Metabolic Engineering of Saccharomyces cerevisiae for Enhanced Carotenoid Production From Xylose-Glucose Mixtures

Co-utilization of xylose and glucose from lignocellulosic biomass is an economically feasible bioprocess for chemical production. Many strategies have been implemented for efficiently assimilating xylose which is one of the predominant sugars of lignocellulosic biomass. However, there were few reports about engineering Saccharomyces cerevisiae for carotenoid production from xylose-glucose mixtures. Herein, we developed a platform for facilitating carotenoid production in S. cerevisiae by fermentation of xylose-glucose mixtures. Firstly, a xylose assimilation pathway with mutant xylose reductase (XYL1m), xylitol dehydrogenase (XYL2), and xylulokinase (XK) was constructed for utilizing xylose. Then, introduction of phosphoketolase (PK) pathway, deletion of Pho13 and engineering yeast hexose transporter Gal2 were conducted to improve carotenoid yields. The final strain SC105 produced a 1.6-fold higher production from mixed sugars than that from glucose in flask culture. In fed-batch fermentation with continuous feeding of mixed sugars, carotenoid production represented a 2.6-fold higher. To the best of our knowledge, this is the first report that S. cerevisiae was engineered to utilize xylose-glucose mixtures for carotenoid production with a considerable high yield. The present study exhibits a promising advantage of xylose-glucose mixtures assimilating strain as an industrial carotenoid producer from lignocellulosic biomass.


INTRODUCTION
Glucose from starch hydrolysates was widely used as raw material in industrial biotechnology (Buschke et al., 2013). In order to avoid competing directly with human nutrition derived from limited land resources, lignocellulose biomass containing the main building blocks of glucose and xylose was regarded as the most promising alternative raw material (Jagtap and Rao, 2018). Xylose is the second abundant sugar from lignocellulose which can account for 30-40% of lignocellulosic biomass (Kwak and Jin, 2017). Therefore, it is important to utilize both glucose and xylose from lignocellulosic biomass during manufacturing process for the economically feasible production of valuable chemicals.
Recent years, the microbial transformations of xylose into various chemicals, including xylitol, ethanol, butanol, squalene, 2,3-butanediol and lipid, were extensively studied (Kwak et al., 2019). However, the difficulty of engineering an efficient microbial cell factory for utilizing xylose hindered fully using lignocellulosic biomass (Zhang et al., 2019). Not only naturally xylose-assimilating yeasts, but also non-xylose assimilating yeasts have been developed as a platform for biosynthesis of biofuels and chemicals from lignocellulosic biomass. Especially, nonxylose-assimilating S. cerevisiae has been extensively studied for bioethanol production from xylose in lignocellulosic biomass due to its remarkable alcoholic fermentation (Kwak et al., 2019). Since S. cerevisiae lacked the capability of xylose assimilation, strategies such as engineering endogenous pathway were not viable. Instead, metabolic engineering strategies including introduction of heterogenous XYL1/XYL2 for xylose metabolism, overexpression of xylose isomerase or XK from various sources to obtain a balanced metabolic pathway, improvement of the transport efficiency through engineering heterogeneous or native xylose transporters, have been implemented to facilitate xylose utilization (Kim et al., 2013a). Nevertheless, strains widely used in industry lacked the ability to convert xylose into valuable products, and only few reports were available on making investigation of fermentation for isoprenoid production from xylose-glucose mixtures (Gao et al., 2018). For example, S. cerevisiae was firstly demonstrated to synthetize zeaxanthin from glucose or xylose (or xylan) (Sun et al., 2012).
The aim of the present study is to develop a S. cerevisiae platform that can utilize mixed sugars for carotenoid production (Figure 1). To illustrate improved carotenoid production from xylose-glucose mixtures, lycopene was chosen as a model carotenoid. We selected a basic lycopene-producing S. cerevisiae strain, namely BL03-D-4 (unpublished), as our starting strain. Firstly, XYL1m (NADH-dependent), and XYL2 were introduced into BL03-D-4 to utilize xylose. Then, introduction of PK pathway, deletion of Pho13 and engineering yeast hexose transporter Gal2 were applied. Consequently, a 2.6-fold higher production was obtained from xylose-glucose mixtures than that from glucose in the continuous fermentation. To the best of our knowledge, this is the first report that S. cerevisiae was engineered to utilize xylose-glucose mixtures for carotenoid production with a considerable high yield. A shorter carotenoid synthetic route and an optimized mevalonic acid (MVA) pathway might account for the higher carotenoid production compared with previous work (Sun et al., 2012). From the comparison of the model carotenoid production from mixed sugars and glucose by engineered S. cerevisiae, we conclude that the assimilation of mixed sugars is a practicable approach for the overproduction of carotenoid in S. cerevisiae from lignocellulosederived sugars.

Strains and Culture Conditions
Yeast strains and different plasmids used in this study are listed in Table 1. Primers are listed in Supplementary Table S1. E. coli DH5α was used for cloning and plasmid propagation. E. coli was grown at 37 • C and 200 rpm in Luria-Bertani (LB) medium (5 g L −1 yeast extract, 10 g L −1 tryptone, and 10 g L −1 NaCl; 20 g L −1 agar was added for solid medium). S. cerevisiae was cultivated in YPD medium (10 g L −1 of yeast extract, 20 g L −1 of peptone and 20 g L −1 of glucose).

Quantitative Analysis
Total carotenoid was extracted according to the literature with some modifications (Xie et al., 2015). Briefly, cell cultures were washed after 96 h cultivation, resuspended in 3 mol L −1 HCl and boiling for 4 min, then cooled in baths of ice for 4 min. Next, cells debris were washed twice and resuspended in acetone vortexed followed by centrifugation. The acetone supernatant was transferred into a new tube for measuring total carotenoid. The total carotenoid production of different strains was determined by UV/Vis spectrometer (PerkinElmer Lambda 45) at 470 nm (Supplementary Figure S1). The extinction coefficient was adopted in acetone using an A 1% 1cm of 3450 (Scott, 2001).
The growth curve of different strains was determined by monitoring OD 600 using a UV/Vis spectrometer. Dry cell weights (DCW) of different strains on glucose or mixed sugars were determined through drying the cells at 80 • C for 24 h. The yields of carotenoids were expressed as mg per g dry cell weight (mg g −1 DCW) (Xie et al., 2015). Concentrations of glucose and xylose in culture broth were analyzed by Glucose assay Kit and D-Xylose assay kit according to the protocol, respectively (Nangjing Jiangcheng Bioengineering Institute, Nanjing, China). The secretory xylitol was analyzed using D-sorbitol/xylitol enzymatic test kits (Megazyme Assay Kits, Megazyme International Ireland Limited, Wicklow, Ireland). Acetic acid concentration was measured with an acetic acid assay kit [Acetic Acid (GK) Kit, Megazyme].

Rational Engineering a Xylose-Assimilating Strain for Carotenoid Biosynthesis
In order to produce xylose-based carotenoid, the xyloseassimilating pathway including XYL1/XYL2 was introduced into the lycopene-producing strain S. cerevisiae BL03-D-4 to use xylose as a carbon source, which was necessary for xylosebased production. Meanwhile, the endogenous XK was also overexpressed through replacing promoter P XK with a relative strong promoter SSA1 to obtain SC101 (Figure 1). Then, shakeflask fermentation was performed for strain SC101 in 50 mL modified YP broth with 30 g L −1 xylose and 10 g L −1 glucose. The growth curve and xylose consuming were detected throughout the cultivation. The results (Figure 2) demonstrated that SC101 could assimilate xylose and obtain a slight higher biomass than BL03-D-4 but far below than the previous work . Approximately 18 g L −1 of xylose was consumed, and this result implied that a restrained metabolic flux through nonoxidative pentose phosphate pathway (PPP) existed in the present xylose-assimilating pathway (Gao et al., 2018). However, SC101 produced a 2.4-fold higher carotenoid than BL03-D-4 from mixed sugars. This result indicated that a xylose-assimilating pathway had a positive effect on carotenoid production.
As shown in Figure 3, the xylose-based carotenoid titer of SC101was similar with the glucose-based titer. These results revealed that the xylose consuming might be a rate-limiting step for carotenoid synthesis (Katahira et al., 2017). Thus, a mutant XYL1 (K271N, NADH preferred) (Bengtsson et al., 2009) was overexpressed to obtain SC102 for increasing xylose-based carotenoid production. This, in turn, more NAD might be available for the XDH reaction, and resulted in higher xylose assimilating rate. Although, we found that carotenoid yield could be increased in some extent (1.25-fold), the consumption of xylose was even slower than SC101 with the wild-type XYL1 ( Table 2). This was due to the fact that xylitol which was derived from xylose was accumulated in the strain with the wild-type XYL1 FIGURE 3 | (A) Product titers and (B) yields of different strains from glucose (light color) or xylose-glucose mixtures (dark color) in flasks fermentation. Cultures were carried out in modified YP medium with 40 g L −1 glucose or 10 g L −1 glucose and 30 g L −1 xylose at 30 • C for 96 h. Data represent the average ± SD of three biological replicates. Statistically significant differences are denoted **P < 0.01 and ***P < 0.001 (two-tailed Student's t-test).

