Engineering Escherichia coli for improved ethanol production from gluconate
Abstract
We report on engineering Escherichia coli to produce ethanol at high yield from gluconic acid (gluconate). Knocking out genes encoding for the competing pathways (l-lactate dehydrogenase and pyruvate formate lyase A) in E. coli KO11 eliminated lactate production, lowered the carbon flow toward acetate production, and improved the ethanol yield from 87.5% to 97.5% of the theoretical maximum, while the growth rate of the mutant strain was about 70% of the wild type. The corresponding genetic modifications led to a small improvement of ethanol yield from 101.5% to 106.0% on glucose. Deletion of the pyruvate dehydrogenase gene (pdh) alone improved the ethanol yield from 87.5% to 90.4% when gluconate was a substrate. The growth rate of the mutant strain was identical to that of the wild type. The corresponding genetic modification led to no improvements on ethanol yield on glucose.
1. Introduction
As the global demand for energy continues to rise and the consequences of reliance on fossil fuels intensify, biofuels attract more attention as a clean, renewable alternative energy source. Cellulosic biomass is a promising resource for biofuel production due to its low cost, widespread abundance, and distinction from food crops. The hydrolysate of cellulosic biomass contains both hexose and pentose sugars, as well as some sugar acids, all of which can potentially be converted to biofuels and chemicals by microbial fermentation.
Ethanol is one of the leading biofuels which is widely pursued in industry. While naturally ethanologenic organisms such as Saccharomyces cerevisiae and Zymomonas mobilis have historically been used in commercial ethanol operations, they are limited in their ability to utilize pentose sugars , which constitute up to 90% of the monomers in hemicellulose in cellulosic biomass. In contrast, Escherichia coli lacks the native ability for homoethanol production, but can metabolize a wide variety of sugars, an important feature in obtaining economical yields from lignocellulose hydrolysates. Recombinant ethanol production pathways have been introduced to E. coli, resulting in conversion of various sugars to ethanol at efficiencies approaching the theoretical maximum . The recombinant E. coli has become a very attractive host for ethanol production from cellulosic biomass. In addition to sugars, E. coli is able to metabolize sugar aldonic acids such as glucuronic acid, which also exist in the hemicellulose hydrolysate . In a recent study, we proposed an alternative route for ethanol production from cellulosic biomass in which cellobionate instead of sugars is produced as the reactive intermediate. Both of the hydrolysis products from cellobionate, glucose and gluconate, can be used efficiently by E. coli
E. coli KO11 was constructed in 1990 by integrating the Zymmonas mobilis ethanol production pathway (PET operon), which includes pyruvate decarboxylase (pdc) and alcohol dehydrogenase II (adhII), into the E. coli W (ATCC 9637) chromosome (originally thought to be E. coli B) at the pyruvate formate lyase (pflB) locus. In addition, the fumarase reductase (frd) gene was also knocked out in E. coli KO11 to prevent succinate production. However, it was reported that a single crossover event during homologous recombination resulted in a second functional pflB gene downstream of the Z. mobilis pathway. Recently, it was discovered through optical mapping and sequencing that the KO11 genome contains approximately 25 tandem repeats of the pdc-adhII-cat ethanol cassette due to IS10-promoted rearrangements which were introduced when knocking out the frd gene . Hence, E. coli KO11 contains two primary pathways that compete with the Z. mobilis ethanol pathway for pyruvate under anaerobic conditions: one produces acetate through acetyl-CoA from PFL, the other produces lactic acid by lactate dehydrogenase (LDH). An additional pathway utilizing the pyruvate dehydrogenase complex (PDH) connects the glycolysis reactions to the tricarboxylic acid cycle. It was previously thought that this pathway is only active under strictly aerobic conditions, but it has been recently demonstrated that there is some flux through this pathway under anaerobic conditions to support redox balance and provide reducing equivalents .
While ethanol yields of or near 100% can be achieved with E. coli KO11 on glucose, an ethanol yield of approximately 87.5% of the theoretical maximum is obtained when gluconate is used as the sole carbon source, although gluconate utilization is faster than glucose utilization. The formation of lactic and acetic acid byproducts reduces the overall yield. In this study, we investigate whether the ethanol yield from gluconate can be improved by knocking out genes for competing pyruvate dissimilation pathways in E. coli KO11, without deleteriously affecting the yield from glucose, and how the modifications affect the strain performance.
