Cloning, Sequencing, and Expression in Escherichia coli of OxlT, the Oxalate:Formate Exchange Protein of Oxalobacter formigenes *

OxlT is the oxalate/formate exchange protein that rep- resents the vectorial component of a proton-motive metabolic cycle in Oxalobacter formigenes . Here we report the cloning and sequencing of OxlT and describe its expression in Escherichia coli . The OxlT amino acid se- quence specifies a polytopic hydrophobic protein of 418 residues with a mass of 44,128 daltons. Analysis of hy- dropathy and consideration of the distribution of charged residues suggests an OxlT secondary structure having 12 transmembrane segments, oriented so that the N and C termini face the cytoplasm. Expression of OxlT in E. coli coincides with appearance of a capacity to carry out the self-exchange of oxalate and the heter- ologous, electrogenic exchange of oxalate with formate. The unusually high velocity of OxlT-mediated transport is also preserved in E. coli . We conclude that the essen-tial features of OxlT are retained on its expression in E. coli . The Gram-negative anaerobe, Oxalobacter formigenes , de-rives metabolic energy from the decarboxylation of oxalate (1, 2) by using a “proton-motive metabolic cycle” (3, 4). In O. formigenes , which provided the first case study of such a pro-ton-motive cycle (3, 4), entry of divalent oxalate is coupled to the exit of its decarboxylation product, monovalent formate, leading to formation of an internally negative membrane potential. Because intracellular oxalate decarboxylation con-sumes a cytosolic proton, entry of negative charge is accompa-nied in stoichiometric fashion by appearance of internal hydroxyl ion. As a result, the combined activities of the vectorial

The Gram-negative anaerobe, Oxalobacter formigenes, derives metabolic energy from the decarboxylation of oxalate (1, 2) by using a "proton-motive metabolic cycle" (3,4). In O. formigenes, which provided the first case study of such a proton-motive cycle (3,4), entry of divalent oxalate is coupled to the exit of its decarboxylation product, monovalent formate, leading to formation of an internally negative membrane potential. Because intracellular oxalate decarboxylation consumes a cytosolic proton, entry of negative charge is accompanied in stoichiometric fashion by appearance of internal hydroxyl ion. As a result, the combined activities of the vectorial antiport reaction and the scalar decarboxylation step comprise a thermodynamic proton pump (3,4). In this way, O. formigenes establishes the proton-motive force required for both the synthesis of ATP by reversal of a dicyclohexylcarbodiimide-sensitive ATPase (29) and for the support of other membrane reactions requiring a proton-motive force.
Early experiments based on reconstitution of activity from crude detergent extracts suggested the oxalate/formate exchange reaction is mediated by a membrane carrier (3). This reasoning was strengthened by finding that oxalate transport is catalyzed by a single protein, OxlT, whose SDS-PAGE 1 mobility (ϳ38 kDa) resembles that of other bacterial carrier proteins (5). However, it was not possible to complete the argument by examination of the OxlT amino acid sequence. For this reason, the work described here was directed to the cloning and sequencing of OxlT. An additional objective was to determine whether OxlT function is retained after expression in Escherichia coli. If so, future studies of this unusual antiporter could exploit the advantages of a genetically tractable host. The work described here indicates that the amino acid sequence of OxlT conforms to the general pattern found for most membrane carriers, including the presence of twelve likely transmembrane segments. Functional studies further suggest that the main properties of OxlT, including its exceptionally high velocity (3,5), are preserved on its expression in E. coli.

EXPERIMENTAL PROCEDURES
Cells and Plasmids-E. coli strain KW251 (Promega) was used for the screening of an O. formigenes lambda phage library; subcloning of positive restriction fragments was performed using pBluescript II KSϪ (Amp r ) carried in strain XL1 blue (Tet r ) (Stratagene). Strain XL1 blue harboring pMS421 (spec r , LacI q ) was called strain XL3 and was used for expression of OxlT from pBKOxlTSKϩ, a pBluescript II SKϩ derivative in which the gene encoding OxlT is under control of the lac promoter. Cells were grown aerobically at 37°C in Luria Broth with drugs as required (100 g/ml carbenicillin, 12 g/ml tetracycline, and 50 g/ml spectinomycin).
O. formigenes Genomic DNA-Cells of O. formigenes from Dr. M. J. Allison (National Animal Disease Center, Ames IA) were the source of genomic DNA used in preliminary hybridization experiments. DNA was extracted using the Easy DNA extraction kit of Invitrogen.
