Biochimica et Biophysica Acta, 412 (1975) 241-255
© Elsevier Scientific Publishing Company, Amsterdam- Printed in The Netherlands BBA 37198 MYOSIN F R O M A R T E R I A L S M O O T H MUSCLE: ISOLATION F O L L O W I N G A C T I N D E P O L Y M E R I Z A T I O N
JOSEPH MEGERMAN * and RICHARD A. MURPHY Department of Physiology, University of Virginia School of Medicine, Charlottesville, Va. 22901 (U.S.A.)
(Received June 2nd, 1975)
SUMMARY The contractile proteins from arterial smooth muscle are highly soluble, and can be extracted at I = 0.05. However, they can be precipitated by a prolonged dialysis at pH 6 to give an actomyosin with a high, although variable, actin:myosin ratio. The sedimentation behavior of this actomyosin at high ionic strength was examined as a function of pH, protein concentration and composition by preparative ultracentrifugation. Comparisons with synthetic skeletal muscle actomyosins of similar composition demonstrated significant differences in the behaviors of these two systems. It was found that much smooth muscle actomyosin is not dissociated by normally relaxing conditions, and that it sediments at a slower rate than F-actin. The solubility of the supernatant protein (a myosin-enriched actomyosin) in 0.2 M KC1 (pH 7) depended on the pH during centrifugation. A lower solubility was associated only with a higher actin concentration in the supernatant, suggesting a dependence on actin repolymerization. Pure myosin was selectively precipitated from the supernatant by polyethylene glycol-6000, but only when the protein was soluble at low ionic strength. The solubility of purified myosin was similar to that of myosin from striated muscles. A relationship between the presence of depolymerized actin and the high solubility of smooth muscle contractile proteins is suggested.
INTRODUCTION The contractile system of vertebrate smooth muscle differs from that of striated muscle in terms of (i) the mechanical output and efficiency of energy conversion as estimated from experiments on living tissue [1-3], (ii) the biochemical properties of the contractile proteins [1, 2, 4-8], and (iii) the organization of these proteins in a filament lattice [9-12]. In particular, the thick filaments appear to be
* To whom correspondance should be sent at Rosenstiel Basic Medical Sciences Research Center, Brandeis University, Waltham, Mass. 02154, U.S.A. Abbreviation: EGTA, ethyleneglycol-bis-(fl-aminoethylether)-N,N'-tetraacetic acid.
242 highly labile, the contractile system being much more soluble than with striated muscle [1, 4], and thin filaments far outnumber thick filaments [9], the relative amount of actin to myosin being 10-fold greater than in striated muscle [3, 13]. These differences in the contractile proteins contribute directly to the difficulties that are encountered in purifying smooth muscle myosin by methods that work with skeletal muscle myosin. The high solubility of the myosin and its simultaneous extraction with actin at low ionic strength prevent the use of differential extraction and precipitation, and the high ratio of actin to myosin tends to render differential ultracentrigugation ineffective. To overcome these problems, several researchers have resorted to the use of high concentrations of KI, which depolymerizes aetin, to extract actomyosin from muscle [6], or to separate actin and myosin in smooth and non-muscle extracts by gel filtration [14, 15]. Because of the potentially denaturing effect of KI, however, an effective purification method that uses more conventional means is still desirable. In the case of skeletal muscle, the ultracentrifugal separation of actin from myosin is readily accomplished at high ionic strength in the presence of ATP and pyrophosphate [16]. Measurements of ATPase activities under these conditions suggest the dissociation of smooth muscle actomyosin as well [8]. Furthermore, measurements of actomyosin ATPase activities at low ionic strength suggest an inhibition of the actomyosin interaction with relatively low concentrations of ATP [7]. It was, therefore, expected that myosin could be readily prepared from smooth muscle actomyosin by selectively sedimenting F-actin during centrifugation in 0.6 M KC1, 40 mM pyrophosphate and 5 mM ATP. This procedure was used with arterial actomyosin to obtain a soluble myosin-enriched supernatant from which the myosin could be selectively precipitated with polyethylene glycol-6000 [17]. However, the ultracentrifugal separation of myosin from actin was far from satisfactory. Two problems arose, depending on the pH that was maintained during centrifugation. (1) At pH 8, up to 7 0 ~ of the myosin contained in the actomyosin sample would sediment along with actin and was not easily recovered. (2) As a result of lower pH values, the myosin-enriched supernatant was not soluble enough for polyethylene glycol fractionation to be effective. The aim of this study was to determine if these difficulties were due to the incomplete dissociation of arterial actomyosin under conditions known to be effective with skeletal actomyosin, or to the high ratio of actin to myosin that is found in vertebrate smooth muscles [3, 13]. We have found the procedure for preparing myosin to greatly depend on the polymerization properties of actin. In addition, the sedimentation properties of myosin and actin under relaxing conditions diverge significantly from those known for the skeletal muscle proteins. They are not explained by the anomalous ratio of actin to myosin found in smooth muscle, but rather suggest the presence of undissociated actomyosin complexes of intermediate size. METHODS
Preparation of contractile proteins The preparation of smooth muscle proteins is described under Results. Skeletal muscle myosin was prepared from the hind limb muscles of cats or dogs by extraction
243 with 0.3 M KC1, 0.15 M potassium phosphate, 1 mM EDTA, 1 mM ATP, pH 6.5. The protein was precipitated by a 10-fold dilution with water and then centrifuged for 3 h at 130000 x g in 0.5 M KC1, 20 mM Tris buffer (pH 7.4 and 0 °C), 10 mM MgClz, 10 mM ATP. Only the top three-fourths of the supernatant were used. Myosin was precipitated by dialysis against 0.02 M KC1, 1 mM EDTA, centrifuged in 0.6 M KC1 (1 h at 48 000 x g) to sediment any remaining actomyosin or myosin aggregates, and reprecipitated by dialysis. Actin was prepared by the method of Spudich and Watt [18] from acetone-dried muscle powder prepared from the residue that remained after myosin extraction.
Analytical techniques Protein concentrations were determined by the biuret technique [19], standardized by micro-Kjeldahl methods assuming 1 6 ~ protein nitrogen content, or by absorbance of bovine serum albumin at 280 nm. The hydrolysis of ATP was determined by assay of released inorganic phosphate using the method of Rockstein and Herron [20], with absorbances measured at 720 nm. All reactions were run as time courses and rates were calculated from initial linear portions of the curve. The standard conditions for measuring the Ca 2÷activated ATPase activity of myosin were 5 mM CaC12, 5 mM ATP, 18 mM morpholinopropane sulfonic acid buffer (pH 7.0 at 25 °C), KC1 to give I = 0.6, and 0.10.3 mg myosin/ml. These ionic conditions were chosen to eliminate any contribution to measured activity by the Mg 2+-activated actomyosin complex. Actomyosin ATPase activity (used as a check for the presence of actin) was measured at I = 0.1, 5 mM MgC12, 0.5 mM ATP, 0.I mM CaC12, 18 mM morpholinopropane sulfonic acid buffer (pH 7.0 at 25 °C) [7]. The composition of all protein samples was determined by electrophoresis of sodium dodecyl sulfate-treated proteins, using the method of Weber and Osborn [21]. The relative contents of major components were obtained by planimetry of the major peaks of Coomassie blue-stained gel densitometer tracings. These ratios, along with the measured total protein content, provided estimates of the amounts of individual components [3]. Dilute samples that were eluted from gel chromatographic columns were precipitated by 1 0 ~ (final concentration) cold trichloroacetic acid, washed in cold acetone, and dissolved with 8 M urea before treatment with sodium dodecyl sulfate. Molecular sieve chromatography was performed on 4 ~ agarose beads (Biogel A-15 m, Bio-Rad Laboratories), in 1.4 x 60 cm columns having an approximate total volume of 100 ml. Flow rates of 5.5-6.5 ml/h were maintained by a peristaltic pump. Absorbance was monitored at 280 nm using a double beam absorbance monitor (ISCO Model UA-5) with elution buffer in the reference cell. All peaks were analyzed for protein composition by sodium dodecyl sulfate electrophoresis of the trichloroacetic acid-precipitated protein. Molecular weights were estimated from the manufacturer's calibration charts. Solubility was defined as the fraction of protein remaining in the supernatant after centrifugation at 10 000 × g for 30 min. Viscosities were measured with an Ostwald-type capillary viscometer that had an outflow time of about 2.5 min for water at 23 °C. Protein samples were cleared of particulates by centrifugation at 10 000 x g before being tested. Protein concentrations and pH were checked after each viscosity measurement.
244
~ ~
ARTERIES A.
Extract: .05M KCI, imM ATP, 2mM EGTA, pH 7.4
cell
debris
SOLUBLE EXTRACT
B.
Dialyze:
42mM
Potassiumphosphate, pH 6
ACTOMYOSIN C.
Ultracentrifuge: 0.6M KC1, 40mMpyrophosphate
soluble
fractions
(precipitate)
pH 6 ~ ~ p 8
5mMATP, imM EGTA SUPERNATANT ACTOMYOSIN
D.
Dialyze: 0.2M KCI, pH 7
(precipitate) E.
(soJuble)
Fractionate: 2% PEG-6000 MYOSIN
(precipitate)
Fig. i. Procedure for isolating myosin from hog carotid arteries and definition of terms. Normal procedure utilizes the pH 8 pathway for differential ultracentrifugation (step C). PEG, polyethyleneglycol. RESULTS
Preparation of actomyosin The preparation of contractile proteins from the smooth muscle of hog common carotid arteries is outlined in Fig. 1. After grinding, the arterial mince was rinsed in distilled water and then suspended in 7-10 ml/g low ionic strength extraction solution (0.05 M KC1, 18 mM morpholinopropane sulfonic acid buffer (pH 7.4, 0 °C), 2 mM ethyleneglycol-bis-(fl-aminoethyl ether) N,N'-tetraacetic acid (EGTA), and 1 m M ATP), thoroughly homogenized with a Polytron blender (Brinkmann Instruments), and stirred for 2 h. Tissue debris was removed by centrifuging for 20 min at 5000 x g. The supernatant was filtered through cheesecloth to remove lipid aggregates and then stirred for 30 min with 10 mg/ml D o w e x 21K (Bio-Rad Laboratories) anion exchange resin (previously washed and adjusted to pH 7 with KOH) to remove ATP. Approx. 70 % of the tissue content of myosin, actin, and tropomyosin was
245 TABLE I PROTEIN FRACTIONS IN ARTERIAL MYOSIN PREPARATION: YIELD AND COMPOSITION Values are mean ± S.E. N, number of preparations; n, number of determinations of composition by densitometry of sodium dodecyl sulfate electrophoresis gels. Fraction
Total media homogenate*** Soluble extract Actomyosin Supernatant actomyosin * Myosin
Yield: total protein Composition: percent of total protein* N
mg/g artery
n
Myosin
Actin
Tropomyosin**
5 30 29
48.3 :L 7.4 32.8 :]- 1.5 16.7 4- 1.0
11 3 7
16.1 4- 1.0 18.5 4- 0.6 30.9 ± 3.5
54.0 4- 2.0 44.8 4- 1.2 48.2 4- 3.7
30.8 4- 1.5 32.2 4- 4.1 20.0 4- 1.6
8
39.6 4- 3,4 --
23.9 4- 3.4 --
35.2 -k 3.4 --
35 10
3.7 4- 0.2 0.97 4- 0.10
* Myosin + actin + tropomyosin = 100 700for each determination. ** This peak on sodium dodecyl sulfate was operationally defined as tropomyosin. Up to 50 700of this amount may be some other, unidentified, protein (ref. 3). *** Ref. 3. * Includes centrifugation of concentrated protein samples which tend to give lower yields. extracted by this method (Table I). All o f the protein with molecular weight approx. 37 000 on sodium dodecyl sulfate gels is here defined as tropomyosin. However, only half o f this fraction has been verified to be t r o p o m y o s i n [3], the remainder being unidentified. After decanting from the resin beads, the extract was dialyzed against 42 m M potassium phosphate, p H 6, for 24-48 h, with several changes of buffer, to precipitate actomyosin. The insoluble actomyosin was sedimented by centrifugation (30 min at 12 000 × g). The supernatant was saved for actin preparation. The precipitate contained half the extracted proteins, including all the myosin, leaving mostly actin and t r o p o m y o s i n in the supernatant.
Preparation of act& A crude actin preparation was obtained f r o m the supernatant remaining after the initial actomyosin sedimentation (Step B, Fig. 1) by adding polyethylene glycol 6000 to 2 ~ . The resultant precipitate was dissolved in 2 m M Tris. HCI, 0.2 m M ATP, 0.5 m M 2-mercaptoethanol and 0.2 m M CaCI2, p H 8 [18]. This actin was stored without repolymerization; gel electrophoresis showed little contamination with other proteins.
Differential centrifugation of actomyosin in relaxing medium The actomyosin was dissolved in 0.6 M KCI, 40 m M sodium p y r o p h o s p h a t e (pH 8.0), 5 m M A T P and 1 m M E G T A , and centrifuged 5 h at 50 000 rev./min in the Beckman 60 Ti r o t o r (176 000 × g average force). These conditions were initially chosen to dissociate actomyosin, to minimize myosin aggregation, and to sediment F-actin. Analysis o f the resultant supernatant by sodium dodecyl sulfate gel electrophoresis showed that increasing the p H that was maintained during ultracentrifugation f r o m 6 to 8 resulted in less protein remaining in that supernatant. That is, the
246 sedimentation of actomyosin increased with pH. The fraction of myosin that remained in the supernatant showed a negative linear correlation with pH: at pH 6, the supernatant contained most of the myosin of the initial actomyosin, while at pH 8, between 60 and 80 ~ was sedimented. Actin followed a similar trend with limits of 70 ~ at pH 6 and 85 ~o at pH 8, while a constantly low fraction of tropomyosin sedimented regardless of pH. In addition, the pellets obtained after 5 h of centrifugation were noticeably different. At pH 6, the pellet had an opaque cream-colored appearance, while at pH 8, the pellet had a small opaque center overlayed with a large amount of translucent material. Examination by sodium dodecyl sulfate electrophoretic gels showed the opaque centers to consist of nearly pure actin while the translucent overlays contained both actin and myosin. The tendency for actin and myosin to sediment together to form the lighter overlay portion of the pellets as the pH was increased suggested an incomplete dissociation even in the presence of 0.6 M KC1, 40 mM pyrophosphate and 5 mM ATP. Several observations support this interpretation. (1) Using 0.6 M KI or KSCN instead of KC1 increased the total supernatant protein by 79 and 34 ~o, respectively, although the relative amounts of actin and myosin remained unchanged. (2) Increasing the temperature caused a marked decrease in supernatant protein. At 25 °C the supernatant contained virtually no actin or myosin, while the amount of tropomyosin was unaffected. (3) The reduced viscosity of the actomyosin, measured before ultracentrifugation, was consistently higher at pH 8 than at lower values (400 ml/g at pH 8 vs 140 ml/g at pH 6 for a typical experiment). At pH 8, the protein was highly thixotropic, forming a viscous gel on standing at 4 °C. This gelation did not occur at lower pH. The effect of total protein concentration on the sedimentation of arterial actomyosin, at constant pH, is shown in Fig. 2. The increase in sedimentation (decrease in supernatant protein) with increasing concentration is consistent with the
2
E
5
I0
Initial c o n c e n t r a t i o n
20
o f a c t o r n y o s i n ( m g / ml)
Fig. 2. Fraction of total protein remaining in supernatant, after differential ultracentrifugation of actomyosin in 0.6 M KCI, 40 mM sodium pyrophosphate, 5 mM ATP, 1 mM EDTA, pH 8, 0 °C, as a function of the initial protein concentration. Line was fitted by method of least squares and represents significant correlation.
247 100 ¢c-
B
80 X
E
0
I'--
0 0
0--
--
--
-0
U
C 0
cL. 2C
~--~
c o Initiol concentrotion of myosin or octin ( m g / m l )
Fig. 3. Fraction of myosin or actin remaining in the supernatant following differential ultracentrifugation of actomyosin in conditions of Fig. 2, as a function of their initial concentration in actomyosin. Arterial actomyosins (open symbols) and skeletal actomyosins (closed symbols) of different composition were prepared as described in the text. (3 and O, myosin; A, actin. Crosses (x) represent polyethylene glycol-purified arterial myosin without the addition of actin. Solid lines represent significant correlations obtained by method of least squares. presence of protein-protein interaction. To attempt a more detailed description of these interactions, the composition of actomyosin preparations and of their corresponding supernatants after ultracentrifugation, as determined from gel electrophoresis, was used to infer the sedimentation behavior of each component protein. These data were obtained from several samples of "natural" actomyosin, prepared as described above, and of "synthetic" actomyosins that were prepared to test the effect of varying the ratio of actin to myosin (see below). The sedimentation of the myosin component in actomyosin increased significantly (supernatant content decreased) with increasing initial concentration (Fig. 3, open circles). Less than 30 7o remained in the supernatant at initial concentrations of 2.5 mg/ml or greater. Little tropomyosin sedimented regardless of initial concentration. Little of the arterial actin remained in the supernatant, although slightly more remained when higher initial concentrations of actin in actomyosin were used (Fig. 3, triangles). Similar experiments were carried out with synthetic actomyosin that was prepared from skeletal myosin and actin. In this system, the sedimentation of the myosin component was much lower and independent of concentration (Fig. 3, closed circles), while the supernatant content of actin decreased slightly with its initial concentration (not shown). Similarly, when pure arterial myosin was used, less than 30 7o sedimented, although initial concentrations of up to 6.5 mg/ml were used (Fig. 3, crosses). These results were all obtained at p H 8. As described above, decreasing p H resulted in more arterial myosin remaining in the supernatant. In two experiments with pure skeletal myosin, sedimentation was not noticeably different at p H 6.
248 The ratio of actin to myosin (actin:myosin) in arterial smooth muscle is nearly 10-fold greater than in skeletal muscle [3, 13]. When actomyosin is precipitated by dialysis of the muscle extract to low ionic strength, this ratio is approximately maintained. We thus considered the possibility that arterial actin and myosin were intrinsically similar to their skeletal muscle counterparts, but that the unusual behavior of arterial actomyosin upon centrifugation was due to the increased actin:myosin. Actomyosins of different actin:myosin were therefore centrifuged under the same conditions used in purifying the arterial proteins. Arterial actomyosins of lower actin:myosin were produced by adding purified arterial myosin to actomyosin, or by using a preparation obtained from the translucent upper portion of pH 8 ultracentrifugation pellets (see above). Skeletal actomyosins of high actin:myosin were also prepared, by combining pure myosin and actin in desired amounts. Tropomyosin was absent in these latter preparations. The interaction of actin and myosin after mixing was verified by measuring a decrease in the viscosity of the mixture upon the addition of ATP, before centrifugation. The results of centrifugation, already described above, were analyzed with regard to the ratio of actin to myosin. There was no correlation between the sedimentation of actin or myosin and their relative content in actomyosin. MgZ+-ATP is the dissociating agent usually employed with skeletal actomyosin, and its use might have enhanced the separation of skeletal actin and myosin. However, the inclusion of Mg 2;-, equimolar to ATP, in three arterial actomyosin experiments produced no significant differences.
