Factors Affecting D-Lactic Acid Production by Flocculant Saccharomyces cerevisiae Under Non-Neutralizing Conditions

Abstract

Integrating heterogeneous genes is widely used in metabolic engineering to produce D-lactic acid (D-LA), an essential compound in bioplastics and pharmaceuticals. However, research on the effects of integrating various loci on gene expression, especially regarding flocculation behavior, remains limited. This study constructed Saccharomyces cerevisiae strains by incorporating a codon-optimized D-LDH gene from Leuconostoc pseudomesenteroides (LpDLDH) into the specific genomic loci of the CYB2, PDC1, MPC1, PDC6, ADH1, and PDC5 genes to redirect pyruvate toward lactic acid. Strains with the LpDLDH gene integrated at the PDC1 locus achieved the highest D-LA titers (51 g/L) with minimal ethanol byproduct, followed by strains with integrations into the CYB2 locus at 31.92 g/L, the MPC1 locus at 10 g/L, and the PDC6 locus at 0.026 g/L. In contrast, strains with LpDLDH integrated at the ADH1 and PDC5 loci failed to produce detectable levels of D-LA and exhibited a complete loss of flocculation. Gene expression analysis revealed a significant expression of genes related to flocculation (FLO5), stress adaptation (HSP150), and cell wall integrity (YGP1, SED1, and SCW11). The CYB2-integrating strain showed strong flocculant properties, contributing to its robustness. These findings highlight the influence of genomic locus selection on metabolic flux and stress adaptation, offering insights into optimizing D-LA production in flocculant S. cerevisiae yeast.

1. Introduction

The demand for lactic acid has increased significantly owing to its extensive application, particularly as a monomer for producing biodegradable and biocompatible polylactic acid materials. Consequently, enhancing fermentation processes to produce optically pure lactic acid has garnered substantial interest in recent research [1]. Lactic acid (LA) also serves as a key precursor in the production of various chemicals, including acrylic acid, acetaldehyde, pyruvic acid, 2,3-pentanedione, 1,2-propanediol, LA esters, and polylactic acid. Polymer-grade LA is essential for biomedical research and applications and serves as a fundamental raw material for the manufacture of environmentally friendly polylactic acid plastics. The diverse applications of LA underscore its importance as a highly valued product on the global market [2,3]. Blending poly D-LA with poly l-lactic acid through stereo-complexation results in plastics with enhanced thermal stability [4,5]. Large-scale L-lactic acid production by major chemical companies utilizing microbes has been well established [6]. However, limited research on producing D-lactic acid using microbes necessitates a more efficient method.

D-lactic acids are efficiently synthesized in substantial quantities by lactic acid bacteria, including Lactobacillus, Sporolactobacillus, and Leuconostoc [7]. For instance, Sporolactobacillus inulinus YBS1-5 produced 107.2 g/L of D-LA through fed-batch fermentation utilizing corncob residue hydrolysate, yielding 0.85 g/g of glucose equivalent [8]. Additionally, Leuconostoc mesenteroides B512 produced 60.2 g/L of D-LA through shake-flask fermentation using sugarcane juice, yielding 0.51 of glucose [9]. Typically, fermentation is effective at approximately neutral pH, is regulated using neutralizing agents, and produces lactate salts. This presents a challenge in the production process. However, the desired product for the polymerization reaction is undissociated (free) lactic acid, not its salt, necessitating further processing to recover the free lactic acid [4]. To address this issue and streamline production, fermentation without neutralizing agents and with a high yield of D-LA is required to simplify the process. However, this approach comes with its own set of challenges. While these bacteria are efficient producers of D-lactic acid, the growth of lactic acid bacteria is inhibited by a low pH due to high concentrations of lactic acid, resulting in the loss of viability and cellular activity [10], limiting their use in non-neutralizing fermentation conditions.

Saccharomyces cerevisiae exhibits a strong tolerance for acid and is suitable for the industrial-scale production of D-LA under non-neutralizing conditions [11]. Although S. cerevisiae does not naturally produce D-LA, optically pure D-LA can be synthesized through the heterologous expression of stereospecific D-lactate dehydrogenase (D-LDH). To achieve high-yield D-LA production in S. cerevisiae, the ethanol fermentation pathway is typically inactivated by disrupting the genes related to the activities of pyruvate decarboxylase (PDC1, PDC5, and PDC6) and alcohol dehydrogenase (ADH1), which control the initial steps of ethanol synthesis. Furthermore, to improve D-LA titers, it is essential to minimize D-LA consumption by disrupting the gene for mitochondrial inner membrane lactate dehydrogenase (DLD1), which oxidizes D-LA to pyruvate, along with the gene for mitochondrial intermembrane space cytochrome b2 (CYB2), which is vital for lactate utilization [11].

In addition to optimizing D-LDH integration loci, it is crucial to elucidate the physiological and molecular changes associated with D-LA production. No study has examined the relationship between D-LA production and flocculation traits in engineered strains with different locus integrations. The flocculation trait, a key characteristic of S. cerevisiae in industrial applications, is influenced by the intrinsic complexity of yeast flocculation regulation, which is governed by yeast genetics and environmental conditions [12]. The production of D-LA involves significant metabolic redirection and acid stress, which can affect cell wall integrity, stress response mechanisms, and flocculation phenotypes [13]. Investigating these changes through gene expression analysis can reveal the interplay between cell wall components, stress resistance pathways, and fermentation performance.

This research aims to expand existing knowledge by evaluating the flocculant S. cerevisiae yeast suitable for D-LA production and investigating the effects of genome integration loci on the D-LDH transgene expression cassette and D-LA production in this flocculant strain. It emphasizes how these loci influence traits related to flocculation and gene expression associated with cell wall integrity and stress response. Identifying optimal integration sites and elucidating the associated gene expression changes will facilitate more efficient microbial production of D-LA by the flocculant S. cerevisiae yeast strain and contribute to developing sustainable methods for producing biodegradable materials.

