Homer1a regulates Shank3 expression and underlies behavioral vulnerability to stress in a model of Phelan-McDermid syndrome

SUMMARY

Mutations of SHANK3 cause Phelan-McDermid syndrome (PMS), and these individuals can exhibit sensitivity to stress, resulting in behavioral deterioration. Here, we examine the interaction of stress with genotype using a mouse model with face validity to PMS. In Shank3^ΔC/+ mice, swim stress produces an altered transcriptomic response in pyramidal neurons that impacts genes and pathways involved in synaptic function, signaling, and protein turnover. Homer1a, which is part of the Shank3-mGluR-N-methyl-D-aspartate (NMDA) receptor complex, is super-induced and is implicated in the stress response because stress-induced social deficits in Shank3^ΔC/+ mice are mitigated in Shank3^ΔC/+;Homer1a^−/− mice. Several lines of evidence demonstrate that Shank3 expression is regulated by Homer1a in competition with crosslinking forms of Homer, and consistent with this model, Shank3 expression and function that are reduced in Shank3^ΔC/+ mice are rescued in Shank3^ΔC/+;Homer1a^−/− mice. Studies highlight the interaction between stress and genetics and focus attention on activity-dependent changes that may contribute to pathogenesis.

Graphical Abstract

In brief

Lin et al. show that stress induces transcriptomic changes in pyramidal neurons and impairs social preferences in Shank3^ΔC/+ mice. Homer1a knockout restores Shank3 expression and functions and rescues stress-induced social deficits in Shank3^ΔC/+ mice. The results illustrate interactions between stress and genetics and highlight activity-dependent changes that may underlie pathogenesis.

INTRODUCTION

Shank3 is a multidomain postsynaptic scaffold protein that binds actin-related proteins and synaptic proteins, including SAP90/PSD95-associated protein 1 (SAPAP1; also known as GKAP1) and Homer. GKAP1 links Shank3 to N-methyl-D-aspartate receptors (NMDARs), although Homer links Shank3 with group 1 metabotropic glutamate receptors (mGluRs) (Monteiro and Feng, 2017). Mutations of SHANK3 are causal for Phelan-McDermid syndrome (PMS), an autism risk neurodevelopmental disorder resulting from heterozygous 22q13.3 deletion (Monteiro and Feng, 2017). Multiple Shank3 mouse genetic models have been analyzed and demonstrate changes in circuitry and behaviors that replicate aspects of autism spectrum disorder (ASD) (Bidinosti et al., 2016; Chen et al., 2020; Drapeau et al., 2018; Duffney et al., 2015; Han et al., 2013; Lee et al., 2021; Monteiro and Feng, 2017; Peça et al., 2011; Qin et al., 2018; Schmeisser et al., 2012; Vicidomini et al., 2017; Wang et al., 2016, 2020; Won et al., 2012). One aspect of PMS that has not been explored is sensitivity to stress.

Individuals with PMS or ASD can struggle to cope with stimuli, even those as benign as common social interaction or a novel environment (Fuld, 2018; Taylor and Corbett, 2014). ASD is associated with an increased probability of adverse childhood experience, and persons with ASD exhibit a more robust stress response with dysregulated circadian cortisol (Corbett et al., 2009; Spratt et al., 2012). Here, we examine how a disease-relevant genetic model of PMS impacts vulnerability to stress.

RESULTS

Shank3^ΔC/+ mice exhibit an altered transcriptomic response to stress

To examine the vulnerability of Shank3^ΔC/+ mice to stress, we performed a transcriptomic analysis from pyramidal neurons (Figure 1A; Sanz et al., 2009). Mice were subjected to 6 min of forced swim, and cortex was collected 1 h or 4 h later (Figure 1A). Quality control confirmed enrichment of mRNA from pyramidal neurons (Slc17; Figure S1A). Transcriptomes of Ctrl and Shank3^−ΔC/+ mice in the naive condition were not markedly different. By contrast, Ctrl and Shank3^ΔC/+ mice exhibited strikingly different transcriptomic profiles at 4 h after swim stress (Figures 1B and 1C; Table S1). Pathway analysis (gene set enrichment analysis [GSEA]) comparing stress responses in Shank3^ΔC/+ versus Ctrl mice identified 7 major pathway clusters (cutoff, q < 0.0005), including intracellular signaling, actin dynamics, proteasome degradation, and mitochondrial function (Figures 1D, 1E, and S1B). Overrepresented signaling pathways included synaptic receptors and their downstream targets as well as pathways related to transmission across chemical synapses, including those associated with glutamatergic and NMDAR functions (Figures 1D and 1E). SynGO (Koopmans et al., 2019) also confirmed enrichment of synaptic components (Figures S2A and S2B).

Figure 1. Shank3^ΔC/+ mice display altered transcriptomic response to stress.

(A) Schematic figure of RNA-seq experiment design. Cohort of Ctrl and Shank3^ΔC/+ RiboTag mice was subjected to 6 min swim stress (stressed). Naive;Ctrl, n = 4; stressed;Ctrl, n = 5; naive;Shank3^ΔC/+, n = 4; stressed;Shank3^ΔC/+, n = 6 mice.

(B) Volcano plots of transcriptomic changes in Shank3^ΔC/+ mice compared with Ctrl in naive condition (left), as well as interaction analysis of the transcriptomic effects of stress are altered in Shank3^ΔC/+ mice (right).

(C) Heatmap of the 50 most differentially expressed genes in Ctrl and Shank3^ΔC/+ mice with swim stress.

(D) GSEA pathways on swim-stress-induced transcriptome of Ctrl and Shank3^ΔC/+ mice.

(E) Pathway cluster of synaptic receptors and signaling.

(F) Volcano plot of swim-stress-induced transcriptome with SFARI genes (left) or Shank3 complex genes (right) highlighted.

(G) Representative RNA-seq track at Homer1 loci.

(H) Ratio of Homer1a to Homer1 transcripts in RNA-seq at 4 h post-swim stress. Statistics were determined by two-way ANOVA followed by Fisher’s least significant difference (LSD) test. Naive;Ctrl versus stressed;Ctrl, **p = 0.0025; naive;Shank3^ΔC/+ versus stressed;Shank3^ΔC/+, ****p < 0.0001; stressed;Ctrl versus stressed;Shank3^ΔC/+, ****p < 0.0001.

