Introduction
Somatostatin cortical interneurons (SST cINs) constitute ~30% of all inhibitory interneurons in the cerebral cortex. They are crucial for gating the flow of the sensory, motor, and executive information necessary for the proper function of the mature cortex (Fishell and Rudy, 2011; Kepecs and Fishell, 2014; Tremblay et al., 2016). In particular, Martinotti SST cINs, the most prevalent SST cIN subtype, are present in both the infragranular and supragranular layers of the cortex and extend their axons into Layer 1 (L1; Rudy et al., 2011; Lim et al., 2018; Ascoli et al., 2008; Nigro et al., 2018; Pouchelon et al., 2021). They specifically target the distal dendrites of neighboring excitatory neurons, thus providing the feedback inhibition necessary for modulating dendritic integration (Adler et al., 2019; Kapfer et al., 2007; Silberberg and Markram, 2007). These roles are dependent upon the ability of SST cINs to form specific synaptic connections with select excitatory and inhibitory cell types during development (Favuzzi et al., 2019).
The mechanisms responsible for generating the precise functional connectivity of SST cINs are poorly understood. Early neuronal activity has emerged as an important factor in directing the maturation of cINs (Wamsley and Fishell, 2017). In addition, recent work has implicated activity as being centrally involved in alternative splicing (Eom et al., 2013; Furlanis and Scheiffele, 2018; Iijima et al., 2011; Lee et al., 2007; Lee et al., 2009; Mauger et al., 2016; Quesnel-Vallières et al., 2016; Vuong et al., 2016; Vuong et al., 2018; Xie and Black, 2001). However, whether these processes are coupled within SST cINs has not been explored.
The Nova family of RNA-binding proteins (Nova1 and Nova2) have been shown to control the splicing and stability of transcripts encoding a variety of neurotransmitter receptors, ion channels, and transmembrane cell adhesion molecules known to affect synaptogenesis and excitability (Dredge and Darnell, 2003; Eom et al., 2013; Saito et al., 2016; Saito et al., 2019; Ule et al., 2005; Ule et al., 2006; Yano et al., 2010). Notably both Nova1 and Nova2 are strongly expressed within cINs during periods of synaptogenesis and as such represent promising effectors that may direct the maturation of SST cINs.
Here, we report that neuronal activity strongly influences efferent SST cIN connectivity. We show that the conditional loss of
Results
Neuronal activity affects the synaptic development of SST cINs
The cortex exhibits a variety of dynamic network activity patterns during cortical synaptogenesis (Allene and Cossart, 2010; Garaschuk et al., 2000; Yang et al., 2009). These are comprised by both spontaneous and sensory evoked events (Garaschuk et al., 2000; Minlebaev et al., 2011; Yang et al., 2013; Pouchelon et al., 2021; Ibrahim et al., 2021). While inhibitory cortical interneurons (cINs) are recruited by these activities (Cossart, 2011; Le Magueresse and Monyer, 2013), whether this influences somatostatin (SST) cIN development has not been fully established. To address the impact of activity on these cINs, we chose to selectively and cell-autonomously dampen or augment their excitability during the first few weeks of development. This represents a perinatal period in cIN development during nascent circuit formation, where they are robustly forming or losing synaptic contacts (Allène et al., 2008; Minlebaev et al., 2011; Yang et al., 2013; Yang et al., 2009). SST cINs in the primary somatosensory cortex (S1) were targeted using AAV viral injections in
Figure 1.
Neuronal activity affects the synaptic development of SST cINs.
(A) Schematic ( (A) has been adapted from the Research Article Summary Schematic from Bernard et al., 2022) of genetic alleles (left) and experimental approach (middle),
To assess the development of the synaptic efferents of infected SST cINs, we allowed pups to mature until juvenile age. The somatosensory cortex was then subjected to immunohistochemistry (IHC) to visualize pre-synaptic (SST + cIN-mCherry+-Syp-eGFP+) compartments and post synaptic components and subjected to puncta analysis (Figure 1B). We quantified SST efferent synapses identified through the colocalization of the virally mediated mCherry reporter, Syp-eGFP and the postsynaptic marker gephyrin (mCherry+/GFP+/gephyrin +puncta), as a proxy for synaptic contacts (Ippolito and Eroglu, 2010). In SST cINs, KIR2.1 expression resulted in a significant reduction of L1 SST cIN efferent synaptic puncta in comparison to control cells (0.109±0.009 puncta/µm2 CTL vs 0.049±0.006 puncta/um2 KIR2.1; Figure 1C). By contrast, the overexpression of NaChBac within SST cINs resulted in a robust increase in L1 synaptic puncta (0.109±0.009 puncta/um2 ctl vs 0.218±0.030 puncta/um2 NaChBac; Figure 1C). Additionally, when we optogenetically activated SST neurons using a conditional channelrhodopsin mouse line
Neuronal activity influences alternative splicing and Nova expression within SST cINs
A growing number of studies indicate that activity-dependent alterative splicing (AS) contributes to the regulation of gene expression and the fine-tuning of transcriptional programs related to synaptic refinement (Eom et al., 2013; Iijima et al., 2011; Fuccillo et al., 2015, Mauger et al., 2016; Quesnel-Vallières et al., 2016; Vuong et al., 2016). This prompted us to test whether neuronal activity itself changes the level of AS within SST cINs during circuit formation, independent of the changes in gene expression. To do so, we used electro-convulsive shock (ECS) during peak synaptogenesis (P8) in mice with genetically labeled SST cINs (
Figure 2.
Neuronal activity influences alternative splicing and Nova expression within SST cINs.
(A) Schematic of experimental approach: Postnatal day 8 (P8)
Figure 2—figure supplement 1.
Acute increases in neuronal activity induces immediate early gene expression and differential splicing within SST + cINs in vivo.
