Biosynthesis of coenzyme Q in eukaryotes

Abstract

Coenzyme Q (CoQ) is a component of the electron transport chain that participates in aerobic cellular respiration to produce ATP. In addition, CoQ acts as an electron acceptor in several enzymatic reactions involving oxidation–reduction. Biosynthesis of CoQ has been investigated mainly in Escherichia coli and Saccharomyces cerevisiae, and the findings have been extended to various higher organisms, including plants and humans. Analyses in yeast have contributed greatly to current understanding of human diseases related to CoQ biosynthesis. To date, human genetic disorders related to mutations in eight COQ biosynthetic genes have been reported. In addition, the crystal structures of a number of proteins involved in CoQ synthesis have been solved, including those of IspB, UbiA, UbiD, UbiX, UbiI, Alr8543 (Coq4 homolog), Coq5, ADCK3, and COQ9. Over the last decade, knowledge of CoQ biosynthesis has accumulated, and striking advances in related human genetic disorders and the crystal structure of proteins required for CoQ synthesis have been made. This review focuses on the biosynthesis of CoQ in eukaryotes, with some comparisons to the process in prokaryotes.

Graphical abstract

The enzymes involved in CoQ synthesis form a complex in S. cerevisiae. At least nine peptides are required for the synthesis of CoQ.

Key words:

• coenzyme Q
• ubiquinone
• yeast
• human
• Arabidopsis

Introduction

Coenzyme Q (CoQ), also known as ubiquinone or 2,3-dimethoxy-5-methyl-6-polyprenyl-1, 4-benzoquinone, is a well-known component of the electron transport chain that participates in aerobic cellular respiration within the mitochondria of eukaryotes and the plasma membrane of prokaryotes. CoQ exists in both reduced and oxidized forms;¹⁾ conversion between these states allows it to transfer electrons to substrates and act as a cofactor of enzymatic reactions (Fig. 1). CoQ plays the role of the electron donor during disulfide bond formation in Escherichia coli,²⁾ and its reduction is coupled to sulfide oxidation by sulfide–quinone oxidoreductase in Schizosaccharomyces pombe and mammals.³⁾ CoQ also couples with electron transfer by glycerol-3-phosphate dehydrogenase and electron-transferring flavoprotein dehydrogenase, the latter of which is involved in the beta-oxidation of fatty acids in mammals.⁴⁾ Moreover, CoQ is required for the de novo synthesis of UMP as a cofactor of dihydroorotate dehydrogenase in many eukaryotes (Fig. 1).^5,6)

Fig. 1. The electron transfer system and enzymes involved in CoQ biosynthesis. Notes: The position of CoQ in the electron transfer system is shown. Complexes I and II transfer electrons to CoQ from NADH and FADH₂, respectively. In yeast, NADH dehydrogenase replaces Complex I in the first reaction. Electrons are transferred to Complex III from CoQH₂, a reduced form of CoQ, and then further transferred to Complex IV through cytochrome c (Cytc). Protons are transferred to the intermembrane space, and this proton gradient drives ATP production through Complex V. A number of different enzymes are coupled with CoQ in oxidation–reduction reactions: DHODH, dihydroorotate dehydrogenase; SQR, sulfide quinone reductase; ETFDH, electron transfer flavoprotein dehydrogenase; and GPDH, glycerol-3-phosphate dehydrogenase.

CoQ synthesis is divided into two parts: the synthesis of isoprenoid (Fig. 2) and the modification of quinone (Fig. 3). The synthesis of isoprenoids has been studied extensively; however, the mechanism by which quinone is synthesized in eukaryotes requires clarification. Living organisms possess a number of species of CoQ with differing isoprenoid side chain lengths. For example, the human and S. pombe CoQ contains ten isoprene units, whereas CoQ from Arabidopsis thaliana, E. coli, and Saccharomyces cerevisiae contains nine, eight, and six units, respectively. The length of the CoQ side chain is defined by the product generated by polyprenyl diphosphate synthases.^7,8)

