b iframe width height src https www youtube
expressed genes upon stable SRC-2 depletion in MCF-7 breast cancer Trichostatin A (TSA) and used SRC-3 depleted cells for comparison (Fig. 1). Firstly, eﬃcient SRC-2 and SRC-3 knock-down in MCF-7 was confirmed at the mRNA and protein levels (Fig. 1a and b). Interestingly, we observed a compensatory eﬀect, where a successful knock-down of SRC-2 resulted in a significant increase in both SRC-3 mRNA and protein level and a similar eﬀect on SRC-2 when knocking down SRC-3. No such com-pensatory eﬀect was detected on the SRC-1 mRNA level, the last member of the SRC-family (Fig. 1a), and we therefore excluded SRC-1 from further analyses.
Moreover, correspondence analysis of normalized microarray data showed that the three biological groups of samples (shKTR, shSRC-2 and shSRC-3) have definite separate gene expression profiles (Fig. 1c). We further organized all diﬀerentially expressed (DE) genes, extracted by SAM analysis (fold change ≥ 1.5 ), into three major groups of gene regulation - unique to SRC-2, unique to SRC-3, and genes commonly regulated by both proteins, in order to further reveal diﬀerences be-tween SRC-2 and SRC-3 transcriptional regulation (Fig. 1d). Compared to shKTR, SRC-3 knock-down induced changes in 371 genes, while SRC-2 knock-down induced changes in a total of 246 genes, of which 146 genes were unique for SRC-2. Additionally, the fraction of genes com-monly regulated by both SRCs represented a minority in both DE gene sets (40% of the total SRC-2 gene set and 27% of the total SRC-3 gene set). Amongst the commonly regulated genes, there were more genes regulated in the same direction of gene expression change (88 genes), than in the opposite direction (12 genes). Finally, we observed more downregulated than upregulated genes both among the shSRC-2 and
shSRC-3 DE genes. Taken together, these results indicate that, although SRC-2 and SRC-3 may mutually regulate a subset of genes, SRC-2 pre-dominantly regulates expression of separate set of genes in MCF-7 cells, as compared to SRC-3.
3.2. SRC-2 target genes and their function
To gain insight into which biological functions are predominantly aﬀected by SRC-2 knock-down, as compared to SRC-3 knock-down, we first performed a functional classification of seven gene groups based on the term Biological Process, using GO annotation database and bioin-formatic tool PANTHER (Fig. 2). The categorization showed that genes implicated in, Developmental processes, Cell cycle and Cellular component organization are highly aﬀected by independent SRC-2 and SRC-3 de-pletions. Interestingly, SRC-2 upregulated genes show similar over-representation as SRC-3 downregulated genes in these categories. The common SRC-2 and SRC-3 upregulated genes show majority of changes in Response to stimulus and Immune system processes. Group of common SRC-2 and SRC-3 genes oppositely regulated by the two coactivators consisted of 12 genes only and was therefore not further analyzed.
We then analyzed SRC-2 and SRC-3 diﬀerentially expressed genes with a FC ≥ 1.5 using Ingenuity Pathway Analysis (IPA) and, simi-larly to the PANTHER analysis, we found the categories related to de-velopment, cell cycle and cellular morphogenesis enriched in both shSRC-2 and shSRC-3 gene sets (Supplementary Fig. S3). Additionally, the analysis identified 11 SRC-2-specific GO categories where the strongest enrichment was found in the GO term Lipid metabolism. One
Fig. 2. Functional analysis of DE genes changed after knock-down of SRC-2 or SRC-3. SRC-2 unique upregulated genes (column 2), SRC-2 unique downregulated genes (column 3), SRC-3 unique upregulated (column 4), SRC-3 unique downregulated genes (column 5), common SRC-2 and SRC-3 genes with transcriptional changes in the same direction: upregulation (column 6), downregulation (column 7), and common SRC-2 and SRC-3 genes with transcriptional changes in the opposite direction (column 8) were categorized in PANTHER, based on the GO term Biological Process. The numbers below gene group name represent the number of diﬀerentially regulated genes that were compared to the reference (human genome, Ref; column 1). The numbers in the table represent changes in the tested gene groups as compared to the reference. Color intensity indicates the strength of statistical significance (-log p-value) for under-represented (blue) and over-represented (red) GO categories. Only GO categories with the most pronounced changes are shown.