2.1 Gene mining and sequence analysis of PMGL-Ba
The gene mining process for PMGL involved using a high enzyme activity pectin lyase gene as the probe sequence, with key residues serving as the screening conditions. The selection criteria included assessing sequence similarity to the probe template, predicting soluble expression of the protein, and considering the type of microorganisms of origin. Ultimately, the identified putative PMGL-Ba (NCBI Reference Sequence: WP_009329358.1) derived from Bacillus licheniformis was finally selected. The protein consists of 494 amino acids with a relative molecular mass of 54693.2. PMGL-Ba was found to be in the same branch as two Bacillus-derived pectin lyases (GenBank: OMI04834.1, QII50179.1) annotated in NCBI, and both are in the polysaccharide lyase 1 family, confirming the reliability of the results(Fig. 1A). Analysis of the protein structure diagram revealed that PMGL-Ba possesses the typical β-helical structure of pectin lyase proteins, consisting of three parallel β-strands connected through three corners (Zheng et al. 2021) (Fig. 1B). Sequence comparisons indicated that certain sequences of PMGL-Ba overlap with conserved amino acid sequences of the pectin lyase reference sequence or contain key amino acids of the catalytic activity center, suggesting potential functional similarities (Fig. 1C).
Figure 1 Phylogenetic analysis of PMGL-Ba. (A) Selection of signal peptides. (B) The predicted structure of PMGL-Ba. Red represents α-helix, gold represents β-sheet, and green represents random coil structure. (C) Sequence comparison analysis.
2.2 Expression and biochemical characterization of PMGL-Ba
The shake flask fermentation of PMGL-Ba recombinant P. pastoris X33 was carried out, with samples taken at 24 h intervals to assess the growth of the recombinant yeast and the enzyme activity of pectin cleavage in the supernatant of the fermentation broth. With an increase in induced expression time, the OD600, fermentation broth supernatant enzyme activity, and protein yield of X33-PMGL-Ba gradually increased. At 144 h of induced expression time, the enzyme activity and protein concentration with esterified pectin as substrate reached 895.69 U/mL and 0.77 g/L, respectively, indicating the successful heterologous secretory expression of recombinant PMGL-Ba using P. pastoris X33(Fig. 2A). The purified PMGL-Ba was analyzed by SDS-PAGE, revealing an apparent molecular weight of about 55 kDa, consistent with the theoretical molecular weight of 54.6 kDa༈Fig. 2B).
The enzyme activity of PMGL-Ba exhibited an initial increase followed by a decrease with increasing pH, pinpointing the optimal reaction pH as 8.5 (Fig. 2C). PMGL-Ba retained more than 80% of its total activity within the pH range of 8.0-9.5. However, the enzyme activity decreased rapidly beyond pH 9.0 or below 8.0. PMGL-Ba has good pH stability, maintaining 60% activity in the pH range of 5.0–11.0(Fig. 2D). The optimal temperature for PMGL-Ba was 60°C, with relative enzyme activity exceeding 60% at temperatures between 50–70°C༈Fig. 2E). PMGL-Ba activity decreased with increasing temperature, and relative enzyme activity remained above 60% at temperatures below 70°C after 1 h of exposure. As the temperature increased further, the enzyme activity gradually decreased, with only about 20% of the enzyme activity remained at the temperature of 80 ℃༈Fig. 2F). PMGL-Ba showed robust enzyme activity and stability over a broad temperature range of 30 to 80°C, indicating its potential advantages for application in medium and high-temperature industries, such as the paper industry.
Figure 2 Expression and enzymatic properties of PMGL-Ba. (A) Fermentation of PMGL-Ba. (B) SDS-PAGE analysis. (C-F) Enzymatic properties of PMGL-Ba.
2.3 Effect of expression components on recombinant PMGL-Ba protein expression
The secretory expression of heterologous proteins in P. pastoris often involves the utilization of the S. cerevisiae α-factor mating pheromone (α-MF) from Saccharomyces cerevisiae. Existing studies have explored the impact of signaling peptides on the secretion of recombinant proteins, revealing varying efficiencies of signaling peptides for different exogenous proteins. Studies on the effect of signal peptides on exogenous protein secretion can be broadly categorized into two groups: modification of existing signal peptides and the discovery of new signal peptides. The commonly used α-MF comprises a two-stage signal, consisting of a 19-amino-acid pre region and a 66-amino-acid pre region at the N-terminal end, demonstrating effective secretion for most exogenous proteins (Barrero et al. 2018). Several studies have sought to enhance the secretion efficiency of α-MF various modifications, such as αMF-CC (αM), and αopt obtained after optimization and point mutation of the αMF sequence, resulting in improved protein secretion (Ahn et al. 2016; Aza et al. 2021; Ito et al. 2022). Another approach involves utilizing more efficient sequences to replace parts of the αMF. Juan J et al. designed combined secretion signals, such as the one consisting of the Saccharomyces cerevisiae Ost1 signal sequence and the α-factor PRO region, which has demonstrated superior secretion ability compared to the α-factor signal sequence (Barrero et al. 2018). In yeast, secreted proteins are targeted to the endoplasmic reticulum (ER) by either the co-translational pathway or the "post-translational" pathway. The co translational pathway is mediated by the signal recognition particle (SRP), while the "post-translational" pathway, independent of SRP, requires the intact membrane protein Sec62p for ER targeting. However, more hydrophobic signaling sequences utilize the SRP pathway to target the ER in order to increase secretion efficiency (Walter et al. 2011). Ost1p, a type I intact membrane protein with a cleavable signal sequence, plays a role in co-translational translocation to the ER (Willer et al. 2008). Subsequent bioreactor-scale fermentation has further confirmed that specific recombinant signal peptide promotes protein secretion (Barrero et al. 2021). To avoid variability in signal peptide efficiency for different protein secretion, the discovery of new signal peptides is essential (Liang et al. 2012; Sastry et al. 2014). Systematic analysis of yeast genomes and secretomes has led to the identification of novel signal peptides with more efficient secretion than α-MF, such as the signal peptide derived from the P. pastoris PAS_chr3_0030 gene (PC0) and the signal peptide NCW2 in Saccharomyces cerevisiae (Shen et al. 2022; Wang et al. 2015; Zou et al. 2022).
In this study, six distinct types of secreted signaling peptides, namely pPIC-α-MF-PMGL-Ba, pPIC-O-pre-α-pro-PMGL-Ba, pPIC-αopt-PMGL-Ba, pPIC-αM-PMGL-Ba, pPIC-PC0-PMGL-Ba, and pPIC-NCW2-PMGL-Ba, were selected to investigate their effects on PMGL-Ba production. As shown in Fig. 3, results from shake flask fermentation revealed a significant enhancement in pectinase secretion with O-pre-α-pro and aM compared to α-MF. Conversely, PMGL-Ba secretion by αopt and NCW2 was lower than that of α-MF. The signaling peptide PC0 exhibited secretion levels similar to α-MF. Notably, the protein concentration and enzyme activity of Ba secreted by the signal peptide O-pre-α-pro were 1.40 g/L and 1554.26 U/mL, respectively, representing a 130.71% and 98.13% increase compared to the departure strain pPICZαA-PMGL-Ba.
Figure 3 Effects of different signaling peptides on PMGL-Ba expression. (A) Selection of signal peptides. (B) Growth curves of strains corresponding to different signal peptides. (C) Effect of different signal peptides on total protein concentration of PMGL-Ba and SDS-PAGE analysis. (D) Effect of different signaling peptides on PMGL-Ba enzyme activity.
P. pastoris, a methylotrophic yeast, tightly regulates the expression of exogenous proteins when utilizing methanol as a carbon source through the methanol utilization pathway (MUT) promoter. The transcription level of exogenous protein genes is intricately linked to the strength of promoter regulation (Püllmann and Weissenborn 2021). The commonly employed AOX1 promoter is favored for its robust induction by methanol, although the precise mechanism of its methanol-induced action remains not fully elucidated (Berg et al. 2013; Wang et al. 2016).
Pichia pastoris is a methylotrophic yeast, and the expression of exogenous proteins using methanol as a carbon source is tightly regulated by the methanol utilization pathway (MUT) promoter. The transcription level of exogenous protein genes is closely linked to the strength of promoter regulation. The AOX1 promoter is commonly used and is often the promoter of choice for its strong induction by methanol. However, the mechanism of action of the AOX1 promoter strictly induced by methanol is not fully understood.
To unravel the mechanism of AOX1 promoter action, previous studies have utilized mutation screening, followed by sequence analysis, or adjusted regulatory factors of other carbon sources (Hartner et al. 2008; Zhan et al. 2017). In this study, three MUT strong promoters, reported to be more effective than AOX1 in inducing exogenous proteins, were selected. The first, AOXm, involved deleting a transcription factor sequence associated with glucose repression and copy addition of a positive cis-acting element sequence after AOX1 deletion (Hartner et al. 2008; Li et al. 2015). The second, from the study by Thomas Vogl et al., utilized genes from the pentose phosphate pathway (PPP) and the defense against reactive oxygen species (ROS) in the MUT pathway to provide strong promoters that somewhat exceed AOX1, such as the promoter of the gene encoding peroxisomal membrane protein (PMP20) (Yurimoto et al. 2000; Vogl et al. 2015). The third is formaldehyde dehydrogenase 1 (FLD1), a key enzyme in metabolizing methanol as a carbon source (Nakagawa et al. 2004; Wang et al. 2019). The promoters of their encoding genes show similar yields for FLD1 and AOX1 in exogenous protein expression, with FLD1 even surpassing AOX1 in some proteins (Püllmann and Weissenborn 2021; Wang et al. 2016).
