Dynamic localization of penicillin-binding proteins during spore development in Bacillus subtilis

Transcription

1 Microbiology (2005), 151, DOI /mic Dynamic localization of penicillin-binding proteins during spore development in Bacillus subtilis Dirk-Jan Scheffers3 Correspondence Dirk-Jan Scheffers Received 11 October 2004 Revised 16 November 2004 Accepted 19 November 2004 Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, UK During Bacillus subtilis spore formation, many membrane proteins that function in spore development localize to the prespore septum and, subsequently, to the outer prespore membrane. Recently, it was shown that the cell-division-specific penicillin-binding proteins (PBPs) 1 and 2b localize to the asymmetric prespore septum. Here, the author studied the localization of other PBPs, fused to green fluorescent protein (GFP), during spore formation. Fusions to PBPs 4, 2c, 2d, 2a, 3, H, 4b, 5, 4a, 4* and X were expressed during vegetative growth, and their localization was monitored during sporulation. Of these PBPs, 2c, 2d, 4b and 4* have been implicated as having a function in sporulation. It was found that PBP2c, 2d and X changed their localization, while the other PBPs tested were not affected. The putative endopeptidase PbpX appears to spiral out in a pattern that resembles FtsZ redistribution during sporulation, but a pbpx knockout strain had no distinguishable phenotype. PBP2c and 2d localize to the prespore septum and follow the membrane during engulfment, and so are redistributed to the prespore membrane. A similar pattern was observed when GFP PBP2c was expressed in the mother cell from a sporulation-specific promoter. This work shows that various PBPs known to function during sporulation are redistributed from the cytoplasmic membrane to the prespore. INTRODUCTION Bacteria display a high degree of subcellular organization, with proteins localizing to distinct sites in the cell at distinct times in the bacterial cell cycle. Examples include the assembly of cell-division proteins in a ring at the middle of the bacterial cell (Errington et al., 2003), the formation of dynamic helical filaments by the prokaryotic actin homologues MreB and ParM (Errington, 2003b; Gerdes et al., 2004), and the polar localization of, for example, chemoreceptors (Maddock & Shapiro, 1993). A process that involves both redistribution of cellular proteins and timeand compartment-dependent gene expression in bacteria is the formation of endospores. Endospores, which are very tough survival structures formed upon nutrient starvation, can be formed by several ancient lineages of bacteria. The development of spores has been best studied in the Grampositive model organism Bacillus subtilis (for recent reviews see Errington, 2003a; Hilbert & Piggot, 2004; Piggot & Losick, 2001). Recent work from various groups, focusing on the (re)distribution of several proteins during sporulation, has revealed 3Present address: Molecular Microbiology, Faculty of Earth and Life Sciences, Vrije Universiteit, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands. Abbreviations: CH, casein hydrolysate; GFP, green fluorescent protein; PBP, penicillin-binding protein; PG, peptidoglycan. intriguing examples of protein localization and distribution. At the start of sporulation, an asymmetric division septum is formed (a schematic overview of division and engulfment is shown in Fig. 1A). The switch from medial to asymmetric division is accomplished by an upshift in the expression of the key cell-division gene ftsz, and a concomitant redistribution of FtsZ from a medial ring to a spiral that extends to both poles of the cell (Ben-Yehuda & Losick, 2002). The spiral pattern is also observed for FtsA and EzrA (Ben-Yehuda & Losick, 2002), which are other components of the division machinery that localize early in the division process (for a recent overview of B. subtilis cell division see Errington et al., 2003). One of the two possible asymmetric division sites is then committed to division by the action of SpoIIE (Barák & Youngman, 1996; Feucht et al., 1996), resulting in the localization of other cell-division proteins to the asymmetric division site, and subsequent division (see Errington, 2003a; Hilbert & Piggot, 2004). Following asymmetric division, the asymmetric septum is degraded, and the mother cell membrane migrates around the prespore membrane in a phagocytosis-like process called engulfment. Abanes-De Mello et al. (2002) showed that membrane migration during engulfment depends on the activity of three proteins expressed in the mother cell, SpoIID, SpoIIM and SpoIIP, and that these three proteins are associated with the leading edge of the engulfing membrane. SpoIID is a cell wall hydrolase which is required not G 2005 SGM Printed in Great Britain 999

2 D.-J. Scheffers Fig. 1. Schematic overview of sporulation and observed redistribution of GFP PBP2c and GFP PBP2d. The cell wall is shown in grey, membranes in black, and GFP PBP molecules as white ellipsoids. (A) 1, Vegetative cell with random distribution of GFP PBP. 2, 3, Asymmetric division. The number of PBPs ending up in what is going to be the prespore inner membrane is lower than the number of PBPs ending up in the mother cell membrane. 4, Septal thinning and recruitment of GFP PBP to the prespore septum. 5, Engulfment. GFP PBPs follow the engulfing mother cell membrane. 6, After membrane fusion, the outer prespore membrane has a high concentration of GFP PBP2c or 2d. (B) An alternative model. 2, GFP PBPs are recruited to the asymmetric division septum before division is complete. 3, This leads to an increased concentration of GFP PBPs on both sides of the asymmetric septum. 