Synaptic Protein Degradation Controls Sexually Dimorphic Circuits through Regulation of DCC/UNC-40

Summary

Sexually dimorphic circuits underlie behavioral differences between the sexes, yet the molecular mechanisms involved in their formation are poorly understood. We show here that sexually dimorphic connectivity patterns arise in C. elegans through local ubiquitin-mediated protein degradation in selected synapses of one sex but not the other. Specifically, synaptic degradation occurs via binding of the evolutionary conserved E3 ligase SEL-10/FBW7 to a phosphodegron binding site of the netrin receptor UNC-40/DCC (Deleted in Colorectal Cancer), resulting in degradation of UNC-40. In animals carrying an undegradable unc-40 gain-of-function allele, synapses were retained in both sexes, compromising the activity of the circuit without affecting neurite guidance. Thus, by decoupling the synaptic and guidance functions of the netrin pathway, we reveal a critical role for dimorphic protein degradation in controlling neuronal connectivity and activity. Additionally, the interaction between SEL-10 and UNC-40 is necessary not only for sex-specific synapse pruning, but also for other synaptic functions. These findings provide insight into the mechanisms that generate sex-specific differences in neuronal connectivity, activity, and function.

Graphical Abstract

Highlights

• •Sex-specific synapse pruning during development is regulated by the ubiquitin pathway
• •The E3 ligase SEL-10 targets the UNC-40 netrin receptor via binding to a CPD motif
• •UNC-40 degradation leads to synapse removal only in hermaphrodites, not males
• •CPD mutations disrupt synaptic functions of UNC-40, leaving axon guidance intact

Salzberg et al. show that local ubiquitin-mediated protein degradation in specific synapses of one sex generates sexually dimorphic circuits in C. elegans. Degradation of the netrin receptor UNC-40 in hermaphrodite synapses is necessary to establish the proper neuronal activity pattern.

Introduction

In sexually reproducing species, males and females respond differently to environmental cues and transform the input into sexually dimorphic behaviors. Most of the nervous system is common to both sexes, and yet it can be wired differently in males and females [1]. In C. elegans, the full connectome is available for both sexes [2, 3, 4], and many instances of sexually dimorphic synaptic connections between shared neurons have been observed and validated [4, 5, 6]. The emerging concept is that early connections formed during juvenile stages in both sexes can be selectively removed upon sexual maturation in one sex to create a dimorphic wiring pattern in the adult. A case in point involves three sex-shared neurons: the sensory phasmid neuron PHB connects to the head interneurons AVA and AVG in larvae of both sexes, but due to active synaptic pruning, adult hermaphrodites retain the PHB-AVA synapses while males retain the PHB-AVG synapses (Figures 1A and S1A). These dimorphic connections can be readily visualized with transsynaptic labeling techniques such as “GRASP” (GFP reconstitution across synaptic partners), a complementary split-GFP technology that reconstitutes a functional fluorescent GFP reporter between specific pre- and postsynaptic neuronal pairs (Figure 1B) [7], or “iBLINC” (biotin labeling of intercellular contacts), which involves a trans-synaptic enzymatic transfer of biotin and subsequent detection by a fluorescently labeled streptavidin [8]. These findings suggest that dimorphic neuronal circuits emerge from sex-specific synapse pruning events, but it remains mostly unknown how these are regulated in a sex-specific manner.

Figure 1.

Sex-Specific Pruning of Synapses Depends on the Ubiquitin-Proteasome System

(A) Schematic of key neurons and synapses in this study and the pruning they undergo in the two sexes after sexual maturation. Magnification (dashed circle) is of a region in the preanal ganglion where synaptic connections are formed. Pruned synapses are depicted by dashed ovals. See also Figure S1.

(B) Schematic of the GRASP technique (GFP reconstitution across synaptic partners) used to visualize and quantify synaptic contacts in vivo.

(C) Proteasome inhibition using Bortezomib abolishes specific dimorphic pruning events. Number of synaptic GFP puncta were quantified for the designated synapses and sexes. PHB>AVA and PHB>AVG are transsynaptically labeled using GRASP; PHA>AVG synapses are labeled using iBLINC [5, 8].

(D) Fluorescent micrographs of GRASP GFP signal of the PHB>AVA hermaphrodite-specific connection in preanal ganglion region (outlined in the inset) in control and proteasome-inhibited males. Neuronal processes are labeled with cytoplasmic mCherry marker. Scale bar is 10 μm.

(E) uba-1 is required in PHB to mediate the pruning of the male-specific PHB>AVG connection in adult hermaphrodites. uba-1(it129) temperature-sensitive animals were shifted from 15°C to 25°C at the indicated larval stages.

(F) Fluorescent micrographs of GRASP signal of the PHB>AVG male-specific connection in uba-1 mutant hermaphrodites. Gut; gut granules. Scale bar is 10 μm.

In (C), we performed the nonparametric Mann-Whitney test (Wilcoxon rank sum test). In (E), we performed Kruskal-Wallis test followed by Dunn’s multiple comparison test. ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05; NS, non-significant in all panels. In all images, anterior is left and dorsal is up. Blue and red color-coding is used in all figures to distinguish male from hermaphrodite, respectively.

One prominent regulatory mechanism for synapse remodeling is the ubiquitin-proteasome system (UPS). The exquisite specificity of the UPS emerges as a critical determinant of the physiological changes synapses undergo during development and plasticity, as well as during the pathological changes that occur in many neurological diseases [9, 10, 11, 12]. Ubiquitination specificity is determined mainly by the large variety of E3 ligases, which allows for a broad spectrum of substrate selectivity [13]. Many studies have revealed the essential roles of E3 ligases in establishing neuronal function and connectivity [14]. However, the high number of E3 ligases and their potential substrates has complicated the assignment of individual E3 ligases to the regulation of distinct stages of neuronal development. How might the UPS function to shape neuronal circuits in the two sexes from a common template?

Using PHB-to-AVG synapses as a model to elucidate sex-specific synaptic pruning, we show that the spatial organization of the ubiquitin pathway orchestrates dimorphic neuronal connectivity. We find that the evolutionary conserved E3 ligase SEL-10/FBW7 degrades UNC-40 only in PHB in hermaphrodites, leading to pruning of synapses. SEL-10 binds a conserved phosphodegron site of UNC-40 upon phosphorylation by GSK-3b and prolyl-isomerization by PINN-1. In the absence of SEL-10, UNC-40 protein levels are stabilized, and synapses between PHB and AVG are maintained. We further show that while SEL-10-dependent degradation of UNC-40 affects synapse maintenance, it is dispensable for the earlier guidance roles of the netrin pathway. Our results demonstrate that UPS pathways have the potential to reconfigure the nervous system via synapse elimination in a spatial and temporal manner and to accommodate the need for sex-specific circuit structure and function.

Results

Dimorphic Synapse Pruning Occurs through Protein Degradation

To gain a mechanistic understanding of sex-specific synapse pruning, we investigated the potential role of several well-known cellular degradative pathways in synaptic pruning during sexual maturation. We found that the apoptotic cell-death pathway and autophagy pathways are dispensable for the pruning of PHB>AVA synapses in males and PHB>AVG synapses in hermaphrodites (Figures S1B and S1C). We next turned to the UPS. In recent years, the UPS has emerged as a critical player in neuronal development with a conserved, central role in establishing and modulating neural circuits [9, 13, 15]. To test whether an active proteasome is necessary for sex-specific synapse pruning, we grew worms on plates containing the proteasome inhibitor Bortezomib [16] and assayed the number of synaptic puncta between specific pairs of neurons. We found that dimorphic synapse pruning failed to occur for either male-specific connections (PHB>AVG, pruning of hermaphrodite synapses) or hermaphrodite-specific connections (PHB>AVA and PHA>AVG, pruning of male synapses) (Figures 1C and 1D). Thus, proteasome-dependent synapse degradation emerges as a central theme for generating dimorphic connectivity in C. elegans.

Spatio-temporal Control of Ubiquitin-Dependent Synapse Pruning

The E1 ubiquitin-activating enzyme, encoded by the uba-1 gene in C. elegans [17], lies at the heart of the UPS biochemical reaction and is necessary for all subsequent steps [12]. To test whether uba-1 controls synapse pruning, we used a temperature-sensitive mutant of uba-1, it129 [17]. We grew worms at the permissive temperature until the L1 stage and then transferred them to the restrictive temperature until adulthood. We found that the pruning of synapses does not occur in uba-1 mutants and that the adult neuronal state is non-dimorphic (Figure 1E), in accordance with the results for worms grown on Bortezomib. Shifting uba-1 animals to the restrictive temperature at a later developmental stage did not halt synapse pruning (Figure 1E), suggesting that synapse pruning occurs in a limited time window during development or that there is some residual gene product of uba-1 resulting from the later temperature shift [17]. Interestingly, the control of synapse number and maintenance appears to be independent of ubiquitin-mediated protein degradation, as the uba-1 temperature shift did not affect synapses not destined to be pruned (PHB>AVG synapses in males) (Figure 1E). Synaptic pruning of PHB>AVG in hermaphrodites was rescued by cell-specific expression of uba-1 in the presynaptic cell PHB, but not in the postsynaptic partner, indicating that uba-1 is active in presynaptic pathways (Figures 1E and 1F). Taken together, our results demonstrate that sex-specific synapse pruning is dependent upon ubiquitin-mediated proteolysis, which functions in a highly specific spatial and temporal manner to sculpt sex-specific neuronal circuits from a common template.

The E3 Ligase SEL-10/FBW7 Regulates Sex-Specific Synapse Pruning

Multi-ubiquitination is carried out by an enzymatic cascade that includes the activity of ubiquitin-activating E1 enzymes, ubiquitin-conjugating E2 enzymes, and the highly diverse E3 ubiquitin ligases [13, 14]. Out of the estimated ∼500 E3 ligases in C. elegans [18], we decided to focus on the conserved E3 ligase sel-10/Fbw7. SEL-10 is an F-box and WD-repeat-containing protein that has been shown to regulate selective synapse pruning in C. elegans [19]. We found that in sel-10 deletion mutants, PHB>AVG synapses were not eliminated in hermaphrodites (Figure 2A). In accordance with the presynaptic requirement of the UPS (Figure 1E), rescue of sel-10 from the presynapse, but not the postsynapse, restored pruning in hermaphrodites (Figure 2A). We further observed local enrichment of SEL-10 puncta along the overlap region between the PHB neuronal processes and the AVG process of both sexes (Figure 2B), and some of these SEL-10 puncta co-localize with synaptic sites (Figure S2). The connectivity region between PHB and AVG in males is shorter, and accordingly, males exhibit consistently fewer SEL-10 puncta (Figure 2B).

Figure 2.

The F-Box E3 Ligase SEL-10 Mediates the Pruning of the PHB>AVG Synapse in Hermaphrodites

(A) Quantification of PHB>AVG GRASP synaptic puncta in sel-10 mutants and in mutants carrying a rescuing transgenic sel-10 cDNA expressed under pre- or postsynaptic promoters.

(B) Fluorescent micrographs of the contact region between the PHB and AVG processes in L4 animals. Animals express a fosmid carrying an N-terminally tagged SEL-10 (mCherry::sel-10^fosmid, an array driving GLR-1::GFP specifically in AVG and are DiD-stained to label the sensory phasmid neuron PHB. Arrowheads points to SEL-10 puncta found in contact area between AVG and PHB neurites. In the bottom panel, hermaphrodites carry an array forcing expression of UNC-6 in AVG (normally expressed only in adult males [20]). Quantification of SEL-10 puncta is shown. Scale bars represent 5 μm. See also Figure S2.

(C) Quantification of PHB>AVG GRASP synaptic puncta in skr-1 mutant background.

(D) Scoring of PHA>AVG GRASP synaptic puncta in sel-10 mutant background.

In (A–D), we performed Kruskal-Wallis test followed by Dunn’s multiple comparison test. ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01; NS, non-significant. See also Figure S3.

SEL-10 is a component of a complex termed SCF (for SKP, Cullin and F-box protein) [21]. skr-1, the ortholog of vertebrate skp1 [22], was shown to regulate, together with sel-10, synapse pruning of the hermaphrodite-specific HSN neuron [19]. We tested deletion mutants of skr-1 and indeed found defects in synaptic pruning, suggesting that skr-1 is the skp component of the SCF complex that degrades PHB-AVG synapses (Figure 2C). We further found that sel-10 is dispensable for the pruning of male-specific PHA>AVG synapses (Figure 2D). Thus, additional unidentified E3 enzymes regulate other sex-specific synapse pruning events.

While the effect of sel-10 on sex-specific neuronal connections may reflect an indirect outcome of its suggested regulatory role in the sex determination pathway [23], we consider this possibility unlikely, because (1) we found no overt masculinization phenotype for sel-10 deletion mutant hermaphrodites, in agreement with previous observations [23]; (2) we observed no decrease in the levels of TRA-1, the master sexual regulator in C. elegans, in PHB or AVG of sel-10 mutant hermaphrodites or of hermaphrodites with inhibited proteasome function (Figure S3A); (3) the fem genes, the most downstream genetic regulators of tra-1 and proposed SEL-10 targets, do not possess the obligatory CPD (Cdc4 phosphodegron, a high-affinity binding site) sequence necessary for SEL-10 binding; (4) wild-type adult hermaphrodites do not express UNC-6 in AVG, but males do [20], while UNC-6 expression in AVG of sel-10 mutant hermaphrodites is as low as in wild-type hermaphrodites (Figure S3B), strengthening the notion that the nervous system in sel-10 hermaphrodites is not masculinized; and (5) sel-10 does not affect all sex-specifically pruned synapses (Figure 2D), as would have been predicted by a direct effect on the sex-determination pathway. We conclude that the sex-related synaptic pruning defects in sel-10 mutants are likely due to a direct synaptic role of SEL-10, in accordance with its synaptic localization (Figures 2B and S2).

