Identification and Characterization of Nonmuscle Myosin II-C, a New Member of the Myosin II Family*

A previously unrecognized nonmuscle myosin II heavy chain (NMHC II), which constitutes a distinct branch of the nonmuscle/smooth muscle myosin II family, has recently been revealed in genome data bases. We characterized the biochemical properties and expression patterns of this myosin. Using nucleotide probes and affinity-purified antibodies, we found that the distribution of NMHC II-C mRNA and protein (MYH14) is widespread in human and mouse organs but is quantitatively and qualitatively distinct from NMHC II-A and II-B. In contrast to NMHC II-A and II-B, the mRNA level in human fetal tissues is substantially lower than in adult tissues. Immunofluorescence microscopy showed distinct patterns of expression for all three NMHC isoforms. NMHC II-C contains an alternatively spliced exon of 24 nucleotides in loop I at a location analogous to where a spliced exon appears in NMHC II-B and in the smooth muscle myosin heavy chain. However, unlike neuron-specific expression of the NMHC II-B insert, the NMHC II-C inserted isoform has widespread tissue distribution. Baculovirus expression of noninserted and inserted NMHC II-C heavy meromyosin (HMM II-C/HMM II-C1) resulted in significant quantities of expressed protein (mg of protein) for HMM II-C1 but not for HMM II-C. Functional characterization of HMM II-C1 by actin-activated MgATPase activity demonstrated a Vmax of 0.55 + 0.18 s–1, which was half-maximally activated at an actin concentration of 16.5 + 7.2 μm. HMM II-C1 translocated actin filaments at a rate of 0.05 + 0.011 μm/s in the absence of tropomyosin and at 0.072 + 0.019 μm/s in the presence of tropomyosin in an in vitro motility assay.

Myosins are protein molecular motors that bind to filamentous actin in an ATP-dependent manner. They are expressed as numerous classes in all eukaryotic cells (1). Nonmuscle myosins include the ubiquitous conventional myosins (class II), which form filaments at relatively low ionic strength and share a number of biological properties with skeletal, cardiac and smooth muscle myosin (1). They are expressed in both muscle and nonmuscle cells and are hexamers, consisting of a pair of heavy chains (200 kDa) and two pairs of light chains (20 and 17 kDa). Nonmuscle myosin II is one of the main motors interacting with cytoskeletal actin and is involved in regulating cytokinesis, cell motility, and cell polarity in many eukaryotic cells (1,2).
Two different nonmuscle myosin IIs have been described to date that are referred to as myosin II-A and II-B (3)(4)(5)(6)(7)(8)(9). These names are based on their unique heavy chain polypeptides, which are encoded by two different genes, located in humans on chromosomes 22 and 17 and in mice on chromosomes 15 and 11, respectively. It is presently not known whether nonmuscle myosin heavy chain (NMHC) 1 II-A and II-B share the same or different light chain isoforms. There is one known form of NMHC II-A, whereas four alternatively spliced forms of NMHC II-B have been described (1,8,9).
The localization of NMHCs II-A and II-B has been described for a number of different cultured cells and tissues (10 -14), but no consistent pattern has emerged. Depending on the cell type, there appear to be areas that contain a particular myosin II isoform, but there are also areas where they clearly overlap at the light microscopic level. Some cells appear to express only one isoform, and a number of different approaches have been utilized to lower or remove one of the isoforms from both cultured cells (15,16) and mice (17)(18)(19). Specific interactions between nonmuscle myosin II-A or II-B isoforms with a variety of proteins have been reported. These include the interaction of nonmuscle myosin II-B with the PBX family of homeodomain transcription factors (20) and the interaction of nonmuscle myosin II-A with the tumor suppressor protein menin (21), the CXCR4 chemokine receptor (22), and the S100 protein mts1 (23).
Analysis of genomic and cDNA data bases raised the possibility of the existence of previously unrecognized myosins, including myosin IIs (24). Expressed sequence tags (ESTs) have occasionally been annotated as "moderately similar to nonmuscle myosin II-B" (or "II-A"), and a potential NMHC gene has been noted on human chromosome 19 (accession number AC020906) (24). Recently, Leal et al. (25) also found that this gene, MYH14, is transcribed into RNA and that this RNA is particularly abundant in human intestine and skeletal muscle.
The purposes of the present study were: 1) to verify that the putative gene encoding a third NMHC, which we refer to as II-C, is translated into a previously unrecognized protein; 2) to determine what differentiates this protein from other nonmuscle myosin IIs; 3) to examine whether it is expressed in different alternatively spliced forms; and 4) to biochemically characterize its properties as a myosin.