5 and 6, GFP PBPs follow the membrane during engulfment, and end up evenly distributed over the outer and inner prespore membranes. only for septal thinning at the onset of engulfment, but also for the completion of engulfment (Abanes-De Mello et al., 2002). The authors suggested an attractive model in which the cell wall acts as a scaffold along which the hydrolase tracks, dragging the mother cell membrane and other engulfment proteins along (Abanes-De Mello et al., 2002). The DNA translocase SpoIIIE also migrates at the leading edge of the engulfing membrane, and is involved in membrane fusion when the engulfing membranes meet (Sharp & Pogliano, 1999), so that the prespore is now surrounded by an inner membrane and a mother-cellderived outer membrane. Work from the same group has shown that prespore-expressed membrane proteins also localize to the septum, track along with the engulfing mother cell membrane, and then spread out along the inner prespore membrane, sometimes to assemble into regular structures (Rubio & Pogliano, 2004). The tracking of these prespore-expressed proteins is dependent on an unidentified mother-cell-expressed protein, suggesting that the proteins interact with a localized partner protein made in the mother cell or with parts of the septal peptidoglycan (PG) (Rubio & Pogliano, 2004). Dynamic localization of proteins from the mother cell to the prespore has also been shown for the morphogenetic protein SpoIVA, which is involved in assembly of the spore cortex and coat (Lewis & Errington, 1996), and for SpoIVFB, a polytopic membrane protein involved in the activation of a transcription factor (Rudner et al., 2002). In the latter case, Rudner et al. (2002) showed that SpoIVFB is inserted into the cytoplasmic membrane in a dispersed fashion, after which the protein diffuses to, and is captured in, the outer prespore membrane (the diffusion-andcapture model). This work is concerned with the localization of penicillinbinding proteins (PBPs), which are proteins involved in the synthesis of the cell wall PG during vegetative growth, cell division and sporulation (for recent reviews on cell wall synthesis during growth and sporulation, see Foster & Popham, 2001; Popham, 2002) (Table 1). Early work on expression profiles of PBPs during vegetative growth and sporulation indicated roles for PBPs 2B, 3, 4* and 5* in sporulation (Sowell & Buchanan, 1983; Todd et al., 1983). Studies on mutant strains, transcriptional profiling and localization have now identified a number of PBPs as playing (putative) roles during sporulation. The class A bifunctional transglycosylase/transpeptidase PBP1 is part of the division machinery that operates during asymmetric division, and is required for efficient sporulation (Scheffers & Errington, 2004). Two other class A PBPs, 2c and 2d, play a redundant role in spore PG synthesis. A strain in which the genes for these PBPs 2c and 2d (pbpf and pbpg, respectively) are inactivated is incapable of completing sporulation, and shows defects in spore PG synthesis (McPherson et al., 2001). The class B transpeptidase PBP2b, the only essential PBP in B. subtilis, is required for the asymmetric cell division, and localizes to the asymmetric septum (Daniel et al., 2000). Another class B PBP, SpoVD, is essential for spore formation, and is required for the synthesis of cortical PG (Daniel et al., 1994), whereas PBP4b does not seem to have an effect on spore PG, but is expressed under the control of the mother-cell-specific s E factor (Eichenberger et al., 2003; Wei et al., 2004). Finally, 1000 Microbiology 151

3 Localization of PBPs during sporulation Table 1. Summary of B. subtilis PBPs implicated in sporulation Gene Protein Expression/s-factor dependency Expressed in* Dynamic localization during sporulation Class A PBPs (bifunctional transglycosylase/transpeptidase) pona PBP1a/b VegD (Popham & Setlow, 1995) MC Yes, to sporulation septum (Scheffers & Errington, 2004) pbpf PBP2c s A, s G (Popham & Setlow, 1993a) MC, PS Yes, to prespore pbpg PBP2d s F, s G (McPherson et al., 2001) PS Yes, to prespore Class B PBPs (transpeptidase) pbpb PBP2b Veg (Daniel et al., 1996) MC Yes, to sporulation septum (Daniel et al., 2000) pbpc PBP3 Veg, spor (low) (Murray et al., 1996) MC No spovd SpoVD s E (Daniel et al., 1994) MC ND pbpi PBP4b s E (Eichenberger et al., 2003; Wei et al., 2004) MC No Low-M r PBPs (carboxypeptidase) daca PBP5 Veg, spord (Todd et al., 1986) Not known No dacb PBP5* s E (Buchanan & Ling, 1992; Simpson et al., 1994) MC Not detected dacf DacF s F (Schuch & Piggot, 1994; Wu et al., 1992) PS ND Low-M r PBPs (endopeptidase) pbpe PBP4* s A (Popham & Setlow, 1993b), s W (Zellmeier et al., 2003) MC No pbpx PbpX s X (Cao & Helmann, 2004) MC Yes, to both asymmetric septa and the prespore ND, Not determined. *MC, mother cell; PS, prespore. DVeg, expression during vegetative growth. dbased on the presence of protein in exponentially growing and sporulating cells. two low-molecular-mass PBPs are involved in spore PG synthesis. These are the carboxypeptidases 5* and DacF, which have partially redundant roles in regulating the degree of cross-linking of the spore PG, and a double mutant for both proteins has decreased spore heat resistance (Popham et al., 1995, 1999). In this study, I made use of a recently constructed collection (Scheffers et al., 2004) of fusions of green fluorescent protein (GFP) to PBPs of B. subtilis to study the localization of these proteins during sporulation. I found that PBP2c and 2d localize to the prespore division septum, and follow the leading edge of the mother cell membrane during engulfment. Also, a localization pattern was identified for the as yet uncharacterized