Introduction

As one of the B-complex vitamins, pantothenic acid (vitamin B5) is an essential component of mammalian nutrition. It is the precursor for phosphopantotheine and thereby participates in over 100 metabolic reactions, as part of coenzyme A and acyl carrier protein (Jackowski 1996). The pantoate-β-alanine ligase (PanC)-catalyzed condensation of β-alanine and pantoic acid is the final step in pantothenic acid biosynthesis. Pantoic acid is synthesized from α-ketoisovalerate by ketopantoate hydroxymethyltransferase (PanB) and ketopantoate reductase (PanE). Additionally, the second reaction can be carried out by the ilvC gene product, acetohydroxyacid isomeroreductase (Primerano and Burns 1983).

In Escherichia coli, the vast majority of β-alanine is synthesized de novo by the l-aspartate-1-decarboxylase (EC 4.1.1.15; PanD)-catalyzed α-decarboxylation of l-aspartic acid (Begley et al. 2001). While β-alanine formation from aspartic acid seems to be the main pathway in K12 strains (Cronan 1980), E. coli B strains are also capable of degrading uracil to β-alanine (Patel and West 1987), a metabolic pathway usually found in eukaryotes (Gojkovic et al. 2001). Furthermore, the hydrolysis of carnosine by peptidase D (Klein et al. 1986) and possibly the degradation of spermidine (Tabor and Kellogg 1970) might be alternative pathways, but their relevance for pantothenate biosynthesis is unclear.

In recent years, efficient E. coli production strains for pantothenic acid have been developed. Good producers, obtained by random mutagenesis, are further improved by overexpressing the panBCD genes (Hikichi et al. 1996) or additionally ilvGM coding for pyruvate acetohydroxyacid synthase (Moriya et al. 1999). Hikichi and coworkers pointed out that the yield of pantothenic acid is raised dramatically by the addition of high β-alanine concentrations (2% or more) to the medium (Hikichi et al. 1996), indicating that β-alanine synthesis is the rate-limiting step, even in panD-overexpressing strains. Since the currently used fermentation processes are strictly dependent on β-alanine uptake, this transport process may in principle be a bottleneck.

The aim of this work was the characterization of the transport of this building block into E. coli MG1655 and the identification of the responsible carrier-protein for a possible application in an improvement of pantothenic acid-producing strains.

Materials and methods

Bacterial strains and growth

The strains and plasmids used in this study are shown in Table 1. E. coli cells were routinely grown at 37°C in Luria–Bertani (LB) medium supplemented with antibiotics (100 μg ml−1 carbenicillin, 10 μg ml−1 tetracycline, or 20 μg ml−1 chloramphenicol) as appropriate. For uptake measurements, cells were grown in M9 minimal medium (Sambrook et al. 1989) supplemented with 0.5% glucose, 1 mM MgSO4, 100 μM CaCl2, and 0.01% of a trace elements solution containing (per liter): 10 g FeSO4·7H2O, 10 g MnSO4·H2O, 10 g CaCl2·2H2O, 1 g ZnSO4·7H2O, 0.2 g CuSO4, and 0.02 g NiCl2. For pantothenic acid production with strain FV5069/pFV31/pACYC-cycA, M9 minimal medium was supplemented with 5% glucose and 0.5 mg l−1 thiamine. Corynebacterium glutamicum strain R127 (panC∷pK18mob′panC′; Sahm and Eggeling 1999) was grown in CGXII minimal medium (Keilhauer et al. 1993) supplemented with 20 μg ml−1 kanamycin and 4% glucose as a carbon source. If necessary, 50 μg l−1 Ca2+-d-pantothenate was added.

Table 1 Bacterial strains and plasmids used
Full size table

Determination of β-alanine uptake by [1-14C]β-alanine

Cells were grown for 17 h at 37°C in M9 minimal medium, washed, suspended in fresh M9 medium to an optical density at 600 nm (OD600) of 5.0, and stored on ice until used. Aliquots (2.7 ml) were transferred to test tubes and prewarmed in a water bath for 3 min at 37°C with continuous stirring. Uptake was started by the addition of [1-14C]β-alanine (specific activity 250 cpm nmol−1; Biotrend, Germany). If appropriate, inhibitors were added 30 s prior to the substrate. Samples (200 μl) were taken, vacuum-filtrated, and washed twice with 2.5 ml of 100 mM LiCl. Radioactivity incorporated in the cells was measured using a Beckman LS6500 scintillation counter.

