A conserved mechanism drives partition complex assembly on bacterial chromosomes and plasmids

ParB_F distribution pattern around parS_F is similar on chromosome and plasmid DNA

The F plasmid partition complex assembles on a centromere sequence, parS_F, composed of twelve 43‐bp tandem repeats (Helsberg & Eichenlaub, 1986), which contain ten 16‐bp inverted repeat motifs to which ParB_F binds specifically in vitro (Pillet et al, 2011) and in vivo (Sanchez et al, 2015). Partition complex assembly has been investigated using small versions of the F plasmid, either ~10 or ~60 Kbp. To discriminate between the different partition complex assembly models, we used two larger DNA molecules: the native 100‐Kbp F plasmid (F1‐10B; Appendix Table S1) and the 4.6‐Mbp E. coli chromosome with parS_F inserted at the xylE locus, in strains either expressing (DLT1472) or not (DLT1215) ParB_F from an IPTG‐inducible promoter.

We first verified the formation of ParB_F clusters on these two different DNA molecules using the ParB_F‐mVenus fluorescent fusion protein. ParB_F‐mVenus, fully functional in plasmid partitioning (Appendix Table S2), was expressed from the endogenous locus on the F plasmid (F1‐10B‐BmV) or from a low‐copy‐number plasmid under the control of an IPTG‐inducible promoter (pJYB294). In both cases, we observed bright and compact foci in nearly all cells (Fig 1B and D), indicating that the assembly of highly concentrated ParB_F clusters on parS_F from large DNA molecules, plasmid or chromosome, occurs similar to the smaller F plasmid counterparts (Sanchez et al, 2015). The number of foci from parS_F inserted on the chromosome is half of what is observed with the F plasmid, as expected from the twofold difference in copy number (Collins & Pritchard, 1973).

We then performed ChIP‐sequencing using anti‐ParB antibodies and compared the ParB_F patterns from the 100‐Kbp F1‐10B plasmid and the xylE::parS_F chromosome insertion (ChIP‐seq data are summarized in Table EV1). For F1‐10B, we observed a ParB binding pattern extending over 18 Kbp of parS_F‐flanking DNA nearly identical to the one previously observed on the 60‐Kbp F plasmid (Sanchez et al, 2015), with the asymmetrical distribution arising from RepE nucleoprotein complexes formed on the left side of parS_F on incC and ori2 iterons (Fig 1C). Besides the strong ParB binding enrichment in the vicinity of parS_F, no other difference in the pattern between the input and IP samples was observed on the F plasmid and on the E. coli chromosome. When parS_F is present on the chromosome, the ParB_F binding pattern displays a comparable enrichment of xylE::parS_F‐flanking DNA over 15 Kbp (Fig 1E). The ParB_F distribution extends ~9 and 6 Kbp on the right and left sides of parS_F, respectively. The asymmetry does not depend on parS_F orientation as an identical ParB_F binding pattern was observed with parS_F inserted in the reversed orientation (xylE::parS_F‐rev, Appendix Fig S1B and C). Similar patterns were also observed when ParB_F or ParB_F‐mVenus were expressed in trans from a plasmid (Appendix Figs S1D and S3D). To the left side of parS_F, ParB_F binding ends near the yjbE locus that harbors two promoters (locus A; Fig 1E, inset and Appendix Fig S1A), and to the right, ParB_F binding ends at the yjbI gene locus (locus E; Fig 1E and Appendix Fig S1A). A dip in the ParB binding intensity is also observed ~1 Kbp downstream from parS_F spanning ~300 bp, corresponding to a promoter region (locus C; Fig 1E and Appendix Fig S1A). Dips and peaks in this ParB_F binding pattern differ in terms of position and intensity when compared to the one present on the F plasmid. Overall, these data clearly indicate that the global ParB_F binding distribution around parS_F depends neither on the size nor the DNA molecule, plasmid or chromosome, and that the ParB_F binding probability is dependent on the local constraints of each given locus.

The “Nucleation & caging” binding model describes the partition complex assembly from the nucleation point to large genomic distance

Based on a smaller version of the F plasmid, we previously proposed the “Nucleation & caging” model describing ParB stochastic binding at large distance (> 100 bp) from parS due to DNA looping back into the confined ParB cluster. The characteristic asymptotic decay is compatible with a power law with the exponent b = −3/2, a property that is also observed with 100‐Kbp F plasmid (Fig 1C) and with parS_F inserted on the E. coli chromosome (Fig 1E and Appendix Fig S1C). This property is thus an intrinsic parameter of the ParB_F binding profile at distance > 100 bp from parS_F. The abrupt initial drop in ParB_F binding at a shorter genomic distance (< 100 bp) from parS_F is explained by the difference of ParB_F binding affinities between specific parS_F sites (K_d ~2 nM) and non‐specific DNA (K_d ~300 nM; Ah‐Seng et al, 2009). We modeled the DNA molecule by a Freely Jointed Chain (FJC) constituted of N monomers of size a [Kuhn length about twice the persistence length of the corresponding Worm‐like chain (Schiessel, 2013)]. One particle is always attached on parS whereas non‐specific sites are in contact with a reservoir of particles displaying a Gaussian distribution centered on parS. The ParB density was normalized to 1 by the value on the right side of parS and captured for non‐specific sites in the following phenomenological formula as the product of two probabilities integrated over the volume: (1)where is the probability for two DNA loci spaced by a genomic distance as to be at a distance r in space for a Gaussian polymer (de Gennes, 1979); is the equilibrium size of the section of DNA of linear length as; is the probability to find a protein ParB at a radial distance r from the centromere, with κ a normalization constant setting the total number N_t of ParB on the DNA molecule and σ the typical size of the cluster. Note that C(r) is the linearized form of the Langmuir model (Phillips et al, 2012) offering a more compact and intuitive expression for P_NC(s). From (1), we easily calculate (see Box 1 for the details of the calculation): (2)

Note that the decay versus the genomic distance as is asymptotically determined by a power law of exponent −3/2 modulated by an amplitude depending on the concentration of ParB. The model has only three parameters: σ = 75 nm is determined from superresolution microscopy (Lim et al, 2014; Sanchez et al, 2015). The two remaining parameters κ (a function of the total number of proteins N_t) and the Kuhn length a are readily obtained from a fit of ChIP‐seq data (see Box 1 for the calculation and Materials and Methods for the fitting procedure). Note that the relation between κ and N_t depends on the bioinformatics analysis (Appendix Fig S1E). We obtained κ = 0.41 for both F plasmid or parS_F‐chromosomal insertions, leading to 360 and 120 ParB per DNA molecule, respectively, in good agreement with former estimate (Bouet et al, 2005). The last remaining free parameter is the Kuhn length a, estimated to 10 or 22 bp for the F plasmid or parS_F‐chromosomal insertions, respectively, to fully describe the ParB_F DNA binding profiles (Fig 1C and E, and Appendix Fig S1D). These fitted values are lower than expected, likely due to the modeling that does not account for supercoiling and confinement. Nevertheless, using these defined parameters, the refined “Nucleation & caging” model provides a qualitative prediction of the experimental data over the whole range of genomic positions, from a few bp to more than 10 Kbp.

ParB_F DNA binding pattern over a wide range of ParB concentrations favors the “Nucleation & caging” model

The physical modeling for each proposed model (Broedersz et al, 2014; Sanchez et al, 2015) predicts distinct and characteristic responses upon variation of the intracellular ParB concentration (see explanations in Fig EV1B). Briefly, (i) the “1‐D filament” model predicts a rapid decrease of ParB binding followed by a constant binding profile dependent on ParB amount, (ii) the “Spreading & bridging” model predicts linear decays with slopes depending on the ParB amount, and (iii) the “Nucleation & caging” model predicts a binding profile which depends only on the size of the foci. The exponent b = −3/2 of the power law distribution would not change upon ParB amount variation resulting in an overall similar decay at a fixed focus size. In order to discriminate between these three model predictions, we performed ChIP‐seq experiments over a large range of intracellular ParB concentrations. To prevent interference with plasmid stability, we used the chromosomally encoded xylE::parS_F construct expressing parB_F under the control of an IPTG‐inducible promoter (DLT2075).

Without IPTG induction, ParB_F was expressed at ~0.2 of the physiological concentration from F plasmid, as judged by Western blot analyses (Appendix Fig S2B). We also tested an 8‐ and 14‐fold overproduction of ParB_F. Assuming the twofold difference in copy number (Fig 1B and D), these three conditions provided ParB_F/parS_F ratios of 0.4, 16, and 28, relative to the F plasmid one. At these three ratios, ChIP‐seq data revealed that ParB_F binding extended similarly over ~15 Kbp around parS_F. We analyzed the right side of parS_F displaying the longest propagation distance by normalizing each dataset (Fig 2A). It revealed that regardless of ParB_F concentration, (i) the ParB_F distribution in the vicinity of parS_F always displays a good correlation with a power law fitting with an exponent of −3/2, (ii) the ParB_F binding profile ends at the same genomic location, i.e. 9 Kbp from parS_F, and (iii) the location of the dips and peaks in the pattern is highly conserved, as confirmed by correlation analyses (Table EV2). These findings indicate a highly robust ParB_F binding pattern that is invariant over a ~70‐fold variation of the ParB_F amount.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. ParB_F DNA binding pattern is robust over a large range of intracellular ParB_F concentrations

1. Normalized and rescaled ParB_F binding profiles at different ParB_F/parS_F ratio. ChIP‐seq density on the right side of parS_F inserted at xylE was measured in DLT2075 induced (16, 28) or not (0.4) with IPTG (100 and 500 μM), or carrying HCN plasmids pZC302 (0.04) or pJYB57 (0.016), normalized as in Fig 1C and E with the amplitudes of the curves rescaled by the indicated factors (1.2, 10, or 50) to overlap with the curves of highest amplitude. The ParB/parS ratio is calculated relative to the one of F plasmid as determined from Western blot analyses (Appendix Fig S2B). Monte Carlo simulations and analytical formula are plotted with the same parameters as in Fig 1E. Note (i) that the dips at ˜9 Kbp are not visible for the low levels of available ParB since the signal is close to the basal level, and (ii) that the ChIP‐seq data at 100 μM IPTG induction (16) are the same as in Fig 1E. Inset: Same as in the main to display the density without rescaling.

2. ParB_F is dispersed in the cell upon titration by HCN plasmids. ParB_F‐mVenus expressed from pJYB294 was imaged as in Fig 1D in DLT3577 (left) and DLT3576 (right) carrying pZC302 and pJYB57, respectively. The number of extra parS_F per cell, indicated on top of each raw, is estimated from the copy number per cell of HCN plasmids carrying 10 specific binding sites. Scale bars: 1 μm.

3. The size of ParB_F clusters is independent of the intracellular ParB_F concentration. We considered two possible evolutions of the cluster size upon variations of ParB amount in the framework of “Nucleation & caging” with corresponding schematics drawn on the right. For direct comparison with (A), all curves are displayed with a rescaling of the amplitude corresponding to the WT expression level. Top: constant ParB concentration; supposing that clusters are compact, the cluster radius σ would depend on the number m of ParB like σ = m^1/3. Predictions profiles, plotted at different ratio of ParB/parS, vary within the range of the experimental levels tested. Bottom: constant cluster size; ParB concentrations vary but the range of exploration remains the same resulting in overlapping profiles.

To further vary the amount of ParB_F available for partition complex assembly, high‐copy‐number (HCN) plasmids containing the parS_F sequence were introduced into the xylE::parS_F strain to efficiently titrate ParB_F by its binding to the excess of specific binding sites (~200‐ and ~500‐fold on pBR322 and pBSKS derivatives, respectively; Diaz et al, 2015). Epifluorescence microscopy of these strains reveals that all cells display a diffuse ParB‐mVenus fluorescence (Fig 2B) in contrast to concise foci without titration (Fig 1A), suggesting a large reduction of ParB availability to non‐specific sites in the vicinity of parS_F on the chromosome. ChIP‐seq analyses in the two titration conditions revealed that ParB binding in the vicinity of parS_F was dramatically reduced as expected. However, rescaling the signals by a factor of 10 and 50 for the pBR322 and pBSKS parS_F‐carrying derivatives, corresponding to a ParB_F/parS_F ratio of 0.04 and 0.016, respectively, revealed a ParB_F binding pattern above the background level (Fig 2B, inset). In both datasets, ParB_F binding decreases progressively over about the same genomic distance and with a similar power law decay as without titration. Moreover, even with these very low amounts of available ParB_F, the dips and peaks in the profiles are present at similar positions (Table EV2).

The invariance of the overall ParB profile over three orders of magnitude of ParB concentration (Fig 2B, inset) excludes the predictions of both the “1‐D filament” and the “Spreading & bridging” models (Fig EV1). In addition, the conservation in the positions of the dips and peaks indicates that the probability of ParB_F binding at a given location is also not dependent on the amount of ParB_F in the clusters. These results are strongly in favor of the refined “Nucleation & caging” model presented above.

The size of the dynamic ParB/parS cluster is independent of ParB intracellular concentration

In all of the ParB induction levels tested, the genomic distance over which ParB_F binds around parS_F is constant and displays a very similar decay (Fig 2A). This conserved binding behavior could provide information on the cluster size as a function of ParB amount. Indeed, the “Nucleation & caging” model predicts a density per site (see equation (2)). Thus, the P_NC(s) decay is entirely determined by the geometry of the foci and the intrinsic flexibility of the DNA, and the overall amplitude depends on the number of ParB. Varying the ParB amount could lead to two limiting situations: (i) the density of ParB, but not σ, is constant, (ii) σ is fixed and ParB density is variable. We plotted these two situations in the range of ParB/parS ratio considered experimentally (Fig 2C): with (i) the different P_NC(s) strongly varied, and (ii) P_NC(s) was invariant relative to the ParB amount resulting in overlapping profiles. Experimental data (Fig 2A) are in excellent agreement with the latter. From this modeling, we thus concluded that the size of partition complexes is invariant to change in ParB intracellular concentration.

The arginine rich motif (box II) of ParB_F is critical for partition complex assembly

The ability of ParB to multimerize through dimer–dimer interactions is required for the formation of ParB clusters. A highly conserved patch of arginine residues present in the N‐terminal domain of ParB (box II motif; Yamaichi & Niki, 2000) has been proposed to be involved in ParB multimerization (Breier & Grossman, 2007; Song et al, 2017). To examine to what extent the box II motif is involved in vivo in the assembly of ParB_F clusters, we changed three arginine residues to alanine (Appendix Fig S3A). The resulting ParB_F‐3R* variant was purified and assayed for DNA binding activity by electro‐mobility shift assay (EMSA) in the presence of competitor DNA using a DNA probe containing a single parS_F site (Fig 3A). ParB_F‐3R* binds parS_F with high affinity (B1 complex) indicating no defect in parS binding nor dimerization, a property required for parS binding (Hanai et al, 1996). However, in contrast to WT ParB, the formation of secondary complexes (B'2 and B'3), resulting from non‐specific DNA binding and dimer–dimer interaction (Sanchez et al, 2015), was impaired further suggesting the implication of box II in dimer–dimer interaction. A mini‐F carrying the parB_F‐3R* allele (pAS30) was lost at a rate corresponding to random distribution at cell division (Appendix Table S2), indicating that this variant is unable to properly segregate the mini‐F.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. The box II motif of ParB_F is crucial for ParB_F binding in the vicinity of parS_F and cluster formation

1. The formation of secondary ParB_F‐DNA complexes requires the box II motif. EMSA was performed with a 144‐bp ³²P‐labeled DNA fragments (C144) carrying a single 16‐bp parS binding motif. Reaction mixtures containing 100 μg ml⁻¹ sonicated salmon sperm DNA were incubated in the absence (−) or the presence of increasing concentrations (gray triangle; 10, 30, 100, 300, and 1,000 nM) of ParB_F or ParB_F‐3R*. Positions of free and bound probes are indicated on the left. B1 represents complexes involving the specific interaction on the 16‐bp binding site, while B'2 and B'3 complexes represent secondary complexes involving the parS_F site with one or two additional nsDNA‐binding interactions, respectively (Sanchez et al, 2015).

2. ParB_F cluster formation requires the box II motif. Epifluorescence microscopy of ParB_F‐3R*‐mVenus from DLT3566 is displayed as in Fig 1D. Scale bars: 1 μm.

