Proteolytic Separation of the Actin-activatable ATPase Site from the Phosphorylation Site on the Heavy Chain of Acanthamoeba Myosin IA*

had shown that phosphorylation of the heavy chain of Acanthamoeba myosin IA is required for actin activa- tion of its Mg2+-ATPase activity and that, like the phosphorylation site, the catalytic site and the actin binding site are also on the heavy chain. We now show that limited digestion of phosphorylated myosin IA by subtilisin allows separation of the catalytically active pep- tide fragment from the phosphorylated peptide without any significant loss of actin-activated Mg2’-A FPase ac- tivity. A proteolytic fragment with full actin-activated Mg2+-ATPase activity has also been isolated from sub- tilisin digests of nonphosphorylated myosin IA, which, before proteolysis, did not have actin-activated Mg2+- ATPase activity. The simplest interpretation of these data is that, in its nonphosphorylated state, the phosphorylation site of Acanthamoeba myosin IA inhibits the catalytic site and that this inhibition can be re-versed either by phosphorylation of the site or by pro- teolytically separating it from the catalytic site. Alter-natively, phosphorylation and proteolysis may, by un- related mechanisms, induce similar conformational changes in the myosin heavy chain that lead to activa- tion of its actomyosin ATPase activity.

had shown that phosphorylation of the heavy chain of Acanthamoeba myosin IA is required for actin activation of its Mg2+-ATPase activity and that, like the phosphorylation site, the catalytic site and the actin binding site are also on the heavy chain. We now show that limited digestion of phosphorylated myosin IA by subtilisin allows separation of the catalytically active peptide fragment from the phosphorylated peptide without any significant loss of actin-activated Mg2'-A FPase activity. A proteolytic fragment with full actin-activated Mg2+-ATPase activity has also been isolated from subtilisin digests o f nonphosphorylated myosin IA, which, before proteolysis, did not have actin-activated Mg2+-ATPase activity. The simplest interpretation of these data is that, in its nonphosphorylated state, the phosphorylation site of Acanthamoeba myosin IA inhibits the catalytic site and that this inhibition can be reversed either by phosphorylation of the site or by proteolytically separating it from the catalytic site. Alternatively, phosphorylation and proteolysis may, by unrelated mechanisms, induce similar conformational changes in the myosin heavy chain that lead to activation of its actomyosin ATPase activity.
Phosphorylation of the heavy chains of single-headed Acanthamoeba myosins IA and LB is required for actin activation of their Mg"-ATPase activities (1,2). In the accompanying paper (3), we showed that the ATPase catalytic site is on the heavy chain of several myosins including Acanthamoeba myosins IA and IB. Therefore, in contrast to vertebrate smooth muscle and non-muscle myosins which are regulated by phosphorylation of their 20,000 molecular weight light chains (4), both the regulatory and catalytic sites of Acanthamoeba myosins IA and IB, as well as the actin-binding site (2), are on the same (heavy) chain. The catalytic (3) and phosphorylation (5) sites of double-headed Acanthamoeba myosin I1 are also both on the heavy chains but with the important differences (5) that there are two phosphorylation sites per heavy chain and that dephosphorylation, rather than phosphorylation, is required for actin activation of the MgZ+-ATPase activity of Acanthamoeba myosin 11.
In a general sense, modification of the phosphorylation sites of myosins could regulate their catalytic sites by either activation or derepression. In this paper, we show that the phosphorylation and catalytic sites on the heavy chain of Acari-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertzsement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. thamoeba myosin IA can be separated by subtilisin digestion and that the reisolated catalytic site has complete enzymatic activity including actin-activated Mg"-ATPase activity in the absence of the phosphorylation &e.

MATERIALS AND METHODS
Acanthamoeba myosin IA (6) and partially purified Acanthamoeba myosin I heavy chain kinase (1) were isolated as described previously. F-actin was prepared from acetone powders of rabbit skeletal muscle (7). Subtilisin was purchased from Sigma Chemicals, ADP-agarose (type IV) from P-L Biochemicals, Bio-Gel A-0.5m from Bio-Rad Laboratories, and [c~-:~'P]ATP and [y'"P)ATP from New England Nuclear. All other chemicals were reagent grade.
Protein concentrations were measured by the procedure of Lowry et al. (8); sodium dodecyl sulfate polyacrylamide gel electrophoresis was as described by Laemmli (9) and the gels were stained with Coomassie brilliant blue according to the method of Fairbanks et ai. (IO). ATPase assays were carried out by measuring the release of radioactive P, from [y-""P]ATP as described by Pollard and Korn (11). For (K',EDTA)-ATPase activity, the buffer contained 20 mM Tris/chloride, pH 7.5, 1 mM ATP, 0.5 M KCl, and 2 mM EDTA; for Mg"'-ATPase activity the buffer contained 20 mM imidazole chloride, pH 7.5, 1 mM ATP, and 2 mM MgC1.r. Protein-bound ["Plphosphate was determined by the filter paper assay described by Pettit et al. (

