Sporulation during Growth in a Gut Isolate of Bacillus subtilis

Abstract

Sporulation by Bacillus subtilis is a cell density-dependent response to nutrient deprivation. Central to the decision of entering sporulation is a phosphorelay, through which sensor kinases promote phosphorylation of Spo0A. The phosphorelay integrates both positive and negative signals, ensuring that sporulation, a time- and energy-consuming process that may bring an ecological cost, is only triggered should other adaptations fail. Here we report that a gastrointestinal isolate of B. subtilis sporulates with high efficiency during growth, bypassing the cell density, nutritional, and other signals that normally make sporulation a post-exponential-phase response. Sporulation during growth occurs because Spo0A is more active per cell and in a higher fraction of the population than in a laboratory strain. This in turn, is primarily caused by the absence from the gut strain of the genes rapE and rapK, coding for two aspartyl phosphatases that negatively modulate the flow of phosphoryl groups to Spo0A. We show, in line with recent results, that activation of Spo0A through the phosphorelay is the limiting step for sporulation initiation in the gut strain. Our results further suggest that the phosphorelay is tuned to favor sporulation during growth in gastrointestinal B. subtilis isolates, presumably as a form of survival and/or propagation in the gut environment.

INTRODUCTION

Bacillus subtilis is a naturally competent, spore-forming Gram-positive bacterium that can be isolated from many aquatic and terrestrial environments. The capacity to develop highly resistant endospores in response to nutrient scarcity, allied with their dispersal, explains in part the isolation of this organism from a wide range of environmental samples (1–3). However, evidence also suggests that B. subtilis actively grows in soil, plant roots, and the gastrointestinal (GI) tract of various organisms (2, 4–8). In particular, its recurrent isolation from the gut of several invertebrates and vertebrates, including humans (9), suggests that B. subtilis might colonize the GI tract as part of its natural life cycle. Spores appear to be important for the ability of B. subtilis to occupy the GI tract. Spores resist digestion by protozoa and passage through the GI tract of the mouse (10, 11). Moreover, spores are also able to germinate and undergo new cycles of growth and sporulation in the GI tract of mice (11). The ubiquitous nature of B. subtilis is also supported by its genomic diversity, as revealed by a recent study (3, 12). As more strains of this species are sequenced and studied, it is expected that niche-specific genes will be identified (3, 12).

In the laboratory, spore formation by B. subtilis is a cell density-dependent response to nutrient exhaustion. Sporulation is triggered after cells enter the stationary phase in a medium that supports sporulation, and it takes about 8 to 10 h to be completed (13, 14). Central to the decision to enter sporulation is the activation, by phosphorylation, of the response regulator Spo0A. Transfer of phosphoryl groups to Spo0A occurs via a series of components called the phosphorelay, served by five sensor kinases (KinA to -E) that autophosphorylate at a histidine residue (15–19) (Fig. 1A). Phosphate is then transferred to an aspartate residue of the Spo0F response regulator and finally to an aspartate of the response regulator domain of Spo0A, via a histidine residue in the Spo0B phosphotransferase (13, 15, 17, 19) (Fig. 1A). Once on its active form, Spo0A∼P directs the transcription of a regulon of more than 100 genes and operons (20) (Fig. 1A). However, the expression of these genes follows a temporal progression, as Spo0A∼P accumulates gradually during entry into stationary phase (21). Genes such as those involved in cannibalism or biofilm formation (22, 23), which have high-affinity binding sites for Spo0A∼P, are expressed at low levels of Spo0A∼P (low-threshold genes) (24). In contrast, genes such as those required for sporulation have low-affinity binding sites for Spo0A∼P and will only be activated later and only if Spo0A∼P accumulates to sufficiently high levels (high-threshold genes) (21, 24). Thus, cells will attempt several responses to starvation, prior to, as a last resource, and upon embarking into the developmental pathway of sporulation (25, 26).

FIG 1.

Activation of Spo0A and the initiation of sporulation. (A) Simplified representation of the circuits involved in the initiation of sporulation in B. subtilis. The figure shows the membrane-associated KinA and KinB kinases, as well as the cytosolic KinC to -E kinases that feed phosphoryl groups into the phosphorelay formed by the Spo0F and Spo0B proteins. The figure also illustrates the actions of the RapA phosphatase and the PhrA peptide. Other Spo0F∼P phosphatases are indicated by * and other peptides by **. Panel B represents a simplified version of the noise generator model that accounts for the generation of heterogeneity in the levels of phosphorylated Spo0A during entry into the stationary phase. The positive-feedback loops governing the synthesis of Spo0A and Spo0F are “just-in-time” elements that maintain proper levels of both proteins as the levels of Spo0A∼P increase. In contrast, the “time delay” circuit that operates through Spo0E postpones the accumulation of high levels of Spo0A at the onset of sporulation. Finally, the flow of phosphoryl groups from the kinases (only KinA is represented for simplicity) through Spo0F and Spo0B acts as a noise generator that creates heterogeneity in the levels of Spo0A∼P across the population.

The phosphorelay serves as a circuit to integrate nutrient availability, cell density, metabolic, cell cycle, DNA replication, and other signals that influence the level of Spo0A∼P (15, 17, 20, 21, 27). The phosphorelay is also negatively regulated by phosphatases that act on Spo0F∼P or Spo0A∼P. Dephosphorylation of Spo0F∼P is promoted by phosphatases of the Rap family, whose action is antagonized at high cell densities upon reinternalization of extracellular cognate inhibitory peptides (Phr) (28–34) (Fig. 1A). Spo0A∼P, in turn, is the substrate for the Spo0E family of phosphatases (35, 36). During the transition from exponential growth to stationary phase, starvation-sensing pathways (35, 37, 38) converge with quorum-sensing pathways (via the Rap-Phr family of phosphatases) to regulate the phosphorylation of Spo0A (39). The phosphorelay acts as a noise generator, causing Spo0A∼P to be activated in a highly heterogeneous pattern across the cell population (Fig. 1B). Noise generation by the phosphorelay results from fluctuations in the levels of expression of the genes for the phosphorelay components (40, 41) and also from the action of the Rap phosphatases (42–44) (Fig. 1B). Deletion of the genes for the Rap phosphatases reduces noise and allows for increased, more synchronized sporulation (42, 43). Since spore differentiation in B. subtilis is a time- and energy-consuming process that following the asymmetric division of the sporangial cell cannot be reverted, the noisiness of the phosphorelay ensures that not all cells commit to sporulation at once, thus providing a bet-hedging strategy in a potentially changing environment (27, 42, 43).

