Metagenomic-based bioprospecting for novel mycotoxin-degrading enzymes to keep food and feed safe

Many species of fungi capable of growing in grain and animal feed produce secondary metabolites called mycotoxins.

Mycotoxins, such as fumonisin FB₁ (FB₁) and T-2 toxin (Figure 1), are the biggest challenge for animal feed producers and therefore, regular monitoring and remediation is necessary, as these toxins can be very detrimental to humans and animals.

The economic loss attributed to mycotoxins in the agri-food sector is estimated to be billions of dollars, so the economic potential of protection or mycotoxin removal measures is correspondingly high.

There are already different measures available today, such as mechanical, thermic, or chemical approaches.

However, these approaches have, as an associated disadvantage, the possibility of generating unknown or unintended side products that may turn out to be as dangerous as the original toxin.

Enzymes are proteins that catalyze biochemical reactions that have been making life possible since 3.8 billion years ago.

Scientists have been able to “domesticate” them in such a way that enzymes are now key players for the circular economy and for the processing, manufacturing, and consuming everyday products, both for humans and animals².

Repowering the agri-food industry with enzymes capable of efficiently removing mycotoxins can contribute to keeping our food and feed safe and reducing carryover of mycotoxins from animals to humans.

However, the option to find such new enzymes, particularly those to degrade mycotoxins such as FB₁ and T-2 toxin is challenging, costly, and time-consuming.

Indeed, only a few enzymes are known that degrade both toxins¹.

In this contribution, we provide a comprehensive overview of our synergetic platform to better bio-prospect environmental metagenomes for such novel rare enzymes.

We further provide experimental evidence that demonstrates the efficient degradation of FB₁ and T-2 toxin by two novel hydrolases retrieved through naïve and bioinformatic screens.

The option to find new enzymes, although challenging, costly (€30k/enzyme), and time-consuming (15 months/enzyme), it is realistic given the recent technological advances.

To fully exploit the potential of sequencing data and as a source of new enzymes it is essential to combine:

• ⇰ Bioinformatics
  
• ⇰ Machine-learning
  
• ⇰ Accurate protein structure prediction
  
• ⇰ Data-driven artificial intelligence techniques
  
• ⇰ Naïve screens with appropriate substrates

In order to access new enzymes capable of degrading mycotoxins, we have established a comprehensive and synergetic platform to better bio-prospect environmental metagenomes for novel rare enzymes (Figure 2).

It consists in the following steps:

• 1. Extraction and sequencing of DNA from microbial communities (the so-called metagenome) of a large number of environmental samples.
• 2. Bioinformatic processing of the sequences representing the metagenomes to filter for those enzymes in the metagenome that have the potential to degrade FB₁ and T-2 toxin, namely, hydrolases.
• 3. Computational docking analysis to further provided computational quantum/molecular mechanics (QM/MM) analysis to identify, by quantifying the catalytic events, which of the selected enzymes can bind and potentially degrade FB₁ and T-2 toxin.
• 4. Finally, the experimental validation: gene synthesis, production, purification, and characterization.

Through the application of our synergetic platform, we have successfully bio-prospected environmental metagenomes and found several FB₁ and T-2 toxin degrading hydrolases, to highlight:

• ✓ A family VIII esterase, isolated from the metagenomic DNA of microbial communities inhabiting the seashore area of Milazzo Harbor (Sicily, Italy), capable of degrading T-2 toxin (Figure 3) without producing intermediate HT-2 toxin¹.
• ✓ An esterase capable of degrading 2-10 ppm FB₁ (100% degradation) after a 5-minute reaction time at concentrations as low as 5 μg/ml at 37-90 °C3.

Authors