Eur. J. Biochem. 60, 67-72 (1975)
Effects of Specific Chemical Modification of Actin Peter D. CHANTLER and Walter B. GRATZER Medical Research Council Cell Biophysics, Unit King’s College, London (Received July 10 / September 6 , 1975)
G-actin has been nitrated with tetranitromethane in conditions that lead to the modification of one tyrosine residue. The reactive residue was found by earlier workers to be Tyr-69. The nitrated actiii is conformationally similar to native G-actin, as judged by sedimentation velocity and circular dichroism analysis. A small proportion only is in the form of covalently linked dimers and trimers. The nitrated G-actin will polymerise to form filaments, indistinguishable in the electron microscope from those of native F-actin, but the polymerisation process is slower. Reduction of the nitrophenol group to the corresponding aminophenol leaves the properties of the protein in respect of polymerisation unchanged. When a dansyl group is introduced at the same point, however, the ability of the actin to polymerise is lost. The nitrated actin and its reduced counterpart will also bind heavy meromyosin, and the characteristic arrowhead formation of the bound molecules along the filaments can be seen in the electron microscope. Neither of the modified F-actins, however, significantly activates or inhibits the myosin ATPase activity. The fluorescence of nitrated actin is strongly quenched through the presence of the nitrophenol chromophore. In soluble complexes with heavy meromyosin the fluorescence is indistinguishable from the sum of the separate contributions of the two protein components. There is thus no measurable excitation transfer between any tryptophan residues on the myosin heads, such as that inferred to be present in the ATPase site, and the nitrotyrosine in position 69 of the actin sequence. Implications of this observation are considered in relation to the different interaction sites in myosin and in actin. The activation of heavy meromyosin ATPase by copolymers containing actin and nitroactin in different proportions has been measured, and is not proportional to the fraction of native actin. The results are consistent with the view that the function of actomyosin depends on the interaction of the myosin heads with more than one actin subunit.
Skeletal muscle actin is made up of monomers of molecular weight 42000, and its sequence has been determined [l]. This monomer must have distributed about its surface at least six or seven specific binding sites, namely two for adjacent actin molecules in the same helical strand [2], and two for the adjoining molecules in the associated strand, one for the nucleotide, with possibly a separate site for divalent metal ions, one or more for attachment to myosin, and at least one more for association with tropomysin and a troponin component [ 3 ] . As yet very little is known about the identity of these sites and any possible overlap and interactions between them. Selective chemical modification offers a means of studying these aspects, and a few such studies have already been described [4 - 61. One site of a chemical modification has been explicitly identified, and is that of nitration with tetranitromethane, which has been found by Lehrer and Elzinga [7] to occur on Tyr-69. These authors reported loss of Abbreviation. phony].
Dansyl,
5-dimethylaininonaphthalene-l-sul-
ability to polymerise after such modification of G-actin, whereas other workers [4] reported that nitration of actin in the filamentous form caused only a partial breakdown in polymer structure, and that the ability of actin so treated to stimulate myosin ATPase activity was partially retained. Because of the side-reactions into which tetranitromethane often enters [S], such conflict between observations in different laboratories is often more apparent than real, in that it depends on the precise experimental conditions used. We have attempted to establish the effect of nitration of the sensitive tyrosine on the properties of actin, in the hope of further using it to study the nature of the myosin binding-site, and as a basis for the introduction of environment-sensitive chromophoric groups. MATERIALS AND METHODS Actin was variously prepared by the procedures of Martonosi [9] and of Spudich and Watt [lo], starting from acetone powder of rabbit skeletal muscle. The
68
Chemically Modified Actin
product was screened by polyacrylamide gel electrophoresis in the presence of sodium dodecylsulphate [ll], and was in all cases found to be essentially free of other proteins. G-actin preparations also showed no detectable quantities of polymeric material when observed in the ultracentrifuge, using schlieren optics. Concentrations of G-actin were determined from the specific absorbance, A: Frn = 11.0 at 280nm [12]. Rabbit skeletal muscle myosin was prepared according to Perry [13] and heavy meromyosin following Lowey and Cohen [14]. Concentrations were determined spectrophotometrically, using specific absorbances of 5.3 and 6.5 respectively for myosin [13] and heavy meromyosin [14]. ATPase activity was measured with a pH-stat (Radiometer TTT-1) under nitrogen, with 10mM sodium hydroxide as the titrant. An excess of magnesium ions over ATP was maintained [15]. Amino acid analyses were performed with a Beckman 120B analyser. Free sulphydryl groups were determined by spectrophotometric titration, using Ellman’s reagent [16]. Electron microscopy of actin filaments was performed with a Philips EM 200 instrument. Specimens were negatively stained with uranyl acetate. The Factin concentration was in the range 0.1 - 0.3 mg/ml, and for observation of ‘decorated’ filaments [17], the grid was allowed to come into contact with heavy meromyosin solution at 0.1 mg/ml. Circular dichroism was measured with a Cary 61 instrument, and fluorescence with a Hitachi MPF-3L spectrofluorimeter, equipped with a thermostatted cell-housing. The absorbance at the excitation wavelength was not allowed to exceed 0.08. For nitration, G-actin in 0.2mM ATP, 1 mM sodium bicarbonate at pH 8.0 was gently agitated for 30 min at 4 “C with a sixty-fold molar excess of tetranitromethane, added as a 0.84 M solution in ethanol [S]. The resulting solution was then dialysed exhaustively against the same buffer to remove excess reagent and products. The concentration of nitrotyrosine in the product was determined spectrophotometrically on the basis of the absorbance change at 440 nm, which occurs when the chromophore ionises, on chang= 4760 M cm ing the pH from 4 to 10. Taking [18], the concentration of nitrotyrosine is given by 44,,/4760 [18]. The total protein concentration was determined by subjecting aliquots to pronase digestion, followed by spectrophotometric ninhydrin analysis. Unmodified G-actin was used as the standard. The reduction of the nitrotyrosine residues to 3-aminotyrosine was performed by addition of a fifty-fold excess of sodium dithionite [19], followed by dialysis. The reduction was reflected by the loss of absorbance at 420 nm. The introduction of a dansyl chromophore at the 3-amino group in the phenolic ring [20] was accomplished by adding 2% wjw of dansyl chloride in acetone to the protein in 1 mM sodium borate, 0.25 inM ATP, the pH being maintained at pH 6.5 for ~
12 h. Dansylation at pH 5.0 1201 was unsuccessful. At pH 6.5 the &-aminogroups do not react to a measurable extent in the conditions of our experiments, and the N-terminus in actin is protected by an acetyl group [l]. The extent of dansylation was determined spectrophotometrically [20].
