INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

myo-Inositol is abundant in nature, especially in soil. Various microorganisms, including Bacillus subtilis, are able to grow on myo-inositol as the sole carbon source. The B. subtilis iol divergon consisting of the iolABCDEFGHIJ and iolRS operons is involved in inositol catabolism (29, 31). These two iol operons are induced upon the addition of inositol to the medium (29). This induction is negatively regulated by a repressor (IolR) belonging to the DeoR family through its unique interaction with the extended binding regions close to their promoters (31). Inositol dehydrogenase encoded by iolG catalyzes the first reaction of inositol catabolism by B. subtilis (19, 29). The synthesis of this enzyme induced by inositol was repressed on the addition of glucose to the medium (18, 29). Very recently, DNA microarray analysis implied that not only the expression of iolG but also that of the other 11 iol genes was under glucose repression (30).

The well-characterized mechanisms underlying glucose repression are those of catabolite repression and inducer exclusion. Recently, the mechanism underlying catabolite repression in B. subtilis was extensively investigated. These studies revealed that Bacillus, as well as other low-GC gram-positive bacteria, possesses a negative regulatory mechanism for catabolite repression, which is very different from the positive regulatory mechanisms of enteric bacteria involving cyclic AMP and cyclic AMP receptor protein (20, 25). In low-GC gram-positive bacteria, negative regulation of the transcription of catabolite-repressive genes occurs through the binding of the catabolite control protein (CcpA) (10), which interacts with allosteric effectors such as P-Ser-HPr (6) and P-Ser-Crh (7), to their cis-acting catabolite-responsive elements (cre's) (15).

DNA microarray analysis revealed that the expression of the iolABCDEFGHIJ and iolRS operons was under glucose repression, which was partially CcpA dependent (30). The glucose repression of the synthesis of inositol dehydrogenase (Idh) is dependent on both CcpA and IolR (8, 16, 29, 30), implying that catabolite repression and the induction system mediated by IolR are involved in glucose repression of the iolABCDEFGHIJ operon. Since almost no glucose repression was observed in a doubly mutated strain with respect to ccpA and iolR (30), IolR-independent repression is likely to be exerted through catabolite repression mediated by CcpA. When cre sequences were searched for in the B. subtilis genome using a query sequence of WTGNAANCGNWNNCW (W and N stand for A or T and for any base, respectively), 126 putative and known cre's were found (15). One of them is located within the iolB gene, that is, cre-iolB, which has been found to function as a cre in an in vivo cre test system (15).

In this work, we found on deletion and base substitution analysis of the iol region that cre-iolB, which is located far downstream of the iol promoter (Piol), functioned as a cre. Further deletion and base substitution analysis revealed another functional cre, which is located between Piol and iolA and named cre-iiolA; the prefix "i" indicates the location of cre in the intergenic region. Thus, the CcpA-dependent catabolite repression of the iolABCDEFGHIJ operon was due to the two cre's functioning independently. Interestingly, the sequence of cre-iiolA does not match the 3' part of the query sequence used for the genome-wide cre search. cre-iiolA was found to belong to a group of cre's exhibiting similar but distinct consensus sequences.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The B. subtilis strains used are listed in Table 1. Strains FU758, FU759, and FU760 carrying iolR::cat were obtained by the transformation of strains 168, QB5223, and QB7096 with DNA of strain YF244 to chloramphenicol (5 µg/ml) resistance on tryptose blood agar base (TBAB) plates (Difco), respectively. Strain FU761 was obtained by the transformation of strain FU759 with DNA of strain QB7096 to kanamycin (10 µg/ml) resistance on TBAB plates, because strains such as strain QB7102 carrying both the ptsH1 and crh::aphA3 mutations were not transformable.

