Expression of the σ^F-Directed csfB Locus Prevents Premature Appearance of σ^G Activity during Sporulation of Bacillus subtilis

Abstract

During sporulation, σ^G becomes active in the prespore upon the completion of engulfment. We show that the inactivation of the σ^F-directed csfB locus resulted in premature activation of σ^G. CsfB exerted control distinct from but overlapping with that exerted by LonA to prevent inappropriate σ^G activation. The artificial induction of csfB severely compromised spore formation.

Spore formation in Bacillus subtilis has become a paradigm for studying cell differentiation in prokaryotes. Central to sporulation is the compartmentalization of gene expression between the smaller prespore and the larger mother cell that result from a division, which occurs soon after the start of spore formation. The prespore (or forespore) develops into the mature, heat-resistant spore. The mother cell is necessary for spore formation but eventually lyses. The expression of different genes in the two cells is governed in part by the activation of four RNA polymerase sigma factors: σ^F and then σ^G in the prespore, and σ^E and then σ^K in the mother cell (reviewed in reference 7). The first sigma factor to become active, σ^F, does so soon after the formation of the spore septum, and its activation leads rapidly to the activation of σ^E in the mother cell. Development continues, with the mother cell engulfing the prespore. Upon the completion of engulfment, the prespore is entirely within the mother cell and so is no longer in direct contact with the medium. At this stage, σ^G becomes active in the prespore and σ^G in turn triggers the activation of σ^K in the mother cell.

Here, we explore the activation of σ^G, which is encoded by spoIIIG. The productive transcription of spoIIIG is directed by RNA polymerase containing σ^F and so is confined to the prespore (8). The transcription of spoIIIG is delayed compared to that of other σ^F-directed genes (9, 18) and requires a σ^E-directed signal from the mother cell (11). It also requires the expression of the σ^F-directed spoIIQ locus (20). Once active, σ^G directs the transcription of its own structural gene, setting up a positive-feedback loop for σ^G accumulation (7). Placing spoIIIG under a strong σ^F-directed promoter (17) or mutating the spoIIIG promoter (6) can result in σ^E-independent transcription of spoIIIG but not activation of σ^G. The activation of σ^G depends on the completion of engulfment of the prespore (19), which requires the activity of several σ^E-directed genes (1, 5). Other controls acting on σ^G, notably, the anti-sigma factor SpoIIAB, prevent it from becoming active in the mother cell (2). In the prespore, however, SpoIIAB becomes sequestered in a complex with SpoIIAA and so is thought not to block σ^G activity in that compartment (7).

Inactivation of csfB results in premature activation of σ^G.

We have recently shown that when spoIIIG was relocated to an origin-proximal locus and the translocation of the origin-distal 70% of the chromosome into the prespore was blocked by a spoIIIE mutation, σ^E activity was no longer required either for the transcription of spoIIIG directed by σ^F or for the activation of σ^G. Thus, σ^G became active even though σ^E activation was blocked by a spoIIR mutation, and it became active in the prespore after septum formation rather than after the completion of engulfment (3). In pursuing those studies, we have found that the inactivation of the csfB locus greatly enhanced σ^G activity; further, that activity was detected during vegetative growth (compare data for strains SL13229 [csfB] and SL13223 [csfB⁺] in Fig. 1) (strains used in this study are listed in Table 1). The expression of the reporter, P_sspA-lacZ, was blocked by inactivating spoIIIG in SL13229, confirming that the reporter indicated σ^G activity (data not shown).

FIG. 1.

Inactivation of csfB causes hyperactivation of σ^G in a strain with chromosome translocation blocked by a spoIIIE36 mutation and spoIIIG inserted in an origin-proximal locus. Bacteria were grown in modified Schaeffer's sporulation medium at 37°C (12). The activity of σ^G was assessed as specific β-galactosidase activity expressed from an sspA-lacZ transcriptional fusion. Specific activity is expressed as nanomoles of o-nitrophenyl-β-d-galactoside hydrolyzed per minute per milligram (dry weight) of bacteria. Strain SL13223, csfB⁺ (empty circles); SL13229, csfB (filled circles).

TABLE 1.

