Analysis of tnrA Alleles Which Result in a Glucose-Resistant Sporulation Phenotype in Bacillus subtilis

Abstract

Bacillus subtilis cells cannot sporulate in the presence of catabolites such as glucose. During the analysis of Tn10-generated mutants, we found that deletion of the C-terminal region of the tnrA gene, which encodes a global regulator that positively regulates a number of genes in response to nitrogen limitation, results in a catabolite-resistant sporulation phenotype. Analyses of nrg-lacZ and nasB-lacZ, which are activated by TnrA under nitrogen limitation, showed that C-terminally truncated TnrA activates nitrogen-regulated genes constitutively. The relief of catabolite repression of sporulation may result from the uncontrolled expression of the TnrA-regulated genes.

Bacillus subtilis is a gram-positive bacterium that can differentiate into heat-resistant spores under conditions of nutrient deprivation. The initiation of sporulation is controlled by phosphorylation of a single transcription factor, Spo0A (reviewed in reference 9). Accumulation of a high-enough concentration of Spo0A∼P activates the transcription of spoIIA, spoIIE, and spoIIG, which are required for cell-type-specific gene expression during the next stage of sporulation. The Spo0A transcription factor is a member of a response regulator of a two-component system, and there are at least three histidine kinases, encoded by kinA, kinB, and kinC, that are responsible for production of Spo0A∼P. Normally, these kinases do not transfer phosphate directly to Spo0A. Instead, Spo0F, a single-domain response regulator, is phosphorylated by one of the kinases and then transfers the phosphate to Spo0A by means of a response regulator phosphotransferase, Spo0B. This extended version of a two-component system may provide the multiple targets for regulation by various environmental and physiological signals such as nutrient depletion, cell density, and Krebs cycle, DNA synthesis, and DNA damage (13).

It is known that the presence of catabolites such as glucose inhibits sporulation. There are many mutations that lead to the catabolite-resistant sporulation (Crs) phenotype, such as those in pai (12), hpr (24), and rpoD (15). Some mutations in pts (6), degS or degU (8, 16, 21), or gsiA (22) also yield the Crs phenotype. Although the mechanisms underlying most Crs mutations are still unknown, it seems that all known Crs phenotypes produce Spo0A∼P in the presence of glucose. Some mutations in spo0A, such as rvtA11 and sof-1, that bypass the phosphorelay also show the Crs phenotype, indicating that phosphorelay may be a main target for catabolite repression of sporulation in the wild type (10, 14, 27, 28).

In this paper, we show that deletion of the C-terminal region of TnrA also results in the Crs phenotype. TnrA is a regulatory protein that is involved in the activation of nitrogen-regulated genes, such as gabP (γ-aminobutyrate permease), nasB (nitrate assimilatory enzymes), ure (urease), and nrgA (putative ammonium permease), during nitrogen-limited growth (4, 31). Our findings suggest that deletion of the C-terminal region of TnrA resulted in constitutive expression of the nitrogen-regulated genes and that this uncontrolled expression of TnrA-regulated genes results in cells with a Crs phenotype that can sporulate in the presence of a normally inhibitory concentration of glucose.

Isolation of the tnrA::Tn10 mutant.

During the screening of Crs mutants generated by mini-Tn10 mutagenesis of the IS75 strain, we found that a tnrA allele, in which Tn10 integrated into an open reading frame (ORF) of the tnrA gene, showed the glucose-resistant sporulation phenotype. The tnrA::Tn10 mutant efficiently sporulates in DSM (Difco sporulation medium) containing 2% glucose (DSMG), showing about 50% sporulation frequency. The sporulation frequency is decreased to about 10% in DSM supplemented with both 2% glucose and 0.2% glutamine (DSMGQ), indicating that catabolite repression of sporulation is only partially overcome in the presence of both glucose and glutamine. The Crs phenotype resulting from the tnrA::Tn10 allele is not a strain-specific property. We produced the tnrA::Tn10 mutation in two other strains, 168 and JH642, and found that these strains also showed similar sporulation phenotypes (see Table 1 for JH642). We used the JH642 strain for further genetic study.

