To discriminate between long and short LEV-11 isoforms in C. elegans, we raised the antibodies “anti-LEV-11(Long)” against a peptide coded within lev-11 exons 1 and 2 (the starting two exons for all long isoforms) and “anti-LEV-11(Short)” against a peptide coded within exon 3b (the starting exon for all short isoforms) (Fig.1). Based on western blot analysis, these antibodies specifically recognize their respective isoform classes. That is, anti-LEV-11(Long) recognized bacterially expressed and purified long isoform LEV-11A but not short isoforms LEV-11C or LEV-11E, while anti-LEV-11(Short) recognized recombinant LEV-11C and E, but not LEV-11A (Fig.2A). Similarly, anti-LEV-11(Long) recognized in whole worm extracts an endogenous protein of approximately the same molecular weight as recombinant LEV-11A, and also recognized exogenously expressed higher molecular weight amino-terminally N-terminally GFP-tagged LEV-11A (GFP::LEV-11A), but not GFP::LEV-11C or GFP::LEV-11E (Fig.2A). In contrast, anti-LEV-11(Short) recognized in worm extracts endogenous proteins of similar molecular weight as recombinant LEV-11C and E, and also recognized overexpressed GFP::LEV-11C and GFP::LEV-11E, but not GFP::LEV-11A (Fig.2A).

We attempted to knock down expression of all short LEV-11 isoforms using RNAi, simultaneously targeting exon combinations 3a-4a-5a (the first three exons for LEV-11E, T, U) and 3a-4a-5b (the first three exons for LEV-11C). However, western blots probed with anti-LEV-11(Short) and anti-LEV-11(Long) showed no changes in LEV-11 expression (Fig.2A), and we observed no phenotypes among the progeny of treated animals. As an alternative approach, we used CRISPR/Cas9-mediated mutagenesis to insert stop codons in three reading frames into lev-11 exons used by different subsets of LEV-11 isoforms (Fig.1). As previous RNAi-based studies implicated LEV-11C as an essential isoform (Anyanful et al., 2001), we first inserted stop codons into exon 5c to (Fig.1). Such premature termination would be expected to either trigger nonsense mediated mRNA decay (NMD) of LEV-11C-coding mRNA, or in absence of NMD, prevent synthesis of the C-terminal half of LEV-11C, including residues critical for cooperative binding to F-actin (Fig.1). For simplicity, we refer to this allele as lev-11(C-). Surprisingly, animals homozygous for any of three independently isolated lev-11(C-) alleles were viable over multiple generations and exhibited no reduction in the number of eggs produced (Fig.3A,B). By western blot analysis, lev-11(C-) animals still expressed significant amounts of low molecular weight LEV-11(Fig.2B), suggesting LEV-11C is not the predominant short isoform in terms of abundance.

Very different results were obtained when we inserted stop codons in three frames into exon 5a, predicted to disrupt LEV-11E, T, and U, while sparing LEV-11C (Fig.1). We isolated two independently generated lev-11(ETU-) alleles in heterozygous animals, but were unable to recover viable homozygous lev-11(ETU-) progeny from either isolate, suggesting LEV-11E, T, and/or U is essential. We observed that 100% of eggs laid by lev-11(ETU-)/(+) heterozygotes hatched as L1 larvae, but approximately 25% of these failed to grow beyond the L1 stage (Fig.3C). Genotyping of the surviving adult progeny of lev-11(ETU-)/(+) heterozygotes showed all were wild-type or heterozygous for lev-11(ETU-), suggesting lev-11(ETU-) homozygotes die during early larval development. In contrast, we observed complete development to adulthood for approximately 100% of the progeny of wild type siblings of lev-11(ETU-)/(+) animals, or of unrelated wild type animals (Fig.3C).

Considering that previous RNAi targeting maternal and zygotic expression of all short LEV-11 isoforms had implicated these proteins in embryonic development (Anyanful et al., 2001), we were surprised lev-11(ETU-) homozygotes did not arrest during embryogenesis, which would have been reflected in the presence of unhatched eggs (Fig.3C). To test whether the lack of a severe embryonic phenotype was due to compensation by LEV-11C, we used CRISPR/Cas9-mediated mutagenesis to insert stop codons into exon 3b, generating a lev-11(CETU-) allele lacking all four short LEV-11 isoforms (Fig.1). Similar to lev-11(ETU-), we recovered heterozygous animals, but were unable to recover homozygotes for lev-11(CETU-). Again, all eggs of lev-11(CETU-)/(+) heterozygotes hatched as L1 larvae, and approximately 25% of these failed to develop further (Fig.3D). Thus, larval but not embryonic development in C. elegans depends on zygotic expression of a short LEV-11 isoform, although these results did not rule out contributions of maternally provided short isoforms to embryogenesis.

Based on the larval arrest phenotype of lev-11(ETU-) and lev-11(CETU-) animals (Fig.3C,D), and the fact that LEV-11E is highly expressed (Watabe et al., 2018; Ono et al., 2023), we suspected LEV-11E is an essential short Tpm isoform. LEV-11E is expressed at highest levels in the excretory cell and pharyngeal muscles (Watabe et al., 2018). To test whether forced expression of LEV-11E in pharyngeal muscle is sufficient to rescue viability of lev-11(ETU-) and lev-11(CETU-) homozygotes, we cloned the lev-11E cDNA behind the strong myo-2 pharyngeal muscle myosin promoter (Okkema et al., 1993) and an N-terminal mCherry tag. Plasmid DNA injected into the syncytial gonads of C. elegans hermaphrodites can recombine into linear extrachromosomal arrays (ECAs) that can be inherited in subsequent progeny (Mello et al., 1991). We injected plasmid bearing the mCherry-tagged LEV-11E (mCH::LEV-11E)-coding transgene into heterozygous lev-11(ETU-)/(+) and heterozygous lev-11(CETU-)/(+) animals. We refer here to the ECAs thus produced with the shorthand notation, Ex[mCH::lev-11E]. Strikingly, genotyping the adult progeny of injected animals showed we could now recover homozygous lev-11(ETU-) and lev-11(CETU-) animals. Notably, these animals had mCherry-positive pharynges, indicating they had inherited the Ex[mCH::lev-11E] ECA. Furthermore, these homozygous mutants in turn produced viable mCherry-positive progeny, with no discernable loss of viability over multiple generations. Immunoblot of whole worm extracts with anti-LEV-11(Short) confirmed lev-11(CETU-); Ex[mCH::lev-11E] worms lack endogenous short LEV-11 isoforms but express higher molecular weight ectopic mCH::LEV-11E (Fig.2B). Thus, mCherry-tagged LEV-11E can provide for all essential maternal and zygotic functions of short LEV-11 isoforms. Interestingly, immunoblot of extracts from lev-11(ETU-); Ex[mCH::lev-11E] worms also showed little to no endogenous short LEV-11 (Fig.2B), further indicating LEV-11C is a lower abundance isoform.

Among Ex[mCH::lev-11E]-bearing animals, mCH::LEV-11E was strongly expressed in pharyngeal muscle, but there was also weak expression in body wall muscle, and we could not rule out unobserved lower expression in other tissues. To specifically test whether expression of LEV-11E in the pharynx is essential for viability, we took advantage of the propensity of ECAs to be inherited in a mosaic manner (Mello et al., 1991). We fixed L1-stage progeny of lev-11(ETU-); Ex[mCH::lev-11E] and of lev-11(CETU-); Ex[mCH::lev-11E] worms, stained these for F-actin, and inspected for mCH::LEV-11E expression in the pharyngeal muscles pm3 through pm7, ignoring pm1, pm2, and pm8 due to their relatively small sizes. For both strains, progeny at the L1 stage included animals with no pharyngeal mCH::LEV-11E, animals with mCH::LEV-11E in all pharyngeal muscle cells, and animals with mCH::LEV-11E expression in only some muscle cells (Fig.3E). When we performed a similar analysis on the progeny that reached adulthood, > 97% expressed mCH::LEV-11E in all pharyngeal muscle cells (Fig.3E), while the remaining animals lacked mCH::LEV-11E expression in no more than one out of three muscle cells in a single pharyngeal muscle. These results suggest LEV-11E must be expressed in all the major pharyngeal muscles for normal larval development.

Pharyngeal muscle cells also serve as epithelial cells that line the pharyngeal lumen. Most pharyngeal muscle cells have large radially-arranged sarcomeres that span the distance from the apical surface of the muscle cell facing the lumen to the basal surface (Albertson et al., 1997). Thus, thin filaments attach to sarcomere Z-line equivalents at either the apical or the basal plasma membranes. In lev-11(ETU-); Ex[mCH::lev-11E] and lev-11(CETU-); Ex[mCH::lev-11E] L1 larvae or adults stained for F-actin, mCH::LEV-11E overlapped with pharyngeal muscle thin filaments except near the Z-lines at the apical and basal surfaces (Fig.4A, Fig.5). Confirming this, we also performed immunostain with anti-LEV-11(Short) and anti-mCherry. The more rigorous processing required for immunostain, and the lower permeability of worms to antibodies compared to phalloidin, resulted in poorer preservation and decoration of thin filaments, but we observed some radial filaments stained with anti-LEV-11(Short) in wild type larval and adult pharyngeal muscle, and anti-LEV-11(Short) immunostain colocalized with anti-mCherry stain in Ex[mCH::lev-11E]-positive animals (Fig.4B,C). Conversely, neither anti-LEV-11(Short) nor anti-mCherry stained radial filaments in lev-11(CETU-) animals lacking the ECA (Fig.4B). Thus, LEV-11E associates with thin filaments in pharyngeal muscle.

Surprisingly, actin filament distribution was not visibly altered in pharynges of lev-11(ETU-) or lev-11(CETU-) L1 animals that failed to inherit the Ex[mCH::lev-11E] array (Fig.4A, 5A), suggesting LEV-11E is not required to maintain thin filament stability or overall sarcomere organization. To test whether LEV-11E is required for pharyngeal muscle activity, we observed pharynges of live animals over time. In the presence of food, pharyngeal muscle of wild type animals undergoes rhythmic contraction termed “pharyngeal pumping”, which can be observed through the transparent body of worms using a stereomicroscope. We observed pharyngeal pumping in 99% of wild type L1 larvae and in 76% of mCherry-positive lev-11(CETU-); Ex[mCH::lev-11E], but in only 2.7% of mCherry-negative lev-11(CETU-) larvae lacking LEV-11E expression (Fig.3F, Movie 1, Movie 2). Taken together, these results suggest LEV-11E plays little to no role in organizing sarcomere structure of pharyngeal muscles, but does regulate contraction.

Considering the relationship between Tpm and formins in other systems, we set out to test whether LEV-11E functionally interacts with formins in the pharynx. Single cell RNA sequencing of L2 larvae has shown FHOD-1 is the only formin expressed at high levels in pharyngeal muscle during this developmental stage (Cao et al., 2017). Consistent with this, we have observed strong expression of FHOD-1 in pharyngeal muscle by immunostain of endogenous FHOD-1 and through expression of a functional C-terminally GFP-tagged FHOD-1 (FHOD-1::GFP) (Mi-Mi et al., 2012).

