NMR Structure Determination: With The NMR Assignments and Molecular Modeling Tools in Hand

NMR Structure Determination
With The NMR Assignments and Molecular Modeling Tools in Hand:

• All we need are the experimental constraints
 Distance constraints between atoms is the primary structure determination factor.
 Dihedral angles are also an important structural constraint
What Structural Information is available

from an NMR spectra?
How is it Obtained?
How is it Interpreted?
NOE
NOE
- a through space correlation (<5Å)
4.1Å
- distance constraint
2.9Å
Coupling Constant (J)
- through bond correlation J NH CaH
- dihedral angle constraint
Chemical Shift CaH NH

- very sensitive to local changes
in environment D
- dihedral angle constraint
Dipolar coupling constants (D)

- bond vector orientation relative
to magnetic field
- alignment with bicelles or viruses
Nuclear Overhauser Effect (NOE)
Nuclear Overhauser Effect (NOE, h) – the change in intensity of an NMR

resonance when the transition of another are perturbed, usually by saturation.
hi = (I-Io)/Io
where Io is thermal equilibrium intensity
Saturation – elimination of a population difference between transitions

(irradiating one transition with a weak RF field)
irradiate
bb N-d
A X
N
ab ba N
X
aa N+d A
Observed signals only occur
Populations and energy levels of a homonuclear from single-quantum transitions
AX system (large chemical shift difference)
Saturated
(equal population)
saturate
bb N-½d
S I
N-½d N+½d
ab ba
I N+½d
aa S Saturated Observed signals only occur
(equal population) from single-quantum transitions
Populations and energy levels immediately
following saturation of the S transitions
N-½d
bb
W1A W1X Relaxation back to equilibrium can occur through:
N-½d W2
N+½d Zero-quantum transitions (W0)
ab ba
W0 Single quantum transitions (W1)
W1X W1A Double quantum transitions (W2)
aa
N+½d
Intensity of NOE “builds-up” as a
function of the mixing time (tm)
N-½d
bb
W1A W2 W1X
N-½d N+½d
ab ba
W0
W1 X
aa W1A
N+½d
Solomon Equation:
X W2  W0
hi  Steady-state NOE enhancement at spin A is
 A 2W1A  W2  W0 a function of all the relaxation pathways
If only W1, no NOE effect at HA

If W0 is dominant, decrease in intensity at HA  negative NOE
If W2 is dominate, increase in intensity at HA  positive NOE
For homonuclear (X=A), maximum enhancement is ~ 50%

For heteronuclear (X=A), maximum enhancement is ~50%(X/A)
Mechanism for Relaxation
• Dipolar coupling between nuclei
 interaction between nuclear magnetic dipoles
– local field at one nucleus is due to the presence of the other
– depends on orientation of the whole molecule
 in solution, rapid motion averages the dipolar interaction
 in crystals, positions are fixed for single molecule, but vary between molecules
leading range of frequencies and broad lines.

• Dipolar coupling, T1 and NOE are related through rotational correlation time (tc)
o
 recall: rotational correlation is the time it takes a molecule to rotate one radian (360 /2p)
tc ~ 10ns (macromolecule)
Relaxation or energy transfers
only occurs if some frequencies
of motion match the frequency
of the energy transition.
tc ~ 10ps (small molecule)
The available frequencies for a
molecule undergoing Brownian
tumbling depends on tc.
The total “power” available for

1/tc relaxation is the total area under
the spectral density function.
Mechanism for Relaxation
• Spectral density is constant for w << 1/tc
 tc decreases, wo also decreases and T1 increases
 at 1/tc ≈ wo there is a point inflection
– W2 falls off first since it it the sum of two transitions
 relaxation rates via dipolar coupling are:
3t c 3t c
W1  6
A
 6
r (1  w At c )
2 2
r
3t c 2t c NOE is dependent on the
W0  6  6
r (w A  w X ) t c )
2 2
r distance (1/r6) separating the
two dipole coupled nuclei
12t c 12t c
W2  6  6
r (1  (w A  w X ) t c )
2 2
r
Important: the effect is time-averaged!
Extreme narrowing limit: 1/tc >>wo then wo2tc2 <<1)