TABLE 1.

Analysis of sporulation phenotypes of the tnrA mutants^a

Strain	Relevant genotype	% Sporesb in medium:			spoIIA-lacZc expression in medium:
		DSM	DSMG	DSMGQ	DSM	DSMG	DSMGQ
JH642	Wild type	75.2 (2.4 × 108)	3.9 (3.9 × 108)	1.8 × 10−3 (3.3 × 105)	97	27	2.4
BS9830	tnrA::Tn10	63.4 (1.8 × 108)	51.4 (5.8 × 108)	11.2 (1.5 × 108)	NDd	ND	ND
BS9909	ΔtnrA::erm	76.4 (1.9 × 108)	3.7 (2.3 × 108)	2.5 × 10−3 (7.2 × 105)	255	52	18
BS9913	ΔtnrA7	32.8 (1.4 × 108)	25.4 (5.5 × 108)	8.5 (1.2 × 108)	71	84	31
BS9932	ΔtnrA20	59.6 (1.6 × 108)	55.3 (5.4 × 108)	5.4 (9.5 × 107)	122	182	119
BS9914	ΔtnrA34	55.0 (1.7 × 108)	9.7 (6.0 × 107)	7.8 × 10−3 (5.1 × 105)	253	0.7	0.7
BS0001	ΔglnA::spc	30.4e (7.6 × 107)	ND	2.9 (2.2 × 107)	71e	ND	104
BS0002	ΔglnA::spc ΔtnrA::erm	ND	ND	ND	214e	ND	4.5

^aThe JH642 strain is trpC2 pheA1, and all strains used are derivatives of JH642. All tnrA mutations were constructed using PCR and integrated into the tnrA locus with the erythromycin resistance marker, derived from the integrative vector pDG1728 (7). The glnA allele was obtained by integrating plasmid pGLN14 (31) into the chromosome, selecting for spectinomycin resistance (100 μg/ml), and checking for chloramphenicol sensitivity (5 μg/ml) to confirm that there was a double crossover. 

^bPercent spores is percentage of spores out of viable cells per milliliter of culture broth. Sporulation frequencies were measured as described previously (26). The numbers in parentheses are those of heat-resistant spores per milliliter. Each value is the average of two to five determinations. The standard error did not exceed 30% for any values. 

^cThe spoIIA-lacZ fusion, which results from a Cambell-type integration into the spoIIA-lacZ locus, was described elsewhere (19). The specific activities of β-galactosidase were determined in cells harvested 2 h after the end of the exponential growth phase. The assays were performed with toluenized cells, as described by Nicholson and Setlow (23). Each value is the average of two or three determinations. The standard error did not exceed 15% for any values. 

^dND, not determined. 

^eTo analyze the glnA mutant, 0.2% glutamine was added to DSM to assay sporulation and measure β-galactosidase activity. 

To establish a true tnrA-null genotype, we constructed a deletion-insertion mutation where the internal tnrA ORF was replaced with an erm gene cassette, creating strain BS9909 (See Fig. 2). Interestingly, this null allele of the tnrA gene resulted in a glucose-sensitive sporulation phenotype, which is similar to that of the wild-type strain JH642 (Table 1). This fact indicated that the Crs phenotype is not caused by the tnrA null mutation. Wray et al. also reported that the tnrA62::Tn917 mutation did not relieve glucose repression of sporulation in nutrient sporulation medium containing 1% glucose (32). We observed that insertion of Tn10 within the C-terminal region of the tnrA ORF created a chimeric protein in which seven codons of the C-terminal tnrA ORF were replaced with 11 codons provided by the Tn10 sequence (Fig. 1). This suggests that the Crs phenotype resulting from the tnrA::Tn10 mutation is due to chimeric TnrA protein.

FIG. 2.

