Local interactions as a structure determinant for protein molecules: III

Protein solubility: sequence based prediction and experimental verification

MOTIVATION

Obtaining soluble proteins in sufficient concentrations is a recurring limiting factor in various experimental studies. Solubility is an individual trait of proteins which, under a given set of experimental conditions, is determined by their amino acid sequence. Accurate theoretical prediction of solubility from sequence is instrumental for setting priorities on targets in large-scale proteomics projects.

RESULTS

We present a machine-learning approach called PROSO to assess the chance of a protein to be soluble upon heterologous expression in Escherichia coli based on its amino acid composition. The classification algorithm is organized as a two-layered structure in which the output of primary support vector machine (SVM) classifiers serves as input for a secondary Naive Bayes classifier. Experimental progress information from the TargetDB database as well as previously published datasets were used as the source of training data. In comparison with previously published methods our classification algorithm possesses improved discriminatory capacity characterized by the Matthews Correlation Coefficient (MCC) of 0.434 between predicted and known solubility states and the overall prediction accuracy of 72% (75 and 68% for positive and negative class, respectively). We also provide experimental verification of our predictions using solubility measurements for 31 mutational variants of two different proteins.

A molecular model for the retinol binding protein-transthyretin complex

A three-dimensional model for the complex between human serum retinol binding protein and transthyretin (formerly named prealbumin) is presented. The model was obtained by interactive rigid-body computer graphics docking and the characterization of the molecular surfaces in terms of fractal dimension. Available experimental data, as well as results from molecular dynamics calculations, support the proposed model.

The frequency of ion-pair substructures in proteins is quantitatively related to electrostatic potential: a statistical model for nonbonded interactions

A statistical analysis of ion pairs in protein crystal structures shows that their abundance with respect to uncharged controls is accurately predicted by a Boltzmann-like function of electrostatic potential. It appears that the mechanisms of protein folding and/or evolution combine to produce a "thermal" distribution of local nonbonded interactions, as has been suggested by statistical-mechanical theories. Using this relationship, we develop a maximum likelihood methodology for estimation of apparent energetic parameters from the data base of known structures, and we derive electrostatic potential functions that lead to optimal agreement of observed and predicted ion-pair frequencies. These are similar to potentials of mean force derived from electrostatic theory, but departure from Coulombic behavior is less than has been suggested.

Different structures in the amino-terminal domain of the ornithine transcarbamylase leader peptide are involved in mitochondrial import and carboxyl-terminal cleavage

The cytoplasmic precursor of mitochondrial ornithine transcarbamylase (carbamoyl-phosphate:L-ornithine carbamoyltransferase, EC 2.1.3.2) contains an amino-terminal leader peptide of 32 amino acids. Secondary structure and helical-wheel analyses predict that the extreme amino-terminal domain (residues 1-15) forms an alpha-helix. To test this thesis, leucine residues at positions 2, 5, 8, and 9 were systematically replaced by either helix-breaking glycine residues or by helix-preserving alanine residues. Triple substitutions of glycine for leucine in positions 2, 5, and 9 or 5, 8, and 9 abolished the uptake of the rat precursor by intact mitochondria, whereas similar alanine substitutions had much less effect. Theoretical computations predicted that the decreased helical stability of the Gly-5,8,9 substitution could be significantly increased by replacing a serine in position with phenylalanine. The introduction of Phe-3, indeed, restored the mitochondrial uptake of the mutant precursor. These results lend strong support to the hypothesis that an alpha-helix is present at the leader's amino terminus during the import of the precursor by mitochondria. Although the precursors with the triply-substituted leaders were impaired with respect to import, they were still cleaved readily by a protease found in a mitochondrial matrix fraction. Substitution of glycine or alanine for all four leucine residues, however, rendered the leader uncleavable at the carboxyl-terminal cleavage site. These results suggest that the structure of the amino-terminal domain is important for recognition of the carboxyl-terminal cleavage sites by the matrix proteases.

Crystallographic structure of an active, sequence-engineered ribonuclease

X-ray diffraction methods were used to test a synthetic-modeling approach to the sequence engineering of bovine pancreatic ribonuclease. A model of RNase S-peptide (residues 1-20), having a simplified amino acid sequence but retaining elements deduced to be essential for conformation and function, was previously synthesized and found to form a catalytically active and stable complex with native S-protein (residues 21-24). We have now obtained a 3-A-resolution electron density map of this semisynthetic complex which reveals that the conformation of model peptide closely mimics that of native S-peptide, as intended by sequence design. Some small differences from the native structure are observed: Glu-2 and Arg-10 of the model complex are not close enough to form a salt bridge, the position of the His-12 imidazole ring is slightly shifted in the active site, and the peptide's amino terminus is reoriented. Nonetheless, the major structural features predicted to be essential by computer-aided peptide-design analysis are preserved in the model peptide portion of the complex. These include (i) the alpha-helical framework involving residues 3-13, (ii) the catalytically competent orientation of His-12, and (iii) complex-stabilizing non-bonding interactions involving Phe-8 and Met-13 of S-peptide and hydrophobic residues in the cleft region of S-protein. Further, sequence simplification has not introduced any non-native, potentially stabilizing contacts between the model peptide and S-protein. The results emphasize the usefulness, in redesigning native proteins, of categorizing sequence into residues providing conformational framework and those determining intra-and intermolecular surface recognition.

Semisynthetic peptides and proteins

Semisynthesis provides a flexible approach for using chemical synthesis to produce large, biologically active polypeptides and analogues. Currently developing improvements in the basic methods used, including polypeptide fragmentation, peptide synthesis, and reconstitution of synthetic and native components, make this overall approach applicable to a variety of species. Sequence modification through semisynthesis thus provides a flexible route to explore the code of rules whereby primary structure directs higher order properties of folded conformation and biological function of large peptides and proteins. The fruits of this endeavor, an understanding of how these macromolecules work, and therein, a basis for design of new structures that ultimately may be produced directly or by recombinant DNA methods, have begun to emerge.

Yeast enolase: mechanism of activation by metal ions

Yeast enolase as prepared by current procedures is inherently chemically homogeneous, though deamidation and partial denaturation can produce electrophoretically distinct forms. A true isozyme of the enzyme exists but does not survive the purification procedure. The chemical sequence for both has been established. The enzyme behaves in solution like a compact, nearly spherical molecule of moderate hydration. Strong intramolecular forces maintain the structure of the individual subunits. The enzyme as isolated is dimeric. If dissociated in the presence of magnesium ions and substrate, then the subunits are active, but if the dissociation occurs in the absence of metal ions, they are inactive until they have reassociated and undergone a first order "annealing" process. Magnesium (II) enhances association. The interaction between the subunits is hydrophobic in character. The enzyme can bind up to 2 mol of most metal ions in "conformational" sites which then allows up to 2 mol of substrate or some substrate analogue to bind. This is not sufficient for catalysis, but conformational metal ions do more than just allow substrate binding. A change in the environment of the metal ions occurs on substrate or substrate analogue binding. There is an absolute correlation between the occurrence of a structural change undergone by the 3-amino analogue of phosphoenolpyruvate and whether the metal ions produce any level of enzymatic activity. For catalysis, two more moles of metal ions, called "catalytic", must bind. There is evidence that the enzymatic reaction involves a carbanion mechanism. It is likely that two more moles of metal ion can bind which inhibit the reaction. The requirement for 2 mol of metal ion per subunit which contribute in different ways to catalysis is exhibited by a number of other enzymes.