The Conserved ASCL1/MASH-1 Ortholog HLH-3 Specifies Sex-Specific Ventral Cord Motor Neuron Fate in Caenorhabditis elegans

Neural specification is regulated by one or many transcription factors that control expression of effector genes that mediate function and determine neuronal type. Here we identify a novel role for one conserved proneural factor, the bHLH protein HLH-3, implicated in the specification of sex-specific ventral cord motor neurons in C. elegans. Proneural genes act in early stages of neurogenesis in early progenitors, but here, we demonstrate a later role for hlh-3. First, we document that differentiation of the ventral cord type C motor neuron class (VC) within their neuron class, is dynamic in time and space. Expression of VC class-specific and subclass-specific identity genes is distinct through development and is dependent on the VC position along the A-P axis and their proximity to the vulva. Our characterization of the expression of VC class and VC subclass-specific differentiation markers in the absence of hlh-3 function reveals that VC fate specification, differentiation, and morphology requires hlh-3 function. Finally, we conclude that hlh-3 cell-autonomously specifies VC cell fate.

61% identical to ASCL1 (33/54). HLH-3 heterodimerizes with the Class I bHLH transcription factor HLH-2, predicted ortholog of TCF3/TCF4/ TCF12 (Kim et al. 2018;Krause et al. 1997). Our previous work has implicated HLH-3 in the terminal differentiation of the hermaphroditespecific motor neurons, HSNs, a bilateral pair of neurons that function in the egg-laying circuitry (Schafer 2006;Doonan et al. 2008;Raut 2017). Work by others has shown that the gene hlh-3 has diverse functions in the nervous system, it is necessary for the appropriate death of the sisters of the NSMs (Thellmann et al. 2003); it works in combination with other transcription factors to induce the serotonergic program in the HSNs, and moreover, its ortholog, ASCL1, can be a functional substitute (Lloret-Fernández et al. 2018); it promotes neurogenesis of I4 (Luo and Horvitz 2017); it co-regulates the initiation of expression of the terminal selector gene ttx-3 (Murgan et al. 2015); and it regulates expression of the chemoreceptor gene srh-234 (Gruner et al. 2016). The above studies implicated hlh-3 in the development of different types of neurons; however, its role in sex-specific neuron development is not as well documented.
We were the first to report that hlh-3 is expressed in the embryonically generated P cells, ectodermal-like precursors of all post-embryonically generated ventral cord motor neurons. We also showed that by the third larval stage (L3) expression of a truncated translational fusion hlh-3 reporter was restricted to the ventral cord motor neuron type C (VC) class, a hermaphrodite sexspecific type of neuron (Doonan et al. 2008). This expression pattern is consistent with a role in neuroblast specification, a function of canonical proneural genes. However, it remained to be determined whether hlh-3 had a function in the specification of the P cell descendants (including sex-shared as well as sex-specific).
Here we report on the role of hlh-3 in the differentiation of postembryonic ventral cord motor neurons. We show it is necessary for the acquisition and maturation of the hermaphrodite sexspecific VC class only.
The postembryonic ventral cord motor neurons are made up of both sex-specific and sex-shared neurons arising from the anterior descendants of ectodermal-like P blast cells (Pn.a) (Sulston and Horvitz 1977). After two additional cell divisions, the Pn.aap cells give rise to the sex-specific neurons of the ventral cord. In hermaphrodites, the P3-P8.aap cells give rise to VCs ( Figure 1A and 1B), whereas in males Pn.aapa and Pn.aapp (where n = descendant of P3 to P11), give rise to the ventral cord neuron type CA and CP, respectively (Sulston et al. 1980). Their fate acquisition (generation) is influenced by positional cues (Hox genes), differential survival (programmed cell death), and sexual identity (VC vs. CA/CP). The VCs of the hermaphrodite are positioned in the midbody and make up six of the total eight sex-specific neurons. Equivalent lineage descendants (Pn.aap) of P1, P2, and P9-12 cells in hermaphrodites undergo programmed cell death (Clark et al. 1993). These neuron classes (VC, CA, and CP) provide an opportunity to study the molecular mechanisms that drive sex-specific neuron differentiation in the ventral nerve cord.
Little is known about how VCs are generated and how they differentiate. To date, it is not known which factor(s) initiate the differentiation program of VCs to establish a class-wide identity. However, it is known that survival of VCs requires the function of the HOX gene lin-39 and the HOX cofactors encoded by unc-62 and ceh-20 (Clark et al. 1993;Salser et al. 1993). UNC-62, along with LIN-39, promotes survival of the VCs by ensuring CEH-20 localizes to the nucleus; the LIN-39/CEH-20 complex then represses egl-1 transcription (Liu et al. 2006;Potts et al. 2009). Sexual determination of the Pn.aap cells is established by the first larval (L1) stage (as VCs in hermaphrodites and the precursors of CAs and CPs in males) (Kalis et al. 2014). It was also shown that LIN-39 is not required for the expression of the VC terminal differentiation feature ida-1. Moreover, since the surviving descendants from P1, P2, and P9-12 still express ida-1::gfp in lin-39(lf); ced-3(lf) double mutants, it was concluded that the role of LIN-39 is most likely restricted to VC survival, not differentiation (Kalis et al. 2014). However, recent evidence has implicated a role for LIN-39 in the expression of a VC marker srb-16 (Feng et al. 2020).
