Solution structure of a sweet protein: NMR study of MNEI, a single chain monellin

The sweet protein MNEI is a construct of 96 amino acid residues engineered by linking, with a Gly-Phe dipeptide, chains B and A of monellin, a sweet p rotein isolated from Discoreophyllum cuminsii. Here, the solution structure of MNEI was determined on the basis of 1169 nuclear Overhauser enhancement derived distance restraints and 184 dihedral angle restraints obtained fro m direct measurement of three-bond spin coupling constants. The identificat ion of hydrogen bonded NH groups was obtained by a combination of H/H-2 exc hange data and NH resonance temperature coefficients derived from a series of HSQC spectra in the temperature range 278-328 K. The good resolution of the structure is reflected by the Z-score of the quality checking program i n WHAT TF (-0.61). The topology of MNEI, like that of natural monellin and of SCM, another single-chain monellin, is typical of the cystatin superfami ly: an alpha -helix cradled into the concave side of a five-strand anti-par allel beta -sheet. The high resolution (14 restraints/residue) 3D structure of MNEI shows close similarity to the crystal structures of natural monell in and of SCM but differs from the solution structure of SCM. The structure s of SCM in the crystal and in solution differ in some of the secondary str ucture elements, but most of all in the relative arrangement of the element s: the four main P-strands that surround the helix in the crystal structure or SCM, are displaced far from the helix in the solution structure of SCM. These differences were attributed to the fact that SCM is a monomer in sol ution and a dimer in the crystal. This result is at variance with the obser vation that our solution structure, like that of SCM, corresponds to a mono meric state of the protein, as demonstrated by the insensitivity of HSQC sp ectra to extreme dilution (down to 20 muM). On the basis of the solution st ructure of MNEI it is possible to propose that the main glucophores are hos ted on loop L34, whereas the N-terminal and C-terminal regions host two oth er important interaction regions, centered around segments 6-9 and 94-96. ( C) 2001 Academic Press.