PbpX that resembles the redistribution of FtsZ upon the start of sporulation. A pbpx knockout strain has no obvious phenotype. These results show that for some PBPs, sporulation induces a dynamic redistribution of their localization patterns. METHODS General methods. The strains and plasmids used in this study are listed in Table 2. B. subtilis cells were made competent for transformation with DNA either by the method of Kunst & Rapoport (1995), or by the method of Anagnostopoulos & Spizizen (1961) as modified by Jenkinson (1983). DNA manipulations and Escherichia coli DH5a transformations were carried out using standard methods (Sambrook et al., 1989). Solid medium used for growing B. subtilis was nutrient agar (Oxoid), and liquid medium was either casein hydrolysate (CH) medium (Sterlini & Mandelstam, 1969) or S medium (Sharpe et al., 1998) supplemented with 1 % (v/v) CH (S+), with antibiotics added as required. Chloramphenicol was used at 5 mg ml 21, spectinomycin at 50 mg ml 21, erythromycin at 0?5 mg ml 21, lincomycin at 12?5 mg ml 21, and kanamycin at 5 mg ml 21. Media used for growing E. coli were 26 TY (tryptone yeast extract medium; Sambrook et al., 1989) and nutrient agar supplemented with ampicillin (100 mg ml 21 ) as required. Construction of GFP fusions. This was performed as described previously (Scheffers et al., 2004). Approximately one-third of each of the promoter-proximal parts of the pbpg, spovd, dacb, dacf and pbpe genes was amplified using PCR, and cloned into psg4902 (Wu & Errington, 2003). All primers and restriction endonucleases used are listed in Table 3. Transformation of the resulting plasmids (Table 2) into B. subtilis, with selection for chloramphenicol resistance resulted in several strains, each of which carried a gfp fusion to a gene of interest at the chromosomal locus as the only copy of the gene of interest, and under the control of the P xyl promoter. Correct integration at the chromosomal locus was confirmed by PCR. The vectors pmds13 and pmds14, which contain gfp under the control of the prespore-specific P spoiiq promoter and the mothercell-specific P spoiid promoter, respectively (Sharp & Pogliano, 2002), were used for the construction of N-terminal GFP fusions to PBPs that are expressed during sporulation. Full-length pbpf, pbpg and pbpx were amplified by PCR with primer sets DJS163 DJS164, DJS165 DJS166 and DJS175 DJS176, respectively. The PCR products were digested with EagI, and ligated into EagI-digested pmds13 or pmds14. The correct orientation of the inserts was confirmed by PCR and sequencing. The resulting plasmids (Table 2) were transformed

4 D.-J. Scheffers Table 2. Bacterial strains and plasmids Strain/plasmid Relevant characteristics Source/construction B. subtilis 168 trpc2 Laboratory collection BFA1208 trpc2 pbpg ::pmutin4 Kobayashi et al. (2003) PS1869 trpc2 pbpf ::erm Popham & Setlow (1996) 3103 trpc2 pbpa : : psg5043 (cat P xyl gfp* pbpa ) Scheffers et al. (2004) 3104 trpc2 dacc : : psg5044 (cat P xyl gfp dacc ) Scheffers et al. (2004) 3105 trpc2 pbpc : : psg5045 (cat P xyl gfp pbpc ) Scheffers et al. (2004) 3106 trpc2 pbpe : : psg5046 (cat P xyl gfp pbpe ) psg5046r168 (Cm R )D 3107 trpc2 pbpx : : psg5047 (cat P xyl gfp pbpx ) Scheffers et al. (2004) 3108 trpc2 dacb : : psg5048 (cat P xyl gfp dacb ) psg5048r168 (Cm R ) 2081 trpc2 yrrr : : psg1489 (cat P xyl gfp pbpi ) Scheffers et al. (2004) 2082 trpc2 pbpd : : psg1490 (cat P xyl gfp pbpd ) Scheffers et al. (2004) 2083 trpc2 pona : : psg1492 (cat P xyl gfp pona ) Scheffers et al. (2004) 2084 trpc2 pbpf : : psg1491 (cat P xyl gfp pbpf ) Scheffers et al. (2004) 2085 trpc2 daca : : psg1493 (cat P xyl gfp daca ) Scheffers et al. (2004) 3510 trpc2 pbpg : : psg5309 (cat P xyl gfp pbpg ) psg5309r168 (Cm R ) 3905 trpc2 pbpx ::psg5313 (erm P spac pbpx lacz laci) psg5313r168 (Em R ) 3906 trpc2 pbpx ::kan This work 3910 trpc2 pbpx ::kan ::cat P xyl gfp pbpf chr DNAR3906 (Cm R ) 3911 trpc2 pbpx ::kan ::cat P xyl gfp pbpg chr DNAR3906 (Cm R ) 3912 trpc2 pbpg ::pmutin4 ::cat P xyl gfp pbpf chr DNARBFA1208 (Cm R ) 3913 trpc2 pbpg ::pmutin4 ::cat P xyl gfp pbpx chr DNARBFA1208 (Cm R ) 3914 trpc2 pbpf ::erm ::cat P xyl gfp pbpx chr DNARPS1869 (Cm R ) 3915 trpc2 pbpf ::erm ::cat P xyl gfp pbpg chr DNARPS1869 (Cm R ) 3920 trpc2 amye ::P spoiiq gfp pbpf Vcat psg5335r168 (Cm R ) 3921 trpc2 amye ::P spoiid gfp pbpf Vcat psg5336r168 (Cm R ) 3922 trpc2 amye ::P spoiiq gfp pbpg Vcat psg5337r168 (Cm R ) 3923 trpc2 amye ::P spoiid gfp pbpg Vcat psg5338r168 (Cm R ) 3932 trpc2 amye ::P spoiiq gfp pbpx Vcat psg5347r168 (Cm R ) 3933 trpc2 amye ::P spoiid gfp pbpx Vcat psg5348r168 (Cm R ) E. coli DH5a F 2 enda1 hsdr17 supe44 thi-1 l-reca1 gyra96 rela1 D(lacZYA argf)u169 w80 dlaczdm Microbiology 151 Gibco-BRL Plasmids pkm1 kan Laboratory collection pmutin4 bla erm P spac lacz laci Vagner et al. (1998) pmds13 amye ::P spoiiq gfp Vcat Sharp & Pogliano (2002) pmds14 amye ::P spoiid gfp Vcat Sharp & Pogliano (2002) psg5313 bla erm P spac pbpx lacz laci This work psg4902 bla cat P xyl gfp Wu & Errington (2003) psg5308 bla cat P xyl gfp spovd This work psg5048 bla cat P xyl gfp dacb This work psg5323 bla cat P xyl gfp dacf This work psg5046 bla cat P xyl gfp pbpe This work psg5309 bla cat P xyl gfp pbpg This work psg5335 amye ::P spoiiq gfp pbpf Vcat This work psg5336 amye ::P spoiid gfp pbpf Vcat This work psg5337 amye ::P spoiiq gfp pbpg Vcat This work psg5338 amye ::P spoiid gfp pbpg Vcat This work psg5347 amye ::P spoiiq gfp pbpx Vcat This work psg5348 amye ::P spoiid gfp pbpx Vcat This work *gfp, F64L, S65T variant of GFP (GFPmut1; Cormack et al., 1996). DThe arrow indicates transformation to chloramphenicol resistance with plasmid psg5046.

