The actin-like MreB cytoskeleton organizes viral DNAreplication in bacteriaDaniel Munoz-Espína,b, Richard Danielb,c, Yoshikazu Kawaic, Rut Carballido-Lopezd, Virginia Castilla-Llorentea,Jeff Erringtonb,c,1,2, Wilfried J. J. Meijera,1, and Margarita Salasa,1,2

aInstituto de Biología Molecular ‘‘Eladio Vinuela’’ and Centro de Biología Molecular ‘‘Severo Ochoa,’’ Consejo Superior de InvestigacionesCientificas-Universidad Autonoma de Madrid, Canto Blanco, 28049 Madrid, Spain; bSir William Dunn School of Pathology, University of Oxford, South ParksRoad, Oxford OX1 3RE, United Kingdom; cInstitute for Cell and Molecular Biosciences, Newcastle University, Newcastle-upon-Tyne NE2 4HH, UnitedKingdom; and dGenetique Microbienne, Institut National de la Recherche Agronomique, 78352 Jouy-en Josas Cedex, France

Contributed by Margarita Salas, June 12, 2009 (sent for review April 20, 2009)

Little is known about the organization or proteins involved inmembrane-associated replication of prokaryotic genomes. Herewe show that the actin-like MreB cytoskeleton of the distantlyrelated bacteria Escherichia coli and Bacillus subtilis is required forefficient viral DNA replication. Detailed analyses of B. subtilisphage �29 showed that the MreB cytoskeleton plays a crucial rolein organizing phage DNA replication at the membrane. Thus,phage double-stranded DNA and components of the �29 replica-tion machinery localize in peripheral helix-like structures in acytoskeleton-dependent way. Importantly, we show that MreBinteracts directly with the �29 membrane-protein p16.7, respon-sible for attaching viral DNA at the cell membrane. Altogether, theresults reveal another function for the MreB cytoskeleton anddescribe a mechanism by which viral DNA replication is organizedat the bacterial membrane.

Bacillus subtilis � phage �29

Genes of the mreB family encode homologues of eukaryoticactin (1, 2) that form a cytoskeleton in most non-spherical

bacteria (3–6). MreB proteins form filamentous structures fol-lowing a helical path around the inner surface of the cytoplasmicmembrane (1). These actin-like filaments are continuously re-modelled during cell-cycle progression (7–11). Evidence is ac-cumulating that the bacterial MreB cytoskeleton plays key rolesin several important cellular processes such as cell shape deter-mination, chromosome segregation, and cell polarity (1, 3, 8,12–16). Whereas Gram-negative bacteria have a single mreBgene, Gram-positive bacteria often have multiple mreB homo-logues. Bacillus subtilis encodes 3 MreB isoforms: MreB, Mbl,and MreBH (17–19).

For decades, evidence has been provided that replication ofphage DNA, like that of other prokaryotic genomes, occurs atthe cytoplasmic membrane (for review see 20). However, little isknown about the proteins or their organization in membrane-associated replication of viral genomes in bacteria. Phages �29and SPP1 infect the Gram-positive bacterium B. subtilis, andphage PRD1 infects the Gram-negative bacterium Escherichiacoli. Whereas PRD1 and �29 use the protein-primed mechanismof DNA replication, phage SPP1 replicates its DNA initially viathe theta mode and later via a rolling circle mode [reviewed in(21)]. Here we show a key role for the MreB cytoskeleton inphages replicating by different modes in the distantly relatedbacteria E. coli and B. subtilis. Thus, the efficiency of replicationof phage PRD1, and that of phages SPP1 and �29, is severelyaffected in the absence of an intact cytoskeleton.

The underlying mechanism by which the cytoskeleton leads toefficient phage DNA replication was analyzed in detail for B.subtilis phage �29, whose DNA replication has been well char-acterized in vitro. The �29 genome consists of a linear double-stranded DNA (dsDNA) with a terminal protein (TP) covalentlylinked at each 5� end that is the primer for the initiation of phageDNA replication. Hence, initiation of �29 DNA replication

occurs via a so-called protein-primed mechanism (22, see Fig.S1). Phage �29 DNA transcription is divided into early and latestages (see Fig. S1 for a genetic and a transcriptional map).Genes encoding DNA replication proteins such as DNA poly-merase (p2) and TP (p3) are located in the left-side early operon.The right-side early operon contains gene 16.7 that encodes amembrane protein (p16.7) required for optimal in vivo �29 DNAreplication (23, 24). Additional functional, biochemical andstructural studies have provided strong evidence that p16.7 isresponsible for attaching �29 DNA to the bacterial membrane(24–26). Crystallographic resolution of the p16.7 DNA bindingdomain (p16.7C) in complex with dsDNA revealed that 1dsDNA binding unit is formed by 3 p16.7C dimers that arearranged in such a way that they form a deep positively chargedlongitudinal cavity that interacts with the phosphate backbone ofdsDNA (27).

Here we have analyzed the subcellular localization of com-ponents of the phage �29 replication machinery and found that�29 DNA polymerase, protein p16.7, and replicating �29dsDNA localize in a helix-like pattern near the membrane ofinfected B. subtilis cells. In addition, this helical organizationdepends on all 3 host-encoded MreB proteins. Moreover, weshow that MreB interacts directly with p16.7 in vivo. A modelintegrating these results is discussed.

