The -like MreB organizes viral DNA replication in

Daniel Mun˜ oz-Espína,b, Richard Danielb,c, Yoshikazu Kawaic, Rut Carballido-Lo´ pezd, Virginia Castilla-Llorentea, Jeff Erringtonb,c,1,2, Wilfried J. J. Meijera,1, and Margarita Salasa,1,2

aInstituto de Biología Molecular ‘‘Eladio Vin˜uela’’ and Centro de Biología Molecular ‘‘Severo Ochoa,’’ Consejo Superior de Investigaciones Cientificas-Universidad Auto´noma de Madrid, Canto Blanco, 28049 Madrid, Spain; bSir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, United Kingdom; cInstitute for Cell and Molecular Biosciences, Newcastle University, Newcastle-upon-Tyne NE2 4HH, United Kingdom; and dGe´ne´ tique 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 involved in occurs via a so-called -primed mechanism (22, see Fig. membrane-associated replication of prokaryotic genomes. Here S1). Phage ␾29 DNA transcription is divided into early and late we show that the actin-like MreB cytoskeleton of the distantly stages (see Fig. S1 for a genetic and a transcriptional map). related bacteria and Bacillus subtilis is required for encoding DNA replication proteins such as DNA poly- efficient viral DNA replication. Detailed analyses of B. subtilis merase (p2) and TP (p3) are located in the left-side early operon. phage ␾29 showed that the MreB cytoskeleton plays a crucial role The right-side early operon contains 16.7 that encodes a in organizing phage DNA replication at the membrane. Thus, (p16.7) required for optimal in vivo ␾29 DNA phage double-stranded DNA and components of the ␾29 replica- replication (23, 24). Additional functional, biochemical and tion machinery localize in peripheral helix-like structures in a structural studies have provided strong evidence that p16.7 is cytoskeleton-dependent way. Importantly, we show that MreB responsible for attaching ␾29 DNA to the bacterial membrane interacts directly with the ␾29 membrane-protein p16.7, respon- (24–26). Crystallographic resolution of the p16.7 DNA binding sible for attaching viral DNA at the . Altogether, the domain (p16.7C) in complex with dsDNA revealed that 1 results reveal another function for the MreB cytoskeleton and dsDNA binding unit is formed by 3 p16.7C dimers that are describe a mechanism by which viral DNA replication is organized arranged in such a way that they form a deep positively charged at the bacterial membrane. longitudinal cavity that interacts with the phosphate backbone of dsDNA (27). Bacillus subtilis ͉ phage ␾29 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 ␾29 enes of the mreB family encode homologues of eukaryotic dsDNA localize in a helix-like pattern near the membrane of Gactin (1, 2) that form a cytoskeleton in most non-spherical infected B. subtilis cells. In addition, this helical organization bacteria (3–6). MreB proteins form filamentous structures fol- depends on all 3 host-encoded MreB proteins. Moreover, we lowing a helical path around the inner surface of the cytoplasmic show that MreB interacts directly with p16.7 in vivo. A model membrane (1). These actin-like filaments are continuously re- integrating these results is discussed. modelled during cell-cycle progression (7–11). Evidence is ac- cumulating that the bacterial MreB cytoskeleton plays key roles Results in several important cellular processes such as cell shape deter- Efficient Phage DNA Replication Requires an Intact Bacterial MreB mination, chromosome segregation, and cell polarity (1, 3, 8, Cytoskeleton. B. subtilis mreB mutant strains can be propagated 12–16). Whereas Gram-negative bacteria have a single mreB with near wild-type growth rate and cell morphology in growth gene, Gram-positive bacteria often have multiple mreB homo- media supplemented with high concentrations of magnesium logues. Bacillus subtilis encodes 3 MreB isoforms: MreB, Mbl, (28). Therefore, when cytoskeleton mutant strains were used in and MreBH (17–19). this work, media were supplemented with 25 mM MgSO4. For decades, evidence has been provided that replication of To examine a possible role of the bacterial actin-like cytoskel- phage DNA, like that of other prokaryotic genomes, occurs at eton in viral DNA replication we determined the efficiency of the cytoplasmic membrane (for review see 20). However, little is DNA replication of phages ␾29, SPP1, and PRD1 in infected known about the proteins or their organization in membrane- wild-type and mreB mutant cells. ␾29 and SPP1 infect the associated replication of viral genomes in bacteria. Phages ␾29 Gram-positive bacterium B. subtilis but use different modes of and SPP1 infect the Gram-positive bacterium B. subtilis, and DNA replication [reviewed in (21)]. In contrast, PRD1 uses a phage PRD1 infects the Gram-negative bacterium Escherichia similar DNA replication mechanism as ␾29 but infects the coli. Whereas PRD1 and ␾29 use the protein-primed mechanism Gram-negative bacterium E. coli (reviewed in 21). Fig. 1 shows of DNA replication, phage SPP1 replicates its DNA initially via that the efficiency of DNA replication of these phages was the theta mode and later via a rolling circle mode [reviewed in severely affected in the absence of an intact MreB cytoskeleton. (21)]. Here we show a key role for the MreB cytoskeleton in In the case of ␾29, deletion of any of the 3 mreB-like genes phages replicating by different modes in the distantly related