5 Localization of PBPs during sporulation Table 3. Primers Restriction sites are underlined. Gene/primer Primer sequence (5 3 ) Restriction endonuclease pbpg GTCGGATCCTTGTGGATGCAATGACAAAT BamHI AAGCTCGAGGCGCCGTGGCCGTAG XhoI spovd GTCGGATCCTCTTGCGCGTCTCGAATG BamHI AAGCTCGAGGGCCTCAGCGTTATCAAGC XhoI dacb GTCGGATCCCTATGCGCATTTTCAAA BamHI AAGCTCGAGACCAAGCTGCTCGGC XhoI dacf GTCGGATCCAGATGAAACGTCTTTTATCC BamHI AAGCTCGAGTTCTTCTTCAGAGCCGG XhoI pbpe GTCGGATCCCTATGAAGCAGAATAAAAGAAAGC BamHI AAGCTCGAGGCTCATGCCAGATGC XhoI DJS128 CAGAAGCTTGGAATTTATTTTGGGAGG HindIII DJS129 AAGGGATCCATAGCAGTTGCAATCAGCG BamHI DJS149 GATAGGAATCACGCCAAGA DJS150 GTCGGATCCTTCCGCTGAAACCGATCT BamHI DJS151 GCTGAATTCCAAGCGCTGATAGACGGAA EcoRI DJS152 GTGGCTTCCTTCTACTTCAT DJS163 GTACGGCCGATGTTTAAGATAAAGAAAAAGAAAC EagI DJS164 AAGCGGCCGTTAAGAGGAAAACAATTTTGGC EagI DJS165 GTACGGCCGGTGGATGCAATGACA EagI DJS166 AAGCGGCCGTCACAAAATATGTGTAATGC EagI DJS175 GTACGGCCGATGACAAGCCCAACC EagI DJS176 AAGCGGCCGTCATTCTTGATTTAGAAGC EagI km3 GGGGGATCCAAGACGAAGAGGATGAAG BamHI km4 CCCGAATTCAGAGTATGGACAGTTGCG EcoRI into strain 168, with selection for chloramphenicol resistance, and integration at the amye locus was confirmed by screening the transformants for failure to degrade starch. Construction of pbpx mutant strains. Using primer pairs DJS149 DJS150 and DJS151 DJS152 (Table 3), ~1?6 kb PCR fragments were generated containing the first 292 bp of pbpx plus upstream sequences, and the last 257 bp of pbpx plus downstream sequences. These fragments were cut with BamHI and EcoRI, respectively, and ligated to a BamHI EcoRI-digested PCR product containing a neo cassette, which was generated using primers km3 and km4, and plasmid pkm1 as a template. The ligation product was subjected to another PCR reaction using primers DJS149 and DJS152, and the resulting PCR product was transformed into strain 168, with selection for kanamycin resistance, generating strain 3906, which contains a deletion of pbpx codons (out of 391 codons). Correct integration of the ligation product into the chromosome was confirmed by PCR and sequencing. Using primer pair DJS128 DJS129, a fragment of pbpx (bp ) was amplified by PCR. The fragment was cut with HindIII and BamHI, and ligated into HindIII BamHI-digested pmutin4, a vector that allows inactivation of the target gene as well as the monitoring of its expression through a transcriptional lacz fusion (Vagner et al., 1998), generating psg5313. psg5313 was transformed into strain 168 to give strain Correct integration of the plasmid into the chromosome was confirmed by PCR. Sporulation methods. Sporulation was induced by growth to OD 600 ~0?8 in CH, followed by resuspension in a starvation medium (SM; Partridge & Errington, 1993; Sterlini & Mandelstam, 1969). Cell pellets were washed with SM prior to resuspension to remove xylose, unless stated otherwise. Time zero (T 0 ) was defined as the point at which the cells were resuspended in the starvation medium. b-galactosidase activity was assayed as described by Errington (1986). One unit of b-galactosidase catalyses the production of 1 nmol 4-methylumbelliferone min 21. Alkaline phosphatase activity was measured as described by Errington & Mandelstam (1983) and Glenn & Mandelstam (1971). Sporulation efficiency was tested by determining the number of heat-resistant spores formed in the cultures at 10 h (T 10 )or25h(t 25 ). Microscopy. Microscopy was performed essentially as described previously (Scheffers et al., 2004). Image acquisition was done as described by Lewis & Errington (1997), using Metamorph version 6.0 software (Universal Imaging Corporation). DNA was stained with Hoechst (1 mg ml 21 ; Molecular Probes). Membranes were stained with FM95.5 (4 mg ml 21 ; Molecular Probes). Images from a single focal plane were deconvolved using the No Neighbours algorithm from the Metamorph software package. Overlays of micrographs were assembled using Metamorph, before exporting the images to Adobe Photoshop version 6.0. RESULTS Construction of GFP PBP2d and GFP PBP4*, and localization during vegetative growth I set out to study the localization of PBPs during sporulation in B. subtilis. Previously, GFP fusions were constructed

6 D.-J. Scheffers to 11 PBPs expressed during vegetative growth (Scheffers et al., 2004). To include additional PBPs specific for sporulation, a similar strategy was attempted for the pbpg, spovd, dacb, dacf and pbpe genes (encoding PBP2d, SpoVD, PBP5*, DacF and PBP4*, respectively, none of which are essential). This approach was successful for fusions to pbpg and pbpe. A strain containing a genetic fusion to dacb was obtained, but when it was studied under the microscope, the GFP PBP5* fluorescence signal was not detected, so this strain was not studied in more detail. Unfortunately, several attempts to obtain GFP fusions to spovd and dacf using this strategy also failed. When expressed during vegetative growth, GFP PBP2d localized in a dispersed fashion along the membrane, whereas GFP PBP4* localized in a punctate pattern (Fig. 2). These patterns were similar to the dispersed or punctate localization patterns observed with most PBPs expressed during vegetative growth (Scheffers et al., 2004). PBP2c, PBP2d and PbpX localize to the prespore During sporulation, the cell-division-specific PBPs PBP2b and PBP1 localize to the asymmetric septum formed during sporulation (Daniel et al., 2000; Scheffers & Errington, 2004). To determine the localization of other PBPs during sporulation, the GFP PBP fusion strains were grown in the presence of 0?5 % xylose to allow expression of the fusion protein. At OD 600 ~0?8, cells were washed with sporulation salts to remove xylose, since the presence of xylose delays the onset of sporulation. Subsequently, cells were induced to sporulate using the resuspension method (see Methods), and the localization of the GFP PBP fusions was followed. Using this method, PBP2c and PbpX were seen to accumulate specifically at the prespore (Fig. 3A, F; Table 1). PBP2c arrives at the prespore septum, and then follows the mother cell membrane during engulfment. After completion of engulfment, GFP PBP2c is localized around the prespore. Whether GFP PBP2c is predominantly present in the inner or outer prespore membrane Fig. 2. Localization of GFP PBP2d and GFP PBP4* during vegetative growth. Fluorescence micrographs of cells expressing GFP PBP2d (A) and GFP PBP4* (B). Illumination for fluorescence was 1 s. Bar, 5 mm. cannot be distinguished (see Discussion). PbpX seemed to localize earlier during sporulation, being detected at both poles before cells committed to form the asymmetric septum (Fig. 4C), and subsequently