ResultsEfficient Phage DNA Replication Requires an Intact Bacterial MreBCytoskeleton. B. subtilis mreB mutant strains can be propagatedwith near wild-type growth rate and cell morphology in growthmedia supplemented with high concentrations of magnesium(28). Therefore, when cytoskeleton mutant strains were used inthis work, media were supplemented with 25 mM MgSO4.

To examine a possible role of the bacterial actin-like cytoskel-eton in viral DNA replication we determined the efficiency ofDNA replication of phages �29, SPP1, and PRD1 in infectedwild-type and mreB mutant cells. �29 and SPP1 infect theGram-positive bacterium B. subtilis but use different modes ofDNA replication [reviewed in (21)]. In contrast, PRD1 uses asimilar DNA replication mechanism as �29 but infects theGram-negative bacterium E. coli (reviewed in 21). Fig. 1 showsthat the efficiency of DNA replication of these phages wasseverely affected in the absence of an intact MreB cytoskeleton.In the case of �29, deletion of any of the 3 mreB-like genes

Author contributions: D.M.-E., R.D., J.E., W.J.J.M., and M.S. designed research; D.M.-E. andY.K. performed research; V.C.-L. contributed new reagents/analytic tools; D.M.-E., R.D.,R.C.-L., J.E., W.J.J.M., and M.S. analyzed data; and D.M.-E., J.E., W.J.J.M., and M.S. wrote thepaper.

The authors declare no conflict of interest.

1J.E., W.J.J.M., and M.S. contributed equally to this work.

2To whom correspondence may be addressed. E-mail: msalas@cbm.uam.es orjeff.errington@ncl.ac.uk.

This article contains supporting information online at www.pnas.org/cgi/content/full/0906465106/DCSupplemental.

www.pnas.org�cgi�doi�10.1073�pnas.0906465106 PNAS � August 11, 2009 � vol. 106 � no. 32 � 13347–13352

CELL

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

May

14,

202

1

caused a similar deleterious effect on DNA replication. Analysesof phage DNA accumulation by agarose gel electrophoresisshowed that intracellular phage DNA was detected early afterinfection of both wild-type and mutant cells (see Fig. S2),indicating that the cytoskeleton mutations have no or little effecton phage DNA injection. The results show that the bacterialMreB cytoskeleton is required for efficient phage DNA repli-cation in Gram-positive and Gram-negative bacteria, regardlessof the phage DNA replication mechanism.

The role of the cytoskeleton in viral DNA replication wasstudied in detail for B. subtilis phage �29. To get further evidencethat an intact cytoskeleton is required for efficient �29 DNAreplication, phage DNA synthesis was studied using the B. subtilisstrain 2060, which contains a disrupted mreB gene and axylose-inducible copy of c-myc-mreB. The results presented inFig. S3 show that the amount of phage DNA increased rapidlywhen the cells were grown in the presence of xylose. In contrast,the efficiency of �29 DNA replication was low in the absence ofthe inductor. Importantly, efficient phage DNA replication wasrestored when xylose was added to the conditional strain 30 minafter infection, confirming that MreB plays a role in phage DNAreplication.

�29 DNA Polymerase Localizes in a Helix-Like Pattern at the Peripheryof Infected Cells During �29 DNA Replication. To determine thesubcellular distribution of the �29 replication machinery we firststudied the localization of the gene 2-encoded �29 DNA poly-merase fused to the green fluorescent protein (GFP) in bothinfected and non-infected live cells. For this, B. subtilis strainswere engineered that contain N- or C-terminal xylose-induciblefusions of gfp to �29 gene 2. Complementation experimentsshowed that both the N- (GFP-p2) and C- (p2-GFP) terminalfusion proteins were functional (Fig. S4).

As shown in Fig. 2 (D, G), GFP-p2 was distributed uniformlyin xylose-induced non-infected cells. Interestingly, between 20and 50 min postinfection, corresponding to the period of effi-cient �29 DNA replication, GFP-p2 localized in helix-like struc-tures (Fig. 2 E and F) close to the membrane of infected cells(Fig. 2 H and I).

�29 Membrane-Protein p16.7 and �29 dsDNA Localize in Helix-LikePatterns. Using immunofluorescence (IF) techniques, we nextstudied the subcelullar localization of protein p16.7 by decon-volving optical sections through the z axis of the cell. Fig. 2 (J–N)shows phase contrast and IF images of �29-infected cells.Overlay of phase contrast and deconvolved IF images (Fig. 2M)showed that p16.7 localized in peripheral helix-like patterns that

in most cases spanned the entire length of the infected cells.Three-dimensional reconstruction of a set of deconvolved Zsections demonstrated that p16.7 forms helical structures at themembrane of infected cells (see SI Text and Movie S1). To studythe localization pattern of p16.7 in live cells, a B. subtilis strain

Fig. 1. Efficient viral DNA replication requires an intact MreB cytoskeleton. The amount of intracellular accumulated �29 (Left), SPP1 (Center), or PRD1 (Right)phage DNA was quantified by real-time PCR at different times after infection of the wild-type and the indicated mreB mutant strains. In the case of B. subtilis,strains used were DM-010 (control) and the isogenic cytoskeleton mutants DM-011 (�mreB), DM-012 (�mbl), and DM-013 (�mreBH). For E. coli, strains wereDM-040 (control) and the isogenic cytoskeleton mutant DM-041 (�mreB). Samples were taken at different times after infection and processed as described inthe Materials and Methods. The amounts of accumulated phage DNA (�g viral DNA per mL culture) are expressed as a function of time after infection.

A B C

D E F

G H I

J K

L M

N O

Fig. 2. Subcellular localization of GFP-p2 and membrane protein p16.7.Phase contrast, GFP fluorescence and merged images of typical cells express-ing a xylose-induced GFP-p2 fusion protein (B. subtilis strain DM-010) innon-infected (A, D, and G) and in �29 sus2(513)-infected cells (B, C, E, F, H, andI) 20 and 50 min after infection. Fluorescence images correspond to ‘‘maxprojections’’ of a deconvolved stack of optical sections after 3D reconstruc-tion. DM-010 cells were grown at 37 °C in LB medium supplemented with 5mM MgSO4 to an OD600 of 0.4. Next, xylose was added to a final concentrationof 0.5% and the culture was infected with mutant sus2(513) at a multiplicityof 5. Non-infected cells were analyzed 50 min after xylose addition. (J–N) IFmicroscopy. Cells (B. subtilis 168 �spo0A) were grown in LB medium contain-ing 5 mM MgSO4 at 37 °C. At an OD600 of 0.4, the culture was split and half ofit was infected with phage sus14(1242) at a MOI of 5; samples were harvested20 min later and processed for immunodetection. (J) Phase contrast image. (K)Unprocessed fluorescence images of p16.7 distribution in infected cells. (L)Same cells as in K after deconvolution of an image stack, shown as a ‘‘maxprojection.’’ (M) Overlay of J and L. (N) DAPI staining of DNA. IF signals areshown after deconvolution of an image stack, as a ‘‘max projection.’’ (O)Localization of the p16.7-GFP fusion protein in non-infected cells of B. subtilisstrain DM-004 grown in LB medium supplemented with 0.1% xylose.

13348 � www.pnas.org�cgi�doi�10.1073�pnas.0906465106 Munoz-Espín et al.

Dow

nloa

ded

by g

uest

on

May

14,

202

1

(DM-004) bearing a xylose-inducible 16.7-gfp fusion was con-structed. Fluorescence microscopy of xylose-induced DM-004cells showed that p16.7-GFP also localized in a helix-like patternat the membrane of non-infected cells (Fig. 2O), revealing thatits helical distribution is independent of other �29-encodedproteins.

Since structural data indicated that the substrate of p16.7 isdsDNA (27), we studied the localization of �29 dsDNA by IF(see Materials and Methods and Fig. S5). No IF signals wereobtained in non-infected cells (Fig. S5B). Similarly to the �29DNA polymerase distribution patterns, the phage dsDNA alsolocalized in a helix-like configuration (Fig. S5D), which was evenmore evident when a stack of images was collected and decon-volved (Fig. S5E).

Proper Membrane-Associated Localization of Components of the �29Replication Machinery Requires an Intact MreB Cytoskeleton. Toinvestigate whether localization of components of the �29replication machinery and that of �29 dsDNA depends on 1 ormore of the B. subtilis MreB isoforms, their subcellular local-ization was examined in mreB, mbl, and mreBH single deletionstrains. In the case of the �29 DNA polymerase, the GFP-p2signals did not adapt a helical configuration in any of the 3cytoskeleton mutant cells analyzed, in contrast to wild-type-infected cells (Fig. 3A). Instead, the fluorescent signals localizedthroughout the cell, sometimes in combination with a weakpunctate pattern. Also in the case of protein p16.7, the helicaldistribution observed in wild-type-infected cells was lost in theabsence of an intact MreB cytoskeleton (Fig. 3B). Instead, p16.7

localized in a much more uniform pattern throughout the cellperiphery in infected �mreB, �mbl, or �mreBH mutant cells.Protein p16.7 contains an N-terminal membrane anchor that isrequired for its membrane localization (24), perhaps explainingwhy it still localized at the membrane in cytoskeleton mutantcells.

Colocalization Experiments of the Bacterial Actin-Like Cytoskeletonand �29 DNA Replication Components. To test the possibility thatthe bacterial cytoskeleton is directly responsible for the helicalorganization of the �29 DNA replication machinery at the cellmembrane, the localization of MreB with p16.7 or �29 DNApolymerase was studied in single cells. Localization of c-Myc-MreB and p16.7 in infected cells was studied by IF using strain2060, which expresses an ectopic copy of c-Myc-MreB from thePxyl promoter. The results revealed that c-Myc-MreB and p16.7follow similar helical paths at the periphery of the cell (Fig. 4A,left panel). Overlay of the fluorescent signals showed that MreBand p16.7 colocalize substantially.

The localization of MreB and �29 DNA polymerase wasexamined in infected live cells using strain DM-019, whichcontains xylose-inducible cfp-mreB and yfp-p2 fusions. The re-sults revealed that MreB and �29 DNA polymerase also followsimilar helical paths at the cell periphery (Fig. 4A, right panel),and superimposition of the CFP and YFP signals showed apartial colocalization.