Dialysis of the supernatant actomyosin The supernatant that was obtained after differential ultracentrifugation was stirred for 30 min with 40 mg/ml Dowex 21K (equilibrated to pH 7) to remove ATP and pyrophosphate, and then dialyzed against 0.2 M KCI, pH 7. When centrifugation was carried out at pH 8, the protein was highly soluble, even when the concen-
c" 100 f
8°I
60
c_
4o
~d c 20 _o
6
7
8
9
pH Fig. 4. PIecipitation of supernatant actomyosin dr, ring dialysis in 0.2 M KCI (pH 7) as a function of the pH maintained during the preceding ultracentrifugation. Total protein was measured for the soluble fractions before and after dialysis, as well as for the precipitates dissolved in 3 M KCI. Points indicate means - S.E., four preparations. Line is least squares fit of all unaveraged data.
249 tration of KC1 was reduced to as low as 0.05 M. This fraction was used to purify myosin by polyethylene glycol fractionation (see below). However, if the pH during centrifugation was less than 8, a significant amount of supernatant actomyosin precipitated after dialysis to 0.2 M KC1 at pH 7 (Fig. 4). Since all samples were dialyzed against the same neutral solution the effect of p H was clearly indirect. That is, during dialysis, the protein "remembered" the pH at which it had been centrifuged. Skeletal muscle myosin, when treated as above, always pricipitated completely in 0.2 M KC1, regardless of the pH of prior centrifugation. Combining the data of Fig. 4 with the observed increase in supernatant concentration of actin as a function of pH during ultracentrifugation reveals a direct dependence of actomyosin precipitation in 0.2 M KCI (pH 7) on the supernatant actin concentration (Fig. 5). No such correlation was found with the myosin concen-
c-
100
~
8o
~
6O
t/) r = 0.806
4O cO
U_
2O
o
I 0
.2
.4
.6
.8
1.0
1.2
1.4
Actin Concentration ( m g / m l )
Fig. 5. Precipitation of supernatant actomyosin during dialysis in 0.2 M KC1 (pH 7) as a function of actin concentration in the supernatant, r, correlation coefficientfound by the method of least squares. tration. The precipitation of this supernatant actomyosin, after centrifugation at low pH, could b e prevented by dialysing against KI instead of KC1 (both at pH 7). Although the concentration of actin in the supernatant was significant after centrifugation at pH 8 (Table I), the normal tests for the presence of actin gave negative results. ATPase activity in the presence of Mg 2+, at I = 0.1, was extremely low ( < 0.001 #mole P J m g per min, 10 preparations), ATP sensitivity (changes in the viscosity number upon the addition of ATP at I ---- 0.6 [22]) was less than 5 7O, and the clearing response on the addition of ATP, at I = 0.1, was negligible. The apparent lack of myosin binding by the supernatant actin suggested it was depolymerized or denatured. To determine the states of actin aggregation, the ultracentrifuge supernatant was chromatographed on 4 7o agarose (Fig. 6). The presence of a void peak (6-7 h) was Variable and reflected differences in the protein preparations. When a void fraction was present, it contained mostly actin. The intermediate peak (8-t0 h) contained mostly myosin, but with some a'ctin, while the third peak (after 13 h) contained similar amounts of actin and tropomyosin. The high absorbance of this final peak was due to the ATP in the applied sample. Under similar ionic conditions, pure
250 Q2
1.0
Eo.~
o.5,~ A
e~,' ;
M
lo
15
Elution time,h
Fig. 6. Chromatography of supernatant actomyosin on 4 % agarose in 0.6 M KC1, 40 mM sodium pyrophosphate, 1 mM EGTA, pH 8, 0 °C. Sample (2 ml) was applied to the column immediately following ultracentrifugation of actomyosin. The large absorbance of the final peak is due to the ATP that had been contained in the sample, in addition to low molecular weight proteins. Arrow indicates void volume. Letters indicate major composition of peaks as determined by gel electrophoresis: A, actin; M, myosin; T, tropomyosin. skeletal myosin was eluted between 8 and 10 h, which corresponds to a molecular weight of 4 000 000. The presence of actin in the fractions eluting after 13 h, corresponding to a molecular weight of less than 100 000, verifies the significant degree of actin depolymerization at high ionic strength.
Purification of myosin by polyethylene glycolfraetionation After dialysis to 0.2 M KC1 (pH 7), the supernatant was adjusted to 2 mg protein/ml. Myosin was then selectively precipitated from the pH 8 supernatant by adding polyethylene glycol-6000 (Baker Chemicals, 20 % (w/v) solution in 0.2 M KC1) to give a final concentration of 2%. The myosin was collected by centrifugation (10 000 x g), washed twice with 50 mM KC1 and dissolved in 0.6 M KC1. When the supernatant actomyosin was less soluble following ultracentrifugation at lower pH (Fig. 4), purification of the myosin by polyethylene glycol fractionation was no longer effective. Even when this less soluble actomyosin was dissolved in KC1 concentrations of 0.3 M KC1 or higher, the addition of polyethylene glycol resulted in the precipitation of both actin and myosin. It appeared that fractionation by polyethylene glycol was possible only when myosin was soluble at low ionic strength as occurred after centrifugation at pH 8. The addition of polyethylene glycol in solid form (to 2 %) precipitated less myosin than when polyethylene glycol was added as a 20% solution (34.0 i 3.3% vs 85.4 ± 12.6 % of the total myosin, means ± S.E., 6 preparations). Furthermore, the myosin thus obtained may have been denatured as evidenced by a lower Ca 2+ATPase activity. The purity of the myosin that was precipitated by polyethylene glycol-6000 was demonstrated by the presence of a single major protein peak that co-electrophoresed with skeletal and cardiac myosins on sodium dodecyl sulfate gels. One or two faint bands corresponding to molecular weights less than 15 000 were sometimes present, but not consistently enough to identify them as light chains of myosin. The absence of these low molecular weight components was not associated with any decrease in Ca2+-ATPase activity. The average yield of purified myosin was 0.97 4-
251 0.1 mg/g artery (mean ± S.E., 10 preparations). Its Ca 2+-dependent ATPase activity, at I = 0.6, averaged 0.187 :k 0.006 #mol Pi/mg per min (8 preparations), and decreased markedly with decreasing ionic strength. The similar dependence of smooth muscle actomyosin ATPase activity on ionic strength is well documented[5, 6, 8]. A 2-fold increase in activity in the presence of denaturing agents, such as 2 M urea [6, 23] was also found. In the presence of Mg 2+, at I = 0.1, myosin ATPase activity rose from undetected levels to 0.018/~mol PJmg per min upon the addition of either skeletal or arterial actins (approx. 2 mg actin/mg myosin). This value is comparable to those previously reported for native arterial actomyosin [7, 8].
Effect of polyethylene glycol on myosin After being precipitated by polyethylene glycol myosin was insoluble at I _< 0.25, even after being dissolved in high ionic strength. Its solubility characteristics thus became similar to those of skeletal muscle myosin. This loss of solubility was not irreversible, as re-centrifugation of the polyethylene glycol-precipitated myosin at pH 8 (as with actomyosin, step C, Fig. 1) rendered much of it soluble again when dialyzed to low ionic strength. The change that occurs in the solubility of myosin is probably not the result of denaturation, as all of the myosin Ca z+-ATPase activity was retained in the protein precipitated by polyethylene glycol. Furthermore, estimates of the myosin content, from gel electrophoresis, in actomyosin samples that had not been treated with polyethylene glycol were found to agree well with estimates based on the measured CaZ+-ATPase activity of these impure samples, assuming the polyethylene glycolpurified myosin to have 100 ~o activity (Fig. 7). This correlation supports the assumption that, for actomyosin samples, the CaZ+-ATPase activity measured at I = 0.6 is due to myosin alone, and also justifies the use of this assay as a measure of myosin content. It does not, however, rule out the presence of some activating effect of polyethylene glycol on myosin Ca2+-ATPase activity (which would compensate for an
100 i/1 u
0 "~
~
oE V]
0
.-I/1 O
113 t-
~
~g
60
4o
•
r = 0.905
20 eo
o 0
I
I
I
I
i
20
40
60
80
100
% Myosin (Ca-ATPase activity) Fig. 7. Correlation of myosin fractions in actomyosins as determined from Ca2+-ATPase activity measurements and by densitometry of sodium dodecyl sulfate gels. Polyethylene glycol-purified myosin was assumed to be 100 % pure and fully active for comparisons with actomyosin activity and composition, r, correlation coefficient determined by method of least squares.
252 equivalent loss of activity due to denaturation), although the absence of such an effect on skeletal muscle myosin (Megerman, J., unpublished) renders this coincidence unlikely. DISCUSSION This procedure for purifying smooth muscle myosin was developed because the methods that are commonly applied to the preparation of skeletal muscle myosin are rendered ineffective by the unusual solubility of myosin and the high relative content of actin in actomyosin extracts. It is clear that the method permits the effective purification of smooth muscle myosin. However, it suffers from several drawbacks: (i) the selective precipitation of myosin by polyethylene glycol is effective only when the supernatant actomyosin is soluble at low ionic strength (0.2 M KC1, pH 7); (ii) this highly soluble actomyosin occurs only under conditions which result in a sizable fraction of the myosin sedimenting during ultracentrifugation (pH 8); and (iii) the precipitation of myosin by polyethylene glycol may cause significant changes in the solubility of the myosin.
