Inter-chromosomal Contact Properties in Live-Cell Imaging and in Hi-C

Abstract

Imaging (fluorescence in situ hybridization [FISH]) and genome-wide chromosome conformation capture (Hi-C) are two major approaches to the study of higher-order genome organization in the nucleus. Intra-chromosomal and inter-chromosomal interactions (referred to as non-homologous chromosomal contacts [NHCCs]) have been observed by several FISH-based studies, but locus-specific NHCCs have not been detected by Hi-C. Due to crosslinking, neither of these approaches assesses spatiotemporal properties. Toward resolving the discrepancies between imaging and Hi-C, we sought to understand the spatiotemporal properties of NHCCs in living cells by CRISPR/Cas9 live-cell imaging (CLING). In mammalian cells, we find that NHCCs are stable and occur as frequently as intra-chromosomal interactions, but NHCCs occur at farther spatial distance that could explain their lack of detection in Hi-C. By revealing the spatiotemporal properties in living cells, our study provides fundamental insights into the biology of NHCCs.

Keywords: CISTR-ACT; CLING; CRISPR live-cell imaging; FIRRE; Hi-C; NHCCs; inter-chromosomal interactions; lncRNA; loci dynamics.

Figure 1. NHCC frequencies in CLING and in Hi-C

(A) Intra-chromosomal interactions of FIRRE with DXZ4, or with a control locus at 101 Mb (*** p < 0.0001, Mann-Whitney rank-sum test). CLING-determined co-localization frequencies (each quantification in 100 RPE-1 nuclei), in comparison to Hi-C signals (resolution/bin size = 40 kb), or Hi-C normalized read numbers in RPE-1 cells. (B) NHCCs of FIRRE-ATF4 (58 %), and FIRRE-PGR (25 %) in CLING (each quantification in 100 RPE-1 nuclei) in comparison to Hi-C contact probabilities (resolution/bin size = 250 kb), or normalized read counts.

Figure 2. NHCC dynamics

(A) Distances (D, d) between co-localized signals were measured and compared to non-co-localized signals over time (t) in 4D-CLING. (B) The ratios of distances (D_n or d_n/D₁, see panel A), determined that co-localized signals were closer to each other than GAPDH and PGR, and they remained associated over time (*** p < 0.0001, Mann-Whitney rank-sum test, means±SEM). (C) Speed measurements (m/s)⁻⁶ separated in minimum and maximum of time-matched 4D-data. At low speeds, all loci were less mobile than the control loci GAPDH and PGR (*** p < 0.0001, Mann-Whitney rank sum test). (D) Spatial distances between co-localized and non-co-localized signals (*** p < 0.0001, Mann-Whitney rank sum test). (E) Distances between co-localized loci of intra- or inter-chromosomal interactions in RPE-1 nuclei (*** p < 0.0001, Mann-Whitney rank sum test, without outlier). (F) Distributions of co-localization distances of intra- or inter-chromosomal interactions were mostly unimodal, whilst was FIRRE-YPEL4 distances were bimodal (Hartigan’s dip test, ** p < 0.001, black dots = medians). (G) Spearman correlations between intra- (bin size = 100 kb) or inter-chromosomal (bin size = 500 kb) CLING spatial distances, CLING co-localization frequencies (size of data points indicates co-localization frequencies in %), and Hi-C interaction probabilities (z-scores) in RPE-1 Hi-C. Smaller spatial distances correlate with higher Hi-C z-scores. (H) Discrepancies between imaging and Hi-C methodologies. (I) Schemes of spatial organization of two given loci in intra-chromosomal interactions and NHCCs (d = distance)