Skip to main content
Nature Portfolio logoLink to Nature Portfolio
. 2026 Oct 20;33(11):1666–1675. doi: 10.1038/s41594-026-01902-8

Mechanism of M18BP1-mediated activation of condensin II for mitotic DNA compaction

Alessandro Borsellini 1,✉, Sebastian Chamera 1, Valentina Cecatiello 1, Joanna Andrecka 1, Alessandro Vannini 1,✉
PMCID: PMC13759321  PMID: 42901950

Abstract

DNA compaction by human condensin II is crucial for the correct organization of mitotic chromosomes. Mitotic activation of condensin II is a multilayered process, controlled mainly by the opposing activity of the MCPH1 and M18BP1 proteins, acting as an interphase inhibitor and a mitotic activator, respectively. Mitotic phosphorylation favors M18BP1 binding to condensin II, resulting in effective and timely chromosome compaction. However, the molecular mechanisms underlying condensin II activation in driving DNA compaction remain uncharacterized. Here, combining cryo-electron microscopy with single-molecule imaging, we unravel the mechanism of condensin II activation by M18BP1 and elucidate their role in DNA compaction. The specific interaction with phosphorylated M18BP1 induces conformational changes in the condensin II complex, relieving an autoinhibitory conformation and allowing stable DNA binding and DNA compaction. Beyond the activity of isolated condensin II complexes on DNA, we observe the emergence of multimeric species that withstand the forces applied by optical tweezers, suggesting that condensin II multimerization might contribute to the mechanical reinforcement of mitotic chromosomes.

Subject terms: Cryoelectron microscopy, Mitosis, Single-molecule biophysics


Here, the authors show that phosphorylated M18BP1 activates condensin II by relieving an autoinhibitory interface, enabling stable DNA binding and compaction. Additionally, the M18BP1–condensin II complex forms clusters that might mechanically reinforce mitotic chromosomes.

Main

At the heart of eukaryotic cell division lies the dramatic reorganization of the chromatin into compact, rod-shaped chromosomes that can be faithfully divided between the daughter cells. An important step in this reorganization is the volumetric compaction of chromatin, driven by changes in core histone post-translational modifications that stabilize local interactions between chromatin fiber1,2. On top of the volumetric compaction, a series of mechanical activities carried out by condensin I and II, in concert with topoisomerase IIα and the chromokinesin KIF4A are fundamental to organize the sister chromatid axes3–7. Condensin I and II, both members of the structural maintenance of chromosomes (SMC) family, shape the higher-order architecture of chromosomes by forming chromatin loops8–10. However, increasing evidence indicates that additional mechanisms, such as interactions between neighboring condensin complexes, might contribute to mitotic chromosome assembly11,12. While condensin I and II share the core SMC2–SMC4 heterodimer, they differ in their sets of regulatory subunits. Condensin I encompasses the HEAT-repeat-containing subunits CAP-D2 and CAP-G, as well as the kleisin subunit CAP-H, while condensin II contains the closely related orthologs CAP-D3 and CAP-G2 as the HEAT-repeat-containing subunits and CAP-H2 as the kleisin subunit. Condensin II resides constitutively in the nucleus and initiates chromatin loop formation in prophase, setting the stage for the nested loop architecture further refined by condensin I upon nuclear envelope breakdown13–16.

Given its nuclear presence throughout the cell cycle, condensin II is subject to stringent regulation to ensure that chromosome condensation begins precisely at mitotic entry. Two major mechanisms have been identified thus far. First, condensin II is activated at mitosis through the interaction with the phosphorylated form of the M18BP1 protein that outcompetes the inhibitory protein MCPH1 (refs. 17–19). MCPH1 binds to condensin II, inhibiting its activity and preventing premature chromosome condensation during interphase. Second, mitotic phosphorylation of the condensin II CAP-D3 tail (residues 1297–1498) relieves an autoinhibitory interaction with the CAP-G2 subunit, enabling full activation of condensin II (ref. 20).

While the role of M18BP1 as an essential factor that triggers condensin II activity at the onset of mitosis has recently been established18, the molecular mechanisms underlying the M18BP1-dependent activation of condensin II to initiate DNA compaction are still elusive. By combining single-molecule optical tweezers and cryo-electron microscopy (cryo-EM), we elucidate the mechanism of condensin II activation by M18BP1 at the molecular level, unraveling distinct conformational states that underlie condensin II function. Binding of phosphorylated M18BP1 to condensin II stimulates the formation of stable DNA-bound complexes capable of DNA compaction. Phosphorylated M18BP1 binding induces extensive structural rearrangements of the condensin II complex, specifically relieving an autoinhibitory interface to allow productive DNA clamping and compaction.

Results

Phosphorylated M18BP1 promotes stable condensin II–DNA interaction

To investigate whether phosphorylated M18BP1873–1132 (M18BP1P) modulates the DNA-binding properties of condensin II, we used a dual-trap optical tweezer system with confocal fluorescence detection, allowing direct visualization and quantification of individual fluorescently labeled proteins interacting with DNA21–23 (Fig. 1a). Preincubation of condensin II with M18BP1P increased both the number of binding events and the dwelling time of condensin II on the DNA molecule (Fig. 1b–d). The enhanced DNA-binding activity is the direct consequence of the formation of stable condensin II–M18BP1P complexes, as doubly labeled condensin II–M18BP1P complexes colocalize on DNA and the M18BP1P protein does not bind to DNA in isolation (Extended Data Fig. 1c,d). Quantitative analysis of kymographs revealed that M18BP1P binding nearly doubled the mean residence time of condensin II on DNA (5.3 ± 0.5 s to 9.4 ± 0.6 s) (Fig. 1b,c). Mean squared displacement (MSD) analysis revealed that condensin II in the presence of ATP exhibits dynamic movement suggestive of one-dimensional diffusion and transient engagement with DNA (Fig. 1b,d). In contrast, the interaction with M18BP1P appears to limit the mobility of condensin II. The diffusion coefficient of condensin II decreased almost tenfold in the presence of M18BP1P (0.11 ± 0.01 to 0.01 ± 0.003 μm2 s−1), indicating a more stable condensin II–DNA interaction upon M18BP1P binding (Fig. 1b,d). As a control, a truncated M18BP1 mutant that lacks the phosphorylation sites required for condensin II binding (M18BP1983–10457A)18 did not alter the mode of condensin II binding to the DNA, which remained diffusive and transient with a diffusion coefficient similar to that observed for condensin II in isolation (0.11 ± 0.01 and 0.09 ± 0.01 μm2 s−1, respectively) (Fig. 1e). To assess whether DNA tension influenced DNA binding by condensin II, experiments were performed at 1 pN and 0.2 pN. Under both conditions, condensin II–M18BP1P complexes displayed indistinguishable DNA-binding dynamics (Extended Data Fig. 1e,f).

Fig. 1. Condensin II loading on DNA is enhanced by M18BP1P.

Fig. 1

a, Schematic representation and channel composition of the microfluidic device used in the optical tweezer experiments. Channels 1 and 2 were used to trap streptavidin-coated polystyrene beads (diameter: 4.1 µm) and single biotinylated λ-DNA molecules (48.5 kb), respectively. Channels 3 and 4 contained experimental buffer. Channel 5 was used for protein loading and imaging of binding events. The gray arrow represents the path of the beads throughout the experiment. b, Representative kymographs imaged in the protein channel. Red traces indicate the position over time of LD655-labeled condensin II molecules. Top: 0.5 nM LD655-labeled condensin II was incubated with λ-DNA in presence of 3 mM ATP. Bottom: 0.5 nM LD655-labeled condensin II mixed with a 20-fold excess of M18BP1P was incubated with the λ-DNA in presence of 3 mM ATP. Kymographs were recorded for 80 s. c, Histogram from eight different kymographs of condensin II DNA-binding events in presence (blue) and absence (green) of M18BP1P. d, MSD plots of condensin II trajectories along the DNA in the presence (blue) and absence (green) of M18BP1P. e,f, MSD plots of condensin II trajectories along the DNA in presence of M18BP17A and ATP (e; yellow) or M18BP1P and ADP-BeF3 (f; magenta). g, Violin plot representing the distribution of diffusion coefficients extracted from the MSD data in the four different conditions shown in d–f. Statistical significance was assessed using an ordinary one-way ANOVA, followed by Tukey’s multiple-comparisons test. ****P < 0.0001; NS, not significant (P > 0.05).

Source data

Extended Data Fig. 1. Single molecule DNA binding controls.

Extended Data Fig. 1

(a,b) Schematic representation of the five-channel microfluidic device used to test protein binding to individual λ-DNA molecules. Gray arrow represents the different experimental phases. (a) M18BP1P labelled with Alexa Fluor 488 (M18BP1P-488) is introduced into the protein channel in the presence of ATP, without condensin II. (b) M18BP1P-488 is introduced together with condensin II labelled with LD655 (CII-655) in the presence of ATP. (c) Representative kymograph showing no detectable DNA binding of M18BP1P-488 in the absence of condensin II. (d) Representative kymograph showing DNA binding of M18BP1P-488 in the presence of condensin II-655 and ATP. (e,f) Comparison between DNA binding experiments performed at beads distance equal to 14.5 µm or 11 µm. (g, h) Comparison between DNA binding experiments performed in presence of condensin II at 2 nM and ADP-BeF3, with (h) or without (g) M18BP1P. (i, j) Comparison between DNA binding experiments performed in presence of condensin II 5 nM, with (j) or without (i) M18BP1P in absence of nucleotides.

To determine whether ATP hydrolysis by condensin II is required for stable DNA binding, condensin II was incubated with M18BP1P and ADP beryllium fluoride (ADP-BeF3), which locks the ATPase domains in a prehydrolytic state (Fig. 1f). Stable and nondiffusive DNA binding by condensin II was observed with a diffusion coefficient similar to that observed in the presence of M18BP1P and ATP (0.009 ± 0.004 μm2 s−1) (Fig. 1f and Extended Data Fig. 1g,h). In the absence of nucleotides, no DNA-binding events were detected. Limited DNA binding was observed only after increasing the condensin II concentration tenfold; in these conditions, M18BP1P also enhanced condensin II loading onto DNA (Extended Data Fig. 1i,j).

Taken together, these experiments demonstrate that M18BP1P stimulates the formation of abundant, stable, nondiffusing condensin II–M18BP1P–DNA complexes, which do not require ATP hydrolysis for their association with DNA (Fig. 1g).

Stable condensin II–M18BP1P complexes induce DNA compaction

Next, we asked whether the stable M18BP1P–condensin II complexes could induce DNA compaction and performed relax–stretch assays (Fig. 2a), as previously described for meiotic protein SYCP3 and yeast cohesin24,25. In presence of ATP, stable condensin II–M18BP1P complexes consistently generated force–distance profiles indicative of DNA compaction, whereby the measured force between the beads increased at lower distances with a distinctive sawtooth shape (Fig. 2b–d and Extended Data Fig. 2a–d). Interestingly, the fluorescent traces that were initially spatially separated along the DNA consistently converged into a single, more intense fluorescent trace (Fig. 2c and Extended Data Fig. 2a,d). This phenomenon suggests that, in the presence of ATP and M18BP1P, condensin II molecules bound to relaxed DNA coalesce together to form protein–DNA clusters that withstand high forces. Imaging the same experiments in wide-field modality revealed the fast formation of individual protein–DNA clusters that could not be resolved by applying high forces (Fig. 2e and Supplementary Video 1). In absence of M18BP1P, no DNA compaction activity was detected and the resulting force–distance curves did not deviate from those obtained in the presence of the λ-DNA molecule in isolation (Extended Data Fig. 3a). Importantly, the ability to form multimeric protein–DNA clusters appears to be conserved among human condensins. In the same experimental setting, condensin I compacted DNA and formed stable protein–DNA clusters similarly to what was observed for the condensin II–M18BP1P complexes (Extended Data Fig. 3b,c), in agreement with recent data26. Thus, while M18BP1P binding is required to restructure condensin II into a DNA compaction-competent complex, condensin I is able to constitutively compact DNA in the absence of auxiliary factors.

Fig. 2. Stable condensin II–M18BP1P complexes compact the DNA.

Fig. 2

a, Schematic of the relax–stretch assay. After observing DNA-binding events in the protein channel, the tethered λ-DNA is transferred to the buffer channel, ensuring no additional DNA-binding events (Extended Data Fig. 2a). In the buffer channel, the distance between the beads is reduced from 12 to 5 µm, allowing relaxation of the λ-DNA molecule. At 5 µm, the DNA and bound proteins are incubated for 1 min in the presence of 3 mM ATP, after which the DNA is stretched to 17 µm while monitoring force–distance curves. b, Force–distance curve of λ-DNA molecule in absence of proteins. c, Kymographs of the relax–stretch assay performed in the presence of 1 nM condensin II, 10 nM M18BP1P (AlexaFluor488) and 3 mM ATP. Below the kymograph, a plot of the respective force recorded during the experiment is shown. d, Force–distance curves of λ-DNA molecule from eight replicates of the experiment shown in c. e, Snapshots of relax–stretch experiment performed in the same conditions as in c but imaged using wide-field microscopy. The light-blue signal comes from M18BP1P proteins labeled with AlexaFluor488, while the green signal comes from SYTOX dye that intercalates the DNA molecule. f, Force–distance curves of stretch–relax assay at different concentrations of condensin II ranging from 2 nM (dark blue) to 0.25 nM (light green), while keeping the concentration of M18BP1P fixed at 10 nM. Each condition is the average of five replicates. g, Representative kymograph of relax–stretch assay performed with condensin II (0.25 nM), M18BP1P (10 nM) and ATP (3 mM). Below the kymograph, a plot of the respective force recorded during the experiment is shown. h, Quantification of fluorescence intensity of traces from kymograph shown in g, reporting the mean values and Gaussian fits. Black symbols indicate the trace of the kymograph from which the histogram was extracted.

Source data

Extended Data Fig. 2. Quantification of fluorescence intensity traces from relax-stretch assay.

Extended Data Fig. 2

(a,b) Representative kymographs of relax–stretch assays performed with condensin II-655 at 2 nM and M18BP1P 10 nM, in presence (a) or absence (b) of ATP 3 mM. (c) Force-distance curves of experiments shown in panel a-b, in presence (green) and absence (blue) of ATP 3 mM, each repeated up to 20 replicates. (d) Representative kymograph of relax–stretch assay performed with condensin II 2 nM and M18BP1P-488 10 nM and ATP 3 mM. On the right, the respective force-distance curve. (e) Quantification of fluorescence intensity of traces from kymograph shown in panel d, reporting mean values and Gaussian fits. Blue histograms arise from traces before relaxation while green histograms arise from traces after relaxation and DNA compaction.

Source data

Extended Data Fig. 3. DNA compaction by condensin II is concentration dependent.

Extended Data Fig. 3

(a) In magenta, force–distance curves of λ-DNA molecule in presence of condensin II 1 nM, and ATP 3 mM. In blue, force–distance curves of λ-DNA molecule after adding M18BP1P at 10 nM in the same channel of condensin II. (b) Force-distance curves of λ-DNA molecule in presence of condensin I 1 nM with ATP 3 mM from five technical replicates. (c) Same experiment as in panel b, using identical buffer conditions used for condensin II. (d) Kymographs showing relax-stretch assay of single λ-DNA molecules in the presence of AF488-labelled M18BP1P at 10 nM and increasing concentrations of condensin II (CII): 0.25 nM, 0.5 nM, 1 nM, and 2 nM. White dotted lines outline beads position during the experiment. Composition of microfluidic channel is indicated schematically to the left of each kymograph. (e) Force-distance (FD) curves obtained with the relax-stretch assay in panel d. Curves represent DNA force-distance profiles following incubation with the respective condensin II concentrations. DNA compaction correlates with shortening of the DNA molecule, visible as a shift from the typical force-distance profile expected for a DNA molecule of given length (red line).

Source data

Two modes of DNA compaction

The stretch–relax assays carried out in the presence of condensin II and M18BP1P displayed consistent shortening of the λ-DNA contour length (Fig. 2c–f), representative of persistent compaction events that are resistant to the applied force. In addition, smaller sawtooth-shaped peaks could be observed at shorter distances, which might indicate rupture of DNA loops24,25.

To further characterize the DNA compaction behavior observed in the stretch–relax assay in the presence of stable condensin II–M18BP1P complexes, we performed a serial dilution experiment of condensin II (ranging from 0.25 nM to 2 nM) in the presence of a fixed concentration of M18BP1P (10 nM; Fig. 2f). Even at the lowest concentration probed, where individual condensin II–M18BP1P complexes were spatially separated by several kilobases, the force–distance profiles displayed appreciable DNA compaction features (Fig. 2f–h and Extended Data Fig. 3d,e). As expected, the degree of DNA compaction increased proportionally with the number of stable complexes bound to the DNA at higher concentrations (Fig. 2f and Extended Data Fig. 3d,e). To probe whether DNA compaction requires condensin II ATPase activity, a previously characterized ATPase-deficient mutant (condensin IIQL)27 was used. Surprisingly, while condensin IIQL–M18BP1P complexes were unable to drive DNA compaction at low protein concentrations, an appreciable degree of DNA compaction could be detected at higher protein concentrations, even in the absence of ATP hydrolysis (Extended Data Fig. 4). A similar behavior was observed for wild-type condensin II–M18BP1P in the presence of ADP-BeF3 (Extended Data Fig. 4a–c). To assess the effect of ionic strength on DNA compaction, we performed stretch–relax assays at a near-physiological salt concentration (150 mM potassium glutamate). Under these conditions, higher condensin II concentration was required to visualize multiple DNA-binding events. Nevertheless, DNA compaction and protein clustering were still observed (Extended Data Fig. 4d,e). Lastly, to investigate the role of condensin II phosphorylation, λ-phosphatase was used to remove phosphorylation modifications introduced during protein expression in insect cells (Extended Data Fig. 4f). Following phosphatase treatment, DNA compaction activity in the presence of M18BP1P and ATP was reduced, relative to the untreated protein (Extended Data Fig. 4g). Although condensin II purified from insect cells carries multiple phosphorylation sites, as revealed by mass spectrometry analyses18, phosphorylation of the functionally important CAP-D3 residue T1415 was not detected and its potential contribution to the DNA compaction activity observed here remains to be determined20,28.

Extended Data Fig. 4. DNA compaction activity dependency on ATP hydrolysis and protein concentration.

Extended Data Fig. 4

(a,b,c) Force–distance curves of λ-DNA molecule in three different experimental conditions as described in figure panels: (a) condensin IIWT with M18BP1P in presence of ATP. (b) condensin IIQL with M18BP1P and ATP. (c) condensin IIWT with M18BP1P and ADP-BeF3. High concentration (2 nM) and low concentration (0.25 nM) are shown in green and blue, respectively. High concentration experiments were performed in triplicate for each condition, whereas low concentration experiments were repeated eight to thirteen times. (d) Kymograph showing relax-stretch assay performed at physiological salt concentration (150 mM Potassium Glutamate) in the presence of AF488-labelled M18BP1P at 20 nM and condensin II (CII) at 5 nM. (e) Force–distance curves for experiments performed as in panel d. (f) SDS–PAGE analysis of condensin II before (UT) and after (λ) treatment with λ-phosphatase. The gel was first stained with Pro-Q Diamond to detect phosphorylated proteins (left) and subsequently stained with Coomassie to visualize total protein (right). The experiment was independently repeated two times with similar results. (g) Force–distance for experiments performed using condensin II at 2 nM with (blue) or without (green) treatment with λ-phosphatase, in presence of M18BP1P.

Source data

Taken together, these results suggest that two distinct mechanisms contribute to the DNA compaction observed in the stretch–relax assay: (1) an ATP hydrolysis-dependent mechanism that enables individual condensin II complexes to compact the DNA, associated with reversible rupture events, and (2) an ATP hydrolysis-independent self-association mechanism in condensin II–M18BP1P complexes that likely compacts the DNA when the bound proteins are in close proximity, associated with persistent compaction events.

Cryo-EM structure of the condensin II–M18BP1P-clamped DNA complex

Single-molecule assays demonstrated that the binding of M18BP1P strongly activates condensin II binding to DNA. To understand how M18BP1P activates condensin II, we used cryo-EM. Condensin II and M18BP1P were incubated with a 235-bp double-stranded DNA in a 1:2:4 ratio and mixed with the ATP analog ADP-BeF3 before crosslinking with sulfo-NHS-diazirine (sulfo-SDA). As in previous structural studies of SMC complexes29–34, the short DNA substrate used here may bypass aspects of the physiological DNA loading reaction. Individual particles isolated from the micrographs were refined to an average resolution of 6.8 Å, with the core of the complex reaching a local resolution of 5.8 Å (Fig. 3a, Extended Data Fig. 5a–g and Table 1).

Fig. 3. Cryo-EM structures of condensin II complexes.

Fig. 3

a, Cryo-EM map of condensin II–M18BP1P complex bound to clamped DNA. SMC2, green; SMC4, purple; CAP-D3, pale blue; CAP-H2, yellow; DNA, gray. b, Model of condensin II bound to clamped DNA, colored as in a. Dotted yellow lines represent the approximate position of the CAP-H2 disordered regions. c, Cryo-EM map of condensin II bound to two ADP-BeF3 molecules. SMC2 is colored in green, SMC4 in purple, CAP-D3 in pale blue, CAP-G2 in pink and CAP-H2 in yellow. d, Model of condensin II bound to two ADP-BeF3 molecules colored as in c. Dotted yellow and pale blue lines represent the approximate positions of the CAP-H2 disordered regions and CAP-D3 C-terminal tail, respectively.

