We sought to solve a co-crystal structure of C. crescentus ParB nucleating at parS. After screening several constructs with different lengths of ParB and parS, we obtained crystals of a 50 amino acid C-terminally truncated ParB in complex with a 22 bp parS DNA (Figure 1). This protein variant lacks the C-terminal domain (CTD) responsible for ParB dimerization (Figure 1A; Figge et al., 2003). Diffraction data for the ParB∆CTD-parS co-crystal were collected to 2.9 Å resolution, and the structure was solved by molecular replacement (see Materials and methods). The asymmetric unit contains four copies of ParB∆CTD and two copies of the parS DNA (Figure 1—figure supplement 1A,B).

Figure 1 with 2 supplements see all

Download asset Open asset

Each ParB∆CTD subunit consists of an NTD (helices α1–α4 and sheets β1–β4) and a DBD (helices α5–α10) (Figure 1B). Each ParB∆CTD binds to a half parS site, but there is no protein-protein contact between the two adjacent subunits (Figure 1B). We previously reported a 2.4 Å co-crystal structure of the DBD of C. crescentus ParB bound to parS (Jalal et al., 2020b) and elucidated the molecular basis for specific parS recognition, hence we focus on the conformation of the NTD here instead. We observed that helices α3 and α4 are packed towards the DBD and are connected to the rest of the NTD via an α3–β4 loop (Figure 1B,C). While the DBD and helices α3–α4 are near identical between the two ParB∆CTD subunits (root-mean-square deviation [RMSD] = 0.19 Å, Figure 1C), the rest of the NTD, from α1 to β4, adopts notably different conformations in the two subunits (Figure 1C,D). Specifically, NTDs (α1–β4) from the two ParB∆CTD subunits are related by a rotation of approximately 80^o due to changes in a flexible loop in between α3 and β4 (Figure 1D). Furthermore, by superimposing the C. crescentus ParB∆CTD-parS structure onto that of Helicobacter pylori (Chen et al., 2015), we observed that the NTDs of ParB from both species can adopt multiple alternative orientations (Figure 1—figure supplement 2). Taken together, these observations suggest that the ability of the NTD to adopt multiple open conformations is likely a general feature of nucleating ParB.

Next, to gain insight into the spreading state of ParB, we solved a 2.7 Å resolution structure of C. crescentus ParB∆CTD in complex with CTPɣS (see Materials and methods). At this resolution, it was not possible to assign the position of the ligand's sulfur atom. Indeed, the placement of the sulfur atom relative to the terminal phosphorus atom may vary from one ligand to the next in the crystal, leading to an averaging of the electron density. Hence, we modeled CTP, instead of CTPɣS, into the electron density (Figure 2 and Figure 2—figure supplement 1). The asymmetric unit contains two copies of ParB∆CTD, each with a CTPɣS molecule and a coordinated Mg²⁺ ion bound at the NTD (Figure 2A). In contrast to the open conformation of the ParB∆CTD-parS structure, nucleotide-bound NTDs from opposite subunits self-dimerize (with an interface area of 2111 Ų, as determined by PISA; Krissinel, 2015), thus adopting a closed conformation (Figure 2A). Multiple CTPɣS-contacting residues also directly contribute to the NTD self-dimerization interface (summarized in Figure 2—figure supplement 2), indicating a coupling between nucleotide binding and self-dimerization. Furthermore, the C. crescentus ParB∆CTD-CTPɣS structure is similar to that of the CDP-bound B. subtilis ParB∆CTD (RMSD = 1.48 Å) and the CTP-bound M. xanthus PadC∆NTD (RMSD = 2.23 Å) (Figure 2—figure supplement 3A), suggesting that the closed conformation at the NTD is structurally conserved in nucleotide-bound ParB/ParB-like proteins.

Figure 2 with 3 supplements see all

Download asset Open asset

Each CTPɣS molecule is sandwiched between helices α1, α2, α3 from one subunit and helix α3′ from the opposite subunit (Figure 2B). Ten amino acids form hydrogen-bonding contacts with three phosphate groups of CTPɣS, either directly or via the coordinated Mg²⁺ ion (Figure 2C). These phosphate-contacting residues are referred to as P-motifs 1–3, respectively (P for phosphate motif, Figure 2C). Four amino acids at helix α1 and the α1–β2 intervening loop provide hydrogen-bonding interactions to the cytosine ring, hence are termed the C-motif (C for cytosine motif, Figure 2C). Lastly, six additional residues contact the ribose moiety and/or the pyrimidine moiety via hydrophobic interactions (Figure 2C). Nucleotide-contacting residues in C. crescentus ParB and their corresponding amino acids in ParB/ParB-like homologs are summarized in Figure 2—figure supplement 2 and Figure 2—figure supplement 3B. The C-motif forms a snug fit to the pyrimidine moiety, thus is incompatible with larger purine moieties such as those from ATP or GTP. Hydrogen-bonding contacts from the G79 main chain and the S74 side chain to the amino group at position 4 of the cytosine moiety further distinguish CTP from UTP (Figure 2C). Taken all together, our structural data are consistent with the known specificity of C. crescentus ParB for CTP (Jalal et al., 2020c).

A direct comparison of the C. crescentus ParB∆CTD-parS structure to the ParB∆CTD-CTPɣS structure further revealed the conformational changes upon nucleotide binding. In the nucleating state, as represented by the ParB∆CTD-parS structure, helices α3 and α4 from each subunit bundle together (32^o angle between α3 and α4, Figure 3). However, in the spreading state, as represented by the ParB∆CTD-CTPɣS structure, α3 swings outwards by 101^o to pack itself with α4′ from the opposing subunit (Figure 3). Nucleotide binding most likely facilitates this ‘swinging-out’ conformation since both α3 and the α3–α4 loop, that is, P-motif 3 make numerous contacts with the bound CTPɣS and the coordinated Mg²⁺ ion (Figure 2C). The reciprocal exchange of helices ensures that the packing in the α3–α4 protein core remains intact, while likely driving the conformational changes for the rest of the NTD as well as the DBD (Figure 4A). Indeed, residues 44–121 at the NTD rotate wholesale by 94^o to dimerize with their counterpart from the opposing subunit (Figure 4A and Figure 4—figure supplement 1A). Also, residues 161–221 at the DBD rotate upward by 26^o in a near rigid-body movement (Figure 4A and Figure 4—figure supplement 1A). As a result, the opposite DBDs are closer together in the spreading state (inter-domain distance = ~ 27 Å) than in the nucleating state (inter-domain distance = ~ 36 Å) (Figure 4—figure supplement 1B). By overlaying the CTPɣS-bound structure onto the parS DNA complex, it is clear that the DBDs in the spreading state clash severely with DNA, hence are no longer compatible with parS DNA binding (Figure 4B). Our structural data are therefore consistent with the previous finding that CTP decreases C. crescentus ParB nucleation on parS or liberates pre-bound ParB from parS site (Jalal et al., 2020c). Overall, we suggest that CTP binding stabilizes a conformation that is incompatible with DNA-binding and that this change might facilitate ParB escape from the high-affinity nucleation parS site.

Figure 3

Download asset Open asset

Figure 4 with 1 supplement see all

Download asset Open asset

To verify the CTP-dependent closed conformation of ParB, we performed site-specific crosslinking of purified proteins using a sulfhydryl-to-sulfhydryl crosslinker bismaleimidoethane (BMOE) (Soh et al., 2019). Residues Q35, L224, and I304 at the NTD, DBD, and CTD, respectively (Figure 5A), were substituted individually to cysteine on an otherwise cysteine-less ParB (C297S) background (Jalal et al., 2020c), to create ParB variants where symmetry-related cysteines become covalently linked if they are within 8 Å of each other (Figure 5B). We observed that the crosslinking of both ParB (Q35C) and ParB (L224C) was enhanced ~2.5–3-fold in the presence of parS DNA and CTP (Figure 5B), consistent with CTP favoring a conformation when the NTD and the DBD are close together. In contrast, ParB (I304C) crosslinked independently of CTP or parS (Figure 5B), supporting the known role of the CTD as a primary dimerization domain (Figge et al., 2003; Fisher et al., 2017).

Figure 5 with 4 supplements see all

Download asset Open asset

Figure 5—source data 1

Original files, annotation of the full raw gels, and data used to generate Figure 5.

https://cdn.elifesciences.org/articles/69676/elife-69676-fig5-data1-v1.zip

Download elife-69676-fig5-data1-v1.zip

Previously, it was shown that B. subtilis ParB-CTP forms a protein clamp that entraps DNA (Soh et al., 2019); however, the location of DNA within the clamp is not yet clear. To locate such DNA-entrapping compartment, we employed a double crosslinking assay while taking advantage of the availability of crosslinkable cysteine residues in all three domains of C. crescentus ParB (Figure 5A). A C. crescentus ParB variant with crosslinkable NTD and CTD interfaces (Q35C I304C) was first constructed and purified (Figure 5C). ParB (Q35C I304C) could form high molecular weight (HMW) species near the top of the polyacrylamide gel in the presence of CTP, a 3 kb parS plasmid, and the crosslinker BMOE (lane 7, Figure 5C, left panel). The HMW smear on the polyacrylamide gel contained both protein and DNA as apparent from a dual staining with Coomassie and Sybr Green (Figure 5C, left panel). Slowly migrating DNA-stained bands were also observed when resolved on an agarose gel (Figure 5C, right panel). The HMW smear most likely contained DNA-protein catenates between a circular parS plasmid and a denatured but otherwise circularly crosslinked ParB (Q35C I304C) polypeptide. Indeed, a post-crosslinking treatment with Benzonase, a non-specific DNA nuclease (lane 8, Figure 5C, left panel) or the use of a linearized parS plasmid (lane 2 vs. lane 4, Figure 5—figure supplement 1) eliminated the HMW smear, presumably by unlinking the DNA-protein catenates. Lastly, the HMW smear was not observed when a plasmid containing a scrambled parS site was used (lane 10, Figure 5C, left panel) or when CTP was omitted from the crosslinking reaction (lane 6, Figure 5C, left panel), indicating that the DNA entrapment is dependent on parS and CTP. Collectively, these experiments demonstrate that as with the B. subtilis ParB homolog, C. crescentus ParB is also a CTP-dependent molecular clamp that can entrap parS DNA in between the NTD and the CTD.

Employing the same strategy, we further narrowed down the DNA-entrapping compartment by constructing a ParB (L224C I304C) variant in which both the DBD and the CTD are crosslinkable (Figure 5D). We found that crosslinked ParB (L224C I304C) also entrapped circular plasmid efficiently in a parS- and CTP-dependent manner, as judged by the appearance of the HMW smear near the top of the gel (lane 7, Figure 5D, left panel). By contrast, ParB (Q35C L224C) that has both the NTD and the DBD crosslinkable was unable to entrap DNA in any tested condition (Figure 5—figure supplement 2). We therefore hypothesized that ParB clamps entrap DNA within a compartment created by a 20-amino-acid linker in between the DBD and the CTD. To investigate further, we constructed a ParB (L224C I304C)-TEV variant, in which a TEV protease cleavage site was inserted within the DBD-CTD linker (Figure 5—figure supplement 3A). Again, ParB (L224C I304C)-TEV entrapped a circular parS plasmid efficiently in the presence of CTP (the HMW smear on lane 7, Figure 5—figure supplement 3A). However, a post-crosslinking treatment with TEV protease eliminated such HMW smear, presumably by creating a break in the polypeptide through which a circular plasmid could escape (lane 8, Figure 5—figure supplement 3A). We also extracted crosslinked ParB (L224C I304C) from gel slices that encompassed the HMW smear and electrophoresed the eluted proteins again on a denaturing gel to find a single band that migrated similarly to a double-crosslinked protein (lane 9, Figure 5—figure supplement 2B). Therefore, our results suggest that a ParB dimer, rather than ParB oligomers, is the major species that entraps DNA. Taken together, we suggest that C. crescentus ParB dimer functions as a molecular clamp that entraps parS-containing DNA within a DBD-CTD compartment upon CTP binding. This is also consistent with experiments that showed a premature and irreversible closing of ParB clamps, achieved either by an extended preincubation with CTPɣS (Jalal et al., 2020c and Figure 5—figure supplement 4B) or by pre-crosslinking a closed clamp form of ParB (Figure 5—figure supplement 4C), prevented nucleation at parS and DNA entrapment.

Next, we investigated the potential role(s) of CTP hydrolysis. Hydrolysis is unlikely to be required for DNA entrapment and translocation since ParB in complex with CTPɣS can still self-load and slide on DNA (Jalal et al., 2020c; Soh et al., 2019). M. xanthus ParB (N172A) and B. subtilis ParB (N112S) mutants, which bind but cannot hydrolyze CTP, failed to form higher-order protein-DNA complexes inside the cells (Osorio-Valeriano et al., 2019; Soh et al., 2019). However, these ParB variants are already impaired in NTD self-dimerization (Soh et al., 2019), hence the mechanistic role of CTP hydrolysis is still unclear. We postulated that creation of a ParB variant defective in CTP hydrolysis but otherwise competent in NTD self-dimerization would enable us to investigate the possible role of CTP hydrolysis. To this end, we performed alanine scanning mutagenesis on the CTP-binding pocket of C. crescentus ParB (Figure 2C). Eleven purified ParB variants were assayed for CTP binding by a membrane-spotting assay (DRaCALA) (Figure 6A), and for CTP hydrolysis by measuring the release rate of inorganic phosphate (Figure 6B). Moreover, their propensity for NTD self-dimerization was also analyzed by crosslinking with BMOE (Figure 6C and Figure 6—figure supplement 1). Lastly, their ability to nucleate, slide, and entrap a closed parS DNA substrate was investigated by a biolayer interferometry (BLI) assay (Figure 6D and Figure 6—figure supplement 2A). Immobilizing a dual biotin-labeled DNA on a streptavidin-coated BLI surface created a closed DNA substrate that can be entrapped by ParB-CTP clamps (Figure 5—figure supplement 4A; Jalal et al., 2020c). The BLI assay monitors wavelength shifts resulting from changes in the optical thickness of the probe surface during the association/dissociation of ParB with a closed DNA substrate in real time (Figure 6—figure supplement 2).

Figure 6 with 2 supplements see all

Download asset Open asset

Figure 6—source data 1

Original files, annotation of the full raw gels, and data used to generate Figure 6.

https://cdn.elifesciences.org/articles/69676/elife-69676-fig6-data1-v1.zip

Download elife-69676-fig6-data1-v1.zip

Overall, we identified several distinct classes of ParB mutants:

1. Class I: ParB (R60A), (R103A), (R104A), (R139A), (N136A), (G79S), and (S74A) did not bind or bound radiolabeled CTP only weakly (Figure 6A), thus also showed weak to no CTP hydrolysis (Figure 6B) or clamp-closing activity (Figure 6C,D).

2. Class II: ParB (Q58A) and (E135A) that are competent in CTP-binding (Figure 6A), but defective in CTP hydrolysis (Figure 6B) and in entrapping a closed parS DNA substrate (Figure 6D). We noted that ParB (Q58A) and ParB (E135A) already had an elevated crosslinking efficiency even in the absence of CTP (Figure 6C). This premature clamp closing might have resulted in a less than wild-type level of DNA entrapment (Figure 6D).

3. Class III: ParB (E102A) did not hydrolyze CTP (Figure 6B) but nevertheless bound CTP efficiently (Figure 6A) to self-dimerize at the NTD and to entrap DNA to the same level as ParB (WT) at all CTP concentrations (Figure 6C,D).

Upon a closer inspection of the BLI sensorgrams (Figure 6—figure supplement 2B and Figure 7), we noted that the entrapped ParB (E102A) did not noticeably dissociate from a closed DNA substrate when the probe was returned to a buffer-only solution (dissociation phase, k_off = 8.0 × 10^–4 ± 1.9 × 10^–4 s^–1, Figure 6—figure supplement 2B and Figure 7). By contrast, entrapped ParB (WT) dissociated approximately 15-fold faster into buffer (k_off = 1.2 × 10^–2 ± 3.7 × 10^–4 s^–1). Further experiments showed that DNA entrapment by ParB (E102A), unlike ParB (WT), is more tolerant to high-salt solution (up to 1 M NaCl, Figure 7A). Nevertheless, ParB (E102A)-CTP could not accumulate on a BamHI-restricted open DNA substrate (Figure 7B,C; Jalal et al., 2020c), suggesting that ParB (E102A)-CTP, similar to ParB (WT), also form a closed clamp that runs off an open DNA end. Collectively, our results suggest that parS DNA and CTP induced a stably closed clamp conformation of ParB (E102A) in vitro.

Figure 7

Download asset Open asset

Figure 7—source data 1

Data used to generate Figure 7.

https://cdn.elifesciences.org/articles/69676/elife-69676-fig7-data1-v1.zip

Download elife-69676-fig7-data1-v1.zip

To investigate the function of ParB (E102A) in vivo, we expressed a FLAG-tagged version of parB (E102A) from a vanillate-inducible promoter (P_van) in a C. crescentus strain where the native parB was under the control of a xylose-inducible promoter (P_xyl) (Figure 8A). Cells were depleted of the native ParB by adding glucose for 4 hr, subsequently vanillate was added for another hour before cells were fixed with formaldehyde for ChIP-seq. Consistent with the previous report (Tran et al., 2018), the ChIP-seq profile of FLAG-ParB (WT) showed an ~10 kb region of enrichment above background with clearly defined peaks that correspond to the positions of parS sites (Figure 8A). By contrast, the ChIP-seq profile of FLAG-ParB (E102A) is significantly reduced in height but has an extra peak over the parB coding sequence (Figure 8A, asterisk). The instability of FLAG-ParB (E102A) in its native C. crescentus host, and hence the reduced protein level (Figure 8—figure supplement 1A), might explain the overall lower height of its ChIP-seq profile (Figure 8). The reason for an extra peak over parB in the ChIP-seq profile of ParB (E102A) is still, however, unknown. We also noted that expressing ParB (E102A) could not rescue cells with depleted ParB (WT) (Figure 8—figure supplement 2). Again, due to the caveat of a lower ParB (E102A) protein level in C. crescentus (Figure 8—figure supplement 1A), we could not reliably link the in vitro properties of ParB (E102A) to its behaviors in the native host.

Figure 8 with 3 supplements see all

Download asset Open asset

To overcome the caveat of protein instability, we instead investigated the spreading of ParB (WT) vs. ParB (E102A) from parS by analyzing the C. crescentus ParB/parS system in Escherichia coli. E. coli does not possess a ParA/ParB homolog nor a parS-like sequence, thus it serves as a suitable heterologous host. C. crescentus parS sites 3 and 4 were engineered onto the E. coli chromosome at the ygcE locus (Figure 8B). CFP-tagged ParB (WT/E102A) was expressed from a leaky lactose promoter (P_lac, no IPTG was added) on a medium-copy-number plasmid. CFP-ParB (WT/E102A) was produced at the same level, as judged by an immunoblot (Figure 8—figure supplement 1B). We observed by ChIP-seq that CFP-ParB (WT) in an E. coli host spreads asymmetrically ~5 kb around parS sites. By contrast, the shape of the ParB (E102A) distribution was clearly different from that of ParB (WT); the profile was further expanded to both neighboring sides of parS (covering in total ~26 kb) at the expense of the enrichment at parS itself (Figure 8B). The more excessive spreading of ParB (E102A) might suggest that this variant, in the absence of CTP hydrolysis, persisted and perhaps slid further away from the loading site parS in E. coli. The reduced enrichment of ParB (E102A) at parS itself (Figure 8B) might be due to reduced cytoplasmic ParB (E102A) available to re-nucleate at parS and/or due to stably entrapped ParB (E102A) sterically hindering further nucleation events. We also noted that the ChIP-seq profile of CFP-ParB (E102A) in E. coli is highly asymmetrical, with more enrichment in the 2905–2911 kb region than the 2885–2899 kb region (shaded areas, Figure 8B). The asymmetrical spreading is possibly due to an impediment in one direction by roadblocks such as RNA polymerases or DNA-bound proteins, which have been shown previously to be able to interfere with ParB spreading (Balaguer F de et al., 2021; Breier and Grossman, 2007; Jalal et al., 2020c; Murray et al., 2006; Rodionov et al., 1999; Soh et al., 2019).

Lastly, we quantified the fluorescence intensity of CFP-ParB (WT/E102A) foci inside cells and found a higher CFP signal for CFP-ParB (E102A) when compared with CFP-ParB (WT) (Figure 8—figure supplement 3). The higher intensity of the localizations could be due to more DNA-bound ParB (E102A) molecules surrounding the parS locus, which is consistent with the ChIP-seq observation showing CFP-ParB (E102A) occupying a more extended genomic area in E. coli. Altogether, at least in the heterologous E. coli host, the ‘clamp-locked’ phenotype of ParB (E102A) implies a possible role of CTP hydrolysis and/or the release of hydrolytic products in reopening wild-type ParB clamp to release DNA and to recycle ParB.

E. coli and C. crescentus were grown in LB and PYE, respectively. When appropriate, media were supplemented with antibiotics at the following concentrations (liquid/solid media for C. crescentus; liquid/solid media for E. coli [μg/mL]): carbenicillin (E. coli only: 50/100), chloramphenicol (1/2; 20/30), kanamycin (5/25; 30/50), and oxytetracycline (1/2; 12/12).

Plasmid pET21b::parB∆CTD-(his)₆ was introduced into E. coli Rosetta (DE3)-competent cells (Merck) by heat-shock transformation. 40 mL overnight culture was used to inoculate 4 L of LB medium + carbenicillin + chloramphenicol. Cells were grown at 37°C with shaking at 250 rpm to an OD₆₀₀ of ~0.4. The culture was then left in the cold room to cool to 28°C before isopropyl-β-D-thiogalactopyranoside (IPTG) was added at a final concentration of 0.5 mM. The culture was shaken for an additional 3 hr at 30°C before cells were pelleted by centrifugation. Pelleted cells were resuspended in a buffer containing 100 mM Tris-HCl pH 8.0, 300 mM NaCl, 10 mM imidazole, 5% (v/v) glycerol, 1 µL of Benzonase nuclease (Merck), 5 mg of lysozyme (Merck), and an EDTA-free protease inhibitor tablet (Merck). Cells were further lyzed by sonification (10 cycles of 15 s with 10 s resting on ice in between each cycle). The cell debris was removed through centrifugation at 28,000 g for 30 min and the supernatant was filtered through a 0.45 µm sterile filter (Sartorius). The protein was then loaded into a 1 mL HisTrap column (GE Healthcare) that had been pre-equilibrated with buffer A (100 mM Tris-HCl pH 8.0, 300 mM NaCl, 10 mM imidazole, and 5% [v/v] glycerol). Protein was eluted from the column using an increasing (10–500 mM) imidazole gradient in the same buffer. ParB∆CTD-containing fractions were pooled and diluted to a conductivity of 16 mS/cm before being loaded onto a 1 mL Heparin HP column (GE Healthcare) that had been pre-equilibrated with 100 mM Tris-HCl pH 8.0, 25 mM NaCl, and 5% (v/v) glycerol. Protein was eluted from the Heparin column using an increasing (25 mM to 1 M NaCl) salt gradient in the same buffer. ParB∆CTD fractions were pooled and analyzed for purity by SDS-PAGE. Glycerol was then added to ParB∆CTD fractions to a final volume of 10% (v/v), followed by 10 mM EDTA and 1 mM DTT. The purified ParB∆CTD was subsequently aliquoted, snap frozen in liquid nitrogen, and stored at –80°C. ParB∆CTD that was used for X-ray crystallography was further polished via a gel-filtration column. To do so, purified ParB∆CTD was concentrated by centrifugation in an Amicon Ultra-15 3 kDa cutoff spin filters (Merck) before being loaded into a Superdex-200 gel filtration column (GE Healthcare). The gel filtration column was pre-equilibrated with buffer containing 10 mM Tris-HCl pH 8.0 and 250 mM NaCl. ParB∆CTD fractions were then pooled and analyzed for purity by SDS-PAGE.

Other C-terminally His-tagged ParB mutants were purified using HIS-Select Cobalt gravity flow columns as described previously (Jalal et al., 2020b). Purified proteins were desalted using a PD-10 column (Merck), concentrated using an Amicon Ultra-4 10 kDa cutoff spin column (Merck), and stored at –80°C in a storage buffer (100 mM Tris-HCl pH 8.0, 300 mM NaCl, and 10% [v/v] glycerol). Purified ParB mutants that were used in BMOE crosslinking experiments were buffer-exchanged and stored in a storage buffer supplemented with TCEP instead (100 mM Tris-HCl pH 7.4, 300 mM NaCl, 10% [v/v] glycerol, and 1 mM TCEP).

Different batches of proteins were purified by ASBJ and NTT. Both biological (new sample preparations from a stock aliquot) and technical (same sample preparation) replicates were performed for assays in this study.

A 22 bp palindromic single-stranded DNA fragment (parS: GGATGTTTCACGTGAAACA TCC) (100 µM in 1 mM Tris-HCl pH 8.0, 5 mM NaCl buffer) was heated at 98°C for 5 min before being left to cool down to room temperature overnight to form 50 µM double-stranded parS DNA. The core sequence of parS is underlined.

Crystallization screens for the C. crescentus ParB∆CTD-parS complex were set up in sitting-drop vapor diffusion format in MRC2 96-well crystallization plates with drops comprising 0.3 µL precipitant solution and 0.3 µL of protein-DNA complex, and incubated at 293 K. His-tagged ParB∆CTD (~10 mg/mL) was mixed with a 22 bp parS duplex DNA at a molar ratio of 2:1.2 (protein monomer:DNA) in buffer containing 10 mM Tris-HCl pH 8.0 and 250 mM NaCl. The ParB∆CTD-parS crystals grew in a solution containing 20.5% (w/v) PEG 3350, 260 mM magnesium formate, and 10% (v/v) glycerol. After optimization of an initial hit, suitable crystals were cryoprotected with 20% (v/v) glycerol and mounted in Litholoops (Molecular Dimensions) before flash-cooling by plunging into liquid nitrogen. X-ray data were recorded on beamline I04-1 at the Diamond Light Source (Oxfordshire, UK) using a Pilatus 6M-F hybrid photon counting detector (Dectris), with crystals maintained at 100 K by a Cryojet cryocooler (Oxford Instruments). Diffraction data were integrated and scaled using XDS (Kabsch, 2010) via the XIA2 expert system (Winter, 2009) then merged using AIMLESS (Evans and Murshudov, 2013). Data collection statistics are summarized in Table 1. The majority of the downstream analysis was performed through the CCP4i2 graphical user interface (Potterton et al., 2018).

The ParB∆CTD-parS complex crystallized in space group P2₁ with cell parameters of a = 54.3, b = 172.9, c = 72.9 Å, and β = 90.5° (Table 1). Analysis of the likely composition of the asymmetric unit (ASU) suggested that it contains four copies of the ParB∆CTD monomer and two copies of the 22 bp parS DNA duplex, giving an estimated solvent content of ~47%.

Interrogation of the Protein Data Bank with the sequence of the C. crescentus ParB∆CTD revealed two suitable template structures for molecular replacement: apo-ParB∆CTD from Thermus thermophilus (Leonard et al., 2004) (PDB accession code: 1VZ0; 46% identity over 82% of the sequence) and Helicobacter pylori ParB∆CTD bound to parS DNA (Chen et al., 2015) (PDB accession code: 4UMK; 42% identity over 75% of the sequence). First, single subunits taken from these two entries were trimmed using SCULPTOR (Bunkóczi and Read, 2011) to retain the parts of the structure that aligned with the C. crescentus ParB∆CTD sequence, and then all side chains were truncated to Cβ atoms using CHAINSAW (Stein, 2008). Comparison of these templates revealed a completely different relationship between the NTD and the DBD. Thus, we prepared search templates based on the individual domains rather than the subunits. The pairs of templates for each domain were then aligned and used as ensemble search models in PHASER (McCoy et al., 2007). For the DNA component, an ideal B-form DNA duplex was generated in COOT (Emsley and Cowtan, 2004) from a 22 bp palindromic sequence of parS. A variety of protocols were attempted in PHASER (McCoy et al., 2007), the best result was obtained by searching for the two DNA duplexes first, followed by four copies of the DBD, giving a TFZ score of 10.5 at 4.5 Å resolution. We found that the placement of the DBDs with respect to the DNA duplexes was analogous to that seen in the H. pylori ParB∆CTD-parS complex. After several iterations of rebuilding in COOT and refining the model in REFMAC5 (Murshudov et al., 1997), it was possible to manually dock one copy of the NTD template (from 1VZ0) into weak and fragmented electron density such that it could be joined to one of the DBDs. A superposition of this more complete subunit onto the other three copies revealed that in only one of these did the NTD agree with the electron density. Inspection of the remaining unfilled electron density showed evidence for the last two missing NTDs, which were also added by manual docking of the domain template (from 1VZ0). For the final stages, TLS refinement was used with a single TLS domain defined for each protein chain and for each DNA strand. The statistics of the final refined model, including validation output from MolProbity (Chen et al., 2010), are summarized in Table 1.

Crystallization screens for the C. crescentus ParB∆CTD-CTPɣS complex crystal were also set up in sitting-drop vapor diffusion format in MRC2 96-well crystallization plates with drops comprising 0.3 µL precipitant solution and 0.3 µL of protein solution (~10 mg/mL) supplemented with 1 mM CTPɣS (Jena Biosciences) and 1 mM MgCl₂, and incubated at 293 K. The ParB∆CTD-CTPɣS crystals grew in a solution containing 15% (w/v) PEG 3350, 0.26 M calcium acetate, 10% (v/v) glycerol, 1 mM CTPɣS, and 1 mM MgCl₂. Suitable crystals were cryoprotected with 20% (v/v) glycerol and mounted in Litholoops (Molecular Dimensions) before flash-cooling by plunging into liquid nitrogen. X-ray data were recorded on beamline I03 at the Diamond Light Source (Oxfordshire, UK) using an Eiger2 XE 16M hybrid photon counting detector (Dectris), with crystals maintained at 100 K by a Cryojet cryocooler (Oxford Instruments). Diffraction data were integrated and scaled using DIALS (Winter et al., 2018) via the XIA2 expert system (Winter, 2009), then merged using AIMLESS (Evans and Murshudov, 2013). Data collection statistics are summarized in Table 1. The majority of the downstream analysis was performed through the CCP4i2 graphical user interface (Potterton et al., 2018).

The ParB∆CTD-CTPɣS complex crystallized in space group P2₁ with cell parameters of a = 69.5, b = 56.1, c = 71.4 Å, and β = 98.4° (Table 1). Analysis of the likely composition of the ASU suggested that it contains two copies of the ParB∆CTD monomer giving an estimated solvent content of ~50%. Molecular replacement templates were generated from the ParB∆CTD-parS complex solved above. Attempts to solve the structure in PHASER using individual subunits taken from the latter in both conformations did not yield any convincing solutions, suggesting that the subunits had adopted new conformations. Given that the two subunit conformations observed in the previous structure differed largely in the relative dispositions of DBD and NTDs, we reasoned that a better outcome might be achieved by searching for the DBD and NTD separately. This time PHASER successfully placed two copies of each domain in the ASU such that they could be reconnected to give two subunits in a new conformation. The result was subjected to 100 cycles of jelly-body refinement in REFMAC5 before rebuilding with BUCCANEER (Cowtan, 2006) to give a model in which 77% of the expected residues had been fitted into two chains and sequenced. The model was completed after further iterations of model editing in COOT and refinement with REFMAC5. In this case, TLS refinement was not used as this gave poorer validation results. The statistics of the final refined model, including validation output from MolProbity (Chen et al., 2010), are summarized in Table 1.

BLI experiments were conducted using a BLItz system equipped with High Precision Streptavidin 2.0 (SAX2) Biosensors (Molecular Devices). BLItz monitors wavelength shifts (nm) resulting from changes in the optical thickness of the sensor surface during association or dissociation of the analyte. All BLI experiments were performed at 22°C. The streptavidin biosensor was hydrated in a low-salt-binding buffer (100 mM Tris-HCl pH 8.0, 100 mM NaCl, 1 mM MgCl₂, and 0.005% [v/v] Tween 20) for at least 10 min before each experiment. Biotinylated double-stranded DNA (dsDNA) was immobilized onto the surface of the SA biosensor through a cycle of baseline (30 s), association (120 s), and dissociation (120 s). Briefly, the tip of the biosensor was dipped into a binding buffer for 30 s to establish the baseline, then to 1 μM biotinylated dsDNA for 120 s, and finally to a low-salt-binding buffer for 120 s to allow for dissociation.

After the immobilization of DNA on the sensor, association reactions were monitored at 1 μM dimer concentration of ParB with an increasing concentration of CTP (0, 1, 5, 10, 50, 100, 500, 1000 µM) for 120 s. At the end of each binding step, the sensor was transferred into a protein-free binding buffer to follow the dissociation kinetics for 120 s. The sensor can be recycled by dipping in a high-salt buffer (100 mM Tris-HCl pH 8.0, 1000 mM NaCl, 10 mM EDTA, and 0.005% [v/v] Tween 20) for 5 min to remove bound ParB.

For the dissociation step in the BLI experiments in Figure 7A, the probe was returned to either a low-salt-binding buffer (100 mM Tris-HCl pH 8.0, 100 mM NaCl, 1 mM MgCl₂, and 0.005% [v/v] Tween 20) for 30 s or a high-salt buffer (100 mM Tris-HCl pH 8.0, 1 M NaCl, 1 mM MgCl₂, and 0.005% [v/v] Tween 20) for 30 s.

For experiments in Figure 7C, DNA-coated tips were dipped into 300 µL of restriction solution (266 µL of water, 30 µL of 10× buffer CutSmart [NEB], and 3 µL of BamHI-HF restriction enzyme [20,000 units/mL]) for 2 hr at 37°C. As a result, closed DNA on the BLI surface was cleaved to generate a free DNA end.

For experiments in Figure 5—figure supplement 4C, purified ParB (Q35C) was incubated with 1 mM CTPɣS in a binding buffer (100 mM Tris-HCl pH 7.4, 100 mM NaCl, 1 mM MgCl₂) for 30 min before BMOE was added to 1 mM. DTT was then added to the final concentration of 1 mM to quench the reaction. Subsequently, crosslinked ParB (Q35C) was buffer-exchanged into a storage buffer (100 mM Tris-HCl pH 8.0, 300 mM NaCl, and 10% glycerol) using 0.5 mL Zeba desalting columns (ThermoFisher). BLITZ assays were performed using 5 μM dimer concentration of crosslinked ParB (Q35C) ± 1 mM CTP.

All sensorgrams recorded during BLI experiments were analyzed using the BLItz analysis software (BLItz Pro version 1.2, Molecular Devices) and replotted in R for presentation. Each experiment was triplicated, standard deviations were calculated in Excel, and a representative sensorgram is presented in Figure 5—figure supplement 4, Figure 6—figure supplement 2, and Figure 7.

Purified C. crescentus ParB-His₆ (WT and mutants, at final concentrations of 0.7, 1.5, 3.1, 6.2, and 12.5 µM) were incubated with 5 nM radiolabeled P³²-α-CTP (Perkin Elmer), 30 µM of unlabeled CTP (ThermoFisher), and 1.5 μM of 22 bp parS DNA duplex in the binding buffer (100 mM Tris pH 8.0, 100 mM NaCl, and 10 mM CaCl₂) for 10 min at room temperature. 4 μL of samples were spotted slowly onto a nitrocellulose membrane and air-dried. The nitrocellulose membrane was wrapped in cling film before being exposed to a phosphor screen (GE Healthcare) for 2 min. Each DRaCALA assay was triplicated, and a representative autoradiograph was shown. Data were quantified using Multi-Gauge software 3.0 (Fujifilm), the bound fraction were quantified as described previously (Roelofs et al., 2011). Error bars represent standard deviations from triplicated experiments.

CTP hydrolysis was monitored using an EnzCheck Phosphate Assay Kit (ThermoFisher). Samples (100 µL) containing a reaction buffer supplemented with an increasing concentration of CTP (0, 1, 5, 10, 50, 100, 500, and 1000 µM), 0.5 µM of 22 bp parS DNA, and 1 µM ParB (WT or mutants) were assayed in a Biotek EON plate reader at 25°C for 8 hr with readings every minute. The reaction buffer (1 mL) typically contained 740 μL Ultrapure water, 50 μL 20× reaction buffer (100 mM Tris pH 8.0, 2 M NaCl, and 20 mM MgCl₂), 200 μL MESG substrate solution, and 10 μL purine nucleoside phosphorylase enzyme (one unit). Reactions with buffer only or buffer + CTP + 22 bp parS DNA only were also included as controls. The plates were shaken at 280 rpm continuously for 8 hr at 25°C. The inorganic phosphate standard curve was also constructed according to the manual. The results were analyzed using Excel and the CTPase rates were calculated using a linear regression fitting in Excel. Error bars represent standard deviations from triplicated experiments.

A 50 µL mixture of 8 µM ParB mutants (with residues at specific positions in the NTD, DBD, or CTD substituted to cysteine) ± CTP (0–1000 µM) ± 0.5 µM DNA (a 22 bp linear DNA or a 3 kb circular parS/scrambled parS plasmid) was assembled in a reaction buffer (10 mM Tris-HCl pH 7.4, 100 mM NaCl, and 1 mM MgCl₂) and incubated for 5 min at room temperature. BMOE (1 mM final concentration from a 20 mM stock solution) was then added, and the reaction was quickly mixed by three pulses of vortexing. SDS-PAGE sample buffer containing 23 mM β-mercaptoethanol was then added immediately to quench the crosslinking reaction. Samples were heated to 50°C for 5 min before being loaded on 12% Novex WedgeWell Tris-Glycine gels (ThermoFisher). Protein bands were stained with an InstantBlue Coomassie solution (Abcam) and band intensity was quantified using Image Studio Lite version 5.2 (LI-COR Biosciences). The crosslinked fractions were averaged, and their standard deviations from triplicated experiments were calculated in Excel.

For the experiment described in lane 8 of Figure 5C,D and Figure 5—figure supplement 2, crosslinking reactions were performed as described above; however, the reaction was quenched using a quenching buffer (10 mM Tris-HCl pH 7.4, 100 mM NaCl, 1 mM MgCl₂, and 2.3 mM β-mercaptoethanol) instead. Subsequently, 1 µL of a non-specific DNA nuclease (Benzonase, 250 units/µL, Merck) was added, and the mixture was incubated at room temperature for a further 10 min before SDS-PAGE sample buffer was added. Samples were heated to 50°C for 5 min before being loaded on 4–12% Novex WedgeWell Tris-Glycine gels (ThermoFisher).

For the experiments described in lane 8 of Figure 5—figure supplement 3A, crosslinking and quenching reactions were performed as described above before 1 µL of TEV protease (10 units/µL, ThermoFisher) was added. The mixture was incubated at room temperature for a further 30 min before SDS-PAGE sample buffer was added. Samples were heated to 50°C for 5 min before being loaded on 4–12% Novex WedgeWell Tris-Glycine gels.

For experiments described in lane 9 of Figure 5—figure supplement 3B, proteins were released from gel slices by a ‘crush & soak’ method. Briefly, 10 gel slices were cut out from unstained SDS-PAGE gels and transferred to a 2 mL Eppendorf tube. Gel slices were frozen in liquid nitrogen and were crushed using a plastic pestle. The resulting paste was soaked in 500 µL of soaking buffer (10 mM Tris-HCl pH 8, 100 mM NaCl, 1 mM MgCl₂, and 1 µL of Benzonase [250 units/µL]), and the tube was incubated with rotation in a rotating wheel overnight. On the next day, the tube was centrifuged at 13,000 rpm for 5 min and the supernatant was transferred to a new 1.5 mL Eppendorf tube. The sample volume was reduced to ~50 µL using a SpeedVac vacuum concentrator before SDS-PAGE sample buffer was added in. The entire sample was loaded onto a single well of a 4–12% WedgeWell Tris-Glycine gel.

For experiments described in Figure 5—figure supplement 1, a circular parS-harboring plasmid was linearized at an unique HindIII site by HindIII-HF restriction enzyme. After restriction, the linearized DNA was extracted with phenol-chloroform and ethanol precipitated before being used in double-crosslinking experiments.

Polyacrylamide gels were submerged in an InstantBlue Coomassie solution (Abcam) to stain for protein or in a SYBR Green solution (ThermoFisher) to stain for DNA. Denatured samples were also loaded on 1% TAE agarose gels and electrophoresed at 120 V for 40 min at room temperature. Afterwards, agarose gels were submerged in a SYBR green solution to stain for DNA.

α-FLAG ChIP-seq experiments on formaldehyde-fixed C. crescentus cells, and the subsequent data analysis was performed exactly as reported previously (Tran et al., 2018).

For ChIP-seq experiments on fixed E. coli cells, cells harboring pKTN25-cfp-parB (WT) or pKTN25-cfp-parB (E102A) were grown in 50 mL LB at 30°C to mid exponential phase (OD₆₀₀ ∼ 0.4, no IPTG was added). Subsequently, formaldehyde is added to a final concentration of 1% to fix the cells. All following steps are identical to ChIP-seq for C. crescentus, except that α-GFP antibody coupled to sepharose beads (Abcam) was used to immunoprecipitate CFP-tagged ParB–DNA complexes.

Each ChIP-seq experiment was duplicated using biological replicates. For a list of ChIP-seq experiments and their replicates in this study, see Supplementary file 1C.

For western blot analysis, C. crescentus or E. coli cells were pelleted and resuspended directly in 1× SDS sample buffer, then heated to 95°C for 5 min before loading. Total protein was run on 12% Novex WedgeWell gels (ThermoFisher) at 150 V for separation. The same amount of total protein was loaded on each lane. Resolved proteins were transferred to PVDF membranes using the Trans-Blot Turbo Transfer System (BioRad) and probed with either a 1:5000 dilution of α-FLAG HRP-conjugated antibody (Merck) antibody or a 1:5000 dilution of α-GFP HRP-conjugated antibody (Abcam). Blots were imaged after incubation with SuperSignal West PICO PLUS Chemiluminescent Substrate (ThermoFisher) using an Amersham Imager 600 (GE Healthcare). Western blot experiments were duplicated using biological replicates.

TLS3079 and TLS3080 were grown in M9 media supplemented with kanamycin (30 µg/mL) until OD₆₀₀ ~ 0.1 prior to imaging. The expression of cfp-parB (WT/E102A) was induced with 0.25 mM IPTG in culture for 60 min before imaging. Imaging was performed using a wide-field epifluorescence microscope (Eclipse Ti-2E, Nikon) with a 63× oil immersion objective (NA 1.41), illumination from pE4000 light source, Hamamatsu Orca Flash 4.0 camera, and a motorized XY stage. Images were acquired using NIS-elements software (version 5.1). For imaging in the CFP channel, 435 nm excitation wavelength was used with 1 s exposure.

Images were analyzed using ImageJ, and plots were generated in GraphPad Prism 8.0. For extracting information on number of ParB foci per cell as well as intensity of ParB in foci, the following analysis pipeline was implemented: cell masks were generated in ImageJ using analyze particle function on thresholds applied to phase profiles. Separately, even background subtraction function was applied to fluorescence profiles, images were convolved (using ‘subtract background’ and ‘convolve’ functions in ImageJ), and regions of interest (ROIs) for foci were detected via an application of appropriate thresholds. The cell masks and ROIs thus detected were applied to the raw data (after background correction) to extract intensity information for each ROI as well as total cell fluorescence. ROI intensity was plotted as a ratio of intensity within a focus (intensity_loc) normalized to total cell intensity (intensity_total). Along with intensity measurement, number of foci per cell was also recorded. The pipeline was implemented in ImageJ using the following command:

n = roiManager("count");for (j = 0; j < n; j++){roiManager("Select", j);run("Analyze Particles...", "size = 3–10 circularity = 0.40–1.00 display summarize add");}

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

This study was funded by the Royal Society University Research Fellowship (URF\R\201020), BBSRC grant (BB/P018165/1), and a Wellcome Trust grant (221776/Z/20/Z, to TBKL), and a DST-SERB CRG grant 2019/003321 (to AB). ASBJ’s PhD studentship was funded by the Royal Society (RG150448), and NTT was funded by the BBSRC grant-in-add (BBS/E/J/000PR9791 to the John Innes Centre). We thank Diamond Light Source for access to beamlines I04-1 and I03 under proposals MX13467 and MX18565 with support from the European Community’s Seventh Framework Program (FP7/2007-2013) under Grant Agreement 283570 (BioStruct-X). We thank Stephan Gruber, Martin Thanbichler, and Manuel Osorio-Valeriano for sharing unpublished results.

1. Preprint posted: May 5, 2021 (view preprint)
2. Received: May 11, 2021
3. Accepted: August 3, 2021
4. Version of Record published: August 16, 2021 (version 1)

© 2021, Jalal 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.