Enhanced methanol tolerance and utilization in Pichia pastoris revealed by adaptive laboratory evolution Authors: Junyeob Kim, Sang Woo Seo Published: 2025 ABSTRACT BackgroundMethanol is a promising renewable C1 feedstock for biomanufacturing. Native methylotrophs such as Pichia pastoris are considered potential chassis strains for methanol utilization. However, the cytotoxicity of methanol limits its usable concentration, posing a major bottleneck in bioproduction. In this study, we aimed to develop a P. pastoris strain with enhanced methanol tolerance and utilization by adaptive laboratory evolution.ResultsThrough serial passaging under increasing methanol concentrations, we obtained an evolved strain that exhibited a specific growth rate of 0.101 h− 1 in a 7% (v/v) methanol, a condition under which the wild-type strain failed to grow. Genome resequencing identified the mutations, and the introduction of individual mutations into the wild-type background demonstrated that mutations in PSR1 and BFA1 significantly improved growth under methanol stress. Notably, the fastest-growing isolate, designated PM7a, was capable of growing in a minimal medium containing 5% methanol as the sole carbon source, whereas the wild-type strain was unable to do so. Furthermore, the introduction of the β-carotene biosynthesis pathway into both the wild-type and PM7a strains, cultured in a minimal medium containing methanol as the sole carbon source, resulted in a 4.91-fold higher titer in PM7a.ConclusionsTaken together, our findings demonstrate that ALE facilitates the development of strains with enhanced methanol tolerance, growth, and biosynthetic performance. These findings highlight the potential of the evolved P. pastoris strain as a robust chassis for methanol-based biomanufacturing.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13036-025-00586-w. Introduction With the growing global emphasis on sustainable development, biomanufacturing using microbial cell factories is gaining interest [1]. Traditionally, sugar has been the most widely used feedstock for microbial fermentation. However, its direct link to the food supply raises concerns about food security [2]. Given the rising demand for biomanufacturing, developing non-food feedstocks for microbes is becoming essential. C1 compounds have emerged as promising alternatives due to their abundance, low cost, and relevance to renewable energy-particularly because some are greenhouse gases [3, 4]. Among C1 compounds, methanol has attracted particular attention because it can be synthesized electrochemically from CO2 and is miscible with water, thereby mitigating mass transfer limitations commonly encountered with gaseous substrates [5, 6]. Despite these advantages, most microbes cannot utilize methanol as a sole carbon source. To address this limitation, several studies have attempted to engineer conventional model organisms such as Escherichia coli and Saccharomyces cerevisiae by introducing synthetic methanol assimilation pathways [7–10]. While these efforts have yielded promising results, challenges such as cofactor imbalance and formaldehyde toxicity remain unsolved [11]. To circumvent these issues, researchers have also focused on improving methanol metabolism in natural methylotrophs, particularly Pichia pastoris (also known as Komagataella phaffii), which naturally compartmentalizes toxic intermediates such as formaldehyde within the peroxisome [12]. However, this detoxification mechanism has limited capacity, and cultivation under high methanol concentrations still results in prolonged lag phases and reduced specific growth rates due to residual toxicity [13]. As a result, particularly for heterologous protein production using the strong alcohol oxidase 1 (AOX1) promoter, methanol is typically maintained at low concentrations (e.g., 0.5-1.0%) during the induction phase [14]. To enhance methanol tolerance, recent studies have employed adaptive laboratory evolution (ALE) as a promising strategy to improve methanol utilization in P. pastoris under methanol stress. For instance, Wang et al. recently utilized an iterative evolution strategy to isolate a P. pastoris strain capable of tolerating up to 10% (v/v) methanol in liquid BMMY (buffered methanol-complex) medium and 13% on plates, attributing tolerance primarily to membrane phospholipid remodeling, notably phosphatidylcholine levels [15]. Another study by Moser et al. evolved populations at 1% methanol in rich YPM (yeast extract-peptone-methanol) and minimal BMM (buffered minimal methanol) medium, reporting populations with specific growth rates of approximately 0.227 h− 1 in YPM and 0.129 h− 1 in BMM, with recurrent mutations in alcohol oxidase 1 (AOX1), a key enzyme of methanol oxidation [16]. Despite these advances, most reports do not combine high tolerance with robust growth and measurable product accumulation in defined minimal medium under methanol as the sole carbon source. Accordingly, industrial processes are typically run at low methanol concentrations, which limits the available carbon supply and often requires frequent supplementation to achieve and sustain high titers. While industrial processes can tightly control low methanol concentrations through fed-batch strategies, high tolerance offers significant practical advantages. First, it can simplify the overall bioprocess by enabling simpler cultivation methods such as batch or pulse-feeding, thereby reducing operational complexity and costs associated with precise control systems [13]. Second, in large-scale bioreactors, tolerance provides