2.1 Co-culture of S. stipitis CICC1960 and Z. mobilis 8b in 80G40XRM
Artificially synthesized 80G40XRM (80 g/L glucose + 40 g/L xylose) was first used to explore the appropriate mode of this consortium fermentation with the aim of improving the glucose-and-xylose co-utilization efficiency. The ratio between glucose and xylose in this medium was 2:1, aiming to simulate the real ratio in lignocellulosic hydrolysates.
In a previous study (21), when S. stiptis was pre-cultured in the medium with glucose as the sole carbon source (glucose medium), its xylose-metabolic gene expression, such as D-xylose reductase and xylitol dehydrogenase expression, was inhibited. Hence, when S. stiptis was inoculated in the xylose medium (xylose was the sole carbon source in the fermentation medium) later, its xylose-metabolic genes needed to be synthesized from scratch. Therefore, S. stiptis exhibited an apparent lag in fermentation in the xylose medium. In contrast, when S. stiptis was precultured in the xylose medium, the two xylose-metabolic genes were fully expressed and the aforementioned problem of lag in the xylose fermentation was alleviated, especially if S. stiptis was inoculated with high initial density, such as A620 = 40. To this end, we decided to preculture S. stiptis CICC1960 in YP120X (120 g/L xylose) and then inoculate it to the 80G40XRM (fermentation medium) with a 100% (v/v) inoculum size, hoping to alleviate the CCR phenomenon in S. stiptis. However, as shown in Fig. 1a and 1b, S. stipitis CICC1960 still could not assimilate glucose and xylose at the same time. This might have been a result of the glucose repression on xylose transportation into the cells (6), or the inoculum size used in this study (the initial OD600 in fermentation was about 1.8) was not high enough. For economic reasons, the inoculum size was not further increased in this study.
For Z. mobilis 8b, 80G40XRM was used as its seed culture medium, since Z. mobilis 8b cannot grow well in xylose medium (data not shown). The results of the Z. mobilis 8b mono-fermentation are shown in Fig. 1 and Additional file 1: Table S1. During fermentation, Z. mobilis 8b showed a strong ability to utilize glucose and xylose simultaneously with ethanol productivity reaching 5.10 g/L/h. More specially, the CCR in Z. mobilis 8b was largely alleviated, which may be attributed to the high inoculum size used in this study (the initial OD600 was approximately 1.8). In 2004, Ali Mohagheghi et al. found that Z. mobilis 8b (the initial inoculum density was 0.2 at 600 nm) exhibited an apparent lag of xylose utilization, while glucose could be assimilated immediately (22). Since the xylose-metabolic genes were inserted to the genome of Z. mobilis 8b, their expression was generally low in individual cells. When Z. mobilis 8b were inoculated with high density, activities of those enzymes encoding xylose-metabolic genes would be higher than the ones under a low initial density inoculation, and this may contribute to the alleviation of CCR in this study (14). Glucose was completely removed within six hours by Z. mobilis 8b, while xylose still remained at 6.16 g/L at 21 h. The ethanol yield of the Z. mobilis 8b mono-fermentation was 0.48 g/g (the theoretical ethanol yield is 0.51 g/g sugars).
