Metarhizium spp. encode an ochratoxin cluster and a high efficiency ochratoxin-degrading amidohydrolase revealed by genomic analysis

Graphical abstract

Highlights

• •Three Metarhizium species were found to contain an ochratoxin (OT) cluster.
• •M. brunneum and M. robertsii degrade both OTA and OTB.
• •MbAmh1 was traced and demonstrated great potential for industrial application.
• •The working model for MbAmh1 was revealed.
• •Metarhizium spp. contain both OT cluster and OT-degradation enzyme.

Abstract

Introduction

Ochratoxins (OTs) are worldwide regulated mycotoxins contaminating a variety of food-environment and agro-environment. Several Aspergillus and Pencillium species synthesize OTs from a six-gene biosynthetic gene cluster (BGC) to produce the highly toxic final product OTA. Although many studies on OTA-degrading enzymes were performed, high efficiency enzymes with strong stability are extremely needed, and the OTA degrading mechanism is poorly understood.

Objectives

The study aimed to explore the OT-degradation enzyme and investigate its degradation mechanisms in Metarhizium, which contain an OT biosynthetic gene cluster.

Methods

Phylogenomic relationship combined with RNA expression analysis were used to explore the distribution of OT BGC in fungi. Bioactivity-guided isolation and protein mass spectrometry were conducted to trace the degrading enzymes in Metarhizium spp., and the enzymes were heterologously expressed in E. coli and verified by in vitro assays. Structure prediction and point mutation were performed to reveal the catalytic mechanism of MbAmh1.

Results

Beyond Aspergillus and Pencillium species, three species of the distant phylogenetic taxon Metarhizium contain an expressed OT-like BGC but lack an otaD gene. Unexpectedly, no OT BGC products were found in some Metarhizium species. Instead, Metarhizium metabolized both OTA and OTB to their non-toxic degradation products. This activity of M. brunneum was attributed to an intracellular hydrolase MbAmh1, which was tracked by bioactivity-guided proteomic analysis combined with in vitro reaction. Recombinant MbAmh1 (5 μg/mL) completely degraded 1 μg/mL OTA within 3 min, demonstrating a strong degrading ability towards OTA. Additionally, MbAmh1 showed considerable temperature adaptability ranging from 30 to 70 °C and acidic pH stability ranging from 4.0 to 7.0. Identification of active sites supported the crucial role of metal iron for this enzymatic reaction.

Conclusion

These findings reveal different patterns of OT synthesis in fungi and provide a potential OTA degrading enzyme for industrial applications.

Introduction

Ochratoxins (OTs) are a group of toxic secondary metabolites (SMs) contaminating a variety of foodstuffs including grains, legumes, coffee, dried fruits, wine, and meat products [1], as well as soil and agro-environmental matrices, such as manure, sewage sludge, drainage water and sediments [2]. OTs occur in several natural structures (e.g. OTA, OTB and OTC), whereby OTA is the most notorious and toxic due to the chlorine atom on the benzene ring [3], [4]. OTA has been classified as a group 2B human carcinogen by the International Agency for Research on Cancer [5]. OTA has also been proven to be nephrotoxic, hepatotoxic, teratogenic and immunotoxic and causes kidney and liver tumors in several animal species [6], [7].

Until now, the known OTA-producing species include members of the genera Aspergillus and Penicillium [8], [9], [10]. In fungi, genes encoding the proteins required to produce a compound are usually arranged in a contiguous fashion as a biosynthetic gene cluster (BGC) [11], [12]. The OTs biosynthetic gene cluster consists of six adjacent co-regulated genes. A polyketide synthase (PKS) and a non-ribosomal peptide synthase (NRPS) are encoded by otaA and otaB, respectively. otaC encodes a cytochrome P450 monooxygenase and otaD encodes the halogenase which catalyzes OTB to OTA. The pathway-specific transcription factor otaR1 regulates the expression of OTA biosynthetic genes and the production of OTs compounds [13], [14], [15], [16], [17]. More recently, a cyclase otaY located between otaA and otaB has been annotated to putatively catalyze the cyclization of the polyketide backbone [9], [18].

To reduce the risk of exposure to this mycotoxin, reliable approaches are needed to eliminate OTA in food and feed. The mycotoxin detoxification by substrate-specific, environmentally friendly and highly efficient biological methods has attracted increased attention [19], [20]. The most effective mechanism for OTA degradation is to hydrolyze the amide bond in molecular, which is formed by NRPS in the biosynthesis pathway [21]. The degrading products are phenylalanine and OTα, and the toxicity of OTα to livestock is at least 500 times lower or even non-toxic than OTA [22]. A number of species of bacteria have the capacity of degrading OTA, such as the Bacillus amyloliquefaciens strain isolated from grain depot-stored maize [23], the Bacillus subtilis strain isolated from fresh elk droppings [24], the Alcaligene [25], Acinetobacter [26] and Lysobacter spp. isolated from soil [27]. The isolation of degrading microbes primarily involves extensive screening from soil, food, and various microbial environments, often resulting in increased workload due to the lack of direction. Therefore, efficient strategies to explore degrading microbes and enzymes are still needed.

Identifying the degrading enzymes and understanding the molecular degradation mechanism are more important to OTA detoxification development. Initially, several commercial hydrolases, such as carboxy-peptidase A (CPA) [28], amano A [29] and protease A [30] were used to degrade OTA, but they showed a relatively low catalytic efficiency. Afterwards, an ochratoxinase from A. niger was obtained by chromatographic separation and mass spectrometry (MS) identification [31]. Using ochratoxinase as a reference, a structure-based approach to mine new enzymes with potential OTA-degrading activity was established [32]. Recently, a combination of genome screening and heterologous expression was used to excavate potential degrading enzymes. Cp4 [27], DacA [24] and ADH3 [33] have been reported to degrade OTA. Despite numerous studies on OTA-degrading enzymes, there remains a pressing need for highly efficient enzymes with robust stability. Additionally, the mechanism underlying OTA degradation remains poorly understood.

In the current study, it was demonstrated that three Metarhizium species contain an OT-like cluster and possess the capability to degrade OTA/OTB. Moreover, comparative analysis of bioactive protein fractions and subsequent heterologous expression indicated that MbAmh1 is responsible for OTA degradation in M. brunneum. Enzymatic assessment of MbAmh1 showed its potential for OTA detoxification in industrial settings. Protein structure examination and point mutation analysis identified a comparable operational model for OTA-degrading enzymes.

Materials and methods

Strains and culture conditions

The fungal strains were grown at 28 °C on Potato Dextrose Agar (PDA, Becton, Dickinson and Company) plates. For extraction of DNA and RNA, A. westerdijkiae fc-1, M. brunneum ARSEF 3297 and M. robertsii ARSEF 23 were cultivated in the liquid Potato Dextrose Broth (PDB, Becton, Dickinson and Company), shaking at 180 rpm at 28 °C for 5 days. Fungi were cultured under the same conditions for protein extraction, except for a cultivation time of 7 days. Yeast Extract Sucrose (YES, 5 % yeast extract, 40 % sucrose and 10 % agar) and Rice Media (RM, 20 % rice) were used also used to analyze OTs production from M. brunneum and M. robertsii.

OT gene cluster identification in fungi

Putative OT genes were identified by BLAST searches of multiple databases with OT genes from A. niger CBS 513.88 (An15g07880 – An15g07920) and A. westerdijkiae fc-1 (AoFC_09698 – AoFC_09702) as query sequences. A putative OT BGC was identified if the specified conditions were satisfied: 1) The amino acid sequence displayed e-values less than 10⁻⁵ compared to the query sequence; 2) The homologs of six OT genes were continuously organized as a cluster in the target genome; 3) The putative clusters were screened manually.

Phylogenomic analysis

For phylogenomic analysis of fungal species (Table S1), groups of proteins were detected by OrthoFinder v2.4.0 [34]. The single copy orthologues were aligned with MUSCLE [35], and the poor alignment regions were trimmed using Gblocks [36]. The high quality sequences were used for the maximum likelihood phylogeny analysis with RAxML-NG v. 0.5.1b [37]. Bootstrap support value was calculated by analyzing 1,000 replicates. For protein phylogeny, the sequences were aligned by MUSCLE, and then a maximum likelihood tree was constructed by treeBeST (https://treesoft.sourceforge.net/treebest.shtml) using 1,000 bootstrap replicates [38].

DNA isolation and genome fragment cloning

Fungi were cultivated in PDB for 5 days. Mycelia of the Aspergillus and Metarhizium spp. were harvested via filtration. Genomic DNA was isolated using a Genomic DNA Kit (DP305, TIANGEN, China), following the manufacturer’s protocol. Genome regions for OT genes were obtained by PCR amplification with primer pairs listed in Table S2.

RNA isolation and reverse transcription

Fungi were cultivated in the PDB, shaking at 180 rpm at 28 °C for 7 days and mycelium tissues were harvested via filtration. Total RNA was extracted using TRIzol reagent (Invitrogen, USA) according to the manufacturer’s recommendations. The integrity and concentration of RNA were evaluated by agarose gel electrophoresis and a NanoDrop 2000c spectrophotometer (Thermo Fisher Scientific, USA). Complementary DNA (cDNA) was synthesized with a TIANScript Ⅱ RT Kit (TIANGEN, China). The OT gene expression level based on cDNA was analyzed with primers listed in Table S2.

Isolation of crude proteins

Mycelia of M. brunneum, M. robertsii and A. westerdijkiae filtered from PDB culture were ground in liquid nitrogen. Mycelium powder of 1.0 g was dissolved in lysis buffer [10 mM Tris-HCl (pH=7.5), 150 mM NaCl, 0.5 mM EDTA, 0.1 % Triton X-100, 1 mM DTT, 1 mM phenylmethanesulfonyl fluoride (PMSF)], stirred with a vortex mixer, and centrifugated at 12000 rpm for 10 min. The concentration of crude proteins was evaluated by a NanoDrop 2000c spectrophotometer (Thermo Fisher, USA) and adjusted to 10 mg/mL.

Protein mass spectrometry and data acquisition

The crude proteins of M. brunneum and M. robertsii were separated by regenerated cellulose type membranes (Millipore, USA) with molecular weight cut-offs (MWCO) of 100, 50, 30 and 10 kDa according to the instruction. The active fractions of 100 µg were reduced by dithiothreitol at 37 °C for 1 h. Then the sample was diluted by 25 mM ammonium bicarbonate buffer of and enzymolyzed by trypsin (trypsin/protein = 1/50) overnight. Nanoflow LC-MS/MS analysis of tryptic peptides was performed on a quadrupole Orbitrap mass spectrometer (Orbitrap fusion lumos, Thermo Fisher Scientific, Germany) coupled to an EASY nLC 1200 ultra-high pressure system (Thermo Fisher Scientific) via a nano-electrospray ion source. Peptides of 1 µg were loaded on a column of 25 cm length and 150 μm inner diameter, packed using ReproSil-Pur C18-AQ 1.9- µm silica beads (Dr. Maisch, Germany). After, data were exported to Proteome Discoverer suite version 2.4 [39] to identify and quantify peptide spectra and protein list. The spectra of M. brunneum and M. robertsii was searched against the proteomes UP000031181 and UP000002498 from the UniProtKB database. Peptides were assembled into proteins and were further filtered based on the combined probabilities of their constituent peptides to a final false discovery rate (FDR) less than 0.01.