Myosin purification depends on actin depolymerization The effect of polyethylene glycol on solubility depends on the molecular size of a protein; larger proteins are precipitated by lower concentrations of polyethylene glycol [24]. Furthermore, polyethylene glycol is known to enhance the interaction between actin and myosin [25]. It follows, therefore, that actin in the soluble ultracentrifuge supernatant (after pH 8) had a smaller molecular size than myosin, and that this actin did not bind myosin. The absence of myosin binding is supported by the negative results of Mg 2+-ATPase and ATP sensitivity tests. The retarded position of actin upon gel filtration of the soluble supernatant actomyosin (Fig. 6) suggests this actin was depolymerized, or nearly so. But this depolymerized actin was probably not denatured, since the precipitation of actomyosin as a function of actin concentration in the supernatant (Fig. 5) demonstrates the ability of the actin to bind myosin and form an insoluble complex, given that it was present in sufficient amounts. This precipitation of actomyosin did not correlate with the concentration of myosin in the supernatant, even though the concentration of this protein varied along with that of actin as a function of pH during ultracentrifugation. We suggest that Fig. 5 reflects a dependence of the polymerization (and perhaps nucleation) of actin on its concentration [26], and that the process is required to form insoluble actomyosin. The fact that no precipitate was formed when KI was substituted for KC1 is consistent with this hypothesis, as KI is known to prevent actin polymerization. Conversely, the successful application of polyethylene glycol fractionation to the purification of myosin appears to require that any actin that is present be in a depolymerized form, as whenever the supernatant actomyosin became insoluble upon dialysis to 0.2 M KC1 (pH 7), purification by this method was not possible. Sedimentation behavior of actomyosin Even though the arterial actomyosin was centrifuged under conditions which,
253 with skeletal actomyosin, would ensure complete dissociation, the negative correlation between the amount of supernatant protein and the initial concentration of actomyosin (Fig. 2) suggests an interacting protein system. The questions to be answered are (1) which of the possible interactions between actin, myosin, and tropomyosin are dominating, and (2) does the arterial actomyosin system behave in a manner significantly different than could be expected with skeletal muscle proteins. An interaction involving tropomyosin is considered least likely, as this protein is readily dispersed in high ionic strength [27]. This is borne out by the low sedimentation of tropomyosin and the absence of any dependence on pH or concentration. However, the sedimentation dependence of the myosin component of arterial actomyosin on its concentration (Fig. 3) is evidence of myosin interaction, but it does not distinguish between myosin self-aggregation or an interaction with actin. The latter is supported, though by no means proven, by the decreased sedimentation of both actin and myosin at lower pH and temperature, and in the presence of KI and KSCN. Moreover, the low sedimentation ( < 30 ~) of pure arterial myosin, even at an initial concentration (6.5 mg/ml) which would result in its complete sedimentation if it were part of actomyosin (Fig. 3, solid line), virtually eliminates myosin self-aggregation from consideration. This low value is similar to the amount sedimented when pure skeletal myosin was centrifuged under identical conditions, and is consistent with the low sedimentation constant (6 S) obtained for a dispersed monomeric myosin. The sedimentation of skeletal myosin was independent of pH between 6 and 8, whereas that of arterial myosin (in actomyosin) was quite sensitive to pH. It is clear that the low sedimentation of pure arterial myosin reflects a predominantly dispersed, non-aggregated state. When actin is present, it is the sedimentation of some actomyosin complex that must be considered. The apparent molecular weight of myosin under conditions similar to those obtaining during ultracentrifugation, as determined by gel filtration, was 4 000 000. This large value reflects molecular size (Stokes radius) rather than weight [28], and is consistent with the finding from light scattering measurements that myosin aggregates containing 1 to 8 monomers (molecular weight approx. 500 000 each) have similar Stokes radii [29]. The similar elution patterns for skeletal and smooth muscle myosins provide no information, therefore, as to the similarity of aggregation states for these two proteins. Most of the experiments with arterial actomyosin were carried out in the absence of Mg 2÷, as this ion is thought to promote the aggregation of smooth muscle myosin [30]. The inclusion of MgC12 produced no significant differences in the content or composition of the supernatants, indicating that the incomplete dissociation of actin from myosin, in the presence of ATP, was not due to the absence of Mg 2÷. Furthermore, synthetic skeletal actomyosin that was prepared to reflect the composition of arterial actomyosin did not behave as did the latter. Even without Mg 2+, less myosin from skeletal actomyosin sedimented than from arterial actomyosin. The behavior of the smooth muscle proteins is therefore not intrinsically similar to that of the skeletal muscle proteins and is not merely the result of the high ratio of actin to myosin that exists in arterial actomyosin. From these comparisons, we conclude that the observed interaction between arterial actin and myosin reflects some inherent differences in the binding properties of these smooth muscle proteins.
254
Intermediate actomyosin complexes It is not yet possible to describe in detail the actomyosin complex that resists dissociation. The distinct two-part appearance of the pellets following ultracentrifugation at pH 8 (where much myosin is sedimented) compared with pH 6 (where little myosin sediments) suggests this complex has a much smaller sedimentation coefficient than does F-actin. The actomyosin mixture being centrifuged at pH 8 thus contains at least three different species of actin: F-actin which sediments rapidly and is independent of myosin, intermediate actin units which co-sediment with myosin, and depolymerized actin which does not bind myosin. The depolymerization of actin could result from the high pressures generated during ultracentrifugation [26], or may reflect a natural tendency for smooth muscle actin. It is not clear if, at pH 6, the intermediate actin units are fewer in number, if the reduced pH prevents their association with myosin, or if possible actomyosin complexes are simply too small to sediment.
Solubility of myosin and of actomyosin The dialysis experiments in 0.2 M KC1 verify that myosin from arterial smooth muscle tends to be more soluble than from skeletal muscle, as was assumed from the extractability of the former at low ionic strength. It is still to be determined, however, whether this solubility is an intrinsic property of the myosin molecule itself, or whether it is due to an association with some solubilizing factor [2, 17]. For instance, myosin may be bound to small units of non-polymerized actin, which are soluble until the polymerization of this actin occurs. This hypothesis is supported by the striking change in myosin solubility that accompanies its purification by polyethylene glycol, but is opposed by the partial resumption of high solubility following the centrifugation of pure myosin. A decrease in actomyosin solubility in vitro, is associated with an increase in its Mg2+-ATPase activity at low ionic strength [8]. But the greatest values of activity reported so far are low compared to that for skeletal actomyosin, and they require that Mg 2+ be present in concentrations 10-fold greater than with skeletal actomyosin [7]. It has been postulated that this Mg z+ acts to chelate ATP and thus reduce its dissociating effect on actomyosin [7], or to reduce the solubility of myosin [31]. But Mg 2÷ has a polymerizing effect on actin as well [26]. We therefore suggest a possible function for Mg 2+ in promoting the repolymerization of actin, which produces actomyosin complexes that are large, insoluble, and have higher levels of MgZ+-activated ATPase activity. ACKNOWLEDGEMENTS The technical assistance of Mrs Mildred Smythers and the generous donation of arteries by George H. Meyers and Sons, Richmond, Virginia, are gratefully acknowledged. J. Megerman was a predoctoral trainee (N.I.H. Grant HL 05815) during this investigation and many of the results were included in his doctoral dissertation at the University of Virginia. R. A. Murphy is the recipient of a career development award (HL 11747). This research was supported by N.I.H. grant HL14547.
255 REFERENCES 1 Somlyo, A. P. and Somlyo, A. V. (1968) Pharmacol. Rev. 20, 197-272 2 Riiegg, J. C. (1971) Physiol. Rev. 51,201-248 3 Murphy, R. A., Herlihy, J. T. and Megerman, J. (1974) J. Gen. Physiol. 64, 691-705 4 Laszt, L. and Hamoir, G. (1961) Biochim. Biophys. Acta 50, 430-449 5 Needham, D. M. and Williams, J. M. (1963) Biochem. J. 89, 552-561 6 BCtr~iny,M., BCtr~ny, K.. Gaetjens, E. and Bailin, G. (1966) Arch. Biochem. Biophys. 113,205-222 7 Murphy, R. A., Bohr, D. F. and Newman, D. C. (1969) Am. J. Physiol. 217, 666-673 8 Murphy, R. A. (1971) Am. J. Physiol. 220, 1494-1500 9 Somlyo, A. P., Somlyo, A. V., Devine, C. E. and Rice, R. V. (1971) Nat. New Biol. 231,243-246 10 Small, J. V. and Squire, J. M. (1972) J. Mol. Biol. 67, 117-149 11 Cooke, P. J. and Fay, F. S. (1972) Exp. Cell Res. 71,265-272 12 Bois, R. M. (1973) Anat. Rec. 117, 61-78 13 Tregear, R. T. and Squire, J. M. (1973) J. Mol. Biol. 77, 279-290 14 Pollard, T. D., Thomas, S. M. and Niederman, R. (1974) Anal. Biochem. 60, 258-266 15 Nachmias, V. T. (1974) J. Cell Biol. 62, 54-65 16 Weber, A. (1956) Biochim. Biophys. Acta 19, 345-351 17 Megerman, J. and Murphy, R. A. (1973) Biophys. J. 13, 321a 18 Spudich, J. A. and Watt, S. (1971) J. Biol. Chem. 246, 4866--4871 19 GornaU, A. G., Bardawill, C. J. and David, M. M. (1949) J. Biol. Chem. 177, 751 20 Rockstein, M. and Herron, P. W. (1951) Anal. Chem. 23, 1500-1501 21 Weber, K. and Osborn, M. (1969) J. Biol. Chem. 244, 4406--4412 22 Weber, H. H. and Portzehl, H. (1952) Adv. Protein Chem. 7, 161-252 23 Gaspar-Godfroid, A. (1970) Angiologica (Basel) 7, 273-290 24 Juckes, I. R. M. (1971) Biochim. Biophys. Acta 229, 535-546 25 Kaldor, G., Hsu, Q. S;, Wachsberger, P. and Chowrashi, P. (1971) Am. J. Physiol. 220, 639-645 26 Oosawa, F. and Kasai, M. (1971) in Subunits in Biological Systems (Timasheff, S. N. and Fasman, G. D., eds), Part A, pp. 261-322 27 Drabikowsky, W. and Nowak, E. (1968) Eur. J. Biochem. 5, 376-384 28 Ackers, G. K. (1970) Adv. Protein Chem. 24, 343-346 29 Lowey, S. and Holtzer, A. 0959) J. Am. Chem. Soc. 81, 1378-1383 30 Shoenberg, C. F. (1969) Tissue Cell 1, 83-96 31 Russell, W. E. (1973) Eur. J. Biochem. 33, 459-466
© Elsevier Scientific Publishing Company, Amsterdam- Printed in The Netherlands BBA 37198 MYOSIN F R O M A R T E R I A L S M O O T H MUSCLE: ISOLATION F O L L O W I N G A C T I N D E P O L Y M E R I Z A T I O N
JOSEPH MEGERMAN * and RICHARD A. MURPHY Department of Physiology, University of Virginia School of Medicine, Charlottesville, Va. 22901 (U.S.A.)
