Ubiquinone . Biosynthesis of quinone ring and its isoprenoid side chain . Intracellular localization

Ubiquinone, known as coenzyme Q, was shown to be the part of the metabolic pathways by Crane et al. in 1957. Its function as a component of the mitochondrial respiratory chain is well established. However, ubiquinone has recently attracted increasing attention with regard to its function, in the reduced form, as an antioxidant. In ubiquinone synthesis the para-hydroxybenzoate ring (which is the derivative of tyrosine or phenylalanine) is condensed with a hydrophobic polyisoprenoid side chain, whose length varies from 6 to 10 isoprene units depending on the organism. para-Hydroxybenzoate (PHB) polyprenyltransferase that catalyzes the condensation of PHB with polyprenyl diphosphate has a broad substrate specificity. Most of the genes encoding (all-E)-prenyltransferases which synthesize polyisoprenoid chains, have been cloned. Their structure is either homoor heterodimeric. Genes that encode prenyltransferases catalysing the transfer of the isoprenoid chain to para-hydroxybenzoate were also cloned in bacteria and yeast. To form ubiquinone, prenylated PHB undergoes several modifications such as hydroxylations, O-methylations, methylations and decarboxylation. In eukaryotes ubiquinones were found in the inner mitochondrial membrane and in other membranes such as the endoplasmic reticulum, Golgi vesicles, lysosomes and peroxisomes. Still, the subcellular site of their biosynthesis remains unclear. Considering the diversity of functions of ubiquinones, and their multistep biosynthesis, identification of factors regulating their cellular level remains an elusive task. Vol. 47 No. 2/2000

Ubiquinone is a component of the mitochondrial respiratory chain [1], participating in electron transport in NADH-coenzyme Q reductase (complex I), succinate coenzyme Q reductase (complex II) and the cytochrome system.Folkers and his group [2] determined the structure of the quinone moiety which was found identical to that described by Morton and his team [3], and suggested the name "ubiquinone" referring to the ubiquitous occurrence of this compound in various tissues.The name ubiquinone (coenzyme Q) was officially accepted in 1975 by the IUPAC-IUB Commission on Biochemical Nomenclature.
The growing interest in ubiquinones is fully justified.In addition to participating in the respiratory chain, they are involved in the redox processes taking place in cytoplasmic as well as Golgi system membranes.In reduced form, as antioxidants, they efficiently protect membrane phospholipids and lipoproteins from lipid peroxidation, as well as membrane proteins and DNA from free radical-induced oxidative damage [4][5][6].In plants similar roles are played by plastoquinones [7].Some vitamins (E, K) are quinone derivatives.
The development of molecular biology and recombinant DNA technologies, the discovery of restriction enzymes and construction of vectors made new approaches possible to investigate ubiquinones.They allowed the identification and cloning of genes encoding enzymes participating in modifications of the ring moiety of the ubiquinone molecule; methyltransferases: UBIG, UBIE from bacteria and COQ3 from yeast; hydroxylases UBIB, UBIH from bacteria and COQ6 from yeast.Genes encoding prenyltransferases catalyzing the transfer of the isoprenoid chain to para-hydroxybenzoate in bacteria (UBIA) and yeast (COQ2) have also been cloned.Practically all the bacterial genes responsible for the synthesis of hexa-up to decaprenyl diphosphate synthases forming the ubiquinone polyisoprenoid side chain have been cloned.These results will allow identification of the respective human genes, characterization of their structure, level and site of expression which is of vital importance considering the functions of ubiquinones.

