Promiscuous Stimulation of ParF Protein Polymerization by Heterogeneous Centromere Binding Factors

J. Mol. Biol. (2007) 374, 1–8

doi:10.1016/j.jmb.2007.09.025

COMMUNICATION

Promiscuous Stimulation of ParF Protein Polymerization by Heterogeneous Centromere Binding Factors Cristina Machón 1,2 , Timothy J. G. Fothergill 1 , Daniela Barillà 3 and Finbarr Hayes 1,2 ⁎ 1

Faculty of Life Sciences, The University of Manchester, 131 Princess Street, Manchester M1 7DN, UK 2

Manchester Interdisciplinary Biocentre, The University of Manchester, 131 Princess Street, Manchester M1 7DN, UK 3

Department of Biology, University of York, PO Box 373, York YO10 5YW, UK Received 21 June 2007; received in revised form 5 September 2007; accepted 7 September 2007 Available online 14 September 2007

The segrosome is the nucleoprotein complex that mediates accurate segregation of bacterial plasmids. The segrosome of plasmid TP228 comprises ParF and ParG proteins that assemble on the parH centromere. ParF, which exemplifies one clade of the ubiquitous ParA superfamily of segregation proteins, polymerizes extensively in response to ATP binding. Polymerization is modulated by the ParG centromere binding factor (CBF). The segrosomes of plasmids pTAR, pVT745 and pB171 include ParA homologues of the ParF subgroup, as well as diverse homodimeric CBFs with no primary sequence similarity to ParG, or each other. Centromere binding by these analogues is largely specific. Here, we establish that the ParF homologues of pTAR and pB171 filament modestly with ATP, and that nucleotide hydrolysis is not required for this polymerization, which is more prodigious when the cognate CBF is also present. By contrast, the ParF homologue of plasmid pVT745 did not respond appreciably to ATP alone, but polymerized extensively in the presence of both its cognate CBF and ATP. The co-factors also stimulated nucleotide-independent polymerization of cognate ParF proteins. Moreover, apart from the CBF of pTAR, the disparate ParG analogues promoted polymerization of non-cognate ParF proteins suggesting that filamentation of the ParF proteins is enhanced by a common mechanism. Like ParG, the co-factors may be modular, possessing a centromere-specific interaction domain linked to a flexible region containing determinants that promiscuously stimulate ParF polymerization. The CBFs appear to function as bacterial analogues of formins, microtubule-associated proteins or related ancillary factors that regulate eucaryotic cytoskeletal dynamics. © 2007 Elsevier Ltd. All rights reserved.

Edited by I. B. Holland

Keywords: Plasmid; Segregation; ParA; ParF; ParG

*Corresponding author. E-mail address: [email protected]. Present addresses: C. Machón, Centre for Biomolecular Sciences, University of Nottingham, Nottingham NG7 2RD, UK; T. J. G. Fothergill, Department of Biochemistry and Molecular Biology, University of Texas M. D. Anderson Cancer Center, Houston, TX 77030, USA. Abbreviations used: CBF, centromere binding factor; DLS, dynamic light-scattering; kct/s, kilocounts/second; ATPγS, adenosine-5-O-(3-thiotriphosphate).

Cytoskeletal structures in eucaryotes comprise actin-containing microfilaments, microtubules and intermediate filaments that direct a variety of fundamental cellular processes, including chromosome segregation, cytokinesis, cell motility and morphology, as well as trafficking of diverse subcellular cargo.1 Cytoskeletal protein dynamics are modulated by a plethora of accessory factors. For example, profilin augments actin polymerization and depolymerization, whereas formins mediate

0022-2836/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.

2 actin nucleation and formation of linear actin filaments.2,3 Microtubule-associated proteins (MAPs) and other polypeptides influence microtubule assembly, stability and localization. 4 Crosstalk between intermediate filaments and other cytoskeletal elements and organelles are coordinated by intermediate filament-associated proteins.5 Bacteria possess ancestral homologues of actin, tubulin and intermediate filaments, as well as prokaryoticspecific polymerizing proteins of the ParA superfamily. Bacterial polymerizing proteins are implicated in plasmid and chromosome segregation, cell division, and determination of cell shape and polarity.6 Although understanding of the assembly and function of prokaryotic cytoskeletal structures is much more rudimentary compared with their eucaryotic counterparts, accessory proteins are known to modulate the properties of bacterial polymerizing proteins in certain cases. For example, filamentation and nucleotide hydrolysis by the FtsZ septal ring protein are modified by numerous protein cofactors.7 The molecular mechanisms of bacterial DNA segregation are best described for low copy number plasmids.8–10 The segrosome of multidrug-resistance plasmid TP228 consists of ParF and ParG proteins that assemble at the parH centromere. ParF (22.0 kDa) defines one clade of the ParA family of ATPases that are widely involved in plasmid and chromosome segregation.11 ATP binding promotes ParF polymerization into extensive multistranded filaments.11 Although nucleotide hydrolysis is not required for filamentation, ParF is a weak ATPase whose catalytic activity is stimulated ∼ 30-fold by the ParG centromere binding factor (CBF). Furthermore, ParG (7.6 kDa) associates with ParF polymers and modulates ParF polymerization kinetics.12 ParG is dimeric,13 with C-terminal regions that interlock into a ribbon-helix-helix fold, and flexible N-terminal tails. 14 The C-terminal regions are necessary for dimerization, DNA binding, and interaction with ParF.14–16 The mobile tails are also multifunctional. First, the tail modulates interaction of ParG with its operator site during transcriptional autorepression of the parFG genes.15 Second, the N-terminal tail of ParG harbours an arginine fingerlike motif that stimulates nucleotide hydrolysis by ParF,16 analogous to arginine fingers in proteins such as Ras-GAPs.17,18 Stimulation of nucleotide hydrolysis by ParG is likely to be a crucial aspect of the ParF polymerization-depolymerization cycle.16 Third, and distinct from its role in ATPase stimulation, the ParG tail is necessary for enhancement of ParF polymerization.16 ParG may either bundle filaments more extensively, or stabilize ParF protomers within filaments.12,16 The flexible tails within each ParG dimer potentially enwrap adjacent ParF monomers, either on the same or parallel filaments, or might act at points of polymer disassembly. Thus, ParG may be functionally analogous to formins and other factors that modulate growth and retraction of eucaryotic actin filaments, or microtubule-associated proteins that regulate tubulin kinetics. Although the

Stimulation of ParF Protein Polymerization

molecular basis of segregation by ParF-ParG proteins is yet to be fully elucidated, pushing of plasmids by ParF polymer growth or plasmid pulling by filament depolymerization may direct replicated plasmids to either side of the septal plane.12 Certain segrosomes involving close homologues of ParF include heterogeneous dimeric proteins that are phylogenetically unrelated to ParG, or each other (Supplementary Data Figure S1).11,19–21 These disparate analogues recognize the cognate centromeres.20–22 Centromere binding is largely specific, and the ParG analogues do not cross-interact.21 Here, we have investigated the role of ParG and its analogues in stimulation of ParF polymerization: diverse CBFs enhance filamentation of both cognate and non-cognate ParF proteins, indicating that a common mechanism underlying stimulation of polymerization has evolved among heterogeneous CBFs. Nucleotide-induced polymerization of ParF homologues Plasmids pTAR, pVT745 and pB171 specify homologues of the ParF protein of plasmid TP228.11,21 However, the genes for these homologues are flanked by loci encoding diverse CBFs whose primary sequences are unrelated to ParG, or each other (Supplementary Data Figure S1). Although disparate, the CBFs are uniformly small (∼ 8– 10 kDa), dimeric proteins that bind their cognate centromeres, like ParG. Centromere recognition by the analogues is largely specific, as is protein dimerization.21 Thus, among the 12 combinations of relevant CBFs (ParG, ParB TAR, AA15 and ParBB171) tested with the corresponding non-cognate centromeres in electrophoretic mobility-shift assays, no binding was observed in 11 instances, with only the AA15 protein of pVT745 crossreacting with the pB171 centromere. These ParF homologues and accompanying ParG analogues provide an ideal suite of proteins to further probe ParF polymerization, and the role that unrelated CBFs exert in this process. Dynamic light-scattering (DLS), which assesses both particle abundance and size, was used to investigate polymerization of purified ParF homologues (Figure 1). Without nucleotide, both ParATAR and ParAB171 proteins remained stable for N60 min at 30 °C with an average intensity b100 kilocounts/ second (kct/s) and a particle size b 40 nm. In the presence of ATP and under conditions previously used to monitor ParF polymerization,12 count rates gradually rose to ∼ 300–600 kct/s (Figure 1(b) and (c), bottom), accompanied by particle size increases to N 100 nm (Figure 1(b) and (c), top). Thus, ATP modestly stimulates ParATAR and ParAB171 polymerization, supporting recent results for the latter.23 The poorly hydrolyzable adenosine-5-O-(3-thiotriphosphate) (ATPγS) analogue promoted polymerization of ParATAR and ParAB171 to an equivalent extent as ATP, demonstrating that nucleotide hydro-

Stimulation of ParF Protein Polymerization

3 lysis is not required for polymerization. Although ATP-induced polymerization of ParATAR and ParA B171 parallels previous observations with ParF12, filamentation of the homologues is more modest compared to ParF (Figure 1(a)). Furthermore, neither the surge in polymerization with ATPγS nor the gradual nucleotide-independent polymerization observed with ParF12 (Figure 1(a)) is apparent with either homologue. Moreover, the third homologue (ParAVT745) did not respond appreciably to either ATP or ATPγS (data not shown). The nucleotide-binding pockets of different ParF homologues may exhibit subtle conformational variations eliciting dissimilar polymerizing behaviour upon ATP or analogue binding. Alternatively, the polymer assembly interface may vary between ParF homologues resulting in different polymerization propensities or polymer stabilities. Cognate CBFs stimulate ParF homologue polymerization

Figure 1. ATP stimulates the assembly of ParF homologues into polymeric structures. Polymerization of ParF (a), ParATAR (b), and ParAB171 (c) was detected by DLS. The bottom panels illustrate the increase in light-scattering intensity upon nucleotide addition (0.5 mM; arrows). The top panels show the corresponding augmentation in polymer average size. Black, no nucleotide; green, ADP; blue, ATP; red, ATPγS. ParF, ParATAR, and ParAB171 were included at 4 μM, 10 μM, and 5 μM, respectively. Note the different vertical scales in the four panels. His-tagged proteins were overproduced in BL21(DE3) and purified by Ni2+ affinity chromatography essentially according to the Novagen technical manual. ParF protein polymerization was measured by DLS in a Malvern Zetasizer Nano System (He-Ne laser, 633 nm).12 Briefly, proteins were centrifuged for 30 min at 14,000 rpm and the supernatant was collected and used for experiments. A 47 μl sample of protein (in 30 mM Tris–HCl (pH 7.0), 100 mM KCl, 2 mM DTT, 10% (v/v) glycerol) were added to a 50 μl quartz cuvette and incubated at 30 °C in the Zetasizer chamber. ATP, ADP or ATPγS (500 μM) and 5 mM MgCl2 were added to a final volume of 50 μl. Values were obtained every 20 s. The intensity or count rate measures the amount of scattered light expressed as photons detected per second. Intensity is proportional to the size and concentration of the scattering particles.

The ParG protein exerts a profound effect on polymerization of ParF, promoting filamentation both in the presence and in the absence of ATP.12,16 As the ParF homologues under scrutiny here respond relatively weakly to nucleotide, the effect of ParG analogues on polymerization of cognate ParF proteins was tested. ParF homologues were incubated with nucleotides and cognate CBFs for 10 min at 30 °C, centrifuged into pellet and supernatant fractions (Sorvall S100-AT3 fixed angle rotor), and analyzed by SDS-PAGE (Figure 2(a)). If ParF proteins assemble into filaments of sufficiently high molecular mass, they will enter the pellet, whereas unpolymerized or partially polymerized protein will remain in the supernatant. ParATAR, ParAVT745 and ParAB171 entered the pellet fractions when incubated with ATP or ATPγS and cognate ParG proteins, revealing that diverse CBFs stimulate polymerization of the corresponding ParF proteins in the presence of nucleotide. These observations broadly parallel those with ParF-ParG,12,16 and confirm that polymerization stimulation is a unifying characteristic of disparate CBFs. The impact of CBFs on ParF polymerization was examined further in real time by DLS (Figure 3). ParF homologues were equilibrated with 0.5 mM nucleotide for 5 min, followed by the addition of cognate CBF. For ParATAR, inclusion of ParBTAR protein induced an immediate spike in light-scattering (N 50,000 kct/s) accompanied by formation of polymers N2000 nm in size (Figure 3(b)). Equivalent results were obtained with ParAVT745 and ParAB171 and their corresponding CBFs (Figure 3(c) and (d)). ATP and ATPγS generated similar responses in these experiments, except that ParATAR polymerization in the presence of ParBTAR was more pronounced with ATP than with ATPγS (Figure 3(b)), which mirrors the pelleting assay data (Figure 2). Substitution of a non-bridging oxygen atom of the γ-phosphoryl group with sulphur in ATPγS might induce steric alterations in the nucleotide-binding pocket, so that

4

Stimulation of ParF Protein Polymerization

Figure 2. (a) Enhancement of ParF homologue polymerization by cognate CBFs in the presence of nucleotide. ParATAR, ParBTAR, ParAVT745 , AA15, ParA B171 and ParBB171 were included at 6 μM, 6 μM, 8 μM, 40 μM, 10 μM and 20 μM, respectively. Sedimentation assays were performed as described.12 In brief, ParATAR, ParAVT745, and ParAB171 were incubated in 30 mM Tris–HCl (pH 8.0), 100 mM KCl, 2 mM DTT, 10% glycerol in 60 μl in the absence or in the presence of 2 mM nucleotides, 5 mM MgCl2, and cognate CBF, for 10 min at 30 °C. Reactions were centrifuged for 30 min at 4 °C at 80,000 rpm in a Sorvall S100-AT3 fixed angle rotor. Samples (20 μl) of supernatant were collected and the remaining supernatant was carefully aspirated. The pellet was resuspended in 15 μl of water. Protein levels in the supernatant (33%) and pellet (100%) were resolved by SDSPAGE and stained with Coomassie brilliant blue and quantified using ImageQuant TL software. The percentages of ParF homologues in pellets are shown. The results are averages of at least two experimental replicates. (b)–(e) Enhancement of ParF homologue polymerization by cognate CBFs in the absence of nucleotide. Different molar ratios of ParF proteins and cognate CBFs were incubated without added nucleotide, and the reactions were then centrifuged; 100% and 33%, respectively, of the pellet and supernatant fractions were resolved by SDS-PAGE (12% (w/v) polyacrylamide gel) and stained with Coomassie brilliant blue. The amounts of ParF homologues (open bars) and ParG analogues (filled bars) detected in the pellet fractions are shown. The data shown are from typical experiments of at least two replicates. ParF (b), ParATAR (c), ParAVT745 (d), and ParAB171 (e) were included at 8 μM, 6 μM, 8 μM and 10 μM, respectively. Due to limitations in protein volumes that could be added to reactions it was not feasible to test protein ratios N 1:10.

stimulation of ParATAR polymerization by ParBTAR is less proficient with ATPγS than with ATP. ADP does not stimulate ParF homologue polymerization (Figure 1). Accordingly, polymerization of these homologues in the presence of cognate CBF was less extensive with ADP than with triphosphate nucleotides (Figures 2(a) and 3). ParG promotes ParF polymerization by nucleotide-dependent and independent mechanisms.16 As the ParF homologues investigated here polymerize weakly with ATP, but avidly when cognate CBF is included, the effect of ParG analogues on filamentation of their corresponding ParF proteins was assessed without nucleotide (Figure 3, brown lines). In all cases, the analogues induced polymerization of cognate ParF homologues, albeit less extensively than when ATP was present. The impact of CBFs on ParF polymerization was examined further in sedimentation assays by incubating cognate ParF and ParG analogues together in ratios from 10:1 to 1:10 without nucleotide, and assessing amounts of the two species in the pellet and supernatant fractions (Figure 2). Levels of ParATAR in the pellet peaked at a ratio of 1:1 with ParBTAR. By

contrast with ParATAR, different pelleting patterns were observed with ParAVT745 and ParAB171 in the presence of increasing amounts of cognate CBF: ParF homologue levels in pellet fractions continued to rise as further ParG analogue was added to the reactions (Figure 2(b)–(e)). In parallel, the amounts of cosedimenting CBFs increased. In summary, polymerization of the three ParF homologues studied here is promoted by cognate ParG analogues to levels comparable, or greater than, the ParF filamentation triggered by ParG. Considering the relatively weak responses of these homologues to ATP alone, the CBFs potentially exert an even more crucial effect on the polymerization process than with the ParF-ParG pair. CBF specificity in enhancement of ParF polymerization As all four CBFs tested strongly enhanced the filamentation of cognate ParF homologues, the specificity of this stimulation was examined in DLS experiments (Figure 4). Both AA15 and ParBB171 enhanced polymerization of ParF of plasmid TP228, albeit to lesser