RESULTS AND DISCUSSION
To demonstrate the correlation between phosphorylation of Acanthamoeba myosin IA and the increase in actin-activated Mg"-ATPase activity, myosin was incubated with varying concentrations of kinase and either 0.25 mM [y-:'2P]ATP or nonradioactive ATP under otherwise identical conditions. The concentration of kinase was varied, rather than the time of incubation, because this procedure minimized proteolysis of the myosin by proteases contaminating the partially purified kinase. The incubations with radioactive ATP were monitored for incorporation of protein-bound radioactivity and the corresponding nonradioactive samples were assayed for their actin-activated Mg"-ATPase activities (Fig. 1A ). Phosphorylation reached a maximum value of 0.72 mol/mol of myosin and actin-activated Mg"'-ATPase activity reached a maximum value of 1.52 pmol/min/mg of myosin. The ATPase activity was a direct linear function of the level of phosphorylation (Fig. 1R).
In preliminary experiments, to determine the physical relationship between the phosphorylation site and the ATP binding (catalytic) site on the myosin heavy chain, one sample of Acanthamoeba myosin IA was photoaffinity-labeled at, the catalytic site with [a-"'PP]ATP ( 3 ) and a second sample was labeled on the phosphorylation site by incubation with myosin I heavy chain kinase and [y-"'P]ATP. The two samples were then separately incubated with subtilisin under identical conditions and the digestion products were separated by sodium dodecyl sulfate polyacrylamide electrophoresis (Fig. 2  The original myosin IA preparation was either already slightly degraded, or not entirely free from impurities, as shown by the pattern of staining with Coomassie blue (Fig. 2,  Lane A ) . Nonetheless, the radioactivity derived from photoaffinity labeling with [a-:"P]ATP was almost entirely restricted to the 130,000 molecular weight heavy chain (Fig. 2, Lane a). Upon digestion with subtilisin, radioactivity seemed to appear sequentially (Fig. 2, Lanes b to d ) in a 28,000 molecular weight peptide and the gel front, which would contain the radioactive 16,000 molecular weight peptide identified in the accompanying paper (3). A 115,000 molecular weight peptide may have been the fist detectable radioactive degradation product (Fig. 2, Lane b ) .
The 130,000 molecular weight heavy chain of Acanthamoeba myosin IA was also heavily labeled when the myosin was incubated with myosin I heavy chain kinase and [y-"PIATP ( Fig. 2, Lane e ) , but other minor peptides, most probably derived from the partially purified kinase preparation, were also radioactive. Upon subtilisin digestion, the radioactivity originally in the phosphorylated heavy chain seemed to appear sequentially in a 115,000 molecular weight peptide, an 87,000 molecular weight peptide, and then, finally, in unidentified peptides that ran at the gel front (Fig. 2, Lanes f and g). One interpretation of the data in Fig. 2 is that a 115,000 molecular weight peptide contained both sites but an 87,000 molecular weight peptide contained only the phosphorylation site and a 28,000 molecular weight peptide contained only the catalytic site. From this analysis it appeared that subtilisin cleaves the heavy chain of Acanthamoeba myosin IA between the phosphorylation and catalytic sites and that, therefore, it might be possible to separate and isolate the peptides that contain the two sites and study their properties.
For this experiment, Acanthamoeba myosin IA was phosphorylated to the extent of 0.79 mol/mol of heavy chain by incubation with myosin I kinase and [y-"'P]ATP and then separated from radioactive ATP and myosin I heavy chain kinase by sequential chromatography on Bio-Gel A-0.5m and I The abbreviation used is: EGTA, ethylene glycol bis(,&aminoethyl ether)N,N,N',N'-tetraacetic acid. ADP-agarose (1,6). The "P-labeled myosin IA was digested with subtilisin as described in Fig. 3. About 95% of the original (K',EDTA)-ATPase activity was recovered after this limited proteolysis. The subtilisin digest was then fractionated by chromatography on ADP-agarose. About 95% of the applied (K',EDTA)-ATPase activity was recovered from the column: 20% of the enzymatic activity was eluted with 50 mM KC1 and 80% of the enzymatic activity was eluted with 1 M KC1 (Fig.  3).
About 98% of the applied radioactivity was also recovered from the ADP-agarose column but it was distributed between the two fractions differently than the enzymatic activity ( Fig.  3): 70% of the radioactivity was eluted with 50 mM KC1 and only about 30% was eluted with 1 M KC1. Therefore, the fraction that was eluted with 1 M KC1 was about %fold enriched in functional catalytic sites relative to phosphorylation sites. Moreover, since the ratio of subtilisin to myosin IA (1:50) was greater than the maximum ratio used in the experiment described in Fig. 2 (1:200), the catalytic site and the contaminating phosphorylation site were almost certainly in separate peptides. Nevertheless, this fraction retained full actin-activated Mg2'-ATPase activity (Table I), even in the absence of added myosin I heavy chain kinase, indicating that the phosphorylation site is not required for actin-activated enzymatic activity.
In a similar experiment, nonphosphorylated Acanthamoeba myosin IA was incubated with subtilisin and the products of digestion were fractionated by gel fitration on Bio-Gel A-0.5m (Fig. 4). All of the (K',EDTA)-ATPase activity was recovered in two broad, overlapping peaks: ST-I, which was eluted in the molecular weight range of about 17,000 to 68,000, and ST-2, which was eluted a t a position corresponding to a molecular weight of less than 17,000. When analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (Fig. 5), ST-1 was found to contain no major peptide larger than molecular weight about 28,000 and ST-2 contained no peptide detectable by Coomassie blue staining of molecular weight greater than about 17,000. Aliquots of the two fractions, ST-1 and ST-2, were then photoaffinity-labeled with [a-"'PP]ATP (3) and separated by sodium dodecyl sulfate polyacrylamide electropho- Myosin was either photoaffinity-labeled with [a-""PIATP (3) or maximally phosphorylated by incubation with kinase and [y-,'"P]ATP (Fig. 1). The two preparations of radioactive myosins (20 pg) were incubated separately with 25, 50, or 100 ng of subtilisin for 15 min at 35°C in 0.05 ml of 10 mM Tris/chloride, pH 7.5, 2 mM MgClz, 1 mM dithiothreitol, and 1 mM EGTA. The samples were then heated at 95°C for 30 min in 2% sodium dodecyl sulfate, 40 mM dithiothreitol, and aliquots containing 10 pg of myosin or its digestion products were * "

