Abstract
Type II DNA topoisomerases (topo II) flip the spatial positions of two DNA duplexes, called G- and T- segments, by a cleavage-passage-resealing mechanism. In living cells, these DNA segments can be derived from distant sites on the same chromosome. Due to lack of proper methodology, however, no direct evidence has been described so far. The beta isoform of topo II (topo IIβ) is essential for transcriptional regulation of genes expressed in the final stage of neuronal differentiation. Here we devise a genome-wide mapping technique (eTIP-seq) for topo IIβ target sites that can measure the genomic distance between G- and T-segments. It revealed that the enzyme operates in two distinctive modes, termed proximal strand passage (PSP) and distal strand passage (DSP). PSP sites are concentrated around transcription start sites, whereas DSP sites are heavily clustered in small number of hotspots. While PSP represent the conventional topo II targets that remove local torsional stresses, DSP sites have not been described previously. Most remarkably, DSP is driven by the pairing between homologous sequences or repeats located in a large distance. A model-building approach suggested that topo IIβ acts on crossovers to unknot the intertwined DSP sites, leading to chromatin decondensation.
Similar content being viewed by others
Introduction
Several lines of evidence demonstrated that the beta isoform of type II DNA topoisomerase (topo IIβ), which is reviewed recently1,2, is essential for transcriptional regulation of genes expressed in the last stage of neuronal development3,4,5. However, detailed mechanism for this process remains largely unknown. Locating the enzyme’s action sites on the genome would be a logical strategy to elucidate the role of topo IIβ. Type II DNA topoisomerases (topo II) catalyze the interconversion of the spatial positions of two DNA duplexes by a cleavage-passage-resealing mechanism. To discriminate these duplexes, the cleaved (gapped) strand and the other strand that is transferred through the gap are called G-segment and T-segment, respectively6,7,8. Topo II can be cross-linked to the G-segment right at the site of action by treating living cells with etoposide, a topo II-specific ‘poison’-type inhibitor. By taking advantage of this property of the enzyme and immuno-selecting the bound DNA, its action sites on the genome have been mapped in several studies9,10,11,12. Since the cleaved strand is covalently bound to the enzyme at the 5′ end, this procedure is similar to chromatin immunoprecipitation (ChIP) but essentially different in that no chemical cross-linker is used. Instead, cross-linking is based on the arrested enzymatic reaction intermediate. The procedure thus detects DNA sites directly involved in the reaction and not simply associated with the enzyme.
We have successfully used this type of strategy, termed “etoposide-mediated topoisomerase immunoprecipitation” or eTIP, for mapping of topo IIβ action sites (toposites) in selected genomic regions11. In that study we adopted oligonucleotide-tiling arrays to map the bound DNA fragments. In the present study we now extend this technology to a genome-wide scale by massive direct sequencing on NGS (named ‘eTIP-seq’). As a result, mapping resolution was improved significantly and, most importantly, repetitive DNA sequences became a reasonable subject of analysis. Another unique feature of the method is that the immunoprecipitated complex contains not only the covalently linked G-segment but also the T-segment associated with the complex noncovalently, which is released by high-salt treatment and can be analyzed separately afterwards. This provides additional useful information unattainable from conventional mapping techniques.
One of the cellular events where topo II becomes essential is the disentanglement of intertwined chromatid DNAs generated at the final stage of cell division8,13. In vertebrates, topo IIα is exclusively responsible for this reaction, namely decatenation. As for topo IIβ, it is believed to relax torsional stresses accumulated locally from various DNA transactions such as transcription14. In this reaction called relaxation, the enzyme removes writhes of DNA axis, either positive or negative, by passage between nearby DNA segments. In principle, however, distance between G- and T-segments can be much larger, like hundreds of kilo-bases, although no supporting evidence for this to occur has been reported to date. To detect these distant segmental passage events catalyzed by topo IIβ, we introduced additional steps to eTIP-seq, which include ligation between G- and T-segments via an oligomer adaptor attached to their sheared ends.
Using these techniques, we show here that topo IIβ operates in two distinct modes termed proximal strand passage (PSP) and distal strand passage (DSP) depending on the distance between G- and T-segments. PSP and DSP differ significantly not only in mechanistic sense but also in physiological consequences. While PSP sites are concentrated around transcription start sites (TSS) and also distributes throughout the genome to contribute the relaxation of local torsional stresses, DSP sites are heavily clustered in relatively small number of regions (hotspots) on chromosome. Most remarkable finding is that DSP occurs at the DNA crossovers facilitated by pairing of homologous DNA segments located at a long distance. This is an entirely novel perspective on the cellular function of topo II enzyme, in that a ‘DNA-centric’ mechanism may govern the target selection. The present study suggests that DSP is involved in regulation of higher-order chromatin structures to activate suppressed neuronal genes.
Results
Genome-wide identification of topo IIβ target sites by eTIP-seq
The experimental procedure for eTIP-seq analysis is outlined in Fig. 1A. Precursor cells of cerebellar granule neurons (CGN) isolated from infant rats were allowed to differentiate in vitro. Using the same culture system we have shown that topo IIβ is essential for transcriptional induction of a group of genes involved in mature neuronal function4,11. In the present study, cells were treated with etoposide at the second day in culture to trap topo IIβ on target DNA. After lysis of treated cells with a detergent sarkosyl and DNA fragmentation, the topo IIβ-DNA complex was recovered on magnetic beads coated with specific antibody to topo IIβ as described previously for eTIP procedure11.
The procedure hereafter branches into two routes. In the first route, labeled as eTIPa-seq, the complex bound-DNA on magnetic beads was fractionated into P1 and P2 fractions and sequenced separately on NGS to categorize the topo IIβ target sites termed ‘toposites’. The eTIPb-seq on the right branch determines the G- and T-segments bound to the same topo IIβ molecule by paired-end sequencing of chimeric fragments generated by the adaptor-mediated ligation. To eliminate noise, resulting chimeras were selected by checking whether or not their ends overlap with toposites that are detected in eTIPa-seq.
We first present three fundamental conditions or rules that are critical for the reasoning of eTIP-seq methodology (Fig. 1B-D). Mammalian topo II binds and acts preferentially on DNA crossovers formed between two duplexes. Handedness of the crossover favored by the enzyme is predetermined by structural properties. The enzyme preferentially recognizes DNA crossovers with right handedness, which is energetically favored over left handed ones15. In the scheme shown in Fig. 1B left, topo II approaches from the bottom of right-handed DNA crossover, whose wider angle is filled with the enzyme’s identical subunits. Configuration of the two strands illustrated here designates that G-segment is already held by the enzyme and T-segment is approaching to the ‘N-gate’16. The same crossover configuration at the ‘DNA gate’ of topo IIβ has been shown recently by a crystallographic study17 (they use different handedness definition). In the present study, we adopt this model as a presupposition to interpret the data obtained. It is worth noting that the crossover handedness is the same when the enzyme approaches from either side of the cross but the cleaved strand (G-segment) differs in each case (Fig. 1B center and right). Since the DNA fragments in immuno-selected topo IIβ-DNA complex are sheared by sonication, G- and T-segments can be contiguous or segregated (Fig. 1C). These correspond to the events of proximal strand passage (PSP) and distal strand passage (DSP), respectively (Fig. 3D). The etoposide-induced covalent linkage between topo II and G-segment is reversible under the conditions used for eTIP-seq, in that the enzyme is only partially denatured by sarkosyl and etoposide is diluted out significantly in the process (see Supplementary Fig. S2 for experimental evidence). Because of this, the immuno-captured topo II-DNA complex is composed of 3 species (Fig. 1D). It is also worth noting that the resealed G-segment is still bound on the enzyme and eluted in P2 fraction by 0.5 M NaCl treatment.
Three classes of topo IIβ target sites (toposites) are identified
We classified toposites into Ts1, Ts2, and Ts3 by eTIPa-seq (Fig. 1A). The detailed procedure is described in Supplementary Fig. S1. Distribution of toposites in a large scale, a whole chromosome, reveals a strong correlation between toposites and local GC content: Ts1 being located in GC-rich area whereas Ts2 resides in relatively AT-rich area (Fig. 2A). The toposite density plot clearly shows that Ts2 and Ts3 clusters are embedded among Ts1-dense regions in a mutually exclusive manner. Assembly of Ts2/Ts3 in relatively small number of clusters is a feature that distinguishes from Ts1 site. Although Ts2 and Ts3 clusters appear to occupy similar positions in this scale, further analysis demonstrated that Ts2 and Ts3 are distinct toposites with unique characters. As shown later, these clusters constitute hotspots of topo IIβ genomic targets that are likely to be involved in some regulatory processes. Genomic positions of toposites are listed in Supplementary Table S1 and Ts2/Ts3 clusters in Supplementary Table S2. Relative numbers of toposites differ significantly, Ts1 being most prominent (Supplementary Fig. S1).
GC contents of toposites are consistent with the above observation. Ts1 sites are significantly GC-rich as compared to Ts2/Ts3 sites (Fig. 2B). To investigate functional relevance, locations of these toposites were then looked at with respect to characteristic regions: transcription start site (TSS zone), genic region (GR), and intergenic region (IR) (Fig. 2C). Only protein-coding genes are taken into account using the “exRefSeq” which is derived from the rat reference sequence11. TSS zone (± 2 kb region encompassing the transcription start site) was included in the analysis as a functional region associated with genes. TSS zone normally overlaps with transcription control region containing promoter and binding sites for transcription factors. Ts1 was highly concentrated in gene-associated regions especially in the TSS zone. In contrast, Ts2/Ts3 sites, as viewed from the number of clusters, were enriched in intergenic regions.
SP120 (hnRNP U/SAF-A/SP120) is a multifunctional nuclear protein that binds both RNA and DNA18,19,20,21. A recent report shows that the protein binds chromatin-associated RNA and is involved in the organization of chromatin structure in a transcription-dependent manner22. Our previous study demonstrated that SP120 is a partner protein of topo IIβ that activates and stabilizes the enzyme through RNA-dependent association23. We have also shown that randomly cloned genomic targets of topo IIβ are frequently enriched with SP120-binding sites, suggesting that these proteins bind DNA within close distance24. Ts3 toposites frequently overlap with SP120 sites but very little with Ts2 sites, indicating the distinctiveness of these toposites (Fig. 2D).
The SP120 site overlaps also with Ts1 site considerably, implying that at least two distinct genomic sites are involved in the interaction between SP120 and topo IIβ. The eTIPb-seq analysis showed the presence of a looped DNA structure in TSS zone, where Ts1 site is overlapping with SP120 site (Supplementary Fig. S3). Topo IIβ-SP120 complex may somehow be involved in the initiation of transcription.
The eTIP-seq opens a new perspective on the reaction mode of topo II enzyme
To acquire positional information between DNA fragments associated with topo IIβ, we devised a new procedure that includes a ligation step between G- and T-segments via an oligomer adaptor attached to their sheared ends (Fig. 3A). We analyzed the resulting library of paired-read sequences that are referred to as ‘chimeras’. With respect to intra-chromosomal chimeras, various features like chimera length, combination of nearby toposites, read orientations, and genomic locations provide useful information. It should be noted that ‘chimera length’ indicates the map distance between paired reads and not the physical length of chimeric DNA fragments. We first plotted histograms for distribution of chimera length and found that a large proportion of chimera is rather short (51% are shorter than 3 kb). Longer chimeras of certain length, however, do exist and appear as peaks between 3 kb and 2 Mb but chimeras longer than 2 Mb rarely form peaks and their number levels off to a background level. Therefore, only chimeras shorter than 2 Mb will be analyzed hereafter. The data for chimera length distribution are presented elsewhere38.
In theory, chimeras must be formed between toposites and can be classified into 6 groups as to the combination of nearby toposites on both ends. From the number of toposites contributing to chimeras, expected number of chimera for randomly chosen toposite pairs of every combination can be calculated. Observed numbers of chimera were then plotted along with expected ones (Fig. 3B). Significance of toposite combinations were determined statistically by comparison of ‘observed’ versus ‘expected’ and the significance levels are summarized (Fig. 3C). The result clearly indicates that only homologous combinations (Ts1-Ts1, Ts2-Ts2, and Ts3-Ts3) are highly favored, whereas heterologous ones are avoided conversely. The presence of heterologous chimeras is probably caused by occasional errors in peak-calling and toposite assignments. Based on these data, only chimeras of homologous combinations will be taken into account and are referred to simply as Ts1, Ts2, and Ts3 chimeras, respectively.
We next examined what sort of information would be available from orientations of read pairs on chimera ends (Fig. 3D). The situation changed dramatically depending on the range of chimera length. In chimeras shorter than 3 kb (PSP, proximal strand passage), RF is by far the most dominant orientation for Ts1, Ts2, and Ts3 chimeras, suggesting that sonication-resistant loop structure (PSP loop) is still present at the ligation step (Fig. 3F). In chimeras between 3 kb and 2 Mb (DSP, distal strand passage), however, the RF dominance in Ts1 chimera is no longer apparent, whereas FR becomes as dominant as RF in Ts2 and Ts3 chimeras. This appears to indicate that the loop structure is disrupted in this range of length and FR is created as an additional combination for ligation (Fig. 3F). As chimera length exceeds 2 Mb, all classes of chimera show no preference in the read orientation, indicating that long chimeras are nonspecific in nature. They are very likely to be originated from the ligation events between neighboring IP complexes fixed on beads, thus termed inter-complex chimera. The same situation can be applied to the Ts1 chimera in the DSP region, which is thus regarded as “noise”. In summary, the Ts1 chimera reflects only PSP events whereas Ts2/Ts3 chimeras are linked to DSP events. The chimera categories and numbers identified are summarized in Fig. 3E. The Ts2/Ts3 chimeras in PSP were ineligible for further analysis. To avoid any confusion on chimera identity, chimeras will be referred hereafter with suffix (/PSP or /DSP) attached to the toposite name.
As this paper specializes only in DSPs, all the results for Ts1/PSP chimeras are presented and analyzed in Supplementary Fig. S3.
DSP sites are clustered to form a small number of ‘hotspots’
The length distribution showed that Ts2/DSPs are much longer than Ts3/DSPs (Fig. 4A). Genomic distribution of DSP chimeras showed a great deal of clustering, which often coincided with toposite clusters of the same categories, Ts2 or Ts3. A list of these “chimera-rich regions”, termed DSP clusters, is given in Supplementary Table S3. By comparing the position of these clusters on the genome browser, here we present a new concept termed ‘hotspot’ that integrates the toposite clusters and DSP clusters (Supplementary Table S4). Their relationship is clearly demonstrated by a region in chromosome 1 (Supplementary Fig. S4). Entire hotspots are depicted in a karyogram together with toposite clusters and DSP clusters (Supplementary Fig. S4). With the exception of Chr19, hotspots occur at least once in most chromosomes, indicating that they can be regarded as chromosome landmarks for topo IIβ targets, which is largely limited in number. With reference to gene density profile, the Ts2-hotspot usually resides in gene-poor regions, often adjacent to gene-rich region or “gene city”. This is consistent with the low GC content of Ts2 toposites and Ts2/DSP (Figs. 2B and 4B).
Ts3-hotspots are fewer than Ts2-hotspots and contain several interesting loci of tandemly repeated gene clusters such as noncoding RNA gene clusters and clusters of MHC (major histocompatibility complex) genes. Browser views of representative loci presented in Supplementary Fig. S4 and Supplementary Fig. S5 show that read positions of all the Ts3/DSP chimeras mapped in these regions coincided well with Ts3 toposites and SP120 sites. It is also worth noting that in accord with these sites high FAIRE-seq peaks were detected, indicating that these Ts3/DSP sites stay in an open chromatin conformation all the time in culture. When entire genome was searched, a large proportion of Ts3/DSP overlapped with SP120 sites but Ts2/DSP showed very little overlapping (Fig. 4C).
Sequences around the DSP chimera ends are homologous each other
In all the featured Ts3-hotspots, repetitive or homologous sequences were present around the chimera ends. To generalize this observation, sequence similarities and read orientations were examined for all Ts3/DSP and Ts2/DSP chimera ends. We do not know the exact position of the crossover targeted by topo IIβ but can assume that it is relatively close to the chimera ends (Fig. 3F). Therefore, from each chimera ends we cut out 2 kb-genomic sequences that contain the reads in the center and measured the homology between the two. Since homologous sequences can reside either on plus (Watson) or on minus (Crick) strand, similarity alignments were made between both strands and the resulting pair of Smith-Waterman scores were compared (Supplementary Table S5). As expected, in all DSP chimeras (both Ts3 and Ts2) either one of the two alignments showed much higher score compared to the other, indicating that chimera ends are indeed highly homologous. This terminal homology in DSP chimeras was significantly higher than control chimera (Fig. 4D). Ts2/DSP showed higher homology than Ts3/DSP.
Since homologous segments of duplex DNA aligned in parallel direction are likely to associate each other25,26, this may facilitate the formation of crossover for topo IIβ recognition. Both directions of repeats (direct and inverted) were present in the genome and, most importantly, the combination of read orientation precisely correlated with the directionality of repeats: RF/FR for direct repeat and FF/RR for inverted repeat (Supplementary Table S5 and Fig. 4E). Homologous pairing of direct repeats occurs much more frequently than that of inverted repeats for both DSP chimeras (Fig. 4F). The rule observed here between read orientations eliminates the possibility of inter-complex ligations, in which case the read orientations should be random. Therefore, the overlapping of Ts3/DSP with Ts3 sites shown in Supplementary Fig. S4 implies that the adaptor-mediated ligation in eTIPb-seq protocol indeed occurs between the G-segment and the T-segment bound to the same topo IIβ molecule.
