Biosci. Biotechnol. Biochem., 1 (5), 1365 1369, Communication The Positive Effect of the Decreased NADPH-Preferring Activity of Xylose Reductase from Pichia stipitis on Ethanol Production Using Xylose-Fermenting Recombinant Saccharomyces cerevisiae Seiya WATANABE, 1;;3 Seung Pil PACK, ;3 Ahmed Abu SALEH, Narayana ANNALURU, Tsutomu KODAKI, ;3 and Keisuke MAKINO ;3;4;y 1 Faculty of Engineering, Kyoto University, Kyotodaigaku-katsura, Saikyo-ku, Kyoto 615-853, Japan Institute of Advanced Energy, Kyoto University, Gokasho, Uji, Kyoto 611-11, Japan 3 CREST, JST (Japan Science and Technology Agency), Gokasho, Uji, Kyoto 611-11, Japan 4 International Innovation Center, Kyoto University, Yoshidahonmachi, Sakyo-ku, Kyoto 66-851, Japan Received February 1, ; Accepted March 3, ; Online Publication, May, [doi:1.11/bbb.14] We focused on the effects of a mutation of xylose reductase from Pichia stipitis (PsXR) on xylose-toethanol fermentation using recombinant Saccharomyces cerevisiae transformed with PsXR and PsXDH (xylitol dehydrogenase from P. stipitis) genes. Based on inherent NADH-preferring XR and several site-directed mutagenetic studies using other aldo-keto reductase enzymes, we designed several single PsXR mutants. KR showing decreased NADPH-preferring activity without a change in NADH-preferring activity was found to be a potent mutant. Strain Y-KR transformed with KR PsXR and wild-type PsXDH showed a 31% decrease in unfavorable xylitol excretion with 5.1% increased ethanol production as compared to the control in the fermentation of 15 g l 1 xylose and 5 g l 1 glucose. Key words: xylose reductase; site-directed mutagenesis; coenzyme specificity; Saccharomyces cerevisiae; xylose fermentation Xylose is a major fermentable sugar present in lignocellulosic biomass, the second most abundant carbohydrate in nature to glucose. 1) Efficient xylosefermenting strains are required to develop economically viable processes to produce biofuels such as ethanol from biomass. ) Although Saccharomyces cerevisiae is used generally for industrial ethanol production, native S. cerevisiae does not ferment xylose. Therefore, the engineering of S. cerevisiae for xylose utilization has focused on adapting the xylose metabolic pathway from the xylose-utilizing yeast Pichia stipitis. In this organism, xylose is converted to xylulose by two oxidoreductases, NAD(P)H-linked xylose reductase (XR; EC 1.1.1.1) 3) and NAD þ -linked xylitol dehydrogenase (XDH; EC 1.1.1.9), 4) and finally, xylulokinase (XK; EC..1.1) phosphorylates xylulose into xylulose 5- phosphate, which is metabolized further via the pentosephosphate pathway. Although S. cerevisiae transformed native XYL1 and XYL genes encoding XR and XDH from P. stipitis (referred to as PsXR and PsXDH, respectively) was the most potent recombinant strain, it was not sufficient to apply to the industrial bio-process. One of the main reasons is a intracellular redox imbalance caused by the different coenzyme specificity between PsXR and PsXDH. ) Hence, modifying the coenzyme specificity of XR and/or XDH by protein engineering is an attractive challenge for efficient ethanol fermentation from xylose using recombinant S. cerevisiae. In the case of PsXDH, recently we achieved complete reversal coenzyme specificity toward NADP þ of XDH using unique NADP þ - dependent sorbitol dehydrogenase as a reference enzyme. 5) Also, effective ethanol fermentation and a reduction in xylitol excretion were found when a novel NADP þ -dependent XDH mutant was co-expressed with NADPH-preferring PsXR wild-type (WT PsXR) in S. cerevisiae cells (Watanabe et al., unpublished). In the case of PsXR, Jeppsson et al. 6) recently reported that enhanced ethanol yield accompanied by decreased xylitol excretion was found in recombinant S. cerevisiae carrying the KM PsXR mutant with increased K m for NADPH ) together with PsXDH WT. This is the first improvement by protein engineering of PsXR on ethanol fermentation from xylose. On the other hand, Lee et al. 8) cloned the XYL1 gene encoding unique NADH-preferring XR from Candida parapsilosis (CpXR; accession no. AY19316), and employed ethanol production by expressing NADH-preferring XR in y To whom correspondence should be addressed. Tel: +81-4-38-351; Fax: +81-4-38-354; E-mail: kmak@iae.kyoto-u.ac.jp Abbreviations: XR, xylose reductase; XDH, xylitol dehydrogenase; PsXR, XR from Pichia stipitis; PsXDH, XDH from Pichia stipitis; WT, wildtype enzyme; CpXR, XR from Candida parapsilosis; AKR, aldo-keto reductase; CtXR, XR from Candida tenuis
1366 S. WATANABE et al. Candida tropicalis, which contains typical NAD þ - dependent XDH. 9) In fact, XR (AKRB5) is a unique member with dual coenzyme specificity in the aldo-keto reductase (AKR) superfamily, 1) but significant NADPH preference is found in most XRs, similarly to most AKR superfamily enzymes. However, CpXR utilizes NADH 1-fold higher than NADPH in catalytic efficiency. Since CpXR is the only XR with greater NADH preference among native XRs known, its sequence information can be used in designing coenzyme modification of other XRs, in particular, NADPH to NADH preference change. Furthermore, site-directed mutagenetic study of NADH preference using XR from Candida tenuis (CtXR) has recently been reported. 11) This fungal XR shows about % sequential similarity to PsXR, although no effect of generated mutant expression has been reported. In the present study, based on this information, sitedirected mutagenesis was performed to obtain PsXR with decreased NADPH preference. Among several PsXR mutants, a potent PsXR mutant showing decreased NADPH preferring activity without a change in NADH-preferring activity was obtained. By co-expression with NAD þ -dependent PsXDH WT in S. cerevisiae, the effects of the mutated PsXR on ethanol production and xylitol excretion was analyzed in xylose fermentation. The XYL1 gene of WT PsXR was cloned from P. stipitis (Yamadazyma stipitis NBRC 168, from the National Institute of Technology and Evaluation, Chiba, Japan) based on the published sequence of the P. stipitis XYL1 gene (GenBankÔ accession no. X59465) and introduced into pqe-81l (Qiagen, Hilden, Germany), a plasmid vector for conferring N-terminal (His) 6 -tag on expressed proteins, to obtain phis(wt). Site-directed mutagenesis was carried out by the single-round PCR method using synthesis oligonucleotide primers containing appropriate mutations and phis(wt) as a template. The codons used for single mutants were as follows: KR (AAG!AGA), KG (AAG!GGT), S1A (TCC!GCC), ND (AAC!GAC), and T3S (ACT!TCT). Escherichia coli DH5 harboring the expression plasmids for the (His) 6 -tagged WT and mutated PsXRs were cultivated at 3 C, and structural genes of PsXRs were significantly induced at 18 Cby 1mM isopropyl--d-thiogalactopyranoside (IPTG). The expressed proteins were confirmed by SDS PAGE separation and Western blotting using an antibody against (His) 6 -tag, and then WT and mutated PsXRs were purified by the ÄKTA purifier system (Amersham Biosciences, Piscataway, NJ) with a Ni-NTA superflow column (Qiagen) and subsequent gel filtration (data not shown). SDS PAGE analysis revealed that the purities of the recombinant enzymes were almost homogeneous, and that the apparent molecular mass of the PsXRs was about 38, Da, in good agreement with the calculated molecular mass of the enzyme with (His) 6 -tag (3,61. Da) (data not shown). For the purified enzymes, XR activity was measured by monitoring the oxidation of NADPH or NADH at 34 nm at 3 Cina reaction mixture of the following composition: 6 mm potassium phosphate buffer (ph.), 133 mm xylose, and.15 mm NADPH (or NADH). One unit of enzyme activity was defined as the amount of enzyme that reduced or oxidized 1 mmol of NAD(P) þ or NAD(P)H per min, and the protein concentrations were determined by the method of Lowry et al., 1) with bovine serum albumin as the standard. To estimate the effect of the KR PsXR mutant on ethanol fermentation from xylose, each XYL1 gene encoding to WT and KR each was expressed in S. cerevisiae together with the XYL gene encoding WT PsXDH. Initially, both the XYL1 and the XYL gene were ligated to a yeast expression vector (ppgk) with the constitutive PGK promoter. 13) In the case of the XYL1 genes, the PsXR coding region with PGK promoter and terminator was excised from the plasmid and reinserted into the YEpM4 vector. 14) Yeast transformation was performed by the lithium acetate method. 15) S. cerevisiae D45- (MATa leu his3 ura3 can1) 16) was transformed with ppgk-xdh, and then further transformed with YEpM4-XR(WT) and YEpM4-XR(KR) to construct recombinant yeast strains Y-WT and Y- KR respectively. Both empty vectors, ppgk and YEpM4, were transformed into S. cerevisiae D45- to construct a control strain (Y-vector). Recombinant yeasts were compared in terms of their ethanol production and xylitol excretion, which were analyzed in batch fermentation using a 1 ml minimal medium (6. g l 1 yeast nitrogen base without amino acid) with 5 g l 1 glucose plus 15 g l 1 xylose in a ml flask under anaerobic conditions at 3 C at 15 rpm. The ethanol concentration was measured by gas chromatography (model GC-14B, Shimadzu, Kyoto, Japan, fitted with a flame ionization detector) using a glass column packed with Thermon-3 (:m 3:mm, Shimadzu). The concentrations of glucose, xylose, xylitol, glycerol, and acetic acid were estimated by HPLC with a Multi-Station LC-8 model II system (Tosoh, Tokyo). Samples (1 ml) were applied at 3 C to an Aminex HPX-8H Organic Analysis column (3 :8 mm, Bio-Rad, Hercules, CA) linked to an RID-8 refractive index detector (Tosoh) and eluted with 5 mm H SO 4 at a flow rate of.4 ml min 1. Cell growth was monitored by measuring absorbance at 6 nm using a spectrophotometer (model U-1, Hitachi, Tokyo). Generally, the coenzyme recognition pocket of AKR superfamily enzymes including XRs are located at the amino acid sequences equivalent to Lys -Ser 1 - Asn -Thr 3 in PsXR (Fig. 1A). As describe above, CpXR is a unique NADH-dependent enzyme, in which Lys in PsXR (as in other NADPH-preferring XRs) is replaced with a basic arginine. Furthermore, only CpXR possesses serine residue at position. On the other hand, in crystal structures of NADH- or NADPH-CtXR
A Protein Engineering of Xylose Reductase 136 B Enzyme Coenzyme preference Amino acid sequence WT NADPH I P K S N T V P KR R KG G P. stipitis XR S1A This study A ND D T3S S C. parapsilosis XR WT NADH R S P D WT NADPH L P E C. tenuis XR K4R (NADH) R L P E N6D (NADH) D L P E WT NADPH F V R R E,5-DKGRA K3G (NADH) F G V R R E 1 3 WT KR KG S1A ND T3S Specific activity (U mg -1 ) 5 1 15 5 NADH NADPH Fig. 1. Site-Directed Mutagenesis of PsXR. A, Mutation design of PsXR. The gray box indicates the modified part of PsXR. A dot indicates the same amino acids as PsXR WT.,5- DKGRA is,5-diketo-d-gluconic acid reductase from Corynebacterium sp. (GenBankPPPPÔ accession no. AAA83534). A significant improvement in NADH preference was found in the K4R and N6D mutants of CtXR and the K3G mutant of,5-dkgra. B, Specific activities of WT and mutated PsXR. Black and gray bars indicate NADH- and NADPH-preferring XR activities respectively for each PsXR. Average values and standard deviation were obtained by measuring the specific activity three times. complex, Ser 5, a residue equivalent to PsXR Ser 1, interacts originally with NADPH, but not with NADH. Furthermore, it was recently reported that a mutation equivalent to ND in PsXR is efficient for acquisition of NADH-preference in CtXR. 11) On the other hand, the PsXR KG mutant was designed by site-directed mutagenetic study using,5-diketo-d-gluconic acid reductase. 1) This enzyme, a member of the AKR superfamily, also shows inherent NADPH specificity, and a mutation equivalent to KG in PsXR leads to efficient NADH preference. Hence, we compared WT and five mutated PsXRs (KR, KG, S1A, ND, and T3S) by characterizing each specific activity according to the different coenzyme usages (Fig. B). Similarly to most XRs, WT PsXR possessed higher NADPH-preferring activity, although it gave dual coenzyme specificity:. and 15. U mg 1 of specific activities for NADH and NADPH respectively (NADH/ NADPH =.5). Among the five single mutants, KR showed decreased NADPH-preferring XR activity, half of the specific activity of WT PsXR, while it showed no significant change in the NADH-preferring activity of WT PsXR, 6. and. U mg 1 of the specific activities for NADH and NADPH respectively (NADH/NADPH is.9). On the other hand, the ratios of NADH/NADPH to S1A and ND were.3 and.6 respectively, which indicates a higher preference for NADPH. Modifications of KG and T3S were found to induce a loss of XR activity (below 1% of the activity of WT PsXR) in the cases of both coenzymes. Indeed, a significant decrease of expression in E. coli cells was found, as compared with WT and other mutants (data not shown). Therefore, their inactivation may be due to decreases in structural stability. As mentioned above, the KM PsXR mutant with increased K m for NADPH was co-expressed with PsXDH WT and contributed to enhanced ethanol yield accompanied by decreased xylitol excretion in xylose fermentation. 6,) Our mutation strategy is similar to this previous strategy. However, while the KM PsXR mutant was reported to possess lower activity than WT PsXR, the KR PsXR mutant showed activity similar to WT PsXR. The maintained XR activity of KR PsXR is to be expected to induce high XR efficiency in xylose-fermenting S. cerevisiae. The XR and XDH activities in cell-free extracts prepared from recombinant yeast cells were measured spectrophotometrically, as shown in Fig. A (XDHs were assayed by the method described previously 5) ). Little NADPH-preferring reduction of xylose and NAD þ -dependent oxidization of xylitol was observed in the Y-vector, probably due to endogenous XR (YHR14W) 18) and XDH (YLRC). 19) In all the recombinant yeasts, XDH activities were observed at almost the same levels. In contrast, as compared with the results for purified enzymes (Fig. 1B), NADH-preferring XR activity in the Y-KR strains was much higher than NADPH-preferring activity, even higher than the XR activities of Y-WT, an interesting result, although the reason is not clear. These results suggest that PsXR WT and KR were expressed functionally in S. cerevisiae cells. Therefore, Y-WT and Y-KR were characterized in the fermentation of 15 g l 1 xylose and 5 g l 1 glucose. There was no significant difference between these recombinant yeast strains in the rate of cell growth (data not shown). In both Y-WT and Y-
1368 S. WATANABE et al. A B Y-vector Y-WT Y-KR.16 XR.1.8.5 15 Specific activity (U mg -1 ).4.5..15 XDH Xylitol, Ethanol (g l -1 ) 1.5 1 Glucose, Xylose (g l -1 ) 1 5.1.5.5 15 3 45 6 5 Time (h) Fig.. Effect of PsXR Mutant on Ethanol Fermentation. A, Activities of XR (top) and XDH (bottom) in recombinant yeast strains. Y-vector possessed empty vectors of both ppgk and YEpM4. Darkgray and light-gray bars indicate NADH- and NADPH-preferring XR activities respectively. NAD þ was used as a coenzyme in XDH activity. B, Ethanol fermentation in minimal medium supplemented with glucose (5 g l 1 ) and xylose (15 g l 1 ) by recombinant S. cerevisiae, Y-WT (open symbol) and Y-KR (closed symbol). Glucose (square); xylose (triangle); xylitol (circle); ethanol (rhomboid). Glucose, xylose, and xylitol were measured by HPLC, while ethanol was measured by gas chromatography. Small amounts of glycerol (.5 g l 1 >) and acetic acid were detected, mainly the during glucose consumption phase, by HPLC (data not shown). KR, glucose was initially consumed within 35 h, and of xylose, about 4% was fermented in h. The most positive effect on fermentation by Y-KR was to decrease unfavorable xylitol excretion significantly (1. gl 1 ), which is 4% relative to that by Y-WT (1.45 gl 1 ). On the other hand, xylose consumption and ethanol production were similar to those by Y-WT, although ethanol production by Y-KR was usually slightly higher than Y-WT at every sampling time:.15 gl 1 and.6 g l 1 at h respectively (5.1% increase). These tendencies are very similar to that formed in the KM PsXR mutant, in which xylitol excretion decreased about 3% as compared with that using PsXR WT, while only an approximately 1% increase in ethanol production was observed. 6) There have been few reports on the protein-engineering approach with XR and XDH. Hence, our study focused on the engineering of exotic XYL1 and XYL genes in terms of xylose-fermenting recombinant S. cerevisiae. In addition to the protein/metabolic engineering described above, the introduction and overexpression of other endogenous genes, including xylulokinase (XK), transketolase (TKL1), transaldolase (TAL1), and several hexose-transporter (HXT1-) genes have been attempted to enhance the pentose-phosphate pathway and/or xylose-uptake. ) A combination of these strategies and our strategy using the protein-engineered enzyme might achieve more effective ethanol production from xylose by recombinant S. cerevisiae. Acknowledgment This work was supported by a Grant-in-Aid for Young Scientists (B) (no, 18659 to S.W.), the Center of Excellence (COE) program for the Establishment of COE on Sustainable Energy System, a Grant-in-Aid for Scientific Research, and Grants for Regional Science and Technology Promotion Kyoto Nanotechnology Cluster project, from the Ministry of Education, Culture, Sports, Science and Technology of Japan. This work was also supported by CREST and Research for Promoting Technological Seeds of the Japan Science and Technology Corporation. References 1) Jeffries, T. W., Utilization of xylose by bacteria, yeasts, and fungi. Adv. Biochem. Eng. Biotechnol.,, 1 3
(1983). ) Jeffries, T. W., and Jin, Y. S., Metabolic engineering for improved fermentation of pentoses by yeasts. Appl. Microbiol. Biotechnol., 63, 495 59 (4). 3) Verduyn, C., Van Kleef, R., Frank, J., Schreuder, H., Van Dijken, J. P., and Scheffers, W. A., Properties of the NAD(P)H-dependent xylose reductase from the xylosefermenting yeast Pichia stipitis. Biochem. J., 6, 669 6 (1985). 4) Rizzi, M., Harwart, K., Erlemann, P., Buithanh, N. A., and Dellweg, H., Purification and properties of the NAD þ -xylitol-dehydrogenase from the yeast Pichia stipitis. J. Ferment. Bioeng., 6, 4 (1989). 5) Watanabe, S., Kodaki, T., and Makino, K., Complete reversal of coenzyme specificity of xylitol dehydrogenase and increase of thermostability by the introduction of structural zinc. J. Biol. Chem., 8, 134 1349 (5). 6) Jeppsson, M., Bengtsson, O., Katja, F., Lee, H., Hahn- Hägerdal, B., and Gorwa-Grauslund, M. F., The expression of a Pichia stipitis xylose reductase mutant with higher K M for NADPH increases ethanol production from xylose recombinant Saccharomyces cerevisiae. Biotechnol. Bioeng., 93, 665 63 (6). ) Kostrzynska, M., Sopher, C. R., and Lee, H., Mutational analysis of the role of the conserved lysine- in the Pichia stipitis xylose reductase. FEMS Microbiol. Lett., 159, 1 11 (1998). 8) Lee, J., Koo, B., and Kim, S., Cloning and characterization of the xyl1 gene, encoding an NADH-preferring xylose reductase from Candida parapsilosis, and its functional expression in Candida tropicalis. Appl. Environ. Microbiol., 69, 619 6188 (3). 9) Ko, B. S., Jung, H. C., and Kim, J. H., Molecular cloning and characterization of NAD þ -dependent xylitol dehydrogenase from Candida tropicalis ATCC 913. Biotechnol. Prog.,, 18 114 (6). Protein Engineering of Xylose Reductase 1369 1) Ellis, E. M., Microbial aldo-keto reductases. FEMS Microbiol. Lett., 16, 13 131 (). 11) Petschacher, B., Leitgeb, S., Kavanagh, K. L., Wilson, D. K., and Nidetzky, B., The coenzyme specificity of Candida tenuis xylose reductase (AKRB5) explored by site-directed mutagenesis and X-ray crystallography. Biochem. J., 385, 5 83 (5). 1) Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J., Protein measurement with the folin phenol reagent. J. Biol. Chem., 193, 65 5 (1951). 13) Kang, Y. S., Kane, J., Kurjan, K., Stadel, J. M., and Tipper, D. J., Effects of expression of mammalian G and hybrid mammalian-yeast G proteins on the yeast pheromone response signal transduction pathway. Mol. Cell. Biol., 1, 58 59 (199). 14) Nikawa, J., Sass, P., and Wigler, M., Cloning and characterization of the low-affinity cyclic AMP phosphodiesterase gene of Saccharomyces cerevisiae. Mol. Cell. Biol.,, 369 3636 (198). 15) Gietz, D., St. Jean, A., Woods, R. A., and Schiestl, S. H., Improved method for high efficiency transformation of intact yeast cells. Nucleic Acids Res.,, 145 (199). 16) Kurtzman, C. P., Molecular taxonomy of the yeasts. Yeast, 1, 1 14 (1994). 1) Banta, S., Swanson, B. A., Wu, S., Jarnagin, A., and Anderson, S., Alteration of the specificity of the cofactor-binding pocket of Corynebacterium,5-diketo-D-gluconic acid reductase A. Protein Eng., 15, 131 14 (). 18) Träff, K. L., Jönsson, L. J., and Hahn-Hägerdal, B., Putative xylose and arabinose reductases in Saccharomyces cerevisiae. Yeast, 19, 133 141 (). 19) Richard, P., Toivari, M. H., and Penttilä, M., Evidence that the gene YLRc of Saccharomyces cerevisiae encodes a xylitol dehydrogenase. FEBS Lett., 45, 135 138 (1999).