Module 3.6.3.3

Biosynthesis of Menaquinone (Vitamin K₂) and Ubiquinone (Coenzyme Q)

R. MEGANATHAN^1* AND OHSUK KWON²

[SECTION EDITOR: JOHN CRONAN]

Posted December 23, 2009

Department of Biological Sciences, Northern Illinois University, DeKalb, IL 60115,¹ and Omics and Integration Research Center, Korea Research Institute of Bioscience and Biotechnology (KRIBB), 52 Eoeun-dong, Yuseong-gu, Daejeon 305-333, Korea²

*Corresponding author. Phone: (815) 753-7803, Fax: (815) 753-0461, E-mail: This e-mail address is being protected from spambots. You need JavaScript enabled to view it

INTRODUCTION

Facultatively anaerobic gram-negative bacteria, including Escherichia coli and Salmonella enterica serovar Typhimurium, contain the isoprenoid quinones of the benzene and naphthalene series. The structures of these quinones are shown in Fig. 1. According to the IUPAC-IUB recommendations (65), the benzoquinones are termed ubiquinones (Q-n) (structure I in Fig. 1) and the naphthoquinones are termed either menaquinones (MK-n) (structure II in Fig. 1) or demethylmenaquinones (DMK-n) (structure III in Fig. 1). The n refers to the number of prenyl units present in the side chain. It should be pointed out that while MK is considered a vitamin (vitamin K₂), Q is not, because vitamin K is an essential nutrient (cannot be synthesized) for mammals, while Q is not an essential nutrient since it can be synthesized from the aromatic amino acid tyrosine.

Fig. 1Structures of major quinones found in E. coli. In the structure of MK, the A ring and B ring of the naphthoquinone are shown.

The major quinones in E. coli are Q-8, MK-8, and DMK-8; minor amounts of Q-1 to Q-7, Q-9 and MK-6, MK-7, MK-9, and DMK-7 may also be present (23). The prenyl side chains have all-trans configuration (7). In contrast to the extensive investigations on the quinone composition and biosynthesis in E. coli, S. enterica has been studied to a lesser extent. These organisms neither have quinones that have one or more of the prenyl residues of the side chain reduced nor MK with more than one methyl group. Methods for the extraction, purification, identification, and analysis of the quinones have been reviewed extensively (7, 22, 27, 33, 46, 59, 83, 85, 96, 100, 101, 102, 113, 127, 128, 130, 144, 157, 160, 170).

Since the last print edition of Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, the following reviews on the subject have appeared. Reviews on the reaction mechanisms of various enzymes involved in MK and Q biosynthetic pathways have been published by Begley (6) and Meganathan (106). Two short reviews on Q biosynthesis have appeared (107, 152). In most of these reviews, the work on E. coli and, to a lesser extent, on Salmonella predominated because of the ease with which these organisms can be manipulated. However, because of advances in technology, it has become a reality that work on other bacteria can be carried out with ease.

Hence, driven by economic and chemotherapeutic potential, research in certain aspects of the MK and Q biosynthesis in some organisms has moved ahead of that of E. coli. The MK and the phylloquinone biosynthetic pathways (the reactions of biosynthesis are identical with the exception of the prenylation) are unique to bacteria and plants, respectively, and are absent in humans and animals. Hence, there is great commercial interest in discovering chemicals that will inhibit the enzymes of the pathway. Such chemicals will be of use as herbicides and chemotherapeutic agents against pathogens such as multidrug-resistant Mycobacterium tuberculosis and methicillin-resistant Staphylococcus aureus (MRSA). With the idea of designing drugs, the crystal structure of MenB from M. tuberculosis has been solved by two different groups (70, 161) and that of S. aureus has been reported recently (163). Hence, in this review, whenever there is work on organisms that is not available in E. coli it will be included. It is expected that the crystal structures of enzymes will be similar if not identical in most organisms since they perform chemically identical reactions.

Most of the information concerning the biosynthesis of MK and Q was obtained with E. coli by use of isotopic tracers, the isolation of mutants, and accumulation of intermediates and enzyme assays. Due to space limitations, only a general account is given here; for more information, several comprehensive reviews should be consulted (8, 9, 10, 12, 13, 53, 54, 163, 168). Both MK and Q are derived from the shikimate pathway and, as such, have some common structural features. The quinone nucleus of Q is derived directly from chorismate while that of MK is from isochorismate via chorismate. The prenyl side chain on the nucleus of both is derived from prenyl PP_i, and the methyl groups are derived from S-adenosylmethionine. In addition, MK biosynthesis requires 2-ketoglutarate and thiamine PP_i (TPP), coenzyme A, and ATP as cofactors. The biosynthesis of Q under aerobic conditions has the additional requirements for oxygen, flavoprotein, and NADH. Finally, it should be noted that the Q biosynthetic pathway in prokaryotes differs in several respects from that of eukaryotes (71, 106, 107, 122, 152).

Despite the fact that both quinones originate from the shikimate pathway, there are several important differences.

1. In the formation of the quinoid nuclei, the pathway for Q diverges at chorismate with the loss of a pyruvoyl group, due to the action of chorismate lyase, resulting in the formation of a benzenoid aromatic acid that is used as the framework on which the rest of the molecule is constructed. MK biosynthesis diverges at isochorismate by the addition of succinic semialdehyde-TPP anion derived from 2-ketoglutarate resulting in the formation of 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylic acid (SEPHCHC). In the subsequent reaction, the pyruvoyl group is eliminated, resulting in the prearomatic compound 2-succinyl-6-hydroxy-2,4-cyclohexadiene-1-carboxylic acid (SHCHC). This is then aromatized to a benzenoid aromatic acid and used as the framework for the construction of the rest of the molecule as shown in Fig. 2.

2. In Q biosynthesis, the prenyl side chain is introduced at an early stage (second step) with the retention of the aromatic carboxyl group. Conversely, while in MK biosynthesis, prenylation is the next to last step and is accompanied by a decarboxylation.

3. In MK biosynthesis all the enzymes in the pathway up to the prenylation are soluble (next to the last step). In Q biosynthesis the enzymes are membrane bound except for the first enzyme.

4. In MK biosynthesis, methylation of the carbon of the nucleus is the last step while the terminal step of Q biosynthesis is the methylation of a hydroxyl group. In addition, in Q biosynthesis a second O-methylation and C-methylation take place in the middle portion of the pathway. It is surprising that both the C-methylation involved in MK biosynthesis and the C-methylation in Q biosynthesis are carried out by the same C-methyltransferase.

5. Q biosynthesis under aerobic conditions requires the introduction of OH groups by reactions involving oxygen; anaerobic Q and MK biosynthesis utilize oxygen atoms derived from water.

Fig. 2Formation of Q and MK.

