A twist in the protein folding dogma: Probing how the cellular translation machinery can affect protein folding

One of the most fundamental assumptions in biology is that the amino acid sequence defines protein structure, at least for globular proteins, and that this sequence carries no memory of the translation process or specific mRNA codon sequence from which it was translated. Challenging this assumption, we have recently shown, using codon-specific Ramachandran plots, that some synonymous codons (those translating into the same amino acid) are associated with different protein backbone dihedral angles around the encoded amino acid (Rosenberg, A., Marx, A., & Bronstein, A.M, 2022, Nature Communications).

To identify specific positions in proteins which appear sensitive to translation we have screened high resolution structures in the PDB for pairs of homologous proteins harboring identical amino acid sequences in equivalent positions, having the same potential interaction network and yet adopting very different backbone conformations. We have shown that these alternate conformations remain stable in molecular dynamics simulations and that alphafold predictions do not capture this phenomenon. That is, alphafold predictions return models where the region has the same conformation in both of the homologous proteins. We now plan to target these regions for crystallographic experiments, altering the protein (coding) sequence at different local and remote positions and assess which of the conformations are formed.