B. subtilis strains used

Straina	Relevant genotypeb	Source
SL10369	PsspA-lacZ@sspA	Lab stock
SL10969	PsspA-gfp@sspA	Lab stock
SL11801	lonA::tet PsspA-gfp@sspA	Lab stock
SL12657	spoIID::neo PsspA-lacZ@sspA	Lab stock
SL13223	spoIIIE36 amyE::spoIIIG spoIIR::PsspA-lacZ	Lab stock
SL13229	spoIIIE36 amyE::spoIIIG spoIIR::PsspA-lacZ csfBΔ::cat	This study
SL13315	lonA::tet csfBΔ::cat PsspA-gfp@sspA	This study
SL13359	thrC::Pspac(Hy)-csfB PsspA-lacZ@sspA	This study
SL13554	csfBΔ::cat PsspA-gfp@sspA	This study
SL13556	spoIIR::PsspA-lacZ	This study
SL13563	spoIIR::PsspA-lacZ csfBΔ::cat	This study
SL13564	spoIIR::PsspA-gfp csfBΔ::cat	This study
SL13767	csfBΔ::cat PsspA-lacZ@sspA	This study
SL13776	lonA::tet PsspA-lacZ@sspA	This study
SL13778	csfBΔ::cat spoIID::neo PsspA-lacZ@sspA	This study
SL13825	amyE::PspoIID-lacZ thrC::Pspac(Hy)-csfB	This study
SL13835	amyE::PspoIIQ-lacZ thrC::Pspac(Hy)-csfB	This study
SL13914	csfBΔ::cat lonA::erm PsspA-lacZ@sspA	This study
SL14127	amyE::PgerE-lacZ thrC::Pspac(Hy)-csfB	This study
SL14244	spoIIR::PsspA-lacZ csfBΔ::cat spoIIIG::spc	This study
SL14263	csfBΔ::cat lonA::erm PsspA-lacZ@sspA spoIIIG::spc	This study

^aAll strains are in the genetic background of B. subtilis 168 strain BR151 (trpC2 lys-3 metB10). Details of strain construction are available upon request.

^b@ indicates that the fusion has been introduced by single-crossover (Campbell-like) recombination.

The csfB locus was identified as being controlled by sigma F (4). The deletion of csfB had no effect on the sporulation of a spo⁺ strain (4), and we presume that the dramatic effect we observed on σ^G activity depended on the particular mutant background used. A possible explanation of our result is that CsfB directly or indirectly prevents premature σ^G activation. We speculate that there is a basal level of σ^F-directed transcription during vegetative growth. This activity results in a low level of spoIIIG transcription, leading to the formation of some σ^G, but a low basal level of σ^F-directed csfB expression is sufficient to prevent σ^G activation. In the csfB mutant, however, a positive-feedback loop is established in which σ^G, now not inhibited by CsfB, directs spoIIIG transcription, leading to high σ^G activity.

The effects described above were observed using a particular genetic background. The effects of csfB inactivation were more modest when chromosome translocation was unimpaired and spoIIIG was at its natural locus. Nevertheless, csfB inactivation did accelerate the appearance of a low level of σ^G activity during spore formation (compare data for strains SL13767 and SL10369 in Fig. 2). This premature expression of σ^G activity was also observed with the inactivation of spoIIIA or spoIIQ (data not shown), loci whose expression is normally required for the activation of σ^G upon the completion of engulfment (7, 10, 20). σ^G became active independently of the completion of engulfment (compare data for the spoIID mutants SL13778 and SL12657 in Fig. 3) and of σ^E activity (compare data for the spoIIR mutants SL13563 and SL13556 in Fig. 3), although the ultimate level of σ^G activity did not reach that of a spo⁺ strain. In csfB spoIID and csfB spoIIR mutants, σ^G activity (monitored with a gfp fusion) was usually located in the preengulfment prespore (Table 2 and data not shown). The inactivation of spoIIIG abolished the expression of the P_sspA-directed reporters used, confirming that the reporters were monitoring σ^G activity (Fig. 3, strain SL14244, and data not shown).

FIG. 2.

Effect of csfB and lonA inactivation on σ^G activity in spo⁺ strains. Bacteria were grown in modified Schaeffer's sporulation medium at 37°C. The activity of σ^G was assessed as specific β-galactosidase activity expressed from an sspA-lacZ transcriptional fusion. Strain SL10369, parent strain (filled circles); SL13767, csfB (empty circles); SL13776, lonA (empty triangles); SL13914, csfB lonA (filled triangles); SL14263, csfB lonA spoIIIG (empty squares).

FIG. 3.