2. Materials and Methods

2.1. Strains and Medium

In this study, S. cerevisiae F118 (NBRC268), a flocculant strain obtained from the National Biological Resource Center (NBRC) culture collection in Tokyo, Japan, served as the primary host yeast for recombinant construction and fermentation. The Escherichia coli NEB stable (high efficiency) strain (New England Biolabs, NEB, Ipswich, MA, USA) was utilized for routine cloning. Specifically, E. coli was cultivated in Luria–Bertani (LB) medium, which contained 10 g/L of tryptone, 5 g/L of yeast extract, and 10 g/L of sodium chloride, supplemented with 100 µg/mL of ampicillin at 37 °C while shaking at 150 rpm. The yeast was grown in yeast extract peptone dextrose medium (YPD), consisting of 20 g/L of peptone, 10 g/L of yeast extract, and 100 g/L of glucose. Solid media for both E. coli and S. cerevisiae were prepared by adding 15 g/L of agar to the liquid medium. An appropriate antibiotic was supplemented into the medium depending on the selection markers used. Plasmid pAUR101 (TakaraBio Inc., Kusatsu, Japan) was employed to amplify transgene expression cassettes in E. coli, with the antibiotic ampicillin serving as a selection marker. For yeast fermentation, the antibiotics ampicillin and kanamycin were used to prevent E. coli contamination, while the antibiotic geneticin/G418 (Cat. No. ant-gn-1, InvivoGen, San Diego, CA, USA) was used to maintain positive recombinant offspring during successive subcultures. For comparative analysis, S. cerevisiae I231 (NBRC1226), J81 (NBRC1953), and J147 (NBRC103959) from the NBRC culture collection, and CCUG53310 sourced from the Culture Collection University of Gothenburg (Sweden) were also included as flocculant yeast strains for evaluation.

2.2. Construction of Transgene Expression System for D-Lactate Synthesis

To enable S. cerevisiae to produce D-LA from glucose, the expression system for the transgene cassettes involved in D-LA synthesis was constructed as follows (Figure 1). Initially, a codon-optimized D-LDH transgene from Leuconostoc pseumesenteroides (GenBank accession number: MW574953) was arranged in the expression cassette driven by the TDH3 promoter and terminator. This transgene cassette was combined with the KanMX expression cassette to provide an antibiotic selection for the designed recombinant strains. Consequently, the positive recombinant strain could be selected and maintained for further investigation. To introduce the transgene expression cassettes into the chromosomes of the wild-type strain using homologous recombination techniques, left-arm and right-arm sequences containing 5′-CYB2, 5′-DLD1, 5′-PDC1, 5′-PDC5, 5′-PDC6, 5′-MPC1, and 5′-ADH1 partial left sequences and 3′-CYB2, 3′-DLD1, 3′-PDC1, 3′-PDC5, 3′-PDC6, 3′-MPC1, and 3′-ADH1 partial right sequences from the S. cerevisiae chromosomes were cloned from the F118 strain genome by PCR method using the primer sets listed in Supplementary Table S1. These were infused with transgene expression cassettes, as shown in Figure 1, resulting in their genome integration into the CYB2, DLD1, PDC1, PDC5, PDC6, MPC1, and ADH1 loci. The transgene expression cassettes bearing only the KanMX expression cassette were designed using the primer sets provided in Supplementary Table S2 for constructing disrupted strains lacking D-LDH, which served as negative controls in this study. The linear nucleotides were subsequently transformed into yeast cells following the lithium acetate/single-stranded carrier DNA/polyethylene glycol method [14]. Transformants were obtained by incubating the cells on YPD agar with 200 mg/L of geneticin at 30 °C for 3 to 5 days. The transgene expression cassette, which includes the left and right arms of the locus, was inserted into the SmaI restriction site of pAUR101 and transformed into E. coli for further amplification and purification prior to the yeast transformation.

2.3. Fermentation

A single colony of the yeast strain grown on YPD agar was selected and inoculated into 12 mL of YPD100 medium (100 g/L of glucose) in a 100 mL Erlenmeyer flask. The cultures were incubated at 150 rpm for 1 to 2 days to serve as a seed culture. To initiate the main culture, the seed was inoculated into 12 mL of YPD with 100 g/L of glucose (YPD100) and cultivated at 90 rpm until the glucose in the medium was depleted. After harvesting the cells from the seed culture, the initial cell concentration was adjusted to OD_600nm = 5 to ensure that the F118 WT strain, the primary host yeast, could be conditioned for early exponential growth, initiating fermentation and completely consuming 100 g/L of glucose within 24 h. The experiments were conducted with three biological replicates, and the cultivation time varied depending on the strain, ranging from 24 to 144 h.

2.4. Fermentation Product Measurements

Several compounds related to the fermentation, including glucose, D-LA, glycerol, acetic acid, and ethanol in the fermentation samples, were analyzed using high-performance liquid chromatography (HPLC) equipped with a refractive index detector (RID-10A, Shimadzu, Kyoto, Japan). A Coregel-87H column (7.8 mm ID × 300 mm, Transgenomic Inc., New Haven, CT, USA) was used at 80 °C with 5 mM sulfuric acid as the eluent at a 0.6 mL/minute flow rate for 40 min. Fermentation samples were periodically collected and then centrifuged at 14,000× g for 5 min at 4 °C. The supernatants were collected and filtered through a 0.45 μm polytetrafluoroethylene (PTFE) filter (Merck Millipore, Carrigtwohill, Ireland) into HPLC vials.

2.5. Cell Growth Measurement

The growth of the yeast was assessed by measuring the dry cell weight, as the flocculent F118 strains could not be evaluated using optical density. At the start (0 h) and end of batch fermentation, cultures from each sample were centrifuged at 14,000× g for 5 min at 4 °C to harvest the cell pellets, which were washed twice with sterilized water. The pellets were then stored at −80 °C and freeze-dried overnight. Finally, the dry cell weight was measured.

