Condensin compacts DNA in 15 nm steps and requires FACT for extrusion on chromatin

doi:10.21203/rs.3.rs-901482/v1

PDF

Biological Sciences - Article

Condensin compacts DNA in 15 nm steps and requires FACT for extrusion on chromatin

https://doi.org/10.21203/rs.3.rs-901482/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted 05 Oct, 2021

You are reading this latest preprint version

Condensin plays a central role in the organisation of chromosomes by compacting chromatin into loops during mitosis. Condensin achieves this through a loop extrusion mechanism that remains poorly understood. To identify the molecular steps of yeast condensin during loop formation, we used optical tweezers with fluorescence detection. We find that single yeast condensin complexes use ATP to extrude DNA through distinct 15 nm steps, thus advancing ~45 base pairs (bp) per step. Under increasing load, the condensin step size remains constant while step-dwell times increase, and stalls at forces >1 pN. We also show that nucleosome arrays hinder processive condensin extrusion and demonstrate that the histone chaperone FACT is required for compaction of nucleosomal arrays by condensin. Importantly, FACT-assisted compaction on nucleosomes also occurs through distinct 15 nm steps. Finally, we show that FACT is required for correct condensin localisation in vivo. Our results establish that loop extrusion by yeast condensin involves a 45 bp stroke that requires FACT for condensin function on chromatin.

General Biochemistry

Biophysics

yeast condensin

loop formation

loop extrusion

Structural Maintenance of Chromosomes (SMC) complexes are ubiquitous proteins involved in the physical manipulation of DNA during many nuclear processes¹. There are three distinct types of SMC complexes in eukaryotic cells, cohesin, condensin and Smc5/6. These complexes dynamically control the architecture of chromatin and chromosomes during the cell cycle and are involved in diverse processes such as sister chromatid cohesion, chromosome compaction and segregation, long-range transcriptional regulation and DNA repair^1-5. Despite intense research, SMC complexes have eluded mechanistic understanding and are a class of DNA binding proteins for which the core mechanism of action remains uncharacterised. SMC complexes contain a heterodimer of SMC proteins. The SMC core heterodimeric subunits form a large, 50 nm diameter ring which topologically traps DNA and is transiently sealed by ATP-hydrolysing head domains. Depending on the specific SMC and its function, the core heterodimer associates with different regulatory subunits, as well as different loading and releasing factors⁶. In yeast, a single condensin complex exists, which associates with a kleisin subunit, Brn1, and two HEAT-repeat proteins, Ycs4 and Ycg1, forming a pentameric holocomplex⁷. In contrast, higher eukaryotes, including humans, have two condensins, condensin I (CI) and II (CII)^8,9.

In vivo studies using chromosome conformation capture techniques have suggested a role for SMC complexes in genome and chromosome organisation that could originate from an intrinsic ability to form or maintain loops in cis^10-12. Such observations have led to the proposal that SMC complexes might harbour a “loop extrusion” activity where DNA can be reeled in through the SMC structure to generate loops¹³. Loop extrusion was first confirmed in vitro using purified yeast condensin, as well as, purified human CI/CII and human cohesin complexes^14-16.

Loop extrusion by the yeast condensin holocomplex has been demonstrated using λ-DNA molecules tethered at both ends to glass surfaces and imaged by total internal reflection fluorescence (TIRF) microscopy¹⁴. In this assay, extrusion of DNA requires ATP and proceeds at high speeds, on average 200 nm/s, or 600 bp of DNA¹⁴. This is in agreement with magnetic tweezers (MT) studies measuring condensin‐mediated DNA compaction in step sizes of ~200 nm¹⁷. However, this high velocity is surprising considering the relatively low ATP turnover (~2 ATP molecules per second)¹⁴. In addition, current compaction models are largely based on the ability of condensin to undergo large-scale conformational changes of its extended architecture during the ATP hydrolysis cycle¹⁸. Therefore, with a complex length of ~50 nm, a maximum compaction (or extrusion) velocity of 50 nm/s (~147 bp) is expected.

The activity of SMC complexes in vivo during genome folding is further complicated by the fact that DNA is chromatinised inside cells. Loop extrusion by human CI and CII and human cohesin complexes is not affected by the presence of nucleosomes^15,16, however, the diffusion of yeast cohesin on DNA curtains is impeded by nucleosomal particles¹⁹. Whether DNA compaction by yeast condensin is impeded by nucleosomes has not yet been thoroughly investigated.

Cohesin roles on chromatin are facilitated by histone chaperones and remodellers, including FACT (Facilitates chromatin transcription)²⁰. In addition, FACT is one of the minimal core factors, alongside condensin, topoisomerase II and histones necessary to reconstitute chromosomes in vitro²¹, suggesting that FACT might be important to facilitate nucleosome remodelling as chromosomes are compacted through the coordinated action of condensin and topoisomerase II.

Here, we have used single‐molecule optical tweezers with fluorescence detection to investigate the activity of single yeast condensin complexes during their compaction of DNA molecules and chromatin arrays. We observe that condensin performs loop extrusion through distinct 15 nm steps, or 45 bp at a time. Step-sizes are not affected by force, but increasing force results in longer step dwell times, thus decreasing the compaction velocity. We also demonstrate that condensin requires the collaborative activity of the histone chaperone, FACT, to overcome nucleosome arrays and demonstrate that this is important for condensin localisation on chromatin in vivo.

Yeast condensin binds DNA in an ATP-independent manner

To investigate DNA compaction by condensin, we used our previously developed optical tweezers assay²² with purified condensin (Suppl. Fig. 1A). Monodisperse pentameric composition at nM concentrations was confirmed by mass photometry²³ (Suppl. Fig. 1B). Briefly, biotinylated l-DNA molecules (48.5 kb) are caught between two streptavidin-coated polystyrene beads trapped by laser beams that enable their spatial manipulation at defined forces (Fig. 1A). Numerous physical parameters, such as force and distance between beads, can be measured directly. Experiments are performed in a five-channel laminar flow cell where a dumbbell comprised of the beads and captured DNAs can be moved between channels containing different buffers and protein complexes (Fig. 1A). Typically, Channel 1 contains the streptavidin-coated polystyrene beads, which are caught by the two laser beams, and Channel 2 contains the biotinylated DNA (Fig. 1A). Single DNA molecules between the beads and validated by measuring a force-extension (FE) curve that can be fit to the Worm-Like Chain model (Supp. Fig. 2). Next, the captured DNA molecule is transferred to Channels 3-5 to carry out desired experiments.

Previous magnetic tweezers studies have shown that purified yeast condensin can bind DNA in the absence of ATP but its compaction activity is strictly ATP dependent¹⁷. These studies showed that incubation of condensin and DNA followed by removal of any unbound protein, led to subsequent compaction when ATP was added¹⁷. To directly visualize condensin binding to DNA, we labelled the complex with a single fluorophore (ATTO647N) at the N-terminus of the kleisin subunit Brn1 (see materials and methods for details). Conjugation of this fluorophore does not affect the complex activity in vitro¹⁴. To test whether condensin molecules bind DNA in the absence of ATP, we flowed Channel 3 with 125 mM NaCl buffer and Channel 4 with 50 mM NaCl buffer containing 1 nM ATTO647N-labelled condensin holocomplex. We captured a λ-DNA molecule and incubated it in Channel 4 for 30 s (in the presence of condensin), before moving it to the condensin-free Channel 3 for imaging (Fig. 1B, n=25). We observed that condensin complexes were indeed bound between the beads, consistent with previous MT experiments¹⁷. As expected, condensin also remained bound to the DNA in Channel 3 following incubation in Channel 4 with 1 mM ATP (Fig. 1B, n = 25).

Single condensin complexes extrude DNA loops in an ATP-dependent manner

In magnetic tweezers, purified condensin has been shown to compact DNA molecules in the presence of ATP^17,24. Optical tweezers with confocal detection in laminar flow microfluidic cells facilitate rapid moving trapped DNA molecules between different experimental conditions, enabling us to simultaneously measure and visualise compaction activity by single condensin holocomplexes. Previously, we have shown that unlabelled condensin compacts λ-DNA molecules against 1 pN forces²². Here, we sought to further investigate DNA compaction using our ATTO647N-labelled condensin. First, we confirmed that condensin-dependent compaction requires ATP. To this aim, we flowed 1 nM labelled complex in buffer containing 50 mM NaCl without and with 1 mM ATP (Channels 4 and 5, respectively, Fig. 1C). We incubated a captured DNA molecule in Channel 4 (no ATP and no condensin) whilst applying a constant 1pN force between the beads. In the absence of ATP, the bead distance remains constant, indicating a stable force clamp (Fig. 1C, n = 25). In the presence of 1 mM ATP and 1 nM condensin complex (Channel 5), the bead distance decreases rapidly (Fig. 1D, n = 38), confirming that condensin requires ATP to compact DNA, as reported^17,22,24. Under these conditions, the average compaction rate is 140 ± 4 nm/s (Fig. 1F n = 23), or 410 ± 10 bp/s, consistent with the magnetic tweezers data¹⁷.

However, as in the MT experiments, it is impossible to ascertain whether compaction is mediated by a single or multiple holocomplexes under these conditions. To overcome this limitation, we first captured a single holocomplex by incubating with 1 nM condensin complex in 125 mM NaCl with 1 mM ATP for 30 s (Channel 4), and then moved the dumbbell to Channel 3 with condensin-free buffer and 1 mM ATP. Under these conditions, the bead distance decreased more slowly (Fig. 1E, n = 40) with an average speed of 7.5 ± 0.4 nm/s (Fig. 1G; n = 25), equivalent to 22 ± 1 bp/s. Importantly, we verified the presence of intramolecular DNA loops by re-extending the DNA with force and observing the expected sawtooth features in the FE curve (Suppl. Fig. 2)²².

These results raise the interesting possibility that the rapid compaction in the protein channel (Fig. 1D) is due to the combined work of multiple loaded complexes, whereas the slower compaction observed in the condensin-free channel (Fig. 1E) is due to the activity of a single condensin holocomplex. To confirm this, we imaged DNA compaction by a single condensin in real time. We first incubated a trapped DNA molecule for 30 s with 1nM labelled condensin (Channel 4), then moved the DNA to Channel 3 (125 mM NaCl, 1 mM ATP), force clamped at 1 pN and imaged the dumbbell (Suppl. Video 1 and Fig. 2A). As compaction proceeds, a single condensin molecule is seen between the two beads, and the distance between condensin and one of the beads (D_R) decreases, while the distance to the second bead (D_L) remains constant (Suppl. Video 1 and Fig. 2A-C, n = 12). Compaction stops when the mobile bead reaches the DNA-bound condensin. Quantifying the distance between the condensin and two beads (D_R and D_L, Fig. 2B-C) confirms asymmetric compaction activity, as previously observed by TIRF microscopy¹⁴. The resulting compaction rate (7.9 nm/s) is consistent with our prior measurements (Fig. 1E), demonstrating that, under these experimental conditions, we measure the activity of individual condensin complexes.

