Presynaptic neurons self-tune by inversely coupling neurotransmitter release with the abundance of CaV2 voltage-gated Ca²⁺ channels

Significance

Faced with a myriad of internal and external challenges, neurons display remarkable adaptability, adjusting dynamically to maintain synaptic stability and ensure the fidelity of neural circuit function. This homeostatic adaptability is essential not only for neural circuit integrity but also has implications for various psychiatric and neurological pathologies. At the presynaptic terminal, the influx of calcium through specific ion channels is necessary for the exocytosis of neurotransmitter-filled synaptic vesicles. Our studies in the nematode Caenorhabditis elegans reveal a sophisticated mechanism where the abundance of presynaptic calcium channels is negatively regulated by the efficiency of synaptic vesicle exocytosis. This self-regulating mechanism ensures that presynaptic neurotransmitter release is autonomously adjusted, thereby maintaining synaptic function and safeguarding the robustness of neural communication.

Abstract

The abundance of CaV2 voltage-gated calcium channels is linked to presynaptic homeostatic plasticity (PHP), a process that recalibrates synaptic strength to maintain the stability of neural circuits. However, the molecular and cellular mechanisms governing PHP and CaV2 channels are not completely understood. Here, we uncover a previously not described form of PHP in Caenorhabditis elegans, revealing an inverse regulatory relationship between the efficiency of neurotransmitter release and the abundance of UNC-2/CaV2 channels. Gain-of-function unc-2SL(S240L) mutants, which carry a mutation analogous to the one causing familial hemiplegic migraine type 1 in humans, showed markedly reduced channel abundance despite increased channel functionality. Reducing synaptic release in these unc-2SL(S240L) mutants restored channel levels to those observed in wild-type animals. Conversely, loss-of-function unc-2DA(D726A) mutants, which harbor the D726A mutation in the channel pore, exhibited a marked increase in channel abundance. Enhancing synaptic release in unc-2DA mutants reversed this increase in channel levels. Importantly, this homeostatic regulation of UNC-2 channel levels is accompanied by the structural remodeling of the active zone (AZ); specifically, unc-2DA mutants, which exhibit increased channel abundance, showed parallel increases in select AZ proteins. Finally, our forward genetic screen revealed that WWP-1, a HECT family E3 ubiquitin ligase, is a key homeostatic mediator that removes UNC-2 from synapses. These findings highlight a self-tuning PHP regulating UNC-2/CaV2 channel abundance along with AZ reorganization, ensuring synaptic strength and stability.

Neurons exhibit extraordinary adaptability, effectively responding to a wide spectrum of external and internal stimuli and challenges. This ability is pivotal for sustaining the delicate balance and functionality of neural circuits (1, 2). Central to this adaptability is homeostatic synaptic plasticity, a process that recalibrates neuronal excitability and synaptic strength, thereby ensuring the overall stability of neural networks.

Homeostatic plasticity varies based on factors such as neuron type and maturation stage (3–5), manifesting in either presynaptic or postsynaptic neurons. For example, in the mammalian central nervous system, a prolonged decrease in the neuronal firing rate leads to an increase in postsynaptic neurotransmitter receptors—a phenomenon known as synaptic scaling. This process adjusts synaptic strength in a multiplicative manner, proportional to the original strength of the synapse (6). Conversely, presynaptic homeostatic plasticity (PHP) is observed across the nervous systems of different species, especially when postsynaptic receptors are impaired. These presynaptic adjustments affect neurotransmitter release by altering calcium influx, the readily releasable pool of synaptic vesicles (SVs), or the release probability (7). The molecular mechanism of PHP has been extensively studied in the Drosophila neuromuscular junction (NMJ), where disruptions in postsynaptic muscle glutamate receptors trigger a compensatory increase in presynaptic release via transsynaptic retrograde signals (8).

Calcium influx at synapses is a critical determinant of neurotransmitter release and synaptic strength. This calcium influx facilitates the fusion of neurotransmitter-filled SVs to the presynaptic plasma membrane (9–11). The primary conduits of Ca²⁺ ions at synapses are CaV2 voltage-gated calcium channels. The regulation of CaV2 channel abundance at synapses is closely linked to homeostatic adjustments in synaptic strength. These channels localize to presynaptic active zones (AZs) through multivalent, redundant interactions with a wide array of AZ proteins (12–14). These findings underscore the correlation between the abundance of CaV2 channels and the size of the AZ. In the context of the Drosophila NMJ, an increase in homeostatic CaV2 channels is observed together with an increase in certain AZ proteins (15). Likewise, in vertebrates, the overexpression of CaV2 channels causes an expansion of the AZ area (16). These findings collectively suggest a complex interplay between CaV2 channel abundance and the structural dynamics of the presynaptic AZ.

To further explore the relationship between CaV2 channel abundance and neurotransmitter release, we leveraged the nematode Caenorhabditis elegans, renowned for its highly conserved neurotransmitter release machinery and CaV2/UNC-2 channels. In this study, we found that the abundance of presynaptic UNC-2 channels, which are homologous to CaV2 channels in C. elegans, is inversely regulated by the level of SV exocytosis. This regulatory mechanism allows presynaptic neurotransmitter release to maintain an established level independently of postsynaptic inputs. Importantly, our data showed that several key AZ proteins are necessary for the regulation of UNC-2 channel abundance. Additionally, our forward genetic screening identified the wwp-1 gene, a homolog of the WWP1 HECT E3 ubiquitin ligase, as a crucial regulator of UNC-2 channel abundance. Mutations in wwp-1 result in diminished ubiquitination of UNC-2 channels, leading to an increase in their presynaptic levels. This finding adds a unique dimension to our understanding of the molecular mechanisms governing PHP and neurotransmitter release.

Results

The Activity of the UNC-2 Channel Adversely Influences Its Abundance at Synapses.

To determine the copy number of UNC-2/CaV2 channels at the synapse, we compared the puncta intensities of endogenously tagged GFP::UNC-2 channels with those of Saccharomyces cerevisiae GFP-tagged Cse4 proteins. Cse4, a yeast homolog of the centromeric histone H3 variant CENP-A, associates with centromeres in pairs, resulting in a total of 32 molecules in haploid cells (Fig. 1A) (17). C-terminally GFP-tagged Cse4 proteins give rise to additional copies due to slow degradation (Fig. 1A) (18, 19). The average maximum and integrated puncta intensities of GFP::UNC-2 puncta in the dorsal nerve cord (DNC) did not differ from those of CSE4c::GFP(diploid) puncta (Fig. 1 B–E), despite the high variability in GFP::UNC-2 puncta intensity. The intensities in heterozygous cim104[GFP::unc-2] animals were ~50% lower than those in Cse4c::GFP(diploid) (Fig. 1 D and E). Consequently, we estimate that the average copy number of UNC-2 is approximately 100 per synapse, with considerable variation across individual synapses. This estimate is consistent with the number of UNC-2 channels estimated with superresolution microscopy (20).