(A) Schematic of experimental approach: P8 tgLhx6eGFP animals were subjected to ECS (left)’ then following a time course of 1 Hr, 3 hr, or 7 hr the S1 cortex was dissected (middle) and GFP +cINs were isolated by FACS for qPCR, Western blot, and RNAseq analysis and IHC (right). (B) Quantification of relative mRNA expression (RQ) of cFOS (normalized to housekeeping gene PPIA) in ctl/sham animals (black), 1 hr (yellow), 3 hr (orange), and 7 hr (red) following ECS within cINs. (C) Immunostaining of cFOS in sham (red) and eGFP vs ECS-treated animals, showing the expression of cFOS 1.5 hr post ECS activity induction. (D) Representative western blot of ARC protein expression within sham treated (two replicates), 1 hr (two replicates), 3 hr (two replicates), and 7 hr (two replicates) following ECS within cINs (Source Data not available due to loss of data file during lab move). (E) Fold of ARC protein expression (normalized to b-actin) in ctl/sham treated (black), 1 hr (yellow), 3 hr (orange), and 7 hr (red) following ECS within cINs. (F) Magnitude of differential alternative splicing events from the comparison of sham cINs to cINs 1 hr (yellow, 268 events), 3 hr (orange, 349 events), and 7 hr (red, 241 events) following ECS.
Figure 2—figure supplement 2.
Neuronal activity influences alternative splicing and Nova expression within SST cINs.
(A) Normalized counts of SST, Prox1, Emx1 and Vip gene expression from the RNA sequencing analysis from control tissue indicating the purity of our sample preparation to be specific for SST + cINs. (B) Bubble dot plot of gene ontology (GO) most significant terms for the genes subjected to activity-dependent alternative splicing within SST + cINs (false discovery rate (FDR)<0.05), x-axis is the enrichment of the activity-dependent AS genes in the GO category (# of genes in GO category from SST transcriptome/ # of genes activity-dependent AS in category). Color of dot indicates magnitude of significance (-log10 transform FDR, none shown above FDR <0.05) and size corresponds to number of genes in category. (C) Protein-protein interaction (PPI) network formed from 312 activity-dependent spliced genes in SST cINs with Disease Association Protein-Protein Link Evaluator (DAPPLE; Rossin et al., 2011) and performed over 10,000 permutations (pVal <0.00009). Green shading- post-synaptic gene network, pink shading- pre-synaptic gene network.
Utilizing the SST cIN transcriptome as a reference, we performed gene ontology (GO) analysis to ask if the genes subject to alternative splicing (AS) were enriched for specific functional categories within these neurons. GO analysis of the genes that underwent activity dependent AS belong to specific ontological categories, such as synapse maturation, synaptic transmission, and axonal growth (Figure 2C and Figure 2—figure supplement 2B). In addition, and as expected we also observed activity-dependent changes in gene expression (Figure 2D). However, the overlap between the genes subjected to AS vs GE was only ~1.2% of all differentially expressed genes (Figure 1E). When we compared the overlapped genes (those that underwent both GE and AS changes), we observed that for many synaptic genes, the level of AS changes was higher than the changes of the same genes at the transcript level (Figure 1F). This suggests that the changes observed in synaptic genes for AS are independent of their changes in transcription level. For example, we observed and validated (Supplementary file 2b) that within activity-stimulated SST cINs the
We next asked whether the activity-dependent AS genes formed a protein-protein interacting network (PPI) based on previously established direct protein interactions in vivo (Rossin et al., 2011). Notably, the genes subjected to activity-dependent AS within SST cINs form highly connected networks illustrating they likely function together to support pre-synaptic vesicle function (Figure 2—figure supplement 2C, pink), post-synaptic organization and receptor-associated synaptic components (Figure 2—figure supplement 2C, blue; pVal <0.0009, 1000 permutations). These genes among others include:
We next sought to identify RNA-binding proteins (RNABPs) that could mediate activity-dependent AS events within SST cINs. To do so, we utilized the RNAseq experiments described above (control vs. ECS) to perform a motif enrichment analysis that utilizes position probability matrices of binding motifs from 102 RNABPs (e.g. PTBP1/2, FUS, ELAVL4, SRRM4, Rbfox1, FMR1, Nova1, Nova2) (Liu et al., 2017; Park et al., 2016; Yang et al., 2016). Previous HITS-CLIP analysis has revealed that Nova1 and Nova2 share an almost identical RNA-binding domain (YCAY) (Licatalosi et al., 2008; Ule et al., 2006; Yuan et al., 2018). Strikingly, the Nova-binding motif was found to be significantly enriched within activity-dependent targets and at a higher frequency than other neuronal splice factors (e.g. Sam68 (KHDRBS1), SLM2 (KHDRBS2), and Rbfox1) (pVal <0.0001; Figure 2H). This finding implicates Nova proteins as playing a fundamental role in directing SST cIN activity-dependent AS.
Neuronal activity during cortical development influences the expression and localization of Nova proteins in SST cINs
We next examined the expression of Nova1 and Nova2 within SST cINs across development and whether their expression is affected by changes in neuronal activity. Utilizing IHC and genetic fate mapping, we observed that the expression of the Nova family (Nova1 and Nova2) proteins begins within cIN populations soon after they become postmitotic and expressed in 100% of SST and PV cINs by adulthood (Figure 3—figure supplement 1A–B). For comparison, we also examined Nova expression in 5HT3aR cINs (Figure 3—figure supplement 1B) within this same region. To specifically examine the expression of Nova1 and Nova2 during SST cIN synaptogenesis, we performed quantitative-PCR (qPCR) on FACS isolated cINs from the S1 cortex of Tg-Lhx6::eGFP mice at P2, P8, and P15. The Tg-Lhx6::eGFP mice express eGFP in both SST and Parvalbumin (PV) cINs (medial ganglionic eminence derived cINs) soon after they become postmitotic. We found that both
We next investigated whether Nova1 and Nova2 are activity-regulated within SST cINs by examining both their expression and localization during the peak of nascent circuit integration. To do so, we subjected
Figure 3.
