EMBO Practical Course:

Structure, Dynamics and function of biomacromolecules by

solution NMR
Biozentrum, University of Basel, July 20- 27, 2013

NMRView Practical
CONTENTS:
1. Basic functions
2. Analysis of NMR relaxation data
3. Analysis of NMR titration data
4. Protein assignment

Bernd Simon, EMBL Heidelberg & Sonja A. Dames, TU Muenchen

Lit.: Johnson, B. A., and Blevins, R.A. (1994), Journal of Biomolecular NMR, 603.
www.nmrview.com

Programming interface is a general scripting language: Tcl/Tk

1
1. Basic functions
Introduction to some basic features of NMRView: interaction with datasets, spectral
windows, peaklists and molecules.

Change to the directory getting_started and start nmrview by typing:

xxx$ cd getting_started
xxx$ nmrview

Displaying spectrra:
To open and draw the NMR spectra hnhsqc, hnca, cbcaconh, hncacb, select from the
main menu:
Dataset-> Open and Draw Dataset: Double click on name of dataset
Alternatively:
Dataset-> Open Dataset: Double click on name of dataset
Windows-> Add: type a name (e.g. hnhsqc) and define the number of rows and
colums (e.g. 1/1). NOTE: to try this variant after you used already the Open and
Draw command for a certain dataset, close the window if you want to use the same
name for it and omit Open Dataset because the dataset is already loaded.
In the newly created window press the right mouse button, set the cursor to
Attributes and press the left mouse button.
This opens a new window:

2
In this window selecting File->Dataset opens another new window that allows to
select an before opened dataset (e.g. hnhsqc) to be displayed in the currently active
window (e.g. hnhsqc) just click on a dataset in the left half and press the Add
button. The name of the dataset appears then to the right. Press Draw in the spectral
attributes window to get the spectrum displayed.
With following parameters and menu option, the appearance of the spectrum can be
changed:
Lvl adjustment of contour level
CLM & NLvls spacing and number of displayed contour levels
X, Y, Z, Z2 dimensions of the spectrum and the spectral range for each dimension
Colors->Positive (Negative) adjustment of the color for positive (negative) peaks

Managing and referencing datasets:

Select Dataset->Manage Datasets from the main menu. This opens a new window
listing all loaded datasets. To obtain information about the size and referencing
highlight a dataset of interest and press the Reference button. If you change the
values for ref, press the buttons WritePar followed by ReadPar. If you now redraw
the respective dataset in a spectra window, you can see that the referencing has been
accordingly adjusted.

Superposition of 2D spectra:
If you want to superimpose 2D spectra (e.g. 1H-15N-HSQC of free and bound
protein), you can load more than one spectrum into a spectrum window in the same
way as described above. Then do not press Draw but select from the spectra window
menu:
Misc->Overlay2D this opens a new window that allows to adjust the contour levels
and colors for each dataset as well as to switch them on an off.

NOTE: all panels such as spectral attributes, peak attributes etc. refer to the currently
active window! The name of the window is displayed in the top bar of the panels: e.g.
.specAttr: .nhhsqc.0 hnhsqc.nv means spectral attributes panel for window hnhsqc
displaying spectrum hnhsqc

3
Displaying a 3D dataset:
A 3D dataset can be opened and drawn as a 2D dataset.
A 3D is displayed as a 2D with dimensions X and Y, whereas the plane (or the
section) is defined by Z. You can change the dimensions of the dataset (1,2,3) that
should be displayed as X,Y or Z.
If you load the 3D in an already existing spectra window, set the values defining the
range for one of the indirect dimensions (e.g. Z) in the attributes window to the same
value in the two boxes for Z, e.g. for the hnca to 132.2264 to display only the 2D
plane corresponding to 132.2264 ppm of the 15N dimension. With the Open and
Draw Dataset option this is automatically done.
Open and draw the hnca.

