Utilization of d-Ribitol by Lactobacillus casei BL23 Requires a Mannose-Type Phosphotransferase System and Three Catabolic Enzymes

Abstract

Lactobacillus casei strains 64H and BL23, but not ATCC 334, are able to ferment d-ribitol (also called d-adonitol). However, a BL23-derived ptsI mutant lacking enzyme I of the phosphoenolpyruvate:carbohydrate phosphotransferase system (PTS) was not able to utilize this pentitol, suggesting that strain BL23 transports and phosphorylates d-ribitol via a PTS. We identified an 11-kb region in the genome sequence of L. casei strain BL23 (LCABL_29160 to LCABL_29270) which is absent from strain ATCC 334 and which contains the genes for a GlpR/IolR-like repressor, the four components of a mannose-type PTS, and six metabolic enzymes potentially involved in d-ribitol metabolism. Deletion of the gene encoding the EIIB component of the presumed ribitol PTS indeed prevented d-ribitol fermentation. In addition, we overexpressed the six catabolic genes, purified the encoded enzymes, and determined the activities of four of them. They encode a d-ribitol-5-phosphate (d-ribitol-5-P) 2-dehydrogenase, a d-ribulose-5-P 3-epimerase, a d-ribose-5-P isomerase, and a d-xylulose-5-P phosphoketolase. In the first catabolic step, the protein d-ribitol-5-P 2-dehydrogenase uses NAD⁺ to oxidize d-ribitol-5-P formed during PTS-catalyzed transport to d-ribulose-5-P, which, in turn, is converted to d-xylulose-5-P by the enzyme d-ribulose-5-P 3-epimerase. Finally, the resulting d-xylulose-5-P is split by d-xylulose-5-P phosphoketolase in an inorganic phosphate-requiring reaction into acetylphosphate and the glycolytic intermediate d-glyceraldehyde-3-P. The three remaining enzymes, one of which was identified as d-ribose-5-P-isomerase, probably catalyze an alternative ribitol degradation pathway, which might be functional in L. casei strain 64H but not in BL23, because one of the BL23 genes carries a frameshift mutation.

INTRODUCTION

Many bacteria have the capacity to utilize a large number of sugars and sugar derivatives, including sugar alcohols (polyols). For example, hexitols such as mannitol or glucitol are well-established carbon sources for numerous bacteria, including the Gram-negative and Gram-positive model organisms Escherichia coli and Bacillus subtilis. They can either be taken up by ion symporters, such as GutP of B. subtilis (1), or by ABC transporters (2) or can be transported and concomitantly phosphorylated by the phosphoenolpyruvate (PEP):carbohydrate phosphotransferase system (PTS) (3, 4). The three pentitols arabinitol, xylitol, and ribitol are less frequently utilized by bacteria. The first pentitol is also known under the name arabitol and the last one as adonitol. d-Ribitol is present in plants (for example, in Adonis vernalis) (5), and d-ribitol-5-phosphate (d-ribitol-5-P) is also a constituent of teichoic and lipoteichoic acids of certain Gram-positive organisms. Evidence for a d-ribitol transporter and d-ribitol-specific metabolic enzymes was first provided for the bacterium Enterobacter aerogenes (previously called Anaerobacter aerogenes and Klebsiella aerogenes). The d-ribitol dehydrogenase of this organism was purified and characterized in the late 1950s (6, 7). The gene of this enzyme was cloned and its DNA sequence determined (8). It was detected in several other organisms, such as Rhodobacter sphaeroides (9), Rhizobium trifolii (10), Klebsiella pneumoniae (11), Zymomonas mobilis (12), Sinorhizobium meliloti (13), and E. coli (14). A second enzyme was found to be necessary for the metabolism of d-ribitol. It converts d-ribulose into d-ribulose-5-P and was therefore called ribulokinase. The two proteins d-ribitol dehydrogenase and ribulokinase are usually coordinately synthesized (15, 16). After its uptake via an ion symport protein (11, 14) or an ABC transporter (13), d-ribitol is first oxidized to d-ribulose with usually NAD⁺ as an electron acceptor and subsequently phosphorylated in an ATP-dependent reaction to the pentose phosphate intermediate d-ribulose-5-P. In some Gram-positive organisms, d-ribulose-5-P is not metabolized via the pentose phosphate pathway but is reduced in an NADP-requiring reaction to d-ribitol-5-P and subsequently converted together with CTP to CDP ribitol, which is used as a building block for the synthesis of ribitol teichoic and lipoteichoic acids (17). In K. pneumoniae, d-arabinitol was found to be transported and metabolized by principally the same pathway. d-Arabinitol is transported by DalT (11), which resembles GlpT from B. subtilis, and catabolized by the enzymes DalD (d-arabinitol dehydrogenase, which converts d-arabinitol into d-xylulose) and DalK (xylulose kinase) (18).

Transport of pentitols via the phosphoenolpyruvate:carbohydrate phosphotransferase system was first suggested when Jack London and his coworkers succeeded in purifying and characterizing NAD⁺-dependent d-xylitol-5-P and d-ribitol-5-P dehydrogenases from Lactobacillus casei strains Cl83 and Cl16, respectively (19). The latter strain is more commonly known under the name 64H (20), and it lacks ribitol dehydrogenase activity. London and coworkers also succeeded in purifying a soluble d-xylitol-specific PTS protein (called III^Xtl, probably an enzyme IIA [EIIA] component; see below) from L. casei strain Cl83 (21). Finally, a ptsH mutant (ptsH encodes the PTS protein HPr) derived from L. casei 64H was reported to have lost the capacity to utilize d-ribitol (22). Together, these results suggested that another mode of transport and metabolism of the three pentitols might exist in certain bacteria, which was likely to resemble the utilization of hexitols transported by the PTS.

A functional PTS is usually composed of four soluble proteins or protein domains forming a phosphorylation cascade (23). In the first step, enzyme I (EI) autophosphorylates with PEP. P∼EI transfers the phosphoryl group to His-15 in HPr. These two proteins are called the general PTS components. In the next step, P∼His-HPr phosphorylates one of usually several EIIAs present in a bacterium, each being specific for a certain sugar. P∼EIIA transfers the phosphoryl group to an EIIB molecule with the same sugar specificity, which in the last step phosphorylates a carbohydrate molecule bound to the cognate membrane integral EIIC component or, for mannose/glucose-type PTS, to the EIIC and EIID heterodimer (see Fig. 3). Phosphorylation of the sugar is thought to lower its affinity for EIIC, and the phosphorylated carbohydrate is therefore released into the cytoplasm (24). EI, HPr, and the EIIAs as well as the EIIBs of all mannose/glucose-type PTS are phosphorylated at the Nε-1 or Nδ-2 position of a histidyl residue, whereas the EIIBs of all other PTS families are phosphorylated at a cysteyl residue.

