Functional interplay between condensin I and topoisomerase IIα in single-molecule DNA compaction

Tsubota, Yuko; Shintomi, Keishi; Kinoshita, Kazuhisa; Masahara-Negishi, Yuki; Aizawa, Yuuki; Shima, Masami; Hirano, Tatsuya; Nishiyama, Tomoko

doi:10.1038/s41467-025-62600-5

Download PDF

Article
Open access
Published: 06 August 2025

Functional interplay between condensin I and topoisomerase IIα in single-molecule DNA compaction

Nature Communications volume 16, Article number: 7239 (2025) Cite this article

4271 Accesses
4 Citations
7 Altmetric
Metrics details

Subjects

A Publisher Correction to this article was published on 27 August 2025

This article has been updated

Abstract

Condensin I and topoisomerase IIα (topo IIα) are chromosomal ATPases essential for mitotic chromosome assembly. Mechanistically how the two ATPases cooperate to assemble mitotic chromosomes remains unknown. Here we investigate the interplay between condensin I and topo IIα at single-molecule resolution. While condensin I alone exhibits ATP-dependent DNA loop formation, it generates stable, compact structures (“lumps”) in the presence of topo IIα in a manner dependent on its C-terminal domain. These lumps predominantly contain a single condensin I complex and a single topo IIα dimer. The strand passage activity of topo IIα introduces DNA knots within the lumps, rendering them resistant to protease treatment. An ATP hydrolysis-deficient mutant of condensin I forms smaller lumps, in which the probability of DNA knotting is markedly reduced. Our findings demonstrate that topo IIα-mediated strand passage is functionally coupled with condensin I-mediated loop extrusion, providing insights into the mechanism underlying mitotic chromosome assembly.

Theory and simulations of condensin mediated loop extrusion in DNA

Article Open access 07 October 2021

Condensin-driven loop extrusion on supercoiled DNA

Article 14 July 2022

Chromatinization modulates topoisomerase II processivity

Article Open access 27 October 2023

Introduction

Mitotic chromosome assembly is a fundamental cellular process that ensures the faithful segregation of genetic information during the mitotic cell cycle. Extensive studies over the past decades led to a consensus that two distinct classes of ATPases, condensins and topoisomerase IIα (topo IIα), play central roles in this process^1,2. For example, the chromosome scaffold, originally characterized as a proteinaceous structure observed in histone-depleted metaphase chromosomes, was later found to contain topo IIα^3,4 and the subunits of condensin I^5,6. More recent studies have successfully reconstituted the core reaction of this process in vitro by mixing a simple substrate with only six purified proteins, including core histones, three histone chaperones, topo IIα, and condensin I^7,8.

Eukaryotic topo IIα, which belongs to the type IIA topoisomerase family, introduces a double-strand break in one DNA strand and passes through a second DNA strand before rejoining the break in an ATP hydrolysis-dependent manner^9,10,11. This strand passage reaction enables intermolecular catenation/decatenation and intramolecular knotting/unknotting of circular DNA in vitro (Fig. 1a). The direction of the reactions (i.e., topological simplification vs. topological complication) depends on the topological states of the DNA substrates and the reaction environment. Such enzymology of topo IIα is reasonably well understood.

**Fig. 1: Condensin I forms stable lumps in the presence of topoisomerase IIα.**

Much less is known about the mechanisms of action of condensin I. Early studies showed that condensin I has the ability to introduce positive superhelical tension into double-stranded DNA (dsDNA) in vitro in an ATP hydrolysis-dependent manner: relaxed circular DNA is converted into positively supercoiled DNA in the presence of type I topoisomerases¹², whereas nicked circular DNA is converted into knotted DNA in the presence of a type II topoisomerase¹³. Because Cdk1 phosphorylation of condensin I strongly stimulates these activities^13,14, they are thought to be physiologically relevant activities underlying mitotic chromosome assembly. Recent single-molecule studies have shown that condensins have the so-called loop extrusion activity that forms and expands DNA loops in an ATP hydrolysis-dependent manner^15,16. The loop extrusion activity has been proposed to be a fundamental activity that organizes mitotic chromosomes by generating consecutive loops¹⁷.

As a natural consequence of the double-stranded nature of DNA, sister DNAs become entangled after DNA replication (Fig. 1b, left)¹⁸. Substantial evidence has demonstrated that the action of condensins helps topo II to facilitate the resolution of such entanglements (inter-chromatid disentanglement) in preparation for complete sister chromatid separation during mitosis (Fig. 1b, center)^19,20. Interestingly, evidence is now accumulating for a second function of topo IIα in mitotic chromosome assembly: DNA strands within individualized chromatids become self-entangled (intra-chromatid entanglement) through the action of topo IIα (Fig. 1b, right)^8,21,22. It is speculated that this previously underappreciated action of topo IIα is facilitated by the crowded environment created during the process of mitotic chromosome assembly⁸.

Although the individual activities of topo IIα and condensin I have been studied extensively, only a limited number of efforts have been made to address the question of how the two ATPases cooperate to assemble mitotic chromosomes at a mechanistic level²³.

In the current study, we have used total internal reflection fluorescence microscopy to analyze the interplay between condensin I and topo IIα at single-molecule resolution. To this end, topo IIα is added to a standard condensin I-mediated loop extrusion assay. We find that inclusion of topo IIα converts the condensin I-mediated DNA structures into stable compact structures in a manner that is dependent on the C-terminal domain (CTD) of topo IIα. Moreover, the strand passage activity of topo IIα induces DNA knotting within the stable lumps, rendering them resistant to protease treatment. ATP hydrolysis by condensin I stimulates the topo IIα-mediated knotting reaction. Our results demonstrate how topo IIα-mediated strand passage is coupled with condensin I-mediated loop extrusion to generate a compact DNA structure. The functional implications of these observations for mitotic chromosome assembly and stabilization are discussed.

Results

Condensin I forms stable lumps in the presence of topo IIα

To address the question of how condensin I and topo IIα cooperate to change the conformation of DNA at a single molecule resolution, we introduced topo IIα to our condensin I-mediated loop extrusion assay. In brief, we utilized a flow cell-based assay system, which was visualized under a total internal reflection fluorescence microscope (TIRFM) as previously reported²⁴. 48.5-kbp λDNA was biotinylated at both ends and tethered onto a streptavidin-coated coverslip to make U-shape DNA substrates. A mammalian condensin I complex and Xenopus laevis topo IIα were expressed using the baculovirus expression system and purified as previously reported^8,25. The purified condensin I was then fluorescently labeled with Alexa488. Unless otherwise indicated, we used unlabeled topo IIα.

When condensin I alone (1 nM) was injected into the flow cells, U-shape DNA was either converted into a loop (35.6%) as reported previously^15,16,25,26 or a compact structure that consisted of a DNA mass without extruded loop (27.3%), hereafter referred to as a “lump” (Fig. 1c, see cartoon in Fig. 1d). The lumps were quantitatively distinguished from loops by measuring the roundness of DNA structures strongly stained with SYTOX Orange. To quantify roundness, the SYTOX Orange-positive DNA areas were approximated as ellipses and the ratio of the major axis (a) to the minor axis (b), i.e., a/b, was calculated. We defined loops as structures in which the a/b ratio was greater than 1.6 at least two consecutive frames during the observation and defined all other compacted structures as lumps (Supplementary Fig. 1a–c). Injection of topo IIα alone at a low concentration (0.125 nM) did not change the DNA shape. However, when condensin I and topo IIα were injected together at concentrations of 1 nM and 0.125 nM, respectively, lump formation dominated: lumps were formed on ~70% of U-shape DNA (see an example in Fig. 1d, “lump”), whereas loops were formed on less than 10% of DNA (see an example in Fig. 1d, “loop”). Importantly, we found that the lumps formed by condensin I and topo IIα were sustained for far longer time (in most cases, sustained throughout the observation time [20 min]) than those formed by condensin I alone (Fig. 1e), although the kinetics of lump formation were similar under the two conditions (Fig. 1f, g). Time-lapse imaging revealed that DNA began to be converted into lumps immediately after condensin I binding (Fig. 1f, g, and Supplementary Fig. 1d–g, Supplementary Movie 1), and the lumps grew more slowly than the loops (Fig. 1f–h). We did not observe examples of loop-to-lump conversion unless a second condensin is bound to the loops (Supplementary Movie 2), suggesting that the lump formation is initiated before loop growth. In the absence of ATP, we observed only limited numbers of lumps with reduced size (Fig. 2a, b, and Supplementary Fig. 2a, b) and reduced rates of formation (Fig. 2c), indicating that the stable lump formation is ATP-dependent.

