Predictive modelling of the COMATOSE transporter reveals a conserved ligand binding pocket for acyl-CoAs

The ABCD subfamily is comprised of mainly peroxisomal ABC transporter proteins which are found in most eukaryotes. They import a range of substrates of the peroxisomal β oxidation pathway. In plants there is a single peroxisomal ABC transporter AtABCD1 also known as COMATOSE (CTS), PXA1, PED3 or ACN1. Substrates are acyl CoAs and CTS possesses intrinsic acyl CoA thioesterase activity (ACOT), which has subsequently been demonstrated in human and yeast homologues. Even though there is no sequence similarity to any known thioesterases, the CoA moiety is cleaved at some point during transport and most likely released within the peroxisome. While there is a wealth of published biochemical and genetic data on CTS, the precise transport mechanism remains unknown. To gain insights into the CTS transport cycle we have utilised AlphaFold2/3 to produce high confidence models allowing us to explore the apo, post-translocation and post hydrolysis states of the protein, and used the available genetic and biochemical data to validate our models. Docking simulations of these models and C22:0-CoA/ CoA revealed key residues, which we propose are part of the putative binding pocket for the acyl CoAs. The D863/Q864/T867 (DQT) triad likely plays a structural role, while the proximity of S810 to the CoA-binding pocket suggests a potential involvement in substrate recognition or binding. Conservation scores and the alignments of CTS models with the cryoEM determined substrate-bound human homologue structures, supported the key interactors of the pocket which can inform future experiments to understand the transport mechanism of this important protein.