bacteria E. coli and B. subtilis. Thus, the efficiency of replication CELL BIOLOGY of phage PRD1, and that of phages SPP1 and ␾29, is severely Author contributions: D.M.-E., R.D., J.E., W.J.J.M., and M.S. designed research; D.M.-E. and Y.K. performed research; V.C.-L. contributed new reagents/analytic tools; D.M.-E., R.D., affected in the absence of an intact cytoskeleton. 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 the The underlying mechanism by which the cytoskeleton leads to paper. efficient phage DNA replication was analyzed in detail for B. The authors declare no conflict of interest. ␾ subtilis phage 29, whose DNA replication has been well char- 1J.E., W.J.J.M., and M.S. contributed equally to this work. ␾ acterized in vitro. The 29 genome consists of a linear double- 2To whom correspondence may be addressed. E-mail: [email protected] or stranded DNA (dsDNA) with a terminal protein (TP) covalently [email protected]. Ј linked at each 5 end that is the primer for the initiation of phage This article contains supporting information online at www.pnas.org/cgi/content/full/ DNA replication. Hence, initiation of ␾29 DNA replication 0906465106/DCSupplemental.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0906465106 PNAS ͉ August 11, 2009 ͉ vol. 106 ͉ no. 32 ͉ 13347–13352 Downloaded by guest on September 24, 2021 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 were DM-040 (control) and the isogenic cytoskeleton mutant DM-041 (⌬mreB). Samples were taken at different times after infection and processed as described in the Materials and Methods. The amounts of accumulated phage DNA (␮g viral DNA per mL culture) are expressed as a function of time after infection.

caused a similar deleterious effect on DNA replication. Analyses in most cases spanned the entire length of the infected cells. of phage DNA accumulation by agarose gel electrophoresis Three-dimensional reconstruction of a set of deconvolved Z showed that intracellular phage DNA was detected early after sections demonstrated that p16.7 forms helical structures at the infection of both wild-type and mutant cells (see Fig. S2), membrane of infected cells (see SI Text and Movie S1). To study indicating that the cytoskeleton mutations have no or little effect the localization pattern of p16.7 in live cells, a B. subtilis strain on phage DNA injection. The results show that the bacterial MreB cytoskeleton is required for efficient phage DNA repli- cation in Gram-positive and Gram-negative bacteria, regardless of the phage DNA replication mechanism. The role of the cytoskeleton in viral DNA replication was studied in detail for B. subtilis phage ␾29. To get further evidence that an intact cytoskeleton is required for efficient ␾29 DNA ABC J K replication, phage DNA synthesis was studied using the B. subtilis strain 2060, which contains a disrupted mreB gene and a xylose-inducible copy of c-myc-mreB. The results presented in Fig. S3 show that the amount of phage DNA increased rapidly when the cells were grown in the presence of xylose. In contrast, the efficiency of ␾29 DNA replication was low in the absence of D E F L M the inductor. Importantly, efficient phage DNA replication was restored when xylose was added to the conditional strain 30 min after infection, confirming that MreB plays a role in phage DNA replication.

␾29 DNA Polymerase Localizes in a Helix-Like Pattern at the Periphery G H I N O of Infected Cells During ␾29 DNA Replication. To determine the subcellular distribution of the ␾29 replication machinery we first studied the localization of the gene 2-encoded ␾29 DNA poly- merase fused to the green fluorescent protein (GFP) in both infected and non-infected live cells. For this, B. subtilis strains were engineered that contain N- or C-terminal xylose-inducible Fig. 2. Subcellular localization of GFP-p2 and membrane protein p16.7. fusions of gfp to ␾29 gene 2. Complementation experiments Phase contrast, GFP fluorescence and merged images of typical cells express- showed that both the N- (GFP-p2) and C- (p2-GFP) terminal ing a xylose-induced GFP-p2 fusion protein (B. subtilis strain DM-010) in fusion proteins were functional (Fig. S4). non-infected (A, D, and G) and in ␾29 sus2(513)-infected cells (B, C, E, F, H, and As shown in Fig. 2 (D, G), GFP-p2 was distributed uniformly I) 20 and 50 min after infection. Fluorescence images correspond to ‘‘max in xylose-induced non-infected cells. Interestingly, between 20 projections’’ of a deconvolved stack of optical sections after 3D reconstruc- tion. DM-010 cells were grown at 37 °C in LB medium supplemented with 5 and 50 min postinfection, corresponding to the period of effi- ␾ mM MgSO4 to an OD600 of 0.4. Next, xylose was added to a final concentration cient 29 DNA replication, GFP-p2 localized in helix-like struc- of 0.5% and the culture was infected with mutant sus2(513) at a multiplicity tures (Fig. 2 E and F) close to the membrane of infected cells of 5. Non-infected cells were analyzed 50 min after xylose addition. (J–N)IF (Fig. 2 H and I). microscopy. 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 of ␾29 Membrane-Protein p16.7 and ␾29 dsDNA Localize in Helix-Like it was infected with phage sus14(1242) at a MOI of 5; samples were harvested Patterns. Using immunofluorescence (IF) techniques, we next 20 min later and processed for immunodetection. (J) Phase contrast image. (K) studied the subcelullar localization of protein p16.7 by decon- 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 ‘‘max volving optical sections through the z axis of the cell. Fig. 2 (J–N) ␾ projection.’’ (M) Overlay of J and L.(N) DAPI staining of DNA. IF signals are shows phase contrast and IF images of 29-infected cells. shown after deconvolution of an image stack, as a ‘‘max projection.’’ (O) Overlay of phase contrast and deconvolved IF images (Fig. 2M) Localization of the p16.7-GFP fusion protein in non-infected cells of B. subtilis showed that p16.7 localized in peripheral helix-like patterns that strain DM-004 grown in LB medium supplemented with 0.1% xylose.

