How To Convert Pdb Structure Of Protein To Equivalent Hsqc Spectra?

Question

[Pappu]

Is it possible to convert PDB structure of a protein to [15N, 1H] HSQC spectra?

Answer 1

[Jam]

Well it would to easy to use a chemical shift prediction program like SPARTA+ or SHIFTX2, and that would give you predicted 1H and 15N shifts which you could plot like an HSQC, though it won't give you any information on linewidth or peak intensity. The predictions are of ok quality, though 1Hn and 15N shifts don't carry that much structural information.

[Neilfws]

+1 for alerting me to the existence of these software. However, I'm very dubious as to how good the predictions would be.

[Jam]

They can be really good for 13Ca, 13Cb and 1Ha, which correlate quite strongly with secondary structure. 15N predictions can be ok, but are not very useful from a structural perspective since they are primarily determined just be the local sequence environment (ie preceding and following amino acids). The accuracy of chemical shift predictions is what has enabled a lot of the advances in protein structure calculations using only chemical shifts in recent years (e.g. CS-ROSETTA)

[Neilfws]

Interesting; thanks for the info.

Answer 2

[Neilfws]

No it is not.

Think about it. A PDB file contains atomic coordinates derived from an experimental technique, namely x-ray diffraction using crystals of the molecule. NMR spectra represent a completely different experimental technique: magnetic resonance measurements of molecules in solution.

There's no reason why one type of file would contain information from the other, which is what would be required to convert between them. And there's no way to predict what complex NMR spectra (e.g. from proteins) would look like, given known atomic coordinates.

EDIT: well, apparently software does exist to predict NMR spectra from PDB coordinates - see other answer from JAM. I'd be amazed if it were any good.

[Gareth Morgan]

PDB files can contain structural data derived from any source, not just crystallography. Although the shift data is not in there directly, it is possible in theory (and increasingly in practice) to back-calculate shifts from structures.

[Neilfws]

You're right; data from NMR experiments can be in PDB file format. Guess I assumed that the question was about conversion from crystallographic data. Almost tempted to delete my poor answer, but I'll leave it as a reminder to think before posting!

Answer 3

[Gareth Morgan]

If the structure was determined by NMR then the shifts might (should...) be in the Biological Magnetic Resonance Bank: http://www.bmrb.wisc.edu/