Construction of an economical xylose-utilizing Saccharomyces cerevisiae and its ethanol fermentation

Abstract Traditional industrial Saccharomyces cerevisiae could not metabolize xylose due to the lack of a specific enzyme system for the reaction from xylose to xylulose. This study aims to metabolically remould industrial S. cerevisiae for the purpose of utilizing both glucose and xylose with high efficiency. Heterologous gene xylA from Piromyces and homologous genes related to xylose utilization were selected to construct expression cassettes and integrated into genome. The engineered strain was domesticated with industrial material under optimizing conditions subsequently to further improve xylose utilization rates. The resulting S. cerevisiae strain ABX0928-0630 exhibits a rapid growth rate and possesses near 100% xylose utilization efficiency to produce ethanol with industrial material. Pilot-scale fermentation indicated the predominant feature of ABX0928-0630 for industrial application, with ethanol yield of 0.48 g/g sugars after 48 hours and volumetric xylose consumption rate of 0.87 g/l/h during the first 24 hours. Transcriptome analysis during the modification and domestication process revealed a significant increase in the expression level of pathways associated with sugar metabolism and sugar sensing. Meanwhile, genes related to glycerol lipid metabolism exhibited a pattern of initial increase followed by a subsequent decrease, providing a valuable reference for the construction of efficient xylose-fermenting strains.

employs corn as a r aw material, second-gener ation cellulose ethanol made from straw is cheaper and more environmentally friendly (Dos Santos et al. 2016a ).Lignocellulosic raw materials such as corn stover are collected and pretreated to break down the lignin and cellulose firstly, then hydrolyzed through enzymes to release low-carbon sugars such as glucose and xylose, and fermented with industrial yeast strains to produce ethanol subsequentl y (Hahn-Häger dal et al. 2007 ).The y east str ain Sacc harom yces cerevisiae commonl y utilized for first-gener ation ethanol pr oduction is firstl y assessed for fermenting xylose but a ppears to lack the ability (Patiño et al. 2019, Stambuk 2019 ).Ther efor e, geneticall y engineer ed str ains wer e constructed afterw ar ds to enhance the metabolism of xylose (Sharma and Ar or a 2020 ).
The main limitations of xylose metabolism in S. cerevisiae include xylose transportation, cofactor requirements of xylosemetabolizing enzymes and xylitol accumulation.To date, various str ategies hav e been de v eloped to modify the xylose metabolic pathway in S. cerevisiae (K uyper et al. 2005a, b , Runquist et al. 2010, Sedlak and Ho 2004, Zhou et al. 2012 ).Gener all y, ther e ar e two pathways to transform xylose to xylulose in natural microorganisms, including a pathway via xylose isomerase (XI) and a pathway via xylose reductase (XR) and xylitol dehydrogenase (XD) (Fig. 1 A).Engineered industrial S. cerevisiae strains that can effectively ferment and metabolize xylose to produce ethanol are constructed thr ough heter ologous expr ession of xylose metabolism pathway.XR and XD ar e endogenousl y expr essed in S. cerevisiae while they are not as efficient due to redox imbalance.By changing the affinities of XR and XD for NADH and NADPH, it is possible to decrease xylitol accumulation and increase ethanol production (Almeida et al. 2011, Jo et al. 2016 ).In addition, a recent study proposed a strategy that involves integrating metabolic engineering with sugar sensing and signalomics in yeast strains to resolve the xylose par adox (Osir o et al. 2019 ).Compar ed with the r edox pathway, the expression of xylose isomerization pathway would not cause coenzyme imbalance due to accumulation of intermediate xylitol, which is an ideal pathway for xylose metabolism (Brat et al. 2009, Zhou et al. 2012 ).After catal ysis fr om xylose, xylulose is phosphorylated by endogenous xylulokinase (XK) and channelled into the pentose phosphate pathway (PPP) and gl ycol ysis to pr oduce ethanol (Van Maris et al. 2006 ).
Pr e vious r esearc h has identified a number of pr ocesses as bottlenec ks that r estrict the xylose utilization (XU) efficiency, including the uptake of xylose (Runquist et al. 2009 ), the accumulation of xylitol (Zhu et al. 2021 ), the conversion of xylulose (Zhou et al. 2012 ), and the flux of the PPP pathway (Johansson and Hahn-Hägerdal 2002 ).One of the most straightforw ar d strategies to diminish the bottlenecks is the overexpression of xylose utilizing related proteins (Kuyper et al. 2005a, Runquist et al. 2009 ).Increasing the copy number of XI encoding gene xylA has been identified as an important strategy to improve the xylose metabolism and ethanol production efficiency of yeast strains.After artificial domestication of recombinant bacteria with multiple copies of the gene in δ integration site, the gene copy number of the best two strains with xylose as the only carbon source increased to 36 and 26, r espectiv el y, while the contr ol str ain had no significant xylose consumption (Dos Santos et al. 2016b ).In addition, studies sho w ed that the consumption rate of xylose mutants obtained from the domestication of the recombinant strain was significantl y incr eased, along with the ethanol yield of as high as 0.43 g/g xylose .T he detection sho w ed that the cop y number and mRNA le v el of the xylA gene of the two domesticated strains were 8-32 times higher than those of the control strain (Zhou et al. 2012 ).T he abo ve studies indicated that the enzyme activity of XI was still the main limiting factor in XI pathwa y strains .XK encoded by gene XKS1 is another necessary gene on the downstream metabolic pathway for both XR-XDH and XI pathwa ys .Ov er expression of XKS1 gene can significantly reduce the accumulation of xylitol (Hohenschuh et al. 2015, Kim et al. 2013a ).The activity of XKS1 in wild-type S. cerevisiae is too low to meet the basic r equir ements of xylose fermentation for industrial transformation (Toivari et al. 2001 ).Moreover, compared with other yeasts, the metabolic flux of the nono xidati ve phase PPP in S. cerevisiae is m uc h lo w er (Öhgren et al. 2006 ).As the sole pathw ay for d -xylose metabolized to the central metabolism, PPP is critical for pentose metabolism and r equir es higher expr ession (Latimer and Dueber 2017 ).Ov er expr ession of all PPP genes has been demonstrated to impr ov e the xylsoe metabolism (Kobayashi et al. 2017) (Kobayashi et al. 2018 ).In addition to gene manipulation, m uta genesis, or ada ptiv e domestication is also an essential step for increasing XU efficiency and obtaining stable expression hosts (Kuyper et al. 2005a, Shen et al. 2012, Zhou et al. 2012 ).Dir ected e volution tec hniques have been applied to sugar transporter proteins, enabling glucose/xylose cotransport and eliminating glucose inhibition in yeast (Li et al. 2016, Xu et al. 2020 ).Another study emplo y ed adaptive domestication to enhance the xylose consumption kinetics of an engineer ed str ain under aer obic conditions, r esulting in a 70% increase in xylose consumption rate (Thalita Peixoto et al. 2022 ).
Despite a number of studies ac hie ving the pr oduction of cellulose by S. cerevisiae , these strains are still difficult to work in actual industrial pr oduction, mainl y due to two r easons .T he first one is the selection of hosts, most hosts selected are laboratory haploid str ains, whic h ar e less r obust than industrial diploid str ains for fermentation (Karhumaa et al. 2005, 2007, Kuyper et al. 2005b ).Another reason is the expression of proteins, most of them are expression based on plasmids, which are less stable than those integrated into genome (Meinander andHahn-Hägerdal 1997 , Zhang et al. 1996 ).There have been attempts to integrate the XI into genome, but with a laboratory haploid strain (Tanino et al. 2010 ).Another r esearc h integr ating se v en differ ent genes r elated to XU into genome was reported based on an industrial diploid strain (Demeke et al. 2013 ), while the XU efficiency in industrial material such as corn stover hydrolysate still needs improvement.
As mentioned earlier, the main challenge of ethanol production through xylose-fermenting lies in the utilization efficiency of industrial materials.Ho w e v er, effectiv el y expr essing XI in an industrial strain and testing it with real industrial production are r ar el y r eported.To this end, the industrial diploid strain S. cerevisiae CY was chosen as the chassis host, involving a strategy that combines genetic engineering with ada ptativ e e volution engineering.Genes related to XU were selected to be integrated into yeast c hr omosomes.Two expr ession cassettes including heterogenous XI gene Pi-xylA origin fr om Pirom yces , and endogenous genes XKS1 , RPE1 , RKI1 , TAL1 , and TKL1 were constructed for the integration.Subsequentl y, ada ptiv e e v olution w as conducted in xyloserich media under both laboratory conditions and industrial conditions.By utilizing a combinatorial str ategy, the best-e volv ed str ain, ABX0928-0630, demonstrated a superior ability to utilize xylose and a higher sugar-alcohol conversion (SAC) efficiency compared to the commercial yeast strain ABX-CIP1 and ABX-CIP2.

Strains, mediums, and culture conditions
The industrial commercial yeast strain S. cerevisiae CY used in this study was taken from Biotechnology Center, COFCO Nutrition Yeast strains were cultured with either YEPD medium (10 g/l yeast extract, 20 g/l peptone, and 20 g/l glucose), YEPX medium (10 g/l yeast extract, 20 g/l peptone, and 40 g/l xylose), or YNBX medium (0.67% yeast nitrogen base without amino acids and 40 g/l xylose) according to different requirements.Evaluation of flask fermentation was conducted on YEPDX medium (10 g/l yeast extract, 20 g/l peptone, 40 g/l xylose, and 80 g/l glucose).The enzymatic hydr ol ysis (EH) liquid in the processing of ada ptiv e r e volution, and the straw EH liquid extrudate used in the bench-and pilot-scale fermentation were from COFCO Bio-energy (Zhaodong) Co., Ltd.T he stra w gas explosion material, enzyme pr epar ation and water were mixed into the beaker and enzymatically hydr ol yzed in a water bath at 50 • C and pH 5.0 for 72 hours to harvest the straw EH solution.NEB ® 5-alpha Competent E. coli (New England BioLabs Ltd., Ipswich, MA, USA) was used for plasmid cloning in LB medium (10 g/l NaCl, 5 g/l yeast extract, and 10 g/l tryptone).

Plasmid and strain construction
XI encoding gene xylA was deriv ed fr om the Pirom yces sp.Str ain (Lee et al. 2012 ).The plasmids pUC18-FB A1p/TEF2p-2X-N AT (which contains individual expression cassettes for xylA and XKS1 gene) and pRS42K-TTRR-KanMX (which contains four nonoxidative PPP genes) were constructed by conventional cloning methods.For the pUC18-FB A1p/TEF2p-2X-N AT plasmids, either TEF2p (promoter of TEF2 ) or FBA1p (promoter of FBA1 ) was used to drive xylA expression and PGK1p (promoter of PGK1 ) was used to drive XKS1 expr ession.Genetic ma ps of these two plasmids are shown in Figure S1 ( Supporting Information ).
All str ains wer e constructed or domesticated fr om the industrial strain S. cerevisiae CY.The process of strain construction is de picted in Fig. 1 .Gi v en that FPS1 integr ation site is a single copy number integr ation site, whic h affects gl ycer ol efflux pr otein, in-activating the FPS1 should inhibit gl ycer ol efflux and adjust the direction of carbon metabolism to pyruvate .T he repeats of the δ-sites in the genome help with multiple integrations of expression cassettes (Cho et al. 1999, Yamada et al. 2010 ), the plasmids pRS42K-TTRR-KanMX and pUC18-FB A1p/TEF2p-2X-N AT w ere orderl y r ecombined into homologous FPS site and δ-site of the S. cerevisiae CY genome to avoid plasmid burden.

Adapti v e evolution of engineered strains
For the first 96 generations of adaptive evolution, single colonies wer e taken fr om the YEPD solid plate and inoculated into a YEPX (100 ml/250 ml) shake flask containing 40 g/l xylose.After 24 hours of shaking at 30 • C, 200 rpm, 1 ml of culture medium was inoculated into fresh YEPX medium.For the later 96 generations of ada ptiv e e volution, 1 ml of cultur e medium was taken fr om the former YEPX medium and inoculated into a shaker bottle containing 40 g/l xylose straw gas explosion material (40 ml/100 ml), After 48 hours of shaking at 30 • C, 200 rpm, 1 ml of culture medium was taken to inoculate fr esh str aw gas explosion material.Conditions of strain domestication at various stages are listed in Table S1 ( Supporting Information ).After e v ery 1-2 months of acclimatization, the fermented liquid is diluted and coated on YNBX plates to isolate the better-performing str ain.At eac h turn, 48 str ains were taken to conduct fermentation evaluation to select strains with high XU efficiency and high SAC efficiency for subsequent domestication.