(I and J) Representative blots (I) and statistical analysis (J) of Homer1a levels in the cortical lysate of WT and Shank3^ΔC/+ mice in 4 h after 6 min swim stress. Statistics was determined by two-way ANOVA followed by Bonferroni’s post hoc test; stress effect, *p = 0.0213; naive;Shank3^ΔC/+ versus stressed;Shank3^ΔC/+, *p = 0.0248. At least n = 5 mice per group.

To evaluate the connection between stress-induced transcriptomic changes and ASD phenotypes that are related to Shank3^ΔC/+ animals, we examined Simon Foundation Autism Research Initiative (SFARI) gene networks associated with ASD (a database of genes implicated in autism susceptibility; https://gene.sfari.org/) as well as genes within the Shank3 complex, determined by in vivo data. Among these gene sets, Homer1 exhibited dramatically different stress-induced expression (Figure 1F).

Homer1a is super-induced by swim stress in Shank3^ΔC/+ mice

The Homer1 gene encodes long and short transcripts. All transcripts encode an Ena/Vasp Homology 1 (EVH1) domain that binds postsynaptic proteins, including Shank3, the inositol 1,4,5-trisphosphate receptors (IP₃Rs), and mGluR5 (Brakeman et al., 1997; Hayashi et al., 2009; Tu et al., 1999). The long forms, Homer1b/c, also contain a coil-coiled domain that self-oligomerizes and crosslinks protein complexes in the post-synaptic density (PSD) (Hayashi et al., 2009; Tu et al., 1999). The activity-induced short form, Homer1a, lacks the coil-coiled domain and disrupts crosslinking (Park et al., 2013; Sala et al., 2003; Tu et al., 1999; Zeng et al., 2018). Homer1a was not identified in the initial analysis because this did not distinguish splice forms. Accordingly, we examined traces of our dataset on genome browser and found that transcripts within the 5^th intron of the Homer1 gene were induced by stress in both Ctrl and Shank3^ΔC/+ (Figures 1G and S3A). Salmon transcript analysis revealed the stress-induced Homer1a expression in Shank3^ΔC/+ mice was larger than that in Ctrl mice at 4 h, but not 1 h post-swim stress (Figures 1G, 1H, S3A, and S3B). Western blot confirmed that Homer1a was robustly induced by stress in Shank3^ΔC/+ mice, but not in wildtype (WT) mice (Figures 1I and 1J). Other immediate early genes (IEGs), including Arc, Egr1, and Fos, were similarly increased in Ctrl and Shank3^ΔC/+, indicating that the super-induction of Homer1a is unique among IEGs (Figures S3C and S3D).

Shank3^ΔC/+ mice display a stress-induced social behavioral deficit linked to Homer1a

We reasoned that the altered transcriptomic response to stress represents adaptations caused by Shank3^ΔC/+ that could either preserve neuronal function or could be maladaptive. To begin to address this issue, we asked how stress might impact behavioral domains and focused on social interaction because this is known to be altered in Shank3^ΔC/+ mice (Duffney et al., 2015; Qin et al., 2018). Expectedly, deficits in naive Shank3^ΔC/+ mice were readily detectable in a social recognition task (Figures S4A-S4C). However, because deficits in the naive state might obscure effects of stress, we used a social motivation task in which performance of naive Shank3^ΔC/+ mice was not different from WT (Figures S4D-S4F). To alleviate any possible bias and increase relevance to other studies, we monitored three variables to characterize social motivation (Figure 2A). We found that Shank3^ΔC/+ mice displayed significant deficits in social preferences 4 h after a swim stress, as revealed in side investigation and average proximity (Figures 2B-2E). To gain a more comprehensive profile of stress-induced social preferences beyond conventional p values (Betensky, 2019; Johnson, 2013; Wasserstein and Lazar, 2016), we calculated Hedges’ g coefficients of effect size (Brydges, 2019; Lakens, 2013). Effect sizes increased in WT mice but reduced in Shank3^ΔC/+ mice in all three variables by stress, indicating a stress-induced reduction of social preference in Shank3^ΔC/+ mice (Figure 2F).

Figure 2. Homer1a knockout restores stress sensitivity of Shank3^ΔC/+ mice in social motivation task.

(A) Scheme of social motivation task comparing exploration of social (a juvenile mouse, “pup”) and the non-social object (an empty “cup”) and variables used to characterize social preferences.

(B–D) Performance in social motivation task comparing naive WT (n = 12 mice), Shank3^ΔC/+ (n = 13mice), Homer1a^−/− (n = 10 mice), and Shank3^ΔC/+; Homer1a^−/− (n = 12 mice) versus stressed WT (n = 13mice), Shank3^ΔC/+ (n = 18 mice), Homer1a^−/− (n = 18 mice), and Shank3^ΔC/+;Homer1a^−/− (n = 15 mice). Statistics were determined by two-way ANOVAs followed by Tukey’s post hoc tests; interaction, side/Shank3^ΔC/+, #p < 0.05; post hoc test, *p < 0.05 and **p < 0.01.

(E and F) Summary of statistical significance (E; p value, shown as −log₁₀) and effect sizes (F; Hedges’ g ± confidence intervals) for cup versus pup comparisons for variables of social motivation task shown in (C)–(E).

(G) Plasma corticosterone levels in naive WT (n = 3), Shank3^ΔC/+ (n = 4), Homer1a^−/− (n = 5), and Shank3^ΔC/+;Homer1a^−/− (n = 6) and stressed WT (n = 5), Shank3^ΔC/+ (n = 4), Homer1a^−/− (n = 2), and Shank3^ΔC/+;Homer1a^−/− (n = 6) mice. Statistics were determined by a two-way ANOVA followed by LSDs post hoc tests; **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Using this behavioral paradigm, we examined the interaction of Homer1a and Shank3^ΔC/+ and found that Shank3^ΔC/+;Homer1a^−/− mice showed normal social preferences in the naive condition and maintained their preference to social objects following swim stress. Genetic rescue of the stress response was remarkable because Homer1a^−/− mice were deficient in social preferences in the naive condition (Figures 2B-2F). Moreover, motor activity significantly increased after stress in Homer1a^−/− mice (Figures S4G and S4H), but not in Shank3^ΔC/+;Homer1a^−/− mice. Thus, Shank3^ΔC/+ rescues behaviors in Homer1a^−/− mice, providing further evidence for convergence of these pathways. Stress induced a robust increase of corticosterone in all genotypes, indicating that the behavioral restoration in Shank3^ΔC/+;Homer1a^−/− mice did not result from blunted stress responses (Figure 2G).

Homer 1a^−/− rescues Shank3 expression and synaptic function

We asked how Homer1a^−/− might rescue behavior in Shank3^ΔC/+. Shank3 is enriched in Triton X-100 insoluble P2 fractions (PSD-enriched fractions; Figures S5A and S5B). Shank3 expression was reduced in the PSD-enriched fraction of Shank3^ΔC/+ mice consistent with haplo-insufficiency, although Shank3 expression in Shank3^ΔC/+;Homer1a^−/− mice was increased relative to Shank3^ΔC/+. Levels of PSD95, SynGAP1, and Homer1b/c in PSD-enriched fractions were not significantly altered in Shank3^ΔC/+ or Shank3^ΔC/+;Homer1a^−/− mice, indicating selective effects of the genetic manipulations (Figures 3A and 3B).

Figure 3. Homer1a knockout restores Shank3 abundance and synaptic physiology in Shank3^ΔC/+.

(A and B) Representative blots (A) and quantification (B) of Shank3, SynGAP1, PSD95, and Homer1b/c in the PSD-enriched fractions. Statistics were determined by one-way ANOVA followed by Tukey’s post hoc test; Shank3, WT versus Shank3^ΔC/+, ****p < 0.0001; WT versus Shank3^ΔC/+; Homer1a^−/−, **p = 0.0095; Shank3^ΔC/+ versus Shank3^ΔC/+;Homer1a^−/−, *p = 0.0127. WT, n = 7; Shank3^ΔC/+, n = 11; and Shank3^ΔC/+;Homer1a^−/−, n = 9 mice.

(C) Representative (gray) and averaged (black) traces from a series of 20 consecutive EPSCs in cortical slices.

(D) Average NMDA/AMPA ratio in WT (n = 20), Shank3^ΔC/+ (n = 19), Homer1a^−/− (n = 8), and Shank3^ΔC/+;Homer1a^−/− (n = 21) neurons. **p < 0.01, ***p < 0.001, and ****p < 0.0001 were determined by one-way ANOVA followed by Bonferroni’s post hoc test. Scale bar, 100 pA/50 mS.

(E and F) Representative traces (E) and time course (F) of the change in field recording of excitatory postsynaptic potential (fEPSP) slope induced by high-frequency stimulation. WT, n = 9; Shank3^ΔC/+, n = 10; Homer1a^−/−, n = 6; and Shank3^ΔC/+; Homer1a^−/−, n = 10 slices. Statistics were determined by two-way ANOVA followed by Bonferroni’s post hoc test. WT versus Shank3^ΔC/+, *p = 0.0155; Shank3^ΔC/+ versus Shank3^ΔC/+; Homer1a^−/−, ***p = 0.005.

NMDAR function is dependent on Shank3 (Duffney et al., 2015; Peça et al., 2011; Schmeisser et al., 2012; Won et al., 2012). Accordingly, we monitored NMDAR synaptic physiology by examining the Schaeffer-CA1 pyramidal neuron synapse using acute hippocampal slices. We first measured α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor (AMPAR)-mediated excitatory postsynaptic current (EPSC) at −70 mV and then, in the presence of AMPAR antagonist 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(f)quinoxaline (NBQX), voltage clamped the neurons at +40 mV to measure NMDAR-mediated EPSCs. Shank3^ΔC/+ mice exhibited a reduction of the NMDA/AMPA ratio that was restored in Shank3^ΔC/+;Homer1a^−/− mice to near WT (Figures 3C and 3D), without altering membrane resistance, Rm, or series resistance, Rs (Figures S5C and S5D). NMDAR function was further assessed by the ability of high-frequency synaptic stimulation to induce long-term potentiation (LTP). NMDAR-LTP was reduced in Shank3^ΔC/+ and restored in Shank3^ΔC/+;Homer1a^−/− mice (Figures 3E and 3F).

Shank3 expression in vivo is controlled by Homer crosslinking in competition with Homer1a

To assess how Homer1a^−/− might rescue Shank3 expression and function, we began by examining the limiting condition of combined Homer1^−/−;Homer2^−/−; Homer3^−/− (HTKO). Shank3 expression in the forebrain lysate of HTKO mice is reduced by −80%–90% in association with reductions of NMDA receptor subunits GluN1 and GluN2A (Figures 4A and 4B). By contrast, in Homer1a^−/− mice, Shank3 and Homer1c were increased without significant changes in GluN1 or GluN2A (Figures 4C and 4D). To examine the effect of Homer crosslinking on Shank3 expression at synapses, we performed immunostaining assays in primary cortical neuron cultures. Immunostaining for Shank3 revealed a near absence of Shank3-positive puncta co-localizing with presynaptic marker vGlut1 in cultured primary neurons from HTKO (Figures S6A and S6B). Synaptic Shank3 expression was restored by expression of WT Homer1c transgene, but not by Homer1c transgenes containing point mutations that selectively interrupt EVH1 binding (W24A or G89N) necessary for crosslinking (Figures 4E-4G). Reciprocally, overexpression of Homer1a in cultured WT cortical neurons reduced Shank3 puncta along dendrites (Figures 4H-4J).

Figure 4. Homer crosslinking mediates Shank3 abundance in vivo.

(A and B) Representative blots (A) and quantification (B) of Shank3, GluN1, and GluN2A levels in the total lysate of Homer triple knockout (HTKO) brain lysate. *p < 0.05 and **p < 0.01 were determined by unpaired t test. n = 3 mice per group.

(C and D) Representative blots (C) and quantification (D) of Shank3, Homer1, GluN1, and GluN2A levels in the total lysate of Homer1a^−/− brain lysate. *p < 0.05 and ***p < 0.001 were determined by unpaired t test; WT, n = 8 and Homer1a^−/−, n = 10 mice.

(E) Schematic figure showing domains and point mutants for Homer1.

(F and G) Representative images (F) and quantification (G) of Shank3 puncta in primary cortical cultures from HTKO mice with overexpressed Homer1 WT, W24A, or G89N mutants. Statistics were determined by one-way ANOVA followed by Bonferroni’s post hoc test; WT versus W24A, *p = 0.0127; WT versus G89N, **p = 0.0067; at least n = 4 neurons per group from 3 different pups. Scale bar, 5 μm.

(H–J) Representative images (H) and quantification (I and J) of the effect of overexpressing HA-Homer1a (1) on Shank3 levels in primary neurons compared to nontransfected neuron (2). Statistics were determined by Mann-Whitney test (I), p = 0.1299, and Kolmogorov-Smirnov test (J), ****p < 0.0001. At least n = 3 neurons per group, and at least 10 puncta were analyzed per neuron. Scale bar, 5 μm.

