How to add zinc for autodockZn

Question

[Ria]

I am working with a metalloprotein. Model has been downloaded from alphafold database. I am going to do autodock Zn. But there is no zinc in the model. Which is the accurate way to add zinc? Can i transfer the zinc from another homologous PDB (cys-his-cys is conserved, i checked)? Or should i use tools like alphafill, metal3d, mib etc? Which would be accurate?

Answer 1

[Kevin Blighe]

Hey, good question. Since you're dealing with an AlphaFold model and the zinc-binding motif (Cys-His-Cys) is conserved, the most straightforward and accurate approach is to use AlphaFill. It's specifically built for this— it transplants metals like zinc from homologous experimental structures into AlphaFold models based on sequence and structure similarity. You can just search for your protein's UniProt ID on the AlphaFill site (alphafill.eu), and it'll give you a ready-to-go model with the zinc added. I've seen it work well for metalloproteins, though sometimes the exact bond distances might need a quick tweak in something like PyMOL if you're picky for docking.

If AlphaFill doesn't quite nail it or your protein isn't in their DB, transferring the zinc manually from a homologous PDB is a solid backup. Align your AlphaFold model to the homolog using TM-align or just superimpose the conserved residues in PyMOL, then copy over the zinc coordinates. Since the motif is the same, this should be pretty reliable without much hassle.

As for other tools like Metal3D or MIB, they're more for de novo prediction when you don't have homology info. Metal3D uses deep learning and gives good location predictions with confidence scores, but it's overkill here since you already have a homolog. MIB (or the updated MIB2) predicts binding sites and docks the metal, but again, homology-based methods will be more precise for your case.

For AutoDock Zn, once you have the zinc in place, make sure to parameterize it properly—AutoDock handles zinc with specific force field tweaks, but that's downstream.

Hope that helps—let me know if you run into issues.

[Ria]

Thank you so much. It helped a lot!