When S. stipitis CICC1960 and Z. mobilis 8b were co-cultured together with different inoculation ratios (3:1–1:3), the glucose consumption profiles of the consortia did not differ considerably from the Z. mobilis 8b mono-fermentation (Fig. 1a). Glucose was completely removed within nine hours by the consortia. On top of this, no apparent CCR was shown since xylose was rapidly assimilated at the very early stage (Fig. 1b). Although xylose consumption rates by the consortia were lower than that in the Z. mobilis 8b mono-fermentation. It was shown that, in general, the higher the ratio of Z. mobilis 8b applied, the higher the rates of xylose assimilation achieved. In special, when S. stipitis CICC1960 : Z. mobilis 8b = 3:1 (initial inoculum size proportion), the xylose assimilation rate was 0.92 g/L/h; when S. stipitis CICC1960 : Z. mobilis 8b = 1:3, the xylose assimilation rate improved to 2.38 g/L/h. At 21 h, the consortia fermentation reached the endpoint of fermentation. Interestingly, when S. stipitis CICC1960 : Z. mobilis 8b = 1:3, the xylose consumption reached 36.73 g/L, which was significantly (P < 0.01) higher than that in the S. stipitis CICC1960 mono-fermentation and that in the Z. mobilis 8b mono-fermentation (Additional file 1: Table S1). Correspondingly, the ethanol titer of this consortium reached 57.21 g/L, which was 48.37% higher than that of the S. stipitis CICC1960 mono-fermentation and 6.52% higher than that of the Z. mobilis 8b mono-fermentation. These results (S. stipitis CICC1960 : Z. mobilis 8b = 1:3) were comparable with or better than other consortia fermentations listed in Table 1, in terms of xylose removal efficiency, ethanol yield, and ethanol productivity.
Table 1
Comparison of various consortia fermentation profiles in synthetic medium
a Fermentation mode | Initial sugar concentration (g/L) | Xylose removal efficiency (%) | Ethanol yield (g/g) | Ethanol productivity (g/L/h) | Reference |
Two-stage fermentation | | | | | |
Suspended Z. mobilis + suspended S. stipitis | 80 g/L glucose + 40 g/L xylose | 62.5 | - | 1.56 | (15) |
b Suspended S. stipitis + suspended Z. mobilis | 60g/L xylose + 100 g/L glucose | - | 0.474 | 1.416 | (16) |
Suspended Z. mobilis + suspended S. stipitis | 80 g/L glucose + 40 g/L xylose | 0.67 | 0.36 | 0.41 | (19) |
One-stage fermentation | | | | | |
Suspended S. cerevisiae + suspended S. stipitis | 75 g/L glucose + 30 g/L xylose | 79.6 | 0.4 | 1.26 | (37) |
Suspended S. cerevisiae + suspended S. stipitis | 20 g/L glucose + 10 g/L xylose | - | 0.416 | 0.608 | (38) |
c Immobilized Z. mobilis + immobilized S. stipitis | 80 g/L glucose + 40 g/L xylose | 72.5 | 0.37 | 0.87 | (19) |
d Suspended Z. mobilis + suspended S. stipitis | 80 g/L glucose + 40 g/L xylose | 84.95 | 0.50 | 4.99 | This study |
a Two-stage fermentation means the two species were inoculated into the medium sequentially, while one-stage fermentation means the two species were inoculated into the medium simultaneously |
b Xylose was first depleted by S. stipitis. Then, glucose medium and Z. mobilis were added to the same system to initiate the glucose fermentation |
c Z. mobilis and S. stipitis were immobilized separately |
d S. stipitis CICC1960 : Z. mobilis 8b = 1:3 |
-, not available |
Due to the outstanding fermentation ability of this consortium (S. stipitis CICC1960 : Z. mobilis 8b = 1:3), the ratio between S. stipitis and Z. mobilis of 1:3 was later employed in the fermentation of corn stover hydrolysate.
2.2 Genetic engineering of Z. mobilis 8b
In Z. mobilis, ZMO0256 encoding D-lactate dehydrogenase is involved in the production of lactate as a byproduct; ZMO0689 encoding glucose-fructose oxidoreductase participates in xylitol and sorbitol production (23). It was demonstrated that disruption of ZMO0689 could improve xylose fermentation performance of Z. mobilis (24). In order to introduce one copy of xylose metabolic genes (xylA, xylB, tktA, talB) into the Z. mobilis 8b genome and improve its xylose assimilation performance, the engineered strains Z. mobilis FR1 (ZMO0256::Ppdc-talB-tktA) and Z. mobilis FR2 (ZMO0256::Ppdc-talB-tktA; ZMO0689::Ppdc-xylA-xylB) were sequentially constructed.
The fermentation performances of Z. mobilis FR1 and Z. mobilis FR2 were evaluated in 80G40XRM and were compared with their parental stain Z. mobilis 8b. As shown in Fig. 2a, the three strains did not differ in their glucose assimilation profiles, and glucose was depleted within 8.5 h. For the xylose consumption and ethanol production profiles (Fig. 2b and 2c), Z. mobilis FR1 did not show much difference with Z. mobilis 8b. However, Z. mobilis FR2 accelerated its xylose assimilation rate in the mid-to-late fermentation period (Fig. 2b): the xylose consumption and ethanol production achieved by Z. mobilis FR2 were increased by 15.36% and 6.81% respectively at 20.5 h as compared with Z. mobilis 8b (Additional file 1: Table S2). Besides, the ethanol productivity of Z. mobilis FR2 was 5.08 g/L/h, which was significantly (P < 0.01) higher than Z. mobilis 8b (4.84 g/L/h). The ethanol yield of Z. mobilis FR2 was 95.47%, which was comparable with A3 (96.6%) and AD50 (96%), the two best strains developed so far by adaptive laboratory evolution (23, 25).
However, as shown in Fig. 2a and 2b, though CCR was alleviated in Z. mobilis FR2 fermentation, its xylose utilization rate was still lower than its glucose utilization rate, and there was still a slight amount of xylose that remained (approximately 2.5 g/L) at the endpoint. These results agreed with other Z. mobilis strains, including C25, 39676/pZB4L (26), ZM4/AcR (pZB5, pJX1) (27), A3 (23), and AD50 (25), yet no exact reason of the incomplete xylose utilization has been identified thus far. In Z. mobilis, xylose is transported through a glucose facilitated diffusion protein (25), which is a native glucose transporter and has low affinity to xylose. This low affinity xylose transport might be the burden behind the above-mentioned problems. Further investigation into this field would greatly promote the commercialization of cellulosic ethanol.
2.3 Effect of oxygen on Z. mobilis FR2 fermentation
In the above experiments, as S. stipitis could not grow under static cultivation, 150 rpm was applied in the S. stipitis mono-fermentation, the Z. mobilis mono-fermentation, and the consortia fermentation that consisted of the two species. However, as Z. mobilis is a facultative anaerobe, it can ferment under both static and aerobic conditions. Therefore, a further study was conducted to check whether there was any difference in Z. mobilis FR2 fermentation profiles under static and agitated (150 rpm) conditions in 80G40XRM. As shown in Fig. 3, oxygen significantly boosted Z. mobilis FR2’s glucose (P < 0.05) and xylose (P < 0.01) consumption rates and improved ethanol productivity by 54.51% (P < 0.01). However, at fermentation endpoint (27 h), the static fermentation of Z. mobilis FR2 showed an increase in xylose consumption and ethanol production by 1.65 g/L and 3.45 g/L, respectively compared with that in agitated fermentation. To achieve high ethanol productivity, ethanol fermentation under 150 rpm was kept in this study.
In theory, the 1.65 g/L more xylose consumed by Z. mobilis FR2 in static condition should be transformed into 0.84 g/L ethanol (the theoretical ethanol yield is 0.51 g/g xylose). However, the real difference in ethanol titers between the static fermentation and agitated fermentation was as high as 3.45 g/L. This suggests oxygen had a negative effect on ethanol production by Z. mobilis FR2. In 1990, Tanaka et al. found that Z. mobilis produced a little more acetaldehyde (0.28–4.49 g/L) when oxygen was supplied (28). Acetaldehyde was primarily produced by NADH dehydrogenase, one of the key components in the Z. mobilis respiratory chain. This enzyme has the same cofactor (NADH) as ethanol dehydrogenase. During aerobic fermentation, the activity of NADH dehydrogenase was higher than that of ethanol dehydrogenase. Therefore, a large amount of NADH was used to reduce the dissolved oxygen concentration in the medium. Due to the lack of sufficient NADH, the transformation from acetaldehyde to ethanol by ethanol dehydrogenase was inhibited, and thus negatively affected the ethanol production of Z. mobilis under agitated cultivation (29). One evidence for this hypothesis is that for Z. mobilis mutant strains, whose NADH dehydrogenase is defective, this negative effect of oxygen on Z. mobilis ethanol fermentation could be alleviated (30–32).
2.4 Assimilation of corn stover hydrolysates by consortium composed of S. stipitis CICC1960 and Z. mobilis FR2
First, it was evaluated whether Z. mobilis FR2 could outcompete Z. mobilis 8b in corn stover hydrolysate fermentation. As shown in Fig. 4a, a difference was not observed regarding glucose assimilation ability of the two strains: glucose assimilation rates were 2.54 g/L/h and 2.51 g/L/h for Z. mobilis FR2 and Z. mobilis 8b, respectively, and glucose was depleted within 30 h. As for xylose assimilation (Fig. 4b), though the two strains did not exhibit a difference during the first 24 hours, the assimilation rate of Z. mobilis 8b gradually decreased, while the rate for Z. mobilis FR2 remained stable. At 60 h, Z. mobilis FR2 assimilated 14.60 g/L xylose, which was significantly higher (P < 0.05) than that of Z. mobilis 8b (10.60 g/L). Additionally, Z. mobilis FR2 produced 37.13 g/L ethanol in 120 g/L corn stover hydrolysate (concentration here refers to the total amount of glucose and xylose in corn stover hydrolysate before sterilization), while Z. mobilis 8b produced 34.93 g/L ethanol (Fig. 4c). These results agreed with the fermentation results in 80G40XRM (Fig. 2). Due to the better xylose assimilation ability of Z. mobilis FR2, Z. mobilis FR2 was used to replace Z. mobilis 8b in the next consortium fermentation with S. stipitis CICC1960 in corn stover hydrolysates.
As shown in Fig. 5, while consortium fermentation (S. stipitis CICC1960 : Z. mobilis FR2 = 1:3) and Z. mobilis FR2 mono-fermentation did not show any difference in glucose and xylose consumption rates and amounts in the 60 and 90 g/L corn stover hydrolysates fermentation, the consortium produced slightly more ethanol (~ 0.86 g/L) than the Z. mobilis FR2 mono-fermentation (Table 2). Additionally, the consortium fermentation in the two cases was better than the S. stipitis CICC1960 mono-fermentation in terms of glucose assimilation, xylose assimilation, and ethanol production rates and quantities, and did not exhibit strong CCR which was evident in the S. stipitis CICC1960 mono-fermentation (Fig. 5). For the 120 g/L corn stover hydrolysate fermentation, the glucose assimilation rate of the consortium (2.83 g/L/h) was slightly slower than Z. mobilis FR2 (3.24 g/L/h), while xylose assimilation rates were nearly the same prior to 36 h. However, the consortium finally produced 33.05 g/L ethanol at endpoint, which was 1.02 g/L higher than the Z. mobilis FR2 mono-fermentation and 16.7 g/L higher than the S. stipitis CICC1960 mono-fermentation (Table 2).
Table 2
Fermentation profiles in the corn stover hydrolysate
a Corn stover hydrolysate concentration (g/L) | S. stipitis CICC1960 : Z. mobilis FR2 | b Glucose consumed (g/L) | b Xylose consumed (g/L) | b Ethanol |
Titer (g/L) | Yield (g/g) | Productivity (g/L/h) |
60 | 1:0 | 38.27 ± 0.53 | 8.64 ± 0.49 | 13.40 ± 0.75 | 0.29 ± 0.01 | 0.73 ± 0.04 |
0:1 | 38.27 ± 0.53 | 13.50 ± 0.00 | 18.72 ± 0.46 | 0.36 ± 0.01 | 2.08 ± 0.05 |
1:3 | 38.27 ± 0.53 | 13.50 ± 0.00 | 19.58 ± 0.32 | 0.38 ± 0.01 | 2.18 ± 0.04 |
90 | 1:0 | 57.30 ± 1.55 | 7.74 ± 0.25 | 18.45 ± 0.35 | 0.28 ± 0.01 | 0.53 ± 0.01 |
0:1 | 57.30 ± 1.55 | 18.62 ± 0.15 | 26.96 ± 0.24 | 0.36 ± 0.00 | 1.99 ± 0.08 |
1:3 | 57.30 ± 1.55 | 18.87 ± 0.04 | 27.38 ± 0.58 | 0.36 ± 0.01 | 2.01 ± 0.05 |
120 | 1:0 | 67.35 ± 1.52 | 4.88 ± 0.19 | 16.35 ± 2.06 | 0.23 ± 0.03 | 0.19 ± 0.02 |
0:1 | 80.61 ± 4.72 | 12.56 ± 0.74 | 32.03 ± 1.14 | 0.34 ± 0.01 | 1.32 ± 0.03 |
1:3 | 80.61 ± 4.72 | 12.69 ± 0.60 | 33.05 ± 0.79 | 0.35 ± 0.01 | 1.11 ± 0.02 |
Data are mean ± standard error from three replicates |
a The corn stover hydrolysate concentration refers to the total concentrations of glucose and xylose in the hydrolysate before autoclave sterilization |
b All data were calculated based on the real sugar concentrations |
In 2014, Lalit K. Singh et al. separated kans grass hydrolysate into a xylose-rich portion and a glucose-rich portion through organic solvent extraction and then used S. stipitis and Z. mobilis to ferment each sugar (54 g/L xylose and 100 g/L glucose) sequentially. The ethanol productivity in their study was 0.723 g/L/h, which was lower than that in our study (1.11 ~ 2.18 g/L/h) (16). This was because the two-stage fermentation Lalit K. Singh et al. employed led to an increase in the fermentation time. In 2020, Ferdian Wirawan et al. used immobilized Z. mobilis and S. stipitis to sequentially ferment 50 g/L sugarcane bagasse (17). Though the productivity in their study was high (1.868 g/L/h), the actual sugar concentration in the hydrolysate was only 11 g/L glucose, 4 g/L xylose, and 4 g/L cellobiose, which is impractical in real applications. These comparisons demonstrated that the consortium fermentation mode utilized in this study has an edge in ethanol production compared with other existed modes of S. stipitis and Z. mobilis co-fermentation.
We note that the ethanol titer and productivity values in the 120 g/L corn stover hydrolysate were lower than the values in 80G40XRM (Table 2, Additional file 1: Tables S1 and S2). This was because many of the inhibitors, like phenols, were presented in the corn stover hydrolysate (33). These inhibitors negatively affected the microbial fitness in the lignocellulosic hydrolysate and thus reduced the ethanol yield and productivity. Although it has been shown that the immobilization of microbes could alleviate the negative effect to some degree, the exact mechanisms are not clear (34).
Duong Thi Thuy Nguyen et al. found that the presence of living Z. mobilis cells negatively affected the xylose assimilation performance of S. stipitis, suggesting that there might be an amensalism relationship between the two species (19, 35). Similarly, in our study, when the initial inoculum OD600 was controlled to 0.1 in the 80G40XRM fermentation, no improvement was observed in the consortium fermentation profiles compared with the Z. mobilis mono-fermentation profiles (data not shown). In contrast, when a high inoculum size (an initial OD600 of approximately 1.8) was used, the consortium of S. stipitis and Z. mobilis assimilated more xylose and produced more ethanol in both the 80G40XRM and 120 g/L corn stover hydrolysate. This implied a commensalism or cooperation relationship between the two species during fermentation (35). Perhaps this positive relationship could only be exhibited under specific conditions, such as under a high cell density.
Exploring the interactions between S. stipitis and Z. mobilis under various cell densities, as well as the ethanol fermentation performance in corn stover hydrolysate by co-immobilized cells of the two species, are our future research goals. Indeed, a deep understanding of the interactions between S. stipitis and Z. mobilis and the mechanisms behind the protective role of co-immobilization on cells will pave the way to further enhance the consortium fermentation performance while lowering the inoculum size, and will help promoting the industrial progression of cellulosic ethanol production.