**Fig. 2: Tracking of single DNA behaviors in the presence of condensin I and topo IIα.**

To identify how many molecules of condensin I and topo IIα are involved in the formation of stable lumps, we measured the intensities of the fluorescently labeled condensin I (Alexa488) and topo IIα (Alexa546) that bind to each lump (Supplementary Fig. 2c–e). The fluorescence intensities of single molecules of condensin I^A488 and topo IIα^Alexa546 were determined by photobleaching experiments, in which individual molecules were bleached in one step (Fig. 2e, f). By comparing these single-molecule intensities with those of condensin I and topo IIα bound to lumps, we estimated that condensin I was predominantly present as a single complex on each lump (48.1%), with smaller populations of two (24.7%) and three (11.2%) complexes (Supplementary Fig. 2c, d). Similarly, topo IIα was predominantly present as a single dimer (75%), with a minority existing as two dimers (25%) (Fig. 2g, h and Supplementary Fig. 2e–j).

Condensin I/topo IIα-mediated lumps are resistant to protease treatment

Although the lumps formed by condensin I alone were far less stable than those formed by both condensin I and topo IIα (Fig. 1e), we found that condensin I was continued to be detectable on DNA throughout the observation time [20 min] even after the lumps were dissolved (Supplementary Fig. 3a). Thus, condensin I binds stably to DNA even in the absence of topo IIα.

To further clarify the differences between condensin I-mediated and condensin I/topo IIα-mediated lumps, we then treated the two types of lumps with a protease (Proteinase K). As expected, the fluorescent signals of both condensin I and topo IIα disappeared from DNA upon protease treatment (Supplementary Fig. 3b, c). We found that the lumps and loops formed by condensin I alone were immediately dissolved upon this treatment (Fig. 3a upper, and 3b, c, Supplementary Movie 3). Remarkably, however, the lumps formed by both condensin I and topo IIα remained even after protease treatment although their size decreased slightly (Fig. 3a lower, and 3b, c). These lumps are hereafter referred to as “protease-resistant lumps”. We also found that the small, condensin I/topo IIα-mediated lumps observed in the absence of ATP (Fig. 2b) were protease-sensitive (Fig. 3c). The protease-resistant lump formation appeared to be sensitive to DNA tension, as its efficiency was decreased under continuous buffer flow, during which the DNA was continuously extended (Supplementary Fig. 3d) As the flow conditions were identical in the experiments and the distance between the tethered ends of the DNA molecules did not significantly vary, differences in DNA tension are not expected to account for the observed changes (Supplementary Fig. 3e, f). Although ~80% of the lumps formed by condensin I and topo IIα were protease-resistant, ~20% were judged to be protease-sensitive (Fig. 3c). We then compared the DNA binding kinetics of condensin I and topo IIα between the protease-resistant and protease-sensitive lumps, and found a tendency for the lumps to become more protease-resistant when condensin I bound to DNA before topo IIα (Supplementary Fig. 4a). These results suggested that condensin I creates a preferred DNA structure for topo IIα to bind and thereby generates the protease-resistant lumps. This idea was further supported by experiments involving the sequential addition of condensin I and topo IIα (Supplementary Fig. 4b).

**Fig. 3: Condensin I/topo IIα-mediated lumps are resistant to protease treatment.**

Finally, we examined changes in the size of protease-resistant lumps (after deproteinization) in response to DNA tension imposed by different flow rates. If the DNA within the lumps was topologically constrained, then it was expected that (1) the lumps would be maintained even as DNA tension increased; and (2) the apparent lump size would decrease as DNA tension increased, but it would recover as tension decreased. We evaluated the apparent lump size by measuring the cross-sectional intensity of the lumps and comparing either the highest DNA intensity on the lump (Max) or the sum of the DNA intensities on the lump (Area Under the Curve [AUC]), and found that both Max and AUC values decreased when the flow rate was increased, but they were restored when the flow rate was decreased again (Supplementary Fig. 5). The lump size was approximately 2–3-fold greater in compacted tips than that in uncompacted tips both before and after extension. These results strongly suggested that the condensin I/topo IIα-mediated lump, but not the condensin I-mediated lump, contains a topologically constrained DNA structure.

DNA strand passage by topo IIα is essential for protease-resistant lump formation

We next directly tested whether the DNA strand passage activity of topo IIα is required for protease-resistant lump formation. We expressed and purified Xenopus laevis topo IIα harboring the strand passage-deficient mutation Y803F (Supplementary Fig. 6a)^27,28. We confirmed that the mutant topo IIα (topo IIα^Y803F) was catalytically null in the standard decatenation assay in vitro (Supplementary Fig. 6b). As expected, topo IIα^Y803F bound to chromatin but failed to assemble mitotic chromosomes in Xenopus egg extracts (Supplementary Fig. 6c).

Topo IIα^Y803F was then subjected to our single-molecule assay. We found that topo IIα^Y803F supported lump formation similarly to wild-type topo IIα (topo IIα^WT) in terms of formation efficiency (Fig. 4a), kinetics (Fig. 4b), stability (Fig. 4c), size (Fig. 4d), and formation rate (Fig. 4e). These results are consistent with the observation that topo IIα^Y803F retains its ability to bind to chromatin in Xenopus egg extracts as described above. Remarkably, however, the lumps formed by topo IIα^Y803F were largely protease-sensitive, as opposed to those formed by topo IIα^WT (Fig. 4f). Thus, the strand passage activity is essential for protease-resistant lump formation, but not for stable lump formation. However, despite the fact that the topo IIα^Y803F is catalytically null, a non-negligible fraction (~10%) of lumps were judged to be protease-resistant under the current condition (Fig. 4f). We reasoned that the DNA present in the residual protease-resistant lumps might be constrained in a non-topological manner. In fact, when a higher extension force was imposed on the DNA by increasing the flow rate (from 20 μl/min to 1 ml/min; Supplementary Fig. 7), the residual protease-resistant lumps formed by topo IIα^Y803F completely disappeared (Supplementary Fig. 6d, YF), although the protease-resistant lumps formed by topo IIα^WT were maintained (Supplementary Fig. 6d, WT).

**Fig. 4: DNA strand passage by topo IIα is essential for protease-resistant lump formation.**

We next treated deproteinized lumps with freshly added topo IIα and found that most of the lumps diminished or disappeared (Fig. 4g). It is noteworthy that in most cases, DNA intensities were decreased in a stepwise manner (Fig. 4g), implying that these lumps were dissolved by one or a few step(s) of the strand passage reaction. Based on these results, we concluded that the protease-resistant lumps generated by condensin I and topo IIα^WT contain DNA knots, as reported in the previous bulk biochemical studies^13,29.

