Limited proteolysis reveals a structural difference in the globular head domains of dephosphorylated and phosphorylated Acanthamoeba myosin II.

Phosphorylation at three sites at the tip of the tail of myosin II from Acanthamoeba castellanii inactivates the actin-activated Mg(2+)-ATPase activity of filamentous myosin and the in vitro motility activity of both monomeric and filamentous myosin. To seek a structural explanation for these effects, we examined the susceptibilities of dephosphorylated and phosphorylated myosins II to endoproteinases. Endoproteinase Arg-C cleaved myosin II preferentially at two sites in the globular head, Lys-621 and Arg-638, producing an NH2-terminal fragment of about 67,000 Da and a COOH-terminal fragment of about 112,000 Da. Dephosphorylated monomers and filaments were cleaved about 3 times more rapidly than their phosphorylated counterparts principally because of a much greater rate of cleavage at Arg-638; the ratio of cleavage at Arg-638:Lys-621 was about 3 for dephosphorylated myosins and about 0.5 for phosphorylated myosins. These data demonstrate that phosphorylation at the tip of the tail of Acanthamoeba myosin II causes a conformational change in the globular head that contains the catalytic sites; therefore, this conformational change may be related to the different catalytic and motile activities of the dephosphorylated and phosphorylated enzymes.

the amino-terminal region of the heavy chains forms two globular heads while their carboxyl-terminal tails interact to form a coiled-coil a-helical rod through which the monomers can self-associate into bipolar filaments. A bend (hinge) occurs in the helical rod about 40% of the distance from the carboxyl terminus to the head-rod junction.
Phosphorylation of three sites in a short, 29-amino acid, non-helical region at the tip of the tails of the heavy chains inactivates the actin-activated Mg2"ATPase activity of myosin I1 filaments but has no effect on the actin-activated Me-ATPase activity of myosin I1 monomers (1). These, and many more, data are consistent with the concept that the enzymatic activity of each myosin molecule in a filament is * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked ''advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. not determined by its own phosphorylation state but by the level of phosphorylation of the filament as a whole.
Despite the fact that phosphorylated monomeric myosin I1 has full actin-activated Mg2+-ATPase activity, phosphorylation inhibits the ability of monomeric as well as filamentous myosin I1 to support movement of actin filaments in an in vitro motility assay (1). This result is surprising because the motility activity has been thought to be a property solely of the globular head (2), and it is difficult to imagine how this amino-terminal domain could be affected by phosphorylation at the tip of the tail.
In the present study, we report the results of limited digestion of Acantharnoeba myosin I1 by endoproteinase Arg-C. Both filamentous and monomeric myosin were cleaved preferentially within a limited region in the globular head, and both forms of dephosphorylated myosin I1 cleaved more rapidly than their phosphorylated counterparts. These results provide independent evidence that phosphorylation at the tip of the tail of monomeric Acantharnoeba myosin I1 affects the conformation of the globular head despite the fact that they are separated by approximately 90 nm of coiled-coil a-helix.

