The results of minimal model calculations indicate that the stability and the kinetic accessibility of the native state of small globular proteins are controlled by few "hot" sites. By means of molecular dynamics simulations around the native conformation, which describe the protein and the surrounding solvent at the all-atom level, an accurate and compact energetic map of the native state of the protein is generated. This map is further simplified by means of an eigenvalue decomposition. The components of the eigenvector associated with the lowest eigenvalue indicate which hot sites are likely to be responsible for the stability and for the rapid folding of the protein. The comparison of the results of the model with the findings of mutagenesis experiments performed for four small proteins show that the eigenvalue decomposition method is able to identify between 60% and 80% of these (hot) sites.Keywords: protein folding; protein stability; molecular dynamics; local elementary structures The study of how the structure and stability of a protein is connected with its amino acid sequence has been, during recent years, a major focus of research. In particular, the problem of protein folding and its relation to stability has been studied through a variety of experimental and theoretical techniques. These studies have highlighted the central role the free energy landscape plays in determining the properties displayed by proteins. The discussion has then been turned to the problem of determining this landscape for specific proteins (Shea and Brooks 2001).Experimental techniques have yielded detailed information on macroscopic features of protein dynamical behavior, such as folding times and stability, and on some specific issues at the level of amino acids, such as the sensitivity to mutations (cf. Fersht 1999). However, there is still no experimental procedure capable of providing insight at the amino acid level into either the folding process in its completeness or the stabilization determinants of proteins. To gain insight into these questions, one must turn to theoretical and computational methods.Because of the high computational costs involved, realistic models that provide an exhaustive description of the free energy landscape have been used in the study of only short peptides. Ferrara and Caflish (2000), for instance, could reconstruct the whole free energy landscape of a small designed ␤-sheet peptide with an all-atom representation of the solute and an implicit model of the solvent. Daura et al. (1998) were able to demonstrate the reversible folding of a small (seven residues) helix forming ␤-peptide in methanol solvent using long molecular dynamics (MD) simulations.On the other hand, minimal models (Chan and Dill 1991;Sali et al. 1994;Mirny and Shakhnovich 2001;Micheletti 2003) have provided important results about the general features of the free energy landscape of proteins. They give an approximate description of both the interaction energy among amino acids (usually through a contact potential en-

Protein evolution is crucial for organismal adaptation and fitness. This process takes place by shaping a given 3-dimensional fold for its particular biochemical function within the metabolic requirements and constraints of the environment. The complex interplay between sequence, structure, functionality, and stability that gives rise to a particular phenotype has limited the identification of traits acquired through evolution. This is further complicated by the fact that mutations are pleiotropic, and interactions between mutations are not always understood. Antibiotic resistance mediated by ␤-lactamases represents an evolutionary paradigm in which organismal fitness depends on the catalytic efficiency of a single enzyme. Based on this, we have dissected the structural and mechanistic features acquired by an optimized metallo-␤-lactamase (M␤L) obtained by directed evolution. We show that antibiotic resistance mediated by this enzyme is driven by 2 mutations with sign epistasis. One mutation stabilizes a catalytically relevant intermediate by fine tuning the position of 1 metal ion; whereas the other acts by augmenting the protein flexibility. We found that enzyme evolution (and the associated antibiotic resistance) occurred at the expense of the protein stability, revealing that M␤Ls have not exhausted their stability threshold. Our results demonstrate that flexibility is an essential trait that can be acquired during evolution on stable protein scaffolds. Directed evolution aided by a thorough characterization of the selected proteins can be successfully used to predict future evolutionary events and design inhibitors with an evolutionary perspective.antibiotic resistance ͉ enzyme ͉ fitness landscape ͉ metalloproteins ͉ epistasis P rotein evolution is crucial for organismal adaptation and fitness (1-8). This process takes place by shaping a given 3-dimensional fold for its particular biochemical function within the metabolic requirements and constraints of the environment.

Understanding the conformational transitions that trigger the aggregation and amyloidogenesis of otherwise soluble peptides at atomic resolution is of fundamental relevance for the design of effective therapeutic agents against amyloid-related disorders. In the present study the transition from ideal alpha-helical to beta-hairpin conformations is revealed by long timescale molecular dynamics simulations in explicit water solvent, for two well-known amyloidogenic peptides: the H1 peptide from prion protein and the Abeta(12-28) fragment from the Abeta(1-42) peptide responsible for Alzheimer's disease. The simulations highlight the unfolding of alpha-helices, followed by the formation of bent conformations and a final convergence to ordered in register beta-hairpin conformations. The beta-hairpins observed, despite different sequences, exhibit a common dynamic behavior and the presence of a peculiar pattern of the hydrophobic side-chains, in particular in the region of the turns. These observations hint at a possible common aggregation mechanism for the onset of different amyloid diseases and a common mechanism in the transition to the beta-hairpin structures. Furthermore the simulations presented herein evidence the stabilization of the alpha-helical conformations induced by the presence of an organic fluorinated cosolvent. The results of MD simulation in 2,2,2-trifluoroethanol (TFE)/water mixture provide further evidence that the peptide coating effect of TFE molecules is responsible for the stabilization of the soluble helical conformation.

