Understanding Functional Roles of Native Pentose-Specific Transporters for Activating Dormant Pentose Metabolism in Yarrowia lipolytica

ABSTRACT

Pentoses, including xylose and arabinose, are the second most prevalent sugars in lignocellulosic biomass that can be harnessed for biological conversion. Although Yarrowia lipolytica has emerged as a promising industrial microorganism for production of high-value chemicals and biofuels, its native pentose metabolism is poorly understood. Our previous study demonstrated that Y. lipolytica (ATCC MYA-2613) has endogenous enzymes for d-xylose assimilation, but inefficient xylitol dehydrogenase causes Y. lipolytica to assimilate xylose poorly. In this study, we investigated the functional roles of native sugar-specific transporters for activating the dormant pentose metabolism in Y. lipolytica. By screening a comprehensive set of 16 putative pentose-specific transporters, we identified two candidates, YALI0C04730p and YALI0B00396p, that enhanced xylose assimilation. The engineered mutants YlSR207 and YlSR223, overexpressing YALI0C04730p and YALI0B00396p, respectively, improved xylose assimilation approximately 23% and 50% in comparison to YlSR102, a parental engineered strain overexpressing solely the native xylitol dehydrogenase gene. Further, we activated and elucidated a widely unknown native l-arabinose assimilation pathway in Y. lipolytica through transcriptomic and metabolic analyses. We discovered that Y. lipolytica can coconsume xylose and arabinose, where arabinose utilization shares transporters and metabolic enzymes of some intermediate steps of the xylose assimilation pathway. Arabinose assimilation is synergistically enhanced in the presence of xylose, while xylose assimilation is competitively inhibited by arabinose. l-Arabitol dehydrogenase is the rate-limiting step responsible for poor arabinose utilization in Y. lipolytica. Overall, this study sheds light on the cryptic pentose metabolism of Y. lipolytica and, further, helps guide strain engineering of Y. lipolytica for enhanced assimilation of pentose sugars.

IMPORTANCE The oleaginous yeast Yarrowia lipolytica is a promising industrial-platform microorganism for production of high-value chemicals and fuels. For decades since its isolation, Y. lipolytica has been known to be incapable of assimilating pentose sugars, xylose and arabinose, that are dominantly present in lignocellulosic biomass. Through bioinformatic, transcriptomic, and enzymatic studies, we have uncovered the dormant pentose metabolism of Y. lipolytica. Remarkably, unlike most yeast strains, which share the same transporters for importing hexose and pentose sugars, we discovered that Y. lipolytica possesses the native pentose-specific transporters. By overexpressing these transporters together with the rate-limiting d-xylitol and l-arabitol dehydrogenases, we activated the dormant pentose metabolism of Y. lipolytica. Overall, this study provides a fundamental understanding of the dormant pentose metabolism of Y. lipolytica and guides future metabolic engineering of Y. lipolytica for enhanced conversion of pentose sugars to high-value chemicals and fuels.

INTRODUCTION

Lignocellulosic biomass derived from agricultural residues, municipal solid wastes, and woody residues can be biologically upgraded to produce high-value chemicals and fuels (1). Biomass feedstocks typically contain high contents of cellulosic and hemicellulosic components, up to 80 to 90% in some biomasses, such as corn cobs and nut shells (2). While cellulose is a homopolymer of glucose linked by beta-1,4-glycosidic bond with cellobiose as a structural unit, hemicellulose is a heteropolymer of pentose (C₅) sugars (e.g., d-xylose [Xyl] and l-arabinose [Ara]), hexose (C₆) sugars (e.g., d-glucose, d-mannose, and d-galactose), and sugar acids (3). After pretreatment and saccharification, these C₅ and C₆ complex sugars are released as monomers, which are subsequently fermented by microorganisms to produce target products (4). For industrial biocatalysis, microbial catalysts capable of coutilizing all complex sugars efficiently without carbon catabolite repression are desirable.

Yarrowia lipolytica, a generally regarded as safe (GRAS) oleaginous yeast, has emerged as a biomanufacturing platform for production of high-value chemicals and biofuels (5, 6). It has an efficient and robust native metabolism to produce high levels of organic acids and neutral lipids. Combining metabolic engineering and fermentation optimization, engineered Y. lipolytica strains have been generated for large-scale production of citric acid (titer, 55 g/liter; rate, 0.6 g/liter/h; yield, 0.83 g/g glucose) (7), alpha-ketoglutaric acid (186 g/liter; 1.75 g/liter/h; 0.36 g/g glycerol) (6), succinic acid (160.2 g/liter; 0.4 g/liter/h; 0.4 g/g crude glycerol) (8), and lipids (98.9 g/liter; 1.3 g/liter/h; 0.27 g/g glucose) (9). In addition, Y. lipolytica exhibits exceptional robustness in tolerating the inhibitory environments associated with biomass pretreatments. For instance, ionic liquids, such as 1-ethyl-3-methyl imidazolium acetate (EMIM OAc), are promising solvents for pretreatment of recalcitrant lignocellulosic biomass (10) but are very inhibitory to microbial health at a level as low as 5% (vol/vol) (11). Remarkably, native Y. lipolytica can thrive and produce high-yield (>90%) alpha-ketoglutaric acid from cellulose in media containing up to 10% (vol/vol) EMIM OAc (12). By screening the genetic diversity of a large collection of 45 Y. lipolytica strains isolated from different habitats, it was found that certain Y. lipolytica strains can thrive in 90% undetoxified, dilute-acid-pretreated switchgrass hydrolysate and yield relatively high rates of lipid production (13).

One current limitation is that native Y. lipolytica is very inefficient at utilizing pentose sugars available from biomass hydrolysates. Figure 1 shows the pentose assimilation pathways in yeasts. Xylose is degraded via the oxidoreductase pathway. Upon being transported intracellularly, xylose reductase (XYL1) converts xylose into d-xylitol, which is further transformed into d-xylulose by xylitol dehydrogenase (XYL2). Next, xylulokinase (XYL3) converts d-xylulose into d-xylulose-5-phosphate, which enters the pentose phosphate pathway (PPP) and is further assimilated for cell growth. On the other hand, l-arabinose is first reduced into l-arabitol by NAD(P)H-dependent arabinose reductase (ARD), which is then converted into l-xylulose by NAD(P)⁺-dependent arabitol dehydrogenase (ADH). l-Xylulose is then converted to d-xylitol by NAD(P)H-dependent xylulose reductase (XLR) (14), which is further assimilated to d-xylulose-5-phosphate, a precursor for PPP. While Y. lipolytica is well known for assimilating hexose sugars (e.g., glucose), its pentose metabolism is dormant and poorly understood.