SYNTHESIS OF QUINONE MOIETY
The benzene moiety is derived mainly from tyrosine (in some cases from phenylalanine) converted to para-hydroxybenzoate [8,9] (Fig. 1) which in turn is condensed with all-E polyisoprenoid diphosphate.A number of subsequent modifications of the ring are required for the completion of ubiquinone [10] (Fig. 2).
Modifications of para-hydroxybenzoate (4-OH-benzoate) condensed with a polyisoprenoid side chain start with C-hydroxylation, followed by O-methylation and decarboxylation.Two additional C-hydroxylations and one O-methylation are necessary for the final synthesis of ubiquinone.The first data on methylation and O-methylation of the benzene ring come from the elegant chemical work of Olson et al. [11].The above sequence of events has been established in a bacterial system [12].However, the results obtained by Kang et al. [13] suggest that in an animal system decarboxylation may occur prior to the first methylation [13].Genes UBIG and COQ3 encode an S-adenosylmethionine O-methyl transferase involved in O-methylation in bacteria and yeast, respectively (Fig. 2) [14,15].The UBIH gene encodes a mono-oxygenase thought to contain flavin adenine nucleotide that catalyses the conversion of the 6-octaprenyl-2-methoxyphenol to 6-octaprenyl-2-methoxy-1,4-benzoquinone [16].Recently a novel gene ohb1 encoding a reversible parahydroxybenzoate decarboxylase from Clostridium hydroxybenzoicum was cloned and characterized [17].Its amino-acid sequence shows 57% identity and 74% similarity to that of hypothetical proteins deduced from open reading frames in genomes from bacteria and archaea, suggesting the possible existence of a novel gene family.These genes could have encoded an ancient type of decarboxylase, which was replaced during evolution by a more efficient one [17].
Not all the genes that encode enzymes participating in ubiquinone biosynthesis in higher eukaryotic cells have been cloned.It seems however, that they are similar to the yeast genes.Indeed, a rat cDNA homologue to UBIG and COQ3 has been cloned by functional complementation of a yeast coq3 mutant with a rat testis cDNA library [18].In plants the biosynthesis of ubiquinones starting from isopentenyl diphosphate and parahydroxybenzoate also requires a similar set of reactions i.e. hydroxylation, decarboxylation, O-methylation and methylation but neither the proteins nor the corresponding genes have been characterized so far [19].

SYNTHESIS OF THE POLYPRENYL SIDE CHAIN
The polyprenyl (isoprenoid) side chain of ubiquinone is synthesized from acetyl-CoA through a sequence of reaction named the mevalonic acid pathway, leading to the formation of farnesyl diphosphate (FPP) (Fig. 3).
The study of the isoprenoid biosynthesis was brought to an enzymatical level by the discovery made independently by the groups of  Lynen and Bloch in 1958 [20,21] of isopentenyl diphosphate (IPP) as the active isoprene unit.
It is worth mentioning that Rohmer et al. [22] discovered a novel biosynthesis pathway of IPP in which mevalonate is not involved.On the basis of labeling patterns of bacterial hopanoids and ubiquinones derived from 13 C-labeled metabolites of glycolysis as well as from [ 13 C]acetate, they proved that glyceraldehyde 3-phosphate occupies the branch point of the glycolysis leading to the non-mevalonate pathway synthesizing IPP [22,23].
The discovery of IPP isomerase, which converts IPP to dimethylallyl diphosphate (DMAPP) [24], as well as the prenyltransfe-rase that catalyses the head-to-tail condensations of IPP with DMAPP and with geranyl diphosphate (GPP), to form (E,E)-farnesyl diphosphate (FPP) (Fig. 4) has led to the recognition that a tremendous number of isoprenoid compounds exists in Nature.The chain length of prenyl diphosphates varies ranging from geraniol (C 10 ) to natural rubber whose carbon chain length extends to several millions.To date 16 prenyltransferases (enzymes that catalyze the transfer of prenyl groups to an acceptor which is usually IPP, but may also be an aromatic compound, a protein etc) with different catalytic functions have been characterized.
In contrast to dolichols composed of 14 to 23 isoprene units in yeast and man, the length of the side chain of ubiquinone varies to a much more limited extent.For instance, Escherichia coli produces mainly ubiquinones having 8 isoprene units, Saccharomyces cerevisiae, 6 isoprene units.Eukaryotic cells contain ubiquinones with 6-10 isoprene residues but in every species only one chain length dominates.
The short-chain prenyl diphosphates are also biosynthetic precursors of various isoprenoids including steroids, carotenoids and prenylated proteins.It is noteworthy that  GGPP (geranylgeranyl diphosphate) synthase occurs not only in bacteria and plants but also in mammals.FPP (Fig. 5) and GGPP synthases both have a homodimeric structure in which the subunits are tightly bound to each other [25].
In spite of the fact that all eukaryotic cells probably contain E-prenyltransferase activity, to date only a few investigations have been concerned with the characterization of this enzyme in animal cells.On the other hand, bacteria are good sources of prenyltransferases because they produce ubiquinones and menaquinones with polyprenyl side chains of chain lengths varying in a species-specific manner [26].

Hexaprenyl diphosphate synthases
It has been found that the quinone side chain consisting of 6 isoprene units is synthesized by hexaprenyl diphosphate (HexPP) synthase [EC 2.5.1.30]which has a novel subunit system (Fig. 6).The enzyme catalyzes the synthesis of (all-E)-HexPP by adding three molecules of IPP to FPP, but it can not catalyze the steps of synthesis of GPP or FPP from DMAPP and IPP [27], FPP is supplied by FPP synthase which occurs in all bacteria.HexPP synthase has been cloned from Micrococcus luteus B-P 26, and sequenced [28].Within a 2130 bp DNA which expresses HexPP synthase in E. coli cells, there are three consecutive open reading frames (ORF1 -143 aa, ORF2 -246 aa and ORF3 -325 aa).It has been established that the two dissociable components A and B of HexPP synthase are products of ORF1 and ORF3, respectively 1 .These genes were similarly located in a newly identified gene cluster participating in menaquinone biosynthesis [30].
Ashby & Edwards [31] have isolated from a plasmid containing a wild-copy genomic DNA fragment a gene (Coq1) that is able to complement the yeast mutant in coenzyme Q biosynthesis and restore its HexPP synthase activity.This enzyme which has conserved regions characteristic of prenyltransferases, seems to correspond to the larger protein component of the two heterodimeric components of bacterial HexPP synthase.However, it is not known whether the yeast protein acts as HexPP synthase by itself or in association with another gene product similar to the smaller protein component.

Heptaprenyl diphosphate synthases
Heptaprenyl diphosphate (HepPP) synthase [EC 2.5.1.30]has been found in Bacillus subtilis which produces exclusively menaquinone with a side chain of 7 isoprene units [32].It catalyzes the synthesis of (all-E) HepPP by adding four molecules of IPP to FPP but like HexPP synthase it does not synthesize C 5 ® C 10 ® C 15 compounds.On the basis of the highly conserved aminoacid sequences of prenyltransferases two genes encoding two protein components of HepPP synthase from Bacillus stearothermophilus were identified [33].One of the proteins constituting an enzyme system for the synthesis of HepPP has 323 amino-acid residues and shows 32% sequence similarity to the FPP synthase from the same bacterium [34].This protein designated as component II¢ has highly conserved regions typical of prenyltransferases.The other protein (component I¢) is composed of 220 amino acids and has no such similarity nor has any similar protein been found within the protein entries in data bases.Therefore, it seems likely that component II¢ carries the active sites for substrate binding and catalysis, while component I¢ plays an auxiliary role.A protein database search for amino-acid sequences similar to that of component II¢ of the HepPP synthase from B. stearothermophilus [33] yielded the GerC3 protein of B. subtilis which is encoded in a cluster of three ORFs, gerC1, gerC2 and gerC3.Two of the gerC products, GerC1 and GerC3, correspond to the two dissociable components I¢ and II¢ which constitute the HepPP synthase from B. stearothermophilus [35].

Octaprenyl diphosphate synthases
Bacterial (all-E) octaprenyl diphosphate synthase (forming an 8 isoprene residue polyprenyl chain) has been found and partially purified from E. coli [36].This enzyme is a homodimeric protein that is functionally active by itself.Analysis of genes in the E. coli chromosome performed by Choi et al. [37] revealed an open-reading frame that showed significant similarity to the ispA gene which encodes FPP synthase of E. coli [38].Jeong et al. [39] determined the entire sequence of this open reading frame and found a high similarity of the gene product to the HexPP synthase of S. cerevisiae [31] and the GGPP synthases of various organisms.Asai et al. [40] identified the ispB (cel) gene encoding the octaprenyl diphosphate synthase in E. coli.The deduced amino-acid sequence of this enzyme also shows the presence of the seven conserved regions typical of prenyltransferases [34].
Okada et al. [41] proved that E. coli octaprenyl diphosphate synthase (IspB) having a mitochondrial import signal is expressed in S. cerevisiae.Yeast cells produced ubiquinone 8 in addition to the originally existing ubiquinone 6.When IspB was expressed in a S. cerevisiae COQ1 defective strain, IspB complemented the defect of growth on a non-fermentable carbon source.

Nonaprenyl and decaprenyl diphosphate synthases
(all-E) Nonaprenyl (solanesyl) diphosphate synthase (SPP) [EC 2.5.1.11]was isolated from M. luteus [42] as an enzyme catalyzing chain elongation from GPP up to C 45 -PP.The enzyme is a homodimeric protein [43].However, it requires a protein factor to maintain efficient catalytic turnover.This factor is a high molecular mass component of a soluble fraction of the same bacterium.Possibly it acts by removing from the active site the hydrophobic products which otherwise would inhibit the reaction.
Decaprenyl diphosphate synthase has been found and partially purified from Paracoccus denitrificans [44].It is also a homodimeric protein utilizing GPP as substrate.Suzuki et al.
[45] and Takahashi et al. 2 have cloned the gene encoding the decaprenyl diphosphate synthase of the fission yeast Schizosaccharomyces pombe and P. denitrificans, respectively.
Both nona-and decaprenyl diphosphate synthases are present in plants.They have been characterized in spinach leaves [47].As does the bacterial enzyme, they utilize GPP as substrate.
In all human tissues the predominant side chain of ubiquinones is decaprenol and only small proportion (2-7%) of nonaprenol is present.

