SOX9 chromatin folding domains correlate with its real and putative distant cis-regulatory elements

Abstract

Evolutionary conserved transcription factor SOX9, encoded by the dosage sensitive SOX9 gene on chromosome 17q24.3, plays an important role in development of multiple organs, including bones and testes. Heterozygous point mutations and genomic copy-number variant (CNV) deletions involving SOX9 have been reported in patients with campomelic dysplasia (CD), a skeletal malformation syndrome often associated with male-to-female sex reversal. Balanced and unbalanced structural genomic variants with breakpoints mapping up to 1.3 Mb up- and downstream to SOX9 have been described in patients with milder phenotypes, including acampomelic campomelic dysplasia, sex reversal, and Pierre Robin sequence. Based on the localization of breakpoints of genomic rearrangements causing different phenotypes, 5 genomic intervals mapping upstream to SOX9 have been defined. We have analyzed the publically available database of high-throughput chromosome conformation capture (Hi-C) in multiple cell lines in the genomic regions flanking SOX9. Consistent with the literature data, chromatin domain boundaries in the SOX9 locus exhibit conservation across species and remain largely constant across multiple cell types. Interestingly, we have found that chromatin folding domains in the SOX9 locus associate with the genomic intervals harboring real and putative regulatory elements of SOX9, implicating that variation in intra-domain interactions may be critical for dynamic regulation of SOX9 expression in a cell type-specific fashion. We propose that tissue-specific enhancers for other transcription factor genes may similarly utilize chromatin folding sub-domains in gene regulation.

Keywords: chromatin looping; long distance gene regulation; non-coding variants; structural variants; tissue-specific enhancers.

Figure 1.

Schematic representation of Hi-C in cis genomic interactions within an ∼3 Mb genomic region flanking SOX9 at 17q24.3. (A) Hi-C profiles at 25 kb resolution around SOX9 in HMEC, HUVEC, NHEK, and IMR90 cell lines. An ∼1.87 MbSOX9 topologically associated domain (TAD) extending between 68.67 to 70.54 Mb (hg19) is designated by the black dashed lines. (B) Hi-C Juicebox view profiles at 5 kb resolution around SOX9 in HMEC, HUVEC, NHEK, and IMR90 cell lines. (C) Location of the protein coding genes KCNJ2, SOX9 (red), and SLC39A11 is shown. (D, E) Histograms depict CTCF ChIP-seq enrichment levels and chromatin states (ChromHMM output) (active TSS - red, transcribed - green, enhancer - yellow, low – gray, and heterochromatin - purple). (F) Arcs represent “arrowhead” chromatin folding domains reported by Rao et al.

Figure 2.

(A) Hi-C contact map around the SOX9 locus in NHEK cells at 25 kb resolution. Black dashed lines depict the SOX9 TAD. Vertical arrows show borders of TADs present in NHEK cells (black) and other cells (dotted). (B) The structural variants breakpoint cluster intervals mapping 5′ to SOX9 and related to campomelic dysplasia (CD, violet box), Pierre Robin sequence (PRS, blue box), acampomelic campomelic dysplasia (ACD, navy box), and sex reversal (XXSR, red box; XYSR, gray box) are shown. (C) Regulatory elements shown as ellipses found to be active in: mandibular mesenchyme (blue), testes (red), node, notochord, gut, bronchial epithelium, and pancreas - E1, migrating cranial neural crest cells - E3, fore- and midbrain E7 (brown), chondrosarcoma cells (navy), heart (green), craniofacial and palatal tissue (gray), somatic tissues (orange), and cartilage (black). Note the correlation of these regulatory elements and breakpoint cluster intervals with the SOX9 TADs. (D) Hi-C Juicebox view contact map around the SOX9 locus in GM12878 cells at 5 kb resolution. Black dashed lines depict the SOX9 TAD. Please note that the SOX9 gene maps inside a CTCF/cohesin loop and that there are several enhancers contained within this loop that interact with the SOX9 promoter.