I am working with an enzyme that uses Fe4S4 iron-sulfur clusters as co-factors. I recently built a model (using other software) and want to use ROSETTA for optimisation and enzyme design. I am very new to the software suite, and have managed to relax the protein together with the substrate ligand, for which I have a mol2 file. I would also like to include the iron-sulfur cluster, for which I have PDB coordinates, as well as AMBER PREP and FRCMOD files.
Is there a way of incorporating the cluster, and if so, how do I do this? The cluster is coordinated by cysteines - how do I tell the program to bond the CYS residues to the cluster?
Secondly, I have a second iron-sulfur cluster in the enzyme for which there is only PDB coordinates (no parameters). How would I go about incorporating this cluster?
I am very new to the ROSETTA suite, and any help is greatly appreciated.
Rosetta doesn't use AMBER topology files - it has it's own topology file format. You can (semi) automatically generate this file by using the molfile_to_params.py script, found at Rosetta/main/source/src/python/apps/public/molfile_to_params.py Run with just the "-h" option to get a description of all the paramers you can use. Basically, you give it a .mol2 or .mol/.sdf file of your ligand, and it will give you back a Rosetta .params file and a normalized .pdb file for the ligand. (The PDB file is useful as molfile_to_params.py will sometimes/often rename your atoms, and Rosetta uses atom names in the input PDBs to match up atoms in the ligand. It also requires the three letter residue name to match in the PDB file and the IO_STRING line of the params file.) You can then use the resulting params file with the option -extra_res_fa to get any Rosetta protocol to read in your new ligand.
The covalent connections are a little more complicated. What I would suggest is to use the enzdes-style constraint(*) file (https://www.rosettacommons.org/docs/latest/match-cstfile-format.html) to specify the covalent geometry. In the distanceAB line, the "periodicity" column serves double duty as a marker for covalent constraints. If set to 1 (as opposed to 0), the interaction is considered to be covalent, and extra code is run to make sure the 1-4 steric interactions aren't penalized, etc. Note that not all protocols will obey an enzdes-style constraint file. Enzyme design will, of course, as will RosettaScripts, if you use the appropriate mover to add the constraints (https://www.rosettacommons.org/docs/latest/Movers-RosettaScripts.html#Ad...).
For the other cluster, I'd recommend converting your PDB file to an sdf/mol or mol2 file with a small molecule structure editing program. I typically use the (freely available) programs OpenBabel and Avogadro, but pretty much any small molecule editing program should be able to convert your PDB into a mol/sdf or mol2 file. One caveat is that molfile_to_params will require a completely connected ligand, so you'll need to add bonds
A final caveat is that Rosetta hasn't been extensively profiled with modeling iron-sulfur clusters, so you'll want to examine the results carefully, and you may need to tweak some internal parameters to adjust how Rosetta models the clusters.
*) Rosetta nomenclature departs from standard MD ones here. When we say "constraint" in Rosetta, we're typically referring to what MD calls "restraints" - that is, the degrees of freedom can be sampled, but a scoring penalty is given to off-ideal values.