Direct glia-to-neuron transdifferentiation gives rise to a pair of male-specific neurons that ensure nimble male mating.

Sexually dimorphic behaviours require underlying differences in the nervous system between males and females. The extent to which nervous systems are sexually dimorphic and the cellular and molecular mechanisms that regulate these differences are only beginning to be understood. We reveal here a novel mechanism by which male-specific neurons are generated in Caenorhabditis elegans through the direct transdifferentiation of sex-shared glial cells. This glia-to-neuron cell fate switch occurs during male sexual maturation under the cell-autonomous control of the sex-determination pathway. We show that the neurons generated are cholinergic, peptidergic, and ciliated putative proprioceptors which integrate into male-specific circuits for copulation. These neurons ensure coordinated backward movement along the mate's body during mating. One step of the mating sequence regulated by these neurons is an alternative readjustment movement performed when intromission becomes difficult to achieve. Our findings reveal programmed transdifferentiation as a developmental mechanism underlying flexibility in innate behaviour.

Introduction

The coordinated execution of innate, stereotyped sexual behaviours, such as courtship and mating, requires sexually dimorphic sensory-motor circuits that are genetically specified during development (reviewed in Auer and Benton, 2016; Barr et al., 2018; Yang and Shah, 2014). Studies in the nematode Caenorhabditis elegans, in which the development and function of neural circuits can be interrogated with single cell resolution, have revealed two general developmental mechanisms underlying sexual dimorphism in the nervous system. The first involves the acquisition of sexually dimorphic features in sex-shared neurons during sexual maturation, which include changes in terminal gene expression, such as odorant receptors, neurotransmitters and synaptic regulators (Hilbert and Kim, 2017; Jarrell et al., 2012; Oren-Suissa et al., 2016; Ryan et al., 2014; Serrano-Saiz et al., 2017a; Serrano-Saiz et al., 2017b; Pereira et al., 2019; Weinberg et al., 2018). The second mechanism involves the generation of sex-specific neurons (Sammut et al., 2015; Sulston and Horvitz, 1977; Sulston et al., 1980). Sex-specific neurons are primarily involved in controlling distinct aspects of reproductive behaviours, such as egg-laying in the hermaphrodite and mating in the male (reviewed in Emmons, 2018). Generation of sex-specific neurons requires sex-specific cell death (Sulston et al., 1983) or neurogenesis events resulting from sex differences in the cell division patterns and neurodevelopmental programmes of post-embryonic cell lineages (reviewed in Barr et al., 2018). Here we identify a novel way in which sex-specific neurons are generated in the nervous system: through a direct glia-to-neuron transdifferentiation of sex-shared cells.

In one of his seminal papers, John Sulston described a sexual dimorphism in the phasmid sensilla of adult animals (Sulston et al., 1980). The phasmid sensillum is one of the seven classes of sense organs that are common to both sexes in C. elegans. These sense organs are organised in sensilla which are concentrated in the head and the tail (Bird and Bird, 1991; Ward et al., 1975; White et al., 1986; Doroquez et al., 2014). Each sensillum is composed of the dendrites of one or more sensory neurons enveloped by a channel, usually composed of a single sheath glial cell and a single socket-glial cell. These sensilla can be viewed as part of an epithelium, continuous with the skin, and are shaped by mechanisms shared with other epithelia (Low et al., 2019). Socket-glial cells are highly polarised and adhere to the hypodermis at the distal end of their process where they form a small, ring-like hollow pore in the cuticle through which the sensory dendrites can access the outside world. The bilateral phasmid sensilla (Figure 1), situated in the tail, are unusual in that they are each composed of two socket-glial cells (PHso1 and PHso2). John Sulston observed that in juvenile animals (L2 stage) of both sexes, PHso1 forms the primary pore (Sulston et al., 1980; Figure 1A). At adulthood, the hermaphrodite retains a similar structure (Figure 1B). In males, however, it is PHso2 that forms the main pore and PHso1 was described as having retracted from the hypodermis and to protrude into the phasmid sheath (Figure 1C). It was also described to contain basal bodies, a structural component of cilia, in the region enveloped by the phasmid sheath. As sensory neurons are the only ciliated cells in C. elegans (Perkins et al., 1986), this is suggestive of neuronal fate, yet Sulston observed no other neuronal characteristics (Sulston et al., 1980). We previously showed that in the amphid sensillum (a similar organ located in the head) the amphid socket-glial cell (AMso) acts as a male-specific neural progenitor during sexual maturation, dividing to self-renew and generate the MCM neurons (Sammut et al., 2015). We therefore sought to investigate the PHso1 cells in more detail.

Figure 1.

The phasmid sensillum.

Diagram of the phasmid sensillum in either sex at the L2 larval stage (A), in adult hermaphrodites (B) and in adult males (C). The socket-glial cells (PHso1 and PHso2) are coloured in light pink; the sheath glial cells (PHsh) in green; and the ciliated dendrites of the phasmid sensory neurons, in dark pink. The adherens junctions are depicted as black lines between cells. Axonemes and cilia are marked as black bars and black lines inside the dendrite tips. Each phasmid opens to the exterior on the extreme right (posterior), where grey lines mark the cuticle borders of the phasmid pore and fan. Hypodermis (hyp), seam (se). Diagram has been modified from and is used with permission from http://www.wormatlas.org.

We find that during sexual maturation (L4 stage), the pair of sex-shared PHso1 glial cells acquire sexually dimorphic function by undergoing a direct (without cell division) glia-to-neuron transdifferentiation that results in the generation of male-specific neurons, which we term the phasmid D neurons (PHDs). This cell-fate plasticity is regulated by the sex-determination pathway, likely cell-intrinsically. However, we find the cell-fate plasticity does not depend on certain molecular mechanisms known to regulate the only other well-described direct transdifferentiation in C. elegans (Kagias et al., 2012). This bilateral pair of previously unnoticed neurons are putative proprioceptors that regulate male locomotion during specific steps of mating. One of these steps is a novel readjustment movement performed when intromission becomes difficult to achieve. Our results reveal sex-specific direct transdifferentiation as a novel mechanism for generating sex-specific neurons and also show the importance of proprioceptive feedback during the complex steps of mating for successful reproduction.

Results

The sex-shared PHso1 cells undergo glia-to-neuron transdifferentiation in males

Using a lin-48/OVO1 reporter transgene to identify and visualise the PHso1 cell bodies both before and during sexual maturation (Johnson et al., 2001; Wildwater et al., 2011), we observed no distinguishable differences between the hermaphrodite and male PHso1 cells at the L3 stage (Figure 2A). The PHso1 cells display a polarised morphology and a visible socket structure in both sexes. In hermaphrodites, this morphology is maintained during the transition to adulthood and PHso1 cells elongate as the animal grows. In males, by contrast, the PHso1 cells display morphological changes that can be observed during the L4 stage, the last larval stage before adulthood (Figure 2A). We find that during this stage PHso1 retracts its socket process and extends a short dendrite-like posterior projection into the PHsh cell and a long axon-like anterior process projecting towards the pre-anal ganglion region. For this reason and those described below, we use the name PHD to refer to the PHso1 cells after these morphological changes take place. A time-lapse of an individual male reveals that socket retraction is initiated at early L4, when the male gonad has almost reached the tail, and that the nascent axon is first observed at mid-L4, when the vas deferens has joined with the cloaca (Figure 2B). The axon-like outgrowth is complete by the end of tail-tip retraction, when ray precursor cells start to fuse into the tail seam syncytium (Figure 2—figure supplement 1). These observations corroborate those of John Sulston and identify the stage of sexual maturation as the time during which the PHso1 cells undergo radical remodelling in males. This remodelling involves quite a remarkable change in morphology from a socket-glial morphology to a neuron-like morphology.

Figure 2.

The sex-shared PHso1 cells undergo glia-to-neuron morphological changes in males.

(A) Expression of lin-48 and the seam (wrt-2 reporter transgenes in PHso1 of hermaphrodites (left panel) and males (right panel) at the third (L3) and fourth (L4) larval stages and in adults. The images show the morphological transformation of male PHso1 into the PHD neuron during sexual maturation. Arrowheads label the axonal process extending from the PHD into the pre-anal ganglion. Asterisks indicate the dendritic process of PHD. (B) DIC and fluorescent images of a time-lapse of PHso1-to-PHD remodelling in an individual male (see Materials and methods). The top two time-points show the late L3 stage after the gonad has looped back and early L4 after the gonad has crossed over itself (see cartoon at the bottom left side of the left panel). The subsequent time-points range from early-to-mid-L4, when the vas deferens has joined with the cloaca, to late mid-L4, when tail-tip retraction is almost complete. The dashed boxes on the DIC images indicate the position of the fluorescent images. Arrowheads indicate the nascent axon.

DIC and fluorescent images of L4 male tails from a staged population. The outgrowth of the anterior axon-like process of the PHso1/PHD cell is followed using lin-48::mCherry expression until late L4, when it reaches the pre-anal ganglion. The dashed boxes on the DIC images indicate the position of the fluorescent images. Arrowheads indicate the growing tip of the nascent axon.

Figure 2—figure supplement 1.

PHD axon outgrowth.

DIC and fluorescent images of L4 male tails from a staged population. The outgrowth of the anterior axon-like process of the PHso1/PHD cell is followed using lin-48::mCherry expression until late L4, when it reaches the pre-anal ganglion. The dashed boxes on the DIC images indicate the position of the fluorescent images. Arrowheads indicate the growing tip of the nascent axon.

To determine whether the PHso1 cells also acquire neuronal characteristics at the level of gene expression, we analysed and compared the expression of glial and neuronal markers in the PHso1 and PHso2 cells of males and hermaphrodites. In adult hermaphrodites, we find that both cells express the panglial microRNA mir-228 (Wallace et al., 2016) and the AMso and PHso glial subtype marker grl-2 (a Hedgehog-like and Ground-like gene protein: Hao et al., 2006) at similar levels (Figure 3A). In males, by contrast, mir-228 and grl-2 expression is noticeably dimmer in PHso1 than in PHso2 by the mid-to-late L4 stage and completely absent by day 2 of adulthood (Figure 3A, B and C). Importantly, expression of these reporters appears equal in brightness in both cells at the late L3 stage and only becomes noticeably dimmer in PHso1 following the morphological changes, such as the retraction of the socket, that occur during the L4 stage (Figure 3B and C, Figure 3—figure supplement 1). In contrast, in hermaphrodites, no morphological changes of PHso1 are observed and the reporters appear equal in brightness through to adulthood (Figure 3A, Figure 3—figure supplement 1). After the posterior hypodermal cells begin to retract leaving the characteristic fluid-filled extracellular space at the tail, PHso1 begins to express the pan-neuronal marker rab-3 (a synaptic vesicle associated Ras GTPase: Stefanakis et al., 2015) in males but not in hermaphrodites, and this expression persists throughout adulthood (Figure 3A, B and C). In addition to rab-3 expression we also observe sequential expression of other neuronal markers such as: the vesicle acetylcholine transporter unc-17 (Alfonso et al., 1993; Pereira et al., 2015; Figure 3D); the phogrin orthologue for dense-core vesicle secretion ida-1 (Zahn et al., 2001; Figure 3E); the intraflagellar transport component required for proper sensory cilium structure osm-6 (Collet et al., 1998; Figure 3F); and the IG domain containing protein oig-8 (Howell and Hobert, 2017; Figure 6A). None of these neuronal markers are observed in PHso2 or in the hermaphrodite PHso1. The switch in gene expression in PHso1 is therefore initiated concomitantly with morphological changes during sexual maturation and in a male-specific manner. Together, these data demonstrate that the male PHso1 bilateral pair of glial cells transdifferentiate into a novel class of cholinergic and likely peptidergic ciliated neurons, which we have termed phasmid D neurons (PHDs).

Figure 3.

The sex-shared PHso1 cells undergo glia-to-neuron molecular changes in males occasionally accompanied by a division.

(A) Expression of the glial marker reporter transgenes mir-228 and grl-2 and pan-neuronal marker rab-3 in the phasmid socket cells (PHso1 and PHso2) and the PHD neuron in adult male animals. (B) Bar chart showing the percentage of PHso1/PHD cells expressing the pan-glial marker reporter transgene mir-228 and the pan-neuronal marker reporter transgene rab-3 scored concomitantly in males at different stages of development and at adulthood. Intensity of the mir-228 reporter transgene in the PHso1/PHD cells was assessed by eye in comparison with PHso2: dark green indicates PHso1/PHD = PHso2, light green indicates PHso1/PHD < PHso2 and white is non-detectable in PHso1. (C) As B but for the subtype-specific glial marker grl-2 scored concomitantly with rab-3. (D) Expression of the acetylcholine vesicle uploader reporter transgene unc-17 in the PHD neurons of adult males. The lin-44 reporter transgene has been coloured red. (E) Expression of an osm-6 reporter transgene in the PHD neurons of adult males, which are co-labelled with a lin-48 transgene. (F) Expression of an ida-1 reporter transgene in the PHD neurons of adult males, which are co-labelled with a lin-48 transgene. (G) EdU staining to assess PHso1 division. Left panel: quantification of EdU labelling in cells per side in adult males. The AMso division that gives rise to AMso and MCM cells was scored as a positive control. Right panel: representative images of EdU DNA labelling (green) present in the nuclei of the AMso socket cell and MCM neuron (head), and absent in the PHD neuron (tail), unless two cells per side are observed (PHD1 and PHD2). All cells scored were labelled with a lin-48 transgene.