Heterologous expression

The candidate genes for OTA degradation were amplified from cDNA with primers containing restriction enzyme sites listed in Table S2. The vector pET28a (+) with His₆-tags at both the N and C terminus of the inserted gene was linearized and then ligated with the target gene through restriction-ligation. The recombinant plasmids were heat-shock transformed into E. coli strain BL21 (Transgen Biotech, China). The positive transformants were incubated in LB broth (tryptone 10 g/L, yeast extract 5 g/L, NaCl 10 g/L) at 37 °C with agitation at 180 rpm for protein expression. After OD₆₀₀ reached to 0.6, the cells were induced with 1 mM isopropyl-β-D-thiogalactoside (IPTG) at 16 °C for 14 h. Then, the cells were harvested by centrifugation at 4000 rpm for 10 min and resuspended in 30 mL W1 buffer (50 mM Tris-HCl, pH 8.0, 300 mM NaCl) containing 1 mM PMSF, 10 mM MgCl₂, 20 μg/mL Lysozyme and 1 μg/mL DNase I). The cells were homogenized at a high pressure of 1000 Mpa and 4 °C for 15 min, followed by centrifugation at 12000 rpm for 45 min. The expression of target proteins was identified by using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The predicted molecular weight (MW) of His₆-tagged MbAmh1 was 55.4 kDa.

Protein purification

Cell lysate containing His₆-tagged proteins was transformed and incubated with Ni-NTA agarose (Qiagen, Germany). The mixture was uploaded onto a pre-equilibrated column and then stepwise washed by buffer W1 (50 mM Tris-HCl pH 8.0, 300 mM NaCl), W2 (50 mM Tris-HCl pH 8.0, 500 mM NaCl, 20 mM imidazole) and W3 (50 mM Tris-HCl pH 8.0) at the flow rate of 2 mL/min. The target proteins were collected by Ni-Elute buffer (50 mM Tris-HCl, pH 8.0, 300 mM imidazole) at the flow rate of 1 mL/min, followed by ion-exchange chromatography (Source 15Q, GE Healthcare Life Sciences, USA). The purified proteins were concentrated by ultrafiltration (30 kDa cut-off) and stored with Superdex buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 2 mM DTT) at −80 °C for degradation assays. Protein concentrations were determined with a spectrophotometer (SpectraMax 190, Molecular Devices) and adjusted to 5.0 mg/mL.

OTs degradation assays

OTs degradation test was conducted in a 1.0 mL reaction system. 100 μL crude protein was added to the reaction tube to reach a final concentration of 1.0 mg/mL, while 0.5 μL purified protein MbAmh1 was added to reach a final concentration of 2.5 μg/mL. OTA or OTB standard (1.0 mg/mL) was added to the reaction tube with the final concentration of 1 μg/mL. The reaction system was supplemented with Tris-HCl buffer (10 mM, pH 6.0) to a final volume of 1.0 mL. The degradation tubes were incubated at 28 °C for 48 h, followed by the addition of 1 mL methanol to stop the reaction at different time points. The reaction solution was filtered through a 0.22 μm filter into a vial. The degradation products were further identified by high performance liquid chromatography (HPLC)/mass spectrometry (MS).

Enzymatic activity assays

Reactions containing MbAmh1 of 2.5 μg/mL and OTA of 1 μg/mL were performed to test the influence of pH, temperature and metal ions. The optimum pH for MbAmh1 was tested within the pH range of 3.0 to 10.0 at 28 °C. The temperature ranges from 20 to 90 °C at the interval of 10 °C was determined for OTA degradation at pH 6. The influence of metal ions (50 mM) was determined by adding sulfate (i.e. Zn²⁺, Mn²⁺, Fe²⁺, Fe³⁺, Ni²⁺, Cu²⁺ and Mg²⁺) and chloride (i.e. Li⁺ and Ca²⁺) to the reaction tube, and the reaction was performed at 28 °C and pH 6 for 20 min.

OT metabolite profiling

OTA, OTB, OTα and OTβ were detected by HPLC with fluorescence detection (FLD) with an Agilent HPLC system (Agilent Technologies, USA). The analysis was performed on an Agilent ZorbaxSB-C18 (4.6 mm × 250 mm, 5 μm) column that was coupled with a fluorescence detector at excitation and emission wavelengths of 333 nm and 460 nm, respectively. OT metabolites were eluted with mobile phase acetonitrile/water/acetic acid (99/99/2, v/v/v) at a flow rate of 1 mL/min. Further confirmation of degrading metabolites was carried out on an Agilent QQQ Mass Spectrometer equipped with an electrospray ionization (ESI) interface. The ion source was operated in positive (OTA and OTB) and negative (OTα and OTβ) ionization modes at 380 V.

Protein structure prediction and point mutations

The AlphaFold was employed to deduce the 3D architecture of the presumed MbAmh1 protein [40]. The prediction of catalytic sites was performed by conducting sequence alignment of homologous proteins using ClustalW (https://www.genome.jp/tools-bin/clustalw) and then performing structural comparison using PyMOL software (https://pymol.org/2/). Substitutions of the potential amino acids with alanine were carried out via the Quick-Change mutation method [41]. The mutated genes were individually expressed using E. coli BL21, and the OTA degrading capability of the mutated protein was assessed with MbAmh1 as control. The activity of mutants was detected at 20 min from the reaction of 2.5 μg/mL purified enzyme and 1 μg/mL OTA at 28 °C and pH 6.