RESULTS Nitration with a sixty-fold molar excess of tetranitromethane was selected as standard procedure for the preparation of nitroactin, since this corresponds to a plateau of reactivity in terms of the number of nitrotyrosine residues formed, whereas at higher concentrations of reagent insoluble products begin to appear. In a series of preparations the nitrotyrosine content, spectrophotometrically determined, was in the range 0.8 - 1.2 residues/molecule. Lehrer and Elzinga [7] have already shown that there is a single readily nitrated tyrosine residue in actin, and identified it as Tyr-69. It is well known [8] that tetranitromethane in some instances gives rise to oxidative side-reactions. In the case of actin these are clearly not extensive. The amino acid analysis reveals no contaminating components, and the absorption spectrum of the aminotyrosyl actin, generated by reduction of nitroactin, shows that there has been little damage to the four tryptophan residues, although a minor degree of destruction cannot be ruled out. Finally, dodecylsulphate-gel electrophoresis of the nitrated actin shows only traces (not more than some 5 %) of covalent dimer and trimer. This is too small a proportion to be unambiguously detectable in the analytical ultracentrifuge. To a good approximation, therefore, we can regard the product of reaction in our conditions as being actin functionally modified at Tyr-69. The titration of thiol groups reveals that one of these is lost in the course of nitration. This is presumably the single rapidly reacting group, the modification of which affects neither polymerisation nor activation of myosin ATPase [21]. When we subject G-actin to prior treatment with Ellman’s reagent to block the reactive thiol, nitration then causes no further drop in thiol titre. Moreover, the product behaves in all respects similarly to F-nitroactin which has not been treated with thiol reagent. Conformationally the nitrated actin is evidently very similar to the native molecule, judged by the sedimentation coefficient and the circular dichroism in the region of peptide absorption, where the new chromophore does not appreciably interfere. Fluorescence emission spectra showed a peak at 338 nm, with a much reduced intensity compared with native G-actin. Polymerisation of the nitrated actin gave rise to a very little change in the fluorescence at neutral pH. At pH 8 there is a red-shift in the G-nitroactin emission of 2 nm, which does not occur in F-nitroactin, and at pH 6 there
69
P. D. Chantler and Mi. B. Gratzer
Fig. 1. Electron micrographs of ( A ) F-actin and ( B ) F-nitroactin, negatively stained with uranyl acetate. Magnification x 75 000
is a small blue-shift with a drop in intensity. Evidently the conformational lability responsible for these changes is largely inhibited in the polymerised form. Denaturation with EDTA [21] leads to a red-shift in Gnitroactin, just as in native G-actin, in agreement with the report by Lehrer and Elzinga [7]. The circular dichroism in the aromatic region is profoundly affected by nitration. G-actin shows aromatic Cotton effects [23], with a prominent negative extremum at 295 nm and a shoulder at about 287 nm, both almost certainly due to contributions of tryptophan residues [24]. In the nitrated species the extremum remains apparent only as a shoulder, and a relatively large negative extremum (ellipticity- 3500 deg. cm’ dmol-’) emerges at 284 nm. The difference between the nitrated and the native species shows predominantly positive ellipticity at 296 nm and a negative effect at 285 nm, with lower activity at longer wavelengths. When the ionic strength of the solution of G-nitroactin was increased (50 mM potassium chloride, 2 mM magnesium chloride), birefringence was developed between crossed polaroids on gentle swirling. This occurred more slowly than in native G-actin similarly treated. Centrifugation yielded a yellow pellet and examination of the supernatant showed that the great bulk of the actin was in all cases incorporated into the polymers. Nitration in the manner described does not
destroy the ability of the protein to polymerise. Moreover, the polymers, when examined in the electron microscope, showed the characteristic helical threadlike structure of F-actin (Fig. l), from which it is indistinguishable. When heavy meromyosin is added to the F-nitroactin, sedimentation at 80000 x g for 3 h leads to its complete association with the F-actin pellet. When it is added to F-nitroactin on electron microscope grids, ‘decorated’ filaments with their characteristic arrowhead pattern appear, which again are indistinguishable from those formed by acto-heavy-meromyosin (Fig. 2). Measurements of the fluorescence spectrum of nitroacto-heavy-meromyosin in solution gave a result identical to the sum of the contributions of the F-nitroactin and the heavy meromyosin taken separately. There is thus no measurable transfer of excitation energy from aromatic residues of the myosin heads to the nitrophenol chromophore. The effect of F-nitroactin on the ATPase activity of myosin was tested in the pH-stat. In a typical experiment heavy meromyosin gave a turnover rate of 3.6 mol Pi . mol . protein-’ . min-l. After addition of an equivalent amount of F-nitroactin a value of 4.3 resulted. This is practically within the experimental error, and we conclude that though myosin heads associate with F-nitroactin with high affinity, the ATPase
Chemically Modified Actiii
Fig. 2. Electron micrographs ofjilarnerits of ( A ) F-uctin and ( B ) F-nitroactin, both decorated with heavy rneromjosin. Magnification x 75000
activity is not significantly enhanced and neither is there any inhibition. ATPase measurements were also performed on complexes of heavy meromyosin with polymers produced from mixtures of G-actin and Gnitroactin. In this case ATPase enhancement was observed, but throughout at a lower level than would be produced by the same concentration of actin in the absence of nitroactin. Reciprocal plots reflecting the ATPase stimulation [25]are shown in Fig. 3, and take the form of a set of straight lines intersecting at different points, all in the first quadrant of the graph. This unusual effect is discussed below. The aminotryrosyl actin behaves in most respects similarly to the nitro derivative. The absorption spectrum shows that the nitrophenol has been reduced, but there is a slight trail to the long-wavelength side of the absorption band. The fluorescence emission spectrum resembles that of native G-actin. The aminotyrosyl actin is able to polymerise, and both the bare filaments and those decorated with heavy meromyosin are indistinguishable from the corresponding structures formed by unmodified actin. The ability to activate myosin ATPase is not restored on reduction of the nitro group. The dansylaminoactin showed an absorption spectrum reflecting the presence of the dansyl chromo-
l/F-actin concn (rnlirng)
Fig. 3. Reciprocal plots showing the activation of myosin subfiagment-1 ATPase by F-uctin and copolymers ofuctin andnitroactin. The ordinate is reciprocal of ATPase velocity enhancement over that of subfragment-1 alone, and the abscissa (in inl/mg) refers to the concentration of the native actin fraction. Concentrations of the 0.01 (a),0.02 (0) and 0.03 (m) mg/ml nitroactin component are 0 (o),
phore (absorption maximum 325 nm), and when excited at 325 nm gave a strong fluorescence band at 495 nm. This derivative showed no propensity to polymerise when the ionic strength was increased.
