Cloning and Characterization of Helicobacter pyloriSuccinyl CoA:Acetoacetate CoA-transferase, a Novel Prokaryotic Member of the CoA-transferase Family*

Sequencing of a fragment of Helicobacter pylori genome led to the identification of two open reading frames showing striking homology with Coenzyme A (CoA) transferases, enzymes catalyzing the reversible transfer of CoA from one carboxylic acid to another. The genes were present in all H. pyloristrains tested by polymerase chain reaction or slot blotting but not inCampylobacter jejuni. Genes for the putative A and B subunits of H. pylori CoA-transferase were introduced into the bacterial expression vector pKK223-3 and expressed inEscherichia coli JM105 cells. Amino acid sequence comparisons, combined with measurements of enzyme activities using different CoA donors and acceptors, identified the H. pylori CoA-transferase as a succinyl CoA:acetoacetate CoA-transferase. This activity was consistently observed in differentH. pylori strains. Antibodies raised against either recombinant A or B subunits recognized two distinct subunits ofM r ∼26,000 and 24,000 that are both necessary for H. pylori CoA-transferase function. The lack of α-ketoglutarate dehydrogenase and of succinyl CoA synthetase activities indicates that the generation of succinyl CoA is not mediated by the tricarboxylic acid cycle in H. pylori. We postulate the existence of an alternative pathway where the CoA-transferase is essential for energy metabolism.

Helicobacter pylori is a Gram-negative, microaerophilic bacterium that establishes long term chronic infections in the human stomach. It has been shown to lead to gastritis, peptic ulcers, and cancer (1)(2)(3)(4). However, despite the importance of H. pylori as a pathogen, little is known about its metabolic and biosynthetic pathways. Information about the metabolism and substrate utilization by H. pylori is essential to understand bacterial colonization and survival and to design new drugs to treat the infection it causes.
The acetate (butyrate)-acetoacetate CoA-transferase (EC 2.8.3.9) found among Clostridia acts mainly to detoxify the medium by removing the acetate and butyrate excreted earlier in the fermentation. This enzyme therefore has a role fundamentally different from other CoA-transferases, usually involved in the uptake of substrates for energy and structural use (6).
The succinyl CoA:3-oxoacid CoA-transferase (SCOT, EC 2.8.3.5) is responsible for the formation of acetoacetyl CoA by transfer of a CoA moiety from succinyl CoA to a 3-oxoacid, usually acetoacetate (23). This enzyme has been found to have the highest activity in heart and kidney of various mammals (18,24). In the mitochondrion of these tissues, acetoacetate is converted to acetoacetyl CoA, which is further broken down to two acetyl CoA molecules capable of entering the tricarboxylic acid cycle. Therefore, this enzyme plays a crucial role in ketone body metabolism (18) as exemplified by the inborn error in humans (25).
In this study, we report the cloning and characterization of the two subunits of a CoA-transferase from H. pylori. A high degree of homology was found by comparing the deduced amino acid sequences from the two H. pylori genes with the pig and human CoA-transferases as well as with different bacterial CoA-transferases. These comparisons combined with measurements of enzyme activities identify H. pylori CoA-transferase as a succinyl CoA:acetoacetate CoA-transferase. Our data show that the bacterial enzyme is a dimeric protein in contrast to the monomeric eukaryotic homologues and indicate that it is rather a unique feature of H. pylori to harbor such an enzy-matic activity, which is not present in Escherichia coli or Campylobacter jejuni.

EXPERIMENTAL PROCEDURES
Bacterial Strains-The H. pylori strains are strain 69A (Department of Medical Microbiology, Amsterdam University, The Netherlands) isolated from a patient with nonulcer dyspepsia, strain 888 -0 (Department of Medical Microbiology and Immunology, Hamburg University, Germany) isolated from a patient with a duodenal ulcer, strain NCTC 11637, and strains Ly2, Ly3, and Ly13 (Division of Gastroenterology, CHUV Lausanne) isolated from patients with a family history of gastric cancer.
Media and Cell Culture Conditions-H. pylori strains were grown in solid medium containing 3.6% GC agar base (Life Technologies, Inc.) supplemented with 1% Isovitale X (Baltimore Biological Laboratories) and 10% donor horse serum (Biological Industries, Kibbutz Beth Haemek, Israel) and maintained in a microaerophilic atmosphere (85% N 2 /10% CO 2 /5% O 2 ) at 37°C for 2 days. C. jejuni and H. felis were cultivated in blood-agar plates and maintained in the same conditions. E. coli strains were grown in Luria-Bertani (LB) medium at 30 or 37°C. Solid medium was prepared by addition of 1.5% bacto-agar (Difco). The antibiotics used were: 50 g/ml ampicillin, 12.5 g/ml tetracyclin, and 40 g/ml kanamycin (Sigma).
Genomic DNA Isolation-Bacterial DNA was prepared according to the method described by Hua et al. (26).
Amplification of H. pylori DNA by the Polymerase Chain Reaction-Southern blotting and screening of H. pylori DNA library (see below) were performed with oligonucleotides ICT 14 (5Ј-GATAAAACCG-GCACC-3Ј) and ICT 20 (5Ј-GCGGGCGCGTCGTT-3Ј). Optimal PCR conditions were established using the PCR optimization kit (Boehringer Mannheim) to yield a 1000-bp DNA fragment. PCR was carried out in 50 l containing 500 ng of H. pylori genomic DNA, 50 pmol of each primer, 200 M of each dNTP (Boehringer Mannheim), and a 28:1 mixture of TaqStart antibody (CLONTECH) ϩ Taq DNA polymerase (Boehringer Mannheim) (final concentrations, 56 and 2 pM, respectively), in 10 mM Tris-HCl, pH 9.2, 50 mM KCl, and 1.5 mM MgCl 2 . The cycling program was one cycle consisting of denaturation at 94°C for 3 min; hybridation at 50°C for 2 min; extension at 72°C for 3 min followed by 35 cycles at 94°C for 30 s; 50°C for 30 s; 72°C for 1 min and one cycle of 94°C for 20 s; 50°C for 20 s; 72°C for 5 min in the microprocessor controlled incubation system Crocodile III from Appligene. For subcloning in the XbaI and ClaI sites of pBluescript KS Ϫ (Stratagene), the oligonucleotides used were ICT 21 (5Ј-GCTCTAGAG-CGATAAAACCGGCACC-3Ј) and ICT 22 (5Ј-CCATCGATG-GGCGGGCGCGTCGTT-3Ј).
To reconfirm the sequences, the putative open reading frames (ORFs) of the H. pylori CoA-transferase subunits were subcloned in the XbaI and ClaI sites of pBluescript KS Ϫ using as primers the oligonucleotides: ICT 22 (see above) and ICT 32 (5Ј-GCTCTAGAGCCTCT-CATTTCGCGCTCCTTGTCG) for the subunit A and ICT 31 (5Ј-CCATCGATATCACGACAAGGAGCGCGAAATGA-3Ј) and ICT 33 (5Ј-GCTCTAGAGCGCTATAGGTGCACTTCAAATTCGG-3Ј) for the subunit B. The PCR amplification of H. pylori CoA-transferase gene with the primers ICT 22 and ICT 33 were used as a probe for slot blotting.
To perform the CoA-transferase activity assays (see below), the putative ORF1 (705 bp), ORF2 (624 bp), or ORF1 and ORF2 (1322 bp) were obtained by PCR and inserted into the EcoRI and PstI sites of pKK223-3 (Pharmacia) using as primers ICT 42 (5Ј-GGAATTCATGA-ACAAGGTTATAACCG-3Ј) and ICT 93 (GGAATTCTGCAGCTCTCAT-TTCGCGCTCCTTGTCG-3Ј) for the A subunit, ICT 94 (5Ј-GGAATTC-ATGAGAGAGGCTATCATTAAAAG-3Ј) and ICT 43 (5Ј-GGAATTCTG-CAGCTATAGGTGCACTTCAAATTCG-3Ј) for the B subunit, and ICT 42 and ICT 43 for the A and B subunits. The reaction was carried out under the same conditions except that 2.5 units of Taq DNA polymerase alone was used and that the pH was set at 8.3. All PCR products were purified using the Geneclean kit (Bio 101, Inc.).
