Pyruvate:flavodoxin oxidoreductase

The oxidative decarboxylation of pyruvate is an important reaction in archaea, bacteria, and eukaryotes alike, generating acetyl-CoA necessary for CAC reactions, fatty acid biosynthesis, and many other reactions requiring acyl-CoA. In many aerobic bacteria and also mammalian systems this reaction is catalyzed by the pyruvate dehydrogenase multi-enzyme complex (43). Under anaerobiosis in bacteria capable of mixed acid fermentation, the generation of acetyl-CoA from pyruvate is catalyzed by pyruvate:formate lyase (31). Biochemical and genomic data indicate that both of these enzymes are absent in H. pylori, and instead this reaction is catalyzed by an unusual four-subunit pyruvate:flavodoxin oxidoreductase (POR) enzyme (22, 23). POR enzymes are generally associated with an anaerobic-type metabolism, and there has been growing interest recently in this class of enzymes as targets for the development of novel anti-anaerobe compounds (48). Indeed, the first crystal structure of a POR enzyme has recently been published (9). These enzymes can be broadly grouped into three types: single-subunit POR, found for example in Clostridium sp. (37), Anabeana sp. and Klebsiella sp. (49), Desulfovibrio africanus (45), and some anaerobic protozoa (14, 48, 55); two-subunit POR, so far identified only in the aerobic archaeon, Halobacterium halobium (29); and the four-subunit POR commonly associated with the members of the archaea and hyperthermophilic eubacteria (4, 5, 52). POR has been purified from H. pylori and shown to belong to this ancestral, four-subunit group of enzymes (22), and was the first example of a four-subunit POR found in a mesophilic eubacterium. Characteristically, these enzymes catalyze the reduction of low-potential electron acceptors in vivo that are commonly either ferredoxin or flavodoxin proteins. There is evidence that the in vivo electron acceptor of the H. pylori POR enzyme is a flavodoxin (FldA; HP1161/JHP1088) (22, 23, 26). The molecular masses of the subunits are as follows: PorA, 47 kDa; PorB, 36 kDa; PorC, 24 kDa; and PorD, 14 kDa (22). The structural genes for POR are located adjacent to each other in the chromosomes of both strains 26695 and J99 in the gene order porCDAB (HP1108–1111/JHP1035–1038). The molecular mass of the native POR was estimated to be 240 kDa; however, the stoichiometry of the native complex is unknown. The PorB subunit displays the conserved amino acid binding motif for thiamine pyrophosphate (TPP) in common with other POR enzymes (23). The PorD subunit is strikingly similar to bacterial ferredoxins, displaying the characteristic cysteine-rich Fe-S binding motifs. Unlike pyruvate dehydrogenases, POR enzymes do not utilize either lipoic acid or flavin adenine dinucleotide (FAD) as cofactors, as the oxidative decarboxylation reaction involves a free-radical mechanism (30, 44, 50). The H. pylori POR was found to be highly oxygen labile, and the presence of this enzyme, along with the related oxygen-labile OOR enzyme (see below), may play an important role in the micro-aerophily of H. pylori (23, 28). Insertion-inactivation of the porB gene was unsuccessful, implying that POR is essential for the growth of H. pylori (23).

It is not known how the POR-reduced flavodoxin is reoxidized in H. pylori, but in cell-free extracts a POR-dependent reduction of NADP (but not NAD) can be observed, which would be consistent with reduced flavodoxin reducing NADP via a hypothetical flavodoxin:NADP oxidoreductase (23, 28). Alternatively, reoxidation of FldA may involve other electron acceptors, including the electron transport chain.

Finally, there is indirect evidence for in vivo expression of POR, through identification in the sera of H. pylori-infected individuals of the immunogenic PorA subunit (36).