Fig. 1.

Image analysis of UNC-2/CaV2 channels at the presynaptic terminal revealed their copy number and distinct AZ localization, separate from endocytic zones. (A) Representative images of the yeast centromeric protein Cse4 tagged with GFP. Cse4i::GFP and Cse4c::GFP denote GFP insertions at the middle (Cse4i) and C terminus (Cse4c) of Cse4, respectively. (B) Schematic diagram of the synapse in the DNC, where images were acquired. VNC: ventral nerve cord. (C) Representative 30 μm segments of DNC images of UNC-2/+, UNC-2, and ELKS-1 at synapses. The average puncta intensity of ELKS-1 is twofold greater than that of UNC-2. The images in both A and C were all adjusted to an identical dynamic range for accurate comparison. (D and E) The average maximum puncta intensities (D) and integrated puncta intensities (E) of yeast GFP-tagged Cse4 and C. elegans endogenously GFP-tagged UNC-2/+, UNC-2, and ELKS-1. (F) A single-plane image of endogenously tagged UNC-11::mScarlet and GFP::UNC-2 from the DA and DB synapses shows UNC-2 localization distinct from that of the endocytic marker UNC-11. (Scale bar, 5 μm.)

To investigate the spatial relationship between UNC-2 channels and synaptic architecture, particularly whether they localize to AZs distinct from endocytic zones, we concurrently imaged UNC-2 channels and UNC-11, a homolog of AP180/CALM involved in endocytosis. Given the densely packed synapses in the DNC, which could obscure distinct zones, we employed the FRT-FLP recombination system (21) to selectively label endogenous UNC-11 and UNC-2 with mScarlet and GFP, respectively, in a subset of DNC neurons. UNC-11 did not colocalize with UNC-2 and tended to encircle UNC-2 puncta (Fig. 1F). A similar observation was made with filamentous actin (F-actin) localized to the endocytic zone in mouse hippocampal neurons (22). These findings reinforce the notion that UNC-2 channels are specifically localized to AZs and are distinct from the regions associated with endocytosis.

Building on the reliability and accuracy of our UNC-2 channel quantitative analysis, we aimed to explore the effect of UNC-2 channel activity on its synaptic abundance. Previously, we and others showed that the S240L substitution mutation in the cytosolic link between transmembrane segment four and five of repeat domain I, whose orthologous mutation in humans is a cause of familial hemiplegic migraine (FHM), results in an increase in channel conductance and synaptic function (Fig. 2A) (13, 23). The puncta intensities of UNC-2SL in cim112[GFP::unc-2SL] animals were considerably lower than those of wild-type (WT) UNC-2 channels (Fig. 2 H–J). This pattern was mirrored in an FHM knock-in mouse model, ref. 24 lends further support to our findings that an inverse relationship exists between the functional activity of channels and their abundance at synapses.

Fig. 2.

The activity of UNC-2/CaV2 channels determines channel abundance independently of postsynaptic acetylcholine receptors. (A) A schematic drawing of the UNC-2 structure. The asterisks denote the positions of the CRISPR/Cas9-mediated amino acid substitution sites, S240L (purple) and D726A (red), and the insertion sites for superecliptic pHluorin (orange) and GFP (green). (B) Alignment of the pore domains from four repeats (I–IV) in UNC-2 and mammalian CaV2.1α₁. The four conserved selectivity filter glutamic acids (E) are highlighted in cyan. Substitution of aspartic acid (D) to alanine (A) at position 1 of the repeat II domain is denoted in red. (C–E) (C) 40-stack maximum projection images of the head region in WT GFP::UNC-2 (cim104) and GFP::UNC-2DA (cim147) animals. (D) DNC image of a cim147[GFP::unc-2DA];cim125[mScarlet::rimb-1] animal showing colocalization of UNC-2DA and RIMB-1. (Scale bar, 5 μm.) (E) UNC-2DA animals showed a reduction in lateral swimming comparable to that of unc-2(e55) null mutants. (F and G) The evoked responses of unc-2(e55) and unc-2(cim150[unc-2DA]) animals were comparable. Representative traces (F) and amplitude responses (G) of evoked responses in WT, unc-2(e55) null, and unc-2DA animals at 1 mM extracellular Ca²⁺ ions. One-way ANOVA, Tukey’s post hoc analysis. ns, not significant. (H–J) GFP::UNC-2DA exhibited increased puncta intensity, while GFP::UNC-2SL showed decreased puncta intensity. (H) Representative DNC images of GFP::UNC-2, GFP::UNC-2DA, and GFP::UNC-2SL. The average puncta intensities (I) and cumulative frequencies (J) showed the same trend. One-way ANOVA with Tukey’s post hoc test. (K–M) (K) Representative images of GFP::UNC-2 puncta in WT and acr-16(ok789);unc-38(x20) acetylcholine receptor mutant animals. (Scale bar, 5 μm.) The average puncta intensity (L), cumulative frequency distribution (M), and puncta number (L) are shown. (Scale bar, 5 μm.) Unpaired Student’s t test. ns, not significant.

To further investigate the relationship between CaV2 channel function and abundance, we sought to identify a loss-of-function mutation that does not cause a trafficking defect and consequential decrease in presynaptic channel levels. Employing CRISPR/Cas9, we introduced a series of mutations based on human pathogenic variants and prior functional studies. The pore of CaV channels is formed by pore-lining loops (P1 and P2) situated between transmembrane helices 5 and 6 across all four repeat domains (I–IV) (Fig. 2A) and features four conserved glutamates within the selectivity filter (25). Notably, the pore domain of repeat II contains a highly conserved, unique aspartic acid following the selectivity filter glutamate (Fig. 2B) (26). The substitution of this aspartic acid with alanine, cim147[GFP::unc-2DA], did not cause channel accumulation in cell bodies adjacent to the nerve ring (Fig. 2C). Moreover, the GFP::UNC-2DA channels were appropriately localized to the AZs, as evidenced by colocalization with the AZ marker RIMB-1 (Fig. 2D). Despite this normal localization pattern, cim147 animals displayed severe movement defects indistinguishable from those of unc-2(e55) null mutants (Fig. 2E). Moreover, when evoked responses were recorded in postsynaptic muscle cells using patch clamp electrophysiology, cim147 animals showed a reduced evoked response comparable to that of unc-2 null mutants (Fig. 2 F and G). Intriguingly, our quantitative puncta analysis revealed that UNC-2DA channels in the DNC had greater intensity than WT UNC-2 channels without a notable change in the number of puncta (Fig. 2 H–J). This observation, together with our finding regarding UNC-2SL abundance, indicates that UNC-2 channel activity inversely correlates with presynaptic UNC-2 channel abundance.