71
P. D. Chantler and W. B. Gratzer
DISCUSSION The nitration of G-actin at Tyr-69 evidently leaves the conformation of the protein essentially undisturbed, as judged by the evidence of sedimentation velocity and circular dichroism. The quenching of the tryptophan emission by the nitrotyrosine chromophore (which is relieved when the latter is reduced to aminotyrosine) is sufficiently extensive that it must be ascribed to excitation transfer from one or more tryptophan residues in close proximity (of the order of 0.5 nm away, with a favourable orientation). A tryptophan close to Tyr-69 in the sequence [I] is Trp-74. The new extrinsic Cotton effects that accompany the introduction of the nitro group are predominantly in the region of the near-ultraviolet tryptophan transitions, rather than at the wavelengths at which only the nitrophenol group absorbs. It therefore seems probable that they arise from interaction of the transition dipoles of a tryptophan and a nitrotyrosine almost in van der Waals contact [26]. As regards the effect of nitration at this low level on the polymerisability of the actin, there is some ambiguity in the literature, one group of workers stating that this capacity remains on nitration (albeit of F-actin [S]), whereas a brief report on nitrated G-actin [7] indicates that polymerising ability is destroyed. The discrepancy inay be due to differences in the reaction conditions, or those of the polymerisation assays; one effect could arise from ionisation of the nitrotyrosine group, for example, which will change with pH in the neutral range. At all events our results are unequivocal in the sense that in our solvent conditions elevation of the ionic strength leads to polymerisation of all the nitroactin. Since the reactivity of Tyr-69 and the absence of any appreciable conformational changes on nitration, indicate that this residue lies on the surface of the G-actin molecule, the failure of nitration to inhibit polymerisation is not especially surprising, and indicates only that Tyr-69 is not directly involved in any of the four probable actin-actin interaction sites [2]. Neither is it surprising that when the nitro group is reduced to the much smaller amino group, the polymerising capacity still survives. On the other hand, the introduction of the dansyl group, which is about 0.8 nm across compared with about 0.3 nm for the nitro group, leads to total inhibition of polymerisation. We cannot determine whether the dansyl group physically obstructs an association site, causes a local conformational disturbance, which perhaps impedes the change normally engendered by salt to trigger the polymerisation, or introduces an unfavourable coulombic contribution from the positive charge. The inhibition of polymerisation by modification at the Tyr-69 position is evidently in any case a finely balanced process. The polymerised nitroactin retains its ability to bind the myosin heads, and there is nothing to indicate
that the strength of binding is in any way diminished. The arrowheads seen in the electron micrographs (Fig. 2) are characteristic of the functional actomyosin interaction. The absence of any accompanying enhancement of the myosin ATPase activity is striking, but not unique, for other chemical modifications, not necessarily of very defined specificity, have been found that do not prevent actomyosin formation but do considerably inhibit the activation of the myosin ATPase [27- 291. (The electron microscopic ‘decoration’ technique was not at that time available to establish whether the interaction was of the ‘native’ type). From the separation of the binding and activation functions it would appear to follow that actin does not activate the myosin ATPase by direct displacement of the bound products of hydrolysis [30]. Indeed, monomeric matrixbound actin in the ionic conditions corresponding to maximum activation by F-actin will also cause little activation [31], and the possibility therefore arises that the interaction with myosin involves two or several actin monomers in the thin filament, one energetically dominant interaction being sufficient for binding, while another is required to elicit ATPase activation. The latter site would be envisaged as involving Tyr-69, which is evidently not a part of the primary binding site. Other explanations would need to invoke small conformational changes in actin associated with the activation process, which would be inhibited by a local chemical modification. There are strong indications [32 - 341 that a tryptophan residue is present in the nucleotide-binding site of the myosin. If the actin attached at or near this site one would expect that energy transfer would occur to the nitrotyrosine group on the actin. An estimate of the lower limit of the distance involved is readily made, with the aid ofthe Forster sixthpower relation for resonance transfer, viz. Rg, the distance for 50 probability of excitation transfer within the lifetime of the excited state of the donor chromophore, is given by [35] : Rg
=
91n lox2 qJ 1 2 8 7 ~n4~N ’
where J is the overlap integral for the emission spectrum of the donor and the absorption spectrum of the acceptor, q the quantum yield of this donor, here taken as 0.14 for a typical protein tryptophan residue [36], n the refractive index of the protein medium, taken as 1.5, N the Avogadro number, and x, the orientation factor, is 2/3 if the two chromophores are in a state of free rotation. The fractional transfer efficiency at distance R is then (Ro/R)6/[1 (Ro/R)6].Since no transfer is seen, it must be supposed that unless the angle between the potential donor and acceptor chroniophores is fixed, and by chalice very close to 90”, these residues are in fact at least 1.5nm or so apart. On the other hand, Harrington and his associates [37,38]
+
72
have assembled strong evidence that the actin-binding site and the ATPase site of myosin are identical or contiguous ; this would be compatible with a location of Tyr-69 remote from the myosin-binding site. A satisfactory description of the actin binding and enzymic activation mechanism is evidently not yet available. The reciprocal plots of Fig. 3 , in which the concentration refers to native actin, are incompatible with a simple inhibitory function, whether competitive or non-competive, of the copolymerised nitroactin. The kinetics of the actomyosin ATPase reaction are much more complex than their representation in terms of a reciprocal plot implies, and Lymn [39] has shown that in the general case such plots will be curved, and for appropriate combinations of the individual rate constants can intersect in the first quadrant of the reciprocal plot [40]. The data shown in Fig. 3 are most readily compatible with a conformational effect, whereby a steric feature of the F-actin helix, or a change in conformation contingent on combination with myosin, is inhibited by the foreign species present in the polymer. Tawada and Oosawa [41] have noted that in copolymers of actin and carboxymethylactin (which activates myosin ATPase with identical apparent V , but lower K,,,), the activation is slightly but consistently less than that of a mixture of the homopolymers. Our effect, which is gross by comparison, may be related to this observation. However, whereas their data can be quantitatively accounted for if it is assumed that two adjacent monomers on the actin helix are responsible for ATPase activation, our results do not appear to be compatible with such a simple scheme. If the fraction of unmodified actin in the copolymer is CI and the concentration in protein monomers is 2,then the effective concentration of unmodified dimers in the chain (in molar units of monomers) will be a2A, of trimers a3A, and so on. We find that oui- data are not very satisfactorily fitted in terms of an active species of unmodified dimer or of trimer. We cannot, however, exclude non-random copolymerisation, which would vitiate this argument. Apart from this reservation, the nature of the reciprocal plots suggests that the conformation of the entire filament is changed in a manner that varies continuously with the proportion of the active copolymeric component, or that there are domains of the correct and incorrect conformations for activation, depending in frequency on the composition of the polymer. Both types of situation have been observed in protein aggregates (see, e. g., Asakura et al. [42]). We are grateful to R. Craig for help with electron microscopy, R. Hynes for amino acid analysis, and the Science Research Council for a Training Scholarship for P.D.C.
P. D. Chantler and W. B. Gratzer: Chemically Modified Actin
REFERENCES 1. Elzinga, M., Collins, J. H., Kuehl, W. M. & Adelstein, R. S. (1973) Proc. Nut1 Acud. Sci. U.S.A. 70, 2687-2691. 2. Oosawa, F. & Kasai, M. (1971) in Subunits in Biological Systems, ( S . N. Timasheff & G. D. Fasman, eds) part A, pp. 261 - 322, M. Dekker, New York. 3. Ebashi, S. (1974) Essays Biochem. 10, 1 - 36. 4. Miihlrad, A,, Corsi, A. & Granata, A. L. (1968) Biochim. Biophys. Acta, 162, 435 - 443. 5. Tawada, K., Asai, H. & Gerber, B. R. (1969) Biochim. Biophys. Acta. 194,486 - 497. 6. Cheung, H. C., Cooke, R. & Smith, L. (1971) Arch. Biochem. 142, 333- 339. 7. Lehrer, S . S . & Elzinga, M. (1972) Fed. Proc. 31, 502. 8. Riordan, J. F. & Vallee, B. L. (1972) in Methods Enzy.mo1. 25B, 515-521. 9. Martonosi, A. (1962) J . B i d . Chem. 237, 2795-2803. 10. Spudich, J. A. & Watt, S. (1971)J. Bid. Chem. 246,4866-4871. 11. Weber, K. & Osborn, M. (1969) J. Biol. Chem. 244,4406-4412. 12. Rees, M. K. & Young, M. (1967) J . Biol. Chem. 242, 44494458. 13. Perry, S. V. (1955) Methods Enzymol. 2, 582-588. 14. Lowey, S. & Cohen, C. (1962) J. Mol. Biol. 4, 293-308. 15. Reisler, E., Burke, M. & Harrington, W. F. (1974) Biochemistry, 13,2014-2022. 16. Ellman, G. L. (1959) Arch. Biochem. 82, 70-77. 17. Huxley, H. E. (1963) 1. Mol. Bid. 7, 281 - 308. 18. Beaven, G. H. & Gratzer, W. B. (1968) Biochim. Biophys. Acta, 168,456-462. 19. Sokolovsky, M., Riordan, J. F. & Vallee, B. L. (1967) Biochem. Biophys. Res. Commun. 27,20-25. 20. Kenner, R. A. & Neurath, H. (1971) Biochemistry, 10, 551 557. 21. Lusty, C. J. & Fasold, H. (1969) Biochemistry, 8, 2933-2940. 22. Lehrer, S. S. & Kerwar, G. (1972) Biochemistry, 11,1211- 1217. 23. Murphy, A. J. (1971) Biochemistry, 10, 3723-3727. 24. Strickland, E. H., Horwitz, J. & Billups, C. (1969) Biochemistry, 8, 3205-3213. 25. Eisenberg, E. & Moos, C. (1968) Biochemistry, 7, 1486-1489. 26. Tinoco, I. (1963) Radiat. Res., 20, 133-139. 27. Birany, M. & Barany, K. (1959) Biochim. Biophys. Acta, 35, 293 - 309. 28. Perry, S. V. & Cotterill, J. (1964) Biochem. J . 92, 603-608. 29. Kaldor, G., Gitlin, J., Westley, F. & Volk, B. W. (1964) Biochemistry. 3, 1137-1145. 30. Taylor, E. W. (1972) Annu. Rev. Biochem. 41, 577-616. 31. Chantler, P. D. & Gratzer, W. B. (1973) FEBS Lett. 34,lO- 14. 32. Yoshino, H., Morita, F. & Yagi, K. (1972)J. Biochem. (Tokyo) 72, 1227- 1235. 33. Werber, M. M., Szent-Gyorgyi, A. G . & Fasman, G. D . (1972) Biochemistry, 11, 2872-2882. 34. Mandelkow, E. M. & Mandelkow, E. (1973) FEBS Lett. 33, 161- 166. 35. Forster, T. (1959) Disc. Furaduy Soc. 27, 7- 17. 36. Eisinger, J., Feuer, B. & Lamola, A. A. (1969) Biochemistry, 8, 3908- 3915. 37. Burke, M., Reisler, E. & Harrington, W. F. (1973) Proc. Natl Acad. Sci. U.S.A. 70, 3793-3796. 38. Reisler, E., Burke, M. & Harrington, W. F. (1974) Biochemistry, 13,2014-2022. 39. Lymn, R. W. (1974) J. Theor. Biol. 43, 313-328. 40. Lymn, R. W. (1975) J. Theor. Biol. 49, 425-429. 41. Tawada, K. & Oosawa, F. (1969) J . Mol. Biol. 44, 309-317. 42. Asakura, S. & Iino, T. (1972) J. Mol. Biol. 64, 251-268.
P. D. Chantler, Department of Biology, Brandeis University, Waltham, Massachusetts, U.S.A. 02154 W. B. Gratzer, Medical Research Council Cell Biophysics Unit, King’s College, 26-29 Drury Lane, London, Great Britain WC2B 5RL
Effects of Specific Chemical Modification of Actin Peter D. CHANTLER and Walter B. GRATZER Medical Research Council Cell Biophysics, Unit King’s College, London (Received July 10 / September 6 , 1975)
G-actin has been nitrated with tetranitromethane in conditions that lead to the modification of one tyrosine residue. The reactive residue was found by earlier workers to be Tyr-69. The nitrated actiii is conformationally similar to native G-actin, as judged by sedimentation velocity and circular dichroism analysis. A small proportion only is in the form of covalently linked dimers and trimers. The nitrated G-actin will polymerise to form filaments, indistinguishable in the electron microscope from those of native F-actin, but the polymerisation process is slower. Reduction of the nitrophenol group to the corresponding aminophenol leaves the properties of the protein in respect of polymerisation unchanged. When a dansyl group is introduced at the same point, however, the ability of the actin to polymerise is lost. The nitrated actin and its reduced counterpart will also bind heavy meromyosin, and the characteristic arrowhead formation of the bound molecules along the filaments can be seen in the electron microscope. Neither of the modified F-actins, however, significantly activates or inhibits the myosin ATPase activity. The fluorescence of nitrated actin is strongly quenched through the presence of the nitrophenol chromophore. In soluble complexes with heavy meromyosin the fluorescence is indistinguishable from the sum of the separate contributions of the two protein components. There is thus no measurable excitation transfer between any tryptophan residues on the myosin heads, such as that inferred to be present in the ATPase site, and the nitrotyrosine in position 69 of the actin sequence. Implications of this observation are considered in relation to the different interaction sites in myosin and in actin. The activation of heavy meromyosin ATPase by copolymers containing actin and nitroactin in different proportions has been measured, and is not proportional to the fraction of native actin. The results are consistent with the view that the function of actomyosin depends on the interaction of the myosin heads with more than one actin subunit.
Skeletal muscle actin is made up of monomers of molecular weight 42000, and its sequence has been determined [l]. This monomer must have distributed about its surface at least six or seven specific binding sites, namely two for adjacent actin molecules in the same helical strand [2], and two for the adjoining molecules in the associated strand, one for the nucleotide, with possibly a separate site for divalent metal ions, one or more for attachment to myosin, and at least one more for association with tropomysin and a troponin component [ 3 ] . As yet very little is known about the identity of these sites and any possible overlap and interactions between them. Selective chemical modification offers a means of studying these aspects, and a few such studies have already been described [4 - 61. One site of a chemical modification has been explicitly identified, and is that of nitration with tetranitromethane, which has been found by Lehrer and Elzinga [7] to occur on Tyr-69. These authors reported loss of Abbreviation. phony].
Dansyl,
5-dimethylaininonaphthalene-l-sul-
ability to polymerise after such modification of G-actin, whereas other workers [4] reported that nitration of actin in the filamentous form caused only a partial breakdown in polymer structure, and that the ability of actin so treated to stimulate myosin ATPase activity was partially retained. Because of the side-reactions into which tetranitromethane often enters [S], such conflict between observations in different laboratories is often more apparent than real, in that it depends on the precise experimental conditions used. We have attempted to establish the effect of nitration of the sensitive tyrosine on the properties of actin, in the hope of further using it to study the nature of the myosin binding-site, and as a basis for the introduction of environment-sensitive chromophoric groups. MATERIALS AND METHODS Actin was variously prepared by the procedures of Martonosi [9] and of Spudich and Watt [lo], starting from acetone powder of rabbit skeletal muscle. The
68
Chemically Modified Actin
product was screened by polyacrylamide gel electrophoresis in the presence of sodium dodecylsulphate [ll], and was in all cases found to be essentially free of other proteins. G-actin preparations also showed no detectable quantities of polymeric material when observed in the ultracentrifuge, using schlieren optics. Concentrations of G-actin were determined from the specific absorbance, A: Frn = 11.0 at 280nm [12]. Rabbit skeletal muscle myosin was prepared according to Perry [13] and heavy meromyosin following Lowey and Cohen [14]. Concentrations were determined spectrophotometrically, using specific absorbances of 5.3 and 6.5 respectively for myosin [13] and heavy meromyosin [14]. ATPase activity was measured with a pH-stat (Radiometer TTT-1) under nitrogen, with 10mM sodium hydroxide as the titrant. An excess of magnesium ions over ATP was maintained [15]. Amino acid analyses were performed with a Beckman 120B analyser. Free sulphydryl groups were determined by spectrophotometric titration, using Ellman’s reagent [16]. Electron microscopy of actin filaments was performed with a Philips EM 200 instrument. Specimens were negatively stained with uranyl acetate. The Factin concentration was in the range 0.1 - 0.3 mg/ml, and for observation of ‘decorated’ filaments [17], the grid was allowed to come into contact with heavy meromyosin solution at 0.1 mg/ml. Circular dichroism was measured with a Cary 61 instrument, and fluorescence with a Hitachi MPF-3L spectrofluorimeter, equipped with a thermostatted cell-housing. The absorbance at the excitation wavelength was not allowed to exceed 0.08. For nitration, G-actin in 0.2mM ATP, 1 mM sodium bicarbonate at pH 8.0 was gently agitated for 30 min at 4 “C with a sixty-fold molar excess of tetranitromethane, added as a 0.84 M solution in ethanol [S]. The resulting solution was then dialysed exhaustively against the same buffer to remove excess reagent and products. The concentration of nitrotyrosine in the product was determined spectrophotometrically on the basis of the absorbance change at 440 nm, which occurs when the chromophore ionises, on chang= 4760 M cm ing the pH from 4 to 10. Taking [18], the concentration of nitrotyrosine is given by 44,,/4760 [18]. The total protein concentration was determined by subjecting aliquots to pronase digestion, followed by spectrophotometric ninhydrin analysis. Unmodified G-actin was used as the standard. The reduction of the nitrotyrosine residues to 3-aminotyrosine was performed by addition of a fifty-fold excess of sodium dithionite [19], followed by dialysis. The reduction was reflected by the loss of absorbance at 420 nm. The introduction of a dansyl chromophore at the 3-amino group in the phenolic ring [20] was accomplished by adding 2% wjw of dansyl chloride in acetone to the protein in 1 mM sodium borate, 0.25 inM ATP, the pH being maintained at pH 6.5 for ~
12 h. Dansylation at pH 5.0 1201 was unsuccessful. At pH 6.5 the &-aminogroups do not react to a measurable extent in the conditions of our experiments, and the N-terminus in actin is protected by an acetyl group [l]. The extent of dansylation was determined spectrophotometrically [20].