(K) Schematic figure of sleep deprivation (SD) and recovery (R). Mice were subjected to either 4 h sleep deprivation (SD, blue) or 4 h sleep deprivation plus 2.5 h recovery (SD/R, purple). Control animals were housed in home cages for either 4 h or 6.5 h.

(L and M) Representative blots (L) and quantification (M) of Shank3, mGluR5, Homer1, and Homer1a levels in the cortical lysates of mice after SD and R. Statistics were determined by one-way ANOVA followed by Bonferroni’s post hoc test; Homer1a, Ctrl(4 h) versus SD(4 h), ***p = 0.0008; Shank3, Ctrl(4 h) versus SD(4 h), *p = 0.0288; at least n = 7 mice per group.

Dynamic balance of Shank3 and Homer1a with sleep homeostasis

To assess the general relevance of Homer regulation of Shank3 expression, we examined a natural model in which Homer1a is upregulated as a consequence of sleep deprivation (SD) and monitored associated changes in Shank3 expression. As reported, Homer1a was robustly upregulated by 4 h SD. In these same lysates, Shank3 was significantly reduced (Figures 4K-4M). Following SD, Homer1a returned to baseline levels after subsequent 2.5 h recovery sleep (SD/R), and this was paralleled by return of Shank3 levels to baseline. mGluR5 and Homer1c levels were not altered with SD or SD/R, indicating that changes were selective for Homer1a and Shank3 (Figures 4K-4M).

DISCUSSION

The present study reveals an interaction of genetics and stress that alters pyramidal neuronal gene expression in a mouse model with face validity to PMS. The altered transcriptomic response in Shank3^ΔC/+ mice includes genes involved in synaptic function, cell signaling, actin dynamics, and proteasome function and genes implicated in ASD (SFARI) and the “Shank complex.” Consistent with an impact on synaptic function, stress induces a de novo social motivation deficit in Shank3^ΔC/+ mice. Shank3^ΔC/+ mice exhibit multiple behavioral phenotypes (Duffney et al., 2015; Kloth et al., 2015; Matas et al., 2021), and our approach selected a social behavioral phenotype that emerged with stress and thereby afforded an opportunity to examine genetic modifiers and molecular mechanisms. We note that phenotype analysis in neurogenetic models is typically performed in naive mice, which may miss important experience-dependent changes. Indeed, pathway analysis of the naive state would not have implicated the Shank pathway, providing precedent for omics studies seeking to identify critical genes and pathways related to cognitive disease. Stress features prominently in the experience of individuals with ASD, and its effect will be important to incorporate into an understanding of basic pathophysiology and clinical presentations.

Do changes in the transcriptional response to stress in Shank3^ΔC/+ mice contribute to neuronal health and function or do they contribute to disease pathogenesis? Homer1a is identified as one of the most amplified of the stress-induced genes in Shank3^ΔC/+ pyramidal neuron transcriptome and encompasses SFARI- and Shank3-related gene sets. Our studies demonstrate that genetic deletion of Homer1a in Shank3^ΔC/+ rescues Shank3 expression and function and prevents a stress-induced social motivation deficit. Rescue is consistent with a model in which Shank3 expression and synaptic localization are dependent on Homer crosslinking. Homer1a expression may contribute to other neurodevelopmental diseases because deletion of Homer1a can rescue biochemical and behavioral domains in mouse models of fragile X syndrome (Aloisi et al., 2017; McBride et al., 2005; Ronesi et al., 2012).

The model of Homer crosslinking as a regulator of Shank3 expression appears broadly relevant and rationalizes dynamic downregulation of Shank3 in association with Homer1a upregulation and sleep deprivation in WT animals. Actions of Homer1a at other targets include agonist-independent activation of mGluR1/5 (Ango et al., 2001), uncoupling of mGluR1/5 from intracellular store channels (Tu et al., 1999), and gating of transient receptor potential canonical (TRPC) channels (Yuan et al., 2003). Accordingly, downregulation of Shank3 in response to Homer1a appears to be a distinct and selective response that likely engages protein turnover pathways and contributes to synapse homeostasis and plasticity.

Why is Homer1a super-inducible in Shank3^ΔC/+? Several reports suggest an overall increase in neuronal activity in Shank3-deficient mice (Chen et al., 2020; Wang et al., 2016). However, in vivo recordings from freely behaving Shank3^−/− mice do not reveal evidence of increased neuronal activity (Tatavarty et al., 2020). Moreover, if Homer1a super-inducibility was simply related to increased activity, other IEGs would be expected to be similarly super-induced. Shank3-deficient mice reportedly express increased levels of histone deacetylases and histone methyltransferases, and chemical or genetic inhibition of these epigenetic regulators can rescue social deficits (Qin et al., 2018; Wang et al., 2020). In addition to chromatin changes, we note a current model suggests that Homer1a is induced as a consequence of dynamic depletion of U1 small nuclear ribonucleoprotein particle (snRNP), a splice factor critical for 5⁰ splice site recognition (Berg et al., 2012). Additional studies are required.

Sociability deficits are linked to dysfunction of NMDAR-dependent synaptic plasticity in the corticostriatal circuit and implicated in mouse ASD models (Bentea et al., 2020; Lin et al., 2021; Peça et al., 2011; Won et al., 2012). Our data support an association between reduced synaptic NMDAR function and vulnerability to stress in Shank3^ΔC/+ mice because this is mitigated by Homer1a^−/− and restoration of NMDAR function. Glutamatergic and NMDAR functions are among the overrepresented pathways induced by stress in Shank3^ΔC/+ mice, suggesting this axis is particularly vulnerable to environmental stress conditions. Although these associations are intriguing, the complex of behavioral deficits in Shank3^ΔC/+ mice suggests that a detailed understanding of mechanisms relating stress, Shank3, NMDARs, and behavior will require circuit-specific analyses.

Limitations of the study

Our study focuses on the genetic interaction of Shank3^ΔC/+ and Homer1a^−/− and motivates future studies that would continuously monitor Shank3 expression at defined synapses and relate this to specific behaviors predicting that superinduction of Homer1a might result in greater reductions of Shank3 in Shank3^ΔC/+ mice and associated behavioral deficits. Alternatively, the action of Homer1a^−/− to rescue Shank3^ΔC/+ phenotypes might represent a cumulative developmental action because Homer1a is critical for normal brain development (Chokshi et al., 2019) and synaptic plasticity (Diering et al., 2017; Hu et al., 2010; Park et al., 2013).