Determination of the OxlT N-Terminal Sequence and Preparation of a Peptide-directed Antibody-OxlT was purified as described (5). After removing lipid (5) from the peak activity fraction appearing on CM-Sepharose chromatography, 100 g of purified OxlT was subjected to SDS-PAGE and transferred at 4°C to an Applied Biosystem ProBlott polyvinylidene difluoride membrane at 100 V for 1 h, using a transfer solution containing 25 mM Tris, 10 mM glycine, and 0.5 mM dithiothreitol. The membrane with adsorbed OxlT was washed four times with distilled water and provided to the Harvard Microchemistry Facility (Cambridge, MA), which reported NNPQTGQSTGLLGNRWFYLV (single-letter amino acid code) as the probable N-terminal sequence; there was indication of a ragged N terminus. A synthetic peptide of this same sequence was synthesized by the Peptide Core Facility (Department of Biological Chemistry, Johns Hopkins Medical School). After conjugation of the peptide to bovine serum albumin (6), rabbit polyclonal antibody was raised against the material by Hazelton Research Products (Denver, PA).
Oligonucleotide Probes-Based on the N-terminal amino acid sequence noted above, we prepared two degenerate oligonucleotide probes. Oligo1 had the nucleotide sequence AA(C/T)AA(C/T)CCICA(A/ G)ACIGGICA (where I indicates inosine), corresponding to amino acid residues 1-7 (NNPQTGQ); Oligo2 had the sequence AA(C/T)(A/C)GIT-GGTT(C/T)TA(C/T)(C/T)T and corresponded to residues 14 -19 (NRW-* These studies were supported by Research Grant MCB-9220823 from the National Science Foundation and United States Public Health Service Grant GM24195 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  FYL). In preliminary work, Southern hybridization at 42°C with Oligo2 gave an unique hybridization band using O. formigenes genomic DNA digested completely by EcoRI, HindIII, or PstI. Oligo1 showed this same pattern but also hybridized to the -HindIII markers. Because the O. formigenes library was housed in lambda phage, we used Oligo2 for our initial screens.
Cloning the Gene Encoding OxlT-Lung et al. (7) had constructed an O. formigenes genomic DNA library using the Promega -GEM11 XhoI half-site arm vector. We screened this library with Oligo2 using the general procedures outlined by Sambrook et al. (8), observing a positive clone for every 3,000 -5000 plaques. After three rounds of plaque purification, insert fragments of 10 -16 kb were identified following SacI digestion. SacI fragments were further digested with PstI and subcloned to eventually yield a 3.2-kb PstI-SacI fragment that showed hybridization to both Oligo1 and Oligo2. This positive fragment was placed in pBluescript II KSϪ, giving pBKOxlTKSϪ, and a nested deletion series was prepared for sequencing. After identification of the gene encoding OxlT, the DNA sequence in this region was confirmed by second strand sequencing using plasmids selected from this same deletion series. Site-directed mutagenesis (9, 10) was then used to introduce an XbaI site 23-base pairs upstream from the likely OxlT start site, ATG (see Fig. 2), producing pBKOxlT-Xb. pBKOxlT-Xb was subsequently digested with XbaI plus HindIII to give a 1.4-kb fragment containing the complete gene encoding OxlT. This 1.4-kb fragment was ligated into the XbaI-HindIII site of pBluescript II SKϩ to generate pBKOxlTSKϩ, where expression of OxlT was regulated by the lac promoter. As a final step, pBKOxlTSKϩ was placed in E. coli XL3 for functional tests.
Sequencing-Double-stranded DNA was sequenced by the DNA Core Facility of the Johns Hopkins Medical School, using the dideoxy chain termination procedure of Sanger et al. (11). The nested deletion series was sequenced using universal primers for pBluescript II KSϪ; as primers for sequencing the oxlT opposite strand, we designed appropriate complementary synthetic oligonucleotides.
Expression of OxlT in E. coli-An overnight preculture of E. coli XL3 carrying pBKOxlTSKϩ was diluted 100-fold in fresh medium, and 1 mM IPTG was added 1 h later; IPTG-induced cells and uninduced control cells were harvested after an additional 4 h of growth.
SDS-PAGE and Immunoblots-SDS-PAGE with 12% acrylamide was performed as outlined by Laemmli (12). For routine immunoblots, protein was transferred to nitrocellulose using standard techniques (13), and after exposure to immune serum diluted 1/2500, binding of the primary antibody was detected by chemiluminescence (Amersham Corp.) (14).
