Newly Identified Nucleoid-Associated-Like Protein YlxR Regulates Metabolic Gene Expression in Bacillus subtilis

Expression of genes encoding NAPs is often temporally regulated. According to results from single-cell analysis, the ylxR gene is induced by glucose and expressed in a bistable mode. These characteristics have not previously been reported for NAP gene expression. Transcriptional profiling of the ylxR disruptant revealed a change in the expression levels of approximately 400 genes, including genes for synthesis of 12 amino acids and 4 nucleotides, in addition to the SigX/M regulons. Thus, YlxR is a critical regulator of glucose response in B. subtilis.

G lucose is the most favorable carbon source for the majority of bacteria, and therefore bacteria have several glucose-responsive gene networks (1). In Grampositive bacteria, including Bacillus subtilis, the transcription factor CcpA is the master carbon catabolite regulator (1,2). The incorporation of glucose accelerates carbon flow in glycolysis, leading to an increase of fructose 1,6-bisphosphate. This increase is thought to trigger the phosphorylation of Ser46 of HPr, a phosphocarrier protein in the sugar phosphotransferase system (P-Ser-HPr). P-Ser-HPr associates with and activates CcpA, leading to global positive and negative effects on the transcriptional network, including for genes encoding carbon metabolism enzymes. Moreover, there are several additional glucose-responsive transcription factors, such as CcpC, CcpN, CggR, and GlcT (2). In Escherichia coli, catabolite gene-activator protein (CAP [Crp]) has been considered a conventional transcription factor responding to glucose. However, recent genomic analyses led to a new idea-that CAP is a nucleoid-associated protein (3).
Bacterial chromosomal DNA had been thought to lack histones. However, bacterial histone-like proteins such as HU and IHF have been found, and their roles were clarified in phage recombination and gene transcription (4). Accumulated studies revealed that this type of proteins has the distinct nature of histones. Thus, proteins which are not structurally related to histones but have similar functions to histones have been found in bacteria. The group of proteins related to bacterial chromatin structure are called nucleoid-associated proteins (NAPs) (5). NAPs have many roles in transcription, recombination, and chromosome condensation, rearrangement, maintenance, and segregation (3). NAPs generally have DNA-binding activity, which is sequence specific and/or sequence independent, or NAPs recognize local DNA structure (3). The modes of transcriptional regulation of NAPs are diverse: for example, H-NS inhibits RNA polymerase (RNAP) progression on DNA, while Fis regulates transcription through various modes of interaction with RNAP (5). Some NAPs play roles in nutrient-dependent transcriptional regulation. For example, in E. coli, leucine-responsive regulatory protein (Lrp) regulates about 10% of all genes (6); Lrp activity is potentiated, inhibited, or unaffected by leucine for different target genes.
Recently, we found that in B. subtilis, glucose induces expression of the extracellular sigma factor genes sigX/M (7). To explore the factors affecting this phenomenon, we performed a transposon mutagenesis screen for mutants with no glucose induction (GI) of sigX-lacZ and identified ylxR (7). In this report, we confirmed the widely conserved ylxR gene in eubacteria as a required factor for GI of sigX/M. Further analysis revealed that ylxR is induced by glucose addition. In vitro DNA-binding and cytological studies of YlxR suggested that YlxR is a NAP in B. subtilis. Thus, we performed transcriptome sequencing (RNA-Seq) analysis to evaluate the impact of ylxR disruption on the B. subtilis transcriptome and observed that YlxR has a profound impact on metabolic gene expression, including on genes involved in synthesis of 12 amino acids and 4 nucleotides.

RESULTS
Identification of ylxR with no GI of sigX-lacZ. In transposon mutagenesis screening for mutants with no GI of sigX-lacZ, we identified uncharacterized ylxR, as well as cshA, the latter encoding a DEAD box RNA helicase that has been reported to be associated with RNAP (7-9). Thus, we previously presented the model shown in Fig. 1A. Considering the structure of the ylxR-containing operon, Tn insertion into ylxR could have a polar effect on downstream genes such as ylxQ (Fig. 1C). Thus, a ylxQ disruption mutant was constructed and introduced into a strain bearing sigX-lacZ. Examination of ␤-galactosidase (␤-Gal) activity in the resultant strain showed no effect of ylxQ disruption: i.e., GI of sigX-lacZ in the ylxQ disruptant was similar to that in the wild-type strain (Fig. 1B, left), where GI was 3-fold (7). Next, a ylxR disruptant with xylose-inducible ylxR in the amyE locus was constructed, and the ␤-Gal activity was examined. Without xylose, the strain showed a little GI, probably due to leaky expression of the xyloseinducible promoter (Fig. 1B, right). In the presence of xylose, the strain showed significant GI of sigX-lacZ. These results indicate that ylxR is involved in the GI of sigX. We note that the ylxR disruption in strain OAM735 is in fact a ylxR depletion mutation due to the probable leaky expression of IPTG (isopropyl-␤-D-thiogalactopyranoside)inducible Pspac-driven intact ylxR (see Fig. S1A in the supplemental material). The phenotype of this mutant without IPTG, however, could not be distinguished from that of the Tn-inserted ylxR mutant. (The depleted mutant was used in Fig. 1B, and see Fig. 6B below, where the depleted and Tn mutants are compared.) Thus, we consider both to be similar mutants.
The ylxR gene encodes a small, basic protein (molecular weight [MW], 10.3 kDa; isoelectric point, 10.09) with uncharacterized domain DUF448, and it is widely conserved in eubacteria (see Fig. S2 in the supplemental material). According to the Pfam database, 2,052 species have a YlxR ortholog. The structure of Streptococcus pneumoniae YlxR has been resolved, and it is proposed to be able to bind nucleic acids (10). In the B. subtilis genome, ylxR is associated with the essential genes nusA and infB, respectively, encoding transcription terminating factor and translation initiation factor B in a single mRNA (11,12).
GI of PylxS by CshA. As sigX/M expression is induced by glucose, it was expected that ylxR expression driven by a major sigma factor, SigA, is also induced by glucose.
Indeed, we observed that the PylxS promoter, which drives the transcription of ylxR, was induced by glucose ( Fig. 2A, left). This GI was not dependent on glucose concentration (0.1 to 2%), as shown in Fig. S1A. A previous report showed that addition of glucose and glutamine to sporulation medium repressed expression of several genes regulated by the transition state regulators SigH and AbrB (13). The addition of glucose resulted in "overflow" metabolism, leading to a decrease in external pH. Buffering of this pH decrease by addition of Tris-HCl (pH 7.5) rescued the repression, which means external pH controls SigH/AbrB-mediated gene expression. Thus, we tested a similar possibility with respect to the mechanism of GI of PylxS: i.e., whether buffering the decrease in pH by addition of Tris-HCl (pH 7.5) affected the GI. We observed that glucose addition resulted in a pH decrease, and addition of Tris-HCl (pH 7.5) significantly buffered this pH decrease, as expected (Fig. S1B, right). However, we still observed GI of PylxS when the external pH was buffered (Fig. S1B, left), excluding external pH control of gene expression as a mechanism for this GI, probably because no glutamine was added. We note that in the ylxR disruptant, external pH similarly decreased to that in the wild-type strain, when glucose was added to the medium (data not shown).
