Researchers Develop Enzyme to Efficiently Break Down Polyurethane

A team of researchers has successfully developed a new enzyme capable of breaking down polyurethane, a widely used polymer found in products such as foam cushioning. This advancement is significant in addressing the complex challenges posed by plastic pollution, which is increasingly recognized as a multifaceted environmental crisis.

According to a recent study published in the journal Science, the production of polyurethane reached a staggering 22 million metric tons in 2024. Traditional methods for breaking down this polymer have proven inadequate due to its intricate chemical structure. The urethane bonds in polyurethanes comprise a nitrogen atom bonded to carbon, which in turn is linked to two oxygen atoms, creating a robust and complex arrangement that is difficult for enzymes to target.

Efforts to decompose polyurethanes have often relied on a chemical called diethylene glycol, which can partially break down these materials under elevated temperatures. Unfortunately, this process results in a byproduct that is often deemed hazardous and typically incinerated, rather than being converted back into usable materials. The research team aimed to find a more effective solution that could integrate with existing processes involving diethylene glycol.

To identify a suitable enzyme, the researchers initially tested 15 enzymes previously documented for their capacity to degrade polyurethanes. The results were discouraging, with only three demonstrating reasonable efficacy. The focus then shifted to the enzyme exhibiting the highest activity, leading the team to explore related proteins in public databases, including the AlphaFold database for protein structure predictions.

The research team employed a neural network called Pythia-Pocket to determine the likelihood of various amino acids in a protein binding to specific chemicals. They also utilized another neural network, simply named Pythia, which predicts the stability of protein structures. By combining these tools, the researchers aimed to design an enzyme that would maintain sufficient flexibility while retaining the necessary structural order for enzymatic activity.

This innovative approach led to the development of a software tool named GRASE, which stands for graph neural network-based recommendation of active and stable enzymes. The results were remarkable; of the 24 proteins evaluated, 21 exhibited catalytic activity, with eight outperforming the previously known best enzyme. The top-performing enzyme demonstrated an incredible 30 times the activity of its predecessor.

The enzyme’s effectiveness was further enhanced when combined with diethylene glycol and heated to 50° C. Under these conditions, it achieved over 450 times the activity of the best natural enzyme, successfully breaking down 98 percent of the polyurethane within a 12-hour timeframe. Notably, the enzyme maintained its stability, allowing for two additional uses before its effectiveness began to diminish.

When scaled up to kilogram-level tests, the enzyme continued to perform exceptionally, breaking down more than 95 percent of the polyurethane into its constituent materials. The researchers emphasize that their approach not only focuses on structural similarities but also incorporates functional aspects, such as stability and the amino acids likely to interact with the target material.

This breakthrough in enzyme design signifies a critical step towards more sustainable methods for managing polyurethane waste, with implications that extend beyond this specific polymer. The tools and methodologies developed in this research may pave the way for future innovations in enzyme engineering, potentially transforming how we approach plastic pollution on a global scale.