You are here

De Novo backbone generation vs Rosetta Match for enzdes and ligand binding.

3 posts / 0 new
Last post
De Novo backbone generation vs Rosetta Match for enzdes and ligand binding.
#1

Howdy Yall, 

So I have not been in the Rosetta world for very long, and most of what I have learned I have had to parse from the Rosetta Commons Demos/Documenation and various publications since no one else at my institution does this sort of work. So while I have learned quite a bit, there is some fundamental questions about approach/methodology that I feel blind towards.

One of the biggest questions I have is about the design of enzyme active sites/ligand binding pockets and their relationship to the underlying backbone. In virtually every publication I have read, the primary way this is done is the creation of a theozyme/theoretical minimal binding interactions, and then using Rosetta Match to place these sidechains onto an existing PDB. 

My question is this: can one work without Rosetta Match? Can one design a theozyme or binding pocket, and then de novo design a backbone that can accomodate the necessary interactions? What is the primary benefit to the traditional protocol of matching to an exsting PDB? It would appear at first that designing a completely novel backbone to accomodate the active site might allow more freedom to the designer.

I know this question is more theoretical than the typical question about the software itself, but any guidance on this would help me to develop my understanding of protein design as a whole.

Thanks,

Cooper

Category: 
Post Situation: 
Thu, 2022-06-30 09:55
csvajda

You're right that desigining a de novo backbone around a theozyme is a "holy grail" of enzyme design. However, we aren't there quite yet. I think there's some preliminary work in the Baker Lab with respect to taking some of the new machine learing approaches for backbone generation/protein hallucination and adding in functionality constraints like theozyme-based geometries (see https://pubmed.ncbi.nlm.nih.gov/34853475/ and https://pubmed.ncbi.nlm.nih.gov/35332283/ for related work), but I'm unaware of a solid success in that area for enzyme catalysis.

So the traditional approach (matching a theozyme to a pre-generated backbone) is still the prefered approach for de novo enzyme design. But note that you don't have to limit yourself to natural protein scaffolds. The Baker lab has some work where they've done matching against de novo designed protein. The proteins weren't designed with the theozyme in mind, but they were designed with the concept of having pockets which would be ameanable to fitting in a ligand.

Thu, 2022-06-30 12:45
rmoretti

Thank you for your response! I will for sure look at those publications and try to learn more

Thu, 2022-06-30 16:26
csvajda