Structure-function studies on Acanthamoeba myosins IA, IB, and II

Myosins IA and IB are globular proteins with only a single, short (for myosins) heavy chain (140,000 and 125,000 daltons for IA and IB, respectively) and are unable to form bipolar filaments. The amino acid sequence of IB heavy chain shows 55% similarity to muscle myosins in the N-terminal 670 residues, which contain the active sites, and a unique 500-residue C-terminus highly enriched in proline, glycine, and alanine. The C-terminal region contains a second actin-binding site which allows myosins IA and IB to cross-link actin filaments and support contractile activity. Myosins IA and IB are regulated solely by phosphorylation of one serine on the heavy chain positioned between the catalytic site and the actin-binding site that activates ATPase. Myosin II is a more conventional myosin in composition (two heavy chains and two pairs of light chains), heavy chain sequence (globular head 45% identical to muscle myosins and a coiled-coil helical tail), and structure (bipolar filaments). The tail of myosin II is much shorter than that of other conventional myosins, and it contains a 25 amino acid sequence in which helical structure is predicted to be weak or absent. The position of this sequence corresponds to the position of a bend in the monomer. Myosin II heavy chains also have a 29-residue nonhelical tailpiece which contains three regulatory, phosphorylatable serines. Phosphorylation at the tip of the tail regulates ATPase activity in the globular head apparently through an effect on filament structure.     

Amino-acid sequence of the hinge region in chicken myosin subfragment-2

The CNBr fragments of the hinge region in the carboxyterminal portion of long subfragment-2 derived from adult chicken pectoralis muscle myosin were isolated and sequenced by conventional methods. The alignment of these fragments was deduced from the homology of their sequences with those of other myosins, so that the sequence of the hinge region consisting of 127 amino-acid residues was determined. A comparison of this sequence with that of chicken embryonic skeletal muscle, chicken gizzard muscle and rabbit cardiac muscle (alpha-myosin) shows degrees of 95%, 36% and 82% sequence identities, respectively. Furthermore, the frequency with which hydrophobic residues are present at position "a" in seven-residues repeats of this region was significantly lower than the other portions of the rod.     

Amino-acid sequence of the short subfragment-2 in adult chicken skeletal muscle myosin

The amino-acid sequence of a short subfragment-2 in the amino-terminal portion of subfragment-2 (S-2) derived from adult chicken skeletal muscle myosin was completely determined. Peptides cleaved by cyanogen bromide and by lysyl endopeptidase of S-carboxymethylated S-2, and hydrolytic peptides obtained with trypsin or dilute acetic acid of larger CNBr fragments were isolated and sequenced. This region was composed of 257 amino-acid residues, and hydrophobic and charged residue repeat units were found highly conserved and with a periodicity in 7 or 28 residues. This sequence of the short S-2 fragment of chicken skeletal muscle myosin was compared with the sequence of chicken and rat embryonic skeletal muscle myosins, rabbit skeletal and rabbit cardiac muscle myosin (alpha-myosin heavy chain), and 95.3%, 86.8%, 89.9% and 94.2% sequence identities were observed, respectively.     

Isolation of a non-muscle myosin heavy chain gene from Acanthamoeba

We have isolated a non-muscle myosin heavy chain gene from Acanthamoeba castellanii using as a heterologous probe a sarcomeric myosin heavy chain gene from Caenorhabditis elegans. The amoeba genomic clone has been tentatively identified as containing a myosin II heavy chain gene based on hybridization to a 5300-nucleotide RNA species, hybrid selection of a mRNA encoding a 185-kDa polypeptide, specific immunoprecipitation of this polypeptide with antiserum to myosin II, and an exact match between the DNA sequence and a carboxyl-terminal myosin II peptide previously sequenced by protein chemical methods (C?té, G.P., Robinson, E.A., Appella, E., and Korn, E. D. (1984) J. Biol. Chem. 259, 12781-12787). We also sequenced a region of the gene whose deduced amino acid sequence shows strong homology with that region of muscle myosins which is thought to be involved in nucleotide binding. These results indicate that the amoeba genomic clone contains at least 90% of the coding information for the 185-kDa heavy chain polypeptide and that the bulk of the gene contains very little intron DNA. Genomic blots of amoeba DNA probed with a portion of this myosin gene indicate the presence of additional highly related sequences within the amoeba genome.     