RESULTS Nitration with a sixty-fold molar excess of tetranitromethane was selected as standard procedure for the preparation of nitroactin, since this corresponds to a plateau of reactivity in terms of the number of nitrotyrosine residues formed, whereas at higher concentrations of reagent insoluble products begin to appear. In a series of preparations the nitrotyrosine content, spectrophotometrically determined, was in the range 0.8 - 1.2 residues/molecule. Lehrer and Elzinga [7] have already shown that there is a single readily nitrated tyrosine residue in actin, and identified it as Tyr-69. It is well known [8] that tetranitromethane in some instances gives rise to oxidative side-reactions. In the case of actin these are clearly not extensive. The amino acid analysis reveals no contaminating components, and the absorption spectrum of the aminotyrosyl actin, generated by reduction of nitroactin, shows that there has been little damage to the four tryptophan residues, although a minor degree of destruction cannot be ruled out. Finally, dodecylsulphate-gel electrophoresis of the nitrated actin shows only traces (not more than some 5 %) of covalent dimer and trimer. This is too small a proportion to be unambiguously detectable in the analytical ultracentrifuge. To a good approximation, therefore, we can regard the product of reaction in our conditions as being actin functionally modified at Tyr-69. The titration of thiol groups reveals that one of these is lost in the course of nitration. This is presumably the single rapidly reacting group, the modification of which affects neither polymerisation nor activation of myosin ATPase [21]. When we subject G-actin to prior treatment with Ellman’s reagent to block the reactive thiol, nitration then causes no further drop in thiol titre. Moreover, the product behaves in all respects similarly to F-nitroactin which has not been treated with thiol reagent. Conformationally the nitrated actin is evidently very similar to the native molecule, judged by the sedimentation coefficient and the circular dichroism in the region of peptide absorption, where the new chromophore does not appreciably interfere. Fluorescence emission spectra showed a peak at 338 nm, with a much reduced intensity compared with native G-actin. Polymerisation of the nitrated actin gave rise to a very little change in the fluorescence at neutral pH. At pH 8 there is a red-shift in the G-nitroactin emission of 2 nm, which does not occur in F-nitroactin, and at pH 6 there
69
P. D. Chantler and Mi. B. Gratzer
Fig. 1. Electron micrographs of ( A ) F-actin and ( B ) F-nitroactin, negatively stained with uranyl acetate. Magnification x 75 000
is a small blue-shift with a drop in intensity. Evidently the conformational lability responsible for these changes is largely inhibited in the polymerised form. Denaturation with EDTA [21] leads to a red-shift in Gnitroactin, just as in native G-actin, in agreement with the report by Lehrer and Elzinga [7]. The circular dichroism in the aromatic region is profoundly affected by nitration. G-actin shows aromatic Cotton effects [23], with a prominent negative extremum at 295 nm and a shoulder at about 287 nm, both almost certainly due to contributions of tryptophan residues [24]. In the nitrated species the extremum remains apparent only as a shoulder, and a relatively large negative extremum (ellipticity- 3500 deg. cm’ dmol-’) emerges at 284 nm. The difference between the nitrated and the native species shows predominantly positive ellipticity at 296 nm and a negative effect at 285 nm, with lower activity at longer wavelengths. When the ionic strength of the solution of G-nitroactin was increased (50 mM potassium chloride, 2 mM magnesium chloride), birefringence was developed between crossed polaroids on gentle swirling. This occurred more slowly than in native G-actin similarly treated. Centrifugation yielded a yellow pellet and examination of the supernatant showed that the great bulk of the actin was in all cases incorporated into the polymers. Nitration in the manner described does not
destroy the ability of the protein to polymerise. Moreover, the polymers, when examined in the electron microscope, showed the characteristic helical threadlike structure of F-actin (Fig. l), from which it is indistinguishable. When heavy meromyosin is added to the F-nitroactin, sedimentation at 80000 x g for 3 h leads to its complete association with the F-actin pellet. When it is added to F-nitroactin on electron microscope grids, ‘decorated’ filaments with their characteristic arrowhead pattern appear, which again are indistinguishable from those formed by acto-heavy-meromyosin (Fig. 2). Measurements of the fluorescence spectrum of nitroacto-heavy-meromyosin in solution gave a result identical to the sum of the contributions of the F-nitroactin and the heavy meromyosin taken separately. There is thus no measurable transfer of excitation energy from aromatic residues of the myosin heads to the nitrophenol chromophore. The effect of F-nitroactin on the ATPase activity of myosin was tested in the pH-stat. In a typical experiment heavy meromyosin gave a turnover rate of 3.6 mol Pi . mol . protein-’ . min-l. After addition of an equivalent amount of F-nitroactin a value of 4.3 resulted. This is practically within the experimental error, and we conclude that though myosin heads associate with F-nitroactin with high affinity, the ATPase
Chemically Modified Actiii
Fig. 2. Electron micrographs ofjilarnerits of ( A ) F-uctin and ( B ) F-nitroactin, both decorated with heavy rneromjosin. Magnification x 75000
activity is not significantly enhanced and neither is there any inhibition. ATPase measurements were also performed on complexes of heavy meromyosin with polymers produced from mixtures of G-actin and Gnitroactin. In this case ATPase enhancement was observed, but throughout at a lower level than would be produced by the same concentration of actin in the absence of nitroactin. Reciprocal plots reflecting the ATPase stimulation [25]are shown in Fig. 3, and take the form of a set of straight lines intersecting at different points, all in the first quadrant of the graph. This unusual effect is discussed below. The aminotryrosyl actin behaves in most respects similarly to the nitro derivative. The absorption spectrum shows that the nitrophenol has been reduced, but there is a slight trail to the long-wavelength side of the absorption band. The fluorescence emission spectrum resembles that of native G-actin. The aminotyrosyl actin is able to polymerise, and both the bare filaments and those decorated with heavy meromyosin are indistinguishable from the corresponding structures formed by unmodified actin. The ability to activate myosin ATPase is not restored on reduction of the nitro group. The dansylaminoactin showed an absorption spectrum reflecting the presence of the dansyl chromo-
l/F-actin concn (rnlirng)
Fig. 3. Reciprocal plots showing the activation of myosin subfiagment-1 ATPase by F-uctin and copolymers ofuctin andnitroactin. The ordinate is reciprocal of ATPase velocity enhancement over that of subfragment-1 alone, and the abscissa (in inl/mg) refers to the concentration of the native actin fraction. Concentrations of the 0.01 (a),0.02 (0) and 0.03 (m) mg/ml nitroactin component are 0 (o),
phore (absorption maximum 325 nm), and when excited at 325 nm gave a strong fluorescence band at 495 nm. This derivative showed no propensity to polymerise when the ionic strength was increased.
