A mutation in F-actin polymerization factor suppresses the distal arthrogryposis type 5 PIEZO2 pathogenic variant in Caenorhabditis elegans

PIEZO proteins are cation channels that transduce mechanical stimuli into physiological responses. PIEZO1 and PIEZO2 trimers form a propellor-like structure with C-terminal domains of each PIEZO monomer assembling to create the central ion pore (Saotome et al., 2018; Wang et al., 2019). Dysfunction in both human PIEZO1 and PIEZO2 causes a variety of physiological disorders and diseases in humans (Alper, 2017). Gain-of-function mutations in PIEZO1 cause hereditary xerocytosis/dehydrated stomatocytosis (DHS) in red blood cells (Zarychanski et al., 2012; Albuisson et al., 2013; Glogowska et al., 2017; Andolfo et al., 2013), while loss-of-function mutations in PIEZO1 could lead to congenital lymphatic dysplasia (Fotiou et al., 2015). In addition, some PIEZO2 gain-of-function and loss-of-function mutations cause pathologies in distal limb and neuronal systems, which lead to joint contractures, muscle atrophy and proprioception deficits (Chesler et al., 2016; Coste et al., 2013; Ma et al., 2023). These pathogenic effects may occur due to insufficient or excessive responses of PIEZO channels to mechanical stimuli. At the cellular level, dysfunctional PIEZO channels may disrupt intracellular calcium homeostasis and interfere with downstream calcium-dependent signaling pathways (Alper, 2017; Beech and Kalli, 2019). Despite the electrophysiological characterization of PIEZO channel variants, at the functional level, the molecular and cellular mechanisms of PIEZO-associated diseases remain elusive.

The C. elegans spermatheca is a multicellular tube that can stretch and dilate. Oocytes enter the spermatheca through a distal neck, are fertilized within a central bag, and exit the spermatheca through the spermatheca-uterine (sp-ut) valve. The spermathecal valves and bag-cells are spatiotemporally coordinated to control oocyte entry during ovulation, as well as oocyte exit after fertilization in the spermatheca. The contractability of the spermathecal tissue is driven by the coordinated activation of myosin (Kelley and Cram, 2019; Kelley et al., 2020a,b). After ovulation, the fertilized oocytes are expelled from the spermatheca and into the uterus, pushing many sperm with them. These sperm must navigate back into the spermatheca for subsequent ovulations. The oocytes and somatic sheath cells secrete chemoattractant molecules and prostaglandins into the uterus to guide the sperm back to the spermatheca (Edmonds et al., 2010; Hoang et al., 2013). The pezo-1 gene is the sole ortholog of human PIEZO1 and PIEZO2 in C. elegans. Dysfunctional PEZO-1 resulted in reduced brood size and low ovulation rate, as well as sperm guidance and navigational defects (Bai et al., 2020a). Cytoskeletal components play a crucial role in coordinating the contractibility of the spermathecal cells (Kelley and Cram, 2019; Kelley et al., 2020a,b). Furthermore, PIEZO channels activity is regulated by the cytoskeleton integrity (Atcha et al., 2021; Wang et al., 2022; Shmukler et al., 2015; Ellefsen et al., 2019; Eijkelkamp et al., 2013; Pardo-Pastor et al., 2018). Based on these findings, we hypothesize that the PIEZO channel plays a functional role during ovulation and contractility within the spermathecal tissue by regulating cytoskeletal activity, specifically actin polymerization. However, the exact role of PEZO-1 channel in governing the cytoskeletal activity during ovulation remains elusive.

In this study, we established a disease model of human PIEZO channels by generating orthologous variants in the C. elegans pezo-1 gene using CRISPR/Cas9 gene editing. All homozygous pezo-1 pathogenic variants reduced brood size and ovulation rate. Two of the mutants, pezo-1(R2405L) and pezo-1(R2405P), caused sperm attraction defects, resulting in some of the sperm failing to migrate back to the spermatheca. Using a forward genetic screen, we found a suppressor allele in the gene gex-3 that belongs to NCKAP1 (NCK-associated protein 1), a subunit of Wiskott-Aldrich syndrome protein (WASP)-verprolin homologous protein (WAVE/SCAR) complex. Depletion of gex-3 by RNAi and the suppressor allele gex-3(L353F) alleviated the small brood size and low ovulation rate, as well as the crushed oocyte phenotypes of the pezo-1(R2405P) mutant. Using an auxin-inducible degradation (AID) system, we drove tissue-specific degradation of GEX-3 in the pezo-1(R2405P) background. Reduced brood size was only restored in the somatic-specific degradation strain, suggesting that the loss of gex-3 in somatic tissues primarily contributed to the suppression of the pezo-1 gain-of-function mutation (R2405P). Finally, we found that actin organization was disrupted and distorted in the pezo-1 mutants, and that gex-3(L353F) partially alleviated actin defects in the pezo-1(R2405P) background. Overall, we used C. elegans to study the pathophysiological effect of an ortholog of human PIEZO2 mutation causing distal arthrogryposis type 5 (DA5) and identified the putative suppressor gex-3(L353F). Here, we provide new insights into the physiological roles of PIEZO channels and their interaction with cytoskeletal organizational elements.

