Dear Bikash,

Every method has its utility and the domain of results that has a unique advantage. Precisely for these reasons the method is used. Solid State NMR is more difficult than solution NMR. Both methods cannot be even compared to the maximum model fidelity and accuracy that can be obtained in high resolution X-ray crystallography. However, both of them have significant advantages.

1) They do not need a crystal.

2) They can use a variable concentration (even though some uses would require an almost equal to that in crystallography).

3) They can use much more variable material in terms of purity (even though the purity counts)

4) They are much more flexible in presence of additives such as in complexes, membrane presence etc.

5) Solid state can imitate some native conditions such as in integral membrane proteins placed in a bilayer.

6) It can estimate better the internal mobility.

7) Better suitable for functional studies (kinetics, activation etc.)

Unfortunately, the general methodology of the NMR does not allow it to reach the fidelity and accuracy of X-ray crystallography.

1) A protein of a size ~10KDa would produce around 10000 reflections (data points) in a X-ray experiment. The same size protein would produce around 1000 data points in the NMR experiment (interpretable and assigned resonances).

2) Constraints are very well investigated and applied in X-ray crystallography.

In NMR methods we have to rely on MD forcefields or crude forcefield approximations.

3) In X-ray the data is very well converted to the structural features. In the NMR there is not even a consensus what should be recognized as usable data.

4) By definition (Fourier transform) the X-ray crystallography is a global technique.

In contrast the NMR is by definition a local technique (proximity measures) that suffer when distant points are being involved.

In essence, the answer is NO but we have sufficient advantages and new sources of information that will use it for foreseeable future.

Dear Bikash,

Every method has its utility and the domain of results that has a unique advantage. Precisely for these reasons the method is used. Solid State NMR is more difficult than solution NMR. Both methods cannot be even compared to the maximum model fidelity and accuracy that can be obtained in high resolution X-ray crystallography. However, both of them have significant advantages.

1) They do not need a crystal.

2) They can use a variable concentration (even though some uses would require an almost equal to that in crystallography).

3) They can use much more variable material in terms of purity (even though the purity counts)

4) They are much more flexible in presence of additives such as in complexes, membrane presence etc.

5) Solid state can imitate some native conditions such as in integral membrane proteins placed in a bilayer.

6) It can estimate better the internal mobility.

7) Better suitable for functional studies (kinetics, activation etc.)

Unfortunately, the general methodology of the NMR does not allow it to reach the fidelity and accuracy of X-ray crystallography.

1) A protein of a size ~10KDa would produce around 10000 reflections (data points) in a X-ray experiment. The same size protein would produce around 1000 data points in the NMR experiment (interpretable and assigned resonances).

2) Constraints are very well investigated and applied in X-ray crystallography.

In NMR methods we have to rely on MD forcefields or crude forcefield approximations.

3) In X-ray the data is very well converted to the structural features. In the NMR there is not even a consensus what should be recognized as usable data.

4) By definition (Fourier transform) the X-ray crystallography is a global technique.

In contrast the NMR is by definition a local technique (proximity measures) that suffer when distant points are being involved.

In essence, the answer is NO but we have sufficient advantages and new sources of information that will use it for foreseeable future.

Excellent post, but I would quibble with a few things. The very best solid state structures actually have a resolution comparable to the best x-ray structures; however, it is every difficult to get the number of constraints necessary for that kind of resolution on a typical sample.

Also, while in principle solid-state can preserve native-like conditions, in practice many of the samples are not near native conditions (powders, microcrystals, and low lipid-to-protein ratios for membrane proteins).

http://www.rcsb.org/pdb/explore/explore.do?structureId=2lgi

Dear Jeff,

Thank you for your post. I will study this paper as I was unaware of such quantal progress in this field. I know the past attempts but not their final result.

Can you get closer into the definition of resolution in this case? I think you mean that the geometry is comparable to ...

I think that the word resolution is used in the paper as a marketing trick rather than carefully defined entity.

By the way in high resolution crystallography (the highest reported resolution is around 0.44Å) the accuracy of positions in individual atoms would be of the order of 0.005Å. When I go through the constraints for this structure (from the PDB) I see traditional strong, medium, weak, assignments. Without quantitation and introducing continuous scale the NMR will never be able to compete with accuracy of X-ray. There are at least three other reasons why the NMR is lagging methodologically. We can talk about it off line.

There are several differences between solid-state NMR and solution NMR to consider:

1. The individual distance constraints obtained by dipolar recoupling experiments in solid-state NMR are much more accurate than corresponding the individual NOESY constraints in solution NMR

2. Orientational constraints from CSA tensor correlations in solid-state NMR can give the dihedral angle directly. J-coupling measurements and chemical shifts can give similar information in solution NMR , but again the measurement is more precise for solid-state NMR.

Solid-state NMR therefore gives fairly precise individual constraints in comparison to solution NMR. The problem is the number of constraints is usually very few. The paper shows that by using uniformly labeled sample you can gather more constraints, enough to get a structure that is comparable to a good (probably not the best) X-ray structure.

I agree with you that resolution is not defined in NMR. As you mention, the structure is obtained by finding a geometry consistent with a number of local constraints. In many cases, it is difficult to tell if differences in the backbone RMSD results from actual different conformations in the structure or from imprecisions in the structure determination. Certain relaxation experiments help in making this determination.

Hi.  NMR structures have advantages over X Ray and vice versa as well  , apart from the crystal aspect. Read this paper fron Baker Lab if you are working in NMR field:

Mao, B., Tejero R., Baker D., & Montelione G. T. (2014).  Protein NMR Structures Refined with Rosetta Have Higher Accuracy Relative to Corresponding X-ray Crystal Structures.. Journal of the American Chemical Society. Abstract  Download: Mao_JACS_2014.pdf (2.16 MB)

@Sumit

The title of the paper is very confusing and not reflective of the actual results. It should be

 Protein NMR Structures Refined with Rosetta Have Higher Accuracy Relative to Corresponding X-ray Crystal Structures compared to standard refinement protocols

... because Rosetta provides more and better HEURISTIC constraints.

One small step for Baker, large step for humanity dealing with NMR.

Hype,  hype , hype.

The main problem is that Rosetta's selling point is that it does not need any real data. The authors and the entire entourage would like to claim that the success rate is around 50% while in reality is closer to around 10%. Everything depends how one constructs the measures of success.