Pulse Sequence and Analysis Scheme for 3D CBCACONH-J(CH)
                         James Chou and Ad Bax
                             Feb 14, 2001

CITATION

JACS 2001, submitted 
Modified from the original pulse scheme by
Grzesiek and Bax  JACS 1992, 114, 6291-6293 

Pulse Sequence File "cbcaconh-ch.jjc"

SUMMARY

This sequence does three interleaved 3D experiments to measure both CaHa and
(CbHb2 + CbHb3) J coupling by the quantitative J approach. The three
intensities correspond respectively to 1.93, 3.73, and 7.22 ms of effective H-C 
J-modulation.  Example data for calmodulin (0.3 mM) is provided, along 
with NMRPipe conversion and processing scripts.  A special-purpose 
fitting program "SFIT" is also included, for calculating the couplings 
from peak evolution curves.  This particular procedure is suitable for
measuring couplings only from well-resolved non-overlapping peaks.

SOFTWARE AND EXAMPLE DATA

The conversion, processing, and analysis schemes described here require
the latest version of the NMRPipe software system.  For the latest NMRPipe 
FTP instructions and password information, send email to delaglio@nih.gov

The example data can be found in the file "pub/jch.tar.Z" at the FTP site. 
Once downloaded, the sample data directory, called "jch", can be extracted 
with the command:

   zcat jch.tar.Z | tar xvf -

MEASUREMENT of CB-HB and CA-HA J-COUPLINGS

This experiment measures CB-HB and CA-HA J-couplings by Quantitative-J
approach using three interleaved CBCACONH experiments. The pulse scheme
is based on the original CBCACONH experiment. C-H J-modulations (tau) of 
1.93, 3.73, and 7.22 ms are introduced respectively in the three interleaved
experiments. Note that the above tau delays are calculated based on fact
that Hsec.3 refocusing pulse (p16) is 500 us long at 150MHz. Proper
calibration should be done if p16 is very different from this value.  

The typical processing and analysis steps are as follows:

1. fid.com

   The interleaved data is converted by a special scheme which
   separates the three 3D experiments.

2. xyz.com

   Each of the three 3D experiments is processed in the usual way.

3. nmrDraw 

   The first 3D spectrum is used to establish a 3D peak table
   for the analysis.  The demo directory includes example peak tables 
   "A-ca.tab" and "A-cb.tab" for the C-alpha and C-beta regions,
   with residue ID assignments recorded for all peaks of interest.

4. series3D.com

   This script uses the "seriesTab" program to extract peak evolutions
   over the series of three spectra.  The results are stored in the
   output file "series.tab"

5. sfit.com

   This script uses the special-purpose program "SFIT" to extract
   the peak evolution curves from the "series.tab" file.
   SFIT uses SIMPLEX optimization to fit three measured intensities 
   to the function A*sin(PI*J*tau). The two unknowns A (intensity) 
   and J (coupling) are thus determined. 

For accurate extraction of J values, one must pay attention to the 
Signal-to-Noise (S/N) ratio of the reference spectrum, which determines 
the error of measurement. In general, error of measurement can be 
estimated by the relation 

   ERR(Hz) = 50/(SN). 

The pulse scheme is designed primarily to measure the CB-HB coupling. Hence, 
the CB peaks are more sensitive than CA peaks. Nevertheless, this is a very
sensitive experiment. For 1mM sample, one should get acceptable (> 20:1)
S/N for both CB and CA cross peaks. 

The quality of a given fit is indicated by the RMS of the residual
given in the output of SFIT.  This will typically be less than 0.1 
for a good fit.  Most bad fits are result of either low S/N or strong 
C-H relaxation interference. However the latter will often
be roughly cancelled out when measuring dipolar couplings 
(aligned - isotropic).