Referencing of the HNCA to fit the 1H-15-HSQC & zooming in and out:
- Draw the HN-N projection of the hnca spectrum:
o Change the dimensions of the display so that the 3rd dimension = 13C
1H=1 and 15N=2 (click on 3 or 2 or 1 to change the order of
dimensions)
o Make sure that for Z=13C the full spectrum is selected: click on Z
select full ( all 3D planes are displayed simultaneously)
o Change Lvl to a higher value maybe 0.1
o Click Draw (takes a while). You may have to increase Lvl more &
redraw
o Overall the selected HN-N projection of the 3D HNCA should look
like the 1H-15N-HSQC.

- Zoom into a region of the HN-N projection around a peak for which you
choose to determine the correct referencing :
o The window you will expand is defined by the 2 crosshairs: the black
one is moved with the left mouse button, the red one is moved with the
middle one. Use both crosshair cursors to define a square enclosing a
certain region of the 2D plane.

4
 o Click right mouse to get the spectra window menu, select View-
>Expand
- Expand the same region/peak in the hnhsqc spectral window
- Center black crosshairs on a selected hsqc-peak (left mouse button)
- Center red crosshairs on the equivalent HN-N projection peak (middle mouse
button)
- Click right mouse button and select menu option Procedures-> reref (in the
hnca window!)
- Redraw the HN-N projection and check the referencing
- To draw the full projection and the full hnhsqc:
o Click the right mouse button to get the menu, select View->Full, do
the same in the hnhsqc window. If a crosshair cursor is now moved, it
should be on the center of a selected peak in both windows.

Loading an amino acid sequence:

- Molecules-> Read Topology -> Sequence file choose the file paz.seq
- Assign-> Atoms lists all residues and the respective nuclei, using special
commands (e.g. getpkppm) assignments from a peak list can be loaded

Peak pick the 1H-15N-HSQC spectrum:

- Draw the dataset hnhsqc at a level which shows all real peaks but no noise
- Click with right mouse in the spectrum and chose menu option Peak->Pick
o Enter a name for the list (e.g. paz_nsqc), and change options as
required: Window all peaks in the displayed window are picked
Box only peaks in region defined by the crosshairs in the spectral window
Select New if you create a new peak list. For picking a 3D set the thickness to 1.
- Look at the peaklist: Assign-> Peaks
- In the new peak list window select from the menu: List-> your listname (e.g.
paz_nsqc)
- To save your peak list as separate text file, e.g. to edit it or to use the
information in excel or another program: File->Write List opens a new
window in which you can select the directory where to put the file and type

5
 the name, now including the extension .xpk, e.g. paz_nsqc.xpk (you can also
save it as .txt if you want to import it in excel etc.). The .xpk file can be
opened with any text editor.
- To open and display a stored peaklist named hnhsqc_part.xpk do:
File->Read List select hnhsqc_part.xpk from the getting_started directory
- Display the new peaklist in the spectral window of hnhsqc: right mouse button
Peak-> PeakAttributes, then select the list hnhsqc_part.xpk and click Draw.
- Display the peak assignments instead of the numbers: click right mouse button
Peak-> PeakAttributes, then select Label With: Residue and click Draw.

Stripplots for a 3D dataset based on a assigned HSQC spectrum:

To display the peaks of a 3D data set for a certain residue/HSQC peak, e.g. the CA-
HN correlations from a 3D HNCA for a certain 1HN-15N HSQC crosspeak:
Windows ->Strips in the small new window type a name, e.g. hnca_str and
press create. This generates a new window. Press Params to define, which
datasets and peaklists to use
Dataset press on the arrow button and select hnca.nv
Read - press on the arrow button and select hnca_part (= reference HSQC
peak list)
Display not used now, here you can select a peak list for the respective 3D
spectrum to be displayed in the strip plots
MaxWindows number of strips, e.g. 5
DimX select 1H
DimZ select 15N
then press accept and close
Press All in the strip window to load all strips and double click on the first one
to get them displayed.
To adjust the countour level set the cursor to one strip, press the right mouse
button to get the spectral attributes window. Adust the contour level (e.g from
0.5 to 0.3) and press Unify in the strip window to get the same setting for all
strips.