followed the mother cell membrane during engulfment (Fig. 3F). None of the other PBPs tested, PBP4, PBP2a, PBP3, PbpH, PBP4b, PBP5, PBP4a and PBP4*, showed a similar localization pattern (Fig. 3, and data not shown). Also, GFP PBP2d fluorescence was not detected above background levels at 2h(T 2 ) and 3 h (T 3 ), even though PBP2d plays a role in sporulation (Fig. 3B; McPherson et al., 2001). However, when cells were resuspended in the presence of xylose, PBP2d was readily detected at the prespore in a pattern similar to that of PBP2c (Fig. 3C). The localization patterns of all of the other PBPs studied were not affected by the addition of xylose to the sporulation medium. Western blotting of cultures resuspended in the absence of xylose confirmed that the levels of GFP PBP2c, GFP PbpX, GFP PBP4b and GFP PBP4* were not altered significantly 2 h into sporulation, but that GFP PBP2d could no longer be detected, indicating that it is subject to proteolysis (data not shown). Also, when resuspended in the absence of xylose, strains expressing GFP PBP2c, GFP PBP2d, GFP PBP4b and GFP PbpX had similar spore counts compared to wild-type, showing that the GFP PBPs do not interfere with sporulation (data not shown). GFP PbpX appears to spiral out to both asymmetric sporulation division sites GFP PbpX localizes to the division septum during vegetative growth (Scheffers et al., 2004). After resuspending GFP PbpX-expressing cells in sporulation medium, a quick change in its localization pattern was detected. Typically, between T 1 and T 2, before a complete asymmetric division septum was formed, the PbpX signal became dispersed along the membrane, with differences in signal intensity along the membrane (Fig. 4A). In some cases, these signals looked like spirals (Fig. 4B) similar to the spirals observed for FtsZ, FtsA and EzrA in sporulating cells (Ben-Yehuda & Losick, 2002). The spiral pattern was more apparent when the image was deconvolved. However, these spiral structures were only observed in rare cases. Interestingly, GFP PbpX was seen to appear at both poles, often with a slight difference in intensity on either pole (Fig. 4C). Polar localization of the cell-division-specific PBPs PBP1 and PBP2b was seen only in cells that appeared to already have committed one of the potential polar division sites to become the asymmetric septum (Daniel et al., 2000; Scheffers & Errington, 2004). Thus, GFP PbpX localizes to both poles, similar to cell-division proteins like FtsZ, and the sporulation-specific SpoIIE (Levin & Losick, 1996; Levin et al., 1997). A pbpx knockout strain has no distinguishable phenotype Given the interesting localization pattern observed with GFP PbpX, I decided to study pbpx in more detail. pbpx 1004 Microbiology 151

7 Localization of PBPs during sporulation Fig. 3. Localization of GFP PBP fusions during sporulation. Fluorescence micrographs of cells expressing GFP PBP2c (A), GFP PBP2d (B, C), GFP PBP4b (D), GFP PBP4* (E) and GFP PbpX (F) are shown as overlays of the GFP image (green) and DNA staining (red). Images were taken directly after resuspension of the cells in sporulation medium (T 0 ), and 2 h (T 2 ) and 3 h (T 3 ) after resuspension. Cells were resuspended in sporulation medium without xylose, except for cells expressing GFP PBP2d, which were resuspended in the absence (B) and presence (C) of xylose. Bar, 5 mm. (A; T 2 ), (B; T 0 and T 3 ) and (D; T 0 and T 2 ) are composite images. was identified as a gene encoding an endopeptidase based upon sequence similarity (Foster & Popham, 2001), and it has recently been described as part of the s X regulon (Cao & Helmann, 2004). pbpx was inactivated in two ways: by replacing 624 internal bases from the gene with a neo resistance marker, and by use of the pmutin-4 vector, which generates a lacz transcriptional fusion to pbpx allowing the determination of the pbpx expression pattern (Methods). Both knockout strains grew at an identical rate, and with similar spore counts compared to wild-type B. subtilis (Table 4). Correct formation of the asymmetric sporulation septum was followed by expression of s E - dependent genes, since activation of this sigma factor is dependent on septation (Piggot & Losick, 2001). The s E - dependent synthesis of alkaline phosphatase was measured for both strains, and was found to be indistinguishable from wild-type (result not shown), showing that deletion of pbpx has no effect on septation during sporulation. The appearance of the DpbpX strain was indistinguishable from that of the wild-type (Table 4). The transcriptional activation of pbpx followed a pattern typical for weak expression during vegetative growth, with no induction upon sporulation (result not shown). pbpx is part of the s X regulon, which is thought to modulate aspects of cell envelope metabolism, possibly through regulation of cell-surface modification (Cao & Helmann, 2004; Huang & Helmann, 1998). A sigx-null mutant is impaired in its ability to survive at high temperature (Huang et al., 1997). To determine whether this effect is mediated by pbpx, the survival of the wild-type and the DpbpX strain was tested, after transferring exponentially growing cells to 54 uc for 30 min. Survival of the DpbpX strain was similar to that of the wild-type (Table 4), showing

8 D.-J. Scheffers Fig. 4. GFP PbpX appears to form spirals, and localizes to both polar division sites early in sporulation. GFP fluorescence of GFP PbpX is dispersed along the membrane (A) and sometimes appears to spiral out (B) from the midcell division septum to the poles upon sporulation. Cells were photographed 80 min after resuspension in sporulation medium. From left to right, a fluorescence micrograph, a deconvolved image of the micrograph, and an overlay of the deconvolved GFP image (green) and DNA staining (red). (C) GFP PbpX is present at both polar division sites (cells indicated with arrows). Fluorescence micrographs and an overlay of the GFP image (green) and DNA staining (red) are shown for cells 60 min (i) and 80 min (ii) after resuspension in sporulation medium. Bars (notice size difference in B), 5 mm. that pbpx is not responsible for the decrease of heat resistance observed in a sigx null strain. It is concluded that pbpx is a non-essential gene in B. subtilis, with no obvious phenotype during vegetative growth or sporulation. PBP2c and PBP2d localize independently of each other PBP2c and PBP2d play redundant roles during sporulation, and the presence of at least one of these PBPs is required for the synthesis of the spore germ cell wall (McPherson Table 4. Properties of pbpx knockout strains Strain Phenotype T d (min)* Mean cell length (mm)d Mean cell width (mm) Heat-resistant spores at T 10 (c.f.u. ml 1 )d Survival at 54 6C 168 Wild-type 33 2?23±0?47 (n=156) 0?93±0?08 (n=156) 3? ?6% 3905 pbpx ::pmutin4 33 ND ND 2? ND 3906 pbpx ::kan 32 2?23±0?48 (n=170) 0?90±0?08 (n=170) 2? ?1% ND, Not determined *Doubling times (T d ) were determined for exponentially growing cells in CH medium. DCell membranes were stained with FM95.5, and cell length and width were determined. dheat resistance was determined by incubating sporulating cells 10 h (T 10 ) after resuspension for 20 min at 80 uc. Cells from an exponentially growing culture at 37 uc were split, and incubated for a further 30 min at either 37 or 54 uc. The c.f.u. in each culture were determined, and survival is expressed as the percentage of c.f.u. present in the culture grown at 54 uc, with the culture grown at 37 uc set at 100 % Microbiology 151

9 Localization of PBPs during sporulation Fig. 5. PBP2c and PBP2d localize to the prespore independently of each other. Fluorescence micrographs of cells from strain 3912 containing an insertional inactivation of pbpg, and expressing GFP PBP2c (A), and from strain 3915 containing a pbpf knockout, and expressing GFP PBP2d (B). Images were taken 2 h after resuspending the cells for sporulation. Xylose was added to the sporulation medium for strain Bar, 5 mm. (A) is a composite image. et al., 2001). The finding that both GFP PBP2c and GFP PBP2d localize to the prespore was in line with this observation. The localization of both GFP fusion proteins in the absence of the other PBP was studied. As shown in Fig. 5, each of the proteins GFP PBP2c and GFP PBP2d was able to localize in the absence of the other protein. Again, for GFP PBP2d, xylose had to be present in the sporulation medium for continued synthesis. This shows that these proteins are not dependent on each other for correct localization, as might be expected from the fact that their functions appear to be redundant during sporulation (McPherson et al., 2001). The sporulation efficiency of a DpbpF strain with gfp pbpg under the control of P xyl (strain 3915), when grown and sporulated without xylose, was less than 0?3 % of that for the wild-type and the DpbpF strain (Table 5). A similar result was obtained with a DpbpG strain with gfp pbpf under control of P xyl (strain 3912). This was expected, since either PBP2c or PBP2d is required for efficient sporulation (McPherson et al., 2001). Xylose fully restored sporulation for strains 3912 and 3915, showing that the presence of GFP PBP2c or GFP PBP2d in the absence of PBP2d or PBP2c, respectively, is sufficient for sporulation (Table 5). This shows that GFP PBP2c and GFP PBP2d are fully functional. The presence of xylose in both the growth and the sporulation medium increased the spore count in the single-knockout strains (PS1869 and BFA1208). This was because the extra carbon source in the sporulation medium allowed the cells to continue growing, causing both a delay of the initiation of sporulation and an increase in the number of cells at the initiation of sporulation. Localization of GFP PBPs under control of sporulation-specific promoters Localization of GFP PBP2c, GFP PBP2d and GFP PbpX to the prespore was observed with protein expressed in the predivisional cell from the P xyl promoter. I wanted to test whether similar patterns could be observed when the proteins were expressed only during sporulation, with expression switched on in either the mother cell or the prespore compartment. Compartment-specific expression makes it possible to distinguish targeting to the outer prespore membrane from targeting to the inner prespore membrane. Vectors were used that allow the expression of N-terminal GFP fusion proteins under control of the prespore-specific P spoiiq promoter or the mother-cell-specific P spoiid promoter (kindly provided by Dr K. Pogliano; see Sharp & Pogliano, 2002). Expression from both promoters resulted in very strong fluorescence signals that required twodimensional deconvolution to provide sufficient resolution. As shown in Fig. 6, expression of GFP PBP2c, GFP PBP2d and GFP PbpX under control of the prespore-specific Table 5. GFP PBP2c and GFP PBP2d are functional Strain Phenotype Xylose added* Heat-resistant spores at T 25 (c.f.u. ml 1 )D 168 Wild-type 2 2? PS1869 pbpf ::erm 2 2? ? pbpf ::erm; P xyl gfp pbpg 2 5? ? BFA1208 pbpg ::pmutin4 2 2? ? pbpg ::pmutin4; P xyl gfp pbpf 2 3? ? *Xylose (0?5 %) was added to both the growth medium and the sporulation medium. D, Heat resistance was determined by incubating sporulating cells 25 h (T 25 ) after resuspension for 20 min at 80 uc

10 D.-J. Scheffers Fig. 6. GFP PBP2c is targeted to the prespore outer membrane when expressed under control of P spoiid. Fluorescence micrographs of cells are shown as overlays of the deconvolved GFP image (green) and DNA staining (red). Cells were expressing GFP PBP2c (A), GFP PBP2d (B) and GFP PbpX (C) under the control of the prespore-specific promoter P spoiiq, and under the control of the mother-cell-specific promoter P spoiid. Cells were processed for microscopy 3 h (T 3 ) and 4h(T 4 ) after resuspension in sporulation medium. Bar, 5 mm. promoter P spoiiq resulted in rather uniform labelling of the inner prespore membrane, although for GFP PBP2d, signal throughout the prespore was also observed, which is attributed to the instability of GFP PBP2d (Fig. 6B, T 4 ). This instability would also explain the uniform fluorescence in the cytoplasm when GFP PBP2d was expressed under control of the mother-cell-specific promoter P spoiid (Fig. 6B). In contrast to this, GFP PBP2c, when expressed in the mother cell, localized to the outer prespore membrane in a pattern that shows that GFP PBP2c follows the mother cell membrane during engulfment. GFP PbpX, when expressed in the mother cell, was found uniformly distributed along the cytoplasmic membrane and the outer prespore membrane. The early redistribution of GFP PbpX, and localization to the asymmetric septa, is precluded by the fact that expression from P spoiid occurs only after septation has occurred (see Hilbert & Piggot, 2004). These results confirm that the GFP PBP2c localization to the outer prespore membrane is specific, since in this experiment GFP PBP2c is expressed after closure of the sporulation septum, which separates the inner prespore membrane from the mother cell membrane that will develop into the outer prespore membrane. DISCUSSION The developmental switch from vegetative growth to sporulation in bacteria causes major changes in the subcellular organization of proteins, and gene expression. Work from several groups has shown that proteins involved in asymmetric cell division (Ben-Yehuda & Losick, 2002; Daniel et al., 2000; Scheffers & Errington, 2004), engulfment (Abanes-De Mello et al., 2002; Sharp & Pogliano, 1999), spore coat and cortex assembly (Price & Losick, 1999; Van Ooij et al., 2004), and mother-cell- or prespore-specific gene expression (Rubio & Pogliano, 2004; Rudner et al., 2002) localize to the prespore in a distinct and dynamic fashion. A striking feature of this localization is that, at least for membrane proteins that are targeted to the outer prespore membrane, a shift from a random membrane distribution to a highly organized localization occurs in accordance with a model in which these proteins diffuse 1008 Microbiology 151

11 Localization of PBPs during sporulation laterally through the membrane to their site of action, where they are captured by (an) as yet unknown factor(s) (Rudner et al., 2002). This paper is concerned with the localization of PGsynthesizing proteins during sporulation (summarized in Table 1). A collection of GFP PBP fusion proteins constructed earlier (Scheffers et al., 2004) was used, in addition to new fusions to include PBPs implicated in sporulation. Expression of the GFP PBP fusion proteins is driven by the P xyl promoter, which is switched on during vegetative growth, but switched off during sporulation by removal of xylose from the sporulation medium. This procedure reveals the localization of membrane proteins, and their redistribution upon sporulation, with no newly synthesized GFP PBPs in the prespore. Out of 11 GFP PBP fusions tested, only three showed a change in localization patterns during sporulation, and two of these fusions are to PBPs known to be involved in sporulation. This strongly suggests that the pattern changes are not caused by artefacts. A striking change in protein localization was observed with GFP PbpX, which started by localizing to the division septum at midcell, and then appeared to spiral out in a pattern resembling FtsZ (Ben-Yehuda & Losick, 2002), and was then found at both asymmetric potential division sites. Although GFP PbpX spirals were rare, it should be noted that spiralling seems to be less obvious for membrane proteins than for cytosolic proteins (compare EzrA GFP to FtsZ GFP and FtsA GFP in Ben-Yehuda & Losick, 2002). GFP PbpX then appeared at both asymmetric septa, with unequal distribution of fluorescence intensity, as observed for SpoIIE by Wu et al. (1998). This observation adds credibility to the redistribution observed for GFP PbpX, since other cell-division proteins, notably PBPs (Daniel et al., 2000; Scheffers & Errington, 2004), only localize to the asymmetric septum when one of the potential division sites has been committed to division (see Hilbert & Piggot, 2004). The unequal distribution possibly reflects which asymmetric division site is chosen for septum formation. This question could be resolved by following GFP PbpX distribution in individual sporulating cells with time. Despite this striking fluorescence pattern, pbpx does not play a critical role in B. subtilis. pbpx was inactivated in two ways, but effects on cell growth, cell shape or sporulation efficiency were not detected. Thus, PbpX cannot be an essential component of the cell-division machinery, but it could be associated with (a) component(s) from the division machinery, which it follows from the midcell division site to both asymmetric cell-division sites. A possible role for the endopeptidase PbpX is the quick removal of PG that connects two cells after vegetative division, or thinning of the sporulation septum prior to engulfment. GFP PBP2c and GFP PBP2d, which are randomly distributed along the membrane in vegetative cells (Scheffers et al., 2004), are redistributed during sporulation: both proteins localized to the sporulation septum, followed the engulfing membrane, and were finally concentrated in the prespore membrane. There are two ways in which this redistribution can be achieved (see Fig. 1). First, GFP PBP2c/d in the mother cell is recruited to the septum after septum closure, and then follows the mother cell membrane during engulfment. As a result, the majority of GFP PBP2c/d will be located in the outer prespore membrane (Fig. 1A). Alternatively, GFP PBP2c/d can be recruited to the sporulation septum during septum formation, after which GFP PBP2c/d is found on both sides of the asymmetric septum. Following the membrane during engulfment results in GFP PBP2c/d being distributed throughout both inner and outer prespore membranes (Fig. 1B). The fact that GFP PBP2c localizes to the engulfing membrane when expressed from the mothercell-specific P spoiid, which is switched on only after closure of the sporulation septum, argues in favour of the first model. PBP2c is expressed during both vegetative growth and sporulation, under control of s G (Popham & Setlow, 1993a). The observed redistribution presumably reflects the behaviour of the vegetatively expressed PBP2c during wild-type sporulation. Interestingly, since a pbpf pbpg double mutant has no detectable defects in its spore germ cell wall, which is synthesized from the surface of the inner prespore membrane, but is severely affected in its cortical PG, which is synthesized from the outer prespore membrane (McPherson et al., 2001), it has been suggested that the more important site of PBP2c action is in the outer prespore membrane (Popham, 2002). The observed pattern reflects this mode of action of PBP2c. PBP2d expression is specific to the prespore (Pedersen et al., 2000), but when expressed as a GFP fusion protein during vegetative growth, GFP PBP2d localized in a dispersed fashion along the membrane, similar to various other PBPs described earlier (Scheffers et al., 2004). To follow GFP PBP2d during sporulation, it was necessary to keep xylose present in the sporulation medium. Interestingly, even though GFP PBP2d was not expressed in the compartment in which it is expressed naturally, it did seem to recognize a targeting signal that guides it to the prespore septum and the engulfing membrane. Unfortunately, when expressed from sporulation-specific promoters, either in the prespore or in the mother cell, GFP PBP2d was degraded rapidly, making it impossible to confirm the localization of GFP PBP2d when expressed in the mother cell, or to study the localization of GFP PBP2d in the prespore in detail. PBP2c and PBP2d play redundant roles in sporulation (McPherson et al., 2001), and in agreement with this, do not depend on each other for their localization to the prespore. PBP2c and 2d show localization patterns that are similar to patterns observed for SpoIVFB, a protein that localizes to the prespore outer membrane by diffusion-and-capture (Rudner et al., 2002). We see two possibilities for the diffusion-and-capture of PBP2c and PBP2d. First, it is possible that PBP2c and PBP2d are actively targeted to the