While the mechanisms that regulate VC class specification have yet to be determined, the mechanisms that regulate VC subclass identity are better understood. Within the VC class, two VC subclasses are distinguished spatially by their proximity to the vulva, categorized as proximal VCs or distal VCs (Schafer 2006). The two VC neurons that flank the vulva are categorized as "proximal" (VC 4 and VC 5), whereas the other four VCs are "distal" to the vulva (VC 1-3, and VC 6) ( Figure 1C). Genetic analysis of unc-4 has revealed that the proximal VC subclass identity is determined by spatial cues. Specifically, the expression of unc-4 as a VC proximal subclass identity gene requires the secretion of LIN-3/EGF (Epidermal Growth Factor) from vulval tissue (vulF cells) (Zheng et al. 2013). EGF signaling promotes proximal VC subclass fate by de-repression of unc-4 in the proximal VCs only. Thus, a non-cell Figure 1 The ventral cord type C motor neuron class. A: Illustration of the position of the six VCs along the ventral nerve cord in the midbody region of an adult hermaphrodite. Anterior is to the left, ventral is down, and the gray triangle on the ventral surface indicates the location of the vulva. B: Diagram of the reiterative post-embryonic cell divisions produced by the P3.a to P8.a neuroblasts that give rise to the VCs (adapted from Sulston and Horvitz 1977) C: Diagram for VC classification includes two sub-classes: proximal (VC autonomous mechanism mediates one aspect of VC differentiation, specifically in proximal VCs. Here we build on the current knowledge of neural specification in C. elegans and discover that the proneural-like bHLH factor, HLH-3, mediates specification and differentiation of the VC sex-specific motor neurons, that is, it is needed early and late in development. By using molecular markers to assess VC differentiation in the absence of hlh-3 function, we find that VC class and subclass identity, as well as morphology, is compromised. Our work identifies a new role for the Achaete-Scute ortholog, HLH-3, in the ventral cord of C. elegans, that is the control of sex-specific motor neuron development. We conclude that HLH-3 is necessary for the expression of the earliest VC class-specific transcriptional regulator (lin-11) examined and is required for the expression of later acting VC class-specific genes.

Strain maintenance
All strains were maintained at 22°on nematode growth media using standard conditions (Brenner 1974). Some strains were provided by the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). hlh-3(tm1688) was isolated by the National Bioresource Project of Japan. cccIs1 was kindly shared by Dr. Jennifer Ross Wolff, uIs45 was kindly shared by Dr. Martin Chalfie, otIs456 was kindly shared by Dr. Oliver Hobert, and both otIs564 and otIs619 were kindly shared by Dr. Paschalis Kratsios. See Supplemental Table 1 for a complete list of strains used in this study.

Integration of extrachromosomal arrays
The transgenic strain harboring icIs270 was generated by exposing L4 hermaphrodites to UV-TMP (350microJoules · 100 on Stratagene UV Stratalinker; 0.03 microgram/microliter TMP. Irradiated animals were placed onto seeded NGM plates and transferred the next day to fresh seeded NGM plates (3 Po/plate). These were followed to clone F1s (150) and subsequently to clone three F2s per F1.
Construction of HLH-3::GFP CRISPR-Cas9 engineered line hlh-3(ic271) Construction of the CRISPR-Cas9 engineered line required modification of two plasmids: the single guide RNA or sgRNA plasmid, pDD162 (Addgene #47549), and the repair template plasmid, pDD282 (Addgene #66823) (Dickinson et al. 2015). The target sequence GCTAT-GATGATCACCAGAAG was selected using the CRISPR design tool on Flybase consisting of a high optimal quality score (96). The sgRNA was cloned into pDD162 to create pLP1. The 59 homology arm was designed as a gBlock containing a silent mutation at the PAM site to prevent Cas9 off-targeting. The gBlock was PCR amplified with primers acgttgtaaaacgacggccagtcgccggca and CATCGATGCTCCT-GAGGCTCCCGATGCTCC and cloned into pDD282. The 39 homology arm was designed via PCR using the primers CGTGATTACAA-GGATGACGATGACAAGAGATAATCTGTTAAGTTGTACC and ggaaacagctatgaccatgttatcgatttccaaggagctggtgcacaag. The PCR product was purified and cloned into pDD282 to create pLP2. The modified constructs pLP1and pLP2, as well as the co-injection plasmid pGH8 (Addgene #19359) were co-injected into an N2 strain: sg-RNA plasmid (pLP1) at 50ng/uL; hlh-3 repair template plasmid (pLP2) at 10ng/uL, and pGH8 at 2.5ng/uL. Screening was carried out according to the published protocol (Dickinson et al. 2015).

