Baculovirus expression of chicken nonmuscle heavy meromyosin II-B. Characterization of alternatively spliced isoforms.

We have expressed two truncated isoforms of chicken nonmuscle myosin II-B using the baculovirus expression system. One of the expressed heavy meromyosins (HMMexp) consists of two 150-kDa myosin heavy chains (MHCs), comprising amino acids 1-1231 as well as two pairs of 20-kDa and 17-kDa myosin light chains (MLCs) in a 1:1:1 molar ratio. The second HMMexp was identical except that it contained an insert of 10 amino acids (PESPKPVKHQ) at the 25-50-kDa domain boundary in the subfragment-1 region of the MHC. These 10 amino acids include a consensus sequence (SPK) for proline-directed kinases. Expressed HMMs were soluble at low ionic strength and bound to rabbit skeletal muscle actin in an ATP-dependent manner. These properties afforded a rapid purification of milligram quantities of expressed protein. Both isoforms were capable of moving actin filaments in an in vitro motility assay and manifested a greater than 20-fold activation of actin-activated MgATPase activity following phosphorylation of the 20-kDa MLC. HMMexp with the 10-amino acid insert was phosphorylated by Cdc2, Cdk5, and mitogen-activated protein kinase in vitro to 0.3-0.4 mol of PO4/mol of MHC. The site phosphorylated in the MHC was identified as the serine residue present in the 10-amino acid insert and its presence was confirmed in bovine brain MHCs. Characterization of the baculovirus expressed noninserted and inserted MHC isoforms with respect to actin-activated MgATPase activity and ability to translocate actin filaments in an in vitro motility assay produced the following average values following MLC phosphorylation: noninserted HMMexp, Vmax = 0.28 s−1, Km = 12.7 μM; translocation rate = 0.077 μm/s; inserted HMMexp, Vmax = 0.37 s−1, Km = 15.1 μM; translocation rate = 0.092 μm/s.

and can be subdivided into a number of different isoforms present in striated muscle, smooth muscle, and nonmuscle cells. Our recent interest has focused on vertebrate cytoplasmic myosin II, which is present in both muscle and nonmuscle cells. To date, two separate genes, located on two different human chromosomes (2,3), have been shown to encode vertebrate nonmuscle myosin heavy chains (MHCs), 1 which we refer to as MHC II-A (22q11.2) and MHC II-B (17p13). Since it is now clear that some of these two isoforms are present in distinct locations in a single cell (4,5), there is reason to believe that each isoform may have a specific function, in addition to possibly overlapping functions. Moreover, previous studies have shown that these two isoforms are expressed in a tissue-dependent manner with brain and testes being particularly enriched for MHC II-B and spleen and intestines containing mostly MHC II-A, while human platelets and rat basophil leukemic cells contain only MHC II-A (6 -8).
MHC subfragment-1 (S-1) can be proteolytically cleaved at two sites, one located about 25 kDa and the second about 75 kDa from the amino-terminal end, giving rise to 25-, 50-, and 20-kDa tryptic fragments (9). These two proteolytically susceptible sites correspond to regions of the molecule that were not resolved in the three-dimensional crystallographic structure of chicken skeletal myosin S-1 and probably are present as disordered surface loops (10). The region at the junction of the 25/50-kDa fragments is termed loop 1 and is near the ATP binding domain. The region at the junction of the 50/20-kDa kDa fragments is termed loop 2 and is near an actin binding domain (10 -12). Recently, we and others have demonstrated that these regions serve as sites for alternative splicing of mRNA to produce isoforms of nonmuscle MHC II-B (13-15) and smooth muscle MHCs (16 -18). mRNA encoding nonmuscle MHC-B has been shown to generate two different insertions in loop 1, referred to as MHC II-B1, and one in loop 2, referred to as MHC II-B2. 2 The insert in loop1 starts just after amino acid 211, and consists of either 10 or 16 amino acids (see Fig. 1) (14). Sequence of human genomic DNA from this area revealed the presence of two exons, one encoding 10 amino acids and the second encoding 6 amino acids, consistent with the idea that these sequences are generated by alternative splicing of pre-mRNA and that the 16-amino acid insert differs from the 10-amino acid insert in that it is encoded by both exons instead of just one (19) (Fig. 1). The 10-amino acid inserted isoform has been shown to be highly expressed in mammalian cerebral cortex and retina, whereas mRNA encoding the 16-amino acid inserted isoform has only been detected in human retinoblastoma cell lines, * 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 1 The abbreviations used are: MHC, myosin heavy chain; myosin S-1, myosin subfragment-1; MLC, myosin light chain; MLC 20 , 20-kDa regulatory light chain; MLC 17 , 17-kDa essential light chain; HMM, heavy meromyosin; MOPS, 3-(N-morpholino)propanesulfonic acid; MAP kinase, mitogen-activated protein kinase; PMSF, phenylmethylsulfonyl fluoride. 2 We refer to the noninserted MHC as MHC II-B, the MHC with the 10-amino acid insert as MHC II-B1, the expressed HMM with the insert as HMM i (inserted) and without the insert as HMM n (noninserted). HMM exp generically refers to both HMM i and HMM n .
Expression of MHC II-B1 in nonmuscle cells appears to vary in both a species-and tissue-dependent manner. In Xenopus, an almost identical insert of 16 amino acids in MHC II-B, starting after amino acid 211, is present in all cells examined to date (20). Unlike avian and mammalian cells, MHC II-B (lacking the insert) is not expressed in Xenopus. In contrast to its ubiquitous expression in Xenopus cells, expression of MHC II-B1 (with the 10-amino acid insert) in avian and mammalian cells is almost always confined to neuronal tissue and neuronal cell lines, where it is accompanied by expression of the noninserted isoform. Of note, the constitutively expressed Xenopus MHC II-B1 isoform has been shown to be phosphorylated by cyclin p34 cdc2 (Cdc2) kinase both in vitro and in situ within the inserted region (21).
To date, there has been no careful characterization of the differences between myosin containing MHC II-B1 and MHC II-B. This has largely been due to the inability to obtain sufficient quantities of pure myosin II-B1 from avian and mammalian cells. In this paper, we focus on the differences between nonmuscle MHC II-B and MHC II-B1, the isoform containing the 10-amino acid insert. We used the baculovirus system to express a heavy meromyosin (HMM)-like product of these two alternatively spliced isoforms along with both the 20-kDa and 17-kDa myosin light chains. We then characterized each isoform with respect to its ATPase activity and ability to propel actin filaments in an in vitro motility assay. The MHC II-B1, but not the MHC II-B isoform, contains a putative phosphorylation site for proline-directed kinases, which we demonstrate can be phosphorylated by cyclin-dependent and mitogen-activated protein (MAP) kinases. The site phosphorylated is within the 10-amino acid insert. We also show that MHC purified from bovine brains contains this site and can also be phsophorylated.

EXPERIMENTAL PROCEDURES
Construction of Transfer Vectors-Three different transfer vectors were employed for transfection of Sf9 cells. cDNA encoding the chicken MHC II-B HMM-like isoform was obtained by utilizing clone S-1, previously derived from a chicken brain library (13), which is a NotI fragment that contains 74 nucleotides of untranslated mRNA at the 5Ј end and terminates at nucleotide 3,693. Single-stranded oligonucleotide adaptors with the following sequences: 5Ј-CT AGC GAT CAG CGT TAG CAC TAT TC-3Ј; 3Ј-G CTA GTC GCA ATC GTG ATA AGC CGG-5Ј; 5Ј-G GCC GAA TAG TGC TAA CGC TGA TCG-3Ј; and 3Ј-CTT ATC ACG ATT GCG ACT AGC GAT C-5Ј were annealed and ligated to the NotI sites. In addition to providing NheI sites for ligation into pBlueBac, the double-stranded oligonucleotides also provided multiple stop sites.
