Purification and Characterization of a Myosin I Heavy Chain Kinase from Acanthamoeba castellanii *

In previous work from this laboratory, a partially purified protein kinase from the soil amoeba Acanthamoeba castellanii was shown to phosphorylate the heavy chain of the two single-headed Acanthamoeba myosin isoenzymes, myosin IA and IB, resulting in a 10to 20-fold increase in their actin-activated Mg2+ATPase activities (Maruta, H., and Korn, E. D. (1977) J. Biol. Chem. 252, 8329-8332). A myosin I heavy chain kinase has now been purified to near homogeneity from Acanthamoeba by chromatography on DE52 cellulose, phosphocellulose, and Procion red dye, followed by chromatography on histone-Sepharose. Myosin I heavy chain kinase contains a single polypeptide of 107,000 Da by electrophoretic analysis. Molecular sieve chromatography yields a Stokes radius of 4.1 nm, consistent with a molecular weight of 107,000 for a native protein with a frictional ratio f approximately 1.3:l. The kinase catalyzes the incorporation of 0.9 to 1.0 mol of phosphate into the heavy chain of both myosins IA and IB. Phosphoserine has been shown to be the phosphorylated amino acid in myosin IB. The kinase has highest specific activity toward myosin IA and IB, about 3-4 @mol of phosphate incorporated/min/ mg (30 “C) at concentrations of myosin I that are well below saturating levels. The kinase also phosphorylates histone 2A, isolated smooth muscle light chains, and, to a very small extent, casein, but has no activity toward phosvitin or myosin 11, a third Acanthamoeba myosin isoenzyme with a very different structure from myosin IA and IB. Myosin I heavy chain kinase requires Mg2+ but is not dependent on Ca2+, Ca2+/calmodulin, or CAMP for activity. The kinase undergoes an apparent autophosphorylation.

light chain of 17,000 Da (3, 4). Myosin IB is also a singleheaded enzyme with a native molecular weight of about 150,000, but contains one 125,000-Da heavy chain and a light chain of 25,000 Da (4). Both myosin I isoenzymes are obtained with variable amounts (always less than 0.5 mol/mol by Coomassie blue stain) of a peptide of 14,000 Da (4). Myosin IA and IB are both globular molecules (3,4) with no detectable tail region' and no known ability to self-associate.' Peptide mapping (7) and immunochemical analysis (8, 9) support the conclusions that myosins IA, IB, and I1 are separate gene products and that the myosin I isoenzymes as isolated are identical with the native molecules in the cell.
For all three Acantharnoeba myosins, the magnitude of their actin-activated Mg2+-ATPase activity is governed by the level of heavy chain phosphorylation (4, 10-12). Myosin I1 possesses three heavy chain phosphorylation sites, all of which have been localized to a 9000-Da chymotryptic peptide isolated from near the COOH terminus of the molecule (13, 14). The actin-activated Mg"-ATPase activity of myosin I1 is inversely correlated with the phosphorylation state of the enzyme, i.e. the fully dephosphorylated enzyme has the highest actomyosin I1 Mg'+-ATPase activity (10, 13). In parallel with the effect of phosphorylation on ATPase activity, dephosphorylated myosin I1 associates more readily into filaments than the phosphorylated molecule (6). A myosin I1 heavy chain kinase which phosphorylates all three sites has been partially purified (13) and a protein phosphatase active toward myosin I1 has been highly purified (15).