Microscopy
Animals were mounted on 3% agarose pads containing droplets of 10mM levamisole. Fluorescent images were acquired with AxioVision on Zeiss Axioskop 2 microscope. Following the collection of images, some conversions were made with FIJI version 2.0.0 (grayscale images were converted with Lookup tables, Red or Green, and merged using Merge Channels) and processed into Adobe Illustrator version 23.1.1 for formatting. Fluorescent reporters were observed under confocal microscopy for the detection of a fluorescent protein signal (presence or absence) in transgenic lines. This study does not report quantification of intensity for any fluorescent reporter observed.

Data availability
Strains and plasmids are available upon request. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article, figures, and tables. Supplemental material available at figshare: https://doi.org/10.25387/g3.12996965.

RESULTS
The Class II bHLH protein HLH-3 is expressed and localized to the nuclei of VCs from L1 through adulthood We have previously shown that in hermaphrodites, hlh-3 is expressed in the postembryonic descendants of the ectodermal-like P cells as well as the HSNs (Doonan et al. 2008). We also have shown that hlh-3 function is cell-autonomously required for normal axon pathfinding and terminal differentiation of the HSNs (Doonan 2014;Doonan et al. 2008;Raut 2017). In those studies, analysis of the expression of a translational fusion reporter with only the first eight amino acids of HLH-3 fused to GFP revealed that expression was widespread in the Pn.a descendants, dynamic, and with time, restricted to the VCs (Pn.aap) and HSNs. To confirm the endogenous spatiotemporal expression pattern of hlh-3 we created hlh-3(ic271[hlh-3::gfp]) [herein called hlh-3::gfp], a CRISPR-Cas9 engineered fluorescent tag at the C terminus of the hlh-3 genomic locus ( Figure 2A) following established genome-engineering protocols (Dickinson et al. 2015). The gene hlh-3 is composed by two exons, the C terminal tag was engineered after the last amino acid in exon 2. The allele hlh-3(tm1688) is a null allele that eliminates the majority of the bHLH domain and transcription start site ( Figure 2B). Our analysis dissects both hlh-3::gfp endogenous expression and the consequence of the absence of hlh-3 function in the context of the ventral cord motor neurons and primarily the VCs. First, in characterizing hlh-3::gfp, our analysis supports both our initial findings (Doonan 2014;Doonan et al. 2008), the recently reported observation that hlh-3 expression reappears in the HSNs at the L4 developmental stage (Lloret-Fernández et al. 2018), and expands our understanding of its role in the VCs (Doonan 2014;Raut 2017). We confirmed that hlh-3 is expressed post-embryonically in the P cells and their descendants and becomes restricted to the terminally differentiated VCs present in adults. After hatching, animals show the expression of hlh-3::gfp in the descendants of the P cells throughout the ventral nerve cord (VNC). We highlight the expression of hlh-3::gfp in an early L1 animal wherein Pn.p expression extinguishes faster than that in Pn.a and its descendants ( Figure 2C, left panel). As development proceeds (L2 and L3), expression is extinguished from other descendants of the Pn.a cells and restricted to the VCs ( Figure  2C, middle and right panels). While fluorescent reporter intensity was not quantified, hlh-3::gfp expression appears to be down-regulated in a window of the fourth larval stage (L4) development ranging from mid L4 to late L4, before increasing in adulthood ( Figure 2D, middle and right panels). To ensure that the detected nuclei are those of VCs, we characterized whether there was co-expression of hlh-3::gfp with a known VC marker, plin-11::mCherry (otIs456 [pmyo-2::GFP; plin-11::mCherry]) ( Figure 2E, bottom left). We find that the hlh-3::gfp positive nuclei are also plin-11::mCherry positive throughout the L4 developmental stage ( Figure 2E and 2F). Interestingly, low levels of hlh-3::gfp expression is also observed in a pair of vulval cells during mid-late substages of L4 development, suggesting a role for hlh-3 in these lineages (data not shown). Expression of hlh-3::gfp in VCs from their birth in L1 through their terminally differentiated stage in adulthood prompted us to investigate the role of hlh-3, as a factor required for an early role in promoting VC fate and required for maintenance of VC fate throughout development. Throughout we will use the allele hlh-3(tm1688), further referred to in this paper as hlh-3(lf).

