INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Penicillin-binding proteins (PBPs), which catalyze the polymerization and cross-linking of bacterial peptidoglycan, can be divided into three classes based on their amino acid sequence: the low-molecular-weight PBPs and the high-molecular-weight (HMW) class A and class B PBPs (8). The HMW class B PBPs are monofunctional transpeptidases, some of which have essential functions in septation and maintenance of cell shape (8), while the HMW class A PBPs have both transglycosylase and transpeptidase activities (14, 42) and appear to be somewhat functionally redundant (17, 34). The Bacillus subtilis ponA gene, coding for the HMW class A PBP1, is transcribed predominantly during log-phase growth (34). Previous work with B. subtilis mutants lacking one or several of the three known HMW class A PBPs (PBP1, PBP2c, and PBP4) showed that (i) lack of PBP1 results in slower growth, increased cell length, and decreased cell diameter and (ii) PBP1 is functionally more important than PBP2c and PBP4 (34). It was also recently demonstrated that PBP1 localizes to cell division sites and plays an important role in the formation of the peptidoglycan division septum in vegetative cells of B. subtilis (30).

The ponA gene is part of a two-gene operon that also includes prfA (PBP-related factor A [note that in the B. subtilis genome database, prfA is called recU]), which is located immediately upstream of and cotranscribed with ponA (33). prfA codes for a putative 206 residue, basic protein (pI of 10.1), which has no significant sequence homology to proteins with known functions. However, DNA sequencing has indicated that genes encoding similar proteins are present in a large number of gram-positive bacteria: a BLAST search (2) of prfA against completed and unfinished microbial genomes (preliminary sequence data were obtained from The Institute of Genomic Research Website at http://www.tigr.org) produced sequences with significant homology from 10 different gram-positive organisms, while no prfA homologs were detected in gram-negative bacteria by this analysis. The former organisms include Enterococcus faecalis (54% identity in a 192-amino-acid overlap), Staphylococcus aureus (58% identity in a 167-amino-acid overlap), Streptococcus pyogenes (51% identity in a 195-amino-acid overlap), S. pneumoniae (49% identity in a 124-amino-acid overlap), Deinococcus radiodurans (49% identity in a 124-amino-acid overlap), and Mycoplasma genitalium (32% identity in a 187-amino-acid overlap). For at least two of these species, namely, S. pneumoniae and S. aureus, the prfA genes are also located upstream of ponA homologs (25, 32). Although the prfA gene product has not yet been identified in B. subtilis, it is known that inactivation of prfA results in cells that grow ~50% more slowly than do wild-type cells and vary significantly in cell length. This phenotype is exacerbated greatly by the additional loss of PBP1 but not by the loss of either PBP2c or PBP4 (34). Finally, it has been shown that a mutation in prfA (recU::cat) renders cells more sensitive to DNA-damaging agents and decreases the efficiency of transformation, suggesting a possible role for prfA in DNA repair and homologous recombination (7). Given that PBP1 plays an important role in cell division in B. subtilis (30) and that a prfA mutation has a clear phenotype (34), we have investigated the function of prfA in detail. Our results indicate that prfA is required for proper chromosome segregation in B. subtilis.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmids and bacterial strains. The plasmids and bacterial strains used in this study are listed in Tables 1 and 2, respectively. Note that strains PS2061 (prfA::spc) and PS2123 (prfA) have deletions removing ~60 and ~42% of the prfA coding region, respectively, and are therefore almost certainly prfA null mutants (33; D. L. Popham and P. Setlow, unpublished results).

Growth of B. subtilis. B. subtilis strains were grown overnight at 30°C on 2× SG (18) or Luria-Bertani (LB) (10 g of tryptone per liter, 5 g of yeast extract per liter, 10 g of NaCl per liter, 1 mM NaOH) agar plates with or without appropriate antibiotics, inoculated into 2× YT medium (16 g of tryptone per liter, 10 g of yeast extract per liter, 5 g of NaCl per liter) or 1× Penassay broth (PAB) (Difco), and grown at 30 or 37°C. In some cases, 1% (wt/vol) xylose was included in the media. Growth rates reported are those for cells in log-phase growth.

