Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsYes. A Nature study published on 14 October 2026 reports genuinely de novo cysteine proteases: proteins built with newly generated folds that hydrolyze peptide-backbone amide bonds in a sequence-dependent way. Of 69 designs tested, 13 were active (19%), the best accelerated the reaction by up to 3 × 107 over the uncatalyzed background, and the designed proteins cleaved intended targets in mammalian cells.
What the 2026 study achieved
De novo protein design means creating a catalytic protein without starting from a known natural protease scaffold. Choi, Bauer, Coventry and colleagues describe enzymes that place an activated cysteine nucleophile, its supporting catalytic residues and a substrate-binding environment inside computationally generated backbones. Their abstract summarizes the result as the de novo design of enzymes that use an activated cysteine nucleophile to hydrolyze the polypeptide backbone in a sequence-dependent manner.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Lippincott® Illustrated Reviews: Biochemistry (Lippincott Illustrated Reviews Series) | $80.48 | Buy on Amazon |
| 2 |
|
Lehninger Principles of Biochemistry | $259.99 | Buy on Amazon |
| 3 |
|
Biochemistry For Dummies | $23.67 | Buy on Amazon |
| 4 |
|
Biochemistry | $107.51 | Buy on Amazon |
| 5 |
|
Textbook of Biochemistry with Clinical Correlations | $116.05 | Buy on Amazon |
As an Amazon Associate I earn from qualifying purchases.
The result is significant because the proteins are not simply edited versions of papain-like or other familiar proteases. Their folds are very different from natural proteases, with TM scores below 0.50 in the authors’ comparison. That supports a stronger claim than successful protein remodeling: the computational method produced new catalytic architectures that function experimentally.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Headline measurements
| Measurement | Reported result | What it shows |
|---|---|---|
| Designs experimentally tested | 69 | The size of the tested design set |
| Active designs | 13 (19%) | Design is productive but far from guaranteed |
| Maximum rate enhancement | Up to 3 × 107 over the uncatalyzed background reaction | The best enzymes strongly accelerate peptide-bond hydrolysis |
| Crystal structures matching computational models | 6 structures; Cα RMSD below 1.2 Å | Several proteins adopted the intended three-dimensional geometry |
| Fold similarity to natural proteases | TM score below 0.50 | The active designs use substantially different folds |
These figures come from the Nature version-of-record study published 14 October 2026. The rate enhancement is measured against the uncatalyzed reaction, not against the fastest native cysteine proteases; the authors state that the best designs still lag the most active natural enzymes.
#1 Best Overall
How the de novo design pipeline works
The workflow combines backbone generation, sequence design, structure prediction and constrained relaxation. The objective is to make the catalytic geometry a property of a new scaffold rather than to search only among known protease families.
- Generate candidate backbones and active sites. RFD2-MI creates protein backbones and proposes active-site geometries that place an activated cysteine, the catalytic support system and a substrate-contacting region in a reactive arrangement.
- Design amino-acid sequences. ProteinMPNN assigns sequences compatible with each proposed backbone and catalytic environment.
- Preserve the intended chemistry during optimization. Rosetta FastRelax constraints keep the catalytic geometry from drifting while the sequence and structure are refined.
- Predict and filter structures. AlphaFold 3 (AF3) evaluates whether the designed sequence is likely to fold into the intended architecture. Designs that fail structural or active-site filters are discarded.
- Test survivors experimentally. The remaining proteins are produced and assayed for peptide-backbone cleavage, sequence dependence and cellular activity.
This iterative filtering matters because a sequence can look plausible in isolation yet fail to fold, misalign the catalytic residues or leave the substrate inaccessible. The study’s 13 active designs show that the pipeline can find working examples, while the 56 inactive designs show why computational selection cannot replace biochemical testing.
Why cysteine-protease chemistry is difficult
Peptide-backbone amide bonds are unusually resistant to hydrolysis in neutral water. A working protease must lower that barrier by bringing the substrate into a reactive pose and making a cysteine sulfur sufficiently nucleophilic to attack the amide carbonyl. In these designs, the active site preorganizes an activated cysteine and positions the accompanying catalytic residues, including the cysteine/histidine system identified by the authors’ analysis.
