RefZ and Noc Act Synthetically to Prevent Aberrant Divisions during Bacillus subtilis Sporulation

ABSTRACT

During sporulation, Bacillus subtilis undergoes an atypical cell division that requires overriding mechanisms that protect chromosomes from damage and ensure inheritance by daughter cells. Instead of assembling between segregated chromosomes at midcell, the FtsZ-ring coalesces polarly, directing division over one chromosome. The DNA-binding protein RefZ facilitates the timely assembly of polar Z-rings and partially defines the region of chromosome initially captured in the forespore. RefZ binds to motifs (RBMs) located proximal to the origin of replication (oriC). Although refZ and the RBMs are conserved across the Bacillus genus, a refZ deletion mutant sporulates with wild-type efficiency, so the functional significance of RefZ during sporulation remains unclear. To further investigate RefZ function, we performed a candidate-based screen for synthetic sporulation defects by combining ΔrefZ with deletions of genes previously implicated in FtsZ regulation and/or chromosome capture. Combining ΔrefZ with deletions of ezrA, sepF, parA, or minD did not detectably affect sporulation. In contrast, a ΔrefZ Δnoc mutant exhibited a sporulation defect, revealing a genetic interaction between RefZ and Noc. Using reporters of sporulation progression, we determined the ΔrefZ Δnoc mutant exhibited sporulation delays after Spo0A activation but prior to late sporulation, with a subset of cells failing to divide polarly or activate the first forespore-specific sigma factor, SigF. The ΔrefZ Δnoc mutant also exhibited extensive dysregulation of cell division, producing cells with extra, misplaced, or otherwise aberrant septa. Our results reveal a previously unknown epistatic relationship that suggests refZ and noc contribute synthetically to regulating cell division and supporting spore development.

IMPORTANCE The DNA-binding protein RefZ and its binding sites (RBMs) are conserved in sequence and location on the chromosome across the Bacillus genus and contribute to the timing of polar FtsZ-ring assembly during sporulation. Only a small number of noncoding and nonregulatory DNA motifs are known to be conserved in chromosomal position in bacteria, suggesting there is strong selective pressure for their maintenance; however, a refZ deletion mutant sporulates efficiently, providing no clues as to their functional significance. Here, we find that in the absence of the nucleoid occlusion factor Noc, deletion of refZ results in a sporulation defect characterized by developmental delays and aberrant divisions.

INTRODUCTION

Chromosome inheritance depends on precise division site selection. Abnormal divisions can result in aneuploidy, including total chromosome loss. Eukaryotes employ cell cycle checkpoints to ensure replication and segregation are complete before cytokinesis initiates. In contrast, bacteria often segregate DNA concurrently with division, so mechanisms to coordinate these processes are critical. To ensure faithful transmission of genetic material to progeny, bacteria segregate replicated DNA to opposite cell halves and divide between chromosome masses (nucleoids). Division at midcell is initiated by polymerization and bundling of membrane-tethered FtsZ protofilaments into the “Z-ring,” a dynamic structure subject to both positive and negative regulation (1–20). The Z-ring facilitates recruitment of additional factors needed for division at midcell, including peptidoglycan remodeling enzymes (2, 14–20).

The Min system and Nucleoid Occlusion (NO) are redundant, but mechanistically distinct systems for ensuring Z-rings assemble at midcell, between chromosomes (21, 22). In B. subtilis, the MinCD complex localizes in the immediate vicinity of the nascent septum and inhibits additional Z-rings from forming (23–26); following division, Min inhibition persists at the newly formed, nucleoid-free poles (27, 28). NO, by contrast, prevents assembly of division-competent Z-rings over the bulk of the nucleoid (29, 30). In E. coli, NO is mediated by an inhibitor of FtsZ polymerization, SlmA. SlmA is also a DNA-binding protein, with specificity for motifs (SBSs) enriched throughout the chromosome except in the terminus (ter) region (30, 31). Following chromosome replication and segregation, ter is localized at midcell. The coincident segregation of SlmA away from midcell leads to release of NO, a condition more favorable to, but not sufficient for, Z-ring assembly. The NO protein of B. subtilis, Noc, also binds to motifs enriched distal to ter (32); however, unlike SlmA, Noc has not been shown to interact with FtsZ directly. Instead, Noc is hypothesized to block Z-ring nucleation sites by tethering the chromosome to the membrane (33). More recent, high resolution microscopy experiments suggest that rather than inhibiting Z-ring formation over the nucleoid, Noc promotes shifting of nonmedial Z-ring intermediates toward midcell in a process described as “corralling” (34). Notably, neither Min nor NO are required for midcell localization of FtsZ, though each contributes to the efficiency of medial division (21).