EXPERIMENTAL PROCEDURES
Identification of cDNA Sequence-BLAST searches of EST and genomic data bases were used to search for new nonmuscle myosins. The LEADS TM platform for clustering and assembly of DNAs and ESTs (www.labonweb.com) was used to cluster and assemble ESTs and for the search for alternative splices. This analysis led to the identification of human and mouse RNAs and ESTs encoding a previously unrecognized NMHC that we termed II-C.
Sequences used to identify NMHC II-C included the human genomic sequence corresponding to nucleotides 63986279 -64255191 of chromosome 19 (based on a BLAT search at genome.ucsc.edu/cgi-bin/hgBlat) and to accession numbers AC020906, AC010515, and AC019157 and the mouse genomic sequences AC073782 and AC073806, as well as different ESTs and annotated RNA sequences clustered in Unigene clusters Hs.115412/Hs.250389 and Mm.7124 (mouse chromosome 7). Comparison of the human ESTs BG829206 and BG468611 indicated that there are two alternatively spliced forms of NMHC II-C, the longer of which contains a cassette insert of 24 nucleotides derived from an additional exon. Electronic gene expression data were retrieved from the websites www.ncbi.nlm.nih.gov/SAGE and bodymap.ims.u-tokyo. ac.jp.
RNA Isolation and RT-PCR-Total RNA was extracted from mouse tissues using Trizol B Reagent (Tel-Test Inc., Friendswood, TX). Poly(A ϩ ) RNA was isolated from total RNA using the NucleoTrap mRNA purification kit (BD Clontech, Palo Alto, CA). Reverse transcription was carried out in a final volume of 25 l using 2 g of total RNA and 2.5 units of Superscript II reverse transcriptase (Invitrogen) in the buffer supplied by the manufacturer, in the presence of 10 pmol of oligo(dT) 15 and 10 units of RNasin (Promega). PCR was carried out using 2 l of the RT reaction in the presence of 2 mM dNTPs, 10 pmol of primers, and 2.5 units of DNA polymerase (Expand high fidelity PCR System; Roche Applied Science). The PCR primers used to verify the alternative splicing of NMHC II-C in human skeletal muscle were 5Ј-ATGCTGCAGGATCGTGAGGACC-3Ј (forward) and 5Ј-ATGAATTT-GCCGAATCGGGAGG-3Ј (reverse).
RNA Blot Analysis-Multiple Tissue Northern blots and Multiple Tissue Expression arrays were used (MTN 12 lane and MTE-2; Clontech, Palo Alto, CA), containing poly(A ϩ ) RNA samples from different human and mouse tissues. A PCR product derived principally from the 3Ј-untranslated region of NMHC II-C (nucleotides 5965-6395 from the first ATG; see Fig. 1) was cloned, sequenced, and used as a specific probe. The probe was labeled using the Random Primer DNA labeling mix (Sigma) and [␣-32 P]dCTP (PerkinElmer Life Sciences). Hybridization was carried out in the ExpressHyb-Hybridization Solution (Clontech) at 68°C for 18 h. The membranes were rinsed twice with 2ϫ SSC, 0.1% SDS at room temperature, followed by two washes with 0.1ϫ SSC, 0.1% SDS at 50°C and exposed to a PhosphorImager screen (Molecular Dynamics, Sunnyvale, CA) and to x-ray films. The results were confirmed using additional blots, with probes corresponding to nucleotides 1-759 of the noninserted NMHC II-C sequence.
Analysis of Hybridization Results from Human Multiple Tissue Arrays-To enable a proper comparison of the general pattern of tissue distribution among the different NMHC mRNAs, PhosphorImager signal intensities were transformed in the following manner: 1) The NMHC hybridization signal intensity in each spot was divided by the ubiquitin hybridization signal intensity. 2) The data were normalized, so that the value of 100 defines the median of the signal intensities of the following organs: brain, heart, skeletal muscle, colon, thymus, spleen, kidney, liver, placenta, uterus, lung, and peripheral blood lymphocytes. 3) Each hybridization was repeated twice, with different multiple tissue array membranes and different probes.
Determination of the Ratio of Inserted and Noninserted mRNA Encoding Mouse NMHC II-C-RT-PCR was carried out using the Gene-Amp RNA PCR core kit (Applied Biosystems). The primer sets flanking the inserted region were as follows: forward primer, 5Ј-GCCCATGTG-GCATCATCTCCA-3Ј; and reverse primer, 5Ј-CTCCCACGATGTAGC-CAGCA-3Ј. Total RNA from the various tissues of 6-week-old mice was reverse transcribed using random hexamers and Moloney murine leukemia virus reverse transcriptase. The resulting cDNA was amplified by PCR using the above primers. The reaction profile included 35 cycles total of denaturation at 94°C for 1 min and annealing and extension at 65°C for 1 min for the first four cycles followed by denaturation at 94°C for 30 s and annealing and extension at 60°C for 45 s for the remaining 31 cycles. RNA samples subjected to PCR amplification in the absence of reverse transcriptase were used as a negative control for contamination by any DNA. The PCR product from liver was subcloned into the pCR2.1 TOPO vector for sequencing using the TOPO-TA cloning kit (Invitrogen), which confirmed the presence of the inserted isoform. Products generated by RT-PCR were analyzed on a 2% agarose gel, and the presence of the inserted isoform in the RT-PCR products was confirmed by Southern blotting using an oligonucleotide corresponding to the 24-nucleotide inserted sequence as a probe.
Preparation and Characterization of Antibodies-Affinity-purified rabbit polyclonal antibodies against specific NMHC II-C peptide sequences were generated. The sequences used for antibody generation and for affinity purification using an Affi-Gel 15 column (Bio-Rad) were: mouse N-terminal sequence, VTMSVSGRKVASRPGP (amino acid 4 -19; see Fig. 1); mouse C-terminal sequence, APGQEPEAPPPATPQ (amino acids 1986 -2000); and human C-terminal sequence, RQVFR-LEEGVASDEEAEE (equivalent to amino acids 1962-1979 in mouse). The last of these antibodies appeared to cross-react with a non-myosin filamentous protein in cultured peripheral neurons when used for immunofluorescence studies. 2 The mouse sequence antibodies were raised in rabbits by Robert Wysolmerski (St. Louis University School of Medicine, St. Louis, MO) and affinity-purified by us. Antibodies to NMHC II-A and II-B have been described previously (7,10,13).
Immunoblot Analysis-Tissue extracts from mouse organs were prepared using a buffer of 80 mM MOPS (pH 7.4), 60 mM KCl, 10 mM MgCl 2 , 5 mM ATP, 4 mM EDTA, 1 mM dithiothreitol, 1% Nonidet P-40, and protease inhibitors (chymostatin, 1-chloro-3-tosylamido-7-amino-2-heptanone, L-(tosylamido-2-phenyl)-ethyl chloromethyl ketone, leupeptin, phenylmethylsulfonyl fluoride, aprotinin, benzamidine, and pepstatin A), at 4°C. The lysates were sedimented at 10,000 ϫ g for 10 min, and the supernatant was used. In samples subjected to immunoprecipitation, the nonmuscle myosins were first purified by actin selection (see below). For immunoprecipitation, a 1:20 (v:v) ratio of affinity-purified antibody to the actin-selected sample was used as described previously (11). The supernates and relevant immunoprecipitates were fractionated by SDS-PAGE on 6% Tris-glycine gels or 8 -16% gradient Trisglycine gels (Invitrogen), transferred onto Immobilon polyvinylidene difluoride membranes (Millipore, Bedford, MA), and subjected to immunodetection using the antibodies described above at a concentration of 0.025 g/ml for NMHC II-A, 0.045 g/ml for II-B and 0.13 g/ml for II-C. Antibodies to nonmuscle actin (Sigma) and glyceraldehyde-3-phosphate dehydrogenase (Biodesign, Kennebunk, ME; Ref. 18) were used to assure equal loading of samples. For immunodetection of baculovirusexpressed, mouse FLAG-tagged, heavy meromyosin (HMM) II-C and HMM II-C1, both monoclonal anti-FLAG antibody (Sigma) and the polyclonal antibody raised to a sequence near the N terminus of the mouse NMHC II-C were used (see above). Peroxidase-conjugated goat anti-rabbit or anti-mouse IgGs were used as secondary antibodies. The proteins were visualized with the SuperSignal system (Pierce). The protein concentrations were determined using a Bio-Rad protein assay kit.