13348 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0906465106 Mun˜oz-Espín et al. Downloaded by guest on September 24, 2021 localized in a much more uniform pattern throughout the cell periphery in infected ⌬mreB, ⌬mbl,or⌬mreBH mutant cells. Protein p16.7 contains an N-terminal membrane anchor that is required for its membrane localization (24), perhaps explaining why it still localized at the membrane in cytoskeleton mutant cells.

Colocalization Experiments of the Bacterial Actin-Like Cytoskeleton and ␾29 DNA Replication Components. To test the possibility that the bacterial cytoskeleton is directly responsible for the helical organization of the ␾29 DNA replication machinery at the cell membrane, the localization of MreB with p16.7 or ␾29 DNA polymerase was studied in single cells. Localization of c-Myc- MreB and p16.7 in infected cells was studied by IF using strain 2060, which expresses an ectopic copy of c-Myc-MreB from the Pxyl promoter. The results revealed that c-Myc-MreB and p16.7 follow similar helical paths at the periphery of the cell (Fig. 4A, left panel). Overlay of the fluorescent signals showed that MreB and p16.7 colocalize substantially. The localization of MreB and ␾29 DNA polymerase was examined in infected live cells using strain DM-019, which contains xylose-inducible cfp-mreB and yfp-p2 fusions. The re- sults revealed that MreB and ␾29 DNA polymerase also follow similar helical paths at the cell periphery (Fig. 4A, right panel), Fig. 3. The helical pattern of GFP-p2 and protein p16.7 is lost in mreB and superimposition of the CFP and YFP signals showed a cytoskeleton mutant strains. (A) Wild-type (strain DM-010), ⌬mreB (strain partial colocalization. DM-011), ⌬mbl (strain DM-012), and ⌬mreBH (strain DM-013) B. subtilis cells containing a xylose-inducible copy of gfp-p2 at the amyE locus were grown to MreB Interacts Directly with Protein p16.7. The observation that mid-exponential phase in LB medium supplemented with 25 mM MgSO at 4 p16.7-GFP (Fig. 2O), but not p2 (Fig. 2G), localized in a 37 °C. At an OD600 of 0.4, the cultures were infected with sus2 (513) mutant phage at a MOI of 5 and supplemented with 0.5% xylose. Samples were helix-like pattern at the membrane in non-infected cells sug- withdrawn and analyzed at 50 min after infection. (B) Strains 168 ⌬spo0A gested that p16.7 might be associated with the cytoskeleton, (wild-type), DM-001 (⌬mreB), DM-002 (⌬mbl), and DM-003 (⌬mreBH) were either directly or indirectly. To study this possibility we first grown in LB medium containing 25 mM MgSO4 at 37 °C. At an OD600 of 0.4, the performed pull-down assays using a B. subtilis strain expressing cultures were infected with phage sus14(1242) at a MOI of 5. Samples were His-tagged MreB. Cultures infected with phage ␾29 were sub- harvested 20 min postinfection and subjected to IF analysis using polyclonal jected to in vivo cross-linking and MreB complexes were purified antibodies against p16.7 (see Materials and Methods). Cells are shown after (see Materials and Methods). After SDS/PAGE, the presence of deconvolution of an image stack, as a ‘‘max projection.’’ ␾29 proteins p16.7, DNA polymerase (p2) and TP (p3) was analyzed by western blotting (Fig. 4B). Importantly, protein (DM-004) bearing a xylose-inducible 16.7-gfp fusion was con- p16.7 was readily detected in both the whole-cell extracts and in structed. Fluorescence microscopy of xylose-induced DM-004 the fraction corresponding to purified MreB complexes (lanes 2 cells showed that p16.7-GFP also localized in a helix-like pattern and 3, and lanes 5 and 6, respectively). In contrast, DNA at the membrane of non-infected cells (Fig. 2O), revealing that polymerase and TP were either not detected or only present in trace amounts (lanes 5 and 6), indicating that they are not closely its helical distribution is independent of other ␾29-encoded associated with MreB. As expected, similar results were obtained proteins. when cells were infected with the lysis-delayed sus14(1242) Since structural data indicated that the substrate of p16.7 is mutant phage (lanes 8 and 9, and lanes 11 and 12) and no p16.7 dsDNA (27), we studied the localization of ␾29 dsDNA by IF signal was observed when cells were infected with a sus14(1242)/ (see Materials and Methods and Fig. S5). No IF signals were ␾ sus16.7(48) mutant phage (lanes 13–18). These results show that obtained in non-infected cells (Fig. S5B). Similarly to the 29 MreB and p16.7 are physically associated in a complex. DNA polymerase distribution patterns, the phage dsDNA also To test whether MreB and p16.7 interact directly, a bacterial localized in a helix-like configuration (Fig. S5D), which was even 2-hybrid system was used (29, see Fig. 4C). Analysis of possible more evident when a stack of images was collected and decon- interactions involving Mbl and MreBH was not possible as the volved (Fig. S5E). Mbl and MreBH fusions were non-functional. Controls, in which only 1 fusion construct was present or harbouring the 2 empty ␾ Proper Membrane-Associated Localization of Components of the 29 vectors, gave white colonies. The blue colonies displayed by Replication Machinery Requires an Intact MreB Cytoskeleton. To transformants containing pair-wise combination of MreB or ␾ investigate whether localization of components of the 29 p16.7 fusions confirmed that these proteins self interact. Blue replication machinery and that of ␾29 dsDNA depends on 1 or colonies were also observed when the MreB and p16.7 fusions

