Practical Relax

Practical Manual
General outline to analyze relaxation data

I. Estimate a correlation time from protein molecular weight.
II. Screen data for residues undergoing large amplitude fast motion.
III. Screen data for residues undergoing slow conformational exchange.
IV. Obtain global diffusion parameters by fitting T1/T2 from the reduced data set with
various diffusion models. For non-isotropic diffusion the diffusion tensor must be
defined. One needs to know the structure of the molecule to this properly.
V. Use statistical test to decide which diffusion model is appropriate for the molecule
under study.
VI. Obtain local diffusion parameters by fitting, T1, T2, and NOE from the full data set
using the global parameters obtained above.
I. Estimate a correlation time from protein molecular weight.
An estimate of a rotational diffusion rate of a spherical particle in liquid can be made

using the Debye-Stokes-Einstein relationship. The rotational diffusion rate Dr is a
function of temperature, viscosity of the solution and size of the particle given below:
kT
Dr =
6"V
where:
k is the Boltzman constant = 1.3806504 x 10-23 kg.m2/s2.K
T is the temperature.
! viscosity of the liquid. Viscosity of water at 300K is 0.8513 x 10-3 kg/m.s
is the
V is the volume of the sphere.
Rotation correlation time is given by: c = 1/(6Dr).
Exercise:
1. Calculate the volume of Ubiquitin with a molecular weight of 8000 Daltons.
Protein density = 1.35 gr/cm3. One Dalton = 1.66024 x 10-24 gr.
2. Estimate the rotation correlation time, c, for Ubiquitin.
II. Identify residues undergoing large amplitude fast motions or chemical exchange.
Residues undergoing large amplitude fast motion can easily be identified with low NOE
values (< 0.65). While residues undergoing chemical exchange will have unusually low
T2 values. However this Low T2 value has to be distinguished from the one due to
anisotropy. This can be achieved by looking at the deviation of its T1 value as well. The
fraction deviation of T1 values from average due to anisotropic diffusion would be
correlated to the deviation in the T2 value.
Use the following criteria to eliminate anomalous T2:
<T2> - T2,n > SD
(T2,n - <T2>) / T2,n > 3 (<T1> - T1,n) T1,n,
where <T1> and <T2> are the averages over the residues that have NOE > 0.65 and
SD is the standard deviation of the T .
2
Exercise:
1. Go to the subdirectory relax_fit and find the file: data_600_all.27
2. Type: cp data_600_all.27 data.work.
3. Type barbato.awk data.work > exclude.list. Residues with large amplitude
fast motion as well as those with chemical exchange are listed in the file
exclude.list.
4. Open the file data.work in a text editor. Delete residues listed in exclude.list
from the data.work file. Save and close the new data.work file.
5. Count the number of residues in the data.work file.
III. Define the molecular coordinate system to describe the fitted diffusion tensor
In order to calculate the diffusion tensor when the rotational motion is not isotropic one
needs to define the molecular coordinate system in which the tensor will be referenced.
This can be done by obtaining the directional cosines for the dipolar interaction vector
corresponding to the measured relaxation from the protein structure.
Exercise:
1. Go to the subdirectory relax_fit and find the file ubq.pdb.

2. Type NH.awk ubq.pdb > ubq.cos.
3. The directional cosines for all N-H vectors in ubiquitin are now contained in
ubq.cos file.
IV. Fitting T1/ T2 values to obtain global diffusion parameters.
Dipolar relaxation rates for two spins 1/2 (such as N-H) are given as:
1/T1 = d2 [ J(H - N) + 3 J(N) + 6 J(H + N) ] + c2 J(N)
1/T2 = 0.5 d2 [4 J(0) + J(H - N) + 3 J(N) + 6 J(H) + 6 J(H + N)] +

(1/6) c2 [ 3 J(N) + 4 J(0)]
NOE = 1 + ( H /N ) d2 (6 J(H + N) - J(H - N)] T1
where
d2 = 0.1 [(H N h) / (2 r3NH)]2

c2 = (2/15) [N 2 ( //- )2]
Spectral density is given by the fourier transform of the correlation function.
J() = - C(t) e-it dt
For an isotropic diffusion:
J() = [ Sf2 c / (1+(c)2) + (1-Sf2) / (1+(')2 ]
1/ = 1/ c + 1/ e
For an anisotropic diffusion (axial symmetry):
J() = S2 ( A1 1 / [1 + (1)2] + A2 2 / [1 + (2)2] + A3 3 / [1 + (3)2] ) + (1-S2)

/ [1 + ( )2]
A1 = 0.75 sin4 1/ 1 = ( 4 D// + 5 D )

A2 = 3 sin2 cos2 1/ 2 = ( 5 D// + D )
A3 = (1.5 cos2 - 1) 1/ 3 = ( 6 D )
where is the angle between the dipolar vector and the principal axis of diffusion tensor.
To a first approximation the ratio of T1 and T2 is not sensitive to fast motion (S2 and )
and variation in CSA values. The error function is given by:
2
E2 = [(T1/ T2)meas (T1/ T2)calc]
Exercise:
1. Go to the subdirectory relax_fit.
2. Type tc_sweep data.work 56 isotropic.out 3.5 4.5 0.01. This will result in a grid
search for an optimal c. Experimental errors for T1, T2, and NOE is 1%, 1%, and
2%, respectively based on reproducibility of the relaxation values. The data was
obtained at 600.13 MHz.
3. Note the optimal c and E/N values. How does this c value compare to the one
estimated using Debye-Stokes-Einstein relationship.
4. Type woessner_2x_powell data.work 56 ubq.cos. Use the following starting
values: = 40, = 40, c = 4.0, D///D = 1.1

