Summary Points
Machine learning is enabling de novo protease design despite complex catalytic sites.
Two Nature studies report computational designs of cysteine proteases and metalloproteases using machine-learning tools.
Of 69 cysteine protease designs, 13 were active and cleaved their target peptides.
They formed an aspartate–histidine–cysteine triad, unlike natural folds, and functioned in mammalian cells.
Designing protein cutters from scratch
Proteases cut proteins into smaller pieces. Designing them from scratch is difficult because their active sites—the regions that drive the reaction—must work with precise chemical coordination. Now, machine-learning tools are helping researchers build these molecular cutters.
In one study, researchers created 69 designs for cysteine proteases. Thirteen cut their intended targets. Their active sites contained an aspartate–histidine–cysteine trio, and their overall shapes differed from those of natural proteases. Separately, another team reported designs for metalloproteases, which rely on metal ions.
From computer models to living cells
The cysteine protease designs were not just computer predictions. Structural tests showed that they formed the intended shapes. Further experiments found that they worked inside mammalian cells. That is an important step: an enzyme must function in a complex biological setting, not only in a test tube.
Still, early success is not the same as a ready-to-use treatment. Most designs did not prove active in the initial screen, and more testing is needed to assess reliability, safety and control.
A promising tool, not an instant cure
In time, designed proteases could help scientists study disease, develop new medicines or break down unwanted proteins. For everyday life, that might mean better ways to address some health problems. But those benefits remain possibilities, not current products.
The advance is meaningful for another reason. It suggests that researchers can create useful protein shapes beyond those found in nature. With careful testing, that ability could expand biology’s toolkit—and, eventually, improve quality of life.
Discover More Technology Insights
Explore the future of technology with our detailed insights on Artificial Intelligence.
Explore groundbreaking research and scientific advancements at the National Science Foundation (NSF).
Sci-BioV1
