Genome-wide analyses of chromatin interactions after the loss of Pol I, Pol II and Pol III

Background The relationship between transcription and the 3D genome organization is one of the most important questions in molecular biology, but the roles of transcription in 3D chromatin remain controversial. Multiple groups showed that transcription affects global Cohesin binding and genome 3D structures. At the same time, several studies have indicated that transcription inhibition does not affect global chromatin interactions. Results Here, we provide the most comprehensive evidence to date to demonstrate that transcription plays a marginal role in organizing the 3D genome in mammalian cells: 1) degraded Pol I, Pol II and Pol III proteins in mESCs, and showed their loss results in little or no changes of global 3D chromatin structures for the first time; 2) selected RNA polymerases high abundance binding sites-associated interactions and found they still persist after the degradation; 3) generated higher resolution chromatin interaction maps and revealed that transcription inhibition mildly alters small loop domains; 4) identified Pol II bound but CTCF and Cohesin unbound loops and disclosed that they are largely resistant to transcription inhibition. Interestingly, Pol II depletion for a longer time significantly affects the chromatin accessibility and Cohesin occupancy, suggesting RNA polymerases are capable of affecting the 3D genome indirectly. So, the direct and indirect effects of transcription inhibition explain the previous confusing effects on the 3D genome. Conclusions We conclude that Pol I, Pol II, and Pol III loss only mildly alter chromatin interactions in mammalian cells, suggesting the 3D chromatin structures are pre-established and relatively stable.


Background
The relationship between transcription and the 3D chromatin structures is one of the most fundamental questions in the post-genomic era [1][2][3][4]. Mounting evidence has shown that transcription activities correlate with more DNA interactions in development and diseases [5][6][7][8]. Topologically associating domains (TADs) boundaries or insulated neighborhoods or CTCF loop domains usually enrich active transcription with both protein machinery and noncoding RNAs [9][10][11].
Recent studies combining knock out with inhibition showed that transcription could relocate Cohesin over mammalian chromatin [15], implicating that Pol II may regulate the 3D genome via its impact on Cohesin. Blocking of transcription elongation in mammalian cells increases Cohesin binding and loop formation at CTCF binding sites within the gene bodies [16]. Inhibit D. melanogaster transcription alters chromatin interactions both within and between domains [12][13][14]. However, it is unclear whether Pol II regulates 3D chromatin landscapes via Cohesin directly.
Many early development-related studies have revealed insights for the transcription and 3D chromatin landscape. Pol II transcription inhibition during the early development of mouse embryo does not affect global TAD structures [17,18], while it is difficult to interpret because of the relatively low sequencing depth of these experiments and developmental arrest after transcription inhibition. Also, the chromatin organizations of transcriptionally inactive mature oocytes and sperms are quite similar to the embryonic stem cells [17,[19][20][21], implicating that it might not be transcription per se, but transcription-associated proteins contribute to the 3D genome organization.
The unchanged phenotype after transcription inhibition in mammalian cells is usually based on aggregate analyses of all chromatin loops [17,18,[21][22][23]. It is hard to evaluate the contribution of transcription, as CTCF and Cohesin play a predominant role in the 3D chromatin landscape, and they occupy most of the loops in mammalian cells [24][25][26]. It is immature to conclude that transcription has no impact on chromatin interactions in mammals because one of the critical evidence is still missing in the field, which is to precisely evaluate the roles of Pol II in the 3D genome in the absence of CTCF and Cohesin.
RNA polymerases synthesize RNA, forming DNA, RNA, and protein ternary complexes in the nucleus [27][28][29]. Pol I, Pol II, and Pol III are three distinct DNAdependent RNA polymerases that function together with thousands of transcriptional and chromatin regulators to synthesize rRNA, mRNA, and tRNA, respectively [9,[30][31][32][33][34][35][36]. Although the structure and function of RNA polymerases have been studied for 50 years [37], in particular, the roles of Pol I and Pol III in 3D chromatin organizations are poorly investigated.
To explore the function of RNA polymerase proteins in 3D chromatin organization, we specifically degraded the largest subunits of Pol I, Pol II, and Pol III in murine embryonic stem cells. Global chromatin interactions and higher resolution interactions among regulatory elements remain unchanged after Pol II depletion. Chromatin organization reforms during mitotic exit in the absence of Pol II, supporting the pre-established model for the 3D genome. We then identified Pol II bound but Cohesin and CTCF unbound loops, and they are largely resistant to transcription inhibition both in our and public datasets from different labs.
Additionally, acute depletion of Pol I and Pol III also does not result in global and local changes for the 3D genome. These results collectively demonstrate that RNA polymerases play a marginal role in organizing 3D chromatin landscapes in mammalian cells. Interestingly, long-time depletion of Pol II decreases chromatin accessibility and Cohesin occupancy over chromatin. We propose that immediate transcription inhibition does not affect 3D chromatin structures, but the indirect effects of transcription inhibition do.

