Rad21 is the core subunit of the cohesin complex involved in directing genome organization

doi:10.21203/rs.3.rs-1120640/v1

PDF

Article

Rad21 is the core subunit of the cohesin complex involved in directing genome organization

https://doi.org/10.21203/rs.3.rs-1120640/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted 14 Dec, 2021

You are reading this latest preprint version

The ring-shaped cohesin complex is an important factor regulating genome structure. It is thought to mediate the formation of chromatin loops and topologically associating domains (TADs) by loop extrusion. However, the regulation of association between cohesin and chromatin is poorly understood. In this study, we directly visualized cohesin loading after up-regulation of cohesin subunit Rad21 by identifying the formation of vermicelli-like structures via live cell super-resolution imaging. We also reveal that cohesin loading can be promoted by Rad21-loader interactions and accumulated contacts were shown at TAD corners while inter-TAD interactions increased after vermicelli formation, indicating that Rad21 is an important determinant of chromatin structure. Moreover, we find that cohesin saddle on topologically associating domains by FISH assay, which is consistent with the CTCF/cohesin-anchored chromatin loop model. Importantly, expression of Rad21 is strictly controlled, and aberrant expression of Rad21 leads to the formation of Rad21 “beads” in the nucleus. In summary, our observations provided important new biological insights into the mechanism of cohesin loading and its functions.

In eukaryotes, nuclear chromatin in the interphase is organized in a hierarchical manner. Previous studies using sequencing methods such as Hi-C detected genome-wide interactions, identified the plaid pattern of chromatin¹, and divided chromatin into A and B compartments based on their different eigenvectors. The large compartments are composed of smaller TADs in the 0.2–1.0 Mb range, which are distinguished by the presence of more abundant intra-TAD interactions compared with inter-TAD interactions². High-resolution Hi-C data revealed that chromatin within some TADs is organized into small, stable contact domains anchored by CTCF/cohesin, known as loop domains^3,4.

The insulator protein CTCF and ring-shaped cohesin complex (also referred as “cohesin” below) were found to be located in the boundaries of TADs and loop domains ^2,3,5. In particular, it was demonstrated that CTCF and cohesin are important for the maintenance of loop domains using rapid protein depletion assays based on auxin-inducible degron tagging. These assays found a loss of insulation at most TAD boundaries after CTCF depletion while loop domains rapidly disappeared after cohesin degradation^4,6. Previous studies have suggested that cohesin-DNA interaction plays an important architectural role in determining chromatin structure during the interphase. Multiple studies, including simulation, in vitro biochemical reconstitution, and single molecule imaging studies, have supported the loop extrusion model to explain the formation of loops and TADs as a process mediated by cohesin progressively extruding DNA until it encounters blocks formed by convergent dimeric CTCF^7–12.

The cohesin complex was originally identified as a tripartite ring playing a role in sister chromatid cohesion in the metaphase, which is constructed by the combination of a V-shaped SMC1/3 heterodimer and the kleisin subunit Rad21^13,14. The association between chromatin and cohesin can be regulated by many factors. For example, cohesin antagonist Wapl can drive the release of cohesin from chromatin^15–17. Tedeschi et al. depleted Wapl to stabilize the cohesin complex on chromatin and found a vermicelli-like distribution of chromatin, indicating that cohesin plays a role in chromatin condensation¹⁸. Moreover, high-resolution Hi-C data showed that chromatin loops are enlarged in cells with Wapl deficiency¹⁹, suggesting that the formation of the vermicelli-like distribution is likely caused by excessive chromatin extrusion with the help of over-loaded cohesin. Inversely, PDS5 is necessary for the maintenance of cohesin²⁰. A loading complex called cohesin loader complex NIPBL-MAU2 (also known as SCC2/SCC4, respectively) is required for cohesin loading prior to DNA replication¹⁷. The loading complex is composed of the head, body and hook domains, and the latter two are sufficient to facilitate cohesin loading²¹. NIPBL was also shown to promote cohesin loading onto circular DNA topologically in vitro by direct interaction at multiple sites on cohesin subunits²². After the loading process, the loading complex is found to hop between different cohesin molecules rapidly in living cells²³. Using cryo-electron microscopy, a recent study revealed the structure of cohesin bound to NIPBL and DNA, which provided the structural basis of cohesin-NIPBL-dependent DNA entrapment. However, the manner in which cohesin and NIPBL work cooperatively to promote chromatin condensation is unclear.

Here, we found that up-regulated expression of Rad21, but not other cohesin subunits, leads to excessive chromatin condensation with a vermicelli-like morphology. To investigate the mechanism of cohesin loading and extrusion in vivo, we used super-resolution imaging and single-molecule imaging to measure the distribution of cohesin subunits. Fluorescence recovery after photo-bleaching and Co-IP results showed that Rad21-loader interaction promotes the formation of a vermicelli-like distribution by facilitating the cohesin loading process. Hi-C data showed that the short distance contact frequency increased in the presence of Rad21 up-regulation, while the long-distance contact frequency decreased. Upon Rad21 up-regulation, chromatin interactions between A and B compartments increased and compartmentalization strength was reduced significantly. Correspondingly, accumulated contacts were shown at TAD corners and inter-TAD interactions increased. These observations support a cohesin loop extrusion model and lead us to propose the unique role of Rad21-loader interaction in cohesin loading and further cohesin extrusion. Moreover, we combined transcriptome changes and chromatin structure alterations upon Rad21 up-regulation and revealed that Rad21 plays a role in breast cancer by affecting chromatin architecture.

Cohesin is arranged in axial chromosomal domains in cells with up-regulated Rad21

To investigate the contributions of different cohesin subunits in the cohesin extruding process, we tagged the cohesin subunits SMC1A, SMC3, Rad21, SA1 and SA2 with HaloTag in HeLa cells. The HaloTag-fused cohesin subunits were fluorescently labeled by Janelia Fluor 549 (JF549) in living cells and imaged by a super-resolution spinning disk confocal system (live SR CSU W1 Nikon). We observed that Rad21, as well as chromatin DNA, were both arranged in a vermicelli-like pattern in cells upon Rad21 up-regulation (Fig. 1a and Supplementary information, Fig. S1a, b), suggesting that excessive Rad21 could cause chromatin DNA condensation. In contrast, over-expression (OE) of the other four cohesin subunits, SMC1A, SMC3, SA1 and SA2, resulted in even distributions (Supplementary information, Fig. S1c). The super-resolution images of the vermicelli pattern of Rad21 revealed a “beads on a string” pattern (Fig. 1a), which indicates that Rad21 forms clusters on chromatin. Interestingly, the Rad21 “beads” were distributed regularly along chromatin with an average inter-bead distance of 0.34 µm (Fig. 1b). To elucidate the relationship between vermicelli formation and up-regulated Rad21, we observed the distribution pattern of Rad21 in live cells that expressed exogenous Rad21 over different time courses. The data showed that the extent of DNA condensation (Fig. 1c) was dependent on the expression level of Rad21. We also calculated the heterogeneity level, which revealed that the distribution of Rad21 became more heterogeneous as its expression level increased (Fig. 1d).

To examine whether Rad21 in the vermicelli are bound to chromatin, we performed inversed fluorescence recovery after photobleaching (iFRAP) experiments and observed that the fluorescence intensity of the unbleached region decayed very slowly (Fig. 1e, f). These results suggested that Rad21 binds stably on chromatin in the vermicelli.

To clarify the integrity of the cohesin ring structure bound to chromatin, SMC1A, SMC3 and SA1 were immuno-stained in the absence and presence of Rad21 up-regulation. We found that endogenous H3 were distributed homogeneously, while SMC1A and SMC3 were colocalized with Rad21 and enriched in vermicelli-like structures after Rad21 up-regulation, which confirmed that the vermicelli-like distribution was caused by the whole cohesin complex rather than Rad21 alone (Supplementary information, Fig. S1d and Fig. 1g, h). In addition, SA1 was less enriched in the vermicelli-like structures in comparison with SMC1A and SMC3, which is in line with the finding that SA1 does not participate in the construction of ring structures directly^14,24. To further dissect the underlying molecular mechanism, we constructed three human Rad21 mutants (F597R, L601R and Q613K) (Supplementary information, Fig. S1e) that are unable to associate stably with SMC1/3 dimers as identified in fission yeast²⁵. We found that none of the three Rad21 mutants could condense chromatin into vermicelli-like structures (Fig. 1i). Moreover, iFRAP led to a rapid decrease in the fluorescence intensity of the whole cell, indicating that these Rad21 mutants could not bind to chromatin without the formation of a cohesin complex (Supplementary information, Fig. S1f). Taken together, these results also suggest that formation of the vermicelli requires the whole cohesin complex.

Rad21-loader interaction facilitates vermicelli-like structure formation by promoting cohesin loading

The vermicelli-like phenotype observed upon Rad21 up-regulation could have two explanations. First, Rad21 may act as the limiting factor for cohesin formation so that up-regulation of Rad21 leads to an increased pool of cohesin. Alternatively, Rad21 may promote cohesin loading on chromatin and thus bias the loading/unloading balance of cohesin for excessive condensation of chromatin. To distinguish between these two possibilities, we performed salt extraction (Materials and Methods) to quantitatively measure the amount of cohesin that are mobile or bound weakly to chromatin (unloaded cohesin) and that bound stably (loaded cohesin).

Western blotting data showed that up-regulation of Rad21 led to reduced expression of the endogenous Rad21 (Fig. 2a). Importantly, we found that the fraction of SMC3 bound to chromatin increased upon Rad21 up-regulation, while the fractions of bound and unbound Rad21 remain nearly unchanged (Fig. 2a, b). When cohesin complex were isolated from whole cell extract by immunoprecipitation against SMC3 in the present or the absent of Rad21 over-expression (Supplementary information, Fig. S2a), we found that up-regulation of Rad21 increased the amount of Rad21 by 2.5-fold while the total amount of cohesin showed almost no difference (Fig. 2c). Moreover, salt extraction and subsequent co-IP experiments showed that Rad21 up-regulation increased loaded cohesin, while the abundance of unloaded cohesin decreased (Fig. 2d). As shown in lanes 6 and 7, excessive Rad21 did not increase the proportion of cohesin in the unbound SMC3 pool. Considering that the amount of SMC3 with weak chromatin binding decreased, as shown in Fig. 2a, we inferred that up-regulated Rad21 reduced the amount of unloaded cohesin. As shown in lanes 11 and 12, the proportion of cohesin in the chromatin-bound SMC3 pool was significantly elevated in Rad21-OE cells. These data collectively suggested that excessive Rad21 promoted cohesin loading on chromatin rather than increasing the abundance of the cohesin complex.

To further support the hypothesis that excessive Rad21 increases Rad21-loader interaction and thus promotes cohesin loading on chromatin, we constructed Rad21 mutants with disrupted interaction with the loader. Previous studies identified the Mis4Scc2–Ssl3Scc4 interaction site between cohesin subunits in S. pombe^22,26. To test this hypothesis in human cells, we constructed cells with Rad21-Loader-Interaction-Site (LIS) deletion and poly-Ala mutations (Supplementary information, Fig. S2b). Salt extraction and subsequent co-IP experiments showed that LIS-Ala mutants accumulated more unbound Rad21 molecules in comparison with wild-type cells and Rad21 over-expressing HeLa cells, and also showed a lack of augmentation of cohesin loading without a significant difference in cohesin formation (Fig. 2d). These results suggested that cohesin loading accumulation is dependent on Rad21-loader interaction, and cohesin with LIS-mutant-Rad21 functions as an intact complex before loading onto chromatin. FRAP experiments further showed that the vermicelli pattern disappeared when LIS of Rad21 was deleted or replaced by poly-Ala in HeLa cells with Rad21-GFP over-expression (Fig. 2d and Supplementary information, Fig. S2c), while mutant LIS1 showed less modulation compared with LIS2 in FRAP assays (Fig. 2e, f).

To test whether SMC1A and SMC3-loader interaction dysfunction could disrupt Rad21-mediated cohesin loading, we constructed LIS poly-Ala mutants for SMC1A and SMC3, respectively (Supplementary information, Fig. S2d). In such conditions, the Rad21-vermicelli pattern persisted and colocalized with mutant SMC1A, but not mutant SMC3, which suggested that SMC1A interaction with the loader is dispensable for the cohesin loading process (Fig. 2g).

To quantify the cohesin loading capability of cohesin subunits, we applied single molecule tracking of Rad21, with H2A as a reference, in living HeLa cells. For single-molecule imaging, we used highly inclined illumination microscopy to reduce background noise. The data showed that the majority of cohesin subunit molecules were stationary and only a small fraction was mobile, suggesting that up-regulated Rad21 was prone to bound stably to chromatin (Supplementary information, Fig. S2e).

Excessively loaded cohesin are anchored by CTCF and saddle on topologically associating domains

Previous studies have suggested that cohesin complexes condense chromatin DNA by progressively extruding DNA with the help of NIPBL and MAU2²³, meanwhile CTCF works as the boundary to halt the extrusion process resulting in the appearance of TADs. Notably, CTCF was enriched in vermicelli-like structures and colocalized with Rad21, which confirmed the association between CTCF and cohesin when extruding loops met their ends at CTCF sites (Fig. 3a, b). The distributions of NIPBL and MAU2 showed no significant difference between wild-type and RAD21-OE cells (Fig. 3a and Supplementary information, Fig. S3a), which was consistent with the dynamic motion of NIPBL during cohesin loading. We also calculated the co-localization ratio of CTCF and Rad21, and we found that CTCF was partially enriched in vermicelli-like structures, while most Rad21 was co-localized with CTCF (Fig. 3c), which corresponded with the ability of CTCF to function as a boundary element when chromatin over-extrusion occurs.

To visualize how TADs are organized in vermicelli-like structures, three TADs (EMC7 chr7:5260000-5860000, ACTB chr7:5260000-5860000, CD28 chr2:203600000-204400000) were labeled after Rad21 over-expression using fluorescence in situ hybridization (FISH) (Materials and Methods). The Rad21 “beads” were found to be localized at the middle of a TAD (Fig. 3d and Supplementary information, Fig. S3b). Interestingly, the normalized intensity profile of the boxed region measured across the center of the EMC7 TAD and Rad21 (n=20 per condition) detected that excessive cohesin saddled on TADs (Fig. 3e), which is consistent with the CTCF/cohesin-anchored chromatin loop model^{4,7,8,12,19,24,27,28} (Fig. 3f). We also calculated the volume of EMC7 TAD and found that the median volume of TAD became larger in the presence of Rad21 up-regulation (Fig. 3g), suggesting that a larger range of intra-TAD interactions resulted from chromatin over-extrusion. Furthermore, we applied 1-hr EdU-Cy5 metabolic labeling^29,30 and found that the area of EdU clusters enlarged significantly after Rad21 was up-regulated (Supplementary information, Fig. S3c, d), which indicated overall enlargement of chromatin domains. To measure the 3D genome organizational change at a larger scale, we performed chromosome painting assays. Chromosome 2 was found to exhibit a fiber-like pattern in the presence of Rad21 over-expression (Fig. 3h, i), indicating that a global conformational effect within a single chromosome was conferred by over-loaded cohesin. We also found that the volume of chromosome 2 became slightly larger after Rad21 over-expression during the formation of vermicelli-like structures (Supplementary information, Fig. S3e). Taken together, these data suggest that over-loaded cohesin could facilitate the formation of larger TADs through the processive extrusion of nearby small TADs and therefore affect macro-scale chromosome organization.