DNA compaction by condensin is predicted to occur by loop extrusion whereby condensin anchors onto DNA and reels it in from one side forming a progressively larger loop¹⁴. To visualize the DNA during condensin-mediated compaction in our set-up, we used SYTOX Orange. First, we captured and incubated a λ-DNA molecule with 1 nM labelled condensin complex and 1 mM ATP for 30 s (Channel 4). Then, we moved the DNA to Channel 5, containing 50 nM SYTOX Orange and 1 mM ATP, where compaction could be directly imaged under 1 pN force clamp (Suppl. Video 2 and Fig. 2D, condensin in red and DNA in green). SYTOX fluorescence intensity increases in the vicinity of the condensin molecule, consistent with accumulation of DNA. To image the accumulated DNA, we applied perpendicular flow, which extends and reveals the extruded DNA loop with the condensin complex localised at the base (Fig. 2E). These results confirm that condensin-mediated compaction in our force clamp experiments occurs by asymmetric loop extrusion. Moreover, we also confirmed this loop extrusion activity in our home-built single-molecule TIRF set up²⁵ (Suppl. Fig. 3), as previously described¹⁴.

Condensin-dependent loop extrusion occurs in discrete 15 nm steps

Condensin loop extrusion on doubly-tethered λ-DNA molecules has been estimated to occur at an average velocity of 600 bp/s¹⁴, involving distinct 200 nm steps¹⁷. Our compaction trajectories under single molecule conditions (Fig. 1E) reveal a much lower compaction rate, of 22 bp/s (Fig. 1G). To identify individual compaction steps, we improved the experimental resolution by using smaller beads (1.7 mm) and a shorter DNA substrate (8.4 kb), which reduce the viscous drag and the dumbbell’s Brownian motion. Calibration experiments demonstrate our ability to detect steps with ≤5 nm resolution at 1 pN (Suppl. Fig. 4). We first incubated a tethered 8 kb DNA molecule with 1 nM ATTO647N-labelled condensin and 1 mM ATP (Channel 4) for 30 s and then force clamped at 1 pN in buffer containing 1 mM ATP (Channel 3). The resulting compaction trajectories reveal distinct compaction steps (Fig. 3A; Suppl. Fig. 5A, n = 35). The majority of steps (97%, n = 65) were downward (i.e. compaction), with rare backward steps (Fig. 3A). We fit the compaction trajectories with a step finding algorithm to determine the step sizes and the dwell time between step (see Methods). The resulting narrow step-size distribution yields an average step size of 14.8 ± 0.2 nm (Fig. 3B; Suppl. Fig. 5A; n = 70) and an average dwell time of 1.8 ± 0.1 s between steps (Fig. 3C; Suppl. Fig. 5A). This result demonstrates that condensin takes steps of ~45 bp on DNA under 1 pN tension. Condensin was unable to compact DNA at ≥1.5 pN forces (not shown) suggesting that condensin-induced loop extrusion stalls between 1 and 1.5 pN.

Previous estimates of condensin steps were measured using magnetic tweezers at lower forces (0.3-0.5 pN)¹⁷. To determine the effect of force on the step-size, we collected compaction trajectories at 0.5 pN (n_trajectories = 34) and 0.3 pN (n_trajectories = 32) forces (Fig. 3a; Suppl. Fig. 5b-c). Analysis of compaction traces at these lower forces shows that the average step size remains constant at 14.8 ± 0.3 nm (n_steps = 126) and 14.8 ± 0.4 nm (n_steps = 93), respectively (Fig. 3B; Suppl. Fig. 5B-C). In contrast, the average stepping dwell time increased linearly with increasing force (1.1 ± 0.1 s at 0.3 pN and 1.3 ± 0.1 s at 0.5 pN, Fig. 3C; Suppl. Fig. 5C). We approximate the stepping rates as the inverse of the average stepping dwell time, and fit their force dependence to the Bell-Evans equation (Fig. 3D)²⁶ to obtain the zero-force stepping rate, k(0) = 1.23 ± 0.05 s^-1, (Fig. 3D) and the distance to the transition state, Dx = 3.3 ± 0.3 nm. Interestingly, comparison of the zero-force stepping rate to the bulk ATP hydrolysis rates (~2-3 ATPs s^-1) ²⁷^-²⁹ suggests that each step likely involves the hydrolysis of two ATP molecules. The distance to the transition state indicates that the rate limiting step for condensin translocation involves an initial conformational change ~11 bp long.

We also sought to image single condensin complexes compacting by loop extrusion on the 8 kb DNA template (Fig. 3E). As for λ-DNA (Fig. 2A), a single condensin holocomplex extrudes DNA loops asymmetrically (Fig. 3E-G; Suppl. Video 3) at 0.5 pN by distinct 14.9 ± 0.4 nm steps (Fig. 3H; n = 49) and 1.1 ± 0.03 s dwell times (Fig. 3I, n = 49), confirming our prior observations (vide supra).

Nucleosomes hinder condensin compaction

In cells condensin work on chromatin, therefore, condensin must be able to processively loop nucleosome-bound DNA in its physiological setting. Previous studies on DNA curtains have shown that human CI and CII and human cohesin complexes can compact λ-DNA molecules containing a few nucleosomes^15,16. Here, we sought to investigate whether the presence of nucleosomes impacts the step size or ability of yeast condensin to extrude DNA loops. We expressed and purified recombinant yeast histone octamers with Alexa532-labeled H2A (Suppl. Fig. 1) and optimised incubation conditions to reliably incorporate one nucleosome along λ-DNA molecules (Suppl. Materials). Nucleosome binding was confirmed by imaging bound nucleosomal particles (Fig. 4A), and by monitoring the characteristic decrease in distance, 26 ± 3 nm, between the beads during nucleosome assembly (Fig. 4B), in addition to extending DNA molecules to destabilize the single bound histone octamers (Suppl. Fig. 6). The measured step size confirms the wrapping of DNA around the histones to form nucleosome as previously reported³⁰.