71
P. D. Chantler and W. B. Gratzer
DISCUSSION The nitration of G-actin at Tyr-69 evidently leaves the conformation of the protein essentially undisturbed, as judged by the evidence of sedimentation velocity and circular dichroism. The quenching of the tryptophan emission by the nitrotyrosine chromophore (which is relieved when the latter is reduced to aminotyrosine) is sufficiently extensive that it must be ascribed to excitation transfer from one or more tryptophan residues in close proximity (of the order of 0.5 nm away, with a favourable orientation). A tryptophan close to Tyr-69 in the sequence [I] is Trp-74. The new extrinsic Cotton effects that accompany the introduction of the nitro group are predominantly in the region of the near-ultraviolet tryptophan transitions, rather than at the wavelengths at which only the nitrophenol group absorbs. It therefore seems probable that they arise from interaction of the transition dipoles of a tryptophan and a nitrotyrosine almost in van der Waals contact [26]. As regards the effect of nitration at this low level on the polymerisability of the actin, there is some ambiguity in the literature, one group of workers stating that this capacity remains on nitration (albeit of F-actin [S]), whereas a brief report on nitrated G-actin [7] indicates that polymerising ability is destroyed. The discrepancy inay be due to differences in the reaction conditions, or those of the polymerisation assays; one effect could arise from ionisation of the nitrotyrosine group, for example, which will change with pH in the neutral range. At all events our results are unequivocal in the sense that in our solvent conditions elevation of the ionic strength leads to polymerisation of all the nitroactin. Since the reactivity of Tyr-69 and the absence of any appreciable conformational changes on nitration, indicate that this residue lies on the surface of the G-actin molecule, the failure of nitration to inhibit polymerisation is not especially surprising, and indicates only that Tyr-69 is not directly involved in any of the four probable actin-actin interaction sites [2]. Neither is it surprising that when the nitro group is reduced to the much smaller amino group, the polymerising capacity still survives. On the other hand, the introduction of the dansyl group, which is about 0.8 nm across compared with about 0.3 nm for the nitro group, leads to total inhibition of polymerisation. We cannot determine whether the dansyl group physically obstructs an association site, causes a local conformational disturbance, which perhaps impedes the change normally engendered by salt to trigger the polymerisation, or introduces an unfavourable coulombic contribution from the positive charge. The inhibition of polymerisation by modification at the Tyr-69 position is evidently in any case a finely balanced process. The polymerised nitroactin retains its ability to bind the myosin heads, and there is nothing to indicate
that the strength of binding is in any way diminished. The arrowheads seen in the electron micrographs (Fig. 2) are characteristic of the functional actomyosin interaction. The absence of any accompanying enhancement of the myosin ATPase activity is striking, but not unique, for other chemical modifications, not necessarily of very defined specificity, have been found that do not prevent actomyosin formation but do considerably inhibit the activation of the myosin ATPase [27- 291. (The electron microscopic ‘decoration’ technique was not at that time available to establish whether the interaction was of the ‘native’ type). From the separation of the binding and activation functions it would appear to follow that actin does not activate the myosin ATPase by direct displacement of the bound products of hydrolysis [30]. Indeed, monomeric matrixbound actin in the ionic conditions corresponding to maximum activation by F-actin will also cause little activation [31], and the possibility therefore arises that the interaction with myosin involves two or several actin monomers in the thin filament, one energetically dominant interaction being sufficient for binding, while another is required to elicit ATPase activation. The latter site would be envisaged as involving Tyr-69, which is evidently not a part of the primary binding site. Other explanations would need to invoke small conformational changes in actin associated with the activation process, which would be inhibited by a local chemical modification. There are strong indications [32 - 341 that a tryptophan residue is present in the nucleotide-binding site of the myosin. If the actin attached at or near this site one would expect that energy transfer would occur to the nitrotyrosine group on the actin. An estimate of the lower limit of the distance involved is readily made, with the aid ofthe Forster sixthpower relation for resonance transfer, viz. Rg, the distance for 50 probability of excitation transfer within the lifetime of the excited state of the donor chromophore, is given by [35] : Rg
=
91n lox2 qJ 1 2 8 7 ~n4~N ’
where J is the overlap integral for the emission spectrum of the donor and the absorption spectrum of the acceptor, q the quantum yield of this donor, here taken as 0.14 for a typical protein tryptophan residue [36], n the refractive index of the protein medium, taken as 1.5, N the Avogadro number, and x, the orientation factor, is 2/3 if the two chromophores are in a state of free rotation. The fractional transfer efficiency at distance R is then (Ro/R)6/[1 (Ro/R)6].Since no transfer is seen, it must be supposed that unless the angle between the potential donor and acceptor chroniophores is fixed, and by chalice very close to 90”, these residues are in fact at least 1.5nm or so apart. On the other hand, Harrington and his associates [37,38]
+
72
have assembled strong evidence that the actin-binding site and the ATPase site of myosin are identical or contiguous ; this would be compatible with a location of Tyr-69 remote from the myosin-binding site. A satisfactory description of the actin binding and enzymic activation mechanism is evidently not yet available. The reciprocal plots of Fig. 3 , in which the concentration refers to native actin, are incompatible with a simple inhibitory function, whether competitive or non-competive, of the copolymerised nitroactin. The kinetics of the actomyosin ATPase reaction are much more complex than their representation in terms of a reciprocal plot implies, and Lymn [39] has shown that in the general case such plots will be curved, and for appropriate combinations of the individual rate constants can intersect in the first quadrant of the reciprocal plot [40]. The data shown in Fig. 3 are most readily compatible with a conformational effect, whereby a steric feature of the F-actin helix, or a change in conformation contingent on combination with myosin, is inhibited by the foreign species present in the polymer. Tawada and Oosawa [41] have noted that in copolymers of actin and carboxymethylactin (which activates myosin ATPase with identical apparent V , but lower K,,,), the activation is slightly but consistently less than that of a mixture of the homopolymers. Our effect, which is gross by comparison, may be related to this observation. However, whereas their data can be quantitatively accounted for if it is assumed that two adjacent monomers on the actin helix are responsible for ATPase activation, our results do not appear to be compatible with such a simple scheme. If the fraction of unmodified actin in the copolymer is CI and the concentration in protein monomers is 2,then the effective concentration of unmodified dimers in the chain (in molar units of monomers) will be a2A, of trimers a3A, and so on. We find that oui- data are not very satisfactorily fitted in terms of an active species of unmodified dimer or of trimer. We cannot, however, exclude non-random copolymerisation, which would vitiate this argument. Apart from this reservation, the nature of the reciprocal plots suggests that the conformation of the entire filament is changed in a manner that varies continuously with the proportion of the active copolymeric component, or that there are domains of the correct and incorrect conformations for activation, depending in frequency on the composition of the polymer. Both types of situation have been observed in protein aggregates (see, e. g., Asakura et al. [42]). We are grateful to R. Craig for help with electron microscopy, R. Hynes for amino acid analysis, and the Science Research Council for a Training Scholarship for P.D.C.