Introduction of Phosphoketolase Pathway for Carotenoid Production
Acetyl-CoA is an important precursor for carotenoid synthesis and its biosynthesis was primary catalyzed by pyruvate decarboxylase (PDC), acetaldehyde dehydrogenase (Ald) and acetyl-CoA synthetase (ACS) in S. cerevisiae cytoplasm (van Rossum et al., 2016). Pyruvate was produced from glucose or xylose through EMP pathway, and then pyruvate was used as the precursor for acetyl-CoA synthesis via the PDH bypass (Kwak et al., 2019). There were many steps for converting glucose or xylose to acetyl-CoA. Therefore, a large proportion of xylose would be consumed for cell growth in the fermentation of xyloseglucose mixtures. Alternatively, in this study, strain SC103 was constructed for improving carotenoid yield through introduction of PK pathway which included xylulose-5-phosphate phosphoketolase (xPK) and phosphotransacetylase (PTA) (Kocharin et al., 2013;Meadows et al., 2016). Xylose-5-phosphate was converted to acetyl-CoA through PK pathway which could greatly retrench the pathway for acetyl-CoA synthesis from xylose (Figure 1). In the fermentation of mixed sugars, xylose was ideally consumed for biosynthesis of target products, while glucose was used for growth and supply cofactor (Su et al., 2015). To achieve this, the native transaldolase (TAL1) and transketolase (TKL1) which could greatly improve xylose assimilation were not overexpressed in this study (Gao et al., 2018). As shown in Table 2, the xylose consumption shown no improvement in SC103 compared to SC102. Meanwhile, overexpression of PK pathway exerted slight inhibition on the growth of strain SC103. The decreased cell growth might ascribe to that the expression of PK pathway was coupled with an increased acetate production ( Table 2; Bergman et al., 2019). The titers of carotenoid reached 64.9 and 75.9 mg L −1 from glucose and mixed sugars, respectively (Figure 3). Meanwhile, the yield of carotenoid from xylose-glucose was 17.1 mg g −1 DCW, which represented a 1.4-fold higher than that from glucose (12.2 mg g −1 DCW).

Deletion of Pho13 Exerts No Significant Effect on Xylose Consumption
In order to further improve xylose utilization, we assessed the effect of Pho13 deletion on xylose assimilating pathway. Previous study has reported that Pho13 deletion played an important role in improving xylose consumption rate and tolerance to toxic chemicals (Van Vleet et al., 2008). Herein, we deleted Pho13 gene in strain SC103 described above to generate strain SC104. By deleting Pho13, as shown in Table 2, SC104 obtained slightly increase of the cell growth and xylose assimilation. We were surprised by this result, for that XK was also overexpressed using a strong promoter SSA1 in SC104. This might due to that the facilitation of Pho13 deletion was only effective in strains which harbored xylose-assimilating pathway with quite high overexpression of XK (Kim et al., 2013b). For that Pho13 might cause a futile cycle by dephosphorylating xylulose-5-phosphate which would result in ATP waste and flux blockage along with highly active XK.

Engineering Hexose Transporter Gal2 for Improving Xylose Assimilating and Carotenoid Production
Strain SC104 was engineered to carotenoid biosynthesis from xylose-glucose, but xylose assimilating rate was very slow and might be a limitation for efficient carotenoid production. There were no significant improvements of intracellular xylose conversion through overexpression of PK pathway or Pho13 deletion. Thus, transport might be a bottleneck, especially in the xylose-glucose mixtures. In previous studies, engineering transports were used to improve engineered strains further (Patino et al., 2019). To facilitate xylose metabolism, the mutant hexose transporter Gal2 (N376F, not inhibited by glucose) (Farwick et al., 2014) was overexpressed with the promoter replacement with SSA1 in strain SC104 to obtain SC105. The subsequent strain SC105 exerted an increasing cell growth and xylose assimilating. About 20 g L −1 of xylose was consumed, which was significantly improved compared with SC104 ( Figure 2F). Carotenoid titers of SC105 from mixed sugars reached 117.5 mg L −1 which was 1.6-fold of that from glucose (Figure 3). The carotenoid yield from mixed sugar was 20.2 mg g −1 DCW, which was 1.5-fold higher than that from glucose (13.2 mg g −1 DCW). Meanwhile, we found that SC105 produced much lower carotenoid (20.2 mg L −1 ) when xylose was used as the sole carbon source than using mixed sugars (Figure 4). This result was different from that obtained in a previous work where the production yield from xylose shown remarkable improvement than that from glucose Lane et al., 2020). We also found that strains almost could not grow on YP medium without sugars ( Figure 5A). Furthermore, the ration of xylose and glucose had no obvious effect on the production of carotenoid in flask fermentation (Figure 5). These results were probably due to that glucose was exhausted quickly in the flask fermentation and there was insufficient cofactor for carotenoid synthesis subsequently (Su et al., 2015). This result also clearly indicated that the addition of glucose had an obvious effect on promoting xylose assimilation and this was similar with our previous work that E. coli was engineered for xylitol production from xylose-glucose mixtures (Su et al., 2015). Although our yield was far from the theoretical mass yields (25%, on consumed sugar), SC105 could FIGURE 4 | Product titers and yields of strain SC105 from different ration of xylose and glucose in flasks fermentation. Xylose: xylose as sole carbon source at 40 g L −1 ; 1:1: 20 g L −1 xylose: 20 g L −1 glucose; 2:1, 26 g L −1 xylose: 13 g L −1 glucose. Each value represents the average ± SD of three biological replicates. Statistically significant differences are denoted ***P < 0.001 (two-tailed Student's t-test).
be used as a starting strain for further systemic engineering (Rude and Schirmer, 2009).

Fed-Batch Fermentation in a 5-L Bioreactor
Since strain SC105 with mutant hexose transporter Gal2 showed much higher carotenoid contents from mixed sugars than that from glucose or xylose (Figure 5). We speculate that continuous feeding of glucose may facilitate xylose utilization for carotenoid production. As a proof of concept, a fed-batch fermentation with continuous feeding of mixed sugars was performed to increase the production of carotenoid. Strain SC106 recovered with modules of His3-Leu2-Ura3 was inoculated with 10% and cultured with YPD medium. When the initial glucose was consumed, xylose concentration was maintained below 50 g L −1 through continuous feedings of mixed sugar (600 g total sugars, xylose/glucose = 3:1). A supplementary fed-batch medium (100 g yeast extract with 10 g L −1 KH 2 PO 4 , 2.5 g L −1 MgSO 4 , 3.5g L −1 K 2 SO 4 , and 0.25 g L −1 Na 2 SO 4 ) was added together with mixed sugars through the fermentation. Strain SC106 could grow rapidly even though xylose was accumulated at high concentrations in the medium. Finally, in fed-batch fermentation with serial feeding of mixed sugar, the cell density of the engineered strain reached to 183, and 903 mg L −1 of carotenoid was produced, which was 2.6-fold of that from glucose fermentation (Figure 6).

DISCUSSION
Lignocellulosic chemicals are economically feasible target because lignocellulosic biomass is renewable, inexpensive, and abundant. As one of the mostly used host in industrial biotechnology, S. cerevisiae has an incomparable capacity to transform glucose from lignocellulosic sugars to bioethanol. Nevertheless, wild-type S. cerevisiae cannot assimilate xylose, the second most abundant sugar in lignocellulosic biomass. To unveil the potential advantages of utilizing mixed sugars over glucose by engineered S. cerevisiae for carotenoid production, lycopene was chosen as an example of heterologous carotenoid. After several rounds of genetic modification in previous lycopene-producing S. cerevisiae BL03-D-4, we have developed a platform capable of assimilating xylose-glucose mixtures for carotenoid production.
Natural xylose-utilizing yeasts can biosynthesize xylulose from xylose through the oxidoreductase pathway which is comprised of xylose reductase (XR) and xylulose dehydrogenase (XDH) (Kwak et al., 2019). Mostly native XRs prefer NADPH as cofactor, but XDH uses only NAD. Thus, a bottleneck in the XR/XDH pathway was generated by the cofactor imbalance between the NADPH-preferred XR and the NAD-dependent XDH. Subsequently, xylitol accumulation and lower production yields was generated (Su et al., 2013). Thus, a mutant XYL1m which was NADH preferred was used for increasing xylose-based carotenoid production in this study and a 25% improvement was obtained. However, Pho13 deletion slightly increased the cell growth and xylose consumption in our strain. This result might be ascribed to the fact that the positive effect of Pho13 deletion only appeared when gene expression responsible for xylulose synthesis was sufficiently high (Lee et al., 2014). Since constitutive promoter TDH 3 was employed for overexpression of XK, xylose consumption was increased significantly through Pho13 deletion (Ni et al., 2007;Van Vleet et al., 2008;Bamba et al., 2016). However, our observation was consistent with the previous report that Pho13 deletion had no significant effect on xylose consumption in strains with the expression of xylose isomerase under the control of TEF 1 promoter (Shen et al., 2012).
Apart from above engineering strategies, engineering xylose transporters were carried out in S. cerevisiae to restore the low capacity of endogenous hexose transporters Wang et al., 2016). Xylose-specific transporter with highly active was much-needed for economically feasible production of chemicals from lignocellulosic sugars (Kim et al., 2013a). However, most sugar transporters of S. cerevisiae have much higher specificity toward glucose than xylose (Hou et al., 2017). Xylose uptake might be one of the limited steps in the pathway for xylose assimilation by S. cerevisiae in the presence of glucose. For simultaneous utilization of xyloseglucose, an engineered transporter that was xylose specific and not inhibited by glucose was highly desirable. Previous studies have been exactly dedicated to eliminating glucose inhibition in hexose transporters while retaining their xylose transportation capacity (Farwick et al., 2014;Li et al., 2016;Wang et al., 2016). A mutant hexose transporter without glucose inhibition was overexpressed in our strain and the corresponding strain SC105 exhibited 26 and 23% improvement on cell growth and xylose consumption, respectively. And a 42% improvement on carotenoid production was obtained.
There were few studies on developing platform of yeast for carotenoids production from xylose except that native xylose-utilizing Phaffia rhodozyma and engineered S. cerevisiae (Rodriguez-Saiz et al., 2010;Sun et al., 2012). P. rhodozyma was intensively studied for biosynthesis of carotenoid from xylose which convert sugars into a wide variety of carotenoids despite limited consumption rate and growth on xylose. Meanwhile, S. cerevisiae has been engineered to synthesize zeaxanthin using xylan from birchwood. Previous work has demonstrated that xylose assimilating in engineered S. cerevisiae exerted improvement of isoprenoid production and highlighted great advantages of using xylose as sole carbon source for isoprenoid production . However, it would be a complicated situation for mixed sugar, because lignocellulosic biomass was mainly composed of glucose and xylose. Although the carotenoid yield obtained in our study was far from the values of previous work (Chen et al., 2016;Ma et al., 2019;Shi et al., 2019). High carotenoid yield was not the main purpose of this study, because a basic strain with a relatively low carotenoid production (with not many genetic modifications) was chose as the starting strain. Our main goal was to develop a platform that could efficiently utilize xylose-glucose mixtures for carotenoid production.

DATA AVAILABILITY STATEMENT
The datasets analyzed in this manuscript are not publicly available. Requests to access the datasets should be directed to zhuhh@gdim.cn.
DATA SHEET S2 | Raw Data.
Frontiers in Bioengineering and Biotechnology | www.frontiersin.org