For myosins IA and IB, heavy chain phosphorylation regulates the actin-activated Mg"-ATPase activity in a manner opposite to that observed for myosin 11, i e . fully phosphorylated myosin I has the highest actomyosin I MgZ+-ATPase activity (4, 12, 16). Maruta and Korn (16) showed that a partially purified cofactor, found by Pollard and Korn (17) to be required for actin activation of myosin I Mg2'-ATPase, is a specific myosin I heavy chain kinase and that the magnitude of actomyosin I Mg"-ATPase activity is directly proportional to the extent of myosin I heavy chain phosphorylation (12). The myosin I kinase obtained by Maruta and Korn was quite impure, containing at least 12 prominent bands on electrophoresis gels (16) and being contaminated with proteases (12). The purpose of the present work was to purify myosin I heavy chain kinase to homogeneity in order to allow its characterization and also to facilitate physical and kinetic studies of the mechanism by which phosphorylation of the heavy chain of myosin I affects its enzymatic activity. In this paper we describe the purification to near homogeneity of a myosin I heavy chain kinase and the initial characterization of this ~-~ " enzyme. In the accompanying paper (18), the effects of phosphorylation of myosin IA and IB on their interaction with Factin are reported.

EXPERIMENTAL PROCEDURES
Materials-DE-52 cellulose and P11 phosphocellulose were obtained from Whatman. Bio-Gel A-0.5m (200-400 mesh), Bio-Gel P-100 (50-100 mesh), and the dye reagent used for the Bradford assay were purchased from Bio-Rad. Procion red dye chromatography heads, ohtained from Amicon Corp., were washed with 8 M urea (2 column volumes) and equilibration buffer (8 column volumes) before use. Histone-Sepharose was prepared using histone 2A (Sigma H-9250) and CNBr-activated Sepharose 4B (Pharmacia) by the procedure described in the Affinity Chromatography Manual from Pharmacia. Coupling was performed a t 4 "C for 24 h using 7.5 mg of histone 2A per ml of swollen packed gel (>98% of the protein was coupled). The remaining active groups were blocked by a 2-h incubation a t 21 "C with 1 M ethanolamine (pH 8.0). The following proteins were obtained as gifts: Acantharnoeba myosin 11, purified by the method of Collins and Korn (11) and partially purified Acanthamoeba myosin I1 heavy chain kinase (13)  Purification of Myosin I Isoenzymes-Myosin IA and IB were purified as described in the accompanying paper (18). Myosin IB was routinely >90% pure; the purity of myosin IA used in these experiments varied between 60-90% and was used less frequently. In all preparations, the purified myosin isoenzymes were shown to he devoid of kinase activity (see helow). A small amount of myosin I heavy chain kinase still associated with myosin IA and IB after DE-52 chromatography was removed from the myosins by chromatography on ADP-agarose as described previously (4). Myosins I A and IB were stored in 20 mM imidazole (pH 7.5), 100 mM KCI, 25-40% glycerol, 1 mM dithiothreitol, and 0.02% sodium azide a t 4 "C. Myosins IA and IB were stable in terms of enzymatic activity and SDS-PAGE profiles for between 1 and 2 weeks; routinely, the myosins were used within less than 1 week. Myosin I isoenzymes as isolated contain negligible phosphate.:' Purification of Myosin I Heavy Chain Kinase, DE-52 Chromatography-Approximately 1 kg of Acanthamoeba castellanii was grown, harvested, and washed as described by Pollard and Korn (3). The cell pellet was disrupted (10 strokes in a tight-fitting glass Dounce homogenizer) in 2 volumes of 30 mM imidazole (pH 7.0), 75 mM KC1, 12 mM sodium pyrophosphate, 5 mM dithiothreitol, 0.1% leupeptin, 1% pepstatin, and 0.6 mM PMSF and the homogenate centrifuged at 100,000 X g for 3 h (Beckman type 30 rotor). All procedures were performed at 4 "C. The supernatant (about 2 liters) was titrated to pH 8.0 with 1 M Tris, dialyzed for 12 h against 28 liters of buffer containing 10 mM Tris (pH 8.0), 7.5 mM sodium pyrophosphate, 1 mM dithiothreitol, and 0.6 mM PMSF, and centrifuged at 40,000 X g for ' 0 min. The supernatant was loaded onto a DE-52 column (5 X 80 cm) equilibrated with 10 mM Tris (pH 8.0), 10 mM KCI, and 1 mM dithiothreitol. The material collected during loading plus 1.5 liters of column wash were used for purification of myosin I heavy chain kinase while the myosin I isoenzymes were eluted from the column as described in the accompanying paper (18).
Phosphocellulose Chromatography-Solid ammonium sulfate (to 2 M ) was added to the material which did not adsorb to DE-52, and the precipitate was collected and resuspended in 150 ml of 20 mM TES (pH 7.0), 50 mM KCI, 5% glycerol, 1 mM dithiothreitol, 0.1% leupeptin, 1% pepstatin, and 0.6 mM PMSF, and dialyzed overnight against 2 liters of the same buffer. This material was applied to a phosphocellulose P-11 column (5 X 20 cm) equilibrated with 20 mM TES (pH 7.0), 25 mM KCI, and 1 mM dithiothreitol. The column was washed with the equilibration buffer, eluted with a linear KC1 gradient (1.2 liters, 25 to 600 mM KCI), and fractions were assayed for protein and myosin I kinase activity (see below and Fig. 1).
Assay ofActin-activated Mg'+-ATPase of Myosin I-Myosin I heavy chain kinase was assayed indirectly during its purification by its ability to increase the actin-activated Mg"+-ATPase of myosin I. The ATPase assay mixture contained 15 mM imidazole (pH 7.5), 2 mM MgCl,, 1 mM EGTA, and 2 mM [y-:"P]ATP (0.5 pCi/pmol) and the following additions: 5-10 pg of myosin IA or IB, 50 pg of skeletal muscle F-actin, and 1-10 g1 of the column fraction to he assayed in a '"P, from the [y-"lP]ATP as described by Pollard  a reaction mixture pre-equilibrated to 30 "C. Aliquots of 40 p1 were discs (Whatman, grade 3 " ) which were then dipped into 10% removed at intervals of 30 or 45 s, spotted onto 2.3-cm filter paper trichloroacetic acid containing 5% sodium pyrophosphate to terminate the reaction, and immediately washed on a suction manifold with 15 ml of 10% trichloroacetic acid. At the end of the incubation, all the filter paper discs were washed four times for 20 min with gentle agitation in 200 ml of 10% trichloroacetic acid plus 5% sodium pyrophosphate, washed once for 5 min with absolute ethanol, once for 5 min with absolute ether, air dried, and counted in 15 ml of Aquasol (New England Nuclear) in a Beckman model 250 scintillation counter.
The amount of myosin I heavy chain kinase added was such that less than 15% of the total substrate was phosphorylated. Under these conditions, the incorporation of "P into substrate was linear with time and proportional to the amount of kinase added. Kinase activity (micromoles of phosphate incorporated/min) was calculated by linear regression analysis of the phosphorylation time course. Control reactions containing only substrate or only kinase showed negligible phosphorylation. Autophosphorylation of myosin I kinase did not contribute significantly to the measured values. Finally, autoradiography of SDS-polyacrylamide gels of myosin I phosphorylated under these conditions showed that all of the 32P was incorporated into the intact myosin I heavy chain.
Stoichiometry of Phosphate Incorporation into Myosin I-The maximal extent of phosphorylation of myosins IA and IB by myosin I heavy chain kinase was determined by the filter paper assay essentially as described above but with higher concentrations of kinase. In calculating the phosphate content of myosin I, corrections were made for '"P incorporated into proteins other than myosin I. This correction was based on densitometric scans of autoradiograms of SDS-polyacrylamide gels of maximally phosphorylated myosin I, where the radioactivity in bands other than the intact myosin I heavy chain was less than 15% of the total protein-bound "P. The amount of myosin I per assay was estimated from the protein concentration corrected for the percentage of Coomassie blue stain in SDS-polyacrylamide gels not present in the intact myosin I heavy chain (usually not more than 10%).