MreB Interacts Directly with Protein p16.7. The observation thatp16.7-GFP (Fig. 2O), but not p2 (Fig. 2G), localized in ahelix-like pattern at the membrane in non-infected cells sug-gested that p16.7 might be associated with the cytoskeleton,either directly or indirectly. To study this possibility we firstperformed pull-down assays using a B. subtilis strain expressingHis-tagged MreB. Cultures infected with phage �29 were sub-jected to in vivo cross-linking and MreB complexes were purified(see Materials and Methods). After SDS/PAGE, the presence of�29 proteins p16.7, DNA polymerase (p2) and TP (p3) wasanalyzed by western blotting (Fig. 4B). Importantly, proteinp16.7 was readily detected in both the whole-cell extracts and inthe fraction corresponding to purified MreB complexes (lanes 2and 3, and lanes 5 and 6, respectively). In contrast, DNApolymerase and TP were either not detected or only present intrace amounts (lanes 5 and 6), indicating that they are not closelyassociated with MreB. As expected, similar results were obtainedwhen cells were infected with the lysis-delayed sus14(1242)mutant phage (lanes 8 and 9, and lanes 11 and 12) and no p16.7signal was observed when cells were infected with a sus14(1242)/sus16.7(48) mutant phage (lanes 13–18). These results show thatMreB and p16.7 are physically associated in a complex.

To test whether MreB and p16.7 interact directly, a bacterial2-hybrid system was used (29, see Fig. 4C). Analysis of possibleinteractions involving Mbl and MreBH was not possible as theMbl and MreBH fusions were non-functional. Controls, in whichonly 1 fusion construct was present or harbouring the 2 emptyvectors, gave white colonies. The blue colonies displayed bytransformants containing pair-wise combination of MreB orp16.7 fusions confirmed that these proteins self interact. Bluecolonies were also observed when the MreB and p16.7 fusionswere co-expressed, providing strong support that MreB andp16.7 interact directly in the complexes detected in the pull-downexperiment.

DiscussionContrary to the long-standing view that the cytoskeleton isunique to eukaryotes, it has now become clear that most bacteriacontain structural and functional homologues of the 3 majoreukaryotic cytoskeletal families: intermediate filaments, tubulin,and actin [for review see (5)]. Viruses are believed to have

Fig. 3. The helical pattern of GFP-p2 and protein p16.7 is lost in mreBcytoskeleton mutant strains. (A) Wild-type (strain DM-010), �mreB (strainDM-011), �mbl (strain DM-012), and �mreBH (strain DM-013) B. subtilis cellscontaining a xylose-inducible copy of gfp-p2 at the amyE locus were grown tomid-exponential phase in LB medium supplemented with 25 mM MgSO4 at37 °C. At an OD600 of 0.4, the cultures were infected with sus2 (513) mutantphage at a MOI of 5 and supplemented with 0.5% xylose. Samples werewithdrawn and analyzed at 50 min after infection. (B) Strains 168 �spo0A(wild-type), DM-001 (�mreB), DM-002 (�mbl), and DM-003 (�mreBH) weregrown in LB medium containing 25 mM MgSO4 at 37 °C. At an OD600 of 0.4, thecultures were infected with phage sus14(1242) at a MOI of 5. Samples wereharvested 20 min postinfection and subjected to IF analysis using polyclonalantibodies against p16.7 (see Materials and Methods). Cells are shown afterdeconvolution of an image stack, as a ‘‘max projection.’’

Munoz-Espín et al. PNAS � August 11, 2009 � vol. 106 � no. 32 � 13349

CELL

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

May

14,

202

1

co-evolved with their hosts, and therefore, it is not surprising thatthey exploit specific aspects of their host cells. Thus, in eu-karyotes it is known that the cytoskeleton is exploited bynumerous viruses modulating its intrinsic dynamic behavior.This allows them to reach their site of replication and to establisha route for the newly assembled progeny to leave the infected cell(for review see ref. 30). However, although emerging studiesincreasingly highlight the importance of the bacterial actin-likecytoskeleton for a number of cellular processes, its exploitationby phages has not been reported. Here we show that an intactbacterial MreB cytoskeleton is required for efficient DNAreplication of phages PRD1, SPP1, and �29. Hence, these resultsindicate that the bacterial cytoskeleton plays a key role in phageDNA replication regardless of host or DNA replicationmechanism.

Compelling evidence has accumulated for replication of pro-karyotic DNA, including that of resident plasmids or phages,occurring at the cytoplasmic membrane (31). To produce highnumbers of progeny within a narrow time window during theirlytic cycle, phage DNA replication must occur simultaneously at

multiple templates and sites, implying that phage DNA replica-tion needs to be highly organized. Several features of thecytoskeletal MreB filaments make them suitable to fulfil apivotal role in the organization of membrane-associated phageDNA replication. First, they are located at the inner surface ofthe cytoplasmic membrane, where efficient phage DNA repli-cation takes place. Second, they typically follow a helical pathspanning the entire length of the cell; that is, they occupy a ratherextensive surface of the cell periphery, allowing organization ofphage DNA replication at multiple sites. Finally, they have adynamic behavior that may generate a driving force to dispersephage DNA replication from initial to additional sites.

