You are here

how can i refine protein loop region contain nonstandard amino acid morespecifically Phosphoserine and phosphothreonine

7 posts / 0 new
Last post
how can i refine protein loop region contain nonstandard amino acid morespecifically Phosphoserine and phosphothreonine

Recenly i have modeled one protein pknQ, its a kinase and contains phosphoserine and phoshothrionien at their activation loop region. I have used homology modeling methods to predict the 3D structure of pknQ and further refined the loop. I have modified the serine and threonine to phosphoserine and phosphothrenine now to minimize the effect of post translation modification i want to further refine loop having nonstandard amino acid (phosphorylated serine and threionine). How can i do this not getting any idea . please help

Post Situation: 
Thu, 2014-08-07 01:49

Rosetta should be able to automatically recognize serine and threonine residues as phosphorylated, if the phosphate atom is present in the input PDB. (Named " P " in both cases. If that fails, you should be able to use the three-letter codes "SEP" and "TPO" for phospho serine and phosphothreonine, respectively.

Otherwise, some more details on how you've structured your inputs and which protocol you're hoping to use would be needed to further debug.

Tue, 2014-08-12 08:04

Thanks for your valuable reply, actually i am very new user of Rosetta and trying to learn it with myself, i have modeled kinase with homolgy modeling and now trying to refined the activation loop that contains both phasphoserine and phasphothrionine residues, unfortunately never find a single article which have introduced this problem before and hence not getting any idea which protocol would be best for my problem.
kindly please suggest any solution or suitable protocol for the same regards.


Tue, 2014-08-12 22:51

I don't think any of the programs on the ROSIE server ( will help you. (FYI, you've posted in the ROSIE section of the forums.)

What you'll want to use depends somewhat on how extensively you want to remodel your loops. Typically with homology modeling the loops already go through some remodeling, especially those regions which don't have alignments to your template structure. (BTW, what protocol did you follow to obtain your homology models?) Depending on the protocol you used, you could potentially just repeat the loop remodeling stage, this time specifying that the loop of interest needs to be remodeled. The classic (single template) Rosetta homology modeling protocol has a separate loop modeling stage. There is documentation here: and there are some tutorials at (Click "Rosetta Tutorials" and look at the Comparative/Homology Modeling and Loop Modeling/Building tutorials.) I believe (but have not tested) that those loop remodeling stages should work with the phosphorylated residues.

Alternatively, if you're anticipating only minor backbone structural changes from the phosphorylation, you might be able to get away with just relaxing the protein with the new residues. The most straightforward approach is to use a text editor to change the names of the residues of interest to SEP/TPO as appropriate, and then run the relax application ( ) on the edited protein. That should add the phosphate groups and attempt to change the structure to accomidate them. You'll likely want to produce several hundred models, and then examine the lower energy ones.

Wed, 2014-08-13 08:00

Thank you again,
I have installed rosetta 3.4, before going to the tutorials you have mentioned in your previous post i followed some Demos which have provided inside the rosetta bundle. Demo was run successfully using following command for a 12 residue long loop region.

/home/administrator/rosetta3.4/rosetta_source/bin./loopmodel.linuxgccrelease -database /home/administrator/rosetta3.4/rosetta_database -loops:input_pdb 1L3R.pdb -loops:loop_file 1L3R.loop -loops:remodel perturb_kic -loops:refine refine_kic -ex1 -ex2 -nstruct 1 -loops:max_kic_build_attempts 100 -in:file:fullatom

during the entire run i have never come across any error massage so i assumed that this method is able to accept the phosphoserine (initially i tried with one PTM only)
after completion the task it has produced these final lines (bellow):
and generated a pdb file " _0001.pdb"

protocols::checkpoint: Deleting checkpoints of LoopMover
protocols.looprelax: Refinetime: 603
protocols.looprelax: ====================================================================================
protocols.looprelax: ===
protocols.looprelax: === Getting Statistics
protocols.looprelax: ===
protocols.looprelax: ===
protocols::checkpoint: Deleting checkpoints of Loopbuild
protocols.loop_build.LoopBuild: loop_cenrms: 0
protocols.loop_build.LoopBuild: loop_rms: 0
protocols.loop_build.LoopBuild: total_energy: 1970.91
protocols.loop_build.LoopBuild: chainbreak: 0

With the given out put i can expect that the run was successful and apart from these basic parameter value if i will introduce some exhaustive parameter i can get batter result.

Please suggest me best parameter for this protocol for initial run (though i can understand that some time parameter varies with the complexity of problem but i don't know the minimal setup for a quality run for any given case)

Thu, 2014-08-14 00:23

If you get output structures, the run is "successful". That doesn't mean that you're getting the results you want, though.

I'd recommend producing a small number of output structures and seeing if they look reasonable. Are all the parts that are moving supposed to move? Are you getting a decent range of sampling? Do the structures look reasonable from the experimental and intuitional information you have about the system? If something looks off about the structures, then you can look at altering the parameters to correct for it.

The demos provided are a reasonable start, but I'd also recommend looking at other related demos/tutorials (like the Meiler Lab ones) as well as published papers (especially the SupMat) for people doing similar things. This way you can get a better sense of what "knobs" there are to turn, and what sort of things people are looking for in the results.

A final suggestion is to do a "positive control" run. That is, find a similar system where you have both the starting structure and the desired result. You can then run your protocol on this case, and see if the results you get are consistent with the "known good" result.

Fri, 2014-08-15 07:40

Thanks for replying,

Actually i have run domo with a positive control only, its a crystal structure of kinase and have resolved activation loop with phasphorthrionine residue.
Before starting the run i have distorted the loop region which i am looking for and after run i have calculated the RMSD value between crystal and predicted loop region (11 residues ) which is appearing about 4.7 Angstroms (much higher).

So now as per your suggestion i will explore the tutorials and papers which have addressed this kinda problem.

thanks for your valuable support and time.

regards j

Fri, 2014-08-15 08:02