Purification and characterization of a third isoform of myosin I from Acanthamoeba castellanii

A third isoform of myosin I has been isolated from Acanthamoeba and designated myosin IC. Peptide maps and immunoassays indicate that myosin IC is not a modified form of myosin IA, IB, or II. However, myosin IC has most of the distinctive properties of a myosin I. It is a globular protein of native Mr approximately 162,000, apparently composed of a single 130-kDa heavy chain and a pair of 14-kDa light chains. It is soluble in MgATP at low ionic strength, conditions favoring filament assembly by myosin II. Myosin IC has high Ca2+- and (K+,EDTA)-ATPase activities. Its low Mg2+-ATPase activity is stimulated to a maximum rate of 20 s-1 by the addition of F-actin if its heavy chain has been phosphorylated by myosin I heavy chain kinase. The dependence of the Mg2+-ATPase activity of myosin IC on F-actin concentration is triphasic; and, at fixed concentrations of F-action, this activity increases cooperatively as the concentration of myosin IC is increased. These unusual kinetics were first demonstrated for myosins IA and IB and shown to be due to the presence of two actin-binding sites on each heavy chain which enable those myosins I to cross-link actin filaments. Myosin IC is also capable of cross-linking F-actin, which, together with the kinetics of its actin-activated Mg2+-ATPase activity, suggests that it, like myosins IA and IB, possesses two independent actin-binding domains.     

Critical regions for assembly of vertebrate nonmuscle myosin II

Myosin II molecules assemble and form filaments through their C-terminal rod region, and the dynamic filament assembly-disassembly process of nonmuscle myosin II molecules is important for cellular activities. To estimate the critical region for filament formation of vertebrate nonmuscle myosin II, we assessed the solubility of a series of truncated recombinant rod fragments of nonmuscle myosin IIB at various concentrations of NaCl. A C-terminal 248-residue rod fragment (Asp 1729-Glu 1976) was shown by its solubility behavior to retain native assembly features, and two regions within it were found to be necessary for assembly: 35 amino acid residues from Asp 1729 to Thr 1763 and 39 amino acid residues from Ala 1875 to Ala 1913, the latter containing a sequence similar to the assembly competence domain (ACD) of skeletal muscle myosin. Fragments lacking either of the two regions were soluble at any NaCl concentration. We referred to these two regions as nonmuscle myosin ACD1 (nACD1) and nACD2, respectively. In addition, we constructed an alpha-helical coiled-coil model of the rod fragment, and found that a remarkable negative charge cluster (termed N1) and a positive charge cluster (termed P2) were present within nACD1 and nACD2, respectively, besides another positive charge cluster (termed P1) in the amino-terminal vicinity of nACD2. From these results, we propose two major electrostatic interactions that are essential for filament formation of nonmuscle myosin II: the antiparallel interaction between P2 and N1 which is essential for the nucleation step and the parallel interaction between P1 and N1 which is important for the elongation step.     

Complete primary structure of a scallop striated muscle myosin heavy chain. Sequence comparison with other heavy chains reveals regions that might be critical for regulation

We have determined the primary structure of the myosin heavy chain (MHC) of the striated adductor muscle of the scallop Aequipecten irradians by cloning and sequencing its cDNA. It is the first heavy chain sequence obtained in a directly Ca(2+)-regulated myosin. The 1938-amino acid sequence has an overall structure similar to other MHCs. The subfragment-1 region of the scallop MHC has a 59-62% sequence identity with sarcomeric and a 52-53% identity with nonsarcomeric (smooth and metazoan nonmuscle) MHCs. The heavy chain component of the regulatory domain (Kwon, H., Goodwin, E. B., Nyitray, L., Berliner, E., O'Neall-Hennessey, E., Melandri, F. D., and Szent-Gy?rgyi, A. G. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 4771-4775) starts at either Leu-755 or Val-760. Ca(2+)-sensitive Trp residues (Wells, C., Warriner, K. E., and Bagshaw, C. R. (1985) Biochem. J. 231, 31-38) are located near the C-terminal end of this segment (residues 818-827). More detailed sequence comparison with other MHCs reveals that the 50-kDa domain and the N-terminal two-thirds of the 20-kDa domain differ substantially between sarcomeric and nonsarcomeric myosins. In contrast, in the light chain binding region of the regulatory domain (residues 784-844) the scallop sequence shows greater homology with regulated myosins (smooth muscle, nonmuscle, and invertebrate striated muscles) than with unregulated ones (vertebrate skeletal and heart muscles). The N-terminal 25-kDa domain also contains several residues which are preserved only in regulated myosins. These results indicate that certain heavy chain sites might be critical for regulation. The rod has features typical of sarcomeric myosins. It is 52-60% and 30-33% homologous with sarcomeric and nonsarcomeric MHCs, respectively. A Ser-rich tailpiece (residues 1918-1938) is apparently nonhelical.     