In the cshA mutant, GI of sigX was abolished; thus, it is possible that cshA disruption may also abolish GI of PylxS (7) (Fig. 1A). In fact, when cshA was disrupted, the basal transcription levels of PylxS were reduced, and no GI was observed (Fig. 2B, left). Proteomic analysis of B. subtilis revealed that CshA, a DEAD box helicase, is acetylated at K244 and K296 (14). Protein acetylation often modifies protein function (15). We recently found acetylated CshA-dependent GI of genes for SigX and SigM (7). The GI of PylxS may also require CshA acetylation, like sigX/M, as the GI of PylxS was abolished in We then tested the effect of addition of two carbon sources to the mediumglycerol and succinate-at an equal molar concentration of glucose. The former is incorporated into glycolysis and the latter into the tricarboxylic acid (TCA) cycle (2). We observed induction of PylxS by glycerol, but not by succinate ( Fig. 2A, right). This observation is also consistent with the above hypothesis involving CshA acetylation, because the addition of glycerol but not succinate leads to increase in acetyl-CoA. Moreover, addition of 3% succinate had no effect on the PylxS expression, and 0.1% glucose induced this promoter (Fig. S1A), excluding the possibility that high osmotic pressure caused the GI of PylxS.
To test the acetylation hypothesis, we constructed a strain in which native cshA was disrupted but artificial induction of wild-type cshA was feasible from a xylose-inducible promoter. First we observed that cshA induction by addition of xylose complemented the expression and GI of PylxS in the cshA disruptant (Fig. S1D). Induction of a mutant form of cshA with K-to-R substitutions at two acetylated lysine residues did not complement the GI of PylxS. We made His-tagged versions of the wild type and the K-to-R mutant in B. subtilis and purified the proteins: similar amounts of the proteins were obtained, suggesting that the wild-type and mutant proteins had similar stability (data not shown). Collectively, these observations strongly suggest a requirement for CshA acetylation for the GI of PylxS.
GI of PylxS among the cell population. To observe GI of PylxS at the single-cell level, we constructed a transcriptional PylxS-gfp fusion and performed microscopic analysis. Since even in the presence of glucose, fluorescence of green fluorescent protein (GFP) was very low and observed from a part of the cell population in log phase (data not shown), the cells were observed in stationary phase. Sporulation is known to be under catabolite repression (16), and thus, only in the phase-contrast images of cells grown without glucose was the progress of spore formation observed (Fig. 2B). Among cells grown without glucose, GFP-positive cells were scarcely observed, while among cells grown with glucose, one-third were GFP positive. Based on these results, we concluded that under the conditions tested, glucose induced the expression of PylxS in a bistable mode, where the distribution of fluorescence is bimodal. Since PylxS drives transcription of two essential genes, it is possible that basal expression of PylxS occurs in cells without glucose. In fact, such basal expression was observed through the PylxS-lacZ fusion in the absence of glucose; using the PylxS-gfp fusion, it was also observed in nearly all cells in Luria-Bertani (LB) medium (see Fig. S3 in the supplemental material). We note that in a former report where a NusA-GFP translational fusion was analyzed, fluorescence seemed to be homogeneously detected from all the cells observed (17). The cause of this difference is not known; however, some experimental differences can be pointed out, such as the media used and the fusion construction strategy.
NAP-like protein YlxR. The structure of YlxR has been resolved and suggested its probable DNA/RNA-binding activity; however, this has not been experimentally verified (10). To test this possibility, we purified intact YlxR and then applied it to electrophoretic mobility shift assay (EMSA). As shown in Fig. 3A, YlxR was able to bind to the pET28b(ϩ) vector, suggesting that the DNA-binding activity of YlxR has low sequence Fluorescence of GFP and DAPI was visualized using WIB and WU filter sets (Olympus), respectively. Image processing and data analysis were performed using Adobe Photoshop CS5. Representative images are shown. Indicated merged photos are in pseudocolor. (C) The chromosomal structure of OAM817 is depicted (symbols as in Fig. 1). specificity. YlxR binding slightly preferred the supercoiled form of this vector DNA compared to its linear form (compare lane 2 to lane 7 in Fig. 3A).