PCR and cloning procedures. To express prfA in B. subtilis from a xylose-inducible promoter, the putative ribosome-binding site and coding region of prfA (from bp 651 to 1300) (33) was amplified by PCR using primers prfA-SphI (5'-GCATGCGTCATGATTAGTTTAATAAGG-3' [underlined nucleotides denote an SphI site]) and prfA-B/S (5'-GGATCCTCAACTAGTACCTTTCGCACCAGATGATGG-3' [underlined nucleotides denote BamHI and SpeI sites; note that the SpeI site results in the addition of two extra amino acid residues at the C terminus of PrfA]) and chromosomal DNA from wild-type B. subtilis (strain PS832) as a template. The PCR product (670 bp) was ligated into pCR 2.1 to generate plasmid pLP78, and the insert was sequenced, removed by digestion with BamHI and SphI, and ligated into pRDC19 digested with the same enzymes to generate plasmid pLP79, which was used to transform B. subtilis (Table 2).

Growth and induction of recombinant E. coli, sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), and N-terminal protein sequencing. Recombinant E. coli was grown at 37°C in 20 or 50 ml of 2× YT medium containing chloramphenicol (20 µg/ml) and kanamycin or ampicillin (both used at 50 µg/ml) to an optical density at 600 nm (OD₆₀₀) of ~0.5, protein expression induced with 1 mM isopropyl--D-thiogalactopyranoside (IPTG), and cultures were incubated at 37°C. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis and N-terminal protein sequencing were carried out essentially as described previously (29, 31).

Fluorescence microscopy. To stain cell walls and nucleic acids of B. subtilis cells, 0.5 ml of a culture grown in 2× YT or PAB medium was fixed for 20 min at room temperature with 4.4% (wt/vol) paraformaldehyde-0.017% glutaraldehyde-28 mM sodium phosphate (pH 7), washed twice with phosphate-buffered saline (PBS) (8 g of NaCl per liter, 0.2 g of KCl per liter, 1.44 g of Na₂HPO₄ per liter, 0.24 g of KH₂PO₄ per liter [pH 7.4]), and resuspended in 100 µl of GTE (50 mM glucose, 20 mM Tris-HCl [pH 7.5], 10 mM EDTA). Lysozyme was added to 2 mg/ml, and the cells were immediately applied to poly-L-lysine-coated microscope slides. After 30 s, excess fluid was removed and bound cells were washed twice with PBS and allowed to dry completely. Following rehydration with PBS, the slides were blocked for 20 min at room temperature with 2% (wt/vol) bovine serum albumin in PBS and incubated for 1 h at room temperature in PBS containing 0.1% (wt/vol) bovine serum albumin, 2 µg of 4',6'-diamino-2-phenylindole (DAPI; Sigma) per ml, and 2 µg of Oregon Green-conjugated wheat germ agglutinin (WGA; Molecular Probes) per ml or 10 µg of propidium iodide (PI) per ml. Finally, the cells were washed six times with PBS and the slides were mounted using the SlowFade antifade kit from Molecular Probes. They were visualized with a Zeiss Axiovert 100 fluorescence microscope equipped with a Plan-APOCHROMAT 100x or a Plan-NEOFLUAR 63x oil immersion lens (Zeiss) and a standard filter block for visualizing Oregon Green, DAPI, and PI. Images were captured with a cooled charged-coupled device camera using exposure times of 1 to 2 s for Oregon Green, 2 s for DAPI, and 0.5 to 2 sec for PI (due to variable staining efficiency, it was necessary to use different exposure times for different strains).

