Unveiling a catalytically promiscuous feruloyl esterase from Clostridium acetobutylicum

Abstract

Abstract

Clostridium species have garnered increasing attention for their ability to convert lignocellulosic biomass into renewable fuels and platform chemicals. Among the enzymes involved in lignin degradation, feruloyl esterases (FAEs) cleave ester bonds between ferulic acid and polysaccharide side chains, thereby facilitating the disruption of lignin-carbohydrate complexes. However, the biochemical and structural properties of Clostridium FAEs remain poorly characterized, with activity studies largely limited to model substrates rather than native lignin-derived compounds. Here, we report the functional and crystallographic characterization of a novel FAE (CaFaeA) from Clostridium acetobutylicum. CaFaeA exhibits broad catalytic activity toward a range of hydroxycinnamate esters as well as bis(2-hydroxyethyl) terephthalate (BHET), distinguishing it from typical carboxyl esterases. Furthermore, the 2.45 Å crystal structure of CaFaeA reveals a canonical α/β-hydrolase fold with a unique lid domain of three α-helices and two antiparallel β-strands partially covering the active site. Mutagenesis identified two gatekeeper residues, Y151 and E168, that regulate substrate access and catalytic performance. Remarkably, CaFaeA demonstrates exceptional tolerance to organic solvents, retaining or even enhancing activity in the presence of 25% dimethyl sulfoxide and n-hexane. With insoluble wheat arabinoxylan (I-WAX) as substrate, its unique lid architecture enabled efficient cleavage of ferulic acid–arabinose ester linkages, resulting in a release of free ferulic acid by 5.39 mg·μmol⁻¹·h⁻¹, representing a high activity within the range reported for FAEs. These findings not only provide mechanistic insights into microbial FAE function but also highlight CaFaeA as a promising candidate for lignocellulosic biomass utilization.

Key points

• Characterization of CaFaeA from Clostridium acetobutylicum

• Substrate promiscuity, gating residues, and solvent tolerance

• High-efficiency hydrolysis of I-WAX for ferulic acid release

Supplementary Information

The online version contains supplementary material available at 10.1007/s00253-026-13756-7.

Introduction

Clostridium acetobutylicum is a well-characterized model organism in industrial biotechnology, primarily known for its capacity to produce solvents such as butanol, acetone, and ethanol through anaerobic fermentation (Oehlenschläger et al. 2025; Suresh et al. 2024; Wu et al. 2025). With increasing emphasis on sustainable biomanufacturing, C. acetobutylicum has emerged as a promising microbial chassis for biorefinery applications, including waste valorization, and synthetic biology-driven metabolic engineering (Du et al. 2022; Liu et al. 2020; Survase et al. 2020). Notably, its broad metabolic versatility makes it particularly attractive for the efficient conversion of lignin-rich biomass (Luo et al. 2020; Mondal et al. 2022).

Among the structural components of lignocellulose, lignin represents a significant bottleneck due to its heterogeneous, aromatic polymeric nature and resistance to microbial degradation (Farmanbordar et al. 2024; Wu et al. 2023). Efficient depolymerization of lignin and disruption of lignin–carbohydrate linkages are pivotal to unlocking the full potential of biomass utilization. In addition to xylanases, feruloyl esterases (FAEs) play an essential role in this process by hydrolyzing ester bonds between ferulic acid and arabinose-containing polysaccharide within plant cell wall polymers, thereby facilitating the breakdown of the lignin-carbohydrate matrix (Usman et al. 2025).

FAEs are classified within the α/β-hydrolase superfamily, a structurally conserved yet functionally versatile group of enzymes that includes lipases, esterases, proteases, and transferases. Members of this superfamily typically share a canonical α/β-fold and a catalytic triad consisting of serine, histidine, and an acidic residue (Asp/Glu). Despite these structural similarities, α/β-hydrolases display substantial variation in substrate recognition, reaction specificity, and turnover efficiency. As such, deciphering the structure–function relationships that govern these variations is essential for guiding protein engineering and the rational design of biocatalysts with tailored functionalities. Although several microbial FAEs have been structurally and biochemically characterized, considerable differences remain in their lid regions and substrate recognition mechanisms (Phienluphon et al. 2024; Yamada et al. 2025). In particular, α/β-hydrolases from industrially relevant microbes such as C. acetobutylicum remain largely unexplored. Homologous enzymes from phylogenetically related organisms often exhibit distinct catalytic profiles due to differences in substrate binding architecture or active-site dynamics (Shi et al. 2025; Yang et al. 2025). Gaining structural and mechanistic insight into FAEs from C. acetobutylicum is not only essential to disrupt lignin-cellulose-hemicellulose network but also for enabling rational design of enzymes tailored for biomass conversion.

In this study, we report the structural and functional characterization of an α/β-hydrolase from C. acetobutylicum. Crystallographic analysis, supported by sequence and structural clustering, identifies the enzyme as a member of the feruloyl esterase family. Enzymatic assays reveal its catalytic activity towards ferulic acid and lignin-derived substrates, suggesting a degree of substrate promiscuity indicative of functional divergence within the FAE family. Our findings expand the current understanding of FAE diversity and provide a structural framework for exploring the catalytic potential of the biocatalysts.

