A Portrait of Ribosomal DNA Contacts with Hi-C Reveals 5S and 45S rDNA Anchoring Points in the Folded Human Genome

Abstract

Ribosomal RNAs (rRNAs) account for >60% of all RNAs in eukaryotic cells and are encoded in the ribosomal DNA (rDNA) arrays. The rRNAs are produced from two sets of loci: the 5S rDNA array resides exclusively on human chromosome 1, whereas the 45S rDNA array resides on the short arm of five human acrocentric chromosomes. The 45S rDNA gives origin to the nucleolus, the nuclear organelle that is the site of ribosome biogenesis. Intriguingly, 5S and 45S rDNA arrays exhibit correlated copy number variation in lymphoblastoid cells (LCLs). Here we examined the genomic architecture and repeat content of the 5S and 45S rDNA arrays in multiple human genome assemblies (including PacBio MHAP assembly) and ascertained contacts between the rDNA arrays and the rest of the genome using Hi-C datasets from two human cell lines (erythroleukemia K562 and lymphoblastoid cells). Our analyses revealed that 5S and 45S arrays each have thousands of contacts in the folded genome, with rDNA-associated regions and genes dispersed across all chromosomes. The rDNA contact map displayed conserved and disparate features between two cell lines, and pointed to specific chromosomes, genomic regions, and genes with evidence of spatial proximity to the rDNA arrays; the data also showed a lack of direct physical interaction between the 5S and 45S rDNA arrays. Finally, the analysis identified an intriguing organization in the 5S array with Alu and 5S elements adjacent to one another and organized in opposite orientation along the array. Portraits of genome folding centered on the ribosomal DNA array could help understand the emergence of concerted variation, the control of 5S and 45S expression, as well as provide insights into an organelle that contributes to the spatial localization of human chromosomes during interphase.

Keywords: Hi-C; concerted variation; genome structure; nuclear structure; nucleolus; rDNA.

Fig. 1.—

Organization of the 5S rDNA array in multiple assemblies. The red arrows represent Alu elements; black solid arrows represent 5S units. Coordinates are for the segment in chromosome 1 of the HuRef and CHM1_1.1 assemblies or for the segment in the contigs of the CHM1tert assembly.

Fig. 2.—

Detailed structure of the 5S repeat unit and copy number along the unit. The red arrows represent Alu elements; black solid arrows represent 5S elements with 100% identity to the reference. The graph shows the profile of normalized read-depth (copy number) along the 5S rDNA repeat unit for two individuals (NA18916 and NA19108).

Fig. 3.—

Structure of the genomic segment containing the 5S rDNA array. The figure displays the 5S array (∼40 kb) and upstream and downstream segments (40 kb on each side of the array) in the 1q42 region of hg19. (A) Line1 and Alu sequences in the 5S rDNA array region and normalized read-depth (copy number) along the segment for two individuals (NA18916 and NA19108). Red vertical bars below the graph denote the position of Alu sequences; the black rectangle denotes a Line1 insert. The 5S array is located between nucleotides 40,000 and 80,000. (B) Dot plot of the region. The plot displays segments of local similarity in the 5S rDNA arrays and adjacent sequences. One sequence is represented on each axis and significant matching regions are distributed along diagonals in the matrix. Green lines represent sequences that align on the forward strand and red lines represent for sequences that match on the reverse strand. The green box is indicative of a ∼20 kb low complexity region adjacent to the 5S array (the low complexity region is also observed in contigs from the CHM1tert PacBio assembly). Gene RNF187 is upstream the low complexity region and gene RHOU is downstream of the 5S array.

Fig. 4.—

Structure of the 45S rDNA array. (A) Profile of normalized read-depth (copy number) along the 45S rDNA repeating unit for two individuals (NA18916 and NA19108). (B) Dot plot for the 45S rDNA array unit (∼45 kb). We used a reference 45S rDNA unit with 45,337 bp nucleotides, which includes the promoter region, ETS1, 18S, ITS1, 5.8S, ITS2, 28S, ETS2, and IGS. Green lines represent sequences that match on the forward strand and red lines represent for sequences that match on the reverse strand. Lines off the main diagonal represent sequence similarity between different parts of the 45S rDNA repeat unit.

Fig. 5.—

Reproducibility of 45S rDNA contacts in experimental replicates and cell lines across 6 Hi-C datasets. Each dot represents the number of contacts identified in each chromosome. Red lines in the lower panels are loess smoothers. Upper panels show the spearman rank correlation between experiments. Red boxes show the correlation between replicates for the same enzyme. Yellow boxes show correlation between experiments using two different enzymes (HindIII and NcoI). Blue boxes represent correlations between the two cell lines (LCL genotype GM06990 and K562). All correlations are statistically significant (P < 0.001).

Fig. 6.—

rDNA contacts identified in LCL and K562 cells. (A) Contacts between the 5S array and each chromosome in experiments with two enzymes (Mbol and DpnII) and two cell lines (LCL and K562). (B) Contacts between the 45S array and each chromosome in experiments with two enzymes (Mbol and DpnII) and two cell lines (LCL and K562). Each dot represents the number of contacts identified in each chromosome. Red lines in the lower panels are loess smoothers. Upper panels show the spearman rank correlation between datasets. All correlations are statistically significant (P < 0.001).

Fig. 7.—

Variable number of 45S rDNA contacts in each human chromosome. The black bars represent the number of contacts observed between the 45S rDNA and the human genome (hg19) for each chromosome in two cell lines (LCL and K562). The orange bars represent the number of contacts expected between the 45S rDNA and the human genome (hg19) for each chromosome. The expected numbers were calculated from the total number of reads mapped to the rDNA in each cell line and the size of each chromosome. Chromosome 2 and chromosome 21 are significantly enriched for rDNA contacts in both cell lines (P < 0.001 in each cell line, Fisher’s exact test), whereas chromosome 16 and 17 are enriched in LCL and K562 cells (P < 0.001 in each case, Fisher’s exact test), respectively. Data summary for the 6 Hi-C experiments shown in figure 5.

Fig. 8.—

Visualization of rDNA contacts in specific segments and chromosomes. (A) The panel shows the segment in which the 5S rDNA arrays are located with no evidence of contact with the 45S rDNA arrays. The yellow box indicates the location of the 5S array. Genes RNF187 and RHOU are located upstream and downstream of the 5S rDNA arrays, respectively. (B) The panel shows contacts between the 45S rDNA and chromosome 21 for LCLs and K562 cells. Note that the rDNA contacts are dispersed across the entire chromosome. (C) The panel shows contacts between the 45S rDNA and a segment in chr9 (interval 150,000–47,520,051). Lanes show 45S rDNA contacts in K562 (two biological replicates shown below it) and in the LCLs (four biological replicates shown below it). (D) The panel shows 45S rDNA contacts on a segment in chromosome chr17 (interval 22,245,045–22,263,225) in the expanded view mode. Each blue bar represents evidence from one 45S rDNA contact. Note the higher number of contacts in K562.