As a first test for interaction between FHOD-1 and LEV-11E, we performed immunoprecipitations (IP) from whole worm lysates using polyclonal anti FHOD-1 (Mi-Mi et al., 2012), followed by mass spectrometry (MS) analysis. Lysates were isolated from worms expressing FHOD-1::GFP at an ectopic locus (resulting in roughly 2–3-fold overexpression) or worms bearing a putative null deletion allele fhod-1(tm2363), referred to hereafter as fhod-1(∆). Asynchronous populations of worms, as well as worm populations synchronized at late larval stages L3 to L4 were sampled. Based on total spectrum counts after MS analysis of pre-IP extracts, lysates contained an abundance of proteins from body wall muscle (e.g., actin, body wall muscle myosin II heavy chains UNC-54/MYO-4 and MYO-3, and paramyosin/UNC-15) pharyngeal muscle (e.g., actin and pharyngeal muscle myosin II heavy chains MYO-1, MYO-2), and intestine (e.g., vitellogenins) (Supplementary Table S1). FHOD-1 (also called Formactin) was not detected in pre-IP extracts, suggesting the formin is of low abundance, but FHOD-1 was strongly enriched after IP from extracts of FHOD-1-over-expressing worms but not of fhod-1(∆) mutants (Supplementary Tables S2, S3). Many sarcomere proteins were enriched after FHOD-1 IP in samples from FHOD-1-overexpressing worms as well as from fhod-1(∆) worms (Table 1), indicating there was significant non-specific precipitation. Despite this high background, short isoforms of Tpm/LEV-11 were strikingly recovered after FHOD-1 IP only from extracts of FHOD-1-overexpressing worms and not from fhod-1(∆) worms. This was true for asynchronous worms and from late-stage larvae (Table 1). Notably, MS analysis of pre-IP extracts showed short Tpm1/LEV-11 isoforms were present at similar levels in fhod-1(∆) and FHOD-1-over-expressing animals (Table 1), suggesting their failure to precipitate from fhod-1(∆) extracts was not simply due to scarcity. Thus, one or more short LEV-11 isoforms are specifically pulled down with FHOD-1 from worm extracts.

Unfortunately, short LEV-11 isoforms have extensive shared sequences that prevented MS analysis from identifying specific isoforms, but the strong expression of FHOD-1 in pharyngeal muscle and relative absence from neurons or intestine (where LEV-11C is expressed) supported LEV-11E as a likely candidate. To probe for association between FHOD-1 and LEV-11E in intact pharyngeal muscle, we crossed the FHOD-1::GFP-expressing transgene into the lev-11(ETU-); Ex[mCH::lev-11E] background. In L1 larvae expressing mCH::LEV-11E, FHOD-1::GFP localized at the ends of mCH::LEV-11E-decorated filament bundles near the Z-lines at the apical and basal plasma membranes (Fig.6). This was consistent with our previous observations of FHOD-1::GFP in wild type animals (Mi-Mi et al., 2012). Notably, there was little to no overlap between FHOD-1::GFP and mCH::LEV-11E (Fig.6, arrows), suggesting the two proteins bind distinct domains on pharyngeal thin filaments. Based on this, we suggest FHOD-1 and LEV-11E likely coprecipitate through an interaction bridged by thin filaments that have not fully depolymerized.

The absence of colocalization of FHOD-1 and LEV-11E did not rule out the possibility the two proteins influence the localization of each other. To test this, we examined FHOD-1::GFP in lev-11(ETU-) animals that failed to inherit the mCH::LEV-11E-expressing ECA. Notably, pharyngeal FHOD-1::GFP localization was not changed, including its restriction to a territory proximal to the plasma membrane, suggesting FHOD-1 localization does not depend on LEV-11E (Fig.6). Based on F-actin stain, pharyngeal muscle thin filaments appear normal in fhod-1(∆) animals (Mi-Mi et al., 2012), but it remained possible FHOD-1 might promote assembly of LEV-11E onto thin filaments. To test this, we crossed fhod-1(∆) into the lev-11(CETU-); Ex[mCH::lev-11E] genetic background. In these fhod-1(∆) animals, mCH::LEV-11E localized normally to thin filaments (Fig.6A,B). Moreover, such animals exhibited pharyngeal pumping, were viable, and grew to adulthood. Thus, FHOD-1 does not appear to influence the function or localization of LEV-11E in pharyngeal muscle. Finally, to determine whether simultaneous loss of FHOD-1 and LEV-11E has a synthetic effect on pharyngeal sarcomere organization, we examined F-actin in fhod-1(∆) lev-11(CETU-) L1 larvae that failed to inherit the mCH::LEV-11E-expressing ECA. Again, pharyngeal F-actin appeared normal in these animals (Fig.6A). Thus, FHOD-1 and LEV-11E associate with different domains of sarcomeric thin filaments in pharyngeal muscle, but neither protein affects the localization of the other, and both proteins are dispensable to establishing the overall actin cytoarchitecture of pharyngeal muscle.

FHOD-1 is also expressed in body wall muscle, where it localizes to small bodies at the edges of muscle cells and in faint striations that overlap dense bodies, which serve as sarcomere Z-lines (Mi-Mi et al., 2012). Thus, we considered the possibility that some co-precipitation of short LEV-11 isoforms we observed of FHOD-1 might arisen from an interaction in body wall muscle. Long LEV-11 isoforms are highly expressed in body wall muscle, but no short isoform is known to be expressed at high level in that tissue (Barnes et al., 2018; Watabe et al., 2018). Previous immunostain with a pan-LEV-11 antibody has shown LEV-11 decorates thin filaments except for a segment proximal to the dense bodies/Z-lines (Ono & Ono, 2002; Barnes et al., 2018). Interestingly, GFP-tagged short isoform LEV-11T (GFP::LEV-11T) exogenously expressed in body wall muscle adopts a similar distribution on thin filaments, while LEV-11U, which has a much lower affinity for F-actin, forms aggregates when overexpressed in body wall muscle (Ono et al., 2023). To test the ability of the other short isoforms to bind body wall muscle thin filaments, we generated ECAs expressing GFP-tagged short isoforms LEV-11C or E, or as a control long isoform LEV-11A, using the strong body wall muscle-specific myo-3 promoter (Waterston, 1989). For ease of analysis, we expressed these constructs in worms in which the endogenous dense body marker α-actinin/ATN-1 has been tagged with mCherry (ATN-1::mCH) (Kimmich et al., 2024). As expected, GFP::LEV-11A colocalized with thin filament F-actin except in a zone proximal to the dense bodies (Fig.7A). We also observed this localization for GFP::LEV-11C and GFP::LEV-11E (Fig.7A), suggesting most short isoforms of LEV-11 have the capacity to bind thin filaments in body wall muscle sarcomeres when expressed exogenously.