Acute degradation of Pol I, Pol II, and Pol III in mESCs with auxin-inducible degron technology
To investigate the roles of RNA polymerase proteins in 3D chromatin organization, we subjected mESCs to RNA polymerases degradation with degron technology. RNA polymerase-mediated transcription could be inhibited with different inhibitors, while they either lost the capacity to distinguish different RNA polymerases or had specificity for RNA polymerase (such as alpha-amanitin) but worked at high concentrations and take a long time [38,39]. The degron system was applied to RNA polymerase subunits because it degrades protein rapidly and accurately (Fig. 1a) [40,41]. The auxin-inducible degron system was set up in murine embryonic stem cells: exogenously expressed OsTIR1 was inserted into the Rosa26 locus under the Tet-on promoter, and then the protein of interest was tagged at its C terminal with mAID-GFP by CRISPR genome editing followed by single clone selection (Fig. S1a). The mAID tag could be recognized and degraded by OsTIR1 in the presence of auxin (Fig. 1a). The C-terminal domain of the largest subunit of RNA polymerase was tagged with mAID-GFP. This system allows us to specifically investigate endogenous Pol I (Rpa1), Pol II (Rpb1), and Pol III (Rpc1) with the GFP tag and the phenotype for loss of function with the mAID tag.
We next confirmed that the RPA1, RPB1, and RPC1 proteins could be depleted as an mAID-GFP fusion protein in murine embryonic stem cells. To determine the time point for Pol II loss of function, we first induced OsTIR1 expression for 24hr, added auxin, and then performed western blot analyses at different time points. Indeed, RNA polymerases could be degraded rapidly and accurately after the addition of auxin (Fig. 1b-c). RNA polymerases ChIP-Seq datasets were generated after depletion. ChIP-Seq heatmap analyses confirmed that protein depletion is efficient with the auxin-inducible degron technology (Fig.   1d). Furthermore, the global mature mRNA level measured by polyA RNA-Seq does not change dramatically after Pol II depletion for 6hr (Fig. S1b, Table S1). This rapid degradation system will allow us to specifically investigate the immediate roles of three RNA polymerases on genome structures for the first time.
To determine whether the rapid depletion of RNA polymerase causes pleiotropic effects on mESCs, we compared the cellular and molecular properties of the engineered and wild-type mESCs under our experimental conditions. The cell viability, cell cycle, caspase 3/7 activities, and γH2AX level are comparable between the wild-type and the engineered cells upon treatment with doxycycline and auxin (Fig. S1c-e, S1g). These engineered cells behaved similarly to the vehicle-treated wild-type cells, indicating that the rapid depletion of RNA polymerases did not cause measurable effects on the cells at the time point that we checked and that mAID-GFP tagging did not interfere with the physiological properties of mESCs ( Fig. S1c-g). A degradation time of 6hr for Pol II, 24hr for Pol I, and Pol III were chosen for the downstream Hi-C analyses because the proteins were degraded quickly. The depletion does not cause noticeable changes in the protein level of other chromatin structural regulators (i.e., SMC1 and CTCF) as well as the largest subunits of Pol I, Pol II and Pol III (Fig. 1b), although a slight destabilization of endogenously tagged proteins is observed as reported previously [40]. These results suggest that mAID-GFP fusion supports the essential functions of RNA polymerases.