3. ParB_F in vivo DNA binding in the vicinity of parS_F sites requires the box II motif. ChIP‐seq was performed on DLT3726 carrying parS_F in the xylE chromosomal locus and expressing ParB_F‐3R* variant. ParB_F‐3R* DNA binding profile displayed the number of nucleotide reads as a function of the Escherichia coli genomic coordinates. The peak at parS_F covered approximately 950 bp, which corresponds to the 402 bp between the 1^st and 10^th specific binding sites and ˜280 bp on each sides (representing the average size of the DNA library; see Appendix Fig S3E). No ParB_F‐3R* enrichment was found on parS_F‐flanking DNA and elsewhere on the chromosome. Inset: Zoom in on the right side of parS_F over 5 Kbp with the ParB density, normalized to 1 at the first bp after the last parS binding repeat, plotted as a function of the distance from parS_F. Note that a highly similar DNA binding pattern is obtained with ParB_F‐3R*‐mVenus (strain DLT3566; Appendix Fig S3C).

The ParB_F‐3R* variant was then expressed in native or fluorescently tagged (ParB‐R3*‐mVenus) forms, from pJYB303 or pJYB296, respectively, in the xylE::parS_F strain. By imaging ParB_F‐3R*‐mVenus, we observed only faint foci in a high background of diffuse fluorescence (Fig 3B). These barely detectable foci may correspond to ParB_F‐3R*‐mVenus binding to the 10 specific sites present on parS_F and, if any, to residual ParB_F cluster formation. We then performed ChIP‐seq assays with ParB_F‐3R* present in ~25‐fold excess (relative ParB_F/parS_F ratio compared to the F plasmid one; Appendix Fig S3B). The resulting DNA binding profile displayed enrichment only at parS_F with a total absence of ParB_F binding on parS_F‐flanking DNA (Fig 3C). Indeed, no residual ParB_F binding to non‐specific DNA was detected when the size of the DNA fragments in the IP library is taken into account (Appendix Fig S3E). This pattern differs from those observed in conditions of ParB_F titration (Fig 2A; inset), indicating that the ParB_F‐3R* box II variant is fully deficient in clustering in vivo. The same patterns were also observed with ParB_F‐3R*‐mVenus (Appendix Fig S3C and F) indicating that the mVenus fluorescent‐tag fused to ParB_F does not promote cluster assembly.

Together, these results indicate that the box II variant is specifically deficient in ParB_F cluster assembly but not in parS_F binding, and thus reveal that the box II motif is critical for the auto‐assembly of the partition complex.

ParB also propagates stochastically from native chromosomal parS sites

ParABS systems are present on most bacterial chromosomes (Gerdes et al, 2000). To determine whether chromosomal ParB‐parS partition complexes also assembled in vivo in a similar manner to the F plasmid, we investigated the bacterium V. cholerae, whose genome is composed of two chromosomes. We focused on the largest chromosome to which ParB_Vc1 binds to three separated 16‐bp parS sites comprised within 7 Kbp (Saint‐Dic et al, 2006; Baek et al, 2014; Fig 4A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. ParB of Vibrio cholerae assembled in cluster similarly to ParB_F

1. Schematic representation of the genomic locus of the chromosome 1 of V. cholerae with the three parS sites, named parS1‐3. The rRNA operon (blue rectangle) spans the genomic coordinates 53,823–59,123.

2. ChIP‐seq performed on strain N16961 is displayed as the number of nucleotide reads in function of the genomic coordinates. Correspondence to the parS1‐3 location represented in (A) is indicated by gray dotted lines. The number of reads in the input sample (gray line) is normalized to the total number of reads in the IP sample.

3. We modeled the ChIP‐Seq data as in Fig 1C–E by means of MC simulations with a Freely jointed chain of size N = 2,000 monomers of size a = 16 bp. Data are normalized after background subtraction to the read value at parS1 (genomic coordinate 62,438). The best fit was achieved with σ = 25 nm and an amplitude κ = 0.15 leading to N_t˜50 ParB on the chromosome. In the MC simulation, we accounted for the finite width of the distribution around parS sites by including the average fragment size of the DNA library (304 bp; for comparison, a simulation without is provided in Appendix Fig S4B).

We purified ParB_Vc1 antibodies against his‐tagged ParB_Vc1 and performed ChIP‐seq assays on exponentially growing cultures. The ParB_Vc1 DNA binding pattern covered ~18 Kbp and displayed three peaks at the exact location of the three parS_Vc1 sites (Fig 4B). No other ParB binding was observed over the Vibrio genome. Each peak exhibits a distinct but reproducible difference in intensity that might correspond to the slight differences in parS_Vc1 sequences (Appendix Fig S4A). An asymmetry in the binding pattern was observed on the left side of parS1 with the limit of ParB_Vc1 binding corresponding to the end of the rRNA operon located ~4 Kbp upstream from parS1 (Fig 4B). This suggests that highly transcribed genes might significantly interfere with the extent of ParB binding.

We modeled ParB_Vc1 DNA binding profile with the framework of the refined “Nucleation and caging” model (see above). The simulations consider three non‐interacting spheres centered on each of the parS sites and take into account (Fig 4C) the average fragment size of the DNA library to account for the width of the peaks around each parS (same modeling as displayed in Appendix Fig S1E for the F plasmid). Simulations are found in good agreement with the ChIP‐seq data with the following parameters: σ = 25 nm, a = 16 bp, and κ = 0.15 leading to N_t~50 ParB proteins on the chromosome (see Materials and Methods for the fitting procedure). Overall, these parameters are of the same order of magnitude as those used for E. coli. The maxima in the ParB binding profile depends on the parS sites (Fig 4C) and are interpreted as a difference in binding affinity. In the simulations, the ParB density is normalized to 1 by the value on the right of parS1. The relative density of the two other parS sites is fixed according to the values read on the ChIP‐seq plot (3 and 29% lower affinity for parS2 and parS3 compared to parS1, respectively). We also noticed a clear difference at the minima of ParB binding on either side of parS2 (64.2 and 68 Kbp; Appendix Fig S4B). In the case of a single cluster constraining the three parS, the profile would only depend on the genomic distance from parS2 resulting in a symmetrical pattern, while in the case of three independent clusters, an absence of symmetry due to the occupation of the specific sites is expected. This indicates that the system displays three independent clusters nucleated at each parS sites. However, the possibility that these clusters mix together at a frequency dependent on the genomic distance between parS sites is not excluded. At larger distances from parS sites, differences between the experimental data and the simulation probably arise from strong impediments to ParB binding, such as the presence of the rRNA operon.