Metallo-␤-lactamases (M␤Ls) constitute an increasingly serious clinical threat by giving rise to ␤-lactam antibiotic resistance. They accommodate in their catalytic pocket one or two zinc ions, which are responsible for the hydrolysis of ␤-lactams. Recent x-ray studies on a member of the mono-zinc B2 M␤Ls, CphA from Aeromonas hydrophila, have paved the way to mechanistic studies of this important subclass, which is selective for carbapenems. Here we have used hybrid quantum mechanical/ molecular mechanical methods to investigate the enzymatic hydrolysis by CphA of the antibiotic biapenem. Our calculations describe the entire reaction and point to a new mechanistic description, which is in agreement with the available experimental evidence. Within our proposal, the zinc ion properly orients the antibiotic while directly activating a second catalytic water molecule for the completion of the hydrolytic cycle. This mechanism provides an explanation for a variety of mutagenesis experiments and points to common functional facets across B2 and B1 M␤Ls.

The main problems found in designing drugs are those of optimizing the drug-target interaction and of avoiding the insurgence of resistance. We suggest a scheme for the design of inhibitors that can be used as leads for the development of a drug and that do not face either of these problems, and then apply it to the case of HIV-1-PR. It is based on the knowledge that the folding of single-domain proteins, such as each of the monomers forming the HIV-1-PR homodimer, is controlled by local elementary structures (LES), stabilized by local contacts among hydrophobic, strongly interacting, and highly conserved amino acids that play a central role in the folding process. Because LES have evolved over many generations to recognize and strongly interact with each other so as to make the protein fold fast and avoid aggregation with other proteins, highly specific (and thus little toxic) as well as effective folding-inhibitor molecules suggest themselves: short peptides (or eventually their mimetic molecules) displaying the same amino acid sequence of that of LES (p-LES). Aside from being specific and efficient, these inhibitors are expected not to induce resistance; in fact, mutations in HIV-1-PR that successfully avoid the action of p-LES imply the destabilization of one or more LES and thus should lead to protein denaturation. Making use of Monte Carlo simulations, we first identify the LES of the HIV-1-PR and then show that the corresponding p-LES peptides act as effective inhibitors of the folding of the protease.

The zinc enzymes metallo beta-lactamases counteract the beneficial action of beta-lactam antibiotics against bacterial infections, by hydrolyzing their beta-lactam rings. To understand structure/function relationships on a representative member of this class, the B2 M beta L CphA, we have investigated the H-bond pattern at the Zn enzymatic active site and substrate binding mode by molecular simulation methods. Extensive QM calculations at the DFT-BLYP level on eleven models of the protein active site, along with MD simulations of the protein in aqueous solution, allow us to propose two plausible protonation states for the unbound enzyme, which are probably in equilibrium. Docking procedures along with MD simulations and QM calculations suggest that in the complex between the enzyme and its substrate (biapenem), the latter is stable in only one of the two protonation states, in addition it exhibits two different binding modes, of which only one agrees with previous proposals. In both cases, the substrate is polarized as in aqueous solution. We conclude that addressing mechanistic issues on this class of enzymes requires a careful procedure to assign protonation states and substrate docking modes.

Molecular dynamics simulations were used to study the effect of mechanical and thermal stimuli on the electrostatic properties of collagen model helices. Our model sequences were based on glycine proline and hydroxyproline amino acids. We find that longitudinal mechanical strain induces significant variation of the polarization of the collagen fibril. Such a phenomenon is determined by reorientation of the backbone polar groups, which are free to respond to the mechanical solicitation. This non-negligible effect is facilitated by the peculiar folding structure of the collagen helix, which is characterized by the absence of an extended hydrogen-bond network. The stretching/compression of the helix requires a concomitant winding/unwinding motion of the global structure; therefore, the shear components of the stress tensor are the components that most effectively induce structural modification associated to the piezoelectric response. The present calculations also report a pyroelectric response to thermal activation. Model calculations indicate that the pyroelectric effect is dominated by secondary components associated with the piezoelectric tensor.

Amyloid fibers are aggregates of proteins. They are built out of a peptide called β-amyloid (Aβ) containing between 41 and 43 residues, produced by the action of an enzyme which cleaves a much larger protein known as the Amyloid Precursor Protein (APP). X-ray diffraction experiments have shown that these fibrils are rich in β-structures, whereas the shape of the peptide displays an α-helix structure within the APP in its biologically active conformation. A realistic model of fibril formation is developed based on the seventeen residues Aβ12-28 amyloid peptide, which has been shown to form fibrils structurally similar to those of the whole Aβ peptide. With the help of physical arguments and in keeping with experimental findings, the Aβ12-28 monomer is assumed to be in four possible states (i.e., native helix conformation, β-hairpin, globular low-energy state and unfolded state). Making use of these monomeric states, oligomers (dimers, tertramers and octamers) were constructed. With the help of short, detailed Molecular Dynamics (MD) calculations of the three monomers and of a variety of oligomers, energies for these structures were obtained. Making use of these results within the framework of a simple yet realistic model to describe the entropic terms associated with the variety of amyloid conformations, a phase diagram can be calculated of the whole many-body system, leading to a thermodynamical picture in overall agreement with the experimental findings. In particular, the existence of micellar metastable states seem to be a key issue to determine the thermodynamical properties of the system.