more of the B. subtilis MreB isoforms, their subcellular local- were co-expressed, providing strong support that MreB and CELL BIOLOGY ization was examined in mreB, mbl, and mreBH single deletion p16.7 interact directly in the complexes detected in the pull-down strains. In the case of the ␾29 DNA polymerase, the GFP-p2 experiment. signals did not adapt a helical configuration in any of the 3 cytoskeleton mutant cells analyzed, in contrast to wild-type- Discussion infected cells (Fig. 3A). Instead, the fluorescent signals localized Contrary to the long-standing view that the cytoskeleton is throughout the cell, sometimes in combination with a weak unique to , it has now become clear that most bacteria punctate pattern. Also in the case of protein p16.7, the helical contain structural and functional homologues of the 3 major distribution observed in wild-type-infected cells was lost in the eukaryotic cytoskeletal families: intermediate filaments, , absence of an intact MreB cytoskeleton (Fig. 3B). Instead, p16.7 and actin [for review see (5)]. Viruses are believed to have

Mun˜oz-Espín et al. PNAS ͉ August 11, 2009 ͉ vol. 106 ͉ no. 32 ͉ 13349 Downloaded by guest on September 24, 2021 Fig. 4. Colocalization and interaction studies of MreB and ␾29 DNA replication proteins in infected cells. (A, left panel) B. subtilis strain 2060, containing a xylose-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 was infected with ␾29 phage sus14(1242) at a MOI of 5. Samples, harvested 20 min postinfection, were processed for IF using poly- and monoclonal antibodies against p16.7 and c-Myc, respectively. Fluorescence signals are shown after deconvolution of an image stack, as a ‘‘max projection.’’ Typical cells are imaged by separate green (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-side or merged as indicated. For clarity the CFP and YFP signals are false coloured green and red, respectively. (B) B. subtilis strain YK827 (⌬spo0A ⌬mreB amyE::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 ␾29 wild-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 blotting using anti-p2, anti-p3, and anti-p16.7 antiserum, respectively (see Materials and Methods). The same amounts of total proteins were loaded for purified complexes. (C) Bacterial 2-hybrid analysis of MreB and p16.7. Images show plated E. coli primary transformants covering pair-wise combinations of proteins p16.7 and 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 onto nutrient agar plates containing the selective antibiotics and X-Gal (see Materials and Methods for details).