Neuronal activity during cortical development influences the expression and localization of Nova proteins in SST cINs.
(A) Volcano plot of RNA seq data showing Nova1 and Nova2 upregulation in SST-cINs in ECS vs control. (B) Upper panel, western blot showing Nova1 and Nova2 protein expression in control (lanes 2 and 3) versus ECS induced SST cINs (lanes 4 and 5). Lower panel, same western blot showing expression of b-actin across lanes. Right, Quantification of the western blot data. Nova1 and Nova2 protein expression relative to β-actin in control versus ECS induced SST cINs (n=3 mice, S1 cortex only; *pVal = 0.038, Nova1; *pVal = 0.022, Nova2; Source Data not available due to loss of data file during lab move). (C) Representative scoring criteria for Nova1/2 localization within SST cINs: IHC of Nova1/2 (blue, anti-Nova1/2) in selective SST + cINS exemplifying the Nova1/2 expression in: cytoplasm only (top), nucleus only (middle) and in both cytoplasm and nucleus (bottom). (D) Left, representative images of Nova1/2 expression (blue) in SST cINs (red) under normal versus ECS. Right, Quantification of the ratio of nuclear to cytoplasmic localization of Nova1/2 in SST + cINs of control animals (grey) and ECS animals (green) (n=3 mice, S1 cortex; **pVal = 0.001). (E) Left, representative images of Nova1/2 expression (blue) in SST cINs (red) using control mCherry versus Kir2.1-mCherry virus injection. Right, Quantification of the ratio of nuclear to cytoplasmic localization of Nova1/2 in SST + cINs of control AAV2/1-Syn-DIO-mCherry (grey) versus AAV2/1-Syn-DIO-NaChBac-P2A-mCherry (pink) versus AAV2/1-Syn-DIO-Kir2.1- P2A-mCherry (blue) injected animals. (n=11 mice, S1 cortex,~30 cells each; ***pVal = 0.0004, NachBac; ***pVal = 0.0001, KIR2.1).
Figure 3—figure supplement 1.
Nova 1 and 2 alternative splicing factors expression within cortical interneurons (cINs).
(A) Representative IHC image of brain section at P21 from Dlx6aCre;RCEeGFP, Lhx6::eGFP and 5HT3aR::eGFP. Anti-Nova1/2 (red); GFP (green). (B) Quantification of eGFP cells expressing Nova1/2. 100% of Lhx6::eGFP cells express Nova1/2. (C) Relative gene expression of Nova1 (orange) and Nova2 (pink), normalized to house-keeping gene Peptidyl prolyl isomerase A (PPIA) using qPCR from Lhx6-eGFP sorted cINs at Postnatal age (P) P2, P8, and P15 (n=4 mice each, S1 cortex only). (D) Fold change of the relative expression of Nova1 and Nova2 between cINs and excitatory neurons (cExt) showing an enrichment of Nova expression in cINs at early developmental ages (n=4 mice each, S1 cortex only). (E) Relative expression of Nova1 and Nova2 genes (using qPCR) of ECS induced SST cINs relative to controls (n=4–6 mice, S1 cortex only; **pVal = 0.002, Nova1; **pVal = 0.005, Nova2). (F) Quantification of the number of Nova1/2-expressing SST +cINs of control AAV2/1-Flex-mCherry (grey) and AAV2/1-Flex-Kir2.1- P2A-mCherry (blue) injected animals. (n=7, S1 cortex,~27 cells each; ***pVal = 0.0001). (G) Representative images of Nova1/2 expression, Left: control SST cIN (injected with mCherry), Right: KIR2.1+SSt cIN at P21. Right, Quantification of Nova1/2 protein pixel intensity (normalized to area) from ctl SST cINs (grey) and KIR2.1+SST cINs (blue) (n=10; **pVal = 0.006).
Following seizure activity Nova proteins have been shown to translocate into the nucleus within excitatory neurons (Eom et al., 2013). We next sought to explore whether manipulating activity also influences intracellular localization of Nova proteins within SST cINs (Figure 3C). We hypothesized that an activity-mediated increase would direct Nova proteins to the nucleus. We therefore analyzed the ratio of Nova expression within the nucleus versus the cytoplasm of SST cINs following ECS (Figure 3D) or after constitutive activity-modulation across the first postnatal month (DIO: AAV injections of KIR2.1 or NaChBac into
Nova1 and Nova2 control distinct AS networks within SST cINs
To address whether Nova1 and Nova2 differentially affect connectivity and maturation, we asked what AS networks they control within SST cINs during development. Given that they share a very similar RNA-binding motif and are found associated with one another in vivo, they were thought to function cooperatively (Licatalosi et al., 2008; Racca et al., 2010; Yuan et al., 2018). However recently, it has been shown that in addition to their synergistic roles, Nova1 and Nova2 proteins each control distinct AS gene networks (Saito et al., 2019; Saito et al., 2016). We thus chose to examine changes in AS within SST cINs in Nova1, Nova2 or Nova1/2 compound conditional knockout (cKO) mice (Saito et al., 2019; Yuan et al., 2018). Using FAC sorting, we isolated SST cINs from
Figure 4.
Nova1 and Nova2 control distinct alternative splicing (AS) networks within SST cINs.
(A) Schematic demonstrating the mouselines used for FACS sorting and subsequent RNA sequencing and splice variant analysis: SST cINs from
Figure 4—figure supplement 1.
Nova2 controls most of the gene expression and splicing events of the Nova1/2 family within SST + cINs and these events coalesce into GO categories and PPI networks related to pre- and post-synaptic development of SST cINs.
(A–C) Gene ontology analysis of differentially expressed synaptic genes in
Figure 4—figure supplement 2.
Overlap in Alternative Splice events between SST-Nova 1, SST-Nova 2 and SST-Nova1/2 dKO.
(A) Quantification of the overlap of
We next assessed the overlap of changes in AS events observed within each mutant (Figure 4—figure supplement 2A–C). We found the number of alterations in
To infer their specific biological functions, we performed GO analysis on the altered AS events from each mutant and then asked whether the affected AS events form direct PPI networks. GO analysis of the
To confirm our predictions from the AS analysis of conditional
Figure 5.
(A) SST +cINs efferent structure: IHC of anti-RFP (red), anti-VGAT (green), and anti-Gephyrin (blue) to label the SST +cIN axonal synaptic puncta (RFP+/VGAT+/Gephyrin +puncta, white) in L1 S1 cortex of
Figure 5—figure supplement 1.
Conditional loss of
(A) Survival plot of conditional knockouts within
We also investigated whether the density of excitatory synapses onto SST cINs is affected by the loss of
Nova RNA binding proteins control-activity-dependent AS in SST cINS during development
Given that activity increases the expression level and nuclear localization of both Nova proteins, we hypothesized that their loss would result in changes in activity-dependent AS. To this end, we repeated our investigation of how Nova-dependent AS isoforms are altered in mutant mice. This time we examined the changes specifically following ECS within SST cINs during synaptogenesis in vivo. Two to 3 hr following ECS, we isolated SST cINs from
Figure 6.
Nova RNA binding proteins control-activity-dependent AS in SST cINS during development.
(A) Schematic of experimental approach: Control and
Figure 6—figure supplement 1.
Nova RNA binding proteins control-activity-dependent AS in SST cINs during development.
(A) Bubble dot plot of the most significant GO terms for the genes undergoing Nova1/2 activity-dependent AS splicing within SST +cINs (all GO terms shown FDR <0.05). (B) Schematic of an SST +cIN presynaptic inhibitory axonal puncta (top right) and a SST +cIN excitatory post-synaptic density (middle) overlaid on top of the significant DAPPLE generated PPI direct network from the 356 genes undergoing Nova1/2-dependent activity induced AS (***pVal = 0.00009, 10,1000 permutations). (C) Example RT-PCR validation of alternative splicing (AS) events of activity- and Nova1/2-dependent alternative exon usage within the gene
Figure 6—figure supplement 2.
Nova1/2 controls the activity-dependent splicing of large and unique pool of mRNAs compared to Rbfox1 within SST cINs and SST-specific Nova2 AS genes overlap well with pan-cIN Nova2 AS genes.
(A) Quantification of the overlap of
We found the majority of genes which undergo activity-induced Nova-dependent differential splicing were significantly enriched for GO categories such as pre-synaptic vesicular function, synapse organization, synaptic transmission, and neuronal growth (Figure 6C and Figure 6—figure supplement 1A). Many of the genes within these categories are known to have important functions for axon organization and synaptogenesis such as,
In sum, the activity-mediated Nova-dependent AS changes within SST cINs are central for fine-tuning of synaptic development. We previously found that another important RNABP, Rbfox1, influences axonal development and also shuttles from the cytoplasm to the nucleus upon increase in activity in SST cINs (Lee et al., 2009; Wamsley et al., 2018). However, upon comparing the activity-dependent splicing programs within SST cINs of Rbfox1 (69 activity-dependent events) to Nova1/2 (346 activity-dependent events), we found Nova proteins control a much larger number of activity-dependent splicing events. This supports our hypothesis that Nova proteins are key players in the control of activity-dependent alternative splicing (Figure 6—figure supplement 2).
Augmenting activity in Nova2 KO fails to enhance SST inhibitory output
Activity increases both the expression of Nova proteins as well as synapse formation, while conversely loss of Nova function causes a striking decrease in synaptogenesis and SST inhibitory output. Moreover, from our analysis of SST cIN KOs, it was evident that of the two Nova proteins,
Figure 7.
Augmenting activity in Nova2 KO fails to enhance SST inhibitory output.
(A) Left, experimental model: Injection of AAV-Syn-DIO-NachBac-P2A-mCherry (activating) in either control mice or
Conversely, we examined whether over-expression (OE) of Nova2 alone could phenocopy the observed changes in connectivity within SST cINs and whether that was affected by reducing the activity level of the cell (using Kir2.1). To that end, we either overexpressed
We next also examined whether suppressing activity while overexpressing Nova2 impacts the inhibitory output of SST neurons (Figure 7G left). The dual expression of Nova2-OE and KIR2.1 within SST cINs prevented the small increase of peak IPSC amplitude observed with Nova2-OE alone. Perhaps most strikingly, as with our initial KIR2.1 experiment, the levels of Nova2 protein despite being constitutively OE were reduced in cells co-expressing KIR2.1 (Figure 7E). This provides strong evidence that the stability and nuclear localization of Nova protein is dependent on the level of basal activity within SST cINs. Therefore, a certain level of activity is needed to maintain Nova protein function, and conversely, Nova proteins are needed to mediate activity-dependent changes in alternative splicing of synaptic proteins.
Discussion
In the present study, we have examined the interacting contributions of neuronal activity and the Nova RNABPs on synaptogenesis of SST cINs. Our analysis began with the observation that activity levels strongly influence the maturation of SST cINs. Acutely evoking activity during circuit integration with ECS resulted in both transcriptional and translational upregulation of Nova proteins and promoted their localization to the nucleus. This was accompanied by a striking change in both the GE and AS of synaptic genes and culminated in enhanced synaptogenesis within SST cINs. We then systematically examined the interdependence between these three observations.
