Actin-Latrunculin A Structure and Function
2000; Elsevier BV; Volume: 275; Issue: 36 Linguagem: Inglês
10.1074/jbc.m004253200
ISSN1083-351X
AutoresElena G. Yarmola, Thayumanasamy Somasundaram, Todd A. Boring, Ilan Spector, M Bubb,
Tópico(s)Advanced Fluorescence Microscopy Techniques
ResumoLatrunculin A is used extensively as an agent to sequester monomeric actin in living cells. We hypothesize that additional activities of latrunculin A may be important for its biological activity. Our data are consistent with the formation of a 1:1 stoichiometric complex with an equilibrium dissociation constant of 0.2 to 0.4 μm and provide no evidence that the actin-latrunculin A complex participates in the elongation of actin filaments. Profilin and latrunculin A bind independently to actin, whereas binding of thymosin β4 to actin is inhibited by latrunculin A. Potential implications of this differential effect on actin-binding proteins are discussed. From a structural perspective, if latrunculin A binds to actin at a site that sterically influences binding by thymosin β4, then the observation that latrunculin A inhibits nucleotide exchange on actin implies an allosteric effect on the nucleotide binding cleft. Alternatively, if, as previously postulated, latrunculin A binds in the nucleotide cleft of actin, then its ability to inhibit binding by thymosin β4 is a surprising result that suggests that significant allosteric changes affect the thymosin β4 binding site. We show that latrunculin A and actin form a crystalline structure with orthorhombic space group P212121and diffraction to 3.10 Å. A high resolution structure with optimized crystallization conditions should provide insight regarding these remarkable allosteric properties. Latrunculin A is used extensively as an agent to sequester monomeric actin in living cells. We hypothesize that additional activities of latrunculin A may be important for its biological activity. Our data are consistent with the formation of a 1:1 stoichiometric complex with an equilibrium dissociation constant of 0.2 to 0.4 μm and provide no evidence that the actin-latrunculin A complex participates in the elongation of actin filaments. Profilin and latrunculin A bind independently to actin, whereas binding of thymosin β4 to actin is inhibited by latrunculin A. Potential implications of this differential effect on actin-binding proteins are discussed. From a structural perspective, if latrunculin A binds to actin at a site that sterically influences binding by thymosin β4, then the observation that latrunculin A inhibits nucleotide exchange on actin implies an allosteric effect on the nucleotide binding cleft. Alternatively, if, as previously postulated, latrunculin A binds in the nucleotide cleft of actin, then its ability to inhibit binding by thymosin β4 is a surprising result that suggests that significant allosteric changes affect the thymosin β4 binding site. We show that latrunculin A and actin form a crystalline structure with orthorhombic space group P212121and diffraction to 3.10 Å. A high resolution structure with optimized crystallization conditions should provide insight regarding these remarkable allosteric properties. Latrunculin A, isolated from the Red Sea sponge Negombata magnifica, was initially identified as an inhibitor of actin polymerization by its morphological effects and by the effects it had on actin filament distribution in cultured nonmuscle cells (1Spector I. Shochet N.R. Kashman Y. Groweiss A. Science. 1983; 214: 493-495Crossref Scopus (625) Google Scholar). Based on the effects of latrunculin A on the steady state level of F-actinin vitro, the effects of the drug were thought to be consistent with sequestration of monomeric actin in a 1:1 molar complex with equilibrium dissociation constant of 0.2 μm (2Coué M. Brenner S.L. Spector I. Korn E.D. FEBS Lett. 1987; 213: 316-318Crossref PubMed Scopus (661) Google Scholar). The binding site of latrunculin has not been conclusively identified, but based on the study of the effects of specific mutations of yeast actin on latrunculin A binding, it has been inferred that latrunculin A may bind to actin near or in its nucleotide binding cleft (3Ayscough K.R. Stryker J. Pokala N. Sanders M. Crews P. Drubin D.G. J. Cell Biol. 1997; 137: 399-416Crossref