MK BIOSYNTHESIS

The MK biosynthetic pathway has been elucidated on the basis of tracer experiments, isolation of mutants blocked in the various steps, isolation and identification of intermediates accumulated by the mutants, and enzyme assays. Early isotopic tracer experiments with various bacteria established that methionine and prenyl PP_i contribute to the methyl and prenyl substituents of the naphthoquinone. The early isotopic tracer studies and other work have been reviewed by Bentley and Meganathan (12). In 1964, Cox and Gibson observed that [G-¹⁴C]shikimate was incorporated into both menaquinone and ubiquinone by E. coli, thus providing the first evidence for the involvement of the shikimate pathway (25). Chemical degradation of the labeled isolated menaquinone (MK-8) showed that essentially all of the radioactivity was retained in the phthalic anhydride. It was concluded that “the benzene ring of the naphthoquinone (sic) portion of vitamin-K₂ arises from shikimate in E. coli” (25). The authors further suggested that shikimate was first converted to chorismate before incorporation into MK. A more complete chemical degradation of the menaquinone derived from labeled shikimate established that all seven carbon atoms were incorporated (18). The remaining three carbon atoms of the naphthoquinone ring were shown to be derived from the middle three carbons of 2-ketoglutarate with the loss of both carboxyl groups (17, 132, 133).

These studies established the immediate precursors of menaquinone as shikimate and the noncarboxyl carbon atoms of 2-ketoglutarate forming the naphthoquinone nucleus. The methyl and isoprenoid side chains were also shown to be derived from S-adenosylmethionine and an isoprenyl alcohol pyrophosphate ester, respectively. Subsequently it was shown that the benzenoid aromatic compound o-succinylbenzoate (OSB) (29) and the naphthalenoid aromatic compound 1,4-dihydroxy-2-naphthoate (DHNA) (12, 134) were incorporated into the naphthoquinone ring of menaquinone. This work was confirmed by the demonstration that menB and menA mutants of E. coli excrete OSB and DHNA, respectively, into the culture medium (173). During a study of the biosynthesis of OSB by growing cultures of E. coli menB it was demonstrated that carbon atom one of the glutamate (2-ketoglutarate) was lost and consequently not incorporated into OSB (109). The isotopic labeling pattern is summarized in Fig. 3.

Fig. 3Primary biosynthetic precursors of menaquinones.

Formation of Isochorismate (Compound IV → V)

The first synthesis of OSB (IX), from chorismate (IV) and 2-ketoglutarate in the presence of TPP by cell-free extracts of E. coli, was obtained by Meganathan (103) (Fig. 4). However, it had been suggested that isochorismate (V) was a much more attractive precursor than chorismate on chemical grounds (29, 53). Evidence in support of this hypothesis was provided (11, 36, 167).

Fig. 4Menaquinone biosynthetic pathway. Each compound in the pathway is identified by its abbreviation and a Roman numeral. CHA, chorismate; ICHA, isochorismate; SS-TPP, succinic semialdehyde-TPP; R₁, pyrimidine component of TPP; R₂, CH3; R₃, CH₂CH₂ OP₂O₆³⁻; SEPHCHC, 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylate; SHCHC, 2-succinyl-6-hydroxy-2,4-cyclohexadiene-1-carboxylate; OSB, o-succinylbenzoate; OSB-CoA; o-succinylbenzoyl-CoA; DHNA-CoA, 1,4-dihydroxy-2-napththoyl-CoA; DHNA, 1,4-dihydroxy-2-napththoate; DMK-8, demethylmenaquinone (may be initially formed as a quinol); MK-8, menaquinone; SAM, S-adenosylmethionine; SAH, S-adenosylhomocysteine. The genes encoding the enzymes are shown for each reaction followed by their location on the chromosome in min. The gene encoding the thioesterase for the conversion of DHNA-CoA (compound XI) to DHNA (compound XII) remains to be identified and is shown as men ??.

Isochorismate is a common intermediate in the biosynthesis of the siderophore enterobactin and MK. The conversion of chorismate to isochorismate in enterobactin biosynthesis is mediated by the enzyme isochorismate synthase encoded by the entC gene (119, 162). The dual role of isochorismate led to the question as to whether the entC-encoded isochorismate synthase (EntC) was supplying the isochorismate required for both pathways. Kaiser and Leistner (72) reported the isolation of a Tn10 insertion in the entC gene that had lost simultaneously the ability to form enterobactin and MK. It is generally accepted that the entC gene is derepressed under iron deficiency and repressed under iron sufficiency (47, 116). Enterobactin is required only under aerobic conditions because of the poor solubility and the consequent unavailability of iron in the Fe³⁺ form. When E. coli is grown anaerobically, iron is present in the highly soluble Fe²⁺ form. Hence, the synthesis of enterobactin is unnecessary for the acquisition of iron by the cell under anaerobic conditions (47, 116).

In contrast, MK is required under anaerobic conditions (12). Further, when the organism is grown with fumarate, trimethylamine-N-oxide (TMAO), or dimethyl sulfoxide (DMSO) as electron acceptor, the presence of MK is obligatory (12, 48, 104, 112). When oxygen or nitrate are the electron acceptors, the aerobic quinone, ubiquinone is used by E. coli (84). Thus, while the conditions that favor the biosynthesis and function of Q are compatible with the biosynthesis of enterobactin, they are incompatible with the biosynthesis of MK.

These apparent contradictions raised some intriguing questions. How does E. coli, growing aerobically under iron deficiency when entC is fully derepressed, prevent the synthesis of MK? Further, under anaerobic conditions, how does E. coli prevent the synthesis of enterobactin when MK synthesis is induced? This paradox might be resolved if the entC gene is regulated by iron in the presence of oxygen and by MK requirement in the absence of oxygen. To study the regulation of the entC gene, an entC-lacZ operon fusion was constructed and the expression of β-galactosidase monitored under various conditions. It was found that the β-galactosidase was fully derepressed at low concentrations of iron and repressed at high iron concentrations under both aerobic and anaerobic growth conditions (91, 92).

These results raised the question as to how E. coli is able to synthesize MK anaerobically when growing in the presence of high concentrations of iron. How does the organism prevent the excess production of MK under iron-deficient aerobic conditions when entC is fully derepressed? To answer these questions, anaerobic growth of an entC::Tn5 mutant was tested on glycerol medium with TMAO, DMSO, or fumarate as electron acceptor. The mutant was able to grow at the same rate as the parent, even in the presence of high concentrations of iron. Further, the mutant produced as much MK as the parent (91, 92, 105). These results provided clear evidence for the presence of an alternate isochorismate synthase specifically involved in MK biosynthesis. As a first step in locating and identifying the gene encoding this alternate isochorismate synthase involved in MK biosynthesis, further sequencing upstream of the 5’ region of the menD gene was carried out. An open reading frame encoding a 430-amino-acid protein exhibiting about 20% amino acid identity with EntC was identified as MenF (30, 31, 91).

The isochorismate synthase (MenF) encoded by the menF gene has been overexpressed and purified to homogeneity. The purified enzyme had a relative M_r of 48,000 (30, 31) as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The native M_r, as determined by gel filtration chromatography, was 98,000, thus establishing that the native enzyme is a homodimer (30). The enzyme showed a requirement for Mg²⁺ for maximal activity.

It is expected that the origin of the hydroxyl groups and the mechanism of the reaction of MenF and EntC will be identical. Four different mechanisms were proposed for the EntC enzyme (165). The origin of the hydroxyl could be from three possible sources: (i) molecular oxygen, (ii) intramolecular transfer of the hydroxyl, or (iii) the solvent H₂O. While the incorporation of molecular oxygen is possible only in the case of aerobic organisms, intramolecular transfer or the incorporation of hydroxyl from water can be carried out by both aerobes and anaerobes. Because of the reported absolute requirement of enterobactin for the chelation of iron during aerobic growth under iron deficiency (47, 116), one would expect the incorporation of molecular oxygen into the hydroxyl group. However, the absence of redox cofactor rules out the involvement of oxygen, and evidence has been obtained demonstrating the incorporation of the C-6 hydroxyl from the solvent H₂O for EntC (165). Consistent with this result is the demonstration of anaerobic biosynthesis of enterobactin in E. coli (91).

Recently, the 3D structure of MenF has been determined and the catalytic mechanism probed by site-directed mutagenesis and biochemical studies (80). Lys-190 has been identified as the active site base that assists in the attack by water at the C₂ carbon. An S_N2” reaction results in the rearrangement of 1-2, 5-6 double bonds resulting in the elimination of the C4 hydroxyl group (80). These findings are in complete agreement with a common mechanism proposed for the three chorismate-utilizing enzymes, anthranilate synthase (AS), 4-amino-4-deoxychorismate synthase (ADCS), and isochorismate synthase (IS) by He et al. (55).

Formation of Succinic Semialdehyde-TPP (SS-TPP) Anion and Addition to Isochorismate (compound VI + V → VIII)

During the studies on the biosynthesis of OSB (IX), cell extracts of two groups of mutants designated as menC and menD blocked in the formation of OSB and requiring OSB for anaerobic growth on glycerol-fumarate medium were examined. Cell extracts of either mutant alone did not form OSB from chorismate (IV) and 2-ketoglutarate in the presence of thiamin pyrophosphate. However, extracts from both mutants in combination produced OSB, and extracts of menC mutants accumulated an intermediate, which was converted to OSB by extracts of menD mutants (110). The intermediate was found to be unstable, and on mild acid treatment yielded OSB and succinylbenzene. On the basis of these properties and nuclear magnetic resonance data, the intermediate was identified as 2-succinyl-6-hydroxy-2,4-cyclohexadiene-1-carboxylate (SHCHC) (VIII) (36).

It has been postulated that the 2-ketoglutarate undergoes a TPP-dependent decarboxylation, with the formation of succinic semialdehyde anion of TPP (VI) (17, 103) and a requirement for TPP in the reaction was shown (110). The mechanism of decarboxylation is identical to that catalyzed by the first enzyme of the 2-ketoglutarate dehydrogenase complex (KGDH complex) (Fig. 5) (11, 99). By use of a sucA mutant (which lacks the first enzyme of the KGDH complex), and by selective removal of the KGDH complex, it was established that the 2-ketoglutarate decarboxylase (KDC) involved in OSB synthesis is a separate enzyme (13, 99, 167).

Fig. 5Proposed mechanism of succinic semialdehyde-TPP anion formation and SHCHC (VIII) synthesis. Only the thiazole ring of the TPP is shown since it is the active site of the molecule. For R₁, R₂, and R₃ see the legend to Fig. 4. The reactions from (XIII) to SEPHCHC (VII) are carried out by menD and the conversion of SEPHCHC (VII) → SHCHC (VIII) is by menH.

Subsequent studies established that the succinic semialdehyde anion (VI) of TPP reacted with isochorismate (V) resulting in the formation of SHCHC (VIII) (11, 36, 167), as had been postulated previously (29, 53). A mechanism for this reaction has been proposed (Fig. 5) (10, 11, 106). When the complete nucleotide sequence of the menD gene was determined it was discovered that both SHCHC synthase and KDC activities are encoded by a single gene (120). This conclusion was further strengthened by overexpression and purification of the MenD protein and by showing that both activities copurified during various steps of the purification process (14, 89).

However, recently, Guo and colleagues have shown that the formation of SHCHC from isochorismate and 2-ketoglutarate is a two-step process requiring two different enzymes. The first enzyme, MenD, decarboxylates the 2-ketoglutarate and adds the resulting succinic semialdehyde anion of TPP (VI) to isochorismate (V), resulting in the formation of 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylate (SEPHCHC) (VII) (67). The mechanism of formation of succinic semialdehyde anion of TPP and its addition to isochorismate is shown in Fig. 5. The stereochemistry of SEPHCHC was determined and shown to be (1R,2S,5S,6S)-2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylic acid (68). On the basis of these results, MenD was designated as SEPHCHC synthase (67).

SEPHCHC is an unstable compound that, in mildly basic solutions, spontaneously undergoes a 2,5-elimination reaction resulting in the formation SHCHC and pyruvate. Crystallization and a preliminary X-ray analysis of MenD has been reported (148).

The in vivo conversion of SEPHCHC to SHCHC (compound VII to VIII) is carried out by (1R,6R)-2-succinyl-6-hydroxy-2,4-cyclohexadiene-1-carboxylate (SHCHC) synthase encoded by the menH gene (Fig. 4 and 5). Surprisingly, MenH contains a Ser-His-Asp catalytic triad, which is typical of many proteases. This triad plays a critical role in enzyme activity since replacing any one of the three amino acids by alanine results in a dramatic decrease in catalytic activity (69). The structure of MenH from the enteric pathogen Vibrio cholerae has been determined (PDB ID code IR3D).

Aromatization of SHCHC → OSB (Compound VIII → IX)

Enzymatic removal of the elements of water from SHCHC (VIII) leads to the formation of the benzenoid aromatic compound OSB (IX) (Fig. 4). The first evidence for the presence of such an enzyme was obtained by the demonstration that cell-free extracts of a menD mutant converted SHCHC (designated as “X” at the time) to OSB (110). This enzyme was subsequently designated as OSB synthase (126, 141). The gene encoding OSB synthase was cloned, and its complete nucleotide sequence has been reported (141).

The enzyme was overexpressed and purified to homogeneity, and its properties have been investigated. The enzyme required a divalent metal ion for activity like the other members of the enolase superfamily. The enzyme was shown to carry out the dehydration of SHCHC to OSB very efficiently with K_cat of (19 ± 1 s⁻¹) and a K_cat/K_m of (1.6 ± 0.3 × 10⁶ M⁻¹ s⁻¹) (121). OSB synthase was classified as a member of the enolase superfamily. Members of this superfamily carry out reactions initiated by abstraction of the α-proton from a carboxylate anion substrate to generate a stabilized enolate anion intermediate (3). As pointed out above, the reaction catalyzed by OSB synthase is a dehydration. It was proposed that the α-proton of the carboxylate substrate (SHCHC) is likely abstracted by a basic catalyst (one lysine) followed by the elimination of the β-hydroxyl group presumably by the assistance of an acid catalyst (a second lysine) (121).

The structure of OSB synthase from E. coli in complex with Mg²⁺ and o-succinylbenzoate was determined. It was found that OSB synthase is the only monomeric member of the enolase superfamily. The product OSB was found to be sandwiched between Lys-133 and Lys-235 located at the ends of the second and sixth β-strands. In addition, one carboxylate oxygen of the substrate is coordinated to the Mg²⁺ (159). Subsequently, the structure of OSB synthase from an inactive K133R mutant in complex with the substrate SHCHC was determined. It was found that Lys-133 is the single base/acid catalyst for the dehydration with the transient Mg²⁺-coordinated enolate anion intermediate. The dehydration was shown to follow a syn-stereochemical course (74). The mechanism and specificity of various members of the enolase superfamily including OSB synthase have been reviewed and should be consulted for further details of the reaction mechanism (42).