Our results indicate that during circuit formation, activity levels within SST cINs correlate with changes in AS and together act to regulate the formation of afferent/efferent connectivity. These events appear to be tightly linked to Nova function, as the expression, localization and splicing activity of both Nova1 and Nova2 proteins are strongly modulated by activity. Examination of how splicing events are impacted by
With regards to AS in particular, Nova function is a core regulator of alternative splicing in many cell types, including SST cINs. It however represents only one of a host of RNABPs within the CNS. Indeed, a recent study demonstrated that within the mature brain many classes of neurons, including SST cINs, can be classified both by their expression levels of RNABPs and their corresponding repertoire of alternatively spliced mRNAs (Furlanis et al., 2019). Comparison of this work to our present findings illustrate that both the expression of RNABPs and the patterns of AS are strongly regulated across development, a phenomenon that may reflect developmental changes in neuronal activity. Consistent with this RNA binding splice factors have previously been shown to promote alternative splicing of synaptic proteins in response to neuronal depolarization and Ca2+ signaling (Eom et al., 2013; Mauger et al., 2016; Quesnel-Vallières et al., 2016; Vuong et al., 2016), For example, previous research demonstrated that the splicing of neurexins, a gene family known to function in synaptogenesis, are mediated through the actions of the SAM68 splicing factor (Iijima et al., 2011). Similarly, It has also been illustrated that neuronal activity reduces the expression of the SRRM4 RNA-binding protein, which resulted in altered RNA splicing and a corresponding decrease in excitatory synapses (Quesnel-Vallières et al., 2016). As such AS represents a largely unexplored but central genetic mechanism, capable of directing cell-type development and synaptic formation specifically.
Understanding both the repertoire of splice factors and the cell-specific patterns of splicing across development will undoubtedly provide further insight into how AS influences cIN development. One could imagine systematically examining the role of these differential splice mRNA variants through combinatorial knockdown or over-expression. However, this would face enormous technical challenges, even if restricted to only those that are Nova-dependent. As we show here many of these genes have been shown to function together (PPI networks). As such AS appears to coordinately target specific biological mechanisms. Given that the abundance of the specific splice forms of different genes within SST cINs is relative rather than absolute, it appears that AS has been coopted by development as an effective mechanism to fine-tune particular biological phenomena. The flexibility of AS to regulate the composition and levels of genes allows cells to adjust their biological function in accordance with both their identity and state (e.g. developmental period, neuronal activity, etc.). As a result, the abundance of specific splice forms co-varies as a function of both transcription and AS. Taken together, this argues that conditional removal of RNABPs, such as Nova2, provides an effective approach for understanding the role of AS within discrete cell types. Additionally, Nova proteins have a yet unexplored role in regulating gene expression, most likely through their ability to regulate the stability of RNA molecules.
In sum, our results show a clear interdependence between activity, Nova function and synaptic formation/strength in SST cINs. The interaction between activity and Nova function is bidirectional. Activity regulates the RNA, protein levels and intracellular localization of Nova proteins within SST cINs, while Nova proteins are in turn required for the activity-dependent regulation of synaptic formation and function (see model Figure 7H). When SST cIN activity is increased with ECS or with NaChBac expression,
We and others have shown that activity regulates programmed cell death (Priya et al., 2018; Denaxa et al., 2018; Wong et al., 2018). However, we observed no indication that the loss of
Contact for reagent and resource sharing
Please contact GF or LAI for reagents and resources generated in this study.
Materials and methods
Key resources table
| Reagent type (species) or resource | Designation | Source or reference | Identifiers | Additional information |
|---|---|---|---|---|
| Strain, strain background ( | SST-Cre | Jackson Laboratories | 13044 | |
| Strain, strain background ( | RCE-GFP | Jackson Laboratories | 032037-JAX | |
| Strain, strain background ( | tgLhx6;eGFP | MMRC | 000246-MU | |
| Strain, strain background ( | Nova1LoxP/LoxP | https://elifesciences.org/articles/00178 | Gift from Darnell Lab | |
| Strain, strain background ( | Nova2 LoxP/LoxP | https://elifesciences.org/articles/00178 | Gift from Darnell Lab | |
| Strain, strain background ( | TRE-Bi-SypGFP-tdTomato | Jackson Laboratories | 12345 | |
| Strain, strain background ( | Rosa-tTA LoxP/LoxP | Jackson Laboratories | 8600 | |
| Strain, strain background ( | Ai9 LoxP/LoxP | Jackson Laboratories | 7909 | |
| Strain, strain background ( | Ai32 LoxP/LoxP | Jackson Laboratories | 24109 | |
| Antibody | Anti-GFP, Chicken Polyclonal IgY | Abcam | Ab13970 | |
| Antibody | Anti-RFP (5 F8), Rat monoclonal | ChromoTek | 5 f8-100 | |
| Antibody | Anti-mCherry, Goat polyclonal | Origene | AB0040-200 | |
| Antibody | Anti-Somatostatin (YC7), Rat monoclonal | EMD Millipore | MAB354 | |
| Antibody | Somatostatin 14, Rabbit | Peninsula Labs | T-4103.0050 | |
| Antibody | Homer 1 c, Rabbit polyclonal | Synaptic systems | 160 023 | |
| Antibody | Vglut 1, Guinea pig polyclonal | Sigma | ab5905 | |
| Antibody | Gephyrin, Mouse IgG monoclonal | Synaptic systems | 147 011 | |
| Antibody | VGAT, Rabbit polyclonal | Synaptic systems | 131 003 | |
| Antibody | Nova1/2, Human polyclonal | pan-Nova (anti-Nova paraneoplastic human serum) | Gift from Darnell Lab | |
| Antibody | tagBFP, Rabbit polyclonal | Evrogen | AB233 | |
| Antibody | Anti-cFOS (4), Rabbit polyclonal | Santa Cruz Biotechnology | SC-52 | |
| Viral Vector | AAV-Syn-DIO-NachBac-P2A-mCherry | NYUAD | This paper | |
