Background

Lignocellulosic biomass is produced by plant photosynthesis from solar energy and is the most abundant renewable feedstock in the world. The bio-conversion of lignocellulosic biomass into ethanol is viewed as one of the most promising ways to partially replace traditional fossil fuels, since the combustion of ethanol produces less particulate matter, carbon monoxide, and hydrocarbons than fossil fuels [1]. In addition, as bioethanol has a high octane number, mixing bioethanol with gasoline could improve the anti-detonating quality of transportation fuel.

Lignocellulosic biomass is primarily degraded into glucose and xylose after pretreatment and enzymatic hydrolysis. The simultaneous and efficient bioconversions of the two sugars into ethanol are the prerequisites for large-scale production of cellulosic ethanol. However, due to carbon catabolite repression (CCR), a considerable amount of wild microbes and engineered microbes having exogenous xylose–metabolic pathways, such as Zymomonas mobilis, Escherichia coli, Saccharomyces cerevisiae and Bacillus amyloliquefaciens, prefer to use glucose, and therefore, their xylose utilization generally lags behind glucose utilization [2,3,4,5]. This greatly hampers the large-scale application of cellulosic ethanol in industry.

Extensive studies have been attempted to relax CCR. For example, as all known xylose transporters are suppressed by glucose, many researchers have tried to engineer glucose-insensitive xylose transporters by evolutionary engineering, error-prone PCR, and site-directed mutagenesis. By this way, researchers have successfully built Gal2-N376F, CiGXS1 FIVFH497* and AN25-R4.18 [6,7,8]. In addition, adaptive evolution, computation simulation, and rational design have been used to find appropriate intracellular targets to alleviate CCR, such as the phosphoenolpyruvate transferase system (PTS), the cyclic AMP receptor protein (CRP), and the xylose operon regulatory protein [9, 10]. However, the complex nature of CCR makes it difficult to entirely reveal how CCR functions. Moreover, those above engineered strains developed cannot co-utilize glucose and xylose with high efficiency [9].

An alternative method is to build artificial consortia to co-ferment glucose and xylose. A common practice is to build a consortium consisting of a xylose-specific strain due to the deficiency in the PTS system and a wild strain that utilizes glucose because of CCR [9]. In this way, 20.82 g/L butanol with a yield of 0.35 g/g was produced from glucose and xylose using two E. coli strains. These results were comparable with butanol titers and yields produced in previous studies from glucose alone [11]. Another strategy is to use two wild species to ferment the glucose and xylose mixture. For example, when Scheffersomyces stipitis and S. cerevisiae were co-fermented, the xylose removal efficiency and ethanol production showed remarkable improvement than in their mono-fermentation [12]. In comparison with two-stage fermentation (glucose and xylose are consumed in a separate fashion), the consortium fermentation was advantageous in terms of assimilating glucose and xylose concomitantly and shortening fermentation time. Compared with strain improvement, the consortia might realize the exchange and transmission of substances, signals and energy between different cells, improve the adaptability and robustness of the population to complex environment. The synergistic interaction of their metabolic pathways also should be more useful than using each one separately. However, the challenge of consortia is selecting suitable strains which minimize the probable negative interactions.

Zymomonas mobilis is an excellent ethanol-producing species whose ethanol production efficiency can reach as high as 98%, which is higher than S. cerevisiae [13]. However, the wild Z. mobilis cannot utilize xylose unless it has been transformed by the exogenous xylose metabolic pathway, such as Z. mobilis 8b [14]. In addition, S. stipitis is recognized as one of the best microbes in nature in terms of its xylose assimilation ability, but it has a severe CCR phenomenon. In recent studies, the consortium of S. stipitis and Z. mobilis has been studied for the co-fermentation of glucose and xylose. However, these studies generally involved two-stage fermentation, or the total sugar concentrations in lignocellulosic hydrolysate medium were low, which is unrealistic in large-scale fermentation [15,16,17,18]. Even for simulated medium containing pure sugars, the xylose removal efficiency still requires improvement [19].

The present work focuses on investigating the potential of consortium fermentation consisting of S. stipitis and Z. mobilis in a simulated medium and corn stover hydrolysate, with a great emphasis on alleviating CCR, increasing the sugar removal efficiency, and increasing the bioethanol production. The corn stover was chosen since it is one of the most common annual agricultural wastes produced in China [20]. Transforming corn stover into industrial products, such as ethanol, not only realizes the resource utilization of this waste, but also contributes to the progress of global warming mitigation.

Results and discussion

Co-culture of S. stipitis CICC1960 and Z. mobilis 8b in 80G40XRM

Artificially simulated 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 [21].

In a previous study [22], 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 OD620 = 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 “high density” and a “low amount” of inoculum fermentations, hoping to alleviate the CCR phenomenon in S. stiptis. However, as shown in Fig. 1 (with “high density” inoculums) and Additional file 1: Fig. S1 (with “low amount” inoculums), S. stipitis CICC1960 still could not simultaneously assimilate glucose and xylose. 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.

Fig. 1
figure 1

Fermentation profiles of consortia consisting of S. stipitis CICC1960 and Z. mobilis 8b in 80G40XRM with “high density” inoculums of fermentations. a Glucose assimilation profiles. b Xylose assimilation profiles. c Ethanol production profiles. Data are mean ± standard error from four replicates

Full size image

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 and Fig. S1. During fermentation with “high density” of inoculum, 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 [23]. 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 6 h 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) with “high density” of inoculum, 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 9 h 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. Specially, 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. To explain the best results observed were not due to the inoculum size of Z. mobilis, but to the presence of S. stipitis, the results of “S. stipites: Z. mobilis = 2:0” were compared with “S. stipitis: Z. mobilis = 1:1”. The results (Fig. 1) showed that the xylose utilization and ethanol yield of higher initial inoculation amount of S. stipitis (2:0) were better than that of 1:0 inoculation amount; however, it was still not better than that of the consortium (S. stipitis: Z. mobilis = 1:1) and other consortia.

Table 1 Comparison of various consortia fermentation profiles in simulated medium
Full size table

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.

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 [24]. It was demonstrated that disruption of ZMO0689 could improve xylose fermentation performance of Z. mobilis [25]. 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, c), 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 higher xylose-to-ethanol transforming rate in Z. mobilis FR2 fermentation could be attributed to the extra introduction of four xylose metabolic genes (xylA, xylB, tktA, talB). As the expression of the four genes are all controlled under the constitutive promoter pdc, the expression levels of their protein products (xylose isomerase, xylulokinase, transketolase, transaldolase) in Z. mobilis FR2 are higher than its parental strain Z. mobilis 8b. Therefore, xylose was transformed into fructose-6-P and flyceraldehyde-3-P and then entered the Entner–Doudoroff pathway in a greater speed (Additional file 1: Fig. S2), and finally contributes to the improvement of ethanol productivity in Z. mobilis FR2 fermentation. 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 [24, 26]. However, as shown in Fig. 2a, b, 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, 39,676/pZB4L [27], ZM4/AcR (pZB5, pJX1) [28], A3 [24], and AD50 [26], 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 [26], 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.

Fig. 2
figure 2

Fermentation profiles of Z. mobilis 8b, Z. mobilis FR1, and Z. mobilis FR2 in 80G40XRM and in corn stover hydrolysate. a, b, c Glucose, xylose, and ethanol production, respectively, in 80G40XRM; d, e, f Glucose, xylose and ethanol production in corn stover hydrolysate. The total concentration of glucose and xylose in the hydrolysate was 120 g/L before autoclave sterilization. Data are mean ± standard error from four replicates

Full size image

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.

Fig. 3
figure 3

Fermentation profiles of Z. mobilis FR2 in 80G40XRM with different rotation speeds. a Glucose assimilation profiles. b Xylose assimilation profiles. c Ethanol production profiles. Data are mean ± standard error from four replicates

Full size image

Since the theoretical ethanol yield is 0.51 g/g xylose, the 1.65 g/L more xylose consumed by Z. mobilis FR2 in static condition, in theory, should be transformed into 0.84 g/L more ethanol. 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 [29]. 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 [30]. 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 [31,32,33].

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. 2d 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. 2e), though the two strains did not exhibit a difference during the first 24 h, 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). In addition, 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. 2f). These results agreed with the fermentation results in 80G40XRM (Fig. 2a, b, c). 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. stipitisCICC1960 in corn stover hydrolysates.

As shown in Fig. 4, 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). In addition, 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. 4). 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).

Fig. 4
figure 4

Consortium fermentation profiles in variousconcentrations of corn stover hydrolysates. a, b, c Glucose, xylose and ethanol fermentation profiles in 60 g/L of corn stover hydrolysate. d, e, f Glucose, xylose and ethanol fermentation profiles in 90 g/L of corn stover hydrolysate. g, h, i Glucose, xylose and ethanol fermentation profiles in 120 g/L of corn stover hydrolysate.The concentration of corn stover hydrolysates represents the total amount of glucose and xylose in the hydrolysates before autoclave sterilization. For consortium fermentation, the inoculation ratio between S. stipitis CICC1960 and Z. mobilis FR2 was 1:3. Data are mean ± standard error from three replicates

Full size image
Table 2 Fermentation profiles in the corn stover hydrolysate
Full size table

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, such as phenols, were presented in the corn stover hydrolysate [34]. 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 [35].

