You are here

known disulfides in Remodel sequence

2 posts / 0 new
Last post
known disulfides in Remodel sequence

I have a protein structure solved by iTasser against a distant relative. But it has a loop that was not solved correctly and it has two cysteines that probably form a disulfide, so I want to remodel the loop with a disulfide (Rosetta Remodel 3.9 on Linux).

I want to keep the sequences as is, so the disulfide scanner is not for me.

I tried with a PDB template with the following lines:

SSBOND   1 CYS     89    CYS    110                          1555   1555  2.05

But that did nothing for the disulfide. Likewise using the 'in_fix_disulf'. I have no idea where I got the REMARK BONE MOTIF line, but another file of mine had it with cystines.

When I run the following (on a template without any headers):

$ROSETTA/remodel.$ROSETTAEXT -database $ROSETTADB -s PLPPR5.r.pdb -no_optH false -ex1 -ex2 -ex1aro -mute basic -in:fix_disulf ${task}.txt -num_trajectory 10 -nstruct 5 -suffix ${task}_R117C -remodel:blueprint cysloop_${task}.blu -enzdes:cstfile dis.cst -generic_aa C;

I get this error:

core.chemical.residue_support: [ WARNING ] Atom type 'SH1' on atom ' SG ' from residue type 'CYS:MP-SG-connect' does not have a centroid mode equivalent: assuming CAbb.
core.chemical.residue_support: [ WARNING ] Atom type 'HS' on atom ' HG ' from residue type 'CYS:MP-SG-connect' does not have a centroid mode equivalent: assuming CAbb.
core.chemical.residue_support: [ WARNING ] Cannot automatically convert CYS:MP-SG-connect to centroid, as it is polymeric/has connections.
core.util.switchresiduetypeset: [ ERROR ] can not find a residue type that matches the residue CYS:MP-SG-connect at position 89

The '-generic_aa C' flag is there because it cannot find atom SG on valine, the default modelling amino acid.

The constraint file looks right:

TEMPLATE:: ATOM_MAP: 1 atom_name: SG
TEMPLATE:: ATOM_MAP: 1 residue3: CYS

TEMPLATE:: ATOM_MAP: 2 atom_name: SG
TEMPLATE:: ATOM_MAP: 2 residue3: CYS

CONSTRAINT:: distanceAB: 2.10 1.00 50.00 1

When I change the 4th value from one (bound) to zero (unbound) it works. As the error message is to do with hydrogens, I grepped out all hydrogens (just in case the no_optH argument was not yet applied, but it happens nevertheless.

I have never managed to get a remodel run with NCAA to work in general, so I have not tried '89 D NC EMPTY CYX' in my blueprint, but that seems a very odd approach. I have not tried LINK line either. But before further embarking on odd tests, I thought I best ask it seems like something that should be simple...


Super-Thanks in advance!!

Post Situation: 
Mon, 2018-09-10 06:43

So what I found is that doing a non-covalent constraint, followed by (i) addition of SSBOND linkage in the PDB (ii) along with removal of 'HG   CYS' and (iii) addition of -in:fix_disulf xxx resolves the disulfides. Which is a tad inelegant. I have not tried each of the parts alone, but I suspect that the grep -v of the gamma-proton is redundant as I as passing -no_optH false anyway.

Non-canonical amino acid remodel. Remodel with   '109 C D EMPTY NC CYX' with '-extra_res_cen CYX.params -in:file:extra_res_fa CYX.params' where CYX.params is straight from the database folder, fails at the centroid build stage with "ERROR: unrecognized atom_type_name 'S'" or 'SH1' if I change the Nbb in the topology file to match the one in CYS. The pdb template has the line "REMARK 0 BONE TEMPLATE   CYX 109 MATCH MOTIF   CYX 110 1''. If it worked it would have been one step, but as I mentioned I have never had a sub-zero score with nn NC remodel as I did not persevere.

CYD residue remodel. Got weird results with CYD, the normal-ish PDB annotation for disulfides. I think there may be something else active in residue_types.txt that is something else.

Thu, 2018-09-13 06:21