P. D. Chantler and W. B. Gratzer: Chemically Modified Actin
REFERENCES 1. Elzinga, M., Collins, J. H., Kuehl, W. M. & Adelstein, R. S. (1973) Proc. Nut1 Acud. Sci. U.S.A. 70, 2687-2691. 2. Oosawa, F. & Kasai, M. (1971) in Subunits in Biological Systems, ( S . N. Timasheff & G. D. Fasman, eds) part A, pp. 261 - 322, M. Dekker, New York. 3. Ebashi, S. (1974) Essays Biochem. 10, 1 - 36. 4. Miihlrad, A,, Corsi, A. & Granata, A. L. (1968) Biochim. Biophys. Acta, 162, 435 - 443. 5. Tawada, K., Asai, H. & Gerber, B. R. (1969) Biochim. Biophys. Acta. 194,486 - 497. 6. Cheung, H. C., Cooke, R. & Smith, L. (1971) Arch. Biochem. 142, 333- 339. 7. Lehrer, S . S . & Elzinga, M. (1972) Fed. Proc. 31, 502. 8. Riordan, J. F. & Vallee, B. L. (1972) in Methods Enzy.mo1. 25B, 515-521. 9. Martonosi, A. (1962) J . B i d . Chem. 237, 2795-2803. 10. Spudich, J. A. & Watt, S. (1971)J. Bid. Chem. 246,4866-4871. 11. Weber, K. & Osborn, M. (1969) J. Biol. Chem. 244,4406-4412. 12. Rees, M. K. & Young, M. (1967) J . Biol. Chem. 242, 44494458. 13. Perry, S. V. (1955) Methods Enzymol. 2, 582-588. 14. Lowey, S. & Cohen, C. (1962) J. Mol. Biol. 4, 293-308. 15. Reisler, E., Burke, M. & Harrington, W. F. (1974) Biochemistry, 13,2014-2022. 16. Ellman, G. L. (1959) Arch. Biochem. 82, 70-77. 17. Huxley, H. E. (1963) 1. Mol. Bid. 7, 281 - 308. 18. Beaven, G. H. & Gratzer, W. B. (1968) Biochim. Biophys. Acta, 168,456-462. 19. Sokolovsky, M., Riordan, J. F. & Vallee, B. L. (1967) Biochem. Biophys. Res. Commun. 27,20-25. 20. Kenner, R. A. & Neurath, H. (1971) Biochemistry, 10, 551 557. 21. Lusty, C. J. & Fasold, H. (1969) Biochemistry, 8, 2933-2940. 22. Lehrer, S. S. & Kerwar, G. (1972) Biochemistry, 11,1211- 1217. 23. Murphy, A. J. (1971) Biochemistry, 10, 3723-3727. 24. Strickland, E. H., Horwitz, J. & Billups, C. (1969) Biochemistry, 8, 3205-3213. 25. Eisenberg, E. & Moos, C. (1968) Biochemistry, 7, 1486-1489. 26. Tinoco, I. (1963) Radiat. Res., 20, 133-139. 27. Birany, M. & Barany, K. (1959) Biochim. Biophys. Acta, 35, 293 - 309. 28. Perry, S. V. & Cotterill, J. (1964) Biochem. J . 92, 603-608. 29. Kaldor, G., Gitlin, J., Westley, F. & Volk, B. W. (1964) Biochemistry. 3, 1137-1145. 30. Taylor, E. W. (1972) Annu. Rev. Biochem. 41, 577-616. 31. Chantler, P. D. & Gratzer, W. B. (1973) FEBS Lett. 34,lO- 14. 32. Yoshino, H., Morita, F. & Yagi, K. (1972)J. Biochem. (Tokyo) 72, 1227- 1235. 33. Werber, M. M., Szent-Gyorgyi, A. G . & Fasman, G. D . (1972) Biochemistry, 11, 2872-2882. 34. Mandelkow, E. M. & Mandelkow, E. (1973) FEBS Lett. 33, 161- 166. 35. Forster, T. (1959) Disc. Furaduy Soc. 27, 7- 17. 36. Eisinger, J., Feuer, B. & Lamola, A. A. (1969) Biochemistry, 8, 3908- 3915. 37. Burke, M., Reisler, E. & Harrington, W. F. (1973) Proc. Natl Acad. Sci. U.S.A. 70, 3793-3796. 38. Reisler, E., Burke, M. & Harrington, W. F. (1974) Biochemistry, 13,2014-2022. 39. Lymn, R. W. (1974) J. Theor. Biol. 43, 313-328. 40. Lymn, R. W. (1975) J. Theor. Biol. 49, 425-429. 41. Tawada, K. & Oosawa, F. (1969) J . Mol. Biol. 44, 309-317. 42. Asakura, S. & Iino, T. (1972) J. Mol. Biol. 64, 251-268.
P. D. Chantler, Department of Biology, Brandeis University, Waltham, Massachusetts, U.S.A. 02154 W. B. Gratzer, Medical Research Council Cell Biophysics Unit, King’s College, 26-29 Drury Lane, London, Great Britain WC2B 5RL
