Saccharomyces Cerevisiae Engineered for Xylose Metabolism

Home , Xylose metabolism

Yeast Yeast 2011; 28: 645–660. Published online 1 August 2011 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/yea.1893 Research Article Saccharomyces cerevisiae engineered for xylose metabolism requires gluconeogenesis and the oxidative branch of the pentose phosphate pathway for aerobic xylose assimilation§

Ronald E. Hector1*, Jeffrey A. Mertens1, Michael J. Bowman1, Nancy N. Nichols1, Michael A. Cotta1 and Stephen R. Hughes2 1Bioenergy Research Unit, National Center for Agricultural Utilization Research, Peoria, IL, USA 2Renewable Product Technology Research Unit, National Center for Agricultural Utilization Research, Peoria, IL, USA

*Correspondence to: Abstract Ronald E. Hector, Bioenergy Research Unit, National Center Saccharomyces strains engineered to ferment xylose using Scheffersomyces stipitis for Agricultural Utilization xylose reductase (XR) and xylitol dehydrogenase (XDH) genes appear to be limited Research, Agricultural Research by metabolic imbalances, due to differing cofactor speciﬁcities of XR and XDH. The Service, US Department of S. stipitis XR, which uses both NADH and NADPH, is hypothesized to reduce the Agriculture†,1815North cofactor imbalance, allowing xylose fermentation in this yeast. However, unadapted University Street, Peoria, IL S. cerevisiae strains expressing this XR grow poorly on xylose, suggesting that 61604, USA. metabolism is still imbalanced, even under aerobic conditions. In this study, we E-mail: investigated the possible reasons for this imbalance by deleting genes required for [email protected] NADPH production and gluconeogenesis in S. cerevisiae. S. cerevisiae cells expressing †Mention of trade names or the XR–XDH, but not a xylose isomerase, pathway required the oxidative branch of commercial products in this the pentose phosphate pathway (PPP) and gluconeogenic production of glucose-6-P article is solely for the purpose of for xylose assimilation. The requirement for generating glucose-6-P from xylose was providing scientiﬁc information also shown for Kluyveromyces lactis. When grown in xylose medium, both K. lactis and does not imply and S. stipitis showed increases in enzyme activity required for producing glucose-6- recommendation or endorsement P. Thus, natural xylose-assimilating yeast respond to xylose, in part, by upregulating by the US Department of enzymes required for recycling xylose back to glucose-6-P for the production of Agriculture. USDA is an equal NADPH via the oxidative branch of the PPP. Finally, we show that induction of opportunity provider and employer. these enzymes correlated with increased tolerance to the NADPH-depleting compound diamide and the fermentation inhibitors furfural and hydroxymethyl furfural; §This article is a US Government S. cerevisiae was not able to increase enzyme activity for glucose-6-P production when work and is in the public domain grown in xylose medium and was more sensitive to these inhibitors in xylose medium in the USA. compared to glucose. Published in 2011 by John Wiley & Sons, Ltd. Keywords: xylose; fermentation; Saccharomyces; gluconeogenesis; NADPH; Received: 16 March 2011 imbalance Accepted: 14 June 2011

Introduction ﬁve-carbon sugar, is the second most abundant sugar in biomass hydrolysates; only glucose is Development of economical processes for con- present at a higher concentration. Although C6 and verting lignocellulosic feedstocks to renewable C5 sugars could be potentially split into separate products will require efﬁcient utilization of all carbon streams, many of the proposed processes available sugars (Sassner et al., 2008). Xylose, a for production of biofuels and other renewable

Published in 2011 by John Wiley & Sons, Ltd. 646 R. E. Hector et al. products require the combined use of xylose and phosphate pathway (PPP) intermediate xylulose- glucose. Saccharomyces cerevisiae would be the 5-phosphate (Figure 1). Most xylose reductases preferred organism for biofuels production, due to identified show a strict dependence on NADPH as the availability of a robust genetic transformation the cofactor, while xylitol dehydrogenase is specific system along with a long history of use in for NAD+. This imbalance of cofactors can lead to industrial fermentation processes. Unfortunately, depletion of NADPH and excess NADH (Van Vleet xylose metabolism presents a unique challenge for and Jeffries, 2009). Aerobically, excess NADH S. cerevisiae. can be re-oxidized by the mitochondria. Under Most natural xylose-assimilating yeasts, such as Scheffersomyces stipitis (formerly known as Pichia anaerobic conditions, excess NADH accumulates stipitis; Kurtzman and Suzuki, 2010) metabolize and xylose utilization slows, or in some cases xylose by a reduction/oxidation pathway. Xylose is stops (Bruinenberg et al., 1984). NADPH must be first converted to xylitol using a xylose reductase regenerated through metabolic routes. When glu- (XR). Xylitol dehydrogenase (XDH) converts xyl- cose is available, most of the NADPH required itol to xylulose, which can then be phosphorylated for anabolic reductive reactions and tolerance to by xylulokinase (XK) to generate the pentose inhibitors is generated through the oxidative branch

A Oxidative PPP Glucose 2 XYLOSE ATP NADPH (XR) XYL1 NADPH + CO NADP+ NADPH NADP+ NADP+ 6-phospho- 6-phospho- Ribulose-5-P xylA Glucose-6-P gluconate Xylitol ZWF1 glucono SOL3 GND1 (XI) (G6PD) lactone SOL4 GND2 PGI1 RKI1 RPE1 NAD+ (XDH) XYL2 ATP NADH Fructose-6-P ATP Fructose-6-P non-oxidative Xylulose Pentose FBP1 PFK1 XKS1 PFK2 Glyceraldehyde-3-P Phosphate (XK) Pathway Fructose-1,6-BP

FBA1 aromatic Histidine and Dihydroxy Glyceraldehyde-3-P acetone-P TPI1 amino nucleic acids acids

glycerol

ETHANOL

B NADP+ NADPH NADP+ NADPH Acetaldehyde Acetate Isocitrate α-ketoglutarate ALD6 IDP2

Figure 1. (A) Representation of glucose and xylose metabolism via glycolysis, pentose phosphate pathway and gluconeogenesis. Abbreviations/gene names and associated pathways are as follows. Glycolysis: PGI1, phosphoglucose isomerase; PFK1 and PFK2, phosphofructokinase; FBA1, fructose 1,6-bisphosphate aldolase; TPI1, triose phosphate isomerase. Oxidative PPP: ZWF1, glucose-6-phosphate dehydrogenase (G6PD); SOL3 and SOL4, 6-phosphogluconolactonase; GND1 and GND2, 6-phosphogluconate dehydrogenase. Non-oxidative PPP: RKI1, ribose-5-phosphate ketol-isomerase; RPE1, D-ribulose-5-phosphate 3-epimerase; TKL1, transketolase; TAL1,transaldolase.Initial xylose metabolism: xylA,xylose isomerase (XI); XYL1, xylose reductase (XR); XYL2, xylitol dehydrogenase (XDH); XKS1,xylulokinase(XK).Gluconeogenesis: PGI1, phosphoglucose isomerase; FBP1, fructose 1,6-bisphosphate phosphotase; FBA1, fructose 1,6-bisphosphate aldolase. (B) Minor sources of NADPH in S. cerevisiae. ALD6, cytosolic NADP+-dependent aldehyde dehydrogenase; IDP2, cytosolic NADP+-dependent isocitrate dehydrogenase

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea S. cerevisiae requires the PG11 and ZWF1 genes for xylose assimilation 647 of the PPP (Lagunas and Gancedo, 1973; Nogae for genes of the oxidative branch of the PPP during and Johnston, 1990). anaerobic growth on xylose, suggesting that regula- S. stipitis appears to be unique among xylose- tion of these pathways may be important for xylose assimilating yeasts as under oxygen-limited con- metabolism (Runquist et al., 2009). This idea was ditions, it can ferment xylose to produce ethanol. further enhanced when it was found that deletion The ability of S. stipitis to ferment xylose has been of an activating transcription factor for PPP genes, attributed to the ability of its XR to use NADH Stb5p,inS. cerevisiae led to impaired NADPH almost as well as NADPH, thus avoiding an imbal- production (Cadiere` et al., 2010). ance of cofactors (Bruinenberg et al., 1984). The To gain further insight into the requirement S. stipitis XR–XDH pathway has been expressed of these pathways for xylose metabolism in in S. cerevisiae and confers the ability to metabo- S. cerevisiae, we deleted genes for recycling lize xylose (Figure 1; reviewed in Hahn-Hagerdal¨ fructose-6-phosphate (fructose-6-P) back to et al., 2007). Natural xylose-assimilating yeasts do glucose-6-phosphate (glucose-6-P) and through the not exhibit (under aerobic conditions) a cofac- oxidative branch of the PPP. We also deleted tor imbalance and grow well on xylose medium. genes for alternative sources of NADPH and anal- S. stipitis can grow in synthetic xylose medium ysed each strain’s ability to assimilate xylose. with an aerobic growth rate of µ = 0.45 h−1 and We show that S. cerevisiae engineered for xylose quickly reach optical densities >12 OD660. While metabolism via an XR–XDH pathway, but not a some further adapted and modified Saccharomyces xylose isomerase (XI) pathway, requires the pro- yeasts with the XR–XDH pathway show signif- duction of glucose-6-P from fructose-6-P for the icant increases in growth on xylose, unadapted generation of NADPH by the oxidative branch S. cerevisiae strains expressing this pathway grow of the PPP. Requirement for the oxidative branch poorly on xylose, indicating that even under aer- of the PPP for xylose metabolism is not unique obic conditions, metabolism is not properly bal- to S. cerevisiae. Both of the xylose-assimilating anced. For example, the unadapted S. cerevisiae yeasts Scheffersomyces stipitis and Kluyve- laboratory strain CEN.PK2-1C, expressing the romyces lactis induce glucose-6-P dehydrogenase S. stipitis XR–XDH pathway, has a much slower and phosphoglucose isomerase activity when grown maximum aerobic growth rate (µ = 0.08 h−1)on on xylose as the sole carbon source, whereas xylose medium and rarely gets above 3 OD660 S. cerevisiae does not. Additionally, K. lactis is (Hector et al., 2010). Other studies with CEN.PK- shown to depend on the oxidative branch of the derived strains with integrated XR–XDH genes PPP for growth on xylose. had even slower growth rates (µ = 0.007 h−1) (Karhumaa et al., 2007). Since excess NADH can be reoxidized to restore the NAD+: NADH ratio Materials and methods under aerobic conditions, slow growth is likely due to NADPH limitation, as well as inefficient flux Strains, media, and general methods of metabolites through the non-oxidative branch of Escherichia coli strains DH10B, TOP10 (Invitro- the PPP. gen, Carlsbad, CA, USA) and NEB10-beta [New Theoretical studies suggest that NADPH require- England Biolabs (NEB), Ipswich, MA, USA] were ments for xylose-assimilating yeasts are mostly met used for routine maintenance and preparation of via the oxidative branch of the PPP for xylose plasmids and were grown in LB medium (Sam- metabolism (Bruinenberg et al., 1983b). How- brook and Russell, 2001). Strains and plasmids ever, additional mechanisms of NADPH regen- used in this study are listed in Table 1. DNA eration may exist in these yeasts that do not was transformed into yeast cells using a stan- exist in S. cerevisiae, either by induction of genes dard lithium acetate method (Gietz and Woods, with unique properties (Verho et al., 2002) or 2002). Synthetic complete (SC) medium consisted by the ability to regulate genes common to of 6.7 g/l Difco yeast nitrogen base (YNB; United S. cerevisiae in a specific response to xylose (Jef- States Biological; Marblehead, MA, USA) and fries and Van Vleet, 2009). A recent study using a was supplemented with amino acids as necessary xylose-fermenting S. cerevisiae strain observed an (Amberg et al., 2005). Sterile glucose or xylose increase in mRNAs for gluconeogenic genes and was added separately.

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea 648 R. E. Hector et al.

Table 1. Strains and plasmids used in this study

Strain Genotype Reference

Y-7124 Scheffersomyces stipitis ARSa CEN.PK2-1C MATa ura3-52 trp1-289 leu2-3,112 his3∆1MAL2-8C SUC2 EUROSCARF YRH501 CEN.PK2-1C [pRH315]b This study YRH528 CEN.PK2-1C zwf1∆::LEU2 [pRH315] This study YRH530 CEN.PK2-1C fbp1∆::LEU2 [pRH315] This study YRH532 CEN.PK2-1C ald6∆::LEU2 [pRH315] This study YRH534 CEN.PK2-1C idp2∆::LEU2 [pRH315] This study YRH536 CEN.PK2-1C stb5∆::URA3 [pRH315] This study YRH562 CEN.PK2-1C [pRH195, pRH218] This study YRH624 CEN.PK2-1C stb5∆::URA3 [pRH315, pRH377, pRH378] This study YRH634 CEN.PK2-1C zwf1∆::LEU2 [pRH195, pRH218] This study YRH741 CEN.PK2-1C pgi1∆::LEU2 [pRH315] This study YRH742 CEN.PK2-1C pgi1∆::LEU2 [pRH195, pRH218] This study YRH834 CEN.PK2-1C zwf1∆::LEU2 gre3∆::HIS3 [pRH315] This study MW179-1D Kluyveromyces lactis MATα metA1 ade− leu2 trp ura3 (Saliola, et al., 2007) Klzwf1 MW179-1D Klzwf1∆::KanMX4 (Saliola, et al., 2007) JW1841-1 E. coli F-, ∆(araD-araB)567, ∆lacZ4787(::rrnB-3), &lambda−, (Baba et al., 2006) ∆zwf-777::kan, rph-1, ∆(rhaD-rhaB)568, hsdR514 CGSC3121 E. coli C-1 F- (Bertani and Weigle, 1953)

Plasmid Description pRS403 pBluescript II SK+, HIS3 (Christianson, et al., 1992) pRS404 pBluescript II SK+, TRP1 (Christianson, et al., 1992) pRS413 pBluescript II SK+, HIS3, CEN6, ARSH4 (Christianson, et al., 1992) pRS414 pBluescript II SK+, TRP1, CEN6, ARSH4 (Christianson, et al., 1992) pRS415 pBluescript II SK+, LEU2, CEN6, ARSH4 (Christianson, et al., 1992) pRS416 pBluescript II SK+, URA3, CEN6, ARSH4 (Christianson, et al., 1992) pRS426 pBluescript II SK+, URA3, 2 µ origin (Christianson, et al., 1992) pRH143 pRS415 + PPDC1-MCS-TADH1 This study pRH153 pRS413 + PHXT7-MCS-THXT7 This study pRH167 pRS426 + PHXT7-MCS-THXT7 (Hector, et al., 2010) pRH193 pCR2.1-TOPO + Piromyces sp. E2 ORF This study pRH195 pRS414 + PHXT7-XKS1-THXT7 (Hector, et al., 2010) pRH211 pRS416 + PPGK1-XYL1-TPGK1;PADH1-XYL2-TADH1 (Hector, et al., 2010) pRH218 pRS426 + PHXT7-Piromyces XI-THXT7 This study pRH274 pRS416 + PPGK1-XYL1-TPGK1;PADH1-XYL2-TADH1;PHXT7-XKS1-THXT7 This study pRH315 pRS414 + PPGK1-XYL1-TPGK1;PADH1-XYL2-TADH1;PHXT7-XKS1-THXT7 This study pRH375 pCR2.1 + ZWF1 ORF This study pRH376 pCR2.1 + GND1 ORF This study pRH377 pRS413 + PHXT7-ZWF1-THXT7 This study pRH378 pRS415 + PPDC1-GND1-TADH1 This study pSUMOduo/URA/HisXI Assembled Piromyces sp. E2 xylose isomerase (+His-tag) in pSUMOduo (Hughes, et al., 2009) a ARS Culture collection at the National Center for Agricultural Utilization Research. b Data in brackets indicate plasmids contained in the strain.

Construction of Saccharomyces cerevisiae gene. DNA oligonucleotides used in this study deletion strains are listed in Table 2. The PGI1 gene was deleted Gene deletions were constructed by gene replace- using primers 133 and 134 to amplify the LEU2 ment with selectable markers (Christianson et al., gene and transformed into S. cerevisiae to replace 1992). Deletion of the STB5 and ZWF1 genes the PGI1 gene. Strains lacking the PGI1 gene was previously described (Hector et al., 2009). were confirmed by PCR amplification from whole Each gene deletion was confirmed by PCR ampli- cells, using primers 135/36 and 35/136. Primers fication of DNA fragments flanking the disrupted 36 and 35 anneal within the LEU2 gene, used

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea S. cerevisiae requires the PG11 and ZWF1 genes for xylose assimilation 649

Table 2. DNA oligonucleotides used in this study

No. Sequence

13 5-GGAAAATACTGTAATATAAATCGTAAAGGAAAATTGGAAATTTTTTAAAGAGATTGTACTGAGAGTGCAC-3 14 5-TGAGGGAAATTTGTTCATATCGTCGTTGAGTATGGATTTTACTGGCTGGACTGTGCGGTATTTCACACCG-3 15 5-GGGGGCCTATCAAGTAAATT-3 16 5-ATGACACCACAGGCAGAAAA-3 17 5-GGAGCTC AACCAGGACGTAAAGGGTA-3 18 5-GGACTAGTTTGATTGATTTGACTGTGTT-3 19 5-TAGGAGCTC TTTCGGGCCCCTGC-3 20 5-AGCGTCTTGTGACTAGTTTTGATT-3 21 5-GGTCGAC GCGAACACTTTTATTAATTC-3 22 5-GCTCGAGTATTTGTGAATAACAGTGCGGTC-3 29 5-GGTCGAC TAAATAAGCGAATTTCTTATGAT-3 30 5-GCTCGAGCGACCTCATGCTATACCTGA-3 35 5-CCAACGTGGTCACCTGGCAA-3 36 5-GTACCACCGAAGTCGGTGAT-3 133 5-ATTCCTCTAGTCTTGCAAAATCGATTTAGAATCAAGATACCAGCCTAAAAAGATTGTACTGAGAGTGCAC-3 134 5-TTTGCTTATAATATAGCTTTAATGTTCTTTAGGTATATATTTAAGAGCGACTGTGCGGTATTTCACACCG-3 135 5-CAGTGAATTTTAATACATATTCCT-3 136 5-CTTGGACGCTGTTCAATAATGT-3 137 5-GACTAGTCAT ATGGCTAAGGAATATTTCCCA-3 138 5-GGTCGAC TTATTGGTACATGGCAACAA-3 264 5-AAGAATAACAGTGCGAACATATAAGAAACATCCCTCATACTACCACACATAGATTGTACTGAGAGTGCAC-3 265 5-CGTACTAAAGTACAGAACAAAGAAAATAAGAAAAGAAGGCGATCATTCTGTGCGGTATTTCACACCG-3 266 5-GGATCGTCCTATGTATAGGCA-3 267 5-ACTTCCATCCCATTCCATTC-3 272 5-AACATCAAGAAACATCTTTAACATACACAAACACATACTATCAGAATACAAGATTGTACTGAGAGTGCAC-3 273 5-TATATGAAAGTATTTTGTGTATATGACGGAAAGAAATGCAGGTTGGTACACTGTGCGGTATTTCACACCG-3 274 5-CACATCAAAACACCGTTCGA-3 275 5-CAAGCCTGTTCTCTCTTTTTC-3 276 5-CACGAGAATAGGAGGTAAGAAGGTAACGTACGTATATATATAAAATCGTAAGATTGTACTGAGAGTGCAC-3 277 5-ATAAAAAGGGAATATATAATATAAATAAATCAATCTGTCATTGAGTAAGTCTGTGCGGTATTTCACACCG-3 278 5-CAAAAACCGGACTTAACATTATTGG-3 279 5-CGAAGAGGGAAAAGAAGAAAGAAAA-3 357 5-GGACTAGTAAGATGAGTGAAGGCC-3 358 5-GCGTCGAC TCTAATTATCCTTCGTATC-3 359 5-GGACTAGTAAATGTCTGCTGATTTCG-3 360 5-GCGGATCC TCCTTTAAGCTTGGTATG-3 pRS 5-AGATTGTACTGAGAGTGCAC-3 universal S pRS 5-CTGTGCGGTATTTCACACCG-3 universal AS

Restriction endonuclease sites are shown italicized and underlined. Start codons are shown in bold. to replace PGI1, while primers 135 and 136 following primer combinations were used to con- anneal to sequences ﬂanking the PGI1 gene. Addi- ﬁrm the deletions: GRE3 (15/16); FBP1 (266/36, tionally, the S. cerevisiae PGI1 deletion strains 267/35); ALD6 (274/36, 275/35); IDP2 (278/36, were not able to grow on glucose medium, as 279/35). previously reported (Dickinson et al., 1995). SC + 2% fructose/0.1% glucose was used to main- Generation of plasmids for expressing tain pgi1∆ cells. The GRE3, FBP1, ALD6 and xylose-utilization genes in S. cerevisiae IDP2 genes were deleted in a similar manner, using primer pairs 13/14, 264/265, 272/273 and 276/277, Plasmid pRH315 for expressing the S. stipitis respectively, to amplify the marker gene. The XYL1 and XYL2 genes, and the S. cerevisiae

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea 650 R. E. Hector et al.

XKS1 gene, was created as follows. The BssHII Enzyme assays fragment from pRH195, containing the PHXT7- Clarified cell lysates were prepared from mid-log XKS1-THXT 7 fragment, was blunted with T4 DNA polymerase (NEB) and cloned into the blunt and phase cells grown in synthetic complete medium dephosphorylated XhoI site of pRH211 (Hec- using either glucose or xylose as the carbon tor et al., 2010), resulting in vector, pRH274. source. Cells were collected by centrifugation, To convert the URA3 marker to TRP1, pRH274 washed once with ice-cold sterile water and cen- was digested with NsiI (which cuts once near trifuged again to pellet the cells. Cell pellets − ◦ the URA3 ORF). The NsiI-cut pRH274 was co- were stored at 80 C for later use. Cells were transformed into CEN.PK2-1C with excess PCR- resuspended in an appropriate amount of Y-PER amplified TRP1 from pRS404. The TRP1 marker reagent (Pierce, Rockford, IL, USA) plus protease was PCR-amplified, using primers pRS, universal inhibitors (Complete, mini, EDTA-free protease S and AS. Recombination-based repair of NsiI- inhibitor cocktail, Roche, Indianapolis, IN, USA) cut pRH274 resulted in replacement of the URA3 and processed according to the manufacturer’s marker with the TRP1 marker, creating pRH315. instructions. Protein concentrations were deter- Plasmid pRH315 was rescued from yeast into mined with the Quick Start Bradford Protein Assay NEB10-beta cells. (Bio-Rad; Hercules, CA, USA) against a bovine Plasmid pRH218 for expressing the Piromyces serum albumin standard. sp. E2 xylose isomerase (XI ) gene was cre- The lack of glucose-6-phosphate dehydrogenase ated by PCR amplification of the XI ORF from (G6PD) activity was confirmed for all strains in pSUMOduo/URA/HisXI (Hughes et al., 2009). this study that were deleted for the ZWF1 gene. Primers 137/138 were used to generate the XI G6PD activity was assayed in buffer containing 33 mM sodium phosphate, pH 8.0, 10 mM MgCl2, fragment (without the His-tag). The XI ORF + was cloned into pCR2.1-TOPO (Invitrogen) and 1mM NADP and an appropriate amount of cell sequenced, creating pRH193. Plasmid pRH193 was lysate. Reactions were started by the addition of digested with SpeI and SalI sites that were included glucose-6-P (Sigma, St. Louis, MO, USA) to a in the oligonucleotide sequences for amplification. final concentration of 10 mM and reactions were The fragment containing the XI was isolated and monitored at 340 nm using a Cary 50 Bio UV- cloned into the SpeI/SalI sites of pRH167 (pRS426 Visible spectrophotometer (Varian, Palo Alto, CA, USA). Specific activity (µmole/min/mg lysate) was + PHXT 7-MCS-THXT 7) to create pRH218 (pRS426 determined using the molar absorption coefficient, + PHXT 7-Piromyces sp. E2 XI-THXT 7). To create plasmids for expressing ZWF1 and ε340,of6.22mM/cm for NAD(P)H. Specific activ- GND1, the genes were first PCR-amplified using ities reported were proportional to the amount of Phusion Hot Start II High-Fidelity DNA Poly- lysate added in a dilution series. Each assay was merase (Finnzymes, Vantaa, Finland) from done using lysates prepared from three independent S. cerevisiae genomic DNA, using primers 357/358 cultures. Phosphoglucose isomerase (PGI) activ- and 359/360, respectively. Amplified DNA frag- ity was determined using a modified G6PD assay ments were first cloned into pCR2.1 (Invitrogen) in which 10 mM fructose-6-P was used instead of and sequenced, creating pRH375 and pRH376. 10 mM glucose-6-P, and 0.5 U recombinant G6PD These plasmids were digested with SpeI and (Sigma) was incorporated. The absence of phos- SalI sites that were included in the oligonu- phoglucose isomerase (PGI) activity was confirmed cleotide sequences for amplification. The frag- for all PGI1 deletions used in this study (data not ments containing the ZWF1 and GND1 genes shown). were isolated and cloned into the SpeI/SalI sites + of pRH153 (pRS413 PHXT 7-MCS-THXT 7) and Aerobic growth kinetics pRH143 (pRS415 + PPDC 1-MCS-TADH 1), creating pRH377 and pRH378. pRH153 and pRH143 were Yeast pre-cultures were grown to mid-log phase made by subcloning sequence-verified promoters in synthetic complete medium with 20 g/l glucose and terminators for PHXT 7 (19/20), THXT 7 (21/22), and washed with sterile water prior to inocula- PPDC 1 (17/18), TADH 1 (29/30), using the primers tion. Synthetic complete medium + 50 g/l xylose listed in parentheses, into pRS413 and pRS415. (SC5X) was used to determine each yeast strain’s

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea S. cerevisiae requires the PG11 and ZWF1 genes for xylose assimilation 651 ability to assimilate xylose aerobically. SC5X cul- (Cadiere` et al., 2010; Hector et al., 2009; tures (25 ml in a 125 ml Erlenmeyer flask) were Larochelle et al., 2006). Here we show that the ◦ started at OD660 0.05 and incubated at 30 C, transcription factor Stb5p, while not essential for shaking at 200 rpm. E. coli strains were grown growth on glucose, is required for growth on at 37 ◦C in defined mineral medium (Nichols and xylose medium (Figure 2A). Stb5p regulates genes Mertens, 2008), using either glucose or xylose at for alternative sources of NADPH as well as 20 g/l as the carbon source. The data shown rep- non-oxidative PPP genes (Cadiere` et al., 2010; resent the average values from experiments that Larochelle et al., 2006), the latter of which are were repeated in triplicate using independent cul- known to influence xylose metabolism. To deter- tures. Standard deviations (SDs) were <10% of the mine whether STB5-deleted cells were impaired for mean. growth on xylose medium simply due to NADPH deficiency resulting from failure to express Inhibitor tolerance growth kinetics NADPH-producing genes of the oxidative branch of the PPP, we overexpressed these genes in a Log phase cultures grown in synthetic complete strain, YRH536, lacking the STB5 gene (stb5∆), medium with either 20 g/l glucose or 50 g/l xylose attempting to restore this part of the pathway. were washed once with sterile water and resus- The two NADPH-generating enzymes of the oxida- = × pended to final OD660 0.2into2 concen- tive branch of the PPP are glucose-6-phosphate trated medium. To initiate stress conditions, 100 µl dehydrogenase (G6PD, encoded by ZWF1 ) and diluted cells were added to 96-well microplates 6-phosphogluconate dehydrogenase (the major iso- containing 100 µl sterile water containing vary- form encoded by GND1 ) (Figure 1). Overexpres- ing amounts of either diamide (Sigma) or fur- sion of the ZWF1 and GND1 genes in the fural (Fischer Scientific, Pittsburg, PA, USA), STB5-deleted strain did not restore growth on resulting in a starting OD660 of 0.1 for each ◦ xylose (data not shown), indicating that dele- well. The microplates were incubated at 30 C tion of the STB5 gene likely affects both the for 48 h, shaking at 200 rpm. Prior to reading oxidative and non-oxidative branches of the PPP. the OD660 in a Benchmark Plus microplate spec- These results showing that the transcription factor trophotometer (Bio-Rad), the microplates were Stb5p is essential for growth on xylose medium mixed to thoroughly resuspend any cells that may indicate that STB5 may be an excellent can- have settled. All strains and each condition tested didate for global transcription machinery engineer- were performed in triplicate and mean values are ing (gTME) (Alper and Stephanopoulos, 2007) for reported. improving xylose metabolism. However, they did not provide additional information about NADPH Statistical analyses regeneration mechanisms during growth on xylose. For experiments with three or greater biological replicates, probability analyses were performed S. cerevisiae cells expressing the S. stipitis using Student’s t-test with a two-tailed distribution XR–XDH pathway rely on NADPH generated and compared to the appropriate control strain. via the PPP to assimilate xylose p < 0.05 was considered significant for this study. The ZWF1 gene is the entry point to the oxida- Statistical analysis was performed using Microsoft tive branch of the PPP, which is considered the Excel. main source of NADPH during growth on glucose. ZWF1 codes for the enzyme glucose-6-P dehydrogenase (G6PD). We observed that G6PD activity Results and discussion is essential for xylose assimilation under aerobic conditions (Figure 2A). During aerobic conditions, The transcription factor Stb5p is essential + excess NADH can be re-oxidized to NAD by for xylose assimilation mitochondria, using oxygen as the terminal elec- It was previously shown that the transcription fac- tron acceptor. Unlike NAD+, NADPH must be tor Stb5p modulates NAPDH availability and reg- supplied via central metabolic routes, regardless of ulates pentose phosphate pathway (PPP) enzymes growth conditions. To allow this study to focus

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea 652 R. E. Hector et al.

A 4 under aerobic conditions (considered optimal for S. cerevisiae WT (XR-XDH pathway) xylose assimilation), NADPH production through zwf1∆ the oxidative branch of the PPP is required in engi- 3 ) stb5∆ neered S. cerevisiae.

660 S. cerevisiae expresses an endogenous aldo-keto reductase, Gre3p, which has xylose reductase activ- 2 ity (Toivari et al., 2004) and deletion of GRE3 has been shown to reduce xylitol production in recom-

Biomass (OD binant S. cerevisiae (Traff¨ et al., 2002). Unlike the 1 S. stipitis XR, Gre3p is speciﬁc for NADPH and does not use NADH (Kuhn et al., 1995). Thus, 0 Gre3p would be expected to consume NADPH 0 24 48 72 96 120 144 168 during the conversion of xylose to xylitol, pos- Time (hours) sibly leading to the strain’s inability to grow in xylose medium when ZWF1 is deleted. To deter- B 4 mine whether endogenous NADPH-dependent XR activity from Gre3p resulted in the ZWF1-deleted 3 strains’ inability to assimilate xylose, a strain was ) constructed that had both GRE3 and ZWF1 genes 660 deleted. This strain (gre3∆,zwf1∆) was also unable 2 to grow in xylose medium (data not shown), indicating that the inability of the zwf1∆ strain to WT assimilate xylose was associated with S. stipitis XR

Biomass (OD ∆ 1 ald6 activity and not a result of endogenous NADPH- idp2∆ dependent XR activity. To determine whether other potential sources 0 contribute to the NADPH supply during growth on 0 24 48 72 96 120 144 168 xylose medium, we deleted the ALD6 and IDP2 Time (hours) genes and determined whether growth on xylose + Figure 2. Glucose-6-P dehydrogenase is essential for was affected (Figure 1B). Ald6p is a NADP - aerobic growth on xylose. Growth curves for S. cerevisiae specific aldehyde dehydrogenase shown to supply strains engineered to metabolize xylose via expression of NADPH in ZWF1-deleted cells grown on glucose the S. stipitis xylose reductase and xylitol dehydrogenase (Grabowska and Chelstowska, 2003). Ald6p was genes and the S. cerevisiae xylulokinase gene. (A) Parent (WT) strain and strains lacking the gene for glucose-6-P also demonstrated to increase in cells grown in dehydrogenase (zwf1∆) or lacking the transcription factor xylose medium, suggesting that it may also pro- Stb5p that regulates genes involved in NADPH regeneration vide NADPH during growth on xylose (Salusjarvi¨ (stb5∆). (B) Parent (WT) strain and strains lacking genes for et al., 2003). Idp2p is a cytosolic NADP+-specific alternative sources of NAPDH. ald6∆ cells lack the ALD6 isocitrate dehydrogenase that catalyses oxidation of gene encoding an NADP+-specific aldehyde dehydrogenase. idp2∆ cells lack the IDP2 gene encoding a cytosolic isocitrate to α-ketoglutarate. Idp2 levels are ele- NADP+-specific isocitrate dehydrogenase. For both (A) and vated during growth on non-fermentable carbon (B), cells were cultured in synthetic complete medium (SC) sources and reduced during growth on glucose plus 50 g/l xylose. Growth curves shown represent the (Minard and McAlister-Henn, 2005). Although averages from experiments performed in triplicate these enzymes are considered minor sources of NADPH with glucose as the carbon source, their on the effect of NADPH availability on xylose contribution during xylose metabolism was metabolism, aerobic conditions were used through- unknown. out. Poor aerobic xylose assimilation of the ZWF1- Strains deleted for either the IDP2 or ALD6 deleted strain (zwf1∆) is consistent with reports genes were impaired, with a significant lag in showing that modulating G6PD activity affects growth on xylose medium (Figure 2B). This result xylose fermentation (Jeppsson et al., 2003; Verho suggests that alternative sources of NADPH can et al., 2003). It further demonstrates that, even contribute NADPH for growth on xylose. Since

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea S. cerevisiae requires the PG11 and ZWF1 genes for xylose assimilation 653

Zwf1p, Ald6p and Idp2p are not essential for 4 S. cerevisiae growth on glucose, the specific requirement for WT (XR-XDH pathway) these genes for growth on xylose suggests that fbp1∆ 3 NADPH regeneration is insufficient in recombinant ) pgi1∆ strains expressing the XR and XDH genes for 660 xylose metabolism. These data also confirm that the main source of NADPH in engineered S. cerevisiae 2 cells growing on xylose is the oxidative branch of

the PPP and indicate that, although other sources do Biomass (OD contribute NADPH, they do not provide sufficient 1 levels of NADPH to support growth on xylose medium. 0 The fact that ZWF1 is required for xylose 0 24 48 72 96 120 144 168 metabolism for strains engineered with the XR– Time (hours) XDH pathway demonstrates that recombinant S. cerevisiae must make glucose-6-P, the substrate Figure 3. Contribution of gluconeogenesis genes PGI1 for Zwf1p, in order to meet NADPH demands. and FBP1 to aerobic growth on xylose. Growth curves for S. cerevisiae strains engineered to metabolize xylose via It further suggests that the use of NADPH by expression of the S. stipitis xylose reductase and xylitol the S. stipitis XR is the main contributing factor dehydrogenase genes and the S. cerevisiae xylulokinase to the NADPH deficiency. Recently, a microar- gene. Comparison of the parent (WT) strain and strains ray study of S. cerevisiae engineered with the lacking the genes required for gluconeogenic production XR–XDH pathway, grown under anaerobic condi- of glucose-6-P. fbp1∆ cells lack the FBP1 gene encoding fructose 1,6-bisphosphotase. pgi1∆ cells lack the PGI1 tions on xylose medium, reported an increase in the gene for phosphoglucose isomerase. Cells were cultured in level of mRNAs for genes involved in the oxida- synthetic complete medium (SC) plus 50 g/l xylose. Growth tive PPP and gluconeogenesis (Runquist, et al., curves shown represent the averages from experiments 2009). To further explore the need for gluconeo- performed in triplicate genesis during growth on xylose, we determined whether strains lacking the PGI1 gene could grow that the main source of NADPH for recombinant on xylose (Figure 3). PGI1 codes for phospho- S. cerevisiae is through production of glucose-6-P glucose isomerase; during gluconeogenesis Pgi1p for use in the oxidative branch of the PPP. Thus, converts fructose-6-P to glucose-6-P for glucan gluconeogenesis is essential for xylose metabolism synthesis for the cell wall and NADPH produc- in engineered S. cerevisiae. tion via the oxidative branch of the PPP. We also tested a strain in which the gluconeogenesis The requirement for gluconeogenesis gene FBP1 was deleted. Fbp1p is one step down for xylose assimilation is not unique from Pgi1p and catalyses the dephosphorylation to S. cerevisiae of fructose-1,6-bisphosphate (F1,6BP) to fructose- 6-P (F6P). In the strain lacking the FBP1 gene We next asked whether the oxidative branch (fbp1∆ strain), glyceraldehyde-3-P (G3P) from the of the PPP is required for growth on xylose PPP cannot be used for gluconeogenesis. F6P medium in natural xylose-assimilating organisms, from xylose metabolism through the non-oxidative such as K. lactis or E. coli. It is possible that branch of the PPP can still be used to make natural xylose-assimilating organisms have addi- glucose-6-P (see Figure 1). Accordingly, cells lack- tional sources of NADPH regeneration and do ing Pgi1p could not grow on xylose, while cells not entirely depend on the oxidative branch of lacking Fbp1p could grow, but at a reduced rate the PPP for NADPH on xylose medium, like (Figure 3). The fact that cells lacking FBP1 were recombinant XR–XDH expressing S. cerevisiae negatively affected on xylose medium suggests that strains. For example, K. lactis has been shown G3P from xylose metabolism is also converted to to induce expression of an NADP+-dependent glucose-6-P. glyceraldehyde-3-P dehydrogenase (NADP- The sum of these data indicates that the XR– GAPDH) when grown on xylose, which could con- XDH pathway results in an NADPH deficiency and tribute to the NADPH supply on xylose medium

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea 654 R. E. Hector et al.

(Verho, et al., 2002). This enzyme could play a key A 4 K. lactis role in xylose metabolism for K. lactis and reduce WT its dependency on recycling carbon back through zwf1∆ the oxidative branch of the PPP. ) 3 K. lactis has been shown to naturally metab- 660 olize xylose via a XR–XDH pathway, using an 2 NADPH-speciﬁc XR (Billard et al., 1995). Dele- tion of the KlZWF1 gene has been studied previously in K. lactis and strains deleted for the Biomass (OD 1 KlZWF1 gene exhibit the expected increased sensitivity to H2O2, but also showed decreased growth 0 on fermentative carbon sources such as glucose 0 24487296120 and fructose (Saliola et al., 2007). We tested the Time (hours) KlZWF1 deletion strain for its ability to assimilate xylose. Deletion of the KlZWF1 gene completely B 3.0 E. coli inhibited growth on xylose medium (Figure 4A), WT indicating that in K. lactis the majority of NADPH 2.5 zwf1∆

for xylose assimilation is provided via the oxida- ) 2.0 tive branch of the PPP. This result also sug- 660 gests that the xylose-inducible NADP–GAPDH in K. lactis does not provide enough NADPH 1.5 for cell growth in the absence of the KlZWF1 1.0 gene. Biomass (OD One possible explanation for the dependency 0.5 on G6PD activity of both the engineered Saccha- romyces and native xylose-metabolizing yeasts is 0.0 that the ﬁrst enzyme of the pathway used, xylose 0 12243648 reductase, consumes NADPH, leading to a deﬁ- Time (hours) ciency (Bruinenberg et al., 1983a). In compari- Figure 4. Contribution of the oxidative branch of the PPP son, most bacteria, including E. coli, metabolize to aerobic growth on xylose for natural xylose-assimilating xylose using a xylose isomerase. This pathway is organisms. Growth curves for natural xylose-assimilating hypothesized not to suffer from the redox imbal- organisms, comparing the parent strain to a strain lacking ance imposed by the XR–XDH pathway (Van glucose-6-P dehydrogenase (zwf1∆ cells). (A) Kluyveromyces lactis cells were cultured in synthetic complete medium (SC) Vleet and Jeffries, 2009). To determine whether plus 50 g/l xylose. (B) Escherichia coli cells were cultured in the need for G6PD was a result of a XR–XDH basal medium plus 20 g/l xylose. Growth curves shown pathway, we next looked at xylose-assimilation of represent the averages from experiments performed in an E. coli strain lacking G6PD activity. Deletion triplicate of the ZWF1 gene in E. coli had little effect on growth on xylose medium (Figure 4B), suggesting Xylose assimilation via a xylose isomerase that the oxidative branch of the PPP is not required in pathway S. cerevisiae relieves the need for xylose metabolism in E. coli. This result is most for G6PD, but not for gluconeogenesis likely due to the use of an isomerization pathway. Additionally, transhydrogenases are present To determine whether NADPH consumption by the in bacteria (Boonstra et al., 1999) and they may S. stipitis XR is a contributing factor to the strain’s facilitate NADPH regeneration during growth on dependence on ZWF1 for growth on xylose, we xylose. However, heterologous expression of trans- replaced the XR–XDH pathway with a xylose hydrogenases in yeast indicate that they may not isomerase (XI) pathway. Cells expressing the XI function in the needed direction (i.e. their expres- pathway convert xylose to xylulose in a single sion depletes NADPH) (Jeun et al., 2003; Nissen step, without an NADPH requirement (Kuyper et al., 2001). et al., 2004). Cells metabolizing xylose via a xylose

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea S. cerevisiae requires the PG11 and ZWF1 genes for xylose assimilation 655

4 either furfural or HMF (Gorsich et al., 2006). Addi- S. cerevisiae tionally, NADPH-dependent furfural reduction has WT (XI pathway) been observed in S. cerevisiae (Bowman et al., 3 ∆ ) zwf1 2010; Liu and Moon, 2009). We hypothesized that

660 pgi1∆ when cells are provided xylose as the only carbon source, sensitivity to these compounds will 2 be increased. This is important because: (a) most xylose-utilizing strains (engineered or natural) con-

Biomass (OD sume glucose prior to consuming xylose; and 1 (b) some strategies call for splitting the C6 and C5 carbon streams and processing them separately. 0 To test this hypothesis, we grew the yeast strains 0 24 48 72 96 120 144 168 in glucose or xylose medium in the presence of Time (hours) increasing inhibitor concentration. Tolerance to the thiol-oxidizing agent diamide Figure 5. S. cerevisiae strains expressing a xylose isomerase was assayed ﬁrst (Figure 6A). Diamide has been pathway do not require the oxidative branch of the shown to deplete NADPH levels; cells with PPP. Growth curves for S. cerevisiae strains engineered to metabolize xylose via expression of the Piromyces reduced ability to regenerate NADPH are hypersen- sp. E2 xylose isomerase and the S. cerevisiae xylulokinase sitive to diamide (Kosower et al., 1969; Larochelle genes. Comparison of the parent (WT–XI pathway) strain et al., 2006). The engineered S. cerevisiae strain and strains lacking the genes required for gluconeogenic expressing the XR–XDH pathway showed in- production of glucose-6-P (pgi1∆ cells) or the oxidative creased sensitivity when grown on xylose com- branch of the pentose phosphate pathway (zwf1∆ cells). Cells were cultured in synthetic complete medium (SC) pared to glucose medium (Figure 6A). This result plus 50 g/l xylose. Growth curves shown represent the is consistent with decreased availability of NADPH averages from experiments performed in triplicate during growth on xylose. Interestingly, on xylose medium at lower concentrations of diamide (0.8– 1.6 mM), cell growth was enhanced by approxi- isomerase pathway were slightly affected by dele- mately 30%. Exposure to 1.5 mM diamide has been tion of ZWF1, but clearly did not require ZWF1 shown to upregulate PPP genes via the transcrip- for growth like strains using the XR–XDH path- tion factor Stb5p (Larochelle et al., 2006). These way (Figure 5). Interestingly, S. cerevisiae cells enzymes are also essential for xylose metabolism expressing the XI pathway still required the pro- and increased expression of the PPP genes via duction of glucose-6-P via PGI1. Since these cells modiﬁed promoters has been previously exploited did not require ZWF1, we believe that the glucose- to increase xylose fermentation (Hahn-Hagerdal¨ 6-P supplied by PGI1 is required in this strain to et al., 2007). Exposure to low levels of diamide produce the glucan component of the yeast cell mimics this effect (i.e. increasing the expression wall. In this respect, a low level of gluconeogene- of PPP genes required for xylose metabolism), sis would be required for cell wall synthesis for all resulting in enhanced xylose metabolism through organisms when xylose is the only available carbon the PPP and increased cell growth at low diamide source. concentrations. At higher concentrations, however, cells grown in xylose medium were more sensitive Implications of NADPH limitation to diamide, compared to the same cells grown in for engineered S. cerevisiae strains glucose medium. The above results suggest that engineered Natural xylose-metabolizing yeasts were also S. cerevisiae strains would show decreased toler- assayed to determine whether they are hypersen- ance to fermentation inhibitors and NADPH stress sitive to diamide when grown on xylose medium. when cultured on xylose medium. Inhibitors in S. stipitis did not show increased sensitivity to lignocellulosic hydrolysates include furfural and diamide on xylose compared to glucose medium hydroxymethyl furfural (HMF). NADPH gener- (Figure 6A). We believe this result demonstrates ated via the oxidative branch of the PPP has been that S. stipitis is capable of regenerating NADPH shown to be required for growth in the presence of just as well on xylose as on glucose medium. In

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea 656 R. E. Hector et al.

S.c.-glucose S.c.-xylose general, the K. lactis strain used was more sen- S.s.-glucose S.s.-xylose sitive than S. cerevisiae and S. stipitis to diamide K.l.-glucose K.l.-xylose when grown on glucose or xylose. Despite the over- A 140 all increased sensitivity, we did not see a signiﬁ-

120 cant difference in sensitivity when grown in xylose medium compared to glucose. 100 Most reported tests for tolerance to fermenta- 80 tion inhibitors, either tested independently or as a 60 hydrolysate, have been performed using glucose as the main carbon source (Almeida et al., 2009; Gor- 40

Relative growth (%) sich et al., 2006; Jeppsson et al., 2003; Palmqvist 20 et al., 1999). Knowledge of how the various yeasts 0 tolerate inhibitors on xylose medium is lacking. 01234 Diamide concentration (mM) We repeated the sensitivity assay as above, but this time using furfural and HMF. Furfural and HMF B 120 are common byproducts of many lignocellulose 100 pre-treatments. Again, S. stipitis tolerated furfural and HMF somewhat equally, regardless of the car- 80 bon source (Figure 6B, C). K. lactis also showed

60 a similar tolerance to furfural on glucose and xylose, although its overall tolerance was reduced 40 compared to the other yeasts. Unlike S. stipitis, Relative growth (%) 20 K. lactis was more sensitive to HMF when cultured in xylose medium (Figure 6C). The results from 0 Figure 6B, C show that recombinant S. cerevisiae 0 10203040 Furfural concentration (mM) is more sensitive to the fermentation inhibitors furfural and HMF when cultured using xylose as C 120 the carbon source, compared to glucose. This lat- 100 ter result is especially important when considering splitting the C6 and C5 streams for processing the 80 pentose sugars separately. Based on the data pre- 60 sented, processing a split C5 sugar stream with

40 S. cerevisiae would require increased dilution or prior detoxiﬁcation. Alternatively, these data show Relative growth (%) 20 that a split C5 stream containing mainly xylose

0 should be well-tolerated by S. stipitis,atleastwith 0 10 20 30 40 50 respect to furfural or HMF. HMF concentration (mM)

Figure 6. Inhibitor sensitivity assays. Cells from mid-log Growth on xylose medium induces G6PD phase growth, pre-grown on synthetic medium plus glucose and PGI activity in natural xylose-assimilating or xylose, were diluted into various concentrations of yeasts, but not in engineered S. cerevisiae three different inhibitors. Relative growth was determined after culture in the presence of inhibitor. S. cerevisiae The results from the inhibitor sensitivity assays engineered for xylose metabolism (YRH501) showed suggest that engineered S. cerevisiae strains are increased sensitivity to NADPH depletion by diamide and to furfural or HMF exposure when cultured on xylose medium. not capable of regulating gluconeogenesis to sup- (A) The effect of diamide on xylose assimilation. (B) The ply glucose-6-P to the oxidative branch of the effect of furfural on xylose assimilation. (C) The effect PPP in response to xylose. To test this hypothe- of hydroxymethyl furfural (HMF) on xylose assimilation. sis, cells were grown to mid-log phase in either Results shown are the average values of three independent glucose or xylose medium and assayed for G6PD cultures. Standard deviations (SDs) were typically <10%. Relative growth was calculated as [100 × (culture + activity. We also measured PGI activity, which inhibitor OD660/culture without inhibitor OD660)] would be required if cells responded to xylose

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea S. cerevisiae requires the PG11 and ZWF1 genes for xylose assimilation 657

by upregulating genes to make glucose-6-P. Our A 1.5 results show that, compared to S. stipitis, engi- Glucose neered S. cerevisiae had lower levels of G6PD Xylose activity on glucose (Figure 7A). More importantly, on xylose medium, G6PD activity in S. cerevisiae 1.0

S. stipitis mole/min/mg) decreased, while G6PD activity for and µ K. lactis increased. With respect to PGI activity, both S. stipitis and K. lactis responded to xylose by 0.5 increasing PGI activity (Figure 7B). S. cerevisiae grown on xylose was not able to increase PGI activity. An increase in both PGI and G6PD activities G6PD activity ( would likely increase ﬂux through the oxidative 0.0 S. cerevisiae S. stipitis K. lactis branch of the PPP. The xylose-assimilating yeast, Candida utilis, has also been shown to increase B Glucose enzymes in the oxidative branch of the PPP in 4.0 Xylose response to xylose, although PGI activity was not assayed in this yeast (Bruinenberg et al., 1983a). 3.0 Our results with G6PD and PGI activity show that S. cerevisiae does not respond in the same mole/min/mg) µ manner as do natural xylose-assimilating yeasts, 2.0 and highlight an area for further improving xylose metabolism in engineered S. cerevisiae strains. 1.0

Dual NADPH and NADH cofactor use PGI activity ( 0.0 by the S. stipitis xylose reductase S. cerevisiae S. stipitis K. lactis The xylose reductase from S. stipitis has been shown to utilize both NADPH and NADH in vitro Figure 7. Saccharomyces cerevisiae strains engineered to metabolize xylose do not induce glucose-6-P dehydrogenase (Rizzi et al., 1988; Verduyn et al., 1985). It is or phosphoglucose isomerase activity on xylose medium this ability to use NADH that is hypothesized like natural xylose-assimilating yeasts. Cells were grown to to balance cofactors, allowing S. stipitis to fer- mid-log phase on either synthetic complete medium with ment xylose (Bruinenberg et al., 1984; van Dijken glucose or xylose as the carbon source. (A) Glucose-6-P and Scheffers, 1986). The S. stipitis XR expressed dehydrogenase activity. (B) Phosphoglucose isomerase activity. Data shown represents the average enzyme activity in S. cerevisiae has also been shown to func- measured from freshly prepared cell lysates from three tion (in vitro) with NADH (Amore et al., 1991). independently grown cultures. Error bars represent SDs. Our results with S. cerevisiae showing a complete Increased activity on xylose for S. stipitis and K. lactis was dependence on NADPH regeneration via G6PD in statistically signiﬁcant (p < 0.05) the strain expressing the S. stipitis XR–XDH pathway indicate that NADH is not used in vivo by the S. stipitis XR. Further questioning the use of and substrate, and in the absence of other cofactors NADH by the S. stipitis XR, a recent study com- and substrates. It is possible that the conditions paring a Saccharomyces strain engineered with an where NADH can be utilized by the S. stipitis NADPH-dependent XR from Neurospora crassa vs XR are never attained in vivo in S. cerevisiae. the dual cofactor accepting S. stipitis XR showed The majority of NAD(H) in the cell is typically + that the NADPH-dependent XR worked just as well maintained as NAD and the NADPH : NADH as the S. stipitis XR under aerobic and fermentative ratio in S. cerevisiae has been estimated to be ∼3 conditions (Bera et al., 2011). These results raise (Watanabe et al., 2007). Additionally, the cytosolic the important question, why doesn’t the S. stipitis free NAD+: NADH ratio (∼101–320, depending XR expressed in S. cerevisiae use NADH effec- on pH) has been estimated to be considerably tively in vivo? higher than the total NAD+: NADH ratio (7.5) Activity with NADPH and NADH measured (Canelas et al., 2008). Coupled with an ∼2-fold in vitro is typically done in excess of both cofactor lower afﬁnity for NADH (Verduyn et al., 1985),

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea 658 R. E. Hector et al. under most conditions free NADH may not be vector, the E. coli Genetic Stock Center for providing available at high enough levels for it to be used the E. coli ZWF1 deletion strain, and Katherine Card and in vivo by the XR. Additionally, NADP+ was Tricia Windgassen for their excellent technical assistance shown to inhibit NADH-linked XR activity much throughout this study. more than NADPH-linked activity (Verduyn, et al., 1985). Under conditions where NADPH is limiting, significant amounts of NADP+ would be present, References thus inhibiting NADH-linked XR activity in vivo. Our results with aerobic cultures of S. stipitis Almeida JR, Karhumaa K, Bengtsson O, et al. 2009. Screening show increases in both G6PD and PGI activity, of Saccharomyces cerevisiae strains with respect to anaerobic suggesting that in vivo for S. stipitis under these growth in non-detoxified lignocellulose hydrolysate. Bioresour conditions, the XR uses NADPH. Further studies Technol 100: 3674–3677. Alper H, Stephanopoulos G. 2007. Global transcription machinery with S. stipitis strains deficient in these activities engineering: a new approach for improving cellular phenotype. (e.g. G6PD and PGI), or expressing an NADPH- Metab Eng 9: 258–267. specific XR, will be needed to further investigate Amberg BC, Burke DJ, Strathern JN. 2005. Methods in Yeast the mechanisms of redox balance in this yeast. Genetics: A Cold Spring Harbor Laboratory Course Manual (2005 edn). Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY. Amore R, Kotter¨ P, Kuster C, et al. 1991. Cloning and expression Conclusions in Saccharomyces cerevisiae of the NAD(P)H-dependent xylose reductase-encoding gene (XYL1 ) from the xylose-assimilating In general, the sum of these results suggests yeast Pichia stipitis. Gene 109: 89–97. that natural xylose-assimilating yeasts respond to Baba T, Ara T, Hasegawa M, et al. 2006. Construction of xylose by redirecting carbon toward the produc- Escherichia coli K-12 in-frame, single-gene knockout mutants: tion of glucose-6-P and that S. cerevisiae lacks this the Keio collection. MolSystBiol2: 2006.0008. Bengtsson O, Hahn-Hagerdal¨ B, Gorwa-Grauslund MF. 2009. ability. This inappropriate regulation in response to Xylose reductase from Pichia stipitis with altered coenzyme xylose resulted in decreased capacity of engineered preference improves ethanolic xylose fermentation by recombi- S. cerevisiae to metabolize xylose and increased nant Saccharomyces cerevisiae. Biotechnol Biofuels 2:9. sensitivity to diamide, furfural and HMF when Bera AK, Ho NW, Khan A, et al. 2011. A genetic overhaul of xylose was the only carbon source. Carbon recy- Saccharomyces cerevisiae 424A(LNH-ST) to improve xylose cling (i.e. from xylose to glucose-6-P) also reduces fermentation. J Ind Microbiol Biot 38: 617–626. Bertani G, Weigle JJ. 1953. Host controlled variation in bacterial the amount of carbon remaining for cell mass and viruses. J Bacteriol 65: 113–121. end-products, since for every glucose-6-P cycled Billard P, Menart S, Fleer R, et al. 1995. Isolation and through the oxidative branch of the PPP, carbon is characterization of the gene encoding xylose reductase from Kluyveromyces lactis. Gene 162: 93–97. lost as CO2. Expression of a xylose isomerase pathway reduced dependence on the oxidative branch Boonstra B, French CE, Wainwright I, et al. 1999. The udhA gene of Escherichia coli encodes a soluble pyridine nucleotide of the PPP. Engineering a xylose reductase with transhydrogenase. J Bacteriol 181: 1030–1034. increased specificity for NADH should also lead Bowman MJ, Jordan DB, Vermillion KE, et al. 2010. Stereochem- to decreased dependence on G6PD and studies that istry of furfural reduction by a Saccharomyces cerevisiae alde- have followed this approach have shown increased hyde reductase that contributes to in situ furfural detoxification. xylose metabolism (Bengtsson et al., 2009; Jepps- Appl Environ Microbiol 76: 4926–4932. Bruinenberg PM, de Bot PHM, van Dijken JP, et al. 1984. son et al., 2006; Krahulec et al., 2010; Watanabe NADH-linked aldose reductase: the key to anaerobic alcoholic et al., 2007). Alternatively, process modifications, fermentation of xylose by yeasts. Appl Microbiol Biotechnol 19: such as simultaneous saccharification and fermen- 256–260. tation or fed-batch fermentation, which keep glu- Bruinenberg PM, van Dijken JP, Scheffers WA. 1983a. An cose available at low concentration, should provide enzymic analysis of NADPH production and consumption in Candida utilis. J Gen Microbiol 129: 965–971. sufficient NADPH to enhance xylose utilization Bruinenberg PM, Van Dijken JP, Scheffers WA. 1983b. A and increase tolerance to inhibitors. theoretical analysis of NADPH production and consumption in yeasts. J Gen Microbiol 129: 953–964. Acknowledgements Cadiere` A, Galeote V, Dequin S. 2010. The Saccharomyces cerevisiae zinc factor protein Stb5p is required as a basal We thank Dr Michele Saliola for providing us with the regulator of the pentose phosphate pathway. FEMS Yeast Res K. lactis KlZWF1 deletion strain and KlZWF1 expression 10: 819–827.

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea S. cerevisiae requires the PG11 and ZWF1 genes for xylose assimilation 659

Canelas AB, van Gulik WM, Heijnen JJ. 2008. Determination Krahulec S, Petschacher B, Wallner M, et al. 2010. Fermentation of the cytosolic free NAD/NADH ratio in Saccharomyces of mixed glucose-xylose substrates by engineered strains of cerevisiae under steady-state and highly dynamic conditions. Saccharomyces cerevisiae: role of the coenzyme specificity of Biotechnol and Bioeng 100: 734–743. xylose reductase, and effect of glucose on xylose utilization. Christianson TW, Sikorski RS, Dante M, et al. 1992. Multifunc- Microb Cell Fact 9: 16. tional yeast high-copy-number shuttle vectors. Gene 110: Kuhn A, Vanzyl C, Vantonder A, et al. 1995. Purification 119–122. and partial characterization of an aldo-keto reductase Dickinson JR, Sobanski MA, Hewlins MJ. 1995. In Saccharomyces from Saccharomyces cerevisiae. Appl Environ Microb 61: cerevisiae deletion of phosphoglucose isomerase can be 1580–1585. suppressed by increased activities of enzymes of the hexose Kurtzman C, Suzuki M. 2010. Phylogenetic analysis of ascomy- monophosphate pathway. Microbiol 141: 385–391. cete yeasts that form coenzyme Q-9 and the proposal of the new Gietz D, Woods RA. 2002. Transformation of yeasts by the lithium genera Babjeviella, Meyerozyma, Millerozyma, Priceomyces,and acetate/single-stranded carrier/polyethylene glycol method. Scheffersomyces. Mycoscience 51: 2–14. Method Enzymol 350: 87–96. Kuyper M, Winkler AA, van Dijken JP, et al. 2004. Minimal Gorsich SW, Dien BS, Nichols NN, et al. 2006. Tolerance to metabolic engineering of Saccharomyces cerevisiae for efficient furfural-induced stress is associated with pentose phosphate anaerobic xylose fermentation: a proof of principle. FEMS Yeast pathway genes ZWF1, GND1, RPE1,andTKL1 in Saccha- Res 4: 655–664. romyces cerevisiae. Appl Microbiol Biotechnol 71: 339–349. Lagunas R, Gancedo JM. 1973. Reduced pyridine-nucleotides Grabowska D, Chelstowska A. 2003. The ALD6 gene product balance in glucose-growing Saccharomyces cerevisiae. Eur is indispensable for providing NADPH in yeast cells lacking J Biochem 37: 90–94. glucose-6-phosphate dehydrogenase activity. JBiolChem278: Larochelle M, Drouin S, Robert F, et al. 2006. Oxidative stress- 13984–13988. activated zinc cluster protein Stb5 has dual activator/repressor Hahn-Hagerdal¨ B, Karhumaa K, Jeppsson M, et al. 2007. Meta- functions required for pentose phosphate pathway regulation and bolic engineering for pentose utilization in Saccharomyces NADPH production. MolCellBiol26: 6690–6701. cerevisiae. Adv Biochem Eng Biotechnol 108: 147–177. Liu ZL, Moon J. 2009. A novel NADPH-dependent aldehyde Hector RE, Bowman MJ, Skory CD, et al. 2009. The Saccha- reductase gene from Saccharomyces cerevisiae NRRL Y-12632 romyces cerevisiae YMR315W gene encodes an NADP(H)- involved in the detoxification of aldehyde inhibitors derived specific oxidoreductase regulated by the transcription factor from lignocellulosic biomass conversion. Gene 446: 1–10. Stb5p in response to NADPH limitation. N Biotechnol 26: Minard KI, McAlister-Henn L. 2005. Sources of NADPH in yeast 171–180. vary with carbon source. JBiolChem280: 39890–39896. Hector RE, Dien BS, Cotta MA, et al. 2010. Engineering indus- Nichols NN, Mertens JA. 2008. Identification and transcriptional trial Saccharomyces cerevisiae strains for xylose fermentation profiling of Pseudomonas putida genes involved in furoic acid and comparison for switchgrass conversion. J Ind Microbiol Biot metabolism. FEMS Microbiol Lett 284: 52–57. in press. doi: 10.1007/s10295-010-0896-1. Nissen TL, Anderlund M, Nielsen J, et al. 2001. Expression of Hughes SR, Sterner DE, Bischoff KM, et al. 2009. Engineered a cytoplasmic transhydrogenase in Saccharomyces cerevisiae Saccharomyces cerevisiae strain for improved xylose utilization with a three-plasmid SUMO yeast expression system. Plasmid results in formation of 2-oxoglutarate due to depletion of the 61: 22–38. NADPH pool. Yeast 18: 19–32. Jeffries TW, Van Vleet JR. 2009. Pichia stipitis genomics, Nogae I, Johnston M. 1990. Isolation and characterization of the transcriptomics, and gene clusters. FEMS Yeast Res. 9: 793–807. ZWF1 gene of Saccharomyces cerevisiae, encoding glucose-6- Jeppsson M, Bengtsson O, Franke K, et al. 2006. The expression phosphate dehydrogenase. Gene 96: 161–169. Palmqvist E, Grage H, Meinander NQ, et al. 1999. Main and of a Pichia stipitis xylose reductase mutant with higher Km for NADPH increases ethanol production from xylose in interaction effects of acetic acid, furfural, and p-hydroxybenzoic recombinant Saccharomyces cerevisiae. Biotechnol Bioeng 93: acid on growth and ethanol productivity of yeasts. Biotechnol 665–673. Bioeng 63: 46–55. Jeppsson M, Johansson B, Jensen PR, et al. 2003. The level of Rizzi M, Erlemann P, Bui-Thanh N-A, et al. 1988. Xylose glucose-6-phosphate dehydrogenase activity strongly influences fermentation by yeasts. Appl Microbiol Biotechnol 29: 148–154. xylose fermentation and inhibitor sensitivity in recombinant Runquist D, Hahn-Hagerdal¨ B, Bettiga M. 2009. Increased Saccharomyces cerevisiae strains. Yeast 20: 1263–1272. expression of the oxidative pentose phosphate pathway and Jeun Y-S, Kim M-D, Park Y-C, et al. 2003. Expression gluconeogenesis in anaerobically growing xylose-utilizing of Azotobacter vinelandii soluble transhydrogenase perturbs Saccharomyces cerevisiae. Microb Cell Fact 8: 49. xylose reductase-mediated conversion of xylose to xylitol by Saliola M, Scappucci G, De Maria I, et al. 2007. Deletion of recombinant Saccharomyces cerevisiae. JMolCatalBEnzym the glucose-6-phosphate dehydrogenase gene KlZWF1 affects 26: 251–256. both fermentative and respiratory metabolism in Kluyveromyces Karhumaa K, Fromanger R, Hahn-Hagerdal¨ B, et al. 2007. High lactis. Eukaryot Cell 6: 19–27. activity of xylose reductase and xylitol dehydrogenase improves Salusjarvi¨ L, Poutanen M, Pitkanen¨ JP, et al. 2003. Proteome xylose fermentation by recombinant Saccharomyces cerevisiae. analysis of recombinant xylose-fermenting Saccharomyces Appl Microbiol Biotechnol 73: 1039–1046. cerevisiae. Yeast 20: 295–314. Kosower NS, Kosower EM, Wertheim B, et al. 1969. Diamide, a Sambrook J, Russell DW. 2001. Molecular Cloning: A Laboratory new reagent for the intracellular oxidation of glutathione to the Manual, 3rd edn. Cold Spring Harbor Laboratory Press: Cold disulfide. Biochem Biophys Res Commun 37: 593–596. Spring Harbor, NY.

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea 660 R. E. Hector et al.

Sassner P, Galbe M, Zacchi G. 2008. Techno-economic evaluation Verduyn C, Van Kleef R, Frank J, et al. 1985. Properties of of bioethanol production from three different lignocellulosic the NAD(P)H-dependent xylose reductase from the xylose- materials. Biomass Bioenerg 32: 422–430. fermenting yeast Pichia stipitis. Biochem J 226: 669–677. Toivari MH, Salusjarvi L, Ruohonen L, et al. 2004. Endogenous Verho R, Londesborough J, Penttila M, et al. 2003. Engineering xylose pathway in Saccharomyces cerevisiae. Appl Environ redox cofactor regeneration for improved pentose fermentation Microbiol 70: 3681–3686. in Saccharomyces cerevisiae. Appl Environ Microbiol 69: Traff¨ KL, Jonsson LJ, Hahn-Hagerdal¨ B. 2002. Putative xylose 5892–5897. and arabinose reductases in Saccharomyces cerevisiae. Yeast 19: Verho R, Richard P, Jonson PH, et al. 2002. Identiﬁcation of 1233–1241. the ﬁrst fungal NADP–GAPDH from Kluyveromyces lactis. van Dijken JP, Scheffers WA. 1986. Redox balances in the Biochemistry 41: 13833–13838. metabolism of sugars by yeast. FEMS Microbiol Rev: Watanabe S, Abu Saleh A, Pack SP, et al. 2007. Ethanol 199–224. production from xylose by recombinant Saccharomyces Van Vleet JH, Jeffries TW. 2009. Yeast metabolic engineering for cerevisiae expressing protein-engineered NADH-preferring hemicellulosic ethanol production. Curr Opin Biotechnol 20: xylose reductase from Pichia stipitis. Microbiol 153: 300–306. 3044–3054.

Published in 2011 by John Wiley & Sons, Ltd. Yeast 2011; 28: 645–660. DOI: 10.1002/yea