ABSTRACT
Fourteen genes encoding putative secondary amino acid transporters were identified in the genomes of Lactococcus lactis subsp. cremoris strains MG1363 and SK11 and L. lactis subsp. lactis strains IL1403 and KF147, 12 of which were common to all four strains. Amino acid uptake in L. lactis cells overexpressing the genes revealed transporters specific for histidine, lysine, arginine, agmatine, putrescine, aromatic amino acids, acidic amino acids, serine, and branched-chain amino acids. Substrate specificities were demonstrated by inhibition profiles determined in the presence of excesses of the other amino acids. Four knockout mutants, lacking the lysine transporter LysP, the histidine transporter HisP (formerly LysQ), the acidic amino acid transporter AcaP (YlcA), or the aromatic amino acid transporter FywP (YsjA), were constructed. The LysP, HisP, and FywP deletion mutants showed drastically decreased rates of uptake of the corresponding substrates at low concentrations. The same was observed for the AcaP mutant with aspartate but not with glutamate. In rich M17 medium, the deletion of none of the transporters affected growth. In contrast, the deletion of the HisP, AcaP, and FywP transporters did affect growth in a defined medium with free amino acids as the sole amino acid source. HisP was essential at low histidine concentrations, and AcaP was essential in the absence of glutamine. FywP appeared to play a role in retaining intracellularly synthesized aromatic amino acids when these were not added to the medium. Finally, HisP, AcaP, and FywP did not play a role in the excretion of accumulated histidine, glutamate, or phenylalanine, respectively, indicating the involvement of other transporters.
INTRODUCTION
The lactic acid bacterium (LAB) Lactococcus lactis is widely used for the production of cheese and buttermilk and is therefore of great commercial importance. Lactic acid bacteria have adapted to nutrient-rich environments and lack various biosynthetic pathways. Most dairy and laboratory strains of L. lactis are auxotrophic for a number of amino acids, for example, branched-chain amino acids and histidine. It is believed that the capacity to synthesize these amino acids was lost during adaptation to milk (1, 2), which provides all essential amino acids in the form of its major protein constituent, casein (3). Casein is degraded by an efficient proteolytic system that involves protein hydrolysis into a range of peptides of different lengths, the transport of some of these peptides via the oligopeptide uptake system Opp, and further degradation to free amino acids by intracellular peptidases (4, 5). Nevertheless, dairy strains of L. lactis are still able to grow in media containing free amino acids as the sole source of amino acids (6), implying that transport systems for essential, free amino acids have been conserved. A range of transport systems for essential as well as nonessential amino acids that use proton motive force (PMF) (by proton symport), electrochemical gradients of the substrate and countersubstrate (exchange), or ATP hydrolysis as the driving force has been described (7). The exchange systems are in fact not used to supply amino acids for protein biosynthesis but are part of simple metabolic pathways that generate energy in the form of ATP or PMF by the deimination or decarboxylation of the precursor amino acid combined with precursor/product exchange across the membrane (8–10).
Most of the knowledge about amino acid transport in L. lactis and other lactic acid bacteria was obtained from studies performed around 1990. Transport studies using whole cells or membrane vesicles revealed secondary uptake systems specific for branched amino acids and methionine (11, 12), alanine and glycine (13), lysine (14), serine and threonine, histidine, cysteine (15), and tyrosine and phenylalanine (16). ATP hydrolysis-driven transport systems were found for glutamate and glutamine (17) and possibly proline (18). Only a few amino acid transporters have since been cloned and characterized. For L. lactis, only bcaP and glnPQ were cloned and experimentally shown to encode a branched-chain amino acid transporter (19) and an ATP-driven glutamate/glutamine transporter (20), respectively. Other genes have been annotated as amino acid transporters but mainly on the basis of their homology with other known transporters or their genetic context.
Most described bacterial and fungal amino acid transporters that use proton symport or precursor/product exchange as an energy-coupling mechanism are members of the amino acid-polyamine-organocation (APC) superfamily (TC 2.A.3 in the classification described previously by Saier [21]). Here we set out to clone and functionally express the APC family members of L. lactis and to determine their substrate specificity. Four mutants, lacking one of the identified transporter genes, were constructed and used to demonstrate the role of the transporters in the physiology of the cell.
MATERIALS AND METHODS
Media and growth conditions.L. lactis NZ9000 (22) and JP9000 (23), strains derived from strain MG1363 carrying the nisRK genes in the pepN and pseudo_10 loci, respectively, were used for the nisin-inducible expression of amino acid transporter genes and as parental strains of transporter deletion mutants. L. lactis cells were grown at 30°C in M17 medium supplemented with 25 mM glucose (here referred to as GM17 medium) and containing 5 μg/ml chloramphenicol, when appropriate, or in SA medium, a chemically defined medium (6) containing free amino acids, unless stated otherwise. Cells of Escherichia coli, used as a cloning host for deletion constructs, were grown in LB medium containing 150 μg/ml erythromycin.
Plasmid and strain construction.Genes encoding putative amino acid transporters were amplified from chromosomal DNA of L. lactis MG1363 and IL1403 (in the case of aguD), using primers listed in Table S1 in the supplemental material, and cloned behind the nisin-inducible promoter PnisA in pNZ8048 (24), using NcoI or PagI restriction sites at the translational start of the genes and an XbaI restriction site approximately 200 bp downstream of the genes. The resulting plasmids were transformed into L. lactis NZ9000 for nisin-inducible expression.
The markerless deletion of lysP, lysQ, llmg_1452 (ylcA in strain IL1403 [see Table S2 in the supplemental material]), and llmg_2011 (ysjA in IL1403) was performed with a two-step integration-and-excision system using plasmid pCS1966 (25), which is based on the positive selection of an erythromycin resistance marker and negative selection against a 5-fluoroorotic acid transporter encoded by oroP. For the deletion of lysQ, up- and downstream flanking regions of approximately 700 bp, overlapping the first 50 bp and last 100 bp of the gene, were amplified from chromosomal DNA using primers lysQ-fl1-fw, lysQ-fl2-rv, lysQ-fl2-fw, and lysQ-fl2-rv (see Table S1 in the supplemental material) and cloned adjacently into pCS1966 using E. coli DH5α as a cloning host. The resulting plasmid was transformed into L. lactis NZ9000. Erythromycin-resistant colonies were tested for integration at the lysQ site by using PCR. The correct clones were grown overnight in nonselective medium to allow for recombination and the excision of the plasmid backbone including either the wild-type or the disrupted lysQ gene, resulting in the lysQ deletion or a return to the wild-type situation, respectively. Cells were spread onto SA medium agar plates containing 10 μg/ml 5-fluoroorotic acid for selection against the oroP gene, and colonies were tested by colony PCR for the deletion of lysQ. For the deletions of lysP, llmg_2011, and llmg_1452, approximately 2.6-kb fragments of the respective genes, including approximately 700 bp of the up- and downstream regions, were amplified by using primers lysP-fl1-fw/2011-fl1-fw/1452-fl1-fw and lysP-fl2-rv/2011-fl2-rv/1452-fl2-rv (see Table S1 in the supplemental material) and cloned into pCS1966. From the resulting constructs, approximately 1.2-kb internal fragments of the transporter genes were deleted by amplifying the whole constructs except the internal fragments by using primers lysP-fl1-rv/2011-fl1-rv/1452-fl1-rv and lysP-fl2-fw/2011-fl2-fw/1452-fl2-fw (see Table S1 in the supplemental material). The PCR products were self-ligated after digestion with restriction enzymes, resulting in the deletion constructs, which were transformed into E. coli DH5α for propagation. The plasmids were transformed into L. lactis JP9000, an MG1363 derivative containing the nisRK genes in the pseudo_10 locus (23). The procedure was continued as described above for the deletion of lysQ.
Amino acid transport assays.L. lactis strains harboring expression constructs or the empty vector pNZ8048 were grown in GM17 medium to the mid-exponential phase (optical density at 600 nm [OD600] of 0.6) and induced with 5 ng/ml nisin. After 1 h, cells were harvested, washed, resuspended to an OD600 of 2 in ice-cold 100 mM potassium phosphate buffer (pH 6.0) containing 0.2% glucose, and kept on ice until use. Samples of 100 μl of the cell suspension were incubated for 5 min at 30°C with continuous stirring, followed by the addition of 14C-labeled amino acids (Ala, Arg, Asn, Asp, Glu, His, Ile, Leu, Lys, Phe, Pro, Ser, Tyr, and Val) or 14C-labeled putrescine (PerkinElmer) to final concentrations ranging from 1 to 5 μM, depending on the amino acid. In the case of putrescine uptake by cells expressing aguD, 50 μM (final concentration) unlabeled agmatine was added 1 min after allowing cells to accumulate [14C]putrescine (present at a concentration of 4.7 μM). Uptake was stopped by the addition of 2 ml ice-cold 0.1 M LiCl and filtration through a 0.45-μm-pore-size nitrocellulose filter (BA85; Schleicher & Schuell GmbH). The filter was washed once with 2 ml 0.1 mM 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).
Measurement of amino acid and dipeptide concentrations using reverse-phase high-performance liquid chromatography (RP-HPLC).Samples were run on a Shimadzu High-Speed HPLC Prominence UFLC 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 buffer (pH 6) containing 10 mM glucose and histidine-leucine, glycine-glutamate, or phenylalanine-valine at concentrations of 5 mM were centrifuged, and supernatants were subjected to derivatization by diethylethoxymethylenemalonate (DEEMM), as described previously by Pudlik and Lolkema (26). The detection of aminoenone derivatives was performed by the use of an Alltech Platinum EPS C18 column with dimensions of 250 by 4.6 mm, operated at 25°C through a binary gradient using eluent A (25 mM acetate [pH 5.8], 0.02% sodium azide) and eluent B (an 80:20 mixture of acetonitrile and methanol), as described previously (27). External standards were prepared by mixing the dipeptide and the two corresponding free amino acids at concentrations of 4 mM and 0.2 mM or 1 mM and 50 μM, respectively.
RESULTS
Secondary amino acid transporters of Lactococcus lactis.The genomes of L. lactis strains IL1403, MG1363, KF147, and SK11 were screened for homologues of amino acid transporters by BLAST searches using different members of the amino acid-polyamine-organocation (APC) superfamily (21) as query sequences. Fourteen proteins that showed significant homology (more than 20% amino acid sequence identity) with known APC family members were identified (Table 1; see also Table S2 in the supplemental material). Twelve of the proteins were found in all four strains, whereas YlcA (IL1403 nomenclature) was not present in strain SK11, and YrfD was present only in L. lactis subsp. lactis strains IL1403 and KF147. Sequence identities between the putative transporter proteins varied from 64% (LysQ and LysP) down to insignificant values below 20% (e.g., ArcD1 and YlcA) (see the phylogenetic tree in Fig. 1). All 14 protein sequences shared the hydropathy profile typical of APC family members, indicating the same fold containing 12 transmembrane segments and with the N and C termini located in the cytoplasm (28, 29). The ArcD2 and ArcD1 proteins have one and two additional transmembrane segments, respectively. The arcD1, arcD2, gadC, and yrfD genes are located in clusters encoding a putative arginine deiminase (ADI) pathway (arcD1 and arcD2) (30), a glutamate decarboxylation pathway (gadC) (31), and an agmatine deiminase (AgDI) pathway (yrfD) (9), and although not demonstrated experimentally, their functions were annotated as arginine/ornithine, glutamate/γ-aminobutyric acid, and agmatine/putrescine exchangers, respectively. In contrast, 8 of the other putative amino acid transporter genes appear to be monocistronic, with neighboring genes of unknown function or functions that could not be directly related to amino acid metabolism. The adjacent ydgB and ydgC genes are separated by a 71-bp noncoding fragment, which does not contain a potential terminator or clear promoter sequence. These genes may be transcribed as one mRNA.
Amino acid transporters of L. lactis subsp. lactis and subsp. cremoris
Phylogenetic tree of all APC family members and BrnQ (LIVCS family) of L. lactis subsp. lactis (IL1403 and KF147) and L. lactis subsp. cremoris (MG1363 and SK11). The nomenclature for strain IL1403 was used, except for BrnQ. Newly assigned names are shown in parentheses. Protein alignment and tree building were carried out by using ClustalX. Amino acid sequence identities between proteins are indicated.
The brnQ gene was identified in strains MG1363, SK11, and KF147 but was missing in IL1403. BrnQ was proposed previously to be a branched-chain amino acid transporter (19, 32) but is not related in amino acid sequence to the APC family of transporters. Instead, BrnQ belongs to the LIVCS family of branched-chain amino acid/cation symporters (TC 2.A.26) (21). Nevertheless, the hydropathy profile alignment suggested that the two families are distantly related and belong to the same structural class (MemGen class ST[2]) (28, 29). No other homologues of the LIVCS family were present in any of the four L. lactis strains. Members of the DAACS family (TC 2.A.23), another family that contains many amino acid transporters, were completely absent from the L. lactis genomes.
Cloning and characterization of the amino acid transporters.The putative amino acid transporter genes, except for ctrA (also known as bcaP), which was well characterized as a branched-chain amino acid transporter previously (19), were expressed in the MG1363-derived strain NZ9000 using the NICE expression system (24). Cells expressing the genes were screened for enhanced uptake by comparing the uptake of 14 amino acids and putrescine (only in case of yrfD) to the uptake by control cells harboring the empty vector pNZ8048. Uptake was measured after 10 s and with the amino acids present at concentrations ranging from 0.9 μM to 5 μM. The expression of 10 transporter genes resulted in increased uptake of one or more amino acids (Table 1), while the remaining four genes (yagE, yibG, yshA, and gadC) showed no increased activity with any of the amino acids (data not shown). The positive clones were studied in more detail by monitoring the uptake of the identified substrate(s) with time (Fig. 2, left, and 3) and by measuring the inhibition of uptake by excess unlabeled amino acids (Fig. 2, right).
Substrate specificities of amino acid transporters. Cells of L. lactis NZ9000 containing the empty vector pNZ8048 or pNZ8048 with lysP (A), lysQ (B), ydgC (C), ydgB (D), ylcA (E), ysjA (F), arcD1 (G), or arcD2 (H) cloned behind the nisin-inducible promoter were induced for 1 h with 5 ng/ml nisin and tested for the uptake of 14C-labeled amino acids at concentrations of 1 to 5 μM (Table 1). Left panels indicate time-dependent uptake in cells harboring the empty vector (closed symbols) or cells expressing a transporter (open symbols). The substrates are indicated on the y axes. In panel E, glutamate uptake is indicated by circles, and aspartate uptake is indicated by triangles. In panel F, phenylalanine uptake is indicated by circles, and tyrosine uptake is indicated by triangles. Right panels indicate substrate specificities determined by the inhibition of the substrate (indicated on the y axes) by 1 mM unlabeled amino acids, indicated on the x axes with one-letter codes. Uptake was determined after 10 s of incubation with the labeled substrate. The value for the control (Co), i.e., without the addition of unlabeled amino acids, was set to 100% (last bar). Error bars represent standard deviations of the means of data from two experiments.