Genome-wide homology search revealed that Ts3/DSP sites are enriched in the loci for MHC genes (RT1), noncoding RNA genes (4.5SH RNA, BC1 RNA, and Snord116) in addition to satellite I (Fig. 4G). The strong bias toward RF/FR orientation indicates that the homologous direct repeats are preferred target of topo IIβ, which was confirmed by examining individual loci for Ts3/DSP (Supplementary Fig. S4 and Fig. S5). The RT1 locus is an exception because FF/RR is a favored combination, which can be explained by the presence of long Ts3/DSP chimeras between the two RT1 gene clusters that originated from a duplication/inverted insertion event, occurred during rodent evolution27 (Supplementary Fig. S5). Major repeats like LINE or SINE were clearly minor constituents in Ts3/DSP. In contrast, however, LINE was the prevalent repeat associated with Ts2/DSP (Fig. 4G and Supplementary Fig. S5). It should be noted that a large proportion of chimera ends, 35% for Ts3/DSP and 66% for Ts2/DSP, are not attributable to any known repeats although these end-pairs are also highly homologous each other. We have not pursued their identity any further but they appear to be composed of several sequence groups that tend to cluster in separate DSP hotspots.
Topo IIβ-dependent alteration of nuclear structure during CGN differentiation
Analyses described hitherto are based on the data from eTIP-seq experiments that are performed at the day 2 of CGN culture, when differentiating cells are most active and abundant4. Hereafter, we will analyze the data sets obtained in three conditions: before differentiation at culture day 1 (D1), after differentiation at day 5 in culture (D5), and at day 5 cultured in the presence of ICRF-193, a specific inhibitor of topo II (D5 +). This type of experiment, as described in our previous study11, is designed to identify differentiation-dependent changes (D1 vs. D5) and to determine whether or not topo IIβ is required in the process (D5 vs. D5 +).
To assess changes in nuclear shape and chromatin state, cellular DNA was stained with Hoechst. When galleries of nuclear images from D1, D5, and D5 + were compared (Fig. 5A), the nuclear shape after 5 days in culture appeared more round and number of heterochromatic regions that are stained brightly decreased to leave a big blob in the center. However, this change does not occur when topo IIβ was suppressed by ICRF-193 (D5 +), suggesting that the enzyme is involved in the rearrangement of chromatin structure. This result was corroborated more objectively by a machine learning program called Wndchrm for image analysis28. The similarity matrix shows that the nuclear appearance is very different between D1 and D5 but it is more similar between D1 and D5 + (Fig. 5B). A 3D image analysis showed that nuclear volume enlarged by 20% during differentiation in a topo IIβ-dependent manner, which probably reflects the chromatin dispersion (Fig. 5C).
To investigate the distribution of nuclear DNA more directly, spectrum of DNA staining intensity was measured (Fig. 5D). The graph represents the histogram of brightness (256 levels in gray scale) obtained from 100 nuclear images for each condition. The three curves had an isosbestic point at the brightness 92–95 (arrow), which was regarded as a boundary between bright (right side) and dark (left side) regions that correspond to condensed (heterochromatic) and dispersed (euchromatic) chromatin compartments, respectively. The ratio of condensed chromatin was calculated from the total voxel number of condensed area and plotted (inset). The result indicates that overall chromatin state shifts toward decondensation in topo IIβ-dependent manner during differentiation.
Taken together, there exists a certain genomic region where repetitive sequences are accumulated as DSP hotspots that serve as a platform for distant chromatin interactions mediated by topo IIβ. Functional consequence of this is likely to be a chromatin decondensation leading to the alteration of transcriptional states of nearby genes (see Discussion).
Discussion
We have shown in the present study that topo IIβ have two distinctive modes of action, PSP and DSP, on genomic DNA. DSP is a novel mode of action in that DNA crossovers targeted by the enzyme are likely to be triggered by association of homologous DNA segments located at a distance.
The homology-sensing can occur between chromatin segments, not necessarily between naked DNA, because oligo-nucleosomes with identical DNA sequence were shown to associate selectively in solution29. Pairing of homologous sequences is an essential step for important biological processes such as meiotic pairing of homologous chromosomes. However, little evidence has been published on the physical basis for the association between duplexes with homologous sequences. Only recently, some theoretical approaches suggested that interaction between the two duplexes of DNA that are aligned side-by-side is most stable if the pair has a certain sequence homology and the direction of alignment is parallel instead of anti-parallel25,30. This mechanism may be adopted to explain our results indicating the terminal homology of DSP chimeras (Fig. 4D).
Here we propose a model that implies the presence of two distinct topo II-involved processes: formation and disruption of higher order chromosomal structures (Fig. 6). Initial pairing is likely to occur in ‘paranemic’ mode25. Since major grooves of paired duplex face each other at the left-handed crossover16, additional hydrogen bonds can be formed between identical base pairs in the paired DNA as suggested to occur in four-stranded DNA structure25,26. Using computational methods, Mazur proposed a convincing model for the mutual recognition between two homologous duplex DNAs through direct complementary hydrogen bonding of major groove surfaces in parallel alignment25. At the contact area, a quadruplex structure is formed between the duplexes associated with left-handed crossover configuration. The quadruplex can extend three to four consecutive base pairs called ‘recognition unit’ that have to be spaced by at least several helical turns. Therefore, the pairing requires relatively long stretches of DNA, but only partial homology. A right-handed crossover is generated between the recognition units, which is a preferred target of topo II. After strand passage between the direct repeat (homologous pairs are on the same strand), number of recognition units are doubled and the pairing mode turns to intertwined mode introducing knots into the looped domain between the homologous pairs (Fig. 6, upper section). In contrast, when homologous pairs are on opposite strands (inverted repeat), negative supercoils, instead of knots, are produced within the looped domain (Fig. 6, lower section).
The paranemic association between homologous pairs should be quite unstable without the consecutive action of topo II. Although, in principle, the topo II enzyme involved here can be either topo IIα or topo IIβ, topo IIα is the feasible one since it is the major enzyme in cycling cells that is required for the mitotic chromosome condensation in G2/M phase31. We speculate that the knotting reaction between LINEs, an abundant repetitive element in mammals, could significantly contribute to the chromatin condensation along the chromosome axis. Roles of repetitive elements in organization of higher-order chromatin structures have been proposed32.
