Portal de Periódicos da CAPES

Mutational Analysis of the Switch II Loop ofDictyostelium Myosin II

1998; Elsevier BV; Volume: 273; Issue: 32 Linguagem: Inglês

10.1074/jbc.273.32.20334

1083-351X

Naoya Sasaki, Takashi Shimada, Kazuo Sutoh,

Cellular Mechanics and Interactions

A loop comprising residues 454–459 ofDictyostelium myosin II is structurally and functionally equivalent to the switch II loop of the G-protein family. The consensus sequence of the "switch II loop" of the myosin family is DIXGFE. In order to determine the functions of each of the conserved residues, alanine scanning mutagenesis was carried out on theDictyostelium myosin II heavy chain gene. Examination ofin vivo and in vitro motor functions of the mutant myosins revealed that the I455A and S456A mutants retained those functions, whereas the D454A, G457A, F458A and E459A mutants lost them. Biochemical analysis of the latter myosins showed that the G457A and E459A mutants lost the basal ATPase activity by blocking of the isomerization and hydrolysis steps of the ATPase cycle, respectively. The F458A mutant, however, lost the actin-activated ATPase activity without loss of the basal ATPase activity. These results are discussed in terms of the crystal structure of the Dictyosteliummyosin motor domain. A loop comprising residues 454–459 ofDictyostelium myosin II is structurally and functionally equivalent to the switch II loop of the G-protein family. The consensus sequence of the "switch II loop" of the myosin family is DIXGFE. In order to determine the functions of each of the conserved residues, alanine scanning mutagenesis was carried out on theDictyostelium myosin II heavy chain gene. Examination ofin vivo and in vitro motor functions of the mutant myosins revealed that the I455A and S456A mutants retained those functions, whereas the D454A, G457A, F458A and E459A mutants lost them. Biochemical analysis of the latter myosins showed that the G457A and E459A mutants lost the basal ATPase activity by blocking of the isomerization and hydrolysis steps of the ATPase cycle, respectively. The F458A mutant, however, lost the actin-activated ATPase activity without loss of the basal ATPase activity. These results are discussed in terms of the crystal structure of the Dictyosteliummyosin motor domain. In the Dictyostelium motor domain designated as S1dC (1Itakura S. Yamakawa H. Toyoshima Y.Y. Ishijima A. Kojima T. Harada Y. Yanagida T. Wakabayashi T. Sutoh K. Biochem. Biophys. Res. Commun. 1993; 196: 1504-1510Crossref PubMed Scopus (107) Google Scholar), a bound nucleotide is surrounded by three loops whose sequences are highly conserved among the myosin family (2Mooseker M.S. Cheney R.E. Annu. Rev. Cell Dev. Biol. 1995; 11: 633-675Crossref PubMed Scopus (384) Google Scholar): the P-loop (residues 179–186 of Dictyostelium myosin II) and the two loops in the 50K segment (residues 233–240 and 454–459 ofDictyostelium myosin II) (see Fig. 1 A). One of the loops in the 50K segment (residues 233–240) is homologous to a loop in the switch I region of GTPases judging from the topological similarity (3Smith C.A. Rayment I. Biophys. J. 1996; 70: 1590-1602Abstract Full Text PDF PubMed Scopus (213) Google Scholar) and has the consensus sequence NXNSSRFG (NNNSSRFG in Dictyostelium myosin II). Residues in the loop are aligned along the ATPase pocket, and some of the side chains form hydrogen bonds with the bound nucleotide. The other loop in the 50K segment has the consensus sequence, DIXGFE (DISGFE inDictyostelium myosin II) and is functionally and structurally equivalent to a loop in the switch II region of GTPases (3Smith C.A. Rayment I. Biophys. J. 1996; 70: 1590-1602Abstract Full Text PDF PubMed Scopus (213) Google Scholar). In GTPases, the switch II loop connects the GTPase site and the switch II α-helix, which is part of the effector binding region. Information on the nucleotide state at the GTPase site is transmitted to the effector binding region partially through this switch II loop. In myosin, the switch II loop connects the ATPase pocket and a long conserved α-helix embedded in the lower 50K subdomain (4Rayment I. Rypniewski W.R. Schmidt B.K. Smith R. Tomchick D.R. Benning M.M. Winkelmann D.A. Wesenberg G. Holden H.M. Science. 1993; 261: 50-58Crossref PubMed Scopus (1850) Google Scholar, 5Fisher A.J. Smith C.A. Thoden J.B. Smith R. Sutoh K. Holden H.M. Rayment I. Biochemistry. 1995; 34: 8960-8972Crossref PubMed Scopus (629) Google Scholar). Recent x-ray crystallographic studies on Dictyostelium S1dC complexed with various nucleotides and nucleotide analogs revealed that the switch II loop undergoes a significant conformational change during ATP hydrolysis (Fig. 1 B) like the loop in GTPases. When S1dC is complexed with MgADP/Vior MgADP/AlFx (the Vi structure), the switch II loop is closer to the ATPase pocket, although the loop moves away from the ATPase pocket when S1dC is complexed with MgADP/BeFx, MgAMPPNP, MgATPγS, 1The abbreviations used are: MgATPγSadenosine 5′-O-(thiotriphosphate)MOPS4-morpholinepropanesulfonic acidNTAnitriloacetic acidHPLChigh performance liquid chromatography.1The abbreviations used are: MgATPγSadenosine 5′-O-(thiotriphosphate)MOPS4-morpholinepropanesulfonic acidNTAnitriloacetic acidHPLChigh performance liquid chromatography. or MgADP (the BeFx structure) (5Fisher A.J. Smith C.A. Thoden J.B. Smith R. Sutoh K. Holden H.M. Rayment I. Biochemistry. 1995; 34: 8960-8972Crossref PubMed Scopus (629) Google Scholar, 6Smith C.A. Rayment I. Biochemistry. 1996; 35: 5404-5417Crossref PubMed Scopus (511) Google Scholar, 7Gulick A.M. Bauer C.B. Thoden J.B. Rayment I. Biochemistry. 1997; 36: 11619-11628Crossref PubMed Scopus (178) Google Scholar). The observed changes in the switch II loop arise from the main chain rotation at the two pivoting residues, Ile-455 and Gly-457. In the Vi structure, Gly-457 and Glu-459 are close to the bound nucleotide. Gly-457 forms a hydrogen bond with an oxygen atom of the Vi moiety of the bound MgADP/Vi. Glu-459 coordinates with a water molecule that is expected to attack the bound ATP (6Smith C.A. Rayment I. Biochemistry. 1996; 35: 5404-5417Crossref PubMed Scopus (511) Google Scholar) and also forms an ionic bond with R238 in such a way as to close the exit ("backdoor") through which Pi may be released on ATP hydrolysis (8Yount R.G. Lawson D. Rayment I. Biophys. J. 1995; : 44S-47SPubMed Google Scholar) (Fig.1 B). In the BeFx structure, however, Gly-457 and Glu-459 are located away from the bound nucleotide. Moreover, the ionic bond between Arg-238 and Glu-459 is broken, and thus the exit is open (Fig.1 B). This conformational change of the loop is accompanied by rigid body motion of the upper and lower 50K subdomains to open and close the 50K cleft. This opening and closure of the cleft may then trigger the swinging of the lever arm, a long α-helix with bound light chains. Thus, the switch II loop of myosin seems to plays a critical role in the conversion of energy derived from ATP hydrolysis into sliding and force generation. adenosine 5′-O-(thiotriphosphate) 4-morpholinepropanesulfonic acid nitriloacetic acid high performance liquid chromatography. adenosine 5′-O-(thiotriphosphate) 4-morpholinepropanesulfonic acid nitriloacetic acid high performance liquid chromatography. A lower eukaryote, Dictyostelium discoideum, has a single copy of the heavy chain gene of myosin II (9DeLozanne A. Lewis M. Spudich J.A. Leinwand L.A. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 6807-6810Crossref PubMed Scopus (26) Google Scholar). Knock-out of this gene generates myosin II-null cells (10Manstein D.J. Titus M.A. De L.A. Spudich J.A. EMBO J. 1989; 8: 923-932Crossref PubMed Scopus (229) Google Scholar), which show myosin-specific defects in growth and development. These myosin-specific phenotypic defects can be reversed by introducing a multicopy plasmid bearing the wild-type heavy chain gene of myosin II (11Ruppel K.M. Egelhoff T.T. Spudich J.A. Ann. N. Y. Acad. Sci. 1990; : 147-155Crossref PubMed Scopus (19) Google Scholar). Using this Dictyostelium discoideum system established by Spudich and coworkers (10Manstein D.J. Titus M.A. De L.A. Spudich J.A. EMBO J. 1989; 8: 923-932Crossref PubMed Scopus (229) Google Scholar, 11Ruppel K.M. Egelhoff T.T. Spudich J.A. Ann. N. Y. Acad. Sci. 1990; : 147-155Crossref PubMed Scopus (19) Google Scholar, 12DeLozanne L.A. Spudich J.A. Science. 1987; 236: 1086-1091Crossref PubMed Scopus (751) Google Scholar, 13Manstein D.J. Ruppel K.M. Spudich J.A. Science. 1989; 246: 656-658Crossref PubMed Scopus (66) Google Scholar, 14O'Halloran T.J. Spudich J.A. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 8110-8114Crossref PubMed Scopus (18) Google Scholar, 15Egelhoff T.T. Brown S.S. Spudich J.A. J. Cell Biol. 1991; 112: 677-688Crossref PubMed Scopus (77) Google Scholar, 16Kubalek E.W. Uyeda T.Q. Spudich J.A. Mol. Biol. Cell. 1992; 3: 1455-1462Crossref PubMed Scopus (34) Google Scholar, 17Uyeda T.Q. Spudich J.A. Science. 1993; 262: 1867-1870Crossref PubMed Scopus (103) Google Scholar, 18Ruppel K.M. Uyeda T.Q. Spudich J.A. J. Biol. Chem. 1994; 269: 18773-18780Abstract Full Text PDF PubMed Google Scholar, 19Uyeda T.Q. Ruppel K.M. Spudich J.A. Nature. 1994; 368: 567-569Crossref PubMed Scopus (187) Google Scholar, 20Patterson B. Spudich J.A. Genetics. 1995; 140: 505-515Crossref PubMed Google Scholar, 21Ruppel K.M. Spudich J.A. Mol. Biol. Cell. 1996; 7: 1123-1136Crossref PubMed Scopus (71) Google Scholar), alanine scanning mutagenesis of the switch II loop was carried out to determine how residues in the switch II loop are involved in the energy conversion. The effects of mutations were studied by examining the phenotypes of cells expressing the mutant myosins and also by examining the in vitro motor functions of the purified mutant myosins. Each residue from Asp-454 to Glu-459 was changed to alanine by site-directed mutagenesis (22Kunkel T.A. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 488-492Crossref PubMed Scopus (4881) Google Scholar, 23Kunkel T.A. Roberts J.D. Zakour R.A. Methods Enzymol. 1987; 154: 367-382Crossref PubMed Scopus (4540) Google Scholar) of the Dictyosteliummyosin II heavy chain gene (9DeLozanne A. Lewis M. Spudich J.A. Leinwand L.A. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 6807-6810Crossref PubMed Scopus (26) Google Scholar). The mutant myosin heavy chain genes were ligated to the Dictyostelium actin-15 promoter andDictyostelium actin-6 terminator to drive their expression in Dictyostelium cells. They were finally inserted into a multicopy extrachromosomal vector, pBIG (24Patterson B. Spudich J.A. Genetics. 1996; 143: 801-810Crossref PubMed Google Scholar). Plasmids carrying the mutant myosin heavy chain genes were introduced intoDictyostelium myosin-null cells in which the myosin II heavy chain gene had been knocked out by means of homologous recombination (10Manstein D.J. Titus M.A. De L.A. Spudich J.A. EMBO J. 1989; 8: 923-932Crossref PubMed Scopus (229) Google Scholar). Dictyostelium cells transformed by electroporation were selected in a medium supplemented with 20 μg/ml of G418 on plastic dishes for a week. The transformed cells thus obtained expressed the mutant myosin II molecules. The truncated myosin heavy chain gene corresponding to S1dC (1Itakura S. Yamakawa H. Toyoshima Y.Y. Ishijima A. Kojima T. Harada Y. Yanagida T. Wakabayashi T. Sutoh K. Biochem. Biophys. Res. Commun. 1993; 196: 1504-1510Crossref PubMed Scopus (107) Google Scholar) was manipulated as above. The C terminus of S1dC was truncated at Arg-761 by introducing a stop codon at the corresponding location. For easier purification of S1dC, a 6-histidine tag (His6) was attached at the N terminus of S1dC by inserting the corresponding DNA sequence between the start codon and the coding sequence of the myosin heavy chain. The resulting transformation vectors were introduced intoDictyostelium AX2 cells. Random mutagenesis of the 459th residue was carried out by polymerase chain reaction with the E459A mutant myosin gene as a template. The E459A myosin gene was used as the template instead of the wild-type myosin gene to avoid excess representation of the wild-type gene in the library of mutagenized genes. The random representation of codons in the library was directly confirmed by sequencing some clones. A collection of mutant myosin genes was then inserted into a multicopy Dictyosteliumtransformation vector bearing the blasticidin-resistance gene, pBIGBsr (25Sutoh K. Plasmid. 1993; 30: 150-154Crossref PubMed Scopus (167) Google Scholar). By means of electroporation, myosin-null cells were transformed with pBIGBsr-based plasmids bearing myosin genes with all possible codons for the 459th residue. The electroporated cells were cultured overnight and then plated onto agar plates including blasticidin (20 μg/ml) together with Escherichia coli cells. After several days at 22 °C, plaques generated by Dictyostelium cells became visible on the E. coli lawns. Starting from 107 myosin-null cells, 672 plaques were obtained. Two types of plaques were easily distinguishable because of the clear difference in their diameters. Cells were cloned from all of the larger plaques (36 plaques). Plasmid DNA was rescued from randomly chosen clones (20 clones) and then sequenced. The growth rates were measured by determining the numbers of cells cultured in suspension. The incubator was shaken at 150 rpm at 22 °C. Development of the transformed cells was examined on agar plates covered with a lawn ofE. coli cells. Dictyostelium cells (1.2 × 104) were suspended in 10 mm Tris-Cl, pH 7.5, and then spotted onto the bacterial lawn. When Dictyosteliumcells had cleared the bacterial cells, they entered the developmental stage. Phosphorylated myosin was prepared as described previously (32Shimada T. Sasaki N. Ohkura R. Sutoh K. Biochemistry. 1997; 36: 14037-14043Crossref PubMed Scopus (77) Google Scholar). For experiments involving low concentrations of fluorescent nucleotides (mant-ATP and Cy3-ATP), myosin was purified further to remove the contaminating nucleotides. Myosin eluted from the HPLC column was dialyzed against a solvent comprising 50 mmNaCl, 10 mm MOPS, pH 7.4, and 2 mmMgCl2 to form filaments. The filaments were collected by centrifugation at 560,000 × g for 30 min and then dissolved in 0.35 m NaCl and 10 mm MOPS, pH 7.4. The wild-type or mutant S1dC bearing an N-terminal histidine tag was extracted from the transformed Dictyostelium cells, and precipitated as an actoS1dC complex by dialysis against a solvent comprising 50 mm NaCl, 10 mm MOPS, pH 7.4, and 2 mm MgCl2. S1dC was then extracted from the precipitate with a solvent comprising 10 mm MOPS, pH 7.4, 0.25 m NaCl, 7 mm MgCl2, and 5 mm ATP. The extract was directly applied to an NTA-Ni column (Qiagen), an affinity column for histidine-tagged proteins. After washing the column with a column volume of the above solvent containing 1 mm ATP, S1dC was eluted with a linear gradient of imidazole, pH 7.4, from 10 mm to 0.5 m. The eluted protein was dialyzed against a solvent comprising 50 mm NaCl, 10 mm MOPS, pH 7.4, and 2 mm MgCl2. Since the majority of the expressed E459A S1dC remained unbound to F-actin during the formation of actoS1dC, the supernatant obtained on high speed centrifugation of the actoS1dC complex was passed through the NTA-Ni column. The bound E459A S1dC was eluted with imidazole as above. In this alternative way, the major portion of the expressed E459A S1dC was recovered. The amount of E459A S1dC purified with the standard procedure was ∼10% of that of the protein recovered with the alternative procedure. Relative concentrations of the proteins were determined by Coomassie Protein Assay Reagent (Pierce), whereas their exact values were calculated photometrically by the method of Gill and von Hippel (36Gill S.C. von Hippel P.H. Anal. Biochem. 1989; 182: 319-326Crossref PubMed Scopus (5010) Google Scholar). Both methods gave consistent results. Cy3-ATP (1 μm) was added to the wild-type (0.2 μm), G457A (0.5 μm), or E459A (0.5 μm) myosin in 50 mm NaCl, 10 mm MOPS, pH 7.4, and 1 mm MgCl2 at 25 °C. After various times, a part of the reaction mixture was taken out and mixed with a 0.01 volume of PCA to stop the reaction. The resulting solution was centrifuged at 10,000 × g for 10 min to remove insoluble materials. The supernatant was applied directly to a reverse-phase HPLC column (Waters, Nova-pack C18). Elution was carried out with 100 mm potassium phosphate buffer, pH 6.8, and 11% acetonitrile. Under these elution conditions, Cy3-ATP and Cy3-ADP were well resolved. Both nucleotides were detected with a fluorimeter. The ratio of the concentrations of Cy3-ATP and Cy3-ADP was calculated to determine the amount of hydrolyzed Cy3-ATP. To compensate for the small amount of Cy3-ADP present in the Cy3-ATP preparation as a contaminant, myosin was first mixed with PCA and then mixed with Cy3-ATP. The resulting solution was treated as above. As for the D454A myosin, reliable measurement of the single turnover of Cy3-ATP was repeatedly hampered by its low yield. All fluorescence measurements were performed with a Perkin-Elmer LS50B Luminescence Spectrophotometer. The binding of mant-ATP to the D454A, G457A, or E459A myosin (0.25 μm) was determined in 150 mm NaCl, 20 mm MOPS, pH 7.4, and 5 mm MgCl2. Under these conditions,Dictyostelium myosin II remained in a soluble state. Since the myosins used for the measurements had lost their ATPase activities, mant-ATP was not hydrolyzed to an appreciable extent during the measurements. The nucleotide concentration was increased by stepwise addition of mant-ATP. The base line was determined using the above solvent without myosin. Fluorescence intensity was measured with excitation at 365 nm and emission at 440 nm. The actin-S1dC interaction was followed using pyrene-labeled F-actin (0.25 μm) (28Kouyama T. Mihashi K. Eur. J. Biochem. 1981; 114: 33-38Crossref PubMed Scopus (717) Google Scholar) and S1dC (0.25 μm) in a solvent comprising 50 mm NaCl, 10 mm MOPS, pH 7.4, and 2 mm MgCl2. Excitation was at 365 nm and emission at 410 nm. Intrinsic tryptophan fluorescence spectra were recorded using the wild-type or mutant S1dC (0.5 μm) in a solvent comprising 50 mm NaCl, 10 mm MOPS, pH 7.4, 2 mm MgCl2, and in the presence and absence of 0.1 mm ATP. Excitation was at 290 nm. Tryptophan fluorescence spectra of completely denatured proteins were also recorded after denaturing them in 6 m GuHCl, and normalized to each other to confirm that observed difference in the tryptophan fluorescence actually arose from different conformations. E459A S1dC purified by the standard or alternative procedure (1 μm) in 50 mm NaCl, 10 mm MOPS, pH 7.4, and 2 mm MgCl2 was incubated with 20 μmmant-ATP. The mixture was passed through a gel filtration HPLC column (Ashahipack) equilibrated with a solvent comprising 0.5 mNaCl and 10 mm MOPS, pH 7.4. The fluorescence intensities of the bound and free mant-ATP (excitation, 365 nm; emission, 440 nm) as well as the absorption of S1dC (at 280 nm) were monitored using two tandemly arranged flow cells. After 2 days of incubation with mant-ATP, the mixture was passed through the gel filtration column to purify the S1dC·mant-ATP complex. At various times after the purification, the purified E459A S1dC·mant-ATP complex was again passed through the column to determine the amount of the fluorescent nucleotide released. Actin-activated and basal MgATPase activities were measured as described (32Shimada T. Sasaki N. Ohkura R. Sutoh K. Biochemistry. 1997; 36: 14037-14043Crossref PubMed Scopus (77) Google Scholar). In vitro motility assays were carried out as described (29Kron S.J. Spudich J.A. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 6272-6276Crossref PubMed Scopus (701) Google Scholar, 30Toyoshima Y.Y. Kron S.J. McNally E.M. Niebling K.R. Toyoshima C. Spudich J.A. Nature. 1987; 328: 536-539Crossref PubMed Scopus (397) Google Scholar, 31Harada Y. Sakurada K. Aoki T. Thomas D.D. Yanagida T. J. Mol. Biol. 1990; 216: 49-68Crossref PubMed Scopus (449) Google Scholar). Dictyostelium myosin II-null cells could not undergo normal cytokinesis and only slowly grew in suspension up to the density of ∼1 × 106 cells/ml, becoming multinucleated cells (Fig. 2,Null), consistent with the previous reports (10Manstein D.J. Titus M.A. De L.A. Spudich J.A. EMBO J. 1989; 8: 923-932Crossref PubMed Scopus (229) Google Scholar, 12DeLozanne L.A. Spudich J.A. Science. 1987; 236: 1086-1091Crossref PubMed Scopus (751) Google Scholar). When a multicopy plasmid bearing the wild-type heavy chain gene of myosin II was introduced into myosin-null cells, the defect in cytokinesis was reversed so that the resulting transformants (designated as "wild-type cells") could grow in suspension as mononucleated cells up to the density of ∼2 × 107 cells/ml (Fig. 2,WT). The other transformants expressing the mutant myosins could be grouped into two types according to their behavior in suspension culture (Fig. 2). I455A and S456A cells grew like the wild-type cells, although D454A, G457A, F458A, and E459A cells did not grow or grew only slowly, like myosin-null cells. When transformedDictyostelium cells were allowed to develop on agar plates covered with E. coli cells, the I455A and S456A cells, which grew well in suspension, formed fruiting bodies like the wild-type cells. However, the D454A, G457A, F458A, and E459A cells, which had a defect in suspension culture, could not develop beyond the mound stage, like myosin-null cells. The actin-activated and basal MgATPase activities of the purified myosins were measured (Fig. 3). The D454A, G457A, F458A, and E459A myosins exhibited very low Vmax values for the actin-activated ATPase activity. Among them, the D454A, G457A, and E459A myosins also exhibited very low basal ATPase activities, suggesting that these three mutant myosins had almost completely lost their ability to hydrolyze ATP. In contrast to these myosins, the F458A myosin retained high basal activity similar to that of the I455A or S456A myosin. The Vmax value for the actin-activated MgATPase activity of the F458A myosin was, however, almost the same as that of its basal activity, indicating the complete loss of the actin-activated ATPase activity. The I455A and S456A myosins, which complemented the defects of myosin-null cells, retained their actin-activated and basal ATPase activities. In vitro motility assays were carried out on the purified myosins (Fig. 4). As expected from the phenotypes of cells expressing the mutants, the D454A, G457A, F458A, and E459A myosins could not drive the sliding of actin filaments, whereas the I455A and S456A myosins drove the sliding. As described above, the D454A, G457A, and E459A mutations almost completely abolished the ATPase activity. To determine which step of the ATPase cycle was blocked by the mutations, the binding of mant-ATP to these myosins was measured. Since all these myosins lost their ATPase activities, it was possible to carry out stepwise titration without appreciable hydrolysis of mant-ATP during the measurements. As shown in Fig. 5, the G457A myosin bound mant-ATP tightly, whereas the D454A and E459A myosins bound it more weakly. As shown below, however, it seems that the weak binding of mant-ATP to the E459A myosin was because of the slow binding of the fluorescent nucleotide to the mutant, not to its intrinsically low affinity. These results suggest that the D454A, G457A, and E459A myosins could bind mant-ATP with various affinities, implying that these mutations did not block the ATP binding to the mutant myosins, but blocked the ATP hydrolysis step, like the R238A mutation in the switch I loop (32Shimada T. Sasaki N. Ohkura R. Sutoh K. Biochemistry. 1997; 36: 14037-14043Crossref PubMed Scopus (77) Google Scholar). To further investigate the step blocked by the G457A and E459A mutations, the single turnover of the ATPase reaction was followed using a fluorescent ATP analog, Cy3-ATP. As shown in Fig. 6, the wild-type myosin showed an "ADP burst" (∼0.7 mol/mol of ATPase-site), reflecting the quick ATP hydrolysis step followed by the rate-limiting Pirelease step. Unlike the wild-type myosin, the G457A myosin did not show such an "ADP burst" (Fig. 6), indicating that the G457A mutation blocked the ATP hydrolysis step, not the Pirelease step. Similar results were recently reported for smooth muscle myosin (33Onishi H. Morales M.F. Kojima S. Katoh K. Fujiwara K. Biochemistry. 1997; 36: 3767-3772Crossref PubMed Scopus (27) Google Scholar). Like the G457A myosin, the E459A myosin did not show an ADP burst (Fig. 6). This result supports, but does not necessarily confirm, that the mutation blocked the hydrolysis step, given the fact that the rate of mant-ATP binding to E459A myosin was unexpectedly slow, as shown below. The ATP-dependent association and dissociation of the Dictyostelium myosin motor domain, S1dC (1Itakura S. Yamakawa H. Toyoshima Y.Y. Ishijima A. Kojima T. Harada Y. Yanagida T. Wakabayashi T. Sutoh K. Biochem. Biophys. Res. Commun. 1993; 196: 1504-1510Crossref PubMed Scopus (107) Google Scholar), with pyrene-labeled F-actin was studied, the pyrene fluorescence being followed (28Kouyama T. Mihashi K. Eur. J. Biochem. 1981; 114: 33-38Crossref PubMed Scopus (717) Google Scholar). S1dC was used mainly because more reproducible data were obtained using this soluble, single-headed fragment. As previously shown, the pyrene fluorescence decreased when the wild-type S1dC formed a rigor complex with the pyrene-labeled actin (34Kurzawa S.E. Manstein D.J. Geeves M.A. Biochemistry. 1997; 36: 317-323Crossref PubMed Scopus (76) Google Scholar). On the addition of ATP, S1dC was transiently dissociated from the F-actin, and then reassociated with it after ATP had been completely hydrolyzed to ADP and Pi. Thus, the pyrene fluorescence transiently increased and then decreased (Fig. 7 A). F458A S1dC also formed a rigor complex with F-actin in the absence of ATP, as indicated by the decrease in the pyrene fluorescence (Fig. 7 B). The rigor complex was transiently dissociated on the addition of ATP and was formed again after ATP had been exhausted. Complete dissociation of the rigor complex was achieved only when a large excess of ATP was added because F458A S1dC retained high basal ATPase activity and, therefore, quickly consumed ATP. Thus, F458A S1dC exhibited normal ATP-dependent dissociation-association with F-actin even though it had lost its actin-activated ATPase activity. When G457A S1dC was mixed with the pyrene-labeled F-actin in the absence of ATP, a rigor complex was formed, as judged from the decrease in the pyrene fluorescence. On the addition of a small amount of ATP (even 1 mol/mol of S1dC), almost complete dissociation of the rigor complex occurred (Fig. 7 C), indicating that G457A S1dC entered in a weak-binding state when it bound ATP. Because of the lack of ATPase activity, G457A S1dC remained in this weak-binding state. Unlike these mutants, however, E459A S1dC, purified by either the standard or the alternative procedure (see "Experimental Procedures"), failed to form a rigor complex with F-actin even in the absence of ATP, as judged from the fact that the pyrene fluorescence of F-actin never decreased on the addition of the purified S1dC. The intrinsic tryptophan fluorescence of the wild-type S1dC increased on the addition of excess ATP (Fig. 8 A), as previously reported for a similar fragment of Dictyostelium myosin (34Kurzawa S.E. Manstein D.J. Geeves M.A. Biochemistry. 1997; 36: 317-323Crossref PubMed Scopus (76) Google Scholar). On the addition of MgADP, a slight decrease in the fluorescence intensity was observed. Unlike that of the wild-type S1dC, however, the tryptophan fluorescence of G457A S1dC did not change on the addition of ATP or ADP (Fig. 8 B). The fluorescence intensity remained at the same level as that of the wild-type S1dC in the absence of ATP. The tryptophan fluorescence of E459A S1dC also did not respond to ATP or ADP. The fluorescence intensity remained higher than that of G457A S1dC, being similar to that of the wild-type S1dC in the presence of ATP (Fig. 8 B). On the addition of ADP to F458A S1dC, the tryptophan fluorescence slightly decreased, as in the case of the wild-type S1dC (Fig. 8 C). However, on the addition of ATP, the tryptophan fluorescence did not increase, but decreased further (Fig. 8 C), indicating that F458A S1dC took on a unique steady state in the presence of ATP. As mentioned above, the purified E459A S1dC did not form a rigor complex with F-actin. One possibility for this unexpected result is that the E459A mutation induced large conformational changes that made the E459A myosin and S1dC unable to form the rigor complex. Another possibility is that the ATPase site of the E459A mutant was occupied by a tightly trapped endogenous ATP. The latter possibility is much more likely because the intensity of the tryptophan fluorescence of E459A S1dC remained higher than that of G457A S1dC in the presence and absence of ATP, as if it bound ATP or ADP·Pi (35Bagshaw C.R. Trentham D.R. Biochem. J. 1974; 141: 331-349Crossref PubMed Scopus (334) Google Scholar). When ATP was tightly trapped at the ATPase site, it would be chased only slowly by the free nucleotide. To test if this was the case, E459A S1dC was incubated with a 20-fold molar excess of mant-ATP. After various times, the mixture was passed through a gel filtration HPLC column to separate the bound and free nucleotides. As shown in Fig. 9 A, the amount of the fluorescent nucleotide incorporated into the protein very slowly increased with increasing incubation time. After a week, ∼60% of the protein had incorporated the fluorescent ATP. Thus, it seems that the mant-ATP slowly chased ATP which was tightly trapped at the ATPase site of E459A S1dC and then was bound there. It must be noted that the binding of mant-ATP to other mutants such as G457A S1dC took place within several seconds under the same conditions (data not shown). To further confirm the tight trapping of the bound nucleotide, the rate of release of mant-ATP trapped in E459A S1dC was measured. At various times after the purification, the purified complex of E459A S1dC and mant-ATP was passed through the HPLC column again to separate the bound from the released mant-ATP. As shown in Fig. 9 B, the bound fluorescent nucleotide was only slowly released, confirming the notion that ATP (or its analog) was tightly trapped at the ATPase site of the E459A mutant, once it had been incorporated there. After the incorporation of mant-ATP, E459A S1dC was purified on the gel filtration column as above, and then the bound fluorescent nucleotide was released with 0.1% PCA. When analyzed on a reverse-phase HPLC column, only mant-ATP was detected in the released nucleotide, directly showing that the ATP hydrolysis step was blocked in the E459A S1dC. The 459th residue was randomly mutagenized using the E459A myosin gene as a template. Then the mixture of mutagenized myosin genes was introduced into myosin-null cells. When transformed cells were allowed to grow in the presence ofE. coli cells, they formed two types of plaques that were easily distinguishable from their diameters. The diameter of the larger plaques was ∼2-fold larger than that of the smaller ones, which was almost the same as that of myosin-null cells. Among the 672 plaques generated on the transformation of 107 cells, 36 plaques were of the larger type. The diameter of a plaque is a good indicator of the in vitro motor functions of myosin (20Patterson B. Spudich J.A. Genetics. 1995; 140: 505-515Crossref PubMed Google Scholar, 24Patterson B. Spudich J.A. Genetics. 1996; 143: 801-810Crossref PubMed Google Scholar). Plasmids bearing the myosin genes were retrieved from Dictyosteliumcells isolated from the larger plaques. The sequencing of randomly chosen plasmids thus rescued (20 plasmids) showed that the codon of the 459th residue was either GAG or GAA (14 GAG and 6 GAA), indicating that the 459th residue of a functional myosin must be glutamic acid. Mutant myosins generated by alanine scanning mutagenesis of the switch II loop can be classified into two groups according to theirin vivo phenotypes. One group, comprising the I455A and S456A myosins, fully reverse the myosin-specific defects of myosin-null cells. The other group, comprising the D454A, G457A, F458A, and E459A myosins do not reverse any of the defects. Although I455 is a highly conserved residue in almost all myosins (2Mooseker M.S. Cheney R.E. Annu. Rev. Cell Dev. Biol. 1995; 11: 633-675Crossref PubMed Scopus (384) Google Scholar) and functions as a pivoting residue for the main chain rotation of the switch II loop during the transition from the Vi structure to the BeFx structure (5Fisher A.J. Smith C.A. Thoden J.B. Smith R. Sutoh K. Holden H.M. Rayment I. Biochemistry. 1995; 34: 8960-8972Crossref PubMed Scopus (629) Google Scholar, 6Smith C.A. Rayment I. Biochemistry. 1996; 35: 5404-5417Crossref PubMed Scopus (511) Google Scholar), the I455A myosin retained most of its motor functions. It must be noted that when the other pivoting residue, G457, was replaced with alanine, the motor functions were completely lost. It seems that the I455A mutation did not block the main chain rotation of the switch II loop, unlike the G457A mutation. Another mutant in the first group, the S456A myosin, exhibited normal motor functions, as expected from the facts that this residue is not conserved among myosins and that alanine occupies this position in some myosins (2Mooseker M.S. Cheney R.E. Annu. Rev. Cell Dev. Biol. 1995; 11: 633-675Crossref PubMed Scopus (384) Google Scholar). The mant-ATP titration and single turnover measurements showed that the G457A mutant was unable to hydrolyze ATP because the ATP hydrolysis step (Fig. 10) was blocked. The G457A mutant could bind ATP tightly, however, as judged from the mant-ATP binding to the mutant myosin. Consistent with the result, the actin·G457A S1dC complex was readily dissociated on addition of an equivalent amount of ATP, whereas the tryptophan fluorescence of the mutant S1dC did not change. These results lead us to conclude that G457A S1dC tightly binds ATP and enters in a state (M′·ATP in Fig. 10) distinguishable from the M·ATP or M*·ATP state (35Bagshaw C.R. Trentham D.R. Biochem. J. 1974; 141: 331-349Crossref PubMed Scopus (334) Google Scholar). The simple collision complex (MATP) is expected to bind ATP much more weakly, and the M*·ATP complex is expected to exhibit more enhanced tryptophan fluorescence than the G457A S1dC·ATP complex. The position of Gly-457 relative to the γ-phosphate of ATP in the ATPase pocket may change on rotation of the main chain of the switch II loop, depending on the state of the nucleotide. Thus, Gly-457 inDictyostelium myosin seems to function like the "γ-phosphate sensor" glycine in GTPases (for example, Gly-60 in Ras). It is likely that the G457A mutation blocked this rotation of the main chain because of steric hindrance. Given the fact that the G457A mutant was trapped before the isomerization step when it bound ATP, it is tempting to speculate that the isomerization step is coupled with the rotation of the main chain of the switch II loop (33Onishi H. Morales M.F. Kojima S. Katoh K. Fujiwara K. Biochemistry. 