Unmasking Novel Sporulation Genes in Bacillus subtilis

Abstract

The Bacillus subtilis transcription factor σ^E directs the expression of a regulon of 262 genes, but null mutations in only a small fraction of these genes severely impair sporulation. We have previously reported that mutations in seven σ^E-controlled genes cause a mild (2- to 10-fold) defect in sporulation. In this study, we found that pairwise combinations of some of these seven mutations led to strong synthetic sporulation phenotypes, especially those involving the ytrHI operon and ybaN. Double mutants of ybaN and ytrH and of ybaN and ytrI had >10,000-fold lower sporulation efficiencies than the wild type. Thin-section electron microscopy revealed a block in cortex formation for the ybaN ytrH double mutant and coat defects for the ybaN single and ybaN ytrI double mutants. Sporulating cells of a ybaN ytrI double mutant and of a ybaN ytrHI triple mutant exhibited a pronounced loss of dipicolinic acid (DPA) between hours 8 and 24 of sporulation, in contrast to the constant levels seen for the wild type. An analysis of the spore cortex peptidoglycans of the ybaN ytrI and ybaN ytrHI mutants showed striking decreases in the levels of total muramic acid by hour 24 of sporulation. These data, along with the loss of DPA in the mutants, suggest that the developing spores were unstable and that the cortex underwent degradation late in sporulation. The existence of otherwise hidden sporulation pathways indicates that functional redundancy may mask the role of hitherto unrecognized sporulation genes.

A central challenge in the field of microbial development is to identify and characterize the functions of all of the genes that govern morphogenesis during cellular differentiation. An attractive system with which to address this challenge is the process of sporulation in the bacterium Bacillus subtilis, in which a growing cell is transformed via a series of well-defined morphological stages into a dormant cell known as a spore (or, more properly, an endospore) (23, 30). Soon after the commitment to sporulate, a two-chamber cell is formed, consisting of a large compartment called the mother cell and a small compartment called the forespore. Initially, the mother cell and the forespore lay side by side, but at an intermediate stage of development the forespore is wholly engulfed by and therefore resides within the mother cell (23). The engulfed forespore is surrounded by two membranes, one derived from the forespore and the other derived from the mother cell membrane that engulfs the forespore. Later in development, a thick layer of peptidoglycan known as the cortex is deposited in the space between the two membranes (24). Distinctive features of the cortex include the modified sugar muramic δ-lactam and a reduced level of muramic acid residues with peptide cross-links to other glycan chains (10). At approximately the same time as cortex formation, a thick protein shell is deposited around the outside of the forespore to create the spore coat (5). Ultimately, the mature spore is released by lysis of the mother cell.

Classical genetic as well as more recent genomic approaches have made it possible to identify genes that are activated during sporulation on a genome-wide scale. Strikingly, only a small subset of these genes appear to be essential for sporulation, based on analyses of null alleles (7, 8, 13, 26, 35). Most likely, in many cases the apparent lack of a sporulation phenotype is due to the mutations having effects that are undetectable under laboratory conditions but significant in the wild. However, it is also the case that the roles of some genes are largely masked by other genes. In other words, to observe the phenotype of a mutation in a given gene, a second mutation is required. An example of this type of redundancy is provided by the genes spoIIB and spoVG, in which mutations only mildly impair sporulation on their own but have a severe impact in combination (i.e., they have a synthetic phenotype) (18). Another synthetic effect involves a double mutant of cwlC and cwlH in which mother cell lysis is strongly blocked, while the corresponding single mutants exhibit little defect (21). A third example involves the activation of the forespore-specific transcription factor σ^F, which is governed by two independent regulatory pathways that are largely redundant (6).

In an effort to determine whether functional redundancy masks the role of other developmentally regulated genes in B. subtilis, we focused on genes under the control of the mother-cell-specific regulatory protein σ^E. The σ^E factor directs the transcription of an unusually large regulon, consisting of 262 genes, which are organized into 163 transcription units (excluding genes and operons that are activated under the dual control of σ^E and the DNA-binding protein SpoIIID) (8). We recently created null mutations in 98 newly identified genes and operons in the regulon. Strikingly, in only three cases did the mutation cause a severe defect in sporulation. In seven cases, however, the mutation caused a slight (2- to 10-fold) but reproducible defect in the efficiency of sporulation, as judged by the production of heat-resistant spores. We wondered whether the impact on sporulation of some of these seven genes is partially masked by redundancy with one or more other genes under the control of σ^E. As a test of this idea, we created strains bearing mutations in pairs of the seven genes. Here we report that certain pairwise combinations of mutations resulted in striking synthetic phenotypes in which cortex and coat formation were severely impaired.