Hydrophobicity variations along the surface of the coiled-coil rod may mediate striated muscle myosin assembly in Caenorhabditis elegans

Caenorhabditis elegans body wall muscle contains two isoforms of myosin heavy chain, MHC A and MHC B, that differ in their ability to initiate thick filament assembly. Whereas mutant animals that lack the major isoform, MHC B, have fewer thick filaments, mutant animals that lack the minor isoform, MHC A, contain no normal thick filaments. MHC A, but not MHC B, is present at the center of the bipolar thick filament where initiation of assembly is thought to occur (Miller, D.M.,I. Ortiz, G.C. Berliner, and H.F. Epstein. 1983. Cell. 34:477-490). We mapped the sequences that confer A-specific function by constructing chimeric myosins and testing them in vivo. We have identified two distinct regions of the MHC A rod that are sufficient in chimeric myosins for filament initiation function. Within these regions, MHC A displays a more hydrophobic rod surface, making it more similar to paramyosin, which forms the thick filament core. We propose that these regions play an important role in filament initiation, perhaps mediating close contacts between MHC A and paramyosin in an antiparallel arrangement at the filament center. Furthermore, our analysis revealed that all striated muscle myosins show a characteristic variation in surface hydrophobicity along the length of the rod that may play an important role in driving assembly and determining the stagger at which dimers associate.     

Myosin I heavy-chain genes of Acanthamoeba castellanii: cloning of a second gene and evidence for the existence of a third isoform

We have determined the complete sequence and structure of a second myosin I heavy-chain gene from Acanthamoeba castellanii. This gene, which we have named MIL, spans approx. 6kb, is split by 17 introns, encodes a 1147-aa polypeptide, and is transcribed in log-phase cells. The positions of six of the introns are conserved relative to a vertebrate muscle myosin gene. Similar to the previously characterized MIB heavy-chain gene, the deduced MIL heavy-chain aa sequence reveals a 125-kDa protein composed of a myosin globular head domain joined to a novel, approx. 50-kDa C-terminal domain that is rich in glycine, proline and alanine residues. There are differences, however, between MIL and MIB in the sequence organization of their unconventional C-terminal domains. We conclude from this and other data that Acanthamoeba express at least three myosin I heavy-chain isoforms: MIL, plus MIA and MIB, whose purifications have been published previously. Amoeba genomic DNA blots probed with a short, highly conserved sequence whose position is transposed between MIB and MIL indicate that the Acanthamoeba myosin I heavy-chain gene family may actually contain as many as six genes. Finally, we compared the myosin I sequences with those of two related proteins, Drosophila NinaC and the bovine myosin I-like protein, and found that a portion of the unconventional C-terminal domains of the amoeba myosins I and the bovine protein appear to be related.     

Human embryonic myosin heavy chain cDNA. Interspecies sequence conservation of the myosin rod, chromosomal locus and isoform specific transcription of the gene

A 3.6 kilobase cDNA clone coding for the human embryonic myosin heavy chain has been isolated and characterized from an expression library prepared from human fetal skeletal muscle. The derived amino acid sequence for the entire rod part of myosin shows 97% sequence homology between human and rat and a striking interspecies sequence conservation among the charged amino acid residues. The single copy gene is localized to human chromosome 17 and its expression in fetal skeletal muscle is developmentally regulated. The sequence information permits the design of isoform-specific probes for studies on the structure of the gene and its role in normal and defective human myogenesis.     

