B

Journal of Biotechnology 144 (2009) 64–69 Contents lists available at ScienceDirect Journal of Biotechnology journa l homepage: www.e lsev ier .com Synerg n coenzy eri Agroba Jin-Ho Ch eo a a Department o nivers Kwanak-gu, Se b Department o a a r t i c l Article history: Received 30 Ja Received in re Accepted 23 A Keywords: Coenzyme Q10 Escherichia col Decaprenyl diphosphate synthase Octaprenly diphosphate synthase 1-Deoxy-d-xylulose synthase Fed-batch fermentation enzy ologo f end ase (D ed pr f CoQ E. col medium with 20 g l−1 initial glucose and by the glucose-feeding strategy of pH-stat. Finally, 99.4 mg l−1 CoQ10 concentration, 1.41 mg g−1 specific CoQ10 content and 3.11 mg l−1 h−1 productivity were obtained in 33 h of the fermentation, which were 78, 1.9, and 19 times higher than those for E. coli BL21(DE3)/pAP1 without the ispB deletion and dxs overexpression. © 2009 Elsevier B.V. All rights reserved. 1. Introduc For deca cal roles of electron in oxidative st 2001). On t a useful sup cial interest diseases an 2007; Pepe in cellular dized CoQ10 NADH cytoc tase (DT-dia cycle (Villa by genetic d glutaryl-Co administrat hence preve ∗ Co-corres ∗∗ Correspon E-mail add 0168-1656/$ – doi:10.1016/j.j tion des, numerous reports have demonstrated the biologi- coenzyme Q10 (CoQ10) in the human body: delivery of the oxidative phosphorylation system, protection from resses, and participation in signal transduction (Crane, he basis of these functions, CoQ10 has been treated as plement in food and medical industries. Recently, spe- s were also paid to the benefits against cardiovascular d neurodegenerative diseases (Galpern and Cudkowicz, et al., 2007). CoQ10 is a lipid-soluble antioxidant located membranes. Different from other antioxidants, oxi- is reduced by membrane-associated enzymes such as hrome b5 reductase and NAD(P)H:quinine oxidoreduc- phorase) in order to maintain the oxidation–reduction lba and Navas, 2000). CoQ10 deficiency can be caused efects, aging, and drugs such as 3-hydroxy-3-methyl- A reductase inhibitors (Lopez et al., 2006). But its oral ion can restore the plasma CoQ10 concentration and nt the CoQ10 deficiency (Bhagavan and Chopra, 2007). ponding author. Tel.: +82 2 880 4889; fax: +82 2 873 5260. ding author. Tel.: +82 2 880 4855; fax: +82 2 873 5095. resses: ycpark@snu.ac.kr (Y.-C. Park), jhseo94@snu.ac.kr (J.-H. Seo). Chemical and biotechnological methods have been developed for the production of CoQ10. Chemical processes need an expen- sive and purified starting material such as solenosol extracted from tobacco (Lipshutz et al., 2002). For biotechnological produc- tion, natural CoQ10 producers including Agrobacterium tumefaciens, Paracoccus denitrificans and Rhodobacter sphaeroides were chemi- cally mutated and selected against inhibitors such as l-ethionine, menadione and daunomycin (Yoshida et al., 1998). Optimization of pH, dissolved oxygen content and sucrose concentration led to the establishment of coenzyme Q10 production process using A. tumefaciens KCCM 10413 (Ha et al., 2007a,b). In addition to chemical mutagenesis, irradiation by a low-energy ion beam with nitrogen ion was applied for strain development (Gu et al., 2006). Escherichia coli, a work horse in metabolic engineering areas, was also engineered by recombinant DNA technology for the syn- thesis of CoQ10 via the introduction of decaprenyl diphosphate synthase (Dps) from various microorganisms including A. tume- faciens and Gluconobacter suboxydans (Zahiri et al., 2006; Park et al., 2005). These recombinant strains were further modified by overexpression of several upstream enzymes in order to increase the metabolic flux toward CoQ10. Overexpression of 1-deoxy- d-xylulose 5-phosphate synthase (Dxs) in the non-mevalonate pathway showed positive effects on CoQ10 production (Kim et al., 2006; Seo et al., 2007). Coexpression of the UbiA and UbiC enzymes involved in the synthesis of a phenolic ring in CoQ10 also increased