Paxlovid works by targeting a viral enzyme known as the main protease, or M^pro. This enzyme acts like molecular scissors, cutting long viral proteins into smaller pieces the virus needs in order to replicate. If M^pro is blocked, the virus cannot properly process its proteins and replication slows dramatically.

In the first study, the team examined exactly what kinds of protein sequences M^pro prefers to cut. Using high-resolution profiling, they mapped the enzyme’s preferred “substrate sequence” in exceptional detail. These experiments were made possible by a patented platform developed in Iverson’s lab called YESS 2.0.

Highly engineered yeast become fluorescently labeled to indicate how efficiently the protease cuts different sequences. Combined with flow cytometry and next-generation sequencing, this allows millions of protease–substrate interactions to be measured quantitatively in a single experiment.

What they found was striking: M^pro is highly selective. It recognises a very specific sequence pattern that is deeply embedded in the virus’s biology. That specificity creates a problem for the virus. If it mutates the enzyme too much in order to avoid a drug, it risks losing its ability to cut its own proteins correctly. In other words, the enzyme cannot easily change without breaking itself.

A few months later, the researchers followed up by looking directly for resistance mutations after years of Paxlovid use. They examined mutated M^pro sequences and tested mutations that might reduce the drug’s effectiveness while still preserving enzyme function.

They did not find strong, widely spreading resistance mutations. Some changes could slightly reduce drug sensitivity in laboratory experiments, but these often weakened the enzyme’s normal activity. The virus appeared to pay a biological cost for trying to escape.

The two studies fit neatly together. The strict substrate preferences identified in the first study help explain why major resistance has not emerged in the second.

Paxlovid closely mimics the natural sequence that M^pro normally cuts. Because of this, any mutation that prevents the drug from binding would also interfere with the enzyme’s normal function. The enzyme’s structure and function leave limited evolutionary room for manoeuvre.

Resistance is always possible, and continued monitoring remains essential. But the absence of significant resistance after years of global use is encouraging. It suggests that Paxlovid targets a viral weak point that the virus cannot easily redesign – a rare and welcome piece of good news in antiviral medicine.