The results obtained in this work, using B. subtilis phage �29as a model system, show that the bacterial actin-like cytoskeletonis indeed exploited by the phage to organize its DNA replicationat the membrane of infected cells. Besides the observation thatthe efficiency of �29 DNA replication is dramatically impairedin each of the mreB mutant strains, we found that �29 DNApolymerase, membrane protein p16.7, and replicating �29dsDNA become rapidly organized in helix-like patterns at the

Fig. 4. Colocalization and interaction studies of MreB and �29 DNA replication proteins in infected cells. (A, left panel) B. subtilis strain 2060, containing axylose-inducible copy of c-myc-mreB at amyE, was grown in LB medium supplemented with 25 mM MgSO4 and 0.5% xylose. At an OD600 of 0.4, the culture wasinfected with �29 phage sus14(1242) at a MOI of 5. Samples, harvested 20 min postinfection, were processed for IF using poly- and monoclonal antibodies againstp16.7 and c-Myc, respectively. Fluorescence signals are shown after deconvolution of an image stack, as a ‘‘max projection.’’ Typical cells are imaged by separategreen (c-Myc-MreB) and red (p16.7) channels, with the images displaced to lay side-to-side or merged as indicated. (A, right panel) B. subtilis strain DM-019(yfp-p2/cfp-mreB) was grown in LB medium supplemented with 25 mM MgSO4 until an OD600 of 0.4. Then, 0.5% xylose was added, the culture was infected with�29 at a MOI of 5 and analyzed by fluorescence microscopy 40 min later. Typical cells analyzed by yellow (YFP) or blue (CFP) channel are represented side-to-sideor merged as indicated. For clarity the CFP and YFP signals are false coloured green and red, respectively. (B) B. subtilis strain YK827 (�spo0A �mreBamyE::Pxyl-mreB-his) was grown in LB medium containing 10 mM MgSO4 and 0.25% xylose at 37 °C. At an OD600 of 0.4–0.5, the cultures were infected with �29wild-type (lanes 1–6), �29 sus14(1242) (lanes 7–12), or �29 sus14(1242)/sus16.7(48) (lanes 13–18), respectively, at a MOI of 5. Samples were harvested at 0(non-infected samples), 10 and 30 min postinfection and subjected to in vivo cross-linking and purification of MreB complexes. Whole-cell extracts (lanes 1–3,7–9, and 13–15) and purified MreB complexes (lanes 4–6, 10–12, and 16–18) were separated, and proteins p2, p3, and p16.7 were visualized by western blottingusing anti-p2, anti-p3, and anti-p16.7 antiserum, respectively (see Materials and Methods). The same amounts of total proteins were loaded for purifiedcomplexes. (C) Bacterial 2-hybrid analysis of MreB and p16.7. Images show plated E. coli primary transformants covering pair-wise combinations of proteins p16.7and MreB after incubation at 30 °C for 48 h. To assay for interactions, 10-�L aliquots of the cotransformations of each test pair of plasmids were spotted ontonutrient agar plates containing the selective antibiotics and X-Gal (see Materials and Methods for details).

13350 � www.pnas.org�cgi�doi�10.1073�pnas.0906465106 Munoz-Espín et al.

Dow

nloa

ded

by g

uest

on

May

14,

202

1

cell periphery after infection, and that this depends on an intactMreB cytoskeleton. Recent data have provided evidence that the3 MreB homologues form a single helical structure in the cell andprobably interact directly with each other (11, 15, 32). This is alsosupported by the observation that dynamic behavior of MreB isrequired for the functionality of Mbl (11). This may explain whyefficient �29 DNA replication and proper helical distribution ofthe analyzed �29 DNA replication components is affected in theabsence of any of the 3 MreB isoforms.

The MreB-dependent helical organization of the �29 DNAreplication machinery could be explained if 1 or more �29proteins interact with the cytoskeleton. Extensive evidence existsthat p16.7 is responsible for attaching replicating �29 DNA tothe membrane of infected cells by binding directly to phagedsDNA (24–27). The observation that a GFP fusion of the �29membrane protein p16.7 (p16.7-GFP) displayed a helical local-ization pattern in non-infected cells suggested that p16.7 mightinteract with the cytoskeleton. This possibility was furthersupported by the observation that p16.7 and the MreB cytoskel-eton colocalize substantially. Additionally, we have gained in-sights into the mechanism by which the MreB cytoskeletonorganizes �29 DNA replication. Results from pull-down exper-iments and bacterial 2-hybrid system provide strong evidencethat protein p16.7 interacts directly with MreB (for a model seeFig. 5). The observations that; (i) proper helical p16.7 distribu-tion requires an intact cytoskeleton; (ii) �29 DNA replicationefficiency is severely affected in mreB mutants; and (iii) p16.7interacts directly with MreB, imply that MreB is ultimatelyresponsible for the proper helical configuration and functionalityof p16.7 and, by extension, for the helical configuration of DNApolymerase and �29 dsDNA (see Fig. 5). Accordingly, when thecells are infected with a mutant phage in gene 16.7, the helicallocalization of the �29 DNA polymerase is also affected (see Fig.S6). These conclusions explain why the helical configurations of

phage DNA polymerase and dsDNA are lost in the absence ofan intact cytoskeleton, enhancing the view of MreB as theprimary organizer of the �29 DNA replication machinery.