Rad21 up-regulation affects genome contacts by stimulating cohesin activity

To study the effect of Rad21 over-expression on genome organization, we performed Hi-C assays in HeLa cells in the presence and absence of Rad21 up-regulation to study the macro-scale consequences on chromosome architecture. The relative contact probability indicated that the genome-wide/global chromosomal contact frequency increased for short distance interactions (0.3–6 Mb) and decreased for long distance interactions (>6 Mb) in Rad21-OE cells (Fig. 4a). For intra-chromosomal interactions, taking chromosome 2 as an example, Hi-C contact matrices showed that the number of far-cis contacts markedly decreased after Rad21 was up-regulated, whereas the number of short-distance contacts increased (Fig. 4b). The inter-chromosome interaction ratio was significantly decreased after Rad21 up-regulation (Supplementary information, Fig. S4a). These findings are in line with the notion that chromosome condensation and segregation occur with the formation of vermicelli-like structures in Rad21-OE cells.

We next explored the effect of Rad21 over-expression on nuclear compartmentalization, namely compartments A and B, which are defined by the first principal component (PC1) of Hi-C correlation matrices. In Rad21 up-regulated cells, we detected decreased intensity of “checkerboard” patterns, and the distribution of the PC1 scores was less strongly bimodal (Fig. 4c), suggesting a decrease in the compartmentalization level of chromosomes in Rad21-OE cells. To quantify the change of compartmentalization for each chromosome, we calculated the ratio of the interaction frequency between different classes of compartments (AB) versus that between the same classes of compartments (AA and BB)³¹. These ratios were significantly increased in Rad21-OE cells (Fig. 4d), indicating less strict segregation between the A and B compartments. We also calculated the average contact frequency enrichment of compartments ranked by PC1 values, which indicated increasing contact enrichment between different compartment categories and decreasing contact enrichment between similar compartment categories (Fig. 4e). Moreover, 1.9% of genomic regions switched from the A compartment to the B compartment after Rad21 was up-regulated, while 4.3% of genomic regions exhibited the opposite change (Fig. 4f).

We further applied an imaging approach to investigate the effect of Rad21 over-expression on chromatin structure. To visualize the effects of Rad21 over-expression on key epigenomic features associated with genome folding directly, we labeled two histone modifications, histone H3 lysine 4 trimethylation (H3K4me3) and histone H3 lysine 27 trimethylation (H3K27me3), which are associated with active and repressive chromatin in compartments A and B, respectively¹^,³^,³². Rad21 over-expression had little effect on these features (Fig. 4g). Notably, the distribution of HP1α was unchanged in the Rad21-OE cells, indicating that the formation of vermicelli-like structures was not dependent on heterochromatin (Fig. 4h).

To obtain a better understanding of the increase in the physical volume of TADs after Rad21 was up-regulated (Fig. 3g), we performed TAD calling on whole-genome contact maps and calculated TAD length. Notably, there was no significant change in TAD length after Rad21 was up-regulated (Supplementary information, Fig. S4b). In addition, no significant difference was observed in the number of long (>500 kb) or short (<500 kb) TADs (Supplementary information, Fig. S4c), and the insulation scores of TAD borders were similar (Supplementary information, Fig. S4d). These findings indicate that the insulation potential remained unchanged following up-regulation of Rad21. Therefore, over-loaded cohesin could lessen TADs and thus increase the accessibility of their associated chromatin.

To further probe how Rad21 over-expression regulates TAD structure, we visualized the interaction heat map at 40 kb resolution. After Rad21 was over-expressed, Hi-C contacts accumulated in TAD corners where CTCF binding sites are located (Fig. 4i). Aggregated TAD analysis (ATA) showed that this effect was global (Fig. 4j). Interestingly, the increase in the signal at the corners of TADs was accompanied by a decreased TAD score (contact frequency ratio between intra-TADs and inter-TADs) (Fig. 4k). Taken together, our results suggest that over-loaded cohesin facilitates the formation of chromatin loops that span the maximal distance within TADs through processive extrusion, which led to increased inter-TAD contact frequency.

Rad21 up-regulation is involved in breast cancer by affecting chromatin architecture

Next, we explored the pathological relevance of genome structural disorders caused by Rad21 up-regulation. We first studied clinical data from The Cancer Genome Atlas (TCGA). Through Kaplan-Meier survival analysis, we found that patients with higher Rad21 expression levels did not survive as long as patients with low levels (Fig. 5a). In addition, the Rad21 expression levels in the breast cancer tissues of patients with four cancer subtypes were significantly higher than those measured in normal tissues (Fig. 5b). Moreover, several studies have found that enhanced Rad21 transcription is correlated with increased gene copy number in breast cancer tissue, which is associated with poor prognosis and resistance to chemotherapy³³^,³⁴. Another study proposed that Rad21 could be a target for breast cancer drugs, because Rad21-targeted siRNA increased the sensitivity of cells to two chemotherapeutic drugs³⁵. Therefore, we explored the relationship between genome organization disorders associated with enhanced Rad21 expression and breast cancer oncogenesis and prognosis.

To assess Rad21 aggregation in breast cancer cells with high Rad21 expression levels, we obtained three human breast cancer cell lines with high levels of Rad21 expression³⁴^,³⁵, and the MCF10A cell line was used as a reference. The Rad21 level of each cell line was measured by western blotting (Supplementary information, Fig. S5a). Super-resolution imaging was used to reveal the endogenous distribution of Rad21. Compared with MCF10A cells, Rad21 was more aggregated and showed a high level of heterogeneity in the three selected breast cancer cell lines, indicating a tendency for the formation of vermicelli-like structures (Fig. 5c, d and Supplementary information, Fig. S5b).

To explore the influence of Rad21 up-regulation, we analyzed gene expression in control and Rad21-OE cells. This analysis revealed 649 significantly up-regulated (fold change > 2, p-value < 0.05) and 138 down-regulated genes (fold change < 0.5, p-value < 0.05) in Rad21-OE cells (Fig. 5e and Supplementary information, S5c). The top 8 Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways for differentially expressed genes were obtained (Supplementary information, Fig. S5d). Among the top 8 KEGG pathways, cytokine-cytokine receptor interaction, the TNF signaling pathway and the NOD-like receptor signaling pathway are known to play roles in breast cancer³⁶^-³⁸. In addition, to obtain insight into the relationship between up-regulated Rad21 and breast cancer, we subjected breast cancer cell lines to gene set enrichment analysis (GSEA). The GSEA identified significant enrichment of mutant TP53, over-expressed KRAS and up-regulated EGFR target genes following Rad21 up-regulation (Supplementary information, Fig. S5e). These gene sets contain several genes that have been widely studied in the context of cancer, including WNT16, CEBPD and EGFR.

Additional analysis was performed to clarify the association between genome structural disorders caused by vermicelli-like structures and differential gene expression. We found that the compartment switch from B to A coincided with higher expression levels, while the switch from compartment A to B was not associated with a significant change (Fig. 5f). One of the up-regulated genes in the compartment B-to-A switch region was cell migration inducing hyaluronidase 1 (CEMIP), which has been reported to promote the proliferation and migration of breast cancer cells³⁹^,⁴⁰. A high CEMIP protein expression level is significantly associated with poor patient survival⁴¹. These results suggest that compartment switching as a result of Rad21 up-regulation can lead to up-regulation of cancer-related genes.

In addition, TAD insulation was also correlated with gene expression. We calculated the TAD score within differentially and stably expressed genes, revealing that both up-regulated and down-regulated genes had reduced TAD score. These results indicated that stronger inter-TAD interactions could affect the expression of genes within the interacting domains (Fig. 5g). These results revealed the connection between breast cancer and genome structural changes upon Rad21 up-regulation.