In this investigation, we characterized the sporulation of an undomesticated isolate of B. subtilis, called BSP1, isolated from the broiler chicken gut (8). We show that BSP1 initiates sporulation during growth and hence bypasses the cell density, nutritional, and other signals that normally restrain the flow of phosphoryl groups to Spo0A and delay sporulation. We show that sporulation is initiated during growth because of increased activation of Spo0A and that this primarily results from the absence of the genes for two Rap phosphatases. Our results also suggest that the increased activation of Spo0A is carefully balanced in BSP1, so that growth is not compromised. Because different combinations of Rap phosphatases appear to be absent from other gut strains of B. subtilis, which also show increased sporulation, we posit that contrary to laboratory strains in which sporulation is delayed, the initiation of spore formation during growth may represent an adaptation of B. subtilis to the gut environment.

MATERIALS AND METHODS

General methods.

Luria-Bertani (LB) medium was used for the routine growth of both B. subtilis and Escherichia coli. Sporulation of B. subtilis was induced by nutrient exhaustion in Difco sporulation medium (DSM), and its efficiency (defined as the ratio of heat-resistant spores relative to the total viable cell count) was assessed during the growth and stationary phases as described previously (45). For B. subtilis and when appropriate, the following antibiotics were used: erythromycin (1 μg ml⁻¹), neomycin (1 μg ml⁻¹), chloramphenicol (5 μg ml⁻¹), kanamycin (10 μg ml⁻¹), tetracycline (10 μg ml⁻¹), spectinomycin (100 μg ml⁻¹), or erythromycin (1 μg ml⁻¹) plus lincomycin (25 μg ml⁻¹), for macrolide-lincosamide-streptogramin B (MLS)-resistant strains.

Microarray-based comparative genomic hybridization.

B. subtilis 168-specific oligonucleotide microarrays were used to identify genes that were absent from or divergent in strain BSP1, as described before (12). Each array contains 3,722 gene-specific 60- to 70-mer oligonucleotides representing approximately 91% of the genes predicted for strain 168 (12). A gene was considered absent or divergent if the log₂ ratio of fluorescence for the standard sequences of strain 168 relative to BSP1 at any given gene spot was >1 (12). Normalization was achieved by adjusting all ratios so that the median of all log₂ ratios for a given experiment was zero (12).

Plasmid and strain construction.

The B. subtilis standard laboratory strain PY79, the gastrointestinal isolate BSP1, and all of their congenic derivatives constructed during the course of this work are listed in Table S1 in the supplemental material and described in detail in the supplemental material. Plasmid construction is also described in the supplemental material and listed in Table S2 in the supplemental material. Oligonucleotide primers used for PCR, mutagenesis, or sequencing are listed in Table S3 in the supplemental material. Competent cells of B. subtilis were prepared and transformed as described previously (46). In crosses involving chromosomal DNA, the lowest concentration allowing efficient transformation was used (around 1 μg ml⁻¹ of a typical suspension of competent cells) to avoid congression (46).

Fluorescence microscopy.

For routine visualization of fluorescence, 600 μl of sporulating cultures was removed at the appropriate times, centrifuged for 30 s at 3,000 × g, and resuspended in 200 μl of 1× phosphate-buffered saline (PBS). When specified, cells were first stained for DNA visualization with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI [Sigma]) at a final concentration of 2 μg ml⁻¹ for 1 min at room temperature in the dark. They were next labeled with the membrane dye FM4-64 (Molecular Probes) at a final concentration of 10 μg ml⁻¹. For mixed experiments (see Fig. 3), cellular suspensions of PY79 were stained with DAPI, while cells of BSP1 were not. Following incubation, both cell suspensions were washed twice with 600 μl of 1× PBS and resuspended in 200 μl of 1× PBS. Cell membranes were then stained with FM4-64 before 50 μl of each strain was mixed for observation. In all experiments, 2 μl of cell suspensions was mounted on agarose pads (1.7% in H₂O) before microscope observation. Phase-contrast and fluorescence images were acquired with a Leica DMRA2 microscope equipped with a ×63 magnification objective and a CoolSNAP HQ Photometrics camera (Roper Scientific). Images were acquired with Leica FW4000 software and prepared for publication using Adobe Photoshop. Quantification of the signal from P_spo0A-yfp (Fig. 3C; also see Fig. 7B) was done in areas of 4 by 4 pixels, using the “measure” option of Leica FW4000 software. Around 500 cells were analyzed for each strain at each time point tested. Quantification of the green fluorescent protein (GFP) signal (see Fig. S2B in the supplemental material) was done for 500 cells of each strain using the ImageJ software (http://rsb.info.nih.gov/ij/).

FIG 3.

Increased expression of P_spo0A-yfp during growth of BSP1. (A) Immunoblot analysis of Spo0A accumulation for BSP1 and PY79 during growth in DSM (1 h before the onset of stationary phase, or h 0) and as cells enter stationary phase (h 0 and 1). The samples were also probed for the levels of the β′ subunit of RNA polymerase, as a loading control. (B) Samples were collected from DSM cultures of BSP1 and the laboratory strain PY79 bearing a P_spo0A-yfp fusion (AH7325 and AH7326, respectively), at the indicated times before and after the end of exponential growth (T₀). Cells of strain PY79 were labeled with the DNA stain DAPI and mixed in equal proportions with cells of BSP1 (which were not stained with DAPI). The cell mixture was then stained with the membrane dye FM4-64 and observed by phase-contrast and fluorescence microscopy. The first column shows phase-contrast images, the middle column shows images of the same cells stained with DAPI and FM4-64, and the right column shows P_spo0A-yfp expression. Vegetative cells of BSP1 derivative (cells labeled with arrowheads) show stronger spo0A-yfp expression at all times compared with the same cells in the PY79 background (cells labeled with stars). Expression levels are higher (and confined to the mother cell) in cells that have overcome the asymmetric septation, with a maximum at T₀ (cell marked with “1”) and a decrease by h 1 (cell marked with “2”). Scale bar, 2 μm. (C) Quantitative analysis of the YFP signal in cells of BSP1 (AH7325) or PY79 (AH7326) at the indicated times (in hours) relative to T₀ in DSM. The quantification was based on images such as those in panel A; the cells scored in each class (y axis) did not show morphological signs of sporulation (see Materials and Methods). The dotted boxes represent the percentage of cells with a YFP signal higher than an arbitrary threshold value, defined as the value at T₋₁ above which the laboratory strain PY79 had 0% of cells.