Fig 3.

Schematic presentation of d-ribitol transport and catabolism in L. casei BL23. d-Ribitol is transported and phosphorylated by a mannose-type PTS (PTS^Rtl). The phosphoryl group is provided by PEP and is transferred to d-ribitol bound to the integral membrane proteins EIIC^Rtl and EIID^Rtl via a phosphorylation cascade formed by the proteins EI, HPr, EIIA^Rtl, and EIIB^Rtl. In the first metabolic step, intracellular d-ribitol-5-P is oxidized to d-ribulose-5-P in an NAD⁺-requiring reaction catalyzed by the enzyme d-ribitol-5-P 2-dehydrogenase. The enzyme d-ribulose-5-P 3-epimerase converts d-ribulose-5-P into d-xylulose-5-P, which is subsequently cleaved in a P_i-requiring reaction catalyzed by the enzyme d-xylulose-5-P phosphoketolase into d-glyceraldehyde-3-P (GAP) and acetyl-P. d-Glyceraldehyde-3-P metabolized via glycolysis is converted into PEP used for ribitol phosphorylation, which closes the cycle. The gene designations of the different proteins are indicated in italics.

Despite the remarkable work carried out by the group of Jack London on PTS-catalyzed pentitol metabolism, the detailed mechanism of d-ribitol utilization by certain L. casei strains remained obscure, because neither the PTS components involved in its transport nor the enzymes catalyzing the metabolism of d-ribulose-5-P could be identified. We noticed that L. casei strain BL23, the genome of which was sequenced in our laboratory (25), was able to ferment d-ribitol similarly to strain 64H, whereas strain ATCC 334 was not. We identified an operon present in strain BL23, but not ATCC 334, which we suspected to encode the enzymes necessary for d-ribitol transport and metabolism. We deleted the gene encoding one of the PTS components, which, similar to EI inactivation, led to the loss of d-ribitol fermentation. We also cloned four metabolic genes of the ribitol region into His tag expression vectors, expressed them in E. coli, purified the encoded enzymes, and determined their activity. These results allowed us to propose a detailed mechanism for d-ribitol transport and catabolism by L. casei strain BL23.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

L. casei strain BL23 as well as the ptsI and rtlB mutants derived from it were grown at 37°C under static conditions in MRS medium (Difco) or in MRS fermentation medium (Scharlau), which was supplemented with 0.5% d-ribitol. The fermentation medium contains chlorophenol red as a pH indicator, and acid production was determined by monitoring the color change of the medium from red (pH 6.9) to yellow (pH 5.1). For more quantitative assays, the pH was also measured by using a pH electrode. The E. coli strains NM522 (Stratagene) and BL23(DE3) were used for cloning experiments and for protein overproduction. NM522 and NM522(pREP4-GroES/EL) (26) transformed with plasmid pQE30 (Qiagen) harboring various genes of the ribitol region were grown at 37°C under agitation in Luria-Bertani (LB) medium supplemented with 100 μg/ml ampicillin and with 100 μg/ml ampicillin and 30 μg/ml kanamycin, respectively. Lactococcus lactis NZ9000, which was used as an intermediary cloning host for the vector pT1NX and its derivatives, was grown at 30°C under static conditions in M17 medium (Oxoid) containing 0.5% glucose. L. lactis transformants were selected with 5 μg/ml erythromycin.

Construction of the rtlB deletion mutant.

In order to delete the rtlB gene of L. casei strain BL23, two DNA fragments corresponding to the regions upstream and downstream from rtlB were amplified by PCR by using chromosomal DNA of strain BL23 as the template and the primer pairs RtlAamForEco-RtlAamRevNot and RtlCavForNot-RtlCavRevSac (Table 1). The PCR product corresponding to the rtlB upstream region was cut with EcoRI and NotI and inserted into plasmid pRV300 (27) cut with the same enzymes. The second PCR product was inserted into the resulting plasmid after it was cut with NotI and SacI. The resulting plasmid, pRV300-deltaRtlB, was used to transform L. casei BL23. A single-crossover integrant was selected based on its resistance to erythromycin, and the correct insertion was confirmed by PCR using the reverse M13 universal primer and the oligonucleotide RtlBverifSac (Table 1). Subsequently, this integrant was grown in MRS medium without erythromycin for approximately 200 generations. Cells were plated on solid MRS medium and replica plated on MRS medium containing erythromycin. Antibiotic-sensitive clones were isolated, and one of them was selected (strain BL375) in which a second recombination event had caused the excision of the plasmid, leading to the deletion of rtlB. The correct deletion of the EIIB^Rtl-encoding rtlB gene was confirmed by PCR amplification using the primers RtlDhFor and RtlIICRev (Table 1) and DNA sequencing of the PCR product.

Table 1.

Primers used in this study

Name	Sequencea	Use
RtlAamForEco	GTGGGAATTCTGTCTTAATGGGTGTAAGCGAAGAC	rtlB deletion
RtlAamRevNot	CCCCGCGGCCGCTCATGATTGCGGGTTACGTGATTTC	rtlB deletion
RtlCavForNot	TGATGCGGCCGCAGACGTTAAAAATCTGCTGAAATAA	rtlB deletion
RtlCavRevSac	ATAAGAGCTCCGAAAATGGCAATCGGCAAGATAGC	rtlB deletion
RtlBverifEco	CTGCAACCATGCAGTCATGACC	rtlB verification
RtlBverifSac	GTGCTTCTGATGATGAGAAAGC	rtlB verification
RtlDhFor	GCATTTGAAGCTATTGTCGC	rtlB deletion
RtlIICRev	GACTTGTTTAACTGGACCTG	rtlB deletion
RtlBbglII	CAATAGATCTACTGAGGGGGAAATTATAATGTC	rtlB complementation
RtlBspeI	TATCACTAGTCCTCCATTTTTATTTCAGC	rtlB complementation
tdhForBamHI	AGGCTGGATCCATGTTAAAGAACCCTGACGTATCAAG	RtpD purification
tdhRevPtsI	ATGTTCTGCAGTTACCATTCAAATTTAAGAACAG	RtpD purification
R5PepiBamF	GGAGGATCCATGTCTATTGAGATTGCACCC	Rpe purification
R5PepiSalR	TAAGTCGACCCGTCTAGGAATTTGGTCATC	Rpe purification
PketolBglF	ATGAGATCTATGGACCAAACAATTGAGAAAAC	Xpk purification
PketolhinR	ATGAAGCTTTTACATGTGCGATGACGTGCTG	Xpk purification
R5PIsoBamF	TGCGGATCCATGAATCAAAATGACTTGAAGC	RpiA purification
R5PIsoHinR	TAAAAGCTTTTGAATATTATGACCCAGATAA	RpiA purification
2D5PAldBamF	TTGGGATCCTTGAGTAAATTGAGACGATTG	DeoC-like purification
2D5PAldhinR	ATGAAGCTTACCCTCCTTATCTAGTCATTTG	DeoC-like purification
R5PDHBamF	GTGGGATCCATGACAAAGATAAAAGCTGCTG	LCABL_29170 purification
R5PDHHinR	AGCAAGCTTGAATCCGTCTTTTCTTACTATTAG	LCABL_29170 purification