Pol II, and Pol III
Previous studies on the relationships between transcription and 3D genome usually focused on Pol II [3,[42][43][44], here we separately investigated Pol I, Pol II, and Pol III for their roles in the 3D genome. To generate high-resolution chromatin structure data after RNA polymerases depletion, we used our recently developed BAT Hi-C (Bridge linker-Alul-Tn5 Hi-C) [45], which could delineate chromatin conformational features such as DNA loops efficiently. Our BAT Hi-C optimized the conventional in situ Hi-C procedure by 1) fragmentation of the chromatin with the blunt four cutter restriction enzyme Alu1; 2) proximity T:A ligation with a biotinylated DNA linker; 3) isolation of chromatin to enrich chromatin-mediated interactions; and 4) construction of the library with Tn5 tagmentation (see Methods). A combination of the BAT Hi-C technique and RNA polymerase rapid degradation system will allow us to adequately investigate the relationship between specific RNA polymerases and chromatin structures.
Pol I, Pol II, and Pol III degron cells were subjected to degradation and collected for BAT Hi-C analyses. Two biological Hi-C replicates for both untreated (total reads=277 million) and Pol II degraded (total reads=300 million) cells were obtained (Table S2). The Hi-C data are reproducible (Fig. S2a) and consistent with the mESC Hi-C data quality matrix, as well as A/B compartments, published previously ( Fig. S2b-c) [46]. We pooled the data and acquired a 25 kb resolution Hi-C dataset for both untreated and degraded Pol II conditions. The quality of Hi-C datasets after Pol I or Pol III degradation is similar to the Pol II degron Hi-C datasets (data not shown). Since the Hi-C detected structures are sensitive to sequencing depth, we sampled our datasets to the same sequencing depth for comparison (Table S2) (Table S3). We then compared the Hi-C-detected TAD structures between untreated and degraded conditions. The TAD boundaries identified in our Hi-C dataset confirm the TAD boundaries reported previously [46].  S2f). For comparison, Hi-C datasets after CTCF degradation were reanalyzed with the same methods and indicated that CTCF degradation causes a significant decrease in intra-TAD interactions (Fig. 2d-e). These results indicate that the 3D genome organization persists after Pol I, Pol II, and Pol III depletion.

Chromatin structures re-establish during mitotic exit after Pol II depletion
Our evidence suggested that RNA polymerases have little or no impact on maintenances of the 3D genome, but RNA polymerases might function during the process of 3D genome establishment. To explore this possibility, we performed Pol II degradation followed by chromatin structure analyses during mitotic exit.
Previous studies have reported that TAD structures disappear in mitosis and reestablish in the early G1 phase [47]. We synchronized our cells into the M phase and simultaneously degraded Pol II and then collected both the untreated and Pol II degraded cells for Hi-C analyses during mitotic exit. The cell cycle analyses indicate that approximately 90% of cells are synchronized into M phase following previously published protocols ( Fig. S2g) [48]. By comparing the cells with or without Pol II during mitotic exit, we found that degradation of Pol II during mitotic exit does not cause changes for the A/B compartments and TAD structures (Fig.   2f). These results reveal that the Pol II protein is nonessential for both the maintenance and establishment of TAD structures in mESCs.