TABLE I Actin-activated Mg2'-ATPase activity of the catalytic fragment isolated from the subtilisin digest ofphosphorylated
Acanthamoeba myosin IA The fraction eluted from ADP-agarose with 1 M KC1 following subtilisin digestion of phosphorylated Acanthamoeba myosin IA (Fig.  3) was dialyzed against 10 mM Tris/chloride, pH 7.5, 1 mM EGTA, 1 mM dithiothreitol, 10% sucrose, and its (K',EDTA)-ATPase and Mg'*-ATPase activities, in the presence and absence of F-actin (0.2 mg/ml) and with and without Acanthamoeba myosin I heavy chain kinase (2 mg/ml), were compared to the activities of the undegraded mvosin. resis. Only two radioactive peptides, with molecular weights of 28,000 and 16,000, were detected (data not shown). These radioactive peptides corresponded in electrophoretic mobilities to those observed when myosin IA was photoaffnitylabeled before digestion with subtilisin (3).

Md'
Even though they were derived from myosin IA that had neither been phosphorylated nor exposed to ATP prior to the enzymatic assay, both ST-1 and ST-2 contained appreciable actin-activated Me-ATPase activity in the absence of myosin I heavy chain kinase (Table 11). Thus, these data show that Acanthamoeba myosin IA can be degraded by subtilisin  Fig. 4 were separated on 16% polyacrylamide gels after heating for 30 min at 95OC in 2% sodium dodecyl sulfate, 40 mM dithiothreitol. The gels were stained with Coomassie blue.
to enzymatically active fragments with molecular weights by gel fitration as low as approximately 20,000. The peptide that contains the catalytic site in these active fractions can have a molecular weight as low as 16,000 by sodium dodecyl sulfate polyacrylamide gel electrophoresis.
The fact that proteolytic fragments of Acanthamoeba myosin IA can have full actin-activated Mg2"ATPase activity in

Actin-activated Mg2+-ATPase activity of the fractions isolated from the subtilisin digestion of nonphosphorylated Acanthamoeba myosin IA
Fractions ST-1 and ST-2 from the experiment described in Fig. 4 were dialyzed against 10 mM Tris/chloride, pH 7.5, 1 mM EGTA, 1 mM dithiothreitol, 10% sucrose, and their enzymatic activities compared to those of undegraded myosin as in Table I. ~ the absence of phosphorylation by myosin I heavy chain kinase, and when separated from the peptide that contains the phosphorylation site, proves that the phosphorylation site is not required for enzymatic activity. Therefore, it seems most likely that, for intact Acanthamoeba myosin IA, phosphorylation of the heavy chain is necessary in order to overcome the inhibition of actin-activated M$'-ATPase activity that occurs when the phosphorylation site is present in its nonphosphorylated form. According to this interpretation, the actin-activated M$'-ATPase activity of the catalytic site can also be derepressed by proteolytic removal of the peptide that contains the phosphorylation site. Another possibility, which seems to us less likely, is that phosphorylation and proteolysis are alternative ways to induce the conformational changes necessary for actin-activated Acanthamoeba myosin IA Mg"-ATPase activity. Similar conclusions have been reached for the regulation of smooth muscle myosin (13)(14)(15)(16) where the phosphorylation and catalytic sites reside on different peptides. One of the next tasks is to determine the nature of structural alterations in the myosin heavy chains induced by phosphorylation, dephosphorylation, and proteolysis.