The overexpression of lysP resulted in a strongly increased initial rate of uptake of lysine (Fig. 2A). The inhibition profile showed that LysP has a narrow substrate specificity; none of the other amino acids inhibited the uptake of lysine significantly. A 10-fold increase in the initial rate of uptake of histidine was observed upon the overexpression of lysQ (Fig. 2B). The inhibition profile indicated that LysQ is mainly a histidine transporter but has a low affinity for the other basic amino acids arginine and lysine. The transporter was renamed HisP. The adjacent genes ydgC and ydgB resulted in 20-fold- and 5-fold-increased initial rates of uptake of serine, respectively (Fig. 2C and D). Uptake by cells expressing ydgC was followed by the rapid efflux of label from the cells that was not observed in the case of ydgB or in the control cells. The nature of this efflux is unclear. The inhibition studies revealed a different substrate specificity of the two transporters. YdgC showed a high affinity for threonine and cysteine, in addition to serine, while serine uptake by YdgB was most strongly inhibited by alanine and glycine, whereas threonine and cysteine had much less of an effect. The transporters YdgC and YdgB were renamed SerP1 and SerP2, respectively. Initial rates of uptake of aspartate and glutamate increased 30-fold and 7-fold, respectively, in cells expressing ylcA (Fig. 2E). The inhibition profile of glutamate uptake revealed that the transporter is specific for the two acidic amino acids. The transporter was renamed AcaP. Strongly increased rates of uptake of phenylalanine and tyrosine were observed upon the expression of ysjA. The inhibition profile of tyrosine uptake showed that YsjA also has an affinity for tryptophan and that the transporter is highly specific for aromatic amino acids. YsjA was renamed FywP. The arcD1 and arcD2 genes found in the ADI cluster resulted in 3- to 4-fold-increased initial rates of arginine uptake (Fig. 2G and H). Also, the rate of uptake of ornithine, the end product of the ADI pathway and the countersubstrate in arginine/ornithine exchange, was increased more than 10-fold (Table 1). The inhibition profiles of both transporters indicated a main substrate specificity for arginine with a low affinity for lysine and histidine. The branched-chain amino acids isoleucine, leucine, and valine were taken up in cells expressing brnQ at a 2- to 4-fold-higher initial rate than in control cells (Fig. 3A). Although the affinity for other amino acids was not determined, BrnQ appears to be a branched-chain amino acid transporter, as suggested previously (19) and similar to the characterized homologue in Lactobacillus delbrueckii (32). Recently, the yrfD gene, present in strain IL1403 but not in strain MG1363, was proposed to encode the precursor/product exchanger AguD, which would mediate the uptake of extracellular agmatine in exchange for intracellular putrescine, the product of the AgDI pathway (9). While control cells did not show significant putrescine uptake activity, as observed previously (33), the expression of aguD enabled the cells to take up putrescine with an initial rate of approximately 1.4 nmol/min/mg protein (Fig. 3B). The addition of a 10-fold excess of unlabeled agmatine to cells that had accumulated putrescine led to a rapid efflux of the latter, demonstrating the exchange mode of transport.
Amino acid uptake via BrnQ (A) and AguD (B). (A) Cells carrying brnQ (open symbols) or harboring the empty vector pNZ8048 (closed symbols) were incubated with [14C]leucine (squares), [14C]isoleucine (triangles), or [14C]valine (circles) at the concentrations shown in Table 1. (B) Cells carrying aguD or harboring pNZ8048 were incubated with 4.7 μM [14C]putrescine. To demonstrate agmatine/putrescine exchange, 50 μM unlabeled agmatine was added to cells which were allowed to accumulate [14C]putrescine for 1 min (triangles).
Deletion of lysP kills high-affinity lysine uptake without affecting growth on free amino acids.The lysP gene was deleted from L. lactis JP9000, an MG1363 derivative, using a two-step integration-and-excision system (34). The lysine uptake activity in resting cells of strains JP9000 and JP9000ΔlysP grown to the mid-exponential phase in GM17 medium was measured at a lysine concentration of 1.6 μM (Fig. 4A). The initial rate of uptake in the mutant had decreased drastically, by approximately 20-fold, to 0.5 nmol/min/mg protein, suggesting that LysP is the major lysine transporter in L. lactis at low lysine concentrations but that at least one other, possibly low-affinity, uptake system for lysine is present.
Phenotype of the lysP deletion mutant. (A) Uptake of [14C]lysine (at a concentration of 1.6 μM) in wild-type cells (strain JP9000) (open circles) or lysP mutant cells (closed circles). Cells were pregrown to the mid-exponential phase in GM17 medium. (B) Growth of JP9000 (triangles) and JP9000ΔlysP (circles) cells in SA medium containing 56 μM lysine (closed symbols) or no lysine (open symbols). Shown are data from a representative experiment of at least two independent experiments.
GM17 medium is a rich medium containing approximately 2.5 mM lysine and most other free amino acids at concentrations in the millimolar range as well as peptides derived from casein by partial hydrolysis (35). The growth characteristics of the wild-type and mutant strains in GM17 medium were the same (not shown). The growth of wild-type strain JP9000 in SA medium, a chemically defined medium containing free amino acids (6), was slightly affected when the standard 1.4 mM lysine was omitted from the medium. The cells grew slightly slower and to a 10%-lower optical density, indicating that the lysine biosynthesis capacity is not a bottleneck in growth (Fig. 4B). In line with these results, the growth of the JP9000ΔlysP strain was similar to that of the wild-type strain under the two conditions. A concentration of lysine as low as 56 μM in the medium was sufficient to overcome the small growth defect in the wild-type strain, while the mutant did not respond to this concentration (Fig. 4B). Apparently, the uptake of lysine by LysP was responsible for the growth enhancement at low lysine concentrations observed for the parent strain.
HisP is essential for growth on free amino acids at low histidine concentrations.A hisP deletion mutant was constructed as described above for lysP. Histidine uptake by JP9000ΔhisP cells grown in GM17 medium was strongly reduced to a low but still significant level when measured at a concentration of 1.5 μM (Fig. 5A). As described above for LysP, it appears that HisP is the main high-affinity histidine transporter but that a second transport system, which is expressed at a low level and/or has a low affinity for histidine, seems to be present.
Phenotype of the hisP deletion mutant. (A) Uptake of [14C]histidine (at a concentration of 1.5 μM) in wild-type cells (open circles) or hisP mutant cells (closed circles). Cells were pregrown to the mid-exponential phase in GM17 medium. (B and C) Growth of NZ9000 (B) and NZ9000ΔlysQ (C) cells in SA medium containing 1 mM (●), 250 μM (○), 50 μM (▼), 10 μM (△), or no (■) histidine. Shown are data from a representative experiment of at least two independent experiments.
In rich GM17 medium, which contains approximately 86 μM histidine (35) as well as peptides derived from casein, no significant difference in growth between the wild-type and mutant strains was observed (not shown). In chemically defined SA medium that contains 250 μM histidine, JP9000ΔhisP grew at a rate approximately 50% of that of the wild type, while the biomass yield was approximately 70% of that of the wild type (Fig. 5B and C). Increasing the concentration of histidine in the medium to 1 mM almost completely restored the growth defect of the mutant, which is in line with the presence of a second system able to transport histidine but with a much lower affinity. With no histidine in the medium, both wild-type and ΔhisP mutant cells were unable to grow, demonstrating the inability of L. lactis to synthesize histidine. A histidine concentration as low as 10 μM allowed the wild-type strain to grow, but the biomass yield was limited to 40%. A concentration of 50 μM resulted in the same growth as that in standard SA medium (Fig. 5B). The mutant strain hardly showed any growth at these two concentrations, demonstrating that the uptake of histidine by HisP is essential for growth at low concentrations (Fig. 5C).
In the absence of free histidine, the growths of both strains in SA medium could be fully restored by the addition of 250 μM dipeptide histidine-leucine (not shown), demonstrating separate systems for the uptake of free histidine and the His-Leu dipeptide, which is probably transported by the di- and tripeptide transporter DtpT (36).
AcaP is essential for growth on free amino acids in the absence of glutamine.Resting cells of ΔacaP mutant strain JP9000ΔacaP grown in SA medium showed a strongly reduced initial rate of uptake of aspartate at a concentration of 2.2 μM, while the initial rate of uptake of glutamate at a concentration of 1.9 μM was only marginally affected (Fig. 6A). Since AcaP was shown to transport both aspartate and glutamate (Fig. 2E), the result indicates that a second transporter that transports glutamate but not aspartate is present and that this transporter is responsible for the uptake of glutamate under the conditions of the experiment. Most likely, the transporter is GlnPQ, an ABC-type glutamate/glutamine transporter (20).
Phenotype of the acaP deletion mutant. (A) Uptake of [14C]glutamate (circles) or aspartate (triangles) in wild-type cells (open symbols) or acaP mutant cells (closed symbols). Glutamate and aspartate were present at concentrations of 1.9 and 2.2 μM, respectively. Cells were grown to the mid-exponential phase in SA medium. (B) Growth of JP9000 cells in standard SA medium (closed circles) and SA medium from which either glutamate (closed triangles), glutamine (open circles), or both (open triangles) were omitted. (C) Growth of JP9000ΔacaP cells in standard SA medium (closed triangles), SA medium without glutamine (open circles), and SA medium without glutamine and containing 0.4 mM glutamate rather than the standard concentration of 2.1 mM (open triangles). Shown are data from a representative experiment of at least two independent experiments.
In GM17 medium rich in all amino acids, no difference in growth between the wild type and the acaP deletion mutant was observed. Similarly, in standard SA medium, the growths of the wild-type and mutant strains were not significantly different (compare Fig. 6B and C). SA medium contains 2.1 mM glutamate but no aspartate. Clearly, AcaP plays no role in the uptake of aspartate for biosynthetic purposes during growth. The omission of glutamate from the medium did not affect the growth of the wild type. The omission of glutamine, which is present at a concentration of 0.7 mM, resulted in only a marginally lower biomass yield. When both glutamate and glutamine were omitted from the medium, no growth was observed (Fig. 6B). Apparently, L. lactis has the capacity to efficiently take up and interconvert glutamate and glutamine. In line with this, the growth of the ΔacaP mutant was not affected when glutamate was omitted from the medium (not shown). The cells are likely to take up glutamine from the medium by an appropriate transporter and internally convert it to glutamate. In contrast, in medium without glutamine, initially, the cells were not able to take up glutamate at a rate high enough to support cell growth, demonstrating the role of AcaP as a glutamate uptake system during the growth of wild-type cells under the same conditions (Fig. 6C). Apparently, the rate of uptake of glutamate by the mutant cells observed at micromolar concentrations (Fig. 6A) is not high enough to support the need for glutamate and glutamine in biosynthesis under the growth conditions. Remarkably, after a long lag phase of approximately 5 h, the ΔacaP mutant started growing at approximately half of the wild-type rate (Fig. 6C). The lag phase indicates an adaptation process, possibly involving the upregulation of a glutamate uptake system other than AcaP (glnPQ). The role of the AcaP transporter during growth on free amino acids in the absence of glutamine is further emphasized by the complete lack of growth of the mutant when, in addition, the concentration of glutamate was reduced to 0.4 mM (Fig. 6C). The growth of the wild type under these conditions was only marginally affected (not shown).
Deletion of fywP reduces the efficiency of growth at low concentrations of the aromatic amino acids.The deletion of the fywP gene from strain JP9000 resulted in a 5-fold-lower initial rate of uptake of phenylalanine in cells grown in GM17 medium and when measured at a concentration of 1 μM (Fig. 7A). It follows that FywP is the main transporter for phenylalanine uptake under these conditions.
Phenotype of the fywP deletion mutant. (A) Uptake of [14C]phenylalanine (at a concentration of 1 μM) in wild-type cells (open circles) or ΔfywP mutant cells (closed circles). Cells were grown to the mid-exponential phase in GM17 medium. (B) Growth of JP9000 (closed symbols) and JP9000ΔfywP (open symbols) cells in SA medium containing 0.24 mM, 0.06 mM, and 0.1 mM (circles) or 48 μM, 12 μM, and 20 μM (triangles) phenylalanine, tyrosine, and tryptophan, respectively, or without any of these amino acids (squares). Shown are data from a representative experiment of at least two independent experiments.
No difference in growth was observed between the wild type and mutant in rich GM17 medium or chemically defined SA medium, which contains 1.2 mM phenylalanine, 0.3 mM tyrosine, and 0.5 mM tryptophan. Although these amino acids are known to not be essential for L. lactis MG1363, their presence usually stimulates growth (6, 16), suggesting that the rate of cellular biosynthesis is limited. Nevertheless, the growths of both strains were not affected when the concentrations of all three amino acids in SA medium were decreased 5-fold, and even when the concentrations were decreased 25-fold into the micromolar range, the wild-type strain grew at the same rate but with a slightly lower biomass yield (Fig. 7B). The ΔfywP mutant then showed an approximately 30%-lower growth rate and a 20%-lower biomass yield, indicating that FywP is important for optimal growth efficiency in an environment with low concentrations of aromatic amino acids. Interestingly, in SA medium with none of the aromatic amino acids present, a lower growth rate for the fywP mutant was also observed, which may indicate that FywP plays a role in the recovery of aromatic amino acids that have leaked out of the cells by some other means.
HisP, AcaP, and FywP play no significant role in the excretion of accumulated histidine, glutamate, and phenylalanine.The ΔhisP, ΔacaP, and ΔfywP mutants and the wild type were grown to the mid-exponential growth phase in SA medium and incubated in a 100 mM potassium phosphate (pH 6.0) buffer containing 25 mM glucose and 5 mM dipeptides His-Leu (strains NZ9000ΔhisP and NZ9000), Gly-Glu (JP9000ΔacaP and JP9000), and Phe-Val (JP9000ΔfywP and JP9000). The consumption of the dipeptide and the excretion of the corresponding free amino acids were analyzed by HPLC (Fig. 8). In all cases, the dipeptide was taken up efficiently during the first 5 min at rates of 0.17, 0.3, and 0.4 μmol/min/mg protein for His-Leu, Phe-Val, and Gly-Glu, respectively. Thereafter, the rates dropped considerably, especially for the Gly-Glu and Phe-Val peptides. The free amino acids produced by the intracellular peptidase activity appeared readily in the medium at a high rate during the first 5 min, after which the rate decreased, in line with the disappearance of the peptide from the medium. The His-Leu and Phe-Val peptides were stoichiometrically converted to free amino acids by the cells. With the Gly-Glu peptide, the rate of glutamate excretion was 2 times lower than that of glycine excretion during the first 5 min, suggesting that part of the free intracellular glutamate was further converted. No products other than glycine, glutamate, and glycine-glutamate were detected in the medium. More importantly, no difference was observed between the wild-type strains and the transporter mutants for any of the conversions, suggesting that neither HisP, AcaP, nor FywP played a role in the excretion of accumulated amino acids.
Dipeptide uptake and amino acid excretion by wild-type and transporter mutant strains. Cells grown to the mid-exponential phase in SA medium were harvested and resuspended to an OD600 of 2 in 100 mM KPi (pH 6.0) buffer containing 25 mM glucose and 5 mM histidine-leucine (A), glycine-glutamate (B), or phenylalanine-leucine (C) and incubated at 30°C. Concentrations of free amino acids and dipeptide in the buffer were determined from samples taken at the indicated time points. Open symbols, wild-type strains NZ9000 (A) and JP9000 (B and C); closed symbols, NZ9000ΔlysQ (A), JP9000ΔylcA (B), and JP9000ΔysjA (C); squares, dipeptide concentrations; circles, histidine (A), glycine (B), and phenylalanine (C) concentrations; triangles, leucine (A), glutamate (B), and valine (C) concentrations. Amino acid concentrations are indicated on the left y axes, and dipeptide concentrations are indicated on the right y axes.
DISCUSSION
Since the emergence of genome sequencing and with the improvement of recombinant DNA techniques, surprisingly little work has been carried out on the identification of amino acid transport systems in lactic acid bacteria, more specifically in the model LAB L. lactis. Most of the knowledge still originates from extensive transport studies of whole cells and membrane vesicles derived from wild-type L. lactis strains that were performed in the late 1980s and early 1990s (8, 14, 16, 17, 37). Only two amino acid transport systems have since been cloned for expression and functional characterization: the ABC-type glutamate/glutamine transporter GlnPQ, encoded by glnP and glnQ (20), and BcaP, a secondary branched-chain amino acid transporter, encoded by bcaP (19). Here 13 genes of L. lactis MG1363 and 1 gene of L. lactis IL1403 (aguD), with all but one (brnQ) encoding members of the APC family (21) and annotated either as unknown or as amino acid transporters based on homology, were cloned and overexpressed in L. lactis NZ9000. Substrates were identified (Fig. 2 and 3), and the specificities for all proteinogenic amino acids were determined (Fig. 2, right). Four transporters, SerP1, SerP2, AcaP, and FywP, for which no function was assigned based on homology, are entirely newly identified amino acid transporters. SerP1 and SerP2 transport serine, and AcaP and FywP transport acidic amino acids and aromatic amino acids, respectively. FywP has the same substrate specificity but is not closely related in sequence to the aromatic amino acid transporter AroP, hence the different names. LysP and HisP (formerly LysQ) were identified as lysine and histidine transporters, respectively, which was predicted previously based on regulatory sequences in their promoter regions (38, 39). The predicted functions of AguD as an agmatine/putrescine exchanger (9) and ArcD1 and ArcD2 as arginine/ornithine exchangers (30) were also confirmed. BrnQ, which is not homologous to the other APC family transporters, was shown to be a branched-chain amino acid transporter, in agreement with transport studies with a brnQ knockout strain described previously (19). No activity could be determined for four genes, as overexpression did not result in an increased uptake of any of the amino acids. The specificity of GadC, which was predicted to be a glutamate/γ-butyric acid exchanger (31), could not be confirmed in this study.
Much of the data presented here are in agreement with data from previous reports on the transport of amino acids in wild-type cells or membrane vesicles of L. lactis. The presence of separate secondary transport systems for lysine (14), histidine (15), serine and threonine (15), and phenylalanine and tyrosine (16) is confirmed here by the identification and characterization of LysP, HisP, SerP1, SerP2, and FywP. In contrast, a secondary transporter for aspartate and glutamate uptake, i.e., AcaP, was not reported previously. Instead, the ABC transporter GlnPQ, with a high affinity for glutamate and glutamine and a low affinity for aspartate, was described (20, 40). The physiological relevance for having both an ATP-driven glutamate transporter and a secondary, probably proton symport-driven glutamate transporter is unclear but may be related to the importance of glutamate and glutamine in nitrogen metabolism in the cell.
The physiological role of the LysP, HisP, AcaP, and FywP transporters was demonstrated by using knockout mutants. Uptake in resting cells demonstrated that the lysP, hisP, acaP, and fywP genes encode the major transporters for lysine, histidine, aspartate, and tyrosine, respectively, when the substrates were present at micromolar concentrations. For all these substrates, unknown, low-affinity transport systems appeared to be present as well. Glutamate uptake in the acaP mutant, although a substrate of AcaP, was similar to that in the wild type, indicating that a second transport system, most likely GlnPQ, is the major glutamate transport system under these conditions. In chemically defined medium containing free amino acids at concentrations ranging from 0.3 to 3.4 mM but no protein and peptides as additional sources of amino acids (SA medium [6]), the growth of the mutants was similar to that of the wild-type strain, which is in line with the presence of additional low-affinity uptake systems. The role of the transporters in growth became apparent at low concentrations, but the effect was strongly dependent on the capacity of L. lactis to synthesize the particular amino acid. While HisP was essential for growth at low concentrations of histidine (Fig. 5B), LysP restored only a small growth defect observed for the wild type when no lysine was present (Fig. 4B). The phenotype of the fywP mutant was in between those phenotypes: a significantly lower growth rate was observed at low concentrations of the aromatic amino acids and even without them (Fig. 7B). FywP transports aromatic amino acids that can be synthesized by L. lactis itself. The effect at low concentrations is in agreement with previously reported observations that the addition of phenylalanine and, in general, the addition of some nonessential free amino acids stimulate growth of L. lactis MG1363 in defined medium (6, 16). Jensen et al. (27) demonstrated previously that in the intracellular amino acid pool of L. lactis MG1363 cells growing in SA medium (containing free amino acids), only aspartate was derived to a substantial proportion (38%) from de novo biosynthesis and that all other amino acids were taken up largely from the medium. This probably reflects regulation to avoid the high energy costs of biosynthesis compared to the uptake of external free amino acids. The observed difference between the wild type and the fywP mutant in SA medium without aromatic amino acids suggests that FywP is involved in retaining these amino acids in the cytoplasm. Aromatic amino acids that would then leak out of the cells by passive diffusion would be taken up again by FywP. The excretion or leakage of phenylalanine and tyrosine was indeed observed previously by Whipp et al. (41) in an E. coli mutant strain that lacks aromatic amino acid transporters. The situation in milk, with casein as the main amino acid source, is more complicated. While casein is preferred over free amino acids as a source of amino acids using its efficient proteolytic system (3), it has been demonstrated at the same time that some strains of L. lactis do not reach their maximal growth rate in milk and that this can be overcome by the addition of certain amino acids (42). The implication is that in complex medium containing casein as well as free amino acids (for example, released by other LAB in mixed fermentations), amino acid transporters still play a role. In this light, it is not surprising that dairy strains of L. lactis that have lost their biosynthetic pathways for histidine and branched-chain amino acids during their adaptation to milk (1, 2) have retained the transporters that can import these as free amino acids, to not become fully dependent on casein as an amino acid source.
During fermentation in milk, some amino acids are excreted by L. lactis (43–45), indicating that the amino acid composition of the peptides taken up after the hydrolysis of casein by PrtP is not matching the requirements for biosynthesis. The excretion of excess amino acids was proposed to be coupled to the excretion of a proton (proton symport), which would contribute to the generation of metabolic energy in the form of proton motive force (PMF) (energy-recycling model). A potential role for HisP, AcaP, and FywP as the secondary transporters involved in this process was investigated by feeding the cells dipeptides and measuring the release of the free amino acids in the wild type and mutants. The dipeptides histidine-leucine, glycine-glutamate, and phenylalanine-valine were efficiently taken up, and the free amino acids produced intracellularly by hydrolysis were readily excreted. No difference between the knockout mutants and the wild type was observed, suggesting that either a different transport system or passive diffusion was responsible for the release from the cells. Passive diffusion was previously shown to be a mechanism of excretion of accumulated neutral amino acids (12, 46); however, it is unlikely that glutamate diffuses across the membrane to a significant extent. Instead, it was demonstrated previously that in Corynebacterium glutamicum, glutamate excretion is mediated by NCgl1221, a mechanosensitive channel which can also mediate betaine transport (34, 47, 48). A deletion mutant accumulated up to 10-fold-higher intracellular glutamate concentrations. In that same organism, other transport systems were found to mediate the excretion of lysine and arginine (LysE) (49), threonine and serine (ThrE) (49), and branched-chain amino acids and methionine (BrnFE) (50, 51). In L. lactis MG1363, only distant homologues of NCgl1221 (the llmg_0881 gene) and BrnFE (the azlC-llmg_0882 gene pair) can be found. Interestingly, llmg_0881 is located downstream of the gadCB operon (but transcribed in the opposite direction), putatively encoding the glutamate/γ-aminobutyrate pathway proteins (31). Apart from homologues of these systems, other types of carriers that mediate the excretion of amino acids may be present.
ACKNOWLEDGMENTS
We thank Agata Pudlik for assistance with the HPLC analysis of amino acids and dipeptides.
This work was supported by the European Community's Seventh Framework Programme, grant agreement no. 211441-BIAMFOOD, and by the Netherlands Organization for Scientific Research (NWO-ALW).
FOOTNOTES
- Received 3 October 2012.
- Accepted 3 November 2012.
- Accepted manuscript posted online 9 November 2012.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.01948-12.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