Extended Data Fig. 5. Cryo-EM data analysis of condensin II bound to DNA.

Extended Data Fig. 5

(a) Representative micrograph from three independent cryo-EM data collections with similar results. (b) 2D class averages from full dataset. (c) Fourier Shell Correlation (FSC) curves for raw map obtained with RELION. (d) Model to map Fourier Shell Correlation (FSC) curves obtained with phenix. (e) Schematic representation of the cryo-EM data processing pipeline, with final map colored by local resolution. (f) Model fit to map for the whole condensin II tetramer (SMC2 green, SMC4 purple, CAP-D3 lightblue and CAP-H2 yellow) bound to DNA (black). (g) Orientation distribution in final set of refined particles. (h) Model of the CAP-H2 N-terminal domain (yellow), in the conformation of human condensin II in the clamped DNA conformation. (i) Model of RAD-21 N-terminal domain (yellow), in the conformation of human cohesin in the clamped DNA conformation (PDB: 6WG3). (j) Model of Brn1 N-terminal domain (yellow), in the conformation of yeast condensin in the clamped DNA conformation (PDB: 7Q2X).

Table 1.

Cryo-EM data collection, refinement and validation statistics

Condensin II + ADP-BeF3 (EMD-56221, PDB 9TSZ) Condensin II + ADP-BeF3 + DNA (EMD-56247, PDB 9TTP) Condensin II + M18BP1P (EMD-56273, PDB 9TUU)
Data collection and processing
Magnification 105,000 79,000 165,000
Voltage (kV) 300 200 300
Electron exposure (e− per Å2) 50 60 60
Defocus range (μm) 0.8–2.0 0.8–2.0 0.5–1.5
Pixel size (Å) 1.2 1.589 0.748
Symmetry imposed C1 C1 C1
Initial particle images (no.) 3.7 million 1.2 million 388,000
Final particle images (no.) 49,000 70,000 52,000
Map resolution (Å) 4.6 6.8 4.1
 FSC1 threshold 0.143 0.143 0.143
Map resolution range (Å) 4.2 to ~7 5.8 to ~10 3.8 to ~6.5
Refinement
Initial model used (PDB code) Model Angelo + AlphaFold-multimer predictions AlphaFold monomer predictions PDB 9F5W
Model resolution (Å) 3.9 6.2 4.1
 FSC threshold 0.143 0.143 0.143
Model resolution range (Å) 3.9 to ~7 5.8 to ~10 4.1 to ~6.5
Map sharpening B factor (Å2) −80 −50 −90
Model composition
 Nonhydrogen atoms 25,049 21,173 26,942
 Protein residues 3,097 2,413 3,342
 Ligands 4 3 0
B factors (Å2)
 Protein 41.48 52.79 110.96
 Ligand 4.66 40.29 NA
Root-mean-square deviations
 Bond lengths (Å) 0.009 0.007 0.005
 Bond angles (°) 1.149 1.271 1.202
Validation
 MolProbity score 1.41 1.59 1.43
 Clashscore 6.73 6.51 5.76
 Poor rotamers (%) 1.77 1.48 1.34
Ramachandran plot
 Favored (%) 97.84 96.58 97.42
 Allowed (%) 2.16 3.3 2.55
 Disallowed (%) 0.04 0.13 0.03

1FSC, Fourier shell correlation.

The resulting structure displays typical features of DNA-bound SMC complexes, with the SMC2 and SMC4 ATPase domains, CAP-D3 and CAP-H2, that form a ‘clamped’ head module29–32. The SMC domains are connected through the ATPase heads and the coiled coils are open in divergent directions (Fig. 3b). The DNA lies on top of the ATPase heads and is held in place by the CAP-D3 subunit that interacts with both SMC2 and SMC4 neck domains and clamps the DNA from above (Fig. 3b). The rest of the SMC coiled coil and the hinge domain remain flexible and are unresolved in the density map. The CAP-H2 N-terminal domain is bound to the SMC2 neck domain, further encircling the DNA molecule, in a conformation that resembles the ‘closed gate’ observed in other SMC complexes bound to DNA (Extended Data Fig. 5h–j). Despite being present during sample preparation, no EM density was obtained for the CAP-G2 subunit. Accordingly, M18BP1P, which binds to condensin II through the CAP-G2 subunit, is also not visible in the final density map. This phenomenon is reminiscent of what is observed in structures of yeast condensin bound to the DNA29–31, where Ycg1, the yeast CAP-G2 homolog, is also found only flexibly connected to the head module.

Cryo-EM structure of condensin II in the autoinhibited state

While attempting to solve the cryo-EM structure of a condensin II–M18BP1P bound to DNA in the presence of ADP-BeF3, most particles aggregated on the grid and were unusable for image processing (Extended Data Fig. 6a). However, we identified a subpopulation of individual particles lacking both M18BP1P and DNA. The resulting structure of condensin II bound to two ADP-BeF3 molecules was refined to an average resolution of 4.6 Å and locally to 4.2 Å (Fig. 3c and Extended Data Fig. 6a–g). The SMC2 and SMC4 ATPase heads are engaged and adopt a conformation competent for ATP hydrolysis, with the two coiled coils of SMC2 and SMC4 that separate at the neck domains but reconnect after ~25 residues, forming a zipped-up coiled-coil structure (Fig. 3d and Extended Data Fig. 6f,h). The elbow and hinge domains are only visible at low contour level because of intrinsic flexibility (Extended Data Fig. 6e). The CAP-G2 subunit hooks both SMC subunits by binding simultaneously to the SMC2 head and the SMC4 coiled-coil region. Additionally, the CAP-H2 N-terminal domain is bound between the CAP-G2 subunit and the SMC2 ATPase head and undergoes substantial structural rearrangement when compared to the DNA-bound condensin II structure (Fig. 3d and Extended Data Fig. 6i,j). The CAP-H2 C-terminal domain binds the SMC ATPase heads in a conformation reminiscent of other SMC structures29,30,33, while its long unstructured domain interacts with both HEAT-repeat subunits (residues 148–202 bound to CAP-D3 and residues 316–367 bound to CAP-G2) (Extended Data Fig. 7a). Lastly, the CAP-D3 subunit is bound to CAP-G2 through its HEAT docker helices and remains on the edge of the complex without interacting with the SMC moieties (Fig. 3c,d). Additional CAP-D3 C-terminal residues (1454–1465 and 1469–1476) could be fitted and refined into the density map after modeling with AlphaFold3 (ref. 35) (Extended Data Fig. 7b,c). Interestingly, this region of the CAP-D3 C-terminal tail has been shown to mediate an autoinhibitory interaction with the CAP-G2 subunit20. In agreement with this observation, the cryo-EM structure reveals that the CAP-D3 C-terminal tail anchors the N-terminal domain of CAP-H2 and the CAP-G2 subunit (around residues 386–418), stabilizing this tripartite interaction (Extended Data Fig. 7b,c). Thus, the cryo-EM structure of condensin II in the presence of ADP-BeF3 reveals a novel conformation characterized by an unusual arrangement of HEAT-repeat-containing subunits that leads to an autoinhibited conformation of the condensin II complex, as highlighted by the superimposition with the structure of the DNA-bound condensin II–M18BP1P complex (Extended Data Fig. 7d).

Extended Data Fig. 6. Cryo-EM data analysis of condensin II bound to two ADP-BeF3 molecules.

Extended Data Fig. 6

(a) Representative micrograph from one cryo-EM data collection. (b) 2D class averages from full dataset. (c) Fourier Shell Correlation (FSC) curves for raw map obtained with RELION. (d) Model to map Fourier Shell Correlation (FSC) curves obtained with phenix. (e) Schematic representation of the cryo-EM data processing pipeline. (f) Model fit to map for the whole condensin II pentamer (SMC2 green, SMC4 purple, CAP-D3 lightblue, CAP-G2 pink, CAP-H2 yellow), with zoom in onto the regions with highest resolution. (g) Orientation distribution in final set of refined particles. (h) Detail of the ABC ATPase domain with the two ADP-BeF3 molecules represented as sticks. (i) Detail of the structure and position of the CAP-H2 N-terminal domain sandwiched between the SMC2 neck and the CAP-G2 subunit. (j) Upper panel, detail of the CAP-H2 N-terminal domain sandwiched between the SMC2 neck and CAP-G2 subunit in the structure of condensin bound to two ADP-BeF3 molecules. Lower panel, detail of CAP-H2 N-terminal domain rearranged in the condensin II DNA bound conformation.

Extended Data Fig. 7. Condensin II in the inhibited conformation.

Extended Data Fig. 7

(a) Detail of the CAP-H2 interaction sites with condensin II subunits. (b) Cryo-EM density map of condensin II bound to two ADP-BeF3 molecules zoomed in the region of interaction of the CAP-D3 C-terminal tail with CAP-G2 and the CAP-H2 N-terminal helices. (c) Model of condensin II bound to two ADP-BeF3 molecules, with highlighted residues of CAP-D3 tail that interact with CAP-G2 and the CAP-H2 N-terminal helices. (d) Superimposition of condensin II structure in the autoinhibited conformation with condensin II structure bound to DNA colored in grey. Dashed line represents potential clash between the SMC4 subunit of condensin II and the DNA molecule. (e) AlphaFold3 prediction of the CAP-G2 - M18BP1 interaction, with relative PAE plot in panel (f). (g,h) Comparison between the human condensin II autoinhibited conformation and the yeast head segment plus ATP conformation (PDB: 6YVD).

M18BP1P relieves condensin II autoinhibition

To address how M18BP1P activates condensin II, we obtained a novel cryo-EM structure of the condensin II–M18BP1P complex at a higher resolution (local resolution: 3.8 Å) (Fig. 4a and Extended Data Fig. 8). While the overall architecture closely resembles the previously published condensin II–M18BP1P complex18, the higher-resolution map reveals additional density for M18BP1P on the surface of the CAP-G2 subunit (Fig. 4a,c). Comparison of the structure of condensin II in its autoinhibited conformation to the structure of the condensin II–M18BP1P complex highlights the dramatic rearrangement of CAP-G2 and CAP-D3 subunits, while the overall conformation of the SMC2, SMC4 and CAP-H2 subunits remains largely unchanged. In the presence of M18BP1P, the two HEAT-repeat-containing subunits act as a rigid body, undergoing a rototranslational movement spanning ~60 Å (Fig. 4b,c and Supplementary Video 2). This movement disrupts the contacts between CAP-G2 and both the SMC2 ATPase head and the SMC4 coiled coil. Importantly, the novel SMC2 head–CAP-G2 interface observed in the apo structure is in close proximity of the M18BP1P-binding site for CAP-G2 (Fig. 4c). The cryo-EM map validates an extended interaction interface between M18BP1P and CAP-G2 predicted by AlphaFold3, including M18BP1 residues whose phosphorylation enhances binding to condensin II (ref. 18) (Extended Data Fig. 7e,f). This suggests that M18BP1P binding to the CAP-G2 subunit could destabilize the CAP-G2–SMC2 interaction, promoting the structural rearrangement of the HEAT-repeat-containing subunits. To test this hypothesis, we performed microscale thermophoresis (MST) experiments to directly assess the interaction between the SMC2/4 dimer and CAP-G2. Titration of SMC2/4 against a fixed concentration of fluorescently labeled CAP-G2 resulted in a measurable affinity (dissociation constant KD equal to 1.3 ± 0.6 μM) (Fig. 4d). In contrast, no binding was detected in the presence of 2 µM M18BP1P, demonstrating that M18BP1P binding prevents the interaction between the SMC2/4 dimer and CAP-G2 (Fig. 4d). As expected, such inhibition of binding was not observed in presence of the M18BP17A mutant, which is impaired in condensin II binding. This observation, together with the structural rearrangements observed in the condensin II–M18BP1P complex in isolation and when bound to DNA, suggests that M18BP1P binding facilitates the detachment of CAP-G2 from its autoinhibitory position against the SMC2 ATPase head, thereby enabling the transition of condensin II into the DNA-bound clamped conformation (Fig. 4e).

