Structure-filtered search of enzyme variants

Area of Science:

• Biotechnology
• Computational Biology
• Enzymology

Background:

• Enzyme performance is crucial for applications but is governed by complex sequence-structure-function relationships.
• Limited insights into these relationships necessitate screening numerous enzyme variants, hindering industrial strain development for smaller labs.
• Current methods often rely heavily on sequence similarity, yielding vast, unmanageable candidate pools.

Purpose of the Study:

• To develop and demonstrate an in silico method for refining enzyme candidate sets for experimental screening.
• To leverage structural similarity filtering combined with sequence homology searches to identify functionally equivalent orthologs.
• To reduce the experimental workload in enzyme discovery and accelerate industrial strain development.

Main Methods:

• Implemented an in silico strategy combining homologous sequence search with structural similarity filtering.
• Utilized AlphaFold-predicted protein structure models for assessing structural similarity.
• Applied the method to identify variants of aspergillic acid synthetase (asaC) and chrysogine synthetase (chyA).

Main Results:

• Initial sequence similarity searches yielded tens of thousands of potential enzyme candidates.
• Structural similarity filtering and active site assessment drastically reduced candidates to 24 for asaC and 1 for chyA.
• The method effectively narrowed down the search space, demonstrating significant efficiency gains.

Conclusions:

• Filtering enzyme candidates by structural similarity is an efficient strategy to reduce experimental screening efforts.
• This in silico approach accelerates the identification of suitable enzyme variants for industrial applications.
• The method facilitates strain development, particularly for laboratories with limited resources compared to biofoundries.