Cyclization of OSB to DHNA (Compound IX → XII)

The conversion of the benzenoid aromatic compound OSB (IX) to the naphthalenoid aromatic compound DHNA (XII) was demonstrated by Bryant and Bentley (16). The process showed an absolute requirement for ATP and CoA. Hence, OSB-CoA (X) was suggested as an intermediate. With use of extracts of Mycobacterium phlei, evidence was obtained for the presence of two enzymatic activities (OSB-CoA synthetase and DHNA synthase). The OSB-CoA was found to be an unstable intermediate that spontaneously hydrolyzed to the spirodilactone form of OSB (Fig. 6, compound XIX). Further, it was shown that, during the formation of OSB-CoA, ATP was hydrolyzed to AMP and PP_i, which is typical of ligases forming CoA esters (108).

Fig. 6Proposed mechanism of formation of acyl adenylate of OSB (XVIII) and its subsequent conversion to OSB-CoA (X). The conversion of OSB-CoA (X) to spirodilactone of OSB (XIX) is nonenzymatic.

The CoA moiety was suggested to be on the aromatic carboxyl group (16, 108) and evidence in support of this suggestion was obtained (56, 57). However, in subsequent publications, it was reported that the CoA is located on the aliphatic carboxyl group (81, 82).

A group of E. coli mutants responding to DHNA (XII), but not to OSB (IX), for anaerobic growth on glycerol-fumarate medium were analyzed for their ability to convert OSB to DHNA (143). None of the mutant extracts formed DHNA. However, when the cell extracts from different mutants were mixed with each other, one of the mutant extracts complemented with extracts of each one of the other three mutants and formed DHNA. To identify the nature of the enzymatic defect in these mutants, cell extracts from each one of these mutants were complemented with OSB-CoA synthetase and DHNA synthase from M. phlei described above, and assayed for DHNA formation. The single mutant whose extract was complemented by OSB-CoA synthetase, and therefore, lacking this enzyme, was designated menE. The other three mutants, whose extracts were complemented by DHNA synthase, were designated menB (143).

The menE gene was cloned and sequenced (140). The gene was overexpressed, and the enzyme was purified to homogeneity. The purified enzyme had subunits of M_r 49,000 and a native M_r of 185,000. Thus, the native enzyme appears to be a homotetramer. The K_m values for OSB, ATP, and CoA were 16, 73.5, and 360 μM, respectively (88). By chemical inactivation and site-directed mutagenesis studies, an essential histidine residue (His-341) located in the ATP binding region has been identified as necessary for catalytic activity of the enzyme (15). Sequence analysis combined with the fact that OSB-CoA synthetase hydrolyzes ATP to AMP and PP_i and requires CoASH for the reaction earns it membership in the acyl-adenylate/thioester-forming superfamily of enzymes (19, 20). A mechanism for the reaction has been proposed (Fig. 6).

The menB gene was cloned and its complete nucleotide sequence determined (142). When the gene was overexpressed and the protein purified to homogeneity, the subunits were found to have a M_r of 32,000, while the native protein had a M_r of 112,000 as determined by gel filtration. Thus, the enzyme is a homotetramer (106).

As discussed above, the substrate for MenB, OSB-CoA, is highly unstable. Hence, for the assay of MenB, the required OSB-CoA is generated in vitro by coupling the reaction with the MenE reaction (12, 108). However, for the coupled assays, crude cell-free extracts were always used. It is surprising that when the overexpressed and purified MenE and MenB enzymes were used in the coupled assay, DHNA formation was not observed. To determine the reasons for the lack of formation of DHNA, small amounts of a crude cell-free extract of E. coli were added to the reaction mixture, and this resulted in the restoration of activity in the incubation mixture. Hence, it appeared that either a cofactor or another protein might be involved in the reaction.

On the basis of the alignment and analysis of the sequence, MenB was included in the enoyl-CoA hydratase/isomerase (crotonase) superfamily (172). Because of the failure of the purified MenB to form DHNA in the complementation assay discussed above and its membership in the enoyl-CoA hydratase superfamily (where other members form CoA esters), it was suggested that the product of MenB is DHNA-CoA (XI) rather than DHNA (XII) (106). Evidence in support of this prediction has been obtained in M. tuberculosis where the product of MenB was identified by mass spectrometry as DHNA-CoA (XI) (161). The crystal structure of MenB as the native enzyme and in complex with acetoacetyl-CoA and DHNA-CoA, respectively, has been reported by Truglio et al. (161) and Johnston et al. (70). The highly conserved active site of MenB contained a deep pocket lined with Asp-192, Tyr-287, and hydrophobic amino acids. Site-directed mutagenesis studies have established that Asp-192 and Tyr-287 are essential for enzymatic catalysis. On the basis of structural and mutagenesis studies, the authors have proposed a possible mechanism for cyclization of OSB-CoA to DHNA-CoA (161).

On the basis of amino acid sequence homology to thioesterases, an unidentified orf152 (yfbB) was postulated to carry out the conversion of DHNA-CoA → DHNA and designated as menH (106, 121). Evidence in support of the proposal was provided by experimental demonstration of thioesterase activity of the protein (86).

However, as discussed above, the MenH protein has been unequivocally demonstrated to carry out the conversion of newly discovered intermediate SEPHCHC to SHCHC and has been christened as SHCHC synthase (69). Thus, the enzyme responsible for the conversion of DHNA-CoA to DHNA remains to be identified.

Prenylation of DHNA to DMK (Compound XII → III)

The conversion of DHNA (XII) to DMK (III) in extracts of E. coli was shown by Bentley (9). Shineberg and Young (146) were able to isolate a membrane-bound 1,4-dihydroxy-2-octaprenyltransferase. The menA gene encoding the enzyme has been cloned (155). The enzyme (MenA) has many features in common with 4-hydroxybenzoate octaprenyltransferase (UbiA) involved in the biosynthesis of ubiquinone. The two enzymes share a common pool of membrane-bound octaprenyl diphosphate (146). The conversion of DHNA to DMK requires replacement of the carboxyl with the isoprenoid side chain. Prenylation and decarboxylation may occur in a single active site, since symmetry experiments exclude 1,4-naphthoquinone as an intermediate (4). Moreover, there has been no evidence for two separate reaction steps or enzymes. A carbocation mechanism based on the dimethylallyl tryptophan synthase reaction (40) has been proposed for the reaction (106). In addition, a quinol to quinone oxidation is required in which demethylmenaquinol is a likely intermediate; the oxidation to DMK is thought to be spontaneous.

Methylation of DMK to MK (Compound III → II)

DMK (III) is methylated to MK (II) by a methyltransferase, which uses S-adenosylmethionine as the methyl donor. In experiments with whole cells it was shown that all three hydrogen atoms of the methyl group of methionine are transferred to DMK (66). The conversion of DMK-3 to MK-3 was demonstrated by Bryant and Bentley with cell extracts by using S-[¹⁴CH₃]-adenosyl-l-methionine (16). A ubiA mutant of E. coli was found to accumulate DMK but not MK. This mutant is believed to be defective in methylation of DMK to MK, suggesting that this is a double mutant (35). In a subsequent study, it was shown that a ubiE mutant, blocked in the methylation of the ubiquinone biosynthetic intermediate 2-octaprenyl-6-methoxy-1,4-benzoquinol (OMB) (XXV) to 2-octaprenyl-3-methyl-6-methoxy-1,4-benzoquinol (OMMB) (XXVI) (174), accumulated DMK but not MK (171). Consistent with this observation is the simultaneous loss of C-methyltransferase activity toward both OMB and DMK and its restoration by a plasmid containing the ubiE gene (95, 171).

Q BIOSYNTHESIS

The ubiquinone biosynthetic pathway was elucidated largely because of the work of Gibson, Cox, Young, and colleagues (45, 46). In 1964, Cox and Gibson observed that [G-¹⁴C]shikimate was incorporated into ubiquinone, thus establishing that the quinone was derived from the shikimate pathway (25). Gibson and colleagues reasoned that, since ubiquinone is required for aerobic electron transport, mutants deficient in its biosynthesis would grow fermentatively on glucose, but not aerobically on oxidizable substrates such as malate or succinate, as the sole source of carbon and energy. Mutagenized cultures were screened for the desired phenotype and potential mutants were analyzed for the presence or absence of ubiquinone (45). By use of this procedure, a number of mutants were isolated and it was found that these mutants accumulated sufficient amounts of intermediates so that their structure could be determined by mass spectrometry and magnetic resonance spectrometry (45, 154).

The biosynthesis of the quinonoid ring and the various ring modification reactions in E. coli have been reviewed (71, 73, 105, 152). A mechanistic perspective on the various reactions has been provided (6, 106). As pointed out in the introduction, in E. coli and Salmonella, the first committed step in the biosynthesis of Q is the formation of 4-hydroxybenzoate from chorismate by the cytoplasmic enzyme chorismate lyase. The 4-hydroxybenzoate formed is attached to the membrane-bound octaprenyl diphosphate by a membrane-bound octaprenyltransferase. For the subsequent reactions, all the substrates and enzymes are in a membrane-bound complex. In this chapter, the intermediates, genes, and enzymes involved in the various reactions are presented first. This will be followed by a brief description of the membrane-bound multienzyme complex and the reported interactions of certain enzymes in the complex with each other.

Conversion of Chorismate to 4-Hydroxybenzoate (Compound IV → XX)

The elimination of pyruvate from chorismate (IV) results in the formation of 4-hydroxybenzoate (4-HB) (XX) (Fig. 7). This aromatizing reaction is the first committed step in the biosynthesis of Q and is catalyzed by the enzyme, chorismate lyase, encoded by the ubiC gene (94). The ubiC gene has been cloned; the enzyme was overexpressed and purified to homogeneity. The UbiC is a monomer of 165 amino acids from which the N-terminal methionine is posttranslationally removed resulting in the functional enzyme. The enzyme has a molecular weight of 18,800 and functions as a monomer. The K_m was reported to be around 6 to 10 μM (117, 147). The purified enzyme failed to accept isochorismate as a substrate, but did convert 4-amino-4-deoxychorismate to 4-aminobenzoate (117). Thus, it appears that the enzyme is unable to distinguish between the hydroxyl group and the amino group at the C-4 position. Walsh et al. (165) have proposed a 1,2-elimination of the elements of pyruvate for the aromatization similar to that of anthranilate synthase reaction. The C-4 H of chorismate is abstracted by the enzyme and loss of the C-3-enolpyruvyl group then results in the formation of the 4-HB. It has been reported that the enzyme was inhibited by 4-HB but not pyruvate (147). In a subsequent study, Holden et al. circumvented the rapid influence of product inhibition on the initial reaction rate by using progress curve analysis of stopped-flow kinetic measurements. Under these conditions the K_m increased by about 3-fold to 29 μM (60). The enzyme releases the pyruvate quickly and retains the 4-HB with a 10-fold higher affinity (K_p = 2.1 μM) (39).

Fig. 7Ubiquinone biosynthetic pathway. Each compound in the pathway is identified by a Roman numeral. Under anaerobic conditions, there are alternate hydroxylases for the three enzymes incorporating molecular oxygen (UbiB, UbiH, and UbiF). It should be noted that, in compound (XXI), the chemical numbering system locates the prenyl side chain at the C-3 carbon; in compound (XXII) and subsequent intermediates, the prenyl side chain is assigned to C-2. Compounds (XXV), (XXVI), and (XXVII), are drawn in the quinol form. Some authors draw these structures in the quinone form. For other abbreviations, see the legend to Fig. 4. The chemical names for the intermediates of the pathway are as follows: (IV), chorismate; (XX), 4-hydroxybenzoate; (XXI), 3-octaprenyl-4-hydroxybenzoate; (XXII), 2-octaprenylphenol; (XXIII), 2-octaprenyl-6-hydroxyphenol; (XIV), 2-octaprenyl-6-methoxyphenol; (XXV), 2-octaprenyl-6-methoxy-1,4-benzoquinol; (XXVI), 2-octaprenyl-3-methyl-6 methoxy-1,4-benzoquinol; (XXVII), 2-octaprenyl-3-methyl-5-hydroxy-6-methoxy-1,4-benzoquinol; (XXVIII), ubiquinol; (I), Q, ubiquinone. The conversion of (XXVIII) to (I) is thought to be nonenzymatic.

The crystal structure of UbiC protein has been solved. The wild-type enzyme tended to aggregate and precipitate even in the presence of reducing agents and salt. To circumvent this problem, two surface-accessible cysteines at sequence positions 14 and 81 were converted by site-directed mutagenesis into serine. This mutant enzyme C14S/C81S, designated as CCSS, showed greatly improved solubility and stability with minimal effect on the catalytic properties (60). The crystal structure of the enzyme from the double mutant at 1.4 Å and the wild-type enzyme at 2.0 Å in complex with the product, 4-HB, was determined. The core of the chorismate lyase consisted of six-stranded antiparallel β-sheet without spanning helices and novel connectivity. The product, 4-HB, was shown to be bound in an internal cavity behind two flaps which completely covers and shields the product from the solvent. Three hydrogen bonds link the product to the internally charged side chains of Arg-76 and Glu-155 and two additional hydrogen bonds link it to the flap atoms 34 N and 114 N. These five hydrogen bonds play a direct role in binding the product. There are three additional hydrogen bonds that link the flaps together and further enhance product retention (150).

To further clarify and understand the unusual ligand binding and the mechanism of reaction, additional structures of mutant enzymes, enzyme inhibitor complexes, and mutant enzyme inhibitor complexes were studied. When a high-resolution crystal structure (1.0 Å) of the enzyme substrate complex was examined, a substrate-sized internal cavity was found behind flaps near the product binding site. The crystal structure (2.4 Å) of the enzyme complexed with the inhibitor vanillate showed that the flaps were partly opened when compared with the product bound enzyme.

An active-site mutant enzyme G90A complexed with the product was examined at a resolution of 2.0 Å. It was found that the presence of the additional methyl group in the mutant enzyme resulted in the enlargement of the 4-HB binding pocket by about 1 Å. However, all the 8-hydrogen bonds involved in product binding in the wild-type enzyme are maintained. When the kinetic properties of the mutant enzyme were compared with the wild-type enzyme it was found that the product inhibition increased by about 40%. The wild-type enzyme had a K_p value of 1.5 (± 0.2) μM versus 0.9 (± 0.1) μM for the mutant. The increase in product inhibition in the mutant is attributed to the presence of the additional methyl group acquired in the conversion of glycine to alanine. The K_m values did not change, while the K_cat value of the mutant decreased to 0.9 (± 0.4) s⁻¹ from 1.4 (± 0.2) s⁻¹.