Electron microscopy. Log-phase B. subtilis cells were fixed, processed, and analyzed by electron microscopy as described previously (37).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Chromosome segregation defect in prfA deletion mutants. Previously it was shown that cells of a prfA deletion mutant (prfA::spc; strain PS2061) produce normal levels of PBP1 but grow more slowly and vary significantly in cell length compared to wild-type cells (33, 34). An identical phenotype was observed for strain PS2123, which contains an in-frame deletion of prfA (33). To analyze these prfA strains in more detail, log-phase cells grown at 30 or 37°C in PAB or 2× YT medium were fixed, stained with DAPI and PI or Oregon Green-conjugated WGA, and subjected to fluorescence microscopy to visualize DNA, septa, and/or cell walls. This analysis revealed that in addition to exhibiting an abnormal division pattern, cultures of both prfA strains contained significant numbers of anucleate cells, ranging from 0.9 to 3% of total cells depending on the precise conditions of the experiment (Table 3). An example of an anucleate cell is shown in Fig. 1B (arrow iv; the cell visible in the right-hand panel does not stain with DAPI) (see also Fig. 4B, arrows). Under similar conditions, cultures of wild-type cells (strain PS832) contained less than 0.1% anucleate cells (Table 3). Cultures of both prfA strains also contained a large proportion of cells (~34%) with abnormal nucleoid-staining patterns (Fig. 1A and B). The abnormal nucleoid-staining patterns observed include nucleoids that are asymmetrically positioned in the cell (Fig. 1A, arrow i; the arrow points to a large region of a cell lacking chromosomal DNA), nucleoids that appear bisected by or impinge upon the septum (Fig. 1A, arrow ii; note also the cell immediately above arrow ii), and large aggregates of nucleoids occupying an extensive part of the cell (Fig. 1B, arrow iii). Such abnormal nucleoid-staining patterns were essentially absent (i.e., were present at <1%) in wild-type cells analyzed in parallel (Fig. 1C) (see also Fig. 4A), indicating that these abnormal staining patterns were not a fixation artifact. Some of the abnormalities in the appearance of nucleoids in prfA cells were even more apparent upon electron microscopy (Fig. 2a to d). For example, nucleoids that were bisected by the septum were clearly visible in some dividing prfA cells (Fig. 2a and d, arrows), while other cells had aggregates of nucleoids that appeared stretched out (Fig. 2b and c). When wild-type cells were analyzed in parallel, no such defects were observed (Fig. 2e). These results therefore suggest that the prfA mutants are defective in chromosome segregation.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Abnormal nucleoid staining of prfA::spc (strain PS2061) mutant cells. Log-phase cells (OD600  0.5) of a prfA::spc mutant (A and B) or a wild-type strain (C) grown at 37°C in 2× YT medium were fixed, stained with DAPI and Oregon Green-conjugated WGA (lectin), and analyzed by fluorescence microscopy to visualize nucleoids (pseudocolored red) and septa and cell walls (green). Arrows: (i) a cell with asymmetrically positioned nucleoids; (ii) a nucleoid that impinges upon the septum; (iii) a cell containing extended, aggregated nucleoids; (iv) an anucleate cell. Bar, 10 µm.

View larger version (138K): [in this window] [in a new window]   FIG. 2.   Electron micrographs of cross-sectioned prfA (a to d) and wild-type (e) cells grown to log phase (OD600  0.5) at 37°C in 2× YT medium. The prfA strain used was PS2061 (prfA::spc). Bars, 0.6 µm. Nucleoids (n) appear as light, fibrillar material in the cytoplasm. Arrows show nucleoids bisected by septa (a and d) and an abnormal cell wall cluster (c).

Membrane and wall defects of a prfA deletion mutant. In our electron microscopy analysis, we also occasionally observed prfA mutant cells with abnormal morphology, like the tortuous cell shown in Fig. 2d, and we found that ~16% (n = 32) of the cells had abnormal clusters or defects in the cell wall (Fig. 2c, arrow) resembling those previously observed in cells lacking PBP1 (30). In some cross-sections of prfA cells (~19%; n = 32) abnormal membrane structures or swirls were observed across the diameter of the cell (Fig. 3); these membrane swirls seemed to be formed by inward growth of the membrane at potential division sites, because wall ingrowths could sometimes be seen at similar positions (Fig. 3a and c). Presumably, septum formation was initiated at these sites but failed to go to completion due to uncoupling of wall and membrane ingrowth. Collectively, these morphological defects could reflect an additional role for PrfA in membrane and/or cell wall synthesis, or they could be secondary effects due to defective chromosome segregation.

View larger version (97K): [in this window] [in a new window]   FIG. 3.   Electron micrographs of cross-sectioned prfA (strain PS2061) cells showing abnormal membrane ingrowths or swirls across the diameter of the cell. Note the presence of wall ingrowths in the periphery of the cells shown in panels a and c. The cells were grown to an OD600 of ca. 0.5 at 37°C in 2× YT medium. Bar, 0.25 µm.

Complementation of the prfA::spc mutation. To rule out the possibility that some or all of the effects of the prfA::spc mutation were due to polar effects on the expression of the downstream ponA gene, we placed a copy of prfA fused to a xylose-inducible promoter (Pxyl) at the thrC locus of the prfA::spc mutant strain to generate strain LP85. As a control, prfA::spc cells transformed with vector alone were used (strain LP88). When these cells were grown in 2× YT or PAB medium supplemented with 1% xylose, the growth, morphology, and chromosome segregation defects associated with the prfA::spc mutation were fully complemented in strain LP85 but not in strain LP88, suggesting that these defects were indeed caused by inactivation of prfA and not by a polar effect on the downstream ponA gene (Table 3 and data not shown). This is consistent with previous findings that the level of PBP1 is normal in strain PS2061 (33). In the absence of xylose, LP85 cells grew like prfA cells and had a similar but slightly less severe nucleoid segregation defect compared to prfA cells (Table 3 and data not shown). We therefore suspect that some PrfA protein (but less than the wild-type level) is present in strain LP85 in the absence of xylose.

A conditional prfA smc double mutant. The smc gene (for "structural maintenance of chromosomes") is required for chromosome structure and partitioning in B. subtilis and inactivation of smc produces cells with chromosome segregation defects similar to but more severe than those observed for prfA cells. For example, smc null mutant cells grown at 23 or 30°C contain about 10 to 15% anucleate cells compared to ~0.9 to 3% for prfA cells (Table 3) (6, 26). We attempted to generate a prfA smc double mutant by transforming PS2061 (prfA::spc) cells with chromosomal DNA from strain RB35 (smc::kan) and selecting transformants on LB plates supplemented with kanamycin at 30°C (strain RB35 is temperature sensitive for growth in rich medium [6]). Although we obtained a number of very small colonies after 2 days of incubation, detailed analysis of several of these transformants suggested that they most probably contained suppressor mutations (results not shown). To avoid the problem of potential suppressor mutations, we therefore generated a conditional prfA smc mutant by transforming strain LP85 (prfA::spc thrC::[Pxyl-prfA xylR erm]) cells with chromosomal DNA from strain RB35 (smc::kan) and selecting transformants at 30°C on LB-kanamycin plates with or without 1% xylose. Interestingly, ~30-fold more colonies were produced when the transformants were selected on xylose-containing plates than on xylose-free plates (744 and 25 colonies on average, respectively, in two independent experiments), indicating that the combined effects of the prfA and smc mutations are detrimental to the cell. One of the transformants (strain LP101; prfA::spc thrC::[Pxyl-prfA xylR erm] smc::kan) selected on a xylose-containing plate was chosen for further analysis. Analysis of LP101 cells grown at 30°C in PAB medium revealed no significant difference in the growth rate of cells grown with or without 1% xylose, which in both cases was similar to that of smc (LP99) cells (Table 3). However, LP101 cultures grown without xylose contained significantly more anucleate cells (~24%) than did cultures grown in the presence of xylose (~15% [Table 3]), suggesting that the lack of PrfA exacerbates the smc phenotype. Cultures of smc (LP99) cells grown under similar conditions contained ~13% anucleate cells (Table 3), which is consistent with previous reports (~10 to 15% anucleate cells [6, 26]). We also examined the LP101 cells grown to stationary phase in PAB medium with or without xylose. This analysis revealed the presence of extremely long chains of LP101 cells in the culture grown without xylose (Fig. 4F) while stationary-phase LP101 cells grown in the presence of 1% xylose looked like smc cells (Fig. 4E and data not shown).