Stimulation of ParF Protein Polymerization

5

Figure 3. Enhancement of ParF homologue polymerization by cognate CBFs in the presence of ATP. Polymerization of (a) ParF, (b) ParATAR, (c) ParAVT745 and (d) ParAB171 (D) was detected by DLS. The bottom panels illustrate the increase in light-scattering intensity upon addition of nucleotide to 0.5 mM after 5 min (open arrows) and, 5 min later, upon addition of cognate CBF (filled arrows). The top panels show the corresponding augmentation in polymer average size. Black, without nucleotide or CBF; green, ADP + cognate CBF; blue, ATP + cognate CBF; red, ATPγS + cognate CBF; brown, cognate CBF only added at 5 min. Note the difference in vertical scale in (a) compared to other panels. ParF, ParATAR, ParAVT745, and ParAB171 were included at 4 μM, 10 μM, 10 μM and 5 μM, respectively, and cognate CBFs at equimolar concentrations. Experiments were performed as outlined in the legend to Figure 1. When included, ParG analogues were added at the 10 min time-point, i.e. 5 min after nucleotide, or were added to non-nucleotide reactions at the 5 min timepoint. In (c) and (d), the traces (brown) for ParAVT745 and ParAB171 with the cognate CBFs in the absence of ATP are significantly higher than the traces (black) for ParAVT745 and ParAB171 alone, but have been compacted to the bottoms of the graphs because of the high polymerization responses in the presence of nucleotides. ParAVT745 and ParAB171 with the cognate CBFs in the absence of ATP maximally reached an average intensity of ∼ 5000 and ∼ 2000 kcts/s, respectively, whereas both proteins showed b 100 kcts/without the cognate CBF or ATP (see Figure 1).

extents than the cognate ParG protein (∼10,000 kct/s for AA15 and ParBB171 compared to ∼20,000 kct/s for ParG). ParBTAR did not promote ParF polymerization above the levels elicited by ATP alone (Figure 4(a)). Strikingly, all three non-cognate ParG analogues stimulated filamentation of the ParATAR protein (Figure 4(b)). ParBB171 was as efficient as cognate ParBTAR protein in promoting ParATAR polymerization, with AA15 and ParBB171 also evoking strong responses, but less than that conferred by the cognate CBF. By contrast, ParAVT745, which polymerizes extensively when incubated with ATP and cognate AA15 protein (Figure 3(c)), reacted modestly with ParG, less strongly with ParBB171, and did

not polymerize detectably with ParBTAR (Figure 4(c)). Polymerization of the fourth ParF homologue tested, ParAB171, was not enhanced by ParBTAR but was stimulated modestly by both other non-cognate CBFs, ParG and AA15 (Figure 4(d)). In summary, although the CBFs are diverse, they exhibit remarkable versatility in polymerization stimulation, enhancing filamentation of non-cognate ParF proteins in nine of the 12 combinations tested. Interestingly, the three non-productive pairings involved the ParBTAR protein, which promoted polymerization of the cognate ParATAR protein only, suggesting that ParBTAR possesses highest specificity for its cognate ParF.

6

Stimulation of ParF Protein Polymerization

Figure 4. Enhancement of ParF homologue polymerization by non-cognate CBFs in the presence of ATP assessed by DLS. The increase in light-scattering intensity is shown upon addition of ATP to 0.5 mM after 5 min (open arrows) and, 5 min later, upon addition of non-cognate CBF (filled arrows). Blue, ATP only; red, ATP + ParG; black, ATP + ParBTAR; green, ATP + AA15; brown, ATP + ParBB171. The traces for each ParF homologue with cognate CBF are the same as those shown in Figure 3. Note the different vertical scales in the four panels. ParF, ParATAR, ParAVT745, and ParAB171 were included at 4 μM, 10 μM, 10 μM and 5 μM, respectively, and cognate CBFs at equimolar concentrations. Experiments were performed as outlined in the legend to Figure 1. When included, ParG analogues were added 5 min after the nucleotide.