To study underlying molecular mechanisms of PIEZO-associated diseases, we used CRISPR/Cas9 gene editing to generate four disease-relevant PIEZO alleles in the C-terminal domains of C. elegans pezo-1, the sole worm ortholog of human PIEZO1 and PIEZO2 (Fig. 1A). Each of the alleles were orthologous mutations associated with three different diseases: dehydrated hereditary stomatocytosis (DHS), distal arthrogryposis subtype 5 disorder (DA5) and Gordon syndrome (GS) (Table 1). All homozygous pezo-1 mutations caused a 27-81% reduction in brood size compared with the wild-type control (Fig. 1B). In particular, non-conservative amino acid residue mutations pezo-1(R2405P) and pezo-1(R2405L) decreased brood size by 70%, whereas a more conservative (polar) mutation, pezo-1(R2405Q), decreased the brood size by ∼20% (Fig. 1B).

Our previous study demonstrated that the reduced brood size in pezo-1 mutants was due to lower ovulation rate as well as oocyte crushing while exiting the spermathecal-uterine (sp-ut) valve (Bai et al., 2020a). To test whether ovulation was affected in each of the current pezo-1 mutants, we performed live imaging to record the ovulation process for at least 1 h and analyzed the ovulation performance in each mutant. Indeed, we determined a low ovulation rate in all tested pezo-1 mutants Fig. 1C,D). Of note, the ovulation rate was reduced to nearly zero in older animals (day-2 adults), whereas wild-type controls continued to ovulate ∼1.83 times per hour per gonad arm (Fig. 1D). We also observed that ∼30% of the ovulated oocytes from day 1 pezo-1(R2405L/P/Q) variants were crushed when transiting through the sp-ut valve, whereas no crushed oocytes were observed in the wild-type animals (Fig. 1E). Overall, our data suggest that pezo-1 pathogenic variants caused a reduction in brood size, low ovulation rates and crushed oocytes.

Various pezo-1 null mutants exhibited defective sperm guidance (Bai et al., 2020a). Either self-sperm or male sperm failed to navigate back to the spermatheca after each ovulation, thus depleting the spermathecal sperm. To test whether each pezo-1 missense mutant fails to attract sperm back to the spermatheca, male sperm navigational performance was assessed in vivo by staining wild-type male sperm with a vital fluorescent dye, MitoTracker CMXRos, which efficiently stains sperm mitochondria in live animals (Hoang et al., 2013). The stained wild-type males were placed for 30 min together with the label-free pezo-1 mutant hermaphrodites. We isolated the hermaphrodites and allowed the sperm to navigate through the uterus to the spermatheca for 1 h. We quantified the sperm distribution by counting the number of fluorescent-labeled sperm in each zone (Fig. 2A-G) (Bai et al., 2020a). In wild-type hermaphrodites, 90.3% of fluorescent sperm navigated through the uterus and accumulated in the spermatheca (Fig. 2B,B′,H). The sperm distribution rate in the pezo-1(T1997M), pezo-1(R2373H) and pezo-1(R2405Q) hermaphrodites were identical to wild-type worms [93.3% for pezo-1(T1997M), 98.2% for pezo-1(R2373H) and 79.4% for pezo-1(R2405Q)] (Fig. 2C-D′,G-H). However, in the pezo-1(R2405L) and pezo-1(R2405P) reproductive tracts, the fluorescence-labeled sperm displayed defective sperm navigation. Only 56.1% and 57.5% of the labeled sperm in the pezo-1(R2405L) and pezo-1(R2405P) hermaphrodites, respectively, were able to migrate back to zone 3, which was the zone adjacent to and including the spermatheca (Fig. 2E-F′,H); the rest of the labeled sperm failed to reach the spermatheca and remained throughout zones 1 and 2, which are furthest from the spermatheca (Fig. 2E-F′,H).

The pezo-1(R2405P) allele caused the most severe defects in the C. elegans reproductive tract (Figs 1 and 2). This mutation is equivalent to the mouse PIEZO1 R2514 mutation, which resides at the end of the α1 helix of the C-terminal domain. We determined that the longest isoform of pezo-1 (isoform G; wormbase.org v. WS280) encodes a mechanosensitive ion channel (Millet et al., 2022). Of note, the behavior observed in the pezo-1(R2405P) mutant was similar to various pezo-1 null mutants previously described (Bai et al., 2020a), suggesting that loss- or gain-of-function pezo-1 mutations are equally pathogenic, similar to their human counterparts (Chesler et al., 2016; Coste et al., 2013; Ma et al., 2023). To determine the effect of pezo-1(R2405P) on PEZO-1 channel function, we heterologously expressed C. elegans wild-type and pezo-1(R2405P) mutant constructs in a Spodoptera frugiperda (Sf9) cell line and measured their mechanical response in the whole-cell patch-clamp configuration while stimulating with a piezo-electrically driven glass probe (Fig. 3A-C). Current densities for the pezo-1(R2405P) mutant were larger than non-infected cells, although not significantly different from wild-type pezo-1 (Fig. 3D). Noteworthy, substituting an arginine with a proline at position 2405 results in mechanocurrents that become inactivated more slowly than the wild-type channel, as reflected by the large time constant of inactivation (Fig. 3E). We also determined that the R2405P construct requires less mechanical stimulation to open than wild-type pezo-1 (Fig. 3F). Taken together, our results suggest that pezo-1(R2405P) is a gain-of-function allele.

The small brood size in the pezo-1 mutants was a simple readout to perform a forward genetic screen to isolate genetic suppressors of PIEZO disease mutations. To better understand the mechanism by which PEZO-1 regulates nematode reproduction, we conducted a genetic suppressor screen of the gain-of-function mutant pezo-1(R2405P), which has the lowest brood size of all the alleles we tested. Previously, we showed that pezo-1 interacted genetically with an ER calcium regulator SERCA pump (sca-1) (Bai et al., 2020a). To increase the sensitivity of our forward screening, we shifted the synchronized EMS-treated F2 population of pezo-1(R2405P) embryos to sca-1 RNAi food and maintained the animals at 25°C until the candidate suppressor lines were isolated (Fig. S1A). Under these sensitive screening conditions, wild-type hermaphrodites produced viable animals with a significant reduction in brood size (Fig. S1B), while sca-1 RNAi treatment of our pezo-1(R2405P) mutant led to sterility or extremely low brood sizes (greater than 80% reduction compared with wild-type worms) after one or two generations (Fig. S1B). We screened ∼150,000 haploid genomes and isolated one stable suppressor line that partially restored the reduced brood size (n=161.1) when compared with pezo-1(R2405P) or pezo-1(R2405P) on sca-1 RNAi food for one generation (n=87.4 and n=34.6, respectively) (total brood size=34.6) (Fig. S1B).

Using a high-throughput genomic mapping strategy, which involves the molecular inversion probes genomic mapping (MIP-MAP) strategy (Mok et al., 2017), we identified genetic modifiers in the suppressor line sup1 (Fig. S2). To maintain the pezo-1(R2405P) allele during the mapping process, we generated pezo-1(R2405P) in the mapping strain VC20019 background by CRISPR/Cas9 gene editing and named the strain pezo-1(R2405P^MM) (Fig. S2A) (Mok et al., 2017). This strain appeared to have similar phenotypes to that of our original pezo-1(R2405P) mutant. We carefully pooled suppressed F2 progeny from the cross and allowed the F2 population to expand for ten generations for MIP-MAP analysis (Fig. S2A). As the flanking regions of the modifiers originated from a pezo-1(R2405P) background, the occurrence frequency of the MIP-MAP probes at the loci responsible for suppression would drop to nearly zero (Mok et al., 2017). Theoretically, reading the frequency of MIP-MAP probes would allow us to narrow down the target regions bearing the putative mutation. After MIP-MAP analysis and whole-genome sequencing, we mapped a single genomic region in the suppressor line pezo-1(R2405P); sup1 (Fig. S2B), which contained candidate genes located on chromosome IV (9.1-12.0 Mb) (Fig. S2).

Assuming the candidate suppressors are loss-of-function alleles, we used RNAi to deplete expression of these candidate genes in the pezo-1(R2405P) mutant. From a RNAi trial experiment, we identified a candidate modifier gene, gex-3, on chromosome IV (Fig. 4A and Fig. S2B). gex-3 is the ortholog of human NCKAP1 (NCK-associated protein 1), one of the five WAVE regulatory complex subunits, which regulates the formation of the actin cytoskeleton via the Arp2/3 complex (Patel et al., 2008). To assess the knockdown efficiency of gex-3 RNAi, we first fed gex-3 (RNAi) bacteria to mNG::GEX-3 animals and quantified the fluorescence intensity of mNG::GEX-3 after 24 h of RNAi treatment (Fig. S3) (Dickinson et al., 2017). The mNG::GEX-3 levels in the germline and spermatheca were dramatically reduced when compared with RNAi-negative control (Fig. S3A-C), suggesting that gex-3 RNAi effectively knocked down gex-3 gene expression in vivo. We then depleted gex-3 by RNAi in wild-type worms and all pezo-1 pathogenic mutants (Fig. 4A,B). gex-3(RNAi) caused significant embryonic lethality in all tested strains (Fig. 4B); however, most embryos did hatch. gex-3(RNAi) also led to reduced brood sizes in both wild-type worms and the pezo-1(T1997M) mutant (Fig. 4A). The brood size in other tested mutants, including pezo-1(R2373H), pezo-1(R2405L) and pezo-1(R2373Q), were not significantly affected by gex-3(RNAi). Of note, the reduced brood size was significantly restored in pezo-1(R2405P) (n=113.1) by gex-3(RNAi) (n=179.3) (Fig. 4A). Additionally, depletion of gex-3 by RNAi caused reduced embryonic lethality in the pezo-1(R2405P) mutant when compared with gex-3 (RNAi) in the wild-type worms, indicating that pezo-1(R2405P) moderately suppressed the phenotype caused by gex-3 (RNAi) (Fig. 4B). Furthermore, live imaging indicated that treatment with gex-3 (RNAi) resulted in similar ovulation rates but fewer crushed oocytes in the pezo-1(R2405P) background (Fig. 4C-E). Using RNAi, we next tested whether other WAVE regulatory complex subunits suppressed the defects in the pezo-1(R2405P) mutant by depleting abi-1 and wve-1, orthologs of human Abelson interactor gene (ABI1) and human Wiskott-Aldrich syndrome protein family member 1 gene (WASF1), respectively, using RNAi. However, the smaller brood size of the pezo-1(R2405P) mutant is not restored with either RNAi treatment (Fig. S4A,B).