YlxR is a small basic protein with DNA-binding activity that shows low levels of sequence dependency, which suggested that YlxR shares certain characteristics with some NAPs (3). To examine whether YlxR is associated with nucleoids in the cell, we constructed ylxR-gfp, where gfp is fused to the C terminus of ylxR, and observed its localization. Cells grown with glucose underwent microscopic observation, and the fluorescence derived from YlxR-GFP seemed to be associated with nucleoids, as was observed for Rok and other proteins associated with the nucleoid in B. subtilis (18,19) (Fig. 3B). The strain with the fusion at the original locus (Fig. 3C) grew in synthetic minimal medium (data not shown [see Fig. 6B]), suggesting that YlxR-GFP would be functional. We observed heterogeneous distribution of YlxR-GFP among the cells, as well as in the case of PylxS-gfp ( Fig. 3B and data not shown). Thus, we consider YlxR to be a NAP-like protein. We note that the apparent observed dissociation constant of YlxR for DNA (about 100 nM) is within the range reported for E. coli NAPs (20).
Transcriptome analysis of ylxR disruptant in the presence of glucose. It is known that many NAPs affect the transcriptome (3,5). To test this possibility for YlxR, we performed RNA-Seq analysis of the ylxR disruptant in the presence of glucose using the wild-type strain as the reference with three biological replicates of early-stationaryphase cells. Expression of 128 and 265 genes was, respectively, downregulated (Ͻ1/2) and upregulated (Ͼ2.0) ( Fig. 4; see Table S1 in the supplemental material).
Sporulation-related genes. Spo0A is a master regulator of sporulation initiation (21). In the ylxR disruptant, the expression of 10 Spo0A-activated transcription units (22 genes) was upregulated (from skfB at 8.9-fold to putB at 2.1-fold [see Table S2a in the supplemental material]). In the presence of glucose, YlxR represses the expression of two Spo0A-activated sigma genes required for sporulation initiation (sigE and sigF). It should be noted that increased SigE and SigF regulon expression in the ylxR disruptant leads to increased SigK and SigG regulon expression due to the so-called "sigma cascade" (21). Thus, the ylxR gene may be involved in the catabolite repression of sporulation.
Sigma regulons. Initial Tn mutagenesis showed that in the ylxR disruptant in the presence of glucose, the expression of sigX/M genes decreased. As expected, the expression of some SigX/M regulon genes was decreased ( Table S2a). The RNA-Seq analysis showed that expression of the motility-related SigD regulon and nitrogen starvation-regulated SigL-regulon increased in the ylxR disruptant in the presence of glucose without increased expression of the sigD and sigL genes themselves (Table S2a) (22). This may be due to some changes in the competition status of RNAP for sigma factors, from the point of view of the so-called "sigma cycle" (23). The changes would be caused by enhanced expression of sigE and sigF and reduced expression of sigM. (For an unknown reason, the expected decrease of sigX expression was not detected in the RNA-Seq analysis.) Metabolic genes. RNA-Seq analysis and subsequent LacZ analysis revealed differentially expressed genes involved in the synthesis of 12 amino acids (Arg, Asn, Cys, Glu, Gln, His, Ile, Leu, Met, Pro, Tyr, and Val) and 4 nucleotides (UMP, GMP, IMP, and AMP) in the ylxR disruptant in the presence of glucose (Fig. 5 to 7; see Fig. S4 in the supplemental material). Thus, YlxR may be a regulator for adaptation to the new metabolic state caused by glucose addition, especially of nitrogen metabolism, including biosynthesis/degradation of amino acids. Expression of several genes was confirmed using lacZ fusions. We