Screening of H. pylori lambda ZAP II Library-The lambda ZAP® II custom H. pylori genomic library (Stratagene) was titered and screened according to the supplier's protocols except that probes were labeled with fluorescein-dUTP (see below). Fourteen positive clones were isolated from the screening of 10 6 phages, and the corresponding pBluescript SK Ϫ plasmids were excised from the lambda ZAP II using the ExAssist/SORL system according to the manufacturer's instructions (Stratagene).
Characterization of H. pylori DNA Clones-One of the positive clones, pGB1, was deleted using the Exo III/Mung Bean nuclease deletion kit (Stratagene) (27). Two different combinations of restriction enzymes (SacI-XbaI and ApaI-ClaI) were used to create 3Ј-and 5Јoverhanging ends for nested deletions from the T3 or T7 promoter elements of pBluescript SK Ϫ . The sequential deletions were generated at intervals of 30 s at 34°C. After blunt ligation and transformation, sizes of different species obtained were determined by digestion with XhoI or BamHI, respectively.
Sequencing-All plasmid DNAs for sequencing were purified using Qiagen tips according to the manufacturer's instructions. 5 g of each plasmid were sequenced on an automatic ALF sequencer using fluoresceinated universal and reverse primers combined with dideoxynucleotides as recommended by the manufacturer (Pharmacia).
Detection of CoA-transferase Genes in Bacteria and PCR Detection-The presence of the operon including the A and B subunits of H. pylori CoA-transferase was detected by PCR on isolated bacterial genomic DNA using a combination of primers ICT 45 (5Ј-CGGGATCCCG-ATGAACAAGGTTATAACCG-3Ј) and ICT 48 (5Ј-GGAATTCGTC-GACGCTATAGGTGCACTTCAAATTCG-3Ј).
DNA Slot Blotting-Prior to application to nitrocellulose membranes (Bio-Rad), 2 g of genomic DNA samples were denatured in 0.4 M NaOH, 10 mM EDTA, pH 8.0, then heated at 100°C for 10 min, and neutralized by addition of an equal volume of cold 2 M ammonium acetate, pH 7.0. Treatment of the membranes and vacuum filtration of denatured DNA samples were performed according to the manufacturer's instructions (Bio-Dot® SF Microfiltration, Bio-Rad). After filtration, DNA was cross-linked to the membranes (UV Crosslinker; Hoefer Scientific Instruments).
Southern Blotting-HaeIII digestions of 10 g of bacterial genomic DNA were run in 0.7% agarose gels in 0.5 ϫ 90 mM Tris borate, 2 mM EDTA buffer and transferred to nylon membranes (Boehringer Mannheim) using a semi-dry electrophoresis transfer cell (Trans-Blot® SD, Bio-Rad) according to the manufacturer's instructions (Bio-Rad). After denaturation, DNA was cross-linked to the membranes as above.
Probe Labeling, Hybridization, and Chemoluminescence (ECL) Detection Conditions-The different PCR fragments used as probes were labeled with fluorescein-dUTP using the random prime labeling system from Amersham Corp. Membranes were prehybridized for at least 30 min in DIG Easy Hyb solution (Boehringer Mannheim). After addition of the labeled probe and overnight hybridization at room temperature or at 37°C, membranes were washed at different stringency conditions (twice in 2 ϫ SSC, 0.1% SDS for 5 min at 21°C and twice in 0.1 ϫ SSC, 0.1% SDS for 15 min at the indicated temperature) with constant agitation. Blocking of the membrane background and ECL detection were performed according to the manufacturer's instructions (ECL detection system, version II; Amersham Corp.). Exposure times of all membranes to x-ray films (X-OmatAR, Kodak) were chosen to visually optimize the chemoluminescent signals.
Preparation of Antibodies against H. pylori Gene Products-The two H. pylori putative ORFs were subcloned separately into pQE11 (Qiagen) with primers ICT 45 (see above) and ICT 46 (5Ј-GGAATTCGTC-GACTCTCATTTCGCGCTCCTTGTCG-3Ј) for the A subunit and ICT 47 (5Ј-CGGGATCCCGATGAGAGAGGCTATCATTAAAAG-3Ј) and ICT 48 (see above) for the B subunit. The histidine tagged-fusion proteins were isolated from inclusion bodies and purified under denaturing conditions by affinity chromatography on Ni 2ϩ -nitrilotriacetic acid resin as described in the manufacturer's protocols (Qiagen). The purified products (H. pylori subA-His tag or H. pylori subB-His tag) were used to produce high titer rabbit antisera, recognizing the protein products of H. pylori ORF1 (A subunit) or ORF2 (B subunit), respectively.
