Association between Triplex-Forming Sites of Cardiac Long Noncoding RNA GATA6-AS1 and Chromatin Organization

Abstract

This study explored the relationship between 3D genome organization and RNA-DNA triplex-forming sites of long noncoding RNAs (lncRNAs), a group of RNAs that do not code for proteins but are important factors regulating different aspects of genome activity. The triplex-forming sites of anti-sense cardiac lncRNA GATA6-AS1 derived from DBD-Capture-Seq were examined and compared to modular features of 3D genome organization called topologically associated domains (TADs) obtained from Hi-C data. It was found that GATA6-AS1 triplex-forming sites are positioned non-randomly in TADs and their boundaries. The triplex sites showed a preference for TAD boundaries over internal regions of TADs. Computational prediction analysis indicated that CTCF, the key protein involved in TAD specification, may interact with GATA6-AS1, and their binding sites correlate with each other. Examining locations of repeat elements in the genome suggests that the ability of lncRNA GATA6-AS1 to form triplex sites with many genomic locations may be achieved by the rapid expansion of different repeat elements. Some of the triplex-forming sites were found to be positioned in regions that undergo dynamic chromatin organization events such as loss/gain of TAD boundaries during cardiac differentiation. These observed associations suggest that lncRNA-DNA triplex formation may contribute to the specification of TADs in 3D genome organization.

Keywords: CTCF; GATA6-AS1; RNA–DNA triplex; topologically associated domains.

Figure 1

Enrichment of lncRNA GATA6-AS1 triplex sites in specific regions in the context of 3D genome organization. The distribution of expected coverage (blue) compared to the observed coverage (vertical red line) in TAD domains and TAD boundaries are shown in panels (A,B), respectively. The x-axis and y-axis represent coverage (in kb) and frequency, respectively. (C) Enrichment of GATA6-AS1 triplex sites at different regions in TADs. A TAD was divided into 10 regions or bins based on distance from the boundary. The y-axis represents the fraction of the size of the bin that was occupied by GATA6-AS1 triplex sites (or a control set of randomized sites). The error bars indicate the standard error of the mean.

Figure 2

GATA6-AS1 and dynamic loss/gain of TAD boundaries and A/B compartment switching. (A) Fourteen unique possibilities of a genomic site based on whether it was associated with the gain or loss of a TAD boundary during cardiac differentiation stages (ES, MES, CP, and CM). The counts for GATA6-AS1 and the control set of randomized sites annotated as one of these unique possibilities are shown and were compared using the Chi-Square test. p-value < 0.001 is indicated by ***. (B) As in panel (A), 14 unique possibilities of a genomic site based on whether it was associated with switching between the two types of compartments, A and B, during the cardiac differentiation stages (ES, MES, CP, and CM). The counts for GATA6-AS1 and the control set of randomized sites annotated as one of these unique possibilities are shown and were compared using the Chi-Square test. p-value < 0.001 is indicated by ***. (C) Gene ontology analysis of targets of triplex-forming sites of GATA6-AS1. The x-axis and y-axis show the top enriched GO “Biological Process” terms and −log10 p-value (enrichment score), respectively.

Figure 3

Characteristics of GATA6-AS1 triplex sites in relationship to CTCF, repeat elements, and conservation. (A) The probability of binding of CTCF (y-axis) to different regions of the GATA6-AS1 sequence (x-axis) is shown. (B) Boxplots comparing the amount of CTCF signal at GATA6-AS1 triplex sites and a control set (genome). *** indicates p-value < 0.001. (C) Fraction (y-axis) of the length of GATA6-AS1 triplex sites occupied by four groups of repeat elements (LTR, SINE, LINE, and DNA transposons) is compared to a control set (genome). GATA6-AS1 triplex sites are further grouped into two groups based on whether it is associated with TAD boundaries or non-TAD boundaries. For each type of repeat element, six statistical comparisons were tested and only the ones which showed significant differences are marked as ***. (D) Conservation score per base pair (y-axis) is shown for GATA6-AS1 triplex sites and a control set (genome). six statistical comparisons were tested and only the ones which showed significant differences are marked as ***.