Transsynaptic PHP can modulate the abundance of presynaptic CaV2 channels (27, 28). Typically, a compromise in postsynaptic receptor function triggers the release of retrograde signals, which in turn act in presynaptic neurons to increase neurotransmitter release in part by increasing CaV2 channel levels (29–31). To determine whether a similar PHP mechanism operates in C. elegans, we assessed UNC-2 channel levels in acr-16;unc-38 double mutants lacking both nicotine- and levamisole-sensitive acetylcholine receptors (Fig. 2 K–M). We did not observe any significant differences in UNC-2 puncta number or intensity between WT and acr-16;unc-38 mutant animals, indicating the absence of the transsynaptic PHP mechanism characteristic of vertebrate and Drosophila NMJs in C. elegans. Furthermore, neither a reduction in SV acetylcholine content nor loss of muscarinic ACh-R function changed UNC-2 levels (SI Appendix, Fig. S1), indicating that presynaptic cholinergic output or signaling does not influence overall UNC-2 channel levels.

Deficits in the Exocytosis of SVs Increase the Abundance of UNC-2 Channels at Synapses.

An increase in UNC-2 channel function enhances neural excitability, Ca²⁺ influx, and SV exocytosis. To determine whether a decrease in SV exocytosis elevates UNC-2SL channel levels, we focused on key SNARE complex regulators, namely unc-18/Munc18 and unc-13/Munc13. While UNC-18 forms a complex with closed, inactive UNC-64/syntaxin to regulate synaptic UNC-64 levels and localization, UNC-13 opens the conformation of UNC-64, facilitating SNARE complex assembly and subsequent exocytosis (32–36). The absence of UNC-18 or UNC-13 blocks SV exocytosis and thus both evoked and spontaneous synaptic release (37–39). Impairing SV exocytosis by introducing the unc-18(md299) or unc-13(s69) null mutation into cim112[GFP::unc-2SL] animals increased UNC-2SL channel levels to match those of WT UNC-2 channels (Fig. 3 A and B). These results strongly suggest that the reduction in UNC-2SL channel abundance stems from a sustained increase in SV exocytosis, which is mediated by elevated UNC-2SL channel activity.

Fig. 3.

A deficit in either the SNARE complex or endocytic components increases the abundance of UNC-2SL channels in the presynaptic AZ. (A and B) Loss of UNC-13 and UNC-18 function increased GFP::UNC-2SL puncta intensity. Representative DNC images (A), average intensities (B), and puncta numbers (B) of GFP::UNC-2SL in unc-18(md299) and unc-13(s69) animals. (Scale bar, 5 μm.) One-way ANOVA with Tukey’s post hoc analysis. (C and D) Loss of UNC-64/syntaxin function increased the puncta intensity of GFP::UNC-2SL. Representative DNC images (C), average intensities (D), and puncta numbers (D) of GFP::UNC-2SL in unc-64(js21) animals. (Scale bar, 5 μm.) Unpaired Student’s t test. ns, not significant. (E and F) Loss of RIC-4/SNAP25 function increased the puncta intensity of GFP::UNC-2SL in the DNC. Representative DNC images (E), average intensities (F), and puncta numbers (F) of GFP::UNC-2SL in ric-4(md1088) animals. (Scale bar, 5 μm.) Unpaired Student’s t test. ns, not significant. (G and H) Loss of UNC-57/endophilin and UNC-26/synaptojanin function increased the puncta intensity of GFP::UNC-2SL in the DNC. Representative DNC images (G), average intensities (H), and puncta numbers (H) of GFP::UNC-2 and GFP::UNC-2SL in WT, unc-57(ok310), and unc-26(e345) animals. One-way ANOVA with Tukey’s post hoc test. (I and J) Representative images (I) of unc-2(cim104);cimIs61 animals before (pre) and after (post) 20 min of exposure to 590 nm light. (Scale bar, 5 μm.) After illumination, the UNC-2 puncta intensity (J) decreased in animals cultured with retinal, without altering the number of puncta (J). One-way ANOVA with Tukey’s post hoc test.

The potential role of SV exocytosis in regulating UNC-2 channel abundance prompted us to investigate the SNARE complex. The SNARE complex is assembled by vesicle-associated v-SNARE (SNB-1/synaptobrevin) and target membrane-associated t-SNAREs (RIC-4/SNAP-25 and UNC-64/syntaxin-1) and is involved in priming and fusion pore formation during SV exocytosis (40–43). We found that hypomorphic unc-64(js21) and ric-4(md1088) mutations increased the levels of UNC-2SL channels (Fig. 3 C–F), although hypomorphic snb-1(js17) did not (SI Appendix, Fig. S2 A and B). Interestingly, we also did not observe an increase in UNC-2SL channels in snt-1(md290) mutant animals, which have a null mutation in a C. elegans homolog of the calcium ion sensor synaptotagmin-1 at synapses (SI Appendix, Fig. S2 C and D). Although we cannot completely rule out the role of SNB-1 and SNT-1 in regulating UNC-2 channel levels due to their potential genetic redundancy, these results indicate that SV exocytosis inversely correlates with UNC-2SL channel levels.

UNC-57/endophilin A and UNC-26/synaptojanin are needed for clathrin-mediated endocytosis and SV retrieval from the plasma membrane and endosomes (44–46). In unc-57 and unc-26 mutants, the total number of SVs is reduced, and the frequency of SV exocytosis is decreased by more than 80% (47). This is presumably due to inefficient retrieval and repacking of SVs from the fused vesicle membrane. To test whether changes in the efficiency of SV endocytosis affect UNC-2SL channel levels, we introduced unc-57(ok310) or unc-26(e345) null mutations into cim112 animals. Mutations in unc-57 and unc-26 led to a considerable increase in UNC-2SL channel levels, indicating that reduced SV turnover can elevate UNC-2SL channel levels (Fig. 3 G and H). Together, our data indicate that elevated SV exocytosis, as opposed to channel activity, in cim112 animals results in lower levels of UNC-2SL than WT UNC-2. This phenomenon is not unique to UNC-2SL, as we found that a reduction in SV exocytosis increases the abundance of WT UNC-2 channels. Compared with those of WT animals, the GFP::UNC-2 puncta intensities of unc-64(js21), ric-4(md1088), unc-13(s69), and unc-18(md299) mutants were elevated (SI Appendix, Fig. S3 A–F). These results support the notion that impeding SNARE complex formation, and thus SV exocytosis, directly increases the abundance of presynaptic UNC-2 channels. Furthermore, a compromise in SV endocytosis elevated the levels of WT UNC-2 channels, as unc-26 and unc-57 mutants showed increased levels of UNC-2 channels (SI Appendix, Fig. S3 G and H). We noted that the effect of the unc-26 and unc-57 mutations on UNC-2 channel levels was more pronounced than that of the SV exocytosis mutants, implying that UNC-26 and UNC-57 are involved in both the SV and UNC-2 endocytic pathways.

