Specific and non-specific interactions of ParB with DNA: implications for chromosome segregation

The segregation of many bacterial chromosomes is dependent on the interactions of ParB proteins with centromere-like DNA sequences called parS that are located close to the origin of replication. In this work, we have investigated the binding of Bacillus subtilis ParB to DNA in vitro using a variety of biochemical and biophysical techniques. We observe tight and specific binding of a ParB homodimer to the parS sequence. Binding of ParB to non-specific DNA is more complex and displays apparent positive co-operativity that is associated with the formation of larger, poorly defined, nucleoprotein complexes. Experiments with magnetic tweezers demonstrate that non-specific binding leads to DNA condensation that is reversible by protein unbinding or force. The condensed DNA structure is not well ordered and we infer that it is formed by many looping interactions between neighbouring DNA segments. Consistent with this view, ParB is also able to stabilize writhe in single supercoiled DNA molecules and to bridge segments from two different DNA molecules in trans. The experiments provide no evidence for the promotion of non-specific DNA binding and/or condensation events by the presence of parS sequences. The implications of these observations for chromosome segregation are discussed.


ParB-DNA association kinetics
For assays monitoring the kinetics of ParB binding to Cy3-labelled DNA via changes in fluorescence, ParB (at double the concentration indicated) was prepared in a buffer containing 50 mM Hepes•KOH (pH 7.5), 100 mM KCl, 2.5 mM MgCl 2 , 0.1 mg/mL BSA and 1 mM DTT. A 50 µl aliquot of this sample was then rapidly mixed in a stopped-flow apparatus (TgK Scientific) with an equal volume of 40 nM 147 bp Cy3-labelled DNA substrate (see Supplementary Table 1), which was prepared in a buffer identical to the one described above. This yielded a final solution with 20 nM 147 bp Cy3-labelled DNA and the indicated concentration of ParB. The sample was excited at 549 nm and the emission recorded with a bandpass filter of >570 nm over time. Three repeats were performed for each concentration of ParB and the mean values plotted.

Plasmid topology assay
To determine if binding of ParB to DNA had a significant effect on DNA twist or writhe a topoisomerase I-linked assay was performed. The construct of plasmid pSP73 containing a parS site was relaxed by the addition of wheat germ Topoisomerase I (Inspiralis Ltd). Relaxed plasmid was then incubated for a further 30 min at 37 °C with 2 U wheat germ topo I and a titration of ParB. The reaction was stopped by the addition of 0.1% SDS and the proteins removed by phenol:chloroform extraction before the samples were run on a 1% agarose gel at 80 V in TBE buffer for 3 hrs. The DNA was then visualised by ethidium bromide staining, and the samples inspected for a ParB-dependent change in plasmid topology.

Calculation of the duration and extent of condensation
The final extension (Z end ) ( Figure 6C) was the mean extension in a time window of two seconds at the end of a condensation event. We considered that a condensation event finished when there are no visible changes in extension for several tens of seconds. The condensation time ( Figure 6B) was determined as the time to reach an extension given by Z end -2 σ end , where σ end is the standard deviation of the extension at the end of the condensation process. This criterion eliminated small condensation events occurring after long periods of time that would otherwise skew the time distribution to longer values.

Condensation force
To determine the force at which the process of condensation starts (Figure 7C), the magnet was alternated between positions that apply a large stretching force (3.9 pN) or a variable smaller force that was gradually decreased. The low force was maintained for 35 s and the mean extension for the first (E 0 ) and the last five seconds (E f ) of the measurement determined. The condensation force is defined as the critical force that satisfies the following criterion: where σ 0 is the standard deviation of E 0 at 1 Hz bandwidth . This threshold value (10σ 0 ) allowed us to discriminate between real condensation events and oscillations that are sometimes visible at forces near condensation.

Force-extension curves
For each DNA molecule, a value of the extension at a given force was calculated ( Figure 7D). Then, a mean value and standard deviation were determined for multiple DNA molecules. In the absence of protein, data were fitted to the worm-like chain model (1). From the fits, we obtained the contour length (L 0 ) of the molecule, which was subsequently used to normalize the data, and the persistence length (P). We measured a mean L 0 and P of 2.2 µm and 33 nm. The value of P was smaller than previously reported (2). This could be due to the presence of Mg 2+ in the buffer (3) and/or to the variability in the attachment positions of the DNA to the magnetic bead in parallel measurements. The molecule previously characterized in the absence of protein was then studied in the presence of 1 μM ParB 2 . ParB force extension data was calculated as above but was not fitted to any model. The solid red line in Figure 7D is simply included as a guide for the eye.

Freely-Orbiting MT (FOMT) experiments
We used a magnet arrangement similar to that described in (4). In this setup, superparamagnetic beads are free to rotate in a plane perpendicular to the DNA axis, thus providing a measurement of the torque induced by the binding of proteins to torsionally constrained DNA molecules. Thus, prior to any attempt to measure a signal induced by the binding of a protein in FOMT, we checked that the DNA molecules were torsionally constrained using the standard MT described in the Materials and Methods section. As in the standard MT, this setup also provides the extension of the DNA in real time. As a control experiment, we reproduced the assay reported in (4) with RecA protein using the buffer 20 mM Hepes pH 6.0, 50 mM NaCl, 1 mM MgCl 2 , 0.1 mM ATPgS, 1 mM DTT, 100 µg/ml BSA, 0.1% TWEEN 20.   Reverse  TCT TTG TTT CAC GTG GAA CAT TCT ACA GGT  TCT TTT TGA TTC AAC TGC TG   100 bp parS  Forward  5ꞌ-Hex  TGT TCG CCA ACT TGA GCA GCT GAT TCA GCA  GTT GAA TCA GAA TGT TCC ACG TGA AAC  AAA GAA AAA AGA ACC TGT GAA AGA TGC   GGT TCT AAA AGA ACG G   100 bp parS  Reverse  CCG TTC TTT TAG AAC CGC ATC TTT CAC AGG  TTC TTT TTT CTT TGT TTC ACG TGG AAC ATT  CTG ATT CAA CTG CTG AAT CAG CTG CTC AAG  TTG GCG         Decondensation by addition of competitor DNA randomly disrupts ParB-DNA interactions such that (E) decondensation only occurs when a critical interaction at the edge of a condensed loop (purple circle) is released by chance, and a much bigger step is likely to be observed.