The CTD of topo IIα is required for protease-resistant lump formation

To get additional insights into the mechanism of protease-resistant lump formation, we next asked if stable binding of topo IIα to DNA is required for this process. Previous studies have shown that the intrinsically disordered CTD of topo IIα is required for stable localization of topo IIα to reconstituted mitotic chromosomes and their thickening⁸. We purified topo IIα lacking its CTD (topo IIα^ΔCTD; Fig. 5a) and found that the addition of topo IIα^ΔCTD had little, if any, impact on condensin I-mediated structural changes of DNA in our single-molecule assay (Fig. 5b). Similar to the lumps formed with condensin I alone, the lumps formed with condensin I and topo IIα^ΔCTD were all protease-sensitive (Fig. 5c). Thus, stable DNA binding mediated by the CTD of topo IIα is required for protease-resistant lump formation, even though topo IIα^ΔCTD is catalytically fully active as judged by the standard DNA decatenation assay⁸. Consistently, we confirmed that freshly added topo IIα^ΔCTD can dissolve deproteinized lumps containing knots (Supplementary Fig. 8).

**Fig. 5: The C-terminal domain of topo IIα is required for protease-resistant lump formation.**

ATP hydrolysis by condensin I strongly stimulates protease-resistant lump formation

Next, we investigated how condensin I’s ATPase activity affects lump formation. For this purpose, we prepared a fluorescently labeled, condensin I mutant complex deficient in ATP hydrolysis (Smc2[E1114Q] and Smc4 [E1218Q], henceforth referred to as condensin I^EQ)²⁵. Consistent with previous studies in all types of eukaryotic SMC complexes^16,30,31,32, condensin I^EQ alone showed no loop extrusion activity (Fig. 6a). Moreover, topo IIα-independent lump formation was also significantly suppressed (Fig. 6a). On the contrary, in the presence of topo IIα, condensin I^EQ formed lumps with a high frequency and stability indistinguishable from those observed with condensin I^WT (Fig. 6b, c).

Whereas condensin I^EQ, like condensin I^WT, bound to DNA as single complexes (Supplementary Fig. 9a, b), we noticed that condensin I^EQ formed lumps more slowly than condensin I^WT (Fig. 6d). The formation rate (Fig. 6e) and the size (Fig. 6f) of the lumps formed by condensin I^EQ were significantly smaller than those formed by condensin I^WT. Furthermore, most of the lumps formed by condensin I^EQ in the presence topo IIα^WT ( ~ 90%) were protease-sensitive (Fig. 6g). To test whether the DNA present in the residual protease-resistant lumps (~10%) is topologically constrained, we applied a higher extension force by increasing the flow rate. We found that the residual protease-resistant lumps formed by condensin I^EQ and topo IIα^WT were fully tolerant to the high extension force (Supplementary Fig. 9c), in contrast to those formed by condensin I^WT and topo IIα^Y803F (Supplementary Fig. 6d). This observation suggests that condensin I^EQ can introduce DNA knots in the presence of topo IIα^WT, but at a much lower frequency than condensin I^WT.

Discussion

The establishment of a chromatid reconstitution assay has allowed us to identify the minimal set of protein components required for the core reaction of mitotic chromosome assembly^7,8. Among the six essential components identified, condensin I and topo IIα are the only two ATPases. Although the two ATPases have extensively been studied using various single-molecule assays, they have never been studied in combination. The current study represents an attempt to study the interplay between condensin I and topo IIα at single-molecule resolution, with the aim of bridging the gap in our knowledge between the reconstitution and single-molecule assays.

Our results reported in the current study are summarized in Fig. 7a. In our current experimental setup, condensin I alone generates loops and unstable lumps on the U-shape DNA substrates (Fig. 1). We speculate that the observed unstable lumps are a precursor of the loops (while these precursors were also observed in previous studies, their frequency may vary depending on experimental conditions, including the concentration of SYTOX dye, differences in protein preparations, and the level of torsional tension imposed on the DNA^15,33,34). Topo IIα alone, at the low concentration used in the current study, has little or no impact on the conformation of the DNA substrates. However, when condensin I and topo IIα are added together, stable lump formation dominates over loop formation (Fig. 1). Importantly, this condensin I/topo IIα-mediated lump formation depends on the CTD of topo IIα (Fig. 5), but not on its catalytic activity (Fig. 4). Although we still do not know how the condensin I/topo IIα-mediated lumps are formed, the requirement for the CTD suggests that topo IIα must stay on the DNA for a long time through CTD-mediated binding to a specific DNA structure created by condensin I (e.g., DNA crossover or bent DNA). It is also possible that topo IIα recognizes a chiral DNA structure generated by condensin I, as human topo IIα has been proposed to preferentially relax positive supercoils in a CTD-dependent manner³⁵. We do not exclude the possibility that topo IIα physically interacts with condensin I on DNA, although there is no direct evidence for condensin-topo II interactions in solution (i.e., without DNA) in eukaryotic systems (e.g., ref. ³²).

**Fig. 7: Interplay between condensin I and topo IIα.**

The condensin I/topo IIα-mediated lumps become protease-resistant when topo IIα is catalytically active (Fig. 4). The following two lines of evidence strongly suggest that the protease-resistant lumps contain knotted DNA (Fig. 7a). First, the topo IIα^Y803F mutant supports the formation of stable lumps but fails to render them protease-resistant. Second, the protease-resistant lumps, after deproteinization, can be resolved with freshly added topo IIα. Most of the protease-resistant lumps are dissolved in a single step, suggesting that most of the knots generated under the current conditions are simple knots, such as 3-noded knots, similar to those observed in the previous bulk biochemistry using condensin I and a type II topoisomerase¹³.

A recent preprint using magnetic tweezers has reported that a low concentration (~0.1 nM or less) of human topo IIα displays the ability to compact DNA, possibly through a mechanism of a polymer-collapse phase transition³⁶. The compaction activity is fully dependent on the CTD of topo IIα, but is independent of ATPase activity of condensin I. Thus, the topo IIα-mediated DNA compaction reported in the preprint is clearly distinct from the condensin I/topo IIα-mediated lump formation described in the current study. Nevertheless, it is important to note that both studies highlight the hitherto underappreciated role of the topo IIα CTD in the conformational changes of DNA that is separable from its strand passage activity.

How might condensin I-mediated loop extrusion be coupled to topo IIα-mediated knotting? A mutant condensin I complex harboring the ATP hydrolysis-deficient SMC subunits (condensin I^EQ) generates lumps at the same frequency as condensin I^WT in the presence of topo IIα (Fig. 6). We notice that the size of condensin I^EQ /topo IIα-mediated lumps was substantially smaller than that of condensin I^WT /topo IIα-mediated lumps and that the former was far more protease-sensitive than the latter. Here, we interpret our observations based on the recently updated model of condensin-mediated loop extrusion³⁷. According to this model, upon ATP binding and head-head engagement, a small loop is formed in the SMC lumen (i.e., the “feeding” loop). Upon ATP hydrolysis and head-head disengagement, the feeding loop is transported into the kleisin lumen, where it is merged into the “extruding” loop (Supplementary Fig. 10). We speculate that the extruding loop generated by condensin I^WT would become an excellent substrate for topo IIα-mediated knotting. In contrast, condensin I^EQ forms a feeding loop but is unable to convert it into an extruding loop. This could explain why condensin I^EQ generates a small lump in which topo IIα-mediated knotting occurs only at a low frequency. The question of how the condensin-mediated loop extrusion process imposes superhelical tension on DNA has only recently been recognized in the field^29,38,39. The currently available data are somewhat fragmentary, and our interpretation is partially, if not completely, consistent with the emerging picture. Clearly, future studies are needed to address this important question.