Construction of strains for deletion analysis of cre's of the iol operon. To construct strains FU709 and FU713 carrying iolR::neo and transcriptional fusions of iol regions (nucleotides [nt] 107 to +2270 and +2474, respectively; +1 is the transcription initiation nucleotide of the iol operon) to lacZ, we first replaced the EcoRI site of plasmid pCRE-test (15) with an XbaI site, which is absent from the iol region (nt 107 to +2474). This was done through its digestion with EcoRI, blunt ending, attachment of an XbaI linker, digestion with XbaI, and subsequent self-ligation, resulting in plasmid pCRE-test2.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Location of cre-iiolA and cre-iolB. The upper part of the figure shows the sequences of cre-iiolA and cre-iolB, which are aligned with a consensus sequence (15). The base substitutions of the indicated bases in the cre sequences which caused the knockout of their function are shown; nt +1 is the iol transcription initiation base. Strains carrying a series of deletion derivatives of Piol (iol promoter)-iolAB'-lacZ in the amyE locus, used for identification of iol-cre's, were constructed as described in the text except for the following. The primers for amplifying the respective iol regions by PCR for construction of strains FU709, FU713, FU719, FU720, FU721, FU722, FU723, and FU724 were a forward one (nt 105 to 85) with an adapter sequence of GTCCTCTAGA and reverse primers (+2448 to +2474, +2250 to +2270, +1173 to +1192, +693 to +712, +279 to +298, +91 to +110, +58 to +82, and +27 to +47) with an adapter sequence of GATAGGATCC; the underlined sequences are XbaI and BamHI sites, respectively.

Base substitutions in cre sequences within the Piol-iol-lacZ fusions. Strains FU715 and FU716 carrying base substitutions of +2405GT and +2410CT in the cre-iolB sequence (Fig. 1) within the Piol-iolAB'(nt 107 to +2474)-lacZ fusion of strain FU709 were constructed as follows. Introduction of base substitutions was performed by means of recombinant PCR using DNA of strain 168 and the following primers, as described previously (31). The common upstream and downstream primers were 5'-GTCCTCTAGACCTTCTCTTACTTCTCTTACTTG-3' (the XbaI site is underlined) and 5'-GATAGGATCCTCATTTAATTAGTAGGATGTGTATCCG-3' (the BamHI site is underlined), respectively. The overlapping primers for +2405GT were 5'-GGGAAATGAAAACTTTGTCATCG-3' and 5'-CGATGACAAAGTTTTCATTTCCC-3', and those for +2410CT were 5'-AACGTTGTTATCGTTCCTGCGG-3' and 5'-CCGCAGGAACGATAACAACGTT-3' (each substituted base is underlined). The resultant recombinant PCR products were digested with XbaI and BamHI and then ligated with the XbaI-BamHI arm of plasmid pCRE-test2. The resultant plasmids were linearized with PstI and then integrated into amyE of strain FU704 through a double-crossover event, resulting in strains FU715 and FU716 being produced.

Strain construction for further functional analysis of cre-iiolA and cre-iolB. The respective iol regions containing cre-iiolA and cre-iolB (nt +63 to +121, and +2375 to +2430) were amplified by PCR using DNA of strain 168 and the following two primer pairs to generate flanking BamHI sites. For amplification of the cre-iiolA region, the upstream and downstream primers were 5'-GATAGGATCCCGCCATTTATTTTTTTGGTG-3' (nt +63 to +82) and 5'-GATAGGATCCCCACTTTTCAGCAAGCCAAG-3' (nt +102 to +121), and for that of the cre-iolB sequence, they were 5'-GATAGGATCCGACGAGACAATGACTGTGGG-3' (nt +2375 to +2394) and 5'-GATAGGATCCTGGTATCCCGCAGGAACGAT-3' (nt +2411 to +2430) (the respective BamHI sites are underlined). The PCR products were digested with BamHI and then cloned into the BamHI site of plasmid pCRE-test (15). The ligated DNA was used for the transformation of E. coli strain JM109 to ampicillin resistance. After confirming the correct orientations and sequences of the Pspac-cre-lacZ fusions in the resulting plasmids pCRE-iiolA and -iolB by sequencing, they were linearized with PstI and used for the double-crossover transformation of strains GM122, 168, QB5223, QB7096, 1A250, and 1A147 to chloramphenicol resistance. The respective resultant strains (FU738, FU726, FU728, FU748, FU742, and FU743) carried the Pspac-(cre-iiolA)-lacZ fusion in their amyE locus, whereas the other strains (FU727 from 168, FU729 from QB5223, FU749 from QB7096, FU744 from 1A250, and FU745 from 1A147) contained the Pspac-(cre-iolB)-lacZ fusion. Strains FU750 and FU751 were obtained by the transformation of strains FU728 and FU729 with DNA of strain QB7096 to kanamycin (10 µg/ml) resistance on TBAB plates, respectively.

