Enzyme design

What is enzyme design?

Enzyme design is the process of creating or modifying proteins to catalyze a specified chemical transformation. Computational workflows represent catalytic residues, transition-state or substrate geometry, cofactors, and surrounding structural constraints, then generate scaffolds and sequences intended to support that arrangement. Success requires more than a folded model: catalytic rate, selectivity, stability, and the proposed reaction mechanism must be measured experimentally.

De novo enzyme design starts from reaction geometry rather than from an existing enzyme sequence. Other projects redesign a natural scaffold while preserving catalytic residues. In either case, ligand and cofactor coordinates must remain chemically meaningful throughout motif scaffolding and ligand-aware sequence design.

The computational funnel should reject candidates that distort catalytic geometry, bury unsatisfied polar atoms, clash with substrates, or fail to refold. Laboratory assays then distinguish true catalysis from background reaction, nonspecific binding, contamination, or an unintended mechanism.

When to use enzyme design

• Best fit. Creating catalytic scaffolds or redesigning activity, selectivity, and substrate scope
• Required starting evidence. Catalytic motif geometry, ligand or cofactor context, mechanistic assumptions, and a quantitative assay

Benefits of enzyme design

• Focused search. Targets explicit catalytic geometry
• Connected evidence. Connects structure and sequence design
• Testable candidates. Can explore new scaffold space

Primary limitations

• Model scope. Transition-state models may be wrong
• Score uncertainty. Folded designs may be inactive
• Experimental requirement. Catalytic validation is experimentally demanding

Enzyme design methods

Motif-scaffolding methods place catalytic atoms within generated protein geometry. The motif is only as meaningful as its chemical model, so protonation, cofactors, substrate pose, and geometric tolerances should be documented.

Ligand-aware sequence design assigns residues around both the backbone and non-protein atoms. It can preserve a pocket context but cannot prove that the designed electrostatics, dynamics, or reaction pathway will produce catalysis.

How to run enzyme design online

Use the workflow as an inspectable computational funnel. Preserve the native output of each method, apply explicit acceptance gates, and keep the evidence behind every selected and rejected candidate.

1. Define reaction. Specify the chemical transformation, catalytic residues, ligand or transition-state model, and assay controls.
2. Encode active site. Prepare a motif PDB whose atom names, bonds, cofactors, and HETATM context are correct.
3. Generate scaffolds. Generate scaffolds that preserve the required active-site geometry and reject distorted motifs.
4. Design sequences. Assign ligand-aware sequences, refold candidates, and inspect pockets, clashes, and stability.
5. Test catalysis. Express selected designs and measure rate, background, selectivity, and mechanism-relevant controls.

How to evaluate enzyme design results

Inspect motif RMSD, catalytic distances and angles, pocket accessibility, clashes, sequence confidence, refold agreement, and alternate ligand poses. Keep native enzyme and no-enzyme controls when available.

Experimental evaluation should report expression, folding, turnover, catalytic efficiency, substrate scope, and selectivity. Controls must rule out spontaneous reaction and adventitious catalysis.

Experimental validation and handoff

Preserve motif atoms, ligand/cofactor coordinates, geometric constraints, sequence-design settings, and rejected candidates. Validate activity with quantitative assays and controls.

Export structures, sequences, settings, scores, logs, and selection criteria together. A reproducible handoff makes computational assumptions visible to the team planning synthesis, expression, biophysical characterization, and functional assays.

How enzyme design works

Generate scaffolds around an active-site motif with RFdiffusion 2, assign ligand-aware sequences with LigandMPNN, and refold with ESMfold.

1. Define reaction. Specify the chemical transformation, catalytic residues, ligand or transition-state model, and assay controls.
2. Encode active site. Prepare a motif PDB whose atom names, bonds, cofactors, and HETATM context are correct.
3. Generate scaffolds. Generate scaffolds that preserve the required active-site geometry and reject distorted motifs.
4. Design sequences. Assign ligand-aware sequences, refold candidates, and inspect pockets, clashes, and stability.
5. Test catalysis. Express selected designs and measure rate, background, selectivity, and mechanism-relevant controls.

Inputs and outputs

Check formats before running, then inspect and download the result from every workflow step.

Inputs

• Design input. PDB FASTA JSON TXT An active-site motif PDB with catalytic residues and ligand or cofactor HETATM records.

Outputs

• Design and review outputs. PDB FASTA CSV JSON Scaffold PDB files, ligand-aware FASTA sequences, refolded models, design scores, and validation files.