Shank promotes action potential repolarization by recruiting BK channels to calcium microdomains

Mutations altering the scaffolding protein Shank are linked to several psychiatric disorders, and to synaptic and behavioral defects in mice. Among its many binding partners, Shank directly binds CaV1 voltage activated calcium channels. Here, we show that the Caenorhabditis elegans SHN-1/Shank promotes CaV1 coupling to calcium activated potassium channels. Mutations inactivating SHN-1, and those preventing SHN-1 binding to EGL-19/CaV1 all increase action potential durations in body muscles. Action potential repolarization is mediated by two classes of potassium channels: SHK-1/KCNA and SLO-1 and SLO-2 BK channels. BK channels are calcium-dependent, and their activation requires tight coupling to EGL-19/CaV1 channels. SHN-1’s effects on AP duration are mediated by changes in BK channels. In shn-1 mutants, SLO-2 currents and channel clustering are significantly decreased in both body muscles and neurons. Finally, increased and decreased shn-1 gene copy number produce similar changes in AP width and SLO-2 current. Collectively, these results suggest that an important function of Shank is to promote microdomain coupling of BK with CaV1.


Introduction
Shank is a synaptic scaffolding protein (containing SH3, PDZ, proline-rich and SAM domains) (Grabrucker et al., 2011). Mammals have three Shank genes, each encoding multiple isoforms (Jiang and Ehlers, 2013). Several mouse Shank knockouts have been described but these mutants exhibit inconsistent (often contradictory) synaptic and behavioral defects (Jiang and Ehlers, 2013), most likely resulting from differences in the Shank isoforms impacted by each mutation. The biochemical mechanism by which Shank mutations alter synaptic function and behavior has not been determined.
In humans, Shank mutations and CNVs are linked to Autism Spectrum Disorders (ASD), schizophrenia, and mania (Durand et al., 2007;Peça et al., 2011). Haploinsufficiency for 22q13 (which spans the Shank3 locus) occurs in Phelan-McDermid syndrome (PMS), a syndromic form of ASD (Phelan and McDermid, 2012). PMS patients exhibit autistic behaviors accompanied by hypotonia, delayed speech, and intellectual disability (ID) (Bonaglia et al., 2011). Heterozygous inactivating Shank3 mutations are found in sporadic ASD and schizophrenia (Durand et al., 2007;Peça et al., 2011). A parallel set of genetic studies suggest that increased Shank3 function also contributes to psychiatric diseases. 22q13 duplications spanning Shank3 are found in ASD, schizophrenia, ADHD, and bipolar disorder (Durand et al., 2007;Failla et al., 2007;Han et al., 2013). A transgenic mouse that selectively over-expresses Shank3 exhibits hyperactive behavior and susceptibility to seizures (Han et al., 2013). Taken together, these studies suggest that too little or too much Shank3 can contribute to the pathophysiology underlying these psychiatric disorders.
Given its link to psychiatric disorders, there is great interest in determining how Shank regulates circuit development and function. Shank is highly enriched in the post-synaptic densities of excitatory synapses; consequently, most studies have focused on the idea that Shank proteins regulate some aspect of synapse formation or function. Through its various domains, Shank proteins bind many proteins (Lee et al., 2011;Sakai et al., 2011), thereby potentially altering diverse cellular functions. Shank proteins have been implicated in activity induced gene transcription (Perfitt et al., 2020;Pym et al., 2017), synaptic transmission (Zhou et al., 2016), synapse maturation (Harris et al., 2016), synaptic homeostasis (Tatavarty et al., 2020), cytoskeletal remodeling (Lilja et al., 2017), and sleep (Ingiosi et al., 2019). Each of these defects could contribute to the neurodevelopmental and cognitive deficits observed in ASD and schizophrenia.
Several recent studies suggest that an important function of Shank is to regulate the subcellular localization of ion channels. Shank mutations decrease the synaptic localization of NMDA and AMPA type glutamate receptors (Peça et al., 2011;Won et al., 2012). Other studies show that Shank proteins promote delivery of several ion channels to the plasma membrane, including HCN channels (Yi et al., 2016;Zhu et al., 2018), TRPV channels (Han et al., 2016), and CaV1 voltage activated calcium channels (Pym et al., 2017;Wang et al., 2017). Of these potential binding partners, we focus on CaV1 because human CACNA1C (which encodes a CaV1 α-subunit) is mutated in Timothy Syndrome (TS), a rare monogenic form of ASD (Splawski et al., 2005;Splawski et al., 2004), and polymorphisms linked to CACNA1C are associated with multiple psychiatric disorders (Psychiatric Genomics, 2013). For this reason, we asked how Shank regulates the coupling of CaV1 channels to their downstream effectors.
C. elegans has a single Shank gene, shn-1. The SHN-1 protein lacks an SH3 domain but has all other domains found in mammalian Shank proteins. Mammalian Shank proteins directly bind CaV1 channels through both the SH3 and PDZ domains (Zhang et al., 2005). We previously showed that the SHN-1 PDZ domain directly binds to a carboxy-terminal ligand in EGL-19/CaV1 (Pym et al., 2017). CaV1 channels are tightly coupled to multiple downstream calcium activated pathways. C. elegans and mouse Shank proteins have been shown to promote CaV1-mediated activation of the transcription factor CREB (Perfitt et al., 2020;Pym et al., 2017).
Here, we test the idea that SHN-1 regulates CaV1 coupling to a second effector, calcium activated potassium currents (which are mediated by BK channels). BK channels are activated by both membrane depolarization and by cytoplasmic calcium. At resting cytoplasmic calcium levels (~100 nM), BK channels have extremely low open probability. Following depolarization, cytoplasmic calcium rises thereby activating BK channels. BK channels bind calcium with a K d ranging from 1 to 10 μM (Contreras et al., 2013); consequently, efficient BK channel activation requires tight spatial coupling to voltage activated calcium (CaV) channels. BK channels associate with all classes of CaV channels (Berkefeld et al., 2006). The co-clustering of BK and CaV channels allows rapid activation of hyperpolarizing potassium currents following depolarization. BK channels decrease action potential (AP) durations, promote rapid after hyperpolarization potentials, decrease the duration of calcium entry, and limit secretion of neurotransmitters and hormones in neurons and muscles (Adams et al., 1982;Edgerton and Reinhart, 2003;Petersen and Maruyama, 1984;Storm, 1987). Thus, BK channels have profound effects on circuit activity.

Results
SHN-1 acts in muscles to regulate action potential duration EGL-19/CaV1 channels mediate the primary depolarizing current during body muscle APs (Jospin et al., 2002;Liu et al., 2011). Because SHN-1 directly binds EGL-19 (Pym et al., 2017), we asked if SHN-1 regulates muscle AP firing patterns. In WT animals, body muscles exhibit a pattern of spontaneous AP bursts (~10 APs/burst; burst frequency 0.1 Hz) ( Figure 1A). Within a burst, APs became progressively wider ( Figure 1B). A similar pattern of progressive AP broadening during burst firing has been reported for many neurons (Geiger and Jonas, 2000;Jackson et al., 1991). Outward potassium currents were progressively decreased during repetitive depolarization ( Figure 1C), suggesting that progressive AP broadening most likely results from accumulation of inactivated potassium channels during bursts, as seen in other cell types (Geiger and Jonas, 2000;Kole et al., 2007). Occasionally, WT muscles also exhibited prolonged depolarizations ( > 150ms), which are hereafter designated plateau potentials (PPs). PPs often occur at the end of an AP burst ( Figure 1A).
In shn-1 null mutants, PP rate and AP widths were significantly increased, AP amplitudes were decreased, resting membrane potential (RMP) was depolarized, while AP frequency and input resistance were unaltered ( Figure 1D-  , and input resistance (R in , I) are compared in WT and shn-1 null mutants. All shn-1 data were obtained from shn-1(nu712) except for R in (I), which were from shn-1(tm488). Values that differ significantly from wild type controls are indicated (ns, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001). Error bars indicate SEM.