sel-10 and the Netrin Pathway Interact to Regulate Dimorphic Connectivity

SEL-10/FBW7 targets substrates for ubiquitin-mediated degradation through the CPD site (Figures 3A and 3B) [21]. To identify SEL-10 substrates, we screened substrates that have been previously inferred computationally [24] and focused on neuronal genes. We found that the UNC-6/netrin receptor UNC-40/DCC (Deleted in Colorectal Cancer) contains a predicted conserved CPD site within its cytoplasmic tail and is thus a potential SEL-10 substrate (Figure 3B). Netrin signaling has been associated with diverse processes in multiple tissues, including cell adhesion and survival, axon guidance, regulation of synaptogenesis, and synaptic function [28, 29, 30]. UNC-6 has been recently shown to be required for male-specific PHB>AVG synapse maintenance [20]. Through sex-specific transcriptional regulation, unc-6 expression is downregulated in AVG in hermaphrodites during sexual maturation, while in males expression persists in AVG. It remains unclear, however, how synapses are eliminated in hermaphrodites in the absence of UNC-6 secretion. To determine how the UPS interacts with the netrin pathway, we performed an epistasis analysis between sel-10, unc-6, and unc-40 (Figures 3C and 3D). In sel-10;unc-40 double mutant animals, the PHB>AVG synapses were not maintained even in males, as in unc-40 single mutants (Figure 3C). In contrast, in sel-10;unc-6 double mutants, male synapses were maintained even in hermaphrodites, similar to sel-10 single mutants (Figure 3D). Moreover, blocking protein degradation stabilizes synapses in males even in the absence of unc-6 (Figure 3E). Together, the epistasis analysis places sel-10 genetically downstream of unc-6 but upstream of unc-40. This suggests a genetic model in which unc-6 negatively regulates the inhibitory interaction between sel-10 and unc-40 (Figure 3F). Mechanistically, we propose that UNC-6 serves to maintain the PHB>AVG synapse in males by preventing SEL-10-mediated degradation of UNC-40. In line with this proposed mechanism, force-expressing UNC-6 in hermaphrodite AVG was sufficient to significantly reduce the amount of SEL-10 protein at the PHB-AVG contact region (Figure 2B).

Figure 3.

sel-10 and the Netrin Pathway Interact to Regulate Dimorphic Connectivity

(A) Schematics of the SCF complex and a recruited phosphorylated CPD-containing substrate bound to SEL-10.

(B) Alignment of the UNC-40 predicted CPD to other CPDs of validated FBW7 substrates [21, 24]. Obligatory residues are highlighted in bold red; phospho-residues are marked with an asterisk.

(C) Epistasis analysis of sel-10 and unc-40. Scoring of PHB>AVG GRASP synaptic puncta in the designated genotypes and sexes.

(D) Epistasis analysis of sel-10 and unc-6. Scoring of PHB>AVG GRASP synaptic puncta in the designated genotypes and sexes.

(E) Blocking protein degradation in unc-6 mutant males by feeding worms with 20 μM Bortezomib is sufficient for synapse maintenance.

(F) Suggested genetic pathway for the regulation of sexually dimorphic pruning by SEL-10 and the netrin pathway.

(G) Quantification of PHB>AVG GRASP synaptic puncta in wild-type, pinn-1 mutant hermaphrodites, animals with 200 μM of the potent PINN-1 inhibitor Sulfopin [28], and unc-40;pinn-1 mutants.

(H) Fluorescent micrographs of animals expressing PINN-1 tagged with GFP (knockin allele, see Key Resources Table and [29]) and DiD-stained to label the PHB neurons (dashed circle).

(I) Quantification of PHB>AVG GRASP synaptic puncta in control or GSK-3 inhibitor-treated hermaphrodites.

(J) Fluorescent micrographs of animals expressing GSK-3::mNeonGreen (knockin allele [30] and DiD-stained to label the PHA/PHB neurons.

Scale bars in (H and J) represent 10 μm. In (C, D, G, and I), we performed Kruskal-Wallis test followed by Dunn’s multiple comparison. ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05; NS, non-significant. See also Figures S3 and S6.

The CPD Kinase gsk-3 and the Prolyl Isomerase pinn-1 Regulate Dimorphic Synapse Pruning

Cdc4 phosphodegrons are subject to tight regulation by kinases and prolyl isomerases that post-translationally modify the conserved Ser/Thr and Pro residues, respectively (Figure 3B) [31]. The Glycogen Synthase kinase Gsk3 and the prolyl isomerase Pin1 have been implicated in CPD regulation in mammals [32, 33, 34, 35]. The pSer/Thr-Pro motif is specifically recognized by the proline isomerase Pin1, which catalyzes the cis/trans isomerization of proline (Figures 3A and 3B). We therefore examined whether the worm orthologs gsk-3 and pinn-1 play a role in dimorphic synaptic pruning. Hermaphrodites with a loss-of-function deletion in pinn-1 fail to prune the PHB>AVG synapses, similar to the phenotype of sel-10 and unc-40 CPD mutants (Figure 3G). In addition, feeding worms a selective small-molecule covalent inhibitor for human Pin1, Sulfopin [28], is sufficient to block synapse degradation (Figure 3G). The same occurs when animals are grown on the gsk-3 inhibitor CHIR-99021 (Figure 3I). To test whether the role of gsk-3 and pinn-1 depends on unc-40, we analyzed double mutants of unc-40;pinn-1 and unc-40 animals grown on plates containing GSK-3 inhibitor. In both cases, unc-40 loss-of-function mutation suppresses the synaptic pruning defects (Figures 3G and 3I). Moreover, using fluorescently labeled translational fusions, we observed expression in the presynaptic neuron PHB for both PINN-1 and GSK-3, placing them in direct cellular context to target UNC-40 (Figures 3H and 3J, respectively). Thus, PINN-1 and GSK-3 regulate sex-specific synapse removal in hermaphrodites, possibly by targeting the UNC-40 CPD, and perhaps involving other substrates.

SEL-10 Degrades UNC-40 through a Conserved Phosphodegron Site

The genetic interactions inferred by the epistasis analysis, together with the predicted degradation site in the UNC-40 cytosolic domain, suggest a role for SEL-10 in regulating UNC-40 stability and the UNC-6/UNC-40 ligand/receptor interaction. To test the in vivo effects of SEL-10 on UNC-40 protein levels, we used a PHB-specific UNC-40::GFP transgene [36]. We found that in hermaphrodites, UNC-40 is expressed during development, but its levels decrease significantly and are undetectable by the young adult stage (Figures 4A, S4A, and S4B). In males, UNC-40 protein levels are significantly higher at the young adult stages compared with hermaphrodites. However, loss of sel-10 results in a significant increase in UNC-40 levels in hermaphrodites and in a moderate increase in males, suggesting that sel-10 regulates UNC-40 protein levels in vivo sex specifically (Figure 4A). At earlier larval stages, UNC-40 levels are non-dimorphic and are unaffected by loss of sel-10 (Figure S4A).

Figure 4.

CPD-Mediated Degradation of UNC-40 by SEL-10

(A) Left: fluorescence micrographs of UNC-40::GFP expression in hermaphrodites of the designated genotypes and developmental time points. Expression was driven by a PHB-specific promoter in a multi-copy array. Scale bars represent 10 μm. Right: quantification of fluorescence intensity in both sexes at L4 and adult developmental stages, binned as high expression (on), low expression (dim), or no expression (off). For statistical comparison, high and dim expression were binned together. See also Figures S4A and S4B.