2.6. Cell Morphological Observation

The morphological changes in the cells resulting from gene disruption and the expression of LpDLDH transgene cassettes in the F118 strain were observed using a BIOREVO BZ-9000 digital microscope (Keyence, Osaka, Japan) at the magnification power of a ×100 objective lens.

2.7. Cell Wall Hydrophobicity Test

The cell wall hydrophobicity of the cultivated cells was determined using a modified microbial adhesion to hydrocarbon assay [13,15]. The cells were deflocculated with 50 mM ethylenediaminetetraacetic acid (EDTA, pH 7) and resuspended in 2 mL of 0.9% sodium chloride solution. The initial absorbance (A₀) was measured at 600 nm using a spectrophotometer (UVmini-1240, Shimadzu). Subsequently, 0.4 mL of octane was added to the cell suspension, and the mixture was vortexed at maximum speed for 60 s. After standing at room temperature for 10 min, the aqueous phase containing non-hydrophobic cells was separated, and its absorbance (A₁) was measured at 600 nm. The cell wall hydrophobicity was calculated using Formula (1).

$$\mathrm{Cell} \mathrm{wall} \mathrm{hydrophobicity}=(1-{\displaystyle \frac{ A_1}{A_0}}\times 100)$$

(1)

2.8. Quantitative PCR for Gene Expression Analysis

Yeast cells were collected during the mid-exponential phase of aerobic fermentation, as this is when cells show their peak metabolic activity. Conducting gene expression analysis at this stage reduces the impact of stationary-phase stress responses and offers insights into how metabolic pathways are regulated under optimal growth conditions. In some cases, gene expression changes were assessed after 6 h of cultivation to capture early metabolic shifts. At this point, cells actively metabolize glucose, and the expression of genes involved in D-LA biosynthesis can be observed before significant environmental stress or nutrient depletion affects transcriptional regulation. To do that, a volume of 0.5–1 mL of cell suspension was centrifuged at 8000 × g for 10 min at 4 °C. The pellets were washed once with cold sterilized water and stored at -80 °C until RNA extraction. RNA was extracted using the NucleoSpin RNA kit (Macherey-Nagel, Düren, Germany), with DNA contamination minimized by rDNase treatment during column digestion, following the manufacturer’s instructions. Reverse transcription was performed using ReverTra Ace qPCR RT Master Mix with gDNA remover (Toyobo, Osaka, Japan), and the resulting cDNA was used for quantitative PCR (qPCR). Gene expression levels were quantified using KOD SYBR qPCR Mix (Toyobo) with 0.4 µM of each forward and reverse primer and 10 ng/µL of diluted cDNA. qPCR was conducted on a Mx3005P system (Agilent Technologies Ltd., Tokyo, Japan). The PCR program included an initial denaturation at 95 °C for 10 min, followed by 60 cycles of 30 s at 95 °C, one minute at 55 °C, and one minute at 72 °C for elongation. The final step included one cycle of one minute at 95 °C, 30 s at 55 °C, and 30 s at 95 °C.

The ACT1 gene was used as an internal reference due to its stable expression across all samples, with consistent Ct values. In some cases, the TUB1 gene was used as an internal reference instead of the ACT1 gene. The expression levels of target genes were calculated as fold changes using the formula 2^−ΔΔCt, where ΔΔC_t = ΔC_t sample − ΔC_t control and ΔC_t sample = C_t-sample − C_t-ACT1 and based on calculations from the MxPro QPCR software v4.10 (Agilent Technologies Ltd., Tokyo, Japan). Primer sequences used in the analysis were designed using the NCBI Primer-BLAST tool (http://www.ncbi.nlm.nih.gov/tools/primer-blast/) (accessed on 1 March 2025), with a target product size of approximately 100 bp, as listed in Supplementary Table S3.

2.9. Statistical Analysis

All experiments on cultivations were conducted in a single independent cultivation with three technical replicates per strain. The mean and standard deviation (SD) were calculated for each dataset. Statistical significance between strain groups of D-LA-producing strains and disrupted strains was determined using an unpaired two-tailed Student’s t-test. A p-value of <0.05 was considered statistically significant.

2.10. Differential Expression Analysis

Differential gene expression analysis was conducted using the DESeq2 package in R (version 4.4.2). Count data were imported into DESeq2, and genes with fewer than 10 total counts across all samples were filtered out to reduce noise. The count matrix was normalized using DESeq2’s median of ratios method to account for differences in sequencing depths across samples.

Statistical testing was performed using the DESeq function, with a negative binomial model used to estimate fold changes and significance. Genes were considered significantly differentially expressed (DEGs) if they met the following criteria: adjusted p-value (P_adj) < 0.05 and absolute log₂ fold change (Log₂FC) > 1. DEGs were visualized using a volcano plot, highlighting significantly upregulated and downregulated genes. The volcano plot was generated in R to visualize the distribution of DEGs, with upregulated genes (Log₂FC > 1, P_adj < 0.05) highlighted in red and downregulated genes (Log₂FC < −1, P_adj < 0.05) highlighted in blue. Genes that did not meet the significance thresholds are shown in black.

3. Results

3.1. D-LA Production Profiles by Engineered Flocculant S. cerevisiae Strains

To evaluate D-LA production by engineered flocculant S. cerevisiae strains, the strains were cultivated in Erlenmeyer flasks with the medium containing 100 g/L of glucose at 150 rpm and 30 °C until glucose depleted. Since the strains have a strong flocculation trait, the initial cell concentration could not be adjusted accurately; therefore, the same wet cell weight (approximately equal to OD_600nm = 5) was used, as mentioned in Section 2.3.