The differentiation of VC class and VC subclass motor neurons is dynamic
Before we analyzed the role of hlh-3 in VCs, we first characterized differentiation features of VCs by defining VC class vs. VC subclassspecific terminal identity. Here we took advantage of fluorescent transcriptional reporter genes that serve as markers of VC fate. We examined the expression of the known VC class-specific markers lin-11, ida-1, and glr-5 encoding a LIM homeodomain transcription factor (Freyd et al.1990), a protein tyrosine phosphatase-like receptor protein homolog of IA2 (Cai et al. 2001(Cai et al. , 2004Zahn et al. 2001), and a glutamate receptor subunit (Brockie et al. 2001), respectively. We confirmed that expression of lin-11 in VCs is observed as early as the second larval stage (L2), and through adulthood (data not shown) (Hobert et al. 1998;Zheng et al. 2013).
Unlike lin-11, a transcriptional regulator, the expression of the VC class terminal identity genes ida-1 and glr-5, was only detected later in development, arising at the L4 developmental stage ( Figure 3A and 3B). Analysis of these VC class differentiation markers throughout L4 substages revealed distinct spatiotemporal patterns suggesting different pathways regulate them. Expression of ida-1 and glr-5 is not equivalent across all six VCs during L4 development ( Figure 3A and 3B). Classification of the L4 substages (early, mid, and late) is based on the vulval L4 morphology as previously described (Mok et al. 2015). We noted that ida-1 expression in early L4 development is robust in the posterior VCs (VC 4-6), whereas, the anterior set of VCs show expression in only half of the sampled individuals. By mid L4, all individuals express ida-1 equivalently in VCs (100% animals); therefore, the individuals in mid and late substages are grouped together ( Figure 3A, right panel). In contrast to ida-1, glr-5 expression is restricted to the proximal VC subclass in the early L4 substage and only fully expressed in all six VCs in the late L4 substage ( Figure 3B). Thus, while the six VCs terminally express their class-specific terminal differentiation genes ida-1 and glr-5, the initiation of transcription is distinct across the sub-stages of L4 development.
Next, we characterized the expression pattern of the VC subclassspecific terminal identity genes unc-4 and unc-17. Others have shown that unc-4 expression requires lin-11 and vulval EGF signaling (Zheng et al. 2013). We corroborate that unc-4 expression is detected after the mid-L4 stages and is maintained throughout adulthood only in VC 4 and VC 5 ( Figure 3C). The expression of UNC-17, in turn, is known to require a posttranscriptional step mediated by UNC-4 (Lickteig et al. 2001). Therefore, we analyzed the expression of two transcriptional unc-17 reporters. To our surprise, and in contrast to work by others, we only detect the expression of these unc-17 reporters in VC 4 and VC 5 at the adult stage regardless of which reporter we characterized (Supplemental Figure 1A-E, and F top left panel). However, our work is different from others (Pereira et al. 2015) in that we did not assess a translational reporter. Instead, we looked at two transcriptional reporters vsIs48 (punc-17::gfp) and mdEx865 (unc-17p::NLS::mCherry + pha-1(+)) and did not observe unc-17 expression in the distal VCs 1-3 and 6 with either reporter. Although we do not see the unc-17 reporters in the distal VCs, we do see expression in proximal VCs, as expected. However, expression is not always detectable in both proximal VCs in exactly the same individual (Supplemental Figure 1D and 1F). Our observations are also consistent with previous reports that anti-UNC-17 immunoreactivity is robust in VC 4 and VC 5, but rarely detectable in distal VCs (Lickteig et al. 2001;Duerr et al. 2008) and possibly only in the second larval (L2) stage (Alfonso et al. 1993).
mir-124 is a novel VC subclass-specific identity feature To expand the molecular repertoire of VC class and subclass markers we looked at several other genes known to be expressed in VCs (flp-11, ntc-1). However, flp-11 and ntc-1 expression is not unique to VCs in the ventral cord (Kim and Li 2004;Garrison et al. 2012). Furthermore, we examined the list of predicted targets of HLH-3/ HLH-2 heterodimers (Grove et al. 2009) to identify genes that may be expressed in VCs too. The gene mir-124 appeared as a logical candidate because it is a predicted target of hlh-3 (Grove et al. 2009), and its expression is detected in a variety of sensory neurons, as well as the CEMs and HSNs, sex-specific neurons in males and hermaphrodites, respectively (Clark et al. 2010). We found mir-124, the highly conserved non-coding microRNA, as a novel VC subclass-specific represent expression in VC nuclei (middle and right panels). Intense gut autofluorescence is apparent dorsally, above the ventral cord, because of overexposure in order to capture low levels of HLH-3::GFP. D: Representative images of the midbody ventral cord of hermaphrodites harboring hlh-3::gfp over distinct L4 developmental stages (early, mid, and late). Larval substages (top panels) are classified by vulva morphology (Mok et al. 2015). Intense gut autofluorescence is apparent above the ventral cord because of overexposure in order to capture low levels of HLH-3::GFP. Filled arrowheads point to the proximal VCs (bottom panels). E: Overlapping expression (merge, top right) of otIs456, containing the VC marker plin-11::mCherry (bottom left), and hlh-3::gfp (top left), in an animal at the L4 molt (bottom right). Filled arrowheads point to co-labeled proximal VCs. F: Overlapping expression (merge) of otIs456, and hlh-3::gfp in a mid L4 animal. Filled arrowhead points to co-labeled proximal VCs. Expression of otIs456 is also detectable in vulval cells. differentiation feature. We characterized mjIs27 [mir-124p::gfp + lin-15(+)] expression across postembryonic development, and only see it in a restricted window. We found mir-124 is expressed from early L4 larval substages through early adulthood, but not in mature gravid egg-laying hermaphrodites ( Figure 3D), which suggests it may be required for the maturation of the VCs but not for maintenance of VC fate. This expression pattern is unlike that of other proximal VC identity features, unc-4 and unc-17, which are expressed throughout adulthood ( Figure 3C; Supplemental Figure 1). Therefore, we classify mir-124 as a novel VC subclass-specific feature expressed during early differentiation. In summary, we conclude that mir-124 can be added to the list of VC identity features belonging to the proximal class (Figure 4).

