Meiosis-Based Laboratory Evolution of the Thermal Tolerance in Kluyveromyces marxianus

Kluyveromyces marxianus is the fastest-growing eukaryote and a promising host for producing bioethanol and heterologous proteins. To perform a laboratory evolution of thermal tolerance in K. marxianus, diploid, triploid and tetraploid strains were constructed, respectively. Considering the genetic diversity caused by genetic recombination in meiosis, we established an iterative cycle of “diploid/polyploid - meiosis - selection of spores at high temperature” to screen thermotolerant strains. Results showed that the evolution of thermal tolerance in diploid strain was more efficient than that in triploid and tetraploid strains. The thermal tolerance of the progenies of diploid and triploid strains after a two-round screen was significantly improved than that after a one-round screen, while the thermal tolerance of the progenies after the one-round screen was better than that of the initial strain. After a two-round screen, the maximum tolerable temperature of Dip2-8, a progeny of diploid strain, was 3°C higher than that of the original strain. Whole-genome sequencing revealed nonsense mutations of PSR1 and PDE2 in the thermotolerant progenies. Deletion of either PSR1 or PDE2 in the original strain improved thermotolerance and two deletions displayed additive effects, suggesting PSR1 and PDE2 negatively regulated the thermotolerance of K. marxianus in parallel pathways. Therefore, the iterative cycle of “meiosis - spore screening” developed in this study provides an efficient way to perform the laboratory evolution of heat resistance in yeast.

A Novel Potent Carrier for Unconventional Protein Export in Ustilago maydis

Recombinant proteins are ubiquitously applied in fields like research, pharma, diagnostics or the chemical industry. To provide the full range of useful proteins, novel expression hosts need to be established for proteins that are not sufficiently produced by the standard platform organisms. Unconventional secretion in the fungal model Ustilago maydis is an attractive novel option for export of heterologous proteins without N-glycosylation using chitinase Cts1 as a carrier. Recently, a novel factor essential for unconventional Cts1 secretion termed Jps1 was identified. Here, we show that Jps1 is unconventionally secreted using a fusion to bacterial β-glucuronidase as an established reporter. Interestingly, the experiment also demonstrates that the protein functions as an alternative carrier for heterologous proteins, showing about 2-fold higher reporter activity than the Cts1 fusion in the supernatant. In addition, Jps1-mediated secretion even allowed for efficient export of functional firefly luciferase as a novel secretion target which could not be achieved with Cts1. As an application for a relevant pharmaceutical target, export of functional bi-specific synthetic nanobodies directed against the SARS-CoV2 spike protein was demonstrated. The establishment of an alternative efficient carrier thus constitutes an excellent expansion of the existing secretion platform.

Determination of Putative Vacuolar Proteases, PEP4 and PRB1 in a Novel Yeast Expression Host Meyerozyma guilliermondii Strain SO Using Bioinformatics Tools

Meyerozyma guilliermondii strain SO, a newly isolated yeast species from spoilt orange, has been used as a host to express the recombinant proteins using methylotrophic yeast promoters. However, as a novel yeast expression system, the vacuolar proteases of this yeast have not been determined, which may have contributed to the low level of heterologous protein secretions. Thus, this study aimed to determine intra- and extracellular proteolytic activity and identify the putative vacuolar proteases using bioinformatics techniques. A clear zone was observed from the nutrient agar skimmed milk screening plate. Proteolytic activity of 117.30 U/ml and 75 U/ml were obtained after 72 h of cultivation for both extracellular and intracellular proteins, respectively. Next, the Hidden Markov model (HMM) was used to detect the presence of the vacuolar proteases (PEP4 and PRB1) from the strain SO proteome. Aspartyl protease (PEP4) with 97.55% identity to Meyerozyma sp. JA9 and a serine protease (PRB1) with 70.91% identity to Candida albicans were revealed. The homology with other yeast vacuolar proteases was confirmed via evolutionary analysis. PROSPER tool prediction of cleavage sites postulated that PEP4 and PRB1 might have caused proteolysis of heterologous proteins in strain SO. In conclusion, two putative vacuolar proteases (PEP4 and PRB1) were successfully identified in strain SO. Further characterization can be done to understand their specific properties, and their effects on heterologous protein expression can be conducted via genome editing.

