EpCAM and Trop2 are paralogous homodimeric transmembrane proteins that were first discovered as an antigen on the surface of colorectal cancer cells and an antigen on the surface of trophoblast cells, respectively. EpCAM is present in all vertebrates, while Trop2 arose by retrotransposition of EPCAM and is present in reptiles, birds and mammals. They are expressed on epithelial cells and participate in epithelial development, regulate cell adhesion and participate in proliferative signalling. Using Alphafold Multimer, we predicted the structures of EpCAM dimers from house mouse, chicken, African clawed frog and zebrafish and performed long molecular dynamics simulations of their extracellular parts, together with the simulation of the crystal structure of the human extracellular part of EpCAM. Using Alphafold Multimer we also predicted the structures of human EpCAM homodimer, Trop2 homodimer and EpCAM:Trop2 heterodimer. We performed short molecular dynamics simulations of extracellular parts of these structures, alongside simulations of crystal structures of extracellular parts of human EpCAM and Trop2 homodimers, and calculated the free energy difference upon dimer formation for these structures using MM/GBSA. We determined that the N-terminal domain of African clawed frog EpCAM exhibits greater flexibility compared to EpCAMs from other species. Dimer interface was smaller in African clawed frog EpCAM compared to others, while thyroglobulin loop in zebrafish EpCAM formed smallest interface with the opposite subunit. Interaction between subunits is the strongest in chicken and the weakest in zebrafish. Hydrophobic binding pocket in zebrafish EpCAM is smaller than others, while in African clawed frog it is more rigid – both differences could result in different binding properties. We also discovered that the subunits in the putative heterodimer are more strongly bound than in EpCAM homodimer and approximately as strong as in Trop2 homodimer. Consequently, heterodimer formation could be thermodynamically favourable. In the laboratory, we didn’t observe the heterodimer using native electrophoresis and crosslinking, but we couldn’t conclusively exclude the possibility of heterodimer formation due to the limitations of our approach.
|