Materials and methods

Strains and materials

The gene encoding CaFaeA (GenBank accession: KHD35464) from Clostridium acetobutylicum was cloned into the pET28a vector using the NdeI and XhoI restriction sites, resulting in the recombinant plasmid pET28a_CaFaeA. PrimeSTAR Max DNA Polymerase was obtained from Takara Bio (Beijing, China). Kanamycin was purchased from Sangon Biotech Co., Ltd. (Shanghai, China). Methyl ferulate (MFA) and other chemicals were obtained from Sigma-Aldrich Trading Co., Ltd. (Shanghai, China).

Gene cloning, mutation, and expression

The recombinant plasmid pET28a_CaFaeA was transformed into E. coli BL21(DE3) and grown in Luria–Bertani (LB) medium supplemented with 50 µg·mL⁻¹ kanamycin at 37 °C with shaking at 200 rpm until the optical density at 600 nm reached 0.6. Protein expression was induced by the addition of 0.1 mM isopropyl-β-D-thiogalactopyranoside, followed by incubation for an additional 14 h at 16 °C and 200 rpm. Site-directed mutagenesis of pET28a_CaFaeA was performed using PrimeSTAR Max DNA polymerase with synthetic oligonucleotide primers listed in Table S1, and the results are shown in Fig. S1A. The PCR products were treated with DpnI to degrade the methylated parents and subsequently transformed into E. coli DH5α cells. Successful mutations were confirmed by DNA sequencing.

Purification and analysis of recombinant protein

Harvested E. coli cells carrying recombinant CaFaeA or its variants were resuspended in lysis buffer (50 mM Tris-HCl pH 8.0, 5% glycerol [v/v], and 1 mM β-mercaptoethanol) and disrupted by ultrasonication on ice. The lysate was centrifuged at 12,000 × g for 30 min at 4 °C, and the resulting supernatant was subjected to Ni–NTA affinity chromatography followed by gel filtration (Fig. S1B). The molecular mass of the native protein was determined by gel filtration chromatography using a HiLoadTM 16/600 SuperdexTM 200 column (Cytiva, Uppsala, Sweden). Protein concentrations were determined using Bradford assay with BSA as standard. Enzyme samples were stored at −80 °C until use in subsequent assays.

Enzyme assays

The specific activities of purified CaFaeA and its mutants were determined using various substrates. Enzyme samples (0.1 mg) were incubated separately with 1 mM hydroxycinnamic acid esters, including methyl ferulate (MFA), methyl sinapate (MSA), methyl caffeate (MCA), and methyl p-coumarate (MpCA), in 100 mM phosphate buffer (pH 8.0) at 30 °C for 10 min. The same amount of enzyme was incubated with 1 mM bis(2-hydroxyethyl) terephthalate (BHET) or p-nitrophenyl palmitate (p-NPP) in 100 mM phosphate buffer (pH 8.0) at 30 °C for 12 h. Following incubation, reaction mixtures were centrifuged, and the resulting supernatants were mixed with an equal volume of methanol and filtered through a 0.22 μm organic membrane filter. One unit of enzyme activity (U) was defined as the amount of product released by 1 mg of purified enzyme per minute under the assay conditions.

Quantification was performed using an Agilent 1260 HPLC system equipped with a variable wavelength detector (VWD). Hydrolysis products were quantified by HPLC using external standards. For hydroxycinnamic acid esters, separation was achieved on a Unitary C18 column (4.6 × 150 mm, 5 μm). The mobile phase consisted of 0.1% (v/v) formic acid in water and acetonitrile. Samples (5 μL) were eluted isocratically with 60% acetonitrile at a flow rate of 0.8 mL min⁻¹, and detection was monitored at 322 nm. Activity was determined by quantifying the formation of the corresponding free acids. For BHET and mono(2-hydroxyethyl) terephthalate (MHET) analysis, a ZORBAX SB-C18 column (4.6 × 150 mm, 5 μm) was employed. The mobile phase comprised 0.1% (v/v) trifluoroacetic acid in water and acetonitrile. Samples (5 μL) were eluted isocratically with 80% acetonitrile at a flow rate of 0.8 mL min⁻¹, with detection at 240 nm. Activity toward BHET was determined by quantifying the formation of the hydrolysis product, MHET. The retention times and chromatographic profiles of relevant substrates and product standards, along with the standard calibration curves, are provided in the supplementary information (Fig. S2). The specific activity toward p-NPP was determined using a spectrophotometric method as described (Rodríguez-Alonso et al. 2023).

Enzyme similarity network analysis

A total of 2191 FAEs were retrieved from the BRENDA database and previous reports. Sequences shorter than 120 amino acids or longer than 900 amino acids were excluded. Redundant sequences were removed by performing 100% sequence identity clustering using CD-HIT. The selected sequences were used to construct a sequence similarity network (SSN) for FAEs using the Enzyme Function Initiative-Enzyme Similarity Tool (EFI-EST) (Oberg et al. 2023). The obtained network was visualized and annotated using Cytoscape 3.9.1 software, and subclusters containing known FAEs were finally isolated. The default E-value was set to 5, and the filter value was set to 30.