| Viral Vector | AAV-Syn-Kir2.1-P2A-mCherry | NYUAD | This paper | |
| Viral Vector | AAV-Syn-DIO-Nova2- | NYUAD | This paper | |
| Viral Vector | VTKS2 Backbone | NYUAD | Addgene_170853 | |
| Software, algorithm | BEDTools | Quinlan Lab | v2.17.0 | |
| Software, algorithm | Picard tools | Broad Institute | http://broadinstitute.github.io/picard/ | |
| Software, algorithm | DESeq2 | Bioconductor | R studio package | |
| Software, algorithm | rMATS | Xing Lab | v3.0.9 | |
| Software, algorithm | Rstudio | Rstudio.com | Version 1.1.456 | |
| Software, algorithm | Custom code | This paper | https://github.com/IbrahimLab-23/Nova-proteins-and-synaptic-integration-of-Sst-interneurons; Laboratory of Neural Circuits, 2023 | |
| Software, algorithm | ImageJ 2.0.0 Java 1.8.0_66 | National Institute of Health | https://imagej.net/; RRID:SCR_003070 | |
| Software, algorithm | Clampfit 10.7 (pClamp) | Molecular Devices | ||
| Software, algorithm | https://www.moleculardevices.com/products/software/pclamp.html; RRID:SCR_011323 | |||
| Software, algorithm | ||||
| Software, algorithm | Prism 9.1.2 | Graphpad Software | https://www.graphpad.com/; RRID:SCR_002798 | |
| Software, algorithm | ||||
| Software, algorithm | Zen Blue | Zeiss | ||
| Software, algorithm | https://www.zeiss.com/microscopy/en_us/products/microscope-software/zen.html; RRID:SCR_013672 | |||
| Software, algorithm |
Mouse maintenance and mouse strains
All experimental procedures were conducted in accordance with the National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committee of the NYU School of Medicine and Harvard Medical School. Generation and genotyping of
Immunochemistry and imaging
Embryos, neonate, juvenile, and adult mice were perfused inter cardiac with ice cold 4% PFA after being anesthetized on ice (neonates) or using sodium pentobarbital anesthesia in adults. Brains that were processed for immunofluorescence on slides were post-fixed and cryopreserved in 30% sucrose. Sixteen µm coronal sections were obtained using Cryostat (Leica Biosystems) and collected on super-frost coated slides, then allowed to dry and stored at –20 °C until use. For immunofluorescence, cryosections were thawed and allowed to dry for 5–10 min and rinsed in 1 x PBS. They were incubated at room temperature in a blocking solution of PBST (PBS-0.1%Tx-100) and 10% normal donkey serum (NDS) for 1 hr, followed by incubation with primary antibodies in PBS-T and 1% NDS at 4 °C overnight or 2 days. Samples were then washed 4 times with PBS-T and incubated with fluorescence-conjugated secondary Alexa antibodies (Life Technologies) in PBS-T with 1% NDS at room temperature for 1 hr. Slides were incubated for 5 min with DAPI, washed three times with PBS-T. Then slides were mounted with Fluoromount G (Southern Biotech) and imaged.
Brains that were processed for free-floating immunofluorescence were first post-fixed in 4% PFA overnight at 4 °C. Fifty-µm-thickness brain slices were taken on a Leica vibratome and stored in a cryoprotecting solution (40% PBS, 30% glycerol and 30% ethylene glycol) at –20 °C. For immunofluorescence, floating sections were blocked for 1 hr at RT in normal donkey or goat serum blocking buffer and incubated for 2–3 days at 4 °C with primary antibodies in blocking buffer. Sections were washed 4x30 min at RT in PBST, incubated overnight at 4 °C with secondary antibodies and DAPI in blocking buffer, washed 4x30 min at RT in PBST before being mounted on super-frost plus glass slides. Primary antibodies are listed in Key Resource Table.
Nova1/2 localization
To quantify the Nova localization in SST cINs, mCherry+/SST cIN, KIR2.1+/SST cINs or NaChBac+/SST cIN (n=27 cells from 3 mice each); control/SST cIN or ECS+/SST cINs (n=27 cells from 3 mice each); Nova2OE/SST cIN or Nova2OE +KIR2.1/SST cINs (n=20 cells from 3 mice) were binned into two categories based on the cell compartment Nova1/2 protein was localized to: Cytoplasmic restricted or Nuclear-expressing (comprised of nuclear restricted or whole soma localization). The number of Nuclear-expressing cells was then divided by the number of cytoplasmic restricted cells to obtain a ratio for Nova localization from either mCherry+/SST cIN or KIR2.1+/SST cINs. This was collected from at least three tissue sections from at least three animals.
Electroconvulsive Shock
Electroconvulsive stimulation (ECS) was administered to animals with pulses consisting of 1.0 s, 50 Hz, 75 mA stimulus of 0.7ms delivered using the Ugo Basile ECT unit Model 57800, as previously described (Guo et al., 2011; Ma et al., 2009). Control/sham animals were similarly handled using the exact same procedure but without the current administration.
Confocal imaging and synaptic puncta analysis
Animals were perfused as described above. Post-fixation incubation prior to cryopreservation was skipped. Cryostat sections (16 μm) were subjected to IHC as described above. Images were taken within the S1 cortex of at least three different sections from at least three different animals per genotype with a Zeiss LSM 800 laser scanning confocal microscope. Scans were performed to obtain four optical Z-sections of 0.33 μm each (totaling ~1.2 μm max projection) with a 63 x/1.4 Oil DIC objective. The same scanning parameters (pinhole diameter, laser power/offset, speed/averaging) were used for all images. Maximum projections of four consecutive 0.33 μm stacks were analyzed with ImageJ (NIH) puncta analyzer plugin (Ippolito and Eroglu, 2010) to count the number of individual puncta consisting of pre-synaptic and post-synaptic markers that are close enough together to be considered a putative synaptic puncta. Synaptic puncta density per image was calculated by normalization to total puncta acquired for each individual channel accounted in each image for each condition. Puncta Analyzer plugin for ImageJ is written by Barry Wark and is available for download (https://github.com/carina-block/Puncta-analyzer/tree/v1.0; Wark et al., 2023). Nova protein intensity was performed as: Cryostat sections of 20 µm were immunostained with goat anti-mCherry and human anti-pan Nova (from Darnell Lab). Images were analyzed using Fiji/ImageJ and Nova1/2 protein intensity levels were assessed normalized against area of the cells expressing the AAV.