View larger version (26K): [in this window] [in a new window]   FIG. 2.   Alignment of the cre-iiolA sequence with known cre sequences that were not revealed with a query sequence of WTGNAANCGNWNNCW. The upper part of the figure shows the knockout substitutions of the cre-iiolA sequence. The sequences of five known cre sequences that were not revealed with a query sequence of WTGNAANCGNWNNCW (15) are aligned with that of cre-iiolA; the known cre's are cre-ixynB (7), cre-iaraA (22), cre-iaraE (23), cre-ilevD (12), cre-iacuA (9), and cre-iacoR (1). The consensus sequence carries two thymines (TT) in the consensus sequence, which are indicated with a dot in this alignment. R represents A or G, and Y represents C or T.

Idh and -Gal assay. Cells were grown to an absorbance level at 600 nm (A₆₀₀) of 0.6 in S6 medium (5) containing 0.5% Casamino Acids (Difco) and supplemented with required amino acids (50 µg/ml) with and without a 10 mM concentration of inositol and/or glucose. In addition, neomycin (15 µg/ml), chloramphenicol (5 µg/ml), and kanamycin (5 µg/ml) were added to the media for the growth of strains carrying iolR::neo, iolR::cat or with cat integration into amyE, and crh::aphA3, respectively. The cells (A₆₀₀ unit = 3.6) were harvested and then lysed by lysozyme treatment and brief sonication (18). Idh activity in crude cell lysates was spectrophotometrically assayed as described previously (18). -Galactosidase (-Gal) activity in crude cell extracts was spectrophotometrically assayed as previously described (15). The amounts of protein in cell extracts were determined by the method of Bradford (3) with bovine serum albumin as a standard.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

B. subtilis genes involved in glucose repression of Idh synthesis. Glucose repression of the synthesis of Idh encoded by iolG is known to occur through catabolite repression mediated by CcpA (8, 16, 30) as well as a regulation involving IolR, probably through inducer exclusion (29, 30). We first investigated the effects of the ptsH1 mutation causing the replacement of Ser46 of HPr with alanine and crh::aphA3 disruption on glucose repression of Idh synthesis. As shown in Table 2, Idh synthesis was severely repressed by glucose in strain 168trpC2 (wild type) (repression ratio, >461). Idh synthesis was not relieved from this glucose repression in strain QB7096 (crh::aphA3) at all (repression ratio, >470), but it was partially relieved from catabolite repression in strain QB5223 (ptsH1) (repression ratio = 79). Idh synthesis was more relieved from glucose repression in strain QB7102 carrying the ptsH1 and crh::aphA3 mutations (repression ratio = 7.3) and strain 1A147 carrying the ccpA1 mutation (repression ratio = 4.9). Moreover, this CcpA-independent glucose repression observed in strain 1A147 (ccpA1) was almost completely abolished in strain FU707 carrying the ccpA1 and iolR::neo mutations (repression ratio = 1.1) (Table 2). Part of the results are consistent with those reported by Galinier et al. (8), who found that Crh involvement in catabolite repression of Idh synthesis was solely observed in the ptsH1 background, because HPr alone is likely to be sufficient to cause this catabolite repression. However, it was reported that either the ptsH crh double mutant or the ccpA::spec mutant was completely relieved from glucose repression of Idh synthesis (8). We cannot explain this discrepancy properly, but it might possibly be due to the difference in the media for cell cultivation; we used S6 (this work; 16, 30) or DSM medium (30), whereas they used CSK medium (8).