(B) CRISPR-Cas9-mediated mutagenesis of the UNC-40 CPD. Genomic nucleotide changes are indicated in red, as well as the two residues (Ser and Thr) mutated to Ala. The other two nucleotide changes facilitated genotyping and prevented re-cutting by Cas9.

(C) UNC-40 protein structure including the intracellular position of the CPD.

(D) Quantification of PHB>AVG GRASP synaptic puncta in the designated genotypes and sexes.

(E) Western blot showing the SEL-10 mediated degradation of an UNC-40 intracellular fragment (UNC-40^endo, residues 1108–1415) expressed in HEK293t cells. SEL-10 was tagged with a myc tag and UNC-40^endo with a FLAG tag. See also Figure S4C.

(F) Quantification of PHB>AVG GRASP synaptic puncta in hermaphrodites overexpressing a 12aa-long peptide corresponding to the UNC-40 CPD sequence with 3aa overhangs, specifically in PHB.

In (A), we preformed Fisher’s exact test. In (D), we performed Kruskal-Wallis test followed by Dunn’s multiple comparison test. In (F), we performed Mann-Whitney test. ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05; NS, non-significant.

Is UNC-40 directly degraded by SEL-10 via the CPD? To address this question, we generated a putative undegradable unc-40 gain-of-function allele by mutating the conserved phospho-acceptor Ser/Thr residues in the UNC-40 CPD into Ala residues (henceforth named mCPD, for mutant CPD) using the CRISPR-Cas9 system (Figures 4B and 4C) [37]. In unc-40(mCPD) animals, dimorphic synapses were retained in both sexes (Figure 4D), suggesting that the CPD motif is necessary to mediate sex-specific synapse removal. Additionally, we expressed both proteins in cultured cells. Co-expressing SEL-10 with a cytoplasmic UNC-40 tail fragment resulted in degradation of the UNC-40 fragment (Figure 4E). Addition of the proteasome inhibitor Velcade restored UNC-40 levels (Figure S4C), supporting the conclusion that UNC-40 degradation is ubiquitin dependent. Therefore, SEL-10 promotes UNC-40 proteasomal degradation in vitro. To demonstrate in vivo that the binding of SEL-10 to UNC-40 is via the CPD site, we overexpressed a 12aa-long peptide carrying the UNC-40 CPD sequence specifically in PHB (Figure 4F). We reasoned this peptide might function to inhibit SEL-10 binding to the endogenous UNC-40, thereby stabilizing UNC-40 and thus the PHB>AVG synapse. Indeed, hermaphrodites expressing this PHB-specific CPD peptide failed to prune the PHB>AVG synapses. This result lends further support (albeit indirect) to the importance of the CPD motif in mediating the binding of SEL-10 to UNC-40 binding.

Surprisingly, PHB>AVG synapses were retained in unc-40(mCPD) hermaphrodites even in a unc-6 loss-of-function background (Figure 4D), implying that UNC-6 secretion from the postsynaptic cell AVG is dispensable for UNC-40 activation in PHB, and instead functions to protect UNC-40 from degradation. Strengthening this conclusion, we found that in the absence of unc-6, the UNC-40 signal is completely lost in L4 and adult male PHB neurons (Figure 4A).

The Synaptic and Guidance Functions of UNC-40 Are Separable

UNC-6/UNC-40 interactions mediate both local paracrine synaptic events [38, 39, 40, 41] and long-range cell migration and axon navigation processes [42]. Reflecting its role as a global organizer of nervous system architecture, unc-40 loss-of-function mutant animals display a smaller body size and uncoordinated movement (Figures 5A–5C). We noticed that unc-40(mCPD) animals, while displaying synaptic defects, retain normal body length and exhibit only mildly compromised locomotion (Figures 5A–5C), raising the possibility that the CPD is dispensable for the guidance functions of UNC-40. We therefore explored whether unc-40(mCPD) animals display axon guidance defects. In wild-type worms, the two PHB tail neurons send anterior processes that meet on the ventral side, where they fasciculate with the AVG neurite (Figure 5D, top panel). As expected, unc-40(e271) loss-of-function mutants show guidance errors whereby at least one of the two PHB axons does not reach the ventral nerve cord (Figure 5D, middle panel). In contrast, unc-40(mCPD) animals do not display PHB guidance defects (Figure 5D, bottom panel). Lack of guidance defects in mCPD mutants was confirmed also for the HSN neurite (Figure S5). Together, these observations suggest that the CPD motif of UNC-40 mediates UNC-40’s synaptic functions but is dispensable for its role as a guidance receptor. Thus, the two functions of UNC-40 can be uncoupled.

Figure 5.

Axon Guidance Occurs Normally in UNC-40 CPD Mutants

(A–C) Comparison of worm body length (A), speed (B), and covered track distance (C) between wild-type and unc-40 loss-of-function or CPD mutants, measured using an automated multi-worm tracker.

(D) Left: fluorescence micrographs of the PHB processes in hermaphrodites of the designated genotypes. Expression of mCherry was driven by a PHB-specific promoter in a multi-copy array. Scale bars are 10 μm. Right: percentage of animals displaying PHB guidance errors in the designated genotypes and sexes.

In (A–C), we performed Kruskal-Wallis test followed by Dunn’s multiple comparison test. In (D), we preformed Fisher’s exact test. ^∗∗∗∗p < 0.0001, ^∗∗p < 0.01, ^∗p < 0.05; NS, non-significant. See also Figure S5.

Interaction between sel-10 and unc-40 in Other Synaptic Contexts

We next assessed whether sel-10 and unc-40 interact to shape additional synapses in C. elegans. Colon-Ramos et al. have shown that the synaptic organization between the AIY-RIA interneurons is shaped in an unc-40-dependent manner [41]. In wild-type animals, a rich synaptic area in AIY marks the contact region with RIA and localizes to the bend point of the AIY process (Figure S6A). We analyzed the localization and intensity of AIY presynaptic vesicles and found that loss of unc-40 results in altered synaptic distribution, mostly in zone 3, indicating that synapses are more evenly spread along the AIY process in the absence of unc-40 (Figures S6B and S6C). The same analysis in sel-10 loss-of-function mutants revealed a stronger synaptic signal in zones 2 and 3 compared with wild type (Figures S6B and S6C), suggesting that sel-10 is required for proper synapse distribution and that it functions to limit the number of presynaptic vesicles in both zones. Moreover, unc-40 mCPD mutants show a similar trend to that of sel-10 mutants (Figures S6A–S6C). Hence, we propose that the interaction between sel-10 and unc-40 may be a broad mechanism for the regulation of synaptic pruning.

Another context in which sel-10 and unc-40 interaction might be inferred is the pruning of synapses in the secondary synaptic region of the HSN motor neuron (Figures S6D and S6E). The developmental pruning of these synapses depends on the SKR-1 and SEL-10 complex [19]. To explore whether UNC-40 interacts with SEL-10 to promote HSN synaptic pruning, we perturbed UNC-40 activity by performing RNAi at the L1 stage (see STAR Methods) and then analyzed the animals at the adult stage. By applying RNAi after the L1 stage, we bypassed the early guidance roles unc-40 plays in HSN migration [43]. We did not find any synaptic defects in unc-40-RNAi-treated animals, nor in unc-40(mCPD) animals (Figures S6F–S6H). Thus, the SEL-10 substrates required for HSN synapse pruning remain unidentified.