2D NOESY (Nuclear Overhauser Effect)
Relative magnitude of the cross-peak is related to

the distance (1/r6) between the protons (≥ 5Ǻ).
NOE is a relaxation factor that builds-up during
The “mixing-time (tm)
How DO We Go From the NOESY Data to A Structure?
2D NOESY Spectra at 900 MHz Lysozyme Ribbon Diagram

We Need to Assign Each NOE Cross-Peak to A Specific
1H-1H Pair from Assignment Table.
Assignment Table
Resi due N CO Ca Cb Ot her s
D1 120.1 (8.08) 179.1 53.8 (4.37) 39.9 (3.00,0.58)
E2 128.9 (9.93) 176.6 56.0 (4.55) 30.9 (2.11,1.78) Cg 36.2(2.75,2.41)
D3 116.1 (8.90) 176.5 56.2 (4.73) 38.9 (2.70,2.34)
E4 113.3 (7.45) 175.6 52.9 (4.71) 27.2 (1.23,0.63) Cg 34.8(2.57,1.79)
R5 123.8 (7.97) 173.1 54.3 (4.43) 28.9 (1.84,1.47) Cg 26.2(1.59,1.16);Cd 42.6(3.09,3.02)
W6 127.8 (9.27) 177.1 55.4 (5.50) 30.8 (3.11)
T7 109.9 (9.32) 174.8 59.6 (4.85) 71.3 (4.34) Cg 20.2(0.73)
N8 115.2 (8.42) 174.5 50.9 (5.06) 38.7 (3.13,2.70)
N9 116.7 (7.88) 173.3 52.1 (4.94) 38.0 (3.33,3.01)
F10 110.1 (7.57) 177.2 58.2 (4.46) 38.5 (2.63,2.67)
R11 121.2 (8.03) 175.0 55.8 (4.06) 29.7 (1.83,1.67) Cg 27.4(1.45,1.35);Cd 43.1(3.13)
E12 119.0 (8.56) 177.9 52.3 (3.79) 27.4 (1.16,-0.05) Cg 36.6(2.19,1.97)
Y13 122.1 (8.28) 173.2 61.9 (3.80) 41.2 (2.99,2.52)
N14 121.5 (8.45) 175.7 54.7 (4.89) 39.2 (2.26)
L15 127.7 (9.02) 176.9 58.0 (4.48) 41.2 (1.96,1.64) Cg 27.2(1.20);Cd 27.8(0.87);Cd 27.8(0.78)
.
.
.
What H assignments
10.5 ppm to 7.65 ppm
for the protein are
consistent with 10.5
ppm & 7.65 ppm?
In some cases, the chemical shifts associated with the NOE cross-peak are unique and an assignment
is straight-forward. More likely is the occurrence that the assignment is ambiguous  multiple
possible assignments to one or both of the chemical shifts.
We Need to Assign Each NOE Cross-Peak to A Specific
1H-1H Pair from Assignment Table.
Assignment Table
Resi due N CO Ca Cb Ot her s
D1 120.1 (8.08) 179.1 53.8 (4.37) 39.9 (3.00,0.58)
E2 128.9 (9.93) 176.6 56.0 (4.55) 30.9 (2.11,1.78) Cg 36.2(2.75,2.41)
D3 116.1 (8.90) 176.5 56.2 (4.73) 38.9 (2.70,2.34)
E4 113.3 (7.45) 175.6 52.9 (4.71) 27.2 (1.23,0.63) Cg 34.8(2.57,1.79)
R5 123.8 (7.97) 173.1 54.3 (4.43) 28.9 (1.84,1.47) Cg 26.2(1.59,1.16);Cd 42.6(3.09,3.02)
W6 127.8 (9.27) 177.1 55.4 (5.50) 30.8 (3.11)
T7 109.9 (9.32) 174.8 59.6 (4.85) 71.3 (4.34) Cg 20.2(0.73)
N8 115.2 (8.42) 174.5 50.9 (5.06) 38.7 (3.13,2.70)
N9 116.7 (7.88) 173.3 52.1 (4.94) 38.0 (3.33,3.01)
F10 110.1 (7.57) 177.2 58.2 (4.46) 38.5 (2.63,2.67)
R11 121.2 (8.03) 175.0 55.8 (4.06) 29.7 (1.83,1.67) Cg 27.4(1.45,1.35);Cd 43.1(3.13)
E12 119.0 (8.56) 177.9 52.3 (3.79) 27.4 (1.16,-0.05) Cg 36.6(2.19,1.97)
Y13 122.1 (8.28) 173.2 61.9 (3.80) 41.2 (2.99,2.52)
N14 121.5 (8.45) 175.7 54.7 (4.89) 39.2 (2.26)
L15 127.7 (9.02) 176.9 58.0 (4.48) 41.2 (1.96,1.64) Cg 27.2(1.20);Cd 27.8(0.87);Cd 27.8(0.78)
.
.
. To determine the structure, an
assignment has to be made for each of
the thousands of observed NOE cross
peaks  A lot of manual effort!
Ambiguity in assignments also arises from errors in peak position and correlation to assignment table.
Need to “match” assignments with an error range that is dependent on the resolution associated with
each axis. Typically, 1H errors may range from 0.02 to 0.06 ppm and 13C, 15N from 0.25 to 1.0 ppm.
An NOE cross-peak of 4.25 ± 0.02 ppm to 2.21 ± 0.02 ppm may be consistent with a dozen or more
possible combinations!
Peak-Picking is Very Challenging in A Protein
NOESY Spectra
• Spectra Can be very crowded with a number of
overlapping peaks
• Automatic peak-picking fails in these crowded regions
• the quality of the structure is inherently dependent on
the quality of the peak-picking
 assignments are made from the peak lists
 wrong peak position or picked noise peaks results
in a mis-assignment that results in an incorrect