^aRestriction sites are in italics.

Complementation of the rtlB mutant with plasmid-encoded rtlB.

The coding region of rtlB was amplified by PCR using L. casei BL23 chromosomal DNA as the template and primers RtlBbglII and RtlBspeI (Table 1). The amplified DNA fragment was digested with BglII and SpeI and cloned into the replicative vector pT1NX (28), which had been previously cut with the same enzymes. In the resulting plasmid, pT1-rtlB, the rtlB gene was expressed under the control of the lactococcal P1 constitutive promoter. This plasmid was used to transform the L. casei rtlB mutant BL375. One transformant was selected and named PL47. For control experiments, a transformant carrying empty pT1NX was also isolated.

Cloning of the different metabolic genes of the ribitol regulon into His tag expression vectors.

The ribitol region present in strain BL23 contains six genes encoding presumed catabolic enzymes (Fig. 1). Five of these genes were amplified by PCR using chromosomal DNA of strain BL23 as the template and the appropriate primer pair (Table 1). Together, the LCABL_29180 and _29190 genes of strain BL23 encode a homologue of DeoC from E. coli and B. subtilis (Fig. 1). Owing to a frameshift mutation, this gene was disrupted into two open reading frames (ORFs). A DNA fragment with a sequence nearly identical to that of the LCABL_29180 and LCABL_29190 genes but without the frameshift mutation could be amplified by PCR using chromosomal DNA of L. casei strain 64H as the template and the primer pair 2D5PAldBamF-2D5PAldhinR (Table 1). The six PCR products corresponding to the various genes were cut with the appropriate restriction enzymes and cloned into the His tag expression vector pQE30 (Qiagen). The correct sequence of each gene was confirmed by DNA sequencing. The resulting plasmids carrying the rtpD gene of L. casei BL23 or the deoC-like gene amplified from strain 64H were used to transform E. coli strain NM522. The proteins encoded by the other four genes could not be obtained in soluble form because they formed inclusion bodies when produced in strain NM522. The pQE30-derived plasmids carrying the rpiA, LCABL_29170, rpe, and xpk genes were therefore used to transform the E. coli strain NM522(pREP4-GroES/EL). This strain carries the plasmid pREP4 (Qiagen) containing the genes for the chaperone GroES/GroEL (26). Indeed, overproduction of three of the encoded proteins in strain NM522(pREP4-GroES/EL) rendered them soluble and allowed their purification (Fig. 2). Only the protein encoded by LCABL_29170 remained exclusively present in inclusion bodies.

Fig 1.

Schematic presentation of the ribitol region of L. casei strain BL23. A ribitol region identical to that of strain BL23 is also found in strains W56, 32G, CRF28, BD-II, and LC2W but not in ATCC 334. The region contains 11 genes, 4 of which code for a mannose-type PTS, 1 for a transcription regulator, and 6 for enzymes presumably involved in ribitol metabolism. The location of the frameshift mutation in the deoC-like gene of strain BL23 creating two ORFs (LCABL_29180 and LCABL_29190) is indicated with an arrow (for details, see legend to Fig. 4). In strain BL23, the ribitol region is inserted between the genes LCABL_29150 and LCABL_29280, which correspond to LSEI_2739 and LSEI_2743 in the genome sequence of strain ATCC 334, in which only three genes encoding a mannitol/fructose-type PTS are inserted at this locus. The L. casei strains 21/1, 12A, and Zhang have a gene arrangement identical to that of ATCC 334. In contrast, similar to BL23, strain M36 as well as Lpc-37, A2-362, and T71499 also contains a large inserted region coding not only for the components of a galactitol-type PTS specific probably for l-ribitol or l-arabinitol but also for the enzymes necessary for the metabolism of l-ribulose-5-P. In these strains, the insertion also occurred between homologues of LCABL_29150 and LCABL_29280, and compared to strains ATCC 334 and BL23, the integration site is shifted by only 3 bp at the beginning and 7 bp at the end of the inserted region. “DH” indicates the presumed l-ribitol or l-arabinitol dehydrogenase present in strains M36, Lpc-37, A2-362, and T71499.

Fig 2.

Electrophoretic separation of five His-tagged proteins encoded by genes of the ribitol region of L. casei strain BL23 (lanes a to d) or 64H (lane e) on a 0.1% SDS-12.5% polyacrylamide gel. For each protein, 5 μg was loaded on lanes a to e. Their calculated molecular weight (MW) (including the His tag) are as follows: ribitol-5-P 2-dehydrogenase (lane a), 41,330; d-ribulose-5-P 3-epimerase (lane b), 24,771; d-xylulose-5-P phosphoketolase (lane c), 91,165; d-ribose-5-P isomerase (lane d), 26,407; and presumed d-2-deoxyribose-5-P aldolase (lane e), 30,410. The d-xylulose-5-P phosphoketolase seemed to be partly degraded during purification, because several minor small fragments were coeluted with the enzyme during purification by ion chelate affinity chromatography on an Ni-NTA agarose column (lane c).

Production and purification of His-tagged enzymes.