For construction of the transfer vector containing the 30-nucleotide insert between T633 and G634, the MHC II-B cDNA encoding the HMM fragment was removed from pBlueBac using BamHI and subcloned into pBluescript. Clone S4, which is derived from the same chicken brain library and contains the 30-nucleotide insert flanked by NsiI sites (13), was restricted and the resulting 539-base pair fragment was ligated into the same site in pBluescript containing the truncated clone MHC II-B (13). The cDNA encoding the MHC II-B1 HMM-like fragment was then ligated into pBlueBac II at the NheI site.
cDNAs encoding bovine nonmuscle MLC 17B (22)(a gift of Dr. David Hathaway, Bristol Myers Squibb Co., Princeton, NJ) and chicken nonmuscle MLC 20 (23)(a gift of Dr. C. Chandra Kumar, Schering-Plough Corp., Bloomfield, NJ) were cloned into the pAcUW51 transfer vector using polymerase chain reaction-derived clones. cDNA encoding MLC 20 was cloned into the BglII site, placing it under control of the p10 promoter, and cDNA encoding MLC 17 was cloned into the BamHI site, under control of the polyhedrin promoter. Orientation and sequence of both light chains were verified using the dideoxy chain termination method (24).
Transfection-Transfection of 3 ϫ 10 6 Sf9 insect cells with either the MHC II-B or MHC II-B1 truncated constructs was carried out using 3 g of plasmid DNA, 1 g of linear AcMNPV DNA, 1 ml of Grace's medium, and 20 l of a cationic liposome solution as per the manufacturer's instructions (Invitrogen, San Diego, CA). The vector containing both the MLC 20 and MLC 17 was transfected using 2 g of plasmid DNA and 0.5 g of BaculoGold DNA using the Ca 3 (PO 3 ) 2 method as per the manufacturer's instructions (PharMingen, San Diego, CA). Plaque assay purification and viral amplification was carried out according to the manufacturer's instructions. Some plaque assays, amplifications, and Sf9 infections were carried out by Cell Trends (Middletown, MD).
Infection and Preparation of Myosin-1 ϫ 10 9 Sf9 cells were coinfected at a multiplicity of infection of 5 with respect to both the MHC virus (II-B or II-B1) and with the virus containing both MLCs. Infected cells were harvested by sedimentation after 72 h of growth, and the pellet was washed twice with phosphate-buffered saline, quick-frozen in liquid nitrogen, and stored at Ϫ80°C.
Actin Selection-Final purification of the HMMs made use of the ability of actin to bind myosin in the absence, but not the presence, of ATP. The 40 -60% (NH 4 ) 2 SO 4 fraction (10 ml) was made 5 M with respect to F-actin and 5 M with respect to phalloidin and sedimented at 543,000 ϫ g for 15 min in a Beckman TLX Ultracentrifuge. The pellet was suspended in 2-3 ml of Buffer A containing 5 M phalloidin and the suspension homogenized in a Teflon glass homogenizer. Following sedimentation, the resulting pellet was solubilized in 0.5-1.0 ml of Buffer A, 5 M phalloidin, 1 mM ATP, and 5 mM MgCl 2 . Resedimentation at 543,000 ϫ g for 15 min results in a supernatant that contains the purified HMM exp . Practically all of the HMM exp appears to be soluble and cycles appropriately with actin in this procedure. Typically, there was no contamination with endogenous Sf9 cell myosin (see Fig. 2). Prior to characterization, the purified HMM was dialyzed in a 500-fold excess of 25 mM KCl, 10 mM MOPS (pH 7.2), 1 mM MgCl 2 , 0.1 mM EGTA, 3 mM NaN 3 , 0.1 mM PMSF, and 5 mM dithiothreitol.
Purification of Bovine Brain Myosin-Blood vessels and meninges were carefully removed from fresh bovine brains and the brains washed with ice-cold physiological saline solution. The tissue was homogenized in a buffer containing 20 mM MOPS (pH 7.2), 2 mM EDTA, 2 mM MgCl 2 , 3 mM NaN 3 , 2 mM dithiothreitol, 0.1 mM PMSF, 5 g/liter leupeptin. The mixture was sedimented (10,000 ϫ g for 60 min) and myosin was extracted from the pellet in the presence of 5 mM ATP, 30 mM MOPS (pH 7.2), 1 M NaCl, 2 mM EGTA, 4 mM EDTA, 3 mM NaN 3 , 2 mM dithiothreitol, 0.1 mM PMSF, and 5 g/liter leupeptin. Following sedimentation (80,000 ϫ g for 60 min), the supernatant was subjected to ammonium sulfate fractionation, and the 35-60% fraction was solubilized and chromatographed on a Sepharose CL-4B column. Fractions containing myosin were pooled, concentrated, and used for phosphorylation.