Fig. 4. M18BP1P binding relieves an autoinhibitory conformation of condensin II.

Fig. 4

a, Cryo-EM map of condensin II bound to M18BP1P. b, Model of condensin II bound to two ADP-BeF3 molecules, with detailed insight into the movement of the CAP-G2 and CAP-D3 subunits in the transition to the M18BP1P-bound condensin II (right). Arrows indicate the rototranslational movement of the heat repeats. c, Comparison between the CAP-G2 interaction surface in the presence (left) and absence (right) of M18BP1P. d, MST curves with titration of SMC2/4 from 10 µM to 240 pM over a fixed concentration (10 nM) of CAP-G2-647. Blue and magenta data points represent experiments performed in the presence of 2 µM M18BP1P or M18BP17A, respectively. Red data points represent the experiment performed in the presence of 2 µM BSA. The curve fit was obtained by combining the data from three independent experiments. e, Model of condensin II bound to two ADP-BeF3 molecules, with arrows representing the movement of the SMC4, CAP-G2 and CAP-D3 subunits in the transition to the DNA-bound condensin II (right). f, Illustrations highlighting the conformational changes happening upon M18BP1P binding or M18BP1P in combination of ATP and DNA.

Source data

Extended Data Fig. 8. Cryo-EM data analysis of condensin II bound to M18BP1P.

Extended Data Fig. 8

(a) Representative micrograph from two independent cryo-EM data collections with similar results. (b) 2D class averages from full dataset. (c) Schematic representation of the cryo-EM data processing pipeline, with final map colored by local resolution. (d) Fourier Shell Correlation (FSC) curves for raw map obtained with RELION. (e) Model to map Fourier Shell Correlation (FSC) curves obtained with phenix. (f) Model fit to map for the whole condensin II - M18BP1P complex (SMC2 green, SMC4 purple, CAP-D3 lightblue, CAP-G2 pink, CAP-H2 yellow, M18BP1P blue), with zoom in onto the regions with highest resolution. (g) Orientation distribution in final set of refined particles.

Discussion

The activation of condensin II at mitotic onset is a multilayered process that involves specific phosphorylation of its subunits by mitotic kinases20,28 and the simultaneous binding of the activator M18BP1 protein, which outcompetes the inhibitory MCPH1 protein, a process also regulated by phosphorylation. Here, we provide detailed insights into the molecular mechanism defining the activation of condensin II by M18BP1P and the resulting effects on DNA compaction activity. Optical tweezers combined with fluorescence microscopy experiments revealed that binding of condensin II to DNA in the presence of ATP is increased in both frequency and duration when M18BP1P is present. Moreover, the presence of M18BP1P favors the transition from a diffusive condensin II to a nondiffusive, stably anchored complex on DNA. These results align and explain previous in vivo evidence showing that M18BP1P is essential for condensin II localization on chromatin18,19.

The cryo-EM structural data of DNA-bound and apo M18BP1P–condensin II complexes provide a potential molecular mechanism for the change in behavior of condensin II observed with the optical tweezer experiments (Fig. 5). In the absence of both M18BP1P and DNA, condensin II adopts a conformation where the CAP-G2 subunit bridges the ATPase heads of the SMC subunits, forming a closed architecture that represents an autoinhibited state not competent for DNA entrapment. Binding of M18BP1P promotes extensive remodeling of the condensin II architecture, displacing CAP-G2 from its inhibitory position and priming the complex for DNA engagement. This structural transition is more evident in the clamped DNA-bound conformation, solved in the presence of both ADP-BeF3 and M18BP1P. In this state, CAP-G2 disengages from the core of the complex, likely remaining associated through the interaction with the flexible CAP-H2 linker. Phosphorylation of the CAP-D3 C-terminal tail has also been characterized to be important for releasing CAP-G2 from its autoinhibited state20. In agreement with this evidence, the cryo-EM structure of the condensin II complex in its autoinhibited conformation reveals the intimate connection linking both the CAP-D3 heat docker domain and the CAP-D3 C-terminal tail with the CAP-G2 subunit (Fig. 3c,d and Extended Data Fig. 7b,c). It is interesting to note that the CAP-D3 tail is not present in the map of the condensin II–M18BP1P complex. Hence, M18BP1P binding could favor the dissociation of the CAP-D3 tail from the CAP-G2 subunit, overriding to an extent the need for phosphorylation of the CAP-D3 tail to activate condensin II. This hypothesis is in line with the evidence that M18BP1P is required and sufficient to trigger premature chromosome condensation in the absence of MCPH1 and mitotic phosphorylation18. Thus, M18BP1P binding likely represents the major activating mechanism of condensin II at mitotic onset, while CAP-D3 C-terminal tail phosphorylation might additionally contribute by disfavoring rebinding of the CAP-G2 subunit in its autoinhibitory conformation.

Fig. 5. Model of condensin II activation.

Fig. 5

(I) Condensin II in the autoinhibited conformation, with CAP-G2 (pink) that bridges the ATPase heads. (II) The autoinhibited conformation of condensin II is released upon M18BP1P binding. (III) Phosphorylation of the CAP-D3 tail, release of the CAP-G2 subunit, ATP-mediated opening of the SMC2/4 coiled coil and DNA binding induce the formation of the DNA-clamped conformation.

In summary, two autoinhibitory redundant mechanisms provide a stringent control of condensin II activation. Together, the binding of M18BP1P to CAP-G2 and the phosphorylation of the CAP-D3 tail trigger the transition of condensin II into an active complex capable of clamping the DNA, with the CAP-G2 subunit disengaging from the core complex. Once CAP-G2 is displaced from its inhibitory position, entrapment of the DNA could happen. It has been proposed that the SMC head engagement promoted by ATP binding enables the powerstroke motion of the CAP-D3–kleisin to swing the DNA within the SMC ring30.

The CAP-G2 autoinhibitory mechanism is unique to condensin II because, in yeast, the CAP-G2 ortholog Ycg1 interacts with the SMC2 head domain without interfering with DNA binding36 (Extended Data Fig. 7g,h). Moreover, Ycg1 disengages from the SMC2 head domain upon DNA binding, in the absence of auxiliary factors29,30. This mechanism likely holds true for human condensin I, which exhibits a higher degree of conservation with yeast condensin. Accordingly, in our optical tweezer experiments, human condensin I does not require additional factors to bind and compact DNA (Extended Data Fig. 3b,c). Nevertheless, recent studies have shown that interaction between the CAP-G subunit and KIF4A can further stimulate the ATPase and loop extrusion activities of condensin I (ref. 37), highlighting how higher eukaryotes have evolved additional regulatory layers to fine-tune condensin activation. A similar principle applies to human cohesin, whose activity is modulated through interactions between the STAG1/2 subunits and multiple cofactors that promote distinct outcomes, such as cohesin unloading (that is, WAPL), loop stalling and localization (that is, CTCF, CENP-U) or protection from release (that is, SGO1)38–41.

The activity of SMC complexes has been largely associated with the ability to form loops of DNA, as demonstrated by single-molecule assays where individual complexes actively form loops of DNA in a process that is ATP dependent8,10,42–44. The optical tweezer experiments shown here highlight how the formation of multiprotein assemblies in both condensin I and condensin II have an important role in compaction of the DNA molecule. This agrees with recent models that propose how loop extrusion by individual complexes is needed at the initial stages of mitotic chromosome folding and is then complemented by an additional mechanism where neighboring nonextruding condensins interact with each other to form functional assemblies11,12,15,45. Further work will be required to define the molecular determinants of the multiprotein assemblies observed here and to assess whether preventing their formation influences the architecture and stability of mitotic chromosomes.

Methods

Protein expression

Human condensin II, condensin II subcomplexes, condensin II Q-loop mutant and condensin I were assembled into biGBac vectors as described previously10,46,47. Recombinant bacmids were generated by Tn7 transposition in DH10EMBacY (Geneva Biotech) cells and transfected into Sf9 cells (Thermo Fisher Scientific, 11496015). Virus-containing supernatant was harvested after 3 days and further amplified in Sf9 cells for another 3 days. For protein expression, amplified virus was used to infect High Five insect cells (Thermo Fisher Scientific, B85502). High Five insect cells were harvested after 3 days, centrifuged at 5,000g for 30 min at 4 °C and flash-frozen in liquid nitrogen. MBP–M18BP1873–1132 and MBP–M18BP1983–10457A were expressed in Escherichia coli Rosetta (DE3) (Millipore Sigma (Novagen), 71401-4). Cells were grown in Terrific Broth medium at 37 °C. Upon reaching an optical density of ~4, cells were induced for 2 h at 30 °C, then centrifuged at 5,000g for 30 min at 4 °C and flash-frozen in liquid nitrogen.

Protein purification

Cell pellets containing condensin II complexes were resuspended in 20 mM HEPES pH 8.0, 300 mM KCl, 5 mM MgCl2, 1 mM DTT and 10% glycerol, supplemented with one Pierce EDTA-free protease inhibitor tablet (Thermo Scientific) and 25 U per ml Benzonase (Sigma). Cells were lysed using a Dounce homogenizer followed by sonication for 5 min (10 s on and 20 s off) and clarified by centrifugation for 40 min at 4 °C using a Beckman Coulter F20 rotor at 50,000g. Clarified lysates were loaded onto a Strep-Tactin XT column (IBA Lifesciences), washed with lysis buffer and eluted with the same buffer supplemented with 50 mM biotin (Sigma). Protein fractions were pooled and diluted threefold in buffer A (20 mM HEPES pH 8.0, 5 mM MgCl2, 5% glycerol and 1 mM DTT). The diluted sample was loaded onto a HiTrap Heparin HP column (Cytiva), washed with buffer A containing 250 mM NaCl and eluted with buffer A containing 500 mM NaCl. Lastly, protein was further purified by size-exclusion chromatography using a Superose 6 Increase 16/600 (Cytiva) equilibrated in 20 mM HEPES pH 8, 300 mM KCl, 5 mM MgCl2, 1 mM DTT and 10% glycerol. Fractions containing purified complexes were pooled, concentrated and flash-frozen in liquid nitrogen.

Condensin II was dephosphorylated by incubation with λ-protein phosphatase (1:50 molar ratio relative to condensin II) for 30 min at 30 °C in the size-exclusion chromatography buffer supplemented with 10 mM MnCl2. Dephosphorylation was verified by SDS–PAGE followed by sequential staining with Pro-Q Diamond phosphoprotein stain (Thermo Scientific) and InstantBlue Coomassie protein stain (Abcam).

Bacterial pellets expressing MBP–M18BP1873–1132 and MBP–M18BP1983–10457A constructs were resuspended in 30 mM HEPES pH 7.5, 200 mM NaCl, 10% glycerol and 2 mM DTT and lysed by sonication on ice at 60% amplitude, 30 s on and 30 s off, for a total of 5 min. After centrifugation for 30 min at 4 °C using a Beckman Coulter F20 rotor at 50,000g the supernatant was injected on a HisTrap column and eluted with an increasing concentration of imidazole. The HisTrap purification was followed by MBP affinity purification and size-exclusion chromatography using Superdex 200 16/600 column (Cytiva). Phosphorylation of M18BP1 was obtained by treating the purified protein overnight at 4 °C as previously described48.

For human condensin I, pellets were resuspended in lysis buffer (25 mM HEPES pH 8.0, 150 mM NaCl, 5 mM MgCl2, 1 mM DTT and 10% glycerol), supplemented with a protease inhibitor cocktail (Calbiochem) and 25 U per ml Benzonase (Sigma). Cells were lysed using a Dounce homogenizer, followed by sonication for 5 min (15 s on, 15 s off), and the lysate was clarified by centrifugation for 60 min at 4 °C using a Beckman Coulter F20 rotor at 50,000g. The clarified lysate was loaded onto a StrepTrap HP 5-ml column (Cytiva) equilibrated with lysis buffer and proteins were eluted with the same buffer supplemented with 50 mM biotin (Sigma). Eluted protein fractions were pooled and applied to a HiTrap Heparin HP column (Cytiva) equilibrated with heparin buffer (25 mM HEPES pH 8.0, 150 mM NaCl, 5% glycerol and 1 mM DTT). Proteins were eluted using a linear NaCl gradient from 150 to 1,000 mM. Fractions of the highest purity were pooled and loaded onto a Superose 6 Increase 10/300 column (Cytiva) equilibrated with gel filtration buffer (25 mM HEPES pH 8.0, 150 mM NaCl and 1 mM DTT). Selected fractions were concentrated, flash-frozen in liquid nitrogen and stored at −80 °C.

Protein labeling

Labeling of condensin II complexes was achieved by covalently attaching fluorescent dyes conjugated to coenzyme A (CoA) to the YBBR tag of SMC2 subunit. Briefly, YBBR tagged condensin II at 1 µM was incubated with Lumidyne 655 fluorophore conjugated to CoA (LD655–CoA, Lumidyne Technologies) at 5 µM, in the presence of 0.5 µM SFP (4′-phosphopantetheinyl transferase) for 1 h at 4 °C in the dark. Subsequently, the reaction mixture was purified over a Superose 6 Increase 10/300 GL column equilibrated in 30 mM HEPES pH 7.5, 150 mM NaCl, 2 mM DTT, 0.05% (v/v) Tween-20 and 5% glycerol (Extended Data Fig. 9a). Fractions (50 μl) were collected and analyzed by SDS–PAGE using 4–12% NuPAGE Bis–Tris gels (Invitrogen) running in MOPS buffer at 200 V for 45 min. Imaging of the gel was performed on Chemidoc with optimal settings for imaging at 655 nm, after which the gel was stained with Instant Blue Coomassie protein stain (Abcam) (Extended Data Fig. 9b). Labeled condensin II fractions were pooled, aliquoted and flash frozen in liquid nitrogen at a concentration of 200 nM. The degree of labeling (DOL) was calculated with the following formula, yielding a labeling efficiency of 64%.

DOL=A655×εprotein(A280−A655×CF)×εdye×100=64%
A655=0.131nanodropreadoutat650forLD655
A280=0.38nanodropreadoutat280
εprotein=44,5210M−1cm−1
εdye=239,000M−1cm−1
CF=0.03correctionfactorforLumidyne655at280nm

MBP–M18BP1873–1132 was labeled by covalently attaching AlexaFluor dyes to cysteine residues present on the protein. To remove DTT, MBP–M18BP1873–1132 at 12 µM was buffer-exchanged using ZebaSpin columns (volume: 0.5 ml) into 30 mM HEPES pH 7.5, 150 mM NaCl and 5% glycerol. Following buffer exchange, the protein was incubated with AlexaFluor488 C5 maleimide at 50 µM for 1 h at 4 °C in the dark. The reaction was then quenched by the addition of Tris buffer (pH 7.5) to a final concentration of 50 mM and subsequently loaded onto a Superdex 200 Increase 10/300 GL column (Cytiva) equilibrated in 30 mM HEPES pH 7.5, 150 mM NaCl, 2 mM DTT, 0.05% (v/v) Tween-20 and 5% glycerol (Extended Data Fig. 9c). Fractions (50 μl) were collected and analyzed by SDS–PAGE using 4–12% NuPAGE Bis–Tris gels (Invitrogen) running in MES buffer at 200 V for 30 min. Imaging of the gel was performed on Chemidoc with optimal settings for imaging AlexaFluor488, after which the gel was stained with Instant Blue Coomassie protein stain (Abcam) (Extended Data Fig. 9d). Labeled MBP–M18BP1873–1132 fractions were pooled, aliquoted and flash-frozen in liquid nitrogen at a concentration of 1 µM. The DOL, calculated with the previous formula using a correction factor of 0.11 (for AlexaFluor488 at 280 nm), was 89%.

Extended Data Fig. 9. Condensin II and M18BP1P labelling.

Extended Data Fig. 9

(a) Size-exclusion chromatography profile (Superose 6 Increase, 24 mL column volume) of condensin II following labeling with LD655-CoA and Sfp transferase. Grey and red traces indicate absorbance at 280 nm and 650 nm, respectively. (b) SDS–PAGE analysis of fractions from size exclusion chromatography in (a). Left panel, Coomassie blue staining. Right panel, fluorescence image acquired upon excitation at 650 nm. Labelling of condensin II was performed independently twice, with similar results. (c) Size exclusion profile (Superdex 200 increase, 2.4 mL volume) of M18BP1P following labelling reaction with AlexaFluor 488-maleimide. Grey and blue traces indicate absorbance at 280 nm and 495 nm, respectively. (d) SDS–PAGE analysis of unlabeled (A) and labeled (B) M18BP1P. Left panel, Coomassie blue staining. Right panel, fluorescence image acquired upon excitation of Alexa Fluor 488. Labelling of M18BP1P was performed independently twice, with similar results.

Source data

Single-molecule DNA-binding experiments

Single-molecule DNA-binding experiments were performed on a C-trap (Lumicks) integrating optical tweezers, confocal fluorescence microscopy and microfluidics controlled by BlueLake software (Lumicks). Every channel of the microfluidic device was first cleaned with 2% Helmanex and MQ water and then equilibrated in 20 mM HEPES pH 7.5, 25 mM KCl, 5 mM MgCl2, 1 mM TCEP and 0.04 % (v/v) Tween-20. Streptavidin-coated polystyrene beads, 0.005% w/v (4.35 μm, Spherotech), were injected into channel 1. Biotin-labeled λ-DNA molecules (about 2 pM) were flowed into channel 2. Channels 3 and 4 contained experimental buffer supplemented with 3 mM ATP. Channel 5 contained labeled protein diluted in experimental buffer. The optical trap was calibrated to achieve a trap stiffness of 0.2–0.3 pN nm−1. After trapping two beads in channel 1, the DNA was captured in channel 2. Successful tether formation was verified in channel 3 by acquiring force–distance profiles, by extending the distance between the beads to 16.5 µm. The DNA was then transferred in the protein channel (channel 5) and the proteins were allowed to incubate for 1 min at tension below 1 pN. The incubation step was carried out in static conditions with all channel valves closed. During the incubation step, confocal fluorescence imaging was performed with laser excitation at 488 or 638 nm for AlexaFluor488 or LD655, respectively. Emission was collected through bandpass filters at 512/25 nm (blue) and a long-pass filter at 640 nm (red). Kymographs were generated by line scanning along the DNA axis with a dwell time of 0.2 ms per pixel, a pixel size of 100 nm, a line width of 27.3 nm and a line time of 60 ms. Kymographs for DNA-binding experiments were recorded with total acquisition times ranging from 50 to 200 s. All measurements were conducted at room temperature (21 °C). Five to six technical replicates were acquired per experimental condition.

For condensin I, experiments were carried out in the same manner, with slight changes in buffer composition that were previously optimized for condensin I stability. In this case, the microfluidic device was equilibrated with buffer containing 50 mM HEPES pH 7.0, 50 mM NaCl, 1 mM MgCl2, 1 mM DTT and 0.02% (v/v) Tween-20. As a control, we also performed experiments with condensin I under the exact same conditions used for condensin II, using a buffer containing 20 mM HEPES (pH 7.5), 25 mM KCl, 5 mM MgCl2, 1 mM TCEP and 0.04% (v/v) Tween-20.

Kymograph analysis and extraction of single-molecule trajectories

Kymographs were analyzed in Python using custom Jupyter Notebooks. Kymograph files were imported using the lumicks.pylake library and processed with the KymoWidgetGreedy tool to identify individual fluorescence trajectories. Traces were detected automatically using the greedy line-tracking algorithm (KymoWidgetGreedy) with the following parameters: pixel intensity threshold, 0.5; Gaussian smoothing sigma, 0.4; window size, 6 pixels; minimum track length, 8 pixels; maximum display intensity, 8; origin correction enabled. Manual correction of the identified traces was performed to remove false positives and to connect traces when needed.

MSD analysis

For each detected trace, the position along the kymograph (in µm) was extracted as a function of frame index and converted to a time point using the known temporal calibration of the acquisition (0.3 s per pixel). Traces shorter than three points or with insufficient temporal coverage were discarded. The MSD was calculated in one dimension for each trajectory using a standard time-averaging approach. Briefly, for a given trace with positions x(t), the MSD at a lag time τ was computed as follows:

MSD(τ)=〈[x(t+τ)−x(t)]2〉t,

where the average is taken over all valid time points t within the trace. Lag times were sampled from one frame up to half the trajectory length to ensure sufficient statistics at longer lags. For each trace, the MSD as a function of lag time was saved to CSV files, together with the corresponding trace index and lag times.