FIG 7.

Transplantation of rapE-phrE, rapI-phrI, and rapK-phrK reduces sporulation of BSP1. (A, top) The rapE-phrE operon, present at 228° in the genetic map of B. subtilis 168, was introduced at the amyE locus of BSP1 (at 25°, as represented), whereas rapI-phrI (at 47° in strain 168) and rapK-phrK (at 176°) were introduced at the nonessential thrC locus (284°). (A, bottom) Kinetics of heat-resistant spore formation in DSM for strain BSP1 and congenic derivatives bearing the following rap-phr operons: rapE-phrE (AH7324), rapK-phrK (AH7369), rapI-phrI (AH7375), rapE-phrE and rapK-phrK (AH7370), or rapE-phrE and rapI-phrI (AH7376). The rapE-phrE operon was introduced at the amyE locus, whereas rapK-phrK or rapI-phrI was introduced at thrC, as indicated. The titer of heat-resistant spores was measured at the indicated times (in hours) before or after the onset of sporulation. The dashed line indicates the sporulation level of BSP1 at T⁻¹. The results shown are the averages of results from three independent experiments with error bars representing the standard deviation. (B) Distribution of the YFP signal in BSP1 (AH7325) and the BSP1 derivatives AH7324 (rapE⁺), AH7323 (rapK⁺), and AH7424 (rapE⁺ rapK⁺), bearing a P_spo0A-yfp fusion. The various cultures were grown in DSM, and samples were collected at the onset of sporulation (T₀) for fluorescence microscopy and quantitative analysis (Materials and Methods). The cells scored in each class (y axis) did not show morphological signs of sporulation (see Materials and Methods); dotted boxes show the percentage of those cells with a fluorescence signal higher than an arbitrary threshold number, defined as the value shown by more than 60% of the BSP1 cells. (C) Percentage of cells with no morphological signs of sporulation (predivisional cells) and with a visible asymmetric septum or other morphological features indicative of later stages of sporulation (postdivisional cells) for the indicated strains.

Immunoblotting.

Cultures of strains BSP1, PY79, and the derivatives of BSP1 (AH7446) and PY79 (AH7445) carrying P_xylA-spo0A fusions were collected 30 min after induction with different concentrations of xylose (0, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, and 1%). Cultures of strains AH7490 and AH7491 (derivatives of PY79 and BSP1, respectively, carrying P_xylA-spo0A and P_spoIIG-cfp [expressing cyan fluorescent protein] fusions) were grown in DSM to an optical density at 600 nm (OD₆₀₀) of ≈0.9 and induced with 0.005% xylose, and samples were collected at various times thereafter (10, 20, 30, 45, 60, 120, 180, 240, and 300 min). The derivatives of BSP1 (AH7431) and PY79 (AH7430) carrying a P_xylA-gfp fusion were induced for 30 min with various concentrations of xylose (as described above). In all cases, samples of 10 ml were collected. Preparation of whole-cell lysates, electrophoretic resolution of proteins by SDS-PAGE (12%), and immunoblotting were performed as described before (47). Anti-Spo0A, anti-σ^A, and anti-GFP polyclonal antibodies were used at concentrations of 1:5,000, 1:10,000 and 1:2000, respectively. Quantification of immunoblot signals was done with the ImageJ software (http://rsb.info.nih.gov/ij/).

RESULTS

Sporulation during growth in a gut isolate of B. subtilis.

When grown on sporulation plates (DSM), strain BSP1 formed opaque colonies, a symptom of spore formation (22), faster than a reference laboratory strain such as PY79 did (not shown). This observation prompted us to test whether BSP1 sporulated faster than PY79 on liquid DSM. We used phase-contrast and fluorescence microscopy to examine progress through the morphological stages of sporulation for derivatives of both strains bearing a ywcE-gfp translational fusion. The ywcE gene codes for a holin-like protein that localizes to the cell and spore membranes, allowing their visualization throughout sporulation (48). The results in Fig. 2A show that at the onset of stationary phase in DSM (time zero [T₀], normally defined as the onset of sporulation), many cells of the BSP1 derivative showed engulfment initiation. In contrast, such morphogenetic signs were only detected for the PY79 derivative 2 h after T₀. Moreover, phase-bright spores were observed at 4 h (T₄) for BSP1 but only at T₈ for the PY79 derivative (Fig. 2A). To test whether morphogenesis was faster in the BSP1 derivative, we determined the time at which at least 40% of the cell population had undergone septation, showed complete engulfment of the forespore by the mother cell, or presented phase-gray or phase-bright spores. The results in Fig. 2B show that progress through these stages of sporulation occurred at the same pace for the two strains, in spite of the time difference relative to T₀ at which they were reached. This suggested that BSP1 initiated sporulation during growth. To have a more quantitative description of the process of sporulation, we measured the titer of heat-resistant spore formation for the two strains when grown on liquid DSM. The maximum titer of heat-resistant spores (about 3 × 10⁸ per ml of culture) was attained for strain PY79 10 h after the entry into the stationary phase, and the threshold of 10⁸ spores/ml was first observed by h 6 (Fig. 2C). For BSP1 however, 10⁸ spores per ml were already observed at h 3, and the maximum value of about 9 × 10⁸ spores/ml was reached at h 7. Moreover, the titer of heat-resistant spores during exponential growth (2 h before T₀) was higher for BSP1 (≈10⁵ spores per ml of culture) than for PY79 (10² spores/ml) (Fig. 2D). The data are consistent with the conclusion that morphogenesis proceeded at the same pace for the two strains (see Fig. S1A in the supplemental material). Also, in agreement with the inference that BSP1 initiated sporulation efficiently during growth, when late-log-phase BSP1 cultures were diluted, the spore titer reached the original, predilution number in about 2 h (see Fig. S1B). In contrast, for PY79, the number of spores did not increase following dilution. This implies the presence in BSP1 cultures of a high proportion of spores close to the stage of becoming heat resistant (see Fig. S1). We conclude that BSP1 initiates sporulation during growth, bypassing the nutritional and cell density signals that normally delay sporulation until entry into the stationary phase.

FIG 2.