see front matter © 2009 Elsevier B.V. All rights reserved. biotec.2009.04.010 istic effects of chromosomal ispB deletio me Q10 production in recombinant Esch cterium tumefaciens dps gene oi a, Yeon-Woo Ryu b, Yong-Cheol Park a,∗, Jin-Ho S f Agricultural Biotechnology and Center for Agricultural Biomaterials, Seoul National U oul 151-921, Republic of Korea f Molecular Science and Technology, Ajou University, Suwon 442-749, Republic of Kore e i n f o nuary 2009 vised form 14 April 2009 pril 2009 i a b s t r a c t For biotechnological production of co manipulations were performed: heter Agrobacterium tumefaciens, deletion o pression of 1-deoxy-d-xylulose synth in E. coli BL21(DE3)�ispB/pAP1 allow gene increased the specific content o of CoQ10, fed-batch fermentation of / locate / jb io tec and dxs overexpression on chia coli expressing ,∗∗ ity, San 56-1, Shillim-dong, me Q10 (CoQ10) in recombinant Escherichia coli, three genetic us expression of decaprenyl diphosphate synthase (Dps) from ogenous octaprenyl diphosphate synthase (IspB), and overex- xs). Expression of the dps gene and deletion of the ispB gene oduction of CoQ10 only. Furthermore, coexpression of the dxs 10 from 0.55–0.89 mg g−1 to 1.40 mg g−1. For mass production i BL21(DE3)�ispB/pAP1 + pDXS was carried out in a defined shiwang Underline shiwang Underline shiwang Underline J.-H. Choi et al. / Journal of Biotechnology 144 (2009) 64–69 65 the specific content of CoQ10 (Cluis et al., 2007; Zhang et al., 2007). In our previous report, CoQ10 production was demonstrated in recombinant E. coli harboring the ddsA gene originated from G. sub- oxydans and more enhancement of CoQ10 content was also achieved by overexpression of the dxs gene from Pseudomonas aeruginosa (Park et al., 2005; Kim et al., 2006). Concomitant production of other CoQs such as CoQ8 and CoQ9, however, reduced CoQ10 pro- ductivity significantly. In order to construct recombinant E. coli systems able to synthesize CoQ10 only, we deleted the endogenous octaprenyl diphosphate synthase gene (ispB) in the chromosome of E. coli and expressed the decaprenyl diphosphate synthase gene (dps) from A. tumefaciens in recombinant E. coli deficient in the ispB gene. Furthermore, the effects of the ispB gene knock-out and over- expression of the dxs gene coding for E. coli 1-deoxy-d-xylulose 5-phosphate synthase were demonstrated in flask and fed-batch fermentations. 2. Materials and methods 2.1. Bacterial strains and plasmids E. coli DH5� and TOP10 strains were used for genetic manip- ulation and E. coli BL21(DE3) was used for CoQ10 production. Plasmid pBlu-dps containing the A. tumefaciens dps gene was constructed previously (Lee et al., 2004). Expression of the dps gene was controlled under a constitutive promoter in plasmid pUCmodII (Park et al., 2005). Plasmids pACYC184 and pACYCDuet- 1 with low copy number (10–12) were purchased from New England Biolabs (Beverly, MA, USA) and Novagen (Darmstadt, Germany), respectively. E. coli strains harboring the plasmids used for ge by the E. All plasmid Table 1. 2.2. Construction of dps gene expression system PCR amplification of a 1.2 kb DNA fragment including the 100 base pairs of the upstream region of the dps gene was performed by using pBlu-dps plasmid and two primers of 5′-dps and 3′-dps which contained two recognition sites of XbaI and NdeI restric- tion enzymes, respectively. The PCR product was digested and ligated into pUCmodII plasmid digested by the same enzymes. This new plasmid was named pAP1 containing the constitutive promoter and the replicon derived from the pUC family (copy number 500–700). And then, a 1.3 kb PCR fragment consisting of the constitutive promoter and dps gene was amplified by using pAP1 plasmid as a template and two DNA oligomers of 5′-subdps and 3′-subdps, which contained two recognition sites of BamHI and AvaI restriction enzymes, respectively. The PCR product was cloned into plasmid pACYC184 and hence, plasmid pJC105 harboring the dps gene combined with the constitutive promoter and P15A ori was constructed. The expression of the dps gene in recombinant E. coli was controlled under the consti- tutive promoter. All PCR primers used in this study are listed in Table 2. 2.3. Construction of dxs expression plasmid The dxs gene from P. aeruginosa was amplified by using the two primers of 5′-dxs and 3′-dxs, which was cloned into pACYCDuet-1 after their digestion with NcoI and EcoRI. The constructed plasmid, pDXS was designed to express the dxs gene under the IPTG- inducible T7 promoter. 2.4. Deletion of chromosomal ispB gene in E. coli BL21(DE3) dele anne ne in E3) Table 1 List of plasmid Plasmid pUCmodII expres pACYC184 pACYCDuet-1 pJC105 5A ori pAP1 C ori pDXS pKD46 pKD4 pCP20 eplico AAGCA TGAC ATGT ATGT GA GCTG CGCG CGAG ion en ecomb ne deletion (pKD46, pKD4 and pCP20) were provided coli Genetic Resource Center (Yale University, USA). s used and constructed in this study are listed in The and W dps ge BL21(D s used in this study. Characteristic AmpR, high copy number vector, modified Plac constitutive ChlR, TetR, low copy number vector, P15A ori ChlR, low copy number vector, Dual T7 promoter, P15A ori pACYC184 derivative, constitutive dps expression vector, P1 pUCmodII derivative, constitutive dps expression vector, pU pACYCDuet-1 derivative, dxs expression by T7 promoter Phage � red recombinase, temperature sensitive replicon AmpR, KanR, oriR6K� AmpR, ChlR, yeast Flp recombinase, temperature sensitive r Table 2 List of DNA oligomers used in this study. Name Sequence 5′-dpsa 5′-AACAACTAGTCTAGACAGCGAAGGACAG 3′-dpsa 5′-AATCCGCATATGTCAGTTGAGACGCTCGA 5′-subdpsa 5′-AATAATCGCGGATCCGCGCAACGCAATTA 3′-subdpsa 5′-AATAATCGCGGATCCGCGCAACGCAATTA 5′-dxsa 5′-AATAATCCATGGTGGCCAAGACGCTCCAT 3′-dxsa 5′-AACAACGAATTCCTACTGCCGGTCGAGAC 5′-�ispBb 5′-ATGAATTTAGAAAAAATCAATGAGTTAAC 3′-�ispBb 5′-TTAACGATCGCGTTGAACAGCGATGTGCG 5′-F-check 5′-TGCCATTTTTTCAGTACAATCACCC 3′-F-check 5′-GCGGTCCGCCACACCCAGCC 5′-R-check 5′-CGGTGCCCTGAATGAACTGC 3′-R-check 5′-GATGCACATCCCTATTTTTCAGGTG a The italicized sequences indicate the recognition sites of the specific restrict b The underlined nucleotides present the gene fragment for the homologous r tion process followed the previous report (Datsenko r, 2000). Because of the lethality of the ispB gene, the plasmid pJC105 or pAP1 was introduced into E. coli before the deletion of the ispB gene. For the expres- Source sion promoter Park et al. (2005) New England Biolabs Novagen This study This study This study Datsenko and Wanner (2000) Datsenko and Wanner (2000) n Datsenko and Wanner (2000) CTG GAG GAG CAAGATATGGCGGGTGTGTGTAGGCTGGAGCTGCTTC GCCGATGAGTGCTTCTCATGGGAATTAGCCATGGTCC zyme. ination of the chromosomal ispB gene. shiwang Highlight shiwang Highlight shiwang Highlight shiwang Highlight shiwang Highlight 66 J.-H. Choi et al. / Journal of Biotechnology 144 (2009) 64–69 sion of � red recombinase, plasmid pKD46 was purified from E. coli BW25113/pKD46 and introduced to E. coli BL21(DE3). Plasmid pKD4 containing the kanamycin resistance gene was used as a template for the construction of the ispB deletion cassette. Two PCR primer 50 nucleoti somal ispB kanamycin (Table 1). A the ispB ge into E. coli homologou E. coli BL21 containing tion of the aid of a he gene. By th 37 ◦C overn and the he tion were of 5′-F-che For the con procedure selection of ampicillin. 2.5. Culture For DNA (5 g l−1 yeas 50 mg l−1 k 500 ml baffl NY, USA). B Fed-batch c previously ( cose in the 800 g l−1 gl by a pH-sta acidity of cu temperatur tained thro 0.1 mM IPTG 15 h of fed- 2.6. Analysi concentratio Optical d spec 2000, which was c pre-determ and acetic a mance liqu system, Agi an RI detect (Phenomen of H2SO4 w According t the cells w Korea) com seido Co., T Absolute et at 1 ml/min CoQ10 was tent of CoQ cell mass. PLC a 3) stra ults fects the e onta and into h of fl ble t strat 21(D 10 es higher than that for E. coli BL21(DE3)/pJC105 (12(±2)%). ild type E. coli produced CoQ8 as expected. However, the nous CoQ8 still remained as a major byproduct in the recom- E. coli systems. nstruction of recombinant E. coli deficient in chromosomal ne produce CoQ10 without CoQ8 and CoQ9 accumulation, the nous ispB gene located in the E. coli chromosome was and the dps gene expression systems were introduced. n of the chromosomal ispB gene without its complemen- is impossible because of the lethality of this gene as reported (Okada et al., 1997). Intracellular exonuclease activities of should be inhibited for the homologous recombination of get gene. Therefore, two plasmids were inserted to the E. 21(DE3) host strain: pJC105 containing the dps gene and expressing the � red recombinase. After introduction of B gene deletion cassette and homologous recombination, nsformants survived against triple antibiotics (ampicillin, phenicol and kanamycin) were selected and their chromo- DNA was subjected to PCR for verification of the endogenous letion. Analysis of the PCR products on agarose gels indi- the correct deletion of the ispB gene and proper location kanamycin resistance gene (Fig. 2a). After removal of the s, 5′-�ispB and 3′-�ispB, were designed to contain des of the 5′- and 3′-end sequences of the chromo- gene and to possess the recognition sites for the resistance gene according to the previous procedure 0.9 kb PCR fragment containing the two regions of ne and kanamycin resistance gene was introduced BL21(DE3)/pJC105 + pKD46 by electroporation. After s recombination, pKD46 plasmid was cured and then (DE3)�ispB/pJC105 was selected on LB agar medium 50 mg l−1 kanamycin and chloramphenicol. Elimina- kanamycin resistance gene was performed by the lper plasmid, pCP20 containing the FLP recombinase e incubation of the cells containing plasmid pCP20 at ight, the kanamycin resistance gene was eliminated lper plasmid was cured. Gene insertion and dele- verified by the PCR method using four PCR primers ck, 3′-F-check, 5′-R-check and 3′-R-check (Table 1). struction of E. coli BL21(DE3)�ispB/pAP1, the same for BL21(DE3)�ispB/pAP1 was used except for the the recombinant cell using 50 mg l−1 kanamycin and conditions manipulation and flask culture, 100 ml of LB medium t extract, 10 g l−1 bacto-tryptone and 10 g l−1 NaCl) with anamycin or 34 mg l−1 chloramphenicol was used in a ed flask (Nalgene 4110-0500, Nalge Nunc Int., Rochester, atch flask culture was performed at 250 rpm and 37 ◦C. ulture was accomplished in a defined medium reported Park et al., 2005). After depletion of 20 g l−1 initial glu- fed-batch fermentation, a feeding solution containing ucose and 20 g l−1 MgSO4·7H2O was fed intermittently t operation mode, which was designed to control the lture broth automatically (Ha et al., 2007b). A culture e of 37 ◦C, pH 6.8 and 1 vvm of air flow rate were main- ughout the cultivation. When expressing the dxs gene, induction was performed at 2 h of flask cultivation and batch cultivation. s of dry cell mass, carbohydrate and CoQ10 n ensity was measured with a spectrophotometer (Ultra- Pharmacia Biotech, Piscataway, NJ, USA) at 600 nm, onverted into dry cell mass by the multiplication of the ined conversion factor, 0.3 (Son et al., 2007). Glucose cid concentrations were determined by a high perfor- id chromatography system (Agilent 1100 series HPLC lent Technologies, Waldbronn, Germany) equipped with or (G1362A) and the RezexTM ROA Organic Acid column ex, Torrance, CA, USA). Five millimolar concentration as flowed into the column at 0.6 ml/min flow rate. o the previous report (Park et al., 2005), CoQ10 inside as analyzed by an HPLC system (Younglin Co., Seoul, plemented with a CAPCELL PAK C18 MG column (Shi- okyo, Japan) and a UV detector (Younglin Co., Korea). hanol was used as a solvent and its flow rate was fixed . The absorbance values were measured at 275 nm. analyzed independently in triplicate. The specific con- 10 was calculated by dividing its concentration by dry Fig. 1. H BL21(DE 3. Res 3.1. Ef For mids c pJC105 formed after 8 were a demon coli BL 5.8 tim The w endoge binant 3.2. Co ispB ge To endoge deleted Deletio tation before E. coli the tar coli BL pKD46 the isp the tra chloram somal ispB de cated of the nalysis of CoQ10 inside the wild type (a) and recombinant E. coli ins containing plasmid pJC105 (b) and pAP1 (c). of dps expression on coenzyme Q10 population xpression of the dps gene from A. tumefaciens, two plas- ining different origins of replication such as P15A for pUC ori for pAP1 were constructed and were trans- E. coli BL21(DE3). As shown in HPLC analysis of CoQ10 ask cultures, two recombinant E. coli BL21(DE3) strains o synthesize CoQ10 as well as CoQ8 and CoQ9, which ed the functional expression of the dps gene (Fig. 1). E. E3)/pAP1 showed 69(±3)% of CoQ content which was shiwang Underline shiwang Underline shiwang Underline shiwang Underline shiwang Underline shiwang Highlight J.-H. Choi et al. / Journal of Biotechnology 144 (2009) 64–69 67 Fig. 2. Scheme of deletion of the E. coli chromosomal ispB gene (a and b) and their confirmation using agarose gel electrophoresis (c). FRT and Kan indicate the FLP recognition target and kanamycin resistance gene, respectively. PCR conditions of template and primers were presented as follows: lane C, wild type strain; lane F, �ispB::kan strain, 5′-F-check and 3′-F-check; lane R, �ispB::kan strain, 5′-R-check and 3′-R-check; lane D, �ispB strain, 5′-F-check and 3′-R-check. PCR primers located under or above the arrows are informed in Table 2. marker gene and curing plasmid pCP20, the ispB gene-deficient strain, E. coli BL21(DE3)�ispB/pJC105 was constructed, which was confirmed by PCR. A 0.8 PCR product on the agarose gel represented the correct deletion of the ispB gene deletion on the E. coli chromo- some (Fig. to construc production nant E. coli d As shown in only and th 3.3. Batch f Batch fla investigate plasmid sp Fig. 3. HPLC a cient in the ch (b). four strains, the crude extracts of the cells were subjected to CoQ10 analysis an culture of B tration and high alue trati binan ene ( ne re creas 1(D E3)� d enh with furth as co sitiv ready 2b). Instead of plasmid pJC105, pAP1 was introduced t E. coli BL21(DE3)�ispB/pAP1. To confirm the exclusive of CoQ10, simple flask fermentations of two recombi- eficient in the ispB gene were carried out in LB medium. Fig. 3, the ispB gene deletion mutants produced CoQ10 e other CoQ species were not detected. ermentations sk fermentations using LB medium were carried out to the quantitative effects of the ispB gene deletion and ecies on CoQ10 production. After 8 h of culture of the times these v concen recom ddsA g ispB ge also in coli BL2 BL21(D 1.2-fol strains For gene w The po was al nalysis of CoQ10 inside the recombinant E. coli BL21(DE3) strains defi- romosomal ispB gene and containing plasmid pJC105 (a) and pAP1 Fig. 4. Batch baffled flask w CoQ10 was an black one indi d their results were displayed in Figs. 4 and 5. The flask L21(DE3)/pAP1 resulted in 1.28 mg l−1 CoQ10 concen- 0.75 mg g−1 its specific content, which were 6.1 and 6.3 er than those of BL21(DE3)/pJC105 (Fig. 4). Moreover, s corresponded to 1.3- and 1.9-fold increases in CoQ10 on and its specific content compared with those for the t E. coli BL21(DE3) strain expressing the G. suboxydans the pACDdsA system) (Park et al., 2005). Deletion of the sulted in not only changing the main CoQ species but ing CoQ10 production. The specific content of CoQ10 in E. E3)�ispB/pJC105 was obtained at 0.55 mg g−1 and E. coli ispB/pAP1 showed 0.98 mg g−1, which were 4.6- and ancement relative to those for the same recombinant out the ispB deletion (Fig. 5). er increment of CoQ10 production, the P. aeruginosa dxs expressed in recombinant E. coli BL21(DE3)�ispB/pAP1. e effects of Dxs coexpression on CoQ10 production known to be caused by the enhancement of isopen- fermentations of the recombinant E. coli BL21(DE3) strains