Biosynthesis and applications of prenylquinones

Abstract

Prenylquinones are isoprenoid compounds with a characteristic quinone structure and isoprenyl tail that are ubiquitous in almost all living organisms. There are four major prenylquinone classes: ubiquinone (UQ), menaquinone (MK), plastoquinone (PQ), and rhodoquinone (RQ). The quinone structure and isoprenyl tail length differ among organisms. UQ, PQ, and RQ contain benzoquinone, while MK contains naphthoquinone. UQ, MK, and RQ are involved in oxidative phosphorylation, while PQ functions in photosynthetic electron transfer. Some organisms possess two types of prenylquinones; Escherichia coli has UQ₈ and MK₈, and Caenorhabditis elegans has UQ₉ and RQ₉. Crystal structures of most of the enzymes involved in MK synthesis have been solved. Studies on the biosynthesis and functions of quinones have advanced recently, including for phylloquinone (PhQ), which has a phytyl moiety instead of an isoprenyl tail. Herein, the synthesis and applications of prenylquinones are reviewed.

Distribution of prenylquinones in living organisms

Keywords:

• Ubiquinone
• menaquinone
• plastoquinone
• rhodoquinone
• phylloquinone

Abbreviations:

• DHFL: dehypoxanthinylfutalosine
• DHNA: 1,4-dihydroxy-2-naphthoate
• DMK: demethylmenaquinone
• DMAPP: dimethylallyl pyrophosphate
• DXP: 1-deoxy-D-xylulose-5-phosphate
• FPP: farnesyl pyrophosphate
• GPP: geranyl pyrophosphate
• GGPP: geranylgeranyl pyrophosphate
• HGA: homogentisate
• IPP: isopentenyl pyrophosphate
• MEP: 2C-methyl-D-erythritol-4-phosphate
• MK: menaquinone
• MVA: mevalonate
• PDS: prenyl diphosphate synthase
• PHB: p-hydroxybenzoate
• PhQ: phylloquinone
• PQ: plastoquinone
• RQ: rhodoquinone
• UQ: ubiquinone

Isoprenoids are compounds built from two common precursors, isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP). More than 50,000 isoprenoid compounds are found in nature [1]. Among them, isoprenylated quinones, in which the length of the isoprenoid side chain or tail varies, are widely distributed in almost all living organisms, and they function in electron transfer. Living organisms must acquire energy through oxidative phosphorylation or photosynthetic phosphorylation, and these processes require lipid molecules to transfer electrons and protons between protein complexes. Typically, ubiquinone (UQ) transfers electrons from Complex I or II to Complex III in oxidative phosphorylation, while plastoquinone (PQ) transfers electrons from photosystem II to the cytochrome b₆f complex in photosynthesis [2].

The isoprenoid side chain is responsible for the lipid-soluble nature of quinones, and anchors them in membrane lipid bilayers, while the electron transfer capacity is derived from the quinone head. The quinone ring undergoes a two-step reversible oxidation/reduction between reduced and oxidized forms. This common property allows electrons and protons to shuttle between different protein complexes in biological membranes, allowing it to function as both a cofactor in enzyme reactions, and as an antioxidant.

Widely distributed (major) and more restricted (minor) quinones are present in living organisms. UQ, menaquinone (MK), PQ, and rhodoquinone (RQ) are major quinones. UQ and RQ are distributed in prokaryotes and eukaryotes, while MK is found in bacteria and archaea, and PQ is restricted to cyanobacteria and plants. Minor quinones include thermoplasmaquinone, methionaquinone, chlorobiumquinone, sulfolobusquinone, and caldariellaquinone, and are found in bacteria and archaea [3].

The length of the isoprenoid side chain and the type of quinone are variable (side chain length is annotated in subscript in this review). For example, bacteria such as Bacillus subtilis produce MK₇, Escherichia coli synthesize UQ₈ and MK₈, and Synechocystis spp. generate PQ₉. Yeasts such as Saccharomyces cerevisiae and Schizosaccharomyces pombe produce UQ₆ and UQ₁₀, respectively. Plants such as Arabidopsis thaliana produce UQ₉ and PQ₉, while Nicotiana tabacum synthesize UQ₁₀ and PQ₉. Nematodes such as Caenorhabditis elegans produce UQ₉ and RQ₉, and higher animals such as Mus musculus and Homo sapiens make UQ₉ and UQ₁₀, respectively (Figure 1). The types of prenylquinones in organisms are highly variable; hence they have been used for classification of microbes [4,5].

Figure 1. Distribution of four major prenylquinones in different organisms. Ubiquinone (UQ) is present in almost all living organisms from bacteria to higher eukaryotes. Menaquinone (MK) is distributed in bacteria and archaea, and therefore considered the oldest type of prenylquinone, first synthesized in primitive living organisms. Plastoquinone (PQ) occurs in cyanobacteria, and was presumably subsequently transferred to plants. Rhodoquinone (RQ) is the most recently evolved quinone, and is synthesized from UQ.

The biosynthesis of prenylquinones has been extensively studied, and despite significant knowledge accumulated, some biosynthetic reactions remain poorly understood. In this review, the biosynthesis of four major prenylquinones and phylloquinone (PhQ) is summarized in detail.

Isoprenoid side chains are synthesized via 2C-methyl-D-erythritol-4-phosphate (MEP) and mevalonate (MVA) pathways

The isoprenoid side chains of prenylquinones are synthesized by prenyl diphosphate synthase (PDS) from DMAPP, geranyl pyrophosphate (GPP), geranylgeranyl pyrophosphate (GGPP), or farnesyl pyrophosphate (FPP) by condensation of IPP. IPP and DMAPP are synthesized from either the 2C-methyl-D-erythritol-4-phosphate (MEP) pathway [6–8] or the mevalonate (MVA) pathway [9].The MEP pathway is present in most prokaryotes, and the MVA pathway occurs in archaea and eukaryotes [10]. Plants and Streptomycetes possess both pathways.