Classification of VC identity
Thus far, we have shown that the VC class of neurons acquire classspecific features via mechanisms that differ in time and space. We have also shown that not all six VCs are identical in their repertoire of transcriptional activity. In Figure 4, we summarize the spatiotemporal expression pattern of VC identity features. Two genes encoding presumptive transcription factors, hlh-3 and lin-11, are Figure 3 The spatiotemporal expression of VC class and subclass-specific identity features is dynamic. A: Expression pattern of the VC class differentiation feature ida-1 in early and mid-late L4 developmental stages. Representative early L4 hermaphrodite expresses inIs179 [pida-1::gfp] in four VCs, VC 3 -VC 6 (indicated by arrowheads). Graphs report the percent of animals expressing inIs179 [pida-1::gfp] (early L4, n = 20; mid-late L4, n = 40) in each VC. Since all VCs express pida-1::gfp by mid-L4, quantification of pida-1::gfp expression for mid and late L4 stages is pooled. Asterisk highlights expression in the uV1 cells too. B: Expression pattern of the VC class differentiation feature glr-5 in early and late L4 development. Image shows an early L4 hermaphrodite expressing icIs270 [pglr-5::gfp] in the proximal VCs (indicated by arrowheads). Graphs report the percent of animals expressing icIs270 in early L4 (n = 10), mid L4 (n = 15), and late L4 (n = 15) developmental stages in each VC. Asterisk highlights expression in the HSNs too. C: Quantification of expression of the VC subclass feature, unc-4, from L4 development through adulthood. Image shows the expression of uIs45 [punc-4::MDM2::GFP] in an adult (indicated by arrowheads). The percent of animals expressing the VC subclass marker in both cells (VC 4 and VC 5) at early L4 (n = 8), late L4 (n = 12), and adult (n = 19) stages is listed adjacent to the image. D: Quantification of expression of VC subclass feature, mir-124, from L4 development through adulthood. Image shows expression of mjIs27 [mir-124p::gfp + lin-15(+)] in the proximal VCs of an early L4 hermaphrodite. The percent of animals expressing the VC subclass marker in both cells (VC 4 and VC 5) at early L4 (n = 8), late L4 (n = 13), and adult (n = 10) substages is listed adjacent to the image. Asterisk highlights expression in the HSNs too.
VC class-specific and hlh-3 expression precedes that of lin-11 ( Figure  4A-C). We classify the ida-1 and glr-5 genes as VC class-specific as well, as they are observed in all six VCs from L4 through adulthood. In contrast, mir-124 is not expressed in adulthood, as the rest of the VC terminal identity features are. Finally, unc-17 was observed in the proximal VCs only. It is worth emphasizing that aside from the analysis of the CRISPR-Cas9 engineered hlh-3::gfp line, our analysis is predominantly based on the characterization of transcriptional reporters ( Figure 4; Supplemental Table 1).
hlh-3 function is required for the acquisition of VC class and VC subclass identity features Previously, hlh-3 has been shown to be required for HSN terminal differentiation (Doonan et al. 2008;Raut 2017;Lloret-Fernández et al. 2018). To address whether hlh-3 has a role in VC differentiation we first examined reporters of VC terminal identity the genes ida-1 and glr-5, in hlh-3(lf) ( Figure 5). We find that expression of the terminal VC class markers ida-1 and glr-5 is eliminated in the distal VCs of one-day-old hlh-3(lf) adult hermaphrodites, synchronized as L4s the day before ( Figure 5A and 5B). A defect in the expression of these differentiation markers is less severe in the proximal VCs, VC 4 differentiation is slightly affected but VC 5 differentiation is almost like WT. However, since we show that ida-1 and glr-5 expression is detectable in early L4 wild type individuals ( Figure 3A and 3B) we also examined their expression in hlh-3(lf) L4 hermaphrodites. We find that expression of the differentiation features ida-1 (Supplemental Figure 2) and glr-5 (not shown) in the earlier stages of L4 development (L4.0 to L4.3) is completely absent in hlh-3(lf) hermaphrodites. This indicates that from mid L4 through adulthood, an hlh-3 independent mechanism regulates expression of these genes in the proximal subclass.