Immunofluorescence Microscopy-Distribution of NMHC II isoforms in developing mouse embryos was visualized by immunofluorescence staining. The embryos were collected in phosphate-buffered saline and then directly immersed in 4% paraformaldehyde overnight at room temperature. For better fixation, the brain and abdominal cavity were partially exposed. The fixed embryos were embedded in paraffin and sectioned at a thickness of 5 m. For antibody staining, the samples were blocked with phosphate-buffered saline containing 0.1% bovine serum albumin and 5% normal goat serum for 1 h at room temperature, incubated with polyclonal antibodies against NMHC II-A (0.25 g/ml), II-B (0.29 g/ml), and II-C (1.3 g/ml) overnight at 4°C, followed by incubation with fluorescein isothiocyanate-conjugated goat anti-rabbit IgG (1:200; Jackson ImmunoResearch) for 1 h at room temperature. After washing, the coverslips were mounted using a Prolong antifade kit (Molecular Probes). The images were collected using a Leica SP confocal microscope (Leica).
Baculovirus Expression of FLAG-tagged HMM II-C and HMM II-C1-Nucleotides 1-4071 of the coding region of HMM II-C, followed by the sequence coding for the FLAG peptide (DYKDDDDK) and an XbaI restriction site at the C terminus, were obtained by RT-PCR reactions using a FLAG-tagged oligonucleotide primer and cloned into baculovirus transfer vector pVL 1392 (BD Pharmingen, San Diego, CA). To generate a clone of the 8-amino acid inserted isoform, we used a PCRderived clone of nucleotides 1-1227 in the TOPO TA vector (Invitrogen) and introduced the 24-nucleotide cassette insert into it by mutagenesis (QuikChange; Stratagene). The clone with the insert together with the 3Ј portion of the HMM II-C was then introduced into pVL1392 and the virus was cotransfected along with a virus containing both light chains 2 P. Bridgman, personal communication.
in Sf9 cells (26). Infected cells were harvested by sedimentation after 72 h of growth, and the pellet was washed twice with phosphatebuffered saline, quick-frozen in liquid nitrogen, and stored at Ϫ80°C. Extraction and purification of HMM was as previously described (27) with the exception that 0.5 M NaCl was used in place of 0.2 M NaCl. The material eluted from the FLAG column was concentrated using a Sepharose Q column or used directly following elution from the FLAG column.
Actin Selection of Tissue Extracts and of Baculovirus-expressed Mouse Heavy Meromyosin II-C-The baculovirus-expressed HMMs were dialyzed overnight in a buffer of 0.5 M NaCl, 10 mM MOPS, pH 7.2, 0.1 mM EGTA, 3 mM NaN 3 , and 1 mM dithiothreitol (buffer A). F-actin (3 M), stabilized with 5 M phalloidin, was added to the expressed HMM, or to lung extracts prepared as outlined above, and the solution was sedimented for 15 min at 470,000 ϫ g to pellet the actomyosin complex. The resulting pellet was solubilized in buffer A containing 5 mM ATP, 5 mM MgCl 2 , and 2 M phalloidin and resedimented to separate the released HMM from the polymerized actin (26).
ATPase Assay-The actin-activated MgATPase activity was measured at 37°C using the NADH-coupled assay in a Beckman DU640 spectrophotometer in a buffer containing 2 mM MgCl 2 , 0.1 mM EGTA, 1 mM ATP, 0.2 mM CaCl 2 , 1 mM calmodulin, and 10 ng/ml myosin light chain kinase as previously described (28).

Identification and Sequencing of Mouse NMHC II-C-Based
on the analysis of genomic and cDNA sequence data in the different gene banks, we predicted the amino acid sequences of the human and mouse NMHC II-C in two alternatively spliced forms. We cloned and verified the sequence of the mouse NMHC II-C using RT-PCR and mRNA from mouse lung and colon. Each PCR segment was generated and sequenced at least three times to verify the accuracy of the sequence. The resulting sequence of the cDNA encoding mouse NMHC II-C is presented in Fig. 1 (GenBank TM accession number AY205605, and AY363100 for the alternatively spliced isoform). The protein sequence derived from this cDNA contains the characteristic domains of a myosin II heavy chain (underlined in Fig. 1 and see legend in Ref. 31 to Fig. 1). The cDNA encodes a gene product of 2000 amino acids with the inserted exon and 1992 without it.