structure distance constraints that results in a local
distortion in the structure
How Many Peaks?

What is each Peak’s Position?
Which are Noise Peaks?
How Do We Resolve These Peak-Picking and Ambiguity Issues?
• Spread out the data into 3D and 4D
 symmetry can help remove ambiguities
• Use 1H-15N and 1H-13C connections
• Iterative analysis of the NOE data
 use structure and distance filter to remove ambiguities
Iterative Analysis of NOESY Data, How Does It Work?
• Assign All unique or unambiguous NOEs
• Calculate Initial Structure with All the Data possible
• Use the Structure to Filter Ambiguous Assignments
 possible assignment has to be ≤ 6 Ǻ
 removes all possible assignments with distances ≥ 6Ǻ
• Calculate New Structure With New constraints

 identify & correct violated constraints
 repeat NOE analysis
 repeat process until all NOEs correctly assigned and a
quality structure is obtained

Using PIPP to Analyze NOESY Data
• read in your initial structure(s)
• read in your NMR assignment list
• click on a peak and PIPP tells you:
 the possible assignments
 chemical shift errors relative to assignment table
 minimum, average and standard deviation distances

After Assigning an NOE Peak to 1H-1H pair, Need to Assign a Distance Constraint
• There are Two Generally Accepted Approaches:
 “Two-Spin” Approximation
 Relaxation Matrix Approach
“Two-Spin” Approximation – the observed volume or intensity of a NOE cross-peak is directly

related to the distance between the 1H-1H pair
1
V (I ) 
r6
This approximation only holds true for the linear part

of the NOE build-up curve (short-mixing times)
when spin diffusion is not a significant component of
the observed volume (intensity)
Spin Diffusion – in the limit of wtc >> 1 (biomolecules), the rate of transfer of the spin energy
between nuclei becomes much larger than the rate of transfer of energy to the lattice.
The observed NOE cross-peak