newly identified GI or glucose repression (GR) of some genes ( Fig. 6 and 7; Fig. S4). The expression of PproBA is positively regulated by YlxR irrespective of glucose addition, because ylxR disruption severely reduced proBA expression (Fig. 6A, left). This decrease in proBA expression in the ylxR disruptant was complemented by artificial expression of ylxR from Pxyl, demonstrating the role of ylxR in the proBA expression (Fig. 6A, right). We had observed previously that proBA Genes whose expression decreased or increased in the RNA-Seq analysis of the ylxR strain in the presence of glucose are shown in blue and red, respectively. Genes in black in parentheses are those whose expression was not changed in RNA-Seq analysis; however, genes whose expression was not significantly changed in RNA-Seq but changed in lacZ analysis of the promoter for the gene are shown in black without parentheses. Genes not tested with lacZ analysis, but whose expression should be changed due to the gene position in the operon, are included in this category. Genes presented as consecutive characters indicate operons. # and $ indicate genes whose expression was not analyzed by lacZ analysis and were assigned to opposite categories (i.e., both up-and downregulation) in three replicates of RNA-Seq. F6P, fructose-6-phosphate; F1,6P, fructose-1,6-biphosphate; 3-PG, 3-phosphoglycerate; PEP, phosphoenolpyruvate; PYR, pyruvate; A-CoA, acetyl-CoA; ACP, acetylphosphate; PRPP, phosphoribosyl pyrophosphate; HCys, homocysteine; SAM, S-adenosyl methionine; PRE, prephenate; CIT, citric acid; 2-OG, 2-oxoglutarate; OAC, oxaloacetate. disruption led to proline auxotrophy (24); thus, it was expected that the ylxR disruptant would be unable to grow in synthetic minimal medium. This was the case, and the observation that the ylxQ disruptant was able to grow in this medium reinforced the role of ylxR in growth, but not the involvement of the downstream gene ylxQ (Fig. 6B). Proline addition restored the growth of the ylxR disruptant, which is consistent with the proline auxotrophy of the strain. We note that proJ expression increased in the ylxR mutant with glucose (Fig. 5); however, proJ expression was very weak (about 5% of that of proBA in the ylxR disruptant with glucose).
The ilvBHC-leuABCD operon, whose products are involved in synthesis of branchedchain amino acids, is regulated by the catabolite control protein A (CcpA) and CodY/ TnrA, which sense branched-chain amino acid/nitrogen availability, respectively (25,26). Expression of this operon is induced by glucose in a CcpA-dependent manner, and ylxR disruption further increased the expression irrespective of glucose addition (Fig. 7A, left). YlxR may function at PilvB cooperatively with CodY and TnrA. Thus, we examined the effect of ylxR disruption on this operon expression in codY or tnrA disruptants. In the tnrA disruptant, the enhancing effect of ylxR disruption was clearly observed, while in the codY disruptant, no effect of ylxR disruption was observed (Fig. 7A, middle and right). These results suggest that YlxR functions at this promoter in a CodY-dependent manner. YlxR regulates about 10% of the total genes in B. subtilis (Table S1), but 27% of the CodY regulon (14/51 operons) is regulated by YlxR, which is a notable difference in proportion (Table S1). However, only 16% of the TnrA regulon (6/37 operons) is regulated by YlxR. In a ccpA background, basal PilvB expression was very low (around 8 Miller units), but a significant effect of ylxR disruption was observed (data not shown). This might be consistent with the observation that 10% of the CcpA regulon transcription units belong to the YlxR regulon, suggesting no specific relationship between CcpA and YlxR.
Next, we observed that expression of the asnH-containing operon was significantly enhanced in the ylxR disruptant only in the presence of glucose, indicating ylxR negatively regulates this operon (Fig. 7B, left). The promoter of the asnH-containing operon is also known to be bound by CodY (27). Thus, we explored the possibility that at this promoter, YlxR may require CodY for its function. In a codY disruptant, expression of this operon was similarly enhanced by the introduction of the ylxR disruption both with and without glucose (Fig. 7B, right). Moreover, the enhancement ratio by the ylxR disruption was larger than that in the codY ϩ cells. Thus, codY disruption enhanced YlxR function, suggesting that CodY may weaken YlxR function. This role of CodY is different from the case in the ilv-leu operon, where YlxR could enhance CodY function, suggesting that YlxR works with CodY in a context-dependent manner. We note no regulatory relationship between transcription of codY and ylxR with each other (Table S1 and data not shown). It has been reported that there are three genes encoding asparagine synthesis enzymes in B. subtilis (28). Thus, we examined asnB expression in RNA-Seq and found that the decrease of asnB expression in the ylxR disruptant was on the threshold of being a significant change in gene expression (0.53-fold [ratio of the expression levels in the ylxR disruptant versus strain 168; designated here "ylxR/168"]). We confirmed substantial levels of decrease of fusion expression by using PasnB-lacZ (about 0.5-fold [ Fig. S4G]). Interestingly, the asnB gene is cotranscribed with the upstream gene metK, which is involved in methionine metabolism (Fig. 5).  To measure actual gene expression in the ylxR disruptant, we constructed a PgltAB-lacZ fusion and examined its ␤-Gal activity (Fig. S4E). Expression of the fusion was significantly decreased in the ylxR disruptant in the presence of glucose, indicating that YlxR positively regulates PgltAB. PgltAB is negatively regulated by glutamine-bound RocG (glutamate dehydrogenase) through sequestration of the positive transcription factor GltC (29) by RocG. rocGA expression is negatively regulated by CcpA in the presence of glucose (29). Thus, glucose addition finally results in the enhancement of PgltAB activity. rocGA is a candidate gene regulated by YlxR because the operon expression was reduced in RNA-Seq analysis, although the effect was not statistically significant. The disruption of ylxR counteracted glucose repression of ProcGA to some extent in lacZ analysis, leading to abolition of GI of PgltAB (Fig. S4A and E). According to the RNA-Seq analysis, the mRNA abundance of glnRA was also inconsistent among replicates (for glnA, 0.41-, 2.07-, and 1.24-fold [ylxR/168]). Thus, we undertook ␤-Gal analysis of PglnRA-lacZ (Fig. S4B). In the ylxR disruptant, PglnRA expression decreased in the presence of glucose.
The expression of the six-gene operon for glycolysis cggR-gapA-pgk-tpi-pgm-eno is repressed by CggR, and glucose addition resulted in GI of these genes (30). Other glycolysis genes include the pfk-pyk operon, whose expression was mildly induced by glucose (31), and we confirmed GI of the pfk-pyk operon (Fig. S4D). Moreover, we observed that ylxR disruption significantly weakened this GI, suggesting the involvement of YlxR in regulation of this operon. Furthermore, ylxR disruption increased the expression of five genes encoding redox enzymes in the electron-transport system (ctaEDFG and qcrB) (Table S1). Finally, ylxR disruption affected five genes involved in the synthesis of four cofactors (Table S2b).

DISCUSSION
The initial goal in this study was to identify causes for GI of sigX/M. Glucose-induced YlxR would change the competition status of several sigma factors for binding to the RNAP core enzyme, maybe leading to abolition of GI of sigX/M. However, more interestingly, we identified a new glucose-responsive system that includes protein lysine acetylation of CshA ( Fig. 1A; Fig. S1D), bistable expression of NAP-like protein YlxR (Fig. 2B), and transcriptional regulation of metabolic genes by YlxR (Fig. 5). Some NAPs have a role in nutrient-responsive transcription regulatory networks (3). Thus, this work enriches knowledge about such NAPs, by showing that YlxR plays a role in a glucose-responsive transcription network. We observed that the effects of the ylxR disruption were lost or strengthened in the codY disruptant, depending on the target gene, suggesting some relationship or interaction between YlxR and CodY at the promoter region of relevant genes.
Single-cell analysis of PylxS-gfp revealed the bistable mode of its expression: i.e., in some cells GFP was produced, while in the other cells, little or no GFP was produced. Typically, the bistable expression of genes is generated by positive-feedback regulation (32). In the case of PylxS, NusA-dependent negative-feedback regulation was reported (12), which, however, does not generate bistable expression. Glucose-induced and ylxR-dependent expression of some metabolic genes was observed, such as gltAB, pfk-pyk, and proBA. Recently, the customary view of metabolic gene expression as homogeneous has been challenged because bistable or heterogeneous metabolic gene expression has been reported (33)(34)(35). Thus, it is an interesting question whether the observed enhanced expression is universal or heterogeneous within the cell population under glucose-rich conditions due to the bistable expression of ylxR.
Several NAPs have been identified in B. subtilis (HBsu, LrpC, and Rok). However, knowledge of the impact of the two former proteins on the transcriptome is not comprehensive (18,(36)(37)(38). HBsu and LrpC regulate their own genes (39,40). Rok regulates the competence master regulator gene comK directly and the biofilm-related bslA gene indirectly (41,42), in addition to other genes, including the mobile and foreign genetic element genes in ICEbs1. However, Rok only regulates 39 genes (41). In contrast, YlxR has a profound impact on the B. subtilis transcriptome, although the YlxR-regulated gene list contains genes regulated by direct DNA binding and through indirect effects. We note that YlxR downregulates many SPbeta phage genes like Rok does ICEbs1 (23 genes in Table S1). However, many such genes are missing from Table S1 since the levels of the phage gene expression were very low, leading to large fluctuations in expression and resultant high P values in our analysis. YlxR is widely conserved in eubacteria-for example, in species belonging to the most deeply rooted phylum, Aquificae (Desulfobacteraceae in Fig. S2). Thus, it will be interesting to determine whether YlxR also functions as an NAP regulating metabolic gene expression in other eubacteria.
YlxR affects regulation of many metabolic genes, especially in the presence of glucose. However, this does not straightforwardly lead to YlxR-mediated changes of the cellular metabolome in response to glucose-rich conditions. After the transcription of genes encoding metabolic enzymes, the corresponding mRNA must be translated into protein, which may be further modified, for example, by phosphate and/or acetyl moieties (43)(44)(45)(46)(47). Moreover, catalytic activity of enzymes is regulated by allosteric binding of end products and/or in vivo substrate concentrations in addition to posttranslational modifications. However, in the transition between carbon sources (glucose and malate), changes of transcription levels of Ͼ2,000 genes were observed (48). YlxR may play a critical role in such changes because of the observed large changes in the transcriptome of the ylxR disruptant. In addition, metabolic gene regulation is an important factor determining cellular metabolic state, because metabolic gene regulation underlies the rapidly changing metabolome that responds to extracellular environments, including nutritional status. Thus, YlxR is an important factor for adaptation of B. subtilis cells to a glucose-rich environment.

MATERIALS AND METHODS
Strains and media. All B. subtilis strains used in this study are listed in Table 1 and in Table S3 in the supplemental material. One-step competence medium (MC) (49), Schaeffer's sporulation medium (SM) (16), LB medium (Difco, Lennox), and Spizizen's minimal medium (50) were used. Antibiotic concentrations were described previously (51,52). Synthetic oligonucleotides were commercially prepared by Tsukuba Oligo Service (Ibaraki, Japan) and are listed in Table S4 in the supplemental material.
Strain construction. To construct a strain carrying the amyE::PylxS-gfp fusion (OAM818), first the gfp-amyE[front] unit was PCR amplified from strain OAM-N41 carrying amyE::PtapA-gfp using the oligonucleotides gfp(SD)-F/amyE-FF (53). Second, the amyE[back]-Cm r -PylxS unit was PCR amplified using the oligonucleotides amyE-RR/PylxR-(SD)-gfp-R from a ligated reaction mixture of a PCR product amplified from chromosomal DNA using the oligonucleotides ylxR-Eco/ylxR-Hin treated with EcoRI/HindIII and pIS284 treated with the same enzymes (54). These fragments were combined in a final PCR using the oligonucleotides amyE-FF/amyE-RR. Final PCR products were transformed into B. subtilis 168. The ylxQ::Tc r unit in OAM840 was constructed using PCR. Briefly Tc r from pBEST304 (55) and the upstream and downstream regions of ylxQ with overlapping regions to Tc r were amplified using primers listed in Table S4 and then combined by PCR. The unit was transformed into B. subtilis 168. Total DNA was taken from the resultant Tc r strain for PCR-based confirmation of the expected chromosomal structure.
Plasmid construction. The plasmids used in this study are listed in Table 1 and Table S3. For PCR, chromosomal DNA was used as the template. To construct pX-ylxR, the PCR product was amplified using the oligonucleotides pX-ylxR-Spe/pX-ylxR-Bam, digested by SpeI/BamHI, and cloned into pX treated with SpeI/BamHI (56). To construct pDG1729-ylxS, the PCR product amplified by using the oligonucleotides ylxR-Eco/ylxR-Hin was digested with EcoRI/HindIII and cloned into pDG1729 treated with the same enzymes (57). To construct pTYB11-ylxR, PCR product amplified by using the oligonucleotides ylxRchitin-F (Sap)/ylxR-chitin-R (Xh) was digested with SapI/XhoI and cloned into pTYB11 treated with the same enzymes (New England Biolabs, Ipswich, MA). To construct pSG1194-ylxR, gfp amplified from pMF20 using the oligonucleotides gfp-F/gfp-Xba-R and the initial PCR product with genomic regions overlapping the gfp gene amplified using ylxR-gfp-F/ylxR-gfp-R were combined via a second PCR using ylxR-gfp-F/gfp-Xba-R (58). The final PCR product was digested with XbaI and BamHI and cloned into the large fragment of plasmid pSG1194 lacking the Discosoma sp. red fluorescent protein DsRed, which was obtained by treatment with the same restriction enzymes, thereby generating the plasmid of interest (59).
Purification of YlxR. E. coli strain ER2566 bearing pTYB11-ylxR was grown in 600 ml of LB medium (100 g/ml ampicillin) at 30°C. At an optical density at 600 nm (OD 600 ) of Ϸ0.8, 0.2 mM IPTG was added, and cells were further incubated for 20 h at 20°C. Chitin-binding domain-and intein-fused YlxR was purified by using chitin-coupled resin and then autoactivating the intein with dithiothreitol (DTT) according to the manufacturer's recommendations (New England Biolabs). After SDS-PAGE analysis of the fractions, the purified protein was dialyzed against buffer containing 10 mM Tris-HCl (pH 8.0), 100 mM KCl, 10 mM MgCl 2 , 1 mM DTT, and 10% glycerol. Aliquots of purified protein were stored at Ϫ80°C.
Electrophoretic mobility shift assay. Purified YlxR was added to the same buffer used for dialysis containing the DNA probe in a final volume of 12 l. Immediately after adding the protein, the reaction mixture with 2 l of loading buffer (40% glycerol, 1ϫ Tris-acetate-EDTA, and 2 mg/ml bromophenol blue) was applied to a 1% agarose gel, and electrophoresis was performed in Tris-acetate-EDTA buffer. DNA was detected with UV light.
Microscopic observations. Cells were grown in SM with or without 2% glucose, 100 l of the culture was centrifuged, and 80 l of the supernatant was removed. The cells were then resuspended in the remaining 20 l. Portions (2 l) of each sample were mounted on glass slides treated with 0.1% (wt/vol) poly-L-lysine (Sigma). If necessary, 4=,6-diamidino-2-phenylindole (DAPI) solution (1 mg/ml in water) was added to the cell suspension at 1 g/ml. Microscopy was performed with an Olympus BX51 phasecontrast and fluorescence microscope with a 100ϫ Plan-N objective (Olympus, Tokyo, Japan). Images