Expression of the glial marker reporter transgene grl-2 and the PHso1 marker reporter transgene lin-48 in the phasmid socket cells (PHso1 and PHso2) and the PHD neuron in L3 males and L4 males and hermaphrodites. Arrowheads label the process extending from the PHD into the pre-anal ganglion. Asterisks indicate the dendritic process of PHD.

(A) Bar chart showing the percentage of PHD neurons per side that express the pan-neuronal marker rab-3 transgene (otIs291) in mutant animals for the apoptosis-inducing gene ced-4 (strain: ced-4(n1162) dpy-17(e164)). The lin-48 reporter transgene was used to identify the PHD neurons. Scorings for control single dpy-17(e164) mutants are included, as are a negative (MCM lineage; Sammut et al., 2015) and positive (PVDR lineage; Sulston and Horvitz, 1977) control for cell death. Cell death in this mutant background is not sex-specific as demonstrated by PVDR scorings in both male and hermaphrodite animals. Male animals from late L4 to adult stages were scored. No statistical difference is observed between any of the groups using Fisher’s exact test. Of note, wildtype animals and ced-4 dpy-17 animals are homozygous siblings obtained from the same cross. Control dpy-17 animals, however, were built from an independent strain. (B-G) Bar charts showing the percentage of cells per side expressing a battery of reporter transgenes before, during and after PHso1 remodelling. Male animals were subdivided into five categories on the basis of gonad migration and tail morphogenesis: late L3 – from the stage at which the gonad begins its posterior migration to the point at which the gonad crosses itself (~28–32 hr); early L4 – from the stage at which the gonad crosses itself to full extension towards the tail (~32–36 hr); mid-L4 – from the start to 50% tail morphogenesis (i.e. complete tail-tip retraction and beginning of ray precursor cell fusion into the tail seam syncytium) (~36–41 hr); late L4 – from 50% to complete tail morphogenesis (i.e. fully developed fan, rays and syncytium;~41–45 hr); adult – from completion of tail-tip retraction to 1-day adult. (B) Quantification of the lin-48 reporter transgene which was used to label in red PHso1 in larvae and PHD in adults. Graphs C-G represent the number of lin-48 expressing cells that concomitantly express different green reporter transgenes. (C) Quantification of the glia subtype marker reporter transgene grl-2 Note that grl-2 is lost in most animals after the remodelling of PHso1 into the PHD neuron (white bar). (D) Quantification of the pan-neuronal marker reporter transgene rab-3 (synaptic vesicle associated Ras GTPase). Note that neuronal reporters are turned on as PHso1 remodels into the PHD neuron. (E) Quantification of the neuron subtype marker reporter transgene unc-17 (acetylcholine vesicle uploader). (F) Quantification of the neuron subtype marker reporter transgene ida-1 (phogrin orthologue for dense-core vesicle secretion). (G) Quantification of the neuron subtype marker reporter transgene oig-8 (sensory-neuron related IG domain containing protein). (H) Quantification of PHD1 and PHD2 expressing the oig-8 (drpIs4) and the grl-2 (drpIs1) reporter transgenes per side, with and without the lin-48 (drpIs3) transgene in the background. As grl-2 is expressed in PHD and PHso2, to identify PHD unambiguously, rab-3 (otIs356) was included in the background and only double positive cells were scored as PHD. The first lin-48 is repeated from B for comparison. Fisher’s exact test was used to compare all categories between genotypes and only statistically significant differences from the ‘2 cell’ phenotype are indicated (*p≥0.05, **p≥0.01, ***p≥0.001).

DIC and fluorescent images of a time-lapse of PHso1-to-PHD remodelling in an individual male (see Materials and methods). The first time-point shows the late-L3 stage after the gonad has looped back. The subsequent time-points, progress through the L4 stage, until mid-late L4, when tail-tip retraction is underway. The dashed boxes on the DIC images indicates the position of the fluorescent images. Arrowheads indicate nascent axon. Images show the right PHso1 with a wildtype morphology and socket structure. In early-to-mid-L4 the PHso1 divides, socket morphology is no longer visible and two cells are observed, which appear to send to a process anteriorly. After the division, DIC images focus on the nucleus of the anterior daughter of the right PHso1. The left PHso1 in this animal does not undergo division and directly assumes the PHD morphology (not shown) as in Figure 2B.

Figure 3—figure supplement 1.

Expression of glial markers is downregulated in the PHso1 of males.

Expression of the glial marker reporter transgene grl-2 and the PHso1 marker reporter transgene lin-48 in the phasmid socket cells (PHso1 and PHso2) and the PHD neuron in L3 males and L4 males and hermaphrodites. Arrowheads label the process extending from the PHD into the pre-anal ganglion. Asterisks indicate the dendritic process of PHD.

PHso1-to-PHD transdifferentiation is usually but not always direct

We sought to further validate that PHso1 transdifferentiation is direct. Division followed by programmed cell death of one of the daughters is common throughout C. elegans embryonic and post-embryonic development (Sulston and Horvitz, 1977; Sulston et al., 1983). We therefore explored the possibility that PHso1 could divide asymmetrically, to give rise to the PHD neuron and a sister cell that undergoes rapid apoptosis. For this purpose, we analysed mutants of the canonical cell death protein CED-4, homologue to the apoptotic protease activating factor 1 (Apaf-1), required to activate the CED-3 effector caspase (Ellis and Horvitz, 1986; Yuan and Horvitz, 1992; reviewed in Conradt et al., 2016). Strong loss-of-function mutations in ced-4 result in the survival of somatic cells that normally undergo programmed cell death during development. We observe no statistically significant increase in the frequency of ‘survivor cells’ in ced-4(n1162) mutants compared to controls, when scoring the number of cells per side of each animal. This suggests there is no cell death in the PHso1 lineage (Figure 3—figure supplement 2A). Moreover, we never observed an apoptotic cell body in the vicinity of the PHD neuron.

Figure 3—figure supplement 2.

PHso1 divides at low frequency in a background-dependent manner.

(A) Bar chart showing the percentage of PHD neurons per side that express the pan-neuronal marker rab-3 transgene (otIs291) in mutant animals for the apoptosis-inducing gene ced-4 (strain: ced-4(n1162) dpy-17(e164)). The lin-48 reporter transgene was used to identify the PHD neurons. Scorings for control single dpy-17(e164) mutants are included, as are a negative (MCM lineage; Sammut et al., 2015) and positive (PVDR lineage; Sulston and Horvitz, 1977) control for cell death. Cell death in this mutant background is not sex-specific as demonstrated by PVDR scorings in both male and hermaphrodite animals. Male animals from late L4 to adult stages were scored. No statistical difference is observed between any of the groups using Fisher’s exact test. Of note, wildtype animals and ced-4 dpy-17 animals are homozygous siblings obtained from the same cross. Control dpy-17 animals, however, were built from an independent strain. (B-G) Bar charts showing the percentage of cells per side expressing a battery of reporter transgenes before, during and after PHso1 remodelling. Male animals were subdivided into five categories on the basis of gonad migration and tail morphogenesis: late L3 – from the stage at which the gonad begins its posterior migration to the point at which the gonad crosses itself (~28–32 hr); early L4 – from the stage at which the gonad crosses itself to full extension towards the tail (~32–36 hr); mid-L4 – from the start to 50% tail morphogenesis (i.e. complete tail-tip retraction and beginning of ray precursor cell fusion into the tail seam syncytium) (~36–41 hr); late L4 – from 50% to complete tail morphogenesis (i.e. fully developed fan, rays and syncytium;~41–45 hr); adult – from completion of tail-tip retraction to 1-day adult. (B) Quantification of the lin-48 reporter transgene which was used to label in red PHso1 in larvae and PHD in adults. Graphs C-G represent the number of lin-48 expressing cells that concomitantly express different green reporter transgenes. (C) Quantification of the glia subtype marker reporter transgene grl-2 Note that grl-2 is lost in most animals after the remodelling of PHso1 into the PHD neuron (white bar). (D) Quantification of the pan-neuronal marker reporter transgene rab-3 (synaptic vesicle associated Ras GTPase). Note that neuronal reporters are turned on as PHso1 remodels into the PHD neuron. (E) Quantification of the neuron subtype marker reporter transgene unc-17 (acetylcholine vesicle uploader). (F) Quantification of the neuron subtype marker reporter transgene ida-1 (phogrin orthologue for dense-core vesicle secretion). (G) Quantification of the neuron subtype marker reporter transgene oig-8 (sensory-neuron related IG domain containing protein). (H) Quantification of PHD1 and PHD2 expressing the oig-8 (drpIs4) and the grl-2 (drpIs1) reporter transgenes per side, with and without the lin-48 (drpIs3) transgene in the background. As grl-2 is expressed in PHD and PHso2, to identify PHD unambiguously, rab-3 (otIs356) was included in the background and only double positive cells were scored as PHD. The first lin-48 is repeated from B for comparison. Fisher’s exact test was used to compare all categories between genotypes and only statistically significant differences from the ‘2 cell’ phenotype are indicated (*p≥0.05, **p≥0.01, ***p≥0.001).

To our surprise, we observed the presence of an extra cell that co-expressed lin-48 and rab-3 in 13–24% of sides of both ced-4 mutant and control animals (Figure 3—figure supplement 2A). This extra cell had a similar morphology to PHD, was located immediately anterior to PHD and was detected from the early L4 stage onwards. It also expressed all the markers known to be expressed by the PHD neurons (Figure 3—figure supplement 2B–G). We realised, however, the percentage occurrence of this extra cell was variable between strains suggesting its presence/absence is background-dependent (Figure 3—figure supplement 2H). In the cases where we did observe an extra cell, we wanted to distinguish between the following possibilities: additional expression of the transgenes in a highly similar neuronal subtype, perduring expression of the transgenes from earlier in the lineages of the phasmid sockets or an infrequent division of PHso1. To this end, we performed 5-ethynyl-2´-deoxyuridine (EdU) staining in a lin-48 background. EdU is a thymidine analogue that is incorporated into the replicating DNA of dividing cells. EdU was fed to a synchronised population of animals at late L2-to-early L3 (22 hr after L1 arrest), after the PHso1 cells are born and before they remodel into the PHD neurons, and adult animals were scored. As a positive control, the divison of the bilateral AMso cells that give rise to the MCM neurons were also scored, as it occurs at a similar time as the PHso1-to-PHD transition (Sammut et al., 2015). In the majority of cases (79% of sides scored), no EdU signal was detected, indicating that PHso1 mainly transdifferentiates directly into the PHD neuron. In 21% of sides however, EdU signal was detected in two lin-48-expressing cells per side in the tail of male animals (Figure 3G) but never in hermaphrodites (data not shown). Importantly, EdU is never present in the PHD neuron unless observed in two cells on the same side, eliminating the possibility of cell death via non-apoptotic machinery, as is the case of the male linker cell (Abraham et al., 2007). In addition, we were able to observe the division of PHso1 in a single animal in a time-lapse of an individual male (Figure 3—figure supplement 3).

Figure 3—figure supplement 3.

Live division of PHso1 in a single-animal time-lapse.

DIC and fluorescent images of a time-lapse of PHso1-to-PHD remodelling in an individual male (see Materials and methods). The first time-point shows the late-L3 stage after the gonad has looped back. The subsequent time-points, progress through the L4 stage, until mid-late L4, when tail-tip retraction is underway. The dashed boxes on the DIC images indicates the position of the fluorescent images. Arrowheads indicate nascent axon. Images show the right PHso1 with a wildtype morphology and socket structure. In early-to-mid-L4 the PHso1 divides, socket morphology is no longer visible and two cells are observed, which appear to send to a process anteriorly. After the division, DIC images focus on the nucleus of the anterior daughter of the right PHso1. The left PHso1 in this animal does not undergo division and directly assumes the PHD morphology (not shown) as in Figure 2B.

All together, these observations indicate that PHso1 can sometimes divide symmetrically to give rise to two neurons, which we term PHD1 and PHD2. However, in the vast majority of cases, the PHso1 cells undergo direct remodelling, without wholesale DNA replication or cell division, during which they directly acquire neuronal fate.

Biological sex regulates PHso1-to-PHD transdifferentiation cell-autonomously

We next addressed whether the PHso1-to-PHD cell fate switch is regulated by the genetic sex of the cell rather than the sex of the rest of the animal, in a manner similar to that of several other sexual dimorphisms in C. elegans (Fagan et al., 2018; Hilbert and Kim, 2017; Lee and Portman, 2007; Oren-Suissa et al., 2016; Ryan et al., 2014; Sammut et al., 2015; Weinberg et al., 2018; White and Jorgensen, 2012; White et al., 2007). To uncouple the sex of PHso1 from the rest of the animal, we drove expression of fem-3 in PHso1 (and PHso2) under the grl-2 promoter. fem-3 inhibits the expression of tra-1, a downstream target of the sex-determination pathway that activates hermaphrodite development and inhibits male development (Hodgkin, 1987 and reviewed in Zarkower, 2006). Therefore, cell-specific expression of a fem-3 transgene will masculinise a cell in an otherwise hermaphroditic background. We find that fem-3 expression specifically in the PHso1, and not PHso2, transforms PHso1 into a PHD-like neuron, resulting in the upregulation of ida-1 and rab-3 expression and the acquisition of neuronal morphology (Figure 4A–C). This indicates that the competence of PHso1 to transdifferentiate into PHD is likely cell-intrinsic and based on genetic sex.

Figure 4.

PHso1-to-PHD plasticity is intrinsically regulated.