enhanced robustness against unavoidable transient fluctuations in local methanol concentrations, ensuring process stability [17]. Finally, the ability to utilize high concentrations of methanol as a carbon source is directly linked to achieving higher cell densities, a key factor in maximizing the final product titer and productivity. Therefore, the development of a P. pastoris strain capable of robust growth and efficient methanol utilization at elevated concentrations is crucial for advancing methanol-based biomanufacturing. In this study, we developed a Pichia pastoris strain with enhanced methanol tolerance and utilization through long-term adaptive laboratory evolution (ALE). The evolved strains exhibited tolerance to 7% (v/v) methanol, a concentration at which the wild-type strain failed to grow, and showed a markedly improved specific growth rate. To identify the genomic mutations, we performed whole-genome resequencing and evaluated the individual effects of identified mutations by introducing them into the wild-type background. We confirmed that methanol was the primary selection pressure through cultivation in a minimal medium. Furthermore, the introduction of a β-carotene biosynthesis pathway led to an increase in titer compared to the wild-type strain. The evolved strain represents a promising chassis for methanol-based biomanufacturing. Materials and methods Strain cultivation and adaptive laboratory evolution The wild-type Pichia pastoris GS115 strain (Invitrogen) was used in this study. A single colony was picked from a yeast extract peptone dextrose (YPD) plate and inoculated into YPD medium in test tubes for overnight cultivation. Cells were transferred to 40 mL of YPM medium at an initial optical density (OD600) of 0.1 in a 250 ml baffled shake flask. The YPM medium consisted of 10 g/L yeast extract (Conda), 20 g/L peptone (BD Difco), and the designated methanol (Sigma-Aldrich) concentration (initially 5% v/v). Cultures were incubated at 30 ℃ and 250 rpm for approximately 48 h until the OD600 reached 2–3, after which they were transferred to a fresh medium at an initial OD600 of 0.1. This process was defined as a single passage and repeated until the 35th passage, at which point the methanol concentration in the YPM medium was increased to 7% (v/v) to impose higher selection pressure. An additional 45 passages were carried out, concluding with the 80th passage. Contamination was monitored at regular intervals by streaking cultures on YPD plate. Specific growth rate measurements For specific growth rate measurements of various P. pastoris strains, a single colony was picked from a YPD plate and inoculated into YPM medium in test tubes containing the designated methanol concentrations (0.5% [v/v] as the default). Once the seed culture reached a sufficient cell density, it was transferred to 40 mL of fresh YPM medium in 250 mL baffled shake flasks at an OD600 of 0.1. OD600 was monitored at regular intervals, and the maximum specific growth rate (µmax) was determined from the exponential growth phase. It was calculated by applying a linear regression to the natural logarithm of the OD600 values plotted against time. Genome resequencing Genomic DNA was extracted from overnight YPD cultures using the Exgene Cell SV kit (Geneall). Sequencing libraries were prepared using the KAPA HyperPlus Kit (Roche), and 150 bp paired-end read sequencing was performed on an Illumina MiniSeq platform. A reference genome of the GS115 laboratory strain was constructed using Breseq based on a previously published genome assembly (GenBank accession: GCA_001746955.1) [18]. The resulting reference genome was used to map sequencing reads from the evolved clones. Mutations consistently identified across clones were validated by Sanger sequencing of PCR products amplified from genomic DNA. Primers used for validation are listed in Supplementary Table S1. Mutant strain cloning To introduce each mutation, BB3cK_pGAP_23*_pPFK300_Cas9 was used, which was a gift from Brigitte Gasser (Addgene plasmid # 104910; http://n2t.net/addgene:104910; RRID: Addgene_104910) [19]. sgRNA sequences were designed using Benchling, and primers were generated accordingly. Primers and sgRNA sequences are listed in Supplementary Tables S1 and S2. sgRNA cassettes were amplified using Phusion Polymerase (Thermo Fisher Scientific) and were assembled with Golden Gate assembly using T4 DNA ligase (Bioneer) and BbsI (New England Biolabs). The plasmids constructed in this study are listed in Supplementary Table S3. Donor sequences were either directly amplified from the genomic DNA of mutant P. pastoris strains or cloned into a pUC19 vector backbone using Gibson assembly (Codex) for deletion constructs. Mach T1 competent Escherichia coli was used for plasmid cloning. Cloned Cas9 plasmid and PCR amplified donor cassette were electroporated into wild-type P. pastoris. Transformants carrying the target mutations were screened by colony PCR using EmeraldAmp (Takara) and mutation-specific mismatched primers [20]. The strains constructed and used in this study are listed in Supplementary Table S4. Formaldehyde concentration measurement The concentration of formaldehyde in the culture was measured using the Nash reaction method adapted for a 96-well plate format [21]. Briefly, cell cultures were centrifuged, and 125 µl of supernatant was transferred to a new 96-well plate. A standard curve was prepared using a formaldehyde solution (Sigma-Aldrich) with final concentrations ranging from 0 to 800 µM. An equal volume of Nash reagent (5 M NH4OAc and 50 mM