INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Citrate is present in milk at concentrations of 8 to 9 mM and is cometabolized with sugars by many strains of lactic acid bacteria (LAB), including L. lactis subsp. lactis biovar diacetylactis (3). The breakdown of citrate results in production of carbon dioxide (responsible for the texture of some cheeses) and production of the flavor compound diacetyl, which is essential for the quality of dairy products such as butter, buttermilk, and cottage cheese. Citrate utilization by LAB requires not only the enzymes responsible for citrate metabolism (reviewed in reference 9), but also citrate permease P (CitP), which catalyzes the uptake of this compound (4). Thus, citrate transport limits the rate of citrate utilization and may affect the yield of aroma compounds from citrate. We have previously shown that the external pH drastically influences the citrate uptake mediated by CitP in L. lactis subsp. lactis biovar diacetylactis. The highest uptake rates were observed at pH 4.5 to 5.5, whereas at pH values above 6.5 only basal levels of citrate transport were detected (19). These observations are consistent with a previous report of Van der Rest et al. (29), who demonstrated that activity of the plasmidic CitH from Klebsiella pneumoniae requires an acidic external pH. In addition, our results support the observation that the specific rates of citrate utilization by growing cells increase about sixfold when the pH decreases from 6.5 to 4.5 (2, 10). Therefore, the dependence of citrate metabolism on pH in L. lactis subsp. lactis biovar diacetylactis seems to be related to the narrow pH optimum for CitP activity. This is presumably because the divalent anionic species of citrate is the preferred substrate for the permease (19). These observations suggest that the form of citrate recognized by the transporter is determined by the pH of the medium. L. lactis subsp. lactis biovar diacetylactis grows at pH 7.0, but CitP does not function at this pH. In milk fermentations, L. lactis subsp. lactis biovar diacetylactis converts lactose to lactate, which results in acidification of the medium to pH values as low as 4.0. This acidification is one of the main factors that lead to the arrest of cell multiplication and possibly cell death (24). It has recently been reported that L. lactis exhibits inducible acid tolerance at low pHs in the exponential phase of growth, which requires de novo protein synthesis. These results indicate that acidic conditions induce expression of genes required for acid adaptation (24). However, little information concerning the mechanism(s) and gene(s) involved in acid resistance in L. lactis is available. We have recently described the characterization and the results of a detailed transcriptional analysis of the citQRP cluster involved in the transport of citrate in L. lactis subsp. lactis biovar diacetylactis (16, 17). In this paper, we report that in lactococci both transcription of citP and citrate uptake increase when cells are grown at low pHs. This increase in citrate transport leads to more efficient glucose utilization, which results in a growth advantage for L. lactis subsp. lactis biovar diacetylactis at acid pHs.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, growth media, plasmids, and plasmid transfer. The bacterial strains and plasmids used in this work are listed in Table 1. Lactococcal strains were grown at 30°C without shaking in M17 medium (7) adjusted to various pHs with HCl. M17 medium was supplemented with either 1% glucose (M17G medium) or 0.4% citrate (M17C medium) or with both carbon sources (M17GC). Transfer of plasmids to L. lactis was performed by electroporation by using the procedure of Dornan and Collins (5). Streptococcus pneumoniae 708 (end-1 exo-2 trt-1 hex-4 malM594) (11) was grown and transformed as previously described (20). Transformants were selected in agar medium containing 5 µg of chloramphenicol per ml.

Adaptation of the bacterial cultures to acid pHs. L. lactis strains were grown overnight in M17 medium adjusted to various pHs and supplemented as indicated below. Stock cultures, previously grown at pH 7.0 and kept frozen at 70°C, were used as inocula. The following morning cells were sedimented by centrifugation and concentrated 10-fold by resuspension in saline solution. Then, appropriate aliquots of the cultures were used to inoculate fresh medium to give an A₆₆₀ of approximately 0.04. Dilution was performed with the medium utilized for the overnight cultures. Finally, the cultures were grown until they reached the absorbance values indicated below and then were utilized for analysis of RNA, citrate transport, or chloramphenicol acetyltransferase (CAT) activity. The growth and dilution conditions used allowed the latent period of the cultures to be reduced and the contributions of the mRNA, CitP, and CAT present in the overnight cultures, which were already induced by acidification of the medium, to be minimized.