Phosphorylation-The following kinases were used to phosphorylate myosin and HMM: Cdc2 kinase, MAP kinase, Cdk5 kinase, protein kinase C, cAMP-dependent kinase, and calmodulin-dependent kinase II (CAM kinase II). Human HeLa cell Cdc2 kinase was a gift of Fumio Matsumura (Rutgers University, NJ), sea star oocyte Cdc2 kinase (5 ng/ml) was purchased from Upstate Biotechnology Inc. (Lake Placid, NY), and baculovirus expressed human Cdc2 kinase was purchased from New England Biolabs (Beverly, MA). Purified chicken HMM exp or bovine brain myosin purified from bovine brain cerebral cortex (0.2-1 mg/ml), was incubated at 30°C for 1-3 h in 50 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 0. Phosphorylation of myelin basic protein or the synthetic peptide was used as a positive control. Conditions for assay of cAMP-dependent protein kinase, protein kinase C, and CAM kinase II have been published (25). Reactions were terminated by addition of Laemmli sample buffer (26) and boiled for 2 min. Samples were subjected to electrophoresis in SDS 8 -16% gradient polyacrylamide Tris-glycine gels (Novex, San Diego, CA) or as described in figure legends. The stoichiometry of MHC-B phosphorylation was determined by extraction of the 32 P-phosphorylated MHCs from the gels using Solvable (DuPont NEN), followed by liquid scintillation counting. In some cases, the stoichiometry was confirmed by eluting the phosphorylated MHC from the SDSpolyacrylamide gels using tryptic digestion (see below), followed by liquid scintillation counting of an aliquot of the extracted peptides.
Tryptic Digestion and Mapping of Phosphopeptides-Expressed MHC, MLC, and purified bovine brain MHC were excised from the SDS-polyacrylamide gel following Coomassie Blue staining and digested as outlined previously (27). The digested peptides were dried repeatedly to remove NH 4 HCO 3 and then subjected to isoelectric focusing (LKB 2117 Multiphor) at 25 watts/1000 V for 1 h at 4°C using 1 M H 3 PO 4 as the anode buffer and 1 M NaOH as the cathode buffer (28). The gels were dried in a heated vacuum dryer and subjected to autoradiography. A peptide standard was synthesized based on the expected tryptic phosphopeptide, DHNIPPESPKPVK, assuming that the KP peptide bond would not be cleaved by trypsin. This synthetic peptide was found to be a substrate for proline-directed kinases Cdc2 kinase, Cdk5 kinase, and MAP kinase and was used as a peptide standard in the isoelectric focusing gels.
Assays for Actin-activated MgATPase Activity-Assays were carried out in a final volume of 100 l using actin concentrations from 6 to Aliquots were removed for at least four time points, and P i release was measured as described (29). V max and K m were determined by fitting the data to the Michaelis-Menten equation.