Statistical analysis

All the data were analyzed using GraphPad Prism version 5.01 and IBM SPSS statistics version 20. The values were presented as means and standard errors of three biological replicates. Statistical significance was calculated using one-way ANOVA and means were compared by least significant difference (LSD) and Duncan’s test. The difference was regard to be statistic significant at p < 0.01.

Results

A conserved OT-like cluster existence in three Metarhizium genome

We surveyed the distribution of OT genes in phylogenetically diverse fungi by BLAST analysis using the amino acid sequences of OTA encoding genes from A. niger CBS 513.88 and A. westerdijkiae fc-1 as queries against 278 Aspergillus and 54 Penicillium genomes from Joint Genome Institute (JGI) and the non-redundant protein sequences database at National Center for Biotechnology Information (NCBI). This search identified an OT BGC – or remains thereof – in 18 Aspergillus species and two Penicillium species, which was consistent with previous results [9]. A gene rearrangement has taken place in A. albertensis and A. alliaceus among otaC, otaD and otaR1. Interestingly, we found that three Metarhizium species from the distant fungal taxon Sordariomycetes, namely, Metarhizium robertsii, M. brunneum and M. anisopliae, all contained an OT-like cluster included otaA (PKS), otaY (Cyclase), otaB (NRPS), otaC (P450) and otaR1 (TF) (Fig. 1), but lacked the halogenase otaD. The five genes of the OT-like cluster are highly conserved among three Metarhizium. The identities of these five genes between Metarhizium sp. and A. westerdijkiae were 48 %, 45 %, 46 %, 49 % and 41 %, respectively (Fig. S1).

Fig. 1.

Distribution of OTA genes in fungi. Phylogenomic relationships were inferred from 2,279 single-copy orthologues present in all genomes, and all the values of bootstrap support were more than 50. The species were manually clustered into five clades (red numbers 1–5) based on their genetic relationship. The OT genes were colored according to their functions and the length scale (kb) was placed on top. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

To provide an overview of the relationships among the OT BGC-containing species, we constructed a phylogenomic tree using single-copy homologous proteins from the fungal genomes (Table S1 and Fig. 1). Phylogenomic analysis supported the species clustered into five clades. Clades 1, 2 and 3 contain fungi from Aspergillus section Circumdati, Flavi and Nigri, respectively. Two Penicillium (section Fasciculata) formed clade 4. Three closely related Metarhizium species formed a sole branch (clade 5) in Sordariomycetes. The putative OT BGCs are shown in Fig. 1.

Verification of OTB cluster in Metarhizium at DNA and RNA level

OTs and the related metabolites have never been reported to be produced in Metarhizium, although the OT-like clusters are present in three Metarhizium species as mentioned above (Fig. 1). Based on the biosynthetic pathway of OTA, an otaD is required for the final chlorination, the absence of which would lead to OTB accumulation in Metarhizium spp. To verify the existence of the sequences, Polymerase Chain Reactions (PCRs) were used to amplify the 21,962 bp cluster from the genome of M. brunneum strain ARSEF 3297 and the 22,283 bp cluster from M. robertsii strain ARSEF 23. As shown in Fig. 2A, six fragments with overlaps cover the full length of OT BGC of the two species, respectively. The transcripts of five biosynthetic gene were also successfully amplified by Reverse Transcription qPCR (RT-qPCR), with A. westerdijkiae serving as the positive control (Fig. 2B). Unfortunately, we were unable to detect the presence of OTB in the crude extracts from M. brunneum and M. robertsii, grown under three OT permissive conditions (PDA, YES and RM) despite observing apparent transcriptional activation (Fig. 2B). Thus we wondered if OTB was being made, either in amounts too small to measure or if, possibly, the product was being degraded as both OTA and OTB have been reported to be degraded to OTα and OTβ by some bacteria, respectively.

Fig. 2.

Identification of OT genes in Metarhizium. (A) Genomic OTA sequence cloned from M. brunneum and M. robertsii. M represents the DNA marker, and lanes 1–6 represented six overlap fragments. (B) RT-qPCR amplification of the genes in OTA cluster from A. westerdijkiae, M. brunneum and M. robertsii. RG indicated reference gene GADPH.

The degradation of OTA/OTB by Metarhizium species

To determine if Metarhizium is indeed degrading OTs, the extra- and intra- cellular enzyme fractions of M. brunneum, M. robertsii and A. westerdijkiae were tested for their degrading activity towards OTA and OTB. After a 24 h co-incubation of intracellular enzymes (1 mg/mL) from M. brunneum and M. robertsii with OTA, HPLC-FLD analysis showed the absence of the OTA peak and the presence of a new peak in the elution profile (Fig. 3A). The retention time and fluorescence spectra of the produced compound were identical to OTα standard. Furthermore, HPLC-MS demonstrated the peak as a molecular ion [M-H]⁻ at m/z 255, which was identical to the molecular weight of OTα (Fig. S2). The intracellular enzyme from OT producing A. westerdijkiae fc-1 also demonstrated a weak degrading ability towards OTA. Noteworthy, the supernatant (extracellular enzymes) of both Metarhizium and A. westerdijkiae did not degrade OTA. Similarly, intracellular but not extracellular enzymes of Metarhizium and A. westerdijkiae were able to degrade OTB, whereas Metarhizium spp. showed higher activity (Fig. 3B and Fig. S2).

Fig. 3.