Here, we provide key insights into the molecular mechanism by which Rad21 facilitates the cohesin loading process and thus directing genome organization using live cell super-resolution imaging and biochemical approaches. We observed that excess Rad21 led to intensive chromatin condensation and vermicelli-like distribution of cohesin. Importantly, we revealed that this process was driven by enhanced loading of the cohesin complex, while the amount of unloaded cohesin was unaffected. These results suggest that Rad21 acts as a crucial factor in cohesin loading. Mechanistically, Rad21 mutants that are unable to bind the cohesin loader complex cannot lead to the vermicelli-like pattern, indicating that Rad21-loader interaction is an important determinant in the cohesin loading process. After Rad21 up-regulation, accumulated contacts were shown at TAD corners and inter-TAD interactions increased. At the subchromosome scale, chromatin interactions between A and B compartments increased and compartmentalization strength was reduced significantly (Fig. 6).

According to previous studies, cohesin was found to form a thread-like distribution that has been described as “vermicelli-like”, and chromatin condensation was shown to be enhanced in mouse embryonic fibroblasts (MEFs) after cohesin was stabilized on chromatin by Wapl depletion. In contrast, our findings showed that over-expression of cohesin subunit Rad21 led to accumulation of cohesin on chromatin and a vermicelli thread-like cohesin distribution in the nucleus of human HeLa cells, while cohesin release from chromatin mediated by Wapl remained unchanged. We therefore inferred that Rad21 is an important factor that affects cohesin loading independent of Wapl.

Previous studies have established that the loader complex SCC2/SCC4 is essential for cohesin loading onto chromatin^17,42−44. In addition, Murayama and Uhlmann showed that Rad21 interacts directly with the loader complex in S. pombe using tiling peptide arrays²². In vitro translated Scc1 was found to bind GST-Scc2, and the binding region of Scc1 was mapped to residues 126–230⁴⁵. The different impact of the LIS mutant of each cohesin subunit suggests that Rad21-LIS is crucial for cohesin loading on DNA and further activation. A recent cryo-EM study of the human cohesin-NIPBL-DNA complex shed new light on the structural details of Rad21-loader interactions²⁴. The cohesin-NIPBL-DNA complex consists of the V-shaped SMC1-SMC3 heterodimer, the U-shaped NIPBL and the SMC1-SMC3 hinge domains with or without SA1. The complex is stabilized by DNA bound at the central tunnel and surrounding flexible Rad21. The LIS of cohesin is missing from the structure with the exception of LIS2 of Rad21, which mediates Rad21-MAU2 interaction. Rad21 LIS2 is bound by U-shaped SA1 and is located close to the DNA binding interface of SA1, while the missing Rad21 LIS1 is close to the external surface of the NIPBL HEAT repeat (R6, R7, R10), which directly interacts with DNA via its internal surface. Our colocalization results (Supplementary information, Fig. S1d) indicate that SA1 is not a constitutive component of cohesin, in line with poorly resolved cryo-EM results, which may due to its low binding occupancy.

In addition, Rad21-loader specific interaction may be important in cohesin regulation. For example, recombinant Chaetomium thermophilum (Ct) Pds5 competes with Ct Scc2 for binding to Scc1 in vitro in a dose-dependent manner, indicating that it possesses the ability to release Scc2 for the next round of cohesin loading⁴⁵. It was also reported that most (16 of 19) Ct Scc2 mutants were deficient in binding to the N-terminal of Ct Scc1in patients with the human developmental disorder termed Cornelia de Lange syndrome (CdLS)⁴⁵. Considering our findings, it is possible that Scc2 mutants impair cohesin loading and disrupt enhancer-promoter chromatin loop formation at genes with important functions in development.

It is widely accepted that CTCF and cohesin work as boundary elements for chromatin loop extrusion based on the observations that many CTCF binding sites are enriched in the bases of chromatin loop domains in GM12878 cells³ and the boundary regions of TADs in IMR90 cells². The association of CTCF and cohesin clusters was demonstrated by two-color direct stochastic optical reconstruction microscopy (dSTORM) super-resolution imaging in mES cells, in which these clusters overlap were found to largely overlap⁴⁶. Consistent with this study, our results showed colocalization of CTCF and cohesin in the presence of vermicelli-like patterns. Furthermore, cohesin was found to be essential for the formation and maintenance of TADs and loops. For example, the loop domains disappeared after cohesin was depleted by auxin, and most loop domains recovered in 1 h after auxin was withdrawn in the auxin-inducible degron HCT-116 cell line⁴. Remarkably, Both TADs and loops were greatly reduced after 180 min of auxin treatment in Scc1-mEGFP-AID HeLa cells²⁷. Our results showed the distribution of cohesin and TADs directly in the nucleus for the first time. We found that cohesin complexes were enriched in the bases of TADs, which was consistent with the hypothesized model that multiple loop-extrusion cohesin complexes extrude longer chromatin loops constantly within the TAD, the complexes were halted by CTCF binding in a convergent direction at the boundary region of TADs⁴⁷. Longer loops also led to stronger inter-TAD interactions after Rad21 up-regulation.

A previous chromosome spreading study using yeast revealed the presence of 100 or more Scc1 foci⁴⁸. In another study, quantitative western blotting revealed 5–20 cohesin complexes at each cohesin binding site, indicating that cohesin may associate with chromatin in clusters⁴⁹. Similarly, 5–15 cohesin molecules were found in cohesin clusters in mES cells by live cell STED imaging⁵⁰. Recently, a new type of phase separation mediated by the SMC protein of the cohesin complex, termed bridging-induced phase separation (BIPS), was reported in a study that showed formation of cohesin-DNA clusters after the application of 10 nM yeast cohesin holocomplexes to double-tethered λDNA. Following the application of cohesin holocomplexes, cluster fusion and rapid recovery after photobleaching were observed, which suggested liquid-like behavior⁵⁰. Consistent with the studies described above, we observed the formation of bead-like clusters in living cells following Rad21 over-expression, and these clusters were located close to TADs labeled by FISH. It is possible that the accumulated cohesin complexes at the boundary regions of TADs mediate the formation and maintenance of chromatin loop domains in the interphase by self-association. However, whether the distribution of cohesin complexes at TAD boundaries is liquid-like cluster or vermicelli-like under physiological conditions is unknown.

Somatic mutations and amplification of Rad21 have been widely reported in both human solid and hematopoietic tumors. Rad21 alterations are relatively common according to TCGA PanCancer atlas studies (7% of all queried patients), which show elevated Rad21 in 20% of ovarian cancer cases and 13% of breast cancer cases. Over-expression of Rad21 has been linked with epithelial breast cancer and has been found to be correlated with poor disease outcome and resistance to chemotherapy. Although the correlation between Rad21 expression and cancer risk is relatively well established, little is known regarding the causes and consequences of Rad21 over-expression in tumorigenesis. In our study, we monitored the outcomes of Rad21 over-expression in HeLa cells and compared the effects of elevated Rad21 on gene expression with TCGA BRCA studies. Unfortunately, we were unable to generate a stable Rad21 over-expression cell line, perhaps due to activation of the TNF-α apoptosis signaling pathway. By comparing transcriptome changes and chromatin structure alterations upon Rad21 up-regulation, we demonstrated that the relationship between Rad21 up-regulation and breast cancer is likely mediated by changes in chromatin architecture, and we found that chromatin alterations and transcriptional changes following Rad-21 up-regulation were consistent. Taken together, our findings suggest that Rad21 plays a key role in regulating chromatin structure and acts as an oncogene. This study provides important information describing the relationship between chromatin status and the etiology of cancers, which may reveal new targets for clinic treatments.