TRANSFER OF POLYPRENYL DIPHOSPHATES TO para-HYDROXYBENZOATE
The E. coli gene UBIA encoding prenyltransferase catalysing the transfer of the isoprenoid side chain to para-hydroxybenzoate has been isolated [51].The UBIA of E. coli and COQ2 of S. cerevisiae genes have been analyzed and cloned [52][53][54].Lutke-Brinkhaus et al. [55] have described PHB-polyprenyltransferase activity in plant mitochondria.The enzyme has a very broad substrate specificity.

LOCALIZATION OF QUINONE SYNTHESIS
Knowledge about the intracellular localization of the ubiquinone synthesizing enzymes is scant.They are believed to be membranebound or membrane associated [56].It is thought that ubiquinone synthesis in cells of higher eukaryotes reflects the yeast pathway.
In plant cells, the precise localization of this biosynthetic pathway is a matter of controversy.Lutke-Brinkhaus et al. [55] and Lutke-Brinkhaus and Kleinig [57] found that plant mitochondria purified from potato tubers catalyze several reactions in the biosynthesis of ubiquinones.In contrast, OEwie¿ewska et al. [47] found nonaprenyl-4-hydroxybenzoate and nonaprenyl-2-methylquinol transferase activities in microsomal and Golgi preparations, and proposed therefore, that ubiquinone biosynthesis occurs in the endoplasmic reticulum-Golgi system, followed by a selective transfer of ubiquinone to the mitochondria.
In animal systems the enzymes of the mevalonic acid pathway which supply polyprenyl side chain have a multilocal intracellular distribution.In addition to mitochondria [49] they are present in the endoplasmic reticulum [58], peroxisomes [59], Golgi system, lysosomes, plasma mem-Vol.47 Ubiquinone biosynthesis 475 The enzyme catalyzes the synthesis of (all-E) hexaprenyl diphosphate by adding three molecules of isopentenyl diphosphate (IPP) to farnesyl diphosphate (FPP) [27].
brane [60] and cytosol [61].According to Dallner and his group [62] ubiquinone synthesis begins in the endoplasmic reticulum and is completed in the Golgi system, after which ubiquinone becomes distributed to other cellular organelles.

REGULATION OF UBIQUINONE SYNTHESIS
The factors influencing the regulation of ubiquinone cellular level are numerous.In view of the complexity of their function, structure and localization of synthesis, ubiquinones might be the subject of another review article.Only a few general points will be mentioned here.
Ubiquinone levels are regulated through the mevalonate pathway.Moreover, there is an intricate interplay among the three major biosynthetic products of mevalonate metabolism, i.e. ubiquinone, dolichol and sterol (Fig. 3) [63,64].
In E. coli the composition of the ubiquinone pool is highly influenced by the degree of oxygen availability: aerobically grown E. coli cells contain significantly more ubiquinone 8.The mechanism of this regulation is not yet shown [65].In yeast under glucose derepression the level of ubiquinone 6 synthesized increases.
Much less is known about the physiological regulation of ubiquinone level.No data is available concerning plants.In animal systems it has been observed that the ubiquinone level increases upon various forms of oxidative stress (physical exercise, cold adaptation, thyroid hormone treatment) and decreases with age.Relatively little is known about the mechanism involved in biodegradation of ubiquinone.The turnover rate of ubiquinone in various tissues is rather similar, ranging from 50 to 125 h [66].This is in contrast to cholesterol and dolichol for which the turnover rates are several orders of magnitude higher in the liver than in the brain [67].
The decrease in ubiquinone content with increasing age is consistent with the free radical theory of aging [68], as reflected by an inverse correlation between longevity and peroxide-producing potential in mammalian tissues [69].This decrease may also account for the age-related increase in the extent of oxidative damage to proteins [70] and DNA [71], the latter especially to mitochondrial DNA [72,73].Increase of ubiquinone concentration was found in neurodegenerative conditions in the brain, such as in Alzheimer's disease [74] and prion disease in mice [75].It is not yet clear whether these diseases are related to an alteration in the bioenergetic capacity and/or in the antioxidant status of the tissues concerned.

Figure 5 .
Figure 5. Yeast farnesyl diphosphate synthase homodimer model.The model was constructed according to the avian crystal structure[25] by D. P³ochocka from Bioinformatic Dept. of Inst.Biochem.Biophys.PAS, Warszawa.One monomer is in ribbon representation, the other is shown as a backbone.

Figure 6 .
Figure 6.The dynamic formation and dissociation of the ternary complex of the heteromeric components of hexaprenyl diphosphate synthase.