Hello, what NMR experiments can detect hydrogen bonds directly?
Is there anything that can work with non-labeled samples at the natural abundance of isotopes?
We have a cyclic peptide that might form a beta-sheet made of two anti-parallel strands.
The problem is that it is made of two pieces of the same sequence closed into a cycle and in such arrangement NOE's from the "middle" part (if it is indeed forming two anti-parallel beta-strands) will be hard to distinguish as being intra- or inter-strand. That's why I think direct observation of hydrogen bonds would be very helpful.
Hydrogen bonded amide NH have a slower rate of H/D exchange with solvent compared to non hydrogen bonded amide NH.
The magnitude of the temperature dependence of the chemical shift is smaller for hydrogen bonded amide NH than for those that are not hydrogen bonded.
Direct detection of Hydrogen bonds: The H(N)CO-TROSY experiment has been used to detect scalar coupling across hydrogen bonds. However, it requires isotope labeling.
Speculation: An, E.COSY type experiment correlating proton (HN-selective) with carbon (Carbonyl-selective) with a long tau delay (optimized to minimize the signal from the sequential two bond and intra residue three bond HN and C' coupling) on unlabeled peptide MAY be able to detect scalar coupling across hydrogen bonds.
answered Mar 14 '10 at 09:43
There are two other ways that come to mind for establishing hydrogen bondedness and h-bond strength, though they don't identify H-bonded partners.
First is measuring H/D fractionation factors. Without going into great detail, one integrates the putatively H-bonded hydrogen in a series of H2O/D2O mixtures. If the hydrogen is only h-bonded with solvent, it will integrate to 0.5 in a 50/50 H2O/D2O mixture.
If the H-bond is strong, the hydrogen will be selectively enriched because of differential zero point energies. I haven't kept up on this literature, but you can check out the following references on the subject. It was a hot subject in Madison for awhile.
Second, there's this old reference that you may find particularly useful:
The main message is that if you compare 1H and 2H spectra of hydrogen-bonded systems, the chemical shifts of the signals will differ in proportion to their h-bond strength. In principle, if you took 1D 1H and 2H spectra of your peptide in 50/50 H/D2O, you'd observe slight differences between the frequencies of the same residue's amide "H" depending on whether it was 1H or 2H. Never tried it myself, but thought it might be useful.
The following more recent references may also be of interest with regard to direct detection of H-bonded 15N and 13C scalar coupling: