Expressing IrrE to enhance yeast tolerance to acetic acid
To study the effect of IrrE on the inhibitor tolerance of S. cerevisiae, we first evaluated the role of IrrE on the strain tolerance to acetic acid, a primary inhibitor in lignocellulose hydrolysates. In the presence of acetic acid, the strain BY4742/IrrE entered the exponential phase at about 30 h and reached the stationary phase at about 54 h with the maximal OD600 = 3.56. In comparison, the growth of control strain was recovered until 48 h and reached a maximum OD600 = 3.12 at 67 h (Additional file 2: Fig. S1). The results showed that the heterologous expression of IrrE slightly enhanced the ability of S. cerevisiae cells against acetic acid shock, but it did not confer strain desired acid tolerance. Besides, we observed that the growth ability of strain BY4742/IrrE was also slightly increased in the presence of the mixture of three inhibitors (furfural, acetic acid and phenol) (Fig. 2). Previous studies have shown that engineered IrrE can enhance acetic acid tolerance of Zymomonas mobilis and yeast[25, 34]. Therefore, it is necessary to take some measures to engineer IrrE in S. cerevisiae to optimize strain FAP tolerance further.
Directed evolution of IrrE to improve strain tolerance
The directed evolution strategy has been widely used in protein modification and it showed that engineered IrrE conferred E. coli enhanced tolerances toward lignocellulosic hydrolysates inhibitors[22]. To further optimize strain tolerance to lignocellulose-derived inhibitors, especially the mixture of multiple inhibitors FAP, directed evolution was employed to modify IrrE in S. cerevisiae. As described in the part of methods, the IrrE library of about 105 mutants were generated. The transformants were initially selected in the 96-well plates based on strain growth in SC-Ura medium with FAP (Fig. 1). The growth rate of the isolated mutants was further verified in the tube and flask (Fig. 1).
Accordingly, in the first round of mutagenesis and selection, two mutants BY4742/I12 and BY4742/I24, were isolated. Compared to the strain BY4742/pRS416 and the strain BY4742/IrrE, the significant growth increase of the two mutants was observed in the presence of FAP (Fig. 2). Meanwhile, strain BY4742/I24 exhibited higher FAP tolerance than the strain BY4742/I12 (Fig. 2). To further improve performance, the mutant I24 from strain BY4742/I24 was used as the template for the second round of directed evolution, and strain BY4742/I37 was obtained. Compared to the strain BY4742/I24, the growth ability of strain BY4742/I37 under FAP stress was also slightly improved (Fig. 2). The experiment reflected the potential role of IrrE in regulating the transcriptional level of genes in S. cerevisiae.
Evolved IrrE to improve strain tolerance to the mixture of multiple inhibitors
As the top level of the hierarchical regulatory network in microorganisms, the global regulators control strain phenotypes by regulating large numbers of genes expression. The engineering of global regulators has been proved to be a highly efficient approach conferring cells desired complex phenotypes in both prokaryotic and eukaryotic microbes[15, 35]. This study has evolved IrrE to improve strain tolerance to the mixture of multiple inhibitors. The fermentation ability of the IrrE mutants was then comparatively analyzed in the SC-Ura medium supplemented with or without multiple inhibitors (0.8 g/L furfural, 3.0 g/L acetic acid and 0.3 g/L phenol). In the absence of multiple inhibitors, the heterologous expression of IrrE in S. cerevisiae slightly weakened strain fermentation. The strain BY4742/pRS416 finished the fermentation at about 15 h, while 8.5 g/L glucose was still left in the culture of strain BY4742/IrrE (Fig. 3a). A slight decrease in the biomass concentration, specific growth rate, glucose consumption rate, the final ethanol titer, ethanol productivity and ethanol yield were also observed in the strain BY4742/IrrE (Table 1). These results indicated the extra metabolic burden of expressing heterologous IrrE gene to S. cerevisiae. However, the mutation of IrrE significantly increased strain biomass yield, glucose consumption rate, specific growth rate, the final ethanol titer, ethanol productivity and ethanol yield (Fig. 3a, Table 1).
