Improving Xylose Fermentation in Saccharomyces cerevisiae by Expressing Nuclear-Localized Hexokinase 2

Understanding the relationship between xylose and the metabolic regulatory systems is a prerequisite to enhance xylose utilization in recombinant S. cerevisiae strains. Hexokinase 2 (Hxk2p) is an intracellular glucose sensor that localizes to the cytoplasm or the nucleus depending on the carbon source. Hxk2p interacts with Mig1p to regulate gene transcription in the nucleus. Here, we investigated the effect of nucleus-localized Hxk2p and Mig1p on xylose fermentation. The results show that the expression of HXK2S14A, which encodes a constitutively nucleus-localized Hxk2p, increased the xylose consumption rate, the ethanol production rate, and the ethanol yield of the engineered yeast strain by 23.5%, 78.6% and 42.6%, respectively. The deletion of MIG1 decreased xylose utilization and eliminated the positive effect of Hxk2p. We then performed RNA-seq and found that the targets of Hxk2pS14A on xylose were mainly genes that encode RNA-binding proteins. This is very different from the known targets of Mig1p and supports the notion that the Hxk2p-Mig1p interaction is abolished in the presence of xylose. These results will improve our understanding of the interrelation between the Snf1p-Mig1p-Hxk2p glucose signaling pathway and xylose utilization in S. cerevisiae and suggests that the expression of HXK2S14A could be a viable strategy to improve xylose utilization.


Introduction
Lignocellulosic material is considered to be a sustainable source for the production of biofuels and other chemicals. The utilization not only of glucose but also of xylose, which is the second most abundant sugar component in the hydrolysates of lignocellulosic materials, will undoubtedly bring economic benefits [1,2]. Saccharomyces cerevisiae is a robust, well studied model organism that is generally recognized as safe and has a strong capacity to metabolize glucose. Furthermore, S. cerevisiae is considered a very competitive microbial cell factory for the production of biofuels and chemicals, especially ethanol [1][2][3]. However, S. cerevisiae cannot utilize xylose because it does not have the initial metabolic pathway that converts xylose to xylulose, and also because of inefficient xylulose metabolism [1,2,4].
Many previous studies in metabolic and evolutionary engineering have attempted to solve this problem. Mainly, genes from another species of yeast, Scheffersomyces(Pichia) stipitis, that encode xylose reductase (XR) and xylitol dehydrogenase (XDH) [1,[5][6][7][8] or xylose isomerases (XI) of bacterial or Under high-glucose conditions, Snf1p is dephosphorylated by the Glc7p-Reg1p protein phosphatase, and Hxk2p is essential for this process. The dephosphorylated form of Snf1p is inactive. Hxk2p and Mig1p are in their dephosphorylated forms and function in glucose repression. In a low glucose environment, Snf1p is phosphorylated by Sak1p/Elm1p/Tos3p. The phosphorylated Snf1p is active and phosphorylates Hxk2p at serine 14, which affects the shuttling of Hxk2p. The active form of Snf1p also phosphorylates Mig1p, and this directly leads Mig1p to move from the nucleus to the cytoplasm. The Mig1p in the cytoplasm is ubiquitinated and degraded by proteases [26][27][28][29].

Yeast Strains and Plasmids
The S. cerevisiae strains and plasmids used in the present study are listed in Table 1. The primer sequences are given in Table A1. Strain BSPC039 [5], which overexpresses the XKS1 and PPP genes, was cultured in yeast extract-peptone-dextrose (YPD) medium consisting of 20 g L −1 peptone, 10 g L −1 yeast extract, and 20 g L −1 glucose. Strain BSL01 [25] harboring the plasmid pJX7, which contains an expression cassette for xylose isomerase, was derived from BSPC039 and was used as a control in the present study. To introduce the mutation S14A in HXK2, we amplified and cloned FRAGMENTs 1 and 4 from BSPC039 genomic DNA using the primer pairs Hxk2p-F/DT-hxk2-R(S14A) and DT-hxk2-F(S14A)/Hxk2p-R, respectively, and we amplified and cloned FRAGMENT 2 and 3 from plasmid pUG72 using primer pairs ura-2-F/ura-2-R and ura-1-R/ura-1-F, respectively. We then fused FRAGMENTs 1 and 2 to obtain FRAGMENT 5, fused FRAGMENTs 3 and 4 to obtain FRAGMENT 6, and then transformed BSPC039 with both FRAGMENTs 5 and 6. The transformants were selected on SC-Ura medium (1.7 g L −1 YNB-AA/AS, 5 g L −1 (NH 4 ) 2 SO 4 , 0.77 g L −1 CSM-Ura (Sunrise Science Products, USA), and 20 g L −1 glucose). The desired transformants (confirmed by sequencing the PCR product of the Hxk2 locus) were then cultured overnight in YPD medium and selected on SC-Ura agar medium supplemented with an additional 50 mg L −1 uracil and 1 g L −1 5-FOA. The clone without the KI-URA expression cassette was then transformed with plasmid pJX7 resulting in the strain BSHA01. More details of the construction process are shown in the diagram in Figure A1.
The MIG1 knockout cassette was amplified from the plasmid YEp-CH [32] using the primers KMig-F and KMig-R. It consists of the homologous recombination arm of MIG1 and the hygromycin B expression cassette containing the resistance gene hygB. We transformed the BSL01 and BSHA01 strains with the MIG1 knockout cassette. The transformants were selected on SC-Ura medium containing 400 mg L −1 hygromycin B to obtain the hygromycin B-resistant strains BSHM01 and BSHM02.

Fermentation
The yeast cells were activated in SC-Ura medium containing 20 g L −1 glucose as carbon source. A single colony was inoculated into a 250-mL shake flask containing 30 mL medium and grown at 30 • C with shaking at 200 rpm. After overnight incubation, the cells were transferred to fresh medium with an initial OD 600 of 0.2 and cultured for another 12-14 h. The activated cells were then collected, washed with sterile water, and inoculated into 30 mL SC-Ura medium containing 20 g L −1 xylose as the sole carbon source; the initial biomass was 0.23 g L −1 dry cell weight (DCW). Fermentation was performed at 30 • C and 200 rpm in 100-mL shake flasks, and there were three biological replicates for each strain [5].

Analytical Methods
Fermentation samples were collected at specific time intervals. The density of the yeast cells (OD 600 ) was determined with a UV-visible spectrophotometer (Eppendorf, Germany). The biomass levels were estimated based on the measured OD 600 -dry weight correlation. One OD 600 unit corresponded to 0.230 g of DCW L −1 for BSPC039 and its derivative strains [5]. The concentrations of xylose and ethanol were determined using a high-performance liquid chromatography (HPLC) system (Shimadzu, Japan) fitted with an Aminex HPX-87H ion exchange column (300 × 7.8 mm) (Bio-Rad, Hercules, CA, USA). H 2 SO 4 (5 mmol L −1 ) was used as the mobile phase with a flow rate of 0.6 mL min −1 , and the temperature of the column oven was 45 • C [5]. The specific growth rate (µ) was the regression coefficient of the log-linear regression of the OD 600 versus time during the exponential growth phase [5]. The xylose consumption rate (r xylose ) and ethanol production rate (r ethanol ) were the amounts of xylose consumed and ethanol produced per hour per liter, respectively. The t-tests were applied to evaluate the differences between means.

Transcriptome Analysis
Samples for RNA sequencing (RNA-seq) were taken from batch fermentations growing on xylose as the sole carbon source at 14 h. The cells in each sample were collected by centrifugation at 5000 rpm at 4 • C for 5 min and then frozen in liquid nitrogen. Total RNA was extracted using a UNIQ-10 Trizol RNA Purification Kit (Sangon Biotech, Shanghai, China) and then fragmented. Contaminating DNA was removed by digestion with DNase I, and cDNA was synthesized by using short mRNA fragments as templates. Three independent RNA extractions were assayed for each strain. The resulting sample libraries were sequenced using an Illumina HiSeq TM 2000 instrument (BGI, Shenzhen, China). Significant differences were indicated by p values ≤ 0.001 and an absolute fold-change threshold of ≥1.5. The information for the Mig1p target genes was obtained from YEASTRACT+ (http://www.yeastract.com/formrankbytf.php) [34].

Nucleus-Localized Hxk2p S14A Enhances Xylose Fermentation
To investigate the role of nucleus-localized Hxk2p in xylose metabolism, the gene HXK2 in strain BSL01 was replaced by a different allele, HXK2 S14A , that encodes the nucleus-localized mutant Hxk2p S14A [21], giving strain BSHA01. The xylose fermentation results ( Figure 2, Table 2) showed that the xylose consumption rate (r xylose ), the ethanol production rate (r ethanol ), and the ethanol yield (Y ethanol ) of BSHA01 were 0.341 ± 0.027 g L −1 h −1 , 0.118 ± 0.002 g L −1 h −1 , and 0.277 ± 0.032 g g −1 xylose, respectively, which are 23.5%, 78.6%, and 42.6% higher than in the control BSL01, respectively. These results suggest that the increase in nucleus-localized Hxk2p has a positive effect on xylose fermentation in recombinant S. cerevisiae strains. Table 2. Xylose fermentation characteristics of the parental and engineered yeast strains.  xylose consumption rate (rxylose), the ethanol production rate (rethanol), and the ethanol yield (Yethanol) of BSHA01 were 0.341 ± 0.027 g L −1 h −1 , 0.118 ± 0.002 g L −1 h −1 , and 0.277 ± 0.032 g g −1 xylose, respectively, which are 23.5%, 78.6%, and 42.6% higher than in the control BSL01, respectively. These results suggest that the increase in nucleus-localized Hxk2p has a positive effect on xylose fermentation in recombinant S. cerevisiae strains.

Deletion of MIG1 Eliminated the Effect of Nucleus-Localized Hxk2p S14A
In high glucose conditions, Mig1p interacts with Hxk2p to form a repressor complex, which represses the expression of genes encoding sugar transporters and the genes needed for the utilization of alternative fermentable carbon sources [21,26]. In order to investigate the effect of Mig1p and the Mig1p-Hxk2p complex on xylose utilization, we deleted the MIG1 gene in strains BSL01 (control) and BSHA01 (BSL01 derivative carrying the mutant allele HXK2 S14A ) to give the strains BSHM01 and BSHM02, respectively. The xylose fermentation results ( Figure 2, Table 2) showed that the xylose consumption rate (r xylose ) of BSHM01 was 0.205 ± 0.010 g L −1 h −1 , which is 25.7% lower than that of BSL01 (0.276 ± 0.034 g L −1 h −1 ). Correspondingly, the ethanol production rate (r ethanol ) of BSHM01 was 0.036 ± 0.004 g L −1 h −1 , which is 45.5% lower than that of BSL01 (0.066 ± 0.003 g L −1 h −1 ). The deletion of MIG1 also decreased the ethanol yield (Y ethanol ) of the strain BSHM01 by 30.4% (from 0.194 ± 0.015 to 0.135 ± 0.023 g g −1 xylose). These results suggest that the Mig1p is not phosphorylated and degraded in the presence of xylose like it is for non-fermentative carbon sources, since the deletion of MIG1 significantly changed the xylose fermentation characteristics. The strain BSHM02 showed similar xylose fermentation characteristics, but the ethanol yield was higher than in BSHM01. This suggests that the positive effects of Hxk2p S14A on xylose fermentation were eliminated by the deletion of MIG1. However, the mutant Hxk2p S14A protein still improved the ethanol production independent of Mig1p.

Transcriptional Profile of the Yeast Strains Expressing Nucleus-Localized Hxk2p S14A
To explore how Hxk2p S14A affects xylose metabolism in yeast, we compared the transcriptomic differences between strains BSHA01 and BSL01 cultured in SC-Ura medium with xylose as the sole carbon source. The results showed that 118 genes were up-regulated and 32 genes were down-regulated, respectively, in BSHA01 compared to BSL01 (Table S1).
Analysis of the differentially expressed genes showed that only a small number of the Hxk2p S14A -regulated genes are related to carbohydrate metabolism (Table 3). Among them, the only up-regulated gene was XKS1 that encodes xylulokinase, an enzyme that converts D-xylulose and ATP to xylulose 5-phosphate and ADP. The importance of XKS1 for xylose fermentation by recombinant S. cerevisiae strains has been well established [14]. This could be one reason that Hxk2p S14A enhances xylose fermentation. Furthermore, the down-regulation of COX2, COX3, and AI2, which are involved in respiration, and GPP1, which encodes glycerol-3-phosphate phosphatase, might be the reason that Hxk2p S14A increases the ethanol yield of the recombinant strains, since either decreased respiration or decreasing the glycerol production reduced the carbon flux of products other than ethanol. Table 3. Hxk2p S14A -regulated genes that were annotated to GO terms related to carbohydrate metabolism in the 'Biological Process' GO category. We then mapped the significant shared GO terms of the genes regulated by Hxk2p S14A using the Gene Ontology Term Finder tool provided by the Saccharomyces Genome Database (http://www. yeastgenome.org). In the 'molecular function' GO category (Figure 3), 51.7% of the up-regulated genes (61 of 118) encode nucleic acid binding (GO:0003676) proteins, and most of these (52 of 118 genes, 44.1%) are RNA-binding (GO:0003723) proteins ( Figure 3). This result suggested that the transcription and translation processes are more active in yeast strains expressing the nucleus-localized Hxk2p S14A protein compared to the wild type strain. This may be another way to explain the enhanced xylose utilization of the strain BSHA01.  . Clustered GO (gene ontology) terms for genes that show significant differential expression in the yeast strain expressing HXK2 S14A compared to its parent strain. (A) Down-regulated genes' clustered GO terms; (B) Up-regulated genes' clustered GO terms.

The Regulatory Targets of Hxk2p S14A in Cells Grown on Xylose Are Different than the Known Targets of Mig1p
Our aim was to analyze the correlation between Mig1p and Hxk2p S14A . We compared the regulatory target genes of Hxk2p S14A in cells cultured in xylose, which we obtained from our transcriptome analysis, with the known target genes of Mig1p, which were obtained from YEASTRACT+ (http://www.yeastract.com/formrankbytf.php) [34]. The results (Figure 4) showed that there are only a very few overlaps in the targets of Mig1p and Hxk2p S14A . The relative expression of several genes show opposite regulatory profiles by Mig1p and Hxk2p S14A . This suggested that, in yeast cells cultured in xylose, the regulatory targets of Hxk2p S14A are totally different than the known targets of Mig1p.

Discussion
Metabolic and evolution engineering strategies have been performed in S. cerevisiae, which is a robust and safe cell factory, to improve its ability to produce ethanol from xylose, which is the second abundant monosaccharide component in lignocellulose hydrolysates [1,2,4]. The state of yeast cells during xylose fermentation, such as the not fully activated glycolytic and pentose phosphate pathway, is very different to that during glucose fermentation. This is due to the response of yeast to xylose being greatly unalike to its response to glucose. Consistent with this, some perturbations to the signaling pathways have shown positive effects on xylose utilization in S. cerevisiae [20,25,35,36]. Further metabolic engineering work will require a better understanding of the relationship between xylose and the metabolic regulatory systems in S. cerevisiae. Hxk2p is not only a hexokinase but is also an intracellular glucose sensor that plays a significant role in the glucose Snf1p circuit and affects the PKA pathway in S. cerevisiae. About 15% of Hxk2p is located in the nucleus under high glucose conditions. It is known that nuclear-localized Hxk2p interacts with Mig1p to form a complex that then regulates the expression of glucose repression genes together. Under low-glucose conditions, Hxk2p moves out of nucleus and does not function as a regulator [29]. In this work, we studied the effects of expressing HXK2 S14A , which encodes the nucleus-localized Hxk2p mutant, and the deletion of MIG1 on xylose metabolism in recombinant yeast strains. In addition, we also investigated the relationship between the Snf1p-Mig1p-Hxk2p glucose signaling pathway and xylose.
In vitro experiments confirmed that Hxk2p becomes irreversibly inactivated by xylose through an autophosphorylation mechanism in the presence of ATP [37]. Due to the prominent role of Hxk2p in glucose repression signaling, Bergdahl et al. (2013) hypothesized that the inactivation of Hxk2p by xylose leads to a reduced regulatory capability of Hxk2p. They also confirmed that overexpressing the anti-xylose mutation Hxk2p F159Y resulted in an increased rate of xylose uptake and a faster initial growth rate on xylose [29]. Our results show that the expression of Hxk2p S14A enhanced both xylose utilization and ethanol production. This result suggests that the nucleus-localized mutant Hxk2p S14A retains its regulatory function in the presence of xylose, even though it may lose its enzyme activity. The transcriptomic differences between strains expressing Hxk2p S14A and the wild-type Hxk2p also support the notion that Hxk2p S14A possesses a regulatory function. The expression level of XKS1 in BSHA01 was up-regulated, which may account for the increased xylose utilization, and the expression levels of COX2, COX3, AI2, and GPP1 were down-regulated, which may explain the increase in ethanol production.
The deletion of MIG1 decreases xylose utilization, which means that Mig1p is not degraded in the presence of xylose like it is for non-fermentative carbon sources. Otherwise, the mig1∆ strain would show no difference compared to the wild-type strain when cultured in xylose. The deletion of MIG1 eliminated the positive effect on xylose utilization, suggesting that Mig1p is more important than Hxk2p to xylose utilization. Moreover, we found that the regulatory targets of Hxk2p S14A for cells growing on xylose are very different to the known targets of Mig1p. This result supports the hypothesis based on GST pulldown assays that the Hxk2p-Mig1p interaction is abolished by xylose [26]. Considering both of the conclusions of our study ((1) Mig1p is not degraded in xylose, and (2) the Mig1p does not interact with Hxk2p in cells growing on xylose), and the fact that Mig1p is phosphorylated by active Snf1p and then degraded when Mig1p separates from Hxk2p, we further conclude that the activity of Snf1p is not (at least not fully) activated in our strain when cultured in xylose-containing medium so that Mig1p is not phosphorylated by Snf1p and then degraded even it does not interact with the Hxk2p.

Conclusions
We found that the expression of the nucleus-localized mutant Hxk2p S14A increased xylose utilization and ethanol production in our engineered yeast strain. This may be due to the coordinated up-regulation of XKS1 and the down-regulation of COX2, COX3, AI2, and GPP1. The deletion of MIG1 decreased xylose utilization and eliminated the positive effect of Hxk2p S14A . Most regulatory targets of Hxk2p S14A are RNA-binding proteins in cells grown on xylose, and they are totally different from the known Mig1p targets. Our results reveal the regulatory profile of nucleus-localized Hxk2p S14A , which is independent of Mig1p in yeast cells cultured on xylose.