Amino acid sequence of the active site of Acanthamoeba myosin II

We have used the substrate [5,6-3H]UTP for direct photoaffinity labeling of the active site of the heavy chain of myosin II from Acanthamoeba castellanii. The only labeled peptide in a total tryptic digest had the sequence of Thr-Glu-Asn-Thr-Me2Lys-Lys (where Me2Lys represents dimethyllysine) with the substrate covalently bound to the Glu residue. This sequence differs at only one position from the sequence of residues 184-189 of nematode myosin heavy chain (Me2Lys----Lys), a post-translational modification, and at two additional positions from residues 185-190 of rabbit skeletal muscle myosin (Glu----Val and Lys----Arg). The partial sequence of a larger labeled peptide derived from total chymotryptic digestion was compatible with and extended this sequence. A 20-residue sequence that contains the active site, tryptic hexapeptide is otherwise identical in Acanthamoeba and rabbit skeletal muscle myosins and has only one more difference in nematode myosin. Because UTP is a substrate for myosin II and a "zero-length" probe, we believe that it identifies amino acid residues that are very close to the substrate during the catalytic cycle.     

Segregated assembly of muscle myosin expressed in nonmuscle cells.

Skeletal muscle myosin cDNAs were expressed in a simian kidney cell line (COS) and a mouse myogenic cell line to investigate the mechanisms controlling early stages of myosin filament assembly. An embryonic chicken muscle myosin heavy chain (MHC) cDNA was linked to constitutive promoters from adenovirus or SV40 and transiently expressed in COS cells. These cells accumulate hybrid myosin molecules composed of muscle MHCs and endogenous, nonmuscle, myosin light chains. The muscle myosin is found associated with a Triton insoluble fraction from extracts of the COS cells by immunoprecipitation and is detected in 2.4 +/- 0.8-micron-long filamentous structures distributed throughout the cytoplasm by immunofluorescence microscopy. These structures are shown by immunoelectron microscopy to correspond to loosely organized bundles of 12-16-nm-diameter myosin filaments. The muscle and nonmuscle MHCs are segregated in the transfected cells; the endogenous nonmuscle myosin displays a normal distribution pattern along stress fibers and does not colocalize with the muscle myosin filament bundles. A similar assembly pattern and distribution are observed for expression of the muscle MHC in a myogenic cell line. The myosin assembles into filament bundles, 1.5 +/- 0.6 micron in length, that are distributed throughout the cytoplasm of the undifferentiated myoblasts and segregated from the endogenous nonmuscle myosin. In both cell lines, formation of the myosin filament bundles is dependent on the accumulation of the protein. In contrast to these results, the expression of a truncated MHC that lacks much of the rod domain produces an assembly deficient molecule. The truncated MHC is diffusely distributed throughout the cytoplasm and not associated with cellular stress fibers. These results establish that the information necessary for the segregation of myosin isotypes into distinct cellular structures is contained within the primary structure of the MHC and that other factors are not required to establish this distribution.     

The carboxyl-terminal isoforms of smooth muscle myosin heavy chain determine thick filament assembly properties.

The alternatively spliced SM1 and SM2 smooth muscle myosin heavy chains differ at their respective carboxyl termini by 43 versus 9 unique amino acids. To determine whether these tailpieces affect filament assembly, SM1 and SM2 myosins, the rod region of these myosin isoforms, and a rod with no tailpiece (tailless), were expressed in Sf 9 cells. Paracrystals formed from SM1 and SM2 rod fragments showed different modes of molecular packing, indicating that the tailpieces can influence filament structure. The SM2 rod was less able to assemble into stable filaments than either SM1 or the tailless rods. Expressed full-length SM1 and SM2 myosins showed solubility differences comparable to the rods, establishing the validity of the latter as a model for filament assembly. Formation of homodimers of SM1 and SM2 rods was favored over the heterodimer in cells coinfected with both viruses, compared with mixtures of the two heavy chains renatured in vitro. These results demonstrate for the first time that the smooth muscle myosin tailpieces differentially affect filament assembly, and suggest that homogeneous thick filaments containing SM1 or SM2 myosin could serve distinct functions within smooth muscle cells.     