What are the functional implications of the current observations for mitotic chromosome assembly and stabilization? Our finding that the CTD of topo IIα is essential for the formation of protease-resistant lumps, hence DNA knotting, would provide an important clue (Fig. 1a). It has been shown that topo IIα^ΔCTD is proficient in topo IIα-catalyzed decatenation of circular DNA but is deficient in its efficient catenation (Fig. 1a)^8,40. In addition, topo IIα^ΔCTD is proficient in inter-chromatid disentanglement but is deficient in intra-chromatid entanglement that is predicted to support chromatid thickening (axial shortening) in the reconstitution assay (Fig. 1b)⁸. Importantly, unlike topo IIα^WT, topo IIα^ΔCTD is not detectable on chromosomes during the assembly processes. These observations have been interpreted to mean that the long residence time imposed by the CTD helps topo IIα to catalyze catenation^8,40 and intra-chromatid entanglement in the reconstitution assay⁸. We speculate that the same is true for the topo IIα-stimulated lump formation and DNA knotting reported in the current study. Although direct evidence for the occurrence of intra-chromatid entanglements (self-entanglements) in mitotic chromosomes is lacking, an early study using micromanipulation approach²¹ and a recent study using high-throughput chromosome conformation capture (Hi-C) and polymer simulations²² have provided data consistent with this idea.

We consider two scenarios for the role of intra-chromatid entanglements within mitotic chromosomes (Fig. 7b). The first is the entanglement within single loops (i.e., intra-loop entanglement), which would facilitate lateral compaction and stabilization of mitotic chromosomes. The second is the entanglement between adjacent loops (i.e., inter-loop entanglement), which would promote axial shortening and stabilization of mitotic chromosomes. The products observed in the current study (DNA knots) are roughly equivalent to those resulting from intra-loop entanglements. However, we consider that it may be possible to recapitulate inter-loop entanglements by modifying the current setup of single-molecule analyses. In any case, it is reasonable to speculate that the two reactions hypothesized here, which are not mutually exclusive, would readily occur in highly crowded environments created during mitotic chromosome assembly. Our current results do not exclude the possibility that topo IIα also uses a mechanism of protein-based “locking” of two DNA duplexes to stabilize mitotic chromosomes after their assembly^41,42.

Finally, it should be added that, although the DNA present within the protease-resistant lumps is most likely a simple form of DNA knots (i.e., three-noded knots), it is uncertain whether they have specific chiralities. In this sense, it is worth mentioning that a mitotically phosphorylated form of native condensin I complex purified from Xenopus egg extracts generates DNA knots with a specific chirality (i.e., positive 3-noded knots) in the presence of a type II topoisomerase¹³. Because the recombinant condensin I complexes used in the current study are not subjected to mitotic phosphorylation, the knots discussed in this study may not have specific chiralities. A better characterization of the interplay between condensins and topo IIα in vitro and on mitotic chromosomes in situ is an important direction in the future.

Methods

Preparation of recombinant condensin I holocomplexes

For the expression of recombinant condensin complexes in insect cells, the Bac-to-Bac Baculovirus Expression System (Thermo Fisher Scientific) was used as described previously²⁵. The ATP hydrolysis-deficient mutations of mSMC2 (E1114Q) and mSMC4 (E1218Q) were introduced as described previously⁴³. These mutations were referred to as transition-state (TR) mutations in Ref#38, whereas the same mutations are referred to as EQ mutations in the current study. Purification and labeling of Halo-tagged versions of the condensin I holocomplexes were performed as described previously²⁵.

To determine fluorescence labeling efficiency of condensin I^A488, anti-FLAG M2 antibodies (Sigma Aldrich, F1804) were nonspecifically attached to the glass, then 1 nM FLAG-tagged condensin I^A488 in ELB + + was introduced at 30 μl/min for 1 min and incubated for 5 min with 5 μl/min flow. After incubation, the flow cells were washed with ELB + + at 30 μl/min for 2 min, and anti-mSmc4 antibody (AfR326-3L) in ELB-Casein (10 mM HEPES-KOH [pH7.7], 50 mM KCl, 2.5 mM MgCl_2, 0.4 mg/ml Casein) was introduced at 10 μl/min for 18 min, and washed with ELB + + at 20 μl/min for 5 min. Then 2 μg/ml Alexa Fluor 647-conjugated anti-rabbit IgG antibodies (Invitrogen) in ELB + + were introduced at 10 μl/min for 10 min and washed with ELB + + at 20 μl/min for 5 min. Finally, proteins were visualized with the 488 and 640 lasers, and the percentage of signal colocalization was calculated.

Preparation of recombinant topo IIα proteins

Plasmid DNAs encoding Xenopus laevis topo IIα (topo IIα^WT [residues 1-1579] and topo IIα^ΔCTD [residues 1-1,205]) with an N-terminal 3xFLAG-tag and a C-terminal Strep II-tag, were constructed as described previously⁸. To prepare topo IIα^Y803F, an amino acid substitution of Y803F was introduced into the original plasmid DNAs using the QuikChange XL site-directed mutagenesis kit (Agilent, 200517). An N-terminally SNAP-tagged topo IIα construct was created with a synthetic DNA fragment encoding the SNAP tag (codon-optimized for Trichoplusia ni). Using these plasmid DNAs, recombinant bacmid DNAs and baculoviruses were produced as described previously⁸. Recombinant proteins were expressed in HighFive insect cells by infecting them with amplified baculovirus. For purification of the full-length versions of topo IIα (topo IIα^WT and topo IIα^Y803F), cell lysate was subjected to tandem chromatography using Strep-Tactin Sepharose resin (IBA, 2-1201-010) and a Capto HiRes S 5/50 cation exchanger column (Cytiva, 29275877). For purification of topo IIα^ΔCTD, a combination of Strep-Tactin Sepharose resin and HiTrap Q HP anion exchanger column (Cytiva, 17115301) was used. Purified SNAP-topo IIα derivatives were fluorescently labeled with the SNAP-Surface AlexaFluor 546 ligand (New England Biolabs, S9132) according to the manufacturer’s instructions. Unconjugated ligand was removed by an additional round of cation exchange chromatography. Protein and conjugated ligand concentrations were determined by measuring absorbance at 280 nm and 558 nm, respectively. The labeling efficiency of Alexa Fluor 546–topo IIα used in this study, calculated as the ratio of the conjugated ligand concentration to the protein concentration, was estimated to be 62.5 %. All purified topo IIα proteins were dialyzed against KHG200/50 buffer (20 mM HEPES-KOH [pH 7.7], 200 mM KCl, 50% glycerol), aliquoted, snap-frozen in liquid nitrogen, and stored at −80 °C until use. The enzymatic activity of each recombinant topo IIα derivative was validated through kinetoplast DNA decatenation assays.

Preparation of microfluidic device

Single-molecule imaging was performed in cruciform flow-cell channels. For preparation of coverslips (Matsunami 22 × 22 mm No.1-S), the coverslips were sonicated in 1 M KOH for 20 min, rinsed with water, sonicated again in methanol, and left to stand for 1–3 nights in 2% 3-Aminopropyltriethoxysilane (Sigma Aldrich, 440140) in methanol. The coverslips were then treated with 100 mM NaHCO₃ containing 100 mg/ml PEG (Laysan Bio, #MPEG-SVA-5K) and 6.6 mg/ml BioPEG (Laysan Bio, # BIO-PEG-SVA-5K) overnight. The coverslips were then gently washed with water, dried, and stored at −80° C. Before use, the coverslips were treated with 80 μM Streptavidin (Sigma Aldrich, S4762) in ELB + + (10 mM HEPES-KOH [pH7.7], 50 mM KCl, 2.5 mM MgCl₂, 1 mg/ml BSA [Sigma Aldrich, 2905-OP, OmniPur BSA, Fraction V]) for at least 30 min. Excess unbound streptavidin was removed by washing with water.