MATERIALS AND METHODS
The preparation of dephosphorylated and phosphorylated Acanthamoeba myosin I1 and rabbit skeletal muscle F-actin, ATPase assays, sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and protein assays are all described or referenced in the accompanying paper (1). Myosin was incubated with endoproteinase Arg-C from mouse submaxillary gland (Boehringer Mannheim) at 30 "C in buffer containing 20 mM Tris-HC1, pH 7.5, 1 mM dithiothreitol, and either 300 or 600 mM KC1 (for monomers) or 5 or 10 mM MgCIZ (for filaments). Digestions were stopped by the addition of L-l-tosylamido-2-phenylethyl chloromethyl ketone (50 pg/ml) and then heating at 100 "C for 3 min in the electrophoretic sample buffer. After electrophoresis, the gels (7.5%) were stained with Coomassie Blue and scanned in an LKB Ultroscan XL laser densitometer connected to an LKB 2400 Gelscan XL software package.
For NHp-terminal sequencing of the cleavage products, the polypeptide fragments were transferred from the acrylamide gels to polvinylidene difluoride microporous membranes (Millipore) using buffer containing 50 mM Tris, 358 mM glycine, 0.1% sodium dodecyl sulfate, and 20% methanol at 4 mA/cm2 for 180 min (3). After the transfer was complete, the transfer membrane was stained with Ponceau S (Sigma), and segments of the bands of interest were sequenced in an Applied Biosystems model 470A gas-phase sequenator equipped with a model 120A on-line phenylthiohydantoin analyzer. The 470 Blott protocol supplied by the manufacturer was used with either the standard cartridge blocks or the Blott cartridge blocks (kindly loaned to us by M. Raum, Applied Biosystems). The background was markedly lower with the Blott cartridge blocks.
For sedimentation velocity analysis, monomeric phosphorylated and dephosphorylated myosin I1 were prepared by dialysis against buffer containing 50 mM Tris, pH 7.0, 300 mM NaC1, 1.0 mM dithiothreitol. Protein concentrations were adjusted to 0.95 mg/ml using an extinction coefficient of 0.56 cm2/mg at 280 nm (4). The samples

20905
Phosphorylation Affects Proteolysts of Acanthamoeba Myosin I I were centrifuged at 46,600 rpm a t 20 "C in a AN-D rotor in a Beckman model E analytical ultracentrifuge. Difference scans were obtained a t 4-min intervals using a schlieren optical system.
A 2" deviation of light path in one cell allowed for simultaneous measurements of the two samples.
Sedimentation coefficients were calculated using a buffer viscosity relative to water a t 25 "C of 1.045, as measured in an Oswald viscometer, and a partial specific volume of 0.735 cc/g estimated from the amino acid composition (4).