To compare multiple kymographs and biological repeats, MSD curves from all traces and experiments belonging to the same condition were pooled. This was achieved by reading all CSV files in condition-specific folders (for example, for different nucleotide or protein conditions) and plotting the MSD versus lag time for each individual trace. This provided a visual assessment of the spread and shape of MSD curves across all trajectories and conditions.

Estimation of diffusion coefficients

Diffusion coefficients were extracted from the initial, approximately linear regime of the MSD–lag time curves, assuming one-dimensional Brownian diffusion. For each trajectory, the MSD versus lag time data were first filtered to remove nonpositive lag times and NaN values and sorted by increasing lag time. Only traces with at least three lag time points and a total duration of at least 1 s (maximum lag time ≥ 1 s) were retained for fitting. The first k points (with k between 4 and 6, depending on trace length) were then used to perform an ordinary least-squares linear regression with an intercept, MSD(τ)=2Dτ+b, yielding a slope and intercept for each trace. The one-dimensional diffusion coefficient D was obtained as half of the fitted slope (D = slope/2). For each experimental condition, the per-trace diffusion coefficients were compiled into a single table and exported as a CSV file. The number of traces, mean diffusion coefficient, s.d. and s.e.m. were computed per condition. To visualize the distribution of diffusion coefficients across conditions, violin plots were generated using Prism. Statistical analysis was performed in GraphPad Prism using an ordinary one-way ANOVA to assess differences between experimental conditions. These plots allowed a direct comparison of the distribution and central tendency of diffusion coefficients obtained from multiple kymographs and independent experiments under the different biochemical conditions.

Single-molecule relax–stretch experiments

Single-molecule relax–stretch experiments were performed under the same conditions as the DNA-binding assays, except for the protein binding and relax–stretch procedure. A custom script was used to automate the experimental sequence. At the start of each measurement, all channel valves were closed and the presence of a single DNA tether was verified in channel 3 by recording a force–distance profile. The tether was then transferred into channel 5 for protein binding, where it was held at a bead-to-bead distance of 14 µm under a force of <1 pN and incubated for 3–5 s. The tether was subsequently moved to channel 4 and an additional 5 s waiting period was included in this channel to confirm the presence of bound proteins. Protein activity was then allowed by decreasing the bead-to-bead distance to 5 µm and incubating for 1 min, after which the distance was increased to 17 µm. A speed of 1 µm s−1 and 0.5 µm s−1 were used to respectively decrease and increase the distance between the beads during the stretch–relax assay. Throughout the entire procedure, the applied force was continuously monitored and confocal fluorescence imaging was conducted with the same settings as described for the single-molecule DNA-binding experiments. Kymographs for relax–stretch assays were recorded with total acquisition times of about 120 s.

Analysis of force–time and force–distance data

Force data were extracted directly from C-Trap.h5 files using the lumicks.pylake Python library. For each dataset, the force channel (force LF/force 2×) was used to export the force values from the time window of the stretch–relax assay. Time (s) and force (pN) values were retrieved from the ‘seconds’ and ‘data’ attributes, respectively. The force–time trace was exported and plotted using GraphPad Prism, with fixed axis limits and a 0-pN reference line for consistent visualization.

Force–distance curves were obtained by reading the corresponding force (force LF/force 2×) and distance (distance/distance 1) channels from each h5 dataset over a defined time window. These values were exported to a CSV file and force–distance curves were plotted using GraphPad Prism, with force on the y axis and distance on the x axis.

Cryo-EM sample preparation

Sample for the structure of condensin II bound to the DNA was obtained by mixing purified condensin II at 0.5 µM and M18BP1873–1132 (previously treated with CDK1–cyclin B–CKS1) at 1 µM in the presence of 0.5 mM ADP, 0.5 mM BeSO4, 5 mM NaF and a 235-bp DNA molecule (Extended Data Table 1) at 2 µM in 30 mM HEPES pH 7.5, 25 mM NaCl, 5 mM MgCl2, 2 mM DTT, 0.03% β-glucopyranoside and 0.05% Tween-20. After 10 min incubation, sulfo-SDA (Thermo Scientific) at 1 mM was added to the mixture and incubated for 30 min at 25 °C to allow the NHS moiety to react with lysine residues. The crosslinked protein material was then treated with ultraviolet light (wavelength: 330–370 nm) on ice for 10 min. The crosslinking reaction was quenched with 30 mM Tris pH 7.5 and the freshly prepared sample was applied to Quantifoil R1.2/1.3 300-mesh copper holey carbon grids, previously glow-discharged using a GloQube (Quorum) set at 30 mA for 60 s. Grids were blotted for 3 s under 100% humidity at 4 °C before being plunged into liquid ethane using a Vitrobot Mark IV (Thermo Scientific).

Extended Data Table 1.

DNA sequence used for cryo-EM sample preparation

PRIMER TCAGCTTCCACTGGAGAATCCCGGTGCCGAGGCCGCTCAATTGGTCGTAGACAGCTCTAGCACCGCTTAAACGCACGTACGCGCTGTCCCCCGCGTTTTAACCGCCAAGGGGATTACTCCCTAGTCTCCAGGCACGTGTCACATATATACATCCTGTAAGTATTTCGTCCACTATTTTCGGCTACTATAAAAAAATGTTTTTTTCGCAACTATGTGTTCGCGAAGTAACCCTTCG
TEMPLATE CGAAGGGTTACTTCGCGAACACATAGTTGCGAAAAAAACATTTTTTTATAGTAGCCGAAAATAGTGGACGAAATACTTACAGGATGTATATATGTGACACGTGCCTGGAGACTAGGGAGTAATCCCCTTGGCGGTTAAAACGCGGGGGACAGCGCGTACGTGCGTTTAAGCGGTGCTAGAGCTGTCTACGACCAATTGAGCGGCCTCGGCACCGGGATTCTCCAGTGGAAGCTGA

Identical procedure was followed to prepare the sample for the structure of condensin II in the autoinhibited conformation, with the only difference that condensin II was added at 1 µM instead of 0.5 µM before crosslinking. Under these conditions, we could identify a subpopulation of individual particles lacking both M18BP1P and DNA.

Sample for the structure of condensin II–M18BP1P complex was obtained by mixing purified condensin II at 1 µM and M18BP1873–1132 (previously treated with CDK1–cyclin B–CKS1) at 3 µM in presence of 0.5 mM ADP, 0.5 mM BeSO4 and 5 mM NaF for 30 min in buffer containing 30 mM HEPES pH 7.5, 200 mM NaCl, 2 mM DTT and 0.04% Tween-20. The sample was then run on a glycerol (10–25%) and glutaraldehyde (0–0.1%) gradient. Ultracentrifugation with a Sw60 rotor set on a G-force of 29,000 was performed for 16 h at 4 °C. After the GraFix fixation, the complex was run on a Superose 6 Increase 10/300 GL column equilibrated in 30 mM HEPES pH 7.5, 200 mM NaCl, 2 mM DTT and 0.04% Tween=20. The peak fractions were concentrated and applied to Quantifoil R1.2/1.3 300-mesh copper holey carbon grids, previously glow-discharged using a GloQube (Quorum) set at 30 mA for 60 s. Grids were blotted for 3.5 s under 100% humidity at 4 °C before being plunged into liquid ethane using a Vitrobot Mark IV (Thermo Scientific).

Cryo-EM data acquisition

Micrographs for the condensin II–DNA structure were acquired as EER movies on a Glacios (Thermo Scientific) transmission EM instrument operated at 200 kV with a Falcon 4i (Thermo Scientific) direct electron detector and Selectris X (Thermo Scientific) energy filter set with a slit width of 10 eV. EPU (Thermo Scientific) software was used for automated data collection following standard procedures. A magnification of ×79,000 was used for imaging, yielding a pixel size of 1.598 Å on images. The defocus range was set from −0.5 to −2 μm. Each micrograph was dose-fractionated to 60 frames, with a total dose of about 60 e− per Å2.

Micrographs for condensin II in autoinhibited conformation were acquired as EER videos on a Titan Krios G4 (Thermo Scientific) transmission EM instrument operated at 300 kV with a Falcon 4i (Thermo Scientific) direct electron detector and Selectris X (Thermo Scientific) energy filter set with a slit width of 10 eV. EPU (Thermo Scientific) software was used for automated data collection following standard procedures. A magnification of ×130,000 was used for imaging, yielding a pixel size of 0.955 Å on images. The defocus range was set from −0.5 to −2 μm. Each micrograph was dose-fractionated to 60 frames, with a total dose of about 60 e− per Å2.

Micrographs for condensin II–M18BP1P structure were acquired as EER videos on a Titan Krios G4 (Thermo Scientific) transmission EM instrument operated at 300 kV with a Falcon 4i (Thermo Scientific) direct electron detector and Selectris X (Thermo Scientific) energy filter set with a slit width of 10 eV. EPU (Thermo Scientific) software was used for automated data collection following standard procedures. A magnification of ×165,000 was used for imaging, yielding a pixel size of 0.748 Å on images. The defocus range was set from −0.5 to −2 μm. Each micrograph was dose-fractionated to 60 frames, with a total dose of about 60 e− per Å2.

Cryo-EM image processing

Details of processing pipelines and refinement statistics are shown in Extended Data Figs. 5, 6 and 8 and Table 1. All steps of image processing were performed using RELION-4 (refs. 49,50). Motion correction was performed using RELION’s implementation of the MotionCor2 program51 and the contrast transfer function (CTF) parameters of the micrographs were estimated using the CTFFIND program52.

Cryo-EM data for the condensin II–DNA complex were collected from three independent datasets comprising a total of 5,006 micrographs. Initial particle picking was performed using Gaussian picking on dataset 1 and particles belonging to the best three-dimensional (3D) classes were subsequently used to train Topaz53 for autopicking on the combined datasets. Approximately 1.2 million particles were extracted using a box size of 256 pixels and binned twofold, resulting in a calibrated pixel size of 3.2 Å per pixel. After iterative rounds of two-dimensional (2D) classification, particles corresponding to well-defined classes were subjected to two rounds of 3D classification, yielding a final set of 70,000 particles belonging to a structurally homogeneous class. A 3D initial reference procedure in RELION was used to generate the initial model required for 3D classification steps. The final particle set was reextracted using a box size of 256 pixels without binning and refined to generate a map at 7.3-Å resolution. Bayesian polishing and CTF refinement were then performed to improve particle alignment and optical parameter estimation. Specifically, two rounds of CTF refinement were used to correct for anisotropic magnification, beam tilt, per-particle defocus and astigmatism. Final 3D refinement yielded a reconstruction at an overall resolution of 6.8 Å, with local resolution reaching 5.8 Å.

