Systematic Evaluation of Soluble Protein Expression Using a Fluorescent Unnatural Amino Acid Reveals No Reliable Predictors of Tolerability

Improvements in genetic code expansion have made preparing proteins with diverse functional groups almost routine. Nonetheless, unnatural amino acids (Uaas) pose theoretical burdens on protein solubility, and determinants of position-specific tolerability to Uaas remain underexplored. To broadly examine associations, we systematically assessed the effect of substituting the fluorescent Uaa, acridonylalanine, at more than 50 chemically, evolutionarily, and structurally diverse residues in two bacterial proteins: LexA and RecA. Surprisingly, properties that ostensibly contribute to Uaa tolerabilitysuch as conservation, hydrophobicity, or accessibilitydemonstrated no consistent correlations with resulting protein solubility. Instead, solubility is closely dependent on the location of the substitution within the overall tertiary structure, suggesting that intrinsic properties of protein domains, and not individual positions, are stronger determinants of Uaa tolerability. Consequently, those who seek to install Uaas in new target proteins should consider broadening, rather than narrowing, the types of residues screened for Uaa incorporation.