The MVA pathway was discovered in the 1960s and consists of seven enzyme-catalyzed reactions. It performs several key functions within cells, and is an important central metabolic pathway in all higher eukaryotes. The MVA pathway of S. cerevisiae is shown in Figure 2 as representative of eukaryotes. Formation of acetoacetyl-CoA from two acetyl-CoA molecules by acetyl-CoA acetyltransferase (Erg10) is followed by the synthesis of 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) by Erg13. In the third step, MVA is generated by reduction of HMG-CoA by HMG-CoA reductase (Hmg), the target of the famous ‘statin’ drugs [11]. MVA is phosphorylated by Erg12 to generate phosphomevalonate, and further phosphorylated by Erg8. Finally, diphosphomevalonate is used by Erg19/Mvd1 to generate IPP or DMAPP. Idi isomerizes between IPP and DMAPP, and DMAPP and IPP are further utilized in condensation reactions for the biosynthesis of isoprenoids. The reactions of all enzymes in the MVA pathway and their three-dimensional structures have been summarized previously [9].

Figure 2. The mevalonate (MVA) pathway in Saccharomyces cerevisiae. The MVA pathway consists of seven enzyme-catalyzed reactions. The first step is the formation of acetoacetyl-CoA from two acetyl-CoA molecules by Erg10 (acetyl-CoA acetyl transferase). Subsequently, Erg13 (HMG-CoA synthase), Hmg1/Hmg2 (HMG-CoA reductase), Erg12 (mevalonate kinase), Erg8 (phosphomevalonate kinase), and Erg19 (diphosphomevalonate decarboxylase) lead to the production of isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP). Idi1 isomerizes between IPP and DMAPP.

The MEP pathway for the biosynthesis of IPP and DMAPP was discovered in the 1990s and consists of eight enzyme-catalyzed reactions. The MEP pathway of E. coli is shown in Figure 3. Condensation of pyruvate and D-glyceraldehyde-3-phosphate to form 1-deoxy-D-xylulose-5-phosphate (DXP) is catalyzed by Dxs (Figure 3). The second step is the conversion of DXP to MEP by IspC, and 4-diphosphocytidyl-2C-methyl-D-erythritol (CDP-ME) is then generated from MEP and CTP by IspD. The fourth step is the phosphorylation of CDP-ME to generate 4-diphosphocytidyl-2C-methyl-D-erythritol-2-phosphate (CDP-ME2P) by IspE, and IspF subsequently removes CMP from CDP-ME2P to generate 2C-methyl-D-erythritol-2,4-cyclodiphosphate (MEcPP). In the sixth step, IspG catalyzes the ring opening of the cyclic pyrophosphate and C3-reductive dehydration of MEcPP to generate 4-hydroxy-3-methylbut-2-enyl diphosphate (HMB-PP). Finally, IspH generates IPP or DMAPP from HMB-PP by reduction. Idi catalyzes isomerization between IPP and DMAPP in the eighth step, and some organisms lack this enzyme. Enzymes of this MEP pathway are attractive targets for the development of drugs against infectious diseases because this pathway occurs in pathogenic prokaryotes but is absent in humans. The antimalarial drug fosmidomycin, which inhibits Dxr, is one of the best-known examples of a drug that targets the MEP pathway. Three-dimensional structures of E. coli Dxs, IspC, IspD, IspE, IspF, IspG, and IspH have been solved, and their precise structure-based catalytic mechanisms have been described [6]. Further details about these biosynthetic pathways synthesizing IPP can be found in previous reviews [6–10].

Figure 3. The 2C-methyl-D-erythritol-4-phosphate (MEP) pathway in Escherichia coli. The MEP pathway consists of eight enzyme-catalyzed reactions. The first step is the condensation of pyruvate and glyceraldehyde-3-phosphate to form 1-deoxy-D-xylulose-5-phosphate by Dxs (DXP synthase). Subsequent steps catalyzed by IspC (DXP reductoisomerase), IspD (MEP cytidyltransferase), IspE (CDP-ME kinase), IspF (MECDP synthase), IspG (4-hydroxy-3-methylbut-2-enyl diphosphate synthase), and IspH (4-hydroxy-3-methylbut-2-enyl diphosphate reductase) lead to the production of IPP and DMAPP.

Prenyl diphosphate synthase

The side chain of prenylquinone is supplied by polyprenyl diphosphate synthase (PDS) and determines the side chain length of prenylquinones (Figure 4) [12,13]. Numerous PDS enzymes have been analyzed [5,14–17], and all consist of seven conserved regions including two DDXXD motifs involved in binding substrates such as FPP (GPP or GGPP) and IPP. PDS occurs in both homomeric and heteromeric forms. PDSs in Gram-negative bacteria are mostly heteromeric, while those in Gram-positive bacteria are mostly heteromeric [15,18]. Eukaryotes have both homomeric (e.g. Coq1 and SPS1) and heteromeric types (e.g. human PDSS1 and PDSS2) [19–21]. The distribution of these PDSs differs among organisms, and among different components in organisms. For example, E. coli contains only one polyprenyl diphosphate synthase (IspB) and shares the side chain of UQ and MK [22,23]. Meanwhile, plants such as Arabidopsis possess three different PDSs, in this case three solanesyl diphosphate synthases (SPS1, SPS2, and SPS3), which are localized to different subcellular organelles (the ER, chloroplasts, and mitochondria, respectively) [24,25]. Humans contain one PDS comprising two subunits, PDSS1 and PDSS2 [21] similar to S. pombe PDS, which also consists of a heteromeric complex and served as the basis for analysis of the human enzyme [20,21]. A heteromeric form of PDS was probably evolved in S. pombe or earlier organisms and succeeded to humans. The result that artificial heteromeric PDSs between Coq1 and Dps1 or IspB and Dps1 is functional [19,26] supported an idea that heteromeric form was evolved from homomeric form.