To probe for endogenous short LEV-11 isoforms in body wall muscle, we immunostained wild type animals using anti-LEV-11(Short) and, as control, anti-LEV-11(Long), as well as an antibody that recognizes the dense body protein ATN-1 (Fig.7B). As expected, anti-LEV-11(Long) immunostain recapitulated our observations with GFP::LEV-11A (Fig.7B) and with previous pan anti-LEV-11 immunostain (Ono & Ono, 2002). However, immunostain with anti-LEV-11(Short) did not resemble over-expressed GFP-tagged short isoforms, but instead decorated faint striations in body wall muscle that overlapped dense bodies (Fig.7C). To test the specificity of this immunostain, we also stained anti-LEV-11(Short) in lev-11(CETU-) animals bearing a modified ECA, Ex[mCH::lev-11E + myo::gfp]. This ECA expresses mCH::LEV-11E from the pharyngeal myo-2 promoter, but at a lower level than previously to reduce the leaky body wall muscle expression. Additionally, the ECA carries a GFP transgene driven by the myo-3 promoter for strong body wall muscle expression, allowing the presence of GFP to indicate inheritance of the ECA by individual body wall muscle cells. Thus, GFP-negative body wall muscle cells would lack the mCH::LEV-11E-coding transgene, and therefore would lack all short LEV-11 isoforms. Anti-LEV-11(Short) immunostain produced the same striated pattern in all body wall muscle cells of this strain, regardless of whether cells were GFP-positive or GFP-negative (Fig.7C). These results strongly suggest anti-LEV-11(Short) immunostain in body wall muscle is non-specific.

Our failure to detect endogenous short LEV-11 isoforms in body wall muscle sarcomeres did not rule out the possibility that an undetected population functionally interacts with FHOD-1 in that tissue. FHOD-1 promotes new sarcomere formation in body wall muscle. As such, body wall muscle cells in fhod-1(∆) mutants assemble striations more slowly than in wild type animals, resulting in narrow body wall muscles with fewer striations per muscle cell (Mi-Mi et al., 2012; Sundaramurthy et al., 2020; Kimmich et al., 2024). Examining body wall muscle size in the three lev-11(C-) strains, we observed that one strain had narrower body wall muscle muscles compared to age-matched wild type animals, similar to, but of lesser magnitude than, fhod-1(∆) mutants (Fig.8A). But no such difference was observed for the other two lev-11(C-) strains, suggesting this effect was unrelated to lev-11(C-). Additionally, no lev-11(C-) strain had significantly fewer striations per muscle cell than wild type, unlike fhod-1(∆) mutants (Fig.8B). Thus, LEV-11C is not required for FHOD-1-dependent body wall muscle growth. We chose not make a similar comparison for mutants lacking LEV-11E, as partial defects in pharyngeal pumping and feeding can lead to reduced body growth, which negatively impacts body wall muscle growth (Sundaramurthy et al., 2020).

However, FHOD-1 also promotes dense body assembly. Thus, while dense bodies in wild type animals appear relatively uniform in size, shape, and spacing, those in fhod-1(∆) mutants appear irregular in shape and spacing and prone to deformation during muscle contraction (Sundaramurthy et al., 2020; Kimmich et al., 2024). To examine dense body organization in presence or absence of short LEV-11 isoforms, we immunostained for ATN-1 and the ECA marker GFP in lev-11(CETU-) animals with mosaic inheritance of Ex[mCH::lev-11E + myo::gfp]. Dense bodies appeared normal in all body wall muscle cells, regardless of the presence or absence of the ECA (Fig.8C). Thus, absence of all short LEV-11 isoforms does not affect dense body morphology in a manner similar to fhod-1(∆). In summary, we did not detect endogenous short LEV-11 isoforms in body wall muscle sarcomeres, or find evidence short LEV-11 isoforms function with FHOD-1 in body wall muscle development.

lev-11E cDNA was isolated from Yuji Kohara clone yk1035c01 with a single point mutation (T667A coding V225A) that we corrected by site-directed mutagenesis using QuikChange Lightning (Agilent, Santa Clara, CA). lev-11C cDNA was isolated from Yuji Kohara clone yk1213f01. lev-11A cDNA was isolated from the Pro-Quest C. elegans cDNA library for yeast two hybrid screening (discontinued product from Invitrogen). Plasmids L4440-lev-11C and L4440-lev-11E for dsRNA production for RNAi were generated by amplifying the first three exons of lev-11C and lev-11E, respectively, with flanking KpnI and NheI sites, and subcloning into L4440. Plasmid pLMM19 was generated by introducing flanking BamHI sites to lev-11C cDNA by PCR and subcloning into pGEX-6p-3. Plasmid pLMM25 was generated by introducing flanking SmaI sites to lev-11E cDNA and subcloning into pGEX-6p-2. Plasmid pGEX-6p-3-lev-11A was generated by seamlessly cloning lev-11A cDNA into pGEX-6p-3 by In-Fusion cloning (Takara, Shiga, Japan). For expression of mCherry-tagged LEV-11 isoforms in body-wall muscle, lev-11A, C, and E cDNAs were appended at their 5’ ends with sequence encoding 5-amino acid Ala-Gly-Ala-Gly-Ser linker, and flanked by BamHI sites using PCR before insertion by In-Fusion cloning into pCFJ104 (Addgene #19328; Watertown, MA), between the body wall muscle-specific myo-3 promoter driving mCherry, and the unc-54 3’ untranslated region (UTR). Resultant plasmids were named pCP-Cherry-lev-11A, pCP-Cherry-lev-11C, and pCP-Cherry-lev-11E. For pharyngeal expression of LEV-11E, lev-11E cDNA from pCP-Cherry-lev-11E, together with linker and BamHI sites, was amplified and inserted by In-Fusion cloning between myo-2 promoter driving mCherry and unc-54 3’UTR of pCFJ90 (Addgene #19327; Watertown, MA) to generate plasmid pKM3. Plasmid myo-3::gfp was created by fusion PCR of the myo-3 promoter, gfp (from pPD95.75), and the unc-54 3’ UTR, followed by TOPO-TA cloning of the resultant myo-3p::gfp::unc-54 3’utr product into plasmid pCR2.1 (Thermo Fisher Scientific, Waltham, MA). Plasmids myo-2::gfp and rab-3::gfp were created in a similar manner, using the myo-2 or rab-3 promoters, respectively, in place of the myo-3 promoter. For expression of GFP-tagged LEV-11 isoforms in body-wall muscle, lev-11A, C, and E cDNAs each appended at the 5’ end with a Gly-Ala-Gly-Ala-Ser-coding linker sequence and flanked by NheI sites using PCR, and inserted into plasmid myo-3::gfp by In-Fusion cloning, generating plasmids pCP-GFP-lev-11A, pCP-GFP-lev-11C, and pCP-GFP-lev-11E, respectively. All construct sequences were verified through Sanger sequencing (Genewiz, South Plainfield, NJ).