(Received June 2nd, 1975)
SUMMARY The contractile proteins from arterial smooth muscle are highly soluble, and can be extracted at I = 0.05. However, they can be precipitated by a prolonged dialysis at pH 6 to give an actomyosin with a high, although variable, actin:myosin ratio. The sedimentation behavior of this actomyosin at high ionic strength was examined as a function of pH, protein concentration and composition by preparative ultracentrifugation. Comparisons with synthetic skeletal muscle actomyosins of similar composition demonstrated significant differences in the behaviors of these two systems. It was found that much smooth muscle actomyosin is not dissociated by normally relaxing conditions, and that it sediments at a slower rate than F-actin. The solubility of the supernatant protein (a myosin-enriched actomyosin) in 0.2 M KC1 (pH 7) depended on the pH during centrifugation. A lower solubility was associated only with a higher actin concentration in the supernatant, suggesting a dependence on actin repolymerization. Pure myosin was selectively precipitated from the supernatant by polyethylene glycol-6000, but only when the protein was soluble at low ionic strength. The solubility of purified myosin was similar to that of myosin from striated muscles. A relationship between the presence of depolymerized actin and the high solubility of smooth muscle contractile proteins is suggested.
INTRODUCTION The contractile system of vertebrate smooth muscle differs from that of striated muscle in terms of (i) the mechanical output and efficiency of energy conversion as estimated from experiments on living tissue [1-3], (ii) the biochemical properties of the contractile proteins [1, 2, 4-8], and (iii) the organization of these proteins in a filament lattice [9-12]. In particular, the thick filaments appear to be
* To whom correspondance should be sent at Rosenstiel Basic Medical Sciences Research Center, Brandeis University, Waltham, Mass. 02154, U.S.A. Abbreviation: EGTA, ethyleneglycol-bis-(fl-aminoethylether)-N,N'-tetraacetic acid.
242 highly labile, the contractile system being much more soluble than with striated muscle [1, 4], and thin filaments far outnumber thick filaments [9], the relative amount of actin to myosin being 10-fold greater than in striated muscle [3, 13]. These differences in the contractile proteins contribute directly to the difficulties that are encountered in purifying smooth muscle myosin by methods that work with skeletal muscle myosin. The high solubility of the myosin and its simultaneous extraction with actin at low ionic strength prevent the use of differential extraction and precipitation, and the high ratio of actin to myosin tends to render differential ultracentrigugation ineffective. To overcome these problems, several researchers have resorted to the use of high concentrations of KI, which depolymerizes aetin, to extract actomyosin from muscle [6], or to separate actin and myosin in smooth and non-muscle extracts by gel filtration [14, 15]. Because of the potentially denaturing effect of KI, however, an effective purification method that uses more conventional means is still desirable. In the case of skeletal muscle, the ultracentrifugal separation of actin from myosin is readily accomplished at high ionic strength in the presence of ATP and pyrophosphate [16]. Measurements of ATPase activities under these conditions suggest the dissociation of smooth muscle actomyosin as well [8]. Furthermore, measurements of actomyosin ATPase activities at low ionic strength suggest an inhibition of the actomyosin interaction with relatively low concentrations of ATP [7]. It was, therefore, expected that myosin could be readily prepared from smooth muscle actomyosin by selectively sedimenting F-actin during centrifugation in 0.6 M KC1, 40 mM pyrophosphate and 5 mM ATP. This procedure was used with arterial actomyosin to obtain a soluble myosin-enriched supernatant from which the myosin could be selectively precipitated with polyethylene glycol-6000 [17]. However, the ultracentrifugal separation of myosin from actin was far from satisfactory. Two problems arose, depending on the pH that was maintained during centrifugation. (1) At pH 8, up to 7 0 ~ of the myosin contained in the actomyosin sample would sediment along with actin and was not easily recovered. (2) As a result of lower pH values, the myosin-enriched supernatant was not soluble enough for polyethylene glycol fractionation to be effective. The aim of this study was to determine if these difficulties were due to the incomplete dissociation of arterial actomyosin under conditions known to be effective with skeletal actomyosin, or to the high ratio of actin to myosin that is found in vertebrate smooth muscles [3, 13]. We have found the procedure for preparing myosin to greatly depend on the polymerization properties of actin. In addition, the sedimentation properties of myosin and actin under relaxing conditions diverge significantly from those known for the skeletal muscle proteins. They are not explained by the anomalous ratio of actin to myosin found in smooth muscle, but rather suggest the presence of undissociated actomyosin complexes of intermediate size. METHODS
Preparation of contractile proteins The preparation of smooth muscle proteins is described under Results. Skeletal muscle myosin was prepared from the hind limb muscles of cats or dogs by extraction
243 with 0.3 M KC1, 0.15 M potassium phosphate, 1 mM EDTA, 1 mM ATP, pH 6.5. The protein was precipitated by a 10-fold dilution with water and then centrifuged for 3 h at 130000 x g in 0.5 M KC1, 20 mM Tris buffer (pH 7.4 and 0 °C), 10 mM MgClz, 10 mM ATP. Only the top three-fourths of the supernatant were used. Myosin was precipitated by dialysis against 0.02 M KC1, 1 mM EDTA, centrifuged in 0.6 M KC1 (1 h at 48 000 x g) to sediment any remaining actomyosin or myosin aggregates, and reprecipitated by dialysis. Actin was prepared by the method of Spudich and Watt [18] from acetone-dried muscle powder prepared from the residue that remained after myosin extraction.
Analytical techniques Protein concentrations were determined by the biuret technique [19], standardized by micro-Kjeldahl methods assuming 1 6 ~ protein nitrogen content, or by absorbance of bovine serum albumin at 280 nm. The hydrolysis of ATP was determined by assay of released inorganic phosphate using the method of Rockstein and Herron [20], with absorbances measured at 720 nm. All reactions were run as time courses and rates were calculated from initial linear portions of the curve. The standard conditions for measuring the Ca 2÷activated ATPase activity of myosin were 5 mM CaC12, 5 mM ATP, 18 mM morpholinopropane sulfonic acid buffer (pH 7.0 at 25 °C), KC1 to give I = 0.6, and 0.10.3 mg myosin/ml. These ionic conditions were chosen to eliminate any contribution to measured activity by the Mg 2+-activated actomyosin complex. Actomyosin ATPase activity (used as a check for the presence of actin) was measured at I = 0.1, 5 mM MgC12, 0.5 mM ATP, 0.I mM CaC12, 18 mM morpholinopropane sulfonic acid buffer (pH 7.0 at 25 °C) [7]. The composition of all protein samples was determined by electrophoresis of sodium dodecyl sulfate-treated proteins, using the method of Weber and Osborn [21]. The relative contents of major components were obtained by planimetry of the major peaks of Coomassie blue-stained gel densitometer tracings. These ratios, along with the measured total protein content, provided estimates of the amounts of individual components [3]. Dilute samples that were eluted from gel chromatographic columns were precipitated by 1 0 ~ (final concentration) cold trichloroacetic acid, washed in cold acetone, and dissolved with 8 M urea before treatment with sodium dodecyl sulfate. Molecular sieve chromatography was performed on 4 ~ agarose beads (Biogel A-15 m, Bio-Rad Laboratories), in 1.4 x 60 cm columns having an approximate total volume of 100 ml. Flow rates of 5.5-6.5 ml/h were maintained by a peristaltic pump. Absorbance was monitored at 280 nm using a double beam absorbance monitor (ISCO Model UA-5) with elution buffer in the reference cell. All peaks were analyzed for protein composition by sodium dodecyl sulfate electrophoresis of the trichloroacetic acid-precipitated protein. Molecular weights were estimated from the manufacturer's calibration charts. Solubility was defined as the fraction of protein remaining in the supernatant after centrifugation at 10 000 × g for 30 min. Viscosities were measured with an Ostwald-type capillary viscometer that had an outflow time of about 2.5 min for water at 23 °C. Protein samples were cleared of particulates by centrifugation at 10 000 x g before being tested. Protein concentrations and pH were checked after each viscosity measurement.
244
~ ~
ARTERIES A.
Extract: .05M KCI, imM ATP, 2mM EGTA, pH 7.4
cell
debris
SOLUBLE EXTRACT
B.
Dialyze:
42mM
Potassiumphosphate, pH 6
ACTOMYOSIN C.
Ultracentrifuge: 0.6M KC1, 40mMpyrophosphate
soluble
fractions
(precipitate)
pH 6 ~ ~ p 8
5mMATP, imM EGTA SUPERNATANT ACTOMYOSIN
D.
Dialyze: 0.2M KCI, pH 7
(precipitate) E.