For the production of His-tagged proteins, E. coli NM522 strains carrying the pQE30-derived plasmids containing the genes rtpD or deoC-like were grown at 37°C to an optical density at 600 nm (OD₆₀₀) of 0.5 to 0.6 before 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) was added. Growth was continued for 3 h before the cells were harvested by centrifugation. NM522(pREP4-GroES/EL) strains carrying the pQE30-derived plasmids containing the rpiA, LCABL_29170, rpe, and xpk genes were grown at 37°C to an OD₆₀₀ of 0.5 to 0.6 before the temperature of the growth medium was lowered to 28°C and 1 mM IPTG was added. Growth continued at 28°C overnight before the cells were harvested by centrifugation. Preparation of crude extracts from the various transformants and purification of the His-tagged proteins by ion chelate affinity chromatography on nickel-nitrilotriacetic acid (Ni-NTA) agarose columns was carried out as previously described (29).

Spectrophotometric enzyme activity assays.

In order to measure the activities of the purified enzymes, we set up a series of spectrophotometric assays. For measuring the activity of d-ribitol-5-P 2-dehydrogenase (RtpD; EC 1.1.1.137), we used a 0.5-ml assay mixture containing 50 mM Tris-HCl (pH 7.4), 5 mM MgCl₂, 0.5 mM NADH (Sigma-Aldrich, Saint Quentin Fallavier, France), and 0.5 mM d-ribulose-5-P (Sigma-Aldrich). After preincubation for 10 min, the reaction was started by adding 9 μg d-ribitol-5-P 2-dehydrogenase. The disappearance of NADH was monitored by measuring the absorption at 340 nm by using an UVIKON 9X3 W double-beam spectrophotometer (Kontron Bio-Tek) with the “Autorate” program. In order to confirm the dependence of RtpD on divalent cations, EDTA was added at final concentrations between 5 and 50 mM and preincubated for 10 min before the reaction was started. To determine the activity of d-ribulose-5-P 3-epimerase (EC 5.1.3.1), we used d-xylulose-5-P (Sigma-Aldrich) as an initial substrate, which this enzyme transforms into d-ribulose-5-P. The subsequent NADH-requiring reduction of d-ribulose-5-P was measured as described above. The corresponding 0.75-ml assay mixture contained 50 mM Tris-HCl (pH 7.4), 5 mM MgCl₂, 0.5 mM NADH, 0.5 mM d-xylulose-5-P, and 18 μg d-ribitol-5-P 2-dehydrogenase, which corresponds to 1.3 enzyme units. The reaction was started by adding 15 μg d-ribulose-5-P 3-epimerase, and the disappearance of NADH was monitored by measuring the absorption at 340 nm. A similar assay was used to determine the activity of d-ribose-5-P isomerase (EC 5.3.1.6) by starting from d-ribose-5-P (Sigma-Aldrich). The 0.75-ml assay mixture contained 50 mM Tris-HCl (pH 7.4), 5 mM MgCl₂, 0.5 mM NADH, 0.5 mM d-ribose-5-P, and 18 μg (1.3 U) d-ribitol-5-P 2-dehydrogenase. The reaction was started by adding 22 μg d-ribose-5-P isomerase, and the disappearance of NADH was monitored by measuring the absorption at 340 nm. d-Xylulose-5-P phosphoketolase (EC 4.1.2.9) is a lyase that uses inorganic phosphate (P_i) to split d-xylulose-5-P into d-glyceraldehyde-3-P and acetylphosphate. Its activity was determined by measuring the formation of d-glyceraldehyde-3-P, which after its transformation into dihydroxyacetonephosphate by triosephosphate isomerase (Sigma-Aldrich) was reduced to glycerol-3-P in an NADH-requiring reaction catalyzed by glycerol-3-P dehydrogenase (Sigma-Aldrich). The 0.75-ml assay mixture contained 50 mM potassium phosphate buffer (pH 7.4), 5 mM MgCl₂, 0.5 mM NADH, 0.5 mM d-xylulose-5-P (Sigma-Aldrich), 50 μg triosephosphate isomerase (5 U), and 2.5 μg d-glycerol-3-P dehydrogenase (25 U). In order to demonstrate the importance of phosphate for the d-xylulose-5-P phosphoketolase reaction, the phosphate buffer was replaced with 50 mM Tris-HCl (pH 7.4). A possible effect of thiamine pyrophosphate was tested by including this compound in the assay mixture at concentrations between 0.15 and 0.5 mM. The reaction was started by adding 10 μg d-xylulose-5-P phosphoketolase, and the disappearance of NADH was monitored by measuring the absorption at 340 nm.

A similar assay based on the formation of d-glyceraldehyde-3-P was used to determine the activity of the presumed 2-deoxy-d-ribose-5-P aldolase (EC 4.1.2.4) of strain 64H, which was supposed to cleave 2-deoxy-d-ribose-5-P into d-glyceraldehyde-3-P and acetaldehyde. The 0.75-ml assay mixture contained 50 mM Tris-HCl (pH 7.4), 5 mM MgCl₂, 0.5 mM NADH, 0.5 mM d-2-deoxyribose-5-P, 5 μg triosephosphate isomerase (5 U), and 5 μg d-glycerol-3-P dehydrogenase (50 U). However, when various amounts of the presumed purified 2-deoxy-d-ribose-5-P aldolase derived from L. casei strain 64H were added, no decrease of the absorption at 340 nm was detected.

Nucleotide sequence accession number.

The DNA sequence of the entire deoC-like gene of L. casei strain 64H has been deposited at the EMBL database under accession number HF562365.

RESULTS

L. casei strain BL23 can utilize d-ribitol and transports it via a PTS.

L. casei strain 64H has previously been reported to be able to utilize d-ribitol (19). When carrying out fermentation assays at 37°C under static conditions in MRS fermentation medium containing 0.5% d-ribitol, we observed that L. casei strain BL23 was also able to ferment this carbon source, whereas strain ATCC 334 was not. d-Ribitol was suggested to be transported in L. casei strain 64H via a specific PTS, because in contrast to other d-ribitol-utilizing bacteria, this organism contains a d-ribitol-5-P dehydrogenase and not a d-ribitol dehydrogenase (19). In addition, a ptsH mutant had lost the capacity to utilize ribitol (22). We therefore presumed that strain BL23 might also transport d-ribitol via a PTS. In order to test this hypothesis, we carried out fermentation studies with a ptsI deletion mutant (lacking the general PTS component enzyme I [Fig. 3]) derived from L. casei BL23 (30). Indeed, the ptsI mutant had lost the capacity to utilize d-ribitol as a carbon source. The pH in the ribitol-containing MRS fermentation medium dropped to about 5 during growth of the wild-type strain, whereas it remained close to the initial 6.9 for the ptsI mutant (Table 2), indicating that strain BL23 takes up the pentitol via a PTS.

Table 2.

Acidification of the fermentation medium after 48 h of growth of the L. casei wild-type strain BL23 and mutants derived from it

Strain (genotype)	pH value reached after growth for 48 ha
BL23	5.05 ± 0.12
BL126 (ptsI)	6.77 ± 0.09
BL375 (rtlB)	6.89 ± 0.02
PL47 (rtlB) complemented	5.33 ± 0.04

^aThe mean values of three independent experiments together with the calculated standard deviations are presented.

We also carried out growth studies with the wild-type strain and the ptsI mutant. However, growth studies turned out to be difficult to perform because the cells grew very slowly and when the wild-type strain reached an OD₅₉₅ of about 1, the cells started to lyse (see Fig. S1 in the supplemental material). Similar problems were encountered when L. casei strains were grown on other polyols such as myo-inositol (31). Because polyols require an additional oxidation step compared to sugars in order to be converted into glycolytic intermediates, a perturbation of the NADH/NAD⁺ balance might be responsible for the growth difficulties. There was nevertheless a reproducible difference between the growth behavior of the wild-type strain and that of the ptsI mutant. While after about 28 h of growth the wild-type strain transiently reached an OD₅₉₅ of more than 1, the ptsI mutant barely exceeded an OD₅₉₅ of 0.6 (see Fig. S1).

The LCABL_29210-LCABL_29240 genes encode the d-ribitol-specific mannose-type PTS components.

d-Ribitol-5-P 2-dehydrogenase of L. casei strain 64H had been purified to homogeneity (19). It was reported to possess a molecular mass around 40 kDa when migrating on an SDS-polyacrylamide gel. Streptococcus pneumoniae contains a protein belonging to the medium-chain dehydrogenase/reductase/zinc-dependent alcohol dehydrogenase family. This enzyme was reported to use NADPH to reduce d-ribulose-5-P to d-ribitol-5-P, which is subsequently converted into CDP-ribitol, used for the synthesis of ribitol-containing teichoic and lipoteichoic acids of this organism (17). This enzyme also has a molecular mass of about 40 kDa, and we therefore suspected that the streptococcal anabolic d-ribitol-5-P 2-dehydrogenase and the catabolic L. casei enzyme might be related. We carried out a BLAST search with the amino acid sequence of the streptococcal protein in the genome sequence of L. casei BL23 (25) in order to possibly detect the d-ribitol-specific genes of this organism. We indeed picked up one L. casei BL23 protein (LCABL_29250) that exhibits 35% sequence identity to the streptococcal protein. The L. casei protein was annotated as threonine dehydrogenase (Tdh). It is part of an 11-kb region that is absent from strain ATCC 334 (Fig. 1). In BL23, the tdh-like gene is followed by 9 other genes transcribed in the same direction and presumably forming an operon. Four of them encode the sugar-specific EIIA, -B, -C, and -D components of a PTS of the mannose/glucose family. In order to test whether the mannose-type PTS is indeed involved in d-ribitol transport, we deleted the gene encoding the EIIB component (LCABL_29230) as described in Materials and Methods. The resulting mutant, BL375, had indeed lost its capacity to utilize d-ribitol as a carbon source. During growth in ribitol-containing MRS fermentation medium, the pH remained close to 7 (Table 2), indicating that the genes LCABL_29240 to LCABL_29210 code for the d-ribitol-specific PTS (PTS^Rtl), and they were therefore renamed rtlA, rtlB, rtlC, and rtlD (Fig. 1).

Complementation of the rtlB mutant restores ribitol fermentation.

In order to complement the rtlB mutant BL375 with the rtlB gene, the latter was amplified by PCR and inserted into vector pT1NX (28). In the resulting plasmid, pT1-rtlB, the rtlB gene is expressed under the control of the constitutive lactococcal P1 promoter (see Materials and Methods). Transformation of the rtlB mutant with pT1-rtlB indeed restored ribitol fermentation to about the same level as was observed for the wild-type strain BL23 (Table 2), whereas the rtlB mutant transformed with empty pT1NX was not able to ferment ribitol (data not shown). We also carried out growth studies with the rtlB mutant BL375 and the complemented strain PL47. While the rtlB mutant behaved similarly to the ptsI mutant, the complemented strain PL47 grew almost identically to the wild-type strain, with a transient increase of the OD₅₉₅ to more than 1 (see Fig. S1 in the supplemental material). These results further established that the PTS encoded by the rtlA, rtlB, rtlC, and rtlD genes is essential for ribitol uptake by L. casei strain BL23.

LCABL_29250 encodes an NADH-dependent d-ribitol-5-P 2-dehydrogenase.

The similarity of LCABL_29250 to the S. pneumoniae d-ribitol-5-P 2-dehydrogenase suggested that this protein might have the same catalytic function. We therefore cloned the gene in the His tag expression vector pQE30, purified the protein (Fig. 2, lane a), and carried out a spectrophotometric assay using d-ribulose-5-P as the substrate as described in Materials and Methods. LCABL_29250 indeed reduced d-ribulose-5-P to d-ribitol-5-P, thus confirming that this protein is a d-ribitol-5-P 2-dehydrogenase. A specific activity of 76.8 μmol per min and mg enzyme (Table 3) was calculated for LCABL_29250, which was renamed RtpD. However, in contrast to the streptococcal enzyme, LCABL_29250 used exclusively NADH as a reducing cofactor. No reaction occurred with NADPH. Absolute specificity for NADH has also been reported for d-ribitol-5-P 2-dehydrogenase isolated from L. casei strain 64H (19). The enzyme of strain BL23 was also very specific for its sugar substrate and could not reduce d-xylulose-5-P, a stereoisomer of d-ribulose-5-P, to d-xylitol-5-P. Although the enzyme is a member of the Zn²⁺-dependent alcohol dehydrogenase family, adding Zn²⁺ ions at various concentrations did not increase the activity of the enzyme. It is possible that RtpD binds Zn²⁺ with high affinity, thus preventing a significant loss of the cofactor during purification. This assumption is in agreement with the observation that it required relatively high concentrations of EDTA (25 mM) to see an about 3-fold inhibition of the activity of RtpD (data not shown). The RtpD activity could be restored by adding 25 mM Zn²⁺. In contrast, no increase in enzyme activity was observed when 25 mM Mg²⁺ or Mn²⁺ ions was added to the EDTA-containing assay mixture. Complete inhibition of RtpD occurred at 50 mM EDTA. The enzyme therefore behaves similarly to previously characterized members of the Zn²⁺-dependent alcohol dehydrogenase family (32).

Table 3.

Specific activities of purified L. casei BL23 enzymes related to the metabolism of ribitol in this organism

Sp act (μmol substrate formed/min and mg protein)
RtpD, ribitol-5-P 2-dehydrogenase	Rpe, ribulose-5-P 3-epimerase	Xpk, xylulose-5-P phosphoketolase	RpiA, ribose-5-P isomerase
76.8	24.5	10.4	18.2

A d-ribulose-5-P 3-epimerase catalyzes the second step of d-ribitol catabolism.

The protein encoded by LCABL_29200, located downstream from the genes coding for the d-ribitol-specific PTS components, was annotated as d-ribulose-5-P 3-epimerase, an enzyme of the pentose phosphate pathway that catalyzes the interconversion of d-ribulose-5-P and d-xylulose-5-P. In order to test whether this protein exhibits indeed the predicted activity, we cloned the LCABL_29200 gene into the His tag expression vector pQE30, purified the protein (Fig. 2, lane b), and set up a coupled spectrophotometric assay as described in Materials and Methods. LCABL_29200 was indeed able to convert d-xylulose-5-P into d-ribulose-5-P, which was subsequently reduced to d-ribitol-5-P in an NADH-consuming reaction. A specific activity of 24.5 μmol per min and mg enzyme (Table 3) was calculated for LCABL_29200. The gene encoding this protein was therefore renamed rpe, for d-ribulose-5-P 3-epimerase (Fig. 1).

A gene encoding a repressor is located upstream from rtpD.

LCABL_29260, the gene located upstream from rtpD and oriented in the opposite direction, encodes a putative repressor of the DeoR family resembling, among others, GlpR from E. coli and IolR from B. subtilis (48% sequence similarity). Binding of repressors of this family to often two or more imperfect direct repeats is usually prevented by the interaction with a phosphorylated carbohydrate or catabolic intermediate, as has been shown for the E. coli glp (33, 34) and the B. subtilis iol regulons (35). A perfect 10-bp direct repeat (TTTGCAAACTTTTGCTTTAACCATTTTTGCAAACT) containing an 8-bp perfect palindrome (in italics) is located in the middle of the noncoding region between the divergently oriented rtpD gene and the presumed regulator gene. It is therefore likely that in response to the presence or absence of ribitol, the protein encoded by LCABL_29260 binds to the direct repeats and thereby controls the expression of the divergently oriented genes of the ribitol region, including LCABL_29260 and LCABL_29270 (Fig. 1). We therefore renamed the LCABL_29260 gene rtlR.

A d-xylulose-5-P phosphoketolase catalyzes the third step of d-ribitol catabolism.

LCABL_29270, the gene downstream from rtlR, encodes a protein exhibiting significant sequence similarity to d-xylulose-5-P phosphoketolase (EC 4.1.2.9). Although the gene encoding the presumed d-xylulose-5-P phosphoketolase in L. casei BL23 is oriented in opposite direction to the ribitol genes, it probably belongs to the d-ribitol regulon because it is part of the about 11-kb region present in strain BL23 but absent from strain ATCC 334 (Fig. 1). This enzyme uses inorganic phosphate (P_i) to cleave d-xylulose-5-P into d-glyceraldehyde-3-P and acetyl-P. d-Xylulose-5-P phosphoketolase was first described for heterofermentative lactobacilli (36, 37), but it is present also in other organisms, such as Fibrobacter (38) and Clostridium acetobutylicum (39). This phosphoketolase is involved in the metabolism of at least part of the xylulose taken up and phosphorylated by these organisms. It was therefore likely that in L. casei BL23, this enzyme cleaves d-xylulose-5-P formed by the d-ribulose-5-P 3-epimerase. In order to test this hypothesis, the LCABL_29270 gene was cloned into plasmid pQE30, the His-tagged protein was purified (Fig. 2, lane c), and its presumed d-xylulose-5-P phosphoketolase activity was measured by setting up a coupled spectrophotometric assay (see Materials and Methods). LCABL_29270 indeed functions as d-xylulose-5-P phosphoketolase, with a specific activity calculated to be 10.4 μmol per min and mg enzyme (Table 3). d-Xylulose-5-P phosphoketolases use thiamine pyrophosphate as a cofactor (40), which is usually covalently bound to the enzyme, and a loss during purification was therefore unlikely. We nevertheless included this cofactor in the assay mixture at various concentrations (0.15 to 0.5 mM), which, however, did not increase the activity of d-xylulose-5-P phosphoketolase (data not shown). As expected, the activity drastically dropped when the phosphate buffer in the assay mixture was replaced with 50 mM Tris-HCl. No activity at all was observed when d-xylulose-5-P was replaced with d-ribulose-5-P. However, when both d-ribulose-5-P and d-ribulose-5-P 3-epimerase were present in the assay mixture, formation of d-glyceraldehyde-3-P could be detected again. In summary, these results suggest that d-ribitol transported by the PTS and converted into d-ribitol-5-P is first intracellularly oxidized to d-ribulose-5-P by the enzyme d-ribitol-5-P 2-dehydrogenase (RtpD) (Fig. 3). d-Ribulose-5-P is subsequently transformed into d-xylulose-5-P by the enzyme d-ribulose-5-P 3-epimerase. In the last step, d-xylulose-5-P phosphoketolase cleaves d-xylulose-5-P into the glycolytic intermediate d-glyceraldehyde-3-P and acetyl-P. This allows L. casei to synthesize one PEP and two ATP molecules when metabolizing d-ribitol, with the PEP molecule being used for the uptake and phosphorylation of the pentitol.

Three additional genes are associated with the ribitol utilization region of BL23.

The d-ribitol region encoding the d-ribitol PTS components contains at its end four additional ORFs (LCABL_29160 to LCABL_29190) which possibly encode the enzymes for an alternative pathway for d-ribitol catabolism. However, this pathway is not operative in L. casei strain BL23, because one of the genes appeared to be split into two ORFs (LCABL_29180 and LCABL_29190). The two protein fragments exhibit significant sequence similarity (65%) to aldolases of the DeoC/LacD family present in, for example, E. coli, lactobacilli, and clostridia. DeoC is a d-2-deoxyribose-5-P aldolase (41), which cleaves d-2-deoxyribose-5-P into d-glyceraldehyde-3-P and acetaldehyde. Because L. casei strain 64H is also able to utilize d-ribitol (14), we PCR amplified the corresponding gene from this organism. The resulting PCR product exhibited a DNA sequence nearly identical to that of LCABL_29180 and LCABL_29190, except that a deletion of 10 bp had occurred in strain BL23, which introduced a frameshift leading to the two ORFs observed in strain BL23 (Fig. 4). The same deletion was found in L. casei strains W56 (42), BD-II (43), and LC2W (44), whereas the two strains 32G and CRF28 contained an intact deoC-like gene. However, the two latter strains are probably not able to utilize ribitol, because both contain a frameshift mutation in the d-ribitol-5-P 2-dehydrogenase-encoding rtpD gene. In order to test whether the DeoC-like enzyme functions indeed as a d-2-deoxyribose-5-P aldolase, we cloned the amplified 64H gene into the His tag plasmid pQE30 and purified the encoded protein (Fig. 2, lane e). However, when d-2-deoxyribose-5-P was incubated with the L. casei 64H-derived protein, no formation of d-glyceraldehyde-3-P could be detected in a coupled spectrophotometric assay. One possible explanation for this negative result might be that the enzyme becomes inactivated during purification. It is also possible that the gene encoding this protein in strain 64H carries a mutation leading to the inactivation of the enzyme or, finally, that it does not possess the presumed d-2-deoxyribose-5-P aldolase activity.

Fig 4.

Frameshift-creating deletion in the deoC-like gene of strain BL23 encoding a presumed d-2-deoxyribose-5-P aldolase. The deletion of 10 bp in the deoC-like gene of strain BL23 compared to the deoC-like gene of strain 64H occurred at position 200 and led to the formation of two ORFs (LCABL_29180 and LCABL_29190). Five other L. casei strains also possess a ribitol region containing a deoC-like gene followed by a homologue of LCABL_29170 and an rpiA-like gene (see Fig. 1). In three of these strains (W56, BD-II, and LC2W), the same frameshift mutation is found in deoC, suggesting that these strains are closely related to BL23. In contrast, the frameshift mutation is absent from strains 32G and CRF28, which contain an intact deoC gene identical to that of strain 64H. The numbers refer to the position in the deoC-like gene of strain 64H and to the position in the genome sequence of strain BL23 (25).

The protein encoded by the last gene of this operon, LCABL_29160, exhibits significant similarity to the pentose phosphate pathway enzyme d-ribose-5-P isomerase. This enzyme transforms d-ribulose-5-P into d-ribose-5-P. In order to determine whether LCABL_29160 exhibits indeed the presumed d-ribose-5-P isomerase activity, the corresponding gene was cloned into plasmid pQE30. The encoded protein was subsequently purified (Fig. 2, lane d), and its activity was determined by using a coupled spectrophotometric assay (see Materials and Methods). LCABL_29160 was indeed able to convert d-ribose-5-P into d-ribulose-5-P, and its specific activity was determined to be 18.2 μmol per min and mg enzyme. Finally, the protein encoded by the penultimate gene of the region from LCABL_29250 to LCABL_29160 is annotated as Zn-containing alcohol dehydrogenase. When cloned into various His tag expression vectors, this gene was only barely expressed in E. coli and formed inclusion bodies. We therefore were not able to purify this protein in an active form and to determine its function. However, it is likely that this protein reduces d-ribose-5-P formed from d-ribulose-5-P by the enzyme d-ribose-5-P isomerase (LCABL_29160) to d-2-deoxyribose-5-P, which might subsequently be cleaved by the presumed d-2-deoxyribose-5-P aldolase (LCABL_29180 and LCABL_29190) into d-glyceraldehyde-3-P and acetaldehyde.

DISCUSSION

While the hexitols mannitol, glucitol (also called sorbitol), and galactitol are often transported via a PTS, pentitols seem to be less frequently taken up via this transport system. The occurrence of PTS-catalyzed ribitol transport was suggested by the identification of several pentitol phosphate dehydrogenases. The first pentitol phosphate dehydrogenase characterized was a d-ribitol-5-P 2-dehydrogenase from Lactobacillus plantarum (45). However, genome sequencing revealed that in several L. plantarum strains, the gene encoding a d-ribitol-5-P 2-dehydrogenase is located in the teichoic acid biosynthesis region, and it probably catalyzes the anabolic conversion of ribulose-5-P into ribitol-5-P, which is subsequently integrated into teichoic acids, similar to what has been reported for S. pneumoniae (17). The first strong support for PTS-catalyzed ribitol transport was obtained for L. casei strain 64H, because this strain was able to ferment ribitol and contains a ribitol-5-P 2-dehydrogenase (19), and a ptsH mutant derived from it had lost the capacity to ferment ribitol (22). A xylitol-specific PTS component (21) and a xylitol-5-P dehydrogenase (19) were purified from L. casei Cl83, confirming that this organism transports xylitol via a PTS. Finally, evidence for PTS-catalyzed transport of d-arabinitol, the third pentitol, was obtained for Listeria monocytogenes. A transposon mutant had been isolated which was no longer able to utilize d-arabinitol and which had the transposon integrated into a gene encoding an EIIC component of a PTS of the galactitol family (46). The genes encoding the three PTS components are followed by two genes encoding potential d-arabinitol-5-P dehydrogenases. A d-arabinitol-phosphate dehydrogenase oxidizing d-arabinitol-1-P to l-xylulose-5-P and d-arabinitol-5-P to ribulose-5-P was also detected in Enterococcus avium (47). This enzyme uses NAD⁺ as well as NADP⁺ as cofactors and depends on Mn²⁺ and not Zn²⁺. The gene encoding this enzyme is located in an operon containing also the genes for the components of a galactitol PTS (47). In addition, the enterococcal enzyme exhibits 64% sequence identity to one of the presumed listerial d-arabinitol-phosphate dehydrogenases (Lmo2664) and 31% identity to the enzyme encoded by the neighboring gene. In conclusion, there was strong evidence that all three pentitols can be transported via a PTS. However, in no case has an in-depth study of PTS-catalyzed pentitol transport and subsequent catabolism of the pentitol phosphate been carried out.