Sporulation during growth of BSP1. (A) Derivatives of strains BSP1 (AH7350) and PY79 (AH7351) bearing a ywcE-gfp fusion were induced to sporulate in DSM, and samples were collected at T₀ and at the indicated times (in hours) thereafter. The cells were observed by phase-contrast (PC) and fluorescence (FL) microscopy. YwcE-GFP labels the cell and spore membranes (48) and was used to visualize the spore membranes following completion of engulfment. Signs of asymmetric division (ad), engulfment initiation (ei), engulfment completion (ec), phase gray (pg), phase bright (pb), and free spores (sp) are highlighted with arrowheads. Scale bar, 2 μm. (B) Time (in hours, after the onset of the stationary phase in DSM, at which at least 40% of the BSP1 or PY79 population reach the indicated morphological stages of sporulation. The top row shows phase-bright or phase-gray spores, the middle row engulfment completion, and the bottom row asymmetric septation. The quantification is based on the analysis of images like those of panel A (see Materials and Methods). (C and D) Kinetics of heat-resistant spore formation by strains BSP1 and PY79 in DSM. T₀ indicates the end of the exponential growth phase. The spore titer was measured at the indicated times (in hours before or after T₀). Panel C shows the appearance of at least 10⁸ heat-resistant spores per ml of culture from T₀ to T₁₀, whereas panel D shows the evolution of the spore titer from T₋₃ to T₃. The results shown are the averages of results from three independent experiments with error bars representing the standard deviation.

Spo0A is active in a large fraction of BSP1 cells.

Since sporulation initiation depends on the activation, by phosphorylation, of Spo0A (13), we reasoned that perhaps this protein was active in a larger fraction of the BSP1 population of growing cells compared to the laboratory strain PY79. We first used immunoblot analysis to look at the levels of Spo0A in cultures of BSP1 and PY79 entering the stationary phase in DSM. We found the levels of Spo0A around the onset of sporulation (T₋₁, T₀, and T₁) to be slightly higher for BSP1 than for PY79 (Fig. 3A). However, the levels of a control protein, the β′ subunit of RNA polymerase, dropped significantly during the stationary phase for BSP1 (Fig. 3A). While this is most likely due to the high proportion of spores found for BSP1 early in the stationary phase, and from which proteins are more difficult to extract, it became clear that immunoblot analysis could not provide an adequate comparison between the levels of Spo0A between the two strains. Moreover, Spo0A-yellow fluorescent protein (YFP) accumulates to high levels in the mother cell, following asymmetric division, further complicating the comparison between cells entering sporulation in the two strains (21). We therefore turned to fluorescence microscopy to quantify expression of a P_spo0A-yfp fusion in cells of BSP1 and PY79 (AH7325 and AH7326, respectively) during growth and entry into the stationary phase in DSM. Samples were collected, labeled with the membrane dye FM4-64, and mixed in equal proportions after additional staining of the AH7326 (the PY79 derivative bearing the P_spo0A-yfp fusion) sample with the DNA stain DAPI. This allowed the identification of cells in which asymmetric division had not yet occurred and the quantification of the YFP signal in those cells under the same viewing conditions. Since Spo0A is autoregulatory, the fusion reports both transcription of spo0A as well as the activity of Spo0A (21). Vegetative cells of the BSP1 derivative show stronger P_spo0A-yfp expression at all times examined compared with cells of the PY79 derivative (Fig. 3B). To compare the distributions of the YFP signal between the two strains, we defined an arbitrary threshold value above which, 1 h before the onset of the stationary phase (T₋₁), no PY79 cells could be found. In contrast, 44.8% of BSP1 cells were above this value (Fig. 3C). At the onset of the stationary phase (T₀), the percentages of cells above the threshold increased to 96.5% for BSP1 and 24.8% for PY79, while at T₁, 99.1% of the BSP1 cells and 65.1% of the PY79 cells were above the threshold (Fig. 3C). The distribution of the YFP signal among the population of vegetative cells did not reveal a bistable pattern but rather was heterogeneous, in line with the conclusion of recent studies (42, 43). Also importantly, no differences were seen in the coding sequence or the regulatory region of the spo0A gene between BSP1 and PY79 (not shown). Since BSP1 shows a high level of sporulation during growth, and because the distribution of the YFP signal for PY79 when it enters sporulation (at T₁) resembles that of BSP1 during growth (T₋₁), it is likely that the cells above the arbitrary threshold have levels of active Spo0A sufficient to activate transcription of the high-threshold genes required for sporulation (21, 24). In agreement with this inference, the fraction of BSP1 cells entering the stationary phase and expressing a fusion of the Spo0A-dependent spoIIG promoter to cfp (P_spoIIG-cfp) was higher than that for PY79 cells (e.g., over 50% of the population for BSP1 at T₀ but only about 15% for PY79) (Fig. 4). Thus, Spo0A reaches levels sufficient to trigger sporulation in a larger fraction of the BSP1 population reaching the stationary phase.

FIG 4.

Increased expression of P_spoIIG-cfp during growth of BSP1. (A) Expression of a Spo0A-dependent fusion of the P_spoIIG promoter to CFP (P_spoIIG-cfp) in BSP1 or PY79 derivatives carrying a P_spo0A-yfp fusion (AH7348 and AH7349, respectively), at the onset of the stationary phase (T₀) in DSM. The cells were stained with DAPI and FM4-64 and analyzed by phase-contrast (PC) and fluorescence microscopy. Scale bar, 2 μm. (B) Quantification of the number of BSP1 and PY79 cells showing expression of P_spoIIG-CFP in the experiment illustrated in panel A, which was extended for 5 h after T₀.

The activity of Spo0A is increased per BSP1 cell.