Protein kinase C phosphorylates nonmuscle myosin-II heavy chain from Drosophila but regulation of myosin function by this enzyme is not required for viability in flies

Conventional myosins (myosin-IIs) generate forces for cell shape change and cell motility. Myosin heavy chain phosphorylation regulates myosin function in simple eukaryotes and may also be important in metazoans. To investigate this regulation in a complex eukaryote, we purified the Drosophila myosin-II tail expressed in Escherichia coli and showed that it was phosphorylated in vitro by protein kinase C(PKC) at serines 1936 and 1944, which are located in the nonhelical globular tail piece. These sites are close to a conserved serine that is phosphorylated in vertebrate, nonmuscle myosin-IIs. If the two serines are mutagenized to alanine or aspartic acid, phosphorylation no longer occurs. Using a 341 amino acid tail fragment, we show that there is no difference in the salt-dependent assembly of wild-type phosphorylated and mutagenized polypeptides. Thus, the nonmuscle myosin heavy chain in Drosophila, which is encoded by the zipper gene, appears to be similar to rabbit nonmuscle myosin-IIA. In vivo, we generated transgenic flies that expressed the various myosin heavy chain variants in a zipper null or near-null genetic background. Like their wild-type counterparts, such variants are able to completely rescue the lethal phenotype due to severe zipper mutations. These results suggest that while the myosin-II heavy chain can be phosphorylated by PKC, regulation by this enzyme is not required for viability in Drosophila. Conservation during 530-1000 million years of evolution suggests that regulation by heavy chain phosphorylation may contribute to nonmuscle myosin-II function in some real, but minor, way.     

Complete amino-acid sequence of subfragment-2 in adult chicken skeletal muscle myosin

Although the complete amino-acid sequence of the short subfragment-2 (short S-2) and the partial sequence of the hinge region derived from adult chicken skeletal muscle myosin have been reported previously, the sequence of the N-terminal portion of subfragment-2 (S-2) and the connective portion between the above two regions could not be determined. In this study, the amino-acid sequence of these undetermined portions were completely sequenced. Furthermore, overlaps of cyanogen bromide (CNBr) peptides in the hinge region were also isolated and sequenced. Peptides obtained by hydrolysis with dilute formic acid and by digestion with lysyl endopeptidase of S-2 were purified and sequenced. These results established the complete amino-acid sequence of S-2 composed of 429 amino-acid residues. This sequence of adult chicken skeletal muscle myosin was compared with that of chicken embryonic skeletal muscle, chicken gizzard muscle and rabbit cardiac muscle myosin (alpha-myosin heavy chain) and shows degrees of 96%, 38% and 84% sequence identities, respectively. The frequency with which hydrophobic residues are present at position "a" in seven-residues repeats of the hinge region was markedly reduced when compared to the short S-2 sequence of the chicken skeletal muscle myosin.     

Genetic analysis of myosin assembly inCaenorhabditis elegans

The established observations and unresolved questions in the assembly of myosin are outlined in this article. Much of the background information has been obtained in classical experiments using the myosin and thick filaments from vertebrate skeletal muscle. Current research is concerned with problems of myosin assembly and structure in smooth muscle, a broad spectrum of invertebrate muscles, and eukaryotic cells in general. Many of the general questions concerning myosin assembly have been addressed by a combination of genetic, molecular, and structural approaches in the nematode Caenorhabditis elegans. Detailed analysis of multiple myosin isoforms has been a prominent aspect of the nematode work. The molecular cloning and determination of the complete sequences of the genes encoding the four isoforms of myosin heavy chain and of the myosin-associated protein paramyosin have been a major landmark. The sequences have permitted a theoretical analysis of myosin rod structure and the interactions of myosin in thick filaments. The development of specific monoclonal antibodies to the individual myosins has led to the delineation of the different locations of the myosins and to their special roles in thick filament structure and assembly. In nematode body-wall muscles, two isoforms, myosins A and B, are located in different regions of each thick filament. Myosin A is located in the central biopolar zones, whereas myosin B is restricted to the flanking polar regions. This specific localization directly implies differential behavior of the two myosins during assembly. Genetic and structural experiments demonstrate that paramyosin and the levels of expression of the two forms are required for the differential assembly. Additional genetic experiments indicate that several other gene products are involved in the assembly of myosin. Structural studies of mutants have uncovered two new structures. A core structure separate from myosin and paramyosin appears to be an integral part of thick filaments. Multifilament assemblages exhibit multiple nascent thick filament-like structures extending from central paramyosin regions. Dominant mutants of myosin that disrupt thick filament assembly are located in the ATP and actin binding sites of the heavy chain. A model for a cycle of reactions in the assembly of myosin into thick filaments is presented. Specific reactions of the two myosin isoforms, paramyosin, and core proteins with multifilament assemblages as possible intermediates in assembly are proposed.     