Effects of pH and temperature on enzyme activity and kinetic parameters

The effect of pH on enzyme activity was determined using the following buffers (100 mM): citrate buffer (pH 3.0–5.0), NaH₂PO₄-Na₂HPO₄ buffer (pH 6.0–8.0), and glycine–NaOH buffer (pH 9.0–10.0). The effect of temperature was assessed over a range of 20 to 80 °C. pH stability was evaluated by incubating the enzyme in the above buffers at 4 °C for 30 min, followed by the measurement of residual activity under the optimal conditions. Thermal stability was assessed by pre-incubating the enzyme at 20–80 °C for 30 min, followed by the determination of residual activity under the optimal assay conditions. Purified CaFaeA (0.1 mg) was incubated with 1 mM methyl ferulate for 10 min under the specified conditions. Enzyme activity was determined by quantifying the released ferulic acid. The maximum activity was defined as 100%. Kinetic parameters were determined under the optimal temperature and pH conditions for each enzyme by varying the concentration of methyl ferulate from 0.2 to 10 mM (Table S2). High-performance liquid chromatography was used to confirm the enzymatic product profile. The kinetic parameters (k_cat, K_m) were calculated by fitting the experimental data to the Michaelis–Menten equation using GraphPad Prism 8.3 program.

Effect of organic solvent on the enzyme activity

The effect of organic solvents on enzyme activity was evaluated by incubating purified CaFaeA (0.1 mg) with 25% (v/v) of various solvents, including methanol, ethanol, n-butanol, dimethyl sulfoxide (DMSO), n-hexane, ethyl acetate, acetonitrile, isopropanol, and acetone at 4 °C for 60 min. Residual enzyme activity was then measured by incubating the enzyme with MFA at 30 °C for 10 min. The amount of ferulic acid was calculated using the HPLC method described previously. The activity of CaFaeA incubated in 100 mM PBS buffer (pH 8.0) is defined as 100%.

Ferulic acid production from insoluble wheat arabinoxylan

Twenty milligrams of insoluble wheat arabinoxylan (I-WAX) was incubated with 100 μg CaFaeA or its mutants in 1 mL 100 mM Na₂HPO₄-NaH₂PO₄ buffer (pH 8.0) at 30 °C, 300 rpm for 12 h. Ferulic acid was quantified using an Agilent 1290 Infinity II liquid chromatography system coupled to an Agilent 6545B quadrupole time-of-flight mass spectrometer. Separation was performed on a Unitary C18 column (4.6 × 150 mm, 5 μm) at 25 °C. The mobile phase consisted of 0.1% formic acid in water and acetonitrile, delivered isocratically (40:60, v/v) at a flow rate of 0.8 mL min⁻¹. Mass spectrometric detection utilized a Dual Agilent Jet Stream Electrospray Ionization source in negative centroid mode, scanning from m/z 100 to 500. Quantification of ferulic acid was based on the extracted ion chromatogram (EIC) of the deprotonated molecular ion (m/z 193.052). Quantification of the product was performed via an external standard curve (Fig. S3). The specific activity was calculated as the amount of ferulic acid released per hour per micromole of enzyme using 100 mg of I-WAX as substrate.

Crystallization, data collection, and structure determination

Purified CaFaeA at a concentration of 14 mg·mL⁻¹ was crystallized by the hanging-drop vapor diffusion method at 20 °C. Crystals formed within 2 weeks in a solution containing 0.1 M Bis–Tris (pH 6.5), 25% (w/v) PEG 3350, and 0.2 M MgCl₂·6H₂O. Prior to data collection, crystals were cryoprotected by soaking in mother liquor supplemented with 25% (w/v) glycerol and rapidly flash-cooled in liquid nitrogen at 100 K. X-ray diffraction data were collected at the BL18U1 beamline of the Shanghai Synchrotron Radiation Facility (SSRF). Data processing was performed using the HKL3000 software (Minor et al. 2006). The structure was solved by molecular replacement with Phaser MR using the structure of PDB entry 7Z2U as the search model. Model building and iterative refinement were conducted using Phenix and Coot (Emsley et al. 2010). Final refinement statistics and structural details are provided in Table S3. The atomic coordinates and structure factors have been deposited in the Protein Data Bank under accession code 9W30.

Molecular dynamics simulations and computational procedures

All molecular dynamics simulations were conducted using GROMACS (v2022.02) with the Amber14SB force field. Prior to simulation, water molecules were removed from the crystal structures. Enzymes were solvated in a cubic box of TIP3P water with a minimum distance of 10 Å between the protein and the box edges. The systems were neutralized by adding Na ions. Energy minimization was carried out using the steepest descent algorithm. Both NVT and NPT equilibration were performed before production simulations were conducted for 100 ns at temperatures of 300 K.

To further support the detailed structural analysis, the homologous enzyme sequences were retrieved from the NCBI and PDB databases, aligned using Clustal X 2.0, and visualized with the ESPript server (Larkin et al. 2007; Robert and Gouet 2014). For enzymes lacking experimental structures, homology models were generated with the SWISS-MODEL server (Waterhouse et al. 2018). Structural visualization was performed using PyMOL, and active-site cavity volumes were calculated with CavityPlus (Wang et al. 2023; Xu et al. 2018).