Construction of plasmid pFL40. Plasmid pFL12 was digested with Bst11071 and BglII, and the 8.5-kb fragment containing the vector region of the plasmid and the citP-cat fusion was purified. Then the 5' overhangs of the fragment were filled with the Klenow fragment of Escherichia coli DNA polymerase I and blunt end ligated to the 1.0-kb ScaI fragment of pFL12, including the P1 cit promoter. The resulting plasmid, pFL40, was established in S. pneumoniae after selection for chloramphenicol resistance and then transferred to L. lactis by electroporation. This plasmid contains the citP-cat fusion under the control of P1 and lacks the citQ and citR genes.

RNA isolation and primer extension. L. lactis strains were grown in M17G medium to an A₆₆₀ of 0.2, and RNA was isolated as previously described (16). The RNAs were checked for the integrity and yield of the rRNAs in all samples. The patterns of the rRNAs were similar in the various preparations. The total RNA concentrations were determined by UV spectrophotometry. Primer extension analysis was performed as previously described (16). The primers used to detect the start site of mRNA1 or mRNA2 were 5'-GAAATTAGAGATGATAC-3' and 5'-AGGGTTTTGTTTTTGGTT-3', which were complementary to mRNA1 from nucleotides 234 to 218 or to both mRNAs from nucleotides 1251 to 1233 (for coordinates see reference 16). One picomole of either primer was annealed to 15 µg of RNA. Primer extension reactions were performed by incubating the annealing mixture with 20 U of avian myeloblastosis virus reverse transcriptase (Promega) at 42°C for 30 min. The sizes of the reaction products were determined by using an 8% polyacrylamide gel containing 7 M urea. Bands labelled with ³²P were detected by autoradiography on Kodak X-Omat S film and were directly quantitated with a PhosphorImager system (Molecular Dynamics).

Determination of CAT activity. L. lactis MG1363 harboring various plasmids was grown as described below. Cultures were sedimented by centrifugation, washed by suspension in buffer A (50 mM Tris-HCl [pH 7.8], 1 mM disodium EDTA, 0.1 mM phenylmethylsulfonyl fluoride, 1 µM dithiothreitol), centrifuged again, and concentrated 10-fold by suspension in buffer A. Total extracts were prepared by passing cells through a French pressure cell at 12,000 lb/in² and removing the cell debris by centrifugation at 20,000 × g for 15 min. CAT activity was determined as previously described by Shaw (27). One unit of CAT activity was defined as the amount of enzyme that catalyzed the acetylation of 1 nmol of chloramphenicol per min at 37°C.

Determination of plasmid copy number. L. lactis MG1363 harboring various plasmids was grown as indicated below. Total DNA extracts containing chromosomal and plasmid DNA were prepared from 1.5-ml cultures essentially as previously described for Bacillus subtilis (12), with the following modifications. Cells were resuspended in 0.1 ml of a solution containing 25% sucrose, 0.1 M NaCl, 50 mM Tris-HCl (pH 8.0), 25 mM disodium EDTA, 2 µg of pancreatic RNase, and 1 mg of lysozyme, and they were incubated at 37°C for 10 min prior to lysis with 10 µl of 10% sodium dodecyl sulfate. Total DNA was analyzed by electrophoresis in a 0.8% agarose gel. DNA bands were revealed by staining with ethidium bromide at a concentration of 0.5 µg/ml. Quantitation of the bands was performed by scanning the gels with a Molecular Analyst (Bio-Rad), with precautions to ensure the linearity of the determinations. DNA samples were run at least three times, or several dilutions of the DNA samples were electrophoresed. The plasmid copy number (N) was calculated as described by Projan et al. (23) by using the following equation: N = (Dp1 + 1.36 Dp2) × Mc/Dc × Mp, where Dp1 and Dp2 are the values obtained from densitometric quantification of open circular and covalently closed circular forms of the plasmid, respectively, Dc is the value obtained from densitometric quantification of chromosomal DNA, Mc is the genome size of L. lactis MG1363 (2.56 × 10⁶ bp according to Le Bourgeois et al. [13]), and Mp is the plasmid size expressed in base pairs.

