An A257V Mutation in the Bacillus subtilis Response Regulator Spo0A Prevents Regulated Expression of Promoters with Low-Consensus Binding Sites

Abstract

In Bacillus species, the master regulator of sporulation is Spo0A. Spo0A functions by both activating and repressing transcription initiation from target promoters that contain 0A boxes, the binding sites for Spo0A. Several classes of spo0A mutants have been isolated, and the molecular basis for their phenotypes has been determined. However, the molecular basis of the Spo0A(A257V) substitution, representative of an unusual phenotypic class, is not understood. Spo0A(A257V) is unusual in that it abolishes sporulation; in vivo, it fails to activate transcription from key stage II promoters yet retains the ability to repress the abrB promoter. To determine how Spo0A(A257V) retains the ability to repress but not stimulate transcription, we performed a series of in vitro and in vivo assays. We found unexpectedly that the mutant protein both stimulated transcription from the spoIIG promoter and repressed transcription from the abrB promoter, albeit twofold less than the wild type. A DNA binding analysis of Spo0A(A257V) showed that the mutant protein was less able to tolerate alterations in the sequence and arrangement of its DNA binding sites than the wild-type protein. In addition, we found that Spo0A(A257V) could stimulate transcription of a mutant spoIIG promoter in vivo in which low-consensus binding sites were replaced by high-consensus binding sites. We conclude that Spo0A(A257V) is able to bind to and regulate the expression of only genes whose promoters contain high-consensus binding sites and that this effect is sufficient to explain the observed sporulation defect.

Bacillus subtilis cells will sporulate in response to high cell density and nutrient deprivation following exhaustion of alternate survival strategies (12, 43). The master regulator controlling the onset of sporulation is Spo0A. Spo0A lies at the terminus of a phosphorelay signal transduction system that integrates metabolic and environmental information (reviewed in references 16, 19, and 45). When activated by phosphorylation, Spo0A initiates a genetic program that divides the cell asymmetrically and builds the endospore (reviewed in references 12, 17, 32, and 43). Microarray experiments suggest that Spo0A regulates the expression of over 120 genes directly (28) and over 500 genes in total, nearly 12% of the B. subtilis genome (10).

Spo0A∼P (the activated phosphorylated form of Spo0A) regulates transcription by binding to specific DNA sequences termed “0A boxes” via its C-terminal domain (44, 46). At promoters activated by Spo0A, binding of the protein to 0A boxes 5′ of the transcription start site increases the rate of transcription initiation (5, 34). At promoters repressed by Spo0A, binding of the protein to 0A boxes 3′ of the transcription start site eliminates transcription (44, 46).

Spo0A is a response regulator, and like other response regulators, its activity is modulated by phosphorylation, which drives a conserved rearrangement in a domain characteristic of this class of proteins, the receiver domain (3, 20, 22, 23). It has been suggested that N-terminal interactions after phosphorylation drive dimerization of the N-terminal domain (25). In addition to crystal structures of the phosphorylated and unphosphorylated N-terminal receiver domain of Spo0A, structures of the C-terminal DNA binding domain have also been determined (24, 47). The DNA binding domain is composed of six α-helices joined by short segments of polypeptide. It contains a helix-turn-helix motif (αC and αD) responsible for sequence-specific DNA binding and an α-helix (αE) that has been shown through genetic analysis to be required for the activation of σ^A-dependent promoters (7, 18). The crystal structures of two DNA binding domains in complex with tandem 0A boxes (from the abrB gene) revealed that Spo0A binds DNA as a head-to-tail dimer (47).

Spo0A activates transcription by two distinct mechanisms. At the spoIIA promoter, Spo0A interacts with RNA polymerase (RNAP) containing σ^H, while at the spoIIG promoter, Spo0A interacts with σ^A (1, 7, 18, 21, 38). At spoIIG, Spo0A compensates for overlong spacing between the conserved promoter elements to which RNAP binds (27, 39). The 0A boxes overlap the −35 element at spoIIG, and sequence-specific interactions between Spo0A and σ^A position and stabilize RNAP, enabling it to initiate transcription efficiently (40). At spoIIA, the 0A boxes are found as single sites or multiples rather than tandem pairs, and in the opposite orientation to those found at the spoIIG promoter. This would imply that Spo0A presents a different surface to RNAP at σ^H-dependent promoters than at σ^A-dependent promoters, one that includes the C-terminal helix, αF.

A large number of mutations within Spo0A have been identified. A mutant that contains a substitution of valine for alanine at position 257, Spo0A(A257V), within αF of the C-terminal domain is particularly unusual (31). Cells containing this mutant do not activate transcription at either spoIIA or spoIIG yet proficiently repress the abrB promoter (31, 33). These data suggest that Spo0A(A257V) binds DNA normally but is somehow defective in stimulating transcription. In addition to possibly acting as a contact surface for σ^H, the C-terminal domain crystal structure suggests that the A257V mutation could indirectly weaken intermolecular contacts between Spo0A monomers bound to tandem 0A boxes (47). However, this does not correlate easily with the observation that cells containing spo0A(A257V) repress abrB normally. Even the requirement that this intermolecular interface be conserved is not entirely clear, since 0A boxes presumed to be involved in transcription regulation do not always occur in pairs, such as at the Spo0A-activated spoIIA and the sporulation-specific spo0A promoter (44), or with the same spacing as the tandem 0A boxes observed at the best-understood spoIIG and abrB promoters.

Since we are interested in how Spo0A functions, we have examined transcriptional regulation by wild-type Spo0A and Spo0A(A257V) both in vivo and in vitro. The results provide an explanation for how Spo0A(A257V) fails to activate transcription from key stage II promoters yet retains the ability to repress the abrB promoter, but they raise additional questions about how Spo0A regulates transcription.

MATERIALS AND METHODS

Bacterial strains and media.