6
 Peaks in the 3D can be picked as in a 2D:
in one strip frame a peak by the cursors and do as above: press right mouse
button, Peak->Pick type name e.g. paz_hnca, select Box and set the thickness
to 1. Then press Parameters and in the parameter window select under
Display the newly generated list paz_hnca. If you now move to the next
window and press the right mouse and select from the menu Cursor-
>PeakAdd you can pick further peaks. To get the crosshair mode back press
the right mouse button and select Cursor-> Crosshairs.
The addition of new peaks can also be checked in the peakPanel window:
under List select hnca_paz. If you want to type in assignments, the labeling
must be as in the atoms assignment panel (e.g. 55.hn, 55.ca, 55.n, the residue
name will automatically appear based on the information from the loaded
sequence file if you click to the next peak and back again using the arrow
buttons next to the peak number.

Dualstripwindow for the parallel display of corresponding strips form 2 different 3Ds:

If you want to display strips from two different 3Ds for each residue/HSQC reference
peak next to each other, e.g. from an HNCAB and a CBCACONH, you can test the
dualstrips tool: Windows->Strips2
Works overall similar as the Strips tool, however has its own attributes button
(Attrib) to enable separate adjustment of the contour level for each 3D. Maybe try
this at the end of the practical or at home.

Make your life easier by defining some own keys:

We make for the first time use of the programming environment. All commands are
accessible via the console or can be bound to keys. New definitions are achieved by
typing or by reading from a file:
- Type in console: bind Spectrum <e> { nv_win expand }
then you can expand a spectral region by just typing e
- Check the definition of the keys: press the right mouse button in an spectral
window and select Misc ->KeyBinding

7
 - Type in the NMRView console: xxx% source keys.tcl
If you look again at the key binding list it shows now more options

NOTE: you can also source other tcl-scripts to add new functions to NV or edit a copy
of a NV tcl-script to change its functionality (e.g. to modify the bin definition for
generating distance restraints from NOE restraints in the noe.tcl script). The tcl scripts
included in NV can be found in the NV program directory (at least for the
LINUX/UNIX version). Self made or modified tcl-scripts can for example be stored
in a directory such as tcl_user in your nmrview directory.

Save the current status of NV:

It is very important to make frequently backups of your achievements.

File -> Save state (saves open datasets and spectral views, however does not work
well for windows containing superimposed data sets and not for all types of stripplots
etc.)
File -> Write star file (saves peaklists and molecules)
To retrieve the save information after restarting NV use:
File -> Read Star File
File -> Load State

Exit nmrview:
File -> Quit

Program Preferences:
File->Prefs
allows to define data paths and more

8
 o 2. Relaxation Experiments

Preparing Spectra & Peaklists:

Process spectra with enough zero filling (final data matrix at least 1k x 1k
points)
The names of the datasets of a series need to follow the convention:
MatrixRootNumber.MatrixExtension (Number = 1 .. n) (e.g. t1rel1.nv
t1reln.nv)
The relaxation time corresponding to a Number is specified in a two column
text file with the format Number Time (see for example file T1times.txt)
Check that the peaklist matches the spectrum exactly
Delete peaks that overlap
Check that your spectra overlay:
If you want to checke if all the spectra of a set (T1 or T2 or NOE)
superimpose, you can load the other spectra of the set with File -> Dataset
followed by File -> Data -> Attributes -> Misc -> Overlay2D (see also 1.
Basic Functions)

Determination of the relaxation times

Change to directory relaxation and start a new nmrview session:

xxx$ cd relaxation
xxx$ nmrview

open all T1, T2 and hetNOE data in NV

File -> Load state (opens needed datasets and spectral views)

read in peaklist
Open the peakPanel:
Assign-> Peaks
in peakPanel: File -> Read List : st_t2.xpk

display peaklist in spectral window:

right mouse click, select Peak -> Peak Attributes
select Peak List, Color by Assigned, Label With Residue, press
Draw

source some additional scripts and key bindings in the NMRView console:
xxx% source ../tcl/source_relaxation

place the cross-hairs to enclose a region without peaks in the spectra to be

analyzed type in the NMRviewconsole:
xxx% getval

9
 the sdev value will be used for the rate analyis tool as the standard deviation
of the noise (on the other of 0.1) do this for 3 data sets (t1rel, t2rel, ) one
after another and note the values. NOTE: when I tried to set the crosshair
cursors when the rate analysis tool was open NV crashed!

open the relaxation analysis script

Analysis-> Rate Analysis

Setup the parameters for T1 relaxation:

- Matrix Root: t1rel

- Matrix Extension: nv
- Peaklist: st_t2
- click the button Setup and choose the file T1times.txt
- click button FitPar choose Equation 1, , StDev. Method choose Input an type
the sdev value from above in the following box (in picture above set to 0.15)

Click through the peaklist an have a look at some decay curves.