In this study, all four promoters mentioned above were individually employed for PMGL-Ba expression, and the most suitable promoter for PMGL-Ba gene was selected. As depicted in Fig. 4, AOX1 demonstrated obvious advantages in PMGL-Ba expression. Biomass analysis further indicated a notable decreasing trend for strains PMP20 and FLD1, suggesting that AOX1 remains the most suitable promoter for PMGL-Ba gene expression in this context.
Figure 4 Effects of different promoters on PMGL-Ba expression. (A) Selection of promoters. (B) Growth curves of strains corresponding to different promoters. (C) Effect of different promoters on total protein concentration of PMGL-Ba and SDS-PAGE analysis. (D) Effect of different promoters on PMGL-Ba enzyme activity.
2.4 Effect of gene dosage on recombinant PMGL-Ba protein expression
In the expression of exogenous proteins in P. pastoris, constructing high-copy strains is often a straightforward and effective approach to enhance protein production. Various methods exist for constructing multicopy strains, with a common strategy involving the in vitro construction of multiple expression cassettes in tandem, subsequently integrated onto the P. pastoris genome. In order to obtain higher copy number strains, different screening markers can be employed for sequential integration into the P. pastoris genome (Wang et al. 2021; Yang et al. 2016). Another prevalent method involves the gradual increase of antibiotic concentration, with commonly used antibiotics such as zeocin and antibiotic G418; the screening concentration is usually higher than 100 µg/mL, with some reaching up to 2000 µg/mL (Kuo et al. 2015; Sha et al. 2013; Song et al. 2019). However, zeocin is a relatively expensive antibiotic, and its screening process is intricate and time-consuming. Another approach involves the Cre/lox recombination system, using a screening marker for multiple integration, but this method suffers from a prolonged cycle time (Li et al. 2017). In addition, Xia et al. constructed a novel double plasmid system using a combined strategy of genomic integration and heterologous expression (Xia et al. 2021).
In this study, homologous recombination was employed to construct three strains with different screening markers, namely GS115-2PMGL-Ba, GS115-3PMGL-Ba, and GS115-4PMGL-Ba, which were sequentially integrated onto the Picot yeast genome using different screening markers. This method effectively reduced the time required for constructing multiple copies by 2–3 days, streamlining the identification process post-integration into the yeast genome. As illustrated in Fig. 5, GS115-2PMGL-Ba exhibited improvements over GS115-PMGL-Ba, showcasing increased protein concentration and enzyme activity. However, GS115-3PMGL-Ba demonstrated a significant decrease in yield compared to GS115-2PMGL-Ba. In shake flask fermentation, the protein concentration and enzyme activity of GS115-2PMGL-Ba reached 1.50 g/L and 1658.01 U/mL, respectively, representing a 6.7% and 6.69% increase compared to X33-PMGL-Ba.
Figure 5 Effects of gene dosage on PMGL-Ba expression. (A) Schematic of different gene doses. (B) Growth curves of strains corresponding to different gene doses. (C) Effect of different gene doses on total protein concentration of PMGL-Ba and SDS-PAGE analysis. (D) Effect of different gene doses on PMGL-Ba enzyme activity.
2.5 Effect of overexpression of cofactors on recombinant PMGL-Ba protein expression
In the realm of protein expression pathways, alterations in chaperone protein levels can exert influence across multiple stages, encompassing transcription, translation, folding, and secretion. However, the impacts of such alterations are often protein-specific, and determining the singularly most crucial chaperone or the optimal combination remains ambiguous (Samuel et al. 2013). In this study, we expanded on GS115-2Ba by overexpressing 11 cofactors previously investigated for their roles in promoting protein expression.
A pivotal step in exogenous protein expression in P. pastoris is the translation process, where translation initiation serves as the rate-limiting step in the translation process. Various translation initiation-associated factors have been targeted to study the impact of translation efficiency. Notably, Jennifer Staudacher et al. demonstrated that overexpression of certain translation factors, particularly Pab1 and eIF4G, significantly enhances exogenous protein production in P. pastoris. The synergistic effect was more pronounced when all four loop-closing related factors, eIF4E, eIF4A, eIF4G and Pab1, were simultaneously overexpressed (Staudacher et al. 2022). In a previous study in our laboratory, Zheng et al. also proposed a novel yeast transcription factor, tentatively named Fhl1p, implicated in the regulation of rRNA processing genes and genes related to ribosomal small/large subunit biogenesis genes. This factor was found to facilitate the translation efficiency of exogenous proteins (Zheng et al. 2019).
Correct protein folding is a critical step in P. pastoris, particularly during post-translational translocation in the endoplasmic reticulum (ER). Newly synthesized peptides entering the ER must maintain an unfolded or loosely folded form after release from the ribosome, preventing aggregation before translocation. This process involves binding to cytoplasmic chaperones such as Ssa1 and Ydj1. In contrast, the ATPase activity of the ER luminal chaperone Kar2 powers post-translational translocations, promoting protein folding as well as targeting misfolded proteins to ERAD (Delic et al. 2013).
In response to endoplasmic reticulum stress induced by the overexpression of exogenous proteins, especially in multicopy strains, a common approach to alleviate these stress effects is the overexpression of cofactors. The HSP30 gene has been shown to play a crucial role in mitigating the repression of target gene transcript levels induced by high-intensity promoters. Besides, HSP30 gene expression is simultaneously regulated by the transcription factors HSF1, MSN2, and MSN4. Notably, the transcription factor HSF1 positively affects alleviating unfolded protein stress (Cui et al. 2023). Overexpression of the transcription factor Hac1p is frequently selected during chaperonin assisted secretory protein folding. In multicopy strains, the overexpression of exogenous proteins induces cellular stress responses, such as the common unfolded protein response (UPR), activated by unconventional splicing of HAC1 mRNA in yeast. The spliced HAC1 mRNA encodes an active transcription factor, Hac1p, which binds to UPR response elements in the promoters of UPR target genes, and it induces the expression of other chaperone proteins that assist in protein folding and protein transport (Liu et al. 2020; Guerfal et al. 2010).
In this study, we systematically overexpressed the 11 cofactors previously discussed, building upon the foundation of GS115-2PMGL-Ba. Interestingly, the Fig. 6 indicates that only overexpression of Hac1p increased the protein expression, while overexpression of the other ten cofactors decreased the protein expression. In shake flask fermentation, the protein concentration and enzyme activity of strain GS115-2PMGL-Ba: Hac1p were 1.81 g/L and 1835.53 U/mL, respectively. Compared with GS115-2PMGL-Ba, the protein concentration and enzyme activity increased by 20.85% and 10.7%.
Figure 6 Position of different protein cofactors in the protein secretory expression pathway. (A) Schematic of different gene doses. (B) Growth curves of strains corresponding to overexpression of different protein cofactors. (C) Effect of overexpression of different protein cofactors on total protein concentration of PMGL-Ba and SDS-PAGE analysis. (D) Effect of overexpression of different protein cofactors on PMGL-Ba enzyme activity.
2.6 5L fermenter refill batch fermentation
After identifying the highest PMGL-Ba yielding strain, GS11-2PMGL-Ba-Hac1p, through the different strategies mentioned above, it was activated and cultured into a seed solution. This seed solution was then introduced into a 5 L fermenter containing 2 L of BSM basal medium to initiate the batch culture phase. The glycerol feed-batch phase commenced after a dissolved oxygen rebound at 28 h. Glycerol replenishment was stopped when the yeast density reached OD600 = about 300, indicating complete glycerol consumption. Subsequently, the methanol feed-batch phase commenced, with samples taken every 8 h, continuing until the completion of 72 h.
As shown in the Fig. 7, the high-density fermentation of GS115-2PMGL-Ba-Hac1p in a 5 L fermenter for 72 h resulted in a substantial increase in protein concentration to 12.49 g/L, representing a remarkable 591.51% improvement compared to shake flask fermentation. Simultaneously, enzyme activity increased to 12668.12 U/mL, reflecting a 595.53% increase compared to shake flask fermentation. Previous experimental findings indicated protein degradation in the late stage of fermentation. To expedite the fermentation process, we increased the starting cell density of the induction stage of methanol flow addition from OD600 = 200 to OD600 = 300, and reduced the fermentation cycle from 144 h in shake flasks to 72 h.
Figure 7 PMGL-Ba high density fermentation. (A) 5L type fermenter high density fermentation. (B) Changes in biomass and total protein concentration over time in high-density fermentations. (C) SDS-PAGE analysis of supernatants after 20-fold dilution at different times in high-density fermentation. (D) Changes in PMGL-Ba activity at different times in high-density fermentation.