Electrophysiological recordings
Slice preparation
Acute brain slices (300 μm thick) were prepared from P18-P22 mice. Mice were deeply anesthetized with isofluorane. The brain was removed and placed in ice-cold modified artificial cerebrospinal fluid (ACSF) of the following composition (in mM): 87 NaCl, 26 NaHCO3, 2.5 KCl, 1.25 NaH2PO4, 0.5 CaCl, 4 MgCl2, 10 glucose, 75 sucrose saturated with 95% O2, 5% CO2 at pH = 7.4. Coronal sections were cut using a vibratome (Leica, VT 1200 S). Slices were then incubated at 34 C for 30 minutes and then stored at room temperature until use.
Recordings
Slices were transferred to the recording chamber of an up-right microscope (Zeiss Axioskop) equipped with IR DIC. Cells were visualized using a 40 X IR water immersion objective. Slices were perfused with ACSF of the following composition (in mM): 125 NaCl, 25 NaHCO3, 2.5 KCl, 1.25 NaH2PO4, 2 CaCl2, 1 MgCl2, 20 glucose, saturated with 95% O2, 5% CO2 at pH = 7.4 and maintained at a constant temperature (31 °C) using a heating chamber. Whole-cell recordings were made from randomly selected tdTomato-positive SST interneurons or tdTomato negative pyramidal cells from layer II-III or layer V of the somatosensory cortex. Miniature synaptic currents were recorded in the presence of 1 uM TTX in ACSF. Recording pipettes were pulled from borosilicate glass capillaries (Harvard Apparatus) and had a resistance of 3–5 MΩ when filled with the appropriate internal solution, as reported below. Recordings were performed using a Multiclamp 700B amplifier (Molecular Devices). The current clamp signals were filtered at 10 KHz and digitized at 40 kHz using a Digidata 1550 A and the Clampex 10 program suite (Molecular Devices). Miniature synaptic currents were filtered at 3 kHz and recorded with a sampling rate of 10 kHz. Voltage-clamp recordings were performed at a holding potential of 0 mV. Current-clamp recordings were performed at a holding potential of –70 mV. Cells were only accepted for analysis if the initial series resistance was less than 40 MΩ and did not change by more than 20% throughout the recording period. The series resistance was compensated online by at least ~60% in voltage-clamp mode. No correction was made for the junction potential between the pipette and the ACSF.
Passive and active membrane properties were recorded in current clamp mode by applying a series of hyperpolarizing and depolarizing current steps and the analysis was done in Clampfit (Molecular Devices). The cell input resistance was calculated from the peak of the voltage response to a 50 pA hyperpolarizing 1 s long current step according to Ohm’s law. Analysis of the action potential properties was done on the first spike observed during a series of depolarizing steps. Threshold was defined as the voltage at the point when the slope first exceeds a value of 20 V.s-1. Rheobase was defined as the amplitude of the first depolarizing current step at which firing was observed. Analysis of miniature inhibitory events was done using Clampfit’s template search.
Pipette solutions
Solution for voltage-clamp recordings from pyramidal cells (in mM): 125 Cs-gluconate, 2 CsCl, 10 HEPES, 1 EGTA, 4 MgATP, 0.3 Na-GTP, 8 Phosphocreatine-Tris, 1 QX-314-Cl and 0.4% biocytin, equilibrated with CsOH at pH = 7.3. Solution for current clamp recordings from SST cINs (in mM): 130 K-Gluconate, 10 KCl, 10 HEPES, 0.2 EGTA, 4 MgATP, 0.3 NaGTP, 5 Phosphocreatine and 0.4% biocytin, equilibrated with KOH CO2 to a pH = 7.3.
Optogenetic stimulation
Blue-light (470 nm) was transmitted to the slice from an LED placed under the condenser of an up-right microscope (Olympus BX50). IPSCs were elicited by applying single 1ms blue-light pulses of varying intensities (max. stimulation intensity ~0.33 mW/mm2) and directed to L2/3 or L5 of the slice in the recording chamber. Light pulses were delivered every 5 s. The LED output was driven by a TTL output from the Clampex software of the pCLAMP 9.0 program suite (Molecular Devices).
Isolation of cortical interneurons from the developing mouse cerebral cortex
Cortical interneurons were dissociated from postnatal mouse cortices (P8) as described (Wamsley et al., 2018). We collected at least 3–5 KO and 3–5 ctl brains and maintained overall balanced numbers of females and males within each condition, in order to avoid sex- related gene expression biases. Following dissociation, cortical neurons in suspension were filtered and GFP +or TdTomato + fate-mapped interneurons were sorted by fluorescence activated-cell sorting (FACS) on either a Beckman Coulter MoFlo (Cytomation), BD FACSAria II SORP or Sony SY3200. Sorted cINs were collected and lyzed in 200 µl TRIzol LS Reagent, then thoroughly mixed and stored at –80 °c until further total RNA extraction.
