pushing and pulling in prokaryotic dna segregation

5/12/2018 Pushing and Pulling in Prokaryotic DNA Segregation - slidepdf.com

http://slidepdf.com/reader/full/pushing-and-pulling-in-prokaryotic-dna-segregation 1/16

Leading Edge

Review

Cell 141, June 11, 2010 ©2010 Elsevier Inc. 927

Introduction

How replicated DNA is segregated prior to cell division is ocentral importance to cell biology. In eukaryotic cells, spectac-ular advances have been made in understanding tubulin-basedmechanisms o DNA segregation. Until recently, attaining asimilar level o insight into the mechanisms o DNA segregationin prokaryotes has been impeded by the lack o a conspicu-ous intracellular cytoskeleton and the small cell size o themodel organisms. Over the past decade, new developmentshave completely changed this picture—our understanding o

prokaryotic DNA segregation is now ast catching up with thestate o play in eukaryotes.Technical advances have greatly improved the sensitivity and

resolution o uorescence microscopy, making it easible tovisualize the dynamics o subcellular structures in prokaryoticcells. Moreover, the recent advent o electron cryotomographyholds great promise or a proound mechanical understandingo prokaryotic cell ultrastructure (Li and Jensen, 2009). Mostimportantly, however, bacterial cell biology has been revolu-tionized by the recent discovery that bacteria have the threetypes o cytoskeletal proteins ound in eukaryotic cells (Fletcherand Mullins, 2010) and, in addition, have numerous cytoskel-etal proteins unique to prokaryotes (Michie and Löwe, 2006;Löwe and Amos, 2009). The discovery o tubulin-like flaments

(Bi and Lutkenhaus, 1991) and more recently proteins similar toactin (Jones et al., 2001; Møller-Jensen et al., 2002) and inter-mediate flaments (Ausmees et al., 2003) have been importantlandmarks in this progress. Clearly, the once-held view thatbacteria are eatureless bags o enzymes has become well andtruly obsolete. Instead, it is becoming evident that the interioro prokaryotic cells is exquisitely organized. A central part othis organization is generated by cytoskeletal proteins, and, asdescribed in this review, these proteins are also central playersin the machines that segregate bacterial DNA.

The three known types o bacterial DNA segregationloci, also known as par loci, encode, respectively, actin-like

ATPases, Walker A cytoskeletal P loop ATPases (ParAs), and

tubulin-like GTPases (Gerdes et al., 2000; Larsen et al., 2007). As shown in Figure 1A, par loci have strikingly similar geneticorganizations. Besides a cytoskeletal NTPase (ATPase orGTPase), par loci encode a DNA-binding protein that interactswith the NTPase and simultaneously binds site-specifcallyto one or more cognate centromeres. The adaptor proteinsthereore unction as a link between the NTPase proteins andthe centromere DNA, as well as playing crucial regulatoryroles in segregation processes. As we will see, despite thesimilarities in their genetic organization, the dierent types

o plasmid-encoded par loci unction by entirely dierentmolecular mechanisms. This conclusion is consistent withthe dierent three-dimensional olds o actin, tubulin, and Ploop ATPases, structures that hint at independent evolution-ary origins (Koonin et al., 2000).

The feld o plasmid- and chromosome-encoded partition-ing loci is developing rapidly, and several excellent reviewshave recently described structural aspects o plasmid segre-gation components (Schumacher, 2007, 2008). Except whereessential, we will thereore not dwell on these aspects in thisreview. Rather, we will ocus on the spatiotemporal dynamicso par loci proteins and potential explanations or the observeddynamics using simplifed molecular models.

Actin-like Filaments and Plasmid Segregation

In eukaryotic cells, actin flaments perorm a plethora o unc-tions, including internalizing membrane vesicles, acilitatingthe movement o cells over suraces, ormation o the cytoki-netic ring, and the intracellular propulsion o some pathogenicbacteria (reviewed in Pollard and Cooper, 2009). Bacteria alsocontain dynamic actin-like fbers that are involved in cell shapedetermination (Jones et al., 2001), plasmid DNA segregation(Møller-Jensen et al., 2002), cell wall synthesis (Daniel and Err-ington, 2003), and subcellular alignment o magnetosomes inmagnetotactic bacteria (Komeili et al., 2006). In bacterial DNAsegregation, these actin-like fbers act in a simple processreminiscent o mitosis.

Pushing and Pulling in Proaryotic DNA

SegregationKenn Gerdes,1,* Martin Howard,2 and Florian Szardenings1

1Centre or Bacterial Cell Biology, Institute or Cell and Molecular Biosciences, Newcastle University, Newcastle upon Tyne NE2 4AX, UK2Department o Computational and Systems Biology, John Innes Centre, Colney, Norwich NR4 7UH, UK*Correspondence: [email protected] 10.1016/j.cell.2010.05.033

In prokaryotes, DNA can be segregated by three dierent types o cytoskeletal flaments. Thebest-understood type o partitioning ( par ) locus encodes an actin homolog called ParM, which

orms dynamically unstable flaments that push plasmids apart in a process reminiscent o mitosis.However, the most common type o par locus, which is present on many plasmids and mostbacterial chromosomes, encodes a P loop ATPase (ParA) that distributes plasmids equidistant

rom one another on the bacterial nucleoid. A third type o par locus encodes a tubulin homolog(TubZ) that orms cytoskeletal flaments that move rapidly with treadmill dynamics.



928 Cell 141, June 11, 2010 ©2010 Elsevier Inc.

The parMRC locus o Escherichia coli plasmid R1 was dis-covered 25 years ago (Gerdes et al., 1985). As shown in Figure1B, the parMRC locus encodes actin homolog ParM (Motor),the adaptor protein ParR (Repressor) and the centromere-like region parC (Centromere) (Gerdes and Molin, 1986; Damand Gerdes, 1994). More recently, parMRC homologous locihave been identifed on plasmids rom both Gram-negativeand -positive bacteria (Ebersbach and Gerdes, 2001; Beckeret al., 2006; Schumacher et al., 2007). So ar, parMRC locihave not been ound on bacterial chromosomes.

The parMRC promoter is located in the middle o parC (Fig-ure 1B) and cooperative binding o ParR to parC autoregulatestranscription o the par operon (Dam and Gerdes, 1994). In

act, all known par operons are autoregulated, either by the ATPase or by the centromere-binding proteins (Jensen et al.,1994; Carmelo et al., 2005; Ringgaard et al., 2007a). Autoregu-lation is the most common mode o gene control in bacteria—

in E. coli, approximately 40% o all genes are autoregulated atthe level o transcription or translation. However, it is likely thattranscriptional autoregulation plays only a supporting role inplasmid segregation, given that par operon transcription couldin all cases tested be driven by constitutive oreign promoterswithout loss o unction (Ogura and Hiraga, 1983; Abeles et al.,1985; Jensen et al., 1994). Thus, transcriptional autoregulationmay simply ensure that the levels o the par -encoded proteinsremain within unctional limits.

How parMRC stabilizes plasmids remained obscureor many years. However, the discovery that ParM ormsdynamic, actin-like flaments that segregate plasmids in amitotic-like process (Møller-Jensen et al., 2002; Møller-Jensen et al., 2003) held the promise that a sophisticatedmechanistic understanding o the system would be possible.Cytological investigations show that plasmids are invari-ably located at ParM flament ends, immediately suggestingthat the mechanism involves pushing o plasmids to the cellpoles. In this way, the plasmids would be located in separatecell halves at the time o septum closure (Møller-Jensen etal., 2003). The observation that ParR can pair parC-carryingplasmids gave the frst hint that the parMRC locus stabilizesplasmids by a mechanism that is analogous to chromosomesegregation in eukaryotic cells (Jensen et al., 1998). Plas-mid pairing and active plasmid movement to the cell polesare also consistent with the subcellular symmetric pattern oplasmids carrying the parMRC locus (Jensen and Gerdes,

1999). ParM flaments are observed in?

40% o exponen-tially growing cells, as expected or rapid turnover o theflaments (Møller-Jensen et al., 2002). ParR and parC areboth required or ParM flaments to orm. Filament orma-tion could not be achieved solely by ParM overproduction,implying that the ParR/ parC complex controls ParM flamentdynamics as described urther be low.

During the initial cytological studies o the parMRC sys-tem, immunouorescence microscopy was used to detect thecomponents (ParM and plasmids) in fxed cells. A more recentstudy extends these studies to living cells (Campbell and Mul-lins, 2007). By time-lapse microscopy, these authors show thatthe parMRC spindle orces paired (or clustered) plasmids apartand moves them rapidly, within seconds, to opposite cell poles.

Ater reaching the poles, the segregated plasmids resumediusive motion in the polar region while the ParM flamentsrapidly and completely depolymerize. Typically, the plasmidsundergo several rounds o segregation during a single cellcycle, indicating that plasmid segregation is not coupled to thecell cycle. In cells with three plasmid oci, repeated rounds osegregation develop into a stable pattern in which one plasmidis ejected rom a polar ocus, moves rapidly to the oppositepole (within 10–30 s), and uses with the ocus at that pole.Such repeated ping pong-like cycles o movement ensure thateach cell pole is continuously populated by a plasmid, thusensuring proper plasmid segregation at cell division (Campbelland Mullins, 2007).

Figure 1. Genetic Organization of par Loci(A) Generic organization o a par locus.(B) par loci encoding actin homologs (such as those o plasmids R1, pB171,pTAR, pSK41, and pBET131).(C) par loci encoding large ParAs with N-terminal DNA-binding domains(such as plasmids P1, F, and RK2).(D) par loci encoding small ParAs (such as plasmids pB171, TP228,pSM19035, and bacterial chromosomes). In the latter cases, the parAB genes are located very close to oriC and the centromere sites ( parS ) areoten spread out in the origin-proximal region (Livny et al., 2007).(E) TubZ-encoding par loci (such as those o plasmids pBtoxis and pXO1).Solid arrows represent genes that encode NTPases (orange) and centromere-binding proteins (purple), respectively. Centromere-like sites are shown as blackbars. Arcs indicate DNA-binding properties o par gene products: solid arcs,regulation o promoter activity; dashed arcs, ormation o partitioning complex.




An important question is whether the ParM flamentsobserved in vivo consist o one or multiple flaments. Evidenceor bundles o flaments comes rom the observation that ParMflaments do not always depolymerize in a single step (Camp-bell and Mullins, 2007). This view is supported by the identif-cation o ParM flament bundles by cryo-electron microscopy(cryo-EM), the frst direct in vivo images o a cytoskeletal fla-ment (Salje et al., 2009). The authors unequivocally identiysmall bundles o three to fve ParM flaments in cross-sectionso cells harboring a plasmid carrying the parMRC locus. TheseParM bundles are consistently located near the edge o thenucleoid. The observed number o flaments may be roughlyrelated to the copy number o the plasmid, and it is suggested

that one ParM flament carries one plasmid at each end. In vitrostudies using DNA-gold nanocrystal conjugates show thateach end o a single polar ParM flament can bind to a singleParR/ parC gold complex, supporting the notion that one ParMflament carries one plasmid at either end (Choi et al., 2008).

Comparison of ParM and Actin Filaments

The crystal structure o the ParM monomer reveals that it isclosely related to those o actin and MreB (van den Ent et al.,2002). ParM has the characteristic actin old (Kabsch and Hol-mes, 1995): two domains with a nucleotide-binding pocket inthe interdomain clet. Upon binding o ADP (van den Ent et al.,2002) or GDP (Popp et al., 2008), the interdomain clet closes,

with domains I and II approaching eachother as rigid bodies. The domain move-ment associated with nucleotide bindingis ?25°, with the connecting interdomainhelix acting as a mechanical hinge. ParMis convenient to work with in vitro andorms flaments that have been analyzed

by several dierent EM techniques. Ini-tially, EM images o negatively stained flaments showed thatParM orms a two-stranded helix very similar to that o actin. Inthese images, the crossover distance, corresponding to a halhelical turn, is 300 Å, whereas the distance between the mono-mers is 49 Å (van den Ent et al., 2002). For actin, these num-bers are 360 and 55 Å, respectively. These close resemblancesraised the possibility that the ParM helix is right handed, asis the case or actin flaments. Surprisingly, two groups haveshown that the ParM helix is, in act, let handed (Orlova etal., 2007; Popp et al., 2008). Reassuringly, however, the maindierences between ParM and actin monomers are in regionswith predicted involvement in the flamentous subunit-subunitinterace o a right-handed helix (Orlova et al., 2007). The oppo-

site handedness o ParM and actin flaments raises the possi-bility that evolution has developed actin-like flaments at leasttwice.

ParM Polymers Exhibit Dynamic Instability

Unexpectedly, ParM flaments polymerize bidirectionally invitro, with equal rates o monomer addition at both ends (Gar-ner et al., 2004; Popp et al., 2007). This is surprising, giventhe polarity o the ParM polymer (van den Ent et al., 2002;Garner et al., 2004; Orlova et al., 2007; Popp et al., 2008).

At physiological concentrations o 10 µM, ParM polymerizesat a rate that would extend a ParM flament between thetwo cell poles in approximately 10 s (Garner et al., 2004).

Figure 2. Dynamic Instability and Groth of

ParM Filaments(A) A molecular model o dynamic instability. ParMflament is shown as a stack o circles, where eachcircle represents a ParM protomer. Blue circles

represent ParM-ATP, green circles represent ParM- ADP, and blue circles with a deleted sector repre-sent the intermediate ParM-ADP-Pi state. Shortlyater polymerization, the ParM flament is com-posed o ATP subunits (1), which over time turninto the ADP-Pi state (2) ollowed by the ADP state(3). The integrity o the flament in (1)–(3) is pre-served by the ATP-cap. Inorganic phosphate re-lease inhibits the ormation o the protective ATP-cap (4). The ADP subunits are then exposed (5),and fnally the flament depolymerizes (6) (adaptedrom Galkin et al., 2009).(B) Schematic drawing describing the proces-sive ParM polymerisation mechanism. The ParRC clamp binds to the two terminal ParM-ATP subunitsand allows addition o only one subunit to one pro-toflament at a time because o steric constraints.Hydrolysis o ATP to ADP releases the ParRC helix

on one side only (causing processivity) and reat-taches to the newly added subunit, causing trans-location. The reattachment causes the ParRC helixto “rock” and to allow addition o a new subunitto the second protoflament in exactly the sameway. This scheme has been called the “staircase”model (adapted rom Salje and Löwe, 2008).




Even more strikingly, the symmetrical bidirectional growthtransitions to an even aster, unidirectional disassembly othe flaments. This behavior is less actin like, and insteadresembles the dynamic instability o microtubules (Mitchi-

son and Kirschner, 1984), where individual microtubule endsalternate between bouts o growth and shrinkage (Mitchisonand Kirschner, 1984). Dynamic microtubule instability is ahighly regulated process, and its prevention is critical orproper eukaryotic chromosome segregation (Higuchi andUhlmann, 2005).

Importantly, ParM exhibits cooperative ATPase activity invitro (Jensen and Gerdes, 1997; Møller-Jensen et al., 2002),and, moreover, the switch rom ParM polymer elongation toshortening is regulated by ATP hydrolysis as ParM flamentsormed in the presence o nonhydrolysable ATP analogs arestable (Møller-Jensen et al., 2002; Garner et al., 2004). Themeasured rate o flament disassembly ar exceedes that o

ATP hydrolysis, consistent with the proposal that the fla-ments quickly depolymerize when nucleotide hydrolysiscatches up with polymerization at one o the flament ends. Insupport o this hypothesis, the addition o a small amount oan ATP hydrolysis-defcient ParM mutant to the polymeriza-tion reaction leads to the stabilization o the flaments (Garneret al., 2004). GTP, like ATP, stimulates the onset o bidirec-tional ParM polymerization, and ParM flaments ormed in thepresence o GTP also exhibit dynamic instability (Popp et al.,2008). However, ParM binds ATP with a 10-old higher afnitythan GTP, supporting the conclusion that ATP is the primarysubstrate or ParM in vivo (Galkin et al., 2009). Interestingly,ParM protomers exist in two orms within the flaments: aclosed orm, avored by ATP, in which the nucleotide-binding

clet closes around the nucleotide, and an open orm, avoredby ADP + Pi (Galkin et al., 2009). This work also presentsatomic structures o both types o flaments. In both cases,the interaces between ParM protomers are completely di-erent rom those ound in F-actin. Again, these results areconsistent with the more microtubule-like dynamic instabilityound in ParM flaments.

Figure 2A shows a molecular model or how dynamic insta-bility o ParM flaments may be regulated. Within the flaments,ParM can bind to ATP, ADP + Pi, or ADP. Only the ATP-boundorm o ParM allows polymerization. Thereore, the ends ogrowing flaments will be capped by ParM-ATP that, as arguedabove, prevents flament depolymerization. I, on the otherhand, ATP hydrolysis within the flament reaches the flament

end, no urther ParM-ATP monomers can be added, the pro-tective ParM-ATP cap is lost, and catastrophic depolymeriza-tion rom the flament end ensues. The model is described inurther detail in the fgure legend.

The ParM homolog AlA o B. subtilis NATTO plasmid pLS32also orms helical, two-stranded flaments with a let-handedpitch (Polka et al., 2009). However, despite these similarities withParM flaments, the positioning o the AlA monomers in the fla-ments is radically dierent rom that o ParM, and, unexpectedly,

AlA flaments do not exhibit dynamic instability. These resultssuggest that the dynamics o AlA flaments is dierent romthat o ParM (Polka et al., 2009). However, urther work will berequired to ully substantiate this potential dierence.

Reconstitution of an Active DNA Segregation Apparatus

Important progress on the parMRC system has come rom anin vitro reconstitution o the plasmid R1 DNA segregation appa-ratus (Garner et al., 2007). Such in vitro experiments are vital

in determining the essential components necessary or reliableoperation o a biological system. To mimic plasmids carrying

parC, uorescently labeled parC DNA ragments are coupledto spherical beads and incubated with ParR, and uorescentlylabeled ParM is then added to the beads. In the presence o

ATP, ParM flaments, reminiscent o microtubule asters, ormaround isolated centrosomes. The asters are dynamic, withflaments growing and shrinking rom the sur ace o the bead,reaching a maximum length o ?3 µm. In the presence o non-hydrolyzable ATP analogs, the flaments grow much longer,indicating that the dynamic instability o ParM limited flamentlength even when one end o the flament is bound to the ParR/

parC complex.In addition to dynamic asters, ParM flaments orm connec-

tions between ParR/ parC beads, similar to the bipolar struc-tures seen in vivo (Møller-Jensen et al., 2003; Garner et al.,2007; Campbell and Mullins, 2007). In time-lapse series, theParR/ parC-coated beads are pushed in opposite directionsat a constant speed, even over long distances o up to 120µm. Stable flament ormation is seen only between pairs oParR/ parC-coated beads, indicating that the attachment othe centromere complex at both flaments ends prevents cata-strophic decay o the ParM polymer (Garner et al., 2007). Byemploying uorescence recovery ater photobleaching (FRAP)and “speckle microscopy,” Mullins and coworkers have shownthat ParM monomers are added at flament ends, most likelyat the junction between the flament end and the ParR/ parC

complex. Free ParM flaments and flaments attached to ParR/ parC, elongate at similar rates, adding ?12 monomers o ParMper second.

The ull in vitro reconstitution o a unctional parMRC spindleapparatus represents an important step orward and providesstrong evidence that additional host cell-encoded actorsare dispensable or the plasmid segregation process. Takentogether, these results show that dynamic instability o ParMflaments is critical or plasmid segregation: flaments stabilizedat one end by ParR/ parC can search the cytoplasmic space orplasmids with a ParR/ parC complex; flaments capped at bothends are actively segregated by ParM polymerization; and,fnally, the dynamic instability o ree ParM flaments providesParM monomers to drive elongation o the stabilized flaments

within the spindle (Garner et al., 2007).

Conserved Structures of Two Divergent Centromere

Complexes

A critical issue in the parMRC system is how the ParR pro-tein interaces with the parC sequence to orm the ParR/ parC complex. In previous work, we showed that ParR o plasmidR1 can pair parC-carrying DNA molecules in vitro (Jensenet al., 1998). In that study, contour-length-measurements oDNA ragments indicate that parC wraps around a core oParR. Direct evidence or such wrapping has now come romtwo recent independent studies (Møller-Jensen et al., 2007;Schumacher et al., 2007).




In the frst o these studies, using EM, it is ound that ParRorms ring-shaped complexes on DNA containing parC (Møller-Jensen et al., 2007). The ring-shaped complexes have diam-eters o 15–20 nm, consistent with the structural analyses o

homologous centromere complexes described below. Some othe ParR/ parC rings are ound as paired structures.

The structure o ParR rom E. coli plasmid pB171, whichexhibits 44% similari ty with ParR o R1, has been solved (Møller-Jensen et al., 2007). It reveals dimers-o-dimers orming ribbon-helix-helix (RHH) DNA-binding motis (Figure 3A). Thus, ParRo pB171 belongs to the MetJ/Arc superamily o DNA-bindingproteins. The RHH motis are ollowed by C-terminal domainsconsisting o a three helix “cap.” This C-terminal cap is pro-posed to strengthen the tight dimerization o the RHH2 motisand to stabilize the dimer-dimer interactions (Møller-Jensen etal., 2007). As seen in Figure 3B, ParR assembles into a continu-ous helical structure consisting o six dimers-o-dimers per ull360° turn (Møller-Jensen et al., 2007). ParR dimers are arrangedwith their N termini acing outward and their C termini point-ing toward the helix center. The N-terminal RHH2 DNA-bindingdomains align on the helix exterior and orm regularly spacedbasic patches. The act that the RHH2 DNA-binding sites arepositioned 3.5 nm apart, which corresponds to one helical turno B helix DNA, and are in register with the repeat structure othe parC centromere DNA strongly argues or a biological rel-evance o the structure. Thus, the crystal structure suggests apartition complex architecture in which the centromere wrapsaround a helical scaold ormed by ParR.

Direct evidence or this possibility came rom a second studythat detailed the structure o the N-terminal domain o ParR opSK41 bound to its cognate parC centromere (Schumacher et

al., 2007). Plasmid pSK41 rom Staphylococcus aureus carriesa parMRC region that shares little sequence similarity with par-

MRC rom plasmid R1. Strikingly, however, ParR rom pSK41 isalso packed in the crystal in a continuous helix consisting o sixdimers-o-dimers in one turn (Schumacher et al., 2007). Again,the N-terminal basic DNA-binding domains orm a continuoussurace on the helix exterior that wrap the centromere DNAabout itsel to create a unique structure that can be describedas a large superhelix (Figure 3C). Within this structure, theC-terminal parts o ParR ace inwards, toward the center othe helix (Schumacher et al., 2007), where they play a role inthe interaction between the centromere complex and ParM(see below). The high degree o conservation o the ParR/ parC structures rom distantly related organisms strongly suggests

that the par loci o plasmids R1 and pSK41 unction by similarmolecular mechanisms.

Molecular Model Explaining Plasmid Segregation by

parMRC

The initial in vivo observations demonstrated that ormation olong ParM flaments required the presence o the ParR/ parCcom-plex (Møller-Jensen et al., 2002), a conclusion later confrmed bythe in vitro reconstitution experiments discussed above (Garneret al., 2007). Thus, the mechanism by which the ParR/ parC com-plex stabilizes ParM flaments is key to understanding the overallprocess. Salje and Löwe (2008) have proposed a mechanismas to how this may occur, based on a combination o elegant

Figure 3. Structure of ParR and ParR/ parC ComplexesParR is a ribbon-helix-helix (RHH) DNA-binding protein that orms dim-ers-o-dimers. Six dimers-o-dimers associate into a continuous helix inwhich the RHH

2domains orm regularly spaced basic pa tches on the helix

exterior.(A) The right panel shows a side view o two interacting ParR dimers o

pB171. The C termini are inside the helix whereas the RHH2 DNA-bindingmotis are on the exterior o the helix. The let panel shows the RHH 2 motisrom above. The distance between adjacent RHH

2 β-ribbon structures cor-

responds to one helical turn o the DNA double helix.(B) ParR o pB171 assembles into a helix with a 13 nm translation per turn(let) and a 15 nm diameter when viewed along the screw axis (right). RHH2 domains orm regularly spaced basic patches on the helix exterior.(C) The let panel shows a ribbon diagram o the N-terminal part o ParRo pSK41 in complex with centromere DNA. The view is looking down thesuper-helix axis. N-terminal ParR dimer-o-dimers are white, cyan, yellow,magenta, green, and blue. The right panel shows an electrostatic suracerepresentation o the centromere complex. Blue and red represent electro-positive and electronegative suraces, respectively.Images rom (A) and (B) are reprinted by permission rom MacMillan Pub-lishers Ltd: EMBO (Møller-Jensen et al., 2007), copyright 2007. Imagesrom (C) are reprinted by permission rom MacMillan Publishers Ltd: Na-ture (Schumacher et al., 2007), copyright 2007.




biochemical and EM experiments: the open mouth o the ParR/ parC helix orms a clamp around the end o elongating ParMflaments, and each ParR/ parC complex caps a single ParM fla-ment (Figure 2B). Via the ParR C termini in the helix interior, the

ParR/ parC clamp binds to the two terminal ParM-ATP subunits.Due to steric constraints, the resulting complex allows the addi-tion o only one ParM-ATP subunit to one protoflament end ata time. Hydrolysis o ATP to ADP releases the ParR/ parC helixon one side only. This side o the helix reattaches to the newlyadded ParM subunit, causing translocation. The reattachmentleads to the generation o the space necessary or the ParR/

parC helix to allow addition o a new subunit to the second pro-toflament in exactly the same way. The process can then iterate,thereby generating processivity.