Comparisons of the Sequence of NMHC II-C to That of Other Myosin IIs-Comparative analyses of both the phylogenetic tree of the motor domain of NMHC IIs (Fig. 2 and Refs. 24 and 32), as well as the entire heavy chain sequence (Table I), suggests that NMHC II-C constitutes a distinct branch of the nonmuscle/smooth muscle myosin heavy chain II family. The identity and similarity levels among NMHC II-A (MYH9), NMHC II-B (MYH10), and smooth muscle myosin heavy chain (MYH11) are higher than between any one of them and NMHC II-C. On the other hand, NMHC II-C is more related to this group of myosin heavy chains than to any other myosin heavy chain subfamily (Fig. 2). Furthermore, the percentage of amino acid identity of each myosin isoform between human and FIG. 2. The phylogenetic tree of the myosin heavy chain II family in humans with a scale of amino acid identity. The tree structure is derived from analysis of the motor domain of the myosin heavy chains, as previously described (24,32). Then the overall identity in the amino acid sequences between the myosin IIs was determined. The vertical axis defines the overall identity between myosin heavy chains at each branching point of the tree.  mouse ranges between 93.3% and 98.7%, which is considerably higher than the percentage of amino acid identity between the different isoforms within the same species.
NMHC II-C mRNA Tissue Distribution-RNA blot analysis of human tissue poly(A ϩ ) RNA with either a probe from the 3Ј-untranslated region of human NMHC II-C (Fig. 3) or from the 5Ј portion of the cDNA (data not shown) showed the presence of a single band of ϳ7 kb, with highest expression in skeletal muscle. The gene is also expressed in brain, heart, colon, kidney, liver, small intestine, and lung. No detectable message is found in thymus, spleen, placenta, and lymphocytes (Fig. 3). This result is in agreement with the analysis of the tissue distribution of the human mRNA signal using the relative frequency of specific sequences in 3Ј-directed cDNA libraries (BodyMap website: bodymap.ims.u-tokyo.ac.jp), which also showed that NMHC II-C is most abundant in skeletal muscle (score 4.0) and is absent in thymus, spleen, placenta, and lymphocytes.
Hybridization of a probe from the 3Ј-untranslated region of human NMHC II-C to a human multiple tissue array blot generally agreed with the results in the RNA blot (Fig. 3) and revealed additional differences between the distribution of NMHC II-C and other nonmuscle NMHC mRNAs (Table II). The table shows that NMHC II-C mRNA is highly expressed in the corpus callosum, which is particularly enriched for glial cells. This finding is confirmed by the BodyMap tissue distribution website (bodymap.ims.u-tokyo.ac.jp), which also shows high levels of this message (GS14638) in the corpus callosum (score ϭ 4.1, which is comparable with a score of 4.0 in skeletal muscle; Fig. 3). Table II also shows that expression of NMHC II-C is low in organs composed mainly of smooth muscle, such as the aorta, uterus, and urinary bladder. NMHCs II-A and II-B are relatively abundant in these tissues (Table II and Ref.  7), whereas the mRNA encoding NMHC II-C is barely detectable. Finally, Table II shows that, in general, human fetal tissue contains little or no NMHC II-C mRNA compared with adult tissues (compare the values for fetal tissue in the last  1, 4, and 7). Only NMHC II-C is precipitated by the antibody to II-C (compare lane 9 with lanes 3 and 6). The antibody used for immunoprecipitation was raised to the C-terminal mouse myosin II-C sequence. The other antibodies for immunoblot analysis (II-A and II-B) are described previously (7). Antibodies used for immunoblotting are indicated below the panels. IP, immunoprecipitate.

TABLE II
Levels of NMHC II-A, II-B, and II-C mRNA in various human organs, using the multiple tissue array (Clontech) The data are adjusted to the signals generated by a ubiquitin probe and normalized so that the value of 100 defines the median ratio of the signals of 12 different tissues of each NMHC (brain, heart, skeletal muscle, colon, thymus, spleen, kidney, liver, placenta, uterus, lung, and peripheral blood lymphocytes). column with those found for adult organs). This hybridization experiment was repeated twice, with two different blots, yielding similar results. We used RT-PCR analysis to determine the ratio of the inserted and noninserted mRNA for the 24 nucleotides encoding 8 amino acids in loop 1 of the NMHC. Fig. 4 shows the distribution of the inserted and noninserted mRNA in a variety of mouse tissues. The figure shows that although the inserted mRNA is the predominant one in adult mouse liver, kidney, and testis, brain and lung contain approximately equal amounts of both mRNAs, and skeletal muscle and heart contain very low amounts of the inserted message. Spleen is a negative control, because NMHC II-C is not expressed in this organ. Comparison of the inserted amino acid sequence between mice and humans showed a difference in one amino acid (ASVSTMSY for mice and ASVSTVSY for humans).