We find that the early differentiation subclass-specific feature mir-124 (mjIs27 [mir-124p::gfp + lin-15(+)]) is completely absent in hlh-3(lf) ( Figure 5D). We followed up with an analysis of unc-4 expression. Others have shown that the expression of this VC subclass-specific terminal identity gene is de-repressed in WT animals after EGF signaling in mid-L4 development (Zheng et al. 2013). Here, we find that the absence of hlh-3 function reduces the frequency of unc-4 expression ( Figure 5E). Since unc-4 expression is required for unc-17 expression (Lickteig et al. 2001), not surprisingly we find that expression of unc-17 is missing the proximal VCs in hlh-3(lf) individuals (Supplemental Figure 1F).

hlh-3 is required for normal axon branching of proximal VCs
Consistent with the reduced expression of VC terminal identity molecular markers in hlh-3(lf), proximal VCs have abnormal axonal branching in the vulval ring ( Figure 5F). This defect suggests that proximal VC function may be impaired in hlh-3(lf), as axonal branching is required for synaptic connections to the egg-laying circuitry (Schafer 2006). Thus, growth and maturation of VC axons require hlh-3 function, as it is the case for the HSNs (Doonan 2014;Doonan et al. 2008;Raut 2017).
VCs survive in the absence of hlh-3 function Our analysis of VC class and VC subclass markers indicate that the expression of VC differentiation markers is compromised in hlh-3(lf) individuals. To ensure VC survival occurs we next sought to eliminate the possibility that VCs degenerate via necrosis or inappropriately undergo programmed cell death in hlh-3(lf). In C. elegans necroticlike cell death results from excess Na + influx through abnormal sodium permeant channels or gain of function mutations in Ga s (Xu et al. 2001). Programmed cell death (PCD) is a conserved  Table 1 for the list of strains containing these markers). With the exception of mir-124, all reported features are maintained through adulthood (denoted by an asterisk).

Figure 5
VCs require HLH-3 to acquire class-specific and subclass-specific differentiation features and normal axon morphology. A-B: Representative images of WT or hlh-3(lf) one-day-old adults harboring the indicated reporters. Filled arrowheads point to detectable VCs in either genotype. Graphs report the percent of animals expressing each reporter in each VC of WT (black bars) and hlh-3(lf) (red bars) one-day-old adults. The expression of the ida-1 marker (inIs179) was quantified in WT (n = 15) and hlh-3(lf) (n = 35) individuals (panel A). The expression of the glr-5 marker (icIs270) was quantified in WT (n = 15) and hlh-3(lf) (n = 30) individuals (panel B). C-D: Representative images of WT and mutant hermaphrodites at different stages of development and harboring the indicated reporters of VC subclass features mir-124 (mjIs27) and unc-4 (uIs45). Fluorescent images of the vulval region of DIC imaged hermaphrodites (top panels) only revealed expression in the proximal VCs of WT individuals (indicated by filled arrowheads). Quantification of the percent of animals with detectable reporter expression of mir-124 (mjIs27) in VC 4 or VC 5 is pathway executed by CED-3, a caspase that functions as the final determinant in the cell death pathway (Conradt et al. 2016). Inhibition of this pathway, by impairment of ced-3 function, results in the survival of cells destined to die. In the context of the ventral nerve cord, the cells P1-P2.aap and P9-12.aap will survive ( Figure 6A). Therefore, we introduced a ced-3(n717) null mutation into hlh-3(lf) mutants and analyzed the expression of a VC differentiation marker, glr-5, in ced-3(lf) and ced-3(lf); hlh-3(lf) individuals. Unlike ced-3(lf) hermaphrodites, which express glr-5 in all VCs, including the surviving P2.aap cell, we find that ced-3(lf); hlh-3(lf) mutants do not express glr-5 in most VCs. This defect is most notable in the distal class, but it is also apparent in the surviving VC-like cell P2.aap, which appears to follow the terminal differentiation program of the proximal VCs ( Figure 6B-D). Therefore, we conclude that the reason VC neurons do not express glr-5 in the absence of hlh-3 function is that they need HLH-3 to fully differentiate and not because they undergo inappropriate PCD.