We carried out nucleosome assembly in Channel 4 and following visual confirmation that a single nucleosome particle had been loaded, the template was moved to Channel 5 and incubated with 1 nM ATTO647N-labelled condensin and 1 mM ATP for 30 s. Following this brief incubation, we moved the template to Channel 3 with 125 mM NaCl and 1 mM ATP containing buffer, and force clamped at 1 pN while simultaneously imaging condensin and nucleosomes during compaction (Fig. 4C; Suppl. video 4). As condensin loop extrusion proceeded, we observed the nucleosome particle approaching condensin (Fig. 4C; Suppl. video 4). Condensin was unable to bypass the nucleosome and compaction stalled (Fig. 4E; Suppl. video 4). This result was intriguing in the light of previous studies that have shown that nucleosomes do not prevent compaction by SMC complexes^15,16. Since 1 pN is on the upper range of stretching forces tolerated by condensin, we repeated the experiment clamping at 0.5 pN to ensure that nucleosome stalling was not due to a non-permissive applied force (Fig. 4D-E). We recorded images before and after compaction (Fig. D-E), as well as compaction traces for individual experiments (Fig. 4F). In 80% of measured cases (n = 24), condensin was found to stall at the nucleosome (Fig. 4D, F). In a minority of cases, 20% (Fig. 4F), condensin fully bypassed the nucleosome (Fig. 4E, n = 6). In such cases, as condensin arrived at the nucleosome, the nucleosome signal travelled downwards in relation to condensin (Fig. 4F), into an equivalent area to where the extruded loop was previously observed (Fig. 2C).

These results demonstrate that condensin only rarely bypasses individual nucleosomes, and therefore in the presence of nucleosome arrays, its physiological template, it would be reasonable to assume that condensin would not be able to extrude loops with the processivity required. To test this experimentally, we assembled a nucleosome array by maximally compacting the DNA in the presence of histones (Fig. 5A). The presence of nucleosome arrays was confirmed by imaging the nucleosomes on the template (Fig. 5A-insert). As expected, incubating these arrays with 1 nM of labelled condensin and 1 mM ATP, we observed no change in the distance between beads, indicating that condensin is not able to promote further compaction of chromatinised templates (Fig. 5B). Importantly, we observed condensin binding to the nucleosome array (Fig. 5B), demonstrating that the presence of tightly packed nucleosomes only inhibits loop extrusion processivity, not condensin access to the DNA.

FACT is required for condensin to compact nucleosome arrays

Our results demonstrate that nucleosome arrays prevent loop extrusion by yeast condensin, implicating that condensin extrusion in vivo likely requires additional factors. We have previously demonstrated that, in yeast cells, loop extrusion by the related SMC complex cohesin, is facilitated by the conserved histone chaperone FACT²⁰. Moreover, chromosome reconstitution from purified components has indicated that FACT is essential for chromosome assembly since condensin, topoisomerase II and FACT are all required when nucleosomes are included in chromosome assembly assays²¹. Therefore, FACT is a good candidate to facilitate loop extrusion of condensin through the nucleosomes and so we investigated whether addition of purified FACT allowed condensin to overcome nucleosomes bound to our DNA templates.

To test this hypothesis, we expressed and purified yeast FACT (Pob3 and Spt16) (Suppl. Fig. 1) to conduct our nucleosome array experiments in the presence of FACT. After confirming that FACT does not destabilise nucleosomes assembled on DNA templates (Suppl. Fig. 7), we assembled nucleosome arrays on our 8 kb DNA templates in Channel 4 by maximal compaction of the template before moving them to Channel 5 (containing 1 nM labelled condensin and 1 nM FACT (Spt16/Pob3) in 125 mM NaCl and 1 mM ATP-containing buffer). After 30 s of incubation, we moved the dumbbell to the condensin-free Channel 3 (containing 125 mM NaCl and 1 mM ATP buffer) and force-clamped at 1 pN. The nucleosome arrays were further compacted in the presence of FACT and condensin (Fig. 5C). Analysis of the compaction trajectories (n_trajectories = 155) reveals the presence of discrete 14.8 ± 0.4 nm steps (Fig. 5D, n_steps = 155) with a 1.5 ± 0.1 s average dwell time (Fig. 5E, n_dwells = 154), in remarkable agreement with the step sizes observed for naked DNA templates (Fig. 3B-C), demonstrating that FACT allows processive loop extrusion by condensin in the presence of nucleosome arrays. Interestingly, we also detected a small number of 12 ± 3 nm back steps (Fig. 5D), showing that occasionally condensin does not fully overcome nucleosome obstacles and backtracks on the template.

To directly visualize FACT-assisted condensin bypass on single nucleosomes in real-time, we simultaneously imaged condensin and a single nucleosome during compaction (Fig. 5G and Suppl. Video 5). The data clearly show condensin fully bypassing the nucleosome particles in the presence of FACT: As condensin reaches the nucleosome, the nucleosome signal is maintained through the bypass, and the nucleosome translocates downwards into the extruded loop area. FACT enables condensin bypassing single nucleosomes in 100% of the trajectories observed (Fig. 5H, n = 13), thereby demonstrating that FACT facilitates condensin extrusion through nucleosomes.

FACT facilitates condensin function at yeast centromeric regions

To further explore a possible functional relationship between FACT and condensin in vivo, we investigated the activities of both complexes at centromeric regions of yeast chromosomes. Condensin localises to pericentromeric chromatin post-replication³¹. FACT interacts with cohesin and is required for its binding at centromeres and discrete sites on chromosome arms²⁰. Therefore, FACT is enriched at pericentromeric regions where it facilitates SMC complex function²⁰. In turn, we sought to test whether inactivation of FACT caused any changes in condensin localisation to pericentromeric regions. We used an auxin (IAA)-inducible degron of Spt16 (SPT16-AID)²⁰ to deplete FACT from cells arrested in metaphase. We then performed calibrated ChIP-seq analysis for the condensin subunit Smc2. In the presence of Spt16, condensin enrichment at centromeric regions was detected (Fig. 6A), consistent with previous studies³¹. In contrast, when we degraded Spt16 the accumulation of condensin around pericentric regions was decreased (Fig. 6A). We found that binding at the core CEN sequence was not as affected at pericentric regions (Fig. 6A) suggesting that condensin is able to load at the CEN sequences, but has difficulties to move, or loop extrude, away from the core CEN sites. Moreover, we detected interaction between condensin and FACT in metaphase arrests using cells that expressed SMC2-6HA and SPT16-9MYC (Suppl. Fig. 9B) confirming that condensin physically interacts with this histone chaperone in yeast cells. These results are consistent with our in vitro data and demonstrate that FACT facilitates condensin function on yeast chromosomes.

It has been established that condensin compacts mitotic chromosomes using its ability to extrude loops on chromatin¹⁴. The exact mechanisms used to produce such loops, however, remains poorly understood, as well as whether condensin requires additional factors to operate on chromatinised templates. In this study, we have identified the size of the steps taken by individual condensin complexes during loop extrusion using high resolution optical tweezers. We have demonstrated that condensin complexes extrude loops by advancing ~45 bp at a time along the DNA contour. We have also shown that increasing tension on the DNA stalls condensin by increasing the dwell time between steps, but its 15 nm (or 45 bp) step size remained constant under all conditions tested, indicating that this is the basic step size for yeast condensin. As previously demonstrated¹⁴, condensin loop extrusion proceeds asymmetrically on the template. The presence of DNA-bound nucleosomes acts as a barrier, and prevented condensin from further reeling in DNA, and instead complexes stalled. We found that the histone chaperone FACT allows condensin to bypass bound nucleosomes without disassembly of the nucleosomal particle. Finally, we demonstrate that FACT facilitates condensin-dependent compaction of DNA containing nucleosome arrays and provide in vivo evidence demonstrating that FACT is important for condensin function at yeast centromeres.

Condensin takes defined steps compacting 15 nm at a time

Recent structures of the yeast condensin holocomplex in different functional states suggest that loop extrusion activity might be mediated by large conformational changes driven by ATP binding and hydrolysis¹⁸. Importantly, extended condensin holocomplexes are ~50 nm long, and therefore, it is reasonable to assume that step sizes cannot be larger than this distance. Previous studies using MT have provided estimates for condensin’s step size at ~80 nm²⁴ and 200 nm¹⁷. However, it was unclear in these studies whether the measurements corresponded truly to steps by single condensin complexes. Most MT compaction assays are typically performed in the presence of unbound complexes in solution, and do not allow for simultaneous imaging of the template-bound complexes, therefore, the observed step sizes could arise from multiple condensing complexes acting on the DNA template. Our setup allows compaction assays to be performed while simultaneously imaging individual condensins on the DNA by rapidly moving between different reagent-containing channels i.e. with and without condensin in solution. Our data shows that condensin takes defined steps, with a narrow distribution of step sizes at 15 nm (Fig. 3B, Suppl. Fig. 5). Another expectation for a mechanistically defined step is that it should not be dependent on small changes in the force stretched DNA. Indeed, we observed very similar step size distributions when applying forces ranging from 0.3 to 1 pN (Fig. 3B).

Our loop compaction/extrusion velocities under conditions where a single condensin is bound (Fig. 3) were 15 – 17 nm/s (or 45 – 51 bp/s), a value significantly lower than the 600-1500 bp/s previously reported¹⁴. However, given that condensin hydrolyses ~2 ATP molecules/s and that the fully extended complex is only 50 nm long, it is difficult to conceive a stepping mechanism that could produce the speeds reported for loop extrusion assays¹⁴. In contrast to loop extrusion, yeast condensin has been previously shown to translocate at ~60 bp/s²⁷ in line with the values obtained here. Since loop extrusion velocities are calculated under conditions when perpendicular flow is applied to the DNA, it is possible that the flow causes some slippage yielding faster apparent velocities.

Presently it seems speculative to propose details of the exact conformational changes that generate the 15 nm steps we observe. Structural analysis of condensin holocomplexes has revealed an exchange of the HEAT-repeat subunits Ycg1 and Ycs4 binding to the SMC ATPase head domains¹⁸. This exchange could itself be at the core of the 15 nm steps observed in our compaction traces. In addition, condensin’s long coiled-coils are fully bent at a discontinuity point called the “elbow” allowing the fold-back of the hinge and its contact with the coiled-coil region above the SMC heads¹⁸. AFM studies suggests that condensin dynamically shuttles between an elongated (O) and a folded (B) state³². MukBEF and cohesin have also been shown to have the ability to fold at an elbow region and to adopt extended and folded conformations³³. A model has been proposed to explain the transition between folded and extended conformations where a twist in the arms generated by the ATP cycle imposes mechanical strain up the SMC arms causing the extension/folding of the elbow³³. Interestingly, the distance between the elbow and hinge is approximately 15 nm¹⁸, it will therefore be of upmost importance to explore whether the dynamic extension and bending of the elbow, and hence the distal segment of the coiled-coils (between the elbow and the hinge), is involved in the conformational change that generates the 15nm steps observed here (Fig. 6B).

FACT facilitates condensin compaction in nucleosome arrays

Chromatin is the physiological substrate of condensin activity. Previous studies demonstrated that nucleosomes do not represent a barrier to loop extrusion for human CI and CII¹⁵. Surprisingly nucleosomes do not impede loop extrusion by human cohesin either¹⁶, although they represent a barrier for the ATP-independent diffusion of yeast cohesin¹⁹. Here, we report that yeast condensin cannot overcome nucleosome arrays as it compacts DNA by loop extrusion. It is surprising that human and yeast complexes behave differently with respect to their ability of achieving nucleosomal bypass. DNA compaction assays on DNA curtains did not follow the fate of individual SMC complexes encountering a nucleosome particle and the experiments were carried out in the presence of buffers containing excess of and unbound holocomplexes, where new binding events cannot be excluded. Moreover, the templates tested were sparse in nucleosomes and the experiments were done in the presence of DNA dyes^15,16, which alter core structure and properties³⁴.

In our experiments, compaction experiments were performed in condensin-free buffers, in the absence of DNA dyes, and with both low and high densities of nucleosomes. We only observed bypass in a minority (20%) of cases when single nucleosome barriers were used (Fig. 4F). It is therefore likely that differences in the experimental setup could explain the differing observations reported. In our assays we are not only imaging individual condensin complexes but also measuring their lone compaction activity. Importantly, the use of optical tweezers in multi-channel flow cells with fluorescence microscopy capabilities provides unparalleled control including the ability to ensure truly single molecule conditions are achieved as the DNA template can be rapidly moved between protein-free buffers and protein-containing channels.

The results presented here indicate that yeast condensin cannot compact nucleosome-bound DNA. And in turn, the question arises as to how is loop extrusion mediated by condensin in vivo on chromatinised chromosomes? Previously, we demonstrated that cohesin-dependent TAD (topologically associating domain)-like structures in G₁ and metaphase chromosomes are reduced in the absence of the histone chaperone FACT²⁰. FACT is a heterodimeric complex formed by Spt16 and Pob3³⁵ and conserved among eukaryotes. It interacts with the histones H2A-H2B and H3-H4 as well as with DNA³⁶. In vitro and in vivo studies have demonstrated that FACT can mediate nucleosome disassembly and reassembly thus facilitating processes that require access to DNA such as transcription and replication³⁶. FACT has been shown to be one of the core components required for the in vitro assembly of chromosomes from a set of purified factors, that besides FACT included condensin, topoisomerase II, nucleoplasmin, Nap1 and histones²¹. Therefore, we considered the possibility that the addition of purified FACT to our compaction assays with nucleosome-bound DNA templates could facilitate nucleosome bypass during condensin loop extrusion. Indeed, we observed that FACT allowed condensin to overcome nucleosomes (Fig. 5D). Although FACT is thought to destabilise the H2A-H2B dimer³⁶, we did not observe loss of H2A signal during bypass of single nucleosomes (Fig. 5H), demonstrating disassembly of H2A-H2B dimers is not the means for condensin overcoming nucleosomes during loop extrusion. Interestingly, recent studies have shown that FACT interacts extensively with nucleosomal-bound DNA while tethering H2A-H2B³⁷. It is therefore possible that FACT allows condensin loop extrusion through nucleosomes by loosening interactions between DNA and the core particle, which might allow condensin to access the 146bp of DNA normally tightly associated to the octamers.

At present, we do not understand the travelling path of DNA (and potentially histones) as it is extruded by condensin. However, it has been shown that condensin topologically entraps DNA³⁸ and therefore it is likely that DNA needs to travel through compartments of defined sizes within the structure, at least transiently. The presence of various sub-compartments has also been demonstrated for yeast cohesin³⁹. Therefore, it is possible that intact nucleosomes (~9nm in size) do not physically fit through the compartments that DNA needs to travel through and hence this prevents condensin from extruding nucleosome-bound DNA segments efficiently. Interestingly, FACT generates a partially disassembled subnucleosome organisation where FACT–H2A–H2B are docked onto the H3-H4 tetramer and DNA³⁷. A partial disassembly might facilitate passage through condensin compartments by reducing the physical dimensions of the particles.

Another possibility could be that FACT binding to nucleosomes generates a geometry of the nucleosomal-FACT-bound DNA where the entry and exit points of the DNA are close enough for condensin to take a step between them (Fig. 6C), rather than extruding through the entire 146 bp of DNA wrapped around histone octamers. It is interesting to note that in such a scenario, the presence of FACT and nucleosomes would speed up the linear DNA compaction by condensin, since a single condensin step overtaking a nucleosome would not only compact the length of the step itself, that is 45 bp, but also the 146 bp of DNA wrapped around the nucleosomal particle. In this case, FACT and nucleosomes would act as accelerators of the compaction achieved by the condensin-dependent steps. Therefore, the velocity of condensin-mediated loop extrusion on chromosomes would be significantly increased compared to naked DNA.

Previously we have shown that in yeast chromosomes, the formation of cohesin-dependent TAD-like structures requires FACT²⁰. To test whether our in vitro observations where we established a key role for FACT in facilitating condensin loop extrusion through nucleosomes (Fig. 5) translates to the in vivo situation, we investigated whether condensin localisation to pericentromeric regions of yeast metaphase chromosomes³¹ is influenced by FACT. Degradation of FACT had a significant impact on the ability of condensin to localise at pericentromeric regions (Fig. 6B), demonstrating that FACT complex is a mediator of condensin stability in vivo.

In the presented study, we have used optical tweezers to visualise single condensin complexes as they compact DNA molecules by extrusion of DNA loops. We found that condensin takes distinct steps of ~15 nm and that the dwell times between steps depends on the tension to which the DNA template is subjected to. We also demonstrate that yeast condensin alone cannot extrude loops on DNA containing nucleosome arrays and show that the histone chaperone FACT is required, demonstrating that FACT is an important factor facilitating loop extrusion of chromatin.

Acknowledgements

We thank A. Llauro (LUMICKS) for technical help. We thank our laboratory members for discussion and critical reading of the manuscript. We thank D. D’Amours, C. Haering, and J. Diffley for sharing plasmids. We thank Paul Girvan for helping us in setting up the TIRFM.

Funding: Work in the Aragon laboratory is supported by the Medical Research Council (UKRI MC-A652-5PY00) and the Wellcome Trust (100955/Z/13/Z). Work in the Rueda laboratory was funded by the Medical Research Council (UKRI MC-A658-5TY10) and the Wellcome Trust (P67153). We thank the Institute of Cancer Research Structural Biology Division and Stephen Hearnshaw for use of the Refeyn One Mass Photometer.

Author contributions: SS, GF and PG-E expressed and purified yeast condensin. CC expressed and purified yeast FACT and histones. EC collected mass photometry data. S.S. collected optical tweezers and TIRF datasets. JG-L and SS performed chromatin immunoprecipitations. LA and D.S.R. conceived and supervised the project. LA and DSR wrote the manuscript. LA, DSR and SS revised the manuscript.

Competing interests: The authors declare that they have no competing interests.

Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

1 Hassler, M., Shaltiel, I. A. & Haering, C. H. Towards a Unified Model of SMC Complex Function. Curr Biol 28, R1266-R1281, doi:10.1016/j.cub.2018.08.034 (2018).

2 Hirano, T. SMC protein complexes and higher-order chromosome dynamics. Curr Opin Cell Biol 10, 317-322, doi:10.1016/s0955-0674(98)80006-9 (1998).

3 Dowen, J. M. & Young, R. A. SMC complexes link gene expression and genome architecture. Curr Opin Genet Dev 25, 131-137, doi:10.1016/j.gde.2013.11.009 (2014).

4 Wu, N. & Yu, H. The Smc complexes in DNA damage response. Cell Biosci 2, 5, doi:10.1186/2045-3701-2-5 (2012).

5 Aragon, L. The Smc5/6 Complex: New and Old Functions of the Enigmatic Long-Distance Relative. Annu Rev Genet 52, 89-107, doi:10.1146/annurev-genet-120417-031353 (2018).

6 van Ruiten, M. S. & Rowland, B. D. SMC Complexes: Universal DNA Looping Machines with Distinct Regulators. Trends Genet 34, 477-487, doi:10.1016/j.tig.2018.03.003 (2018).

7 Freeman, L., Aragon-Alcaide, L. & Strunnikov, A. The condensin complex governs chromosome condensation and mitotic transmission of rDNA. J Cell Biol 149, 811-824, doi:10.1083/jcb.149.4.811 (2000).

8 Ono, T. et al. Differential contributions of condensin I and condensin II to mitotic chromosome architecture in vertebrate cells. Cell 115, 109-121, doi:10.1016/s0092-8674(03)00724-4 (2003).

9 Hirano, T. Condensins: universal organizers of chromosomes with diverse functions. Genes Dev 26, 1659-1678, doi:10.1101/gad.194746.112 (2012).

10 Gibcus, J. H. et al. A pathway for mitotic chromosome formation. Science 359, doi:10.1126/science.aao6135 (2018).

11 Gassler, J. et al. A mechanism of cohesin-dependent loop extrusion organizes zygotic genome architecture. EMBO J 36, 3600-3618, doi:10.15252/embj.201798083 (2017).

12 Rao, S. S. P. et al. Cohesin Loss Eliminates All Loop Domains. Cell 171, 305-320 e324, doi:10.1016/j.cell.2017.09.026 (2017).

13 Dekker, J. & Mirny, L. The 3D Genome as Moderator of Chromosomal Communication. Cell 164, 1110-1121, doi:10.1016/j.cell.2016.02.007 (2016).

14 Ganji, M. et al. Real-time imaging of DNA loop extrusion by condensin. Science 360, 102-105, doi:10.1126/science.aar7831 (2018).

15 Kong, M. et al. Human Condensin I and II Drive Extensive ATP-Dependent Compaction of Nucleosome-Bound DNA. Mol Cell 79, 99-114 e119, doi:10.1016/j.molcel.2020.04.026 (2020).

16 Kim, Y., Shi, Z., Zhang, H., Finkelstein, I. J. & Yu, H. Human cohesin compacts DNA by loop extrusion. Science 366, 1345-1349, doi:10.1126/science.aaz4475 (2019).

17 Eeftens, J. M. et al. Real-time detection of condensin-driven DNA compaction reveals a multistep binding mechanism. EMBO J 36, 3448-3457, doi:10.15252/embj.201797596 (2017).

18 Lee, B. G. et al. Cryo-EM structures of holo condensin reveal a subunit flip-flop mechanism. Nat Struct Mol Biol 27, 743-751, doi:10.1038/s41594-020-0457-x (2020).

19 Stigler, J., Camdere, G. O., Koshland, D. E. & Greene, E. C. Single-Molecule Imaging Reveals a Collapsed Conformational State for DNA-Bound Cohesin. Cell Rep 15, 988-998, doi:10.1016/j.celrep.2016.04.003 (2016).

20 Garcia-Luis, J. et al. FACT mediates cohesin function on chromatin. Nat Struct Mol Biol 26, 970-979, doi:10.1038/s41594-019-0307-x (2019).

21 Shintomi, K., Takahashi, T. S. & Hirano, T. Reconstitution of mitotic chromatids with a minimum set of purified factors. Nat Cell Biol 17, 1014-1023, doi:10.1038/ncb3187 (2015).

22 Gutierrez-Escribano, P. et al. A conserved ATP- and Scc2/4-dependent activity for cohesin in tethering DNA molecules. Sci Adv 5, eaay6804, doi:10.1126/sciadv.aay6804 (2019).

23 Sonn-Segev, A. et al. Quantifying the heterogeneity of macromolecular machines by mass photometry. Nat Commun 11, 1772, doi:10.1038/s41467-020-15642-w (2020).

24 Strick, T. R., Kawaguchi, T. & Hirano, T. Real-time detection of single-molecule DNA compaction by condensin I. Curr Biol 14, 874-880, doi:10.1016/j.cub.2004.04.038 (2004).

25 Zhao, R. & Rueda, D. RNA folding dynamics by single-molecule fluorescence resonance energy transfer. Methods 49, 112-117, doi:10.1016/j.ymeth.2009.04.017 (2009).

26 Mora, M., Stannard, A. & Garcia-Manyes, S. The nanomechanics of individual proteins. Chem Soc Rev 49, 6816-6832, doi:10.1039/d0cs00426j (2020).

27 Terakawa, T. et al. The condensin complex is a mechanochemical motor that translocates along DNA. Science 358, 672-676, doi:10.1126/science.aan6516 (2017).

28 Kschonsak, M. et al. Structural Basis for a Safety-Belt Mechanism That Anchors Condensin to Chromosomes. Cell 171, 588-600 e524, doi:10.1016/j.cell.2017.09.008 (2017).

29 Hassler, M. et al. Structural Basis of an Asymmetric Condensin ATPase Cycle. Mol Cell 74, 1175-1188 e1179, doi:10.1016/j.molcel.2019.03.037 (2019).

30 Spakman, D., King, G. A., Peterman, E. J. G. & Wuite, G. J. L. Constructing arrays of nucleosome positioning sequences using Gibson Assembly for single-molecule studies. Sci Rep 10, 9903, doi:10.1038/s41598-020-66259-4 (2020).

31 Leonard, J. et al. Condensin Relocalization from Centromeres to Chromosome Arms Promotes Top2 Recruitment during Anaphase. Cell Rep 13, 2336-2344, doi:10.1016/j.celrep.2015.11.041 (2015).

32 Ryu, J. K. et al. The condensin holocomplex cycles dynamically between open and collapsed states. Nat Struct Mol Biol 27, 1134-1141, doi:10.1038/s41594-020-0508-3 (2020).

33 Burmann, F. et al. A folded conformation of MukBEF and cohesin. Nat Struct Mol Biol 26, 227-236, doi:10.1038/s41594-019-0196-z (2019).

34 Newton, M. D., Fairbanks, S. D., Thomas, J. A. & Rueda, D. S. A Minimal Load-and-Lock Ru(II) Luminescent DNA Probe. Angew Chem Int Ed Engl 60, 20952-20959, doi:10.1002/anie.202108077 (2021).

35 Chen, P. et al. Functions of FACT in Breaking the Nucleosome and Maintaining Its Integrity at the Single-Nucleosome Level. Mol Cell 71, 284-293 e284, doi:10.1016/j.molcel.2018.06.020 (2018).

36 Formosa, T. & Winston, F. The role of FACT in managing chromatin: disruption, assembly, or repair? Nucleic Acids Res 48, 11929-11941, doi:10.1093/nar/gkaa912 (2020).

37 Liu, Y. et al. FACT caught in the act of manipulating the nucleosome. Nature 577, 426-431, doi:10.1038/s41586-019-1820-0 (2020).

38 Cuylen, S., Metz, J. & Haering, C. H. Condensin structures chromosomal DNA through topological links. Nat Struct Mol Biol 18, 894-901, doi:10.1038/nsmb.2087 (2011).

39 Chapard, C., Jones, R., van Oepen, T., Scheinost, J. C. & Nasmyth, K. Sister DNA Entrapment between Juxtaposed Smc Heads and Kleisin of the Cohesin Complex. Mol Cell 75, 224-237 e225, doi:10.1016/j.molcel.2019.05.023 (2019).

There is NO Competing Interest.

SupplVideo1.mp4
Supplementary Video 1
SupplVideo2.mp4
Supplementary Video 2
SupplVideo3.mp4
Supplementary Video 3
SupplVideo4.mp4
Supplementary Video 4
SupplVideo5.mp4
Supplementary Video 5
SupportinginformationFinal.docx
Supplementary Information

PDF

Version 1

posted 05 Oct, 2021

You are reading this latest preprint version

Condensin compacts DNA in 15 nm steps and requires FACT for extrusion on chromatin

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Condensin takes defined steps compacting 15 nm at a time

FACT facilitates condensin compaction in nucleosome arrays

Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1