Detection of H. pylori Gene Products-For detection of H. pylori gene products in Helicobacter or transformed E. coli, cellular lysates (see below) prepared in the presence or the absence of 100 M phenylmethylsulfonyl fluoride and 100 M EDTA were run on 15% SDS-polyacrylamide gels and transferred onto nitrocellulose membranes by electroblotting. After blocking overnight in 5% nonfat dry milk in Tris-buffered saline at pH 7.4, filters were incubated 60 min with rabbit serum directed against H. pylori subA-His tag or H. pylori subB-His tag and washed twice in Tris-buffered saline, pH 7.4, and twice in Tris-buffered saline containing 0.05% Nonidet P-40 (Sigma). Filters were then incubated for 60 min with goat anti-rabbit IgG antibodies coupled to horse-radish peroxidase (Amersham Corp.), washed as above, and developed using the chemoluminescence detection kit (ECL) from Amersham Corp.
Preparation of Cellular Lysates-E. coli JM105 cells (Pharmacia) were transformed with the plasmid pKK223-3 containing the 1322-bp fragment encoding the complete putative H. pylori CoA-transferase (pCoAT), the 705 bp and the 624 bp corresponding to the H. pylori CoA-transferase ORF1 (pCoATA) and ORF2 (pCoATB), respectively, or the H. pylori Urease A gene (pUreA) (28). Cells were grown overnight at 37°C in 25 ml of LB/50 g/ml ampicillin, transferred to 200 ml of fresh medium, and grown up to A 600 Ϸ 0.25. Cells were stimulated with 100 M isopropyl-␤-D-thiogalactopyranoside (Stratagene) for 1 h and harvested by centrifugation at 3000 ϫ g for 10 min.
For ␤-ketoadipate CoA and succinyl CoA:acetoacetate CoA-transferase assays, bacteria were washed once with M9 salts minimal medium and harvested by centrifugation at 3000 ϫ g for 10 min, and pellets were stored at Ϫ20°C until just prior to disruption. Thawed bacterial pellets were resuspended in 1 ml of 50 mM phosphate buffer, pH 6.8, 1 mM dithiothreitol (Merck). Bacterial suspensions were disrupted by sonication, and debris was removed by centrifugation. For acetateacetoacetate CoA-transferase assay, the buffer used for bacterial wash and pellets resuspension was 50 mM MOPS, pH 7.0, 0.5 M (NH 4 ) 2 SO 4 , 20% (v/v) glycerol, 1 mM EDTA. For measuring the generation of succinyl CoA, bacterial pellets were resupended in 50 mM Tris buffer, pH 7.4. For the determination of ␣-ketoglutarate dehydrogenase activity, bacterial pellets were resuspended in phosphate-buffered saline, pH 7.4, and disrupted by mild sonication. Protein concentration in cellular lysates was determined using the Bio-Rad protein assay with ␥ globulin as standard according to the supplier's instructions.
Acetate Acetoacetate CoA-transferase Activity Assay-Activity was measured by monitoring the decrease in A 310 due to the disappearance of acetoacetyl CoA as described previously (6,21). The assay was performed in 100 mM Tris-HCl, pH 7.5, containing 5% (v/v) glycerol, 40 mM MgCl 2 , 50 l of cellular lysate, 100 M acetoacetyl CoA, and 150 mM potassium acetate as carboxylic acid source.
Succinyl CoA-Acetoacetate CoA-transferase Activity Assay-Activity was measured by monitoring the increase in A 310 corresponding to the formation of acetoacetyl CoA as described previously (29). The assay contained 67 mM lithium-acetoacetate, 300 M succinyl CoA, and 15 mM MgCl 2 in 50 mM Tris-HCl, pH 9.1.
