Physiology and Substrate Specificity of Two Closely Related Amino Acid Transporters, SerP1 and SerP2, of Lactococcus lactis

Abstract

The serP1 and serP2 genes found adjacently on the chromosome of Lactococcus lactis strains encode two members of the amino acid-polyamine-organocation (APC) superfamily of secondary transporters that share 61% sequence identity. SerP1 transports l-serine, l-threonine, and l-cysteine with high affinity. Affinity constants (K_m) are in the 20 to 40 μM range. SerP2 is a dl-alanine/dl-serine/glycine transporter. The preferred substrate appears to be dl-alanine for which the affinities were found to be 38 and 20 μM for the d and l isomers, respectively. The common substrate l-serine is a high-affinity substrate of SerP1 and a low-affinity substrate of SerP2 with affinity constants of 18 and 356 μM, respectively. Growth experiments demonstrate that SerP1 is the main l-serine transporter responsible for optimal growth in media containing free amino acids as the sole source of amino acids. SerP2 is able to replace SerP1 in this role only in medium lacking the high-affinity substrates l-alanine and glycine. SerP2 plays an adverse role for the cell by being solely responsible for the uptake of toxic d-serine. The main function of SerP2 is in cell wall biosynthesis through the uptake of d-alanine, an essential precursor in peptidoglycan synthesis. SerP2 has overlapping substrate specificity and shares 42% sequence identity with CycA of Escherichia coli, a transporter whose involvement in peptidoglycan synthesis is well established. No evidence was obtained for a role of SerP1 and SerP2 in the excretion of excess amino acids during growth of L. lactis on protein/peptide-rich media.

INTRODUCTION

The lactic acid bacterium (LAB) Lactococcus lactis is extensively used for the manufacture of buttermilk and cheese. L. lactis originally lived on plants, and dairy strains are currently derived by reductive evolution (1, 2). The change to nutrient-rich environments like milk is believed to be the reason that dairy strains of L. lactis lost biosynthetic pathways for various amino acids like branched-chain amino acids and histidine (3, 4). In milk, casein provides a rich source of all amino acids. It is degraded by an efficient proteolytic system, which consists of a proteinase, several peptidases, and a peptide transport system. The proteinase PrtP is attached extracellularly to the cell wall and hydrolyzes casein to peptides varying in length from 5 to more than 30 amino acid residues. The peptides are transported into the cell via the oligopeptide uptake system Opp and further degraded in the cell by intracellular peptidases (5, 6). The system generates free amino acids in the cytoplasm following uptake of peptides without the need for specific transporters to take up free amino acids from the environment. Nevertheless, media containing free amino acids as the sole source of amino acids do support growth of L. lactis (7), indicating that transport systems for single amino acids have been preserved. The physiological roles of many of these transporters during growth on different media are not clear.

Transport studies in L. lactis and other LAB, mainly performed in the late 1980s, using whole cells or membrane vesicles demonstrated proton gradient-driven uptake activity for branched-chain amino acids and l-methionine (8, 9), l-alanine and glycine (10), l-lysine (11), l-serine, l-threonine, l-histidine, and l-cysteine (12), and l-tyrosine and l-phenylalanine (13). l-Glutamate, l-glutamine (14), and possibly l-proline (15) transport was shown to be driven by ATP hydrolysis. More recent work has identified many of the genes encoding the transporters. bcaP and gnlPQ were identified in L. lactis as encoding a secondary branched-chain amino acid transporter (16) and an ATP-driven l-glutamate/l-glutamine ABC transporter (17), respectively. Trip et al. (18) cloned and expressed all 14 members of the APC (amino acid-polyamine-organocation) superfamily (19) identified in the genome of L. lactis and screened the recombinant cells for enhanced uptake of 14 proteinous amino acids and putrescine. LysP, HisP (formerly LysQ), AcaP (formerly YlcA), and FywP (formerly YsjA) were identified as the major transporters for uptake of l-lysine, l-histidine, l-aspartate, and l-phenylalanine/l-tyrosine, respectively. The substrate specificities of six more secondary amino acid transporters were determined. Two of these, SerP1 and SerP2, were both identified as l-serine transporters. The genes llmg_0376 (serP1) and llmg_0375 (serP2) of L. lactis MG1363, encoding these two transporters, are located adjacently on the chromosome, separated by a 71-bp noncoding fragment, which does not contain a potential terminator or promoter sequence (18). The two genes encode proteins containing 459 (SerP1) and 456 (SerP2) amino acid residues that share 61% sequence identity. Inhibition studies following expression of SerP1 and SerP2 in the wild-type background showed different substrate specificities for the two transporters. l-Serine uptake by SerP1 was inhibited by l-threonine and l-cysteine, while l-serine uptake by SerP2 was inhibited by l-alanine and glycine (18).

Here, the physiological functions of SerP1 and SerP2 in L. lactis were determined by growth and transport studies using ΔserP1 and ΔserP2 deletion strains and recombinant expression strains in the double mutant background. We demonstrated that SerP1 is the main l-serine transporter responsible for optimal growth of L. lactis in media containing free amino acids as the sole source of amino acids. l-Threonine and l-cysteine are additional substrates of SerP1. SerP2 transports dl-alanine, dl-serine, and glycine, which overlaps the substrate specificity of CycA transporters that are involved in cell wall synthesis through the uptake of the precursor d-alanine.

MATERIALS AND METHODS

Strains, media, and growth conditions.

Lactococcus lactis JP9000, derived from strain MG1363 and carrying the nisRK genes in the pseudo_10 locus (20), was used as the parent for construction of the mutants containing the deletions ΔserP1, ΔserP2, and ΔserP1P2. The ΔserP1P2 double mutant was used as the host for the nisin-inducible expression of the SerP1 and SerP2 transporters in the P1ΔserP1P2 and P2ΔserP1P2 strains, respectively. L. lactis was grown at 30°C in M17 supplemented with 28 mM glucose (referred to as GM17) and in the presence of 3 μg/ml erythromycin or 5 μg/ml chloramphenicol, when appropriate, or in chemically defined SA medium containing 19 free amino acids and supplemented with 56 mM glucose and nisin, when appropriate (7). Escherichia coli DH5α, used for routine cloning to create the deletion constructs, was grown in LB medium containing 150 μg/ml erythromycin. Bacterial strains were summarized in Table S2 in the supplemental material.

Growth curves of L. lactis strains were recorded using a Biotek Powerwave 340 96-well plate reader. Overnight cultures in GM17 medium were diluted to an optical density at 600 nm (OD₆₀₀) of 0.05 in 200 μl of the indicated medium and covered with 50 μl of silicon oil (1:4, silicon oil M20 and M200) to prevent evaporation. The optical density at 600 nm was measured for 20 h every 10 min with 30 s of shaking before each measurement.

Construction of deletion and expression mutants.

Oligonucleotides used in this study are listed in Table S1 in the supplemental material, and plasmids and constructs are in Table S2. The L. lactis JP9000 genes llmg_0376 (serP1) and llmg_0375 (serP2) were deleted with a two-step integration-and-excision system using plasmid pCS1966 (21), creating the ΔserP1 and ΔserP2 single-knockout strains and the ΔserP1P2 double-knockout strain. The latter was constructed by a single recombination event. Two 800-bp fragments, corresponding in sequence to the up- and downstream regions of the genes or gene pair with 100 to 140 bp overlapping with the genes, were amplified by PCR using specific primer pairs and subsequently digested with specific restriction endonucleases (see Table S1). The amplicons were ligated into appropriately digested pCS1966, creating pCSdelSERP1, pCSdelSERP2, and pCSdelSERP1P2. The resulting plasmids were transformed into L. lactis JP9000, generating the markerless ΔserP1, ΔserP2, and ΔserP1P2 deletion mutants (18, 21).

The P1ΔserP1P2 and P2ΔserP1P2 expression strains were constructed by transformation of plasmids pNZ376 and pNZ375 (18), encoding SerP1 and SerP2 under the control of the nisin-inducible promoter P_nisA, respectively, into L. lactis ΔserP1P2 using a standard electroporation protocol.

Amino acid transport assays.

Cells were grown in GM17 medium containing nisin when appropriate (see below) to the mid-exponential phase (OD₆₀₀, 0.6 to 0.8). Cells were harvested, washed, resuspended to an OD₆₀₀ of 2 in ice-cold 100 mM potassium phosphate (pH 6.0) buffer containing 0.2% glucose, and kept on ice until use. An aliquot of 100 μl of cells was preincubated for 5 min at 30°C with continuous stirring, followed by the addition of ¹⁴C-labeled l-serine, l-threonine, or l-alanine, to final concentrations of 6 μM, 1.7 μM, and 3 μM, respectively. Uptake was stopped by addition of 2 ml ice-cold 0.1 M LiCl, and the suspension was filtered over a 0.45-μm-pore-size nitrocellulose filter (BA85; Schleicher & Schuell GmbH). The filter was washed once with 2 ml of ice-cold 0.1 M LiCl and submerged in Emulsifier Scintillator Plus scintillation fluid (Packard Bioscience). Radioactivity was measured by scintillation counting with a Tri-Carb 2000CA liquid scintillation analyzer (Packard Instruments).

Strains P1ΔserP1P2 and P2ΔserP1P2 were used for the determination of kinetic constants. The cells were grown in GM17 medium to an OD₆₀₀ of 0.5, after which expression was induced by adding 0.025 and 0.1 ng/ml of nisin to the P1ΔserP1P2 and P2ΔserP1P2 cultures, respectively, followed by an additional 1 h of incubation. The nisin concentration was fine-tuned to obtain a maximal uptake after 10 s of approximately 10% of total radiolabel in the experiments. Initial rates were inferred from the 6- and 10-s time points. Uptake measurements were performed using ¹⁴C-labeled l-serine and l-alanine at a concentration of 3 μM as described above.

Measurement of amino acid and dipeptide concentrations using reverse-phase high-performance liquid chromatography (RP-HPLC).

Samples were run on a Shimadzu high-speed Nexera HPLC and later analyzed by using LC Solutions 1.24 SP1 software from Shimadzu (Kyoto, Japan). Samples taken at different time points from cell suspensions in 100 mM potassium phosphate (pH 6.0) buffer containing 0.2% glucose and 5 mM l-serine–l-valine or l-alanine–l-leucine were centrifuged. Supernatants were subjected to derivatization by diethylethoxymethylenemalonate (DEEMM), as described previously (22). Aminoenone derivates were detected using a Shim-pack XR-ODS column (Shimadzu) with dimensions of 75 by 3.0 mm, operated at 40°C through a linear gradient using eluent A (0.1% formic acid in water) and eluent B (acetonitrile) at a flow rate of 1 ml/min. For the measurements of l-Ser–l-Val and corresponding free amino acids, the gradient started at 14% eluent B, reaching 36% eluent B in 12 min. For the measurements of l-Ala–l-Leu and corresponding free amino acids, the gradient started at 21% eluent B, reaching 43% eluent B in 12 min.

RESULTS

Growth of SerP1 and SerP2 deletion mutants on free amino acids.

The physiological role of SerP1 and SerP2 in L. lactis strain JP9000 was studied by deleting the serP1 and serP2 genes, yielding the ΔserP1 and ΔserP2 mutant strains, respectively, using a two-step integration-and-excision system (21). In addition, the ΔserP1P2 double mutant, in which both genes were deleted, was produced.

No difference in growth between the wild type and the deletion mutants was observed in rich GM17 medium (not shown), suggesting that the two transporters do not play a role when L. lactis grows on a medium containing peptides. In contrast, when grown on SA medium that contains only free amino acids as a source of amino acids, the ΔserP1 mutant showed a significant growth defect. The ΔserP2 mutant was only marginally affected. In keeping with this, the ΔserP1P2 double mutant showed a slightly stronger growth defect than that observed for the ΔserP1 mutant (Fig. 1A; also, see Table S3 in the supplemental material). The growth defect involved both a lower growth rate and a lower biomass yield. It follows that predominantly SerP1 is responsible for the uptake of one or more amino acids that are essential for optimal growth on a medium containing free amino acids.

FIG 1.

(A) Growth of the wild type (◼) and the ΔserP1 (◻), ΔserP2 (●), and ΔserP1P2 (○) mutants in standard SA medium. (B) Growth of the wild type in standard SA medium (●) and SA medium lacking either l-threonine (○), l-cysteine (◼), l-alanine (◻), l-serine (▲), or glycine (△). (C) Growth of the wild type (◼ and ▲) and the ΔserP1 mutant (◻ and △) in SA medium (◼ and ◻) and SA medium from which l-alanine (3.4 mM) and glycine (2.7 mM) were omitted (▲ and △). Values are the means from at least two independent experiments.

Inhibition studies had previously identified l-serine, l-threonine, and l-cysteine as substrates of SerP1 and l-serine, l-alanine, and glycine as substrates of SerP2 (18). Growth of L. lactis JP9000 on SA medium was similar to growth on SA medium lacking l-threonine, l-cysteine, l-alanine, and glycine (Fig. 1B). Apparently, the biosynthetic capacity of these four amino acids is high enough to support optimal growth. In contrast, the absence of l-serine from SA medium severely affected growth (Fig. 1B), indicating that the rate of l-serine biosynthesis in L. lactis JP9000 is a limiting factor and that l-serine transport into the cell is crucial for optimal growth. The growth defect observed in the absence of l-serine in the medium was similar to the one observed for growth of the ΔserP1P2 double mutant in complete SA medium (Fig. 1A).

Growth of the three deletion mutants on SA medium from which the standard 2.9 mM l-serine was omitted showed the same growth defect as observed for the wild-type strain on the same medium (see Fig. S1 in the supplemental material). Adding a concentration of 1 mM l-serine to the medium restored growth of the wild type at a rate close to the rate observed in standard SA medium and resulted eventually in the same biomass yield. The characteristics of the ΔserP2 mutant were similar (see Fig. S1A and B). In contrast, for the ΔserP1 and ΔserP1P2 mutants, the addition of 1 mM l-serine resulted in only a marginal increase of the rate, while the yield was still significantly lower after 10 h of growth (see Fig. S1C and D). The inability to restore the growth of the ΔserP1 mutant suggests the inability of SerP2 to transport l-serine effectively, which would be at variance with an earlier report (18). However, the ability of SerP2 to transport l-serine was evident from growth experiments in SA medium from which both l-alanine and glycine were omitted (Fig. 1C, closed symbols). Growth of the wild type was slightly diminished in the absence of these two amino acids. Remarkably, growth of the ΔserP1 mutant was significantly stimulated in the absence of l-alanine and glycine in the medium, demonstrating that in the absence of competing substrates, SerP2 does support growth by l-serine uptake (Fig. 1C, open symbols). Apparently, l-serine is not the main substrate of SerP2.

Taken together, these data strongly suggest that SerP1 is mainly responsible for l-serine uptake into the cell, with a much smaller contribution of SerP2. The role of l-serine as the growth-limiting factor in media without l-serine or in cells without a l-serine transporter (ΔserP1) was further supported by restored optimal growth in both cases by providing l-serine in the form of the dipeptide l-Ser–l-Val (not shown).

Substrate specificity of SerP1 and SerP2.

The uptake activities of [¹⁴C]l-serine, [¹⁴C]l-threonine, and [¹⁴C]l-alanine by resting cells of L. lactis JP9000 and the ΔserP1, ΔserP2, and ΔserP1P2 deletion mutants grown to the mid-exponential phase in GM17 medium and energized by glucose were measured at concentrations in the micromolar range (Fig. 2). No uptake was observed in the absence of glucose (data not shown). In the presence of glucose, the wild-type cells took up l-serine at an initial rate of 6.0 nmol/min · mg, while uptake by the ΔserP1P2 double mutant was completely abolished. Almost all uptake activity of the wild type could be attributed to SerP1, since in the ΔserP1 mutant, the uptake was reduced by 95%, while uptake by the ΔserP2 mutant was not significantly different from that of the wild type. Clearly, under these conditions, SerP1 is the main l-serine transporter. The same pattern was observed for l-threonine uptake. No uptake was observed with ΔserP1 or ΔserP1P2, whereas ΔserP2 showed an initial rate close to the rate observed for the wild type (1.5 nmol/min · mg), suggesting that l-threonine is transported exclusively by SerP1 (Fig. 2B). The opposite pattern was observed for l-alanine uptake. The wild type took up l-alanine at an initial rate of 1.1 nmol/min · mg, and similar to what was observed for the other two amino acids, the ΔserP1P2 mutant showed no significant l-alanine uptake. However, in contrast to the other two substrates, the uptake activity was retained in the ΔserP1 mutant and diminished in the ΔserP2 mutant, indicating that l-alanine is transported into the cell mainly by SerP2 (Fig. 2C). The activities of SerP1 with l-serine and l-threonine and of SerP2 with l-alanine show that these transporters are functionally expressed in the ΔserP2 and ΔserP1 deletion mutants, respectively, when the strains are grown in rich GM17 medium. This result in addition to the observation that SerP2 can support growth of ΔserP1 in the absence of l-alanine and glycine in the medium by l-serine uptake (Fig. 1C, open symbols) shows that the deletion of serP1 does not affect expression of serP2.

FIG 2.

Uptake of [¹⁴C]l-serine (A), [¹⁴C]l-threonine (B), and [¹⁴C]l-alanine (C) at concentrations of 6, 1.7, and 3 μM, respectively, in JP9000 (■) and the ΔserP1 (◻), ΔserP2 (●), and ΔserP1P2 (○) mutants. Values are the means from at least two independent experiments.

The ΔserP1P2 double mutant was completely devoid of uptake activity of either l-serine, l-threonine or l-alanine, indicating that no transporters other than SerP1 or SerP2 for these substrates are present in the membrane of L. lactis (Fig. 2). This makes the double mutant the perfect host for the recombinant expression of SerP1 and SerP2 to determine the characteristics of both transporters independently of each other. Uptake of [¹⁴C]l-serine, [¹⁴C]l-alanine, and [¹⁴C]l-threonine by resting cells of the ΔserP1P2 strains producing recombinant SerP1 and SerP2 (P1ΔserP1P2 and P2ΔserP1P2, respectively) was measured under the conditions described above using the nisin-controlled expression (NICE) system (Fig. 3) (23). P1ΔserP1P2 took up l-serine at a 2.5-fold-higher initial rate than observed with the corresponding ΔserP2 mutant, indicating higher levels of expression in the recombinant strain. Uptake rates of l-threonine and l-alanine were 20% and 3% of the rate of l-serine by the same cells, respectively. Expression levels of P2 were 17-fold higher in the recombinant strain relative to the ΔserP1 deletion strain based on the comparison of the initial rates of uptake of l-alanine (Fig. 3B). In support of the results of Trip et al. (18), SerP2 transported l-serine, but at an initial rate that was 3-fold lower than that observed for l-alanine. l-Threonine was not transported by the P2ΔserP1P2 recombinant strain (Fig. 3B).

FIG 3.

Uptake of [¹⁴C]l-serine (■), [¹⁴C]l-alanine (●), and [¹⁴C]l-threonine (▲) at concentrations of 6 μM, 3 μM, and 1.7 μM, respectively, in P1ΔserP1P2 (A) and P2ΔserP1P2 (B). P1ΔserP1P2 and P2ΔserP1P2 were induced with nisin concentrations of 0.025 and 0.1 ng/ml, respectively. Values are the means from at least two independent experiments.

Kinetic constants derived from initial rate measurements of uptake of different substrates by the P1ΔserP1P2 and P2ΔserP1P2 strains are summarized in Table 1. SerP1 transports l-serine and l-threonine with high affinity (K_m of 18 and 30 μM, respectively) and similar maximal rates (V_max of 57 and 54 nmol/min · mg, respectively). Inhibition studies of the rate of [¹⁴C]l-serine uptake identified l-cysteine as an additional high-affinity substrate with a K_i of 43 μM. The rate of uptake of [¹⁴C]l-alanine was too low to allow determination of the kinetic parameters. Inhibition studies revealed affinities for l-Ala and Gly that were at least one order of magnitude lower than observed for the three high-affinity substrates. SerP1 was stereoselective, with a strong preference for l-serine. The affinity for d-serine was in the millimolar range. The P2ΔserP1P2 strain expressing SerP2 transported l-serine efficiently as well, but it did so with a lower affinity and higher maximal rate than the P1ΔserP1P2 strain (K_m of 356 versus 18 μM and V_max of 216 versus 57 nmol/min · mg, respectively). The higher maximal rate may be a consequence of a higher level of expression or a higher turnover number of SerP2. The highest affinity of SerP2 observed was that for l-alanine, with a K_m of 20 μM, which is in the range observed for the high-affinity substrates of SerP1. Inhibition studies of [¹⁴C]l-alanine uptake showed that SerP2 also recognizes glycine with relatively high affinity (K_i of 100 μM), while SerP2 had poor affinity for the other two high-affinity substrates of SerP1, l-threonine and l-cysteine, with inhibition constants of >50 mM and 2.3 mM, respectively. In contrast to SerP1, SerP2 is not stereoselective. The affinities for the d- and l- isomers of alanine and serine were found to be in the same range.

TABLE 1.

Kinetic parameters of the amino acid transporters SerP1 and SerP2^a

Substrate	SerP1			SerP2
	Km (μM)	Vmax (nmol/min · mg)	Kib (μM)	Km (μM)	Vmax (nmol/min · mg)	Kic (μM)
l-Ser	18 ± 7	57 ± 12		356 ± 105	216 ± 10
l-Ala			510 ± 120	20 ± 6	115 ± 14
l-Thr	30 ± 8	54 ± 21				>50,000
l-Cys			43 ± 9			2,320 ± 170
Gly			877 ± 375			100 ± 48
d-Ala			3,200 ± 1,200			38 ± 1
d-Ser			1,600 ± 870			127 ± 65

^aValues are means and standard deviations from at least two independent replicates.

^bInhibition of [¹⁴C]l-serine uptake.

^cInhibition of [¹⁴C]l-alanine uptake.

To summarize, with the caveat that inhibition does not necessarily reflect transport, SerP1 seems to transport l-serine, l-threonine, and l-cysteine with high affinity, while SerP2 seems to transport dl-alanine, dl-serine, and glycine.

Physiology of SerP2.

The ability of SerP2 to not only recognize and bind but also transport d-serine and d-alanine was demonstrated by growth experiments, which, in addition, demonstrated possible physiological roles of SerP2. d-Serine is a bacteriostatic agent that slows growth and results in lower cell densities, most likely by inhibiting l-serine and pantothenate biosynthesis pathways (24). d-Serine added at a concentration of 50 mM to SA medium did not significantly alter the growth characteristics of wild-type L. lactis JP9000 (Fig. 4A). However, omitting l-alanine and glycine from SA medium, which by itself does not affect growth (Fig. 1C), resulted in a severe inhibition of growth, showing that in the absence of the high-affinity substrates of SerP2, the cells become susceptible to the bacteriostatic effect of d-serine. The lack of growth inhibition of the ΔserP2 mutant in the same experiment when l-alanine and glycine were omitted from the medium confirms that SerP2 is exclusively responsible for the uptake of d-serine into the wild-type cell (Fig. 4B).

FIG 4.

(A and B) Growth of the wild type (A) and the ΔserP2 mutant (B) in SA medium (■), SA medium plus 50 mM d-serine (●), and SA medium without l-alanine and glycine plus 50 mM d-serine (▲). (C and D) Growth of the wild type (C) and the ΔserP2 mutant (D) in SA medium without l-alanine (■), SA medium without l-alanine plus 0.5 mM d-cycloserine (●), and SA medium without l-alanine plus 0.5 mM d-cycloserine and 2 mM d-alanine (▲). Values are the means from at least two independent experiments.

The substrate specificity of SerP2 (dl-alanine/dl-serine/glycine) overlaps with the specificity of CycA transporters (d-alanine/d-serine/glycine) (24, 25, 26, 27, 28), which are involved in bacterial cell wall biosynthesis through the uptake of the precursor d-alanine, which is essential for growth of bacteria (29). d-Cycloserine, a cyclic, structural analog of d-alanine, is a potent inhibitor of cell wall biosynthesis by competitively inhibiting alanine racemase, which converts l-alanine into d-alanine, and d-alanine ligase, which catalyzes the first step in the incorporation of d-alanine into peptidoglycan (29, 30, 31). Growth of L. lactis JP9000 in SA medium from which l-alanine was omitted to challenge the endogenous synthesis of d-alanine by alanine racemase was effectively inhibited at a concentration of 0.5 mM d-cycloserine (Fig. 4C). The inhibition was largely overcome by including 2 mM d-alanine in the medium. The ΔserP2 mutant was as sensitive to the inhibitor as the wild type, indicating that SerP2 is not essential for transporting d-cycloserine into the cell (Fig. 4D). However, in contrast to what was observed for the wild type, addition of 2 mM d-alanine did not resolve the inhibition of the ΔserP2 mutant by d-cycloserine, indicating that SerP2 was responsible for d-alanine uptake by the wild-type cells (Fig. 4D).

Excretion of internally generated l-serine and l-alanine.

L. lactis JP9000 cells were incubated in 100 mM potassium phosphate (pH 6.0) buffer containing 0.2% glucose with 5 mM concentrations of the dipeptides l-Ser–l-Val and l-Ala–l-Leu. The disappearance of the dipeptides and the simultaneous appearance of the corresponding free amino acids was analyzed by HPLC after the cells had been centrifuged at regular intervals (Fig. 5). The cells consumed l-Ser–l-Val at a rate of 55 μM/min, which was constant during the 20 min of the experiment. At the same rate, l-valine appeared in the supernatant, indicating stoichiometric conversion of the peptide to the amino acid by the cells. l-Serine appeared at a lower rate than expected based on stoichiometric conversion, indicating cytoplasmic conversion into an unknown compound. This compound was excreted, as evidenced by the appearance of an unidentified peak in the HPLC chromatogram. The uptake rate of the l-Ala–l-Leu peptide was 75 μM/min, and both l-alanine and l-leucine appeared stoichiometrically as free amino acids in the supernatant.

FIG 5.

Dipeptide uptake and amino acid excretion by wild-type and mutant strains. Cells were incubated with 5 mM concentrations of the dipeptides l-Ser–l-Val (A) and l-Ala–l-Leu (B). (A) l-Ser–l-Val consumption (▲ and △) and l-Ser (■ and ◻) and l-Val (● and ○) production by JP9000 (■, ▲, and ●) and the ΔserP1 mutant (◻, △, and ○). (B) l-Ala–l-Leu consumption (▲ and △) and l-Ala (■ and ◻) and l-Leu (● and ○) production by JP9000 (■, ▲, and ●) and the ΔserP2 mutant (◻, △, and ○). The data points of JP9000 (■) and ΔserP1 (◻) for l-Ser production overlap.

The experiments were repeated with the ΔserP1 and ΔserP1P2 mutants incubated with the l-Ser–l-Val dipeptide and the ΔserP2 and the ΔserP1P2 mutants with the l-Ala–l-Leu dipeptide. The results for the ΔserP1 and ΔserP2 mutants (Fig. 5) were not significantly different from the results for the wild-type cells, indicating that neither SerP1 or SerP2 was involved in the uptake of the dipeptide or, more importantly, in the excretion of the amino acids.

DISCUSSION

Previously (18), the products of two closely related genes found next to each other on the chromosome of L. lactis were shown to transport l-serine, hence their names, SerP1 and SerP2. The genes were overexpressed in a wild-type background, resulting in enhanced uptake activity of l-serine. Already, in the same study using the same experimental system, inhibition studies suggested that the substrate specificities of the two transporters were not exactly the same. Here, the substrate specificities and kinetic parameters of the two transporters were determined by expressing them separately in a double knockout of the two structural genes, i.e., the P1ΔserP1P2 and P2ΔserP1P2 strains. SerP1 is a stereoselective transporter for l-serine, l-threonine, and l-cysteine with affinities for the substrates in the range of 20 to 40 μM. The affinities for l-alanine and glycine were at least one order of magnitude lower, and those for the isomers d-serine and d-alanine were even lower (Table 1). The highest affinities of SerP2 were observed for d- and l-alanine, also in the 20 to 40 μM range. Slightly lower affinities were observed for d- and l-serine and for glycine (100 to 350 μM). The SerP1 substrates l-threonine and l-cysteine were only poorly recognized by SerP2. It follows that in spite of the high sequence identity of 61%, the two transporters have evolved quite different substrate specificities. In line with the previous study, both SerP1 and SerP2 transport l-serine, but with 20-fold-different affinities (18 and 356 μM, respectively). The name SerP2 is unfortunate, since d- and l-alanine appear to be the preferred substrates for this transporter.

The different but partly overlapping substrate specificities of SerP1 and SerP2 correlate with different physiological roles of the transporters. SerP1 is the main transporter responsible for the uptake of l-serine into the cells during growth on media containing only free amino acids as a source of amino acids. The biosynthetic capacity of l-serine of L. lactis is growth rate limiting, making l-serine transport by SerP1 a crucial step for optimal growth (Fig. 1A and B). In principle, SerP2 could also do the job, but l-serine transport by SerP2 is largely inhibited by the presence of the high-affinity substrates l-alanine and glycine. Only when these two amino acids are not present in the medium can SerP2 replace SerP1 in supporting optimal growth (Fig. 1C). The main physiological function of SerP2 though is most likely in cell wall biosynthesis. The substrate specificity of SerP2, dl-alanine/dl-serine/glycine, overlaps that of CycA transporters, d-alanine/d-serine/glycine (24–28). SerP2 shares 42% sequence identity with CycA of E. coli. CycAs are involved in cell wall biosynthesis by taking up d-alanine from the medium. d-Alanine is an essential peptidoglycan precursor that is incorporated into the peptide cross-links between the glycan layers (29). Besides uptake from the medium, the cell can also synthesize d-alanine from l-alanine, a conversion catalyzed by alanine racemase. In L. lactis, deletion of the single alanine racemase gene (alr) makes the cells strictly dependent on an external supply of d-alanine (32). Cytoplasmic d-alanine is converted to d-Ala-d-Ala, a reaction which is catalyzed by d-alanine ligase, the first step in the synthesis of the pentapeptide building blocks. The latter two enzymes are the targets of the antibiotic d-cycloserine (30, 31), a structural analog of d-alanine and a broad-spectrum antibiotic used as a second-line drug against Mycobacterium tuberculosis (33) to treat tuberculosis. The two enzymes are competitively inhibited by the drug, and the effectiveness depends on the availability of d-alanine. In SA medium from which l-alanine was omitted to challenge the endogenous synthesis of d-alanine, growth of L. lactis was completely inhibited by 0.5 mM d-cycloserine (Fig. 4C and D). Growth was restored when 2 mM d-alanine was included in the medium, strongly suggesting a mechanism of action of the antibiotic that is the same as in E. coli and M. tuberculosis. Growth was not restored in the SerP2 deletion mutant, demonstrating that SerP2 was solely responsible for the uptake of d-alanine in the wild-type strain. Both in E. coli and M. tuberculosis it is believed that the CycA transporter is also responsible for the uptake of d-cycloserine into the cell, even though some doubt was raised recently (34, 35). The SerP2 deletion mutant was equally sensitive to d-cycloserine, suggesting that at least in this mutant of L. lactis, another transporter is responsible for the uptake of the drug.

The serP1 and serP2 genes are found adjacent on the L. lactis chromosome separated by only 71 bp, which do not contain a potential terminator or promoter sequence (16). They most likely originate from a recent gene duplication event. The pair is found in all sequenced strains of L. lactis and in Lactococcus garvieae but not outside the genus Lactococcus. The encoded transporters are members of the amino acid transporter (AAT; TC 2.A.3.1) family, which is part of the amino acid-polyamine-organocation (APC) superfamily of transporters (19). The AAT family contains the CycA transporters of E. coli and M. tuberculosis and many other proteins annotated as CycA, mostly from the phyla Actinobacteria, Firmicutes, and Proteobacteria. Phylogenetic analysis of the abundant members of the AAT family shows that SerP1 and SerP2 split off from a single branch in the tree, i.e., they share a most recent ancestor that is not shared with any of the other members (see Fig. S2 in the supplemental material). Sequence analysis indicates that SerP2 is more closely related to the members outside the genus Lactococcus than SerP1, suggesting that the SerP1 and SerP2 transporters may have evolved through the following sequence of events. The common ancestor of SerP1 and SerP2 in the primordial Lactococcus species was a CycA-like transporter involved in cell wall synthesis. The gene duplicated, after which one of the genes evolved to encode SerP2 and largely retained the original substrate specificity and function. The other gene evolved to encode SerP1, with a different substrate specificity, to have a different physiological function, i.e., l-serine uptake to complement the limited capacity of endogenous l-serine biosynthesis. The SerP1, SerP2, and CycA transporters demonstrate that sequence identity does not necessarily correspond with substrate specificity.

Growth of L. lactis in nitrogen rich media like milk during fermentation or GM17 in the laboratory does not depend on the capacity to take up free amino acids from the medium. For instance, in milk, casein is the main nitrogen source, with only 2% of the nitrogen required for growth being derived from uptake of free amino acids (36). In agreement, deletion of the SerP1 and SerP2 transporters did not affect growth on GM17 medium. Nevertheless, uptake of l-serine by SerP1 and l-alanine by SerP2 in resting cells grown in GM17 medium demonstrated significant levels of the transporters in the membrane (Fig. 2). CodY is a global transcriptional regulator of nitrogen metabolism that represses expression of genes involved in the proteolytic system and amino acid transport and metabolism in nitrogen-rich media (37). Repression follows by binding to an upstream located consensus sequence AATTTTCWGAAAATT (the CodY box) that is located in one or more copies upstream of the target gene. Copy number and sequence conservation of the CodY box are most likely important for the level of repression. The serP1 gene was previously identified as a putative target for CodY (37). The upstream region contains a single copy of a putative CodY box located 31 bp upstream of the start codon (AATCATCTGATAATT; boldface indicates the consensus CodY box). Close examination of the upstream region of serP2 identified a putative CodY box located 28 bp upstream of the start codon (CAGTCTCTGAAAATC). The expression of serP1 was reported to be upregulated only 1.5 times in a codY mutant (37), which would be in line with the significant SerP1 (and SerP2) transport activity observed here in cells grown in GM17 medium. Apparently, the consensus sequences are too weak to result in significant repression of expression of the serP1 and serP2 genes under these conditions.

Alternatively, amino acid transporters may play a role in the excretion of amino acids during growth on peptide-rich media. The composition of the peptides likely does not match the cellular need for particular amino acids, and the excess amino acids generated in the cytoplasm by hydrolysis of the imported peptides are excreted from the cells. One way to address a role for an amino acid transporter in this process is to feed resting cells with a peptide of particular amino acid composition and monitor the release of the amino acids from the wild-type cells and cells missing the transporter. In a previous study (18), this approach was used successfully in L. lactis using the dipeptides His-Leu, Gly-Glu, and Phe-Val to target the function of the histidine transporter HisP, the glutamate transporter AcaP, and the phenylalanine transporter FywP, respectively. Here, the dipeptides l-Ser–l-Val and l-Ala–l-Leu were used in combination with the ΔserP1 and ΔserP2 mutants. In none of these cases did deletion of the transporter affect the release of the free amino acids from the cells. The mechanism by which the excess amino acids are released from the cell during growth on peptides remains enigmatic.