Greetings - especially Paul Schlosser and Anurag Khetan
My first ever bionet post, so forgive me a blatant plug for my own paper among others relevant to
the topic raised by Anurag and attracting Paul's comments.
NMR is inherently insensitive, that's why most in vivo stuff concentrates on major metabolites in
major tissues - phosphate and ATP in muscle, etc. The limitation is fundamental - changing the
data processing from FT to MEM only helps at the margins, and doesn't change the ballpark you're
playing in. Sometimes the margins matter, sure. But the biggest improvements in signal to noise
ratio come from having higher fields, bigger samples and better probes. You may see stuff about
using probes at liquid helium temperatures, but with biological samples the source of the noise is
usually the sample itself (because its salty) and you can't do much about that either. So there's
another limit.
Detecting hundreds of micromolar concentrations (cytoplasmic concentration) is quite possible even
in 1g of mammalian cells, but at concentrations much less than that then the experiment duration
gets up around the many-hours mark. Even that assumes you can resolve your signal from others,
which is another critical limitation. You can't usually decouple cell samples because of the
marked heating it causes, and the inherent linewidths mean it wouldn't always help anyway. It's
the sample inhomogeneity that cause the problem and you can't help that if you don't want to
homogenise your sample !
I think the most successful measurements of micromolar metabolites are likely to be made
indirectly - either via observing abundant metabolites in equilibrium with your inabundant target,
or using a probe/reporter molecule.
The equilibrium approach has been very successfully used with the creatine kinase equilibrium to
measure free ADP in cells (Brosnan et al, JBC 265 20849- 1990)
Probes have also been used - hemoglobin nmr to measure pH in vivo, and myoglobin to measure pO2
(Brown et al Febs Letts. 82, 12-16 1977 and Kreutzer et al PNAS 89, 4731-4733 1992). Synthetic
small-molecule probes analagous to the fluorescent dyes have also been used, e.g. fluorobapta to
measure calcium in the heart (Kirschenlohr et al PNAS 85, 9017-9021 1988).
I'm most interested in big-molecule probes, proteins, i.e. getting a ligand-binding protein to
report on its ligand concentration. This has been done bioluminescently, fluorescently and by NMR
- read my paper, Williams et al., Biochemistry 32(18), 4895-4902 1993. Each has its own problems
and limitations, nothing is perfect that's for sure. As far as I'm aware, all these techniques
require calibrating with solutions in vitro, assuming things about the intracellular environment
that are probably right BUT...
Right, back to the nmr machine to see how my cells are getting along ....
Simon Williams
Biochemistry Dept.,
Cambridge University,
England
PS - Paul, didn't we meet in Canterbury last January ? I'd like to get in touch again about the
PFK plasmid curing stuff you discussed a bit with me and Kevin Brindle (the boss).
