A Region of the Myosin Rod Important for Interaction With Paramyosin in Caenorhabditis elegans Striated Muscle

Corresponding author: Pamela E. Hoppe, Department of Genetics, Box 8232, Washington University School of Medicine, 4566 Scott Ave., St. Louis, MO 63110., phoppe{at}genetics.wustl.edu (E-mail)

Communicating editor: P. ANDERSON

	ABSTRACT

TOP ABSTRACT The C. elegans thick... Interaction between the... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The precise arrangement of molecules within the thick filament, as well as the mechanisms by which this arrangement is specified, remains unclear. In this article, we have exploited a unique genetic interaction between one isoform of myosin heavy chain (MHC) and paramyosin in Caenorhabditis elegans to probe the molecular interaction between MHC and paramyosin in vivo. Using chimeric myosin constructs, we have defined a 322-residue region of the MHC A rod critical for suppression of the structural and motility defects associated with the unc-15(e73) allele. Chimeric constructs lacking this region of MHC A either fail to suppress, or act as dominant enhancers of, the e73 phenotype. Although the 322-residue region is required for suppression activity, our data suggest that sequences along the length of the rod also play a role in the isoform-specific interaction between MHC A and paramyosin. Our genetic and cell biological analyses of construct behavior suggest that the 322-residue region of MHC A is important for thick filament stability. We present a model in which this region mediates an avid interaction between MHC A and paramyosin in parallel arrangement in formation of the filament arms.

WE are interested in the mechanisms that guide the assembly of myosin and the related protein paramyosin into the highly ordered striated muscle thick filament. Whereas models have been proposed for the molecular alignments involved in myosin-myosin and paramyosin-paramyosin assembly, the interaction between myosin and paramyosin has not been as thoroughly explored. Formation of a thick filament of appropriate structure requires molecular information encoded in the myosin heavy chain (MHC) or paramyosin molecule itself, as well as the activity of associated proteins (reviewed in LIU et al. 1997 ; BARRAL and EPSTEIN 1999 ). The observation that purified MHC or purified paramyosin protein assembles in vitro to generate the distinct periodicities found in MHC and paramyosin structures in vivo demonstrated that the intermolecular interactions that give rise to the fundamental repeating unit of the filament structure are encoded within the molecules themselves (HUXLEY 1963 ; reviewed in SQUIRE 1981 ). Similarly, experiments in which purified Caenorhabditis elegans MHC and paramyosin were mixed prior to assembly produced cofilaments containing a myosin cortex surrounding a paramyosin core (HARRIS and EPSTEIN 1977 ). Thus, the differential location of the myosin and paramyosin molecules within the filament is also directed, at least in part, by sequences within the molecules themselves. In Drosophila, the diverse cell-type-specific morphologies exhibited by thick filaments derived from different tissue sources are likely specified through interactions with other proteins (WELLS et al. 1996 ).

Although MHC and paramyosin assemble with different intermolecular axial staggers both in vitro and in vivo (reviewed in SQUIRE 1981 ), MHC and paramyosin are homologous proteins (Fig 1A) that share several features of protein sequence thought to be important for driving assembly and specifying molecular overlaps. Analysis of MHC and paramyosin rod sequences revealed a remarkable pattern of alternating concentrations of charge associated with a 28-residue repeat. Interactions between these segments of opposite charge are thought to play a major role in the assembly of both of these proteins into the thick filament (PARRY 1981 ; MCLACHLAN and KARN 1982 ; KAGAWA et al. 1989 ). In addition, MHC and paramyosin share a conserved pattern in the distribution of hydrophobic residues along the rod surface that may act to specify intermolecular overlap and drive assembly (HOPPE and WATERSTON 1996 ). Third, MHC and paramyosin contain a small region at the rod C terminus, called the assembly competent domain (ACD), which has been proposed to play a critical role in conferring assembly competence in MHC (SOHN et al. 1997 ). This 29-amino-acid domain is part of a 63-residue region, conserved in MHC and paramyosin, which displays distinctive sequence features. Compared to the rest of the coiled coil, the ACD has a unique charge profile, a relatively neutral total charge, and a high proportion of large apolar residues in surface positions (COHEN and PARRY 1998 ).

View larger version (26K): In this window In a new window Download PPT slide   Figure 1. (A) The major proteins of the invertebrate thick filament are MHC and paramyosin. Both MHC and paramyosin form dimers through a long -helical coiled-coil domain or "rod," represented here by a long rectangle. The MHC rod is divided into the 40 28-residue zones described by MCLACHLAN and KARN 1982 , and the shorter paramyosin molecule contains a subset of these zones (KAGAWA et al. 1989 ). Paramyosin is homologous to the C-terminal three-fourths of the MHC rod, which roughly corresponds to light meromyosin (LMM), the filament-forming MHC domain (LOWEY et al. 1969 ). The C. elegans proteins contain a homologous, small (30 residue) nonhelical region, which is known to be phosphorylated in paramyosin (SCHRIEFER and WATERSTON 1989 ). Interestingly, the homologous nonhelical sequences are located at opposite termini in paramyosin and MHC. The locations of the functional domains identified in previous studies are indicated. Regions 1 and 2 were defined in C. elegans as domains important for filament initiation (HOPPE and WATERSTON 1996 ). Studies with vertebrate myosin defined the ACD, which is important for ordered assembly (SOHN et al. 1997 ). (B) Filament initiation in C. elegans occurs through the association of MHC A (gray molecules) in a tail-to-tail or "antiparallel" organization. (C) Elongation of the filament arms from each side of the initiation center involves the assembly of MHC A (gray) and MHC B (white) in parallel arrangement. Paramyosin (black) lies in the filament core. Although paramyosin is shown here underlying the MHC A-containing region, this has not been demonstrated in C. elegans (see text). The orientation of paramyosin within the filament is drawn the same as that of MHC (N terminus is marked by the nonhelical tailpiece) but this has not been established.

	The C. elegans thick filament

TOP ABSTRACT The C. elegans thick... Interaction between the... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

C. elegans body wall thick filaments contain two isoforms of MHC, MHC A and MHC B, and the single paramyosin isoform, which is expressed in all muscle cell types of the worm (ARDIZZI and EPSTEIN 1987 ). During assembly, these three proteins segregate to three distinct compartments within the thick filament (Fig 1B and Fig C). Paramyosin lies in the filament core upon which MHC assembles (DEITIKER and EPSTEIN 1993 ). MHC A, the minor isoform encoded by the myo-3 gene, is present in the central 1.8-µm region of the 10-µm-long thick filament (MILLER et al. 1983 ). The major isoform, MHC B, is encoded by unc-54 (EPSTEIN et al. 1974 ) and is present only in the filament arms.

Genetic analysis has shown that the two MHC isoforms are functionally distinct. The center of the MHC A-containing region is composed of MHC dimers assembled in an antiparallel tail-to-tail fashion and is the site where filament initiation is thought to occur (MILLER et al. 1983 ). In addition, the MHC A-containing portion of the filament contains extensive regions in which MHC A is assembled in parallel, forming the base of each filament arm. Null mutations at the myo-3 locus cause the complete lack of normal thick filaments, resulting in embryonic paralysis and death (WATERSTON 1989 ). Thus MHC A appears uniquely capable of antiparallel association or some other aspect of filament initiation. Mutations that eliminate MHC B cause severe defects in motility, correlated with a reduction in the number of thick filaments present. The residual filaments can be of normal length, and they contain MHC A along their entire length (EPSTEIN et al. 1986 ). Increasing the expression of the minor isoform, MHC A, through introduction of a myo-3 transgene can restore an unc-54 (MHC B-deficient) animal to near wild type (FIRE and WATERSTON 1989 ), consistent with earlier results using chromosomal duplications (RIDDLE and BRENNER 1978 ; WATERSTON et al. 1982 ; OTSUKA 1986 ; MARUYAMA et al. 1989 ). Thus, in addition to being required to initiate the filament center, MHC A can assemble upon the paramyosin core in parallel to form the entire length of the filament arms.

Paramyosin is encoded by the unc-15 gene (WATER- STON et al. 1977). Mutant animals that lack paramyosin are paralyzed and contain hollow thick filaments composed of myosin, which are found generally near the ends of muscle cells, rather than in the region of the contractile apparatus (WATERSTON et al. 1977 ). These myosin filaments are more fragile (MACKENZIE and EPSTEIN 1980 ) and appear disordered, as they do not exhibit the distinct localization of MHCs A and B seen in wild type (EPSTEIN et al. 1986 ). Thus paramyosin appears to be necessary for normal thick filament stability and structure. However, because the paramyosin null mutant is viable, thick filaments capable of supporting contraction must be formed in the absence of paramyosin.

There is no direct evidence that paramyosin is present under the MHC A-containing region at the center of the thick filament in C. elegans. In molluscs, paramyosin is clearly present in the central region as well as in the length of the filament arms. In these species, MHC can be removed from isolated thick filaments, leaving an intact paramyosin core that can be shown by negative staining to be bipolar (SZENT-GYORGYI et al. 1971 ). Similar experiments cannot be done in C. elegans, however, because MHC A remains associated with the filament core under conditions that remove MHC B and a portion of the paramyosin (DEITIKER and EPSTEIN 1993 ). Staining of C. elegans adult muscle with paramyosin antibodies gives a pattern of localization much like that of MHC B, where staining is detected in the region of the A-band occupied by the filament arms, but a central gap in staining is seen in the medial MHC A-containing region (EPSTEIN et al. 1993 ). Future work is required to settle this issue.

Several thick filament-associated proteins have been identified in C. elegans, but it is not yet known whether their primary role is specification of intermolecular pattern or thick filament stability. The filagenins were identified biochemically as components of the thick filament core (DEITIKER and EPSTEIN 1993 ; LIU et al. 1998 ). The UNC-45 protein (VENOLIA et al. 1999 ), a component of thick filaments in C. elegans (AO and PILGRIM 2000 ), was identified through mutation (EPSTEIN and THOMSON 1974 ; VENOLIA and WATERSTON 1990 ). The protein is important for thick filament stability and may play a role in assembly (BARRAL et al. 1998 ).

	Interaction between the paramyosin mutation e73 and MHC A

TOP ABSTRACT The C. elegans thick... Interaction between the... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To elucidate the molecular events required to form a wild-type thick filament, we have investigated a unique genetic interaction between MHC A and a missense paramyosin mutant, unc-15(e73). The e73 mutation results in a single charge change, glutamic acid to lysine, which is thought to increase the affinity of paramyosin for itself during assembly (GENGYO-ANDO and KAGAWA 1991 ). Animals homozygous for the e73 allele move poorly and exhibit severely disrupted muscle structure with needle-like, brightly birefringent aggregates at the edges and ends of muscle cells. EM analysis shows that like the null mutant e1214, e73 contains hollow thick filaments of abnormal diameter in disorganized arrangement. In addition, e73 muscle contains dense aggregates that exhibit the periodicities typical of paramyosin paracrystals (WATERSTON et al. 1977 ). Antibody staining of isolated mutant assemblages and adult worms has shown that the aggregates contain paramyosin as well as associated MHC B and perhaps a small amount of MHC A; further, these aggregates correspond to the brightly birefringent bodies (EPSTEIN et al. 1987 ; GENGYO-ANDO and KAGAWA 1991 ). Antibody staining also revealed a diverse array of paramyosin-containing assemblages that are not detected by polarized light microscopy (EPSTEIN et al. 1993 ).

Genetic screens for second mutations that could restore motility in an e73 homozygote produced sup-3 alleles, which are mutations at the myo-3 locus that increase the expression of MHC A (RIDDLE and BRENNER 1978 ; WATERSTON et al. 1982 ; OTSUKA 1986 ; MARUYAMA et al. 1989 ). Although the sup-3 alleles can improve motility of some unc-54 and unc-15 mutants, the mechanism of suppression appears to be different for the two genes. Extra MHC A can compensate for unc-54 null mutations by directly replacing MHC B in elongation of the thick filament arms. However, extra MHC A poorly suppresses dominant and point mutations in unc-54, where mutant MHC B protein is made, which presumably interferes with the function of the wild-type proteins (RIDDLE and BRENNER 1978 ).

In contrast, extra MHC A poorly suppresses the paramyosin null e1214, but very effectively suppresses the point mutation e73. Failure to suppress the null indicates that increased levels of MHC A cannot directly substitute for paramyosin. Instead, suppression of the missense allele occurs because increased MHC A levels somehow lead to the increased incorporation of the e73 mutant paramyosin into normal thick filaments, so that suppressed animals now exhibit regions of muscle lattice that contain normal thick filaments with electron-dense cores (RIDDLE and BRENNER 1978 ; OTSUKA 1986 ).

To elucidate the molecular events involved in specifying the MHC-paramyosin interaction during thick filament formation, we are interested in defining the mechanism by which extra MHC A is able to shift the e73 mutant paramyosin from aggregate formation into thick filament formation. As a first step toward defining the MHC-paramyosin interaction at the molecular level, we worked to identify the sequences within the MHC A molecule responsible for e73-suppression activity. Our results define a 322-residue region of the rod that is sufficient for robust e73-suppression activity. The active region contains residues previously implicated as important for filament assembly, but is larger, and thus does not correlate with previously defined activities. While this region is critical for suppression activity, sequences outside this region appear to be required for full activity, arguing that the A-specific suppression involves sequences along a large length of the rod. Our data support a model in which e73 suppression occurs through a parallel MHC A-paramyosin interaction during elongation of the filament arms. Further, our results suggest that an interaction between paramyosin and the 322-residue region of MHC A plays an essential role in thick filament stability.

	MATERIALS AND METHODS

TOP ABSTRACT The C. elegans thick... Interaction between the... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

DNA constructs:
All clone fragments generated through PCR were sequenced unless otherwise indicated. Italicized bases in oligonucleotides do not match wild-type sequence. Generation of the chimeric myosin constructs has been described (HOPPE and WATERSTON 1996 ).

A truncated MHC A lacking the C-terminal nonhelical tailpiece was constructed by replacing the two consecutive P residues that mark the end of the coiled coil with two in-frame ochre codons. A 300-bp NcoI-KpnI genomic fragment from pPHpucAA was replaced with a PCR fragment amplified with primers TCAGCTTCCATGGCTTAATAAGATGGTTTCCCAATG and CTGCAGTGCCCATACATTGC.

An epitope-tagged version of the MHC B head with the hemagglutinin (HA) tag YPYDVPDYA inserted after the initiator M was generated in unc-54 subclone pPH24aP1 using a PCR technique called splice overlap extension (SOE; HORTON et al. 1989 ). A 438-bp MluI-XbaI genomic fragment was replaced with a PCR product amplified using SOE primers CCATACGACGTCCCAGACTACGCCGAGCACGAGAAGGACCCA and GTCTGGGACGTCGTATGGGTACATGATTTCTCGCTTCTT with outside primers TTCCCGATCTTACCAACTCC and GGGAAAGATGATCTTGAAGC.

A coinjection marker detectable in paralyzed worms was constructed using the rol-6 collagen gene promoter region to drive hypodermal green fluorescent protein (GFP) expression. Like the rol-6 message (PARK and KRAMER 1994 ), our marker is detectable in L1 through adult stages. Oligonucleotides GCTTTACACTTTATGCTTCC and TATATGGTACCCTGGAAATTTTCAGTTAGATC were used to amplify an 2 kbp fragment from pRF4 (MELLO et al. 1991 ). This fragment was digested with KpnI and cloned into KpnI-cut pPD65_67 (A. FIRE, S. XU, J. AHNN and G. SEYDOUX, personal communication), selecting for original orientation to the ATG, to make pPHgfp1. The construct was not sequenced.

Generation of transgenic lines:
Transformed lines carrying extrachromosomal arrays were generated as described (MELLO et al. 1991 ), using a 50:1 ratio of rol-6 pRF4/myosin or a 25:25:1 ratio of pPHgfp-1/Bluescript/myosin at 200 ng/µl in 10 mM Tris, 1 mM EDTA, pH 8. All constructs, except for chimera 10, were tested for suppression of unc-15 (e73) by direct injection into the homozygous mutant strain, CB73. If suppression of e73 was not obtained, the negative result was confirmed by injecting unc-54 (e190) homozygotes (CB190) to establish lines with improved motility and crossing these arrays into an e73 background (below). Chimera 10 was tested by first isolating a rescuing array in CB190 and then crossing into CB73.

Testing transgenic arrays in a paramyosin mutant background:
Most constructs were tested using the dominant rolling associated with the rol-6 plasmid pRF4 as marker in genetic crosses. Some arrays were marked with a plasmid driving GFP expression in the pharynx under the myo-2 promoter, a gift from Aguan Wei. At least two independent arrays were tested for each construct. The arrays were moved from the unc-54 (e190) background in which they were selected into wild type. Then, gld-1(q485) males were mated to wild-type transgenic hermaphrodites, and the transgenic male progeny were used to cross into unc-15(e73) homozygotes. Transgenic progeny were picked singly and allowed to self to identify the phenotypically wild-type balanced line, unc-15 (e73) + / + gld-1(q485). The q485 allele causes a tumorous germline in the adult hermaphrodite (FRANCIS et al. 1995 ). Once a balanced line was obtained, transgenic progeny, including Unc rollers identified by prodding animals during L3–L4 stages, were picked singly to identify any e73 homozygotes carrying the array through scoring the phenotypes of the self progeny. If all transgenic e73 homozygotes were Unc, we concluded that the construct did not have suppression activity.

To generate arrays that could be scored in completely paralyzed animals, the constructs were injected into unc-54(e190) with the rol-6::GFP coinjection marker described above. The lethality of transgenic arrays in animals homozygous for the unc-15 alleles e73, e1214, and su228 was tested by examination of all progeny in a 1-day collection from a single transgenic animal heterozygous for an unc-15 allele balanced by gld-1 (q485). Eggs were allowed to develop for 1 day, and then all transgenic animals were removed from the plate using a fluorescence dissecting microscope and sorted by phenotype. Green Unc animals were picked singly to score viability and brood size. Unhatched eggs were grouped and reexamined the following day; dead animals were mounted on microscope slides with 2% agar pads for viewing on a compound microscope, using fluorescence and Nomarski optics. The remaining progeny were allowed to grow for 1–2 days, at which time gravid animals (e73 +/ + q485) were picked singly for genotyping, and q485 homozygotes (+ q485) were counted.

Antibody staining:
Embryos were fixed with paraformaldehyde and methanol and stained using the methods of HRESKO et al. 1994 . Monoclonal antibody 5-14 was a gift of Henry Epstein and Irving Ortiz. The HA epitope was detected using anti-HA high affinity (a rat monoclonal) and fluoresceinated goat anti-rat IgG (both from Roche Molecular Biochemical).

	RESULTS

TOP ABSTRACT The C. elegans thick... Interaction between the... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Sequences within the MHC A rod confer e73-suppression activity in chimeric myosins:
Genetic screens for suppressors of the motility defects associated with the paramyosin mutant unc-15(e73) produced sup-3 mutations, which are genomic duplications of the myo-3 locus that increase MHC A expression (see Introduction). No comparable mutations were isolated at the unc-54 locus, suggesting that excess MHC B could not supply e73-suppression activity. To test this hypothesis directly, we generated transgenic arrays of wild-type unc-54 genomic sequence, both by direct injection into e73 homozygotes and by crossing arrays selected for unc-54(e190) rescue into an unc-15(e73) background (see MATERIALS AND METHODS). Overexpression of MHC B produced no improvement in motility of e73 homozygotes, nor did it cause any apparent adverse effects.

The differential function of the two MHC isoforms as revealed by the MHC A-specific suppression of unc-15(e73) provided an opportunity to probe the interaction of the MHC and paramyosin proteins during thick filament formation in vivo. To identify the MHC A sequences essential for e73 suppression activity, we generated transgenic arrays using chimeric MHC constructs, which contained different regions of MHC A and MHC B, and tested these arrays for e73 suppression in the manner described above. Because sup-3 alleles (genomic duplications that increase expression of MHC A) suppress the unc-15(e73) motility defects more strongly than they suppress unc-54(e190) defects (RIDDLE and BRENNER 1978 ), transgenic arrays selected for e190 rescue should have sufficient expression levels to suppress e73 if the critical MHC A residues are present. Further, selection for e190 rescue ensures that the level of transgene expression is not too high, since overexpression of MHC may cause an Unc phenotype (FIRE and WATERSTON 1989 ). All constructs used in this study rescue e190 and are thus functional myosins.

The results from the first set of constructs, diagrammed in Fig 2, indicate that the MHC A sequences encoding the rod confer e73 suppression activity in chimeric myosins. Chimera 2, which contains the N-terminal half of MHC A, including transcriptional regulatory elements, the motor domain, and the light-chain-binding sites, fails to suppress e73, indicating that these sequences are not sufficient for activity. Sequences encoding the MHC A C-terminal nonhelical tailpiece are not required, since replacement with MHC B sequences in chimera 4, or removal to produce a truncated MHC A in construct T1, does not affect suppression activity. In contrast, chimera 4, which contains the entire MHC A rod, exhibits suppression activity comparable to that of the wild-type MHC A construct in both motility and cell structure (Fig 3), indicating that suppression is mediated through the coiled-coil sequences that constitute the filament-forming region of the molecule.