Thermococcus sp. KS-1 PPIase as a fusion partner improving soluble production of aromatic amino acid decarboxylase

Peptidyl-prolyl cis-trans isomerase (PPIase, EC 5.2.1.8) catalyzes the racemization reaction of proline residues on a polypeptide chain. This enzyme is also known to function as a molecular chaperon to stabilize protein conformation during the folding process. In this study, we noted FK506 binding protein (FKBP)-type PPIase from a hyperthemophilic archaeon Thermococcus sp. strain KS-1 (PPIase KS−1) to improve the solubility of Pseudomonas putida aromatic amino acid decarboxylase (AADC) that is an indispensable enzyme for fermentative production of plant isoquinoline alkaloids. AADC fused N-terminally with the PPIase KS−1 (PPIase KS−1-AADC), which was synthesized utilizing Escherichia coli host, showed improved solubility and, consequently, the cell-free extract from the recombinant strain exhibited 2.6- to 3.4-fold elevated AADC activity than that from the control strain that expressed the AADC gene without PPIase KS−1. On the other hand, its thermostability was slightly decreased by fusing PPIase KS−1. The recombinant E. coli cells expressing the PPIase KS−1-AADC gene produced dopamine and phenylethylamine from L-dopa and phenylalanine by two- and threefold faster, respectively, as compared with the control strain. We further demonstrated that the efficacy of PPIase KS−1-AADC in solubility and activity enhancement was a little but obviously higher than that of AADC fused N-terminally with NusA protein, which has been assumed to be the most effective protein solubilizer. These results suggest that PPIase KS−1 can be used as one of the best choices for producing heterologous proteins as active forms in E. coli.

Characterization and modulation of endoplasmic reticulum stress response target genes in Kluyveromyces marxianus to improve secretory expressions of heterologous proteins

Background Kluyveromyces marxianus is a promising cell factory for producing bioethanol and that raised a demand for a high yield of heterologous proteins in this species. Expressions of heterologous proteins usually lead to the accumulation of misfolded or unfolded proteins in the lumen of the endoplasmic reticulum (ER) and then cause ER stress. To cope with this problem, a group of ER stress response target genes (ESRTs) are induced, mainly through a signaling network called unfolded protein response (UPR). Characterization and modulation of ESRTs direct the optimization of heterologous expressions. However, ESRTs in K. marxianus have not been identified so far. Results In this study, we characterized the ER stress response in K. marxianus for the first time, by using two ER stress-inducing reagents, dithiothreitol (DTT) and tunicamycin (TM). Results showed that the Kar2–Ire1–Hac1 pathway of UPR is well conserved in K. marxianus. About 15% and 6% of genes were upregulated during treatment of DTT and TM, respectively. A total of 115 upregulated genes were characterized as ESRTs, among which 97 genes were identified as UPR target genes and 37 UPR target genes contained UPR elements in their promoters. Genes related to carbohydrate metabolic process and actin filament organization were identified as new types of UPR target genes. A total of 102 ESRTs were overexpressed separately in plasmids and their effects on productions of two different lignocellulolytic enzymes were systematically evaluated. Overexpressing genes involved in carbohydrate metabolism, including PDC1, PGK and VID28, overexpressing a chaperone gene CAJ1 or overexpressing a reductase gene MET13 substantially improved secretion expressions of heterologous proteins. Meanwhile, overexpressing a novel gene, KLMA_50479 (named ESR1), as well as overexpressing genes involved in ER-associated protein degradation (ERAD), including HRD3, USA1 andYET3, reduced the secretory expressions. ESR1 and the aforementioned ERAD genes were deleted from the genome. Resultant mutants, except the yet3Δ mutant, substantially improved secretions of three different heterologous proteins. During the fed-batch fermentation, extracellular activities of an endoxylanase and a glucanase in hrd3Δ cells improved by 43% and 28%, respectively, compared to those in wild-type cells. Conclusions Our results unveil the transcriptional scope of the ER stress response in K. marxianus and suggest efficient ways to improve productions of heterologous proteins by manipulating expressions of ESRTs.

Heterologous production of active form of beta-lytic protease by Bacillus subtilis and improvement of staphylolytic activity by protein engineering

Background Most of the proteases classified into the M23 family in the MEROPS database exhibit staphylolytic activity and have potential as antibacterial agents. The M23 family is further classified into two subfamilies, M23A and M23B. Proteases of the M23A subfamily are thought to lack the capacity for self-maturation by auto-processing of a propeptide, which has been a challenge in heterologous production and application research. In this study, we investigated the heterologous expression, in Bacillus subtilis, of the Lysobacter enzymogenes beta-lytic protease (BLP), a member of the M23A subfamily. Results We found that B. subtilis can produce BLP in its active form. Two points were shown to be important for the production of BLP in B. subtilis. The first was that the extracellular proteases produced by the B. subtilis host are essential for BLP maturation. When the host strain was deficient in nine extracellular proteases, pro-BLP accumulated in the supernatant. This observation suggested that BLP lacks the capacity for self-maturation and that some protease from B. subtilis contributes to the cleavage of the propeptide of BLP. The second point was that the thiol-disulfide oxidoreductases BdbDC of the B. subtilis host are required for efficient secretory production of BLP. We infer that intramolecular disulfide bonds play an important role in the formation of the correct BLP conformation during secretion. We also achieved efficient protein engineering of BLP by utilizing the secretory expression system in B. subtilis. Saturation mutagenesis of Gln116 resulted in a Q116H mutant with enhanced staphylolytic activity. The minimum bactericidal concentration (MBC) of the wild-type BLP and the Q116H mutant against Staphylococcus aureus NCTC8325 was 0.75 μg/mL and 0.375 μg/mL, respectively, and the MBC against Staphylococcus aureus ATCC43300 was 6 μg/mL and 3 μg/mL, respectively. Conclusions In this study, we succeeded in the secretory production of BLP in B. subtilis. To our knowledge, this work is the first report of the successful heterologous production of BLP in its active form, which opens up the possibility of industrial use of BLP. In addition, this study proposes a new strategy of using the extracellular proteases of B. subtilis for the maturation of heterologous proteins.

From sequence to yield: deep learning for protein production systems

Recent progress in laboratory automation has enabled rapid and large-scale characterization of strains engineered to express heterologous proteins, paving the way for the use of machine learning to optimize production phenotypes. The ability to predict protein expression from DNA sequence promises to deliver large efficiency gains and reduced costs for strain design. Yet it remains unclear which models are best suited for this task or what is the size of training data required for accurate prediction. Here we trained and compared thousands of predictive models of protein expression from sequence, using a large screen of Escherichia coli strains with varying levels of GFP expression. We consider models of increasing complexity, from linear regressors to convolutional neural networks, trained on datasets of variable size and sequence diversity. Our results highlight trade-offs between prediction accuracy, data diversity, and DNA encoding methods. We provide robust evidence that deep neural networks can outperform classic models with the same amount of training data, achieving prediction accuracy over 80% when trained on approximately 2,000 sequences. Using techniques from Explainable AI, we show that deep learning models capture sequence elements that are known to correlate with expression, such as the stability of mRNA secondary structure. Our results lay the groundwork for the more widespread adoption of deep learning for strain engineering across the biotechnology sector.

Increased expression of recombinant chitosanase by co-expression of Hac1p in the yeast Pichia pastoris.

Background: Pichia pastoris is one of the most popular eukaryotic hosts for producing heterologous proteins, while increasing secretion of target proteins is still a top priority for their application in industrial fields. Recently, the research effort to enhance protein production therein has focused on up-regulating the unfolded protein response (UPR). Objective: We evaluated the effects of activated UPR via Hac1p co-expression with the promoter AOX1 (PAOX1) or GAP (PGAP) on expression of recombinant chitosanase (rCBS) in P. pastoris. Method: The DNA sequence encoding the chitosanase was chemically synthesized and cloned into pPICZαA and the resulted pPICZαA/rCBS was transformed into P. pastoris for expressing rCBS. The P. pastoris HAC1i cDNA was chemically synthesized and cloned into pPIC3.5K to give pPIC3.5K/Hac1p. The HAC1i cDNA was cloned into pGAPZB and then inserted with HIS4 gene from pAO815 to construct the vector pGAPZB/Hac1p/HIS4. For co-expression of Hac1p, the two plasmids pPIC3.5K/Hac1p and pGAPZB/Hac1p/HIS4 were transformed into P. pastoris harboring the CBS gene. The rCBS was assessed based on chitosanase activity and analyzed by SDS-PAGE. The enhanced Kar2p was detected with western blotting to evaluate UPR. Results: Hac1p co-expression with PAOX1 enhanced rCBS secretion by 41% at 28°C. Although the level of UPR resulted from Hac1p co-expression with PAOX1 was equivalent to that with PGAP in terms of the quantity of Kar2p (a hallmark of the UPR), substitution of PGAP for PAOX1 further increased rCBS production by 21%. The methanol-utilizing phenotype of P. pastoris did not affect rCBS secretion with co-expression of Hac1p or not. Finally, Hac1p co-expression with PAOX1 or PGAP promoted rCBS secretion from 22 to 30°C and raised the optimum induction temperature. Conclusion: The study indicated that Hac1p co-expression with PAOX1 or PGAP is an effective strategy to trigger UPR of P. pastoris and a feasible means for improving production of rCBS therein.