Cell Culture

The human HeLa-S3 immortalized cell line (Catalog No. 30-2004) and human SK-BR-3 cell line (Catalog No. 30-2007) were grown at 37°C with 5% CO2 in high-glucose DMEM (Thermo Scientific, 11965084) containing 10% (v/v) fetal bovine serum (HyClone, SV30087.02) and 100 U/mL Penicillin Streptomycin (Thermo Scientific,15140163). The human MCF10A cell line (Catalog No. CC-3150) was grown at 37°C with 5% CO2 in MEGM Mammary Epithelial Cell Growth Medium (Lonza, CC-3150) containing supplements required for growth. The human MDA-MB-157 cell line (Catalog No. 30-2008) was grown at 37 ℃ without CO2 equilibration in Leibovitz's L-15 Medium (Thermo Scientific, 11415-064) supplemented with 10% (v/v) fetal bovine serum and 100 U/mL Penicillin Streptomycin. The human HCC1395 cell line was cultured in RPMI 1640 Medium (Thermo Scientific,11875093) containing 10% (v/v) fetal bovine serum and 100 U/mL Penicillin Streptomycin at 37°C with 5% CO2.

Live cell imaging and analysis

The Human SMC1A, SMC3, Rad21, and SA1/2 genes were tagged with Halo tag or EGFP and cloned into the pcDNA3.1 vector by Gibson assembly. Cells were plated in glass-bottom dishes, transfected with 2 μg of the indicated plasmids, and allowed to grow for 24–36 h. Cells were then labelled with 1 μM JF549 (Promega, 6147/5) for more than 15 min and rinsed twice with PBS before imaging. The dishes were then incubated at 37 ℃ with 5% CO₂ in the incubation chamber of the microscope. Super resolution live-cell imaging was performed using a spinning disk confocal system (Nikon, live SR CSU W) with an EMCCD (iXon DU-897E) mounted on a Nikon Ti-E microscope with a CFI Apo TIRF 100 × Oil (N.A. 1.49) objective.

For FRAP assays, the cells were transfected with 2 μg of plasmids containing a gene labelled by an EGFP tag for 24–36 h. One pre-bleach image was acquired, after which one half of a nucleus was bleached with the 488 nm laser (100% laser power), and recovery images were acquired by a confocal microscope (Carl ZEISS, LSM880).

For single molecule imaging assay, cells were labelled with 0.1 nM JF549 for more than 15 min and rinsed twice with PBS before imaging. In order to collect a sparse single molecule signal, cells were photo-bleached with a strong laser. The time-lapse imaging data was acquired on a custom-built microscope (Olympus, IX83) with a 10 ms exposure time.

The single-molecule signal was recognized and tracked using the FIJI (Fiji Is Just ImageJ) plugin for Trackmate software. The diffusion coefficient was then obtained as previously described⁵¹ using the following formula:

where Δx and Δy represent the step size during the time interval Δt.

The distribution of the diffusion coefficient value was fitted with a Gaussian function using Origin software.

Immunofluorescence

Cells were fixed using 4% (w/v) paraformaldehyde (Electron Microscopy Sciences, 157-8) in PBS (Thermo Scientific, 14190144) for 15 min, followed by three washes in PBS. Permeabilization of cells was performed using 0.5% Triton X-100 (Sigma-Aldrich, T8787-50 ML) in PBS for 10 min, followed by blocking with 5% IgG-free bovine serum albumin (BSA, Jackson, 001-000-162) for 30 min. The indicated primary antibody was added to 5% BSA in PBS at an appropriate dilution and incubated overnight at 4 ℃. After 3 rinses in PBS, cells were incubated with a secondary antibody labeled with an Alexa Fluor dye at a dilution of 1:200 in PBS with 5% BSA for 1 h. The cells were then rinsed with PBS 3 times and fixed using 4% PFA in PBS for 10 min. After 2 rinses in PBS, the cells were finally stored in PBS. Imaging data were acquired using a spinning disk confocal system (Nikon, live SR CSU W1) with an EMCCD (iXon DU-897E) mounted on a Nikon Ti-E microscope with a CFI Apo TIRF 100 × Oil (N.A. 1.49) objective.

For STED ( Leica, TCS SP8 STED 3X) imaging, the sample preparation method was identical to the procedure described above, except that the secondary antibody was a donkey anti-Rabbit lgG secondary antibody (Alexa Fluor 594).

ANTIBODY	IDENTIFIER	SOURCE
Anti-Histone H2B	ab1790	abcam
Anti-Histone H3	ab1791	abcam
Anti-Histone H3 (tri methyl K4)	ab8580	abcam
Anti-Histone H3 (tri methyl K27)	ab6002	abcam
Anti-CTCF	ab128873	abcam
Anti-SMC1A	ab133643	abcam
Anti-SMC3	ab9263	abcam
Anti-SA1	ab4455	abcam
Anti-NIPBL	ab220952	abcam
Anti-Scc4	ab183033	abcam

ANTIBODY	IDENTIFIER	SOURCE
Donkey anti-Rabbit IgG (H+L) Secondary Antibody, Alexa Fluor 488	A-21206	Thermo Scientific
Donkey anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 488	A-21202	Thermo Scientific
Donkey anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 594	A-21207	Thermo Scientific
Donkey anti-Mouse IgG (H+L) ReadyProbes™ Secondary Antibody, Alexa Fluor 594	R37115	Thermo Scientific

Heterogeneity level analysis

The nucleoli and extranuclear regions were manually selected from the image and the intensity (gray value) of these parts was set to 0. The average intensity of each image was set to the same value to eliminate cell heterogeneity. Considering the radial distance of the vermicelli-like structures was approximately 5 pixels, we used a 5 × 5-pixel box traversing the image and calculated the intensity of the box. If the intensity was 0, we discarded the value, which represented the nucleoli or extranuclear region. To quantify the intensity of different samples, we fitted the distribution of intensity values with a double Gaussian function. The two Gaussian peaks represented the distributions of the sparse regions and dense regions respectively. The heterogeneity level was acquired by calculating the distance between the two peaks. A custom-written MATLAB program was used for the analysis.

Co-immunoprecipitation (co-IP) assay

GFP-positive HeLa cells transfected with the indicated plasmids were collected using FACS (BeckMan Coulter, Astrios EQ). Generally, 10⁶ cells were sorted, washed twice with PBS, and lysed in a buffer containing 20 mM Tris-HCl (pH 7.5), 100 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.5% NP-40, 0.1%SDS and protease inhibitors. After incubation at 4 °C for 10 min, the soluble fraction (containing 0.1 M NaCl) was separated from the chromatin fraction by centrifugation at 15,000 × g for 10 min. The chromatin pellet was then incubated in the above buffer with the addition of 0.5 M NaCl at 4 °C for 1 h, followed by centrifugation at 15,000 × g for 10 min. The final pellet was resuspended in 1 M NaCl of the above buffer, sonicated for 20 s (1 s on and 1 s off), incubated at 4 °C for 1 h, and then centrifuged. These lysate fractions were subjected to Co-IP with an antibody against SMC3 (abcam, ab9263). Further immunoblot analyses were performed with the following antibodies: H3 (Ruiying Biological Technology, RLM-3038), ACTB (ABclonal, AC026), Rad21 (Thermo Scientific, PA5-54128), and SMC3 (abcam, ab9263).