View larger version (15K): In this window In a new window Download PPT slide   Figure 2. MHC A-specific e73-suppression activity maps to the MHC A coiled-coil rod. The schematic myosin molecules are positioned with N termini to the left and show MHC A residues in black and MHC B residues in white. The first line represents previously published results demonstrating that overexpression of MHC A suppresses (SUP) the motility defects of e73 homozygotes (see Introduction). The suppression activity of MHC B, three chimeric constructs, and a truncated MHC A are presented in the lines below. Failure to produce a change in the e73 phenotype is indicated by a minus sign. Taken together, these results demonstrate that sequences within the MHC A rod are necessary and sufficient for e73 suppression. Except for construct T1, the ability to supply filament-initiation function in a myo-3 homozygote has been previously published (HOPPE and WATERSTON 1996 ) and is included here only for comparison. All constructs in this and subsequent figures rescue the motility defects of the MHC B null, unc-54(e190).

View larger version (129K): In this window In a new window Download PPT slide   Figure 3. A chimeric MHC containing the MHC A rod dramatically improves muscle cell structure in an e73 homozygote, as visualized by polarized light microscopy of living animals. (A) Body wall muscle of the wild-type N2 strain exhibits brightly birefringent A-bands, which correspond to the thick-filament-containing region of the sarcomere. Shown is much of a single muscle cell, where arrowheads indicate the cell boundary. The ends of this cell are outside the area of the figure. The arrows point to the overlapping ends of two adjacent muscle cells. (B) The body wall muscle cell of an e73 homozygote has a shrunken appearance and shows little birefringence in the area of the contractile apparatus. The brightly birefringent, needle-like structures found near the ends and sides of the cell (arrow) are paramyosin paracrystals (see Introduction). (C) The striking improvement in motility in an e73 homozygote suppressed by an extrachromosomal array of chimera 4 is associated with amelioration of the structural defects in body wall muscle. Although far from wild type, the cells are larger and show increased birefringence in the region of the contractile apparatus such that some longitudinal striations can be discerned. No brightly birefringent aggregates are present. Arrowheads point to the cell boundary, and the arrow indicates a region where the ends of two muscle cells overlap. (D) The e73 homozygotes suppressed by expression of chimera 11 also show a striking improvement in motility, but are noticeably slower than e73 animals suppressed by chimera 4. Examination of the muscle cell phenotype shows that compared to e73, the cells are larger and exhibit more birefringence in the contractile apparatus. However, suppression is not as robust as seen with chimera 4 since paracrystals remain (arrows). This phenotype is similar to the suppression obtained with the genomic duplication sup-3(e1407) (BROWN and RIDDLE 1985 ).

A 322-residue region of the C-terminal MHC A rod is sufficient for suppression:
To elucidate the molecular events underlying the genetic interaction between MHC A and paramyosin, we used chimeric myosin constructs to identify the critical MHC A rod residues. Results from constructs included in Fig 4 define a 322-amino-acid region of the C-terminal rod, contained in chimera 11, that is sufficient for marked suppression of the motility and structural defects associated with the e73 mutation. However, although the suppression activity of chimera 11 is sufficient to restore motility and egg laying, none of the obtained transgenic lines could provide the degree of improvement in muscle cell structure and motility that was easily obtained with chimera 4, which contains the entire MHC A rod. As shown in Fig 3D, the best suppression obtained with chimera 11 showed significant improvement in muscle cell structure, but small birefringent aggregates remained in every cell.

View larger version (14K): In this window In a new window Download PPT slide   Figure 4. A 322-residue region of the C-terminal rod is sufficient for e73 suppression activity. The schematic myosin molecules are positioned with N termini to the left and show MHC A residues in dark gray and MHC B residues in white. The smallest region of MHC A, which in chimeric myosins gives robust suppression of e73 defects, is the 322-residue region of the C-terminal MHC A rod, indicated in chimera 11. The smaller region of MHC A residues in chimera 13, which contains the ACD, does not confer e73-suppression activity. The MHC A residues in chimera 13 correspond to region 2, one of the two MHC A rod domains that are independently sufficient in chimeric myosins for filament initiation function.

Suppression activity is lost upon division of the 322-residue region into two segments of 152 and 170 residues in chimeras 12 and 13, respectively. The 322-amino-acid minimal region contains two smaller domains that have been proposed to play important roles in myosin and paramyosin assembly: the 170-residue region sufficient for myo-3 rescue (HOPPE and WATERSTON 1996 ) as well as the 29-residue ACD (SOHN et al. 1997 ; COHEN and PARRY 1998 ). However, the results from chimeras 12 and 13 show that while both of these smaller regions of MHC A are required for e73 suppression function, neither is sufficient in chimeric myosins to improve motility of e73 homozygotes.

MHC chimeras that act as dominant enhancers of e73:
To determine the ability of rod sequences outside the 322-residue region to confer suppression activity, we tested chimeras containing N-terminal portions of the MHC A rod. Chimeras 2, 9, and 10 were tested and found to have no effect on the motility of e73 homozygotes (Fig 5A). As was found with constructs containing C-terminal rod regions of MHC A (Fig 4), testing more N-terminal MHC A rod regions revealed that the ability of a chimera to suppress e73 did not correlate with MHC A-specific filament initiation activity, indicating that these are separable functions.

View larger version (32K): In this window In a new window Download PPT slide   Figure 5. (A) The 322-residue region of the C-terminal MHC A rod is required for e73 suppression activity. All five rod chimeras diagrammed here contain MHC B sequences within the critical region at the rod C terminus, and all fail to suppress e73. Chimeras 7 and 8, which contain the largest number of N-terminal MHC A rod residues, instead cause an enhancement (ENH) of the severity of the e73 phenotype. Three of the chimeras contain region 1 and are therefore able to rescue myo-3. Neither the ability of a chimera to suppress e73 nor the ability of a chimera to enhance e73 correlates with myo-3 filament initiation function. (B) Enhancement of the e73 phenotype by rod chimeras containing MHC B residues within the 322-residue region. Transgenic progeny from a single hermaphrodite unc-15(e73) +/+ gld-1(q485) carrying a chimeric myosin array were analyzed to determine the effect of the transgene on the e73 phenotype (see MATERIALS AND METHODS). Two independent arrays (designated stExN) selected for unc-54(e190) rescue were used for each construct tested. The number of transgenic animals of each phenotype or genotype in one brood is presented in a column, and the percentage of the total transgenic progeny from that brood is given in parentheses. As expected, heterozygous parent hermaphrodites that carried no array produced progeny that were 25% Unc (e73 +), 50% phenotypically wild-type heterozygotes (e73 +/+ q485), and 25% tumorous germline (+ q485). A very similar distribution of phenotypes was obtained upon examination of the transgenic progeny of an e73 +/+ q485 hermaphrodite carrying a chimera 2 array, indicating that the presence of chimera 2 had no discernible effect on the e73 phenotype. In contrast, hermaphrodites carrying a chimera 7 transgene showed a specific loss of transgenic progeny with the Unc phenotype and the appearance of a new class with lethal or sterile phenotypes. Homozygous e73 animals carrying the chimera 8 transgene showed a similar decrease in viability, although array stEx89 produced a far less penetrant enhancement of e73 defects. (a) The most severe terminal phenotype found in the transgenic e73 animals was the Pat phenotype (paralyzed, arrested elongation at twofold), indicating a severe disruption in embryonic muscle function. Less severe phenotypes included Unc animals that failed to unfold after hatching and died in larval stages. (b) No animals had the Pat phenotype. (c) Includes 12 Pat animals. (d) Includes 13 Pat animals. (e) Includes 14 Pat animals. (f) Includes 2 Pat animals.

However, in performing genetic crosses designed to move chimera 7 arrays selected for unc-54(e190) rescue into an e73 background (see MATERIALS AND METHODS), we were unable to recover a transgenic e73 homozygote. Close examination of self progeny from e73/+ balanced hermaphrodites carrying the chimera 7 transgene revealed dead L1 animals arrested at the twofold stage of elongation (Pat phenotype), which is characteristic of mutations that cause severe defects in body wall muscle function (WATERSTON 1989 ; BARSTEAD and WATERSTON 1991 ; WILLIAMS and WATERSTON 1994 ).

To more easily follow the transgenic arrays in inviable or completely paralyzed animals, we constructed a coinjection marker that drives GFP expression in the hypodermis (see MATERIALS AND METHODS), which can be visualized using a fluorescence dissecting microscope. Using this marker, we tested chimeras 7, 2, and 8 in a balanced strain heterozygous for e73 and closely examined all transgenic progeny (see MATERIALS AND METHODS). The results (Fig 5B) indicate that chimeras containing large regions of the N-terminal MHC A rod, but lacking the C-terminal MHC A region necessary for e73 suppression, act as enhancers of the e73 phenotype, resulting in either lethality or a more severe Unc phenotype. The strongest, most penetrant enhancement of e73 defects was obtained with transgenic arrays of chimera 7, which contains MHC B sequences in the 322-residue C-terminal rod and MHC A residues in the remaining rod. The majority of transgenic e73 homozygotes showed the Pat phenotype. Chimera 8, which also contains MHC B residues in the 322-residue region but only a subset of the MHC A residues found in chimera 7, produced an enhancement of the e73 phenotype, but the effects were more variable and less penetrant.

The observed synthetic lethality of the chimera 7 transgene in an unc-15(e73) background occurs in animals that are wild type for both MHC genes and thus presumably have normal levels of wild-type MHCs A and B. This indicates that the construct has "antimorphic" or "dominant negative" character, acting as a poison to the thick filament. In a wild-type unc-15 background, the chimera 7 transgenes are able to rescue null mutations in either the myo-3 or unc-54 loci, indicating that the chimera can function in both filament initiation and elongation, respectively. To determine whether the dominant negative action of chimera 7 was due to an allele-specific interaction with the e73 mutant protein, we tested the chimera 7 array in other paramyosin mutant backgrounds. The chimera caused similar lethality in two other severe paramyosin mutants (Fig 6), indicating that the interaction is not allele specific. Instead, a more general loss of paramyosin function, coupled with the expression of chimera 7, is responsible for the synthetic lethality.

View larger version (101K): In this window In a new window Download PPT slide   Figure 6. (A) Chimera 7 acts as a dominant enhancer of other severe paramyosin mutations. The mutant lesions in the three alleles tested are shown (GENGYO-ANDO and KAGAWA 1991 ). (B) Analysis of the transgenic progeny of a single unc-15 +/+ q485 hermaphrodite revealed a loss of the Unc class and the appearance of dead and sterile animals, indicating that chimera 7 causes synthetic lethality in combination with two other unc-15 alleles, including the null e1214. (C and D) Progeny from a single e1214 +/+ q485 hermaphrodite were videotaped during embryogenesis and then photographed at the L1 stage. The Nomarski micrograph shows two larvae that exhibit the terminal Pat phenotype, surrounded by their siblings that have elongated normally and hatched from their eggshells. The pair of arrows indicates the tail and nose of one Pat larva, which has arrested at the twofold stage of elongation. The second larva has progressed slightly beyond the twofold stage. (D) The presence of the rol-6::GFP marker confirms that the two arrested larvae (arrows) contain the chimera 7 transgene.

Synthetic lethality is associated with defects in embryonic movement:
To gain insight into the cause of the synthetic lethality of the myosin chimera in paramyosin mutants, we used time-lapse video microscopy through Nomarski optics to determine the stage at which defects in embryonic movement appeared. For comparison, we first characterized the movement of unc-15 homozygous embryos that did not express the chimeric myosin. Although both unc-15 mutants e73 and e1214 exhibit a striking paralysis in larval and adult stages, the movement of the mutant animals within the eggshell is not obviously different from wild type. Like wild type (WILLIAMS and WATERSTON 1994 ), body wall twitching begins at the 1.5-fold stage of development, and the embryo is able to roll in the eggshell by the 2-fold stage. During later stages, there is no qualitative difference in the movement of the mutant embryos compared to wild type, as the mutants continue to turn and roll within the eggshell. We did not determine if quantitative differences exist. Because unc-15 mutant embryos elongate prior to hatching, these observations are consistent with the hypothesis that coordinated movement is required for successful embryonic elongation (WILLIAMS and WATERSTON 1994 ) and suggest that movement within the eggshell has less stringent requirements than locomotion through the environment, which occurs after hatching.

To determine the synthetic lethal phenotype of chimera 7 in an unc-15 background, collections of eggs from a transgenic heterozygous unc-15/ + balanced mother were videotaped during embryogenesis. Although the onset of muscle twitches occurred correctly at the 1.5-fold stage, the e73 animals carrying the chimera 7 array never exhibited vigorous movement or rolling and arrested elongation at the 2-fold stage (Fig 6). Thus the synthetic lethality is associated with a severe disruption in early muscle function.

Interestingly, e1214 null homozygous embryos that carry the chimera 7 transgene showed the same terminal elongation defect as transgenic e73 homozygotes, arresting at the twofold stage. However, these animals exhibited a less severe defect in early movement. Compared to the transgenic e73 homozygotes, these animals move more vigorously and show clear attempts at coordinated movement, including incomplete rolling. This is the opposite of what would be expected on the basis of severity of the paramyosin alleles alone, since e1214 homozygotes are more severely paralyzed than e73 homozygotes.

Synthetic lethality is associated with filament instability:
To gain insight into the cause of the synthetic lethality, we used an epitope-tagged version of chimera 7 to follow the behavior of the molecule in vivo. Because attaching the GFP tag to either the N or C terminus of MHC A resulted in impaired function in vivo (P. E. HOPPE and R. H. WATERSTON, unpublished observations), we fused the smaller HA tag to the N terminus of chimera 7. Analysis of transgenic lines demonstrated that the tagged protein rescues the motility defects of the unc-54 null and assembles normally during embryogenesis (Fig 7A and Fig B). However, most arrays containing an HA-tagged construct could not be passed through males in genetic crosses, indicating that the HA tag may be causing subtle behavioral defects.