The bacterial strains, plasmids, and oligonucleotides used in this study are listed in Tables 1, 2, and 3. Escherichia coli cultures were grown in Luria Bertani (LB) medium supplemented with 100 μg/ml ampicillin. B. subtilis cultures were grown in LB (35) or Schaeffer's sporulation medium (SSM) (37) and transformed as described previously (31) The antibiotics used for selection in Bacillus when necessary were chloramphenicol (5 μg/ml) and kanamycin (5 μg/ml). Standard genetic techniques, enzymatic reactions, and DNA manipulations were performed as described previously (35) or as recommended by the manufacturer.

TABLE 1.

Bacterial strains used in this study

Strain	Description	Source/Reference
E. coli
DH5α	Cloning strain	Invitrogen
BL21(λDE3)(pLysS)	IPTGa-inducible protein over expression strain	Novagen
MC0A	Spo0A overexpression	8
BTA9V	Spo0A(A257V) overexpression	This study
B. subtilis
JH642	trpC2 phe-1	J.A. Hoch, Scripps Research Institute, La Jolla, Cal
JH695	trpC2 phe-1 spo0A(A257V)	11
GBS101	crsA47 trpC2 phe-1 amyE::(spoIIG-lacZ Kmr)	9
GBS106	crsA47 trpC2 phe-1 amyE::(spoIIA-lacZ Cmr)	9
GBS1011	JH695 crsA47 trpC2 phe-1 amyE::(spoIIG-lacZ Kmr)	This study
GBS1061	JH695 crsA47 trpC2 phe-1 amyE::(spoIIA-lacZ Cmr)	This study
JH16304	JH642 amyE::(spoIIG1-lacZ) Cmr	31
SS642IIG1	JH642 amyE::(spoIIG1-lacZ) Cmr	This study
SS642IIG2	JH642 amyE::(spoIIG1-lacZ) Cmr	This study
SS642IIG3	JH642 amyE::(spoIIG1-lacZ) Cmr	This study
SS642IIG4	JH642 amyE::(spoIIG1-lacZ) Cmr	This study
SS642IIG5	JH642 amyE::(spoIIG1-lacZ) Cmr	This study
DR2001	JH16304 amyE::(spoIIG-lacZ) spo0A(A257V) Cmr	33
SS695IIG1	JH695 amyE::(spoIIG3-lacZ) Cmr	This study
SS695IIG2	JH695 amyE::(spoIIG3-lacZ) Cmr	This study
SS695IIG3	JH695 amyE::(spoIIG3-lacZ) Cmr	This study
SS695IIG4	JH695 amyE::(spoIIG3-lacZ) Cmr	This study
SS695IIG5	JH695 amyE::(spoIIG3-lacZ) Cmr	This study

^aIPTG, isopropyl-β-d-thiogalactopyranoside.

TABLE 2.

Plasmids used in this study

Plasmid	Description	Source/reference
pGEM-T	PCR cloning vector	Promega
pGEMA9V	pGEM-T spo0A(A257V) gene	This study
pGEMSpo0A	pGEM-T with spo0A sequence from RBS to E263	This study
pET16b	Protein expression vector	Novagen
pET16b0A	pET16b with an 848-bp NcoI-BamHI insert bearing the spo0A gene	8
pET16bA9V	pET16b with an 848-bp NcoI-BamHI insert bearing the spo0A(A257V) gene	This study
pUCIIGtrpA	pUC19 with spoIIG promoter and proximal transcript fused to trpA terminator	36
pJM5134	Cloning vector with abrB promoter	30
pUCIIGAtrpASS1	pUCIIGtrpA with a single mutation in site 2 0A boxes	This study
pUCIIGAtrpASS2	pUCIIGtrpA with two mutations in site 2 0A boxes	This study
pUCIIGAtrpASS3	pUCIIGtrpA with three mutations in site 2 0A boxes	This study
pUCIIGAtrpASS4	pUCIIGtrpA with four mutations in site 2 0A boxes	This study
pUCIIGAtrpASS5	pUCIIGtrpA with five mutations in site 2 0A boxes	This study
pDH32	amyE integrative vector for lacZ fusions	42
pDH32SSIIG1	pDH32 containing an EcoRI-BamHI fragment from pUCIIGtrpASS1	This study
pDH32SSIIG2	pDH32 containing an EcoRI-BamHI fragment from pUCIIGtrpASS2	This study
pDH32SSIIG3	pDH32 containing an EcoRI-BamHI fragment from pUCIIGtrpASS3	This study
pDH32SSIIG4	pDH32 containing an EcoRI-BamHI fragment from pUCIIGtrpASS4	This study
pDH32SSIIG5	pDH32 containing an EcoRI-BamHI fragment from pUCIIGtrpASS5	This study

TABLE 3.

Oligonucleotides used in this study

Name	Sequencea
0A-4	CGGGATCCAAAGACGTTTGAT
0A-5	CGCCATGGAGAAAATTAAAGTTTGTGTTG
abrB-F	AAGGATTTTGTCGAATAATGACGAA
abrB-R	TCTTCGTCATTATTCGACAAAATCC
EcoRIF	ATGATTGAATTCAAGCTTCCTCGACAA
BamHIR	ATGATTGGATTCCTTCTTGCTTCATAA
IIG1F	TCTCAACATTATTTGACAGACTTTCCC
IIG1R	AAAGTCTGTCAAATAATGTTGAGAGGA
IIG2F	TCTCAACATTATTTGACAAACTTTCCCACA
IIG2R	TGTGGGAAAGTTTGTCAAATAATGTTGAGAGGA
IIG3F	GTATTTTCCTCTCGACATTATTTGACAAACTTTCCCACA
IIG3R	TGTGGGAAAGTTTGTCAAATAATGTCGAGAGGAAAATACAAT
IIG4F	GTATTTTCCTTTCGACATTATTTGACAAACTTTCCCACA
IIG4R	TGTGGGAAAGTTTGTCAAATAATGTCGAAAGGAAAATACAAT
IIG5F	GTATTTTCCTTTCGACATTATTCGACAAACTTTCCCACA
IIG5R	TGTGGGAAAGTTTGTCGAATAATGTCGAAAGGAAAATACAAT
AbrB+1	AAGGATTTTGTCGAATTAATGACGAAGA
AbrB+2	AAGGATTTTGTCGAATATAATGACGAAGA
AbrB+3	AAGGATTTTGTCGAATAATAATGACGAAGA
AbrB-1	AAGGATTTTGTCGAAAATGACGAAGA
AbrB-2	AAGGATTTTGTCGAAATGACGAAGA
AbrBΔR	TCTTCGTCATT

^aOligonucleotides are listed 5′-3′.

