Enhanced Production of Recombinant Thermophilic Xylanase X11P in Ogataea polymorpha via In‐Silico Signal Peptide Discovery and Fed‐Batch Fermentation

ABSTRACT

Efficient secretion of heterologous proteins is essential for advancing yeast‐based bioprocesses, yet signal peptide (SP) optimization in the thermotolerant methylotrophic yeast Ogataea polymorpha remains limited. This study integrates in‐silico SP discovery, experimental validation, and bioprocess engineering to enhance secretion of the thermophilic xylanase X11P under sucrose‐inducible expression. Genome‐wide screening of 5184 O. polymorpha proteins using SignalP, Phobius, DeepLoc, WoLF PSORT, and ProP identified 11 high‐confidence SP candidates. Comparative analysis with Komagataella phaffii endogenous proteins guided selection of seven SPs for experimental evaluation. Among these, the novel O. polymorpha α‐mating factor‐like peptide FUN_005010 exhibited strong secretion‐promoting activity, with its prepro‐sequence yielding the highest extracellular xylanase levels and outperforming the classical Saccharomyces cerevisiae α‐MF. To evaluate industrial applicability, sucrose‐based fermentation strategies were systematically optimized in a 5‐L bioreactor. Controlled sucrose feeding and balanced C/N ratios were found to be critical for maximizing maltase (MAL) promoter‐driven expression. A stepwise increasing sucrose feed combined with induction at 30°C enabled X11P titers up to 770 U/mL, representing a 15‐fold improvement over shake‐flask cultures. This work demonstrates that the combination of SP evaluation and optimized sucrose‐inducible fed‐batch operation significantly enhances X11P production in O. polymorpha. The identified FUN_005010‐prepro SP and the refined process framework provide valuable tools for developing O. polymorpha as a high‐performance industrial expression platform.

Genome‐wide signal peptide screening identified a novel Ogataea polymorpha α‐mating factor‐like peptide (FUN_005010) that markedly improved recombinant thermophilic xylanase X11P secretion. Combined with optimized sucrose fed‐batch fermentation, X11P titers reached 770 U/mL, representing a 15‐fold increase over shake‐flask cultures.

Summary

• Genome‐wide in‐silico screening identifies strong SP candidates.

• SP evaluation outperforms classical Saccharomyces cerevisiae α‐MF.

• Novel FUN_005010 SP boosts xylanase secretion in Ogataea polymorpha.

• Optimized sucrose‐based fermentation raises X11P titers to 770 U/mL.

• Integrated SP evaluation and process engineering enhances protein production.

1. Introduction

The efficient production of heterologous proteins is fundamental to a wide array of biotechnological applications, including those in the pharmaceutical, food and feed, diagnostic, and renewable energy sectors. Yeasts have long been favored as eukaryotic expression platforms due to their ability to perform post‐translational modifications, rapid growth on inexpensive media, and suitability for high‐cell‐density fermentation. Among them, methylotrophic yeasts are notable for their strong inducible promoters and metabolic versatility. Although Pichia pastoris (syn. Komagataella phaffii) has been the dominant host for decades, other methylotrophic species are increasingly recognized for their unique advantages. Ogataea polymorpha (syn. Hansenula polymorpha) has emerged as a promising alternative due to its thermotolerance (up to 50°C), broad substrate utilization, and ability to secrete a range of recombinant proteins, highlighting its potential as a next‐generation chassis for industrial biotechnology (Xie et al. 2024).

In secretory expression systems, a major determinant of recombinant protein yield is the efficiency of the secretory pathway. Signal peptides (SPs), short N‐terminal sequences that mediate co‐translational translocation of nascent proteins into the endoplasmic reticulum (ER), are essential for initiating secretion. Once translocated, proteins undergo folding and processing before being exported to the extracellular environment. The choice of SP strongly influences secretion efficiency, ER targeting fidelity, and proper signal cleavage. The Saccharomyces cerevisiae α‐mating factor (Scα‐MF) remains widely used across yeast hosts; however, it is not universally optimal. Its long pre‐pro region, multi‐step maturation, and potential for inefficient cleavage can impede secretion, especially in non‐Saccharomyces yeasts with distinct processing machinery (Massahi and Çalık 2016; Dai et al. 2024).

Consequently, considerable attention has been directed toward the discovery and evaluation of alternative SPs. In K. phaffii, the integration of SP library screening with computational tools such as SignalP, Phobius and ProP has enabled the identification of high‐performing SPs tailored to specific target proteins. Massahi and Çalık (2015, 2016) evaluated 56 SPs from diverse organisms, revealing several that enhanced secretion of recombinant human growth hormone. Duan et al. (2019) used genome‐ and secretome‐guided SP mining to uncover endogenous candidates capable of improving secretion up to 172‐fold for multiple model proteins. Zou et al. (2022) demonstrated that α‐MF variants from Wickerhamomyces ciferrii and endogenous SPs from K. phaffii could outperform the classical Scα‐MF in driving EGFP secretion. More recently, Karaoğlan (2023) assessed SPs from Kluyveromyces lactis, K. marxianus, and O. polymorpha, and showed that several of these peptides, especially those from O. polymorpha PIR proteins, significantly improved α‐amylase and xylanase secretion in K. phaffii. These studies underscore that SP performance is host‐ and protein‐specific, emphasizing the need for SP optimization in non‐conventional yeasts such as O. polymorpha.

Despite its growing relevance, systematic SP optimization in O. polymorpha remains scarce. In practice, the α‐MF pre‐pro leader sequence derived from S. cerevisiae is the most commonly used SP for heterologous protein secretion in O. polymorpha (Eilert et al. 2013; Tsuda and Nonaka 2024). This limitation is notable given the organism's suitability for high‐temperature fermentations, which are advantageous for producing thermostable enzymes (Vasylyshyn et al. 2026). Thermophilic xylanases, for instance, are valuable in biomass conversion, food processing, and feed formulations, where high‐temperature activity provides both stability and process efficiency. The thermophilic xylanase X11P, derived from a metagenomic library isolated from sugarcane bagasse, exhibits strong catalytic performance at elevated temperatures and oligosaccharide specificity. Removal of its carbohydrate‐binding module and subsequent random mutagenesis further improved its catalytic efficiency (Boonyapakron et al. 2021), making X11P an ideal reporter for evaluating secretion efficiency in O. polymorpha.

Promoter selection is another key component of an effective production system. Although methanol‐inducible promoters such as MOX and AOX are widely used in methylotrophic yeasts, reliance on methanol introduces safety, environmental, and cost concerns. Recent advancements have demonstrated that sucrose‐inducible expression systems offer a viable alternative. Puseenam et al. (2018) developed a sucrose‐based induction platform in O. thermomethanolica, enabling robust expression without methanol. This system enabled high‐cell‐density fed‐batch fermentation of O. thermomethanolica for phytase production, where sucrose induction significantly increased phytase yield, productivity, and specific activity compared with methanol, supporting cost‐effective enzyme production under optimized conditions (Boonchoo et al. 2019). Most recently, transcriptomic profiling by Likhitrattanapisal et al. (2025) revealed that sucrose triggers strong induction of metabolic and secretory pathways in O. polymorpha TBRC 4839, further supporting its potential for industrial bioprocesses.

Given the lack of SP‐focused studies in O. polymorpha, the industrial importance of thermostable enzymes, and the growing relevance of sucrose‐inducible expression, there is a pressing need to improve recombinant protein secretion in this host. In this study, we integrate in‐silico SP prediction, experimental SP evaluation, and fed‐batch fermentation to enhance secretion of the thermophilic xylanase X11P in O. polymorpha. Using computational tools including SignalP, Phobius, DeepLoc, WoLF PSORT, and ProP, we identified endogenous SP candidates based on predicted cleavage efficiency, secretion probability, subcellular localization, and propeptide processing. Selected SPs were experimentally validated under a sucrose‐inducible promoter and benchmarked against the canonical Scα‐MF. To assess industrial applicability, fed‐batch cultivation was performed in a 5‐L bioreactor.

2. Materials and Methods

2.1. Strains and Plasmids

The methylotrophic yeast O. polymorpha TBRC 4839 was obtained from the Thailand Bioresource Research Center (TBRC) and used as the expression host throughout this study. The strain was routinely cultured at 37°C on yeast extract peptone dextrose (YPD) medium consisting of 1% (w/v) yeast extract, 2% (w/v) peptone, and 2% (w/v) glucose. For propagation and maintenance of plasmids, Escherichia coli DH5α was employed and cultivated in Luria‐Bertani (LB) medium supplemented with 25 μg/mL hygromycin. Yeast transformants were selected on YPD agar plates containing 200 μg/mL hygromycin.

Plasmid construction was carried out using the in‐house developed expression vector, pOpSwitch‐X11P. This plasmid features a sucrose‐inducible promoter (pMAL) and a thermophilic bacterial xylanase gene (X11P) as a secretion reporter. The plasmid also includes a hygromycin resistance marker to facilitate selection in both E. coli and O. polymorpha.

2.2. In‐Silico Identification of Candidate SPs

To identify novel SP for efficient secretion in O. polymorpha TBRC 4839, a genome‐wide bioinformatics analysis was conducted using a dataset of 5184 protein‐coding sequences derived from the O. polymorpha genome database (Likhitrattanapisal et al. 2025; Supporting Information S1: Data Table S1). SP prediction was initially performed using two software tools: SignalP version 6.0 (Teufel et al. 2022) and Phobius version 1.01 (Käll et al. 2004). Proteins were considered putative secretory candidates when both tools predicted consistent N‐terminal SP regions. The candidates were further filtered using more stringent criteria: SignalP 6.0 likelihood scores above 0.99 and cleavage site probability scores exceeding 0.95.

To assess the presence of pro‐sequence motifs, which are important for post‐translational processing, ProP version 1.0c was used to identify cleavage sites in the putative secretory protein candidates (Duckert et al. 2004). Moreover, the presence of conserved motifs such as KR or KREA boxes near or adjacent to the predicted cleavage sites was also considered. The protein candidates with pro‐sequence motifs and KR or KREA boxes located within the first 150 amino acids from the N‐terminus (ProP score cutoff > 0.15) were selected.

To confirm the extracellular localization of the selected candidates, subcellular localization prediction was carried out using DeepLoc 2.0 (Thumuluri et al. 2022) and WoLF PSORT 2.0 (Horton et al. 2007). Only the protein candidates of which localizations were predicted to be extracellular by both DeepLoc and WoLF PSORT were chosen for the experimental validation. The well‐characterized α‐mating factor signal sequence from S. cerevisiae (Scα‐MF) was included as a positive control throughout the study.

2.3. Construction of SP‐Reporter Expression Plasmids

Seven candidate SP sequences, including the FUN_005010 prepro‐sequence and both the pre‐ and prepro‐sequences of Scα‐MF (Supporting Information Data Table S2), each flanked by BlnI and EcoRI restriction sites at the 5′ and 3′ ends, respectively, were commercially synthesized (GenScript, USA) and initially cloned into the pUC57 cloning vector. These SP fragments were excised using BlnI and EcoRI and subsequently ligated into the pOpSwitch‐X11P vector upstream of the X11P reporter gene. The resulting recombinant plasmids were transformed into E. coli DH5α and selected on LB agar containing 25 μg/mL hygromycin. Correct insertion and orientation of the SP fragments were verified by automated DNA sequencing.

Transformation of O. polymorpha with the constructed expression plasmids was performed via LiAc/single‐stranded carrier DNA/PEG transformation method (Holkenbrink et al. 2018) using 1 µg of NotI‐linearized plasmid. Yeast transformants were selected on YPD agar plates containing 200 μg/mL hygromycin. Integration of the recombinant plasmids into the yeast genome was confirmed by colony PCR using plasmid‐specific primers. The resulting recombinant strains were used for subsequent expression and secretion analysis of X11P.

2.4. Expression of X11P Under Sucrose‐Inducible Conditions

To evaluate the expression of X11P under the control of the sucrose‐inducible promoter, a single colony of each recombinant strain was inoculated into 5 mL YPD supplemented with 200 µg/mL hygromycin and grown overnight at 37°C with shaking at 200 rpm. Overnight cultures were then transferred into 5 mL fresh YPD containing the same antibiotic to an initial OD₆₀₀ of 0.2 and incubated for 24 h to allow biomass accumulation. Protein expression was subsequently induced by adding sucrose to a final concentration of 2% (w/v), and cultures were incubated for an additional 48 h prior to sampling.

2.5. Fed‐Batch Fermentation of Recombinant X11P Xylanase

Recombinant X11P xylanase was produced in O. polymorpha using a fed‐batch fermentation strategy with sucrose as the induction carbon source (Boonchoo et al. 2019). Seed culture of recombinant yeast was prepared by inoculating a single colony or 0.2 ml of −80°C glycerol stock into 20 mL yeast extract peptone sucrose (YPS) medium consisting of 1% (w/v) yeast extract, 2% (w/v) peptone, and 2% (w/v) sucrose and incubated in a shaker with shaking speed of 250 rpm at 37°C for 24 h. A 20 mL aliquot of this pre‐culture was transferred into 180 mL of Syn6 synthetic medium containing sucrose as the carbon source (Stöckmann et al. 2003; Eilert et al. 2012). This seed was incubated under the same conditions (37°C, 250 rpm) for an additional 24 h. This second‐stage seed culture was then used to inoculate the bioreactor.

Fed‐batch fermentation was performed in a 5‐L stirred‐tank bioreactor containing 2 L of Syn6 medium supplemented with 20 g/L sucrose. The batch phase was conducted at 37°C with the pH controlled at 5.0 using 24% (w/w) ammonium hydroxide, which also served as a nitrogen source. Dissolved oxygen (DO) was controlled at 30% air saturation using cascade control, with agitation speed adjusted between 500 and 1000 rpm and oxygen enrichment applied at a constant aeration rate of 1 vvm.

