INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Acetic acid is one of the major by-products of carbon metabolism detectable during the growth of Bacillus subtilis in rich media. Two metabolic pathways are known to be involved in the conversion of pyruvate to acetate. The first, with acetyl coenzyme A (acetyl-CoA) and acetyl phosphate (acetyl~P) as intermediates, is considered to be the main source of acetate excretion during the exponential growth of B. subtilis in the presence of an excess of carbohydrates (14, 38). The second operates during the stationary phase, and acetate is generated via butanediol (38). Overacidification of the medium due to pyruvate and acetate accumulation during exponential growth is prevented by the conversion of pyruvate to uncharged acetoin (aerobic conditions) or uncharged 2,3-butanediol (anaerobic conditions), both of which are also excreted (38). During late exponential growth phase, acetoin undergoes oxidative dissimilation to acetaldehyde and finally to acetate (23). Acetate and acetoin excreted into the growth medium can be reused during stationary growth phase when other carbon sources have been depleted and can therefore be regarded as carbon storage compounds (16). The acuABC genes and the aco operon have been shown to be involved in acetoin utilization during growth and sporulation (14, 17a), whereas acetyl-CoA synthetase, the product of the acsA gene, is responsible for acetate utilization (16).

This study focused on the phosphotransacetylase (pta) gene, which together with the acetate kinase (ackA) gene encodes the enzymes that catalyze the conversion of acetyl-CoA to acetate via an acetyl~P intermediate. In most bacteria, these two genes are organized into a single operon, and they have been shown to be involved in the maintenance of the intracellular acetyl-CoA and acetyl~P pools (1, 2, 22, 32, 37). As in Mycoplasma genitalium (6), the B. subtilis genes pta and ack are located at distant loci on the chromosome: at 326° (pta) and 263° (ack). CcpA (catabolite control protein A), a member of the LacI-GalR family of repressors, is a key regulator of carbon flow in B. subtilis. It acts as either a negative regulator of the expression of carbon utilization genes or a positive regulator of the expression of genes involved in the excretion of excess carbon, such as ackA (17, 19). CcpA mediates glucose control by binding to the DNA operator sequence known as the catabolite response element (CRE) (17, 42). The phosphoenolpyruvate:sugar phosphotransferase system (PTS) is responsible for the uptake of various sugars in bacteria (31) and is also involved in carbon catabolite repression (CCR) in B. subtilis. The phosphocarrier protein HPr and its homologue Crh (catabolite repression HPr) are both phosphorylated at the regulatory Ser-46 site by the ATP-dependent HPr kinase (10, 11). This enzyme is stimulated by glycolytic intermediates such as fructose-1,6-bisphosphate (FBP). Replacement of Ser-46 with alanine, which abolishes the ATP-dependent phosphorylation of HPr (ptsH1 mutant), or additional disruption of the crh gene (ptsH1 crh::aphA3 double mutant) causes the partial or complete relief from CCR of several systems (5, 10). This finding indicates that in addition to CcpA, both Ser-46-phosphorylated HPr (HPrSer-P) and Ser-46-phosphorylated Crh (CrhSer-P) are involved in carbon catabolite control.

An interaction between CcpA and HPrSer-P has been demonstrated in vitro, and in some cases this interaction has been found to be stronger in the presence of glycolytic intermediates such as FBP (4). The resulting protein complex interacts specifically with the CRE of the gluconate, xylose, -xylosidase, and levanase operons (8, 9, 13, 24a). Glucose-6-phosphate also stimulates the binding of CcpA to the CRE sequence of the xyl and gnt operons (13, 28). While binding of CcpA to the CRE of the -amylase gene has been observed in the presence of the corepressor FBP, NADP has been shown to stimulate CcpA-dependent inhibition of transcription from the amyE promoter (20).

The mechanism of CCR has been extensively studied in B. subtilis, but much less is known about catabolite activation of gene expression. For the ackA gene of B. subtilis (15), catabolite activation was abolished in a ccpA mutant, in a ptsH1 crh double mutant, or after removal of the second of the two CRE sequences identified upstream from the ackA promoter (41). CcpA is also required to induce the expression of the alsSD operon involved in acetoin biosynthesis (34).

