There is a reasonable experiment you can do: Take the unit cell of your (presumed) crystal complex and put the search model into the until cell in a way that seems reasonable (use coot, or your favorite program). Make structure factors (with CCP4 or your favorite program) and throw away the phases. Then rotate your MR search model randomly and put the search model and structure factors into the MR program. See if it finds a solution. You can repeat it with structure factor data to which noise has been added. You can use this method to optimize your MR search parameters.
Don't be surprised if you cannot find a MR solution at all (even if you know for sure it should exist in your artificial problem). Once upon a time I tried to solve the structure of a small poly-peptide and with the method above I proved to myself that it was not possible to find a solution (with the programs that were available at that time, when Dinosaurs still roamed the Earth). I found that MR is very finicky when applied to small peptides. At least you will be able to determine what the optimal parameters are (resolution, search radius) and whether it can succeed at all. To pursue Se-Met is smart. Small peptide crystals resist heavy atom soaking. If your resolution is high enough, you can also try direct methods. On such a small peptide that should be easy (provided your crystals are well-behaved). Frequently the diffraction resolution of small peptide crystals is high enough that direct methods work very well. Mark -----Original Message----- From: Sean Seaver <[email protected]> To: [email protected] Sent: Tue, Nov 24, 2009 5:13 pm Subject: Re: [ccp4bb] Methylation of macromolecular complexes >I have a small complex, one component is 13 kDa with structure available and the other is 7 kDa, which could not be able to grew crystals after lots of efforts. It grew crystals after methylation and diffracted to 2.3 A, however, I could not be able to solve the structure using the structure of the big component as a template for molecular replacement, and heavy atoms soaking was not successful. I plan to do selenomethionine expression next. Does the methylation change protein structures a lot? otherwise, why does molecular replacement not work? I would much appreciate any idea and suggestions how to solve the structure using the data and the template avaliable. --- A couple of questions that I would consider in regards to the molecular replacement not working: Is the binding between the proteins 1:1? Does the Matthews coefficient reflect this? MR maybe difficult if you have a number of 7 kDa proteins binding to a 13 kDa protein. Hope that Helps, Sean