Worms were maintained under standard growth conditions at 20°C on nematode growth medium (NGM) plates with OP50–1 Escherichia coli food (Brenner, 1974). Age synchronization was done by dissolving gravid adults in 1:2 ratio reagent grade bleach:5 M NaOH to liberate embryos, and hatching embryos overnight at 20°C in M9 buffer (Ausubel et al., 1992).

Worms were grown in liquid culture for IP-MS experiments. Briefly, ten adult worms were bleached to release embryos, which were grown at 20°C on NGM plates with OP50–1 until they had reached adulthood. Twenty adult worms were then moved to five new plates and grown at 20°C. After 84 hr, worms were washed off plates with sterile M9 buffer and transferred into a 2.8 L beveled flask containing OP50–1 bacteria suspended in 500 mL of S-Complete medium (Sulston & Hodgkin, 1988), and grown at 20°C while shaking (~ 200 rpm). When the majority of worms in the liquid culture had become adults, sucrose gradient centrifugation was applied to the culture to separate the live adult worms from the debris. Worms were washed with ice-cold sterile M9 buffer then 0.1 M NaCl solution, and then treated with alkaline bleach solution for 5–7 min to release embryos. Embryos were washed with ice-cold sterile water and allowed to hatch in M9 buffer overnight at 20°C. The next day, all newly hatched and starved L1-stage worms were transferred into a 2.8 L Flask containing OP50–1 bacteria in 500 mL of S-Complete medium and allowed to grow as an age-synchronized culture at 20°C until they reached the desired developmental stage.

Worms were treated for RNAi against short LEV-11 isoforms as we have done previously for other genes (Hegsted et al., 2019). Briefly, wild type worms were grown for three generations on IPTG-induced HT115 E. coli (Wang and Barr, 2005) expressing dsRNA from plasmids L4440-lev-11C and L4440-lev-11E.

The full genotypes of all worm strains used in this study are listed in Supplementary Table S4. Strains N2 [wild type] (RRID:WB-STRAIN:WBStrain00000001), CB1467 [him-5(e1461)] (RRID:WB-STRAIN:WBStrain00004339) that has a high incidence of male progeny, and KR2151 [hIn1[unc-54(h1040)]] (RRID:WB-STRAIN:WBstrain00023951) bearing a chromosomal inversion on chromosome I, were obtained from the Caenorhabditis elegans Genetics Center (University of Minnesota, Twin Cities, MN). Worm strains XA8001 [fhod-1(tm2363)] (RRID:WB-STRAIN:WBStrain00049243) bearing a putative null allele of fhod-1 (Mi-Mi et al., 2012), DWP3 [qaIs8001[fhod-1::gfp]] bearing a GFP-tagged fhod-1 transgene (Mi-Mi et al., 2012), and RSL62 [atn-1(ftw35 atn-1::mCherry::ICR::GFPnls)] tagged with mCherry at the endogenous atn-1 locus (Kimmich et al., 2024) were previously isolated. For production of male animals bearing fhod-1(tm2363) for genetic crosses, XA8001 was crossed with CB1467 to generate XA8032 [fhod-1(tm2363); him-5(e1461)].

Strategies for our first two CRISPR-Cas9-mediated mutagenesis targets, lev-11 exons 5a and 5c, were developed and conducted in collaboration with InVivo Biosystems (Eugene, OR), resulting in alleles lev-11(ups174), called lev-11(C-) in the text, which introduces a three-frame stop and novel XhoI site into exon 5c (targeting LEV-11C), and lev-11(ups175)/lev-11(ETU-), which introduces a three-frame stop into and eliminates an endogenous SalI site from exon 5a (targeting LEV-11E, T, and U). Worm strains DWP262, DWP263, and DWP264 are three independent isolates of lev-11(ups174), while strains DWP265 and DWP266 are two independent isolates that are heterozygous for lev-11(ups175). Crossing of DWP265 [lev-11(ups175)/(+)] with KR2151 produced the stably balanced heterozygous strain DWP267 [lev-11(ups175)/hIn1]. Additional CRISPR-Cas9 mutations targeting lev-11 were generated based on previously published protocols (Smith et al., 2020). For this, Cas9 protein (Integrated DNA Technologies, IDT, Coralville, IA), tracrRNA (IDT), crRNAs (IDT) and ssODNs (IDT) were injected into the syncytial gonad for homology-directed repair. Allele lev-11(ups184)/lev-11(CETU-) was generated by inserting a 3-frame stop and silently adding a BamHI site into exon 3b (targeting all four short isoforms LEV-11C, E, T, and U) using crRNA [AAU GUC GAA GGU AAA CAA GGA GG] and ssODN [ACA CCT CTG ACT TCT GCA GAA TGT CGA AGG TAA ACT AAA TAA ATA AAG GGA TCC ACA TCG CTT CTT GAT GTC CTC AAG AAG AAG ATG CG]. Two independent isolates of lev-11(184) were crossed with KR2151 to produce the stably balanced heterozygous strains DWP280 and DWP281 [lev-11(ups184)/hIn1]. Allele lev-11(ups182) was generated with the intent of disrupting short isoforms LEV-11C, T, and U and long isoform LEV-11D by inserting a 3-frame stop and silently adding a XhoI site into exon 9b, using crRNA [UAU UUC AGA ACU CCG UGA CGC GG] and ssODN [TAT AAA TAA CAG TGT GAC TCA ATA TTT CAG AAC TCT AAA TAA ATA AAC TCG AGA AAG CAC GCC AAT TAC AAG ATG AGC TTG ACC ATA T]. Strains DWP276, DWP277, and DWP278 are three viable independent isolates of lev-11(ups182). However, western blot analysis showed no dramatic loss of endogenous short LEV-11 protein (Fig.2B), possibly due to the close proximity of the mutation to the endogenous stop codon in exon 9b (Fig.1). The genomic sequences of all CRISPR-generated lev-11 alleles were verified by Sanger sequencing.