NMHC II-C Protein Is Detected by Specific Antibodies-The selectivity of the antibody raised to the mouse C-terminal sequence is demonstrated in Fig. 5. This figure shows an immunoblot of a crude myosin fraction containing all three nonmuscle myosin II isoforms prepared by actin selection from mouse lungs (Fig. 5, lanes 1, 4, and 7; see "Experimental Procedures" and Ref. 26). Only NMHC II-C is immunoprecipitated using the antibody raised to the II-C peptide as shown by immunoblotting the immunoprecipitate with antibodies to II-A, II-B, and II-C (Fig. 5, lanes 3, 6, and 9). The absence of NMHC II-A and II-B from the immunoprecipitate also shows that NMHC II-C is a homodimer with respect to the myosin heavy chains rather than a heterodimer of II-C with either II-A or II-B. Fig. 6 is an immunoblot comparing the protein expression of NMHC II-A, II-B, and II-C. Note that two separate blots are included for each panel: one utilizing 10 tissue samples and one with three. Equal amounts of protein, as determined by Coomassie Blue staining and detection of nonmuscle actin (left panels) and glyceraldehyde-3-phosphate dehydrogenase by antibodies (right panels), were used to assure equal protein loading (data not shown). The blots show that NMHC II-C differs from II-A in being low or nondetectable in seminal vesicles, liver, bladder, and spleen. The presence of NMHC II-C in stomach and colon and its absence in bladder distinguishes its expression pattern from that of II-B. The antibody raised to the human NMHC II-C C-terminal sequence was used for this blot; the antibody to the mouse C-terminal sequence gave similar results.
Differences in the Localization of the Three NMHCs-Immunofluorescence microscopy was used to examine the distribution of NMHC II-C, II-B, and II-A in mouse embryos at E11.5 and E16.5. Fig. 7 (A, E, and I) shows the widespread distribution of all three isoforms in the mouse embryo. (The apparent concentration of II-A and II-C in the liver at E11.5 most likely is due to autofluorescence secondary to the presence of blood cells.) Panels B, F, and J of Fig. 7 are enlargements of the mouse brains from the previous sections as indicated. Note the enhanced staining for II-C in the developing pituitary at E11.5 (Fig. 7B), for II-B at the pial and ventricular surfaces of the brain (Fig. 7F), and for II-A in the vasculature (Fig. 7J).
Panels C, G, and K of Fig. 7 show differences in the localization of the isoforms in the inner ear. Fig. 7C shows intense staining for II-C in the developing sensory area of the cochlea, whereas II-B is more widely expressed in the epithelial cells and the surrounding mesenchymal cells (Fig. 7G). Again the staining for II-A is most intense in the vessels, although clearly the isoform is also present in both the epithelial and mesenchymal cells (Fig. 7K). Panels D, H, and L of Fig. 7 show that II-C staining is most intense in the apical area of the epithelial cells in the E16.5 mouse intestine (Fig. 7D), and, although staining for II-A is also relatively intense in these cells, it is not localized to the apical part of the cell, as is II-C (7L). Fig. 7H shows that II-B stains most intensely in the serosal cells surrounding the intestine but not in the epithelial cells.
Expression of Recombinant FLAG-tagged HMM II-C and HMM II-C1-We used the Baculovirus expression system to generate both alternatively spliced forms of FLAG-tagged heavy meromyosin II-C along with the 20-and 17-kDa myosin light chains. Both HMM II-C and the 8-amino acid inserted isoform HMM II-C1 bound to skeletal muscle actin and detached from it in the presence of MgATP (shown for HMM II-C in Fig. 8A, lane 4). The figure shows a Coomassie Blue-stained gel depicting purified HMM II-C (Fig. 8A, lane 4) and purified HMM II-C1 (Fig. 8B, lanes 3-8). In this case, the HMM II-C1 was used for characterization studies directly after the FLAG column, although in some cases, it was subjected to actin selection prior to characterization. The finding that both HMM II-C and II-C1 were capable of binding to actin in the absence of MgATP and were released in the presence of MgATP demonstrates that both isoforms share a basic biological property of myosins. However, despite repeated efforts, we were only able to obtain sufficient quantities of the inserted isoform, HMM II-C1, for characterization of the actin-activated MgATPase FIG. 6. Immunoblot analysis of the distribution of NMHCs II-C, II-A, and II-B in adult mouse organs. Protein extracts from different adult mouse tissues were electrophoresed in SDS Tris-glycine gels, transferred to Immobilon polyvinylidene difluoride membranes, and probed with specific antibodies against NMHC II-C, II-A, or II-B as indicated. Antibodies raised to the C-terminal sequences of the three NMHCs were used. Note the relative absence of II-C in liver and spleen in contrast to II-A and its presence in colon and stomach in contrast to II-B. MW indicates molecular mass markers. activity and in vitro motility assay. These values are shown in Table III and are compared with those obtained previously for NMHC II-B and II-A HMM. The table shows that HMM II-C1 has a V max between that of II-B and II-A and translocates actin filaments at approximately the same rate as HMM II-B, considerably more slowly than HMM II-A. DISCUSSION In this study, we present evidence for a previously unrecognized myosin II that is expressed at the protein level in both muscle and nonmuscle tissues. We refer to the new myosin heavy chain as NMHC II-C. We present here the nucleotide and derived amino acid sequence of two alternatively spliced forms of NMHC II-C, which contains all the characteristic domains of conventional myosin heavy chains. Using specific antibodies, we provide evidence that this sequence is expressed as a protein.