In common with the ParG analogues under investigation here, YefM is a dimeric protein that binds DNA, is of similar size (9.3 kDa monomer), and possesses both structured and unstructured domains.24 However, YefM fulfils a different function as the antitoxin component of a toxin–antitoxin complex. Significantly, YefM failed to induce polymerization of all ParF homologues in the presence of ATP (data not shown). Similar results were obtained with bovine serum albumin. Thus, proteins that are functionally unrelated to CBFs fail to enhance ParF homologue polymerization, supporting that crossstimulation of polymerization by disparate CBFs (Figure 4) reflects bona fide interactions between these proteins and ParF homologues. Conclusions The ParF protein of plasmid TP228 polymerizes extensively in response to ATP, assembling into multistranded filaments with a distinctive morphology.12 Here, we demonstrate that ParF homologues encoded by plasmids pTAR and pB171 also respond to added nucleotide under identical experimental conditions, albeit more moderately than ParF. Although ParF polymerization is stimulated more strongly by ATPγS than by ATP, ParATAR and ParAB171 both showed only modest response differ-

ences with ATP and the non-hydrolyzable analogue. By contrast, the ParF homologue of plasmid pVT745 did not polymerize appreciably with either ATP or ATPγS. Although the Walker-type ATP-binding motifs in these homologues are highly-conserved,11 these gradations in polymerization responses among ParF proteins may reflect either variations in the conformations of the ATP-binding pockets, which affect nucleotide binding and/or stability, or differences in the monomer–monomer interface in the various homologues resulting in different polymerization behaviour. ParF also has an intrinsic tendency to polymerize without nucleotide which is repressed by ADP,12 but neither ParATAR, ParAVT745, nor ParAB171 exhibited this property. Thus, ParF may be primed for nucleotide-induced polymerization more than the other proteins tested. In addition to members of the ParF subgroup, other recent observations in vivo and/or in vitro have revealed that ATP-induced filamentation is a defining characteristic of ParA segregation proteins.25–29 Bidirectional plasmid pulling by filament depolymerization or plasmid pushing by ParA polymerization have been suggested as mechanisms by which ParA proteins may propel plasmids away from the septal plane before cytokinesis.8,12 Chromosome pulling in Vibrio cholerae by the retraction of ParA filaments has been described recently.28

Stimulation of ParF Protein Polymerization

The ParAB171 protein was shown recently to generate oscillating helical filaments that extend over the nucleoid in vivo, and to filaments in response to ATP in vitro.23,30,31 The results here confirm these in vitro observations although diverge on the response of ParAB171 to ATPγS: here, ATP and ATPγS elicited analogous filamentation responses (Figure 1(c)), whereas Ebersbach et al. found that ATPγS could compete with ATP for binding to ParAB171, but did not promote protein polymerization.23 This led to the proposal that ATP hydrolysis was required for ParAB171 polymerization. By contrast, ATP binding, but not hydrolysis, is sufficient and necessary for polymerization of ParF,11 as well as for ParATAR polymerization (Figure 1(b)). Although the reasons for inconsistencies in the effect of ATPγS on ParAB171 polymerization are unclear, the results presented here demonstrate that ParBB171 has a profound effect on ParAB171 polymerization. Previous observations have shown that oscillation of ParAB171 in vivo requires the ParBB171 protein,30,31 and that ParAB171 and ParBB171 interact in twohybrid assays.23 ParF and ParG of plasmid TP228 also associate in two-hybrid assays, 12 but the segregation proteins from TP228 and pB171 show no cross-interaction (and data not shown).21 ParF polymerization is modulated by the CBF, ParG.12,16 Analogously, polymerization of the ParF homologues studied here is promoted by the cognate ParG analogues to levels comparable, or greater than, with the ParF-ParG pair. Considering the comparatively weak responses of the ParF homologues to ATP alone, the CBFs arguably exert an even more crucial role in the polymerization process than with the ParF-ParG pair. The CBFs are phylogenetically heterogeneous, do not cross-interact, and demonstrate DNA binding specificity for the cognate centromeres.21 Disparate CBFs stimulate polymerization of the cognate ParF proteins by both nucleotide-dependent and independent processes, as noted with ParG.16 Moreover, although the CBFs are diverse, they exhibit remarkable elasticity in polymerization stimulation: the CBFs enhance polymerization of non-cognate ParF proteins in nine of 12 combinations tested in the presence of ATP. Interestingly, the three non-productive combinations involved the ParBTAR protein, which failed to promote polymerization of any ParF homologue tested, except for the cognate ParATAR protein. Although the mechanism by which the CBFs stimulate ParF polymerization remains to be fully elucidated, the flexible N-terminal tails in the ParG dimer are required for promoting polymerization of its cognate ParF.16 ParG does not appear to extend the length of ParF fibres, but the mobile tails instead might enable ParF nucleation and/or bridge ParF monomers within adjacent protofilaments thereby promoting the formation of thicker bundles. Intriguingly, circular dichroism analysis suggests that each of the CBFs described here possesses N 40% of unordered residues.21 In the absence of high-resolution information about the tertiary structures of the ParBTAR, AA15 and ParBB171 proteins, we speculate

7 that, like ParG, these homodimers all possess flexible regions that contact the ParF proteins and modulate their polymerization. Furthermore, considering the promiscuity displayed by the CBFs, we hypothesize that the putative mobile domains adopt analogous conformations, thereby fulfilling equivalent functions in polymerization stimulation. MinD is a member of the ParA superfamily that, together with MinC, prevents placement of the celldivision septum at random locations in Escherichia coli. Inhibition is relieved specifically at midcell by the MinE protein (10.2 kDa), thereby allowing cell division at the correct site.32 MinD polymerizes into filaments33 that are morphologically similar to ParF fibres,12 and whose dynamics are modulated by MinE.33 Thus, there are numerous intriguing parallels between the properties of the MinCDE complex, and ParF homologues and their CBFs. Similarities can also be drawn between CBFs and protein regulators of eucaryotic cytoskeletal elements. For example, formins are a diverse group of homodimeric proteins that nucleate and associate with the fast-growing barbed end of actin. Formins can exist in multiple isoforms that assemble actin filaments involved in distinct cellular processes. Although diverse and acting with different kinetics, formins are mechanistically similar.3 Thus, there are precedents for disparate co-factors acting on similar cytoskeletal elements. High-resolution information about the tertiary structures of CBFs additional to ParG, and of CBF-ParF complexes, will further illuminate the mechanism by which they commonly promote ParF polymerization and mediate DNA segregation.

Acknowledgements This work was supported by grants from The Wellcome Trust and Biotechnology & Biological Sciences Research Council to FH. DB is a Medical Research Council New Investigator. TJGF was supported by a UMIST Life Sciences Initiative Studentship.

Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2007.09.025

References 1. Vale, R. D. (2003). The molecular motor toolbox for intracellular transport. Cell, 112, 467–480. 2. Kovar, D. R. (2006). Molecular details of forminmediated actin assembly. Curr. Opin. Cell Biol. 18, 11–17. 3. Yarmola, E. G. & Bubb, M. R. (2006). Profilin: emerging concepts and lingering misconceptions. Trends Biochem. Sci. 31, 197–205.

8 4. Amos, L. A. & Schlieper, D. (2005). Microtubules and MAPs. Advan. Protein Chem. 71, 257–298. 5. Green, K. J., Bohringer, M., Gocken, T. & Jones, J. C. (2005). Intermediate filament associated proteins. Advan. Protein Chem. 70, 143–202. 6. Shih, Y. L. & Rothfield, L. (2006). The bacterial cytoskeleton. Microbiol. Mol. Biol. Rev. 70, 729–754. 7. Margolin, W. (2005). FtsZ and the division of prokaryotic cells and organelles. Nature Rev. Mol. Cell Biol. 6, 862–871. 8. Hayes, F. & Barillà, D. (2006). The bacterial segrosome: a dynamic nucleoprotein machine for DNA trafficking and segregation. Nature Rev. Microbiol. 4, 133–143. 9. Hayes, F. & Barillà, D. (2006). Assembling the bacterial segrosome. Trends Biochem. Sci. 31, 247–250. 10. Schumacher, M. A. (2007). Structural biology of plasmid segregation proteins. Curr. Opin. Struct. Biol. 17, 103–107. 11. Hayes, F. (2000). The partition system of multidrug resistance plasmid TP228 includes a novel protein that epitomizes an evolutionarily distinct subgroup of the ParA superfamily. Mol. Microbiol. 37, 528–541. 12. Barillà, D., Rosenberg, M. F., Nobbmann, U. & Hayes, F. (2005). Bacterial DNA segregation dynamics mediated by the polymerizing protein ParF. EMBO J. 24, 1453–1464. 13. Barillà, D. & Hayes, F. (2003). Architecture of the ParFParG protein complex involved in procaryotic DNA segregation. Mol. Microbiol. 49, 487–499. 14. Golovanov, A. P., Barillà, D., Golovanova, M., Hayes, F. & Lian, L. Y. (2003). ParG, a protein required for active partition of bacterial plasmids, has a dimeric ribbonhelix-helix structure. Mol. Microbiol. 50, 1141–1153. 15. Carmelo, E., Barillà, D., Golovanov, A. P., Lian, L. Y., Derome, A. & Hayes, F. (2005). The unstructured N-terminal tail of ParG modulates assembly of a quaternary nucleoprotein complex in transcription repression. J. Biol. Chem. 280, 28683–28691. 16. Barillà, D., Carmelo, E. & Hayes, F. (2007). The tail of the ParG DNA segregation protein remodels ParF polymers and enhances ATP hydrolysis via an arginine finger-like motif. Proc. Natl Acad. Sci. USA, 104, 1811–1816. 17. Ahmadian, M. R., Stege, P., Scheffzek, K. & Wittinghofer, A. (1997). Confirmation of the arginine-finger hypothesis for the GAP-stimulated GTP-hydrolysis reaction of Ras. Nature Struct. Biol. 4, 686–689. 18. Nadanaciva, S., Weber, J., Wilke-Mounts, S. & Senior, A. E. (1999). Importance of F1-ATPase residue α-Arg-376 for catalytic transition state stabilization. Biochemistry, 38, 15493–15499. 19. Galli, D. M., Chen, J., Novak, K. F. & Leblanc, D. J. (2001). Nucleotide sequence and analysis of conjugative plasmid pVT745. J. Bacteriol. 183, 1585–1594.

Stimulation of ParF Protein Polymerization 20. Kwong, S. M., Yeo, C. C. & Poh, C. L. (2001). Molecular analysis of the pRA2 partitioning region: ParB autoregulates parAB transcription and forms a nucleoprotein complex with the plasmid partition site, parS. Mol. Microbiol. 40, 621–633. 21. Fothergill, T. J. G., Barillà, D. & Hayes, F. (2005). Protein diversity confers specificity in plasmid segregation. J. Bacteriol. 187, 2651–2661. 22. Kalnin, K., Stegalkina, S. & Yarmolinsky, M. (2000). pTAR-encoded proteins in plasmid partitioning. J. Bacteriol. 182, 1889–1894. 23. Ebersbach, G., Ringaard, S., Moller-Jensen, J., Wang, Q., Sherratt, D. J. & Gerdes, K. (2006). Regular cellular distribution of plasmids by oscillating and filamentforming ParA ATPase of plasmid pB171. Mol. Microbiol. 61, 1428–1442. 24. Kędzierska, B., Lian, L. Y. & Hayes, F. (2007). Toxin-antitoxin regulation: bimodal interaction of YefM-YoeB with paired DNA palindromes exerts transcriptional autorepression. Nucl. Acids Res. 35, 325–339. 25. Lim, G. E., Derman, A. I. & Pogliano, J. (2005). Bacterial DNA segregation by dynamic SopA polymers. Proc. Natl Acad. Sci. USA, 102, 17658–17663. 26. Leonard, T. A., Butler, P. J. & Lowe, J. (2005). Bacterial chromosome segregation: structure and DNA binding of the Soj dimer - a conserved biological switch. EMBO J. 24, 270–282. 27. Adachi, S., Hori, K. & Hiraga, S. (2006). Subcellular positioning of F plasmid mediated by dynamic localization of SopA and SopB. J. Mol. Biol. 356, 850–863. 28. Fogel, M. A. & Waldor, M. K. (2006). A dynamic, mitotic-like mechanism for bacterial chromosome segregation. Genes Dev. 20, 3269–3282. 29. Bouet, J. Y., Ah-Seng, Y., Benmeradi, N. & Lane, D. (2007). Polymerization of SopA partition ATPase: regulation by DNA binding and SopB. Mol. Microbiol. 63, 468–481. 30. Ebersbach, G. & Gerdes, K. (2001). The double par locus of virulence factor pB171: DNA segregation is correlated with oscillation of ParA. Proc. Natl Acad. Sci. USA, 98, 15078–15083. 31. Ebersbach, G. & Gerdes, K. (2004). Bacterial mitosis: partitioning protein ParA oscillates in spiral-shaped structures and positions plasmids at mid-cell. Mol. Microbiol. 52, 385–398. 32. Rothfield, L., Taghbalout, A. & Shih, Y. L. (2005). Spatial control of bacterial division-site placement. Nature Rev. Microbiol. 3, 959–968. 33. Suefuji, K., Valluzzi, R. & RayChaudhuri, D. (2002). Dynamic assembly of MinD into filament bundles modulated by ATP, phospholipids, and MinE. Proc. Natl Acad. Sci. USA, 99, 16776–16781.