All enzymatic activities were monitored every minute for 4 min. One unit of enzyme activity is defined as the amount of enzyme required to convert 1 mol of substrate to product in 1 min under the assay conditions used. The ⌬OD of the first minute was used to estimate specific activities, which were expressed as milliunits/mg of total protein in the cellular lysates.
Other Enzymatic Activity Assays-The activity of ␣-ketoglutarate dehydrogenase was assayed by determining the formation of the reduced form of NADH as described (30) except that the reaction was performed in the presence of 1 mM ADP and 10 mM ␣-ketoglutarate. Measure of succinyl CoA synthesis was performed in the presence of excess hydroxylamine by complexing the succinohydroxamic acid formed to ferric salts (31) or by measuring the activity of succinyl CoA synthetase in presence of succinate, CoA, and ATP according to Bridger et al. (32).

Amplification of H. pylori DNA by the Polymerase Chain
Reaction and Screening of H. pylori lambda ZAP II Library-H. pylori 69A chromosomal DNA was amplified by PCR using primers ICT 14 and ICT 20 initially designed to select P-ATPase genes; the resultant PCR product of 1000 bp was partially sequenced. Sequence analysis revealed a striking homology with part of the genes encoding the CoA-transferase family (EC 2.8.3).
To verify the specificity of the PCR fragment as a probe, chromosomal DNA from H. pylori strains 69A and NCTC 11637 was cleaved with the restriction enzyme HaeIII and hybridized with the fluorescein-dUTP-labeled PCR product. Strong positive signals were observed with both strain DNAs (Fig. 1).
The probe was then used for the screening of an H. pylori Zap II library. Phages were amplified in E. coli XL-1 Blue cells, transferred to nylon membranes, and hybridized with the labeled PCR product. Fourteen positive clones were isolated after the third screening. Following in vivo excision, one plasmid named pGB1 was selected for further analysis.
Cloning and Analysis of the Nucleotide Sequence-Sequencing of pGB1 led to the identification of a 1395-bp fragment of H. pylori genomic DNA. Computer analysis of DNA sequence revealed two adjacent open reading frames designated ORF1 and ORF2 (Fig. 2). The largest open reading frame, ORF1, begins at nucleotide 73, terminates at nucleotide 771, and potentially encodes 233 amino acids. The second one, ORF2, located downstream, extends from nucleotide 771 to nucleotide 1391 and potentially encodes 207 amino acids. Both ORFs are oriented in the same direction. The position of the putative ribosome binding sites (Shine Dalgarno sequences) and the first methionine codons are underlined and written in bold type, respectively.
Sequence Comparison with Known Proteins-The predicted amino acid sequences were used to search through data bases by using the Blast programs from the University of Wisconsin Genetics Computer Group package. Significant homologies were found with five amino acid sequences, corresponding to CoA-transferases from very diverse organisms. The sequences of the different CoA-transferases were subsequently compared with each other using the Bestfit program. The protein most closely related to H. pylori ORFs is a B. subtilis protein. This protein of unknown function, listed as a probable ␤-ketoadipate CoA-transferase, exhibits 75 and 87% similarity with ORF1 and ORF2, respectively. The next most closely related to H. pylori ORFs is the succinyl CoA:3-oxoacid CoA-transferase from pig heart mitochondria (22) with 69 and 74% similarity. The human counterpart was recently cloned, and the amino acid sequence was published (18); the homology between the human and the H. pylori enzyme is 69% for ORF1 and 74% for ORF2. A significant sequence similarity was also displayed between H. pylori ORFs (60 -65% for ORF1 and 69 -73% for ORF2) and the ␤-ketoadipate:succinyl CoA-transferase of P. putida (12) and A. calcoaceticus (33) and the acetate-acetoacetate CoA-transferase from C. acetobutylicum (21).
Amino acid sequence alignments prepared by using the Pileup program (University of Wisconsin Genetics Computer Group) confirmed the phylogeny of the different CoA-trans- ferases and showed that H. pylori ORF1 and ORF2 align with the N and C termini, respectively, of the putative Bacillus subtilis, the pig and the human proteins (Fig. 3). Although the three latter enzymes are monomeric, all the other cloned CoAtransferases are made of two different subunits, named A and B.
Detection of CoA-transferase Genes in Bacteria-The ubiquity of the putative CoA-transferase gene(s) was monitored on chromosomal DNA by PCR amplification of the whole operon from nucleotides 73 to 1394 (Fig. 4A). All H. pylori strains tested (lanes 7-11) presented a strong positive signal. In contrast, no specific band was observed when using genomic DNA of E. coli JM 105, C. jejuni, or H. felis. To confirm the presence or the absence of similar sequences to that of H. pylori CoAtransferase in other bacteria, the bacterial DNAs were analyzed by slot blot and hybridized to the CoA-transferase AB probe using different stringency conditions (Fig. 4B). At the highest stringency conditions used (68°C), only the CoA-transferase genes of H. pylori strains were detected; however, signals observed at 42°C indicated that genes homologous to H. pylori CoA-transferases were present in H. felis and E. coli JM 105 genomes.
Detection of the CoA-transferase Protein(s) in H. pylori-To confirm that the H. pylori enzyme was made of two subunits as suggested by the analysis of the nucleotide sequences, the H. pylori genes encoding the two ORFs were separately cloned into pQE11, overexpressed in E. coli M15 with six extra histidine residues at the amino end and purified by affinity chromatography using Ni 2ϩ -nitrilotriacetic acid beads. The purified recombinant proteins were injected into rabbits to yield specific antibodies against each ORF product. The anti-ORF1 and the anti-ORF2 antisera were then used to detect the endogenous H. pylori CoA-transferase. The antisera recognized two separate proteins of M r ϳ26,000 (A subunit) and 24,000 (B subunit) (Fig.  5, lanes 1, A and B), respectively, confirming the dimeric structure of the putative CoA-transferase. The same results were obtained when cellular lysates were prepared in the presence of phenylmethylsulfonyl fluoride (Fig. 5, lanes 2, A and B), thus ruling out a serine protease type digestion of a monomeric precursor as reported for the eukaryotic homologue (22).
Expression of H. pylori CoA-transferase Genes in E. coli-The whole DNA fragment containing the CoA-transferase operon of H. pylori was then inserted into pKK223-3 to yield a plasmid named pCoAT. Following isopropyl-␤-D-thiogalactopyranoside induction of transformed E. coli cells, two distinct bands corresponding to proteins of approximately M r 26,000 and 24,000 were detected by Western blotting with the anti-A and the anti-B subunit antibodies independent of the presence of phenylmethylsulfonyl fluoride (Fig. 5, lanes 3 and 4, A and B), confirming the fact that H. pylori putative CoA-transferase is  . subtilis (B. subt), and pig and human heart mitochondria (Pig and Human). Identical residues among all aligned proteins are in bold type. The glycine cluster of each protein is underlined. The conserved active site glutamate is indicated with a bold asterix above the sequence. Numbers refer to the amino acid residues of the human heart SCOT. made of two independent subunits. No cross-reaction was observed between the anti H. pylori CoA-transferase antibodies and E. coli lysates (Fig. 5, lanes 5 and 6, A and B).
Measure of H. pylori CoA-transferase Enzymatic Activity in E. coli-Based on protein homologies with other CoA-transferases (Fig. 3), three different pairs of substrates were tested in the resulting bacterial lysates: ␤-ketoadipate ϩ succinyl CoA, acetate ϩ acetoacetyl CoA, and acetoacetate ϩ succinyl CoA. The enzymatic activity baseline was determined using cellular lysates of E. coli transfected with pUreA as a negative control (pKK223-3 containing 719 bp of H. pylori urease A gene). All three enzymatic activities tested were detected over basal levels. E. coli crude lysates expressing pCoAT showed a 10-, 22-, and 11,000-fold increase in ␤-ketoadipate CoA, acetate-acetoacetate CoA, and SCOT activities, respectively, compared with cells expressing pUreA (Table I).
To determine whether the two H. pylori CoA-transferase subunits were required for protein function, they were expressed separately in E. coli. Although synthesis of each subunit was evidenced by Western blot (data not shown), no SCOT activity could be detected (Table II). However, when equal amounts (measured as A 600 ) of A and B subunit expressing cells were mixed before preparing the lysates, part of the activity, corresponding to 1146 Ϯ 24 milliunits/mg, could be recovered.
Measure of Succinyl CoA:Acetoacetate CoA-transferase Activity in H. pylori-The only CoA-transferase activity detected in H. pylori lysates was the one leading to the formation of acetoacetyl CoA from succinyl CoA and acetoacetate in the presence of Mg 2ϩ (Fig. 6). The rate of formation of acetoacetyl CoA in H. pylori lysates was lower than the one measured in transformed E. coli and was not linear upon time (Fig. 6A), indicating that acetoacetyl CoA was processed concomitantly. The addition of 2 mM iodoacetamide, an inhibitor of acetoacetyl CoA thiolase (34) resulted in a 2.5-fold increase of CoA-transferase specific activity (60 Ϯ 2 milliunits/mg versus 25 Ϯ 2 milliunits/mg without iodoacetamide) (Fig. 6B) and in a linear accumulation of acetoacetyl CoA (Fig. 6A).
Demonstration of the Ability of H. pylori to Generate Succinyl CoA-In most bacteria, succinyl CoA is generated through the oxidative or the reductive arm of the tricarboxylic acid cycle, by an ␣-ketoglutarate dehydrogenase and by a succinate synthetase, respectively. We therefore have determined whether the corresponding activities were present in H. pylori. Although both enzymatic activities could be meas-  ured in control E. coli K12 lysates, they were below detection in H. pylori (Table III).
When the accumulation of succinyl CoA was measured in the presence of excess hydroxylamine by complexing the succinohydroxamic acid formed to ferric salts, 8.2 mol of succinyl CoA/mg of bacterial lysate was detected in H. pylori compared with 5.3 mol/mg in E. coli after 30 min. In contrast to what was observed for E. coli, the generation of succinyl CoA in H. pylori was independent of the addition of succinate, CoA, or ATP, the substrates of succinyl CoA synthetase, thus demonstrating that in H. pylori, succinyl CoA is generated by other enzyme(s). DISCUSSION We describe here the gene cloning and biochemical characterization of a novel enzyme of H. pylori. On the basis of comparative analyses of the amino acid sequences of identified CoA-transferases and reconstitution of the activity in E. coli, the H. pylori enzyme is a succinyl CoA:acetoacetate CoA-transferase. This study constitutes the first report of a bacterial SCOT.
Sequence comparisons between the H. pylori CoA-transferase and other members of the family showed that both subunits of H. pylori CoA-transferase present a high degree of homology with the SCOT isolated from pig (22), human heart mitochondria (18), and an hypothetical protein with no established function of B. subtilis (GenBank). Furthermore, a significant degree of homology was detected with the other cloned members of the CoA-transferase family such as the butyrate-acetoace-tate CoA-transferase (21) and the ␤-ketoadipate CoA-transferase (12), confirming the remarkable amino acid sequence conservation between CoA-transferases from very diverse organisms (12). If we exclude the Bacillus protein with no known function, all the bacterial CoA-transferases, including the H. pylori one, are made of two subunits, whereas the eukaryotic CoA-transferases are monomeric enzymes. Parales and Harwood (12) already noted that the A subunits of the prokaryotic proteins align with the N-terminal half of the mammalian protein, whereas the B subunits align with the C-terminal portion and that strong sequence similarity is seen throughout the length of the alignment with very few gaps. Interestingly, the pig heart enzyme is susceptible to two proteolytic events. One of them is produced in the hydrophilic region and generates N-and C-terminal fragments that retain full catalytic activity (22). This highly hydrophilic region falls between the alignments of H. pylori CoA-transferase A and B subunits (Fig. 3), as already observed (12). Our results directly support the hypothesis of Parales and Harwood, who suggested then that a gene fusion probably occurred at some time during evolution of CoA-transferases. The second cleaved form of the pig heart CoA-transferase involves an autolytic fragmentation at the active site thiol ester, glutamate 344 (35), known to be conserved in all sequenced CoA-transferases (Fig. 3, bold asterix). The other region with striking homology between the mammalian enzyme and the A subunits of other CoA-transferases is that of glycine clusters (Fig. 3, underlined), which may be implicated in CoA binding to the transferase (12).
Although CoA-transferases are very conserved proteins, mechanistically and functionally similar in catalyzing the reversible transfer of one CoA from one carboxylic acid to another, their substrate ranges and their role in metabolism appear to be very different. Although other CoA-transferases are present in E. coli (36), no succinyl CoA:acetoacetate CoAtransferase activity was detected in these bacteria, allowing us to monitor the H. pylori enzyme activity in reconstitution experiments after transfection of the H. pylori genes. Our results show that in the presence of succinyl CoA, the H. pylori enzyme is able to very efficiently convert acetoacetate into acetoacetyl CoA. No enzymatic activity could be detected when the structural genes encoding for the A and B subunits were expressed separately in E. coli, demonstrating that the two subunits are required for the function of the enzyme. It is of interest to note that genetic linkage of the H. pylori SCOT subunits is not required for the effective association of the two subunits into a functional heterodimer because activity can be restored by mixing E. coli cells expressing each subunit separately.
The succinyl CoA:acetoacetate CoA-transferase is made constitutively in H. pylori when the bacteria are grown in vitro, on plates or in liquid cultures. We still do not know, however, the importance of the H. pylori CoA-transferase for Helicobacter survival/pathogenicity in vivo. Although evidence for several major catabolic pathways has been obtained in H. pylori, little is known about its metabolic activities. The changes in substrate levels during culture of H. pylori and the effect of oxygen supply have been investigated by Chalk et al. (37). Their results suggest that H. pylori has the capability of both aerobic and anaerobic substrate utilization and are consistent with, during batch culture, an initial partially anaerobic phase of substrate utilization followed by a period of aerobic catabolism. However, the source of carbon used for energy metabolism by H. pylori is not yet elucidated. Homologues of enzymes involved in the utilization of carbohydrates, enzymes of the Entner Doudoroff and pentose phosphate pathways are present in H. pylori genome, consistent with recent observations that H. pylori has the capacity to metabolize glucose (38). A number of tricarboxylic acid cycle enzyme homologues and associated enzymes have been identified for which activities have been previously detected (39). However, Mendz et al. (38) suggested an absence of a functional citric acid cycle, indicated by the presence of fumarate reductase, which reduces fumarate to succinate rather than oxidizing succinate to fumarate as is done by a flavoprotein in the citric acid cycle (40). Our results concerning the lack of ␣ketoglutarate dehydrogenase and succinate synthetase activities in H. pylori, unambiguously demonstrate that succinyl CoA is not issued from the tricarboxylic acid cycle in these bacterium. Although we were able to evidence the ability of H. pylori to generate succinyl CoA at a metabolically significant rate, further investigations will be required to identify the enzyme(s) that is/are involved in this process.
Finding more evidence for induction of either enzymes that use acetoacetyl CoA to oxidize NADH or enzymes that use acetyl CoA to phosphorylate ADP via acetyl phosphate might help us to understand the role of H. pylori CoA-transferase. However, because we have evidence that the genes encoding the CoA-transferase are preceded by a gene showing high homology to a thiolase (data not shown), we postulate that the bacteria convert the substrate to acetoacetyl CoA, which is then cleaved to two acetyl CoA (Fig. 7). These two molecules of acetyl CoA may then provide energy by a metabolic cycle in which one of them regenerates succinyl CoA from succinate and the other provides the anhydride energy required for ATP synthesis. Gene disruption experiments should allow us to establish whether a mutant without these genes is able to survive/grow and infect hosts.
As already mentioned, succinyl CoA:3-oxoacid CoA-transferases have been identified in mammalian tissues such as the brain and the heart where ketone bodies such as acetoacetate and ␤-hydroxybutyrate constitute a major metabolic fuel. In rats, it was found that in the gastric glandular mucosa, the enzyme activity in mitochondria was as high as that in heart and kidney, and two to four times greater than in other regions of the gastrointestinal tract. It was suggested that acetoacetate metabolism might support acid secretion on refeeding after a period without food (41,42). Acetoacetate is definitively a substrate that is available in the gastric mucosa; however, our results do not allow us yet to assign a precise role to the succinyl CoA:acetoacetate CoA-transferase in H. pylori metabolism. A great deal of information about the kinetic of the mammalian enzyme and about the action of metabolic inhibitors is available (17,43) that should allow us to better understand the bacterial counterpart and use it as a potential target for new therapies against Helicobacter infection.