In this study, we demonstrate that the pta (formerly ywfJ) gene, sequenced in the framework of the B. subtilis genome project (12), encodes a phosphotransacetylase. We also report the identification of the pta promoter and of a regulatory CRE sequence upstream from this promoter. Our results show that pta (like ackA and alsSD) is a CcpA-activated gene involved in excess carbon excretion pathways in B. subtilis. In addition, we show that both HPrSer-P and CrhSer-P are involved in the catabolite activation of pta expression.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and growth conditions. The B. subtilis strains used in this study are listed in Table 1. Escherichia coli XL1 Blue and TG1 were used for plasmid preparations. B. subtilis and E. coli strains were grown in Luria-Bertani (LB) medium or in sporulation medium. Antibiotics were added at the following concentrations: ampicillin, 100 µg ml¹; chloramphenicol, 5 µg ml¹; kanamycin, 5 µg ml¹; and spectinomycin, 60 µg ml¹. The transformation procedures used for E. coli and B. subtilis were as described by Sambrook et al. (35) and Kunst and Rapoport (21), respectively. For carbon regulation studies, B. subtilis strains containing the pta-lacZ fusion were grown in CSK medium (24) in the absence or presence of 0.4% glucose. For Pta activity assay, B. subtilis strains were grown on LB medium supplemented with 1% glucose. -Galactosidase specific activity in mutants containing the various lacZ fusions was measured as previously described (39). One unit of -galactosidase was defined as the amount of enzyme producing 1 nmol of o-nitrophenol per min at 28°C. Protein concentrations were determined by using a Bio-Rad protein assay kit. The integration of DNA fragments into the amyE locus of B. subtilis by double crossover was assessed by monitoring the loss of amylase activity on tryptose blood agar base (Difco) supplemented with 10 g of hydrolyzed starch (Connaught) per liter. Starch degradation was detected by sublimating iodine onto the plates.

DNA and RNA manipulations. Standard procedures were used to extract plasmids from E. coli (35). Restriction enzymes, phage T4 DNA polymerase, phage T4 DNA ligase, and T4 polynucleotide kinase were used as recommended by the manufacturers. The PCR products were purified by using a Qiaquick kit (Qiagen). RNA was extracted from B. subtilis 168 grown in CSK medium with or without 0.4% glucose and harvested at an optical density at 600 nm (OD₆₀₀) of 0.2. RNA was extracted as previously described (3). The ³²P-labelled oligonucleotide 5'CAATTTTAACGTCTTTTCCAGCTAC3' (labelled with T4 polynucleotide kinase; Biolabs) was used to map the pta promoter by primer extension. DNA was sequenced by the dideoxy-chain termination method of Sanger et al. (36).

Plasmid constructs. The main characteristics of the plasmids used in this study are shown in Fig. 1. A 6,187-bp region of the B. subtilis chromosome containing the pta (formerly ywfJ) gene (12) was inserted into the E. coli plasmid pDIA5304 to give pDIA5373. pDIA5374 was obtained by inserting a BamHI-BglII DNA fragment containing a kanamycin resistance cassette into the unique BglII restriction site of pDIA5373. The EcoRI-BglII and HindIII-BglII fragments from pDIA5373 were inserted into the integrative plasmid pJM783 (30), giving pDIA5375 and pDIA5376, respectively (Fig. 1). These two plasmids were integrated by Campbell recombination into the B. subtilis pta locus, producing a transcriptional pta-lacZ fusion in a pta⁺ (pDIA5375) or pta (pDIA5376) genetic background. Three fragments, 109 to +306 (415 bp) containing the CRE site of pta (CREpta), 49 to +306 (355 bp) with the CRE deleted (CREpta), and 22 to +306 (328 bp) with the CRE and the 35 promoter region deleted (Ppta), were amplified by PCR using pDIA5373 as a template (numbering is relative to the transcription start site). Oligonucleotides allowing the creation of an EcoRI restriction site near the 5' end and of a BamHI restriction site near the 3' end were used. After digestion with BamHI and EcoRI, the PCR products were inserted in plasmid pAC6 (39), yielding pDIA5377, pDIA5378, and pDIA5381, respectively (Fig. 1). The resulting pta-lacZ fusions were subsequently integrated at the B. subtilis amyE locus. To carry out the footprinting experiments, a fragment containing the CRE region of the pta gene (from 109 to +139) and a fragment in which the CRE was absent (from 49 to +139) were inserted between the HindIII and EcoRI sites of the high-copy-number plasmid pUC18 (Fig. 1).

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Plasmid construction. Various B. subtilis chromosomal DNA fragments (horizontal lines) containing the pta gene (bold arrow or bar) were inserted into the vectors listed at the right. Designations for the resulting plasmids are given at the left. The B. subtilis chromosomal DNA fragments are numbered relative to the identified transcriptional start site of the pta gene. The promoter region of the pta gene (P) and the CRE (open box) are indicated. Some restriction sites are also shown.

Isolation and purification of proteins used in protein-DNA interaction assays. CcpA-His₆, Crh-His₆, HPr-His₆, and HPr kinase-His₆ were purified on Ni-nitrilotriacetic acid-agarose columns as previously described (10). Crh-His₆ and HPr-His₆ were phosphorylated with HPr kinase-His₆ in the presence of ATP as described by Galinier et al. (11), such that about 90% of Crh-His₆ and HPr-His₆ were phosphorylated. HPr kinase was then denatured by incubation for 10 min at 80°C.

DNase I footprinting. The DNA probe was prepared as follows. pDIA5379 and pDIA5380 were linearized with EcoRI and treated with the Klenow fragment of DNA polymerase I (Boehringer) in the presence of a mixture of dGTP, dCTP, dTTP (0.5 mM), and [-³²P]dATP. A phenol-chloroform extraction was performed with the plasmids labelled at one end followed by a second digestion with HindIII. The EcoRI-HindIII labelled DNA fragments were purified on a 6% polyacrylamide gel. Binding of either CcpA, HPr, HPrSer-P, Crh, or CrhSer-P to these DNA fragments was assessed in 20-µl reaction mixtures containing about 0.03 pmol of one of the ³²P-labelled DNA fragments (150,000 to 275,000 cpm) and 1 µg of poly(dI-dC) in 100 mM KCl-10 mM HEPES (pH 7.6)-0.1 mM EDTA-2 mM MgCl₂-1 mM dithiothreitol-10% glycerol. The DNA binding reaction was performed in the presence of 2 µM CcpA and 10 µM either HPr, HPrSer-P, Crh, or CrhSer-P by incubating the assay mixture for 10 min at room temperature. The concentrations of MgCl₂ and CaCl₂ were adjusted to 1 and 0.5 mM, respectively, and 20 ng of DNase I (Worthington Biochemical, Freehold, N.J.) was added. The mixture was incubated at room temperature for 1 min, and the reaction was stopped by phenol extraction followed by the addition of 4 volumes of stop buffer (0.4 M sodium acetate, 50 µg of calf thymus DNA/ml, 2.5 mM EDTA). A+G Maxam and Gilbert reactions (26) were carried out with the same DNA fragments.

Acetate excretion measurements and phosphotransacetylase assay. For measurement of acetate production, the B. subtilis strains were grown in CSK medium supplemented with 0.4% glucose. Production of acetate was detected by high-pressure liquid chromatography using a Perkin-Elmer (Norwalk, Conn.) 3B liquid chromatograph. Metabolites were separated on an Aminex HPX 87H strong cation-exchange column (Bio-Rad Laboratories, Richmond, Calif.), protected with a cation H microguard column (Bio-Rad Laboratories). Samples were filtered through SJHV 0.45-µm-pore-size filter units (Millipore Corp., Bedford, Mass.). Elution from the column was performed at 40°C with 0.01 N H₂SO₄ at a flow rate of 0.7 ml/min. Elution of the separated products was followed with a Perkin-Elmer LC25 refractometer equipped with a Sigma 15 integrator. Each metabolite was identified and quantified comparing its retention time and peak surface with those of the corresponding standards.

1

1

1

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of the pta gene. As part of the B. subtilis genome sequencing project, an open reading frame, ywfJ, encoding a protein with a sequence very similar to that of the phosphotransacetylases of E. coli (25) and Methanosarcina thermophila (22) was identified. The B. subtilis ywfJ gene encodes a protein composed of 323 amino acids, with a calculated molecular weight of 34,800. Similar proteins (40 to 46% identical residues) are found in Mycoplasma capricolum, M. genitalium, M. pneumoniae, Pseudomonas denitrificans, Clostridium thermosaccharolyticum, C. acetobutylicum, and C. glutamicum. Longer variants of phosphotransacetylase have been found in E. coli (25), Haemophilus influenzae, and Synechocystis sp. These longer proteins are composed of about 700 amino acids, and the B. subtilis phosphotransacetylase is similar only to the C-terminal part of the longer variants of phosphotransacetylase.

Inactivation of the pta gene. To demonstrate that ywfJ does indeed encode a phosphotransacetylase involved in acetate production, the gene was disrupted with a kanamycin cassette (Fig. 1). The phosphotransacetylase activity of the wild-type strain 168 and ywfJ mutant (BSIP1171) was measured after growth on LB medium in the presence of 0.4% glucose. The formation of acetyl-CoA from acetyl phosphate via a coupled reaction (32) was tested. No significant activity was detected in the ywfJ mutant, whereas the phosphotransacetylase activity of the wild-type strain was 56 U/mg of protein. Therefore, ywfJ clearly encodes the phosphotransacetylase, and the gene was therefore renamed pta.

Regulation of expression of the pta gene. We investigated pta expression under various conditions by generating transcriptional pta-lacZ fusions. The integrative plasmids pDIA5375 and pDIA5376 (Fig. 1) are derivatives of pJM783 carrying a DNA fragment containing the promoter region plus the beginning of the pta gene and an internal fragment of pta, respectively. The integration of these plasmids at the B. subtilis pta locus by Campbell-type recombination resulted in strains containing a transcriptional fusion between pta and lacZ and an intact (BSIP1104) or disrupted (BSIP1105) pta gene (Table 1).

Analysis of the pta promoter region. The DNA sequence of the pta promoter region is presented in Fig. 2. The translation initiation codon is probably a GTG codon preceded by a potential ribosome-binding site (Fig. 2). The 5' end of the pta transcript was identified by primer extension analysis using total RNA extracted from a B. subtilis wild-type strain grown on CSK medium in the presence and absence of glucose (Fig. 3, lanes 1 and 2, respectively). Glucose clearly stimulates pta transcription (lanes 1 and 2). The 5'-end-labelled primer (Fig. 2), gave a 95-base extension product which identified the adenine designated nucleotide +1 as the 5' end of the pta mRNA. The putative 10 and 35 regions of the promoter are indicated in Fig. 2. pDIA5381 (Fig. 1), a derivative of pAC6 carrying a pta fragment (from 22 to +306 [P in Fig. 2]) with only the 10 region fused to lacZ, was constructed to confirm the position of the promoter. This fusion was integrated at the amyE locus of B. subtilis and gave a much lower level of -galactosidase activity (2 to 6 U/mg of protein) than the fusion containing the complete promoter region (109 to +306) (250 U/mg of protein) (Fig. 4). Deletion of the 35 part of the promoter abolished transcription of the pta-lacZ fusion. A typical CRE sequence (18), TGAAAGCGCTATAA, is located between positions 62 and 49 relative to the transcription start site of the pta gene (boxed in Fig. 2). This sequence contains one mismatch (the adenine at position 50) compared with the CRE consensus sequence (18). To investigate the role of this sequence in catabolite regulation of the pta gene, two pta-lacZ fusions were constructed by inserting into pAC6 DNA fragments with and without the CRE region of the pta gene (Fig. 1). The start points of fragments CREpta (109) and CREpta (49) are indicated in Fig. 2. The pta-lacZ fusions present in the resulting plasmids pDIA5377 (CRE, 109 to +306) and pDIA5378 (CRE, 49 to +306) were integrated as single copies at the amyE locus of B. subtilis 168. -Galactosidase activity was measured for the various strains grown in CSK medium in the presence or absence of 0.4% glucose (Fig. 4). For the CREpta-lacZ fusion, we observed a 3-fold activation of pta-lacZ expression in the mid-exponential growth phase in CSK medium supplemented with 0.4% glucose. No such activation was observed with the CREpta-lacZ fusion. This indicates that the DNA fragment located between positions 109 and 49 is responsible for catabolite activation of the pta gene and suggests that the predicted CRE sequence is involved in this activation.

