Defining the role of ATP hydrolysis in mitotic segregation of bacterial plasmids

Abstract

Hydrolysis of ATP by partition ATPases, although considered a key step in the segregation mechanism that assures stable inheritance of plasmids, is intrinsically very weak. The cognate centromere-binding protein (CBP), together with DNA, stimulates the ATPase to hydrolyse ATP and to undertake the relocation that incites plasmid movement, apparently confirming the need for hydrolysis in partition. However, ATP-binding alone changes ATPase conformation and properties, making it difficult to rigorously distinguish the substrate and cofactor roles of ATP in vivo. We had shown that mutation of arginines R36 and R42 in the F plasmid CBP, SopB, reduces stimulation of SopA-catalyzed ATP hydrolysis without changing SopA-SopB affinity, suggesting the role of hydrolysis could be analyzed using SopA with normal conformational responses to ATP. Here, we report that strongly reducing SopB-mediated stimulation of ATP hydrolysis results in only slight destabilization of mini-F, although the instability, as well as an increase in mini-F clustering, is proportional to the ATPase deficit. Unexpectedly, the reduced stimulation also increased the frequency of SopA relocation over the nucleoid. The increase was due to drastic shortening of the period spent by SopA at nucleoid ends; average speed of migration per se was unchanged. Reduced ATP hydrolysis was also associated with pronounced deviations in positioning of mini-F, though time-averaged positions changed only modestly. Thus, by specifically targeting SopB-stimulated ATP hydrolysis our study reveals that even at levels of ATPase which reduce the efficiency of splitting clusters and the constancy of plasmid positioning, SopB still activates SopA mobility and plasmid positioning, and sustains near wild type levels of plasmid stability.

Figure 1. Effect of SopB R36 and R42 mutations on mini-F stability.

A. sopAB operon and sopC in natural configuration (cis; see sketches): white bars - strain DLT1900 carrying pDAG218 and derivatives, grown at 30°C in minimal-glycerol-leucine medium (t∼160 min); shaded bars - strain DH10B carrying pDAG173 and derivatives, grown in minimal-glycerol-Casamino acids medium (t∼80 min). Error bars are standard deviations of ≥3 loss-rate assays. Note 10-fold scale change for sopC ⁺ plasmids. B. sopAB and sopC on separate plasmids, pAM138 and mini-F respectively (trans), in strain MC1061 grown in minimal-glycerol-leucine medium: white bars – loss rates. Shaded bars show ATPase deficits as the inverse of ATPase stimulation by each SopB in vitro, determined previously .

Figure 2. Effect of R36 and R42 mutations on mini-F clustering.

Derivatives of strain DLT2583 (para-tetR::gfp) carrying the tetO-array mini-F, pDAG848, and producing SopA and mutant SopBs in trans were grown in MGC medium (A) or LB broth (B), and the mini-F plasmid foci were visualized as described in Materials and Methods and counted. Total cells scored were: A - 756 (wt), 626 (R36A) 415 (R36A, R42A); 344 (ΔsopA); 368 (ΔsopA, R36A); B - 2118 (wt), 3966 (R36A), 1874 (R36K), 1585 (R42A), 1736 (R42K). The wt and R36A focus number distributions in panel A were shown by the Student t-test to be different at a confidence interval of 95%. Sop protein concentrations relative to wt mini-F were measured by immuno-blot assay to be 1.9- and 2.2-fold (SopA) and 1.7 and 2.4-fold (SopB) higher in cells grown in MGC and LB respectively. C. Brightness of the foci in LB-grown cells, measured using Microbe Tracker. The distributions are different at the 95% confidence level by the Student t-test. Because focus intensity responds non-linearly to plasmid content (as seen elsewhere; e.g. [51]), foci in R36A cells were on average 1.3-fold brighter than in SopB-wt cells while consisting of 2.4 times as many molecules. D. Localization of pDAG848 without SopA. Cells were grown in MGC. The shortest pole-focus distances in single-focus cells were sorted into 1/10 cell-length classes. Dotted-line bars show distributions of single foci in the presence of SopA (Figure 5B). Distributions of focus brightness in ΔsopA cells containing wt and R36A SopB proteins were not significantly different at the 95% confidence level, by the same test as applied to the brightness data of Figure 2C (not shown).

Figure 3. Effect of R36 and R42 mutations on SopA movement.

A. Time-lapse image series of typical DLT2687 (plac-sopA::xfp)/pDAG848 (mini-F)) cells carrying pYAS47 (sopA⁺B⁺) or pYAS64 (sopA⁺B ^R36A) grown in MGC-IPTG (0.1 µM) medium and applied to MGC-IPTG agarose slides. Images were acquired with an Olympus microscope at 70-second intervals for 29 minutes using 0.5 sec exposure. B. Kymograph of the SopA::Xfp foci in A. Each point represents the position of the brightest pixel in the image. C. Distribution of oscillation rates in cells producing the wt and the four mutant SopBs. Rate data are binned into 0.5 oscillation/30 min intervals. The levels of SopA::Xfp and SopA were, respectively, about 3-fold and 0.3-fold the wt mini-F levels (Figure S1C); repression of the plasmid-borne plac promoter is stronger than that of the chromosomal plac in the conditions used.

Figure 4. Effect of SopB-R36A on SopA relocation at high resolution.

SopA::Xfp movement in DLT2740 (plac-sopA::xfp)/pDAG848 (mini-F)) cells carrying pYAS47 (sopA⁺B⁺) or pYAS64 (sopA⁺B ^R36A) grown in MGC with 1 µM IPTG was monitored as in Figure 3 but with a Nikon microscope at 20 sec intervals for 20 min using 100 msec exposure. To facilitate comparison, kymograph widths representing cell length have been equalized and brightness has been adjusted. The kymograph at bottom has been expanded to show differentiation of SopA “wait” (w) and “migrate” (m) periods; wait is defined as an exposure in which >80% of the pixels are on one side of the mid-line, migrate as the interval between two waits.

Figure 5. Limited effect of R36A mutation on mini-F positioning.

The distance between the pole nearest a focus and the foci in MGC-grown cells of the one- and two-focus classes was measured. A: Pole-focus distances plotted against cell-length. The wt and R36A data-sets are superimposed. Blue symbols denote the four cases (out of 107, all R36A) in which both foci were unambiguously in the same cell-half. B: Pole-focus distances sorted into 1/10-cell intervals. In the right-hand panel distances to foci beyond the mid-point are measured from the same pole as the closer foci.

Figure 6. Effect of R36A mutation on mini-F movement.

A. Cumulative displacement of mini-F, from images acquired every 5 secs for 5 mins, drawn using the Z project plug-in of Image J. B. Average cumulative displacement in 30 cells as illustrated in A; error bars are standard deviations. C. Illustrative individual kymographs of 1-focus cells; width corresponds to cell length (not equalized). D. As for C, 2-focus cells.

Figure 7. Steps in partition at which SopB stimulates SopA to hydrolyse ATP.

Mini-F plasmids are shown oval with SopB bound to sopC in the partition complex as dark green (wt) or yellow (R36A) and SopB spread in cis or onto the nucleoid (grey) in the corresponding lighter shade. No distinction is made here between direct binding of SopB to the nucleoid and binding via SopA tethers; the latter option is favoured by recent observations of partition complex movement as a function of SopA-ATP disassociation from a DNA carpet . Distinct functional forms of SopA are shown in the key as: SopA*-ATP, able to bind to non-specific DNA (corresponding to ParA*₂-ATP₂; [23]); SopA^S-ATP, modified by SopB to be unable to bind non-specific DNA ; SopA, without ATP following hydrolysis; SopA-ATP, associated with ATP but not having undergone transition to the DNA-binding form . A. Model of SopB-stimulated splitting of clusters. SopB spreads from clustered partition complexes, creating a mat which attracts SopA-ATP and shelters it from the nucleoid to allow SopA polymerization . For clarity only one polymer (chain of red dots) and its interactions is shown, at the left of the SopB mat. Polymer contact (horizontal arrow) with SopB-sopC initiates a hydrolysis-driven transfer of each partition complex to successive polymer subunits which generates the force to break inter-plasmid links (shown as hooks). Hydrolysis stimulation by SopB-R36A is too low to drive this process efficiently. Following splitting, the SopB mat shrinks, attenuating polymerization and this mode of plasmid separation. The separation mode that positions plasmid molecules (e.g. diffusion-ratchet; , ; shown at right of SopB mats) is mostly independent of ATP hydrolysis and is activated by both wt and R36A SopB complexes. B. Model of SopA priming for relocation. Relatively slow conformational transitions must raise the level of SopA-ATP competent to start relocation from the pole. Most of the SopA-ATP dislodged by the R36A partition complex does not hydrolyse its ATP, does not need the transition, and so reverses direction without delay. C. Model of restricted F plasmid movement. The extended partition complex of a segregated plasmid encounters nucleoid-bound or relocating SopA molecules at random and moves as dictated by local SopA dynamics. Wt SopB complexes stimulate ATP hydrolysis, disrupting SopA-SopB contact and curtailing movement, whereas R36A complexes do not, leading to a wider ambit (longer reverse arrow), accounting for the exaggerated position adjustments seen in Figure 6.