Chromosome painting

Cells were plated on a coverslip, transfected with 2 μg of the indicated plasmids, and allowed to grow for 24 h. After fixation in 3:1 (v/v) methanol/acetic-acid, the cells were washed 3 times in PBS. Cell permeabilization was performed using 0.5% Triton X-100 (Sigma-Aldrich, T8787-50 ML) in PBS for 10 min. Next, 10 μL of a commercial probe mixture was applied (Sigma-Aldrich, D-0302-100-FI XCP 2), and the slide was sealed with rubber cement. The cells and probe were denatured by heating the slide on a hotplate at 75 ℃ for 2 min, followed by hybridization in a humidified chamber at 37 ℃ overnight. The slide was then removed from the chamber, and the coverslip was washed in 0.4× SSC (Sigma-Aldrich, S6639-1L) (pH=7.0) at 72 ℃ for 2 min. The coverslip was dried, washed in 2× SSC with 0.05% Tween-20 (Solarbio, T8220) (pH=7.0) at room temperature for 30 seconds, rinsed briefly in distilled water, and allowed to air-dry. Next, 10 µL DAPI was applied to a clean slide and covered with the coverslip, followed by sealing with rubber cement. The slide was imaged with a confocal microscope (Nikon, Live SR CSU W1) with a 100× objective. The imaging data were then processed using Fiji Is Just ImageJ (FIJI).

Probe design and construction

Probe design and construction were performed with previously described methods³⁰^,⁵². In brief, oligonucleotide probes targeting specific genomic regions were designed following online instructions and the methods of Oligominer (https://github.com/brianbeliveau/Oligominer). The probes were amplified from complex oligonucleotide pools (Hongxun Biotech) by limited cycles of amplification. The templates in the oligonucleotide pools were designed to contain a 32-mer targeting region that is complementary to the genomic sequence, a 30-mer flanking sequence to be hybridized by the secondary probes, as well as two 20-mer primer binding sequences to amplify the probes.

For probe construction, we firstly amplified the oligonucleotide pools via 26 cycles of PCR to generate templates for in vitro transcription (Phanta Max Super-Fidelity DNA Polymerase, #P505-d2). The PCR product was column-purified (Zymo DNA Clean and Concentrator, DCC-5) and then converted into RNA via in vitro transcription (HiScribe™ T7 High Yield RNA Synthesis Kit NEB, #E2040S). RNA templates were converted back to DNA via reverse transcription (MAN0012047 TS Maxima H Minus Reverse Transcriptase, Thermo Fisher #EP0751), and the single-stranded DNA products were then purified by alkaline hydrolysis (50 μL of 0.25 M EDTA and 0.5 M NaOH) and column purification (Zymo Research, #D4006). The concentrations of the composites were described in a previous work from our laboratory.

Fluorescence in situ hybridization

For FISH experiments, the cell sample preparation methods were identical to those used for the immunofluorescence experiments. After fixation using 4% (w/v) paraformaldehyde for 15 minutes, the samples were incubated in 1x PBST (1xPBS +1% (v/v) Triton X-100) for 10 minutes and then rinsed twice with 1x PBST. Each sample was then incubated in 100 μg/mL RNaseA (TransGen Biotech, GE101-01) to remove RNA followed by incubation in 0.1M HCl in 1x PBST for 10 minutes, washed 3 times in 1x PBST, washed 3 times in 2x SSCT (2x saline sodium citrate + 1% (v/v) Triton X-100) at RT, and incubated in 2x SSCT + 50% (v/v) formamide at 4 ℃ overnight. For sample prehybridization, the samples were incubated in 50% formamide (Sigma-Aldrich, 47671) + 2x SSCT at 78 ℃ for 10 minutes and then dehydrated by incubation in 70%, 85%, and 100% ice-cold ethanol successively, for 1 min each. For probe prehybridization, synthesized primary probes (5 μL) and secondary probes (1μL, 100 μM) were mixed with 100% formamide (35 μL), after which the probe mixture was incubated in a mixer for 15 min at 37 ℃. Next, pre-warmed 20% (w/v) dextran (Sigma-Aldrich, D8906-10G) (35 μL) was added to the mixture, which was incubated in a mixer for 30 min at 37 ℃. Finally, the mixture was incubated at 86 ℃ for 3 minutes and cooled on ice immediately.

Samples were then denatured for 3 minutes at 86 °C and hybridized at 37 °C in a humidified chamber overnight. The hybridized samples were then rinsed twice for 15 minutes in pre-warmed 2x SCCT at 60 °C, followed by washing for 10 minutes in 2x SSC at room temperature, and stored at 4 °C in 2x SSC before imaging.

Cell synchronization and nucleotide labelling

Cells were grown on glass-bottom dishes and synchronized at the G1/S transition by 2 mM thymidine (Sigma-Aldrich, T1895) for 15 h, followed by culturing in fresh DMEM for 10 h, and treatment with 2 μg/mL aphidicolin (abcam, ab142400) for 15 h.

To incorporate EdU in HeLa cells, the cells were bathed with growing medium containing EdU for 30 min after releasing them from the G1/S transition for 1 h, followed by growth in fresh DMEM for 2 days. The cy5 dye was then conjugated to EdU using the Click-iT EdU Imaging Kit (Thermo Scientific, C10340) according to the manufacturer’s instructions. Imaging data were acquired using the spinning disk confocal system described above.

Radius of gyration analysis

3D chromosome morphology was first segmented from 3D images using the Imaris Volume module with its default parameters, after which we labeled each 3D chromosome according to its connectivity and obtained the coordinates and intensity (gray value) of each voxel within each chromosome using custom-written MATALB scripts. The radius of gyration (Rg) was obtained as the following formula:

Cluster size analysis

Clusters were recognized using the Surfaces module including with the Imaris software package. The “diameter of largest sphere” was set to 0.492 μm, the threshold value was set to 25, and the remaining parameters were set to their default values. When the search was complete, we then erased the signal outside the nuclei manually. Finally, the area and other cluster properties were exported by Imaris automatically for the subsequent statistical analysis.

In situ Hi-C

The in situ Hi-C libraries was generated as previously described (Rao et al., 2014). Briefly, cells were grown to approximately 70–80% confluence, washed with PBS, fixed in 1% formaldehyde, and suspended in Hi-C lysis buffer to which 100 U MobI restriction enzyme (NEB, R0147) was added for overnight chromatin digestion. Free ends were labeled with biotin and then ligated together in situ. Crosslinks were reversed, the DNA was sheared to produce 300–500 bp fragments, and then biotinylated ligation junctions were recovered with streptavidin beads. Hi-C libraries were amplified using PCR, constructed according to the NEBnext library preparation protocol (NEB, E7335), and sequenced on the Illumina HiSeq X Ten platform.

Hi-C data analysis

Hi-C data were processed by HiC-Pro{Servant, 2015 #718}⁵³. Briefly, reads were first aligned on the hg19 reference genome. Uniquely mapped reads were normalized using Iterative Correction and Eigenvector (ICE) decomposition and library size. For compartment A/B analysis, HiTC⁵⁴ was used to visualize the interaction matrix and calculate the PC1 (at 150-kb resolution). Compartment switches were defined by comparing the PC1 values between Rad-OE cells and control cells, using zero as the PC1 cutoff. For TAD analysis, ICE-normalized 40-kb resolution matrices were used to detect TAD with a script described by Crane et al (https://github.com/dekkerlab/crane-nature-2015). Insulation scores were calculated for each 40-kb bin, and the valleys of the insulation score curves were defined as TAD boundaries⁵⁵. CTCF annotation was genera{van der Weide, 2021 #751}ted by GENOVA⁵⁶. Aggregate TAD analysis was performed using cooltools (v0.4.0)⁵⁷.{van der Weide, 2021 #751}{van der Weide, 2021 #751}{van der Weide, 2021 #751}{van der Weide, 2021 #751}{van der Weide, 2021 #751}{van der Weide, 2021 #751}{van der Weide, 2021 #751}

RNA-seq experiments

Total RNA was extracted using the MolPure Cell RNA Kit (YEASEN, 19231ES50). RNA sequencing libraries were constructed using the NEBNext Ultra RNA Library Prep Kit for Illumina® (NEB England BioLabs). RNA-seq paired-end reads were sequenced on the Illumina NovaSeq 6000 platform.

