Effect of increasing cell density on metabolic dynamics
Our previous study showed that increasing the initial cell density ofS. cerevisiae resulted in an improved glucose uptake rate and enhanced lactic acid accumulation (Pangestu et al., 2022). The present findings align with our previous observations and are even further confirmed with more evidence. As depicted in Figure 2b, the augmentation of cell density expedited the rate of glucose consumption in both strains, with a more pronounced effect observed in the BTCC3L strain (red vs. yellow straight lines) than in the F118L strain (blue vs. green straight lines). Consistent with the screening results, the accumulation of lactic acid in the F118L strain also vastly surpassed that in the BTCC3L strain during cultivation (Figure 2a). However, increasing cell density in the BTCC3L strain led to an 18-fold increase in lactic acid yield (from 0.004 ± 0.002 to 0.07 ± 0.00 g·g glucose-1), which was still lower than the values observed in the F118L strain across both low- and high-density cell cultures (0.23 ± 0.05 g·g glucose-1 and 0.27 ± 0.03 g·g glucose-1, respectively). Notably, in comparison to our prior research, this BTCC3L strain (LDH +, CYB2 -) displayed slightly higher lactic acid accumulation than the BTCC3LX5 strain (LDH +, PDC5 -), though at a 2-fold lower level than that of the BTCC3LX1 strain (LDH +, PDC1 -) (Pangestu et al., 2022), thus revealing the order of how effective the disruption of those genes is in improving L-lactic acid production. In addition to lactic acid, the accumulation of other major metabolites, namely, ethanol and glycerol, also exhibited elevation in the BTCC3L strain when cultivated at a higher cell density. To provide stronger evidence, we also attempted to perform an identical genetic modification strategy in the BY4741 strain, a common laboratory strain exhibiting no flocculation behavior, and observed a similar trend, i.e., enhanced accumulation of major metabolites as a result of amplifying the cell density (Figure S1d). In contrast, the concentration of lactic acid in the F118L strain experienced a relatively negligible change, albeit a slight improvement, in response to altering the initial cell concentration, implying a lesser cell-density dependence in lactic acid accumulation by the flocculant yeast.
Gene expression analysis revealed significant changes as the cell density increased (L vs. H ). Specifically, in the BTCC3L strain, 31 genes were upregulated, with the expression of theHSP26 gene being the highest, and 1 gene, ACS1 , was downregulated (|LFC| ≥ 2, -log10adjusted p -value ≥ 1.30) (Figure 3a). Figure 3b illustrates the interaction network of DEGs in the BTCC3L strain. Notably, the majority of DEGs are related to glycolysis and pyruvate metabolism, represented by gray nodes. Additionally, six glycolytic genes (FBA1 ,PFK1 , PGK1 , PGM2 , TDH3 , and TPI1 ) connect the gray group with other clusters associated with the ETC (yellow nodes), cell growth (blue nodes), and stress response (red nodes). This observation supports the interconnectedness of metabolic turbulence in glycolysis with other pathways. While the PTK2 gene does not appear to have direct connections with other clusters, a deeper exploration of interactions at the second shell of the network reveals its connection to the TPK1 gene (not measured in this study), which is directly linked to the HXK1 , PGM2 , RAS2 , and SOL4 genes (Figure S2).
The dramatic upregulation of the HSP26 gene, the most upregulated DEG in the BTCC3L strain (shown in Figure 3c), was likely due to the higher lactic acid accumulation, as its transcription is known to be repressed under less acidic conditions (Carmelo & Sá-Correia, 1997). The other top five genes with the highest LFC (PGK1 ,SPI1 , TDH1 , and TDH3 ) are associated with glycolysis, highlighting the strong impact of increasing cell density on this pathway. Along with glycolysis, as shown in Figure 3d, other DEGs are also involved in amino acid biosynthesis (10 genes) and PPP (6 genes), with predominant expression in the cytosol (7 genes), mitochondria (7 genes), or both compartments (11 genes) according to the functional enrichment analysis (Raudvere et al., 2019). The results also indicate that altering the cell density affected proton transmembrane transport, ATP and pyruvate metabolism, NAD and NADP binding, and GAPDH (NAD+ and NAD(P)+) phosphorylation.
Figure 4 provides an overview of the impact of increasing cell density on the metabolic profile of the BTCC3L strain. The treatment resulted in the upregulation of genes associated with hexose transport (i.e.,HXT7 ), glycolysis (i.e., HXK1 , PGM2 , PFK1 ,FBA1 , TPI1 , TDH1, TDH2, TDH3, PGK1 , GPM2 andENO1 ), PPP (i.e., ZWF1 , SOL4 and TKL2 ) and ethanol metabolism (PDC5 and ADH3 ). These upregulations led to an increased glucose consumption rate and the generation of fermentation products. For comparison, a separate study employed a combined global metabolic engineering strategy and CRISPR-δ integration (GMES/CRISPR) approach to randomly introduce 12 glycolysis-related genes, complemented with the LDH gene, into the genome, aiming to amplify the activity of the pathway (Mitsui et al., 2020). In this study, we attained a similar outcome by only performing a higher-density cell cultivation without the need for extensive genetic modifications.
Additionally, the relative gene expression of the exogenous LDHgene was still low due to the intact PDC and ADH genes. However, we observed an increase in lactic acid accumulation (Figure 2a) and a noticeable enhancement of LDH gene expression (LFC = 5.83, Figure S3a) upon increasing cell density. Notably, the gene was not included as one of the DEGs in Figure 3c due to its relatively high variance among the three measured biological replicates, resulting in the -log10 adjusted p-value of 0.71 (lower than the predefined significance threshold of 1.30), yet this increased expression might be adequate to explain the observations. It is noteworthy that L-lactate can also be produced from methylglyoxal (Takatsume et al., 2006), a minor byproduct of glycolysis, as indicated by the upregulation of theGRE2 gene (Figure 4). However, further investigation is required to determine the significance of this pathway considering that the compound is not typically generated in large quantities inside the cell. Nevertheless, in high-cell density cultures, there is a possibility that the generation of methylglyoxal and lactaldehyde, its conversion product by the GRE gene, is substantial. If so, it presents a promising strategy to rewire glucose conversion to lactate via a nonpyruvate route, since the pyruvate-to-lactate route competes with several reactions, such as alcohol metabolism and the TCA pathway. It is important to note that cultivating the wild-type BTCC3 strain (without the introduction of the LDH gene) in high-cell density culture did not lead to an increase in lactic acid accumulation (Pangestu et al., 2022), which suggests that the increase in lactic acid production observed in this study was likely attributed to the genetic engineering strategies employed in the BTCC3 strain rather than solely the effect of high-cell density culture. It is also important to further study the affinity of the lactate dehydrogenase enzyme to methylglyoxal and lactaldehyde, as these compounds share chemical structure similarities with pyruvate.
The increased glycolytic activity resulted in an enhancement of the energy production capacity, as indicated by the significant upregulation of genes associated with the ETC, specifically the NDI1 ,CYC1 and CYT1 genes, which are located in the inner mitochondrial membrane and are responsible for facilitating the movement of electrons from respiratory complex I to IV (Figure 4). As electrons pass through these complexes, protons (H+) are pumped from the mitochondrial matrix to the intermembrane space, generating a concentration gradient (Mitchell, 1961). The intermembrane space becomes more positively charged compared to the mitochondrial matrix, and subsequently, the protons flow back to the mitochondrial matrix through the F0 domain, driving the rotation of a rotor within ATP synthase, which induces conformational changes in the F1 domain, leading to the conversion of ADP and inorganic phosphate to ATP. The synthesized ATP is then transported to supply energy for various processes, including numerous key metabolic pathways and cellular adaptation to changing environments. For instance, elevated lactic acid generation can lead to a depletion in intracellular pH. Vacuolar H+-ATPase (V-ATPase), consisting of multiple subunits, including VMA2 and VMA4 , plays a crucial role in acidifying the vacuole by pumping protons (H+) into the vacuolar lumen (Forgac, 2007) and maintaining the proper cytosolic pH for optimum enzymatic activities.
The upregulation of RAS2 in high-density cell culture, on the other hand, may be attributed to the increased accumulation of fructose-1,6-bisphosphate (F1,6BP) (Peeters et al., 2017) and the response to nutrient-limited conditions (Bhattacharya et al., 1995; Pothoulakis & Ellis, 2018) caused by the high cell-density conditioning. These changes then stimulate adenylate cyclase, an enzyme that converts ATP into cyclic adenosine monophosphate (cAMP), which then activates protein kinase A (PKA), which phosphorylates and regulates the activity of various downstream targets, including transcription factors that control the expression of ACT1 , a gene encoding actin involved in multiple cellular processes such as cell division, cell motility, intracellular transport, cell shape and structure, and signal transduction (Hubberstey & Mottillo, 2002; Pardee, 2010). Additionally, PKA activation influences the expression of the cell wall protein gene YGP1(Arnthong et al., 2022; Howard et al., 2001) and various stress-response genes, such asHSP26 , HSP150 , and SPI1 , via the RIM101 pathway, HOG pathway, Msn2/4 transcription factor regulation, or other possible routes (Ferguson et al., 2005; Ma & Liu, 2010; Martínez-Pastor et al., 1996; Mira et al., 2009; Schüller et al., 1994). HSP26 functions as a molecular chaperone (Haslbeck & Vierling, 2015), aiding in protein folding and stability, whereas HSP150acts as a cell wall stabilizer (Kapteyn et al., 1999). Meanwhile, SPI1 is involved in cell wall biogenesis, contributing to cell wall strength and protection (Simões et al., 2006). Therefore, these interactions provide a rational explanation for the correlations observed between glycolytic rate, cell cycle and growth, and stress tolerance, as these processes are regulated by PKA-mediated signaling pathways.
Surprisingly, in the F118L strain, none of the examined genes showed significant differential expression (Figure 3a). This finding suggests that the observed metabolic dynamics described earlier occurred in the F118L strain without the need for high-cell density conditioning. However, to corroborate this hypothesis, further analysis comparing the gene expression patterns in both strains at different cell densities is needed, as will be detailed in the subsequent cross-strain analysis.