These data strongly support that the partition complex assembly mechanism is conserved on plasmid and chromosome ParABS systems.

Nucleoprotein complexes, but not active transcription, are the major determinants for the impediment of ParB stochastic binding

The major dips in the ParB_F DNA binding signal are often found at promoter loci (Appendix Fig S1A). To investigate the link between gene expression and the impediment to ParB propagation, we reproduced the ChIP‐seq assays using the xylE::parS_F strain grown in the presence of rifampicin, an inhibitor of RNA synthesis that traps RNA polymerases at promoters loci in an abortive complex unable to extend RNAs beyond a few nucleotides (Herring et al, 2005). We did not observe significant changes to the ParB signal on either side of parS_F (Fig 5A; compare red and blue curves). Notably, the ParB signal still strongly drops in promoter regions (e.g., loci A, C, and E) and the dips and peaks are present at the same locations (Fig 5B and Table EV2). This indicates that active transcription by RNA polymerase is not a major impediment to ParB binding, but rather that RNA polymerases bound or stalled at the promoter could.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. Robust dips and peaks signatures in ParB DNA binding profiles

ChIP‐sequencing assays were performed on DLT2075 (xylE::parS_F) expressing ParB_F grown in exponential (expo) or stationary (stat) phases with addition of rifampicin when indicated (+Rif). The assays have been performed in duplicate for the +Rif and once for the stationary phase experiments.

1. ParB_F DNA binding around parS_F is independent of active transcription. The color‐coded ParB_F profiles are represented over 50 Kbp as the relative ParB density normalized to 1 at the first bp after the last parS_F binding site. Loci A, C, E, and F are defined in Appendix Fig S1A.

2. The dips and peaks are highly similar in the three indicated conditions. Same as in (A) with zoom in on the right side of parS_F up to 9 Kbp and normalization to 1 at genomic coordinate 230. The dotted line corresponds to the analytics description of “Nucleation and caging” (see details in Fig 1C–E).

3. ParB_F binding profile upstream of the locus A. Same as in (A) with zoom in from −6.5 to −16.5 Kbp by normalization to 1 at genomic coordinate −6.5 Kbp (upstream of the dip at the locus A). The ParB_F DNA binding profile remains compatible to a power law, represented by the analytics description (dotted line), upstream of the locus A in stationary phase (black) and in exponential phase (blue). Also, the dips and peaks are highly similar in both conditions. These data are not in favor of the “1D‐spreading” or the “Spreading and bridging” models that predicts a basal uniform distribution or a linear decrease after a barrier, respectively (Broedersz et al, 2014).

4. The promoter region at locus A prevents ParB_F DNA binding. ChipIP‐seq assays were performed in isogenic xylE::parS_F strains (DLT2075; black curve) in which the locus A is replaced by a kanamycin gene (DLT3651; red curve). The assay in the Δ(locus A) genomic context has been performed once. The relative ParB density as a function of the distance from parS_F is drawn and normalized as in (A). The promoter region is depicted as in Appendix Fig S1A.

We also measured the ParB binding profile in stationary phase, a growth condition in which gene expression is strongly reduced. On the right side of parS_F, ParB distribution was similar to all other tested conditions (Fig 5A), thus confirming the robustness of the binding pattern. On both sides, the strong reduction of ParB binding at loci A, C, and E was still observed. However, in contrast to the other conditions, ParB binding recovers after these loci and extends up to ~18 Kbp on both sides, resulting in the location of parS_F in the middle of a ~36 Kbp propagation zone. Interestingly, the ParB binding profiles after these recoveries are still compatible with a power law exhibiting the same characteristics as at lower genomic distances (Fig 5C). In stationary phase, the reduced intracellular dynamics (Parry et al, 2014) and the higher compaction of the DNA (Meyer & Grainger, 2013) may stabilize the partition complex revealing the ParB_F bound at larger distances from parS_F. Also, in higher (stationary phase) or lower (rifampicin‐treated cells) DNA compaction states (Appendix Fig S5A), the ParB_F DNA binding pattern is not altered, exhibiting a similar profile of dips and peaks (Fig 5B). This indicates that the assembly of the partition complex is not perturbed by variation in DNA compaction level within the nucleoid.

To further demonstrate the impediment of ParB_F binding in promoter regions, we constructed a strain in which the locus A, carrying two promoters, an IHF and two RcsB binding sites, is replaced by a kanamycin resistance gene (Fig 5D). The measured ParB_F binding pattern remained highly comparable except at the locus A where the dip is absent. This result clearly indicates that site‐specific DNA binding proteins are the main factors for restricting locally ParB_F binding.

ParB molecules exchange rapidly between partition complexes

Single molecule in vivo localization experiments have shown that over 90% of ParB_F molecules are present at any time in the confined clusters (Sanchez et al, 2015). However, stochastic binding of most ParB_F on non‐specific DNA suggests that partition complexes are highly dynamic. To unravel ParB_F dynamics, we performed fluorescence recovery after photobleaching (FRAP) on two‐foci cells for measuring ParB_F dynamics between partition complexes. By laser‐bleaching only one focus, we could determine whether ParB_F dimers could exchange between clusters and measure the exchange kinetics. As ParB_F foci are mobile, we choose to partially bleach (~50%) the focus enabling immediate measurement of fluorescence recovery (Fig 6A and B). A few seconds after bleaching, the fluorescence intensity recovers while it decreases in the unbleached focus. This exchange is progressive and the intensity between the two foci equilibrated in ~80 s on average (between 50 and 120 s for most individual experiments). We estimate that, when exiting a cluster, each ParB_F dimer has the same probability to reach any of the two clusters. Therefore, the time of equilibration between the two foci corresponds to the exchange of all ParB_F. These results thus indicate that the partition complexes are dynamic structures with a rapid exchange of ParB_F molecules between clusters.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6. ParB dynamics between partition complexes

ParB_F exchange between foci was measured by FRAP and FLIM (fluorescence lifetime imaging microscopy) from two‐foci cell of DLT1215 carrying pJYB234.

1. Representative images of a photobleached cell during a FRAP experiment. The 488 nm laser was pulsed (Bleach) on one of the two foci at ˜2.4 s (black arrow). Red and blue arrows correspond to the bleached and unbleached focus, respectively. Time is indicated in seconds (upper right). The cell outline is drawn in red. Scale bar: 1 μm.

2. Quantification of ParB_F‐mVenus fluorescence intensity over time. The dynamics of fluorescence intensity is shown from averaging 18 independent measurements of the bleached (FRAP, red line) and unbleached (FLIM, blue line) foci. Foci fluorescence intensity in each experiment was normalized to the average intensity of each focus before photobleaching. The Appendix Fig S6 displays the three pre‐bleaching and the first post‐bleaching data points on an expanded scale. Natural bleaching during the course of the experiments (green curve) was estimated for each measurement by averaging the fluorescence intensity of 15 foci present in each field of view. Error bars correspond to standard deviation (mean ± SD).

Bacterial strains and plasmids

Escherichia coli and V. cholerae strains and plasmids are listed in Appendix Table S1. Cultures were essentially grown at 37°C with aeration in LB (Miller, 1972) containing thymine (10 μg ml⁻¹) and chloramphenicol (10 μg ml⁻¹) as appropriate. For microscopy and stability assays, cultures were grown at 30°C with aeration in MGC (M9 minimal medium supplemented with 0.4% glucose, 0.2% casamino acids, 1 mM MgSO₄, 0.1 mM CaCl₂, 1 μg ml⁻¹ thiamine, 20 μg ml⁻¹ leucine, and 40 μg ml⁻¹ thymine).

Strain DLT1471 was constructed in several steps. First, the Eco47III–ApaI sopOPAB DNA fragment from plasmid F was inserted into a rep^ts plasmid‐borne 933‐codon 3′ fragment of the lacZ gene (pJYB50). The resulting plasmid, pJYB52, was introduced into DLT1215 and proceeded for allele exchange by selecting double recombination events as described (integration–excission assay; Cornet et al, 1994), yielding to the following chromosome fusion lacZ’::PparAB_F::’lacZ in strain DLT1471. Strain DLT1472 expressing ParB_F from the chromosome fusion lacZ’::PparB_F::’lacZ was described previously (Bouet et al, 2005).

Strains DLT2073 and DLT2075 are DLT1215 and DLT1472 derivatives, respectively, in which a 538‐bp DNA fragment carrying the entire parS_F site has been introduced at the EcoRV restriction site of the xylE gene (91 min on the E. coli chromosome) in the forward orientation. Briefly, the PCR‐amplified xylE gene was introduced onto a pFC13 derivative (Cornet et al, 1994) using the HindIII and XhoI restrictions sites, leading to pJYB102. A parS_F DNA fragment, PCR‐amplified from pDAG114 using the following oligonucleotides SopC‐5′RV (5′‐TCCTTTGATATCGGCCAGAAAGCATAACTG‐3′) and SopC‐3′RV (5′‐GCCGATATCAGGAATTCATGGAATCGTAGTCTC‐3′), was introduced into the EcoRV restriction of the xylE locus on pJYB102, leading to pJYB103.1 and pJYB103.2 depending on the orientation of parS_F insertion. These plasmids were introduced into DLT1472 and subjected to the integration–excision procedure as above. Strains DLT2074 and DLT2076 are identical to DLT2073 and DLT2075, respectively, with parS_F inserted in xylE in the reverse orientation (parS_F‐rev). Insertions of parS_F on the E. coli chromosome at locations other than xylE were performed using lambda red recombineering and selecting for the FRT‐Kan^R‐FRT cassette amplified from pDK4 (Datsenko & Wanner, 2000). The hupA‐mcherry allele used for live imaging of the nucleoid was inserted in various E. coli strains by P1 transduction from strain DLT3053 (Le Gall et al, 2016).

The original plasmid F, F1‐10 (gift from C. Lesterlin), was converted to ccdB⁻ to allow for performing stability assays. The removal of the CcdB toxin from the addiction system was performed along with the introduction of the chloramphenicol resistant gene, using lambda red recombineering (Datsenko & Wanner, 2000), by inserting the corresponding loci from the mini‐F pDAG114 (Lemonnier et al, 2000) into F1‐10, leading to F1‐10B. When necessary, the excision of the FRT‐kan‐FRT selection cassette was performed using the pCP20 plasmid (Datsenko & Wanner, 2000). F1‐10B ΔAB and F1‐10B‐BmV derivatives were constructed by lambda red recombineering using plasmids pDAG209 and pJYB234, respectively, as substrates to generate the linear DNA fragment that include the cat gene for selecting the recombinants.

