De novo protein design: towards fully automated sequence selection

Area of Science:

• Biochemistry
• Computational Biology
• Structural Biology

Background:

• Systematic, quantitative methods are being developed for general protein design algorithms.
• Existing methods often focus on specific protein parts or folds.
• There is a need for objective, quantitative design algorithms based on physical principles.

Purpose of the Study:

• To expand computational protein design by developing quantitative methods for all protein residue types (core, surface, boundary).
• To create an objective, quantitative design algorithm applicable to various protein folds and motifs.
• To test the design methodology using the betabetaalpha motif, exemplified by zinc finger DNA-binding domains.

Main Methods:

• Utilized previously published sequence scoring functions from combined experimental and computational approaches.
• Employed the Dead-End Elimination theorem to efficiently search for optimal protein sequences.
• Applied these methods to design a betabetaalpha motif, targeting 20 out of 28 positions.

Main Results:

• Successfully designed a novel peptide sequence (pda8d) with <40% homology to known sequences, lacking metal-binding sites or cysteine residues.
• The designed peptide (pda8d) exhibited high solubility, monomeric behavior, and confirmed folding into the target betabetaalpha motif via circular dichroism and NMR spectroscopy.
• Structural analysis revealed well-defined secondary and tertiary structures, with excellent agreement between the designed and target structures (backbone RMSD of 0.55 Å, atomic RMSD of 1.04 Å).

Conclusions:

• The developed quantitative design methodology is effective for creating novel protein structures with specific motifs.
• The designed peptide pda8d serves as a validated example of successful de novo protein design.
• This approach holds promise for developing general algorithms in computational protein design and protein engineering.