View larger version (81K): In this window In a new window Download PPT slide   Figure 7. Embryos stained for the HA-tagged chimera 7 and for endogenous MHC A. The 5-14 antibody recognizes an MHC A epitope in the head domain (D. MILLER, personal communication). We confirmed that the epitope was absent from chimera 7 by the lack of 5-14 staining in adults and embryos of the genotype myo-3(st386); stEx68 (not shown). (A and B) Embryos from a homozygous unc-54(e190) line rescued with the HA-tagged chimera 7 array stEx63 were stained with anti-HA antibody to visualize protein distribution. (A) The HA-tagged chimera 7 assembles like wild type (EPSTEIN et al. 1993 ; HRESKO et al. 1994 ; WILLIAMS and WATERSTON 1994 ) during embryogenesis, as shown in this 1.75-fold embryo. (B) Wild-type organization is maintained following the onset of movement, as shown in this 3-fold embryo. Arrows indicate where the four A-bands per muscle quadrant can be counted. (C and D) Embryos from a balanced line e73/+ carrying the HA-tagged chimera 7 array stEx96. Because the q485 balancer is sterile, one-fourth of the transgenic embryos identified by the anti-HA antibody are e73 homozygotes. (C) A 3-fold embryo stained with anti-HA exhibits near-wild-type structure and is thus presumably e73/+ or +/+. No defects due to overexpression of MHC are apparent. (D) The same embryo stained with 5-14 to visualize endogenous MHC A. (E) An embryo exhibiting the synthetic lethal phenotype (presumably e73; stEx96) stained for the HA tag. The arrow points to a gap in MHC staining, which is adjacent to an area containing brightly staining material in disorganized arrangement. This portion of the muscle quadrant has detached from the body wall (position of the body wall indicated by arrowheads). (F) Endogenous MHC A shows a similar gap in staining.

The observation that synthetic lethality is associated with early defects in movement is consistent with either a disruption in early organization or a defect in filament stability, where structures assemble correctly but cannot be maintained. To distinguish between these possibilities, we collected eggs from a balanced line e73/+ carrying a transgenic copy of the HA-tagged chimera 7 and stained for the HA tag and for endogenous MHC A. Because the q485 balancer is sterile, one-fourth of the transgenic embryos identified by the anti-HA antibody are e73 homozygotes. Prior to the twofold stage, both endogenous MHC A and the chimera 7 protein exhibit an organized staining pattern that is no worse than that of the e73 homozygote, which has been described (EPSTEIN et al. 1993 ). In subsequent developmental stages, a portion of the transgenic progeny exhibits gaps in A-band staining as well as large, brightly staining masses (Fig 7E and Fig F), suggesting that the onset of vigorous contraction caused breakage and subsequent collapse of thick filament material. Within a single animal, one can find regions that contain near normal MHC staining and regions exhibiting highly disrupted structure (not shown).

	DISCUSSION

TOP ABSTRACT The C. elegans thick... Interaction between the... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We have exploited an isoform-specific interaction between MHC and paramyosin to identify a region of MHC critical for interaction with paramyosin in vivo. Expression of chimeric myosins in an unc-15(e73) mutant background has defined a 322-residue region of the C-terminal MHC A rod that is sufficient for robust suppression of the phenotypic defects associated with this paramyosin missense allele. The 322-residue region is necessary, since all chimeras containing MHC B sequences in this domain fail to suppress e73. Further, chimeras containing MHC B sequences in this C-terminal domain and MHC A sequences in a larger part of the N-terminal rod have the opposite effect, acting as dominant enhancers of the e73 phenotype, resulting in severe embryonic paralysis and death. These data strongly implicate the 322-residue region as a site critical for functional MHC A-paramyosin interactions.

MHC A sequences along a large length of rod contribute to isoform-specific function:
The ability of a chimeric myosin to suppress e73 depends both upon its expression level and the amount of MHC A rod residues contained within the chimera. Transgenic arrays of chimera 4, which contains the entire MHC A rod, produced better suppression of the e73 structural and motility defects than sup-3(e1407), the strongest sup-3 allele previously published (BROWN and RIDDLE 1985 ). The allele sup-3(e1407) is associated with an approximately two-fold increase in MHC A accumulation (WATERSTON et al. 1982 ). The observed improvement in suppression with transgenic arrays, which allow higher expression levels, is consistent with the known dose dependence of sup-3 alleles (RIDDLE and BRENNER 1978 ).

Our results show that at least the 322-residue region at the C terminus of the MHC A rod must be present in a chimeric myosin for suppression to occur. However, comparison of the degree of improvement in motility and muscle cell structure achieved by arrays expressing chimera 4 (which contains the entire MHC A rod) and arrays expressing chimera 11 (which contains only the minimal MHC A 322-residue active region) suggests that the intrinsic suppression activity of these two constructs is different (Fig 3). Because the chimera 11 arrays tested in e73 are able to dramatically improve movement in unc-54(e190), which requires higher levels of MHC A than e73 suppression, we believe the reduced suppression obtained with chimera 11 as compared to chimera 4 reflects a difference in suppression activity of the chimeras, rather than a failure to obtain sufficient protein expression. Therefore, our data suggest that rod residues outside the critical region also play a role in the isoform-specific interaction of wild-type MHC A with paramyosin. However, the unstable nature of the transforming arrays precludes sufficiently accurate protein quantification to confirm this hypothesis.

A second observation supporting the hypothesis that sequences along the length of the rod contribute to isoform-specific function comes from our analysis of the dominant negative effects resulting from expression of myosin chimeras that lack MHC A residues in the C-terminal rod in a paramyosin mutant background. The severity of the defects produced by a given chimera correlates with the number of MHC A residues contained in the N-terminal rod. Chimera 7, containing MHC A residues in all 762 rod positions outside the 322-residue region, produced the most severe enhancement of the e73 defects. Chimera 8, with 532 MHC A residues in the N-terminal rod, also increased the severity of e73 defects, but its effects were less penetrant. Division of the 762-residue MHC A rod region contained in chimera 7 into three subdomains in chimeras 2 (230 MHC A residues), 9 (264), and 10 (268) produced no detectable effects in e73. This argues for participation of residues along the length of the rod in determining myosin behavior in vivo.

The mechanism of e73 suppression:
Our genetic and cell biological analyses offer some insight into the mechanism by which overexpression of MHC A results in increased incorporation of mutant paramyosin into normal thick filaments. Our cell biological study of the HA-tagged chimera 7 construct is consistent with the hypothesis that the 322-residue MHC A domain plays an important role in filament stability. In the presence of chimera 7, MHC staining in e73 animals appears organized in early stages of development. However, staining of later animals exhibiting the synthetic lethal phenotype shows a marked disruption in thick filament organization (Fig 7E and Fig F), suggesting that the lattice was not maintained after the onset of vigorous contractions.