(A) Expression of the ida-1 and rab-3 reporter transgenes in adult hermaphrodites carrying the masculinising array grl-2 in PHso1 (right panel) and in non-array-carrying hermaphrodites (left panel). (B) Bar chart showing the percentage of PHso1 and PHso2 cells expressing the ida-1 reporter transgenes in adult hermaphrodites carrying or not carrying the masculinising grl-2 array. Of note, the grl-2 promoter fragment is also expressed in the AMso, excretory pore and excretory duct cells in the head (Hao et al., 2006). Fisher’s exact test was used to compare all categories between genotypes and only statistically significant differences from the non-array carrying animals are indicated (*p≤0.05, **p≤0.01, ***p≤0.001). (C) Bar chart showing the percentage of PHso1 cells expressing the rab-3 reporter transgene in adult hermaphrodites carrying or not carrying the masculinising grl-2 array. Fisher’s exact test was used to compare all categories between genotypes and only statistically significant differences from non-array carrying animals are indicated (*p≤0.05, **p≤0.01, ***p≤0.001).

Factors required for Y-to-PDA transdifferentiation are largely dispensable for PHso1-to-PHD and AMso-to-AMso+MCM transdifferentiation

The first well-described transdifferentiation event in C. elegans is the direct conversion of the rectal epithelial cell Y into the motor neuron PDA (Jarriault et al., 2008). As with PHso1-to-PHD, Y-to-PDA occurs directly, without a division and in a sex-specific manner. However, Y-to-PDA happens exclusively in hermaphrodites (in males the Y blast cell divides and PDA is generated from the anterior daughter of this division) and during the late L1 stage, much earlier than sexual maturation. In Y-to-PDA, the erasure of the original epithelial fate requires the action of a complex of conserved nuclear factors (SOX-2/Sox, CEH-6/Oct, SEM-4/Sall, and EGL-27/Mta) and the acquisition of the final motor neuron fate involves the cholinergic terminal selector EBF transcription factor UNC-3, in addition to specific combinations of histone modifier complexes (Kagias et al., 2012; Richard et al., 2011; Zuryn et al., 2014). We investigated how broadly the molecules involved in the initiation of Y-to-PDA transdifferentiation could be acting in C. elegans, as we hypothesised that the factors involved in the re-specification would depend on the specific terminal neuronal cell fate. Hence, we analysed loss-of-function mutants for sem-4, egl-27, and sox-2 in PHso1-to-PHD transdifferentiation, and in the previously described AMso-to-AMso+MCM cell fate switch (Sammut et al., 2015).

We observed no statistically significant defect in the number and cell identity of MCM and PHD neurons in sem-4(n1971) null mutants (Figure 5A). In egl-27(ok1671) strong loss-of-function mutants, although a significant defect is observed in the number of lin-48-positive MCM neurons, normal rab-3 expression suggests a defect in lin-48::tdTomato reporter transgene expression levels but not in the division of the AMso or in the acquisition of neuronal fate by the MCM. No lin-48 expression is ever observed in the region of PHso1/PHD in egl-27 mutants at any stage or in either sex (Figure 5A and data not shown). This is consistent with previous results that show that egl-27 affects the division pattern of the T lineage and consequently the production of PHso1 (Herman et al., 1999). This early role of egl-27 in the PHso1 lineage precludes its study in the generation of the PHD neuron using loss-of-function mutants.

Figure 5.

Factors required for Y-to-PDA transdifferentiation are largely dispensable for PHso1-to-PHD and AMso-to-AMso+MCM transdifferentiation.

(A) Bar chart showing the percentage of AMso, MCM, and PHD cells expressing the lin-48 transgene (drpIs3; left panel) and the pan-neuronal marker rab-3 reporter transgene (otIs291; right panel), in sem-4(n1971) putative null and egl-27(ok1670) strong loss-of-function mutant animals. The presence and morphology of cells was assessed by lin-48 that is expressed, in the head, in the MCM mother (the AMso) and retained in the AMso and MCM daughters after the division (two cells per side). In the tail, lin-48 is expressed in PHso1 before the cell remodelling and after, in the PHD neuron (one cell per side unless PHD1 and PHD2 are observed). Cells per side were scored in male animals from late L4 to adult stages. Neuronal identity was assessed by rab-3. For the PHD neuron rab-3 was scored concomitantly with lin-48 due to the high number of rab-3-expressing neurons in the tail. Fisher’s exact test was used to compare all categories between genotypes and only statistically significant differences from the wildtype phenotype are indicated (*p≤0.05, **p≤0.01, ***p≤0.001). Of note, 4/7 cells lacking rab-3 expression in the tail of sem-4 mutants retained a socket morphology, which is never observed in the cells lacking the reporter in the control strain. This could suggest a block to the initiation of transdifferentiation in PHso1 or that the severe morphological defects of male tails in these mutant animals impair the remodelling process. (B) Bar chart showing the percentage of MCM and PHD neurons expressing rab-3 after RNAi-knockdown of sox-2. L4440 empty vector was used as a negative control and GFP RNAi-knockdown as a positive control. Fisher’s exact test was used to compare all categories between genotypes and statistically significant differences from the wildtype phenotype are indicated (*p≤0.05, **p≤0.01, ***p≤0.001). (C) Bar chart showing the percentage of MCM and PHD neurons expressing the ida-1 neuron subtype marker in sox-2(ot460) null mutant animals rescued for lethality with a sox-2 fosmid-based extrachromosomal array: (otEx4454[sox-2(fosmid)::mCherry + elt-2]). A mixed population of of sox-2 mutant (mosaic or non-rescued) and wildtype (mutant-rescued) cells were scored. No statistical difference is observed between any of the groups using Fisher’s exact test.

As in Figure 5, cells per side were scored in animals from late-L4 to adult stages. Fisher’s exact test was used to compare all categories between genotypes and only statistically significant differences from the wildtype phenotype are indicated (*p≥0.05, **p≥0.01, ***p≥0.001). (A) Bar chart showing the percentage of AMso, MCM, and PHD cells expressing the lin-48 reporter transgene (drpIs3) after RNAi-knockdown of sox-2 in male animals. In hermaphrodites AMso was scored, as it never divides to produce the MCM neuron. Of note, scored lin-48::tdTomato-positive cells correspond to the same animals scored in Figure 5B and 5 SB. (B) Bar chart showing the percentage of MCM and PHD neurons expressing the ida-1 reporter transgene (inIs179) after RNAi knock-down of sox-2. Ectopic AMso cells are observed in sox-2 depleted F1 animals but not in sox-2 depleted P0 animals. (C) Bar chart showing the percentage of AMso, MCM and PHD cells expressing the lin-48 reporter transgene (drpIs3) in sox-2(ot460) null mutant animals rescued from lethality with a sox-2 fosmid-based extrachromosomal array (otEx4454[sox-2(fosmid)::mCherry + elt-2p::DsRed]). A mixed population of of sox-2 mutant (mosaic or non-rescued) and ‘wildtype’ (mutant-rescued) cells were scored.

As sox-2 null mutations cause L1 lethal larval arrest, we used two parallel strategies to investigate the role of sox-2 in PHso1-to-PHD and AMso-to-AMso+MCM: RNAi knock-down and analysis of a rescued sox-2 null allele. For the former, we made use of the RNAi-sensitised strain nre-1(hd20) lin-15B(hd126) (Schmitz et al., 2007). No statistically significant defect is observed in the PHD neuron in sox-2 RNAi-treated animals (Figure 5B). An increase in the number of rab-3-expressing MCM cells was detected (Figure 5B) but analysis of lin-48 expression revealed that the extra rab-3 positive cells derive from ectopic AMso cells that divide efficiently to produce MCM neurons (Figure 5—figure supplement 1A). Interestingly, the ectopic MCM cells recapitulate normal development as they not only express the pan-neuronal marker rab-3 but also the neuron subtype marker ida-1 (Figure 5—figure supplement 1B). Moreover, although ectopic AMso cells are observed in both sexes, they exclusively divide in males and not in hermaphrodites (Figure 5—figure supplement 1A). This, together with the fact that ectopic lin-48-positive cells are induced in RNAi-treated animals only in the F1 generation (and not the P0 generation – see Materials and methods), suggests an early role of sox-2 in an unknown lineage to supress the production of supernumerary AMso cells.

Figure 5—figure supplement 1.

Factors required for Y-to-PDA transdifferentiation are largely dispensable for PHso1-to-PHD and AMso-to-AMso+ MCM transdifferentiation – additional data.

As in Figure 5, cells per side were scored in animals from late-L4 to adult stages. Fisher’s exact test was used to compare all categories between genotypes and only statistically significant differences from the wildtype phenotype are indicated (*p≥0.05, **p≥0.01, ***p≥0.001). (A) Bar chart showing the percentage of AMso, MCM, and PHD cells expressing the lin-48 reporter transgene (drpIs3) after RNAi-knockdown of sox-2 in male animals. In hermaphrodites AMso was scored, as it never divides to produce the MCM neuron. Of note, scored lin-48::tdTomato-positive cells correspond to the same animals scored in Figure 5B and 5 SB. (B) Bar chart showing the percentage of MCM and PHD neurons expressing the ida-1 reporter transgene (inIs179) after RNAi knock-down of sox-2. Ectopic AMso cells are observed in sox-2 depleted F1 animals but not in sox-2 depleted P0 animals. (C) Bar chart showing the percentage of AMso, MCM and PHD cells expressing the lin-48 reporter transgene (drpIs3) in sox-2(ot460) null mutant animals rescued from lethality with a sox-2 fosmid-based extrachromosomal array (otEx4454[sox-2(fosmid)::mCherry + elt-2p::DsRed]). A mixed population of of sox-2 mutant (mosaic or non-rescued) and ‘wildtype’ (mutant-rescued) cells were scored.

We additionally analysed a strain containing a sox-2(ot640) null allele that was not available at the time of Y-to-PDA characterisation (Vidal et al., 2015). To bypass lethality, this strain contains an extrachromosomal sox-2 fosmid-based rescue array. The presence of this array in individual worms can be tracked by a fluorescent co-injection transgene, elt-2. Within an individual animal the array is inherited in a mosaic fashion and this is validated by the presence of dead embryos positive for the array. As with RNAi-knockdown, no statistically significant defect is observed in the PHD neuron (Figure 5C). In the head, one ectopic ida-1expressing MCM neuron is overserved in 3/266 sides per animal, coinciding with an extra AMso (Figure 5C and Figure 5—figure supplement 1C) and phenocopying the F1 RNAi results. However, four out of 269 total lin-48 expressing MCM cells lack ida-1 expression (Figure 5C and Figure 5—figure supplement 1C), suggesting a possible minor role of sox-2 in AMso-to-AMso+MCM, not detected in the RNAi-knockdown experiments.

Mutants in the same genes implicated in Y-to-PDA transdifferentiation were employed in this work and with the exception of sox-2, identical alleles were used. However, while the Y cell completely fails to transdifferentiate into PDA in these mutants (70–100% of the cases - Kagias et al., 2012), the PHso1-to-PHD and AMso-to-AMso+MCM cell fate switches show no or only very mild defects. Altogether these results suggest that different molecular mechanisms from the ones regulating Y-to-PDA must govern the PHso1-to-PHD and AMso-to-AMso+MCM cell fate switches.

PHDs are sensory neurons of male-specific copulation circuits

To further establish the neuronal characteristics of PHDs, we examined their synapses and ultrastructure, and mapped their full wiring diagram. We identified a ~1 kb promoter region that specifically drives the expression of oig-8 in the PHD neurons. Expression of a rab-3 translational fusion under the control of this oig-8 promoter reveals that PHDs form synapses in the pre-anal ganglion, where the synaptic vesicle associated mCherry-tagged RAB-3 protein can be observed (Figure 6A). At the ultrastructural level, we observed both synaptic and dense-core vesicles in PHD, proximal to the cell body (Figure 6B). Ultrastructural analysis of the PHD dendrite in seven different animals independently confirms John Sulston’s original observation on the presence of cilia, including the basal body and axonemes (Sulston et al., 1980). The basal body of PHD lies just dorsal and anterior to the phasmid sheath channel containing the PHA and PHB cilia. It has 2, 1, or 0 short cilia extending within the phasmid sheath very close to PHA and PHB cilia, but within a separate sheath channel. Interestingly, the PHD cilia can lie in variable positions relative to the PHA and PHB cilia: medial, lateral, dorsal, or ventral. In addition, we observe that the dendrite is more elaborate than previously described and has a number of unciliated finger-like villi within the phasmid sheath cell, proximal to the basal body (Figure 6C and Figure 6—figure supplement 1). Through reconstruction of serial electron micrographs all the synaptic partners of PHD were identified (Supplementary file 1; Jarrell et al., 2012). PHD neurons project from the dorsolateral lumbar ganglia anteriorly and then ventrally along the posterior lumbar commissure and into the pre-anal ganglion. In the pre-anal ganglion, these establish chemical synapses and gap junctions with sensory neurons and interneurons, most of which are male-specific (Figure 6D). Previously, the axonal process of PHD was attributed to the R8B ray neuron (Jarrell et al., 2012). A re-examination of the lumbar commissure allowed us to disambiguate PHD’s axon. PHDs receive synaptic input from male-specific ray sensory neurons involved in the initiation of the mating sequence in response to mate contact (Barr and Sternberg, 1999; Liu and Sternberg, 1995; Koo et al., 2011). The main PHD output is to the male-specific EF interneurons (35.4% of chemical synapses), both directly and through their second major post-synaptic target, the sex-shared PVN interneuron. The EF interneurons are GABAergic (Serrano-Saiz et al., 2017b) and synapse onto the AVB pre-motor interneurons which drive forward locomotion. Other PHD outputs include the PVV (male-specific) and PDB (sex-shared but highly dimorphic in connectivity; Jarrell et al., 2012; Cook et al., 2019) interneurons, which synapse onto body-wall muscle, and the cholinergic male-specific interneurons PVY and PVX whose output is to the AVA pre-motor interneurons, which drive backward locomotion. PVN, PVY, and PVX form disynaptic feed-forward triplet motifs that connect PHD strongly to the locomotion circuit interneurons AVB and AVA. The pattern of connectivity of PHD suggests a possible role in male mating behaviour.

