Using an enzyme cocktail formulation for lignocellulosic substrate hydrolysis, such as cellulases from P. echinulatum, A. niger, and T. reesei, is a well-known strategy for cellulose pretreatment in SSF. The Trichoderma cellulase mixture is one of the most commonly used enzymes in practical applications. However, it often requires a cellulase cocktail supplemented by BGL of Aspergillus for promotion of a synergistic effect to produce fermentable glucose from raw biomass material [11, 20, 21]. To design a practical enzyme cocktail formulation, expressing a group of cellulases with suitable proportions might be efficient for a specific cellulose substrate hydrolysis. In this study, an enzyme cocktail strategy was applied to choose several synergistic cellulases from different fungi and a designer concept was achieved for a cellulase cocktail production in a single host, using a synthetic biology technique.
Presently, degradation of lignocellulose requires large quantities of cellulases to release glucose from plant cell wall [2, 27]. The high cost of cellulases is the current bottleneck of biofuel industry. Moreover, the special culturing and induction condition required by a fungal host is a limiting step of traditional enzyme production technologies [7, 10]. In this study, to reduce the culturing time and cost, a yeast engineering approach was used to transform cellulase genes from different fungi into the KY3 host genome. In this study, a recombinant strain KR7 with five different cellulases from Trichoderma, Aspergillus, and Neocallimastix, was constructed. Although the specific FPA activity of the secreted enzymes of KR7 was still lower than the commercial enzymes of Celluclast 1.5 L from Trichoderma, this yeast system has many advantages as a cell factory for enzyme production, such as a high growth rate, enzyme secretion, and high temperature fermentation. However, it still needs further improvement before industry application. For example, elevation of the enzyme production could be achieved by improving the enzyme secretion ability with better secretion leader sequence  or by reforming the protein folding with related chaperon gene co-expression [29, 30]. Furthermore, the enzyme activity can be further improved by increasing the thermostability by addition of specific ions in the growth media .
Although KR5 possessed the cbhI gene of T. reesei, it could not efficiently hydrolyze the filter paper. There might be two possible reasons: First, as the cbhI gene in KR5 was regulated by the KlGapDH promoter, which was the weakest promoter we used with the PGASO method, KR5 might not have been able to express sufficient cellobiohydrolases for the filter paper hydrolysis. Second, the enzyme cocktail results showed that the addition of CBHI alone only slightly improved the cellulolytic ability of KR5, but the addition of CBHII resulted in a significant increase in the total FPA activity, probably because of the joint effect of CBHI and CBHII as KR5 already possessed the cbhI gene. This result is similar to a previous enzyme co-expression study , which suggested that the joint effect of different cellulases is important for cellulose hydrolysis. We will study this possibility in the future.
To improve the cellulose consumption ability of a host, the cellulose hydrolysis enzyme system should be able to coordinate with their cellulose hydrolyte transporting mechanisms. Although K. marxianus could use the lactose permease (Lac12) for cellobiose transport , it was not efficient for ethanol conversion. Other fungus, such as N. crassa, relies on a high-affinity cellodextrin transport system for rapid growth on cellulose . We have shown that the yeast strain KY3-NpaBGS-CDT, which co-expressed the cellodextrin transport (CDT-1) gene of N. crassa and the NpaBGS gene of Neo. patriciarum, could increase the growth rate under a cellodextrin substrate. Compared with KY3-NpaBGS, which possessed a single enzyme, KY3-NpaBGS-CDT grew slightly faster in the first five days, while KR5, which expressed a combination of cellulases, could digest more cellodextrins and showed a higher growth mass in the seven-day culture. KR7 provided a better solution to assemble the two functional features via the PGASO method in a single yeast for cellulolytic ethanol production.