Click the button FitAll and give a filename to store your T1 data, e.g. t1_result
Click the button Reset

10
Repeat the procedure for the T2 relaxation data. ( use t2rel, T2times.txt)

Analyze the hetronuclear NOE data:

Select from main menu Analysis-> HetNOE Analysis

- Peaklist: st_t2
- NOE Data: t1rel
- Reference Data: nv
- Std Dev of Noise: 0.05
- click Analyze asks for a filename, type e.g. noe_result and press Save

We have now files for T1, T2 and heteronuclear NOE for each peak (e.g. t1_result.txt
& t1_result.dat, t2_result.txt & t2_result.dat, noe_result). Open a Linux shell and
look at the files we just created. To do a quick analysis of some average values and to
estimate the overall correlation time of the molecule the following python scripts may
be used. Alternatively the data can be used for programs such as TENSOR2 to be
analyzed according to the Lipari & Szabo model-free formalism.

e.g. t1_result.txt:
st_t2.0 3.hn 874.116821 875.105831 873.057900 5.773692
st_t2.1 4.hn 823.842283 825.077653 822.824321 4.062249

e.g. t1_result.dat:
3 0.874116821 0.001058921 0.00098901
4 0.823842283 0.001017962 0.00123537 ...

e.g. t2_result.txt:
st_t2.0 3.hn 149.765015 149.968879 149.619557 9.487787
st_t2.1 4.hn 141.739993 141.889432 141.506924 6.594122

e.g. t2_result.dat:
3 0.149765015 0.000145458 0.000203864
4 0.141739993 0.000233069 0.000149439 ...

e.g. noe_result:
st_t2.0 3.HN -1.55176354011 0.00415108151296
st_t2.1 4.HN -1.47967042021 0.00510848667948

11
Additional remarks:

Depending on the quality of the data you can choose different modes of peak
integration. Jitter takes the maximum/minimum intensity of the peak in a region
around the peak center. Instead you can try to use Volume or Evolume. For not
completely overlapping peaks you may want to use some deconvolution routine, as
for example nlinLS from the nmrPipe package. To convert the peaklists to nmrDraw
peak table format use the routine nv2nmrDraw (bsconversion.tcl) , (Before using the
standard nmrPipe scripts fit.com and model. Com you need to cluster the peaks using
clustTab.tcl)

Use of the python scripts:

For all python scripts: If you copy them to another computer, open the script with a
text editor and correct the path to the bin directory for python. If you look at the
scripts with an editor you can also figure out what exactly they are doing. To get
information about the usage type the name of the python script. Usage examples:

xxx$ ./Hnrelax_analysis.py t1_result.txt t2_result.txt noe_result

Generates different files, among them some xmgrace files that contain different plots
(t1, t2, and noe values as well as t1 over t2 as function of the sequence. In additional
several calculated values are displayed in the shell, where you started the script.

xxx$ ./correlation_time.py 298 45

Calculates the rotational correlation timec using the provided parameters
(temperature in K and number of residues).

xxx$ ./Get_Size.py 298 3.5

Provides protein size in terms of number of residues for a certain temperature and
rotational correlation timec .

./relaxation_rates.py
asks for the resonance frequency of 15N, e.g. ~60 MHz on a ~600 MHz spectrometer
and the rotational correlation timec to provide estimates of the to be expected T1
and T2 times in ms.

