INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Catabolite repression (CR) is a ubiquitous phenomenon whereby the presence of a rapidly metabolizable carbon source such as glucose inhibits the expression of genes encoding proteins concerned with the utilization of alternative carbon sources (32, 33). In Escherichia coli this phenomenon is mediated by two transcriptional regulatory proteins and their cognate cytoplasmic signaling molecules, the cyclic AMP receptor protein which binds cyclic AMP to effect positive control of operon expression, and the catabolite repressor/activator (Cra) protein, which effects positive control except when a metabolite such as fructose-1-phosphate or fructose-1,6-bisphosphate is bound to it (34, 35). In low-GC-content gram-positive bacteria, such as Bacillus subtilis, an entirely different CR mechanism is operative. This process involves catabolite control protein A (CcpA) (18, 19), which interacts with metabolites and a phosphorylated regulatory protein [HPr(Ser-P) or Crh(Ser-P)] to effect CR by a negative control mechanism (9, 20, 27, 36). Although the general features of this gram-positive bacterial CR mechanism are recognized, the fine details have yet to be elucidated.

We recently reported the isolation of a novel B. subtilis mutant exhibiting partial resistance to CR due to the insertion of a mini-Tn10 element into a gene which proved to encode a protein that was 30% identical to CcpA (4). We designated this protein catabolite control protein B (CcpB). While CcpA solely mediates CR during growth in liquid media with a high level of agitation, both CcpA and CcpB mediate CR during growth under the same conditions but with a low level of agitation or during growth on solid media (4).

The ccpB mutation was isolated by mini-Tn10 insertional mutagenesis of a B. subtilis strain carrying a translational gluconate (gnt) operon fusion (gntRK'-'lacZ), selecting for blue colonies on solid media containing gluconate plus glucose plus X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside) (4). The success of this method in identifying the ccpB gene prompted us to isolate and characterize additional mutants by the same procedure. In this paper we describe the isolation and characterization of three classes of mutants, two of which exhibit full resistance to glucose-promoted gnt operon CR and the other of which exhibits partial resistance. The former two classes of mutants proved to carry mini-Tn10 elements in (i) the ptsG gene encoding the well-characterized glucose permease (glucose enzyme II) of the phosphoenolpyruvate-sugar phosphotransferase system (PTS) and (ii) the glcT gene encoding the transcriptional antiterminator that positively controls ptsG gene expression (39). The third class of mutants proved to carry the mini-Tn10 element in a novel, previously unsequenced gene that we have designated glcP. The product of glcP, the GlcP protein, is shown to be a glucose/mannose:H⁺ symport permease, a member of the fucose/glucose/galactose family of the major facilitator superfamily (28). We present evidence suggesting that expression of glcP is glucose inducible and that induction depends upon the integrity of the PTS.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and growth conditions. The B. subtilis and E. coli strains used in this study are listed in Table 1. B. subtilis and E. coli strains were routinely grown in Luria-Bertani (LB) medium supplemented with 1% sugar when appropriate. Growth rates of B. subtilis strains were determined with C minimal medium (1) supplemented with 1% glucose. B. subtilis was transformed as described by Kunst et al. (22). E. coli was transformed by standard processes as described by Sambrook et al. (37). Antibiotics were used for selection in B. subtilis at the following concentrations: phleomycin, 0.25 µg/ml; kanamycin, 5 µg/ml; and spectinomycin, 100 µg/ml; in E. coli the following concentrations were used: ampicillin, 100 µg/ml; and spectinomycin, 150 µg/ml.

Recombinant DNA techniques. Standard molecular cloning techniques were performed as described by Sambrook et al. (37). PCR amplification was performed with Taq DNA polymerase (Stratagene) in accordance with the manufacturer's recommendations. Synthetic oligonucleotides were obtained commercially from either GenSet or Research Genetics. The following primers were used to clone the glcP gene in E. coli: glc1, ATACATATGAATGTTAAGAGGGACATATTTATTTGG; and glc2, GTCGACGAATTCCTACTGTGTTTTTGATGCTTTGTTTTCAAG. Nucleotide sequencing was performed by PCR dye terminator sequencing on an ABI 373 sequencer (Applied Biosystems).

-Galactosidase assays after growth in liquid media. B. subtilis cells from a single colony were grown overnight at 37°C in 5 ml of LB medium containing a 1% concentration of the desired sugar and the appropriate antibiotic(s). Then, 1.5 ml of culture was harvested by centrifugation. Cells were suspended in 1 ml of Z buffer (24) supplemented with 40 µg of lysozyme and 6.25 µg of DNase per ml and incubated for 10 min at 37°C. Cell debris was eliminated by centrifugation. -Galactosidase activity was measured as described by Miller and expressed in Miller units (24). Protein concentrations were measured with the Coomassie blue reagent supplied by Bio-Rad (3), with bovine serum albumin as a standard.

Transport assays. E. coli or B. subtilis cells were harvested in the mid-log growth phase, washed twice, and resuspended in 50 mM Tris-maleate buffer (pH 7.0). Sugar uptake was examined by using a 100-µl cell suspension incubated at 37°C. ¹⁴C-sugar (glucose, mannose, 2-deoxyglucose, methyl -D-glucopyranoside, or glucitol) was added at 0 min at the concentrations noted. The inhibitors carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone (FCCP) and sodium arsenate were added where noted at concentrations of 10 µM and 10 mM, respectively. Experiments using sodium arsenate were performed in 50 mM potassium phosphate buffer (pH 7.0). Cold sugars were added as inhibitors at the concentrations indicated in Table 3. Samples of 20 to 30 µl were removed at appropriate intervals, filtered through 25-mm-diameter membrane filters (0.45-µm pore size; Millipore Corp., Bedford, Mass.), and washed with cold 50 mM Tris-maleate buffer (pH 7.0). Washed filters were submerged in vials containing 10 ml of scintillation fluid, and the radioactivity was quantified in a liquid scintillation counter.

Nucleotide sequence accession number. The complete sequence of the B. subtilis glcP gene and surrounding regions reported in this paper has been submitted to the GenBank database and is available under accession number AF002191.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Isolation of mutants resistant to glucose-promoted gnt operon CR. The B. subtilis gnt operon, encoding the catabolic enzyme system responsible for gluconate utilization (7, 29), is subject to CR mediated by a mechanism involving (i) two catabolite-responsive element sequences present within and upstream of the coding region of gntR (25-27), (ii) the CcpA protein (8), and (iii) HPr(Ser-P) or glucose-6-P (5, 27). CR of the gnt operon was examined in this study by using B. subtilis ST100, which carries a gntRK'-'lacZ translational fusion inserted into the chromosomal amyE gene (4, 5). Transposon mutants of ST100 were generated with a mini-Tn10 transposon, which carries a spectinomycin resistance gene and a ColE1 origin of replication (38), as previously described (4). Three classes of mutants which were resistant to glucose-promoted gnt operon CR but sensitive to mannitol-promoted gnt operon CR were isolated. The first of these classes was partially resistant to glucose CR and was represented by strain ST141; the second class was fully resistant to glucose CR and was represented by strains ST143 and ST144; the third class was almost fully resistant to glucose CR and was represented by strain ST145 (Table 2). To ensure that only a single transposon insertion was present in each of these strains and that the insertion was responsible for the relief of CR observed, a backcross was performed. ST100 was transformed with chromosomal DNA from each of the strains, and transformants were selected with spectinomycin. In each case, 100% of the transformants displayed the same CR phenotype as the original mutants.

Cloning and sequencing of transposon inserts in ptsG. Chromosomal DNA was purified from strains ST143 and ST144 and digested with HindIII, which does not cut within the mini-Tn10 transposon. The mini-Tn10 transposon insert and flanking DNA from each of these strains were cloned into E. coli TG1 by self-religating the HindIII-digested DNA. The mini-Tn10 transposon carries a ColE1 origin of replication. Sequencing the cloned inserts from ST143 and ST144 revealed that the transposon was located within the same gene, ptsG, which encodes the glucose-specific enzyme II of the PTS (11, 30). In ST143, the transposon insert occurred at codon 521, which corresponds to the linker region located between the IIB and IIA domains of enzyme II. In ST144, the transposon insert was located at codon 341, which corresponds to a site within the IIC domain of enzyme II. We also examined five additional independently isolated transposon insertional mutants in ptsG, which had identical phenotypes. The transposon inserts in these mutants were cloned into E. coli TG1, and the sites of insertion were determined approximately by restriction analysis. The mutations occurred randomly throughout ptsG, arguing that ptsG does not include a region that serves as a hot spot for mini-Tn10 transposon insertion.