Expression and purification of Spo0A(A257V).

Chromosomal DNA was isolated from B. subtilis strain JH695 (33) as described previously (31). The spo0A(A257V) gene was amplified using the upstream primer 0A-5 and the downstream primer 0A-4 (Table 3). The resulting 848-bp fragment was cloned, creating plasmid pGEMA9V, and the A257V substitution was confirmed by sequencing (Nucleic Acid Protein Sequencing Unit, University of British Columbia). An expression construct was created by subcloning the NcoI/BamHI fragment of pGEMA9V [including the full spo0A(A257V) sequence, native stop, and transcription terminator] into pET16b (Novagen) to create pET16bA9V. pET16bA9V was transformed into E. coli BL21(λDE3)(pLysS), creating strain BTA9V.

Spo0A and Spo0A(A257V) were expressed and purified from strains MC0A and BTA9V as described previously (41) up to and including heparin-agarose affinity purification. The eluate was dialyzed overnight at 4°C against 2 liters of buffer C (20 mM sodium phosphate [pH 8.0], 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride) containing 50 mM NaCl. Following dialysis, the protein was loaded directly onto a 30-ml DNA-cellulose column equilibrated with buffer C plus 50 mM NaCl and eluted with a 150-ml linear gradient of 50 to 850 mM NaCl in buffer C. Samples were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and those containing Spo0A or Spo0A(A257V) were pooled and concentrated. The concentrated protein was dialyzed overnight at 4°C into buffer C containing 150 mM NaCl, 0.1 mM dithiothreitol, and 30% glycerol. The purified Spo0A and Spo0A(A257V) were aliquoted and stored at −20°C. The final products were approximately 95 to 98% pure as determined by densitometry of silver-stained SDS-PAGE gels and were used directly for in vitro studies.

In vitro phosphorylation reactions.

Steady-state phosphorylation of Spo0A and Spo0A(A257V) was carried out as described previously (4, 6). The in vitro phosphorylation rates of Spo0A and Spo0A(A257V) were determined by incubating phosphorelay components (0.75 μM KinA, 0.1 μM Spo0B, 0.5 μM Spo0F, 25 μM ATP, and 25 μCi of [γ-³²P]ATP [6,000 Ci/mM; Amersham Biosciences]) in 1× transcription buffer at 25°C. Aliquots were removed at various times after the addition of 0.5 μM Spo0A or Spo0A(A257V), added to 5 μl of 2× SDS-PAGE buffer (100 mM Tris-Cl, pH 6.8, 4% SDS, 0.2% bromophenol blue, 20% glycerol, 2 mM β-mercaptoethanol), and separated on a 15% gel by SDS-PAGE. Radiolabeled phosphates were detected using a Molecular Dynamics PhosphorImager SI system and quantified using ImageQuant 5.2 software.

In vitro transcription reactions.

Standard transcription reactions were carried out as described previously (15). For kinetic assays, 4 nM spoIIG template DNA, RNA polymerase containing σ^A (RNAPσ^A), 0 to 1,200 nM Spo0A or Spo0A(A257V), and ATP and GTP were incubated, allowing RNAPσ^A to initiate transcription. After 0 to 90 s, the initiated complexes were provided with UTP and CTP to allow transcript elongation and with heparin to limit transcription to a single round. The reactions were terminated after 5 min by the addition of stop buffer (7 M urea, 0.1% bromophenol blue, and 0.1% xylene cyanol in 0.5× TBE). The transcripts were separated by electrophoresis through an 8% denaturing polyacrylamide gel and detected using a Molecular Dynamics PhosphorImager SI system. The level of promoter activity was quantified using ImageQuant 5.2 software.

Western blotting and in vivo expression.

Strains GBS1011 and GBS1061 were created by transforming GBS101 and GBS1061 (9), respectively, with genomic DNA from JH695 and assaying the transformants for a Spo⁻ phenotype. Western blots were performed according to standard procedures using a 1:5,000 dilution of anti-Spo0A (a gift from C. Moran), developed with ECL Western Blotting Detection Reagents (GE Healthcare), and visualized using a VersaDoc MP 4000 system (Bio-Rad Laboratories). Protein levels were estimated based on intensity relative to a Spo0A standard.

Construction of spoIIG promoter mutants and fusion constructs.

Forward and reverse overlapping primers (Table 3) were designed to mutate the promoter-proximal 0A boxes in the spoIIG promoter in the plasmid pUCIIGtrpA (36) using the QuikChange method (Stratagene) and the 0A box sequence identified by Liu et al. (26). The plasmids encoding mutant spoIIG promoters are listed in Table 2. Mutations were confirmed by sequencing. To test activity in vivo, the mutant spoIIG promoters were amplified using primers EcoRIF (upstream primer) and BamHIR (downstream primer), and the resulting product was cloned into the EcoRI and BamHI sites of pDH32 (42), creating spoIIG promoter-lacZ translational fusions. The recombinant plasmid was linearized with PstI and transformed into either strain JH642 or strain JH695, selecting for chloramphenicol resistance. The resulting strains were used to determine spoIIG-lacZ expression during growth and sporulation. β-Galactosidase activity in cells was assayed as described by Ferrari et al. (11).

Electrophoretic mobility shift assay (EMSA).