For the condensin II complex bound to two ADP-BeF3 molecules, 8,294 micrographs were collected. Following an initial round of Gaussian picking, particle extraction and 2D classification, particles with the best signal-to-noise ratio were used for Topaz training and autopicking. A total of 734,000 particles were extracted using a 400-pixel box size and binned fourfold to a pixel size of 4.8 Å per pixel. Following one round of 2D classification, 567,000 particles were retained for downstream processing. Particles were subjected to 3D classification with image alignment, resulting in a subset of 133,000 particles that was refined to an initial reconstruction at 7.2-Å resolution. A 3D initial reference procedure in RELION was used to generate the initial model required for 3D classification steps. Further 3D classification without image alignment was performed to obtain a final set of 56,000 homogeneous particles. These particles were reextracted using a 480-pixel box size and rescaled to 360 pixels, corresponding to a pixel size of 1.6 Å per pixel. A first round of Bayesian polishing followed by 3D refinement produced a reconstruction at 6.9-Å resolution. To improve the density corresponding to the core and hinge modules, signal subtraction and focused refinement approaches were applied. The hinge and elbow domains did not yield interpretable density, likely because of strong preferred orientation and possible denaturation at the air–water interface. However, 2D classification revealed a small subset of approximately 2,300 particles displaying features resembling 2D projections of the hinge and elbow domains of condensin II in a zipped-up conformation. In contrast, focused refinement of the core complex after signal subtraction, followed by anisotropic magnification estimation, achieved a resolution of 5.6 Å. An additional round of Bayesian polishing, followed by per-particle defocus and astigmatism correction and beam tilt estimation, further improved the map resolution to 4.2 Å.

Cryo-EM datasets for the condensin II–M18BP1P complex were collected from two datasets comprising 12,273 and 22,728 micrographs, respectively. Following an initial round of Gaussian picking, particle extraction and 2D classification, particles with the best signal-to-noise ratio were used for Topaz training and autopicking on both datasets. The two datasets were processed in parallel, yielding 158,000 and 864,000 initial particles, which were extracted using a box size of 400 pixels at a calibrated pixel size of 0.935 Å per pixel. After subsequent rounds of 2D and 3D classification, particles from both datasets were combined, yielding a subset of 70,000 particles that was refined to an initial reconstruction at 6.2-Å resolution. A 60-Å filtered map obtained from a previously published structure of human condensin II bound to M18BP1 (PDB 9F5W) was used as a reference for initial 3D classification. Bayesian polishing followed by 3D classification and 3D refinement was used to obtain a map of the entire complex at 6.0-Å resolution. In parallel, signal subtraction focused on the HEAT-repeat-containing subunits was carried out, followed by 3D classification without alignment. A subset of 52,000 particles was selected for subsequent rounds of Bayesian polishing and CTF refinement, including anisotropic magnification estimation, beam tilt estimation and per-particle defocus and astigmatism correction, followed by 3D refinement to obtain a final density map with local resolution reaching 3.8 Å.

Model building

Initial models for the condensin II DNA-bound complex were obtained by rigid-body fitting of the individual subunit of the condensin II complex into the density map using ChimeraX54. As starting structures, the subunits isolated from the previously published structure of condensin II bound to M18BP1 (PDB 9F5W) were used, except for the N-terminal domain of CAP-H2 subunit, for which the AlphaFold3 prediction of the first 88 aa was used. The DNA molecule was generated in Coot55 using the sequence used for sample preparation.

Initial models for the condensin II complex bound to two ADP-BeF3 molecules was obtained first by running ModelAngelo56 in RELION. The resulting model provided the initial backbone tracing and sequence assignment for the well-resolved regions of the map. Missing residues and incompletely modeled regions were subsequently completed by rigid-body fitting of individual condensin II subunits into the density map using ChimeraX and Coot. AlphaFold3 predictions for the interactions between the C-terminal tail of CAP-D3 and CAP-G2 or CAP-H2 were generated and subsequently fitted into the corresponding cryo-EM densities.

Initial models for the condensin II–M18BP1P complex was initiated by rigid-body fitting of the previously solved structure of human condensin II bound to M18BP1 (PDB 9F5W) into the cryo-EM density map using ChimeraX. Additional residues of M18BP1P that were not present in the previously determined structure were modeled using AlphaFold3 predictions as reference models and subsequently fitted into the corresponding cryo-EM densities.

For all structures, self-restraints for the complete assembly were generated in Coot using a 4.3-Å cutoff and applied during iterative rounds of real-space refinement and manual curation in Coot. The refined model was then used for a final round of real-space refinement and validation in PHENIX57. All structural figures were prepared using ChimeraX and PyMOL.

MST

MST experiments were performed using fluorescently labeled CAP-G2 (10 nM) as the ligand and the SMC2/4 dimer as the titrant. CAP-G2 was labeled through the strep tag with an AlexaFluor647-labeled streptavidin. A 16-point 1:2 serial dilution of SMC2/4 was prepared starting from 20 µM in buffer containing 30 mM HEPES pH 7.5, 150 mM KCl, 0.5 mM TCEP and 0.05% (v/v) Tween-20. Samples were mixed 1:1 (v/v) with labeled CAP-G2, incubated for 10–15 min at room temperature and loaded into premium-coated capillaries. Measurements were performed using high MST power on a Monolith instrument (NanoTemper Technologies). To assess the effect of M18BP1, parallel experiments were carried out in the presence of 2 µM M18BP1P, M18BP17A or BSA (control). Data were analyzed using GraphPad Prism and binding curves were fitted with a one-site-specific binding model to determine dissociation constants.

Reporting summary

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

Online content

Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41594-026-01902-8.

Supplementary information

Reporting Summary (1.4MB, pdf)
Peer Review File (568.5KB, pdf)
Supplementary Video 1 (16.3MB, mov)

Single-molecule wide-field video of λ-DNA compaction in the presence of unlabeled condensin II (1 nM), AlexaFluor488-labeled M18BP1P (10 nM), ATP (3 mM) and SYTOX orange (1 nM).

Supplementary Video 2 (24.7MB, mov)

Morphing between the structure of condensin II in the autoinhibited conformation and the structure of condensin II bound to M18BP1P.

Source data

Source Data Fig. 1 (3.7MB, xlsx)

Lumicks force–distance curves.

Source Data Fig. 2 (2.2MB, xlsx)

Lumicks force–distance curves.

Source Data Fig. 4 (240.1KB, xlsx)

MST source data.

Source Data Extended Data Fig. 2 (642.1KB, xlsx)

Lumicks force–distance curves.

Source Data Extended Data Fig. 3 (2.3MB, xlsx)

Lumicks force–distance curves.

Source Data Extended Data Fig. 4 (4.6MB, xlsx)

Lumicks force–distance curves.

Source Data Extended Data Fig. 4 (103KB, pdf)

Uncropped gels.

Source Data Extended Data Fig. 9 (67KB, pdf)

Uncropped gels.

Acknowledgements

We thank the Cryo-EM and Biophysics Units of the National Facility for Structural Biology at Human Technopole for technical support and assistance.

Extended data

Author contributions

Conceptualization, A.B. and A.V. Methodology, A.B. and J.A. Investigation, A.B., S.C., J.A. and V.C. Visualization, A.B. Funding acquisition, A.V. Project administration, A.B. and A.V. Supervision, A.V. Writing—original draft, A.B. Writing—revised draft, A.B. and A.V. Writing—review and editing, A.B., S.C., J.A., V.C. and A.V.

Peer review

Peer review information

Nature Structural & Molecular Biology thanks the anonymous reviewers for their contribution to the peer review of this work. Peer reviewer reports are available. Primary Handling Editor: Sara Osman, in collaboration with the Nature Structural & Molecular Biology team.

Funding

This work was supported by the Human Technopole in coordination with the Italian Ministry of Economy and Finance, Cancer Research UK Program Foundation CR-UK C47547/A21536 (A.V.), Wellcome Trust Investigator Award 200818/Z/16/Z (A.V.) and European Molecular Biology Organization Long-Term Fellowship ALTF 552-2022 (A.B.).

Data availability

The cryo-EM map and atomic coordinates of the condensin II complex in autoinhibited conformation were deposited to the EM Data Bank and Protein Data Bank with accession codes EMD-56221 and PDB 9TSZ. The cryo-EM map and atomic coordinates of the condensin II–DNA complex were deposited to the EM Data Bank and Protein Data Bank with accession codes EMD-56247 and PDB 9TTP. The cryo-EM map and atomic coordinates of the condensin II–M18BP1P complex were deposited to the EM Data Bank and Protein Data Bank with accession codes EMD-56273 and PDB 9TUU. Raw Lumicks data are available from Zenodo (https://doi.org/10.5281/zenodo.20187436, LUMICKS_DATASET-Borsellini_et-al_2026_part1; https://doi.org/10.5281/zenodo.20187685, LUMICKS_DATASET_Borsellini_et-al_2026_part2)58,59. Correspondence and requests for materials should be addressed to A.V. (alessandro.vannini@fht.org). Source data are provided with this paper.

Code availability

Code for single-molecule data analysis is available from GitHub (https://github.com/aborsellini92/Borsellini-et-al-2026).

Competing interests

The authors declare no competing interests.

Footnotes

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

Contributor Information

Alessandro Borsellini, Email: Alessandro.Borsellini@fht.org.

Alessandro Vannini, Email: Alessandro.Vannini@fht.org.

Extended data

is available for this paper at https://doi.org/10.1038/s41594-026-01902-8.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41594-026-01902-8.