Mutations in parB_F were first introduced into pYAS6 by mutagenic primer‐directed replication using the Stratagene QuikChange kit and subsequently integrated into mini‐F derivatives or the pAM238 expression vector by PCR amplification followed by allelic replacement using appropriate restriction enzymes. The mVenus gene was constructed by introducing the monomeric A207K mutation in the venus‐Yfp gene (Sanchez et al, 2015). All plasmid constructs were verified by DNA sequencing (MWG). Mini‐F and F plasmids derivatives were introduced in strains by CaCl₂‐ or electro‐transformation and conjugation, respectively.

pJYB322 was constructed by introducing a 171‐bp DNA fragment carrying three consensus parS sites (5′‐TGTTTCACGTGAAACA‐3′), called 3x‐parS_chr, into the NheI and ClaI restriction sites of pJYB263, the ΔparA derivative of pJYB234 (Le Gall et al, 2016).

Plasmid stability assays

Stability of mini‐F and plasmid F derivatives was assayed in strain DLT1215 grown over 25 generations in MGC at 30°C or over 20 generations in LB at 37°C, and subsequently plated on LB agar medium, replica plating to medium with chloramphenicol, and calculating loss rates from the fractions of each sample resistant to chloramphenicol, as previously described (Sanchez et al, 2013).

Epifluorescence microscopy

Exponentially growing cultures were deposited on slides coated with a 1% agarose buffered solution and imaged as previously described (Diaz et al, 2015), using an Eclipse TI‐E/B wide field epifluorescence microscope. Snapshots were taken using a phase contrast objective (CFI Plan Fluor DLL 100X oil NA1.3) and Semrock filters sets for YFP (Ex: 500BP24; DM: 520; Em: 542BP27) and Cy3 (Ex: 531BP40; DM: 562; Em: 593BP40) with an exposure time range of 0.1–0.5 s. Nis‐Elements AR software (Nikon) was used for image capture and editing.

ChIP‐sequencing assay and analysis

ChIP‐seq was performed as previously described (Diaz et al, 2017) with minor modifications, using polyclonal antibodies raised against WT ParB_F or his‐tagged ParB_Vc1. ParB_F and ParB_Vc‐1 antibodies were affinity‐purified from anti‐ParB_F (our own) and anti‐ParB_Vc‐1 (gift donated from the lab of D. Chattoraj) polyclonal serums, using purified ParB_F and ParB_Vc‐1‐his₆, respectively. The optimization step for determining the amount of antibodies needed to pull down all ParB_Vc1 in the IP samples was fully described (Diaz et al, 2017).

A maximum of 2 ng of ChIP DNA was used to synthetize the library as described in manufacturer's instructions (Ion ChIP‐Seq Library Preparation on the Ion Proton™ System—revision B—Step10). The last size selection was modified by a double size selection (0.55×/0.25×) of binding to AMPure^® XP beads followed by a step of wash and elution. After qualification and quantification (Bioanalyzer—Agilent, Santa Clara, CA), libraries were diluted to 100 pM and pooled in a ratio of 20% input and 80% IP. For all the subsequent analyses, the measured size distribution of the DNA fragments was subtracted of the linker size added for sequencing (93 bp). Template preparation (clonal library amplification on sequencing bead) was made using the Ion PI™ Template OT2 200 Kit version 3 following the manufacturer's instructions. Emulsion PCR was performed in the Ion OneTouch™ 2 Instrument. DNA‐positive ISPs were then recovered and enriched in the Ion OneTouch™ ES according to standard protocols. Sequencing of samples was conducted on Ion Proton and PI chips according to the Ion PI™ 200 Sequencing Kit Protocol (version 3).

The sequence reads were counted and aligned as described (Diaz et al, 2017). The quality of reads was assessed with FastQC program (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and they were mapped with TMAP from Torrent Suite software 5.0.4 (https://www.thermofisher.com/fr/fr/home/life-science/sequencing/next-generation-sequencing/ion-torrent-next-generation-sequencing-workflow/ion-torrent-next-generation-sequencing-data-analysis-workflow/ion-torrent-suite-software.html). The counting of reads over the genomic sequences was performed using bedtools 2.26.0 (http://bedtools.readthedocs.io/en/latest/) with the tool genomecov and the option –d and –fs or –bga for the bedgraph files production for peak visualization. This allows to display the totality of ChIP‐seq data with the possibility to view several datasets simultaneously using IGV (Integrated Genome Viewer, version 2.3; http://software.broadinstitute.org/software/igv/). Graphing the DNA portion of interest from ChIP‐seq data was done using Excel or R softwares. Cognate input and IP samples were normalized by the number of total reads for direct comparison. For the ParB density plots, the data were normalized after background subtraction and set to the value of 1 at the last bp of the 10^th repeat of parS_F, allowing to display the results of Monte Carlo simulations on the same graph.

Fit of the parameters

The “Nucleation and caging” model contains only three parameters: σ, a, and N_t. One parameter is known experimentally: σ ~ 75 nm (Lim et al, 2014; Sanchez et al, 2015). We know the order of magnitude of N_t ~ 300 (Bouet et al, 2005) as a benchmark of the fitted value obtained from κ from ChIP‐seq data. The value of σ²/a is fitted from ChIP‐seq, which gives access to the value of a (σ being known). We use the following trial function to perform a non‐linear fit of the ChIP‐seq data between 0 and 10 Kbp: where A₀ and A₁ are the two fit parameters allowing to obtained the two free parameters of the model: κ (related to N_t) and a. Note that the length unit to analyze the ChIP‐seq data is the base pair (bp). We identify A₀ = κ and A₁ = a/3σ². For the F plasmid, we find A₀ = 0.41 and A₁ = 1.96 × 10⁻⁴ nm⁻¹. From A₁, using σ = 75 nm, we get a ~ 3.4 nm (~10 bp). From A₀, we get κ ~ 0.41 leading to N_t ~ 360. For the chromosome, using the same analysis, we get A₀ = κ ~ 0.41 leading to a smaller number of proteins on the DNA, N_t ~ 120, yet of the same order of magnitude as the F plasmid. The second variable of the fit A₁ = 4.33 × 10⁻⁴ nm⁻¹ leads to a ~ 7.5 nm (~22 bp). For the estimation of the parameters for V. cholerae, the experimental values for σ and a are not available, and κ (or N_t) could not be read directly from the ChIP‐seq data because of the fragment size library and the presence of barrier (rRNA operon) that impede the ParB binding signal at large genomic distance. Thus, given the order of magnitude observed in E. coli (between 10 and 20 bp for the F plasmid and the chromosome, respectively), it is reasonable to assume a same range for V. cholerae, with a = 16 bp corresponding to the footprint of a ParB. In order to conserve the height and peaks of the ChIP‐seq data, we take σ = 25 ± 5 nm and κ = 0.15 ± 0.05. The probability to form a foci is chosen to be P_parS1 = 1, P_parS2 = 0.9, and P_parS3 = 0.6 in order to match the height of each peak observed in ChIP‐seq. These simulations are only semi‐quantitative, in order to show on very general physical ground that the “Nucleation and caging” is able to explain the long range decay observed in ChIP‐seq.

Monte Carlo simulations

The Monte Carlo procedures used an explicit polymer modeled by a Freely Jointed Chain (FJC). The polymer is in contact with a ParB reservoir. To reproduce the confinement of ParB around parS, the particles are not simulated explicitly. Instead, we modeled ParB binding to the polymer with a probability decreasing as a Gaussian function of the locus distance from parS. The Monte Carlo procedures were performed using the following scheme:

1. Build a Freely Jointed Chain (FJC) of N monomers of size a.

2. Define a particular site on the FJC as parS (or potentially many parS for V. cholerae) and define a Gaussian distribution of particles .

3. For each monomer label by an index i, choose a random number ran. If ran < C(r), then a particle is attached to the site i.

4. Start again at step 1 until the statistics is good enough.

This fast procedure allows to get very good statistics. In Figs 1C–E and 2A, and Appendix Fig S4B, we have analytical expressions for P_NC(s), which serve as a benchmark for simulations at the bp resolution. In Fig 4C (xylE::parS_F), Appendix Fig S1E (plasmid F), and Appendix Fig S4B (V. cholerae) where we included the average DNA fragments size in the ChIP libraries—as well as three parS sequences for V. cholerae—the analytical expressions were not obvious, we therefore used simulations.

Modeling the ChIP‐seq data with the integration of the average fragments size of the DNA library

The ChIP‐seq assay is based on the sequencing of sonicated fragments (of size ~250 bp) with an intrinsic unknown on the precise location of ParB (footprint of 16 bp). Here, the bioinformatic convention to build the ParB profile is to count +1 read at each bp of the fragments with at least one detected bound ParB. From the simulation perspective, this leads to two averages: (i) over the bound ParB positions, and (ii) over all fragment positions. When a bound ParB is detected, the average over the fragments positions consists in adding to the simulated profile a triangular function centered at the actual ParB position (where it takes the value of 1) and decreasing linearly down to 0 at a genomic distance ± the fragment size.

Correlation analysis

The correlation analyses were performed using the following formula:

Western immunoblotting

The determination of ParB_F relative intracellular concentrations and antibody purifications was performed as described (Diaz et al, 2015). When indicated, samples were diluted in DLT1215 extract to keep constant the total amount of proteins.

EMSA and proteins purification

Electro‐mobility shift assay was performed as described (Bouet et al, 2007) in the presence of sonicated salmon sperm DNA as competitor (100 mg ml⁻¹), using 1 nM radiolabeled 144‐bp DNA probe containing a single parS_F site generated by PCR. ParB_F and ParB_F‐3R* proteins were purified using an intein strategy as previously described from plasmids pYAS6 and pYAS25, respectively (Ah‐Seng et al, 2009). ParB_Vc‐1‐his₆ was affinity‐purified in a single step using a 10–1,000 nM imidazole gradient from strain DLT3431 (Castaing et al, 2008).

ParB_F and ParB_F‐3R* proteins were purified as previously described (Ah‐Seng et al, 2009).

FRAP and FLIM assays

Escherichia coli cells (Stellar) carrying pJYB213 (ParBF‐eGfp), grown in mid‐exponential phase in MGC medium and spotted on microscope slides coated with 1% MGC‐buffered agarose, were subjected to laser‐bleaching. Each field was imaged three times (pre‐bleached step) before photobleaching (at ~2.4 s) a single ParB_F focus with a 488 nm laser into two‐foci cells. ROI (region of interest) were 0.2 × 0.2 or 0.3 × 0.3 corresponding to 5 or 9 pixels, respectively. The laser power was set between 67 and 74 Hz to ensure partial bleaching, thus enabling to follow fluorescence recovery on a time scale of second right after photobleaching. Images were taken using an EMCCD camera with a 0.13 μm per pixel resolution (Hamamatsu). To follow recovery dynamics, images were taken every 5, 10, and 20 s up to 38, 108, and 169 s, respectively. Overall, photobleaching in the field of view during the time course of each FRAP experiment (green curves in Fig 6B and Appendix Fig S6A) was averaged from 15 unbleached foci from each field. Normalization to 1 was performed by averaging the focus fluorescence intensity from the three pre‐bleached images.