To confirm the localization of UNC-2 channels to the surface plasma membrane, we used genome editing to insert superecliptic pHluorin, a pH-sensitive GFP variant, into the second extracellular loop of UNC-2, creating the cim165[pHluorin::unc-2] strain (Fig. 2A). Although the puncta intensity of pHluorin::UNC-2 was lower than that of GFP::UNC-2 due to the intrinsically low pHluorin brightness, its average intensity was significantly greater in unc-13 mutants, confirming that UNC-2 channels are indeed localized to the plasma membrane of AZs in SV exocytosis mutants (SI Appendix, Fig. S4).

Optogenetic Enhancement of Neuronal Activity Decreases the Abundance of UNC-2 Channels at Synapses.

To bolster our conclusion that SV exocytosis has an inverse relationship with UNC-2/CaV2 channel levels, we asked whether increasing neuronal excitability using optogenetic stimulation would change the abundance of UNC-2. We first generated an integrated transgenic line, cimIs61[rgef-1p::CeChrimson], expressing the red light-activated channelrhodopsin Chrimson under the control of the panneuronal rgef-1 promoter (48). The strain cimIs61;cim104[GFP::unc-2] was exposed to 590 nm wavelength light at very low intensity (~1.5 mW/cm²) for durations ranging from 1 to 20 h and then evaluated for UNC-2 levels. While 1 h of optogenetic stimulation was not sufficient to impact UNC-2 levels, sessions exceeding 2 h led to a decrease in UNC-2 channel abundance without affecting the number of puncta (SI Appendix, Fig. S5 A and B). Moreover, this decrease was not dependent on the quantal contents, as unc-17 (vesicular acetylcholine transporter) mutants showed analogous decreases (SI Appendix, Fig. S5 C and D). Importantly, compared with unexposed animals, control animals exposed to the same light conditions but without retinal supplementation exhibited no significant difference in UNC-2 levels, confirming the specific effect of optogenetic activation. This result supports the notion that increased neuronal excitability and subsequent elevated SV exocytosis reduce UNC-2 channel abundance.

To further determine the duration needed for UNC-2 channel turnover at the synapse, we subjected single cimIs61 animals to 590 nm light stimulation at a high intensity (30 to 60 mW/mm²) (48). A comparison of pre- and postexposure images from the same animal showed that 20 min of exposure to light significantly reduced the puncta intensities compared to those of control animals without retinal supplementation (Fig. 3 I and J). These results illustrate that UNC-2 channel levels are reduced with enhanced neuronal activity and consequentially elevated SV exocytosis at a turnover rate of approximately 20 min.

An Increase in SV Exocytosis Antagonizes the Abundance of UNC-2 Channels at Synapses.

UNC-2DA channels exhibited greater abundance than did WT UNC-2 channels (Fig. 2 H–J). What drives this increased abundance? Because our data showed that SV exocytosis mutations cause an increase in UNC-2 and UNC-2SL channel abundance, we hypothesize that the increase in UNC-2DA channel abundance results from reduced SV exocytosis. Specifically, inactive UNC-2DA channels could hinder calcium influx into synapses, a crucial step for SV fusion, thereby impeding SV exocytosis. This reduction in exocytosis may then trigger a compensatory increase in UNC-2DA channel levels as a homeostatic response. If this hypothesis holds true, enhancing synaptic transmission in cim147[GFP::unc-2DA] animals should lead to a reduction in UNC-2DA channels. To test this hypothesis, we introduced mutations enhancing SV exocytosis into cim147 animals and assessed their UNC-2DA levels.

The large-conductance, calcium-activated potassium K⁺ (BK) channel localizes to presynaptic AZs alongside the CaV2 channel (49–51) and functions as a negative regulator of synaptic transmission (52). In the absence of the C. elegans BK channel ortholog SLO-1, UNC-2 channel function is up-regulated, thus enhancing SV exocytosis. We found that the introduction of the slo-1(eg142) null mutation into cim147[GFP::unc-2DA] animals reduced the intensity of the UNC-2DA puncta to levels comparable to those of WT animals (Fig. 4 A and B), indicating that an increase in synaptic transmission has a negative effect on UNC-2 channel abundance.

Fig. 4.

An increase in the abundance of UNC-2DA channels is accompanied by an increase in core AZ protein levels. (A and B) Increased SV exocytosis reduced the abundance of GFP::UNC-2DA channels. Representative DNC images (A), average intensities (B), and puncta numbers (B) of GFP::UNC-2 or GFP::UNC-2DA in the WT, slo-1(eg142), unc-64(cim152), and tom-1(ok285). (Scale bar, 5 μm.) One-way ANOVA with Tukey’s post hoc test. ns, not significant. (C–J) Representative DNC images, average intensities, and puncta numbers of UNC-10::GFP (C and D), GFP::SYD-2 (E and F), GFP::RIMB-1 (G and H), and CLA-1::GFP (I and J) in WT and unc-2(cim150[unc-2DA]) animals. (Scale bar, 5 μm.) Student’s t test

TOM-1/Tomosyn negatively regulates SV exocytosis through its interaction with the SNARE complex (34, 35, 53, 54). Specifically, TOM-1/Tomosyn competes with SNB-1/synaptobrevin and binds to the SNARE proteins UNC-64/Syntaxin and RIC-4/SNAP-25, thereby limiting SNARE complex formation. Loss of TOM-1 function enhances vesicle priming and exocytosis (54, 55). The introduction of the tom-1(ok285) mutation into cim147[GFP::unc-2DA] animals considerably decreased the puncta intensity compared to that in cim147 animals alone (Fig. 4 A and B), indicating that elevating the efficiency of SV exocytosis leads to a reduction in UNC-2 channel levels.

