Control of the galactose-to-glucose consumption ratio in co-fermentation using engineered Escherichia coli strains

Marine biomasses capable of fixing carbon dioxide have attracted attention as an alternative to fossil resources for fuel and chemical production. Although a simple co-fermentation of fermentable sugars, such as glucose and galactose, has been reported from marine biomass, no previous report has discussed the fine-control of the galactose-to-glucose consumption ratio in this context. Here, we sought to finely control the galactose-to-glucose consumption ratio in the co-fermentation of these sugars using engineered Escherichia coli strains. Toward this end, we constructed E. coli strains GR2, GR2P, and GR2PZ by knocking out galRS, galRS-pfkA, and galRS-pfkA-zwf, respectively, in parent strain W3110. We found that strains W3110, GR2, GR2P, and GR2PZ achieved 0.03, 0.09, 0.12, and 0.17 galactose-to-glucose consumption ratio (specific galactose consumption rate per specific glucose consumption rate), respectively, during co-fermentation. The ratio was further extended to 0.67 by integration of a brief process optimization for initial sugar ratio using GR2P strain. The strategy reported in this study will be helpful to expand our knowledge on the galactose utilization under glucose conditions.

Scientific RepoRtS | (2020) 10:12132 | https://doi.org/10.1038/s41598-020-69143-3 www.nature.com/scientificreports/ In this study, as a proof-of-concept, we tried to finely control the galactose-to-glucose consumption ratio during the co-fermentation of both sugars by engineering the sugar metabolism in E. coli on the basis of our recent study 19 , followed by a brief process optimization for sugar ratio and concentration.

Results
Galactose-to-glucose consumption ratio in E. coli GR2, which lacked the galactose operon repressors, during glucose consumption. In an effort to finely control the galactose consumption rate in the co-fermentation of glucose and galactose using E. coli, we first constructed the galRS mutant, E. coli strain GR2 ( Fig. 1 and Table 1), which was not able to produce the galactose operon repressors encoded from the galRS genes [21][22][23][24][25][26][27] . We expected that the resulting strain would consume galactose at a higher rate than the parent strain in the co-fermentation of glucose and galactose, although we anticipated that this rate would still be fairly low.
To determine the galactose consumption rate of the galRS mutant, we cultured GR2 cells in R/2 medium supplemented with 4 g/L glucose and 4 g/L galactose. Parent strain W3110 was cultured under the same conditions as a control. In the co-fermentation of glucose and galactose, E. coli strain GR2 had completely consumed the glucose at 12 h of culture, yielding a specific glucose consumption rate of 1.3629 g/gDCW/h ( Fig. 2a and Table 2). During this period, GR2 consumed 0.36 g/L galactose, for a specific consumption rate of 0.1262 g/gDCW/h. Under the same conditions, parent strain W3110 showed a similar level of glucose consumption ( Fig. 2b) but had a lower specific galactose consumption rate at 0.0373 g/gDCW/h ( Table 2). Strain GR2 yielded R gal/glu = 0.09, whereas parent strain W3110 yielded R gal/glu = 0.03 (Table 2).

Figure 1.
Control of the galactose-to-glucose consumption ratio in the co-fermentation of both sugars using the engineered E. coli. 'X' indicates gene knockout. The relevant genes and their encoded enzymes are as follows: galR, DNA-binding transcriptional repressor; galS, DNA-binding transcriptional isorepressor; galK, galactokinase; galT, galactose-1-phosphate uridylyltransferase; pfkA, 6-phosphofructokinase I; pfkB, 6-phosphofructokinase II; zwf, glucose-6-phosphate dehydrogenase. Abbreviations: Gal1P, galactose-1phosphate; G1P, glucose-1-phosphate; G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; FBP, fructose-1,6-phosphate; PYR, pyruvate; 6PG, 6-phosphate-gluconate; Ru5P, ribulose-5-phosphate; X5P, xylulose-5phosphate; E4P, erythorse-4-phosphate; KDPG, 2-keto-3-deoxy-phosphogluconate. www.nature.com/scientificreports/ Galactose-to-glucose consumption ratio in E. coli GR2P, which lacked the galactose operon repressors and 6-phosphofructokinase, during glucose consumption. Embden-Meyerhof-Parnas pathway (EMPP) was chosen as the next target to be engineered, because a metabolic intermediate glucose-6-phosphate is formed from both glucose and galactose. However, glucose-6-phosphate formation from both sugars are quietly differ from each other. To yield glucose-6-phosphate, glucose is catabolized by one-step reaction, while galactose is metabolized via Leior pathway 28 . Thus, it was expected that glucose and galactose consumption could be individually affected by modulation of EMPP. To further modulate R gal/glu , rather than overexpressing the galactose operon, we next knocked out the pfkA gene, which encodes ATP-dependent 6-phosphofructokinase, in E. coli strain GR2. As 6-phosphofructokinase is a key enzyme in the EMPP, we expected that the specific glucose consumption rate could be negatively affected by the lack of 6-phosphofructokinase in the resulting GR2P cells under the co-fermentation of glucose and galactose. In flask cultures, GR2P showed a maximum specific growth rate (µ max ) of 0.2993/h with a short lag time, which was rather lower than the 0.5653/h and 0.4469/h obtained from cultures of W3110 and GR2, respectively ( Fig. 3 and Table 2). This indicates that the additional knockout of the pfkA gene negatively affected cell growth. E. coli GR2P had completely consumed the glucose at 24 h of culture, yielding a specific glucose consumption rate of 0.4714 g/gDCW/h ( Fig. 3 and Table 2). Meanwhile, the specific galactose consumption rate was 0.0552 g/gDCW/h, yielding an R gal/glu of 0.12 (Table 2). Thus, our results indicate that the specific glucose consumption rate was negatively influenced by regulation of the EMPP in the engineered E. coli, whereas the specific galactose consumption rate was less sensitive to this manipulation.
Galactose-to-glucose consumption ratio in E. coli GR2PZ, which lacked the galactose operon repressors, 6-phosphofructokinase, and glucose-6-phosphate dehydrogenase, during glucose consumption. A previous report found that zwf-mutant E. coli, which lacked glucose-6-phosphate dehydrogenase, showed carbon flow through the oxidative pentose phosphate pathway (OPPP) 29 . To additionally modulate the glucose consumption rate of our engineered strain, we knocked out the zwf gene in E. coli GR2P.
As glucose-6-phosphate dehydrogenase is involved in converting glucose-6-phosphate to 6-phosphogluconate in the OPPP, we expected that the glucose consumption rate would be further reduced in the resulting GR2PZ strain, compared to the parent strain, GR2P. In the co-fermentation of glucose and galactose, GR2PZ showed a maximum specific growth rate of 0.2076 /h with a lag time, and thus had the lowest µ max among the strains used in this study ( Fig. 4 and Table 2). Furthermore, the sudden drop in cell mass was also observed at the very early stage (Fig. 4), which might be caused from the simultaneous knockout of the pfkA and zwf genes in E. coli. GR2PZ had completely consumed the glucose at 36 h of culture, yielding a specific glucose consumption rate of  www.nature.com/scientificreports/ 0.4419 g/gDCW/h ( Fig. 4 and Table 2). The specific galactose consumption rate was 0.0737 g/gDCW/h, for an R gal/glu of 0.17.

Galactose-to-glucose consumption ratio was further modulated by integration of a brief process optimization for initial sugar ratio.
We demonstrated the control of glucose-to-galactose consumption ratio during the co-fermentation of glucose and galactose by engineering central sugar metabolism. Next, one of the engineered E. coli strains was applied to the process optimization for initial sugar ratio and concentrations, to see if the galactose-to-glucose consumption ratio could be further modulated by different initial sugar ratio and concentration. Among the engineered three strains, GR2P was randomly selected, and cultured in R/2 medium supplemented with sugars at the different initial ratio and concentration, such as 4 g/L glucose and 2 g/L galactose (4glu-2gal); and 4 g/L glucose and 8 g/L galactose (4glu-8gal) (Fig. 5). The culture parameters using GR2P in R/2 medium supplemented 4 g/L glucose and 4 g/L galactose (4glu-4gal) were obtained from Fig. 3. All parameters obtained from the culture conditions, such as 4glu-2gal, 4glu-4gal, and 4glu-8gal were listed in Table 3. E. coli GR2P achieved an R gal/glu of 0.12, 0.12, and 0.67 under the condition of 4glu-2gal, 4glu-4gal, and 4glu-8gal, respectively (Table 3). Taken together, our results indicate that galactose-toglucose consumption ratio was further modulated by integration of a brief process optimization together with the modulation of sugar metabolism.

Discussion
Galactose produced by a marine biomass is a promising fermentable sugar for the production of biofuels and biochemicals. A recent report described an engineered E. coli strain capable of consuming galactose at the same rate as glucose 20 . In the same study, an R gal/glu of nearly 1.0 was achieved by overexpressing the galactose operon in a mutant E. coli lacking the galactose operon repressor 20 . However, the fine-tuned control of galactose consumption by E. coli has not previously been reported in a co-fermentation with glucose. In this study, we demonstrate that the galactose consumption rate may be finely controlled during co-fermentation with glucose by comparing our engineered E. coli strains and their parent W3110. Strains W3110, GR2, GR2P, and GR2PZ achieved R gal/glu values of 0.03, 0.09, 0.12, and 0.17, respectively, during the co-fermentation of these sugars. Meanwhile, R gal/glu of 0.67 was achieved by a brief process optimization using GR2P strain. We found that the galactose-to-glucose consumption ratio could be slightly modulated by disrupting the galactose operon repressors encoded by the galRS genes, and further modulated by regulating the central metabolic pathway. The galRS-mutant E. coli strain, GR2, had an R gal/glu ratio of 0.09, compared to the 0.03 ratio of parent strain W3110. Products of the galRS genes have been reported to negatively regulate galactose metabolism by binding upstream of the galactose operon 19,[30][31][32] . Thus, the inactivation of galactose operon repressors could logically increase the specific galactose consumption rate in E. coli while leaving the specific glucose consumption rate unchanged. Indeed, in the co-fermentation of glucose and galactose, GR2 and W3110 showed specific galactose consumption rates of 0.1262 g/gDCW/h and 0.0373 g/gDCW/h, respectively, while yielding specific glucose consumption rates of 1.3629 g/gDCW/h and 1.3706 g/gDCW/h (Fig. 2, Table 2).
In the co-fermentation of glucose and galactose using strains GR2P and GR2PZ, in which the central metabolic pathway was further modified by knockout of the pfkA and pfkA-zwf genes, respectively, we obtained R gal/glu ratios of 0.12 and 0.17, respectively. For GR2P and GR2PZ, the specific consumption rates for both sugars were decreased compared to that of parent strain GR2. Glucose consumption was more sensitive than galactose consumption to modulation of the central metabolic pathway: The specific glucose consumption rate decreased from 1.3629 g/gDCW/h (GR2) to 0.4714 g/gDCW/h (GR2P; 65% decrease) and 0.4419 g/gDCW/h (GR2PZ; 68% decrease), while the specific galactose consumption rate decreased from 0.1262 g/gDCW/h (GR2) to 0.0552 g/ gDCW/h (GR2P; 56%) and 0.0737 g/gDCW/h (GR2PZ; 41%). R gal/glu ratios of 0.12 and 0.17 were obtained for the co-fermentation of glucose and galactose by strains GR2P and GR2PZ, respectively (Fig. 4, Table 2). We do not yet know why glucose consumption was more sensitive than galactose consumption to the pfkA and pfkA-zwf mutations. However, this may relate to a previous report that deletion of the zwf gene altered the metabolism of the mutant cells by changing the redox balance [33][34][35] .
The engineered GR2P and GR2PZ strains utilized both sugars simultaneously in the presence of the major CCR. Under the galRS double-knockout condition, it seems that the repression of galactose metabolism by the  www.nature.com/scientificreports/ CCR was slightly leaky, which was supported by Table 2 and Fig. 2. The galRS were knocked-out in GR2P and GR2PZ strains also, which means that two strains were leaky in CCR, too. Thus, we believe that the simultaneous utilization of both sugars was not resulted from the complete deregulation of CCR. However, the simultaneous utilization of both sugars in GR2P and GR2PZ strains could be explained by dramatic decrease of specific glucose consumption rate compared to the parent GR2 strain (Table 2), which was more sensitive compared to the decrease of specific galactose consumption.
In conclusion, we herein report the regulation of the galactose-to-glucose consumption ratio by engineering E. coli strains for the co-fermentation of these sugars. We achieved R gal/glu values of 0.03, 0.09, 0.12, and 0.17 in the co-fermentation of glucose and galactose using E. coli strains W3110, GR2, GR2P, and GR2PZ, respectively. The R gal/glu value of 0.67 was achieved by a brief process optimization for initial sugar ratio using GR2P strain. In the engineered E. coli strains, the galactose-to-glucose consumption ratio could be changed by activation of the galactose operon and modulation of central metabolic pathways (EMPP and OPPP) in the co-fermentation of both sugars. The strategy reported in this study can help regulate the galactose-to-glucose consumption ratio in a microbial co-fermentation of both sugars from galactose-rich biomass. Furthermore, the repression of carbohydrate metabolism might be applicable to enhance acetyl-CoA pool from another carbon source which is uncommon to E. coli, such as polyethylene.

Materials and methods
Microbial strains and plasmids. Details on the utilized engineered E. coli strains and their parent strain, K-12 W3110, are listed in Table 1. Plasmid pCW611, which harbored the genes encoding lambda red recombinase and Cre-recombinase, was used for gene knockout in E. coli 36 . Plasmid pMtrc9, which contained the lox66cat-lox77 cassette, was used as a PCR template to amplify the DNA fragment for target gene knockout 37 . Details on the utilized plasmids are presented in Table 1.
Media and culture conditions. Escherichia coli cells were cultured in R/2 medium supplemented with the indicated carbon sources. R/2 medium (pH 6.8) contained 2 g (NH 4 ) 2 HPO 4 , 6.75 g KH 2 PO 4 , 0.7 g MgSO 4 ·7H 2 O, 0.85 g citric acid, and 5 mL trace metal solution per liter 38  To prepare the seed culture, a colony grown on an LB agar plate was inoculated into 25-mL test tubes containing 5 mL R/2 medium. The test tubes were incubated in a shaking incubator (IST-4075, JEIOTECH, Korea) at 37 °C and 200 rpm. The next day, 2.2 mL of seed culture was transferred to each 500-mL flask containing 220 mL R/2 medium. For co-fermentation experiments, 3 g/L glucose and 3 g/L galactose were supplemented in the medium as carbon sources.
Determination of the galactose-to-glucose consumption ratio. To determine the galactose-toglucose consumption ratio during glucose consumption, E. coli cells were cultured in 500-mL flasks containing R/2 medium supplemented with 4 g/L glucose and 4 g/L galactose, at 37 °C and 200 rpm. One milliliter of culture broth was taken at 0, 3, 6, 9, 12, 16, 24, 28, 32, 36, and 48 h and used to determine the residual glucose and galactose concentrations (see below for analytical methods). Another 1 mL of broth was taken at the same time and used to determine the dry cell weight. From these analyses, the specific glucose consumption rate (R glu ) was determined, while the specific galactose consumption rate (R gal ) was determined during glucose consumption. The galactose-to-glucose consumption ratio, R gal/glu , was determined. All flask fermentations were independently performed in triplicate.
Analytic methods. Cell growth was monitored by determining the absorbance at 600 nm using a spectrophotometer (HIACHI, JP/U-1900, Tokyo, Japan). For determination of the glucose and galactose concentrations, culture broth supernatants were prepared by centrifugation followed by filtration using a 0.22-μm pore syringe filter. Samples were applied to high performance liquid chromatography (HPLC; 1100 series, Agilent Technologies, Santa Clara, CA, USA) equipped with a refractive index detector (RID; RI-101, Shodex, Klokker-