The cruciform flow-cell channels were developed by putting glass slides and coated coverslips with double-sided tape (Grace Bio Labs GBL620001-1EA) with a cruciform channel of 2-mm width. Inlet (Natsume KN-392-1-SP-19) and outlet tubes (Natsume KN-392-1-SP-55) were connected to each channel, and the outlet tubes were connected to a syringe pump (Harvard Apparatus Pump 11 Elite) to withdraw solutions. The channels were washed with ELB + +.

Biotin-conjugation at both ends of λDNA (N3011, New England Biolabs) was performed as follows: 0.15 pmol λDNA and two biotinylated oligonucleotides, which are complementary to the cos sites (oligo #1, 5’-GGGCGGCGACC[biotinylated-T]−3’, 2 pmol, custom made by Integrated DNA Technologies); oligo #2, 5’-AGGT[biotinylated-T]CGCCGCCC-3’, 10 pmol, custom made by Eurofins Genomics, were phosphorylated with 0.5 μl T4 polynucleotide kinase (M0201, New England Biolabs) in the presence of 1 mM ATP at 37°C in 12 μl volume for 3 h. Oligo #1 (13.3x molar excess) and λDNA was diluted with 84 μl T4 ligation buffer (B0202, New England Biolabs), annealed by cooling the mixture gradually from 65 to 10 °C, and then ligated with 0.75 μl T4 DNA ligase (M0202, New England Biolabs) and 10 mM ATP at 16 °C for 2 h. Finally, oligo #2 (71.4 x molar excess) was annealed to another end of λDNA and ligated in the same way.

The resultant biotin-conjugated λDNA was injected into the perpendicular channel of the flow-cell at a final concentration of 11.5 pM in DNA buffer (50 mM Tris-HCl [pH7.5], 50 mM NaCl, 5 mM MgCl₂) with either 22 nM SYTOX Orange (Thermo Fisher Scientific, S11368, inFigs. 1c–h, 2a–c, 3–6, Supplementary Figs. 1, 2a–d, 3a, c–f, 4b, 5, 6d, 8, 9) or SYTOX Green (Thermo Fisher Scientific, S7020, in Fig. 2d, g, h, Supplementary Figs. 2e, 3b, 4a). The linear distance between both ends of the DNA was found to be 1 to 3 μm. Unbound λDNA was washed out first with DNA buffer and then with imaging buffer (IB: 50 mM Tris-HCl [pH7.5], 50 mM NaCl, 5 mM MgCl₂, 3 mM ATP, 1 mM DTT, 22 nM SYTOX Orange). In the experiments using Alexa546-labeled topo IIα (topo IIα^A546), high salt-imaging buffer (50 mM Tris-HCl [pH7.5], 100 mM NaCl, 5 mM MgCl₂, 3 mM ATP, 1 mM DTT, 22 nM SYTOX Green) was used to prevent non-specific binding of topo IIα^A546 (Fig. 2d, g, h, Supplementary Figs. 2e, 4a). To observe U-shape DNA, the flow direction was switched from the perpendicular channel to the horizontal channel at a flow rate of 20 μl/min (estimated tension is <0.2 pN).

Lump formation assay

Condensin I and unlabeled topo IIα were simultaneously introduced into U-shape DNA-tethered flow cells with imaging buffer at a flow rate of 20 μl/min for 100 s. Condensin I was introduced at 1 nM in all experiments and topo IIα was introduced at either 125 pM (Figs. 1c–h, 2a–c except sample without ATP, 3, 6, Supplementary Figs. 1, 2a–d, 3a, b (left), c, e, f, 9a-b) or 88 pM (Fig. 2a–c, sample without ATP, Supplementary Figs. 3d, 4b, 5, 6d, 8, 9c). Unbound proteins were washed out with imaging buffer. To prevent air from entering the flow cells, the buffer flow was transiently stopped for approximately ten seconds when exchanging buffers. When ATP dependency was tested, we used imaging buffer without ATP instead (Figs. 2a–c, 3c). When both condensin I^A488 and topo IIα^A546 were visualized, due to the overlap of the fluorescent wavelengths of Alexa488 and SYTOX Green, DNA was first detected by SYTOX Green to identify the places of U-shape DNA. After washing out SYTOX dye, DNA binding of condensin I ^A488 and topo IIα^A546 was monitored in real time. At the end of monitoring, DNA was stained again with SYTOX Green to confirm lump formation. To test protease resistance, lumps were treated with 0.5 mg/ml Proteinase K (ProK, Roche, #03115873001) in imaging buffer at a flow rate of 20 μl/min. When topo IIα^A546 was visualized, the lumps were treated by ProK after the 2nd SYTOX staining (Fig. 2d, g, h Supplementary Figs. 2e, 4a). In topo IIα reinjection assay (Fig. 4g, Supplementary Fig. 8), ProK was washed out with imaging buffer at 20 μl/min for 100 s, and then 250 pM topo IIα^WT or topo IIα^ΔCTD was reinjected with imaging buffer at 20 μl/min. In the intense wash assay (Supplementary Figs. 6d, 9c), after washing out ProK, the flow rate was increased from 20 μl/min to 1 ml/min for 50 s, and then decreased to 20 μl/min again to see whether the lumps remained. To calculate the number of folds of DNA intensity, protease-resistant lumps were formed as described above, then the flow rate was increased in a stepwise manner: at 20, 100, 200, 400 μl/min for 100 s, then at 1000 μl /min for 25 s, and finally at 20 μl/min for 100 s. To clarify the effect of flow on lump formation, proteins and DNA were incubated with and without flow. Condensin I at a concentration of 1 nM and unlabeled topo IIα at 88 pM were simultaneously introduced into the U-shape DNA binding flow cell with imaging buffer at a flow rate of 20 μl/min for 100 s. The flow cells were incubated with or without flow for another 100 s. Without changing the flow rate, the inlet tube carrying the proteins was quickly switched to a protein-free imaging buffer, and the flow cells were washed for 100 s. Similarly, the inlet tube was quickly switched to imaging buffer containing ProK without changing the flow rate. Finally, the percentage of ProK-resistant lump was measured.

In all imaging experiments except the bleaching assay, condensin I^A488 and SYTOX Green-stained DNA were visualized using a 350 μW 488-nm laser (Coherent Sapphire LP 488 ± 2, 150 mW), and topo IIα^A546 and SYTOX Orange-stained DNA were visualized using a 100 μW 561-nm laser (Coherent Sapphire LPX 561 ± 2, 480 mW). All experiments were performed at room temperature (25 ± 2 °C).

DNA lengthening using optical tweezers

Optical tweezers experiments were performed using a C-trap system (LUMICKS) to estimate the extension force on DNA. Two streptavidin-coated polystyrene beads (4.35 μm in diameter; Spherotech SVP-40-5) were laser-trapped, and topologically unconstrained biotin-conjugated λDNA was captured between the beads. After trapping, the bead-DNA complex was placed to a channel filled with imaging buffer containing SYTOX Orange (22 nM). The beads were first settled at a relaxed position (5 μm) and then moved to an extended position till double-strand denaturing (25 μm) at a rate of 0.5 μm/s. DNA force-extension (FE) curves were measured based on the external force acting on the beads and the distance between them. Experiments and data acquisition were performed under the control of BlueLake software (LUMICKS). Under extensive flow conditions (1000 μl/min), DNA extension was measured to be 13.8 μm. From the obtained FE curve, the corresponding force was estimated to be 1.2 ± 0.33 pN (Supplementary Fig. 7a).