PubMed Scopus (644) Google Scholar, 4Belmont L.D. Patterson G.M.L. Drubin D.G. J. Cell Sci. 1999; 112: 1325-1336PubMed Google Scholar). The observation that latrunculin affects nucleotide exchange has been offered as support of this conclusion (3Ayscough K.R. Stryker J. Pokala N. Sanders M. Crews P. Drubin D.G. J. Cell Biol. 1997; 137: 399-416Crossref PubMed Scopus (644) Google Scholar). These data, however, are inconclusive in light of the fact that many actin-binding proteins with binding sites that are spatially distant from the nucleotide cleft are also able to affect nucleotide exchange (5Safer D. Sosnick T.R. Elzinga M. Biochemistry. 1997; 36: 5806-5816Crossref PubMed Scopus (97) Google Scholar) and that actin demonstrates several additional allosteric properties that serve as a precedent for the transmission of structural alterations to distant sites (6Kuznetsova I. Antropova O. Turoverov K. Khaitlina S. FEBS Lett. 1996; 383: 105-108Crossref PubMed Scopus (44) Google Scholar, 7Prochniewicz E. Thomas D.D. Biochemistry. 1997; 36: 12845-12853Crossref PubMed Scopus (44) Google Scholar, 8De La Cruz E.M. Ostap E.M. Brundage R.A. Reddy K.S. Sweeney H.L. Safer D. Biophys. J. 2000; 78: 2516-2527Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar). When latrunculin A is employed in studies of cell biology, the observed effects are consistent with depolymerization of actin filaments consequent to sequestration of monomeric actin by latrunculin (9Spector I. Shocet N.R. Blasberger D. Kashman Y. Cell Motil. Cytoskeleton. 1989; 13: 127-144Crossref PubMed Scopus (487) Google Scholar). A previous preliminary report (2Coué M. Brenner S.L. Spector I. Korn E.D. FEBS Lett. 1987; 213: 316-318Crossref PubMed Scopus (661) Google Scholar) did not rule out the possibility that latrunculin A has effects related to the polymerization of actin in addition to monomer sequestration, and these possibilities are explored in our current studies. Other effects of latrunculin A on the cytoskeleton are possible, however, and evidence has been reported that latrunculin can affect the expression of actin and possibly of other actin-binding proteins by a feedback mechanism that may sense the cellular concentration of actin monomers, resulting in more complicated outcomes than that predicted by monomer sequestration alone (10Bershadsky A.D. Gluck U. Denisenko O.N. Sklyarova T.V. Spector I. Ben-Zéev A. J. Cell Sci. 1995; 108: 1183-1193Crossref PubMed Google Scholar). To characterize the surface interactions of latrunculin A and actin, we examined whether latrunculin A affected the interaction of actin with other actin-monomer-binding proteins. To our surprise, latrunculin A inhibited binding by thymosin β4 but not binding by profilin or DNase I. Because thymosin β4 has been postulated to perform functions related to wound healing (11Frohm M. Gunne H. Bergman A.C. Agerberth B. Bergman T. Boman A. Liden S. Jornvall H. Boman H.G. Eur. J. Biochem. 1996; 237: 86-89Crossref PubMed Scopus (195) Google Scholar), apoptosis (12Niu M. Nachmias V.T. Cell Adhes. Commun. 2000; 7: 311-320Crossref PubMed Scopus (36) Google Scholar), and the inflammatory response (13Young J.D. Lawrence A.J. MacLean A.G. Leung B.P. McInnes I.B. Canas B. Pappin D.J.C. Stevenson R.D. Nat. Med. 1999; 5: 1424-1427Crossref PubMed Scopus (171) Google Scholar), augmentation of the concentration of free thymosin β4 by latrunculin A could potentiate these responses. Our results imply that actin-binding marine natural products may have effects other than those predicted solely by their effects on actin polymerization and, by inference, that marine natural products may exist that affect actin-binding protein function without directly affecting actin polymerization. Finally, our results illustrate a novel mechanism by which pharmacological agents that bind actin could be used to modulate the function of actin-binding proteins. Rabbit skeletal muscle actin was prepared from frozen muscle (Pel-Freez, Rogers, AR) in Buffer G (5.0 mmTris, 0.2 mm ATP, 0.2 mm dithiothreitol, 0.1 mm CaCl2, and 0.01% sodium azide, pH 7.8) (15Kang F. Laine R.O. Bubb M.R. Southwick F.S. Purich D.L. Biochemistry. 1997; 36: 8384-8392Crossref PubMed Scopus (107) Google Scholar), and pyrenyl-actin (actin labeled on Cys-374 withN-(1-pyrene)iodoacetamide) was prepared with 0.7–0.95 mol of label/mol of protein using the method of Kouyama and Mihashi (14Kouyama T. Mihashi K. Eur. J. Biochem. 1981; 114: 33-38Crossref PubMed Scopus (760) Google Scholar). Recombinant human profilin was purified as described previously (15Kang F. Laine R.O. Bubb M.R. Southwick F.S. Purich D.L. Biochemistry. 1997; 36: 8384-8392Crossref PubMed Scopus (107) Google Scholar). Beef pancreatic DNase I (molecular biology grade; Worthington Biochemical Corp., Freehold, NJ) was reconstituted from lyophilized powder. Rat thymosin β4cDNA (which codes for an amino acid sequence identical to that of human thymosin β4) in a pcDNA3 (Invitrogen, Carlsbad, CA) vector was a gift from Dr. Vivianne Nachmias (University of Pennsylvania School of Medicine). Oligonucleotides were designed so as to add a cysteine residue to the C terminus, and both strands of the cloned products were sequenced to verify the outcome. After cloning into an pET-12a expression vector, the Escherichia colistrain BL21(DE3) was transformed with plasmid. Latrunculin A was stored as a 2 or 10 mm stock in Me2SO and was diluted to 100 μm in Buffer G for the in vitroexperiments. Cells containing wild-type or cysteine-modified thymosin β4 constructs were grown at 37 °C in M9 medium and harvested 3 h after induction with 1 mmisopropyl β-d-thiogalactopyranoside. Cell pellets were dissolved in 0.5 m cooled perchloric acid, sonicated for 2 min in ice, and centrifuged for 30 min at 4 °C (130,000 × g). The supernatant was adjusted to pH 7.0–7.5 with KOH and centrifuged to remove KClO4. After the adjustment of pH to 4.0 with formic acid, the supernatant was rapidly heated to 80 °C for 10 min, chilled on ice for 10 min, centrifuged for 30 min at 4 °C, dialyzed against 20 mmformic acid, pH 4.0, and loaded on a SP-Hi Trap column (Amersham Pharmacia Biotech). The thymosin β4 was eluted with a linear gradient of NaCl (0–2 m) in 20 mm formic acid, pH 4.0). The fractions were neutralized with 2M Tris base as soon as eluted and dialyzed against P buffer (5 mm Tris-HCl, 40 mm KCl, 0.2 mmdithiothreitol, 0.02% sodium azide, pH 7.9). Thymosin β4concentration was determined using the BCA protein assay (Bio-Rad). For labeling, thymosin β4 was dialyzed in 50 mm sodium borate buffer, and then tetramethylrhodamine-5-maleimide (T-6027, Molecular Probes Inc.) was added in four aliquots to a final molar ratio of dye to thymosin β4 of 2:1. After 8 h of stirring at 33–34 °C, the sample was chilled on ice and left overnight. The reaction was stopped by addition of dithiothreitol, and the sample was dialyzed against P buffer. Thymosin β4 was then gel-filtered with Superose-12 column, and the concentration of thymosin β4was determined by BCA protein concentration assay. Extent of labeling was determined using extinction coefficients for dye of ε541 = 115 mm−1 and ε280 = 32.5 mm−1. Modification of the C terminus with addition of an acetylated cysteine has previously been shown not to affect the actin binding properties of thymosin β4 (16Carlier M.-F. Didry D. Erk I. Lepault J. Van Troys M.L. Vanderkerckhove J. Perelroizen I. Yin H. Doi Y. Pantaloni D. J. Biol. Chem. 1996; 271: 9231-9239Abstract Full Text Full Text PDF PubMed Scopus (70) Google Scholar). Actin (4% pyrenyl-labeled) was converted to Mg2+-actin by the addition of 125 μm EGTA and 50 μmMgCl2, and after 15 min, it was polymerized by the addition of MgCl2 to a final concentration of 2.0 mmwith varying KCl (or at 10 mm KCl and varying latrunculin A). Individual steady state samples were prepared by dilution of 10 μm F-actin without a change in buffer conditions, and steady state fluorescence readings were obtained at 24 h as described previously (17Bubb M. Spector I. Beyer B.B. Fosen K.M. J. Biol. Chem. 2000; 275: 5163-5170Abstract Full Text Full Text PDF PubMed Scopus (428) Google Scholar). Equilibrium dissociation constants were calculated assuming that the x intercepts reflected the total amount of unpolymerized actin, either as monomer or as a complex of latrunculin A and actin. The analysis assumes that fluorescence intensity is proportional to F-actin concentration. A seeded polymerization assay using gel-filtered cross-linked F-actin seeds was used to measure elongation rates of 4.0 μmMg2+-actin as described previously (17Bubb M. Spector I. Beyer B.B. Fosen K.M. J. Biol. Chem. 2000; 275: 5163-5170Abstract Full Text Full Text PDF PubMed Scopus (428) Google Scholar). Preliminary data confirmed that the initial rate of polymerization was proportional to both the concentration of added seeds and to the concentration of free actin. Excess free ATP was removed using AG 1-X8 anion exchange resin (Bio-Rad) as described previously (18Mockrin S.C. Korn E.D. Biochemistry. 