We used a CRISPR knock-in green fluorescent reporter mNeonGreen::GEX-3 (mNG::GEX-3) to determine the subcellular localization of GEX-3 in C. elegans (Dickinson et al., 2017). mNG::GEX-3 was widely expressed from embryonic stages to adulthood (Fig. 5). Of note, the mNG::GEX-3 was strongly expressed in various tissues subjected to mechanical stimulation, including the pharyngeal-intestinal valve and spermatheca (Fig. 5A-B‴). Notably, mNG::GEX-3 also co-localized with mScarlet::PEZO-1 at the pharyngeal-intestinal valve (Fig. 5A-A‴), spermathecal membrane (Fig. 5B-B‴) and early embryonic membranes (Fig. 5C-C‴).

After confirming that the of loss of function of gex-3 suppressed the phenotypes in pezo-1(R2405P) animals, we generated a candidate suppressor allele by changing Leu353 to phenylalanine by CRISPR/Cas9 gene editing on the wild-type background. This allele was renamed as gex-3(L353F). To assess its suppression during reproduction, this newly generated allele was then introduced onto the pezo-1(R2405P) background by CRISPR/Cas9 gene editing. The predicted GEX-3 structure from AlphaFold suggests that the residue Leu353 was located at the N terminus of a helical segment (Fig. 6A) (Varadi et al., 2022; Jumper et al., 2021). The homozygous gex-3(L353F) strain exhibited a smaller brood size when compared with wild-type worms (Fig. 6B). Unlike gex-3(RNAi) treatment, there was no embryonic lethality observed in the gex-3(L353F) mutant generated by CRISPR/Cas9 (Fig. 6C). To determine whether this missense mutation disrupted the temporal and/or spatial expression pattern of GEX-3, we generated the gex-3(L353F) allele using CRISPR/Cas9 on the mNG::GEX-3 reporter strain, which was named mNG::gex-3(L353F). Endogenously tagged mNG::GEX-3 was expressed in multiple C. elegans tissues, including embryos, the spermatheca, the germline and pharynx (Fig. 5, Fig. S5A,C). Similar localization patterns were found in the mNG::gex-3(L353F) mutant (Fig. S5B,D). Overall, these data suggest that the gex-3(L353F) allele displays a weak effect, which may only partially compromise GEX-3 protein function, without altering the trafficking and cellular localization of GEX-3 in vivo.

The gex-3(L353F); pezo-1(R2405P) double mutant rescued the reduced brood size and subtle embryonic lethality observed in the pezo-1(R2405P) mutant alone (Fig. 6B,C). In addition, the double mutant reduced the crushed oocyte rate (from 31.4% to 11.1%) and slightly increased the ovulation rate to 0.25±0.62 ovulations per hour in the day 2 adults, when compared with pezo-1(R2405P) alone (0.0±0.0 UNITS) (Fig. 6D-F). The double mutant displayed a higher ovulation rate in day 1 animals (1.57±0.84 UNITS) than gex-3(L353F) alone (0.76±0.60 UNITS) (Fig. 6D), suggesting that pezo-1(R2405P) and gex-3(L353F) mutually suppress one another's reproductive defects.

The small brood size of the pezo-1(R2405P) mutant might arise from the high number of sperm that fail to navigate back to the spermatheca. To test whether the gex-3(L353F) allele restored sperm attraction back to the spermatheca, we assessed the MitoTracker-stained male sperm navigational performance in both gex-3(L353F) and gex-3(L353F); pezo-1(R2405P). In gex-3(L353F) hermaphrodites mated with stained wild-type males, over 90% of fluorescent sperm navigated through the uterus and accumulated in the spermatheca (Fig. S6A,A′,C) within 1 h of mating. The sperm distribution rate in the double mutant hermaphrodites was 43.9%, which was similar to the pezo-1(R2405P) single mutant (57.5%, Fig. 2G), suggesting that gex-3(L353F) had no effect on sperm attraction behavior. Overall, our genetic interaction data suggest that both gex-3 (RNAi) and the gex-3(L353F) mutant could partially suppress the reproductive defects of the pezo-1(R2405P) strain, but only at the ovulation and brood size levels. Meanwhile, the pezo-1(R2405P) allele also partially alleviated the embryonic lethality (Fig. 4B) and increased low ovulation rate (Fig. 4C) caused by gex-3 (RNAi), suggesting that pezo-1(R2405P) and gex-3 (RNAi) are mutual suppressors.