Although attractive, the model proposed by Salje and Löwedoes have at least one potential drawback, arising rom stericconsiderations. One remarkable eature o the ParR/ parC com-plex is its ability to interact with both ends o the asymmet-ric ParM flament, thereby promoting bidirectional elongation(Møller-Jensen et al., 2003; Garner et al., 2007). This eatureis radically dierent rom anything that has been observed oreither conventional actin or microtubules. Thereore, or thismodel to work, the subdomains o ParR that interact with ParMmust be ree to rotate by at least 180° or the complex to attachat both ends o the ParM flament. An alternative model, pro-posing that the ParR/ parC helix (Figures 3B and 3C) encirclesthe ParM flament, does not suer rom this steric problem(Popp et al., 2010). This latter solution has a precedent romeukaryotes, where the Dam1 complex rom budding yeast isthought to encircle microtubules during chromosome segrega-tion (Eremov et al., 2007; Grishchuk et al., 2008).

Partitioning Loci Encoding P loop ATPases

par loci encoding cytoskeletal ATPases with a variant Walker A box moti were discovered almost 30 years ago (Ogura andHiraga, 1983; Austin and Abeles, 1983), but mechanistic insighthas begun to emerge only rather recently. MinD ATPases(described later) possess the same variant Walker A box motiand share a number o important characteristics with the par -encoded ParAs (Michie and Löwe, 2006). So ar, this type o Ploop ATPase has only been ound in prokaryotes, but based onstructural similarities (Michie and Löwe, 2006) it has been sug-gested that the ParA/MinD amily o proteins may be relatedto septins that orm cytoskeletal-like structures involved in thedivision o eukaryotic cells (Bertin et al., 2008; McMurray and

Thorner, 2009). However, urther work is necessary to test thisintriguing possibility.

ParAs can be subdivided into two subtypes: those possess-ing an N-terminal DNA-binding helix-turn-helix moti and thosethat lack this moti. We reer to these subtypes as large andsmall ParAs (Gerdes et al., 2000), as illustrated in Figures 1Cand 1D. par loci encoding large ParAs also encode large ParBsand are ound only on plasmids, whereas small ParAs lackingan N-terminal DNA-binding domain are present on many plas-mids and most bacterial chromosomes (Livny et al., 2007). Theoverall picture now emerging indicates that ParAs encodedby plasmids segregate DNA by similar mechanisms. Chromo-some-encoded ParAs are involved in both chromosome repli-

cation control and segregation, as well as other developmentalprocesses (Murray and Errington, 2008). Unless stated other-wise, we will describe plasmid-encoded ParAs (both large andsmall) under the same headings, ollowed by a separate sec-

tion on chromosome-encoded loci.

ParAs from Plasmids

The par loci encoded by the F and P1 plasmids have beenstudied extensively. Both loci encode two trans-acting proteinsand a cis-acting centromere analog located downstream o the

sop / par operons (Figure 1C). The P loop ATPases encoded byF and P1 (SopA and ParA) contain N-terminal HTH motis andautoregulate transcription o their operons via binding to oper-ators that overlap with the sop / par promoters (Friedman and

Austin, 1988; Hirano et al., 1998; Dunham et al., 2009). The par loci o plasmids TP228 and pB171, parFGH and par2, have alsobeen intensely studied and encode small ParAs (Gerdes et al.,2000; Hayes, 2000; Ebersbach and Gerdes, 2001).

Dynamic and Unusual Properties of ParA

Multiple studies have shown that ParA proteins move dynami-cally over the nucleoid (Hirano et al., 1998; Quisel et al., 1999;Marston and Errington, 1999; Ebersbach and Gerdes, 2001; Limet al., 2005; Hatano et al., 2007; Pratto et al., 2008; Castaing etal., 2008). Observations with par2 o pB171 initially suggestedthat ParA oscillates in spiral-shaped structures between thetwo ends o the nucleoid (Ebersbach and Gerdes, 2004; Eber-sbach et al., 2006), although more recent results suggest morecomplex spatiotemporal dynamics (see below). Other ParAsalso orm cytoskeletal-like structures in vivo (Lim et al., 2005;Hatano et al., 2007; Pratto et al., 2008). Some o the cytoskel-

etal-like structures observed in vivo have been assembled invitro, or ParAs rom both plasmids and chromosomes (Leon-ard et al., 2005; Barillà et al., 2005; Lim et al., 2005; Ebersbachet al., 2006; Bouet et al., 2007; Pratto et al., 2008; Batt et al.,2009). Importantly, ParAs also bind DNA cooperatively andnonspecifcally in vitro, consistent with in vivo nucleoid asso-ciation (Leonard et al., 2005; Hester and Lutkenhaus, 2007;Pratto et al., 2008; Ringgaard et al., 2009). Very recently, elec-tron microscopy and three-dimensional reconstruction sug-gest that ParA2 o Vibrio cholerae chromosome II orms helicalflaments on double-stranded DNA in a sequence-independentmanner, thus urther validating ParAs as cytoskeletal, flament-orming proteins (Hui et al., 2010). ParA2 orms asymmetricflaments in the presence o either ADP or ATP, but the two

nucleotides induce flaments with clear ly distinguishable heli-cal pitches (204 versus 120 Å, respectively).

In most cases, nonspecifc DNA stimulates ParA polymer-ization and bundling, but in the case o the F plasmid, DNAinhibits polymerization (Bouet et al., 2007). The physiologicalrelevance o the disparities seen in vitro or dierent ParAs isnot yet understood and may reect genuine mechanistic dier-ences or artiacts (Castaing et al., 2008).

ParAs orm dimers in vitro (Leonard et al., 2005; Pratto et al.,2008; Dunham et al., 2009), and it is the ATP-bound dimers thatassociate cooperatively with nonspecifc DNA (Leonard et al.,2005; Pratto et al., 2008). The observation that mutant ParAsunable to bind ATP do not associate nonspecifcally with DNA




has led to the suggestion that the state o the nucleotide bounddetermines the subcellular localization o ParA (Ebersbach andGerdes, 2001; Castaing et al., 2008; Murray and Errington,2008). The role o ATP hydrolysis in controlling the subcellularlocalization o ParAs appears central to the understanding othese par loci.

ParBs Promote the Dynamic Relocation of ParAs

ParB proteins bind site specifcally and cooperatively to centro-meric DNA that consists o either multiple direct repeats (suchas parC and sopC o pB171 and F, respectively) or invertedrepeats ( parS o P1 and bacterial chromosomes). In caseswhere the centromeres consist o inverted repeats, the ParB

proteins can spread several kb into neighboring DNA, thusorming large nucleoprotein complexes (Rodionov et al., 1999;Murray et al., 2006). ParBs also interact directly with cognateParAs, and it has been suggested that the ormation o theselarge nucleoprotein complexes increases the efciency withwhich ParAs execute the DNA segregation process. This viewis supported by deletion analyses o centromere regions (Eber-sbach and Gerdes, 2001; Ringgaard et al., 2007a), although asingle centromere repeat o F can still mediate plasmid segre-gation (Biek and Shi, 1994). Furthermore, as described later,spreading o ParB rom parS may also unction to increase theefciency with which ParB loads bacterial condensin (structuralmaintenance o chromosome; SMC) on the chromosome originregion (Gruber and Errington, 2009; Sullivan et al., 2009).

In all cases investigated, cognate ParBs are required or thedynamic relocation o ParAs in the cell (Quisel et al., 1999; Mar-ston and Errington, 1999; Ebersbach and Gerdes, 2001; Limet al., 2005; Pratto et al., 2008; Ringgaard et al., 2009). ParBsinteract with ParAs via their N-terminal ends (Autret et al., 2001;Leonard et al., 2005; Barillà et al., 2007; Ringgaard et al., 2009;Gruber and Errington, 2009) and stimulate ParA ATPase activityin vitro in the presence or absence o nonspecifc DNA (Daviset al., 1992; Leonard et al., 2005; Barillà et al., 2007; Pratto etal., 2008; Ah-Seng et al., 2009). Change o a conserved lysinein the N terminus o ParB simultaneously abolishes dynamicrelocation o ParA in vivo and stimulation o ATPase activity invitro, strongly arguing that relocation depends on ATP hydro-

Figure 4. DNA Segregation by a par Locus Encoding ParA(A) Schematic fgure showing the relationship between regular plasmid distri-bution over the nucleoid and plasmid segregation at cell division. The drawingshows division o cells with two or our plasmids and visualizes how regularplasmid distribution over the nucleoid leads to ordered plasmid transmis-

sion rom the mother cell to the two daughter cells. Note that the plasmidsare tethered to the nucleoid by the nonspecifc DNA-binding activity o theParA flaments, one end o which interacts with the ParB/ parC complex othe plasmid. In a cell with one plasmid, the ParA flaments oscillate betweenthe ends o the nucleoid. The ParA-mediated pulling o the plasmid leads toa time-averaged positioning o the plasmid at mid-cell, as observed in Figure4B. In cells with two or more plasmids, the ParA dynamics is more compli-cated and result in a regular distribution o the plasmids over the nucleoid.

At cell division, the regular distribution on, and tethering o the plasmids to,the nucleoid DNA results in ordered segregation o the plasmid copies to thedaughter cells.(B) Visualization o the subcellular localization o a ully unctional ParA-GFPusion (green) and an R1 plasmid carrying the ParA-encoding par locus opB171 ( par2 ) (red) in a time lapse series (Ringgaard et al., 2009).




lysis (Autret et al., 2001; Barillà et al., 2007; Ringgaard et al.,2009). These fndings suggest that ParB promotes the con-version o ParA-ATP to ParA-ADP, to stimulate detachment oParA rom the nucleoid. These observations also indicate thatdierent ParBs control ParA activity by very similar i not identi-cal molecular mechanisms.

ParAs Position Plasmids at Regular Intervals

To understand how par loci encoding P loop ATPases stabi-lize plasmids, the subcellular localization and movement oplasmids have been analyzed in an important series o experi-ments using uorescent in situ hybridization (FISH) or taggingo plasmids with uorescent DNA-binding proteins. F and P1plasmids localize at mid-cell in newborn cells and migrate to

approximately quarter-cell positions ater duplication (Gordonet al., 1997; Niki and Hiraga, 1997). Plasmids lacking par lociare distributed randomly in spaces not occupied by nucleoids.These initial observations led to the suggestion that par locitethered plasmids at cell-quarter positions. However, morerecent observations with improved detection techniques showthat these par loci position plasmids regularly over the nucle-oid (Ebersbach et al., 2005, 2006; Adachi et al., 2006; Hatanoet al., 2007; Bertin et al., 2008; Ringgaard et al., 2009; Sen-gupta et al., 2010). Such equidistant positioning is observedeven ater plasmid copy number amplifcation, strongly sug-gesting that the cells do not carry a receptor that tethers plas-mids at specifc subcellular sites (Ebersbach et al., 2006). The

observation o unctional par loci in het-erologous host cells is also consistentwith an apparent lack o host-encodedreceptors (Yamaichi and Niki, 2000;Godrin-Estevenon et al., 2002). The reg-ular plasmid distribution also fts in wellwith the initial observation that newborncells contain one plasmid that is approxi-

mately at mid-cell (Figure 4A). A ter replication o the plasmid,the plasmids move to positions that are equidistant rom themiddle o the cell in opposite cell halves. Most importantly,the regular plasmid positioning ensures that whenever morethan one plasmid is present, there will be at least one plas-mid copy on each side o the cell division plane (Figure 4A).Thus, regular plasmid positioning over the nucleoid explainshow ParAs mediate stable transmission o plasmids at cell divi-sion (as explained urther in the legend to Figure 4A). It is notyet ully understood how the ParA flaments generate regularplasmid distributions, although simple molecular and math-ematical models to explain this fnding have very recently beendeveloped (see below). However, it is already clear that theplasmid segregation mechanisms mediated by ParA and ParM

are entirely dierent, with respect to both the subcellular local-ization o the segregated plasmids (nucleoid versus cell poles)and the mechanism o orce generation (see below).