Analytical methods. L-Lactic acid and D-lactic acid contents were determined enzymatically with L-lactate and D-lactate dehydrogenases (Boehringer Mannheim type 1112 821 test kit) by using frozen supernatants of cultures. Citric acid was assayed enzymatically with citrate lyase, L-malate dehydrogenase, and L-lactate dehydrogenase (Boehringer Mannheim type 139076 test kit). Glucose was assayed enzymatically with glucose oxidase and peroxidase (Sigma type 510A test kit).

Citrate transport assay. Bacterial cultures were grown in M17G to an A₆₆₀ of 0.2, sedimented by centrifugation, and washed in buffer B (25 mM sodium phosphate buffer, pH 5.5). Cultures were resuspended (to give 1 × 10⁹ to 2 × 10⁹ cells/ml) in 650 µl of buffer B. Transport was assayed over a 5-min period with 12 µM [1,5-¹⁴C]citrate (83.8 mCi/mmol) at 30°C. Incorporation of [¹⁴C]citrate was measured as previously described (26).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

pH of the medium influences expression of the citP gene. To determine whether expression of citP was affected by the acidity of the growth medium, we measured the activity of CAT encoded by plasmid pFL12 (Table 1 and Fig. 1). This plasmid contains citQ, citR, and a citP-cat translational fusion at the first codon of citP, and these three genes are under the control of the strong P1 and weak P2 cit promoters (17). Batch cultures of L. lactis MG1363(pFL12) were grown at the initial pHs indicated in Fig. 2. Under these conditions, the acidification of the medium due to bacterial growth was minimized, and a decrease in pH of less than 0.5 U was observed (data not shown). Moreover, all of the cultures were in the exponential phase of growth. Lack of growth of MG1363(pFL12) below pH 4.5 prevented analysis of expression of citP-cat under more acidic conditions. CAT activity slightly decreased from 582 ± 47 to 343 ± 22 U when the pH was decreased from 7.0 to 6.0 (Fig. 2A), but increased at lower pHs, reaching a maximum value (4,016 ± 49 U) at pH 4.5 (Fig. 2A). The copy numbers of pFL12 in L. lactis MG1363(pFL12) cultures grown at different pHs (Fig. 2B) differed by less than twofold. Thus, the 12-fold increase in CAT activity when the pH was changed from 6.0 to 4.5 could not be due to an increase in gene dosage of the citP-cat fusion in the cells. It has been proposed that the levels of the chaperones DnaK and GroEL increase in L. lactis during acid stress (25). Therefore, the high CAT levels detected in MG1363(pFL12) grown at pH 4.5 could be ascribable to an increase in the enzymatic activity as a consequence of greater chaperonin-assisted folding of the CAT protein at low pHs rather than to enhanced synthesis of this enzyme. In fact, the pH of the medium also affected the levels of CAT activity encoded by plasmid pFL20 (Fig. 2A), in which the cat gene is under control of transcriptional and translational signals unrelated to the citP gene (Table 1 and Fig. 1). However, the acid stress resulted in only a twofold increase in CAT levels, and this effect was also observed in cells carrying plasmids pFL16 and pFL40 (see below). Plasmids pFL16, pFL20, and pFL40, as well as pFL12, are based on the pLS1 replicon. The plasmid copy number in cells grown at various pHs varied within a twofold range (from 14 to 28 copies per chromosome equivalent) for all four plasmids and under all conditions tested (Fig. 2B). Consequently, the external pH seems to have only a slight effect on the copy number of pLS1-based plasmids and/or on the specific activity of CAT.

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Physical map of the inserts present in recombinant plasmids pFL12, pFL16, pFL20, and pFL40. Genes are transcribed in the directions indicated by the arrows. P1 and P2, cit promoters; Ptet, promoter of pLS1 tetL gene; PpolA, promoter of pneumococcal polA gene; SDcitP, Shine-Dalgarno sequence of citP; ATGcitP, translational start codon of citP; Tcat, transcriptional terminator of pC194cat gene; solid boxes, DNA segments from pCIT264 lactococcal plasmid; open boxes, pC194 cat gene; box with left diagonal cross-hatching, pneumococcal insert; box with right diagonal cross-hatching, DNA segment including Shine-Dalgarno (SD10) and translational start codon (ATG10) of the T7 gene 10; loop, putative secondary structure at which processing of cit mRNA takes place. Only relevant restriction sites are shown.

