/ . Biochem. 84, 1453-1458 (1978)
Myosin and Actin from Escherichia coli K12 C600 Kayoko NAKAMURA, 1 Koui TAKAHASHI,* and Shizuo WATANABE Department of Chemistry, Faculty of Science, Tokyo Institute of Technology, Meguro-ku, Tokyo 152, *Department of Animal Science, Faculty of Agriculture, Hokkaido University, Sapporo, Hokkaido 060 Received for publication, June 9, 1978
Myosin-like protein and actin-like protein from E. coli formed filaments very similar in structure to those of myosin and actin from skeletal muscle. At 0.2 M KC1, a large number of "thick filaments" of uniform size (about 0.6-0.7 ftm long and about 20 nm wide) was present. These thick filaments aggregated as the KC1 concentration decreased to less than 0.2 M. Filaments of actin-like protein were decorated with muscle heavy meromyosin, showing "arrowheads". The arrowhead structure disappeared in the presence of ATP. A mixture of E. coli myosin-like protein and rabbit skeletal actin exhibited a gelation phenomenon on the addition of ATP. The phenomenon was reversible and showed ATP specificity. However, the gelation phenomenon was not observed with the mixture of E. coli actin-like protein and E. coli myosin-like protein. These results provide compelling evidence that the E. coli myosin-like protein and actinlike protein we isolated are essentially identical to myosin and actin, respectively.
In the previous paper (/), we reported the isolation of myosin-like protein and actin-like protein from E. coli. We identified the myosin-like protein on three counts: its mobility in SDS-gel electrophoresis, its ability to bind with rabbit skeletal actin, and its ATPase activities. We also identified the actin-like protein on four counts: its mobility in SDS-gel electrophoresis, its ability to bind with rabbit skeletal myosin, its polymerizing-depoly1
merizing ability, and its ability to activate MgATPase of myosin. In the present paper, we report that the two proteins can form filaments very similar in structure to those of myosin and actin from skeletal muscle. Moreover, we also report that the mixture of E. coli myosin-like protein and rabbit skeletal actin exhibits a gelation phenomenon on the addition of ATP, instead of superprecipitation.
MATERIALS AND METHODS Present address: Tokyo Metropolitan Institute of Medical Science, Department of Radiology, Bunkyo-ku, Actin-Like Protein and Myosin-LJke Protein Tokyo 113. Abbreviations: E. coli, Escherichia coll; SDS, sodium from E. coli—Actin-like protein and myosin-like dodccyl sulfate; EGTA, glycoetherdiamine-Af,A^,Ar'^v"- protein were prepared from E. coli by the methods described in our previous paper (7). The protein tetraacetic acid. Vol. 84, No. 6, 1978
1453
K. NAKAMURA, K. TAKAHASHI, and S. WATANABE
1454
concentration was determined by the biuret method (2), by the method of Lowry et al. (J), or by measuring the absorbance at 280 nm, using bovine serum albumin as a standard for each determination. Electron microscopy: The procedure for the electron microscopic study was essentially the same as those described by Huxley (4). Samples placed on carbon-coated grids were stained with 1 % uranyl acetate and examined under a Hitachi 11-B electron microscope, with an acceleration voltage of 75 kV. HMM: Skeletal heavy meromyosin (HMM) was kindly supplied by Dr. F. Morita of Hokkaido University. Diethylaminoethyl-cellulose chromatography: Protein solutions were dialyzed against a buffer solution containing 0.35 M KC1 and 20 mM TrisHC1 (pH 7.8), and applied to the column, which had been fully equilibrated with the same solution. Seven mg of protein was applied to a 1.5 cm x 5 cm column containing about 1.5 g of diethylaminoethyl (DEAE)-cellulose. Elution was carried out at a flow rate of 0.3 ml/s. All chromatographic procedures and all procedures for protein preparation were carried out at 0-4°C. DEAE-cellulose was purchased from Whatman and R. Balton Ltd. RESULTS /. DEAE-Cellulose Chromatography—In the previous paper (7), our preparation of myosin-like protein contained some non-protein substances. In fact, our preparation of E. coli myosin-like protein had a very high UV absorption at 260 nm (see Fig. l-2a). In that report (7), we estimated the protein concentration from the absorbance at 280 nm. We now determined the protein content of the same solution of myosin-like protein preparation by three different methods: measurement of the 280 nm absorbance, the biuret method (2), and Lowry's method (J). The protein concentration was found to be 1.6 mg/ml, 0.9 mg/ml, and 0.93 mg/ml, respectively. Therefore the low ATPase activities of E. coli myosin-like protein reported in the previous paper (7) should be increased by a factor of 1.8. We attempted DEAEcellulose chromatography by a modification of Mueller's method (5) to remove the non-protein substances. However, Fig. 1-1 shows that the
II
III IV V
1.0
9 O
s
Q.
CM
O CD
1.0
0.5
I
2 3 4 5
6 7 8
Fr.No Fig. 1-1. Chromatography of the myosin-like protein preparation on diethylaminocthyl cellulose. Protein solution (7 mg in 6 ml) in 0.35 M KC1, 20 mM Tris-HCl (pH 7.8) was applied to a column, 1.5cmx5cm. Elution was carried out at a flow rate of 0.3 ml/s with a solution containing 0.35 M KC1 and 20 mM Tns-HCl (pH 7.8). O, Absorbance at 280 nm; • , absorbance at 260 nm; • , the protein concentration determined by Lowry's method (3). absorbance of every fraction at 260 nm remained high, indicating that the myosin-like protein could not be fully separated from non-protein substances. Judging from the UV spectrum (Fig. l-2b), the content of non-protein substances was reduced, though only slightly, in fraction n . Therefore, fraction II was used for the following experiments. II. Thick Filament Formation—Electron microscopy revealed that out myosin-like protein preparation can form thick filaments. The thick filaments shown in Fig. 2 were obtained in the following way: 0.2 ml of E. coli myosin-like protein solution (0.8 mg/ml E. coli myosin-like protein in 0.6 M KCI, 10 mM imidazole-HCl, pH 7.2) was transferred into a test tube of 1.5 cm diameter. With constant stirring and taking care to avoid foaming of the solution, 0.4 ml of a diluting solution (10 mM imidazole-HCl, pH 7.2) was added / . Biochem.
MYOS1N AND ACTIN FROM E. coli
I5
20 .
SI 5 I Q5
1.0
260 280 300 Vtawltngth (run)
0 0.
260 280 300 WavfUngth (nm>
Fig. 1-2 UV spectra of the myosin-like protein preparation (a), and of the fraction obtained by DEAEcellulose chromatography of the myosin-like protein (b).
1455 slowly, i.e., taking 40 s. Aggregates were scarcely observed at 0 2 M K G , and a large number of " thick filaments" of uniform size (about 0.60.7 f*m long and about 20 nm wide) was seen. The electron micrograph also shows a small number of filaments with apparent bipolar structure. The nature of particles (8-9 nm in diameter) shown in Fig. 2 remains unclear (6). We observed that at 0.1 M K G , all thick filaments of E. coli myosin-like protein were packed into large aggregates (data not shown), and that the aggregates decreased in number as the K G concentration was increased to 0.2 M. These properties of E. coli myosin-like protein were reported by Mabuchi (7) with thick filaments of egg-cell myosin. ///. Thin Filament Formation—Electron microscopic study revealed that the actin-hke protein of E. coli can form long, fine filaments (Fig. 3-1). The width of the filaments was estimated to be approximately 6-9 nm. It is possible to see the presence of 365 A periodicity along the length of the filaments. The actin-hke protein preparation of E. coli
Fig. 2. Electron micrograph of thick filaments of E. coli myosin-like protein. The concentration of myosinlike protein was 0.25 mg/ml and the medium contained 0.2 M KQ and 10 ITIM imidazole-HCl (pH 7.2). The scale bar represents 0.5 ftm. Magnification: x54,000. Vol. 84, No. 6, 1978
1456
K. NAKAMURA, K. TAKAHASHJ, and S. WATANABE
Fig. 3-1. Electron micrograph of actm-hke filaments. The concentration of "actin-equivalent component" was 0.035 mg/ml and the medium contained 0.1 M KG and 5 IHM K phosphate (pH 7.2). Magnification: x 71,000, insert, x 140,000.