Cloning and sequencing of a transposon insert in glcT. The transposon insert and flanking DNA in strain ST145 were cloned into E. coli and sequenced as described above. The transposon insert proved to be in codon 110 of the glcT gene, recently shown to encode transcriptional antiterminator GlcT, which controls expression of the ptsG gene (39). The GlcT protein is homologous to a family of antiterminators that are regulated by PTS-mediated phosphorylation (40). The ptsG operon and GlcT activity appear to be negatively autoregulated by PtsG, probably because PtsG phosphorylates GlcT (39). Loss of GlcT prevents antitermination and hence prevents expression of ptsG, thus preventing PtsG-mediated CR (Table 2).

Cloning and sequencing of a transposon insert in a novel transporter gene, glcP. The transposon insert and flanking DNA in strain ST141 were cloned into E. coli and sequenced as described in the previous two sections. The site of transposon insertion did not have significant nucleotide sequence similarity to any sequence in the databases. Hence, we determined the nucleotide sequences on both strands of a 2,962-bp region flanking the insert by use of a primer walking approach. The complete sequence is shown in Fig. 1A. Four open reading frames (ORFs) were detected in this region (Fig. 1B). The transposon insert occurred at codon 141 within the second of these ORFs. The product of this ORF is predicted to be a 401-amino-acyl residue protein which proved to show sequence similarity with the fucose permease of E. coli (14, 15) and the glucose/galactose permease of Brucella abortus (6, 31). These permeases constitute a unique family within the major facilitator superfamily (28). We designated this novel B. subtilis ORF glcP (glucose permease). Hydropathy plots of the GlcP protein sequence suggested a 12-transmembrane -helical spanner topology, as for its homologs. Both upstream and downstream of glcP are inverted repeats which might serve as transcriptional terminators. No obvious promoter precedes the glcP gene.

View larger version (44K): [in this window] [in a new window]   FIG. 1.   (A) Nucleotide sequence of the glcP gene and the surrounding region of the B. subtilis chromosome and translated protein sequences of GlcP, partial OrfA, OrfB, and partial OrfC. The position of the mini-Tn10 transposon insert is shown white on black. Arrows indicate the directions of transcription. Inverted repeats that precede and follow glcP and that may serve as transcriptional terminators are also indicated (opposing arrows). (B) Schematic depiction of sequenced genes presented in panel A together with functional assignments when available. The number of nucleotides separating each ORF is indicated in parentheses. ST, stem-loop structures that may serve as transcriptional terminators.

CR resistance in ptsG and glcP mutants. The -galactosidase activities of the gntRK'-'lacZ fusion in strains ST100, -141, -143, -144, and -145 were examined on solid media in the presence of gluconate and various repressing sugars (Table 2). In the absence of a repressing sugar, all strains exhibited uniformly high levels of -galactosidase activity. In wild-type strain ST100, glucose, mannitol, and fructose repressed -galactosidase synthesis to the background level. Sucrose was somewhat less effective, and glycerol proved to be nonrepressive. glcP mutant strain ST141 was partially relieved for glucose and sucrose CR, but no relief was observed when the repressing sugar was mannitol or fructose. ptsG mutant strains ST143 and ST144 exhibited maximal -galactosidase activities after growth in the presence of glucose, but essentially no activity when mannitol was the repressing sugar. The glcT mutant behaved like the ptsG mutants except that very slight repression was observed after growth in the presence of glucose. Thus, the glcP, ptsG, and glcT mutations specifically but differentially relieve CR when glucose or a glucose-generating sugar such as sucrose is the repressing sugar.

Effects of glcP, ptsG, and glcT mutations on growth of B. subtilis. Growth rates of the various B. subtilis strains were examined on solid minimal media, and some of the results obtained were confirmed with quantitative growth rate measurements in liquid media. The ptsG, glcT, and glcP mutant strains exhibited depressed growth rates on minimal glucose plates as estimated by colony size. The decrease in colony size was much greater for the ptsG and glcT mutants than for the glcP mutant, in agreement with the CR results reported in Table 2. The glcT mutant grew somewhat better than the ptsG mutant, suggesting that some expression of the ptsG gene occurs in the absence of GlcT function. No effect of these mutations was observed on minimal mannitol plates. The doubling times of wild-type strain ST100 and glcP mutant ST141 were 95 and 145 min, respectively, in liquid minimal media containing 1% glucose as the sole carbon source (see Materials and Methods).