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, 36]. 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 [36]. 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.

Analysis of gene expression by qRT-PCR

As for the outstanding fermentation ability of the inoculum ratio 1:3 between S. stiptis and Z. mobilis with “high density”, genes of xylose and glucose transporters in two strains and carbon catabolite repressor (CCR) in S. stipitis were chosen to qRT-PCR analysis. Samples were taken at 3, 9, 21 and 27 h, respectively, and the gene expressions in the consortia were compared with that of corresponding mono-fermentation at the same time. As shown in Fig. 5, at the early stage of fermentation (3 h), genes related to xylose and glucose transporters in S. stipitis (XUT4, XUT5, XUT7, SNF3, QUP2, RGT2, HGT1, HGT2) and Z. mobilis (ZMO 0366, ZMO 0293) had no significant differences between the consortium (1:3) and mono-fermentation. However, these genes in the consortium (1:3) were up-regulated to some different extent compared with that in mono-fermentation at 9 h and 21 h. For example, XUT5 and XUT7 related xylose transporter of S. stipitis in the consortium (1:3) had a 155.1- and 5.8-fold enhanced expression at 9 h, respectively, compared to the mono-fermentation of S. stipitis, even 549.5-fold for XUT5 at 21 h. Among glucose transporter related genes including SNF3, QUP2, RGT2 and HGT1 of S. stipitis in the consortium (1:3) were up-regulated about 17.1-, 68.9-, 7.9-, and 185.6-fold at 9 h, respectively, compared to S. stipitis mono-fermentation. At the same time, the gene ZMO0366 (glucose facilitated diffusion protein) and ZMO0293 (sugar porter family MFS transporter) came from Z. mobilis also had about 9.9-, 13.6-fold up-regulation at 9 h in the consortium (1:3), respectively. These results above were in consistent with the results of fermentation ability, which could explain why the utilizations of xylose and glucose in the consortium (1:3) were better than that of S. stipitis mono-fermentation and Z. mobilis mono-fermentation (Fig. 1). The higher the expression of xylose or glucose transporter, the more xylose or glucose could be transported for utilization. In addition, the gene ZMO 0366 and ZMO 0293 also had about 7.9-, 14.8-fold up-regulation at 9 h in the consortium of 3:1, respectively, which were comparable to the consortium of 1:3. The finding showed that the up-regulation of genes related xylose and glucose transporters came from S. stipitis played an important role in the consortium of 1:3. It has been reported that Mig2 was identified as a repressor that collaborates with Mig1 to cause glucose-induced repression of SUC2 gene [37]. Mig2 could be also involved in the cross talk between the nucleus and the mitochondria through Ups1 to regulate mitochondrial morphology in a glucose dependent manner [37, 38]. Interesting, the expression of Mig1 gene came from S. stipitis was down-regulated at 9 h and 21 h in the consortium compared to S. stipitis mono-fermentation, while Mig2 were up-regulated. Therefore, it could be speculated that the down-regulation of Mig1 could contribute to reducing CCR, and Mig2 might play a more important role compared to Mig1 in the consortium. From Fig. 5, we also found that the increase of initial inoculation of S. stipitis in mono-fermentation didn’t cause changes in the expression levels of xylose and glucose transporters and CCR genes, which was consistent with the fermentation results, that is, the increase of S. stipitis initial inoculation didn’t improve the fermentation performance of mono-fermentation. Gao et al. have reported that the initial ratio can influence the structure and, more importantly, the function and the bacterial interaction of a community [39]. Our results suggested that there might be an interaction between S. stipitis and Z. mobilis and the initial ratio of 1:3 is able to regulate the expression of genes of xylose and glucose transporters in two strains to better utilize glucose and xylose simultaneously.

Fig. 5
figure 5

Fold changes in xylose/glucose transporters and CCR gene expression levels of consortia and control mono-fermentation. a, b, c Xylose transporter in S. stipitis; d, e, f, g, h Glucose transporter in S. stipitis; i, j Carbon catabolite repressor (CCR) in S. stipitis; k, l Glucose/xylose transporter in Z. mobilis.

Full size image

Conclusions

This study focused on evaluating a fermentation method to efficiently transform glucose and xylose to ethanol via artificial consortia composed of suspended S. stipitis and suspended Z. mobilis. By fermentation process optimization and genetic engineering, the consortium built here exhibited enhanced xylose assimilation ability and ethanol production performance in both 80G40XRM and corn stover hydrolysates, and did not experience evident CCR phenomenon. Hence, this study proved that S. stipitis and Z. mobilis could be co-cultured together in suspension in cellulosic ethanol production and provided a novel strategy for further application.

Materials and methods

Strains, plasmids and primers

S. stipitis CICC1960 was purchased from the China Center of Industrial Culture Collection. Z. mobilis 8b was kindly given by Shihui Yang, Hubei University, and was used as the starting strain for further strains constructions. Plasmid Pmini was kindly given by Nan Peng, Huazhong Agricultural University. All primers were synthesized by Tsingke Biotechnology Co., Ltd. (Chengdu, China) and purified via polyacrylamide gel electrophoresis.

All strains, plasmids and sgRNAs are listed in Additional file 1: Table S3. Primers are listed in Additional file 1: Table S4.

Genetic engineering and strain development

The CRISPR-Cas system was used to engineer Z. mobilis FR1 and Z. mobilis FR2 in this study. Specifically, the plasmids Pmini-Ppdc-talB-tktA and Pmini-Ppdc-xylA-xylB were constructed using Gibson Assembly® Protocol (New England BioLabs, Ipswich, US) and were transformed into E. coli trans110 (TransGen Biotech, Beijing, China). After verification of the plasmid sequences by sequencing, the two plasmids were transformed into Z. mobilis 8b sequentially to construct Z. mobilis FR1 (ZMO0256::Ppdc-talB-tktA) and Z. mobilis FR2 (ZMO0256::Ppdc-talB-tktA; ZMO0689::Ppdc-xylA-xylB). ZMO0256up500, pdc promoter, talB, tktA, ZMO0256down500, ZMO0689up540, xylA, xylB, and ZMO0689down500 were all amplified from Z. mobilis 8b, using the corresponding primers listed in Additional file 1: Table S4. Only the engineered strains confirmed by both gel electrophoresis and gene sequencing were used in the experiments.

Preparation of the corn stover hydrolysates

Alkaline pretreatment [40]

Corn stover was collected from Jianyang, Sichuan, China. After air-drying, milling and passing through a sieve with an aperture size of 380 μm, the corn stover was mixed with 1.34% (w/v) NaOH solution (The solid loading was 10% (w/v)), loaded into reaction kettles, and transferred into a drying oven sequentially. The treatment parameters were 140 °C and 6 h.

The pretreated corn stover was washed intensively with water or diluted HNO3 until the pH of the washing water turned neutral. The washed corn stover was oven dried and stored in sealed bags at room temperature.

Enzymatic hydrolysis

Pretreated corn stover (10 g) was mixed with 100 mL citric acid buffer (8.823 g/L tri-sodium citrate dihydrate, 3.843 g/L citric acid, pH 4.8) and 5 mL cellulase (Sigma-Aldrich, Saint Louis, US). Afterwards, the mixture was placed in an incubator at 50 °C and 150 rpm for 72 h. When the enzymatic hydrolysis was finished, the hydrolysate was centrifuged twice at 3000 g for 25 min each time. The hydrolysate was then centrifuged again at 10,000 g for 5 min for the further removal of the remaining solids. Sugars in the lignocellulosic hydrolysate were concentrated using a rotary evaporator to attain 60 g/L, 90 g/L, and 120 g/L hydrolysates with regard to the concentrations of total glucose and xylose. Subsequently, 10 g/L yeast extract, 2 g/L KH2PO4, 1 g/L (NH4)2SO4, and 2 g/L MgSO4·7H2O were added to the hydrolysates before adjusting the pH of the hydrolysates to 5.6 by 5 M NaOH.

Growth and fermentation conditions

Preparation of the seed cultures of S. Stipitis and Z. mobilis

Scheffersomyces stipitis was streaked in a YPD plate (20 g/L glucose, 10 g/L yeast extract, 20 g/L peptone, and 15 g/L agar) and cultured at 30 °C for approximately 1 day. A single colony was transferred from the YPD plate into 5 mL of YP120X medium (120 g/L xylose, 10 g/L yeast extract, and 20 g/L peptone) and cultivated at 30 °C and 150 rpm for 19–36 h. Afterwards, the entire pre-seed culture was inoculated into a 100 mL of fresh YP120X medium and cultivated at 30 °C and 150 rpm for 22 h. At this point, the OD600 value of the seed culture was approximately 1.8.