Nucleic acid extraction, RNA amplification, cDNA library preparation, and RNA sequencing
Total RNAs from sorted SST cINs (P8 mouse S1 cortices for Figure 2, Figure 2—figure supplement 2C, Figure 4—figure supplement 1, Figure 4—figure supplement 2 and Figure 5) were extracted using TRIzol LS Reagent and PicoPure columns (if <20 K cells were recovered) or PureLink RNA Mini Kit (if >20 K cells were recovered), with PureLink DNase for on-column treatment, following the manufacturers’ guidelines. RNA quality and quantity were measured with a Picochip using an Agilent Bioanalyzer and only samples with high quality total RNA were used (RIN: 7–10). 20 ng of total RNA was used for cDNA synthesis and amplification, using NuGEN Ovation RNA-Seq System V2 kit (NuGEN part # 7102). A total of 100 ng of amplified cDNA were used to make a library using the Ovation Ultralow Library System (NuGEN part # 0330). The samples were mulitplexed and subjected to 50-nucleotide paired-end read rapid with the Illumina HiSeq 2500 sequencer (v4 chemistry), to generate >50 million reads per sample. Library preparation, quantification, pooling, clustering and sequencing was carried out at the NYULMC Genome Technology Center. qRT-PCR (quantitative RT-PCR) was performed using SYBR select master mix (Thermo Fisher Scientific) on cDNA synthesized using SuperScript II reverse transcriptase and oligo(dT) primers.
List of RT- and qRT-PCR primers:
| Primer name | Sequence |
|---|---|
| Adam22-FAM-fw |
|
| Adam22-Rv |
|
| Anks1b-FAM-Fw |
|
| Anks1b-FAM-Fw |
|
| Sez6-FAM-Fw |
|
| Sez6-Rev |
|
| Dlg3-FAM-Fw |
|
| Dlg3-Rev |
|
| Syngap1-FAM-Fw |
|
| Syngap1-Rev |
|
| Prkrir-FAM-Fw |
|
| Prkrir--Rev |
|
| Sorbs2-FAM-Fw |
|
| Sorbs2-Rev |
|
| Nrxn1-FAM-Fw |
|
| Nrxn1-Rev |
|
| Ezh2-FAM-Fw |
|
| Ezh2-Rev |
|
| Triobp-FAM-Fw |
|
| Triobp-Rev |
|
| Itch-FAM-Fw |
|
| Itch-Rev |
|
Bioinformatics
Downstream computational analysis were performed at the NYULMC Genome Technology Center and at KAUST. All the reads were mapped to the mouse reference genome (mm10) using the STAR aligner (Dobin et al., 2013). Quality control of RNAseq libraries (i.e. the mean read insert sizes and their standard deviations) was calculated using Picard tools (v.1.126, RRID:SCR_006525) (http://broadinstitute.github.io/picard/). The Read Per Million (RPM) normalized BigWig files were generated using BEDTools (v2.17.0) (Quinlan and Hall, 2010) and bedGraphToBigWig tool (v4). For the SST cIN P8 ECS, approx. 60E6-80E6 reads were aligned per sample; for P8
We used rMATS (v3.0.9) to quantify the AS event types (i.e. Skipped exons (SE), alternative 3' splice sites (A3SS), alternative 5' splice sites (A5SS), mutually exclusive exons (MXE) and retained introns (RI)). rMATS uses a counts-based model, it detects AS events using splice junction and exon body counts and calculates an exon inclusion level value ψ for each event in each condition. It then determines the differential |∆ψ| value across conditions (cut-offs for significance were placed at FDR <0.05 and |∆ψ|≥0.1). To compare the level of similarity among the samples and their replicates, we used two methods: classical multidimensional scaling or principal-component analysis and Euclidean distance-based sample clustering. The downstream statistical analyses and generating plots were performed in Rstudio (Version 1.1.456) (http://www.r-project.org/).
To assess the enrichment for the Nova-binding motif in the differentially regulated exons we utilized rMAPS (Park et al., 2016). We utilized the raw output from rMATS analysis (6 RNAseq experiments of SST cINs +ECS vs SST cINs ctl) with significant splicing events cut off at FDR >50%. rMAPS performs position weight analysis to assess the enrichment of RNA-binding protein binding motifs in the exonic and flanking intronic regions of up-regulated or down-regulated exons and plots the motif density along with a given pValue in comparison to unregulated exons.
We performed GO analysis using the DAVID online Bioinformatics Resources 6.8 at FDR >0.05 (unless otherwise specified) (Huang et al., 2009) and tested PPI networks by utilizing DAPPLE at 10,000 permutations (Rossin et al., 2011). The GO categories were assigned to each group of genes, and after that, we used ClusterProfiler, the R function that helps with gene functional annotation and to perform GO enrichment analysis.
Validation of SST-cINs AS activity-dependent exons by RT-PCR
Total RNAs from sorted cINs from wt/ctl SST cINs, ECS SST cINs, and ECS
Quantification and statistical analysis
No statistical method was used to pre-determine sample sizes, but our sample sizes were similar to those reported in previous publications in the field. In all figures: *, p-value <0.05; **, p-value <0.01; ***, p-value <0.001; ****, p-value <0.0001. Statistical analyses for motif enrichment were performed by rMAPS and differential alternative splicing changes were performed using rMATS. Percentages were compared with repeated t-tests in
All values presented in the manuscript are average ± standard error of the mean (SEM). The statistical values for the intrinsic physiology are obtained using one-way ANOVA with Bonferroni correction for multiple comparisons between the different genotypes: Controls,
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2023, Ibrahim, Wamsley et al. This work is published under https://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
Somatostatin interneurons are the earliest born population of cortical inhibitory cells. They are crucial to support normal brain development and function; however, the mechanisms underlying their integration into nascent cortical circuitry are not well understood. In this study, we begin by demonstrating that the maturation of somatostatin interneurons in mouse somatosensory cortex is activity dependent. We then investigated the relationship between activity, alternative splicing, and synapse formation within this population. Specifically, we discovered that the Nova family of RNA-binding proteins are activity-dependent and are essential for the maturation of somatostatin interneurons, as well as their afferent and efferent connectivity. Within this population, Nova2 preferentially mediates the alternative splicing of genes required for axonal formation and synaptic function independently from its effect on gene expression. Hence, our work demonstrates that the Nova family of proteins through alternative splicing are centrally involved in coupling developmental neuronal activity to cortical circuit formation.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer