ABSTRACT
In Bacillus subtilis, a ParB-like nucleoid occlusion protein (Noc) binds specifically to Noc-binding sites (NBS) around the chromosome to help coordinate chromosome segregation and cell division. Noc does so by binding to cytidine triphosphate (CTP) to form large membrane-associated nucleoprotein complexes to physically inhibit the assembly of the cell division machinery. The site-specific binding of Noc to NBS DNA is a prerequisite for CTP-binding and the subsequent formation of a membrane-active DNA-entrapped protein complex. Here, we solve the structure of a truncated B. subtilis Noc bound to NBS DNA to reveal the conformation of Noc at this crucial step. Our structure reveals the disengagement between the N-terminal CTP-binding domain and the NBS-binding domain of each DNA-bound Noc subunit, this is driven, in part, by the swapping of helices 4 and 5 at the interface of the two domains. Site-specific crosslinking data suggest that this conformation of Noc-NBS exists in solution. Overall, our results lend support to the recent proposal that parS/NBS-binding catalyzes CTP-binding and DNA-entrapment by preventing the re-engagement of the NTD and DBD from the same ParB/Noc subunit.
INTRODUCTION
Cells must couple chromosome segregation and division to reproduce efficiently. In Firmicutes, such as Bacillus subtilis, the nucleoid occlusion protein Noc contributes to the coordination between chromosome segregation and the initiation of cell division1–4. Noc helps direct the assembly of the cell division machinery towards the middle of a dividing cell where the concentration of DNA is the least, thus increasing cell division efficiency2–4. Critical to this function of Noc is its ability to recruit chromosomal DNA to the cell membrane to form large Noc-DNA-membrane complexes which inhibit the FtsZ formation over the nucleoid and/or to corral the FtsZ ring towards the mid-cell position5,6. Noc is a paralog of a chromosome partitioning protein ParB, and is also a CTPase enzyme that binds cytidine triphosphate (CTP) to form a protein clamp that can slide and entrap DNA6–8. Apo-Noc first binds to nucleate on 16-bp NBS (Noc-binding site) sites scattering along the chromosome6,9,10. The nucleation at NBS promotes CTP-binding and the subsequent engagement of N-terminal domains from opposing subunits of a Noc homodimer to form a clamp-closed complex that can escape from NBS to slide and spread to the neighboring DNA while still entrapping DNA6. The DNA-entrapped Noc-CTP complexes are also active at binding to the cell membrane due to the liberation of a 10-amino-acid membrane-targeting amphipathic helix6. As a result, Noc-CTP brings the entrapped chromosomal DNA close to the cell membrane to form large Noc-DNA-membrane complexes that are inhibitory to the assembly of nearby cell division machinery5,6.
Previously we solved two X-ray crystallography structures of the CTP-binding domain and DNA-binding domain of a Geobacillus thermoleovorans Noc to better understand the molecular mechanism of this protein family6. Nevertheless, it remains unclear how the Noc-NBS binding event mechanistically promotes the N-terminal domain engagement to form a closed-clamp Noc. To investigate further, in this study, we solve a structure of a B. subtilis Noc-NBS DNA complex to reveal the conformation of a nucleating Noc. Through comparisons to other available structures of Noc, and its paralog ParB, and by in-solution site-specific crosslinking, we provide evidence for the extended conformation of nucleating Noc.
RESULTS
Co-crystal structure of B. subtilis Noc with NBS DNA reveals that the N-terminal CTP-binding domain of each Noc subunit is disengaged from its DNA-binding domain
To gain insight into the nucleating state of Noc, we sought to determine a co-crystal structure of a Noc-NBS complex from B. subtilis. After screening several constructs with various lengths of Noc and NBS, we solved a 2.9 Å resolution crystal structure of B. subtilis NocΔCTD in complex with 16-bp NBS DNA duplex (see Materials and Methods) (Table 1). The NocΔCTD variant lacks the 41-amino-acid C-terminal domain (CTD) responsible for Noc dimerization (Figure 1A-B)5,6. The asymmetric unit contains two copies of NocΔCTD bound to a single 16-bp NBS DNA duplex (Figure 1B).
Each NocΔCTD subunit consists of an N-terminal CTP-binding domain (NTD) (helices α1 to 4 and sheets β1 to 4) and a DNA-binding domain (DBD) (helices α5 to 11) (Figure 1C). The electron density for the first 27 amino acids that contains the membrane-targeting peptide was poorly resolved, and thus this region was absent from the model (Figure 1A). Each NocΔCTD subunit is bound to a half NBS site, the NBS DNA adopts a conformation whereby in one strand the 5’ base was flipped out, and in the other, the 3’ base was flipped out, enabling a sticky-ended interaction (with a one-base overhang) between the duplexes in adjacent asymmetric units (Figure 1B and Supplementary Figure 1). We previously solved a structure of only the DBD of Noc with NBS (2.23Å, PDB: 6Y93) to elucidate the molecular basis for NBS-binding specificity10. Given that the conformation of the DBD and the core NBS site are similar between the previous structure and the structure in this work (root-mean-square deviation RSMD = 0.46 Å), we describe the conformation of the NTD in-depth here instead. By structural alignment of the two NocΔCTD subunits, we noted that the DBD and helices α4-5 are highly similar (RSMD = 0.27 Å) while the rest of the NTD (β1-β4) is orientated in a different direction (approx. 30° apart, owing to the flexible loop in between α4 and β4) (Figure 2A). The multiple alternative orientations at the NTD is likely a common feature of all nucleating ParB family proteins, including Noc. This was the case for the NTD of Caulobacter crescentus ParB bound to parS DNA11, and is also evidential from the superimposition of the B. subtilis NocΔCTD-NBS structure onto that of ParBΔCTD-parS from Helicobacter pylori and C. crescentus (Figure 2B)11,12. Multiple alternative conformations of nucleating ParB/Noc family members suggest flexibility at the N-terminal CTP-binding domain.
The most notable feature of the NocΔCTD-NBS structure is the disengagement of the NTD and DBD (Figure 1C), which is likely driven by the swinging-out conformation of α4-α5 (Figure 3A-B). Helices α4 and α5 from the same NocΔCTD subunit are not packed together, instead α4 swings outwards by approx. 100° to pack against α5’ from the adjacent NocΔCTD subunit (Figure 3A). This swinging-out conformation has not been observed in the previous structures of DNA-bound C. crescentus or H. pylori ParBΔCTD, of Thermus thermophilus ParBΔCTD-apo, or G. thermoleovorans NocΔCTD-apo (Figure 3B and Supplementary Figure 2)6,11–13. In previous structures of apo- or DNA-bound ParB/Noc, the equivalent helix α4 consistently folds back to pack with α5 from the same protein subunit (the folding-back conformation) (Figure 3B and Supplementary Figure 2). The swinging-out conformation of helices α4-5 is often associated with the nucleotide-bound state of ParB/Noc instead (Figure 3B and Supplementary Figure 2)6,7,11,14. It has been suggested that CTP-binding most likely facilitates the swinging-out conformation of ParB/Noc since nucleotides have been observed to make numerous contacts to both the equivalent α4 and the α4-α5 connecting loop in various ParB proteins7,8,11,14. The observation of a swingingout conformation in DNA-bound Noc is therefore surprising, given that CTP was not included in the crystallization drop and that CTP-binding is incompatible with high-affinity binding at the nucleation site NBS6. We reason that the swinging-out conformation might be thermodynamically possible in the DNA-bound nucleating ParB/Noc, and that CTP-binding, instead of facilitating, further stabilizes the swinging-out conformation.
Site-specific cysteine-cysteine crosslinking suggests the swinging-out conformation of Noc-NBS in solution
To test if the swinging-out conformation of α4-α5 is possible in NBS-bound Noc in solution, we employed site-specific chemical crosslinking with the cysteine-specific bismaleimide compound BMOE15. Based on the structures of apo-NocΔCTD6 and NocΔCTD-NBS, we engineered a dual cysteine substitution at E112 and H143 on an otherwise cysteine-free B. subtilis Noc to create a Noc (E112C H143C) variant (Figure 3A). In the folding-back conformation where helices α4 and α5 from the same Noc subunit pack together, crosslinking of E112C to H143C would generate an intramolecular crosslinked species (Noc IntraXL), while a swinging-out conformation would give rise to intermolecularly crosslinked species (a singly-crosslinked Noc InterXL and a doubly-crosslinked Noc Inter2XL) which are twice the theoretical molecular weight of a Noc monomer (Figure 4A). Crosslinking of apo-Noc (E112C H143C) only resulted in a prominent band that migrated faster in a denaturing acrylamide gel than noncrosslinked protein (Figure 4B, lane 1 vs. 2), this is most likely a Noc IntraXL species. Little of Noc InterXL or Inter2XL species was observed (~4.4% crosslinked fraction) suggesting that the swinging-out conformation is unfavored in apo-Noc (Figure 4B, lane 1 vs. lane 2). The addition of only CTP did not promote the swinging-out conformation noticeably (Figure 4B, lane 2, ~4.4% vs. lane 4, ~8.7% crosslinked fraction). The singly (InterXL) and the doubly (Inter2XL) crosslinked species appeared more prominently when NBS only (Figure 4B, lane 2, ~4.4% vs. lane 3, ~19.3% crosslinked fraction) or NBS + CTP were preincubated with Noc (Figure 4B, lane 2, ~4.4% vs. lane 5, ~31.5% crosslinked fraction). The InterXL/2XL fraction further increased when NBS was used in a molar excess to Noc (E112C H143C) (Supplementary Figure 3). We were able to assign different bands to either being InterXL or Inter2XL by performing crosslinking reactions of Noc (E112C H143C) + NBS + CTP with an increasing concentration of the BMOE crosslinker (Figure 4C). The assumption is that a singly-crosslinked InterXL preferably forms at a lower concentration of a crosslinker. Overall, our result here suggests that the swinging-out conformation of α4-5 is possible in solution and is promoted when Noc is bound to the NBS DNA.
DISCUSSION
In B. subtilis, noc resulted from parB via a gene duplication and neo-functionalization event10,16, and both Noc and ParB are CTP-dependent molecular switches7,8,17–20. CTP-binding switches nucleating ParB/Noc (bound at a high-affinity parS/NBS site) from an openclamp conformation (Figure 5A-B) to a closed-clamp conformation that can escape from parS/NBS to slide to neighboring DNA while still entrapping DNA (Figure 5C)6,7,14,17,18. The closed-clamp conformation is possible due to the new dimerization interface between the two adjacent N-terminal CTP-binding domains of ParB/Noc (the so-called NTD-NTD engagement, Figure 5C)6,7,14,17. Here, our NocΔCTD-NBS structure represents an openclamp conformation because there is no protein-protein contact between the majority of two adjacent NTDs of Noc, except for the swapping helices α4 and α4’ (Figure 1B and Figure 5B).
It has been observed that, without parS/NBS, CTP is unable to efficiently promote the NTD-NTD engagement to close the ParB/Noc clamp6,7,14,17. To rationalize this phenomenon, Antar et al (2021) noted that two ParB subunits would not be able to occupy a parS site if they were to adopt a conformation similar to apo-ParB (in which the NTD and the DBD of the same ParB subunit fold back on each other) because of a severe clash between opposing ParB subunits15. Antar et al (2021) proposed that, to avoid this potential clash, the NTD and the DBD from each parS-bound ParB must be untethered/disengaged from each other15. The DBD-NTD disengagement later favors the two opposing NTDs to dimerize in the presence of CTP to form a clamp-closed complex15. In sum, parS serves a catalyst in a reaction that favors the formation of the product (the closed clamp) by inhibiting the reversion to the substrate (the open clamp apo-ParB). Our structure of DNA-bound Noc here lends support to this hypothesis because the conformation of the DNA-bound Noc subunit is drastically different from that of apo-Noc, especially with the swinging-out helices α4-α5 disengaging the NTD and DBD from each other (Figure 5Bi). It is possible that NBS-bound Noc might exist as an ensemble of states with helices α4-α5 in either a folding-back (Figure 5Bii) or a swinging-out conformation (Figure 5Bi), and that the NocΔCTD-NBS structure here represents a snapshot of this dynamic process. The swinging-out conformation of α4-α5 might be rare in solution, given that the crosslinking reaction of Noc (E112C H143C) + NBS produced IntraXL as the major species. Nevertheless, the proportion of InterXL and Inter2XL increased substantially when NBS (Figure 4B, lane 3) is included in comparison to apo-Noc only (lane 2) or Noc + CTP only conditions (lane 4). Moreover, the proportion of InterXL and Inter2XL also increased when NBS was added in excess (Supplementary Figure 3). The proximity of adjacent Noc subunits and the restriction in movement by a DNA-fixated DBD may increase the likelihood of swapping helices α4-α5 in Noc. This might in part contribute to further promoting the NTD-NTD engagement upon CTP-binding (Figure 5C), and might additionally explain how NBS serves as a catalysis for NTD-NTD engagement and thus clamp closure for a ParB-like protein Noc. However, it is also worth noting that ParB, in the presence of parS, does not undergo α4-α5 swapping as readily as Noc-NBS (S. Gruber, personal communication)15. It is still unclear why this is the case and how it is related to the biological functions of ParB vs. Noc, but it explains why helices α4-α5 in all previous X-ray crystallography structures of ParB-parS complex are all in the folding-back conformation11,12.
MATERIALS AND METHODS
Plasmid and strain construction
Construction of pET21b:: Bacillus subtilis NocΔCTD-his6 and pET21b::noc (E112C H143C)-his6
The coding sequence of a 41-amino-acid C-terminally truncated B. subtilis Noc was amplified by PCR using a forward primer (aactttaagaaggagatatacatatgaagcattcattctctcg tttcttc) and a reverse primer (gtggtgctcgagtgcggccgcaagcttatctctgctgaatgc tttgcgtctc), and pET21b::B. subtilis Noc-his66 as template. The resulting PCR product was gel-purified and assembled into an NdeI-HindIII-cut pET21b using a 2x Gibson master mix (NEB). Gibson assembly was possible owing to a 23-bp sequence shared between the NdeI-and-HindIII cut pET21b backbone and the PCR amplified fragment. The 23-bp homologous region was introduced during the synthesis of the above primers.
A double-stranded DNA (dsDNA) fragment containing a B. subtilis noc (E112C H143C) gene was chemically synthesized (gBlocks, IDT). The gBlocks fragment was assembled into an NdeI-HindIII-cut pET21b using a 2x Gibson master mix to result in pET21b::noc (E112C H143C)-his6. All plasmids were verified by Sanger sequencing (Eurofins, Germany).
Protein overexpression and purification
B. subtilis NocΔCTD-His6 were purified through a 3-column (HisTrap, Heparin, Superdex-75 gel filtration) procedure as described previously6. Purified NocΔCTD-His6 was stored at −80°C in storage buffer (10 mM Tris-HCl pH 8.0 and 250 mM NaCl) before crystallization.
Noc (E112C H143C)-His6 was purified through a 2-column (His-Select Cobalt Affinity Gel, Superdex-200 gel filtration) procedure using the following buffers: buffer A-HisTrap (100 mM Tris-HCl pH 7.4, 300 mM NaCl, 10 mM imidazole, 5% (v/v) glycerol), buffer B-HisTrap (100 mM Tris-HCl pH 7.4, 300 mM NaCl, 500 mM imidazole, 5% (v/v) glycerol), and gel filtration buffer (100 mM Tris-HCl pH 7.4 and 300 mM NaCl). Purified protein was concentrated using an Amicon Ultra-4 10 kDa cut-off spin column, and stored at −80°C in storage buffer (100 mM Tris-HCl pH 7.4, 300 mM NaCl, 10 (v/v) glycerol, and 0.1 mM TCEP).
In vitro crosslinking using a sulfhydryl-to-sulfhydryl crosslinker bismaleimidoethane (BMOE)
Noc (E112C H143C)-His6 (4 μM final concentration) was incubated on ice either alone or with 1 mM CTP, or 1 μM 22-bp NBS DNA duplex (or with a twofold increasing concentration of NBS from 0 to 5 μM), or both in a crosslinking buffer (100 mM Tris-HCl pH 7.4, 130 mM NaCl, 5 mM MgCl2) for 10 min. Then, 20 mM DMSO solution of the crosslinking reagent (BMOE, ThermoFisher) was added to the reaction to the final concentration of 2 mM. The mixture was incubated at room temperature for 5 min before the crosslinking reaction was quenched by SDS-PAGE loading dye + β-mercaptoethanol. Samples were heated to 90°C for 10 min before being loaded on 4-12% Bis-Tris polyacrylamide gels (ThermoFisher). Each experiment was triplicated. Polyacrylamide gels were stained in an InstantBlue Coomassie solution (Abcam) and band intensity was quantified using Image Studio-Lite (LICOR Biosciences). Raw gel images were deposited to the Mendeley repository: doi: 10.17632/6sp26rm6zy.1
Reconstitution of NBS DNA for X-ray crystallography
A 16-bp NBS DNA fragment (5’-TATTTCCCGGGAAATA-3’) (3.6 mM in buffer containing 10 mM Tris-HCl pH 8.0 and 250 mM NaCl) was heated to 98°C for 5 min before being left to cool at room temperature overnight to form double-stranded NBS DNA (final concentration: 1.8 mM).
Protein crystallization, structure determination, and refinement
B. subtilis NocΔCTD-His6 (~10 mg/mL) was mixed with the 16-bp NBS DNA at a molar ratio of 1:1.2 (protein: DNA) in the gel filtration elution buffer (10 mM Tris-HCl pH 8.0, 250 mM NaCl). Crystallization screens were set up in sitting-drop vapor diffusion format in MRC2 96-well crystallization plates with drops comprised of 0.3 μL precipitant solution and 0.3 μL of protein and incubated at 293 K. After optimization of initial hits, the best crystals of the complex grew in a solution containing 17% (w/v) PEG3350, 0.25 M magnesium acetate and 10% (v/v) sucrose. These were cryoprotected in the crystallization solution supplemented 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 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 DIALS21 via the XIA2 expert system22 then merged using AIMLESS23 to a resolution of 2.9 Å in space group P212121 with cell parameters of a = 70.5, b = 99.3, c = 99.4 Å. Data collection statistics are summarized in Table 1. Analysis of the likely composition of the asymmetric unit (ASU) suggested that it contained two copies of the 29.5 kDa NocΔCTD monomer plus the 16-bp NBS duplex, giving an estimated solvent content of 51%.
The majority of the downstream analysis was performed through the CCP4i2 graphical user interface24. For molecular replacement, a template was constructed from the structure of the B. subtilis NOC DNA-binding domain (DBD) complexed to an NBS duplex (PDB accession code 6Y93)10. Initially, PHASER25 was run using the protein and DNA components of this entry comprising two copies of the DBD and one DNA duplex, although the latter was truncated from a 22mer to a 16mer. This yielded a good solution and, in common with the template structure, the DNA formed a pseudo-continuous filament spanning the crystal due to base-pair stacking between DNA fragments in adjacent ASUs. However, there was only sufficient space to accommodate 15 bp per ASU within this filament. For the time being, the DNA model was truncated to the central 14 bp NBS site in COOT26 before real space refining using “chain refine”. The model was subsequently refined with REFMAC527, using jelly body refinement giving Rwork and Rfree values of 0.363 and 0.404, respectively, to 2.9 Å resolution. Inspection of the electron density at this stage revealed evidence for the missing N-terminal domains (NTDs). A template for these was generated using SCULPTOR28 from the Geobacillus thermoleovorans NOC structure (PDB accession code 7NFU)6, where the corresponding domain shares 67% sequence identity with B. subtilis. After quickly tidying the output of the REFMAC5 job in COOT, this was put back into PHASER as a search model together with two copies of the NTD template. However, PHASER was only able to place one of the latter sensibly. After further jelly body refinement of this partial model (giving Rwork and Rfree values of 0.313 and 0.351, respectively, to 2.9 Å resolution) the electron density was inspected again in COOT, at which point it was possible to manually dock the missing domain into fragmented density. Following restrained refinement in REFMAC5, the density for the DNA was much clearer, enabling the missing DNA bases to be fitted. In one strand the 5’ base was flipped out, and in the other, the 3’ base was flipped out, enabling a sticky-ended interaction (with a one-base overhang) between the duplexes in adjacent ASUs. After further iterations of model building in COOT and restrained refinement in REFMAC5, the final model was produced with Rwork and Rfree values of 0.230 and 0.277, respectively, to 2.9 Å resolution. Refinement and validation statistics are summarized in Table 1.
ACKNOWLEDGEMENTS
This study was funded by the Royal Society University Research Fellowship Renewal (URF\R\201020 to T.B.K.L) and BBSRC (BBS/E/J/000PR9791 to the John Innes Centre). K.V.S is supported by Wellcome grant (221776/Z/20/Z). A.S.B.J’s PhD studentship was funded by the Royal Society (RG150448). We thank Diamond Light Source for access to beamline I04 under proposal MX18565. We thank Clare Stevenson for assistance with X-ray crystallography. We thank S. Gruber, H. Antar, and T. McLean for discussions and comments on the manuscript.