Dimorphic Activity Pattern of AVG Is UNC-40 Dependent

Does UNC-40 CPD-mediated synaptic degradation have any functional effect in vivo on dimorphic neuronal activity? To answer this question, we recorded calcium responses from AVG to an aversive stimulus applied to either hermaphrodite or male worm tails in a microfluidic chamber (Figure 6A). In wild-type animals, AVG exhibits a dramatic dimorphic response to 2 M glycerol: male AVGs respond to the stimulus with a single peak (continuous rise in activity throughout the 20 s of stimulus delivery), while hermaphrodite AVGs show a biphasic on and off response, with an early “on” peak (∼10 s after stimulus delivery) followed by an “off” peak upon stimulus removal (Figures 6B and 6C). Importantly, hermaphrodite AVG response in CPD mutants is significantly attenuated (Figures 6C and 6D, green), implying that the ectopic PHB>AVG synapse (and potentially additional male-specific synapses to AVG) in these animals compromised the AVG response. This demonstrates that the degradation of the presynaptic UNC-40 receptor in hermaphrodites is necessary to establish the proper activity pattern of the postsynaptic cell AVG. We also recorded the responses of AVG in sel-10 mutant hermaphrodites and found that sel-10 loss of function attenuates the hermaphrodite response to a similar, yet not identical, extent as the unc-40(mCPD) (Figure S7). Thus, circuit activity is directly affected by dimorphic synaptic connections.

Figure 6.

UNC-40 CPD Is Required to Evoke the Full Response of AVG to an Aversive Stimulus

(A) A worm is placed in a microfluidic chip [44], positioned to enable stimulus delivery to the tail while imaging AVG in the head. Black arrow indicates flow direction of the stimulus, 2 M glycerol, dyed in rhodamine B. Inset represents a magnified region of the head showing a GCaMP6s signal in AVG (scale bar is 10 μm). Chip scale bar is 100 μm.

(B) Individual worms’ AVG calcium traces of wild-type males, hermaphrodites, and unc-40(ety1) CPD mutant hermaphrodites in response to 2 M glycerol. Normalized GCaMP6s signal ΔF/F0 is color coded (color bar shows color range according to normalized fluorescence intensities). Stimulus was delivered at 20 to 40 s; imaging duration was 2 min.

(C) Average GCaMP6s signals of AVG neurons of wild-type males (blue), hermaphrodites (red), and UNC-40 CPD mutant hermaphrodites (green) in response to the stimulus. The colored lines and shaded area surrounding them indicate the mean values of ΔF/F0 and SEM, respectively.

(D) Peak On and Off responses of wild-type and unc-40 CPD mutant hermaphrodites. The dots represent the maximum intensity of single AVG neurons at 20 to 40 s (On response) and 40 to 60 s (Off response). We performed Mann-Whitney test. ^∗p < 0.05; NS, non-significant. See also Figure S7.

Discussion

Here, we found how the UPS is recruited sex specifically to create a dimorphic neuronal circuit. There is growing evidence that E3 ubiquitin ligases localize to distinct subcellular compartments and regulate the proper development of neuronal circuits, also in a dimorphic context [12, 14, 45]. In Drosophila, proteasome-dependent sex-specific cleavage of the transcription factor Lola results in dimorphic neurite branching and patterning of a courtship circuit [45]. We show that precise synaptic connectivity is regulated by UPS-dependent activity of SEL-10, which actively removes synapses by degrading the netrin receptor DCC/UNC-40. SEL-10 localizes to the dimorphic synaptic region in the pre-anal ganglion, and cell-specific rescue experiments suggest that it functions from the presynaptic compartment to cell-intrinsically control sex-specific synapse pruning. Our data in Figure 4D imply that UNC-40 can transmit the synaptic maintenance signal in male PHB even in the absence of the AVG-released UNC-6 cue, as long as the UPS degradation signal is inhibited. Therefore, in this synaptic context, UNC-40 can act without UNC-6 bound to it, in line with previous reports [46, 47, 48, 49]. We suggest that UNC-6 prevents the access of SEL-10 to UNC-40 in an unknown manner, which may involve the protective modification of UNC-40, the inhibitory modification of SEL-10, or both (Figure 7). Whether UNC-40 is extracted from the plasma membrane by intermembrane cleavage and proteolysis or whether it’s taken up by vesicles and degraded by the lysosome remains to be elucidated [9, 50].

Figure 7.

Suggested Model for SEL-10-Dependent Sex-Specific Synapse Degradation

In adult hermaphrodites (left), no UNC-6/netrin is secreted from AVG, thus leading to the ubiquitylation and degradation of UNC-40/DCC in the pre-synaptic cell PHB upon the combined action of SKR-1, PINN-1, GSK-3, and SEL-10. In males, however, the secretion of netrin from AVG protects UNC-40/DCC from SEL-10 binding in an unknown manner, resulting in the stabilization and maintenance of UNC-40/DCC and, consequently, of the synapse with PHB (right).

The early guidance roles of the netrin pathway in brain patterning [51, 52, 53, 54] complicate our ability to fully assess its role in synaptogenesis during later development. While in C. elegans, UNC-6 is known to regulate both long-range guidance of neurons and short-range synaptic events, whether the two functions are mediated by the same molecular determinants is unknown. The CRISPR strain we generated, which targets the CPD of UNC-40 specifically, exhibits synaptic defects while retaining normal axon guidance. As such, the mutant CPD allele provides an elegant tool to bypass the guidance roles of the netrin pathway and focus on the synaptic functions of UNC-6 and UNC-40 signaling per se. Previous work in C. elegans has shown that in some cases, synaptic components such as CLEC-38 and the E3 ligase RPM-1 can interact with UNC-40 and modulate the netrin pathway to terminate axon outgrowth [55, 56], suggesting crosstalk between synaptic and axon navigation pathways. Since we do not observe PHB guidance defects in CPD mutant animals, the possibility of autoregulation of netrin pathway-dependent axon guidance of PHB by synaptic UNC-40 seems unlikely. Netrin-DCC interactions have been shown to strengthen synaptic contacts and promote synaptogenesis in multiple models [20, 38, 39, 41]. Moreover, stabilized UNC-40 can even cluster post-synaptic receptors as a mechanism to maintain synaptic connections [40]. Our work is the first demonstration of local regulation on DCC by an E3 ligase that controls only its synaptogenic activities. It is likely that the vertebrate netrin receptor DCC is regulated by the SEL-10 homolog FBW7, since both human and mouse DCC contain a CPD site (Figure 3B). Additionally, DCC is known to be regulated by the UPS in neurons by other F-box proteins such as TRIM9 [57]. Plooster et al. propose that netrin inhibits the ubiquitination of DCC and modulates axon branching, similar to our synaptic model. Thus, similar regulatory mechanisms may operate in vertebrate synaptic pruning.

We further show a unique example of an identifiable neuron, AVG, with an activity pattern that differs dramatically between the two sexes in response to the same stimulus (Figure 6). Interestingly, the male-specific connection between PHB and AVG is ectopically stabilized in unc-40(mCPD) hermaphrodites but does not lead to full masculinization of the AVG response in hermaphrodites; it only diminishes the amplitude of the bi-phasic hermaphrodite activity pattern. This implies that additional CPD-independent inputs must be integrated by male AVG, which is missing in CPD-mutated hermaphrodites. Indeed, according to the published connectome, AVG is predicted to make several other male-specific connections [3, 4]. Thus, UNC-40-dependent synapses do not account entirely for the dimorphic difference between male and hermaphrodite AVG response. Interestingly, AVG activity in sel-10 mutant hermaphrodites is even closer to the male pattern (having lost the double ON/OFF peaks) (Figure S7), suggesting additional, UNC-40-independent roles for SEL-10. This may not be surprising, given that SEL-10 has additional synaptic targets [24].

In a recent meta-analysis of genome-wide association studies (GWASs) for eight different psychiatric disorders such as autism spectrum disorder, bipolar disorder, major depression, and schizophrenia, the only genetic variant to be associated with all diseases was found in the DCC locus [58]. DCC also appears to be differentially expressed in autism spectrum disorder patients’ post-mortem brains [59]. Together with our work, these studies point to a conserved pleiotropic role for DCC in the proper formation and maintenance of brain circuits. As many neurological disorders exhibit sexual dimorphism in disease mechanisms [60, 61], understanding the genetic mechanisms that modulate dimorphic circuits could prove beneficial in the development of novel gender-specific therapies.

STAR★Methods

Key Resources Table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Monoclonal ANTI-FLAG® M2-Peroxidase (HRP) antibody	Sigma-Aldrich	cat#A8592
Monoclonal Anti-β-Actin−Peroxidase antibody	Sigma-Aldrich	cat#A3854
Anti-Myc tag antibody [9E10] (HRP)	Abcam	cat#ab62928
Bacterial and Virus Strains
E. coli OP50	Caenorhabditis Genetics Center	N/A
Chemicals, Peptides, and Recombinant Proteins
Bortezomib (PS-341)	SelleckChem	cat#S1013
CHIR-99021	MedChemExpress	cat#HY-10182
Sulfopin	Nir London lab’s, Weizmann Institute of Science, Israel	[28]
Experimental Models: Cell Lines
HEK293t Human embryonic kidney cells	ATCC	cat#CRL-3216™
Experimental Models: Organisms/Strains
otIs630 X [srg-13::BirA::nrx-1 25ng/ul, inx-18p::AP::nlg-1 25 ng/ul, unc-122::streptavidin::2xsfGFP 25 ng/ul, inx-18p::wcherry 10ng/ul, pRF4 50 ng/ul]; him-5(e1490)V	[6]	OH14099
uba-1(it129) IV	Caenorhabditis Genetics Center	RV110
uba-1(it129)IV; him-5(e1490)V; otIs614	This study	MOS5
uba-1(it129)IV; him-5(e1490)V; otIs614; etyEx1[gpa-6p::uba-1 cDNA 10ng/ul, myo-2::gfp 50ng/ul, pBS 50ng/ul]	This study	MOS22
uba-1(it129)IV; him-5(e1490)V; otIs614; etyEx3[inx-18p::uba-1 cDNA 5ng/ul, myo-2::gfp 20ng/ul, pBS 80ng/ul]	This study	MOS25
otIs612 [MVC12 (flp-18::NLG-1::GFP11) 15ng/ul, MVC6 (gpa-6::NLG-1::GFP1-10) 15ng/ul, MVC11 (flp-18::mcherry) 10ng/ul, MVC15 (nlp-1::mcherry) 10ng/ul, pRF4 50ng/ul]; him-5(e1490)V	[5]	OH13575
sel-10(ok1362) V	Courtesy of Iva Greenwald	GS6156
him-8(e1489) IV	Caenorhabditis Genetics Center	CB1489
him-5(e1490) V	Caenorhabditis Genetics Center	CB4088
sel-10(ok1632)V; him-8(e1489) IV	This study	MOS35
sel-10(ok1632)V; him-8(e1489 IV; otIs612	This study	MOS45
sel-10(ok1632)V; him-8(e1489)IV; otIs614	This study	MOS155
sel-10(ok1632)V; him-8(e1489)IV; otIs614; etyEx12[inx-18::sel-10 5ng/ul, myo-2::gfp 10ng/ul, pBS 80ng/ul]	This study	MOS106
sel-10(ok1632)V; him-8(e1489)IV; otIs614; etyEx13 [gpa-6::sel-10 5ng/ul, myo-2::GFP 10ng/ul, pBS 80ng/ul]	This study	MOS109
sel-10(ok1632)V; him-8(e1489)IV; otIs630	This study	MOS54
pha-1(e2123); arEx1776[fosmid mCherry-sel-10]	[24]	GS6914
pha-1(e2123); him-5(e1490); arEx1776; otIs582[inx-18p::GLR-1::GFP, ttx-3::mcherry]	This study	MOS72
him-5(e1490); otIs614; otEx6914[inx-18p::UNC-6::SL2::NLS::YFP::H2B 50 ng/uL, ttx-3::GFP 30 ng/uL, pBS 50 ng/uL]	[20]	OH14845
pha-1(e2123); him-5(e1490); arEx1776; otIs582; otEx6914	This study	MOS87
otIs614[inx-18p::nlg-1::gfp11 30ng/ul, MVC6 (gpa-6::nlg-1::gfp1-10) 30ng/ul, MVC15 (nlp-1::mcherry) 5ng/ul, inx-18p::wcherry 10ng/ul, pRF4 50ng/ul]; him-5(e1490)V	[5]	OH13577
unc-6(ev400) X, him-5(e1490); otIs614	[20]	OH14844
sel-10(ok1632), unc-6(ev400), him-8(e1489); otIs614	This study	MOS160
unc-40(e271), him-8(e1489); otEx6913[inx-18p::NLG-1::GFP11 30ng/ul, MVC6 gpa-6::NLG-1:::GFP1-10 30ng/ul, MVC15 nlp-1::mcherry 5ng/ul, inx-18::wcherry 10ng/ul, pRF4 50ng/ul]	This study	MOS49
sel-10(ok1632), unc-40(e271), him-8(e1489); otEx6913	This study	MOS47
pinn-1(tm2235) II, glt-3(bz34) IV; nuIs5[glr-1::GFP + glr-1::G(alpha)s(Q227L) V + lin-15(+) V]	Caenorhabditis Genetics Center	IMN30
pinn-1(tm2235) II, him-5(e1490); otIs614	This study	MOS114
pinn-1(ju1504[pinn-1::gfp::loxp::3xflag ll])	[29]	CZ25245
gsk-3(cp251[gsk-3::mnG-C1ˆ3xFlag])	Caenorhabditis Genetics Center	LP538
unc-40(ety1[T1268A; S1272A]) I	This study	MOS76
unc-40(ety1); otEx6913	This study	MOS91
unc-40(ety1), unc-6(ev400), him-8(e1489); otEx6913	This study	MOS103
wyEx2871[MVC114a (gpa-6p::unc-40::GFP) 35 ng/ μl, unc-122p::RFP 20 ng/μl]	[36]	TV7675
him-8(e1489) IV; wyEx2871	This study	MOS107
sel-10(ok1632)V; him-8(e1489) IV; wyEx2871	This study	MOS108
him-5(e1490)V; etyEx31	This study	MOS118
unc-40(ety1) I; him-5(e1490)V; etyEx31	This study	MOS129
ced-3(n717)IV; him-5(e1490)V; otIs614	This study	MOS8
gsnl-1(ok2979)V; him-5(e1490)V; otIs612	This study	MOS9
epg-8(bp251); him-5(e1490)V; otIs612	This study	MOS2
tra-1(ez72[biotag::GFP::TEV::3xflag::tra-1]) III	Caenorhabditis Genetics Center	DZ840
tra-1(ez72) III; him-5(e1490) V; otIs460[inx-18p::wcherry; pha-1(+)]	This study	MOS27
sel-10(ok1632) V; tra-1(ez72) III; otIs460	This study	MOS29
him-8(e1489) IV; otIs638[unc-6fosmid::SL2::NLS::YFP::H2B 10 ng/uL, ttx-3::mCherry 3 ng/uL, OP50 genomic 100 ng/uL)]	[20]	OH14852
sel-10(ok1632) V, him-8(e1489) IV; otIs638	This study	MOS36
kyIs262 IV[unc-86::myr GFP + odr-1::RFP]	[43]	CX5974
unc-40(e271)I; kyIs262 IV	This study	MOS146
unc-40(ety1) I; kyIs262 IV	This study	MOS94
wyIs45 [ttx-3p::GFP::rab-3 + unc-122p::RFP] X	[41]	TV392
unc-40(e271) I; wyIs45 X	This study	MOS104
sel-10(ok1632) V; wyIs45 X	This study	MOS121
unc-40(ety1) I; wyIs45 X	This study	MOS130
syg-1(ky652) X; kyIs235 [unc-86::snb-1::YFP + unc-4p::lin-10::RFP(intron) + odr-1::RFP ]V	[62]	CX652
kyIs235 V	This study	MOS77
sel-10(ok1632) kyIs235 V	This study	MOS78
unc-40(ety1) I; kyIs235 V	This study	MOS90
unc-40(e271) I, pinn-1(tm2235) II; him-8(e1489) IV; otEx6913	This study	MOS244
skr-1(ok1696) I	Caenorhabditis Genetics Center	VC1241
skr-1(ok1696) I; him-8 (e1489) IV; otEx6913	This study	MOS16
unc-6(ev400); wyEx2871; him-5(e1490) V	This study	MOS240
arEx1776 [fosmid mCherry-sel-10]; iyEx85 [30ng/ul MVC142A/gpa-6::syd-2::YFP, 20ng/ul::RFP/unc-122:RFP]	This study	MOS241
sel-10(ok1632) V; him-8(E1489) IV; etyEx31[inx-18b::GCaMP6s PCR fusion, sra-6::wrmscarlet, MVC15 (PHB::mCherry)]	This study	MOS233
otIs614; him-5(e1490) V; etyEx60[gpa-6::unc-40 CPD peptide(12aa) 20ng/ul; myo-2::GFP 30ng/ul; pBS 50 ng/ul]	This study	MOS227
Oligonucleotides
See Table S1		N/A
Recombinant DNA
inx-18p::uba-1 cDNA	This study	pMO59
gpa-6p::sel-10 cDNA	This study	pMO61
inx-18p::sel-10 cDNA	This study	pMO62
gpa-6p::CPDUNC-40	This study	N/A
pcDNA3.1-FLAG-UNC-40 dNT (endodomain)	This study	pMO64
pcDNA3.1-myc-sel-10-GFP	[19]	N/A
gpa-6p::uba-1 cDNA	This study	pMO58
Software and Algorithms
ImageJ, Fiji		https://imagej.net/Fiji
Zen	Zeiss	N/A
Wormlab	MBF Biosciences	https://www.mbfbioscience.com/wormlab
Prism 8	GraphPad	https://www.graphpad.com/
Illustrator/Photoshop	Adobe	N/A

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Meital Oren-Suissa (meital.oren@weizmann.ac.il).

Materials Availability

Unique strains generated in this study have been deposited at the Caenorhabditis Genetics Center. Requests for other strains and plasmids should be directed to the Lead Contact.

Data and Code Availability

The code for processing of calcium imaging data was generated using MATLAB and is available upon request.

Experimental Model and Subject Details

Wild-type strains were C. elegans variety Bristol, strain N2. Worms were maintained according to standard methods. Worms were grown at 20°C on nematode growth media (NGM) plates seeded with bacteria (E. coli OP50) as a food source, with the exception of temperature sensitive alleles, which were maintained at 15°C. Sex and age of animals used for each experiment are indicated in the corresponding figures and legends. All transgenic strains used in this study are listed in the Key Resources table.

Method Details

Molecular cloning

To generate the uba-1 rescue constructs pMO58 (gpa-6p::uba-1 cDNA), pMO59 (inx-18p::uba-1 cDNA) and pMO60 (flp-18p::uba-1 cDNA), UBA-1 cDNA was amplified from the C. elegans ORFeome library (Source BioScience) and cloned, using RF cloning [63], into MVC6 (gpa-6::nlg-1::gfp1-10, swapping the nlg-1::gfp1-10 fragment), pMO10 (inx-18 2^nd intron::mCherry, swapping the mCherry fragment) and MVC12 (flp-18p::nlg-1::gfp11, swapping the nlg-1::gfp-11 fragment), respectively.

To generate the sel-10 rescue constructs pMO61 (gpa-6p::sel-10 cDNA) and pMO62 (inx-18p::sel-10 cDNA), sel-10 was amplified from an N2 mixed-stage cDNA library and cloned into MVC6 (gpa-6::nlg-1::gfp1-10, swapping the nlg-1::gfp1-10 fragment) and pMO19 (inx-18 2^nd intron::BirA::nrx-1, swapping the BirA::nrx-1 fragment), respectively, using Gibson assembly [64].

pcDNA3.1-GFP-sel-10-myc was a kind gift from Mei Ding [19]. pAC5-unc-40 constructs were a kind gift from Claire Benard [65]. The unc-40 endodomain fragment was amplified from a pCB310 template (carrying FLAG-tagged unc-40 endodomain [65]) and inserted into a pcDNA3.1 backbone using Gibson assembly to create pMO64.

To insert the UNC-40 CPD peptide under the PHB promoter, two complementary oligonucleotides encompassing the UNC-40 CPD with 3aa overhangs (MLRGTPPNSSAA, CPD in bold, see Table S1) were used in a site-directed insertion protocol.

AVG calcium imaging: inx-18p and GCaMP6s-SL2-tagRFP-unc-54 3′UTR fragments were amplified and linked by PCR fusion [66].

