Development of efficient D-lactic acid-producing Saccharomyces cerevisiae strains by evolutionary and rational metabolic engineering

Abstract

There is an increasing demand for microbial production of optically pure lactic acid (LA) as a monomer for biodegradable poly lactic acid (PLA). Saccharomyces cerevisiae, having high acid tolerance, has emerged as a promising LA-producing host. In this dissertation, efficient D-LA-producing strains of S. cerevisiae were developed using rational metabolic engineering and adaptive evolution. Firstly, to generate D-LA-producing strain, highly stereospecific D-lactate dehydrogenase gene (ldhA, designated Lm.ldhA) from Leuconostoc mesenteroides subsp. mesenteroides ATCC 8293 was selected and expressed in S. cerevisiae lacking natural LA production activity. To prevent D-LA utilization upon glucose depletion, DLD1 encoding D-lactate dehydrogenase and JEN1 encoding monocarboxylate transporter were disrupted in combination. Ethanol formation was reduced by deleting PDC1 and ADH1 genes encoding major pyruvate decarboxylase and alcohol dehydrogenase, respectively. In addition, glycerol production was eliminated by deleting GPD1 and GPD2 genes encoding glycerol-3-phosphate dehydrogenase, resulting in a gradual increase in D-LA production level up to 12.9 g/L with a yield of 0.65 g/g glucose. Next, in order to overcome growth inhibition by LA in the engineered LA-producing strain adaptive evolution was carried out by gradual increase in LA concentrations in the culture medium. As a result, the evolved strain showed higher glucose consumption rate with an improved D-LA production level compared with the unevolved strain. By additional genome-integration of the Lm.ldhA gene into the evolved strain, a D-LA production level was further improved up to 38.3 g/L in acidic fermentation condition without pH control. In neutralization condition, this strain produced 118.6 g/L D-LA, showing an increased yield (0.79 g/g glucose). Through whole genome sequencing analysis, mutations of five genes, including two nonsense mutations (GSF2 and SYN8) and three point mutations (STM1, SIF2, and BUD27) were detected in the evolved strain. Deletion of GSF2 in the unevolved strain led to improved growth rate and glucose consumption ability, resulting in D-LA production level comparable to that of the evolved strain. It was also demonstrated that deletion of STM1 or SIF2 increased resistance to LA in S. cerevisiae as well as deletion of SYN8, which is already known to increase LA tolerance. Lastly, D-LA production level was further improved by integrating an additional copy of HAA1 into the evolved strain. This Haa1-overexpressing strain consumed 62.2 g/L glucose and produced 48.9 g/L D-LA with a yield of 0.79 g/g glucose under acidic fermentation. In a flask fed-batch fermentation under neutralizing conditions, this final strain showed the highest productivity of 2.2 g/(L∙h). In addition, it was demonstrated that Haa1 is phosphorylated by CK2 in S. cerevisiae. Although the function of CK2-dependent phosphorylation should be further elucidated, this result suggests that CK2 might be implicated in LA stress adaptation through regulating Haa1.