12
3. Titration Experiments
Titration module from Kevin Gardner (http://freedom7.swmed.edu/titration.htm)

Preparing Spectra & Peaklists see Relaxation Experiments

Calculating Kd-values

Change to directory titration and start a new nmrview session

xxx$ cd titration
xxx$ nmrview

Open all HSQC spectra in the order of the titration, has already been done and
saved, therefore only do:
File -> Load state (opens datasets and spectral views)
Read in peaklist
Open the peakpanel:
Assign-> Peaks
in peakpanel: File -> Read List select embo.xpk
Display peaklist in spectral window:
right mouse click in spectrum window : Peak -> Peak Attributes etc.

13
 Source the titration scripts and key bindings:
xxx% source ../tcl/source_titration

The titration analysis tool is opened by typing the following in the NMRView
console:

xxx% NvMkTitrationWin

Enlarge the window with the mouse (to see Close button in lower right corner)

Modify the setup in the Titration window:

- Hmax 1.3 Nmax 8

- Dataset for enumeration: select hsqc11mer_ref.nv
- Peaklist: select embo
- choose contour level suitable for peak picking: Level: 0.14 (do not click Draw
danger of crashing when pressing the PickAll button)
- adjust contour level of individual spectra (click arrow buttons next to Change active
levels by )
- click Redraw
- enter the protein concentration Protein Concent: 0.089
- to provide the ligand concentrations for the different titration steps, type in the
NMRView console:
xxx% readConcentration Concentrations.txt
- place the crosshairs cursors to enclose all peaks in the separate window for the
reference spectra that has been loaded. If I done in the spectrum within the titration
tool NV may crash.

14
 - click PickAll
o all spectra will be picked and a new list is loaded for each spectrum
- click LinkPeaks
o the result depends on the selected Chemical Shift Cutoff ( 0.1 0.4)
(= minimum amount of movement of peak to be graphed for KD calc)

- click Calculate KDs

o xmgrace is opened and the shifting peaks are fitted to the titration
equation
o the results of the fit are shown in the NMRView console and several
files are written to the directory where you started NV (KD.dat,
KD2par.txt, KDVals.txt, graceconsole.txt, PeakLink.txt, titrafit.fit)

Remarks and additional routines that may be used as recommended by Bernd Simon:
(in NMRView console)
- Instead of getting all fits in one window inside xmgrace you can fit the
different peaks individually. Try the routines
o CalcKDcurr
o CalcKDall
- Instead of peak picking all spectra and tracing the peaks you can use the
routine:
o emboPickAll

15
 to perform the peak picking, array formation etc.. This routine copies the
reference peak list and moves the peaks for each titration point. To fit the Kds
in the end you have to read in the reference peak list again before doing the
fitting.
- If the automatic peak tracing fails, copy the original peak list and adjust all
titration points by going through each peak and adjust with the command
MovePeakInterplMaxNew (needs bspeak.tcl and bspeak3D.tcl to be
sourced). The lists are then arranged with the commands initKGpklist and
initKGlinkedarr and continue with the peak fitting as above.

16
4. Protein assignments
Change to the directory cbca and start nmrview.

Restore prepared nmrview session:

- File -> Read Star File select cbca_start.str after setting the path to the
current nmrview directory (.../nmrview/cbca) and open. This opens your
peaklists and loads the protein sequence (and structures).
- File -> Load State this opens datasets and spectral views previously stored
with the SaveState command in the current directory of nmrview! (check with
pwd in the console). If you get an error message, make sure that you really
started NV from the directory cbca of the nmrview practical. Save state and
Load state write/read the file .nmrview in the current directory.
- Source some user defined scripts for defining keys and procedures used later
in the practical: type source sourceall in the NMRView console.

CBCA Analysis:
Introduction to the backbone assignment tool. To give some hands on experience you
are going to connect and assign a ~20 amino acid stretch of a medium size protein.
For the assignment, three spectra are going to be used: HNCA, HNCACB and
CBCA(CO)NH. See also the nmrview manual for more information: Help -> Users
Manual (Internal or External Viewer), Select CBCA analysis

The basic workflow for the backbone assignment is:

1) create peaklists for all spectra including the 2D 1H-15N-HSQC, which serves
as reference list containing the HN and N shifts for all residues but prolines.
2) Cluster peaks of the 3D-spectra based on common HN/N frequencies. Match
adjacent clusters based on CA and CB chemical shifts. Place matched
fragments onto your protein sequence based on CA and CB chemical shifts.
Assign the clusters and transfer the assignments to the chemical shift database.

17
For the practical, you make use of already prepared peak lists for all spectra and you
start right away with the the clustering.
Setting up the cbca preferences:
- Open cbca window (Windows->CBCA) and open the setup preferences in the
cbca window (File -> Preferences)
o To make sure that you dont do a typo, you are going to open the cbca-
preferences that have been prepared for you: File -> Restore. This
command loads and anlyses the file NvCBCA in the current directory,
that has been created previously with the command File -> Save.
Check all entries for the preference.

o Make sure that the Labels are the same as in the peak lists (to check
this, select in the peak panel Edit->Reference), check for example that
the proton dimension is not called HN but 1H etc.
o The Tolerances (Interlist) should be something like 1H 0.05, 13C 0.6,
15N 0.2 (if you make sure that all peaks belonging to one amino acid

18
 are all picked in the same 15N plane you can choose tighter tolerance
on 15N)
o Define the peaklists and spectra with the push-buttons:
- InterResidue: Filename Select cbcaconh.nv Peaklist Select
caconh
- IntraResidue: Filename Select hnca.nv Peaklist Select hnca
- Select the spectrum region to show only the CA region (Limits
41.4 74.3, Full is not selected)
o Check that Show row is selected, change the Row selection from 0 to
1
o Matchlists: enter the names of the peaklists you loaded/selected above
in the following order: hnca, caconh, hncacb, cbcaconh
o The Matchscript is MatchCBCA4
o Close the CBCA-Preferences window.
Peak picking and cleaning of peak lists (examine the results)
- In the main CBCA window select Mode -> Examine Peaks
- Each Row of the CBCA window displays 4 strips (a strip is defined by a
peak in the hsqc: the 15N shift defines from which Z plane of the 3D a strip is
extracted, the HN shift corresponds to the center of a small defined region of
the X axis, whereas the full carbon region is displayed on the Y-axis)
- The central 2 strips of each row have the same Z and X dimensions but show
different spectra, the outer 2 strips of each row change in what they are
showing depending on the current Mode
- In our current setting concentrate on the 4 strips in the center of the display
(the 2 middle strips of each row). The spectra displayed are:
o Top left: ca (and cb) of residue i-1,
cbcaconh
o Bottom left: cb (and ca) of residue i-1,
cbcaconh
o Top right: ca of residue i (and residue i-1) hnca
o Bottom right: cb (and ca) of residue i (and residue i-1) hncacb

19
NOTE: the cbcaconh is used twice to get information about the CA and CB
nuclei of residue i-1, alternatively other 3D data sets (HNCOCA etc.) may be
used in addition or as replacement
If you see nothing with the peak no. set to 0 just scroll through the following
peaks using the arrow buttons next to Peak. If this does not help, adjust the the
contour level settings.

20
Cluster peaks:
- Cluster peaks according to spin-systems/ common HN/N frequencies (cbca
window: Mode -> Cluster Peaks)
- Press Link intralist followed by Link interlist to start the clustering
- Examine the result of clustering with Mode->Edit Custers (HN-C) and scroll
through the clusters using the arrow up (and down) button(s)
Before editing the clusters it is recommended to vary the matching tolerances (
File-> Preferences) to optimize the performance of the automatic clustering. Most
of the clusters should contain 4-6 peaks. If the tolerances are chosen too tight you
will get many incomplete clusters with fewer than 4 peaks. If the tolerances are
chosen too loose, clusters with more than 6 peaks correlating two amino acids will
be created. The optimum value will depend on the quality of the spectra and the
peak picking.

Match clusters:
- Match clusters (cbca window: Mode -> Match Cluster): automatically looks
for possible connections ( chemical shift matches) between the different
spin-systems if the button Match is pressed
- Verify connections (cbca window: Mode -> Edit Matches): check automatic
assignments and confirm correct links. In this mode the CBCA window
displays 8 strips (2 rows of 4 strips):

i-1 i i i+1
CA shift info hnca cbcaconh hnca cbcaconh
CB shift info hncacb cbcaconh hncacb cbcaconh

o The central 4 strips belong to the current cluster that corresponds to

the 1H-15N peak of residue i . Top left and bottom left: cbcaconh
information (CA and CB shift of the preceding residue = i-1), top right
hnca and bottom right hncacb (CA and CB shift of current residue i,
displays also that of i-1).