RNA-seq data analysis

The raw RNA sequences were cleaned using TrimGalore (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/) and mapped to human reference genome hg19 by STAR (v2.7.1a) with default parameters. All mapped bam files were converted to bigwig using bedtools (v2.24.0) ⁵⁸. for visualization in IGV. High-quality mapped reads were quantified using htseq-count (v0.11.3)⁵⁹. Differentially expressed genes were analyzed by DEseq2⁶⁰. Functional enrichment of previously reported gene sets in the transcriptomes between Rad21-OE and control cells was determined using the GSEA software package^61,62. GO enrichment analysis was performed using Enrichr⁶³.

Data availability

All Hi-C and RNA-seq datasets have been deposited in GEO under the accession number GSE183186.

Acknowledgements

We thank Dr. Wei Guo from University of Pennsylvania for providing HeLa-S3 cell line. We also thank Dr. Chunyan Shan from Core Facilities at School of Life Sciences, Peking University for assistance with super-resolution imaging. We thank the flow cytometry Core at National Center for Protein Sciences at Peking University, particularly Dr. Hongxia Lv, Liying Du and Jia Luo, for technical help. This work is supported by grants from This work is supported by grants from the National Key R&D Program of China for Y.S. (No. 2017YFA0505300) and the National Science Foundation of China for Y.S. (21825401).

Author Contributions

Y.S. and Y.A.S. conceived of the study. Y.S. supervised the study. Y.S., Y.A.S., X.X., and W.X.Z. designed the experiments. Y.A.S conducted all imaging experiments with help from K.Y.C, X.T.W. X.X. performed co-IP and western blotting experiments. W.X.Z. performed bioinformatics analysis. Y.Z. prepared the Hi-C library. Y.Z.L, and M.L.Z. designed and constructed the probe. Y. Hou contributed technical assistance/suggestions. Y.S., Y.A.S., X.X., and W.X.Z wrote the manuscript with input from all authors.

Conflict of Interest

The authors declare no competing financial interests.