reported in the graph below the images. Fluorescence in mid-L4 hermaphrodites of the indicated genotypes and harboring mjIs27 was either detectable (on) or not detectable (off); WT mid L4s (n = 17) and hlh-3(lf) mid L4s (n = 14). Quantification of the percent of animals with detectable reporter expression of uIs45 reporter expression in VC 4 or VC 5 is shown in the graph below the images. Fluorescence was either bright (on), dim, or not detectable (off) for expression of uIs45 in WT (n = 66) and hlh-3(lf) (n = 61) one-day-old adults. E: Quantification of proximal VC axon branching in WT and hlh-3(lf) individuals. Normal axons branch into a vulval ring, as observed with uIs45 in the WT genotype highlighted by the filled arrows (left panel). In contrast, hlh-3(lf) hermaphrodites display abnormal axon branching highlighted by the unfilled arrows (right panel). The numbers to the right represent the percent of individuals with abnormal branching in one-day-old adult WT (n = 15) and hlh-3(lf) (n = 24) hermaphrodites. Figure 1A) (Wormbase: Curatorial remark). The transcriptional reporter otIs564 [punc-47::mChOpti] marks all GABAergic neurons expressing a vesicular GABA transporter (within the VNC this includes DD and VD neurons; Supplemental Figure 3A) (Gendrel et al. 2016). We find that the total number of cholinergic neurons anterior to the vulva is equivalent between WT and hlh-3(lf) individuals (Supplemental Figure 1C). Likewise, the total number of GABAergic neurons is equivalent between WT and hlh-3(lf) hermaphrodites (Supplemental Figure 3B). These analyses demonstrate that the cholinergic and GABAergic sex-shared ventral cord motor neurons acquire their terminal neurotransmitter fate. Thus, hlh-3 function is not necessary for the acquisition of the terminal fates in sex-shared neurons, rendering its function specific to the terminal differentiation of sex-specific ventral cord VC neurons.

Supplemental
The male-specific ventral cord motor neurons do not require hlh-3 function We wondered whether the male-specific ventral cord motor neuron differentiation was also dependent on hlh-3 function. The CA and CP pairs of male-specific motor neurons arise from the division of the Pn.aap neuroblast, anteriorly (type CA) and posteriorly (type CP) as shown in Supplemental Figure 4A (Sulston et al. 1980). We tracked differentiation of the CAs 1-9 and the CPs 1-6 with the differentiation markers for ida-1 and tph-1, respectively (Supplemental Figure 4B). We find that hlh-3(lf) one-day-old adult males, when compared to WT one-day-old adult males show expression of differentiation markers in all CA and CP neurons, nearly at equivalent proportions (Supplemental Figure 4C and 4D). This suggests that hlh-3 does not have a role in promoting the differentiation of these neurons. Notably, we did not quantify CP0 (descendant of the P2 lineage) although we did observe expression of pida-1::gfp in both WT and hlh-3(lf) males (data not shown).

DISCUSSION
hlh-3 specifies VC fate Our work identified a new role for the proneural gene hlh-3 as a regulator of sex-specific motor neuron differentiation in the postembryonic VNC of the hermaphrodite. Both terminal and nonterminal identity features associated with the sex-specific motor neurons, VCs, are reduced or absent in animals that lack hlh-3 function. We show that VCs do not undergo inappropriate programmed cell death ( Figure 6). We confirmed that the VCs are present, in particular the distal VCs, because expression of otIs619 [unc-11 promoter 8 ::NLS::mCherry] is detected in the VCs (Supplemental Figure 5). One-day-old adult hermaphrodites from both genotypes express the pan-neuronal marker at a similar frequency, suggesting that although these cells do not express VC features in hlh-3(lf), they have a neuronal ground state. While most of our analysis measures transcriptional gene activity of VC identity genes, we also demonstrate that the morphology of the VC subclass is affected. In summary, we implicate hlh-3 in the specification of the VC motor neuron class; all six VCs require hlh-3 function.
Differentiation of the proximal VCs involve hlh-3 dependent and hlh-3 independent mechanisms Our analysis also revealed that in the absence of hlh-3 function, the differentiation of proximal VCs is less affected than that of distal VCs. We have gained some insight into these differences with the analysis of markers that are expressed in early L4 vs. later L4 substages (Figure 3). Expression of VC class and VC subclass-specific identity features ida-1, glr-5, and mir-124, is seen in VCs in early L4 substages in a WT context, yet, is completely absent from these early substages through adulthood in animals that lack hlh-3 function (Figure 3; Figure 5; Figure 8A; Supplemental Figure 2). This indicates hlh-3 function is required in all six VCs before L4 development through adulthood. However, our findings also support that a hlh-3 independent mechanism regulates expression of features in the proximal subclass (see below). Lastly, our findings support that hlh-3 function is necessary for expression of VC features but it does not address whether the role is a direct one.
As mentioned before, in the course of our studies we also learned that in the mutant context, and during later stages of L4 development, expression of VC identity features appeared in just a few VCs, the proximal ones. This suggests that there may be a parallel pathway, which can promote VC differentiation. Since the proximal VCs are less affected in their expression of the terminal identity genes that arise after mid-L4 development (unc-4 and unc-17), we propose that this alternative pathway acts by mid L4 but not sooner. We infer that the hlh-3 independent parallel pathway is mediated by EGF, a cue secreted by vulF cells as early as mid L4, already shown to be required for expression of unc-4 in proximal VCs ( Figure 8B; Zheng et al. 2013). The presence of this parallel pathway could ensure that at least proximal VCs retain some function, as they are primary contributors to egg laying by providing feedback to HSNs and vulva muscles (Schafer 2006).