During the batch phase, cells were grown until sucrose was completely depleted, as indicated by a rise in DO and confirmed by offline sugar analysis. The process then transitioned to the fed‐batch phase, in which sucrose‐based feed medium was supplied at a controlled rate to support biomass accumulation and simultaneously induce the production of recombinant enzyme.

2.6. Enzyme Activity Assay

The xylanase activity in culture supernatants was quantified using a 1% (w/v) solution of beechwood xylan (Sigma‐Aldrich, USA) as the substrate. The enzymatic reaction was carried out under optimized conditions as described previously (Boonyapakron et al. 2021). The release of reducing sugars was determined using the 3,5‐dinitrosalicylic acid (DNS) method (Miller 1959). Absorbance was measured at 540 nm using a spectrophotometer. One unit (U) of xylanase activity was defined as the amount of enzyme releasing 1 μmol of reducing sugars per minute under the assay conditions. Enzyme activity values were normalized to the optical density of the culture at 600 nm (OD₆₀₀), and results were expressed as U/OD to account for variations in cell density among samples.

2.7. Sugar and Metabolite Analysis by UHPLC

Metabolites generated during fermentation were analyzed using ultra‐high‐performance liquid chromatography (UHPLC). Culture supernatants collected from the recombinant yeast cultivation were examined for residual sugars and key metabolites, including ethanol, acetate, and pyruvate, to support optimization of fermentation conditions for enhanced X11P production. Quantification was performed using a SUGAR SH1011 column (Shodex; 8.0 mm i.d. × 300 mm) connected to a Dionex UHPLC system. The column temperature was maintained at 30°C, and 3 mM perchloric acid (HClO₄) was used as the mobile phase at a flow rate of 0.4 mL/min. Samples were filtered prior to injection, and metabolites were identified and quantified based on retention times and calibration curves of authentic standards.

2.8. Statistical Analysis of Enzyme Activity

Enzyme activity data were evaluated by one‐way analysis of variance (ANOVA), followed by Tukey's post hoc multiple comparison test. Results are expressed as mean ± standard deviation (SD) based on three independent replicates (n = 3). Differences were considered statistically significant at p < 0.05.

3. Results and Discussion

3.1. In‐Silico Identification of SPs from O. polymorpha

To improve the efficiency of heterologous protein secretion in O. polymorpha, a genome‐wide search was performed to identify potential high‐performance SP from its genome (Likhitrattanapisal et al. 2025). A total of 5184 annotated protein sequences obtained from the O. polymorpha genome were subjected to in‐silico SP prediction. Candidate SPs were initially selected based on high probability scores, as suggested in previous studies (Massahi and Çalık 2015).

SignalP 6.0 and Phobius identified 197 proteins containing predicted SPs. The selection criteria were: (1) an SP probability score greater than 0.99 and (2) a cleavage site probability score greater than 0.95, both derived from SignalP 6.0. Among these, 123 proteins met the criteria and were therefore selected as high‐confidence candidates for secretion. These sequences were further examined for subcellular localization using DeepLoc 2.0 (Thumuluri et al. 2022) and WoLF PSORT 2.0 (Horton et al. 2007), resulting in a set of 63 protein candidates predicted to be localized extracellularly.

Subsequently, ProP version 1.0c (Duckert et al. 2004) was used to identify potential pro‐sequence cleavage sites within the first 150 amino acids from the N‐terminus, further narrowing the list to 38 proteins. These candidates were then filtered for the presence of conserved KR or KREA motifs, which are commonly associated with efficient pro‐protein processing, resulting in 11 candidate proteins (Supporting Information S1: Data Figure S1 and Table S3). All of which are predicted to possess strong secretion potential in O. polymorpha.

The stepwise bioinformatic screening approach ensured that only SPs with high secretion likelihood, confirmed extracellular localization, suitable pro‐sequence features, and characteristic processing motifs were advanced for downstream experimental validation. This systematic selection process not only minimized the number of candidates requiring laboratory testing but also maximized the probability of identifying SPs with superior secretion efficiency in O. polymorpha.

3.2. Comparative Analysis of Candidate SPs With K. phaffii Endogenous Proteins

The 11 candidate proteins identified from O. polymorpha (FUN_000056, FUN_000291, FUN_000707, FUN_000802, FUN_001823, FUN_002850, FUN_003195, FUN_004163, FUN_004188, FUN_004380, and FUN_005010) were compared against the 32 endogenous proteins of K. phaffii reported in a previous study (Zou et al. 2022). Sequence identity analysis revealed that eight of these O. polymorpha proteins shared significant similarity with K. phaffii endogenous proteins. In contrast, three proteins FUN_001823 (hypothetical protein), FUN_004163 (ECM14 putative metallocarboxypeptidase), and FUN_005010 (putative α‐mating factor protein) displayed no detectable sequence similarity to K. phaffii proteins (Supporting Information S1: Data Table S4).