View larger version (33K): [in this window] [in a new window]   FIG. 4.   Fluorescence micrographs of prfA, smc, and spoOJ strains grown in PAB medium at 30°C without (A, B, C, and F) or with (E) 1% xylose or in 2× YT at 37°C (D). The cells were fixed in log phase (A to D) or stationary phase (E and F) and stained with PI (Ai to Fi) to visualize the cytoplasm (red) and septa (dark lines between cells) and DAPI (Aii to Fii) to visualize nucleoids (pseudocolored green). (A) PS832 (wild type); (B) PS2061 (prfA::spc); (C) LP99 (smc::kan); (D) LP105 (prfA spoOJ::spc); (E and F) LP101 (prfA::spc thrC::[Pxyl-prfA xylR erm] smc::kan). Cells of strain PS2123 (prfA) appear identical to those of strain PS2061 (B); cells of strain LP102 (spoOJ::spc) appear similar to the wild-type cells (A) Arrows indicate anucleate cells. Bar, 10 µm.

Phenotype of a prfA spoOJ mutant. Another gene involved in DNA segregation in B. subtilis is spoOJ. SpoOJ is similar to the ParB family of plasmid-encoded partition proteins and is required for proper chromosome segregation during vegetative growth (13) and for correct chromosome positioning during sporulation (38). SpoOJ colocalizes with the origin region of the chromosome (9, 20, 22) and binds to specific sites located in the origin-proximal ~20% of the chromosome (21). Furthermore, the presence of one of these latter sites, called parS, on an otherwise unstable plasmid stabilizes the plasmid in a SpoOJ-dependent manner, indicating that parS acts as a partitioning site (21).

prfA spoOJ

Overexpression of prfA causes nucleoid condensation in E. coli. We next tried to overexpress prfA in B. subtilis by generating strain LP89 (thrC::[Pxyl-prfA xylR erm]) and growing the cells in 2× YT medium supplemented with 1% xylose. While LP89 cells grew normally under these conditions and were indistinguishable from wild-type cells, by SDS-PAGE analysis we could not detect any differences in protein profiles between xylose-induced LP89 and prfA::spc cells (data not shown), suggesting that PrfA is not very abundant even in induced LP89 cells. We therefore cloned prfA in the expression vector pET9d and introduced it into E. coli BL21(DE3)/pLysS cells (41) to generate strain LP77. Induction of cells of strain LP77 with IPTG led to significant overproduction of a ~24.5-kDa protein, whose identity as PrfA was confirmed by N-terminal sequencing (data not shown). The overexpressed PrfA was found largely in inclusion bodies, although a significant amount appeared to be cytoplasmic and the protein was toxic for E. coli (data not shown).

View larger version (112K): [in this window] [in a new window]   FIG. 5.   Nucleoid condensation in E. coli expressing prfA. Cells were grown in 2× YT medium with appropriate antibiotics at 37°C and induced with IPTG for 0, 30, or 60 min, and the nucleoids were stained with DAPI and examined by fluorescence microscopy. (A) cells of strain PS2602 (pET11a); (B) cells of strain LP77 (pET9d-prfA); (C) mixture of PS2602 and LP77 cells; (D) cells of strain PS2599 (pET11a-dacC). Bar, 10 µm. Arrows show condensed nucleoids in cells expressing prfA.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Following DNA replication, bacterial cells segregate daughter chromosomes into daughter cells with high fidelity, resulting in the presence of fewer than 0.03% anucleate cells in growing cultures of E. coli or B. subtilis (11, 13). The events involved in chromosome segregation and partitioning include resolution of chromosome dimers resulting from recombinational crossovers between sister chromosomes, decatenation of interlinked daughter chromosomes, and movement of daughter chromosomes away from each other (43). Recent evidence suggests that bacterial chromosome segregation is an active mitosis-like process and that the origin of replication (oriC) of the E. coli and B. subtilis chromosome has a specific orientation during the cell cycle (9, 10, 20, 44). Thus, in newborn cells oriC is oriented toward a cell pole; after replication of this region, one of the two origins moves rapidly toward the opposite pole of the cell while the termination region remains centrally located (10, 44).