To further ascertain the effect of SV exocytosis on UNC-2 channel levels, we focused on UNC-64/syntaxin. Syntaxin assumes a “closed” conformation, but for SNARE complex formation, it requires an “open” conformation to allow its interaction with SNAP-25 and synaptobrevin (32). A constitutively open form of UNC-64 bypasses SV priming by UNC-13/Munc13 and thus enhances SV exocytosis and evoked neurotransmitter release (33, 56). Using CRISPR/Cas9, we generated the constitutively open form of unc-64(cim152) containing the L166A and E167A mutations and crossed it with cim147[GFP::unc-2DA] animals. The intensity of the GFP::UNC-2DA puncta was reduced to levels comparable to those in WT animals (Fig. 4 A and B). In line with the findings in GFP::UNC-2DA, the puncta intensity of GFP::UNC-2 was slightly but significantly lower in unc-64(cim152) mutants than in WT animals (SI Appendix, Fig. S6). Together these results indicate that enhancing the efficiency of SV exocytosis with the open syntaxin mutation resulted in a reduction in UNC-2 channel abundance.

A Homeostatic Increase in UNC-2 Channel Abundance Requires an Increase in Select Core AZ Proteins.

Having established that the function of the SV exocytosis machinery influences UNC-2 channel levels, we considered the role of core AZ proteins in determining UNC-2 channel abundance. Core AZ proteins form a complex network through multivalent, redundant interactions and organize CaV2 channels, SVs, and SV exocytosis components, thereby mediating SV priming and fusion (57–59). Although core AZ proteins are highly redundant in function, certain AZ proteins are more critical for localizing CaV2 channels at the presynaptic AZ than others are. In a recent study, we demonstrated that UNC-10/RIM and SYD-2/Liprin-α play crucial roles in UNC-2 synaptic localization, with an additive effect of RIMB-1/RIM-BP or ELKS-1 (13). Recognizing their role in UNC-2 channel localization, we assessed the levels of core AZ proteins in unc-2(cim150[unc-2DA]) mutant animals. We used endogenously tagged AZ proteins to compare their abundance in WT and cim150 animals. We found that the puncta intensities of UNC-10, SYD-2, RIMB-1, and CLA-1 (a Bassoon homolog) were considerably greater in unc-2(cim150[unc-2DA]) animals than in WT animals, but there were no changes in puncta numbers (Fig. 4 C–J). However, the levels of ELKS-1 and UNC-13 did not significantly change in unc-2(cim150[unc-2DA]) animals (SI Appendix, Fig. S7). These results indicate that the homeostatic increase in UNC-2DA channels is accompanied by an increase in select core AZ proteins.

The increase in select core AZ proteins in unc-2(cim150[unc-2DA]) animals prompted us to determine whether these proteins are necessary for the increase in UNC-2DA channel abundance. We introduced unc-10(md1117), syd-2(ok217), rimb-1(ce828), and elks-1(ok2762) null mutations into cim104[GFP::unc-2] and cim147[GFP::unc-2DA] animals and compared the UNC-2 channel levels. Mutations in unc-10, syd-2, and rimb-1 caused UNC-2DA channel levels to decrease to levels comparable to those of WT UNC-2 channels with the corresponding mutations, indicating that UNC-10, SYD-2, and RIMB-1 are critical not only for normal UNC-2 localization but also for the increase in UNC-2 channel abundance in response to a deficit in synaptic transmission (Fig. 5 A–F). However, even though the elks-1 mutation reduced UNC-2DA channel levels, these levels were still greater than those of WT UNC-2 channels (Fig. 5 G and H), supporting that ELKS-1 is not linked to the homeostatic increase in UNC-2DA channels. Therefore, the increase in UNC-2DA abundance requires the reorganization of select AZ components.

Fig. 5.

A homeostatic increase in UNC-2DA channels requires the AZ proteins UNC-10/RIM, SYD-2/Liprin-α, and RIMB-1/RIM-BP. Mutations in unc-10 (A and B) and syd-2 (C and D) reduced the average intensities and puncta numbers of both GFP::UNC-2 and GFP::UNC-2DA to similar levels. (E and F) The rimb-1(ce828) mutation reduced the average intensities of both GFP::UNC-2 and GFP::UNC-2DA to similar levels. (G and H) The elks-1 mutation reduced the average intensities of GFP::UNC-2 and GFP::UNC-2DA to different levels, indicating the dispensability of ELKS-1 in the homeostatic increase in UNC-2DA. One-way ANOVA with Tukey’s post hoc test. ns, not significant.

Forward Genetic Screening Identified wwp-1/WWP1 as a Critical Regulator of the Degradation of UNC-2/CaV2.

Next, we sought to identify the components controlling the abundance of UNC-2 channels. The CaV2 channel consists of one pore-forming α1 subunit and two auxiliary subunits, a cytosolic β and an extracellular α2δ. Likewise, C. elegans UNC-2, which encodes the pore-forming α1 subunit, associates with CCB-1/β and UNC-36/α2δ. In a recent study, we demonstrated that UNC-2 channels recruit these auxiliary subunits to maintain stability at synapses (60). In the absence of these auxiliary subunits, UNC-2 channels are prone to degradation. In particular, the loss of UNC-36 function dramatically reduced the number and intensity of UNC-2 channels at synapses, thus preventing UNC-2 from being readily detectable in the nerve ring, a synapse-rich axon bundle structure that, together with the head ganglia, is considered to be the C. elegans brain (Fig. 6A). By leveraging these findings, we designed a forward genetic screen to identify the components controlling CaV2 channel abundance. We reasoned that a loss-of-function mutation in a gene that mediates UNC-2 degradation in unc-36 mutants would restore GFP::UNC-2 signals in the nerve ring. We screened the F2 progeny of mutagenized unc-36;cim104[GFP::unc-2] mutants to isolate suppressor mutants with increased GFP signals in the nerve ring. In this screen, we identified three independent alleles, cim73, cim74, and cim76, of the spad-1 gene (suppressor of unc-36/α₂δ). We identified spad-1 as a previously named gene called wwp-1. The cim73, cim74, and cim76 alleles had Q220X, R717W, and G724R mutations, respectively, in the predicted WWP-1 coding sequence (Fig. 6B). R717 and G724 are positioned within the highly conserved and structurally critical region of the ubiquitin ligase domain. The wwp-1 gene encodes a HECT E3 ubiquitin ligase with four WW domains. Based on sequence homology, its mammalian homologs are WWP1, WWP2, and ITCH; all have four WW domains, which confer substrate specificity, and one C2 domain, which interacts with membranes. The introduction of a WT copy of the wwp-1 gene into neurons of unc-36;wwp-1(cim73) mutants reduced the endogenous UNC-2 levels to those of unc-36 null mutant animals (Fig. 6A), confirming that WWP-1 regulates UNC-2 channel levels in neurons. Quantification of DNC also revealed an increase in UNC-2 puncta intensity and number in wwp-1;unc-36 double-mutant animals compared to unc-36 single mutant animals (Fig. 6 C and D). Taken together, these results indicate that WWP-1 is a crucial regulator of UNC-2 channel levels at the presynaptic AZ in unc-36 mutant animals.