GEX-3 is expressed in all reproductive tissues, including the spermatheca, germline, oocytes, and embryos (Fig. 5B,C). To better understand the role of GEX-3 in suppressing the subfertility of the pezo-1(R2405P) mutant, we used an auxin-inducible degradation system (AID) to degrade GEX-3 in somatic tissues or the germ line (Zhang et al., 2015). We knocked in a cassette with degron and GFP coding sequence at the gex-3 N-terminus using CRISPR/Cas9 gene editing (named AID::GFP::GEX-3) so that the degron strain could be visualized and quantified by a GFP fluorescent signal (Fig. S7). The somatic- and germline-specific AID strains were driven by Peft-3 and Pmex-5, respectively (Ashley et al., 2021), and activated when the animals were exposed to 2 mM auxin (indole-3-acetic acid, IAA). The GFP fluorescence intensities in the induced animals were significantly reduced compared to vehicle control (Fig. S7B-B″,D-D″,E,F). The strain expressing the degron interactor transgene Pmex-5::tir-1::BFP::AID led to a 2-3 fold reduction in fluorescence intensity of AID::GFP::GEX-3 in the germline and oocytes (Fig. S7B,B″,E), however, the intensity was not affected in the somatic tissues (Fig. S7E). The Peft-3::tir-1::BFP::AID led to an approximate 1.5 fold reduction of fluorescent intensity of AID::GFP::GEX-3 in the sheath, spermathecal cells, and germline (Fig. S7D,D″,F), suggesting the somatic degron strain affected the AID::GFP::GEX-3 in both somatic and germline cells.

To assess the tissue-specific suppression of GEX-3 in pezo-1(R2405P), we introduced the pezo-1(R2405P) allele into each degron strain by CRISPR/Cas9 gene editing. We exposed L4 animals to either 0.25% ethanol as a control or 1-2 mM IAA, then brood sizes and embryonic lethality rate were determined 0-60 h post-L4 (Fig. 7A,B). Interestingly, the brood size of the pezo-1(R2405P) mutant was significantly restored in the somatic tissue-specific AID::GFP::GEX-3 strain (Fig. 7A). Meanwhile, there were no significant changes in brood size in germline-specific AID::GFP::GEX-3 driven by the Pmex-5 promoter (Fig. 7A). Additionally, depletion of GEX-3 in both somatic and germline tissues led to severe embryonic lethality (Fig. 7B). Unlike the partial rescue of embryonic lethality of gex-3 (RNAi) by pezo-1(R2405P), embryonic lethality was close to 100% for pezo-1(R2405P); AID::GFP::GEX-3 strains, regardless of the promoter used for the tir-1::BFP::AID cassette (Fig. 7B). Therefore, degradation of GEX-3 in the somatic tissue suppressed the small brood size of pezo-1(R2405P), likely owing to the role of gex-3 in the spermatheca or sheath contraction. This seems likely as we observed an improvement in ovulation performance when combining gex-3 RNAi and gex-3(L353F) with pezo-1(R2405P) (Figs 4 and 6).

The WAVE regulatory complex is essential for the organization of the actin cytoskeleton and its dynamics. Therefore, we predicted that gex-3 may coordinate with pezo-1 actin organization and orientation in the spermatheca. Previous studies indicated that spermathecal contractility was tightly associated with proper actin organization (Kelley and Cram, 2019; Kelley et al., 2020a,b). To further investigate the functional contribution of pezo-1 and gex-3 to spermathecal contractility and actin organization, we used a spermatheca-specific GFP marker GFP::ACT-1 for labeling actin. In mature wild-type animals, the entire actin network was tightly compacted in the contracting spermathecal cells (Fig. 8A,A′). In the dilated spermatheca occupied by an ovulating oocyte, each spermathecal cell contains prominent parallel actin bundles at the cellular cortex that run along the basal cell edges; these are known as peripheral actin bundles (Fig. 8B,B′). To test whether actin organization was affected by pezo-1 and gex-3, we crossed the actin reporter strain into the following four strains: (1) pezo-1(R2405P), (2) pezo-1 full deletion mutant pezo-1Δ, (3) gex-3(L353F) and (4) the pezo-1(R2405P) gex-3(L353F) double mutant. We found a variety of defects in actin bundle distribution and orientation in the spermathecal cells of the pezo-1(R2405P) and pezo-1Δ mutants (Fig. 8C-E′,H,I). In these cells with actin bundle orientation defects, the actin bundles within the cells were aligned, but ran perpendicular to the long cell axis, termed perpendicular actin (Fig. 8C-D′,I). We also observed aligned actin that was densely associated with thicker and brighter actin bunches in pezo-1(R2405P) and pezo-1Δ, which were defined as bunched actin (Fig. 8D-E′,H) (Wirshing and Cram, 2018). The gex-3(L353F) mutant caused only mild actin defects, including bunching and perpendicular defects (Fig. 8F,F′,H). The pezo-1(R2405P) gex-3(L353F) double mutant significantly rescued bunching actin defects but did not affect perpendicular defects (Fig. 8G,G′,I). Collectively, these results indicate that pezo-1 is crucial to spermathecal contractility, likely through influencing actin cytoskeletal organization and orientation. gex-3(L353F) partially suppressed the actin defects that may contribute to the alleviated ovulation defects and brood size (Fig. 9).

The PIEZO proteins are associated with at least 26 human disorders and diseases (Alper, 2017). More than 100 variants of PIEZO have been identified to cause physiological disorders. Therefore, it is crucial to understand the molecular mechanism whereby dysfunctional PIEZO alters physiological processes, as well as to identify molecular and genetic determinants that may affect PIEZO activity. Complete knockout of Piezo1 and Piezo2 in mice causes embryonic lethality and fetal cardiac defects (Ranade et al., 2014; Nonomura et al., 2017). The PIEZO mutants in other systems, such as Drosophila or zebrafish, only lead to mild phenotypes (Kim et al., 2012; Duchemin et al., 2019; Shmukler et al., 2015), which limits the possibility of performing forward genetic screens to identify genetic determinants of PIEZO in those systems in vivo. Our previous study demonstrated that PEZO-1 channel influenced a series of reproductive processes, including ovulation, the expulsion of the fertilized oocyte into the uterus and sperm navigation (Bai et al., 2020a). Dysfunction of PEZO-1 causes a severe reduction in the ovulation rate, defective sperm navigation behavior and small brood size. The severe reduction in brood size in the pezo-1 mutants provided an easy and reproducible readout for a forward genetic screen.

Caenorhabditis elegans is a powerful model in which gene editing and behavior are becoming an attractive system for precision modeling of human genetic diseases. In this study, we tested five PIEZO pathogenic variants in the C. elegans pezo-1 gene, all of which displayed similar ovulation phenotypes as our pezo-1 deletion mutants, suggesting that these alleles compromise PEZO-1 function and/or channel activity. All five variants were localized to the predicted pore of the PEZO-1 channel (Fig. 1A). The conserved human variants cause diseases in various organs or tissues, such as DHS in red blood cells or DA5, and Gordon syndrome in joint and distal extremities (Alper, 2017) (Table 1). The phenotypes observed in C. elegans do not reflect these PIEZO-derived disease symptoms, yet our disease modeling pipeline demonstrated the usefulness of the C. elegans reproductive tract for investigating the physiological contribution and molecular mechanisms of these PIEZO-based diseases mutations.

The severe reduction in brood size of the pathogenic mutants allowed us to screen putative genetic determinants for PIEZO suppressors in vivo. Our chemical mutagen-mediated forward genetic screening combined with MIP-MAP genomic mapping facilitated the discovery of the suppressor alleles. In this study, we successfully identified the cytoskeletal regulator WAVE/SCAR complex subunit GEX-3, which suppressed the defective phenotypes caused by the gain-of-function pezo-1(R2405P) allele. To our knowledge, this is the first genetic suppressor of PIEZO that has been identified through a forward genetic screening approach.

Actin-binding and regulatory proteins are crucial for proper spermathecal contractility and actin organization, which are necessary for achieving proper ovulation (Wirshing and Cram, 2018). Our actin imaging data revealed prominent parallel actin bundles, referred to peripheral actin bundles, at the basal cell surface (Fig. 9). In contrast, pezo-1 mutants disrupted and distorted the actin organization and orientation in the spermatheca during ovulation (Fig. 9), likely contributing to the observed contractility and ovulation defects, such as crushed oocytes. PIEZO channel function relies on the effective communication and physical interactions between the channels and cytoskeletal components, including those between actin filaments and/or focal adhesions and the extracellular matrix (Wang et al., 2022; Ellefsen et al., 2019; Romero et al., 2020, 2023; Nourse and Pathak, 2017).

GEX-3 is a component of the WAVE complex, which controls actin cytoskeletal organization and dynamics by triggering the activity of the actin polymerization regulator Arp2/3 complex (Chen et al., 2010). Loss of gex-3 resulted in a disrupted actin cytoskeleton during embryogenesis and axon migration in C. elegans (Bernadskaya et al., 2012). Based on our proposed model (Fig. 9), the GEX-3 suppressor allele could affect actin polymerization and actin organization in a way that partially alleviates the actin defects caused by the PEZO-1 gain-of-function mutation, thereby mitigating mechanotransduction deficiencies and improving spermathecal contractility during ovulation. These findings are consistent with recent studies showing that PIEZO activity could enhance actin polarization by physically tethering to the cadherin-β-catenin mechanotransduction complex (Atcha et al., 2021; Wang et al., 2022) (Fig. 9). Overall, the link between PEZO-1 and the actin cytoskeleton is likely part of a proposed feedback mechanism (Nourse and Pathak, 2017), in which the activation of PEZO-1 influences the dynamics and formation of cytoskeletal components, while the cytoskeleton affects PEZO-1 activation during mechanical activities such as ovulation.

In addition, the suppressor gex-3 that we identified is likely to be pezo-1(R2405) allele specific, as it did not suppress the other four disease alleles generated in this study. We also could not rule out the possibility of an additional suppressor allele within the suppressor background, as we observed that the CRISPR-edited gex-3 suppressor strain had weaker effects compared wit the original suppressor strain. Using this genetic screening pipeline will allow us to identify more suppressors in other pathogenic mutants, as well as uncover the molecular mechanism and functional contribution of each pathogenic allele to PEZO-1 activity. These screens may aid in defining the cellular mechanism that modulates PIEZO channel function and pave the way for further therapeutic approaches.