Our CBP ethanol production study showed co-expressing exogenous fungal genes in a yeast host can convert cellulose to ethanol. Since the avicel was pretreated with steam sterilization in the autoclave, smaller cellulose fragments, such as cellodextrins, might have been released in the culture medium. However, there was no detectable reducing sugar in this medium. Compared to KR5, KR7 expressed the cdtI-gfp gene on the cell membrane, and also two additional cellulase genes (Figure 6). The presence of cellodextrin transporter might have helped the uptake of these smaller cellulose fragments in KR7 in the first three days of fermentation. When cellulases were secreted for cellulose digestion, the extra EglA and CBHII expressed in KR7 might have contributed to a higher cellulolytic efficiency than KR5 in the second and the third day of fermentation. The crystal avicel might have been gradually hydrolyzed by the synergistic cellulolytic enzyme reaction in the fourth day and the fifth day in KR7 (Figure 6). Although it was difficult to quantify the amount of substrate being degraded during fermentation with the saturated substrate level of 10% avicel, our study demonstrated that assembling an enzyme cocktail in an engineered yeast can improve the cellulosic ethanol productivity.
In the PGASO concept, each gene cassette contains the gene sequence linked, at the 5’ end, to a promoter sequence, and a sequence at the 3’ end of the gene cassette is identical to the 5’ end of the adjacent cassette. The promoter sequence in a gene cassette should be different from those of all other gene cassettes and the accuracy of gene cassette assembly was based on site-specific homologous recombination. In our previous study, the KR5 strain was derived from the transformation of five gene cassettes with a total length of ~15 Kb, and the procedure had 63% accuracy with the predesigned order assembly . In this study, the KR7 strain was derived from the transformation of seven gene cassettes with a total length of ~22 Kb and the accuracy was decreased to 3%. Moreover, the quantitative PCR data indicated that more copies of the cbhI, eglA, and npabgs gene cassettes were inserted than the other gene cassettes, probably via the non-homologous end-joining (NHEJ) pathway . NHEJ might cause an unanticipated result in genetic engineering. For example, the BGL activity in the supernatant of KR7 was 2.5-fold lower than KR5 (Figure 5B), which might be due to the 2-fold lower copies of the npabgs gene inserted in the KR7 genome . Since NHEJ is a fairly common phenomenon in Kluvyveromyces marxianus, random gene insertion and screening for activity might be a good strategy for host engineering . However, the random gene insertion approach is not convenient for transforming multiple genes into a host with desired gene copy numbers, and it is difficult to change the relative expression levels of different genes after they had been transformed. To increase the transformation accuracy of PGASO and to reduce the chance of an unanticipated result, we will try to reduce the NHEJ effect in KY3.
With K. marxianus KY3 as the host, the PGASO method has a high potential for practical applications. The fast growth rate and several other traits of KY3 make it desirable as a cell factory for commercial enzyme production. However, due to the saturated substrate level of 10% avicel, it was difficult to quantify the amount of substrate being degraded during the time course of fermentation of KR7. We measured the ethanol concentration at time zero of each culture in YP medium with avicel, and detected no significant ethanol residue. Also, a negative control was conducted by culturing KY3 in YP medium without avicel using OD 0.1 inoculums and OD 20 inoculums, and 0 g/l (OD 3) and roughly 0.3 g/l ethanol equivalent were detected after 5 days of culturing (data not shown). The ethanol equivalents and cell growth of KY3 might have been produced from the alternative metabolic pathway from peptone and amino acid in YP medium. Thus, our measurement of ethanol production was not accurate, but the data did show that KR7 could indeed convert avicel to ethanol, whereas KY3 and KR5 could not. Although this implies that <2% of the avicel was converted to ethanol under the current condition, it can be improved by elevating the reaction temperature or by other strategies. A better CBP fermentation parameter, such as SSR (solid-to-solution ratio), ORP (oxidation-reduction potential), and HRT (hydraulic retention time) will be considered to monitor the cellulose conversion rate in the future. Furthermore, to engineer K. marxianus KY3 to be a better CBP host for ethanol production, a number of issues still need to be resolved, such as expanding the variety of cocktail enzymes, increasing the number of gene cassettes, improving the pentose assimilating ability, decreasing the ethanol consumption rate and employing the cell surface engineering technology [24, 34]. However, from the engineering point of view, this cellulolytic efficiency of CBP is not as “efficiently” as SHF (separate hydrolysis and fermentation) and SSF (simultaneous saccharification and fermentation), but is more economical in the enzymes used and operational cost. Although KR7 can convert cellulose to ethanol in one step and has many potential characters as a CBP strain, it is not good enough for industry ethanol production. Thus, further improvement is being pursued. Increasing the knowledge of the biology and the genome of K. marxianus will facilitate its applications in industry [13, 35, 36].