In Vitro Motility Assay-The sliding actin filament in vitro motility assay was conducted as described previously with some modifications (30). Unphosphorylated HMM exp (0.1-0.2 mg/ml) was adsorbed to the nitrocellulose-coated surface for 1 min. Unbound HMM exp was removed and the surface blocked by washing and incubating with 0.5 M NaCl, 10 mM MOPS (pH 7.0), 0.1 mM EGTA, 5 mM dithiothreitol, 1 mg/ml bovine serum albumin. HMM exp bound to the surface was phosphorylated by incubating for 2 min with 20 mM KCl, 20 mM MOPS (pH 7.2), 5 mM MgCl 2 , 1 mM ATP, 0.1 mM EGTA, 0.2 mM CaCl 2 , 5 mM dithiothreitol, 10 Ϫ7 M calmodulin, 5 ϫ 10 Ϫ8 M myosin light chain kinase in the presence of 5 M unlabeled F-actin to improve the quality of movement (30). Following washout of this solution and addition of rhodamine phalloidin actin buffer (20 mM KCl, 20 mM MOPS, 5 mM MgCl 2 , 0.1 mM EGTA, 5 mM dithiothreitol, and 20 nM rhodamine phalloidin actin), the movement was initiated by the addition of assay buffer containing 80 mM KCl, 20 mM MOPS (pH 7.2), 5 mM MgCl 2 , 1 mM ATP, 0.1 mM EGTA, 200 nM smooth muscle tropomyosin, 50 mM dithiothreitol, 2.5 mg/ml glucose, 0.1 mg/ml glucose oxidase, 2 g/ml catalase, 0.7% methylcellulose, 30°C. The data were recorded as 2 min fields on sVHS tapes using image averaging and minimal illumination levels. Actin filament movement was quantified according to Homsher et al. (31). Data points were taken at 10-s intervals, which resulted in centroid displacements of 4 -5 pixels. Data were taken from individual preparations only if 80 -100% of the actin filaments in a field were moving simultaneously.
Miscellaneous Procedures-Rat brain calmodulin (32), chicken gizzard myosin light chain kinase (33), rabbit skeletal muscle actin (34), smooth muscle HMM (35), and rabbit skeletal muscle tropomyosin (36) were purified as described previously. Protein concentrations were determined using the Bio-Rad Protein Assay. Smooth muscle HMM was used as a standard.

Purification of HMM exp -As detailed under "Experimental
Procedures," Sf9 cells were coinfected with two viral constructs, one containing cDNA encoding an HMM-like fragment (amino acids 1-1231) of chicken nonmuscle MHC II-B and a second virus containing cDNA encoding both MLC 20 and MLC 17 . In separate experiments, Sf9 cells were coinfected with viruses containing cDNA encoding an HMM-like fragment of MHC II-B1, which contains 10 extra amino acids (PESPKPVKHQ) starting after amino acid 211 along with the same MLCs. Our initial experience had shown that expression of the MHC 150 fragment (150 kDa) alone or the MHC 150 and the MLC 17 lead to an insoluble, aggregated product, which sedimented at velocities of 47,000 ϫ g (data not shown). In contrast, coexpression of the MHC 150 along with both MLCs resulted in the expressed product being soluble at both high (0.6 M) and low (0.02 M) NaCl concentrations following sedimentation at 300,000 ϫ g for 1 h.
Purification of HMM without the insert (HMM n ) and with the insert (HMM i ) is shown in Fig. 2. The only difference between these two isoforms is the presence of the 10 amino acids in loop 1 starting after residue 211 in the MHC. Lanes 1 and 4 show the pattern of polypeptide staining following the initial extraction of Sf9 cells. Examination of both the low speed and high speed (300,000 ϫ g) pellets revealed that most of the MHC exp was soluble under these conditions (data not shown). Lanes 2 and 5 show the polypeptide pattern following fractionation of the extract supernatant with 40 -60% ammonium sulfate. Lanes 3 and 6 show the purified HMM exp following release from F-actin by MgATP. Virtually all of the HMM exp bound to actin in the absence of ATP and was released into the supernatant in the presence of ATP (data not shown). Thus, most of the expressed HMM heavy chains combined with light chains was soluble and bound to actin in an ATP-dependent manner. Although both the extract and 40 -60% ammonium sulfate fraction show overexpression of MLC 20 compared to MHC 150 , scanning of the purified HMM exp gave a molar ratio of 1:1:1 for the MHC 150 and two MLCs. Using the procedure outlined above and detailed under "Experimental Procedures," we are able to purify between 0.4 and 2 mg of purified HMM from 10 9 infected cells (650 ml of 1.5 ϫ 10 6 cells/ml).