5. Note the global diffusion parameters as well as the E/N values.
6. Type woessner_3x_powell data.work 56 ubq.cos. Use the following starting
values: = 70, Dy/Dx = 1.0.
7. Again note the derived parameters and E/N values.
V. Statistical test for proper diffusion model.
In the last section you have fitted your data to three diffusion models. With each
progression in the complexity, thus increased number of variables, one should see a
reduction in the overall error in the fit (E/N). However a distinction has to be made
between the reduction in error of the fit due to increasing fitting variables versus a better
model used in the fit. One such test is the F-test:
E(m) = E(m) / (N-m)

F = [E(m) - E(m+k)] / [k E(m+k)]
P(F; k, N-m-k) < 0.01
Exercise:
1. Based on the E/N values obtain for the isotropic and axially symmetric diffusion
calculate the F value using the above equations.
2. Compare the F value that you obtain to the F value cutoff for a given P(F) in the
table
3. Do the same calculation for axially symmetric to fully asymmetric fitting.
4. Based on the calculated F and P(F) decide on the proper diffusion model for
ubiquitin.
VI. Calculating local diffusion parameters.
Now that you have the global diffusion parameters the order parameter as well as the fast
correlation time (e) can be calculated for every residue in the full data set using the
following error function:
E2 = ([T1c - T1m] / T1m )2 + ([T2c - T2m ] / T2m )2 + ([NOEc - NOEm ] / NOEm)2
1. Type woessner_3x_2 data_600_all.27 69 ubq.cos sf_te.list.
2. Input the global parameters that you obtained in the previous sections.
3. Once the calculation is finished, check the sf_te.list file.
4. Take a look at Sf and e values for residues that were identified as undergoing
conformational exchange in exercise for part II.
4. See if you can correlate the fast internal parameters to the structure of ubiquitin.

Practical Relax

Uploaded by

Copyright:

Available Formats

Practical Relax

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Practical Relax

Uploaded by

Copyright:

Available Formats

Practical Manual

General outline to analyze relaxation data

I. Estimate a correlation time from protein molecular weight.

An estimate of a rotational diffusion rate of a spherical particle in liquid can be made

Rotation correlation time is given by: c = 1/(6Dr).

1. Calculate the volume of Ubiquitin with a molecular weight of 8000 Daltons.

Protein density = 1.35 gr/cm3. One Dalton = 1.66024 x 10-24 gr.

2. Estimate the rotation correlation time, c, for Ubiquitin.

Use the following criteria to eliminate anomalous T2:

<T2> - T2,n > SD

(T2,n - <T2>) / T2,n > 3 (<T1> - T1,n) T1,n,

1. Go to the subdirectory relax_fit and find the file: data_600_all.27

2. Type: cp data_600_all.27 data.work.

3. Type barbato.awk data.work > exclude.list. Residues with large amplitude

5. Count the number of residues in the data.work file.

1. Go to the subdirectory relax_fit and find the file ubq.pdb.

IV. Fitting T1/ T2 values to obtain global diffusion parameters.

1/T1 = d2 [ J(H - N) + 3 J(N) + 6 J(H + N) ] + c2 J(N)

1/T2 = 0.5 d2 [4 J(0) + J(H - N) + 3 J(N) + 6 J(H) + 6 J(H + N)] +

NOE = 1 + ( H /N ) d2 (6 J(H + N) - J(H - N)] T1

d2 = 0.1 [(H N h) / (2 r3NH)]2

Spectral density is given by the fourier transform of the correlation function.

J() = - C(t) e-it dt

For an isotropic diffusion:

J() = [ Sf2 c / (1+(c)2) + (1-Sf2) / (1+(')2 ]

For an anisotropic diffusion (axial symmetry):

J() = S2 ( A1 1 / [1 + (1)2] + A2 2 / [1 + (2)2] + A3 3 / [1 + (3)2] ) + (1-S2)

A1 = 0.75 sin4 1/ 1 = ( 4 D// + 5 D )

1. Go to the subdirectory relax_fit.

obtained at 600.13 MHz.

estimated using Debye-Stokes-Einstein relationship.

4. Type woessner_2x_powell data.work 56 ubq.cos. Use the following starting

values: = 40, = 40, c = 4.0, D///D = 1.1

5. Note the global diffusion parameters as well as the E/N values.

6. Type woessner_3x_powell data.work 56 ubq.cos. Use the following starting

values: = 70, Dy/Dx = 1.0.

7. Again note the derived parameters and E/N values.

V. Statistical test for proper diffusion model.

E(m) = E(m) / (N-m)

calculate the F value using the above equations.

3. Do the same calculation for axially symmetric to fully asymmetric fitting.

VI. Calculating local diffusion parameters.

E2 = ([T1c - T1m] / T1m )2 + ([T2c - T2m ] / T2m )2 + ([NOEc - NOEm ] / NOEm)2

1. Type woessner_3x_2 data_600_all.27 69 ubq.cos sf_te.list.

3. Once the calculation is finished, check the sf_te.list file.

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.