21
 o The two windows on the left display the hnca and hncacb of the
clusters that match best with the cbcaconh of the current cluster (->
possible residue i-1)
o The two windows on the right display both the cbcaconh of the clusters
that match best with hnca and hncacb of the current cluster ( ->
possible residue i+1)
o You can easily check and compare all automatically detected matches
for i-1 and i+1 by clicking through the lists (left i-1, right i+1) in the
little cbcaCntrl window. The fewer contents are in the list, the less
assignment options exist. The numbers next to a cluster give some
information about the matching cluster. See the nmrview manual for
detailed information: Help -> Users Manual (Internal or External
Viewer),
Select CBCA analysis->Confirm best cluster links

22
Example for how to use of a Tcl-script to incorporate an external c-program to help
backbone assignment:
The chemical shifts of the individual clusters or series of clusters are examined to
identify certain types of amino acids to facilitate sequence specific assignments. For
the localization within the protein sequence an external program (seq_prob: S.
Grzesiek and A. Bax (1993) J. Biomol. NMR, 3, 185-204) is used within NMRView:
- You can find the procedures used in the file ./tcl/bscluster.tcl, which has been
sourced with the other scripts an the beginning of the session (to verify this
you can open the sourced file sourceall with a text editor). You can have a
look at the newly defined procedures by typing in the NMRView console:
more ./tcl/bscluster.tcl. (The procedures underlying a specific command can
generally be seen by typing more commandname in the NMRView console.)
- Find potential Serin or Alanin clusters:
o In the NMRView console you type FindAla, FindSer, FindAlaSuc
(which clusters 'suceed'/ have as preceding residue an alanine) see also
procs in bscluster, more FindAla, examples:

(cbca) 3 % FindAla
Cluster 5 is Alanine ? NN CA=55.79800 CB=18.13500
Cluster 9 is Alanine ? NN CA=55.97700 CB=18.53100

(cbca) 5 % FindAlaSuc
Cluster 17 is succeding Alanine ? NN CA=56.06400 CB=18.85300
Cluster 27 is succeding Alanine ? NN CA=55.82100 CB=18.51000

- Assign the previously connected clusters to your protein sequence:

o Type ExtractGrShifts to get statistical probabilities for the current
cluster, e.g . no. 8:

(cbca) 8 % ExtractGrShifts

23
Cluster Number 8
aa1 45.39600 0.0
aa2 53.16700 33.74900
Label ca cb ALA ARG ASN ASP CYS GLN GLU GLY HIS HIH
ILE LEU LYS MET PHE PRO SER THR TRP TYR VAL
aa1 45.40 0.00 0.0 0.0 0.0 0.0 0.0 0.0 0.0 100.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
aa2 53.17 33.75 0.0 16.2 1.1 0.0 4.4 4.6 11.2 0.0 24.5 2.8 0.0 0.0 8.1
24.0 0.0 0.0 0.0 0.0 3.1 0.0 0.0

prob.: 6.351e+01: GLY76: LYS77

prob.: 3.634e+01: GLY121: GLN122
prob.: 6.138e-02: ALA2: MET3
prob.: 6.138e-02: ALA4: MET5
prob.: 2.068e-02: ALA18: LYS19

o Type ExtractGrShiftsFrag 3 5 10 to locate the fragment of adjacent

clusters 3, 5, and 10 to you protein sequence:

(cbca) 9 % ExtractGrShiftsFrag "3 5 10"