ParA-Mediated Plasmid Movement

The regular plasmid localization, and rapid movement aterplasmid duplication, set demanding constraints on themechanism o ParA-mediated plasmid movement. In partic-ular, ParA proteins must be able to move the plasmids whileconstantly adjusting the interplasmid distance according tothe number o plasmids present. Clearly, two possibilitiesarise or the orce generating mechanism: it could either bea pushing mechanism, as in the case o ParM, or a pulling

Figure 5. The ParA Pulling Mechanism A molecular model o the ParA pulling mechanismbased on Ringgaard et al. (2009). ParA-ATP dim-ers bind cooperatively to nucleoid DNA, leading tothe ormation o ParA flaments. Formation o fla-

ments begins with a nucleating core rom whichrapid polymerization proceeds (1). Subsequently,a growing flament contacts a plasmid via ParBbound to parC centromere DNA (2). ParBs boundto parC on the plasmid stimulate the ATPase ac-tivity o ParA-ATP at the end o the flament (3).By this reaction, ParA-ATP is converted to its ADPorm and released rom the DNA, leaving a newParA-ATP flament end accessible or interactionwith the partition complex. For each depolymer-ization event, the plasmid can either detach (4 ′ )or remain attached to the end o the depolymer-izing ParA flament (4). The moving plasmid leavesbehind it a ParA-ree nucleoid zone (5). Eventu-ally, the ParA-ATP subunits released by ParB/

parC assemble into a new flamentous structurein this zone that polymerizes toward the plasmidrom the opposite side. Upon contact ormation,

this flament will move the plasmid in the oppositedirection. In this way, a plasmid will jiggle aroundits position in between two other plasmids or be-tween a plasmid and the nucleoid end. Finally,ree ParA-ADP is rejuvenated to ParA-ATP and thecycle repeats (6).




mechanism, as in tubulin-mediated movement o eukaryoticchromosomes (Eremov et al., 2007; McIntosh et al., 2008;Grishchuk et al., 2008; Wan et al., 2009). To analyze howParA o pB171 moves plasmids over the nucleoid, we devel-

oped a triple-labeling system that allowed the simultane-ous visualization in a time-lapse series o: (1) ParA, (2) theplasmids, and (3) the bacterial nucleoid (Ringgaard et al.,2009). Our results suggest that ParA moves plasmids by apulling mechanism. In the presence o a single plasmid, asseen in Figure 4B, ParA nucleates away rom the plasmidand subsequently polymerized toward it. Upon reachingthe plasmid, ParA polymerization reverses to depolymer-ization, and the plasmid ollows the retracting ParA signal,thus suggesting a mechanism based on pulling. Ater a shortperiod, ParA nucleates on the opposite side and the cycleis repeated. These observations rom living cells, togetherwith the biochemical data obtained with dierent ParAs,support a molecular model that explains how the energy o

ATP hydrolysis is converted into mechanical orce capableo pulling plasmids (see Figure 5). Direct evidence or thismodel would require the in vitro reconstitution o an activeParA-encoding par system, an important, i challenging, goalor uture experiments.

The par locus o F ( sopABC ) has also been investigated withrespect to SopA and plasmid dynamics in living cells (Hatanoet al., 2007). The observed patterns o plasmid movement andlocalization are in most respects similar to those or pB171.However, the dynamics o SopA-green uorescent protein(GFP) dier rom that o ParA-GFP. SopA-GFP orms a largeocus that oscillates rom one end o the nucleoid to the other.While oscillating, the F plasmid ocus ollows the SopA ocus.

However, there is no apparent contact between the SopA andplasmid oci. SopA-GFP also orms a static flamentous struc-ture that spans the entire length o the nucleoid without beingshortened as the plasmid migrates (Hatano et al., 2007). Thisstatic SopA scaold may tether plasmids via the centromerecomplex, thus explaining why plasmids apparently stay associ-ated with the nucleoid in the absence o SopA/ParA-mediatedmovement (Hatano et al., 2007; Ringgaard et al., 2009).

Importantly, the proposed molecular model in Figure 5does not yield a straightorward explanation or how ParA-encoding par loci generate the observed time-averagedequidistant positioning o plasmids. However, mathematicalmodeling suggests that i the plasmid detachment rate (step4′ in Figure 5) varies with flament length such that the detach-

ment rate decreases with increasing flament length, then asimple stochastic lattice model could predict the observedplasmid distributions seen in living cells (Ringgaard et al.,2009). The plasmid migration distances observed in experi-ments are indeed positively correlated with the initial lengtho the ParA flament, consistent with the prediction that theplasmid detachment rates should be lower or longer ParAflaments. How the par components manage to vary plasmiddetachment rates is not yet known. However, one possibilityis that ParB controls the detachment rate by regulating ParAflament bundling or by aecting the strength with which ParAadheres to the ParB-bound centromere. These possibilitiescertainly call or urther experimental work.

ParB Mediates Centromere Pairing

The model in Figure 5 does not include plasmid pairing at cen-tromeres. However, several independent lines o evidence sug-gest that ParB can pair plasmids via centromere-binding, both

in vivo (Edgar et al., 2001; Bouet et al., 2006; Sengupta et al.,2010) and in vitro (Schumacher and Funnell, 2005; Ringgaard etal., 2007b; Pratto et al., 2008). In vitro, the N terminus o ParB isrequired or centromere pairing (Ringgaard et al., 2007b; Prattoet al., 2009). Interestingly, ParB-mediated centromere pairingmay be regulated by ParA, given that the presence o ParA-ATPavors the ParB-mediated in vitro pairing reaction (Pratto et al.,2009). These observations raise the possibility that ParA mayregulate the release o paired complexes. The observation thatParA can disassemble the P1 partition complex at low concen-trations o ParB (Bouet and Funnell, 1999) is consistent with thelatter conjecture.

Based on data derived rom time-lapse image analysis o alarge number o cells, it has been proposed that plasmid pai r-ing enhances the efciency o P1 plasmid segregation (Sen-gupta et al., 2010). The repeated pairing and active separationo plasmid oci may enhance the fdelity o plasmid segrega-tion by ensuring that sister plasmids in close proximity to oneanother are recognized by the partitioning apparatus, pairedup, and are again moved apart. Mathematical modeling othis process yields plasmid distributions very similar to thoseobserved in vivo, raising the possibility that plasmid pairingconstitutes an important step in the segregation process.Whether this mechanism by itsel is sufcient to generate theobserved regular positioning o P1 plasmids requires moredetailed inormation as to how the process is regulated at themolecular level and, in particular, on the unction o ParA in

the process. In summary, the role o centromere pairing isincompletely understood, and the phenomenon deserves ur-ther study.

Comparison of ParA-Encoding par Loci and minCDE

The minCDE locus o E. coli encodes three componentsrequired or placement o the division septum at mid-cell(de Boer et al., 1989; Lutkenhaus, 2007): (1) MinD, a P loop

ATPase that oscillates rapidly in spiral-shaped structuresrom one cell pole to the other, (2) MinC, a cell division inhibi-tor that associates with MinD and prevents ormation o theFtsZ ring, and (3) MinE, a topological specifcity actor thatstimulates the ATPase activity o MinD and thereby generatesthe dynamic, oscillating behavior o the MinCD complex. The

cellular oscillation o MinCD prevents ormation o the Z ringat the cell pole and thereby helps the Z ring to localize at mid-cell, where the MinC concentration is lowest. The similaritiesbetween the minCDE system o E. coli and ParA-encoding

par loci are striking: ParAs and MinD are similar variant Ploop ATPases that orm dynamic patterns on a cellular sur-ace (nucleoid and cell membrane, respectively) (Raskin andde Boer, 1999; Ebersbach and Gerdes, 2001; Hu et al., 2002;Shih et al., 2003); ATP-bound dimers o ParAs and MinD areproposed to orm cytoskeletal-like flaments on DNA andmembranes, respectively (Hu and Lutkenhaus, 2003; Leonardet al., 2005; Pratto et al., 2008); the nucleotide state deter-mines the subcellular locations o ParAs and MinD (Lackner et




al., 2003; Leonard et al., 2005; Prat to et al., 2008); the ATPaseactivities o the proteins are stimulated by the N termini otheir dimeric partner proteins (ParB and MinE, respectively),partners that are required or protein dynamics in living cells

(Radnedge et al., 1998; Ma et al., 2004; Leonard et al., 2005;Barillà et al., 2007; Pratto et al., 2008); and flament dynamicsposition another cellular structure (Z ring or plasmid, respec-tively). These observations show that evolution has solvedtwo very dierent spatial and mechanical problems relatedto cell division (DNA segregation and septum placement) byanalogous molecular mechanisms. Moreover, or both theMin and Par systems, mathematical modeling (Meinhardt andde Boer, 2001; Howard et al., 2001; Howard and Kruse, 2005;Doubrovinski and Howard, 2005) has proven invaluable inuncovering the undamental mechanisms at work and in rig-orously demonstrating that realistic dynamics can be gener-ated by simple, but experimentally-motivated, computationalmodels.

Regular Subcellular Positioning by Orphan ParA

Homologs

Many prokaryotic chromosomes encode parA homologs thatlack an obvious ParB DNA-binding partner and a centrom-ere region (Gerdes et al., 2000). Some o these “orphan” parA genes are located within large operons that encode specializedmetabolic unctions. Here, we give three important exampleso orphan ParAs, together with their respective roles in subcel-lular positioning.

Carbon fxation in cyanobacteria occurs within carboxy-somes, structures that consist o an icosahedral proteinaceousshell enclosing the carbon-fxing enzymes. Very recently, uo-

rescent labeling has demonstrated that the carboxysomes arespatially ordered in a linear ashion, very similar to plasmidscarrying ParA (Savage et al., 2010). As a consequence, cellsundergoing division evenly segregate carboxysomes in a non-random process. Mutation o the cytoskeletal protein ParAspecifcally disrupts carboxysome order, promotes randomcarboxysome segregation during cell division, and impairscarbon fxation. Moreover, ParA orms flament-like structuresthat oscillate rom one cell pole to the other, suggesting thatthe mechanisms or plasmid and carboxysome positioning arerelated.

Rhodobacter sphaeroides, a chemotactic photosyntheticbacterium, contains chemoreceptor clusters located in itscytoplasm. Interestingly, clusters containing the chemorecep-

tor TlpP are located either at mid-cell or at the cell quarters(Thompson et al., 2006). However, disruption o ppfA, encod-ing an orphan ParA homolog reduces the number and altersthe regular positioning o the TlpP-containing clusters. At thesame time, the disrupted cells lost the corresponding chemot-actic response. These observations show that ParA can alsoposition protein complexes in the cytoplasm and that suchpositioning can be vital or unction.

In enterobacteria, a ParA homolog (BcsQ) localized to thepole is essential or cellulose biosynthesis. Immunogold detec-tion o cellulose at the BcsQ-labeled pole o individual bacteriastrongly suggests a role or BcsQ in the polar localization ocellulose biosynthesis (Le and Ghigo, 2009).