volume between Hi & Hj is
potentially increased by a
HiHkHj & HiHlHj
energy transfer.
Effectively, the spin energy diffuses through all possible paths between all possible
spin systems. Spin diffusion is always present, the magnitude depends on the mixing
time and the efficiency of any particular pathway. The longer the path, the smaller the
energy that is transferred
As We Discussed Before, One Common Approach Uses a Qualitative Binning of
NOE Intensities
• generally cluster NOE volumes into strong, medium, weak and very weak
• The following rules apply:
Strong 2.5 0.7 0.2  for NH-NH constraints use: 2.5 0.7 0.6
Medium 3.0 1.2 0.3  for NOEs with NH use: 3.0 1.2 0.5
Weak 4.0 2.2 1.0
Very Weak 5.0 2.0 1.0
the lower limit is always set to slightly less than twice the hydrogen
van der Waals radius (1.8Å)
• NOEs for methyls are scaled down by 1/3
• Uses reasonably short mixing-time (100-150msec) and allows quantity of distance constraints to
correct for spin-diffusion effects
 remove or move to lower bin violated constraint that may arise from spin-diffusion
 alternative is to leave all constraints as observed with corresponding contribution to
violation energy and potential structure distortion  point of debate in NMR community
NOEs are observed between the Ala methyl

and both Leu methyls. Structure indicates a
violation to one methyl-methyl pair  NOE
probably a result of spin diffusion
Violated constraint
Two-Spin Approximation
• Instead of binning into strong, medium, weak and very weak, can assign a relative distance
• Use a reference volume with a known fixed distance to calibrate all volumes
 2.52 Å for Leu Ha-Hb
 2.52 Å for Phe or Tyr Hd-He
Serious Problems: obtain a highly precise distance that ignores all the inherent errors associated
with the accuracy of measuring volumes, spin-diffusion and dynamics.
Short, inaccurate distance constraints cause severe local structure distortions
What would happen to the

structure if both Leu d had to
be 3Å to the Ala b?
Violated constraint
Relaxation Matrix Approach Takes Into Account Spin-Diffusion
• removes manual and potentially biased approach to identify spin-diffusion issues
 From the structure, calculates a spin relaxation matrix to correct for spin-diffusion
contributions to observed volumes
 There are a number of programs that perform this analysis (CORMA, FIRM, MardiGras,
MORASS, etc)
Calculate from Experimental

structure Volumes from
NOESY spectra
J. Am. Chem. Soc. 1990, 112, 6803-6809

Relaxation Matrix Approach Takes Into Account Spin-Diffusion
• Merge the cross-relaxation rates calculated from the structure with experimental volumes
 obtain distance matrix to calculate new structure
 iterate process
Can relate cross-relaxation rates(G) with experimental NOE volumes (V)

Problems with Relaxation Matrix Approach
• Errors and Failures with Matrix Calculations
 Do not obtain complete experimental NOESY volume matrix
Very difficult to accurately measure diagonal peaks
 Significant errors in measuring experimental volumes
 Peak overlaps and degenerate assignments
 Missing peaks, limits of S/N
 Noise
• Assume Uniform Dynamics (tc)

Poor Assumption
 Different regions of protein structure have very different local dynamics
 Contributions to relaxation rates by dynamics can be much more significant than
spin-diffusion
• Output of calculation is a very specific distance constraint
 But, may have high errors (volume, dynamics)
 Large structure distortion that is propagated through iteration

Two Very Important Facts to Remember
• NOEs Reflect the Average Distance
• Protein Structures Are Dynamic
In reality, protein undergoes wide-ranges

of motions (snapshots of 100 BPTI
conformations)
We visualize protein
structures as a static image
J. Mol. Biol. (1999) 285, 727±740
2D, 13C, 15N Labeled Protein
We have already
discussed labeling the
protein, data
collection and the
resonance assignment
NMR Data Collection
Next Step, is to identify which residues
adopt which secondary structure present
Resonance Assignments Secondary Structure
X-ray or Homology Model
C.S
Low Resolution Structure Dock Ligands
NOE
High Resolution Structure