The model predicts that at the time of eTIP-seq analysis (day 2 in CGN culture) topo II is engaged in the removal of knots between direct repeats or removal of supercoils between inverted repeats because the combination of read orientations detected in DSP chimeras was exclusively RF/FR for direct repeats and FF/RR for inverted repeats (Supplementary Table S5). The left-handed crossover in the knot will be converted to right-handed configuration by thermal wobbling before the action of topo IIβ15,16. We can state definitively that topo IIβ is the enzyme involved in these reactions since β-specific antibody was used for eTIP-seq and topo IIα is not expressed in differentiating neuronal cells. We emphasize that eTIP-seq can discriminate the directionality of topo II reaction, knotting or unknotting, by determining the read orientations on the chimera between direct repeats. The favored read orientations, RF/FR over FF/RR, in DSP chimeras with direct repeats suggests that unknotting is the preferred reaction of topo IIβ probably because knots are detrimental to transcription33,34. The unknotting between distant chromatin sites would certainly lead to an opening or decondensation of global chromatin structures as observed in this study (Fig. 5). The reason why inverted repeats are only minor target of the enzyme would be that negative supercoils are constantly removed by the action of topoisomerase I. There might be some protein factors involved in additional stabilization of DNA crossovers at DSP sites. Since about 68% of Ts3/DSP is overlapped with SP120 sites (Fig. 4C), the protein may serve as a crossover stabilizer at these sites. But it is not so at Ts2/DSP sites where SP120 is almost absent. Although the possibility that other protein is responsible for the crossover stability cannot be eliminated, we suggest that homologous pairing at Ts2/DSP sites can be conducted purely by ‘DNA-centric’ manner without any assistance from proteins.
To date, transcriptional regulation by topo IIβ has been attributed to its action within relatively narrow DNA regions like promoters or enhancers35,36. Related observation in the present study would be that Ts1 toposites are highly enriched in TSS zone (Fig. 2C). When compared to other mapping technique for topo IIβ-generated DSBs (END-seq)10, about 46% of END-seq peaks in mouse cortical neurons overlapped with Ts1 toposites (see Supplementary). Furthermore, the pattern of Ts1 distribution around TSS (shown in Supplementary Fig. S3) was very similar to that of DSB detected by END-seq in human cells37.
It is unable to exclude the possibility that the observed inhibition of chromatin de-compaction by ICRF-193 (Fig. 5) is partly caused by the inhibition of topo IIβ-dependent removal of the torsional stress in highly expressed genes. However, we suggested by mRNA-seq analysis that the alteration of chromatin compaction may not be explained solely by the effects of inhibitor on transcription in terminally differentiating neurons (see Supplementary). As to the involvement of topo IIβ in ongoing transcription, the present study implicated that the relaxation of transcription-generated supercoils occurs at the Ts1/PSP sites within genic regions (Supplementary Fig. S3). In contrast, Ts2/DSP and Ts3/DSP sites are likely to be involved in the de-compaction of global chromatin structure during differentiation that is inhibited by ICRF-193 (Fig. 5). More complete analyses on the relationship between chromatin higher-order structures and gene expression are described elsewhere38.
Methods
Animals
All animal care and experimental details have been approved by the Animal Experiment Committee of Okayama University (authorization number: OKU-2012108, OKU-2014084, and OKU-2018265). All animal experiments were conducted in accordance with the Guidelines specified by the Regulation for Animal Experiments of Okayama University.
Primary culture
The primary culture was started with cerebellar tissue obtained from Wistar rats of 8 days after birth as described previously11. Cells maintained in culture are mostly postmitotic cerebellar granule neurons (CGN) that continue to differentiate in vitro. When it is required, topo IIβ activity was inhibited specifically throughout the culture period by daily addition of 10 µM ICRF-193 that also degrades the enzyme39 (first addition was made at 12 h). Cells were harvested either at day1 (D1), day 2 (D2), day 5 (D5) or day 5 in the presence of ICRF-193 (D5 +) depending on the experiment.
eTIPa-seq
Samples were prepared by modifying the protocol described previously11. CGN cells cultured in 100-mm dishes for 2 days were treated with 0.5 mM etoposide (VP-16) in serum-free medium for 15 min. The cells on dish were lysed with 1% sarkosyl and the lysate was passed through a 23 G needle. After adding concentrated CsCl to a final concentration of 0.5 M, the mixture was sonicated with a microprobe, and diluted with 3 volumes of TEST-100 buffer (10 mM Tris–HCl: pH 7.5, 10 mM EDTA, 100 mM NaCl, 0.1% Triton X-100, and protease inhibitors). Topo IIβ cleaved complex in the supernatant was immuno-captured overnight at 4 °C with Dynabeads Protein G coated with a specific monoclonal antibody to topo IIβ (3B6)4. Beads were washed three times with TEST-200 (the same as TEST-100 except that NaCl concentration is 200 mM).
Washed beads were treated with TEST-500 buffer (the same as TEST-100 except that NaCl concentration is 500 mM) at 4 °C to elute the noncovalently associated DNA (P2 fraction) from the bead-bound DNA (P1 fraction). After digesting with RNase A and proteinase K, DNA was purified by phenol/chloroform extraction followed by ethanol-precipitation with carrier glycogen.
Purified DNA fractions (P1 and P2) were sheared to ~ 200 bp using Covaris S2 sonicator. The eTIPa-seq libraries for NGS sequencing were prepared using the TruSeq DNA LT Sample Preparation Kit (Illumina) according to manufacturer’s instructions. After 10 cycles of PCR amplification, quality of the library was monitored by Agilent 2100 BioAnalyzer. Paired end sequencing (100 bp each) was performed on Illumina HiSeq 2000.
eTIPb-seq
Entire procedure is illustrated schematically in Fig. 3A. The immuno-captured topo IIβ-DNA complex (IP complex) bound to magnetic beads was prepared as in the eTIPa-seq procedure. After washing with TEST-200 buffer, sheared DNA ends were blunted by T4 DNA polymerase and dA-tailing of 3′ ends was done with NEBNext module mix (Klenow fragment, proteinase inhibitors in NEB reaction buffer). To attach the adaptor to the ends of topo IIβ-associated DNA fragments, beads were incubated for 6 h at 22 °C in adaptor ligation mix (12,000 units of T4 DNA ligase, 5% PEG-8000, 1.5 µM ligation adaptor DNA, proteinase inhibitors in NEB ligation buffer) and the 5′ end of ligation adaptor was phosphorylated by T4 polynucleotide kinase. The ligation between adaptors (2nd ligation) by T4 DNA ligase was performed over night at 16 °C with continuous rotation. The reaction was terminated on ice with 10 mM EDTA and the bead-bound DNA (eTIPb-seq library) was purified as in eTIPa-seq procedure. The ligation adaptor DNA was prepared by annealing complementary oligomer DNAs (Operon): 5′-GTTGGATCCGATA[Bio-dT]CGC-3′ (top strand) and 5′-GGCCGCGATATCGGATCCAACT-3′ (bottom strand). Top strand contains a biotinylated deoxy thymidine (Bio-dT) and bottom strand has 3′-dT tail and 5′ overhang of GGCC for the 2nd ligation between adaptors. To anneal the oligomers, equimolar amounts of oligos were placed in boiling water for 5 min and then allowed to cool slowly to room temperature. The completion of annealing was confirmed by polyacrylamide gel electrophoresis.
To remove the biotinylated nucleotide from terminal adaptors, the eTIPb-seq library DNA was incubated with T4 DNA polymerase in the presence of dATP and dTTP. Purified DNA was sheared to ~ 200 bp by Covaris sonicator. The biotinylated DNA fragments were captured with Dynabeads MyOne Streptavidin C1 (LifeTechnologies) and the beads were washed 3 times with Tween wash buffer. The fragmented DNA on beads was end-repaired, dA-tailed and ligated with paired-end sequencing adaptor, followed by PCR amplification (15 cycles with PCR primer cocktail and PCR master mix, Illumina). Amplification products were purified using AMPure XP beads (Beckman) according to manufacturer’s instructions. Quality of the resulting library was monitored using Agilent 2100 BioAnalyzer. Paired end sequencing (100 bp each) was performed on Illumina HiSeq 2000.
ChIP-seq
A standard procedure was used for SP120 ChIP. Briefly, the CGN cells at the 2nd day in vitro (D2) were fixed with 1% formaldehyde for 10 min and quenched with 125 mM glycine for 5 min at 25˚C. Washed cells were scraped off, collected by centrifugation and stored at − 80 °C until use. The frozen cell pellets were resuspended in SDS-lysis buffer containing 0.8% SDS and incubated on ice for 10 min. The cell suspension was diluted with 3 volumes of the lysis buffer without SDS (ChIP dilution buffer) and sonicated for 2 min with intermittent poses by ultrasonic disruptor (UD-201, TOMY) attached with a special microtip (power setting at 5). After twofold dilution with ChIP dilution buffer, pre-clearing with normal mouse IgG, the final lysate was used for immunoprecipitation with anti-SP120 monoclonal antibody (1-67D23) bound to Dynabeads Protein G. The beads were washed consecutively with RIPA buffer, 500 mM-NaCl RIPA buffer, LiCl buffer, and 0.1% Triton in 10 mM Tris–HCl (pH8). Elution was performed twice with an elution buffer containing 1% SDS. After adding 5 M NaCl to a final concentration of 0.2 M, the combined eluate was heated at 65 °C for 4 h to reverse the cross-link. Sixty nanograms of purified DNA was subjected to size selection (100–600 bp range) by AMPure XP beads (Beckman) and fragmented by Covaris to generate ~ 200 bp segments. The fragmented DNA was end repaired, dA-tailed and ligated with paired-end sequencing adapter. Purified DNA was amplified with 14 cycles of PCR using PE PCR 1.0 and PE PCR 2.0 primers (Illumina) and the libraries were paired-end sequenced (100 bp each) on Illumina MiSeq platforms.
FAIRE-seq
We used the FAIRE-seq technique to map the genomic regions that are devoid of associated proteins and thus accessible to protein factors including topo II. Except for some modifications, we followed the experimental procedures for FAIRE (formaldehyde-assisted isolation of regulatory elements) described previously40. NGS protocols were the same as used for eTIPa-seq.
Mapping and peak-calling of sequence data
For eTIPa-seq and ChIP-seq analyses, the FASTQ paired-end read data were mapped by BWA against rat reference genome (UCSC rn4) and then filtered to remove frequently encountered experimental artifacts. Only paired reads of 100 bp with correct orientation and reasonable distance (< 500 bp) were selected. To eliminate repetitive sequences that are mapped multiply on genome, only uniquely mapped pairs on both ends were used for peak calling. R-package ZINBA version 2.02.03 was used to find peaks by setting ‘refinepeaks’ to 1, ‘threshold’ to 0.5, ‘extension’ to 200 and other factors to default41. Selected fragments that are flanked by uniquely and correctly mapped read pairs were used to construct wiggle track formats. Pileup data were calculated by ‘basealigncount’ function of ZINBA with extension of 200.
For eTIPb-seq analysis, the FASTQ paired-end read data were mapped by BWA to rn4 separately as ‘read 1′ and ‘read 2′. Sequence reads with the ligation adaptor sequence at 5′ end were discarded. Only uniquely mapped pairs were used for further analyses.
For FAIRE-seq analysis, the wiggle data were formatted through the same procedure as used in eTIPa-seq.
Microscopy and image processing
CGN cells at day1 (D1), day 2 (D2), day 5 (D5) or day 5 in the presence of ICRF-193 (D5 +) were fixed with paraformaldehyde and stained with Hoechst 33,342. Fluorescence images were acquired by using AxioVision 4.5 software through AxioCam MRm camera installed on an Axiovert 200 M inverted fluorescence microscope with an ApoTome optical sectioning system (Carl Zeiss).
Serial 25 optical sections were collected from 100 nuclei at 0.275 µm z-axis intervals. Measurements were performed by ImageJ software. The z-stack was cropped to obtain regions of interest. The cropped z-stack was then converted to 8-bit gray scale and subjected to gray-level histogram quantification as follows. The stacks were normalized by linearly stretching the voxel intensities over a full range of 256 Gy-levels with the ‘Histogram’ tool (termed normalized-stacks) and smoothed by a mean filter (filtered-stacks). Using the normalized-stacks, three-dimensional nucleus segmentation was automatically performed by the aid of the NucleusJ plugin ‘Nucleus segmentation’ function42 (segmented-stacks). The corresponding filtered- and segmented-stacks were combined by ‘Subtract in the Image calculator’ (combined-stacks). By using the ‘Histogram’ tool, intensities of all voxels within the combined-stacks were computed, expressed as histogram plots, and exported as text files.