MATERIALS AND METHODS

Strain construction.

The B. subtilis strains used for this study are listed in Table 1. Standard techniques were used for strain construction (11). All deletion strains were generated by the technique of long-flanking-homology PCR (33). Chromosomal DNAs obtained from these strains were analyzed by PCR to confirm the integration of the resistance cassette at the expected locus. The sequences of the primers used for the constructions are available upon request. To obtain strains that were doubly mutant, we transformed chromosomal DNA from one mutant strain into competent cells containing a replacement of a different gene with a second antibiotic cassette. The marker for JS31 was switched from spectinomycin to chloramphenicol by using plasmid pDAG32. The chromosomal DNAs of the mutants were transformed into the PS832 background for peptidoglycan analysis experiments.

TABLE 1.

Strains and plasmids used for this study

Strain or plasmid	Genotype	Source or Reference
B. subtilis strains
PY79	Wild type	34
PS832	Prototrophic revertant of strain 168	D. Popham
JS31	ypjBΔ::spc	This study
JS32	ytrIΔ::spc	This study
JS33	yunBΔ::kan	This study
JS35	ypjBΔ::spc::cat (pDAG32)	This study
JS37	ybaNΔ::tet	This study
JS38	ytvIΔ::tet	This study
JS39	ytrHΔ::spc	This study
JS40	ytvIΔ::erm	This study
JS42	ytrH ytrIΔ::spc	This study
JS43	ybaNΔ::tet ytvIΔ::erm	This study
JS44	ytrIΔ::spc ypjBΔ::cat	This study
JS45	ytrIΔ::spc yhbHΔ::erm	This study
JS46	ytrIΔ::spc ytvIΔ::tet	This study
JS47	ytrIΔ::spc ybaNΔ::tet	This study
JS48	ypjBΔ::spc ytvIΔ::tet	This study
JS49	ypjBΔ::spc yhbHΔ::erm	This study
JS50	ypjBΔ::spc ybaNΔ::tet	This study
JS51	yhbHΔ::spc	This study
JS52	ypjBΔ::spc yunBΔ::kan	This study
JS53	yhbHΔ::spc ytvIΔ::tet	This study
JS54	ytvIΔ::tet ytrHΔ::spc	This study
JS55	yhbHΔ::spc yunBΔ::kan	This study
JS56	ypjB::cat ytrHΔ::spc	This study
JS57	ybaNΔ::tet ytrHΔ::spc	This study
JS58	yhbHΔ::spc ybaNΔ::tet	This study
JS59	yhbHΔ::erm	This study
JS60	yunBΔ::kan ytrHΔ::spc	This study
JS61	yhbHΔ::erm ytrHΔ::spc	This study
JS62	yunBΔ::kan ybaNΔ::tet	This study
JS63	yunBΔ::kan ytvIΔ::tet	This study
JS69	ytrIΔ::spc yunBΔ::kan	This study
JS162	ytrH ytrIΔ::spc amyE::PytrHI-ytrH (cat)	This study
JS163	ytrH ytrIΔ::spc amyE::PytrHI-ytrI (cat)	This study
JS164	ytrH ytrIΔ::spc amyE::ytrHI (cat)	This study
JS176	ybaNΔ::tet ytrIΔ::spc amyE::PytrHI-ytrI (cat)	This study
JS177	ybaNΔ::tet ytrHΔ::spc amyE::PytrHI-ytrI (cat)	This study
JS224	ybaNΔ::tet amyE::PybaN-gfp-ybaN (cat)	This study
JS225	ytrIΔ::spc amyE::PytrHI-gfp-ytrI (cat)	This study
JS233	ybaNΔ::tet amyE::PybaN-gfp-ybaN (cat) cotE::erm	This study
JS234	ybaN::tet amyE::PybaN-gfp-ybaN (cat) spoIVA::erm	This study
PE477	ytrH-gfp	8
DPVB293	[PS832] ytrI::spc	This study
DPVB296	[PS832] ybaN::tet	This study
DPVB297	[PS832] ytrH ytrI::spc	This study
DPVB298	[PS832] ytrI::spc ybaN::tet	This study
DPVB299	[PS832] ytrH ytrI::spc ybaN::tet	This study
E. coli strain
DH5α	Cloning host	Laboratory stock
Plasmids
pDG364	Permits recombination into the chromosome at amyE	12
pDAG32	Converts spc to cat	Laboratory stock; similar to that discussed in reference 28
pJS1	pDG364 with ytrH	This study
pJS2	pDG364 with ytrHI	This study
pJS37	pDG364 with PytrH ytrI	This study
pJS58	pDG364 with PybaN gfp ybaN	This study
pJS59	pDG364 with PytrH gfp ytrI	This study
pJS60	pDG364 with gfp ybaN	This study
pJS61	pDG364 with gfp ytrI	This study
pPE56	pCV0119 with ytrH	8, 31
pCV0119	Permits Campbell integration of C-terminal GFP	8

gfp fusion genes were generated by in-frame fusions to appropriate coding sequences. 3′ gfp fusions were obtained by a Campbell-type (single reciprocal) recombination of plasmids generated by cloning XhoI- and BamHI-digested PCR fragments of the gene of interest into pCV0119, which carries gfp (8, 31). 5′ gfp fusions were constructed by placing the appropriate promoter in front of gfp, which in turn was fused in frame with the corresponding gene. The gfp fragment used for the 5′ fusion constructs was amplified by PCR from pKL47 (16) and digested with HindIII and XhoI. The PCR fragments for the open reading frames were digested with XhoI and BamHI and were ligated to gfp and pDG364 digested with HindIII and BamHI to create pJS60 and pJS61. pJS60 and pJS61 and the promoter PCR fragments were digested with EcoRI and HindIII and ligated to obtain pJS58 and pJS59. All plasmid constructions were performed in Escherichia coli DH5α by standard methods.

Sporulation efficiency.

One colony was inoculated into 5 ml of Difco sporulation (DS) medium (27). After growth for 30 h at 37°C, the number of spores was determined by heat killing for 10 min at 80°C.

Phase-contrast and fluorescence microscopy.

For phase-contrast microscopic observations, one colony from each mutant was inoculated into 5 ml of DS medium at 37°C overnight. A 0.3-ml sample of each culture was centrifuged briefly and resuspended in 10 μl of 1× phosphate-buffered saline. For fluorescence microscopic observations, strains were grown in hydrolyzed casein growth medium at 37°C and induced to sporulate by the resuspension method (11, 29). At the appropriate times, 0.3-ml aliquots were centrifuged and resuspended in 10 μl of 1× phosphate-buffered saline supplemented with the membrane stain FM4-64 (see Fig. 2) or FM1-43 (for engulfment experiments [not shown]) (Molecular Probes) at a final concentration of 1.5 μg ml⁻¹. Three-microliter samples of concentrated cell suspensions were placed on microscope slides and immobilized with poly-l-lysine-treated coverslips.

FIG. 2.

Subcellular localization of GFP-YbaN. Cells were collected after 3 h of sporulation at 37°C, treated with the vital membrane stain FM4-64, and observed by fluorescence microscopy. Bar, 1 μm.

Electron microscopy.

Cultures were induced to sporulate by exhaustion in DS medium at 37°C and samples were fixed after 24 h of sporulation. Thin-section electron microscopy was performed as previously described by Catalano et al. (3).

Spore peptidoglycan analysis.

Analyses of spore peptidoglycan synthesis and structure were performed in the B. subtilis strain PS832 background to maximize sporulation synchronization and peptidoglycan recovery. Mutations were introduced into the PS832 background as previously described (1) and sporulated by nutrient exhaustion in 2× SG medium (15) in the absence of antibiotics. Alkaline phosphatase and dipicolinic acid (DPA) accumulation were assayed as previously described (20). Total muramic acid accumulation was determined by an amino acid-sugar analysis, and the forespore peptidoglycan structure was quantified after muramidase digestion and high-performance liquid chromatography separation of the muropeptides as previously described (19).

RESULTS AND DISCUSSION

Synergy between mutations in σ^E-controlled genes.

Twenty-one strains were created that were doubly mutant for all possible pairs of the seven null mutations (Table 2). We reasoned that pairs of mutations for genes with redundant functions would act synergistically to impair sporulation. Synergy was defined as a sporulation efficiency that was ≥10-fold lower for a double mutant than that predicted by the simple product of the sporulation efficiencies of the two corresponding single mutants (6). By this criterion, 9 of the 21 pairs of mutations exhibited a synergistic effect (Table 2). For two cases (both involving ybaN), the impairment caused by two mutations was >2 (mutations of ybaN and ytrH) and almost 3 (mutations of ybaN and ytrI) orders of magnitude more severe than that predicted from the product of the effects of the two single mutations. In both cases, the double mutants exhibited a >10,000-fold defect in sporulation compared to the wild type. A triple mutant of all three genes exhibited a similarly severe defect in sporulation. Henceforth, we restrict our analysis to the interaction of ybaN with ytrH and ytrI.

TABLE 2.

Synergistic effects of pairs of mutations on sporulation efficiency

Genotype	Predicted sporulation efficiencyb	Actual sporulation efficiencyc	Synergistic effectd
WTa		1.0
yunB		0.5
ytvI		0.29
yhbH		0.25
ypjB		0.25
ytrI		0.14
ybaN		0.12
ytrH		0.11
yhbH ytvI	0.07	0.15	0.5
yhbH yunB	0.13	0.19	0.7
ytrH ytrI	0.02	0.05	0.4
ypjB yunB	0.13	0.12	1
yhbH ytrH	0.03	0.03	1
ypjB yhbH	0.06	0.04	1.5
yunB ytvI	0.15	0.13	1.2
yunB ybaN	0.06	0.02	3
yunB ytrH	0.06	0.02	3
ypjB ytrH	0.03	0.01	3
ytrI yhbH	0.07	0.018	4
yhbH ybaN	0.03	0.005	6
ytrI ytvI	0.04	0.003	13
ytvI ytrH	0.03	0.002	15
ytrI yunB	0.07	0.004	18
ytrI ypjB	0.04	0.002	20
ypjB ytvI	0.07	0.003	23
ypjB ybaN	0.03	0.001	30
ybaN ytvI	0.04	0.0008	50
ybaN ytrH	0.01	0.00006	167
ytrI ybaN	0.02	0.00003	667
ytrH ytrI ybaN	0.006	0.00006	100

^aWT (wild type) is strain PY79.

^bThe product of the sporulation efficiencies observed for the corresponding single mutants of the two indicated genes.

^cRatio of heat-resistant spores produced by the indicated mutant, after 30 h of culturing at 37°C, to those produced by the wild type. Values reported are the average of at least three experiments.

^dThe predicted sporulation efficiency divided by the observed sporulation efficiency.

Because ytrH is directly upstream of and in an operon with ytrI, it was possible that the strong sporulation defect of the ybaN ytrH double mutant was due to a polar effect. Therefore, we inserted a copy of ytrI that had been joined to the promoter for the operon into the chromosome at the amyE locus (resulting in strains JS176 and JS177) to test for its capacity to complement the sporulation defects of the ybaN ytrH and ybaN ytrI double mutants. The results showed that the presence of ytrI at amyE substantially alleviated the sporulation defect of the ybaN ytrI double mutant but not that of the ybaN ytrH double mutant. This finding indicates that the severe sporulation defect of the ybaN ytrH double mutant was not due to a polar effect, at least not exclusively so. Also, the mild sporulation defect of a deletion of the entire ytrHI operon was alleviated by the presence of a copy of the entire operon at amyE (in strain JS164) but not by the presence of ytrI at amyE. We concluded that both ytrH and ytrI contribute to proper sporulation and interact genetically with ybaN.

To investigate the nature of the sporulation defects caused by single and double mutants of ybaN, ytrH, and ytrI, we performed phase-contrast and fluorescence microscopy. Phase-contrast microscopy revealed that the ybaN and ytrI single mutants and the ytrHI operon mutant were all capable of producing phase-bright spores, but at efficiencies of only about 10 to 20% that of the wild type. In contrast, the ybaN ytrH and ybaN ytrI double mutants produced <1% phase-bright spores. As spores develop, they first appear dark gray by phase-contrast microscopy (phase dark), but as the cortex is deposited the spores then begin to refract light and appear bright white (phase bright), which indicates spore maturation. In the case of the ybaN ytrI double mutant (but rarely in the case of the ybaN ytrH mutant), some immature spores could be detected that appeared phase dark and were generally not released from the sporangium. When examined by fluorescence microscopy after treatment with the vital membrane stain FM1-43, none of the mutants, including the two double mutants, exhibited a conspicuous defect in asymmetric division or in engulfment. In toto, these results indicate that the strong block in sporulation that occurs in the absence of ybaN and either ytrH or ytrI takes place after engulfment.

Mutant spore morphology.

We examined the morphology of the mutant spores by a thin-section electron microscopic analysis of cells after 24 h of sporulation. The ytrI and ytrHI operon mutants produced free, mature spores that resembled those of the wild type. However, in contrast to the wild type, the mutants exhibited many sporangia in which spores were still sequestered in the mother cell (Fig. 1B and C). These sporangia were at various stages of development between stage IV, the point at which the cortex is first detectable but the coat is not yet present, and full maturity.

FIG. 1.

Thin-section electron micrographs of mutant spores. Cells were fixed after 24 h of sporulation. (A) PY79 (wild type); (B) JS32 (ytrI::Spc); (C) JS42 (ytrH ytrI::Spc); (D) JS37 (ybaN::Tet); (E) JS47 (ybaN::Tet ytrI::Spc); (F) JS57 (ybaN::Tet ytrH::Spc). OC, outer coat; IC, inner coat; Cx, cortex; Cr, core. Bars, 0.125 μm (A), 1 μm (B and F), 0.5 μm (C and E), and 0.25 μm (D).

Sporulating cells of the ybaN mutant exhibited several defects. First, similar to the observations with the ytrI and ytrHI mutants, even after 24 h of sporulation a significant number of sporangia had failed to release their prespores, many of which had not yet assembled coats. Unlike the ytrI and ytrHI mutants, however, ybaN mutant spores had defective outer coats. In contrast to wild-type spores, whose outer coats are composed of one or more uniformly dark-staining layers (Fig. 1A), ybaN mutant spore coats were thin and appeared mottled. Most strikingly, however, a significant number of released ybaN mutant spores possessed ridges that were much more deeply folded than those of wild-type spores (Fig. 1D). This could have been due to defects in either the coat or the cortex. Finally, the ybaN ytrI double mutant resembled the ybaN single mutant, producing immature spores with defective coats (Fig. 1E). Many of the ybaN ytrH double mutant cells appeared to have lysed during sporulation, and the remainder were blocked at the stage of forespore engulfment prior to cortex synthesis (Fig. 1F). These data corroborate the sporulation blocks seen with phase-contrast and fluorescence microscopy.

Spore peptidoglycan analysis.

To investigate whether the mutants were defective in cortex formation, we analyzed muramic acid accumulation and peptidoglycan structures during sporulation. Before performing this analysis, we confirmed that the time course of sporulation for the mutants was similar to that for the wild type by measuring the onset of the accumulation of alkaline phosphatase, a marker for early sporulation, and DPA, a marker of late sporulation. All of the mutants produced alkaline phosphatase with the same timing and abundance as the wild type. All of the mutants also accumulated DPA to wild-type levels after 4 to 7 h of sporulation, but from 8 to 24 h, the level of DPA in the mutants decreased whereas that of the wild type remained approximately constant. The loss of DPA was most severe for the ybaN ytrHI triple and ybaN ytrI double mutants, which exhibited a sixfold decrease in DPA levels relative to the wild type at 24 h, a finding that might indicate that the mutants produced unstable spores that were unable to retain their cytoplasmic contents.

Next, cells were collected from hours 2 to 8 and at hour 24 of sporulation and were analyzed for their peptidoglycan abundance and composition. The muramic acid contents and spore peptidoglycan structures of the mutants were similar to those of the wild type from hours 2 to 6 of sporulation, at which time the wild type had produced about 50% of its total spore peptidoglycan content (data not shown). From hours 6 to 8 of each sporulation, the mutant strains exhibited defects in spore peptidoglycan accumulation. The ytrI and ytrHI mutants accumulated >60%, the ybaN mutant accumulated >40%, and the ybaN ytrI double mutant and the ybaN ytrHI triple mutant accumulated <40% of the amount of muramic acid accumulated by the wild type. In each case, the structure of the spore peptidoglycan was similar to that of the wild type, except for slight decreases in the percentages of muramic acid in the lactam form and with single l-Ala side chains and slight increases in the percentages of muramic acid residues with tripeptide side chains (Table 3). These structural changes are consistent with a decreased synthesis or degradation of spore peptidoglycan, leaving a larger percentage of the spore peptidoglycan as the germ cell wall (19). After 24 h, the ytrI and ytrHI mutants exhibited ∼25% and the ybaN mutant exhibited 4% of the amount of muramic acid accumulated by the wild type. The most striking results were observed at 24 h for the ybaN ytrI double mutant and the ybaN ytrHI triple mutant, which exhibited <2% of the amount of muramic acid accumulated by the wild type. In toto, the results shown in Table 3 suggest that the double mutants do make a spore cortex with a relatively normal structure but that the developing spores are unstable and the cortex is degraded late in sporulation.

TABLE 3.

Analysis of spore peptidoglycans

Strain	% of muramic acid with indicated side chainsa								% of muramic acid with cross-links		% Total muramic acid
	δ-Lactam		Ala		TriP		TP
	T8c	T24c	T8	T24	T8	T24	T8	T24	T8	T24	T8	T24
WTb	45.9	48.3	25.9	26.8	2.8	2.1	25.4	22.9	3.0	2.7	70.3	71.9
ytrI	46.2	46.8	26.9	27.5	3.8	3.5	23.1	22.1	3.1	2.7	45.3	17.8
ybaN	41.2	42.4	18.4	16.7	6.1	7.4	34.2	33.5	3.1	3.2	30.5	2.7
ytrHI	44.2	45.8	25.3	26.7	4.9	4.2	25.5	23.2	3.4	2.9	44.8	18.4
ybaN, ytrI	40.4	35.5	20.6	20.4	7.7	16.5	31.4	27.6	3.2	3.9	26.2	0.9
ybaN, ytrHI	41.1	35.3	19.0	16.3	7.0	16.2	32.9	32.2	3.1	3.5	21.6	0.4

^aAbbreviations: TriP, tripeptide; TP, tetrapeptide.

^bWT, wild type (PS832).

^cT₈ and T₂₄, 8 and 24 h after sporulation, respectively.

Subcellular localization of YbaN.

The effects of ytrH, ytrI, and ybaN mutations on cortex stability raised the question of whether any of the three gene products are localized in the mother cell membrane that surrounds the engulfed forespore. Indeed, the predicted amino acid sequences of all three gene products are consistent with the idea that they are membrane proteins. To investigate the subcellular localization of the proteins, we attempted to create functional, in-frame fusions of the genes to gfp. We found that fusions of gfp to the 5′ or 3′ end of ytrI or to the 3′ end of ytrH were nonfunctional (data not shown). However, a fusion of gfp to the 5′ end of ybaN was functional. We initially detected fluorescence from this fusion after 2 h of sporulation, and it ultimately appeared as a ring around the forespore (Fig. 2). Since YbaN is synthesized only in the mother cell (as a result of σ^E-dependent expression of this gene), we inferred that YbaN is associated with, or very close to, the outer forespore membrane. This localization was not dependent on SpoIVA or CotE, which are morphogenetic proteins that are required for the assembly of the spore coat around the outer forespore membrane (5). Since a spoIVA mutation results in the release of the entire coat from the forespore (22), we took the lack of dependence of YbaN assembly on spoIVA to indicate that YbaN is not a coat protein. There are several other SpoIVA-independent, outer forespore-associated proteins that affect spore formation, including SpoVM (17, 25, 32), YabP, and YabQ (2, 31). The expression of ybaN and ytrHI is under the negative control of SpoIIID, which is the case for most integral membrane proteins under σ^E control for which the subcellular localization has been determined (7). A gel shift analysis has confirmed that SpoIIID binds to the promoter of ybaN (7). We concluded that YbaN is located in the outer forespore membrane, from which it may participate in the formation or stability of the spore cortex.