Actual involvement of cre-iolB in catabolite repression of the iol operon. Upon a search for cre sequences in B. subtilis, 126 putative and known cre's were revealed (15). Among them, cre-iolB was found to function as a cre in an in vivo cre test system (15). The iol operon is most likely transcribed from only one promoter, Piol (29). Hence, to determine whether or not cre-iolB, which is located approximately 2,400 bp downstream of Piol, is actually involved in the catabolite repression of the iol operon, we constructed a transcriptional fusion, Piol-iolA-iolB'-lacZ, possessing an iol region (nt 107 to +2474) with cre-iolB (nt +2397 to +2411), which expresses lacZ under the direction of Piol (Fig. 1), and integrated it into the chromosomal amyE locus of strain FU704 carrying iolR::neo. The resulting strain, FU709, produced a high level of -Gal constitutively even on growth without inositol, which was repressed 3.5-fold on the addition of glucose to the medium (Table 3).

Identification of another cre for catabolite repression of the iol operon. To find another cre for the iol operon, which might be located between nt 107 and +2270, we further constructed a series of deletion derivatives of strain FU713, from nt +2270 toward the 5' direction in the Piol-iolA-iolB'-lacZ fusion (Fig. 1), and determined the catabolite repression ratios of -Gal synthesis (Table 3). The catabolite repression ratios of -Gal synthesis (2.3 to 2.4) obtained with strains FU719, FU720, FU721, and FU722 carrying iol regions from nt 107 to nt +1192, +712, +298, and +110, respectively, were almost the same as that with strain FU713 carrying one from nt 107 to +2270 (2.4) (Table 3), indicating that no other cre is located in the iol region between nt +111 and +2270 covering iolA and iolB'. However, -Gal synthesis was completely relieved from catabolite repression in strains FU723 and FU724 carrying iol regions from nt 107 to +82 and +47, respectively (repression ratio = 0.9). These results suggest that another cre for the iol operon is located between nt +83 and +110.

Functional analysis of cre-iiolA by means of base substitutions. Fig. 2 lists several cre's, such as cre-ixynB (7), cre-iaraA (22), cre-iaraE (23), cre-ilevD (12), cre-iacuA (9), and cre-iacoR (1), which have been reported to function or proposed to function as cre's but were not revealed in our previous cre search (15). Interestingly, the last C of the consensus sequence of WTGNAANCGNWNNCW is not conserved in the sequences of these cre's as well as cre-iiolA (Fig. 2), nevertheless the substitution of this C abolished the cre function almost completely, as in the cases of CT and G in cre-iamyE (27), CT in cre-gntR (6), CT in cre-hutP (28), and CT in cre-iolB (Table 3). Instead, the WN bases underlined in the above sequence are TT in all of these cre sequences (Fig. 2). Thus, we examined whether or not the positioning of the TT bases is essential for cre-iiolA function.

Analysis of HPr and Crh involvement in catabolite repression exerted through cre-iiolA and cre-iolB Not only HPr but also Crh is involved in catabolite repression of Idh synthesis (Table 2) (8), which is likely exerted through cre-iiolA and cre-iolB (Table 2). The sequence of cre-iiolA was found to be distinct from that of cre-iolB (Fig. 2). So, we examined the HPr and Crh involvement in the catabolite repressions exerted through cre-iiolA and cre-iolB. We constructed a series of isogenic strainsFU726 (wild type), FU728 (ptsH1), FU748 (crh::aphA3), and FU750 (ptsH1 crh::aphA3)carrying the Pspac-(cre-iiolA)-lacZ fusion in the amyE locus and another series of isogenic strainsFU727 (wild type), FU729 (ptsH1), FU749 (crh::aphA3), and FU751 (ptsH1 crh::aphA3)carrying the Pspac-(cre-iolB)-lacZ fusion in this locus.