View larger version (23K): [in this window] [in a new window]   FIG. 2.   Influence of a shift in the external pH on expression of the citP-cat fusion (A) and plasmid copy number (B). L. lactis MG1363 harboring plasmid pFL12 (), pFL16 (), pFL40 (), or pFL20 () was grown to an A660 of 0.2 in M17G medium adjusted to the initial pHs indicated. Crude extracts were prepared from these cultures, and the CAT activities, as well as the plasmid copy numbers, were determined as described in Materials and Methods. Each value is the average of the values from at least three independent experiments.

Transcription of the citQRP operon is induced at low pH. The response to acid observed in cells carrying pFL12 apparently results from increased expression of citP and could be the result of greater transcription and/or translation of the citP-cat fusion. To test these possibilities, the influence of external pH on CAT levels in cells carrying pFL16 was analyzed (Fig. 2A). Plasmid pFL16 contains the same citQ, citR, and citP-cat genes as pFL12, but they are under control of the pLS1 tetL and the cit P2 promoters (Table 1 and Fig. 1). Replacement of the P1 cit promoter by the tetL promoter in pFL16 reduced the effect of acidity to the levels observed with cells carrying the control plasmid, pFL20 (Fig. 2A). These results indicated that neither the strong tetL promoter (15) nor the weak P2 promoter (17) yield greater transcription in an acid environment. The P2 promoter is not involved in the acid response, and induction of citP could be due to greater transcription of the cit operon from the P1 promoter. To test this possibility, the influence of growth pH on the levels of citP-cat mRNA encoded by pFL12 from P1 were analyzed by performing a primer extension study (Fig. 3A). We have previously shown that the cit transcript synthesized from P1 starts at two A residues (16), and, as expected, the corresponding two extended products (bands designated 5'-end in Fig. 3) were detected in the four primer extension reaction mixtures. The levels of these RNA species, which contained the expected 5' ends, were approximately 14-fold higher at pH 4.5 than at pH 6.5 (Fig. 3B). Approximately 50% of the molecules have their 5' ends at this location. In addition, other bands (Fig. 3A, bands a through e) were detected. These extra bands were also more abundant (18.7-fold ± 5-fold more abundant) in RNAs from cultures grown at pH 4.5 (Fig. 3A, lanes 2 and 4) than in RNAs from cultures grown at pH 6.5 (lanes 1 and 3). Band d, representing 20% ± 2% (at pH 4.5) and 10% ± 3% (at pH 6.5) of the total radioactivity, could correspond to a transcript with a different 5' end. However, the DNA sequence of the region preceding the putative 5' end does not conform to known Lactococcus promoter sequences. Each of the other extra bands accounted for less than 12% of the total radioactivity at each pH, indicating that these bands may represent products of dissociation of reverse transcriptase during DNA synthesis. The increase in the levels of the transcripts correlated with the CAT activity in the same cultures (data not shown), supporting the hypothesis that transcription from the P1 promoter is responsible for the acid response. We have previously shown that transcription from the P2 promoter contributes only 20% of the expression of the cit operon in plasmid pFL12 (17) when cells are grown in a medium buffered at neutral pH. There was no increase in the primer extension product for the P2 promoter in cultures of MG1363(pFL12) at pH 4.5 compared to cultures at pH 6.5 (data not shown). Thus, it seems that P2 does not contribute to the stress response observed in MG1363(pFL12), as expected from the low level of acid induction detected in MG1363(pFL16) (Fig. 2A).

View larger version (48K): [in this window] [in a new window]   FIG. 3.   Induction of transcription from the P1 promoter at acid pH. Four cultures of L. lactis MG1363(pFL12) were grown to an A660 of 0.2 in M17G medium adjusted to an initial pH of 6.5 (lanes 1 and 3) or 4.5 (lanes 2 and 4). Total RNAs from these cultures were prepared, and the 5' end of the cit transcript was determined by primer extension. Then, appropriate volumes of each sample were applied to the gel to give (per well) 2.5 µg of RNA (lanes 1 and 2), 5 µg (lane 3), or 1 µg (lane 4). The extension products are indicated by arrows. DNA sequence ladders (lanes A, C, G, and T), which were used as size standards, were generated with the oligonucleotide used for primer extension. (A) Gel. (B) Scans of the lanes of the gel obtained with the PhosphorImager system.

Acidification of the medium by growth induces CitP synthesis. The results described above showed that growth at different pHs affects expression of citP. However, under normal growth conditions, lactococci sense gradual acidification by extracellular accumulation of lactic acid produced by the cells. Thus, we investigated expression of the citP-cat fusion and changes in the extracellular pH during growth of MG1363(pFL12) in medium having an initial pH of 6.5 or 4.5 (Fig. 4). To prolong the exponential phase of growth, twofold-concentrated M17 medium supplemented with 1% glucose was used in these experiments. Acidic conditions greatly reduced growth. At pH 4.5, cells grew with a doubling time of 145 min and they reached the stationary phase at an absorbance of 0.4, whereas at pH 6.5 the doubling time was 50 min and the culture reached an absorbance of 2.0 during the exponential phase of growth. As expected, acidic growth conditions resulted in higher levels of CAT activity encoded by pFL12. Moreover, an increase in the expression of the citP-cat fusion was detected during growth of both cultures. However, different patterns of induction were observed in the two cultures. At pH 4.5, an increase in CAT activity was observed during the exponential phase, accompanied by a slight decrease in pH. This activity levelled off when the pH stabilized and decreased during the late stationary phase. The latter behavior was presumably due to a lack of synthesis of CAT and to proteolytic degradation of the preexisting enzyme. At pH 6.5, the level of CAT did not increase until the pH of the medium dropped below 5.0. Then, induction started to take place during the late exponential growth phase, and the maximum levels of CAT were observed at the beginning of the stationary phase. Later, a decrease in CAT activity was detected until the activity reached a plateau. These experiments indicated that transcriptional induction is related to acidification of the medium rather than to the stage of growth or to the depletion of nutrients.

View larger version (23K): [in this window] [in a new window]   FIG. 4.   Induction of expression of citP-cat fusion by acidification of the growth medium. L. lactis MG1363(pFL12) was grown in M17G medium adjusted to an initial pH of either 6.5 or 4.5. At the times indicated, the absorbances of the cultures (), the pHs of the external medium (), and the CAT activities () encoded by pFL12 were measured.

The citrate transport system of L. lactis subsp. lactis biovar diacetylactis CRL264(pCIT264) is regulated by acid stress. Utilization of L. lactis subsp. lactis biovar cremoris MG1363 and plasmid pFL12 allowed us to determine the influence of pH on expression of citP in the absence of synthesis of CitP. However, it did not prove that induction of active CitP, encoded by parental plasmid pCIT264, takes place in the natural host, L. lactis subsp. lactis biovar diacetylactis CRL264. Thus, synthesis of citQRP mRNA and citrate transport activity were investigated by using exponential cultures of CRL264(pCIT264) grown in M17G medium at pH 7.0 or 4.5. To detect levels of the citQRP transcript, total RNA was prepared from the lactococcal cultures and analyzed by Northern blot hybridization as previously described (16). This analysis revealed that a shift from pH 7.0 to 4.5 resulted in eightfold-higher levels of the full-length transcript (data not shown). This transcriptional induction of the citQRP operon resulted in higher citrate transport activity catalyzed by the citrate permease of resting lactococcal cells (Fig. 5). The initial rates of citrate uptake and the levels accumulated were significantly and reproducibly higher (at least two- to fourfold higher) in cells previously grown at pH 4.5. These results suggest that induction of transcription at acid pHs results in an increase in CitP levels. This conclusion is supported by the results obtained with the citP-cat fusion present in pFL12 (Fig. 2A and 4). However, we cannot rule out the possibility that CitP activity is stimulated by some alteration produced in the whole cell or in the cell membrane as a consequence of the low-pH stress. Therefore, to rigorously establish that increased synthesis of CitP occurs at acid pH, it would be necessary to quantify the CitP levels. In any case, our results show that growth at low pHs results in enhanced citrate transport activity catalyzed by CitP.