The results obtained during this study confirm that L. casei strain BL23 is able to take up the pentitol d-ribitol via a PTS of the mannose/glucose family composed of four ribitol-specific proteins. In strain BL23, catabolism of d-ribitol-5-P formed during PTS-catalyzed transport is achieved by the three additional enzymes d-ribitol-5-P 2-dehydrogenase, d-ribulose-5-P 3-epimerase, and d-xylulose-5-P phosphoketolase (Fig. 3). A BLAST search revealed that the genes encoding the seven proteins required for d-ribitol uptake and metabolism are absent not only from strain ATCC 334 but also from most of the other 18 L. casei strains for which the genome sequence has been determined (25, 42–44, 48–50). In fact, in addition to BL23, the entire ribitol region is present in only five other L. casei strains: W56, 32G, CRF28, BD-II, and LC2W. These five strains also possess the three additional genes located upstream from the LCABL_29250-LCABL_29160 gene cluster. Three of them contain a deletion of a small sequence in the presumed deoC gene identical to the deletion detected in BL23 leading to a frameshift mutation (Fig. 4). Only the two recently sequenced strains, CRF28 and 32G (49), possess a deoC-like gene resembling LCABL_29180 and LCABL_29190, which however lacks the frameshift-inducing deletion; they therefore contain a gene nearly identical to the presumed deoC present in strain 64H. In these three strains the entire ORF is translated, but it remains nevertheless questionable whether the encoded protein is functional. Using a coupled spectrophotometric assay, we failed to demonstrate that the L. casei 64H protein possesses the presumed d-2-deoxyribose-5-P aldolase activity.

Two L. casei strains used for the production of marketed probiotic drinks, L. casei Shirota (Yakult) and L. casei defensis (Danone), were recently also shown to ferment ribitol (51). However, although this work has been published, the genome sequence data are not yet accessible. Surprisingly, in the same study, L. casei strain BL23 was also tested and was found not to ferment ribitol when using the API CH 50 kit (bioMerieux, Marcy l'Etoile, France), although in our assays with this kit, BL23 always gave a positive fermentation signal for ribitol. We have no explanation for this discrepancy. Homologues of the proteins encoded by the genes from LCABL_29260 to LCABL_29200 are present in the same order in several Lactobacillus salivarius strains. Interestingly, in L. salivarius strains UCC1118 and ECT 5713, the ribitol genes are present on plasmids pMP118 and pHN3, respectively, but in each organism, the last three genes corresponding to LCABL_29250-LCABL_29160 are lacking. This observation further supports the assumption that these three genes are not essential for the metabolism of d-ribitol but possibly constitute a second pathway for ribitol catabolism.

Two genes exhibiting significant sequence similarity to either LCABL_29270 (xpk), which encodes the enzyme d-xylulose-5-P phosphoketolase, or LCABL_29240 (rtpD), which codes for d-ribitol-5-P 2-dehydrogenase, are also found in L. casei strains M36, Lpc-37, A2-362, and T71499. The corresponding operon in these four organisms also contains the gene for an l-ribulose-5-P 4-epimerase (EC 5.1.3.4), which resembles AraD of E. coli and B. subtilis (more than 55% amino acid sequence identity). AraD converts l-ribulose-5-P into d-xylulose-5-P during l-arabinose utilization (52, 53). The presence of a gene in the xpk-araD region of the four L. casei strains that encodes an enzyme resembling d-ribitol-5-P 2-dehydrogenase of strain BL23 suggests that this enzyme might form l-ribulose-5-P from either l-ribitol or l-arabinitol transported and phosphorylated by a PTS. The three genes located at the beginning of this region encode indeed the EIIA, -B, and -C components of a galactitol-type PTS (Fig. 1), and the gene located at the end of the inserted region encodes a second galactitol-type EIIB component. It is therefore tempting to assume that this PTS transports and phosphorylates either l-ribitol or l-arabinitol, which would subsequently be oxidized to d-ribulose-5-P. Indeed, the dehydrogenase of the xpk-araD region shows the most significant similarity to the E. avium arabinitol-phosphate dehydrogenase (47). Pentitols with the l-configuration are very rare in nature (54), but at least l-arabinitol has recently been shown to be utilized by plant-symbiotic bacteria (13). In the last catabolic step, d-xylulose-5-P formed from l-ribulose-5-P by the enzyme l-ribulose-5-P 4-epimerase is probably cleaved by the d-xylulose-5-P phosphoketolase into d-gylceraldehyde-3-P and acetyl-P. The catabolic pathway would therefore largely resemble the one operative in strain BL23 for the utilization of d-ribitol (Fig. 3). A region of identical gene composition is also present in the two Lactobacillus rhamnosus strains LRHMPDP2 and LRHMPDP3. In addition to being part of the d- or l-ribitol or l-arabinitol region, d-xylulose-5-P phosphoketolase is sometimes also associated with an arabinose region. This is the case in several L. plantarum strains, in which the xpk gene is located next to an l-arabinose region composed of four genes encoding an l-arabinose transporter, an l-arabinose isomerase, a ribulokinase, and an l-ribulose-5-P 4-epimerase.

Interestingly, identical to the ribitol region of strain BL23, the presumed l-ribitol or l-arabinitol operon in the four L. casei strains M36, Lpc-37, A2-362, and T71499 is also inserted between homologues of LCABL_29150 (encoding an l-ascorbate-6-phosphate lactonase belonging to a presumed ascorbate-specific PTS operon) and LCABL_29280 (encoding a sugar kinase-like transcription regulator). As outlined in Fig. 1, in the various strains of L. casei, at least three different PTS-related regions have been inserted at this specific locus. BL23 and related strains contain a mannose-type PTS specific for d-ribitol, M36 and its related strains a galactitol-type PTS presumed to transport l-ribitol or l-arabinitol, and ATCC 334 and like strains a mannitol/fructose-type PTS. The insertion points in the three types of strains are nearly identical and differ only by 3 bp at the beginning and by 7 bp at the end of the inserted fragment. It is interesting to note that the region comprising about 35 kb upstream and 80 kb downstream from the d-ribitol region contains a very large number of carbohydrate transport and utilization operons. However, the gene composition of these operons suggests that the specificity of these carbohydrate transporters exhibits a great variability. This entire region therefore seems to be a hot spot for the integration of carbohydrate transport and metabolism operons, which might have been one factor that allowed L. casei strains in the course of evolution to efficiently adapt to the carbon sources prevailing in the changing environments to which they are exposed (25, 42–44, 48–50).