Figure 4. Biosynthetic pathway of the isoprenoid tail of prenylquinones. Polyprenyl diphosphate synthase synthesizes trans-polyprenyl diphosphate of a certain length. S. cerevisiae Coq1 (hexaprenyl diphosphate synthase) forms products from six isoprene units, E. coli IspB (octaprenyl diphosphate synthase) synthesizes products with eight isoprene units, Arabidopsis SPS1, SPS2, and SPS3 (solanesyl diphosphate synthase) generate products with nine isoprene units, and human and Schizosaccharomyces pombe decaprenyl diphosphate synthase (DPS; a heteromer of PDSS1 and PDSS2 or Dps1 and Dlp1, respectively) catalyzes the formation of products with ten isoprene units. S. cerevisiae Coq2 (PHB-hexaprenyl diphosphate transferase), E. coli UbiA (PHB-octaprenyl diphosphate transferase), and human COQ2 (PHB-decaprenyl diphosphate transferase) or S. pombe Ppt1 (Coq2; PHB-decaprenyl diphosphate transferase) condense p-hydroxybenzoate (PHB) with trans-polyprenyl diphosphate to form UQ₆, UQ₈, and UQ₁₀, respectively. MenA prenylates DHNA, and homogentisate solanesyl transferase (HST) prenylates homogentisate (HGA). DPP, decaprenyl diphosphate; HexPP, hexaprenyl diphosphate; NPP, nonaprenyl diphosphate; OPP, octaprenyl diphosphate.

The three-dimensional structure of PDS has been solved, and octaprenyl diphosphate synthase (IspB) from E. coli consists of 14 α-helices [27]. Recent co-crystallization of IspB with its substrates (FPP and IPP) revealed aspartate-rich motifs surrounding the binding regions of substrates, and indicated a product pocket that determines the chain length. The three-dimensional structure of the heteromeric heptaprenyl diphosphate synthase from Staphylococcus aureus was solved [28], revealing a regulatory subunit that does not resemble the catalytic subunit.

E. coli ispB is essential for growth, while ispA encoding FPS is not [22], presumably because ispB replaces the function of ispA [29]. Since ispB is required for the synthesis of the side chain of both UQ and MK, E. coli cannot survive without both quinones. Coq1 in S. cerevisiae [19] and Dps1 (or Dlp1) in S. pombe are not essential for growth [20,30], while C. elegans coq1 and PDSS1 (or PDSS2) in mouse are essential for development [31,32].

Prenylquinones

Different groups of prenylquinones such as UQ, MK, PQ, and RQ (Figure 1) are present in different taxonomic groups, and prenylquinone profiling is a useful taxonomic tool [4]. Exactly why such a wide variety of quinones are found in nature is an interesting question. The isoprenoid side chain gives these molecule their lipid-soluble character, the quinone ring defines the redox mid-potential, and organisms have evolved the optimal quinone types for survival. The redox potential (E⁰) of MK is −74 mV, compared with −63 mV for RQ and + 100 mV for UQ. The lower redox potential of MK and RQ explains why they are used in electron transfer systems under anaerobic conditions, while UQ is employed in aerobic conditions. Differences in the natural environment of living organisms probably affect selection of the preferred quinone type. Evolutionally, MK probably arose in archaea, while PQ and UQ evolved later in bacteria and became distributed in eukaryotes, and RQ forms evolved most recently. However, some researchers believe that PQ may have evolved first in cyanobacteria, and was then distributed to other organisms, because the pathway for UQ biosynthesis is similar to that of PQ, and PQ has the simplest structure among prenylquniones [33].

Some organisms possess two types of quinones, such as E. coli that synthesizes UQ₈ and MK₈, and C. elegans that has UQ₉ and RQ₉. Possessing different types of quinones may be beneficial for adapting to changing environmental conditions. In E. coli, the level of UQ₈ is 4–5 times higher than that of MK₈ and demethyl menaquinone (DMK)₈ when growing under aerobic conditions, whereas UQ is three times less abundant than MKs under anaerobic conditions. In Euglena gracilis, RQ₉ is present at a similar concentration to UQ₉ under aerobic conditions, but is more abundant under anaerobic culture conditions. In C. elegans, UQ₉ is 3.56-fold more abundant than RQ₉, indicating a preference for aerobic growth, although anaerobic growth also occurs [34].

The significance of the length of the side chain of prenylquinones remains contentious, and only UQ has been thoroughly investigated. The side chain of UQ is determined by the supplied prenyl diphosphate synthesized by PDS [12]. Genetically engineered S. cerevisiae produce UQ₅ to UQ₁₀ and grow well, but the native form (UQ₆ in this case) is preferred for better growth [13]. E. coli producing UQ₆ to UQ₁₀ also grow well, but a longer side chain is preferable for better growth [22]. C. elegans clk-1 mutant, which lacks the penultimate enzyme (Coq7) in UQ synthesis, lives longer than wild type. When engineering E. coli producing UQ₆ to UQ₁₀ were used as diet, they reverse the longevity of this mutant, but the effect is different [35]. C. elegans prefers longer UQs such as UQ₈ to UQ₁₀, and the preference for a certain length may reflect the affinity for binding proteins or the membrane lipid composition. Why plants prefer the isoprene unit 9 form in PQs, while animals prefer the isoprene unit 4 form in MKs (or PQs) is also interesting.