To express fluorescently tagged LEV-11 proteins, adult hermaphrodites were injected in the syncytial gonads with plasmids for expression of GFP- or mCherry-tagged LEV-11 isoforms and for expression of co-injection markers for visual identification of transgenic progeny or individual muscle cells. For over expression of GFP-tagged LEV-11 isoforms used to validate antibodies for western blot analysis (Fig.2A), N2 worms were injected with 1 ng/μl pCP-GFP-lev-11A, C, or E, together with 90 ng/μl pRS315 (Sikorski & Hieter, 1989) carrier DNA, and 10 ng/μl pGH8 (neuronal mCherry expression; RRID:Addgene_19359), 5 ng/μl pCFJ104 (body wall muscle mCherry expression; RRID:Addgene_19328), and 2.5 ng/μl pCFJ90 (pharyngeal mCherry expression; RRID:Addgene_19327) co-injection markers. Resulting strains DWP112 [upsEx62[myo-3p::gfp::lev-11E]], DWP115 [upsEx65[myo-3p::gfp::lev-11A]], and DWP120 [upsEx70[myo-3p::gfp::lev-11C]] express high levels of GFP-tagged LEV-11E, LEV-11A, and LEV-11C, respectively, in body wall muscle. To compare the distribution of GFP-tagged LEV-11 isoforms to dense bodies in body wall muscle (Fig.8A), RSL62 worms were injected with 1 ng/μl pCP-GFP-lev-11A, C, or E, together with 90 ng/μl pRS315, 10 ng/μl rab-3::gfp, and 2.5 ng/μl myo-2::gfp, generating strains DWP243 [atn-1(ftw35[atn-1::mCH]); upsEx161[myo-3p::mCH::lev-11A]], DWP250 [atn-1(ftw35[atn-1::mCH]); upsEx168[myo-3p::mCH::lev-11C]], and DWP251 [atn-1(ftw35[atn-1::mCH]); upsEX169[myo-3p::mCH::lev-11E]], respectively. To drive pharyngeal expression of mCherry-tagged LEV-11E in homozygous lev-11(ups175) animals, strain DWP267 was injected with 2.5 ng/μl pKM3, 90 ng/μl pRS315, and 10 ng/μl rab-3::gfp plasmid, resulting in a viable homozygous lev-11(ups175) strain DWP275 [lev-11(ups175); upsEx178[myo-2p::mCH::lev-11E]]. Similarly, DWP280 [lev-11(ups184)/hIn1] was injected with 2.5 ng/μl pKM3, 37.5 ng/μl pRS315, 10 ng/μl rab-3::gfp plasmid, and 50 ng/μl myo-3::gfp plasmid, resulting in a viable homozygous lev-11(ups184) strain, DWP282 [lev-11(ups184); upsEx185[myo-2p::mCH::lev-11E]]. For comparison of FHOD-1::GFP and mCherry::LEV-11E distributions in pharyngeal muscle, DWP275 and DWP282 were each crossed with DWP3, resulting in strains DWP312 [lev-11(ups175); qaIs8001[fhod-1::gfp]; upsEx178[myo-2p::mCH::lev-11E]] and DWP313 [lev-11(ups184); qaIs8001[fhod-1::gfp]; upsEx185[myo-2p::mCH::lev-11E]], respectively. To examine the combined effects of loss of FHOD-1 and LEV-11E, DWP282 was crossed with XA8032, and progeny selected for loss of him-5(e1461) to generate DWP323 [fhod-1(tm2363) lev-11(ups184); upsEx185[myo-2p::mCH::lev-11E]].

Worms were grown in the liquid culture as synchronized populations until the majority of animals were at least larval stages L3 and L4 for late-stage larval extracts, or as asynchronous populations for mixed-stage extracts. Worms were suspended in 30% sucrose and immediately centrifuged at 1500 × g, 5 min to separate lower-density live worms floating at the top from denser debris that pellets to the bottom. Collected worms were washed with ice-cold M9 buffer and 0.1 M NaCl solution before storage at −80°C overnight. Frozen worm pellets were thawed on ice in a cold room and resuspended in two volumes of ice-cold homogenization buffer (50 mM NaCl, 20mM Tris-HCl, pH 8.0, 1mM EDTA, 1mM DTT, 1mM PMSF), drop-frozen in liquid nitrogen, and pulverized into fine powder using a pre-chilled coffee grinder. The crude homogenate was subject to sonication on ice (30% amplitude, 50% duty cycle; twelve times 15 sec on/45 sec off, with 2 min off after every fourth run) before centrifugation for 10 min at 15,000 rpm, 4°C using a pre-chilled TLA 120.2 rotor. Pellets were washed twice with an equal volume of homogenization buffer to remove most actin monomers, free actin filaments, and non-body wall tissues. Actin and myosin filaments stably incorporated into myofibrils of body wall muscle were then extracted using an equal volume of extraction buffer (0.6 M KCl, 20 mM Tris-HCl, pH 8.0, 5 mM ATP, 5 mM MgCl₂, 1 mM DTT, 1 mM PMSF, 0.2 mM EGTA). Extracted supernatant samples were clarified 10 min at 15,000 rpm, 4°C, and then flash frozen with liquid nitrogen and stored at −80°C.

