The Caenorhabditis elegans nonmuscle myosin genes nmy-1and nmy-2 function as redundant components of the let-502/Rho-binding kinase and mel-11/myosin phosphatase pathway during embryonic morphogenesis

Morphogenesis transforms the C. elegans embryo from a roughly spherical ball into a long, thin worm. In contrast to most other animal systems, C. elegans morphogenesis occurs in the absence of widespread cell growth, proliferation or movements (reviewed by Chin-Sang and Chisholm, 2000; Simske and Hardin, 2001). Embryonic elongation occurs when actin filaments within a subset of epidermal cells (the lateral epidermal or `seam' cells) shorten, causing the cells, and the embryo, to lengthen. Actin is initially disorganized within the lateral epidermal cells; however, at the onset of elongation, actin organizes into filaments that shorten as elongation proceeds(Priess and Hirsh, 1986).

let-502/Rho-binding kinase (ROK) and mel-11/myosin phosphatase targeting subunit (MYPT or myosin-binding subunit MBS) regulate the cell shape changes within the lateral epidermal cells to drive elongation. Based on genetic data coupled with the known functions of their vertebrate homologs in the regulation of smooth muscle contraction and focal adhesions formation (Kaibuchi et al.,1999; Pfitzer,2001; Piekny et al.,2000; Shelton et al.,1999; Somlyo and Somlyo,2000; Wissmann et al.,1997), let-502 and mel-11 probably regulate elongation by altering the activity of myosin. Contraction is triggered by phosphorylation of regulatory myosin light chain (rMLC, mlc-4 in C. elegans) by myosin light chain kinase or integrin-linked kinase(Pfitzer, 2001; Somlyo and Somlyo, 2000). Myosin phosphatase in turn removes the activating phosphate from rMLC,preventing contraction. Activation of ROK by the small GTPase Rho results in an inhibitory phosphorylation of the targeting subunit MYPT of myosin phosphatase, releasing the block to contraction. let-502/ROK and mlc-4/rMLC mutants fail to elongate, while mel-11/MYPT mutants hypercontract.

The Drosophila homologs Drok/ROK and MYPT similarly regulate morphogenetic processes. Nonmuscle MHC zipper and rMLC spaghetti-squash, Drok and Mypt/Dmbs are required for dorsal closure(Mizuno et al., 2002; Tan et al., 2003; Winter et al., 2001; Young et al., 1993). These molecules also are involved in other developmental processes such as epidermal planar polarization (Winter et al.,2001). Therefore, nonmuscle myosin, ROK and MYPT probably comprise a conserved `cassette' that is used for many developmentally-specific contractile processes during animal development.

To identify new components of the C. elegans elongation pathway,we screened for suppressors of the mel-11 hypercontraction phenotype(Piekny et al., 2000). Here,we describe nmy-1, which encodes a nonmuscle myosin heavy chain(MHC). Genetic evidence is consistent with nmy-1 acting as a partner for MLC-4/rMLC, the downstream target for LET-502/ROK and MEL-11/MYPT. However, the nmy-1 null mutant phenotype is less severe than mlc-4 or let-502, suggesting that nmy-1 functions redundantly. Indeed, we found that nmy-2, previously known only for its requirement in the early embryo (Cuenca et al., 2003; Guo and Kemphues, 1996) acts in concert with nmy-1. In addition,NMY-1 antisera reveal that this protein forms filamentous-like structures in the lateral epidermal cells during elongation.

The expression pattern of LET-502 and MEL-11, determined by GFP transcriptional reporters, is consistent with their functioning as regulators of elongation (Wissmann et al.,1999). LET-502, the trigger for contraction, is expressed in the same lateral epidermal cells as MLC-4 that drive elongation(Shelton et al., 1999; Wissmann et al., 1999). MEL-11, the negative regulator of contraction, is also expressed within these cells; however, at lower levels than in the neighboring dorsal and ventral epidermal cells (Wissmann et al.,1999). A caveat is that these reporters revealed only zygotic transcription, although either maternal or zygotic activity of the two genes suffices for viability. Therefore we do not know precisely in which cells the genes act nor the mechanism by which LET-502 and MEL-11 regulate epidermal cell shape changes. For example, they could regulate myosin contraction and/or tether actin cables to the membrane during morphogenesis. Using antisera against LET-502 and MEL-11, we show that both proteins are expressed in the lateral epidermal cells during elongation. Although LET-502 localizes near myosin throughout elongation, MEL-11 becomes sequestered away from the contractile apparatus in response to LET-502 activity. Thus, LET-502 and MEL-11 act at the level of myosin contractility.

C. elegans were maintained under standard conditions(Brenner, 1974). Strains were constructed using standard procedures, and cis-linked morphological markers were often used to follow the mutations used through crosses. Homozygous lethal or sterile mutations were maintained as heterozygous stocks balanced either with appropriate crossover suppressors or normal chromosomes with flanking morphological markers. To determine hatching rates of different genetic combinations, as many L4 hermaphrodites as needed were brooded at the appropriate temperatures until they ceased laying fertilized embryos; a minimum of 300 progeny were scored unless otherwise noted(Mains et al., 1990).

Nomenclature follows that of Horvitz et al.(Horvitz et al., 1979). Genes,alleles and balancer chromosomes listed are below; descriptions can be found in Edgley et al. (Edgley et al.,1995), Hodgkin (Hodgkin,1997) and in Wormbase(www.wormbase.org).

Linkage Group I: let-502(ca201, sb106ts and sb108), dpy-5(e61), rde-2(ne221). Linkage Group II: mel-11(it26ts and sb55), unc-4(e120),sqt-1(sc13). Linkage Group III: mlc-4(or253). Linkage Group V: sma-1(e30). Linkage Group X: unc-115(e2225), dpy-8(e130),unc-78(e1217), spc-1(ra409), nmy-1(sb92, sb113 and sb115), unc-2(e55). Balancer chromosomes: the crossover suppressor mnC1 II was used to balance mel-11.

Single nucleotide polymorphism mapping was performed as described by Wicks et al. (Wicks et al., 2001). The Dumpy (Dpy) and Lumpy phenotype of sb113 was mapped to the left arm of LGX by crossing with the Hawaiian strain CB4856 and using primers indicated in Wicks et al. (Wicks et al., 2001). Another mutant, sb115, was mapped to the left arm of LGX between dpy-8 and unc-115 via classical genetics. sb92 has no phenotype on its own but was mapped between unc-2 and unc-78 based on suppression of mel-11. sb115 dpy-8 was crossed to CB4856, and Lumpy non-Dpy and Dpy non-Lumpy recombinants were selected, mapping sb115 between polymorphisms Y41G9A and F52E4 (data submitted to Wormbase). Cosmid F52B10 is within this region and contains nmy-1. Two independent PCR products spanning nmy-1 in sb92, sb113 and sb115, were sequenced(University of Calgary Core DNA Sequencing Laboratory)(Piekny et al., 2000) and were found to contain mutations (see Results).

Embryos were dissected from gravid hermaphrodites that had been incubated without food for 1 hour. This induces animals to hold their eggs, enriching for mid-stage embryos. Embryos were mounted in M9 solution(Horvitz and Sulston, 1980) or larva and adults were placed in anaesthetizing solution (0.1% tricaine, 0.01%tetramisole) on 3% agarose pads. Specimens were examined by Nomarski optics on a Zeiss Axioplan 2 microscope, and images were photographed using the Hamamatsu Orca ER digital camera and collected using Axiovision 3.0 for Windows NT.

For indirect immunofluorescence, mid-staged embryos were placed on polylysine-coated slides in M9. A cover slip was placed on each slide, and the embryos were frozen on dry ice for at least 30 minutes. The cover slips were cracked off, and slides were placed in –20°C methanol for 10 minutes, followed by 20 minutes in –20°C acetone, then rehydrated with 90% ethanol for 10 minutes at –20°C, 60% ethanol for 10 minutes at –20°C, then 30% ethanol for 10 minutes at room temperature. For actin and some NMY-1 staining, freeze-cracked slides were placed in a 3.7%formaldehyde solution (w/v) in 1× phosphate-buffered saline (PBS) at room temperature for 10 minutes and then placed in 1× PBS with 0.1%Tween-20 (PBT). For all methods, the slides were placed directly into 1×PBT for a minimum of 1 hour, then incubated with appropriate dilutions of antisera in PBT with 20% normal goat or donkey serum (Jackson Immunoresearch Laboratories). F-actin was stained with phalloidin-Alexa 488(Molecular Probes; 10 μl/slide was placed into a tube and the methanol was evaporated overnight, then 1×PBT was added with 20% normal goat serum and other antisera if needed). Rabbit anti-NMY-2 polyclonal antibodies(Guo and Kemphues, 1996) were used at a 1:50 dilution, rabbit anti-MEL-11 polyclonal antibodies at 1:50(Piekny et al., 2002), rat anti-LET-502 polyclonal antibodies at 1:50 (Piekny et al., 2002), rat anti-NMY-1 polyclonal antibodies (see below) at 1:200 and rabbit anti-NMY-1 polyclonal antibodies (see below) at 1:1000. All slides were incubated with primary antibodies for 1 hour to overnight at room temperature and washed three times with PBT. Anti-rat IgG conjugated to indocarbocyanine(Cy-3; Jackson Immunoresearch Laboratories) and anti-rabbit IgG conjugated to Alexa 488 (Molecular Probes) were used at a 1:100 dilution in PBT, and slides were incubated at room temperature for 1 hour. Slides were washed three times with PBT and soaked in 1 μg/ml DAPI (Roche) for 1 minute at room temperature, washed with PBT, then mounted with a drop of Slowfade Light Antifade solution (Molecular Probes). Images were collected using Axiovision 3.0 for Windows NT as stacks of 20×1 μm from a Zeiss Axioplan 2 microscope using a 63× oil objective with the Hamamatsu Orca ER digital camera. Stacks were digitally deconvolved using the constrained iterative algorithm of Axiovision 3.0 for Windows NT. Superficial stacks corresponding to epidermal cells or internal stacks corresponding to the pharyngeal/intestinal region were analyzed with Adobe Photoshop version 4.0 for Windows.

Double-stranded RNA (dsRNA) was generated for nmy-1 and nmy-2. The primers for nmy-1 were made to the C-terminal region of the coding sequence (sequence identity <80% with nmy-2). The primers were: forward primer (including the T3 promoter)5′AATTAACCCTCACTAAAGGGGCAACATCAACTGACGAG3′ and reverse primer(including the T7 promoter)5′TAATACGACTCACTATAGGGAGCATCGAGAAGATCGTC3′. Primers for nmy-2 similarly were made to the C-terminal portion of the coding sequence showing lower sequence homology to nmy-1. The primers were:forward primer (including the T3 promoter)5′AATTAACCCTCACTAAAGGGGACGAGACTCGATGCTGA3′ and reverse primer(including the T7 promoter)5′TAATACGACTCACTATAGGGATCTCTGGAGAGTGTCTC3′. dsRNA was made from the PCR products and used for soaking worms as described in Piekny and Mains(Piekny and Mains, 2002). dsRNA to nmy-1, nmy-2 and both together were injected as described by Fire et al. (Fire et al., 1998)with concentrations varying from 10^-6 μg/μl to 1 μg/μl for each gene. Some nmy-2 RNAi experiments were performed using bacteria expressing dsRNA from the feeding library described by Fraser et al.(Fraser et al., 2000). L4 or young adult were cultured for 1-2 days until first broods were obtained and were passaged to plates seeded with normal E. coli for subsequent broods.

Rat polyclonal antibodies were generated using a GST-NMY-1 fusion encoding C-terminal region amino acids 941-1133, which was amplified using the following primers: forward (containing the BamHI restriction site)5′TACAGGATCCGAGAAACCGTCCGTGATCTC3′ and reverse (containing the EcoRI restriction site)5′CGAGGAATTCCCTTGTCATTTCTGCCTTAT3′ and cloned directionally into the pGEX-3X vector (Pharmacia). Rat injections were performed as described previously (Piekny and Mains,2002). Rabbit polyclonal antibodies were raised against a GST-NMY-1 fusion by the Peptide Group from Macromolecular Resources. All antisera were purified by subtracting against a glutathione-sepharose 4B column (Pharmacia). Western blot analysis showed that all of the antisera recognized a predominant band above 200 kDa (expected ∼230 kDa) with gravid adult hermaphrodite extracts obtained by 1×PBS and with extracts further solubilized with 1 M NaCl, implying that pools of NMY-1 are both cytoplasmic and associated with the cytoskeleton (data not shown). The band was blocked by adding excess GST-NMY-1 to the antisera (data not shown) and immunostaining was not detectable in nmy-1(sb115) (data not shown but see results), which truncates the protein before the region used for immunization.

Previously we performed a screen for suppressors of mel-11 to identify new genes in the let-502/mel-11 pathway(Piekny et al., 2000). sb92, sb113 and sb115 are semi-dominant suppressors(Table 1) that failed to complement for Dpy, Lumpy and Roller (Rol) phenotypes (genetic properties will be described below). sb115 was mapped near cosmid F52B10 (Materials and methods), which contains a single gene, nmy-1. As diagrammed in Fig. 1, nmy-1 is predicted to encode a 1963 amino acid nonmuscle MHC, with an N-terminal SH3-like domain, a myosin motor head followed by a myosin coiled-coil tail. sb92 included a Glu(285)Val mutation, sb113 Gly(647)Glu and sb115 Gln(480)stop. All are in highly conserved residues within the head domain (Fig. 1). The myosin head and tail are highly conserved between NMY-1 and the Drosophila nonmuscle MHC, Zipper, human MHCB and another C. elegans MHC, NMY-2 (Fig. 1; see Fig. S1)(Guo and Kemphues, 1996; Mansfield et al., 1996; Philips et al., 1995; Young et al., 1993). Overall,NMY-1 shares highest amino acid identity (48%) with Drosophila Zipper and human MHCB. NMY-1 shares 45% identity with C. elegans NMY-2,suggesting that they are paralogs.

RT-PCR revealed one predominant SL1-spliced transcript of ∼6000 bp,similar to the predicted length of 6383 nucleotides (data not shown). Sequencing confirmed the predicted intron/exon structure of nmy-1found at Wormbase(www.wormbase.org). In addition, microarray data predicts that nmy-1 is most highly expressed during embryogenesis (Hill et al., 2000).

A previous large-scale RNAi survey reported sterility, lethality and Dpy and uncoordinated (Unc) phenotypes for nmy-1(Kamath et al., 2003). However, short regions of high DNA sequence identity to nmy-2 in the myosin head domain could result in RNAi crossreactivity. We performed nmy-1 gene-specific RNAi (Materials and methods) and observed little embryonic lethality but did see Dpy Lumpy phenotypes identical to those of our nmy-1 alleles sb113 and sb115. Like our nmy-1 alleles, injection of nmy-1 dsRNA suppressed mel-11 lethality (Table 1).

nmy-1 alleles behave as semi-dominant suppressors of mel-11 with stronger suppression when homozygous(Table 1). Although suppressed mel-11; nmy-1 embryos hatch, a large proportion of the early larva with sb113 and sb115 arrested with an elongation phenotype similar to let-502 and mlc-4. For example, 84% of mel-11(it26); nmy-1(sb113) larva that hatched at 25°C displayed L1 arrest without elongation, in comparison to nmy-1(sb113)alone, which displayed 1-4% L1-L3 larva arrest/slow growth(Table 2). All escapers showed the sb113 Dpy, Lumpy phenotype, indicating that mel-11 did not reciprocally suppress the nmy-1 morphological defects. All mel-11(it26); nmy-1(sb115) larva displayed L1 arrest compared with 30% for sb115 alone (data not shown).

Our alleles of nmy-1 are the first reported for this gene. sb92 has no obvious phenotypes on its own, while sb113 and sb115 are Dpy Lumpy Rol with some embryonic lethality(Table 2). All sb113and sb115 L1 larva display bulging at the anterior(Fig. 2A,D), which usually disappears as the animals grew. sb113 had a small percentage (1-4%)of slow growth or larval arrested between L1-L3 while the value was 30% for sb115. sb113 and sb115 adults are Dpy, Lumpy and sometimes Rol, consistent with non-lethal elongation defects, and adult hermaphrodites have low brood sizes (Table 2). Larva and adults have small pharynxes (Fig. 2G,H) and possible intestinal defects as bacteria accumulate in their guts. In addition, the adults show varying degrees of Unc, although whether this is a secondary consequence of the elongation defects remains to be determined. Gonad arms sometimes fail to migrate correctly (e.g. fail to turn in the DV axis), but this is probably a secondary consequence of other elongation defects as the distal tip cell migration is normal in L4 larva(data not shown).

All of the nmy-1 alleles behave as loss-of-function alleles. The phenotypes of sb113 and sb115 are unchanged when heterozygous with each other or a chromosomal deficiency of the region (data not shown). sb92 is more severe in trans to either allele or the deficiency than when homozygous, indicating that it represents a partial loss-of-function. sb115 is likely a null allele based on its phenotypes and molecular lesion (a truncation in the N-terminal myosin head domain). sb113 and sb115 phenotypes were rescued by crossing mutant hermaphrodites to wild-type males, indicating zygotic activity of the genes (Table 2). sb115/sb92 and sb113/sb92 hermaphrodites show more severe phenotypes when the stronger alleles are maternally inherited, indicating some maternal activity, even though the NMY-1 protein is not detected in early embryos (data not shown).

Although NMY-1 is predicted to partner with MLC-4, nmy-1(null) is viable while the mlc-4(null) is lethal(Shelton et al., 1999); all mlc-4 mutant larva arrest at L1 with failed elongation(Fig. 2B). This result implies that nmy-1 is not the only myosin target of the let-502/mel-11 pathway. NMY-2 is a nonmuscle MHC with high levels of sequence similarity to NMY-1 (Fig. 1) and has been characterized for its role in anterior/posterior polarity and cytokinesis using RNAi (Cuenca et al., 2003; Guo and Kemphues, 1996). Because RNAi results in early, lethal defects,the role of nmy-2 in elongation has not been explored.

To determine if nmy-2 could potentially function redundantly with nmy-1, we stained wild-type embryos with NMY-2 antisera to determine if the protein persists through elongation. NMY-2 expression levels decreased with time, but it is presented at low levels in all cells, including epidermal cells, until after the onset of elongation, and this pattern was not altered in nmy-1 mutant (Fig. 3A-I).

To establish redundancy between nmy-1 and nmy-2, we performed nmy-2(RNAi) in wild-type and nmy-1 mutant worms. To avoid the block of nmy-2(RNAi) at the one-cell stage, we used feeding as the means of dsRNA delivery, which gradually eliminates gene function over several days, permitting us to examine nmy-1interactions. The broods on the first day after reaching adulthood displayed 52% embryonic viability (Table 3), indicating that nmy-2 was incompletely eliminated. However, L1 arrest phenotypes were never observed. nmy-2(RNAi);nmy-1(sb113 or sb115) displayed the same level of embryonic lethality as nmy-2(RNAi) alone, but half of the hatching embryos showed L1 arrest similar to mlc-4 (Fig. 2B,E,F; Table 3). Although we normally see low levels of L1-L3 slow growth/arrest in both nmy-1(sb113) and nmy-1(sb115) alone, the L1 arrest phenotype seen in the double mutants was much stronger and distinctly like that seen in mlc-4 (and let-502). The enhancement of nmy-1elongation defects by nmy-2 suggests redundancy. Finally, rde-2(ne221), which shows partial resistance to RNAi(Tabara et al., 1999) and thus may bypass the early block, produced 13% embryos with failed elongation when grown on nmy-2(RNAi) bacteria compared with none when fed normal bacteria (Table 3).