ParA Proteins and Chromosome Segregation

Most bacterial chromosomes encode a ParA type par locusand so ar these loci are the only type identifed in bacterialchromosomes. Because o the pivotal role played by par loci

in plasmid segregation, it has been suggested that the chro-mosome-encoded par loci mediate chromosome segregation(Ireton et al., 1994; Sharpe and Errington, 1996; Glaser et al.,1997; Gerdes et al., 2000; Fogel and Waldor, 2006; Jakimowiczet al., 2007; Livny et al., 2007; Toro et al., 2008). Chromosome-encoded centromeres where the ParAB proteins act are usu-ally denoted parS. The frst experimentally verifed parS sitesbound by ParB were discovered in B. subtilis, with eight suchsites located in the 20% origin-proximal region o its chromo-some (Lin and Grossman, 1998; Breier and Grossman, 2007). Arecent comprehensive bioinormatics analysis o 400 prokary-otic chromosomes shows that 69% o all bacteria contain parS sites related to parS o B. subtilis, making parS arguably thebest-conserved DNA sequence moti identifed in bacteria todate (Livny et al., 2007). In the majority o cases, the parS sitesare located in the origin-proximal region. Almost all chromo-somes encoding parS sites also encode corresponding parAB genes, usually located very close to the origin o replication.Only Archaea, two branches o γ-proteobacteria (including E.

coli ), and one branch o Firmicutes (including Mycoplasma )do not contain obvious parABS loci, arguing that these locievolved early in the bacterial phylum and that their absencemay reect gene loss (Livny et al., 2007).

In some organisms, chromosome-encoded parABS loci playa direct role in chromosome segregation, as in Vibrio cholerae (Fogel and Waldor, 2006) and Caulobacter crescentus (Toro etal., 2008), two crescent-shaped bacteria with similar patterns

o origin movements. V. cholerae has two circular chromo-somes that are segregated independently (Fogel and Waldor,2005). Each chromosome has a par locus encoding a ParA,ParB, and multiple origin-proximal parS sites to which thecognate ParB protein binds (Yamaichi et al., 2007). In particu-lar, ParAI encoded by chromosome I o V. cholerae exhibits adisassembly pattern suggestive o orce generation by pulling(Fogel and Waldor, 2006). ParAI extends rom the new cell poletoward the old pole, where ParBI is localized. Upon ParBI/ ori ocus splitting, ParBI/ ori moves across the cell, trailing retract-ing ParAI structures. This observation suggests that ParAIorms a dynamic structure that pulls the origin region to thenew pole through interactions with ParBI/ parSI (Fogel and Wal-dor, 2006). ParA type par loci rom plasmids and chromosomes

may thereore generate the orce used or DNA movement bysimilar molecular mechanisms (Figure 5).

In other bacteria, both Gram-positive and -negative, the roleplayed by parABS loci in chromosome segregation is less clear.

Although deletion o parB o B. subtilis increases by ?100-oldthe ormation o cells during vegetative growth that lack a chro-mosome, the majority o the cells (>98%) exhibit a normal pat-tern o chromosome segregation (Ireton et al., 1994). Moreover,deletion o parA has no detectable eect on chromosome seg-regation but is required or parS-mediated plasmid stabiliza-tion (Lin and Grossman, 1998; Marston and Errington, 1999).Consequently, the role o parABS o B. subtilis in chromosomesegregation during vegetative growth appears to be indirect.




Similarly, deletion o parABS in some Gram-negative bacteriaresults in only mild deects in chromosome segregation (God-rin-Estevenon et al., 2002; Lewis et al., 2002). In contrast to

parABS, it is well established that the absence o SMC o B.

subtilis, or the SMC homolog MukB o E. coli, leads to chromo-some decondensation and chromosome segregation deects(Niki et al., 1991; Britton et al., 1998; Moriya et al., 1998). Recentevidence urther supports an indirect role or parABS in chro-mosome segregation via SMC (Gruber and Errington, 2009;Sullivan et al., 2009). These groups show that ParB recruitsSMC to the origin-proximal region o the chromosome via adirect interaction between the two proteins and that this inter-action is involved in chromosome organization and efcientsegregation. Moreover, deletion o parA greatly enhances thechromosome segregation deect o an smc null mutation sug-gesting another mechanistic link between SMC and ParAB thathas yet to be characterized (Lee and Grossman, 2006).

As described above, there is little evidence that ParA playsan essential role in B. subtilis chromosome segregation dur-ing vegetative growth. ParA may play a similarly indirect roleduring sporulation by acting in parallel with RacA-mediatedchromosome segregation (Wu and Errington, 2003). RacA oB. subtilis is a DNA-binding protein crucial or chromosomesegregation during dierentiation (Ben-Yehuda et al., 2003). Itbinds to a 14 bp inverted repeat present at 25 locations o theorigin-proximal region o the chromosome (Ben-Yehuda et al.,2005). RacA is urthermore recruited to the cell pole by DivIVA(Edwards et al., 2000; Ben-Yehuda et al., 2003; Lenarcic etal., 2009), thereby orming an important bridge between thechromosome and the cell pole that is required or chromosomesegregation. The act that deletion o ParA greatly enhances

the deleterious eect o racA mutants on chromosome segre-gation hints at overlapping unctions or ParA and RacA in thisprocess (Wu and Errington, 2003).

Terminus Domain Organization and Segregation

Bacterial chromosome organization and dynamics have beenanalyzed extensively (Niki et al., 2000; Viollier et al., 2004; Valenset al., 2004; Wang et al., 2005, 2006; Nielsen et al., 2006). Asmentioned above, E. coli lacks an obvious parABS locus, rais-ing questions as to how DNA segregation is achieved. The E.

coli chromosome has been divided cytologically and geneti-cally into our macrodomains called the Ori, Ter, Right, and Letmacrodomains, as well as two more loosely defned, nonstruc-tured regions called NSR and NSL, located to the right and let

o the Ori macrodomain. Using P1 parS sites and a uores-cently labeled ParB usion protein, Espeli and coworkers exam-ined how E. coli chromosomal regions segregate ater replica-tion: duplicated markers belonging to macrodomains show acosegregation step, whereas cosegregation is not apparentin the nonstructured regions (Espeli et al., 2008). They reportthat chromosome segregation occurs in three phases: frst, theorigin-proximal hal o the chromosome consisting o the Orimacrodomain and the two nonstructured regions segregateconcomitantly in a short period o time. Second, the Right andLet macrodomains segregate progressively according to thegenetic map. Third, the Ter macrodomain is rapidly segregated

just prior to cell division.

In most cases, the actors that organize the macrodomainsare unknown. However, the system organizing the Ter macro-domain has recently been identifed using an elegant statisticalapproach in which the parS sites o B. subtilis are used as a

test example (Mercier et al., 2008). The reasoning behind theuse o parS is the observation that ParAB and eight parS siteso B. subtilis organize the origin-proximal region into a domainprofcient or segregation. This approach predicted a “macro-domain signature moti” in the Ter macrodomain ( matS ) thatwas ound to be repeated 23 times exclusively in the 800 kblong terC-proximal region o the chromosome. The matS motiswere subsequently used to identiy the matS-binding proteinMatP, which is crucial or organization and proper segregationo the Ter macrodomain (Mercier et al., 2008). In E. coli, the Terregion segregates a ew minutes beore cell division (Li et al.,2003; Bates and Kleckner, 2005; Espeli et al., 2008; Adachi etal., 2008). Thus, ater completion o replication, duplicated Termacrodomains are kept together by MatP close to mid-cell untillate in the cell cycle, suggesting that their segregation might beunder the control o actors that destabilize the colocalizationdependent on MatP. Consistent with this proposal, deletion o

matP results in loss o the Ter macrodomian structure and inpremature segregation o Ter macrodomain markers, whereassegregation o the other macrodomains is unaected. Theseresults point toward an apparatus that actively segregates theTer macrodomain (Mercier et al., 2008), as proposed previouslyby genetic and cytological observations (Niki et al., 2000).

ParAs and Control of Replication Initiation

In addition to their indirect role in chromosome segregation,several studies have implicated the parABS system o B. sub-

tilis in the control o chromosome replication. Intriguingly, thisfnding invalidates the long-accepted hypothesis that par lociunction independently o their native replicon, as frst sug-gested by Ogura and Hiraga (1983) and later reinorced by oth-ers (Austin and Abeles, 1983; Gerdes et al., 1985). parB nullmutant cells have an increased origin and chromosomal copynumber, and they lose synchrony in the initiation o chromo-some replication (Lee et al., 2003; Ogura et al., 2003). A double

parAB mutant possessed similar eatures (Lee and Grossman,2006), with this latter work also showing that parAB mutantsail to separate replicated oriC regions properly.

A recent report rom the Errington group has provided mech-anistic insight into the role played by ParA in replication control:ParA modulates the activity o DnaA, the replication-initiator

actor controlling ormation o the replication initiation com-plex at oriC (Murray and Errington, 2008). Using two classeso mutant derivatives o ParA, these authors showed that oneclass (ParAD40A ) stimulates initiation o replication, whereas asecond class (ParAG12V ) inhibits initiation. Importantly, bothstimulation and inhibition o oriC activity require DnaA. Inhibi-tion o initiation by the ParAG12V is severe but could be com-pletely bypassed by point mutations in dnaA. Cells with ParAG12V-GFP orm DnaA-dependent oci that colocalize with oriC even in the absence o ParB, suggesting a direct interactionbetween ParA and DnaA. This putative in vivo interaction isconfrmed by crosslinking and two-hybrid analysis. Interest-ingly, although ParAG12V interacts strongly with DnaA, ParAD40A




interacts only weakly. This dierence in apparent bindingstrengths suggests that ParA may utilize distinct mechanismsor negative and positive regulation o DnaA (Murray and Err-ington, 2008).

par Loci Encoding Tubulin/FtsZ-like Homologs

Although much is now known about prokaryotic DNA segre-gation, the repertoire o possible segregation mechanisms islikely ar rom exhausted, as highlighted by the recent discov-ery o a unique par locus in Gram-positive bacteria. Numer-ous plasmids rom B. thuringensis (such as pBtoxis) and B.

anthracis (such as pDSW208 and pXO1) encode a cytoskeletalprotein, TubZ, and a small DNA-binding protein, TubR (Lar-sen et al., 2007), that together may constitute a third type o

par locus (Figure 1E). TubZ was initially discovered as a actorrequired or replication o plasmid pXO1 o B. anthracis andwas thereore called RepX (Tinsley and Khan, 2006). However,RepX possesses properties very similar to those o the other

TubZs, as discussed below (Anand et al., 2008).TubZs belong to the tubulin/FtsZ super amiliy o GTPases but

are only distantly related to both tubulin and FtsZ. TubZs o B.

thuringiensis and B. anthracis are again quite distantly related, yetthey exhibit very similar characteristics in vitro (Chen and Erick-son, 2008), including both high GTPase activities and assemblyinto double-stranded helical flaments in vitro (Chen and Erickson,2008). In the presence o GTP, ormation o TubZ flaments wasdynamic (allowing both assembly and disassembly) and substoi-chiometric amounts o a nonhydrolyzable GTP analog stabilizedthe flaments (Chen and Erickson, 2008). This behavior is remi-niscent o eukaryotic tubulin polymers consisting o a GDP corestabilized by a small GTP cap (Erickson and O’Brien, 1992; Desai

and Mitchison, 1997). Furthermore, TubZ o the B. thuringiensis plasmid orms dynamic flaments in vivo that move rapidly alongthe cell membrane in a treadmilling-like pattern (Larsen et al.,2007), again reminiscent o eukaryotic microtubules.

Microtubules exibit dynamic instability, both in vitro (Mitchi-son and Kirschner, 1984) and in vivo (Cassimeris et al., 1988;Sammak and Borisy, 1988; Schulze and Kirschner, 1988). Inliving cells, when the minus ends o microtubules are tightlyanchored at centrosomes, they are thought to exchange sub-units by polymerization/depolymerization at their plus endsvia the dynamic instability mechanism. However, ree micro-tubules can also exhibit a treadmilling behavior both in vitro(Margolis et al., 1978) and in vivo (Rodionov and Borisy, 1997).TubZ exhibits a similar treadmilling behavior both in the pres-

ence and absence o TubR, a protein that may anchor TubZflaments to plasmids (Larsen et al., 2007). It is likely that thetreadmilling behavior o TubZ is critical or plasmid segregationand thus that the tubZR locus represents a third mechanism orprokaryotic plasmid segregation. Moreover, it is possible thatTubZs, like B. subtilis ParA proteins, unction both in plasmidsegregation and replication control. Further investigations othis intriguing system are most defnitely warranted.

Prospects

The present state o the feld raises a number o importantquestions or urther investigation. For the actin-encoding par loci, it would be highly inormative to obtain the crystal struc-

ture o ParM in combination with the ParR/ parC centromerecomplex, which may help to discriminate between the “stair-case” model (Figure 2B) and other possibilities (Popp et al.,2010). An even more ambitious goal would be to reconstitute

in vitro the molecular machineries o both plasmid and chro-mosomal par loci encoding P loop ATPases. A successulreconstitution would allow a rigorous test o the proposedpulling model (Figure 5) that was derived rom a combinationo in vivo data and mathematical modeling (Ringgaard et al.,2009). Crystal structures o ParA bound to nonspecifc DNAare also eagerly awaited, given that these structures couldconfrm the assumed cytoskeletal nature o ParA flaments,as suggested by 3D flament reconstruction (Hui et al., 2010).The unction o the pairing reaction seen with both actin- andParA-encoding par loci should also be addressed, mainly by invitro approaches but, i possible, also in vivo. Finally, it shouldbe within reach to dissect the mechanism o tubZR-mediatedplasmid stabilization.

The picture now emerging is that bacterial plasmids haveevolved dierent segregation mechanisms by recruiting dier-ent proteins belonging to cytoskeletal protein amilies (actin-like, tubulin-like, ParA). These cytoskeletal systems are highlydynamic and exhibit unexpected and intriguing properties thatcan be very dierent rom those o their eukaryotic equivalents.Moreover, these systems ensure stable plasmid transmissionduring cell division by highly diverse mechanisms. Although abasic understanding o some o these systems is now withinreach, much more remains to be discovered. Consequently,the feld o prokaryotic DNA segregation will, in our opinion,remain richly rewarding or many years to come.

ACkNOwLEDGMENTS

We are grateul to S. Gruber, J. Errington, H. Murray, and L. J. Wu or criticalcomments on the manuscript and to the Errington and Gerdes groups orstimulating discussions. This work was supported by the Biotechnology andBiological Sciences Research Council o the UK by a grant to K.G. and bycore support to M.H. M.H. is also supported fnancially by The Royal Society.

REFERENCES

Abeles, A.L., Friedman, S.A., and Austin, S.J. (1985). Partition o unit-copyminiplasmids to daughter cells. III. The DNA sequence and unctional organi-zation o the P1 partition region. J. Mol. Biol. 185, 261–272.

Adachi, S., Hori, K., and Hiraga, S. (2006). Subcellular positioning o F plas-mid mediated by dynamic localization o SopA and SopB. J. Mol. Biol. 356,850–863.

Adachi, S., Fukushima, T., and Hiraga, S. (2008). Dynamic events o sisterchromosomes in the cell cycle o Escherichia coli. Genes Cells 13, 181–197.

Ah-Seng, Y., Lopez, F., Pasta, F., Lane, D., and Bouet, J.Y. (2009). Dual role oDNA in regulating ATP hydrolysis by the SopA partition protein. J. Biol. Chem.

284, 30067–30075.

Anand, S.P., Akhtar, P., Tinsley, E., Watkins, S.C., and Khan, S.A. (2008). GTP-dependent polymerization o the tubulin-like RepX replication protein encod-ed by the pXO1 plasmid o Bacillus anthracis. Mol. Microbiol. 67 , 881–890.

Ausmees, N., Kuhn, J.R., and Jacobs-Wagner, C. (2003). The bacterial cy-toskeleton: an intermediate flament-like unction in cell shape. Cell 115,705–713.

Austin, S., and Abeles, A. (1983). Partition o unit-copy miniplasmids todaughter cells. I. P1 and F miniplasmids contain discrete, interchangeable




sequences sufcient to promote equipartition. J. Mol. Biol. 169, 353–372.

Autret, S., Nair, R., and Errington, J. (2001). Genetic analysis o the chromo-some segregation protein Spo0J o Bacillus subtilis: evidence or separatedomains involved in DNA binding and interactions with Soj protein. Mol. Mi-crobiol. 41, 743–755.

Barillà, D., Rosenberg, M.F., Nobbmann, U., and Hayes, F. (2005). BacterialDNA segregation dynamics mediated by the polymerizing protein ParF. EMBOJ. 24, 1453–1464.

Barillà, D., Carmelo, E., and Hayes, F. (2007). The tail o the ParG DNA segre-gation protein remodels ParF polymers and enhances ATP hydrolysis via anarginine fnger-like moti. Proc. Natl. Acad. Sci. USA 104, 1811–1816.

Bates, D., and Kleckner, N. (2005). Chromosome and replisome dynamics inE. coli: loss o sister cohesion triggers global chromosome movement andmediates chromosome segregation. Cell 121, 899–911.

Batt, S.M., Bingle, L.E., Daorn, T.R., and Thomas, C.M. (2009). Bacterialgenome partitioning: N-terminal domain o IncC protein encoded by broad-host-range plasmid RK2 modulates oligomerisation and DNA binding. J. Mol.Biol. 385, 1361–1374.

Becker, E., Herrera, N.C., Gunderson, F.Q., Derman, A.I., Dance, A.L., Sims,J., Larsen, R.A., and Pogliano, J. (2006). DNA segregation by the bacterialactin AlA during Bacillus subtilis growth and development. EMBO J. 25,5919–5931.

Ben-Yehuda, S., Rudner, D.Z., and Losick, R. (2003). RacA, a bacterial proteinthat anchors chromosomes to the cell poles. Science 299, 532–536.

Ben-Yehuda, S., Fujita, M., Liu, X.S., Gorbatyuk, B., Skoko, D., Yan, J., Marko,J.F., Liu, J .S., Eichenberger, P., Rudner, D.Z., and Losick, R. (2005). Defning acentromere-like element in Bacillus subtilis by Identiying the binding sites orthe chromosome-anchoring protein RacA. Mol. Cell 17 , 773–782.

Bertin, A., McMurray, M.A., Grob, P., Park, S.S., Garcia, G., 3rd, Patanwala, I.,Ng, H.L., Alber, T., Thorner, J., and Nogales, E. (2008). Saccharomyces cerevi-siae septins: supramolecular organization o heterooligomers and the mecha-nism o flament assembly. Proc. Natl. Acad. Sci. USA 105, 8274–8279.

Bi, E.F., and Lutkenhaus, J. (1991). FtsZ ring structure associated with divisionin Escherichia coli. Nature 354, 161–164.

Biek, D.P., and Shi, J. (1994). A single 43-bp sopC repeat o plasmid mini-F issufcient to allow assembly o a unctional nucleoprotein partition complex.Proc. Natl. Acad. Sci. USA 91, 8027–8031.

Bouet, J.Y., and Funnell, B.E. (1999). P1 ParA interacts with the P1 partitioncomplex at parS and an ATP-ADP switch controls ParA activities. EMBO J.18, 1415–1424.

Bouet, J.Y., Bouvier, M., and Lane, D. (2006). Concerted action o plasmidmaintenance unctions: partition complexes create a requirement or dimerresolution. Mol. Microbiol. 62, 1447–1459.

Bouet, J.Y., Ah-Seng, Y., Benmeradi, N., and Lane, D. (2007). Polymerizationo SopA partition ATPase: regulation by DNA binding and SopB. Mol. Micro-biol. 63, 468–481.

Breier, A.M., and Grossman, A.D. (2007). Whole-genome analysis o the chro-mosome partitioning and sporulation protein Spo0J (ParB) reveals spreadingand origin-distal sites on the Bacillus subtilis chromosome. Mol. Microbiol.64, 703–718.

Britton, R.A., Lin, D.C., and Grossman, A.D. (1998). Characterization o aprokaryotic SMC protein involved in chromosome partitioning. Genes Dev.12, 1254–1259.

Campbell, C.S., and Mullins, R.D. (2007). In vivo visualization o type II plas-mid segregation: bacterial actin flaments pushing plasmids. J. Cell Biol. 179,1059–1066.

Carmelo, E., Barillà, D., Golovanov, A.P., Lian, L.Y., Derome, A., and Hayes,F. (2005). The unstructured N-terminal tail o ParG modulates assembly o aquaternary nucleoprotein complex in transcription repression. J. Biol. Chem.

280, 28683–28691.

Cassimeris, L., Pryer, N.K., and Salmon, E.D. (1988). Real-time observationso microtubule dynamic instability in living cells. J. Cell Biol. 107 , 2223–2231.

Castaing, J.P., Bouet, J.Y., and Lane, D. (2008). F plasmid partition depends oninteraction o SopA with non-specifc DNA. Mol. Microbiol. 70, 1000–1011.

Chen, Y., and Erickson, H.P. (2008). In vitro assembly studies o FtsZ/tubulin-like proteins (TubZ) rom Bacillus plasmids: evidence or a capping mecha-nism. J. Biol. Chem. 283, 8102–8109.

Choi, C.L., Claridge, S.A., Garner, E.C., Alivisatos, A.P., and Mullins, R.D.(2008). Protein-nanocrystal conjugates support a single flament polymeriza-tion model in R1 plasmid segregation. J. Biol. Chem. 283, 28081–28086.

Dam, M., and Gerdes, K. (1994). Partitioning o plasmid R1. Ten direct repeatsanking the parA promoter constitute a centromere-like partition site parC,that expresses incompatibility. J. Mol. Biol. 236, 1289–1298.

Daniel, R.A., and Errington, J. (2003). Control o cell morphogenesis in bacte-ria: two distinct ways to make a rod-shaped cell. Cell 113, 767–776.

Davis, M.A., Martin, K.A., and Austin, S.J. (1992). Biochemical activities o theparA partition protein o the P1 plasmid. Mol. Microbiol. 6, 1141–1147.

de Boer, P.A., Crossley, R.E., and Rothfeld, L.I. (1989). A division inhibitorand a topological specifcity actor coded or by the minicell locus determineproper placement o the division septum in E. coli. Cell 56, 641–649.

Desai, A., and Mitchison, T.J. (1997). Microtubule polymerization dynamics. Annu. Rev. Cell Dev. Biol. 13, 83–117.

Doubrovinski, K., and Howard, M. (2005). Stochastic model or Soj relocationdynamics in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 102, 9808–9813.

Dunham, T.D., Xu, W., Funnell, B.E., and Schumacher, M.A. (2009). Structuralbasis or ADP-mediated transcriptional regulation by P1 and P7 ParA. EMBOJ. 28, 1792–1802.

Ebersbach, G., and Gerdes, K. (2001). The double par locus o virulence ac-tor pB171: DNA segregation is correlated with oscillation o ParA. Proc. Natl.

Acad. Sci. USA 98, 15078–15083.

Ebersbach, G., and Gerdes, K. (2004). Bacterial mitosis: partitioning proteinParA oscillates in spiral-shaped structures and positions plasmids at mid-cell.Mol. Microbiol. 52, 385–398.

Ebersbach, G., Sherratt, D.J., and Gerdes, K. (2005). Partition-associated in-compatibility caused by random assortment o pure plasmid clusters. Mol.Microbiol. 56, 1430–1440.

Ebersbach, G., Ringgaard, S., Møller-Jensen, J., Wang, Q., Sherratt, D.J.,and Gerdes, K. (2006). Regular cellular distribution o plasmids by oscillat-ing and flament-orming ParA ATPase o plasmid pB171. Mol. Microbiol. 61,1428–1442.

Edgar, R., Chattoraj, D.K., and Yarmolinsky, M. (2001). Pairing o P1 plasmidpartition sites by ParB. Mol. Microbiol. 42, 1363–1370.

Edwards, D.H., Thomaides, H.B., and Errington, J. (2000). Promiscuous tar-geting o Bacillus subtilis cell division protein DivIVA to division sites in Es-

cherichia coli and fssion yeast. EMBO J. 19, 2719–2727.Eremov, A., Grishchuk, E.L., McIntosh, J.R., and Ataullakhanov, F.I. (2007). Insearch o an optimal ring to couple microtubule depolymerization to proces-sive chromosome motions. Proc. Natl. Acad. Sci. USA 104, 19017–19022.

Erickson, H.P., and O’Brien, E.T. (1992). Microtubule dynamic instability andGTP hydrolysis. Annu. Rev. Biophys. Biomol. Struct. 21, 145–166.

