Manipulation of prenyl chain length determination mechanism of cis-prenyltransferases Yugesh Kharel*, Seiji Takahashi*, Satoshi Yamashita and Tanetoshi Koyama Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai, Japan In the biosynthesis of isoprenoids, which are the most structurally diverse and abundant among natural prod- ucts, all carbon skeletons are biosynthesized using sequential condensation of isopentenyl diphosphate (IPP, C 5 ) and allylic diphosphates by actions of prenyl chain elongating enzymes, commonly called prenyl- transferases. Prenyltransferases can be classified into two major groups, i.e. trans- and cis-prenyltransferases, according to the geometry of the prenyl chain units in the products [1,2]. The reactions catalysed by prenyl- transferases start by the formation of allylic cation after the elimination of pyrophosphate ion to form an allylic prenyl diphosphate, followed by addition of an IPP with stereospecific removal of a proton at the 2-position. The only difference between the reaction catalyzed by trans- and cis-prenyltransferase is the prochirality of the proton, i.e. pro-R for trans- prenyltransferase and pro-S for cis-prenyltransferase (Fig. 1a). However, recent molecular analysis and crys- tal structure determination of Micrococcus luteus B-P 26 undecaprenyl diphosphate (UPP, C 55 ) synthase, which catalyses cis-condensation to synthesize UPP, Keywords chain-length determination mechanism; cis- prenyltransferase; isoprenoid biosynthesis; Micrococcus luteus B-P 26; undecaprenyl diphosphate synthase Correspondence T. Koyama, Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Katahira 2-1-1, Aoba-ku, Sendai, 980-8577, Japan Fax: +81 22 217 5620 Tel:: +81 22 217 5621 E.mail: koyama@tagen.tohoku.ac.jp Note *Both authors contributed equally to this work. (Received 28 October 2005, revised 9 December 2005, accepted 12 December 2005) doi:10.1111/j.1742-4658.2005.05097.x The carbon backbones of Z,E-mixed isoprenoids are synthesized by sequential cis-condensation of isopentenyl diphosphate (IPP) and an allylic diphosphate through actions of a series of enzymes called cis-prenyltrans- ferases. Recent molecular analyses of Micrococcus luteus B-P 26 undecapre- nyl diphosphate (UPP, C 55 ) synthase [Fujihashi M, Zhang Y-W, Higuchi Y, Li X-Y, Koyama T & Miki K (2001) Proc Natl Acad Sci USA 98, 4337–4342.] showed that not only the primary structure but also the crystal structure of cis-prenyltransferases were totally different from those of trans-prenyltransferases. Although many studies on structure–function rela- tionships of cis-prenyltransferases have been reported, regulation mecha- nisms for the ultimate prenyl chain length have not yet been elucidated. We report here that the ultimate chain length of prenyl products can be controlled through structural manipulation of UPP synthase of M. luteus B-P 26, based on comparisons between structures of various cis-prenyl- transferases. Replacements of Ala72, Phe73, and Trp78, which are located in the proximity of the substrate binding site, with Leu ) as in Z,E-farnesyl diphosphate (C 15 ) synthase ) resulted in shorter ultimate products with C 20)35 . Additional mutation of F223H resulted in even shorter products. On the other hand, insertion of charged residues originating from long- chain cis-prenyltransferases into helix-3, which participates in constitution of the large hydrophobic cleft, resulted in lengthening of the ultimate prod- uct chain length, leading to C 60)75 . These results helped us understand reaction mechanisms of cis-prenyltransferase including regulation of the ultimate prenyl chain-length. Abbreviations DedolPP, dehydrodolichyl diphosphate; E,E-FPP, E,E-farnesyl diphosphate; FARM, first aspartate-rich motif; GPP, geranyl diphosphate; IPP, isopentenyl diphosphate; UPP, undecaprenyl diphosphate; Z,E-FPP, Z,E-farnesyl diphosphate; Z,E-DecPP, Z,E-mixed decaprenyl diphosphate. FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS 647 showed that not only the primary but also the three- dimensional structure of cis-prenyltransferase were totally different from those of trans-prenyltransferases [3-5]. Homologous genes for cis-prenyltransferases have been identified in various organisms [6–13], revealing five highly conserved regions in the primary structure of all cis-prenyltransferases [3]. On the other hand, cis-prenyltransferases are classified into three subfamilies with respect to product chain length, i.e. short-chain (C 15 ), medium-chain (C 50)55 ), and long- chain (C 70)120 ) cis-prenyltransferases (Fig. 1b). Z,E- farnesyl diphosphate (Z,E-FPP, C 15 ) synthase from Mycobacterium tuberculosis, Rv1086, is the only enzyme identified as a short-chain cis-prenyltrans- ferase, which catalyses cis-condensation of one IPP with geranyl diphosphate (GPP, C 10 ) [11]. Medium- chain cis-prenyltransferases are represented by UPP synthase which catalyses cis-condensation of eight IPPs with E,E-FPP (C 15 ). This enzyme is responsible for biogenesis of undecaprenyl phosphate, an indispens- able glycosyl carrier lipid in bacterial cell wall biosyn- thesis. Z,E-mixed decaprenyl diphosphate (Z,E-DecPP, C 50 ) synthase from M. tuberculosis, Rv2361, is also categorized in this subfamily [11]. Most of dehydro- dolichyl diphosphate (DedolPP) synthases in eukaryo- tes catalysing synthesis of precursors of sugar carrier lipid dolichol during biosynthesis of glycoproteins, are categorized as long-chain cis-prenyltransferases. Srt1p and Rer2p from Saccharomyces cerevisiae [7,8] and HDS from human [12] are included in this sub- family. One of the most interesting research topics on cata- lytic mechanisms of prenyl chain elongating enzymes is to understand mechanisms by which individual pre- nyltransferases recognize prenyl chain lengths of allylic substrates and products. Using a series of effective ran- dom chemical mutagenesis, Ohnuma et al. showed that the ultimate product chain length of trans-prenyltrans- ferases was regulated by bulky residues located upstream of the first Asp-rich motif (FARM), which is one of the most highly conserved regions among trans- prenyltransferases [2,14,15]. The bulky residues func- tion as a floor of the catalytic pocket for the growing isoprenoid chain to block further elongation. Based on crystal structure and mutagenetic studies of the avian FPP synthase, Tarshis et al. concluded that allylic A B Fig. 1. Classification of prenyltransferases. (A) Schematic drawing of the reactions catalysed by trans-prenyltransferase and cis-prenyltrans- ferase. (B) Classification of cis-prenyltransferases with respect to product chain lengths. These enzymes catalyse cis-condensation of IPP (isoprene unit, C 5 ) onto an allylic diphosphate, all-E-prenyl diphosphate. Grey lines indicate numbers of isoprene units of representative allylic substrates for each cis-prenyltransferase. Red arrows indicate numbers of isoprene units of representative ultimate products, which are condensed with cis-configuration by each enzyme. Product chain length of cis-prenyltransferases Y. Kharel et al. 648 FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS diphosphate bound through Mg 2+ to Asp residues in the FARM motif [16]. In contrast, mechanisms for determination of the ultimate product chain length of cis-prenyltransferases have not yet been determined although mutational analysis of highly conserved resi- dues and determination of crystal structures of UPP synthase have enabled us to understand basic catalytic mechanisms of cis-prenyltransferases [5,17–24]. In order to investigate regions important for determination of product chain length in cis-prenyl- transferases, we searched characteristic residues of short- or long-chain cis-prenyltransferase subfamilies, based on comparisons between primary structures of cis-prenyltransferases identified, and crystal structures of UPP synthases [5]. Introduction of mutations in regions of M. luteus B-P 26 UPP synthase, which correspond to characteristic residues, resulted in dras- tic alteration of the ultimate product chain length of prenyl products. Amino acid residues located in close proximity to the substrate binding site play an essen- tial role in the synthesis of short-chain products such as Z,E-FPP, while some charged residues on the side-wall of the large hydrophobic cleft are important to determine the ultimate chain length of polyprenyl products. These findings enabled us to understand reaction mechanisms of cis-prenyltransferases, and manipulate enzymes to produce various lengths of Z,E-mixed polyisoprenoids. Results In order to identify characteristic residues of short- or long-chain cis-prenyltransferase subfamilies, primary structures of proteins that were identified as cis-prenyl- transferases were compared. Because Rv1086 from M. tuberculosis is the only enzyme cloned and identified as a short-chain cis-prenyltransferase, Z,E- FPP synthase [11], we attempted to identify Rv1086- specific residues that will help name the important amino acid residues for ultimate chain-length determin- ation. As shown in Fig. 2A, only Rv1086 possesses three Leu residues at positions 84, 85, and 90 instead of the corresponding Ala, Phe, and Leu ⁄ Trp in the conserved region III of other cis-prenyltransferases. Fig. 2. Amino acid residues characteristic for short-chain cis-prenyl- transferases functioning in the termination mechanism of cis-prenyl chain elongation to produce short-chain Z,E-mixed prenyl diphos- phates. (A) Multiple alignment of amino acid sequences of seven cis-prenyltransferases: Rv1086, Z,E-FPP synthase from Mycobacterium tuberculosis (Accession No.; D70895); Rv2361c, DecPP synthase from M. tuberculosis (H70585); M. luteus, UPP synthase from M. luteus B-P 26 (BAA31993); E. coli, UPP synthase from E. coli (Q47675); Rer2p, DedolPP synthase from Saccharo- myces cerevisiae (BAA36577); Srt1p, polyprenyl PP synthase from S. cerevisiae (NP_013819); HDS, DedolPP synthases from human (BAC57588). Here, parts of sequences in conserved regions III and V are shown. Multiple alignment was obtained using the CLUSTAL W program, and was edited using SeqVu. Residues with more than 70% identity are boxed, and all residues are coloured as follows: non- polar (G, A, V, L, I, P, F, M, W, C), yellow; uncharged polar (N, Q, S, T, Y), green; acidic (D, E), red; basic (K, R, H), blue. Red triangles indicate residues that are characteristic for short-chain cis-prenyl- transferase. (B) TLC autoradiograms of polyprenyl alcohols derived from products of cis-prenyl chain elongation with wild-type and mutant UPP synthases, LL, LLL, F223H, and LLLH, using E,E-FPP (left panel) of GPP (right panel) as allylic substrates. Spots of authentic standard prenyl alcohols are as follows: C 15 , E,E-farnesol; C 20 , all-E-geranylgeraniol; C 55 , undecaprenol. Ori; origin, S.F.; solvent front. Y. Kharel et al. Product chain length of cis-prenyltransferases FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS 649 Furthermore, His-237 of Rv1086 downstream of the conserved region V was found at the corresponding position for Leu ⁄ Phe residue in other cis-prenyltrans- ferases. To investigate whether these Rv1086-characteristic residues function in the reaction mechanism, we replaced the corresponding residues of M. luteus B-P 26 UPP synthase with these Rv1086-specific residues to construct mutant enzymes with single-, double-, triple-, and quadruple mutations, i.e. F223H, A72L ⁄ F73L (LL), A72L ⁄ F73L ⁄ W78L (LLL), and A72L ⁄ F73L ⁄ W78L ⁄ F223H (LLLH), respectively. These mutants were expressed in Escherichia coli, and puri- fied to analyse prenyltransferase activity in vitro. Product distribution patterns of the mutants were com- pared with those of the wild-type UPP synthase, which produces C 55 and C 60 prenyl products in vitro. The TLC autoradiogram clearly showed that LL produced shorter polyisoprenoids (C 25)40 as major products), and LLL produced even shorter polyisoprenoids (C 20)35 as major products) than wild-type UPP syn- thase when E,E-FPP and GPP were used as allylic sub- strates (Fig. 2B). In our previous study, site-directed mutagenesis at the highly conserved Phe73 in region III of M. luteus B-P 26 UPP synthase resulted in a 32-fold increased K m value for IPP, and a 16-fold decreased k cat value [20]. These results indicated that Phe73 in region III was important for binding of IPP in the proper direc- tion. We also observed the shortening of the ultimate chain length of prenyl products in mutants of M. luteus B-P 26 UPP synthase, i.e. F73A and S74A [20]. To analyse functions of Ala72 and Trp78, which are located in the vicinity of Phe73 in region III, kin- etic constants of the mutants were determined. A72L ⁄ F73L showed an 88-fold higher K m value for IPP compared with the wild-type enzyme. However, A72L ⁄ F73L ⁄ W78L showed only a six-fold increased K m value for IPP (Table 1), indicating that shortening of the ultimate chain length of products by the triple mutation, A72L ⁄ F73L ⁄ W78L was not simply caused by a decrease in the affinity of the catalytic domain for IPP. In addition, effects of mutations at Ala72, Phe73, and Trp78 on K m values for E,E-FPP were not signifi- cant (within threefold). On the other hand, replacement of Phe223 with His in mutants F223H or LLLH, dramatically decreased catalytic activity when E,E-FPP was used as allylic substrate (Fig. 2B, left panel). However, when GPP was used, mutants showed certain cata- lytic activity producing C 45-55 and C 15 (Fig. 2B, right panel). These results suggested that His237 of Rv1086, corresponding to Phe223 of M. luteus B-P 26 UPP synthase, might function to prefer GPP as an allylic substrate. We next investigated the characteristic residues of long-chain cis-prenyltransferase subfamilies to identify the important region for producing polyisoprenoid chains longer than C 55 . Multiple alignment of cis-pre- nyltransferases revealed that long-chain cis-prenyl- transferases such as Srt1p and Rer2p from S. cerevisiae [7,8] and HDS from human [12], have three to seven extra amino acid residues downstream of the conserved region III (Fig. 3A). This position corresponds to helix-3 of the M. luteus B-P 26 UPP synthase, which participates in the constitution of the hydrophobic cleft. The hydrophobic cleft is composed of helix-2, helix-3, sheet-2, and sheet-4, and is consid- ered to accommodate the elongated prenyl intermedi- ates [5]. In order to indicate the significance of these residues, we constructed two mutants of M. luteus B-P 26 UPP synthase, i.e. EKE and RAKDY, which contained insertions corresponding exactly to the extra amino acid residues of DedolPP synthases from human (HDS, positions 107–109) and yeast (Srt1p, positions 148–152), respectively (Fig. 3A). Products from prenyltransferase reaction with these UPP syn- thase mutants in vitro were analysed by reversed- phase TLC. As shown in Fig. 3B, EKE produced relatively longer prenyl products with carbon chain lengths of C 55)70 . Furthermore, RAKDY gave even longer prenyl products with chain lengths of C 60)75 . In contrast, the control mutant with an insertion of five Ala residues instead of extra residues did not show such effects (Fig. 3B). Introduction of extra amino acid residues caused a several times increase of K m values for IPP (Table 1). In contrast, mutants EKE and RAKDY showed a 1.4–2.8-fold lower K m values for FPP, suggesting that the change in struc- ture of helix-3 moderately affected binding affinity for FPP. To confirm the significance of the extra resi- dues found in long-chain cis-prenyltransferases, we Table 1. Kinetic constants for wild-type and mutant UPP synthases of M. luteus B-P 26. Enzymes K m (IPP) (lM) a K m (FPP) (lM) Wild-type 7.8 ± 2.8 8.3 ± 1.3 F73A b 252 ± 62 9.3 ± 3.3 A72L ⁄ F73L (LL) 690 ± 242 26 ± 6 A72L ⁄ F73L ⁄ W78L (LLL) 49 ± 25 14 ± 2 EKE 39 ± 10 6 ± 2 RAKDY 16 ± 4 3 ± 0.3 a For reactions with FPP. b Previous data from [3]. Product chain length of cis-prenyltransferases Y. Kharel et al. 650 FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS constructed a deletion mutant of Srt1p lacking the five extra residues, Srt1p-delta (Fig. 3A). Because Srt1p expressed in E. coli does not show significant prenyltransferase activity, wild-type Srt1p and Srt1p- delta were expressed in the yeast mutant strain SNH23-7D which is deficient in the DedolPP synthase gene, RER2, to analyse the reaction products clearly. Product analysis of the prenyltransferase reaction with yeast crude proteins indicated that the mutant Srt1p-delta produced shorter prenyl products with a chain length of C 65)95 as major products (Fig. 3C), while wild-type Srt1p produced C 75)110 prenyl prod- ucts as reported previously [8]. Taken together, the region in helix-3 is also very important for the ulti- mate chain length determination mechanism for cis- prenyltransferases. Discussion In nature, a variety of Z,E-mixed polyisoprenoids with different carbon chain lengths from Z,E-FPP (C 15 )to natural rubber (C >10,000 ) are produced, and distribu- tion of carbon chain lengths of Z,E-mixed polyisopre- noids seems to depend on the origin of organisms. Most bacteria produce only UPP (C 55 ), while eukaryo- tes such as yeasts and mammals produce longer Z,E-mixed polyisoprenoids including DedolPP (C 70)100 ). In higher plants, various Z,E-mixed polyisoprenoids including dolichol and polyprenol are biosynthesized with two different distribution patterns of carbon chain length, encompassing C 50)60 and C 70)120 . More- over, rubber producing plants such as Hevea brasilien- sis produce high molecular weight cis-1,4-polyisoprene, Fig. 3. Extra amino acid residues character- istic for long-chain cis-prenyltransferases are important in producing long-chain Z,E-mixed prenyl diphosphates. (A) Multiple alignment of amino acid sequences of seven cis-pre- nyltransferases. In this figure, a part of the sequence of helix-3 is shown. Similar nota- tions as in Fig. 2 are used for sequences and colour usage. The red upper line indi- cates the extra residues that are characteris- tic for long-chain cis-prenyltransferase. Some parts of sequences of the mutants of M. luteus B-P 26 UPP synthase (EKE, RAKDY) and of Srt1p (Srt1p-delta) are shown. (B) TLC autoradiograms of polypre- nyl alcohols corresponding to products of cis-prenyl chain elongation with wild-type and mutant UPP synthases, EKE, RAKDY, and AAAAA, using E,E-FPP as allylic sub- strate. C 55 indicates the spot of authentic undecaprenol. Ori, origin; S.F., solvent front. (C) TLC autoradiograms of polyprenyl alcohols corresponding to the products of cis-prenyl chain elongation with wild-type and mutants Srt1p, Srt1p-delta, using E,E-FPP as substrate. Spots of authentic standard prenyl alcohols are as follows: C 15 , E,E-farnesol; C 55 , undecaprenol; C 75 , C 75 -polyprenol; C 100 ,C 100 -polyprenol. Ori, origin; S.F., solvent front. Y. Kharel et al. Product chain length of cis-prenyltransferases FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS 651 i.e. natural rubber. Several lines of evidence suggest that product chain lengths of cis-prenyltransferases can be altered by environmental factors. Yamada et al. [25,26] reported differences in chain lengths of dolichol among tissues in the rat such as liver (C 90)95 ) and tes- tis (C 85)90 ). Matsuoka et al. [27] reported that chain length distribution of products from DedolPP synthase in microsomal fractions of rat liver could be affected by reaction conditions in vitro such as detergents and phospholipids. However, recent reports on cloning and characterization of cis-prenyltransferases from various organisms showed that specificities of product chain lengths mainly depended on different enzymatic prop- erties attributable to structural diversity of each enzyme. Based on these facts, we investigated charac- teristic residues of short- or long-chain cis-prenyl- transferase subfamilies to elucidate chain length determination mechanisms. Leu residues at positions 84, 85, and 90 in the Z,E-FPP synthase Rv1086 were found to play a very important role in shortening the ultimate chain length of prenyl products (Fig. 2). The corresponding residues in M. luteus B-P 26 UPP synthase, i.e. Ala72, Phe73, and Trp78, are located in the highly conserved region III, and were suggested to be located close to the bind- ing site for IPP, which had been identified by a series of site-directed mutagenesis studies [20]. In the crystal structure of UPP synthase from M. luteus B-P 26, these residues are located at the edge of the large hydrophobic cleft. Recently, crystal structure of E. coli UPP synthase bound with the allylic substrate E,E-FPP [28] has been revealed. In this structure, Ala69, which corresponds to Ala72 of M. luteus B-P 26 UPP synthase, is located close to the x-end carbon, C-14 of E,E-FPP at a distance of 3.2 A ˚ . More recently, crystal structure of the D26A mutant E. coli UPP syn- thase, which shows about an 800-fold lower k cat value than the wild-type enzyme, bound with IPP, has also been described [29]. In order to understand the molecular mechanisms of chain length determination, we built structural models of Rv1086 and of LLLH mutant of M. luteus UPP synthase using the E. coli UPP synthase-E,E-FPP complex structure as template, based on multiple sequence alignments of cis-prenyl- transferases previously identified. Then, structures of E,E-FPP and IPP were superimposed on the structural models (Fig. 4). In these proposed models, Leu residue which corresponds to Ala69 of E. coli UPP synthase is closer to the C-14 of E,E-FPP, suggesting that the bulky alkyl group of Leu72 in LLLH may interfere with cis-addition of IPP onto E,E-FPP or Z,E-FPP. In contrast, Phe73 and Trp78 of M. luteus B-P 26 UPP synthase, corresponding to Leu85 and Leu90 of Rv1086, respectively, are not close to the binding site for E,E-FPP [5,28]. In the crystal structure of M. luteus B-P 26 UPP synthase without substrates, res- idues from Ser74 to Val85 downstream of sheet-2 in the conserved region III could not be defined because of high flexibility [5]. In addition, substitution of the highly conserved Phe73 and Ser74 in region III of M. luteus B-P 26 UPP synthase into Ala resulted in 32- and 16-fold increases in K m value for IPP, and 16- and 12-fold decreases in k cat value, respectively [20]. These results indicated that the flexible domain in the conserved region III was important for binding of IPP in the proper direction, and for catalytic function. The double mutant A72L ⁄ F73L prepared in the pre- sent study also showed an 88-fold higher K m value for IPP compared with the wild-type enzyme. However, the triple mutant A72L ⁄ F73L ⁄ W78L showed a sixfold Fig. 4. Overall catalytic centre of a structural model for the LLLH mutant. The model was built based on the crystal structure of E,E- FPP complex of UPP synthase from E. coli (Protein Data Bank No. 1V7U). Then, structures of E,E-FPP and IPP, which were identified from the complex structure of IPP and the D26A mutant of E. coli UPP synthase (Protein Data Bank No. 1X09), were superimposed on the structural models. Amino acid residues mutated in the LLLH mutant are indicated in red, and are overlapped with residues in the wild-type UPP synthase of M. luteus B-P 26. The structural P-loop motif, proposed to function in binding of allylic substrates such as E,E-FPP, and charged residues including Arg197 and Arg203 are indicated. Product chain length of cis-prenyltransferases Y. Kharel et al. 652 FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS increased K m value for IPP (Table 1). These results suggested that replacement of Trp78 for Leu was necessary for constitution of a proper IPP binding domain when Ala72 and Phe73 were replaced with Leu residues which corresponded to Leu84, -85, and -90, respectively, in Rv1086. Substitution of Phe223 with His dramatically decreased the catalytic activity of UPP synthase when E,E-FPP was used as allylic substrate. However, GPP could be accepted by these mutants as allylic substrate to produce shorter prenyl products, i.e. C 45)55 and C 15 (Fig. 2B). In structural models of Rv1086 and the LLLH mutant of M. luteus UPP synthase (Fig. 4), His237 of Rv1086 is located in proximity of the struc- tural P-loop motif, which is thought to recognize the diphosphate group of allylic substrates such as E,E-FPP [5,23]. Substitution of His for Phe223 may affect binding affinity for the allylic substrate with the structural P-loop motif. This hypothesis agrees with the fact that in the prenyltransferase reaction by Rv1086 in vitro, GPP and neryl diphosphate (C 10 ) were the only functional allylic substrates among the five allylic substrates tested, including dimethylallyl diphos- phate (C 5 ) E,E-FPP and E,E,E-geranylgeranyl diphos- phate (C 20 ) [30]. On the other hand, three to seven extra amino acid residues found downstream of the conserved region III were shown to be important for the production of long chain Z,E-mixed polyisoprenoids by the long-chain cis- prenyltransferase subfamily. According to the crystal structure of UPP synthase of M. luteus B-P 26, these extra residues are located on the side wall of the large hydrophobic cleft [5]. Ko et al. reported that replace- ment of Leu137 of E. coli UPP synthase, which is located at the ‘bottom’ of the hydrophobic cleft, with Ala resulted in elongation of the ultimate chain length of Z,E-mixed polyisoprenoids, producing C 55 and C 60 as major products in the presence of 0.1% Triton X-100 [21]. Moreover, they indicated that the mutant L137A produced C 70 and C 75 as major products in a reaction without Triton X-100 for 96 h [21]. Based on these results, they proposed that Leu137 functioned as the floor of the tunnel to block further elongation of polyprenyl products. This proposed model seems to be analogous to the chain-length determination mechan- ism of trans-prenyltransferases, in which bulky residues located upstream of FARM play an important role in determination of the ultimate prenyl chain length, composing a suitable size for the pocket for growing of the isoprenoid chain [2,15,16]. According to this model, we constructed and analysed a mutant of M. luteus B-P 26 UPP synthase L140A, which corres- ponded to the E. coli UPP synthase mutant L137A, and found that the mutant also produced C 55 and C 60 as major products in the presence of Triton X-100 (data not shown). However, our results obtained in the present study couldn’t be explained by the model pro- posed by Ko et al. because Leu137 of E. coli UPP synthase did not correspond to the domain where we introduced the extra amino acid residues, and the mutants EKE and RAKDY produced even longer polyprenyl products (C 60)75 ) than the E. coli UPP syn- thase mutant L137A in the presence of Triton X-100. Insertion of five Ala residues instead of the peptides RAKDY did not cause a significant change in the length of prenyl products, indicating a requirement for insertion of specific amino acid residues for lengthen- ing the ultimate product chain. Moreover, this sugges- ted the significance of some charged or polar residues at the proper positions rather than expansion of the interior space of the hydrophobic cleft. In the crystal structure of E. coli UPP synthase bound with E,E-FPP, helix-3 is kinked to be closer to the E,E-FPP binding domain compared with the struc- ture without substrates [28], indicating that the structure of UPP synthase shifts to the closed conformation when the substrate binding site is shared with the allylic sub- strate. The interior of the large hydrophobic cleft, sur- rounded by sheet-2, sheet-4, helix-2, and helix-3, mainly consists of hydrophobic residues [5]. Hydrophobic resi- dues on kinked helix-3 in the closed conformation may function as a guide rail to introduce the elongating pre- nyl chain in the proper direction (Figs 5A and B). Although most of residues constituting the hydrophobic cleft are highly conserved among cis-prenyltransferases, residues localized on helix-3 show a wide diversity (Fig. 3A). Moreover, the domain in which we intro- duced extra amino acid residues corresponded to the hinge region of the kinked helix-3 (Fig. 5B). Therefore, we proposed that charged residues inserted at the hinge region of helix-3 might control the bending direction of the growing hydrophobic prenyl chain along the hydro- phobic interior of helix-3 so that the hydrophobic cleft could accommodate the bulk of the prenyl chain to fit a suitable size during enzymatic elongation. In conclusion, we identified critical regulatory domains in cis-prenyltransferases for determination of the ultimate product chain length, and proposed a model for chain-length determination. Further elucida- tion and manipulation of chain-length determination mechanisms of cis-prenyltransferase are not only attractive but also very important as a biotechnological aspect because biological materials composed of the polymer of IPP with cis-configuration such as natural rubber, can be used for the development of novel func- tional materials. Y. Kharel et al. Product chain length of cis-prenyltransferases FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS 653 Experimental procedures Materials and general procedures Nonlabelled IPP and E,E-FPP were synthesized according to Davisson et al. [31]. 1- 14 C-labelled IPP (1.95 TBqÆmol )1 ) was from Amersham Biosciences (Princeton, NJ, USA). Restriction enzymes and other DNA modifying enzymes were from TaKaRa Bio (Otu, Japan) and TOYOBO (Osaka, Japan). Potato acid phosphatase was from Sigma. Precoated reversed phase TLC plate, LKC-18 was pur- chased from Whatman (Brentford, UK). The yeast strain SNH23-7D, mutant allele rer2-2 [7], and pRS316 containing SRT1 [8] were kindly provided by A. Nakano and M. Sato (RIKEN, Japan). The expression vector for yeast, pJR1133 was kindly provided by A. Ferrer (University of Barcelona, Spain). Restriction enzyme digestions, transformations, and other standard molecular biological techniques were carried out as described by Sambrook et al. [32]. All other chemi- cals were of analytical grade. Multiple alignments and homology modelling of mutant enzymes Amino acid sequences of cis-prenyltransferases already identified were aligned using the clustal w Multiple Sequence Alignment Program [33]. The resulting alignment was edited by seqvu (Shareware alignment programme, Garvan Institute, Sydney, Australia). Three-dimensional models for the mutants, Rv1086 and Srt1p were obtained by The swiss-model alignment interface (http://swissmodel. expasy.org/) using multiple alignment of cis-prenyltrans- ferases and crystal structures of UPP synthase from M. luteus B-P 26 [5] or E. coli [21,28,29] as structural templates. Expression vector system and site-directed mutagenesis The expression plasmid, pMluUEX for M. luteus B-P 26 UPP synthase [20] was used as template for preparation of the mutants. Site-directed mutagenesis was carried out according to protocols for the Gene Editor in vitro Site- Directed Mutagenesis System (Promega, Madison, MI, USA). The single stranded wild-type UPP synthase gene used as template in the mutagenesis reaction was prepared by infection of E. coli JM109 cells (TaKaRa) harboring pMluUEX with R408 helper phages. Mutagenic oligonucleo- tides designed to produce the desired mutant enzymes were: 5¢-CAGTTGACAATAAGTACAGCG-3¢ (for A72L ⁄ F73L); 5¢-CTTTAGGTCGACTCAAATTTTCAGTTGACAATAA GTACAGCG-3¢ (for A72L ⁄ F73L ⁄ W78L); 5¢-CCGGCCAG TGTTCATCGATAAATAC-3¢ (for F223H); 5¢-CTTTAGG TCGACTCAAAT TTTCAG TTGACA ATAAGTACA GCG-3 ¢ and 5¢-CCGGCCAGTGTTCATCGATAAATAC-3¢ (for A72L ⁄ F73L ⁄ W78L ⁄ F223H). For insertion of three or five Fig. 5. Large hydrophobic cleft of M. luteus B-P 26 UPP synthase with or without FPP. Models were built based on the crystal structure of E,E-FPP complex of UPP synthase from E. coli (Protein Data Bank No. 1V7U). Then, the structure of E,E-FPP was superimposed on the models. Side chains are coloured as follows: nonpolar, grey; uncharged polar, yellow; acidic, red; basic, blue. (A) Large hydrophobic cleft of M. luteus B-P 26 UPP synthase (Protein Data Bank No. 1F75) consisting of helix-2, helix-3, sheet-2, and sheet-4. Only side chains facing towards the inside of the hydrophobic cleft are shown. (B) Structural model of the large hydrophobic cleft of M. luteus B-P 26 UPP synthase with E,E-FPP. Only side chains facing towards the inside of the hydrophobic cleft are shown. The white arrow indicates a predicted direction of the prenyl chain elongation. The circle indicates regions in which the extra amino acid residues, EKE or RAKDY were introduced. Product chain length of cis-prenyltransferases Y. Kharel et al. 654 FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS amino acid residues between Pro101 and Glu102 of M. luteus B-P 26 UPP synthase, oligonucleotides used were: 5¢-CAAT GAGTTCCTCCTTCTCCGGTAAAAATGTG-3¢ (for EKE); and 5¢-AACATTTTTTTCAATGAGCTCATAGTCCTTG GCTCTCGGTAAAAATG-3¢ (for RAKDY). Introduction of mutations was confirmed by sequencing whole nucleotide sequences using the dideoxy chain-termination method with a DNA sequencer (LI-COR, model 4200, ALOKA, Tokyo, Japan). Construction of deletion mutant for Srt1p The SRT1 expression plasmid was constructed with the SRT1 fragment amplified by PCR using appropriate prim- ers. The sense and antisense primers, 5¢-CGTTTCTGGGT ACCATAATGAAAATGC-3¢ and 5¢-CTAGTTGTCGACT TTTACTTATTCATC-3¢, respectively, were designed to create a KpnI site upstream the starting codon ATG, and a SalI site downstream the stop codon TAA, respectively. The PCR product was digested with KpnI and SalI, and separated on a 1% agarose gel. The desired band was elut- ed, and ligated into the pBluescript II SK(–) vector, then digested with the same restriction enzymes to give pBS-Srt1. Five amino acids deletion construct of Srt1p (Srt1p-delta) was generated using the Gene Editor in vitro Site-Directed Mutagenesis System (Promega) with the primer 5¢-GATGAATTCGCGAAGAAGGATCCCTTAT AC-3¢ to obtain pBS-Srt1p-delta. Wild-type SRT1 and the mutant Srt1p-delta were digested with KpnI and SalI, subcloned into pJR1133 vector, and digested with the same restriction enzymes to give pJR1133-Srt1p and pJR1133- Srt1p-delta, respectively. Overproduction and purification of UPP synthase mutant Each construct for expression of target proteins was introduced into E. coli BL21(DE3) cells, and cells were cultured in Luria–Bertani or M9YG medium. The proce- dures us for overproduction and purification of UPP syn- thase mutants and the wild-type enzyme were essentially similar to those described in a previous paper [20,24]. Purity of the mutant enzymes was analysed by SDS ⁄ PAGE with Coomassie brilliant blue staining. Pro- tein concentrations were measured by the method of Bradford with BSA as standard. Expressions of Srt1p and Srt1p-delta in yeasts pJR1133-Srt1p and pJR1133-Srt1p-delta plasmids were introduced into yeast rer2-2 mutant strain, SNH23-7D (MATa rer2 mfa1::TRP1::HIS3 ura3 trp1 ade2 leu2 his3 lys2), which is deficient in the activity of DedolPP synthase [8]. Transformants were selected as Ura + colonies, were cultured at 23 °C on agar plates containing minimal medium (0.16% yeast nitrogen base without amino acids, 0.5% ammonium sulfate, 2% glucose, supplemented with 60 lgÆmL )1 Leu and 30 lgÆmL )1 Lys). Yeast cells were grown in 100 mL minimal medium until the late-logarithmic phase (OD 600 of 1.0). Preparation of yeast membrane fraction proteins from wild-type Srt1p and Srt1p-delta mutants Yeast cells were harvested by centrifugation at 3000 g for 5 min, suspended in 300 lL zymolyase buffer (50 mm Tris ⁄ HCl pH 7.5, 10 mm MgCl 2 ,1m sorbitol, 30 mm dithiothreitol, 2 mgÆmL )1 zymolyase 100T), and incubated at 30 °C for 30 min. The spheroplast-formed cells were collected by centrifugation at 1000 g for 5 min, and resus- pended in 400 lL breakage buffer (50 mm KH 2 PO 4 pH 7.5, 1 m dithiothreitol, 1 mm phenylmethylsulfonyl fluoride, 1 lgÆmL )1 aprotinin, 1 lgÆmL )1 leupeptin, and 1 lgÆmL )1 pepstatin A). Using an equal volume of glass beads, cell suspensions were vortexed for 30 s, followed by incubation in an ice bath for 1 min, repeated five times. Cell lysates were centrifuged at 300 g for 5 min to remove unbroken cells. Supernatants were further centrifuged at 13 000 g for 10 min to separate membrane and soluble fractions. Cis-Prenyltransferase assay and product analysis UPP synthase activity was measured by determining amounts of [1- 14 C]IPP incorporated into butanol-extracta- ble polyprenyl diphosphates. The standard assay mixture contained 100 mm Tris ⁄ HCl pH 7.5, 0.5 mm MgCl 2 , 10 lm E,E-FPP, 10 lm 1- 14 C-labelledIPP (37 MBqÆmol )1 ), 0.05% (w ⁄ v) Triton X-100, and a suitable amount of enzyme solution in a final volume of 200 lL. For the yeast cis-prenyltransferase assay, the standard reaction mixture contained (final volume of 100 lL) 25 mm phos- phate buffer pH 7.5, 20 mm 2-mercaptoethanol, 20 m m potassium fluoride, 4 mm MgCl 2 ,50lm [1- 14 C]IPP, 10 lm E,E-FPP, and 50 lg proteins of membrane frac- tions. After incubation at 30 °C for 30 min, reaction products were extracted with 1-butanol saturated with water, and radioactivity in the butanol extract was meas- ured with an Aloka LSC-1000 liquid scintillation counter. For product analysis, radioactive prenyl diphosphate products in the reaction mixture were hydrolysed to the corresponding alcohols with potato acid phosphatase according to the method of Fujii et al. [34]. Product alcohols were extracted with pentane, and analysed by reversed phase TLC with a solvent system of acet- one ⁄ water (19 : 1) (for bacterial cis-prenyltransferase) or acetone ⁄ water (39 : 1) (for yeast cis-prenyltransferase). Positions of authentic standards were visualized with Y. Kharel et al. Product chain length of cis-prenyltransferases FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS 655 iodine vapour, and distribution of radioactivity was ana- lysed with a Fuji BAS 1000 Mac Bioimage Analyzer. Acknowledgements We are grateful to Dr A. Nakano and Dr M. Sato (RIKEN, Japan) for kindly providing the yeast strain SNH23-7D, and Dr A. Ferrer (University of Barce- lona, Spain) for kindly providing the vector pJR1133. This work was supported in part by Grants-in-Aid for Scientific Research from JSPS, Japan Society for the Promotion of Science, and by the Foundations of Takeda Science, Eno Science, The Yuasa International, and by the Sumitomo Foundation. References 1 Ogura K & Koyama T (1998) Enzymatic aspects of iso- prenoid chain elongation. Chem Rev 98, 1263–1276. 2 Wang KC & Ohnuma S (2000) Isoprenyl diphosphate synthases. Biochim Biophys Acta 1529, 33–48. 3 Kharel Y & Koyama T (2003) Molecular analysis of cis-prenyl chain elongating enzymes. Nat Prod Report 20, 111–118. 4 Shimizu N, Koyama T & Ogura K (1998) Molecular cloning, expression, and purification of undecaprenyl diphosphate synthase. No sequence similarity between E- and Z-prenyl diphosphate synthases. J Biol Chem 273, 19476–19481. 5 Fujihashi M, Zhang Y-W, Higuchi Y, Li X-Y, Koyama T & Miki K (2001) Crystal structure of cis-prenyl chain elongating enzyme, undecaprenyl diphosphate synthase. Proc Natl Acad Sci USA 98, 4337–4342. 6 Apfel CM, Takacs B, Fountoulakis M, Stieger M & Keck W (1999) Use of genomics to identify bacterial undecaprenyl pyrophosphate synthetase: cloning, expres- sion, and characterization of the essential uppS gene. J Bacteriol 181, 483–492. 7 Sato M, Sato K, Nishikawa S, Hirata A, Kato J & Nakano A (1999) The yeast RER2 gene, identified by endoplasmic reticulum protein localization mutations, encodes cis-prenyltransferase, a key enzyme in dolichol synthesis. Mol Cell Biol 19, 471–483. 8 Sato M, Fujisaki S, Sato K, Nishimura Y & Nakano A (2001) Yeast Saccharomyces cerevisiae has two cis-pre- nyltransferases with different properties and localiza- tions. Implication for their distinct physiological roles in dolichol synthesis. Genes Cells 6, 495–506. 9 Oh SK, Han KH, Ryu SB & Kang H (2000) Molecular cloning, expression, and functional analysis of a cis-pre- nyltransferase from Arabidopsis thaliana. Implications in rubber biosynthesis. J Biol Chem 275, 18482–18488. 10 Cunillera N, Arro ´ M, Fore ´ s O, Manzano D & Ferrer A (2000) Characterization of dehydrodolichyl diphosphate synthase of Arabidopsis thaliana, a key enzyme in doli- chol biosynthesis. FEBS Lett 477, 170–174. 11 Schulbach MC, Brennan PJ & Crick DC (2000) Identifi- cation of a short (C 15 ) chain Z-isoprenyl diphosphate synthase and a homologous long (C 50 ) chain isoprenyl diphosphate synthase in Mycobacterium tuberculosis. J Biol Chem 275, 22876–22881. 12 Endo S, Zhang Y-W, Takahashi S & Koyama T (2002) Identification of human dehydrodolichyl diphosphate synthase gene. Biochim Biophys Acta 1625, 291–295. 13 Asawatreratanakul K, Zhang Y-W, Wititsuwannakul D, Wititsuwannakul R, Takahashi S, Rattanapittayapron A & Koyama T (2003) Molecular cloning, expression and characterization of cDNA encoding cis-prenyltrans- ferases from Hevea brasiliensis. A key factor participat- ing in natural rubber biosynthesis. Eur J Biochem 270, 4671–4680. 14 Ohnuma S, Narita K, Nakazawa T, Ishida C, Takeuchi Y, Ohto C & Nishino T (1996) A role of the amino acid residue located on the fifth position before the first aspartate-rich motif of farnesyl diphosphate synthase on determination of the final product. J Biol Chem 271, 30748–30754. 15 Wang K & Ohnuma S (1999) Chain-length determina- tion mechanism of isoprenyl diphosphate synthases and implications for molecular evolution. Trends Biochem Sci 24, 445–451. 16 Tarshis LC, Proteau PJ, Kellogg BA, Sacchettini JC & Poulter CD (1996) Regulation of product chain length by isoprenyl diphosphate synthases. Proc Natl Acad Sci USA 93, 15018–15023. 17 Fujikura K, Zhang Y-W, Yoshizaki H, Nishino T & Koyama T (2000) Significance of Asn-77 and Trp-78 in the catalytic function of undecaprenyl diphosphate synthase of Micrococcus luteus B-P 26. J Biochem (Tokyo) 128, 917–922. 18 Pan J-J, Yang L-W & Liang P-H (2000) Effect of site- directed mutagenesis of the conserved aspartate and glu- tamate on E. coli undecaprenyl pyrophosphate synthase catalysis. Biochemistry 39, 13856–13816. 19 Pan J-J, Chiou S-T & Liang P-H (2000) Product distri- bution and pre-steady-state kinetic analysis of Escheri- chia coli undecaprenyl pyrophosphate synthase reaction. Biochemistry 39, 10936–10942. 20 Kharel Y, Zhang Y-W, Fujihashi M, Miki K & Koyama T (2001) Identification of significant residues for homoallylic substrate binding of Micrococcus luteus B-P 26 undecaprenyl diphosphate synthase. J Biol Chem 276, 28459–28464. 21 Ko TP, Chen YK, Robinson H, Tsai PC, Gao YG, Chen AP, Wang AH & Liang PH (2001) Mechanism of product chain length determination and the role of a flexible loop in Escherichia coli undecaprenyl-pyrophosphate synthase catalysis. J Biol Chem 276, 47474–47482. Product chain length of cis-prenyltransferases Y. Kharel et al. 656 FEBS Journal 273 (2006) 647–657 ª 2006 The Authors Journal compilation ª 2006 FEBS [...]... the determination of dolichyl phosphates in tissues Annal Biochem 156, 380–385 27 Matsuoka S, Sagami H, Kurisaki A & Ogura K (1991) Variable product specificity of microsomal dehydrodolichyl diphosphate synthase from rat liver J Biol Chem 266, 3464–3468 28 Chang SY, Ko TP, Chen AP, Wang AH & Liang PH (2004) Substrate binding mode and reaction mechanism of undecaprenyl pyrophosphate synthase Product chain. .. undecaprenyl pyrophosphate synthase Product chain length of cis-prenyltransferases 29 30 31 32 33 34 deduced from crystallographic studies Protein Sci 13, 971–978 Guo RT, Ko TP, Chen AP, Kuo CJ, Wang AH & Liang PH (2005) Crystal structures of undecaprenyl pyrophosphate synthase in complex with magnesium, isopentenyl pyrophosphate, and farnesyl thiopyrophosphate: roles of the metal ion and conserved residues... AH & Liang PH (2002) Probing the conformational change of Escherichia coli undecaprenyl pyrophosphate synthase during catalysis using an inhibitor and tryptophan mutants J Biol Chem 277, 7369–7376 23 Fujikura K, Zhang Y-W, Fujihashi M, Miki K & Koyama T (2003) Mutational analysis of allylic substrate binding site of Micrococcus luteus B-P 26 undecaprenyl diphosphate synthase Biochemistry 42, 4035– 4041... Kharel Y, Zhang Y-W, Fujihashi M, Miki K & Koyama T (2003) Significance of highly conserved aromatic residues in Micrococcus luteus B-P 26 undecaprenyl diphosphate synthase J Biochem (Tokyo) 134, 819–826 25 Yamada K, Yokohama H, Abe S, Katayama K & Sato T (1985) High-performance liquid chromatographic method for the determination of dolichols in tissues and plasma Annal Biochem 150, 26–31 26 Yamada K,... York Thompson JD, Higgins DG & Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice Nucleic Acids Res 22, 4673–4680 Fujii H, Koyama T & Ogura K (1982) Efficient enzymatic hydrolysis of polyprenyl pyrophosphates Biochim Biophys Acta 712, 716–718 FEBS Journal 273 (2006) 647–657... Barontini S, Papi C, Minutolo F, Bertini S, Brennan PJ & Crick DC (2001) Purification, enzymatic characterization, and inhibition of the Z-farnesyl diphosphate synthase from Mycobacterium tuberculosis J Biol Chem 276, 11624– 11630 Davisson VJ, Woodside AB & Poulter CD (1985) Synthesis of allylic and homoallylic isoprenoid pyrophosphates Methods Enzymol 110, 130–144 Sambrook J, Fritsch EF & Maniatis T (1989) . determination of the ultimate product chain length, and proposed a model for chain- length determination. Further elucida- tion and manipulation of chain- length determination mechanisms. Manipulation of prenyl chain length determination mechanism of cis-prenyltransferases Yugesh Kharel*, Seiji Takahashi*,