1980; 19: 5359-5362Crossref PubMed Scopus (210) Google Scholar). Actin (1.7 μm) and profilin (0.2 μm) were incubated in a glass cuvette with Buffer G without ATP and various concentrations of latrunculin A. A mixture of εATP and KCl (final concentrations, 3.37 μm and 50 mm, respectively) was added to start the reaction. After mixing, samples were placed in spectrofluorometer, and the time course of fluorescence changes was recorded (15Kang F. Laine R.O. Bubb M.R. Southwick F.S. Purich D.L. Biochemistry. 1997; 36: 8384-8392Crossref PubMed Scopus (107) Google Scholar). Exchange rates were obtained by fitting the time course to a single exponential. Data were then fit to the following equilibrium dissociation constants:K d P, for profilin to actin,K d L for latrunculin A to actin, andK d LP for profilin to the complex of actin and latrunculin A, and also to k A,k AP, k AL, andk ALP, the rate constants of ATP dissociation from actin, profilin, actin-latrunculin A, and actin-latrunculin A-profilin ternary complex, respectively. Actin at a concentration of 2.9 μm was incubated in Buffer G with or without 3.4 μm thymosin β4 in the presence and absence of 40 μm latrunculin A. Solutions were incubated for 35 min before loading on gel. Native gels were equilibrated in buffer containing 0.1 mm CaCl2, 0.01% sodium azide, 0.2 mm ATP, 0.2 mm dithiothreitol, and 25 mm Tris-Tricine, pH 8.2. In experiments with labeled thymosin β4, the picture of the fluorescent gel was taken before staining with Coomassie. Data were collected on a Photon Technology International (South Brunswick, NJ) spectrofluorometer. Tetramethylrhodamine-5-maleimide-labeled thymosin β4 was excited with vertically polarized light at 546 nm. The horizontal and vertical components of the emitted light were detected at 568 nm. Solutions of labeled thymosin β4 (0.1 μm) in Buffer G were titrated with Mg2+-actin in the presence or absence of a constant amount of latrunculin A (or with latrunculin A in the presence or absence of a constant amount of Mg2+-actin). Data were fit globally as described by Vinson et al. (19Vinson V.K. De La Cruz E.M. Higgs H.N. Pollard T.D. Biochemistry. 1998; 37: 10871-10880Crossref PubMed Scopus (131) Google Scholar), with the inclusion of a term for the formation of a ternary complex between actin, latrunculin, and thymosin. Fitting parameters included the equilibrium dissociation constants for thymosin β4 to actin (K d T), for latrunculin A to actin (K d L), and for thymosin β4 to the complex of actin and latrunculin A (K d LT) and the terms indicating the anisotropy of free thymosin β4 (r f ) and the anisotropy of the complex of thymosin β4 with actin or with actin-latrunculin A complex (r b ). Assuming that the concentration of free thymosin, [T], is low relative to K d T, (or strictly, [T]/K d T ≪ (1+ [L]/K L), equations for the observed fluorescence anisotropy, r, can be written as a function of the total actin, [A] t , and total latrunculin A, [L] t , concentrations as follows, r=rf+(rb−rf)(1−KdLT[A]([L]t/([A]+KdL)+KdLT/KdT)+KdLT)Equation 1 [A]=((KdL−[A]t+[L]t)2+4[A]tKdL)1/2−(KdL−[A]t+[L]t)2Equation 2 where [A] is free actin concentration. Sedimentation equilibrium experiments were performed using absorption optics with data collected at 535 nm (the absorption maximum for labeled thymosin β4) in a Beckman XLA centrifuge. All samples contained 1.6 μm labeled thymosin β4. Samples of 110 μl in Buffer G reached equilibrium in 42 h at 13,900 rpm (after initially overspeeding to 15,100) at 4 °C. Buffer density was determined by pyknometry, and partial specific volumes were as previously reported for actin or calculated from amino acid sequence for thymosin β4 (20Kirschner M.W. Schachman H.K. Biochemistry. 