Engineering a protein-protein interface using a computationally designed library.

Area of Science:

• Protein engineering
• Computational biology
• Biochemistry

Background:

• Protein design algorithms face limitations in sampling conformational and sequence space.
• Experimental screening of protein libraries offers broad sequence exploration but lacks computational guidance.
• Integrating computational and experimental methods can overcome individual approach limitations.

Purpose of the Study:

• To combine computational protein design with experimental screening to create a variant of ubiquitin-ligase E6AP that binds the nonnatural partner Ubc12.
• To evaluate the efficacy of naive, semidirected, and directed libraries in identifying high-affinity binders.
• To determine if computational screening for well-folded sequences is sufficient for successful protein design.

Main Methods:

• Construction of three E6AP mutant libraries: naive, semidirected (disallowing destabilizing mutations), and directed (using docking and sequence optimization).
• Screening of libraries using a split-dihydrofolate reductase complementation assay.
• Characterization of binding affinity (Kd) for selected binders.

Main Results:

• Directed libraries showed >30-fold enrichment over the naive library after one screening round.
• Multiple tight binders (Kd < 100 nM) were identified after four rounds of selection from directed libraries.
• Naive libraries failed to yield binders with Kd < 50 μM even after four rounds of selection.

Conclusions:

• Computational protein design simulations can generate directed libraries enriched in tight binders.
• Combining computational design with experimental screening is an effective strategy for protein engineering.
• In some cases, computationally screening for well-folded sequences is sufficient for successful protein design without explicit binding calculations.