Nonradioactive determination of amino acid uptake

Cultivation and treatment of the cells was the same as for radioactive uptake measurements. Experiments were started by the addition of substrate and samples (200 μl) were taken at intervals. The cells were immediately centrifuged through a silicon oil layer into 30 μl 20% perchloric acid (Klingenberg and Pfaff 1967) to separate them from the medium. After the addition of 45 μl H20, the cells were lyzed by ultrasonic treatment. Cell debris was precipitated by adding 5 M KOH/1 M triethanolamine (25 μl), incubating on ice for 30 min, and centrifuging for 10 min at 20,000 g. The aqueous phase was collected for HPLC analysis.

HPLC analysis of β-alanine and other amino acids

Amino acids (including β-alanine) in the samples were pre-column-derivatized with o-phthalaldehyde (Lindroth and Mopper 1979) and separated on a Hewlett Packard HP1100 HPLC system, using a Hypersil ODS5 125.0×4.6 mm column (CS-Chromatographie-Service, Germany). A gradient ranging from 97% buffer A (20 mM sodium acetate, 0.036% triethylamine, 0.3% tetrahydrofurane) to 100% buffer B (20% 100 mM sodium acetate, 40% methanol, 40% acetonitrile) was used for elution and the derivatives were detected by their fluorescence at 230 nm.

Microbiological determination of pantothenic acid

Pantothenic acid concentrations in samples were determined by monitoring the growth of the pantothenic acid auxotrophic strain C. glutamicum R127 (panC∷pK18mob′panC′). The final cell density of this strain in CGXII minimal medium is strictly dependent on the pantothenic acid content and can thereby be used for quantification. For details, see Sahm and Eggeling (1999).

Calculations

Intracellular solute concentrations were calculated assuming a cytoplasmic volume of 0.69 μl and 1.82 μl of extracytoplasmic bound medium per OD600 unit and per milliliter, respectively (Stock et al. 1977; Cayley et al. 1991). For transport rate calculations, 1 ml of cell suspension at OD600=1.0 was assumed to contain 0.34 mg dry weight of bacteria.

Cloning of cycA-expressing plasmids

For the cloning of pTrc-cycA, the 1,413 bp cycA gene of MG1655 was amplified from genomic DNA by PCR, using the primers 5′-AGAGAATTCCATGGTAGATCAGGTAAAAGTCG-3′ and 5′-GCAGAAGCTTATTTCCGCAGTTCAGCAGC-3′. After digestion with EcoRI and HindIII (restriction sites shown in italics), this product was inserted into the EcoRI/HindIII sites of vector pTrc99a. For the generation of pACYC-cycA, first the cycA promoter of MG1655 was amplified by PCR, using the primers 5′-TCGTTGGTATACAATATTAACTGCTGGAAATCCTCTAAGC-3′ and 5′-GATCTACCATGGTTTTTTCTTCCTG-3′. Then, product was ligated to the pDRIVE vector (Qiagen, Hilden, Germany) by A-T cloning. lacI q and the trc promoter were cut out of pTrc-cycA using AccI and NcoI and the remaining plasmid was purified. With AccI and NcoI (sites shown in italics) the cycA promoter was cut out of pDRIVE and ligated with the pTrc-cycA fragment. This plasmid was then cut with SspI and a 2.3-kb fragment (consisting of the cycA promoter and cycA) was isolated. This fragment was ligated with XmnI-cut pACYC184, leading to the low-copy expression vector pACYC-cycA.

Results

Biochemical characterization of β-alanine uptake

For the biotechnological production of pantothenic acid, the uptake of β-alanine is crucial (Hikichi et al. 1996). Therefore, we first characterized this transport process in E. coli MG1655. Uptake measurements with [1-14C]-labeled substrate showed an intracellular accumulation of β-alanine at 1 mM external substrate (Fig. 1). Using the H+-specific ionophore carbonylcyanide-m-chlorophenylhydrazone (CCCP), the energetics of the transport process was investigated. As a potent uncoupler, CCCP causes the dissipation of the electrochemical H+ potential, a driving force for energy-dependent transport systems. When CCCP was added 30 s prior to [1-14C]β-alanine, the uptake rate dropped dramatically and the intracellular concentration did not exceed 1 mM during the experiment. Transport experiments with varying substrate concentrations revealed Michaelis–Menten kinetics, with a K m of 2.4 mM and a maximal transport rate (V max) of 46 nmol mg−1 dry weight min−1 under our experimental conditions (data not shown).

Fig. 1
figure 1

Dependency of β-alanine transport on the electrochemical H+ potential as measured by [1-14C]β-alanine. β-Alanine concentration was 1 mM and uptake was examined with (open triangles) and without (open squares) the addition of CCCP to a final concentration of 50 μM. Independent uptake experiments were carried out in duplicate with identical results

Full size image

Identification of the β-alanine uptake system

During the transport measurements, a strong inhibition of β-alanine uptake by traces of LB medium was observed. To test whether this might result from competitive inhibition of β-alanine transport by an amino acid in the LB medium, two cultures of strain MG1655 were incubated in M9 minimal medium containing a final concentration of 1% LB medium and β-alanine was added to one of the cultures, to a final concentration of 10 mM. After 1 min of incubation, cells of both cultures were separated from the medium by silicon oil centrifugation and the cytoplasmic amino acids pools were determined by reversed-phase HPLC. In the presence of β-alanine, significantly less l-alanine was transported into the cells (data not shown). From this result, it was suggested that β-alanine uptake is mediated by an l-alanine uptake carrier. In E. coli, two different transport systems were described for l-alanine uptake, the cycA gene product and the LIV-I system (Robbins and Oxender 1973), which differ in their energetics and substrate specificities. Besides l-alanine, the CycA permease mediates uptake of glycine, d-alanine, d-serine, and d-cycloserine, whereas LIV-I is a primary active ABC transporter for the import of the branched-chain amino acids l-leucine, l-isoleucine, and l-valine. These properties were used to differentiate between CycA and LIV-I. Uptake of 1 mM β-alanine was strongly inhibited by an equimolar amount of the CycA substrate d-cycloserine, whereas l-isoleucine had no effect (Fig. 2). In an additional experiment, glycine revealed a strong inhibitory effect but the LIV-I substrate l-leucine did not (data not shown). These data suggest CycA may be the β-alanine carrier of E. coli. To confirm this assumption of β-alanine transport in strain CAG12073 (Singer et al. 1989), a MG1655 derivative carrying a Tn10 insertion in cycA was measured. Here, the uptake was reduced to 12.6% of the wild type activity at 1 mM of external substrate (1.4 nmol mg−1 dry weight min−1 vs 11.1 nmol mg−1 dry weight min−1) and could not be further inhibited by glycine or l-leucine (data not shown). Efficient β-alanine uptake obviously depends on a functionally active CycA protein.

Fig. 2
figure 2

Uptake of β-alanine by E. coli MG1655. The substrate concentration was 1 mM and transport was measured with [1-14C]β-alanine in the presence of 1 mM l-alanine (open circles), d-cycloserine (open upright triangles), l-isoleucine (open downward triangles), or without the addition of a competitor (open squares)

Full size image

For the overexpression of cycA, the plasmid pTrc-cycA was generated. This high copy number vector is a derivative of pTrc99a (Pharmacia Biotech) and contains the 1,413-bp cycA gene of MG1655. Although the strong trc promoter is repressed by the plasmid-encoded LacI repressor, the leakiness of the system allowed sufficient gene expression even without induction. As expected, transformation of strain CAG12073 by pTrc-cycA restored its β-alanine uptake activity, while vector pTrc99a without insert had no effect (data not shown).

Enhancement of β-alanine uptake by cycA-overexpression

For an enhancement of β-alanine uptake, strain MG1655 was transformed with pTrc-cycA and with pTrc99a as a control. Figure 3 shows β-alanine uptake upon the addition of 10 mM β-alanine (data obtained by HPLC analysis). In MG1655/pTrc-cycA, β-alanine uptake was much faster than in the control strain; and the internal concentration rose up to 300 mM in 10 min. Interestingly, in the cycA-overexpressing strain, net import stopped at about 300 mM and declined only slightly. Without β-alanine feeding, endogenously synthesized β-alanine was not detectable in any strain, indicating a concentration below 0.5 mM, which is the detection limit of the HPLC method used. From these results, we conclude that the overexpression of cycA is an effective tool for the enhancement of β-alanine uptake in E. coli.

Fig. 3
figure 3

β-Alanine uptake of MG1655 carrying either plasmid pTrc99a (open squares) or pTrc-cycA (open circles) at a substrate concentration of 10 mM. Cells were separated by silicon oil centrifugation and cytoplasmic β-alanine was determined by HPLC analysis

Full size image

CycA overexpressing strains become sensitive to β-alanine

During the investigation of strain MG1655/pTrc-cycA, a severe growth inhibition by β-alanine was observed. When cells were cultivated to OD600=1.3 in minimal medium, the addition of β-alanine to 1 mg l−1 significantly delayed further growth (Fig. 4); and at 10 mg l−1 it stopped growth completely. As seen in other experiments, 891 mg l−1 β-alanine (equivalent to 10 mM) even caused cell lysis (data not shown). In contrast, strains MG1655/pTrc99a and MG1655 grew normally at these concentrations and tolerated more than 18 g l−1 β-alanine. As seen in additional experiments, the CycA substrates glycine, d-alanine, and l-alanine did not inhibit the growth of MG1655/pTrc-cycA (data not shown). Since β-alanine per se is not toxic (Fig. 3) and growth inhibition occurred only in pTrc-cycA strains, we assumed that a high level of CycA protein or its activity might cause sensitivity.

Fig. 4
figure 4

Growth inhibition of E. coli MG1655/pTrc-cycA by β-alanine. Cells were grown in M9 minimal medium to OD600=1.3 and then subjected to the following concentrations of β-alanine: 10 mg l−1 (open squares), 3 mg l−1 (open circles), 1 mg l−1 (open upright triangles) and 0.3 mg l−1 (open downward triangles). As a control, MG1655/pTrc-cycA (open diamonds) and MG1655/pTrc99a (open left-pointing triangles) were grown without the addition of β-alanine

Full size image

To decrease cycA expression, we first replaced the strong trc promoter of pTrc-cycA with the native cycA promoter of E. coli MG1655. The promoter-cycA construct was cut-out of the high-copy vector and ligated into the low-copy plasmid pACYC184, leading to the expression plasmid pACYC-cycA. Subsequently, E. coli MG1655 was transformed with pACYC-cycA and tested for β-alanine susceptibility. In the resulting strain, MG1655/pACYC-cycA, the addition of 10 mM β-alanine alone slowed down growth drastically, while in the presence of β-alanine other CycA substrates such as glycine, l-alanine, or d-alanine restored growth (data not shown)—probably by a competitive displacement of β-alanine from the transporter. Interestingly, d-alanyl–d-alanine, a precursor in cell wall biosynthesis, was also able to suppress the inhibitory effect of β-alanine at least partially (data not shown). This hints at an interference in β-alanine transport and cell wall biosynthesis.

Consequence of cycA overexpression in E. coli FV5069 derivatives for pantothenic acid production

E. coli FV5069 and FV5069/pVF31 (containing a panBCD expression plasmid) are efficient pantothenic acid-producing strains (Hikichi et al. 1996). To investigate the consequence of cycA overexpression, these strains were transformed by the cycA expression plasmids and subsequently tested for pantothenate production. As FV5069/pTrc-cycA turned out to be even more sensitive to β-alanine than strain MG1655, we used only 25 μM β-alanine for the production assay in M9 minimal medium (Fig. 5a). Compared with the control, FV5069/pTrc-cycA showed a prolonged lag phase and slower growth during the exponential phase. Nevertheless, because of the similar amount of pantothenic acid produced, its specific productivity (units of product per unit biomass) was significantly higher (Fig. 5a).

Fig. 5
figure 5

a,b Effect of cycA overexpression on pantothenic acid production. a Specific productivity (units of pantothenic acid per unit biomass) of FV5069/pTrc99a (open squares) and FV5069/pTrc-cycA (open circles). The experiment was carried out in duplicate. b Pantothenic acid production by FV5069/pFV31/pACYC184 (open squares) and FV5069/pFV31/pACYC-cycA (open circles) in M9 minimal medium after the addition of 2% (224 mM) β-alanine

Full size image

The transformation of strain FV5096/pFV31 with the low-copy expression vector pACYC-cycA also led to β-alanine sensitivity, but cell lysis was not observed (data not shown). For the production assay, cells were initially grown in M9 minimal medium until they reached the stationary phase (Fig. 5b). Then, β-alanine was added to a final concentration of 2% (224 mM). After the addition, FV5069/pFV31/pACYC-cycA and FV5069/pFV31/pACYC184 both started to produce pantothenic acid, but the amount in the FV5069/pFV31/pACYC-cycA culture was even lower than in the control.