Solubilization and Reconstitution of OxlT and Assays of Transport-IPTG-induced cells and uninduced control cells (each 5 mg of protein) were harvested by centrifugation, resuspended in 5 ml of lysozyme (300 g/ml) and DNase (40 g/ml), and incubated at 37°C for 10 min. Cells were repelleted and then resuspended in 5 ml of water. The resulting ghosts were spun down and resuspended in 0.5 ml of ice-cold solubilization solution (15) (25 mM MOPS/K, 20% (v/v) glycerol, 0.4% acetone/ ether purified E. coli phospholipid, 1 mM dithiothreitol, 1.25% octyl-␤-D-glucoside, 0.75 mM phenylmethylsulfonyl fluoride, and 10 mM oxalate). After incubation at 4°C for 20 min, the suspension was clarified by centrifugation at 4°C in an Eppendorf refrigerated microfuge (15,000 ϫ g for 15 min) to give a crude detergent extract that was stored at Ϫ80C until use.
OxlT transport activity was monitored by reconstitution of protein into proteoliposomes (3,15,16). In a final volume of 250 l, 50 -100 l of a detergent extract was mixed with 1.36 mg of bath-sonicated liposomes, additional detergent (to 1.25%), and either 50 mM MOPS/K or 50 mM MOPS/NMG (pH 7). After incubation at 4°C for 20 min, proteoliposomes were formed at 23°C by the addition of 5 ml of a dilution and loading buffer (pH 7). For estimates of oxalate self-exchange (Table I), the loading buffer contained 100 mM potassium oxalate, 50 mM MOPS/K, and 1 mM dithiothreitol. To assess oxalate/formate exchange (see Fig. 5), the loading buffer was either 100 mM potassium formate or 100 mM NMG formate, along with 50 mM MOPS/K or 50 mM MOPS/ NMG and 1 mM dithiothreitol. Formation of proteoliposomes was complete within 20 min, at which point we used one of two protocols to assess OxlT activity. In a rapid filtration assay (17) to monitor oxalate self-exchange (Table I), 0.2 ml of the proteoliposomal suspension was applied directly, under vacuum, to the center of a 0.22-m GSTF Millipore filter. The external medium was removed by two 5-ml rinses with assay buffer (100 mM K 2 SO 4 and 50 mM MOPS/K, pH 7), and on release of the vacuum the assay began as proteoliposomes were covered with 0.25 ml of assay buffer containing 100 M [ 14 C]oxalate. The reaction was terminated 3 min later by filtration and three quick rinses with assay buffer. Alternatively (see Fig. 5), formate-loaded proteoliposomes were isolated by centrifugation (16) and resuspended in a small volume of their K-or NMG-based loading buffers. Subsequently, they were diluted 120-fold into either NMG-or K-based assay buffers, as above, containing 100 M [ 14 C]oxalate, with or without 1 M valinomycin. In this way, it was possible to generate a membrane potential whose polarity was either interior positive (potassium outside, NMG inside) or interior negative (NMG outside, potassium inside). As an additional basis for comparison, proteoliposomes were loaded with NMG-formate and tested using the NMG-based assay buffer.
Protein Estimation-Protein content was estimated using a modification of the procedure of Schaffner and Weissman (18).

Cloning of the Gene Specifying
OxlT-In the absence of a positive selection procedure, we cloned OxlT by obtaining Nterminal sequence information from the purified protein and designing appropriate oligonucleotide probes with which to screen an O. formigenes library (see "Experimental Procedures"). Initial screening of a -GEM library gave positive inserts of 10 -16 kb; subcloning eventually yielded a 3.2-kb fragment whose sequence included two overlapping open reading frames on opposite DNA strands of a 1.4-kb interval. One of these open reading frames specified a hydrophobic protein whose N-terminal sequence matched that determined for purified OxlT (Fig. 1), and we tentatively identified this gene as our desired target, oxlT; later functional tests (below) verified the assignment. The oxlT sequence has been submitted to Gen-Bank (accession number U40075).
The DNA sequence of the gene, oxlT, indicates that expression of its encoded protein (OxlT) follows patterns well established for bacterial systems. Thus, a likely promoter having Ϫ35 and Ϫ10 sequences of TTGAAA and TTCAAT, respectively, occupies a 29-base interval ending 70 nucleotides upstream of the initiating codon, AUG. Transcriptional termination is probably mediated by a 31-base stem-loop structure (AAAAAAGCCCGGCTTTCCGCCGGGCTTTTTT) that begins 72 nucleotides from the first of two in-frame stop (UAA) codons.
Characteristics of the Cloned Protein-Analysis of the deduced OxlT amino acid sequence (Fig. 2) reveals a novel hydrophobic protein of 418 amino acid residues having a predicted mass of 44,128 daltons. No proteins with significant homology to OxlT were found in a EMBL BLITZ search of the Swiss Protein Data Base using the Smith and Waterman (19) algorithm; similarly, we found no proteins related to the hypothetical hydrophilic protein specified on the OxlT noncoding strand.
