Developing pea (Pisum sativum L.) seeds were chosen to evaluate the pe
rformance of various nuclear magnetic resonance (NMR) and magnetic res
onance imaging (MRI) methods of detecting sucrose in plants. The metho
ds included chemical shift selective imaging (CHESS), heteronuclear co
rrelation via C-13-H-1 coupling (HMQC), and homonuclear correlation vi
a H-1-H-1 coupling (DQF). The same experiments were also performed on
sucrose phantom samples to evaluate the methods in the absence of the
line broadening observed in plant systems. Using the spin echo techniq
ue for multi-slice imaging, we could discern the detailed internal str
ucture of the intact seed with a resolution of tens of microns. The pr
oton spin-lattice relaxation time and linewidth as a function of the a
ge of the seed were measured to optimize the efficiency of the NMR and
MR experiments. The age-dependent changes in these NMR parameters are
consistent with the accumulation of insoluble starch as age increases
. Both the NMR and MRI results are in accord with the results of chemi
cal analysis, which reveal that the sucrose concentration is higher in
the embryo than in the seed coat, and glucose is at low concentration
throughout the seed. Of the three methods for proton observation, the
enhanced version of the CHESS approach (CD-CHESS) provides the best c
ombination of sucrose detection and water suppression. Direct observat
ion of C-13 is preferable to indirect detection using HMQC because of
water signal bleed-through in samples with large (>200 Hz) linewidths.