Citrate synthase

Citrate synthase (EC 4.1.3.7) catalyzes the first step in the oxidative branch of the CAC in which acetyl-CoA and oxaloacetate are condensed to generate citrate and CoA. Citrate synthases in other bacteria are encoded by gltA, and a single copy of a gltA homolog is present in both H. pylori 26695 (HP0026) and J99 (JHP0022) strains. The purification of the H. pylori citrate synthase enzyme has been described (15). In common with other gram-negative citrate synthase enzymes, the native enzyme exists as a hexamer, with an approximate molecular mass of 300 to 340 kDa and a subunit size of 50 kDa. However, regulation of the H. pylori enzyme is more characteristic of citrate synthases from gram-positive bacteria and eukaryotes. For example, unlike citrate synthases from gram-negative bacteria, the H. pylori enzyme is not inhibited by NADH (46). This has been related to the role of the CAC in H. pylori being directed toward biosynthesis rather than energy generation. Also, 2-oxoglutarate, succinyl-CoA, AMP, and ADP, which inhibit other bacterial citrate synthases, do not inhibit the H. pylori enzyme. However, ATP was found to act as a strong competitive inhibitor with respect to acetyl-CoA (46).

Aconitase

The next step in the oxidative branch of the cycle is carried out by aconitase (EC 4.2.1.3), which catalyzes the reversible isomerization of citrate to generate isocitrate via cis-aconitate. Unlike E. coli, which contains two differentially regulated aconitase enzymes, AcnA and AcnB (12), only a single copy of an acnB homolog has been identified in the H. pylori genomes (HP0779/JHP0716). Aconitase activity has been detected in the cytosolic fraction of H. pylori cells both by NMR and spectrophotometric assays (46). However, full activity was only detected if the cytosolic fraction was recombined with the cell membrane fraction, indicating that interaction with the cell envelope is required for optimum activity.

Isocitrate dehydrogenase

Isocitrate dehydrogenase (EC 1.1.1.42) catalyzes the NAD(P)-dependent oxidative decarboxylation of isocitrate to generate 2-oxoglutarate and CO₂. In E. coli this enzyme acts as a critical branch point between the CAC reactions and the glyoxylate bypass during growth on C2 compounds like acetate. Regulation is achieved through reversible phosphorylation, leading to inactivation of this enzyme by the AceK protein, an isocitrate dehydrogenase kinase/phosphatase (33). The gene icd (HP0027/JHP0023), encoding isocitrate dehydrogenase, is located immediately downstream of the citrate synthase gene in H. pylori (1, 52). It is unlikely that the H. pylori isocitrate dehydrogenase enzyme acts as a branch point in isocitrate metabolism for two reasons: H. pylori probably lacks a functional glyoxylate bypass (see discussion below), and no homologs of aceK have been identified in the genome sequence. Pitson et al. (46) have reported that the H. pylori enzyme is NADP-specific and cannot utilize NAD as the electron acceptor, whereas Hoffman et al. (21) detected low levels of NAD-dependent activity. Sigmoidal kinetics with isocitrate and NADP indicates that H. pylori isocitrate dehydrogenase may be subject to allosteric regulation, and the activity is slightly stimulated by AMP (45).

2-Oxoglutarate metabolism

The next step in the oxidative branch of the conventional CAC is the oxidative decarboxylation of 2-oxoglutarate in the presence of CoA to generate succinyl-CoA with the release of CO₂. This reaction is commonly catalyzed by the 2-oxoglutarate dehydrogenase multi-enzyme complex. Both enzymatic data (11, 21, 23, 46) and genome sequence analysis (1, 53) indicate that H. pylori lacks this enzyme complex. Instead, oxidation of 2-oxoglutarate is carried out by an oxygen-labile 2-oxoglutarate:acceptor oxidoreductase enzyme (23), which shares a number of biochemical and structural features in common with the POR enzyme described above. The OOR enzyme has been purified from H. pylori (23) and is composed of four subunits, OorA (43 kDa), OorB (33 kDa), OorC (21 kDa), and OorD (10 kDa). In the chromosome, the structural genes for the OOR enzyme are located together in the order oorDABC (23) and correspond to gene numbers HP0588–0591 and JHP0536–0539 in H. pylori strains 26695 and J99, respectively (1, 53). OorB and OorD display sequence motifs for TPP and Fe-S cluster binding in common with POR and related enzymes. Interestingly, however, the pair-wise amino acid identity of the corresponding POR and OOR polypeptides is relatively low (25% or less), and in contrast to POR, the OorA and OorB subunits are more similar to the two-subunit puruvate:ferredoxin oxidoreductase of H. halobium than to the four-subunit enzymes commonly found in extremophiles. This strongly suggests independent evolutionary origins for these oxidoreductases in H. pylori. Mutants in oorA could not be generated, suggesting that OOR is essential for H. pylori (23).

Corthesy-Theulaz et al. (11) concluded that succinyl-CoA was not generated via CAC reactions in H. pylori owing to the absence of the 2-oxoglutarate dehydrogenase complex, but it is now clear that OOR effectively substitutes for this enzyme. The product of the OOR reaction is succinyl-CoA, and Hughes et al. (23) showed that the activity of the purified enzyme was strictly 2-oxoglutarate and CoA-dependent with K_m values of 0.3 and 13 μM, respectively. However, Pitson et al. (46) studied 2-oxoglutarate oxidation in H. pylori cell extracts using ¹H-NMR spectroscopy and reported an FAD or benzyl viologen-dependent 2-oxoglutarate "oxidase" activity not absolutely dependent on, but strongly stimulated by, the addition of CoA in an apparently allosteric manner. Unusual for an allosteric activator, 2-mM CoA increased the apparent K_m for 2-oxoglutarate and FAD to nonphysiological values (26 and 28 mM, respectively) despite the V_max also being raised. The product of this reaction was succinate, whereas succinyl-CoA production could not be detected. These results are difficult to reconcile with the properties of purified OOR reported by Hughes et al. (23) but may be due to the use of cell extracts in the work of Pitson et al. (46).

Interconversion of succinate and SCS

SCS catalyzes the sole reaction of the CAC in which a nucleotide triphosphate is generated from the conversion of succinyl-CoA to succinate and CoA. The enzyme can also operate in the reverse direction to generate succinyl-CoA. There is no biochemical evidence for the presence of SCS activity in H. pylori (11, 46), and no homologs of the sucCD genes have been identified in the either of the H. pylori genome sequences. The absence of this enzyme would be consistent with other evidence that the CAC in H. pylori consists of a reductive branch ending in succinate and an oxidative branch ending in succinyl-CoA, with no physiological need to connect the two branches. A succinyl-CoA:acetoacetate CoA-transferase (SCOT) enzyme has been identified by Corthesy-Theulaz et al. (11), which in principle could substitute for SCS (10), although this would be dependent on acetoacetate or acetoacetyl-CoA. The role of this enzyme is unclear and is further considered below.

Fumarate reductase

In E. coli, the interconversion of fumarate and succinate can be carried out by two enzymes: succinate dehydrogenase, which is expressed under aerobic conditions, or fumarate reductase, which is induced under anaerobiosis (25, 51). In the absence of oxygen, fumarate can be used as an alternative terminal electron acceptor for the proton-translocating electron transport chain and is important in ATP generation in many anaerobic bacteria (32). H. pylori lacks succinate dehydrogenase (1, 46, 53), whereas the presence of a fumarate reductase has been confirmed by biochemical analysis (3, 38, 46) and sequence data (1, 17, 53). H. pylori fumarate reductase is closely related to that of the related anaerobe Wolinella succinogenes and is encoded by three genes, frdCAB (HP0193–0191/jhp179–177), encoding polypeptides of 27, 81, and 31 kDa, respectively (17). FrdA and FrdB display the amino acid motifs for FAD and Fe-S binding, respectively, in common with other fumarate reductases and also succinate dehydrogenases. The frdC gene encodes a hydrophobic, di-heme cytochrome b, which may serve as a membrane anchor for FrdA and B. Like the fumarate reductase of W. succinogenes, activity was localized in the membrane fraction (3, 17). Mutants in frdA have been generated, indicating this enzyme is not essential for H. pylori. However, the mutants demonstrated a prolonged lag phase on standard growth medium under microaerobic conditions (17). FrdA is also immunogenic, and was recognized in 55% of serum samples from H. pylori-infected individuals (3). Expression of fumarate reductase activity appears to be constitutive in H. pylori, as unlike E. coli, levels of activity did not markedly change in cells grown under varying O₂ concentrations (13). Also, although the presence of fumarate reductase provides evidence of anaerobic-type respiration, H. pylori has not been successfully cultured under anaerobic conditions, even in the presence of additional fumarate as a terminal electron acceptor. Mendz et al. (40) found that the H. pylori enzyme was inhibited by three known inhibitors of fumarate reductase, morantel, oxantel, and thiabendazole, although high MICs would preclude their use as chemotherapeutic agents.

Fumarase

Fumarase (EC 4.2.1.2) catalyzes the reversible generation of malate from fumarate, and homologs of the structural gene for a type II fumarase, fumC, have been identified in both of the H. pylori genome sequences (HP1325/JHP1245). Fumarase activity has been investigated by Pitson et al. (46). The K_m and V_max values strongly indicate that the enzyme functions preferentially in the reductive direction, i.e., in the formation of fumarate from malate (fumarate K_m, 121 mM; V_max, 4.1 μmol min⁻¹ mg of protein ⁻¹; malate K_m, 7.3 mM; V_max, 10.8 μmol min⁻¹ mg of protein⁻¹). Again, this is consistent with fumarase operating in a reductive CAC arm leading to succinate via fumarate reductase.

Malate dehydrogenase (NAD-linked) and malate quinone oxidoreductase

Only very low specific activities (less than 10 nmol min⁻¹ mg of protein⁻¹) of NADH-dependent malate dehydrogenase were detected in H. pylori by Hoffman et al. (21), but much higher activities (around 100 nmol min⁻¹ mg of protein⁻¹ assayed in the reductive direction) were measured by Pitson et al. (46). It is difficult to reconcile these values except as strain or assay differences. No homologs of a typical NAD-MDH have been identified in either of the H. pylori genome sequences, suggesting the enzyme responsible for the activity is very divergent or otherwise novel.

Davison et al. (13) reported the presence in H. pylori of a loosely membrane-bound dye-linked (NAD-independent) malate dehydrogenase activity, and l-malate-dependent cytochrome reduction. The enzyme responsible was predicted by Kelly (28) to be a flavoprotein malate:quinone reductase (Mqo), which functions as an electron donor to the quinone pool, analogous to the better known d-lactate dehydrogenase (41). Kather et al. (27) have confirmed this and have shown that the HP0086 ORF encodes a protein with Mqo activity that is only distantly related to known Mqo enzymes. Why should H. pylori possess two types of malate-oxidizing enzymes? It is likely that these function differentially; the Mqo is involved in the use of malate as a respiratory chain electron donor, whereas the NAD-MDH could function in the opposite (reductive) direction to synthesize malate as part of the reductive C4 branch of the CAC pathway.

Aspartase

The assimilation of amino acids may play an important role in the provision of biosynthetic intermediates and energy conservation in H. pylori in vivo, as the bacterium is capable of growth in defined media supplemented with only amino acids as the sole carbon source (39). H. pylori was found to metabolize aspartate rapidly, and the initial product of aspartate metabolism was identified as fumarae (39), indicating a role for aspartase in fumarate formation. Aspartase is encoded by the aspA gene, and homologs have been identified in both genome sequences (HP0649/jhp594).

Enzymes of the glyoxylate bypass

Malate synthase and isocitrate lyase together catalyze the two reactions of the glyoxylate bypass, which is an anaplerotic mechanism for C4 dicarboxylic acid synthesis that specifically operates during growth on C2 compounds as sole carbon source. Malate synthase catalyzes the conversion of glyoxylate and acetyl-CoA to malate and CoA, and is encoded by aceB in E. coli. Although significant malate synthase activity has been detected in H. pylori (21, 46), no homologs of aceB have been identified in the genome sequences, possibly suggesting that the H. pylori enzyme may have diverged from other malate synthases. Hoffman et al. (21) have reported very low isocitrate lyase activity in some strains whereas Pitson et al. (46) could not detect this enzyme activity either spectrophotometrically or by ¹H-NMR. As no homologs of isocitrate lyase genes have been identified in the genome sequence, and the phenylhydrazine assay is prone to interference from many carbonyl-containing compounds, it seems unlikely that a functional glyoxylate bypass is operative in H. pylori.

Indications from the genome sequence

Few studies have examined the synthesis of fatty acids in H. pylori, and most of our understanding has come from analysis of the genome sequences (1, 53). H. pylori utilizes the type II or dissociated fatty acid synthesis pathway, typical of most bacteria and plants, in which discrete proteins catalyze individual steps in the pathway. Table 2 lists the relevant genes in strains 26695 and J99, and Figure 2 shows the predicted pathway for fatty acid biosynthesis in H. pylori, largely based on the E. coli model for fatty acid biosynthesis (reviewed in reference 47).

Table 2

Summary of genes associated with fatty acid synthesis in H. pylori.

Figure 2

. Predicted pathways for fatty acid and phospholipid biosynthesis in H. pylori. During the initiation phase of fatty acid biosynthesis, acetyl-CoA is carboxylated to generate malonyl-CoA, which is then converted to malonyl-ACP. Malonyl-ACP is also required (more...)

The first committed step of fatty acid biosynthesis, in which malonyl-CoA is generated from acetyl-CoA, is catalyzed by acetyl-CoA carboxylase (ACC) (see below). Malonyl-CoA reacts with the acyl-carrier protein (ACP) to generate malonyl-ACP, which is catalyzed by the fabD gene product, malonyl-CoA:ACP transacylase. Malonyl-ACP is required not only for initiation of fatty acid biosynthesis, but also for each subsequent round of elongation of the fatty acid chain. ACP is a small, soluble protein that plays a critical role in fatty acid biosynthesis in most bacteria by covalently binding the intermediates of the pathway. The structural gene for ACP (acpP) has been identified in both genome sequences. Interestingly, there is an additional ACP homolog found only in H. pylori strain 26695 (HP0962), which encodes a larger protein, extended at the N terminus. To function in fatty acid biosynthesis, the apo-ACP protein must first be activated by transfer of the 4′-phospho-pantotheine from CoA, and this reaction is predicted to be catalyzed by holo-ACP synthase, encoded by acpS in H. pylori.

To initiate fatty acid synthesis, acetoacetyl-ACP is generated from malonyl-ACP, and in E. coli there are three potential pathways leading to the formation of acetoacetyl-ACP (47). First, malonyl-ACP can undergo condensation with acetyl-CoA to generate acetoacetyl-ACP. In E. coli, this reaction is catalyzed by β-ketoacyl-ACP synthase III (FabH), which also catalyzes an alternate pathway in which acetyl-CoA is converted to acetyl-ACP. Acetyl-ACP is then condensed with malonyl-ACP to generate acetoacetyl-ACP, which in E. coli is catalyzed by β-ketoacyl-ACP synthase I (FabB) and II (FabF). FabB can also function to decarboxylate malonyl-ACP to generate acetyl-ACP, which then undergoes condensation with a second malonyl-ACP molecule. In the case of H. pylori, fabF and fabH homologs are present, whereas fabB has not been identified. Further biochemical studies are required to confirm the function, and to ascertain the substrates and side reactions associated with these enzymes in H. pylori.

Acetoacetyl-ACP is converted to β-hydroxyacyl-ACP by the β-ketoacyl-ACP reductase enzyme (FabG), and the product of this reaction is then dehydrated. In E. coli, this dehydration reaction can be carried out by two enzymes, FabA and FabZ. However, only FabA catalyzes the isomerization of trans-2 decenoyl-ACP to cis-3-decenoyl-ACP, the intermediate for unsaturated fatty acid synthesis in E. coli. A homolog of fabZ is present in the H. pylori genome sequences, whereas fabA has not been identified. It is currently unknown whether the H. pylori FabZ has a similar role to E. coli FabA, as the branch point between the synthesis of saturated and unsaturated fatty acids. FadR, which has a dual role in E. coli as a positive activator for transcription of fabA and a repressor of fatty acid degradation, has not been identified in H. pylori.

The final step in the cycle is catalyzed by enoyl-ACP reductase (FabI) to generate acyl-ACP, which can then enter another round of elongation or, alternatively, enter the pathway for membrane phospholipid synthesis. The first step of phospholipid biosynthesis is the transfer of the acyl chain to the 1-position of glycerol-3-phosphate, the scaffold for this synthesis, by glycerol-3-phosphate acyltransferase (PlsB) (19). A plsB homolog has not been identified in the H. pylori genome. However, a homolog of the gene encoding PlsC, which catalyzes the transfer of a second acyl group to glycerol-3-phosphate, is present. H. pylori also contains a homolog of plsX, which by mutational analysis plays an as yet poorly defined role in phospholipid biosynthesis in E. coli (34). Genes associated with the production of CDP-diglycerol (cdsA), phosphatidylserine (pssA), and phosphatidylethanolamine (psd) from phosphatidic acid have been identified in H. pylori, and Ge and Taylor (16) have found that insertion-inactivation mutants of pssA could not be generated, indicating this enzyme is essential in H. pylori. A homolog of pgsA encoding the enzyme for production of phosphatidylglycerophosphate is present, whereas the genes encoding the subsequent enzymes of this pathway, leading to the production of phosphatidylglycerol and cardiolipin (pgp and cls, respectively), have not been identified.

As discussed above, H. pylori contains cyclopropane fatty acids, which in other bacteria are generated through modification of membrane phospholipids. This modification involves methylenation of the phospholipid by S-adenosylmethionine and is predicted to be catalyzed by the product of the H. pylori cfa gene. H. pylori also contains unusual cholesteryl glucosides, which are synthesized though the action of UDP-glucose:sterol glucosyl-transferases in other organisms. No sequence similarity to genes encoding these enzymes has been identified in H. pylori.

ACC

H. pylori ACC has been characterized by Burns et al. (6), and the possible role of this enzyme in the capnophilic phenotype of H. pylori is discussed elsewhere (see chapter 10). Genome sequence data (1, 53) indicate that H. pylori ACC is typical of other prokaryotic enzymes, in that it is composed of four individual proteins: biotin carboxylase, biotin carboxyl carrier protein, and two subunits constituting the carboxytransferase. NMR spectroscopy has confirmed that the end product of the ACC reaction is malonyl-CoA, and that the reaction is reversible (6).

SCOT and thiolase

Several enzymes have been identified in H. pylori that may contribute to the production of acetyl-CoA in the bacterium. Corthesy-Theulaz et al. (11) have identified a SCOT enzyme in H. pylori, which catalyzes the reversible transfer of CoA from succinyl-CoA to acetoacetate to generate acetoacetyl-CoA. H. pylori SCOT is composed of two subunits of molecular masses of 26 kDa (encoded by HP0691/JHP0637) and 24 kDa (HP0692/JHP0636). SCOT may play a role in the metabolism of acetoacetate in H. pylori, forming acetyl-CoA from acetoacetyl-CoA through the action of a thiolase enzyme, encoded by fadA (HP0690/JHP0638), which lies immediately upstream of the SCOT (scoAB) genes. However, the precise physiological role of SCOT in H. pylori has not been clearly defined, and a role in succinate/succinyl-CoA interconversion has also been suggested (10, 11).