acetylacetone) was added to each well. The plate was then incubated at 37 ℃ for 1 h, and the absorbance was subsequently measured at 412 nm using a Hidex Sense Microplate Reader (Hidex). The measured formaldehyde concentration was then normalized to the OD600 of the corresponding timepoint. Culture in minimal medium Single colonies were inoculated into 5 mL of YPM medium containing 3% (v/v) methanol in test tubes. Once the seed cultures reached a sufficient cell density, the cells were harvested, washed with distilled, deionized water, and resuspended in a buffered minimal methanol medium (BMM). The BMM composition included 1.34% yeast nitrogen base without amino acids (Sigma Aldrich), 0.1 M potassium phosphate buffer (pH = 6), 4 × 10− 5% biotin (Sigma Aldrich), and 3% (v/v) methanol. Histidine (Sigma Aldrich) was supplemented at 14.8 mg/L to support the histidine auxotrophy of the GS115 strain based on the yeast complete supplement mixture. Once the secondary seed cultures reached sufficient growth, cells were transferred to a 40 mL of fresh BMM medium containing 5% (v/v) methanol in 250 mL baffled shake flasks at an initial OD600 of 0.1. OD600 was monitored at regular intervals. Residual methanol concentrations were measured using a 1260 Infinity II LC system (Agilent) equipped with a Hi-Plex-H column (Agilent). The mobile phase was 5 mM H2SO4, with a flow rate of 0.5 mL/min and a column temperature of 50 ℃. β-carotene biosynthesis and measurement The coding sequences of CarRP (AJ250827.1) and CarB (AJ238028.1) from Mucor circinelloides and GGPP synthase (DQ016502.1) from Xanthophyllomyces dendrorhous were previously synthesized in-house [22]. Each sequence was amplified and assembled into the pPIC3.5 K (Thermo Fisher Scientific) using Gibson assembly. The resulting plasmid was linearized with SacI (New England Biolabs) and integrated into the AOX1 locus via electroporation. Single colonies of recombinant strains were inoculated into a BMM medium containing 2% (v/v) methanol. Once the seed cultures reached sufficient growth, the cells were harvested and transferred to fresh 40 mL BMM medium containing 3% (v/v) methanol in 250 mL baffled shake flasks, followed by cultivation for 120 h. Carotenoids were extracted from harvested cell pellets as previously described [23]. Extracted β-carotene and analytical standards (Sigma-Aldrich) were analyzed by 1260 Infinity II LC system (Agilent) equipped with an InfinityLab Poroshell 120 EC-C18 column (Agilent) under the conditions described previously [24]. Results Long-term ALE in YPM to improve methanol utilization The methanol utilization capacity of wild-type Pichia pastoris was evaluated by measuring its specific growth rate in a YPM medium containing increasing concentrations of methanol. The GS115 strain exhibited a specific growth rate of 0.231 h− 1 in 1% (v/v) YPM, comparable to 0.266 h− 1 in YPD. Growth gradually decreased as methanol concentration increased, with 5% being the maximum concentration at which GS115 was able to grow, exhibiting a specific growth rate of 0.0582 h− 1 (Fig. 1A). Fig. 1Long-term adaptive laboratory evolution of P. pastoris. (A) The specific growth rate of wild-type P. pastoris in YPM medium with increasing methanol concentration. Growth in glucose-fed YPD medium is shown as a red dotted line. (B) Schematic of ALE passage progression. Methanol concentration was increased in the 35th passage to apply higher selection pressure. (C) Changes in specific growth rate (µmax) over passage number. (D) Specific growth rate of isolated strains at different methanol concentrations. PM5 and PM7 indicate strains isolated from the final passage of 5% YPM and 7% YPM, respectively. N.G., no growth observed. Values represent the average of two biological replicates (n = 2) in (A) and (D). The error bar indicates the standard deviation, and the white dots indicate individual data points Long-term adaptive laboratory evolution of P. pastoris. (A) The specific growth rate of wild-type P. pastoris in YPM medium with increasing methanol concentration. Growth in glucose-fed YPD medium is shown as a red dotted line. (B) Schematic of ALE passage progression. Methanol concentration was increased in the 35th passage to apply higher selection pressure. (C) Changes in specific growth rate (µmax) over passage number. (D) Specific growth rate of isolated strains at different methanol concentrations. PM5 and PM7 indicate strains isolated from the final passage of 5% YPM and 7% YPM, respectively. N.G., no growth observed. Values represent the average of two biological replicates (n = 2) in (A) and (D). The error bar indicates the standard deviation, and the white dots indicate individual data points To expand the range of methanol concentrations supporting growth, we conducted adaptive laboratory evolution (ALE) to select strains with enhanced methanol tolerance and improved methanol utilization through long-term adaptation. ALE was initiated in shake flasks using 5% YPM medium, the highest methanol concentration that supported wild-type growth (Fig. 1B). Cells were serially passaged at approximately 48-hour intervals during exponential growth, with each passage initiated at an OD600 of 0.1. Once the specific growth rate stabilized within a specific range, the methanol concentration was increased to 7% at the 35th passage to apply higher selection pressure. ALE was continued for an additional 45 passages, concluding at the 80th passage (Fig. 1C). At the final passage in 5% YPM, ALE resulted in a specific growth rate of 0.166 h− 1, representing a 6.19-fold higher than that