I would like to know whether the enzyme design protocol of Rosetta Suite can be used for increasing the catalytic activity of the enzyme. Here, I don't have to design the active site from scratch, but I need to design the active site in such a way so that the catalytic activity of the enzyme is increased. Also, the enzyme I am dealing with does not have data for transition state modeling for the substrate that I am dealing with.
Please provide you comments and suggestions
In the standard de novo enzyme design procedure (as used in the Kemp eliminase, retro-aldolase and Diels-Alderase papers from the Baker lab) there's two distinct steps in Rosetta. The first is the matcher step, which finds positions and orientations of the catalytic residues and substrate on suitable protein scaffolds. The second is the design step, which redesigns the region surrounding the catalytic site to best accommodate the substrate and catalytic residues. Each of these steps can be used separately. In your case you already have an appropriate scaffold with properly positioned catalytic residues, so you can skip the matcher step and go straight to the redesign step, optimizing the surrounding residues for your substrate of interest. (Keep in mind, though, that evolution probably has done a decent job of optimizing the active site toward the wildtype substrates, so opportunities for improvement may be limited. Alternative substrates, or increasing specificity is something that has been done, though.)
Not having a transition state model is not too much of an issue. Rosetta's optimization function is not "quantum mechanically correct", so it's not overly sensitive to the sorts of things that QM modelers of transition states worry about. It's entirely possible to get good results by mocking up a reasonable transition state-like complex. You can even go further, if you like. For the retroaldolase, the reaction was multistep. In order to get decent results throughout the reaction, they actually made a "transition state analog" that was a hybrid of the theorized transition states and intermediate states. A quantum mechanically modeled transition state is not needed for Rosetta to suggest designs - a rough approximation often suffices.
I am still looking into the TS design part, however, I have a query regarding the docking of substrate into the active site. I have performed the docking of ligands into the active site using Rosetta docking and I want to know how to analyze the results particularly when we don't have native structure with a docked ligand to compare with. In my case, I have selected the top 10 lowest scoring structures to look for the interactions. I have done docking with Autodock parallely as well to check if I get some match between the scores.
This another question is also related to docking and design. I was wandering whether Rosetta provides flexibility of docking+design? For eg, If I want to mutate certain residues around to active to all 20 amino acids in order to design a more efficient binding site and do the docking simultaneously. I think this can be done and don't know how to club the two procedures together. Can you please help??
The evaluation for ligand docking normally uses multiple criteria. First you cut off a certain fraction of the docked conformations by total score. (Standard protocol is to leave only the best 5%.) This gets rid of the really bad structures, or those structures where the protein is contorted in order to bind the ligand. After that, you rank the docked conformations by ligand interaction energy. The lowest ligand interaction energy structure is the one which is probably the best representative of the docked conformation. See the RosettaLigand papers (Meiler Proteins. 2006 Nov 15;65(3):538-48.; Davis J Mol Biol. 2009 Jan 16;385(2):381-92; Lemmon Methods Mol Biol. 2012;819:143-55. as well as others) to see what techniques are commonly employed. You'll definitely want to look at the selected structure and see if it makes sense. Depending on your system, you may want to add other filters/cutoffs to get rid of system-specific issues.
Regarding design, that is easily accomplished with the RosettaScripts framework. Basically, you allow design instead of just repacking with your TaskOperations, or you add an extra PackRotamersMover step or even an EnzymeDesign step (see https://www.rosettacommons.org/docs/latest/ligand-centric-movers.html) which does design in the middle of the docking protocol. (e.g. dock, design, redock). The Meiler Lab tutorials have a docking+design example in some of the ligand docking tutorials. I'd go with the version found here: http://structbio.vanderbilt.edu/comp/workshops/rosetta_13/tutorials_2013...
For evaluating design, you'll want to use the same techniques as evaluating docking, but you'll want to add more filters. Hydrogen bonds, packing, repulsive interactions to the ligand, etc. The goal is not necessarily to get great values on all of these, but to get a good binding energy where none of the other metrics is too extreme - you're eliminating bad designs more so than identifying good ones.
Thank you for your response. This really solved my problem.