Table 1
Fermentation parameters for the strains carrying the improved IrrE in the presence and absence of FAP. Results represent the mean of duplicate experiments.
FAP
|
Strains
|
µ (h− 1)
|
rglu (g/L/h)
|
Ethanol titer (g/L)
|
reth (g/L/h)
|
Ethanol yield (%)
|
-
|
BY4742/pRS416
|
0.51
|
1.11
|
8.27
|
0.39
|
82.6
|
BY4742/IrrE
|
0.45
|
0.95
|
7.97
|
0.38
|
79.6
|
BY4742/I12
|
0.64
|
1.34
|
8.48
|
0.71
|
84.7
|
BY4742/I24
|
0.69
|
1.34
|
8.62
|
0.72
|
86.1
|
BY4742/I37
|
0.67
|
1.34
|
8.57
|
0.71
|
85.6
|
+
|
BY4742/pRS416
|
0.079
|
0.13
|
8.65
|
0.057
|
86.4
|
BY4742/IrrE
|
0.075
|
0.17
|
8.27
|
0.054
|
82.6
|
BY4742/I12
|
0.073
|
0.42
|
8.70
|
0.18
|
86.1
|
BY4742/I24
|
0.089
|
0.48
|
8.79
|
0.18
|
87.0
|
BY4742/I37
|
0.10
|
0.61
|
8.80
|
0.18
|
87.1
|
When the mixture of three inhibitors was added in SC-Ura medium, the fermentation performance of five strains was all significantly affected. The second-round mutant strain BY4742/I37 grew into the stationary phase and exhausted all the glucose at about 33 h (Fig. 3b). The first-round mutants BY4742/I12 and BY4742/I24 finished the fermentation within 48 h and 42 h respectively (Fig. 3b). In contrast, the fermentation of strain BY4742/IrrE was extended to approximately 115 h, while strain BY4742/pRS416 was still in the early exponential phase at that time and finished the fermentation until 153 h. The glucose consumption rate and the ethanol productivity were in parallel with the growth rate (Fig. 3b). The mutant BY4742/I37 exhibited the highest glucose consumption rate of 0.61 g/L/h, which was about 4.7-fold and 3.6-fold of the strain BY4742/pRS416 and strain BY4742/IrrE respectively (Table 1). The ethanol productivities of the three mutants were all 0.18 g/L/h, which were about 3.2-fold and 3.3-fold of the strain BY4742/pRS416 and strain BY4742/IrrE respectively (Table 1). The above data demonstrated that although the ethanol productivity of the mutants was markedly increased, the five strains generated similar ethanol yield in the presence of multiple inhibitors.
Sequence-based analysis of the mutants
Based on the above results, the great tolerance enhancement occurred after two rounds of error-prone PCR. Thus, the three IrrE mutants, I12, I24 and I37 screened from these two rounds were sequenced, and the resultant amino acid sequences were aligned with the wild type IrrE (Fig. 4a and 4b). Mutant I12 differed from IrrE at seven amino acid loci, while mutant I24 and mutant I37 had four mutations and nine mutations respectively. According to the mutations identified in I12, I24 and I37, single-site mutants were generated to investigate their effect on yeast tolerance to multiple FAP (Fig. 4c). For the mutants from I12, I103T, S133R, P162S and V204A showed a negative impact on strain growth under FAP-free condition. When the multiple inhibitors were added, the growth of strain I103T, P162S and V204A exhibited the moderate improvement compared to that of strain BY4742/IrrE, and the mutation V299A and A300V significantly increased strain growth ability, while M74T and S133R had only a little effect on strain sensitivity to FAP. Compared to the mutant I12, A300V was more tolerant to FAP stress, while I103T, P162S and V299A appeared more sensitive. For the four mutants from I24, the growth of E119V and L160F was decreased compared to strain BY4742/IrrE in the FAP-free medium. Under the FAP stress condition, the growth of A52E was rarely affected compared to that of the strain BY4742/IrrE. In contrast, E119V, L160F and R244G all rescued strain growth with inhibitors. Meanwhile, these three mutants appeared more sensitive to FAP than the mutant I24, suggesting that these three mutants work synergistically. In addition to the same four mutants as I24 has, I37 has five additional mutants and their growth was all decreased compared to strain BY4742/IrrE in FAP-free medium. M169V, E271K and Base 824 Deletion mutants showed a little increased tolerance compared to the control strain under FAP stress conditions, while the mutant L65P showed a moderately enhanced inhibitor tolerance. The improved tolerance of strain I37 may be the result of the combined action of nine site mutations. Through this comprehensive analysis, eleven mutants of IrrE protein, including L65P, I103T, E119V, L160F, P162S, M169V, V204A, R244G, Base 824 Deletion, V299A and A300V, were identified with an essential role in yeast FAP tolerance. As the reported structure of Deinococcus deserti IrrE, three domains are identified, including one zinc peptidase-like domain in the N terminal domain, one helix-turn-helix motif in the middle domain, and one GAF-like domain in the C-terminal domain[36]. Based on the sequence alignment and the homology model, the domain boundaries of D. radiodurans IrrE were determined as follows: the N-terminal domain covering residues1-161; the middle domain covering residues 162–203; and the C-terminal domain covering residues 204–328[21]. Through the mutant analysis, Vujicic-Zagar et al. indicate that the first and third domains are also critical regions for radiotolerance in strain D. deserti[36]. In this work, except for P162S and M169V, other mutations identified in the strain BY4742/I12, BY4742/I24 and BY4742/I37 all presented at the first and third domains, suggesting the critical role of these two domains in regulating S. cerevisiae to tolerant FAP stress.
Uncovering the global perturbation generated by IrrE in S. cerevisiae in response to the mixture of multiple inhibitors
To study the regulation mechanism of IrrE in S. cerevisiae cells in response to the mixture of multiple inhibitors, transcriptome sequencing and metabolite analysis were carried out. We examined the profiles of the two strains under normal growth conditions to study the genetic background differences between strain BY4742/IrrE and strain BY4742/pRS416. Samples in the middle of the lag phase of both strains under multiple inhibitors (0.8 g/L furfural, 3.0 g/L acetic acid and 0.3 g/L phenol) were collected for RNA-seq to compare the transcriptional profile changes between them. To identify genes related to IrrE regulation in response to FAP, we focused on genes that displayed more than 2-fold (log2foldchange > 1.0 and p value < 0.05) increase or decrease in treated BY4742/IrrE strain compared with treated BY4742/pRS416 strain but not in untreated BY4742/IrrE strain compared with untreated BY4742/pRS416 strain. Also, we tested changes of some intracellular metabolites to understand how they responded to FAP stress. IrrE may regulate the external defense and internal repair system to increase strain inhibitors tolerance.
Acetic acid, furfural, and phenol are the main components of inhibitors produced by lignocellulose pretreatment and have been proved to cause accumulation of reactive oxygen species (ROS) in yeast[10, 32]. As shown in Fig. 5a, the ROS level in untreated strain BY4742/IrrE was slightly lower than that in untreated strain BY4742/pRS416. However, the ROS level in strain BY4742/IrrE was reduced by about 61.4% compared to that in the control strain after exposure to FAP, despite the simultaneous increase of ROS levels in both strains. These results implied that the strain BY4742/IrrE suffered less ROS damage. The same protection role of IrrE was also verified in E. coli and Brassica napus under salt shock and in E. coli and A. simplex under ethanol shock[19, 20, 30, 37]. The transcription levels of the genes encoding enzymes related to ROS detoxification were upregulated in treated strain BY4742/IrrE (Additional file 1: Table S1). We further determined the activities of superoxide dismutases and catalases, and the results showed that these ROS scavenging enzymes exhibited higher activity in the strain BY4742/IrrE compared to strain BY4742/pRS416 in the presence of FAP (Fig. 5b and 5c). These results indicated that the IrrE might reduce the level of intracellular ROS by increasing the activities of the related enzymes.
