A question about rotamers

[Si Yan]

Hello Charles and others,

I have refined some structures from solid-state NMR restraints. When we
compare the NMR structure of our protein with X-ray structures with
different binding partners, we find some PHE's aromatic rings present
rotamer difference between NMR structure and the X-ray structures. It is
hard for us to judge that this rotamer difference is a result of real
difference in structures or it is just because of lacking of restraints to
constrain the side-chain torsion angles.
We only used distance restraints and backbone torsion angle restraints
for the structure refinements. In our NMR restraints list for the
structure refinement, there is no restraints from the atoms of the aromatic
rings.
My question is there any way to constrain the side-chain torsion angle
in Xplor-NIH?
I saw a function named

*correctSymmetricSidechains in selectTools module.*
But I am not sure if this the correct thing I should include in my
refinements. If this is what I could use, how can I use it properly?


Thanks very much!

Best
Si

[Bermejo, Guillermo (NIH/CIT) [E]]

Hi Si,

Something that may help is using a statistical torsion angle potential term. Although no substitute for experimental restraints, it will at least help getting rid of rotamer outliers, so that you are left with plausible ones. (Note that, in the standard approach, this term applies to all the residues. It also affects the backbone, improving the Ramachandran statistics.)

You can access the latest version of this potential, called "torsionDBPot" [Bermejo et. al. (2012) Protein Sci. 21: 1824], through the Python interface. For an example on how to set it up you can take a look at xplor/eginput/gb1_rdc/refine.py, where "xplor" refers to your Xplor-NIH directory.

Additionally, you may want to explore TALOS-N, which now predicts chi1 torsion angles.

I hope it helps.

Best,

Guillermo

________________________________________
From: Si Yan [***@udel.edu]
Sent: Wednesday, April 09, 2014 5:06 PM
To: xplor-***@nmr.cit.nih.gov
Subject: [Xplor-nih] A question about rotamers

Hello Charles and others,

I have refined some structures from solid-state NMR restraints. When we compare the NMR structure of our protein with X-ray structures with different binding partners, we find some PHE's aromatic rings present rotamer difference between NMR structure and the X-ray structures. It is hard for us to judge that this rotamer difference is a result of real difference in structures or it is just because of lacking of restraints to constrain the side-chain torsion angles.
We only used distance restraints and backbone torsion angle restraints for the structure refinements. In our NMR restraints list for the structure refinement, there is no restraints from the atoms of the aromatic rings.
My question is there any way to constrain the side-chain torsion angle in Xplor-NIH?
I saw a function named correctSymmetricSidechains in selectTools module.

But I am not sure if this the correct thing I should include in my refinements. If this is what I could use, how can I use it properly?


Thanks very much!

Best
Si


--
Si Yan
Department of Chemistry and Biochemistry
University of Delaware
Newark, DE, 19716
Email to ***@udel.edu<mailto:***@udel.edu>
Phone: [phone]

[Si Yan]

Hi Guillermo,

Thanks very much for your suggestions. I will try this torsion angle
potential term and the TALOS-N.

Best
Si


On Wed, Apr 9, 2014 at 5:54 PM, Bermejo, Guillermo (NIH/CIT) [E] <

Post by Bermejo, Guillermo (NIH/CIT) [E]
Hi Si,
Something that may help is using a statistical torsion angle potential
term. Although no substitute for experimental restraints, it will at least
help getting rid of rotamer outliers, so that you are left with plausible
ones. (Note that, in the standard approach, this term applies to all the
residues. It also affects the backbone, improving the Ramachandran
statistics.)
You can access the latest version of this potential, called "torsionDBPot"
[Bermejo et. al. (2012) Protein Sci. 21: 1824], through the Python
interface. For an example on how to set it up you can take a look at
xplor/eginput/gb1_rdc/refine.py, where "xplor" refers to your Xplor-NIH
directory.
Additionally, you may want to explore TALOS-N, which now predicts chi1 torsion angles.
I hope it helps.
Best,
Guillermo
________________________________________
Sent: Wednesday, April 09, 2014 5:06 PM
Subject: [Xplor-nih] A question about rotamers
Hello Charles and others,
I have refined some structures from solid-state NMR restraints. When
we compare the NMR structure of our protein with X-ray structures with
different binding partners, we find some PHE's aromatic rings present
rotamer difference between NMR structure and the X-ray structures. It is
hard for us to judge that this rotamer difference is a result of real
difference in structures or it is just because of lacking of restraints to
constrain the side-chain torsion angles.
We only used distance restraints and backbone torsion angle restraints
for the structure refinements. In our NMR restraints list for the
structure refinement, there is no restraints from the atoms of the aromatic
rings.
My question is there any way to constrain the side-chain torsion angle in Xplor-NIH?
I saw a function named correctSymmetricSidechains in selectTools module.
But I am not sure if this the correct thing I should include in my
refinements. If this is what I could use, how can I use it properly?
Thanks very much!
Best
Si
--
Si Yan
Department of Chemistry and Biochemistry
University of Delaware
Newark, DE, 19716
Phone: [phone]

