INTRODUCTION

Top Abstract Introduction Materials & Methods Results & Discussion References

Under oxic growth conditions, most Escherichia coli strains are not able to utilize citrate due to the lack of a functional transport system. This is a key characteristic of E. coli among enterobacteria (15). Some E. coli strains capable of aerobic growth on citrate possess plasmid-encoded citrate uptake systems. The citrate carrier genes from two of these plasmids were cloned and sequenced (13, 29). The deduced proteins, which consisted of 431 amino acids and differed in six positions only, exhibited 93% sequence identity to CitA from Salmonella typhimurium (30) and 66% to CitH from Klebsiella pneumoniae (35), both of which are chromosomally encoded. Citrate transport by CitH was shown to occur in symport with protons (36), and a similar mechanism is likely to apply for the plasmid-encoded CitA carriers from E. coli.

Under anoxic conditions, E. coli can utilize citrate if an oxidizable cosubstrate is present. The corresponding fermentation pathway is shown in Fig. 1. After uptake into the cell, citrate is split by citrate lyase to acetate and oxaloacetate. The latter is subsequently converted via malate and fumarate to succinate by malate dehydrogenase, fumarase, and fumarate reductase. The reducing equivalents required for this conversion must be provided by the oxidation of the cosubstrate, e.g., glucose or glycerol (17). Two enzymes must be specifically induced for anaerobic citrate dissimilation, i.e., a citrate uptake system and citrate lyase. The latter enzyme has been purified from E. coli and was shown to consist of three subunits ( [55.5 kDa],  [35 kDa], and  [12.5 kDa]), similar to citrate lyase from other bacterial species (25).

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Cosubstrate-dependent citrate fermentation by E. coli. The enzymes involved in the conversion of citrate to succinate are (1) citrate lyase, (2) malate dehydrogenase, (3) fumarase, and (4) fumarate reductase.

After the completion of the E. coli genome sequence (2), a gene cluster which showed a high degree of similarity to the citCDEFG operon of K. pneumoniae was identified between 13.9 and 14.2 min (Fig. 2). These genes encode the three subunits of citrate lyase (CitD, CitE, and CitF), a ligase required for the acetylation of the 2-(5"-phosphoribosyl)-3'-dephosphocoenzyme-A prosthetic group (CitC), and a protein (CitG) which presumably is involved in the biosynthesis or the covalent attachment of the prosthetic group (4). The proteins deduced from the E. coli genes citC, citD, citE, citF, and citG exhibited 50.6, 44.9, 65.1, 70.7, and 48.3% sequence identity to the corresponding K. pneumoniae proteins, respectively. A noticeable difference between the two gene clusters was the presence of an additional open reading frame (designated citX in Fig. 2) between the E. coli citF and citG genes. The citAB genes located upstream and divergent to E. coli citC encode proteins which are most closely related to the K. pneumoniae CitA-CitB two-component signal transduction system (42.0 and 48.7% identity, respectively). The K. pneumoniae CitA and CitB proteins are essential for expression of the genes specifically involved in citrate fermentation, including citCDEFG (5). The sensor kinase CitA was proposed to function as a citrate sensor (5), and the response regulator CitB was shown to bind to two sites extending from 50 to 96 upstream of the citC transcription start site and from 55 to 89 upstream of the citS transcription start site (20). Phosphorylation led to 10- to 100-fold increase of the apparent binding affinity (20). The E. coli citB gene has also been designated criR, because the deduced amino acid sequence is identical to the sequence of the CriR protein from Shigella flexneri, which has been implicated in the regulation of the ipa genes (24). Since E. coli does not possess an invasion plasmid carrying ipa genes, the primary function of the citAB gene products is presumably the regulation of the citrate lyase genes, as in K. pneumoniae. This supposition is supported by the similarity of the DNA-binding helix-turn-helix motifs of the two CitB response regulators and by the similarity of the citC upstream regions. With respect to the gene designations, one should be aware that in E. coli citA denotes both a plasmid-linked gene encoding a citrate carrier and a chromosomally located gene encoding a histidine sensor kinase. Similarly, citB denotes a gene associated with the plasmid-linked citA and also a chromosomally located gene encoding the cognate response regulator of the sensor kinase CitA.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   E. coli genes required for citrate fermentation and comparison with the corresponding K. pneumoniae genes. Genes shaded in dark gray represent those K. pneumoniae genes involved in citrate fermentation which are also present in E. coli; genes shaded in light gray are those present only in the E. coli or only in the K. pneumoniae cluster. All other indicated genes are presumably not directly involved in citrate fermentation.