SHN-1 binding to EGL-19 promotes AP repolarization
Because SHN-1 has multiple binding partners, we sought to confirm that prolonged APs result from decreased SHN-1 binding to EGL-19. To address this question, we recorded APs in strains containing mutations that disrupt this interaction ( Figure 3A). PP frequency was significantly increased by a deletion removing the EGL-19 carboxy-terminal PDZ ligand [egl-19(nu496 ΔVTTL)] ( Figure 3B). AP widths were significantly increased by a deletion removing the SHN-1 PDZ domain [shn-1(nu542 ΔPDZ)] and by the egl-19(nu496 ΔVTTL) mutation ( Figure 3C). Furthermore, the shn-1(nu712 null) and egl-19(nu496 ΔVTTL) mutations did not have additive effects on PP rate and AP widths in double mutants ( Figure 3-figure supplement 2). Taken together, these results support the idea that SHN-1 binding to EGL-19/CaV1 accelerates AP repolarization.
The online version of this article includes the following figure supplement(s) for figure 4: Consistent with this idea, AP duration was significantly increased in mutants lacking SHK-1/KCNA ( Figure 4E-F), as previously reported Liu et al., 2011). Contradictory results have been reported for AP firing patterns in slo-1 and slo-2 BK mutants Liu et al., 2011). These studies used intracellular solutions that alter BK channel function. In (Liu et al., 2011), an intracellular solution containing high chloride levels was used, thereby exaggerating SLO-2's contribution to AP repolarization (Yuan et al., 2000). In , an intracellular solution containing a fast calcium chelator (BAPTA) was used, which inhibits BK activation thereby minimizing their impact on APs. We re-investigated the effect of SLO channels on APs using intracellular solutions with low chloride and a slow calcium chelator (EGTA), finding that AP durations were increased to a similar extent in slo-1 and slo-2 single mutants ( Figure 4E-F). Taken together, these results confirm that SHK-1/KCNA and SLO/BK are the primary channels promoting AP repolarization in body muscles.
SLO-1 and SLO-2 function together to promote AP repolarization SLO-1 and SLO-2 subunits are co-expressed in muscles and could potentially co-assemble to form heteromeric channels. To determine if channels containing both SLO-1 and SLO-2 regulate AP repolarization, we analyzed AP widths in slo-1; slo-2 double mutants. AP widths in slo-1; slo-2 double mutants were not significantly different from those found in the single mutants ( Figure 4F). Because slo-1 and slo-2 mutations did not have additive effects on AP widths, these results support the idea that heteromeric SLO-1/2 channels mediate rapid repolarization of muscle APs.
We did several experiments to further test the idea that SLO-1 and SLO-2 function together in heteromeric channels. First, we recorded voltage-activated potassium current in body muscles and found that Ik hiCl was modestly reduced in slo-1 mutants, was dramatically reduced in slo-2 mutants, and was not further reduced in slo-1; slo-2 double mutants ( Figure 5A-B). These results suggest that Ik hiCl is mediated by heteromeric channels (containing both SLO-1 and SLO-2 subunits) and by SLO-2 homomers.
As a final test of this idea, we asked if subcellular localization of SLO-1 and SLO-2 subunits requires expression of both subunits. For this analysis, endogenous SLO-1 and SLO-2 subunits were labeled with split GFP. Using CRISPR, we introduced the eleventh β-strand of GFP (GFP 11 ) into the endogenous slo-1 and slo-2 genes and visualized their expression by expressing GFP 1-10 in body muscles. Strains containing the GFP 11 tagged alleles exhibited wild-type AP widths, RMP, and Ik hiCl currents, indicating that the tag did not interfere with SLO channel function ( Figure 5-figure supplement 1). Using these alleles, we find that SLO-2 puncta intensity was significantly reduced in slo-1 null mutants, indicating that channels containing SLO-2 subunits require SLO-1 for their trafficking ( Figure 5C-D). By contrast, SLO-1 puncta intensity was unaffected in slo-2 mutants, suggesting that BK channels lacking SLO-2 were trafficked normally ( Figure 5E-F). Collectively, these results suggest that rapid muscle repolarization following APs is mediated by SLO-1/2 heteromeric channels and by SLO-2 homomers. Two prior studies also suggested that SLO subunits form heteromeric channels when heterologously expressed in Xenopus oocytes. SLO-1 currents were inhibited by a dominant-negative SLO-2 construct (Yuan et al., 2000). Similarly, mammalian SLO2 subunits (KCNT1 and 2) co-assemble to form heteromeric channels (Chen et al., 2009). Our results suggest that endogenously expressed SLO subunits also form heteromeric channels in native tissues.

SHN-1 controls AP duration through BK channels
SHN-1's impact on AP duration could be mediated by changes in either SHK-1 or SLO channels. To determine if SHN-1 acts through SLO channels, we asked if slo-2 mutations block SHN-1's effects on AP widths. Consistent with this idea, AP widths in slo-2 single mutants were not significantly different from those in slo-2 double mutants containing shn-1(nu712 null), shn-1(nu542 ΔPDZ), or egl-19(nu496 ΔVTTL) mutations ( Figure 6A). These results suggest that SHN-1 controls AP duration by regulating SLO-2 channels.
SHN-1 promotes microdomain coupling of SLO-2 with EGL-19/CaV1 channels BK channels bind calcium with affinities ranging from 1 to 10 μM (Contreras et al., 2013). As a result of this calcium dependence, BK channels have very low open probability at resting cytoplasmic calcium levels (~100 nM) and efficient BK activation typically requires close spatial coupling to calcium channels (Barrett et al., 1982). We next asked if body muscle BK channels are functionally coupled to EGL-19/CaV1 channels. Consistent with this idea, Ik hiCl current was significantly decreased by nemadipine, an EGL-19/CaV1 antagonist ( Figure 7A-B; Kwok et al., 2008). The inhibitory effect of nemadipine on Ik hiCl was eliminated in slo-2 mutants ( Figure 7A-B), suggesting that the nemadipinesensitive potassium current was mediated by SLO-2.
Is EGL-19 coupling to SLO-2 mediated by microdomain signaling? To test this idea, we compared Ik hiCl and AP widths recorded with intracellular solutions containing fast (BAPTA) and slow (EGTA) calcium chelators ( Figure 7C-E). We found that Ik hiCl recorded with BAPTA was significantly smaller than that recorded with EGTA ( Figure 7C). Similarly, AP widths recorded with BAPTA were significantly longer than those recorded with EGTA ( Figure 7D-E). The effect of BAPTA on Ik hiCl and AP widths was eliminated in slo-2 mutants ( Figure 7C and E), suggesting that the BAPTA sensitive potassium current is mediated by SLO-2. BAPTA's effect on Ik hiCl and AP widths was reduced but not eliminated in shn-1 mutants ( Figure 7C and E), consistent with the partial loss of SLO-2 current in these mutants ( Figure 6D-E). Taken together, these results suggest that SHN-1 promotes SLO-2 microdomain coupling to EGL-19/CaV1.
The online version of this article includes the following figure supplement(s) for figure 5: Figure supplement 1. Analysis of GFP 11 tagged slo-1 and slo-2 alleles.
The online version of this article includes the following figure supplement(s) for figure 6:

SHN-1 regulates BK channel activation in motor neurons
Thus far, our results suggest that SHN-1 promotes EGL-19 to SLO-2 coupling in muscles. We next asked if SHN-1 also promotes coupling in motor neurons. To test this idea, we analyzed Ik hiCl in cholinergic motor neurons and found that it was significantly reduced in shn-1 null mutants ( Figure 9A-C). The slo-2 and shn-1 mutations did not have additive effects on Ik hiCl in double mutants, suggesting that the shn-1 mutation selectively decreases SLO-2 current in motor neurons ( Figure 9A-C). Consistent with decreased SLO-2 currents, we observed a corresponding decrease in axonal SLO-2(nu725 GFP 11 ) puncta fluorescence in shn-1 mutant motor neurons ( Figure 9D-E). Thus, our results suggest that SHN-1 promotes EGL-19/CaV1 to SLO-2 coupling in both body muscles and cholinergic motor neurons.