�29 DNA replication is a highly dynamic process involvingmultiple transient interactions between various �29 proteins andreplicating phage DNA. Similarly, the bacterial cytoskeleton isconstantly remodeled during growth (7). Taking into account thedynamic behavior of both processes, it is unlikely that compo-nents of the �29 replication machinery and the MreB cytoskel-eton will have a permanent interaction, probably explaining whyMreB and p16.7 colocalized substantially but not completely.Besides functioning as a scaffold for the organization of thephage DNA replication, the intrinsic dynamics of the MreBcytoskeleton may play an active role in the rapid distribution ofphage DNA to multiple sites at the membrane of infected cells.Although previous studies have proposed a role for MreB ingenerating the force for chromosomal segregation in differentbacterial organisms [for review see (5)], a direct function in B.subtilis chromosome segregation appears not to be evident (28).Using the MreB cytoskeleton as a motor-like force to organize�29 DNA replication would enhance the efficiency of thisprocess by allowing simultaneous replication of multiple tem-plates at various sites at the membrane. Additionally, we cannotexclude the possibility that the MreB cytoskeleton contributes toefficient �29 DNA replication by recruiting, besides p16.7, otherprotein(s) involved in membrane-associated �29 DNA replica-tion.

Previous work has demonstrated the importance of the MreBactin-like cytoskeleton in various key cellular processes (1, 3, 8,12–16). Here we report that the bacterial cytoskeleton is alsorequired for efficient phage DNA replication. Furthermore, weshow that this is true in both Gram-positive and Gram-negativebacteria and irrespective of the mechanism of phage DNAreplication. Using B. subtilis phage �29 as a model, we show notonly that the phage uses the bacterial MreB cytoskeleton as ascaffold to organize replication of its DNA at the membrane ofthe infected cell, but we also identified membrane protein p16.7as the �29 protein that directly interacts with the cytoskeletonand which, by extension, is responsible for cytoskeleton-dependent helix-like organization of �29 DNA polymerase andviral dsDNA. Thereby this work also highlights the plasticity withwhich prokaryotic viruses, like their eukaryotic counterparts,have evolved efficient mechanisms to take advantage of thecytoskeleton of their hosts for their own benefit.

Materials and MethodsGeneral Methods. Since phage �29 DNA replication is inhibited by Spo0A (33,34), spo0A deletion strains were used when indicated. Unless stated other-wise, the lysis-delayed mutant phage �29 sus14(1242) (35) was used. Themutation in gene 14 has no effect on phage DNA replication or phagemorphogenesis but allows examination of phage protein and DNA localiza-tion at late infection times. In the case of B. subtilis, to avoid possible polareffects on the mreCD morphogenes located downstream mreB (1, 17, 28, 36),experiments were performed in strain DM-011 containing an in-frame dele-tion of mreB. Unless stated otherwise, mid-logarithmically growing B. subtiliscells were infected with �29 or SPP1 at a multiplicity of infection (MOI) of 5 andcell samples were harvested and processed at the indicated times after infec-tion. In the case of E. coli, cells were infected with PRD1 at a MOI of 25.

DNA Techniques and Plasmid Construction. All DNA manipulations were carriedout according to Sambrook et al. (37). Cloning was performed by standardmethods. Plasmids used are listed in Table S3. See SI Text for details.

Bacterial Strains, Phages, and Growth Conditions. Bacterial strains used arelisted in Table S1. Phages used are listed in Table S2. Phage plaque assays weredone by standard methods (37). See SI Text for details.

Bacterial Transformation, Conjugation, and DNA Labeling. B. subtilis cells weretransformed by standard procedures (28, 38). In DNA labeling experiments,the thymine analogue BrdU (Sigma) was used. See SI Text for details. E. coli

Fig. 5. Model of membrane-associated �29 DNA replication organized bythe MreB cytoskeleton. MreB, Mbl, and MreBH are shown to form a putativetriple helical structure closely associated with the inner surface of the mem-brane. Each dimeric unit of protein p16.7 is represented by a yellow hexagon.Although we only have evidence of a direct interaction between p16.7 andMreB we cannot exclude the possibility that p16.7 also interacts directly with1 or both of the other MreB analogs (Mbl and MreBH). As demonstrated bycrystallographic data, 3 p16.7 dimers arrange side by side, defining a deepdsDNA-binding cavity, forming a functional dsDNA binding unit (27). Thetridimeric p16.7 units form oligomers in a helix-like localization at the cellmembrane. Initiation of �29 DNA replication starts with the recognition of theorigin of replication by a TP/DNA polymerase heterodimer. After a transitionstep, �29 DNA polymerase dissociates and continues processive DNA elonga-tion coupled to strand displacement. The proteins and DNA are not drawn toscale. For simplicity, other viral proteins involved in DNA replication are notdrawn.

Munoz-Espín et al. PNAS � August 11, 2009 � vol. 106 � no. 32 � 13351

CELL

BIO

LOG

Y

Dow

nloa

ded

by g

uest

on

May

14,

202

1

K529 donor strain was conjugated with MC1000 or MC1000�mreB recipientstrains by standard methods (37).

Immunofluorescence and Epifluorescence Microscopy. Samples were fixed afterthe indicated times of infection and processed essentially as described (39)with some modifications detailed in SI Text. For live cell imaging, cells wereimmobilized on microscope slides covered with a thin film of 1% agarose inwater, essentially as described (15, 40). For GFP detection, filter 49002(Chroma) was used. CFP and YFP fluorescence were detected with a dualCFP/YFP-ET filter (89002, Chroma). For dual color acquisition the YFP channelwas imaged first, followed immediately by the CFP channel. See SI Text fordetails.