Analysis of OxlT hydropathy according to the method of Kyte and Doolittle (20) (Fig. 3) suggests the presence of 12 hydrophobic segments, each of sufficient length to constitute a transmembrane ␣-helix (TM1-12). A similar analysis according to Rost et al. (21) predicts 11 transmembrane ␣-helices, including TM1 and TM3-12 ( Fig. 3) but excluding TM2, whose peak hydropathy value is the lowest of the 12 putative transmembrane segments (Fig. 3). Although membrane carriers with 11 transmembrane segments have been described in bacteria (22,23), it is more typical to find examples with 10 or 12 transmembrane regions (4, 23, 24). For this reason, our initial model of OxlT topology (Fig. 3) assumes the 12 transmembrane segments suggested by analysis of hydropathy. This initial model FIG. 1. OxlT N-terminal sequences. The N-terminal sequence determined by microsequencing of purified OxlT (top) is compared with the N-terminal sequence specified by the cloned gene, oxlT (bottom). also conforms to the common finding (4, 23, 24) of a central cytoplasmic loop that separates the regions containing TM1-6 and TM7-12.
To orient the proposed OxlT structure with respect to cytoplasmic and extracellular phases, we used the observation of von Heijne (25) that transmembrane segments often have an excess of positively charged residues at their cytoplasmic ends, especially in bacterial systems. It is evident that in our proposed structure (Fig. 3, top), charged residues are assigned to either the extracellular (net charge of Ϫ1) or cytoplasmic (net charge of ϩ13) surfaces, with the exception of the single lysine residue (Lys 355 ) that appears within TM11 (Fig. 3).
Expression of OxlT in E. coli-To determine whether the gene tentatively identified as oxlT specifies the OxlT transport protein, we constructed a vector (pBKOxlTSKϩ) that brings protein expression under control of the lac promoter. Expression of the encoded protein and assays of its function were then carried out in XL3, a strain also carrying a middle copy compatible plasmid (pMS421) encoding the gene for LacI q . This gave strong repression in the absence of IPTG and allowed us to propagate pBKOxlTSKϩ without selective pressures that might accompany constitutive or otherwise unregulated protein expression.
The experiments described in Figs. 4 and 5 document that the gene identified as oxlT specifies the OxlT transport protein and that the main features of OxlT function are retained in E. coli. Thus, antibody directed against the OxlT N terminus reported expression of OxlT in IPTG-induced cells carrying pBKOxlTSKϩ but not in uninduced cells or in cells carrying the parent pBluescript II SKϩ (with or without IPTG) (Fig. 4). It is also evident that SDS-PAGE profile of OxlT when expressed in E. coli resembles that of authentic OxlT, from O. formigenes, including the presence of both monomeric (ϳ35 kDa) and dimeric (ϳ65 kDa) forms of the protein (Fig. 4) (5).
Equally important, in this same experiment we showed that appearance of OxlT immunoreactivity coincides with acquisition by induced cells of a capacity to catalyze both the oxalate self-exchange reaction and the electrogenic exchange of oxalate and formate. For such functional tests, we prepared detergent extracts from both induced and uninduced cells (Fig. 4). To examine oxalate self-exchange, oxalate-loaded proteoliposomes were washed free of external substrate by filtration on Millipore filters (0.22-m pore size), and then, while still affixed to the filters, they were covered for 3 min with an assay medium containing 100 M [ 14 C]oxalate, followed by a final filtration and wash. This test (Table I) gave no indication of oxalate transport by cells bearing pBluescript II SKϩ (Ϯ IPTG) (0.02 mol/mg protein). By contrast, uninduced cells with pBKOx-lTSKϩ displayed a low but significantly positive signal (0.14 mol/mg protein), whereas IPTG induction led to markedly increased accumulation of label (2.3 mol/mg protein) ( Table I).
The detergent extract from IPTG-induced cells was also used to prepare proteoliposomes loaded with the potassium or NMG salts of formate so as to monitor the exchange of oxalate with formate (Fig. 5). The particles were diluted into media containing NMG sulfate or potassium sulfate, respectively, so that addition of valinomycin established an electrical gradient, internally negative or positive in polarity. Such trials were un-  Cloning, Sequencing, and Expression of OxlT in E. coli ambiguous in their findings: imposition of an internally positive electrical potential strongly stimulated the oxalate transport observed in controls not treated with the ionophore, whereas imposition of an internally negative potential completely inhibited the reaction. Proteoliposomes prepared and assayed in the absence of potassium were unaffected by valinomycin and showed oxalate transport virtually identical to that found for the potassium-or NMG-loaded proteoliposomes not exposed to valinomycin (Fig. 5 legend). Similar findings had been reported earlier (3,5) for OxlT reconstituted from O.
formigenes. In both instances, the pattern of response indicates that the exchange of oxalate and formate is electrogenic, with negative charge moving in parallel with oxalate. Because the pK a 's for oxalate are 1.23 and 3.83, the simplest model is that the OxlT transporter, whether expressed in O. formigenes or E. coli, mediates exchange of divalent oxalate and monovalent formate.

DISCUSSION
The work summarized here had as its main goal the cloning and sequencing of OxlT, the oxalate/formate antiport protein of O. formigenes. Several criteria show this goal has been met. In particular, the cloned gene specifies the N-terminal sequence found in authentic OxlT (Fig. 1), and expression of this gene confers upon E. coli the capacity to mediate both the homologous self-exchange of oxalate and the heterologous, electrogenic exchange of oxalate with formate (Table I and Fig. 5). We therefore conclude that this antiport protein retains its most important functional properties when expressed in E. coli. It is likely the main physical characteristics of OxlT are also preserved in E. coli, because the OxlT SDS-PAGE profiles in E. coli and O. formigenes are equivalent (Fig. 4) and because the positive response to an N-terminal peptide-directed antibody suggests OxlT retains its natural N terminus (Fig. 4).
Analysis of the OxlT amino acid sequence reveals a polytopic hydrophobic protein (Fig. 3) whose general structure resembles that of known membrane carriers in the several respects (4,22,24): (i) the presence of 12 (or 11) presumed transmembrane segments; (ii) N-and C-terminal regions facing the cytoplasm (presuming an even number of transmembrane segments); (iii) the finding of a cytoplasmic loop midway along the sequence (residues 190 -219), separating the region containing TM1-6 from that containing TM7-12; and (iv) an excess of positively charged residues at the presumed cytoplasmic surface. Although there is no apparent sequence homology between OxlT and known membrane carriers (or other transporters), these general features, along with the earlier biochemical characterization, are sufficient to classify OxlT as a conventional secondary transport protein.
The OxlT predicted structure has two additional features deserving of specific comment. First, we note the presence of a single charged residue (Lys 355 ) within TM11 (Fig. 3, top). Because OxlT substrates are anionic (oxalate 2Ϫ and formate 1Ϫ ), the presence of this apparently uncompensated positive charge in the hydrophobic sector prompts the hypothesis that Lys 355 forms part of an anionic binding center within the substrate translocation pathway. Preliminary tests are compatible with this idea, because several uncharged substitutions at position 355 give variants that fail to transport, whereas the K355R derivative retains activity. 2 A second finding of interest is that OxlT has only two cysteine residues (Cys 28 and Cys 271 ). Because neither of these cysteines is required for function, 2 OxlT presents an attractive target for cysteine scanning mutagenesis, an approach that has proven valuable to the study of several membrane transport systems (26 -28).
Evaluation of oxalate transport (Table I and Fig. 5) supports the idea that the main features of OxlT selectivity are retained in E. coli. Moreover, calculations using these data suggest that the unusually high velocity of OxlT is also preserved in this expression system. Thus, detergent extracts from induced E. coli yielded a potential-stimulated oxalate/formate antiport rate of 24 mol/min/mg protein (Fig. 5), whereas for the same conditions we found an exchange rate of 16 mol/min/mg protein using O. formigenes (3). And in further work (not given), we found the kinetic parameters of oxalate self-exchange to be 2 K. Abe and P. C. Maloney, unpublished results.  Table I) were harvested. Expression of OxlT was monitored by an immunoblot using antibody directed against the OxlT N terminus (see "Experimental Procedures").  (Fig. 4 legend) was used to prepare proteoliposomes (or liposomes) loaded with potassium formate or NMG formate, as described under "Experimental Procedures." To begin the reaction, proteoliposomes (or liposomes) were diluted into NMG-or potassium-based assay medium (as shown) containing 100 M [ 14 C]oxalate with 1 M valinomycin or the equivalent amount of carrier ethanol as shown. Samples were taken for filtration and washing at the indicated times. The presence of external potassium (K out ) or internal potassium (K in ) is indicated on the graph. Proteoliposomes (and liposomes) loaded and assayed using only NMG-based solutions were also tested, but these data have been omitted for clarity. Transport by these particles (with or without valinomycin) was essentially identical in rate and extent to the ionophore-untreated controls shown here.