To assess binding of phosphorylated and unphosphorylated Spo0A and Spo0A(A257V) to 0A boxes, 20 pmol of the abrB-F oligonucleotide was end labeled using [γ-³²P]ATP (6,000 Ci/mM; Amersham Biosciences) and T4 polynucleotide kinase. The labeled abrB-F was annealed to abrB-R to form a 23-bp duplex. The duplex sequence matched that of the abrB gene from +3 to +29 and contained two 0A boxes separated by 3 bp. Competitor DNAs for the abrB set of experiments were prepared by PCR amplification of a forward primer containing the desired sequence change and the reverse primer abrBΔR. Competitor DNAs for the spoIIG set of experiments were prepared by amplification of the promoter regions of plasmids containing the mutated sequences. The sequences of all primers used can be found in Table 3. Competitors were quantified by comparison to DNA standards of known mass using an AlphaImager gel documentation system (AlphaInnotech Inc.) or by SYBR green staining and scanning using a Typhoon PhosphorImager (Molecular Dynamics).

EMSAs were performed by mixing labeled abrB duplexes and various amounts of unlabeled competitor and adding the DNAs to a reaction mixture containing Spo0A or Spo0A(A257V) under previously described conditions (41). After 2 min at 37°C, the samples were loaded onto a 4.5% polyacrylamide gel and separated at 10 V/cm for 1.5 h. The gels were dried, and protein-DNA complexes and free DNA were detected using a Molecular Dynamics PhosphorImager SI system and quantified using ImageQuant software. The competition EMSAs were interpreted by analogy to enzyme kinetics with a competitive inhibitor, treating Spo0A∼P as the enzyme (E), bound radioactive DNA-enzyme complex as the enzyme-substrate complex (DNA_bound), free labeled DNA as the substrate (DNA_free), and unlabeled DNA as the inhibitor, using the equation (DNA_free + DNA_bound)/DNA_bound = (K_S/E K_I)I + b, where K_S is the dissociation constant for the labeled DNA-Spo0A∼P complex and K_I is the dissociation constant for the unlabeled competitor DNA-Spo0A∼P complex, and b is the y intercept. The data were plotted as (DNA_free plus DNA_bound)/DNA_bound versus the competitor input, I. The slope (K_S/K_I E) of the competition curve with mutant unlabeled competitor, decreases when K_I increases, indicating reduced stability of the unlabeled competitor DNA-Spo0A∼P complex. Thus, a lower slope indicates weaker binding to the competitor.

The average of three separate experiments was used to determine the slopes of the competition curves. The slopes found for Spo0A and Spo0A(A257V) are reported in Table 4. Slopes of less than 1 cannot be distinguished from one another. Each experiment used a range of unlabeled competitor DNA (approximately 5 to 40 μM). The concentrations of competitors were compared by electrophoresis of aliquots and subsequent staining, and the competitors were adjusted to the same relative concentrations. In the figures, the concentrations were normalized to a maximum input of 1. As we were interested in the relative affinities of Spo0A∼P and Spo0A(A257V)∼P for wild-type and mutant oligonucleotides reflected in the slopes, analysis of y-axis intercepts was not included.

TABLE 4.

Effectiveness of competition by promoter variants

Promoter name	Slope of competition curvea		Ratio
	Spo0A∼P	Spo0A(A257V)∼P
Wild-type abrB	74.0	67.0	1.1
abrB+1	6.7	3.8	1.8
abrB+2	4.0	0.7	5.7
abrB+3	1.7	0.2	8.5
abrB-1	3.5	2.3	1.5
abrB-2	5.0	2.1	2.4
Wild-type spoIIG	3.7	1.2	3.0
spoIIG1	5.8	0.8	7.3
spoIIG2	4.8	0.6	8.0
spoIIG3	5.4	1.3	4.1
spoIIG4	6.4	3.6	1.7
spoIIG5	38.0	33	1.2

^aThe slopes were derived from the slopes of competition curves, such as those shown in Fig. 4. A higher slope indicates better competition and therefore higher affinity for the competitor DNA.

RESULTS

A single amino acid substitution within the DNA binding domain of Spo0A, A257V, was previously reported to cause a sporulation-deficient phenotype (11). In vivo, this mutant protein retained its ability to repress transcription at the abrB promoter yet lost its ability to activate transcription from both σ^A- and σ^H-dependent promoters (31, 33). To understand the differential loss of function, we investigated whether (i) the mutation affected the ability to regulate transcription in an in vitro system and (ii) how alterations of DNA binding site sequence affected the activity of the mutant protein.

Phosphorylation and transcription regulation by Spo0A(A257V) in vitro.

Phosphorylation of Spo0A increases its affinity for DNA binding sites (2). One possible explanation for the differential loss of function observed with Spo0A(A257V) could be that the mutant protein was not phosphorylated, permitting binding to and regulation of only those promoters containing high-consensus 0A boxes, like abrB, and not promoters with low-consensus 0A boxes, like spoIIG and spoIIA. To test this possibility, we purified wild-type and mutant Spo0A proteins and directly quantified phosphorylation by a reconstituted phosphorelay. We found that both proteins showed the same initial rate and final level of phosphorylation (Fig. 1). This was the expected result, since the mutation and site of phosphorylation lie in different domains, and it indicated that decreased phosphorylation could not explain the absence of transcription activation in vivo (31, 33).

FIG. 1.

Spo0A(A257V) is efficiently phosphorylated. Spo0A and Spo0A(A257V) were incubated with the phosphorelay components KinA, Spo0F and Spo0B, and [γ-³²P]ATP for various times, as indicated, before the reaction was terminated to determine steady-state phosphorylation (A and B) or the rate of phosphorylation (C and D). (A and C) The radiolabeled components were separated by 15% SDS-PAGE, and ³²P-labeled Spo0A or Spo0A(A257V) was detected by exposure to a phosphor screen. (B and D) The extent of phosphorylation was determined by PhosphorImager analysis. Symbols: Spo0A∼P, white bars in panel B, filled circles in panel D; Spo0A(A257V)∼P, gray bars in panel B, open circles in panel D. WT, wild type. Representative images are shown. The error bars indicate standard deviations.

Previous studies have shown that both Spo0A and Spo0A(A257V) are capable of repressing abrB transcription to similar levels in vivo (31, 33). As a control for in vitro activity of the mutant protein, we compared the abilities of phosphorylated and unphosphorylated Spo0A(A257V) and Spo0A to repress abrB using a previously described single-round transcription assay (15). We found that Spo0A(A257V)∼P inhibited transcription, but slightly less effectively than Spo0A∼P (Fig. 2). For example, 800 nM of Spo0A∼P reduced transcription by 84%, while 1,200 nM of Spo0A(A257V)∼P reduced transcription by 83%. Phosphorylation of both the mutant and wild-type proteins enhanced their abilities to repress abrB transcription: 3.3-fold for Spo0A∼P but only 1.3-fold for Spo0A(A257V). These data agreed with in vivo experiments showing that the A257V substitution did not substantially affect the ability of Spo0A to repress abrB transcription (15, 29).

FIG. 2.

Spo0A(A257V) represses transcription from the abrB promoter in vitro. Phosphorylated or unphosphorylated mutant and wild-type (WT) Spo0A proteins were incubated with the initiating nucleotides ATP, UTP, and GTP and a linear DNA fragment encoding both the P1 and P2 transcription initiation sites of the abrB promoter. CTP and heparin were added 3 minutes after the addition of RNAPσ^A to permit transcript elongation. The reactions were terminated after 5 minutes, and the transcripts were separated by electrophoresis through an 8% denaturing polyacrylamide gel. ³²P-labeled abrB transcripts were detected by exposure to a phosphor screen (A), and the level of transcripts produced from the abrB promoter was determined by PhosphorImager analysis (B). Filled circles, Spo0A∼P; filled triangles, Spo0A; open circles, Spo0A(A257V)∼P; open triangles, Spo0A(A257V). Representative PhosphorImages are shown. The error bars indicate standard deviations.

We next measured the abilities of Spo0A(A257V) and Spo0A to activate spoIIG using the assay described by Spiegelman et al. (44). Using a fixed initiation time of 2 min, increasing amounts of Spo0A∼P stimulated transcription over the range of 200 to 800 nM protein (Fig. 3). Unexpectedly we found that Spo0A(A257V)∼P was capable of stimulating transcription initiation, although only to levels approximately half that of the wild type protein. Both wild-type and mutant Spo0A were more effective activators when phosphorylated. These data demonstrated that Spo0A(A257V) was capable of stimulating transcription from a σ^A-dependent promoter in vitro, at least at high inputs.

FIG. 3.

Spo0A(A257V)∼P stimulates spoIIG transcription in vitro. Phosphorylated or unphosphorylated Spo0A or Spo0A(A257V) was incubated with a linear DNA fragment encoding the spoIIG operon promoter and the initiating nucleotides ATP and GTP. RNAPσ^A was added to the mixture and allowed to initiate transcription for 2 min prior to addition of the remaining nucleotides and heparin. Following a 5-minute incubation, the elongated transcripts were separated by electrophoresis through an 8% denaturing polyacrylamide gel. ³²P-labeled spoIIG transcripts were detected by exposure to a phosphor screen (A), and the level of transcription was determined by PhosphorImager analysis (B). Filled circles, Spo0A∼P; filled triangles, Spo0A; open circles, Spo0A(A257V)∼P; open triangles, Spo0A(A257V). WT, wild type. A representative phosphorimage is shown. The values reflect the average of three independent experiments. The error bars indicate standard deviations.

Previous studies have indicated that Spo0A∼P activates transcription at the spoIIG promoter primarily by stimulating the rate at which promoter-bound, inactive RNAP converts to an active form capable of polymerizing nucleoside triphosphates (5, 34). Since the defect in transcription activation by Spo0A(A257V) could be caused by a low rate of conversion from a closed to an open complex, we measured the rates of initiation driven by Spo0A(A257V) and Spo0A (Fig. 4). Phosphorylated and unphosphorylated Spo0A and Spo0A(A257V) were incubated with template DNA and the initiating nucleotides ATP and GTP. RNAPσ^A was added, and the proteins were allowed to form initiated complexes for 0 to 120 s before a single round of elongation was permitted by adding heparin and the remaining two nucleotides. The number of transcripts increased rapidly in the presence of Spo0A∼P and reached a maximum after 90 s of incubation. Compared to the wild-type protein, we found that Spo0A(A257V)∼P also stimulated the conversion of RNAP to an active form, though threefold less than that stimulated by Spo0A∼P. The unphosphorylated forms of wild-type and mutant proteins also stimulated the conversion of RNAP to an active form, albeit at a much reduced rate relative to the phosphorylated proteins, as previously noted for wild-type Spo0A (34).

FIG. 4.

Spo0A(A257V)∼P stimulates the rate of transcription initiation at the spoIIG promoter. Transcription was initiated by adding RNAPσ^A to a reaction mixture containing a linear DNA fragment encoding the spoIIG operon promoter, the initiating nucleotides (ATP and GTP), and either Spo0A or Spo0A(A257V) (phosphorylated or unphosphorylated). At various times, aliquots were removed and added to heparin plus CTP and UTP. After the reactions were terminated, ³²P-labeled spoIIG transcripts were separated by electrophoresis through an 8% denaturing polyacrylamide gel and detected by exposure to a phosphor screen (A), and their levels were determined by PhosphorImager analysis (B). The symbols indicate the addition of Spo0A∼P (filled circles), Spo0A (filled triangles), Spo0A(A257V)∼P (open circles), and Spo0A(A257V) (open triangles). A representative phosphorimage is shown. The values reflect the average of three independent experiments. The error bars indicate standard deviations.

Expression of stage II promoter fusions in the presence of increased levels of Spo0A∼P and Spo0A(A257V)∼P.

Expression of spoIIG and spoIIA require elevated levels of activated Spo0A∼P (13). Spo0A levels increase in part by means of a feedback loop whereby Spo0A∼P stimulates expression of both σ^H and the σ^H-dependent promoter of the spo0A gene (14). If A257V disrupts the interaction between Spo0A and σ^H, then the levels of Spo0A would remain low and fail to activate expression of spoIIG and spoIIA. Since we have been unable to purify RNAPσ^H to test this directly, we relied on the overexpression of spo0A and spo0H (the gene encoding σ^H) in crsA47 mutants (9) to determine if elevated levels of Spo0A(A257V) restore the expression of spoIIG and spoIIA. We first performed a time course Western blot to follow the induction of Spo0A and Spo0A(A257V) in wild-type and crsA47 mutant backgrounds. In a wild-type background, Spo0A levels increased after the onset of stationary phase and peaked 2 h later, while Spo0A(A257V) levels increased only slightly over time (Fig. 5A and B). In particular, 2 to 3 h after the onset of stationary phase, a time when spoIIG and spoIIA expression normally peaks, the levels of Spo0A(A257V) were less than one-fourth that of the wild-type protein. In a crsA47 mutant background, the levels of both wild-type Spo0A and Spo0A(A257V) were increased, although the levels of Spo0A exceeded those of Spo0A(A257V). However, the levels of Spo0A(A257) in a crsA47 background closely matched the levels of Spo0A in a wild-type background. These data indicate that after the onset of stationary phase, the levels of Spo0A(A257V) should be sufficient to restore the expression of spoIIG and spoIIA in a crsA47 mutant background.

FIG. 5.

Elevated amounts of Spo0A(A257V) cannot stimulate transcription of spoIIG-lacZ or spoIIA-lacZ promoter fusions in vivo. B. subtilis strains were grown in SSM, and samples were collected and harvested during the transition from an exponential to a stationary phase of growth; time zero indicates the end of exponential growth. (A) Whole-cell extracts were separated by electrophoresis through 12% SDS-PAGE, transferred onto nitrocellulose membranes, subjected to immunoblot analysis using an anti-Spo0A polyclonal antibody, and detected using a VersaDoc MP 4000 system. (B) Protein levels were quantified by densitometry using purified Spo0A as a standard. (C) Transcription of spoIIG-lacZ and spoIIA-lacZ promoter fusions in B. subtilis Spo0A overexpression strains was determined by assaying β-galactosidase activity in samples of cells grown in SSM. Closed circles, spo0A; closed triangles, spo0A crsA47; open circles, spo0A(A257V); open triangles, spo0A(A257V) crsA47.

We tested the effects of increased levels of Spo0A and Spo0A(A257V) on the expression of spoIIG- and spoIIA-lacZ fusions. Consistent with the increased levels of Spo0A at early time points, expression of spoIIG-lacZ and spoIIA-lacZ peaked earlier than normal (Fig. 5C). Despite the elevated protein levels, we found that Spo0A(A257V) was unable to stimulate expression of the spoIIG- or spoIIA-lacZ fusion. These data indicate that the low in vivo levels of Spo0A(A257V) are insufficient to account for the loss of regulation of spoIIG and spoIIA in vivo.

Binding of Spo0A∼P and Spo0A(A257V)∼P to tandem 0A boxes.

The results described above led us to seek an alternative explanation for the spo0A(A257V) phenotype. One possibility consistent with the A257V substitution disturbing the packing of adjacent Spo0A molecules is that Spo0A(A257V) could regulate genes containing high-consensus 0A boxes, like abrB, but not genes containing low-consensus 0A boxes, like spoIIG and spoIIA. We tested the relative affinities of Spo0A and Spo0A(A257V) by EMSA for a series of tandem 0A boxes.

EMSAs using DNA fragments that have low-consensus 0A boxes have often indicated very low binding affinities for Spo0A∼P. However, DNase I protection assays showed that binding is selective. Since low levels of binding by EMSA could reflect either low-affinity interaction or protein-DNA instability during electrophoresis, we used a competition assay to measure relative DNA binding to different sequences to circumvent this problem. In this assay, we mixed a radioactively labeled oligonucleotide duplex containing the tandem 0A boxes from the abrB promoter with various concentrations of unlabeled DNA fragments. The DNAs were then incubated with Spo0A∼P or Spo0A(A257V)∼P for 2 min at 37°C before bound DNA fragments were separated from unbound DNA fragments by polyacrylamide gel electrophoresis and the fraction of labeled oligonucleotide bound was determined. Since binding to the tandem abrB 0A boxes is very efficient, we assumed complexes bound to this DNA did not dissociate in the gel. To compare templates, we measured the efficiency of competition by plotting the data in a reciprocal form that was expected to produce a linear relationship {(DNA_bound + DNA_free)/DNA_bound versus [competitor]}, as described in Materials and Methods. The slopes of these lines reflect the efficiency of competition by the unlabeled competitor and are reported in Table 4. Figure 6 shows examples of the EMSA gels used to generate these results, along with the curves generated for competition with the wild-type and AbrB+1 and AbrB+3 templates.

FIG. 6.

Binding of wild-type and mutant Spo0A proteins to abrB 0A box spacing variants. Spo0A∼P and Spo0A(A257V)∼P were incubated with a mixture of double-stranded ³²P-labeled oligonucleotides that contained the consensus 0A boxes from the abrB promoter and different inputs of unlabeled competitor DNAs with consensus 0A boxes separated by spacers of various lengths. The proteins were allowed to bind DNA for 2 minutes prior to challenge with calf thymus DNA in glycerol and were separated on a running 8% nondenaturing polyacrylamide gel. (A) ³²P-labeled DNA was detected by exposure to a phosphor screen, and the levels of bound versus unbound DNA were determined by PhosphorImager analysis. (B) The data are plotted as (DNA_free + DNA_bound)/DNA_bound versus the normalized input for wild-type abrB, abrB+1, and abrB+3 competitor DNAs. Each data point represents the average from three separate experiments. The slopes of the curves are reported in Table 4. Filled circles, Spo0A∼P; open circles, Spo0A(A257V)∼P.

Because the A257V mutation is expected to affect the packing of adjacent Spo0A molecules, we first tested competition between labeled abrB 0A boxes and unlabeled duplex oligonucleotides in which the sequences of 0A boxes at the abrB promoter were maintained while the spacing between them was altered. We found that increasing the 3-bp spacing of the wild-type abrB 0A boxes to 4 bp (abrB+1) decreased the effectiveness of competition by a factor of 10 for wild-type Spo0A∼P. Further increases of the spacing between the 0A boxes further reduced the effectiveness of the competitors. Decreasing the spacing by 1 or 2 bp had roughly the same effect. In both cases, the oligonucleotides with reduced spacing were poor competitors.

By comparison, we found that the abrB binding site variants were much less effective competitors for Spo0A(A257V)∼P. Spo0A(A257V)∼P bound competitors with one, two, or three additional base pairs between the 0A boxes two-, five-, and eightfold more poorly than did Spo0A∼P. Fragments with reduced spacing were only modestly less effective competitors for Spo0A(A257V)∼P binding than the wild-type protein. These data indicated that the A257V substitution rendered the protein less tolerant of alterations in the spacing of tandem 0A boxes.

To explore the effects of sequence changes within the 0A boxes on the binding of Spo0A∼P and Spo0A(A257V)∼P, we tested competition between labeled abrB 0A boxes and unlabeled duplex oligonucleotides in which the sequences of the 0A boxes were varied from low consensus (spoIIG-like) to high consensus (abrB-like) while the spacing between them was held constant. We found that competitor DNAs with low-consensus 0A boxes (spoIIG and spoIIG1 through spoIIG4) were poor competitors for Spo0A∼P or Spo0A(A257V)∼P binding. However, the low-consensus 0A boxes were two- to sixfold less effective competitors for Spo0A(A257V)∼P than for Spo0A∼P. Only when the unlabeled duplex contained consensus 0A boxes (spoIIG5) did it effectively compete for binding of Spo0A∼P and Spo0A(A257V)∼P with the labeled abrB 0A boxes. These data supported the conclusion that the A257V substitution rendered the protein less tolerant of alterations from the consensus sequence of tandem 0A boxes.

Expression of spoIIG-lacZ fusions with high-consensus 0A boxes.

The preceding results indicated that the specific loss of transcription activation at the spoIIG promoter is due to the low consensus of the site 2 0A boxes. We tested this prediction in vivo by making lacZ promoter fusions using the same spoIIG promoter mutations tested in vitro in the DNA binding competition assay. The promoter fusions were introduced into B. subtilis strains containing wild-type spo0A (JH642) or mutant spo0A(A257V) (JH695). Previous work had shown that increasing the consensus of the site 2 0A boxes resulted in elevated levels and an earlier onset of transcription (2). We saw similar results in strain JH642 (Fig. 7). Each successive mutation of the site 2 0A boxes toward consensus increased the maximal level of transcription and activated the promoter at earlier time points. These data support the conclusion that high-consensus binding sites facilitate promoter activation at lower levels of Spo0A.

FIG. 7.

Spo0A(A257V) stimulates transcription from spoIIG promoters with high-consensus 0A boxes. B. subtilis strains JH642 (spo0A) and JH695 [spo0A(A257V)] were transformed with DNA encoding spoIIG-lacZ promoter mutants (spoIIG-1 through spoIIG-5) (Table 5). The strains were grown in SSM, and samples were collected and harvested during the transition from an exponential to a stationary phase of growth; time zero indicates the end of exponential growth. Filled squares, spoIIG1; open triangles, spoIIG2; filled triangles, spoIIG3; open circles, spoIIG4; filled circles, spoIIG5.

In strains containing spo0A(A257V), however, only the spoIIG-lacZ promoter fusion with consensus 0A boxes was effectively transcribed. When the 0A boxes were identical to those at the abrB promoter, spoIIG was activated by Spo0A(A257V) just as the abrB promoter was effectively repressed. These data indicated that Spo0A(A257V) was capable of both positively and negatively regulating transcription, but only at promoters containing consensus binding sites.

DISCUSSION

Spo0A is an interesting transcription factor, as it has been identified as a repressor and as an activator for forms of RNAP associated with different σ factors. One mutation in Spo0A, A257V, is unusual because it appears to separate the positive and negative regulatory functions of the protein. In B. subtilis strains encoding this mutant Spo0A, the abrB gene is repressed while the spoIIG operon is not activated (31, 33).

We undertook this analysis to resolve the differential effects of the mutation on these functions. In this study, we showed that Spo0A(A257V) was phosphorylated at the same rate and to the same extent as Spo0A (Fig. 1A and B) and was capable of activating and repressing transcription with only modestly reduced efficiency. After ruling out low Spo0A(A257V) protein levels in vivo as a possible cause for misregulation of stage II genes (Fig. 4 and 5), we reasoned that the mutation could compromise DNA binding. Competitive EMSA experiments showed that Spo0A and Spo0A(A257V) bound nearly identically to high-consensus tandem 0A boxes, but when the sites deviated from consensus, we found large differences in relative affinities (Fig. 6 and Table 4). Consistent with this, Spo0A(A257V) could activate transcription of spoIIG in vivo, but only when the low-consensus 0A boxes near the promoter were replaced with near-consensus 0A boxes (Fig. 7). We conclude that the in vivo phenotype of strains containing Spo0A(A257V) reflects reduced binding of the mutant protein to low-consensus binding sites; Spo0A(A257V) regulates only those promoters containing consensus tandem 0A boxes. Since the A257V mutation is located at the dimer interface distant from the DNA binding surface at tandem 0A boxes (47), these data must reflect a diminished interaction between Spo0A monomers. By inference, the wild-type Spo0A dimer interface must compensate for the reduced interactions between the protein and the DNA to stabilize binding at low-consensus 0A boxes.

The observation that the level of similarity to tandem consensus 0A boxes has such a dramatic effect on regulation by Spo0A(A257V) raises interesting questions about how the Spo0A regulon is constructed. A search of the Bacillus genome for “ideal” tandem 0A boxes (5′-TTGCACAN_3-5TTGCACA-3′) showed that there are five such sites that have only one mismatch; there are no perfect matches. Two of these sites are near the abrB gene, but only the site downstream of the initiation site appears to be in a position relevant to regulation. The ywsB gene has a tandem 0A box at position −87 and was reported to be a low-Spo0A-threshold repressed gene (13). The function of YwsB has not been examined. The divergently transcribed genes yppF and yppG have a high-consensus tandem 0A box within the intergenic region, but these genes are also unstudied and are not known to be responsive to Spo0A. The ywmC gene is preceded by high-consensus tandem 0A boxes, but they are located nearly 200 bp from the translation start site and may be irrelevant to expression, since the location of the promoter for the gene is not known.

