Fig 2: Regulation of neuronal gene expression by ZMYND11.(a) Schematic showing stereotactic injection of AAV expressing either Cre recombinase or non-functional ΔCre into opposite hemispheres of primary visual cortex (V1) of 4-week-old Zmynd11fl/fl mice. (b) Volcano plot of gene expression changes in Zmynd11fl/fl AAV-Cre vs. AAV-ΔCre cortex. Plot shows DESeq2 log2 fold-change RNA expression vs. −log10 p-value with Benjamini-Hochberg false discovery rate (FDR) adjustment. Dashed lines show cutoff for log2 fold-change (± 0.25) and FDR-adjusted p-value (0.1). Tissue from primary visual cortex (n = 8 individual mice). (c) Gene ontology analysis of DESeq2 significantly up- and down-regulated genes in Zmynd11fl/fl AAV-Cre vs. AAV-ΔCre cortex, GO:BP (Biological Process) gene set, analysis by gProfiler2. (d) Schematic showing infection of DIV7 primary cortical neurons from Zmynd11fl/fl embryos with lentivirus expressing either Cre or ΔCre. Samples were collected for total RNA and nascent RNA (PRO-Seq) processing. (e) Scatter plot showing changes in total and nascent RNA expression in Zmynd11fl/fl Lenti-Cre vs. Lenti-ΔCre DIV 7 cortical cultures. Plot shows log2 fold-change total RNA-seq vs. log2 fold-change PRO-Seq; dotted line shows linear fit with Pearson correlation coefficient labeled. n = 3 independent neuron dissections (RNA-Seq), and n = 4 independent neuronal dissections (PRO-Seq). (f) Aggregate plots showing the average distribution of PRO-Seq signal across genes in wild-type primary cortical neurons. Gene body and transcription start site (TSS) regions highlighted for computation of Pausing Index. (g) Box plot of PRO-seq counts at the transcription start site (left) and gene body (middle) of significant upregulated, downregulated, and unchanged genes by DESeq2 in Zmynd11fl/fl Lenti-Cre vs. Lenti-ΔCre. (Right) Box plot showing pausing index (ratio of PRO-Seq counts at TSS/GB) at up-, down-, and unchanged genes. *** = p < 1E-30, * = p < 1E-05, p-values calculated by Wilcoxon rank-sum test (n = 4 independent neuronal dissections).

Fig 3: ZMYND11 interacts with KMT2A in the brain.(a) Schematic showing experimental approach of immunoprecipitation followed by protein mass spectrometry for the discovery of proteins interacting with endogenously FLAG-HA-tagged ZMYND11. (b) Scatter plot showing proteins identified by mass spectrometry in eluates from anti-FLAG immunoprecipitation of ZMYND11-FH from whole brain tissue, ranked by enrichment in ZMYND11-FH tissue over wild-type; n = 3 independent replicates per condition. (c) Representative western blot showing co-immunoprecipitation (co-IP) of ZMYND11 and KMT2A from nuclear lysates of brain tissue from Zmynd11FH/FH (FH) and wild-type littermates (WT). FT: Flow-through, IP: anti-FLAG immunoprecipitation. Data are representative of 3 independent experiments. (d) Heatmap of CUT&RUN signal (scaled read coverage per bp) across genes ranked by ZMYND11 occupancy for KMT2A (N- and C-terminal fragments), RBBP5, ASH2L, and ZMYND11 in 6-week-old mouse cortex; n = 3 individual mice per condition. (e) Venn diagram showing overlap of TSS-localized ZMYND11 peaks and MLL1 consensus peaks. (f) Box plots showing H3K4 mono- (H3K4me1) and tri-methylation (H3K4me3) levels at top and bottom quartiles of ZMYND11-bound transcription start sites, showing the difference in CUT&RUN density between wild-type and Zmynd11-cKO in 6-week-old mouse cortex. * = p < 0.01, p-values calculated using Wilcoxon rank-sum test; n = 3 independent replicates per condition. (g) Box plots showing differential gene expression by DESeq2 analysis of RNA-seq in Zmynd11fl/fl AAV-Cre vs. AAV-ΔCre, with genes separated into quartiles based on enrichment of WT/cKO ZMYND11 CUT&RUN signal at TSS (low to high). * = p < 0.05, ** = p < 0.01, *** = p < 0.00001, p-values calculated using Wilcoxon rank-sum test; n = 3 independent replicates per condition.

Fig 4: ZMYND11 regulates brain development and mouse behavior.(a) Heatmap showing relative levels of Zmynd11 transcript in mouse (Mouse ENCODE transcriptome data) and human tissues (RNA sequencing of total RNA from 20 human tissues, UConn Health Center). (b) PCR genotyping showing Cre-mediated excision of Zmynd11 exon 7 in Zmynd11fl/fl brain tissue with expression of Baf53b-Cre. (c) Western blot showing expression of ZMYND11 and β-Actin in cell lysates of cortical tissue. Zmynd11 protein is depleted in cortex from Zmynd11fl/fl; Baf53b-Cre (Zmynd11-cKO) mice. (d) Box plot showing the distribution of body mass of 6-week-old Zmynd11fl/fl; Baf53b-Cre (Zmynd11-cKO) mice and wild-type (Zmynd11wt/wt; Baf53b-Cre) littermates. ** = p < 0.001, p-values calculated using Wilcoxon rank-sum test, n = 13 WT Female, n = 11 cKO Female, n = 20 WT Male, and n = 12 cKO Male. (e) Representative brightfield image (minimum-intensity plot of confocal image stack) of Golgi-stained cortex from Zmynd11-cKO mouse at 10X magnification. Scale bar = 200 μm. Data are representative of 6 individual mice per condition. (f) Representative traces of individual Layer 2/3 cortical pyramidal neurons of 6-week-old Zmynd11fl/fl; Baf53b-Cre and Zmynd11wt/wt; Baf53b-Cre mice. Live tracing of brightfield images from Neurolucida software; scale bar = 200 μm. Data are representative of 6 cells per mouse, 6 individual mice per condition. (g) Sholl analysis of dendritic branch complexity in basal and apical dendrites of Layer 2/3 pyramidal neurons from 6-week-old Zmynd11fl/fl; Baf53b-Cre and Zmynd11wt/wt; Baf53b-Cre mice. Plot shows mean values ± SEM. p-values calculated using linear mixed-effects mode; n = 6 cells per mouse, 6 individual mice per condition. (h) Comparison of dendritic spine density in basal and apical dendrites of Layer 2/3 pyramidal neurons from 6-week-old Zmynd11fl/fl;Baf53b-Cre and Zmynd11wt/wt; Baf53b-Cre mice. * = p < 0.05, p-values calculated using Wilcoxon rank-sum test; n = 6 cells per mouse, 6 individual mice per condition. (i) Quantification of freezing behavior exhibited by Zmynd11-cKO mice and wild-type littermates (10–12 weeks old) upon contextual fear conditioning during acquisition (foot shock in novel context, final minute of 5-minute acquisition period), recall (repeat of context without additional foot shock, first 5 minutes in shock context cage), and extinction (after 25 minutes in shock context cage). p-values calculated using mixed model ANOVA, n = 7 individual mice (WT), n = 5 individual mice (Het), n = 8 individual mice (cKO). (j) Duration of mouse interactions with a novel and habituated object present in the same housing context, short-term recall. ** = p < 0.001, p-values calculated using Wilcoxon rank-sum test, n = 19 individual mice (WT), n = 21 individual mice (cKO). (k) Quantification of distance travelled by mice during the novel object recognition task in (j). *** = p < 0.0001, p-values calculated using Wilcoxon rank-sum test, n = 19 individual mice (WT), n = 21 individual mice (cKO). (l) Number of marbles buried by mice in open arena. * = p < 0.01, p-values calculated using Wilcoxon rank-sum test, n = 6 individual mice (WT), n = 9 individual mice (cKO).

Fig 5: ZMYND11 binds active chromatin across the genome.(a) Representative tracks from IGV (Integrated Genomics Viewer) showing CUT&RUN signal for ZMYND11 and H3K36me3 in wild-type (Zmynd11wt/wt; Baf53b-Cre) and Zmynd11-cKO (Zmynd11fl/fl; Baf53b-Cre) cortex at two example loci. (b) (Above) Heatmaps of CUT&RUN signal from cortical tissue (scaled read coverage per bp) across all genes for ZMYND11 and H3K36me3. (Below) Aggregate plots showing the average distribution of CUT&RUN signal for ZMYND11 and H3K36me3 across genes separated into quartiles by level of expression. ZMYND11 CUT&RUN is normalized by subtraction of signal present in Zmynd11-cKO; n = 2 independent replicates per condition. (c) Aggregate plots showing average distribution of ZMYND11 CUT&RUN signal centered around significant ZMYND11 peaks in Zmynd11fl/fl;Baf53b-Cre (cKO) and Zmynd11wt/wt; Baf53b-Cre (WT) brain tissue (MACS2 no-control p-value < 1e-05, ZMYND11 CUT&RUN WT/cKO > 2). (d) Annotation of the genomic localization of ZMYND11 peaks by Homer annotatePeaks (MACS no-control p-value < 1e-05, ZMYND11 CUT&RUN WT/cKO > 2). (e) Proportion of genes with significant differential expression in Zmynd11fl/fl AAV-Cre vs. AAV-ΔCre, grouped by the number of ZMYND11 CUT&RUN peaks per gene.