Acetyl‐CoA Electron Density: Acyl‐CoA Reactivity and Crystallographic Data

Abstract

Acetyl‐CoA and other acyl‐CoAs are the building blocks used in carbon‐carbon bond forming reactions underlying fatty acid and polyketide biosynthesis and many other biological activities. Since many enzymes use acetyl‐CoA, understanding how the substrate binds in the active site is key to determining interactions leading to enzymatic catalysis and substrate selectivity. We examined all of the PDB deposited crystal structures that report acetyl‐CoA bound and report stats for the reactive acetyl‐thioester moiety. We quantified this data by generating the real‐space R‐factor (RSR) and real‐space correlation coefficient (RSCC) by running a few rounds of refinement with Refmac to normalize the data. We examined the modeled thioester ends in comparison to the rest of CoA by creating a Linux script that splits acetyl‐CoA into the nucleic acid, pantetheine, and acetyl‐thioester moieties. We also analyzed the 3’phosphate of acetyl‐CoA to examine how splitting the molecule affects the RSR/RSCC values of small portions of the molecules along with the ribonucleotide containing substrates NAD+, NADH, NADP, and NADPH. Preliminary analysis of the output data reveals the nucleic acid and pantetheine RSR/RSCC values indicate good fits of models to electron density, but there is a portion of structures where RSR/RSCC values indicate the acetyl‐thioester moiety is poorly fit. We also find that the RSR method of calculation shows a significant difference between the acetyl‐CoA thioester and the 3’phosphate where the 3’phosphate is better than the average of the entire structure while the thioester is poorly fit. This is also consistent with the RSR values of NADP and NADPH 3’ phosphates. Conversely, the RSCC values of the acetyl‐CoA thioester end and the 3’phosphate were both equally lower than the entire structure. It appears there is an overfitting of the acetyl‐thioester fitting in crystal structures, likely due to a pressure to claim mechanistic inference from the structures.