Bhargav, P. | Search bioRxiv

bioRxiv2025

Engineering Escherichia coli FAD synthetase for the biosynthesis of FAD nucleotide analogues results in aminoglycoside antibiotic tolerance

Cofactors such as flavin adenine dinucleotide (FAD), nicotinamide adenine dinucleotide, coenzyme A, and S-adenosylmethionine cofactors share a common adenosine moiety that is generally not considered to contribute directly to catalysis. In FAD biosynthesis, the adenosine group is installed by the flavin mononucleotide adenylyltransferase (FMNAT) domain of the bifunctional FAD synthetase using FMN and ATP. This study explores the nucleotide selectivity of the Escherichia coli FMNAT domain and physiological effects of FAD nucleotide analogues synthesized with cellular nucleotides other than ATP. First, we engineer the FAD synthetase enzyme at an active site loop to produce FAD nucleotide analogues. Chromosomal substitution of the wild-type E. coli FAD synthetase with the mutated version results in the intracellular synthesis of these noncanonical FAD nucleotide analogues and confers tolerance to aminoglycoside antibiotics. Upon investigating their potential downstream metabolic roles, we find that their biosynthesis perturbs cellular metabolism and they could bind to a range of E. coli enzymes. These findings validate the FMNAT domain loop as a viable site for cofactor engineering, with potential for synthetic biology applications. Finally, the observed correlation between FAD analogue production and increased tolerance to aminoglycosides reveals novel metabolic mechanisms underlying antibiotic sensitivity.

microbiology↗

bioRxiv2024

AlphaMut: a deep reinforcement learning model to suggest helix-disrupting mutations

1Helices are important secondary structural motifs within proteins and are pivotal in numerous physiological processes. While amino acids (AA) such as alanine and leucine are known to promote helix formation, proline and glycine disfavor it. Helical structure formation, however, also depends on its environment, and hence, prior prediction of a mutational effect on a helical structure is difficult. Here, we employ a reinforcement learning algorithm to develop a predictive model for helix-disrupting mutations. We start with a toy model consisting of helices with only 30 AA and train different models. Our results show that only a few mutations lead to a drastic disruption of the target helix. We further extend our approach to helices in proteins and validate the results using rigorous free energy calculations. Our strategy identifies amino acids crucial for maintaining structural integrity and predicts key mutations that could alter protein function. Through our work, we present a new use case for reinforcement learning in protein structure disruption.

bioinformatics↗