DNA double-strand breaks (DSBs) are involved in many cellular mechanisms, including replication, transcription, and genome rearrangements. The recent observation that hot spots of DSBs in human chromosomes delimit DNA domains that possess coordinately expressed genes suggests a strong relationship between the organization of transcription patterns and hot spots of DSBs. In this study, we performed mapping of hot spots of DSBs in a human 43-kb ribosomal DNA (rDNA) repeated unit. We observed that rDNA units corresponded to the most fragile sites in human chromosomes and that these units possessed at least nine specific regions containing clusters of extremely frequently occurring DSBs, which were located exclusively in non-coding intergenic spacer (IGS) regions. The hot spots of DSBs corresponded to only a specific subset of DNase-hypersensitive sites, and coincided with CTCF, PARP1, and HNRNPA2B1 binding sites, and H3K4me3 marks. Our rDNA-4C data indicate that the regions of IGS containing the hot spots of DSBs often form contacts with specific regions in different chromosomes, including the pericentromeric regions, as well as regions that are characterized by H3K27ac and H3K4me3 marks, CTCF binding sites, ChIA-PET and RIP signals, and high levels of DSBs. The data suggest a strong link between chromosome breakage and several different mechanisms of epigenetic regulation of gene expression.

Genome instability plays a key role in multiple biological processes and diseases, including cancer. Genome-wide mapping of DNA double-strand breaks (DSBs) is important for understanding both chromosomal architecture and specific chromosomal regions at DSBs. We developed a method for precise genome-wide mapping of blunt-ended DSBs in human chromosomes, and observed non-random fragmentation and DSB hot spots. These hot spots are scattered along chromosomes and delimit protected 50–250 kb DNA domains. We found that about 30% of the domains (denoted forum domains) possess coordinately expressed genes and that PARP1 and HNRNPA2B1 specifically bind DNA sequences at the forum domain termini. Thus, our data suggest a novel type of gene regulation: a coordinated transcription or silencing of gene clusters delimited by DSB hot spots as well as PARP1 and HNRNPa2B1 binding sites. The study of chromosomal domains is hampered both by the complexity of chromosomal architecture and by the lack of experimental approaches for identification of structural and functional elements in chromosomes. The primary nucleosomal level in organization of chromosomes was detected by digestion of chromatin with exogenous nucleases. Many years ago, we detected 50–250 kb chromosomal domains that are released when genomic DNA is prepared under mild conditions, now known as forum domains. We assumed that this large-scale fragmentation of chromosomes by endogenous nucleases reflected some higher-order level of organization in chromosomes. However, both the existence of large 50–250 kb protected regions and short regions possessing hot spots of DNA breakage might reflect specific features in organization of chromosomes. Here, we used a procedure for genome-wide mapping of terminal regions of forum domains in human chromosomes. Our results show that hot spots of DNA breaks are not randomly distributed along chromosomes, that adjacent to these regions are binding sites of PARP1 and HNRNPA2B1, and that forum domains contain coordinately expressed gene clusters.

Human rDNA clusters form numerous contacts with different chromosomal regions as evidenced by chromosome conformation capture data. Heterochromatization of rDNA genes leads to heterochromatization in different chromosomal regions coupled with the activation of the transcription of genes related to differentiation. These data suggest a role for rDNA clusters in the regulation of many human genes. However, the genes that reside within the rDNA-contacting regions have not been identified. The purpose of this study was to detect and characterize the regions where rDNA clusters make frequent contacts and to identify and categorize genes located in these regions. We analyzed the regions that contact rDNA using 4C data and show that these regions are enriched with genes specifying transcription factors and non-coding RNAs involved in differentiation and development. The rDNA-contacting genes are involved in neuronal development and are associated with different cancers. Heat shock treatment led to dramatic changes in the pattern of rDNA-contacting sites, especially in the regions possessing long stretches of H3K27ac marks. Whole-genome analysis of rDNA-contacting sites revealed specific epigenetic marks and the transcription sites of 20–100 nt non-coding RNAs in these regions. The rDNA-contacting genes jointly regulate many genes that are involved in the control of transcription by RNA polymerase II and the development of neurons. Our data suggest a role for rDNA clusters in the differentiation of human cells and carcinogenesis.

Endogenous hot spots of DNA double-strand breaks (DSBs) are tightly linked with transcription patterns and cancer genomics^1,2. There are nine hot spots of DSBs located in human rDNA units^3–6. Here we describe that the profiles of these hot spots coincide with the profiles of γ-H2AX or H2AX, strongly suggesting a high level of in vivo breakage inside rDNA genes. The data were confirmed by microscopic observation of the largest γ-H2AX foci inside nucleoli in interphase chromosomes. In metaphase chromosomes, we observed that only some portion of rDNA clusters possess γ-H2AX foci and that all γ-H2AX foci co-localize with UBF-1 binding sites, which strongly suggests that only active rDNA units possess the hot spots of DSBs. Both γ-H2AX and UBF-1 are epigenetically inherited and thus indicate the rDNA units that were active in the previous cell cycle. These results have implications for diverse fields, including epigenetics and cancer genomics.

RNAi has been suggested for use in gene therapy of HIV/AIDS, but the main problem is that HIV-1 is highly variable and could escape attack from the small interfering RNAs (siRNAs) due to even single nucleotide substitutions in the potential targets. To exhaustively check the variability in selected RNA targets of HIV-1, we used ultra-deep sequencing of six regions of HIV-1 from the plasma of two independent cohorts of patients from Russia. Six RNAi targets were found that are invariable in 82%-97% of viruses in both cohorts and are located inside the domains specifying reverse transcriptase (RT), integrase, vpu, gp120, and p17. The analysis of mutation frequencies and their characteristics inside the targets suggests a likely role for APOBEC3G (apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like 3G, A3G) in G-to-A mutations and a predominant effect of RT biases in the detected variability of the virus. The lowest frequency of mutations was detected in the central part of all six targets. We also discovered that the identical RNAi targets are present in many HIV-1 strains from many countries and from all continents. The data are important for both the understanding of the patterns of HIV-1 mutability and properties of RT and for the development of gene therapy approaches using RNAi for the treatment of HIV/AIDS.

Chromosomes are organized into 3D structures that are important for the regulation of gene expression and differentiation. Important role in formation of inter-chromosome contacts play rDNA clusters that make up nucleoli. In the course of differentiation, heterochromatization of rDNA units in mouse cells is coupled with the repression or activation of different genes. Furthermore, the nucleoli of human cells shape the direct contacts with genes that are involved in differentiation and cancer. Here, we identified and categorized the genes located in the regions where rDNA clusters make frequent contacts. Using a 4C approach, we demonstrate that in Drosophila S2 cells, rDNA clusters form contacts with genes that are involved in chromosome organization and differentiation. Heat shock treatment induces changes in the contacts between nucleoli and hundreds of genes controlling morphogenesis. We show that nucleoli form contacts with regions that are enriched with active or repressive histone marks and where small non-coding RNAs are mapped. These data indicate that rDNA contacts are involved in the repression and activation of gene expression and that rDNA clusters orchestrate large groups of Drosophila genes involved in differentiation.