Real-Time PCR. Cells corresponding to 1-mL aliquots of B. subtilis or E. colicultures, withdrawn at different times after infection, were harvested, pro-cessed, and analyzed by real-time PCR essentially as described (41).

Purification of MreB Complexes. Strain YK827 was grown in LB mediumcontaining 10 mM MgSO4 and 0.25% xylose at 37 °C. When the cells reachedan OD600 of 0.4 – 0.5, the culture was infected with phages �29, �29sus14(1242), or �29 sus14(1242)/sus16.7(48) and then treated with 1% form-aldehyde for 10 min. Glycine was added at a final concentration of 150 mM toquench the reaction. Purification of protein complexes with His-tagged MreBproteins was performed as described (42). To identify proteins in the complex,

eluates were separated by SDS/PAGE after heating (60 min at 90 °C). Sepa-rated proteins were analyzed by western blotting. Proteins p2, p3, and p16.7were detected with anti-p2, -p3, and �p16.7 antiserums (1/1,000 dilution),respectively.

Bacterial 2-Hybrid Assay. The method used was that of Karimova et al. (29). SeeSI Text for details.

ACKNOWLEDGMENTS. We thank Alex Formstone, Ian Selmes, Ying Li,Veronica Labrador, Carlos Sanchez, and Jose M. Lazaro for their help and JuanCarlos Alonso (Centro Nacional de Biotecnologıa, Madrid, Spain), DennisBamford (University of Helsinki, Helsinki, Finland), Kenn Gerdes (Institute forCell and Molecular Biosciences, Newcastle, United Kingdom), and Peter Grau-mann (University of Freiburg, Freiburg, Germany) for phage SPP1, phagePRD1, strains MC1000 and MC1000�mreB, and strains JS64 and JS65, respec-tively. This work was supported by Spanish Ministry of Education and ScienceGrants BFU2005-00733 (to M.S.) and BFU2005-01878 (to W.J.J.M.), SpanishMinistry of Science and Innovation Grant BFU2008-00215 (to M.S.), SpanishMinistry of Education and Science Grant Consolider-Ingenio 2010 24717 (toM.S.) and by an Institutional grant from Fundacion Ramon Areces to theCentro de Biología Molecular ‘‘Severo Ochoa.’’ D.M.-E. and V.C.-L. are holdersof an I3P contract from the Spanish National Research Council and a predoc-toral fellowship from the Spanish Ministry of Education and Science, respec-tively. Work in the Errington laboratory was supported by grants from theBiotechnology and Biological Sciences Research Council and the Human Fron-tier Science Program.

1. Jones LJ, Carballido-Lopez R, Errington J (2001) Control of cell shape in bacteria: helical,actin-like filaments in Bacillus subtilis. Cell 104:913–922.

2. Van den Ent F, Amos LA, Lowe J (2001) Prokaryotic origin of the actin cytoskeleton.Nature 413:39–44.

3. Shih Y, Rothfield L (2006) The bacterial cytoskeleton. Microbiol Mol Biol Rev 70:729–754.

4. Michie KA, Lowe J (2006) Dynamic filaments of the bacterial cytoskeleton. Annu RevBiochem 75:467–492.

5. Carballido-Lopez R (2006) The bacterial actin-like cytoskeleton. Microbiol Mol Biol Rev70:888–909.

6. Gitai Z (2007) Diversification and specialization of the bacterial cytoskeleton. Curr OpinCell Biol 19:1–8.

7. Carballido-Lopez R, Errington J (2003) The bacterial cytoskeleton: In vivo dynamics ofthe actin-like protein Mbl of Bacillus subtilis. Dev Cell 4:19–28.

8. Figge RM, Divakaruni AV, Gober JW (2004) MreB, the cell shape-determining bacterialactin homologue, coordinates cell wall morphogenesis in Caulobacter crescentus. MolMicrobiol 51:1321–1332.

9. Gitai Z, Dye N, Shapiro L (2004) An actin-like gene can determine cell polarity inbacteria. Proc Natl Acad Sci USA 101:8643–8648.

10. Kim SY, Gitai Z, Kinkhabwala A, Shapiro L, Moemer WE (2006) Single molecules of thebacterial actin MreB undergo directed treadmilling motion in Caulobacter crescentus.Proc Natl Acad Sci USA 103:10929–10934.

11. Defeu Soufo HJ, Graumann PL (2006) Dynamic localization and interaction with otherBacillus subtilis actin-like proteins are important for the function of MreB. MolMicrobiol 62:1340–1356.

12. Daniel RA, Errington J (2003) Control of cell morphogenesis in bacteria: Two distinctways to make a rod-shaped cell. Cell 113:767–776.

13. Gitai Z (2005) The new bacterial cell biology: Moving parts and subcellular architecture.Cell 120:577–586.

14. Carballido-Lopez R (2006) Orchestrating bacterial cell morphogenesis. Mol Microbiol60:815–819.

15. Carballido-Lopez R, Formstone A, Li Y, Noirot P, Errington J (2006) Actin homologMreBH governs cell morphogenesis by localization of the cell wall hydrolase LytE. DevCell 11:399–409.