of the initial wild-type strain passage. The final passage in 7% YPM achieved a specific growth rate of 0.0958 h− 1, whereas the wild-type strain had previously failed to grow under these conditions. To assess changes in methanol tolerance and utilization following ALE, final passage populations from each methanol concentration were streaked to isolate single colonies. Among the randomly selected colonies, three fast-growing isolates from each condition were selected and designated as PM5a-c for those isolated from 5% YPM and PM7a-c for 7% YPM, respectively. To evaluate whether exposure to higher methanol concentrations confers a greater selective advantage, specific growth rates of evolved strains were compared under the same conditions as during ALE, namely 5% and 7% YPM. The evolved strains PM5a-c exhibited specific growth rates of 0.153, 0.166, and 0.142 h− 1, respectively, while PM7a-c showed specific growth rates of 0.179, 0.182, and 0.162 h− 1. Their average specific growth rates, 0.154 and 0.174 h− 1, correspond to 2.65- and 2.99-fold higher, respectively, than that of wild-type GS115 in 5% YPM (Fig. 1D). Notably, under 5% YPM, PM7a-c showed no significant growth advantage compared to PM5a-c. In contrast, under 7% YPM, PM7a-c exhibited specific growth rates of 0.101, 0.0895, and 0.0938 h− 1, respectively, with an average of 0.0948 h− 1. Identification of beneficial mutations for enhanced growth on methanol medium To identify the genetic changes responsible for the improved methanol tolerance and growth observed in ALE strains, genome resequencing was performed on the evolved isolates and the wild-type GS115 strain. Through resequencing, between 3 and 11 mutations were identified depending on the strain (Supplementary Table S5). Mutations located in intergenic regions or rDNA were excluded from further analysis, and only those in coding regions were summarized, as convergent mutations across multiple strains are most likely to contribute to the observed phenotypes (Table 1). Mutations in eIF4G (eukaryotic translation initiation factor 4G) and PSR1 (plasma membrane associated protein phosphatase), initially identified in strain PM5b, were conserved in all three strains isolated from the 7% YPM condition. Similarly, mutations in a USP-family deubiquitinase (AT250_GQ6801843), MIOX (myo-inositol oxygenase), and BFA1 (subunit of a GTPase-activating protein) were consistently observed in all three isolates. eIF4G and the USP-family deubiquitinase carried missense single nucleotide variants (SNVs), whereas PSR1 and MIOX acquired premature stop codons due to insertions or deletions (indels). Additionally, BFA1 harbored a premature stop codon resulting from a point mutation. Table 1Identified mutations in the coding regionStrainTypeMutationGeneAnnotationsPM5aSNVD137E (GAT→GAG) SAP30 Subunit of histone deacetylase complexPM5bSNVP116S (CCC→TCC) AT250_GQ6805313 Protein with a Plectin/S10 domainPM7aSNVQ430R (CAA→CGA) AT250_GQ6801955 Pho85p cyclin of the Pho80p subfamilyIndel(G)10→11 (928/996 nt) HAC1 Unfolded protein response transcription factorPM7bIndelΔ2bp (268/744 nt) UBC6 Ubiquitin-conjugating enzyme in the ERAD pathwayPM7cSNVK21R (AAA→AGA) SUP45 Translation termination release factorPM5c, PM7cSNVE347* (GAA→TAA) SSM4 Membrane-embedded ubiquitin-protein ligase and retrotranslocasePM5b, PM7a, b, cSNVG951S (GGC→AGC) eIF4G Eukaryotic translation initiation factor 4GIndel(A)7→8 (7/1137 nt) PSR1 Plasma membrane associated protein phosphatasePM7a, b, cSNVD138Y (GAT→TAT) AT250_GQ6801843 Ubiquitin-specific protease (USP)SNVE208* (GAA→TAA) BFA1 Subunit of a GTPase-activating proteinIndelΔ2bp (GA) (306/1044 nt) MIOX Myo-inositol oxygenaseLocus Tag is shown when a curated gene name is unavailable. Type abbreviations: SNV, single nucleotide variant; Indel, insertion/deletion Table 1Identified mutations in the coding regionStrainTypeMutationGeneAnnotationsPM5aSNVD137E (GAT→GAG) SAP30 Subunit of histone deacetylase complexPM5bSNVP116S (CCC→TCC) AT250_GQ6805313 Protein with a Plectin/S10 domainPM7aSNVQ430R (CAA→CGA) AT250_GQ6801955 Pho85p cyclin of the Pho80p subfamilyIndel(G)10→11 (928/996 nt) HAC1 Unfolded protein response transcription factorPM7bIndelΔ2bp (268/744 nt) UBC6 Ubiquitin-conjugating enzyme in the ERAD pathwayPM7cSNVK21R (AAA→AGA) SUP45 Translation termination release factorPM5c, PM7cSNVE347* (GAA→TAA) SSM4 Membrane-embedded ubiquitin-protein ligase and retrotranslocasePM5b, PM7a, b, cSNVG951S (GGC→AGC) eIF4G Eukaryotic translation initiation factor 4GIndel(A)7→8 (7/1137 nt) PSR1 Plasma membrane associated protein phosphatasePM7a, b, cSNVD138Y (GAT→TAT) AT250_GQ6801843 Ubiquitin-specific protease (USP)SNVE208* (GAA→TAA) BFA1 Subunit of a GTPase-activating proteinIndelΔ2bp (GA) (306/1044 nt) MIOX Myo-inositol oxygenaseLocus Tag is shown when a curated gene name is unavailable. Type abbreviations: SNV, single nucleotide variant; Indel, insertion/deletion Identified mutations in the coding region Locus Tag is shown when a curated gene name is unavailable. Type abbreviations: SNV, single nucleotide variant; Indel, insertion/deletion To evaluate the contributions of individual mutations to the evolution of the phenotype, each mutation was introduced into the wild-type GS115 background, generating the strains eIF4Gm, PSR1m, USPm, BFA1m, and MIOXm, respectively (Fig. 2A). For PSR1m, the gene was deleted because a viable point mutant could not be obtained; the original ALE-derived mutation introduced a premature stop codon at the 13th amino acid. The mutant strains were cultured in 5% and 7% YPM to assess their