Fig 3-2. The binding of skeletal heavy meromyosin (HMM) to E. colt actin-like protein in the absence of ATP. The final concentrations of "actin-equivalent component" and HMM were 0.035 mg/ml and 0 25 mg/ml, respectively. The medium contained 0.1 M KCI and 10 m.M Tris-maleate (pH 6.8). Magnification: x 71,000. J.
Biochem.
MYOSIN AND ACTIN FROM E. coli
1457
Fig 3-3. Dissociation of E. coli actm-Iike protein and skeletal heavy meromyosin (HMM) in the presence of 1 mM ATP. The conditions were the same as those described in Fig. 3-2 except that the ATP concentration was 1 mM Magnification: x71,000.
ITT
Fig. 4
Vol. 84, No. 6, 1978
TV
was mixed with rabbit muscle HMM (approximately 0.25 mg/m!) in buffer solution of 0.1 M K G and 10 mM Tris-maleate (pH 6 8) or in this buffer solution plus I mM ATP. In the absence of ATP, many of the thin filaments became decorated with muscle HMM, showing "arrowheads", all of which pointed in the same direction (Fig. 3-2). The diameter of the decorated filaments (around 24 nm), the polarity indicated by the arrowheads and the repeating distance (36 nm) of these arrowheads resembled those reported for filaments of actin from other sources (7, 8). In the presence of
Fig. 4. Gelation phenomena Every sample contained O25mg/ml of E coli myosin-like protein (fraction II) and 0.13 mg/ml of rabbit skeletal actin. The medium contained 0 . 1 6 M KC1, 20 mM Tns-maleate (pH 6.8), I mM MgClj, together with (I) 0 I mM CaCl, and no ATP, (II) 0.1 mM CaCl, and 0.1 mM ATP, (III) 0.1 mM CaCl, and 0.1 mM AMP-PNP, or (IV) I mM EGTA and 0.1 mM ATP. The samples were kept standing without stirring at 20°C. The photos was taken 30 min after the addition of ATP or AMP-PNP (a), and three hours after the addition of ATP or AMP-PNP (b).
K NAKAMURA, K. TAKAHASHI, and S. WATANABE
1458
1 mM ATP, HMM dissociated from the filaments, those of rabbit skeletal myosin, but are somewhat leaving nearly bare filaments (Fig. 3-3) The wider (20 nm as compared with 15 nm). diameter of the bare thin filaments was the same As indicated by arrows in Fig. 3-1, we observed as that of the thin filaments in Fig. 3-1 short filamental structure in an electron micrograph of the E. coli actin-like protein. The structures IV. Gelation Phenomenon—We reported in the previous paper (/) that the E. coli myosin-like were larger in width (about 9 nm) and shorter in length than actin filaments. Moreover, the strucprotein in combination with rabbit skeletal actin tures were not decorated by rabbit muscle HMM. failed to show superprecipitation activity. We The decorated filaments of E. coli actin-like protein investigated whether the DEAE-cellulose chroare similar to those of muscle actin of Mycoplasma matographed preparation of E. coli myosin in pneumomae (8). The arrowhead structure of combination with skeletal actin could superHMM-decorated thin filaments of E coli actinprecipitate, using a medium containing 0 . 1 6 M like protein disappeared on addition of ATP. KCI, 20 mM Tris-maleate (pH 6.8), 0.1 mM CaCI2 or 1 mM EGTA, 1 mM MgCI,, and 0.1 mM ATP. These results provide compelling evidence that the thin filaments of E. coli actin-like protein described We again failed to observe superprecipitation are indeed those of actin. activity. However, we found that the mixture of E. coli myosin-like protein preparation (fraction A gelation phenomenon was reported to occur II) and rabbit skeletal actin formed a solid gel with actin preparations of non-muscle cells (9). upon addition of ATP (at 20°C) (Fig. 4a). The However, in our experiments, the gelation phenomenon was not observed with E. coli actin-like gelation phenomenon also occurred on the addiprotein, but was observed with skeletal actin and tion of 4-ATP, but did not occur on the addition E. coli myosin-like protein. Moreover, the gelation of ADP of AMP-PNP The protein mixture in the same medium containing calcium (0 1 mM phenomenon was found to be reversible and to have ATP specificity. Therefore, the gelation we CaCli) formed a solid gel within about 20 min observed was associated with the E. coli myosinafter the addition of ATP, and it recovered to a like protein rather than the E. coli actin-like sol a few hours later, probably when the ATP in protein. the medium had been exhausted (Fig. 4b). Addition of ATP produced a solid gel again. When We wish to express our thanks to Dr. Kazuo Sudo of the medium contained no calcium (1 mM EGTA), the University of Tokyo for his interest and encourageit took more than 20 min to form a solid gel. ment in connection with our electron microscope The time required for gel formation was also experiments dependent on the ATP concentration, that is, the time required increased as the ATP concentration REFERENCES was increased. DISCUSSION Mabuchi (7) reported that sea-urchin egg myosin forms an aggregated form of thick filaments in 0.1 M KCI, whereas rabbit skeletal myosin forms an unaggregated form of thick filaments in 0.1 M KCI. The thick filaments of E. coli myosin-like protein also are in the aggregated form in 0.1 M KCI, but are in the unaggregated form in 0.2 M KCI, where thick filaments of sea-urchin egg myosin are still in the aggregated form. E. coli myosin-like protein thick filaments are not very different in length (0.6-0.7 pm as compared with 0.5—1.0 pm) from
I. Nakamura, K. & Waianabe, S. (1978) J. Biochem. 83, 1459-1470 2 Gornall, A.G , Badawill, C J., & David, M.M. (1949) J. Biol. Chem. 177, 751-766 3. Lowry, O.H., Rosebrough, N J., Farr, A.L., & Randall, R J. (I95J) J. Biol. Chem. 193, 265-275 4. Huxley, H E (1963) J. Mol. Biol 7, 281-308 5. Mueller, H. & Perry, S V. (1961) Biochem. J. 80, 217-223 6 Ishihama, A., Ikeuchi, T., Matsumoto, A., & Yamamoto, S. (1976) J. Biochem 79, 927-936 7. Mabuchi, 1. (1976) J Mol. Biol. 100, 569-582 8. Neimark, H.C. (1977) Proc. Natl. Acad. Sci. U.S. 74, 4041^(045 9. Pollard, T.D. (1976) J Cell Biol. 68, 579-601
J. Biochem.
Myosin and Actin from Escherichia coli K12 C600 Kayoko NAKAMURA, 1 Koui TAKAHASHI,* and Shizuo WATANABE Department of Chemistry, Faculty of Science, Tokyo Institute of Technology, Meguro-ku, Tokyo 152, *Department of Animal Science, Faculty of Agriculture, Hokkaido University, Sapporo, Hokkaido 060 Received for publication, June 9, 1978
Myosin-like protein and actin-like protein from E. coli formed filaments very similar in structure to those of myosin and actin from skeletal muscle. At 0.2 M KC1, a large number of "thick filaments" of uniform size (about 0.6-0.7 ftm long and about 20 nm wide) was present. These thick filaments aggregated as the KC1 concentration decreased to less than 0.2 M. Filaments of actin-like protein were decorated with muscle heavy meromyosin, showing "arrowheads". The arrowhead structure disappeared in the presence of ATP. A mixture of E. coli myosin-like protein and rabbit skeletal actin exhibited a gelation phenomenon on the addition of ATP. The phenomenon was reversible and showed ATP specificity. However, the gelation phenomenon was not observed with the mixture of E. coli actin-like protein and E. coli myosin-like protein. These results provide compelling evidence that the E. coli myosin-like protein and actinlike protein we isolated are essentially identical to myosin and actin, respectively.
In the previous paper (/), we reported the isolation of myosin-like protein and actin-like protein from E. coli. We identified the myosin-like protein on three counts: its mobility in SDS-gel electrophoresis, its ability to bind with rabbit skeletal actin, and its ATPase activities. We also identified the actin-like protein on four counts: its mobility in SDS-gel electrophoresis, its ability to bind with rabbit skeletal myosin, its polymerizing-depoly1
merizing ability, and its ability to activate MgATPase of myosin. In the present paper, we report that the two proteins can form filaments very similar in structure to those of myosin and actin from skeletal muscle. Moreover, we also report that the mixture of E. coli myosin-like protein and rabbit skeletal actin exhibits a gelation phenomenon on the addition of ATP, instead of superprecipitation.