Lieberman-Aiden, E. et al. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science 326, 289–293, doi:10.1126/science.1181369 (2009).
Dixon, J. R. et al. Topological domains in mammalian genomes identified by analysis of chromatin interactions. Nature 485, 376–380, doi:10.1038/nature11082 (2012).
Rao, S. S. et al. A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. Cell 159, 1665–1680, doi:10.1016/j.cell.2014.11.021 (2014).
Rao, S. S. P. et al. Cohesin Loss Eliminates All Loop Domains. Cell 171, 305-320 e324, doi:10.1016/j.cell.2017.09.026 (2017).
Vietri Rudan, M. & Hadjur, S. Genetic Tailors: CTCF and Cohesin Shape the Genome During Evolution. Trends Genet 31, 651–660, doi:10.1016/j.tig.2015.09.004 (2015).
Nora, E. P. et al. Targeted Degradation of CTCF Decouples Local Insulation of Chromosome Domains from Genomic Compartmentalization. Cell 169, 930-944 e922, doi:10.1016/j.cell.2017.05.004 (2017).
Alipour, E. & Marko, J. F. Self-organization of domain structures by DNA-loop-extruding enzymes. Nucleic Acids Res 40, 11202–11212, doi:10.1093/nar/gks925 (2012).
Fudenberg, G. et al. Formation of Chromosomal Domains by Loop Extrusion. Cell Rep 15, 2038–2049, doi:10.1016/j.celrep.2016.04.085 (2016).
Nasmyth, K. Disseminating the genome: joining, resolving, and separating sister chromatids during mitosis and meiosis. Annu Rev Genet 35, 673–745, doi:10.1146/annurev.genet.35.102401.091334 (2001).
Ganji, M. et al. Real-time imaging of DNA loop extrusion by condensin. Science 360, 102–105, doi:10.1126/science.aar7831 (2018).
Davidson, I. F. et al. DNA loop extrusion by human cohesin. Science 366, 1338–1345, doi:10.1126/science.aaz3418 (2019).
Kim, Y., Shi, Z., Zhang, H., Finkelstein, I. J. & Yu, H. Human cohesin compacts DNA by loop extrusion. Science 366, 1345–1349, doi:10.1126/science.aaz4475 (2019).
Gruber, S., Haering, C. H. & Nasmyth, K. Chromosomal cohesin forms a ring. Cell 112, 765–777 (2003).
Haering, C. H., Lowe, J., Hochwagen, A. & Nasmyth, K. Molecular architecture of SMC proteins and the yeast cohesin complex. Mol Cell 9, 773–788 (2002).
Kueng, S. et al. Wapl controls the dynamic association of cohesin with chromatin. Cell 127, 955–967, doi:10.1016/j.cell.2006.09.040 (2006).
Gandhi, R., Gillespie, P. J. & Hirano, T. Human Wapl is a cohesin-binding protein that promotes sister-chromatid resolution in mitotic prophase. Curr Biol 16, 2406–2417, doi:10.1016/j.cub.2006.10.061 (2006).
Ciosk, R. et al. Cohesin's binding to chromosomes depends on a separate complex consisting of Scc2 and Scc4 proteins. Mol Cell 5, 243–254, doi:10.1016/s1097-2765(00)80420-7 (2000).
Tedeschi, A. et al. Wapl is an essential regulator of chromatin structure and chromosome segregation. Nature 501, 564–568, doi:10.1038/nature12471 (2013).
Haarhuis, J. H. I. et al. The Cohesin Release Factor WAPL Restricts Chromatin Loop Extension. Cell 169, 693-707 e614, doi:10.1016/j.cell.2017.04.013 (2017).
Panizza, S., Tanaka, T., Hochwagen, A., Eisenhaber, F. & Nasmyth, K. Pds5 cooperates with cohesin in maintaining sister chromatid cohesion. Curr Biol 10, 1557–1564, doi:10.1016/s0960-9822(00)00854-x (2000).
Chao, W. C. et al. Structural Studies Reveal the Functional Modularity of the Scc2-Scc4 Cohesin Loader. Cell Rep 12, 719–725, doi:10.1016/j.celrep.2015.06.071 (2015).
Murayama, Y. & Uhlmann, F. Biochemical reconstitution of topological DNA binding by the cohesin ring. Nature 505, 367–371, doi:10.1038/nature12867 (2014).
Rhodes, J., Mazza, D., Nasmyth, K. & Uphoff, S. Scc2/Nipbl hops between chromosomal cohesin rings after loading. Elife 6, doi:10.7554/eLife.30000 (2017).
Shi, Z., Gao, H., Bai, X. C. & Yu, H. Cryo-EM structure of the human cohesin-NIPBL-DNA complex. Science 368, 1454–1459, doi:10.1126/science.abb0981 (2020).
Haering, C. H. et al. Structure and stability of cohesin's Smc1-kleisin interaction. Mol Cell 15, 951–964, doi:10.1016/j.molcel.2004.08.030 (2004).
Higashi, T. L. et al. A Structure-Based Mechanism for DNA Entry into the Cohesin Ring. Mol Cell 79, 917-933 e919, doi:10.1016/j.molcel.2020.07.013 (2020).
Wutz, G. et al. Topologically associating domains and chromatin loops depend on cohesin and are regulated by CTCF, WAPL, and PDS5 proteins. EMBO J 36, 3573–3599, doi:10.15252/embj.201798004 (2017).
Schwarzer, W. et al. Two independent modes of chromatin organization revealed by cohesin removal. Nature 551, 51–56, doi:10.1038/nature24281 (2017).
Su, Q. P. et al. Superresolution imaging reveals spatiotemporal propagation of human replication foci mediated by CTCF-organized chromatin structures. Proc Natl Acad Sci U S A 117, 15036–15046, doi:10.1073/pnas.2001521117 (2020).
Li, Y. et al. Transcription-coupled structural dynamics of topologically associating domains regulate replication origin efficiency. Genome Biol 22, 206, doi:10.1186/s13059-021-02424-w (2021).
Du, Z. et al. Allelic reprogramming of 3D chromatin architecture during early mammalian development. Nature 547, 232–235, doi:10.1038/nature23263 (2017).
Zhang, J. et al. DONSON and FANCM associate with different replisomes distinguished by replication timing and chromatin domain. Nat Commun 11, 3951, doi:10.1038/s41467-020-17449-1 (2020).
Porkka, K. P., Tammela, T. L., Vessella, R. L. & Visakorpi, T. RAD21 and KIAA0196 at 8q24 are amplified and overexpressed in prostate cancer. Genes Chromosomes Cancer 39, 1–10, doi:10.1002/gcc.10289 (2004).
Xu, H. et al. Enhanced RAD21 cohesin expression confers poor prognosis and resistance to chemotherapy in high grade luminal, basal and HER2 breast cancers. Breast Cancer Res 13, R9, doi:10.1186/bcr2814 (2011).
Atienza, J. M. et al. Suppression of RAD21 gene expression decreases cell growth and enhances cytotoxicity of etoposide and bleomycin in human breast cancer cells. Mol Cancer Ther 4, 361–368, doi:10.1158/1535-7163.MCT-04-0241 (2005).
Sonnessa, M. et al. NLRP3 Inflammasome From Bench to Bedside: New Perspectives for Triple Negative Breast Cancer. Frontiers in Oncology 10, doi:ARTN 158710.3389/fonc.2020.01587 (2020).
Wang, L. M. et al. Identification of Alternatively-Activated Pathways between Primary Breast Cancer and Liver Metastatic Cancer Using Microarray Data. Genes-Basel 10, doi:ARTN 75310.3390/genes10100753 (2019).
Cruceriu, D., Baldasici, O., Balacescu, O. & Berindan-Neagoe, I. The dual role of tumor necrosis factor-alpha (TNF-alpha) in breast cancer: molecular insights and therapeutic approaches. Cell Oncol 43, 1–18, doi:10.1007/s13402-019-00489-1 (2020).
Chen, Y., Li, L. H. & Zhang, J. F. Cell migration inducing hyaluronidase 1 (CEMIP) activates STAT3 pathway to facilitate cell proliferation and migration in breast cancer. J Recept Sig Transd 41, 145–152, doi:10.1080/10799893.2020.1800732 (2021).
Banach, A. et al. CEMIP upregulates BiP to promote breast cancer cell survival in hypoxia. Oncotarget 10, 4307–4320, doi:10.18632/oncotarget.27036 (2019).
Xu, J. et al. Association between KIAA1199 overexpression and tumor invasion, TNM stage, and poor prognosis in colorectal cancer. Int J Clin Exp Patho 8, 2909–2918 (2015).
Gillespie, P. J. & Hirano, T. Scc2 couples replication licensing to sister chromatid cohesion in Xenopus egg extracts. Curr Biol 14, 1598–1603, doi:10.1016/j.cub.2004.07.053 (2004).
Lengronne, A. et al. Cohesin relocation from sites of chromosomal loading to places of convergent transcription. Nature 430, 573–578, doi:10.1038/nature02742 (2004).
Watrin, E. et al. Human Scc4 is required for cohesin binding to chromatin, sister-chromatid cohesion, and mitotic progression. Current Biology 16, 863–874, doi:10.1016/j.cub.2006.03.049 (2006).
Kikuchi, S., Borek, D. M., Otwinowski, Z., Tomchick, D. R. & Yu, H. Crystal structure of the cohesin loader Scc2 and insight into cohesinopathy. Proc Natl Acad Sci U S A 113, 12444–12449, doi:10.1073/pnas.1611333113 (2016).
Hansen, A. S., Pustova, I., Cattoglio, C., Tjian, R. & Darzacq, X. CTCF and cohesin regulate chromatin loop stability with distinct dynamics. eLife 6, doi:10.7554/eLife.25776 (2017).
Bonev, B. & Cavalli, G. Organization and function of the 3D genome. Nat Rev Genet 17, 661–678, doi:10.1038/nrg.2016.112 (2016).
Michaelis, C., Ciosk, R. & Nasmyth, K. Cohesins: Chromosomal proteins that prevent premature separation of sister chromatids. Cell 91, 35–45, doi:Doi 10.1016/S0092-8674(01)80007-6 (1997).
Weitzer, S., Lehane, C. & Uhlmann, F. A model for ATP hydrolysis-dependent binding of cohesin to DNA. Curr Biol 13, 1930–1940, doi:10.1016/j.cub.2003.10.030 (2003).
Ryu, J. K. et al. Bridging-induced phase separation induced by cohesin SMC protein complexes. Sci Adv 7, doi:10.1126/sciadv.abe5905 (2021).
Shao, S. et al. TagBiFC technique allows long-term single-molecule tracking of protein-protein interactions in living cells. Commun Biol 4, 378, doi:10.1038/s42003-021-01896-7 (2021).
Boettiger, A. N. et al. Super-resolution imaging reveals distinct chromatin folding for different epigenetic states. Nature 529, 418–422, doi:10.1038/nature16496 (2016).
Servant, N. et al. HiC-Pro: an optimized and flexible pipeline for Hi-C data processing. Genome Biology 16, doi:ARTN 25910.1186/s13059-015-0831-x (2015).
Servant, N. et al. HiTC: exploration of high-throughput 'C' experiments. Bioinformatics 28, 2843–2844, doi:10.1093/bioinformatics/bts521 (2012).
Crane, E. et al. Condensin-driven remodelling of X chromosome topology during dosage compensation. Nature 523, 240-U299, doi:10.1038/nature14450 (2015).
van der Weide, R. H. et al. Hi-C analyses with GENOVA: a case study with cohesin variants. NAR Genom Bioinform 3, lqab040, doi:10.1093/nargab/lqab040 (2021).
Flyamer, I. M., Illingworth, R. S. & Bickmore, W. A. Coolpup.py: versatile pile-up analysis of Hi-C data. Bioinformatics 36, 2980–2985, doi:10.1093/bioinformatics/btaa073 (2020).
Quinlan, A. R. BEDTools: The Swiss-Army Tool for Genome Feature Analysis. Curr Protoc Bioinformatics 47, 11 12 11–34, doi:10.1002/0471250953.bi1112s47 (2014).
Anders, S., Pyl, P. T. & Huber, W. HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169, doi:10.1093/bioinformatics/btu638 (2015).
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15, 550, doi:10.1186/s13059-014-0550-8 (2014).
Mootha, V. K. et al. PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nature genetics 34, 267–273, doi:10.1038/ng1180 (2003).
Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102, 15545–15550, doi:10.1073/pnas.0506580102 (2005).
Kuleshov, M. V. et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res 44, W90-97, doi:10.1093/nar/gkw377 (2016).

There is NO Competing Interest.

20211118SupplementaryInformation.docx
FigureS1, FigureS2, FigureS3, FigureS4, FigureS5
accessionnumbersandreviewertoken.pdf
Supplementary Info File #1
nrreportingsummaryflatten.pdf
Supplementary Info File #2

PDF

Version 1

posted 14 Dec, 2021

You are reading this latest preprint version

Rad21 is the core subunit of the cohesin complex involved in directing genome organization

Status:

Version 1

Abstract

Figures

Introduction

Results

Cohesin is arranged in axial chromosomal domains in cells with up-regulated Rad21

Discussion

Materials And Methods

Cell Culture

Live cell imaging and analysis

Immunofluorescence

Heterogeneity level analysis

Co-immunoprecipitation (co-IP) assay

Chromosome painting

Probe design and construction

Fluorescence in situ hybridization

Cell synchronization and nucleotide labelling

Radius of gyration analysis

Cluster size analysis

In situ Hi-C

Hi-C data analysis

RNA-seq experiments

RNA-seq data analysis

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1