Spatiotemporal 7q11.23 protein network analysis implicates the role of DNA repair pathway during human brain development

Abstract

Recurrent deletions and duplications of chromosome 7q11.23 copy number variants (CNVs) are associated with several psychiatric disorders. Although phenotypic abnormalities have been observed in patients, causal genes responsible for CNV-associated diagnoses and traits are still poorly understood. Furthermore, the targeted human brain regions, developmental stages, protein networks, and signaling pathways, influenced by this CNV remain unclear. Previous works showed GTF2I involved in Williams-Beuren syndrome, but pathways affected by GTF2I are indistinct. We first constructed dynamic spatiotemporal networks of 7q11.23 genes by combining data from the brain developmental transcriptome with physical interactions of 7q11.23 proteins. Topological changes were observed in protein-protein interaction (PPI) networks throughout different stages of brain development. Early and late fetal periods of development in the cortex, striatum, hippocampus, and amygdale were observed as the vital periods and regions for 7q11.23 CNV proteins. CNV proteins and their partners are significantly enriched in DNA repair pathway. As a driver gene, GTF2I interacted with PRKDC and BRCA1 to involve in DNA repair pathway. The physical interaction between GTF2I with PRKDC was confirmed experimentally by the liquid chromatography-tandem mass spectrometry (LC-MS/MS). We identified that early and late fetal periods are crucial for 7q11.23 genes to affect brain development. Our results implicate that 7q11.23 CNV genes converge on the DNA repair pathway to contribute to the pathogenesis of psychiatric diseases.

Figure 1

A flow-chart shows the plan of work involved in this research study. (A) Twenty-three 7q11.23 CNV genes expressed in the brain were identified. (B) Physical protein–protein interaction dataset combined with 7q11.23 CNV genes to construct CNV protein–protein interactions (PPIs). (C) 7q11.23 CNV PPIs combined with the Human brain transcriptome dataset. (D) 7q11.23 spatiotemporal co-expression PPIs network was established. (E) Functional enrichment analysis and functional module analysis were performed.

Figure 2

The 7q11.23 co-expressed interacting protein pairs are significantly enriched in three spatiotemporal intervals. The fractions of protein pairs from 7q11.23 CNV co-expressed and interacting with HIBE proteins (red line), all co-expressed and interacting HIBE proteins (black line), proteins from 1000 Genome Project CNVs co-expressed and interacting with HIBE proteins (dark gray line), and 7q11.23 CNV proteins co-expressed with all brain-expressed human genes (aquamarine line). Thirty-one spatiotemporal intervals of brain development are shown on the x-axis. 7q11.23 co-expressed interacting protein pairs are significantly enriched in spatiotemporal intervals (indicated by star symbol) compared with control networks. The statistical enrichment was calculated using Fisher’s exact test, and P values were FDR-corrected for multiple comparisons.

Figure 3

Functional convergence of the 7q11.23 spatiotemporal networks. The overlap of 7q11.23 genes (left Venn diagram) and their co-expressed interacting partners (right Venn diagram) are across three significant spatiotemporal intervals. Top 20 significant enriched biological process GO terms of shared proteins are showed.

Figure 4

Difference between the 7q11.23 spatiotemporal networks. Spatiotemporal networks were compared across different brain regions within the same developmental period (P1R1 and P1R3) and cross different development periods within the same brain region (P1R3 and P4R3). 7q11.23 genes are shown as red nodes, their co-expressed interacting partners as gray node, and the PPIs between co-expressed genes at a particular developmental period are shown as gray edges. The nodes that lost all edges were removed from the corresponding networks. Significant differences are observed across developmental periods but not across brain regions (ANOVA statistics shown below the graphs).

Figure 5

Functional analyses of proteins within three significant intervals, P1R1, P1R3 and P4R3. Dot plot shows top 10 enriched GO terms of biological process for CNV proteins and their partners within three significant intervals.

Figure 6

Spatiotemporal networks implicate GTF2I-PRKDC-DDR pathway and proteomic investigation of interaction between GTF2I and PRKDC. (A) Dynamic spatiotemporal networks of the GTF2I, GTF2IRD1, and RFC2 interacted with their hub partners. 7q11.23 genes are shown as red nodes, their co-expressed hub partners as gray node, and the PPIs between co-expressed genes at a particular developmental period are shown as gray edges. The nodes that lost all edges were removed from the corresponding networks, and the PPIs between co-expressed genes at a particular developmental period are showed as colored edges (P1R1, blueviolet; P1R3, green; P4R3, turquoises). (B) GTF2I and PRKDC were identified and the amino acids highlighted are peptides identified by immunoprecipitation (IP) and LC–MS/MS.