An important difference between the K. pneumoniae and E. coli cit gene clusters is the lack of genes encoding a Na⁺-dependent citrate carrier (citS) and oxaloacetate decarboxylase (oadGAB) in the latter species (Fig. 2). In fact, these genes are not present on the whole E. coli chromosome. The absence of oxaloacetate decarboxylase provides an explanation for the different fermentation pathways in these organisms (3), and the absence of a CitS-type protein necessitates the use of a different citrate uptake system. Inspection of the E. coli DNA region downstream of citG revealed a gene (ybdS) encoding a highly hydrophobic protein with 34% sequence identity to the 2-oxoglutarate/malate translocator from spinach chloroplasts. The ybdS start codon is located only 50 bp downstream of the stop codon of citG, indicating that ybdS is cotranscribed with citCDEFXG. In this report, we present evidence that the protein encoded by ybdS functions as a citrate carrier; we therefore renamed the gene citT, for citrate transporter.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results & Discussion References

Bacterial strains and growth conditions. E. coli DH5 (Bethesda Research Laboratories) was routinely used as the host for the cloning procedures. E. coli JM83 (38) was used for the preparation of chromosomal DNA by the method of Marmur (18). E. coli BL21(DE3), which contains the phage T7 RNA polymerase gene under the control of the lacUV5 promoter (33), served as the host for the expression of citT and citS from pET-derived plasmids (Novagen). The strains were routinely grown at 37°C in Luria-Bertani (LB) medium (21) with shaking at 180 rpm. For testing the ability to grow with citrate as the sole carbon and energy source, we used either Simmons' citrate agar (31) supplemented with 12 µM thiamine or a liquid medium (adjusted to pH 6.8 with Na⁺-poor KOH) that contained 7 mM citric acid, 20 mM KH₂PO₄, 11 mM (NH₄)₂SO₄, 1 mM MgSO₄, 12 µM thiamine, 0.4% (vol/vol) trace elements solution (7), and either 25 mM Na₂SO₄ or 25 mM K₂SO₄. The latter medium contained less than 50 µM Na⁺ as determined by atomic absorption spectroscopy. The antibiotics ampicillin (100 to 200 µg/ml) and kanamycin (50 µg/ml) were used as appropriate. Cells used for growth studies with citrate minimal medium were pregrown in LB medium and washed with Na⁺-poor minimal medium before use as inoculum.

Recombinant DNA work. For routine work with recombinant DNA, established protocols were used (27). For the construction of a citT expression plasmid, the citT gene was amplified from chromosomal E. coli DNA by using the oligonucleotides ec-citT-for (5'-GATTCGAAGCTTCATATGTCTTTAGCAAAAGATAATATATGG-3') and ec-citT-rev (5'-CCGCGAATTCTTAGTTCCACATGGCGAGAATCGGCCAG-3'). In ec-citT-for, the ATG start codon of citT is part of an NdeI restriction site, which is preceded by a HindIII site and five additional nucleotides, allowing increased restriction efficiency. In ec-citT-rev, a BamHI restriction site is introduced after the citT stop codon. The PCR mixture contained 500 ng of genomic DNA of E. coli JM83, 0.5 µM each primer, 0.2 mM deoxynucleoside triphosphates, 1× buffer for cloned Pfu DNA polymerase, and 2.5 U of Pfu DNA polymerase (Stratagene). After an initial denaturation step (2 min at 95°C), 30 cycles consisting of 15 s at 95°C, 15 s at 62°C, and 4 min at 72°C were performed, followed by a terminal elongation step (4 min at 72°C). The complete PCR mix was subsequently separated on a 1% agarose gel, and the expected 1.49-kb fragment was isolated with Qiaex (Qiagen). After restriction with HindIII and BamHI, the PCR product was purified with a QIA-Quick spin column and ligated with HindIII/BamHI-restricted pUC19 (38), resulting in pUC19-citT. Since the citT gene in pUC19-citT was not preceded by a well-conserved ribosome binding site, a 1.46-kb NdeI/BamHI fragment from pUC19-citT was cloned in pET24b (Novagen) restricted with the same enzymes, resulting in pET24-citT. Plasmid pCitS_His-3 is a derivative of pET16b (Novagen) and is used for synthesis of the K. pneumoniae CitS citrate carrier modified with an N-terminal His₁₀ tag (23).

DNA sequence analysis. The sequence of the citT gene present in plasmid pET24-citT was determined by the dideoxynucleotide chain termination method (28), using the protocols and equipment for automated DNA sequencing (Sequencer 310 and PRISM Ready Reaction Dye-Deoxy terminator cycle sequencing kit from Applied Biosystems). For this purpose, pET24-citT was purified with a Qiagen Tip-500 column. A primer-walking strategy involving eight primers derived from citT and two primers derived from pET24b was applied. Computer-assisted DNA and protein sequence analysis was performed with the software package of the University of Wisconsin Genetics Computer Group. Prediction of the transmembrane helices indicated in Fig. 3 was performed with the TopPred II software (6) and the TMpred software (12).

Transport experiments. For transport experiments, E. coli BL21(DE3) transformed with either pET24b, pET24-citT, or pCitS_His-3 was grown in LB medium with appropriate antibiotics to an optical density at 600 nm (OD₆₀₀) of ca. 0.7 to 1.0. Subsequently, cells were washed once with 50 mM morpholineethanesulfonic acid-Tris buffer (pH 7.0) and concentrated 10-fold in the same buffer. The protein concentration of the resulting cell suspension was calculated by assuming that an OD₆₀₀ of 1.4 corresponds to 10⁹ cells/ml and those 10⁹ cells contain 150 µg of protein (21). At time zero, 98 µl of the concentrated cell suspension preincubated at 25°C was added to 2 µl of 1.2 mM [1,5-¹⁴C]citrate (145 cpm/pmol). After various times at 25°C, transport was terminated by the addition of 0.9 ml of ice-cold 0.1 M LiCl followed by rapid filtration through 0.45-µm-pore-size cellulose nitrate filters (diameter, 25 mm; Sartorius). The filters were washed once with 1 ml of ice-cold 0.1 M LiCl and then placed into scintillation vials. Immediately afterwards, 4 ml of scintillation fluid (Irga-Safe Plus; Packard) was added, and the entrapped [1,5-¹⁴C]citrate was determined by liquid scintillation counting. The values obtained in this way were corrected for a time zero value obtained as follows: 98 µl of cell suspension was first mixed with 0.9 ml of ice-cold 0.1 M LiCl and then applied to 2 µl of 1.2 mM [1,5-¹⁴C]citrate. This mixture was rapidly filtered as described above.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials & Methods Results & Discussion References

Aerobic growth of E. coli harboring a citT expression plasmid on citrate as the sole carbon and energy source. The open reading frame located at 13.9 min on the E. coli chromosome (designated citT), starting 50 bp downstream of the citG stop codon, encoded a protein of 487 amino acids (53.1 kDa) that was very hydrophobic and contained 12 putative transmembrane helices (Fig. 3). The physical proximity of citT to the citrate lyase genes and the fact that CitT showed 34% amino acid sequence identity to the 2-oxoglutarate/malate translocator from spinach chloroplasts (37) suggested to us that CitT might function as a citrate carrier. To test this assumption, the citT gene was amplified by PCR from chromosomal DNA of E. coli and ligated as a 1.5-kb HindIII/BamHI fragment in pUC19, which allows transcription of citT from the vector-encoded lac promoter. The ligation mixture was transformed into E. coli DH5 and plated on Simmons' citrate agar. After 48 h at 37°C, several citrate-positive colonies were identified by the color change of the agar from green to blue. This color change of the pH indicator bromthymol blue is observed only with cells able to utilize citrate as a carbon and energy source, which results in alkalinization of the medium. Cells unable to utilize citrate, such as E. coli DH5 containing only the vector pUC19, form only very small colonies (diameter, <1 mm), presumably by using residual carbon sources present in the agar, and bromthymol blue remains green. Restriction analysis of plasmid DNA isolated from several Cit⁺ clones showed that all contained pUC19 with the 1.5-kb HindIII/BamHI insert carrying citT. One of the pUC19-citT plasmids was transformed again into E. coli DH5 and plated on Simmons' citrate agar. In this case, all transformants were able to utilize citrate, confirming that citT is responsible for the Cit⁺ phenotype. Besides citrate, isocitrate could also be used as the sole carbon and energy source by E. coli DH5 harboring pUC19-citT.

View larger version (89K): [in this window] [in a new window]   FIG. 3.   Amino acid sequence alignment of the E. coli (Ec) CitT protein with the 2-oxoglutarate/malate translocator from spinach (Spinacia oleraceae [So]) chloroplasts and five eubacterial gene products. Identical residues present in at least four of the sequences are framed in black; conservative exchanges are framed in gray. Asterisks indicate residues conserved in all sequences. Putative transmembrane helices within the CitT sequence are overscored. Details on the aligned sequences can be found in the text and in Table 1. Hi, Haemophilus influenzae; Bs, Bacillus subtilis; Hp, Helicobacter pylori.

Transport studies with cell suspensions. The growth experiments described above confirmed our suggestion that CitT catalyzes the uptake of citrate. Further evidence was obtained by transport studies with cell suspensions. As shown in Fig. 4, cell suspensions of E. coli BL21(DE3)/pET24-citT catalyzed [1,5-¹⁴C]citrate uptake with an initial rate of 1.4 nmol min¹ (mg of protein)¹, whereas cells containing only the vector pET24b did not show any [1,5-¹⁴C]citrate accumulation. Citrate uptake by cells containing CitT was independent of Na⁺ ions in the concentration range tested (10 µM to 50 mM; data not shown). The similarity of CitT to the 2-oxoglutarate/malate translocator from spinach chloroplasts and the fact that the final product of citrate fermentation in E. coli is succinate led us to assume that CitT may function as a citrate/succinate antiporter. Indeed, the addition of 10 mM succinate to cells that previously had taken up [1,5-¹⁴C]citrate led to a complete efflux of the ¹⁴C label with an initial rate of 3.7 nmol min¹ (mg of protein)¹ (Fig. 4). The same effect was obtained by the addition of 10 mM citrate. Fumarate (10 mM) caused a comparable effect, but the rate was somewhat lower and efflux was not complete (data not shown). In a control experiment, uptake and efflux of [1,5-¹⁴C]citrate were analyzed with E. coli BL21(DE3)/pCitS_His-3. These cells contain the citrate carrier CitS from K. pneumoniae modified by an N-terminal His₁₀ tag (22, 23). Since the CitS carrier was shown to be highly specific for citrate as a substrate (1), we expected that only citrate, not succinate or fumarate, would be able to trigger an efflux of [1,5-¹⁴C]citrate previously taken up. As shown in Fig. 5, cells containing CitS catalyzed the uptake of citrate with an initial rate of 5.6 nmol min¹ (mg of protein)¹. Addition of 10 mM citrate led to a partial efflux of the ¹⁴C label from [1,5-¹⁴C]citrate-loaded cells [initial rate, 7.3 nmol min¹ (mg of protein)¹], whereas succinate or fumarate did not elicit such a response. This result confirms that the succinate- and fumarate-induced efflux observed with the E. coli/pET24-citT cells (Fig. 4) is catalyzed by the CitT protein rather than by other carriers present in the cytoplasmic membrane. The fact that only a partial efflux of the ¹⁴C label was observed in the experiment shown in Fig. 5 can be explained by the fact that part of [1,5-¹⁴C]citrate had been converted to other intermediates of the tricarboxylic acid cycle within the 30 s before addition of the unlabeled citrate. These other intermediates are transported not by CitS but apparently by CitT, as indicated by the complete efflux shown in Fig. 4.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Uptake and efflux of [1,5-14C]citrate by cell suspensions of E. coli BL21(DE3) containing either pET24b or pET24-citT. The transport experiments were performed as described in Materials and Methods.

View larger version (17K): [in this window] [in a new window]   FIG. 5.   Uptake and efflux of [1,5-14C]citrate by cell suspensions of E. coli BL21(DE3) containing pCitSHis-3. The transport experiments were performed as described in Materials and Methods.

1

1

1

1

1

1

View larger version (19K): [in this window] [in a new window]   FIG. 6.   Uptake and efflux of DL-[1,4-14C]tartrate by cell suspensions of E. coli BL21(DE3) containing either pET24b or pET24-citT. The transport experiments were performed as described in Materials and Methods.

Proteins related to CitT. A search for proteins related to CitT from E. coli led to the identification of five bacterial polypeptides, none of which has been functionally characterized hitherto, and the 2-oxoglutarate/malate translocator (SOT1) from Spinacia oleraceae (37). The SOT1 protein is located in the inner envelope membrane of spinach chloroplasts, where it catalyzes the import of 2-oxoglutarate in exchange for stromal malate. The protein has recently been purified and functionally reconstituted into liposomes (19). Besides malate, succinate, fumarate, and 2-oxoglutarate can be used as counterions (37). From the sequence alignment shown in Fig. 3, it is obvious that the chloroplast protein is related to CitT and the other bacterial proteins described below. Moreover, preliminary characterization of the catalytic properties of CitT indicates that the transport mechanism of this carrier is similar to that of the SOT1 protein.