Discussion
Our results lead to six principal conclusions. First, we show that SHN-1 acts cell autonomously in muscles to promote rapid repolarization of APs. Second, heteromeric BK channels containing both SLO-1 and SLO-2 subunits promote AP repolarization. Third, SHN-1 limits AP duration by promoting BK channel activation. Fourth, shn-1 mutants have decreased SLO-2 channel clustering and decreased SLO-2 currents. Fifth, increased and decreased SHN-1 gene dosage produce similar defects in AP durations and SLO-2 currents. And sixth, SHN-1 also promotes SLO-2 activation in motor neurons. Below we discuss the significance of these findings.
2018; Latorre et al., 2017). Thus, Shank mutations could broadly alter neuron and muscle function via changes in CaV-BK coupling. It will be very interesting to determine if this new function for Shank is conserved in other animals, including humans.

Implications for understanding neurodevelopmental disorders
Several studies suggest that Shank3 deletions and duplications are both linked to ASD and schizophrenia, suggesting that opposite changes in Shank3 levels produce similar or overlapping psychiatric phenotypes (Bonaglia et al., 2006;Durand et al., 2007;Failla et al., 2007;Gauthier et al., 2010;Han et al., 2013). It remains possible that more detailed analysis will reveal phenotypic differences between gain and loss of human Shank3. In either case, it is currently unclear how opposite changes in Shank3 levels produce psychiatric phenotypes. Different (potentially opposite) biochemical defects arising from decreased and increased Shank dosage could produce psychiatric traits, perhaps by circuit level mechanisms (Antoine et al., 2019;Peixoto et al., 2016). For example, Shank duplications and hemizygosity could act in different cells or circuits to produce psychiatric traits. Our results provide support for a second possibility. We find that increased and decreased shn-1 gene dosage produce similar cell autonomous CaV1-BK coupling defects. Two prior studies suggest that bidirectional changes in Shank produce similar defects in Wnt signaling and CaV1 current density (Harris et al., 2016;Pym et al., 2017). Collectively, these results suggest that some biochemical functions of Shank exhibit this unusual pattern of dose sensitivity and consequently could contribute to the pathophysiology of human Shankopathies (i.e. Shank3 mutations, CNVs, or PMS). The role of human Shank in CaV1-BK coupling has not been tested. Nonetheless, it seems plausible that this new physiological function could contribute to neuropsychiatric or co-morbid phenotypes associated with human Shankopathies. Consistent with this idea, PMS and human KCNMA1/ BK mutations are associated with several shared phenotypes including: autism, developmental delay, intellectual disability, hypotonia, seizures, and gastrointestinal defects (i.e. vomiting, constipation, or diarrhea) (Bailey et al., 2019;Laumonnier et al., 2006;Phelan and McDermid, 2012;Soorya et al., 2013;Witmer et al., 2019). Moreover, BK channels are regulated by two high confidence ASD genes (UBE3A and hnRNP U). BK channels are degraded by the ubiquitin ligase UBE3A (Sun et al., 2019), mutations in which cause Angelman's syndrome. The RNA binding protein hnRNP U promotes translation of slo-2 mRNA . Collectively, these results support the idea that disrupted CaV1-BK channel coupling could play an important role in shankopathies and that BK channels may represent an important therapeutic target for treating these disorders.

Strains
Strain maintenance and genetic manipulation were performed as described (Brenner, 1974). Animals were cultivated at room temperature (~22 °C) on agar nematode growth media seeded with OP50 bacteria. Alleles used in this study are described in Table 2 and are identified in each figure legend. All strains utilized are listed in the Key Resources Table. Transgenic animals were prepared by microinjection, and integrated transgenes were isolated following UV irradiation, as described (Dittman and Kaplan, 2006). Single copy transgenes were isolated by the MoSCI and miniMoS techniques (Frøkjaer-Jensen et al., 2008;Frøkjaer-Jensen et al., 2014).