200 μL of Affi-Prep Protein A beads (Bio-Rad, Hercules, CA) were washed four times with 1 mL of PBST (1xPBS, 0.1% Tween-20) and centrifuged at 6000 rpm (3500 × g), 30 sec before being resuspended in 1 mL of PBST buffer with an additional 100 μL (~100 μg) of affinity-purified FHOD-1 antibody DPMSP2 (Mi-Mi et al., 2012) (RRID:AB_3399731) and incubated overnight at 4°C with agitation. The following day, beads were washed three times with 1 mL of PBST, three times with 1 mL of 0.2 M sodium borate (pH 9.0), then resuspended in 900 μL of 0.2 M sodium borate (pH 9.0) with 20 mM final concentration of fresh cross-linker, dimethyl-pimelimidate, and gently rotated 30 min at room temperature. The cross-linker was inactivated by washing beads with 1 mL of 0.2 M ethanolamine, 0.2 M NaCl (pH 8.5) twice before incubating with 1 mL of the same for 1 hr rotating, and storage in the same solution at 4°C overnight.

For mixed-stage and late-stage larvae extracts, two 100-μl aliquots of cross-linked FHOD-1 antibody-coupled beads were briefly washed 3 times with 1 mL of 0.5 M glycine (pH 2.5) to eliminate uncoupled antibody and reduce background precipitation, and then immediately washed three times with lysis buffer (100 mM KCl, 50 mM HEPES (pH 7.4), 1 mM EGTA, 1 mM MgCl₂, 0.5 mM DTT, 10% Glycerol, 0.05% NP-40). Two immunoprecipitation reactions were performed in parallel by incubation of 75 μL of beads with 1 mL of muscle enriched extracts of DWP3 [fhod-1(+) with fhod-1::gfp] or XA8001 [fhod-1(tm2363)] worms for 5 hr at 4°C. Beads were then washed three times with 1 mL of lysis buffer with EDTA-free protease-inhibitor cocktail (Roche, Indianapolis, IN), and then three times with 1 mL of lysis buffer without NP-40 or protease-inhibitor cocktail. Immunoprecipitated proteins were eluted from beads with sample buffer without reducing agent, and afterwards supplemented with DTT to a final 100 mM and stored at −80°C until use.

Since SDS-PAGE tends to fail at resolving low abundance proteins or proteins with extreme molecular weights, the entire precipitation for each group as well as pre-immunoprecipitation muscle extracts of mixed stage worms were analyzed by LTQ-Orbitrap Tandem Mass Spectrometry (Han et al., 2008) at the Arizona Proteomics Consortium (University of Arizona, Tucson, AZ). Mass spectrometry data from the Arizona Proteomics Consortium were analyzed using Scaffold Viewer software (Version 5.3.2, Proteome Software Inc.). Minimum characteristics for protein identification confidence were 99.9% probability of correct protein, and 3 minimum peptides per protein with 95% probability of correct peptide identification.

To generate antibodies specific for short LEV-11 isoforms, guinea pigs (gp27–57, gp27–58) were immunized against keyhole limpet hemocyanin-conjugated peptide K-V-N-K-E-G-A-Q-Q-T-S-L-L-D-V-L-K-K-K, corresponding to amino acids 3–21 shared by all short LEV-11 isoforms but not long LEV-11 isoforms. To generate antibodies specific for long LEV-11 isoforms, chickens (ch10–42, ch10–43) were immunized against similarly conjugated peptide A-I-K-K-K-M-Q-A-M-K-I-E-K-D-N-A-L-D-R-C, corresponding to amino acids 3–22 of all long isoforms. Peptide conjugation and immunizations were performed by the Pocono Rabbit Farm and Laboratory, Inc. (Canadensis, PA).

For antibody affinity purification and characterization, LEV-11A, LEV-11C, and LEV-11E were produced as recombinant glutathione-S-transferase (GST) fusion proteins from E. coli bearing pGEX-6p-3-lev-11A, pLMM19, or pLMM25, respectively, and released from GST by cleavage with GST-3CPro Precision protease, as described previously (Amin et al., 2007). For affinity purification of antibodies for short LEV-11 isoforms, 2.5 mg LEV-11C was resolved by SDS-PAGE and stained in-gel 3 min with Coomassie brilliant blue to allow excision of LEV-11C-containing portions of polyacrylamide gel. LEV-11C was electroeluted from gel portions into coupling buffer (0.1 M NaHCO₃, pH 8.3, 0.5 M NaCl). CNBr-Sepharose beads (1 g) were hydrated in water for 10 min before activation in 40 ml of 10 mM HCl, followed by further continuous wash with 210 ml of 10 mM HCl over a sintered glass filter, and a final wash with 40 ml of ice-cold coupling buffer. The activated and washed CNBr-Sepharose beads were resuspended in 8 ml of coupling buffer containing electroeluted LEV-11C, and mixed 2 hr at room temperature to allow coupling. LEV-11C-coated beads were then incubated over night at 4˚C in 0.2 M glycine, pH 8.0 to quench further CNBr-reactivity. Beads were then washed with 20 ml of ice-cold acetate buffer (0.1 M Na-acetate, pH 4.0, 0.5 M NaCl) followed by 20 ml of ice-cold coupling buffer. This sequence of washes was repeated four additional times. LEV-11C-coated beads were then suspended in 2 ml of gp27–57 crude antiserum with 5 mM NaN₃ overnight at 4˚C. Antibody-coated beads were washed with 40 mL of ice-cold coupling buffer, and then eluted with 6 mL of 0.1 M glycine-HCl, pH 2.5 as 500-μL fractions that were each immediately neutralized with 120 μL of Tris-HCl, pH 8.5. Fractions 4–8 were pooled as ~ 0.2 mg/ml affinity purified anti-LEV-11(Short) DPMSP13. These were supplemented to 5 mM NaN₃ and 2 mg/ml BSA for storage at −80˚C. LEV-11C-coupled beads were stored at 4˚C in coupling buffer with 5 mM NaN₃.