Pol I, Pol II and Pol III loss mildly alter local chromatin interactions
RNA polymerases are not required for global chromatin organizations; we next asked whether they play a crucial role in organizing local chromatin structures associated with active transcription. To test whether they organize local chromatin structures, we first identified Pol I, Pol II and Pol III binding hotspots or clusters with the ROSE algorithm and found mild differences in some specific regions ( Fig. 3a-c). Overall, the Hi-C contact frequency decreased 2.3%, 12.6%, and 1.5% at their binding clusters, respectively for Pol I, Pol II, and Pol III (Fig. 3d), also noticed mild differences in some specific regions (Fig. 3a-c). Interestingly, Pol III binding hotspots seem to be more clustered after Pol III depletion (Fig. 3c), which is in agreement with the previously known insulator functions of tRNA elements [49][50][51], and is also consistent with the small but increased chromatin interactions across tRNA cluster after Pol III depletion ( Fig. 3c-

d). Further correlation analyses among
Hi-C interaction changes and different types of functional genomic datasets were performed, and we found that Hi-C interaction changes have a better correlation with the corresponding ChIP-Seq signals (Fig. 3e). These results lead us to conclude that Pol I, Pol II, and Pol III play a modest role in structuring local chromatin interactions, and Pol II seems to contribute more. We then focused on Pol II in the following study.

Higher-resolution chromatin interaction analyses indicate that Pol II contributes no or little to organize small loop domains
The chromatin interaction analyses are highly dependent on the resolution of structures. To obtain higher-resolution information on the Pol II-dependent intra-TAD structures, we performed H3K27Ac HiChIP and Ocean-C analyses after Pol II depletion. H3K27Ac HiChIP maps chromatin loops-associated with active promoters and enhancers [52,53] and Ocean-C, a recently developed antibody free chromosome conformation capture technique, combines FAIRE-seq with Hi-C to map hubs of open chromatin interactions [54]. Chromatin loops were identified by hichipper and HiCCUPS [55,56] (Table S4). The histogram shows that most of the loop strength does not change much after Pol II degradation for loops identified by both algorithms with the Hi-C, HiChIP, or Ocean-C dataset (Fig. S3). Scatter plot and box plot of HiCCUPS-identified chromatin loops with all three datasets consistently show a slight decrease after Pol II depletion (Fig. 4a).
We next sought to identify features of the chromatin loops that are sensitive to Pol II depletion. Previous studies reported that active transcription defined the small compartmental domains throughout Eukarya, but transcription itself could not predict 3D chromatin organization in mammals at specific loci [13]. Our analyses of chromatin structures after Pol II degradation were used to explore the existence of Pol II transcription-dependent loop domains. Loop length analyses indicate that Pol II degradation mostly affects the interaction strength for small loop domains (Fig. 4b), usually less than 250 kb, which is reminiscent of small compartmental domains in Drosophila [13]. Considering the current resolution of Hi-C is not sufficient to make a conclusion, we analyzed two other higher resolution datasets, HiChIP and Ocean-C, and both of them exhibited consistent trends as Hi-C did. The HiChIP and Ocean-C datasets show 14% and 15% changes more significantly than the Hi-C dataset (4%) because they preferentially capture interactions connecting open chromatin, and a similar sequencing depth would have a higher resolution than Hi-C.
If Pol II is involved in structuring small chromatin loops, we will anticipate that active and silent regions concerning Pol II transcription would behave differently after Pol II depletion. To test this idea, we ranked averaged GRO-Seq (nascent transcript) signals for loop domains and their upstream and downstream 100 kb windows in decreasing order. Indeed, the top GRO-Seq signal-associated loop domains are smaller and showed a more significant decrease in chromatin interactions than the bottom GRO-Seq signal-associated domains in the three different datasets (Fig. 4c). These results indicate that although Pol II has no functions on the global genome structures, it mildly contributes to actively transcribed small loops.
If Pol II contributes to local chromatin organization, then Proximity Ligation-Assisted ChIP-seq (PLAC-seq) analyses for Pol II should show that Pol II is preferentially associated with short-range chromatin loops [57]. We reanalyzed previously published Pol II PLAC-Seq in mESCs. The results show that Pol IIassociated interactions connect promoters, gene bodies, terminators, and enhancers (Fig. 4d, Table S5). Most of these interactions are within 100kb (Fig.   4e), constrained within CTCF loop domains [58,59]. Some Pol II-associated interactions involve chromatin clusters among genes and their potential regulatory elements, as observed in the Bclaf1 locus (Fig. 4e). These results indicate that Pol II mediates short-range chromatin loops around gene regulatory elements and that Pol II depletion mildly affects these loops.