Recently, Leal et al. (25) have studied the region of the human chromosome 19q13.3 and have reported the genomic structure and mRNA distribution of NMHC II-C, as evidenced FIG. 7. Detection of NMHC II isoforms in mouse tissues during development. Sections of paraformaldehydefixed mouse tissues were probed with antibodies to NMHC II-C (A-D), II-B (E-H), and II-A (I-L). A, E, and I are saggital sections from an E11.5 mouse and show that all three proteins are widely distributed throughout the embryo. B, F, and J are enlarged areas of the brain (indicated in A, E, and I) and show increased staining for II-C in the pituitary (Pi) at E11.5, intense staining at the pial and ventricular surfaces for II-B, and enhanced vascular staining for II-A. C, G, and K show staining of the mouse inner ear at E16.5. Whereas II-C is particularly intense in the developing sensory cells of the cochlea, II-B is expressed in both the mesenchysmal and epithelial cells, whereas II-A staining is most intense in the vasculature. D, H, and L show staining of the mouse small intestine at E16.5. Both II-C and II-A are intensely stained in the epithelial cells, but II-C is particularly concentrated at the apical border (Ap) of these cells. In contrast, II-B appears more intense in the surrounding serosal cells. by an RNA blot. Their computer data analysis overlooked the alternative spliced variant NMHC II-C1 described above but suggested the existence of shorter alternative spliced forms of this myosin lacking major components of the motor domain. In our analysis, we found similar data but could not confirm it by cloning. Our RNA distribution analysis suggests that NMHC II-C is more widely distributed than the presentation by Leal et al. (25). Fig. 3 provides a comparison of the results of two different approaches to determine the level of expression of NMHC II-C mRNA: the classical approach of an RNA blot and the emerging approach of estimating the level of certain mRNAs according to the number of specific gene tags in a 3Ј-directed cDNA library. Information about gene expression patterns using the latter approach is available from different websites for many genes, including ones that are not yet identified and characterized. In both approaches, we found that NMHC II-C is abundantly expressed in human skeletal muscle and absent in spleen, thymus, placenta, and lymphocytes. With the exception of the expression level in the lung, the ranking of the expression levels in both approaches is similar, although the ratios among them are different. Given the progress in cDNA data accumulation, the information about gene expression is becoming more accurate, and we show that one can gain valuable information about patterns of gene expression from these data bases.
We further examined the patterns of gene expression of NMHCs in human tissues at the RNA level using multiple tissue arrays and developed a simple approach to compare patterns of expression of the different NMHCs (Table II). Because probes of different genes vary in the efficiency of hybridization, the absolute levels of mRNAs cannot be compared. However, the tissue distribution of mRNA levels around the mean or median can serve as a tool for comparison among different genes in a quantitative manner. Of note is that the data shown in Table II for NMHC II-A and II-B are in agreement with that previously reported based on standard techniques (3)(4)(5)(6)(7).
The tissue distribution of NMHC II-C is qualitatively different from other myosin II heavy chains in several respects and suggests that it has a different role from other cytoskeletal myosins. First, NMHC II-C is more abundant in human adult tissues than in human fetal tissues. Unlike NMHC II-A and II-B, it is not expressed in mouse embryonic stem cells. 3 Furthermore, Buxton et al. (33) found that the differentiation stimulus of histone hyperacetylation induces the expression of NMHC II-C, which further supports the idea that myosin II-C is more abundant in differentiated tissues. In addition, NMHC II-C is abundant in specific human brain areas (corpus callosum and pons; Table II) that are rich in glial cells and fiber tracts compared with neuronal cell bodies. On the other hand, the levels of NMHC II-C are very low in organs with abundant smooth muscle, such as uterus, urinary bladder, and aorta, distinguishing it from other nonmuscle myosin IIs.