Division in bacteria is not always medial nor occluded by the presence of nucleoid. For example, the earliest stage of B. subtilis sporulation is characterized by an asymmetric septation over one chromosome, resulting in two cell compartments with transiently different genetic complements (35). The larger “mother” cell contains a complete copy of the chromosome, while the smaller future spore (forespore) initially captures only a segment of the origin-proximal region of a second chromosome (36–38); bisection of the forespore-destined chromosome is avoided because the DNA is threaded through the FtsK-like ATPase, SpoIIIE, which pumps the remainder of the chromosome into the forespore (39–41). The polar division of sporulation is of considerable interest because it requires bypass of the Min and NO systems active during vegetative growth (40, 42, 43). The redistribution of FtsZ from midcell to the pole is known to be facilitated by enhanced transcription of ftsAZ from a SigH-dependent promoter (P2) (44) and by expression of the FtsZ-associated protein SpoIIE (45). Remarkably, increasing ftsAZ copy number and artificially inducing spoIIE is sufficient to shift the Z-rings of nonsporulating cells toward the poles (46).

RefZ (Regulator of FtsZ) is a DNA-binding protein expressed by Spo0A-P during the early stages of sporulation (47). refZ is also repressed by the glucose repressor CcpA (48), activated under conditions of phosphate limitation (49), and is upregulated in stationary phase by SigH (50). During sporulation, RefZ binds to DNA motifs (RBMs) located near oriC (RBM_O) and on the left and right chromosomal arms (RBM_L1, RBM_L2, RBM_R1, and RBM_R2) (51, 52). The left and right arm RBMs fall at the boundary demarcating the segment of chromosome localized in the forespore at the time of polar septation (51). Cells lacking refZ or the RBMs are more likely to capture DNA regions generally excluded from the forespore at the time of polar division, indicating that RefZ-RBM complexes somehow influence the position of the septum relative to the chromosome (51). During sporulation, a ΔrefZ mutant also exhibits, on average, a delay in polar Z-ring formation (52) and a small shift of the septum toward midcell (53). These results suggest the ΔrefZ mutant’s “overcapture” phenotype may be at least partially attributable to changes in chromosome organization or increased forespore dimensions that arise due to the delay.

There is likely a strong selective advantage to maintaining RBM positioning on the chromosome, as the arrangement of the RBMs on the left and right chromosome arms is remarkably conserved across the entire Bacillus genus (51); however, deleting refZ or introducing mutations into the RBMs that prevent RefZ binding does not reduce sporulation efficiency under standard laboratory growth conditions (51, 54). The mechanism underlying RefZ’s observed effects on FtsZ, and the selective advantage conferred to Bacillus by maintaining refZ and the RBMs are not known.

In this work, we performed a candidate-based screen to determine if deleting refZ in combination with other genes previously implicated in FtsZ regulation and/or chromosome capture would result in synthetic sporulation phenotypes. We found that combining ΔrefZ with deletions of ezrA, sepF, parA, or minD did not detectably reduce sporulation efficiency. In contrast, reduced sporulation of a ΔrefZ Δnoc mutant was evident in a plate-based assay. Using reporters of sporulation progression, we determined that the sporulation defect occurred after Spo0A activation, but prior to late sporulation. A subpopulation of ΔrefZ Δnoc mutant cells failed to divide polarly or, following division, to activate the first forespore sigma factor. In addition, the ΔrefZ Δnoc mutant exhibited aberrant divisions indicative of dysregulated Z-ring assembly. Our results reveal an epistatic relationship between RefZ and Noc consistent with the two proteins contributing synthetically to cell division regulation and spore development.

RESULTS

Deletion of refZ and noc results in a synthetic sporulation defect.

