Effect of Concomitant Lycopene Biosynthesis on CoQ10 Accumulation in Transformed Escherichia coli Strains

Document Type : Research Paper

Authors

1 Department of Molecular Genetics, National Institute of Genetic Engineering and Biotechnology (NIGEB), P.O. Box 14965/161, Tehran, I.R. Iran

2 Department of Biochemistry, Payame Noor University, P.O. Box 19569, Tehran, I.R. Iran

3 Department of Agricultur, Science and Research Campus of Azad University, P.O. Box 1477893855, Tehran, I.R. Iran

Abstract

CoQ10 and lycopene are isoprenoid compounds with nutraceutical and pharmaceutical benefits. In this study, the effect of concomitant lycopene biosynthesis on CoQ10 accumulation in transformed Escherichia coli DH5α was studied. A lycopene production pathway including geranylgeranyl diphosphate synthase (crtE), phytoene synthase (crtB), and phytoene desaturase (crtI) from Erwinia herbicola was constructed in two CoQ10-producing E. coli strains. E. coli Ba and E. coli Br containing dds orthologs encoding for decaprenyl diphosphate synthase (Dds), respectively from Agrobacterium tumefaciens and Rhodobacter sphaeroides were transformed by the lycopene pathway resulting in E. coli Ba-lyc and E. coli Br-lyc. The lycopene pathway in E. coli Br-lyc interestingly resulted in a significant increase in CoQ10 production from 564 ± 28 to 989 ± 22 mg /g DCW. To confirm that the improvement of CoQ10 production in E. coli Br-lyc was due to lycopene biosynthesis and not just geranylgeranyl diphosphate formation in the lycopene pathway, crtE was only introduced into E. coli Ba and E. coli Br strains. Surprisingly, crtE expression had adverse effects on CoQ10 production in both strains. The results shed light on the Dds-catalyzed reaction as a bottleneck controlled by precursors; and the efficiency of a parallel lycopene pathway to streamline the flow of metabolites.

Keywords


INTRODUCTION

Ubiquinone or coenzyme Q (CoQ) is a lipid-soluble molecule of the electron transport chain located in the mitochondrial inner membrane of human cells as well as the cytoplasmic membrane of bacteria. This coenzyme is essential in energy generation and some other processes including formation of disulfide bonds in proteins, regulation of gene expression and detoxification of reactive oxygen species (ROS) (Groneberg et al., 2005; Søballe and Poole, 2000; Alleva et al., 1995; Ernster et al., 1995; Kontush et al., 1995).
CoQ is generally composed of a benzene ring and an isoprenoid side chain comprised of isoprene units (isopentenyl diphosphate, IPP) (Fig. 1). The structure of the ring is common; however, the length of the side chain is a characteristic of various species of CoQs in different organisms. For example, the human isoprenoid side chain is comprised of 10 isoprene units resulting in the formation of CoQ10. Other organisms may use different kinds of CoQs as electron and proton transporters in their own respiratory chains. For example, Escherichia coli contains CoQ8 while Saccharomyces cerevisiae has CoQ6 (Okada et al., 1998; Suzuki et al., 1997; Meganathan, 1996; Ernster, 1995; Grunler, 1994). The synthesis of the side chain of CoQ10 is catalyzed by decaprenyl diphosphate synthase; while in the case of CoQ8, the octaprenyl diphosphate synthase is the responsible enzyme. Accordingly, all that is required for the production of a given ubiquinone in a living cell is the expression of a heterologous gene coding for the corresponding polyprenyl diphosphate synthase (Kawamukai, 2002; Okada et al., 1996).
      The attachment of decaprenyl diphosphate to a benzene ring and some subsequent ring modifications result in CoQ10 production. Dds may use IPP, farnesyl diphosphate (FPP) and geranylgeranyl diphosphate (GGPP) as substrates for addition of more IPP molecules to make decaprenyl diphosphate.  
A condensation reaction between an isopentenyl diphosphate (IPP) and its allelic isomer, dimethylalyl diphosphate, results in the formation of geranyl diphosphate (GPP) with 10 carbons. The addition of a third IPP to GPP results in the formation of farnesyl diphosphate (FPP) with 15 carbons. Likewise, one more IPP may be added to FPP to make a 20-carbon molecule called geranylgeranyl diphosphate (GGPP). The latter two reactions are catalyzed by FPP synthase and GGPP synthase, respectively. IPP, GPP, FPP and GGPP are potential substrates for subsequent additions of IPP units by decaprenyl diphosphate synthase during CoQ10 biosynthesis (Fig. 1). Several decaprenyl diphosphate synthases have been purified and characterized from various organisms (Takahashi et al., 2003; Søballe and Poole, 2000; Suzuki et al., 1997). FPP and GGPP have been shown to be first and second most appropriate substrates for all the enzymes (Kawamukai, 2002).            
Lycopene is a natural pigment which is biosynthesized from isoprene units as is CoQ10. Two GGPP molecules are condensed head to head by phytoene synthase to make a 40-carbon molecule. The resulting phytoene molecule is then converted to lycopene via two tandem reactions catalyzed by an enzyme called phytoene desaturase (Sandmann, 2002) (Fig. 2).
Both CoQ10 and lycopene are known as natural antioxidants beneficial in keeping health and prevention of aging and ailments such as cardiovascular and neurodegenerative diseases and cancer. The separate production of CoQ10 and lycopene in E. coli has been investigated through several studies; but, there is very limited, if any, study on co-production of both materials in the same cell. The strategy of co-production of two metabolites in the same cell is indeed of importance from a biotechnology view point. However, the co-production of lycopene and CoQ10 was not the main target of this study. Instead, the current study was conducted to reveal the effect of concomitant lycopene biosynthesis on CoQ10 accumulation in transformed E. coli. To this end, two CoQ10-producing E. coli strains containing orthologs of decaprenyl diphosphate synthase from Agrobacterium tumefaciens and Rhodobacter sphaeroides were used for concomitant lycopene production.