Discussion

E. coli MG1655 is able to accumulate β-alanine more than 20-fold in the cytoplasm compared with the external medium. In combination with the observed transport inhibition by CCCP, this demonstrates the presence of an active β-alanine uptake system. While the determined V max of 46 nmol mg−1 dry weight min−1 is in the range of other amino acid transporters, the K m of 2.4 mM is rather high. This indicates that β-alanine is presumably not the preferred substrate of this uptake system.

The β-alanine transporter was identified as CycA. The CycA protein is a secondary transporter which mediates the uptake of l-alanine, d-alanine, glycine, d-serine, and d-cycloserine (Wargel et al. 1970; Cosloy 1973). It belongs to the amino acid transporter family (Jack et al. 2000), a subgroup within the amino acid/polyamine/organocation superfamily of transporters. The 470-amino-acid transmembrane protein is predicted to comprise 12 transmembrane helices and to utilize H+ co-transport as the energy source (Swiss-Prot data base; http://www.expasy.org/sprot; accession no. P39312). Together with fklB, cycA is thought to form the fklB-cycA operon, whose expression is regulated by the nitrogen assimilation control (Nac) protein (Zimmer et al. 2000). Additionally, the obvious functionality of the expression plasmid pACYC-cycA implicates the activity of a predicted promoter located directly upstream of cycA.

Uptake measurements using the cycA mutant CAG12073 revealed a strongly reduced uptake rate of only 12.6%, compared with the wild type. Since this residual activity could not be further inhibited by glycine or l-leucine, it is not caused by CycA or LIV-I and may rather be due to the unspecific action of other amino acid transport systems or simply diffusion. The restored uptake activity of CAG12073/pTrc-cycA shows that CycA in fact is the main β-alanine carrier of MG1655. Moreover, the drastically increased transport rates in strain MG1655/pTrc-cycA suggest that cycA overexpression is a suitable tool for the improvement of β-alanine uptake.

However, this basically positive effect was counteracted by the observed β-alanine sensitivity caused by cycA overexpression. β-Alanine concentrations above 1 mg l−1 (11.2 μM) inhibited the growth of MG1655/pTrc-cycA significantly and 10 mM caused cell lysis. The following calculation illustrates that the observed growth inhibition is unlikely to be caused by high intracellular β-alanine concentrations. In the experiment shown in Fig. 4, β-alanine was added to the MG1655/pTrc-cycA culture at OD600=1.3, corresponding to approximately 0.9 μl cytoplasm ml−1 cell suspension. If the cells took up the entire substrate at a concentration of 1 mg l−1 (11.2 μM), which is already growth-inhibitory, the intracellular β-alanine pool would have been 12.5 mM. As can be seen in Fig. 3, the control strain MG1655/pTrc99a reached ten-fold higher intracellular concentrations without any growth inhibition when exposed to 10 mM β-alanine.

Although the reason for this effect is not yet clear, some results indicate an impairment of cell wall biosynthesis. In E. coli, the antibiotic d-cycloserine is taken up by CycA (Wargel et al. 1971). It acts by competitive inhibition of the d-alanyl–d-alanine ligase and thereby blocks cell wall biosynthesis (Lugtenberg 1972). β-Alanine could have a similar effect in the overexpressing strains, if a direct channeling of substrates from CycA to the d-alanyl–d-alanine ligase could be assumed. If, as in the overexpressing strains, the ratio of CycA to d-alanyl–d-alanine ligase was too high, this might result in a block of all dipeptide-forming enzymes, while a low transporter:ligase ratio in the wild type would prevent this detrimental effect. This assumption is supported by the finding that addition of the dipeptide d-alanyl–d-alanine was found to partially restore the growth of MG1655/pACYC-cycA in the presence of β-alanine.

As observed for the MG1655 derivatives, the cycA-overexpressing pantothenic acid production strains also turned out to be sensitive to β-alanine. Since high β-alanine concentrations in the medium are a prerequisite for pantothenic acid production (Hikichi et al. 1996), this fact excludes the direct application of cycA for strain improvement. Although the increase of specific productivity in FV5069/pTrc-cycA proves that the concept of cycA overexpression is in principle valid, the applicable substrate concentrations are too low to obtain commercially interesting yields. Since the yield of FV5069/pFV31/pACYC-cycA (Fig. 5b) was even lower than the control, we conclude that a benefit of cycA overexpression for pantothenic acid production is rather doubtful.