To estimate the force under gentle flow conditions (20 μl/min), the obtained FE curve was fitted using the worm-like chain (WLC) model⁴⁴, with or without polynomial correction⁴⁵, assuming a persistence length (Lp) of 50 nm.

$$F=\frac{{k}_{B}T}{{L}_{p}}\left[\frac{1}{4{\left(1-z/{L}_{c}\right)}^{2}}-\frac{1}{4}+\frac{z}{{L}_{c}}+\sum \limits_{i=2}^{i\le 7}{\alpha }_{i}{\left(\frac{z}{{L}_{c}}\right)}^{i}\right]$$

(1)

where, the external force, Boltzmann constant, absolute temperature, the DNA extension, and contour length are denoted as F, k_B, T, z, and L_C, respectively. The order of the polynomial is denoted as arbitrary i, where α₂ = -0.5164228, α₃ = -2.737418, α₄ = 16.07497, α₅ = -38.87607, α₆ = 39.49944, and α₇ = -14.17718. The best fit was obtained using the least squares method with i = 4 and L₀ = 16.7 μm (Supplementary Fig. 7b). DNA extensions (z) under buffer flow at 20 μl/min was measured to be 6.0 ± 0.7 μm. By substituting these values into the Eq. (1), the force was calculated to be 0.07 ± 0.02 pN.

For extensible WLC (eWLC) model (2)^46,47, we tested to fit FE curve with Lc = 16.7 μm, Lp = 50 nm, and different K values (either 500, 1000 or 10,000 pN).

$$z={L}_{c}\left(1-{\frac{1}{2}\left(\frac{{k}_{B}T}{F{L}_{p}}\right)}^{\frac{1}{2}}+\frac{F}{K}\right)$$

(2)

where, K denotes the elasticity modulus. We realized it was less fit the FE curve compared to WLC with 4th polynomial correction (Supplementary Fig. 7c). All calculation and visualization were performed using Anaconda and Spyder software. For curve fitting, the curve_fit function from the scipy.stats package in the Scipy library was used. The source code used for the analysis is available at https://doi.org/10.5281/zenodo.15789118.

Bleaching assay

For the bleaching assay to detect single-molecule intensities of condensin I^A488 and topo IIα^A546 (Figs. 2e, f), 1 nM condensin I^A488 or 75 pM topo IIα^A546 was non-specifically attached to coverslips. The fluorescently labeled proteins were then bleached by either a 3.5 mW 488-nm laser or 5.0 mW 561-nm laser, the images were taken every 0.5 s, and the bleaching step was detected. The step size was determined by calculating the difference in average intensity of five frames before and after the bleaching.

DNA decatenation assay

One hundred nanograms of a catenated DNA substrate (kinetoplast DNA; TG2013, TopoGEN) were mixed with 40 ng of recombinant topo IIα in a 10-µl volume (20 mM HEPES-KOH [pH 7.7], 80 mM KCl, 5 mM MgCl₂, and 2 mM ATP), and incubated at 22 °C. At various times, aliquots were taken and treated with SDS (0.5%) and proteinase K (1.0 mg/ml; P-4032, Sigma) at 37 °C for 1 h. The resultant DNAs were purified with phenol and separated by gel electrophoresis on a 0.8% agarose gel in TAE. After being stained with ethidium bromide, fluorescent images were acquired with an image analyzer (Amersham Imager 680 [version 2.0.0], Cytiva).

Depletion and add-back assays using recombinant topo IIα in Xenopus egg extracts

The high-speed supernatant of metaphase-arrested Xenopus egg extracts (M-HSS) and demembranated Xenopus sperm nuclei were prepared as described previously^48,49. A rabbit polyclonal antibody was raised against a synthetic peptide corresponding to the C-terminal amino acid sequence of Xenopus laevis topo IIα (GRQKKPVTYLEDSDDDF) and affinity-purified (referred to here as anti-topo IIα, in-house identifier AfR474-2). One hundred microliters of M-HSS were incubated with Dynabeads Protein A (Thermo Fisher Scientific, 10002D) conjugated with 20 µg of anti-topo IIα. After two successive rounds of incubation on ice (30 min each), the supernatants were isolated from the beads under a magnetic field and used as topo IIα-depleted M-HSS for subsequent assays. Control mock-depleted M-HSS was prepared using non-immune rabbit IgG (Sigma-Aldrich, I-5006) instead of anti-topo IIα. To assess the ability of recombinant topo IIα (wild-type and Y803F mutant) to support mitotic chromatid assembly in egg extracts, sperm nuclei were incubated with either mock-depleted or topo IIα-depleted M-HSS that had been supplemented with buffer (KHG200/50) or recombinant topo IIα at 22 °C for 120 min. The resultant chromatin structures were fixed, stained with DAPI, labeled with anti-topo II monoclonal antibody (M052-3, MBL, RRID: AB_592894), and analyzed by fluorescence microscopy as described previously⁸.

Immunofluorescence labeling of FLAG-topo IIα

After lump formation by condensin I and FLAG-tagged topo IIα, the lumps were washed with imaging buffer in the presence or absence of ProK for 100 s, and then ProK was washed out by another 100-s wash. The channels were washed with ELB + + at 20 μl/min for 5 min, 1 μg/ml anti-FLAG M2 antibodies (Sigma Aldrich, F1804) in ELB + + was introduced at 10 μl/min for 18 min, and washed with ELB + + at 20 μl/min for 5 min. Then 2 μg/ml Alexa Fluor 647-conjugated anti-mouse IgG antibodies (Thermo Fisher Scientific, A-21235) in ELB + + were introduced at 10 μl/min for 10 min and washed with ELB + + at 20 μl/min for 5 min. Finally, DNAs were visualized with 11 nM SYTOX Green in ELB + +, and the localization and intensity of FLAG-topo IIα on the lump were analyzed.

Data collection and processing

Experiments were performed on a total internal reflection fluorescence microscope (Nikon Ti) and images were acquired every 0.5 or 5 s using a CCD camera (Andor DU-888 X-7952) under the control of NIS software (NIS-Elements AR 4.40 and NIS-Elements AR Analysis 5.40.01). All images were analyzed by Fiji (Image J ver 1.52p). The intensities of condensin I and topo IIα on all U-shape DNA were measured, and only those with intensities equivalent to a single molecule were used for the analyses of lump formation. Lumps and loops on U-shape DNA were classified by visual observation. The difference in shape between loops and lumps was also determined as follows. A threshold was set to distinguish loops and lumps from other parts of DNA, and these regions were selected using the wand tool of Fiji. An ellipse was then fitted to each selected region, and the ratio of the major axis to the minor axis was measured. Measurements were taken every 5 s for 60 s following the initial binding of condensin I to the DNA. Lump stability was defined as the time from the initiation of lump formation to either the disappearance of the lump or the end of the observation (up to 20 min) under a flow rate of 20 μl/min. The fluorescent intensities of DNA-bound condensin I and topo IIα were measured every 5 s, and the background was subtracted. The fluorescence intensitiy (30–50 s after the initial binding of condensin I to the DNA, as shown in Fig. 2a) was measured to estimate the number of condensin I complexes and topo IIα dimers forming each lump. The distribution of fluorescence intensities was fitted using a five-component Gaussian mixture model, with mean and variance corresponding to 1–5 fluorescent molecules, as determined by bleaching experiments. We also simulated the expected distribution of fluorescence intensities, based on the labeling efficiency, assuming different ratios of molecule numbers. All calculations and visualizations were performed using Microsoft® Excel® for Microsoft 365 MSO (ver. 2412), GraphPad Prism (Prism 9.5.1), Anaconda (conda 24.11.3), and Spyder software (Spyder IDE 5.5.1). For fitting, the norm function from the scipy.stats package in the Scipy library and the GaussianMixture function from from sklearn.mixture package in Scikit-learn library were used (https://doi.org/10.5281/zenodo.15789118). Visualization was performed using the Matplotlib library. The lump size was calculated by normalizing the intensity of lumps to the intensity of full-length λDNA and converting it to DNA length by multiplying by 48.5 kbp. The tip of U-shape DNA, where a lump is typically formed, exhibits naturally higher DNA intensity compared to the dsDNA in the side arm of U-shape DNA. Hence, the size of the tip without protein was also calculated for comparison. The lump and loop formation rates were determined based on the average growth rate as long as they continued to increase. Relative DNA length was calculated by dividing the DNA length at each frame by the average DNA length from the first five frames taken before protein injection. To determine the fold increase in lump intensities, line profiles were taken across the DNA cross-sections at the tip regions with or without lump, and at the side arms. From the resulting line profiles, the relative sum intensities (Area Under the Curve, AUC) and the relative maximum DNA intensities (Max) were calculated by subtracting background signals from the integrals of the tip and side arm DNA profiles. All images shown in the figures were background-subtracted, smoothed, and resized for clarity.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All data analyzed in this study will be made available by the authors upon request. Source data are provided with this paper.