MATERIALS AND METHODS

Media and culture conditions: Escherichia coli DH5a was grown on Luria-Bertani (LB) medium. In CoQ10 production experiments, 5ml of 2YT medium (1% yeast extract, 1.6% tryptone, 0.5% NaCl) was used for overnight precultivation of E. coli strains in a test tube at 37ºC. The main culture growth was conducted by using the same medium (7 ml) supplemented with 0.5% glycerol and 0.01% 4-hydroxybenzoic acid (sodium salt, Sigma, USA) (2YTGH), in 25 ml tubes Ampicillin (100 mg/ml) and kanamycin (30 mg/ml) were added to the culture media as required. Main cultures were inoculated to an initial OD600 of 0.1 and incubated at 30°C for 24 h in a rotary shaking incubator at 200 rpm (GFL, 3033). Cell growth was monitored by measuring turbidity at 600 nm, followed by conversion to dry cell weight (g DCW/l) using a standard curve.

Plasmids construction: E. coli DH5a [F´/endA1 hsdR17 (rK– mK+) glnV44 thi-1 recA1 gyrA (Nalr) relA1 D (lacIZYA-argF) U169 deoR 5(j80dlac D (lacZ) M15] was used as a  host for plasmid construction and CoQ10 production (Table 1). Restriction enzymes were purchased from Fermentas (Germany). T4 DNA ligase was purchased from New England Biolabs. Pfu DNA polymerase was purchased from Promega. Plasmid preparation, gel extraction and PCR purification kits were procured from Qiagen.
The genes encoding for decaprenyl diphosphate synthase (dds) were isolated from Rhodobacter sphaeroides 2.4.1 (ATCC 17023) and Agrobacterium tumefaciens (ATCC 33970) as reported elsewhere (Zahiri et al., 2006a). The dds gene of Rhodobacter sphaeroides 2.4.1 (rsdds) was obtained by double digestion of pTrsdds with EcoRI and SacI (Table 1). The resulting fragment was inserted into a pBBR1MCS2 plasmid which was previously linearized by the same enzymes, thus giving rise to pBrsdds (Fig. 3). The atdds gene, coding for decaprenyl diphosphate synthase of A. tumefaciens, was isolated from pTatdds by PCR using pfu DNA polymerase and oligonucleotides atdds-3 and atdds-4 as forward and reverse primers, respectively (Table 1). In the design of atdds-F a silent mutation was introduced to remove an inside EcoRI site from atdds. The PCR product was double digested by XmaI and inserted into pBBR1MCS2 to generate pBatdds (Fig. 3 and Table 1). The gene encoding for geranylgeranyl diphosphate synthase (crtE) was isolated from the Erwinia herbicola (ATCC 21434) genome using a PCR with crtE-F and crtE-R primers (Hundle et al., 1994). The PCR product was double digested with EcoRI and XbaI enzymes and ligated into pTrc99A linearized with the same enzymes. The resulting plasmid designated as pTcrtE was subsequently used in the construction of an artificial operon for lycopene production. The overlapping genes encoding for phytoene synthase (crtB) and phytoene desaturase (crtI) were isolated from E. herbicola as a single DNA fragment by a PCR using crtIB-F and crtIB-R primers (Hundle et al., 1994). The fragment was then inserted into pTcrtE after both the insert and vector had been restricted with XbaI and HindIII. The resulting plasmid was designated as pTcrtEIB.  The recombinant plasmids were then introduced into E. coli cells using a modified chemical transformation method (Casali and Preston, 2003).

Determination and quantification of ubiquinones: The production of various coenzyme Qs in E. coli strains was analyzed by high performance liquid chromatography (HPLC). CoQs were extracted from E. coli cells with hexane as explained elsewhere with some modifications. Briefly, a   500 ml sample of each culture was centrifuged at 13000 rpm (Heraeus, Biofuge) and the resulting pellet was washed once with 1 ml of distilled water and then with 1 ml of 20mM Tris-HCl, pH 7.6. The washed cells were lysed by being resuspended in lysis buffer [8% sucrose, 5% Triton X-100, 50 mM Tris-HCl (pH 8.0), 50 mM EDTA (pH 8.0) and 1 mg/ml of lysozyme] and incubated at 37°C for 30 min. CoQs were extracted from the cell lysate with 900 ml of a hexane/propanol (5:3 v/v) mixture and then with 500 ml of hexane. Following each extraction, the hexane organic phase containing CoQ10 was separated from the aqueous phase by centrifugation at 14,000 rpm and pooled in a clean tube. The hexane extract was dried in a vacuum evaporator (speedvac®, AES 1010). The remaining pellet was dissolved in 500 ml of absolute ethanol (HPLC grade, Fisher Scientific); whereof 10 ml was injected into a HPLC machine (Shimadzu 10A system) equipped with a Symmetry® C18 column (Waters). The chromatography was run at room temperature using a ethanol and methanol mixture (70:30 v/v, HPLC grade, Fisher Scientific) as a mobile phase at a rate of 1 ml/min. The detection and quantification of CoQs were performed using a UV detector at 275 nm. The corresponding peaks of CoQ8, CoQ9 and CoQ10 in the HPLC chromatograms were identified using authentic standards (Sigma).

Quantification of lycopene: Lycopene production in transformed E. coli strains was quantified through extraction by a mixture of methanol/chloroform (70:30 v/v) and determination of the optical density of the extracts at 474 nm (OD474). Cells were harvested by centrifugation of 1 ml of culture broth at 5000 rpm for 3 min. The pellet was then washed with water and resuspended in 400 ml of the methanol/chloroform mixture by vortexing for 5 min. After centrifugation, the supernatant was transferred to a clean tube and the extraction procedure was repeated once more. The supernatants were pooled together and the final volume of lycopene extract was raised up to one ml by adding more solvent. The cell debris remained as a white pellet at the end of extraction.  The optical density of the extract was converted to lycopene concentration by using a standard curve.         


RESULTS

Coenzyme Q10 production in E. coli: Decaprenyl diphosphate synthase-encoding genes of A. tumefaciens (atdds) and R. sphaeroides (rsdds) were used for transformation of E. coli into a CoQ10 producing organism. The heterologous genes were cloned in pBBR1MCS2. The resulting plasmids, pBatdds and pBrsdds, were used for transformation of E. coli DH5a. E. coli is naturally a CoQ8 producing organism. As a result of functional expression of the heterologous genes, the transformed E. coli DH5a cells, designated as E. coli Ba and E. coli Br, could produce CoQ10 in addition to CoQ8. All strains were grown in 2YTGH medium under the same experimental conditions, as described above. The CoQs, including CoQ8 and CoQ10, were extracted using hexane and determined quantitatively by HPLC. The expression of the heterologous genes in E. coli Ba and E. coli Br led to a significant decrease in  CoQ8 production as compared to that of E. coli DH5a (Fig. 4). The CoQ10 production of E. coli Ba was reproducibly higher than that of E. coli Br. On average, the CoQ10 content of E. coli Ba was approximately 629 ± 40 mg/gDCW, while that of E. coli Br approximately 564 ± 28 μg/gDCW under the described experimental conditions.

Expression of a heterologous lycopene pathway in CoQ10 producing E. coli: An artificial operon representing the lycopene production pathway was constructed in the pTrc99A plasmid. This operon was comprised of geranylgeranyl diphosphate synthase (crtE), phytoene synthase (crtB) and phytoene desaturase (crtI) encoding genes. The resulting plasmid, designated as pTcrtEIB, was used for transformation of E. coli DH5a resulting in E. coli Tlyc (Fig. 3). The genome of E. coli lacks a counterpart gene of crtE encoding for geranylgeranyl diphosphate (GGPP) synthase. As a result of the expression of crtE, 20-carbon GGPP molecules are synthesized from isopentenyl diphosphate precursors. Two GGPP molecules, in turn, are combined head to head by phytoene synthase resulting in the formation of a 40-carbon phytoene. Phytoene molecules undergo tandem steps of modification catalyzed by phytoene desaturase resulting in the formation of lycopene (Fig. 2).
The expression of the heterologous pathway in E. coli Tlyc was confirmed by the formation of red colonies by the harboring cells. Lycopene production in E. coli Tlyc was approximately 140 ± 17 mg/l. This construct was then used for transformation of E. coli Ba and E. coli Br. The resulting strains were designated as E. coli Ba-lyc and E. coli Br-lyc, respectively. The latter strains were cultured for the purpose of CoQ10 production in 2YTG medium under the same conditions, as described previously. Lycopene production in E. coli Ba-lyc did not significantly affect CoQ10 production. However, in the case of E. coli Br-lyc, CoQ10 production was significantly improved from approximately 564 ± 28 to 796 ± 38 μg/g DCW. Induction of the cells by IPTG (0.05 mM) elevated CoQ10 production even further to approximately 989 ± 22 μg/g DCW (Fig. 5). E. coli Ba-lyc concomitantly produced lycopene up to 98 ± 4 mg/l, which was decreased to 62 ± 8 mg/l after induction by 0.05 mM IPTG. Lycopene production in E. coli Br-lyc was quantified as 108 ± 13 mg/l, which was decreased to 59 ± 10 mg/l after induction by 0.05 mM IPTG (Fig. 5). Growth was not significantly affected in both strains under concomitant lycopene production conditions (Fig. 5).

The effect of crtE expression in E. coli Ba and E. coli Br: The expression of crtE results in the formation of GGPP as an intermediate metabolite in the lycopene pathway. GGPP is the essential metabolite in lycopene production as well as the second most suitable substrate after FPP for CoQ10 production (Zahiri et al., 2006b; Takahashi et al., 2003). E. coli contains an ispA gene for FPP biosynthesis but does not have a counterpart for crtE. One may presume that the improvement of CoQ10 production in E. coli Br-lyc is due to GGPP formation as a result of crtE expression and does not have a direct relation with lycopene biosynthesis as a whole pathway.  
To reveal the effect of crtE alone on CoQ10 production, the genes crtI and crtB were removed from pTcrtEIB. The resulting plasmid, pTcrtE, was used for transformation of E. coli Ba and E. coli Br. The resulting strains, E. coli Ba-crtE and E. coli Br-crtE, were cultured for coenzyme Q production under the same conditions, as described previously. Ubiquinones, including CoQ8 and CoQ10, were extracted and quantified by HPLC. The results showed that the expression of crtE was highly inhibiting for both CoQ8 and CoQ10 production in E. coli Ba-crtE and E. coli Br-crtE, even in the absence of the inducer (Fig. 6). However, the inhibition of CoQ10 production was more severe in E. coli Ba-crtE. Growth of both strains was not significantly affected by crtE expression (Fig. 6). It seemed that in E. coli Ba-crtE, respiration was mainly supported by the residual amount of CoQ8 production that is sufficient for supporting growth.
DISCUSSION

The IPP concentration has been suggested as a source of limitation in the biosynthesis of isoprenoids (Sandmann, 2002; Huang et al., 2001). Various metabolic engineering strategies have so far been attempted by researchers to improve isoprenoid production in E. coli. These attempts mainly focus on increasing IPP availability through manipulation of the native MEP pathway or introduction of a heterologous mevalonate pathway (Rodrígues-Villalthe όn et al., 2008; Kim and Keasling, 2003; Martin et al., 2003; Matthews and Wurtzel, 2000; Harker and Bramley, 1999). In the case of CoQ10, manipulation of the MEP pathway through overexpression of DXS has been found to result in a slight increase in CoQ10 production by up to 940 mg/g DCW (Kim, 2006). In contrast, researchers have found that a heterologous mevalonate pathway results in significant, but not yet satisfactory, improvements in CoQ10 production by up to 2400 mg /g DCW (Zahiri et al., 2006b). When compared with amorphadiene production by engineering a heterologous mevalonate pathway in E. coli, the similar strategy has not been as efficient for CoQ10 production as expected (Newman, 2006). This probably is an indication of other bottlenecks than IPP availability in the pathways leading to CoQ10 biosynthesis (Cluis, 2007).
In contrast to IPP biosynthesis, little attention has so far been paid to the consumption rate of IPP in the cell. It is well known that the production and consumption rates of precursors are critical to achieve the maximum productivity of a biosynthetic pathway. Especially that accumulation of precursors may pose new bottlenecks via the inhibition of some enzymes of the pathway.
In this study the effect of concomitant lycopene production was studied on CoQ10 accumulation in E. coli. Isopentenyl diphosphate (IPP) is a common precursor for the biosynthesis of ubiquinones as well as lycopene. Two, three and four IPP molecules may combine together by specific enzymes to form GPP, FPP and GGPP, respectively.  All the four molecules are potential substrates for the addition of more IPP molecules by decaprenyl diphosphate synthase in a step wise manner, in order to make the long 50-carbon chain of CoQ10 (Fig. 1). Experiments have revealed that FPP and GGPP are respectively the preferred substrates of decaprenyl diphosphate synthase, resulting in maximum in vitro reaction rates of the enzyme (Zahiri, 2006a; Lee et al., 2004; Takahashi et al., 2003). On the other hand, GGPP is the essential substrate for the biosynthesis of lycopene. Therefore, ubiquinone and lycopene biosynthesis pathways share the same pool of precursors.
In the current work, a heterologous lycopene production pathway was engineered in two CoQ10-producing E. coli strains. The concomitant lycopene production did not significantly affect CoQ10 production in E. coli Ba-lyc carrying decaprenyl diphosphate synthase of A. tumefaciens. In contrast, CoQ10 production in E. coli Br-lyc harboring decaprenyl diphosphate synthase of R. sphaeroides was remarkably improved as a result of the concomitant production of lycopene. The main difference between these two strains is the biochemical characteristics of decaprenyl diphosphate synthases being used for CoQ10 production. The Dds enzymes of A. tumefaciens and R. sphaeroides have 56% similarity according to amino acid alignments (Zahiri et al., 2006a). The relatively low similarity between the enzymes may reflect in the difference in their biochemical characteristics. In both strains, lycopene production did not harmfully affect growth. It may be concluded that the net result of lycopene biosynthesis is synergistic for CoQ10 production in E. coli Br-lyc.
It is important to confirm that improved CoQ10 production in E. coli Br-lyc is due to lycopene production and not GGPP formation as an intermediate molecule of the lycopene pathway, which is a suitable substrate for CoQ10 production as well.  
     The expression of crtE in CoQ10 producing E. coli strains resulted in significant decreases in both CoQ8 and CoQ10 production. However, growth was not significantly affected by the expression of crtE. As the gene codes for geranylgeranyl diphosphate synthase, it seems that the decrease in ubiquinone biosynthesis occurs as a result of the inhibition of both octa- and deca-prenyl diphosphate synthases by GGPP. This is the first report of the in vivo inhibition of prenyl diphosphate synthases by GGPP. To make sure of such an inhibitory effect of substrates on prenyl diphosphate synthases, crtE in pTcrtE was replaced with ispA encoding for farnesyl diphosphate synthase. The resulting plasmid, pTispA, was used for transformation of E. coli Ba and E. coli Br. The same results as with crtE were achieved with ispA. Ubiquinone biosynthesis was strongly inhibited as a result of ispA expression (data not shown). Here again, growth was not significantly affected.
The results may show that the inhibition of decaprenyl diphosphate synthase by the precursors is a major bottle neck affecting CoQ10 accumulation. The Dds of R. sphaeroides is less sensitive to inhibition by GGPP. Therefore, GGPP at tolerated levels may be used by this enzyme as a substrate for CoQ10 production. R. sphaeroides is a photosynthetic organism that produces high levels of CoQ10 in addition to isoprenoid red pigments. It can be concluded that to take example of natural systems may be an efficient approach along with rational and computational approaches in designing metabolic engineering strategies. The concomitant production of biomaterials in the same cell is important from a biotechnological viewpoint. This strategy may result in higher accumulation of important products, which has applications in the pharmaceutical, food and feed industries.

Alleva R, Tomasetti M, Battino M, Curatola G, Littarru GP, Folkers K (1995). The roles of coenzyme Q10 and vitamin E on the peroxidation of human low density lipoprotein subfractions. Proc Natl Acad Sci. 92: 9388-9391.
Casali N, Preston A (2003). E. coli Plasmid Vectors. Methods in Molecular Biology. Humana Press Inc. 235: 49-53.
Cluis CP, Burja AM, Martin VJJ (2007). Current prospects for the production of coenzyme Q10 in microbes. Trends Biotechnol. 25: 514-521.
Ernster L, Dallner G (1995). Biochemical, physiological and medical aspects of ubiquinone function. Biochim Biophys Acta. 1271: 195-204.
Groneberg DA, Kindermann B, Althammer M, Klapperc M, Vormannd J, Littarrue GP, Döring F (2005). Coenzyme Q10 affects expression of genes involved in cell signalling, metabolism and transport in human CaCo-2 cells. Int J Biochem Cell B. 37: 1208-1218.
Grunler J, Ericsson J, Dallner G (1994). Branch-point reactions in the biosynthesis of cholesterol, dolichol, ubiquinone and prenylated proteins. Biochim Biophys Acta. 1212: 259-277.
Harker M, Bramley PM (1999). Expression of prokaryotic 1-deoxy-D-xylulose-5-phosphatases in Escherichia coli increases carotenoid and ubiquinone biosynthesis. FEBS Lett. 448: 115-119.
Huang QL, Roessner CA, Croteau R, Scotta AI (2001). Engineering Escherichia coli for the Synthesis of Taxadiene, a Key Intermediate in the Biosynthesis of Taxol. Bioorgan Med Chem. 9: 2237-2242.
Hundle B, Alberti M, Nievelstein V, Beyer P, Kleinig H, Armstrong GA, Burke DH, Hearst JE (1994). Functional assignment of Erwinia herbicola Eho10 carotenoid genes expressed in Escherichia coli. Mol Gen Genet. 245: 406-416.
Kawamukai M (2002). Biosynthesis, bioproduction and novel roles of ubiquinone. J Biosci Bioeng. 94: 511-517.
Kim SJ (2006). Amplification of 1-deoxy-D-xyluose 5-phosphate (DXP) synthase level increases coenzyme Q10 production in recombinant Escherichia coli. Appl Microbiol Biotechnol. 72: 982-985.
Kim SW, Keasling JD (2001). Metabolic engineering of nonmevalonate isopentenyl diphosphate synthesis pathway in E. coli enhances lycopene production. Biotechnol Bioeng. 72: 408-415.
Kontush A, Hubner C, Finckh B, Kohlschutter A, and Beisiegel U (1995). Antioxidative activity of ubiquinol-10 at physiologic concentrations in human low density lipoprotein. Biochim Biophys Acta. 1258: 177-187.
Kovach ME, Elzer PH, Hill DS, Robertson GT, Farris MA, Roop RM, Peterson KM (1995). Four new derivatives of the broadhost- range cloning vector pBBR1MCS, carrying different antibioticresistance cassettes. Gene. 166: 175-176.
Lee JK, Her G, Kim SH, Seo JH (2004). Cloning and Functional Expression of the dps Gene Encoding Decaprenyl Diphosphate Synthase from Agrobacterium tumefaciens. Biotechnol Prog. 20: 51-56.
Martin VJJ, Pitera DJ, Withers ST, Newman JD, Keasling JD (2003). Engineering a mevalonate pathway in E. coli for production of terpenoids. Nat Biotechnol. 21: 796-802.
Matthews PD, Wurtzel ET (2000). Metabolic engineering of carotenoid accumulation in E. coli by modulation of the isoprenoid precursor pool with expression of deoxyxylulose phosphate synthase. Appl Microbiol Biotechnol. 53: 396-400.
Meganathan R (1996). Biosynthesis of the isoprenoid quinones menaquinone (vitamin K2) and ubiquinone (coenzyme Q). In: Escherichia coli and Salmonella: cellular and molecular biology, 2nd ed. Neidhardt FC, Curtiss R, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE: ASM Press,Washington DC. PP. 642-656.
Newman JD (2006). High-level production of amorpha-4, 11-diene in a two-phase partitioning bioreactor of metabolically engineered Escherichia coli. Biotechnol Bioeng. 95: 684-691.
Okada K, Kainou T, Matsuda H, Kawamukai M (1998). Biological significance of the side chain length of ubiquinone in Saccharomyces cerevisiae. FEBS Lett. 431: 241-244.
Okada K, Suzuki K, Kamiya Y, Zhu X, Fujisaki S, Nishimura Y, Nishino T, Nakagawa T, Kawamukai M, Matsuda H (1996). Polyprenyl diphosphate synthase essentially defines the length of the side chain of ubiquinone. Biochim Biophys Acta. 1302: 217-223.
Rodrígues-Villalόn A, Pérez-Gil J, Rodríguez-Concepciόn M (2008). Carotenoid accumulation in bacteria with enhanced supply of isoprenoid precursors by upregulation of exogenous or endogenous pathways.  J Biotechnol. 135: 78-84.
Sandmann G (2002). Combinatorial biosynthesis of carotenoids in a heterologous host: a powerful approach for the biosynthesis of novel structures. Chem Bio Chem. 3: 629-635.
Søballe B, Poole RK (2000). Ubiquinone limits oxidative stress in Escherichia coli. Microbiology. 146: 787-796.
Suzuki K, Okada K, Kamiya Y, Zhu XF, Nakagawa T, Kawamukai M, Matsuda H (1997). Analysis of the decaprenyl diphosphate synthase (dps) gene in fission yeast suggests a role of ubiquinone as an antioxidant. J Biochem. 121: 496-505.
Takahashi S, Nishino T, Koyama T (2003). Isolation and expression of Paracoccus denitrificans decaprenyl diphosphate synthase gene for production of ubiquinone-10 in Escherichia coli. Biochem Eng J. 16: 83-190.
Zahiri HS, Noghabi KA, Shin YC (2006a). Biochemical characterization of the decaprenyl diphosphate synthase of Rhodobacter sphaeroides for coenzyme Q10 production. Appl Microbiol Biotechnol. 73: 796-806.
Zahiri HS, Yoon SH, Keasling JD, Lee SH, Kim SW, Yoon SC, Shin YC (2006b). Coenzyme Q10 production in recombinant Escherichia coli strains engineered with a heterologous decaprenyl diphosphate synthase gene and foreign mevalonate pathway. Metab Eng. 8: 406-416.