MATERIALS AND METHODS Present address: Tokyo Metropolitan Institute of Medical Science, Department of Radiology, Bunkyo-ku, Actin-Like Protein and Myosin-LJke Protein Tokyo 113. Abbreviations: E. coli, Escherichia coll; SDS, sodium from E. coli—Actin-like protein and myosin-like dodccyl sulfate; EGTA, glycoetherdiamine-Af,A^,Ar'^v"- protein were prepared from E. coli by the methods described in our previous paper (7). The protein tetraacetic acid. Vol. 84, No. 6, 1978
1453
K. NAKAMURA, K. TAKAHASHI, and S. WATANABE
1454
concentration was determined by the biuret method (2), by the method of Lowry et al. (J), or by measuring the absorbance at 280 nm, using bovine serum albumin as a standard for each determination. Electron microscopy: The procedure for the electron microscopic study was essentially the same as those described by Huxley (4). Samples placed on carbon-coated grids were stained with 1 % uranyl acetate and examined under a Hitachi 11-B electron microscope, with an acceleration voltage of 75 kV. HMM: Skeletal heavy meromyosin (HMM) was kindly supplied by Dr. F. Morita of Hokkaido University. Diethylaminoethyl-cellulose chromatography: Protein solutions were dialyzed against a buffer solution containing 0.35 M KC1 and 20 mM TrisHC1 (pH 7.8), and applied to the column, which had been fully equilibrated with the same solution. Seven mg of protein was applied to a 1.5 cm x 5 cm column containing about 1.5 g of diethylaminoethyl (DEAE)-cellulose. Elution was carried out at a flow rate of 0.3 ml/s. All chromatographic procedures and all procedures for protein preparation were carried out at 0-4°C. DEAE-cellulose was purchased from Whatman and R. Balton Ltd. RESULTS /. DEAE-Cellulose Chromatography—In the previous paper (7), our preparation of myosin-like protein contained some non-protein substances. In fact, our preparation of E. coli myosin-like protein had a very high UV absorption at 260 nm (see Fig. l-2a). In that report (7), we estimated the protein concentration from the absorbance at 280 nm. We now determined the protein content of the same solution of myosin-like protein preparation by three different methods: measurement of the 280 nm absorbance, the biuret method (2), and Lowry's method (J). The protein concentration was found to be 1.6 mg/ml, 0.9 mg/ml, and 0.93 mg/ml, respectively. Therefore the low ATPase activities of E. coli myosin-like protein reported in the previous paper (7) should be increased by a factor of 1.8. We attempted DEAEcellulose chromatography by a modification of Mueller's method (5) to remove the non-protein substances. However, Fig. 1-1 shows that the
II
III IV V
1.0
9 O
s
Q.
CM
O CD
1.0
0.5
I
2 3 4 5
6 7 8
Fr.No Fig. 1-1. Chromatography of the myosin-like protein preparation on diethylaminocthyl cellulose. Protein solution (7 mg in 6 ml) in 0.35 M KC1, 20 mM Tris-HCl (pH 7.8) was applied to a column, 1.5cmx5cm. Elution was carried out at a flow rate of 0.3 ml/s with a solution containing 0.35 M KC1 and 20 mM Tns-HCl (pH 7.8). O, Absorbance at 280 nm; • , absorbance at 260 nm; • , the protein concentration determined by Lowry's method (3). absorbance of every fraction at 260 nm remained high, indicating that the myosin-like protein could not be fully separated from non-protein substances. Judging from the UV spectrum (Fig. l-2b), the content of non-protein substances was reduced, though only slightly, in fraction n . Therefore, fraction II was used for the following experiments. II. Thick Filament Formation—Electron microscopy revealed that out myosin-like protein preparation can form thick filaments. The thick filaments shown in Fig. 2 were obtained in the following way: 0.2 ml of E. coli myosin-like protein solution (0.8 mg/ml E. coli myosin-like protein in 0.6 M KCI, 10 mM imidazole-HCl, pH 7.2) was transferred into a test tube of 1.5 cm diameter. With constant stirring and taking care to avoid foaming of the solution, 0.4 ml of a diluting solution (10 mM imidazole-HCl, pH 7.2) was added / . Biochem.
MYOS1N AND ACTIN FROM E. coli
I5
20 .