As nmy-1 and let-502 both suppress mel-11, we expected to see enhancement between nmy-1 and let-502(Piekny et al., 2000). let-502(sb108), a weak loss-of-function allele, shows no larval arrest, while approximately one-third of nmy-1(sb115) do so. However, let-502(sb108); nmy-1(sb115) showed 82% L1 arrested with failed elongation (Table 1). Similar results were seen with other nmy-1 and let-502allelic combinations (data not shown). Therefore, loss of nmy-1(which results in a partial loss of total NMY protein) synergizes with a partial loss of let-502, consistent with both acting in the same pathway.

Our genetic results predict NMY-1 expression in the lateral epidermis during elongation. NMY-1 antisera revealed that it is indeed highly expressed in these cells at the onset of elongation(Fig. 4B), similar to that observed by Shelton et al. (Shelton et al., 1999) for MLC-4. NMY-1 was also present at the adherens junctions (AJs) of the developing pharynx (data not shown). As elongation proceeded, NMY-1 became more punctate (Fig. 4D) and organized into filamentous-like structures(Fig. 4F). Embryos of similar stages stained for actin showed a similar pattern, albeit filamentous structures become apparent earlier (Fig. 4A,C,E; embryos were not co-stained because NMY-1 and actin require incompatible fixation methods). Both actin and myosin organize into filaments that run perpendicular to the axis of elongation and are thus potentially arrayed to draw the lateral epidermal cells together in the dorsal/ventral axis, causing embryonic elongation.

The immunolocalization patterns of MEL-11 and LET-502 during morphogenesis is consistent with the hypothesis that they regulate the activity of myosin. Prior to elongation, LET-502 and MEL-11 showed extensive colocalization in the cytoplasm with punctate staining at epidermal cell boundaries(Fig. 5A-C). As elongation proceeded, MEL-11 became enriched and was restricted to cell boundaries while LET-502 remained primarily cytoplasmic(Fig. 5D-F). When elongation was completed, MEL-11 again increased in the cytoplasm, with decreased,punctate localization at cell boundaries, similar to the pre-elongation pattern (Fig. 5J-L). Both proteins were also expressed at the AJs of the developing pharynx and intestine (Fig. 5G-I) and in the germline precursor cells (data not shown).

These data suggest that LET-502, but not MEL-11, is in a region where it can influence contractions of actin/myosin filaments during elongation. Indeed, as elongation proceeded, punctate anti-LET-502 staining interspersed with the NMY-1 filaments (Fig. 6A-D). By contrast, MEL-11 was membrane bound, away from NMY-1(Fig. 6E-H). The vertebrate MEL-11 homolog, MYPT, moves to the plasma membrane after phosphorylation by ROK, where it remains sequestered from actin/myosin to allow contraction(Shin et al., 2002). To determine if MEL-11 was similarly influenced by let-502 activity, we examined MEL-11 localization in let-502 mutants. Indeed, using the strong let-502(ca201) allele, MEL-11 remained mostly cytoplasmic with punctate cell membrane localization (Fig. 7A,B), similar to its distribution wild-type embryos prior to elongation (Fig. 5B). Because let-502(ca201) mutant embryos arrest at the onset of elongation,disrupted staining could arise as a secondary consequence of developmental arrest. Therefore, we examined MEL-11 localization in a partial loss-of-function let-502(sb106) mutant, which does elongate. Although less dramatic compared with the strong allele let-502(ca201), we again found higher levels of cytoplasmic MEL-11 in the in let-502(sb106) embryos compared with wild type(Fig. 7C,D). This suggests that the disappearance of the cytoplasmic MEL-11 pool, which occurs at the onset of elongation and which could potentially inhibit myosin activity, is at least partly dependent on let-502(+) activity. This is consistent with the antagonistic genetic interactions of let-502 and mel-11.

Although the above evidence supports the model that LET-502 and MEL-11 regulate myosin activity to drive elongation, another (not mutually exclusive)model proposes that LET-502 and MEL-11 regulate anchoring of actin cables to the membranes of the lateral epidermal cells. Elongation-defective phenotypes are observed in genes such as sma-1 (β_H-spectrin), spc-1 (α-spectrin), hmr-1 (cadherin) and hmp-1 and hmp-2 (catenins), the products of which are required to tether actin filaments to the membranes of the lateral epidermal cells (Costa et al., 1998; McKeown et al., 1998; Norman and Moerman, 2002). We found that spc-1 and sma-1 are genetically epistatic with mel-11, suggesting that mel-11 hypercontraction requires the anchoring of actin cables by these genes(Table 1). However, if LET-502 is involved in anchoring these cables, then we would expect high levels of LET-502 (a possible positive regulator of actin anchoring) and low levels of MEL-11 (a possible negative regulator of anchoring) at the membrane. However,we saw the exact opposite staining pattern(Fig. 5). As reported by Norman and Moerman (Norman and Moerman,2002), we observed normal actin filaments in let-502 and mlc-4 mutants, as well as in nmy-1(sb113) and mel-11(it26) embryos (data not shown). Our results thus suggest that LET-502 and MEL-11 are not required to anchor actin filaments, but likely regulate the contraction of filaments. However, LET-502 and MEL-11 are present at the AJs of the pharynx and gut, and could be required for some aspect of AJ formation or function in these organs.

Drosophila Zipper/nonmuscle myosin requires Spaghetti squash/rMLC for localization (Edwards and Kiehart,1996; Jordan and Karess,1997; Wheatley et al.,1995). We found that NMY-1 similarly failed to form filamentous-like structures in mlc-4-null mutant embryos, with NMY-1 collapsing into large foci (Fig. 8A).

As myosin is the target of MEL-11, we expected that MEL-11 localization might change in nmy-1 and mlc-4 mutants. However, mlc-4 embryos displayed wild-type patterns of MEL-11(Fig. 8B,C) and F-actin distribution (data not shown) (Norman and Moerman, 2002). We also examined nmy-1(sb113);nmy-2(RNAi) embryos, which lack all MHC and arrest at a similar stage to mlc-4. MEL-11 no longer localized to membranes(Fig. 8D-F). When MHC was partly eliminated using nmy-1(sb113), membrane localization of MEL-11 was frequently decreased and occasionally was absent. An extreme case is shown in Fig. 8G-I, which may correspond to those rare sb113 embryos where elongation failed. These results may indicate that the LET-502 interaction that directs MEL-11 to the membrane likely requires MEL-11 binding to MHC.

Elongation of the C. elegans embryo occurs due to actin-mediated contractions within the lateral epidermal (seam) cells(Priess and Hirsh, 1986). We previously identified a pathway that regulates elongation that includes LET-502/ROK and MEL-11/MYPT (Piekny et al., 2000; Piekny and Mains,2002; Wissmann et al.,1997; Wissmann et al.,1999). MEL-11 prevents contraction by dephosphorylating MLC-4/rMLC. LET-502 in turn inhibits MEL-11, releasing the block to elongation. We describe a new component of the pathway, NMY-1/MHC, identified by mutations that suppress mel-11(–) hypercontraction. Based on genetic and molecular analyses, NMY-1 is an MHC partner for MLC-4/rMLC. Fig. 9 models how NMY-1,LET-502 and MEL-11 regulate elongation. At the onset of morphogenesis, actin and NMY-1 each form filamentous-like structures that probably constitute the contractile apparatus that drives elongation(Fig. 4). Similar to MYPT during smooth muscle contraction (Shin et al., 2002), MEL-11 moves from the cytoplasm to the membrane, away from the contractile apparatus, when contraction begins (Figs 5 and 6). After elongation is complete, low levels of cytoplasmic MEL-11 reappear. This MEL-11 membrane sequesterization depends at least in part on let-502 activity,consistent with LET-502 triggering elongation by MEL-11 inhibition(Fig. 7). Membrane localization was also disrupted when total NMY levels are reduced or eliminated, suggesting that the MEL-11/LET-502 interaction occurs on myosin(Fig. 8). Thus, the defective elongation seen in nmy-1 mutants may not only stem from decreased amounts of MHC, but also from an inhibition of myosin activity by the increased levels of cytoplasmic MEL-11.

In contrast to MEL-11 membrane sequesterization during elongation, LET-502 is interspersed with the myosin filaments(Fig. 6). In addition to triggering contraction by MYPT inhibition, ROK can directly phosphorylate rMLC in some in vitro systems (Amano et al.,1996). The localization of LET-502 with the NMY-1 filaments is consistent with this possibility.

Similar to what is seen in Drosophila(Edwards and Kiehart, 1996; Jordan and Karess, 1997; Wheatley et al., 1995), NMY-1 fails to form filaments in mlc-4 mutants but instead forms discrete foci, suggesting that MLC-4 is required for the proper myosin organization. In higher eukaryotes, rMLC is not required for MHC to assemble into contractile units, but rMLC phosphorylation establishes contacts between myosin and actin filaments (Trybus, 1996). Perhaps actin association is required to assemble NMY-1 into the ordered arrays we observe in C. elegans embryos. In the absence of mlc-4/rMLC, NMY-1 assembly is more random, forming the large foci rather than parallel arrays.

In addition to a role for LET-502 and MEL-11 in regulating the contraction of actin, an alternative model would suggest that like their mammalian homologs in focal adhesion formation, LET-502 and MEL-11 could regulate the attachment of actin cables at the apical surfaces of lateral epidermal cells. The C. elegans α and β_H-spectrin genes, spc-1 and sma-1, respectively, display elongation defects not unlike those seen in let-502 or mlc-4 mutants(McKeown et al., 1998; Norman and Moerman, 2002), and we observed that spc-1 and sma-1 are epistatic to mel-11 hypercontraction (Table 1). However, Norman and Moerman(Norman and Moerman, 2002)showed that the actin cables were not disrupted in let-502 and mlc-4 mutants, and we found that this was also the case for mel-11 and nmy-1. Moreover, the immunolocalization patterns for LET-502 and MEL-11 are opposite to that predicted by the actin anchoring model: LET-502, which would be predicted to tether actin to the membrane,remains cytoplasmic, while MEL-11, which would release the cables, is found at the membrane. Therefore, these results support a role for the LET-502/MEL-11 pathway in regulating the contraction of actin/myosin rather than anchoring of actin filaments.

Our analysis of nmy-1 suggests that it functions redundantly with another gene during embryonic elongation. The nmy-1 null phenotype, a mis-shapen adult, is much less severe than expected if it were the only MHC to regulate this process as mlc-4, the only rMLC known to function during elongation, displays arrest as unelongated L1 larva(Shelton et al., 1999). We demonstrated that a second MHC gene, nmy-2, functions redundantly with nmy-1. nmy-2(RNAi) disrupts AP polarity and cytokinesis in the early embryo (Cuenca et al.,2003; Guo and Kemphues,1996), which prevented examining nmy-2 function later in embryogenesis. However, we found that NMY-2 is present in epidermal cells at the onset of elongation, where it likely assembles into filamentous structures similar to NMY-1. Furthermore, nmy-2(weak RNAi); nmy-1double mutants displayed elongation-defective phenotypes similar to mlc-4 that were not seen for either nmy-2(weak RNAi) or nmy-1 alone (Table 3). Therefore, we have uncovered a role for nmy-2 in elongation.

To summarize, we have isolated a new component of the let-502/mel-11 pathway, nmy-1, which encodes a nonmuscle MHC that functions with mlc-4/rMLC to regulate the contractile process of embryonic elongation. nmy-1 functions redundantly with a second nonmuscle MHC gene, nmy-2. Together the two MHCs ensure the successful elongation of embryos into the characteristic long, thin worm.

We thank M. Walsh, W. Brook and J. McGhee for advice and comments on the manuscript, S. McWilliams for technical assistance, K. Kemphues for NMY-2 antibodies, K. Norman and D. Moerman for spc-1, and M. Glotzer for advice and reagents. Some strains were obtained from the Caenorhabditis Genetics Center, funded by the National Institutes of Health National Center for Research Resources. This work was supported by Canadian Institutes for Health Research and Alberta Heritage Foundation for Medical Research grants to P.E.M.