Similarly, Z. mobilis was streaked in a RM plate (20 g/L glucose, 10 g/L yeast extract, 2 g/L KH2PO4, 1 g/L (NH4)2SO4, 2 g/L MgSO4·7H2O, and 15 g/L agar) and cultivated at 30 °C for 2 days. A single colony was transferred from the RM plate into 5 mL of 80G40XRM (80 g/L glucose, 40 g/L xylose, 10 g/L yeast extract, 2 g/L KH2PO4, 1 g/L (NH4)2SO4, and 2 g/L MgSO4·7H2O) and cultivated statically at 30 °C for 19–36 h. The entire culture was then inoculated into 100 mL of the fresh 80G40XRM and cultivated statically at 30 °C for 22 h. The OD600 value of the seed culture was approximately 1.8 at this point.

Fermentation conditions

For the “high density” inoculation fermentation, the seed cultures of S. stipitis and Z. mobilis were centrifuged at 4 °C with 4000 rpm for 5 min, washed once with sterile ddH2O, and suspended with 1/25 (v/v) new fermentation medium, respectively (for 80G40XRM fermentation, 80G40XRM was used to suspend cells, while sterile ddH2O was used for corn stover hydrolysates fermentation). The concentrated cells were then inoculated into the corresponding fermentation medium with different ratios of S. stipitis and Z. mobilis. The total inoculum size was 100% (v/v, the ratio of seed culture volume/fermentation culture volume). For the “low amount” inoculation fermentation, after centrifuging and washing, the seed cultures of S. stipitis and Z. mobilis were added proper 80G40XRM to suspend cells to make OD600 value about 1.0.The total fermentation volume was 50 mL for 80G40XRM fermentation and 10 mL for corn stover hydrolysates fermentation. The fermentation conditions were 30 °C and 150 rpm, unless there was further indication in the portion for exploring the effect of oxygen on Z. mobilis fermentation (as seen in Section “Effect of oxygen on Z. mobilis FR2 fermentation”).

Four replicates were performed for the 80G40XRM fermentation, while three replicates were performed for the corn stover hydrolysates fermentation.

Analytical methods

Determination of glucose, xylose and ethanol concentrations during fermentation

The glucose, xylose, and ethanol titers in the fermentation samples were analyzed by an Agilent 1200 Series HPLC system (Agilent Technologies, Santa Clara, US) equipped with a Bio-Rad HPX-87H column (Bio-Rad Laboratories, Richmond, US). The mobile phase was 5 mM H2SO4. The operating parameters were 20 μL injection volume, 0.6 mL/min rate, and 35 °C. The ethanol productivity and yield were calculated using the following formulas.

  • (1) Ethanol productivity = Ethanol titer/fermentation time.

  • (2) Ethanol yield = Ethanol titer/glucose and xylose consumed.

The theoretical ethanol yield is 0.51 g/g sugars consumed.

Quantitative real-time reverse transcription PCR

Total cellular RNA was extracted from cells grown for 3 h, 9 h, 21 h and 24 h using Trizol reagent (Invitrogen). Reverse transcription step was carried out using Goldenstar RT6 cDNA Synthesis Kit Ver 2 (TsingKe, Chengdu, China) with random primer mix following the manufacturer’s manual. Quantitative real-time reverse transcription PCR was performed in the Bio-Rad iQ5 real-time PCR detection system with SuperReal PreMix (SYBR Green) (TIANGEN, Beijing, China). All optimized primers are shown in Table S4 and were designed using primer software to amplify approximately 100 bp from the 3′ end of the target genes. PCR conditions were 1 min at 95 °C, followed by 40 cycles of heating at 95 °C for 15 s and 60 °C for 15 s, and 72 °C for 30 s, and final extension at 72 °C for 5 min. PCR amplification was detected by SYBR Green. The ratios of the cycle threshold (Ct) values were determined from the Bio-Rad iQ5 Optical System Software provided. To analyze the gene expression level, the ΔΔCt method was chosen and standard curves of each primer were plotted to ensure similar amplification efficiency compared with the reference gene. The rrsA gene, encoding the 16S RNA, and tdh2 gene served as an endogenous control to normalize for differences in total RNA for Z. mobilis 8b, and S. stipitis, respectively.

Statistical analysis

Data are presented as mean ± standard error. All figures were prepared using Prism 8 (GraphPad Software, LLC).

Significant differences were statistically analyzed using IBM® SPSS® Statistics (Version 22, US). If P > 0.05 in the homogeneity of variance tests, a one-way ANOVA followed by T test was used. Otherwise, a nonparametric test (Kruskal–Wallis H) was used.