2.1. Adaptive laboratory evolution of recombinant strain Y. lipolytica XYL+
2.1.1. Setting up the right conditions for evolution
To perform D-glucose analog 2-deoxyglucose (hereafter abbreviated as dG) tolerance evolution to the yl-XYL+, we firstly test the toxicity of dG to confirm the initial condition of evolution. dG is toxic to yl-XYL+ in both YNBdG10 medium (dG as sole carbon source) and YNBdG10X10 medium (with dG and xylose as carbon sources). It caused 88.8% and 100% death of yl-XYL+ in YNBdG10 medium within 24 and 32 h, respectively, while in YNBdG10X10 medium, it resulted in 50.9% and 58.9% death (Fig.2A). This suggested that dG blocked xylose transportation or metabolism, but it was not totally lethal to yl-XYL+, which made it possible to drive yl-XYL+ evolution to consume xylose in YNBdG10X10 medium.
After confirming that dG inhibited the xylose metabolism of the yl-XYL+, a small amount of D-glucose was added to the YNBdGxX10 medium to help yl-XYL+ grow to a baseline biomass so that it can evolve to uptake xylose in the presentence of dG. Impacts of different D-glucose addition were investigated in Fig.2B, it was found that 2 g/L of D-glucose can ensure that the yl-XYL+’s growths to about 1.0 of OD600 within 24 h, which basically satisfied a baseline biomass requirement. Then we investigated how much xylose could be consuming within 24 h when the baseline biomass was set as 1.0 of OD600. We found that the original strain could consume 10 g/L xylose within 24 h with an initial OD600 at 1.0 (Fig.2C). To further confirm this result, we tested the consumption of 10 g/L xylose in 24 h with different initial inoculum size (Fig.2D). It was also found that baseline biomass at 1.0 of OD600 could consume 10 g/L xylose, while less inoculum could not. Therefore, the initial conditions of evolution were set as follows: the concentrations of dG, D-glucose, and xylose were 1, 2, and 10 g/L in the YNB medium, respectively; the evolution period was 48 h (determined after tests), in which, the first 24 hours were assumed to be a proliferative period (using D-glucose as carbon source) and the second 24 hours were assumed to be the evolution period (using xylose as sole carbon source in the presentence of dG).
2.1.2. Adaptive laboratory evolution selects strains with the capacity of consuming xylose
Based on the above initial conditions, we started an adaptive laboratory evolution of yl-XYL+. The evolution went through four stages for a total of 64 days, as shown in Fig.3A. It shows the measured biomass (OD600) and residual xylose in YNBD2dGxX10 medium. When residual xylose concentration is less than 0.5 g/L, an evolution stage is ended. The evolution conditions remain unchanged except dG concentration increasing from 1 g/L to 2, 5 and 10 g/L in stage 1, 2, 3 and 4. In stage 1, overall increasing biomass and decreasing residual xylose are obversed, which implies that the strains have evolved to utilize xylose at the presentence of 1 g/L of dG. Evolution of stage 1 completed on 20th days and the culture populations from the last batch were harvested and spread on solidified medium for clones selection and phenotype test. Strain yl-XYL+*01*3 with best growth and xylose consumption phenotype was selected from ten randomly colonies picked up from the plate (Additional file 1: Figure S1A), and was used for subsequent evolution. Strains yl-XYL+*02*9, yl-XYL+*03*5 and yl-XYL+*04*10 were obtained in the same way (Additional file 1: Figure S1B, C and D).
2.1.3. Qualitative characterization of domesticated strains
At the end of each round of domestication, one of the best colonies was selected for the next round of domestication. After 4 rounds of domestication, we obtained 4 domesticated strains, which have a closely derivative relationship with high values to be characterized and compared with each other. dG tolerance of the domesticated and control strains was evaluated. The domesticated strain and the control strain were spread on various solid mediums including YNBdG1, YNBdG1X10, YNBD10X10 and YNBX10, respectively (Additional file 1: Figure S2A). It is obvious that the evolved strains can grow on YNBdG1X10 medium, but the original strain yl-XYL+ cannot. As expected, strains with stronger dG resistance showed better growth phenotype in YNBdG1/X10 medium. Morphology of parent and evolved strains in optical microscope was also investigated, but no obvious difference was observed (Additional file 1: Figure S2B).