View larger version (21K): [in this window] [in a new window]   FIG. 2.   pta promoter region. The nucleotide sequence of a 291-bp-long DNA fragment containing the pta promoter and the beginning of the pta gene is presented. The vertical arrow indicates the position of the transcription start site, +1. The primer used for the primer extension experiment is indicated by the long horizontal arrow. The 10 and 35 regions, corresponding to the RNA polymerase binding site, are indicated. The potential ribosome-binding site (Shine-Dalgarno [SD]) and the CRE sequence are boxed. The bent arrows at positions 109, 49, and 22 indicate the start of the CRE, CRE, and P fragments used for the construction of the pta-lacZ fusions in pDIA5377, pDIA5378, and pDIA5381 (Fig. 1).

View larger version (37K): [in this window] [in a new window]   FIG. 3.   Mapping of the transcription start site of the pta gene by primer extension. Total RNAs was extracted from B. subtilis 168 grown in CSK medium in the presence (lane 1) or absence (lane 2) of 0.4% glucose. In lane 3, the labelled oligonucleotide was loaded as a control. The labelled primer used for reverse transcription is indicated in Fig. 2. Sequencing reactions were performed with the same oligonucleotide as a primer and pDIA5373 as the template.

View larger version (17K): [in this window] [in a new window]   FIG. 4.   Importance of the CRE and 35 promoter region for pta expression. We monitored the expression of pta-lacZ expression over time for strains BSIP1114 (), BSIP1115 (), and BSIP1116 (). These strains contain the CREpta, CREpta, and Ppta fragments (Fig. 2) fused to the lacZ gene inserted at the amyE locus. The strains were grown in CSK medium in the absence (open symbols) or presence (closed symbols) of 0.4% glucose.

Effects of trans-acting CCR proteins on pta expression. CcpA, HPr, and Crh are known to play a role in catabolite activation of genes involved in carbon excretion pathways (41). In addition, HPr kinase, which phosphorylates both HPr and Crh at Ser-46, is necessary for CCR (11). We investigated whether these trans-acting factors were involved in the regulation of expression of the pta gene. Chromosomal DNA from a ccpA or hprK mutant was used to transform a pta-lacZ strain, and chromosomal DNA from strain BSIP1104 (pta-lacZ) was used to transform the ptsH1 (QB5223), crh::aphA3 (QB7096), and ptsH1 crh::aphA3 (QB7102) strains (Table 1). Expression of the pta-lacZ fusion was assessed in the various mutants after growth in CSK medium in the presence or absence of 0.4% glucose. The results are summarized in Table 3.

Interaction of trans-acting CCR proteins with the pta regulatory region. To test binding of the transcriptional activator/repressor CcpA to the CRE sequence identified in the pta promoter region, DNase I footprinting experiments were performed. The key regulator CcpA, the PTS protein HPr, and its homologue Crh (in both dephosphorylated and Ser-46-phosphorylated forms) were purified. Two DNA fragments containing the pta promoter region from positions 109 to +139 or 49 to +139 (Fig. 1) were labelled with ³²P at the EcoRI site. The first DNA fragment (CREpta) contained the CRE site identified in vivo. This site was deleted in the second fragment (CREpta). The results of the DNase I footprinting experiments are shown in Fig. 5.