Fig. 2. Biosynthetic pathway of the isoprenoid tail of CoQ. Notes: Isopentenyl diphosphate (IPP) is synthesized via the mevalonate (MVA) pathway in eukaryotes and the 2-C-methyl-D-erythritol 4-phosphate (MEP) pathway in prokaryotes and plants. In each organism, trans-polyprenyl diphosphate of a certain length is synthesized by polyprenyl diphosphate synthase. S. cerevisiae Coq1 synthesizes six isoprene units, E. coli IspB synthesizes eight isoprene units, and H. sapiens or S. pombe decaprenyl diphosphate (DPP; a heteromer of PDSS1 and PDSS2 or Dps1 and Dlp1, respectively) synthesize ten isoprene units. S. cerevisiae Coq2, E. coli UbiA, and H. sapiens COQ2 or S. pombe Ppt1(Coq2) condense PHB with trans-polyprenyl diphosphate to form CoQ6, CoQ8, and CoQ10, respectively. DMAPP, dimethylallyl diphosphate; GPP, geranly diphosphate; FPP, farnesyl diphosphate; IPP, isopentenyl diphosphate; HexPP, Hexaprenyl diphosphate; OPP, Octaprenyl diphosphate.

Fig. 3. Overview of the proposed CoQ biosynthetic pathway. Notes: The CoQ biosynthetic pathways of prokaryotes (represented by E. coli) and eukaryotes (represented by S. cerevisiae). In general, the gene names ubi* and COQ(coq)* are used in prokaryotes and eukaryotes, respectively; however, the nomenclature of genes can differ among species. After condensation of PHB with trans-polyprenyl diphosphate, the ring structure is modifed. In E. coli, decarboxylation by UbiD or UbiX is followed by hydroxyation by UbiI, O-methylation by UbiG, hydroxyation by UbiH, C-methylation by UbiE, a final hydroxylation by UbiF, and then O-methylation by UbiG. At least eight genes are responsible for CoQ biosynthesis in E. coli. In S. cerevisiae, pABA is also used as a substrate in addition to PHB. In this species, the first ring is modifed by hydroxyation by Coq6, followed by O-methylation by Coq3; however, the subsequent decarboxylation and hydroxylation steps are unclear. The ring is then modified further by C-methylation by Coq5, a final hydroxylation by Coq7, and O-methylation by Coq3.

In prokaryotes, the biosynthetic pathway that converts para-hydroxybenzoate (PHB) into CoQ consists of at least eight steps, and the order of reactions is believed to be different in eukaryotes (Fig. 3). These steps include the condensation and transfer of the isoprenoid side chain to PHB, followed by hydroxylation, methylation, and decarboxylation. To date, this pathway has been studied most extensively in E. coli and S. cerevisiae using genetic and biochemical analyses. At least 11 genes (ubiA, B, C, D, E, F, G, H, I, J, and X) are involved in CoQ biosynthesis in E. coli, and 11 genes (YAH1, ARH1, and COQ1-COQ9) are required for this process in S. cerevisiae. Involvement of para-amino benzoic acid (pABA) in CoQ biosynthesis is only shown in S. cerevisiae.

Following on from my previous reviews published in 2002⁹⁾ and 2009,¹⁾ the current review provides an overall picture of the biosynthetic pathway of CoQ, with a special focus on recent advances in structurally based insights and the relationship between CoQ and human genetic disorders. Because bacterial biosynthesis of CoQ has been comprehensively reviewed by Aussell et al.¹⁰⁾ this review focuses on the process in eukaryotes.

Genes responsible for CoQ biosynthesis

Genes involved in the CoQ biosynthetic pathway have been studied most extensively in two model organisms, namely E. coli and S. cerevisiae, which are representatives of prokaryotes and eukaryotes, respectively.^1,9) E. coli ubi (ubiA–J and ubiX) mutants and S. cerevisiae coq (coq1–coq9) mutants, which are unable to synthesize CoQ, were used to define the biosynthetic pathway of CoQ.^11–14) Until recently, all corresponding genes for E. coli ubi mutants and S. cerevisiae coq mutants have been identified. While the functions of all E. coli ubi genes except ubiB have been completely assigned, the functions of three S. cerevisiae COQ genes (COQ4, COQ8, and COQ9) remain elusive.^1,12) Table 1 shows a comparison of the CoQ biosynthetic genes from seven species, namely Homo sapiens, Mus musculus, Caenorhabditis elegans, A. thaliana, S. cerevisiae, S. pombe, and E. coli. The order of the first three steps of the CoQ biosynthetic pathway differs in E. coli and S. cerevisiae (Fig. 3). The decarboxylation stage of the ring modification procedure precedes the hydroxylation and methylation stages in E. coli, but the order of these reactions is thought to be different in S. cerevisiae. The species-specific phenotypes of all coq gene-knockout S. pombe^3,15) and C. elegans¹⁶⁾ strains have been reported. It is generally considered that the CoQ biosynthesis process is common in eukaryotes (Table 1); however, some differences have been reported: (i) polyprenyl diphosphate synthase forms a homodimeric or homotetrameric structure in S. cerevisiae,¹⁷⁾ but the homologous enzyme forms a heterotetramer structure composed of two subunits in S. pombe, mouse, and human^18,19); (ii) orthologs of Coq7 are missing in plant genomic databases; and (iii) Coq9 is missing in C. elegans.

