Repeatedly displaying IM7 proteins on the P. pastoris cell surface
In traditional yeast cell-surface display methods, the dockerin-cohesion pair from bacterial cellulosomes is adopted, in which the dockerin is roughly a 10 kDa calcium-binding module that non-covalently associates with the scaffoldin cohesion at affinity in the sub-nM (~ 10− 6 M) range [16]. In this work, the IM7/CL7 protein pair (Fig. 1) with a much higher binding affinity (KD ~10− 14-10− 17 M), was used instead [26]. In principle, the ultra-strong protein-protein interaction between IM7 and CL7 could significantly enhance the cellulosome assembly efficiency. Taking advantage of this IM7/CL7 system, we recently developed an indirect P. pastoris surface display method that can display ~ 2.8 × 106 enzyme molecules per yeast cell [24], observing a display efficiency ten times higher than that of traditional S. cerevisiae display methods. As shown in Fig. 1, the yeast surface anchor protein SED1 from S. cerevisiae without its signal sequence was fused to the repeated IM7 units. The surface localization of IM7 proteins was verified with immunofluorescence microscopy and flow cytometric analysis (FACS) in Fig. 2. As a control, the strain Y-IM0 without modification was not immunostained, whereas the variant Y-IM1, Y-IM2 and Y-IM3 were all in green color in the presence of mouse anti-HA monoclonal antibodies and FITC- conjugated goat anti-mouse antibodies. These results indicated that all recombinant yeasts displayed the IM7 proteins on the surface, and the display efficiency was elevated when increasing IM7 repeats. In addition, the engineered yeasts were all in red color in the presence of CL7 tagged mCherry fluorescent proteins (Fig. 3). The FACS data (Fig. 3) further demonstrated that repeatedly anchoring the IM7 proteins enhanced the display efficiency of CL7-mCherry. Compared with other yeast surface display systems, in which less than 50% of cells were positively stained by immunofluorescence [18], our yeast surface display system has a much better display level of over 90%, possibly owning to the ultra-strong protein-protein interaction between CL7 and IM7.
In vitro assembling minicelluolosome on the P. pastoris cell surface
Previously, researchers assembled functional minicellulosome in vitro on the S. cerevisiae yeast cell surface by incubation of the engineered yeast that has a chimeric scaffoldin [9, 10] or two miniscaffoldins [20, 21] with exogenous recombinant cellulases. In comparison to in vivo assembling mode, the in vitro assembled cellulosome showed higher bioethanol production, possibly because the metabolic load was lowered in these strains. In this work, we also chose the in vitro assembly strategy to construct minicelluolosomes on the P. pastoris cell surface. Specially, the different cellulases, including CBH from Yarrowia lipolytica, EG from Clostridium thermocellum DSM1237, BGL from Thermoanaerobacterium thermosaccharolyticum DSM 571, and CBD from Thermobifida fusca, were fused with a N-terminal CL7 tag and recombinantly expressed from E. coli. Those purified cellulases were finally incubated with yeast strain Y-IM2 or Y-IM3.
To prove the successful construction of minicellulosome, Avicel, PASC (86.2), or CMC was utilized as the substrate for enzyme activity assay [31]. We adjusted the ratio of EG, CBH, BGL and CBD at 1:1:1:1, 2:4:2:7 and 1:3:6:10, respectively. Meanwhile, the free cellulases (1:1:1:1) were used as a control [31]. The data (Fig. 4) indicate that the enzyme activity of Y-IM2 and Y-IM3 is comparable to or higher than that of free cellulases. Interestingly, both minicellulosome and the free cellulases showed higher activity toward CMC and PASC than Avicel. However, the improvement of the enzyme activity causing by minicellulosome toward Avicel was more obviously (~ 2.6-fold). Based on these results, we chose 1:1:1:1 and 2:4:2:7 as the optimized ratios for the following ethanol fermentation experiments.
Lyophilization of the engineered P. pastoris as compound cellulases
The commercial cellulase is often supplied as the compound of three-type cellulases including EG, CBH and BGL. The annual consumption of cellulases is huge in various industrial fields [1], especially for breeding industry. As it is known to all, P. pastoris is considered as a GRAS (generally recognized as safe) microorganism by FDA and has been employed to produce diverse human peptides and proteins. In addition, P. pastoris has a strong cell wall and outer membrane in structure capable of serving as a stable biomaterial for enzyme immobilization. Inspired by these works, we thought that the above engineered P. pastoris with surface-display minicellulosome might be suitable for industrial applications. Initially, we tested if the induced but dead yeast with cellulosome has an enzyme activity to hydrolyze the cellulose. The results (Fig. S1) demonstrated that the catalytical activity was unchanged within at least three months when the yeast strains were stored at -20oC. Thereafter, the Y-IM2 or Y-IM3 variant with minicellulosome was lyophilized as powders for long-term storage. When recovered in solution, no cellulase activity changes were observed (Fig. S1), further proving that the lyophilized P. pastoris can be used as the compound cellulases. Most importantly, such kind of lyophilized P. pastoris with cellulosome can be rapidly produced at large scale and low cost, showing great commercial potential in industry.
Direct fermentation of the cellulose to ethanol
Direct ethanol fermentations from Avicel, PASC (86.2) or CMC was examined using E. coli lysates treated Y-IM2 and Y-IM3 (Fig. 5a and Fig. 5b). The data demonstrated that the ethanol titer quickly increased within 60 hours for all the substrates. Besides, Y-IM2 was better as the ethanol producer than Y-IM3 toward Avicel and PASC, with the highest ethanol yield of 2.5 g/L for Avicel and 1.2 g/L for PASC (Fig. 5c), respectively. In the previous S. cerevisiae fermentation studies, Avicel or PASC was always the better substrate than CMC, yet the average ethanol production was lower than 2 g/L. Our bioethanol production is higher than or at least comparable to those of works. Surprisingly, CMC was the best carbon source for both Y-IM2 and Y-IM3 variants, achieving a highest ethanol yield of 5.1 g/L. With an ultra-high viscosity, CMC was thought to weaken the diffusion of hydrolysis products and influence the S. cerevisiae fermentation. Very low ethanol production was observed when CMC was used in yeast CBP studies [18, 20, 21]. Comparingly, the P. pastoris has been found able to achieve a very high cell density in fermentation. Therefore, it may utilize the high-viscosity CMC for ethanol fermentation better than S. cerevisiae, which is consisted with the observations here. Meanwhile, the Y-IM0 variant without minicellulosome was found to have no ability to produce ethanol.