent paracrystals formed from a 1:16 ratio of labeled to unlabeled actins exhibited even more contrast than the fluorescent bundles in living Physarum when photographed under exactly the same conditions (Fig. 4 a and b). DISCUSSION The molecular cytochemical method introduced in this paper must be treated with caution to minimize the possibility of artefacts. We believe that the following set of criteria should be used to evaluate these types of experiments: (a) purity and source of proteins, (b) specificity of labeling and separation of labeled proteins according to functional activity, (c) comparison of some functional activity of the labeled proteins with unlabeled controls, (d) investigation of the incorporation of labeled proteins in both cell-free extracts and intact cells, (e) incorporation of labeled proteins into cells in solutions with known effects on the cytoplasm, and (f) quantitative analysis of spectroscopic parameters performed only after the intracellular environmental effects are investigated. The present experiments were performed with electrophoretically pure rabbit psoas muscle actin that was labeled with 5-4AF, which exhibits sulfhydryl group specificity. A homog-

Proc. Natl. Acad. Sci. USA 75 (1978)

enous population of labeled actins was prepared by repeated polymerization and depolymerization of the actin sample. Furthermore, it has been demonstrated that both the polymerizability and ability to activate myosin ATPase activity has been maintained. The results with the cell extracts suggested that the labeled actin could at least copolymerize with the endogenous actin pool, because the labeled actin has been added to the extracts at final concentrations below the standard critical concentration (Fig. 1). The association of functional labeled actin with the contracted pellets was specific because the controls (denatured actin and albumin) did not become incorporated into the contracted pellets. Furthermore, similar percentages of labeled actin and endogenous amoeba actin were concentrated in the contracted pellets, suggesting that the labeled actin was a good reporter of actin activity. It is not presently understood why only about 20% of the actin became part of the contracted pellets under the conditions used. The two single-cell model experiments indicated that the functional labeled actin became part of the cellular actin pool, while the denatured actin or albumin diffused nonspecifically in the cells or cell models. The fluorescence intensity in the intracellular contracted knots (Fig. 2) further indicated the ability of labeled actin to participate in intracellular contractions and become incorporated into actin filaments. The absence of discrete fluorescence bundles in C. carolinensis is consistent with the diffuse pattern of birefringence detected (31). The fact that contractions can be induced in any region of the endoplasm or ectoplasm in intact cells (24) is also consistent with our present observations on the distribution of actin. The presence of fluorescent microspikes in some amoebae is interesting, but requires further investigation. The appearance of discrete fluorescent bundles that matched the birefringent bundles in intact microplasmodia of Physarum indicated that the labeled actin could be incorporated into normal structures containing actin. The endogenous bundles of actin are oriented both parallel to the long axis and circularly around the plasmodium in the ectoplasm (31). In accordance with the observations of Nakajima and Allen (31), large bundles were detected in regions of relatively low streaming velocity. However, no distinct circularly arranged birefringent and/or fluorescent bundles were detected in the microplasmodia under the conditions used. It was significant that the birefringent fibers developed in the injected cells only about 30 min after injection. Thus, the labeled actin appeared to require a "cycling" time in the microplasmodia. The fluorescence intensity of the paracrystals formed from a 1:16 ratio of labeled to unlabeled actin was even greater than the fluorescence intensity of the bundles in Physarum. These results indicated that actin bundles could be readily detected when labeled actin was diluted by more than an order of magnitude. Therefore, the fluorescent bundles in Physarum could represent the copolymerization and aggregation of endogenous Physarum actin with the exogenous labeled actin. Furthermore, the absence of fluorescent bundles in Chaos, where they are not expected, supports the suggestion that the fluorescent bundles observed in Physarum are not artefactual. The initial results with Chaos and Physarum suggest that actin as well as other contractile and cytoskeletal proteins can be incorporated into nonmuscle cells. The future development of molecular cytochemical methods will include spectroscopic measurements of protein interactions in vivo as well as in vitro. It should be possible to investigate the conformation and supramolecular structure of specific proteins in vivo. Thus, the

Cell Biology:

Taylor and Wang

response of single living cells to different experimental conditions can be analyzed. The authors would like to thank S. B. Hellewell, J. Lupo, and J. Z. Miller for technical assistance. J. Heiple supplied us with P. polycephalum. The helpful discussions with Dr. E. Haas are gratefully acknowledged. This work was supported by the Institute of Arthritis, Metabolism and Digestive Diseases Research Grant AM 18111 and a Research Career Development Award to D.L.T. 1. Ishikawa, H., Bischoff, R. & Holtzer, H. (1969) J. Cell Biol. 43, 312-328. 2. Aronson, J. F. (1965) 1. Cell Biol. 26,293-298. 3. Sanger, J. S. (1975) Proc. Natl. Acad. Sci. USA 72, 1913-1916. 4. Goldman, R. D. (1975) J. Histochem. Cytochem. 23,529-542. 5. Goldman, R. D., Yerna, M. & Schloss, J. A. (1976) J. Supramol. Struct. 5, 155-183. 6. Buckley, I. D. & Porter, K. R. (1974) J. Mlcrosc. (Oxford) 104, 107-120. 7. Lazarides, E. & Weber, K. (1974). Proc. Natl. Acad. Sci. USA 71, 2268-2272. 8. Radda, G. K. (1975) in Methods in Membrane Biology, ed. Korn, E. D. (Plenum Press, New York), Vol. 4, pp. 97-188. 9. Dandliker, W. B. & Portman, A. J. (1971) in Excited States of Proteins and Nucleic Acids, eds. Steiner, R. F. & Weinryb, I. (Plenum Press, New York), pp. 199-275. 10. Cheung, H. C., Cooke, R. & Smith, L. (1971) Arch. Biochem.

Biophys. 142,333-39. 11. Kawasaki, Y., Mihashi, K., Tanaka, H. & Ohnuma, H. (1976) Biochim. Blophys. Acta 446,166-178. 12. Sekine, T., Ohyashiki, T., Machida, M. & Kanaoka, Y. (1974) Biochim. Blophys. Acta 351, 205-213. 13. Ohyashiki, T., Sekine, T. & Kanaoka, Y. (1974) Biochim. Biophys. Acta 351, 214-223.

Proc. Nttl. Acad. Sci. USA 75 (1978)

861

14;. Thaer, A. A. & Sernetz, M. (1973) Fluorescence Techniques in Cell Biology (Springer-Verlag, New York). 15. West, S. S. (1969) in Physical Techniques in Biological Research, ed. Pollister, A. W. (Academic Press, New York) Vol. 3, pp. 253-321. 16. Fernandez, S. M. & Berlin, R. D. (1976) Nature 264,411-415. 17. Nihei, T., Mendelson, A. & Botts, J. (1974) Biophys. J. 14, 236-242. 18. Daniel, J. W. & Baldwin, H. H. (1964) in Methods in Cell Physiology, ed. Prescott, D. M. (Academic Press, New York) Vol. l,p.9. 19. Taylor, D. L., Condeelis, J. S., Moores, P. L. & Allen, R. D. (1973) J. Cell Biol. 59, 378-394. 20. Loomis, W. F. (1971) Exp. Cell Res. 64,484-486. 21. Taylor, D. L., Condeelis, J. S. & Rhodes, J. A. (1977) in Cell Shape and Surface Architecture, eds. Revel, J. P., Henning, U. & Fox, F. (Alan R. Liss, New York), Vol. 17, pp. 581-603. 22. Condeelis, J. S. & Taylor, D. L. (1977) J. Cell Biol. 74, 901927. 23. Taylor, D. L., Rhodes, J. A. & Hammond, S. A. (1976) J. Cell Biol.

70,123-143. 24. Taylor, D. L. (1977) Exp. Cell Res. 105,413-426. 25. Spudich, J. A. & Watt, S. (1971) J. Biol. Chem. 246, 48664871. 26. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 27. Chen, R. F. (1969) Arch. Biochem. Blophys. 133,263-276. 28. Pollard, T. D. & Korn, E. D. (1973) 1. Biol. Chem. 248,46824690. 29. Allen, R. D. (1971) Exp. Cell Res. 72,34-45. 30. Allera, A., Beck, R. & Wohlfarth-Bottermann, K. E. (1971) Cytoblologie 4, 437-450. 31. Nakajima, H. & Allen, R. D. (1965) 1. Cell Biol. 25,361-374. 32. Nagai, R., Yoshimoto, Y. & Kamiya, N. (1975) Proc. Jpn. Acad.

51,38-43.

Molecular cytochemistry: incorporation of fluorescently labeled actin into living cells.

Proc. Nati. Acad. Sci. USA Vol. 75, No. 2, pp. 857-861, February 1978 Cell Biology Molecular cytochemistry: Incorporation of fluorescently labeled ac...
2MB Sizes 0 Downloads 0 Views