[Charles Schwieters]

Hello Si--

    I have refined some structures from solid-state NMR
restraints. When we compare the NMR structure of our protein with
X-ray structures with different binding partners, we find some PHE's
aromatic rings present rotamer difference between NMR structure and
the X-ray structures. It is hard for us to judge that this rotamer
difference is a result of real difference in structures or it is
just because of lacking of restraints to constrain the side-chain
torsion angles.    We only used distance restraints and backbone
torsion angle restraints for the structure refinements.  In our NMR
restraints list for the structure refinement, there is no restraints
from the atoms of the aromatic rings.    My question is there any
way to constrain the side-chain torsion angle in Xplor-NIH? I saw a
function named correctSymmetricSidechains in selectTools module.
But I am not sure if this the correct thing I should include in my
refinements. If this is what I could use, how can I use it properly?

Guillermo's suggestions of trying torsionDB and TalosN may well help
you immensely.

To complete the conversation, however, I wanted to follow up with a
bit about symmetric sidechains. The presence of symmetric sidechains
(PHE, TYR, ASP, ARG, and GLU) can degrade the apparent overall heavy
atom precision if atoms which are chemically equivalent, but named
differently are in close proximity. The chi2, chi3 or chi5 torsion
angles which define the symmetry operations are given reduced ranges
to remove the degeneracy. correctSymmetricSidechains will make the
appropriate changes such that the torsion angles are in the correct
ranges, and it is automatically called when you load a structure using
protocol.loadPDB or protocol.initCoords. Additionally,
protocol.initDihedrals by default adds additional restraints such that
the symmetry-reduced angle ranges are maintained throughout a
calculation.

So while torsionDB will add knowledge-based information,and TalosN can
give you experimentally based restraints, one should be aware of the
symmetry issue so that conformations which are very close don't appear
to have chi angles which differ by 180 degrees.

best regards--
Charles

[Si Yan]

Hi Charles,

Thanks!
This information is very helpful.
I have tried the torsionDB term, and the RAMA energy does decrease a
little. I will include the torsion angles from TALOS-N for the refinement.

Best
Si

Post by Charles Schwieters
Hello Si--

Post by Si Yan
I have refined some structures from solid-state NMR
restraints. When we compare the NMR structure of our protein with
X-ray structures with different binding partners, we find some PHE's
aromatic rings present rotamer difference between NMR structure and
the X-ray structures. It is hard for us to judge that this rotamer
difference is a result of real difference in structures or it is
just because of lacking of restraints to constrain the side-chain
torsion angles. We only used distance restraints and backbone
torsion angle restraints for the structure refinements. In our NMR
restraints list for the structure refinement, there is no restraints
from the atoms of the aromatic rings. My question is there any
way to constrain the side-chain torsion angle in Xplor-NIH? I saw a
function named correctSymmetricSidechains in selectTools module.
But I am not sure if this the correct thing I should include in my
refinements. If this is what I could use, how can I use it properly?

Guillermo's suggestions of trying torsionDB and TalosN may well help
you immensely.
To complete the conversation, however, I wanted to follow up with a
bit about symmetric sidechains. The presence of symmetric sidechains
(PHE, TYR, ASP, ARG, and GLU) can degrade the apparent overall heavy
atom precision if atoms which are chemically equivalent, but named
differently are in close proximity. The chi2, chi3 or chi5 torsion
angles which define the symmetry operations are given reduced ranges
to remove the degeneracy. correctSymmetricSidechains will make the
appropriate changes such that the torsion angles are in the correct
ranges, and it is automatically called when you load a structure using
protocol.loadPDB or protocol.initCoords. Additionally,
protocol.initDihedrals by default adds additional restraints such that
the symmetry-reduced angle ranges are maintained throughout a
calculation.
So while torsionDB will add knowledge-based information,and TalosN can
give you experimentally based restraints, one should be aware of the
symmetry issue so that conformations which are very close don't appear
to have chi angles which differ by 180 degrees.
best regards--
Charles