12 D.-J. Scheffers prespore, or captured at the prespore membrane, via an unidentified protein pathway. This active targeting could make sense for PBP2c, which is expressed in both mother cell and prespore, but not for PBP2d, which is expressed in the prespore alone in the wild-type situation. So, if this model were true, under our experimental conditions, GFP PBP2d redistribution should not be observed, unless the protein factor that is recognized is present in the space between inner and outer prespore membranes, and accessible to PBPs present in either membrane. Secondly, PBP2c and PBP2d may be recruited to the prespore by the presence of substrate or substrate analogues. GFP PBP2c and GFP PBP2d follow the engulfing membrane, which contains SpoIID, SpoIIM and SPoIIQ at its leading edge (Abanes-De Mello et al., 2002). SpoIID is a cell wall hydrolase, which is suggested to use the cell wall as a track to drag along the membrane during engulfment (Abanes- De Mello et al., 2002). The hydrolase activity of SpoIID would release PG building blocks that PBP2c and PBP2d could recognize as substrates, and maybe recycle by using them for synthesis of the spore germ wall or cortex, even before engulfment is complete. Recent work in Staphylococcus aureus (Pinho & Errington, 2005) has shown that some high-molecular-mass PBPs depend on the presence of substrate for their correct localization. This has also been suggested for PBP localization in Streptococcus pneumoniae (Morlot et al., 2004). Targeting of PBP2c and PBP2d to the prespore by the availability of substrate is an attractive model, although the question would remain of why there is a difference in substrate binding between PBPs 2c and 2d and the other high-molecular-mass PBPs. ACKNOWLEDGEMENTS I thank Jeff Errington for his support, stimulating discussions and critical reading of the manuscript; Mariana Pinho for valuable comments on the manuscript; other members of the laboratory for helpful discussions and advice; Kit Pogliano and Aileen Rubio (University of California, San Diego) for plasmids pmds13 and pmds14. This work was supported by a Marie Curie Postdoctoral Fellowship (HPMF-CT ), and a grant from the Biotechnology and Biological Sciences Research Council. REFERENCES Abanes-De Mello, A., Sun, Y. L., Aung, S. & Pogliano, K. (2002). A cytoskeleton-like role for the bacterial cell wall during engulfment of the Bacillus subtilis forespore. Genes Dev 16, Anagnostopoulos, C. & Spizizen, J. (1961). Requirements for transformation in Bacillus subtilis. J Bacteriol 81, Barák, I. & Youngman, P. (1996). SpoIIE mutants of Bacillus subtilis comprise two distinct phenotypic classes consistent with a dual functional role for the SpoIIE protein. J Bacteriol 178, Ben-Yehuda, S. & Losick, R. (2002). Asymmetric cell division in B. subtilis involves a spiral-like intermediate of the cytokinetic protein FtsZ. Cell 109, Buchanan, C. E. & Ling, M.-L. (1992). Isolation and sequence analysis of dacb, which encodes a sporulation-specific penicillin-binding protein in Bacillus subtilis. J Bacteriol 174, Cao, M. & Helmann, J. D. (2004). The Bacillus subtilis extracytoplasmic-function s X factor regulates modification of the cell envelope and resistance to cationic antimicrobial peptides. J Bacteriol 186, Cormack, B. P., Valdivia, R. H. & Falkow, S. (1996). FACS-optimized mutants of the green fluorescent protein (GFP). Gene 173, Daniel, R. A., Drake, S., Buchanan, C. E., Scholle, R. & Errington, J. (1994). The Bacillus subtilis spovd gene encodes a mother-cellspecific penicillin-binding protein required for spore morphogenesis. J Mol Biol 235, Daniel, R. A., Williams, A. M. & Errington, J. (1996). A complex fourgene operon containing essential cell division gene pbpb in Bacillus subtilis. J Bacteriol 178, Daniel, R. A., Harry, E. J. & Errington, J. (2000). Role of penicillinbinding protein PBP 2B in assembly and functioning of the division machinery of Bacillus subtilis. Mol Microbiol 35, Eichenberger, P., Jensen, S. T., Conlon, E. M. & 8 other authors (2003). The s E regulon and the identification of additional sporulation genes in Bacillus subtilis. J Mol Biol 327, Errington, J. (1986). A general method for fusion of the Escherichia coli lacz gene to chromosomal genes in Bacillus subtilis. J Gen Microbiol 132, Errington, J. (2003a). Regulation of endospore formation in Bacillus subtilis. Nat Rev Microbiol 1, Errington, J. (2003b). Dynamic proteins and a cytoskeleton in bacteria. Nat Cell Biol 5, Errington, J. & Mandelstam, J. (1983). Variety of sporulation phenotypes resulting from mutations in a single regulatory locus, spoiia, inbacillus subtilis. J Gen Microbiol 129, Errington, J., Daniel, R. A. & Scheffers, D.-J. (2003). Cytokinesis in bacteria. Microbiol Mol Biol Rev 67, Feucht, A., Magnin, T., Yudkin, M. D. & Errington, J. (1996). Bifunctional protein required for asymmetric cell division and cellspecific transcription in Bacillus subtilis. Genes Dev 10, Foster, S. J. & Popham, D. L. (2001). Structure and synthesis of cell wall, spore cortex, teichoic acids, S-layers, and capsules. In Bacillus subtilis and its Closest Relatives: from Genes to Cells, pp Edited by L. Sonenshein, R. Losick & J. A. Hoch. Washington, DC: American Society for Microbiology. Gerdes, K., Moller-Jensen, J., Ebersbach, G., Kruse, T. & Nordstrom, K. (2004). Bacterial mitotic machineries. Cell 116, Glenn, A. R. & Mandelstam, J. (1971). Sporulation in Bacillus subtilis 168. Comparison of alkaline phosphatase from sporulating and vegetative cells. Biochem J 123, Hilbert, D. W. & Piggot, P. J. (2004). Compartmentalization of gene expression during Bacillus subtilis spore formation. Microbiol Mol Biol Rev 68, Huang, X. & Helmann, J. D. (1998). Identification of target promoters for the Bacillus subtilis s X factor using a consensusdirected search. J Mol Biol 279, Huang, X., Decatur, A., Sorokin, A. & Helmann, J. D. (1997). The Bacillus subtilis s X protein is an extracytoplasmic function sigma factor contributing to survival at high temperature. J Bacteriol 179, Jenkinson, H. F. (1983). Altered arrangement of proteins in the spore coat of a germination mutant of Bacillus subtilis. J Gen Microbiol 129, Kobayashi, K., Ehrlich, S. D., Albertini, A. & 96 other authors (2003). Essential Bacillus subtilis genes. Proc Natl Acad Sci U S A 100, Microbiology 151