RESULTS
When fully dephosphorylated and phosphorylated filamentous myosin I1 in 10 mM MgCI, were digested by endoproteinase Arg-C, the heavy chains were cleaved to two major products of approximately 112,000 and 67,000 Da (Fig. l), as had been observed previously for cleavage by trypsin (5, 6). Notably, the initial rate of cleavage of dephosphorylated filaments was more than 3 times faster than for phosphorylated filaments (Fig. 1). The same results were obtained for filaments in 5 mM MgCI, (data not shown). Very similar results were obtained for endoproteinase Arg-C digestion of monomeric myosin I1 in 300 mM NaCl ( Fig. 2) or 600 mM NaCl (data not shown); the same two major fragments were formed, and dephosphorylated monomers were cleaved more rapidly than phosphorylated monomers. The same relative rates of cleavage of dephosphorylated and phosphorylated myosin I1 were obtained when the buffer contained 1 mM ATP and when 8 ~L M F-actin was added to the filamentous endoproteinase Arg-C (pr0teinase:myosin = 1:25, w/w = 1 3 , rnol/mol heavy chain) in buffer containing 10 mM M&I, a t 30 "C, and samples were removed at the indicated times for analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis as described under "Materials and Methods." The inset shows the relevant portion of the Coomassie Blue-stained gels. The only major hands ohserved were the undigested heavy chain, the 112,000-Da COOH-terminal fragment, and the 67,000-Da NH,-terminal fragment. The first 5 lanes from left to right are aliquots from the digestion of dephosphorylated myosin I1 removed a t 0, 30, 60, 120, and 240 min. The first 5 lanes from right to left are the corresponding samples from the digestion of phosphorylated myosin 11. The phosphorylated 112,000-Da fragment migrates more slowly than the dephosphorylated fragment. The percent of the heavy chain cleaved was calculated independently for each lane from gel scans assuming equal staining of the three hands per unit mass. This experiment has heen repeated 3 times (and 2 additional times in 5 mM MgCIJ with essentially identical results. D, dephosphorylated myosin 11; P, phosphorylated myosin 11. myosin, although in the latter case the rates of digestion were appreciably lower. By autoradiography of gels prepared from digests of "Tlabeled phosphorylated myosin 11, it was found that, as with limited trypsin digests (5, 61, the 112,000-Da pol-ypeptide contained the phosphorylation sites (data not shown), Le. it was derived from the carboxyl-terminal end of the heavy chain. Thus, the initial cleavage by endoproteinase Arg-C was well within the NH,-terminal globular head about 67,000 Da from the amino terminus. T o determine the precise site(s) of endoproteinase cleavage within the globular head, we performed NH2-terminal sequence analysis of both major fragments.
As expected, no sequence was obtained from the 67,000-Da fragment consistent with its being derived from the amino terminus of the heaw chain which is blocked. Unexpectedly, however, two sequences were obtained for the 112,000-Da fragment a t all time points for both dephosphorylated and phosphorylated, monomeric, and filamentous myosins. The data in Table I show the results for the first eight cycles for monomeric dephosphorylated and phosphorylated myosin I1 after digestion for 4 h. The data for the dephosphorylated sample are most easily interpreted because one sequence was present at about 3 times the concentration of the other. The major sequence corresponds to residues Ser-639 through Gly-646 (7), which would he produced by cleavage at Arg-638. The minor sequence corresponds to residues Ala422 through Ala-629 ( 7 ) , which would result from cleavage at Lys-621. The ratio of cleavage at Arg-638:Lys-621 was about 31. The same two amino acids were obtained at each of the corresponding cycles when the phosphorylated sample was sequenced, hut the ratio of cleavage at Arg-638:Lys-621 was only about O.6:l (Table I). Sequence analysis of the 112,000-Da fragments from filamentous dephosphorylated and phosphorylated myosins gave similar results (Fig. 3). Thus, the principal effect TABLE I Amino-terminal sequences of the 112,000-Da cleavage fragment of monomeric dephosphorylated and phosphorylated myosin 11 The 112,000-Da bands produced after digestion of monomeric dephosphorylated (Dephos.) and phosphorylated (Phos.) myosin I1 with endoproteinase Arg-C for 4 h were subjected to NHz-terminal sequencing as described under "Materials and Methods." Approximately 360 pmol of myosin I1 heavy chain were applied to each lane of the electrophoretic gel, but only a portion of the recovered peptides was sequenced. Two amino acids were detected at each cycle for each sample. It was apparent from the relative yields of each residue for the dephosphorylated sample that two sequences were present corresponding precisely to segments Ser-639 to Gly-646, produced by cleavage at Arg-638, and Ala-622 to Ala-629, produced by cleavage at Lys-621. The relative yields of the two fragments cannot be compared between the dephosphorylated and phosphorylated samples, but the relative yields of the two sequences within each sample individually are meaningful. The same two sequences were obtained at all time points and for filamentous as well as monomeric myosins.  FIG. 3. The ratio of cleavage of monomeric and filamentous dephosphorylated and phosphorylated myosin I1 by endoproteinaee Arg-C at Arg-638:Lys-621. The ratios were calculated from the data in Table I  of phosphorylation on endoproteinase Arg-C digestion of myosin I1 was to reduce substantially the rate of cleavage at Arg-638 and perhaps increase slightly the rate of cleavage at Lys-621 with a resulting net decrease in the overall rate of cleavage.
Sedimentation velocity analysis gave sz0+, values of 7.87 for both phosphorylated and dephosphorylated myosins in 300 mM NaC1, in agreement with Sinard et al. (8) who reported s values of 7-8 monomers under these conditions. T o evaluate the possibility that the myosin monomers might be in equilibrium with a small fraction of higher oligomers that were actually the substrate for the endoproteinase, we determined the effect of myosin concentration on the rate of cleavage. If, for example, the substrate were dimers or higher oligomers, the initial rate of cleavage would be expected to be proportional to an exponential power of 2 or greater of the myosin concentration. No evidence for this was found either for a function of endoproteinase concentration. The conditions were the same as in the experiment described in Fig. 3 except that myosin 11 at a concentration of 200 pg/ml (0.5 p~) was incubated with endoproteinase Arg-C at proteinase:myosin ratios of 1:25, 1:10, and 1:5 (w/w) = 1:3, 1:1.5, and 1:0.6 (mol/mol heavy chain). D, dephosphorylated myosin II; P, phosphorylated myosin 11. dephosphorylated or phosphorylated myosin I1 (Fig. 4).
It appeared from the data in Figs. 1, 2, and 4 that the rate of myosin digestion slowed substantially when about 50% of the heavy chains had been cleaved. To examine this further, the ratio of endoproteinase Arg-C to myosin I1 was increased as much as &fold to a molar ratio (endoproteinase:myosin I1 heavy chain) of 1.5:l. Although, as expected, the initial rate of cleavage increased with protease concentration, it slowed appreciably when cleavage reached about 50% (Fig. 5). More extensive cleavage occurred when the molar ratio was increased to 4:l (endoproteinase:myosin I1 heavy chain), but dephosphorylated myosin I1 was still cleaved much more rapidly than phosphorylated (data not shown).

DISCUSSION
The two cleavage sites, Arg-638 and Lys-621, are well within the subfragment 1-like globular head of myosin I1 and fairly distant from the head-tail junction which is conventionally assigned to Pro-847 but which Rimm et al. (9) suggest may be near residue 900 in Acantharnoeba myosin 11. Both sites are in the region of the 5O-kDa/2O-kDa tryptic cleavage site characteristic of myosins and which occurs at Arg-642 in Acanthamoeba myosin 11. 'The selectivity of cleavage by both trypsin and endoproteinase Arg-C is interesting given that there are 3 Lys and 3 Arg residues between Lys-621 and Lys-M. A. L. Atkinson, personal communication.

(7)
and that endoproteinase Arg-C generally has a strong preference for cleavage at arginines. It is also of interest that endoproteinase Arg-C did not cleave at either of the 2 Arg residues in the non-helical tail (data not shown) that were the preferred sites in an earlier study (10) in which slightly different incubation conditions were used (18 h at 0 "C).
Phosphorylation of one of the two pairs of light chains of smooth muscle myosin and certain vertebrate non-muscle myosins converts monomers from a folded 10 S conformation (in which the tail interacts with the head) into an extended 6 S conformation (11-13), which is the immediate precursor of filaments. The 6 and 10 S conformations also show different susceptibilities to protease digestion (14, 15). A similar conformational change cannot, however, be the basis of the different susceptibilities of dephosphorylated and phosphorylated myosin I1 to endoproteinase Arg-C cleavage because phosphorylated and dephosphorylated monomers have the same sedimentation coefficient and because endoproteinase digestion of filaments, in which the tails cannot be folded back on the heads, is similarly affected by phosphorylation.
Recent electric birefringence studies (16) have shown that filaments of dephosphorylated myosin I1 are much more rigid than filaments of phosphorylated myosin 11. This has been rationalized on the assumption that the hinge region in each monomer in the filament lies near the phosphorylation sites of other molecules in the filament (8, 17), and thus the state of phosphorylation at the tips of the tails could influence the conformation around the hinge. The observations reported in this paper must have a different explanation, however, because phosphorylation affects the rate of endoproteinase Arg-C digestion of monomeric as well as filamentous myosin 11, and the phosphorylation sites and the hinge region within the same molecule are too far apart for one to directly affect the other.
Thus, if, as seems most likely, there is a single explanation for the very similar results obtained for monomeric and filamentous myosin 11, it would seem that it can be neither a large change in shape (such as the 6 S = 10 S interconversion) nor a more subtle change in conformation at the hinge region in the rod. It is as if the consequences of phosphorylation at the tip of the tail of Acantharnoeba myosin I1 are projected through the coiled-coil, helical tail to the globular head. Whatever the specific mechanism, it may be caused by the significant change in charge that results from the addition of 3 a.