Pol II bound but CTCF and Cohesin unbound loops are resistant to Pol II depletion
Our RNA polymerases degron system can degrade Pol II within 1hr (Fig. 1b), which has the unique advance to separate the impact of transcription from the presence of RNA polymerases. So, we performed BAT Hi-C analyses after Pol II depletion for 1hr and 6hr. These Hi-C datasets are highly reproducible, as we showed before (Fig. S2a, S4a). The Compartment and TAD structure analyses reproducibly show no obvious changes after Pol II depletion for 1hr and 6hr (Fig.   2b, 2f, S4b-e). Taken together, our results demonstrate that Pol II proteins are dispensable for global chromatin interactions in mESCs.
Previous transcription inhibition followed by chromatin structure analyses also exhibited no change through aggregated analyses [17,18,22]. These analyses did not exclude the possibility that subsets of chromatin loops might change after transcription inhibition. On the other side, most of the chromatin loops in mammals are occupied by CTCF and Cohesin, so it is better to select Pol II bound, but CTCF and Cohesin unbound loops to evaluate the contributions of Pol II in 3D genome organization. We then selected Pol II bound loops in the absence of Cohesin and found CTCF binding almost invisible at these loops (Fig. 5a). A similar classification was applied to our Pol II degron Hi-C datasets, and these loops are largely preserved after Pol II degradation for 1hr and 6hr (Fig. 5b). Then, we also identified Pol II bound but both CTCF and Cohesin unbound loops and found they also preserved after Pol II depletion (Fig. S5a). We then performed the same analyses with the previous transcription inhibition Hi-C dataset in early embryo development and activated B cell [18,22]. The analyses consistently show that Pol II predominantly bound loops are resistant to transcription inhibition (Fig. 5b, S5a).
In contrast, CTCF bound loops significantly decrease interaction strength after CTCF depletion with the same analyses pipeline (Fig. S5b).
Super-enhancers loci are usually associated with extremely high levels of transcription [60,61]. Therefore, we examined these regions for the changes in chromatin interactions. The target-centered maps indicate that the super-enhancer regions mildly increase chromatin interactions after Pol II degradation at both 1hr and 6hr (Fig. 5c). Specifically, the interactions around the key pluripotency gene Esrrb and housekeeping gene Dhx9 adjacent to super-enhancer-associated loop domains are also increased (Fig. 5d, S5c). Moreover, this observation could be further independently validated by 4C-Seq analyses after Pol II depletion (Fig. 5e).
We then inhibited transcription with actinomycin D, DRB, or flavopiridol and observed a consistent increase trend in the interaction frequency (Fig. 5e). Recent live-cell imaging analyses indicate that Pol II transcription restrains the dynamics of chromatin [62]. This explains our observation that Pol II degradation causes a slight increase in chromatin interactions of super-enhancer regions. These results further suggest that Pol II is not directly involved in chromatin organization but restrain the dynamics of chromatin, underlying a mechanism for the increase of chromatin interactions after Pol II depletion.

Immediate depletion of Pol II does not perturb Cohesin chromatin binding, but more prolonged depletion does
If Pol II indeed is nonessential in 3D genome organization, how to explain the previous finding that transcription-mediated Cohesin chromatin binding and transcription elongation-mediated chromatin structures? We speculated that Pol II might indirectly regulate Cohesin. Further taking advantage of the Pol II degron system, we can degrade Pol II at different time points, especially at 1hr time point (Fig. 1b). The single locus (Fig. 6a) and meta-gene analyses (Fig. 6b-c)

