D-Lactic Acid Production from Sugarcane Bagasse by Genetically Engineered Saccharomyces cerevisiae

Lactic acid (LA) is a promising bio-based chemical that has broad applications in food, nutraceutical, and bioplastic industries. However, production of the D-form of LA (D-LA) from fermentative organisms is lacking. In this study, Saccharomyces cerevisiae harboring the D-lactate dehydrogenase (DLDH) gene from Leuconostoc mesenteroides was constructed (CEN.PK2_DLDH). To increase D-LA production, the CRISPR/Cas12a system was used for the deletion of gpd1, gpd2, and adh1 to minimize glycerol and ethanol production. Although an improved D-LA titer was observed for both CEN.PK2_DLDHΔgpd and CEN.PK2_DLDHΔgpdΔadh1, growth impairment was observed. To enhance the D-LA productivity, CEN.PK2_DLDHΔgpd was crossed with the weak acid-tolerant S. cerevisiae BCC39850. The isolated hybrid2 showed a maximum D-LA concentration of 23.41 ± 1.65 g/L, equivalent to the improvement in productivity and yield by 2.2 and 1.5 folds, respectively. The simultaneous saccharification and fermentation using alkaline pretreated sugarcane bagasse by the hybrid2 led to an improved D-LA conversion yield on both the washed solid and whole slurry (0.33 and 0.24 g/g glucan). Our findings show the exploitation of natural yeast diversity and the potential strategy of gene editing combined with conventional breeding on improving the performance of S. cerevisiae for the production of industrially potent products.


Introduction
Environmentally sustainable platforms for the production of commodity and specialty chemicals are needed to replace current fossil-based platforms. Carbon-neutral renewable resources, including lignocellulosic waste streams, are seen as particularly promising for the environmentally sustainable production of bio-based chemicals. Lactic acid is a biobased chemical used in the food, chemical, and healthcare industries. Lactic acid is used to make polylactic acid (PLA), a biodegradable polymer that is widely used in packaging material, mulch film, garbage bags, and medical materials [1,2]. PLA is typically produced from racemic mixtures, although it can also be produced from pure enantiomers to make acids and other inhibitors and can withstand harsh fermentation conditions, which make it a good candidate for D-LA production from cellulosic hydrolysates. However, D-LA production from cellulosic biomass using wild-type and engineered S. cerevisiae has not been studied extensively so far [24].
In this study, we combined rational engineering and yeast mating to create an intraspecific hybrid S. cerevisiae strain that is able to produce D-LA efficiently from sugarcane bagasse hydrolysates. Combining these strategies provided an effective way to confer two distinct beneficial traits: weak acid tolerance (from the natural strain) and D-LA production (from the engineered strain). This could not have been achieved using the individual techniques alone.

Plasmid Construction
The TDH3 promoter and CYC1 terminator were PCR-amplified from genomic DNA of S. cerevisiae (strain CEN.PK2-1C) using primer pairs DeltaUp_KpnI_TDH3pro_F and TDH3Pro_R, and CYC1ter_F and CYC1ter_R, respectively. The S. cerevisiae codon-optimized Lm.ldhA gene (GenBank accession number MW574957) was purchased from Genscript as a plasmid (pUC57-Lm.ldhA). Lm.ldhA and HIS3 selectable markers were PCR-amplified from pUC57-Lm.ldhA and pRSII405 by using the primer pairs LmLDH_F and LmLDH_R, and CYC1_BamHI_HIS3pro_F and DeltaDown_EcoRI_HIS3ter_R, respectively. These four DNA fragments were assembled into a single fragment by overlap-extension PCR (OE-PCR). The resulting cassette was ligated to pJET1.2 vector (Thermo Fisher Scientific, Waltham, MA, USA) to yield pJET-DeltaUp-LmLDH-His3-DeltaDown. To construct pRPR1-gRNA-delta for genome integration, the gRNA-delta fragment was PCR-amplified from pRPR1-gRNA handle-RPR1t using primers delta_gRNA1_HindIII_F and gRNA_Rev. The 0.13-kb PCR amplicon was gel-purified and ligated via the HindIII/XhoI sites of pRPR1-gRNA handle-RPR1t to yield pRPR1-gRNA-delta. To construct pArray plasmids for specific gene deletion, the gpd1, 2, and adh1 specific crRNA for each gene deletion was designed using the CRISPR RGEN tools [29], shown in Table S2. The crRNA array, which comprises gene-specific crRNA for each target gene as well as direct repeats (5 -AATTTCTACTGTTGTAGAT-3 ) and flanking sequences homologous to plasmid backbone pUDE735, was synthesized by GenScript and amplified by using primers crRNA-F and crRNA-R. The plasmid backbone pUDE735 was prepared by PCR amplification of pUDE735 using the primers tSUP4-F and pCAS9-R. Homologous recombination of the synthesized fragment with the plasmid backbone pUDE735 resulted in pArray-GPD1,2 plasmid for gpd1 and gpd2 gene deletion and pArray-ADH1 plasmid for adh1 gene deletion.

CRISPR-Cpf1-Mediated Gene Deletion
The markerless CRISPR/Cas12a system was employed to generate deletion strains. Donor DNA for each gene deletion was obtained by PCR using complementary primers. The donor DNA was gel purified and then transformed into competent CEN.PK2-1C_DLDH cells along with the corresponding pArray plasmid and pUDC175, which contains the Cpf1 (Cas12a) gene from Francisella tularensis under the control of the TEF1 promoter. Transformants were selected on SC medium with L-histidine/L-tryptophan dropout containing 200 µg/mL G418. Colony PCR using colony PCR primers were performed to verify the gene deletions.

Yeast Mating
The wild-type S. cerevisiae BCC39850 (MATalpha; weak acid-tolerant) and CEN.PK2-1C_DLDH∆gpd (MATa) strains were crossed in YPD agar plate and grown overnight at 30 • C. After mating, a loop full of the mated population was suspended in SC broth and streaked on SC agar plate (without amino acid), and then incubated at 30 • C for 3-5 days until diploid colonies appeared. Diploid colonies were taken and re-streaked on sporulation agar plates (10 g/L potassium acetate, 1 g/L yeast extract, 0.5 g/L dextrose, 20 g/L agar) and incubated at 30 • C for 3-5 days to confirm tetrad formation under microscope.

Batch Fermentation, Growth Curve, and Lactic Acid Production under Non-Neutralized Conditions
D-LA fermentation was carried out in 250 mL Erlenmeyer flasks containing 50 mL of YPD containing 100 g/L glucose with S. cerevisiae cells suspended at an initial density OD600 of 0.05. The synthetic inhibitor cocktail was formulated to mimic the composition of hydrolysate toxins present in sugarcane bagasse from the alkaline pretreatment process as modified from van der Pol and coworkers [30]. To evaluate the production of D-LA in the presence of synthetic hydrolysate toxins, an equal volume of synthetic hydrolysate toxins was added to 50 mL of YPD (100 g/L glucose), resulting in a medium containing 0.35 g/L formic acid, 1.65 g/L acetic acid, 0.025 g/L levulinic acid, 0.047 g/L vanillin, and 0.018 g/L syringaldehyde with the same initial OD 600 of S. cerevisiae cells. The cultures were incubated at 30 • C with rotary shaking at 200 rpm. For time-course experiments, samples were taken every 8 h for D-LA production analysis.

Alkaline Pretreatment of Sugarcane Bagasse and Simultaneous Saccharification and Fermentation
Sixty grams of sugarcane bagasse was pretreated by adding 180 mL of 50 g/L NaOH and heated at 90 • C for 90 min under air atmosphere. The whole slurry of the pretreated sugarcane bagasse was prepared by adding 400 mL of water, and pH was adjusted to neutral with sulfuric acid. The total volume was made up to 1 L to achieve the final concentration of 6% loading of whole slurry. The washed solid was also obtained from the same pretreatment condition, in which the solid fraction was separated and washed with tap water until neutral pH was reached and the solid sample was then dried at 50 • C. Chemical compositions of the pretreated biomass samples were analyzed according to National Renewable Energy Laboratory (NREL) analytical procedures [31]. The composition of washed solid was reported as 55.1% cellulose, 18.1% hemicellulose, and 8.7% lignin. To evaluate the D-LA production efficiency of S. cerevisiae strains, the washed solid and whole slurry of alkaline-pretreated sugarcane bagasse (6% solid loading) were used as the carbon substrates for simultaneous saccharification and fermentation (SSF). YP medium (10 g/L yeast extract and 20 g/L peptone) containing sugarcane bagasse solid fraction or whole slurry supplemented with 30 filter paper units (FPU)/g biomass commercial cellulase (Cellic Ctec2; Novozyme Inc., Franklinton, NC, USA) was inoculated with S. cerevisiae cells at OD 600 of 0.5 and incubated at 30 • C for 48 h with shaking at 200 rpm. The amounts of sugar degradation byproducts, including carboxylic acids, furfural, and hydroxymethyl furfural present in the whole slurry are shown in Table S3.

Product Analysis
Samples from both batch fermentation and simultaneous saccharification and fermentation were harvested by centrifugation (10,000× g for 5 min at 4 • C) and filtered through 0.2-micron filter cellulose acetate membranes (Millipore; Milford, MA, USA). The yields of sugars, lactic acid, ethanol, and glycerol were determined using a high-performance liquid chromatographic (HPLC) system (Shimadzu Prominence LC-20 equipped with a refractive index detector (Shimadzu Corporation, Kyoto, Japan) and an Aminex-HPX-87H Column (Bio-Rad, Hercules, CA, USA)). The column temperature was maintained at 65 • C and 5 mM H 2 SO 4 was used as the mobile phase at a flow rate of 0.5 mL/min.

Engineering S. cerevisiae to Produce D-Lactic Acid
To engineer the yeast S. cerevisiae to produce high titers of D-LA, we first introduced the D-lactate dehydrogenase (D-LDH) gene from Leuconostoc mesenteroides, a natural D-LA producer, into the laboratory strain CEN.PK2-1C. To increase the number of chromosomally integrated D-LDH expression cassettes, we employed a CRISPR-based strategy developed by Shi and coworkers [32], whereby the Lm.ldhA gene construct of interest is integrated into multiple Ty retrotransposon delta sites spread throughout the yeast genome. The resulting strain, named CEN.PK2_DLDH, was able to produce D-LA at a titer of 2.3 ± 0.2 g/L after three days of fermentation in 50 mL Falcon tubes ( Figure 1). In addition to the desired product, D-LA, the strain also produced ethanol as a co-product (1.7 ± 0.2 g/L) and a small amount of glycerol (0.042 ± 0.01 g/L). To increase the production level of D-LA and minimize the production of the side-products ethanol and glycerol, we targeted gpd1, gpd2, and adh1 for deletion by employing the CRISPR/Cas12a (previously called Cpf1) system. We observed the highest D-LA titer in the triple deletion strain (CEN.PK2_DLDH∆gpd∆adh1), which produced D-LA at a titer of 14.4 ± 1.8 g/L, a 6.3-fold titer improvement over the level observed in strain CEN.PK2_DLDH. Similarly, we saw improvement in the double deletion strain (CEN.PK2_DLDH∆gpd), which produced D-LA at a titer of 9.6 g/L ± 0.4 g/L. The triple deletion strain also produced less ethanol (0.5 ± 0.1 g/L) and an undetectable level of glycerol. However, the improvements in the D-LA titer and yield in the triple deletion strain were accompanied by severe growth retardation ( Figure 1D) and a reduction in the maximum specific growth rate (Table S4). Given the unfavorable growth phenotype of the triple deletion strain, we chose the double deletion strain CEN.PK2_DLDH∆gpd for further improvement by yeast breeding.

Engineering S. cerevisiae to Produce D-Lactic Acid
To engineer the yeast S. cerevisiae to produce high titers of D-LA, we first introduced the D-lactate dehydrogenase (D-LDH) gene from Leuconostoc mesenteroides, a natural D-LA producer, into the laboratory strain CEN.PK2-1C. To increase the number of chromosomally integrated D-LDH expression cassettes, we employed a CRISPR-based strategy developed by Shi and coworkers [32], whereby the Lm.ldhA gene construct of interest is integrated into multiple Ty retrotransposon delta sites spread throughout the yeast genome. The resulting strain, named CEN.PK2_DLDH, was able to produce D-LA at a titer of 2.3 ± 0.2 g/L after three days of fermentation in 50 mL Falcon tubes ( Figure 1). In addition to the desired product, D-LA, the strain also produced ethanol as a co-product (1.7 ± 0.2 g/L) and a small amount of glycerol (0.042 ± 0.01 g/L). To increase the production level of D-LA and minimize the production of the side-products ethanol and glycerol, we targeted gpd1, gpd2, and adh1 for deletion by employing the CRISPR/Cas12a (previously called Cpf1) system. We observed the highest D-LA titer in the triple deletion strain (CEN.PK2_DLDHΔgpdΔadh1), which produced D-LA at a titer of 14.4 ± 1.8 g/L, a 6.3-fold titer improvement over the level observed in strain CEN.PK2_DLDH. Similarly, we saw improvement in the double deletion strain (CEN.PK2_DLDHΔgpd), which produced D-LA at a titer of 9.6 g/L ± 0.4 g/L. The triple deletion strain also produced less ethanol (0.5 ± 0.1 g/L) and an undetectable level of glycerol. However, the improvements in the D-LA titer and yield in the triple deletion strain were accompanied by severe growth retardation ( Figure 1D) and a reduction in the maximum specific growth rate (Table S4). Given the unfavorable growth phenotype of the triple deletion strain, we chose the double deletion strain CEN.PK2_DLDHΔgpd for further improvement by yeast breeding.

Improving D-LA Production of CEN.PK2_DLDHΔgpd by Conventional Yeast Mating
The double deletion knockout strain expressing D-LDH (CEN.PK2_DLDHΔgpd) was crossed with the weak acid-tolerant strain S. cerevisiae BCC39850 ( Figure S1). Over two-hundred diploid progenies were screened for D-LA production before choosing the representative hybrid strains for further analysis. The presence of an integrated Lm.ldhA gene in the three selected putative progeny strains (hybrid2, hybrid35, and hybrid36) and the parental strain CEN.PK2_DLDHΔgpd was demonstrated by PCR ( Figure 2A). The status of the gpd genes was assessed by PCR using the gpd1 and gpd2 primers, which revealed the presence of both the intact and deleted copies of gpd1 and gpd2 in hybrid2 ( Figure 2B). Hybrid 35

Improving D-LA Production of CEN.PK2_DLDH∆gpd by Conventional Yeast Mating
The double deletion knockout strain expressing D-LDH (CEN.PK2_DLDH∆gpd) was crossed with the weak acid-tolerant strain S. cerevisiae BCC39850 ( Figure S1). Over twohundred diploid progenies were screened for D-LA production before choosing the representative hybrid strains for further analysis. The presence of an integrated Lm.ldhA gene in the three selected putative progeny strains (hybrid2, hybrid35, and hybrid36) and the parental strain CEN.PK2_DLDH∆gpd was demonstrated by PCR ( Figure 2A). The status of the gpd genes was assessed by PCR using the gpd1 and gpd2 primers, which revealed the presence of both the intact and deleted copies of gpd1 and gpd2 in hybrid2 ( Figure 2B). Hybrid 35 contained intact gpd1 and deleted gpd2, whereas intact gpd1 and gpd2 genes were found in hybrid36 ( Figure 2B). The selected hybrid strains were also evaluated for lactic acid tolerance by spotting on YPD plates containing 0, 30, 40, 50, and 60 g/L of DL-lactic acid (DL-LA). Growth was observed at day 1 and day 3 after incubating at 30 • C ( Figure 3). The growth of the double knockout CEN.PK2_DLDH∆gpd parental strain was markedly slower than hybrid2 and hybrid35 on unsupplemented YPD. In the presence of 40-50 g/L of DL-LA, no growth was observed for CEN.PK2_DLDH∆gpd, whereas the other strains still showed a varying degree of growth. Based on the growth phenotype, hybrid2 was considered to be the most tolerant strain to DL-LA as it was the only one able to overcome a DL-LA concentration of 50 g/L ( Figure 3). The Lm.ldhA copy number among the strains was also quantified by qPCR ( Figure S2). The Lm.ldhA copy number of hybrid2 was only 0.93 while hybrid35 and hybrid36 contained about 2.98, which could result from the different patterns of meiotic segregation. Next, the D-LA productivity in each strain was assessed under aerobic fermentation conditions in YPD broth with or without synthetic hydrolysate toxins.
contained intact gpd1 and deleted gpd2, whereas intact gpd1 and gpd2 genes were found in hybrid36 ( Figure 2B). The selected hybrid strains were also evaluated for lactic acid tolerance by spotting on YPD plates containing 0, 30, 40, 50, and 60 g/L of DL-lactic acid (DL-LA). Growth was observed at day 1 and day 3 after incubating at 30 °C (Figure 3). The growth of the double knockout CEN.PK2_DLDHΔgpd parental strain was markedly slower than hybrid2 and hybrid35 on unsupplemented YPD. In the presence of 40-50 g/L of DL-LA, no growth was observed for CEN.PK2_DLDHΔgpd, whereas the other strains still showed a varying degree of growth. Based on the growth phenotype, hybrid2 was considered to be the most tolerant strain to DL-LA as it was the only one able to overcome a DL-LA concentration of 50 g/L (Figure 3). The Lm.ldhA copy number among the strains was also quantified by qPCR ( Figure S2). The Lm.ldhA copy number of hybrid2 was only 0.93 while hybrid35 and hybrid36 contained about 2.98, which could result from the different patterns of meiotic segregation. Next, the D-LA productivity in each strain was assessed under aerobic fermentation conditions in YPD broth with or without synthetic hydrolysate toxins.   contained intact gpd1 and deleted gpd2, whereas intact gpd1 and gpd2 genes were found in hybrid36 ( Figure 2B). The selected hybrid strains were also evaluated for lactic acid tolerance by spotting on YPD plates containing 0, 30, 40, 50, and 60 g/L of DL-lactic acid (DL-LA). Growth was observed at day 1 and day 3 after incubating at 30 °C (Figure 3). The growth of the double knockout CEN.PK2_DLDHΔgpd parental strain was markedly slower than hybrid2 and hybrid35 on unsupplemented YPD. In the presence of 40-50 g/L of DL-LA, no growth was observed for CEN.PK2_DLDHΔgpd, whereas the other strains still showed a varying degree of growth. Based on the growth phenotype, hybrid2 was considered to be the most tolerant strain to DL-LA as it was the only one able to overcome a DL-LA concentration of 50 g/L (Figure 3). The Lm.ldhA copy number among the strains was also quantified by qPCR ( Figure S2). The Lm.ldhA copy number of hybrid2 was only 0.93 while hybrid35 and hybrid36 contained about 2.98, which could result from the different patterns of meiotic segregation. Next, the D-LA productivity in each strain was assessed under aerobic fermentation conditions in YPD broth with or without synthetic hydrolysate toxins.

Batch Fermentation for Lactic Acid Production under Non-Neutralized Conditions
The fermentative abilities of CEN.PK2_DLDH∆gpd, hybrid2, hybrid35, and hybrid36 were analyzed in the YPD medium containing 100 g/L glucose with or without synthetic toxins. The glucose consumption and D-LA and ethanol production (as byproduct) were monitored at 8 h intervals of cultivation time. In both fermentation conditions, no neutralizing agents were added. Figure 4 shows the glucose consumption and the D-LA and ethanol production. In the absence of synthetic hydrolysate toxins, hybrid2 consumed glucose the fastest, in which 50% of the glucose was consumed after 8 h. The glucose consumption was lower for hybrid35 and hybrid36, and CEN.PK2_DLDH∆gpd demonstrated the slowest consumption ( Figure 4C). The maximum production of D-LA and ethanol was measured ( Figure 4B). The highest D-LA production was obtained with hybrid36. The CEN.PK2_DLDH∆gpd strain produced the least D-LA but still produced almost as much ethanol as hybrid35. In addition to D-LA and ethanol, some glycerol was produced in fermentations with hybrid2, hybrid35, and hybrid36 (Table S5). No glycerol product was detected in the fermentations with CEN.PK2_DLDH∆gpd.

Batch Fermentation for Lactic Acid Production under Non-Neutralized Conditions
The fermentative abilities of CEN.PK2_DLDHΔgpd, hybrid2, hybrid35, and hybrid36 were analyzed in the YPD medium containing 100 g/L glucose with or without synthetic toxins. The glucose consumption and D-LA and ethanol production (as byproduct) were monitored at 8 h intervals of cultivation time. In both fermentation conditions, no neutralizing agents were added. Figure 4 shows the glucose consumption and the D-LA and ethanol production. In the absence of synthetic hydrolysate toxins, hybrid2 consumed glucose the fastest, in which 50% of the glucose was consumed after 8 h. The glucose consumption was lower for hybrid35 and hybrid36, and CEN.PK2_DLDHΔgpd demonstrated the slowest consumption ( Figure 4C). The maximum production of D-LA and ethanol was measured ( Figure 4B). The highest D-LA production was obtained with hybrid36. The CEN.PK2_DLDHΔgpd strain produced the least D-LA but still produced almost as much ethanol as hybrid35. In addition to D-LA and ethanol, some glycerol was produced in fermentations with hybrid2, hybrid35, and hybrid36 (Table S5). No glycerol product was detected in the fermentations with CEN.PK2_DLDHΔgpd. The effects of synthetic hydrolysate toxins on microbial growth, glucose consumption, and the production of D-LA and ethanol were also investigated in aerobic batch fermentation. Overall, the addition of the synthetic hydrolysate toxins leads to an initial delay in the glucose uptake ( Figure 4F) followed by reduced consumption. However, there are differences as hybrid35, hybrid36, and CEN.PK2_DLDHΔgpd exhibited longer lag times than hybrid2 ( Figure 4F).
Theoretically, one gram of D-LA can be produced from one gram of glucose catalyzed by D-lactate dehydrogenase (DLDH). Overall, the values of the D-LA productivity and yield of hybrid2, hybrid35, and hybrid36 were markedly greater than that of the parental strain CEN.PK2_DLDHΔgpd (Table 2). When the synthetic hydrolysate cocktail was added, all the strains exhibited the reduced glucose uptake and a reduction in the growth The effects of synthetic hydrolysate toxins on microbial growth, glucose consumption, and the production of D-LA and ethanol were also investigated in aerobic batch fermentation. Overall, the addition of the synthetic hydrolysate toxins leads to an initial delay in the glucose uptake ( Figure 4F) followed by reduced consumption. However, there are differences as hybrid35, hybrid36, and CEN.PK2_DLDH∆gpd exhibited longer lag times than hybrid2 ( Figure 4F).
Theoretically, one gram of D-LA can be produced from one gram of glucose catalyzed by D-lactate dehydrogenase (DLDH). Overall, the values of the D-LA productivity and yield of hybrid2, hybrid35, and hybrid36 were markedly greater than that of the parental strain CEN.PK2_DLDH∆gpd (Table 2). When the synthetic hydrolysate cocktail was added, all the strains exhibited the reduced glucose uptake and a reduction in the growth rate and D-LA productivity. The highest D-LA productivity was observed for hybrid2 under both conditions. Similarly, it was found that synthetic hydrolysate toxins also affected the productivity of ethanol in the same way as D-LA productivity. A marked reduction in ethanol productivity was observed in both hybrid35 and hybrid36, whereas there was little effect on ethanol productivity in CEN.PK2_DLDH∆gpd and hybrid2. Table 2. D-lactic acid (D-LA) productivity and yield in batch fermentation using glucose as a sole carbon source.

Productivity (g/L/h) Yield (g/g)
D-LA Ethanol D-LA Ethanol

D-Lactic Acid Production Using Alkaline-Pretreated Sugarcane Bagasse in SSF
The D-LA production from alkaline-pretreated bagasse was investigated for the strain with the highest productivity (hybrid2) and the parental strain CEN.PK2_DLDH∆gpd as the control ( Figure 5). The SSF of the washed solid fraction ( Figure 5A) showed that the maximum D-LA production was higher for hybrid2. In contrast, CEN.PK2_DLDH∆gpd produced more ethanol than hybrid2. No glucose was detected during the SSF, indicating efficient sugar utilization by both strains during the SSF of the washed solid. The glucose consumption of CEN.PK2_DLDH∆gpd during the SSF of the whole slurry was quite limited which directly affected both D-LA and ethanol production ( Figure 5B). Hybrid2 was able to consume glucose more efficiently and produced more D-LA than CEN.PK2_DLDH∆gpd during the SSF of the whole slurry. Alkaline pretreatment at high temperature can also release a large amount of acetic acid and phenolic compounds into the liquid fraction of the whole slurry [30]. This could inhibit either microbial growth or activities of enzymes that are used for saccharification [33]. The D-LA productivity was greater from the washed solid compared with the whole slurry (Table 3). Because there were no inhibitor byproducts and fermentable sugars detected in the sugarcane bagasse washed solid fraction, almost all of the washed solid fraction was gradually converted to glucose (total 32.7 ± 0.43 g/L) and xylose (12.21 ± 0.16 g/L). The analysis of the degradation byproducts in the whole slurry showed that the most abundant inhibitor was acetic acid (3.06 ± 0.15 g/L), whereas formic acid, levulinic acid, furfural, and hydroxymethy furfural were not found (Table S3). Table 3. D-lactic acid (D-LA) productivity and yield in SSF of washed solid and whole slurry of alkaline-pretreated sugarcane bagasse.

Discussion
In this study, a transgenic strain of S. cerevisiae was established that expresses L. mesenteroides D-LDH genes integrated into the genome. D-LA was produced from this strain as expected, although the co-production of ethanol and glycerol limited the D-LA productivity. The disruption of the gpd1 and gpd2 genes in this strain would eliminate glycerol-3-phosphate dehydrogenase activity and thus the production of glycerol. In S. cerevisiae, Gpd1 and Gpd2 are the rate-controlling enzymes for glycerol synthesis. Both isoforms play important roles in osmotic adaptation, the response to oxidative stress, heat shock protection, and redox balance [34]. Glycerol also participates in the biosynthesis of glycerophospholipids and triacylglycerols by conversion to the intermediate G3P and plays a role in maintaining cytosolic redox balance, all of which are important for S. cerevisiae to retain normal physiological functions and growth [35]. The deletion of the gpd1 and gpd2 genes in S. cerevisiae might be beneficial for the production of high-value chemicals because glycerol is considered to be a major byproduct, accounting for 5% of carbon during cell growth. However, the deletion of both gpd genes had a negative impact on growth ( Figure 1D). Nissen and coworkers [36] reported a similar finding, in which a double deletion gpd mutant showed a 29% reduction in biomass synthesis with a 12.7% increased ethanol yield, whereas single deletion mutants (gpd1Δ and gpd2Δ) showed very little increased ethanol yield under aerobic conditions. Furthermore, the volumetric productivity of ethanol by S. cerevisiae can be increased by fine-tuning GPD expression during fedbatch fermentation [35].

Discussion
In this study, a transgenic strain of S. cerevisiae was established that expresses L. mesenteroides D-LDH genes integrated into the genome. D-LA was produced from this strain as expected, although the co-production of ethanol and glycerol limited the D-LA productivity. The disruption of the gpd1 and gpd2 genes in this strain would eliminate glycerol-3-phosphate dehydrogenase activity and thus the production of glycerol. In S. cerevisiae, Gpd1 and Gpd2 are the rate-controlling enzymes for glycerol synthesis. Both isoforms play important roles in osmotic adaptation, the response to oxidative stress, heat shock protection, and redox balance [34]. Glycerol also participates in the biosynthesis of glycerophospholipids and triacylglycerols by conversion to the intermediate G3P and plays a role in maintaining cytosolic redox balance, all of which are important for S. cerevisiae to retain normal physiological functions and growth [35]. The deletion of the gpd1 and gpd2 genes in S. cerevisiae might be beneficial for the production of high-value chemicals because glycerol is considered to be a major byproduct, accounting for 5% of carbon during cell growth. However, the deletion of both gpd genes had a negative impact on growth ( Figure 1D). Nissen and coworkers [36] reported a similar finding, in which a double deletion gpd mutant showed a 29% reduction in biomass synthesis with a 12.7% increased ethanol yield, whereas single deletion mutants (gpd1∆ and gpd2∆) showed very little increased ethanol yield under aerobic conditions. Furthermore, the volumetric productivity of ethanol by S. cerevisiae can be increased by fine-tuning GPD expression during fed-batch fermentation [35].
We hypothesized that D-LA production and other fermentation characteristics could be improved by mating the CEN.PK2_DLDH∆gpd strain with a wild-type strain. Three progeny strains were isolated. Hybrid35 had a deletion of gdp2 whereas hybrid36 appeared to be wild-type with respect to both gpd genes. Hybrid2 appeared to be the diploid hybrid, as shown by the presence of wild-type and deleted copies of both gdp genes ( Figure 2B). The superior growth and D-LA production characteristics of hybrid2 could be attributed to heterosis, although it is not known which genetic factor(s) from the BCC39850 parent contribute to the superior characteristics of hybrid2 compared with those of the CEN.PK2_DLDH∆gpd parent. The extended lag phase corresponded to a detoxification phase in which S. cerevisiae adapts and responds to the inhibitors [37]. These affected the ethanol and D-LA productivity especially in hybrid35 and hybrid36. Because the genetic backgrounds of CEN.PK2-1C and BCC39850 are different, the hybrid strains constructed in this study responded to inhibitors in different manners. The physiological and metabolic changes, such as the composition of plasma membrane rearrangements, the ability to maintain intracellular pH homeostasis, the detoxification of reactive oxygen species, and the assimilation of inhibitors, are reported to the improved yeast strains [37,38].
Compared with previous reports on LA production in engineered S. cerevisiae (Table 4), the D-LA productivity of hybrid2 was higher except for the strain reported in [39]. Moreover, the D-LA yield of hybrid2 is in the same range as that reported for other S. cerevisiae strains with disruption of ADH and PDC genes [6,19,39,40] in which the metabolic flux was shifted toward LA production. The approach of crossing genetically engineered strains expressing heterologous DLDH with wild-type could be applied for the further improvement of D-LA production yield, for example, in engineered strains with deletions of other genes, including ADH, PDC, and D-Lactate dehydrogenase gene (dld1). The D-LA production from a cellulosic biomass was also demonstrated using the CEN.PK2_DLDH ∆gpd and hybrid2 strains. Alkaline-pretreated sugarcane bagasse was used as a representative feedstock for D-LA production by the SSF to avoid substrate inhibition that could potentially inhibit microbial fermentation and enzymatic saccharification. The major byproduct found in the whole slurry was acetic acid. To cope with acid stress, S. cerevisiae responds to the changes of the intracellular pH from the D-LA and the external acetate/H+ in ways that utilize a large amount of ATP, which leaves less ATP available for cell growth [41]. As a result, the D-LA productivity by hybrid2 from the whole slurry was lower than that obtained from the washed solid fraction.
To our knowledge, there are no other reports of D-LA production from sugarcane bagasse using engineered S. cerevisiae. The D-LA productivity obtained from hybrid2 was comparable to that reported for the fermentation of corn stover hydrolysate using lactic acid bacteria (LABs), which ranged from 0.32 to 1.02 g/L/h [1]. In terms of industrial applicability, D-LA production in S. cerevisiae has advantages over LABs, including the ability to cope with environmental stresses, low pH, and no requirement of complex nutrition for growth [10]. With the advantages of sustainable, environmental, and socio-economic aspects as well as cost-competitiveness, the use of lignocellulosic feedstocks for D-LA production needs to be studied more intensively. Here, we have shown that it is feasible to produce D-LA from sugarcane bagasse, an economically important lignocellulosic feedstock, using engineered yeast.

Conclusions
In this study, S. cerevisiae expressing heterologous LmD-LDH (CEN.PK2_DLDH) was constructed using a CRISPR-Cas9-based strategy. To minimize ethanol and glycerol sideproducts, and thus consequently increase the production of D-LA, we deleted the gpd1, gpd2, and adh1 genes by employing the CRISPR/Cas12a system. This strategy was successful, as shown by the reduction in side-products and increased D-LA production for both the double deletion CEN.PK2_DLDH∆gpd and triple deletion CEN.PK2_DLDH∆gpd∆adh1 strains. However, the improvements in the D-LA titer and yield were accompanied by growth retardation, which was pronounced for the triple deletion strain. Therefore, the double deletion CEN.PK2_DLDH∆gpd strain was crossed with the weak acid-tolerant S. cerevisiae BCC39850 wild-type strain for strain improvement. The D-LA productivity and yield of the isolated hybrid2 were significantly greater under the fermentation conditions containing formulated hydrolysate inhibitors compared with CEN.PK2_DLDH∆gpd. Moreover, hybrid2 also showed superior D-LA production in simultaneous saccharification and fermentation using alkaline-pretreated sugarcane bagasse on washed solid and whole slurry. The work demonstrates the use of combined conventional breeding and gene editing for developing a yeast cell factory with improved performance for D-LA production from cellulosic substrates with industrial potential.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jof8080816/s1, Figure S1: Acetic acid, formic acid, and levulinic acid tolerant test of wild Saccharomyces cerevisiae strains; Figure S2: Quantification of LmldhA copy number among the strains by qPCR; Table S1: Primers used in this study; Table S2: Designing of genespecific crRNA for gene deletion and homologous flanking sequences for homologous recombination with pUDE735; Table S3: Carboxylic acids, furfural, and hydroxymethyl furfural detected in slurry of alkaline-pretreated sugarcane bagasse; Table S4: Maximum specific growth rate of all Saccharomyces cerevisiae strains used in this study; Table S5: Glycerol production by Saccharomyces cerevisiae strains used in this study.