Inactivation of csfB renders σ^G activation independent of σ^E activity and of the completion of engulfment. Bacteria were grown in modified Schaeffer's sporulation medium at 37°C. The activity of σ^G was assessed as specific β-galactosidase activity expressed from sspA-lacZ transcriptional fusions. Strain SL12657, spoIID (filled circles); SL13778, spoIID csfB (empty circles); SL13556, spoIIR (filled triangles); SL13563, spoIIR csfB (empty triangles); SL14244, spoIIR csfB spoIIIG (empty squares).

TABLE 2.

Location of GFP expressed from a σ^G-directed promoter^a

Strain	Description or relevant genotype	Time (h)b	% Of cells displaying GFP fluorescence characterized asc:			Total no. of organisms scored
			PS	MC	WC
SL10969	Parent strain	4	0	0	0	50
		6	66	0	0	50
SL11801	lonA	0	0	0	3d	205
		2	0	0	8d	155
		4	1	0	10d	256
		6	32	0	4	143
SL13315	csfB lonA	0	0	0	8	225
		2	0	0	9	150
		4	23	0	23	120
		6	53	0	2	106
SL13554	csfB	0	0	0	2	220
		2	0	0	4	351
		4	20	0	1	103
		6	51e	0	0	156
SL13564	spoIIR csfB	2	15	0	5	92
		4	13	0	2	183
		6	36	0	3	75

^aMembranes were visualized by staining with FM4-64. σ^G activity was visualized using a P_sspA-gfp fusion.

^bTime after the end of exponential growth.

^cGFP, green fluorescent protein; PS, prespore specific; MC, mother cell specific; WC, whole-cell fluorescence.

^dWeak signal detected.

^eSome bacteria also displayed weak activity in the mother cell.

CsfB and LonA exert overlapping controls of σ^G activation.

There are now several examples of overlapping regulatory mechanisms acting during spore formation (reviewed in reference 7), and we tested to see if the regulation of σ^G by CsfB overlapped that by the LonA protease (14, 16). LonA has previously been shown to prevent inappropriate σ^G activation during stationary phase or during a stress response (14). The inactivation of lonA did not completely deregulate σ^G; indeed, it had little effect on σ^G activity under sporulation conditions and did not affect spore formation (14). Under our conditions, the inactivation of lonA also had little effect on σ^G activity (Fig. 2, strain SL13776). However, the inactivation of both csfB and lonA resulted in substantial σ^G activity extending from exponential growth through sporulation for a spo⁺ strain (SL13914) (Fig. 2), much greater than that resulting from the inactivation of only csfB (SL13767) or lonA. This hyperactivity in the csfB lonA mutant was abolished by the inactivation of spoIIIG (Fig. 2, strain SL14263). Thus, CsfB and LonA appear to exert different but overlapping controls to prevent premature σ^G activation.

The effects of lonA and csfB mutations on the localization of σ^G activity were visualized using an sspA-gfp transcriptional fusion. The expression of this fusion is normally detected only in the prespore after the completion of engulfment (Table 2, strain SL10969). Under the conditions used, engulfment was completed between 4 and 6 h and septation was completed between 2 and 4 h after the end of exponential growth. In different experiments, 30 to 70% of organisms displayed strong prespore-specific green fluorescent protein fluorescence at 6 h whereas no fluorescing organisms were observed in samples of SL10969 at 4 h or earlier. With the csfB mutant SL13554, 1% of bacteria in the 0- and 2-h samples displayed whole-cell fluorescence (Table 2). About 20% of organisms in the 4-h sample of SL13554 displayed fluorescence specific to the preengulfment prespores, consistent with the loss of CsfB's permitting premature activation of σ^G in the prespore. In the 6-h sample of SL13554, about 50% of organisms displayed fluorescence in the prespores, which were by then largely engulfed.

A small proportion of lonA mutant bacteria that had not completed engulfment (strain SL11801 in 0-, 2-, and 4-h samples) displayed whole-cell fluorescence (Table 2). This fluorescence was very weak, which may account for the failure to detect σ^G activity at those times using lacZ as a promoter probe (Fig. 2). In contrast, about 10% of the population of a lonA csfB double-mutant strain displayed strong fluorescence during exponential growth and at early times during spore formation (Table 2). Later in sporulation, the double mutant resembled the other strains in that strong fluorescence was detected in the engulfed prespores of about 50% of the population (Table 2, strain SL13315 at 6 h). These results are consistent with the idea that CsfB and LonA exert different but overlapping controls to prevent premature σ^G activation. Despite the effect on the compartmentalization of σ^G activity, the inactivation of both csfB and lonA had little effect on spore formation under the conditions used (data not shown); however, it is possible that the ∼10% of bacteria displaying premature, strong whole-cell σ^G activity were unable to form spores.