The proposed role of the 322-residue region in filament stability, rather than in filament formation or contractile activity, is consistent with the known behavior of chimera 7 in other mutant backgrounds. Chimera 7 can rescue null mutations in either MHC locus (myo-3 or unc-54) when wild-type paramyosin is present. Therefore, the chimera 7 protein must have motor function and must be capable of assembling in antiparallel and parallel arrangement to form the filament center and arms, respectively. In addition, we found that animals exhibiting the synthetic lethal phenotype can have both cells where chimera 7 is well organized and cells showing drastic defects in both chimera 7 and endogenous MHC A organization. Thus, chimera 7 protein must be capable of assembling into well-organized structures within at least some cells of the e73 homozygote. Coupled with the lack of discernible defects in any early embryos we examined, our data point to a primary defect in filament maintenance rather than filament formation.

The proposal that an MHC A-paramyosin interaction is critical for filament stability fits in well with what is known about the function of the two wild-type proteins. Paramyosin, MHC A, and the filagenins form the stable core structure that remains after removal of MHC B by treatment with high salt (EPSTEIN et al. 1985 ; DEITIKER and EPSTEIN 1993 ). Sequence comparisons of the nematode MHC proteins revealed that MHC A has diverged along the length of the rod in ways that make it more similar to paramyosin (HOPPE and WATERSTON 1996 ): these proteins share sequence features that correlate with the internal, strictly structural role of paramyosin in the thick filament (COHEN et al. 1987 ; KAGAWA et al. 1989 ).

The restoration of motility through the overexpression of MHC A in an e73 mutant is correlated with an increased number of normal thick filaments (OTSUKA 1986 ). In the e73 homozygote, much of the pool of thick filament component proteins is assembled into nonfunctional structures that do not contribute to motility. Much of the MHC B in the cell forms hollow thick filaments that are fragile (MACKENZIE and EPSTEIN 1980 ) and located outside the contractile apparatus (WATERSTON et al. 1977 ). The e73 paramyosin forms abnormal aggregates onto which MHC B assembles (EPSTEIN et al. 1987 ).

How does the action of the 322-residue region in mediating a stable interaction between MHC A and paramyosin lead to the production of an increased number of functional thick filaments? One possible mechanism of e73 suppression is based on the known function of MHC A at the filament center where myosin assembles in antiparallel fashion. In this model, an increased level of MHC A causes the nucleation of more filament centers, increasing the likelihood that a mutant paramyosin molecule will contact and contribute to a growing filament rather than joining a mutant assemblage. However, if MHC A-mediated suppression of e73 acted at the level of filament initiation, we would expect that chimeric MHCs capable of rescuing myo-3 lethality would also be able to suppress e73. Because these two activities require different critical regions within the MHC A molecule, our data suggest that e73 suppression function does not correspond to filament initiation, which presumably requires MHC A to assemble through antiparallel interactions.

A second, related model for suppression activity is that rather than initiating more filament centers, higher levels of MHC A increase the likelihood that an initiation event will succeed in producing a functional filament. In this model (Fig 8A and Fig B), suppression occurs because increased levels of MHC A cause expansion of the central MHC A-containing zone to occupy a larger fraction of the filament arms (EPSTEIN et al. 1986 ) where the additional MHC A is assembled in parallel, replacing MHC B. The more avid MHC A-paramyosin interaction (EPSTEIN et al. 1985 ) then drives an increased recruitment of the mutant paramyosin into these filament arms. This model, in which suppression occurs through the interaction of paramyosin with MHC A that is assembled in parallel, is consistent with the genetic data that filament initiation does not require paramyosin and that MHC A-mediated suppression of unc-15 does require that paramyosin protein is present (see Introduction).

View larger version (34K): In this window In a new window Download PPT slide   Figure 8. (A) Schematic representation of the model proposed for the genesis of the e73 mutant phenotype (GENGYO-ANDO and KAGAWA 1991 ). Mutant paramyosin molecules (black) have an increased tendency for self assembly. Therefore, the filament initiation centers formed by MHC A molecules (gray) are unable to compete for addition of paramyosin. In the worm, the region of MHC A molecules assembled in parallel at the base of the filament arms is much more extensive than shown here. The drawings of the assembled myosin molecules are not meant to represent the true thick filament structure. (B) Model for the mechanism by which overexpression of chimera 11 leads to restoration of motility in the e73 homozygote. MHC B sequences in chimera 11 are represented by white regions, and the rod is marked with the 28-residue zones. The MHC A sequences within the 322-residue region of chimera 11 confer MHC A-like assembly properties, enabling the chimeric protein to assemble with MHC A in the central region of nascent thick filaments. Like MHC A, chimera 11 has a sufficiently avid interaction with the e73 paramyosin to cause a shift in assembly of the mutant paramyosin from abnormal self-assembly to thick filament formation. Further, the 322-residue critical region enables the molecule to establish a sufficiently stable interaction with paramyosin to maintain filament integrity. (C) Model for the mechanism by which chimera 7 causes synthetic lethality in an e73 background. Like chimera 11, chimera 7 contains sufficient MHC A rod sequences to enable it to assemble with endogenous MHC A in the central region of the filament. However, because chimera 7 lacks MHC A sequences in the C-terminal rod, the chimeric protein is unable to bind effectively with paramyosin. The onset of contraction causes filament breakage. (D) Diagram of MHC and paramyosin molecules assembling in parallel with N termini to the left. If MHC and paramyosin associate with the same intermolecular stagger proposed for paramyosin-paramyosin assembly, the 322-residue region coincides with the predicted length of overlap between assembling molecules (GENGYO-ANDO and KAGAWA 1991 ).

The genesis of the synthetic lethal interaction between e73 and chimera 7 can also be explained in the context of this model for suppression activity (Fig 8C). Because chimera 7 contains a large amount of MHC A rod sequences, its partial MHC A-like activity leads to the addition of the chimeric protein to the growing filament centers present in e73 animals in the areas where wild-type MHC A assembles. However, because the chimera lacks MHC A sequences within the 322-residue region, it does not form sufficiently stable contacts with paramyosin. Therefore, the filament breaks when contractions begin, decreasing the number of functional filaments in the cell below the critical number required for viability. Because chimera 7 also causes synthetic lethality in the paramyosin null, although the defects in embryonic movement are less severe, this model requires that chimera 7 forms less stable contacts with surrounding MHC molecules than does MHC A, which also leads to filament breakage.

The more severe phenotype observed in the e73 transgenic animals suggests that the presence of mutant paramyosin is more detrimental in the background of the chimera 7 transgene than the complete absence of paramyosin. The increased severity of the synthetic lethal phenotype in the e73 background, as compared to the e1214 background, may be due to the aberrant aggregates of paramyosin. These aggregates may act to further weaken the contractile apparatus by recruiting thick filament proteins, such as MHC (EPSTEIN et al. 1987 ), the filagenins, and UNC-45, away from any potentially functional thick filaments in the cell.