specific growth rates. In 5% YPM, PSR1m showed the highest specific growth rate of 0.0990 h− 1, a value 1.70-fold higher than that of the wild-type (Fig. 2B). The eIF4Gm and BFA1m strains exhibited moderate increases, with specific growth rates of 0.0698 and 0.0754 h− 1, representing 1.20-fold and 1.30-fold higher rates than the wild-type, respectively. These results suggest that the respective mutations may contribute to cellular fitness, including improved methanol tolerance and utilization. In 7% YPM, BFA1m showed a specific growth rate of 0.0563 h− 1, a 1.15-fold higher rate compared to the 36th passage, whereas the wild-type strain failed to grow (Fig. 2C). Fig. 2Consequences of each identified mutation on methanol tolerance and growth. (A) Schematic overview of individual mutation validation strategy. Individual mutations identified by genome resequencing were introduced into wild-type P. pastoris using CRISPR-Cas9-mediated homologous recombination. Specific growth rates of the resulting mutant strains were measured in 5% YPM (B) and 7% YPM (C). Wild-type GS115 and selected ALE passages, 35th in (B) and 36th and 80th in (C), were included as controls. N.G., no growth observed. Values represent the average of three biological replicates (n = 3). The error bar indicates the standard deviation, and white dots indicate individual data points. P-values were determined by the two-tailed Student’s t-test compared to each initial strain. Asterisks denote the P-value. *P < 0.05, **P < 0.01, ***P < 0.001 Consequences of each identified mutation on methanol tolerance and growth. (A) Schematic overview of individual mutation validation strategy. Individual mutations identified by genome resequencing were introduced into wild-type P. pastoris using CRISPR-Cas9-mediated homologous recombination. Specific growth rates of the resulting mutant strains were measured in 5% YPM (B) and 7% YPM (C). Wild-type GS115 and selected ALE passages, 35th in (B) and 36th and 80th in (C), were included as controls. N.G., no growth observed. Values represent the average of three biological replicates (n = 3). The error bar indicates the standard deviation, and white dots indicate individual data points. P-values were determined by the two-tailed Student’s t-test compared to each initial strain. Asterisks denote the P-value. *P < 0.05, **P < 0.01, ***P < 0.001 To investigate the potential mechanisms behind these improvements, formaldehyde levels in the culture medium were measured for all mutants that showed significant growth benefits. Interestingly, among these mutants, only the BFA1m strain maintained low formaldehyde levels comparable to the evolved PM7a strain (Supplementary Figure S1). Although the growth improvement was modest, these data support a tolerance mechanism involving improved formaldehyde management for the disruption of BFA1, which was not evident for eIF4G or PSR1 mutations. Overall, most single mutations did not substantially improve growth, and several strains failed to grow in 7% YPM. Notably, the few mutations associated with improved growth-such as those in PSR1 and BFA1-introduced premature stop codons. In contrast, the 2-bp indel in MIOX, which causes a frameshift, did not result in noticeable growth changes. Evolved strain enables growth on minimal methanol medium To evaluate the impact of ALE on methanol metabolism, we compared the performance of evolved strains to that of wild-type GS115. The PM7a strain, which exhibited the highest specific growth rate in 7% YPM, was selected as the representative ALE strain. While ALE was performed in the YPM medium, in which glucose was replaced with methanol in the commonly used yeast culture medium YPD, we used buffered minimal methanol medium (BMM) to isolate the effects of methanol metabolism by minimizing the influence of yeast extract and peptone. Because GS115 is a histidine auxotroph, histidine was supplemented in 5% BMM based on the composition of the yeast complete supplement mixture to ensure consistent nutrient availability. For seed culture preparation, cells were first grown in YPM and then harvested at mid-log phase, then transferred to histidine-supplemented BMM to minimize the lag phase associated with medium transition (Fig. 3A). Fig. 3Biomass accumulation and methanol consumption during cultivation in minimal medium. (A) Schematic of the minimal medium culture procedure. Wild-type GS115 and the fastest-growing 7% YPM isolate, PM7a, were first inoculated in 3% YPM and then transferred to a buffered minimal methanol medium (BMM) containing 3% methanol to minimize the lag phase by adaptation to the medium. (B) Changes in OD600 (filled circles) and residual methanol concentration (open circles) during cultivation in BMM containing 5% methanol. Values represent the average of three biological replicates (n = 3). The error bar indicates the standard deviation Biomass accumulation and methanol consumption during cultivation in minimal medium. (A) Schematic of the minimal medium culture procedure. Wild-type GS115 and the fastest-growing 7% YPM isolate, PM7a, were first inoculated in 3% YPM and then transferred to a buffered minimal methanol medium (BMM) containing 3% methanol to minimize the lag phase by adaptation to the medium. (B) Changes in OD600 (filled circles) and residual methanol concentration (open circles) during cultivation in BMM containing 5% methanol. Values represent the average of three biological replicates (n = 3). The error bar indicates the standard deviation The wild-type GS115 failed to grow in BMM, likely due to increased sensitivity