NADPH is necessary for the process of furfural detoxification in S. cerevisiae[38]. Pentose phosphate pathway (PPP) is the primary pathway for the production of NADPH in cells and the transcriptome results showed that the PPP-related genes were upregulated in treated strain BY4742/IrrE (Fig. 6d), which is accordance with the higher tolerance to FAP of the strain BY4742/IrrE.
The gene hxk1 and err3 involved in the glycolysis pathway were upregulated in treated strain BY4742/IrrE, and the gene hxk1 is encoding the rate-limiting enzyme HXK1 in this pathway (Fig. 6c). Glycolysis is an energy-producing pathway under anaerobic fermentation conditions and ATP is the main form of energy in yeast cells. The detoxification of inhibitors is an energy-consuming process for yeast, and enhancement of energy supply seems to be necessary for yeast to resist FAP damage. The strain BY4742/IrrE showed a higher ATP yield than the strain BY4742/pRS416 in the presence of FAP (Fig. 5d).
The expression of genes related to trehalose and glycogen biosynthesis was upregulated in strain BY4742/IrrE under FAP stress (Fig. 6a and 6b). Trehalose act as a storage factor and stress protectant in yeast cells. Yeast can synthesize glycogen in response to the stress conditions. In consonance with the transcriptome analysis, the metabolite analysis showed that the trehalose content of the strain BY4742/pRS416 was 20.0% higher than that of the strain BY4742/IrrE under non-stress conditions. In comparison, the trehalose content of the strain BY4742/IrrE was boosted by 51.6% compared to that of the strain BY4742/pRS416 under FAP stress (Fig. 5e). In line with our results, the trehalose contents were significantly increased in E. coli strain with IrrE under osmotic or salt stress[23, 30]. To summarize, the accumulation of trehalose and glycogen was significantly involved in the mechanism of IrrE to improve strain tolerance. Moreover, the transcription of genes related to glycogen degradation was also upregulated (Fig. 6b), which suggests that the accumulated glycogen can be phosphorylated and then enter the glycolysis pathway to release energy.
Meanwhile, the transcriptome results showed that some sets of genes were also upregulated in treated strain BY4742/IrrE, including genes related to DNA repair, some transcription factors/activators, some genes encoding membrane proteins and transport proteins and some genes associated with the ribosome (Additional file 2: Table S4). These sets of genes should play essential roles in FAP stress resistance of strain with IrrE.
From the transcriptome results we can see that IrrE regulated 869 ((log2foldchange > 1.0 and p value < 0.05) genes in S. cerevisiae (Additional file 2: Figure S4), which were much more than the number of genes changed by mutated transcription factor Spt15p(132) and Taf25-3 in the application of gTME in S. cerevisiae under the unstressed conditions and oxidation stress respectively[15, 39]. As illustrated by Fig. 7, the global transcriptional factor IrrE may switch on diverse defense systems in yeast to resist FAP stress. ROS detoxification plays a vital role in enhancing yeast tolerance by reducing oxidative damage caused by FAP. Accumulated glycogen and trehalose act as pressure protectors to enhance yeast tolerance. At the same time, energy is stored in glycogen and released in the form of ATP via the glycolysis pathway for the utilization of energy demand pathways such as substance transport. NADPH produced by PPP can be used as a cofactor that is essential for the detoxification of inhibitors. Although we have analyzed the regulatory network of IrrE protein in yeast in response to inhibitors, the specific genes or proteins regulated by IrrE are still unclear and need further study.