co-evolved with their hosts, and therefore, it is not surprising that multiple templates and sites, implying that phage DNA replica- they exploit specific aspects of their host cells. Thus, in eu- tion needs to be highly organized. Several features of the karyotes it is known that the cytoskeleton is exploited by cytoskeletal MreB filaments make them suitable to fulfil a numerous viruses modulating its intrinsic dynamic behavior. pivotal role in the organization of membrane-associated phage This allows them to reach their site of replication and to establish DNA replication. First, they are located at the inner surface of a route for the newly assembled progeny to leave the infected cell the cytoplasmic membrane, where efficient phage DNA repli- (for review see ref. 30). However, although emerging studies cation takes place. Second, they typically follow a helical path increasingly highlight the importance of the bacterial actin-like spanning the entire length of the cell; that is, they occupy a rather cytoskeleton for a number of cellular processes, its exploitation extensive surface of the cell periphery, allowing organization of by phages has not been reported. Here we show that an intact phage DNA replication at multiple sites. Finally, they have a bacterial MreB cytoskeleton is required for efficient DNA dynamic behavior that may generate a driving force to disperse replication of phages PRD1, SPP1, and ␾29. Hence, these results phage DNA replication from initial to additional sites. indicate that the bacterial cytoskeleton plays a key role in phage The results obtained in this work, using B. subtilis phage ␾29 DNA replication regardless of host or DNA replication as a model system, show that the bacterial actin-like cytoskeleton mechanism. is indeed exploited by the phage to organize its DNA replication Compelling evidence has accumulated for replication of pro- at the membrane of infected cells. Besides the observation that karyotic DNA, including that of resident plasmids or phages, the efficiency of ␾29 DNA replication is dramatically impaired occurring at the cytoplasmic membrane (31). To produce high in each of the mreB mutant strains, we found that ␾29 DNA numbers of progeny within a narrow time window during their polymerase, membrane protein p16.7, and replicating ␾29 lytic cycle, phage DNA replication must occur simultaneously at dsDNA become rapidly organized in helix-like patterns at the

13350 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0906465106 Mun˜oz-Espín et al. Downloaded by guest on September 24, 2021 phage DNA polymerase and dsDNA are lost in the absence of an intact cytoskeleton, enhancing the view of MreB as the primary organizer of the ␾29 DNA replication machinery. ␾29 DNA replication is a highly dynamic process involving multiple transient interactions between various ␾29 proteins and replicating phage DNA. Similarly, the bacterial cytoskeleton is constantly remodeled during growth (7). Taking into account the dynamic 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 why MreB and p16.7 colocalized substantially but not completely. Besides functioning as a scaffold for the organization of the phage DNA replication, the intrinsic dynamics of the MreB cytoskeleton may play an active role in the rapid distribution of phage DNA to multiple sites at the membrane of infected cells. Although previous studies have proposed a role for MreB in generating the force for chromosomal segregation in different bacterial organisms [for review see (5)], a direct function in B. ␾ Fig. 5. Model of membrane-associated 29 DNA replication organized by subtilis chromosome segregation appears not to be evident (28). the MreB cytoskeleton. MreB, Mbl, and MreBH are shown to form a putative Using the MreB cytoskeleton as a motor-like force to organize triple 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. 29 DNA replication would enhance the efficiency of this Although we only have evidence of a direct interaction between p16.7 and process by allowing simultaneous replication of multiple tem- MreB we cannot exclude the possibility that p16.7 also interacts directly with plates at various sites at the membrane. Additionally, we cannot 1 or both of the other MreB analogs (Mbl and MreBH). As demonstrated by exclude the possibility that the MreB cytoskeleton contributes to crystallographic data, 3 p16.7 dimers arrange side by side, defining a deep efficient ␾29 DNA replication by recruiting, besides p16.7, other dsDNA-binding cavity, forming a functional dsDNA binding unit (27). The protein(s) involved in membrane-associated ␾29 DNA replica- tridimeric p16.7 units form oligomers in a helix-like localization at the cell ␾ tion. membrane. Initiation of 29 DNA replication starts with the recognition of the Previous work has demonstrated the importance of the MreB origin of replication by a TP/DNA polymerase heterodimer. After a transition step, ␾29 DNA polymerase dissociates and continues processive DNA elonga- actin-like cytoskeleton in various key cellular processes (1, 3, 8, tion coupled to strand displacement. The proteins and DNA are not drawn to 12–16). Here we report that the bacterial cytoskeleton is also scale. For simplicity, other viral proteins involved in DNA replication are not required for efficient phage DNA replication. Furthermore, we drawn. show that this is true in both Gram-positive and Gram-negative bacteria and irrespective of the mechanism of phage DNA replication. Using B. subtilis phage ␾29 as a model, we show not cell periphery after infection, and that this depends on an intact only that the phage uses the bacterial MreB cytoskeleton as a MreB cytoskeleton. Recent data have provided evidence that the scaffold to organize replication of its DNA at the membrane of 3 MreB homologues form a single helical structure in the cell and the infected cell, but we also identified membrane protein p16.7 probably interact directly with each other (11, 15, 32). This is also as the ␾29 protein that directly interacts with the cytoskeleton supported by the observation that dynamic behavior of MreB is and which, by extension, is responsible for cytoskeleton- required for the functionality of Mbl (11). This may explain why dependent helix-like organization of ␾29 DNA polymerase and efficient ␾29 DNA replication and proper helical distribution of viral dsDNA. Thereby this work also highlights the plasticity with the analyzed ␾29 DNA replication components is affected in the which prokaryotic viruses, like their eukaryotic counterparts, absence of any of the 3 MreB isoforms. have evolved efficient mechanisms to take advantage of the The MreB-dependent helical organization of the ␾29 DNA cytoskeleton of their hosts for their own benefit. replication machinery could be explained if 1 or more ␾29 proteins interact with the cytoskeleton. Extensive evidence exists Materials and Methods that p16.7 is responsible for attaching replicating ␾29 DNA to General Methods. Since phage ␾29 DNA replication is inhibited by Spo0A (33, the membrane of infected cells by binding directly to phage 34), spo0A deletion strains were used when indicated. Unless stated other- dsDNA (24–27). The observation that a GFP fusion of the ␾29 wise, the lysis-delayed mutant phage ␾29 sus14(1242) (35) was used. The membrane protein p16.7 (p16.7-GFP) displayed a helical local- mutation in gene 14 has no effect on phage DNA replication or phage ization pattern in non-infected cells suggested that p16.7 might morphogenesis but allows examination of phage protein and DNA localiza- interact with the cytoskeleton. This possibility was further tion at late infection times. In the case of B. subtilis, to avoid possible polar effects on the mreCD morphogenes located downstream mreB (1, 17, 28, 36), supported by the observation that p16.7 and the MreB cytoskel- experiments were performed in strain DM-011 containing an in-frame dele- eton colocalize substantially. Additionally, we have gained in- tion of mreB. Unless stated otherwise, mid-logarithmically growing B. subtilis sights into the mechanism by which the MreB cytoskeleton cells were infected with ␾29 or SPP1 at a multiplicity of infection (MOI) of 5 and organizes ␾29 DNA replication. Results from pull-down exper- cell samples were harvested and processed at the indicated times after infec- iments and bacterial 2-hybrid system provide strong evidence tion. In the case of E. coli, cells were infected with PRD1 at a MOI of 25. that protein p16.7 interacts directly with MreB (for a model see