16. Madabhushi R, Marians KJ (2009) Actin homolog MreB affects chromosome segrega-tion by regulating topoisomerase IV in Escherichia coli. Mol Cell 33:171–180.

17. Levin PA, Margolis PS, Setlow P, Losick R, Sun D (1992) Identification of Bacillus subtilisgenes for septum placement and shape determination. J Bacteriol 174:6717–6728.

18. Varley AW, Stewart GC (1992) The divIVB region of the Bacillus subtilis chromosomeencodes homologs of Escherichia coli septum placement (MinCD) and cell shape(MreBCD) determinants. J Bacteriol 174:6729–6742.

19. Abhayawardhane Y, Steward GC (1995) Bacillus subtilis possesses a second determi-nant with extensive sequence similarity to the Escherichia coli mreB morphogene. JBacteriol 177:765–773.

20. Bravo A, Serrano-Heras G, Salas M (2005) Compartmentalization of prokaryotic DNAreplication. FEMS Microbiol Rev 29:25–47.

21. Calendar R (2006) in The Bacteriophages (Oxford Univ Press, New York, Oxford).

22. Salas M (1991) Protein-priming of DNA replication. Annu Rev Biochem 60:39–71.23. Meijer WJ, Lewis PJ, Errington J, Salas M (2000) Dynamic relocalization of phage �29

DNA during replication and the role of the viral protein p16.7. EMBO J 19:4182–4190.24. Meijer WJ, Serna-Rico A, Salas M (2001) Characterization of the bacteriophage �29-

encoded protein p16.7: A membrane protein involved in phage DNA replication. MolMicrobiol 39:731–746.

25. Munoz-Espín D, et al. (2004) Phage �29 DNA-replication organizer membrane proteinp16.7 contains a coiled-coil and a dimeric, homeodomain-related, functional domain.J Biol Chem 279:50437–50445.

26. Asensio JL, et al. (2005) Structure of the functional domain of �29 replication orga-nizer. Insights into oligomerization and DNA binding. J Biol Chem 280:20730–20739.

27. Albert A, et al. (2005) Structural basis for membrane anchorage of viral �29 DNA duringreplication. J Biol Chem 280:42486–42488.

28. Formstone A, Errington J (2005) A magnesium-dependent mreB null mutant: Implica-tions for the role of mreB in Bacillus subtilis. Mol Microbiol 55:1646–1657.

29. Karimova G, Pidoux J, Ullmann A, Ladant D (1998) A bacterial two-hybrid system basedon a reconstituted signal transduction pathway. Proc Natl Acad Sci USA 95:5752–5756.

30. Greber UF, Way M (2006) A superhighway of virus infection. Cell 124:741–754.31. Firshein W (1989) Role of the DNA/membrane complex in prokaryotic DNA replication.

Annu Rev Microbiol 43:89–120.32. Defeu Soufo HJ, Graumann PL (2005) Bacillus subtilis actin-like protein MreB influences

the positioning of the replication machinery and requires membrane proteins MreC/Dand other actin-like proteins for proper localization. BMC Cell Biol 6:10.

33. Meijer WJ, et al. (2005) Molecular basis for the exploitation of spore formation assurvival mechanism by virulent phage �29. EMBO J 24:3647–3657.

34. Castilla-Llorente V, Munoz-Espín D, Villar L, Salas M, Meijer WJ (2006) Spo0A, the keytranscriptional regulator for entrance into sporulation, is an inhibitor of DNA replica-tion. EMBO J 25:3890–3899.

35. Jimenez F, Camacho A, de la Torre J, Vinuela E, Salas M (1977) Assembly of Bacillussubtilis phage �29. 2. Mutants in the cistrons coding for the non-structural proteins. EurJ Biochem 73:57–72.

36. Leaver M, Errington J (2005) Roles for MreC and MreD proteins in helical growth of thecylindrical cell wall in Bacillus subtilis. Mol Microbiol 57:1196–1209.

37. Sambrook J, Fritsch EF, Maniatis T (1989) in Molecular Cloning: A Laboratory Manual(Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York).

38. Ward JB, Jr, Zahler SA (1973) Genetic studies of leucine biosynthesis in Bacillus subtilis.J Bacteriol 116:719–726.

39. Lewis PJ, Errington J (1997) Direct evidence for active segregation of oriC regions of theBacillus subtilis chromosome and co-localization with the Spo0J partitioning protein.Mol Microbiol 25:945–954.

40. Glaser P, et al. (1997) Dynamic, mitotic-like behavior of a bacterial protein required foraccurate chromosome partitioning. Genes Dev 11:1160–1168.

41. Gonzalez-Huici V, Salas M, Hermoso JM (2004) The push-pull mechanism of bacterio-phage �29 DNA injection. Mol Microbiol 52:529–540.

42. Ishikawa S, Kawai Y, Hiramatsu K, Kuwano M, Ogasawara N (2006) A new FtsZ-interacting protein, YlmF, complements the activity of FtsA during progression of celldivision in Bacillus subtilis. Mol Microbiol 60:1364–1380.

13352 � www.pnas.org�cgi�doi�10.1073�pnas.0906465106 Munoz-Espín et al.

Dow

nloa

ded

by g

uest

on

May

14,

202

1