Protein Secondary Structure and NOE Patterns
• a-Helix
Sequential NOEs observed in 3D 15N-edited NOESY are indicative of a-helix
• a-Helix
• a-Helix
• a-Helix
• a-Helix
• b-Sheet
Across strand NOEs observed in 3D 15N-edited NOESY and 3D 13C-edited NOESY are
indicative of b-sheet
• b-Sheet
• b-Sheet
• Turns
Sequential NOEs observed in 3D 15N-edited NOESY are indicative of turns
 Similar to a-helix, shorter amino acid stretches connect b-strands
• Turns
Sequential NOEs observed in 3D 15N-edited NOESY are indicative of turns
 Similar to a-helix, shorter amino acid stretches connect b-strands
Protein Secondary Structure and Carbon Chemical Shifts
Chemical shift differences

between Ca,Cb random-coil
values and experimentally
observed values yields
secondary structure chemical
shift.
Helix:
DCa ~ 3 ppm
DCb ~ -1 ppm
b-strand:
DCa ~ -2 ppm
DCb ~ 3 ppm
a1
bIV
bIII
a2 a3 bI
bII
a4
• TALOS +
Shen et al. (2009) J. Biomol NMR 44:213

• TALOS+
 Given the Ca, Cb Chemical shift assignments and primary sequence
 Compares the secondary chemical shifts against database of chemical shifts and associated
high-resolution structure
 comparison based on “triplet” of amino acid sequences present in database structures
with similar chemical shifts and secondary structure

 Provides potential f , y backbone torsion constraints
 Issues: May not provide a unique solution, two or more sets of f , y are possible. Can not
initially use TALOS results if ambiguous. Can add constraint latter if consistent with
structure.
• TALOS+
TALOS may provide relatively tight

error bounds associated with the
predicted f,y.
It is better being more conservative by

using minimal errors of:
f ± 30
y ± 50
c ± 20
Protein Secondary Structure and 3JHNa
• Karplus relationship between f and 3JHNa
 f =180o  3JHNa  ~8-10 Hz  b-strand
 f = -60  JHNa = ~3-4 Hz  a-helix
o 3
Vuister & Bax (1993) J. Am.Chem. Soc. 115:7772

Protein Secondary Structure and 3JHNa
• Karplus relationship between f and 3JHNa
 Measure 3JHNa for a protein using HNHA
 Ratio of cross-peak to diagonal intensity
yields coupling constant
 Common approach to measure coupling
constants in complex protein NMR spectra
J. Am. Chem. Soc. 1993,115, 7772-7777

Protein Secondary Structure and NH Exchange Rates
• Relatively Slower Exchanging NHs are involved in Hydrogen Bonds or Buried
 Hydrogen Bonds are an important component of secondary structures
NH Exchange Rate
Protein Secondary Structure and NH Exchange
Rates
• Measure Relatively Slow Exchanging NHs
Exchange NMR sample into 100% D2O and
measure series of 1H-15N HSQC as a function of
time.
• Observed slowly exchanging NHs are
correlated with secondary structure.
Secondary structures identified from NOEs,
chemical shifts and 3JHNa
 Assign hydrogen-bond distance constraint
J. Mol. Biol. (1997) 271, 472-487

• Measure Fast Exchanging NHs
 CLEANEX-PM experiment selects H2O
 amount of exchange depends on mixing time NH exchange with H2O
during spin-lock
 Missing Peaks are slowly exchanging NHs
• Observed slowly exchanging NHs are

correlated with secondary structure.
 Secondary structures identified from NOEs,
chemical shifts and 3JHNa
 Assign hydrogen-bond distance constraint
2D 1H-15N HSQC CLEANEX-PM
J. Am. Chem. Soc. (1997) 119, 6203

J. Biomol. NMR (1998) 11:221
• Measure Fast Exchanging NHs
 CLEANEX-PM experiment
 amount of exchange depends on mixing time
– Can measure exchange rates by varying mixing time

 Missing Peaks are slowly exchanging NHs
• Assign Hydrogen-Bond Distance Constraints for Slowly Exchanging NHs
For a –helix hydrogen bond constraints:

constraint between Oi & Ni+4 2.8 0.4 0.5
constraint between Oi & HNi+4 1.8 0.3 0.5
For b –sheet hydrogen bond constraints are

Across strands:
constraint between Oi & Nj 2.8 0.4 0.5
constraint between Oi & HNj 1.8 0.3 0.5
Protein Secondary Structure Summary
• Secondary Structure NOEs, slowly exchanging NHs, 3JNHa and secondary
structure chemical shifts provide the foundation for detemining a protein NMR
structure
Protein Tertiary Structure Determination
• Include all the various structural information, disulphide bonds, hydrogen
bonds, dihedral angles, chemical shifts, coupling constants, and NOES (distance
constraints) to calculate (FOLD) the structure using simulated annealing
Depiction of short-range and long range NOEs Anal. Chem. 1990, 62, 2-15
Simply Need to Complete the Assignment of All the Remaining NOEs in an
Iterative Process to Obtain the Structure
Automated Protein Structure Determination
• A number of efforts are on-going to automate the process (ARIA,
AutoStructure, CYANA, Rosetta, etc)
 From assignment tables, NOE peak lists, coupling constants and slowly
exchanging NHs.
Improving the Quality of NMR Structures
• Stereospecific Assignments
Assign Hb1 and Hb2 chemical shifts and determine c1 dihedral angle
Assign Hb1 and Hb2 chemical shifts and
determine c1 dihedral angle.
 Assign Val H1 and H2 chemical shifts and

 HNHB and HN(CO)HB experiments
 Relative intensities of Hb cross-peaks
determines c1 and stereospecific assignment
Can include c1 dihedral

constraints and replace HB#
pseudoatoms with specific
distance constraints.
HNHB
Pseudoatoms require weaker
upper bound constraints that
can be tighten with specific
assignment HN(CO)HB
Improving the Quality of NMR
Structures
Assign Leu Hd1 and Hd2 chemical
shifts and determine c2 dihedral
angle.
 based on c1 assignments
 Assign Ile c2 dihedral angle.

Assign Leu Hd1 and Hd2 chemical shifts and
 Assign Ile c2 dihedral angle.
 Long Range
13C-13C Correlation
 Relative intensities of Hd cross-peaks
compared to diagonal determines 3JCaCd

 Assign Hd1, Hd2 and He1,He2 chemical shifts and determine c2
dihedral angle for Phe and Tyr.
 Inferred assignment from structure.
 Phe and Tyr undergo rapid ring flip so Hd1, Hd2 and He1,He2
are equivalent  degenerate chemical shifts

 c2 should be 90 or -90 , which are symmetrically equivalent
o o
Based on which (90,-90) c2 is closer

to the observed structure, add that c2
constraint and replace Hd#, He#
Hd1 pseudoatoms with stereoassignments
He1
that are consistent with the structure.
 Making stereospecific assignments increase the relative number of distance
constraints while also tightening the upper bounds of the constraints
 There is a direct correlation between the quality of the NMR structure and the
number of distance constraints

 more constraints  higher the precision of the structure
Increasing Number of NOE Based Constraints

• Dipolar Coupling Constants
 Solid state NMR yields a powder pattern or a ensemble of
chemical shifts due to different orientations relative to the
magnetic field.
 CSA –chemical shift anisotropy
 In liquids, isotropic tumbling averages the various
orientations to zero
 Simulated in a solid by spinning the sample
w  ½(3cos2q-1)
Magic Angle (54.7o)  averages to zero
• Residual Dipolar Coupling Constants (RDC)
 In liquids, dipolar coupling can become non-zero if isotropic tumbling is removed.
 proteins can be aligned by using bicelles, virus particles or polyacrylamide gels.
 mechanically impede the random tumbling of protein
 do not want and interaction between the protein and alignment media
 RDC can be measured for each bond vector in the protein
 RDCs are measured from a difference in the coupling constants between
aligned and unaligned media.