Amino acid sequence of a segment of the Acanthamoeba myosin II heavy chain containing all three regulatory phosphorylation sites

The actin-activated Mg2+-ATPase activity of myosin II from the soil amoeba Acanthamoeba castellanii is regulated by phosphorylation of 3 serine residues on the myosin II heavy chain. Partial chymotryptic digestion of 32P-labeled myosin II cleaves from the tail end of the myosin II heavy chain a small peptide which contains all three phosphorylation sites. During purification the phosphorylated peptide is resolved into several different species as a result of heterogeneity both in phosphate content and in size (probably due to chymotryptic cleavage at the carboxyl terminus). However, all forms of the peptide have an identical amino terminus. The sequence of the first 58 residues of the peptide is: N-S-A-L-E-S-D-K-Q-I10-L-E-D-E-I-G-D-L-H- E20-K-N-K-Q-L-Q-A-K-I-A30-Q-L-Q-D-E-I-D-G-T- P40-S-S-R-G-G-S-T-R-G-A50-S-A-R-G-A-S-V-R. The phosphorylated serines are at positions 46, 51, and 56. The first 36 residues of the sequence display a repeating 3-4-3-4 pattern of hydrophobic residues suggesting that this section of the peptide forms an alpha-helical coiled-coil structure. A -Gly-Thr-Pro sequence at residues 38-40 disrupts the alpha-helix and, at the same point, the repeating pattern of non-polar residues is lost. It is likely that the residues extending from Gly-38 to the end of the myosin II tail, which include the 3 phosphorylatable serines, form a randomly coiled or small globular structure. This is the first report of the sequence around the regulatory phosphorylation sites on any myosin heavy chain.     

Isolation of a cDNA encoding the motor domain of nonmuscle myosin which is specifically expressed in the mantle pallial cell layer of scallop (Patinopecten yessoensis)

It has been reported that catch and striated muscle myosin heavy chains of scallop are generated through alternative splicing from a single gene [Nyitray et al. (1994) Proc. Natl. Acad. Sci. USA 91, 12686-12690]. They suggested that the catch muscle type myosin was expressed in various tissues of scallop, including the gonad, heart, foot, and mantle. However, there have been no reports of the primary structure of myosin from tissues other than the adductor muscles. In this study, we isolated a cDNA encoding the motor domain of myosin from the mantle tissue of scallop (Patinopecten yessoensis), and determined its nucleotide sequence. Sequence analysis revealed that mantle myosin exhibited 65% identity with Drosophila non muscle myosin, 60% with chicken gizzard smooth muscle myosin, and 44% with scallop striated muscle myosin. The mantle myosin has inserted sequences in the 27 kDa domain of the head region, and has a longer loop 1 structure than those of scallop striated and catch muscle myosins. Phylogenetic analysis suggested that the mantle myosin is classified as a smooth/nonmuscle type myosin. Western blot analysis with antibodies produced against the N-terminal region of the mantle myosin revealed that this myosin was specifically expressed in the mantle pallial cell layer consisting of nonmuscle cells. Our results show that mantle myosin is classified as a nonmuscle type myosin in scallop.     

Regulation of Dictyostelium myosin I and II

Dictyostelium expresses 12 different myosins, including seven single-headed myosins I and one conventional two-headed myosin II. In this review we focus on the signaling pathways that regulate Dictyostelium myosin I and myosin II. Activation of myosin I is catalyzed by a Cdc42/Rac-stimulated myosin I heavy chain kinase that is a member of the p21-activated kinase (PAK) family. Evidence that myosin I is linked to the Arp2/3 complex suggests that pathways that regulate myosin I may also influence actin filament assembly. Myosin II activity is stimulated by a cGMP-activated myosin light chain kinase and inhibited by myosin heavy chain kinases (MHCKs) that block bipolar filament assembly. Known MHCKs include MHCK A and MHCK B, which have a novel type of kinase catalytic domain joined to a WD repeat domain, and MHC-protein kinase C (PKC), which contains both diacylglycerol kinase and PKC-related protein kinase catalytic domains. A Dictyostelium PAK (PAKa) acts indirectly to promote myosin II filament formation, suggesting that the MHCKs may be indirectly regulated by Rac GTPases.