References

  • 1.Allan, J., Harborne, N., Rau, D. C. & Gould, H. Participation of core histone ‘tails’ in the stabilization of the chromatin solenoid. J. Cell Biol.93, 285–297 (1982). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Zhiteneva, A. et al. Mitotic post-translational modifications of histones promote chromatin compaction in vitro. Open Biol.7, 170076 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Kireeva, N., Lakonishok, M., Kireev, I., Hirano, T. & Belmont, A. S. Visualization of early chromosome condensation: a hierarchical folding, axial glue model of chromosome structure. J. Cell Biol.166, 775–785 (2004). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Samejima, K. et al. Mitotic chromosomes are compacted laterally by KIF4 and condensin and axially by topoisomerase IIα. J. Cell Biol.199, 755–770 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Maeshima, K. & Laemmli, U. K. A two-step scaffolding model for mitotic chromosome assembly. Dev. Cell4, 470–480 (2003). [DOI] [PubMed] [Google Scholar]
  • 6.Hirano, T., Kobayashi, R. & Hirano, M. Condensins, chromosome condensation protein complexes containing XCAP-C, XCAP-E and a Xenopus homolog of the Drosophila Barren protein. Cell89, 511–521 (1997). [DOI] [PubMed] [Google Scholar]
  • 7.Ono, T. et al. Differential contributions of condensin I and condensin II to mitotic chromosome architecture in vertebrate cells. Cell115, 109–121 (2003). [DOI] [PubMed] [Google Scholar]
  • 8.Ganji, M. et al. Real-time imaging of DNA loop extrusion by condensin. Science360, 102–105 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Gibcus, J. H. et al. A pathway for mitotic chromosome formation. Science359, eaao6135 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Kong, M. et al. Human condensin I and II drive extensive ATP-dependent compaction of nucleosome-bound DNA. Mol. Cell79, 99–114 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Kinoshita, K. et al. A loop extrusion-independent mechanism contributes to condensin I-mediated chromosome shaping. J. Cell Biol.221, e202109016 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Forte, G. et al. Bridging condensins mediate compaction of mitotic chromosomes. J. Cell Biol.223, e202209113 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Hirano, T. Condensins: universal organizers of chromosomes with diverse functions. Genes Dev.26, 1659–1678 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Ono, T., Fang, Y., Spector, D. L. & Hirano, T. Spatial and temporal regulation of condensins I and II in mitotic chromosome assembly in human cells. Mol. Biol. Cell15, 3296–3308 (2004). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Walther, N. et al. A quantitative map of human condensins provides new insights into mitotic chromosome architecture. J. Cell Biol.217, 2309–2328 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Hirota, T., Gerlich, D., Koch, B., Ellenberg, J. & Peters, J. M. Distinct functions of condensin I and II in mitotic chromosome assembly. J. Cell Sci.117, 6435–6445 (2004). [DOI] [PubMed] [Google Scholar]
  • 17.Yamashita, D. et al. MCPH1 regulates chromosome condensation and shaping as a composite modulator of condensin II. J. Cell Biol.194, 841–854 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Borsellini, A. et al. Condensin II activation by M18BP1. Mol. Cell85, 2688–2700 (2025). [DOI] [PubMed] [Google Scholar]
  • 19.Wenda, J. M., Prosee, R. F., Gabus, C. & Steiner, F. A. Mitotic chromosome condensation requires phosphorylation of the centromeric protein KNL-2 in C. elegans. J. Cell Sci.134, jcs259088 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Yoshida, M. M., Kinoshita, K., Shintomi, K., Aizawa, Y. & Hirano, T. Regulation of condensin II by self-suppression and release mechanisms. Mol. Biol. Cell35, ar21 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Gross, P., Farge, G., Peterman, E. J. & Wuite, G. J. Combining optical tweezers, single-molecule fluorescence microscopy, and microfluidics for studies of DNA-protein interactions. Methods Enzymol.475, 427–453 (2010). [DOI] [PubMed] [Google Scholar]
  • 22.Candelli, A., Wuite, G. J. & Peterman, E. J. Combining optical trapping, fluorescence microscopy and micro-fluidics for single molecule studies of DNA–protein interactions. Phys. Chem. Chem. Phys.13, 7263–7272 (2011). [DOI] [PubMed] [Google Scholar]
  • 23.Gorman, J. & Greene, E. C. Visualizing one-dimensional diffusion of proteins along DNA. Nat. Struct. Mol. Biol.15, 768–774 (2008). [DOI] [PubMed] [Google Scholar]
  • 24.Syrjänen, J. L. et al. Single-molecule observation of DNA compaction by meiotic protein SYCP3. eLife6, e22582 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Gutierrez-Escribano, P. et al. A conserved ATP- and Scc2/4-dependent activity for cohesin in tethering DNA molecules. Sci. Adv.5, eaay6804 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Tsubota, Y. et al. Functional interplay between condensin I and topoisomerase IIα in single-molecule DNA compaction. Nat. Commun.16, 7239 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Eeftens, J. M. et al. Real-time detection of condensin-driven DNA compaction reveals a multistep binding mechanism. EMBO J.36, 3448–3457 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Abe, S. et al. The initial phase of chromosome condensation requires Cdk1-mediated phosphorylation of the CAP-D3 subunit of condensin II. Genes Dev.25, 863–874 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Lee, B. G., Rhodes, J. & Lowe, J. Clamping of DNA shuts the condensin neck gate. Proc. Natl Acad. Sci. USA119, e2120006119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Shaltiel, I. A. et al. A hold-and-feed mechanism drives directional DNA loop extrusion by condensin. Science376, 1087–1094 (2022). [DOI] [PubMed] [Google Scholar]
  • 31.Hassler, M. et al. Structural basis of an asymmetric condensin ATPase cycle. Mol. Cell74, 1175–1188 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Burmann, F., Funke, L. F. H., Chin, J. W. & Lowe, J. Cryo-EM structure of MukBEF reveals DNA loop entrapment at chromosomal unloading sites. Mol. Cell81, 4891–4906 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Higashi, T. L. et al. A structure-based mechanism for DNA entry into the cohesin ring. Mol. Cell79, 917–933 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Yu, Y. et al. Cryo-EM structure of DNA-bound Smc5/6 reveals DNA clamping enabled by multi-subunit conformational changes. Proc. Natl Acad. Sci. USA119, e2202799119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature630, 493–500 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Lee, B. G. et al. Cryo-EM structures of holo condensin reveal a subunit flip-flop mechanism. Nat. Struct. Mol. Biol.27, 743–751 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Cutts, E. E., Tetiker, D., Kim, E. & Aragon, L. Molecular mechanism of condensin I activation by KIF4A. EMBO J.44, 4 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Li, Y. et al. The structural basis for cohesin–CTCF-anchored loops. Nature578, 472–476 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Garcia-Nieto, A. et al. Structural basis of centromeric cohesion protection. Nat. Struct. Mol. Biol.30, 853–859 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Yan, L. et al. A non-canonical role of the inner kinetochore in regulating sister-chromatid cohesion at centromeres. EMBO J.43, 2424–2452 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Shintomi, K. & Hirano, T. Releasing cohesin from chromosome arms in early mitosis: opposing actions of Wapl–Pds5 and Sgo1. Genes Dev.23, 2224–2236 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Davidson, I. F. et al. DNA loop extrusion by human cohesin. Science366, 1338–1345 (2019). [DOI] [PubMed] [Google Scholar]
  • 43.Kim, Y., Shi, Z., Zhang, H., Finkelstein, I. J. & Yu, H. Human cohesin compacts DNA by loop extrusion. Science366, 1345–1349 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Pradhan, B. et al. The Smc5/6 complex is a DNA loop-extruding motor. Nature616, 843–848 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Hirano, T. Mitotic genome folding. J. Cell Biol.224, e202504075 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Houlard, M. et al. MCPH1 inhibits condensin II during interphase by regulating its SMC2–Kleisin interface. eLife10, e73348 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Weissmann, F. et al. biGBac enables rapid gene assembly for the expression of large multisubunit protein complexes. Proc. Natl Acad. Sci. USA113, E2564–E2569 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Huis In ‘t Veld, P. J. et al. Reconstitution and use of highly active human CDK1:cyclin-B:CKS1 complexes. Protein Sci.31, 528–537 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Zivanov, J. et al. New tools for automated high-resolution cryo-EM structure determination in RELION-3. eLife7, e42166 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Kimanius, D., Dong, L., Sharov, G., Nakane, T. & Scheres, S. H. W. New tools for automated cryo-EM single-particle analysis in RELION-4.0. Biochem. J.478, 4169–4185 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Zheng, S. Q. et al. MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy. Nat. Methods14, 331–332 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Rohou, A. & Grigorieff, N. CTFFIND4: fast and accurate defocus estimation from electron micrographs. J. Struct. Biol.192, 216–221 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Bepler, T. et al. Positive-unlabeled convolutional neural networks for particle picking in cryo-electron micrographs. Nat. Methods16, 1153–1160 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Pettersen, E. F. et al. UCSF ChimeraX: structure visualization for researchers, educators, and developers. Protein Sci.30, 70–82 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Casanal, A., Lohkamp, B. & Emsley, P. Current developments in Coot for macromolecular model building of electron cryo-microscopy and crystallographic data. Protein Sci.29, 1069–1078 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Jamali, K. et al. Automated model building and protein identification in cryo-EM maps. Nature628, 450–457 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Afonine, P. V. et al. Real-space refinement in PHENIX for cryo-EM and crystallography. Acta Crystallogr. D74, 531–544 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 58.Borsellini, A., Vannini, A., Chamera, S., Cecatiello, V. & Andrecka, J. LUMICKS_and_MST_DATASET_Borsellini_etal_2026_part1. Zenodo 10.5281/zenodo.20187436 (2026). [DOI]
  • 59.Borsellini, A., Vannini, A., Chamera, S., Andrecka, J. & Cecatiello, V. LUMICKS_DATASET_Borsellini_etal_2026_part2. Zenodo 10.5281/zenodo.20187685 (2026). [DOI]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Reporting Summary (1.4MB, pdf)
Peer Review File (568.5KB, pdf)
Supplementary Video 1 (16.3MB, mov)

Single-molecule wide-field video of λ-DNA compaction in the presence of unlabeled condensin II (1 nM), AlexaFluor488-labeled M18BP1P (10 nM), ATP (3 mM) and SYTOX orange (1 nM).

Supplementary Video 2 (24.7MB, mov)

Morphing between the structure of condensin II in the autoinhibited conformation and the structure of condensin II bound to M18BP1P.

Source Data Fig. 1 (3.7MB, xlsx)

Lumicks force–distance curves.

Source Data Fig. 2 (2.2MB, xlsx)

Lumicks force–distance curves.

Source Data Fig. 4 (240.1KB, xlsx)

MST source data.

Source Data Extended Data Fig. 2 (642.1KB, xlsx)

Lumicks force–distance curves.

Source Data Extended Data Fig. 3 (2.3MB, xlsx)

Lumicks force–distance curves.

Source Data Extended Data Fig. 4 (4.6MB, xlsx)

Lumicks force–distance curves.

Source Data Extended Data Fig. 4 (103KB, pdf)

Uncropped gels.

Source Data Extended Data Fig. 9 (67KB, pdf)

Uncropped gels.

Data Availability Statement

The cryo-EM map and atomic coordinates of the condensin II complex in autoinhibited conformation were deposited to the EM Data Bank and Protein Data Bank with accession codes EMD-56221 and PDB 9TSZ. The cryo-EM map and atomic coordinates of the condensin II–DNA complex were deposited to the EM Data Bank and Protein Data Bank with accession codes EMD-56247 and PDB 9TTP. The cryo-EM map and atomic coordinates of the condensin II–M18BP1P complex were deposited to the EM Data Bank and Protein Data Bank with accession codes EMD-56273 and PDB 9TUU. Raw Lumicks data are available from Zenodo (https://doi.org/10.5281/zenodo.20187436, LUMICKS_DATASET-Borsellini_et-al_2026_part1; https://doi.org/10.5281/zenodo.20187685, LUMICKS_DATASET_Borsellini_et-al_2026_part2)58,59. Correspondence and requests for materials should be addressed to A.V. (alessandro.vannini@fht.org). Source data are provided with this paper.

Code for single-molecule data analysis is available from GitHub (https://github.com/aborsellini92/Borsellini-et-al-2026).


Articles from Nature Structural & Molecular Biology are provided here courtesy of Nature Publishing Group

RESOURCES