Cluster 3: Caim1:54.89700 Cbim1: 31.50500 Cai: 63.88400
Cbi: 32.10700
Cluster 5: Caim1:64.38000 Cbim1: 36.05100 Cai: 55.79800
Cbi: 18.13500
Cluster 10: Caim1:59.02500 Cbim1: 62.92100 Cai: 58.83500
Cbi: 65.30100
Label ca cb ALA ARG ASN ASP CYS GLN GLU GLY HIS HIH
ILE LEU LYS MET PHE PRO SER THR TRP TYR VAL
aa1 54.90 31.50 0.0 17.2 0.0 0.0 13.2 11.3 16.0 0.0 17.1 2.0 0.0 0.0
5.9 10.7 0.0 0.0 0.0 0.0 6.5 0.0 0.0
aa1 63.88 32.11 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 52.2 0.0 0.0 0.1 0.0 47.7

24
aa2 55.80 18.14 100.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
aa3 58.83 65.30 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 99.4 0.6 0.0 0.0 0.0

prob.: 1.000e+02: PRO46: PRO47: GLN48: SER49

prob.: 3.446e-13: TYR84: PHE85: HIS86: SER87
prob.: 0.000e+00: GLY1: ALA2: MET3: ALA4
prob.: 0.000e+00: ALA2: MET3: ALA4: MET5
prob.: 0.000e+00: MET3: ALA4: MET5: PRO6

Check your assignments by reading the peak list hnhsqc of the session getting_started
and displaying the assigned peaklist in the hnhsqc spectra window.

Create your own view to assign side chains:

The major strength of NMRView is the easy setup of an arbitrary number of spectral
views. As an example you are going to setup a view which helps in identifying side
chain carbon resonances. Four spectral views (strips) are shown next to each other:
CBCA(CO)NH (hni+1), (H)CC(CO)NH-TOCSY(hni+1), and (H)CNH- NOESY(hni+1
and hni)

- Read in the assignments of the molecule: Open assignment panel: Assign-

>Atoms and in the new window load the assignments: File -> ReadPPM.
- Bring the windows for NH detected sidechain experiments (cconh.nv,
chnnoe.nv) to the front, create the axis order X 1H Y 13C Z 15N (click right
mouse button inside spectrum to open the attributes window). Create the same
view as the current strip of your CB(CACO)NH window by clicking in one of
the corresponding strip windows of the cbca window the right mouse button
and selecting Edit-> Copy. Then move the cursor to the cconh spectrum, click
the right mouse button and select Edit-> PasteLimits..
- Create a new spectral window with four columns and one row by selecting
from the main NV menu: Windows-> Add

25
 o Type a name, e.g. sidechain and select 4 columns and 1 row .
- Copy the current strip of your CB(CACO)NH window to the left spectrum:
o Click right mouse button in the cbcaconh (second strip in top row) in
cbca-window and select: Edit-> Copy
o In the same strip click the right mouse button and select Edit->
PasteAll
- Display CC(CO)NH spectrum in second column/strip:
o Click the right mouse button and select Attributes. In the new window
select: File->Dataset in the new window click on cconh.nv and press
Add, change the axis order to have again 13C on y and 15N on z
o Display as CBCA(CO)NH: Right mouse: Edit->Paste Limits
- Display (H)CNH NOESY spectrum (= chnnoe) in strips three and four
analogously as described for the cconh data set.

- To adjust a strip/column of the side chain window to a specific 15N shift:

o Center a strip, e.g. no. 4, to the HN frequency of residue i (not i-1):
o In the assignment table highlight the HN frequency of residue i
o Typing Cntrl-N jumps to the plane corresponding to the 15N shift (see
key bindings)
o Change the 1H-limits of the view (lef/right arrow-keys or in the
spectral attributes window)

- To allow to change the settings for all four columns/strips with one command,
type the command Goccsi 79 in the NMRView console. This changes the
settings of the sidechain window such that the first three columns show the
strips corresponding to HN/N shift information of residue 80 and the last
column corresponding to the one of residue 79 (for more information type
more Goccsi in the NMRView console)
- Check the assignments by comparing them to a chemical shifts database: Go
to residue 79 in the assignPanel (type 79 in the upper right corner and press
return). Click inside a strip of your new window and press the key 1 you see

26
 the standard assignments of the current residue type indicated in the window
(by typing 2 this is removed)
- Try the command GetScBMRB 79 in the NMRView console.

To quit the CACB analysis tool, in the cbca window select File->Close

27