Selective incorporation of the stereospecifically deuteriated sugar moieties (> 97 atom % 2H enhancements at H2', H2'', H3' and H5'/5'' sites, approximately 85 atom % 2H enhancement at H4' and approximately 20 atom % 2H enhancement at H1') in DNA and RNA by the 'NMR-window' approach has been shown to solve the problem of the resonance overlap [refs. 1, 2 & 3]. Such specific deuterium labelling gives much improved resolution and sensitivity of the residual sugar proton (i.e. H1' or H4') vicinal to the deuteriated centers (ref. 3). The T2 relaxation time of the residual protons also increases considerably in the partially-deuteriated (shown by underline) sugar residues in dinucleotides [d(CpG), d(GpC), d(ApT), d(TpA)], trinucleotide r(A2'p5'A2'p5'A) and 20-mer DNA duplex 5'd(C1G2C3-G4C5G6C7G8A9A10T11T12C13G14C15G16C17G18C19G20)(2) 3'. The protons with shorter T2 can be filtered away using a number of different NMR experiments such as ROESY, MINSY or HAL. The NOE intensity of the cross-peaks in these experiments includes only straight pathway from H1' to aromatic proton (i-i and i-i + 1) without any spin-diffusion. The volumes of these NOE cross-peaks could be measured with high accuracy as their intensity is 3 to 4 times larger than the corresponding peaks in the fully protonated residues in the normal NOESY spectra. The structural informations thus obtainable from the residual protons in the partially-deuteriated part of the duplex and the fully protonated part in the 'NMR window' can indeed complement each other.