View this table: [in this window] [in a new window]   TABLE 5.   Effects of the ptsH, crh and ccpA mutations on catabolite repression of -Gal synthesis under the control of the spac promoter and cre-iiolA or cre-iolB

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Glucose repression of the iol operon is known to be exerted through catabolite repression mediated by CcpA and a regulation system involving IolR, probably through inducer exclusion (30). We investigated the catabolite repression of the iol operon under experimental conditions where the involvement of IolR in its repression was eliminated. Deletion and base substitution analysis allowed us to identify two cre's of the iol operon (cre-iiolA and cre-iolB) (Fig. 1 and Tables 3 and 4). cre-iiolA is located between the iol promoter and the iolA gene (nt +86 to +100), while cre-iolB is in iolB (nt +2397 to +2411). The presence of two cre's has been reported in the gnt (14), ackA (26), ara (22), and rbs (15, 24) operons. Our previous in vivo results (14) implied that catabolite repression exerted by cre-igntR (cre_up) was probably independent of that exerted by cre-gntR (cre_down). This study also revealed that cre-iiolA and cre-iolB likely function independently (Table 5). In addition, cre-iolB was found to function at the original location (nt +2397 to +2411), i.e., far downstream of the transcription initiation site (Table 3). Although two cre's of each of the other operons have not been characterized well, it is notable that the relative locations of the two cre's of the ara and rbs operons are very similar to those of iol. These results are well consistent with the idea that the mechanism underlying catabolite repression can be explained by a transcription roadblock if cre is located downstream of the transcription initiation site (6, 13).

Our previous genome-wide cre search using a query sequence of WTGNAANCGNWNNCW (15) failed to reveal cre-iiolA as well as six known cre's (cre-ixynB [7], cre-iaraA [22], cre-iaraE [23], cre-ilevD [12], cre-iacuA [9], and cre-iacoR [1]), although it revealed 126 putative and known cre's, including cre-iolB. Alignment of the sequences of the six cre's not revealed by the search led to another consensus sequence of WTGAAARCGYTTWWN (Fig. 2). The 5' part of this consensus sequence perfectly coincides with the previous query sequence, but the 3' one does not match it well. It is notable that the 3' part of the latter consensus sequence includes conserved TT but is devoid of the last CW of the former consensus sequence. Actually, base substitution analysis of the TT of cre-iiolA indicated that they are indispensable for its function (Fig. 2 and Table 4). Although the sequence of cre-iiolA appeared to be distinct from that of cre-iolB in conserved bases, a protein complex of CcpA with either P-Ser-HPr or P-Ser-Crh likely binds to both cre's to similar extents (Table 5).

Recent analysis of B. subtilis cre sequences led to the following three conclusions (15). (i) Lower mismatching of cre sequences with the query sequence (WTGNAANCGNWNNCW) is required for cre function. (ii) Although cre sequences are partially palindromic, lower mismatching in the same direction as that of transcription of the target genes is more critical for cre function than that in the inverse direction. (iii) Yet, a more palindromic nature of cre sequences is desirable for a better function. Comparison of the above two consensus sequences also implied that the 5' part of cre sequences should be well conserved for their efficient function and that a protein complex of CcpA with P-Ser-HPr or P-Ser-Crh recognizes this part. The last CW (preferably CA) of the query sequence of WTGNAANCGNWNNCW is likely to be required for pairing with TG, resulting in proper binding of the complex. However, this pairing might be compensated for by another pairing of TT of the second consensus sequence of WTGAAARCGYTTWWN with AA. Of course, both pairings appear to be more desirable for efficient cre function.

We searched for cre sequences in the B. subtilis genome with the currently deduced consensus sequence of WTGAAARCGYTTWW through the DNA pattern search program of the SubtiList Web Server (http://genolist.pasteur.fr/SubtiList/). This search revealed 14 more putative cre's without any mismatch: cre-iybgJ, cre-iycbF, cre-iyceK, cre-iydaA, cre-yebB, cre-iopuE, cre-islp, cre-iyqkI, cre-levR, cre-iysfC, cre-yusL, cre-iyvbQ, cre-iyvfK, and cre-iyycE. Our present study implies that a genome-wide search for certain cis-acting elements with a single query sequence will not reveal most of the elements in question because of additional features due to their secondary structure, such as a palindromic nature (this work) and sequence periodicity (31). In the case of a cre search, a genome-wide search with these two query sequences appears to be able to reveal almost all of them.