An enzyme reversed microplastic damage in human cells

A machine went looking through 246 million protein sequences for something nobody had thought to test.

The Giant Problem

Start with the volume. PET production runs past 82 million metric tons a year, going into bottles, packaging, textiles and a great deal else. Less than a tenth of all the plastic we consume gets recycled. Seventy-nine percent of it accumulates in landfills or is dumped into the environment, and somewhere between 4.8 and 12.7 million metric tons reach the ocean annually. On current estimates, making plastic will account for about a fifth of global petroleum use by 2050. Out there, it does not stay in one piece. It fragments into microplastics and then nanoplastics, and those get into people: through food, through drinking water, and through the air indoors, off the ordinary furnishings of a house. Then they stay put. There is no established way to clear microplastics out of human tissue, and the enzymes that break PET down have not worked efficiently at the temperature and chemistry of a living body. This is where this paper starts.

The Science

PET is a polymer: one chemical unit repeated thousands of times and strung end to end, like a chain of paper clips. An enzyme called a PETase pulls the links apart. The best-known one, IsPETase, works at moderate temperatures, but not efficiently. Rather than tinker with IsPETase, Yuxuan Wang, Yajie Wang and their colleagues went looking for a different starting point. They used deep learning to search 246 million protein sequences, a catalogue of proteins found in living things, the overwhelming majority of which nobody has tested for anything at all. The model was aimed at the stretches of that space PET research had never touched. It returned AhPETase. Less than half its sequence matches any known PET-degrading enzyme, which makes it a distant relative rather than another variation on a familiar theme. At 37 °C it broke PET down 7.76 times better than IsPETase, and outperformed all the typical PETases it was measured against. Then they engineered it. The improved variant, AhPETase-M1, kept working for more than 20 days under physiological conditions, meaning body temperature and body chemistry. Enzymes usually give out long before that. They also fed it real rubbish. Not clean laboratory film but microplastics ground from post-consumer PET, plastic that had already been used and thrown away. Loaded with the same amount of enzyme, AhPETase-M1 got through those particles 34 times faster than MG8, a plastic-degrading enzyme the human body already makes on its own, tested here in a lab-produced form called rMG8. Then the cells. Human lung cells and human colon cells, growing in dishes, were exposed to PET microplastics and damaged by them. Add the enzyme, and the damage was reversed. The authors describe this as the first proof of concept for detoxifying microplastics enzymatically, and proof of concept is exactly the right weight for it: two cell types in dishes, in a laboratory in Hangzhou.

Why You Should Care

The obvious reading is a pill that clears microplastics out of you, and that is NOT what this is. Nothing here went into an animal, let alone a person. There is no delivery route, no dose, and no evidence about what a working plastic-cutting enzyme would do to the tissue it was sitting in. What the work does establish is narrower. PET can be taken apart efficiently at body temperature, which had not been shown well before, and doing so can undo measurable harm in human cells. That is a research direction rather than a treatment, and a treatment would be a decade away at best, if the other problems can be solved at all. The nearer uses are outside a body. An enzyme that runs at body temperature and keeps going for weeks is a useful thing to have wherever heating something up is the expensive part of the job. And the search method carries over. If pointing a model at 246 million sequences turns up an enzyme unlike the ones already known, the same trick is available for the plastics nobody can currently degrade at all.