The transcriptional regulation of I24 in yeast
Through the directed evolution method, three mutant strains with improved tolerance were obtained. Here the mutant strain BY4742/I24 was selected as a model to perform transcriptomic analysis and compared the profile with that of the control strain BY4742/pRS416. We first examined the patterns of the two strains under normal growth conditions. The results of transcriptome analysis proved that the protein I24 might accelerate the strain BY4742/I24 growth by modulating genes related to ergosterol biosynthesis and energy metabolism (Fig. 8). Sterols are significant sections of the cytomembrane and are concerned with strain resistance to hydrophobic molecules[40]. From Fig. 8 we can see that most of the genes associated with ergosterol biosynthesis were upregulated. Rapid cell growth requires sufficient energy supply. As the regulatory mechanism of IrrE, the transcriptome results of I24 also showed that some genes involved in the glycolysis pathway were upregulated and genes related to branch pathways of synthetic amino acids are down-regulated (Fig. 8a).
Like the analytical method of the strain BY4742/IrrE, this mutant group was then performed background removal. The results of transcriptome analysis showed that the protein I24 might enhance the tolerance of the strain BY4742/I24 by modulating genes in many aspects. Pma1 is necessary for maintaining cytosolic pH homeostasis and the electrochemical potential at the plasma membrane[41–43], and its encoding gene PMA1 achieved 2.56-fold time increase in treated strain BY4742/I24. The addition of amino acids enhances the tolerance of S. cerevisiae to ethanol and multiple inhibitors. In the meantime, the transcription levels of some amino acid permeases were also shown to be upregulated (Additional file 2: Table S5), which is in accordance with the higher tolerance to FAP of the strain BY4742/I24. Ribosome biogenesis is the core process of cell growth[44]. Many upregulated genes were enriched in the ribosome biogenesis set, which demonstrated that the rapid ribosome assembly in treated strain BY4742/I24 is to accommodate the protein processing required for rapid growth (Additional file 2: Figure S2). Furthermore, the protein I24 also enhanced strain tolerance by modulating some translation initiation factors and general stress response elements. Additional file 2: Table S5 showed that three transcription activators and two transcription factors were upregulated in treated strain BY4742/I24, among which MSN2 and MSN4 are two associated transcription activators and become active under many stress conditions[45, 46].
Therefore, the protein I24 may cause a wide range of perturbations by regulating the transcription level of transcription activators/factors, stabilizing the cell membrane and removing excess H+ to protect the intracellular environment, and enhancing the antioxidant capacity under inhibitor environment. It suggested that on one hand, directed evolution conferred the strain BY4742/I24 the different regulatory mechanisms; on the other hand, it reflected the plasticity of IrrE.
The effect of IrrE on strain thermotolerance
To explore whether the global regulator IrrE from prokaryotes could elicit the tolerance of yeast cells to other industrially relevant stress, the effect of the wild and mutant IrrE on strain thermotolerance was investigated. As shown in Fig. 9, the fermentation ability of the five strains was compared on growth, glucose consumption and ethanol production in SC-Ura medium at 38 ºC and 42 ºC. Under the heat stock of 38 ºC, all the strains were capable of exhausting glucose in the fermentation medium. Notably, the strains expressing wild or mutant IrrE exhibited higher specific growth rate, glucose consumption rate, the final ethanol titer, ethanol productivity and final biomass concentration than the control strain BY4742/pRS416 (Fig. 9a and Table 2). At 42 ºC, the cell growth, sugar consumption, the final ethanol titer, ethanol productivity and biomass production were all significantly affected no matter in the control strain or the four recombinant strains (Fig. 9b and Table 2). The growth of the control strain BY4742/pRS416 was almost completely suppressed. It had approximately 5 g/L glucose consumed and 1 g/L ethanol produced within 35 h, and subsequently no more glucose was utilized until 72 h. However, excellent fermentation advantage was observed in recombinant strains. The mutant strain BY4742/I24 and BY4742/I37 could grow into the stationary phase at about 21 h and approximate 15.5 g/L glucose was consumed in 45 h. The strain BY4742/I37 exhibited a moderate advantage in the maximum biomass, the final ethanol titer, ethanol productivity and the final ethanol yield than those in strain BY4742/I24. The strain BY4742/IrrE and BY4742/I12 exhausted 20 g/L glucose in 21 h and entered the stationary phase with the higher biomass and ethanol production. These results suggested that S. cerevisiae cells could be conferred enhanced tolerance against thermal stress through expressing IrrE or its mutants, which was consistent with the previous findings that IrrE could protect E. coli cells from thermal shocks[20].