1997; 36: 3767-3772Crossref PubMed Scopus (27) Google Scholar), which occurs on the transition from the BeFx structure to the Vistructure, and that G457A S1dC·ATP takes on the BeFx structure. Consistent with this notion, mant-ADP/BeFx was trapped in G457A S1dC, whereas mant-ADP/Vi was not (data not shown). The side chain of Glu-459 is located close to the bound nucleotide in the Vi structure of Dictyostelium S1dC, forming a hydrogen bond with a water molecule suitably positioned to participate in ATP hydrolysis (6Smith C.A. Rayment I. Biochemistry. 1996; 35: 5404-5417Crossref PubMed Scopus (511) Google Scholar). This strategic location of Glu-459 suggests that the residue is crucial for the hydrolysis step. In fact, the E459A mutant was unable to hydrolyze ATP because the ATP hydrolysis step was blocked (Fig. 10). Once ATP was in the ATPase pocket of the E459A mutant, it was almost irreversibly trapped there without hydrolysis, as observed here. The results suggest that the E459A mutant was trapped possibly at the M*·ATP state (Fig. 10) (35Bagshaw C.R. Trentham D.R. Biochem. J. 1974; 141: 331-349Crossref PubMed Scopus (334) Google Scholar). Besides its role in ATP hydrolysis, Glu-459 may also play a role as a "gatekeeper" of the backdoor for Pi release (8Yount R.G. Lawson D. Rayment I. Biophys. J. 1995; : 44S-47SPubMed Google Scholar), opening and closing it through the ionic interaction with Arg-238 (Fig.1 B) (32Shimada T. Sasaki N. Ohkura R. Sutoh K. Biochemistry. 1997; 36: 14037-14043Crossref PubMed Scopus (77) Google Scholar). This notion implies that ATP hydrolysis is tightly coupled with the opening and closing of the backdoor. The crucial importance of Glu-459 was also highlighted by the observation that the motor functions were retained only when glutamic acid occupied the 459th position. The side chain of Asp-454 faces the ATPase pocket, and is coordinated to an Mg ion of the bound nucleotide through a water molecule (5Fisher A.J. Smith C.A. Thoden J.B. Smith R. Sutoh K. Holden H.M. Rayment I. Biochemistry. 1995; 34: 8960-8972Crossref PubMed Scopus (629) Google Scholar, 6Smith C.A. Rayment I. Biochemistry. 1996; 35: 5404-5417Crossref PubMed Scopus (511) Google Scholar). Ser-237 in the switch I loop of Dictyostelium myosin is also directly coordinated to the Mg ion from the other side of the ATPase pocket. Unlike the S237A myosin (32Shimada T. Sasaki N. Ohkura R. Sutoh K. Biochemistry. 1997; 36: 14037-14043Crossref PubMed Scopus (77) Google Scholar), however, the D454A myosin bound mant-ATP, although weakly, indicating that Asp-454 is of secondary importance in retaining the MgATP in the ATPase pocket, whereas Ser-237 is essential for this. In contrast to the D454A, G457A, and E459A myosins, the F458A myosin in the second group retained the basal MgATPase activity although it completely lost the actin-activated ATPase activity. The observedin vivo defects of the F458A myosin arose from the lack of this essential ability to power the motor. The side chain of Phe-458 points away from the ATPase pocket and is buried in a hydrophobic pocket formed by residues such as Asn-472, Asn-475, His-572, Tyr-573, and Ala-574 in the core of the lower 50K subdomain (Fig.1 C). When the main chain of the switch II loop rotates at Gly-457 and Ile-455, the side chain of Phe-458 swings and rotates. It seems that the hydrophobic side chains surrounding Phe-458 follow this motion through the hydrophobic interaction. Thus, the swinging and rotation of Phe-458 seem to trigger the rigid body motion of the lower 50K subdomain, which opens and closes the 50K cleft (5Fisher A.J. Smith C.A. Thoden J.B. Smith R. Sutoh K. Holden H.M. Rayment I. Biochemistry. 1995; 34: 8960-8972Crossref PubMed Scopus (629) Google Scholar). Therefore, disruption of the hydrophobic interaction by the F458A mutation blocks some of the structural changes expected to occur during the ATPase cycle, forcing the mutant to bypass some intermediate states. In fact, in the presence of ATP, F458A S1dC was in a unique steady state quite different from M**·ADP·Pi, as judged from the tryptophan fluorescence intensity. The F458A mutant in this unique steady state failed to interact with F-actin in such a way that it stimulated the actin-activated ATPase activity. Further kinetic and structural studies on the F458A myosin would reveal how F-actin triggers the actin-activated ATPase activity. We thank Reiko Ohkura for her excellent technical assistance. The coordinates of the motor domain ofDictyostelium myosin II were kindly provided by Dr. Rayment (University of Wisconsin). The myosin II heavy chain gene, myosin-null cells, pBIG vector, and recombinant MLCK gene were provided by Dr. Spudich (Stanford University), Dr. Patterson (University of Arizona), and Dr. Uyeda (National Institute for Advanced Interdisciplinary Research, Japan). Mant-ATP and Cy3-ATP were provided by Dr. Hiratsuka (Asahikawa Medical University) and Dr. Oiwa (Kansai Advanced Research Center, Communication Research Laboratory, Japan), respectively.