Solution structures by 1H NMR of the novel cyclic trypsin inhibitor SFTI-1 from sunflower seeds and an acyclic permutant.

SFTI-1 is a recently discovered cyclic peptide trypsin inhibitor from sunflower seeds comprising 14 amino acid residues. It is the most potent known Bowman-Birk inhibitor and the only naturally occurring cyclic one. The solution structure of SFTI-1 has been determined by 1H-NMR spectroscopy and compared with a synthetic acyclic permutant. The solution structures of both are remarkably similar. The lowest energy structures from each family of 20 structures of cyclic and acyclic SFTI-1 have an rmsd over the backbone and heavy atoms of 0.29 A and 0.66 A, respectively. The structures consist of two short antiparallel beta-strands joined by an extended loop containing the active site at one end. Cyclic SFTI-1 also has a hairpin turn completing the cycle. Both molecules contain particularly stable arrangements of cross-linking hydrogen bonds between the beta-strands and a single disulfide bridge, making them rigid and well defined in solution. These stable arrangements allow both the cyclic and acyclic variants of SFTI-1 to inhibit trypsin with very high potencies (0.5 nM and 12.1 nM, respectively). The cyclic nature of SFTI-1 appears to have evolved to provide higher trypsin inhibition as well as higher stability. The solution structures are similar to the crystal structure of the cyclic inhibitor in complex with trypsin. The lack of a major conformational change upon binding suggests that the structure of SFTI-1 is rigid and already pre-organized for maximal binding due to minimization of entropic losses compared to a more flexible ligand. These properties make SFTI-1 an ideal platform for the design of small peptidic pharmaceuticals or pesticides.

Three-dimensional structure of RTD-1, a cyclic antimicrobial defensin from Rhesus macaque leukocytes.

Most mammalian defensins are cationic peptides of 29-42 amino acids long, stabilized by three disulfide bonds. However, recently Tang et al. (1999, Science 286, 498-502) reported the isolation of a new defensin type found in the leukocytes of rhesus macaques. In contrast to all the other defensins found so far, rhesus theta defensin-1 (RTD-1) is composed of just 18 amino acids with the backbone cyclized through peptide bonds. Antibacterial activities of both the native cyclic peptide and a linear form were examined, showing that the cyclic form was 3-fold more active than the open chain analogue [Tang et al. (1999) Science 286, 498-502]. To elucidate the three-dimensional structure of RTD-1 and its open chain analogue, both peptides were synthesized using solid-phase peptide synthesis and tert-butyloxycarbonyl chemistry. The structures of both peptides in aqueous solution were determined from two-dimensional (1)H NMR data recorded at 500 and 750 MHz. Structural constraints consisting of interproton distances and dihedral angles were used as input for simulated-annealing calculations and water refinement with the program CNS. RTD-1 and its open chain analogue oRTD-1 adopt very similar structures in water. Both comprise an extended beta-hairpin structure with turns at one or both ends. The turns are well defined within themselves and seem to be flexible with respect to the extended regions of the molecules. Although the two strands of the beta-sheet are connected by three disulfide bonds, this region displays a degree of flexibility. The structural similarity of RTD-1 and its open chain analogue oRTD-1, as well as their comparable degree of flexibility, support the theory that the additional charges at the termini of the open chain analogue rather than overall differences in structure or flexibility are the cause for oRTD-1's lower antimicrobial activity. In contrast to numerous other antimicrobial peptides, RTD-1 does not display any amphiphilic character, even though surface models of RTD-1 exhibit a certain clustering of positive charges. Some amide protons of RTD-1 that should be solvent-exposed in monomeric beta-sheet structures show low-temperature coefficients, suggesting the possible presence of weak intermolecular hydrogen bonds.

The solution structure of C1-T1, a two-domain proteinase inhibitor derived from a circular precursor protein from Nicotiana alata.

A two-domain portion of the proteinase inhibitor precursor from Nicotiana alata (NaProPI) has been expressed and its structure determined by NMR spectroscopy. NaProPI contains six almost identical 53 amino acid repeats that fold into six highly similar domains; however, the sequence repeats do not coincide with the structural domains. Five of the structural domains comprise the C-terminal portion of one repeat and the N-terminal portion of the next. The sixth domain contains the C-terminal portion of the sixth repeat and the N-terminal portion of the first repeat. Disulphide bonds link these C and N-terminal fragments to generate the clasped-bracelet fold of NaProPI. The three-dimensional structure of NaProPI is not known, but it is conceivable that adjacent domains in NaProPI interact to generate the circular "bracelet" with the N and C termini in close enough proximity to facilitate formation of the disulphide bonds that form the "clasp". The expressed protein, examined in the current study, comprises residues 25-135 of NaProPI and encompasses the first two contiguous structural domains, namely the chymotrypsin inhibitor C1 and the trypsin inhibitor T1, joined by a five-residue linker, and is referred to as C1-T1. The tertiary structure of each domain in C1-T1 is identical to that found in the isolated inhibitors. However, no nuclear Overhauser effect contacts are observed between the two domains and the five-residue linker adopts an extended conformation. The absence of interactions between the domains indicates that adjacent domains do not specifically interact to drive the circularisation of NaProPI. These results are in agreement with recent data which describe similar PI precursors from other members of the Solanaceae having two, three, or four repeats. The lack of strong interdomain association is likely to be important for the function of individual inhibitors by ensuring that there is no masking of reactive sites upon release from the precursor.

Structure of reduced DsbA from Escherichia coli in solution.

The three-dimensional structure of reduced DsbA from Escherichia coli in aqueous solution has been determined by nuclear magnetic resonance (NMR) spectroscopy and is compared with the crystal structure of oxidized DsbA [Guddat, L. W., Bardwell, J. C. A., Zander, T., and Martin, J. L. (1997) Protein Sci. 6, 1148-1156]. DsbA is a monomeric 21 kDa protein which consists of 189 residues and is required for disulfide bond formation in the periplasm of E. coli. On the basis of sequence-specific 1H NMR assignments, 1664 nuclear Overhauser enhancement distance constraints, 118 hydrogen bond distance constraints, and 293 dihedral angle constraints were obtained as the input for the structure calculations by simulated annealing with the program X-PLOR. The enzyme is made up of two domains. The catalytic domain has a thioredoxin-like fold with a five-stranded beta-sheet and three alpha-helices, and the second domain consists of four alpha-helices and is inserted into the thioredoxin motif. The active site between Cys30 and Cys33 is located at the N terminus of the first alpha-helix in the thioredoxin-like domain. The solution structure of reduced DsbA is rather similar to the crystal structure of the oxidized enzyme but exhibits a different relative orientation of both domains. In addition, the conformations of the active site and a loop between strand beta5 and helix alpha7 are slightly different. These structural differences may reflect important functional requirements in the reaction cycle of DsbA as they appear to facilitate the release of oxidized polypeptides from reduced DsbA. The extremely low pKa value of the nucleophilic active site thiol of Cys30 in reduced DsbA is most likely caused by its interactions with the dipole of the active site helix and the side chain of His32, as no other charged residues are located next to the sulfur atom of Cys30 in the solution structure.