Code availability

The Python-based data analysis code is available at https://doi.org/10.5281/zenodo.15789118⁵⁰.

Change history

27 August 2025
A Correction to this paper has been published: https://doi.org/10.1038/s41467-025-63490-3

References

Hirano, T. Condensin-based chromosome organization from bacteria to vertebrates. Cell 164, 847–857 (2016).
Article CAS PubMed Google Scholar
Paulson, J. R., Hudson, D. F., Cisneros-Soberanis, F. & Earnshaw, W. C. Mitotic chromosomes. Semin. Cell Dev. Biol. 117, 7–29 (2021).
Article CAS PubMed PubMed Central Google Scholar
Earnshaw, W. C., Halligan, B., Cooke, C. A., Heck, M. M. & Liu, L. F. Topoisomerase II is a structural component of mitotic chromosome scaffolds. J. Cell Biol. 100, 1706–1715 (1985).
Article CAS PubMed Google Scholar
Gasser, S. M., Laroche, T., Falquet, J., Boy de la Tour, E. & Laemmli, U. K. Metaphase chromosome structure. Involvement of topoisomerase II. J. Mol. Biol. 188, 613–629 (1986).
Article CAS PubMed Google Scholar
Saitoh, N., Goldberg, I. G., Wood, E. R. & Earnshaw, W. C. ScII: an abundant chromosome scaffold protein is a member of a family of putative ATPases with an unusual predicted tertiary structure. J. Cell Biol. 127, 303–318 (1994).
Article CAS PubMed Google Scholar
Maeshima, K. & Laemmli, U. K. A two-step scaffolding model for mitotic chromosome assembly. Dev. Cell 4, 467–480 (2003).
Article CAS PubMed Google Scholar
Shintomi, K., Takahashi, T. S. & Hirano, T. Reconstitution of mitotic chromatids with a minimum set of purified factors. Nat. Cell Biol. 17, 1014–1023 (2015).
Article CAS PubMed Google Scholar
Shintomi, K. & Hirano, T. Guiding functions of the C-terminal domain of topoisomerase IIalpha advance mitotic chromosome assembly. Nat. Commun. 12, 2917 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Nitiss, J. L. DNA topoisomerase II and its growing repertoire of biological functions. Nat. Rev. Cancer 9, 327–337 (2009).
Article CAS PubMed PubMed Central Google Scholar
Vos, S. M., Tretter, E. M., Schmidt, B. H. & Berger, J. M. All tangled up: how cells direct, manage and exploit topoisomerase function. Nat. Rev. Mol. Cell Biol. 12, 827–841 (2011).
Article CAS PubMed PubMed Central Google Scholar
Pommier, Y., Nussenzweig, A., Takeda, S. & Caroline, A. Human topoisomerases and their roles in genome stability and organization. Nat. Rev. Mol. Cell Biol. 23, 407–427 (2022).
Article CAS PubMed PubMed Central Google Scholar
Kimura, K. & Hirano, T. ATP-dependent positive supercoiling of DNA by 13S condensin: a biochemical implication for chromosome condensation. Cell 90, 625–634 (1997).
Article CAS PubMed Google Scholar
Kimura, K., Valentin, V. R., Crisona, N. J., Hirano, T. & Cozzarelli, N. R. 13S Condensin actively reconfigures DNA by introducing global positive Writhe: implications for chromosome condensation. Cell 98, 239–248 (1999).
Article CAS PubMed Google Scholar
Kimura, K., Hirano, M., Kobayashi, R. & Hirano, T. Phosphorylation and activation of 13S condensin by Cdc2 in vitro. Science 282, 487–490 (1998).
Article ADS CAS PubMed Google Scholar
Ganji, M. et al. Real-time imaging of DNA loop extrusion by condensin. Science 360, 102–105 (2018).
Article ADS CAS PubMed PubMed Central Google Scholar
Kong, M. et al. Human condensin I and II drive extensive ATP-dependent compaction of nucleosome-bound DNA. Mol. Cell 79, 99–114 (2020).
Article CAS PubMed PubMed Central Google Scholar
Goloborodko, A., Imakaev, M. V., Marko, J. F. & Mirny, L. Compaction and segregation of sister chromatids via active loop extrusion. Elife 5, e14864 (2016).
Article PubMed PubMed Central Google Scholar
Postow, L., Crisona, N. J., Peter, B. J., Hardy, C. D. & Cozzarelli, N. R. Topological challenges to DNA replication: Conformations at the fork. Proc. Natl. Acad. Sci. USA 98, 8219–8226 (2001).
Article ADS CAS PubMed PubMed Central Google Scholar
Baxter, J. et al. Positive supercoiling of mitotic DNA drives decatenation by topoisomerase II in eukaryotes. Science 331, 1328–1332 (2011).
Article ADS CAS PubMed Google Scholar
Dyson, S., Segura, J., Martinez-Garcia, B., Valdes, A. & Roca, J. Condensin minimizes topoisomerase II-mediated entanglements of DNA in vivo. EMBO J. 40, e105393 (2021).
Article CAS PubMed Google Scholar
Kawamura, R. et al. Mitotic chromosomes are constrained by topoisomerase II-sensitive DNA entanglements. J. Cell Biol. 188, 653–663 (2010).
Article CAS PubMed PubMed Central Google Scholar
Hildebrand, E. M. et al. Mitotic chromosomes are self-entangled and disentangle through a topoisomerase-II-dependent two-stage exit from mitosis. Mol. Cell 84, 1422–1441 (2024).
Article CAS PubMed PubMed Central Google Scholar
Baxter, J. & Aragon, L. A model for chromosome condensation based on the interplay between condensin and topoisomerase II. Trends Genet 28, 110–117 (2012).
Article CAS PubMed Google Scholar
Sakata, R. et al. Opening of cohesin’s SMC ring is essential for timely DNA replication and DNA loop formation. Cell Rep. 35, 108999 (2021).
Article CAS PubMed Google Scholar
Kinoshita, K. et al. A loop extrusion-independent mechanism contributes to condensin I-mediated chromosome shaping. J. Cell Biol. 221, e202109016 (2022).
Article CAS PubMed PubMed Central Google Scholar
Tane, S. et al. Cell cycle-specific loading of condensin I is regulated by the N-terminal tail of its kleisin subunit. Elife. 11, e84694 (2022).
Article CAS PubMed PubMed Central Google Scholar
Berger, J. M., Gamblin, S. J., Harrison, S. C. & Wang, J. C. Structure and mechanism of DNA topoisomerase II. Nature 379, 225–232 (1996).
Article ADS CAS PubMed Google Scholar
Liu, Q. & Wang, J. C. Identification of active site residues in the “GyrA” half of yeast DNA topoisomerase II. J. Biol. Chem. 273, 20252–20260 (1998).
Article CAS PubMed Google Scholar
Martinez-Garcia, B. et al. Condensin pinches a short negatively supercoiled DNA loop during each round of ATP usage. EMBO J. 42, e111913 (2023).
Article CAS PubMed Google Scholar
Davidson, I. F. et al. DNA loop extrusion by human cohesin. Science 366, 1338–1345 (2019).
Article ADS CAS PubMed Google Scholar
Pradhan, B. et al. The Smc5/6 complex is a DNA loop-extruding motor. Nature 616, 843–848 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Ryu, J. K. et al. The condensin holocomplex cycles dynamically between open and collapsed states. Nat. Struct. Mol. Biol. 27, 1134–1141 (2020).
Article CAS PubMed Google Scholar
Golfier, S., Quail, T., Kimura, H. & Brugues, J. Cohesin and condensin extrude DNA loops in a cell cycle-dependent manner. Elife 9, e53885 (2020).
Article PubMed PubMed Central Google Scholar
Shaltiel, I. A. et al. A hold-and-feed mechanism drives directional DNA loop extrusion by condensin. Science 376, 1087–1094 (2022).
Article ADS CAS PubMed Google Scholar
McClendon, A. K. et al. Bimodal recognition of DNA geometry by human topoisomerase II alpha: preferential relaxation of positively supercoiled DNA requires elements in the C-terminal domain. Biochemistry 47, 13169–13178 (2008).
Article CAS PubMed Google Scholar
Wu, M. et al. Human topoisomerase iialpha promotes chromatin condensation via a phase transition. bioRxiv, https://doi.org/10.1101/2024.10.15.618281 (2024).
Dekker, C., Christian, H. H., Peters, J.-M. & Rowland, B. D. How do molecular motors fold the genome?. Science 382, 646–648 (2023).
Article ADS CAS PubMed Google Scholar
Davidson, I. F. et al. Cohesin supercoils DNA during loop extrusion. Cell Rep. 44, 115856 (2025).
Article CAS PubMed Google Scholar
Janissen, R. et al. All eukaryotic SMC proteins induce a twist of -0.6 at each DNA-loop-extrusion step. Sci. Adv. 10, eadt1832 (2024).
Article CAS PubMed PubMed Central Google Scholar
Kawano, S. et al. DNA-binding activity of rat DNA topoisomerase II alpha C-terminal domain contributes to efficient DNA catenation in vitro. J. Biochem. 159, 363–369 (2016).
CAS PubMed Google Scholar
Nielsen, C. F., Zhang, T., Barisic, M., Kalitsis, P. & Hudson, D. F. Topoisomerase IIalpha is essential for maintenance of mitotic chromosome structure. Proc. Natl. Acad. Sci. USA 117, 12131–12142 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Meijering, A. E. C. et al. Nonlinear mechanics of human mitotic chromosomes. Nature 605, 545–550 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Kinoshita, K., Kobayashi, T. J. & Hirano, T. Balancing acts of two HEAT subunits of condensin I support dynamic assembly of chromosome axes. Dev. Cell 33, 94–106 (2015).
Article CAS PubMed Google Scholar
Marko, J. F. & Siggia, E. D. Stretching DNA. Macromolecules 28, 8759–8770 (1995).
Article ADS CAS Google Scholar
Bouchiat, C. et al. Estimating the persistence length of a worm-like chain molecule from force-extension measurements. Biophys. J. 76, 409–413 (1999).
Article ADS CAS PubMed PubMed Central Google Scholar
Wang, M. D., Yin, H., Landick, R., Gelles, J. & Block, S. M. Stretching DNA with optical tweezers. Biophys. J. 72, 1335–1346 (1997).
Article ADS CAS PubMed PubMed Central Google Scholar
Patkowski, J. B. et al. The F-pilus biomechanical adaptability accelerates conjugative dissemination of antimicrobial resistance and biofilm formation. Nat. Commun. 14, 1879 (2023).
Article CAS PubMed PubMed Central Google Scholar
Shintomi, K. & Hirano, T. A sister chromatid cohesion assay using xenopus egg extracts. Methods Mol. Biol. 1515, 3–21 (2017).
Article CAS PubMed Google Scholar
Shintomi, K. & Hirano, T. Reconstitution of mitotic chromatids in vitro. Curr. Protoc. Cell Biol. 79, e48 (2018).
Article PubMed Google Scholar
YukoTsubota YukoTsubota/condtopo: Source code of paper (ver1.01).Zenodo ta. https://doi.org/10.5281/zenodo.15789119 (2025).

Download references

Acknowledgements

We are grateful to Katsunori Fujiki and Katsuhiko Shirahige for technical support in the use of the optical tweezer system. We also thank Keiko Kimura and Mai Shimoura for the preparation of the flow-cell devices, Ryota Sakata and Yoshimi Kinoshita for technical advice on microscopic work, and Makoto M. Yoshida for discussions. The work in the authors’ laboratory was supported by JSPS Grants-in-Aid for Scientific Research/KAKENHI (#19H05755 and #22H02551 to K.S.; #23K05649 to K.K.; #18H05276 and #20H05938 to T.H.; #20H05933 and #20H05937 to T.N.). Y.T was supported by Grant-in-Aid for JSPS Fellows (#24KJ1514).

Author information

Masami Shima
Present address: Neural Circuit of Multisensory Integration RIKEN Hakubi Research Team, RIKEN Center for Brain Science, Wako, Saitama, Japan

Authors and Affiliations

Genome Integrity and Control Laboratory, Division of Biology, Graduate School of Science, Kyoto University, Sakyo, Kyoto, Japan
Yuko Tsubota & Tomoko Nishiyama
Chromosome Dynamics Laboratory, RIKEN, Wako, Saitama, Japan
Keishi Shintomi, Kazuhisa Kinoshita, Yuki Masahara-Negishi, Yuuki Aizawa, Masami Shima & Tatsuya Hirano

Authors

Yuko Tsubota
View author publications
Search author on:PubMed Google Scholar
Keishi Shintomi
View author publications
Search author on:PubMed Google Scholar
Kazuhisa Kinoshita
View author publications
Search author on:PubMed Google Scholar
Yuki Masahara-Negishi
View author publications
Search author on:PubMed Google Scholar
Yuuki Aizawa
View author publications
Search author on:PubMed Google Scholar
Masami Shima
View author publications
Search author on:PubMed Google Scholar
Tatsuya Hirano
View author publications
Search author on:PubMed Google Scholar
Tomoko Nishiyama
View author publications
Search author on:PubMed Google Scholar

Contributions

Y.T., K.S., K.K., T.H., and T.N. designed experiments. Y.T. performed the single-molecule experiments. K.S., K.K., Y.M.-N., Y.A., and M.S. purified topo IIα and condensin I complexes. Y.T., K.S., and T.N., analyzed data. Y.T., T.H., and T.N. wrote the manuscript with input from all authors.

Corresponding authors

Correspondence to Tatsuya Hirano or Tomoko Nishiyama.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information (download PDF )

Description of Additional Supplementary Files (download PDF )

Supplementary Movie 1 (download AVI )

Supplementary Movie 2 (download AVI )

Supplementary Movie 3 (download AVI )

Reporting Summary (download PDF )

Transparent Peer Review file (download PDF )

Source data

Source Data (download ZIP )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, 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 you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. 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-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Tsubota, Y., Shintomi, K., Kinoshita, K. et al. Functional interplay between condensin I and topoisomerase IIα in single-molecule DNA compaction. Nat Commun 16, 7239 (2025). https://doi.org/10.1038/s41467-025-62600-5

Download citation

Received: 24 December 2024
Accepted: 25 July 2025
Published: 06 August 2025
Version of record: 06 August 2025
DOI: https://doi.org/10.1038/s41467-025-62600-5

This article is cited by

Motorized chromosome models of mitotic chromosome folding
- Zhiyu Cao
- Chaoqun Du
- Peter G. Wolynes
Nature Communications (2025)