(soJuble)
Fractionate: 2% PEG-6000 MYOSIN
(precipitate)
Fig. i. Procedure for isolating myosin from hog carotid arteries and definition of terms. Normal procedure utilizes the pH 8 pathway for differential ultracentrifugation (step C). PEG, polyethyleneglycol. RESULTS
Preparation of actomyosin The preparation of contractile proteins from the smooth muscle of hog common carotid arteries is outlined in Fig. 1. After grinding, the arterial mince was rinsed in distilled water and then suspended in 7-10 ml/g low ionic strength extraction solution (0.05 M KC1, 18 mM morpholinopropane sulfonic acid buffer (pH 7.4, 0 °C), 2 mM ethyleneglycol-bis-(fl-aminoethyl ether) N,N'-tetraacetic acid (EGTA), and 1 m M ATP), thoroughly homogenized with a Polytron blender (Brinkmann Instruments), and stirred for 2 h. Tissue debris was removed by centrifuging for 20 min at 5000 x g. The supernatant was filtered through cheesecloth to remove lipid aggregates and then stirred for 30 min with 10 mg/ml D o w e x 21K (Bio-Rad Laboratories) anion exchange resin (previously washed and adjusted to pH 7 with KOH) to remove ATP. Approx. 70 % of the tissue content of myosin, actin, and tropomyosin was
245 TABLE I PROTEIN FRACTIONS IN ARTERIAL MYOSIN PREPARATION: YIELD AND COMPOSITION Values are mean ± S.E. N, number of preparations; n, number of determinations of composition by densitometry of sodium dodecyl sulfate electrophoresis gels. Fraction
Total media homogenate*** Soluble extract Actomyosin Supernatant actomyosin * Myosin
Yield: total protein Composition: percent of total protein* N
mg/g artery
n
Myosin
Actin
Tropomyosin**
5 30 29
48.3 :L 7.4 32.8 :]- 1.5 16.7 4- 1.0
11 3 7
16.1 4- 1.0 18.5 4- 0.6 30.9 ± 3.5
54.0 4- 2.0 44.8 4- 1.2 48.2 4- 3.7
30.8 4- 1.5 32.2 4- 4.1 20.0 4- 1.6
8
39.6 4- 3,4 --
23.9 4- 3.4 --
35.2 -k 3.4 --
35 10
3.7 4- 0.2 0.97 4- 0.10
* Myosin + actin + tropomyosin = 100 700for each determination. ** This peak on sodium dodecyl sulfate was operationally defined as tropomyosin. Up to 50 700of this amount may be some other, unidentified, protein (ref. 3). *** Ref. 3. * Includes centrifugation of concentrated protein samples which tend to give lower yields. extracted by this method (Table I). All o f the protein with molecular weight approx. 37 000 on sodium dodecyl sulfate gels is here defined as tropomyosin. However, only half o f this fraction has been verified to be t r o p o m y o s i n [3], the remainder being unidentified. After decanting from the resin beads, the extract was dialyzed against 42 m M potassium phosphate, p H 6, for 24-48 h, with several changes of buffer, to precipitate actomyosin. The insoluble actomyosin was sedimented by centrifugation (30 min at 12 000 × g). The supernatant was saved for actin preparation. The precipitate contained half the extracted proteins, including all the myosin, leaving mostly actin and t r o p o m y o s i n in the supernatant.
Preparation of act& A crude actin preparation was obtained f r o m the supernatant remaining after the initial actomyosin sedimentation (Step B, Fig. 1) by adding polyethylene glycol 6000 to 2 ~ . The resultant precipitate was dissolved in 2 m M Tris. HCI, 0.2 m M ATP, 0.5 m M 2-mercaptoethanol and 0.2 m M CaCI2, p H 8 [18]. This actin was stored without repolymerization; gel electrophoresis showed little contamination with other proteins.
Differential centrifugation of actomyosin in relaxing medium The actomyosin was dissolved in 0.6 M KCI, 40 m M sodium p y r o p h o s p h a t e (pH 8.0), 5 m M A T P and 1 m M E G T A , and centrifuged 5 h at 50 000 rev./min in the Beckman 60 Ti r o t o r (176 000 × g average force). These conditions were initially chosen to dissociate actomyosin, to minimize myosin aggregation, and to sediment F-actin. Analysis o f the resultant supernatant by sodium dodecyl sulfate gel electrophoresis showed that increasing the p H that was maintained during ultracentrifugation f r o m 6 to 8 resulted in less protein remaining in that supernatant. That is, the
246 sedimentation of actomyosin increased with pH. The fraction of myosin that remained in the supernatant showed a negative linear correlation with pH: at pH 6, the supernatant contained most of the myosin of the initial actomyosin, while at pH 8, between 60 and 80 ~ was sedimented. Actin followed a similar trend with limits of 70 ~ at pH 6 and 85 ~o at pH 8, while a constantly low fraction of tropomyosin sedimented regardless of pH. In addition, the pellets obtained after 5 h of centrifugation were noticeably different. At pH 6, the pellet had an opaque cream-colored appearance, while at pH 8, the pellet had a small opaque center overlayed with a large amount of translucent material. Examination by sodium dodecyl sulfate electrophoretic gels showed the opaque centers to consist of nearly pure actin while the translucent overlays contained both actin and myosin. The tendency for actin and myosin to sediment together to form the lighter overlay portion of the pellets as the pH was increased suggested an incomplete dissociation even in the presence of 0.6 M KC1, 40 mM pyrophosphate and 5 mM ATP. Several observations support this interpretation. (1) Using 0.6 M KI or KSCN instead of KC1 increased the total supernatant protein by 79 and 34 ~o, respectively, although the relative amounts of actin and myosin remained unchanged. (2) Increasing the temperature caused a marked decrease in supernatant protein. At 25 °C the supernatant contained virtually no actin or myosin, while the amount of tropomyosin was unaffected. (3) The reduced viscosity of the actomyosin, measured before ultracentrifugation, was consistently higher at pH 8 than at lower values (400 ml/g at pH 8 vs 140 ml/g at pH 6 for a typical experiment). At pH 8, the protein was highly thixotropic, forming a viscous gel on standing at 4 °C. This gelation did not occur at lower pH. The effect of total protein concentration on the sedimentation of arterial actomyosin, at constant pH, is shown in Fig. 2. The increase in sedimentation (decrease in supernatant protein) with increasing concentration is consistent with the
2
E
5
I0
Initial c o n c e n t r a t i o n
20
o f a c t o r n y o s i n ( m g / ml)
Fig. 2. Fraction of total protein remaining in supernatant, after differential ultracentrifugation of actomyosin in 0.6 M KCI, 40 mM sodium pyrophosphate, 5 mM ATP, 1 mM EDTA, pH 8, 0 °C, as a function of the initial protein concentration. Line was fitted by method of least squares and represents significant correlation.
247 100 ¢c-
B
80 X
E
0
I'--
0 0
0--
--
--
-0
U
C 0
cL. 2C
~--~
c o Initiol concentrotion of myosin or octin ( m g / m l )
Fig. 3. Fraction of myosin or actin remaining in the supernatant following differential ultracentrifugation of actomyosin in conditions of Fig. 2, as a function of their initial concentration in actomyosin. Arterial actomyosins (open symbols) and skeletal actomyosins (closed symbols) of different composition were prepared as described in the text. (3 and O, myosin; A, actin. Crosses (x) represent polyethylene glycol-purified arterial myosin without the addition of actin. Solid lines represent significant correlations obtained by method of least squares. presence of protein-protein interaction. To attempt a more detailed description of these interactions, the composition of actomyosin preparations and of their corresponding supernatants after ultracentrifugation, as determined from gel electrophoresis, was used to infer the sedimentation behavior of each component protein. These data were obtained from several samples of "natural" actomyosin, prepared as described above, and of "synthetic" actomyosins that were prepared to test the effect of varying the ratio of actin to myosin (see below). The sedimentation of the myosin component in actomyosin increased significantly (supernatant content decreased) with increasing initial concentration (Fig. 3, open circles). Less than 30 7o remained in the supernatant at initial concentrations of 2.5 mg/ml or greater. Little tropomyosin sedimented regardless of initial concentration. Little of the arterial actin remained in the supernatant, although slightly more remained when higher initial concentrations of actin in actomyosin were used (Fig. 3, triangles). Similar experiments were carried out with synthetic actomyosin that was prepared from skeletal myosin and actin. In this system, the sedimentation of the myosin component was much lower and independent of concentration (Fig. 3, closed circles), while the supernatant content of actin decreased slightly with its initial concentration (not shown). Similarly, when pure arterial myosin was used, less than 30 7o sedimented, although initial concentrations of up to 6.5 mg/ml were used (Fig. 3, crosses). These results were all obtained at p H 8. As described above, decreasing p H resulted in more arterial myosin remaining in the supernatant. In two experiments with pure skeletal myosin, sedimentation was not noticeably different at p H 6.
248 The ratio of actin to myosin (actin:myosin) in arterial smooth muscle is nearly 10-fold greater than in skeletal muscle [3, 13]. When actomyosin is precipitated by dialysis of the muscle extract to low ionic strength, this ratio is approximately maintained. We thus considered the possibility that arterial actin and myosin were intrinsically similar to their skeletal muscle counterparts, but that the unusual behavior of arterial actomyosin upon centrifugation was due to the increased actin:myosin. Actomyosins of different actin:myosin were therefore centrifuged under the same conditions used in purifying the arterial proteins. Arterial actomyosins of lower actin:myosin were produced by adding purified arterial myosin to actomyosin, or by using a preparation obtained from the translucent upper portion of pH 8 ultracentrifugation pellets (see above). Skeletal actomyosins of high actin:myosin were also prepared, by combining pure myosin and actin in desired amounts. Tropomyosin was absent in these latter preparations. The interaction of actin and myosin after mixing was verified by measuring a decrease in the viscosity of the mixture upon the addition of ATP, before centrifugation. The results of centrifugation, already described above, were analyzed with regard to the ratio of actin to myosin. There was no correlation between the sedimentation of actin or myosin and their relative content in actomyosin. MgZ+-ATP is the dissociating agent usually employed with skeletal actomyosin, and its use might have enhanced the separation of skeletal actin and myosin. However, the inclusion of Mg 2;-, equimolar to ATP, in three arterial actomyosin experiments produced no significant differences.