Construction of tnrA mutations. The DNA fragments used to construct the tnrA mutations and the erythromycin resistance gene are diagrammed. The flanking DNA fragments used for chromosomal integration of tnrA mutations are not shown. The 1.1-kb erm cassette is not drawn to scale. Arrows indicate the coding regions of the genes. C-terminally deleted tnrA mutations were constructed by insertion of stop codons at desired sites in the tnrA ORF. aa, amino acids.

FIG. 1.

Map of the chromosomal region of tnrA. The arrows indicate the coding regions of the genes. The location of the Tn10 transposon insertion is indicated. C-terminal regions of both the TnrA wild type and the TnrA chimera encoded by tnrA::Tn10 are also shown. Amino acids translated by the Tn10 sequence are underlined.

Construction of C-terminally deleted tnrA mutants.

Since it seemed that truncation of the C-terminal region of tnrA may be important for the Crs phenotype that we observed, a set of deletion mutations was made in which C-terminal amino acid residues of the tnrA product were serially removed (Fig. 2). All of these constructs were integrated into the original tnrA locus along with the erm gene cassette. Construction of the mutations was confirmed by both Southern hybridization and sequencing of DNA fragments obtained by PCR. As shown in Table 1, strains BS9913 and BS9932, harboring ΔtnrA7 (seven C-terminal amino acids deleted) and ΔtnrA20 (20 C-terminal amino acids deleted), respectively, showed Crs phenotypes similar to that of the tnrA::Tn10 mutant, though the BS9913 strain had decreased sporulation frequencies in both DSM and DSMG compared to that of BS9932. Strain BS9914 containing ΔtnrA34 (34 C-terminal amino acids deleted) exhibited a glucose-sensitive sporulation phenotype as well. The strain that contains the intact tnrA gene with an erm cassette showed a sporulation phenotype similar to that of strain JH642 (data not shown).

The Crs-type tnrA mutants (BS9913 and BS9932) grew more slowly than did wild-type cells in DSM (generation time of 27 min versus 20 min for JH642) and had a reduced cell yield (final optical density at 600 nm of 1.9 to 2.1 versus 2.4 to 2.6 for JH642). There was no significant difference in the growth rates of the tnrA-null mutant (BS9909) and the wild-type strain in DSM.

The sporulation phenotypes of strains with the tnrA alleles were somewhat variable and seemed to be sensitive to specific growth conditions. To analyze sporulation phenotypes of the tnrA mutants more quantitatively, expression of a spoIIA-lacZ fusion in various tnrA mutants were examined. Since expression of spoIIA reflects production of Spo0A∼P, one of the key transcription factors in sporulation initiation, monitoring β-galactosidase activity from a spoIIA-lacZ strain provides a good indication of the ability of cells to initiate sporulation. As shown in Table 1, expression levels of the spoIIA-lacZ fusions in the wild-type strain (JH642), the tnrA-null mutant (BS9909), and the strain having ΔtnrA7 (BS9914) were repressed by 2% glucose, but strains having ΔtnrA7 (BS9913) and ΔtnrA20 (BS9932) expressed spoIIA-lacZ during growth in DSMG. These results agree with previous data on sporulation frequency. Interestingly, spoIIA-lacZ expression in the ΔtnrA7 and ΔtnrA20 mutants was barely repressed in DSMGQ culture, although sporulation frequencies were reduced in the same medium. This fact suggests that, in DSMGQ culture, Crs mutants can produce enough Spo0A∼P for activation of spoIIA but the sporulation process is still repressed. In summary, it is likely that the Crs phenotype exhibited in DSMG by strains with the truncated tnrA alleles was mainly due to a loss of C-terminal amino acid residues from TnrA.

Analysis of nitrogen regulation of tnrA mutants.