Characterization of the Two Expressed HMMs- Table I compares the V max of the actin-activated MgATPase activity, the K m of HMM exp for actin, and the velocity of actin filament propulsion for both isoforms. The data show that the presence of the 10-amino acid insert has only a modest effect on these parameters. Fig. 3 depicts the difference in the in vitro motility assays for six different preparations of the two isoforms that were prepared at the same time. Although preparations 2, 4, and 5 appear to show some differences in the rate of actin movement between the two isoforms, this difference does not always appear to be significant. The increase in the average mean velocity for HMM i compared to HMM n is about 20% (Table I). Likewise, HMM i shows a similar small increase in the actin-activated MgATPase activity compared to HMM n .
Phosphorylation of HMM exp -The presence of a consensus sequence for proline-directed kinases (SPK) (37) suggested that the single serine residue in the 10-amino acid insert of the expressed chicken brain MHC II-B1 might be phosphorylated by Cdc2 kinase, MAP kinase, and/or the brain specific Cdk5 kinase. Fig. 4 shows a Coomassie Blue-stained gel and corresponding autoradiogram of HMM n and HMM i following phosphorylation by Cdk5 kinase. As can be seen from the autoradiogram, only the isoform containing the inserted sequence is phosphorylated on the heavy chain. In addition, the MLC 20 of both isoforms is also phosphorylated, although to a lesser extent.
In order to identify the site(s) phosphorylated by Cdk5 ki-nase, both the phosphorylated MHC 150 and MLC 20 bands were excised from the gel, digested with trypsin, and subjected to gel isoelectric focusing. Fig. 5 shows that the major phosphopeptide generated by trypsin comigrates with a synthetic phosphopeptide with the amino acid sequence: DHNIPPESPKPVK. This peptide represents the amino acid sequence of the predicted tryptic peptide from the inserted MHC sequence, assuming that trypsin would not cleave the KP peptide bond (Fig. 1). Comigration of the MHC tryptic phosphopeptide with the phosphorylated standard peptide is consistent with serine 214, the only serine (or threonine) present in the inserted sequence being the phosphorylated residue. The minor peptide seen near

FIG. 2. Purification of HMM n and HMM i from baculovirus-infected Sf9 cells.
A Coomassie Blue-stained SDS 8 -16% gradient polyacrylamide gel is shown for HMM n (no insert) and HMM i (containing the 10-amino acid insert). Lanes 1 and 4 show the polypeptide pattern of the Sf9 cell extract prior to sedimentation. A total of 2 l and 3 l out of 50 ml were electrophoresed for HMM n and HMM i , respectively. Lanes 2 and 5 show the 40 -60% ammonium sulfate fraction. Sample size was 3 l out of 15 ml. Lanes 3 and 6 show the purified HMM n and HMM i following addition of ATP and sedimentation to release actin (see "Experimental Procedures" for details). Sample size was 2 l out of 0.9 ml.

TABLE I Comparison of kinetic values for the expressed isoforms
The values for V max and K m are the mean and standard deviation from five parallel preparations of HMM i and HMM n . The values for translocation rate of actin filaments is the mean and standard deviation of the mean from six parallel preparations of HMM i and HMM n . There was no statistical difference between the V max and K m for the two isoforms, but the difference in translocation rate was significantly different (p Ͻ 0.01) by a paired t test. the bottom of the gel is most likely due to partial cleavage of the RKD sequence at the amino-terminal end of the expected tryptic peptide to yield KDHNIPPESPKPVK, in addition to the expected peptide (see Fig. 1).
We studied the ability of two other proline-dependent kinases to catalyze phosphorylation of MHC 150 . Both MAP kinase (data not shown) and Cdc2 kinase were capable of phosphorylating the heavy chain of HMM i , but not HMM n . It was also of interest to see if tissue-purified bovine brain myosin, which had previously been shown to contain both MHC II-B and MHC II-B1 (14), could be phosphorylated. Fig. 6 shows the result of an in vitro phosphorylation assay using bovine brain myosin as well as HMM i as substrate. Panel A is a Coomassie Blue-stained SDS-10% polyacrylamide gel, and panel B is the corresponding autoradiogram. The figure shows that purified bovine brain MHC II-B1 is a substrate for Cdc2 kinase, and panel C shows that the site phosphorylated is identical to that found for HMM i . In addition to proline-directed kinases, we assayed a number of other kinases to see if they could catalyze phosphorylation of HMM i . Neither cAMP-dependent protein kinase, CaM kinase II, nor protein kinase C could catalyze phosphorylation of the heavy chain of this isoform (data not shown).