Statistical analysis

Data were analyzed using Origin software, all the enzymatic activities of CaFaeA and the mutant assessments were performed in triplicate, and experimental data were expressed as mean ± standard deviation. For the relative activity, the highest activity observed was defined as 100%, and all other values were expressed relative to this.

Results

Identification and functional characterization of a feruloyl esterase from Clostridium acetobutylicum

While most functionally characterized FAEs to date have originated from fungi and lactic acid bacteria, the catalytic potential of FAEs derived from Clostridium species remains largely unexplored. Through genome mining of Clostridium acetobutylicum, we identified a putative α/β-hydrolase-encoding gene (GenBank accession: KHD35464) comprising 259 amino acids and a predicted molecular mass of 29.2 kDa. The functional role of this protein had not been previously reported. To investigate its catalytic potential, the gene was cloned and heterologously expressed in E. coli BL21(DE3), followed by purification for biochemical analysis.

The substrate specificity of the purified enzyme was evaluated using a panel of hydroxycinnamate esters and structurally related substrates (Table 1). The enzyme displayed differential activity towards methyl ferulate (MFA) and its structural analogues, including methyl p-coumarate (MpCA), methyl caffeate (MCA), and methyl sinapate (MSA), with specific activities of 64 ± 1.05, 161 ± 5.76, 139 ± 3.56, and 17 ± 0.56 mU·mg⁻¹, respectively. In contrast, no detectable activity was observed towards p-nitrophenyl palmitate (p-NPP), a typical substrate for lipases, indicating that the enzyme lacks lipolytic function. Based on its substrate profile, the enzyme was classified as a feruloyl esterase and designated as CaFaeA. Interestingly, CaFaeA also demonstrated low but reproducible activity towards bis(2-hydroxyethyl) terephthalate (BHET), a poly(ethylene terephthalate) oligomeric intermediate, with a specific activity of 0.6 ± 0.03 mU·mg⁻¹. Additionally, CaFaeA retained activity against complex substrates such as I-WAX, with a specific activity of 15.7 mU·mg⁻¹. This observation highlights the catalytic promiscuity of CaFaeA and suggests a potentially broader substrate range.

Table 1.

Specific activities of CaFaeA on different types of substrates (mU·mg⁻¹)

	MpCA	MCA	MFA	MSA	BHET	p-NPP	I-WAX
Specific activity	161.0 ± 5.76	139.0 ± 3.56	64.0 ± 1.05	17.0 ± 0.62	0.6 ± 0.03	not detected	15.7 ± 0.98

Further biochemical characterization using MFA as a model substrate revealed that CaFaeA exhibits maximal activity at pH 8.0 (Fig. 1A) and retains over 90% activity after 30 min incubation at pH 10.0 (Fig. 1B), indicating substantial alkali tolerance. The enzyme showed an optimal temperature of 40 °C (Fig. 1C), although its thermal stability was limited, with activity decreasing sharply at elevated temperatures and retaining only 26% at 60 °C (Fig. 1D). These results suggest that CaFaeA is well-suited for catalysis under mildly alkaline conditions, indicating its unique advantage in alkaline biomass treatment processes, especially considering that most FAEs exhibit optimal activity at around pH 6.5 (Oliveira et al. 2019).

Fig. 1.

Biochemical characterization of CaFaeA. A Optimal pH for CaFaeA activity. B pH stability of CaFaeA after incubation at various pH conditions for 30 min. C Optimal temperature for CaFaeA activity. D Thermal stability of CaFaeA after incubation at different temperatures for 30 min

Phylogenetic and sequence clustering analysis of CaFaeA

To place CaFaeA within the broader landscape and elucidate its evolutionary relationships, we performed sequence-based clustering using a curated dataset of 1971 non-redundant FAE homologs retrieved from the BRENDA database and previously published literature (Table S4). Unsupervised clustering analysis resolved the FAE sequences into five major groups (Fig. 2A), each corresponding to distinct phylogenetic lineages and presumed functional specializations.

Fig. 2.

Sequence clustering and phylogenetic analysis of CaFaeA. A Unsupervised sequence clustering of 1971 FAEs revealed five major clades, denoted as cluster 1 (red circles), cluster 2 (blue circles), cluster 3 (orange circles), cluster 4 (green circles), and cluster 5 (violet circles). The first neighbors of AmCE1 were marked by blue rhombus, composed of bacterial species. B Maximum-likelihood phylogenetic tree highlighting the evolutionary relationships among representative members from each cluster, with CaFaeA highlighted in boxes and the related enzymes included in the analysis, along with their respective accession numbers: PET46 (8B4U), TtFAE (WP_011025998), LbFAE (7Z2X), GthFAE (MG834533), FaeLfe (AOR52356), FAELam (AOR52353), FAELfa (AOR52355), FAELac (AOR52354), FAE-C1 (MK607948), FAE-C2 (KM974951), FAE-C3 (MH785296), Lj1228 (WP_011162057), Est1E (2WTM), and BfCinA (AAC44493). C Percent identity matrix showing pairwise sequence similarity across selected FAE homologs

Cluster 1 was predominantly composed of FAEs from fungal species, while cluster 3 mainly contained FAEs derived from Streptophyta. Cluster 2 comprised a mixture of FAEs from both prokaryotic and eukaryotic sources. Notably, AmCE1, a fungal FAE, has been previously reported to cluster more closely with bacterial FAEs based on phylogenetic analysis (Kasmaei et al. 2022), and the clustering results similarly showed its greater association with bacterial-derived FAEs.