We then wanted to determine whether the increased activity of Spo0A in the BSP1 strain could be attributed to augmented transcription of spo0A, which relies in part on the activation of Spo0A, as Spo0A is autoregulatory (21), or to increased activation of Spo0A through the phosphorelay. To eliminate the autoregulatory loop acting at the level of spo0A transcription, we replaced the wild-type copy of the gene with a xylose-inducible P_xylA-spo0A allele in BSP1 and PY79 derivatives additionally carrying a P_spoIIG-cfp fusion. We then monitored the accumulation of Spo0A and the level of sporulation as a function of the xylose concentration. The results in Fig. 5A show that for PY79, Spo0A accumulated linearly with the xylose concentration in the range of 0 to 0.005%, whereas for BSP1, Spo0A was only detected at xylose concentrations ≥0.0001% and accumulated linearly with xylose concentrations up to 0.001%. Control experiments, in which we have used a P_xylA-gfp fusion, have shown that the responses of the P_xylA promoter to xylose are very similar in both BSP1 and PY79: not only GFP showed a parallel accumulation for each concentration of xylose tested for the two strains (albeit at slightly lower levels for BSP1), but the distribution of the fluorescence signal across the population was also very similar (see Fig. S2 in the supplemental material). Therefore, it appears that the accumulation of Spo0A is somehow limited in the background of BSP1 (see also below). Importantly, for PY79, 0.0001% was the lowest concentration of xylose for which a titer of heat-resistant spores higher than 10⁶/ml was detected (Fig. 5A). For BSP1, and in sharp contrast, a titer higher than 10⁶spores/ml was detected even in the absence of xylose (Fig. 5A). Clearly, for a xylose concentration of 0.0005%, both strains show a similar spore titer, but only trace amounts of Spo0A could be detected in the BSP1 derivative (Fig. 5A, red bars), while considerably more Spo0A was detected in the PY79 derivative. Microscopic examination of BSP1 and PY79 cultures 5 h after induction with 0.0005% xylose confirmed expression of the P_spoIIG-cfp reporter in both strains and the presence of phase-bright spores in BSP1 (Fig. 5B). The results suggest that Spo0A is more active per cell of the BSP1 strain. Furthermore, and in line with the conclusion of recent studies (42, 43), our results suggest that the activation of Spo0A and not its enhanced production through the positive autoregulatory loop that acts at the level of spo0A transcription is the foremost factor for entry into sporulation. Also in agreement with this conclusion, inactivation of the P_s promoter of the spo0A gene caused a 200-fold reduction in the final titer of spores (measured 24 h after the onset of the stationary phase in DSM) formed by PY79 but only a 5-fold reduction in strain BSP1 (see Fig. S3 in the supplemental material).

FIG 5.

Spo0A is more active per BSP1 cell. (A) Derivatives of BSP1 (AH7446) and PY79 (AH7445) bearing a P_xylA-spo0A fusion were grown in DSM to an OD₆₀₀ of about 0.9, at which point xylose was added to the indicated concentrations. Samples were collected 2 h after addition of xylose and processed for immunoblot analysis with anti-Spo0A and anti-σ^A antibodies (as a loading control). Reference samples were also prepared from the parental strain BSP1 or PY79 (bearing no P_xylA-spo0A fusion [NF]) 30 min after an OD₆₀₀ of ≈0.9 was attained. The bottom part of the panel shows the titer of heat-resistant spores for each of the cultures analyzed in panel A, measured 24 h after addition of xylose. The red bars indicate the results obtained with 0.005% xylose. (B) Expression of a P_spoIIG-cfp in BSP1 or PY79 (NF, no fusion) and in congenic strains bearing P_xylA-spo0A or P_xylA-sad67 fusions, as indicated, when grown in DSM to which 0.005% xylose was added at an OD₆₀₀ of ≈0.9. Samples were examined by phase-contrast (PC) and fluorescence microscopy 5 h after addition of xylose. Scale bars, 2 μm. (C) Immunoblot analysis of the kinetics of Spo0A and σ^A accumulation in extracts from the same strains shown in panel A, following addition of xylose to a final concentration of 0.0005%. Samples were collected and analyzed at the indicated times (in minutes). (D) Immunoblot analysis of Spo0A and σ^A accumulation in extracts from the same strains shown in panel B, collected and prepared 1, 3, and 5 h following addition of xylose to 0.005%. The parental strains, BSP1 or PY79, bearing no P_xylA-spo0A fusion (NF) were included in the analysis. The positions of Spo0A or σ^A are indicated by arrows in panels A, B, and D.

Evidence for a pathway that limits the activity of Spo0A.

In the experiments described above, extracts were prepared for immunoblot analysis 2 h after induction with xylose. Therefore, it seemed possible that Spo0A initially accumulated to high levels in BSP1, triggering sporulation, its levels then rapidly falling. To test this possibility, we monitored the kinetics of accumulation of Spo0A after induction of cultures with 0.0005% xylose. For PY79, the level of Spo0A increased with time until 180 min after induction, reaching a plateau thereafter (Fig. 5C). In BSP1, the level of Spo0A increased with time (but always at lower levels than with PY79), reaching a maximum 120 min following induction, after which its levels receded (Fig. 5C). Because Spo0A is more active per BSP1 cell, the results suggest that a mechanism is in place to limit the accumulation of active Spo0A.

As an independent test of this idea, we replaced the wild-type spo0A gene, in both PY79 and in BSP1, with a fusion of the P_xylA promoter to the sad67 allele, which codes for a constitutively active form of Spo0A. The BSP1 and PY79 derivatives bearing the P_xylA-sad67 construct additionally carried a P_spoIIG-cfp fusion, so that transcription of an Spo0A-dependent sporulation gene could be monitored by microscopy in relation to spore morphogenesis. The accumulation of Spo0A was monitored by immunoblotting 1, 3, and 5 h after induction of BSP1 and PY79 cultures expressing either P_xylA-spo0A or P_xylA-sad67 with 0.005% xylose. For reference, the accumulation of Spo0A was inspected in cultures of BSP1 and PY79, to which xylose was added to the same concentration (0.005%). Spo0A was detected at much lower levels in BSP1 than in PY79 (Fig. 5D). Moreover, while in PY79 Spo0A accumulated to similar levels throughout the experiment, in BSP1 it reached a maximum 1 h after induction, and then its level decreased (Fig. 5D). No difference was observed in the accumulation of Spo0A^sad67, compared to wild-type Spo0A, in PY79. In sharp contrast, in BSP1, Spo0A^sad67 was only detected 5 h after induction, and only in trace amounts (Fig. 5D). Induction of wild-type Spo0A production resulted in expression of P_spoIIG-cfp in both PY79 and BSP1 at h 5 and efficient spore formation in the latter strain. Strikingly, and in contrast to PY79, induction of Spo0A^sad67 in BSP1 suppressed spoIIG-cfp expression and spore formation in BSP1 (Fig. 5B). Consistent with the lack of expression of the high-threshold spoIIG operon, no asymmetric septa, the formation of which requires high levels of Spo0A, were seen in the sad67 derivative of BSP1 (not shown). (Note that this strain, however, will reach a titer of spores identical to that of the PY79 derivative bearing the sad67 allele after 24 h of incubation.) Together, these results suggest that in BSP1, increasing the activity of Spo0A above a certain threshold level triggers a pathway that leads to elimination of the transcription factor.

Involvement of the phosphorelay in sporulation of BSP1.