View larger version (75K): [in this window] [in a new window]   FIG. 5.   DNase I footprinting experiments with the pta promoter region and trans-acting CCR proteins. (A) Footprinting with the 248-bp EcoRI-HindIII DNA fragment containing the CREpta (109 to +139) (Fig. 2); (B) footprinting with the 188-bp EcoRI-HindIII DNA fragment CREpta (49 to +139) (Fig. 2). The CREpta and CREpta fragments were obtained from plasmids pDIA5379 and pDIA5380, respectively, and labelled at the 3' end as described in Materials and Methods. A+G standards for Maxam and Gilbert reactions were made for each fragment. The DNA sequence of the promoter region of the two fragments (from positions 84 to 33 [A] and 14 to +34 [B]) as well as the protected areas are indicated. DNase I digestions were performed as indicated in Materials and Methods with both fragments. In each case, the DNA was digested in the absence of proteins (lane 1) or in the presence of 2 µM CcpA (lane 2), 2 µM CcpA and 10 µM HPr (lane 3), 2 µM CcpA and 10 µM HPrSer-P (lane 4), 2 µM CcpA, 10 µM HPrSer-P, and 20 mM FBP (lane 5), 2 µM CcpA and 10 µM Crh (lane 6), 2 µM CcpA and 10 µM CrhSer-P (lane 7), or 2 µM CcpA, 10 µM CrhSer-P, and 20 mM FBP (lane 8).

View larger version (7K): [in this window] [in a new window]   FIG. 6.   pta promoter regions protected in DNase I footprinting experiments. The sequence of the pta promoter region from positions 70 to +29 is presented. The CRE sequence is boxed, and the CRE consensus sequence is indicated. The bases protected against digestion with DNase I are indicated by asterisks.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Glycolysis and the Krebs cycle, the central pathways of carbon metabolism, have a dual function: providing energy and reducing equivalents in form of ATP and NADH or NADPH, respectively, and facilitating the synthesis of precursors for most biosynthetic pathways. To fulfill both functions and to maximize the growth rate, some of the carbon flow may be directed toward the formation of by-products excreted into the medium. If external electron acceptors such as oxygen or nitrate become limiting, B. subtilis cells growing in a medium containing excess of carbon excrete a large variety of compounds that are either intermediates of the central pathways, such as pyruvate or succinate, or produced by pathways branching off from the central metabolic pathways, such as acetate, acetoin, and lactate (34, 38). In addition to the redox equilibrium and the flow rate through the central pathways, pH changes may also affect by-product synthesis as observed in Lactobacillus plantarum (40). The pH dependence is not surprising, as the two major by-products, pyruvate and acetate, are organic acids. B. subtilis cells have developed complex regulatory systems to control by-product formation so as to keep carbon metabolism balanced.

The absence of phosphotransacetylase activity in ywfJ mutant indicates clearly that this gene encodes phosphotransacetylase. By analyzing acetate production in various mutant strains, we found that acetate is not only formed via the Ack-Pta pathways. The low level of acetate in the pta mutant confirms that this gene encodes phosphotransacetylase. The similar level of acetate synthesis in the alsS pta double mutant indicates that the butanediol cycle is not involved in acetate production. Some unknown pathway is therefore responsible for the significant acetate production in the pta mutant and in the alsS pta double mutant. Acetyl-CoA synthetase may be responsible for the residual acetate synthesis in the pta mutant, although its main function is related to acetate utilization (34).

Although the pta and ackA mutants affect the same pathway for acetate excretion, there is a significant difference between these two mutants. Glucose strongly inhibits the growth of the ackA mutant only (15), suggesting that the low level of acetate excretion is not responsible for this growth inhibition. Acetate kinase catalyses the second step of the acetate excretion pathway, the conversion of acetyl~P to acetate. In the presence of glucose, the ackA mutant probably accumulates acetyl~P. This compound has been shown to be involved in the regulation of signal transduction by the two-component regulatory systems in various bacteria (27). In the ackA mutant, abnormal acetyl~P-mediated regulation of several two-component systems may account for the inhibitory effect of glucose on cell growth. Such growth inhibition is not observed for the pta mutant, which probably contains only low concentrations of acetyl~P.

The location of ack and pta in two independent transcriptional units suggested that acetate kinase and phosphotransacetylase synthesis are differently regulated, but very similar mechanisms of regulation of expression have been observed for these two genes (15, 41). Coregulation is further substantiated by the similar codon preferences in pta and ackA (29). In gram-positive bacteria, catabolite regulation of various operons is mediated by the trans-acting dimeric protein CcpA, a repressor belonging to the LacI-GalR family of bacterial regulatory proteins. CcpA binds to the palindromic operator sequence CRE (17, 42). As observed for the ackA gene (15, 41), expression of the pta gene is increased threefold in the presence of glucose (Table 3 and Fig. 4), and this activation is mediated by CcpA (Table 3). We demonstrated that the products of at least four genes (ccpA, ptsH, crh, and hprK) are involved in catabolite activation of the pta gene. Replacement of the Ser-46 of HPr with an alanyl residue, as in ptsH1, and disruption of the crh gene did not affect glucose activation (Table 3). By contrast, a ptsH1 crh double mutant showed a release from catabolite activation identical to that of the ccpA and hprK single mutants. It therefore seems likely that both HPrSer-P and CrhSer-P exert their effects on catabolite activation via CcpA as for CCR (9, 10).

The pta promoter region (Fig. 2) contains a single ^A-dependent promoter and a CRE sequence, TGAAAGCGCTATAA, located between positions 62 and 49 relative to the transcriptional start site. Deletion of this CRE abolished catabolite activation of pta gene expression (Fig. 4). The location of this CRE (centered on position 55.5) is similar to that of the ackA CRE (56.5), the only other B. subtilis CcpA binding site shown to be involved in catabolite activation (41).

DNase I footprinting experiments demonstrated that HPrSer-P and CrhSer-P increased the specific binding of CcpA to the pta CRE (Fig. 5). However, FBP was required to observe clear protection by the CcpA-HPrSer-P and CcpA-CrhSer-P complexes. FBP, the concentration of which is greatly increased by growth of the cells in glucose-containing medium (7), interferes at several steps in the CCR signal transduction pathway, ultimately facilitating the binding of CcpA to the CRE. FBP stimulates the phosphorylation of HPr and Crh at the regulatory Ser-46 by the ATP-dependent HPr kinase (11) and the interaction between HPrSer-P and CcpA (4). In some cases, FBP probably also increases the in vitro binding of the CcpA-HPrSer-P and CcpA-CrhSer-P complexes to their targets as observed for the xyn CRE (9). FBP may also be responsible for doubling of pta-lacZ expression in mutants carrying a disrupted pta gene. Inactivation of the Pta-Ack acetate excretion pathway in such mutants may lead to an increase in the concentration of FBP and other glycolytic intermediates and hence to altered catabolite regulation.

A second region, located between positions +1 and +14, is also protected by the CcpA-HPrSer-P and CcpA-CrhSer-P complexes against in vitro digestion by DNase I. This second CcpA binding site, which is less similar to the consensus CRE sequence, is not sufficient to confer in vivo glucose activation (Fig. 4). The presence of an auxiliary CRE site has also been shown for the xyl and gnt operons (13, 28). Although the position of this additional CRE sequence is atypical for a positive regulatory site, the binding of CcpA to multiple CRE sites, possibly brought into close contact by DNA looping, may help to cause local changes in DNA conformation or may facilitate CcpA-RNA polymerase interactions. The binding of the CcpA-HPrSer-P and CcpA-CrhSer-P complexes caused also significant alterations in the pattern of DNase I digestion outside the CRE sequence (Fig. 5), suggesting that changes in the DNA structure may be induced.

These results confirm that very similar mechanisms, including protein phosphorylation and protein-protein and protein-DNA interactions, are involved in catabolite activation and carbon catabolite repression in B. subtilis. Further studies are required to elucidate the molecular details of these mechanisms, leading in one case to activation and in the other to inhibition of gene expression.