Extent and Effect of Phosphorylation-The various prolinedirected kinases were capable of incorporating between 0.1 and 0.4 mol of PO 4 /mol of MHC in vitro. Cdc2 kinase gave the highest incorporation (0.3-0.4 mol of PO 4 /mol) and, therefore, this enzyme was used to phosphorylate HMM i and determine whether phosphorylation has an effect on the actin-activated MgATPase activity. In two separate experiments (see Table II), phosphorylation of HMM i had only a slight, if any, effect on the V max of the actin-activated MgATPase activity. As a control, we subjected HMM n to the exact same assay conditions and measured enzymatic activity before and after addition of Cdc2 kinase. In one experiment where the stoichiometry of MHC phosphorylation was 0.4 mol of PO 4 /mol of MHC, no significant effect of this phosphorylation on the velocity of actin sliding was found.

DISCUSSION
The pre-mRNA encoding MHC II-B, but not MHC II-A, is subject to alternative splicing to produce a number of different isoforms that are only expressed in avian and mammalian neuronal cells. The function of the isoforms produced by this splicing is not known and they appear to be confined to cells that are part of the central nervous system. In chicken brain, the B1 isoform, detected by quantitative polymerase chain reaction, is already present by embryonic day 4 (the first day analyzed) and it reaches a peak on embryonic day 10 (14). Using S-1 nuclease analysis, Takahashi Fig. 4. Lane 3 shows the focusing of a synthetic phosphopeptide, DHNIPPESPKPVK, that was phosphorylated using the same kinase. This peptide corresponds to the sequence of amino acids 207-219 in HMM i (see Fig. 1). Lane 4 shows the tryptic phosphopeptides of the MLC from the same sample. The identification of the two light chain phosphopeptides as Ser-1 and Ser-1Ј was based on standard phosphopeptides not shown (44) and is due to partial cleavage by trypsin. The identification of the MHC peptide as containing Ser-214 is based on the sequence of MHC II-B1 (13). The peptide in lanes 1 and 2, migrating nearer to the negative pole, is most likely due to partial cleavage of the Arg-Lys sequence at the amino terminus of the tryptic peptide (see Fig. 1).
FIG. 6. In vitro phosphorylation of bovine brain myosin and HMM i by Cdc2 kinase. Purified bovine brain myosin and HMM i were phosphorylated using Cdc2 kinase. A, Coomassie Blue-stained SDS-12.5% polyacrylamide gel showing the incubation mixture of bovine brain myosin in the absence of kinase (ϪK), kinase alone, bovine brain myosin ϩ Cdc2 kinase (ϩK), HMM i Ϫ kinase (ϪK), and HMM i ϩ kinase (ϩK). B, autoradiogram of A showing phosphorylation of MHC from brain myosin and HMM i . There is also autophosphorylation of Cdc2 kinase and phosphorylation of the MLCs. C, isoelectric focusing gel. The bovine brain MHC and HMM i MHC from the gel in panel A were excised from the gel and digested with trypsin. The tryptic peptides were analyzed by an isoelectric focusing gel, and the major phosphopeptide was identified as containing Ser-214 based on its comigration with standards (see Fig. 5). able to detect small amounts of mRNA encoding MHC II-B1 in the adult chicken brain. On the other hand, human cerebral cortex and retina have been shown to be highly enriched for mRNA encoding MHC II-B1 and have been shown to express the 10-amino acid inserted peptide (14). Kelley et al. (18) found that smooth muscle myosin containing an insert of seven amino acids (QGPSFSY) in the exact same place in loop 1 as that described for the nonmuscle MHC II-B1, translocates actin filaments 2.5 times faster than does smooth muscle myosin, which does not contain the inserted amino acids. Uyeda et al. (38) produced chimeric Dictyostelium myosins in which the amino acid sequence in the loop 2 region was exchanged for the homologous sequence from other types of myosins. These substitutions were found to modulate the actinactivated MgATPase activity of the chimeric myosins in a manner roughly proportional to the rate of the myosin from which the loop was derived, but did not have a similar effect on the rate of in vitro motility. Based on these two biochemical studies and the location of the two inserts in the crystal structure of myosin, Spudich suggested that the sequence in loop 1, which is near the ATP binding pocket, might have a profound effect on the translocation of actin filaments by myosin in the in vitro motility assay, whereas the sequence in loop 2 near an actin binding site may affect the actin-activated MgATPase activity (11).