When the G90A mutant enzyme was bound with the inhibitor vanillate, the structure at 1.9 Å showed two vanillate molecules. One of the vanillate molecules occupied the product site normally occupied by 4-HB, and the second molecule of vanillate occupied an adjacent site or cavity. The two sites were found to be connected by a tunnel that is partly open on both ends. The product binding site was designated as the primary ligand site, and the adjacent site (additional site), where the second vanillate molecule binds, was designated as the secondary ligand site (150). Based on the structural studies summarized above in combination with molecular modeling, molecular dynamics, and binding measurements with inhibitors, a model has been proposed to account for catalytic, product binding, and product release mechanisms.

It has been proposed that the enzyme operates by a two-site or tunnel mechanism (150). According to this mechanism, the enzyme contains bound 4-HB in the primary site (designated as primary ligand site, 1°). When the substrate binds to the second site (designated as secondary ligand site, 2°), it promotes the release of the product from the primary site. As the product 4-HB is released from the primary site, the substrate chorismate moves to the primary site. In the primary site the substrate is unstable and it is rapidly converted to the products 4-HB and pyruvate. Since pyruvate is small, it exits rapidly from the primary site while the 4-HB is retained in the bound state and the process is repeated in a cyclic manner.

Prenylation of 4-Hydroxybenzoate (Compound XX → XXI)

The prenylation of 4-hydroxybenzoate (XX) to 3-octaprenyl-4-hydroxybenzoate (XXI) is carried out by the enzyme 4-hydroxybenzoate octaprenyltransferase encoded by the ubiA gene. The enzyme is membrane bound and requires octaprenyl diphosphate and Mg²⁺ (175). In addition to octaprenyl diphosphate, the enzyme could incorporate geranyl, farnesyl, phytyl, or solanesyl diphosphate as a side-chain precursor (35, 111). This lack of specificity also extends to the aromatic substrate; thus, 4-aminobenzoate can replace 4-hydroxybenzoate as a substrate (35). Recently, it has been shown that the enzyme accepts a wide variety of benzoic acid derivatives as substrates. As already mentioned, replacing the C-4 hydroxyl with an amino group did not affect reactivity. However, replacing the hydroxyl with a methoxy group was not tolerated. Compounds substituted at C-5 with OH, NH₂, Cl, or CO-CH₃ groups were used as substrates by the enzyme. Similarly, compounds with hydroxyl groups at C-4, C-5, and C-6 or hydroxyl group at C-4, C-6, and methyl group at C-5 were substrates (169).

The prenyl transfer reactions are electrophilic substitution reactions. The reaction mechanism probably includes a carbocation (43); evidence for this proposal comes from studies on the related enzyme dimethylallyltryptophan synthase (40).

Formation of 2-Octaprenylphenol (Compound XXI → XXII)

The conversion of 3-octaprenyl-4-hydroxybenzoate (XXI) to 2-octaprenylphenol (XXII) was demonstrated by Cox et al. (26). The enzyme responsible for this conversion was named 3-octaprenyl-4-hydroxybenzoate decarboxylase. The presence of decarboxylase was also observed by El Hachimi et al. (35). The enzyme activity was absent in ubiD mutants (26).

When cell extracts were prepared using a French press, centrifuged at 30,000 × g, and the supernatant further centrifuged at 150,000 × g for 3 h, most of the activity remained in the soluble fraction, establishing that the enzyme separated from the membrane. A 24-fold-purified preparation of the enzyme was obtained. The molecular weight of the enzyme was reported to be M_r 340,000 (98). For optimal activity, the enzyme required Mn²⁺, washed membranes, or an extract of phospholipids, and an unidentified heat-stable factor of molecular weight less than 10,000. The reaction was strongly stimulated by dithiothreitol and methanol. Since the substrate of the enzyme 3-octaprenyl-4-hydroxybenzoate is membrane bound and the enzyme is stimulated by phospholipid, it has been suggested that the enzyme normally functions in association with the cytoplasmic membrane in vivo (98). A reaction mechanism has been suggested (6, 106).

A number of ubiD mutants studied form about 20% of the wild-type levels of Q, indicating that the mutants are leaky or there is an alternate enzyme for the reaction. However, the significance of any alternate carboxylyase in wild-type strains has been questioned (98).

An alternate 3-octaprenyl-4-hydroxybenzoate decarboxylase encoded by the ubiX gene has been described in S. enterica serovar Typhimurium (S. typhimurium) that carries out the same reaction as the ubiD-encoded enzyme (61). A ubiX gene showing 70% homology to the S. enterica serovar Typhimurium) (S. typhimurium) gene has been identified in E. coli (118, 177). The orfs encoding the two enzymes UbiD and UbiX have been identified from E. coli (178). Recently, a report has appeared suggesting that both UbiD and UbiX are required for the decarboxylation of 3-octaprenyl-4-hydroxybenzoate, in particular, during logarithmic phase of growth (49).

It has been reported that several E. coli strains, including the enterohemorrhagic O157:H7, contain, in addition to UbiX, a second paralog, designated Pad1. The amino acid sequence of this paralog was reported to have a 52% identity to UbiX and a slightly higher identity to Saccharomyces cerevisiae phenylacrylic acid decarboxylase Pad1. The exact biochemical role of E. coli Pad remains to be determined (129).

Hydroxylation and Methylation Reactions

In the subsequent steps of the pathway, the 2-octaprenylphenol undergoes three hydroxylation reactions, alternating with three methylation reactions, resulting in the formation of ubiquinol (XXVIII) and then Q (I). For convenience, the hydroxylation reactions are considered together, and this will be followed by a description of the three methylation reactions.

Hydroxylation reactions.

Three flavin-linked monooxygenases are involved in the three hydroxylation reactions of the pathway with three hydroxyl groups being introduced at positions C-6, C-4, and C-5 of the benzene nucleus, respectively. The three reactions are:

Mutants blocked in each of these hydroxylation reactions were isolated and designated ubiB, ubiH, and ubiF, respectively.

Consistent with their metabolic block, ubiB mutants accumulate 2-octaprenylphenol (XXII) (26, 176). However, the predicted product of the UbiB reaction (XXIII) has never been isolated and characterized, and it may not occur as a free intermediate (2, 176).

As part of the genome project, when the sequence was annotated, ubiB was considered identical to that of fre and luxG (28). In subsequent studies, an orf previously designated yigR was identified as ubiB. An insertion mutant was isolated and was shown to accumulate the expected intermediate in the pathway, 2-octaprenylphenol (compound XXII). As mentioned above, the expected product of the reaction 2-octaprenyl-6-hydroxyphenol (XXIII) could not be isolated (125).

Mutants blocked in the methylation of 2-octaprenyl-6-hydroxyphenol (XXIII) to 2-octaprenyl-6-methoxyphenol (XXIV) have been isolated (ubiG::kan) (see “Methylation reactions,” below). However, surprisingly, these mutants also failed to accumulate the expected intermediate before the block 2-octaprenyl-6-hydroxyphenol (XXIII), thus supporting the suggestion that it may not occur as a free intermediate (2, 176).

Mutants unable to convert (XXIV) to (XXV) have been isolated with the gene being designated as ubiH (176). The ubiH gene is identical to the visB gene and confers a photosensitive phenotype due to the accumulation of (XXIV) (115).

The final hydroxylation in Q biosynthesis is the conversion of (XXVI) to (XXVII) and mutants blocked in the reaction were isolated and characterized. As expected, these mutants, designated as ubiF, accumulated (XXVI) which was isolated and identified (174). The ubiF gene was identified as orf391 and the product accumulated by insertion mutants in this orf was found to be (XXVI) (93).

Under aerobic conditions, the origin of the oxygen atoms of Q was determined by ¹⁸O-labeling experiments. Cultures were grown on the oxidizable carbon source succinate, under strictly aerobic conditions in a defined atmosphere of ¹⁸O₂. The Q was isolated from these cultures and subjected to mass spectral analysis. The spectrum showed several prominent peaks with m/z values differing from that of normal Q by +6, establishing that ¹⁸O had been incorporated. Further, it was demonstrated that the ¹⁸O was incorporated at positions 4, 5, and 6 (2).

The nature of the hydroxylation reactions discussed above has been investigated. A hemA mutant defective in the biosynthesis of cytochromes was able to convert 2-octaprenyl-[¹⁴C]phenol to ¹⁴C-labeled Q-8, ruling out the involvement of the cytochrome P-450 monooxygenase system, and suggesting the involvement of flavin-linked monooxygenases in these reactions (75). A mechanism analogous to that proposed for the flavin-dependent tyrosine hydroxylase (166) has been suggested by Begley et al. (6).

When grown anaerobically, with glycerol as a carbon source and fumarate as an electron acceptor, E. coli forms considerable quantities of Q (50 to 70% of aerobically grown cells). Mutants blocked in the various nonhydroxylating reactions of the pathway such as ubiA, ubiD, and ubiE, remain Q deficient under both aerobic and anaerobic conditions, establishing that the same genes and enzymes participate under both aerobic and anaerobic conditions (1).

In contrast, the three groups of mutants blocked in the three oxygenases discussed above, ubiB, ubiH, and ubiF, were able to synthesize Q under anaerobic conditions providing evidence that specific hydroxylases are involved in the anaerobic pathway (1). These hydroxylases likely derive the hydroxyl groups from the solvent H₂O similar to EntC and MenF reactions discussed above.

Methylation reactions.

Two methylations on O and one on C involved in the pathway are:

The methylation steps alternate with the three hydroxylations described above introducing methyl groups at the 6-OH, the ring C-3, and the 5-OH group, respectively. The three methyl groups are derived from methionine (66), with S-adenosylmethionine being the actual methyl donor.

The C-methylase responsible for the methylation of ring C-3 is encoded by the ubiE gene. Mutants blocked in the methylation accumulate the substrate of the enzyme, 2-octaprenyl-6-methoxy-1,4-benzoquinol (XXV) (174). The UbiE enzyme is nonspecific and carries out the methylation of the menaquinone intermediate, DMK (III) → MK (II) in addition to its role in the methylation of (XXV) → (XXVI) (95) (discussed in “MK Biosynthesis,” above).

During the screening for mutants blocked in the O-methylation reactions, mutants blocked in the methylation of 6-OH were not obtained. However, mutants blocked in the methylation of the 5-OH were isolated, designated as ubiG, and shown to accumulate compound (XXVII), which was isolated and characterized (154). Further, the ubiG mutants being leaky formed about 10% of the wild-type levels of Q (1). In subsequent studies, it was reported that the O-methylase encoded by the ubiG gene is nonspecific and that it carries out the methylation of both 6-OH and 5-OH groups (62). This lack of specificity also extends to the presence of other groups on the benzoquinone ring; the enzyme, in addition, methylates 3,4-dihydroxy-5-hexaprenylbenzoquinol to 3-methoxy-4-hydroxy-5-hexaprenylbenzoquinol. The reported leakiness of the ubiG mutant, mentioned above, probably allowed sufficient intermediate (XXIII) to be methylated at the 6-OH resulting in the formation of (XXIV), which was subsequently converted to (XXVII) and methylated at the 5-OH, resulting in the formation of (XXVIII) and Q. A ubiG::kan mutant has been isolated. However, surprisingly, this mutant failed to accumulate the expected intermediate before the block 2-octaprenyl-6-hydroxyphenol (XXVIII). Two possible reasons have been advanced for the failure to detect compound (XXVIII). First, as mentioned above, compound (XXVIII) may not occur as a free intermediate (2, 176). Second, it has been suggested that the compound may be highly reactive due to the presence of the catechol moiety, and hence degraded (62).

Organization of Q Biosynthetic Enzymes into a Complex

As discussed above, not all the enzymes involved in Q biosynthesis have been studied in cell-free extracts. Among the enzymes studied, chorismate pyruvate-lyase (UbiC) is a cytoplasmic enzyme, while 4-hydroxybenzoate octaprenyltransferase (UbiD) is firmly membrane bound. Two other enzymes that have been studied, 3-octaprenyl-4-hydroxybenzoate carboxy-lyase and 2-octaprenyl-3-methyl-5-hydroxy-6-methoxy-1,4-benzoquinol methyltransferase, are considered to be normally associated with the membrane (97, 98). The association of enzymes with membrane is supported by the isolation of a 2-octaprenyl [U-¹⁴C]phenol (XXII) charged enzyme complex of molecular weight M_r of 2 × 10⁶ containing at least 12 proteins ranging from 40,000 to 80,000 M_r from cells grown anaerobically on glycerol/fumarate medium in the presence of 4-hydroxy[U-¹⁴C]benzoate. When this complex was incubated with S-adenosylmethionine, NADH, NADPH, Mg²⁺, and a cytoplasmic enzyme of a molecular weight of about 20,000 (probably a methyltransferase) (77) in the presence of oxygen, all of the ¹⁴C-labeled phenol was converted to Q (76). This complex, therefore, contains the oxygen-dependent Q-8 biosynthetic apparatus. In anaerobically grown cells, this apparatus which is charged with 2-octaprenylphenol may be kept in a standby position. When oxygen becomes available, Q-8 biosynthesis can be effectively turned on (76, 77). Since this complex was isolated without detergent treatment, it was thought that it had broken from the membrane as a distinct and native domain. This complex contains, in addition to a high level of 2-octaprenylphenol and low levels of Q, phospholipid and other membrane proteins (76, 77).

Based on studies with a thiol-sensitive mutant (IS16), it was reported that there is genetic evidence for interaction between UbiX and UbiG proteins (50). The IS16 mutant had point mutations resulting in change of a single amino acid in UbiX (S98R) and UbiG (L132Q) when compared with the sequence of the same two proteins in E. coli K-12. Complementation of this mutant with either ubiX from E. coli K₁₂ strain (ubiX K-12) or ubiG K₁₂ restored the wild-type phenotype. In contrast, while an ubiG insertion mutant was rescued by complementation by ubiG K₁₂ it was not rescued by ubiX K-12 (50). Rescue of Q-deficient phenotypes can be achieved by levels of Q that are significantly lower than those present in the wild-type strains (50). These studies were cited as providing supporting evidence for the polypeptide complex described by Knoll (77) discussed above.

Regulation of Q Biosynthetic Genes

It is known that the quinone composition of E. coli is influenced by the availability of oxygen. Cells grown under vigorous aeration contain 2- to 3-fold higher concentrations of Q compared with DMK and MK. Under anaerobic conditions, the MK and DMK concentrations increase 2- to 3-fold while the concentration of Q decreases (12, 152). The mechanism of this regulation is not completely understood. Shestopalov et al. (145) have shown that chloramphenicol had no effect on these changes, suggesting post-translational regulation of quinone levels. Further, these authors have shown that mutations in the regulatory systems of Fnr and Arc had no effect on the quinone pool. Suzuki et al. (156) studied the regulation of the ubiA gene by using plasmid-borne lacZ fusions and showed that the gene is catabolite repressed by glucose. A similar study on plasmid-borne ubiG gene also showed glucose catabolite repression (44). Søballe and Poole (151) studied the transcriptional regulation of ubiC-lacZ in a monolysogen and showed that the expression was higher aerobically than anaerobically. It was further reported that glucose repressed expression, while anaerobic growth in the presence of alternate electron acceptors, nitrate, and fumarate did not affect expression. Further, it was shown that ubiC was negatively regulated by transcriptional regulators Fnr and IHF (151).

In a recent study, the expression of the operon fusions ubiC′-lacZ⁺, ubiCA′-lacZ⁺, and ubiA′-lacZ⁺ were studied. In glycerol media under aerobic conditions the highest level of expression was observed with the operon fusion ubiC′-lacZ⁺. Compared with the ubiC′-lacZ⁺, the ubiCA′-lacZ⁺ operon fusion showed 26% of the activity while the ubiA′-lacZ⁺ operon fusion had an activity of 1%. Thus, the ubiC gene is regulated by the upstream promoter, while the ubiA gene lacks its own promoter (90). The effect of fermentable and oxidizable carbon sources on the expression of ubiC′-lacZ⁺ was determined. The expression was low in the case of fermentable carbon source glucose; increasing glucose concentration resulted in increased repression. In the presence of oxidizable carbon sources the expression increased 2- to 3-fold. In both fermentable and oxidizable carbon sources, supplementation of the medium with casamino acids resulted in decrease in expression. Aerobically, deficiency in both Q and MK or MK alone resulted in a 2-fold increase in expression compared with wild-type cells. In the strain carrying the arcA mutation, under anaerobic conditions the expression was from 25% to 50% higher than the anaerobically grown wild-type strain, while in the fnr mutant the activities did not change (90). The lack of regulation by FNR is in agreement with the absence of binding site (139). In the case of the narXL mutant, the activity increased 50% anaerobically and 137% in the presence of NO₃⁻ In the presence of other electron acceptors, O₂, fumarate, and TMAO, the activities were from 70% to 90% higher than that of the wild-type (90).

The expressions of the two genes involved in the decarboxylation of 3-octaprenyl-4-hydroxybenzoate, ubiD and ubiX were studied using LacZ operon fusions. During aerobic growth the expression of both genes depended on the carbon source: succinate > glycerol > glucose. Mutations in fnr, arcA or hemA increased the expressions of both genes. During anaerobic growth in LB medium glucose strongly repressed the expression of ubiD but not ubiX (178).

Consequences of Mutations in ubi Genes

Certain pleotropic properties which will be of value in isolating and/or characterizing mutants are described here. As described above, ubi mutants were isolated by their inability to utilize succinate or other reduced compounds as carbon sources.

Hypersensitivity to thiols such as dithiothreitol (DTT), 2-mercaptoethanol, and 1-thioglycerol was demonstrated in ubiX and ubiD mutants (177). Subsequently, it was shown that a ubiCA insertion mutant also exhibited this property (153). Thiol sensitivity is likely a common property of all Q-deficient strains since a respiratory chain is essential for the maintenance of the disulfide bond-forming system (78, 79).

An E. coli mutant resistant to the antibiotic and antitumor agent phleomycin was isolated. The mutant was also found to be resistant to bleomycin and unable to grow on succinate as the sole source of carbon and resistant to the lethal effects of heating at 52°C. The suc⁻ phenotype and mapping data led to the conclusion that the mutant was defective in the ubiF gene. To confirm the observed properties, known ubiA, ubiD, and ubiF mutants were compared with the newly isolated mutants. It was found that they also exhibited these properties (24). Recently, an ubiCA mutant was shown to exhibit the pleotrophic phenotype, being resistant to heat, linolenic acid, and phleomycin. In addition, it has been shown that Q is involved in superoxide scavenging, and in protection against oxidative stress mediated by CuSO₄ or H₂O₂ (153).

A mutant showing partial resistance to streptomycin was found to be defective in the ubiF gene. Membranes of this strain accumulated 2-octaprenyl-3-methyl-6-methoxy-1,4-benzoquinol (XXVI) but not Q. A previously characterized ubiF mutant was found to show reduced uptake of gentamycin. At present, there is no evidence implicating Q in aminoglycoside antibiotic uptake and these observations are attributed to the general impairment of respiratory capacity (114).

Mutations in the Q biosynthetic pathway (ubiD, ubiB, and ubiG) led to the lack of flagellar synthesis and motility (5, 58). A ubiA men⁺ strain was motile anaerobically and nonmotile aerobically, while mutants blocked in Q and MK were found to be nonmobile under both aerobic and anaerobic conditions. Thus, it appears that a functional electron transport system is essential for motility and flagellar synthesis.

Mutants lacking Q, MK, or both have been isolated, and the role of quinones in electron transport to oxygen and nitrate has been studied (164).

Functions of Isoprenoid Quinones MK and Q

The roles of MK in the anaerobic respiratory chains and Q in the aerobic respiratory chains are well established. For details on the role of the quinones, other reviews and EcoSal chapters should be consulted (32, 41, 51, 52, 64, 73, 123, 124, 131, 152). EcoSal chapter 3.2.2 covers the aerobic and anaerobic metabolism and respiratory chains.

Biosynthesis of Isoprenoid Side Chain of MK and Q

E. coli and other gram-negative bacteria synthesize the isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP) by the mevalonate-independent pathway also known as the non-mevalonate pathway (other names are deoxyxylulose phosphate and methylerythritol phosphate pathways). The IPP condenses with DMAPP and the prenyl chain is elongated to the 40-carbon octaprenyl diphosphate (OPP). The details of side chain biosynthesis are beyond the scope of this chapter. A number of reviews are available on the topic (21, 34, 37, 38, 63, 87, 106, 107, 135, 136, 137, 138, 149, 158).

Acknowledgments

We express our sincere thanks to Professor Ronald Bentley for the critical reading of the manuscript and to Jim Bruno for his assistance in the preparation of the manuscript. My thanks are also due to many past and present members of my research group whose names appear in the references cited. Research in the laboratory of R.M. over the years was supported by Public Health Service Grants from the National Institutes of Health.

Research in the laboratory of O.K. was supported by Korea Science and Engineering Foundation Grant R01-2005-000-10288-0, a Molecular and Cellular Biodiscovery Research Program grant, and a 21st Century Frontier R&D Program grant from the Korea Ministry of Science and Technology.