Protein homology.

To look for clues about the functions of ytrH, ytrI, and ybaN, we searched the databases for proteins with significant similarities to the inferred products of these genes. All three genes have orthologs in other endospore-forming bacteria (8), but YtrH and YtrI exhibited little similarity to the products of other genes in these species. Interestingly, orthologs for ytrH and ytrI were found only in the Bacillus genus and not in Clostridium. In all of the currently published genomes for endospore-forming bacteria, both genes are always present in tandem, suggesting that they may be functionally related. We also noted that the inferred stop codon (TGA) for the ytrH open reading frame in B. subtilis overlaps with the inferred start codon (ATG) for ytrI (the overlapping sequence being ATGA), which suggests that translation of the two open reading frames is coupled. If so, this might indicate that the levels of synthesis of YtrH and YtrI are closely coordinated and perhaps that the two proteins interact with each other.

Whereas YtrH and YtrI are not significantly similar to other proteins in the databases, YbaN is paralogous to a known polysaccharide deacetylase in B. subtilis, the pdaA gene product. pdaA is a developmentally regulated gene that is transcribed in the forespore under the control of σ^G (9) and is required for the production of essentially all muramic δ-lactam, which is unique to the spore cortex and serves as a specificity determinant for autolytic enzymes that perform cortex degradation during spore germination (10). The similarity of YbaN to a polysaccharide deacetylase, as well as its sequestration to the outer forespore membrane, suggests that it has some role in spore cortex formation. However, the subtle phenotype of a ybaN mutation indicates a function distinct from that of pdaA.

Nomenclature.

In toto, the results of this study indicate that when they are present individually, mutations of ybaN and of the ytrHI operon impair sporulation at a late stage (VI) (22) of morphogenesis. In other words, the mutations did not block cortex or coat synthesis but did affect the stability of the former and the structure of the latter in the case of the ybaN mutation. On this basis, and following the nomenclature of Piggot and Coote (22), the most recent compilation of spo genes (23), and a recent assignment of a spoVIF gene (14), we propose the designations spoVIE for ybaN and spoVIGA and spoVIGB for ytrH and ytrI, respectively. Of course, as we have seen, pairwise combinations of mutations in ybaN (spoVIE) with mutations in either ytrH (spoVIGA) or ytrI (spoVIGB) severely impaired the production of heat-resistant spores and hence exhibited a block in sporulation at an earlier stage than stage VI.

Conclusions.

The σ^E factor directs the transcription of an especially large regulon, but mutations in relatively few (<20%) members of the regulon cause a severe defect in sporulation. The principal contribution of the present work is the demonstration that some genes in which a mutation causes only a very mild defect in development do in fact play critical roles in sporulation but that these roles are normally masked by redundancy with other genes. Thus, certain pairwise combinations of mutations revealed strong synthetic defects in sporulation. We presume that in these cases the genes in question function in separate pathways that contribute independently to a common step in morphogenesis. Thus, for example, we infer that ybaN (spoVIE) on the one hand and ytrH (spoVIGA) and ytrI (spoVIGB) on the other hand contribute independently to some common aspect of cortex formation or stability. It is conceivable that the functions of other genes in the σ^E regulon are also masked by redundancy. A test of this will require the large-scale creation of all possible double mutants for all members of the regulon for which a mutant phenotype was not observed. We note that describing a gene as redundant does not necessarily imply that it is dispensable but rather that laboratory assays are insufficient to detect its role. This is particularly likely in the case of the cortex, a complex structure that is essential to spore survival. Even a small increase in its protective ability (as may hypothetically be contributed by YbaN [SpoVIE]) is likely to be readily selected in the environment. Complex synergistic interactions between sporulation genes are not restricted to those involved in cortex formation. For example, analyses of strains bearing multiple coat protein gene mutations revealed synergistic interactions that would not be predicted from the corresponding single gene mutations (4). Most likely, more novel functions for sporulation genes, both those that were recently identified and those that have been previously characterized, remain to be revealed by the creation of multiply mutant strains.