SI 5 I Q5
1.0
260 280 300 Vtawltngth (run)
0 0.
260 280 300 WavfUngth (nm>
Fig. 1-2 UV spectra of the myosin-like protein preparation (a), and of the fraction obtained by DEAEcellulose chromatography of the myosin-like protein (b).
1455 slowly, i.e., taking 40 s. Aggregates were scarcely observed at 0 2 M K G , and a large number of " thick filaments" of uniform size (about 0.60.7 f*m long and about 20 nm wide) was seen. The electron micrograph also shows a small number of filaments with apparent bipolar structure. The nature of particles (8-9 nm in diameter) shown in Fig. 2 remains unclear (6). We observed that at 0.1 M K G , all thick filaments of E. coli myosin-like protein were packed into large aggregates (data not shown), and that the aggregates decreased in number as the K G concentration was increased to 0.2 M. These properties of E. coli myosin-like protein were reported by Mabuchi (7) with thick filaments of egg-cell myosin. ///. Thin Filament Formation—Electron microscopic study revealed that the actin-hke protein of E. coli can form long, fine filaments (Fig. 3-1). The width of the filaments was estimated to be approximately 6-9 nm. It is possible to see the presence of 365 A periodicity along the length of the filaments. The actin-hke protein preparation of E. coli
Fig. 2. Electron micrograph of thick filaments of E. coli myosin-like protein. The concentration of myosinlike protein was 0.25 mg/ml and the medium contained 0.2 M KQ and 10 ITIM imidazole-HCl (pH 7.2). The scale bar represents 0.5 ftm. Magnification: x54,000. Vol. 84, No. 6, 1978
1456
K. NAKAMURA, K. TAKAHASHJ, and S. WATANABE
Fig. 3-1. Electron micrograph of actm-hke filaments. The concentration of "actin-equivalent component" was 0.035 mg/ml and the medium contained 0.1 M KG and 5 IHM K phosphate (pH 7.2). Magnification: x 71,000, insert, x 140,000.
Fig 3-2. The binding of skeletal heavy meromyosin (HMM) to E. colt actin-like protein in the absence of ATP. The final concentrations of "actin-equivalent component" and HMM were 0.035 mg/ml and 0 25 mg/ml, respectively. The medium contained 0.1 M KCI and 10 m.M Tris-maleate (pH 6.8). Magnification: x 71,000. J.
Biochem.
MYOSIN AND ACTIN FROM E. coli
1457
Fig 3-3. Dissociation of E. coli actm-Iike protein and skeletal heavy meromyosin (HMM) in the presence of 1 mM ATP. The conditions were the same as those described in Fig. 3-2 except that the ATP concentration was 1 mM Magnification: x71,000.
ITT
Fig. 4
Vol. 84, No. 6, 1978
TV
was mixed with rabbit muscle HMM (approximately 0.25 mg/m!) in buffer solution of 0.1 M K G and 10 mM Tris-maleate (pH 6 8) or in this buffer solution plus I mM ATP. In the absence of ATP, many of the thin filaments became decorated with muscle HMM, showing "arrowheads", all of which pointed in the same direction (Fig. 3-2). The diameter of the decorated filaments (around 24 nm), the polarity indicated by the arrowheads and the repeating distance (36 nm) of these arrowheads resembled those reported for filaments of actin from other sources (7, 8). In the presence of
Fig. 4. Gelation phenomena Every sample contained O25mg/ml of E coli myosin-like protein (fraction II) and 0.13 mg/ml of rabbit skeletal actin. The medium contained 0 . 1 6 M KC1, 20 mM Tns-maleate (pH 6.8), I mM MgClj, together with (I) 0 I mM CaCl, and no ATP, (II) 0.1 mM CaCl, and 0.1 mM ATP, (III) 0.1 mM CaCl, and 0.1 mM AMP-PNP, or (IV) I mM EGTA and 0.1 mM ATP. The samples were kept standing without stirring at 20°C. The photos was taken 30 min after the addition of ATP or AMP-PNP (a), and three hours after the addition of ATP or AMP-PNP (b).