STAR★METHODS

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animals

All procedures involving animals were approved and under the guidelines of The Johns Hopkins University Institutional Animal Care and Use Committee. Shank3^ΔC/+ mice (deposited to JAX, #018398) were generated on a mixed 129/C57BL6J strain background and later backcrossed to a C57BL6JL stain background for at least 8 generations. For transcriptomic (ribotag) studies of pyramidal neurons, Shank3^ΔC/+ mice were crossed with homozygous HA-Rpl22^f/f;Camk2a-Cre^+/+ mice on C57BL6JL stain background to receive hemizygous mice with mutated Shank3 allele, Shank3^ΔC/+;HA-Rpl22^f/+;Camk2a-Cre^+/−. The controls were littermates from the above breeding that lacked mutated Shank3, Shank3^+/+;HA-Rpl22^f/+;Camk2a-Cre^+/−, referred to as controls (Ctrls). For studies of Shank3^ΔC/+ and Homer1a interactions, Shank3^ΔC/+ mice were brought to homozygous Homer1a knockout (Homer1a^−/−) background through crosses of heterozygous Homer1a knockout (Homer1a^+/−) and Shank3^ΔC/+ mice. Homer triple knockout mice (HTKO) was generated by crossing Homer1^−/− (JAX, #023312) mice with Homer2^−/− (JAX, # 023313) and Homer3^−/− (JAX # 023314).

Adult 3-5 month-old male mice were used in all studies. They were housed in groups of three to five littermates in standard individually ventilated cages with ad libitum access to standard chow and water. Mice were kept on a 12 h light/12 h dark cycle with lights on at 7:00 AM. For sleep deprivation/recovery studies, mice were placed separately in a clean mouse cage during light phase and remained there for an additional 4hrs with gentle tapping of the cage or by disturbing the bedding material. Mice were then sacrificed for tissue collection or returned to their home cage for recovery sleep for an additional 2.5hrs. Control mice were undisturbed at their home cage and brain tissues were collected 4hrs or 6.5hrs later. For swim stress studies, behaviorally naive mice were placed in a container (40 cm x 20 cm in diameter) half-filled with water at 25°C. The forced swim stress procedure lasted 6 min, and then mice were returned to their home cages. Stress-naive mice were left in home cages undisturbed. After a four-hour delay, separate cohorts of mice were used in behavioral studies or sacrificed for brain and blood samples collection.

METHOD DETAILS

RiboTag Pulldown

Immunoprecipitation and purification of ribosome-associated RNA from pyramidal neurons was performed on Shank3^ΔC/+;HA-Rpl22^f/+;Camk2a-Cre^+/− and Ctrl mouse brains as described (Sanz et al., 2009) with minor modifications. Briefly, mouse corti were homogenized in lysis buffer (50 mM Tris, 100 mM KCl, 12 mM MgCl₂, 1% NP-40, 100 μg/ml cycloheximide, 10 mM Ribonucleoside Vanadyl Complex, 1 mM DTT, plus RNAsin and Complete^™ EDTA-free protease inhibitor mixture, pH 7.4) on ice. Homogenates were spun at 10, 000 g for 10 min to remove debris. After collecting the supernatants, ribosomes and associated RNA were immunoprecipitated with an anti-HA antibody. Ribosome-associated RNA was purified by RNeasy mini kit (QIAGEN), and then subjected to RNA sequencing or quantitative PCR as described in “RNA Extraction, cDNA synthesis and Quantitative PCR.”

RNA sequencing and analysis

RNA-seq libraries were prepared using Illumina TruSeq stranded mRNA library preparation kit (Illumina). Library quality was validated by high sensitivity DNA analysis kit using Agilent Bioanalyzer 2100, and the concentration of RNA-seq libraries was quantified by qPCR using the KAPA library quantification kit (Kapa Biosystems). Sequencing of RNA-Seq libraries was performed on a HiSeq 2500 instrument (Illumina) with 100 bp (x2) reads according to manufacturer’s instructions. 36 libraries with distinct barcodes were sequenced in 6 lanes (3 flow cells). The depth of RNA-seq was 15 million reads per sample.

Trimmomatic 0.39 (http://www.usadellab.org/cms/?page=trimmomatic) was used to trim adapters and low-quality reads from the raw paired-end fastq files, using default parameters. Next, FastQC analysis (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was used to determine the quality of each file. After quality control, the reads were aligned to mouse reference genome mm10 using the HISAT2 (2.2.0) (https://daehwankimlab.github.io/hisat2/) splice-mapping algorithm with default parameters. Read coverage tables for each gene were created using StringTie (2.1.3) (https://ccb.jhu.edu/software/stringtie/) and read coverage tables for each transcript were created using Salmon (1.4) (https://salmon.readthedocs.io/en/latest/salmon.html). To determine normalized gene counts and differential gene or transcript expression, DESeq2 was used (1.12.3) (https://www.rdocumentation.org/packages/ DESeq2/versions/1.12.3). To find genes for which swim stress-induced gene expression change was significantly different between the genotypes, an interaction term was added to the DESeq2 analysis design parameter. For all analyses, hits with an FDR (q-value) less than 0.05 and log2fold change greater than 1 were considered significant using the Benjamini-Hochberg method for multiple testing.

Following bioinformatics analysis with DESeq2, all genes were ranked by adjusted p value and the ranked list was used for Gene Set Enrichment Analysis (GSEA, http://www.gsea-msigdb.org/gsea/index.jsp) or SynGO (https://syngoportal.org) to investigate significantly enriched gene ontologies and pathways. Heatmaps were constructed based on abundance measurements from htseq-count (https://htseq.readthedocs.io/en/release_0.11.1/count.html). Heatmaps were generated using Morpheus software (https://software.broadinstitute.org/morpheus).