TnrA is a global regulator that positively regulates a number of genes and operons in response to nitrogen limitation. Since the Crs phenotype is apparently caused by truncated TnrA protein, we checked strains with the different tnrA alleles to determine whether their nitrogen-regulating function was normal. Transcriptional lacZ fusions to nrg and nasB, which are activated by TnrA under nitrogen-limited conditions, were constructed in various tnrA mutants, and β-galactosidase activities were monitored during growth in TSS minimal medium. We used 0.2% glutamate as a limiting nitrogen source and 0.2% glutamate plus 0.2% (wt/vol) glutamine as an excess nitrogen source. As previously reported (31), TnrA was responsible for most of the induction of the nrg-lacZ and nasB-lacZ expression during nitrogen-limited growth. While the expression of fusions in the wild-type strain (JH642) was strongly induced during nitrogen-limited growth, induction was barely observed in the strains BS9909 and BS9914 (Table 2). Interestingly, in the strains showing the Crs phenotype (BS9913 and BS9932), the nrg-lacZ and nasB-lacZ expression levels were increased in both nitrogen-limited and excess-nitrogen conditions. This fact suggests that ΔTnrA7 and ΔTnrA20 constitutively activate nitrogen-regulated genes, such as nrg and nasB, irrespective of nitrogen availability.

TABLE 2.

Analysis of nitrogen-regulating function of the tnrA mutants

Strain	Relevant genotype	Expression of nrg-lacZa in mediumb:		Expression of nasB-lacZa in mediumb:
		TSSE	TSSEQ	TSSE	TSSEQ
JH642	Wild type	587	0.3	17	0.2
BS9909	ΔtnrA::erm	0.4	0.1	0.5	0.2
BS9913	ΔtnrA7	721	775	45	49
BS9932	ΔtnrA20	573	758	27	56
BS9914	ΔtnrA34	0.6	0.2	0.3	0.2
BS0001	ΔglnA::spc	NDc	562	ND	85

^aThe upstream regions of nrg (−80 to +213) and nasB (−126 to +259), relative to the transcriptional start site, were amplified with PCR and integrated into the amyE locus in a single copy by means of a double crossover using plasmid pDG1728 (7). The specific activities of β-galactosidase were determined in cells harvested during exponential growth phase (optical density at 600 nm of 0.4 to 0.6). Each value is the average of two or three determinations. The standard error did not exceed 15% for any values. 

^bTSSE, TSS minimal medium (5) containing 0.2% glutamate as a limited nitrogen source; TSSEQ, TSSE plus 0.2% glutamine as an excess nitrogen source. 

^cND, not determined. 

Although the metabolic signal molecule that regulates TnrA activity is not known, it is generally believed that glutamine synthetase (GS) is required for production and/or transduction of the nitrogen signal (31). If the Crs phenotypes exhibited by strains with the ΔtnrA7 and ΔtnrA20 alleles are caused by constitutive activation of nitrogen-regulated genes, then the glnA mutant, which cannot synthesize GS, should also have the Crs phenotype, because nitrogen-regulated genes are activated constitutively in the glnA mutant (Table 2) (31). However, as the glnA mutant is a glutamine auxotroph, we could not examine the sporulation phenotype in DSMG. Instead, expression of spoIIA-lacZ was measured, since the ΔtnrA7 and ΔtnrA20 mutants expressed spoIIA-lacZ during growth in DSMGQ culture even when sporulation was partially repressed. As shown in Table 1, sporulation frequency of the glnA mutant was partially repressed in the DSMGQ culture, but expression of spoIIA-lacZ was induced, similar to the results obtained for the ΔtnrA7 and ΔtnrA20 mutants. Furthermore, we also observed that spoIIA-lacZ expression of the glnA mutant grown in DSMGQ was almost repressed again by mutational inactivation of the tnrA gene (Table 1), indicating that spoIIA-lacZ expression of the glnA mutant resulted from constitutive expression of the nitrogen-regulated genes. All of the results we obtained imply that constitutive expression of nitrogen-regulated genes caused by C-terminally truncated TnrA results in derepression of sporulation in DSMG culture.

