Identiﬁcation of Mutations Responsible for Improved Xylose Utilization in an Adapted Xylose Isomerase Expressing Saccharomyces cerevisiae Strain

: Economic conversion of biomass to biofuels and chemicals requires efﬁcient and complete utilization of xylose. Saccharomyces cerevisiae strains engineered for xylose utilization are still considerably limited in their overall ability to metabolize xylose. In this study, we identiﬁed causative mutations resulting in improved xylose fermentation of an adapted S. cerevisiae strain expressing codon-optimized xylose isomerase and xylulokinase genes from the rumen bacterium Prevotella ruminicola . Genome sequencing identiﬁed single-nucleotide polymorphisms in seven open reading frames. Tetrad analysis showed that mutations in both PBS2 and PHO13 genes were required for increased xylose utilization. Single deletion of either PBS2 or PHO13 did not improve xylose utilization in strains expressing the xylose isomerase pathway. Saccharomyces can also be engineered for xylose metabolism using the xylose reductase/xylitol dehydrogenase genes from Scheffersomyces stipitis . In strains expressing the xylose reductase pathway, single deletion of PHO13 did show a signiﬁcant increase xylose utilization, and further improvement in growth and fermentation was seen when PBS2 was also deleted. These ﬁndings will extend the understanding of metabolic limitations for xylose utilization in S. cerevisiae as well as understanding of how they differ among strains engineered with two different xylose utilization pathways.


Introduction
Biomass-derived sugars such as glucose and xylose can be metabolized by many microorganisms. However, most of them have complex genetic systems that are not amenable to the extensive genome modification required for metabolic engineering to produce biofuels and organic acids at high yields and productivities. Toxicity of lignocellulosic hydrolysates and/or end-product inhibition is also an issue with many microorganisms. Brewer's yeast, Saccharomyces cerevisiae, has been used extensively due to its ease of genetic modification, availability of a variety of promoters and terminators for metabolic engineering, and ability to grow anaerobically at low pH (reviewed in [1]). S. cerevisiae does not naturally metabolize xylose, and to overcome this limitation, strains have been engineered to express three different routes for xylose utilization [2][3][4]. Oxidative metabolism of xylose using the bacterial Weimberg pathway has recently been shown in S. cerevisiae [5]. For the two main pathways investigated in this study (oxido-reductive and isomerization), xylose is converted to xylulose-5P (X5P), which enters central metabolism through the pentose phosphate pathway (PPP) (Figure 1). The reductase/dehydrogenase pathway (XR/XDH) uses xylose reductase and xylitol dehydrogenase, typically from Scheffersomyces stipitis, which can lead to a redox imbalance under anaerobic conditions [6,7]. Alternatively, expression of xylose isomerase (XI) is independent of cofactors and has potential to avoid this redox imbalance [7,8]. Multiple XI genes from bacteria and anaerobic fungi have been used to metabolically engineer xylose utilization in S. cerevisiae. In comparison to strains Numerous studies have shown it is possible to increase xylose utilization in engineered strains by adaptive laboratory evolution. However, many of these studies do not identify the causative mutations [18][19][20][21][22][23][24][25][26][27][28][29][30]. Studies that do identify genetic changes in strains evolved for increased xylose fermentation overwhelmingly show changes associated with increased activity of the PPP, irrespective of route used to convert xylose to X5P. Loss of PHO13 function is commonly seen in adapted strains, as PHO13 deletion increases PPP activity [15,31,32]. Aside from increasing PPP activity, PHO13 deletion also appears to alleviate ATP depletion, which is putatively caused by futile cycling around xylulose and X5P [33]. Many adaptive evolution studies aimed at improving xylose metabolism also start with strains overexpressing either the PPP gene TAL1 (encodes transaldolase) or overexpressing multiple PPP genes. For genetic studies on strains overexpressing PPP genes, mutation to PHO13 was either not identified [14,17,[34][35][36][37][38][39], or, PHO13 deletion was shown to negatively affect xylose utilization [22,40]. Conversely, strains without PPP overexpression often identify PHO13 mutations or deletion as beneficial [15,22,[31][32][33].
When starting with low-copy xylose isomerase, adapted strains frequently show increased copy number of the XI gene. Starting with the Brazilian industrial ethanol production strain PE-2, dos Santos et al. (2016) integrated Orpinomyces XI and extra XKS1 copies, as well as integrated PPP genes (TAL1, TKL1, RPE1, RKI1), and deleted GRE3 [35]. Genome sequencing of the evolved strain revealed approximately 26 copies of the XI gene, as well as additional mutations in ISU1 and SSK2. Several other studies with adapted strains also identified increased XI gene copy number ranging up to 36-fold [14,36,39,41], highlighting the need for increased XI activity for growth on xylose.
Previously, we discovered a xylose isomerase from the rumen bacterium Prevotella ruminicola TC2-24 and expressed both the XI and XK enzymes from this bacterium in S. cerevisiae and evaluated its ability to grow on xylose [16]. Xylose fermentation was initially poor, and the specific growth rate for aerobic xylose cultures was 0.06 h −1 ± 0.005. Adaptive laboratory evolution was performed for ca. 40 generations under microaerobic conditions and the resulting strain, YRH1114, had a 3.8-fold increase in aerobic growth rate to 0.23 h −1 ± 0.024. Under anaerobic conditions, xylose consumption and final ethanol concentration increased 2.7-fold and 3.3-fold, respectively, in the adapted strain. Xylitol yield also decreased 1.7-fold and ethanol yield increased 1.2-fold, suggesting that xylose was more efficiently fermented to ethanol in the adapted strain. Numerous studies have shown it is possible to increase xylose utilization in engineered strains by adaptive laboratory evolution. However, many of these studies do not identify the causative mutations [18][19][20][21][22][23][24][25][26][27][28][29][30]. Studies that do identify genetic changes in strains evolved for increased xylose fermentation overwhelmingly show changes associated with increased activity of the PPP, irrespective of route used to convert xylose to X5P. Loss of PHO13 function is commonly seen in adapted strains, as PHO13 deletion increases PPP activity [15,31,32]. Aside from increasing PPP activity, PHO13 deletion also appears to alleviate ATP depletion, which is putatively caused by futile cycling around xylulose and X5P [33]. Many adaptive evolution studies aimed at improving xylose metabolism also start with strains overexpressing either the PPP gene TAL1 (encodes transaldolase) or overexpressing multiple PPP genes. For genetic studies on strains overexpressing PPP genes, mutation to PHO13 was either not identified [14,17,[34][35][36][37][38][39], or, PHO13 deletion was shown to negatively affect xylose utilization [22,40]. Conversely, strains without PPP overexpression often identify PHO13 mutations or deletion as beneficial [15,22,[31][32][33].
When starting with low-copy xylose isomerase, adapted strains frequently show increased copy number of the XI gene. Starting with the Brazilian industrial ethanol production strain PE-2, dos Santos et al. (2016) integrated Orpinomyces XI and extra XKS1 copies, as well as integrated PPP genes (TAL1, TKL1, RPE1, RKI1), and deleted GRE3 [35]. Genome sequencing of the evolved strain revealed approximately 26 copies of the XI gene, as well as additional mutations in ISU1 and SSK2. Several other studies with adapted strains also identified increased XI gene copy number ranging up to 36-fold [14,36,39,41], highlighting the need for increased XI activity for growth on xylose.
Previously, we discovered a xylose isomerase from the rumen bacterium Prevotella ruminicola TC2-24 and expressed both the XI and XK enzymes from this bacterium in S. cerevisiae and evaluated its ability to grow on xylose [16]. Xylose fermentation was initially poor, and the specific growth rate for aerobic xylose cultures was 0.06 h −1 ± 0.005. Adaptive laboratory evolution was performed for ca. 40 generations under microaerobic conditions and the resulting strain, YRH1114, had a 3.8-fold increase in aerobic growth rate to 0.23 h −1 ± 0.024. Under anaerobic conditions, xylose consumption and final ethanol concentration increased 2.7-fold and 3.3-fold, respectively, in the adapted strain. Xylitol yield also decreased 1.7-fold and ethanol yield increased 1.2-fold, suggesting that xylose was more efficiently fermented to ethanol in the adapted strain.

Growth Analysis
Cells were grown in xylose medium using the Bioscreen C™ automated microbiology growth curve analysis system (Growth Curves USA; Piscataway, NJ, USA), which features 100 micro-well culture plates. Growth assays were performed essentially as described in [16]. Each strain was analyzed in at least quadruplicate using separate biological replicates. Cell mass (OD 600 ) was determined by converting the wideband OD (WB) values from the Bioscreen C™ as described in [16].

Genome Sequencing and Analysis
Genomic DNA isolation, library preparation, and sequencing were performed as previously described [48]. In short, sequencing libraries were constructed from isolated genomic DNA using the Nextera XT Library prep kit following the manufacturer protocol (Illumina, San Diego, CA, USA). The library samples were processed for sequencing and loading using the MiSeq Reagent Kit v3 (Illumina; San Diego, CA, USA) and run per the manufacturer protocol on the MiSeq system with a maximum read length of 2 × 300 bp. Sequencing reads were trimmed to eliminate adaptors, low quality (Q < 20) bases, and ambiguous nucleotides, and reads were filtered to remove bacterial and human DNA contaminants using CLC Genomics Workbench version 20.0.4 (Qiagen; Germantown, MD, USA). Sequence read depth for strains CEN.PK2-1C, YRH1114, and YRH1136 was 85X, 63X, and 30X, respectively. Reads were deposited in the GenBank SRA database under BioProject number PRJNA877627. Trimmed reads of the unadapted CEN.PK2-1C reference strain, the adapted strain containing the plasmid-based xylose utilization pathway, YRH1114, and the strain cured of the xylose pathway plasmids, YRH1136, were mapped to the annotated genome sequence of the Saccharomyces cerevisiae reference strain, CEN.PK113-7D (NCBI BioProject PRJNA393501) using the CLC Genomics Workbench default parameters. Duplicate reads were removed, and local realignments were performed to improve mapping accuracy. Variants (i.e., SNPs and INDELs) of each mapped strain were called using the default parameters in CLC Genomics Workbench. The variants of YRH1114 and YRH1136 were then filtered to remove the variants found in the wild-type CEN.PK2-1C strain. The variant list was then filtered to remove variants with low coverage resulting in 158 variants in the adapted strains. Mutations with potential functional consequence were determined using the GO Enrichment Analysis tool in CLC Genomics Workbench package, leaving a final list of 7 variants.

Mating-Type Testing and MAT Switching
The plasmid pGAL-HOT, containing the HO gene behind a galactose-induced promoter and TRP1 marker, was used to switch the mating type of the haploid strains. After growing overnight in YP + sucrose (20 g/L), YRH1911 cells were resuspended in YP + galactose (20 g/L) and incubated (30 • C, 250 rpm). After 4 h of induction, cells were diluted and plated to obtain approximately 100-200 cells per YPD plate. Mating type was confirmed by multiplex PCR as described in [49]. A MATa isolate of the adapted strain was allowed to lose the pGAL-HOT vector, by passage without selection, to generate YRH1954. YRH1955 was generated by transforming YRH1954 with an empty pRS413 (HIS3) to facilitate diploid selection. Mating pairs (YHR1955 X YRH631 and YRH1955 X YRH1114) were then patched together on YPD plates to allow mating. After incubation for 6 h at 30 • C, cells were transferred to SD-Ura-Trp-His plates to select only for diploid cells. Mating type was again confirmed by multiplex PCR as described in [49], with diploids showing the presence of both MATa and MATα specific genes at the mating type locus.

Tetrad Dissection
The heterozygous diploid strain YRH1981 (from YRH1955 X YRH631 mating) was grown overnight in synthetic complete medium with 20 g/L glucose (SD-Ura-Trp) to an OD 6000 .8. Cells were washed once with sterile water and resuspended in SD-Ura-Trp and allowed to grow overnight in a roller drum at 30 • C. Cells were washed twice with sterile water and resuspended in 1% potassium acetate sporulation medium and incubated at 30 • C for 3 to 5 days. Sporulated cells (500 mL) were washed twice with sterile water and resuspended with 50 uL of digestion buffer (1 mL of 1M sorbitol and 0.5 µL of zymolyase (5 U/mL), ZymoResearch; Irvine, CA, USA). Tetrads were incubated in digestion buffer at 37 • C. After 30 min, 1 mL of sterile water was added, and the tetrads were stored on ice to stop digestion. A 20 µL volume of digested tetrads was transferred to an SD-Ura-Trp plate for dissection. Tetrads were dissected using a SporePlay dissection microscope (Singer Instruments; Somerset, UK). Plates containing the spores were incubated at 30 • C. Each haploid was tested to determine mating type, SNP presence, and ability to grow on xylose medium.

SNP PCR
The presence or absence of each SNP was determined using a one-step PCR assay as described in [50]. Primers were designed with a common mismatched nucleotide at position −2. Primers for amplification of the wild-type and SNP variants have the SNP mutation as the final 3 nucleotide in the primer (position 0). The wild-type primer results in a product only if the wild-type gene is present. Conversely, the SNP primer only results in a PCR product from the mutated gene. Two separate PCR assays using wild-type or SNP primers with a common reverse primer were performed for each haploid from tetrad dissection.

Batch Fermentation
Xylose fermentation was performed using 50 mL cultures of YP medium supplemented with 50 g/L xylose in a 125 mL Erlenmeyer flask incubated at 30 • C with shaking at 150 RPM for 4 days. Cells for inoculation were grown to mid-log phase in either SD-Ura-Trp or SD-Ura medium, depending on the strain, to maintain selection for plasmids. YP5X cultures were inoculated to a starting OD 600 of~1.0. Flasks were sealed with a rubber stopper pierced with a 20-gauge needle with glass wool placed at one end to restrict ingress of air and enable release of CO 2 . Samples were taken at various timepoints to determine cell mass (OD 600 ), xylose, and fermentation products (by high-performance liquid chromatography, HPLC as previously described [51]). Fermentation experiments were performed using three biological repeats and all fermentation data calculations (i.e., yields, rates, and carbon recoveries) were performed as previously described [16]. Probability analyses were performed using Student's t test with a two-tailed distribution. Values with p < 0.05 were considered significant for this study. Statistical analysis was performed using Microsoft Excel. Cell dry weight (CDW) was calculated using an OD-to-CDW conversion factor for the yeast strains used in this study (CDW haploid = 0.65 ± 0.003 g/L/OD 600 , CDW diploid = 0.73 ± 0.007 g/L/OD 600 ). The conversion factor was determined by drying cells at differing OD to constant weight at 100 • C. Cells were washed two times with distilled water prior to drying. OD 600 was measured using a BioMate 3S spectrophotometer (Thermo Fisher Scientific Inc.; Waltham, MA, USA).

Genome Sequence Analysis of the Adapted Strain
Gene amplification is commonly observed in adapted strains [52][53][54], and several studies starting with integrated xylose isomerase indicate that increased copy number of the gene is essential for increased xylose utilization [14,35,36,39,41]. In this study, high copy number was gained by expressing the XI gene from a high-copy number plasmid. XK was also expressed from a separate low-copy plasmid because, while elevated XK levels are required to increase xylose fermentation, excess expression of XK in several studies has been shown to be detrimental to xylose (and xylulose) fermentation [55][56][57][58][59].
First, we determined if the XI gene was integrated into the genome. Starting with the adapted YRH1114 strain, progeny strains were generated that had lost one or both vectors for expressing either the XI or XK genes. Strains that lost the high-copy XI vector also lost the ability to grow on xylose, suggesting that copies of the XI gene did not integrate into the genome. Increased ability to grow on xylose was observed when new plasmids for expressing the XI and XK were reintroduced into the adapted strain lacking its original XI/XK plasmids. DNA sequencing of the rescued expression plasmids also confirmed that mutations were not present in the plasmid sequence. The sum of these results indicated that the causative mutations required for increased growth on xylose occurred within the native genome and not as a result of genome integrated XI or mutation(s) occurring within the expression vectors. Aside from expected regions of high coverage (i.e., rDNA, Ty elements, and sub-telomeric), genome sequencing and read-depth analysis of the adapted YRH1114 strain did not reveal additional regions with increased copy number. Further analysis for insertions/deletions (INDELs) and single-nucleotide polymorphisms (SNPs) did show the presence of seven SNPs located in open reading frames that were not present in the CEN.PK2-1C genome (Table 3). Of the seven SNPs identified in YRH1114, two genes had been discovered in other adapted strains. Transposon mutagenesis and selection for improved growth on xylose led to the discovery of strains with transposons inserted into the PHO13 gene [32]. Separately, Kim et al., (2013) identified two additional PHO13 mutations (i.e., pho13 G166R and pho13 G253D ) in strains evolved for improved xylose fermentation [33]. In both studies, the PHO13 mutations were hypothesized to result in lost or decreased Pho13p function, and analysis of strains with targeted deletion of PHO13 supported their hypotheses. Additional studies of strains with targeted PHO13 deletion showed increased carbon flux through the PPP, which is essential for xylose metabolism [15,31,32]. Like the previously discovered pho13 mutant alleles mentioned above, the pho13 G208C mutation identified in our adapted strain resulted in an altered glycine residue. Glycine residues influence the formation of short loops [60]. Further inspection of the location of the pho13 G166R , pho13 G253D , and pho13 G208C mutations with respect to predicted protein structure [61,62] showed that the mutated glycine residues were all located at the start of a short loop between two secondary protein structures. It is possible that mutation of the glycine residues in these alleles disrupts loop formation, resulting in altered protein folding and decreased function. Regarding the other gene discovered in another adapted strain, ask10 M475R and ask10∆ mutations were identified that resulted in upregulated HSP104 levels [37]. Both ask10 mutants, as well as HSP104 overexpression, were suggested to increase XI activity by facilitating protein folding [37]. The five other genes listed in Table 3 have not been previously associated with xylose metabolism.

Determination of Phenotype Dominance
To identify which of the seven mutations were phenotypic, we generated heterozygous (unadapted/adapted) diploids to first determine if improved xylose utilization was dominant or recessive. The mating type of the adapted haploid strain YRH1114 was switched from MATa to MATα and the adapted MATα strain was mated to an unadapted parent strainYRH631 (MATa) to create the heterozygous diploid strain YRH1981. The MATα version of YRH1114 was mated to the MATa YRH1114 strain to generate the homozygous diploid strain YRH1982. Both diploid strains were grown on xylose and compared to the original unadapted and adapted haploid strains ( Figure 2). The heterozygous diploid YRH1981 grew as poorly as the unadapted YRH631 strain and the homozygous diploid YRH1982 grew nearly as well as the adapted haploid strain, indicating that the genetic changes were recessive.
switched from MATa to MATα and the adapted MATα strain was mated to an unadapted parent strainYRH631 (MATa) to create the heterozygous diploid strain YRH1981. The MATα version of YRH1114 was mated to the MATa YRH1114 strain to generate the homozygous diploid strain YRH1982. Both diploid strains were grown on xylose and compared to the original unadapted and adapted haploid strains ( Figure 2). The heterozygous diploid YRH1981 grew as poorly as the unadapted YRH631 strain and the homozygous diploid YRH1982 grew nearly as well as the adapted haploid strain, indicating that the genetic changes were recessive. ) and heterozygous diploid (YRH1981 ) strains were grown on YP medium + 50 g/L xylose and compared to the unadapted haploid (YRH631 ) and adapted haploid (YRH1114 ) strains. Cell dry weight was used to measure cell mass instead of OD600 to account for differences in dry weight per OD600 between haploid and diploid strains. Plots are the average values from at least triplicate cultures. Error bars show standard deviations.

Tetrad Dissection to Identify Causative Mutations
Most of the SNPs were located on different chromosomes, and mutations located on the same chromosome appeared on opposite arms of the chromosome (Table 3). Since none of the SNPs were closely linked, haploid spores from the heterozygous diploid were analyzed to determine which SNPs were associated with improved growth on xylose medium. Heterozygous diploid strain YRH1981 was sporulated and tetrads with re-assorted wild-type and SNP combinations were dissected onto SD-Ura-Trp plates to maintain selection for the XI and XK expression plasmids ( Figure 3A). Growth of the haploids was compared in xylose cultures ( Figure 3B). The presence of the expected SNP in each haploid was confirmed using a PCR-based SNP detection assay; the results of the tetrad analysis are listed in Table 4.
Each haploid strain from the tetrad analysis that grew well in xylose medium carried both pho13 G208C and pbs2 L363X mutations; these mutations were identified as likely candidates for increased growth on xylose seen in these strains. None of the haploid strains carrying just one of these mutations was able to grow as well on xylose medium, further suggesting that both pho13 G208C and pbs2 L363X mutations are required. The pbs2 L363X mutation results in leucine-to-stop change in the middle of protein, likely rendering it non-functional. Additionally, since loss-of-function mutation, or deletion, of PHO13 has previously switched from MATa to MATα and the adapted MATα strain was mated to an unadapted parent strainYRH631 (MATa) to create the heterozygous diploid strain YRH1981. The MATα version of YRH1114 was mated to the MATa YRH1114 strain to generate the homozygous diploid strain YRH1982. Both diploid strains were grown on xylose and compared to the original unadapted and adapted haploid strains ( Figure 2). The heterozygous diploid YRH1981 grew as poorly as the unadapted YRH631 strain and the homozygous diploid YRH1982 grew nearly as well as the adapted haploid strain, indicating that the genetic changes were recessive. ) and heterozygous diploid (YRH1981 ) strains were grown on YP medium + 50 g/L xylose and compared to the unadapted haploid (YRH631 ) and adapted haploid (YRH1114 ) strains. Cell dry weight was used to measure cell mass instead of OD600 to account for differences in dry weight per OD600 between haploid and diploid strains. Plots are the average values from at least triplicate cultures. Error bars show standard deviations.

Tetrad Dissection to Identify Causative Mutations
Most of the SNPs were located on different chromosomes, and mutations located on the same chromosome appeared on opposite arms of the chromosome (Table 3). Since none of the SNPs were closely linked, haploid spores from the heterozygous diploid were analyzed to determine which SNPs were associated with improved growth on xylose medium. Heterozygous diploid strain YRH1981 was sporulated and tetrads with re-assorted wild-type and SNP combinations were dissected onto SD-Ura-Trp plates to maintain selection for the XI and XK expression plasmids ( Figure 3A). Growth of the haploids was compared in xylose cultures ( Figure 3B). The presence of the expected SNP in each haploid was confirmed using a PCR-based SNP detection assay; the results of the tetrad analysis are listed in Table 4.
Each haploid strain from the tetrad analysis that grew well in xylose medium carried both pho13 G208C and pbs2 L363X mutations; these mutations were identified as likely candidates for increased growth on xylose seen in these strains. None of the haploid strains carrying just one of these mutations was able to grow as well on xylose medium, further suggesting that both pho13 G208C and pbs2 L363X mutations are required. The pbs2 L363X mutation results in leucine-to-stop change in the middle of protein, likely rendering it non-functional. Additionally, since loss-of-function mutation, or deletion, of PHO13 has previously ) and heterozygous diploid (YRH1981 switched from MATa to MATα and the adapted MATα strain was mated to an unadapted parent strainYRH631 (MATa) to create the heterozygous diploid strain YRH1981. The MATα version of YRH1114 was mated to the MATa YRH1114 strain to generate the homozygous diploid strain YRH1982. Both diploid strains were grown on xylose and compared to the original unadapted and adapted haploid strains ( Figure 2). The heterozygous diploid YRH1981 grew as poorly as the unadapted YRH631 strain and the homozygous diploid YRH1982 grew nearly as well as the adapted haploid strain, indicating that the genetic changes were recessive. ) and heterozygous diploid (YRH1981 ) strains were grown on YP medium + 50 g/L xylose and compared to the unadapted haploid (YRH631 ) and adapted haploid (YRH1114 ) strains. Cell dry weight was used to measure cell mass instead of OD600 to account for differences in dry weight per OD600 between haploid and diploid strains. Plots are the average values from at least triplicate cultures. Error bars show standard deviations.

Tetrad Dissection to Identify Causative Mutations
Most of the SNPs were located on different chromosomes, and mutations located on the same chromosome appeared on opposite arms of the chromosome (Table 3). Since none of the SNPs were closely linked, haploid spores from the heterozygous diploid were analyzed to determine which SNPs were associated with improved growth on xylose medium. Heterozygous diploid strain YRH1981 was sporulated and tetrads with re-assorted wild-type and SNP combinations were dissected onto SD-Ura-Trp plates to maintain selection for the XI and XK expression plasmids ( Figure 3A). Growth of the haploids was compared in xylose cultures ( Figure 3B). The presence of the expected SNP in each haploid was confirmed using a PCR-based SNP detection assay; the results of the tetrad analysis are listed in Table 4.
Each haploid strain from the tetrad analysis that grew well in xylose medium carried both pho13 G208C and pbs2 L363X mutations; these mutations were identified as likely candidates for increased growth on xylose seen in these strains. None of the haploid strains carrying just one of these mutations was able to grow as well on xylose medium, further suggesting that both pho13 G208C and pbs2 L363X mutations are required. The pbs2 L363X mutation results in leucine-to-stop change in the middle of protein, likely rendering it non-functional. Additionally, since loss-of-function mutation, or deletion, of PHO13 has previously ) strains were grown on YP medium + 50 g/L xylose and compared to the unadapted haploid (YRH631 switched from MATa to MATα and the adapted MATα strain was mated to an unadapted parent strainYRH631 (MATa) to create the heterozygous diploid strain YRH1981. The MATα version of YRH1114 was mated to the MATa YRH1114 strain to generate the homozygous diploid strain YRH1982. Both diploid strains were grown on xylose and compared to the original unadapted and adapted haploid strains ( Figure 2). The heterozygous diploid YRH1981 grew as poorly as the unadapted YRH631 strain and the homozygous diploid YRH1982 grew nearly as well as the adapted haploid strain, indicating that the genetic changes were recessive. ) and heterozygous diploid (YRH1981 ) strains were grown on YP medium + 50 g/L xylose and compared to the unadapted haploid (YRH631 ) and adapted haploid (YRH1114 ) strains. Cell dry weight was used to measure cell mass instead of OD600 to account for differences in dry weight per OD600 between haploid and diploid strains. Plots are the average values from at least triplicate cultures. Error bars show standard deviations.

Tetrad Dissection to Identify Causative Mutations
Most of the SNPs were located on different chromosomes, and mutations located on the same chromosome appeared on opposite arms of the chromosome (Table 3). Since none of the SNPs were closely linked, haploid spores from the heterozygous diploid were analyzed to determine which SNPs were associated with improved growth on xylose medium. Heterozygous diploid strain YRH1981 was sporulated and tetrads with re-assorted wild-type and SNP combinations were dissected onto SD-Ura-Trp plates to maintain selection for the XI and XK expression plasmids ( Figure 3A). Growth of the haploids was compared in xylose cultures ( Figure 3B). The presence of the expected SNP in each haploid was confirmed using a PCR-based SNP detection assay; the results of the tetrad analysis are listed in Table 4.
Each haploid strain from the tetrad analysis that grew well in xylose medium carried both pho13 G208C and pbs2 L363X mutations; these mutations were identified as likely candidates for increased growth on xylose seen in these strains. None of the haploid strains carrying just one of these mutations was able to grow as well on xylose medium, further suggesting that both pho13 G208C and pbs2 L363X mutations are required. The pbs2 L363X mutation results in leucine-to-stop change in the middle of protein, likely rendering it non-functional. Additionally, since loss-of-function mutation, or deletion, of PHO13 has previously ) and adapted haploid (YRH1114 switched from MATa to MATα and the adapted MATα strain was mated to an unadapted parent strainYRH631 (MATa) to create the heterozygous diploid strain YRH1981. The MATα version of YRH1114 was mated to the MATa YRH1114 strain to generate the homozygous diploid strain YRH1982. Both diploid strains were grown on xylose and compared to the original unadapted and adapted haploid strains ( Figure 2). The heterozygous diploid YRH1981 grew as poorly as the unadapted YRH631 strain and the homozygous diploid YRH1982 grew nearly as well as the adapted haploid strain, indicating that the genetic changes were recessive. ) and heterozygous diploid (YRH1981 ) strains were grown on YP medium + 50 g/L xylose and compared to the unadapted haploid (YRH631 ) and adapted haploid (YRH1114 ) strains. Cell dry weight was used to measure cell mass instead of OD600 to account for differences in dry weight per OD600 between haploid and diploid strains. Plots are the average values from at least triplicate cultures. Error bars show standard deviations.

Tetrad Dissection to Identify Causative Mutations
Most of the SNPs were located on different chromosomes, and mutations located on the same chromosome appeared on opposite arms of the chromosome (Table 3). Since none of the SNPs were closely linked, haploid spores from the heterozygous diploid were analyzed to determine which SNPs were associated with improved growth on xylose medium. Heterozygous diploid strain YRH1981 was sporulated and tetrads with re-assorted wild-type and SNP combinations were dissected onto SD-Ura-Trp plates to maintain selection for the XI and XK expression plasmids ( Figure 3A). Growth of the haploids was compared in xylose cultures ( Figure 3B). The presence of the expected SNP in each haploid was confirmed using a PCR-based SNP detection assay; the results of the tetrad analysis are listed in Table 4.
Each haploid strain from the tetrad analysis that grew well in xylose medium carried both pho13 G208C and pbs2 L363X mutations; these mutations were identified as likely candidates for increased growth on xylose seen in these strains. None of the haploid strains carrying just one of these mutations was able to grow as well on xylose medium, further suggesting that both pho13 G208C and pbs2 L363X mutations are required. The pbs2 L363X mutation results in leucine-to-stop change in the middle of protein, likely rendering it non-functional. Additionally, since loss-of-function mutation, or deletion, of PHO13 has previously ) strains. Cell dry weight was used to measure cell mass instead of OD 600 to account for differences in dry weight per OD 600 between haploid and diploid strains. Plots are the average values from at least triplicate cultures. Error bars show standard deviations.

Tetrad Dissection to Identify Causative Mutations
Most of the SNPs were located on different chromosomes, and mutations located on the same chromosome appeared on opposite arms of the chromosome (Table 3). Since none of the SNPs were closely linked, haploid spores from the heterozygous diploid were analyzed to determine which SNPs were associated with improved growth on xylose medium. Heterozygous diploid strain YRH1981 was sporulated and tetrads with re-assorted wildtype and SNP combinations were dissected onto SD-Ura-Trp plates to maintain selection for the XI and XK expression plasmids ( Figure 3A). Growth of the haploids was compared in xylose cultures ( Figure 3B). The presence of the expected SNP in each haploid was confirmed using a PCR-based SNP detection assay; the results of the tetrad analysis are listed in Table 4.
Each haploid strain from the tetrad analysis that grew well in xylose medium carried both pho13 G208C and pbs2 L363X mutations; these mutations were identified as likely candidates for increased growth on xylose seen in these strains. None of the haploid strains carrying just one of these mutations was able to grow as well on xylose medium, further suggesting that both pho13 G208C and pbs2 L363X mutations are required. The pbs2 L363X mutation results in leucine-to-stop change in the middle of protein, likely rendering it non-functional. Additionally, since loss-of-function mutation, or deletion, of PHO13 has previously been shown to increase xylose utilization [15,[31][32][33], the pho13 G208C mutation in our adapted strain is also likely non-functional. The ability of multiple haploid strains with these two mutations to grow well in xylose medium, regardless of other SNPs, also indicated that the SNPs and INDELs located in intergenic regions do not play a significant role in xylose utilization. been shown to increase xylose utilization [15,[31][32][33], the pho13 G208C mutation in our adapted strain is also likely non-functional. The ability of multiple haploid strains with these two mutations to grow well in xylose medium, regardless of other SNPs, also indicated that the SNPs and INDELs located in intergenic regions do not play a significant role in xylose utilization.  Table 4. Table 4. Tetrad analysis.

Strain
Tetrad Figure 3B for growth ratings. ** (+) indicates presence of the WT allele while (−) indicates presence of the mutant allele.  Table 4. Table 4. Tetrad analysis. Figure 3B for growth ratings. ** (+) indicates presence of the WT allele while (−) indicates presence of the mutant allele.

Validation of PBS2 and PHO13 Genes
Since results from the heterozygous diploid strain indicated that the mutations were recessive, we next expressed wild-type PBS2 and PHO13 genes in the adapted strain and compared growth to that of the original adapted strain YRH1114, as well as the adapted strain with empty vector controls (i.e., YRH1966, YRH2042, and YRH2044). Expression of both PBS2 and PHO13, either separately or combined, reduced growth on xylose medium, but not to the extent of the unadapted strain YRH631 ( Figure 4A). One possible explanation is that episomal expression of the genes does not adequately replicate the expression level from their native genomic locations. Additionally, it was possible that one of the SNPs was exerting a semi-dominant effect in the presence of the wild-type gene.
HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele.  HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele.  HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele. HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele. HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele. YRH1968 (YRH1114 + PBS2 plasmid), mutations of pbs2p lacking the kinase domain still show Hog1p binding activity [63]. HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele. mutations of pbs2p lacking the kinase domain still show Hog1p binding activity [63]. HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele.    [63]. HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele. mutations of pbs2p lacking the kinase domain still show Hog1p binding activity [63]. HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele. The kinase domain of Pbs2p encompasses amino acids 360 to 688, and truncation mutations of pbs2p lacking the kinase domain still show Hog1p binding activity [63]. HOG1 deletion has been shown to increase xylose utilization [17]; the pbs2 L363X allele likely exerts its function through Hog1p (see Results and Discussion, Section 3.8). Thus, it is possible that the pbs2 L363X allele, which creates a stop codon at amino acid 363, could interfere with wild-type PBS2. When both PBS2 and pbs2 L363X are expressed in the same cell, the shortened, non-phosphorylating mutant pbs2p could compete with wild-type Pbs2p for Hog1p binding, resulting in decreased Hog1p phosphorylation. To test this possibility, the wild-type PBS2 gene was re-integrated into the adapted strain YRH1114 at its normal location to make strain YRH1934, thus removing any expression of the mutant pbs2p. Strain YRH1934, with only the pbs2 L363X mutation replaced with PBS2 in the genome and all other SNPs present, had a growth pattern identical to the unadapted strain ( Figure 4B). These data suggest that for strains expressing the XI pathway, mutation in both PBS2 and PHO13 genes is required to improve growth on xylose. The data also suggest that the pbs2 L363X allele is semi-dominant when present with the wild-type PBS2 allele.

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XIexpressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XIexpressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the YRH631 (unadapted strain),

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XIexpressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the YRH1114 (adapted strain); Panel (A),

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XIexpressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the YRH1563 (YRH631 + pho13∆),

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XIexpressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the YRH2021 (YRH631 + pbs2∆),

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XIexpressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the YRH2022 (YRH631 + pbs2∆, pho13∆); Panel (B),

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XIexpressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the YRH2056 (CEN.PK2-1C + pbs2∆),

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XIexpressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the YRH2057 (CEN.PK2-1C + pho13∆),

Analysis of Strains with PBS2 and PHO13 Deletions
While results from the heterozygous diploid strain suggested that at least one of the two mutations responsible for improved growth on xylose was recessive, it was not clear whether complete loss of function could also improve growth on xylose, or if partial function of the mutant allele was required. Starting with the unadapted YRH631 parental strain, we next deleted PBS2 and PHO13, separately or together. Using YRH631, single deletions of PBS2 or PHO13 were unable to improve growth on xylose. In contrast, deletion of both PBS2 and PHO13 did improve xylose growth ( Figure 5A). Upon further analysis of SNPs in YRH1114 versus YRH631, we discovered that the ste24 L418F SNP was present in the parental YRH631 background, and thus propagated throughout the haploids obtained from tetrad dissection. To rule out any involvement of ste24 L418F for improving growth on xylose, we recreated the PBS2 and PHO13 deletions in the parent CEN.PK2-1C background with wild-type STE24 and re-transformed the strains with XI and XK expression vectors. No difference in growth was observed when using the CEN.PK2-1C background ( Figure 5B) compared to strains generated from the YRH631 parent. These results indicated that the ste24 L418F mutation did not facilitate growth on xylose and that the growth advantage on xylose medium stems from loss of function of both PBS2 and PHO13.

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XIexpressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the YRH2058 (CEN.PK2-1C + pbs2∆, pho13∆).

Analysis of PBS2 and PHO13 Requirement for Strains Expressing the XR/XDH Pathway
We next investigated whether the causative mutations discovered in the adapted XI-expressing strain were also functional in a strain expressing the XR/XDH pathway. Starting with the CEN.PK2-1C strain, we deleted PBS2 and PHO13 in strains engineered to express the Scheffersomyces stipitis xylose reductase and xylitol dehydrogenase genes (XYL1 and XYL2), as well as overexpressed S. cerevisiae XKS1. Contrary to cells expressing the XI pathway, cells expressing the XR/XDH pathway showed a significant improvement in growth from only PHO13 deletion (YRH2054) (Figure 6). While single deletion of PBS2 in these cells (YRH2053) did not improve growth on xylose, strain YRH2055 with deletion of both PBS2 and PHO13 showed additional improvement in growth compared to the single PHO13 deletion strain. The implications of this result in comparison to strains expressing the XI/XK pathway are discussed in Section 3.9.
single PHO13 deletion strain. The implications of this result in compa pressing the XI/XK pathway are discussed in Section 3.9.

Xylose Fermentation in Strains with Targeted PBS2 and PHO13 Deletio
To determine if PHO13 and PBS2 mutations were resposibile for t fermentation obserseved with the adapted strain YRH1114 in our pr we determined the effect of both single and double deletion of PHO13 expressing either the XI/XK or XR/XDH pathway. Batch fermentat expressing the XI/XK pathway (Figure 7 panels A, C, E, G) were con from aerobic culture using xylose medium. Single deletion of PHO1 increase growth or ethanol production when grown under microaerob Deletion of both PHO13 and PBS2 resulted in a fermentation prof to that of the adapted YRH1114 strain (Table 5 and Figure 7). R fermentations using strains expressing the XR/XDH pathway were a results from our aerobic growth analyses. As seen under aerobic deletion of PBS2 did not improve fermentation, while single deletion in a significant increase in xylose consumption and ethanol produ Figure 7 panels B, D, F, H). Strain YRH2055 with deletion of both PBS2 showed additional improvement compared to the single PHO13 delet previously, using the XR/XDH pathway results in low ethanol yield d of xylitol production when compared to strains expressing the XI/XK and [16]). The highest ethanol yields obtained in this study were from (YRH1114) and the strain expressing the XI/XK pathway with targete

Xylose Fermentation in Strains with Targeted PBS2 and PHO13 Deletions
To determine if PHO13 and PBS2 mutations were resposibile for the increased xylos fermentation obserseved with the adapted strain YRH1114 in our previouse study [16] we determined the effect of both single and double deletion of PHO13 and PBS2 in strain expressing either the XI/XK or XR/XDH pathway. Batch fermentations using strain expressing the XI/XK pathway (Figure 7 panels A, C, E, G) were consistent with result from aerobic culture using xylose medium. Single deletion of PHO13 or PBS2 did no increase growth or ethanol production when grown under microaerobic conditions. Deletion of both PHO13 and PBS2 resulted in a fermentation profile that was simila to that of the adapted YRH1114 strain (Table 5 and Figure 7). Results from batch fermentations using strains expressing the XR/XDH pathway were also consistent with results from our aerobic growth analyses. As seen under aerobic conditions, singl deletion of PBS2 did not improve fermentation, while single deletion of PHO13 resulted in a significant increase in xylose consumption and ethanol production (Table 5 and Figure 7 panels B, D, F, H). Strain YRH2055 with deletion of both PBS2 and PHO13, again showed additional improvement compared to the single PHO13 deletion strain. As seen previously, using the XR/XDH pathway results in low ethanol yield due to higher level of xylitol production when compared to strains expressing the XI/XK pathway (Table  and [ 16]). The highest ethanol yields obtained in this study were from the adapted strain (YRH1114) and the strain expressing the XI/XK pathway with targeted deletion of both PBS2 and PHO13 (YRH2058). While there was a slight increase in ethanol yield for strain YRH2058 compared to YRH1114, the increase was not statistically significant. Since th ferementation profile of the strain with targeted PHO13 and PBS2 deletions was simila to the adapted strain, we believe that mutations to PHO13 and PBS2 in YRH1114 are th causitive mutations resulting in improved xylose fermentation.

Xylose Fermentation in Strains with Targeted PBS2 and PHO13 Deletions
To determine if PHO13 and PBS2 mutations were resposibile for the increased xylose fermentation obserseved with the adapted strain YRH1114 in our previouse study [16], we determined the effect of both single and double deletion of PHO13 and PBS2 in strains expressing either the XI/XK or XR/XDH pathway. Batch fermentations using strains expressing the XI/XK pathway (Figure 7 panels A, C, E, G) were consistent with results from aerobic culture using xylose medium. Single deletion of PHO13 or PBS2 did not increase growth or ethanol production when grown under microaerobic conditions. Deletion of both PHO13 and PBS2 resulted in a fermentation profile that was similar to that of the adapted YRH1114 strain (Table 5 and Figure 7). Results from batch fermentations using strains expressing the XR/XDH pathway were also consistent with results from our aerobic growth analyses. As seen under aerobic conditions, single deletion of PBS2 did not improve fermentation, while single deletion of PHO13 resulted in a significant increase in xylose consumption and ethanol production ( Table 5 and Figure 7 panels B, D, F, H). Strain YRH2055 with deletion of both PBS2 and PHO13, again, showed additional improvement compared to the single PHO13 deletion strain. As seen previously, using the XR/XDH pathway results in low ethanol yield due to higher levels of xylitol production when compared to strains expressing the XI/XK pathway (Table 5 and [16]). The highest ethanol yields obtained in this study were from the adapted strain (YRH1114) and the strain expressing the XI/XK pathway with targeted deletion of both PBS2 and PHO13 (YRH2058). While there was a slight increase in ethanol yield for strain YRH2058 compared to YRH1114, the increase was not statistically significant. Since the ferementation profile of the strain with targeted PHO13 and PBS2 deletions was similar to the adapted strain, we believe that mutations to PHO13 and PBS2 in YRH1114 are the causitive mutations resulting in improved xylose fermentation.

Xylose Fermentation in Strains with Targeted PBS2 and PHO13 Deletions
To determine if PHO13 and PBS2 mutations were resposibile for the increased xylose fermentation obserseved with the adapted strain YRH1114 in our previouse study [16], we determined the effect of both single and double deletion of PHO13 and PBS2 in strains expressing either the XI/XK or XR/XDH pathway. Batch fermentations using strains expressing the XI/XK pathway (Figure 7 panels A, C, E, G) were consistent with results from aerobic culture using xylose medium. Single deletion of PHO13 or PBS2 did not increase growth or ethanol production when grown under microaerobic conditions. Deletion of both PHO13 and PBS2 resulted in a fermentation profile that was similar to that of the adapted YRH1114 strain (Table 5 and Figure 7). Results from batch fermentations using strains expressing the XR/XDH pathway were also consistent with results from our aerobic growth analyses. As seen under aerobic conditions, single deletion of PBS2 did not improve fermentation, while single deletion of PHO13 resulted in a significant increase in xylose consumption and ethanol production ( Table 5 and Figure 7 panels B, D, F, H). Strain YRH2055 with deletion of both PBS2 and PHO13, again, showed additional improvement compared to the single PHO13 deletion strain. As seen previously, using the XR/XDH pathway results in low ethanol yield due to higher levels of xylitol production when compared to strains expressing the XI/XK pathway (Table 5 and [16]). The highest ethanol yields obtained in this study were from the adapted strain (YRH1114) and the strain expressing the XI/XK pathway with targeted deletion of both PBS2 and PHO13 (YRH2058). While there was a slight increase in ethanol yield for strain YRH2058 compared to YRH1114, the increase was not statistically significant. Since the ferementation profile of the strain with targeted PHO13 and PBS2 deletions was similar to the adapted strain, we believe that mutations to PHO13 and PBS2 in YRH1114 are the causitive mutations resulting in improved xylose fermentation.

Xylose Fermentation in Strains with Targeted PBS2 and PHO13 Deletions
To determine if PHO13 and PBS2 mutations were resposibile for the increased xylose fermentation obserseved with the adapted strain YRH1114 in our previouse study [16], we determined the effect of both single and double deletion of PHO13 and PBS2 in strains expressing either the XI/XK or XR/XDH pathway. Batch fermentations using strains expressing the XI/XK pathway (Figure 7 panels A, C, E, G) were consistent with results from aerobic culture using xylose medium. Single deletion of PHO13 or PBS2 did not increase growth or ethanol production when grown under microaerobic conditions. Deletion of both PHO13 and PBS2 resulted in a fermentation profile that was similar to that of the adapted YRH1114 strain (Table 5 and Figure 7). Results from batch fermentations using strains expressing the XR/XDH pathway were also consistent with results from our aerobic growth analyses. As seen under aerobic conditions, single deletion of PBS2 did not improve fermentation, while single deletion of PHO13 resulted in a significant increase in xylose consumption and ethanol production ( Table 5 and Figure 7 panels B, D, F, H). Strain YRH2055 with deletion of both PBS2 and PHO13, again, showed additional improvement compared to the single PHO13 deletion strain. As seen previously, using the XR/XDH pathway results in low ethanol yield due to higher levels of xylitol production when compared to strains expressing the XI/XK pathway (Table 5 and [16]). The highest ethanol yields obtained in this study were from the adapted strain (YRH1114) and the strain expressing the XI/XK pathway with targeted deletion of both PBS2 and PHO13 (YRH2058). While there was a slight increase in ethanol yield for strain YRH2058 compared to YRH1114, the increase was not statistically significant. Since the ferementation profile of the strain with targeted PHO13 and PBS2 deletions was similar to the adapted strain, we believe that mutations to PHO13 and PBS2 in YRH1114 are the causitive mutations resulting in improved xylose fermentation.

Xylose Fermentation in Strains with Targeted PBS2 and PHO13 Deletions
To determine if PHO13 and PBS2 mutations were resposibile for the increased xylose fermentation obserseved with the adapted strain YRH1114 in our previouse study [16], we determined the effect of both single and double deletion of PHO13 and PBS2 in strains expressing either the XI/XK or XR/XDH pathway. Batch fermentations using strains expressing the XI/XK pathway (Figure 7 panels A, C, E, G) were consistent with results from aerobic culture using xylose medium. Single deletion of PHO13 or PBS2 did not increase growth or ethanol production when grown under microaerobic conditions. Deletion of both PHO13 and PBS2 resulted in a fermentation profile that was similar to that of the adapted YRH1114 strain (Table 5 and Figure 7). Results from batch fermentations using strains expressing the XR/XDH pathway were also consistent with results from our aerobic growth analyses. As seen under aerobic conditions, single deletion of PBS2 did not improve fermentation, while single deletion of PHO13 resulted in a significant increase in xylose consumption and ethanol production ( Table 5 and Figure 7 panels B, D, F, H). Strain YRH2055 with deletion of both PBS2 and PHO13, again, showed additional improvement compared to the single PHO13 deletion strain. As seen previously, using the XR/XDH pathway results in low ethanol yield due to higher levels of xylitol production when compared to strains expressing the XI/XK pathway (Table 5 and [16]). The highest ethanol yields obtained in this study were from the adapted strain (YRH1114) and the strain expressing the XI/XK pathway with targeted deletion of both PBS2 and PHO13 (YRH2058). While there was a slight increase in ethanol yield for strain YRH2058 compared to YRH1114, the increase was not statistically significant. Since the ferementation profile of the strain with targeted PHO13 and PBS2 deletions was similar to the adapted strain, we believe that mutations to PHO13 and PBS2 in YRH1114 are the causitive mutations resulting in improved xylose fermentation. YRH2055 (YRH2040 + pbs2∆, pho13∆).

Xylose Fermentation in Strains with Targeted PBS2 and PHO13 Deletions
To determine if PHO13 and PBS2 mutations were resposibile for the increased xylose fermentation obserseved with the adapted strain YRH1114 in our previouse study [16], we determined the effect of both single and double deletion of PHO13 and PBS2 in strains expressing either the XI/XK or XR/XDH pathway. Batch fermentations using strains expressing the XI/XK pathway (Figure 7 panels A, C, E, G) were consistent with results from aerobic culture using xylose medium. Single deletion of PHO13 or PBS2 did not increase growth or ethanol production when grown under microaerobic conditions. Deletion of both PHO13 and PBS2 resulted in a fermentation profile that was similar to that of the adapted YRH1114 strain (Table 5 and Figure 7). Results from batch fermentations using strains expressing the XR/XDH pathway were also consistent with results from our aerobic growth analyses. As seen under aerobic conditions, single deletion of PBS2 did not improve fermentation, while single deletion of PHO13 resulted in a significant increase in xylose consumption and ethanol production ( Table 5 and Figure 7 panels B, D, F, H). Strain YRH2055 with deletion of both PBS2 and PHO13, again, showed additional improvement compared to the single PHO13 deletion strain. As seen previously, using the XR/XDH pathway results in low ethanol yield due to higher levels of xylitol production when compared to strains expressing the XI/XK pathway (Table 5 and [16]). The highest ethanol yields obtained in this study were from the adapted strain (YRH1114) and the strain expressing the XI/XK pathway with targeted deletion of both PBS2 and PHO13 (YRH2058). While there was a slight increase in ethanol yield for strain YRH2058 compared to YRH1114, the increase was not statistically significant. Since the ferementation profile of the strain with targeted PHO13 and PBS2 deletions was similar to the adapted strain, we believe that mutations to PHO13 and PBS2 in YRH1114 are the causitive mutations resulting in improved xylose fermentation.

Mutation to HOG1 Pathway Kinases Improves Xylose Utilization
While this study is the first to show that mutations to PBS2 can increase xylose utilization, previous studies with strains evolved for improved xylose utilization identified two other members of the mitogen-activated protein (MAP) kinase cascade of which PBS2 is part. PBS2 is a MAP kinase kinase (MAPKK) that is part of the high osmolarity glycerol YRH631 (unadapted XI strain),

Mutation to HOG1 Pathway Kinases Improves Xylose Utilization
While this study is the first to show that mutations to PBS2 can increase xylose utilization, previous studies with strains evolved for improved xylose utilization identified two other members of the mitogen-activated protein (MAP) kinase cascade of which PBS2 is part. PBS2 is a MAP kinase kinase (MAPKK) that is part of the high osmolarity glycerol

Mutation to HOG1 Pathway Kinases Improves Xylose Utilization
While this study is the first to show that mutations to PBS2 can increase xylose utilization, previous studies with strains evolved for improved xylose utilization identified two other members of the mitogen-activated protein (MAP) kinase cascade of which PBS2 is part. PBS2 is a MAP kinase kinase (MAPKK) that is part of the high osmolarity glycerol

Mutation to HOG1 Pathway Kinases Improves Xylose Utilization
While this study is the first to show that mutations to PBS2 can increase xylose utilization, previous studies with strains evolved for improved xylose utilization identified two other members of the mitogen-activated protein (MAP) kinase cascade of which PBS2 is part. PBS2 is a MAP kinase kinase (MAPKK) that is part of the high osmolarity glycerol

Mutation to HOG1 Pathway Kinases Improves Xylose Utilization
While this study is the first to show that mutations to PBS2 can increase xylose utilization, previous studies with strains evolved for improved xylose utilization identified two other members of the mitogen-activated protein (MAP) kinase cascade of which PBS2 is part. PBS2 is a MAP kinase kinase (MAPKK) that is part of the high osmolarity glycerol YRH2054 (YRH2040 + pho13∆),

Mutation to HOG1 Pathway Kinases Improves Xylose Utilization
While this study is the first to show that mutations to PBS2 can increase xylose utilization, previous studies with strains evolved for improved xylose utilization identified two other members of the mitogen-activated protein (MAP) kinase cascade of which PBS2 is part. PBS2 is a MAP kinase kinase (MAPKK) that is part of the high osmolarity glycerol YRH2055 (YRH2040 + pbs2∆, pho13∆). expression [17,34]. Gre3p is an endogenous xylose reductase, and deletion of GRE3 has been shown to decrease production of xylitol 2-to 3-fold in XI expressing strains [68]. Xylitol is also a competitive inhibitor of xylose isomerase, so decreased xylitol concentration leads to increased XI activity [12,69]. Consistent with decreased Gre3p and decreased xylose reductase activity, we previously showed that our adapted strain has 2.9-fold increased XI activity and a 1.7-fold decrease in xylitol yield compared to the unadapted strain [16]. Batch fermentations performed in this study also showed a 1.8-fold and 2.2-fold decrease in xylitol yield for the adapted YRH1114 strain and YRH2058 (CEN.PK2-1C with XI/XK pathway + pbs2∆, pho13∆) when compared to the unadapted strain, respectively.

Model Showing Involvement of PBS2 and PHO13 in Separate Steps of Xylose Utilization
We propose the following model to explain the requirement of both pbs2 and pho13 mutation to improve xylose growth for the engineered strains used in this study (Figure 8). Loss of PHO13 function has been shown to increase expression of PPP enzymes [31], which is required for general xylose metabolism. For strains expressing the XI pathway, cells with just pho13∆ are still limited in xylose conversion to X5P ( Figure 8A). Cells with pbs2∆, which likely acts like a hog1∆ strain because PBS2 is required for HOG1 function [65], have increased xylose conversion to X5P, but are then limited by poor flux through the PPP due to the presence of PHO13. When utilizing the XI pathway, both gene deletions are required to get xylose into central metabolism and allow cell growth. Conversely, for cells expressing the XR/XDH pathway for xylose metabolism, X5P production is not as limiting a factor. Thus, increased flux through the PPP from deletion of only PHO13 would be expected to increase growth on xylose ( Figure 8B). Previous studies using multiple different genetic backgrounds are consistent with this model. These studies show that deleting only PHO13 in strains using the XR/XDH pathway increases xylose utilization [31][32][33]40,70]. Several of these studies utilized different genetic backgrounds. Different XI and XK genes were also used compared to ours, suggesting that the results with respect to PHO13 deletion are not unique to specific strain backgrounds or xylose isomerases.

Model Showing Involvement of PBS2 and PHO13 in Separate Steps of Xylose Utilization
We propose the following model to explain the requirement of both pbs2 and pho13 mutation to improve xylose growth for the engineered strains used in this study ( Figure  8). Loss of PHO13 function has been shown to increase expression of PPP enzymes [31], which is required for general xylose metabolism. For strains expressing the XI pathway, cells with just pho13∆ are still limited in xylose conversion to X5P ( Figure 8A). Cells with pbs2∆, which likely acts like a hog1∆ strain because PBS2 is required for HOG1 function [65], have increased xylose conversion to X5P, but are then limited by poor flux through the PPP due to the presence of PHO13. When utilizing the XI pathway, both gene deletions are required to get xylose into central metabolism and allow cell growth. Conversely, for cells expressing the XR/XDH pathway for xylose metabolism, X5P production is not as limiting a factor. Thus, increased flux through the PPP from deletion of only PHO13 would be expected to increase growth on xylose ( Figure 8B). Previous studies using multiple different genetic backgrounds are consistent with this model. These studies show that deleting only PHO13 in strains using the XR/XDH pathway increases xylose utilization [31][32][33]40,70]. Several of these studies utilized different genetic backgrounds. Different XI and XK genes were also used compared to ours, suggesting that the results with respect to PHO13 deletion are not unique to specific strain backgrounds or xylose isomerases. Strains lacking PBS2 or HOG1, although beneficial for xylose utilization, may not be ideal for growth in lignocellulosic hydrolysates. For example, CRISPRi-identified downregulation of both HOG1 and PBS2 was linked to poor growth in spruce hydrolysate [71]. HOG1 deletion was also shown to affect glucose utilization in corn stover hydrolysate [72]. The adapted YRH1114 strain was previously tested on a glucose and xylose mixture, and glucose utilization was not decreased [16]. However, hydrolysate was not used in the previous study. There is some evidence that PBS2 and HOG1 deletions may not be com- Strains lacking PBS2 or HOG1, although beneficial for xylose utilization, may not be ideal for growth in lignocellulosic hydrolysates. For example, CRISPRi-identified downregulation of both HOG1 and PBS2 was linked to poor growth in spruce hydrolysate [71]. HOG1 deletion was also shown to affect glucose utilization in corn stover hydrolysate [72]. The adapted YRH1114 strain was previously tested on a glucose and xylose mixture, and glucose utilization was not decreased [16]. However, hydrolysate was not used in the previous study. There is some evidence that PBS2 and HOG1 deletions may not be completely identical. Autophosphorylation of Hog1p has been shown to occur in a pbs2∆ strain [73]. In this regard, deletion of PBS2, or co-expression of the pbs2 L363X allele with PBS2, may yield better results than HOG1 deletion, since some level of Hog1p phosphorylation would still be present in these strains. Additional work is needed to determine if there are differences between HOG1 and PBS2 deletions, as well as comparison of PBS2 deletion and the semi-dominant pbs2 L363X allele for negative effects when grown on lignocellulosic hydrolysates.

Conclusions
In the present work, we analyzed the genome sequence of an engineered S. cerevisiae strain expressing codon-optimized Prevotella ruminicola XI and XK genes that was previously adapted for increased xylose fermentation [16]. Tetrad analysis of haploid spores from the heterozygous diploid showed that of the seven ORF mutations present in the adapted strain, only pbs2 L363X and pho13 G208C mutations improved xylose utilization. Strains carrying a single mutated or deleted PBS2 or PHO13 did not show improved growth on xylose, indicating that both mutations were necessary. Double deletion of PBS2 and PHO13 also improved xylose utilization for strains expressing the XR/XDH pathway. However, unlike strains expressing the XI pathway, these strains also showed a significant improvement from deletion of the PHO13 gene alone.

Supplementary Materials:
The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/fermentation8120669/s1: Table S1: DNA oligonucleotides used in this study; Construction of plasmids for expressing PBS2 and PHO13; Construction of PBS2 and PHO13 deletion strains.