Finally, the human ortholog of gex-3, NCKAP1, plays a critical role in neurodevelopment. Disruptive variants of NCKAP1 have been associated with neurodevelopmental disorders, including Coffin-Siris syndrome 1 (CSS1) and autism spectrum disorder (ASD) (Guo et al., 2020; Stelzer et al., 2016). The underlying causes of the NCKAP1-associated diseases are likely due to altered actin dynamics, which interfere with neuronal migration during cortical development (Guo et al., 2020). Patients with Coffin-Siris syndrome 1 variants exhibit symptoms such as aplasia or hypoplasia of the distal phalanges and abnormal facial features, which are similar to those observed in individuals with DA5 (Vergano and Deardorff, 2014). Therefore, by using genetic approaches, we can not only identify suppressors of disease relevant pezo-1 mutants, but also uncover other molecules that contribute to these symptoms in humans. This helps to establish a molecular connection between disease-causing genes and provides valuable insights for future therapeutic advancements.

C. elegans strains were maintained with Golden lab protocols (Bai et al., 2020b). Strain information is listed in Table S1.

AG437 pezo-1(R2405P) early L4 hermaphrodites were washed three times in M9 and soaked in 48 mM ethyl methane sulfonate (EMS) solution for 4 h at room temperature. The EMS-treated animals were washed three times in M9 and were transferred to a fresh 100 mm MYOB plate with OP50 on one side. The animals were allowed to recover for up to 4 h before being picked to individual 100 mm MYOB plates with fresh OP50. Only the recovered animals that were able to crawl across the plates to the OP50 food were transferred to the fresh plates. A total of 70 MYOB plates with 25-30 mid-L4 (P0s) on each were incubated at 20°C. Gravid F1 adult were bleached and F2 embryos were collected after hypochlorite treatment. The F2 embryos were shaken in a glass flask with M9 buffer overnight, and hatched larvae were grown on sca-1 RNAi plates at 25°C for 1 week. Their progeny were screened for viable larvae. Approximately 150,000 mutagenized haploid genomes were scored in this fashion.

The RNAi feeding constructs were chosen from the Ahringer and Vidal libraries (Rual et al., 2004; Fraser et al., 2000). RNAi bacteria were grown until log phase was reached and spread on MYOB plates containing 1 mM IPTG and 25 μg/ml carbenicillin, and incubated for 12-14 h. The seeded RNAi plates were stored at 4°C up to 1 week. To deplete the target genes gex-3, wve-1 and abi-1, mid-L4 hermaphrodites were picked onto plates with the IPTG-induced bacteria. Animals were grown on RNAi plates at 20°C for 36-60 h for brood size and other assays.

Single mid-L4 hermaphrodites were picked onto 35 mm MYOB plates seeded with 5-10 μl of fresh OP50 bacteria and allowed to lay eggs for 36 h (plate one contains the brood size from 0-36 h post mid-L4). The hermaphrodites were transferred to a newly seeded 35 mm MYOB plate to lay eggs for another 24 h and were flamed from the plate (the brood size on this plate was defined as the brood size from 36-60 h post mid-L4). Twenty-four hours after removing the hermaphrodites, the viable larvae were counted for the embryonic viability. Brood sizes were determined at 36 h and 60 h. Percentage of embryonic viability=(the number of hatched larva/the total brood size)×100.

For imaging ovulation, animals were immobilized on 7% agar pads with anesthetic (0.1% tricaine and 0.01% tetramisole in M9 buffer). DIC image acquisition was carried out using a Nikon 60× oil objective with 2-3 μm z-step size; 15-25 z planes were captured. Time interval for ovulation imaging is every 60-90 s, and duration of imaging is 60-90 min. Ovulation rate=(number of successfully ovulated oocytes)/total image duration. Actin imags were captured using a Nikon 60× water objective with 0.5 μm z-step size; 15-20 z planes were captured.

All CRISPR/Cas9 editing was generated into Bristol N2 strain as the wild type unless otherwise indicated. The crRNAs were synthesized from Horizon discovery, along with tracrRNA. Repair template and oligos were purchased from Integrated DNA Technologies (IDT). The CRISPR design followed the standard protocols (Paix et al., 2015). Approximately 20-30 young gravid animals were injected with the CRISPR/Cas9 injection mix. Detailed sequence information of CRISPR design is listed in Table S2.

MitoTracker Red CMXRos (MT) (Invitrogen, M7512) was used to label male sperm following the protocol adapted from previous studies (Hoang et al., 2013; Kubagawa et al., 2006). Wild-type males were incubated in the MT buffer for 2 h in the dark. The stained males were covered with foil to prevent light exposure overnight. About 30 males were placed with 10 anesthetized hermaphrodites (0.1% tricaine and 0.01% tetramisole in M9 buffer) on MYOB plates seeded with a 10 μl OP50 bacteria. After 20-30 min of mating, hermaphrodites were then isolated and allowed to rest on food for at least 1 h. The mated hermaphrodites were then mounted for microscopy on 7% agarose pads with the anesthetic. Image acquisition was carried out using a Nikon 60×1.2 NA water objective with 1 μm z-step size. Sperm distributions were quantified using the ImageJ cell counter.