To estimate the degree of chromatin condensation at various culture days, the staining intensity for 100 nuclei was recorded (8 bit, 256 Gy scale). After normalizing the intensity to total intensity per nucleus, intensity versus median frequency curve was plotted. The isosbestic point for D1, D5, and D5 + curves indicates the border between condensed and dispersed chromatin regions. The condensed region corresponds to the right side of isosbestic point.
Statistical analysis
For the continuous variables of interest, the Mann–Whitney U test with two-sided significance (wilcox.exact function of R-package, exactRankTests) was used if the overall difference was statistically significant. Box plots were drawn by R with default settings. The chi-square test or Fisher’s exact test (chisq.test or fisher.test function of R) was used to examine the differences in the categorical characteristics and changes in the groups. Analyses with R-package were conducted according to R Core Team, R: A language and environment for statistical computing, R Foundation for Statistical Computing, Vienna, Austria (2018), available online at https://www.R-project.org/.
Data availability
The raw data obtained in this study are available from DDBJ Read Archive (https://ddbj.nig.ac.jp//DRASearch/) under accession numbers of DRA007399 for eTIPa-seq, DRA007479 for eTIPb-seq, DRA007375 for SP120 ChIP-seq, DRA002525 for mRNA-seq and DRA007436 for FAIRE-seq.
References
Austin, C. A. et al. TOP2B: the first thirty years. Int. J. Mol. Sci. 19, 2765. https://doi.org/10.3390/ijms19092765 (2018).
Madabhushi, R. The roles of DNA topoisomerase IIb in transcription. Int. J. Mol. Sci. 19, 1917. https://doi.org/10.3390/ijms19071917 (2018).
Lyu, Y. L. et al. Role of topoisomerase IIb in the expression of developmentally regulated genes. Mol Cell Biol 26, 7929–7941. https://doi.org/10.1128/mcb.00617-06 (2006).
Tsutsui, K., Tsutsui, K., Sano, K., Kikuchi, A. & Tokunaga, A. Involvement of DNA topoisomerase IIb in neuronal differentiation. J Biol Chem 276, 5769–5778. https://doi.org/10.1074/jbc.M008517200 (2001).
Yang, X., Li, W., Prescott, E. D., Burden, S. J. & Wang, J. C. DNA topoisomerase IIb and neural development. Science 287, 131–134 (2000).
Chang, C. C., Wang, Y. R., Chen, S. F., Wu, C. C. & Chan, N. L. New insights into DNA-binding by type IIA topoisomerases. Curr Opin Struct Biol 23, 125–133. https://doi.org/10.1016/j.sbi.2012.11.011 (2013).
Dong, K. C. & Berger, J. M. Structural basis for gate-DNA recognition and bending by type IIA topoisomerases. Nature 450, 1201–1205. https://doi.org/10.1038/nature06396 (2007).
Roca, J. Topoisomerase II: a fitted mechanism for the chromatin landscape. Nucleic Acids Res 37, 721–730. https://doi.org/10.1093/nar/gkn994 (2009).
Baranello, L. et al. DNA break mapping reveals topoisomerase II activity genome-wide. Int. J. Mol. Sci. 15, 13111–13122. https://doi.org/10.3390/ijms150713111 (2014).
Canela, A. et al. Genome Organization Drives Chromosome Fragility. Cell 170, 507–521 e518, doi:https://doi.org/10.1016/j.cell.2017.06.034 (2017).
Sano, K., Miyaji-Yamaguchi, M., Tsutsui, K. M. & Tsutsui, K. Topoisomerase IIb activates a subset of neuronal genes that are repressed in AT-rich genomic environment. PLoS ONE 3, e4103. https://doi.org/10.1371/journal.pone.0004103 (2008).
Yu, X. et al. Genome-wide TOP2A DNA cleavage is biased toward translocated and highly transcribed loci. Genome Res 27, 1238–1249. https://doi.org/10.1101/gr.211615.116 (2017).
13Uemura, T. et al. DNA topoisomerase II is required for condensation and separation of mitotic chromosomes in S. pombe. Cell 50, 917–925 (1987).
Salceda, J., Fernandez, X. & Roca, J. Topoisomerase II, not topoisomerase I, is the proficient relaxase of nucleosomal DNA. Embo j 25, 2575–2583. https://doi.org/10.1038/sj.emboj.7601142 (2006).
Timsit, Y. DNA self-assembly: from chirality to evolution. Int. J. Mol. Sci. 14, 8252–8270. https://doi.org/10.3390/ijms14048252 (2013).
Timsit, Y. Local sensing of global DNA topology: from crossover geometry to type II topoisomerase processivity. Nucleic Acids Res 39, 8665–8676. https://doi.org/10.1093/nar/gkr556 (2011).
Chen, S. F. et al. Structural insights into the gating of DNA passage by the topoisomerase II DNA-gate. Nat Commun 9, 3085. https://doi.org/10.1038/s41467-018-05406-y (2018).
Fackelmayer, F. O., Dahm, K., Renz, A., Ramsperger, U. & Richter, A. Nucleic-acid-binding properties of hnRNP-U/SAF-A, a nuclear-matrix protein which binds DNA and RNA in vivo and in vitro. Eur. J. Biochem. 221, 749–757 (1994).
Hasegawa, Y. et al. The matrix protein hnRNP U is required for chromosomal localization of Xist RNA. Dev. Cell 19, 469–476. https://doi.org/10.1016/j.devcel.2010.08.006 (2010).
Kiledjian, M. & Dreyfuss, G. Primary structure and binding activity of the hnRNP U protein: binding RNA through RGG box. Embo j 11, 2655–2664 (1992).
Tsutsui, K. et al. Identification and characterization of a nuclear scaffold protein that binds the matrix attachment region DNA. J Biol Chem 268, 12886–12894 (1993).
Nozawa, R. S. et al. SAF-A Regulates Interphase Chromosome Structure through Oligomerization with Chromatin-Associated RNAs. Cell 169, 1214-1227.e1218. https://doi.org/10.1016/j.cell.2017.05.029 (2017).