CRISPR/Cas9-mediated genome editing

In order to generate the unc-40(ety1) CPD mutated strain, we employed a previously described CRISPR/Cas9 genome engineering protocol [37]. Briefly, tracrRNA and two crRNAs, targeting the dpy-10 and the unc-40 loci, were mixed with a recombinant cas9 (IDT), ssODN repair template to introduce a dominant mutation into the dpy-10 locus [37], and ssODN targeting unc-40 (see Table S1 for sequences).

Rol/dpy F1 progeny were singled and screened by PCR using the primers CTACGTGGAGCACCCCCGAATG and TCTCTCCATTCGATGAACCA with an annealing temperature of 65.5°C, conditions in which only the ety1[mCPD] allele is amplified but not the wild-type unc-40. Plates with a positive PCR signal were Sanger-sequenced to validate homozygosity of the ety1[mCPD] allele.

Pharmacological inhibition of protein degradation

L3 transgenic worms (OH13575, OH13577 and OH14099) were transferred to NGM agar plates containing either 20 μM Bortezomib/Velcade (Selleckchem) or 0.04% DMSO and incubated for 24 h before imaging [16].

To test whether synapse degradation depends on pinn-1 and gsk-3, a selective small-molecule covalent inhibitor for human Pin1 (Sulfopin [28]), or a highly selective inhibitor of glycogen synthase kinase 3 (GSK-3), CHIR-99021 (Axon Medchem) were used. L3 transgenic worms (OH13577, see Table S1) were transferred to plates seeded with OP50 containing 200 μM Sulfopin, CHIR-99021 or DMSO and incubated for 24 h before imaging. In all pharmacological-treatment experiments we observed a slight effect on movement but no lethality or growth defects were observed.

Cell culture and Western

HEK293T cells were seeded in a 6-well plate. Each well was transfected with a total of 2 μg of plasmids (Lipofectamine 2000, Invitrogen). 48 h later proteins were extracted, separated on SDS-PAGE and analyzed by western blotting using antibodies against myc tag (for GFP-SEL-10-myc, Abcam ab62928), FLAG tag (UNC-40-FLAG, Sigma A8592) or actin (Sigma A3854).

RNAi

RNA interference was performed using the feeding method [67]. pos-1 and unc-22 RNAi served as positive controls for assay efficiency. To generate late effect RNAi of unc-40 (for testing HSN SSR synapse elimination), in order to bypass early guidance defects, L1 hermaphrodites were fed HT115 bacteria carrying dsRNA for the relevant genes or a control empty vector and scored at the adult stage.

Microscopy

Worms were anaesthetized, using 100 mM sodium azide (NaN₃), and mounted on 5% agar on glass slides. Worms were analyzed by Nomarski optics and fluorescence microscopy, using a Zeiss 880 confocal laser-scanning microscope. The estimated resolution of the confocal microscope used for imaging GFP was ∼150 nm. Multidimensional data were reconstructed as maximum intensity projections using Zeiss Zen software. GRASP GFP puncta were quantified by scanning the original full Z-stack for distinct dots in the area where the processes of the two neurons overlap. In animals with guidance defects (unc-6 and unc-40 mutants), only animals where guidance of PHBs and AVG was scored as normal GRASP puncta were quantified (only ∼45% of animals show guidance defects of PHB/AVG).

Figures were prepared using Adobe Illustrator v24.0.

Measuring guidance defects and synaptic puncta distribution in the HSN neuron

To observe HSN axon guidance, animals carrying a myristoylated GFP tag in their HSN (unc-86::myr::GFP) [43] were synchronized, and young adults (identified by vulva appearance) were scored for ventral HSN neurites. To measure the number of synaptic puncta in HSN, we used transgenic animals carrying unc-86::snb-1::YFP [62]. Synaptic puncta were quantified in the primary synaptic region (PSR) and secondary synaptic region (SSR) [19] (see Figure S6).

Quantification of GFP::RAB-3 synaptic distribution in AIY neurons

Levels of fluorescence intensity in the AIY neuron were measured and compared between different genetic backgrounds. Confocal Z-stacks were converted to maximum intensity projections using Zen software (Zeiss). In each image, a polygon ROI was selected for zones 1-3 of the AIY axon (after Colon-Ramos et al., 2007 [41]). Then, image-thresholding was used and the mean intensity in each zone was quantified (see Figure S6).

Automated worm tracking

Wild-type, unc-40(e271) and unc-40(ety1) hermaphrodites were analyzed for locomotion-related phenotypes. 1-day-old animals were placed on an NGM plate seeded with diluted 30 μL of OP50 bacteria. To keep the worms in the camera’s field of view, a plastic ring with a 1.5 cm diameter was placed in each plate and five worms were placed inside it. After a 10 min habituation period, a 2-min recording session was performed at ∼22°C (room temperature) with the WormLab automated tracking system (MBF Bio-science) [68]. Videos were segmented to extract the worm contour and skeleton for phenotypic analysis. 12 parameters were measured for each worm. Raw WormLab data was exported to Prism (GraphPad) for further statistical analysis.

Calcium imaging

Calcium imaging on 1-day-old adult worms was performed in microfluidic devices designed according to Chronis et al. [44]. Chip fabrication was carried out at the Nanofabrication Unit of the Weizmann Institute. The tunnel height was 28 μm and the width at the worm’s nose space was 24 μm. Operation of the microfluidic chip was performed using two-channel syringe pumps (Fusion 200, Chemyx), which controlled the flow of buffer and stimulus into the microfluidic chip. The solutions were pushed through a PVC tube and a stainless-steel connector into the tunnels of the chip, and a manual switch determined the arrival of the stimulus to the worm. Tubes were replaced between experiments and connectors cleaned with ethanol. The pumping rate during experiments was 0.005-0.05 ml/min. A 2 min habituation was given prior to imaging. To prevent movement, 10 mM Levamisole was added to all solutions (except the S basal solution used to load the worm). To visualize proper delivery of the stimulus to the worm, 50 μM rhodamine B was added only to the stimulus. If the worm moved or the flow was incorrect, the file was discarded and a second trial was performed with the same worm. No more than two trials were done with the same worm. Imaging was done with a Zeiss LSM 880 confocal microscope using a C-apochromat 40x/n1.2 water objective. The imaging rate was 6.667 Hz, the imaging duration was 2 min, and the stimulus duration was 20 s. The stimulus was given 20 s after the recording started. For analysis, the GCaMP6s fluorescence intensity was measured using FIJI. All files were exported as tiff files, ROIs (regions of interest) of the somas were drawn manually to best represent the signal, and their mean gray values were exported. The ROI size was kept constant for all neurons in each experiment. Downstream data processing was performed using MATLAB, the baseline fluorescent level (F0) was calculated by averaging the mean gray values of 100 frames (15 s) before stimulus delivery. ΔF was calculated by subtracting F0 from the value at each time point/frame/image, and the result was divided by F0 (ΔF/F0), to normalize differences in fluorescence baseline levels.

Quantification and Statistical Analysis

Significance was computed using the GraphPad Prism software (version 8). Bar graphs are a box-and-whiskers type of graph, min to max showing all points. The vertical bars represent the median and “+” represents the mean. Hermaphrodite data is shown in red, male in cyan. Statistical test parameters and outcomes are indicated in figure legends. Reporting on number of animals used in each experiment is found in Table S2.

Published: August 27, 2020