In the study by Zou et al. (2022), the secretion efficiency of EGFP in K. phaffii was evaluated using SPs derived from endogenous proteins. Four of these SPs (PAS_chr3_0076, PAS_chr4_0584, PAS_chr4_0584, and PAS_chr1‐1_0011) demonstrated high EGFP secretion efficiency (Supporting Information S1: Data Table S4). Notably, these four K. phaffii proteins exhibited sequence similarity to the following O. polymorpha candidates: FUN_000291 (cysteine‐rich secretory protein family), FUN_000802 (SAP9 aspartyl protease), FUN_002850 (APR1 aspartic proteinase precursor), and FUN_003195 (hypothetical protein).

Based on these findings, SPs from these four O. polymorpha proteins were selected for further analysis, along with SPs from the three proteins unique to O. polymorpha FUN_001823 (hypothetical protein), FUN_004163 (metallocarboxypeptidase), and FUN_005010 (mating factor protein) that exhibited no sequence similarity to K. phaffii proteins and are therefore considered novel candidates. In total, seven SPs were chosen for experimental evaluation of X11P secretion.

3.3. Evaluation of Candidate SPs for X11P Expression

To evaluate the secretion efficiency of the selected candidate SPs, seven in‐silico identified sequences were fused to the recombinant X11P and expressed in O. polymorpha. For FUN_005010, both the pre‐sequence and the prepro‐sequence were examined. As a reference control, the pre‐sequence and prepro‐sequence of the Scα‐MF were included in the study. The secretion performance was assessed based on extracellular xylanase activity (U/OD). Among all candidates, the FUN_005010 SP displayed remarkable differences between its pre‐ and prepro‐forms (Figure 1, Supporting Information S1: Data Figure S2). While the pre‐sequence alone yielded moderate secretion (0.09 U/OD), the incorporation of its pro‐region significantly enhanced extracellular enzyme levels, resulting in the highest activity observed in this study (0.19 U/OD). In contrast, the classical Scα‐MF SPs exhibited lower performance, with activities of 0.07 U/OD (pre) and 0.09 U/OD (prepro). These results clearly demonstrate that the newly identified SPs FUN_005010‐prepro outperformed the widely used Scα‐MF sequences in directing efficient secretion of X11P.

Figure 1.

Heterologous expression of X11P xylanase in Ogataea polymorpha using various signal peptides under sucrose‐inducible conditions. Transformants were first cultivated in glucose‐containing medium and subsequently induced with sucrose. For each signal peptide, three independent transformants were analyzed for X11P xylanase expression. Data are presented as mean ± SD. Enzyme activity was evaluated by one‐way ANOVA as described in the Materials and Methods section. Among the signal peptides tested, Fun_005010‐prepro exhibited the highest activity, showing a statistically significant increase compared to all others, except FUN_002850. Wt denotes the wild‐type O. polymorpha. Transformants carrying each signal peptide fused to the X11P gene are indicated.

In addition, SDS‐PAGE analysis of intracellular protein fractions (cell lysates) was performed (Supporting Information S1: Data Figure S3). The results show that recombinant yeast strains expressing different SPs exhibit more prominent intracellular protein bands compared with the wild‐type strain. Notably, at the expected molecular weight of the target protein, the strain harboring the FUN_005010‐prepro SP displays a relatively weaker intracellular band than strains carrying other SPs. This observation suggests that the FUN_005010‐prepro SP enhances secretion efficiency, resulting in reduced intracellular accumulation of the target protein. Therefore, the observed differences in extracellular enzyme activity and protein levels are likely attributable, at least in part, to variations in secretion efficiency rather than differences in expression levels alone.

Sequence analysis revealed that FUN_005010 shares notable similarity with yeast α‐mating factor proteins (Supporting Information S1: Data Figure S4). The peptide comprises 85 amino acids, including the conserved motif KREA and a putative mature region (GWGWHRVN/SRNEVIF). FUN_005010 exhibits up to 94% identity with the α‐MF sequence of O. parapolymorpha DL‐1, underscoring its evolutionary conservation. Typically, yeast α‐MF SPs contain a 20−25 residue pre‐region, which is the N‐terminal signal sequence responsible for directing the nascent polypeptide to the secretory pathway and facilitating its translocation across the ER membrane, followed by a 60−100 residue pro‐region, a segment that assists in proper folding, stability, and efficient secretion of the protein (Roongsawang et al. 2016). In comparison, the pre‐sequence of FUN_005010 is 24 amino acids long, consistent with known α‐MF signals, whereas its pro‐region is unusually short, consisting of only 14 amino acids. This is considerably shorter than the pro‐regions in K. pastoris or S. cerevisiae, which range from 60 to 100 amino acids. In addition, KR or KREA motifs are typically recognized as proteolytic processing sites for endoproteases such as Kex2p, which cleave precursor proteins during secretion, while additional motifs such as EAEA may be further processed by Ste13p. However, no KR or EAEA motifs were detected at the C‐terminal end of the FUN_005010 pro‐region, indicating the absence of canonical Kex2p or Ste13p cleavage sites. Consequently, the downstream putative mature region refers to the final functional peptide sequence that remains after the removal of the SP and any pro‐peptide segments, in this case represented by the sequence (GWGWHRVN/SRNEVIF).

Despite its unusually short length and atypical processing signals, the FUN_005010 pro‐region markedly improved secretion performance. This observation aligns with previous findings emphasizing the importance of the pro‐region in promoting efficient protein folding, ER translocation, and trafficking. Roongsawang et al. (2016) reported that the prepro‐sequence of the O. thermomethanolica α‐MF (Otα‐MF) enhanced the secretion of heterologous endoxylanase more effectively than the classical ScαMF signal. Notably, their study also demonstrated that the requirement for a pro‐region is protein dependent; while the Otα‐MF prepro‐sequence strongly improved secretion of glycosylated phytase, the pre‐sequence alone did not.

Consistent with these observations, our results reinforce the concept that optimal SP performance depends on compatibility between SP architecture and the tested protein. The substantial improvement achieved with the FUN_005010‐prepro sequence suggests that even a short pro‐region can enhance secretion when appropriately matched with the tested protein. It is possible that the pro‐region of FUN_005010 plays roles in facilitating proper folding, stabilizing intermediates during ER processing, or improving trafficking efficiency.

3.4. Development of Fed‐Batch Strategies for Sucrose‐Induced X11P Production

Previous studies reported that phytase production in Ogataea thermomethanolica was significantly enhanced by sucrose‐inducible fed‐batch fermentation (Boonchoo et al. 2019). Building on this approach, the present study investigated bioprocess optimization, particularly feeding strategies, to improve recombinant protein production in O. polymorpha. Sucrose‐based fed‐batch fermentation was applied to enhance xylanase X11P production, using sucrose as both carbon source and inducer. Constant feeding of 400 g/L sucrose at 62 mL/h effectively induced X11P expression, resulting in a maximum activity of 142 U/mL (Figure 2). Although this titer was higher than those achieved in shake‐flask experiments (50 U/mL), overall enzyme levels remained relatively low, mirroring the importance of controlled feeding in fed‐batch systems to improve heterologous enzyme yields. Such fed‐batch strategies with optimized feeding profiles have been shown to significantly enhance recombinant enzyme production in yeast systems (Shang et al. 2017).

Figure 2.

Growth profile and X11P production in a 5‐L bioreactor under fed‐batch fermentation with a constant feeding rate of 62 mL/h. The feed compositions were sucrose at 400 g/L. Error bars indicate the standard deviation of three technical replicates (n = 3), reflecting measurement variability.

Based on our observations, two factors are likely to limit productivity. First, feeding began at 13 h when sucrose was already depleted, indicating that nutrient limitation had occurred before induction. Second, the constant feeding rate may not have matched cellular metabolic demand, restricting carbon supply during the induction phase. These observations underscored the importance of initiating feeding earlier, while residual sucrose remains available, and of adjusting feed rates dynamically to balance growth and induction. Adjusting the feeding onset and rate profiles to align with metabolic demand has been emphasized in other dynamic fed‐batch regimes, where feed control based on substrate uptake or residual carbon enhances recombinant protein productivity (Spadiut et al. 2014). Subsequent experiments therefore focused on optimizing both feeding onset and feed‐rate profiles to improve X11P productivity.

To assess the effect of feeding dynamics, two improved strategies were tested: (I) constant sucrose‐plus‐yeast‐extract feeding at 62 mL/h for 7.5 h, and (II) a stepwise increasing feed rate from 62 to 124 mL/h, with an equal total feed volume of 800 mL. Feeding began at 12 h, when residual sucrose remained at 4‐6 g/L. The constant feed strategy yielded the highest xylanase titer of 268 U/mL at 29 h (Figure 3A), whereas the increasing‐rate profile produced 206 U/mL at 29 h (Figure 3B).

Figure 3.

X11P production profiles in fed‐batch fermentation using different feeding strategies. Constant feed rate (A). Stepwise increasing feed rate (B). Error bars indicate the standard deviation of three technical replicates (n = 3), reflecting measurement variability.

Although the final enzyme titer obtained with the increasing feed rate was lower than that of the constant feed rate, the increasing strategy resulted in a higher enzyme production rate (0.154 vs. 0.124 U/mL/h), indicating faster induction during the early fed‐batch phase. This difference may arise because both strategies used the same total feed volume (800 mL) starting at 12 h. Under constant feeding, substrate addition continued until the end of fermentation, allowing sustained enzyme production. In contrast, the increasing feed rate exhausted the feed earlier, after which enzyme activity no longer increased and biomass declined, suggesting cell lysis. Nevertheless, during the active production phase, the increasing feed rate (Figure 3B, 12–26 h) exhibited a higher enzyme production rate than constant feeding (Figure 3A, 12–29 h). Such observations are consistent with reports in yeast fed‐batch cultivations where tailored feeding profiles can alter product synthesis rates and overall titers (Shang et al. 2017). These results highlight the need to balance sucrose supply between supporting biomass formation and ensuring effective induction via the sucrose‐responsive MAL promoter.

3.5. Optimization of Feeding Profiles for Balanced Growth and Induction

Because sucrose serves as both the carbon source and inducer, precise feeding control is essential for high‐level X11P expression. Four feeding strategies were evaluated: constant and stepwise‐increasing feed rates. Constant feeding was performed at 51.5 mL/h for 15.5 h (I) and 29 mL/h for 27.5 h (II). Stepwise feeding increased from 29 to 51.5 mL/h over 9.5 h (III) or 11 h (IV). All strategies used 800 mL of feed and started after 12−12.5 h of cultivation.

The stepwise increasing feed rate (condition III) provided the most effective induction, ultimately yielding xylanase titers comparable to those obtained with the increasing‐rate profiles, reaching approximately 417 U/mL at 54 h (Figure 4A). Although initial production rates were slower, this strategy provided a favorable balance between growth and induction. Similar observations have been reported where precise feeding profiles enhance recombinant protein production by balancing growth and induction rates in fed‐batch yeast cultures (Dietzsch et al. 2011). In contrast, higher or rapidly increasing feed rates primarily promoted biomass accumulation, diminishing recombinant protein synthesis (Figure 4B). These results indicate that excessive sucrose influx skews metabolism toward growth rather than induction, thereby reducing MAL promoter responsiveness.

Figure 4.

Fed‐batch fermentation profiles of X11P production (A) and biomass formation (B) by recombinant Ogataea polymorpha under different feeding strategies. Constant feeding was applied at 51.5 mL/h for 15.5 h (I) or 29 mL/h for 27.5 h (II), while stepwise feeding increased from 29 to 51.5 mL/h over 9.5 h (III) or 11 h (IV). Error bars indicate the standard deviation of three technical replicates (n = 3), reflecting measurement variability.

To further improve productivity, the feed C/N ratio was optimized. Two sucrose‐to‐yeast‐extract ratios, 400:25 (≈16:1) and 400:50 (≈8:1), were tested under a constant 29 mL/h feed rate, compared with the baseline low‐nitrogen feed (400:10; ≈40:1). Increasing nitrogen content (400:50) enhanced biomass formation and boosted xylanase production to 600 U/mL (Figure 5B). Optimization of C/N ratio has been shown to significantly influence recombinant xylanase activity in fed‐batch systems under varying nitrogen sources (Fuzi et al. 2014). However, the gain was not proportional to the twofold increase in yeast extract, and the high‐N feed required heat sterilization, increasing process complexity. The baseline low‐nitrogen feed (10 g/L), sterilized by filtration, achieved 400 U/mL (Figure 4A) and outperformed the intermediate C/N ratio (400:25), which yielded 350 U/mL (Figure 5A). These findings suggest that while additional nitrogen improves performance, diminishing returns and operational constraints necessitate a balanced C/N ratio for economically feasible bioprocessing.

Figure 5.

Growth and X11P production profiles in fed‐batch fermentation using feed media with different carbon‐to‐nitrogen (C:N) ratios. C:N ratio of 400:25 (A) and C:N ratio of 400:50 (B). Error bars indicate the standard deviation of three technical replicates (n = 3), reflecting measurement variability.

3.6. High‐Level X11P Production Under Optimized Fed‐Batch Conditions

Based on insights from feeding optimization studies, fed‐batch fermentation conditions were further refined to maximize X11P production. Cultivation was performed at 35°C, and fed‐batch induction was initiated at 12–14 h when dry cell weight exceeded 5 g/L. The sucrose‐based feed (400 g/L sucrose plus 10 g/L yeast extract) was supplied using a stepwise increasing feed rate, adding 1.5 L of feed over 44.5 h to an initial working volume of 1.5 L. Two separate experiments were conducted to evaluate induction temperature, with induction carried out at 35°C in one experiment and at 30°C in the other.

Although the sucrose consumption rate and xylanase formation rate during the same induction period were higher at 35°C than at 30°C, the final xylanase titer at 72 h was greater under the lower induction temperature. Specifically, induction at 35°C resulted in a xylanase activity of 700 U/mL with a specific activity of 105.47 U/mg, corresponding to a total extracellular protein concentration of 6.64 mg/mL (Figure 6A). In contrast, induction at 30°C further enhanced xylanase production, yielding 770 U/mL with a slightly higher specific activity of 106.64 U/mg and an increased protein concentration of 7.22 mg/mL (Figure 6B). These results indicate that, despite slower metabolic and expression kinetics, prolonged induction at a reduced temperature favored higher overall enzyme accumulation. This improvement is likely attributable to enhanced protein folding efficiency, reduced cellular stress, and decreased proteolytic degradation under lower‐temperature conditions, which collectively promote greater stability and accumulation of the recombinant enzyme. Lower induction temperatures have been shown to alleviate ER stress and improve folding and secretion efficiency in yeast hosts by moderating translation rates and enhancing protein processing capacity (Anasontzis et al. 2014; Dragosits et al. 2009). In P. pastoris, elevated cultivation temperatures have been reported to intensify unfolded protein response (UPR) activation, leading to accumulation of misfolded intermediates and reduced recombinant protein yield, whereas lower temperatures mitigate ER stress and improve secretion efficiency (Zhong et al. 2014). In addition, proteolytic degradation represents a major limitation in yeast‐based secretion systems. In P. pastoris, recombinant proteins can be degraded by both intracellular and extracellular proteases, particularly under stress conditions (Cereghino and Cregg 2000; Macauley‐Patrick et al. 2005). Reduced cultivation temperature has been associated with decreased protease activity and improved protein stability, further contributing to increased final product titers (Zhong et al. 2014). Although the recombinant enzyme X11P is thermostable and might be expected to fold efficiently at elevated temperatures, the observed improvement at lower temperatures suggests that host‐related constraints, particularly limitations in ER processing capacity and proteostasis, as well as protease‐mediated degradation, are the dominant factors. Therefore, the higher titer observed at lower temperature likely reflects a more favorable balance between protein synthesis, folding, secretion, and degradation within the host system, rather than the intrinsic thermal stability of the recombinant enzyme alone. Both conditions led to approximately 14–15 fold improvement compared with shake‐flask production (50 U/mL). The enhanced titers emphasize the critical role of induction temperature, controlled sucrose feeding, and adequate nitrogen supply in enabling high‐level expression from the MAL promoter in O. polymorpha.

Figure 6.

Growth and X11P production profiles under optimal cultivation conditions in fed‐batch fermentation with sucrose induction. Induction at 35°C (A) and 30°C (B). Error bars indicate the standard deviation of three technical replicates (n = 3), reflecting measurement variability.

4. Conclusion

This study developed an integrated strategy to enhance heterologous X11P xylanase secretion and sucrose‐inducible expression in O. polymorpha. Bioinformatic screening identified several effective SPs, with FUN_005010‐prepro exhibiting the highest secretion efficiency. Based on these results, systematic optimization of sucrose‐based fed‐batch fermentation achieved an X11P xylanase titer of 770 U/mL, representing a 15‐fold increase compared with shake‐flask cultures. Overall, the synergistic combination of SP evaluation and bioprocess optimization provides a robust and scalable platform for industrial X11P production in O. polymorpha.

Author Contributions

Research concept and design: Kanokarn Kocharin, Supawadee Ingsriswang and Niran Roongsawang. Acquisition of data: Kanokarn Kocharin, Chitwadee Phithakrotchanakoon, Somsak Likhitrattanapisa, Aekkachai Puseenam, Kitisak Sansatchanon, Jedsadakorn Ninrat, Wannaporn Sagasae, Paopit Siriarchawatana, Warasirin Sornlek, Supawadee Ingsriswang and Niran Roongsawang. Data analysis and interpretation: Kanokarn Kocharin, Chitwadee Phithakrotchanakoon, Somsak Likhitrattanapisa, Aekkachai Puseenam, Kitisak Sansatchanon, Jedsadakorn Ninrat, Wannaporn Sagasae, Paopit Siriarchawatana, Warasirin Sornlek, Supawadee Ingsriswang and Niran Roongsawang. Drafting the manuscript: Kanokarn Kocharin, Chitwadee Phithakrotchanakoon, Somsak Likhitrattanapisa and Niran Roongsawang. Critical revision of the manuscript: Niran Roongsawang. Final approval of the manuscript: Niran Roongsawang. Agreed to be accountable: Kanokarn Kocharin, Chitwadee Phithakrotchanakoon, Somsak Likhitrattanapisa, Aekkachai Puseenam, Kitisak Sansatchanon, Jedsadakorn Ninrat, Wannaporn Sagasae, Paopit Siriarchawatana, Warasirin Sornlek, Supawadee Ingsriswang and Niran Roongsawang.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1

Supporting File 2

Supporting File 3

Supporting File 4

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article and from the corresponding author upon reasonable request.