K NAKAMURA, K. TAKAHASHI, and S. WATANABE
1458
1 mM ATP, HMM dissociated from the filaments, those of rabbit skeletal myosin, but are somewhat leaving nearly bare filaments (Fig. 3-3) The wider (20 nm as compared with 15 nm). diameter of the bare thin filaments was the same As indicated by arrows in Fig. 3-1, we observed as that of the thin filaments in Fig. 3-1 short filamental structure in an electron micrograph of the E. coli actin-like protein. The structures IV. Gelation Phenomenon—We reported in the previous paper (/) that the E. coli myosin-like were larger in width (about 9 nm) and shorter in length than actin filaments. Moreover, the strucprotein in combination with rabbit skeletal actin tures were not decorated by rabbit muscle HMM. failed to show superprecipitation activity. We The decorated filaments of E. coli actin-like protein investigated whether the DEAE-cellulose chroare similar to those of muscle actin of Mycoplasma matographed preparation of E. coli myosin in pneumomae (8). The arrowhead structure of combination with skeletal actin could superHMM-decorated thin filaments of E coli actinprecipitate, using a medium containing 0 . 1 6 M like protein disappeared on addition of ATP. KCI, 20 mM Tris-maleate (pH 6.8), 0.1 mM CaCI2 or 1 mM EGTA, 1 mM MgCI,, and 0.1 mM ATP. These results provide compelling evidence that the thin filaments of E. coli actin-like protein described We again failed to observe superprecipitation are indeed those of actin. activity. However, we found that the mixture of E. coli myosin-like protein preparation (fraction A gelation phenomenon was reported to occur II) and rabbit skeletal actin formed a solid gel with actin preparations of non-muscle cells (9). upon addition of ATP (at 20°C) (Fig. 4a). The However, in our experiments, the gelation phenomenon was not observed with E. coli actin-like gelation phenomenon also occurred on the addiprotein, but was observed with skeletal actin and tion of 4-ATP, but did not occur on the addition E. coli myosin-like protein. Moreover, the gelation of ADP of AMP-PNP The protein mixture in the same medium containing calcium (0 1 mM phenomenon was found to be reversible and to have ATP specificity. Therefore, the gelation we CaCli) formed a solid gel within about 20 min observed was associated with the E. coli myosinafter the addition of ATP, and it recovered to a like protein rather than the E. coli actin-like sol a few hours later, probably when the ATP in protein. the medium had been exhausted (Fig. 4b). Addition of ATP produced a solid gel again. When We wish to express our thanks to Dr. Kazuo Sudo of the medium contained no calcium (1 mM EGTA), the University of Tokyo for his interest and encourageit took more than 20 min to form a solid gel. ment in connection with our electron microscope The time required for gel formation was also experiments dependent on the ATP concentration, that is, the time required increased as the ATP concentration REFERENCES was increased. DISCUSSION Mabuchi (7) reported that sea-urchin egg myosin forms an aggregated form of thick filaments in 0.1 M KCI, whereas rabbit skeletal myosin forms an unaggregated form of thick filaments in 0.1 M KCI. The thick filaments of E. coli myosin-like protein also are in the aggregated form in 0.1 M KCI, but are in the unaggregated form in 0.2 M KCI, where thick filaments of sea-urchin egg myosin are still in the aggregated form. E. coli myosin-like protein thick filaments are not very different in length (0.6-0.7 pm as compared with 0.5—1.0 pm) from
I. Nakamura, K. & Waianabe, S. (1978) J. Biochem. 83, 1459-1470 2 Gornall, A.G , Badawill, C J., & David, M.M. (1949) J. Biol. Chem. 177, 751-766 3. Lowry, O.H., Rosebrough, N J., Farr, A.L., & Randall, R J. (I95J) J. Biol. Chem. 193, 265-275 4. Huxley, H E (1963) J. Mol. Biol 7, 281-308 5. Mueller, H. & Perry, S V. (1961) Biochem. J. 80, 217-223 6 Ishihama, A., Ikeuchi, T., Matsumoto, A., & Yamamoto, S. (1976) J. Biochem 79, 927-936 7. Mabuchi, 1. (1976) J Mol. Biol. 100, 569-582 8. Neimark, H.C. (1977) Proc. Natl. Acad. Sci. U.S. 74, 4041^(045 9. Pollard, T.D. (1976) J Cell Biol. 68, 579-601
J. Biochem.