 Different RDCs can be normalized based on ratios of bond lengths
D AB ( NH )  D AB ( N  H  rNH
3
 /  A B  rAB
3
A )
unaligned aligned
The magnitude of RDC depends on the extent of the
DAB(NH) = 93.3 Hz – 100.2 Hz = - 6.9 Hz alignment of the protein and the orientation of the bond
vector to BO
Alignment frame Orientation of

of the protein bond-vector in
protein alignment
DAB(q,f) = Da(3cos2q- 1) + 3/2 Dr(sin2q cos2f)]
where Da and Dr are the axial (**) and rhombic (2) components of the traceless diagonal
tensor D given by 1/3[DZZ - (Dxx+ Dyy)/2] and 1/3(Dxx- Dyy), with Dzz > Dyy  Dxx
Da and Dr measure the magnitude (intensity) of the alignment.
q is the angle between the bond and the z-axis of the principal alignment frame (x, y, z)
f is the angle between the bond's projection in the x-y plane and the x-axis
To use DAB in a structure calculation, need to know Da and Dr
• Residual Dipolar Coupling Constants (RDC) Dxx most populated
value in histogram
Recreate “Powder
Pattern” by histogram plot
of all normalized RDCs
Dyy average of the Dzz average of the

low extreme values high extreme values
of RDC of RDC
Extremes of the histogram correspond to the alignment tensor components D xx, Dyy, and Dzz
Dxx+Dyy+Dzz =0
Dzz = 2Da
R is the rhombicity given by Dr/Da
Dyy = - Da(1 + 3/2 R)
Dxx = - Da(1 - 3/2R)
- use different normalized sets of RDCs (CH, NH, CaC’, NC , NC’, etc) to create one histogram , etc)
- exclude residues with substantially lower order parameter than average
 RDC provide long-range (> 5 Å) interaction
 RDCs identify angular orientation of bond vector to alignment axis
 No translational orientation of bond vector.
 RDC alone can not define a protein structure
 NOEs are necessary to define translational orientation of bond vectors
 Absolute alignment axis is unknown
 RDC are all relative to an alignment axis
 Allow the structure and RDC to be refined against a coordinate position that
“floats” or changes during the simulation
Refine RDCs against coordinate vector Observed RDC only limits bond vector
position that is free to move during to taco shaped curve on sphere
Without RDCs With RDCs
 Addition of RDCs do not induce dramatic
changes in structure
 improves relative orientation/packing
 very valuable for proper alignment of complexes
Without RDCs With RDCs
Consistency of Structures with Experimental RDCs
Protein Science (2004), 13:549-554

• Water Refinement
 protein structures generally calculated in vacuum.
 water has a significant effect on protein structures
 explicit solvent model
–MD simulation in box of water

– box > 10 Å, keep solvent from edge
– 1000 to 10,000s water molecule
– Computationally expensive not usually done
 implicit solvent model

– treat solvent as a high dielectric continuous
medium (e) around protein
– significantly faster
– different solvation models  generalized Born model (GB)
where :
where q – point charge , r is the distance between the

charges and a is the radius of ion, eo dielectric constant
for a vacuum, e of solvent
http://cmm.cit.nih.gov/intro_simulation/node8.html
 generalized Born model (GB) has been implemented in Xplor
 compare structures in vacuum to water
– no visible difference
Xia et al. (1996) J. Biomol. NMR 22:317

 subtle, but significant improvements
 compare structures in vacuum to water
– improves NH to CO hydrogen bonds