2. Materials and methods
2.1. Strain construction
E. coli KO11 was obtained from the ATCC (ATCC 55124). E. coli KO11 was transformed with the pKD46 plasmid to promote homologous recombination via the Red recombinase system (γ, β, and exo). The pKD4 plasmid, carrying a kanamycin resistance gene flanked by FRT sites, was used to generate the PCR products by using primers with 44–45 nt extensions that are homologous to the regions upstream and downstream of the target gene. Knockout cassettes were subsequently excised utilizing the pCP20 plasmid containing flipase . It is important to note that pflA was knocked out to prevent pyruvate formate lyase activity instead ofpflB. Because of the IS-10 promoted tandem repeats of the PET insertion cassette, knocking out pflB by homologous recombination is rather difficult because of the multiple copies in close proximity to the PET operon. Alternatively, pflA, which encodes for PFL activating enzyme I, is essential for PFL functionality and the deletion of pflA led to similar extent of loss of intracellular PFL activity as the pflBdeletion. Because pflA is present in a single copy located downstream of the PET operon, it could be knocked out without the risk of altering or removing the PET operon and confounding the results. The strains used in this study are listed in Table 1.
-
Table 1. Strains and plasmids used in this study.
E. coli KO11 was obtained from the ATCC (ATCC 55124). E. coli KO11 was transformed with the pKD46 plasmid to promote homologous recombination via the Red recombinase system (γ, β, and exo). The pKD4 plasmid, carrying a kanamycin resistance gene flanked by FRT sites, was used to generate the PCR products by using primers with 44–45 nt extensions that are homologous to the regions upstream and downstream of the target gene. Knockout cassettes were subsequently excised utilizing the pCP20 plasmid containing flipase . It is important to note that pflA was knocked out to prevent pyruvate formate lyase activity instead ofpflB. Because of the IS-10 promoted tandem repeats of the PET insertion cassette, knocking out pflB by homologous recombination is rather difficult because of the multiple copies in close proximity to the PET operon. Alternatively, pflA, which encodes for PFL activating enzyme I, is essential for PFL functionality and the deletion of pflA led to similar extent of loss of intracellular PFL activity as the pflBdeletion. Because pflA is present in a single copy located downstream of the PET operon, it could be knocked out without the risk of altering or removing the PET operon and confounding the results. The strains used in this study are listed in Table 1.
- Table 1. Strains and plasmids used in this study.
2.2. Characterization of the mutants in batch cultures
Pre-culture flasks containing 10 mL of Luria Bertani (LB) media and 2% glucose were inoculated with a single colony and incubated in a rotary shaker for 8 h at 37 °C and 200 rpm. 1 mL of the pre-culture was transferred to a 200 mL seed serum bottle containing 100 mL of LB with 2% glucose. The seed culture was incubated for 16 h at 37 °C and 200 rpm. The main fermentation serum bottles contained a 100 mL working volume of LB with 90 mM of either glucose of sodium gluconate. The serum bottles were pH adjusted to 6.5 with hydrochloric acid. Bottles were purged with argon to obtain anaerobic conditions. All fermentations were run in a minimum of triplicates on a rotary shaker at 37 °C and 200 rpm. Samples were taken at various time internals to analyze the concentrations of metabolites and cell mass.
Growth rate experiments were performed in an identical manner to the fermentation experiments, but with increased sampling during the exponential phase of growth to ensure linear logarithmic data for the maximum growth rate calculation. The maximal growth rate, μmax, was determined by plotting the natural logarithm of the optical density versus time and calculating the slope of the linear data corresponding to the exponential phase of growth.
Pre-culture flasks containing 10 mL of Luria Bertani (LB) media and 2% glucose were inoculated with a single colony and incubated in a rotary shaker for 8 h at 37 °C and 200 rpm. 1 mL of the pre-culture was transferred to a 200 mL seed serum bottle containing 100 mL of LB with 2% glucose. The seed culture was incubated for 16 h at 37 °C and 200 rpm. The main fermentation serum bottles contained a 100 mL working volume of LB with 90 mM of either glucose of sodium gluconate. The serum bottles were pH adjusted to 6.5 with hydrochloric acid. Bottles were purged with argon to obtain anaerobic conditions. All fermentations were run in a minimum of triplicates on a rotary shaker at 37 °C and 200 rpm. Samples were taken at various time internals to analyze the concentrations of metabolites and cell mass.
