One xylose baseline industrial diploid yeast strain constructing and its followed evolutionary engineering
In this study, the combined metabolic and evolutionary engineering strategies were used to improve the metabolic utilization efficiency of glucose and xylose, acetate tolerance of S. cerevisiae from lignocellulosic hydrolysate. Due to the better fermentation performance of diploid yeast strain in vitality and endurance compared to that of haploid strain, two diploid yeasts, S. cerevisiae Angel and S. cerevisiae Henderson, were selected as the starting strains to be modified. As expected, neither the natural Angel yeast nor the Henderson yeast can use xylose. Their comparative fermentation was conducted in the same condition containing 50 g/L glucose and 50 g/L xylose with 6 g/L acetate, which was displayed as below (Fig. 1 a and b). As can be seen from the Fig. 1, two diploid strains both made good use of glucose while only used 3 g/L xylose at the end of fermentation. Ethanol production had been on an upward trend before 12 h, and after that it had been on a downward trend, indicating that the yeast preferred glucose which was consumed before it began to metabolize little xylose [28]. The preferred glucose consumption of S. cerevisiae is the direct result of the sugar specificities of the hexose transporters. Xylitol formation of Henderson yeast was slightly higher than that of Angel yeast, and there was no significant difference in glycerol accumulation. Angel and Henderson yeast strain both consumed most of initial acetate, resulting in an end concentration at 2.3 and 1.2 g/L acetate, respectively.
To enable the ability for xylose utilization, one xylose metabolic pathway carried by two plasmids fps1-nat` (XYL1, XYL2, XKS1) and pUC-TTRR (TAL1, TKL1, RKI1, RPE1) was introduced into Angel and Henderson strain, producing two engineered strains ABN and BBN, respectively [26, 29]. The fermentation performance of each engineered strain was evaluated in mixed sugar medium with acetate. ABN and BBN both consumed all glucose within 12 h, while little differences were observed in xylose consumption and ethanol production after glucose depletion. Between two engineered diploid strains, ABN showed a higher efficiency in xylose consumption and ethanol production with 27.4 g/L ethanol produced from 50 g/L glucose and 50 g/L xylose in 120 h. The amount of residual xylose was dramatically decreased to 28.8 g/L in 120 h. BBN utilized approximately 10.9 g/L xylose and produced 25.0 g/L ethanol (Fig. 1 c and d). Therefore, xylose metabolic efficiency of ABN is superior to BBN accompanied with similar glycerol accumulation and lower xylitol accumulation. Therefore, ABN with better performance was selected to conduct directional evolution experiment under the pressure of acetate. Unexpectedly, the capability of two engineered strains ABN and BBN to use acetate was much lower than that of their own original strain, indicating that metabolism of acetate was inhibited when the overexpressed xylose metabolism was enhanced.
Directed evolution and genetic modification of diploid strains
The presence of acetate can seriously hinder the yeast growth in the process of mixed sugar fermentation and results in lower xylose utilization and ethanol formation. When a large amount of acetate exists in the medium, it leads to intracellular acidification and hinters the utilization of xylose. The improvement of acetate metabolic pathway can effectively accelerate the utilization of acetate. Directed evolution was adopted here to effectively increase the metabolic efficiency of mixed sugar with acetate. Compared with haploid yeast, diploid yeast with larger in volume and stronger in vitality is often used in industrial ethanol fermentation. In this study, the originally diploid strain ABN was evolved to obtain the strain A1 [27]. On this basis, A2 was obtained by evolving A1 in the YP medium with xylose acclimation containing 8 g/L acetate. The xylose consumption and ethanol production of A1 and A2 strains were compared in fermentation medium with 79.1 g/L glucose, 39 g/L xylose and 3.1 g/L acetate (Fig. 1 e and f). Samples were taken every 12 h during 96 h fermentation process. Although the two tested strains consumed glucose within 12 h, there also existed some differences about xylose consumption and ethanol production after glucose depletion. A1 produced 43.6 g/L ethanol with consumption of approximately 17 .9 g/L xylose relative to 45.2 g/L ethanol production with 21.4 g/L xylose consumption of A2, reaching 88.1% and 88.2% of the theoretical yield in ethanol production, respectively. In contrast to A1, A2 consumed 3.5 g/L more xylose and produced 1.6 g/L more ethanol, resulting in its better obvious fermentation advantage. It can be likely that the original strain had less genetic modification, which was not conducive to producing more favorable mutations (Fig. 1 e and f).
The xylose consumption and ethanol production of strains expressing different promoters were compared under glucose and xylose co-fermentation conditions in previous research [27]. The remained xylose after fermentation in strain WXY46 (six gene cluster: XYL1(K270R)-XYL2-TAL1-PYK1-MGT05196-PYK1-MGT05196) expressing xylose pathway genes by constitutive promoters was significantly lower than that of strain WXY48 expressing only xylose pathway genes by HSP promoters and strain WXY63 expressing xylose pathway genes by only TCA promoters. In consistent with that the ethanol yield of WXY46 was also higher than that of strain WXY48 or WXY63. Therefore, we expressed the six gene cluster of higher ethanol production on A2 to obtain a better A21Z strain. We wanted to investigate whether A2 can further promote the ethanol production and general applicability of the six gene cluster. The A22Z was obtained by integrating the 2z-e7 cluster (XYL1 (K270R)- XYL2-TAL1--klPYK1--MGT05196-klPYK1-MGT05196) into A2 as same above. The final xylose consumption and the ethanol yield of A21Z were 37.5 g/L and 53.5 g/L, respectively, reaching 90.7% of the theoretical yield in ethanol production. The A22Z produced ethanol of 51.1 g/L with consuming xylose of 33.2 g/L, reaching 90.1% of the theoretical yield in ethanol production. The amount of residual xylose of A21Z was dramatically decreased to 1.1 g/L in 96 h relative to the control strain A2, whose produced ethanol was also 8.3 g/L less than A21Z. Therefore, compared with A2, A21Z and A22Z showed significant improvement in xylose consumption rate, ethanol production rate and sugar alcohol conversion rate. However, the fermentation effect of A22Z was slightly worse than that of A21Z (Fig. 2 a and b).
Adaptive acclimation of the target strain S. cerevisiae A21Z in hydrolysate at the solids loading of 15% wheat straw stover
The lignocellulose is usually used to produce bio-based chemicals such as ethanol by anaerobic fermentation at present. Because of the complicated components of the wheat straw hydrolysate, inhibitors are accumulated in the pretreatment step. Although some inhibitors are removed after detoxification, the remaining inhibitors still negatively affects the cell growth of the fermentation. The adaptive domestication strategy is the commonly used method so that the target strains can be better adapted to the fermentation environment of industrial hydrolysate. In this study, Fig.3 a and b indicated that the glucose concentration of the mixture of the yeast seed A21Z and treated wheat straw hydrolysate at 10.0% (v/v) was 45.85 g/L; the initial concentration of xylose, ethanol and glycerol were 17.07 g/L, 1.84 g/L, 1.01 g/L, respectively. Stability of commonly used measures of target strain in the 15% solids loading wheat straw hydrolysate mainly referred to the utilization of glucose, xylose, and increasing ethanol and glycerol generation. 15% wheat straw stover hydrolysate was utilized by A21Z to produce ethanol directly (Fig. 3 a, b). In the initial transfer process, target strain A21Z could not utilize 15% wheat straw stover hydrolysate stably to enrich ethanol. While with the adaptive evolution, the releasing glucose and xylose with the increasing ethanol and the glycerol generation kept in a stable range. This means that the strain had adapted to the environment of 15% solid content of wheat straw detoxified hydrolysate, and the simultaneous saccharification and co-fermentation (SSCF) could be carried out to further evaluate the target strain.
Simultaneous saccharification and co-fermentation (SSCF) of S. cerevisiae A21Z and its acclimated A31Z
Wheat straw was dry acid pretreated and bio detoxified before the simultaneous saccharification and co-fermentation (SSCF). The co-fermentation strains using glucose, xylose and other kinds of saccharides were one of the key factors for simultaneous saccharification. SSCF was the better choice for the use of xylose rich lignocellulosic materials to produce ethanol and A21Z was the better target strain for mixed sugar fermentation at laboratory fermentation stage. With the adaptive acclimatization, the further adapted strain A31Z derived from A21Z in wheat straw stover hydrolysate was evaluated by SSCF. In the pre-hydrolysis stage, as shown in Fig.3 c, glucose increased with the saccharification time with a certain cellulase dosage of 9.71% (W/W) in fermentation medium. Xylose was constant in the saccharification time because the majority of xylan was already converted to xylose and oligo-xylan in the pretreatment step. Glucose and xylose were about 77.08 g/L and 35.95 g/L with the saccharification time of 12h. In the consequent SSCF stage, with the function of adapted S. cerevisiae A31Z,the initial glucose was quickly converted into ethanol within 24 h and then started to utilize glucose from cellulose hydrolysis. Xylose conversion decreased with the fermentation time step by step. Finally, at the cellulase dosage of 15 mg total protein /g, adapted S. cerevisiae A31Z produced 56.68 g/L of ethanol, leading to sugar alcohol conversion rate of 63.13%; the xylose conversion rate of A31Z achieved 84.90% (Fig.3 c). Approximately, half ethanol came from the initial glucose released from the pre-hydrolysis, the other half was from xylose and the glucose released during the SSCF [30].
Fermentation performance of different yeasts in enzymatic hydrolysis of Miscanthus, maize and wheat straw
Lignocellulosic biomasses such as Miscanthus, maize and wheat straw are important feed stocks for 2G bioethanol. Here, sugars from enzymatic hydrolysis of Miscanthus, maize and wheat straw were used as substrates for the yeasts to produce ethanol, respectively. Sugar yields (% dry matter) released could be found in Table S1, and relative cell wall composition from these biomasses could be found in Table S2. As shown in Fig. 4, the control strain Angel used in a company for ethanol production, the evolved stain A31Z and our previous reported strain CE7 were chosen to ferment these hydrolyses at 37℃. Ethanol yield and concentration produced could also found in Table 2. It showed that strain A31Z could give a higher production of ethanol in the hydrolysis of Miscanthus. Under the condition of other two hydrolyses, there is no significant difference in the ethanol production among these three yeasts. However, as shown in Table 3, there is a signification improvement of pentose utilization rate of A31Z in all three kinds of hydrolyses.
Fermentation performance of A31Z using corn as feedstocks
The engineered strain A31Z was given a better performance in pentose utilization, and showed a good potential to be used as a cell factory for 2G ethanol. Here we also evaluated the performance of A31Z and a commercial strain Angel as control strain in 1st generation yeast-based production process of bioethanol from corn (Fig. 5a and 5b). Corn contained approximately two-thirds starch, and the corn starch was converted to glucose with the process described in the method part. After 120 h fermentation with extracted glucose from corn starch, the control strain Angel could produce 119.40 g/L ethanol production, while A31Z could produce 122.32 g/L ethanol, as shown in Fig. 5a and 5b. However, the byproduct glycerol was reduced about 20% in A31Z. It was believed that glycerol and ethanol levels are inversely related [31], and the slightly increased ethanol in A31Z could be due to the decreased glycerol.
Corn contained approximately two-thirds starch. When corn is processed by yeast to produce ethanol, other component corn except starch could be recycled into a rich and nutritious feedstock, named DDGS (Dried Distillers Grains with Solubles). Glucose and xylose were the main sugars contained in DDGS. The A31Z was purposely engineered to use pentose, and was firstly evaluated its performance using a mixture of glucose/xylose as a mimic of DDGS hydrolysis (Fig. 5 c and d). As expected, there was a significant improvement when xylose was added in the medium, which is also consistent with our previous data [27]. A31Z made an ethanol production at 63.33 g/L, and Angel only produced 46.21 g/L ethanol. The difference was obviously due to the different xylose consuming ability of these two stains as shown in Fig. 5c and 5d. Then A31Z and Angel were fermented using DDGS hydrolysis directly (Fig. 5e and 5f). The fermentation of A31Z resulted in an ethanol production at 5.54 g/L from a total sugar at 11.17 g/L (contained 7.84 g/L glucose and 3.33 g/L xylose). In contrast, the fermentation of Angel only resulted in an ethanol production at 4.94 g/L from a total sugar at 9.56 g/L (contained 6.74 g/L glucose and 2.82 g/L xylose). It showed that there was a better performance in A31Z for ethanol production under the fermentation using DDGS.
The SSCF of whole corn in integrated ethanol production
It has been reported that the integrated fermentation of whole corn could result in a higher ethanol concentration and could facilitate the introduction of the 2G technology [32, 33]. In the above parts, the constructed strain A31Z had been shown a good fermentation ability from a mixture of glucose and xylose, 1G feedstock (such as corn) and 2G feedstock (such as Miscanthus). In this section, we ferment this strain in an integrated process to simultaneously ferment 1G and 2G feedstock. As shown in Fig. S1, the carbon source was obtained from corn flour (1G feedstock) and corn stover (2G feedstock). The SSCF experiments showed there was a significant improvement in the ability of ethanol production and xylose utilization in the strain A31Z (Table 4). It had been reported that the integrated 1G and 2G ethanol production process will result a better economic result when compared with the stand-alone 1G or 2G plant [34]. Here we showed the technical feasibility of A31Z in the integrated 1G + 2G ethanol production process from whole corn with the current hydrolysis technology. In the future, the ethanol production ability can be further enhanced when advanced hydrolysis technologies were developed, which could result a better economic result and enabled its commercialization.