A number of genes thought to be involved in the orientation and separation of daughter chromosomes have been characterized in E. coli and B. subtilis (reviewed in reference 43). These include E. coli xerC and xerD and their B. subtilis homologs ripX and codV, which code for recombinases involved in site-specific recombination leading to resolution of chromosome dimers (4, 5, 36). In addition, six so-called par genes involved in decatenation of interlinked daughter chromosomes have been characterized, as well as a number of genes thought to be involved in chromosome movement and partitioning (reviewed in reference 43). Among the latter group are the muk, minD, and ftsK genes of E. coli (12, 23, 49) and the smc, spoOJ, and spoIIIE genes of B. subtilis (6, 13, 26, 45). Finally, recent work by Lemon and Grossman (19) indicates that the process of DNA replication itself may contribute to the movement and separation of daughter chromosomes.

prfA is required for proper chromosome segregation. In the present work we have shown for the first time that prfA is required for proper chromosome segregation in B. subtilis. Thus, cultures of prfA cells grown at 30 or 37°C in rich medium were found to contain ~0.9 to 3% anucleate cells and a significant proportion (~34%) of cells with abnormal nucleoid staining patterns (Fig. 1 and 4B and Table 3). By analysis of a conditional prfA smc mutant (strain LP101) and a prfA spoOJ mutant (strain LP105), we found that inactivation of prfA also exacerbates the smc and spoOJ chromosome segregation phenotypes. This suggests that prfA affects chromosome segregation via a pathway different from those used by smc and spoOJ.

How does PrfA affect chromosome segregation? An obvious question arising from this work is what exact role PrfA plays in chromosome segregation. A possible DNA-binding activity of PrfA is supported by the findings that PrfA is very basic (pI of 10.1) and that overexpression of prfA causes nucleoid condensation in E. coli (Fig. 5), although the latter finding could be a result of the basic nature of PrfA. If PrfA binds directly to DNA, the chromosome segregation defect seen in prfA cells could be due to impaired DNA replication, dimer resolution, decatenation, or chromosome movement and positioning. An effect of PrfA on DNA replication seems unlikely because the protein-to-DNA ratio in prfA cells is not significantly different from that of wild-type cells (L. B. Pedersen and P. Setlow, unpublished observations). Furthermore, in addition to anucleate cells, cells with increased DNA content were observed in prfA cultures (Fig. 1). Despite the ability of PrfA to induce DNA condensation when overproduced in E. coli, a general nonspecific DNA-condensing activity of PrfA in vivo also seems unlikely, given the presumed low abundance of this protein: PrfA has so far eluded detection by SDS-PAGE of extracts from wild-type B. subtilis cells, and PBP1, which is coexpressed with PrfA (33), is present in only 450 to 1,000 copies per cell (30).

Localization of PrfA? The C-terminal domain of FtsK is homologous to the SpoIIIE protein of B. subtilis. SpoIIIE is required for postseptational chromosome translocation in sporulating cells (45, 47) and in vegetative cells grown under conditions when normal nucleoid separation or septum positioning is perturbed (39). SpoIIIE localizes near the middle of the asymmetric division septum during sporulation and has been suggested to form a seal between the DNA and the leading edge of the division septum (46). Given that PBP1 localizes to the division septum and plays a role in its formation (30), it is tempting to speculate that PrfA is associated with the division septum and interacts with the chromosome in a manner similar to that proposed for FtsK and SpoIIIE. Determination of the cellular localization of PrfA would clearly help in clarifying these issues. Extensive attempts to localize PrfA by use of a PrfA-green fluorescent protein fusion have so far been unsuccessful (Pedersen and Setlow, unpublished), but efforts are under way to produce an antiserum against PrfA (D. L. Popham, personal communication).

Possible indirect effects of PrfA on chromosome segregation. Finally, it is possible that PrfA affects chromosome segregation indirectly by affecting septum formation, by recruiting other proteins to certain sites, or by acting as a chaperone for proteins involved in DNA-wall-membrane interactions. Although prfA cells vary greatly in length, suggesting a defect in septum placement (34), indirect evidence suggests that this phenotype may be a secondary effect of impaired chromosome segregation rather than vice versa. For example, studies with E. coli parC and mukB mutants have indicated that FtsZ ring formation and hence septation can be prevented to some extent by abnormally large or aberrant nucleoids (24, 48). Furthermore, mutations in minD, which is known to be involved in septum positioning (35), have no overt effect on chromosome segregation in B. subtilis (39), although such an effect has been reported for E. coli (1, 16, 27). Therefore we find it plausible that the impaired septum placement phenotype of prfA mutant cells (34) could be due to the presence of abnormally large or aberrant nucleoids in these cells.