Palmer et al. 1993 JACS 115, 6333-45, measures side chain dynamics in a Zinc
finger peptide using natural 13C abundance but I can not really follow all
the treatment that they did.
Lee et al. 1999 JACS 121, 2891-2902 also measure CH3 dynamics with
ubiquitin.
Still, my doubt is that, due to the rotation of the methyl protons you can
never describe a J(w) for any of the three protons if the molecule has non
isotropic diffusion. Even if you have a CHD2 methyl group, you will never
now the orientation of the CH vector relative to the diffusion tensor.
What I seem to understand from the literature is that methyl rotation about
the symmetry axis is usually fixed to S2=0.111 assuming a perfect
tetrahedral geometry of the methyl groups. So it is not the C-H vector
itself that is question but rather the vector that is aligned with the
methyl rotation symmetry axis. Then, you have the reorientation of the
methyl symmetry axis which also contributes for the motion of the methyl
group. So if you get an S2 much lower than 0.111 it means that there is
additional motion to the CH3 group which can only appear from the
reorientation of the symmetry axis.
Does this make any sense to you?
-----Original Message-----
From: edward.dauvergne@xxxxxxxxx [mailto:edward.dauvergne@xxxxxxxxx] On
Behalf Of Edward d'Auvergne
Sent: quarta-feira, 16 de Setembro de 2009 16:49
To: Tiago Pais
Cc: relax-users@xxxxxxx; Vic Vostr
Subject: Re: CH3 relaxation
Hi,
One can always analyse any collected data, but whether the results are
meaningful is another kettle of fish. So, yes you can use relax - it
won't stop you from doing anything! You just have to be careful and
not fall into the pitfalls already documented in the literature.
There is pseudo-atom capabilities built into the program now that you
might be able to take advantage of. This has not been tested for
standard relaxation data yet, due to the theoretical limitations, but
you're most welcome to try if a pseudo atom is the route you would
like to take. If it fails, just make a bug report
(https://gna.org/bugs/?func=additem&group=relax) and then I or one of
the relax developers should be able to fix it for you.
The issue of orientation is due to the diffusion tensor. If it is
purely isotropic, bond vector orientation is not important. If it is
not, then a false orientation will cause either artificial ns motions
(Schurr, 1994) or artificial Rex (Tjandra et al., 1995) to appear.
But maybe a pseudoatom position between the three proton positions
(i.e. the symmetry axis) would be good enough, if the tensor isn't too
anisotropic? There are also sidechain movements meaning that the
methyl rotation is only one part - yet the most important part - of
the methyl dynamics. Lipari and Szabo used isotropic tensors (well
the looked a little at anisotropy but got that very wrong), so they
never looked at this in detail. The Schurr and Tjandra publications
are what you want. For the interference issues, this will be found in
Kay's papers. The exact reference I don't remember off the top of my
head. Do you have a reference of anyone doing R-CH3 a model-free
analysis before?
Regards,
Edward
P.S. Did you work out what the DLL issue was in the end?
2009/9/16 Tiago Pais <tpais@xxxxxxxxxxx>:
Hi,
The original papers of Lipari and Szabo on model-free address the
interpretation of relaxation data from methyl groups. I found no reference
to the issue of the three reorienting CH vectors. It seems to me that the
axis of symmetry of methyl rotation is used as a reference. Then, a
Woessner
model for the methyl motion is assumed, leading to a theoretical S2 of
about
0.111 for methyl groups with perfect tetrahedral geometry. In this way one
can obtain the S2 of the symmetry axis (eg. Ca-Cb vector in Alanine) by
dividing the experimentally obtained S2 by 0.111.
In any case, even with deuterium partial labeling you would have the
proton
jumping between positions meaning a time dependent vector orientation, no?
I
think the advantage of Deuterium is about the purity of the relaxation
mechanism which is dominated by quadrupolar effects.
Considering that I have acquired all the data in the absence of relaxation
interference, is it feasible to use RELAX to fit the modelfree models?
Regards
Tiago Pais
-----Original Message-----
From: edward.dauvergne@xxxxxxxxx [mailto:edward.dauvergne@xxxxxxxxx] On
Behalf Of Edward d'Auvergne
Sent: quarta-feira, 12 de Agosto de 2009 15:07
To: Tiago Pais
Cc: Vic Vostr; relax-users@xxxxxxx
Subject: Re: value.display()
Hi,
Please see below for answers to your questions.
On Wed, Aug 12, 2009 at 12:53 PM, Tiago Pais<tpais@xxxxxxxxxxx> wrote:
Dear Edward,
I have some questions relating the functioning of relax 1.3.3:
1- When I initiate the relax command a get an error message about a
missing
DLL (python25.dll). However the program manages to start anf get to the
command line. Is this DLL related to the "curve fitting" module? Does the
relax program run properly without this DLL
This is a strange error. It is related to Python and not relax, and
is the main Python library used by all Python programs. Do you have
Python 2.5 installed? Could you cut and paste the error message? Do
you get the same message if you run python from the command prompt as
well?
2- The "relax --test-suite" routine seems to stall after a while. I have
three "F" in the middle of the dotted line. Can this be related to the
DLL
I'm not sure. If you're using Python 2.5 and getting this error, then
relax should not run. This is strange.
The stalling is likely to be on a system test which takes a while to
run. It should be finished in 1-3 min though, unless you are using a
very old machine. I'm currently fixing all the tests on windows in
the 1.3 repository code, and relax 1.3.4 when released should pass
most tests (note that 2 unit tests of the float module are know to
fail).
3- The "TAB" button is not working within relax
This is a windows only problem as there is no special library called
'readline'. If this is too annoying, maybe running this in GNU/Linux
(in a virtual machine if necessary) would be the easiest way to remove
this lack of TAB key.
And now some question related to my specific case:
1- What is the nomenclature to describe the C nucleus in a methyl group?
I
will have to assume an effective CH vector and not three different ones
for
each proton; how do I tell this to the program?
This is slightly problematic because of all the cross-correlated
relation rates (interference). There is no theory to handle this,
that I am aware of. People overcome this by specific labelling so
that there is one hydrogen and two deuteriums (e.g. see Lewis Kay's
papers). Not only are the interference effects fatal, but the three
CH vectors experience different spectral density functions J(w). The
J(w) are direction dependent, and here you have three different
directions which constantly change. Again the theory is completely
missing, as far as I know. What you need is a reference where someone
has handled the 1H-13C relaxation rates from a CH3 group before! Or
you'll need to derive the necessary theory.
2- What is the format for the input R1,R2, and NOE data? Couldn't seem to
find it written in the manual.
relax will handle any column formatted file. See the help for the
relax_data.read() user function in the manual for full details, or
type:
relax> help(relax_data.read)
Sorry for all these questions but I am just starting dealing with this
program.
No problems. If you have any other questions, please don't hesitate to
ask.
Regards,
Edward
_______________________________________________
relax (http://nmr-relax.com)
This is the relax-users mailing list
relax-users@xxxxxxx
To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-users
_______________________________________________
relax (http://nmr-relax.com)
This is the relax-users mailing list
relax-users@xxxxxxx
To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-users
Re: CH3 relaxation