A variety of scarcer prenylquinones other than those widely distributed in nature have been identified. Thermoplasmaquinone and methionaquinone are found in Thermoplasma spp. and Hydrogenobacter thermophilus, respectively [36,37]. Chlorobiumquinone, containing oxygenized isoprenoid in MK, is found in the photosynthetic bacterium Chlorobium limicola and in Leishmania parasitic protozoans [38]. Sulfolobusquinone, caldariellaquinone, and benzodithiophenoquinone, containing sulfur in an additional heterocyclic ring, are found in Sulfolobales, an order of thermophilic and aerobic archaebacteria [39]. Sulfomenaquinone, containing sulfur in the end of the side chain, is found in Mycobacterium tuberculosis [40], and a saturated isoprenoid in UQ is found in Fungi [41]. There are a few known organisms, such as obligatory fermentative bacteria, that lack prenylquinones [4].

Ubiquinone (coenzyme Q)

Ubiquinone (UQ; 2,3-dimethoxy-5-methyl-6-polyprenyl-1,4-benzoquinone) is an essential cofactor in oxidative phosphorylation, present in all eukaryotes and alpha-, beta-, and gamma-proteobacteria [4]. UQ was discovered by F. Crane in 1957, and the structure was determined by K. Folkers the following year [42]. UQ functions in many physiological processes including sulfide oxidation [43,44], first discovered in fission yeast and later in humans, as well as regulation of the mitochondrial permeability transition pore, and the translocation of protons and Ca²⁺ across biological membranes in eukaryotes [45]. UQ is the only lipid-soluble antioxidant produced in humans, and it is present in almost all membranes, ranging from mitochondrial membranes, Golgi, ER, and plasma membranes, to very low density lipoproteins. UQ₁₀ production decreases with aging in humans, as does the antioxidant capability of cells [46]. In humans, the heart, liver, and kidney have higher UQ₁₀ levels than other organs [47].

In model organisms such as E. coli, S. cerevisiae, and S. pombe, UQ deficiency is not lethal, but causes growth defects on minimum medium, and a heightened sensitivity to oxidative stress [17]. In C. elegans [31], UQ deficiency leads to gamma-aminobutyric acid (GABA) neuron degeneration, and in Drosophila melanogaster [48], it can cause mitochondrial stress and neuronal apoptosis. In Arabidopsis, UQ is necessary for seed development [49]. In humans, UQ₁₀ deficiency has been implicated in various diseases involving muscle and neural development, with the severity of the disease correlated with the acuteness of the UQ₁₀ shortfall [50].

The biosynthetic pathway of UQ has been reviewed previously [17,51–55], but important progress has been made in recent years. Biosynthesis of UQ has received greatest attention in E. coli and S. cerevisiae, serving as representative prokaryotes and eukaryotes, respectively (Figure 5). Some variation in UQ biosynthetic enzymes is observed in prokaryotes and eukaryotes; in particular, decarboxylation and C1 hydroxylation enzymes are not defined in eukaryotes, and likely to be different from prokaryotic enzymes [51].

Figure 5. Overview of the proposed UQ biosynthetic pathway. The UQ biosynthetic pathways of E. coli and S. cerevisiae are shown. In E. coli, PHB is first condensed with trans-polyprenyl diphosphate, and the ring structure is then modified. Decarboxylation catalyzed by UbiD (3-octaprenyl-4-hydroxybenzoate decarboxylase) and UbiX (flavin prenyl transferase) follows. The ring is further hydroxylated by UbiI (2-octaprenylphenol hydroxylase), O-methylated by UbiG (2-octaprenyl-6-hydroxy phenol methylase), hydroxylated by UbiH (2-octaprenyl-6-methoxyphenol hydroxylase), C-methylated by UbiE (2-octaprenyl-6-methoxy-1,4-benzoquinone methylase), hydroxylated by UbiF (2-octaprenyl-3-methyl-6-methoxy-1,4-benzoquinone oxygenase), and O-methylated by UbiG (3-demethylubiquinone 3-methyltransferase). In S. cerevisiae, para-amino benzoic acid (pABA) and PHB are used for UQ synthesis. The first ring is modified via hydroxylation by Coq6 (PHB-2-hexaprenyl hydroxylase), followed by O-methylation by Coq3 (2-hexaprenyl-6-hydroxy phenol methyltransferase). After decarboxylation and hydroxylation steps, the ring is further modified via C-methylation by Coq5 (2-hexaprenyl-6-methoxy-1,4-benzoquinone methyltransferase), a final hydroxylation by Coq7 (2-hexaprenyl-3-methyl-6-methoxy-1,4-benzoquinone oxygenase), and O-methylation by Coq3. H. sapiens contains similar enzymes with S. cerevisiae except PDSS1 and PDSS2 [70].

In E. coli, PHB is first condensed with trans-polyprenyl diphosphate by UbiA [56], and the ring structure is then modified. The decarboxylation step is catalyzed by UbiD with the assistance of UbiX, which generates the prenylated FMN cofactor for UbiD [57]. UbiX functions as a flavin prenyltransferase. The ring is further hydroxylated by UbiI [58], O-methylated by UbiG [59], hydroxylated by UbiH, C-methylated by UbiE [60], hydroxylated by UbiF [61], then O-methylated by UbiG. It is reported that ubiK and ubiJ are required for efficient biosynthesis of UQ in E. coli [62]. Hydroxylation and ring formation are reportedly catalyzed by enzymes encoded by ubiM and ubiL in Rhodospirillum [33]. The ubiZ gene product is predicted to be involved in UQ synthesis in Acinetobacter junii, based on genomic analysis of 254 human gut microbes [63]. However, verification of these genes in the biosynthesis of UQ awaits further evidence. An attempt to produce a higher amount of UQ by genetic engineering was first succeeded in E. coli by expressing ubiA, ubiB, ubiC, ubiG, ubiH and ispB [64].

In S. cerevisiae, PHB and para-amino benzoic acid (pABA) are used for UQ synthesis. PHB is synthesized from 4-hydroxybenzaldehyde by Hfd1 in S. cerevisiae [65]. A conserved homolog of Hfd1 is found in humans, but it is still not clear how many steps are required to form 4-hydroxybenzaldehyde from tyrosine [65]. pABA was originally identified as a precursor of ring formation in S. cerevisiae, and we observed that it is also used in S. pombe (unpublished). The first ring is prenylated by Coq2 [66] modified via hydroxylation by Coq6 [67] followed by O-methylation by Coq3 [59]. The enzymes responsible for decarboxylation and hydroxylation remain unclear. The ring is then modified further via C-methylation by Coq5 [68] a final hydroxylation by Coq7 [69] and O-methylation by Coq3. The genes involved in biosynthesis in eukaryotes are well conserved among yeasts, plants, and humans [70], although there is some variation among species. Even between the two model yeasts S. cerevisiae and S. pombe, components of PDS are different [19]. There are at least four genes (COQ4, COQ8, COQ9, and COQ11) responsible for the synthesis of UQ, but their functions are not known. The function of Coq4 is clearly conserved in humans and plants [70]. Conservation of Coq9 in higher eukaryotes is not so obvious, but interestingly, a homolog is also found in some prokaryotes [33]. Coq11 is associated with the UQ synthetic enzyme complex named CoQ synthome, and is required for UQ synthesis in S. cerevisiae [71]. Coq11 is also required for efficient UQ synthesis in S. pombe (unpublished). A deamination step is required for the synthesis of UQ from pABA, and the involvement of Coq9 or Coq6 has been proposed [72,73]. The UbiD and UbiX homologs Pad1 and Fdc1 found in yeasts are not involved in UQ synthesis, but are required for ferulic acid synthesis [74]. How decarboxylation takes place during ring formation in eukaryotes is a long-standing question in UQ synthesis.

In addition to the three-dimensional structure described previously [51], the structure of Coq3 was recently solved [75]. Coq3 forms a typical Class I S-adenosyl methionine methyltransferase (SAM-MTase) fold. Coq3 is a membrane-binding protein specifically binding to liposomes containing phosphatidylglycerol (PG), cardiolipin (CL), or diphosphatidylglycerol (DPPG). The three-dimensional structures of Coq7 and Coq11 are yet to be reported.

How UQ is transported has been a long-standing question. By searching for the binding protein using UQ, three UQ-binding proteins, Coq10, saposin, and voltage-dependent anion channel 1 (VDAC1), were identified [76,77]. Coq10 is localized to mitochondria in eukaryotes, and homologs are found in prokaryotes. Lack of Coq10 results in respiration deficiency in yeasts [76,78]. Coq10 itself is not required for the biosynthesis of UQ, but it is thought to be required for efficient operation of electron transfer systems. The binding site of Coq10 for UQ₁₀ was determined by affinity-purified Coq10 using a UQ analog in S. pombe [77]. Saposin is another protein that binds UQ, but it is only found in mammals. Among different types, saposin B was shown to bind UQ as well as tocopherol [79]. VDAC1, located in the mitochondrial outer membrane of S. cerevisiae, is another UQ-binding protein [80]. The role of VDAC1 in Ca²⁺-induced mitochondrial permeability is affected by binding to UQ, but whether this function is conserved in other organisms is not known.

The unique fission yeast Schizosaccharomyces japonicus produces 200-fold less UQ₁₀ than S. pombe [81]. This fission yeast acquires energy through fermentation and has abandoned respiration. How this yeast survives in its natural environment with such a small amount of UQ₁₀ is interesting and worthy of study.

Menaquinones

Menaquinone (MK; 2-methyl-3-polyprenyl-1,4-naphthoquinone) is found in bacteria, and is the sole quinone in anaerobically growing bacteria [82,83]. MK was discovered in 1939 by E. A. Doisy [84]. MKs are found in archaea and bacteria such as γ-, δ-, and ε-proteobacteria, Gram-positive bacteria, green sulfur bacteria, green filamentous bacteria, and flavobacteria. As MKs have a low midpoint redox potential, they are believed to have appeared early in evolution before UQ, since they function in a reducing atmosphere as was present before the increase in oxygen concentration following the arrival of photosynthetic organisms.

MKs occur in different forms, with the number of isoprene units varying between 4 and 13. Some bacteria such as E. coli possess both UQ₈ and MK₈, and the relative amounts of each depend on oxygen levels; while UQ levels are higher under aerobic conditions, MK₈ is more abundant under anaerobic conditions. Neither MK nor UQ is essential for survival in E. coli, but at least one of these prenylquinones is needed [22]. The side chain of MK is usually fully unsaturated, but it can be also be partially or fully saturated in some organisms [85].

Two pathways are known for the synthesis of MK. The classical pathway involves nine steps catalyzed by MenF, MenD, MenH, MenC, MenE, MenB, MenI, MenA, and MenG (Figure 6) [83,86]. The biosynthesis of MK in E. coli starts from the conversion of chorismate to isochorismate by MenF, and succination is then catalyzed by MenD. In the third step, 2-succinyl-6-hydroxy-2,4-cyclohexadiene-1-carboxylate (SHCHC) is synthesized by MenH, and O-succinylbenzoate (OSB) is then synthesized from SHCHC by MenC. In the fifth step, CoA is adducted to OSB by MenE, and MenB then cyclizes OSB-CoA to form DHNA-CoA. In the seventh step, MenI synthesizes naphthoate [87], which is prenylated by MenA, and the product is methylated by MenG (UbiE) in the final step. Prenylation takes place during the later stages via MenA, in contrast with the synthesis of UQ in which it occurs earlier via UbiA (Figure 5). The methylation enzyme (MenG) for MK is homologous to UbiE functioning in UQ synthesis in E. coli. MenJ works as a reductase of the side chain in Mycobacterium tuberculosis [88]. Further methylation of MK is observed in some bacterium such as Shewanella oneidensis [89].

Figure 6. Menaquinone biosynthesis in E. coli. The biosynthesis of menaquinone in E. coli starts from the conversion of chorismate to isochorismate by MenF. Subsequently, MK is synthesized by MenD (succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexadiene-1-carboxylate synthase), MenH (2-succinyl-6-hydroxy-2,4-cyclohexadiene-1-carboxylate synthase), MenC (O-succinylbenzoate synthase), MenE (O-succinylbenzoic acid-CoA ligase), MenB (naphthoate synthase), MenI (DHNA-CoA thioesterase), MenA (1,4-dihydroxy-2-naphthoate octaprenyltransferase), and MenG (UbiE). *MenJ functions as a reductase of the side chain in Mycobacterium tuberculosis.

The novel pathway for MK synthesis was first discovered in Streptomyces coelicolor and subsequently in Helicobacter pylori and Thermus thermophilus [90,91]. Six enzymes are engaged in 1,4-dihydroxy-6-naphthoate biosynthesis. MqnA converts chorismate to 3-[(1-carboxyvinyl)oxy]benzoic acid, which is condensed with SAM by MqnE, leading to aminofutalosine (AFL). MqnE is a radical SAM enzyme that catalyzes the addition of the adenosyl radical to the double bond of 3-[(1-carboxyvinyl)oxy]benzoic acid. Deamination of AFL is catalyzed by a specific deaminase for which no common gene name has been assigned [92]. MqnB (futalosine hydrolase) then removes hypoxanthine, forming dehypoxanthinylfutalosine (DHFL) [93]. MqnC cyclizes DHFL, and MqnD cleaves the cyclic 1,4-dihydroxy-6-naphthoate to release 1,4-dihydroxy-6-naphthoate (Figure 7). In H. pylori, MqnB directly converts aminodeoxyfutalosine into DHFL [94], indicating an alternative way in the futalosine pathway.

Figure 7. Novel pathway of menaquinone synthesis via futalosine. A novel menaquinone biosynthesis pathway was originally discovered in Streptomyces coelicolor, in which 1,4-dihydroxy-6-naphthoate is synthesized by MqnA (chorismate dehydratase), MqnD (1,4-dihydroxy-6-naphthoate synthase), MqnE (aminofutalosine synthase), AFL deaminase, MqnB (futalosine hydrolase), and MqnC (dehypoxanthine futalosine cyclase). MK synthesis from 1,4-dihydroxy-6-naphthoate is still not clearly understood, but a prenylation step catalyzed by MqnP has been proposed.

The novel genes mqnP, mqnL, and mqnN were predicted to encode enzymes involved in MK synthesis in Helicobacter cinaedi following genomic analysis of 254 human gut microbes. [63]. MqnP is predicted to be involved in prenylation of 1,4-dihydroxy-6-naphthoate, and MqnL and MqnN are likely involved in decarboxylation. Comparing gene clusters can be useful for predicting biosynthetic genes such as those orchestrating the synthesis of MK. However, biochemical analysis is essential for confirming any predictions, and analysis of this futalosine pathway leading to MK is still under investigation.

The three-dimensional structures of MK biosynthetic enzymes MenC from E. coli [95] and Thermosynechococcus elongatus [96], and MenB [97], MenD [98], MenE [99], MenF [100], MenH [101], and MenI [102] from E. coli, have been determined (Figure 8). The structure of MenC is similar to that of other members of the enolase superfamily [95]. The structure of MenB, a crotonase superfamily member, was solved in complex with a substrate analog, revealing an intramolecular Claisen condensation reaction mechanism [97]. MenD is highly dependent on thiamine diphosphate for its structural stability [98]. MenE requires a conserved arginine for binding the OSB carboxylate, and catalyzes CoA ligation via an acyl-adenylate intermediate [99]. Structural and biochemical analyses of MenF revealed Lys190 as the base that activates a water molecule for nucleophilic attack at the chorismate C2 carbon [100]. MenH has an α/β-hydrolase fold with a catalytic triad comprising Ser86, His232, and Asp210 [101]. MenI (YdiI) belongs to the hotdog fold enzyme superfamily [102]. The three-dimensional structures of MqnA [103] and MqnD, and AFL deaminase (Nis0429) functioning in the futalosine-mediated MK pathway, have been solved [92]. The structure of MqnA (DUF178) from Deinococcus radiodurans was originally solved as a domain of unknown function before being identified as MqnA [103]. The structure of MqnD from Thermus thermophilus HB8 comprises two alpha/beta domains, a large domain, and a small domain [104]. The three-dimensional structure of MqnB from H. pylori has a Rossmann fold [93]. The structures of AFL deaminase (Nis0429) from Nitratiruptor sp. and Dr0824 from D. radiodurans reveal that Ser145 interacts with the carboxylate moiety of the substrate [92].

Figure 8. Crystal structures of menaquinone biosynthetic enzymes. (A) MenF from E. coli (PDB ID: 2EUA). (B) MenC from E. coli (PDB ID: 1FHU). (C) MenB from E. coli (PDB ID: 3T89). (D) MenD from E. coli (PDB ID: 3HWX). (F) MenE from E. coli (PDB ID: 5C5H). (G) MenH from E. coli (PDB ID: 4GDM). (H) MenI from E. coli (PDB ID: 4K4B). (I) MqnA from Deinococcus radiodurans (PDB ID: 216E). (J) MqnD from Thermus thermophilus HBB (PDB ID: 3A3U). (K) MqnB from Helicobacter pylori (PDB ID: 4BMX).

Phylloquinones (PhQs, vitamin K₁)

Phylloquinone (VK₁; 2-methyl-3-phytyl-1,4-naphthoquinone) functions as an essential photosynthetic electron transporter in photosystem I, and was discovered by H. C. P. Dam in 1934 as a vitamin [105]. Higher amounts are found in green leafy vegetables because it is directly involved in photosynthesis. Humans rely on PhQ uptake from vegetables as a precursor for the synthesis of MK₄. PhQ is thought to be converted to MK₄ by UBIAD in humans. Experiments performed on rodents showed that at least some of their tissues are able to convert PhQ to MK₄. UBIAD mediates the conversion of PhQ into MK₄, probably by cleaving the side chain of PhQ to generate 2-methyl-1,4-naphthoquinone (menadione; VK₃), then prenylating it with GGPP to form MK₄ [106].

The biosynthetic pathway of PhQ in cyanobacteria and plants is thought to resemble the MK pathway [2,107]. Four genes, menF, menD, menC, and menH, involved in PhQ biosynthesis in Arabidopsis, are fused at a single locus named PHYLLO [108]. The structure of the MenI ortholog AtDHNAT1 (DHNA-CoA thioesterase) has been solved [109]. O-succinylbenzoyl-coenzyme A (OSB-CoA) ligase (a MenE ortholog) encoded by aae14 is essential for PhQ synthesis [110]. A MenG ortholog was identified as the methyltransferase catalyzing the last step of PhQ synthesis in Arabidopsis [111], and a MenB homolog has been identified in the Arabidopsis genome sequence. MenB, MenI, and MenG orthologs localize to the peroxisome [112], while PHYLLO, comprising MenF, MenD, MenC, and MenH orthologs, is localized to chloroplasts [108]. Carboxy-1,4-naphthoquinone phytyltransferase (a MenA ortholog) is involved in PhQ synthesis chloroplasts [113]. The phytyl moiety of PhQ is synthesized either by reduction of GGPP in de novo synthesis, or via the salvage pathway. Recent analysis of Arabidopsis vte6 encoding phytyl phosphate kinase revealed that it performs an essential role in PhQ synthesis [114]. The entire biosynthetic pathway of PhQ in plants is still under investigation.

Plastoquinones

Plastoquinone (PQ; 2,3-dimethyl-1,4-benzoquinone), discovered in 1946 [115] functions in the electron transport chain of oxygenic photosynthesis, and plays an indispensable role in plant growth and development. PQ is found in cyanobacteria and plants. PQ₉ is distributed widely among organisms, while PQ₈ is found in maize. In the biosynthesis of PQ in plants, tyrosine is converted to p-hydroxyphenylpyruvate (PHPP) by tyrosine aminotransferase (TAT) [2]. Homogentisate (HGA) is then synthesized from HPP by p-hydroxyphenylpyruvate dioxygenase (Figure 9). The prenyl tail is synthesized independently from the head group by SPS1, and IPP for prenyl tail synthesis is supplied by the MEP pathway in chloroplasts. Condensation of HGA with the prenyl tail is catalyzed by homogentisate solanesyl transferase (HST) [116]. Finally, a methylation reaction is catalyzed by methyl transferase (Vte3) [117]. The vte genes are required for vitamin E synthesis and perform some functions in PQ synthesis. The lipid-associated protein Fibrillin5 (FBN5), required for PQ synthesis, interacts with SPS1 and SPS2 [118]. Overexpression of SPS1 in Arabidopsis resulted in enhanced phototolerance [119].

Figure 9. The biosynthesis of plastoquinone. In plants, tyrosine is converted to p-hydroxyphenylpyruvate (PHPP) by TAT (tyrosine amino transferase), and the product is oxidized to HGA by HPPD (p-hydroxyphenylpyruvate dioxygenase). HGA is prenylated by HST (homogentisate prenyltransferase), and methylated by VTE3 (MSBQ methyltransferase) to yield PQ. In Synechocystis, PHB is used as a quinone backbone, and after prenylation of PHB by Slr0926, decarboxylation, oxidation, and methylation take place.

Similar to UQ, the biosynthesis of PQ differs in eukaryotes and prokaryotes. Chorismate lyase generates PHB in the synthesis of PQ in the cyanobacterium Synechocystis [120]. PHB is prenylated by Slr0926, and decarboxylated by Slr1099 and Sll0936, then oxygenized and methylated to make PQ (Figure 9).

There are other forms of PQ with shorter side chains such as PQ₃ and PQ₄, as well as analogs such as PQ-B, and PQ-C, which differ in the modification pattern of their side chains [115]. PQ-C contains hydroxyl group in the prenyl chain and PQ-B is a fatty acid ester form of PQ-C [115].

Rhodoquinones

Rhodoquinone (RQ; 2‐methoxy‐3‐amino-5‐methyl-6-polyprenyl-1,4-benzoquinone) was discovered in the bacterium Rhodospirillum rubrum in 1965, and subsequently in other organisms such as Rhodoferax fermentansi [121], E. gracilis [122], C. elegans [34], planaria, parasitic helminths, snails, mussels, lungworms, and oysters. Anaerobically and aerobically grown E. gracilis cells contain similar total amounts of RQ and UQ, but RQ constitutes 43 and 28% of the pool under anaerobic and aerobic conditions, respectively [122]. Helminth parasites can use fumarate as a terminal electron acceptor in the respiratory chain since they possess RQ-fumarate oxidoreductase. C. elegans produce both UQ₉ and RQ₉, and the relative amounts are thought to be of relevance to lifespan [123]. Rhodoplanes serenus produces UQ₁₀ and RQ₁₀ [124]. As these examples demonstrate, organisms possessing RQ also have UQ, and UQ was shown to be required for the biosynthesis of RQ in R. rubrum. A novel gene named rquA was found in R. rubrum that is required only for RQ synthesis but not UQ synthesis [125]. The biosynthesis of RQ is still not fully understood.

Applications of prenylquinones

UQ₁₀ (coenzyme Q₁₀) is popular as a food supplement and sold worldwide in both reduced and oxidized forms. The demand for skin care cosmetics and public awareness of the importance of antioxidants such as UQ₁₀ has increased, and UQ₁₀ is also used therapeutically in Alzheimer’s, Huntington’s, Parkinson’s, and cardiovascular diseases [47]. As UQ₁₀ is naturally produced in humans, and available from foods such as meat and fish, side effects are very rare. Taking statins to reduce the amount of cholesterol also lowers UQ₁₀ levels, and so taking both simultaneously is recommended [47]. The UQ₁₀ commercial market is large, and UQ₁₀ is purified from yeast or photosynthetic bacteria. Several native producers of UQ₁₀ have been investigated to optimize UQ₁₀ production. S. pombe, Sporidiobolus johnsonii, Rhodobacter sphaeroides, and Agrobacterium tumefaciens reportedly produce 1.0, 10.5, 8.7, and 4.5 mg/g dry cell weight (DCW), respectively [126]. However, because these amounts were measured by different groups using different methods, direct comparison is necessary to assess the efficiency of UQ₁₀ production by these microorganisms. Attempts to produce UQ₁₀ in rice and tobacco have proven successful [127,128], and regulation of genetically modified organisms (GMOs) hampers the commercial production of UQ₁₀ in rice.

MK is also sold as a food supplement in the form of MK₄ or MK₇. MK₇ is from B. subtilis, and MK₄ is produced in animals. The Japanese food Natto, fermented by B. subtilis, contains MK₇, and increases bone mineral density and reduces bone fractures. Humans do not biosynthesize MK, but it is utilized for blood coagulation, bone metabolism, and cell-cycle regulation. The major sources of MK in humans are the diet and gut flora. As MK is only synthesized in bacteria, inhibitors of MK synthesis are useful for inhibiting the growth of harmful bacteria such as methicillin-resistant Staphylococcus aureus (MRSA). Using this concept, analogs of each reaction step were synthesized and shown to be efficient inhibitors such as 7-methoxy naphthalene derivatives, methyl 4-oxo-4-phenylbut-2-enoate, and lysocin E of MenA [129], MenB [130], and MenE, respectively [131].

PhQ (VK₁) is used as a vitamin supplement since mammals are not able to synthesize it, and must obtain it from their diet. PhQ is as a cofactor for coagulation factors II, VII, IX, and X, required for the formation of anticoagulant factors protein C and S, and for bone protein formation. PhQ is commonly used to treat warfarin toxicity.

PQ itself is not a commercially useful product, but derivatives such as plastoquinonyl-decyl-triphenylphophonium (SkQ1) and its methylated derivative SkQ3 are under consideration for usage as antioxidants. Mitochondrial-targeted SkQ1 is currently under clinical trial for glaucoma treatment and prevention of dry eye [132].

RQ is only found in a limited number of organisms that are not used for food, and applications for RQ have not been reported.

Concluding remarks

There are a wide variety of prenylquinones in nature, but their synthesis is not fully understood. In this review, the biosynthesis of prenylquinones was summarized, with emphasis on UQ, MK, PhQ, PQ, and RQ. Extensive studies have uncovered in detail MEP and MVA biosynthetic pathways that lead to the synthesis of IPP and DMAPP. Enzymes condensing IPP with DMAPP to produce polyprenyl diphosphate have been particularly well studied. While the synthesis of the isoprenoid side chain is relatively well characterized, many aspects of the modification of prenylquinones remain obscure. UQ synthesis in bacteria is quite well defined, but our knowledge of UQ synthesis in eukaryotes is incomplete. Classical MK synthesis is mostly understood, but the novel MK pathway mediated by futalosine is vague in comparison. Many three-dimensional structures related to UQ and MK synthesis have been solved, but some reactions in PhQ synthesis remain undefined. Furthermore, exactly how RQ is synthesized after UQ remains to be solved. Bioinformatics and genomics approaches are proving useful for predicting the biosynthetic pathways of these prenylquinones, but details of such work fall beyond the scope of this review. Nevertheless, this summary of recent progress on the biosynthesis of prenylquinones should prove useful, and will likely accelerate the characterization of unknown reactions and enzymes.

Disclosure statement

No potential conflict of interest was reported by the author.

Funding

This work was supported by the Ministry of Education, Culture, Sports, Science, and Technology of Japan [grant number 17H03806].

Acknowledgments

I thank Dr. T. Kaino for checking this manuscript.