Effects of the glcP mutation on uptake of [¹⁴C]glucose and [¹⁴C]methyl -glucoside in B. subtilis. Time courses for the uptake of [¹⁴C]glucose and [¹⁴C]methyl -glucoside are shown in Fig. 2A and 2B, respectively. The glcP mutation in the wild-type genetic background reduced the uptake of glucose about 30% after cells were grown in rich medium containing 1% glucose. By contrast, when glucose uptake was measured in the ptsGHI deletion background after growth under the same conditions, the uptake rate was low and the glcP mutation had no discernible effect (Fig. 2A). The residual glucose uptake observed suggested that a third transport system, independent of the PTS and GlcP, must be operative. Methyl -glucoside uptake rates were also reduced by the glcP mutation in the wild-type genetic background when cells were grown under the same conditions as those described above (Fig. 2B). The ptsGHI mutation completely abolished [¹⁴C]methyl -glucoside uptake. The third putative glucose transporter evidently does not accumulate methyl -glucoside.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Uptake of radioactive glucose (A) and methyl -glucoside (B) into various B. subtilis strains. Both studies were conducted with the 14C-sugar present at 50 µM (see Materials and Methods). Strains used were as follows: wild type (); glcP mutant (); ptsGHI mutant (); ptsGHI glcP mutant ().

Cloning of the wild-type glcP gene and expression in E. coli. The glcP gene was amplified by PCR from B. subtilis ST100 DNA with primers specific to the ends of glcP (see Materials and Methods). This PCR-generated product was cloned into E. coli vector pCR2.1 (Invitrogen) and introduced into E. coli INVF' creating construct pGlcP. Insertion of the glcP gene into this plasmid was verified by nucleotide sequencing of the 5' and 3' ends of the gene with primers specific for the vector. E. coli RE707 and RE777 were transformed with pGlcP in which the glcP gene is expressed from the vector lac promoter. These two strains are defective for glucose and galactose transport, respectively (6). The plasmid complemented the glucose-negative RE707 strain, as determined by fermentation on eosin-methylene blue-glucose (1%) plates as well as by growth on minimal glucose (0.5%) plates. Strain RE707 proved to be mannose negative, and plasmid pGlcP allowed partial restoration of that negative phenotype as well. On the other hand, it did not complement the fermentation and growth defects of strain RE777 on galactose plates. These results suggest that GlcP is expressed in E. coli under the conditions used and that it transported both glucose and mannose but not galactose.

¹⁴C-sugar uptake studies in E. coli. Results reported in the previous section strongly indicated that GlcP transports glucose and mannose but not galactose. When the glcP gene was expressed at high levels in E. coli RE707, rapid uptake of [¹⁴C]glucose, [¹⁴C]2-deoxyglucose, and [¹⁴C]methyl -glucoside was observed (unpublished results and Fig. 3). The K_m for methyl -glucoside uptake was estimated to be roughly 20 µM (data not shown). Both 2-deoxyglucose and methyl -glucoside were accumulated against substantial concentration gradients (greater than fivefold). The latter sugar was extracted from the cells with boiling water and was analyzed by passage through Dowex AG1-X2, 50- to 100-mesh columns (0.8 by 5 cm) (21). All of the accumulated sugar was in the free (nonphosphorylated) form. The addition of the protonophore FCCP (10 µM) prevented the accumulation of methyl -glucoside, but the addition of 10 mM sodium arsenate had no effect (Fig. 3). These results suggest that GlcP functions by a proton motive force-driven mechanism.

View larger version (22K): [in this window] [in a new window]   FIG. 3.   Uptake of radioactive 2-deoxyglucose (A) and methyl -glucoside (B) into E. coli RE707 lacking () or bearing (all other symbols) a high-level expression vector, pGlcP, encoding the GlcP protein of B. subtilis. The figure illustrates uptake in the absence of inhibitor () or in the presence of FCCP () or sodium arsenate () (see Materials and Methods). Sugar uptake by strain RE707 was not affected by either inhibitor (data not shown).

View this table: [in this window] [in a new window]   TABLE 3.   Effects of nonradioactive sugars, present at a 1,000-fold excess, on the uptake of [14C]methyl -glucoside into E. coli RE707 bearing the plasmid-encoded glcP gene

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

In this study we have isolated and characterized three classes of insertional mutants, two of which (ptsG and glcT) abolish glucose-specific CR and one of which (glcP) partially relieves CR. The ptsG mutants are defective for the glucose-specific enzyme II of the PTS. These mutants did not display glucose CR, but they exhibited strong repression when mannitol was the repressing sugar. The parental strain was strongly repressed by both sugars. Since -galactosidase activity in the gntRK'-'lacZ fusion strain studied was essentially the same when cells were grown in gluconate- or gluconate-plus-glucose-containing media, we conclude that ptsG mutations completely abolish CR. The comparable effects of the glcT (ptsG antiterminator-encoding) insertional mutation can be explained by the poor expression of ptsG expected in the absence of antitermination (39).

In B. subtilis, ptsG precedes ptsH (HPr), which precedes ptsI (enzyme I), and all three genes are in the same orientation (30). Gonzy-Tréboul et al. (11) provided evidence that ptsH and ptsI form an operon. ptsG presumably constitutes a distinct operon. Evidence for this proposal is provided in this report. Thus, knockout mutations in ptsG or glcT abolished glucose utilization and glucose-promoted CR but had no discernible effect on mannitol utilization or mannitol-promoted CR. Since mannitol is utilized by B. subtilis exclusively via the PTS, enzyme I and HPr must be synthesized at near-wild-type levels in the ptsG and glcT insertional mutants.

Loss of GlcP function due to an insertional mutation in the newly defined glcP gene of B. subtilis resulted in partial relief from CR. The sum of the effects of the ptsG and glcP mutations was less than additive since loss of ptsG function gave full loss of CR. This fact remained enigmatic until we considered the possibility that expression of glcP may require the presence of exogenous glucose and a functional ptsG gene in B. subtilis. Thus, when wild-type cells were grown without glucose or when a ptsG mutant strain was grown with glucose, no GlcP activity was detected. Loss of ptsG may prevent expression of glcP and hence prevent CR mediated by GlcP. Possibly, PtsG promotes glcP gene transcription by allowing accumulation of cytoplasmic glucose-6-P, a potential glcP gene inducer. While this postulate represents an attractive possibility, the proposed PTS-dependent induction mechanism may prove more complex and interesting. Work is in progress to define the details of this mechanism.

The work presented in this paper provides a preliminary description of GlcP, a novel glucose/mannose:H⁺ symporter. This permease has been found to be a member of the fucose/glucose/galactose:H⁺ symporter (FGHS) family, one of the seventeen currently recognized families that constitute the major facilitator superfamily (MFS), also called the uniporter/ symporter/antiporter superfamily (12, 13, 23, 28). The well-characterized sugar porter family of the MFS is an exceptionally large family with over 130 currently sequenced members, many of which transport hexoses by either uniport or proton symport (2, 16, 17, 28). Although the proteins of the small FGHS family also catalyze sugar:H⁺ symport, they are only distantly related to the proteins of the sugar porter family (28). The two previously characterized members of the FGHS family, the fucose permease of E. coli (14, 15) and the galactose/glucose permease of B. abortus (6, 31), both transport sugars of the galacto configuration. Our results suggest that GlcP is not capable of transporting galactose, although it does transport glucose, mannose, 2-deoxyglucose, and methyl -glucoside. These observations suggest that GlcP stereospecifically recognizes the hydroxyl group at position 4 (which must be in the gluco configuration rather than the galacto configuration) but does not recognize the hydroxyl group at position 2 (which can be in the cis or trans configuration or can be absent) or the hydroxyl group at position 1 (which can be methylated). It will be interesting to examine the transport specificities of the other two characterized members of the FGHS family in order to determine their stereospecificity requirements. Further studies may reveal the breadth of functions served by proteins of the FGHS family as additional members of this small MFS family become sequenced and characterized.