Artificial induction of csfB expression blocked spore formation.

The csfB locus was placed under the control of the strong, IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible P_spac(Hy) promoter (13) and integrated at the ectopic thrC locus. The P_spac(Hy)-csfB construct was introduced into spo⁺ strains containing lacZ under the control of σ^G, σ^F, σ^E, or σ^K. The presence of IPTG had no observable effect on growth but reduced spore formation about 1,000-fold compared to that in cultures without IPTG (Table 3). This effect was independent of lonA (data not shown). Examination by microscopy indicated that organisms completed engulfment efficiently in the presence of IPTG but did not go on to form phase-bright spores. The induction of csfB reduced σ^G activity by about 50% and delayed the appearance of activity by about 1 h compared to no induction (Fig. 4, left panel, strain SL13359). The induction of csfB also resulted in about 50% reduction of σ^E activity (Fig. 4, right panel, strain SL13825) but no reduction of σ^F activity (Fig. 4, right panel, strain SL13835). There was a reduction and delay in σ^K activity caused by csfB induction (Fig. 4, left panel, strain SL14127), which may be a secondary consequence of the effect on σ^G. It is not known if CsfB acts directly or indirectly on these sigma factors. CsfB is predicted to be a 64-residue protein (4). It shows no similarity to proteins of known function, and its mode of action is unknown. Given that csfB transcription is ordinarily directed by σ^F, so that csfB is expressed predominantly in the prespore, the effect of CsfB on σ^E (and possibly σ^K) may represent a fail-safe mechanism to help prevent σ^E from becoming active in the prespore. The observed effects on reporters of σ^E, σ^G, and σ^K activities caused by csfB induction (Fig. 4) seem inadequate to explain the dramatic effect on spore formation. However, it may be that CsfB has a strong effect on particular promoters, leading to the block in spore formation.

TABLE 3.

Effect of csfB induction on spore formation

Strain	Description or relevant genotypea	% Of parent level of sporulation in the absence of IPTG as measured:
		Without IPTG	With IPTG
BR151	Parent strain	100b	94
SL13359	Pspac(Hy)-csfB PsspA-lacZ@sspA	58	0.08
SL13825	Pspac(Hy)-csfB amyE::PspoIID-lacZ	68	0.05
SL13835	Pspac(Hy)-csfB amyE::PspoIIQ-lacZ	84	0.04
SL14127	Pspac(Hy)-csfB amyE::PgerE-lacZ	59	0.06

^a@ indicates that the fusion has been introduced by single-crossover (Campbell-like) recombination.

^bEquivalent to 2.1 × 10⁸ heat-resistant spores/ml.

FIG. 4.

Effect of induction of csfB expression from the P_spac(Hy) promoter on the activities of sporulation-specific σ factors, assessed with lacZ transcriptional fusions directed by the different σ factors. Bacteria were grown in modified Schaeffer's sporulation medium at 37°C without IPTG (empty symbols) or with 1 mM IPTG throughout growth and sporulation (filled symbols). (Left panel) σ^G, strain SL13359 (circles); σ^K, strain SL14127 (triangles); (right panel) σ^F, strain SL13835 (circles); σ^E, strain SL13825 (triangles).

CsfB is one of several entities that regulate σ^G activation. The main role for CsfB during spore formation is probably to help prevent premature activation of σ^G in the prespore. In the absence of CsfB, σ^G becomes active after the formation of the sporulation septum rather than after the completion of engulfment (Table 2). This premature activation is independent of the SpoIIIA-SpoIIIJ pathway; CsfB does not appear to be involved in that pathway, which is required for the more substantial activation of σ^G that normally occurs in the prespore following the completion of engulfment (7, 10, 15). CsfB is also not required for the repression of σ^G by SpoIIAB in the mother cell (data not shown). CsfB has a role overlapping with (though distinct from) that of LonA (14) in regulating σ^G, as the loss of both greatly enhanced premature σ^G activity (Fig. 3). Despite this expanding list of regulators, our understanding of σ^G regulation remains incomplete.