View larger version (17K): [in this window] [in a new window]   FIG. 5.   Detection of levels of CitP expression at neutral and acid pHs. L. lactis subsp. lactis biovar diacetylactis CRL264(pCIT264) was grown to an A660 of 0.2 in M17G medium adjusted to an initial pH of 7.0 () or 4.5 (), and the citrate transport catalyzed by CitP was measured.

Citrate and glucose are cometabolized by L. lactis subsp. lactis biovar diacetylactis at acid pH. The results described above suggested that the increase in transcription of citP accompanied by the enhancement of citrate uptake could improve adaptation of L. lactis subsp. lactis biovar diacetylactis to acidic conditions. We studied the effect of plasmid pCIT264 on the growth of L. lactis subsp. lactis biovar diacetylactis. Based on an analysis of the DNA nucleotide sequence of pCIT264, it appears that the citrate transport genes are present in the plasmid but other genes involved in citrate metabolism are not. Therefore, we tested the growth of CRL264(pCIT264) Cit⁺ and isogenic strain CRL30 Cit (cured of pCIT264) in M17 medium supplemented or not supplemented with carbon sources (Table 2). At pH 4.5 the doubling times of both strains were longer than the doubling times at pH 7.0 in all media tested. However, at pH 4.5 CRL264 grew faster and produced greater biomass than CRL30 in medium supplemented with citrate. This effect was more evident when M17GC medium was utilized. Under these conditions CRL264 and CRL30 had doubling times of 75 and 130 min, respectively. Furthermore, this medium supported the growth of CRL264 to an absorbance of 2.8, whereas CRL30 reached a final absorbance of only 0.45. The presence of plasmid pCIT264 did not affect growth at pH 7.0 with any of the media tested or growth at pH 4.5 when M17 or M17G medium was used. These results indicated that cometabolism of citrate and glucose provided the energy for efficient growth of L. lactis subsp. lactis biovar diacetylactis CRL264 at low pHs. To further test this hypothesis, we measured citrate and glucose consumption, as well as lactate accumulation, in the medium during growth of the two lactococcal strains (Cit⁺ and Cit) at low pHs (Fig. 6). Under these conditions, growth of CRL264 resulted in alkalinization of the external medium. This effect was presumably due to metabolism of citrate, since it was not detected with CRL30, and alkalinization was also observed after growth of CRL264 in M17C medium (Table 2). In addition, in the case of CRL264 the slope of the pH curve correlated with consumption of citrate for up to 7.5 h of incubation (Fig. 6). When the external pH was greater than 4.75, the rate of citrate utilization dramatically increased, indicating that rapid metabolism of citrate occurred. This behavior (enhanced consumption of citrate) was accompanied by a diauxic-like growth curve and was not correlated with the production of lactate. Moreover, when the pH reached a value higher than 5.0 (after 7.5 h of growth), very efficient consumption of glucose started to take place, and this consumption was not ascribable to the increase in cellular mass. This enhancement of glucose metabolism was correlated with an increase in production of lactate, whose secretion was presumably responsible for the observed reduction in pH when the incubation times were longer.

View larger version (22K): [in this window] [in a new window]   FIG. 6.   Growth characteristics of L. lactis subsp. lactis biovar diacetylactis CRL264 (A) and CRL30 (B). Cultures were grown in M17GC medium adjusted to an initial pH of 4.5. The growth was monitored by measuring absorbance (). Samples of the cultures were taken at the times indicated and centrifuged, and the pHs of the supernatants were determined (). The supernatants were also analyzed for the presence of glucose (), citrate acid (), and lactic acid ().

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The metabolism of lactose by LAB during dairy fermentations results in lactate accumulation and consequent acidification of the medium (22). Thus, LAB are naturally exposed to acid stress generated by growth. However, little is known about the mechanisms involved in this process (for a review see reference 24).

In this work, we investigated the role of acid stress in the expression of the citrate transport system of L. lactis subsp. lactis biovar diacetylactis. The citQRP operon is induced either by a shift to acidic conditions or by acidification of the medium during growth. Moreover, this induction occurs at the transcriptional level and results in enhanced citrate transport. The maximum levels of CitP were obtained from lactococcal cultures grown at pH 4.5. We have previously observed that the maximum transport activity of CitP requires a pH of 4.5 to 5.5 (19). Therefore, the activation of this transport system, which is a response to acid stress in lactococci, enhances citrate utilization by L. lactis subsp. lactis biovar diacetylactis without any deleterious effect. At neutral pH, the low level and activity of CitP do not permit transport of citrate, thus preventing toxicity from the accumulation of unmetabolized citrate. Furthermore, acidification of the medium during dairy fermentation should result in a gradual increase in citrate transport due to induction of CitP synthesis and to higher activity of the preexisting permease. The result is that the most efficient citrate uptake occurs under optimal conditions for citrate metabolism in L. lactis subsp. lactis biovar diacetylactis.

How is the citQRP operon induced? The P1 promoter is indispensable for the acid response, and an increase in transcription from P1 occurs at low pHs. This enhancement of transcription could be due to utilization of a general stress sigma factor by the lactococcal RNA polymerase, as is the case with ^S in Salmonella typhimurium and E. coli (14, 28) or with ^B in Bacillus subtilis (30). However, in L. lactis only the vegetative sigma factor has been identified, although research to detect a stress sigma factor has been performed (1, 6). Therefore, it seems more likely that it is the vegetative sigma factor of the RNA polymerase that recognizes P1 and that additional factors are responsible for the efficiency of P1 utilization. The induction occurs in the presence of glucose in the growth medium. Thus, it is not likely that the system is subjected to regulation mediated by the CcpA and Hpr signal-transducing pathway, which seems to be the main mechanism for catabolic repression in gram-positive bacteria. Furthermore, an inspection of the DNA sequence of the region surrounding and including P1 did not predict the existence of a catabolic repression element (CRE) operator binding site for the CcpA-Hpr complex (8). The detection of transcriptional induction by acid in the absence of citP gene proves that the citP product (CitP) is not involved in the process. However, an unidentified cis-acting element present or encoded by the CIT⁺ plasmid seems to be required for the induction, since the removal of regions located upstream and downstream of P1 resulted in expression of citP at noninduced levels.

What is the role of citrate utilization in L. lactis subsp. lactis biovar diacetylactis? Hugenholtz et al. (10) demonstrated that L. lactis subsp. lactis biovar diacetylactis is able to grow at an acid pH by utilizing citrate as the only energy source, although a small cell mass increase was detected. We observed that at neutral pH the presence of citrate in M17 medium does not stimulate the growth of either strain CRL264 (Cit⁺) or strain CRL30 (Cit), and, as expected, both strains utilize glucose efficiently as an energy source. By contrast, at pH 4.5 glucose by itself supports poor slow growth of both strains, which is consistent with the observation that acidification inhibits growth and metabolism even if nutrients are still available (21). However, at this pH, the presence of citrate and glucose results in a growth advantage only for the Cit⁺ strain. The same behavior was observed when glucose was replaced by lactose (results not shown). Since growth of this strain was not stimulated by citrate alone, utilization of this compound could not be the main source of energy at low pHs. Therefore, cometabolism of sugar and citrate seems to be important for growth stimulation under acidic conditions. It has been established that metabolism of citrate alkalinizes the external medium and that metabolism of glucose acidifies the external medium (21). Thus, the slow alkalinization of the external medium observed during growth of CRL264 in the presence of citrate and either glucose (Fig. 6) or lactose (results not shown) could be due to citrate metabolism. This increase in the external pH seems to result in efficient citrate utilization accompanied by an increase in sugar metabolism. Therefore, alkalinization of the growth medium by citrate transport is crucial for cometabolism of sugar and citrate by L. lactis subsp. lactis biovar diacetylactis. This phenomenon could provide L. lactis subsp. lactis biovar diacetylactis with a selective advantage, which should prolong its survival in fermented products.