Fig. 5). The observations that; (i) proper helical p16.7 distribu- DNA Techniques and Plasmid Construction. All DNA manipulations were carried CELL BIOLOGY tion requires an intact cytoskeleton; (ii) ␾29 DNA replication out according to Sambrook et al. (37). Cloning was performed by standard efficiency is severely affected in mreB mutants; and (iii) p16.7 methods. Plasmids used are listed in Table S3. See SI Text for details. interacts directly with MreB, imply that MreB is ultimately Bacterial Strains, Phages, and Growth Conditions. Bacterial strains used are responsible for the proper helical configuration and functionality listed in Table S1. Phages used are listed in Table S2. Phage plaque assays were of p16.7 and, by extension, for the helical configuration of DNA done by standard methods (37). See SI Text for details. polymerase and ␾29 dsDNA (see Fig. 5). Accordingly, when the cells are infected with a mutant phage in gene 16.7, the helical Bacterial Transformation, Conjugation, and DNA Labeling. B. subtilis cells were localization of the ␾29 DNA polymerase is also affected (see Fig. transformed by standard procedures (28, 38). In DNA labeling experiments, S6). These conclusions explain why the helical configurations of the thymine analogue BrdU (Sigma) was used. See SI Text for details. E. coli

Mun˜oz-Espín et al. PNAS ͉ August 11, 2009 ͉ vol. 106 ͉ no. 32 ͉ 13351 Downloaded by guest on September 24, 2021 K529 donor strain was conjugated with MC1000 or MC1000⌬mreB recipient eluates were separated by SDS/PAGE after heating (60 min at 90 °C). Sepa- strains by standard methods (37). rated proteins were analyzed by western blotting. Proteins p2, p3, and p16.7 were detected with anti-p2, -p3, and Ϫp16.7 antiserums (1/1,000 dilution), Immunofluorescence and Epifluorescence Microscopy. Samples were fixed after respectively. the indicated times of infection and processed essentially as described (39) with some modifications detailed in SI Text. For live cell imaging, cells were Bacterial 2-Hybrid Assay. The method used was that of Karimova et al. (29). See immobilized on microscope slides covered with a thin film of 1% agarose in SI Text for details. water, essentially as described (15, 40). For GFP detection, filter 49002 (Chroma) was used. CFP and YFP fluorescence were detected with a dual ACKNOWLEDGMENTS. We thank Alex Formstone, Ian Selmes, Ying Li, CFP/YFP-ET filter (89002, Chroma). For dual color acquisition the YFP channel Vero´nica Labrador, Carlos Sa´nchez, and Jose´M.La´ zaro for their help and Juan was imaged first, followed immediately by the CFP channel. See SI Text for Carlos Alonso (Centro Nacional de Biotecnologı´a, Madrid, Spain), Dennis details. Bamford (University of Helsinki, Helsinki, Finland), Kenn Gerdes (Institute for Cell and Molecular Biosciences, Newcastle, United Kingdom), and Peter Grau- mann (University of Freiburg, Freiburg, Germany) for phage SPP1, phage Real-Time PCR. Cells corresponding to 1-mL aliquots of B. subtilis or E. coli PRD1, strains MC1000 and MC1000⌬mreB, and strains JS64 and JS65, respec- cultures, withdrawn at different times after infection, were harvested, pro- tively. This work was supported by Spanish Ministry of Education and Science cessed, and analyzed by real-time PCR essentially as described (41). Grants BFU2005-00733 (to M.S.) and BFU2005-01878 (to W.J.J.M.), Spanish Ministry of Science and Innovation Grant BFU2008-00215 (to M.S.), Spanish Purification of MreB Complexes. Strain YK827 was grown in LB medium Ministry of Education and Science Grant Consolider-Ingenio 2010 24717 (to containing 10 mM MgSO and 0.25% xylose at 37 °C. When the cells reached M.S.) and by an Institutional grant from Fundacio´n Ramo´n Areces to the 4 Centro de Biología Molecular ‘‘Severo Ochoa.’’ D.M.-E. and V.C.-L. are holders an OD of 0.4–0.5, the culture was infected with phages ␾29, ␾29 600 of an I3P contract from the Spanish National Research Council and a predoc- ␾ sus14(1242), or 29 sus14(1242)/sus16.7(48) and then treated with 1% form- toral fellowship from the Spanish Ministry of Education and Science, respec- aldehyde for 10 min. Glycine was added at a final concentration of 150 mM to tively. Work in the Errington laboratory was supported by grants from the quench the reaction. Purification of protein complexes with His-tagged MreB Biotechnology and Biological Sciences Research Council and the Human Fron- proteins was performed as described (42). To identify proteins in the complex, tier Science Program.