RefZ promotes shifting of medial Z-rings toward the pole and the precise capture of chromosome in the forespore, yet deletion of refZ does not affect sporulation efficiency (51, 52), suggesting RefZ function is redundant with other factors affecting cell division and/or chromosome capture. To look for synthetic sporulation defects, we integrated a reporter of late-stage sporulation, P_cotD-lacZ, into the chromosome at an ectopic locus. In this background, cells reaching late-stage sporulation express beta-galactosidase, resulting in accumulation of blue pigment on sporulation medium containing X-gal (Fig. 1). When spotted at equivalent densities, ΔrefZ cells turned blue at a rate indistinguishable from wild type (Fig. 1), consistent with prior results demonstrating that the ΔrefZ mutant does not exhibit a sporulation defect (51, 52). A mutant harboring point mutations that abrogate RefZ binding at the five origin-proximal RBMs (51), RBM_5mu, also sporulated indistinguishably from wild type in the plate-based assay (Fig. 1).

FIG 1.

Plate-based sporulation assay based on lacZ expression from a late-stage sporulation promoter, P_cotD. wt (BAM1323), Δnoc (BAM1321), ΔezrA (BAM1563), ΔsepF (BAM1548), ΔparA (BAM1549), ΔminD (BAM1575), ΔrefZ (BAM1550), ΔrefZ Δnoc (BAM1546), ΔrefZ ΔezrA (BAM1559), ΔrefZ ΔsepF (BAM1577), ΔrefZ ΔparA (BAM1568), ΔrefZ ΔminD (BAM1578), RBM_5mu (BAM1573), RBM_5mu Δnoc (BAM1562), RBM_5mu ΔezrA (BAM1564), RBM_5mu ΔsepF (BAM1547), RBM_5mu ΔparA (BAM1569), RBM_5mu ΔminD (BAM1576).

Next, we assessed expression of P_cotD-lacZ in strains harboring deletions of genes implicated in regulating FtsZ dynamics (noc, ezrA, and sepF) (7, 32, 43, 55, 56), chromosome capture (soj) (56–58), or both (minD) (55, 56). Again, each of the single deletion strains progressed in sporulation comparably to wild type (Fig. 1). Similar results were obtained when each deletion was introduced into a ΔrefZ or RBM_5mu mutant background, with two exceptions: the ΔrefZ Δnoc and RBM_5mu Δnoc mutants did not turn blue during the experimental time course (Fig. 1), consistent with a defect or halt in sporulation progression.

To determine if the mutants produced viable spores, we grew cells in a sporulation medium and determined the number of CFU before and after a heat treatment that kills vegetative cells. The single mutants produced spores at levels within 2-fold of wild-type, consistent with either no or a minor sporulation defect (Table 1). In contrast, the double mutants produced 5–10-fold fewer spores. Notably, the double mutants also exhibited around a 2-fold drop in CFU before heat treatment, suggesting that a subset of cells committed to sporulation (59), but were unable to complete the sporulation process. We conclude that RefZ and the RBMs contribute synthetically with Noc to support wild-type sporulation.

TABLE 1.

Sporulation efficiencies of wild type and mutants^a

Genotype	CFU/mL	Spores/mL	% spores/CFU	% spores/wt spores
wt	2.3 × 108 (± 5.7 × 107)	2.7 × 108 (± 5.4 × 107)	124 (± 26)	100
ΔrefZ	2.2 × 108 (± 3.5 × 107)	2.1 × 108 (± 2.9 × 107)	90 (± 25)	77 (± 12)
RBM5mu	3.4 × 108 (± 3.8 × 107)	1.9 × 108 (± 3.9 × 107)	57 (± 19)	70 (± 14)
Δnoc	2.1 × 108 (± 2.8 × 107)	1.4 × 108 (± 2.2 × 107)	69 (± 11)	55 (± 12)
ΔrefZ Δnoc	7.0 × 107 (± 5.4 × 106)	5.2 × 107 (± 4.7 × 105)	74 (± 6)	20 (± 4)
RBM5mu Δnoc	9.1 × 107 (± 8.5 × 106)	3.3 × 107 (± 3.8 × 106)	36 (± 1)	12 (± 2)

^aValues shown are the average of three experimental and biological replications, with standard deviations in parentheses.

ΔrefZ Δnoc and RBM_5mu Δnoc mutants initiate sporulation.

The sporulation delay observed in the ΔrefZ Δnoc and RBM_5mu Δnoc double mutants could be due to failed or reduced entry into the sporulation program or to a delay or halt at any stage of sporulation prior to cotD expression. To investigate further, we introduced a fluorescent reporter (P_spoIIG-CFP) activated during the earliest stage of sporulation by the sporulation master regulator, Spo0A-P (48, 60). Using P_spoIIG-CFP, we were able to identify cells that had initiated sporulation, including cells without polar septa. Two and a half hours after resuspension in sporulation medium, cells were collected, and CFP expression was monitored by epifluorescence microscopy. The single and double mutants were qualitatively indistinguishable from wild type with respect to the number of CFP-expressing cells (Fig. 2 and Fig. S1), suggesting that the sporulation defect in ΔrefZ Δnoc and RBM_5mu Δnoc mutant cells occurs after sporulation is initiated; however, we do not exclude the possibility that a subset of cells may also fail to initiate.

FIG 2.

Expression of CFP from Spo0A-dependent promoter P_spoIIG. (A) Images were captured 2.5 h following sporulation by resuspension. wt (BAM909), ΔrefZ (BAM1603), RBM_5mu (BAM910), Δnoc (BAM912) ΔrefZ Δnoc (BAM1604), RBM_5mu Δnoc (BAM920). CFP images scaled identically to allow for direct comparison of fluorescence. All images are the same magnification. Yellow arrowheads indicate examples of aberrantly dividing cells (Class II). (B) Cell division classes during sporulation.

While the double mutants displayed no qualitative delay in Spo0A-P activation, it was evident that the frequency of cells expressing CFP but lacking polar septa was increased (Fig. 2 and Fig. S1). To quantitate, the CFP channels of all images were scaled identically, and cells lacking polar septa were scored as either expressing CFP or not (Fig. 3 and Table 2). Notably, while only 5–12% of aseptate cells expressed CFP in wild type and the single mutants, this value increased to 36% and 54% in the ΔrefZ Δnoc and RBM_5mu Δnoc double mutants, respectively (Table 2). The increase was observed across biological and experimental replicates. We conclude that compared to wild type, ΔrefZ Δnoc and RBM_5mu Δnoc double mutants show a greater proportion of cells that initiate sporulation and then fail to progress to polar division during the experimental time course.

FIG 3.

Expression of CFP from Spo0A-dependent promoter P_spoIIG in nondividing cells. Images were captured 2.5 h following sporulation by resuspension. wt (BAM909), ΔrefZ Δnoc (BAM1604), RBM_5mu Δnoc (BAM920). Membranes were stained with TMA. CFP images scaled identically to allow for direct comparison of fluorescence. All images are the same magnification. Yellow dashes indicate examples of nondividing cells scored as CFP (−). Red plus signs indicate examples of nondividing cells scored as CFP (+).

TABLE 2.

Cell division and P_spoIIG-CFP signal classes during sporulation

Genotype	Aseptate n (aseptate/total) %	Class I n (class I/total) % (class I/class I&II) %	Class II n (class II/total) % (class II/class I&II) %	CFP (+) an = all cells (an/total) %	CFP (+) bn = aseptate (bn/total) % (bn/an) % (bn/aseptate) %
wt	355 (29)	844 (70) (99)	6 (<1) (1)	888 (74)	38 (3) (4) (11)
ΔrefZ	345 (27)	945 (73) (99)	11 (1) (1)	998 (77)	42 (3) (4) (12)
RBM5mu	461 (36)	822 (64) (99)	5 (<1) (1)	852 (66)	25 (2) (3) (5)
Δnoc	440 (39)	681 (60) (98)	12 (1) (2)	733 (65)	40 (4) (5) (9)
ΔrefZ Δnoc	518 (44)	628 (53) (94)	40 (3) (6)	856 (72)	188 (16) (22) (36)
RBM5mu Δnoc	628 (57)	448 (40) (92)	37 (3) (8)	821 (74)	336 (30) (41) (54)

^aTotal number of cells in the population (sum of aseptate, class I, and class II) that categorize as CFP (+). All class I & II cells categorize as CFP (+).

^bTotal number of aseptate cells that categorize as CFP (+).

The ΔrefZ Δnoc and RBM_5mu Δnoc double mutants divide aberrantly during sporulation.

At 2.5 h sporulation, the majority of wild-type and single mutant cells expressing P_spoIIG-CFP possessed an asymmetric septum or had progressed to the engulfment stage, when forespores appear rounded (Fig. 2 and Fig. S1). Asymmetric septa and engulfment were also observed in the ΔrefZ Δnoc and RBM_5mu Δnoc mutants; however, unlike the single mutants, cells with abnormal morphological features were also readily observed (Fig. 2 and Fig. S1). The most frequent abnormal phenotype was a cell with two septa, one polar and one at approximately midcell of the presumed mother cell compartment (Fig. 2, yellow arrowheads & Table 2). Notably, while only 1% of wild-type cells and 1–2% of the single mutants divided aberrantly (Table 2, Class II), this percentage increased to 6% and 7% in the ΔrefZ Δnoc and RBM_5mu Δnoc double mutants, respectively. A 3-fold increase in Class II divisions remained evident in the double mutants even after normalizing to the entire population of cells (dividing by the total number of Class I, II, and aseptate cells, regardless of CFP signal). We conclude that there is a synthetic interaction between the activities of RefZ (likely bound at RBMs) and Noc that inhibits aberrant divisions during wild-type sporulation.

The Δnoc ΔrefZ and RBM_5mu Δnoc double mutants activate SigF.

Following polar division, sporulation progression requires activation of SigF, the first forespore-specific sigma factor. To assess if the polar compartments resembling forespores activated SigF in the mutants, we monitored expression of a SigF-dependent reporter, P_spoIIQ-CFP (61–63) using epifluorescence microscopy (Fig. 4). The percentage of cells with activated SigF across the entire population of cells with forespore-like compartments was comparable for wild type (85%), ΔrefZ Δnoc (89%), and RBM_5mu Δnoc (86%) (Table 3). Consistent with our other results, only 1% of wild-type cells (4/659) exhibited aberrant (Class II) divisions, while the proportion was higher for the ΔrefZ Δnoc (6%) and RBM_5mu Δnoc (7%) mutants (Fig. 4 and Table 3). In wild type, 100% (4/4) of Class II cells with apparent forespores exhibited SigF activation. By comparison, this value decreased to 39% and 56% for the ΔrefZ Δnoc and RBM_5mu Δnoc mutants, respectively (Fig. 4, Table 3). We conclude that ΔrefZ Δnoc and RBM_5mu Δnoc cells that divide asymmetrically are still able to activate SigF in the smaller cell compartments, but with reduced frequency compared to wild type.

FIG 4.

Expression from P_spoIIQ-CFP, a SigF-dependent reporter. Images were captured 2.5 h following sporulation by resuspension. WT (BAM1638), ΔrefZ Δnoc (BAM1639), RBM_5mu Δnoc (BAM1640). Membranes were stained with TMA. CFP channels are scaled identically to allow for direct comparison. Images are identical magnification. Yellow arrowheads indicate examples of aberrantly dividing cells (Class II).

TABLE 3.

Cell division and P_spoIIQ-CFP forespore signal (SigF activation) classes during sporulation

Genotype	Class I “normal” n(% class I&II)	Class II “aberrant” n(% class I&II)	CFP (+) n (% class I&II) (% class II)	CFP (−)n (% class I&II) (% class II)
wt	655 (99)	4 (1)	563 (85) (100)	96 (15) (0)
ΔrefZ Δnoc	445 (93)	36 (7)	430 (89) (61)	51 (11) (39)
RBM5mu Δnoc	528 (94)	36 (6)	487 (86) (44)	77 (14) (56)

RefZ’s division regulation activity is required for preventing aberrant septum formation in the absence of Noc.

We previously identified 10 RefZ loss-of-function variants (rLOFs) that retain the ability to bind DNA but are no longer able to perturb Z-ring assembly when artificially induced during vegetative growth (64). Cells harboring the rLOF alleles in place of wild-type refZ miscapture DNA in the forespore indistinguishably from the ΔrefZ mutant, suggesting that RefZ-RBM complexes affect chromosome capture through direct or indirect effects on FtsZ (64). We hypothesized that RefZ’s ability to modulate FtsZ activity would also be required to prevent aberrant divisions in the absence of noc. To test, we first replaced native refZ with each of the 10 rLOF alleles in a Δnoc background and evaluated sporulation using the plate-based P_cotD-lacZ assay. In the presence of wild-type noc, each of the rLOF encoding variants supported sporulation at levels indistinguishable from wild type (wt) or a strain encoding wild-type refZ linked to a chloramphenicol resistance cassette (WT), which is isogenic to the rLOF strains (Fig. 5). Conversely, none of the rLOF variants supported wild-type sporulation in a Δnoc background (Fig. 5), as observed with the ΔrefZ Δnoc and RBM_5mu Δnoc double mutants (Fig. 1). These results suggest RefZ’s ability to affect FtsZ is also required to support wild-type sporulation in the absence of Noc.

FIG 5.

RefZ LOFs phenocopy ΔrefZ with respect to supporting sporulation in the absence of noc. (A) Plate-based sporulation assay based on lacZ expression from a late-stage sporulation promoter, P_cotD. Wild-type (wt) is strain encoding wild-type refZ. WT is refZ with genetic linkage a chloramphenicol resistance cassette (isogenic to the rLOF strains). (B) Images were captured 3.5 h following sporulation by resuspension. Membranes were stained with TMA. All images are at identical magnification. Yellow arrowheads indicate examples of aberrant (Class II) divisions.

To determine whether the sporulation defect observed in the rLOF Δnoc double mutants also resulted in increased aberrant divisions, we monitored division in sporulating cells using fluorescence microscopy. Aberrant divisions were rarely observed in the isogenic wild-type refZ strain or the rLOF mutant strains; however, when paired with Δnoc, each of the rLOF mutants phenocopied the ΔrefZ Δnoc and RBM_5mu Δnoc double mutants (Fig. 5). We conclude that the residues of RefZ that are required to affect division and support wild-type chromosome capture are also required to prevent abnormal divisions during sporulation in the absence of noc.

DISCUSSION

Both RefZ and Noc are DNA-binding proteins previously implicated in FtsZ regulation. RefZ is expressed early in sporulation and facilitates both timely polar Z-ring assembly (52) and precise capture of DNA in the forespore (51). A ΔrefZ mutant sporulates with wild-type efficiency, so the functional significance of RefZ activity during sporulation remains unclear (52). Noc is also expressed during sporulation but to our knowledge has not previously been associated with a sporulation function. Here, we find that deleting both noc and refZ results in a synthetic sporulation defect characterized by aberrant divisions and stalled sporulation progression.

We observed several types of sporulation defects in the ΔrefZ Δnoc mutant, each of which may contribute to the reduced sporulation observed in the plate-based assay. In one type, cells initiated the sporulation program, but failed to divide polarly. In a second, cells initiated sporulation and divided polarly, but failed to activate SigF, the first forespore-specific sigma factor. The largest observed category of aberrant cells divided polarly and activated SigF, but also possessed an extra septum near midcell. Some of these extra septa appear to curve at later time points, reminiscent of early engulfment (several examples can be seen in Fig. 5). The “mother” cell chromosome may be pumped into the forespore-distal compartment in these cells. If so, the curvature might be explained if the center compartment attempted to engulf the abnormally large “twin” via residual SigE programming. We did not investigate the phenotype further in the present study but propose the ΔrefZ Δnoc mutant may be useful for interrogating models of engulfment and spore morphogenesis (65).

Genetic interactions between Noc and other proteins implicated in cell division regulation have been observed previously (66, 67). Under conditions of rapid growth, a Δnoc ΔminD mutant is filamentous and lyses, suggesting interplay between the activities of Noc and MinD. Though we lack a mechanistic understanding of how Noc and RefZ influence Z-ring formation, the fact a sporulation defect was only observed when combining ΔrefZ with Δnoc (Fig. 1), suggests there is some specificity to the interaction. Both RefZ and Noc spread along DNA and require DNA-binding for activity (32, 51, 52). Transcriptomic profiling and identification of the RefZ and Noc binding sites did not reveal obvious regulons (32, 43, 52). Of note, the NO protein SlmA is also not considered to be a transcription factor in E. coli; however, SlmA has been shown to activate chitobiose utilization in another enteric, Vibrio cholerae (68). It may be informative to revisit the effects of RefZ, Noc, and SlmA on transcription under various growth conditions using more modern methods (RNA-seq versus microarrays) and also reexamine the regions where RefZ, Noc, and SlmA bind to look for relationships among the genes aside from position on the chromosome.

The noc gene evolved following a duplication of spo0J (parB) in the Firmicutes (69). In B. subtilis, Noc and Spo0J are 38% identical. Both proteins bind DNA and are regulated by CTP (70–73); however, unlike Spo0J (74–77), Noc does not appear to play a role in chromosome segregation; instead, a Δnoc mutant assembles aberrant Z-rings and/or divides over the chromosome, though only under conditions in which DNA replication and/or organization are perturbed (43, 78, 79). By comparison in Staphylococcus aureus, Δnoc mutants sometimes not only assemble extra Z-rings or divide over chromosomes, but also overinitiate DNA replication (79, 80). Several lines of evidence suggest that, at least in S. aureus, Noc’s influence on Z-ring assembly is sensitive to nucleotide pools. First, in cells lacking comEB (encoding a putative CMP/dCMP deaminase), noc becomes essential (79). Characterized CMP/dCMP deaminases generate dCMP from dUMP, a precursor required for dTTP synthesis. Second, mutations in dnaA that reduce DNA replication initiation suppress Δnoc ΔcomEB synthetic lethality and reduce the aberrant Z-rings associated with Δnoc (79). Third, Δnoc mutants are sensitized to DnaA overexpression compared to wild type (79). These results suggest that Noc has an activity that may buffer the cell against uncoordinated DNA replication and cell division. The reason for RefZ’s synthetic interaction with Noc remains unclear, though it is notable that the aberrant divisions and failures to progress in sporulation also occur at a time in development when new rounds of DNA replication are inhibited (54, 65). Although refZ falls within the spo0A regulon, transcriptional profiling suggests that sporulation is only one context in which refZ is expressed (47–50, 81). Any future models for RefZ function would benefit from incorporating these additional expression contexts, as well as the conservation of the RBMs across the genus.

MATERIALS AND METHODS

General methods.

Strains and details of strain construction can be found in the supplemental materials (Table S1 and Text S1). All B. subtilis strains were derived from B. subtilis 168. For microscopy, 25 mL cultures were grown in 250 mL baffled flasks placed in a 37°C in a shaking water bath. B. subtilis transformations were carried out as previously described (49), unless otherwise indicated. Sporulation was initiated by growing cells in CH medium followed by resuspension in sporulation medium (49). B. subtilis selections were carried out at the following antibiotic concentrations: 100 μg/mL spectinomycin, 7.5 μg/mL chloramphenicol, 10 μg/mL kanamycin, 10 μg/mL tetracycline, and 1 μg/mL erythromycin (erm) plus 25 μg/mL lincomycin for MLS. For transformation and selection in E. coli, antibiotics were included at the following concentrations: 100 μg/mL ampicillin and 25 μg/mL kanamycin.

PcotD-lacZ sporulation assay.

For the spot plate sporulation assays, isolated colonies were used to inoculate 4 mL of DSM broth (49) and cultures were grown at 37°C in a roller drum to midlog phase. All samples were normalized to the lowest recorded culture OD600 and 5 μL from each dilution was spotted on DSM agar plates supplemented with 40 μg/mL X-gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside). Plates were incubated overnight at 37°C prior to imaging with a ScanJet G4050 flatbed scanner (Hewlett Packard) using VueScan software and medium format mode. Images were processed using Adobe Photoshop (versus 12.0) and ImageJ64 (82).

Heat-kill assay.

Heat-kill assays to assess terminal sporulation were performed as described (83) except that DSM cultures were supplemented with 2.0 mg/mL tryptophan.

Fluorescence microscopy.

Three hundred to 500 μL samples were harvested at 6,010 × g for 1 min in a tabletop microcentrifuge. Supernatants were aspirated and pellets were resuspended in 3–5 μL of 1× PBS containing 0.02 mM 1-(4-[trimethylamino] phenyl)-6-phenylhexa-1,3,5-triene (TMA-DPH) (Invitrogen). Cells were mounted on glass slides with polylysine-treated coverslips. Images were captured with NIS Elements Advanced Research software (version 4.10) on a Nikon Ti-E microscope fitted with a CFI Plan Apo lambda DM 100× objective, Prior Scientific Lumen 200 Illumination system, C-FL UV-2E/C DAPI filter cube with a neutral density filter and a C-FL Cyan GFP filter cube using a CoolSNAP HQ2 monochrome camera. Images were captured for 1 s. Images were analyzed in NIS-Elements or ImageJ64 (82). To score cells as either expressing or not expressing CFP, all images were scaled identically, blinded, and manually classified by experimenter as shown in Fig. 3. Cells that could not be individually or fully resolved due to crowding or location on the periphery of an image were not included in the quantitation.