Estimation of the Stokes Radius of Myosin, I Heavy Chain Kinase-Purified myosin I kinase (40 pg) and three standards of known Stokes radii, ovalbumin (Rs = 2.85 nm). bovine serum albumin (Rs = 3.5 nm), and aldolase (Rs Phosphoamino Acid A nalysis-Following maximal phosphorylation of myosin IB (100 pg) by myosin I heavy chain kinase in the presence of [y-,32PJATP, myosin was separated from ATP on a P-100 gel filtratlon column (1 X 40 cm) equilibrated with 0.1 M ammonium bicarbonate (pH 7.5). The radioactive myosin peak in the void volume was pooled, lyophilized, and subjected to partial acid hydrolysis in 6 N HCI a t 105 "C for 30, 60, 90, and 180 min. The hydrolysates were repeatedly lyophilized to remove HCI, mixed with authentic phosphoserine and phosphothreonine standards, and analyzed by electrophoresis at pH 1.9 on cellulose thin layer sheets exactly as described by Cote et al. (13). Miscellaneous Methods-Protein concentrations were determined using the colorimetric assay of Bradford (21), with bovine serum albumin as a standard. The values determined for purified myosin I using the Bradford assay were approximately 20-25% lower than values determined using the protein assay of Lowry et a/. (22) with bovine serum albumin as a standard. KC1 concentrations in column eluates were estimated by conductivity measurements. Discontinuous SDS-polyacrylamide gel electrophoresis was performed as described by Laemmli (23), followed by Coomassie blue staining according to Fairbanks et al. (24). Gels were destained in 10% acetic acid. For autoradiography of "'P-labeled proteins and phosphoamino acids, dried gels and thin layer sheets were exposed to x-ray film (Kodak X-Omat AR-2) beneath an intensifying screen (Dupont, Cronex Lightning Plus). Densitometric scans of autoradiograms and Coomassie blue-stained gels were performed a t 600 nm using a Helena Laboratories Quick-Scan gel scanner. Care was taken to scan samples within the predetermined linear response range of the instrument.

RESULTS
Purification of Myosin Z Heavy Chain Kinase-Myosin I heavy chain kinase was detected during its purification by its ability to activat.e t,he myosin I Mg'+-ATPase in the presence of actin. This assay provided a means of detecting specifically protein kinases that affect a physiologically significant property of myosin I, i.e. activation of its ATPase. Chromatography on DE-52 was the first step in the separation of myosin I kinase from myosin IA and IB, which elute together at 0.1 M KC1 (4). As shown previously by Maruta and Korn (16), the myosin I peak contains some kinase activity. We found, however, that about 85% of the myosin I kinase activity recovered from the column was in the material which did not adsorb to the DE-52 resin. Therefore, the DE-52 flow-through fraction was used for subsequent kinase purification.
Phosphocellulose chromatography of the DE-52 fraction yielded two peaks of activity, a large peak (PC1) at 0.12 M KC1 and a smaller peak (PCZ) at 0.36 M KC1 (Fig. 1). PC1 usually contained 60-70% of the total activity eluted, was well by guest on November 1, 2017 http://www.jbc.org/ Downloaded from separated from the bulk of the eluted protein, and was used for further myosin I kinase purification. While the activity in PC2 was not routinely purified, in one experiment both PC1 and PC2 were further purified by chromatography on histone-Sepharose. These partially purified kinases were found to incorporate maximally 0.8 mol of phosphate per mol of myosin IB individually and when added together (data not shown), suggesting that the two fractions phosphorylated the same site in myosin IB. The phosphocellulose peak (PC1) was further purified by Procion red dye chromatography, which yielded a major peak of activity at 0.9 M KCl, well separated from the bulk of the eluted protein (Fig. 2). In addition to activating the actinactivated Mg2+-ATPase activity of myosin I, this fraction was found to phosphorylate histone 2A with a high specific activity (see Table I). Consequently, histone-Sepharose was used as a purification step.
All the myosin I kinase activity bound to the histone-Sepharose column and was eluted as a single peak with a gradient of Mg"-ATP (Fig. 3). The specific activity of myosin I kinase was essentially constant in fractions across the peak, indicating an essentially homogenous protein preparation (data not shown). SDS-PAGE of the pooled kinase peak revealed more than 94% of the Coomassie blue stain in a 107,000-Da band (average of four separate determinations) (Fig. 4, lane 5).
Molecular sieve chromatography of myosin I kinase on Bio-Gel A-0.5m yielded a single symmetrical peak of kinase activity which eluted with the 107,000-Da band (Fig. 4, lane 6). A Stokes radius of 4.1 nm was estimated, consistent with a native molecular weight of 107,000 for a protein with a frictional ratio of approximately 1.3:l. Furthermore, as is typical of many protein kinases (25), the 107,000-Da polypeptide underwent an apparent autophosphorylation when incubated with Mg"+ and [Y-:'~P]ATP (Fig. 4, lanes 7 and 8). These results, along with the high specific activity of myosin I kinase (see below), all indicate that the histone-Sepharose peak is a highly purified protein kinase of M, = 107,000 which exists in solution as a roughly globular monomer and which phosphorylates both myosin I and histone 2A.
In a typical preparation, 0.2 to 0.5 mg of myosin I heavy chain kinase was obtained from 1000 g of cells (50 g of total protein and 25 g of protein in the 100,000 X g extract). It was difficult to obtain accurate quantitative measurements of kinase specific activity and yield at each step of the purification procedure because the purified myosin I substrate, which was always from a previous preparation, was not stable over the course of the kinase isolation (see under "Discussion"). The SDS-polyacrylamide gel in Fig. 4 indicates the approximate degree of purification of the 107,000-Da band at each chromatographic step (Fig. 4, lanes 2-5).
Myosin I heavy chain kinase was stable in terms of enzymatic activity and SDS-PAGE profile for at least 3 months when stored in 50% glycerol at -20 "C and lost less than 20% of its activity after 6 months of storage. In some fully active preparations, however, instead of the 107,000-Da band, SDS-PAGE showed two equimolar bands of approximately 50,000 Da and 60,000 Da (Fig. 4, lane 9). This material eluted at the same position on gel filtration (data not shown) as the 107,000-Da kinase. It is likely, therefore, that in these preparations the 107,000-Da kinase had been proteolytically cleaved to two peptides that stayed together under nondenaturing conditions.
Purified myosin I kinase was not contaminated by detectable levels of either a protease or a protein phosphatase. When myosins IA and IB were maximally phosphorylated by purified kinase and the proteins separated from ATP on The initial rate of phosphorylation of the indicated substrates was measured in incubations containing 20 mM imidazole (pH 7.5), 50 mM KC1, 4 mM MgCl,, 0.5 mM dithiothreitol, 0.5-1.0 mM [r-"P]ATP (50-100 pCipmol), and the indicated concentrations of substrate and kinase.  ' Acid phosphatase-treated myosin I1 containing t l mol of phosphate/mol of myosin 11. A maximum of 6 mol of phosphate/mol of myosin can be incorporated using myosin I1 heavy chain kinase (13).
Concentration is that of the phosphorylatable 20,000-Da smooth muscle light chain.
Sephadex G-25 and incubated for 2 h at 30 "C, there was no detectable loss of myosin I heavy chain phosphate. Furthermore, incubation of the purified myosins with myosin I kinase (150 ratio of kinase to myosin, w/w) for 30 min at 30 "C followed by SDS-PAGE revealed no proteolysis of the myosin I heavy chain. This latter observation is important because Maruta and Korn (12) previously showed that the actinactivated M e -A T P a s e of myosin I can be activated by proteolysis of the heavy chain, as well as by phosphorylation.
Optimal Conditions for Assay of Myosin I Heavy Chain Kinase-Because of the difficulty in obtaining pure myosin I isoenzymes in substrate concentrations, histone 2A, which is a good substrate (see below), was used in many of the experiments to determine optimal assay conditions. The kinase was relatively insensitive to alterations in pH, exhibiting a broad pH optimum centered about pH 7.5 (Fig. 5 A ) . The activity of myosin I kinase toward myosin IB was inhibited by KC1 concentrations greater than 75 mM and was 50% inhibited at 140 mM KCl, relative to the rate at 40 mM KC1 (Fig. 5B). The sensitivity to increasing ionic strength was affected by the concentration of myosin I used in the assay. When the myosin I concentration was reduced from 2.4 to 1 FM, myosin I kinase activity was significantly inhibited above 55 mM KC1 and was 50% inhibited at about 90 mM KC1 (data not shown). Therefore, for all experiments involving phosphorylation of myosin I, the final KC1 concentration of the reaction was kept between 40-50 mM. The myosin I isoenzymes have been shown to exist as soluble monomers at KC1 concentrations as low as 20 mM (18). Phosphorylation of histone 2A (67 PM) was not as sensitive to ionic strength, being significantly inhibited only above 310 mM KC1 (data not shown).
The activity of myosin I kinase toward histone 2A was dependent on Mg'+ (Fig. 5C), being completely inactive in its absence, and appeared to require low concentrations of free MgZ' for optimal activity, as shown by the increase in activity when total Mg'+ was in excess of ATP. The kinase did not, however, demonstrate a pronounced optimum concentration for free M e , as has been observed for cyclic nucleotidedependent kinases (26). Similarly, the phosphorylation rate with myosin IB as substrate was essentially constant from 1 to 9 mM free M$+ (Fig. 5C). The rate of phosphorylation of myosin IB (2 p~) varied by less than 10% between incubations containing 0.1 mM EGTA (4.36 pmol/min.mg), 20 p~ free Ca" (4.68 pmol/min.mg), 20 p~ free Ca'+ plus 2 p~ calmodulin (4.28 pmol/min.mg), 50 pM CAMP (4.11 pmol/minmg), and 20 pg/ml of CAMP-dependent protein kinase inhibitor (4.70 pmollmin. mg).
The dependence of the rate of phosphorylation of histone 2A by myosin I heavy chain kinase on the concentration of ATP was determined over a concentration range of 5 p~ to 2 mM ATP (Fig. 6). The results obeyed classical Michaelis-Menten kinetics and when plotted by Lineweaver-Burk analysis yielded a K,,, for ATP of 43 p~ (Fig. 6, inset).
Phosphorylation of Myosin I Isoenzymes by Myosin I Heavy Chain Kinase-To determine the site of phosphorylation and the stoichiometry of phosphate incorporation, myosin IA and IB were phosphorylated to a maximum extent by purified myosin I kinase. Both myosin IA and IB showed a maximum of 0.9-1.0 mol of phosphate per mol of myosin (Fig. 7). Autoradiography of SDS-polyacrylamide gels of maximally phosphorylated myosin IB showed all the '"P incorporated into the 125,000-Da heavy chain; no detectable phosphorylation of the 25,000-Da light chain was observed (Fig. 4, lanes  10-13). Similarly, no :' lP was incorporated into the 17,000-Da light chain of maximally phosphorylated myosin IA (Fig. 4,  lanes 14 and 15). Phosphoamino acid analysis of maximally phosphorylated myosin IB revealed a single phosphorylated amino acid, phosphoserine (Fig. 7). These results indicate that highly purified myosin I kinase phosphorylates both myosin IA and IB at a single site, that this site is within the heavy chain, and that, at least in myosin IB, this site is a serine residue. As shown previously (4, 16) and in more detail in an accompanying paper (18)   assays were performed with myosin IB varied from 0.3 to 2.4 P M (Fig. 8). We were limited to this narrow range because the highest concentration of purified myosin IB we obtained was only 0.75 mg/ml. The initial rate of phosphorylation of myosin IB over this concentration range increased in almost direct Acanthamoeba Myosin I Heavy Chain Kinase proportion to the increase in myosin concentration, although the slope of the line appears to begin to fall off above 1.5 PM myosin IB. The data at the very low myosin I concentrations (<l p M ) suggest something other than simple hyperbolic kinetics. Replicate measurements of phosphorylation rates at very low myosin I concentrations showed considerable variability, and it may be that the type of assay performed underestimates the phosphorylation rate when the protein concentration in the assay is very low. Nevertheless, the results in Fig. 8 indicate that at the concentrations of myosin I used in most experiments (1-2 PM), the measured rate is well below the V,,, of myosin I heavy chain kinase for myosin I.
Substrate Specificity of Myosin I Heavy Chain Kinase-The activity of purified myosin I kinase toward myosin IA and IB was compared with its activity toward the general phosphateaccepting proteins, histone 2A, casein, and phosvitin, and toward Acanthamoeba myosin I1 (Table I). As stated above, we were limited in the concentration of myosin I we could use. Nevertheless, even at these low myosin IB concentrations (1-2 PM), the specific activity was on the order of 2-4 pmol/ min.mg. Myosin I kinase phosphorylated the myosin IA isoenzyme at essentially the same rate as myosin IB. Myosin I kinase also phosphorylated histone 2A at a high rate, but phosphorylated casein a t a very low rate, and phosvitin not a t all. Absolutely no phosphorylation of Acanthamoeba myosin I1 by myosin I kinase was observed, even at very high myosin 1 kinase concentrations (40 nM). In separate experiments, no phosphorylation of myosin IA or IB by a partially purified Acanthamoeba myosin I1 heavy chain kinase could be detected, either by filter paper assay or by autoradiography (data not shown). Therefore, Acanthamoeba contains at least two myosin heavy chain kinases, one of which is specific for the myosin I isoenzymes and the other specific for myosin 11. Interestingly, myosin I heavy chain kinase also phosphorylated isolated light chains from turkey gizzard smooth muscle myosin at a significant rate.

DISCUSSION
After the discovery that heavy chain phosphorylation regulates the actin-activated Mg2'-ATPase activity of Acanthamoeba myosin I (4, 12, 16), heavy chain phosphorylation was also found to regulate Acanthamoeba myosin I1 (10, 11, 13, 14), Dictyostelium myosin (28,29) (in both cases phosphorylation inhibits the actin-activatable ATPase activity), and Physarum myosin (30) (where phosphorylation activates the actin-activatable ATPase activity). Heavy chain phosphorylation also occurs in liver macrophage (31), lymphocyte (32), and brain (33) myosins, although the stoichiometry and consequences of these phosphorylations are unknown. The purification and characterization of myosin heavy chain kinases and the mechanism(s) by which heavy chain phosphorylation regulates myosin activity is, therefore, of general importance. This paper is the first report of purification to homogeneity of any myosin heavy chain kinase. That the enzyme is highly purified is strongly supported by SDS-PAGE, by the high specific activity of the purified kinase, and by the coincidence of the protein and activity peaks on gel chromatography. Myosin I heavy chain kinase is an approximately globular protein containing one polypeptide of M , = 107,000. The kinase phosphorylates both myosin IA and IB at a high rate at one site within the heavy chain, which in the case of myosin IB is a serine residue. Although we have never had sufficient myosin I to measure it, the V,,, of myosin I kinase at excess substrate would almost certainly exceed 10 pmol/min.mg making this heavy chain kinase as active as the very active myosin light chain kinases purified from smooth (34), cardiac (35), and skeletal (36) muscles.
It was difficult to determine the yield and degree of purification of myosin I heavy chain kinase a t each chromatographic step because of the low yield and lack of stability of the purified myosin I substrate. Several quantitative measurements of kinase activity in the DE-52 flow-through fraction (1 g of protein) were made, however, using purified myosin IB (0.7 FM) as substrate and determining the initial rate of phosphorylation by the filter paper assay. Myosin I kinase in these DE-52 fractions had a specific activity of about 2.5 nmol/min.mg. Based on the specific activity of purified kinase at the same substrate concentration (about 750 nmol/min. mg; see Fig. 8), the combination of phosphocellulose, red dye, and histone-Sepharose chromatography provided about a 300-fold purification of kinase activity starting with the DE-52 eluate. It was not possible to measure the activity of myosin I kinase in the crude extract because of the contaminating myosin I. If, however, one arbitrarily assumes an 80% recovery of activity applied to the DE-52 column of which 85% was in the nonadsorbed flow-through fraction, then the DE-52 step would provide an additional 20-fold purification. With this assumption, a final yield of 0.5 mg of purified kinase would represent approximately an 11% yield of myosin I kinase from the extract with a 6000-fold purification. On this basis, 1 kg of cells would contain about 5 mg of myosin I kinase to about 150 mg of myosin IA and 100 mg of myosin IB (4, 18). Therefore, there would be sufficient kinase in the cell to phosphorylate all of the myosin I within about 5-10 s, assuming a rate of reaction equal to that in the assay in vitro and using the assumed recovery for the DE-52 chromatographic step.
Myosin I heavy chain kinase phosphorylates smooth muscle myosin light chains and we have recently found that the kinase also phosphorylates intact smooth muscle myosin and smooth muscle heavy meromyosin at high rates (at the same site as does smooth muscle myosin light chain kinase) and fully activates the actin-activated Mg2+-ATPase activity of heavy meromyosin (37). Smooth muscle light chain kinase does not phosphorylate Acanthamoeba myosin I (37). We have been unable to compare myosin I heavy chain kinase carefully to Acanthamoeba myosin I1 heavy chain kinase because of the poor yield and instability of the myosin I1 kinase. But myosin I kinase has no activity toward myosin I1 and partially purified myosin I1 heavy chain kinase has no activity toward myosin I.
We cannot be certain that the purified myosin I heavy chain kinase described in this paper is the only protein kinase in Acanthamoeba able to phosphorylate myosin I heavy chain and regulate its activity. The myosin I heavy chain kinase described by Maruta and Korn (16) was partially purified from the DE-52 fraction that contained myosin I and which we find accounts for about 15% of the total myosin I kinase activity recovered from the column. The two most prominent bands in SDS-polyacrylamide gels of this partially purified kinase were 95,000 and 58,000 Da (16). Also, as mentioned under "Results," we found a second kinase fraction, in addition to the one we purified, eluting from phosphocellulose. These other kinase fractions may contain different enzymes or, as we think more likely, they may be modified forms (proteolytic or phosphorylated products, for example) of the enzyme that has been purified. Myosin I heavy chain kinase is probably identical with the cofactor protein partially purified by Pollard and Korn (17) which was thought to have a molecular weight of about 100,000.
The mechanism by which myosin I heavy chain kinase is regulated in situ is not known. The isolated enzyme is not affected by Ca2+, Ca2+/calmodulin, or CAMP. In contrast, the  38) and nonmuscle sources (39-41) are absolutely dependent on Ca'+/calmodulin. However, several of these myosin light chain kinases have been isolated as proteolytic products of the native enzymes that possess full activity in the absence of Ca'+ (42-44). By analogy, the purified myosin I kinase we isolated might have been similarly deregulated by proteolysis but we have no evidence to suggest this. In preliminary experiments, we have found that at least 0.4 mol of phosphate can be incorporated per mol of myosin I kinase, probably by autophosphorylation. We do not know if this phosphorylation affects myosin I kinase activity. Both CAMPdependent protein kinase (26) and cGMP-dependent protein kinase (26) undergo autophosphorylation and certain properties of both enzymes are altered by autophosphorylation.
The availability of highly purified myosin I heavy chain kinase with high specific activity has allowed us to study the effects of phosphorylation of the heavy chains of myosin IA and IB on their interaction with F-actin. Our initial studies are reported in the accompanying paper (18).