hbond geometry in nmr structures
stein at microvirus.chem.tufts.edu
Tue Mar 18 13:49:57 EST 1997
a duplex of a small, backbone-modifed dinucleotide, still carrying two
protecting groups, in chloroform is what I'm working on for quite a
while now. The problem is that the structure that I got out of the well
documented calculation pathways, e.g. "sa.inp - dgsa.inp - refine.inp",
does not fully explain the obvious proton chemical shift shifts that we
see in the 1D spectrum and that we think are due to some unusual hbonds
in the molecule. So, although the r-factor after a backcalculation falls
well below 0.1, we still do not "see" the hbonds in refined structure,
which is almost certainly because the nmr data is too "fuzzy" to really
lock a particular proton at exactly the position where it is supposed to
be in an ideal or even close-to-ideal hbond geometry.
My question is: How is this usually done in NMR 3D-structure
elucidation? Is it valid to say: "Ok, the heavy atom of the hbond donor
group and the heavy acceptor atom are close to each other in space and I
could move my proton in position by e.g. rotating the OH-group around
the -C-O- bonds - that's enough of an evidence that there is an hbond"?
Or are there any means to take inter-molecular hbonds into account
during the calculation?
In this particular case I think I'm dealing with hbonds of ribose
OH-groups and one non-Watson-Crick hbond of a one of the bases.
Dr. Christoph Steinbeck - PGP available on my homepage
mailto:stein at microvirus.chem.tufts.edu -
Dep. of Chemistry, Tufts Univ., 62 Talbot Ave., Medford, MA 02155, USA
Phone: int + (617) 627-3881 FAX: int + (617) 627-3443
*** Windows 95 - from the people who brought you edlin ***
More information about the X-plor