Flask fermentation evaluation in different stages
Ev aluation of earl y-sta ge domesticated str ains was conducted in synthetic culture medium YEPDX in shake flasks.Yeast seeds were cultured in YEPD at 30 • C for 16 hours and then added into shake flasks with an inoculum OD 600nm = 1 (0.45 g stem cells/l of fermentation liquor) in 500 ml shake flasks containing 200 ml YEPDX medium.Then cultivated under anaerobic conditions at 30 • C for 72 hours.Evaluation of later-stage domesticated strains was conducted in industrial media.Yeast seeds were cultured in YEPX at 30 • C for 16 hours and then added into shake flasks with an inoculum OD 600nm = 1 in 500 ml shake flasks containing 200 ml straw EH liquid extrudate, 0.5 g/l nitrogen, 50 mg/l penicillin, and 10 mg/l stre ptom ycin.Then culti vated under anaerobic conditions at 30 • C for 72 hours .T hr ee par allel essays wer e set for eac h group.Glucose , xylose , and ethanol concentrations were detected by HPLC.Titer for ethanol, glucose , xylose , and xylitol were expressed in g/l, yields were expressed in g/g sugar, rates were expressed in g/l/h.The XU and SAC efficiency were calculated according to the following formulas: XU efficiency r epr esents the consumption of total xylose fr om the initial xylose concentr ation.SAC efficiency r epr esents the ratio of ethanol production with sugar consumption according to stoichiometry.

Pilot-scale fermentation
Pilot-scale fermentation was carried out using 30 m 3 fermenters.Corn stalks were cut into small sections for soaking in a 20-m 3 enzymol ysis de vice for 5 minutes, maintaining a pr essur e of 1.25 M Pa at 190 • C for 3 minutes to obtain a corn stover steam explosion material; adjust the moisture content of the steam explosion material to 53wt%, then add 72wt% sulfuric acid solution to mix e v enl y, add 8wt% of cellulase to EH at 50 • C and pH 5.0 for 72 hours to obtain an enzymaticall y hydr ol yzed matur e mash.The inoculum was controlled at 0.45 g stem cells/l of fermentation liquor to carry out ethanol fermentation of the enzymatically hydrolyzed mature mash with a loading volume of about 16 m 3 .Fermentation par ameters wer e optimized pr e viousl y: at the temper atur e of 32 • C, with no aeration, at pH of about 4.5, and penicillin dosage of 90 g to conduct fermentation for 56 hours.Samples were taken e v ery 8 hours to determine the concentration of glucose , xylose , and ethanol concentrations by HPLC.Three parallel assays were set for each fermentation.

Transcriptome sequencing and data analysis
After incubating in straw EH liquid extrudate at 30 • C, samples for transcriptome sequencing were taken at 36 hours .T he samples wer e stor ed at −80 • C until pr ocessing.The fr a gments wer e sequenced on the BGISEQ-500 platform.Ther e wer e 21.30 M output data with a comparison rate of 93.02% of each sample generated.HISAT(v2.1.0)was used to map the raw reads to yeast genome .T he S288C genome ( https://www.ncbi.nlm.nih.gov/ genome/ ?term=s288c ) was used as the r efer ence .T he ClusterGVis ( https:// github.com/junjunlab/ ClusterGVis ) (Li et al. 2023, Zhang 2022 ) and ClusterPr ofiler pac ka ge (Yu et al. 2012 ) were used for cluster analysis and enrichment analysis to illustrate the trend of relative expression at various domestication stages.

Rational construction of xylose fermentation S. cerevisiae
Due to the lack of efficient XI, those heterologous isomerases deriv ed fr om fungi and thermophilic bacteria with high effi-ciency ar e a pplied for constructing xylose metabolic pathway in S. cerevisiae (Ha et al. 2011 ).Here, a 439-amino acid long XI origin from Piromyces sp. was selected for modification ( Table S2 , Supporting Information : SEQ3 , SEQ4 ).A number of potential loci were selected for mutation tests based on the elucidated catal ytic mec hanism in our pr e vious work (Yi et al. 2019 ).The m utant with an Ala144Thr amino acid substitution was selected and synthesized to construct the XI expression cassette ( Table S2 , Supporting Information : SEQ1 , SEQ2 ).
Subsequentl y, the r ecombinant plasmid containing the modified Pi-xylA gene, XKS1 and four nonoxidase genes encoding the PPP ( RKI1 , RPE1 , TKL1 , and TAL1 ) were introduced into the S. cerevisiae CY for ov er expr ession (Fig. 1 A).The modified strain was cultur ed and scr eened in the medium YEPX.Through screening, six recombinant yeast strains ( δ-TEF2p-2X-NAT-3#, -4#, -5#, δ-FB A1p-2X-N AT-2#, 3#, and 14#) with the potential to co ferment C5 and C6 sugar to produce ethanol were obtained.The gr owth curv e and XU curve of these six strains were studied, and it was found that these engineering strains had poor adaptability to the envir onment ( Figur e S2 , Supporting Information ).In order to solve this problem, we further domesticated these six strains in YEPDX medium.After seven rounds of domestication, the OD value of δ-TEF2p-2X-NAT-5# strain increased significantly and began to produce odor of alcohol.Ther efor e, we selected the se v enth gener ation of δ-TEF2p-2X-NAT-5# as the starting strain for the next step of ada ptiv e e volution and r enamed it ABX0421 (Fig. 1 B).

Adapted to evolution for rapid xylose fermentation
After se v en gener ations of domestication, the gr owth r ate of δ-TEF2p-2X-NAT-5# strain has been greatly improved (ABX0421), but the gr owth r ate and the XU efficiency ar e still slow e v en under aerobic conditions.In order to isolate random mutants with incr eased gr owth r ate and XU efficiency, ABX0421 str ain was continuousl y tr ansferr ed in xylose-supplemented YEPX medium under aerobic conditions.After 10 times of transformation, the SAC efficiency of the strain improved from 0% to 37%, and the XU efficienc y w as impr ov ed fr om 2% to 77% in YEPX medium.But ther e were no significant trend changes in SAC rate and XU rate in the YEPDX medium, the ethanol yield r eac hed 0.224 g/g sugar with a large amount of xylitol produced, the resultant strain was named ABX0601 ( Figure S3 , Supporting Information ).
Then ABX0601 was further domesticated in YEPX medium for 89 generations and the xylitol production decreased significantly (Fig. 1 B; Figure S3 , Supporting Information ).We took sever al samples fr om differ ent sta ges, including the 17th (ABX0601), 50th (ABX0701), 78th (ABX0805) generations, and 96th (ABX0918).These strains were then subjected to shake flask fermentation experiments in YEPDX medium.Yeast cells were withdrawn at 76 hours for HPLC detection to obtain the fermentation curve and calculate the SAC efficiency and XU efficiency.The results sho w ed that after long-term domestication, the ethanol yield incr eased fr om 0.224 g/g sugar to 0.416 g/g sugar, the SAC efficiency incr eased fr om 58% to 88%, and the XU efficiency increased from 25% to 99.71% (Fig. 2 A; Table S3 , Supporting Information ).Even though the strain sho w ed excellent fermentation performance in the YEPDX medium, the performance in the EH medium was still not outstanding enough.Hence, to further enhance the performance in EH medium, domestication in industrial conditions was performed by gr aduall y incr easing the medium pr essur e; str ains in different stages were taken for the monitoring of c har acteristics, and the fermentation liquid was diluted and coated on YNBX plates to isolate the best-performing strain ( Figure S4 , Supporting Information ).The strain ABX0928-0630 sho w ed the best performance and the ethanol yield increased from 0.416 g/g sugar to 0.465 g/g sugar, the SAC efficiency incr eased fr om 71.56% to 92.31%, the XU efficiency incr eased fr om 25.11% to 96.15% in EH medium (Fig. 2 B; Table S3 , Supporting Information ).
To test the superiority of the final domestication strain ABX0928-0630 over a commercialized strain, we performed shakeflask fermentation with the commercial strain ABX-CIP2 as a control in EH medium.The glucose utilization efficiency and the XU efficiency of ABX0928-0630 were higher than those of ABX-CIP2, while the SAC efficiency of ABX0928-0630 was slightly inferior (Fig. 2 C and D), indicating a pr efer ential performance of the domesticated str ain ov er ABX-CIP2.In addition, ABX0928-0630 took less time to r eac h a stable value during fermentation, which can effectiv el y sav e time in commercial industrial cellulose ethanol fermentation.
Besides, based on the performance of strains reported in the last 10 years, the ethanol productivity and XU achieved in this r esearc h demonstr ate a clear adv anta ge compar ed to most of the studies reported in Table 1 .When cultured in the YEPDX medium, the engineer ed str ain shows a high xylose consumption rate and ethanol productivity in 24 hours compared with other studies.Although the ethanol productivity of this strain with straw EH liquid extrudate during the first 24 hours is low, which is caused by the str esses fr om the medium, the engineer ed str ain still demonstr ates a r e w ar ding performance of ethanol yield as w ell as XU and SAC efficienc y.It's w orth mentioning that the xylitol accumulation of the engineered strain in this research is very low no matter in YEPDX medium or industrial straw EH liquid extrudate medium, which helps avoid production inhibition by xylitol.

Pilot-scale fermentation compared with commercial strains
To e v aluate the fermentation performance of ABX0928-0630 in a larger industrial fermentation environment, we conducted a fermentation test of ABX0928-0630 with real material in a pilot-scale fermenter (30 m 3 ) at a biochemical factory (COFCO, Zhao Dong) and compared with tw o commer cial industrial production strains (ABX-CIP1 and ABX-CIP2).Samples were taken every 8 hours to determine the sugar alcohol concentration of the mash ( Figure S5 and Table S4 , Supporting Information ).
The fermentation data of the three strains sho w ed that the final strain ABX0928-0630 after adaptive evolution can convert 56.43 g/l glucose and 24.35 g/l xylose to 38.25 g/l ethanol in 48 hours at an initial OD600 of 1.0 (0.63 g DCW/l) without producing xylitol, the ethanol yield r eac hed up to 0.48 g/g sugar, which is higher or the same le v el as the two commercial strains ABX-CIP1 and ABX-CIP2 (0.42 g/g sugar and 0.48 g/g sugar, r espectiv el y) (Fig. 3 A).The volumetric xylose consumption of ABX0928-0630 during the first 24 hours could r eac h up to 0.76 g/l/h, which is higher than that of ABX-CIP1 and ABX-CIP2 (0.43 g/l/h and 0.53 g/l/h, r espectiv el y) (Fig. 3 B).T he a v er a ge XU efficiency of ABX0928-0630 r eac hed 97.7%, whic h is m uc h higher than that of ABX-CIP1 (65.6%) and ABX-CIP2 (85.6%).The av er a ge SAC efficiency of ABX0928-0630 was 92.9%, while that of the ABX-CIP1 and ABX-CIP2 were 81.9% and 93.3% r espectiv el y (Fig. 3 C).The time-rate curves of both the total SAC efficiency and XU efficiency shown in Fig. 3 (D) show the same trend: ABX0928-0630 has the highest total SAC and XU efficienc y, follo w ed b y ABX-CIP2, and the lo w est was exhibited by ABX-CIP1.The fermentation with ABX0928-0630 in the pilot-scale test was better than fermentations with other strains, indicating that this strain is suitable for pilot production and has production capacity for industrial applications.Sugar: G for glucose and X for xylose; XU: xylose utilization; Y: yield, g/g sugar; P: productivity, g/l/h; and C: consumption rate, g/l/h.For Y xylitol , all yields less than 0.01 were marked as and higher than 0.01 were marked as 4 .

Tr anscriptome anal ysis throughout domestication
Through metabolic engineering, we enhanced the adaptation of S. cerevisiae growth on xylose with glucose.Howe v er, it was still not clear how the cellular metabolism was r esha ped for this change.
To explore the gene expression in the cellulose metabolism path-way at different stages of adaptive evolution, strains at different stages of domestication (represent the initial industrial strain, the initial domesticating strain, the final strain domesticated in synthetic medium and the final strain domesticated in industrial medium, r espectiv el y) wer e selected for tr anscriptome sequencing.Based on pr e viousl y obtained fermentation data, the yeast strains can run out of glucose and begin to utilize xylose between 24 and 48 hours.We presume that this stage may be when cellulose metabolism is most active and the r ele v ant gene expr ession in the pathway is the highest.Thus, we performed shake-flask fermentation in EH medium with these strains and withdr e w the cells at 36 hours for RNA sequencing.
To compr ehensiv el y understand the entir e pr ocess of ada ptiv e e volution and r e v eal pr e viousl y uncov er ed tr aits r elated to xylose metabolism, we used all transcripts to perform principal component analysis.As shown in Fig. 4 (A), good agreement between ABX0401 and CY, ABX0805 and ABX0701 suggesting the transcriptional similarity of physiological status in each pair.To define the temporal characteristics of the complete transcriptome dataset, we performed clustering analysis across ABX0401, ABX0601, ABX0805, ABX0918, and ABX0928-0630.Among the total expressed genes, eight clusters exhibited distinct patterns among the fiv e gr oups wer e identified (Fig. 4 B).The genes in each cluster possessed distinct functions as r e v ealed b y KEGG pathw ay enrichment analysis.
We first analyzed the regulation of relevant genes in several metabolic pathways that are most pertinent to xylose metabolism ( Table S5 , Supporting Information ).Genes in clusters C1 and C7 sho w ed a significant increase during domestication, and the functions of these genes wer e enric hed in galactose metabolism, biosynthesis of nucleotide sugars, starch and sucrose metabolism, amino sugar and nucleotide sugar metabolism, fructose and mannose metabolism, proteasome, and meiosis .T he externally integr ated gene XKS1 , r esponsible for xylulose phosphorylation, sho w ed a significant increase in transcription levels through domestication.Continuing with the conversion of xylulose-5P into the PPP, the externally integrated TTRR ( TKL1 , TAL1 , RPE1 , and RKI1 ) genes also sho w ed a substantial impr ov ement in tr anscription.Subsequentl y, within the metabolic pathway from Gl ycer oldehyde-3P to ethanol, most genes exhibited upregulation, such as PGK1 , PYK1 , and PDC1 and ADH6 (Fig. 4 B).As mentioned abo ve , xylose transportation relies on hexose transporters, genes related to this, including SOR1 and YNR071C sho w ed a significant incr ease whic h suggests an increased efficiency of sugar uptake during domestication.IZH3 , GCR1 , and HSP30 , whic h ar e r elated to sugar sensing and signalomics were upregulated, indicating an enhanced ability of the engineered strains for sugar sensing and ethanol response.
Notably, genes in cluster C6 were enriched in the biosynthesis of secondary metabolites, including 2-o xocarbo xylic acid metabolism, biosynthesis of amino acids , TC A cycle and pyruv ate metabolism, whic h aligns with our pr e vious findings .T his implies that the increased utilization of xylose b y y east had an impact on cellular activity, ultimately leading to a reduction in secondary metabolism.In pathways leading to secondary metabolites, most of the TCA cycle genes show significant downregulation.This downr egulation pr e v ents the accum ulation of acetate byproducts and helps maintain pH levels.While the genes associated with gl ycer olipid metabolism gener all y exhibit an upr egulation trend, although many of them are subsequently downregulated as domestication pr ogr esses.We pr opose that these tr ends are a result of osmotic regulation and energy optimization within the cell.During the ada ptiv e domestication, the composition of the culture medium changed from a mixture of glucose and xylose to xylose alone and then to industrial materials, transitioning fr om r elativ el y mild to complex and se v er e conditions.Consequentl y, the initial incr ease in gl ycer ol pr oduction could be viewed as a pr otectiv e r esponse to the more severe living conditions.As the domestication process continued, the engineered strain graduall y ada pted to the harsher gr owth envir onment and tended to adopt a more energy-efficient production mode to achieve balance, resulting in the downregulation of these genes .T he optimization ability is mainly due to ada ptiv e domestication in industrial conditions and not only maintains a transcription balance but also avoids metabolic disturbances with multiple carbon resources (Infante et al. 2003 ).
Another major problem in simultaneous cofermentation of glucose and xylose is the inhibition effect of carbon decomposition (CCR effect).Xylose transportation relies on hexose transporters, ho w e v er, xylose fermentation is not as efficient as glucose fermentation in S. cerevisiae .The significantl y incr eased tr anscript le v els of a series of GAL genes in galactose metabolism pathway, such as GAL1 , GAL2 , GAL7 , GAL10 , and GAL80 , ar e involv ed in glucose-r epr essible galactose metabolism.Their gratuitous induction suggests a probable loss of glucose r epr ession (Baleja et al. 1997 ).

Discussion
Xylose is the second-highest monosaccharide in lignocellulose hydr ol ysate.In the pr oduction of cellulose ethanol using lignocellulose hydr ol ysate as r aw material, efficient and full utilization of xylose is an essential c har acteristic of fermentation strains.Ho w e v er, natur al br e wing yeast cannot metabolize xylose .T herefor e, obtaining yeast str ains that can efficientl y metabolize xylose through engineering modification of brewing yeast is a k e y issue in cellulose ethanol pr oduction.Pr e vious studies have been focused on hosts of laboratory haploid strains and expression based on plasmids, whic h wer e ineffectiv e in actual industrial pr oduction (Selim et al. 2018 ).In this study, the industrial diploid strain was chosen as the chassis host, and genes related to XU were integrated into yeast chromosomes.In the following, adaptive evolution, real industrial fermentation and transcriptome analysis were conducted sequentially.Our results indicate that the beste volv ed str ain demonstr ated a superior ability to utilize xylose and a higher SAC efficiency compared to the commercial yeast strain.
Transcriptome data sho w ed a significant increase in the expr ession le v el of hexose tr ansporters, indicating an incr eased efficiency in sugar uptake during domestication.The genes related to sugar sensing and signal omics were upregulated, indicating an enhanced ability of the engineered strain to respond to sugar sensing and ethanol.The expression levels of other genes r elated to pr ocesses suc h as galactose metabolism, nucleotide sugar biosynthesis, starch and sucrose metabolism, amino and nucleotide sugar metabolism, fructose and mannose metabolism, proteasome, and meiosis also significantly increased during domestication.The increase in XU by yeast has an impact on cell activity, ultimately leading to a decrease in secondary metabolism.This downregulation can prevent the accumulation of acetate byproducts and help maintain pH levels.With the progress of domestication, genes related to glycerol lipid metabolism show a trend of increasing first and then decreasing, which may be r elated to intr acellular osmotic r egulation and ener gy optimization.We propose that these trends are a result of osmotic regulation and energy optimization within the cell.During the ada ptiv e domestication, the composition of the culture medium changed from a mixture of glucose and xylose to xylose alone and then to industrial materials, tr ansitioning fr om r elativ el y mild to complex and se v er e conditions.Consequentl y, the initial incr ease in gl ycer ol pr oduction could be vie wed as a pr otectiv e r esponse to the mor e se v er e living conditions.As the domestication pr ocess continued, the engineered strain gradually adapted to the harsher gr owth envir onment and tended to adopt a mor e ener gy-efficient production mode to achieve balance, resulting in the downregulation of these genes .T he optimization ability is mainly due to adaptive domestication in industrial conditions and not only maintains a transcription balance but also avoids metabolic disturbances with multiple carbon resources (Infante et al. 2003 ).Galactose transporters can transport xylose without being inhibited by glucose, and xylose transporters can achieve simultaneous consumption of glucose and xylose, which provides the possibility of ac hie ving efficient coutilization of glucose and xylose.

Conclusions
Through genetic modification and continuous domestication, we have obtained the engineered strain ABX0928-0630, and the fermentation of corn stalk-based raw materials with ABX0928-0630 increased the ethanol yield by ∼30% in 30 m 3 fermenters under pilot-scale conditions, decreasing the cost of raw material by about $150-175 per ton ethanol.With a shorter fermentation time and higher ethanol conversion efficiency and XU efficiency, the strain ABX0928-0630 can effectively decrease production cost and time, which indicates a better commercial application prospect in ethanol fermentation.
Compared with the XR-XDH pathw ay, the XI pathw ay has a higher ethanol yield during xylose fermentation.In this study, we constructed a r ecombinant str ain, but further r esearc h is needed on the integration sites of genes on the genome and how to impr ov e enzyme activity.Mor eov er, this study explains the changes in gene expression during the process of modification and domestication at the transcriptome level, and these research findings can be used to guide future genetic modification work.In addition, glucose and xylose are the two main components of lignocellulose .T he simultaneous consumption of glucose and xylose is crucial for the production of fuel and chemicals using lignocellulose hydr ol ysis pr oducts as r a w materials .In futur e r esearc h, further attention needs to be paid to the issue of gene transcription inhibition related to xylose metabolism, as well as the exploration of new xylose transporters, in order to achieve efficient coutilization of glucose and xylose.

Figure 1 .
Figure 1.XI-driven xylose assimilation in engineered yeast.(A) Schematic illustration of the xylose assimilation module.(B) Schematic diagram of strain construction and domestication.

Figure 2 .
Figure 2. Domestication, screening, and fermentation result of engineered strains.(A) XU and SAC efficiency of strains at different stages in YEPDX medium and (B) EH medium.(C) Glucose , xylose , ethanol concentrations , and (D) XU efficiencies, SAC efficiencies of ABX0928-0630 compared with ABX-CIP2 in EH medium.

Figure 3 .
Figure 3. Pilot-scale fermentation of ABX0928-0630 compared with commercial strains ABX-CIP1 and ABX-CIP2.(A) Bar chart of the ethanol yield within 24 and 48 hours and (B) xylose consumption rate within 24 and 48 hours.(C) Bar chart of the XU efficiency and SAC efficiency.(D) The time-rate curve of the total SAC efficiency and XU efficiency.

Figur e 4 .
Figur e 4. (A) T he number of differ entiall y expr essed genes (DEGs) (bcv = 0.2, P < .05)during differ ent sta ges of domestication.(B) Differ ential expression patterns across different domestication stages .T he DEGs between ABX0401, ABX0601, ABX0805, ABX0918, and ABX0928-0630 were classified into eight clusters, C1-C8 r epr esenting differ ent patterns in r esponse to domestication.Gene size indicates the number of genes in eac h pattern.The genes and pathways highlighted with different colors in each cluster possessed distinct functions which were revealed by KEGG pathway enric hment anal ysis.
Initial Ethanol Content 0 .511 × ( Initial Glucose and Xylose Content − End Glucose and Xylose Content )

Table 1 .
Main fermentation performance indexes for the present study and reported studies