Production of R- and S-1,2-propanediol in engineered Lactococcus lactis

1,2-propanediol (1,2-PDO) is a versatile chemical used in multiple manufacturing processes. To date, some engineered and non-engineered microbes, such as Escherichia coli, Lactobacillus buchneri, and Clostridium thermosaccharolyticum, have been used to produce 1,2-PDO. In this study, we demonstrated the production of R- and S-1,2-PDO using engineered Lactococcus lactis. The L- and D-lactic acid-producing L. lactis strains NZ9000 and AH1 were transformed with the plasmid pNZ8048-ppy harboring pct, pduP, and yahK genes for 1,2-PDO biosynthesis, resulting in L. lactis LL1 and LL2, respectively. These engineered L. lactis produced S- and R-1,2-PDO at concentrations of 0.69 and 0.50 g/L with 94.4 and 78.0% ee optical purities, respectively, from 1% glucose after 72 h of cultivation. Both 1% mannitol and 1% gluconate were added instead of glucose to the culture of L. lactis LL1 to supply NADH and NADPH to the 1,2-PDO production pathway, resulting in 75% enhancement of S-1,2-PDO production. Production of S-1,2-PDO from 5% mannitol and 5% gluconate was demonstrated using L. lactis LL1 with a pH–stat approach. This resulted in S-1,2-PDO production at a concentration of 1.88 g/L after 96 h of cultivation. To our knowledge, this is the first report on the production of R- and S-1,2-PDO using engineered lactic acid bacteria.

Among diols, 1,2-PDO is one of the most versatile chemicals with two optical isomers, R-1,2-PDO and fermenter-controlled cultivation conditions (Niu et al. 2019). 1,2-PDO can be produced in engineered microbes via a pathway in which glucose is first converted to D-and L-lactate, followed by the synthesis of R-and S-1,2-PDO from D-and L-lactate, respectively, in which cofactors such as acetyl-CoA, NADH, and NADPH are required (Altras and Cameron 1999;Saxena et al. 2010;Niu and Guo 2015;Niu et al. 2019). This suggests that microbes producing the 1,2-PDO precursor lactate at a high titer are suitable as production hosts for 1,2-PDO production, and that R-and S-1,2-PDO can be separately produced in D-and L-lactic acid producers, respectively. In fact, Niu et al. demonstrated the production of R-and S-1,2-PDO separately using D-and L-lactic acid-producing engineered E. coli, respectively (Niu et al. 2019).
Based on their glycolytic metabolism, LAB can be classified into two groups: homo-and hetero-fermentative LAB (Axelsson 2004;Bintsis 2018). Homofermentative LAB, such as Lactococcus spp., theoretically yield 2 mol of lactate from 1 mol of consumed glucose. LAB can also be divided into three groups: D-lactic acid producers, L-lactic acid producers, and both-isomer producers, which are determined by the catalytic properties of LDH encoded in their genomes (Gao et al. 2012;Rahman et al. 2013;Ghaffer et al. 2014;Eş et al. 2018). These results suggest that the use of homofermentative LAB as production hosts is reasonable for 1,2-PDO production, and that D-and L-lactic acid-producing LAB can be used to separate R-and S-1,2-PDO, respectively. Kuipers et al. constructed a recombinant host for the nisin-controlled gene expression (NICE) system, L. lactis NZ9000, by incorporating the two-component regulatory gene nisRK into the genome of L. lactis MG1363 (Kuipers et al. 1998). Since L. lactis NZ9000 has been used as a superior host for protein expression, this strain would be a suitable host for 1,2-PDO production. Recently, Aso et al. constructed a L. lactis NZ9000 derivative by replacing its major intrinsic L-LDH gene with a heterologous D-LDH gene from Lactobacillus delbrueckii subsp. lactis JCM 1107, resulting in a D-lactic acid producer, L. lactis AH1 (Aso et al. 2019). Using the AH1 strain, D-lactate was produced from starch directly by expression of a heterologous α-amylase gene from Streptococcus bovis NRIC 1535. From this demonstration, it is presumed that Rand S-1,2-PDO can be produced by the expression of pct, pduP, and yahK in L. lactis NZ9000 and AH1, respectively, using the NICE system.
Therefore, the present study demonstrated the production of R-and S-1,2-PDO using engineered L. lactis NZ9000 and AH1, respectively, through an exogenous 1,2-PDO production pathway, according to a previously reported demonstration using engineered E. coli (Niu et al. 2019) (Fig. 1). Additionally, it has been reported that the metabolism of mannitol and gluconate in L. lactis results in enhancement of intracellular amounts of NADH and NADPH, respectively (Ramos et al. 2001;Neves et al. 2002;Wegmann et al. 2007;Linares et al. 2010). This suggests that addition of mannitol and gluconate improves 1,2-PDO production. To prove this, the effect of adding mannitol and gluconate to cultures of engineered L. lactis on 1,2-PDO production was investigated.
To characterize production of L-lactate and S-1,2-PDO, resting cells of L. lactis LL1 were prepared as follows: recombinant L. lactis was cultivated in GM17 medium for 18 h at 30 °C. After centrifugation of the culture for 10 min, the culture supernatant was removed. Collected cells were resuspended in 1 mL of 100 mM sodium phosphate buffer (pH 7.0) containing 1% glucose, 1% sodium gluconate, and/or 1% mannitol after washing three times with distilled water. Resting cells were incubated for 72 h at 30 °C, and then the supernatant was subjected to analysis.
Fermentation controlled with a pH-stat was performed in a 2 L jar fermenter M-1000B (Tokyo Rikakikai Co., Ltd., Tokyo, Japan) with a working volume of 1 L at an agitation speed of 80 rpm at 30 °C without aeration under microaerobic conditions, and the pH of the culture was maintained at 6.8 by automatic addition of 2 M NaOH using a peristaltic pump. The concentration of dissolved oxygen (DO) in the culture was monitored using a field controller, model mk-750 DO (Automatic System Research Co., Ltd., Saytama, Japan).

Analytical methods
Concentrations of lactate, 1,2-PDO, glucose, gluconate, mannitol, acetate, and ethanol in the culture supernatants were measured using a Prominence HPLC system (Shimadzu, Kyoto, Japan) equipped with an Aminex HPX-87H (Bio-Rad, CA, USA) and a refractive index (RI) detector (Shimadzu). These analytes were eluted using a 5 mM sulfuric acid solution at a flow rate of 0.4 mL/ min at 30 °C. The chirality of 1,2-PDO produced by L. lactis was determined using a Prominence HPLC system (Shimadzu) equipped with a CHIRAL ART Cellulose-C S-5 μm column (YMC Co., Ltd., Kyoto, Japan) and monitored using an RI detector (GL Science, Tokyo, Japan) according to a previously described protocol (Niu and Guo 2015). The mobile phase (99% hexane and 1% isopropanol) was eluted at a flow rate of 0.5 mL/min at 25 °C. The chirality of lactate produced by L. lactis was analyzed using a Prominence HPLC system (Shimadzu) equipped with an MCI GEL CRS10W column (Mitsubishi Chemical, Tokyo, Japan) and monitored at 254 nm using 1 mM copper (II) sulfate solution as the mobile phase at a 0.5 mL/min flow rate (Aso et al., 2019). The specific growth rate (μ) was calculated as the slope of the regression line, from a plot between ln (X/X 0 ) and time (t) during the exponential growth period, where X (OD 600 ) and X 0 (OD 600 ) are the cell concentrations at t (h) and at the beginning of the exponential phase, respectively.

1,2-PDO production in engineered L. lactis
Expression of proteins encoded by the pct, pduP, and yahK genes in engineered L. lactis was confirmed by SDS-PAGE analysis (data not shown). This showed that the NICE system induced expression of 1,2-PDO biosynthetic genes in engineered L. lactis. To check 1,2-PDO production in the engineered L. lactis, resting cells of L. lactis LL1 and LL2 were incubated for 72 h in a buffer containing 1% glucose (Table 1). After 72 h of incubation, L. lactis LL1 produced 0.62 ± 0.00 g/L of L-lactate and 0.69 ± 0.01 g/L of S-1,2-PDO, and L. lactis LL2 produced 0.52 ± 0.04 g/L of D-lactate and 0.73 ± 0.06 g/L of R-1,2-PDO. This showed that the introduced genes pct, pduP, and yahK function in both L. lactis NZ9000 and AH1. The optical purities of L-lactate and S-1,2-PDO produced by L. lactis LL1 were 96.8% and 94.4% ee, respectively, and the optical purities of D-lactate and R-1,2-PDO produced by L. lactis LL2 were 85.7% and 78.0% ee, respectively. This showed that R-and S-1,2-PDO can be separately produced in L. lactis LL1 and LL2, respectively.
For characterization of 1,2-PDO production, L. lactis LL1 and LL2 were cultivated for 72 h in M17 medium supplemented with 1% glucose in a flask. After 72 h of cultivation, L. lactis LL1 produced 0.33 ± 0.00 g/L of L-lactate and 0.69 ± 0.01 g/L of S-1,2-PDO. L. lactis LL2 produced 0.35 ± 0.05 g/L of D-lactate and 0.50 ± 0.02 g/L of R-1,2-PDO (Fig. 2). The production profiles were similar to those obtained using resting cells. The 1,2-PDO yield from glucose in L. lactis LL1 was 1.4 times as high as in L. lactis LL2 ( Table 2). The specific growth rates μ of L. lactis LL1 and LL2 were 1.02 ± 0.03 h −1 and 0.99 ± 0.01 h −1 , respectively. There was no significant difference between OD 600 at 72 h of cultivation and μ of L. lactis LL1 and LL2 (P > 0.05), indicating that the production of the optical isomers did not affect cell growth. Acetate and ethanol were detected as the main by-products in the cultures, but these production levels were negligible compared to those of 1,2-PDO ( Table 2).

Supplementation of NADH and NADPH for 1,2-PDO production
To investigate the effect of adding gluconate and mannitol on 1,2-PDO production, resting cells of L. lactis LL1 were incubated in a buffer supplemented with 1% glucose, 1% mannitol, and/or 1% gluconate. The production levels of L-lactate and S-1,2-PDO from 1% glucose were 0.66 ± 0.01 g/L and 0.69 ± 0.01 g/L, respectively Table 1 Production profiles of D-and L-lactate and R-and S-1,2-PDO in the engineered L. lactis after 72 h of incubation with 1% glucose a Yield (%, g-product/g-glucose × 100)
These results indicate that NADH and NADPH were supplied through the metabolism of mannitol and gluconate, and consequently, 1,2-PDO production was improved.
To characterize S-1,2-PDO production with gluconate and mannitol, L. lactis LL1 was cultivated with 1% mannitol and 1% gluconate for 72 h in a flask. The medium was supplemented with 0.3 g/L of acetate to enhance the intracellular level of acetyl-CoA in L. lactis (Puvendran and Jayaraman 2019). The initial pH of the culture was 7.0, but at 24 h of cultivation was 6.3. This was caused by L-lactate production, and resulted in 3.0 g/L of mannitol and 4.6 g/L of gluconate remaining in culture after cultivation. To promote consumption of mannitol and gluconate by L. lactis LL1, the pH of the culture was adjusted to 7.0 every 24 h by adding 1 M MOPS (pH 7.0). After 72 h of cultivation, L. lactis LL1 produced 0.57 ± 0.08 g/L of L-lactate and 1.21 ± 0.01 g/L of S-1,2-PDO from 1% mannitol and 1% gluconate (Fig. 4). The production titers of L-lactate and S-1,2-PDO were 1.7 times as high as those obtained with 1% glucose. The OD 600 after 30 h of cultivation was 6.61 ± 0.06, and μ was 0.24 ± 0.00 h −1 . These properties were comparable to those observed when cells were cultured with 1% glucose. This indicates that supplementation with mannitol and gluconate had similar effects on cell growth. Consumption rates of gluconate Table 2 Production profiles of metabolites in the engineered L. lactis after 72 h of cultivation with 1% glucose a Yield (%, g/g-glucose × 100)

Discussion
Since the biosynthetic pathway for the production of 1,2-PDO has been reported, bioproduction of this compound has been attempted by introducing the biosynthetic pathway in various microorganisms including E. coli (Cameron and Cooney 1986, Niu and Guo 2015, Niu et al. 2019. In this study, we demonstrated production of R-and S-1,2-PDO using engineered L. lactis constructed from L-and D-lactic acid-producing LAB, L. lactis NZ9000, and AH1, respectively, by expression of 1,2-PDO biosynthetic genes with glucose, mannitol, and/ or gluconate (Fig. 1). The advantage of using LAB is that lactic acid is the main organic acid required for LAB to produce the desired product, 1,2-PDO.
The resting cells of L. lactis LL1 produced L-lactate and S-1,2-PDO at concentrations of 0.62 g/L and 0.69 g/L, respectively, with negligible production of other isomers (Table 1). On the other hand, L. lactis LL2 produced an isomer each of lactate and 1,2-PDO (Table 1). This is because L. lactis AH1, which is the parent strain of L. lactis LL2, has been reported to produce D-lactate at an optical purity of 87.5% ee (Aso et al. 2019). We obtained a similar result, in that the optical purity of D-lactate was 85.7% ee when L. lactis LL2 was cultivated in this study. These results indicate that the optical purities of R-and S-1,2-PDO produced by the engineered L. lactis were determined in accordance with those of the D-and L-lactate produced. The low D-lactate optical purity of L. lactis LL2 is due to the fact that L. lactis AH1 has lldh homologue genes (ldhB, ldhX, and hicD) in addition to lldh (Aso et al. 2019). The production of lactate, a primary metabolite, is known to be associated with cell growth. Actually, however, the resting cells of L. lactis LL1 and LL2 produced lactate and 1,2-PDO from glucose. This is may be due to that L. lactis MG1363, which is the parent strain of L. lactis LL1 and LL2, maintains enzymatic activity related to the glycolysis during all phases of cell growth (Palmfeldt et al. 2004). ATP was probably pooled in the resting cells. Because, it has been reported that ATP in the resting cells of L. lactis was increased compared to the growing cells (Palmfeldt et al. 2004).
Production of R-and S-1,2-PDO in the growing cells of L. lactis LL1 and LL2 with 1% glucose at 72 h of cultivation was characterized. Production of 1,2-PDO by L. lactis LL1 was 1.4 times as high as that produced by L. lactis LL2 (Fig. 2). Interestingly, production of 1,2-PDO by L. lactis LL1 was more than 40 times as high as that produced by a pflA-deficient E. coli BW25113, harboring the same set of 1,2-PDO biosynthetic genes, from 1% glucose, as previously reported (Sato et al. 2020). The engineered E. coli produced heterofermentatively D-lactate along with acetate and ethanol at non-negligible concentrations, leading to the decrease of the yield  of D-lactate. On the other hand, engineered E. coli that produced the maximum 1,2-PDO titer, showed 1.05 g/L acetate, and 0.13 g/L ethanol production from 1% glucose (Niu et al. 2019). In contrast, homofermentative LAB, including L. lactis, produce lactate with by-products at negligible concentrations. Therefore, 1,2-PDO production by engineered L. lactis significantly increased compared to that produced by the engineered the engineered pflA-deficient E. coli BW25113. L. lactis effectively produces lactate from glucose under microaerobic conditions (Papagianni et al. 2007). L. lactis is cultured under aerobic conditions, the lactate production is low due to NADH oxidase. This results in the production of acetate (Neves et al. 2002). On the other hand, 1 mol each of acetyl-CoA, NADH, and NADPH are required for 1 mol of 1,2-PDO production. To supply acetyl-CoA, microaerobic conditions were demonstrated in this study. Although this study demonstrated to produce 1,2-PDO under microaerobic condition, the lactate titer of L. lactis LL1 and LL2 was lower than that of L. lactis NZ9000 and AH1 which produced 7.0 g/L and 6.6 g/L of lactate from 1% glucose, respectively (Aso et al. 2019). This result indicated that the overexpression of pct, pduP, and yahK induced the conversion of lactate to 1,2-PDO. Overexpression of phosphofructokinasegene (pfkA) from Aspergillus niger in L. lactis LM0230 improved glucose uptake and enhanced lactate production yields (Papagianni and Avramidis 2011). Therefore, the overexpression of pfkA gene in L. lactis LL1 and LL2 might be improved lactate and 1,2-PDO productions. It is known that additional NADH and NADPH are generated in L. lactis through the metabolism of gluconate and mannitol, and that mannitol-1-phosphate dehydrogenase and 6-phospogluconate dehydrogenase generate NADH and NADPH, respectively (Ramos et al. 2001;Neves et al. 2002;Wegmann et al. 2007;Linares et al. 2010;Zhao et al., 2017). To improve 1,2-PDO production by supplying these cofactors to the 1,2-PDO biosynthetic pathway, effects of adding glucose, mannitol, and/or gluconate on 1,2-PDO and lactate production using the resting cells of L. lactis LL1 were investigated. Although addition of either 1% mannitol or gluconate led to decreased S-1,2-PDO production compared to addition of 1% glucose, addition of both compounds resulted in the highest S-1,2-PDO production (Fig. 3). Mannitol is intracellularly converted to mannitol-1-phosphate, followed by the formation of fructose-6-phosphate. Fructose-6-phosphate is not metabolized via the pentose phosphate pathway, in which NADPH is generated. Therefore, the addition of 1% mannitol alone led to the lowest S-1,2-PDO production. It was suggested that the addition of gluconate alone does not contribute to the cellular redox balance and consequently results in a decrease in S-1,2-PDO production. Interestingly, the production of L-lactate was significantly enhanced by the addition of 1% gluconate. This may be because sodium ions derived from gluconate neutralize the L-lactate produced, and then the resting cells produce a large amount of L-lactate (1.92 g/L).
Production of S-1,2-PDO was characterized by cultivation of L. lactis LL1 with M17 medium supplemented with 1% gluconate and 1% mannitol in flask cultivation. The use of 1% gluconate and 1% mannitol led to 75% enhancement of S-1,2-PDO production than did the use of 1% glucose (Fig. 4). In the metabolic pathway in the engineered L. lactis, 1 mol of NADH are synthesized from 1 mol of mannitol whereas 1 mol of NADH and 1 mol of NADPH are synthesized from 1 mol of gluconate (Niu et al. 2019) (Fig. 1). In the following step, the synthetic pathway for 1,2-PDO production consumes the equivalent numbers of NADH and NADPH. This ensures the redox balance for the whole process. This proves that 1,2-PDO production is enhanced by an additional supply of NADH and NADPH derived from mannitol and gluconate in the engineered L. lactis. Niu and Guo have also mentioned the necessity of supplying NADH and NADPH for 1,2-PDO production in recombinant E. coli (Niu and Guo 2015;Niu et al. 2019).
Generally, lactate production by LAB can be improved by maintaining the pH of the cultures during fermentation (Andersen et al., 2009;Carvaiho et al. 2013). Therefore, jar fermenter cultivation with a pH-stat for S-1,2-PDO production was performed with 5% gluconate and 5% mannitol. As a result, the production of lactate was improved, but the production of 1,2-PDO did not significantly increase (Fig. 5). In flask cultivation with 1% mannitol and 1% gluconate, there is no significant difference between the consumption rates of gluconate and mannitol (gluconate, 0.19 g/L/h; mannitol, 0.18 g/L/h). However, the consumption rate of gluconate was 1.8 times as high as that of mannitol in the jar-fermenter cultivation with 5% mannitol and 5% gluconate during 54 h of cultivation (gluconate, 0.91 g/L/h; mannitol, 0.51 g/L/h). Neves et al. reported that the NADH burden caused by mannitol-1-phosphate dehydrogenase inhibits mannitol metabolism in L. lactis MG1363 cultivated with mannitol (Neves et al. 2002). On the other hand, gluconate seems to be metabolized through the pentose phosphate pathway and glycolysis without such obstruction because NADPH generated from gluconate is smoothly consumed for the synthesis of cell biomass, especially during the exponential growth phase. This suggests that gluconate is more easily metabolized than mannitol when added at a higher concentration (5%), and consequently, gluconate is consumed first. The production rate of S-1,2-PDO was nearly constant after 48 h of cultivation. This