Figure 6.

The PHDs are putative proprioceptive neurons of male-specific copulation circuits.

(A) Expression of the oig-8 and oig-8 reporter in the PHD neurons of adult males. The synapses made by the PHDs in the pre-anal ganglion (PAG) can be observed (ventral view). (B) Electron micrographs of the soma of a PHD neuron of an adult male. sv, synaptic vesicle; dcv, dense-core vesicles; PHDR, right PHD neuron. (C) As B for a PHD dendrite. (D) Diagram depicting the connectivity of the PHD neurons with their main pre-synaptic inputs (ray neurons) and post-synaptic targets. The connections between the ray neurons (RnA/B) and their post-synaptic targets independently of PHD are indicated in grey. Arrows and red lines indicate chemical and electrical synaptic connections, respectively. The thickness of the arrows is proportional to the anatomical strength of their connections (# serial sections). Neurons are colour-coded according to their neurotransmitter: red, cholinergic; yellow, glutamatergic; dark yellow, dopaminergic; blue, GABAergic; green, serotonergic; white, orphan. Note that some neurons (R8B, R9A, PVV) express more than one neurotransmitter. (E) Example traces showing PHD activity as normalised GCaMP/RFP fluorescence ratio in restrained animals. Traces are shown for a wildtype male, an unc-51(e359) mutant with and without a histamine-inducible silencing transgene in muscle (myo-3), a mutant in synaptic transmission (unc-13), and a mutant in dense-core vesicle exocytosis (unc-31, CADPS/CAPS). The proportion of traces where calcium peaks were identified is indicated for each genotype and treatment. n = number of neurons imaged. (F) Plots of frequency values of calcium transients per neuron. Dots represent individual neurons imaged. Tukey box-and-whisker plots indicate the interquartile ranges and median. *p<0.05; One-way ANOVA with multiple comparisons. Two groups were compared: unc-51 genotypes and treatments and another group including wt, unc-13, and unc-31. Only statistically significant comparisons are indicated.

Transmission electron microscopy images from serial transverse thin sections of an adult male tail, from anterior (A) to posterior (E). Dorsal is up. The level at which the sections are taken is labelled on the figure according to the cilia structures visible in PHD (compare to Figure 6). In all sections, the left and right PHD neurons are indicated with a boxed red R(ight) or L(eft). In A, the other phasmid neurons are also labelled and their cilia can be observed in each panel. (A) The basal body of the cilia is visible in PHDL and PHDR. (B) The PHD cilia are first visible in this section. (C) The PHD axonemes are visible in this section. (D) The finger-like villi are labelled in this section with arrowheads (but are also visible in all sections), and can be traced in serial sections to the basal body of PHD. Their appearance is identical to the more numerous villi that extend from the AFD cilium in the amphid (Ward et al., 1975).

Figure 6—figure supplement 1.

Ultrastructure of the male phasmid sensilla.

Transmission electron microscopy images from serial transverse thin sections of an adult male tail, from anterior (A) to posterior (E). Dorsal is up. The level at which the sections are taken is labelled on the figure according to the cilia structures visible in PHD (compare to Figure 6). In all sections, the left and right PHD neurons are indicated with a boxed red R(ight) or L(eft). In A, the other phasmid neurons are also labelled and their cilia can be observed in each panel. (A) The basal body of the cilia is visible in PHDL and PHDR. (B) The PHD cilia are first visible in this section. (C) The PHD axonemes are visible in this section. (D) The finger-like villi are labelled in this section with arrowheads (but are also visible in all sections), and can be traced in serial sections to the basal body of PHD. Their appearance is identical to the more numerous villi that extend from the AFD cilium in the amphid (Ward et al., 1975).

PHDs are putative proprioceptive neurons

We next sought to establish the function of the PHD neurons. We began by asking what sensory stimuli might activate them. To monitor neuronal activity, we co-expressed GCaMP6f and RFP in the PHDs and performed ratiometric measurements of fluorescence. We imaged calcium transients in restrained animals glued to a slide without anaesthetic. To our surprise, without applying any exogenous stimulation, we observed intermittent calcium transients in the PHDs of wildtype animals, every 30 s on average (Figure 6E–F and Video 1). We observed calcium transients in 73% of the neurons imaged and no peaks could be identified in the remaining traces (n = 19). Since in restrained animals there are small tail and spicule movements due to the defecation cycle and sporadic muscle contractions, one explanation could be that PHDs may be proprioceptive. We reasoned that if PHD activity resulted from internal tissue deformation caused by muscle contractions, inhibition of muscle activity should eliminate or reduce calcium transients. To inhibit muscle contractions, we generated transgenic worms in which muscles can be silenced in an inducible manner through the expression of the Drosophila histamine-gated chloride channel HisCl1 (Pokala et al., 2014), under the myo-3 promoter. To increase the efficiency of muscle silencing, we introduced this transgene into an unc-51(e369) mutant background that renders animals lethargic. unc-51 encodes a serine/threonine kinase required for axon guidance and is expressed in motor neurons and body-wall muscle (Ogura et al., 1997). Histamine-treated myo-3 animals were highly immobile and the frequency of calcium transients in the PHDs was strongly reduced (Figure 6E–F and Videos 2 and 3). Transients were completely eliminated in half of the neurons (Figure 6F). The reduction in transient frequency was specific to the silencing of the muscles because in histamine-treated unc-51(e369) and histamine non-treated myo-3 control animals, frequency was similar to that of wildtype animals (Figure 6E,F and Video 4).

Video 1.

Imaging of neuronal activity in PHD neurons with GCaMP6f (left channel) and RFP (right channel) in restrained animals: wildtype male.

Animals are expressing an oig-8 transgene. Videos play at 100 fps (recorded at 20 fps).

Video 2.

Imaging of neuronal activity in PHD neurons with GCaMP6f (left channel) and RFP (right channel) in restrained animals: unc-51(e359) male expressing a histamine-inducible silencing transgene in muscle (myo-3) and treated with 20mM histamine.

Animals are expressing an oig-8 transgene. Videos play at 100 fps (recorded at 20 fps).

Video 3.

Imaging of neuronal activity in PHD neurons with GCaMP6f (left channel) and RFP (right channel) in restrained animals: unc-51(e359) male expressing a histamine-inducible silencing transgene in muscle (myo-3) and treated with 20mM histamine.

Animals are expressing an oig-8 transgene. Videos play at 100 fps (recorded at 20 fps).

Video 4.

Imaging of neuronal activity in PHD neurons with GCaMP6f (left channel) and RFP (right channel) in restrained animals: unc-51(e359) male treated with 20mM histamine.

Animals are expressing an oig-8 transgene. Videos play at 100 fps (recorded at 20 fps).

Although PHDs are sensory neurons, they do receive some small synaptic input from other neurons (Figure 6D) and therefore the activity observed in restrained animals in response to muscle contractions could arise either directly through PHD sensory input or indirectly through pre-synaptic neurons. To test this, we imaged PHD activity in unc-13(e51) mutant males with impaired synaptic neurotransmission (Miller et al., 1996) and in unc-31(e169) mutant males with impaired dense-core vesicle secretion (Speese et al., 2007). In both these mutants, PHD calcium transients persisted at similar frequencies to those in wildtype animals (Figure 6E and F). This suggests that PHD activity does not require chemical input from the network. However, we cannot completely rule out that residual neurotransmission in the hypomorphic unc-13(e51) mutants may be sufficient to trigger wildtype levels of PHD activity in response to muscle contractions. Together, these results suggest that PHDs may respond directly to internal cues arising from muscle contractions, and that they may be proprioceptive neurons. PHDs may sense internal tissue deformations through their elaborate ciliated dendrites, which are deeply encased in the phasmid sheath and not exposed to the outside environment.

A novel readjustment step during male mating

The connectivity of PHD to male-specific neurons in the tail suggests that they may play a role in male reproductive behaviours controlled by tail circuits. In C. elegans, these include food-leaving behaviour, an exploratory strategy to search for mates (Lipton et al., 2004; Barrios et al., 2012; Barrios et al., 2008), and mating behaviour. Mating consists of a temporal sequence of discrete behavioural steps: response (to mate contact); scanning (backward locomotion while scanning the mate’s body during vulva search); turning (at the end of the mate’s body to continue vulva search); location of vulva (stop at vulva); spicule insertion (intromission); and sperm transfer (reviewed in Barr et al., 2018; Figure 7). The coordinated execution of these mating steps relies on sensory cues from the mate (Barr and Sternberg, 1999) and also, presumably, proprioceptive inputs within the male’s copulation circuit as well (Sulston et al., 1980 and reviewed in García, 2014). This sensory information guides the male to either initiate the next step of the sequence or to reattempt the current, unsuccessful step through readjustment of movement and/or posture (reviewed in Barr et al., 2018).

Figure 7.

The PHD neurons are required for coordinated backward locomotion and effective intromission during mating.

Diagram depicting the steps of reproductive behaviours controlled by the male tail circuits. The steps affected by PHD ablation are highlighted in red. Intact (mock) and ablated males carried either an oig-8 or an unc-17 transgene to identify the PHD neurons for ablation. In F, males carrying an oig-8 transgene were used to silence PHD neurons acutely. White arrows indicate the transitions between the mating sequence. Grey arrows indicate the corrective transitions that males perform when they fail to attain the subsequent goal. The corrective transitions that occur upon failure of spicule insertion attempts (between location of vulva and spicule insertion) are always preceded by a displacement from the vulva (not depicted). Behavioural analysis in intact and PHD-ablated males are shown for each step. Calcium imaging in PHD neurons is shown for steps C, F, and H. The black trace shows PHD activity as normalised GCaMP/RFP fluorescence ratio changes averaged for several events and phase locked (red dotted line) to the switch in the direction of locomotion (C and F) or to the start of spicule insertion (H). The grey shadow shows S.E.M. (A) Male exploratory behaviour measured as PL values (probability of leaving food per worm per hour). n, number of males tested. Maximum likelihood statistical analysis was used to compare PL values. n.s., no statistically significant difference, p≥0.05; error bars, S.E.M. (B) Response efficiency, measured as 1/number of contacts with a mate before responding; and hesitation, measured as a switch in direction of locomotion during response. (C) Scanning locomotion during vulva search. Categories: switching (change in the direction of locomotion from backward to forward); pause (stopping during backward scanning); mixture (scans with switches and pauses); continuous (uninterrupted backward movement along the mate’s body). p<0.001 (χ2 test of continuous and discontinuous scans); n = number of events. (D) Turning, measured as proportion of good turns per male. (E) Vulva location efficiency measured as 1/number of vulva encounters before stopping. (F) Proportion of continuous and discontinuous Molina manoeuvres. Categories: switching (a brief change in the direction of movement either during forward locomotion away from the vulva or during backward locomotion returning to the vulva); STOP during forward (stopping during forward movement away from the vulva and then continuing in the same direction); STOP before returning backward (stopping at the transition between forward movement away from the vulva and returning backwards to the vulva); mixture (manoeuvres that displayed more than one of the discontinuities described above); continuous (smooth movement forward away from the vulva and return backwards to the vulva without stopping or switching in between). p<0.001 (χ2 test of continuous and discontinuous manoeuvres); n = number of events. (G) Number of displacements away from the vulva per unit of time spent at vulva. (H) Left bar chart, proportion of males able to insert their spicules; n.s., no statistically significant difference, p≥0.05 (χ2 test); n = animals tested. Right bar chart, sperm transfer efficiency measured as percentage of cross-progeny; p<0.001 (χ2 test); n = total progeny. For B, D, E, and G, bar and dots represent mean and individual animal values, respectively; error bars, S.E.M. n.s., no statistically significant difference, p≥0.05 (Mann-Whitney U test). Worm cartoons were modified with permission from original drawings by Rene García.

(A) DIC and fluorescent images of a one-day adult male expressing the unc-17 transgene in both PHDR and PHDL neurons in the tail, before the laser ablation. (B) Same animal after the PHDR neuron has been ablated. The laser photobleaches all GFP signal in the region. (C) Same animal after 24 hr after the ablation. The GFP signal from the transgene reappears in the non-ablated PHDL and in other structures of the male tail but is specifically lost in the ablated PHDR. Of note, the GFP signal coming from the PHDL is dimmer due to the older age of the animal (2 day adult).

Plots show proportion of events with ectopic prodding in mock and PHD-ablated males. n = events: (A) all Molina manoeuvres; (B) discontinuous Molina manoeuvres; (C) all scans; (D) all pauses while scanning. For A, p<0.05 (χ2 test). For B, C, and D, p≥0.05 (χ2 test).

During our behavioural analysis of wildtype males, we identified a novel readjustment movement that has not been previously described which we have termed the ‘Molina manoeuvre’. This movement occurs when the male has been trying to insert its spicules in the mate’s vulva for a period of time without success and subsequently loses vulva apposition. Unsuccessful spicule insertion attempts occurred in 65% of males, resulting in either a long displacement from the vulva, which led to the re-initiation of the scanning sequence, or in a small displacement (less than two tail-tip length) from the vulva (98% of vulva losses), as previously described (Correa et al., 2012). We find that these small displacements from the vulva lead either to local shifts to relocate the vulva (68% of total vulva losses) or to a Molina manoeuvre (32% of vulva losses) (Figure 7). The Molina manoeuvre consists of the initiation of forward locomotion along the mate’s body away from the vulva to the end of the mate’s body (or at least two male tail-tips distance away from the vulva), at which point the male tail acquires a deeply arched posture, followed by a return to the vulva with backward locomotion along the same route (Video 5). Males perform this movement towards either end (head or tail) of the mate as a smooth, continuous sequence. Although we used paralysed unc-51 mutant hermaphrodites to aid our mating analysis, males also perform this readjustment movement with moving wildtype mates (Video 6).