Rank #2
Preorganization is the central challenge. Small shifts in the cysteine, histidine or substrate can reduce productive collisions, alter proton transfer or allow water to outcompete the intended reaction. The designed proteins therefore need both correct local chemistry and a stable global fold that holds those groups in place.
How sequence specificity is created
The enzymes do not merely cut any exposed peptide bond. They recognize target sequences through a designed binding cleft that positions particular residues of the substrate next to the catalytic center. This follows the established principle that, in papain-like cysteine proteases, specificity is largely controlled by the structure of the substrate-binding cleft. In a de novo enzyme, the cleft must be invented along with the catalytic fold.
Sequence-dependent cleavage is consequently a two-part design problem: activate the cysteine and build a pocket that selects and orients the desired peptide. A highly reactive cysteine without a selective cleft would risk broad, uncontrolled cleavage; a selective pocket without precise catalytic geometry would bind substrate but react slowly.
Rank #3
How closely did the proteins match their computer models?
Six crystal structures had Cα RMSD values below 1.2 Å relative to their computational design models. That level of agreement indicates that the experimentally observed backbones were close to the intended coordinates, including the overall placement of the active site. It is an important validation of the design-and-prediction workflow because catalytic activity depends on three-dimensional positioning, not just on having the right amino-acid list.
Structural agreement does not mean every prediction was correct. The 13-of-69 activity rate shows that many computationally selected candidates still failed to produce measurable catalysis. Nor does a low RMSD by itself establish high catalytic efficiency: the authors attribute the remaining performance gap to imperfect active-site preorganization and substrate positioning.
Do the designed proteases work in living cells?
Yes, the study reports cleavage of intended targets in mammalian cells. That experiment demonstrates that the catalytic concept can operate in a cellular environment rather than only in a purified test tube. It is an important synthetic-biology milestone because cells introduce crowding, competing proteins, degradation pathways and different chemical conditions from a controlled biochemical assay.
Rank #4
Cellular cleavage is not evidence of a therapy. The study does not establish clinical efficacy, delivery to a particular tissue, long-term stability, immunological safety or selective activity in humans. Those questions would require separate animal and clinical investigations.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What still limits activity
The authors’ analysis points to two main bottlenecks:
- Active-site preorganization: the cysteine and histidine system is not held in the optimal geometry or electrostatic environment in every design.
- Substrate positioning: the binding cleft does not always keep the target peptide in a pose that favors attack and bond cleavage.
Future designs will therefore need to optimize catalytic geometry and electrostatics together with the shape, flexibility and contacts of the substrate cleft. Improving one without the other may increase binding while leaving the chemical step slow, or increase reactivity while sacrificing sequence selectivity.
Best Value
How to access the reported structures and sequences
Atomic coordinates and structure factors are deposited in the RCSB Protein Data Bank under accession codes 9YNL, 9YNM, 9YOX, 9YOY, 9YOZ and 9YP0. Protein sequences appear in Supplementary Table 1, and complete plasmid DNA sequences are supplied in the study’s Supplementary Information files. These resources allow researchers to inspect the experimental structures, reproduce expression constructs and compare measured proteins with their computational models.
What this means for computational enzyme design
The study establishes a practical proof of concept rather than a push-button recipe. Computational methods can generalize to catalytic folds that are unlike known proteases, and a measurable fraction of generated designs can work in vitro and in mammalian cells. The immediate benchmarks for progress are not just more active proteins, but higher activity among tested designs, closer matching of predicted and experimental active sites, sharper substrate-sequence selectivity, improved stability and reliable function in the intended cellular context.
For readers asking whether cysteine proteases can be designed from scratch, the evidence is now affirmative. The qualification is equally important: success remains probabilistic, the strongest synthetic enzymes are not yet as active as the best natural counterparts, and cellular activity is an early demonstration rather than a medical result.
Recommended Free Tools
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