As shown in Figure 2, all WT strains exhibit a flocculation trait, with the F118 strain demonstrating more pronounced flocculation behavior than the others. The disruption of the CYB2 gene in all strains (ΔCYB2) does not influence the flocculation traits of the strain. However, after implementing the transgene expression system for D-LA production at the CYB2 locus in their chromosomes, the recombinant F118 strain (ΔCYB2::LpDLDH) retained the flocculation trait, albeit to a lesser degree compared to the WT strain.

In contrast, the other recombinant strains (I231ΔCYB2::LpDLDH, J81ΔCYB2::LpDLDH, J147ΔCYB2::LpDLDH, and CCUGΔCYB2::LpDLDH) lost their flocculation traits. The loss of flocculation traits indicates a cellular response to metabolic stress caused by integrating the D-LA synthesis pathway. Typically, disrupting the CYB2 gene does not impact S. cerevisiae cell growth or phenotype, as its expression is generally suppressed during glucose fermentation. Similarly, CYB2 gene expression decreases under anaerobic conditions while producing lactic acid [16,17]. Research by Pangestu et al. [18] demonstrated that the recombinant F118 strain (ΔCYB2::LcLLDH), which utilized the L-LA production system employing L-LDH from Lactobacillus casei, preserved its flocculation behavior without any decline. This finding was corroborated by the upregulation of flocculation-related genes such as FLO5 and genes associated with cell wall hydrophobicity, including YGP1 and its Haa1 regulon.

In contrast, these genes (FLO5 and HAA1) are downregulated during the production of D-LA, as illustrated in Figure 3. The PMA1 gene, also downregulated by the Haa1 regulon, encodes the H(+)-ATPase Pma1 on the yeast plasma membrane, which expels protons dissociating from weak acid molecules [19]. Pma1 is crucial for yeast adaptation to weak acids, generating an electrochemical proton gradient vital for nutrient uptake and the regulation of intracellular pH [20,21]. Furthermore, the Haa1 regulon significantly contributes to the rapid adaptation of yeast to weak acids, regulating approximately 80% of weak acid-induced gene expressions, both directly and indirectly [22,23]. This indicates that the substantial decrease in HAA1 gene expression observed (Figure 3) contributed to the compromised condition of F118 cells and reduced flocculation levels. Recently, the flocculation behavior of the F118 strain was altered in response to the concentration of inhibitory chemical compounds (ICCs) in the culture medium. This study indicated that Mot3p was involved in this response mechanism based on its range of prionogenicity [13]. In this study, the expression of the MOT3 gene was nearly identical to that of the ΔCYB2 strain. However, the expression levels of the FLO5 and YGP1 genes, which are associated with flocculation and modified cell wall hydrophobicity, respectively, decreased. This suggests that the D-LA produced within the F118 cells triggered a different ICC response than those ICCs present outside the cells. This contrasts with the previously reported phenomenon of L-LA production in the ΔCYB2 strain [18]. This indicates that D-LA is more toxic than L-LA for yeast cells despite the fact that the F118 strain is more resilient than the other evaluated strains. The loss of flocculation traits in recombinant strains (I231ΔCYB2::LpDLDH, J81ΔCYB2::LpDLDH, J147ΔCYB2::LpDLDH, and CCUGΔCYB2::LpDLDH) might be due to low expressions of the FLO5 and YGP1 as well as MOT3 genes in these WT strains. Consequently, the detrimental effect on the flocculation traits and their resilience during the fermentation of D-LA was more pronounced than in the F118 strain, as shown in Figure 4. The disruption of CYB2 in all strains (ΔCYB2) has no significant effect on shape and aggregation properties, as the WT and ΔCYB2 strains are comparable. However, integrating the LpDLDH transgene expression cassette into the CYB2 locus significantly impacts the morphological and aggregation properties of the I231, J81, J147, and CCUG strains. The strain cells decrease in size compared to the WT strains and lose their flocculation traits, except for the F118 strain.

During fermentation, the detrimental effects of weak acids, such as lactic acid, lead to reduced glucose consumption and lower fermentation rates for D-LA production (Figure 5). To verify these effects, fermentations were conducted in a YPD medium containing glucose as a carbon source buffered with MES/NaOH (pH of 5.5, similar to the initial pH) at 100 mM, with or without CaCO₃ at 5 g/L, which served as a neutralizer for the lactic acid released into the medium. Consequently, glucose consumption occurred more rapidly when the F118, J147, and CCUG strains were cultivated in a medium without buffering to a pH of 5.5 and without CaCO₃ neutralization (a < b < c), indicating that these strains are more resilient than the I231 and J81 strains during D-LA production under these conditions. A similar trend was observed in D-LA production by the recombinant strain of F118 (ΔCYB2::LpDLDH), which achieved 56 g/L of D-LA after 12 h of fermentation, yielding 0.56 g of D-LA/g-glucose consumed. Interestingly, the recombinant strains of I231ΔCYB2::LpDLDH, J81ΔCYB2::LpDLDH, J147ΔCYB2::LpDLDH, and CCUGΔCYB2::LpDLDH failed to produce D-LA when cultivated in the medium buffered to a pH of 5.5 and neutralized with CaCO₃. These strains were ultimately capable of producing D-LA when grown in the medium without buffering and CaCO₃ neutralization, though the yield was very low (less than 0.05 g of D-LA/g-glucose consumed). In contrast, ethanol was primarily produced, differing from the culture of strain F118. Thus, the detrimental effects of weak acids, such as lactic acid, on glucose consumption and D-LA production have been clarified, revealing a significant causal relationship with changes in flocculation traits.