Table 2
Fermentation parameters for the strains harboring the improved IrrE at different cultivation temperature. Results represent the mean of duplicate experiments.
FAP
|
Strains
|
µ (h− 1)
|
rglu (g/L/h)
|
Ethanol titer (g/L)
|
reth (g/L/h)
|
Ethanol yield (%)
|
38 ºC
|
BY4742/pRS416
|
0.44
|
1.32
|
8.52
|
0.57
|
85.0
|
BY4742/IrrE
|
0.56
|
1.5
|
8.73
|
0.58
|
87.1
|
BY4742/I12
|
0.78
|
1.64
|
8.84
|
0.74
|
88.3
|
BY4742/I24
|
0.73
|
1.65
|
8.90
|
0.74
|
88.9
|
BY4742/I37
|
0.76
|
1.66
|
8.76
|
0.73
|
87.5
|
42 ºC
|
BY4742/pRS416
|
0.03
|
0.21
|
1.33
|
0.038
|
13.3
|
BY4742/IrrE
|
0.162
|
0.89
|
8.34
|
0.14
|
83.3
|
BY4742/I12
|
0.163
|
0.92
|
8.90
|
0.15
|
88.9
|
BY4742/I24
|
0.106
|
0.32
|
6.26
|
0.11
|
62.5
|
BY4742/I37
|
0.119
|
0.34
|
7.05
|
0.12
|
70.4
|
We further evaluated the fermentation performance of the recombinant strains with the wild or mutant IrrE in the presence of multiple inhibitors under high temperature conditions. When the cultivation temperature was 38 ºC, the addition of FAP completely suppressed the growth of all five strains, which were still in the lag phase until 170 h. Rare glucose was consumed and little ethanol was produced (Additional file 2: Figure S3). Under the condition of 34 ºC, all strains exhibited better growth than those under the condition of 30 ºC in the absence of FAP (Fig. 10). Similarly, the biomass yield, the final ethanol titer, ethanol productivity, glucose consumption rate and specific growth rate of the three mutants was significantly increased compared to the strain BY4742/pRS416, while the slight metabolic burden was observed in strain BY4742/IrrE (Fig. 10). However, at 34 ºC, the presence of FAP remarkably enlarged the lag phase and reduced the glucose consumption rate and the ethanol productivity. The growth of strain BY4742/pRS416 was still in the lag phase until 170 h, while the strain BY4742/IrrE entered the exponential phase at approximately 135 h (Fig. 10). The strain BY4742/I37, which exhibited the best fermentation capacity, could grow into the stationary phase at about 80 h. The strain BY4742/I12 and BY4742/I24 finished the fermentation at about 125 h and 100 h respectively (Fig. 10). The glucose consumption rate and ethanol productivity were consistent with the strain growth (Fig. 10).
We found that the strain BY4742/IrrE exhibited better performance compared with the mutant strains with higher tolerance to FAP stress, suggesting that the mechanism referred to cellular tolerance to FAP differs from that for resistance to thermal stress. Meanwhile, the different stresses have a synergistically negative effect on cell survival. The increase of fermentation temperature made yeast cells face more severe challenges under the same concentration of FAP. The synergy of temperature and inhibitors was also observed during simultaneous saccharification and co-fermentation of pretreated corn stover[47]. However, the engineering of IrrE could achieve the purpose of enhancing the tolerance of yeast cells to two complex phenotypes (FAP tolerance and thermal tolerance) simultaneously. Further studies are ongoing to clarify the molecular mechanisms by which IrrE confers improved thermal tolerance in S. cerevisiae.