The unusual sporulation properties of strain BSP1—in particular, the observation that Spo0A reached sufficient levels of activity to promote sporulation during growth—prompted us to test whether genomic variation could be detected in the genes coding for the components of the phosphorelay. Using microarray-based comparative genomic hybridization (M-CGH) (12), we found that spo0A, spo0B, and spo0F, as well as the genes (kinA to -E) coding for the five histidine kinases controlling the initiation of sporulation, were present and with no detectable variability in the genome of BSP1 compared to strain 168 (see Table S4 in the supplemental material). Of note is the divergence found for the lrpA and lrpB genes, which function in a KinB stimulatory pathway (49). We then tested whether activation of Spo0A still required the phosphorelay. KinA and KinB are the main kinases involved in the initiation of sporulation (42, 50, 51). Mutations in kinA completely abolished spore formation at T₀ for PY79 (from 10⁴ spores/ml) but not for BSP1 (from 4 × 10⁵ to 5 × 10² spores/ml), whereas deletion of kinB caused only a modest 3-fold reduction in the two strains (Fig. 6). No other single mutant had an appreciable effect on spore formation, in either PY79 or BSP1. While a kinA kinB double mutant of PY79 did not form spores (even when measured 24 h after the initiation of sporulation), BSP1 still formed 10⁴ spores/ml at this time (Fig. 6). Therefore, while KinA retains its central role in the initiation of sporulation in BSP1, the results suggest that at least one other kinase is responsible for the residual sporulation of the kinA kinB double mutant. This residual sporulation appears to be the responsibility of KinC, because inactivation of kinC in a kinA kinB double mutant (but not the inactivation of kinD or kinE) caused a 100-fold reduction in the spore titer measured at h 24 (Fig. 6). This conclusion is in agreement with the results of a previous study that implicated KinC in a residual level of sporulation observed for laboratory strains during growth in minimal medium (50). It is also in agreement with the finding that the overproduction of either KinA or KinC can markedly enhance the fraction of cells with polar septa, the formation of which requires the expression of high-threshold Spo0A-dependent genes (42). KinC and KinD are also known to be involved in a low level of phosphorylation of Spo0A during growth, sufficient to suppress the lytic cycle of bacteriophage ϕ29 (52).

FIG 6.

Effect of the phosphorelay kinases on sporulation by BSP1. The kinetics of heat-resistant spore formation was monitored for strains BSP1 (black) and PY79 (white) and derivatives of both strains bearing the indicated single or multiple deletions of the genes coding for the phosphorelay kinases (kinA, kinB, kinC, kinD, and kinE). The titer of heat-resistant spores was measured at the indicated times in hours after the onset of sporulation (T₀) in DSM. The dashed lines indicate the levels of sporulation for the parental strains at T₀. The results shown are the averages of results from three independent experiments with error bars representing the standard deviation.

The genes for three Rap phosphatases are divergent or absent in BSP1.

In contrast, to the genes for the sporulation kinases Spo0B, Spo0F, and Spo0A, the M-CGH analysis revealed that for the rapE, rapI, and rapK genes, the log₂ ratio of B. subtilis 168 fluorescence to BSP1 fluorescence was in all cases >1 (ranging from 3.1 to 3.6) (12), suggesting that these genes were either divergent or not present in BSP1 (see Table S4 in the supplemental material). These genes are located within or close to mobile genetic elements that may be absent from the genome of BSP1 (see Table S4 and the supplemental material for a more detailed description of these results). Also, PCR with primers designed on the basis of the genome sequence of strain 168 (see Fig. S4 in the supplemental material) failed to amplify the rapE, rapI, and rapK genes. The rapI-phrI pair is also missing from the genome of the laboratory strain PY79 but present in other commonly used strains, such as MB24 (data not shown). While the M-CGH and PCR data are only suggestive, the recent sequencing of the genome of strain BSP1 confirms the absence of the rapE, rapI, and rapK genes (see Discussion). Because at least some of the five Rap phosphatases are known to drain phosphate from Spo0A during growth (by dephosphorylating Spo0F) and are antagonized by cognate inhibitory peptides in a cell density-dependent manner (28–33), we hypothesized that the absence of rapE, rapI and rapK could be related to the ability of BSP1 to initiate sporulation during growth. To test this possibility, the rapE-phrE, rapI-phrI, and rapK-phrK operons were introduced in single copy, alone or in combination, at the nonessential amyE or thrC locus of BSP1, to produce strains carrying rapE-phrE (here designated rapE⁺ for simplification), rapK-phrK (here designated rapK⁺), or rapI-phrI (here designated rapI⁺), and the combinations rapEK⁺ and rapEI⁺ (Fig. 7A, top). We then examined the titer of heat-resistant spores in the various strains during entry into the stationary phase in a sporulation medium in comparison with BSP1. While the profile of spore formation did not change appreciably for the rapE⁺ and rapI⁺ strains (Fig. 7A, bottom), a 10-fold reduction in the titer of spores was found for the rapK⁺ and rapEI⁺ strains during growth (T₋₁), and a smaller reduction (6- to 7-fold) was found during entry into the stationary phase (T₀ and T₁) (Fig. 7A, bottom). In contrast, the rapEK⁺ insertion imposed a 100-fold reduction in the titer of spores from T₋₁ through T₁ (Fig. 7A). As expected, transplantation of the rap-phr operons into BSP1 reduced the fraction of cells expressing a P_spo0A-yfp fusion at the onset of the stationary phase (T₀) in DSM above an arbitrary value in which 82% of the BSP1 population could be included (Fig. 7B); it also reduced the fraction of cells with morphological signs of sporulation (Fig. 7C). These results suggest that the absence of the rap-phr operons, in particular that of the rapE-phrE/rapK-phrK pair, is sufficient to increase the ability of the BSP1 strain to initiate sporulation during growth.

The skf operon does not mediate lysis in BSP1.