To computationally locate tandem 0A boxes near most of the genes within the B. subtilis genome reported to be regulated by Spo0A (28), one needs to permit more than four mismatches within the pair. Our in vitro binding data indicate that these are low-affinity sites. Using these computational restraints, over 2,000 potential Spo0A binding sites can be identified. This is greatly in excess of the number of genes identified by chromatin immunoprecipitation microarray experiments whose expression is proposed to be directly regulated by Spo0A (28). In fact, the 35 most highly Spo0A-regulated genes (28) do not appear to have tandem 0A boxes, even allowing up to six mismatches. While these genes may not be directly regulated by Spo0A, if they are, then this implies that the sequence of the 0A boxes is not the primary determinant of Spo0A binding. We observed this at the spoIIG promoter, where we showed that Spo0A∼P is recruited by RNAP to the 0A boxes, at least on linear templates (40). Inhibition of transcription at abrB happens without displacing RNAP, suggesting that Spo0A∼P may also contact the polymerase at this promoter, as well (15).

Liu et al. (26) used a computational approach to derive the Spo0A binding site consensus. They searched the genomes of three Bacillus species for these sites and the transcription units associated with at least a single 0A box. They examined intergenic sequences 80 nucleotides from the start of translation for genes in transcription units. The set of genes that contain one high-weight matrix score 0A box includes one of the genes with high-consensus tandem 0A boxes (yypF), but not the others, in which the tandem 0A boxes are further than 80 nucleotides away from the translation start sites. Liu et al. also noted that there are very few tandem high-consensus 0A boxes and suggested that possibly only a single 0A box is needed, even for a Spo0A dimer to bind DNA (26). In this study, we showed that the strength of Spo0A∼P binding to tandem 0A boxes decreased rapidly when the 0A boxes were separated or changed from “ideal.” We suggest that the lack of high-consensus tandem binding sites reflects the nature of Spo0A regulation. Spo0A regulates the sequential activation and repression of genes at different stages of sporulation. The modifications of the Spo0A binding sites at the spoIIG promoter (Fig. 7) showed that the 0A box sequence is one of the regulatory inputs that determine the timing and relative order of gene expression, along with the amount of Spo0A and its level of phosphorylation.

The observation that the level of consensus of the 0A boxes made such a large difference in the ability of Spo0A(A257V) to activate transcription in vivo was surprising and suggested different modes of DNA binding to high- and low-consensus 0A boxes. This might be related to the currently unsolved mystery of the molecular rearrangements involved in the activation of Spo0A. It has been amply demonstrated that activation is induced by phosphorylation, but it is not yet clear what phosphorylation does. Phosphorylation may, as has been suggested, induce dimer formation of the N-terminal receiver domain (25). One can envision that full-length Spo0A may bind to DNA as a monomer through its C-terminal domain. Binding of a second monomer could stabilize the first contacts via the interface between the two monomers observed in the Spo0A C-terminal domain DNA crystal structure. If phosphorylated, the N terminus could dimerize, strengthening the interaction of the Spo0A pair with the DNA, stimulating productive interaction with RNAP (and possibly other effectors). Since the A257V mutation has been predicted to disrupt the interface between the C-terminal domains (47), Spo0A(A257V) may be particularly dependent on maximizing sequence-specific protein-DNA contacts to drive dimerization, stable binding, and transcriptional regulation in vivo. Such a model is appealing, as the cooperative binding of Spo0A∼P would endow these regulatory events with switch-like behavior, though we lack convincing evidence for cooperativity from in vitro analysis.

An alternative effect of receiver domain phosphorylation could be to reverse inhibition of the C-terminal DNA binding domain. Transcription activation might require repositioning the receiver domains, for example, if the receiver domain interacts with a surface of the C-terminal domain that forms the interface between DNA-bound Spo0A pairs. The interaction between the DNA binding domains and 0A boxes could contribute to the displacement of receiver domains and activation. Mutations that decrease interaction between the C-terminal domains, such as the A257V mutation, could reduce the tendency to displace the receiver domain interaction. Suppressors that increase C-terminal domain interaction (31) would restore activation. This model implies that the receiver domain is in fact inhibitory. A previous analysis of a constitutively active D75S mutation in the receiver domain that appeared to be in a “locked-off” position supports this view (8).

In conclusion, we have shown that the A257V mutation, which fails to activate transcription from key stage II promoters yet retains the ability to repress the abrB promoter, is able to bind to and regulate the expression of only those genes whose promoters contain high-consensus binding sites. This mutation reveals the importance of the interface between adjacent dimers in selecting binding sites and provides insight into how Spo0A might be activated.

TABLE 5.

Sequences of spoIIG and abrB promoter variants

Promoter name	Sequencea
Wild-type spoIIG	CCTCTCAACATTAATTGACAGAC
spoIIG1	CCTCTCAACATTATTTGACAGAC
spoIIG2	CCTCTCAACATTATTTGACAAAC
spoIIG3	CCTCTCGACATTATTTGACAAAC
spoIIG4	CCTTTCGACATTATTTGACAAAC
spoIIG5	CCTTTCGACATTATTCGACAAAC
Wild-type abrB	TTCGTCATTATTCGACA
abrB+1	TTCGTCATTAATTCGACA
abrB+2	TTCGTCATTATATTCGACA
abrB+3	TTCGTCATTATTATTCGACA
abrB-1	TTCGTCATTTTCGACT
abrB-2	TTCGTCATTTCGACA

^aThe non-template strand sequences for spoIIG and spoIIG promoter variants are shown. 0A boxes are underlined. Mutations expected to increase Spo0A binding are shown in boldface and underlined. The template strand sequences for abrB and abrB promoter variants are shown. Base pairs between the 0A boxes are indicated in italics.