Arginine and Citrulline Catabolic Pathways Encoded by the arc Gene Cluster of Lactobacillus brevis ATCC 367

ABSTRACT

High concentrations of l-arginine or l-citrulline in the growth medium provided the wine bacterium Lactobacillus brevis with a significant growth advantage. The arginine deiminase pathway (ADI) arc gene cluster of Lactobacillus brevis contains three genes—arcD, arcE1, and arcE2—encoding putative l-arginine/l-ornithine exchangers. Uptake experiments with Lactococcus lactis cells expressing the genes showed that all three transported l-ornithine with affinities in the micromolar range. Similarly, ArcD and ArcE2 transported l-arginine, while ArcE1 transported l-citrulline, an intermediate of the ADI pathway. Chase experiments showed very efficient exchange of l-arginine and l-ornithine by ArcD and ArcE2 and of l-citrulline and l-ornithine by ArcE1. Low affinities (millimolar range) combined with low translocation rates were found for ArcD and ArcE2 with l-citrulline and for ArcE1 with l-arginine. Resting cells of Lactobacillus brevis grown in the presence of l-arginine and l-citrulline rapidly consumed l-arginine and l-citrulline, respectively, while producing ammonia and l-ornithine. About 10% of l-arginine degraded was excreted by the cells as l-citrulline. Degradation of l-arginine and l-citrulline was not subject to carbon catabolite repression by glucose in the medium. At a high medium pH, l-citrulline in the medium was required for induction of the l-citrulline degradation pathway. Pathways are proposed for the catabolic breakdown of l-arginine and l-citrulline that merge at the level of ornithine transcarbamylase in the ADI pathway. l-Arginine uptake is catalyzed by ArcD and/or ArcE2, l-citrulline by ArcE1. l-Citrulline excretion during l-arginine breakdown is proposed to be catalyzed by ArcD and/or ArcE2 through l-arginine/l-citrulline exchange.

IMPORTANCE Lactobacillus brevis, a bacterium isolated from wine, as well as other food environments, expresses a catabolic pathway for the breakdown of l-citrulline in the medium that consists of the l-citrulline/l-ornithine exchanger ArcE1 and part of the catabolic arginine deiminase (ADI) pathway enzymes. The proposed pathways for l-arginine and l-citrulline breakdown provide a mechanism for l-citrulline accumulation in fermented food products that is the precursor of the carcinogen ethyl carbamate.

INTRODUCTION

The arginine deiminase (ADI) pathway is a bacterial trait found most abundantly among Lactobacillales in the phylum Firmicutes (1, 2). The conversion of l-arginine to l-ornithine serves as a source of energy, carbon, and/or nitrogen and as a mechanism for survival in acidic environments (3–5). The degradation of l-arginine to l-ornithine, ammonia, and carbon dioxide generates 1 mol of ATP per mol of l-arginine consumed. The pathway consists of a single transport step and three metabolic steps (Fig. 1A). The l-arginine/l-ornithine exchanger that catalyzes electroneutral exchange between l-arginine and l-ornithine is responsible for the concomitant uptake of the substrate l-arginine and excretion of the end product l-ornithine (6, 7). Once internalized, l-arginine is converted into l-citrulline and ammonia catalyzed by ADI (encoded by arcA). Subsequently, citrulline is converted into l-ornithine and carbamoyl-P catalyzed by ornithine transcarbamylase (OTC; encoded by arcB). Carbamoyl phosphate is used to phosphorylate ADP, yielding ATP, carbon dioxide, and ammonia, a reaction catalyzed by carbamate kinase (CK; encoded by arcC) (4, 8). Since no metabolic energy is needed for the transport reaction, ATP produced by the ADI pathway can be entirely used for other energy-demanding purposes.

FIG 1.

Schematic representation of the ADI pathway (A) and the ADI gene cluster (B) in Lb. brevis. ArcD/E, l-arginine/l-ornithine exchangers ArcD or ArcE encoded by genes arcD and arcE, respectively; ADI, arginine deiminase encoded by arcA; OTC, ornithine transcarbamylase encoded by arcB; CK, carbamate kinase encoded by arcC; 2CS, two-component regulatory system. Boxed genes represent putative operon structures taken from the DOOR2.0 database (30). Filled triangles point to cre sites found in the RegPrecise 3.0 database (17).

Analysis of arc gene clusters found in the genomes of 124 different bacterial species revealed two types of l-arginine/l-ornithine exchangers that were found at more or less the same abundance (2). The encoding genes arcD and arcE belong to different gene families and the gene products ArcD and ArcE to different structural classes. Detailed functional analysis of ArcD1 of Lactococcus lactis and ArcE of Streptococcus pneumoniae revealed similar kinetic parameters and the ArcE exchanger of S. pneumoniae was shown to recover ADI pathway activity in a L. lactis arcD1 and arcD2 knockout strain. Therefore, the two types of genes and the operon structures in which they are embedded appear to be products of convergent evolution. The most striking difference between the two types of exchangers was the ability of ArcE to efficiently catalyze exchange of l-ornithine and the pathway intermediate l-citrulline which opens up the possibility of l-citrulline catabolism by arcE-containing bacteria.

The majority of the arc gene clusters analyzed contained either the arcD or the arcE gene, but few (3/124) contained both (2). Of these, the ADI gene cluster found in the wine bacterium Lactobacillus brevis is one of the most complex clusters (Fig. 1B). A putative operon arcABDTC contains the arcD gene encoding the l-arginine/l-ornithine exchanger, the three genes encoding the metabolic enzymes ArcA, ArcB and ArcC and the arcT gene, encoding a transaminase of unknown function. The operon is flanked downstream by the regulator gene argR and upstream by the arcE1 gene, encoding an ArcE type of l-arginine/l-ornithine exchanger. In the opposite direction, two genes encoding a two-component regulatory system are followed by yet another arcE type of gene, arcE2, and the dipeptidase-encoding gene arcP, which is found in arc clusters of other bacteria as well. It follows that the arc cluster on the Lb. brevis ATCC 367 chromosome encodes three putative arginine/ornithine exchangers: ArcD, ArcE1, and ArcE2. Here, it is shown that the cluster encodes overlapping pathways for l-arginine and l-citrulline breakdown, both beneficial for growth. l-Arginine breakdown involves l-arginine/l-ornithine exchange catalyzed by ArcD, and l-citrulline breakdown involves l-citrulline/l-ornithine exchange catalyzed by ArcE1. Moreover, the l-arginine pathway provides an excretion pathway for l-citrulline, which is the precursor for the carcinogen ethyl carbamate in fermented food products.

RESULTS

Functional characterization of ArcD, ArcE1, and ArcE2 of Lb. brevis ATCC 367.

The arc gene cluster found on the chromosome of Lb. brevis ATCC 367 contains three genes encoding putative arginine/ornithine exchangers: arcD, arcE1, and arcE2. The genes were cloned and expressed in L. lactis strain ΔarcD1D2 (9) using the nisin controlled expression (NICE) system (10). Resting cells of the resulting strains DΔarcD1D2, E1ΔarcD1D2, and E2ΔarcD1D2, respectively, were assayed for uptake of ¹⁴C-labeled l-arginine, l-citrulline, and l-ornithine at concentrations in the micromolar range in a potassium phosphate pH 6 buffer containing no glucose (Fig. 2A to C). The host strain ΔarcD1D2 lacks the genes arcD1 and arcD2, which encode the two known arginine/ornithine exchangers of the parent L. lactis JP9000, reducing the background uptake activities of l-arginine and l-ornithine essentially to zero (Fig. 2A and B) (9). In addition, no uptake of l-citrulline by the host strain could be detected (Fig. 2C). All three recombinant strains readily accumulated l-ornithine under the conditions of the experiments (Fig. 2B). In contrast, strains DΔarcD1D2 and E2ΔarcD1D2 on the one hand and E1ΔarcD1D2 on the other hand were complementary with respect to l-arginine and l-citrulline uptake. The former strains took up arginine and no citrulline, while the latter strain took up citrulline and no arginine (Fig. 2A and C). The uptake experiments were performed in a buffer containing no glucose, i.e., the cells were not energized. Therefore, the accumulation is a result of exchange activity with an (unknown) substrate in the cytoplasm (see reference 2).

FIG 2.

(A to C) Uptake of l-[¹⁴C]arginine (A), l-[¹⁴C]ornithine (B), and l-[¹⁴C]citrulline (C) at initial concentrations of 1.3, 10, and 10 μM, respectively, by resting cells of L. lactis strains ΔarcD1D2 (♢), DΔarcD1D2 (□), E1ΔarcD1D2 (△), and E2ΔarcD1D2 (○). (D to F) Uptake of l-[¹⁴C]ornithine (△) at an initial concentration of 10 μM by resting cells of L. lactis strains DΔarcD1D2 (D), E1ΔarcD1D2 (E), and E2ΔarcD1D2 (F). At the time point indicated by the arrow, 1 mM unlabeled l-arginine (□), l-ornithine (○), or l-citrulline (♢) or 50 mM unlabeled l-citrulline (◆) was added to the cell suspension.

Affinity constants of the ArcD, ArcE1, and ArcE2 transporters for l-arginine, l-citrulline, and l-ornithine were inferred from initial rate measurements or inhibition experiments (Table 1). The concentration of the inducer nisin present during growth of the cells was adjusted to allow for initial rate measurements. The kinetic affinities for the substrates were all in the micromolar range. The K_m values for the common substrate l-ornithine were found to be similar with 15 ± 3, 27 ± 2, and 15 ± 1 μM for ArcD, ArcE1, and ArcE2, respectively. The affinity of ArcD for l-arginine was an order of magnitude lower than observed for ArcE2 with K_m values of 5.8 ± 2 and 0.4 ± 0.1 μM, respectively, and the K_m value of ArcE1 for l-citrulline was 8.5 ± 1 μM. ArcD and ArcE2 showed no uptake activity with l-citrulline at μM concentrations (Fig. 2A and C). Inhibition studies of l-ornithine uptake by ArcD and ArcE2 revealed affinities for l-citrulline in the millimolar range. A concentration of 10 mM l-citrulline inhibited the uptake of l-ornithine by ArcD by 20%, indicating that the inhibition constant is way over 10 mM. The affinity of ArcE2 for l-citrulline was an order of magnitude higher with a K_i of 1.9 ± 0.4 mM. Using the same approach, the affinity of ArcE1 for l-arginine was determined to be 1 order of magnitude lower than the kinetic affinity for l-citrulline and l-ornithine with a K_i of 0.15 ± 0.01 mM. Estimated maximal rates from the initial rate measurements varied between 12 and 152 nmol/min · mg. The numbers represent a combination of turnover numbers with the different substrates and expression levels of the transporters in the membrane.

TABLE 1.

Kinetic parameters of the ArcD, ArcE1, and ArcE2 transporters of Lb. brevis expressed in L. lactis ΔarcD1D2

Substrate	Mean ± SDa
	ArcD			ArcE1			ArcE2
	Km (μM)	Vmax (nmol/min · mg)	Ki (mM)	Km (μM)	Vmax (nmol/min · mg)	Ki (mM)	Km (μM)	Vmax (nmol/min · mg)	Ki (mM)
l-Arg	5.8 ± 2	63 ± 39				0.15 ± 0.01	0.4 ± 0.1	12 ± 1.5
l-Orn	15 ± 3	152 ± 114		27 ± 1.5	13 ± 2.7		15 ± 0.7	20 ± 2
l-Cit			>10	8.5 ± 1	33 ± 7				1.9 ± 0.4

^aValues represent the means from at least two replicates. K_i values were inferred from the inhibition of l-ornithine uptake at 10 μM. l-Arg, l-arginine; l-Orn, l-ornithine; l-Cit, l-citrulline.

The capacity of the transporters to rapidly exchange substrates across the cytoplasmic membrane was determined by chase experiments. Resting cells containing the transporters were allowed to accumulate l-[¹⁴C]ornithine, after which an excess of unlabeled substrate was added. Rapid release of internalized label by 1 mM unlabeled l-arginine or l-ornithine indicated high l-arginine/l-ornithine and l-ornithine/l-ornithine exchange capacities of ArcD and ArcE2 (Fig. 2D and F). In contrast, a very slow release was observed upon the addition of 1 mM unlabeled l-citrulline (open diamonds). In part, the lack of l-citrulline/l-ornithine exchange by ArcD and ArcE2 might have been due to the low affinity of the two transporters for l-citrulline (>10 and 1.9 mM, respectively) (Table 1). Raising the l-citrulline concentration to 50 mM resulted in significant release of label from the cells (closed diamonds), but this was still at a much lower rate than observed with l-arginine and l-ornithine. ArcE1 rapidly exchanged l-[¹⁴C]ornithine with l-citrulline and l-ornithine and at a much slower, but significant rate with l-arginine (Fig. 2E). Thus, ArcD and ArcE2 are likely to function as l-arginine/l-ornithine exchangers, while ArcE1 is likely to function as an l-citrulline/l-ornithine exchanger.

Growth enhancement of Lb. brevis ATCC 367 by l-arginine and -citrulline.

Growth of Lb. brevis ATCC 367 was followed in modified MRS medium (mMRS) supplemented with 2% glucose, with or without an additional 20 mM l-arginine or l-citrulline (Fig. 3A). The cell density increased similarly up to about 10 h, after which growth in mMRS-glucose slowed down relative to growth in the same medium with extra l-arginine or l-citrulline. Growth was slightly better with l-arginine than observed with l-citrulline while, ultimately, the same cell yield was obtained in the three media (Fig. 3A). The experiment was repeated in the relatively poor basal medium (BM) containing 0.1% glucose where the cell yield was dramatically higher in the presence of either of the two amino acids. Again, growth was slightly better with additional l-arginine than with l-citrulline (Fig. 3B). Clearly, the presence of high concentrations of both l-arginine and l-citrulline gave Lb. brevis a growth advantage in both types of media.

FIG 3.

(A and B) Growth of Lb. brevis in mMRS supplemented with 2% glucose (A) and BM (B) with no further additions (♢) and in the presence of 20 mM l-arginine (□) and 20 mM l-citrulline (△). (C and D) Degradation of l-arginine by resting cells of Lb. brevis grown in mMRS–l-arginine supplemented with 2% glucose (C) and degradation of l-citrulline by cells grown in mMRS–l-citrulline supplemented with 2% glucose (D). The initial concentrations of l-arginine and l-citrulline were 5 mM. Symbols: □, l-arginine; △, l-ornithine; ○, ammonium; ♢, l-citrulline.

Resting cells of Lb. brevis grown in mMRS–l-arginine supplemented with 2% glucose and resuspended in a 100 mM potassium phosphate (pH 6) buffer at an optical density at 600 nm (OD₆₀₀) of 4 consumed l-arginine at a rate of 0.076 mM/min (Fig. 3C). The kinetics was zero order down to >90% conversion, indicating high affinity of the metabolic pathway for the substrate. Concomitant with the disappearance of l-arginine from the supernatant buffer, l-ornithine and double the amount of ammonia were produced. A small amount of the ADI pathway intermediate l-citrulline accounting for around 10% of the original l-arginine concentration was excreted by the cells as well. The same pattern was observed with cells grown in BM–l-arginine at a slightly higher l-arginine consumption rate of 0.090 mM/min (see Fig. S1 in the supplemental material). The pattern is consistent with l-arginine breakdown by the ADI pathway in both media (Fig. 1). Under the same conditions, resting cells grown in mMRS–l-citrulline supplemented with 2% of glucose (Fig. 3D) and BM–l-citrulline (see Fig. S1 in the supplemental material) consumed l-citrulline at rates of 0.073 and 0.083 mM/min, respectively. At the same time, more or less equal amounts of l-ornithine and ammonia were produced by the cells in line with the conversion of l-citrulline by ornithine transcarbamylase (ArcB) and carbamate kinase (ArcC) (Fig. 1). It follows that both l-arginine and l-citrulline present in the medium are metabolized by (part of) the ADI pathway and that the growth advantage observed during growth in the presence of high concentrations of the two amino acids may be largely due to ATP production by substrate level phosphorylation in the pathway. The substrate specificities of the ArcD, ArcE1, and ArcE2 transporters (see above) indicate that the l-arginine pathway involves the ArcD and/or ArcE2 exchangers and the l-citrulline pathway the ArcE1 exchanger.

Differential expression of the l-arginine and l-citrulline pathways.

Differential expression of the pathways was demonstrated by measuring l-arginine and l-citrulline consumption rates by the same cells grown in mMRS–l-arginine medium and, similarly, in mMRS–l-citrulline medium (Fig. 4). As above, cells grown in high l-arginine readily consumed l-arginine at an initial rate of 0.12 mM/min and in the process, in addition to l-ornithine and ammonia (not shown), 1.1 mM l-citrulline was excreted into the buffer. The same cells consumed l-citrulline at a much lower rate of 0.004 mM/min. Consistently, no significant decrease was observed in the amount of l-citrulline excreted during l-arginine consumption after depletion of l-arginine (Fig. 4A). In contrast, cells grown in mMRS–l-citrulline consumed l-citrulline as expected but also l-arginine at rates comparable to cells grown in mMRS–l-arginine (Fig. 4). During the metabolism of l-arginine, l-citrulline was excreted as well, but in this case followed by reuptake after the depletion of l-arginine. It follows that in mMRS-based media, expression of the l-citrulline pathway requires induction by l-citrulline in the medium. In contrast, the l-arginine pathway is constitutively expressed or the basal level of l-arginine present in mMRS is enough to give full induction. Since the ArcE1 exchanger is the only component of the l-citrulline pathway not shared with the l-arginine pathway, the arcE1 gene is likely to be the target for the regulation by l-citrulline. Importantly, excretion of l-citrulline during l-arginine consumption was not subject to regulation by l-citrulline, and excretion and (re)uptake of l-citrulline are therefore different activities catalyzed by different transporters, i.e., by ArcD/ArcE2 and ArcE1, respectively.

FIG 4.

Degradation of l-arginine (□) and l-citrulline (◆) by resting cells of Lb. brevis grown in mMRS–l-arginine (A) and by cells grown in mMRS–l-citrulline (B). ♢, l-citrulline in the supernatant buffer during l-arginine degradation.

Induction of expression of the arcE1 gene by l-citrulline was not required in BM medium. l-Citrulline excreted during l-arginine consumption by cells grown in BM–l-arginine was taken up by the cells again (see Fig. S1 in the supplemental material). The two main differences between the mMRS supplemented with 2% glucose and BM media are the glucose concentrations (2% versus 0.1%) and the pH (pH 7 versus pH 6). The l-arginine and l-citrulline consumption rates by Lb. brevis grown in mMRS supplemented with 2% glucose (Fig. 3C and D) and mMRS without glucose (Fig. 4) were comparable, suggesting no hierarchal control of glucose metabolism over the ADI pathway. This conclusion was confirmed by a more systematic approach in which the rates of l-arginine and l-citrulline consumption and product formation were measured when catalyzed by cells grown in mMRS–l-arginine and mMRS–l-citrulline media to which 0, 0.1, 0.5, 1, and 2% glucose was added (see Fig. S2 in the supplemental material).

l-Citrulline consumption was measured by cells of Lb. brevis grown in mMRS–l-arginine, mMRS–l-citrulline, BM–l-arginine, and BM–l-citrulline with the pH set at 5, 6, and 7. The cells were harvested at a low cell density (OD₆₀₀ of 0.3) to minimize pH changes in the medium caused by cell metabolism. Cells grown in the presence of l-citrulline consumed l-citrulline at rates that were not significantly different for the different pH values in both types of growth media (Fig. 5). In contrast, cells grown in the presence of l-arginine showed an ∼10-fold-higher l-citrulline metabolic rate at pH 5 than observed at pH 7 in both media (Fig. 5). At pH 6, cells grown in BM–l-arginine behaved like cells grown at pH 5, while cells grown in mMRS–l-arginine behaved like cells grown at pH 7. The same result was obtained for l-citrulline reuptake following l-arginine metabolism (see Fig. S4 in the supplemental material). It follows that the pH of the growth medium is the dominant factor for the requirement of induction of the arcE1 gene by l-citrulline, but this appears to be modulated by other differences between the BM and mMRS media.

FIG 5.

Initial rate of l-citrulline consumption by resting cells of Lb. brevis grown in mMRS (A) and BM (B) supplemented with 20 mM l-arginine (hatched bars) and 20 mM l-citrulline (open bars) and set at the indicated pH values. Error bars show the standard deviations based on duplicate experiments.

Excretion of l-citrulline during l-arginine metabolism.

Excretion of l-citrulline during l-arginine metabolism may be due to l-arginine/l-citrulline exchange and/or l-ornithine/l-citrulline exchange catalyzed by ArcD and/or ArcE2. l-Ornithine/l-citrulline exchange would depend on the built up of a l-ornithine concentration in the buffer in the initial phases of l-arginine metabolism, and a delay in l-citrulline excretion may be expected. The rate of l-citrulline excretion was measured at cell OD₆₆₀ values of 2, 1, and 0.5, which are 2-, 4-, and 8-fold lower than the standard densities, respectively, resulting in similarly lower l-ornithine production rates. The concentration of l-arginine was lowered to 1 mM to allow for a better resolution between the l-citrulline and l-arginine peaks in the high-pressure liquid chromatography (HPLC) profile in the initial phases when little l-citrulline is produced. The l-arginine consumption rates and l-ornithine production rates were proportional to the cell density, with average rates of 23 ± 2 and 26 ± 2 μM/min per unit of OD₆₆₀, respectively (Fig. 6A). At all cell densities, the l-citrulline concentration increased from the first point measured at a constant rate of 1.7 ± 0.2 μM/min per unit of OD₆₆₀ (Fig. 6B), strongly suggesting that l-ornithine/l-citrulline exchange is not the mechanism by which l-citrulline is excreted from the cells.

FIG 6.

l-Arginine consumption (filled symbols) and l-ornithine production (open symbols) (A) and l-citrulline production (B) by resting cells of Lb. brevis degrading l-arginine resuspended at an OD₆₀₀ of 2 (circles), 1 (triangles), or 0.5 (squares). The cells were grown in mMRS–l-arginine.

Cells of Lb. brevis grown in mMRS–l-citrulline took up l-[¹⁴C]citrulline at an initial rate of 3.3 ± 0.2 nmol/min · mg when the initial concentration was 50 μM (Fig. 7). A much lower rate was observed with cells grown in mMRS–l-arginine in agreement with the lack of induction of the arcE1 gene by l-citrulline in mMRS at pH 7. Addition of 5 mM l-arginine following the accumulation of l-[¹⁴C]citrulline by the cells grown in the presence of l-citrulline was followed by a slow release of label from the cells, indicating that the translocation of cytoplasmic l-citrulline out of the cells mediated by l-arginine/l-citrulline exchange is not a very efficient process. The rate of exchange was on the same order of magnitude as observed for the chase of internalized l-[¹⁴C]ornithine by unlabeled l-citrulline catalyzed by the recombinant L. lactis strains DΔarcD1D2 and E2ΔarcD1D2 (Fig. 2D and F).

FIG 7.

Uptake of l-[¹⁴C]citrulline at an initial concentration of 50 μM by resting cells of Lb. brevis grown in mMRS–l-arginine (♢) and mMRS–l-citrulline (□). At the 5-min time point, 5 mM unlabeled l-arginine was added to the cell suspension. Error bars show the standard deviations based on two replicates.

DISCUSSION

The arginine deiminase (ADI) pathway converts l-arginine in the medium via the cytoplasmic intermediate l-citrulline into l-ornithine that is excreted back to the medium (Fig. 1A). The pathway provides the cells with an amount of metabolic energy of 1 mol of ATP per mol of l-arginine consumed. Lb. brevis metabolizes l-citrulline present in the medium by part of the ADI pathway, including the ATP-generating step giving the cells a similar growth advantage. l-Citrulline enters the cell in exchange for the end product l-ornithine catalyzed by ArcE1, a dedicated l-citrulline/l-ornithine exchanger. Internalized l-citrulline is converted to l-ornithine and carbamoylphosphate by ornithine transcarbamylase (OTC; arcB), and the latter is converted to ammonia and carbon dioxide, yielding ATP by CK (arcC) (Fig. 8, top). The l-citrulline pathway consists of ArcE1, ArcB, and ArcC. The l-arginine pathway in Lb. brevis makes use of the l-arginine/l-ornithine exchanger ArcD, followed by ADI (arcA), OTC (arcB), and CK (arcC) (Fig. 8, bottom). The two pathways merge at OTC. Associated with the l-arginine pathway of Lb. brevis is the excretion of l-citrulline, an activity observed before in Lb. sakei (11). Excretion in Lb. brevis follows from l-arginine/l-citrulline exchange catalyzed by ArcD (see below). The genes encoding the enzymes required for the l-arginine pathway are likely to be transcribed in one transcriptional unit (the arc operon; see Fig. 1B), which is consistent with the apparent constitutive expression in the different media used in this study. In contrast, the components of the l-citrulline pathway are not embedded in a single operon, and the pathway is subject to differential regulation of expression of the l-citrulline/l-ornithine exchanger gene arcE1. Expression requires the presence of l-citrulline in medium of pH 7, but not at pH 5. The effect of pH on the reuptake of l-citrulline after excretion during l-arginine metabolism observed in Lb. sakei may be explained by the same regulatory mechanism (11). The physiological relevance of the regulation is not clear. Above, the transport activity in the l-arginine pathway in Lb. brevis was assigned to ArcD since the arcD gene is part of the arc operon. However, the functional characteristics of ArcD and ArcE2 are very similar (Table 1), and the possibility that ArcE2 plays a role here as well cannot be excluded. The arcE2 and arcP genes are likely to be under the control of the two-component regulatory system encoded immediately upstream of the two genes, and identifying the effector may facilitate finding their physiological role.

FIG 8.

Models for the l-citrulline (top) and l-arginine (bottom) degradation pathways in Lb. brevis. See the text and the legend to Fig. 1 for further explanation.

In Gram-positive bacteria the CcpA-dependent carbon catabolite repression (CCR) system mediates hierarchical control of glucose over other carbon and energy sources. Briefly, the primary sensor in the regulatory route is HPr kinase (HPrK) that is activated by fructose-1,6-biphosphate. HPrK phosphorylates HPr, a general component of the phosphoenolpyruvate-dependent phosphotransferase system (PTS) that is responsible for the uptake of most carbohydrates in most bacteria. Phospho-HPr binds to the transcriptional regulator CcpA, thereby inducing the binding of the complex to so-called cre sites (“catabolite responsive elements”) in the promoter regions of the target genes (12–14). All of these elements can be found on the Lb. brevis chromosome, but glucose has nevertheless been reported to lack control over other carbohydrates in this organism (15, 16). Even though several putative cre sites were identified in the arc gene cluster (Fig. 1) (17), the data presented here also show no control of glucose metabolism over the breakdown pathways for l-arginine or l-citrulline. Lb. brevis is an obligate heterofermentative lactic acid bacterium that metabolizes glucose following the phosphoketolase pathway, which does not involve fructose-1,6-bisphosphate as intermediate, and carbohydrates are taken up via symport systems rather than the PTS. Both points do not easily fit in the generalized CCR scheme given above and, as a consequence, the functionality of the CCR system in obligate heterofermentative bacteria is currently under debate (18, 19). The functionality of the cre sites in the arc gene cluster of Lb. brevis and, more generally, the transcriptional regulation require further investigations.

A previous study demonstrated that l-arginine/l-ornithine exchange activity in the ADI pathway of different bacteria is catalyzed by either of two evolutionary unrelated proteins, the ArcD or ArcE type (2). ArcD1 of L. lactis and ArcE of S. pneumoniae were shown to catalyze the same l-arginine/l-ornithine exchange reaction. In addition, ArcE of S. pneumoniae catalyzed efficient l-citrulline/l-ornithine exchange, suggesting that the ADI operon of S. pneumoniae encodes both l-arginine and l-citrulline degradation pathways in which the ArcE transporter would be responsible for catalyzing l-arginine/l-ornithine exchange, as well as l-citrulline/l-ornithine exchange. ArcD exchangers characterized thus far do not catalyze l-citrulline/l-ornithine exchange, and it may be expected that bacteria harboring an ADI operon containing ArcD have only the l-arginine degradation pathway. In L. brevis where ArcE1 and ArcE2 are present in addition to ArcD in the ADI operon, ArcE1 catalyzes only l-citrulline/l-ornithine exchange, suggesting that the ability to degrade l-citrulline was added later in evolution to the ADI pathway in this microorganism. ArcE2 has yet another functional profile by catalyzing only l-arginine/l-ornithine exchange like ArcD does. The role of ArcE2 in Lb. brevis is not clear from the present study. Recruitment of the dedicated l-citrulline/l-ornithine exchanger by Lb. brevis suggests the presence of standalone l-citrulline degradation pathways in nature encoded by operons containing arcB, arcC, and arcE, but no arcA. In line with this, the putative operon arcBCE is found in Staphylococcus aureus and Staphylococcus carnosus, and arcECB is found in Staphylococcus carnosus and Staphylococcus epidermis, in Streptococcus agalactiae and Streptococcus iniae, and in Tetragenococcus halophilus. Remarkably, a putative operon arcDCB was identified in Clostridium ljungdahlii, suggesting that the ArcD type may also have evolved to l-citrulline transport.

Ethyl carbamate is a carcinogen that accumulates in many fermented food products and beverages (20, 21). In wine, ethyl carbamate is formed in a reaction of ethanol with the precursor l-citrulline. Accumulation of l-citrulline correlates with the ability to degrade l-arginine in the ADI pathway by, among other species, lactobacilli (22, 23). Several studies showing variations in l-citrulline accumulation in wine by modulation of ADI pathway activity link the excretion process directly to the ADI pathway (24–26). The metabolic scheme shown in Fig. 8 (bottom panel) provides a mechanism for l-citrulline excretion during l-arginine degradation by Lb. brevis. In the model, cytoplasmic l-citrulline is excreted by ArcD in competition with conversion to l-ornithine by ArcB and ArcC. Cells in which the l-citrulline pathway was expressed showed similar rates of l-arginine and l-citrulline consumption (Fig. 4B), suggesting rate limitation in the common part of the pathways, promoting the accumulation of cytoplasmic l-citrulline and therefore favoring excretion. Aspects of the model are supported by the data presented here. (i) Excretion is catalyzed by ArcD/E2, not by ArcE1. Excretion was still observed in the absence of ArcE1, which rules out a major role for the latter in the excretion process. The similar rates in the presence and absence of ArcE1 show that l-ornithine/l-citrulline exchange by ArcE1, a very efficient process (Fig. 2B), is not involved significantly in the excretion process. These observations stress that excretion and uptake of l-citrulline are catalyzed by different entities. Excretion by ArcD is part of the l-arginine degradation pathway, whereas uptake by ArcE1 is a standalone activity in the l-citrulline degradation pathway. (ii) l-Arginine inhibits reuptake of l-citrulline. Excretion of l-citrulline proceeds until l-arginine is exhausted. Consumption of excreted l-citrulline by cells that express the l-citrulline pathway is inhibited by l-arginine because of the relatively high affinity of ArcE1 for l-arginine (K_i = 0.15 mM) (Table 1). (iii) Excretion follows from l-arginine/l-citrulline exchange. In the steady state of l-arginine degradation, l-citrulline was excreted from the beginning without the need for the build up of a concentration of l-ornithine in the medium (Fig. 6), suggesting l-arginine/l-citrulline exchange as the mechanism for the excretion process. (iv) Exchange of internal l-citrulline for external l-arginine is a slow process. Both ArcD and ArcE2 have low affinity for l-citrulline and, even under saturating l-citrulline concentrations, the rate of exchange was poor in the chase experiments (Table 1 and Fig. 2D and F). Nevertheless, the low exchange capabilities of ArcD and ArcE2 are compatible with the observed excretion rates in Lb. brevis since the exchange of cytoplasmic l-citrulline with external l-arginine is also a slow process in Lb. brevis (Fig. 7). Alternatively, l-citrulline is excreted by a transporter not related to the ADI pathway. The lack of uptake of l-citrulline by Lb. brevis not expressing ArcE1 argues against this possibility.

MATERIALS AND METHODS

Strains, media, and growth conditions.

Lactococcus lactis JP9000 (referred to as the wild type), derived from strain MG1363 and carrying the nisRK genes in the pseudo_10 locus (27), is the parent of deletion mutant strain ΔarcD1D2 (9). The double mutant strain ΔarcD1D2 was used as a host for the nisin-inducible expression (10) of the arcD (LVIS_2025), arcE1 (LVIS_2028), and arcE2 (LVIS_2031) transporter genes of Lactobacillus brevis ATCC 367, yielding strains DΔarcD1D2, E1ΔarcD1D2, and E2ΔarcD1D2, respectively (see Table S1 in the supplemental material). L. lactis strains were grown at 30°C in GM17 medium containing 28 mM glucose and supplemented with 5 μg/ml chloramphenicol and nisin at the indicated concentrations when appropriate. Lb. brevis ATCC 367 was grown at 30°C in commercially available MRS medium (MRS) containing 2% glucose, in modified MRS medium (mMRS) (28) containing glucose at concentrations between 0 and 2% (wt/vol) as indicated, and in a basal medium (BM) containing 5 g/liter peptone, 3 g/liter yeast extract, and 1 g/liter glucose. The pH of mMRS was set at 7, and the pH of BM was set at 6 unless otherwise stated. Both mMRS and BM media were supplemented with 20 mM l-arginine or 20 mM l-citrulline when indicated (mMRS–l-arginine, mMRS–l-citrulline, BM–l-arginine, and BM–l-citrulline).

Growth curves of Lb. brevis strains were recorded using a Biotek Powerwave 340 96-well plate reader. Precultures in MRS were diluted to an OD₆₀₀ of 0.05 in 200 μl of the indicated medium and covered with 50 μl of silicon oil (1:4 of silicon oil M20 and M200) to prevent evaporation. The OD₆₀₀ was measured every 20 min for 30 h, with 30 s of shaking before each measurement.

Plasmid and strain construction.

DNA fragments containing the arcD, arcE1, and arcE2 genes were amplified by PCR using Lb. brevis ATCC 367 genomic DNA as the template and the oligonucleotide primers listed in Table S2 in the supplemental material. The fragments were ligated behind the nisin-inducible promoter P_nisA in pNZ8048 (10) after digestion with the NcoI and XbaI restriction enzymes, resulting in plasmids pNZarcD, pNZarcE1, and pNZarcE2, respectively. The plasmids were sent for DNA sequencing for confirmation of the expected sequences. L. lactis strain ΔarcD1D2 was transformed with the above-mentioned plasmids using a standard electroporation protocol, resulting in expression strains DΔarcD1D2, E1ΔarcD1D2, and E2ΔarcD1D2.

Transport assays. (i) Uptake experiments.

L. lactis strains ΔarcD1D2, DΔarcD1D2, E1ΔarcD1D2, and E2ΔarcD1D2 were grown in GM17 supplemented with 0.5 ng/ml nisin (ΔarcD1D2 and DΔarcD1D2) or 5 ng/ml nisin (E1ΔarcD1D2 and E2ΔarcD1D2). Lb. brevis ATCC 367 strain was grown in mMRS–l-arginine or mMRS–l-citrulline. Cells were harvested at an OD₆₀₀ of 0.7 to 0.8, washed once, and resuspended in ice-cold 100 mM potassium phosphate (pH 6.0) buffer to OD₆₀₀ values of 2 and 4 for L. lactis and Lb. brevis, respectively, and kept on ice until use. An aliquot of 100 μl of cells was preincubated for 5 min at 30°C under continuous stirring. Uptake experiments with L. lactis were started by addition of l-[¹⁴C]arginine, l-[¹⁴C]ornithine, or l-[¹⁴C]citrulline in a volume of 1 μl, resulting in final concentrations of 1.3, 10, and 10 μM, respectively. Uptake of l-[¹⁴C]citrulline by Lb. brevis was assayed at an initial concentration of 50 μM. Uptake was stopped at the indicated time points 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).

(ii) Chase experiments.

Cells of L. lactis and Lb. brevis were allowed to take up l-[¹⁴C]ornithine and l-[¹⁴C]citrulline, respectively, as described above. At the indicated time point, an excess of unlabeled l-arginine, l-ornithine or l-citrulline was added to the L. lactis cell suspension in a volume of 1 μl and at the indicated concentration, after which internal radioactivity was determined as described above at the indicated time points. The accumulated l-[¹⁴C]citrulline by Lb. brevis was chased with an excess of unlabeled l-arginine.

Kinetic parameters.

Kinetic affinities for the various substrates were inferred from initial rates of uptake measured over the first 10 s in a range of radiolabeled substrate concentrations. The K_m values of ArcD and ArcE₂ for l-arginine were measured in ranges of 1 to 20 μM and 0.16 to 1.3 μM, respectively. The K_m of ArcD, ArcE1, and ArcE2 for l-ornithine in the range of 1 to 35 μM, and the K_m of ArcE1 for l-citrulline in the range of 1 to 10 μM. The inhibition constants for l-citrulline and l-arginine were inferred from the initial rates of uptake of l-[¹⁴C]ornithine in the presence of unlabeled l-citrulline and l-arginine in the concentration range of 0 to 10 mM and 0 to 500 μM, respectively. The K_i was calculated at zero concentration of the inhibitor using the data presented in Table 1.

ADI pathway activity.

Lb. brevis ATCC 367 cells were grown in the indicated medium to a final OD₆₀₀ between 0.3 and 0.5 and subsequently harvested, washed once, resuspended to an OD₆₀₀ of 4 in 100 mM potassium phosphate (pH 6.0) buffer, and stored on ice until use. An aliquot of 0.7 ml was incubated at 30°C for 10 min. At time zero, the reaction was started by addition of 5 mM l-arginine or l-citrulline. Samples of 50 μl were taken every 10 min, diluted with 50 μl of 100 mM potassium phosphate (pH 6.0) buffer, and immediately centrifuged for 30 s at maximum speed in a tabletop centrifuge. The supernatant was stored on ice until further analysis by reversed-phase HPLC following derivatization of amino acids and ammonium ions by DEEMM (29). Aminoenone derivatives were obtained by mixing 100 μl of the supernatants with 175 μl of 1 M borate (pH 9.0) buffer, 75 μl of methanol, 2 μl of 0.1% (wt/vol) d-aminoadipic acid, and 3 μl of DEEMM in 1.5-ml closed Eppendorf tubes, followed by 30 min of incubation at room temperature in an ultrasound bath. Subsequently, the mixture was incubated at 70°C for 2 h to allow complete degradation of the excess of DEEMM. The samples were run on a Shimadzu high-speed HPLC Nexera UFLC and later analyzed using the LC Solutions 1.24 SP1 software from Shimadzu (Kyoto, Japan). The C₁₈ Shim-pack XR-ODS II (3.0 mm [inner diameter] by 75 mm) was operated at 40°C and run at a flow rate of 1 ml/min. The aminoenone derivatives were eluted by a binary gradient of eluent A (25 mM acetate buffer [pH 5.8] with 0.02% sodium azide) and eluent B (an 80:20 mixture of acetonitrile and methanol). The gradient is given in Table S3 in the supplemental material. The target compounds were identified according to the retention times of control runs with the pure compounds and quantified using a calibration curve.