CRISPR alleles
CRISPR alleles were isolated as described (Arribere et al., 2014). Briefly, we used unc-58 as a co-CRISPR selection to identify edited animals. Animals were injected with two guide RNAs (gRNAs) and two repair templates, one introducing an unc-58 gain of function mutation and a second modifying a gene of interest. Progeny exhibiting the unc-58(gf) uncoordinated phenotype were screened for successful editing of the second locus by PCR. Split GFP and split sfCherry constructs are described in Feng et al., 2017. MiniMOS inserts in which Pmyo-3 drives expression of either GFP 1-10 (nuSi144) or sfCherry 1-10 SL2 GFP 1-10 (nuSi458) were created. Tissue specific shn-1 knockout was performed by introducing LoxP sites into intron 1 and the 3'UTR of the endogenous locus, in shn-1(nu697), and expressing CRE in muscles (pat-10 promoter) or neurons (sbt-1 promoter). Tissue-specific shn-1 rescue was performed by introducing a stop cassette into intron 2 of shn-1 using CRISPR, creating the shn-1(nu652) allele. The stop cassette consists of a synthetic exon (containing a consensus splice acceptor sequence and stop codons in all reading frames) followed by a 3' UTR and transcriptional terminator taken from the flp-28 gene (the 564 bp sequence just 3' to the flp-28 stop codon). The stop cassette is flanked by FLEX sites (which are modified loxP sites that mediate CRE induced inversions) (Schnütgen and Ghyselinck, 2007). In this manner, orientation of the stop cassette within the shn-1 locus is controlled by CRE expression. Expression of shn-1 is reduced when the stop cassette is in the OFF configuration (i.e. the same Table 2. Alleles used in this study. orientation as shn-1) but is unaffected in the ON configuration (opposite orientation). The endogenous flp-28 gene is located in an intron of W07E11.1 (in the opposite orientation). Consequently, we reasoned that the flp-28 transcriptional terminator would interfere with shn-1 expression in an orientation selective manner. A similar strategy was previously described for conditional gene knockouts in Drosophila (Fisher et al., 2017).

Fluorescence imaging
Worms were immobilized on 10% agarose pads with 0.3 µl of 0.1 µm diameter polystyrene microspheres (Polysciences 00876-15, 2.5% w/v suspension). Body muscles just anterior to the vulva were imaged. Images were taken with a Nikon A1R confocal, using a 60 X/1.49 NA oil objective, with Nyquist sampling. Image volumes spanning the muscle surface were collected (~10 planes/volume, 0.15 μm between planes, and 0.06 μm /pixel). Maximum intensity projections for each volume were auto-thresholded, and puncta were identified as round fluorescent objects (area >0.1 μm 2 ), using analysis of particles. Mean fluorescent intensity in each punctum was analyzed in the raw images. All image analysis was done using FIJI.

Electrophysiology
Whole-cell patch-clamp measurements were performed using a Axopatch 200B amplifier with pClamp 10 software (Molecular Devices). The data were sampled at 10 kHz and filtered at 5 kHz. All recordings were performed at room temperature (~19°C-21°C).

Statistical methods
For normally distributed data, significant differences were assessed with unpaired t tests (for two groups) or one way ANOVA with post-hoc Dunn's multiple comparisons test (for >2 groups). For non-normal data, differences were assessed by Mann-Whitney (two groups) or Kruskal-Wallis test with post-hoc Dunn's multiple comparisons test ( > 2 groups). Data graphing and statistics were performed in GraphPad Prism 9. No statistical method was used to select sample sizes. Data shown in each figure represent contemporaneous measurements from mutant and control animals over a period of 1-2 weeks. For electrophysiology, data points represent mean values for individual neuron or muscle recordings (which were considered biological replicates). For imaging studies, data points represent mean puncta fluorescence values in individual animals (which were considered biological replicates). All data obtained in each experiment were analyzed, without any exclusions.