We initially attempted to detect topological interactions between B. subtilis Smc–ScpAB and plasmid DNA using pull-down assays as previously described (Ivanov and Nasmyth, 2005; Ghosh et al., 2009; Cuylen et al., 2011). However, several attempts failed to provide clear evidence for entrapment of small circular DNA by prokaryotic condensin. Conceivably, Smc–ScpAB does not interact with these artificial substrates in a physiological manner. To circumvent this possibility, we established an inverse assay by immobilizing whole chromosomes of B. subtilis in agarose plugs and monitoring their association with covalently closed rings of Smc–ScpA under harsh protein denaturing conditions (Figure 1A). To develop the chromosome entrapment assay we first performed experiments with the replicative sliding clamp, DnaN, in B. subtilis, which is known to entrap DNA in a topological manner. Furthermore, most of cellular DnaN protein is maintained in the vicinity of active replication forks in B. subtilis, presumably by its topological association with leading and lagging strand DNA (Su'etsugu and Errington, 2011).

Figure 1 with 1 supplement see all

Download asset Open asset

Based on the crystal structure of S. pneumoniae DnaN we engineered a pair of cysteine residues (N114C, V313C) into the B. subtilis protein so that DnaN can be cross-linked into covalent rings in the presence of a cysteine-specific cross-linker such as BMOE (Figure 1B). For detection a cys-less variant of the HaloTag (‘HT’) was fused to the C-terminus of DnaN (Figure 1—figure supplement 1B) and the construct was integrated into the genome of B. subtilis via allelic replacement at the endogenous locus. The dnaN-ht genes with and without cysteine mutations supported normal growth of B. subtilis, implying that they encoded functional DnaN proteins (data not shown). In vivo cross-linking of DnaN-HT resulted in two additional, slow migrating bands in SDS-PAGE gels (Figure 1C), corresponding to single and double cross-linked species of DnaN dimers, designated as X-DnaN-HT and XX-DnaN-HT, respectively (Figure 1—figure supplement 1A). We next embedded cells in agarose plugs and disrupted their cell walls by lysozyme digestion. Agarose plugs were then subjected to an electric field in the presence of SDS to denature and remove any unattached proteins from chromosomes. Plugs were finally treated with benzonase to digest genomic DNA and to release any stably entrapped protein. DnaN-HT protein was then analysed by in-gel fluorescence. Non-crosslinkable DnaN-HT was efficiently depleted from agarose plugs during the entrapment assay (Figure 1C). In contrast, the double cross-linked, circular form of DnaN(N114C, V313C)-HT (XX-DnaN-HT) was retained in the agarose plug during electrophoresis with high efficiency (∼50% of input). A minor fraction of single cross-linked DnaN dimer (X-DnaN-HT) was also observed. This is likely generated from XX-DnaN-HT by spontaneous hydrolysis of thiol-malemide adducts during protein isolation (Kalia and Raines, 2007; Baldwin and Kiick, 2013). Importantly, the presence of benzonase during cell lysis eliminated all DnaN from the plug, indicating that circular DnaN is retained in plugs via its interaction with cellular DNA. Furthermore, in the absence of the cross-linker BMOE, no DnaN-HT was detected in the eluate fraction (Figure 1—figure supplement 1C). The chromosome entrapment assay thus specifically detects a topological association of intact chromosomes with DNA sliding clamps and confirms that a major fraction (at least 50%) of DnaN is loaded onto DNA in rapidly growing cells.

Next, we used the newly developed chromosome entrapment assay to test for an association between native chromosomes and Smc–ScpAB complexes. Cysteine pairs were introduced at the Smc–Smc and at both Smc–ScpA interfaces and a HT was fused at the C-terminus of Smc to allow in-gel fluorescence detection (Figure 2A) (Bürmann et al., 2013). Strains bearing the cysteine mutations and the Smc-HaloTag fusion supported normal growth on nutrient rich medium demonstrating the functionality of the modified Smc complex (Figure 2—figure supplement 1A). Cells were treated with BMOE and extracts were analysed by SDS-PAGE. As internal control for the chromosome entrapment assay we employed the DnaN(N114C, V313C) protein, whose double cross-linked form was detected in input and eluate samples by immunoblotting (Figure 2—figure supplement 1B). Various species of Smc–ScpAB were identified in extracts of BMOE cross-linked cells by in-gel fluorescence. These correspond to fully cross-linked Smc–ScpA–Smc rings and several intermediate cross-linking species as reported previously (Figure 2B,C) (Bürmann et al., 2013). To reveal the identity of all species, strains lacking one of six engineered cysteines were used as controls that collectively form several intermediate cross-linked species but no fully cross-linked rings of Smc–ScpAB (Figure 2—figure supplement 1C). In these control samples little or no Smc-HT protein was retained in agarose plugs under denaturing conditions as expected for any non-circular protein (Figure 2B). In the presence of all pairs of cysteine, however, a set of two closely migrating species was consistently detected at significant levels after the chromosome entrapment assay (∼10–20% of input material) (Figure 2B). We argued that the two closely migrating species might correspond to Smc–ScpAB with a single or a double cross-link at the Smc hinge. Consistent with this notion we find that only a single species of Smc₂–ScpA accumulated during the chromosome entrapment assay when a single cysteine residue (R643C) was used to cross-link the Smc hinge domains (Figure 2—figure supplement 2A). These findings strongly suggest that Smc–ScpAB is bound to chromosomes via entrapment of chromosomal DNA. If this were indeed the case, then its retention in agarose plugs should depend on the integrity of Smc–ScpAB rings and of chromosomal DNA. Incubation of agarose plugs with the nuclease benzonase during cell lysis eliminated the Smc-HT and DnaN signal in the sample (Figure 2B, Figure 2—figure supplement 1B). To disrupt covalent Smc–ScpAB rings, we inserted cleavage sites for TEV protease into the linker region preceding the C-terminal WHD of ScpA and incubated cells during lysozyme treatment with recombinant TEV protease to open any circular Smc₂–ScpA species. As expected, little or no Smc-HT signal was detected in agarose plugs after TEV cleavage of ScpA (Figure 2D). To exclude any artefacts due to the presence of the HT on Smc we have repeated the chromosome entrapment assay with an untagged allele of Smc using immunoblotting with anti-Smc antibodies for the detection of cross-linked species, which yielded very similar results (Figure 2—figure supplement 2C). Furthermore, we found that Smc₂-ScpA rings are stably trapped in agarose plugs over extended periods of time in constant or alternating electric fields (data not shown). Thus, our chromosome entrapment assay specifically detects the association between intact chromosomal DNA and rings of Smc–ScpAB in B. subtilis, demonstrating that DNA fibers pass through the Smc ring.

Figure 2 with 2 supplements see all

Download asset Open asset

Next, we established the requirements for the formation of interconnections between Smc–ScpAB rings and chromosomes. The intrinsic ATPase activity of cohesin has previously been implicated in stable association with chromosomes (Arumugam et al., 2003; Weitzer et al., 2003). More specifically, ATP hydrolysis has been hypothesized to transiently open an entry gate for DNA in the cohesin ring during its loading onto chromosomes (Gruber et al., 2006; Hu et al., 2011). To test what steps of the ATP hydrolysis cycle in Smc–ScpAB are involved in the entrapment of chromosomal DNA, we made use of smc alleles harboring mutations that specifically prevent ATP binding (K37I), engagement of Smc head domains (S1090R) or ATP hydrolysis (E1118Q) (Figure 3A) (Hirano and Hirano, 2004). The three mutant proteins are expressed at normal levels in B. subtilis being indicative of proper protein folding (Figure 3—figure supplement 1A). However, they do not support growth on nutrient rich medium similar to smc null mutants, implying that all steps of the ATPase cycle are essential for Smc functionality (Figure 3—figure supplement 1B) (Gruber et al., 2014). For the chromosome entrapment assay, these Smc ATPase mutations were combined with cysteine mutations for BMOE cross-linking. To support their viability, the resulting strains as well as the wild-type controls were grown in minimal medium. The three mutant Smc proteins assembled into normal Smc–ScpAB complexes as judged by Smc–ScpA cross-linking, albeit there is a slight decrease in the fraction of ScpA proteins bridging Smc dimers and a concomitant minor increase in ScpA subunits bound to single Smc proteins (Figure 3—figure supplement 1C, species ‘e’ and ‘d’, respectively) (Bürmann et al., 2013). Intriguingly, the ATP binding and engagement mutants abolished the fraction of covalent ring species retained in the agarose plug during the chromosome entrapment assay (Figure 3B). In case of the ATP hydrolysis mutant Smc(E1118Q) only minute amounts of cross-linked rings were recovered from SDS treated plugs. This small fraction of stably bound condensin conceivably arises as a consequence of residual levels of ATP hydrolysis activity in Smc(E1118Q) (Hirano and Hirano, 2004). Thus, ATP binding and ATP dependent Smc head engagement—and most probably also ATP hydrolysis—are essential for entrapment of chromosomal DNA by condensin in bacteria, as has been supposed for cohesin in yeast. Furthermore, the strict requirement of several steps of the ATPase cycle strongly suggests that entrapment of DNA corresponds to the physiological form of association with the bacterial chromosome.

Figure 3 with 1 supplement see all

Download asset Open asset

What other factors might be required for the loading of condensin onto DNA? The ScpB subunit forms homodimers that bind in an asymmetric manner to the central region of a single ScpA monomer. It thus is in close proximity of the Smc ATPase domains. Together with ScpA it putatively plays a role in the regulation of the Smc ATPase activity (Kamada et al., 2013). Its precise molecular function, however, is not clear yet. To test whether ScpB is involved in the association of Smc–ScpA rings with chromosomes we combined the cysteine mutations in Smc and ScpA with an scpB in-frame deletion (Figure 4—figure supplement 1). Ring formation was only mildly affected by the absence of ScpB as judged by BMOE cross-linking and in-gel fluorescence detection (Figure 4A) (Bürmann et al., 2013). However, Smc complexes lacking ScpB subunits failed to entrap chromosomes altogether demonstrating that ScpB is absolutely required for loading of prokaryotic condensin onto chromosomal DNA.

Figure 4 with 1 supplement see all

Download asset Open asset

ParB proteins—bound to parS sites—are crucial for efficient targeting of Smc–ScpAB to a large region of the chromosome near the replication origin (Gruber and Errington, 2009; Sullivan et al., 2009; Minnen et al., 2011). ParB might act by simply increasing the local concentration of Smc–ScpAB around oriC either before or after its loading onto the chromosome. Alternatively, ParB bound to parS sites might be more directly involved in the loading reaction itself, for example, as catalytic factor, and its absence might thus affect levels of chromosomal condensin. To test this, we performed the chromosome entrapment assay with cells lacking the parB gene. Intriguingly, the levels of Smc–ScpAB entrapping chromosomal DNA were strongly reduced in the parB null mutant as judged by the limited retention of Smc–ScpA species in agarose plugs (Figure 4B). Thus, ParB protein likely promotes the entrapment of chromosomal DNA by Smc–ScpAB. This strongly suggests that most condensin is loaded onto the chromosome at parS sites, where ParB protein is bound. In all other parts of the chromosome entrapment of DNA fibers by Smc–ScpAB might be very inefficient. The cysteine bearing smc allele causes growth defects when combined with ΔparB (Figure 4—figure supplement 1). Therefore, we cannot formally exclude the possibility that the decreased loading of Smc observed in ΔparB are due to the cysteine modifications in Smc and that chromosomal loading of wild-type Smc is not or much less affected by parB deletion.

Previously, two parB point mutations (N112S and R149G), which prevent the formation of Smc-GFP foci, have been isolated in B. subtilis (Gruber and Errington, 2009). We found that these mutations strongly impair loading of Smc onto the chromosome in the entrapment assay similar to ΔparB (Figure 4B). The R149G mutation is positioned on the helix-turn-helix motif of ParB and might thus directly affect binding to parS sites (Leonard et al., 2004). The N112S mutation, however, is located in another highly conserved region, which has been implicated in the ‘spreading’ of ParB protein from parS sequences into adjacent DNA (Leonard et al., 2004; Graham et al., 2014). The spreading of ParB along several kb of DNA is a feature conserved in plasmid and chromosome derived ParB proteins, however, the underlying mechanism is only poorly understood (Rodionov et al., 1999). It might possibly involve the formation of a large nucleoprotein complex (Broedersz et al., 2014). Several other mutants of ParB (including B. subtilis ParB G77S and R80A) have been reported to be defective in spreading from parS sites (Breier and Grossman, 2007; Graham et al., 2014). Intriguingly, also these mutations resulted in largely reduced levels of Smc on the chromosome in our entrapment assay, being comparable to the levels found in a parB deletion mutant (Figure 4B). This implies that ParB spreading from parS sites or formation of large nucleoprotein complexes might be essential for loading of DNA into the Smc ring by ParB. These findings are consistent with the observation that formation of Smc-GFP foci near the origin of replication are affected by the G77S mutation (Sullivan et al., 2009). In summary, these results demonstrate that several factors—including ScpB protein, a ParB/parS nucleoprotein complex and the Smc ATPase cycle—are required to promote efficient loading of condensin rings onto the chromosome.

Smc proteins and fragments thereof exhibit affinity for single- and double-stranded DNA in vitro (Chiu et al., 2004; Hirano and Hirano, 2006; Soh et al., 2015). The physical contacts with DNA might occur once condensin has been successfully loaded onto chromosomes and thus be a permanent feature of chromosomal Smc–ScpAB. Alternatively, the direct association with DNA might be restricted to certain intermediates in the chromosomal loading reaction. To test for interactions between Smc–ScpAB and specific chromosomal DNA fragments, we have affinity-purified endogenous Smc–ScpAB from B. subtilis cell lysates using a short Avitag peptide fused to the C-terminus of the Smc protein, which gets biotinylated when the biotin ligase gene birA is co-expressed (‘Smc-Avitag’). We then examined fractions for the co-purification of fragments of chromosomal DNA—generated by restriction digest with XbaI—using quantitative PCR with primer pairs specific for different parts of the chromosome. Since we worried that Smc–ScpAB might not be sufficiently stable in diluted cell extracts, we cross-linked the three ring interfaces in Smc–ScpAB using BMOE cross-linking of engineered pairs of cysteines. A small fraction of chromosomal DNA was reproducibly co-purified with wild-type Smc-Avitag, whereas the yield of co-purified DNA was significantly improved by the presence of cross-linkable cysteine residues in Smc–ScpAB (Figure 5). In both cases origin-proximal regions (yyaD, parS-359, dnaA and dnaN) of the chromosome were more efficiently enriched than distal regions (amyE, trnS and ter) by the co-purification with Smc implying that the observed Smc-DNA contacts are dependent on chromosomal loading of Smc–ScpAB by ParB protein at parS sites and are thus physiologically relevant (Figure 5, Figure 5—figure supplement 1) (Gruber and Errington, 2009). The association of DNA with wild-type and BMOE cross-linked Smc–ScpAB was highly sensitive to washes with a salt solution (2M NaCl), suggesting that it was dependent on electrostatic contacts between DNA and protein. These DNA contacts are presumably formed by the Smc–ScpAB complex itself. Alternatively, albeit less likely, other chromosomal proteins physically bound to DNA could prevent the release of condensin from DNA by blocking its sliding towards DNA ends.

Figure 5 with 1 supplement see all

Download asset Open asset

In many cases, it is challenging to measure the activity and outcome of biochemical processes in the living cell. Here, we report the establishment of a straight-forward method to determine the physical association of ring-shaped protein complexes with whole bacterial chromosomes. Two examples, the SMC condensin complex and the sliding clamp DnaN, document the significant potential of our simple entrapment approach. In principle, similar assays should also be possible with eukaryotic cells and for many other chromosomal proteins such as for example hexameric helicases and certain transcription factors. Furthermore, analogous procedures might be useful to address biological questions related to other denaturation-resistant cellular structures such as cell wall polymers (e.g., made up of peptidoglycan, chitin or cellulose).

SMC–kleisin complexes are major governors of chromosome superstructure in most branches of the phylogenetic tree. The eukaryotic variants cohesin and condensin have been suggested to work as concatenases, which hold selected stretches of DNA together by simple embracement in their ring (Haering et al., 2008; Cuylen et al., 2011). Whether DNA entrapment is an ancestral and thus fundamentally conserved function of SMC–kleisin complexes, however, remained elusive so far. Furthermore, interaction studies developed for cohesin and condensin are based on small, artificial DNA substrates and might thus not necessarily reflect the mode of binding to native chromosomes. These assays also fall short of providing an estimate for the fraction of SMC complexes involved in interlocked associations with DNA and thus leave open the possibility that DNA entrapment might be an insignificant side reaction. Finally, it has not been tested under physiological conditions, whether the ATPase cycle is required for proper loading of DNA into any SMC–kleisin complex. To provide answers to these questions, we have established the chromosome entrapment assay to determine the association of prokaryotic condensin with native chromosomes. Our results clearly demonstrate that chromosomal DNA is loaded into condensin complexes in B. subtilis—in a manner that depends on the non-SMC subunit ScpB and at least one full cycle of Smc ATPase activity. The chromosome entrapment assay recovers about 10–20% of the fully cross-linked input material. This number probably understates the real proportion of chromosomally entrapped Smc complexes due to the loss of material during protein re-isolation from agarose plugs, due to possible adverse effects of cysteine mutations on Smc–ScpAB loading and because cysteine-maleimide linkages are vulnerable to hydrolytic reversal during and after the entrapment assay (Kalia and Raines, 2007; Baldwin and Kiick, 2013). Interestingly, a recent single-molecule tracking study in B. subtilis revealed two major populations of Smc: 80% of Smc-YFP proteins are displaying highly dynamic behavior on the nucleoid, whereas the other 20% (and most ScpA-YFP protein) are immobile and constrained within a small volume of the cell (Kleine Borgmann et al., 2013). This immobile fraction possibly represents Smc–ScpAB complexes embracing origin proximal DNA after loading at parS sites as observed in the entrapment assay.

Our results show that embracement of chromosomal DNA is a predominant feature of Smc–ScpAB, which has been evolutionarily retained in cohesin and likely all other SMC–kleisin complexes as well. Since the chromosome entrapment assay is based on the immobilization of intact replicating chromosomes, which possibly represent internally knotted and branched DNA structures, it is conceivable that Smc–ScpAB rings are linked to chromosomal DNA by non-topological capture of DNA loops, which themselves might be interlinked (i.e., knotted) with other parts of the chromosome. Therefore, it remains to be determined whether DNA entrapment by Smc–ScpAB is of topological (Figure 6A) and/or non-topological (Figure 6B) nature.

Figure 6 with 1 supplement see all

Download asset Open asset

Smc–ScpAB plays a crucial role in the segregation of replication origins in B. subtilis cells (Gruber, 2014; Wang et al., 2014), presumably by organizing nascent sister chromosomes so that their spatial overlap and entanglement is minimized. It is tempting to speculate that ParB/parS not only enriches Smc in the vicinity of the replication origin but also sets up lengthwise compaction of chromosomes by presenting DNA of a certain topology to the Smc–ScpAB complex. Consistent with this notion, we found that several parB mutants, which are defective in the ability to form large nucleoprotein complexes and to spread from parS sites, fail to promote loading of Smc–ScpAB onto the chromosome. Thus, Smc–ScpAB might capture, stabilize and expand structures–such as DNA loops or coils–that are pre-formed within larger ParB/parS nucleoprotein assemblies (Figure 6A). Alternatively, ParB/parS might serve as an elaborate landing platform on the chromosome, where Smc–ScpAB initiates with the help of ParB the lengthwise compaction of chromosomes by forming and extruding loops of DNA (Figure 6B). Extrusion of DNA might involve the translocation of Smc–ScpAB along DNA fibers made up of naked or ParB coated DNA. These models are not mutually exclusive. A reasonable first step towards understanding the architecture of Smc/ParB/parS assemblies might be the investigation of physical interactions between ParB/parS and Smc–ScpAB and their functional interconnection with the Smc ATPase cycle. In the future, Smc/ParB/parS structures could serve as a relatively simple paradigm for chromosome organization by the more intricately regulated cohesin and condensin complexes in eukaryotes.

ParB serves a supplemental—albeit important—role in the Smc loading process. In contrast, Smc functionality and its loading onto DNA in vivo is critically dependent on the ScpB subunit and the ATPase cycle. It remains unclear though what the exact role of ScpB and ATP hydrolysis in the entrapment of DNA might be. If loading indeed depends on transient opening of a DNA entry gate, the open state would likely represent an energetically unfavorable reaction intermediate (Figure 6A). Timely opening would require energy input as well as tight regulation. We have recently demonstrated that the ATP dependent engagement of Smc head domains—together with DNA binding to the Smc hinge domain—can transform the configuration of the Smc coiled coil from a rod to a more open ring-like conformation (Soh et al., 2015). Hydrolysis of ATP and/or ScpB binding could drive a subsequent conformational change that might open the SMC–kleisin ring (Figure 6A). Alternatively, ScpB and/or ATP hydrolysis might stabilize Smc–ScpAB once loaded onto DNA or promote Smc's sliding along DNA to allow efficient extrusion of DNA loops by the Smc–ScpAB ring (Figure 6B).

Deletion or mutation of the parB gene results in a clear drop in the levels of chromosomally bound condensin (∼5–10 fold less) in our chromosome entrapment assay (Figure 4B, Figure 2—figure supplement 2C). In addition to this loading defect, also the specific recruitment of Smc–ScpAB towards the replication origin is lost in the absence of ParB (Gruber and Errington, 2009; Sullivan et al., 2009). Thus, in parB mutants only a very small proportion of cellular Smc–ScpAB is bound to chromosomes within the replication origin region, where it presumably performs its essential function by promoting the separation of nascent sister chromosomes (Gruber et al., 2014; Wang et al., 2014). Nevertheless, defects in chromosome segregation are rather mild in parB mutants when compared to mutants of smc. Using quantitative blotting of Smc protein and replication origin DNA from cell extracts, we have estimated the average number of Smc protein to be around 30 dimers per replication origin in a fast growing population of cells (Figure 6—figure supplement 1). Assuming that all Smc complexes entrap chromosomal DNA in wild-type cells, only three to six Smc dimers (10–20% of total) are loaded onto the chromosome in parB mutants according to our measurements. Thus, a handful of Smc–ScpAB complexes, which are presumably randomly distributed over the chromosome, appears to be capable of supporting near-normal chromosome segregation under these conditions—when chromosome segregation is already compromised by the loss of the parABS system. Few Smc–ScpAB therefore seem to be able to provide enough organization to the replication origin region and the remainder of the chromosome to prevent lethal accumulation of inter-linked sister chromosomes. It is conceivable that individual Smc–ScpAB complexes are able to organize large chunks of a bacterial chromosome, possible by forming giant loops of DNA. Alternatively, Smc activity might be needed only at a limited number of defined locations on the B. subtilis chromosome and/or for very short periods of time. However, when levels of functional Smc dimers are in addition reduced for example by hypomorphic mutations in the smc gene itself, the loss of ParB protein becomes lethal (Gruber and Errington, 2009).

This work reveals the mode of association of Smc–ScpAB with bacterial chromosomes, highlights its striking evolutionary conservation and demonstrates the involvement of the SMC ATPase cycle in chromosomal loading. Future work must address the underlying biochemical mechanisms to get basic insight into the architectural role of SMC in chromosome biology.

Genetic modifications at smc, scpAB, parB and dnaN loci were generated via double cross-over recombination in strains derived from B. subtilis 1A700 or B. subtilis 168ED. Genotypes of strains used in this study are listed in Supplementary file 1. Cells were transformed with plasmids or B. subtilis genomic DNA using a 2-step starvation protocol as previously described (Hamoen et al., 2002; Bürmann et al., 2013). Transformants were selected by growth on nutrient agar (NA) plates (Oxoid, UK) supplemented with antibiotics as required: 5 µg ml⁻¹ kanamycin, 80 µg ml⁻¹ spectinomycin, 10 µg ml⁻¹ tetracycline, 5 µg ml⁻¹ chloramphenicol, 1 µg ml⁻¹ erythromycin and 25 µg ml⁻¹ lincomycin. Strains displaying a condensin null phenotype were selected on SMG medium instead: SMM salt solution (2 g l⁻¹ ammonium sulphate, 14 g l⁻¹ dipotassium hydrogen phosphate, 6 g l⁻¹ potassium dihydrogen phosphate, 1 g l⁻¹ trisodium citrate, 0.2 g l⁻¹ magnesium sulphate, 6 g l⁻¹ potassium hydrogen phosphate) supplemented with 5 g l⁻¹ glucose, 20 mg l⁻¹ tryptophan and 1 g l⁻¹ glutamate with the respective antibiotics. Strains were single-colony purified and grown in the absence of antibiotics for experiments.

Cells were pre-grown in a 96-well plate in SMG medium for 24 hr at 37°C. Overnight cultures were diluted 9^2-fold (high density spots) or 9^5-fold (low density spots) and spotted onto NA or SMG agar plates. Plates were incubated at 37°C for 12 hr on NA or 24 hr on SMG agar.

Cells were grown in either LB Miller medium (10 g l⁻¹ tryptone, 5 g l⁻¹ yeast extract, 10 g l⁻¹ sodium chloride) or SMG medium to mid-exponential phase at 37°C (in LB Miller medium, OD₆₀₀ of 0.4; in SMG medium, OD₆₀₀ of 0.03). Cells were harvested by centrifugation or vacuum filtration and washed in ice-cold PBS supplemented with 0.1% glycerol (‘PBSG’). Cell aliquots (corresponding to 1 ml at an OD₆₀₀ of 1.25) were re-suspended in PBSG and incubated with the cross-linker BMOE (bis-maleimidoethane, Applichem, Germany) at a concentration of 1 mM (diluted from a 20 mM stock solution in DMSO). After a 10 min incubation on ice the reaction was quenched by addition of 2-mercaptoethanol (‘2-ME’) to a final concentration of 28 mM.

For the preparation of protein extracts (‘input’) a mixture of following components was added to an aliquot of cells: 400 units ready-lyse lysozyme (Epicentre, Madison, WI), 12.5 units benzonase (Sigma-Aldrich, St. Louis, MO) and a protease-inhibitor cocktail (Sigma). In addition, 1 µM HT Oregon Green substrate (Promega, Madison, WI) was added to cell suspensions with HaloTag bearing alleles. Samples were then incubated for 20 min at 37°C protected from light. Finally, the samples were heated to 70°C for 5 min in LDS Sample Buffer (NuPage, Thermo Scientific, Waltham, MA) containing 200 mM DTT and loaded onto a SDS-PAGE gel (see below). Gels with Oregon Green labeled samples were scanned on a Typhoon scanner (GE Healthcare, UK) with Cy3-DIGE filter setup.

Cells were grown, cross-linked and quenched as described above. Lysozyme stock solution, protease inhibitor and HT substrate were added to an aliquot of cells at concentrations given above. The cell suspension was mixed immediately in a 1:1 ratio with a 2% solution of Megabase agarose (BioRad, Hercules, CA) or low-melt agarose (BioRad) and casted into 100 µl agarose plugs using plug molds (BioRad). Agarose plugs were incubated for 20 min at 37°C, protected from light, and then loaded into the wells of a 6% SDS-PAGE Tris-glycine gel. The polyacrylamide mini-gel was run for 60 min at 25 mA protected from light. Agarose plugs were then re-extracted from the gel and transferred into 1.5 ml Eppendorf tubes. 1 ml of Wash Buffer (‘WB’: 0.01 mM EDTA, 0.5 mM Tris, 0.5 mM MgCl₂, 0.01% SDS) was added per agarose plug. Plugs were incubated for 10 min with gentle agitation protected from light. This step was repeated once. Wash buffer was then discarded and replaced by 100 µl fresh WB supplemented with 50 units of benzonase (Sigma). Plugs were incubated at 37°C for 30 min. Plugs were melted at 85°C for 2 min under vigorous agitation. The samples were frozen at −80°C and stored overnight. Samples were then thawed, centrifuged for 10 min at 4°C and 14,000×g and transferred to a 0.45 µm Cellulose acetate spin column (Costar, Tewksbury, MA) and spun for 1 min at 10,000×g. The flow-through was concentrated in a Speed Vac (Thermo Scientific, no heating, 2.5 hr running time). The concentrated sample was resuspended in LDS Sample Buffer (NuPage) containing 200 mM DTT and heated for 3 min at 70°C. Samples were loaded onto Tris-acetate gels (3–8% Novex, Thermo Scientific) and run for 2.5 hr at 35 mA per gel at 4°C. For DnaN detection Bis-Tris gels (8–12% Novex) were run for 1 hr at 200 V at room temperature. Gels were either scanned on a Typhoon scanner (FLA 9000, GE Healthcare) with Cy3-DIGE filter setup or immuno-blotted using antibodies against DnaN or Smc (see below). For cleavage of ScpA(TEVs) or degradation of chromosomal DNA 15 units of His-TEV protease or 12.5 units of benzonase, respectively, was added before casting agarose plugs.

B. subtilis strains containing smc-tev-avitag alleles were grown to OD₆₀₀ of 0.4 in 100 ml LB Miller at 37°C. Part of the culture was fixed with formaldehyde and subjected to chromatin immunoprecipitation (ChIP) as described by (Gruber and Errington, 2009) using a rabbit anti-Smc antiserum. In parallel, 10 ml of the culture were mixed with ice, harvested by centrifugation and washed in cold PBSG. Cells were resuspended in 200 µl PBSG and treated with 0.5 mM BMOE for 10 min on ice. The reaction was quenched with 14 mM 2-ME and cells were washed once in CutSmart buffer (New England Biolabs, Ipswich, MA). Cells were resuspended in 200 µl CutSmart containing 10 kU Ready-Lyse lysozyme (Epicentre), 40 U XbaI (New England Biolabs) and a protease inhibitor cocktail (Sigma). The suspension was incubated for 15 min at 37°C before addition of 1800 µl buffer LS (10 mM Tris/HCl, 150 mM NH₄OAc, 1 mM EDTA, 6 mM 2-ME, 0.05% Tween-20, 0.01% NaN₃, final pH 7.9 at 23°C). Lysates were centrifuged for 5 min at 20,000×g. Subsequently, 1400 µl of the extract were incubated with 100 µl Dynabeads Streptavidin C1 for 30 min at room temperature. Beads were washed once in buffer LS, then split, resuspended either in buffer LS or in buffer HS (10 mM Tris/HCl, 2 M NaCl, 1 mM EDTA, 6 mM 2-ME, 0.05% Tween-20, 0.01% NaN₃, final pH 7.9 at 23°C) and incubated for 15 min at room temperature. Beads were washed twice with buffer LS, and protein/DNA complexes were eluted for 1 hr at 22°C by incubation with 350 µl LS containing TEV protease and 1 mM DTT. DNA from input and eluate fractions was purified by treatment with 0.5 mg/ml Proteinase K for 1 hr at 55°C followed by phenol/chloroform extraction. Samples were analysed by quantitative PCR using the second derivative maximum of a four parameter logistic model similar to the method described by (Zhao and Fernald, 2005).

After gels were scanned for in-gel fluorescence detection, they were immediately transferred onto a PVDF membrane (Immobilon-P, Merck Millipore, Germany) using semi-dry transfer. Membranes were blocked with 3.5% (wt/vol) milk powder in PBS with 0.1% Tween 20. Rabbit polyclonal sera against B. subtilis DnaN (Lenhart et al., 2013), B. subtilis Smc (this paper) and B. subtilis ParB (this paper) were used as primary antibodies for immunoblotting at dilutions of 1:5000 each. The membrane was developed with HRP-coupled secondary antibodies and chemiluminescence (SuperSignal West Femto, Thermo Scientific) and visualized on a LAS-3000 scanner (FujiFilm, Germany).

To estimate DnaN cross-linking kinetics (Figure 1—figure supplement 1A) samples were grown as described above. An aliquot of cells was incubated with the cross-linker BMOE (1 mM) for the indicated length of time before the reaction was quenched with 2-ME (28 mM).

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

We are very grateful to Lyle Simmons for providing B. subtilis dnaN-gfp strains and α-DnaN antiserum and to Mark Dillingham for sharing expression plasmids and purification protocols for untagged Smc protein. We thank Boris Pfander and Stefan Jentsch for sharing resources, Anna Reisenbichler for technical help and all members of the Gruber laboratory for stimulating discussions and helpful comments on the manuscript. This work was supported by an ERC Starting Grant (DiseNtAngle, #260853) to SG and by the Max Planck Society and by the National Research Foundation of Korea (NRF) grant funded by the Korea government (#2014-022694) to B-HO.

© 2015, Wilhelm et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.