Subject: a biomolecular NMR protein structure protocol, any suggestion is welcome
From: ethernet3399
Date: Nov 16 7:05 PM
Protein structure calculation streamline operation protocol and notes
sections:
1.NMR buffer
2.spectrums
3.processing (NMRPipe and Sparky)
4.peak picking (Sparky)
5.backbone assignment (Sparky and MARS)
6.sidechain assignment and NOE assignment (Sparky)
7.NOE analysis and structure calculation (ARIA and CNS)
Section 1 NMR buffer
The common phosphate salt buffer used in our lab:
20mM phosphate salt(Na2HPO4 and NaH2PO4),
50mM NaCl
pH6.5
0.01% NaN3
DTT is optional depends on the protein property
the above solution is 90% in the final NMR buffer and D2O is 10%.
this buffer is used when the backbone amino-hydrogen is needed( almost
all the backbone assignment spectrum). While in 13C sample and record
side chain spectrum(such as: HCCH-TOCSY, C13-HSQC-NOESY), we should
solve the sample in 100% D2O for less solvent peak disturbing.
Section 2 spectrum
backbone assignment:
N15-HSQC
HNCACB
CBCA(CO)NH
HNCO
HN(CA)CO
C(CO)NH
HNHA (optional)
HA(CO)NH (optional)
HNCA (optional)
CA(CO)NH (optional)
side chain assignment:
HCCH-TOCSY
N15-HSQC-NOESY
C13-HSQC-NOESY
Section 3 processing
use NMRPipe some sample file as a reference to process 2D and 3D spectrum
NOTE 1: if some axis is upside down, add
NMRPipe -fn REV
to this axis to make it upside up.
NOTE 2: when export HCCH-TOCSY and C13-HSQC-NOESY to sparky(ucsf)
format, the axis order re-arrange parameter should be used. Because
HCCH-TOCSYs two Hydrogen order is different with other spectrum, or
at least the C13-HSQC-NOESY, while these two spectrum must be aligned
to use side chain assignment to help NOE assignment. If use
pipe2ucsf tools normally, these two spectrum is weird for alignment.
So heres the suggestion:
pipe2ucsf -213 hcch-tocsy.ft hcch-tocsy.ucsf
the parameter -213 can tell the command pipe2ucsf to re-arrange
the axis order. Pre-change the nmrpipe output xxx%03d.ft files axis
order is useless for sparky, though useful to nmrview.
NOTE 3: the function LP (linear prediction) is usually placed in the
last axis, for this functions condition is other axis has already do
the FT. If you want to do two axis LP, its more complicated, read the
NMRPipe LP help carefully.
NOTE 4: when use EXT function with PS(phase adjustment), let EXT in
PSs former is better. For when you change the PS, then EXT, you see
the half spectrum, then use the nmrDraw to adjust the phase with the
half spectrum, that maybe different with whole spectrum, so there
maybe something wrong. So the clever way is put the EXT before PS
operation. Like this:
nmrPipe -fn EXT -left -sw
|nmrPipe -fn PS -p0 0 -p1 0 -di
NOTE 5: the sparky can project a 3D spectrum to a 2D spectrum, there
are two way:
ucsfdata -p1 -o hsqc.ucsf N15-hsqc-noesy.ucsf
or
ucsfdata -s1 256 -r -o hsqc.ucsf N15-hsqc-noesy.ucsf
the result is same, while the former way is faster.
Section 4 peak picking
Read the sparky manual first. Do the peak picking following the
sparkys manual, get the atoms chemical shifts. Maybe frequently used
commands:
ct set contour levels and colors
zo zoom out
zf zoom full
zi zoom in
at assign the select peak
lt show a list of peaks for a spectrum
more commands referred to the sparky manual.
From N15-hsqc and cbca(co)nh the N and Qs side chains amino hydrogen
chemical shifts can be identified.
These are N, Qs side chain graph. In cbca(co)nh spectrum, this part
can be found following the H1-N, H2-Ns peak in N15-hsqc to the
cbca(co)nh. While to N, is Ca-Cb-(CO)-NH1, Ca-Cb-(CO)-NH2. And for Q,
is CB-CG-(CO)-NH1, CB-CG-(CO)-NH2. So compare with N, Qs CaCb or CBCG
chemical shifts, these NH1, NH2 can be identified. These amino
Hydrogen maybe helpful in later NOE assignment.
Section 5 backbone assignment
use the CA, CA-1; CB, CB-1; CO, CO-1 to found the sequential residue,
then line these residues up. As a rule of thumb, CA, CBs chemical
shifts can indicate the residues type.
CA, CB, CO is the ith residue alpha, beta, and C carbon, while
CA-1, CB-1, CO-1 are the i-1th residue alpha, beta, and C carbon.
The program MARS can help, but remember to check it out manually. Do
not pay more trust to computer program than they should receive.
Compare the sequence and residue type with the proteins first
sequence to check.
Note: because in N15-HSQC the residue PRO is not appear, so in these
backbone spectrum, no PROs info can be found. While in C(CO)NH
spectrum, if a residue is following PRO, this PROs side chains
carbon chemical shifts can be found.
Section 6 Side chain assignment
From the HNHA spectrum, HAs chemical shifts can be got. And from
C(CO)NH, the residues side chain carbon chemical shifts can be got.
So in HCCH-TOCSY, other side chains hydrogen chemical shifts can be
identified. The rule is supposed we have know the CA, HA, CBs
chemical shifts, now we can find HB2, HB3s chemical shifts in
HCCH-TOCSY spectrum.
If we havent c(co)nh spectrum, we have to pick all the HA-CA-CB-HB,
HB-CB-CG-HG, HG-CG-CD-HD et als peaks, extract these
information(chemical shifts), then re-arrange these partial
information to a whole residue. This is practical, though more difficulty.
And from N15-hsqc-NOESY, some residues side chain Hydrogen will
appear too, so it can be a reference.
Section 7 NOE analysis and structure calculation
1.sparky xe command can export the resonance shifts and peak list as
the xeasy format. And use a small awk script to filter and adjust the
format more like xeasy.
2.Prepare the sequence file(xxx.seq).
3.Fill in proper string in the file conversion.xml
4.begin conversion by aria2 –convert conversion.xml, remember all
the files needed by conversion in the same directory.
5.Copy the noe_spectrums xml file and sequence xml file to the
directory where aria2 is supposed to run.
6.Use aria2 -g project.xml to fill something or adapt something,
save it.
7.Use aria2 -s project.xml to setup(copy some related files to the
proper directory)
8.use aria2 project.xml to run aria2 structure calculation.