Western blot analysis (Fig.2) was conducted as previously published (Yingling & Pruyne, 2021). Sample sizes were normalized based on whole protein Coomassie brilliant blue stain intensity after SDS-PAGE. Affinity purified anti-LEV-11(Short) DPMSP13 and crude anti-LEV-11(Long) serum ch10–42 were diluted 1:200 in 1% milk/TBS/0.1% Tween 20 and incubated with primary antibody for 2 hr at room temperature. Images were acquired using a Bio-Rad ChemiDoc MP imager, and processed with Image Lab and Photoshop CS4.

F-actin staining of worms (Figures 4A, 5, 7A) was done as previously published (Mi-Mi et al., 2012). Immunostain (Fig.4B, C 7B, C, 8C) was done based on prior protocols (Finney & Ruvkun, 1990), but without the addition of spermidine-HCl and with a methanol concentration of 75% during fixation. Monoclonal antibody MH35 (anti-ATN-1) was developed by R.H. Waterston (Francis & Waterston, 1985) and generously provided as a gift by P. Hoppe (Western Michigan University, Kalamazoo, MI), and rabbit anti-GFP was a gift from Anthony Bretscher (Cornell University, Ithaca, NY). Primary antibody dilutions were MH35 1:10⁴, DPMSP13 affinity-pure anti-LEV-11(Short) 1:10, ch10–42 crude anti-LEV-11(Long) serum 1:100, anti-GFP 1:200, anti-mCherry 1:1000 (#632543 Takara Bio, San Jose, CA). Secondary antibodies donkey anti-Guinea pig Texas Red and goat anti-mouse FITC (Rockland Immunochemicals, Limerick, PA) were diluted 1:500.

Confocal images were acquired with an SP8 Laser Scanning Confocal Microscope (Leica, Wetzlar, Germany) driven by LAS X software (Version 3.5.2, build 4758; Leica) with a HCX Plan Apochromat 63X/NA 1.4 oil immersion lambda objective. Z-stack images were collected at 0.3 μm intervals for body wall muscle images and 0.2 μm for pharyngeal images. Images were deconvolved with Huygens Essential Software (Huygens compute engine 18.10.0, Scientific Volume Imaging B.V.), Classic Maximum Likelihood Estimation deconvolution algorithm, with 40 iterations and a signal to noise ratio of 20. Images were linearly processed to enhance contrast and false colored in Adobe Photoshop CS4 or CC2018 (Adobe, San Jose, CA). Deconvolved confocal z-stacks were used to generate X, Y, Z plane pharyngeal images using Imaris x64 software (version 10.1.0; Bitplane AG, Belfast, United Kingdom).

For analysis of mosaic pharynges (Fig.3E), age synchronized animals were fixed and mounted on slides. Pharynges were considered mosaic if at least one but not all of the major pharyngeal muscles (pm3–7) visibly express mCherry for DWP272 [lev-11(ETU-); Ex[mCH::lev-11E]] n = 86 L1 larvae, n= 70 day-1 adults and for DWP282 [lev-11(CETU-); Ex[mCH::lev-11E]] n = 105 L1 larvae, n = 75 day-1 adults.

Analyses of BWM cell width and striations (Fig.8A,B) were done as previously published (Kimmich et al., 2024). Age synchronized day-1 adult N2, XA8001 [fhod-1(Δ)], DWP262 [lev-11(C-)], DWP263[lev-11(C-)], and DWP264[lev-11(C-)] were fixed and stained with Alexa568-Phalloidin. 15–20 worms were analyzed per strain with two to four muscle cells measured per worm (one muscle cell per muscle). For consistency, only muscle cells within two to four muscle cells from the vulva were analyzed.

L1 larvae were placed on an NGM plate containing OP50–1. Using a Nikon SMZ18 fluorescence stereoscope (Nikon, Tokyo, Japan) equipped with a Sola light engine (Lumencore, Beaverton, OR), presence or absence of pharyngeal pumping in individual worms was determined by visual observation of the grinder bulb of the pharynx, followed by examination of the pharyngeal muscle under fluorescence illumination for mCherry expression (Fig.3F). 50–60 larvae were analyzed per strain in each of three independent trials. Movies of pharyngeal pumping (Movie 1, Movie 2) were captured using a Canon EOS Rebel 600D T3i camera with a Martin Microscope Company MM-SLR Adapter on a Nikon SMZ18 stereoscope. Movie files were edited and processed using iMovie (version 10.4.2, Apple, Cupertino, California).

To analyze viability (Fig.3A, C, D), young adult worms were set onto separate fresh NGM/OP50–1 plates and allowed to lay eggs for 3 hr. Eggs were counted on each plate. The next day (day 2), unhatched eggs (indicative of failed embryogenesis) and L1 stage larvae were counted. On days 2 – 4, viable worms were counted. For wild type N2 worms and strains for which we could isolate homozygous mutants, this analysis was performed with ten initial worms. For strains that we were unable to isolate homozygous animals (DWP267 and DWP280), analysis was performed with 40 adult progeny of heterozygous animals that, after the initial egg laying, were subject to single worm PCR to genotype and identify wild-type and heterozygous animals. Brood Size Analysis

To measure brood sizes (Fig.3B), ten L4 animals were set onto separate NGM/OP50–1 plates and allowed to lay eggs. Every 12 hr, worms were moved to fresh plates, and eggs were counted. This was repeated until 5 days had elapsed.

Analysis of variance was performed on data containing three or more groups. Unpaired T-tests were performed on comparisons of two groups. Differences for which p < 0.05 were considered statistically significant. Graphs were made and post hoc analyses conducted using Prism 10 (version 10.1.1; GraphPad Software, Boston, MA) with data presented as average ± standard deviation.