Combinatorial assembly and design of enzymes

Area of Science:

• Protein engineering and computational biology
• Enzyme design and structure-function relationships
• Machine learning applications in biochemistry

Background:

• Enzyme structure and function are limited by complex long-range interactions required for proper folding.
• Designing novel enzymes with specific functions is challenging due to these folding constraints.

Purpose of the Study:

• To introduce a novel computational strategy, combinatorial assembly and design of enzymes (CADENZ), for designing structurally diverse enzymes.
• To generate stable, low-energy enzyme structures with functional catalytic sites.

Main Methods:

• Utilized an atomistic and machine learning approach for combinatorial enzyme fragment assembly.
• Applied the CADENZ strategy to design endoxylanases.
• Employed activity-based protein profiling to identify and recover functional enzyme designs.

Main Results:

• Successfully recovered thousands of structurally diverse endoxylanases using the CADENZ strategy.
• Identified that functional enzyme designs feature preorganized active sites and compact, stable non-active site packing.
• Achieved a 10-fold improvement in design success rate and recovered over 10,000 enzymes after iterative refinement of CADENZ.

Conclusions:

• The CADENZ strategy enables the combinatorial design of enzyme fragments for generating diverse, stable, and functional enzyme structures.
• Lessons learned from functional designs, such as active site preorganization and improved packing, enhance the CADENZ method.
• This iterative design-test-learn approach is broadly applicable to modular protein families for discovering novel protein functions and design principles.