The degradation of OTA and OTB by M. brunneum, M. robertsii and A. westerdijkiae. The degradaion tubes containing OT stadards, crude protein and Tris-HCl buffer were incubated at 28 °C and pH 6.0. Chromatograms were obtained after a 0–48 h co-incubation of OTA (A)/OTB (B) with intra- and extra- cellular enzymes. Ⅰ, standard; Ⅱ, intracellular enzymes from M. brunneum; Ⅲ: intracellular enzymes from M. robertsii; Ⅳ, intracellular enzymes from A. westerdijkiae; Ⅴ, extracellular enzymes from M. brunneum; Ⅵ, extracellular enzymes from M. robertsii; Ⅶ, extracellular enzymes from A. westerdijkiae. The concentrations of OTA/OTα after degradation by intracellular enzymes from M. brunneum (C), M. robertsii (D) and A. westerdijkiae (E). The concentrations of OTB/OTβ after degradation by intracellular enzymes from M. brunneum (F) M. robertsii (G) and A. westerdijkiae (H). The bars and deviation represent means and standard errors of three biological replicates.

Evaluation of degradation over time showed that the quantities of OTα and OTβ corresponded to the theoretical concentration calculated from the complete OTA and OTB added to the reaction system. The degradation of OTA by intracellular proteins from M. brunneum and M. robertsii was initiated before 0.5 h, and the entirety of OTA (1 μg/mL) had been hydrolyzed to OTα within 12 h (Fig. 3C and D). However, for the reaction by intracellular proteins from A. westerdijkiae, the concentration of OTA did not decrease over time after 12 h and 16 h (Fig. 3E). OTB degradation exhibited similar features to OTA (Fig. 3F, G and H). However, OTB could not be degraded completely by M. brunneum and M. robertsii even after extending the reaction time from 24 h to 48 h. This may indicate that Metarhizium was producing small amounts of OTB under these conditions or simply that the degrading enzyme(s) were not as efficient as for OTA. Together, these data supported the ability of Metarhizium to enzymatically degrade OTs and A. westerdijkiae as well but to a more limited capacity.

Identification of a new amidohydrolase for OTs degradation from Metarhizium

We fractionated and collected the intracellular enzymes of Metarhizium using regenerated cellulose-type membranes with MWCO of 100, 50, 30 and 10 kDa (Fig. 4A). The vast majority of the hydrolytic activity was detected in the fraction of proteins between 100 and 50 kDa (Fig. 4B). We subsequently determined the sequences of proteins present in this fraction in M. brunneum and M. robertsii using Orbitrap Fusion Lumos mass spectrometer. The identified peptides corresponded to 892 proteins in M. brunneum and 673 proteins in M. robertsii. Gene ontology analysis revealed a similar pattern of protein categorization for the two fungi. The enrichment of enzymes associated with proteolysis activity supported the degrading activity towards amides (Fig. 4C). Given that M. brunneum and M. robertsii exhibited identical degradative activity, we speculated one or more conserved proteins were responsible for their degrading activity. The possibility that species-specific OTA-degrading enzymes might also be involved in OTs degradation has been not considered here. A total of 499 proteins were mapped to both M. brunneum and M. robertsii, with protein sequence identity ≥ 80 % as a cut-off (Fig. 4D). Manual categorization suggested that among them, 47 enzymes had the potential to hydrolyze OTA, including amidohydrolases, amidases and peptidases (Table S3).

Fig. 4.

Identification of OTA degrading enzyme in M. brunneum and M. robertsii. (A) Crude proteins were separated by regenerated cellulose-type membranes with MWCO of 100, 50, 30 and 10 kDa to obtain fractions a to e. (B) Chromatograms were obtained after a 24 h co-incubation of OTA with crude protein fractions from M. brunneum (a1–e1) and M. robertsii (a2–e2). (C) Gene ontology analysis of proteins from fraction b1 and b2. (D) The number of proteins shared by fraction b from M. brunneum and M. robertsii. (E) PCR amplification of MbAmh1 from M. brunneum cDNA, M represents DNA marker. (F) SDS-PAGE analysis of the purified MbAmh1, M represents protein marker.

An in vitro strategy was designed to identify degradation enzymes from the more active M. brunneum. We were able to amplify 38 out of the 47 genes from cDNA, and all of them were cloned and expressed in an E. coli BL21 expression vector. Either extra- or intra-cellular fraction of recombinant E. coli was incubated with OTA and OTB of 1 μg/mL at 28 °C for 24 h, respectively. The extracellular fractions for the 38 candidate enzymes were unable to degrade either OTA or OTB. However, the intracellular fraction of KID75740.1, named MbAmh1, was found to completely degrade OTA and OTB, while no degraded product was detected for other candidate enzymes (Fig. S3). The open reading frame (ORF) length of MbAmh1 is 1458 bp (Fig. 4E), which consists of one exon encoding a protein of 485 amino acids. The apparent MW of His₆-tagged MbAmh1 obtained by SDS-PAGE was consistent with the predicted 55.4 kDa (Fig. 4F). MbAmh1 and its homolog in M. robertsii, namely MrAmh1, shared 98 % sequence identity. We speculate they should possess similar activity due to the high sequence identity. MbAmh1 contains a conserved amidohydrolase domain and lacks an N-terminal signaling peptide, corresponding to the non-degrading effect for extracellular enzyme fractions of Metarhizium spp.