Video 5.

Males performing Molina manoeuvres during mating: wildtype male performing a Molina manoeuvre during mating with a paralysed unc-51(e359) hermaphrodite.

Videos are played at 40 fps (sped up x2).

Video 6.

Males performing Molina manoeuvres during mating: wildtype male performing a Molina manoeuvre during mating with a wildtype hermaphrodite.

Videos are played at 40 fps (sped up x2).

The PHD neurons are required for coordinated backward locomotion and effective intromission during mating

To test the hypothesis that PHDs regulate reproductive behaviours, we performed behavioural tests of intact and PHD-ablated males and functional imaging of PHD activity in freely behaving males during mating. Target specificity of the laser ablations was confirmed 1-day after ablation by absence of fluorescence returning from the transgene used to label the PHD neurons (Figure 7—figure supplement 1A–C). We found no defects in food-leaving behaviour, response to mate contact, turning, location of vulva, or spicule insertion (Figure 7A,B,D,E and H). However, these experiments revealed a role for PHDs in initiation and/or maintenance of backward locomotion during scanning and during the Molina manoeuvre. PHD-ablated males often switched their direction of locomotion during scanning, performing fewer continuous backward scans during vulva search compared to intact males (Figure 7C). Consistent with this, we observed higher activity (i.e. Ca2+ levels) in PHD during backward locomotion than during forward locomotion while scanning (Figure 7C). PHD activity also peaked just after the switch to backward locomotion during Molina manoeuvres, when the tail tip displays an acute bent posture, and PHD-ablated males performed defective, discontinuous manoeuvres, often stopping at the transition from forward to backward locomotion to return to the vulva (Figure 7F and Video 7). Importantly, we did not find significant differences between intact and PHD-ablated males in number of events that trigger the initiation of manoeuvres (Figure 7G), indicating that PHDs are specifically required for coordinated locomotion during the manoeuvre itself, at the transition point to backward locomotion. Similar locomotion defects resulting in discontinuous manoeuvres were observed when PHD neurons were acutely silenced with an inducible chloride channel transgene (oig-8 (Figure 7G). Together, these data demonstrate that without intact PHD neurons, backward movement along the mating partner becomes slightly erratic, often interrupted by a switch in direction (during scanning), or its initiation is delayed by a stop (during Molina manoeuvres). The qualitative difference in locomotion defects between these two steps of mating may result from differences in the state of the neural network before and after spicule insertion attempts (Correa et al., 2015; Correa et al., 2012; Liu et al., 2011). Sometimes while pausing, particularly during Molina manoeuvres but also during scanning, wildtype and PHD-ablated males prodded their spicules at locations other than the vulva (Figure 7—figure supplement 2). Ectopic prodding happened rarely but it occurred significantly more often in ablated than in wildtype males (Figure 7—figure supplement 2A). Because ectopic prodding occurred only during pauses and PHD-ablated males produced many more discontinuous manoeuvres than wildtype males, we interpret the increase in ectopic prodding as a secondary consequence of pausing rather than as the main defect of PHD-ablated males. Consistent with this interpretation, there was no statistically significant difference in the occurrence of ectopic prodding events between ablated and wildtype males when only discontinuous manoeuvres were considered (Figure 7—figure supplement 2B).

Figure 7—figure supplement 1.

PHD neuron ablation control.

(A) DIC and fluorescent images of a one-day adult male expressing the unc-17 transgene in both PHDR and PHDL neurons in the tail, before the laser ablation. (B) Same animal after the PHDR neuron has been ablated. The laser photobleaches all GFP signal in the region. (C) Same animal after 24 hr after the ablation. The GFP signal from the transgene reappears in the non-ablated PHDL and in other structures of the male tail but is specifically lost in the ablated PHDR. Of note, the GFP signal coming from the PHDL is dimmer due to the older age of the animal (2 day adult).

Video 7.

Males performing Molina manoeuvres during mating: PHD-ablated male performing a defective, discontinuous Molina manoeuvre.

Videos are played at 40 fps (sped up x2).

Figure 7—figure supplement 2.

Some ectopic prodding occurs during discontinuous Molina manoeuvres and during pauses while scanning.

Plots show proportion of events with ectopic prodding in mock and PHD-ablated males. n = events: (A) all Molina manoeuvres; (B) discontinuous Molina manoeuvres; (C) all scans; (D) all pauses while scanning. For A, p<0.05 (χ2 test). For B, C, and D, p≥0.05 (χ2 test).

In addition to the aforementioned changes in neuronal activity, the highest level of PHD activation was observed during intromission, the penultimate step of the mating sequence, which precedes sperm transfer and involves full insertion of the spicules into the mate’s vulva while sustaining backward locomotion. PHD activity increased two-fold upon spicule insertion and remained high for several seconds while the spicules were inserted (Figure 7H). PHDs are required for the efficiency of these two last steps of mating since PHD-ablated males that were observed to complete the mating sequence, including intromission and ejaculation, during a single mating with one individual mate, produced fewer cross-progeny than intact males in the same conditions (Figure 7H). Intact PHDs may increase the efficiency of sperm transfer by controlling the male’s posture during intromission. While the cues and network interactions that activate PHDs during mating are presently not known, the high levels of PHD activity upon spicule insertion are consistent with a putative proprioceptive role for these neurons.

Discussion

The data presented here extend John Sulston’s original observations and demonstrate that male PHso1 cells undergo a direct glia-to-neuron transdifferentiation to produce a novel class of bilateral cholinergic, peptidergic, and ciliated sensory neurons, the phasmid D neurons. This updates the anatomy of the adult C. elegans male: it now comprises 387 neurons (93 of which are male-specific) and 90 glia (two are lost during the transdifferentiation to the PHDs). This is the second example of neurons arising from glia in C. elegans, confirming that glia can act as neural progenitors across metazoan taxa. In both cases we have demonstrated that fully differentiated, functional glia can retain neural progenitor properties during normal development. The complete glia-to-neuron cell fate switches we observe strongly suggest the production of neurons from glia is a process of natural transdifferentiation. In the first case (Sammut et al., 2015), this is an indirect transdifferentiation as it occurs via an asymmetric cell division, leading to self-renewal of the glial cell and the production of a neuron. In contrast, the work described here characterises a direct transdifferentiation and as such is the second direct transdifferentiation to be described in C. elegans. The first involves the transdifferentiation of the rectal epithelial cell Y into the motor neuron PDA, which occurs in hermaphrodites at earlier larval stages, before sexual maturation (Jarriault et al., 2008). Excitingly, most of the key genes that regulate the Y-to-PDA cell fate switch (Kagias et al., 2012; Richard et al., 2011; Zuryn et al., 2014) appear dispensable for PHso1-to-PHD and AMso-to-AMso+MCM transdifferentiation, indicating that independent molecular mechanisms exist that control cell-fate plasticity in a sex-specific, time-specific and/or a cell-specific manner. Intriguingly, direct conversion of glial-like neural stem cells has been observed in the adult zebrafish brain (Barbosa et al., 2015), suggesting that this may be a conserved mechanism for post-embryonic generation of neurons. Furthermore, it will be interesting to understand the molecular basis underlying the variability in the low frequency, background-dependent division of PHso1 and whether it has any functional or evolutionary consequence. The fact that a cell’s terminal division can be variable is especially surprising in C. elegans, as its cell lineage is considered highly stereotyped and invariable. The only other neurons known to be variable in their generation are the post-embryonic and male-specific DX3/4 and EF3/4 interneurons located in the pre-anal ganglion (Sulston et al., 1983). It is intriguing that the EF neurons are the main post-synaptic partners of PHD (Figure 6D), and as such it is tempting to speculate there may be a possible link between cell division and cell connectivity.

Based on our manipulations and imaging of PHD activity in restrained and freely mating animals, we propose that PHDs may be proprioceptive neurons which become activated by tail-tip deformations and engage the circuits for backward locomotion. The high levels of PHD activity during intromission and upon tail-tip bending during the Molina manoeuvre, steps which require inhibition of forward locomotion and sustained backward movement, are consistent with such a model. PHDs may ensure sustained backward locomotion through excitation of their post-synaptic targets PVY, PVX, and EF interneurons, all of which have been shown to promote backward movement through AVA and AVB during scanning and location of vulva (Sherlekar and Lints, 2014; Sherlekar et al., 2013). As wildtype hermaphrodites are highly uncooperative during mating and actively move away (Kleemann and Basolo, 2007), any erratic and uncoordinated movement by the male is likely to result in the loss of its mate.

Sensory input from the mating partner is essential for successful mating and accordingly, many sensory neurons in copulation circuits are dedicated to this purpose (reviewed in Villella and Hall, 2008 and Barr et al., 2018). However, coordinated motor control in all animals also requires proprioception, sensory feedback from internal tissues that inform the individual about its posture and strength exerted during movement (Sherrington, 1907; Tuthill and Azim, 2018). Several putative proprioceptive neurons have been identified in C. elegans (reviewed in Schafer, 2015). Within the male copulation circuit, a putative proprioceptive role has been attributed only to the spicule neuron SPC, on the basis of the attachment of its dendrite to the base of the spicules (Sulston et al., 1980). The SPC neuron is required for full spicule protraction during intromission (García et al., 2001; Liu and Sternberg, 1995). The presence of proprioceptive neurons in copulation circuits may be a broadly employed mechanism to regulate behavioural transitions during mating. In this light, it would be interesting to determine whether the mechanosensory neurons found in the mating circuits of several other organisms (Liu et al., 2007; Ng and Kopp, 2008; Pavlou et al., 2016) may play a role in self-sensory feedback.

In summary our results provide new insight into the developmental mechanisms underlying the neural substrates of sexually dimorphic behaviour. The glia-to-neuron transdifferentiation that results in PHD represents an extreme form of sexual dimorphism acquired by differentiated sex-shared glial cells during sexual maturation. The PHD neurons, perhaps through proprioception, enable the smooth and coordinated readjustment of the male’s movement along its mate when spicule insertion becomes difficult to attain. This readjustment movement represents an alternative step within the stereotyped mating sequence. These findings reveal the extent to which both cell fate and innate behaviour can be plastic yet developmentally wired.

Materials and methods

C. elegans strains

Nematode culture and genetics were performed following standard procedures (Brenner, 1974; Stiernagle, 2006). For genetic crosses, all genotypes were confirmed using PCR unless the mutant phenotype was obvious (see Key Resource Table for primer sequences). Animals considered wildtype carry him-5 or him-8 mutations to facilitate male analysis. All assays were conducted at 20°C. Strains used in this study are listed in the Key Resource Table.

DNA constructs and transgenic strains

The oig-8 reporter was built by amplifying a 964 bp promoter fragment upstream of the oig-8 ATG adding the restriction sites of SphI and XmaI (see Key Resource Table for these and other primer sequences). This promoter drives expression in PHD neurons and two pairs of sensory neurons in the head which we have not identified. The PCR product was digested and ligated into SphI/XmaI-digested pPD95.75 vector. To generate oig-8, the same oig-8 promoter fragment was ligated into SphI/XmaI-digested pkd-2 (a gift from M. Lázaro-Peña). The oig-8 construct was created by PCR fusion (Hobert, 2002). A ~ 1 kb promoter region of the oig-8 locus was PCR fused with the GCaMP sequence. The GCaMP6f::sl2::rfp fragment was amplified from gateway plasmid pLR306, a kind gift from Luis Rene García. The same promoter region was used to create the oig-8 plasmid by Gibson cloning. The myo-3 construct was created as PCR fusion. A 2.2 kb promoter region of the myo-3 locus was amplified and PCR fused with the HisCl1 sequence (see Key Resource Table for primer sequences). The HisCl1::sl2::mCherry fragment was amplified from plasmid pNP471, a kind gift from the Bargmann lab. The grl-2 reporter was generated by integrating sEx12852 (Hao et al., 2006) to generate drpIs1. The lin-48 construct was a kind gift from Mike Boxem and contains a 6.8 kb promoter fragment upstream of the coding sequence of tdTomato and the unc-54 3’ UTR. It was injected into him-5(e1490) animals and spontaneously integrated generating drpIs3.

PHso1-to-PHD single animal live imaging

Animals containing the lin-48 transgene were synchronised via sodium hypochlorite treatment and allowed to develop until the early L3 stage. Next, they were mounted individually on 5% agarose pads in a droplet of the minimum concentration of sodium azide that paralysed the animals (which varied within experimental replicates) and viewed under a compound microscope to determine their developmental time. A total of 10–20 males in which the gonad had just turned to grow posteriorly were imaged. Care was taken to use the lowest fluorescent light intensity and exposure in order to minimise toxicity. After the initial imaging, animals were gently recovered by aspiration, washed in a droplet of M9 buffer, transferred to individual plates and allowed to develop overnight at 15°C. Each animal was then immobilised and imaged as described above. In those animals where the leading edge of the gonad had crossed back over to the ventral side, a sequence of time-lapse images was taken every 2–4 hr until PHso1-to-PHD remodelling was complete, 10–12 hr later.

Immunohistochemistry and EdU staining

EdU (5-ethynyl-2′-deoxyuridine) feeding was performed as previously described (van den Heuvel and Kipreos, 2012) combined with immunohistochemistry to enhance reporter transgene fluorescence. In short, MG1693 bacteria (Thy-deficient, E. coli stock centre: http://cgsc.biology.yale.edu/) were grown in 100 ml of minimal medium containing 20 mM EdU for at least 24 hr, concentrated by centrifugation and seeded on NGM plates with 100 μg/mL ampicillin. A synchronised population of L1 larvae, obtained by hypochlorite treatment and hatching in M9 buffer, was allowed to feed on OP50 E. coli until early L3 (24–28 hr at 20°C). Worms were washed thoroughly in M9, plated onto MG1693 EdU containing plates and allowed to grow until late L4 – early adult stage (22–24 hr). Hard-tube fixation was chosen in order to preserve the male tail morphology (McIntire et al., 1992). Worms were washed thoroughly, transferred to DNA LoBind tubes (Eppendorf) and fixed for 12 hr in 4% paraformaldehyde at 4°C in a rocking nutator, followed by 5% β-mercapto-ethanol for another 12 hr at 37°C in a rocking nutator. Worms were then incubated for 6 hr in 1 mg/ml collagenase (Sigma Aldrich) at 37°C and 700 rpm, followed by 24 hr with rabbit anti-RFP antibody (1:500, MBL) to detect the tdTomato fluorescence from the drpIs3 reporter transgene. Alexa 555-conjugated donkey anti-rabbit (1:200; Molecular probes) was used as secondary antibody. As EdU reaction can interfere with antibody staining, this was performed after the immunohistochemistry. The Click-IT reaction was performed in a total volume of 50 µl per tube, in the dark, according to the manufacturer's instructions (Click-IT EdU Alexa Fluor 594 kit, Invitrogen). Slides were mounted in Vectashield antifade mounting medium (Vector Laboratories). Samples were analysed the same day or the day after mounting, to minimise fading of fluorescence. Only animals where both head and tail were properly stained for tdTomato (cell body) and EdU (nucleus) were scored. Four independent EdU experiments were performed (see Source Data Figure 3) and all animals scored were plotted in the same bar chart.

RNAi feeding experiments

The RNAi bacterial clone for sox-2 was obtained from the Ahringer RNAi Collection (Source BioScience) (Kamath et al., 2003). RNAi feeding experiments were performed following standard protocols (Kamath et al., 2001). Briefly, isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to a final concentration of 0.6 mM to NGM medium to prepare the RNAi plates. 24 hr later, HT115 bacteria transfected with the sox-2 RNAi clone was seeded onto RNAi plates. Staged early-L4 larvae were transferred to seeded RNAi plates and 4–5 days later their progeny, which had developed under the embryonic effects of RNAi knock-down (F1 generation), were scored at the stage of young adult. To exclusively analyse post-embryonic effects, a synchronised population of L1 larvae was transferred to RNAi seeded plates and young adult males were scored 2–3 days later (P0 generation).

All experiments were performed at 20°C. The nre-1(hd20) lin-15b(hd126) background was used to sensitise worms to the RNAi effects. One experiment was performed with three replicate plates per condition. L4440 empty vector (pPD129.36, Addgene) was used as negative control and gfp RNAi clone as positive control.

Cell-type specific sex transformations

We used two previously described strains each containing an array that drives fem-3 expression from a grl-2 promoter fragment (Sammut et al., 2015). This grl-2 promoter is expressed in PHso1 and PHso2 in the tail, and AMso, excretory pore, and excretory duct in the head (Hao et al., 2006). In the case of oleEx24 the presence/absence of the array in PHso1 and PHso2 was monitored by the mCherry expression from the array itself and the expression of an ida-1 reporter (inIs179), in the background of the strain was used to monitor neuronal fate in PHso1/PHD. In the case of oleEx18, we had difficulty visualising the mCherry from the array in PHD and crossed a lin-48 transgene (drpIs3) into the strain to visualise PHso1/PHD. The presence/absence of the array in whole animals was assessed using the array co-injection marker elt-2. Neuronal fate was monitored using a rab-3 reporter.

Cell ablations

PHD was ablated with a laser microbeam as previously described (Bargmann and Avery, 1995). L4 males were staged and placed on a seeded plate the night before. Ablations were carried out at one-day of adulthood and PHD was identified by oig- or unc-17 reporter expression. Mock-ablated animals underwent the same treatment as ablated males except for laser trigger. Animals were left to recover overnight and behavioural assays were performed the next day. After behavioural assays, animals were checked for lack of GFP expression in the tail-tip region where PHD is normally located to confirm correct cell ablation. The few animals in which PHD had not been efficiently ablated (judged by the presence/return of GFP) were discarded from the data.

Behavioural assays

All behavioural assays were scored blind to the manipulation. Males carried either an oig-8 or an unc-17 transgene to identify the PHD neurons for ablation. For the experiments in which PHD was silenced with a HisCl1 transgene, mating assays were performed in 2-day-old adults, as in the ablations. Before the assay, wildtype or array-carrying animals with RFP-positive PHD neurons were moved to a 20 mM histamine plate with food for 1 hr. As controls, array-carrying animals not treated with histamine were also tested. Histamine plates were prepared as regular NGM plates adding 20 mM histamine when the agar cooled to a temperature below 56°C.

Food leaving

Animals were tested as 3-day-old adults (the day after they had been tested for mating). Assays were performed and scored as previously described (Barrios et al., 2008).

Mating

Assays were performed and scored as previously described (Sammut et al., 2015). Males were tested at two days of adulthood with 1-day-old unc-51(e369) hermaphrodites picked the night before as L4s. Each male was tested once for all steps of mating. Those males that were not successful at inserting their spicules with the first hermaphrodite were tested again with a maximum of three hermaphrodites to control for hermaphrodite-specific difficulty in penetration (Liu and Sternberg, 1995). Assays were replicated at least three times on different days and with different sets of males. Videos of mating events were recorded and scored blind by two independent observers.

Response

A male was scored as responding to mate contact if it placed its tail ventral down on the mate’s body and initiated the mating sequence by backing along the mate’s body to make a turn. The response efficiency was calculated by dividing 1 (response) by the total number of contacts made with the mate before responding. If a male did not respond within three minutes, it was scored as having 0 efficiency. As a more sensitive measure of the quality of response, we scored hesitation during response. Hesitation is a switch in direction between forward and backward locomotion from the time the male establishes contact with the mate to the first turn (or to location of vulva if this occurs without the need of a turn).

Scanning

A single scan was scored as the journey around the mate’s body away from and returning to the vulva position. The first scan was counted as the journey from the point of first contact to the hermaphrodite vulva position. A scan was considered continuous if locomotion was maintained in the backward direction without switching direction or pausing.

Ectopic prodding

Prodding was scored as a visible protrusion of the spicules out of the gubernaculum and/or visible twitching of the spicule muscles. Prodding was considered ectopic if it occurred at regions of the mate’s body other than the vulva.

Molina manoeuvres

A continuous single manoeuvre was scored as the journey away from the vulva in forward locomotion, to a distance bigger than two tail-tip lengths, and return to the vulva in backward locomotion. Since the transition between the forward and backward steps could take up to 3 s in a wildtype male, the category of discontinuous manoeuvre ‘stop before backward step’ was scored if the pause was longer than 3 s. Any other visible pause during forward or backward locomotion was considered a STOP regardless of its duration. The category of discontinuous manoeuvre ‘switching’ was scored as a change in direction of locomotion while travelling away or towards the vulva without reaching it.

Displacements from the vulva

Scored as movements of one or two tail-tip lengths away from the vulva.

Fertility assays

Each individual male was monitored for all mating steps during a single mating until it ejaculated. After disengagement from the mate, the hermaphrodite was picked and placed in a fresh plate to lay progeny. The adult hermaphrodite was transferred to a fresh plate each day during three consecutive days. After three days from the eggs being laid, L4 larvae and adult progeny were counted as Unc self-progeny of Wt cross-progeny.

Microscopy and imaging

Worms were anesthetised using 50 mM sodium azide and mounted on 5% agarose pads on glass slides. Images were acquired on a Zeiss AxioImager using a Zeiss Colibri LED fluorescent light source and custom TimeToLive multichannel recording software (Caenotec). Representative images are shown following maximum intensity projections of 2–10 1 µm z-stack slices and was perfumed in ImageJ.

Ca2+ imaging

Imaging was performed in an upright Zeiss Axio Imager two microscope with a 470 nm LED and a GYR LED (CoolLED) with a dual-band excitation filter F59-019 and dichroic F58-019 (Chroma) in the microscope turret. Emission filters ET515/30M and ET641/75 and dichroic T565lprx-UF2 were placed in the cube of a Cairn OptoSplit II attached between the microscope and an ORCA-Flash four camera (Hamamatsu). Acquisition was performed at 20 fps. Imaging during mating was performed with a 20x long working distance objective (LD Plan-NEOFLUAR numerical aperture 0.4), placing the male on an agar pad with food and 20 hermaphrodites. The ~50 mm per side agar pad was cut out from a regular, seeded NGM plate and placed on a glass slide. The hermaphrodites were placed in a ~ 100 mm2 centre region. A fresh pad was used every two recordings.

Imaging in restrained animals was performed for 2.5–3 min with a 63x objective (LD C- apochromat numerical aperture 1.15). Animals were glued with Wormglu along the body to a 5% agarose pad on a glass slide and covered with M9 or 20 mM histamine (Sigma, H7125) and a coverslip.

Ca2+ imaging analysis

A moving region of interest in both channels was identified and mean fluorescent ratios (GFP/RFP) were calculated with custom-made Matlab scripts (Busch et al., 2012), kindly shared by Zoltan Soltesz and Mario de Bono. For recordings in restrained animals, bleach correction was applied to those traces in which an exponential decay curve fitted with an R square >0.6. Ratios for each recording were smoothened using an 8-frame rolling average. For ΔR/R0 values, R0 for each recording period was calculated as the mean of the lowest 10th percentile of ratio values. Traces that were locked to behavioural transitions had their ΔR/R0 values added or subtracted such that all events had the same value at t = 0.

Peaks were identified manually by an observer blind to the genotype and treatment. Peaks were called as signals above 0.2 ΔR/Rmax (where Rmax was calculated as the highest 5th percentile of ratio values), above 2σ from local basal and a minimum duration of 5 s.

Electron microscopy and serial reconstruction

The samples were fixed by chemical fixation or high-pressure freezing and freeze substitution as previously described (Hall et al., 2012). Several archival print series from wildtype male tails in the MRC/LMB collection were compared to wildtype adult males prepared in the Hall lab, showing the same features overall. Ultrathin sections were cut using an RMC Powertome XL, collected onto grids, and imaged using a Philips CM10 TEM. The PHD cell bodies were identified in the EM sections based on position and morphology. This was followed by serial tracing of the projections to establish their morphology and connectivity. The method of quantitative reconstruction using our custom software is described in detail in Jarrell et al., 2012 and Xu et al., 2013. The connectivity of the PHD neurons was determined from the legacy N2Y EM series (Sulston et al., 1980). Circuit diagrams were generated using Cytoscape (Smoot et al., 2011). PHD neuronal maps and connectivity tools are available at www.wormwiring.org.

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

In this work, the authors add to findings showing that neurons can arise via transdifferentiation from other cell types in C. elegans. Here, they show that a sex-shared socket glial cell transdifferentiates to a male-specific sensory neuron type, and suggest that these neurons play a proprioceptive role in male mating behavior. This work provides insights into mechanisms by which the nervous system generates sexually dimorphic behaviors.

Decision letter after peer review:

Thank you for submitting your article "A direct glia-to-neuron natural transdifferentiation ensures nimble sensory-motor coordination of male mating behaviour" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Ronald Calabrese as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Douglas Portman (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

All reviewers were supportive of this work, noting that the discovery of a sex-specific glia-to-neuron transdifferentiation event, and the correlation of this neuron with behavioral phenotypes, is of great interest. However, since the existence of this cell was previously suggested by Sulston, and both transdifferentiation and the role of glia as neuronal precursors have been reported previously, all reviewers felt (and I concur) that this work requires additional mechanistic insights at this point.

Essential revisions:

1) Strengthen the evidence for transdifferentiation. As currently presented, it is not completely clear whether this is direct transdifferentiation or whether a cell division event occurs. A few experimental avenues were suggested. These include:

a) Examining whether you detect an apoptotic cell in the vicinity of PHso1 during its transdifferentiation. If not, this might help bolster the direct transdifferentiation model, as opposed to division of PHso1 followed by rapid death of one of the daughters. Along these lines, also consider looking at a canonical cell death mutant.

b) Provide time lapse video of the differentiation event so that the cell dynamics can be easily followed. A reviewer felt that the images in Figure 2 were not very clear and suggested using a membrane-tagged reporter for easier visualization.

c) A number of genes have previously been shown to regulate the Y-to-PDA transdifferentiation (Jarriault and others). Determining whether these genes or a subset also play a role in the PHso1 transdifferentiation process would help to establish whether the processes are shared or distinct and begin to outline the underlying mechanisms.

2) Clarify and strengthen the behavioral analysis. The correlation of the PHD neuron with the 'Molina maneuver' is a particular strength of this work.

a) You nicely propose that the PHDs couple arched tail posture to reverse locomotion. However, one would also expect a similar requirement at the beginning of contact-response behavior, where linking an arched tail posture to reverse locomotion is also important. Please determine whether response behavior is less smooth, even if it ultimately occurs just as frequently as in wildtype.

b) In Video 4, it seems that, after moving away from the vulva, the male repeatedly attempts intromission "ectopically", near the tail. It's hard to tell from the video whether spicule prodding is happening at this time, but if it is, could it be that PHD is not required for the Molina manoeuvre itself, but rather to prevent inappropriate spicule prodding at locations away from the vulva? Please look for pausing/prodding at locations other than the vulva during scanning behavior. If this is observed, it could be that the PHDs instead inhibit Lov (location-of-vulva) behavior.

c) Videos 3, 4, S3: it would be helpful to indicate in the video itself exactly when the "Molina manoeuvre" takes place and, in Video 4, when the defect occurs. Also, it should be noted somewhere that these videos are sped up 2x from real time.

d) While the ablation data are nice, there are some issues regarding how you determined that the cells were ablated, and the absence of an orthogonal method of ablation for confirmation. Reviewers suggested a couple of possibilities for experiments: you have a PHD-specific oig-8 promoter version. You could try to feminize driving TRA-2(IC) under this promoter and checking behavior although it is possible that this promoter comes on too late to feminize. Alternatively, you could silence/ablate PHD using this promoter to confirm the behavioral phenotype.

Summary:

All reviewers were supportive of this work, noting that the discovery of a sex-specific glia-to-neuron transdifferentiation event, and the correlation of this neuron with behavioral phenotypes, is of great interest. However, since the existence of this cell was previously suggested by Sulston, and both transdifferentiation and the role of glia as neuronal precursors have been reported previously, all reviewers felt (and I concur) that this work requires additional mechanistic insights at this point.

We have now performed a number of additional experiments – described in more detail below – that provide several additional cellular and mechanistic insights. In summary we find that, at the molecular level, the “plasticity cassette” described by Sophie Jarriault’s lab is not required for either PHso1-to-PHD or AMso-to-AMso’+MCM. This suggests not all transdifferentiations in C. elegans are equivalent and that the genes involved in Y-to-PDA transdifferentiation are not the only plasticity genes. We also now provide data that demonstrate that PHso1 actually divides at low frequency and in a background dependent manner to generate PHD1/PHD2. We find this variability incredibly intriguing of course, as the only other variable lineages in C. elegans are the EF neurons, the DX neurons and P3.p vulval precursor cells. The EFs are the major post-synaptic target of the PHDs. We are continuing our work on both the molecular regulation and variable cell division aspects in our ongoing studies, outside the scope of this manuscript.

Although the potential existence of PHD had indeed been suggested by John Sulston, he also stated “no other neuronal characteristics were observed” aside from the presence of cilia. Here we demonstrate the acquisition of neuronal features at the morphological, ultrastructural and molecular levels as well as the function of the PHD neuron in male mating behaviour. We think that a second report of a glia-to-neuronal cell fate switch actually cements this process as a fundamental mechanistic principle of neural development (which in many ways is very similar to vertebrate adult neurogenesis). This is also only the second well-described example of a direct transdifferentiation in C. elegans (the other being Y-to-PDA). The number of transdifferentiation paradigms described with clear lineage relationships pre/post cell fate switch and at such cellular resolution are few and far between and so should be of interest to eLife readers.

Essential revisions:

1) Strengthen the evidence for transdifferentiation. As currently presented, it is not completely clear whether this is direct transdifferentiation or whether a cell division event occurs. A few experimental avenues were suggested. These include:

As mentioned above, we have now performed a number of additional experiments to address this and we thank the reviewers for pushing us to look into this in much more detail. We find that ~80% of the time PHso1-to-PHD is a direct transdifferentiation. However, in ~20% of cases (variable depending on background) PHso1 actually divides to generate what we have termed PHD1 and PHD2. We note that in all our behaviour experiments we ensured following laser ablation that no PHDs remained and that the drivers we used to silence the cells would also silence PHD2. We find PHD2 indistinguishable from PHD1 in terms of gene expression.

a) Examining whether you detect an apoptotic cell in the vicinity of PHso1 during its transdifferentiation. If not, this might help bolster the direct transdifferentiation model, as opposed to division of PHso1 followed by rapid death of one of the daughters. Along these lines, also consider looking at a canonical cell death mutant.

John Sulston originally described all the apoptotic events in the male tail and in corroboration of his data we do not observe any apoptotic cell corpses in the vicinity of PHso1. However, given that the male tail is undergoing quite dramatic morphogenetic changes at the same time as PHso1-to-PHD transdifferentiation, we have also analysed ced-4(n1164) mutants. We find no evidence for a cell death arising from PHso1 (Figure 3—figure supplement 1). To our surprise however, at a low frequency we did observe the presence of one or two extra cells in the tail (maximum one extra cell per side) co-expressing lin-48 and rab-3 in both mutant and control animals and have now investigated the origin of these cells in more detail (see below).

b) Provide time lapse video of the differentiation event so that the cell dynamics can be easily followed. A reviewer felt that the images in Figure 2 were not very clear and suggested using a membrane-tagged reporter for easier visualization.

We now provide, in an updated version of Figure 2, a time-series of an individual animal imaged every 2–4 h over the 10–12 h period of PHso1-to-PHD transdifferentiation. We think that this experiment unambiguously demonstrates that in the majority of cases PHso1-to-PHD is a direct transdifferentiation.

Following our observation of additional cells expressing lin-48/rab-3 in the vicinity of PHD (see above), we have now characterised in more detail the battery of glial/neuronal markers at our disposal and find the presence of an additional PHD-like cell at a low and background-dependent frequency (13-24%). This data is now presented in Figure 3—figure supplement 1. We find these extra cells indistinguishable from PHD both in terms of morphology and gene expression. To determine the origin of these cells we performed Edu staining experiments, to look for dividing cells, in a lin-48 background. We only detect Edu staining in lin-48 labelled cells in the tail if, and only if, two cells per side are observed. We observe this at a frequency of 21% (Figure 3). In addition, we observed the division of PHso1 “live” in one of our time-lapses of single animals (Figure 3—figure supplement 2). All together this indicates that while PHso1 mainly transdifferentiates directly into PHD it can also, at low frequency, divide symmetrically to generate two neurons, which we now refer to as PHD1 and PHD2.

c) A number of genes have previously been shown to regulate the Y-to-PDA transdifferentiation (Jarriault and others). Determining whether these genes or a subset also play a role in the PHso1 transdifferentiation process would help to establish whether the processes are shared or distinct and begin to outline the underlying mechanisms.

We have now analysed loss-of-function mutants in sem-4 and egl-27, as well as both RNAi and a mosaically-rescued null allele of sox-2. These genes are all core members of the Y-to-PDA “plasticity cassette” which has been shown by the Jarriault lab to be required for the initiation of Y-to-PDA transdifferentiation, specifically in the erasure of Y identity. We find that these factors are largely dispensable for PHso1-to-PHD transdifferentiation. This data is presented in Figure 5, a new figure. Being somewhat surprised (but rather excited) by these results we extended our analysis to AMso-to-AMso’+MCM transdifferentiation and again find these factors are dispensable. All together our results demonstrate that AMso-to-AMso’+MCM indirect transdifferentiation and PHso1-to-PHD direct transdifferentiation are likely governed by distinct mechanisms. This suggests that the “plasticity cassette” are not the only plasticity genes in C. elegans and we are looking forward to discover the molecular mechanisms of glia-to-neuron cell fate switches in our future work.

2) Clarify and strengthen the behavioral analysis. The correlation of the PHD neuron with the 'Molina maneuver' is a particular strength of this work.

a) You nicely propose that the PHDs couple arched tail posture to reverse locomotion. However, one would also expect a similar requirement at the beginning of contact-response behavior, where linking an arched tail posture to reverse locomotion is also important. Please determine whether response behavior is less smooth, even if it ultimately occurs just as frequently as in wildtype.

We have re-analysed the videos from the ablation experiments scoring “hesitation” as a more sensitive measurement of the quality of the Response step. As described in Materials and methods, we have termed “hesitation” a switch in direction between forward and backward locomotion from the time the male establishes contact with the mate to the first turn (or to location of vulva if this occurs without the need of a turn). Although the proportion of PHD-ablated males that hesitate during response is slightly higher than mock males, the difference is not statistically significant. One possible reason for a less pronounced discontinuity in movement during response compared to during scanning may be that we are measuring locomotion during a shorter period of time during response and so the full defects cannot reveal themselves. We have added this data to Figure 7B

b) In Video 4, it seems that, after moving away from the vulva, the male repeatedly attempts intromission "ectopically", near the tail. It's hard to tell from the video whether spicule prodding is happening at this time, but if it is, could it be that PHD is not required for the Molina manoeuvre itself, but rather to prevent inappropriate spicule prodding at locations away from the vulva? Please look for pausing/prodding at locations other than the vulva during scanning behavior. If this is observed, it could be that the PHDs instead inhibit Lov (location-of-vulva) behavior.

We have re-analysed the videos of the ablations experiments and scored ectopic prodding during Molina manoeuvres (Figure 7—figure supplement 2A and B) and during scans (Figure 7—figure supplement 2C and D). We only observed a small but statistically significant increase in ectopic prodding during Molina manoeuvres in PHD-ablated males compared to mocks (Figure 7—figure supplement 2A). However, ectopic prodding occurred exclusively during discontinuous manoeuvres and the proportion of discontinuous Molina manoeuvres with ectopic prodding was similar in mock and PHD-ablated males (Figure 7—figure supplement 2B). Therefore, we conclude that the main defect resulting from the ablation of PHD is discontinuity in locomotion during manoeuvres and that ectopic prodding may be a secondary consequence of pausing and interrupting locomotion. Because PHD-ablated males perform many more discontinuous manoeuvres than mocks, ectopic prodding also increases.

We have also changed the graph showing the percentage of discontinuous scans in Figure 7C and added categories so that the plot matches the one for Molina manoeuvres in Figure 7F.

c) Videos 3, 4, S3: it would be helpful to indicate in the video itself exactly when the "Molina manoeuvre" takes place and, in Video 4, when the defect occurs. Also, it should be noted somewhere that these videos are sped up 2x from real time.

We have added labels in those videos indicating when Molina manoeuvres and discontinuity defects occur. We have also indicated that the videos are sped up 2x in the video legend. We have also renamed the videos so that they are all main and not supplemental videos. Current Videos 1 to 7 were respectively, video 1, 2, S1, S2, 3, S3 and 4

d) While the ablation data are nice, there are some issues regarding how you determined that the cells were ablated, and the absence of an orthogonal method of ablation for confirmation. Reviewers suggested a couple of possibilities for experiments: you have a PHD-specific oig-8 promoter version. You could try to feminize driving TRA-2(IC) under this promoter and checking behavior although it is possible that this promoter comes on too late to feminize. Alternatively, you could silence/ablate PHD using this promoter to confirm the behavioral phenotype.

We respectfully disagree with the reviewer’s view that there were issues regarding how we determine the PHD cells are ablated. We believe laser ablations are the cleanest and most specific way to remove neurons in an otherwise intact animal and we have extensive expertise in larval ablations. Any other type of manipulation that relies on a genetic driver has the caveat of lack of specificity without a single-cell-specific driver. The most specific promoter at our disposal, the oig-8 promoter, is not uniquely expressed in PHD but also in two classes of sensory neurons in the head. To further demonstrate the specificity of our laser ablations and the method by which we ensure PHD cell death, we have included a new ablation control experiment in which we specifically ablate PHDR, leaving PHDL intact (the cells are ~5µm apart). These control experiments are presented in Figure 7—figure supplement 1.

We have additionally pursued a genetic approach as requested by the reviewers. We initially tried driving a reconstituted ced-3 caspase in the PHD neurons but we either failed to get transgenic lines at mid-range concentrations (15 ng/µl) or failed to efficiently kill all PHD neurons in a significant proportion of animals at low concentrations (5 ng/µl). We then generated an inducible silencing transgene driving an histamine-gated chloride channel in PHD neurons, oig-8::HisCl1::sl2::rfp. We could confirm expression of this transgene in PHD neurons by red fluorescence. This approach recapitulated the discontinuity in Molina manoeuvres that we had previously observed in the laser ablation experiments. We have included this data in Figure 7F. However, we would like to note that we did not observe a statistically significant difference in the quality of scans (discontinuous versus discontinuous) in PHD-silenced males compared to controls (data not shown). We don’t know the reasons for a lack of a scan phenotype in the silencing experiments but having extra oig-8 expressing neurons silenced in the head may compensate for the absence of PHD activity. Alternatively, removing the neuron completely through ablation may impact the circuit more severely than just silencing it.

Key resources table

Reagent type (species) or resource	Designation	Source or reference	Identifiers	Additional information
Genetic reagent (C. elegans)	C. elegans: Strain N2	Caenorhabditis Genetics Center	WormBase: N2
Genetic reagent (C. elegans)	AW827	Dr Alison Woollard laboratory		him-5(e1490), heIs63 [wrt-2p::GFP::PH + wrt-2p::GFP::H2B + lin-48p::mCherry] V
Genetic reagent (C. elegans)	CHL56	This study		drpIs3 [lin-48p::tdTomato] I; him-5(e1490) V
Genetic reagent (C. elegans)	BAR37	This study		nsIs198 [mir-228p::GFP + lin-15(+)]; otIs356 [rab-3p::2xNLS::TagRFP], him-5(e1490) V
Genetic reagent (C. elegans)	CHL32	This study		dpy-5(e907) I; otIs356 [rab-3p::2xNLS::TagRFP], him-5 (e1490) V; drpIs1 [grl-2p::GFP, dpy-5(+)] X
Genetic reagent (C. elegans)	OH13083	Pereira et al. Elife. 2015; Caenorhabditis Genetics Center		him-5(e1490) V; otIs576 [unc-17fos::GFP + lin-44::YFP]
Genetic reagent (C. elegans)	CHL36	This study		drpIs3 [lin-48p::tdTomato] I; inIs179 [ida-1p::GFP] II; him-8(e1489) IV
genetic reagent (C. elegans)	CHL57	This study		drpIs3 [lin-48::tdTomato] I; mnIs17 [osm-6p::GFP + unc-36(+)]; him-5(e1490) V
Genetic reagent (C. elegans)	CHL58	This study		drpIs3 [lin-48::tdTomato] I; otIs291 [rab-3p::2xNLS::YFP + rol-6(su1006)], him-5(e1490) V
Genetic reagent (C. elegans)	CHL59	This study		drpIs3 [lin-48p::tdTomato] I; dpy-17(e164) III; otIs291 [rab-3p::2xNLS::YFP + rol-6(su1006)], him-5(e1490) V
Genetic reagent (C. elegans)	CHL60	This study		drpIs3 [lin-48p::tdTomato] I; ced-4(n1162), dpy-17(e164) III; otIs291 [rab-3p::2xNLS::YFP + rol-6(su1006)], him-5(e1490) V
Genetic reagent (C. elegans)	CHL61	This study		drpIs3 [lin-48p::tdTomato] I; him-5(e1490) V; drpIs1 [grl-2p::GFP, dpy-5(+)] X
Genetic reagent (C. elegans)	CHL63	This study		drpIs3 [lin-48p::tdTomato] I; him-5(e1490) V; vsIs48 [unc-17p::GFP] X
Genetic reagent (C. elegans)	CHL64	This study		drpIs3 [lin-48p::tdTomato] I; inIs179[ida-1p::GFP] II; him-5(e1490) V
Genetic reagent (C. elegans)	CHL67	This study		drpIs3 [lin-48p::tdTomato] I; drpIs4 [oig-8p::GFP + pha-1(+)]; him-5(e1490) V
Genetic reagent (C. elegans)	CHL65	This study		drpIs4 [oig-8p::GFP + pha-1(+)]; him-5(e1490) V
Genetic reagent (C. elegans)	BAR77	This study		oleEx24 [grl-2(1 kb)::fem-3::SL2::mCherry (8 ng)+ elt-2p::GFP (40 ng)]; inIs179 [ida-1p::GFP] II; him-8(e1489) IV
Genetic reagent (C. elegans)	CHL62	This study		oleEx18 [grl-2(1 kb)::fem-3::SL2:: mCherry(20 ng) + elt-2p::GFP (40 ng)]; drpIs3 [lin-48p::tdTomato] I; otIs291 [rab-3p::2xNLS::YFP + rol-6(su1006)], him-5(e1490) V
Genetic reagent (C. elegans)	CHL68	This study		sem-4(n19719), drpIs3 [lin-48p::tdTomato] I; otIs291 [rab-3p::2xNLS::YFP + rol-6(su1006)], him-5(e1490) V
Genetic reagent (C. elegans)	CHL69	This study		drpIs3 [lin-48p::tdTomato] I; egl-27(ok1670) II; otIs291 [rab-3p::2xNLS::YFP + rol-6(su1006)], him-5(e1490) V
Genetic reagent (C. elegans)	CHL70	This study		drpIs3 [lin-48p::tdTomato] I; otIs291 [rab-3p::2xNLS::YFP + rol-6(su1006)], him-5(e1490) V; nre-1(hd20), lin-15B(hd126) X
Genetic reagent (C. elegans)	CHL71	This study		drpIs3 [lin-48p::tdTomato] I; inIs179 [ida-1p::GFP] II; him-8(e1489) IV Generated during sox-2 mutant crosses. Distinct background to CHL36.
Genetic reagent (C. elegans)	CHL72	This study		drpIs3 [lin-48p::tdTomato] I; inIs179 [ida-1p::GFP] II; him-8(e1489) IV; otEx4454 [sox-2(fosmid)::mCherry + elt-2p::DsRed]
Genetic reagent (C. elegans)	CHL73	This study		drpIs3 [lin-48p::tdTomato] I; inIs179 [ida-1p::GFP] II; him-8(e1489) IV; sox-2(ot640[unc-119(+)]) X; otEx4454 [sox-2(fosmid)::mCherry + elt-2p::DsRed]
Genetic reagent (C. elegans)	CHL74	This study		drpIs3 [lin-48p::tdTomato] I; inIs179 [ida-1p::GFP] II; him-8(e1489) IV; nre-1(hd20), lin-15B(hd126) X
Genetic reagent (C. elegans)	EM1370	Dr Scott Emmons laboratory		bxEx271 [oig-8p::GFP + oig-8p::mCherry::rab-3]; him-5(e1490) V
Genetic reagent (C. elegans)	EM1371	Dr Scott Emmons laboratory		bxEx272 [oig-8p::GFP + oig-8p::mCherry::rab-3]; him-5(e1490) V
Genetic reagent (C. elegans)	BAR90	This study		him-5(e1490) V; oleEx27 [oig-8p::GCaMP6f::SL2::mRFP(30 ng/µl) + cc::GFP(30 ng/µl)]
Genetic reagent (C. elegans)	BAR115	This study		unc-51(e369), him-5(e1490) V; oleEx27 [oig-8p::GCaMP6f:: SL2::mRFP(30 ng/µl) + cc::GFP(30 ng/µl)]
Genetic reagent (C. elegans)	BAR115	This study		unc-51(e369), him-5(e1490) V; oleEx27 [oig-8p::GCaMP6f:: SL2::mRFP(30 ng/µl) + cc::GFP(30 ng/µl)]; oleEx38 [myo-3::HisCL1::SL2::Cherry(20 ng/µl)]
Genetic reagent (C. elegans)	BAR95	This study		unc-13 (e51) I; him-5 (e1490) V; oleEx27 [oig-8p::GCaMP6f:: SL2::mRFP(30 ng/µl) + cc::GFP(30 ng/µl)]
Genetic reagent (C. elegans)	BAR106	This study		unc-31(e169) IV; him-5(e1490) V; oleEx27 [oig-8p::GCaMP6f:: SL2::mRFP(30 ng/µl) + cc::GFP(30 ng/µl)]
Genetic reagent (C. elegans)	EM1251	Dr Scott W. Emmons laboratory		bxEx201[oig-8p::GFP + pha-1(+)]; him-5(e1490) V - line#1
Genetic reagent (C. elegans)	EM1253	Dr Scott W. Emmons laboratory		bxEx201[oig-8p::GFP + pha-1(+)]; him-5(e1490) V - line#3
Genetic reagent (C. elegans)	BAR94	This study		him-5(e1490) V; lite-1(ce314) X; oleEx30[oig-8p::GCaMP6f::SL2::mRFP(50 ng/µl)]
Genetic reagent (C. elegans)	CB369	Caenorhabditis Genetics Center		unc-51(e369) V
Genetic reagent (C. elegans)	BAR160	This study		him-5(e1490) V; oleEx53 [pAB6(oig-8p::HisCl1::SL2::RFP) (30 ng/uL)+unc-122p::GFP(25 ng/uL)]
Strain, strain background (E. coli)	Strain OP50	Caenorhabditis Genetics Center	OP50
Strain, strain background (E. coli)	Strain MG1693 (Thy-)	E. coli stock centre	MG1693	Strain for EdU staining experiments
Strain, strain background (E. coli)	Strain: HT115 (DE3)	Caenorhabditis Genetics Center	HT115
Tecombinant DNA reagent	Plasmid: pPD95.75	Addgene	#1494	See: DNA constructs and transgenic strains Used to create oig-8prom::GFP and for PCR fusions
Recombinant DNA reagent	Plasmid: oig-8prom::gfp	This study		See: DNA constructs and transgenic strains 964 bp upstream of the oig-8 ATG
Recombinant DNA reagent	Plasmid: pkd- 2prom::mCherry::rab-3	Dr María I. Lázaro-Peña		See: DNA constructs and transgenic strains Replaced promoter by oig-8prom
Recombinant DNA reagent	Plasmid: oig- 8prom::mCherry::rab-3	This study		See: DNA constructs and transgenic strains Created from pkd-2prom::mCherry::rab-3
Recombinant DNA reagent	Plasmid pLR306	Dr Luis Rene García Laboratory		See: DNA constructs and transgenic strains Gateway plasmid used to amplify GCaMP6f::sl2::rfp
Recombinant DNA reagent	Plasmid: oig- 8prom::GCaMP6f::SL2::RFP	This study		See: DNA constructs and transgenic strains Created by PCR fusion (see primers)
Recombinant DNA reagent	Plasmid: pNP471	Dr Cori Bargmann Laboratory		See: DNA constructs and transgenic strains Used to amplify HisCl1::sl2::mCherry
Recombinant DNA reagent	Plasmid: oig- 8prom::HisCl1::sl2::rfp	This study		See: DNA constructs and transgenic strains Created by Gibson
Recombinant DNA reagent	Plasmid: myo- 3prom::HisCl1::sl2::mCherry	This study		See: DNA constructs and transgenic strains Created by PCR fusion (see primers)
Recombinant DNA reagent	Plasmid: lin- 48prom::tdTomato	Dr Mike Boxem Laboratory		See: DNA constructs and transgenic strains 6.8 kb upstream of lin-48 ATG
Recombinant DNA reagent	Plasmid pPD129.36 (L4440)	Dr Andrew Fire Laboratory, Addgene	#1654	Control plasmid for RNAi experiments
Recombinant DNA reagent	Plasmid: sox-2 RNAi clone	Source BioScience		Silence endogenous sox-2 gene
Sequence-based reagent	primer_oig-8 fusion_F	This study	PCR fusion primers	See: DNA constructs and transgenic strains gggagtgacctatgcaaacc
Sequence-based reagent	primer_oig-8 fusion_R	This study	PCR fusion primers	See: DNA constructs and transgenic strains CGACGTGATGAGTCGACCAT tgttttacctgaaatctttt
Sequence-based reagent	primer_myo-3 fusion_F	This study	PCR fusion primers	See: DNA constructs and transgenic strains cgtgccatagttttacattcc
Sequence-based reagent	primer_myo-3 fusion_R	This study	PCR fusion primers	See: DNA constructs and transgenic strains gctagttgggctttgcatGCttctagatggatctagtggtc
Sequence-based reagent	primer_ced-4 (n1162)_F	This study	PCR genotyping primers	See: C. elegans strains gatgctctgcgaaatcgaatgc
Sequence-based reagent	primer_ced-4 (n1162)_R	This study	PCR genotyping primers	See: C. elegans strains tcacctaaatcacacatctcgtcg
Sequence-based reagent	primer_dpy-17 (e164)_F	This study	PCR genotyping primers	See: C. elegans strains aggaggaagcccaatcaacc
Sequence-based reagent	primer_dpy-17 (e164)_R	This study	PCR genotyping primers	See: C. elegans strains cagttggtccttcttctccagc
Sequence-based reagent	primer_sem-4 (n1971)_F	This study	PCR genotyping primers	See: C. elegans strains aaggtgatgcgatgatgtctcc
Sequence-based reagent	primer_sem-4 (n1971)_R	This study	PCR genotyping primers	See: C. elegans strains taatgatcggcttgggtgtgg
Sequence-based reagent	primer_egl-27 (ok1679)_F ext	This study	PCR genotyping primers	See: C. elegans strains tcatcgtttccagtctcttcagg
Sequence-based reagent	primer_egl-27 (ok1679)_R ext	This study	PCR genotyping primers	See: C. elegans strains cgctggttatcaaatgacgcc
Sequence-based reagent	primer_egl-27 (ok1679)_F int	This study	PCR genotyping primers	See: C. elegans strains agacaccagaagctacgaaacc
Sequence-based reagent	primer_egl-27 (ok1679)_R int	This study	PCR genotyping primers	See: C. elegans strains gtttgcatcacggtcttcacg
Sequence-based reagent	primer_nre-1(hd20) lin-15B(hd126)_F	This study	PCR genotyping primers	See: C. elegans strains catgagagctgcgctgaggc
Sequence-based reagent	primer_nre-1(hd20) lin-15B(hd126)_R	This study	PCR genotyping primers	See: C. elegans strains ggttcgggctcgcggtagtc
Chemical compound, drug	Paraformaldehyde	Thermo Fisher Scientific	#28908	Used at 4%
Chemical compound, drug	β-mercapto-ethanol	Sigma Aldrich	M6250	Used at 5%
Chemical compound, drug	Collagenase type IV	Sigma Aldrich	C-5138	Used at 1 mg/ml
Chemical compound, drug	Vectashield antifade	Vector Laboratories	H-1900
Chemical compound, drug	Isopropyl-β-D- thiogalactopyranoside (IPTG)	Generon	GEN-S-02122	Used at 0.6 mM in plates for RNAi knock-down
Chemical compound, drug	Histamine	Sigma	H7125	Used at 20 mM in plates for behavioural assays
Chemical compound, drug	Wormglu	GluStitch, Delta, Canada
Antibody	Anti-RFP pAb (Rabbit polyclonal antibody)	MBL International	PM005	(1:500)
Antibody	Donkey anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 555 (Donkey polyclonal antibody)	Molecular probes	A-31572	(1:200)
Commercial assay or kit	Click-IT EdU Alexa Fluor 594	Invitrogen	C10339
Software, algorithm	ImageJ	ImageJ (http://imagej.nih.gov/ij/)	RRID:SCR_003070
Software, algorithm	GraphPad Prism	GraphPad Prism (graphpad.com)	RRID:SCR_002798
Software, algorithm	Affinity Designer	https://affinity.serif.com/ en-us/designer/	RRID:SCR_016952