Table 1. CoQ biosynthetic genes from various species.

Function	H. sapiens	M. musculus	C. elegans	A. thaliana	S. cerevisiae	S. pombe	E. coli
Polyprenyl diphosphate synthase	PDSS1 + PDSS2	Pdss1 + Pdss2	COQ1	SPS3	COQ1	dps1 + dlp1	ispB
PHB-polyprenyl diphosphate transferase	COQ2	Coq2	COQ2	COQ2 (PPT1)	COQ2	coq2 (ppt1)	ubiA
O-methyltransferase	COQ3	Coq3	COQ3	COQ3	COQ3	coq3	ubiG
Unknown	COQ4	Coq4	COQ4	COQ4	COQ4	coq4	-
C-methyltransferase	COQ5	Coq5	COQ5	COQ5	COQ5	coq5	ubiE
Monooxygenase	COQ6	Coq6	COQ6	COQ6	COQ6	coq6	ubiI
Monooxygenase	COQ7	Coq7 (Clk1)	COQ7	-	COQ7	coq7	ubiF
Protein kinase	ADCK3, ADCK4	Adck3, Adck4	COQ8	COQ8 (ABC1)	COQ8 (ABC1)	coq8	(ubiB)
Unknown	COQ9	Coq9	–	COQ9	COQ9	coq9	–

Recent studies of human inherited diseases led to the identification of genetic disorders related to CoQ biosynthesis (Table 2). The following sections describe the genes involved in CoQ biosynthesis in S. cerevisiae in detail.

Table 2. Genotype–phenotype correlations in inherited deficiencies of CoQ10 biosynthesis in humans.

Gene	Amino acid change caused by mutation	CoQ10 level	Clinical features	References
PDSS1	D308E	3%**	Deafness, bulimia, obesity, optic atrophy, valvulopathy, macrocephathy, peripheral neuropathy	^31^)
PDSS2	Q322X & S382L	2%*, 13%**	Steroid-resistant nephrotic syndrome, leigh syndrome	^30^)
COQ2	Y297C	37%*, 18%**	Encephalomyopathy, nephropathy	^39^)
COQ2	R197H & N228S	36%*	Steroid-resistant nephrotic syndrome	^40^)
	S146N	2%*	Acute renal failure, epileptic encephalopathy
COQ2	N401fsX415	24%**	Neurological distress, liver failure, nephrotic syndrome, anemia, pancytopenia, diabetes, cytolysis, seizures	^31^)
COQ2	R387Q & V343A	63%***	Multiple-system atrophy	^94^)
COQ4	Del &WT	44%**	Mental retardation, encephalomyopathy, dysmorphic features	^47^)
COQ6	G255R	–	Nephrotic syndrome	^52^)
	A353D	–
CABC1/ADCK3	E551K	8%*	Cerebellar ataxia, strabismus, seizures	^95^)
	R213W & G272V	29%*	Seizures, cerebellar ataxia, ptosis,
	G272D & fs (1812insG)	<5%*	Cerebellar atrophy
CABC1/ADCK3	Q167LfsX36	63%**	Cerebellar ataxia, cerebellar atrophy	^62^)
	Y514C & T584X	51%**	Cerebellar ataxia, cerebellar atrophy
COQ8/CABC1/ADCK3	R348X, L379X	–	Cerebellar ataxia, atrophy	^96^)
ADCK4	R178W	9.7%***	Steroid-resistant nephrotic syndrome	^61^)
COQ9	R244X	15%*	Neonatal lactic acidosis, renal tubular dysfunction, seizures, hypertrophic cardiomyopathy, global development delay	^67^)

* muscle.

** fibroblast.

*** lymphoblast; fs, frame shift; del, deletion; ins, insertion.

Coq1/IspB/PDSS1 + PDSS2/polyprenyl diphosphate synthase

Polyprenyl diphosphate synthase (Coq1 in S. cerevisiae, IspB in E. coli, and a heteromer of PDSS1 and PDSS2 in H. sapiens), the enzyme that synthesizes the isoprenoid chain of CoQ, is classified into two types, namely the homomer and heteromer types. S. cerevisiae Coq1 (hexaprenyl diphosphate synthase) catalyzes the condensation of isopentenyl diphosphate and farnesyl diphosphate or geranylgeranyl diphosphate to produce six units of prenyl diphosphate, which are used to produce CoQ6 (Fig. 2).¹⁷⁾ Long-chain trans-polyprenyl diphosphate synthases, such as Coq1, have been studied in many species ranging from bacteria to humans. These enzymes possess seven conserved regions, including two DDXXD motifs that involve binding of substrates in association with Mg²⁺.²⁰⁾ Bacterial polyprenyl diphosphate synthases, S. cerevisiae Coq1, and A. thaliana solanesyl diphosphate synthase are homomeric enzymes.^21–24) Meanwhile, decaprenyl diphosphate synthases from S. pombe and human are heterotetramers consisting of dimer of two proteins, namely Dps1/Dlp1 and PDSS1/PDSS2, respectively.^18,19,25) PDSS2/Dlp1 lacks the aspartate-rich motifs and is bound to PDSS1/Dps1 to form a heteromeric enzyme.^18,19) Recent studies generated functional artificial heteromeric polyprenyl diphosphate synthases comprising S. cerevisiae Coq1 and S. pombe Dps1, or E. coli IspB and S. pombe Dps1, the latter of which (Dps1) is non-functional by itself, suggesting that it may support the activities of polyprenyl diphosphate synthases by partnering with a related protein.^17,26) The crystal structures of octaprenyl diphosphate synthase (IspB) from Thermotoga maritime and E. coli have been solved,^27,28) and analyses of these enzymes in complex with farnesyl diphosphate and isopentenyl diphosphate clearly defined the substrate-binding sites (Fig. 4). Mice carrying mutations in the Pdss2 gene display focal segmental glomerulopathy-like kidney disease that is rescued by CoQ10 supplementation.²⁹⁾ In addition, mutations in the human PDSS1 and PDSS2 genes, such as those associated with Leigh disease and nephropathy, result in poor mitochondrial functionality (Table 2).^30,31)

Fig. 4. Crystal structures of the proteins involved in CoQ synthesis.

Coq2/UbiA/COQ2/PHB-polyprenyl diphosphate transferase

PHB-polyprenyl diphosphate transferase (Coq2 in S. cerevisiae, UbiA in E. coli, and COQ2 in H. sapiens) catalyzes the condensation of PHB with the isoprenoid chain,^32,33) and genetic and biochemical analyses have revealed that this enzyme has broad substrate specificity.³⁴⁾ The yeast coq2 mutant and the E. coli ubiA mutant, which display a complete loss of PHB-polyprenyl diphosphate transferase activity, are complemented by the human and S. cerevisiae COQ2 genes, respectively.³²⁾ The analysis of a T-DNA insertion mutant revealed that the A. thaliana PPT1 gene (a homolog of S. cerevisiae COQ2) is essential for embryo development.³⁵⁾ Alongside hydroxy benzoic acid, the novel precursor pABA can also be used as a substrate for Coq2 in S. cerevisiae,^36,37) but presumably not in humans. Analysis of the crystal structure of UbiA from Aeropyrum pernix (Fig. 4) revealed that it comprises nine transmembrane domains and enabled visualization of the potential entry site of the substrate (PHB).³⁸⁾ Mutations in the human COQ2 gene cause encephalomyopathy, cerebellar ataxia, neurological distress, and other disorders (Table 2).^31,39,40)

Coq3/UbiG/COQ3/O-methyltransferase

Dihydroxypolyprenylbenzoate methyltransferase (O-methyltransferase) is encoded by the ubiG gene in E. coli and the COQ3 gene in S. cerevisiae and H. sapiens.^41,42) During CoQ biosynthesis, this enzyme catalyzes two O-methylation steps at positions 5 and 6 of the ring structure after hydroxylation by Coq6 and Coq7. The amino acid sequences of the proteins encoded by COQ3 homologs contain four regions that are conserved in a large family of methyl transferase enzymes that utilize S-adenosyl methionine (SAM) as a methyl donor. E. coli UbiG complements the function of S. cerevisiae Coq3, indicating the functional similarity of these proteins.⁴³⁾ S. pombe mutants lacking coq3 display the common phenotypes found in other coq mutants.⁴⁴⁾ Homozygous C. elegans coq3 mutants that lack methyltransferase activity display delayed development and a sterile phenotype, and these mutants are lethal at the embryonic stage in the next generation.⁴⁵⁾

Coq4/COQ4

Coq4 is absolutely required for the biosynthesis of CoQ in S. cerevisiae, S. pombe, and C. elegans. Homologs of Coq4 are also found in cyanobacteria; however, the molecular function of the Coq4 protein has not been elucidated in any organism.^15,46) The amino acid sequence of the Coq4 protein does not share any significant homology with protein domains or motifs with known enzymatic activity. The Coq4 protein is peripherally associated with the matrix side of the inner mitochondrial membrane and may play a structural role in the putative polypeptide CoQ biosynthetic complex.⁴⁶⁾ Lack of COQ4 in S. cerevisiae causes the instability of several Coq proteins, such as Coq7 and Coq3. Haploinsufficiency of COQ4 leads to encephalomyopathy in humans.⁴⁷⁾

Coq5/UbiE/COQ5/C-methyltransferase

S. cerevisiae Coq5, E. coli UbiE, and H. sapiens COQ5 catalyze the C-methylation step in the CoQ biosynthetic pathway. These enzymes contain four sequence motifs that are present in a large family of SAM-dependent methyltransferase enzymes. Demonstration of the enzymatic activity of Coq5 methyltransferase using 2-methoxy-6-polypreny-1,4-benzoquinone revealed its exact role in CoQ synthesis.⁴⁸⁾ E. coli UbiE complements the function of Coq5 in S. cerevisiae,⁴⁸⁾ and human COQ5 associates with COQ4. Analysis of the crystal structure of S. cerevisiae Coq5 in both the apo and SAM-bound forms revealed that it forms a dimer and contains a typical SAM domain (Fig. 4).⁴⁹⁾

Coq6/UbiI/COQ6/monooxygenase

S. cerevisiae Coq6, E. coli UbiI, and H. sapiens COQ6 are putative flavin-dependent monooxygenases that are responsible for introducing a hydroxy group to 4-hydroxy-3-polyprenyl benzoic acid. The Coq6 protein contains three conserved regions: an ADP-binding motif, a FAD/NADH-binding motif, and a consensus sequence that binds to the ribityl moiety of FAD.⁵⁰⁾ A recent analysis demonstrated that Coq6 is involved in C-5 oxidation during the early stage of CoQ synthesis. The involvement of mitochondrial ferredoxin (YAH1) and its reductase (ARH1) in CoQ synthesis has also been proposed.⁵¹⁾ Recently, E. coli UbiI was shown to be a counterpart of Coq6.¹⁴⁾ Mutations in the human COQ6 gene cause nephrotic syndrome (Table 2).⁵²⁾

Coq7/UbiF/COQ7/monooxygenase

S. cerevisiae Coq7, E. coli UbiF, and H. sapiens COQ7 proteins are putative flavin-dependent enzymes that catalyze monooxygenation in the penultimate step of CoQ biosynthesis. The Coq7 protein belongs to a family of di-iron-binding oxidases that contain a conserved EXXH motif. The clk-1/coq7 mutant of C. elegans, which accumulates the CoQ precursor demethoxyubiquinone (DMQ), shows a prolonged life span, developmental delay, and low egg production.^53–55) DMQ is also accumulated in the S. pombe coq7 mutant; however, no apparent role of DMQ has been observed in this species.⁴⁴⁾ E. coli UbiF also catalyzes the same step of CoQ biosynthesis as Coq7 and C. elegans Clk-1. The Coq7, Clk-1, and UbiF proteins are highly conserved among different kingdoms, but analyses of DNA sequences revealed no apparent ortholog in plants. PKA-mediated phosphorylation and Ptc7-mediated dephosphorylation of Coq7 and other Coq proteins have been reported in S. cerevisiae.⁵⁶⁾

Coq8/UbiB/ADCK3 (ADCK4)/protein kinase

S. cerevisiae Coq8 and H. sapiens ADCK3 (or ADCK4) are protein kinases involved in CoQ synthesis. In S. cerevisiae, COQ8 has now been approved as an official gene name because the previously used name (ABC1), which was based on a chaperone of the bc1 complex, has been questioned.⁵⁷⁾ The S. pombe coq8 gene is also essential for CoQ biosynthesis.⁵⁸⁾ Coq8 has been classified as a putative protein kinase based on the presence of conserved kinase motifs in its primary structure. A recent study demonstrated that Coq8 is involved in the phosphorylation of Coq3, either directly or indirectly, and now, it is considered to be a regulator of the Coq enzyme complex.⁵⁹⁾ The structure of human ADCK3 has been solved.⁶⁰⁾ There are five Coq8-like kinases (ADCK1–5) in humans; ADCK3 is involved in CoQ synthesis, and mutations in the ADCK4 gene are related to human CoQ10 deficiency.⁶¹⁾ In fact, mutations in the ADCK3 and ADCK4 genes have been linked with a number of human genetic diseases (Table 2);⁶²⁾ however, it is unclear whether the other ADCK proteins are also involved in CoQ biosynthesis.

Coq9/COQ9

The COQ9 gene is absolutely required for CoQ biosynthesis in S. cerevisiae⁶³⁾ and S. pombe;¹⁵⁾ however, the function of the encoded protein is still unknown. Although the Coq9 protein is conserved in eukaryotes, it has no primary sequence homology with known proteins. Furthermore, there is no apparent ortholog of COQ9 in E. coli. The Coq9 protein is a component of the multi-subunit CoQ biosynthetic complex⁶⁴⁾ and is required for the removal of the nitrogen substituent from pABA-derived Q.⁶⁵⁾ Analysis of the crystal structure of human COQ9 (Fig. 4)⁶⁶⁾ revealed that it contains a lipid-binding site and is similar to the bacterial TetR family of transcriptional regulators. Coq9 associates with Coq7 in humans. Notably, the relationship between a human genetic disorder and a mutation in the human COQ9 gene suggests its involvement in the ring modification of CoQ (Table 2).⁶⁷⁾

Coq10/COQ10A and COQ10B/ CoQ-binding proteins

Although Coq10 is not involved in its biosynthesis directly, it is a unique binding partner of CoQ. The Coq10 protein is localized to the mitochondria but does not belong to the respiration complex in S. cerevisiae.⁶⁸⁾ The existence of a CoQ-binding protein in mitochondria challenges the current model that CoQ is a free lipid molecule in membranes. Recently, Coq10 from S. pombe was also characterized as a mitochondrial CoQ-binding protein that is required for proper respiration.⁶⁹⁾ Further characterization of Coq10 suggested that it is essential for proper functioning of the electron transfer system, possibly by assisting in the transfer of CoQ from one site to another in the mitochondrial membranes of eukaryotes.⁷⁰⁾ A photo-affinity labeling experiment revealed that a FVPFCQK sequence in Coq10 is responsible for binding to CoQ.⁷¹⁾

Coq11

Very recently, Coq11 was identified as a protein associated with the Coq biosynthetic complex.⁷²⁾ The COQ11 gene is not absolutely required for CoQ synthesis, but its deletion reduces the CoQ level in S. cerevisiae.

Three-dimensional structures of proteins involved in CoQ synthesis

The three-dimensional structures of nine proteins responsible for CoQ synthesis (IspB, UbiA, UbiC, UbiD, UbiX, UbiI, S. cerevisiae Coq5, H. sapiens ADCK3, and H. sapiens COQ9), as well as a protein that is homologous to Coq4, have been determined. The structures of the eukaryotic proteins or their homologs found in other organisms are shown in Figure 4. E. coli IspB, which is a homolog of Coq1, consists of 14 α-helices.²⁸⁾ The recent cocrystallization of IspB with its substrates (farnesyl diphosphate and isopentenyl diphosphate) revealed that the substrate-binding regions are located around aspartate-rich motifs and identified a pocket that determines the acceptable chain length. Analysis of the crystal structure of A. pernix UbiA, a homolog of Coq2, showed that it comprises nine transmembrane structures and identified the proposed substrate (PHB) binding and entry sites.³⁸⁾ Investigation of a truncated form of E. coli UbiI, which is a homolog of Coq6, revealed that it forms a tetramer that resembles the structure of typical flavin-dependent monooxygenases.¹⁴⁾ The crystal structure of Coq5 from S. cerevisiae resembles that of a typical class-I SAM-dependent methyltransferase,⁴⁹⁾ and the residues involved in the interaction of Coq5 with SAM are located on four loops. Examination of the crystal structure of a truncated form of ADCK3, a member of the UbiB family of protein kinases, revealed an atypical protein kinase with multiple UbiB-specific features positioned to inhibit protein kinase activity.⁶⁰⁾ These inhibitory regions include an N-terminal domain that occupies the typical substrate-binding pocket, and a unique A-rich loop that limits ATP binding.⁶⁰⁾ This structure would explain why in vitro kinase activity was not detected for UbiB. H. sapiens COQ9 displays a striking structural homology to members of the TetR family of regulators and contains a lipid-binding pocket. The protein consists of nine α-helices and forms a dimer, and the deduced crystal structure included a phospholipid in its hydrophobic interface.⁶⁶⁾ In addition to the structures of five proteins that are apparently involved in CoQ synthesis, the crystal structure of the cyanobacteria Alr8543 protein (Nostoc sp. PCC7120), which is homologous to Coq4, has also been solved.⁷³⁾ The direct involvement of the Alr8543 protein in CoQ synthesis has not been proven, but its cocrystallization with geranylgeranyl monophosphate supports a role as a substrate holder during CoQ synthesis.⁷³⁾

In addition to the three-dimensional structures of Coq proteins and their homologs, the crystal structures of the bacterial decarboxylases UbiD (PDB ID: 4IP2)⁷⁴⁾ and UbiX (PDB ID: 1SBZ)⁷⁵⁾ have also been solved. A UbiX homolog named Pad1 exists in yeast, but it is not thought to be involved in CoQ synthesis.⁷⁶⁾ The crystal structure of the bacterial chorismate lyase UbiC (PDB ID: 1TT8) has also been solved, but its counterpart is not found in eukaryotes.⁷⁷⁾

Complex formation

In S. cerevisiae, the enzymes involved in CoQ synthesis are thought to form a large complex termed as CoQ synthome (Fig. 5). In this complex, Coq2 spans the inner membrane and the other enzymes are peripherally associated with this membrane on the matrix side. Native polyacrylamide gel electrophoresis analyses identified a 700 kDa band containing Coq3, Coq5, Coq6, Coq9, and Coq4 as well as a larger 1,300 kDa band containing Coq7 also. Coq4 plays a central role in the function of the CoQ synthome and associates with Coq3, Coq5, Coq6, and Coq7. Coq9 associates weakly with the complex, and Coq2 and Coq8 function separately.⁷⁸⁾ Coq8 is a kinase that is involved in complex formation through the phosphorylation of Coq3, Coq5, and Coq7; however, it is unclear whether Coq8 phosphorylates these proteins directly or indirectly. Several lines of evidence indicate that the human CoQ biosynthetic enzymes also form a complex;^66,79) this topic has been comprehensively reviewed by Gonzalez-Mariscal et al.⁸⁰⁾

Fig. 5. Structure of the CoQ biosynthetic enzyme complex. Notes: The enzymes involved in CoQ synthesis form a complex in S. cerevisiae.⁷⁸⁾ There is also some evidence that this complex exists in humans. This figure is modified from the figure reported by Allan et al.⁷²⁾ Proteins in the figure are not proportional to the actual molecular sizes. The structure of S. cerevisiae Coq5 has been solved; for the other enzymes, the structures of homologs from other species are indicated. E. coli IspB is shown as a Coq1 homolog, A. pernix UbiA is shown as a Coq2 homolog, Alr8543 from Nostoc sp. PCC7120 is shown as a Coq4 homolog, E. coli UbiI is shown as a Coq6 homolog, human ADCK3 is shown as a Coq8 homolog, and human COQ9 is shown as a Coq9 homolog. Coq1 is separated from the complex. Coq2 spans the inner membrane. The positions of the other proteins have not been defined experimentally, although Coq4 seems to be in the center, and Coq7 and Coq8 seem to be located at the edge of the complex.

Diseases caused by CoQ deficiency

The diseases caused by a deficiency in CoQ have been reviewed previously.^81,82) The first mutations related to human CoQ deficiency were identified in COQ2; since then, mutations in seven other CoQ biosynthetic genes (PDSS1, PDSS2, COQ4, COQ6, ADCK3, ADCK4, and COQ9) have been shown to cause various diseases (summarized in Table 2). Encephalomyopathy, nephropathy, cerebellar ataxia, and seizures are common features of CoQ deficiency; these symptoms are associated with mitochondrial disorders. The level of CoQ10 differs between patients and depends on the specific mutations, and some disease symptoms are eased by supplementation with CoQ10. There are also reports that secondary CoQ deficiency can be caused by mutations in genes that are not directly involved in CoQ synthesis, such as oncogene B-RAF and aprataxin (APTX). ^81,82)

Sites of CoQ synthesis

In eukaryotes, CoQ biosynthesis occurs mainly in the mitochondria; indeed, all S. cerevisiae and S. pombe proteins involved in the process localize to this cellular compartment.¹⁵⁾ Human COQ proteins also localize to the mitochondria, but the Golgi-localized enzyme UBIAD1 (prenyl transferase), which is homologous to E. coli UbiA, can also participate in CoQ.⁸³⁾ Loss of UBIAD1 apparently reduces the cytosolic pool of CoQ10. As most of other enzymes involves CoQ localization to mitochondria, the mechanism by which UBIAD1 assists with CoQ production in the Golgi is still unclear.

Regulation of the expression of COQ genes and genes regulated by CoQ

Little is currently known about the regulation of COQ gene expression. The expression of S. cerevisiae COQ5 is reportedly regulated by Mig1, Rtg3, and Hap2.⁸⁴⁾ A microarray analysis revealed that a deficiency of endogenous CoQ in C. elegans clk1 mutants down-regulates a cluster of genes that are important for growth and up-regulates oxidation reactions and protein interactions.⁸⁵⁾ In addition, a microarray analysis demonstrated the induction of specific genes related to cell survival in CoQ-deficient human cells.⁸⁶⁾

Bioproduction of CoQ10 in microorganisms and plants

Because CoQ10 is a commercially sold food supplement, its efficient production by microorganisms has been explored.⁸⁷⁾ E. coli, Agrobacterium, and photosynthetic bacteria such as Rhodobacter spheroides and Rhodobacter capsulatus have been used to produce CoQ10. E. coli produces CoQ8 naturally, but expressing decaprenyl diphosphate synthase in cells that lack the endogenous ispB gene enables them to produce CoQ10.²¹⁾ A previous study generated an engineered E. coli strain that expressed the Agrobacterium tumefaciens decaprenyl diphosphate synthase gene (ddsA) and had a strengthened mevalonate pathway; this strain was capable of producing substantial amounts of CoQ10.⁸⁸⁾ By optimizing the growth medium and conditions, photosynthetic Rhodobacter capable of producing CoQ10 at concentrations up to 8.70 mg/mg dry cell weight was generated.⁸⁹⁾ In addition, Agrobacterium can produce CoQ10 at concentrations up to 11.84 mg/mg dry cell weight.⁹⁰⁾ The fission yeast S. pombe is a good candidate for CoQ10 production, and the yield can be doubled by genetic engineering, although it is not as high as that obtained from other microorganisms.⁹¹⁾ In addition to microbial production, researchers have attempted to produce CoQ10 using other systems; the attempts to produce CoQ10 in rice are worth noting.^92,93)

Concluding remarks

Recent studies of the mechanisms of CoQ synthesis have been very fruitful. In particular, current understanding of the correlation between defective genes involved in CoQ synthesis and human diseases is surprisingly advanced. The three-dimensional structures of proteins involved in CoQ synthesis have also been determined recently. Despite this accumulating knowledge of CoQ biosynthesis, we still do not have a clear picture of the whole biosynthetic pathway in eukaryotes. For example, Coq4 and Coq9 are absolutely required for CoQ biosynthesis in eukaryotes, but their specific functions are unclear. Furthermore, additional factors, such as Coq10 and Coq11, also appear to be involved in CoQ biosynthesis. Besides the electron transport system, reactions involving sulfide-quinone oxidoreductase, glycerol-3-phosphate dehydrogenase, dihydroorotate dehydrogenase, and electron-transferring flavoprotein dehydrogenase require CoQ as a cofactor. Overall, additional work is required to understand the complete pathway of CoQ biosynthesis.

Disclosure statement

No potential conflict of interest was reported by the author.

Acknowledgment

The author would like to thank Dr. T. Kaino for critical reading of the manuscript.

Additional information

Funding

This work was supported by a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan [grant number 24380056].

Notes

Abbreviations: CoQ, coenzyme Q; DMQ, demethoxyubiquinone; pABA, para-amino benzoic acid; PHB, para-hydroxybenzoate; SAM, S-adenosyl methionine.