FIG 1.

d-Xylose (green) and l-arabinose (purple) assimilation pathways in yeast. Abbreviations: XYL1, xylose reductase; XYL2, xylitol dehydrogenase; XYL3, xylulokinase; TKL, transketolase; TAL, transaldolase; ARD, arabinose reductase; ADH, arabitol dehydrogenase; XLR, xylulose reductase.

We have recently activated and elucidated the native metabolism of Y. lipolytica for xylose assimilation. By performing bioinformatic, enzymatic, and transcriptomic analyses, we identified 16 putative pentose-specific transporters and demonstrated that, like XYL2, the xylose transporter is an important rate-limiting step that causes poor xylose assimilation in native Y. lipolytica (15). Several metabolic engineering approaches have been applied to enhance xylose assimilation by Y. lipolytica. The first strategy was to express the native metabolic enzymes, e.g., XYL2 (15) or a combination of XYL2 and XYL3 (16), to improve xylose assimilation. The second approach was to overexpress heterologous enzymes of the xylose assimilation pathway. By overexpressing xyl1 and xyl2 genes from Pichia stipitis, together with implementing a starvation adaptation strategy, the engineered Y. lipolytica produced 15 g/liter of lipid from xylose at a rate of 0.19 g/liter/h (17). Likewise, by overexpressing xyl1 and xyl2 genes from P. stipitis and xyl3 from Y. lipolytica, Ledesma-Amaro et al. generated a recombinant Y. lipolytica that could consume 30 g/liter xylose within 3 days (18) (see Table S3 in the supplemental material). It remains unknown whether Y. lipolytica has a native l-arabinose metabolism and is capable of assimilating this substrate as a sole carbon source for growth.

In this study, we analyzed the functional roles of native pentose-specific transporters for activating a dormant pentose metabolism in Y. lipolytica. We screened a set of 16 putative pentose-specific transporters to identify the best candidates to enhance xylose assimilation and demonstrated that these transporters are specific to not only xylose, but also arabinose. We further activated and elucidated the native arabinose assimilation pathway in Y. lipolytica and shed light on the functional roles of transporters and metabolic enzymes for coutilization of xylose and arabinose. With targeted enzymatic and transcriptomic analyses, we further identified the arabitol dehydrogenase as the rate-limiting step that causes poor arabinose assimilation in Y. lipolytica.

RESULTS AND DISCUSSION

Screening putative pentose-specific transporters for enhanced xylose consumption.

We constructed a set of 16 Y. lipolytica strains, YlSR202, YlSR203, YlSR205, YlSR207 to YlSR209, YlSR212 to YlSR218, YlSR220, YlSR222, and YlSR223, each of which overexpresses a putative pentose-specific transporter and a native xylitol dehydrogenase (Table 1). To identify the best xylose-assimilating candidates, we screened these strains for fast growth on solid-agar plates containing xylose as a carbon source. The results showed that YlSR207 and YlSR223, carrying Y. lipolytica TRP6 (TRP6_Yli) (YALI0C04730g) and TRP22_Yli (YALI0B00396g), respectively, exhibited enhanced growth on xylose within 72 h of incubation, while other Y. lipolytica strains exhibited either poor or no growth (Fig. 2). Poor growth may be attributed to low affinity of the transporters for xylose and/or improper folding of overexpressed transporters.

TABLE 1.

Plasmids and strains used in this study

Plasmid/strain	Genotype	Source
Plasmids
pSR001	pSL16-PTEF-TCYC1::leu2	12
pSR002	pSL16-PTEF-XYL2Yli-TCYC1::leu2	15
pSR008	pSL16-PTEF-TCYC1::ura3	This study
pSR019	pSL16-PTEF-trp1Yli-TCYC1::ura3	This study
pSR020	pSL16-PTEF-trp2Yli-TCYC1::ura3	This study
pSR022	pSL16-PTEF-trp4Yli-TCYC1::ura3	This study
pSR024	pSL16-PTEF-trp6Yli-TCYC1::ura3	This study
pSR009	pSL16-PTEF-trp7Yli-TCYC1::ura3	This study
pSR025	pSL16-PTEF-trp8Yli-TCYC1::ura3	This study
pSR028	pSL16-PTEF-trp11Yli-TCYC1::ura3	This study
pSR029	pSL16-PTEF-trp12Yli-TCYC1::ura3	This study
pSR030	pSL16-PTEF-trp13Yli-TCYC1::ura3	This study
pSR031	pSL16-PTEF-trp14Yli-TCYC1::ura3	This study
pSR032	pSL16-PTEF-trp15Yli-TCYC1::ura3	This study
pSR033	pSL16-PTEF-trp16Yli-TCYC1::ura3	This study
pSR034	pSL16-PTEF-trp17Yli-TCYC1::ura3	This study
pSR036	pSL16-PTEF-trp19Yli-TCYC1::ura3	This study
pSR038	pSL16-PTEF-trp21Yli-TCYC1::ura3	This study
pSR039	pSL16-PTEF-trp22Yli-TCYC1::ura3	This study
pSR041	pSL16-PTEF-adhAoz-TCYC1::leu2	This study
Y. lipolytica strains
YlSR001	MATA ura3-302 leu2-270 xpr2-322 axp2-ΔNU49 XPR2::SUC2	ATCC MYA-2613
YlSR101	YlSR001 + pSR001	15
YlSR102	YlSR001 + pSR002	15
YlSR202	YlSR001 + pSR002 + pSR019	This study
YlSR203	YlSR001 + pSR002 + pSR020	This study
YlSR205	YlSR001 + pSR002 + pSR022	This study
YlSR207	YlSR001 + pSR002 + pSR024	This study
YlSR208	YlSR001 + pSR002 + pSR009	This study
YlSR209	YlSR001 + pSR002 + pSR025	This study
YlSR212	YlSR001 + pSR002 + pSR028	This study
YlSR213	YlSR001 + pSR002 + pSR029	This study
YlSR214	YlSR001 + pSR002 + pSR030	This study
YlSR215	YlSR001 + pSR002 + pSR031	This study
YlSR216	YlSR001 + pSR002 + pSR032	This study
YlSR217	YlSR001 + pSR002 + pSR033	This study
YlSR218	YlSR001 + pSR002 + pSR034	This study
YlSR220	YlSR001 + pSR002 + pSR036	This study
YlSR222	YlSR001 + pSR002 + pSR038	This study
YlSR223	YlSR001 + pSR002 + pSR039	This study
YlSR157	YlSR001 + pSR041	This study

FIG 2.

Growth screening of Y. lipolytica strains expressing putative xylose-specific transporters on SC−Leu−Ura agar plates containing 10 g/liter xylose as a single sugar.

Overall, we identified the two Y. lipolytica strains, YlSR207 carrying TRP6_Yli and YlSR223 carrying TRP22_Yli, that exhibited the most effective growth on xylose. This result provides a basis for detailed characterization of the functional roles of these transporters in activating the native pentose metabolism in Y. lipolytica in subsequent studies.

Elucidating that TRP6_Yli and TRP22_Yli are xylose-specific transporters in Y. lipolytica.

To demonstrate that TRP6_Yli and TRP22_Yli are xylose-specific transporters in vivo, we characterized YlSR207 and YlSR223 in a defined, xylose-containing SC−Leu−Ura (where SC represents “synthetic complete”) liquid medium (see Materials and Methods). As a reference, we chose YlSR202, which overexpressed TRP1_Yli and did not improve xylose assimilation (Fig. 2). The result showed that YlSR202, YlSR207, and YlSR223 grew on xylose as a single carbon source without significant accumulation of xylitol (Fig. 3A and B), because they overexpressed XYL2_Yli, which has been previously shown to be the rate-limiting step of the xylose assimilation pathway in Y. lipolytica (15). Specifically, the reference strain, YlSR202, grew most slowly, with a specific growth rate of 0.014 ± 0.001 reciprocal hour (1/h), and took 192 h to completely consume 10 g/liter xylose with a specific xylose uptake rate of 0.154 ± 0.014 mmol/g cell dry weight (CDW)/h (Table 2). YlSR202 exhibited the same growth phenotype as YlSR102, which overexpressed only XYL2_Yli without any transporter, as previously characterized (15). In contrast, YlSR207 (0.019 ± 0.004 1/h) and YlSR223 (0.023 ± 0.001 1/h) grew 40% and 62% faster than YlSR202, respectively, under the same growth conditions (P < 0.05; n ≥ 6). Both YlSR207 and YlSR223 were also capable of completely consuming 10 g/liter xylose within 108 h, with 23% and 50% increases in the specific xylose uptake rates, respectively (P < 0.05; n ≥ 6).

FIG 3.

Profiles of cell growth and metabolites of Y. lipolytica strains YlSR202, YlSR207, and YlSR223 growing on single xylose. Circles, OD; squares, xylose; triangles, xylitol. Each data point represents an average value ±1 standard deviation from the results of 6 biological replicates.

TABLE 2.

Specific growth rates (μ) and specific sugar uptake rates (r_S) of Y. lipolytica strains growing on single xylose and mixed pentose (Xyl + Ara) sugars

Strain	μ (1/h)		rS (mmol/g CDW/h)
			rxyl		rara (Xyl + Ara)
	Xyl	Xyl + Ara	Xyl	Xyl + Ara
YlSR102	0.014 ± 0.001	0.007 ± 0.001	0.154 ± 0.014	0.095 ± 0.007	0.070 ± 0.007
YlSR207	0.019 ± 0.004	0.018 ± 0.000	0.190 ± 0.039	0.170 ± 0.028	0.043 ± 0.014
YlSR223	0.023 ± 0.001	0.019 ± 0.001	0.231 ± 0.042	0.163 ± 0.011	0.021 ± 0.006

To further validate the notion that the improvements in cell growth and xylose consumption of YlSR207 and YlSR223 were due to overexpression of TRP6_Yli and TRP22_Yli, but not XYL2_Yli, we showed that the strains YlSR202, YLSR207, and YlSR223 exhibited the same xylitol dehydrogenase XYL2 activity regardless of the cofactor (NAD⁺ or NADP⁺) used for the assay (P < 0.05; n ≥ 6) (see Fig. S1 in the supplemental material). By reviving strains YlSR207 and YlSR223 from frozen glycerol stocks, we confirmed their stable growth phenotypes, since their growth rates remained unaffected. Further, YlSR207 and YlSR223 exhibited the same growth rates without lag when they were first cultured in glucose medium and then transferred to fresh xylose medium.

Altogether, the data showed that Y. lipolytica has not only native metabolic enzymes (15), but also xylose-specific transporters for xylose assimilation, as demonstrated here, even though its xylose metabolism remains dormant due to transcriptional repression. This native xylose metabolism can be activated by overexpressing the native pentose assimilation genes of Y. lipolytica without a need to use any heterologous enzymes. We demonstrated that TRP6_Yli and TRP22_Yli are native xylose-specific transporters of Y. lipolytica. In yeast, the specificity of a xylose transporter is determined by the G-G/F-X-X-X-G structural motif and two conserved amino acids, threonine and asparagine (i.e., T213 and N370 in Saccharomyces cerevisiae HXT7) (19, 20). TRP6_Yli (YALI0C04730p) contains a G-F-L-L-F-G structural motif and tyrosine instead of asparagine (N348Y) (15). Although no conserved structural motif is found in TRP22_Yli (YALI0B00396p), valine, a small hydrophobic amino acid, instead of threonine and asparagine contributes to the xylose specificity (15).

Engineering pentose-specific transporters is critical for enhanced pentose assimilation in yeasts, especially in the presence of a competitive glucose substrate. For instance, S. cerevisiae is not known to have xylose-specific transporters, and therefore, it is required to express the heterologous transporter Candida intermedia Gxf1 (21) or to engineer the native glucose-specific transporter HXT7 to be xylose specific (19, 20) for enhanced xylose assimilation. The discovery of the functional roles of TRP6_Yli and TRP22_Yli will be useful for metabolic engineering of other yeasts (such as S. cerevisiae) to enhance xylose assimilation.

Understanding the native l-arabinose assimilation pathway in Y. lipolytica. (i) Degradation of single arabinose.

The existence and function of the native arabinose assimilation pathway in Y. lipolytica are widely unknown. To investigate the native arabinose metabolism of Y. lipolytica, we first performed genome mining. The bioinformatic result showed that Y. lipolytica has the putative enzymes of the arabinose assimilation pathway, including arabinose reductase (ARD), arabitol dehydrogenase (ADH), and xylulose reductase (XLR) (Fig. 1). We identified 11 putative ARD genes, 5 putative XLR genes, and 1 putative ADH gene (see Table S2 in the supplemental material for the gene loci). Interestingly, the arabitol dehydrogenase ADH_Yli gene (YALI0E12463g) is the same gene identified as a xylitol dehydrogenase (XYL2_Yli) gene (15).

To test whether the arabinose assimilation pathway is active in Y. lipolytica, we first cultured YlSR102 (overexpressing XYL2_Yli) in defined liquid medium containing arabinose as a carbon source. The result showed that YlSR102 grew poorly on arabinose and consumed only an insignificant amount of arabinose (Fig. 4A). However, transcriptomics showed that the arabinose assimilation pathway genes were upregulated. Four ARD genes—ARD6, ARD7, ARD8, and ARD9—out of 11 putative ARD genes were upregulated in cell cultures growing on single arabinose by 2.74- ± 1.40-fold, 10.77- ± 2.95-fold, 7.21- ± 1.06-fold, and 5.86- ± 0.23-fold, respectively, compared to cells grown in single xylose (Fig. 5). Two XLR genes—XLR1 and XLR4—out of 5 putative XLR genes were upregulated by 5.13- ± 0.29-fold and 4.32- ± 0.46-fold, respectively. The ADH_Yli (also identified as XYL2_Yli) gene was upregulated by 1.81- ± 0.70-fold. Consistent with the transcriptomic data, we also detected ARD, ADH, and XLR enzyme activities from YlSR102 cultures (Table 3). The ARD and XLR activities, specific to NADPH, were significantly higher for YlSR102 growing on arabinose than for YlSR102 growing on xylose (negative control). In contrast, the ADH activity, specific to NAD⁺, remained similar between cells grown in single arabinose and single xylose. In addition, we characterized YlSR207 and YlSR223 to investigate whether TRP6_Yli and TRP22_Yli helped enhance arabinose assimilation. However, like YlSR102, YlSR207 and YlSR223 poorly consumed arabinose (Fig. 4B and C).

FIG 4.

Profiles of cell growth and metabolites of Y. lipolytica strains YlSR102 (A), YlSR207 (B), YlSR223 (C), and YlSR157 (D) growing on single arabinose. Each data point represents an average value ±1 standard deviation from the results of 6 biological replicates.

FIG 5.

Relative mRNA expression levels (log₂ scale) of putative pentose assimilation pathway genes of Y. lipolytica YlSR102 growing on single arabinose (Ara) and mixed pentose sugars (Xyl+Ara). The reference condition for normalization was growth on single xylose. Each data point represents an average value of 3 biological replicates. Error bars are not included to avoid crowding but are presented in Table S2 in the supplemental material.

TABLE 3.

In vitro enzyme activities of ARD, ADH, XLR, and XYL2 of YlSR102 growing on single xylose, single arabinose, and mixed pentose sugars

Substrate	In vitro enzyme activitya (U)
	ARD		ADH		XLR		XYL2 (NAD+)
	NADH	NADPH	NAD+	NADP+	NADH	NADPH
Xyl	ND	ND	0.07 ± 0.01	ND	ND	0.06 ± 0.02	2.48 ± 0.50
Ara	ND	0.75 ± 0.18	0.07 ± 0.01	ND	ND	0.35 ± 0.05	1.36 ± 0.22
Xyl + Ara	ND	1.62 ± 0.14	0.06 ± 0.00	ND	ND	0.59 ± 0.09	2.10 ± 0.53

^aNAD(P)⁺ and NAD(P)H were used as cofactors for enzyme assays. ND, not detected.

Taken together, the results suggest that Y. lipolytica has endogenous metabolic capability to assimilate arabinose as a carbon source, but levels of gene expression and/or activities of metabolic enzymes and transporters might not be sufficient. The cofactor imbalance, due to ARD and XLR specificity for NADPH and of ADH for NAD⁺, might also have contributed to poor arabinose assimilation in Y. lipolytica.

(ii) Degradation of a mixture of xylose and arabinose by YlSR102.

Since the arabinose and xylose assimilation pathways are interconnected (Fig. 1), especially when Y. lipolytica has only one putative YALI0E12463p enzyme that functions as xylitol dehydrogenase (XYL2_Yli) and arabitol dehydrogenase (ADH_Yli), we investigated whether activation of the native xylose metabolism of Y. lipolytica by growing it on a mixture of xylose and arabinose could enhance arabinose assimilation.

Characterization of YlSR102 showed that it was able to grow in a mixture of xylose and arabinose (Fig. 6A). Upregulation of the xylose- and arabinose-metabolizing genes further validated the active pentose metabolism of Y. lipolytica (Fig. 5). YlSR102 consumed 1.82 ± 0.81 g/liter arabinose with a specific arabinose uptake rate of 0.070 ± 0.007 mmol/g CDW/h (Table 2) and accumulated 0.13 ± 0.00 g/liter arabitol after 216 h. Consistently, we also detected ARD, ADH, and XLR activities from YlSR102 cultures (Table 3). While the ARD and XLR activities were 2-fold higher for YlSR102 growing on mixed pentose sugars than for YlSR102 growing on single arabinose, the ADH activity remained almost the same. This result strongly suggests that Y. lipolytica has an active arabinose assimilation pathway but that arabitol dehydrogenase is likely a rate-limiting step.

FIG 6.

Profiles of cell growth and metabolites of Y. lipolytica strains YlSR102 (A), YlSR207 (B), YlSR223 (C), and YlSR157 (D) growing on mixed pentose sugars. Each data point represents an average value ±1 standard deviation from the results of 6 biological replicates.

In contrast to the enhanced arabinose assimilation, YlSR102 exhibited a 1.62- ± 0.04-fold lower xylose consumption rate for growth on mixed pentose sugars than for growth on single xylose (Table 2). To explain this phenotype, we analyzed the XYL2 and ADH activities of YlSR102 growing on single (control) and mixed sugars, since XYL2_Yli and ADH_Yli were encoded by the same gene, YALI0E12463g (Table 3). We observed that the XYL2 activities were relatively similar for growth on xylose and mixed pentose sugars but lower for growth on arabinose. The ADH activities were similar under all three conditions but significantly lower than the XYL2 activities. Taken together, these results suggest that the competition of XYL2 might have resulted in reduced xylose consumption when YlSR102 grew on mixed pentose sugars. Poor ADH activities were also likely responsible for inefficient arabinose assimilation. In some xylose-fermenting yeasts, such as Neurospora crassa and P. stipitis, xylose reductase (XYL1) has activity for both xylose and arabinose substrates (22, 23). Since xylose reductase of Y. lipolytica is 48.89% and 55.56% identical to those of P. stiptis (UniProt accession no. P31867) and N. crassa (UniProt accession no. Q7SD67), respectively, it is also plausible that the competition of xylose and arabinose substrates for this xylose reductase might have additionally reduced xylose assimilation efficiency.

Overall, the results showed that the native arabinose metabolism of Y. lipolytica is active and that arabinose utilization is synergistically enhanced by xylose. Xylose assimilation, however, is repressed by arabinose due to possible substrate competition of shared intermediate enzymes of the xylose and arabinose assimilation pathways. ADH might be a major bottleneck for efficient arabinose assimilation for growth.

(iii) Degradation of a mixture of xylose and arabinose by YlSR207 and YlSR223.

To further investigate the functional roles of TRP6_Yli and TRP22_Yli in activating pentose metabolism, we grew YlSR207 and YlSR223 on a mixture of xylose and arabinose. The results show that YlSR207 and YlSR223 grew much faster than YlSR102, with >2.7-fold increase in the specific growth rate and >4.4-fold increase in the biomass titer (Fig. 6B and C and Table 2). We also observed that the arabinose consumption of YlSR207 and YlSR223 was synergistically enhanced. YlSR207 and YlSR223 consumed 5.80 ± 1.58 g/liter and 6.14 ± 0.01 g/liter arabinose, respectively, with 1.15 ± 0.21 g/liter and 1.47 ± 0.08 g/liter arabitol accumulation, respectively.

Unlike YlSR102, both YlSR207 and YlSR223 completely consumed 10 g/liter xylose within 192 h, with 1.78-fold and 1.71-fold increases in specific xylose uptake rates, respectively. Strain YlSR207 achieved very similar specific xylose uptake rates when growing on either single xylose or mixed pentose sugars (Table 2). In contrast, the xylose-specific uptake rate of YlSR223 was 1.40-fold lower for growth on mixed pentose sugars than for growth on single xylose. Since arabinose assimilation is xylose dependent, it is difficult to directly compare the efficiencies of TRP6_Yli and TRP22_Yli in enhancing arabinose assimilation. To examine this, we first cultured YlSR207 and YlSR223 in single xylose (phase 1) until substrate completion, followed by addition of 10 g/liter arabinose (phase 2) (see Fig. S2 in the supplemental material). The results showed that there was no statistical significance of the specific arabinose consumption rates for YlSR207 (0.07 ± 0.02 mmol/g CDW/h) and YlSR223 (0.04 ± 0.01 mmol/g CDW/h).

Taken together, the results strongly validated the existence of the dormant pentose metabolism of Y. lipolytica. Overexpressing the native xylitol/arabitol dehydrogenase (XYL2_Yli) and pentose transporters (TRP6_Yli and TRP22_Yli) can activate pentose metabolism. Shared common intermediate enzymes and transporters of xylose- and arabinose-degrading pathways clearly affect the efficiency of pentose assimilation. The low activity of the endogenous arabitol dehydrogenase and cofactor imbalance likely hinder arabinose assimilation. TRP6_Yli and TRP22_Yli have in vivo activities toward both xylose and arabinose, and their overexpression enhances pentose assimilation. Like Y. lipolytica, other yeasts (except S. cerevisiae), including Candida arabinofermentans PYCC 5603T, Pichia guilliermondii PYCC 3012, and Ambrosiozyma monospora, also possess arabinose transporters (24, 25).

(iv) Alleviating the l-arabitol dehydrogenase rate-limiting step in Y. lipolytica.

A combination of bioinformatic and enzymatic analyses and arabitol accumulation clearly suggested that l-arabitol dehydrogenase is the rate-limiting step that causes poor arabinose assimilation in Y. lipolytica. To truly validate this, we constructed strain YlSR157, which overexpressed a heterologous arabitol dehydrogenase of Aspergillus oryzae (ADH_Aoz), using the constitutive TEF promoter. ADH_Aoz was chosen because it exhibited higher activity toward l-arabitol (39.2 mU/mg protein) than d-xylitol (6.54 mU/mg protein) (26).

Unlike YlSR102, YlSR157 grew faster and consumed more sugars when growing on both single arabinose and mixed pentose sugars (Fig. 4D and 6D). In single arabinose, YlSR157 reached a final optical density (OD) of 2.13 ± 0.02 and consumed 3.66 ± 0.13 g/liter arabinose with insignificant accumulation of arabitol after 193 h (Fig. 4D). For growth on mixed pentose sugars, YlSR157 achieved a final OD of 2.23 ± 0.31 and consumed a total of 4.80 ± 0.51 g/liter xylose and 3.40 ± 0.43 g/liter arabinose with insignificant accumulation of xylitol (0.27 ± 0.01 g/liter) and arabitol (0.08 g/liter ± 0.00 g/liter) after 193 h (Fig. 6D). Consistent with the growth phenotypes, YlSR102 yielded higher XYL2 activity than YlSR157 by 3.1-fold (see Fig. S3 in the supplemental material). While YlSR102 barely exhibited any ADH activity, YlSR157 achieved significantly (>50-fold) higher ADH activity. These results suggest that ADH_Aoz exhibited activities toward both d-xylitol and l-arabitol but that it has higher specificity for l-arabitol than for d-xylitol. Like XYL2_Yli, ARD_Aoz is specific for NAD⁺ only (data not shown). This cofactor imbalance might help explain incomplete arabinose assimilation when YlSR157 grew on either single arabinose or mixed pentose sugars.

Altogether, we validated the notion that arabitol dehydrogenase is the rate-limiting step in Y. lipolytica. While overexpressing ADH_Aoz enhanced arabinose assimilation due to its higher activity and specificity for arabitol, cofactor imbalance might have hindered efficient arabinose assimilation. Cofactor engineering will be critical to optimize arabinose assimilation in Y. lipolytica.

Conclusions.

We identified the two most promising pentose-specific transporters, TRP6_Yli and TRP22_Yli, which can enhance xylose consumption in Y. lipolytica, by screening a set of 16 putative transporters. Guided by bioinformatic, enzymatic, and transcriptomic analyses, we discovered that Y. lipolytica has the dormant native pentose metabolism to assimilate both xylose and arabinose. By overexpressing the native XYL2_Yli and TRP6_Yli (or TRP22_Yli), it is possible to activate this dormant pentose metabolism. Shared transporters and intermediate enzymes, such as XYL2_Yli (also ADH_Yli), can negatively affect efficient assimilation of pentose sugars. Due to low activity of arabitol dehydrogenase (ADH_Yli), it is considered to be the rate-limiting step and makes arabinose assimilation xylose dependent. Overall, this study sheds light on the cryptic pentose metabolism of Y. lipolytica and, further, helps guide strain engineering of Y. lipolytica for enhanced assimilation of pentose sugars.

MATERIALS AND METHODS

Plasmids and strains.

Y. lipolytica MYA-2613, a thiamine, leucine, and uracil auxotroph, was purchased from the American Type Culture Collection (ATCC) and used as the parent for genetic modification in our study. Table 1 shows the strains and plasmids used in this study. Table 4 lists the primers used to construct these plasmids and strains. The plasmid pSR008 was constructed by replacing the leucine selection marker gene (leu2) of the plasmid pSR001 (12) with the uracil selection marker gene (ura3). The ura3 gene was amplified from the genomic DNA (gDNA) of Y. lipolytica NRRL YB-423 using the primers URA3_Fwd/URA3_Rev and then assembled with the backbone amplified from pSR001 using the primers pSR008_Fwd/pSR008_Rev via the Gibson assembly cloning method (27).

TABLE 4.

Primers used in this study

Gene/protein/plasmid	Primer	Sequence
Primers used for constructing plasmids
Yli URA3	URA3_Fwd	CGGATATACTTGCTTGAATATACAGTAGTATGCGGCCGCTCGACGCCCAGAGAGCCATTG
	URA3_Rev	GTTTTTTGGGACACAAATACGCCGCCAACCCGGTCTCTCGAGAAACACAACAACATGC
pSR008	pSR008_Fwd	AGAGACCGGGTTGGCGGCGTATTTG
	pSR008_Rev	AGCGGCCGCATACTACTGTATATTCAAG
pSR001, pSR019–pSR039	pSR001_Fwd	TCATGTAATTAGTTATGTCACGCTTAC
	pSR001_Rev	TTTGAATGATTCTTATACTCAGAAG
trp1Yli	TRP1YL_Fwd	GTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGTCCACTAGTGCTATGACC
	TRP1YL_Rev	GGGCGTGAATGTAAGCGTGACATAACTAATTACATGACTAAGAGGACTCGGAGAAGTC
trp2Yli	TRP2YL_Fwd	CGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGGCCATTATTGTGGCTG
	TRP2YL_Rev	GCGTGAATGTAAGCGTGACATAACTAATTACATGACTAATCCGAATCAAATCCAGAATCG
trp4Yli	TRP4YL_Fwd	AGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGATTGGAAACGCTCAAATTAAC
	TRP4YL_Rev	GAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGATTACAATTGAGAGGGAGGG
trp6Yli	TRP6YL_Fwd	CGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGGGCTTCAGAGGCC
	TRP6YL_Rev	GGCGTGAATGTAAGCGTGACATAACTAATTACATGATTAAACATGTCTGGTTTCCTCTTG
trp7Yli	TRP7YL_Fwd	CGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGGGACGAAACTGGCTAG
	TRP7YL_Rev	GAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGATTAAGCTTGAGAAACGTTCTC
trp8Yli	TRP8YL_Fwd	CGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGTTTTCGTTAACGGGC
	TRP8YL_Rev	GAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGATTATACCGGAGGTTGAGGG
trp11Yli	TRP11YL_Fwd	CGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGAAAGACTTCCTCGCC
	TRP11YL_Rev	GAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGACTACGCTGTCTCGATTCGAAC
trp12Yli	TRP12YL_Fwd	GCATTTCCTTCTGAGTATAAGAATCATTCAAAATGATACTTTTTTGGTTACACAGAG
	TRP12YL_Rev	GAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGATTATTGATGAGTGGTGGTGTC
trp13Yli	TRP13YL_Fwd	GTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGAAGCTGTTTAAACGAGAAGC
	TRP13YL_Rev	GAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGACTATCCACGAATAGTGGCAC
trp14Yli	TRP14YL_Fwd	TCGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGTCGCTGGACAAAAACC
	TRP14YL_Rev	GGGCGTGAATGTAAGCGTGACATAACTAATTACATGACTACTTCTTGTAGCCTCTCTTG
trp15Yli	TRP15YL_Fwd	CGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGTCGTCTATATCTTCGTCC
	TRP15YL_Rev	GAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGACTACATGGTCCAAACCTCGG
trp16Yli	TRP16YL_Fwd	CGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGTCTTCCTATCCATCCGAG
	TRP16YL_Rev	GGGGAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGATTAAGCAAGCTCCGCCG
trp17Yli	TRP17YL_Fwd	CGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGGTTTTTGGACGAGAAAAAG
	TRP17YL_Rev	GGGAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGATTAAACGAACTCGGCAGTG
trp19Yli	TRP19YL_Fwd	TAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGTTCTGGAAGAACATGAAAAATG
	TRP19YL_Rev	CGTGAATGTAAGCGTGACATAACTAATTACATGATTAACAATTCTCCACATGAATAACAC
trp21Yli	TRP21YL_Fwd	TAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGTTCTGGAAAAACATGAAGAATG
	TRP21YL_Rev	GAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGACTAACAGGTCTCCACGTGAAC
trp22Yli	TRP22YL_Fwd	CGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGTCGCACCGGCCCTG
	TRP22YL_Rev	GAGGGCGTGAATGTAAGCGTGACATAACTAATTACATGATCACCTATCAGCATTTTCACC
trp23Yli	TRP23YL_Fwd	ATCGTTAAGCATTTCCTTCTGAGTATAAGAATCATTCAAAATGTCAATCAAGTCGCTCTC
	TRP23YL_Rev	AGGGCGTGAATGTAAGCGTGACATAACTAATTACATGACTAGACACCATCTTTAGCAACC
Primers used in RT-PCR
L-ARD1	ARD1_rt Fwd	AGCTGCAGACCGACTACGTT
	ARD1_rt Rev	GCCTCCCAAGTCTCCTCGAC
L-ARD2	ARD2_rt Fwd	CGACACATTGACACCGCCTT
	ARD2_rt Rev	CGGTCGTGGTAAGTGACCCA
L-ARD3	ARD3_rt Fwd	AGCCCAAGACCTGGAAGCAA
	ARD3_rt Rev	CCTGGTTGACAGCAGGGACA
L-ARD4	ARD4_rt Fwd	GGGTGTCAAGCGAGAGGACA
	ARD4_rt Rev	GGCCAGTGGATGAGCAGCAT
L-ARD5	ARD5_rt Fwd	CGTCGACATGCTGCTGATCC
	ARD5_rt Rev	GCCTCCATCTGCTTCCAGGT
L-ARD6	ARD6_rt Fwd	CGCCATCAAGGACGGCTACA
	ARD6_rt Rev	ACAAACAGCTCCTCCCGCTT
L-ARD7	ARD7_rt Fwd	CCCATCGCTGTGGAAATCGC
	ARD7_rt Rev	CCGTTGGCTACAACACGTTCA
L-ARD8	ARD8_rt Fwd	TACTCCTTCCTGCCCAAGCG
	ARD8_rt Rev	CGTTGGACACGCCAATGGAC
L-ARD9	ARD9_rt Fwd	TCGCCTCTAGCTTCCAACGA
	ARD9_rt Rev	TGCCCTTCTGAGTAGCCCATC
L-ARD10	ARD10_rt Fwd	GGCAATGAAGCTGCCACTTGT
	ARD10_rt Rev	AGCTGCATGGTGCCCTCATA
L-ARD11	ARD11_rt Fwd	GGGCATCTCCGAATGGCTGA
	ARD11_rt Rev	CTTGGCCTTGGTCAGAGCCT
L-XLR1	XLR1_rt Fwd	AACGCCAAGCAGGTCGAGA
	XLR1_rt Rev	CCTTGTCCCACTCCTCGTTGT
L-XLR2	XLR2_rt Fwd	GGCTGAGCACCTCGCTAAGA
	XLR2_rt Rev	GCGTTGGCGACGAAGATGTC
L-XLR3	XLR3_rt Fwd	GCAGGACCTGGGTCTGGATG
	XLR3_rt Rev	CGTGCACAACTCGGTCGATG
L-XLR4	XLR4_rt Fwd	GTGGTGGTCTGGGTATCGCA
	XLR4_rt Rev	CGTTGGCGGTCTTGTACCAC
L-XLR5	XLR5_rt Fwd	TTAAGCGAATGCCCGGCAAC
	XLR5_rt Rev	CCGGCGTTAGCGATGACAAC
Primer used to confirm Y. lipolytica strains (PTEF)	PTEF_seq Fwd	GTGGTTGGGACTTTAGCCAAGGG

A set of 16 plasmids containing 16 putative xylose transporter genes identified previously were built (15). Specifically, these transporter genes were amplified from gDNA of YlSR001 using the primer sets shown in Table 4 and assembled with the pSR008 backbone amplified using the primers pSR001_Fwd/pSR001_Rev. Gene locus information for the 16 xylose-specific transporters of native Y. lipolytca are available in Table S1 in the supplemental material. To construct pSR041, we used the heterologous l-arabitol dehydrogenase gene of A. oryzae (GenBank accession no. BAC81768.1), which was codon optimized and synthesized by the gBlocks gene fragment synthesis service from Integrated DNA Technologies, Inc., Coralville, IA (IDT), and then Gibson assembled with the pSR001 backbone (15) amplified using the primers pSR001_Fwd/pSR001_Rev. The codon-optimized sequence of ADH_Aoz is presented in Text S1 in the supplemental material.

Escherichia coli TOP10 was used for molecular cloning. All Y. lipolytica mutants (Table 1) were constructed by transforming YlSR001, a thiamine, leucine, and uracil auxotroph, with the target plasmids via electroporation (28). All constructed Y. lipolytica mutants were screened and maintained on SC−Leu−Ura agar plates containing 10 g/liter xylose as a carbon source. The mutants were verified by performing yeast colony PCR using a TEF promoter binding forward primer (P_TEF_seq Fwd) and gene binding reverse primer (TRPX_Yli_Rev) (Table 4). In Y. lipolytica, genes were expressed under the TEF (404) promoter (29).

Media and cell culturing. (i) Medium preparation.

Lysogeny broth (LB) medium containing 5 g/liter yeast extract, 10 g/liter tryptone, and 5 g/liter NaCl was used for molecular cloning in E. coli. For selection, 100 μg/ml ampicillin was added to the LB medium. For Y. lipolytica characterization experiments, synthetic SC−Leu−Ura medium (pH 5.5) containing yeast nitrogen base (catalog no. Y0626; Sigma-Aldrich, St. Louis, MO, USA); a synthetic dropout amino acid mixture without leucine, uracil, and tryptophan (catalog no. Y1771 [Sigma-Aldrich] or catalog no. AC172110250 [Fisher Scientific, Waltham, MA, USA]); and l-tryptophan (catalog no. AC140590010; Acros Organics, NJ, USA) was used. An additional 30 μg/ml chloramphenicol was added to the medium to prevent bacterial contamination (30). Single sugars or a mixture of sugars, including glucose, xylose, and/or arabinose, were also added to the yeast synthetic medium at an initial concentration of 10 g/liter.

(ii) Isolation of efficient xylose-utilizing mutants.

To isolate the efficient xylose-utilizing mutants, each of the 16 Y. lipolytica mutants, YlSR202, YlSR203, YlSR205, YlSR207 to 209, YlSR212 to 218, YlSR220, YlSR222, and YlSR223, carrying the native XYL2 and putative xylose transporter genes, were first grown in a 15-ml culture tube containing 1 ml SC−Leu−Ura medium with 10 g/liter glucose as a carbon source. When cell growth reached mid-exponential phase (OD at 600 nm, ∼3), the cells were spun down and the cell pellet was washed with 5 ml sterile water three times to remove any residual glucose. Next, the pellet was resuspended in SC−Leu−Ura medium without sugars and diluted 10×, 100×, and 1,000× with the same medium. The diluted cultures of each mutant were then spotted on an SC−Leu−Ura plate containing 10 g/liter xylose. Mutants exhibiting fast growth on xylose were selected for subsequent characterization studies.

(iii) Cell culturing.

Growth characterization of Y. lipolytica in yeast synthetic media containing single sugars or a mixture of xylose and/or arabinose was conducted in a 500-ml baffled flask with a 50-ml working volume at 28°C and 300 rpm. The initial OD was adjusted to 0.2. Time profiles of cell growth and metabolites were measured during cell culturing and used to determine kinetic parameters, such as the specific cell growth rate (μ; 1/h) and specific sugar uptake rate (r_S; mmol/g CDW/h) (31). The gene expression and enzyme activity were also measured for cell cultures collected during the mid-exponential growth phase. All experiments were performed with at least 6 biological replicates.

Bioinformatics.

BLASTP (32) was applied to identify the putative ARD, ADH, and XLR enzymes of the arabinose assimilation pathway of Y. lipolytica. We used ARD_Ani (from Aspergillus niger; UniProt accession no. A2QBD7 [33]), ADH_Ncr (from N. crassa; UniProt accession no. Q7SI09 [34]), NADPH-dependent XLR_Ani (UniProt accession no. G3YG17 [35]), NADPH-dependent XLR_Tre (from Trichoderma reesei; UniProt accession no. G0RH19 [39]), and NADH-dependent XLR_Amo (from Ambrosiozyma monospora; UniProt accession no. Q70FD1 [36]) as BLASTP templates.

Analytical methods. (i) HPLC.

Extracellular metabolites (e.g., sugars and sugar alcohols) were analyzed using a Shimadzu high-performance liquid chromatography (HPLC) system equipped with a refractive index detector (RID) (Shimadzu Scientific Instruments, Inc., MD, USA). Prior to an HPLC run, the culture medium was filtered through a 0.2-μm-pore-size membrane filter. To analyze glucose, we used a Bio-Rad Aminex 87H column (catalog no. 1250140; Bio-Rad Laboratories, CA, USA) set at 48°C and 10 mN H₂SO₄ as a mobile phase operating at 0.6 ml/min (15). To analyze xylose, xylitol, arabinose, and arabitol, we used a Bio-Rad Aminex 87C column (catalog no. 1250095; Bio-Rad Laboratories, CA, USA) set at 85°C and water as a mobile phase running at 0.6 ml/min.

(ii) Transcriptomics by RT-PCR.

Gene expression levels were quantified using real-time PCR (RT-PCR). First, mid-exponential growth phase cells (ODs in a range from 2.0 to 2.5) were harvested. Total RNA was purified using a Qiagen RNeasy minikit (catalog no. 74104; Qiagen Inc., CA, USA), and cDNA was synthesized using a QuantiTect reverse transcription kit (catalog no. 205311; Qiagen Inc., CA, USA). The RT-PCR experiment was performed using a QuantiTect SYBR green PCR kit (catalog no. 204143; Qiagen Inc., CA, USA) and a StepOnePlus real-time PCR system (Applied Biosystems, CA, USA). A relative mRNA expression level of a target gene under a given condition was normalized to that of a housekeeping actin gene (YALI0D08272g), as previously described (15). The primers used for RT-PCR are listed in Table 4. To compare relative gene expression levels of a target gene under two different conditions, we calculated the log₂ ratio of the expression levels for the target gene between condition 1 (e.g., growth on arabinose) and condition 2 (e.g., growth on xylose as a reference condition) (37). A relative mRNA expression level for each gene under a given growth condition was reported as an average ±1 standard deviation from a data set of at least three biological replicates. Student's t test was performed to evaluate statistical significance.

(iii) In vitro enzyme activity assays.

To prepare cell lysates, cell cultures were collected during the mid-exponential growth phase (ODs between 2.0 and 2.5) and suspended in Y-PER yeast protein extraction reagent (catalog no. 78990; Thermo Scientific, IL, USA) including 2× EDTA-free Pierce protease inhibitors (catalog no. 88266; Thermo Scientific, IL, USA). The cells were then lysed by incubation at 28°C and 300 rpm for 60 min. The soluble fraction was separated by centrifugation at 17,000 × g for 10 min and used for enzyme activity assays.

Each in vitro enzyme activity assay for ARD, XYL2, ADH, and XLR was conducted by adding 1.0 μg of whole-cell lysate to a reaction mixture containing 25 mM sodium phosphate buffer (pH 6.0), an appropriate cofactor [e.g., NAD(P)H or NAD(P)⁺], and substrate (e.g., l-arabinose, d-xylitol, l-arabitol, or l-xylulose) in 384-well plates with a 70-μl working volume at 28°C. For the ARD assay, 1 mM NAD(P)H and 300 mM l-arabinose were used, and the reduction rate of NAD(P)H was monitored at 340 nm. For the XYL2 and ADH assays, 1 mM NAD(P)⁺ and 300 mM substrate (i.e., xylitol for XYL2 and arabitol for ADH) were used. For the XLR activity assay, 1 mM NAD(P)H and 30 mM l-xylulose were used. One unit of each enzyme activity was defined as 1 μmol of NAD(P)H generated or reduced per mg protein per min.

Enzyme kinetics was measured using a BioTek Synergy HT microplate reader with associated Gen5 software (BioTek Instruments, Inc., VT, USA). The protein concentration was quantified by the Bradford method (38). All enzyme assay experiments were performed with at least three biological replicates.