Dialysis of the supernatant actomyosin The supernatant that was obtained after differential ultracentrifugation was stirred for 30 min with 40 mg/ml Dowex 21K (equilibrated to pH 7) to remove ATP and pyrophosphate, and then dialyzed against 0.2 M KCI, pH 7. When centrifugation was carried out at pH 8, the protein was highly soluble, even when the concen-
c" 100 f
8°I
60
c_
4o
~d c 20 _o
6
7
8
9
pH Fig. 4. PIecipitation of supernatant actomyosin dr, ring dialysis in 0.2 M KCI (pH 7) as a function of the pH maintained during the preceding ultracentrifugation. Total protein was measured for the soluble fractions before and after dialysis, as well as for the precipitates dissolved in 3 M KCI. Points indicate means - S.E., four preparations. Line is least squares fit of all unaveraged data.
249 tration of KC1 was reduced to as low as 0.05 M. This fraction was used to purify myosin by polyethylene glycol fractionation (see below). However, if the pH during centrifugation was less than 8, a significant amount of supernatant actomyosin precipitated after dialysis to 0.2 M KC1 at pH 7 (Fig. 4). Since all samples were dialyzed against the same neutral solution the effect of p H was clearly indirect. That is, during dialysis, the protein "remembered" the pH at which it had been centrifuged. Skeletal muscle myosin, when treated as above, always pricipitated completely in 0.2 M KC1, regardless of the pH of prior centrifugation. Combining the data of Fig. 4 with the observed increase in supernatant concentration of actin as a function of pH during ultracentrifugation reveals a direct dependence of actomyosin precipitation in 0.2 M KCI (pH 7) on the supernatant actin concentration (Fig. 5). No such correlation was found with the myosin concen-
c-
100
~
8o
~
6O
t/) r = 0.806
4O cO
U_
2O
o
I 0
.2
.4
.6
.8
1.0
1.2
1.4
Actin Concentration ( m g / m l )
Fig. 5. Precipitation of supernatant actomyosin during dialysis in 0.2 M KC1 (pH 7) as a function of actin concentration in the supernatant, r, correlation coefficientfound by the method of least squares. tration. The precipitation of this supernatant actomyosin, after centrifugation at low pH, could b e prevented by dialysing against KI instead of KC1 (both at pH 7). Although the concentration of actin in the supernatant was significant after centrifugation at pH 8 (Table I), the normal tests for the presence of actin gave negative results. ATPase activity in the presence of Mg 2+, at I = 0.1, was extremely low ( < 0.001 #mole P J m g per min, 10 preparations), ATP sensitivity (changes in the viscosity number upon the addition of ATP at I ---- 0.6 [22]) was less than 5 7O, and the clearing response on the addition of ATP, at I = 0.1, was negligible. The apparent lack of myosin binding by the supernatant actin suggested it was depolymerized or denatured. To determine the states of actin aggregation, the ultracentrifuge supernatant was chromatographed on 4 7o agarose (Fig. 6). The presence of a void peak (6-7 h) was Variable and reflected differences in the protein preparations. When a void fraction was present, it contained mostly actin. The intermediate peak (8-t0 h) contained mostly myosin, but with some a'ctin, while the third peak (after 13 h) contained similar amounts of actin and tropomyosin. The high absorbance of this final peak was due to the ATP in the applied sample. Under similar ionic conditions, pure
250 Q2
1.0
Eo.~
o.5,~ A
e~,' ;
M
lo
15
Elution time,h
Fig. 6. Chromatography of supernatant actomyosin on 4 % agarose in 0.6 M KC1, 40 mM sodium pyrophosphate, 1 mM EGTA, pH 8, 0 °C. Sample (2 ml) was applied to the column immediately following ultracentrifugation of actomyosin. The large absorbance of the final peak is due to the ATP that had been contained in the sample, in addition to low molecular weight proteins. Arrow indicates void volume. Letters indicate major composition of peaks as determined by gel electrophoresis: A, actin; M, myosin; T, tropomyosin. skeletal myosin was eluted between 8 and 10 h, which corresponds to a molecular weight of 4 000 000. The presence of actin in the fractions eluting after 13 h, corresponding to a molecular weight of less than 100 000, verifies the significant degree of actin depolymerization at high ionic strength.
Purification of myosin by polyethylene glycolfraetionation After dialysis to 0.2 M KC1 (pH 7), the supernatant was adjusted to 2 mg protein/ml. Myosin was then selectively precipitated from the pH 8 supernatant by adding polyethylene glycol-6000 (Baker Chemicals, 20 % (w/v) solution in 0.2 M KC1) to give a final concentration of 2%. The myosin was collected by centrifugation (10 000 x g), washed twice with 50 mM KC1 and dissolved in 0.6 M KC1. When the supernatant actomyosin was less soluble following ultracentrifugation at lower pH (Fig. 4), purification of the myosin by polyethylene glycol fractionation was no longer effective. Even when this less soluble actomyosin was dissolved in KC1 concentrations of 0.3 M KC1 or higher, the addition of polyethylene glycol resulted in the precipitation of both actin and myosin. It appeared that fractionation by polyethylene glycol was possible only when myosin was soluble at low ionic strength as occurred after centrifugation at pH 8. The addition of polyethylene glycol in solid form (to 2 %) precipitated less myosin than when polyethylene glycol was added as a 20% solution (34.0 i 3.3% vs 85.4 ± 12.6 % of the total myosin, means ± S.E., 6 preparations). Furthermore, the myosin thus obtained may have been denatured as evidenced by a lower Ca 2+ATPase activity. The purity of the myosin that was precipitated by polyethylene glycol-6000 was demonstrated by the presence of a single major protein peak that co-electrophoresed with skeletal and cardiac myosins on sodium dodecyl sulfate gels. One or two faint bands corresponding to molecular weights less than 15 000 were sometimes present, but not consistently enough to identify them as light chains of myosin. The absence of these low molecular weight components was not associated with any decrease in Ca2+-ATPase activity. The average yield of purified myosin was 0.97 4-
251 0.1 mg/g artery (mean ± S.E., 10 preparations). Its Ca 2+-dependent ATPase activity, at I = 0.6, averaged 0.187 :k 0.006 #mol Pi/mg per min (8 preparations), and decreased markedly with decreasing ionic strength. The similar dependence of smooth muscle actomyosin ATPase activity on ionic strength is well documented[5, 6, 8]. A 2-fold increase in activity in the presence of denaturing agents, such as 2 M urea [6, 23] was also found. In the presence of Mg 2+, at I = 0.1, myosin ATPase activity rose from undetected levels to 0.018/~mol PJmg per min upon the addition of either skeletal or arterial actins (approx. 2 mg actin/mg myosin). This value is comparable to those previously reported for native arterial actomyosin [7, 8].
Effect of polyethylene glycol on myosin After being precipitated by polyethylene glycol myosin was insoluble at I _< 0.25, even after being dissolved in high ionic strength. Its solubility characteristics thus became similar to those of skeletal muscle myosin. This loss of solubility was not irreversible, as re-centrifugation of the polyethylene glycol-precipitated myosin at pH 8 (as with actomyosin, step C, Fig. 1) rendered much of it soluble again when dialyzed to low ionic strength. The change that occurs in the solubility of myosin is probably not the result of denaturation, as all of the myosin Ca z+-ATPase activity was retained in the protein precipitated by polyethylene glycol. Furthermore, estimates of the myosin content, from gel electrophoresis, in actomyosin samples that had not been treated with polyethylene glycol were found to agree well with estimates based on the measured CaZ+-ATPase activity of these impure samples, assuming the polyethylene glycolpurified myosin to have 100 ~o activity (Fig. 7). This correlation supports the assumption that, for actomyosin samples, the CaZ+-ATPase activity measured at I = 0.6 is due to myosin alone, and also justifies the use of this assay as a measure of myosin content. It does not, however, rule out the presence of some activating effect of polyethylene glycol on myosin Ca2+-ATPase activity (which would compensate for an
100 i/1 u
0 "~
~
oE V]
0
.-I/1 O
113 t-
~
~g
60
4o
•
r = 0.905
20 eo
o 0
I
I
I
I
i
20
40
60
80
100
% Myosin (Ca-ATPase activity) Fig. 7. Correlation of myosin fractions in actomyosins as determined from Ca2+-ATPase activity measurements and by densitometry of sodium dodecyl sulfate gels. Polyethylene glycol-purified myosin was assumed to be 100 % pure and fully active for comparisons with actomyosin activity and composition, r, correlation coefficient determined by method of least squares.
252 equivalent loss of activity due to denaturation), although the absence of such an effect on skeletal muscle myosin (Megerman, J., unpublished) renders this coincidence unlikely. DISCUSSION This procedure for purifying smooth muscle myosin was developed because the methods that are commonly applied to the preparation of skeletal muscle myosin are rendered ineffective by the unusual solubility of myosin and the high relative content of actin in actomyosin extracts. It is clear that the method permits the effective purification of smooth muscle myosin. However, it suffers from several drawbacks: (i) the selective precipitation of myosin by polyethylene glycol is effective only when the supernatant actomyosin is soluble at low ionic strength (0.2 M KC1, pH 7); (ii) this highly soluble actomyosin occurs only under conditions which result in a sizable fraction of the myosin sedimenting during ultracentrifugation (pH 8); and (iii) the precipitation of myosin by polyethylene glycol may cause significant changes in the solubility of the myosin.
Myosin purification depends on actin depolymerization The effect of polyethylene glycol on solubility depends on the molecular size of a protein; larger proteins are precipitated by lower concentrations of polyethylene glycol [24]. Furthermore, polyethylene glycol is known to enhance the interaction between actin and myosin [25]. It follows, therefore, that actin in the soluble ultracentrifuge supernatant (after pH 8) had a smaller molecular size than myosin, and that this actin did not bind myosin. The absence of myosin binding is supported by the negative results of Mg 2+-ATPase and ATP sensitivity tests. The retarded position of actin upon gel filtration of the soluble supernatant actomyosin (Fig. 6) suggests this actin was depolymerized, or nearly so. But this depolymerized actin was probably not denatured, since the precipitation of actomyosin as a function of actin concentration in the supernatant (Fig. 5) demonstrates the ability of the actin to bind myosin and form an insoluble complex, given that it was present in sufficient amounts. This precipitation of actomyosin did not correlate with the concentration of myosin in the supernatant, even though the concentration of this protein varied along with that of actin as a function of pH during ultracentrifugation. We suggest that Fig. 5 reflects a dependence of the polymerization (and perhaps nucleation) of actin on its concentration [26], and that the process is required to form insoluble actomyosin. The fact that no precipitate was formed when KI was substituted for KC1 is consistent with this hypothesis, as KI is known to prevent actin polymerization. Conversely, the successful application of polyethylene glycol fractionation to the purification of myosin appears to require that any actin that is present be in a depolymerized form, as whenever the supernatant actomyosin became insoluble upon dialysis to 0.2 M KC1 (pH 7), purification by this method was not possible. Sedimentation behavior of actomyosin Even though the arterial actomyosin was centrifuged under conditions which,
253 with skeletal actomyosin, would ensure complete dissociation, the negative correlation between the amount of supernatant protein and the initial concentration of actomyosin (Fig. 2) suggests an interacting protein system. The questions to be answered are (1) which of the possible interactions between actin, myosin, and tropomyosin are dominating, and (2) does the arterial actomyosin system behave in a manner significantly different than could be expected with skeletal muscle proteins. An interaction involving tropomyosin is considered least likely, as this protein is readily dispersed in high ionic strength [27]. This is borne out by the low sedimentation of tropomyosin and the absence of any dependence on pH or concentration. However, the sedimentation dependence of the myosin component of arterial actomyosin on its concentration (Fig. 3) is evidence of myosin interaction, but it does not distinguish between myosin self-aggregation or an interaction with actin. The latter is supported, though by no means proven, by the decreased sedimentation of both actin and myosin at lower pH and temperature, and in the presence of KI and KSCN. Moreover, the low sedimentation ( < 30 ~) of pure arterial myosin, even at an initial concentration (6.5 mg/ml) which would result in its complete sedimentation if it were part of actomyosin (Fig. 3, solid line), virtually eliminates myosin self-aggregation from consideration. This low value is similar to the amount sedimented when pure skeletal myosin was centrifuged under identical conditions, and is consistent with the low sedimentation constant (6 S) obtained for a dispersed monomeric myosin. The sedimentation of skeletal myosin was independent of pH between 6 and 8, whereas that of arterial myosin (in actomyosin) was quite sensitive to pH. It is clear that the low sedimentation of pure arterial myosin reflects a predominantly dispersed, non-aggregated state. When actin is present, it is the sedimentation of some actomyosin complex that must be considered. The apparent molecular weight of myosin under conditions similar to those obtaining during ultracentrifugation, as determined by gel filtration, was 4 000 000. This large value reflects molecular size (Stokes radius) rather than weight [28], and is consistent with the finding from light scattering measurements that myosin aggregates containing 1 to 8 monomers (molecular weight approx. 500 000 each) have similar Stokes radii [29]. The similar elution patterns for skeletal and smooth muscle myosins provide no information, therefore, as to the similarity of aggregation states for these two proteins. Most of the experiments with arterial actomyosin were carried out in the absence of Mg 2÷, as this ion is thought to promote the aggregation of smooth muscle myosin [30]. The inclusion of MgC12 produced no significant differences in the content or composition of the supernatants, indicating that the incomplete dissociation of actin from myosin, in the presence of ATP, was not due to the absence of Mg 2÷. Furthermore, synthetic skeletal actomyosin that was prepared to reflect the composition of arterial actomyosin did not behave as did the latter. Even without Mg 2+, less myosin from skeletal actomyosin sedimented than from arterial actomyosin. The behavior of the smooth muscle proteins is therefore not intrinsically similar to that of the skeletal muscle proteins and is not merely the result of the high ratio of actin to myosin that exists in arterial actomyosin. From these comparisons, we conclude that the observed interaction between arterial actin and myosin reflects some inherent differences in the binding properties of these smooth muscle proteins.
254
Intermediate actomyosin complexes It is not yet possible to describe in detail the actomyosin complex that resists dissociation. The distinct two-part appearance of the pellets following ultracentrifugation at pH 8 (where much myosin is sedimented) compared with pH 6 (where little myosin sediments) suggests this complex has a much smaller sedimentation coefficient than does F-actin. The actomyosin mixture being centrifuged at pH 8 thus contains at least three different species of actin: F-actin which sediments rapidly and is independent of myosin, intermediate actin units which co-sediment with myosin, and depolymerized actin which does not bind myosin. The depolymerization of actin could result from the high pressures generated during ultracentrifugation [26], or may reflect a natural tendency for smooth muscle actin. It is not clear if, at pH 6, the intermediate actin units are fewer in number, if the reduced pH prevents their association with myosin, or if possible actomyosin complexes are simply too small to sediment.
Solubility of myosin and of actomyosin The dialysis experiments in 0.2 M KC1 verify that myosin from arterial smooth muscle tends to be more soluble than from skeletal muscle, as was assumed from the extractability of the former at low ionic strength. It is still to be determined, however, whether this solubility is an intrinsic property of the myosin molecule itself, or whether it is due to an association with some solubilizing factor [2, 17]. For instance, myosin may be bound to small units of non-polymerized actin, which are soluble until the polymerization of this actin occurs. This hypothesis is supported by the striking change in myosin solubility that accompanies its purification by polyethylene glycol, but is opposed by the partial resumption of high solubility following the centrifugation of pure myosin. A decrease in actomyosin solubility in vitro, is associated with an increase in its Mg2+-ATPase activity at low ionic strength [8]. But the greatest values of activity reported so far are low compared to that for skeletal actomyosin, and they require that Mg 2+ be present in concentrations 10-fold greater than with skeletal actomyosin [7]. It has been postulated that this Mg z+ acts to chelate ATP and thus reduce its dissociating effect on actomyosin [7], or to reduce the solubility of myosin [31]. But Mg 2÷ has a polymerizing effect on actin as well [26]. We therefore suggest a possible function for Mg 2+ in promoting the repolymerization of actin, which produces actomyosin complexes that are large, insoluble, and have higher levels of MgZ+-activated ATPase activity. ACKNOWLEDGEMENTS The technical assistance of Mrs Mildred Smythers and the generous donation of arteries by George H. Meyers and Sons, Richmond, Virginia, are gratefully acknowledged. J. Megerman was a predoctoral trainee (N.I.H. Grant HL 05815) during this investigation and many of the results were included in his doctoral dissertation at the University of Virginia. R. A. Murphy is the recipient of a career development award (HL 11747). This research was supported by N.I.H. grant HL14547.
255 REFERENCES 1 Somlyo, A. P. and Somlyo, A. V. (1968) Pharmacol. Rev. 20, 197-272 2 Riiegg, J. C. (1971) Physiol. Rev. 51,201-248 3 Murphy, R. A., Herlihy, J. T. and Megerman, J. (1974) J. Gen. Physiol. 64, 691-705 4 Laszt, L. and Hamoir, G. (1961) Biochim. Biophys. Acta 50, 430-449 5 Needham, D. M. and Williams, J. M. (1963) Biochem. J. 89, 552-561 6 BCtr~iny,M., BCtr~ny, K.. Gaetjens, E. and Bailin, G. (1966) Arch. Biochem. Biophys. 113,205-222 7 Murphy, R. A., Bohr, D. F. and Newman, D. C. (1969) Am. J. Physiol. 217, 666-673 8 Murphy, R. A. (1971) Am. J. Physiol. 220, 1494-1500 9 Somlyo, A. P., Somlyo, A. V., Devine, C. E. and Rice, R. V. (1971) Nat. New Biol. 231,243-246 10 Small, J. V. and Squire, J. M. (1972) J. Mol. Biol. 67, 117-149 11 Cooke, P. J. and Fay, F. S. (1972) Exp. Cell Res. 71,265-272 12 Bois, R. M. (1973) Anat. Rec. 117, 61-78 13 Tregear, R. T. and Squire, J. M. (1973) J. Mol. Biol. 77, 279-290 14 Pollard, T. D., Thomas, S. M. and Niederman, R. (1974) Anal. Biochem. 60, 258-266 15 Nachmias, V. T. (1974) J. Cell Biol. 62, 54-65 16 Weber, A. (1956) Biochim. Biophys. Acta 19, 345-351 17 Megerman, J. and Murphy, R. A. (1973) Biophys. J. 13, 321a 18 Spudich, J. A. and Watt, S. (1971) J. Biol. Chem. 246, 4866--4871 19 GornaU, A. G., Bardawill, C. J. and David, M. M. (1949) J. Biol. Chem. 177, 751 20 Rockstein, M. and Herron, P. W. (1951) Anal. Chem. 23, 1500-1501 21 Weber, K. and Osborn, M. (1969) J. Biol. Chem. 244, 4406--4412 22 Weber, H. H. and Portzehl, H. (1952) Adv. Protein Chem. 7, 161-252 23 Gaspar-Godfroid, A. (1970) Angiologica (Basel) 7, 273-290 24 Juckes, I. R. M. (1971) Biochim. Biophys. Acta 229, 535-546 25 Kaldor, G., Hsu, Q. S;, Wachsberger, P. and Chowrashi, P. (1971) Am. J. Physiol. 220, 639-645 26 Oosawa, F. and Kasai, M. (1971) in Subunits in Biological Systems (Timasheff, S. N. and Fasman, G. D., eds), Part A, pp. 261-322 27 Drabikowsky, W. and Nowak, E. (1968) Eur. J. Biochem. 5, 376-384 28 Ackers, G. K. (1970) Adv. Protein Chem. 24, 343-346 29 Lowey, S. and Holtzer, A. 0959) J. Am. Chem. Soc. 81, 1378-1383 30 Shoenberg, C. F. (1969) Tissue Cell 1, 83-96 31 Russell, W. E. (1973) Eur. J. Biochem. 33, 459-466