Nonmuscle myosin II-C behaves differently from most housekeeping genes including NMHC II-A and II-B as shown by its low expression in early development and the fact that, in some instances, it is an inducible protein (33). Other myosins that are known to be induced are cardiac myosins. Two cardiac myosin heavy chains have been described in the heart: ␣and ␤-myosin heavy chains. Their composition is affected by various physiological stimuli, such as pressure overload and endocrine, paracrine, and autocrine stimuli (34 -36). Significant alterations in cardiac function are associated with isoform shifts, evidenced by changes in the maximum unloaded shortening velocity and ATP consumption by the contractile machinery (34). This is consistent with the different mechanical and enzymatic properties of the two isoforms (1,37,38).
The kinetic mechanisms for NMHC II-A and NMHC II-B are significantly different and suggest that these two myosins probably do not have redundant functions (28,39,40). NMHC II-A behaves as a normal but slow conventional myosin with a low duty cycle, meaning that it spends most of its kinetic cycle detached from actin (39). In contrast, NMHC II-B has a duty cycle intermediate between that of conventional and processive unconventional myosins such as myosin V (28,40). This behavior suggests that NMHC II-B may be better adapted for maintaining tension in a static manner. The kinetic mechanism for NMHC II-C remains to be examined. As expected, both the in vitro motility and the actin-activated MgATPase activity of HMM II-C1 were dependent on phosphorylation of the 20-kDa myosin light chain.
Surprisingly, there was a big difference in the yield of the baculovirus expression of the two alternatively spliced forms of HMM II-C. Although we produced the inserted form by splicing in the insert to a clone of the noninserted form and confirming the sequence afterward, the longer form consistently yielded significantly more protein product, which enabled a more thorough biochemical characterization. The effect of small changes of nucleotide sequence on expression level has been observed before (29).
Unlike cardiac myosins, which form both homodimers (known as V1 and V3) and heterodimers (known as V2) (34), nonmuscle myosins form only homodimers (11). This is exemplified for NMHC II-C since, following immunoprecipitation of NMHC II-C, no NMHC II-A or II-B was found in the precipitate (Fig. 5).
The term "nonmuscle myosin" was generally used to distinguish between the ubiquitous forms of myosin II and the muscle-specific isoforms. However, this term is misleading because these myosins are not only expressed in all types of nonmuscle tissue but also have a significant role in the development and function of muscle tissues such as the heart and smooth muscles (13,17,41,42). Fallon and Nachmias (43) described this inadequacy in myosin nomenclature as early as 1980 and suggested the term "cytoplasmic myosin." However, this term would preclude the possibility that some of these myosins could also localize to the nucleus and does not differentiate these myosins from sarcomeric myosins. Because these myosins are likely to be involved in determining cell shape and polarity and interact with cytoskeletal actin, we suggest the use of the term cytoskeletal myosin II. Still, no name, at present, is without some problems and, until a consensus is reached, we shall continue to use the misnomer "nonmuscle myosin." Using antibodies to all three nonmuscle myosins, we compared the distribution of the various isoforms at E11.5 and E16.5 in mice using immunofluorescence microscopy. Of particular note at E11.5 is the prominence of NMHC II-C in mouse 3 M. A. Conti, unpublished observation. pituitary, the enrichment of NMHC II-B in the brain, particularly at the ventricular and pial cortical areas and the prominence of NMHC II-A in the vasculature of the brain (Fig. 7, B, F, and J). Analysis of the inner ear and epithelial cells of the intestines also showed major differences in the distribution of the three isoforms but clearly indicate cells where the distribution also overlaps ( Fig. 7, C, G, and K; D, H, and L). Indeed, very few cells are actually devoid of any one isoform. For example, neuronal cells do contain some NMHC II-A ( Fig. 7J and Ref. 44). Importantly, however, the presence of both NMHC II-A and II-C in neuronal cells does not compensate for the loss of II-B following ablation of this isoform (17,19) or lowering NMHC II-B to less than 20% of the normal amount (18). Moreover, mutation of single amino acids in NMHC II-A results in kidney, platelet, and other defects in humans (45) and is not compensated for by the presence of NMHC II-B and II-C in the same tissues or even the same affected cells. Also, a single amino acid mutation in mouse NMHC II-B results in defects in ventral body wall closure as well as cardiac myocyte, neuronal cell, and brain abnormalities uncompensated by the other myosin II isoforms. 4 It will thus be of interest to learn the consequences of mutation in NMHC II-C in both humans and mice.