Degrading activity of MbAmh1

MbAmh1 degraded OTA optimally at pH 4.0–7.0, and the reaction rate decreased more dramatically when the pH is lower than 4.0 or higher than 8.0 (Fig. 5A). At pH 4.0–7.0, 1 μg/mL of OTA was almost completely degraded to OTα after incubation with 2.5 μg/L of purified MbAmh1 for 40 min at 28 °C. However, none of OTA was degraded at pH 3 and 10. These data indicated acidic pH was more suitable than alkaline pH for the reaction between MbAmh1 and OTA. The optimal temperature for OTA hydrolysis was observed to be 30–70 °C (Fig. 5B). MbAmh1 could completely degrade 1 μ g/mL OTA within 10 min at pH 6 and 60 °C, and the degradation rate remained 94 % at 70 °C for 10 min. The thermal stability of MbAmh1 is a great advantage for potential applications in food and feed processing. The enzymatic activity of MbAmh1 was also influenced by metal ions (Fig. 5C). The presence of Li⁺, Zn²⁺, Mn²⁺ and Ca²⁺ at a concentration of 50 mM significantly increased the activity of MbAmh1 at 20 min after incubation, while Fe²⁺, Fe³⁺ and Cu²⁺ completely abolished the degradation activity of MbAmh1 under otherwise identical reaction conditions. Ni²⁺ significantly decreased the activity of MbAmh1. At the most optimal pH (6.0) and temperature (60 °C), MbAmh1, at the concentration of 1.0, 2.5 and 5.0 μg/mL, completely degraded OTA within 11, 5 and 3 min, respectively. The strong activity of MbAmh1 supported its primary roles in the degradation ability of Metarhizium (Fig. 5D).

Fig. 5.

Degradation activity of MbAmh1. The degradation rate of OTA was affected by pH (A), temperature (B) and metal irons (C), and asterisk indicated a significant difference (p < 0.01) between the corresponding value and control (Cont). (D) OTA degradation over time at the most optimal pH (pH 6) and temperature (60 °C).

The active sites of MbAmh1 for degrading OTA

A DALI search was conducted, using the MbAmh1 protein model predicted by AlphaFold, to identify structural homologues, revealing that it shares the closest structural similarity with ochratoxinase (PDB: 4C5Y, chain A), exhibiting a Z-score of 60.0 and RMSD of 1.3 Å [31]. This was followed by imidazolonepropionase (PDB: 2G3F, chain A) [42], which displayed a Z-score of 31.0 and RMSD of 2.7 Å, and molinate hydrolase [43] (PDB: 4UB9, chain A), showing a Z-score of 30.8 and RMSD of 3.2 Å. In a recent study, the CryoEM structure of ADH3 was determined (PDB: 8IHQ, chain A), showing a RMSD of 2.0 Å when compared to MbAmh1 [33], [44].

The structure of MbAmh1 showed the classic fold of the amidohydrolase superfamily [45]. It can be split into two regions including a main catalytic domain, and a smaller β-sandwich domain. The catalytic domain comprises residues 112–435 forming (β/α)₈-TIM barrel structural fold, a central twisted eight-stranded mixed β-sheet flanked by eight α-helices (Fig. 6A and B). The β-sandwich domain comprises residues 55–111 from N- terminus and 436–485 from C-terminus. A highly conserved binuclear metal center is deeply embedded within the putative substrate-binding pocket of the catalytic domain, suggesting that MbAmh1 likely employs the same catalytic mechanism as other homologous enzymes.[45]. By comparing with homologous structures and sequences of the known ones (Fig. S4), the α site of the MbAmh1 binds metal ion through amino acids His¹¹⁷, His¹¹⁹, Lys²⁵⁷ and Asp³⁸⁹, while β site is coordinated by His²⁹⁸, His³¹⁸ and Lys²⁵⁷ (Fig. 6C). Residue Asp³⁸⁹ is proposed to function as the catalytic residue for the cleavage of the carbon–nitrogen bond in OTA. The conformational changes of active site were primarily stabilized by Gly²⁶³, Ser²⁶⁶, Asp²⁶⁹ and Gln²⁷⁵, located in equivalent positions of ochratoxinase [31]. Point mutations of 15 candidate amino acid residues from the putative catalytic active center were performed to validate their significant roles. As shown in Fig. 6D, the results suggested His¹¹⁷, His¹¹⁹, Lys²⁵⁷, Gly²⁶³, Ser²⁶⁶, Gln²⁷⁵, His²⁹⁸, His³⁰⁰, His³¹⁸ and Asp³⁸⁹ on MbAmh1 were essential for OTA degradation. The mutation of those residues led to a complete loss of degrading function, while the mutation of His²⁰², Asp²⁶⁹, Val²⁶⁴, and Leu²⁶⁵ significantly reduce the degradation activity. Furthermore, random mutations of residues within the β-sandwich (K61A, D84A, L103A, D442A and G477A) revealed that only the mutation D442A significantly impacts the activity of MbAmh1 (Fig. 6D).

Fig. 6.

Structure and the critical amino acids from MbAmh1. (A) The MbAmh1 structure in cartoon representation. The β-sheets, α-helices and loops of the core TIM-barrel domain are colored red, cyan and magenta, respectively. The β-sandwich domain is colored yellow. The putative substrate-binding pocket of the catalytic domain is indicated by a black outline (B) The MbAmh1 structure in surface representation. (C) Enlarged display of OTA binding pocket and critical residues. (D) The degradation rate of MbAmh1 mutants. Asterisk indicate a significant difference between the corresponding values (p < 0.05) and MbAmh1. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Discussion

The variation of OT BGC in fungi

Previously, OTA was reported to be produced by Aspergillus and Penicillium genera from Eurotiomycetes. Here we provided the evidence that an OT-like cluster also existed in Metarhizium species from the distant fungal taxon Sordariomycetes, which demonstrated a discontinuous distribution of the OT BGC in fungi (Fig. 7A). There are many genetic variations in this cluster among fungi (Fig. 7A). The lack of otaD in Metarhizium could be due to loss of otaD during evolution under the assumption of a six-gene BGC originally existed in Metarhizium. Precisely determining the origin of OT BGC is challenging, given our limited knowledge of microbial genetic resources. Gene rearrangement in OT BGC occurred in A. albertensis and A. alliaceus from section Flavi. In addition, loss of OTA biosynthetic genes was found in Aspergillus species from section Circumdati [46] and Nigri [47]. Accruing data suggests that mycotoxins and other fungal SMs are adaptive factors that help fungi withstand abiotic or biotic stress [48], [49]. Whereas the role of OTs as ecological protectants or weapons of producing fungi is unclear, our data revealing variation in OT BGC makeup and product synthesis suggests an evolved niche for their production.

Fig. 7.

Schematic overview of variation of OT BGC and OTA degradation in fungi. (A) The variation of OT BGC. (B) Disappearence of OTA in fungal species.

The OTA degradation by ochratoxigenic fungi

OTA-degrading microbes have been screened from plenty of fungi and bacteria in previous studies [50]. Here, we find the degrading Metarhizium through screening the OT BGC at the genomic level in fungi. Interestingly, A. westerdijkiae could also weakly degrade OTA, while the most related homolog AoFC_06196 shared an identity of 28 % with MbAmh1. It was also reported OTA producing species A. niger and A. carbonarius had the abilitly to degrade OTA to OTα [51]. Through the fungal ochratoxinase-like model, Leitão and Enguita predicted that filamentous fungi, including OTA producer, are a rich source of hydrolases potentially capable of degrading OTs [32]. In addition, we previously observed the fluctuates in concentration of the mycotoxin aflatoxin in A. flavus [52], possibly indicating an endogenous aflatoxin degradation enzyme(s) in producing fungi that could be fruitful to examine for this mycotoxin. Taken together, aiming at the native producer provides a new thought for looking for mycotoxin-degrading enzymes.

The genus Metarhizium is an ubiquitous entomopathogenic fungal clade containing promising biocontrol agents [53], [54]. Although the OT BGC genes were expressed in Metarhizium, we could not detect either OTB or OTβ in the media tested in this study. It is possible that OTB is not ultimately biosynthesized. Drawing insights from the example of A. niger could shed some light on this possibility. In A. niger, the OTA-degrading enzyme ochratoxinase is produced, despite possessing the OT BGC and being a renowned OTA producer [55]. During the evolutionary process of A. niger, strains with partial deletions in the OT BGC, rendering them unable to produce OTA, have emerged. For example, A. niger strain ATCC1015 lacks 21 kb of the OT BGC, resulting in OTA synthesis deficiency [47]. Similarly, our observation indicates that the OT-like cluster lacks the otaD gene in Metarhizium spp., leading to the loss of OTA. It is plausible that Metarhizium spp. have evolved the degrading-enzyme MbAmh1 to efficiently degrade OTA, while gradually losing the capability to synthesize OTA during evolution (Fig. 7B).

The OTA degrading mechanism

While numerous microorganisms and crude enzymes have been identified to degrade OTA, relatively few investigations have focused on potential endogenous OT degradation mechanisms, as it is challenging to identify the specific enzyme(s) responsible from microbes. Traditional approaches for enzyme identification by precipitation and chromatography are time-consuming and variable. A rare success case was the isolation of ochratoxinase from A. niger [31]. 0.7 μg/mL ochratoxinase had decreased substrate concentrations from 50 to 25 ng/mLwithin 60 min. Additionally, carboxypeptidase PJ_1540 was also found through differential gene expression analysis in the presence or absence of OTA [26]. In this study, MbAmh1 was traced through bioactivity-guided proteomic analysis combined with in vitro reaction, revealing its significant activity and critical role in the degradation of M. robertsii. Thus, multiple methods, including separation techniques, bioinformatics, gene expression and others, should be combined for the effective degrading enzymes discovery.

There is currently limited research on the structure of OTA detoxifying enzymes and their catalytic mechanisms. The crystal structure of ochratoxinase [31] and Cryo-EM structure of ADH3 [44] have been determined and the catalytic mechanism has been discussed. The carboxylated Lys²¹⁰ of ADH3 [44] and Lys²⁴⁶ of ochratoxinase [31] that function as a bridging ligand between two metal ions reveal that they would belong to the subtype Ⅱ of amidohydrolases. MbAmh1 possesses a Lys²⁵⁷ in an equivalent position, that most probably should be carboxylated. And mutations of K257A led to a complete loss of degrading function (Fig. 6C and D). Deduced from the sequence alignment of homlogs, the metal ion at α site of MbAmh1 is coordinated by residues His¹¹⁷/His¹¹⁹/carboxylated Lys²⁵⁷/Asp³⁸⁹, while β site is coordinated by residues His²⁹⁸/His³¹⁸/carboxylated Lys²⁵⁷ (Fig. 6C). The function of residue Asp³⁸⁹ in MbAmh1 is to serve as the catalytic residue that coordinates substrate protonation as the acid/base. Indeed, mutation of Asp³⁸⁹ to Ala completely abrogated MbAmh1 activity. In addition to residues involved in metal ion coordination, His²⁰², Val²⁶⁴, Leu²⁶⁵ and His³⁰⁰ also participate in the formation of the OTA binding pocket [44].

The binuclear metal center is essential for the overall catalytic activity. The OTA carbonyl group is polarized by Lewis acid catalysis through complexation with the β metal ion, leading to activation of the hydrolytic water molecule for nucleophilic attack through interaction with the α metal ion. Typically, two Zn²⁺ ions occupy the active site of the subtype Ⅱ amidohydrolases [45]. Addition of Zn²⁺, Mn²⁺, Ca²⁺, and Li⁺ in the reaction system can enhance the degradation activity of MbAmh1, possibly due to the coordination of these metal ions with His residues in the OTA binding pocket, promoting the activation of the scissile bond and deprotonation of the hydrolytic water molecule.

Structural comparison of MbAmh1 with ADH3 and ochratoxinase, along with our results from the mutation of some key sites, could suggest a catalytic pathway of MbAmh1. At first, the metal β binds with the carbonyl oxygen of the amide group in OTA, inducing polarization of the carbonyl group to facilitate nucleophilic attack. Following this, the Asp³⁸⁹ residue's side chain extracts a proton from a bridging hydroxide that is expected to coordinate with the two metal ions. Lastly, the deprotonated bridging hydroxide initiates the cleavage of the carbon–nitrogen bond by attacking it, with a proton being donated by the protonated Asp³⁸⁹.

Industrial application potential of MbAmh1

Biological detoxification has emerged as a focal point of research due to its non-destructive nature towards nutrients, high efficacy, low toxicity, and remarkable specificity. The development of highly efficient mycotoxin-degrading enzymes is of paramount importance in the food and feed industry. Through our study, we found 5 μg/mL of MbAmh1 could completely degrade 1 μg/mL of OTA to OTα within 3 min. In contrast, the first commercial CPA of 67 μg/mL could only degrade 5 % of OTA after 1 h of incubation with 50 ng/mL OTA, while the more effective ochratoxinase degraded 50 % of OTA compare to CPA [28], [31]. A carboxypeptidase from Bacillus amyloliquefaciens could degrade OTA by 23.9 % after a 12 h incubation, but the degradation rate was relatively low compared to the host-degrading strain [23]. Recombinant DacA in E. coli could degrade 71.3 % OTA to OTα [24] and the degradation ratio for rCP4 was only 36.8 % within 24 h [27]. More recently, an effective hydrolase ADH3 was identified from Stenotrophomonas acidaminiphila, this enzyme was able to completely degrade OTA of 50 μg/mL within 90 s [33]. In conclusion, MbAmh1 demonstrated a stronger degrading activity than most of the enzymes previously found.

It is anticipated that mycotoxin-degrading enzymes will soon be commercially developed and utilized in the food and feed industries. The potential application of these degrading enzymes encompass their utilization in corn bran, corn starch, Distillers Dried Grains with Solubles (DDGS), oil and juice, and even as feed additives to facilitate degradation in the acidic stomach environment [56], [57]. The catalytic activity of degrading enzymes is significantly affected by the complicated application environment and thereby novel catalysts with robustness at different temperatures and pH changes are needed to fulfill the industrial requirements. Generally, natural enzymes rarely have the capacity to withstand high temperatures, which is significantly requested for industrial production [58]. MbAmh1 exhibits remarkable thermal stability across a broad temperature ranges, from 30 to 70 °C, with OTA completely degraded within 20 min at 70 °C. This feature suggests that its detoxification applications for food and feed may not be compromised. Compared to other degrading enzymes such as rCP4 [27] and ADH3 [33], MbAmh1 displayed excellent thermal stability and acidic pH adaptability. In order for an enzyme to be utilized for the degradation of OTA in food and feed, it is essential that no alterations occur to the food matrix, a requirement that can be minimized through the use of highly specific enzymes. It has been reported that members of the binuclear-metallohydrolase family exhibit strong substrate specificity, with loops containing specific residue compositions covering the active site surface, thereby determining the enzyme's specific substrate preference [32], [59], [60]. While we have gained a preliminary understanding of the enzymatic characteristics and structure of MbAmh1, further investigation into its molecular mechanisms of catalyzing OTA/OTB degradation is necessary. This research serves as the foundation for the rational design and molecular modification of the enzyme to meet the requirements of industrial applications.

Conclusions

Overall, our findings imply the distribution of an OT BGC in fungi, and some Metarhizium contain the cluster beyond the known species in Aspergillus and Penicillium. Furthermore, the crude protein extracts of M. brunneum and M. robertsii were found to strongly degrade OTA/OTB and the final degradation product was non-toxic OTα/OTβ, and we traced an amidohydrolase MbAmh1 as the critical working enzyme. MbAmh1 had a strong ability to degrade OTA, and showed considerable temperature adaptability ranging from 30 to 70 °C and acidic pH stability ranging from 4.0 to 7.0. Furthermore, some critical amino acids from catalytic center of MbAmh1 were identified. Collectively, MbAmh1 showed great potential in detoxifying OTA for industrial application, and these results provided new strategies to identify mycotoxin-degrading enzymes. Further research will prioritize structural analysis and molecular modification of MbAmh1, along with evaluating its safety, effectiveness and production cost under industrial conditions.

Compliance with ethics requirements

This article does not contain any studies with human or animal subjects.

CRediT authorship contribution statement

Gang Wang: Conceptualization, Methodology, Formal analysis, Writing – original draft, Funding acquisition. Wenqing Wu: Methodology, Formal analysis. Nancy P. Keller: Conceptualization, Writing – review & editing. Xu Guo: Methodology, Validation, Formal analysis. Erfeng Li: Methodology, Validation. Junning Ma: Software, Validation, Visualization. Fuguo Xing: Conceptualization, Writing – review & editing, Supervision, Funding acquisition.

Appendix A. Supplementary material

The following are the Supplementary data to this article: