The Wittig reaction can be used for late-stage functionalization of proteins and peptides to ligate glycans, pharmacophores, and other natural or unnatural fragments. These fragments can provide additional functionalities that are not  presented by the naturally occurring amino acids. The ammino sequence is one of the determining factors for the efficiency of the Wittig reactions on peptides or proteins . Our group has previously employed the Wittig reaction to modify 160,000 N-terminal aldehydes in Genetically-Encoded (GE) peptide libraries. Previous studies revealed that the reaction rate strongly depends on the nature of the two N-terminal amino acids: specifically peptides with penultimate Pro-Pro exhibit a significantly decreased reaction rate. A plausible mechanistic hypothesis is that the hydrogen bonds in the peptide backbone stabilize the transition state (TS) of the Wittig reaction. DFT() calculation of the geometry of transition states in a model peptide CHO-Ala-Ala indeed detected a contribution of hydrogen bonding between backbone N-H and the betaine in the TS. We used deep-sequencing data from the modification of GE-libraries with Wittig reagents, to train two machine learning models that can predict the effects of peptide sequences on the rate of the Wittig reaction.

Learning Structure Activity Relationship (SAR) of the Wittig Reaction<br>
from Genetically-Encoded substrates

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.

Learning Structure Activity Relationship (SAR) of the Wittig Reaction from Genetically-Encoded substrates