Fig. 6.

The E3 ubiquitin ligase WWP-1 is a regulator of UNC-2/CaV2 channel abundance. (A) Genetic screen for suppressor mutants of unc-36(e251);unc-2(cim104[GFP::unc-2]). GFP::UNC-2 signals in the nerve rings of WT, unc-36, unc-36;wwp-1(cim73), and unc-36;wwp-1(cim73);wwp-1(+neuron) animals. (B) A schematic diagram of the predicted domains of WWP-1 with the mutation sites in three alleles. (C and D) UNC-2 channel levels were elevated when the wwp-1(cim73) mutation was introduced into unc-36(e251) animals. Representative GFP::UNC-2 images (C), average puncta intensities (D), and puncta numbers (D) from the DNC of unc-36 and wwp-1;unc-36 animals. Unpaired Student’s t test. (E) Snapshots of WT, unc-36, and wwp-1;unc-36 animals 40 s after transfer to a spot (red asterisk). (F) The basal speed of unc-36 animals in the absence of bacterial food was improved by the wwp-1 mutation in a UNC-2-dependent manner (1 min videos, 2 fps). One-way ANOVA with Tukey’s post hoc test. (G and H) UNC-2 channel levels were elevated in wwp-1(cim73) animals. Representative GFP::UNC-2 images (G), average puncta intensities (H), and puncta numbers (H) from the DNC of WT and wwp-1 animals. (Scale bar, 5 μm.)

Next, we assessed the mobility of unc-36;wwp-1(cim73) animals to determine whether the observed increase in UNC-2 channel levels leads to a change in their behavioral phenotype. We observed that unc-36;wwp-1(cim73) double-mutant animals moved better than unc-36 single mutant animals did, particularly when mechanically stimulated (Fig. 6F). Quantitative analysis of their off-food movement speed further demonstrated that their movement was significantly greater than that of unc-36 and unc-2 null mutant animals (Fig. 6G). Importantly, this improvement was dependent on UNC-2 as the unc-2(e55);unc-36(e251);wwp-1(cim73) and unc-2(e55);wwp-1(cim73) mutants did not improve (Fig. 6G). These findings showed that increased levels of UNC-2 channels at synapses, even in the absence of the UNC-36/α2δ subunit, improve behavioral and neuromuscular functions.

In addition to its role in preventing UNC-2 degradation in unc-36 mutants (60), WWP-1 may also play a role in maintaining normal levels of presynaptic UNC-2 channels to regulate synaptic transmission. To test this possibility, we investigated whether the increase in UNC-2 levels in the absence of WWP-1 was also observed in the WT background. Comparison of UNC-2 puncta between WT and wwp-1 animals revealed that UNC-2 puncta in wwp-1 animals had greater intensity than those in WT animals (Fig. 6 G and H). Furthermore, the number of UNC-2 puncta in the wwp-1 animals was greater than that in the WT animals, indicating slower turnover in the absence of WWP-1 (Fig. 6H). Taken together, these results demonstrate that WWP-1 regulates presynaptic UNC-2 levels in WT animals.

WWP-1 Ubiquitinates UNC-2/CaV2 Channels for Removal from the Presynaptic Active Zone.

How does WWP-1 control the abundance of UNC-2 channels at synapses? We hypothesized that WWP-1 directly ubiquitinates UNC-2 channels to target them for degradation. To test this hypothesis, we compared the levels of ubiquitinated UNC-2 channels in unc-2(cim104[GFP::unc-2]) and wwp-1(cim73);unc-2(cim104) animals. We immunoprecipitated GFP::UNC-2 with GFP-nanotrap beads and then performed western blot (WB) analysis with an anti-ubiquitin antibody (Fig. 7A). When equal volumes of animals were subjected to lysis and immunoprecipitation (IP), we detected stronger ubiquitin signals in unc-2(cim104) animals than in wwp-1(cim73);unc-2(cim104) animals. Interestingly, however, probing with a GFP antibody after antibody stripping showed stronger UNC-2 signals in wwp-1(cim73);unc-2(cim104) mutants than in unc-2(cim104) animals. Consistent with our imaging data showing that wwp-1(cim73) mutant animals exhibit higher UNC-2 levels than WT animals (Fig. 6 G and H), these results support our hypothesis that WWP-1 directly ubiquitinates UNC-2 channels, leading to their degradation, and that loss of WWP-1 function increases UNC-2 levels.

Fig. 7.

Neural activity controls the abundance of UNC-2/CaV2 channels at the presynaptic AZ. (A) IP and WB analysis showed that wwp-1(cim73) animals have a higher total UNC-2 level but a lower level of ubiquitinated UNC-2 channels than WT animals. (B) The wwp-1(cim73) mutation enhanced cholinergic outputs. wwp-1;zxIs6[unc-17p::ChR2::YFP] animals exhibited uncoordinated, sluggish movement upon low-intensity blue light exposure, while zxIs6 animals did not. The plot shows the average speeds of 10 animals before, during, and after blue light exposure. (C) wwp-1(cim73) animals are more sensitive to aldicarb-mediated paralysis than WT animals. (D and E) Loss of UNC-13 did not further increase the intensity of UNC-2SL channels in wwp-1 animals but suppressed puncta number. Representative GFP::UNC-2SL images (D), average intensities (E), and puncta numbers (E) of wwp-1(cim73), unc-13(s69), and wwp-1(cim73) unc-13(s69) animals. (Scale bar, 5 μm.) One-way ANOVA with Tukey’s post hoc analysis. (F and G) The unc-10(cim166[unc-10K97/99E]) mutation increases UNC-2SL channel levels. Representative GFP::UNC-2SL images (F), average intensities (G), and puncta intensities (G) in the WT and unc-10(cim166) backgrounds. (Scale bar, 5 μm.) Unpaired Student’s t test. ns, not significant. (H) Model of UNC-2/CaV2 degradation. UNC-2 channels maintain constant levels in the AZ through slow delivery and degradation. Elevated SV exocytosis results in lateral compression and displaces UNC-2 channels to the endocytic zone, causing them to be removed. Reduced SV exocytosis diminishes UNC-2 displacement, increasing UNC-2 levels in the AZ.