In summary, we have found that the acquisition of VC class features (shown herein) is impaired in hlh-3(lf) individuals. Of the features we have analyzed, only one subclass differentiation feature, expression of mir-124, is fully dependent on hlh-3 function ( Figure  5C; Figure 8A). Since mir-124 expression is restricted to the VC proximal subclass, it may have a role in promoting VC subclass diversity. However, since the expression of mir-124, is seen prior to the EGF cue, and is completely absent in hlh-3(lf), we believe that it is regulated by hlh-3 and not by the EGF-dependent pathway. Further work has to address whether mir-124 functions as an intrinsic, cellautonomous mechanism to promote VC class diversity.
With this work, we propose that: (1) hlh-3 functions cell-autonomously to specify VC class fate in development (from L1 to adulthood) and (2) during L4 development an EGF-dependent cue promotes proximal VC subclass fate diversity for function in egg laying. Our proposal is consistent with the observation that expression of lin-11, glr-5, ida-1, and unc-4, in the proximal VCs, is not significantly altered in the absence of hlh-3 function. To reiterate, the proposed hlh-3 dependent pathway specifies VC class fate and an hlh-3 independent pathway promotes VC subclass diversity.
The LIM homeodomain transcription factor LIN-11 in VCs is downstream of and positively regulated by hlh-3 As shown by others, the gene encoding LIN-11 is expressed from L2 through adulthood (Hobert et al. 1998). We have observed this as well with the translational reporter wgIs62 (lin-11::TY1::EGFP::3xFLAG + unc-119(+)) (data not shown). Since our analysis indicates that hlh-3 is expressed before lin-11, we characterized the expression of a different lin-11 transcriptional reporter (plin-11::mCherry) in the absence of hlh-3 function. We showed that hlh-3(lf) mutants exhibit reduced lin-11 transcriptional activity in VCs (Figure 7). It is likely that hlh-3 directly targets lin-11, but further work will determine whether this effect is direct or indirect. Interestingly, the ortholog ASCL1 has been shown to directly target the lin-11 ortholog, Lhx1, in Figure 8 Two pathways promote the acquisition and maintenance of VC-class and VC-subclass features. A: Expression of the VC identity features (lin-11, ida-1, glr-5, mir-124) require hlh-3 function. B: The regulation of the VC identity features occurs in a cell-autonomous way prior and independently of EGF signaling during mid-L4 development. The alternative pathway, dependent on EGF, regulates the expression of unc-4 (Zheng et al. 2013). We propose that the function of EGF signaling adds a secondary input to regulate lin-11 levels in the proximal VCs, and affect unc-4 and unc-17, as well as other VC identity features.
Our analysis of lin-11 expression in hlh-3(lf) also revealed that the proximal VCs are less affected than the distal VCs by the absence of hlh-3 function (Figure 7; Figure 8A). The proximal VCs express plin-11::mCherry at higher proportions than the distal VCs. Thus, the presence of lin-11 transcriptional activity may be dependent on a secondary pathway other than one that is mediated by hlh-3. Given that others have shown lin-11 acts downstream of EGF, lin-11 may be targeted by both a hlh-3 dependent pathway and this secondary EGFdependent pathway ( Figure 8B; Zheng et al. 2013).
We propose that the reason lin-11 transcriptional activity is observed in the proximal VCs of hlh-3(lf) individuals is that EGFdependent signaling is acting in parallel to hlh-3. It is known that the proximal VCs acquire this subclass-specific identity feature (unc-4) in a time-dependent manner, occurring after EGF signaling, after mid-L4 development (Zheng et al. 2013). Our analysis suggests the EGF signaling pathway promotes lin-11 transcription too. This would explain why, in the absence of hlh-3, there is still expression of lin-11 ( Figure 7). Lastly, our findings that hlh-3(lf) mutants also exhibit reduced unc-4 transcriptional activity in the proximal VCs is a logical consequence of lower lin-11 expression in the proximal VCs ( Figure  4E; Figure 8A). Our model shows that two pathways affect the expression of lin-11 and other VC identity genes ( Figure 8).
hlh-3 may be a terminal selector of VC fate Terminal selectors are factors that initiate and maintain the expression of effector genes, required in the final determination of neuronal subtype specification (Allan and Thor 2015). hlh-3 meets several criteria to be classified as a gene encoding a terminal selector in the VCs. First, it is expressed from the birth to the maturation of all VC features. Second, in its absence, all known VC class terminal identity features fail to be acquired. Lastly, it functions cell-autonomously. Since more than one terminal selector can function to regulate downstream effector genes, it is possible that another terminal selector may function with hlh-3. To confirm if hlh-3 is a terminal selector, additional work will need to test for the direct regulation of VC identity genes by hlh-3.