Espeli, O., Mercier, R., and Boccard, F. (2008). DNA dynamics vary accordingto macrodomain topography in the E. coli chromosome. Mol. Microbiol. 68,1418–1427.

Fletcher, D.A., and Mullins, R.D. (2010). Cell mechanics and the cytoskeleton.Nature 463, 485–492.

Fogel, M.A., and Waldor, M.K. (2005). Distinct segregation dynamics o thetwo Vibrio cholerae chromosomes. Mol. Microbiol. 55, 125–136.




Fogel, M.A., and Waldor, M.K. (2006). A dynamic, mitotic-like mechanism orbacterial chromosome segregation. Genes Dev. 20, 3269–3282.

Friedman, S.A., and Austin, S.J. (1988). The P1 plasmid-partition system syn-thesizes two essential proteins rom an autoregulated operon. Plasmid 19,103–112.

Galkin, V.E., Orlova, A., Rivera, C., Mullins, R.D., and Egelman, E.H. (2009).Structural polymorphism o the ParM flament and dynamic instability. Struc-ture 17 , 1253–1264.

Garner, E.C., Campbell, C.S., and Mullins, R.D. (2004). Dynamic instability in aDNA-segregating prokaryotic actin homolog. Science 306, 1021–1025.

Garner, E.C., Campbell, C.S., Weibel, D.B., and Mullins, R.D. (2007). Reconsti-tution o DNA segregation driven by assembly o a prokaryotic actin homolog.Science 315, 1270–1274.

Gerdes, K., and Molin, S. (1986). Partitioning o plasmid R1. Structural andunctional analysis o the parA locus. J. Mol. Biol. 190, 269–279.

Gerdes, K., Larsen, J.E., and Molin, S. (1985). Stable inheritance o plasmidR1 requires two dierent loci. J. Bacteriol. 161, 292–298.

Gerdes, K., Møller-Jensen, J., and Bugge Jensen, R. (2000). Plasmid andchromosome partitioning: surprises rom phylogeny. Mol. Microbiol. 37 ,455–466.

Glaser, P., Sharpe, M.E., Raether, B., Perego, M., Ohlsen, K., and Errington, J.(1997). Dynamic, mitotic-like behavior o a bacterial protein required or ac-curate chromosome partitioning. Genes Dev. 11, 1160–1168.

Godrin-Estevenon, A.M., Pasta, F., and Lane, D. (2002). The parAB geneproducts o Pseudomonas putida exhibit partition activity in both P. putidaand Escherichia coli. Mol. Microbiol. 43, 39–49.

Gordon, G.S., Sitnikov, D., Webb, C.D., Teleman, A., Straight, A., Losick,R., Murray, A.W., and Wright, A. (1997). Chromosome and low copy plas-mid segregation in E. coli: visual evidence or distinct mechanisms. Cell 90,1113–1121.

Grishchuk, E.L., Eremov, A.K., Volkov, V.A., Spiridonov, I.S., Gudimchuk, N.,Westermann, S., Drubin, D., Barnes, G., McIntosh, J.R., and Ataullakhanov,

F.I. (2008). The Dam1 ring binds microtubules strongly enough to be a proces-sive as well as energy-efcient coupler or chromosome motion. Proc. Natl.

Acad. Sci. USA 105, 15423–15428.

Gruber, S., and Errington, J. (2009). Recruitment o condensin to replicationorigin regions by ParB/SpoOJ promotes chromosome segregation in B. sub-tilis. Cell 137 , 685–696.

Hatano, T., Yamaichi, Y., and Niki, H. (2007). Oscillating ocus o SopA associ-ated with flamentous structure guides partitioning o F plasmid. Mol. Micro-biol. 64, 1198–1213.

Hayes, F. (2000). The partition system o multidrug resistance plasmid TP228includes a novel protein that epitomizes an evolutionarily distinct subgroup othe ParA superamily. Mol. Microbiol. 37 , 528–541.

Hester, C.M., and Lutkenhaus, J. (2007). Soj (ParA) DNA binding is mediatedby conserved arginines and is essential or plasmid segregation. Proc. Natl.

Acad. Sci. USA 104, 20326–20331.Higuchi, T., and Uhlmann, F. (2005). Stabilization o microtubule dynamics atanaphase onset promotes chromosome segregation. Nature 433, 171–176.

Hirano, M., Mori, H., Onogi, T., Yamazoe, M., Niki, H., Ogura, T., and Hiraga,S. (1998). Autoregulation o the partition genes o the mini-F plasmid and theintracellular localization o their products in Escherichia coli. Mol. Gen. Genet.

257 , 392–403.

Howard, M., and Kruse, K. (2005). Cellular organization by sel-organiza-tion: mechanisms and models or Min protein dynamics. J. Cell Biol. 168,533–536.

Howard, M., Rutenberg, A.D., and de Vet, S. (2001). Dynamic compartmental-ization o bacteria: accurate division in E. coli. Phys. Rev. Lett. 87 , 278102.

Hu, Z., and Lutkenhaus, J. (2003). A conserved sequence at the C-terminus

o MinD is required or binding to the membrane and targeting MinC to theseptum. Mol. Microbiol. 47 , 345–355.

Hu, Z., Gogol, E.P., and Lutkenhaus, J. (2002). Dynamic assembly o MinD onphospholipid vesicles regulated by ATP and MinE. Proc. Natl. Acad. Sci. USA99, 6761–6766.

Hui, M.P., Galkin, V.E., Yu, X., Stasiak, A.Z., Stasiak, A., Waldor, M.K., andEgelman, E.H. (2010). ParA2, a Vibrio cholerae chromosome partitioning pro-tein, orms let-handed helical flaments on DNA. Proc. Natl. Acad. Sci. USA107 , 4590–4595.

Ireton, K., Gunther, N.W., 4th, and Grossman, A.D. (1994). spo0J is requiredor normal chromosome segregation as well as the initiation o sporulation inBacillus subtilis. J. Bacteriol. 176, 5320–5329.

Jakimowicz, D., Zydek, P., Kois, A., Zakrzewska-Czerwińska, J., and Chater,K.F. (2007). Alignment o multiple chromosomes along helical ParA scaoldingin sporulating Streptomyces hyphae. Mol. Microbiol. 65, 625–641.

Jensen, R.B., and Gerdes, K. (1997). Partitioning o plasmid R1. The ParMprotein exhibits ATPase activity and interacts with the centromere-like ParR-parC complex. J. Mol. Biol. 269, 505–513.

Jensen, R.B., and Gerdes, K. (1999). Mechanism o DNA segregation inprokaryotes: ParM partitioning protein o plasmid R1 co-localizes with itsreplicon during the cell cycle. EMBO J. 18, 4076–4084.

Jensen, R.B., Dam, M., and Gerdes, K. (1994). Partitioning o plasmid R1. TheparA operon is autoregulated by ParR and its transcription is highly stimulatedby a downstream activating element. J. Mol. Biol. 236, 1299–1309.

Jensen, R.B., Lurz, R., and Gerdes, K. (1998). Mechanism o DNA segregationin prokaryotes: replicon pairing by parC o plasmid R1. Proc. Natl. Acad. Sci.USA 95, 8550–8555.

Jones, L.J., Carballido-López, R., and Errington, J. (2001). Control o cellshape in bacteria: helical, actin-like flaments in Bacillus subtilis. Cell 104,913–922.

Kabsch, W., and Holmes, K.C. (1995). The actin old. FASEB J. 9, 167–174.

Komeili, A., Li, Z., Newman, D.K., and Jensen, G.J. (2006). Magnetosomes

are cell membrane invaginations organized by the actin-like protein MamK.Science 311, 242–245.

Koonin, E.V., Aravind, L., and Kondrashov, A.S. (2000). The impact o com-parative genomics on our understanding o evolution. Cell 101, 573–576.

Lackner, L.L., Raskin, D.M., and de Boer, P.A. (2003). ATP-dependent interac-tions between Escherichia coli Min proteins and the phospholipid membranein vitro. J. Bacteriol. 185, 735–749.

Larsen, R.A., Cusumano, C., Fujioka, A., Lim-Fong, G., Patterson, P., andPogliano, J. (2007). Treadmilling o a prokaryotic tubulin-like protein, TubZ,required or plasmid stability in Bacillus thuringiensis. Genes Dev. 21,1340–1352.

Le, Q.B., and Ghigo, J.M. (2009). BcsQ is an essential component o the Es-cherichia coli cellulose biosynthesis apparatus that localizes at the bacte-rial cell pole. Mol. Microbiol. Published online April 8, 2009. 10.1111/j.1365-

2958.2009.06678.x.Lee, P.S., and Grossman, A.D. (2006). The chromosome partitioning proteinsSoj (ParA) and Spo0J (ParB) contribute to accurate chromosome partitioning,separation o replicated sister origins, and regulation o replication initiation inBacillus subtilis. Mol. Microbiol. 60, 853–869.

Lee, P.S., Lin, D.C., Moriya, S., and Grossman, A.D. (2003). Eects o thechromosome partitioning protein Spo0J (ParB) on oriC positioning and repli-cation initiation in Bacillus subtilis. J. Bacteriol. 185, 1326–1337.

Lenarcic, R., Halbedel, S., Visser, L., Shaw, M., Wu, L.J., Errington, J., Mar-enduzzo, D., and Hamoen, L.W. (2009). Localisation o DivIVA by targeting tonegatively curved membranes. EMBO J. 28, 2272–2282.

Leonard, T.A., Butler, P.J., and Löwe, J. (2005). Bacterial chromosome segre-gation: structure and DNA binding o the Soj dimer—a conserved biologicalswitch. EMBO J. 24, 270–282.




Lewis, R.A., Bignell, C.R., Zeng, W., Jones, A.C., and Thomas, C.M. (2002).Chromosome loss rom par mutants o Pseudomonas putida depends ongrowth medium and phase o growth. Microbiology 148, 537–548.

Li, Z., and Jensen, G.J. (2009). Electron cryotomography: a new view intomicrobial ultrastructure. Curr. Opin. Microbiol. 12, 333–340.

Li, Y., Youngren, B., Sergueev, K., and Austin, S. (2003). Segregation o theEscherichia coli chromosome terminus. Mol. Microbiol. 50, 825–834.

Lim, G.E., Derman, A.I., and Pogliano, J. (2005). Bacterial DNA segregation bydynamic SopA polymers. Proc. Natl. Acad. Sci. USA 102, 17658–17663.

Lin, D.C., and Grossman, A.D. (1998). Identifcation and characterization o abacterial chromosome partitioning site. Cell 92, 675–685.

Livny, J., Yamaichi, Y., and Waldor, M.K. (2007). Distribution o centromere-like parS sites in bacteria: insights rom comparative genomics. J. Bacteriol.189, 8693–8703.

Löwe, J., and Amos, L.A. (2009). Evolution o cytomotive flaments: the cy-toskeleton rom prokaryotes to eukaryotes. Int. J. Biochem. Cell Biol. 41,323–329.

Lutkenhaus, J. (2007). Assembly dynamics o the bacterial MinCDE systemand spatial regulation o the Z ring. Annu. Rev. Biochem. 76, 539–562.

Ma, L., King, G.F., and Rothfeld, L. (2004). Positioning o the MinE bindingsite on the MinD surace suggests a plausible mechanism or activation o theEscherichia coli MinD ATPase during division site selection. Mol. Microbiol.54, 99–108.

Margolis, R.L., Wilson, L., and Keier, B.I. (1978). Mitotic mechanism based onintrinsic microtubule behaviour. Nature 272, 450–452.

Marston, A.L., and Errington, J. (1999). Dynamic movement o the ParA-likeSoj protein o B. subtilis and its dual role in nucleoid organization and devel-opmental regulation. Mol. Cell 4, 673–682.

McIntosh, J.R., Grishchuk, E.L., Morphew, M.K., Eremov, A.K., Zhudenkov,K., Volkov, V.A., Cheeseman, I.M., Desai, A., Mastronarde, D.N., and Ataul-lakhanov, F.I. (2008). Fibrils connect microtubule tips with kinetochores: amechanism to couple tubulin dynamics to chromosome motion. Cell 135,322–333.

McMurray, M.A., and Thorner, J. (2009). Septins: molecular partitioning andthe generation o cellular asymmetry. Cell Div. 4, 18.

