Simulation of oligopeptide dynamics and folding. The use of NMR chemical shifts to analyse the MD trajectories.

In this paper, a simulation of the folding process, based on a random perturbations of the phi, psi, chi1 dihedral angles, is proposed to approach the formation at the atom level of both principal elements of protein secondary structure, the alpha-helix and the beta-hairpin structures. Expecting to understand what may happen in solution during the formation of such structures, the behaviour of large sets of random conformations that are generated for small oligopeptides was analysed. Different factors that may influence the folding (as conformational propensity, hydrophobic interactions and side-chain mobility) were investigated. The difference between the corresponding theoretical folding and the real conformational diversity that is observed in solution is appraised by a comparison between the calculated and observed NMR secondary chemical shifts. From this study it appears that hydrophobic interactions and mobility represent the principal factors that initiate folding and determine the observed hydrogen-bond pattern, which subsequently allows packing between the peptide side chains.

Crystal structure and biochemical properties of the human mitochondrial ferritin and its mutant Ser144Ala.

Mitochondrial ferritin is a recently identified protein precursor encoded by an intronless gene. It is specifically taken up by the mitochondria and processed to a mature protein that assembles into functional ferritin shells. The full mature recombinant protein and its S144A mutant were produced to study structural and functional properties. They yielded high quality crystals from Mg(II) solutions which diffracted up to 1.38 Angstrom resolution. The 3D structures of the two proteins resulted very similar to that of human H-ferritin, to which they have high level of sequence identity (approximately 80%). Metal-binding sites were identified in the native crystals and in those soaked in Mn(II) and Zn(II) solutions. The ferroxidase center binds binuclear iron at the sites A and B, and the structures showed that the A site was always fully occupied by Mg(II), Mn(II) or Zn(II), while the occupancy of the B site was variable. In addition, distinct Mg(II) and Zn(II)-binding sites were found in the 3-fold axes to block the hydrophilic channels. Other metal-binding sites, never observed before in H-ferritin, were found on the cavity surface near the ferroxidase center and near the 4-fold axes. Mitochondrial ferritin showed biochemical properties remarkably similar to those of human H-ferritin, except for the difficulty in renaturing to yield ferritin shells and for a reduced ( approximately 41%) rate in ferroxidase activity. This was partially rescued by the substitution of the bulkier Ser144 with Ala, which occurs in H-ferritin. The residue is exposed on a channel that connects the ferroxidase center with the cavity. The finding that the mutation increased both catalytic activity and the occupancy of the B site demonstrated that the channel is functionally important. In conclusion, the present data define the structure of human mitochondrial ferritin and provide new data on the iron pathways within the H-type ferritin shell.

The use of NMR chemical shifts to analyse the MD trajectories: simulation of bovine pancreatic trypsin inhibitor dynamics in water as a test case for solvent influences.

In this paper the NMR secondary chemical shifts, that are estimated from a set of 3D-structures, are compared with the observed ones to appraise the behaviour of a known x-ray diffraction structure (of the bovine pancreatic trypsin inhibitor protein) when various molecular dynamics are applied. The results of a 200 ps molecular dynamics under various conditions are analysed and different ways to modify the molecular dynamics are considered. With the purpose of avoiding the time-consuming explicit representation of the solvent (water) molecules, an attempt was made to understand the role of the solvent and to develop an implicit representation, which may be refined. A simulation of hydrophobic effects in an aqueous environment is also proposed which seems to provide a better approximation of the observed solution structure of the protein.

Structural description of the active sites of mouse L-chain ferritin at 1.2 A resolution.

The first ferritin structure refined at the atomic level has been achieved on recombinant mouse L-chain apoferritin (rMoLF) crystals. These latter diffract to 1.2 A resolution under cryogenic conditions. When cryo-cooling the sample, the thermal disorder usually observed at room temperature is reduced and the low-temperature structure reveals several details concerning the protein putative active sites and their properties. Within the pores built up by the molecular three-fold symmetry axes, the iron entry route to the ferritin cavity, residues H118, D131 and E134, exhibit alternate conformations associated with the binding of partially hydrated cadmium ions, a metal used as a crystallization agent. At the mineral ferrihydrite nucleation center, the electron density maps evidence the orientation of E57, E60, E61 and E64 glutamate side chains (whereas they were observed highly disordered in previous ferritin structures determined at room temperature) and allow a description of the site taking into account the binding geometry of four Cd(2+) ions. Moreover, the side chain of residue K140, lying in the vicinity of the ferrihydrite nucleation center, is shown to interact with residue E61. As previously highlighted, this observation confirms the importance of K140 in the rMoLF sequence, as being responsible for the low level of iron incorporation by mousel L-chain ferritin compared to human L-chain ferritin. Finally, the diffusion of small molecules within the ferritin cavity is illustrated here by the presence of ordered molecules of glycerol used as a cryo-protectant, which bind the inner cavity surface of the protein.

Study of the yeast Saccharomyces cerevisiae F1F0-ATPase epsilon-subunit.

The yeast Saccharomyces cerevisiae F1F0-ATPase epsilon-subunit (61 residues) was synthesized by the solid-phase peptide approach under both acidic and basic strategies. Only the latter strategy allowed us to obtain a pure epsilon-subunit. The strong propensity of the protein to produce few soluble dimeric species depending on pH has been proved by size-exclusion chromatography, electrophoresis and mass spectrometry. A circular dichroism study showed that an aqueous solution containing 30% trifluoroethanol or 200 mM sodium dodecyl sulphate is required for helical folding. In both solvents at acidic pH, the epsilon-subunit is soluble and monomeric.