In light of this model that MHC A-mediated suppression of e73 occurs in the filament arms, we find an intriguing correlation between the e73-suppression domain and the model proposed for the intermolecular axial stagger between paramyosin molecules assembled in parallel (Fig 8D). If MHC A-paramyosin assembly occurs with this same stagger, the region of overlap between the two molecules coincides with the 322-residue region. The borders of our minimal 322-residue region could then be explained because MHC A sequences throughout this segment contribute to the avid interaction with paramyosin at the MHC A C terminus. However, as discussed above, our data point to the participation of N-terminal MHC A rod residues in MHC A-paramyosin interaction in vivo. In the context of this model for the intermolecular stagger, the N-terminal MHC A rod sequences would provide the site of contact to a different paramyosin (or myosin) rod region in an adjacent molecule.

The role of the 322-residue region:
How do sequences within the 322-residue region mediate the MHC A-paramyosin interaction? One model is that thick filament-associated proteins bind this region of MHC A and supply the link to the appropriate region of the paramyosin molecule. The observation that sequences along the length of the rod contribute to isoform-specific behavior could result from the distribution of binding sites for one or more associated proteins throughout the coiled coil. The UNC-45 protein has been shown to co-localize with MHC B but not MHC A (AO and PILGRIM 2000 ), so it is an unlikely candidate to mediate an MHC A-paramyosin interaction. The filagenins are relatively good candidates for this putative role in MHC A-paramyosin assembly. These proteins assemble with paramyosin in the filament core in wild type (DEITIKER and EPSTEIN 1993 ) and can associate with MHC B in a paramyosin mutant background (LIU et al. 1998 ).

A second potential mechanism is that the active region of MHC A binds to paramyosin directly. A role in direct MHC-paramyosin interaction is consistent with what is known about the function of the smaller domains contained within the 322-residue region. The vertebrate ACD was defined biochemically using purified rod fragments (SOHN et al. 1997 ) and thus clearly functions in direct myosin-myosin interactions. In C. elegans, where paramyosin and other proteins are present and where the molecular events of filament initiation are not defined, the role of the active MHC A domains is unknown. However, the observation that the MHC A domains that are important for filament initiation exhibit a more hydrophobic rod surface, like that of paramyosin, and the demonstrated resistance of the MHC A-paramyosin interaction to dissociation by increasing salt concentrations (EPSTEIN et al. 1985 ) led to the model that these regions may mediate direct MHC-MHC and/or MHC-paramyosin contacts required to form a stable filament via these hydrophobic residues (HOPPE and WATERSTON 1996 ).

The model that matching hydrophobic rod exteriors and perhaps other shared sequence features support the more avid MHC A-paramyosin interaction can also be invoked to explain the isoform-specific ability of MHC A to suppress the e73 phenotype. In this model, the two MHC A-specific activities, filament initiation and e73 suppression, both result from a more hydrophobic MHC A coat. However, the critical regions for these two activities are different because filament initiation requires molecules to assemble in antiparallel configuration, whereas e73 suppression is mediated through parallel interactions between MHC and paramyosin. Therefore, the active regions we have defined are regions of the molecule that are involved in distinct steps of the assembly pathway. This idea has precedent in Acanthamoeba, where different regions of the rod have been associated with distinct steps in minifilament assembly (SINARD et al. 1989 , SINARD et al. 1990 ).

	ACKNOWLEDGMENTS

We thank Henry Epstein and Irving Ortiz for the 5-14 antibody. Tim Schedl and members of his laboratory provided the gld-1(q485) allele, the use of their microscopes for photography, and many helpful suggestions. We thank Ross Francis for help with polarized light microscopy and photography. We thank Chelly Hresko and Ross Francis for scientific discussions, and Tim Schedl, Chelly Hresko, and Daniela Gerhard for critical reading of the manuscript. We are grateful to Monika Arora for her assistance with the HA-tag construct and with assembly of the manuscript. Mutant strains were received from the Caenorhabditis Genetics Center. This work was supported by U.S. Public Health Service grant GM23833 and a Muscular Dystrophy Association grant awarded to R. H. Waterston, as well as National Research Service Award Fellowship 5 F32 GM12412-03 awarded to P. E. Hoppe.

Manuscript received April 6, 2000; Accepted for publication June 19, 2000.

	LITERATURE CITED

TOP ABSTRACT The C. elegans thick... Interaction between the... MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AO, W. and D. PILGRIM, 2000  Caenorhabditis elegans UNC-45 is a component of muscle thick filaments and colocalizes with myosin heavy chain B, but not myosin heavy chain A. J. Cell Biol. 148:375-384[Abstract/Free Full Text].

ARDIZZI, J. P. and H. F. EPSTEIN, 1987  Immunochemical localization of myosin heavy chain isoforms and paramyosin in developmentally and structurally diverse muscle cell types of the nematode Caenorhabditis elegans.. J. Cell Biol. 105:2763-2770[Abstract/Free Full Text].

BARRAL, J. M. and H. F. EPSTEIN, 1999  Protein machines and self assembly in muscle organization. Bioessays 21:813-823[Medline].

BARRAL, J. M., C. C. BAUER, I. ORITZ, and H. F. EPSTEIN, 1998  Unc-45 mutations in Caenorhabditis elegans implicate a CRO1/She4p-like domain in myosin assembly. J. Cell Biol. 143:1215-1225[Abstract/Free Full Text].

BARSTEAD, R. J. and R. H. WATERSTON, 1991  Vinculin is essential for muscle function in the nematode. J. Cell Biol. 114:715-724[Abstract/Free Full Text].

BROWN, S. J. and D. L. RIDDLE, 1985  Gene interactions affecting muscle organization in Caenorhabditis elegans.. Genetics 110:421-440[Abstract/Free Full Text].

COHEN, C. and D. A. D. PARRY, 1998  A conserved C-terminal assembly region in paramyosin and myosin rods. J. Struct. Biol. 122:180-187[Medline].

COHEN, C., D. E. LANAR, and D. A. D. PARRY, 1987  Amino acid sequence and structural repeats in Schistosome paramyosin match those of myosin. Biosci. Rep. 7:11-16[Medline].

DEITIKER, P. R. and H. F. EPSTEIN, 1993  Thick filament substructures in Caenorhabditis elegans: evidence for two populations of paramyosin. J. Cell Biol. 123:303-311[Abstract/Free Full Text].

EPSTEIN, H. F. and J. N. THOMSON, 1974  Temperature-sensitive mutation affecting myofilament assembly in C. elegans.. Nature 250:579-580[Medline].

EPSTEIN, H. F., R. H. WATERSTON, and S. BRENNER, 1974  A mutant affecting the heavy chain of myosin in Caenorhabditis elegans.. J. Mol. Biol. 90:291-300[Medline].

EPSTEIN, H. F., D. M. MILLER, I. ORTIZ, and G. C. BERLINER, 1985  Myosin and paramyosin are organized about a newly identified core structure. J. Cell Biol. 100:904-915[Abstract/Free Full Text].

EPSTEIN, H. F., I. ORITZ, and L. A. T. MCKINNON, 1986  The alteration of myosin compartmentation in specific mutants of Caenorhabditis elegans.. J. Cell Biol. 103:985-993[Abstract/Free Full Text].

EPSTEIN, H. F., I. ORITZ, and G. C. BERLINER, 1987  Assemblages of multiple thick filaments in nematode mutants. J. Muscle Res. Cell Motil. 8:527-536[Medline]