Evolution of alcohol oxidase for improved methanol bioconversion and formaldehyde tolerance Authors: Xu-Wei Ding, Ya-Juan Sun, Xiao-Yu Hu, Jun-Yi Yang, Jian-He Xu, Gao-Wei Zheng, Chun-Xiu Li Published: 2027 ABSTRACT Methanol oxidation by alcohol oxidases (AOXs) is a key bottleneck in one-carbon (C1) bioconversion due to limited catalytic efficiency and poor formaldehyde tolerance. Here, we report the directed evolution of an alcohol oxidase from Gloeophyllum trabeum, yielding an optimized variant, GtAOXM3. The engineered enzyme exhibits a sixfold increase in catalytic efficiency (7.7 s−1 mM−1), together with enhanced formaldehyde tolerance, thermostability, and high methanol specificity. Molecular dynamics simulations suggest that increased global rigidity and cooperative residue dynamics contribute to the improved performance. When incorporated in multienzyme cascade systems, GtAOXM3 enables efficient conversion of methanol to the value-added chemicals dihydroxyacetone (34.5 mM) and ethylene glycol (23.3 mM). This work establishes GtAOXM3 as an efficient and cost-effective biocatalyst for methanol-based C1 biotransformation. Introduction Climate change and the energy crisis represent two of the most pressing challenges of the 21st century [1]. Conventional chemical manufacturing remains heavily dependent on fossil resources and food-based feedstocks, leading to substantial greenhouse gas emissions, environmental burdens, and intensified competition with the food supply chains [2,3]. In this context, third-generation (3G) biorefineries have emerged as a promising and sustainable alternative [4]. Unlike first- and second-generations, which depend on food crops or lignocellulosic biomass, 3G biorefineries utilize one-carbon (C1) compounds—such as CO, CO2, methane, methanol, and formic acid—as feedstocks for the production of value-added chemicals via biological, chemical, or hybrid catalytic processes [5,6]. These C1 feedstocks are abundant, cost-effective, and independent of agricultural resources, rendering them particularly attractive for sustainable biomanufacturing. Among C1 feedstocks, methanol and formic acid have attracted considerable attention owing to their relative stability and low toxicity [7]. However, the biological utilization of formic acid remains challenging. Owing to its highly oxidized state, formic acid must overcome a substantial energetic barrier and be either reduced to formaldehyde or activated (e.g., via acylation or phosphorylation) prior to assimilation, which limits broader application [8]. In contrast, methanol-based pathways are energetically more favorable and have therefore been intensively explored [9]. Biological methanol oxidation is primarily mediated by three classes of enzymes [10]: O2-dependent alcohol oxidases (AOXs) [11], NAD-dependent alcohol dehydrogenases (ADHs) [12], and pyrroloquinoline quinone (PQQ)-dependent methanol dehydrogenases (MDHs) [13]. NAD-dependent ADHs generally exhibit high recombinant expression levels but catalyze thermodynamically unfavorable reactions that require efficient cofactor regeneration. PQQ-dependent MDHs are constrained by the high cost and limited availability of the PQQ cofactor, hindering large-scale industrial application. By comparison, AOXs utilize molecular oxygen as the terminal electron acceptor and catalyze thermodynamically favorable reactions, making them among the most promising candidates for methanol bioconversion. Nevertheless, AOXs are predominantly derived from fungi and localized in peroxisomes, where methanol oxidation is accompanied by the generation of hydrogen peroxide [14,15]. This may compromise enzyme stability and promote the overoxidation of formaldehyde to formic acid. As a result, AOX-based systems typically require the coexpression or supplementation of catalase (CAT) to decompose hydrogen peroxide in situ and maintain redox balance [16]. To date, several artificial biosynthetic pathways employing methanol as the sole C1 feedstock have been reported. In these systems, methanol—often derived from the electrochemical reduction of CO2—is converted into high-value chemicals, such as starch [17], hexose sugars [18], l-lactic acid [19,20], l-alanine [21], and 2-keto-4-hydroxybutyrate [22], through in vitro multienzyme cascade reactions [23]. Most of these routes rely on commercially available alcohol oxidases from Pichia pastoris (PpAOX), whose high cost hampers large-scale industrial deployment. Moreover, AOXs often suffer from limited catalytic efficiency and formaldehyde inhibition, constraining overall pathway efficiency [24,25]. Notably, several alternative AOXs capable of functional expression in E. coli have been identified, including MrAOX [19], GtAOX [20], PfAOX [22], PcAOX [26], and HpAOX [27] (Table 1). These enzymes provide important catalytic modules for C1 bioconversion, enabling further pathway engineering and optimization. Hence, the