1. Jones LJ, Carballido-Lo´pez R, Errington J (2001) Control of cell shape in bacteria: helical, 22. Salas M (1991) Protein-priming of DNA replication. Annu Rev Biochem 60:39–71. actin-like filaments in Bacillus subtilis. Cell 104:913–922. 23. Meijer WJ, Lewis PJ, Errington J, Salas M (2000) Dynamic relocalization of phage ␾29 2. Van den Ent F, Amos LA, Lowe J (2001) Prokaryotic origin of the actin cytoskeleton. DNA during replication and the role of the viral protein p16.7. EMBO J 19:4182–4190. Nature 413:39–44. 24. Meijer WJ, Serna-Rico A, Salas M (2001) Characterization of the bacteriophage ␾29- 3. Shih Y, Rothfield L (2006) The bacterial cytoskeleton. Microbiol Mol Biol Rev 70:729– encoded protein p16.7: A membrane protein involved in phage DNA replication. Mol 754. Microbiol 39:731–746. 4. Michie KA, Lo¨we J (2006) Dynamic filaments of the bacterial cytoskeleton. Annu Rev 25. Mun˜oz-Espín D, et al. (2004) Phage ␾29 DNA-replication organizer membrane protein Biochem 75:467–492. p16.7 contains a coiled-coil and a dimeric, homeodomain-related, functional domain. 5. Carballido-Lo´pez R (2006) The bacterial actin-like cytoskeleton. Microbiol Mol Biol Rev J Biol Chem 279:50437–50445. 70:888–909. 26. Asensio JL, et al. (2005) Structure of the functional domain of ␾29 replication orga- 6. Gitai Z (2007) Diversification and specialization of the bacterial cytoskeleton. Curr Opin nizer. Insights into oligomerization and DNA binding. J Biol Chem 280:20730–20739. Cell Biol 19:1–8. 27. Albert A, et al. (2005) Structural basis for membrane anchorage of viral ␾29 DNA during 7. Carballido-Lo´pez R, Errington J (2003) The bacterial cytoskeleton: In vivo dynamics of replication. J Biol Chem 280:42486–42488. the actin-like protein Mbl of Bacillus subtilis. Dev Cell 4:19–28. 28. Formstone A, Errington J (2005) A magnesium-dependent mreB null mutant: Implica- 8. Figge RM, Divakaruni AV, Gober JW (2004) MreB, the cell shape-determining bacterial tions for the role of mreB in Bacillus subtilis. Mol Microbiol 55:1646–1657. actin homologue, coordinates cell wall morphogenesis in Caulobacter crescentus. Mol 29. Karimova G, Pidoux J, Ullmann A, Ladant D (1998) A bacterial two-hybrid system based Microbiol 51:1321–1332. on a reconstituted signal transduction pathway. Proc Natl Acad Sci USA 95:5752–5756. 9. Gitai Z, Dye N, Shapiro L (2004) An actin-like gene can determine cell polarity in 30. Greber UF, Way M (2006) A superhighway of virus infection. Cell 124:741–754. bacteria. Proc Natl Acad Sci USA 101:8643–8648. 31. Firshein W (1989) Role of the DNA/membrane complex in prokaryotic DNA replication. 10. Kim SY, Gitai Z, Kinkhabwala A, Shapiro L, Moemer WE (2006) Single molecules of the Annu Rev Microbiol 43:89–120. bacterial actin MreB undergo directed treadmilling motion in Caulobacter crescentus. 32. Defeu Soufo HJ, Graumann PL (2005) Bacillus subtilis actin-like protein MreB influences Proc Natl Acad Sci USA 103:10929–10934. the positioning of the replication machinery and requires membrane proteins MreC/D 11. Defeu Soufo HJ, Graumann PL (2006) Dynamic localization and interaction with other and other actin-like proteins for proper localization. BMC Cell Biol 6:10. Bacillus subtilis actin-like proteins are important for the function of MreB. Mol 33. Meijer WJ, et al. (2005) Molecular basis for the exploitation of spore formation as Microbiol 62:1340–1356. survival mechanism by virulent phage ␾29. EMBO J 24:3647–3657. 