to methanol toxicity under nutrient-limited conditions compared to YPM (Fig. 3B). A slight decrease in residual methanol concentration was observed, which may be attributable to spontaneous evaporation. In contrast, PM7a exhibited clear growth after a short lag phase, with a specific growth rate of 0.0836 h− 1 and reaching an OD600 of 6.20. Correspondingly, the methanol concentration decreased to 3.68%, indicating active consumption coupled with biomass accumulation. Although growth ceased before all the methanol was consumed, this was not due to evaporation, which was found to be minimal (0.07% per 24 h), nor was it due to formaldehyde toxicity as formaldehyde did not accumulate in the culture medium (Supplementary Figure S2). From these data, the biomass yield on methanol was calculated to be approximately 0.172 g DCW/g MeOH, indicating substantial biomass formation from methanol under high-stress conditions. To further characterize the evolved strain and investigate potential fitness trade-offs, we also compared its growth in non-selective YPD and low-stress 1% BMM conditions (Supplementary Figure S3). In YPD medium, the specific growth rates of the two strains were nearly identical, indicating that the evolution for high-methanol tolerance did not reveal an obvious fitness cost during growth on glucose. In 1% BMM, the two strains exhibited distinct growth dynamics; while the wild-type GS115 initially reached a higher cell density, the evolved PM7a strain ultimately achieved a higher final biomass. Enhanced production of β-carotene using methanol by an evolved PM7a strain To evaluate whether ALE-driven adaptations lead to improved production performance, we used β-carotene as a model product. Strain PM7a, the fastest-growing ALE isolate, was selected for pathway integration. The β-carotene biosynthetic pathway was constructed by introducing heterologous CarRP and CarB from M. circinelloides (Fig. 4A) [25]. To enhance the supply of the precursor geranylgeranyl pyrophosphate (GGPP), a GGPP synthase from X. dendrorhous with high reported production efficiency was introduced. All three genes were integrated into the AOX1 locus under native P. pastoris promoters and terminators, yielding the strains GS115-βCar and PM7a-βCar for wild-type GS115 and PM7a, respectively (Fig. 4B). Cultures were conducted in BMM to directly compare production performance under methanol as the sole carbon source. Unlike in previous growth assays, histidine was not supplemented because it was used as the auxotrophic marker for β-carotene pathway integration. The high methanol tolerance of the engineered strain was confirmed by its growth in a 5% BMM medium, a condition under which the wild-type strain could not grow (Supplementary Figure S4). To enable a direct performance comparison with the wild-type strain, the methanol concentration was subsequently reduced to 3%. Fig. 4β-carotene production performance under high methanol conditions. (A) Schematic of the metabolic pathway for β-carotene synthesis from methanol. Introduced heterologous genes are highlighted. GGPPS, geranylgeranyl pyrophosphate synthase; CarRP, phytoene synthase/lycopene cyclase; CarB, phytoene dehydrogenase. (B) Diagram of the integration cassette targeting the AOX1 locus. Native P. pastoris promoters and terminators were used. (C) β-carotene production titers in engineered strains after 120 h of cultivation in BMM containing 3% methanol. GS115-βCar and PM7a-βCar strains denote wild-type GS115 and PM7a strains integrated with β-carotene production cassettes, respectively. Values represent the average of three biological replicates (n = 3). The error bar indicates the standard deviation, and white dots indicate individual data points. Asterisks denote the P-value. *P < 0.05, **P < 0.01, ***P < 0.001 β-carotene production performance under high methanol conditions. (A) Schematic of the metabolic pathway for β-carotene synthesis from methanol. Introduced heterologous genes are highlighted. GGPPS, geranylgeranyl pyrophosphate synthase; CarRP, phytoene synthase/lycopene cyclase; CarB, phytoene dehydrogenase. (B) Diagram of the integration cassette targeting the AOX1 locus. Native P. pastoris promoters and terminators were used. (C) β-carotene production titers in engineered strains after 120 h of cultivation in BMM containing 3% methanol. GS115-βCar and PM7a-βCar strains denote wild-type GS115 and PM7a strains integrated with β-carotene production cassettes, respectively. Values represent the average of three biological replicates (n = 3). The error bar indicates the standard deviation, and white dots indicate individual data points. Asterisks denote the P-value. *P < 0.05, **P < 0.01, ***P < 0.001 The PM7a-βCar strain reached an OD600 of 3.76, compared to 0.999 for the GS115-βCar strain (Supplementary Figure S5). Accordingly, β-carotene titer increased from 0.959 mg/L in GS115-βCar to 4.71 mg/L in PM7a-βCar, representing a 4.91-fold higher titer (Fig. 4C). To further evaluate the metabolic conversion efficiency, we calculated the yield of β-carotene from the consumed methanol. The evolved PM7a-βCar strain exhibited 0.238 mg β-carotene per g methanol, 2.58-fold higher than the GS115-βCar yield of 0.614 mg/g. We also compared specific productivity, expressed as mg β-carotene per g dry cell weight, which was similar between strains (GS115-βCar: 2.61; PM7a-βCar: 2.70 mg/g DCW). Together, these results indicate a growth-linked improvement leading to higher product obtained per unit methanol while per-biomass productivity remained comparable. Discussion In this study, adaptive laboratory evolution (ALE) was conducted to enhance the methanol tolerance and utilization of P. pastoris under high methanol concentrations. The wild-type GS115 strain exhibited a decline in specific growth rate with increasing methanol concentrations, consistent with previous observations that elevated methanol levels are toxic to cells. This is likely attributable to methanol-induced toxicity, such as formaldehyde-mediated crosslinking between nucleic acids and proteins [26]. During the early stages of ALE at both methanol concentrations, a sharp increase in specific growth rate was observed. This trend likely reflects physiological acclimation to high-methanol stress rather than by the immediate fixation of new mutations [27, 28]. After this short-timescale adjustment, growth rate plateaued, indicating a transition to slower, mutation-limited dynamics. Transient declines in specific growth rate were occasionally observed during ALE. Such decreases may reflect population dynamics associated with crossing fitness valleys, where beneficial mutations initially impose a fitness cost before leading to improved adaptation [29]. In this context, methanol-tolerant strains likely outcompeted fast-growing but less tolerant ones by better surviving early cell death under toxic conditions, thereby causing a modest reduction in population-level specific growth rate. At the end of ALE, evolved strains from the 7% methanol condition, PM7a-c, showed similar specific growth rates with those from the 5% condition, PM5a-c, when cultured in 5% YPM medium. This result suggests that ALE at higher methanol concentrations might have exerted greater selective pressure on tolerance, thereby enhancing overall cell fitness. The higher methanol stress may have facilitated the enrichment of genotypes with improved survival mechanisms, even if such mutations were initially costly in terms of growth. The mutations identified through genome resequencing occurred in genes that are not directly involved in the methanol utilization pathway. Accumulation of non-pathway-specific mutations during ALE is not unexpected, as similar cases have been reported in previous studies [30–32]. However, the functional implications of these mutations appear distinct from those in previous ALE studies on P. pastoris. For instance, one study focused on phenotypic adaptations such as changes in cell morphology and membrane composition, while another identified mutation directly within the AOX1, a key enzyme in methanol metabolism [15, 16]. In contrast, our genome resequencing results pinpoint mutations in genes such as PSR1 and BFA1, genes linked to stress signaling and cell cycle control. Functionally, these findings suggest modulation of regulatory circuit that improves fitness under methanol rather than direct changes to the methanol pathway. The functional context of the mutated genes supports their potential contribution to the enhanced phenotype observed in ALE strains. PSR1 encodes a plasma membrane associated protein phosphatase involved in the general stress response [33, 34]. Its loss of function, caused by a premature stop codon, may indirectly modulate cellular responses to methanol-induced stress. Specifically, the cell may be less responsive to methanol-induced stress signals that would normally trigger a protective cell-cycle arrest. This could allow the cell to reduce engagement of growth-inhibiting pathways, thereby shifting resources from prolonged stress programs toward biomass accumulation. However, this strategy exhibits a clear fitness trade-off, as the PSR1m strain, while showing the fastest growth at 5% methanol, failed to grow at the higher concentration of 7%. This pattern suggests that the PSR1-mediated stress response, although a constraint on growth under moderate stress, becomes increasingly important for survival when methanol toxicity is severe. BFA1, a component of a GTPase-activating protein complex, is involved in multiple cell cycle checkpoint pathways [35]. Disruption of BFA1 may modulate cell cycle dynamics in favor of survival or stress resistance under high methanol concentrations. For instance, a weakened mitotic exit checkpoint due to the BFA1 mutation could reduce the time cells spend in checkpoint-induced arrest, thereby contributing to a higher net population growth with a possible trade-off in mitotic fidelity. In addition to this role, the BFA1m strain demonstrated an enhanced capacity for formaldehyde management, maintaining low concentration comparable to the PM7a strain. This result suggests that the BFA1 mutation contributes to methanol tolerance through an additional mechanism involving improved formaldehyde management. The unique conservation of the BFA1 mutation across all PM7 isolates, combined with its demonstrated role in enhancing formaldehyde management, suggests that the acquisition of the BFA1 mutation was a key evolutionary event responsible for the tolerance of the PM7 strains at the high methanol concentrations. eIF4G, a core component of the translation initiation complex, plays a central role in protein synthesis and the assembly of stress granules [36, 37]. The missense mutation in eIF4G may alter translation-initiation dynamics, decreasing stress-induced shutdown or favoring initiation on detoxification and chaperone mRNAs. Such reprogramming could adjust the proteome to improve tolerance under methanol stress. Taken together, these results indicate that while most individual mutations have limited effects, a subset of loss-of-function variants can enhance methanol tolerance and may serve as valuable targets for future strain engineering. One exception may be the 2-bp frameshift mutation in MIOX, which did not yield a detectable growth phenotype but may have been passively fixed during ALE without functional relevance. However, its potential role could emerge in combination with other untested mutations, including those in intergenic regions. Beyond single-locus reconstructions, dissecting the evolved phenotype will require building defined double/triple combinations and quantifying epistasis under methanol selection. Because marker-less homology-directed repair in wild-type P. pastoris is inefficient, future work will deploy higher-efficiency SNP editing strategies (e.g., base editors) where permissible by target transitions. The evolved strain PM7a exhibited robust growth under nutrient-limited minimal conditions, in contrast to the complete growth failure of wild-type GS115. Comparison with previous growth data for X-33 in 1% BMM shows that PM7a achieves 94% of X-33’s specific growth rate, despite being cultured at a higher methanol concentration of 5% [16]. This highlights a significant improvement in methanol tolerance and growth capacity of PM7a under elevated methanol stress. While other components in YPM may have contributed to adaptation during ALE, these results confirm that methanol served as the primary selective pressure driving evolutionary improvement. Despite the cessation of growth, methanol was not completely consumed. Given that evaporation was negligible and no formaldehyde accumulation was detected in the culture medium, the most plausible explanation is the exhaustion of another essential nutrient (e.g., nitrogen or histidine) in the minimal medium. The enhanced methanol tolerance and utilization of PM7a under minimal conditions also translated into enhanced production capacity, as evidenced by increased β-carotene biosynthesis. Although the specific productivity per unit biomass was comparable between the evolved and wild-type strains, the evolved PM7a-βCar strain formed more biomass per unit methanol consumed, which increased the product yield from methanol and final titer compared to the GS115-βCar background. Previous studies have reported β-carotene production in P. pastoris under YPD-based conditions or through methanol feeding with staged induction using BMM medium [38–40]. These strategies typically rely on rich supplements and controlled induction. In contrast, our approach demonstrates substantial titer in a minimal medium, with methanol as the sole carbon source and without induction or feeding steps. These results suggest that ALE-driven mutations improve methanol tolerance and utilization, which promotes biomass accumulation and increases product formation. However, the performance of the evolved strain appeared to be specific to high-stress conditions. When cultured at a moderate methanol concentration of 1.25%, the PM7a-βCar strain exhibited growth comparable to that of the wild-type but produced a slightly lower β-carotene titer (Supplementary Figure S6). This observation indicates a distinct fitness trade-off, wherein mutations conferring robustness at high methanol concentrations do not translate to superior performance under less stressful conditions and may impose a slight metabolic cost on secondary metabolite production. Notably, cultures were performed at a methanol concentration lower than that used during ALE to ensure comparability with the wild-type. It is possible that selecting isolates from intermediate passages with faster growth but lower tolerance might have yielded strains with higher production capacity, albeit potentially less robust under stress. Conclusions Long-term adaptive evolution of P. pastoris GS115 in 5% and 7% YPM resulted in a strain with enhanced methanol tolerance and utilization. The evolved strain achieved a specific growth rate of 0.101 h− 1 in 7% methanol, a condition in which the wild-type strain failed to grow. Genome resequencing identified multiple mutations, and reverse engineering of single mutants demonstrated that mutations in PSR1 and BFA1 notably contributed to improved growth. Among the evolved variants, PM7a exhibited enhanced methanol adaptability, as it was able to grow in a minimal medium with methanol as the sole carbon source, a condition under which GS115 did not grow detectably. Furthermore, the integration of a β-carotene biosynthetic module into PM7a resulted in a 4.91-fold higher product titer compared to the wild-type strain. These findings highlight the potential of ALE to develop P. pastoris as a high-performance methylotrophic chassis. Further investigation of combinatorial effects among identified mutations may enable the development of additional strategies for advanced P. pastoris engineering. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1: Supplementary tables containing data related to cloning constructs and mutation analysis, as well as supplementary figures illustrating growth profiles, methanol consumption, and related analyses. Supplementary Material 1: Supplementary tables containing data related to cloning constructs and mutation analysis, as well as supplementary figures illustrating growth profiles, methanol consumption, and related analyses. Supplementary Material 2: Raw optical density (OD600) data recorded over time for growth analysis of the indicated strains under various methanol concentrations. Supplementary Material 2: Raw optical density (OD600) data recorded over time for growth analysis of the indicated strains under various methanol concentrations.