Since the domesticated strain can grow well with xylose as carbon source in the presence of dG, we examined the fermentation performance of the domesticated strain in YNBD15X15 medium. The samples were taken at 24 h and 36 h. As shown in Fig.3B, the evolved strains can utilize significant amounts of xylose in the presentence of D-glucose, the parental strain barely uses it. Nonetheless, the level of D-glucose utilization seems to be impaired in all evolved strains. Overall, all the evolved strains can utilize xylose in the presentence of D-glucose or dG, which indicated that they can ferment D-glucose and xylose simultaneously. In addition, the more evolved strains performed better than the less evolved ones.
2.2. Evolved strains can simultaneously uptake D-glucose and xylose
2.2.1. Fermentation phenotype
In order to understand the fermentation performance of the domesticated strain, the strains yl-XYL+, yl-XYL+*01*3 and yl-XYL+*04*10 were grown in different medium in order to systematically investigated xylose and D-glucose consumption (Fig.4). In pure xylose medium, xylose utilizations and biomass of all strains were almost identical (Fig.4B, E and H). With D-glucose as sole carbon source, yl-XYL+*01*3 showed similar phenotype with the original strain yl-XYL+, but yl-XYL+*04*10 exhibited 10.19% and 9.78% decrease compared to yl-XYL+ and yl-XYL+*01*3 in the first 12 h, respectively (Fig.4A, D and G). Similar to glucose utilization, the final biomass production of yl-XYL+*04*10 exhibited 13.9% and 8.7% decrease compared to yl-XYL+ and yl-XYL+*01*3, respectively (Fig.4A, D and G).
When the mixed sugars were fermented, all domesticated strains were able to utilize xylose in the presentence of D-glucose at concentration over 25 g/L, while the original strain yl-XYL+ could not until D-glucose was exhausted. As shown in the third column of Fig.4C, F and I, it is obvious that yl-XYL+*04*10 showed better co-fermentation performance than yl-XYL+*01*3, and yl-XYL+*01*3 performed much better than yl-XYL+. Interestingly, in the first 24 h, average xylose utilization rates of the evolved strains in the mixed sugar fermentation were enhanced by 131% and 105% compared to the original strain, but their D-glucose utilization rate was decreased by 28% and 24.6%. This suggests that the simultaneous fermentation might be achieved in a manner that weakened D-glucose uptake but enhanced xylose utilization.
2.2.2. Xylose uptake kinetics in the presentence of dG and D-glucose
In order to investigate whether the xylose uptake became stronger and the D-glucose uptake became weaker during the fermentation of the mixed sugar, we carried out kinetic studies in both YNBdGxXx medium and YNBDxXx medium, as depicted in Table 1 and Fig.5. 1 g/L of dG almost completely inhibited xylose uptake of yl-XYL+ (Fig.5A); while for yl-XYL+*01*3, 1 g/L of dG had no effect. 2 g/L of dG only inhibited less than 30% of its xylose uptake rate. 5 g/L and 10 g/L dG severely inhibited and even completely eliminated xylose uptake (Fig.5B); As for yl-XYL+*04*10, 5 g/L and 10 g/L of dG inhibited the uptake rate by 53.6% and 56.3% respectively when using 30 g/L of xylose (Fig.5C). Inhibition of xylose uptake by D-glucose on parent and evolved strains show a similar pattern. 1 g/L D-glucose can completely inhibit xylose uptake of yl-XYL+, which explained why yl-XYL+ cannot utilize xylose in the presentence of D-glucose (Fig. 5D). 10 g/L D-glucose can inhibit xylose uptake rate of yl-XYL+*04*10 by 69.9% when using 30 g/L of xylose (Fig.5F). Generally, inhibition of xylose uptake by dG and D-glucose are concentration dependent. As the concentration of dG or D-glucose increases, the inhibition is significantly enhanced. Besides, inhibition of xylose uptake by D-glucose is more severe than that by dG.
Through the fermentation phenotype and xylose uptake kinetics analysis, we basically assumed that D-glucose/xylose co-fermentation phenotype of the evolved strains was closely related to xylose uptake enhancement (or weakened inhibition by D-glucose) and lower capacity to utilize D-glucose alone, but the genotype changes responsible for the phenotype is still not clear.
2.3. RNA-seq analysis
RNA-seq was performed to investigate the differences in global gene expression between strain yl-XYL+ and yl-XYL+*04*10. RNA was obtained after 6 h of fermentation in the D-glucose medium and D-glucose/xylose medium respectively. The number of genes significantly regulated (up-regulated by more than 2-fold or decreased by at least 50%) in yl-XYL+*04*10 in D-glucose and mixed sugar medium was 56 and 12, respectively; The ten most upregulated and downregulated genes in yl-XYL+*04*10 are listed in Additional file 1: Table S2 and S3, respectively. For instance, the transcript levels of genes encoding taurine dioxygenase and malate synthase (YALI0_A21439g and YALI0_D19140g, respectively), were 13.1- and 8.2-fold higher in yl-XYL+*04*10. Transcription of YALI0_A21439g and YALI0_C23452g, which encodes a protein of unknown function involved in serine/threonine metabolism, was enhanced by 11.3- and 5.0-fold. Besides, transcript levels of YALI0_D01111g encoding a D-glucose transporter decreased by 63%.
To figure out the putative genes directly responsible to D-glucose/xylose co-utilization, transcriptional level of genes involved in D-glucose and pentose metabolism were carefully investigated and the transcriptional profiling was shown along with metabolic pathway (Fig.6).
Generally, genes involved in hexose utilization including transport and catabolism was down-regulated, among which, genes YALI0_E20427g and YALI0_E20207g were even down-regulated by 67.9% and 35.2%, respectively (Fig.7A, E4G_vs_C0G). It explains to some extent why the rate of D-glucose utilization is reduced in mixed-sugar fermentation. On the contrary, several genes encoding pentose-specific transporters were dramatically up-regulated. For example, gene YALI0_D00363g and YALI0_C04730g were up-regulated by 60% and 3.52-fold, respectively (Fig. 7C and D). Since YALI0_C04730g is also annotated as an arabinose-specific transporter, the fermentation phenotype of original and evolved strains in arabinose medium was investigated (Additional file 1: Figure S3). We found that all strains cannot grow with arabinose with sole carbon source. However, the evolved strains can uptake arabinose in presentence of D-glucose, and yl-XYL+*04*10 can even uptake 12.2 g/L arabinose within 48 h in YPD15A15, while yl-XYL+ cannot (Additional file 1: Figure S3C).
Interestingly, not all the genes annotated as pentose-specific transporters were up-regulated, of which, YALI0_B00396g and YALI0_D01111g were down-regulated by 47.6% and 63% (Fig.7E and F, E4GX_vs_C0GX). This could suggest a misannotation of these transporter and further characterization of their real function in vivo would be required to verify this. Besides, genes involved in pentose phosphate pathway (PPP) such as tkt (YALI0_E06479g, encoding transketolase) and tal (YALI0_F15587g, encoding transaldolase) did not change expression level very much, which is unexpected. It seems that the major responsible changes for the improved uptake of xylose are at the level of transporters efficiency rather than PPP.
RNA-seq did not only reveal global changes in gene expression levels in evolved strains compared to the parent strain, but also helped us to identify meaningful mutations in some key genes (including base insertion, deletion or substitution) (Additional file 1: Table S4). For example, a single nucleotide substitution occurred in gene YALI0_E23287g encoding a D-glucose transporter with adenine (A) replacing cytosine (C) at base 1118, which resulted in an amino acid residue change of N373T. Another important point mutation was found in YALI0_E15488g encoding a hexokinase with cytosine (C) replacing guanine (G) at base 809. The single base substitution lead to a codon change at amino acid 270 from glycine to alanine (G270A). We confirmed above point mutations using PCR amplification and subsequent DNA sequencing (Additional file 1: Figure S4). We hypothesize that such mutation could lead to a reduced use of D-glucose. In order to prove that, we then checked the hexokinase activity of the evolved strain yl-XYL+*04*10 and it resulted to be decreased by 73.4% and 74.4% in YPD30 medium and YPD30X30 medium respectively, compared to the original strain (Fig.8). This explains, at least to some extent, why the utilization of D-glucose is weakened in the evolved strain.