Growth rate experiments were performed in an identical manner to the fermentation experiments, but with increased sampling during the exponential phase of growth to ensure linear logarithmic data for the maximum growth rate calculation. The maximal growth rate, μmax, was determined by plotting the natural logarithm of the optical density versus time and calculating the slope of the linear data corresponding to the exponential phase of growth.
2.3. Sample analysis
The concentrations of sugars and organic acids in the samples were determined by HPLC with an Aminex HPX-87H column and a RI detector. The mobile phase used was 5 mM sulfuric acid with a flow rate of 0.6 mL/min.
The concentrations of sugars and organic acids in the samples were determined by HPLC with an Aminex HPX-87H column and a RI detector. The mobile phase used was 5 mM sulfuric acid with a flow rate of 0.6 mL/min.
2.4. Calculations of the yields and fluxes
Yield calculations were based on the amount of product produced versus amount of substrate consumed, compared to the theoretical maximum based on the stoichiometry. The flux was calculated using the final concentration of the product (mM) at the end of the batch culture divided by the concentration of substrate (glucose or gluconate) consumed.
Yield calculations were based on the amount of product produced versus amount of substrate consumed, compared to the theoretical maximum based on the stoichiometry. The flux was calculated using the final concentration of the product (mM) at the end of the batch culture divided by the concentration of substrate (glucose or gluconate) consumed.
3. Results
3.1. Glucose as the substrate
When glucose was used as the carbon source, all the strains are able to ferment glucose and produce ethanol as the main fermentation product. As shown in Fig. 1, the deletion of pflA led to much slower growth rates. The pflA single deletion strain maintained about 30% of the original growth rate, while the growth rate of the pflA and ldh double deletion strain was roughly 70% of that of the wild type. The deletion of pdh did not appear to lead to a slower growth rate, and the deletion of ldh led to slightly lower growth rates (90% of the wild type). The pdh and ldh double knockout strain had a similar growth rate as the ldh single knockout strain.
-
Fig. 1. Maximum specific growth rate of E. coli KO11 and mutants grown on glucose or gluconate. Error bars indicate standard deviation of the replicates.
Ethanol was the main fermentation product for E. coli KO11 and all the mutants constructed. Greater than 100% theoretical conversion based on glucose to ethanol was achieved. Casamino acids in the LB medium contribute to ethanol production, resulting in greater than 100% of the theoretical yields. The ΔldhΔpflA and Δpdh mutants had a slightly higher ethanol yield than the wild type KO11. Other mutants have similar ethanol yields as the wild type.
There was a small amount of acetate produced along with ethanol by the wild type KO11. Deletion of the pflAgene in mutants (ΔpflA and ΔpflAΔldh) eliminated any acetate production. The Δpdh mutant produced slightly less acetate than the wild type.
There was no detectable amount of lactate produced by wild type and mutants except for the strain ΔpflA.
When glucose was used as the carbon source, all the strains are able to ferment glucose and produce ethanol as the main fermentation product. As shown in Fig. 1, the deletion of pflA led to much slower growth rates. The pflA single deletion strain maintained about 30% of the original growth rate, while the growth rate of the pflA and ldh double deletion strain was roughly 70% of that of the wild type. The deletion of pdh did not appear to lead to a slower growth rate, and the deletion of ldh led to slightly lower growth rates (90% of the wild type). The pdh and ldh double knockout strain had a similar growth rate as the ldh single knockout strain.
- Fig. 1. Maximum specific growth rate of E. coli KO11 and mutants grown on glucose or gluconate. Error bars indicate standard deviation of the replicates.
Ethanol was the main fermentation product for E. coli KO11 and all the mutants constructed. Greater than 100% theoretical conversion based on glucose to ethanol was achieved. Casamino acids in the LB medium contribute to ethanol production, resulting in greater than 100% of the theoretical yields. The ΔldhΔpflA and Δpdh mutants had a slightly higher ethanol yield than the wild type KO11. Other mutants have similar ethanol yields as the wild type.
There was a small amount of acetate produced along with ethanol by the wild type KO11. Deletion of the pflAgene in mutants (ΔpflA and ΔpflAΔldh) eliminated any acetate production. The Δpdh mutant produced slightly less acetate than the wild type.
There was no detectable amount of lactate produced by wild type and mutants except for the strain ΔpflA.
3.2. Gluconate as the substrate
When gluconate was used as the carbon source, E. coli KO11 and all of the mutants produced ethanol and acetic acid as the two main products as shown in Fig. 2. The ethanol yield by E. coli KO11 was only 87.5% of the theoretical yield, while the acetate yield reached 117% of the theoretical amount of acetate with sodium gluconate as the carbon source. Overall, the strains utilized gluconate faster than glucose. The maximal ODs reached by the cells were lower than those of glucose. The growth rates of different mutants on gluconate showed a similar trend as on glucose.
-
Fig. 2. Fermentation characteristics of E. coli KO11 and mutants on gluconate. Error bars indicate standard deviation of the replicates
Knocking out the pflA gene alone led to substantially less acetate production. However the carbon flowed toward more lactate production instead of ethanol production. The ethanol yield (74%) was even lower than that of the wild type and the strain exhibited the highest level of lactate production (4 times that produced by KO11). Knocking out ldh in addition to pflA efficiently directed more carbon toward ethanol production. Mutant strain ΔldhΔpflA obtained the highest ethanol yield on gluconate at 97.5% of the theoretical maximum. Much less acetate was produced, with only 69% of the theoretical yield. No detectable amount of lactate was produced. Knocking out pdh alone led to slightly higher ethanol production, as well as less acetate production. The metabolic pathway is shown in Fig. 3, with fluxes through the pathways for the measured products (ethanol, acetate, and lactate) given in Table 2.
-
Fig. 3. Pathways involved in the fermentative utilization of glucose and gluconate by E. coli KO11. (PDC: pyruvate decarboxylase; ACD: acyl-CoA dehydrogenase; ADHII: alcohol dehydrogenase II; ADHE: alcohol dehydrogenase E, PTA: phosphate acetyltransferase; ACK: acetate kinase).
-
Table 2. Fluxes normalized to pyruvate are for batch anaerobic fermentations with both glucose and gluconate.
When gluconate was used as the carbon source, E. coli KO11 and all of the mutants produced ethanol and acetic acid as the two main products as shown in Fig. 2. The ethanol yield by E. coli KO11 was only 87.5% of the theoretical yield, while the acetate yield reached 117% of the theoretical amount of acetate with sodium gluconate as the carbon source. Overall, the strains utilized gluconate faster than glucose. The maximal ODs reached by the cells were lower than those of glucose. The growth rates of different mutants on gluconate showed a similar trend as on glucose.
- Fig. 2. Fermentation characteristics of E. coli KO11 and mutants on gluconate. Error bars indicate standard deviation of the replicates
Knocking out the pflA gene alone led to substantially less acetate production. However the carbon flowed toward more lactate production instead of ethanol production. The ethanol yield (74%) was even lower than that of the wild type and the strain exhibited the highest level of lactate production (4 times that produced by KO11). Knocking out ldh in addition to pflA efficiently directed more carbon toward ethanol production. Mutant strain ΔldhΔpflA obtained the highest ethanol yield on gluconate at 97.5% of the theoretical maximum. Much less acetate was produced, with only 69% of the theoretical yield. No detectable amount of lactate was produced. Knocking out pdh alone led to slightly higher ethanol production, as well as less acetate production. The metabolic pathway is shown in Fig. 3, with fluxes through the pathways for the measured products (ethanol, acetate, and lactate) given in Table 2.
- Fig. 3. Pathways involved in the fermentative utilization of glucose and gluconate by E. coli KO11. (PDC: pyruvate decarboxylase; ACD: acyl-CoA dehydrogenase; ADHII: alcohol dehydrogenase II; ADHE: alcohol dehydrogenase E, PTA: phosphate acetyltransferase; ACK: acetate kinase).
- Table 2. Fluxes normalized to pyruvate are for batch anaerobic fermentations with both glucose and gluconate.
4. Discussion
There are four routes of pyruvate dissimilation in E. coli KO11: pfl, ldhA, pdh, and the recombinant PET operon from Z. mobilis encoding for the pdc and adhII genes. The enzymes that utilize pyruvate, PDH, PDC, PFL, and LDH, have apparent Km values of 0.4 mM, 0.4 mM, 2.0 mM, and 7.2 mM, respectively. Due to a low apparent Kmvalue, carbon is very efficiently directed toward the PDC pathway for ethanol production. More than the theoretical maximum amount of ethanol is produced by E. coli KO11. PDH was previously thought to have minimal function under anaerobic conditions because it is strongly inhibited by NADH, which is found at higher levels during anaerobic growth and requires NAD+ for activity. Contrastingly, LDH requires NADH for activity. In this way, the organism prevents competition between these two pathways. However, it has been shown that under anaerobic conditions, there is still flux through the PDH pathway for certain strains of E. coli, and it has been proposed that the role of PDH is to maintain redox balance and provide reducing equivalents. It is also noteworthy that in certain strains deficient inpfl, PDH alone cannot support fermentative growth on glucose without supplementation but can enable growth on a more oxidized substrate such as glucuronate under anaerobic conditions
Under anaerobic conditions with glucose, 2 ATP and 2 NADH are generated during glycolysis. An additional ATP per glucose can be achieved by dissimilation of pyruvate through the PFL pathway, which has been demonstrated in E. coli, and the accumulating NADH is used to reduce metabolic intermediates and create fermentative products. With gluconate, a more oxidized substrate, only 1 ATP and 1 NADH are generated through the Entner–Doudoroff pathway. As a result, the cell needs to produce more oxidized products such as acetate, with lactic acid and ethanol production as a means to balance the accumulating NADH. Theoretically, 1.5 moles of ethanol, 0.5 moles of acetate and 1.5 ATP will be produced from 1 mole of gluconate, whereas 2 moles of ethanol and 2 ATP can be produced from 1 mole of glucose.
When glucose is the carbon source, ethanol is primarily produced by E. coli KO11 as well as a small amount of acetic acid. No obvious lactic acid was detected. The acetate produced by KO11 seems to be mainly from the PFL pathway since the deletion of pflA led to no acetate production. In the absence of pflA, the only way for the cell to produce acetyl-CoA is through the PDH pathway, thereby increasing the NADH pool. A small amount of lactic acid is produced to balance the extra reducing power, whereas no lactic acid is produced byE. coli KO11 or the other mutants when pflA is still active under the same conditions. The flux to ethanol does not appear to change. These results are in agreement with previous studies using K-12 strains, which demonstrates that LDH is the preferred pathway for redox maintenance when both LDH and PDH pathways are present.
When gluconate was used as the carbon source, the pflA deletion mutant produced only 40% of the acetic acid level of KO11, and the acetate was produced through the PDH pathway. As a consequence, more NADH was produced. The strain produced more lactic acid instead of ethanol as the means to consume the excess NADH. The lactic acid yield increased 4-fold, while the ethanol yield decreased from 87.5% to 74%.
When the LDH pathway was inactivated, there was no significant change in ethanol production when glucose was the substrate, a result that agrees with previous studies with KO11 and K-12 derivatives. However, the acetate production increased. The same trend was seen when gluconate was used as the substrate. The acetate increase indicates that flux through the PFL pathway increased. It is unlikely that flux through the PDH pathway increased significantly, as this would produce more NADH which would require increased production of ethanol, the only remaining pathway for redox balance. Since ethanol production was slightly lower than wild type KO11, we can deduce that flux through the PFL pathway is responsible for the increased acetate production. It appears that PFL is the preferred route of acetyl-CoA generation over the PDH pathway when the LDH pathway is inactivated. In support of this hypothesis, the fermentation products of the pdh and ldh double mutants are identical to that of the ldhmutants when both glucose and gluconate are used as the carbon source, in terms of both growth rates and the product profile.
When both pflA and ldh were inactivated (AH003), the highest ethanol yields were obtained for both glucose and gluconate. With pflA disrupted, the only way for the cell to produce acetyl-CoA (and subsequently additional ATP) is through the PDH pathway. This will create an excess of NADH, which can only be re-oxidized to NAD+ through the native ethanol pathway or the exogenous PET operon (pdc/adhII), resulting in a purely homoethanol fermentation when glucose is used as the substrate. When gluconate is the substrate, acetate has to be produced through the conversion of acetyl-CoA (produced by PDH) through the PTA and ACK pathways. Production of ethanol is the only means to consume NADH. Hence, ethanol yield is substantially improved compared to the wild type.
When pdh was inactivated (strain AH006), the change of flux to acetate, lactate and ethanol as compared to the wild type (KO11) remained negligible when glucose was used as the substrate. In contrast, when gluconate was the substrate, flux to ethanol increased compared to the wild type KO11, while acetic acid production significantly decreased. When glucose is the carbon source, PDH does not appear to contribute to metabolite production in the wild type when pflA is still present. However, it appears to play a bigger role in contributing to metabolite production when gluconate is the substrate. The decrease in acetate production also indicated that PDH contributes to acetate production in KO11 when gluconate was used as the substrate. The deletion of pdh gene led to the improved ethanol yield due to reduced flux toward acetate.
Although the deletion of competing pathways yields rather minor increases in ethanol production when glucose is used as the substrate, the approach is more successful in diverting carbon flow toward ethanol when gluconate is used as substrate. This difference is best explained by the effect of substrate on the intracellular NADH: NAD+ ratio. When glucose is used, there is a higher ratio of NADH: NAD+ during glycolysis because it is a more reduced substrate. PDH seems to be efficiently inhibited and plays a minimal role in the metabolism with glucose as the substrate. The high NADH: NAD+ ratio positively affects the ethanol production pathway and most of the carbon is diverted toward ethanol production. Knocking out genes for competing pathways has been only marginally successful in improving ethanol yield. The deletion of the pdh gene did not change the metabolite profile. When the more oxidized substrate, gluconate, is used, the NADH:NAD+ ratio is lower, leading to less inhibition of PDH and allowing for a more active role of PDH in the metabolism. Additionally, there is more carbon flow toward the competing pathways instead of the ethanol production pathway. The deletion of competing pathways PFL and LDH successfully eliminated the lactate production, decreased acetate production, and diverted more carbon toward ethanol production. The deletion of pdh alone led to more carbon flow toward ethanol production.
There are four routes of pyruvate dissimilation in E. coli KO11: pfl, ldhA, pdh, and the recombinant PET operon from Z. mobilis encoding for the pdc and adhII genes. The enzymes that utilize pyruvate, PDH, PDC, PFL, and LDH, have apparent Km values of 0.4 mM, 0.4 mM, 2.0 mM, and 7.2 mM, respectively. Due to a low apparent Kmvalue, carbon is very efficiently directed toward the PDC pathway for ethanol production. More than the theoretical maximum amount of ethanol is produced by E. coli KO11. PDH was previously thought to have minimal function under anaerobic conditions because it is strongly inhibited by NADH, which is found at higher levels during anaerobic growth and requires NAD+ for activity. Contrastingly, LDH requires NADH for activity. In this way, the organism prevents competition between these two pathways. However, it has been shown that under anaerobic conditions, there is still flux through the PDH pathway for certain strains of E. coli, and it has been proposed that the role of PDH is to maintain redox balance and provide reducing equivalents. It is also noteworthy that in certain strains deficient inpfl, PDH alone cannot support fermentative growth on glucose without supplementation but can enable growth on a more oxidized substrate such as glucuronate under anaerobic conditions
Under anaerobic conditions with glucose, 2 ATP and 2 NADH are generated during glycolysis. An additional ATP per glucose can be achieved by dissimilation of pyruvate through the PFL pathway, which has been demonstrated in E. coli, and the accumulating NADH is used to reduce metabolic intermediates and create fermentative products. With gluconate, a more oxidized substrate, only 1 ATP and 1 NADH are generated through the Entner–Doudoroff pathway. As a result, the cell needs to produce more oxidized products such as acetate, with lactic acid and ethanol production as a means to balance the accumulating NADH. Theoretically, 1.5 moles of ethanol, 0.5 moles of acetate and 1.5 ATP will be produced from 1 mole of gluconate, whereas 2 moles of ethanol and 2 ATP can be produced from 1 mole of glucose.
When glucose is the carbon source, ethanol is primarily produced by E. coli KO11 as well as a small amount of acetic acid. No obvious lactic acid was detected. The acetate produced by KO11 seems to be mainly from the PFL pathway since the deletion of pflA led to no acetate production. In the absence of pflA, the only way for the cell to produce acetyl-CoA is through the PDH pathway, thereby increasing the NADH pool. A small amount of lactic acid is produced to balance the extra reducing power, whereas no lactic acid is produced byE. coli KO11 or the other mutants when pflA is still active under the same conditions. The flux to ethanol does not appear to change. These results are in agreement with previous studies using K-12 strains, which demonstrates that LDH is the preferred pathway for redox maintenance when both LDH and PDH pathways are present.
When gluconate was used as the carbon source, the pflA deletion mutant produced only 40% of the acetic acid level of KO11, and the acetate was produced through the PDH pathway. As a consequence, more NADH was produced. The strain produced more lactic acid instead of ethanol as the means to consume the excess NADH. The lactic acid yield increased 4-fold, while the ethanol yield decreased from 87.5% to 74%.
When the LDH pathway was inactivated, there was no significant change in ethanol production when glucose was the substrate, a result that agrees with previous studies with KO11 and K-12 derivatives. However, the acetate production increased. The same trend was seen when gluconate was used as the substrate. The acetate increase indicates that flux through the PFL pathway increased. It is unlikely that flux through the PDH pathway increased significantly, as this would produce more NADH which would require increased production of ethanol, the only remaining pathway for redox balance. Since ethanol production was slightly lower than wild type KO11, we can deduce that flux through the PFL pathway is responsible for the increased acetate production. It appears that PFL is the preferred route of acetyl-CoA generation over the PDH pathway when the LDH pathway is inactivated. In support of this hypothesis, the fermentation products of the pdh and ldh double mutants are identical to that of the ldhmutants when both glucose and gluconate are used as the carbon source, in terms of both growth rates and the product profile.
When both pflA and ldh were inactivated (AH003), the highest ethanol yields were obtained for both glucose and gluconate. With pflA disrupted, the only way for the cell to produce acetyl-CoA (and subsequently additional ATP) is through the PDH pathway. This will create an excess of NADH, which can only be re-oxidized to NAD+ through the native ethanol pathway or the exogenous PET operon (pdc/adhII), resulting in a purely homoethanol fermentation when glucose is used as the substrate. When gluconate is the substrate, acetate has to be produced through the conversion of acetyl-CoA (produced by PDH) through the PTA and ACK pathways. Production of ethanol is the only means to consume NADH. Hence, ethanol yield is substantially improved compared to the wild type.
When pdh was inactivated (strain AH006), the change of flux to acetate, lactate and ethanol as compared to the wild type (KO11) remained negligible when glucose was used as the substrate. In contrast, when gluconate was the substrate, flux to ethanol increased compared to the wild type KO11, while acetic acid production significantly decreased. When glucose is the carbon source, PDH does not appear to contribute to metabolite production in the wild type when pflA is still present. However, it appears to play a bigger role in contributing to metabolite production when gluconate is the substrate. The decrease in acetate production also indicated that PDH contributes to acetate production in KO11 when gluconate was used as the substrate. The deletion of pdh gene led to the improved ethanol yield due to reduced flux toward acetate.
Although the deletion of competing pathways yields rather minor increases in ethanol production when glucose is used as the substrate, the approach is more successful in diverting carbon flow toward ethanol when gluconate is used as substrate. This difference is best explained by the effect of substrate on the intracellular NADH: NAD+ ratio. When glucose is used, there is a higher ratio of NADH: NAD+ during glycolysis because it is a more reduced substrate. PDH seems to be efficiently inhibited and plays a minimal role in the metabolism with glucose as the substrate. The high NADH: NAD+ ratio positively affects the ethanol production pathway and most of the carbon is diverted toward ethanol production. Knocking out genes for competing pathways has been only marginally successful in improving ethanol yield. The deletion of the pdh gene did not change the metabolite profile. When the more oxidized substrate, gluconate, is used, the NADH:NAD+ ratio is lower, leading to less inhibition of PDH and allowing for a more active role of PDH in the metabolism. Additionally, there is more carbon flow toward the competing pathways instead of the ethanol production pathway. The deletion of competing pathways PFL and LDH successfully eliminated the lactate production, decreased acetate production, and diverted more carbon toward ethanol production. The deletion of pdh alone led to more carbon flow toward ethanol production.
No comments:
Post a Comment