1971; 10: 1900-1925Crossref PubMed Scopus (51) Google Scholar). The gradient was analyzed according to a method of implicit constraints as described previously (21Bubb M.R. Lewis M.S. Korn E.D. J. Biol. Chem. 1991; 266: 3820-3826Abstract Full Text PDF PubMed Google Scholar). In brief, at 535 nm, only labeled thymosin β4 has a measurable extinction coefficient. The other sample components are invisible. Therefore, at this wavelength, the absorbance at any radius is directly proportional to the sum of the concentration of all allowable thymosin β4-containing species (in a model of noncompetitive inhibition, these include thymosin β4, thymosin β4 bound to actin, and thymosin β4-actin-latrunculin A ternary complex). The species are assumed to be in chemical equilibria at all radii as governed by appropriate equilibrium dissociation constants. Curve fitting is constrained by the initial concentration of all components, and the fitting parameters include only the dissociation constants and the concentration of each component at an arbitrary radius,r b (21Bubb M.R. Lewis M.S. Korn E.D. J. Biol. Chem. 1991; 266: 3820-3826Abstract Full Text PDF PubMed Google Scholar). DNase I (30 nm) was incubated with actin (30 nm) with varying concentrations of latrunculin A for 5 min at room temperature before adding 100 μg/ml DNA. The reaction mixture was in buffer containing 21 mm NaCl, 0.1 mm CaCl 2, 2.0 mm MgCl2, 0.1 mm ATP, and 5 mm Tris, pH 7.9. After 20 min, samples were loaded on 0.7% agarose gels, and the gels were subsequently stained with ethidium bromide. Crystals were grown in hanging droplets containing 1.3–1.5 mammonium sulfate, 3 mm MgCl2, 60 mmimidazole, pH 6.7, with actin concentration at 9 mg/ml and a ratio of 1:1 or 1.2:1 of latrunculin A to actin. Crystals appeared after 2–4 weeks at room temperature. Seemingly identical crystallization conditions produced satisfactory crystals only in approximately 50% of attempts. Crystals with typical dimensions 0.4 × 0.5 × 0.4 mm were wet mounted on glass capillaries at room temperature. The data were collected using Cu K∝ radiation (λ = 1.541Å) from a Rigaku RU-200 x-ray generator (40 kV, 90 mA, 0.3 mm collimator). The generator was equipped with an R-Axis IIc image plate, and data were collected at a crystal-IP distance of 100 mm with an oscillation range of 1.1 degree/frame. Data were integrated and scaled using the two companion programs of the HKL Suite, Denzo and Scalepack (22Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref PubMed Scopus (38782) Google Scholar). Stock 50 or 60% Ficoll solutions in crystallization buffer, prepared according to Ref. 23Westbrook E.M. Methods Enzymol. 1985; 114: 187-196Crossref PubMed Scopus (84) Google Scholar, were mixed in the desired ratios with crystallization buffer to prepare solutions of various Ficoll concentrations. The density of each solution was calculated from the mass of 0.4 ml of solution as measured with a positive-displacement pipette. Ammonium sulfate was used to vary the buffer density for a given concentration of Ficoll. The technique relies on the assumption that ammonium sulfate, but not Ficoll, rapidly enters the solute component of the crystal. Crystals were layered on top of a step gradient of Ficoll in a 6-mm-diameter glass cuvette and immediately centrifuged at 8000 × g for 2 min. After the positions of crystals were located, the cuvette was centrifuged for an additional 1 min to check for changes in position. In control experiments, centrifugation time varied from 1 to 10 min. Some crystals were lightly cross-linked in 0.15% glutaraldehyde for 12 h at room temperature prior to density measurements. Whereas uncross-linked crystals were stable for only 10–15 min at low solvent density, the crystals were stable (with constant density) for several hours after covalent cross-linking. Crystal density as a function of the partial specific volume of the protomer was then calculated as described previously (23Westbrook E.M. Methods Enzymol. 1985; 114: 187-196Crossref PubMed Scopus (84) Google Scholar). The calculated concentration of latrunculin A-actin complex was proportional to the concentration of latrunculin A, consistent with monomer sequestration (Fig.1, inset). The calculated equilibrium dissociation constant (K d L) is similar to that previously reported (K d L = 0.2 μm in very low ionic strength buffer containing 0.1 mm CaCl2 and 2.0 mmMgCl2) (2Coué M. Brenner S.L. Spector I. Korn E.D. FEBS Lett. 1987; 213: 316-318Crossref PubMed Scopus (661) Google Scholar). Latrunculin A did not cause any significant differences in the slopes of the curves for fluorescenceversus actin concentration relative to controls, consistent with 1) the absence of actin-filament capping activity (24Terry D.R. Spector I. Higa T. Bubb M.R. J. Biol. Chem. 1997; 272: 7841-7845Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar), and 2) similar binding affinity to pyrenyl-actin and unlabeled actin (25Lal A.A. Korn E.D. J. Biol. Chem. 1985; 260: 10132-10138Abstract Full Text PDF PubMed Google Scholar). The slight increase (or perhaps absence of change) in affinity at increasing ionic strength implies that electrostatic interactions contribute insignificantly to binding (Fig. 1). Unlike the actin-monomer-binding protein, profilin, elongation data for actin in the presence of latrunculin A can be explained by monomer sequestration alone withK d L of 0.22 ± 0.06 μm (Fig. 2). Notably, although the data suggest that monomer sequestration is the most simple explanation for these data, they fail to prove that latrunculin A-actin complex does not participate in elongation, as any of a number of more complicated models are plausible in which the complex adds and dissociates in a nonproductive manner. Nucleotide exchange on actin was used to indirectly assess the interaction of latrunculin A with actin in the presence or absence of profilin (profilin, by itself, accelerates nucleotide exchange on actin (18Mockrin S.C. Korn E.D. Biochemistry. 1980; 19: 5359-5362Crossref PubMed Scopus (210) Google Scholar)). Latrunculin A alone inhibited nucleotide exchange on 1.7 μm Mg2+-actin, as previously reported (Fig.3, inset) (3Ayscough K.R. Stryker J. Pokala N. Sanders M. Crews P. Drubin D.G. J. Cell Biol. 1997; 137: 399-416Crossref PubMed Scopus (644) Google Scholar). Latrunculin A also inhibited nucleotide exchange in the presence of profilin (Fig.3), implying either that latrunculin A bound competitively with profilin to actin or that nucleotide exchange on actin was inhibited in the ternary complex of latrunculin A, profilin, and actin. Quantitative evaluation of the data eliminated the possibility that binding was competitive. Consistent with fluorescence anisotropy data (data not shown), profilin binds to actin with equilibrium dissociation constant,K d P, of 0.1 μm under these experimental conditions. Assuming thisK d P and a model of competitive binding, the data could not be fit by any possible combination of binding constants of latrunculin A to actin and exchange rates for the complexes of profilin-actin and latrunculin A-actin (Fig. 3,dashed line). Also, the best possible fit required an unreasonable K d L for latrunculin A-actin (0.009 μm) when compared with the other results reported here. In contrast, a model in which profilin and latrunculin A bound independently to actin provided a good fit to the experimental data and yielded a reasonable K d L for latrunculin A-actin (0.28 μm; in TableI, the large error estimate forK d L in the nucleotide exchange experiment relative to the other experimental methods reported here is due to the large number of fitting parameters). Moreover, the best possible global fit to the data was achieved when latrunculin A and profilin were allowed to interact cooperatively with actin, so that the affinity of latrunculin A for actin was increased by a factor of 1.8 when profilin was bound. The fit achieved by allowing this minor degree of positive cooperativity (Hill coefficient of 1.1) was not significantly improved in comparison to a more simple, independent binding model. We conclude that the data rule out competitive binding, but the presence of either slight positive cooperativity or no cooperativity can plausibly explain the experimental results.Table IEquilibrium dissociation constants organized by experimental techniqueType of experimentK d LK d TK d LTK d P1-aK P defined from fluorescence anisotropy experiments (data not shown).μmSteady state (Fig.1)0.15–0.22Seeded polymerization (Fig.2)0.22 ± 0.06Nucleotide exchange (Fig. 3)0.28 ± 0.280.1Fluorescence anisotropy (Fig. 5)0.35 ± 0.050.23 ± 0.027.65 ± 0.74Seeded polymerization (Fig. 5)0.40 ± 0.050.31 ± 0.03Analytical ultracentrifugation (Fig. 5)0.52 ± 0.180.92 ± 0.288.0 ± 1.9Equilibrium dissociation constants are listed for binding latrunculin A to actin (K d L), thymosin β4 to actin (K d T), thymosin β4 to actin saturated with latrunculin A (K d LT), and profilin to actin (K d P). Data used for calculation of each constant are shown in the figures indicated. Error estimates are based on a standard least squares deviation fitting algorithm with 95% confidence limits.1-a K P defined from fluorescence anisotropy experiments (data not shown). Open table in a new tab Equilibrium dissociation constants are listed for binding latrunculin A to actin (K d L), thymosin β4 to actin (K d T), thymosin β4 to actin saturated with latrunculin A (K d LT), and profilin to actin (K d P). Data used for calculation of each constant are shown in the figures indicated. Error estimates are based on a standard least squares deviation fitting algorithm with 95% confidence limits. For the nucleotide exchange experiments, profilin concentration was constant (0.2 μm) and not saturating; therefore, the exchange rate with no latrunculin A (0.023 s−1) is not equal tok AP; rather, k AP was obtained as the best global fit to the data. Assuming noncompetitive binding, the best global fit for the exchange rate constants was obtained with k AP = 0.137 s−1, k A = 0.0015 s−1, k AL = 0.0003 s−1, and k ALP = 0.0011 s−1. Values for k A andk AP are consistent with previous reports (26Selden L.A. Kinosian H.J. Estes J.E. Gershman L.C. Biochemistry. 1999; 38: 2769-2778Crossref PubMed Scopus (67) Google Scholar). The fit curves are insensitive to relatively large changes ink AL and k ALP, and these parameters cannot be distinguished from 0 by the given data. The addition of latrunculin A to samples containing mixtures of thymosin β4 and actin caused less actin to shift to a band corresponding to a high electrophoretic mobility complex of thymosin β4 and actin, implying that the complex is dissociated by latrunculin A (Fig.4, compare lanes 3 and4). Similarly, less fluorescently labeled thymosin β4 shifted to the band corresponding to this complex in the presence of latrunculin A (Fig. 4, compare lanes 1and 2). Previous results have suggested that the extent of binding seen in this assay may not quantitatively reflect the apparentK d for thymosin β4 and actin (27Safer D. Golla R. Nachmias V.T. Biochemistry. 1990; 87: 2536-2540Google Scholar), perhaps because of excluded volume effects, but changes in the amount of shifted protein are qualitatively indicative of the extent of formation of a thymosin β4-actin complex. The data also show that unlabeled thymosin β4 bound as well to actin as labeled thymosin β4 and that binding to actin was inhibited by latrunculin A to the same extent as covalently labeled thymosin β4 (Fig. 4, lanes 7–10). Fluorescence anisotropy of labeled thymosin β4 increased from 0.08 when free to 0.18 when saturated with actin. The anisotropy was lower in the presence of latrunculin A than in its absence at any given actin concentration, indicating that latrunculin A inhibits binding of thymosin β4 to actin (Fig.5 A, top panel). Increasing latrunculin A at fixed actin concentration caused dissociation of thymosin β4 from actin (Fig.5 A, bottom panel); if binding was independent (that is, if K d T is equal to the equilibrium dissociation constant for binding of thymosin β4 to latrunculin A-actin complex,K d LT), then these curves would be flat. In contrast, the best global fit to all four data sets was obtained with K d T = 0.23 ± 0.02 μm, K d L = 0.35 ± 0.05 μm, and K d LT = 7.65 ± 0.74 μm. Therefore, this assay implies inhibition by latrunculin A, with approximately 33 times (the ratio ofK d LT toK d T) weaker affinity of thymosin β4 for actin when latrunculin A is bound to actin. Measurement of the elongation rate after seeding actin polymerization provides additional information regarding the interaction of latrunculin A, thymosin β4, and actin (Fig.5 B). The data obtained for equivalent amounts of latrunculin A and thymosin β4 were nearly indistinguishable, therefore implying that the effects of thymosin β4 on filament elongation, like those of latrunculin A, can be explained by a simple model of monomer sequestration and that the binding constantsK d L andK d T are simi
Referência(s)