We found no major effect on either of these two activities when comparing side by side preparations of baculovirus expressed truncated MHC II-B isoforms that either contained or did not contain the inserted sequence in loop 1. Although both Fig. 3 and Table I show higher values for both the in vitro motility assay and the actin-activated MgATPase activity, these increases are, at best, modest. This suggests that the presence of this insert in neuronal cell myosin may have other functional consequences rather than to alter these two parameters of myosin activity or that this is a subtle modulatory mechanism.
The presence of a consensus sequence for proline-directed kinases in the inserted residues raised the possibility that this MHC might serve as a substrate for a number of kinases, including Cdc2, Cdk5, and MAP kinase. Previous work with Xenopus MHC II-B, which contains a similar, although 6 amino acids longer, inserted region at loop 1, has shown that the insert can be phosphorylated by Cdc2 kinase, but not MAP kinase (21). In this paper, we show that a number of prolinedirected kinases can phosphorylate the 10-amino acid insert in the chicken nonmuscle MHC. This inability of Xenopus MHC-IIB to be phosphorylated by MAP kinase may reflect the difference in sequence (TESPK versus PESPK) between the species (see Fig. 1) and might also be related to the 6 extra amino acids present in the Xenopus insert. We also found that Cdk5 kinase could phosphorylate the MLC 20 at the same site phosphorylated by protein kinase C, in agreement with the observation of Satterwhite et al. (39) using Cdc2 kinase. The extent of MLC phosphorylation was considerably less than that of the MHC with Cdk5, an enzyme that appears to be only active in neuronal tissue (37).
Despite multiple additions of kinase, only 30 -40% of the MHC was phosphorylated. This did not appear to be due to prior phosphorylation of the MHC since 32 PO 4 -labeling of the Sf9 cells just prior to harvesting showed no evidence for labeling of MHC exp following SDS-polyacrylamide gel electrophoresis of a lysed cell extract (data not shown). Failure to obtain more than 40% phosphorylation of Ser-214 in vitro may mean that we have not yet identified the relevant proline-directed kinase, the right conditions for phosphorylation or both. In any case, the partial phosphorylation of the MHC that we observed had no significant effect on the actin-activated MgATPase activity.
What then could be the role of the 10-amino acid insert? Previous work has shown that splicing of the mRNA to introduce the insert is responsive to certain signal transduction pathways. For example, mRNA encoding MHC II-B can be induced to splice in the 30 nucleotides encoding MHC II-B1 by treating rat PC-12 cells with nerve growth factor, but not epithelial growth factor. The cells then cease to divide and initiate neurite outgrowth (14). Since MHC II-B is the major isoform (perhaps the only isoform since the amount of MHC II-A in brain is small and may be expressed only in non neuronal cells) present in neuronal cells, it is conceivable that splicing in the insert acts as a localization mechanism, permitting the myosin II-B1 to be bound in a particular part of the cell. Phosphorylation of the myosin in the inserted sequence might then act to regulate this association. Studies to explore this possibility are presently under way.
Previous investigators have used the baculovirus expression system to express other HMMs, including smooth muscle HMM (40 -42) and cardiac HMM (43). Of particular interest have been studies using site-directed mutagenesis to understand the mechanism of smooth muscle myosin regulation (40). To our knowledge, this is the first report on expression of a vertebrate nonmuscle myosin. The ability to express milligram quantities of an enzymatically active myosin fragment and to introduce discrete mutations is proving to be a powerful technique in understanding all forms of myosins. Our study has shown that the system is also a valuable technique for studying different functions of closely related isoforms of myosin that would prove extremely difficult to obtain in pure form.