– improves f and y angle distributions
 sample Xplor script
.
.
.
topology
Read in the GB parameter and @TOPPAR:topamber.inp
topology files (partial charges) @TOPPAR:amberpatches.pro
end
parameter @TOPPAR:paramber.gb.inp end
segment
name="ASN1"
molecule name=ASN number=1 end
end
patch NASN refe=nil=(resid 1) end
patch CASN refe=nil=(resid 1) end
vector do (charge = charge + .2306) (name ht%)
vector show sum (charge) (all)
Set N- and C-terminal
charges for protein segment
name="ASN1"
molecule name=ASN number=1 end
end
patch NTER refe="+"=(resid 1) end
patch CTER refe="-"=(resid 1) end
vector
.
show sum (charge) (all)
.
.
 sample Xplor script (continued)
.
.
.
parameter
nbonds
atom cdie trunc
Set standard electrostatic e14fac=0.8333333 ! use this to reproduce amber elec
parameters cutnb 500. ctonnb 480. ctofnb 490. ! essentially no cutoff
tolerance=100. ! only build the nonbonded list once
nbxmod 5 vswitch
wmin=1.0
end
end
parameters nbonds
Set GB parameters EPS=1. WEPS=80. offset=0.09 GBHCT ! GB parameters
end
End
@TOPPAR:volumes.amber
Read in volumens and vector do (rmsd = rmsd * 0.9) (all) ! reduce volumes by 10%
flags include gbse gbin end
tell Xplor to use GB in
parameter reduce selection=(all) overwrite=true mode=average end end
calculation
energy end
.
.
.
 sample Xplor script (continued)
.
.
.
! weakly restrain Calpha position
Weakly restrain Ca vector do (harmonic = 0.5) (name ca)
position during dynamics coor disp=reference @asn.-80.10.pdb
constraints harmonic end
Don’t allow large structural flags include harmonic end
.
changes. Just make local .
changes to fix problems .

NMR Structure Determination: With The NMR Assignments and Molecular Modeling Tools in Hand

Uploaded by

Copyright:

Available Formats

NMR Structure Determination: With The NMR Assignments and Molecular Modeling Tools in Hand

Uploaded by

Document Information

Original Description:

Original Title

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

NMR Structure Determination: With The NMR Assignments and Molecular Modeling Tools in Hand

Uploaded by

Copyright:

Available Formats

NMR Structure Determination

With The NMR Assignments and Molecular Modeling Tools in Hand:

What Structural Information is available

Chemical Shift CaH NH

Dipolar coupling constants (D)

Nuclear Overhauser Effect (NOE, h) – the change in intensity of an NMR

Saturation – elimination of a population difference between transitions

If only W1, no NOE effect at HA

For homonuclear (X=A), maximum enhancement is ~ 50%

– depends on orientation of the whole molecule

 in solution, rapid motion averages the dipolar interaction

leading range of frequencies and broad lines.

The total “power” available for

– W2 falls off first since it it the sum of two transitions

 relaxation rates via dipolar coupling are:

Extreme narrowing limit: 1/tc >>wo then wo2tc2 <<1)

Relative magnitude of the cross-peak is related to

2D NOESY Spectra at 900 MHz Lysozyme Ribbon Diagram

 wrong peak position or picked noise peaks results

in a mis-assignment that results in an incorrect

How Many Peaks?

• Calculate New Structure With New constraints

 repeat NOE analysis

 repeat process until all NOEs correctly assigned and a

quality structure is obtained

 minimum, average and standard deviation distances

“Two-Spin” Approximation – the observed volume or intensity of a NOE cross-peak is directly

This approximation only holds true for the linear part

The observed NOE cross-peak

 alternative is to leave all constraints as observed with corresponding contribution to

NOEs are observed between the Ala methyl

Short, inaccurate distance constraints cause severe local structure distortions

What would happen to the

Calculate from Experimental

J. Am. Chem. Soc. 1990, 112, 6803-6809

Can relate cross-relaxation rates(G) with experimental NOE volumes (V)

 Significant errors in measuring experimental volumes

 Peak overlaps and degenerate assignments

 Missing peaks, limits of S/N

• Assume Uniform Dynamics (tc)

 Different regions of protein structure have very different local dynamics

 Contributions to relaxation rates by dynamics can be much more significant than

 Large structure distortion that is propagated through iteration

In reality, protein undergoes wide-ranges

Resonance Assignments Secondary Structure

X-ray or Homology Model

High Resolution Structure

Chemical shift differences

Shen et al. (2009) J. Biomol NMR 44:213

with similar chemical shifts and secondary structure

TALOS may provide relatively tight

It is better being more conservative by

Vuister & Bax (1993) J. Am.Chem. Soc. 115:7772

constants in complex protein NMR spectra

J. Am. Chem. Soc. 1993,115, 7772-7777

J. Mol. Biol. (1997) 271, 472-487

• Observed slowly exchanging NHs are

2D 1H-15N HSQC CLEANEX-PM