INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Lactobacillus brevis is a ubiquitous microorganism that can be isolated from various biotopes, such as milk, fermented vegetables, and the intestinal tracks of animals and that is often found as a spoilage contaminant in beer production (7, 11, 26). Fermented plant materials, one of the predominant ecological niches of L. brevis, are usually rich in hemicellulose fibers and therefore represent an abundant source of D-xylose, from which L. brevis can derive energy for growth. The fermentation of D-xylose is not a common property among Lactobacillus species. Besides L. brevis and Lactobacillus pentosus, most of the other lactobacilli are unable to utilize D-xylose as an energy source. The ability to ferment D-xylose, however, could improve the properties of heterofermentative lactic acid bacteria that are commonly used in fermented-food technology. Lactobacillus plantarum, for instance, is widely used to stimulate the fermentation of silage, sourdough, and diverse vegetables, such as the white cabbages used in the production of sauerkraut. L. plantarum cannot ferment D-xylose, although this property could potentially improve the competitive position of this organism in the fermentation of plant materials and would conform to the food-grade status of recombinant strains. The use of D-xylose metabolism as a food-grade selection marker for heterofermentative lactobacilli has already been proposed by Posno et al. (24). A plasmid, pLP3537-xyl, harboring the D-xylose catabolizing genes of L. pentosus, was used to complement the inability of Lactobacillus casei ATCC 393 to metabolize this pentose. However, the growth of the L. casei transformants carrying pLP3537-xyl was slow compared to that of L. pentosus, which naturally ferments D-xylose. In that study, the authors postulated that the transport of D-xylose in L. casei was the limiting function for growth on this compound. Indeed, pLP3537-xyl lacked a specific transporter for D-xylose, the absence of which could limit its general use.

Consequently, we aimed at the characterization and functional analysis of a D-xylose transporter from a Lactobacillus species that could be used to optimize the food-grade vector based on D-xylose fermentation. So far, no evidence indicating the presence of a specific transporter for D-xylose in L. pentosus has been obtained, but the situation could be different with L. brevis.

The xylA gene (encoding D-xylose isomerase) of L. brevis has previously been cloned and sequenced (2). Sequencing of regions downstream of the xylA gene revealed the presence of another, albeit incomplete, gene: xylB (encoding D-xylulose kinase). The possibility that one or more genes specifying a D-xylose transporter of L. brevis could be located in the surrounding of the xylA and xylB genes was investigated.

We report here the cloning of the xylT gene encoding the proton motive force (PMF)-linked D-xylose transport system of L. brevis and its functional expression in L. plantarum 80 by using a Lactobacillus expression system suitable for developing a food-grade vector based on D-xylose fermentation. We also provide information on the arrangement of the xyl genes in L. brevis, an arrangement that is similar to that which is found in Bacillus megaterium (30) and Tetragenococcus halophila (33), except that a repressor gene, xylR, is lacking in front of xylA in L. brevis. We also show that the L. brevis D-xylose transporter is remarkably similar to XylE of Escherichia coli (8) and T. halophila and to XylT of B. megaterium in its primary sequence and substrate specificity.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, and growth conditions. E. coli DH5 (supE44 lacU169 (80lacZM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1) was used for the propagation of the Lactobacillus-E. coli shuttle vectors, and E. coli TG1 [supE hsd5 thi (lac-proAB) F'(traD36 proAB⁺ lacI^q lacZM15)] was used for the subcloning of the L. brevis xylABT locus. They were maintained on Luria-Bertani broth, and ampicillin was added at a concentration of 100 µg/ml when necessary. Lactobacillus strains were cultivated in MRS medium (Difco Laboratories, Detroit, Mich.) or in M medium (19) supplemented with 0.5 or a 1% (wt/vol) concentration of the indicated sugar. Erythromycin (5 µg/ml) was used for the selection. Identification of L. plantarum 80(pLPA9) transformants was performed on M medium agar plates containing 25 mM D-glucose, 60 µg of 5-bromo-4-chloro-3-indolyl--D-glucuronide per ml (Sigma Chemicals Co., St. Louis, Mo.), and 100 mM potassium phosphate (pH 7.4).

Inverse PCR. Genomic DNA was extracted from L. brevis as described previously (2). DNA was restricted with ApoI, BclI, EcoRV, and Sau3AI, and the restricted fragments were self-ligated with T4 DNA ligase. To amplify the xylT gene and the remainder of the xylB gene, the ligation mixtures were amplified by using Taq DNA polymerase (GIBCO BRL, Gathersburg, Md.) and two sets of primers: either Sen-2 (5'-AGTCTTACGACCAGCGAT-3'), specifying codons 93 to 99 of the xylB gene, and Ansen-4 (5'-GGAGTAACTTAAGCCTTC-3'), specifying anticodons 74 to 79 of the xylB gene, or Sen-3 (5'-GGTGGTTGGTCAGGTTA-3'), specifying codons 230 to 235 of the xylT gene, and Ansen-5 (5'-CCACTTGGTCATGCTTGT-3'), specifying codons 207 to 212 of the xylT gene. The purified PCR fragments were restricted with the appropriate restriction enzymes and ligated into pUC19, yielding plasmids pYCBS0.8 (Sau3AI), pYCBEV1.5 (EcoRV), and pYCBBc2.3 (BclI) or into pUC18, yielding plasmid pYCBA2.0 (ApoI) (see Fig. 1A). Several randomly selected clones were sequenced.

Construction of the Lactobacillus xylT expression plasmid, pLPA9. The construction of plasmid pLPA9 was performed essentially as described previously for the construction of the L. pentosus xylP gene expression vector, pLPA6 (5). The xylT gene was amplified from L. brevis genomic DNA (by using the Expand high-fidelity system; Boehringer Mannheim) and cloned into the cloning vector pTUT-MCS2 (5). The forward primer (5'-CCTTTGGTACCGAACGTCGTAAGGAGCG-3') created a KpnI site (underlined) and a stop codon (boldface) upstream of the xylT original ribosome binding site (RBS [italics]). The reverse primer (5'-TTACCCATGGTGATCCCACCTCTTTCGTAATCG-3') starting 7 nucleotides downstream of the xylT gene stop codon, generated an NcoI site (underlined) overlapping the ATG codon (boldface) of the gusA gene from plasmid pTUT-MCS2. A putative RBS (italics) was also introduced eight nucleotides upstream of gusA and served as a potential translation start for this reporter gene. The resulting plasmid was digested by BglII and XhoI, and the cloning cassette was introduced between the BamHI and XhoI sites of the hybrid lactobacillus-E. coli shuttle expression vector pLP503(t) (25), yielding pLPA9(t). After NotI digestion and religation, yielding plasmid pLPA9, the stop codon introduced in the forward primer was inframe with the first few codons of the ldh gene (which is part of the ldh expression cassette) and permitted translation of the native XylT protein. Then, 5 µg of plasmid pLPA9 DNA was used to transform L. plantarum 80 as described previously (15).

Enzyme assays. The conversion of D-xylose (200 mM) to D-xylulose by D-xylose isomerase was measured according to the sorbitol dehydrogenase method as described previously (3). D-Xylulose kinase activity was determined by measuring the appearance of ADP formed as a result of the D-xylulose kinase reaction (4). Beta-xylosidase activity was assayed with p-nitrophenyl--D-xylopyranoside (5 mM), as described previously for the measurement of -xylosidase activity in L. pentosus (4). Enzyme activities were determined in cell extracts prepared from cells harvested during the logarithmic phase of growth as described earlier (4), except that cells (resuspended in 500 µl of 50 mM potassium phosphate buffer containing 0.5 mM EDTA and 1 mM dithiothreitol) were disrupted by shaking at full speed (IKA-VIBRAX-VXR; IKA-Labortechnik) for 2 h at 4°C with 100 mg of glass beads (ca. 0.1 to 0.3 mm in diameter; Pertorp Analytical). The protein concentration in these samples was determined by the method of Smith et al. (32).

D-Xylose transport assays. L. plantarum 80 cells cultivated in M medium supplemented with 25 mM glucose and 75 mM L-malate were harvested by centrifugation (5,000 × g, 4°C, 10 min), washed twice with uptake buffer (50 mM potassium phosphate buffer [pH 6.5 or 4.5] containing 2 mM MgSO₄), and resuspended at a concentration of approximately 30 mg (dry weight)/ml. For the transport assay, cells were diluted to a concentration of 3 to 5 mg (dry weight)/ml in 800 µl of uptake buffer. Cells were preenergized at 30°C by incubation with either 50 mM L-malate (at an extracellular pH of 4.5) for 2 min or by incubation with 5 mM of glucose (at an extracellular pH of 6.5) for 5 min. Transport was initiated by the addition of D-[U-¹⁴C]xylose (specific activity, 0.4 µCi/mmol; Amersham) at a final concentration of 100 µM. At given time points, 100-µl samples were taken and diluted in 5 ml of ice-cold 0.1 M LiCl. The samples were rapidly filtered through glass fiber filters (Whatman GF/F) and washed with 2 ml of ice-cold 0.1 M LiCl. The radioactivity on the filter was determined by liquid scintillation analysis. To determine the initial rate of D-xylose uptake, the transport reaction was stopped after 20 s by quenching the whole mixture (total volume, 100 µl) in 5 ml of 0.1 M LiCl. Potential competitors or uncouplers were added 5 s before the initiation of the transport reaction, unless indicated otherwise. Accumulation levels of D-xylose in L. plantarum 80 were calculated by assuming an intracellular volume of 1.5 µl/mg (dry weight) (10). For the measurement of D-xylose transport in L. brevis, the cells were grown in M medium supplemented with 25 mM of the indicated sugar and 25 mM of L-arginine. Transport experiments were performed essentially as described above for L. plantarum 80, except that the cells were energized at pH 6.5, with 20 mM of L-arginine or 5 mM glucose, and added 2 and 5 min prior to the start of the transport reaction, respectively. For L. brevis, D-[U-¹⁴C]xylose (specific activity, 0.1 µCi/mmol) was used at a final concentration of 500 µM.

DNA manipulation. All DNA manipulations, including Southern hybridization, ligation, and transformation of E. coli, were done according to standard procedures (28). All enzymes were used according to the specifications of the manufacturers. Double-stranded DNA sequencing was performed according to the method of Wang (34). The Blast comparisons and the amino acid Pileup comparisons were performed by using the Genetic Computer Group programs through the facilities of the CAOS/CAMM center, Nijmegen, The Netherlands.

Nucleotide sequence accession number. The complete nucleotide sequence of the L. brevis xyl locus is available under GenBank accession number AF045552.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Cloning of the D-xylulose kinase gene and flanking regions. The cloning strategy is summarized in Fig. 1A. Initially, a 2.5-kb NlaIII/EcoRI region from L. brevis was cloned using a PCR probe generated with primers deduced from the N-terminal sequence of the L. brevis D-xylose isomerase and a region conserved among virtually all bacterial D-xylose isomerases. Its reported sequence contained a 1,347-bp open reading frame (ORF) coding for XylA (2). Within the 2.5-kb NlaIII/EcoRI sequence, at 107 bp downstream from the xylA stop codon, an ATG codon at the beginning of an ORF was found. The deduced amino acid sequence of this ORF showed considerable homology to the N-terminal amino acid sequence of the E. coli and L. pentosus D-xylulose kinases (2).

View larger version (29K): [in this window] [in a new window]   FIG. 1.   (A) Physical map and organization of the L. brevis xylABT locus. The upper part shows the xylB and xylT cloning strategy. The stem-loop structures indicate the putative transcriptional terminators. The nucleotide (nt) sequences of upstream regions of xylA and xylT are depicted in panels B and C, respectively. The open boxes denote putative 10 and 35 consensus sequences of the promoters. The putative regulatory elements (cre and xylO) are in boldface italic letters. The potential RBSs are indicated by asterisks. The beginning of the deduced amino acid sequence of the xylA and xylT genes is depicted below the nucleotide sequence. The putative transcriptional terminator located downstream of the xylB gene is underlined by thin arrows (panel C). For clarity, only the NlaIII and Sau3AI sites used for cloning are depicted.

Nucleotide sequence of the xylB and xylT genes. The xylB ORF (1,506 nucleotides long) starts with an ATG codon (located 96 nucleotides after the xylA stop codon) and terminates with a TAG codon. The xylB gene codes for a protein of 502 amino acids and showed a high degree of identical amino acids (66%) when compared to the product of the L. pentosus xylB gene (data not shown). Another ORF (1,371 nucleotides long), xylT, begins 222 bp downstream of the xylB stop codon with an ATG codon and terminates with a TAA codon. Both the xylB and the xylT ORFs are preceded by putative Shine-Dalgarno sequences. No obvious promoter motifs were found immediately upstream of xylB. However, starting 110 bp upstream of the xylT start codon, a xylO sequence was observed, which showed similarity to xylO found in the upstream region of the L. brevis and L. pentosus xylA genes and in the upstream region of the L. pentosus xylP gene (Fig. 1C). 35 and 10 promoter elements (TTTCAA and TATGAT), spaced by 17 nucleotides, were found 145 and 128 bp upstream of the xylT start codon. Moreover, a putative cre site overlapping the 35 element was also identified. Potential Rho-independent transcriptional terminator sequences were found within the noncoding regions after xylA (G° = 17.1 kcal mol¹), xylB (G° = 14.6 kcal mol¹), and xylT (G° = 18.4 kcal mol¹).

Sequence homology of XylT with sugar transporters. The Blast computer program (1) was used to search entries in the Swiss-Prot protein database showing similarity to the deduced amino acid sequence of xylT from L. brevis. XylT demonstrated strong similarity throughout the entire sequence to other bacterial monosaccharide transporters of the "major facilitator superfamily" (MFS) (13, 20, 22), especially to the L-arabinose-H⁺ symporter of E. coli, AraE (60%); the D-galactose-H⁺ symporter of E. coli, GalP (59.5%); the glucose facilitator of Zymomonas mobilis, GlfZ (57%); the D xylose-H⁺ symporters of B. megaterium, XylT (58%) and of E. coli, XylE (57%); and the D-xylose-Na⁺ symporter of T. halophila, XylE (57%). An alignment is presented in Fig. 2. Surprisingly, the similarity score shared between XylT of L. brevis and the three bacterial D-xylose-cation symporters was significantly lower than the score shared between XylE of E. coli and T. halophila and XylT of B. megaterium (73 to 81%). Moreover, XylT of L. brevis shared the highest similarity score with the L-arabinose-H⁺ symporter, AraE, of E. coli.

View larger version (98K): [in this window] [in a new window]   FIG. 2.   Comparison of the primary sequences of XylT of L. brevis (XylT-Lb); GlfZ of Z. mobilis (GlfZ-Zm, Swiss-Prot accession number P21906); XylT of B. megaterium (XylT-Bm, EMBL gene bank accession number Z71474); GalP, AraE, and XylE of E. coli (GalP-Ec, Swiss-Prot accession number P37021; AraE-Ec, Swiss-Prot accession number P09830; XylE-Ec, Swiss-Prot accession number P09098); and XylE of T. halophila (XylE-Th, EMBL gene bank accession number AB009593). The alignment was done by using the Pileup program (9), and some gaps were introduced to maximize the alignment. The identical amino acids are shown in white letters on a solid background. The 12 putative transmembrane segments are indicated by arrows above the alignment. The numbers on the left of the alignment correspond to the amino acid positions for each protein.

Regulation of D-xylose uptake and D-xylose catabolism in L. brevis. It is not known whether D-xylose is required for the expression of the L. brevis xyl genes. We could, however, identify putative xylO sequences in the regions upstream of the xylA and xylT genes, suggesting that their expression would be induced by D-xylose. Moreover, the presence of putative cre sites upstream of the xylA and of the xylT genes suggested a possible negative regulation by glucose and other carbohydrates. Therefore, the activity of D-xylose isomerase and D-xylulose kinase were measured in cell extracts of L. brevis grown in the presence of D-xylose, D-ribose, L-arabinose, D-glucose, maltose, and a mixture of D-glucose and D-xylose. Active D-xylose transport was also measured in L. brevis cells in the presence of an exogenous energy source, L-arginine or D-glucose (Table 1). D-Xylose isomerase and D-xylulose kinase activity and the transport of D-xylose were only detected when cells were grown on D-xylose. Moreover, the addition of D-glucose to cells growing on D-xylose decreased the total activity of D-xylose isomerase and D-xylulose kinase by fourfold, and the D-xylose transport activity was decreased by sixfold. In addition, it has been demonstrated that the repressor gene, xylR, of the Bacillus subtilis and L. pentosus xyl regulons is involved in the negative control of a -xylosidase and of an -xylosidase encoding gene, respectively (4, 12, 17). In L. brevis, we could identify a xylR repressor gene homologue, although it was not found within the xylABT locus. This finding prompted us to investigate the presence of - or -xylosidase activities in L. brevis. No -xylosidase activity could be detected, but a low level of -xylosidase activity was detected in cell extracts with most of the growth substrates (see Table 1). However, this activity was increased about 10-fold in the presence of D-xylose, suggesting a mechanism of induction. The -xylosidase activity was reduced about twofold when both D-glucose and D-xylose were added to the growth medium compared to that of cells grown on D-xylose only.

Functional expression of the D-xylose transport gene in L. plantarum 80. The L. brevis strain studied here could not be transformed with plasmid DNA when standard Lactobacillus electrotransformation procedures were used (unpublished observations). Therefore, the xylT gene could not be inactivated by plasmid integration to determine its role in D-xylose uptake in L. brevis. However, we have recently developed a Lactobacillus expression system which was used to characterize the -xyloside transporter of L. pentosus, XylP (5). Consequently, an xylT expression vector was constructed, pLPA9 (Fig. 3) and was used to transform L. plantarum 80, a strain lacking PMF-linked D-xylose transport activity. To demonstrate that the D-xylose transport gene of L. brevis was functionally expressed in L. plantarum 80, D-xylose transport activity with D-[U-¹⁴C]xylose was assayed under conditions in which a PMF was generated from either malolactic fermentation (21) or glucose fermentation. Under both conditions, L. plantarum 80 harboring plasmid pLPA9 could transport and efficiently accumulate significant amounts of D-xylose, whereas the parental wild-type strain could not (Fig. 4). The initial rates of uptake in L. plantarum 80(pLPA9) were approximately 9 and 12 nmol/min per mg (dry weight) when L-malate and D-glucose were the source of the PMF, respectively. The accumulation level of D-xylose ([D-xylose]_in/[D-xylose]_out) after 2 min was around 60 when the PMF was generated by L-malate uptake and metabolism. The accumulation level was 30-fold when 5 mM glucose was the source of metabolic energy. In addition, increasing the glucose concentration to 20 mM resulted in a slower rate of uptake (ca. 3 nmol/min per mg [dry weight]), and in a fourfold decrease of the D-xylose level of accumulation. These findings suggest that the metabolism of glucose can negatively affect the activity of XylT and/or the maintenance of D-xylose inside the cell. Without L-malate or glucose as source of the PMF, D-xylose could not be transported or accumulated in L. plantarum 80(pLPA9). These results clearly indicate that the xylT gene of L. brevis encodes a D-xylose transporter.

View larger version (22K): [in this window] [in a new window]   FIG. 3.   Structure of the xylT lactobacillus-E. coli shuttle expression vector pLPA9(t). The expression cassette of this vector comprises the strong ldh promoter (Pldh from L. casei ATCC 393), the xylT gene, the E. coli -glucuronidase gene (gusA) used as a marker of gene expression, and the terminator sequence of the L. plantarum 80 cbh gene (Tcbh [6]). The two terminator sequences (Tldh from L. casei ATCC 393 [25]) downstream from the strong Pldh promoter are used to circumvent instability of the expression vector in E. coli. These two Tldh sequences can be eliminated by digestion of the plasmid with NotI and religation, yielding pLPA9. After transformation of L. plantarum 80 with plasmid pLPA9, the transformants expressing the gusA reporter gene can be selected as described in Materials and Methods.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   D-Xylose uptake by cells of L. plantarum 80(pLPA9) (; three independent experiments are plotted) or L. plantarum 80 wild type (). Cells were preenergized at 30°C by incubation with 50 mM L-malate for 2 min at an extracellular pH of 4.5 (panel A) or by incubation with 5 mM D-glucose for 5 min at an extracellular pH of 6.5 (panel B). Uptake of 0.1 mM D-[U-14C]xylose by L. plantarum 80(pLPA9) without PMF-generating conditions is shown in both panels (), and the effect of 20 mM glucose on the accumulation of D-xylose by L. plantarum 80(pLPA9) is indicated () in panel B. Each experiment was performed at least in triplicate, and the standard deviation never exceeded 10%.

Substrate specificity and kinetic parameters of XylT. The kinetic parameters of XylT in L. plantarum 80(pLPA9) were determined with an L-malate-generated PMF. The K_m and V_max for D-xylose transport were 215 ± 15 µM and 35 ± 2 nmol/min per mg (dry weight), respectively. The effects of various sugars or sugar analogs on the initial rate of uptake of D-xylose were also tested (Table 2). An excess of the test substrate (50-fold) was added 5 s before the addition of D-[U-¹⁴C]xylose, and the initial velocity of D-xylose transport was measured (20 s). D-Xylose transport by XylT was poorly inhibited (20%) by a 50-fold excess of L-arabinose and methyl--D-xylose, but no inhibition could be detected with methyl--D-xylose, D-ribose, D-fucose, and D-galactose. 6-Deoxy-D-glucose and D-glucose inhibited the transport of D-xylose. The inhibition of D-xylose transport by 6-deoxy-D-glucose was found to be competitive (Fig. 5), with a K_i of 220 ± 3 µM. As described above, the metabolism of D-glucose in L. plantarum 80 affected both the initial rate of uptake and the accumulation level of D-xylose. Since the metabolic inhibition of D-xylose was different with two different concentrations of D-glucose, the role of D-glucose as a potential competitive inhibitor of D-xylose uptake by XylT could not be assessed in our assay conditions.

View larger version (19K): [in this window] [in a new window]   FIG. 5.   Eadie-Hofstee plot of the D-[U-14C]xylose uptake rate by cells of L. plantarum 80(pLPA9) as a function of the D-xylose concentration, without inhibitor () or with 0.5 mM () or 1 mM () 6-deoxy-D-glucose. Cells were preenergized at 30°C by incubation with 50 mM L-malate for 2 min at an extracellular pH of 4.5. Rates were calculated after an uptake of 20 s.

Role of the PMF components on D-xylose uptake. At pH 4.5, the PMF generated by L-malate transport and metabolism (160 mV) is composed of an electrochemical membrane potential, (70 mV) and of an electrochemical proton gradient, ZpH (90 mV), where Z equals 2.3 (RT/F) and R, T, and F have their usual meanings (21). To determine the role of these components in the transport of D-xylose by XylT, the effects of uncoupling agents on the initial rate of uptake and on the accumulation level of D-xylose were studied (Table 3). The ionophore nigericin (H⁺/K⁺ antiporter), which dissipates the pH, decreased the initial rate of uptake and lowered the accumulation level about 80% when used at a concentration of 0.5 µM or higher. A similar but stronger effect was obtained with the protonophore carbonyl cyanide m-chlorophenylhydrazone (CCCP), which collapses the total PMF.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The aim of this study was twofold. First, we wanted to characterize a D-xylose transporter from a Lactobacillus species and, second, we wanted to construct a system enabling efficient expression of the transporter. The expression system should be of value in the development of food-grade vectors for heterofermentative lactobacilli based on D-xylose fermentation.