Discussion
The relationship between transcription and the 3D genome organization is one of the most fundamental questions in modern molecular biology. Here we provided evidence: 1) specific transcription inhibition by degradation of Pol I, Pol II and Pol III result in little or no changes for global 3D chromatin structures as assayed by Hi-C (Fig. 2a-c, 2f); 2) select high abundance Pol I, Pol II and Pol III binding sitesassociated chromatin interactions and found they still persist after degradation (Fig.   3a-c); 3) generate higher resolution chromatin interaction maps and reveal that transcription inhibition mildly alters small loop domains (Fig. 4b-c); 4) separate the impact of transcription from the presence of RNA polymerase by the degradation of Pol II within 1hr and show that Pol II protein is dispensable for global chromatin organization ( Fig. S4b-d); 5) identify Pol II dominant loops in the absence of CTCF and Cohesin and found they are mostly resistant to transcription inhibition (Fig. 5b,   S5a). This evidence collectively demonstrates RNA polymerases play a marginal role in the 3D genome organization. Besides, time-course experiments show that more prolonged depletion causes a decrease of Cohesin binding, but not for the short time treatment (Fig. 6b-c), suggesting that RNA polymerases can regulate the 3D chromatin landscape indirectly. RNA polymerases are core enzymes for transcription that are constitutively expressed in all cell types, so 3D chromatin organization independent of transcription is likely to be general in mammalian cells.
Mounting evidence has shown a strong correlation between transcription and 3D genome structures [12,63,64], while their relationship is still controversial before our study. Transcription inhibition experiments suggest the persistence of chromatin structures after transcription inhibition [17,18] These results argue against this combinatory model for protein factors, but we cannot exclude a possibility that nuclear noncoding RNAs may organize 3D genome structures, which can also be independent of transcription activities.

We show that transcription may indirectly affect 3D genome structures via
Cohesin chromatin binding, and there are many other possibilities. For example, a previous imaging study showed that Pol II transcription constrains the dynamics of chromatin in the nucleus. After transcription inhibition, the highly transcribed regions move fasters in the nucleus, interact more with some regions, and less with the other regions. Since transcription is vital for all genes, it is also likely that transcription affects short half-life chromatin structural proteins.
Previous studies showed that structural factors such as CTCF and Cohesinmediated chromatin interactions create frameworks for loop domains and constrain high-frequency chromatin interactions within them [2,3,72,73]. CTCF and Cohesin depletion destroy most chromatin structures in mammalian cells, while it has little effects on transcription [24,26]. Here we depleted the core subunits of Pol I, Pol II, and Pol III in mESCs and found little changes for chromatin structures. This data suggests that transcription and 3D chromatin structures are largely uncoupled in the nucleus of mESCs. The previous signaling induced and heat shock followed chromatin structure analyses suggested a pre-existing model for the 3D genome [74][75][76]. Our evidence also supports the pre-existing model, for which the mechanisms could be connected to genome sequences or epigenetic modifications.

Conclusions
Our study provides the first comprehensive analyses of the roles of Pol I, Pol II, and Pol III proteins in 3D chromatin organization in mESCs. We demonstrate that RNA polymerases play a minor role in organizing the 3D genome, and we propose that transcription does not regulate the 3D genome directly but is able to regulate the 3D genome indirectly. This explains the confusing effects on the 3D genome after transcription inhibition because it is difficult to separate the direct or indirect effects of transcription inhibition. Our study also implicates that transcription and 3D chromatin organizations are largely uncoupled in the nucleus. Since transcription is nonessential for the 3D genome, further studies of genetics, epigenetics features of 3D genome, and noncoding RNAs may reveal new insights for the 3D genome organization.

Materials and methods
Details see supplementary information file.   Table S1. RNA-seq RPKM values in Pol II Untreated and Degron cells. Table S2. Summary statistics of the Hi-C, HiChIP, and Ocean-C data. Table   S3. Hi-C identified TADs (contact domains).

Competing interests
The authors declare no competing interests.         b) Heatmap illustrating ATAC-seq signals centered at all promoters, enhancers, and CTCF-binding insulators (± 5kb) before (untreated) and after auxin treatment (1hr or 6hr). Those regions were ranked by their level of chromatin accessibility (from high to low) for each catalog in mESCs. c) Same analysis as in Figure 6b but focused on Cohesin ChIP-seq binding.