In this study, the robust constitutive TDH3 promoter was utilized to enhance the high expression of the LpDLDH gene in all engineered strains under elevated glucose conditions. However, its activity diminishes in acidic environments. Once D-LA is produced and released into the medium, extracellular lactic acid can diffuse through the plasma membrane into the cytosol, dissociating into lactate and protons, which promotes intracellular acidification. This results in a decrease in D-LA yield due to reduced cell growth and the repression of TDH3 promoter activity [24]. Nevertheless, the F118ΔCYB2::LpDLDH strain can achieve a high yield of D-LA because it maintains the flocculation trait during fermentation. In contrast, the other flocculating strains (I231ΔCYB2::LpDLDH, J81ΔCYB2::LpDLDH, J147ΔCYB2::LpDLDH, and CCUGΔCYB2::LpDLDH) could not sustain this trait even when the D-LA concentration was low. Consequently, D-LA may diffuse into the cytosol and lower the intracellular pH. This observation indicates that the flocculation trait is crucial for the yeast strains’ ability to tolerate weak acids and support effective lactic acid production, even in conditions lacking a buffering effect and without neutralization. The superior performance of strain F118 under challenging conditions highlights its potential as a strong platform for industrial-scale D-LA production.

3.2. Ideal Integration Spot of D-LDH in Flocculant S. cerevisiae F188 Strain for D-LA Production

The ideal genomic integration sites were considered to optimize D-LDH integration into the F188 strain for D-LA production. The engineered strains were developed to enhance D-LA production by altering the interactions between the ethanol and D-LA pathways (Figure 6). Research focused on lactate-oxidizing enzymes (DLD1, CYB2) has consistently shown higher D-LA yields in engineered yeast strains [25]. Pyruvate decarboxylases, encoded by PDC1 and PDC5, convert pyruvate into acetaldehyde during ethanol fermentation. In S. cerevisiae, most of the catalytic activity for converting acetaldehyde into ethanol is attributed to the cytosolic alcohol dehydrogenase (ADH1) gene [26]. Disrupting these genes results in pyruvate accumulation, increasing the substrate pool for D-LDH to convert pyruvate into D-LA. Lactate production has risen dramatically in strains with PDC gene deletions that overexpress D-LDH due to the greater availability of pyruvate [27].

Mpc1, an enzymatic component of the mitochondrial pyruvate transport system, plays a significant role in this process. Inactivating the MPC1 gene that encodes Mpc1 results in the cytosolic retention of pyruvate, increasing its availability for D-LDH-mediated conversion to D-LA. Studies on mitochondrial pyruvate transport systems have revealed that disrupting the MPC1 gene leads to elevated cytosolic pyruvate levels, promoting increased lactate production. The literature indicates that the absence of the MPC1 gene is expected to cause the most significant reduction in mitochondrial pyruvate uptake, as the formation of both active carrier complexes (MP_COX and MPC_FERM) is hindered without the Mpc1 subunit [28].

In this study, the F118 strain was engineered to incorporate a transgene expression system for D-LA synthesis into the CYB2, PDC1, DLD1, ADH1, PDC5, PDC6, and MPC1 loci, as outlined in Section 2.1 (Figure 1). The integration of the D-LDH transgene expression cassette into the various chromosomal loci of the F118 strain was achieved through homologous recombination. Subsequently, evolved strains were assessed for their ability to produce D-LA and their metabolic characteristics based on the locus of the D-LDH transgene expression cassette in the chromosomes.

D-LA production was performed using a high glucose concentration of 100 g/L to maximize the D-LA titer. The strain with the D-LDH transgene expression cassette integrated at the PDC1 locus (ΔPDC1::LpDLDH) showed the highest D-LA titer, reaching 51.36 g/L, while the strain at the CYB2 locus (ΔCYB2::LpDLDH) produced 31.92 g/L (Figure 7a). Moreover, the formation of ethanol byproducts was lower in the ΔPDC1::LpDLDH strain compared to the ΔCYB2::LpDLDH strain, with a reduction of approximately 35%, indicating a successful swap or disruption of the PDC1 gene, which plays a central role in the ethanol pathway. These findings highlight the critical influence of the integration locus on the balance between D-LA production and ethanol byproduct formation, with the PDC1 locus offering a more favorable environment for efficient D-LA synthesis in the F118 strain. Strains with D-LDH integrated at the ADH1 locus (ΔADH1::LpDLDH) or the PDC5 locus (ΔPDC5::LpDLDH) successfully grew on the media containing G418 as a selection marker and consumed the glucose during fermentation by 48 h. However, they did not produce D-LA, indicating that the transgene expression system for the D-LDH gene, despite being integrated, may not have been functional or active enough to facilitate D-LA production.

Notably, the strain with D-LDH integrated at the PDC6 locus (ΔPDC6::LpDLDH) or the MPC1 locus (ΔMPC1::LpDLDH) was able to grow on the selective media. It produced D-LA, indicating that these integrations were sufficient for transformation and D-LDH gene expression. The ΔPDC6::LpDLDH and ΔMPC1::LpDLDH strains produced D-LA at lower titers of 10 g/L and 0.026 g/L, respectively. Both strains consumed glucose within 24 h and produced ethanol at 45 g/L and 41 g/L, respectively, comparable to the WT strain at 40 g/L. The high ethanol production in these strains suggests an incomplete redirection of the pyruvate flux toward D-LA synthesis, potentially due to the insufficient expression of D-LDH or the dominance of competing pathways.

The disrupted F118 strains lacking D-LDH (∆CYB2, ∆PDC1, ∆DLD1, ∆ADH1, ∆PDC5, ∆PDC6, and ∆MPC1), which served as negative controls (Figure 7b), exhibited varying growth rates, glucose consumption, and ethanol production. Notably, the ∆CYB2, ∆DLD1, and ∆MPC1 strains displayed the fastest growth among the strains, fully consuming glucose within 24 h. These strains also demonstrated similar ethanol production levels, reaching 45.97 g/L, 46.04 g/L, and 45.72 g/L, respectively. In contrast, the ∆PDC5 and ∆PDC1 strains showed slower growth, with glucose consumption delayed until 48 h. The ∆ADH1 strain experienced the slowest growth, taking approximately 144 h to completely consume glucose. Notably, the ∆PDC6 strain could not utilize glucose, with levels remaining around 100 g/L even after 48 h without cell growth. These results highlight the significant differences in growth and metabolic efficiency among the disrupted strains, indicating variations in growth and metabolic efficiency. The ∆CYB2, ∆DLD1, and ∆MPC1 strains demonstrated higher metabolic activities and ethanol production, while ∆ADH1 and ∆PDC6 exhibited impaired glucose utilization.

Based on these results, the CYB2 loci can serve as targets for integrating the D-LDH transgene expression system in the F118 strain without compromising cell growth and metabolic activity. These findings underscore the potential for the further optimization of D-LA production in the F118 strain. Furthermore, the multi-integration of the D-LDH transgene expression system into PDC1 and CYB2 may enhance D-LA production. The engineered strains effectively produce D-LA even at high glucose concentrations. Since the expression of the D-LDH transgene in these strains is driven by the TDH3 promoter, it is positively influenced by the glucose concentration in the medium. Catabolite repression occurs during D-LA production to prevent the utilization of D-LA and other byproducts, such as acetic acid and ethanol, as alternative carbon sources instead of glucose. Additional integration into the ADH1 loci or the disruption of this gene to reduce ethanol byproduct formation is not recommended, as it may severely impact cell growth and metabolic activities, leading to low titers of D-LA production.

3.3. Flocculation Behavior Across the Engineered Strains

To observe the effect of the robustness of the strain, flocculation was analyzed during the stationary phase compared to the WT strain, which functioned as a strong flocculant strain. The ΔCYB2::LpDLDH strain exhibited significant clumping, leading to the formation of large cell aggregates in the culture broth, while the ΔPDC1::LpDLDH strain showed less pronounced flocculation, resulting in smaller clumps (Figure 8c). These observations related to dry cell weight measurements for the ΔPDC1::LpDLDH strain, which indicated the lowest weight at 2.97 g/L. These differences highlight the variations in flocculation behavior between the two strains, with the ΔCYB2::LpDLDH demonstrating a higher degree of flocculation and larger cell aggregates. Meanwhile, the ΔPDC1::LpDLDH strain showed a lower, yet still evident, level of flocculation. Despite the discrepancies in flocculation behavior, the ΔCYB2::LpDLDH and the ΔPDC1::LpDLDH strains exhibited high hydrophobicity values (

Table 1. Summary of D-LDH transgene expression system’s integrated and disrupted strains.

Strain	Time a (hour)	Final D-LA (g/L)	Final Ethanol (g/L)	Final Glycerol (g/L)	Flocculation/ Hydrophobicity (%)	Final Dry Cell Weight (g/L)
Wild type	24	0.04 ± 0.03	40.63 ± 0.64	2.94 ± 0.26	Flocculant/89.32	6.31 ± 0.73
ΔCYB2::LpDLDH	12	31.92 ± 0.98 ***	31.64 ± 0.68 **	0	Flocculant/97.16	4.77 ± 0.22
ΔPDC1::LpDLDH	48	51.36 ± 1.78 ***	20.2 ± 1.69 **	0	Less flocculant/98.38	2.97 ± 0.12
ΔMPC1::LpDLDH	24	10.32 ± 0.35 ***	41.53 ± 0.22 ***	0	Flocculant/N.D.	6.96 ± 0.15
ΔPDC6::LpDLDH	24	0.026 ± 0.01 ***	45.11 ± 0.35 ***	4.25 ± 0.42 **	Flocculant/N.D.	9.42 ± 0.44
ΔDLD1::LpDLDH	48	0 *	33.84 ± 0.42 *	1.39 ± 0.05 ***	Non-flocculant/1.57	5.05 ± 0.059
ΔADH1::LpDLDH	48	0 *	38.29 ± 2.66 *	1.31 ± 0.17 **	Non-flocculant/N.D.	6.9 ± 0.8
ΔPDC5::LpDLDH	48	0 *	41.69 ± 1.33 ***	1.39 ± 0.11 **	Non-flocculant/N.D.	6.63 ± 0.32
∆CYB2	24	0	45.96 ± 1.26	4.7 ± 1.15	Flocculant/N.D.	9.57 ± 1.5
∆PDC1	72	0	43.06 ± 1.07	4.41 ± 0.51	Flocculant/N.D.	3.95 ± 0.95
∆MPC1	21	0	45.72 ± 0.81	5.19 ± 0.61	Flocculant/N.D.	8.98 ± 0.17
∆PDC6	48	0	1.45 ± 2.43	0.57 ± 0.01	Less flocculant/N.D.	N.D. b
∆DLD1	24	0	46.03 ± 1.19	3.95 ± 0.09	Flocculant/N.D.	9.26 ± 0.48
∆ADH1	144	0	39.07 ± 2.1 *	4.78 ± 0.21	Less flocculant/N.D.	3.71 ± 0.33
∆PDC5	48	0	44.61 ± 1.27	3.98 ± 0.11	Flocculant/N.D.	3.46 ± 0.37

^a fermentation time; ^b ΔPDC6 strain exhibited incomplete fermentation, with slow growth and residual glucose (~100 g/L) remaining after 48 h. Statistical significance was assessed using Student’s t-test. Each value is the mean ± standard deviations (SD) (n = 3), * p < 0.05, ** p < 0.01, and *** p < 0.001. N.D., not determined.

). Additionally, ΔPDC6::LpDLDH and ΔMPC1::LpDLDH presented large clumps of cells. Notably, the strain ΔPDC6::LpDLDH displayed the highest dry cell weight. This indicates that cell surface hydrophobicity does not primarily drive the differences in flocculation observed between the strains, as both exhibited high hydrophobicity regardless of their aggregation patterns. Considering the high hydrophobicity noted in both strains, the reduced flocculation in the ΔPDC1::LpDLDH strain may relate to genetic factors rather than surface hydrophobicity, as mentioned in Section 3.1.

In contrast, strains with D-LDH integrated into the DLD1 (ΔDLD1::LpDLDH), ADH1 (ΔADH1::LpDLDH), and PDC5 (ΔPDC5::LpDLDH) loci, which failed to produce D-LA, were observed to be completely non-flocculant. These strains showed no visible clumping, and the cells remained evenly dispersed throughout the culture, indicating a loss or absence of flocculation behavior. Interestingly, the disrupted strains lacking D-LDH (∆CYB2, ∆PDC1, ∆DLD1, ∆ADH1, ∆PDC5, ∆PDC6, and ∆MPC1), which served as negative controls (Figure 8a), still possessed the strong flocculation trait, similar to the WT strain; however, only the ∆ADH1 and ∆PDC6 strains exhibited the weak flocculation trait compared to the other strains.

The disruption of the locus genes in all strains (ΔCYB2, ΔPDC1, ΔPDC6, and ΔMPC1) has no significant effect on cell aggregation properties, as comparable to the WT strain, as shown in Figure 8b. However, integrating the LpDLDH transgene expression cassette into the DLD1, ADH1, and PDC5 loci significantly impacts the morphological and aggregation properties of the ΔDLD1::LpDLDH, ΔADH1::LpDLDH, and ΔPDC5::LpDLDH strains. The strain cells decrease in size compared to the ΔDLD1, ΔADH1, and ΔPDC5 strains and lose their flocculation traits, as shown in Figure 8d.

3.4. Gene Expression Analysis of D-LA-Producing Strains

A gene expression analysis was used to investigate mechanisms influencing lactic acid fermentation. TUB1 and ACT1 genes were evaluated across yeast strains using stability assessment tools to select an appropriate housekeeping gene for qPCR normalization. The chosen gene was used to normalize expression data for target genes related to D-LA synthesis and strain robustness. qPCR experiments were conducted and analyzed to identify differentially expressed genes. Gene expression patterns should be correlated with fermentation profiles and D-LA production levels to elucidate underlying mechanisms.

The results of the Ct values of housekeeping genes are listed in Supplementary Table S4. Results indicated that the ACT1 gene exhibited more consistent Ct values, making it a more reliable reference gene. In contrast, the TUB1 gene displayed variability, with a Ct value of 27.73 in the ΔPDC1::LpDLDH strain, three cycles higher than in the WT strain. The strain capable of producing D-LA showed low Ct values, which is beneficial as it confirms the presence of mRNA levels for the housekeeping genes. In contrast, the non-producing strain exhibited high Ct values, indicating potential issues. This difference suggests a problem with the non-producing strain, as ACT1 or TUB1, being essential housekeeping genes, should usually be expressed in yeast cells; therefore, gene expression analysis can compare the D-LA producing strains (ΔCYB2::LpDLDH, ΔPDC1::LpDLDH, ΔMPC1::LpDLDH, and ΔPDC6::LpDLDH) to control (WT strain) to eliminate the backgrounds of related gene expressions after disruption.

The heatmap analysis (Figure 9a) shows the expression levels of 30 genes across four strains with varying D-LA production levels, as described in the previous section (Figure 7, Table 1). Among these, the LpDLDH gene, encoding the D-LDH, exhibits significantly higher expressions in the ΔPDC6::LpDLDH, ΔMPC1::LpDLDH, and ΔPDC1::LpDLDH strains (red color). However, the ΔCYB2::LpDLDH strain displays a notably low Log₂FC in LpDLDH expression despite producing 30 g/L of D-LA.

We also examined the gene expression changes associated with flocculation behavior by focusing on the FLO5 gene. We evaluated the expression levels of FLO5 across various engineered strains, including ΔCYB2::LpDLDH, ΔPDC1::LpDLDH, ΔMPC1::LpDLDH, and ΔPDC6::LpDLDH, which have previously been shown to produce different levels of D-LA. The heatmap data revealed significant differences in FLO5 expression among the strains. Notably, the ΔCYB2::LpDLDH strain showed no substantial changes in FLO5 expression compared to the other engineered strains, suggesting that the flocculation behavior in this strain was not significantly influenced by the integration of the D-LDH gene at the CYB2 locus. In contrast, other strains exhibited the downregulation of FLO5, with the ΔPDC1::LpDLDH strain displaying the lowest Log₂FC, indicating a more pronounced reduction in FLO5 expression. This downregulation was consistent with the volcano plot analysis (Figure 9b), where FLO5 was identified as one of the most significantly downregulated genes in response to the experimental conditions. Additionally, three genes (LpDLDH, YGP1, SCW11, and SED1) were upregulated. In contrast, the others were downregulated, indicating that integrating LpDLDH at the CYB2 locus can facilitate a high expression of the D-LDH gene for D-LA production and enhance the expression of the YGP1, SCW11, and SED1 genes involved in cell wall hydrophobicity (integrity), which is essential for maintaining cell robustness during fermentation.

These findings highlight the transcriptional impact of integrating the D-LDH gene into different loci, which may influence the efficiency of D-LA production and the metabolic balance in the engineered F118 strains.

4. Discussion

Previous studies have similarly reported that genomic loci can influence both transcriptional activity and metabolic pathway stability, further validating the strategic selection of these sites for metabolic engineering [29,30].

Among the engineered strains, only LpDLDH integrated into CYB2, PDC1, PDC6, and MPC1 loci successfully produced D-LA, while other strains either failed to produce the desired metabolite or exhibited growth defects. Notably, strains with integrations into loci such as DLD1, ADH1, and PDC5 loci either showed no D-LA production or did not grow in a medium containing antibiotics, possibly due to metabolic stress or the disruption of essential pathways. The differential production of D-LA, ethanol, and glycerol across strains highlights the influence of locus-specific metabolic fluxes. For example, strains that produced higher levels of D-LA often exhibited reduced ethanol production, suggesting a successful redirection of carbon flux away from the ethanol pathway. These findings are consistent with prior studies on ethanol pathway disruption, where metabolic rewiring resulted in the improved production of target metabolites [10,31,32]

The fermentation profiles of strains with gene disruptions but lacking the LpDLDH gene expression system revealed significant variations in growth and metabolite production. Strains disrupted in PDC6 exhibited particularly slow growth and failed to complete fermentation, highlighting the essential role of this gene in glycolytic flux and redox balance. In contrast, the disruption of the ADH1 gene resulted in reduced ethanol production, consistent with its central role in ethanol biosynthesis. Additionally, the non-producing strains exhibited a loss of flocculation, indicating a potential link between metabolic stress and cell wall integrity. Conversely, the high-producing ΔPDC1::LpDLDH strain exhibited reduced flocculation, which, while advantageous for fermentation homogeneity, may pose challenges for downstream processing. The producing strain, the ΔPDC1::LpDLDH strain (51 g/L of D-LA), exhibited reduced flocculation, and it has been reported that the replacement of PDC1 with LDH to produce L-lactic acid was efficient [33].

The downregulation of the FLO5 gene observed in most engineered D-LA-producing S. cerevisiae strains suggests a significant shift in cell wall integrity and flocculation behavior. The FLO5 gene is a key gene within the FLO gene family, which encodes cell wall glycoproteins responsible for flocculation by mediating cell-cell adhesion via lectin-like interactions with mannose residues on adjacent cells [15]. The repression of FLO5 gene expression in the engineered strains likely contributes to the loss of flocculant properties, leading to a more dispersed phenotype. A study about Pichia pastoris producing D-LA incorporated flocculation and showed the downregulation of cell wall rigidity [34].

Several studies have highlighted the role of the FLO5 gene in determining the adhesive and flocculant characteristics of yeast cells. The study by [35] demonstrated that the overexpression of the FLO5 gene enhances flocculation, whereas its deletion or downregulation reduces cell aggregation and increases sedimentation times. Other studies stated [36] that the specific adhesion properties are primarily determined by the characteristics of particular flocculin protein machinery rather than general cell wall properties, such as hydrophobicity. Each FLO gene expression contributes to distinct phenotypes and varying phenotype intensities, with FLO1 and FLO5 gene expressions responsible for cell aggregation and flocculation.

Additionally, environmental factors such as pH shifts and osmotic stress—conditions that arise during high D-LA production—are known to modulate FLO5 gene expression [37]. The accumulation of D-LA in the fermentation medium likely contributes to stress responses, further driving the downregulation of flocculation genes as part of the yeast’s adaptation mechanism.

From an industrial perspective, the observed downregulation of FLO5 gene expression may present advantages and challenges. Reduced flocculation can enhance bioprocess efficiency by promoting homogeneous mixing and improved substrate accessibility. However, it may also complicate downstream processing by increasing separation costs and reducing biomass retention in continuous fermentation systems. Future optimization strategies could involve fine-tuning FLO5 gene expression using promoter engineering or laboratory adaptive evolution to balance production and flocculation properties.

The successful production of D-LA from glucose using engineered S. cerevisiae demonstrates the feasibility of utilizing genomic locus selection and transcriptomic insights for pathway optimization. Optimizing fermentation conditions, such as pH control and oxygen availability, could enhance yields and minimize byproducts like ethanol. Research by [38] employed an acid-tolerant strain to efficiently produce L-LA. A combination of metabolic engineering and fermentation regulation in these acid-tolerant strains proves beneficial for advancing LA production by S. cerevisiae. Additional challenges must be addressed before S. cerevisiae can become a commercially viable platform for renewable lactic acid production. Future engineering efforts aimed at increasing S. cerevisiae’s tolerance to low pH conditions and inhibitors present in common feedstock hydrolysates will be crucial for successful commercial applications [39]. The research conducted by [40] examined the transcriptional response of S. cerevisiae to LA enantiomers. The impact of pH on the D-LA response has highlighted the necessity of controlling and optimizing D-LA production in yeast under both neutralizing and non-neutralizing conditions separately.

Finally, further investigations into the relationship between metabolic stress, flocculation, and cell wall integrity under varying environmental conditions could provide a more comprehensive understanding of the cellular dynamics in engineered strains.

5. Conclusions

This study highlights the essential points for genetically engineering S. cerevisiae yeast strains to produce D-LA. Firstly, the integration locations of genes in the yeast genome are discussed. Only specific locations (CYB2, PDC1, PDC6, and MPC1) led to successful D-LA production. Secondly, the metabolic changes resulting from genetic modifications influenced how yeast cells generated various substances. Yeast strains that produced more D-LA often produced less ethanol, indicating a shift in related metabolism. Thirdly, the growth and fermentation processes of disrupted strains are analyzed. Some modified strains grew slowly or could not complete fermentation, emphasizing the significance of certain genes for yeast survival and function. Fourthly, changes in flocculation behavior are observed. Many D-LA-producing strains lost their ability to aggregate (flocculate), which was linked to the reduced expression of the FLO5 gene, essential for cell–cell adhesion. Fifthly, industrial implications must be considered. The loss of flocculation may improve fermentation efficiency but could complicate separating yeast from the liquid afterward. Sixthly, as future directions, this study proposes methods to enhance D-LA production, such as fine-tuning gene expression, optimizing fermentation conditions, and further exploring the relationship between metabolic stress and cell behavior. Overall, this study illustrates how the careful selection of genetic modification sites and a thorough understanding of gene expression changes can help optimize yeast strains for producing desired substances like D-LA.