Similarity between the L. brevis xylABT locus, the xylABT operon of B. megaterium, and the xylABE operon of T. halophila. The cloning and sequencing of the region downstream from the xylA and xylB genes of L. brevis revealed the presence of a potential xyl gene, xylT, encoding a putative membrane-embedded protein. On the basis of homology between the primary structure of XylT from L. brevis (further referred to as XylT^Lb) and some members of the MFS, including the low-affinity D-xylose-cation symporters XylE of E. coli (further referred to as XylE^Ec), XylE of T. halophila (further referred to as XylE^Th) and XylT of B. megaterium (further referred to as XylT^Bm), a D-xylose-H⁺ symport activity has been assigned to XylT^Lb. The organization of the L. brevis xylABT locus showed similarity to that of the B. megaterium xylABT operon (30) and the T. halophila xylABE operon. No tightly linked xylR gene could be found upstream of xylA in L. brevis. This finding does not exclude, however, that the xylR gene in L. brevis may be located in another operon, perhaps serving to regulate a distinct set of genes. Preliminary results have indicated that the expression of the xylABT genes in L. brevis is inducible by D-xylose and is repressed by D-glucose. Putative xylO sequences (XylR binding sites) and cre-like elements (CcpA binding sites) upstream of xylAB and xylT genes strongly suggest a negative regulation by XylR and CcpA homologues. However, the transcriptional regulation of the L. brevis xyl locus is not yet known, and the role of these transcriptional regulators remains to be demonstrated.

Characteristics of XylT^Lb. The L. brevis xylT gene was functionally expressed in L. plantarum 80 by using a recently described Lactobacillus expression system, which confirmed XylT^Lb to be a D-xylose-H⁺ symporter and enabled the determination of some properties of this transporter. It is interesting to note that throughout the whole sequence, XylT^Lb appeared to be more closely related to the E. coli L-arabinose-H⁺ transport protein, AraE, than to the bacterial D-xylose-cation symporters, XylE^Ec, XylE^Th, and XylT^Bm. This difference extended to the large gap found between the third and the fourth putative transmembrane region in the sequence of XylT^Lb, AraE, and GalP introduced to optimize the alignment with XylE^Ec, XylE^Th, GlfZ, and XylT^Bm. Even though such a homology would suggest a higher affinity of XylT^Lb for L-arabinose than for D-xylose, L-arabinose proved to be a very poor competitor of D-xylose uptake by XylT^Lb. This finding suggests that L-arabinose is presumably not a physiological substrate for the XylT^Lb protein. In fact, the pattern of inhibition of D-xylose transport by several sugars or sugar analogs indicated that the properties of the XylE^Ec and XylT^Lb transport proteins are very similar. In contrast to the AraE protein of E. coli, for which L-arabinose, D-fucose, and D-xylose are substrates (13), transport of D-xylose by XylT^Lb is only inhibited by 6-deoxy-D-glucose (6-methyl-D-xylose) and possibly D-glucose. XylT^Lb can also discriminate between D-xylose and D-xylose analogs with a methyl substitution on the C1 of the pyranoside ring, since neither methyl--D-xylose nor methyl--D-xylose were efficient competitors of D-xylose transport. A similar specificity was previously found for XylE^Ec (13). These results indicate that the primary sequence homology between members of the MFS is clearly not sufficient to predict their substrate specificity. Substrate recognition by these proteins may reside in the positions of specific charged residues (mostly histidine, glutamic acid, or aspartic acid) located in the hydrophilic regions, which may help to bind the substrates (for a review, see reference 23). The alignment shown in Fig. 3, however, did not reveal such conserved residues in XylT^Lb, XylT^Bm, XylE^Ec, and XylE^Th, which could discriminate the D-xylose-cation symporters from the other monosaccharide-cation symporters. Nevertheless, the apparent affinity constant of the XylT^Lb protein for D-xylose (215 µM) is substantially higher than the apparent affinity constant of XylT^Bm (100 µM) and of XylE^Ec (60 µM) for D-xylose (36). Finally, the susceptibility of D-xylose transport to the protonophore CCCP, which collapses the PMF, and to the ionophore nigericin, which discharges the pH, indicates that D-xylose transport by XylT^Lb proceeds most likely in symport with a proton.

Development of a D-xylose food-grade vector. In this study, we have shown that the Lactobacillus expression system, pLPA9, allowed the functional expression of a D-xylose transporter in L. plantarum 80. This plasmid could serve as the basis for the construction of a food-grade vector based on the metabolism of D-xylose by cloning of the xylA and xylB genes downstream of the xylT gene. Such a vector, which makes use of a strong and constitutive Lactobacillus promoter (P_ldh from L. casei ATCC 393) may represent a useful tool for improving the catabolic capacity of heterofermentative lactic acid bacteria that are important in the fermentation of plant materials and in the production of fermented food.