Auxin indole-3-acetic acid (IAA) was purchased from Alfa Aesar (A10556). A 400 mM stock solution of IAA was made in ethanol and was added to MYOB medium to a final concentration of 1 or 2 mM auxin. To efficiently degrade the GEX-3 protein, mid-L4 hermaphrodites were picked onto auxin plates. Animals were grown on the plates at 20°C for 24 h for the degradation efficiency test, and for 60 h for brood size assay.

All imaging was carried out on a spinning disk confocal system that uses a Nikon 60×1.2 NA water or oil objectives, a Photometrics Prime 95B EMCCD camera and a Yokogawa CSU-X1 confocal scanner unit. Nikon's NIS imaging software were applied to capture the images. The image data were processed using ImageJ/FIJI Bio-formats plug-in (National Institutes of Health) (Schindelin et al., 2012; Linkert et al., 2010).

To express PEZO-1 in Sf9 cells (a clonal isolate of Spodoptera frugiperda Sf21 cells), we generated recombinant baculoviruses, according to the manufacturer's instructions (Bac-to-Bac expression system; Invitrogen). To generate this baculovirus, we used a pFastBac construct (Epoch Life Science) containing an 8× histidine–maltose binding protein tag and a synthesized pezo-1 isoform G nucleotide sequence (one of the longest isoforms according to RNA sequencing; wormbase.org release WS280). For expression of PEZO-1 R2405P, the construct contained an 8× histidine–maltose binding protein tag and a synthesized pezo-1 isoform G with the R2405P point mutation. We infected Sf9 cells with either wild-type or mutant pezo-1 baculovirus for 48 h as previously described (Millet et al., 2022). Infected cells were plated on glass coverslips coated with a peanut lectin solution (1 mg/ml; Sigma-Aldrich) for patch-clamp experiments.

pezo-1(R2405P^MM) males were mated with the homozygous hermaphrodites of each suppressor line (Fig. S2A). We pooled suppressed F2 progeny from the cross and allowed the F2 population to expand for ten generations for MIP-MAP analysis. Ten generations of self-recombination are sufficient to distribute the MIP-MAP single nucleotide polymorphism (SNPs) (also refer to molecular probes) into the suppressor line background and provide a high molecular resolution to identify the mutated regions. Candidate mutations (defined as novel, homozygous and nonsynonymous) were identified by whole-genome sequencing as described previously (Smith, 2022). Briefly, sequencing libraries were constructed using a Invitrogen Pure Link Genomic DNA Mini Kit (K1820-01) with genomic DNA from homozygous suppressor-bearing strains. The libraries were pooled and sequenced on a HiSeq 3000 instrument (Illumina) to at least 20-fold coverage. Variants were identified with a pipeline of BBMap (Bushnell, 2022), SAMtools (Li et al., 2009), FreeBayes (Garrison and Marth, 2012) and ANNOVAR (Wang et al., 2010). Mapping loci for suppressors were identified using molecular inversion probes (MIPs) to SNPs as described previously (Mok et al., 2017). Briefly, suppressor-bearing strains were mated to SNP mapping strain VC20019 (Thompson et al., 2013), which had been engineered via CRISPR to contain the pezo-1(R2405P) mutation. F1 cross-progeny were allowed to self-fertilize, and a minimum of 50 homozygous F2 progeny were pooled for construction of MIP libraries. SNP allele frequencies were determined using a custom script and plotted with R (R Core Team, 2021) to delimit the mapping interval.

The electrophysiological data and statistical analyses were performed using GraphPad Instat 3 and OriginPro 2018 software. Statistical methods and sample numbers are detailed in the corresponding figure legends. No technical replicates were included in the analyses. Statistical significance for other assays was determined by the P-value from an unpaired two-tailed t-test, one-way ANOVA or Chi-squared test. ns=not significant; *P<0.05; **P<0.01, ***P<0.001; ****P<0.0001. Both the Shapiro-Wilk and Kolmogorov–Smirnov Normality test indicated that all data follow normal distributions.

We thank the Caenorhabditis Genetics Center, which is funded by National Institutes of Health Office of Research Infrastructure Programs (P40OD010440), for providing strains for this study. We thank Dr Erin Cram's generosity for providing the actin reporter strains. We are grateful to the members of the Golden laboratory for productive discussions and preparing reagents. We especially thank Dr Erin Cram, Dr Orna Cohen-Fix, Dr Kevin O’ Connell and Dr Katherine McJunkin for critical inputs on the project and feedback on the manuscript. We thank all members of the Baltimore Worm Club for providing feedback and suggestions for our investigations. In Memoriam: Dr Andy Golden, who provided tremendous both scientific and financial support to this study. He will be sorely missed.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.202214.reviewer-comments.pdf