Cannibalism, which is mediated by the production of a killing factor coded for by the skf and sdp operons, is a well-studied mechanism for delaying sporulation (22, 53). Both the skf and sdp operons belong to the low-threshold class of Spo0A-dependent genes, and skf also codes for a transporter that confers immunity to the Skf producer cells. Spo0A-inactive cells in which the skf operon is not transcribed are sensitive to the toxin and are killed, allowing the resistant cells to feed on the released nutrients and delay entry into sporulation (22, 23, 53). Microarray analysis (see Table S4 in the supplemental material) and PCR with skf- or sdp-specific primers (see Fig. S4 in the supplemental material) confirmed the presence of the skf operon in BSP1. The DNA sequence of a PCR fragment encompassing the entire operons revealed essentially no differences between PY79 and BSP1 at the protein level (see Fig. S5A in the supplemental material). However, and in agreement with a recent reannotation of the genome of strain 168 (54), we found that in both BSP1 and PY79, the skfC and skfD cistrons are fused into a single gene (see Fig. S5A). We have also confirmed that a P_skf-cfp fusion is expressed in BSP1 cells during entry into stationary phase in DSM (at T₁) (see Fig. S5B). skf-mediated killing is seen for the laboratory strain PY79 as a drop in the OD₆₀₀ of cultures entering the stationary phase of growth and by a corresponding decrease in the viable cell count (22). Neither of these manifestations was seen for BSP1, irrespective of whether the operon was deleted (see Fig. S5C and D). Lysis of BSP1 or a congenic derivative bearing a deletion of the skf operon was also investigated by fluorescence-activated cell sorting (FACS) following Live/Dead staining of cells entering the stationary phase (see Fig. S6A in the supplemental material). As previously reported (22), cell lysis was strongly dependent on an intact skf locus for PY79 (see Fig. S6B). For BSP1, however, lysis was independent of the presence of the skf operon (see Fig. S6B). The fluorescence microscopy of the Live/Dead-stained BSP1 cells suggested that part of the signal detected in the red channel by FACS analysis could be accounted for by the high proportion of the cells at a late stage of sporulation, when the mother cell undergoes lysis to release the mature spore (see Fig. S6A). In any case, the results suggest that the skf-mediated delay of entry into sporulation previously documented for PY79 does not operate in BSP1.

rap genes and the skf operon in other wild isolates of B. subtilis.

Strain BSP1 shows robust induction of sporulation in part because it lacks the genes for three of the Rap phosphatases that are thought to help delay sporulation until the stationary phase of growth. Moreover, BSP1 is immune to the delay imposed upon entry into sporulation by expression of the skf operon. These observations motivated us to inspect other wild isolates of B. subtilis for the presence or absence of rap genes and/or the skf cluster. We used PCR, with primers designed based on the sequence of the 168 strains, to test for the presence of the rap genes and of the skf operon in nine additional B. subtilis strains isolated, as for BSP1, from broiler chickens (8). In addition, six other B. subtilis isolates of ovine and bovine origins (our unpublished results) were included in the screen. In parallel, we examined a total of four laboratory strains, including PY79. While rapA was present in all strains tested, most of the strains lacked at least one rap gene. The rapA, rapB, rapE, and rapK genes could be amplified from the laboratory strains PY79, NCBI 3610, JH642, and 168, but rapI was detected only in strain 168 (see Fig. S4 in the supplemental material). Because rapI is known to be present in JH642 and in NCBI 3610 (see Fig. S4), our PCR assay generates false negatives. Nevertheless, the spore titer for poultry strain 210, which lacked the rapB, rapE, rapI, and rapK genes, at h 4 of the stationary phase in DSM was above 10⁸ spores/ml (1.1 × 10⁸ spores/ml), similar to BSP1 (2.1 × 10⁸ spores/ml) (see Fig. S4). In contrast, poultry strain 278, which lacked only rapI, formed 4.1 × 10⁶ spores/ml at h 4, whereas another poultry strain, 285, which lacked both rapE and rapI, showed a somewhat intermediate titer (2 × 10⁷ spores/ml) (see Fig. S4). The titer of spores measured at h 4 of the stationary phase in DSM was between 1.4 × 10⁴ and 1.8 × 10⁵ for all laboratory strains tested (see Fig. S4). The absence of rapE and phrE has been noticed in other isolates of B. subtilis (55). The rapE-phrE operon is located within the skin element, which is also not strictly conserved in strains of B. subtilis (56, 57). Our analysis suggests a correlation between enhanced sporulation and the absence of certain Rap phosphatases. It also illustrates the plasticity of the B. subtilis genome with respect to the presence or absence of rap genes. In fact, this plasticity also extends to the skf operon, which seemed absent (or was divergent) from one of the strains of ovine or bovine origin, and from seven of the nine additionally tested strains derived from broiler chickens (see Fig. S4).

DISCUSSION

Here, we have analyzed entry into sporulation in a poultry GI tract isolate of B. subtilis, BSP1 (8). In this strain, as also observed for B. subtilis strains isolated from the GI tract of other animals (this work) and from the human GI tract (11, 58), a high titer of spores is reached much earlier after the onset of the stationary phase in a medium that supports sporulation than for standard laboratory strains. We show here that for BSP1, and presumably also for the other GI tract B. subtilis strains, spore differentiation is triggered during growth because a significant fraction of the population maintains high levels of Spo0A∼P and expresses high-threshold genes required for sporulation. We also showed that Spo0A was more active in BSP1 cells. M-CGH studies as well as PCR suggested that the operons coding for the RapE, RapI, and RapK phosphatases are absent from BSP1, and the recent sequencing of the BSP1 genome confirms this inference (59). The absence of RapE, RapI, and RapK appears to be the primary cause of the increased sporulation of BSP1, because reintroduction of combinations of the missing rap-phr operons into BSP1 lowers the levels of Spo0A∼P, the fraction of cells expressing high-threshold Spo0A-dependent genes, and the titer of spores during the initial hours of the stationary phase in DSM.

The rapE-phrE/rapK-phrK grouping caused the highest reduction in sporulation. Previous studies have indicated that RapE has an accessory role in sporulation (55). However, the effect of deletion of rapE on the sporulation titer was assessed after overnight growth, when, even for BSP1, the differences compared to a laboratory strain are not as evident as during the earlier times of sporulation. In contrast, the rapK-phrK operon, together with the rapC-phrC and rapF-phrF pairs, has been implicated in the expression of ComA-dependent genes (60) but not in sporulation. Likewise, the rapI-phrI system, which regulates mobilization of the integrative and conjugative element Bs1, has not been connected to sporulation (ICEBs1) (61). Presumably, the role of these operons in sporulation is restricted to a specific niche or genomic context and does not emerge under laboratory conditions for the commonly used strains. In any event, because reintroduction of the rap operons in BSP1 does not reproduce exactly the kinetics of sporulation of the laboratory strain PY79, it is likely that other factors contribute to the initiation of sporulation during growth. For example, while maintaining their central role in the initiation of sporulation in the two strains, BSP1, but not PY79, still forms spores in the absence of KinA and KinB (Fig. 6), suggesting other kinases assume a comparatively more important role in BSP1.

Our results suggest that in BSP1, the phosphorelay is tuned to trigger sporulation during growth. This contrasts with the behavior of laboratory strains, in which several adaptive responses are mounted during the transition to the stationary phase, and entry into sporulation is delayed. Cannibalism, for instance, is not manifested in strain BSP1. The expression of the operon gene skf, coding for the sporulation-killing factor, which mediates killing of Spo0A-inactive sister cells (22, 23, 53, 62), is detected in BSP1. However, Skf-mediated lysis is not perceptible, presumably because of the large fraction of cells in which Spo0A accumulates and that are Skf resistant. Initiation of sporulation during growth may confer an adaptive advantage in the gut ecosystem. There are examples of spore-forming bacteria in which sporulation is central to their survival in the gut ecosystem. Metabacterium polyspora, for instance, is a strict anaerobe that lives in close association with its coprophagous host, the guinea pig (63, 64). In this organism, formation of multiple endospores, which will also protect the organism outside its host, is the primary form of propagation (63, 64). M. polyspora produces multiple (up to nine) endospores per cell, in register with passage through the gastrointestinal tract of the host. Only mature spores survive passage through the mouth and stomach of the guinea pig. The spores will germinate in the intestine and will initiate the next round of sporulation, bypassing binary fission. Several species or morphotypes of Epulopiscium, an intestinal symbiont of the surgeonfish, are also known to use endospore formation as a form of survival and dispersion. For Epulopiscium-like organisms, endospore differentiation is closely coordinated with fish behavior and follows a circadian rhythm, which enhances dispersal to additional host fish (64, 65). Other intestinal symbionts of mammals, termed “segmented filamentous bacteria,” which are related to the Clostridia, may also produce multiple endospores (66–71). B. subtilis is known to complete its entire life cycle in the gut (11), and it is possible that also for this organism sporulation is important for survival and/or propagation in the gut. In keeping with this idea, it is significant that increased sporulation was also found for human GI tract isolates of B. subtilis (11, 58), as well as for other strains associated with the GI tract of various animals (this work), and that all of these strains seem to lack different combinations of rap-phr operons (see Fig. S4 in the supplemental material). It is tempting to suggest that the precise combination of rap-phr deletions matches the particular “sporulation needs” of a given niche. Perhaps significantly, the rap-phr operons are positioned close to or within mobile genetic elements (6, 72) with rapI-phrI on the ICEBs1 element and rapE-phrE found within the 48-kbp skin (σ^K-intervening) element. The skin element is known to be absent from several strains of B. subtilis (12, 55) (see Fig. S4) and may also be absent from BSP1 (see Table S4 in the supplemental material). The M-CGH results suggests that prophages in the vicinity of rapI-phrI and rapK-phrK may be absent from BSP1 (see Table S4), and these, along with several prophage genes, are either absent or divergent among a group of 18 strains that were compared to B. subtilis 168 (12). It thus seems plausible that imprecise prophage excision together with the mobility of other genetic elements may have caused loss of the rap-phr operons.

It is not known whether the phenomenon of sporulation initiation during growth is specific to GI tract isolates of B. subtilis or is a widespread feature of wild strains. However, it seems most likely that the efficiency and kinetics of sporulation is not always increased but rather adjusted to specific niches. One wild isolate of B. subtilis, strain NCIB3610, bears a stop codon within the kinA gene, and is conditional for sporulation (73). In MSgg medium, NCIB3610 forms aerial structures, called fruiting bodies, that are preferential sites of spore formation (74), but it sporulates inefficiently in DSM (73). Spore differentiation is ecologically beneficial, as spores will remain viable under harsh environmental conditions, but it is a time-consuming and energetically costly process, requiring the differential expression of over 10% of the genome over a period of 7 to 8 h (75, 76). The propagation of B. subtilis in the laboratory in the absence of selection for sporulation results in a reduction in the ability to sporulate (75, 76). Thus, it may be that sporulation, like other features of wild isolates of B. subtilis, including swarming motility (77, 78), poly-γ-glutamate synthesis (79), production of antibiotics (80, 81), or the formation of robust biofilms (74, 82), was attenuated during domestication of the organism, either because it has become neutral with respect to fitness or because selection favored its loss (75, 76). The M-CGH analysis showed that about 7% of the predicted coding sequences of strain 168 were absent or highly divergent in BSP1 (see Table S4 in the supplemental material), a number in the same range as that obtained for other B. subtilis subsp. subtilis strains that have been studied in this manner (3, 12). The M-CGH analysis does not detect genes that may be specific to BSP1 (absent from strain 168), and thus the overall genomic diversity between the two strains is likely to be higher. The divergence detected in genes such as cotV, coding for a spore surface protein (83–85) or in the lrp genes, which function in a KinB stimulatory pathway (49), may also enhance the value of sporulation for BSP1 in vivo. It will be interesting to investigate whether this is the case and whether the profile of genomic variability seen for BSP1 is specific to GI tract isolates of B. subtilis.

The activation of Spo0A occurs in a heterogeneous manner, because the phosphorelay acts as a noise generator (40–43) (Fig. 1B). Noise creates asynchrony in the time of entry into sporulation, as not all the cells will reach the required high-threshold concentration of Spo0A∼P. Hence, the flow of phosphate through the relay is the limiting factor for entry into sporulation (Fig. 1B). Importantly, increasing the flow of phosphate through the phosphorelay decreases asynchrony and increases the fraction of cells that form spores (42). Our results are in agreement with these ideas. Not only is Spo0A more active per BSP1 cell, but a larger fraction of the population is above the threshold for entry into sporulation during growth, both features arising from increased flow of phosphate through the relay. It has been suggested that heterogeneity in the activation of Spo0A is advantageous because not all of the cells will commit to sporulation in a possibly changing environment (42). The sharp contrast between a laboratory strain and BSP1 in this regard again suggests that in the gastrointestinal tract, massive, synchronized sporulation is favored. However, an alternative view is that the conditions in the gut environment never promote efficient sporulation by laboratory strains and that BSP1 (and possibly other gut isolates) induces sporulation during growth to ensure that spores are formed.

Finally, our results also suggest that above a certain level of activity, a pathway is activated that leads to proteolysis of Spo0A. The biological significance may be that growth is compromised above a certain level of sporulation. Unraveling the components of this pathway will be a goal of future work.