Meinhardt, H., and de Boer, P.A. (2001). Pattern ormation in Escherichia coli:a model or the pole-to-pole oscillations o Min proteins and the localization othe division site. Proc. Natl. Acad. Sci. USA 98, 14202–14207.

Mercier, R., Petit, M.A., Schbath, S., Robin, S., El Karoui, M., Boccard, F., andEspéli, O. (2008). The MatP/matS site-specifc system organizes the terminusregion o the E. coli chromosome into a macrodomain. Cell 135, 475–485.

Michie, K.A., and Löwe, J. (2006). Dynamic flaments o the bacterial cytoskel-eton. Annu. Rev. Biochem. 75, 467–492.

Mitchison, T., and Kirschner, M. (1984). Dynamic instability o microtubulegrowth. Nature 312, 237–242.

Møller-Jensen, J., Jensen, R.B., Löwe, J., and Gerdes, K. (2002). ProkaryoticDNA segregation by an actin-like flament. EMBO J. 21, 3119–3127.

Møller-Jensen, J., Borch, J., Dam, M., Jensen, R.B., Roepstor, P., and Gerdes,K. (2003). Bacterial mitosis: ParM o plasmid R1 moves plasmid DNA by anactin-like insertional polymerization mechanism. Mol. Cell 12, 1477–1487.

Møller-Jensen, J., Ringgaard, S., Mercogliano, C.P., Gerdes, K., and Löwe, J.(2007). Structural analysis o the ParR/parC plasmid partition complex. EMBOJ. 26, 4413–4422.

Moriya, S., Tsujikawa, E., Hassan, A.K., Asai, K., Kodama, T., and Ogasawara,N. (1998). A Bacillus subtilis gene-encoding protein homologous to eukaryoticSMC motor protein is necessary or chromosome partition. Mol. Microbiol.

29, 179–187.

Murray, H., and Errington, J. (2008). Dynamic control o the DNA replication

initiation protein DnaA by Soj/ParA. Cell 135, 74–84.

Murray, H., Ferreira, H., and Errington, J. (2006). The bacterial chromosomesegregation protein Spo0J spreads along DNA rom parS nucleation sites.Mol. Microbiol. 61, 1352–1361.

Nielsen, H.J., Ottesen, J.R., Youngren, B., Austin, S.J., and Hansen, F.G.(2006). The Escherichia coli chromosome is organized with the let and rightchromosome arms in separate cell halves. Mol. Microbiol. 62, 331–338.

Niki, H., and Hiraga, S. (1997). Subcellular distribution o actively partitioningF plasmid during the cell division cycle in E. coli. Cell 90, 951–957.

Niki, H., Jaé, A., Imamura, R., Ogura, T., and Hiraga, S. (1991). The new genemukB codes or a 177 kd protein with coiled-coil domains involved in chromo-some partitioning o E. coli. EMBO J. 10, 183–193.

Niki, H., Yamaichi, Y., and Hiraga, S. (2000). Dynamic organization o chromo-somal DNA in Escherichia coli. Genes Dev. 14, 212–223.

Ogura, T., and Hiraga, S. (1983). Partition mechanism o F plasmid: two plas-mid gene-encoded products and a cis-acting region are involved in partition.Cell 32, 351–360.

Ogura, Y., Ogasawara, N., Harry, E.J., and Moriya, S. (2003). Increasing theratio o Soj to Spo0J promotes replication initiation in Bacillus subtilis. J. Bac-teriol. 185, 6316–6324.

Orlova, A., Garner, E.C., Galkin, V.E., Heuser, J., Mullins, R.D., and Egelman,E.H. (2007). The structure o bacterial ParM flaments. Nat. Struct. Mol. Biol.14, 921–926.

Polka, J.K., Kollman, J.M., Agard, D.A., and Mullins, R.D. (2009). The structureand assembly dynamics o plasmid actin AlA imply a novel mechanism oDNA segregation. J. Bacteriol. 191, 6219–6230.

Pollard, T.D., and Cooper, J.A. (2009). Actin, a central player in cell shape andmovement. Science 326, 1208–1212.

Popp, D., Yamamoto, A., Iwasa, M., Narita, A., Maeda, K., and Maéda, Y.(2007). Concerning the dynamic instability o actin homolog ParM. Biochem.Biophys. Res. Commun. 353, 109–114.

Popp, D., Narita, A., Oda, T., Fujisawa, T., Matsuo, H., Nitanai, Y., Iwasa, M.,Maeda, K., Onishi, H., and Maéda, Y. (2008). Molecular structure o the ParMpolymer and the mechanism leading to its nucleotide-driven dynamic instabil-ity. EMBO J. 27 , 570–579.

Popp, D., Xu, W., Narita, A., Brzoska, A.J., Skurray, R.A., Firth, N., Goshdas-tider, U., Maéda, Y., Robinson, R.C., and Schumacher, M.A. (2010). Structureand flament dynamics o the pSK41 actin-like ParM protein: implications orplasmid DNA segregation. J. Biol. Chem. 285, 10130–10140.

Pratto, F., Cicek, A., Weihoen, W.A., Lurz, R., Saenger, W., and Alonso, J.C.(2008). Streptococcus pyogenes pSM19035 requires dynamic assembly o

ATP-bound ParA and ParB on parS DNA during plasmid segregation. Nucleic Acids Res. 36, 3676–3689.

Pratto, F., Suzuki, Y., Takeyasu, K., and Alonso, J.C. (2009). Single-moleculeanalysis o proteinxDNA complexes ormed during partition o newly repli-cated plasmid molecules in Streptococcus pyogenes. J. Biol. Chem. 284,30298–30306.

Quisel, J.D., Lin, D.C., and Grossman, A.D. (1999). Control o developmentby altered localization o a transcription actor in B. subtilis. Mol. Cell 4,665–672.

Radnedge, L., Youngren, B., Davis, M., and Austin, S. (1998). Probing the struc-ture o complex macromolecular interactions by homolog specifcity scan-ning: the P1 and P7 plasmid partition systems. EMBO J. 17 , 6076–6085.

Raskin, D.M., and de Boer, P.A. (1999). MinDE-dependent pole-to-pole os-cillation o division inhibitor MinC in Escherichia coli. J. Bacteriol. 181,6419–6424.

Ringgaard, S., Ebersbach, G., Borch, J., and Gerdes, K. (2007a). Regula-tory cross-talk in the double par locus o plasmid pB171. J. Biol. Chem. 282,3134–3145.




Ringgaard, S., Löwe, J., and Gerdes, K. (2007b). Centromere pairing by aplasmid-encoded type I ParB protein. J. Biol. Chem. 282, 28216–28225.

Ringgaard, S., van Zon, J., Howard, M., and Gerdes, K. (2009). Movement andequipositioning o plasmids by ParA flament disassembly. Proc. Natl. Acad.Sci. USA 106, 19369–19374.

Rodionov, V.I., and Borisy, G.G. (1997). Microtubule treadmilling in vivo. Sci-ence 275, 215–218.

Rodionov, O., Lobocka, M., and Yarmolinsky, M. (1999). Silencing o genesanking the P1 plasmid centromere. Science 283, 546–549.

Salje, J., and Löwe, J. (2008). Bacterial actin: architecture o the ParMRCplasmid DNA partitioning complex. EMBO J. 27 , 2230–2238.

Salje, J., Zuber, B., and Löwe, J. (2009). Electron cryomicroscopy o E. colireveals flament bundles involved in plasmid DNA segregation. Science 323,509–512.

Sammak, P.J., and Borisy, G.G. (1988). Direct observation o microtubule dy-namics in living cells. Nature 332, 724–726.

Savage D.F., Aonso B., Chen A.H., and Silver P.A.. (2010). Spatially or-dered dynamics o the bacterial carbon fxation machinery. Science 327 ,1258–1261.

Schulze, E., and Kirschner, M. (1988). New eatures o microtubule behaviourobserved in vivo. Nature 334, 356–359.

Schumacher, M.A. (2007). Structural biology o plasmid segregation proteins.Curr. Opin. Struct. Biol. 17 , 103–109.

Schumacher, M.A. (2008). Structural biology o plasmid partition: uncoveringthe molecular mechanisms o DNA segregation. Biochem. J. 412, 1–18.

Schumacher, M.A., and Funnell, B.E. (2005). Structures o ParB bound to DNAreveal mechanism o partition complex ormation. Nature 438, 516–519.

Schumacher, M.A., Glover, T.C., Brzoska, A.J., Jensen, S.O., Dunham, T.D.,Skurray, R.A., and Firth, N. (2007). Segrosome structure revealed by a com-plex o ParR with centromere DNA. Nature 450, 1268–1271.

Sengupta, M., Nielsen, H.J., Youngren, B., and Austin, S. (2010). P1 plasmidsegregation: accurate redistribution by dynamic plasmid pairing and separa-tion. J. Bacteriol. 192, 1175–1183.

Sharpe, M.E., and Errington, J. (1996). The Bacillus subtilis soj-spo0J locusis required or a centromere-like unction involved in prespore chromosomepartitioning. Mol. Microbiol. 21, 501–509.

Shih, Y.L., Le, T., and Rothfeld, L. (2003). Division site selection in Escheri-chia coli involves dynamic redistribution o Min proteins within coiled struc-tures that extend between the two cell poles. Proc. Natl. Acad. Sci. USA 100,7865–7870.

Sullivan, N.L., Marquis, K.A., and Rudner, D.Z. (2009). Recruitment o SMCby ParB-parS organizes the origin region and promotes efcient chromosomesegregation. Cell 137 , 697–707.

Thompson, S.R., Wadhams, G.H., and Armitage, J.P. (2006). The positioning

o cytoplasmic protein clusters in bacteria. Proc. Natl. Acad. Sci. USA 103,8209–8214.

Tinsley, E., and Khan, S.A. (2006). A novel FtsZ-like protein is involved in rep-lication o the anthrax toxin-encoding pXO1 plasmid in Bacillus anthracis. J.Bacteriol. 188, 2829–2835.

Toro, E., Hong, S.H., McAdams, H.H., and Shapiro, L. (2008). Caulobacterrequires a dedicated mechanism to initiate chromosome segregation. Proc.Natl. Acad. Sci. USA 105, 15435–15440.

Valens, M., Penaud, S., Rossignol, M., Cornet, F., and Boccard, F. (2004).Macrodomain organization o the Escherichia coli chromosome. EMBO J. 23,4330–4341.

van den Ent, F., Møller-Jensen, J., Amos, L.A., Gerdes, K., and Löwe, J.(2002). F-actin-like flaments ormed by plasmid segregation protein ParM.EMBO J. 21, 6935–6943.

Viollier, P.H., Thanbichler, M., McGrath, P.T., West, L., Meewan, M., McAdams,H.H., and Shapiro, L. (2004). Rapid and sequential movement o individualchromosomal loci to specifc subcellular locations during bacterial DNA repli-cation. Proc. Natl. Acad. Sci. USA 101, 9257–9262.

Wan, X., O’Quinn, R.P., Pierce, H.L., Joglekar, A.P., Gall, W.E., DeLuca, J.G.,Carroll, C.W., Liu, S.T., Yen, T.J., McEwen, B.F., et al. (2009). Protein archi-tecture o the human kinetochore microtubule attachment site. Cell 137 ,672–684.

Wang, X., Possoz, C., and Sherratt, D.J. (2005). Dancing around the divisome:asymmetric chromosome segregation in Escherichia coli. Genes Dev. 19,2367–2377.

Wang, X., Liu, X., Possoz, C., and Sherratt, D.J. (2006). The two Escheri-chia coli chromosome arms locate to separate cell halves. Genes Dev. 20,1727–1731.

Wu, L.J., and Errington, J. (2003). RacA and the Soj-Spo0J system combineto eect polar chromosome segregation in sporulating Bacillus subtilis. Mol.Microbiol. 49, 1463–1475.

Yamaichi, Y., and Niki, H. (2000). Active segregation by the Bacillus sub-tilis partitioning system in Escherichia coli. Proc. Natl. Acad. Sci. USA 97 ,14656–14661.

Yamaichi, Y., Fogel, M.A., McLeod, S.M., Hui, M.P., and Waldor, M.K. (2007).Distinct centromere-like parS sites on the two chromosomes o Vibrio spp. J.Bacteriol. 189, 5314–5324.

pushing and pulling in prokaryotic dna segregation

Documents