Kawano, S., Miyaji, M., Ichiyasu, S., Tsutsui, K. M. & Tsutsui, K. Regulation of DNA Topoisomerase IIb through RNA-dependent association with heterogeneous nuclear ribonucleoprotein U (hnRNP U). J Biol Chem 285, 26451–26460. https://doi.org/10.1074/jbc.M110.112979 (2010).
Miyaji, M., Furuta, R., Sano, K., Tsutsui, K. M. & Tsutsui, K. Genomic regions targeted by DNA topoisomerase IIb frequently interact with a nuclear scaffold/matrix protein hnRNP U/SAF-A/SP120. J. Cell. Biochem. 116, 677–685. https://doi.org/10.1002/jcb.25024 (2015).
Mazur, A. K. Homologous Pairing between Long DNA Double Helices. Phys Rev Lett 116, 158101. https://doi.org/10.1103/PhysRevLett.116.158101 (2016).
McGavin, S. Models of specifically paired like (homologous) nucleic acid structures. J Mol Biol 55, 293–298 (1971).
Hurt, P. et al. The genomic sequence and comparative analysis of the rat major histocompatibility complex. Genome Res 14, 631–639. https://doi.org/10.1101/gr.1987704 (2004).
Shamir, L. et al. Wndchrm - an open source utility for biological image analysis. Source Code Biol. Med. 3, 13. https://doi.org/10.1186/1751-0473-3-13 (2008).
Nishikawa, J. & Ohyama, T. Selective association between nucleosomes with identical DNA sequences. Nucleic Acids Res 41, 1544–1554. https://doi.org/10.1093/nar/gks1269 (2013).
30DJ, O. L., Wynveen, A., Albrecht, T. & Kornyshev, A. A. Which way up? Recognition of homologous DNA segments in parallel and antiparallel alignments. The Journal of chemical physics 142, 045101, doi:https://doi.org/10.1063/1.4905291 (2015).
Carpenter, A. J. & Porter, A. C. Construction, characterization, and complementation of a conditional-lethal DNA topoisomerase IIalpha mutant human cell line. Mol. Biol. Cell 15, 5700–5711. https://doi.org/10.1091/mbc.e04-08-0732 (2004).
Tang, S. J. Chromatin Organization by Repetitive Elements (CORE): A Genomic Principle for the Higher-Order Structure of Chromosomes. Genes 2, 502–515. https://doi.org/10.3390/genes2030502 (2011).
Portugal, J. & Rodriguez-Campos, A. T7 RNA polymerase cannot transcribe through a highly knotted DNA template. Nucleic Acids Res 24, 4890–4894 (1996).
Valdes, A., Segura, J., Dyson, S., Martinez-Garcia, B. & Roca, J. DNA knots occur in intracellular chromatin. Nucleic Acids Res 46, 650–660. https://doi.org/10.1093/nar/gkx1137 (2018).
Ju, B. G. et al. A topoisomerase IIb-mediated dsDNA break required for regulated transcription. Science 312, 1798–1802. https://doi.org/10.1126/science.1127196 (2006).
Madabhushi, R. et al. Activity-Induced DNA Breaks Govern the Expression of Neuronal Early-Response Genes. Cell 161, 1592–1605. https://doi.org/10.1016/j.cell.2015.05.032 (2015).
Gittens, W. H. et al. A nucleotide resolution map of Top2-linked DNA breaks in the yeast and human genome. Nat Commun 10, 4846. https://doi.org/10.1038/s41467-019-12802-5 (2019).
Miyaji, M. et al. Topoisomerase IIb targets DNA crossovers formed between distant homologous sites to modulate chromatin structure and gene expression. Preprint at bioRxiv https://doi.org/10.1101/484956 (2019).
Isik, S. et al. The SUMO pathway is required for selective degradation of DNA topoisomerase IIb induced by a catalytic inhibitor ICRF-193(1). FEBS Lett 546, 374–378 (2003).
Simon, J. M., Giresi, P. G., Davis, I. J. & Lieb, J. D. Using formaldehyde-assisted isolation of regulatory elements (FAIRE) to isolate active regulatory DNA. Nat. Protoc. 7, 256–267. https://doi.org/10.1038/nprot.2011.444 (2012).
Rashid, N. U., Giresi, P. G., Ibrahim, J. G., Sun, W. & Lieb, J. D. ZINBA integrates local covariates with DNA-seq data to identify broad and narrow regions of enrichment, even within amplified genomic regions. Genome Biol 12, R67. https://doi.org/10.1186/gb-2011-12-7-r67 (2011).
Poulet, A. et al. NucleusJ: an ImageJ plugin for quantifying 3D images of interphase nuclei. Bioinformatics 31, 1144–1146. https://doi.org/10.1093/bioinformatics/btu774 (2015).
Acknowledgements
This work was supported by Grant-in-aid for Scientific Research (23310133 to KT) from MEXT and Grant-in-aid for Scientific Research on Innovative Areas “Genome Science” (No. 221S0002) from MEXT. We thank S. Sugano, Y. Suzuki, and Y. Kohara for providing us with NGS sequencing opportunities and suggestions on sequence analysis.
Author information
Authors and Affiliations
Contributions
M.M. designed and conducted laboratory experiments for SP120 ChIP-seq, performed bioinformatic analyses on sequence data generated by eTIPa-seq, eTIPb-seq, FAIRE-seq, and mRNA-seq experiments. R.F. conducted eTIPa-seq experiments and optimized the eTIPb-seq procedure. O.H. performed microscopic analyses. K.S. conducted FAIRE-seq experiments. N.H. and R.K. acquired the sequence data for ChIP-seq. J.K. and M.T. contributed to the mapping of sequence reads and ZINBA analysis for eTIP-seq and ChIP-seq. K.M.T. analyzed and interpreted all the results, especially data from eTIP-seq, and revised the manuscript critically to ensure textual consistency. K.T. conceived and designed the study, analyzed and interpreted the results and wrote the manuscript with help from M.M., R.F., O.H. and K.M.T. All authors read and approved the final manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Miyaji, M., Furuta, R., Hosoya, O. et al. Topoisomerase IIβ targets DNA crossovers formed between distant homologous sites to induce chromatin opening. Sci Rep 10, 18550 (2020). https://doi.org/10.1038/s41598-020-75004-w
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-020-75004-w
- Springer Nature Limited