Next, we examined whether the increase in CaV2 channel levels in wwp-1 mutant animals induces a behavioral trait indicative of increased presynaptic neurotransmitter release. We introduced the zxIs6[unc-17p::ChR2::YFP] transgene, which drives the expression of channelrhodopsin in cholinergic motor neurons, into wwp-1(cim73) animals. We then compared the movement of wwp-1(cim73);zxIs6 animals under low-intensity (~1.5 mW/cm²) blue light exposure with that of control zxIs6 animals (Fig. 7B and Movie S1). Although this low light intensity slightly enhanced the movement of zxIs6 animals presumably due to innate blue light avoidance behavior, it impaired coordinated movement or caused partial paralysis in wwp-1(cim73);zxIs6 animals. Consistent with the light-induced paralysis behavior in wwp-1(cim73) animals, we noted that these animals were more sensitive to aldicarb (an acetylcholine esterase inhibitor)-mediated paralysis than were WT animals (Fig. 7C), indicating an increase in presynaptic neurotransmitter release. Collectively, these findings demonstrate that wwp-1 mutations result in an increase in UNC-2/CaV2 channels, thereby enhancing synaptic transmission.

Our findings indicate that while WWP-1 regulates UNC-2 channel abundance at synapses through ubiquitination, SV exocytosis inversely regulates UNC-2 channel abundance. To investigate the potential functional interplay between WWP-1 activity and SV exocytosis, we conducted a genetic epistatic analysis. Introduction of the unc-13(s69) null mutation into wwp-1(cim73);unc-2(cim112[GFP::unc-2SL]) mutant animals did not lead to a further increase in the puncta intensity of UNC-2SL channels compared to that of unc-13(s69);unc-2(cim112) and wwp-1(cim73);unc-2(cim112) mutant animals (Fig. 7 D and E). This observation suggests that UNC-13 and WWP-1 may function in the same pathway. Notably, while we consistently observed a slight but significant increase in the number of UNC-2SL puncta in wwp-1(cim73) mutant animals, this increase was not detected in unc-13(s69) wwp-1(cim73) double-mutant animals. Similar patterns were also noted for WT UNC-2 channels (SI Appendix, Fig. S8). These results strongly suggest that SV exocytosis, mediated by UNC-13, precedes the initiation of WWP-1-mediated UNC-2 channel degradation.

As in mammals (61), UNC-13 features two isoforms, UNC-13L and UNC-13S, which differ chiefly in the presence of an N-terminal C2A domain, although both are involved in mediating SV exocytosis. The longer UNC-13L, equipped with the N-terminal C2A domain, engages with the zinc finger domain of UNC-10/RIM and specifically localizes to the AZ (14, 37, 61). In contrast, the shorter UNC-13S, which lacks this interaction with UNC-10, is more diffusely distributed throughout the presynaptic terminal. Given that UNC-10 plays a pivotal role in anchoring UNC-2 at the AZ (13) and directly interacts with UNC-13L (37, 61), we hypothesized that disrupting the UNC-10 and UNC-13L interaction would reduce the efficiency of SV exocytosis at the AZ. This in turn would decrease the displacement of UNC-2 channels, resulting in diminished degradation of UNC-2 channels by WWP-1. We used CRISPR/Cas9 to target two highly conserved lysine residues (K97/99) in the zinc finger domain of UNC-10/RIM that interact with UNC-13 (61). The resulting unc-10(cim166[unc-10K97/99E]) mutation was introduced into cim112[GFP::unc-2SL]. We observed an increase in UNC-2SL and UNC-2 channel levels in unc-10(cim166) mutant animals (Fig. 7 F and G and SI Appendix, Fig. S9). These results support the idea that the WWP-1-mediated degradation and removal of UNC-2 channels are dependent on SV exocytosis at the AZ.

Discussion

In this study, we identified a presynaptic homeostatic mechanism regulating CaV2/UNC-2 channel abundance at synapses in C. elegans. Our quantitative imaging revealed that the abundance of UNC-2 channels at the presynaptic AZ is negatively regulated by SV exocytosis. Increasing SV exocytosis decreases the abundance of UNC-2 channels and vice versa.

Model for Presynaptic Homeostatic Regulation of UNC-2/CaV2 Channel Abundance by SV Exocytosis.

How does SV exocytosis control the abundance of UNC-2 channels? SV exocytosis produces a membrane load at the plasma membrane of AZs, the magnitude of which depends on the duration and frequency of stimulation (62). Recent EM and superresolution microscopy studies have shown that the flattening of fused vesicles exerts lateral compression in the plasma membrane and drives endocytosis (22). Considering that UNC-2 channels interact with AZ proteins in dense projections (63) and with the extracellular matrix via UNC-36/α₂δ (the extracellular auxiliary CaV2/UNC-2 subunit), they are presumed to remain stable within the AZ. However, the interactions between AZ proteins are highly dynamic and have low affinity (12). Moreover, biochemical assessment of α₂δ proteins indicated a surprisingly weak association with α₁ (64, 65). Our presented data are consistent with the model that UNC-2 channel removal is driven by SV exocytosis (Fig. 7H). An elevated rate of SV fusion, and thus a consequential increase in the membrane load and tension, leads to a high incidence of displacement of UNC-2 channels from the AZ, exposing and making them susceptible to WWP-1-mediated degradation. Conversely, a reduced SV exocytosis rate would increase presynaptic UNC-2 channel levels because UNC-2 channel replenishment tips this balance. We propose that this regulatory mechanism of UNC-2 feedback is a form of presynaptic plasticity that curbs excessive SV exocytosis by limiting UNC-2 channel abundance and conversely corrects insufficient SV exocytosis by increasing UNC-2 channel abundance. We speculate that this self-tuning, negative feedback regulation of UNC-2 channel abundance preserves the relative strength of synapses and maintains functional motor neural circuits that are essential for undulating, elegant sinusoidal movement in C. elegans.

Although most of the proteins regulating SNARE formation altered UNC-2 channel levels, we observed that mutations in snt-1/ synaptotagmin-1 and snb-1/synaptobrevin did not (SI Appendix, Fig. S1). One possible explanation for this observation is that, unlike other C. elegans SNARE genes, the genome encodes seven SNT/synaptotagmin and five SNB/synaptobrevin homologs, and their absence may be substituted by other homologous genes and/or cause a partial decrease in evoked EPSCs that is not sufficient for eliciting homeostatic changes in UNC-2 levels (66, 67). Nonet et al. noted that snt-1 and snb-1 single and snb-1;snt-1 double-mutant animals retain some movement compared to completely paralyzed unc-64 mutant animals and concluded that transmitter release is not completely abolished in the absence of SNB-1 (68). Supporting such genetic redundancy, a recent study demonstrated that C. elegans uses a dual Ca²⁺ sensor system to regulate Ca²⁺-mediated SV exocytosis: SNT-1 functions as a fast Ca²⁺ sensor and SNT-3 functions as a delayed Ca²⁺ sensor (66, 67, 69).