CaFaeA was assigned to a distinct cluster 5, composed primarily of bacterial FAEs. Within this cluster, several members from lactic acid bacteria were co-localized, suggesting a possible shared functional ancestry or environmental specialization (Fig. 2B). Intriguingly, cluster 5 also included several catalytically versatile FAEs, such as enzymes PET46 and Est1E (Fig. 2C). Est1E has demonstrated a broad activity against plant-derived hemicellulose (Goldstone et al. 2010). PET46, an archaeal-derived FAE with broad substrate promiscuity, has been reported to outperform canonical PET hydrolases like LCC and IsPETase in hydrolyzing BHET and MHET (Perez-Garcia et al. 2023). Similarly, CaFaeA exhibits activity against BHET and hydroxycinnamate esters, suggesting that CaFaeA represents a functionally divergent subtype within the feruloyl esterase family, displaying promiscuous catalytic activity towards multiple substrates.

Crystal structure of CaFaeA reveals unique substrate recognition lid

To elucidate the structural determinants of its unusual catalytic promiscuity, we solved the crystal structure of CaFaeA at 2.45 Å resolution. The enzyme adopts a canonical α/β-hydrolase fold, featuring a central eight-stranded parallel β-sheet flanked by eight surrounding α-helices (Fig. 3A). The catalytic triad—Ser113, Asp204, and His233—is arranged in a geometry characteristic of this enzyme family; Ser113 is positioned at the nucleophilic elbow between β5 and α3, while Asp204 and His233 reside in loops preceding α8 and α7, respectively. Strikingly, beyond the conserved catalytic core, CaFaeA harbors a structurally distinct lid domain that partially encloses the active-site cavity. This region, located between β6 and α6, comprises three α-helices and two antiparallel β-strands. The short α-helix segment between β6 and α4 is excluded from the main α-helical framework (Fig. 3B).

Fig. 3.

A Monomer structure of CaFaeA. The lid domain is marked in box, the catalytic triad, comprising Ser113, Asp204, and His233, is depicted as spheres. B Secondary structure topology of CaFaeA. The topology diagram depicts the arrangement of α-helices and β-strands in CaFaeA, with secondary structure elements numbered according to their positions in the primary sequence

Identification of gatekeeper residues of Y151 and E168 in lid region modulating substrate accessibility

Structural comparison between CaFaeA and previously characterized FAEs revealed notable variations in the architecture of the lid domain, particularly at the entrance of the catalytic pocket. Bulky polar residues, such as Y151 and E168, are positioned at strategic sites within this region and are presumed to modulate substrate access. To elucidate the functional relevance of these putative gatekeeper residues, site-directed mutagenesis was performed at positions Y151 and E168. Residue Y151, located in a loop preceding α4, is hypothesized to contribute steric control over substrate entry. To assess the impact of reduced side-chain size, Y151 was substituted with Y151G and Y151V, with the latter retaining moderate hydrophobicity but reduced bulk. Residue E168 is located at the turn between two antiparallel β-strands within the lid domain and is not evolutionarily conserved (Fig. S4). Sequence alignments indicate that this position corresponds to glycine in high-activity FAEs GthFAE and varies in others (corresponding Tyr in LbFAE, His in LaFAE, Asn in PET46, Lys in Lj0536, and Asp in Est1E). To investigate the roles of side-chain size and charge at this site, E168 was substituted with E168G, which removes the side chain entirely, and E168R, which introduces a bulky, positively charged group.

To gain mechanistic insights, molecular dynamics (MD) simulations were conducted. Both E168 variants rapidly reached equilibrium and remained structurally stable, whereas the Y151 variants exhibited elevated root mean square deviation (RMSD) values and increased root mean square fluctuation (RMSF) fluctuations in the substrate entry region, particularly beyond 100 ns, indicative of local destabilization (Fig. 4A). RMSF analyses further revealed that the E168 mutations had minimal impact on the surrounding structural flexibility, while Y151G and Y151V caused significant alterations within the lid loop, implicating this region in dynamic gating of substrate access (Fig. 4B). Structural analysis of the equilibrated E168R variant highlighted two mechanisms likely contributing to the observed activity loss: (i) the elongated guanidinium side chain of R168 forms a new hydrogen bond with Y151, increasing steric hindrance at the substrate entrance and (ii) the introduced positive charge may interact electrostatically with the negatively charged ferulate product, impeding product release. By contrast, the E168G mutation, while preserving overall structural integrity, removes a side chain capable of stabilizing substrate interactions via hydrogen bonding or salt bridges, likely compromising substrate recognition and transition-state stabilization (Fig. 4C). For the Y151 variants, structural inspection of the wild-type enzyme revealed a stabilizing π–π interaction between Y151 and F37, which is disrupted upon substitution with glycine or valine (Fig. 4D). This loss of interaction is consistent with the increased structural fluctuations observed in MD simulations and likely contributes to impaired stabilization of the lid domain, resulting in reduced catalytic performance across all tested substrates, including methyl ferulate (Fig. 4E). Together, these findings identify Y151 and E168 as critical gatekeeper residues that govern substrate accessibility and positioning in CaFaeA.

Fig. 4.

Gatekeeping role of Y151 and E168 in CaFaeA substrate recognition. A Root-mean-square deviation (RMSD) profiles of CaFaeA and mutants over a 100-ns molecular dynamics (MD) simulation. B Root-mean-square fluctuation (RMSF) of backbone atoms in CaFaeA and mutants during the 100-ns MD simulation. C Post-equilibration structural overlay of CaFaeA and Y151 mutants. Wild-type CaFaeA is shown in grey, Y151G in light blue, and Y151V in magenta. D Post-equilibration structural overlay of CaFaeA and E168 mutants. Key residues are shown in ball-and-stick representation. Wild-type CaFaeA is shown in grey, E168R in green, and E168G in salmon. E Relative hydrolytic activities of CaFaeA and its mutants toward hydroxycinnamate esters

Organic solvent tolerance of CaFaeA

Research utilizing non-food biomass such as cellulose and lignocellulosic hydrolysates for organic solvent biosynthesis has achieved notable progress (Cheng et al. 2019). Although higher solvent concentrations could substantially reduce downstream separation costs and energy consumption, most reported FAEs, such as AnFAE from Aspergillus niger, exhibit markedly diminished activity in organic solvents like 20% DMSO, acetone, and dioxane when assayed against methyl p-hydroxycinnamates (Faulds et al. 2011). Considering the prominent role of Clostridium species in acetone-butanol-ethanol fermentation processes, we assessed the tolerance of CaFaeA to various organic solvents.

CaFaeA demonstrated remarkable and solvent-dependent tolerance (Fig. 5). It retained full activity in the presence of 25% DMSO and exhibited an approximately 10% enhancement in activity after incubation with n-hexane, suggesting a potential solvent-mediated activation effect. The enzyme also maintained over 60% of its catalytic activity after exposure to 25% methanol or ethanol, indicating moderate tolerance to these polar solvents. Conversely, exposure to acetone or n-butanol caused a drastic activity decline of nearly 80%, while acetonitrile and isopropanol led to near-complete enzyme inactivation. This exceptional tolerance not only underscores the potential of CaFaeA for direct catalytic action on hydrophobic lignin fragments, thereby circumventing environmental issues associated with conventional acid/base pretreatments, but also highlights its suitability for biocatalytic production of ferulic acid. Indeed, FAE has been directly applied to the enzymatic hydrolysis of raw materials such as wheat bran in high DMSO concentrations to improve ferulic acid yields (Faulds et al. 2011).

Fig. 5.

Effects of organic solvents on CaFaeA enzymatic activity. Relative enzyme activities with MFA as substrate. The activity of CaFaeA incubated in 100 mM PBS buffer (pH 8.0) is defined as 100%

Discussion

To characterize the structural variations among different feruloyl esterase (FAE) clusters, structural alignments were performed on representative enzymes from the major clusters. Comparative structural analysis revealed that both the lid region and the serine-containing catalytic motif vary substantially across different FAEs (Fig. 6). In the cluster 1 enzyme, a putative feruloyl esterase from Aureobasidium pullulans (GenBank accession: THW67961), this domain is comprised entirely of an extended, flexible loop, with the conserved motif typically appearing as GKSNG (Fig. 6A). Conversely, FAE-XynZ, a member of cluster 2 (PDB: 1JJF), lacks a distinct lid element at the corresponding position, leaving the active site fully exposed to form a solvent-accessible, open catalytic cavity. When acting on feruloylated arabinose substrates, surface-exposed aromatic residues specifically recognize and bind the arabinose moiety via aromatic stacking interactions, resulting in high catalytic efficiency. However, it exhibits poor activity toward small-molecule substrates such as ethyl ferulate due to the absence of a specialized substrate recognition mechanism. Furthermore, its conserved motif is also unique, frequently appearing as GFSYG (Fig. 6B). FoFaeC (PDB:6FAT), a representative of cluster 3, features a voluminous lid domain (residues 231–390) positioned directly above the active site, composed of multiple α-helices that define a distinct substrate-binding pocket. Notably, this domain incorporates a calcium ion remote from the catalytic center; while not directly involved in catalysis, this ion coordinates with surrounding residues to maintain the structural rigidity of the lid (Fig. 6C). The cluster 4 enzyme, a putative feruloyl esterase from Medicago truncatula (GenBank accession: RHN64625), is structurally similar to the representative cluster 5 enzyme, CaFaeA, with an RMSD of 1.726 Å. Their lid regions are both composed of multiple α-helices and anti-parallel β-sheets. Notably, the catalytic serine of cluster 4 is located within a highly conserved and well-defined GHSKG motif (Fig. 6D); in contrast, the motif in cluster 5 exhibits lower conservation than that of cluster 4 (Fig. 6E).

Fig. 6.

Structural comparison of representative feruloyl esterases (FAEs) from clusters 1 to 5. The three-dimensional structures illustrate the architectural diversity of the lid domains (highlighted in pink) relative to the conserved core α/β-hydrolase fold. The nucleophilic elbow, containing the catalytic serine, is depicted by the sequence logo below each structure, showing the conservation of the pentapeptide motif (G-X-S-X-G). A Cluster 1 representative (GenBank accession: THW67961). B Cluster 2 representative, FAE-XynZ (PDB: 1JJF). C Cluster 3 representative, FoFaeC (PDB: 6FAT). D Cluster 4 representative (GenBank accession: RHN64625). E Cluster 5 representative, CaFaeA

Given this structural diversity, particularly within the lid domain, we solved the crystal structure of CaFaeA and conducted a comparative analysis focusing on the lid domain characteristics of its closest structural homologs in cluster 5 (Table 2). Structural superposition with LaFAE from Lactobacillus acidophilus indicated that while the overall fold is conserved, subtle architectural differences exist in the lid domain. Specifically, LaFAE features an additional antiparallel β-strand pair formed between α5 and α6, leading to reduced lid flexibility. In GthFAE, rational engineering of the lid domain via the T41N/T150R double mutation enhanced thermostability at the expense of catalytic efficiency, highlighting a trade-off between structural rigidity and functional versatility (Yang et al. 2022). Similarly, in LbFAE, the lid domain (residues 140–186) is composed of three α-helices and four β-strands, forming a relatively rigid framework that correlates with its diminished activity towards arabinose esters. This structural rigidity likely limits conformational adaptability and may contribute to narrow substrate specificity. By contrast, Est1E from Butyrivibrio proteoclasticus possesses a more flexible lid architecture, comprising two short α-helices and two β-sheets interspersed with an extended loop region. This flexible element is suggested to facilitate dynamic substrate accommodation, contributing to the broad substrate spectrum of Est1E (Goldstone et al. 2010).

Table 2.

Unique structural features of the lid domain in feruloyl esterases

α-helices are represented by pink cylinders, β-sheets are indicated with arrows

In CaFaeA, residues 147–152 form an antiparallel loop projecting toward the nucleophilic Ser113, creating a steric barrier that occludes the active site. This occlusive loop is also observed in LbFAE, LaFAE, Lj0536, and Est1E. In contrast, PET46, TtFAE, and GthFAE possess a rigid α-helix in this region, which is oriented parallel to the central β-sheet, resulting in a more solvent-exposed and accessible active site (Fig. 7). The rigid configuration of the lid likely imposes steric constraints on substrate entry, favoring small phenolic esters while the loop (147–152) near Ser113 provides sufficient flexibility to accommodate bulkier substrates like I-WAX. This structural trait suggests a gating mechanism that confers substrate specificity while enabling latent catalytic promiscuity. Overall, for CaFaeA, substrate specificity is dictated by the lid architecture overlying the active site. While this domain imposes steric constraints that favor the fixation of small substrates, the antiparallel loop (residues 147–152) imparts essential structural plasticity. This flexibility allows the enzyme to modulate its active pocket when encountering bulky substrates, thereby satisfying the stereochemical demands of a broad substrate spectrum.

Fig. 7.

Structural comparison of the lid domain (light pink) among feruloyl esterase. The structural region (residue 147–152) of CaFaeA is marked in orange

As natural lignocellulosic substrates provide a more rigorous assessment of enzymatic performance than model compounds such as MFA, we evaluated CaFaeA activity using insoluble wheat arabinoxylan (I-WAX). To highlight the performance disparity between complex and model substrates, activity data from CaFaeA and 17 previously characterized FAEs were compiled and compared using both I-WAX and MFA (Table 3). A normalized parameter, relative specific activity (mg·μmol⁻¹·h⁻¹), was defined to facilitate a standardized comparison of catalytic efficiency across diverse experimental conditions. CaFaeA exhibited a relative specific activity of 5.39 mg·μmol⁻¹·h⁻¹, a value that surpasses that of over 70% of reported FAEs. Notably, MgFae1 features a C-terminal carbohydrate binding module (CBM) domain that likely facilitates stronger binding to long-chain xylans, thereby enhancing the degradation of I-WAX (Fig. 8A). Structurally, FAE-C1, FAE-C2, and FAE-C3 share high similarity with CaFaeA, with RMSD values of 0.76 Å, 0.89 Å, and 0.78 Å, respectively. Compared to CaFaeA, the rigid α4 helix that lies in the lid domain of CaFaeA is replaced by a flexible loop region in FAE-C3 (Fig. 8B). The conformational flexibility of this region more effectively facilitates the access of complex substrates to the catalytic center of FAE-C3. In contrast, both FAE-C1 and FAE-C2 possess an additional lengthy loop region at the N-terminus. The unregulated flexibility of this loop may obstruct substrate access to the active-site cavity.

Table 3.

Kinetic parameters and specific activities of FAEs towards MFA and their hydrolysis capabilities on I-WAX

Enzyme	Specific activityb (mg·μmol−1·h −1)	Specific activityc mU·mg⁻1	kcat/Kme (mM−1·s−1)	References
CaFaeA	5.39	64	2.30	This study
FAE-C1	4.13	n.r.d	n.r	(Wong et al. 2019)
FAE-C2	3.63	n.r	n.r
FAE-C3	9.63	n.r	n.r
AnFae1	0.80	7608	n.r	(Lin et al. 2023)
AoTan1	0.13	15,080	n.r
AoFae1	0.38	5460	n.r
AoTan2	0.012	530	n.r
AoTan3	0.15	3720	n.r
McFae1	0.16	48	n.r
TtFae1	0.50	2110	n.r
HiFae1	7.46	2410	n.r
NcFae1	0.53	3410	n.r
MgFae1	11.26	970	n.r
WtsFae1A	0.82	750	n.r
WtsFae1B	0.017	2510	n.r
BtFaea	0.54	n.r	300	(Du et al. 2023)
BtFaeW180Aa	1.45	n.r	1330	(Wang et al. 2024)
BtFaeW180Da	1.40	n.r	1070

^aFeruloyl esterase was used in conjunction with a commercial xylanase to release ferulic acid from I-WAX

^bThe specific activity was calculated based on previously reported enzymatic activities. Data for esterase and xylanase homologues were excluded from the analysis

^cOne unit of enzyme activity (1U) was defined as the amount of product released by 1 mg of purified enzyme per minute under the assay conditions

^dn.r., not reported

^eEnzymatic kinetics of FAEs toward methyl ferulate

Fig. 8.

A Surface hydrophobicity of MgFae1, CBM is marked in black. B Comparative analysis of the structural features of CaFaeA, FAE-C1, FAE-C2, and FAE-C3. The N-terminus is highlighted in black, and the α4 region of CaFaeA is outlined with a box. C Surface hydrophobicity and the substrate-binding pocket volume of the enzymes. The binding pocket of each enzyme is highlighted by a black frame

Interestingly, fungal FAEs such as NcFae1, AoFae1, and AnFae1 display markedly higher catalytic efficiency than CaFaeA toward small molecule substrates like MFA, with activities 53.2-, 85.3-, and 118.8-fold higher, respectively. However, their relative specific activities toward I-WAX were substantially lower than that of CaFaeA (Table 3). Structural comparisons revealed considerable deviations from CaFaeA, with RMSD values of 2.72 Å, 16.84 Å, and 12.18 Å for NcFae1, AoFae1, and AnFae1, respectively, particularly around the substrate-binding pocket. The pocket volumes of NcFae1, AoFae1, and AnFae1 were calculated to be 322.88 ų, 330.50 ų, and 323.12 ų, respectively, whereas CaFaeA possesses a markedly expanded pocket volume of 856.50 ų (Fig. 8C). This structural feature likely underlies the superior performance of CaFaeA in processing bulky lignocellulosic substrates such as I-WAX. Although the release of ferulic acid resulting from CaFaeA hydrolyzing the ester bonds between ferulic acid and arabinose-containing polysaccharides within plant cell wall polymers could increase toxicity (Cerisy et al. 2017), this underscores the critical synergy between “enzyme engineering” and “strain engineering.” Therefore, the expression of CaFaeA in such pre-adapted, resistant strains represents an ideal strategy: CaFaeA successfully unlocks the carbohydrate backbone for fermentation, while the host’s evolved resistance mitigates the toxicity of the released ferulic acid.

Conclusion

In this study, we characterized CaFaeA from Clostridium acetobutylicum and revealed its broad substrate promiscuity, including activity toward hydroxycinnamate esters and the PET oligomer BHET. This indicates that the FAE family is a potential resource for identifying new enzymes capable of PET degradation. Structural analysis uncovered a distinctive lid domain that partially occludes the active site, with site-directed mutagenesis and molecular dynamics simulations identifying Y151 and E168 as critical gatekeeper residues governing substrate access and catalytic performance. Notably, CaFaeA exhibited remarkable tolerance to organic solvents such as DMSO and n-hexane, maintaining high activity under conditions that typically inactivate other FAEs. CaFaeA efficiently hydrolyzed ferulic acid from I-WAX, surpassing the catalytic efficiency of the majority of reported FAEs on lignocellulosic substrates. Collectively, these findings establish CaFaeA as a structurally and functionally distinct feruloyl esterase with promising potential in lignin utilization and biomass-derived bioprocessing applications.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

SL, SX, and YF conceived and designed research. SL, XH, XC and LQ conducted the research. XL, TF, YF, and XH contributed discussions and suggestions. SL analyzed data. SL and YF wrote the manuscript. All authors read and approved the manuscript.

Funding

This work was supported by grants from the National Key Research and Development Program of China (2022YFC2105602); the Department of Science and Technology of Liaoning Province (2024JH2/102600031); and the Department of Science and Technology of Liaoning Province (2025-MS-005).

Data availability

The datasets and materials from this study can be available upon request to interested researchers.

Declarations

Ethical approval

Not applicable.

Competing interests

The authors declare no competing interests.