12. Daniel RA, Errington J (2003) Control of cell morphogenesis in bacteria: Two distinct 34. Castilla-Llorente V, Mun˜oz-Espín D, Villar L, Salas M, Meijer WJ (2006) Spo0A, the key ways to make a rod-shaped cell. Cell 113:767–776. transcriptional regulator for entrance into sporulation, is an inhibitor of DNA replica- 13. Gitai Z (2005) The new bacterial cell biology: Moving parts and subcellular architecture. tion. EMBO J 25:3890–3899. Cell 120:577–586. 35. Jime´nez F, Camacho A, de la Torre J, Vin˜uela E, Salas M (1977) Assembly of Bacillus 14. Carballido-Lo´pez R (2006) Orchestrating bacterial cell morphogenesis. Mol Microbiol subtilis phage ␾29. 2. Mutants in the cistrons coding for the non-structural proteins. Eur 60:815–819. J Biochem 73:57–72. 15. Carballido-Lo´pez R, Formstone A, Li Y, Noirot P, Errington J (2006) Actin homolog 36. Leaver M, Errington J (2005) Roles for MreC and MreD proteins in helical growth of the MreBH governs cell morphogenesis by localization of the cell wall hydrolase LytE. Dev cylindrical cell wall in Bacillus subtilis. Mol Microbiol 57:1196–1209. Cell 11:399–409. 37. Sambrook J, Fritsch EF, Maniatis T (1989) in Molecular Cloning: A Laboratory Manual 16. Madabhushi R, Marians KJ (2009) Actin homolog MreB affects chromosome segrega- (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York). tion by regulating topoisomerase IV in Escherichia coli. Mol Cell 33:171–180. 38. Ward JB, Jr, Zahler SA (1973) Genetic studies of leucine biosynthesis in Bacillus subtilis. 17. Levin PA, Margolis PS, Setlow P, Losick R, Sun D (1992) Identification of Bacillus subtilis J Bacteriol 116:719–726. genes for septum placement and shape determination. J Bacteriol 174:6717–6728. 39. Lewis PJ, Errington J (1997) Direct evidence for active segregation of oriC regions of the 18. Varley AW, Stewart GC (1992) The divIVB region of the Bacillus subtilis chromosome Bacillus subtilis chromosome and co-localization with the Spo0J partitioning protein. encodes homologs of Escherichia coli septum placement (MinCD) and cell shape Mol Microbiol 25:945–954. (MreBCD) determinants. J Bacteriol 174:6729–6742. 40. Glaser P, et al. (1997) Dynamic, mitotic-like behavior of a bacterial protein required for 19. Abhayawardhane Y, Steward GC (1995) Bacillus subtilis possesses a second determi- accurate chromosome partitioning. Genes Dev 11:1160–1168. nant with extensive sequence similarity to the Escherichia coli mreB morphogene. J 41. Gonza´lez-Huici V, Salas M, Hermoso JM (2004) The push-pull mechanism of bacterio- Bacteriol 177:765–773. phage ␾29 DNA injection. Mol Microbiol 52:529–540. 20. Bravo A, Serrano-Heras G, Salas M (2005) Compartmentalization of prokaryotic DNA 42. Ishikawa S, Kawai Y, Hiramatsu K, Kuwano M, Ogasawara N (2006) A new FtsZ- replication. FEMS Microbiol Rev 29:25–47. interacting protein, YlmF, complements the activity of FtsA during progression of cell 21. Calendar R (2006) in The Bacteriophages (Oxford Univ Press, New York, Oxford). division in Bacillus subtilis. Mol Microbiol 60:1364–1380.

13352 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0906465106 Mun˜oz-Espín et al. Downloaded by guest on September 24, 2021