It is difficult to explain why the constitutive expression of tnrA-regulated genes results in a Crs phenotype in DSMG. Recently, it was reported that, in the histidine utilization (hut) operon, carbon catabolite repression (CCR) was partially relieved in the glnA mutant (33). Since this partial relief of CCR was suppressed by tnrA mutation, it was proposed that the defect in CCR of the hut operon seen in the glnA mutant could result from the inappropriate expression of TnrA-regulated genes. The observation that the nrg-lacZ and nasB-lacZ fusions are also expressed constitutively in the Crs-type tnrA mutants raises the possibility that both the Crs phenotype of the tnrA mutants and the relief of CCR of the hut operon may result from the same metabolic imbalance caused by the uncontrolled activation of TnrA-regulated genes. It is also possible that the intracellular concentration of GTP (and/or GDP), which is closely related to induction of sporulation (17, 18), could be affected by inappropriate expression of TnrA-regulated genes. We observed that the upstream region of a putative xanthine-uracil permease gene (yunJ) has two TnrA-binding consensus sequences (TGTNAN₇TNACA), which are upstream DNA sequences commonly found in TnrA-regulated genes (31). yunJ is followed by three genes, yunK, yunL, and yunM, which may be organized in an operon structure. Since yunL also codes for a uricase-like protein, the gene of this operon may code for a system to take up and degrade purine. As xanthine dehydrogenase activity is also subject to the GS-dependent signaling pathway (3), C-terminally truncated TnrA can stimulate constitutive degradation of purine molecules, resulting in a decrease in the GTP pool and induction of sporulation initiation. Examination of the intracellular GTP pool in the Crs-type tnrA mutant should help to determine the accuracy of this hypothesis.

It is interesting to note that spoIIA-lacZ expression in the BS9909 and BS0002 strains resulted in about a twofold-higher level of β-galactosidase activity than that of the wild-type control. This observation suggests that TnrA protein negatively affects formation of Spo0A∼P. Recently, Wang et al. (30) discovered an operon that encodes a KipI protein that specifically inhibits the autophosphorylation reaction of KinA. Interestingly, TnrA is required for activation of an operon containing a kipI gene. Wang et al. showed that, in the absence of functional TnrA, KipR, a product of another gene of the kipI-containing operon, completely represses the expression of the kipI-containing operon. We do not know what exactly is responsible for stimulation of spoIIA-lacZ by the tnrA null mutation, but because KipI is a potent inhibitor of KinA activity, loss of activation of the kip operon, which may be caused by the tnrA null mutation, should stimulate formation of Spo0A∼P. Strain BS9914 also has a higher level of spoIIA-lacZ expression in DSM than does wild type, but expression was strongly repressed in DSMG and DSMGQ medium. Thus, it seems that ΔtnrA34 is not a null allele, though the phenotypes of sporulation frequency and nitrogen regulation are very similar to those of the null mutant. We do not know the exact reason for strong repression of the spoIIA-lacZ expression of strain BS9914 in DSMG and DSMGQ.

TnrA protein belongs to the MerR family of regulators (29). These MerR-type proteins are characterized by an N-terminal DNA-binding domain which is highly homologous within the group and which shares homology with various C-terminal domains specific to general substrates such as mercuric ions (MerR [20]), thiostrepton (TipA [11]), or oxidative stress (SoxR [1]). Although the effector molecule of TnrA is still unknown, the results of analysis of the tnrA mutants presented here suggest that the C-terminal region of the TnrA protein participates in sensing the hypothetical inducer molecule. As the deletion of the C-terminal region of TnrA allows activation of target genes without the inducer molecule, it seems that separation of the DNA-binding domain from the C-terminal region mimics activation of TnrA naturally occurring under nitrogen-limited conditions. Recently, Baranova et al. reported that the activity of Mta, a MerR-type regulator involved in the regulation of multidrug transporters, is also stimulated by removal of a C-terminal inducer-binding domain (2). Thus, it seems likely that the activity of the DNA-binding domain (possibly having an affinity for the target promoter) is inhibited by the C-terminal domain under repressed conditions and that this inhibition is relieved by the interaction of an inducer molecule with the C-terminal domain under derepressed conditions. This may be a common mechanism for regulation of MerR-type proteins.