Homeostatic adjustment in UNC-2 channel abundance is accompanied by AZ reorganization. The levels of UNC-10/RIM, SYD-2/liprin-α, and RIMB-1 are elevated in cim150[unc-2DA] animals and are required for the increase in UNC-2DA channel abundance. Similarly, a transsynaptic homeostatic increase in Cacophony/CaV2 at Drosophila NMJs is accompanied by parallel increases in select AZ proteins (15, 70). However, specific AZ proteins, which are elevated during the transsynaptic homeostatic response, differ from those involved in the homeostatic increase in UNC-2 abundance in our study, suggesting that distinct homeostatic adjustments may require different mechanisms.

Diverse Mechanisms of PHP at Different Synapses.

The PHP observed at the Drosophila NMJ involves transsynaptic retrograde signaling (27, 28). A compromise in postsynaptic glutamate receptor function triggers the release of retrograde signals, which act in presynaptic neurons to increase neurotransmitter release in part by increasing CaV2 channel levels, thus stabilizing overall synaptic strength (29–31). Notably, the evoked EPSP response is comparable between WT and postsynaptic receptor mutant NMJs. However, C. elegans NMJs do not exhibit this form of presynaptic transsynaptic homeostasis. C. elegans has two types of acetylcholine receptors (ACh-Rs): fast-acting nicotine-sensitive and slow-acting levamisole-sensitive ACh-Rs (71). The absence of levamisole-sensitive ACh-Rs in the muscle does not cause compensatory restoration, resulting in a persistent reduction in evoked EPSCs (72). Moreover, we did not observe an increase in UNC-2 channel abundance in mutants lacking both types of ACh-Rs. Likewise, a lack of postsynaptic GABA receptors does not cause a presynaptic increase in UNC-2 channels (73). What is the explanation for this interspecies difference? Unlike the Drosophila postsynaptic muscle, which receives single excitatory glutamatergic inputs, the C. elegans body wall muscle receives both excitatory cholinergic and inhibitory GABAergic inputs. A muscle excitation deficit leads to a compensatory reduction in GABA input, which in turn increases muscle excitability (74, 75). This property allows the C. elegans body wall muscle to compensate without triggering a transsynaptic homeostatic response. For this reason, it is recognized that C. elegans NMJs more closely resemble multi-input neural circuits than do conventional NMJs (75). The diverse types of homeostatic plasticity at different synapses are not surprising since each synapse has evolved to optimize its function in neural circuits.

Self-Tuning PHP May Not be Unique to C. elegans.

Although the type of PHP we observed in this study has not been explicitly described, several reported observations may result from the same type of self-tuning presynaptic plasticity. Di Guilmi et al. observed that CaV2.1 knock-in mice with the gain-of-function S240L mutation displayed a reduction in channel density in the calyx of Held while showing a greater evoked response (24). This observation is similar to our UNC-2SL level quantification data, which showed a reduction in presynaptic UNC-2SL levels compared to WT UNC-2 levels (Fig. 2 H–J). Studies with cultured hippocampal mouse neurons showed that chronic tetrodotoxin (TTX) treatment in neural networks caused an increase in presynaptic Ca²⁺ transients, while chronic gabazine (a GABA agonist) treatment led to a decrease in Ca²⁺ transients (76–78). These observations are dependent on P/Q-type CaV2.1 channels, even though whether channel abundance or calcium influx per channel opening is responsible has not been completely resolved (77, 78). Additionally, global neural network silencing in cultured rat cortical neurons up-regulates RIM and CaV2.1 (79). These homeostatic effects are presumed to result from the suppression of postsynaptic neurons but may also originate from the presynaptic neurons themselves since the pharmacological reagents used in the experiments can be equally accessible to both pre- and postsynaptic neurons.

Our model predicts that highly active synapses exhibit a high turnover rate of synaptic proteins, such as AZ proteins and CaV2 channels, resulting in elevated degradation of these proteins. Indeed, endocytic degradation of SV proteins in mammals (80, 81) and autophagy of SV and AZ proteins in C. elegans (82) increase in an activity-dependent manner. Furthermore, in mammalian neurons, enhancing and blocking action potentials with bicuculline and TTX increase and decrease synaptic proteasome activity, respectively (83).

WWP1 Family E3 Ubiquitin Ligases as Potential Therapeutic Targets for Altering CaV2 Channel Abundance and Homeostatic Adaptation.

What is the underlying molecular mechanism for the action of WWP-1 on the removal of UNC-2? UNC-2/CaV2 channels are likely to undergo continuous removal and replenishment at AZs. We postulate that WWP-1 initiates the ubiquitination of displaced UNC-2 channels, acting at an initial step in their degradation. A lack of WWP-1 would impede the removal of UNC-2 channels from synapses, leading to slow turnover and accumulation of UNC-2 channels over time. Mutations in the pore-forming subunit of CaV2 channels in humans give rise to a range of neurological disorders, including FHM1, episodic ataxia type 2, cerebellar atrophy, and epileptic seizures (84). These mutations are either gain-of-function, resulting in increased channel activity, or loss-of-function, leading to decreased activity. Our present study indicates that changes in CaV2 channel abundance can trigger homeostatic adjustments in neuronal communication, which may play either a protective or causative role in various neurological and mental disorders. While multiple strategies aimed at reducing the levels of CaV2.2 and CaV2.1 have been explored (85, 86), therapeutic approaches for increasing CaV2 channel levels remain underinvestigated. Targeting mammalian homologs of WWP-1 to adjust CaV2 channel levels at synapses is a promising therapeutic approach for addressing disorders related to specific forms of homeostatic plasticity and CaV2 channelopathies.

Materials and Methods

Detailed information on the worm strains (SI Appendix, Table S1) and handling; genetic screening; CRISPR/Cas9 genome editing in C. elegans; microscopy and image analysis; C. elegans behavioral analysis; electrophysiology; and IP and western blotting is provided in SI Appendix.

Statistical Analysis.

All data are presented as mean ± SEM. We performed the statistical analysis using GraphPad Prism version 9. We applied an unpaired Student’s t test to compare the means of the two groups. For three or more groups, we utilized ordinary one-way ANOVA with Tukey's post hoc test to compare group means.

Supplementary Material

Appendix 01 (PDF)

Movie S1.

Movement of zxIs6[unc-17p::ChR2::YFP] and wwp-1(cim73);zxIs6[unc-17p::ChR2::YFP] animals before and after optogenetic stimulation. wwp-1(cim73) mutants exhibited enhanced cholinergic responses upon exposure to low-intensity (~1.5 mW/cm²) blue light. The results of the quantitative speed analysis before, during, and after blue light exposure are shown in Fig. 7B. The blue light on and off is indicated by a blue circle on the copper ring.

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.