development of cost-effective AOXs with enhanced catalytic activity and formaldehyde tolerance remains highly desirable.Table 1Overview of reported methanol oxidases heterologously expressed in E. coli.Table 1EnzymeOrganismkcat (s−1)KM (mM)kcat/KM (s−1 mM−1)Ref.MrAOXMoniliophthora roreri8.93.72.419GtAOXGloeophyllum trabeum2.88.30.320PfAOXPestalotiopsis fici2.31.02.322PcAOXPhanerochaete chrysosporium182926HpAOXR339AHypoxylon sp.2.63.60.727 Table 1Overview of reported methanol oxidases heterologously expressed in E. coli.Table 1EnzymeOrganismkcat (s−1)KM (mM)kcat/KM (s−1 mM−1)Ref.MrAOXMoniliophthora roreri8.93.72.419GtAOXGloeophyllum trabeum2.88.30.320PfAOXPestalotiopsis fici2.31.02.322PcAOXPhanerochaete chrysosporium182926HpAOXR339AHypoxylon sp.2.63.60.727 Overview of reported methanol oxidases heterologously expressed in E. coli. In this work, we used an alcohol oxidase from Gloeophyllum trabeum (GtAOXM2: V229A/S271G/K297R/L600M/Y605M), which was previously identified in our study [20]. This enzyme had been partially optimized via iterative saturation mutagenesis, improving its catalytic efficiency from 0.34 to 1.38 s−1 mM−1, but it showed limited formaldehyde tolerance. To further enhance its performance, GtAOXM2 was engineered via error-prone PCR and combinatorial mutagenesis to enable efficient methanol oxidation (Fig. 1). The optimized variant, GtAOXM3, harboring three additional mutations (D500G, E526I, and V533L), exhibited a specific activity of 14.1 U/mg, comparable to that of commercially available PpAOX. Notably, GtAOXM3 displayed high soluble expression, along with improved formaldehyde tolerance and thermostability. Substrate scope analysis revealed a strong preference for methanol, with no detectable activity toward pathway intermediate or final products. Finally, integration of GtAOXM3 into multienzyme cascades enabled the synthesis of dihydroxyacetone (DHA) and ethylene glycol (EG) from 150 mM methanol, achieving analytical yields of 69% and 31%, respectively. Collectively, this work provides a high-performance enzymatic module for efficient C1 bioconversion and expands the toolbox for methanol-based synthetic biochemistry.Fig. 1Methanol oxidation to formaldehyde catalyzed by engineeredGtAOX andin vitromultienzyme cascades converting methanol to ethylene glycol (EG) and dihydroxyacetone (DHA). AOX, Alcohol oxidase; CAT, Catalase; BFD, Benzoylformate decarboxylase; BAL, Benzaldehyde lyase; GOX: Alcohol dehydrogenase. Methanol oxidation to formaldehyde catalyzed by engineeredGtAOX andin vitromultienzyme cascades converting methanol to ethylene glycol (EG) and dihydroxyacetone (DHA). AOX, Alcohol oxidase; CAT, Catalase; BFD, Benzoylformate decarboxylase; BAL, Benzaldehyde lyase; GOX: Alcohol dehydrogenase. Methods Reagents and materials Ethylene glycol, dihydroxyacetone, and glycolaldehyde dimer were purchased from Bidepharm (Shanghai, China). 4-Aminoantipyrine and 2,4,6-tribromo-3-hydroxybenzoic acid were purchased from Dibai Biotechnology (Shanghai, China). Plasmid miniprep kits and gel extraction kits were purchased from Aidlab Biotechnologies (Beijing, China). Horseradish peroxidase (HRP, >150 U/mg) was purchased from Macklin Biochemical Technology (Shanghai, China). All genes were synthesized by GenScript (Nanjing, China) and inserted into the pET-28a(+) vector between the NdeI and XhoI restriction sites and expressed as an N-terminally His-tagged fusion protein. Recombinant protein expression and purification The recombinant plasmid harboring the target gene was transformed into E. coli BL21(DE3) chemically competent cells and plated on LB agar containing 50 μg/mL Kan. The single colony was inoculated into 4 mL LB medium and grown at 37 °C and 200 rpm for 8 h. The culture (1%, v/v) was then transferred into TB medium containing Kan and cultivated under the same conditions until the OD600 reached 0.6–0.8. Protein expression was induced with 0.2 mM IPTG, followed by incubation at 16 °C for 20 h. Cells were harvested by centrifugation (12,000 rpm, 4 °C, 3 min) and resuspended in Tris-HCl buffer (50 mM, pH 7.5) containing 300 mM NaCl and 10 mM imidazole. Cell disruption was performed by ultrasonication, and the lysate was clarified by centrifugation (12,000 rpm, 4 °C, 40 min). The supernatant was loaded onto a Ni–NTA affinity column, washed with Tris-HCl buffer (50 mM, pH 7.5) containing 300 mM NaCl and 59 mM imidazole, and eluted with the same buffer supplemented with 255 mM imidazole. Eluted proteins were concentrated using a 10 kDa ultrafiltration cube and buffer-exchanged twice with Tris-HCl buffer (50 mM, pH 7.5) containing 100 mM NaCl and 1 mM DTT. Determination of protein concentration Protein concentrations were determined using two complementary approaches. For rapid screening during directed evolution, protein concentration was estimated by measuring absorbance at 280 nm using a NanoDrop 2000c spectrophotometer, with calculations based on theoretical extinction coefficients obtained from ProtParam (https://web.expasy.org/protparam/). For more accurate determination of functional enzyme concentration used in kinetic and substrate specificity assays, the fraction of FAD-bound protein was measured. Briefly, 1 mL of purified enzyme (0.5–1 mg/mL) was incubated in water at 80 °C for 15 min in the dark, with gentle mixing every 5 min, followed by centrifugation at 12,000 rpm and 4 °C for 10 min. The absorbance at 450 nm of the supernatant was measured using a UV–Vis spectrophotometer, with a blank containing no enzyme. The flavin concentration was calculated based on the molar extinction coefficient of FAD (extinction coefficient: 11,300 M−1 cm−1) [28] and compared to the protein concentration determined by A280 nm. This allowed accurate quantification of the functional FAD-containing fraction of the enzyme. Construction of a high-throughput screening strategy AOX activity was determined using a coupled colorimetric assay (Fig. S4) [29]. In this assay, hydrogen peroxide produced during the enzymatic oxidation of methanol reacts with 4-aminoantipyrine (4-AAP) and 2,4,6-tribromo-3-hydroxybenzoic acid (TBHBA) in the presence of horseradish peroxidase (HRP), yielding a stable magenta quinone-imine derivative with a characteristic absorption at 510 nm (ε = 29.4 mM−1 cm−1). The chromogenic reagent was prepared by dissolving 1 mg HRP (200 U) in 4.87 mL HEPES buffer (100 mM, pH 7.5), followed by the addition of 30 μL 4-AAP (final concentration 33 mM) and 100 μL TBHBA (final concentration 50 mM, DMSO stock). The solution was stored at low temperature prior to use. Determination of formaldehyde Formaldehyde concentration was determined using the acetylacetone colorimetric method (Fig. S5) [30]. In the presence of excess ammonium, acetylacetone reacts with formaldehyde to form a stable yellow compound with a characteristic absorbance at 414 nm. The acetylacetone reagent was prepared by dissolving 0.5 g ammonium acetate, 60 μL acetic acid, and 5 μL acetylacetone in deionized water to a final volume of 1 mL and stored at 4 °C in the dark. For standard curve construction, 1 mL of deionized water or formaldehyde standards (0.2–4.2 μL/mL) were mixed with 100 μL reagent and incubated at 60 °C for 15 min. After cooling and centrifugation at 12,000 rpm for 5 min, absorbance at 414 nm was measured. Formaldehyde concentrations in enzymatic reaction samples were determined following the same procedure after appropriate dilution and calculated based on the standard curve. Directed evolution of GtAOXM2 The GtAOX mutant library was generated using error-prone PCR (epPCR). The recombinant plasmid pET28a(+)-GtAOXM2, containing the GtAOXM2 sequence, served as the template, and primers listed in Table S2 (Entry 1–2) were used for epPCR. The 50 μL reaction mixture contained 2 μL plasmid DNA (1–10 ng), 1.5 μL each of forward and reverse primers, 1.5 μL rTaq DNA polymerase, 2.5 μL MnCl2 (1 M stock), 4 μL dNTPs, 5 μL 10× Mg2+ buffer, 32.5 μL ddH2O, and 1 μL DMSO. PCR conditions were set as follows: 98 °C for 3 min; 30 cycles of 98 °C for 10 s, 55 °C for 15 s, and 72 °C for 2 min; followed by 72 °C for 7 min and storage at 4 °C. In parallel, the pET28a(+) vector was linearized using primers listed in Table S2 (Entry 3–4). The 50 μL reaction mixture included 1 μL template plasmid (1–10 ng), 1.5 μL each primer, 25 μL 2× PrimeSTAR Max mix, 21 μL ddH2O, and 1 μL DMSO. PCR conditions were set as follows: 98 °C for 3 min; 30 cycles of 98 °C for 10 s, 55 °C for 15 s, and 72 °C for 2.5 min; followed by 72 °C for 7 min and storage at 4 °C. PCR products were treated with 2 μL Dpn I for 3 h to remove template DNA and purified via agarose gel electrophoresis. Mutant inserts (10 μL) were assembled with linearized vectors (10 μL) using Exnase II recombinase (20 μL) through a one-step cloning reaction at 50 °C for 30 min. The ligation mixtures were transformed into E. coli BL21(DE3) competent cells and plated on LB agar containing kan. Plates were incubated at 37 °C for 12 h to obtain the mutant library. Single colonies were inoculated into 96-well deep-well plates containing 300 μL LB supplemented with 50 μg/mL kanamycin and cultured at 37 °C, 800 rpm for 12 h. Subsequently, 20 μL of culture was transferred to fresh wells containing 480 μL LB with kan and grown under the same conditions for 4 h. IPTG was then added to a final concentration of 0.2 mM, and protein expression was induced at 16 °C for 20 h. Cells were harvested by centrifugation at 4 °C, 4000 rpm for 10 min and stored at −80 °C overnight. For cell lysis, 200 μL HEPES buffer (100 mM, pH 7.5) containing 1 mg/mL lysozyme was added to each well, followed by shaking at 25 °C, 800 rpm for 2 h. Subsequently, 400 μL HEPES buffer was added, and plates were centrifuged at 4 °C, 4000 rpm for 20 min. The supernatants were collected as crude enzyme extracts. Enzyme activity was measured using the high throughput colorimetric assay described above. Each reaction contained 50 μL chromogenic reagent, 130 μL HEPES buffer (100 mM NaCl, 5 mM MgCl2, pH 7.5), methanol (10 mM final concentration), and 20 μL crude extract. Meanwhile, an additional 100 mM formaldehyde was supplemented as a screening pressure. Reactions were performed at 30 °C for 10 min, and absorbance at 510 nm was recorded. Mutants exhibiting higher activity than the parental enzyme were sequenced and subsequently purified for detailed activity assays. Activity and kinetic assays The enzyme assay mixture (1 mL) contained 10 mM methanol, 0.002 mg/mL purified AOX, 25% (v/v) colorimetric reagent, and 100 mM HEPES buffer (containing 100 mM NaCl, pH 7.5). Reactions were conducted at 30 °C for 3 min, and the absorbance at 510 nm was immediately recorded. Control reactions without AOX were used to correct for background. One unit of enzyme activity (1 U) was defined as the amount of enzyme catalyzing the formation of 1 μmol H2O2 from methanol per minute under the assay conditions. Kinetic parameters were determined by measuring enzyme activity over a substrate concentration range of 0–120 mM methanol. Data were fitted to the Michaelis-Menten equation using GraphPad Prism 8.0.2 to obtain KM and kcat values. Formaldehyde tolerance assay The residual activity of AOX in the presence of 0–50 mM formaldehyde was measured using the same reaction system described in Section 2.7. Different concentrations of formaldehyde were added to the reaction mixture prior to the initiation of the reaction. For the formaldehyde incubation experiment, purified AOX was diluted to a final concentration of 2 mg/mL in a total volume of 1 mL. Formaldehyde was then added to a final concentration of 50 mM, and the mixture was incubated in a water bath at 30 °C. Aliquots were taken at 10-min intervals to determine the residual activity. The activity measured immediately after the addition of formaldehyde (t = 0) was defined as the initial activity. Thermostability measurement The thermostability of AOXM2 and AOXM3 was assessed using a Chirascan circular dichroism (CD) spectrometer (Applied Photophysics, UK). Protein samples were dissolved at a concentration of 0.2 mg/mL and loaded into a 0.1 cm path-length quartz cuvette. The samples were heated from 30 to 80 °C at a rate of 2 °C/min, and CD spectra were recorded over the wavelength range of 200–280 nm with a 1 nm interval. The resulting temperature-dependent CD data were processed using Global 3 analysis software, and the melting temperature (Tm) was determined. Substrate specificity assay Compounds 1–3a and 10–12a were prepared as 200 mM stock solutions in HEPES buffer. Compounds 4–9a were prepared as 200 mM stock solutions in DMSO. The relative activities of GtAOX mutants toward 1–12a were measured in 1 mL reactions containing 50 μL of 1–12a (final concentration 10 mM), purified AOX (0.002 mg/mL for 1–2a or 0.2 mg/mL for 3–12a), 250 μL of the HRP-coupled reagent, and 100 mM HEPES buffer (containing 100 mM NaCl, pH 7.5). Reactions were conducted at 30 °C for 3 min. For each substrate, a control reaction without enzyme was performed to account for background. H2O2 production was measured and analyzed as described in Section 2.7. High-performance liquid chromatography (HPLC) analysis All enzymatic reactions were terminated by the addition of an equal volume of 500 mM H2SO4, followed by centrifugation, and the supernatants were filtered through a 0.22 μm PES membrane prior to HPLC analysis. Analytical yields were calculated based on pre-established standard curves. Dihydroxyacetone (DHA) was quantified using a Shimadzu LC-2030 HPLC system equipped with a UV detector and an Aminex HPX-87H column (300 mm × 7.8 mm, 9 μm). The mobile phase consisted of 10 mM H2SO4 and was delivered at a flow rate of 0.5 mL/min. The column temperature was maintained at 40 °C. DHA was detected at a wavelength of 210 nm, and the injection volume was 10 μL. Ethylene glycol was analyzed using a Shimadzu LC-20A HPLC system equipped with a refractive index detector (RID) and an Aminex HPX-87H column (300 mm × 7.8 mm, 9 μm). The mobile phase was 10 mM H2SO4 with a flow rate of 0.6 mL/min. The column temperature was set to 60 °C, and the injection volume was 20 μL. Molecular dynamics simulation The structure of GtAOXM3 complexed with FAD and methanol was predicted by AlphaFold3 [31] and docked using AutoDock Vina [32]. MD simulations were performed using GROMACS 2023 package [33]. The protein topology was processed using the AMBER99SB-ILDN force field and the TIP3P water model. For the ligands, generalized AMBER force field (GAFF) parameters were generated using Antechamber in AmberTools. Partial charges were assigned via the AM1-BCC method, and topology conversion to GROMACS format was performed using the ACPYPE tool. The complex was solvated in a cubic box with a minimum solute-box distance of 1.0 nm and neutralized with 0.15 M NaCl. Energy minimization was carried out using the steepest descent algorithm. The system was then equilibrated in the NVT ensemble at 300 K using the V-rescale thermostat and in the NPT ensemble at 1 bar using the Parrinello-Rahman barostat, with position restraints applied to the protein backbone and ligand heavy atoms. Following equilibration, a 100 ns production run was performed without restraints. A time step of 2 fs was used, with bond lengths constrained by the LINCS algorithm. Long-range electrostatic interactions were treated using the Particle Mesh Ewald (PME) method. Simulation trajectories were processed using GROMACS tools. The structural stability and local flexibility were evaluated by calculating the Root Mean Square Deviation (RMSD) and Root Mean Square Fluctuation (RMSF), respectively. Dynamic Cross-Correlation Matrix (DCCM) analysis was conducted to characterize correlated motions. Structural visualizations were rendered using PyMOL or UCSF ChimeraX. Results and Discussion Directed evolution of GtAOXM2 GtAOXM2 was chosen as the starting template for directed evolution based on its relatively favorable catalytic activity among alcohol oxidases in our in-house library (Fig. 2A) [11,20]. An error-prone PCR (epPCR) library containing 7200 transformants was generated [34]. To selectively enrich formaldehyde-tolerant mutants, 100 mM formaldehyde was included in the activity assay. Screening the library identified nine variants with single or multiple amino acid substitutions with improved catalytic activity (Fig. 2B). Among these, the E526K/K650R/N43S triple mutant exhibited the highest activity, reaching 6.5 U/mg. To clarify the contributions of individual mutation within the multi-point variants, mutational deconvolution was performed (Fig. 2C). This analysis revealed that the V533A and E526K substitutions conferred only marginal activity improvements, whereas the single D500G mutation doubled catalytic activity to 9.1 U/mg. The remaining mutations did not confer measurable benefits to enzyme performance. Based on these results, single-site saturation mutagenesis was conducted at positions 500, 526, and 533 to identify the optimal residues at each site (Fig. 2D). At position 500, several substitutions (e.g., G, M, and W) outperformed the parental variant, with D500G consistently exhibiting the highest activity. For positions 526 and 533, the optimal substitutions were identified as I and L, yielding relative activities of 110% and 105%, respectively. Subsequent combinatorial assembly of these beneficial mutations generated two optimized variants: D500G/E526I and D500G/E526I/V533L (designated GtAOXM3). These variants exhibited specific activities of 11.4 and 11.3 U/mg, corresponding to an approximately 2.6-fold improvement over the parental enzyme (Fig. 2E).Fig. 2Directed evolution of GtAOXM2. (A) Overview of the directed evolution trajectory. (B) Mutants with improved catalytic activity identified via error-prone PCR. (C) Deconvolution of beneficial mutations. (D) Single-site saturation mutagenesis at positions 500, 526, and 533. (E) Combinatorial mutations at positions 500, 526, and 533. Directed evolution of GtAOXM2. (A) Overview of the directed evolution trajectory. (B) Mutants with improved catalytic activity identified via error-prone PCR. (C) Deconvolution of beneficial mutations. (D) Single-site saturation mutagenesis at positions 500, 526, and 533. (E) Combinatorial mutations at positions 500, 526, and 533. Enzymatic characterization To ensure accurate measurements of the concentration of active, FAD-bound enzyme, the functional fraction of M2 and M3 was determined, yielding FAD occupancies of 80% and 81%, respectively (Fig. S2–S3). Using these adjusted concentrations, the kinetic parameters of the GtAOX variants were systematically evaluated (Table 2 and Fig. S7). GtAOXM2 exhibited a catalytic constant (kcat) of 10.3 s−1 and a Michaelis constant (KM) of 7.5 mM, corresponding to a catalytic efficiency (kcat/KM) of 1.4 s−1 mM−1. Introduction of the single-point mutation D500G resulted in a twofold increase in kcat to 21.0 s−1 accompanied by a reduction in KM to 4.4 mM, suggesting enhanced apparent substrate affinity and catalytic turnover. Subsequent incorporation of the E526I mutation further decreased the KM to 2.7 mM while maintaining high catalytic activity. The addition of V533L mutation slightly increased kcat to 21.6 s−1, yielding a catalytic efficiency of 7.7 s−1 mM−1, representing a sixfold improvement over GtAOXM2.Table 2Determination of kinetic parameters of key GtAOX variants.Table 2EnzymeSpecific activity (U/mg)kcat (s−1)KM (mM)kcat/KM (s−1 mM−1)FoldGtAOXM25.5 ± 0.110.3 ± 0.27.5 ± 0.91.4–GtAOXM2-D500G11.4 ± 0.221.0 ± 0.24.4 ± 0.34.83.4GtAOXM2-D500G/E526I14.2 ± 0.620.1 ± 0.62.7 ± 0.67.45.3GtAOXM314.1 ± 0.421.6 ± 0.62.8 ± 0.57.75.5 Table 2Determination of kinetic parameters of key GtAOX variants.Table 2EnzymeSpecific activity (U/mg)kcat (s−1)KM (mM)kcat/KM (s−1 mM−1)FoldGtAOXM25.5 ± 0.110.3 ± 0.27.5 ± 0.91.4–GtAOXM2-D500G11.4 ± 0.221.0 ± 0.24.4 ± 0.34.83.4GtAOXM2-D500G/E526I14.2 ± 0.620.1 ± 0.62.7 ± 0.67.45.3GtAOXM314.1 ± 0.421.6 ± 0.62.8 ± 0.57.75.5 Determination of kinetic parameters of key GtAOX variants. Time-course experiments were performed to compare formaldehyde production (Fig. 3A). In these assays, hydrogen peroxide was efficiently removed through a coupled cascade with catalase from Bacillus subtilis (BsCAT). Under conditions of 100 mM methanol and 0.2 mg/mL AOX, GtAOXM2 produced 19.9 mM formaldehyde within 1 h, whereas GtAOXM3 reached 28.7 mM, demonstrating enhanced formaldehyde synthesis. Increasing the methanol concentration to 200 mM resulted in only a modest 4 mM increase in formaldehyde, suggesting that the reaction approached saturation under these conditions, likely due to the low KM of GtAOXM3 in combination with other rate-limiting factors such as product accumulation (Fig. 3B). In contrast, increasing the GtAOXM3 concentration to 1.0 mg/mL allowed the production of up to 38.2 mM formaldehyde within 1 h (Fig. 3C).Fig. 3Enzymatic characterization of GtAOX variants. (A) Methanol oxidation to formaldehyde catalyzed by GtAOXM2 and GtAOXM3. Reaction condition: 100 mM Methanol, 0.2 mg/mL GtAOXM2 or GtAOXM3, 200 U/mL CAT, 30 °C, 1 h. (B) Effect of methanol concentration on formaldehyde production. Reaction condition: 50–200 mM Methanol, 0.2 mg/mL GtAOXM3, 200 U/mL CAT, 30 °C, 1 h. (C) Effect of GtAOXM3 concentration on formaldehyde production. Reaction condition: 100 mM Methanol, 0.1–1 mg/mL GtAOXM3, 200 U/mL CAT, 30 °C, 1 h. (D) Residual activities of GtAOXM0, GtAOXM2, GtAOXM3 and PcAOX at 0–50 mM formaldehyde. (E) Residual activity of AOX during incubation with 50 mM formaldehyde. t = 0 min represents the initial activity measured after the addition of formaldehyde. (F) Tm value determined by circular dichroism (CD) spectroscopy. Enzymatic characterization of GtAOX variants. (A) Methanol oxidation to formaldehyde catalyzed by GtAOXM2 and GtAOXM3. Reaction condition: 100 mM Methanol, 0.2 mg/mL GtAOXM2 or GtAOXM3, 200 U/mL CAT, 30 °C, 1 h. (B) Effect of methanol concentration on formaldehyde production. Reaction condition: 50–200 mM Methanol, 0.2 mg/mL GtAOXM3, 200 U/mL CAT, 30 °C, 1 h. (C) Effect of GtAOXM3 concentration on formaldehyde production. Reaction condition: 100 mM Methanol, 0.1–1 mg/mL GtAOXM3, 200 U/mL CAT, 30 °C, 1 h. (D) Residual activities of GtAOXM0, GtAOXM2, GtAOXM3 and PcAOX at 0–50 mM formaldehyde. (E) Residual activity of AOX during incubation with 50 mM formaldehyde. t = 0 min represents the initial activity measured after the addition of formaldehyde. (F) Tm value determined by circular dichroism (CD) spectroscopy. Given the inhibitory effect of formaldehyde on AOX activity, formaldehyde tolerance assays were conducted (Fig. 3D). In the presence of 50 mM formaldehyde, GtAOXM2 retained 57.2% of its activity, whereas GtAOXM3 and PcAOX maintained 70.8% and 66.6%, respectively. After incubation at 30 °C with 50 mM formaldehyde for 1 h, the residual activities of GtAOXM2 and GtAOXM3 were 29.9% and 41.7%, respectively, validating the effectiveness of the directed evolution strategy targeting formaldehyde tolerance (Fig. 3E). In addition, the residual activity of PcAOX was 34.2% under the same conditions, and no significant difference in formaldehyde tolerance was observed between PcAOX and GtAOXM3, as they showed comparable performance at certain time points. Thermostability assessments indicated that the melting temperatures (Tm) of GtAOXM0, GtAOXM2, and GtAOXM3 were 51 °C, 52 °C, and 55 °C, respectively, while PcAOX exhibited a Tm of 52 °C, suggesting a modest increase in structural stability for the evolved variants. (Fig. 3F and S8). In addition, previous studies have reported that AOXs can overoxidize formaldehyde to formic acid [19]. Comparative measurements indicated that GtAOXM3 exhibited only 1.6% overoxidation activity, underscoring its high chemoselectivity toward formaldehyde formation (Fig. S6). Structural and dynamic basis for enhanced stability To gain structural insights into the enhanced performance of GtAOXM3, a homodimeric structural model complexed with FAD and methanol was constructed (Fig. 4A). Structural mapping revealed that the substituted residues G500, I526, and L533 are located on two distal loops regions, spatially separated from the substrate-binding pocket. Notably, these residues cluster within an extended hairpin loop (residues 507–540, Loop 2) that protruding from each monomer and mediates stabilizing inter-subunit interactions within the dimeric assembly [35,36].Fig. 4Structural and dynamic basis for enhanced stability of GtAOXM3. (A) Structural of GtAOXM3. FAD and methanol are shown in stick. The three introduced mutation sites are highlighted as purple spheres. The active-site pocket is outlined in green, and the hairpin loop (residues 507–540) mediating dimerization is shown in dark blue. Loop 1 (residues 481–500, chain B) and loop 2 (residues 507–540, chain B) are depicted in yellow and orange, respectively. (B) RMSD profiles of GtAOX variants. (C–D) DCCM analysis of residue motions. Structural and dynamic basis for enhanced stability of GtAOXM3. (A) Structural of GtAOXM3. FAD and methanol are shown in stick. The three introduced mutation sites are highlighted as purple spheres. The active-site pocket is outlined in green, and the hairpin loop (residues 507–540) mediating dimerization is shown in dark blue. Loop 1 (residues 481–500, chain B) and loop 2 (residues 507–540, chain B) are depicted in yellow and orange, respectively. (B) RMSD profiles of GtAOX variants. (C–D) DCCM analysis of residue motions. Comparative molecular dynamics simulations were performed to assess the global structural stability of the variants. Root-mean-square deviation (RMSD) analysis showed that GtAOXM3 maintained lower deviations compared to GtAOXM2 (Fig. 4B). The RMSD of GtAOXM2 fluctuated mainly between 2.5 Å and 3.5 Å, whereas the GtAOXM3 stabilized around 2.5 Å. This decrease suggests that the introduced mutations collectively enhance the global structural rigidity of the protein, enabling it to better retain its folded conformation in solution, consistent with the experimentally observed improvement in thermostability. To further investigate local dynamics, root-mean-square fluctuation (RMSF) analysis was performed (Fig. S11). Most regions of both variants displayed low fluctuations (RMSF <2.0 Å), consistent with the RMSD results. However, a pronounced increase in RMSF was observed in the hairpin loop region (residues 525–536) of GtAOXM3, with peak values approaching 6.0 Å, compared to 2.0–3.0 Å in GtAOXM2. Although increased local flexibility is often associated with reduced stability, previous studies have demonstrated that enhanced loop dynamics can contribute positively to protein stability by increasing the configurational entropy of the folded state, even at the expense of enthalpic interactions [37]. In this context, the elevated flexibility of the hairpin loop in GtAOXM3 likely reflects a redistribution of conformational freedom, enabled by strengthened core and interfacial interactions that reduce the dependence on loop rigidity for maintaining dimer stability. Dynamic cross-correlation matrix (DCCM) analysis provided further insight into the nature of these loop motions (Fig. 4C–D). Despite the elevated RMSF values within this region, an overall enhancement in positive correlated motions was observed across the entire protein structure (average ΔCorrelation > 0). This indicates that the increased mobility is not a result of random fluctuations but rather reflects highly coordinated, modular motions. Such correlated dynamics are characteristic of functional loop regions and may facilitate conformational adaptability without compromising overall structural stability. Given the distal location of this loop from the active site, these coordinated motions are consistent with a potential long-range structural effect, whereby modulation of loop dynamics indirectly contributes to improved catalytic performance and stability. Substrate specificity of GtAOX variants The substrate specificity of the GtAOX variants was evaluated using a panel of primary alcohols and diols (2–12a, Table 3). As the alkyl chain length increased, a pronounced decline in catalytic activity was observed, with activity decreasing by approximately 1–4 order of magnitude from methanol to higher alcohols. For substrates 4–9a, enzymatic activity was nearly undetectable (<2 mU/mg), indicating that elongation of the carbon chain is poorly tolerated. This narrow substrate scope is consistent with structural features of the GtAOX active site. The substrate-binding pocket is defined by several conserved bulky residues, including F101, F421, and W559, which collectively form a confined cavity that disfavors accommodation of larger alcohols [35,36]. Diol substrates 10–12a also exhibited negligible catalytic activity (<0.1 mU/mg). Notably, ethylene glycol (10a) and dihydroxyacetone (12a), the final products of the multienzyme cascade, were not further oxidized by GtAOX.Table 3Substrate specificity of GtAOX variants.Table 3SubstrateSpecific activity (mU/mg)GtAOXM2GtAOXM31a5537141382a267973a56134a1.90.95a1.31.66an.d.a.n.d.a.7an.d.a.n.d.a.8a0.4n.d.a.9an.d.a.n.d.a.10an.d.a.n.d.a.11an.d.a.n.d.a.12a<0.1<0.1n.d.a.: no detectable activity. Table 3Substrate specificity of GtAOX variants.Table 3SubstrateSpecific activity (mU/mg)GtAOXM2GtAOXM31a5537141382a267973a56134a1.90.95a1.31.66an.d.a.n.d.a.7an.d.a.n.d.a.8a0.4n.d.a.9an.d.a.n.d.a.10an.d.a.n.d.a.11an.d.a.n.d.a.12a<0.1<0.1n.d.a.: no detectable activity. Substrate specificity of GtAOX variants. n.d.a.: no detectable activity. These results demonstrate that GtAOX exhibits a strong preference for methanol. Such intrinsic substrate selectivity is advantageous for its application in multienzyme cascade systems, as it minimizes undesired side reactions and ensures selective C1 bioconversion. Application of GtAOXM3 in multienzyme cascade synthesis GtAOXM3 was incorporated into multienzyme cascade systems for the synthesis of biochemically relevant products, dihydroxyacetone (DHA) and ethylene glycol (EG, Fig. 5 and Fig. S10). In Route A, methanol was converted to DHA by coupling GtAOXM3 with an engineered benzaldehyde lyase from Herbiconiux sp. (HeBALM11) [38]. Under an initial methanol concentration of 150 mM, 34.5 mM DHA was produced within 24 h, corresponding to an analytical yield of 69%. In Route B, methanol-to-EG conversion was achieved using a three-enzyme system comprising GtAOXM3, a benzoylformate decarboxylase from Polynucleobacter necessarius (PnBFDM1) [39], and an NADH-dependent alcohol dehydrogenase from Gluconobacter oxydans 621H (GOX) [40], with NADH regeneration mediated by glucose dehydrogenase from Bacillus subtilis (BmGDH) [41]. Under the same initial methanol concentration (150 mM), 23.3 mM EG was obtained after 24 h, corresponding to an analytical yield of 31%.Fig. 5Multienzyme cascade conversion of methanol to dihydroxyacetone (DHA, Route A) and ethylene glycol (EG, Route B). Reaction condition of route A (1 mL): 150 mM methanol, 1 mg/mL GtAOXM3, 7 mg/mL HeBALM11, 200 U/mL BsCAT, 0.5 mM TPP, 5 mM MgCl2, HEPES buffer (100 mM, pH 7.5), 30 °C, 24 h. Reaction condition of route B (1 mL): 150 mM methanol, 1 mg/mLGtAOXM3, 2.5 mg/mL PnBFD, 0.4 mg/mL GOX, 200 U/mL CAT, 4 U/mL GDH, 0.5 mM TPP, 1 mM NAD+, 100 mM Glucose, 5 mM FeCl2, 20 mM ascorbic acid, HEPES buffer (100 mM, pH 7.5), 30 °C, 24 h. AOX, Alcohol oxidase; CAT, Catalase; BFD, Benzoylformate decarboxylase; BAL, Benzaldehyde lyase; GOX, Alcohol dehydrogenase; GDH, Glucose dehydrogenase; GA, Glycolaldehyde; TPP, Thiamine pyrophosphate. Please note that insufficient CAT activity may lead to partial oxidation of formaldehyde to formic acid by hydrogen peroxide. Multienzyme cascade conversion of methanol to dihydroxyacetone (DHA, Route A) and ethylene glycol (EG, Route B). Reaction condition of route A (1 mL): 150 mM methanol, 1 mg/mL GtAOXM3, 7 mg/mL HeBALM11, 200 U/mL BsCAT, 0.5 mM TPP, 5 mM MgCl2, HEPES buffer (100 mM, pH 7.5), 30 °C, 24 h. Reaction condition of route B (1 mL): 150 mM methanol, 1 mg/mLGtAOXM3, 2.5 mg/mL PnBFD, 0.4 mg/mL GOX, 200 U/mL CAT, 4 U/mL GDH, 0.5 mM TPP, 1 mM NAD+, 100 mM Glucose, 5 mM FeCl2, 20 mM ascorbic acid, HEPES buffer (100 mM, pH 7.5), 30 °C, 24 h. AOX, Alcohol oxidase; CAT, Catalase; BFD, Benzoylformate decarboxylase; BAL, Benzaldehyde lyase; GOX, Alcohol dehydrogenase; GDH, Glucose dehydrogenase; GA, Glycolaldehyde; TPP, Thiamine pyrophosphate. Please note that insufficient CAT activity may lead to partial oxidation of formaldehyde to formic acid by hydrogen peroxide. Conclusion In summary, the alcohol oxidase variant GtAOXM3 was obtained through adaptive directed evolution, exhibiting improved catalytic efficiency, formaldehyde tolerance, and thermostability. GtAOXM3 displayed a specific activity of 14.1 U/mg and a catalytic efficiency of 7.7 s−1 mM−1, comparable to the naturally efficient enzyme PcAOX and commercially available PpAOX. Although the catalytic efficiency and formaldehyde tolerance of GtAOXM3 show no significant difference compared to PcAOX, these results provide a framework for selecting and evolving AOX variants under diverse conditions, offering additional options and strategies for constructing future C1 compound synthesis pathways. Integration of GtAOXM3 into multienzyme cascades facilitated the biosynthesis of dihydroxyacetone (DHA) and ethylene glycol (EG) from methanol, achieving analytical yields of 69% and 31%, respectively. Notably, the formaldehyde condensation step remains rate-limiting, suggesting that further optimization through enzyme screening and directed evolution could improve overall pathway efficiency [42]. Collectively, this work provides a promising biocatalytic tool for methanol-based bioconversion and establishes a versatile platform for the development of efficient C1-based synthetic pathways. CRediT authorship contribution statement Xu-Wei Ding: Writing – original draft, Validation, Methodology, Investigation. Ya-Juan Sun: Writing – original draft, Methodology, Investigation. Xiao-Yu Hu: Software, Investigation. Jun-Yi Yang: Software, Investigation. Jian-He Xu: Writing – review & editing, Supervision, Project administration. Gao-Wei Zheng: Writing – review & editing, Supervision, Funding acquisition. Chun-Xiu Li: Writing – review & editing, Supervision, Project administration, Funding acquisition. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.