Biochemical Fractionation and Western Blots

Brain tissues were lysed with a modified RIPA buffer containing 1% Triton, 0.5% Na-deoxycholate, 0.1% SDS, 50 mM NaF, 10 mM Na₄P₂O₇, 2 mM Na₃VO₄, and protease inhibitor cocktail (Roche, Switzerland) in PBS, pH 7.4. For extracting the Triton X-100 resistant and soluble fraction, mouse brain tissue was homogenized in 20X ice-cold homogenization solution (320mM sucrose, 10mM HEPES pH 7.4, 1mM EDTA, 5mM Na pyrophosphate, 1mM Na3VO4, and protease inhibitor cocktail (Roche, Switzerland)). After 10 min centrifugation at 800 x g, the supernatant (S1) was further centrifuged at 16,000xg for 20min at 4°C to obtain the P2 fraction. The P2 fraction was solubilized in 1% TX-100 buffer containing 50mM Tris (pH 7.4), 150mM NaCl, 1mM EDTA, 1% TX-100 plus protease and phosphatase inhibitor at 4°C for 20 min and centrifuge at max speed (benchtop) for 30min. The supernatant was Triton soluble fraction, and the pellet was Triton insoluble fraction. Since Triton insoluble fraction contains major postsynaptic density (PSD) proteins, we termed it PSD enriched fraction. Protein extracts were separated by 4%–12% SDS-PAGE, transferred to PVDF membranes, blocked with 5% non-fat milk, and then probed with primary antibodies overnight at 4°C. After washes with TBST (TBS with 0.1% Tween-20), membranes were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature (RT). Immunoreactive bands were visualized by the enhanced chemiluminescent substrate (ECL, Pierce), read and quantified using a ChemiDoc Imaging System (BioRad). Actin or GAPDH was used as a loading control. A complete list of antibodies used is available in Key resources table.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal antibody to Homer1a	Hu et al., 2010	N/A
Rabbit polyclonal antibody to Shank3, C-terminal	this paper	N/A
Mouse Anti-Actin, beta Monoclonal Antibody	Abcam	Cat# ab8226, RRID:AB_306371
Rabbit polyclonal antibody to Shank3, N-terminal	this paper	N/A
Anti-GluR1-NT (NT) Antibody	Millipore	Cat# MAB2263, RRID:AB_11212678
Mouse Anti-NMDAR2A Monoclonal Antibody	BD Biosciences	Cat# 612286, RRID:AB_399603
Anti-NR1 Antibody, CT	Millipore	Cat# 05-432, RRID:AB_390129
Rabbit polyclonal antibody to Homer	Hu et al., 2010	N/A
Mouse Anti-Vesicular Glutamate Transporter 1 (VGLUT1)	Millipore	Cat# MAB5502, RRID:AB_262185
Rabbit Anti-mGluR5 Monoclonal Antibody	Abcam	Cat# 2237-1, RRID:AB_1267242
Rabbit Anti-PSD95 Monoclonal Antibody	Cell Signaling Technology	Cat# 3409, RRID:AB_1264242
SynGAP (D20C7) Rabbit mAb antibody	Cell Signaling Technology	Cat# 5539, RRID:AB_10694401
Homer (D-3) antibody (long form Homer1)	Santa Cruz Biotechnology	Cat# sc-17842, RRID:AB_627742
GAPDH antibody	Abcam	Cat# ab9483, RRID:AB_307273
Mouse Anti-Human HA Monoclonal Antibody	Sigma-Aldrich	Cat# H9658, RRID:AB_260092
c-Myc Antibody (9E10)	Santa Cruz Biotechnology	Cat# sc-40 AC, RRID:AB_2857941
Chemicals, peptides, and recombinant proteins
SR 95531 hydrobromide, GABAzine	Tocris	Cat#1262, CAS 104104-50-9
NBQX	Tocris	Cat# 0373, CAS 118876-58-7
Critical commercial assays
RNeasy mini kit	QIAGEN	Cat#74104
TruSeq® Stranded mRNA Library Prep	Illumina	Cat#20020594
KAPA Library Quantification Kits	Roche	Cat#07960140001
Deposited data
Raw and analyzed data	this paper	GEO: GSE174111
Experimental models: organisms/strains
Mouse: Shank3tm1.1Pfw/J	this paper, also deposited to The Jackson Laboratory	Cat# JAX:018398, RRID:IMSR_JAX:018398
Mouse: B6J.129(Cg)-Rpl22tm1.1Psam/SjJ	The Jackson Laboratory	Cat# JAX:029977, RRID:IMSR_JAX:029977
Mouse: Tg(Camk2a-cre)2Gsc	MGI	Cat# 4457404, RRID:MGI:4457404
Mouse: B6N.129(Cg)-Homer1tm1Mhd/PfwJ	The Jackson Laboratory	Cat# JAX:023312, RRID:IMSR_JAX:023312
Mouse: B6.129(Cg)-Homer2tm1 Mhd/PfwJ	The Jackson Laboratory	Cat# JAX:023313, RRID:IMSR_JAX:023313
Mouse: Homer3tm1Mhd/PfwJ	The Jackson Laboratory	Cat# JAX:023314, RRID:IMSR_JAX:023314
Mouse: Homer1a KO	Hu et al., 2010	N/A
Oligonucleotides
See Table S2 for oligonucleotide information		N/A
Recombinant DNA
pRK5-Myc-Homer1c(WT)	Hu et al., 2012	N/A
pRK5-Myc-Homer1c(W24A)	Hu et al., 2012	N/A
pRK5-Myc-Homer1c(G89N)	Hu et al., 2012	N/A
pRK5-HA-Homer1a	Hu et al., 2012	N/A
Software and algorithms
ImageJ	https://imagej.nih.gov/ij/	RRID:SCR_003070
Graph Pad Prism	https://www.graphpad.com/scientific-software/prism/	RRID:SCR_002798
Statistica	https://www.tibco.com/	RRID:SCR_014213
ANY-maze	https://sandiegoinstruments.com/product/any-maze/	RRID:SCR_014289
ZEN	https://www.zeiss.com/microscopy/us/products/microscope-software/zen-lite.html	RRID:SCR_013672
pClamp	https://www.moleculardevices.com/	RRID:SCR_011323
FastQC	https://www.bioinformatics.babraham.ac.uk/projects/fastqc/	RRID:SCR_014583
HISAT2 (2.2.0)	https://daehwankimlab.github.io/hisat2/	RRID:SCR_015530
StringTie (2.1.3)	https://ccb.jhu.edu/software/stringtie/	RRID:SCR_016323
htseq-count	https://htseq.readthedocs.io/en/release_0.11.1/count.html	RRID:SCR_011867
Salmon (1.4)	https://salmon.readthedocs.io/en/latest/salmon.html	RRID:SCR_017036
DESeq2 (1.12.3)	https://www.rdocumentation.org/packages/DESeq2/versions/1.12.3	RRID:SCR_015687
Gene Set Enrichment Analysis (GSEA)	http://www.gsea-msigdb.org/gsea/index.jsp	RRID:SCR_003199
EnrichmentMap	https://www.baderlab.org/Software/EnrichmentMap	RRID:SCR_016052
Cytoscape (3.8)	https://cytoscape.org/	RRID:SCR_003032
GeneMANIA	https://genemania.org/	RRID:SCR_005709

Immunoprecipitation

Mouse cortex was homogenized in 1% Triton X-100 immunoprecipitation buffer (PBS with 1% Triton X-100, PhosSTOP (Roche, Switzerland), and protease inhibitor cocktail (Roche, Switzerland). The homogenate was centrifuged at 12000 x g for 10 min. The supernatant was used as input for immunoprecipitation by rabbit anti-Homer antibody (Santa Cruz Biotechnology, CA) and Protein G Dynabeads (Thermo Fisher Scientific, MA). The eluate was separated by SDS-PAGE gels and the co-immunoprecipitated proteins were analyzed by Western Blotting.

RNA Extraction, cDNA synthesis and Quantitative PCR

Each cortex was homogenized in 1 mL TRIzol (Thermo Fisher Scientific, MA) on ice and 0.2 mL of chloroform was added. The mixture was centrifuged at 12000 g for 15 min at 4°C. The colorless upper aqueous layer containing the RNA was transferred to a new tube and 0.5 mL of isopropanol was added. After a 10 min incubation at room temperature, the mixture was centrifuged at 12000 g at 4°C for 10 min. The pellet was washed with 1 mL 75% ethanol and centrifuged for 5 min at 7500 g at 4°C. The RNA pellet was air-dried and resolved in RNase-free water. The quantity and quality of RNA was measured by Nanodrop (Thermo Fisher Scientific, MA). RNA was reverse transcribed using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, CA). Quantitative PCR was performed with QuantStudio 6 Flex system (Applied Biosystems) using the SYBR green ROX qPCR mastermix in a 384-well optical plate. PCR cycling conditions were as follows: 95°C for 10 min, followed by 40 cycles of 95°C for 30 s, 60°C for 30 s and 72° C for 30 s. A melt curve step was included to determine the specificity of PCR amplification. Gapdh was used as an internal control. A complete list of DNA primer sequences used is available in Key resources table.

Immunocytochemistry

Primary neurons were cultured from pups of embryonic day 15-17. Mouse cortex and hippocampus were dissociated by papain and were plated at poly-L-lysine coated coverslip for imaging. At day 12 post plating, neurons were transfected with plasmid by Lipofectamine 2000 (Thermo Fisher Scientific, MA) and were fixed by paraformaldehyde after 48 h of expression. The fixed neurons were permeabilized by 0.1% Triton X-100 for 10 min and were blocked by 5% normal goat serum for 1 h at 4°C. Primary antibodies were diluted in 5% normal goat serum (blocking reagent) and incubated with neurons overnight at at 4°C. After excessive wash, secondary antibodies were incubated with neurons for 1 h at room temperature. Antibodies used in this experiment are listed in Key resources table. Images were taken by laser scanning confocal microscope (LSM 510 or 800, Carl Zeiss AG, Germany) and were analyzed by ImageJ Fiji or MetaMorph (Molecular Devices, CA) from neurons of at least 3 different embryos.

Brain slice preparation for electrophysiology

Transverse brain slices of the hippocampus were prepared in ice-cold dissection buffer (110 choline chloride, 2.5 KCl, 7 MgCl₂, 0.5 CaCl₂, 2.4 Na-pyruvate, 1.3 Na-ascorbic acid, 1.2 NaH₂PO₄, 25 NaHCO₃, and 20 glucose). Slices were recovered for 3 h at room temperature in ACSF: 124 mM NaCl, 5 mM KCl, 1.25 mM NaH₂PO₄, 26 mM NaHCO₃, 1 mM MgCl₂, 2 mM CaCl₂, and 20 mM D-glucose) saturated with 95% O2, 5% CO2.

Electrophysiology

NMDA/AMPA ratio

Whole cell patch-clamp recordings of CA1 pyramidal neurons were performed in the presence of GABA_A receptor antagonist (10 μM, GABAzine). The pipette solution contained (in mM): 90 Cs-methansulfonate, 48.5 CsCl, 5 EGTA, 2 MgCl₂, 2 Na-ATP, 0.4 Na-GTP, 1 QX 314 bromide, and 5 HEPES (pH 7.2, 290 ± 5 mmol/kg). For measurement of the NMDA/AMPA ratio, cells were voltage-clamped at −70 mV (Vh = −70 mV) for peak AMPA current analysis. Then NMDA currents were recorded at Vh = +40 mV in the presence of the selective AMPAR antagonist 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(f)quinoxaline (NBQX, 10 μM, Tocris). Data were acquired ~3 min after achieving the whole-cell configuration. Series resistances (Rs) of recordings ranged between 8 and 15 MΩ. Cells were rejected from analysis if Rs changed by more than 15%.

In whole-cell patch-configuration, the micropipette is filled with cesium-based internal solution, which prevents current leakage and thereby provides more accurate excitatory postsynaptic current measurements. To measure passive biophysical properties of CA1 pyramidal neurons, negative current was injected to the cell. Input resistance Ri, also called as membrane resistance, Rm, and series resistance, Rs, also called access resistance, Ra, were calculated using steady-state current and capacitive currents, respectively, with −2.5 or −5 mV hyperpolarizing step pulses (100 msec) at V rest = −70 mV without compensation of fast and slow capacitance.

Extracellular recordings

Field recording of excitatory postsynaptic potential (fEPSP) of hippocampal CA1 neurons of postnatal day (P) 8-10 week was performed as described with minor modifications (Park et al., 2008). For recording, slices were placed in an interface-type recording chamber and perfused with 32°C ACSF at a rate of 3 ml/min. fEPSPs were recorded with extracellular recording electrodes (0.5-1.5 MΩ) filled with ACSF and placed in the stratum radiatum of area CA1. Synaptic responses were evoked by a 100 μsec current pulse to Schaffer collateral axons with a 38 concentric bipolar stimulating electrode which had a central core conductor of Epoxylite-coated Pt/Ir wire protruding 125 μm from the surrounding 35 gauge stainless steel tube. Stable baseline responses were collected every 30 s by using a stimulation intensity (10-30 μA) yielding 50%–60% of the maximal fEPSP slope response. fEPSP was measured in Schaffer collateral-CA1 synapses in hippocampal slices derived from 8-10 week old male mice. LTP was induced by high frequency stimulation (HFS) (100 Hz, 1 s).

Behavioral Testing

Mice were handled for several minutes a day for four consecutive days at least one week prior to the start of behavioral testing to familiarize them with handling. All testing was performed in an isolated behavior room at 23–24°C. All behavioral testing occurred in the light phase of the circadian cycle. Before each test, mice were moved to the testing room and allowed to habituate for at least 1 h before behavioral testing. Behaviors were video recorded by a computer-based video tracking system (Any Maze, Stoelting Co, Wood Dale, IL). Behavioral tests were separated by at least 24 h.

Social Recognition Task

To test for social recognition, a habituation-dishabituation paradigm with reciprocal interactions was carried out in a neutral arena consisting of a clean standard mouse cage as previously described (Savonenko et al., 2008). The test mouse was placed in the cage 2 min before a stimulus mouse to avoid an interference of the orientation response to the cage and to the stimulus mouse. Stimulus animals were 28-35 day-old mice of the same sex (males) and strain background (C57Bl6/J) as the test mice, housed in groups of 3-5 mice per cage. In the habituation stage, the test mouse was exposed to the same stimulus mouse over three 2-min-long trials with an inter-trial interval of 25 min. For the fourth, dishabituation, trial, the test mouse was exposed to a novel stimulus mouse for 2 min after the same inter-trial interval (25 min). The time and number of events of social investigation were recorded by well-trained observers blind to the experimental goals and genotypes of mice using observation acquisition coding in Any Maze software. Social investigation toward stimulus mouse was defined as direct, active, olfactory exploration of the stimulus mouse, specifically nosing and sniffing of the head, body or anogenital regions, close following, and pursuit. Each stimulus mouse was used in only one habituation-dishabituation paradigm per day. The test cage was cleaned with 30% ethanol between each of the four testing trials.

Social Motivation Task

To test for sociability, preferences to social and non-social stimuli were assayed as described in Lin et al. (2021) and Moy et al. (2004) with minor modifications. The apparatus was a rectangular box (60 cm x 40 cm x 35 cm high) made from opaque Plexiglas. The novel stimulus mice were mature (2 month old) male mice of C57Bl6/J strain background. Two metal pencil holders [8.5 cm height x 8.5 cm diameter with a net-like wire walls (5x3 mm net cells)] were placed symmetrically from the center and the walls of the apparatus. Before social motivation testing, a 5-min-long orientation trial was conducted to allow each test mouse to investigate the apparatus with no stimulus mice present. During the next trial, a novel stimulus mouse was placed under one of the holders (in a counterbalanced order under the left or right holder) and 2 min later the test mouse was introduced for 5 min of observation. After each test, holders were washed in water followed by 70% ethanol and dried by a paper towel. Testing cage was cleaned by 30% ethanol between trials.

For characterization of social preferences, three variables were automatically recorded while tracking animal’s head using Any-Maze software: 1) total time spent investigating sides of the apparatus (25 cm x 40 cm) containing enclosures; 2) average proximity to each enclosure; and 3) total time spent investigating area of the apparatus centered around each enclosure (15 cm diameter). The choice of these variables for the social motivation task was motivated by an attempt to increase relevance to other studies as well as to avoid possible effects of an arbitrary character of any single behavioral measure. The variable of area investigation time is one of the most used variables for this task. The variable of side investigation is used in a significant number of studies that utilize commercially available equipment tracking the presence of an animal in one side of the apparatus or another. The third variable, average proximity, is analogous to a measure of spatial preference used widely in the water maze tasks (Gallagher et al., 2015). This measure is designed to reflect a spatial bias of animal exploration. In this study, we calculated proximity to each of the two targets (social and non-social enclosures). In addition, motor activity of test mouse was automatically recorded during the orientation trial.

Novelty-induced exploration and anxiety levels in Open Field

Novelty-induced motor activation was tested by placing the mouse in a novel brightly-lid open-field arena that had a diameter of 100 cm and 55 cm high sidewalls. Open field testing was carried out as previously described (Melnikova et al., 2006). Each subject was released near the wall and behavior recorded for 15 min. Activity measures included distance traveled, percent time spent in active exploration (episodes of movement ≥ 5 cm/s), and speed of movement during active exploration. To analyze anxiety levels, the activity measures were broken down into two zones. Based on our previous studies, a 20 cm wide wall zone constituted the most preferred peripheral zone, while the rest of the open field was defined as a central zone comprising ~67% of the arena surface and was most aversive for mice. Time spent in the peripheral zone was used as a measure of thigmotaxis.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical methods used for RNA-seq analyses are described in the appropriate method section. The biochemical and electrophysiological data were analyzed using a two-sided t test or analysis of variance (ANOVA) where appropriate using GraphPad Prism 9 software (GraphPad Software Inc, San Diego, CA). The behavioral data were analyzed using the statistical package STATISTICA 13 (TIBCO Software Inc, Palo Alto, CA). Group comparisons were carried out using mixed design ANOVA with Genotype as a main factor and repeated-measure(s) (RM) (trials, blocks, enclosures, etc). Group x RM interactions were set as orthogonal. The Turkey post hoc test was applied to significant Group or Group x RM interactions to evaluate statistical significance of differences between specific sets of means. An effect size, a standardized (in units of StDev) difference between means, was used to report results from different variables in a comparable scale. Among numerous measures developed to characterize effect sizes, we have chosen Hedges’ g coefficient that successfully incorporates corrections for group sizes as well as for the correlation between repeated-measures in a within-subject design (Lakens, 2013). The latter was particularly important as the concept of social preference can be represented as a difference in investigation of social versus non-social object measured in the same subject. Reporting effect sizes for measures of social motivation was also in compliance with the guidance from the American Statistical Association that the p values, in isolation, do not measure the importance of results (Betensky, 2019; Lakens, 2013). Effect sizes were considered as small, medium, or large ifq ~0.2, 0.5 or more than 0.8, respectively. Sample sizes and details of statistical analyses were presented in figure legends. p < 0.05 denoted statistical significance. Group data were presented as mean ± s.e.m., unless listed otherwise.

Supplementary Material

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2021.110014.

Highlights.

• Stress causes deteriorated social preference behavior in Shank3^ΔC/+ mice

• Stress alters transcriptome and boosts Homer1a in Shank3^ΔC/+ pyramidal neurons

• Shank3 expression in vivo is reduced by Homer1a and depends on Homer crosslinking

• Homer1a^−/− rescues Shank3 level and stress-induced phenotypes in Shank3^ΔC/+ mice

INCLUSION AND DIVERSITY.

One or more of the authors of this paper self-identifies as a member of the LGBTQ+ community. While citing references scientifically relevant for this work, we also actively worked to promote gender balance in our reference list.