Comparative protein structure prediction is limited mostly by the errors in alignment and loop modeling. We describe here a new automated modeling technique that significantly improves the accuracy of loop predictions in protein structures. The positions of all nonhydrogen atoms of the loop are optimized in a fixed environment with respect to a pseudo energy function. The energy is a sum of many spatial restraints that include the bond length, bond angle, and improper dihedral angle terms from the CHARMM-22 force field, statistical preferences for the main-chain and side-chain dihedral angles, and statistical preferences for nonbonded atomic contacts that depend on the two atom types, their distance through space, and separation in sequence. The energy function is optimized with the method of conjugate gradients combined with molecular dynamics and simulated annealing. Typically, the predicted loop conformation corresponds to the lowest energy conformation among 500 independent optimizations. Predictions were made for 40 loops of known structure at each length from 1 to 14 residues. The accuracy of loop predictions is evaluated as a function of thoroughness of conformational sampling, loop length, and structural properties of native loops. When accuracy is measured by local superposition of the model on the native loop, 100, 90, and 30% of 4-, 8-, and 12-residue loop predictions, respectively, had <2 Å RMSD error for the mainchain N, Cα, C, and O atoms; the average accuracies were 0.59 ± 0.05, 1.16 ± 0.10, and 2.61 ± 0.16 Å, respectively. To simulate real comparative modeling problems, the method was also evaluated by predicting loops of known structure in only approximately correct environments with errors typical of comparative modeling without misalignment. When the RMSD distortion of the main-chain stem atoms is 2.5 Å, the average loop prediction error increased by 180, 25, and 3% for 4-, 8-, and 12-residue loops, respectively. The accuracy of the lowest energy prediction for a given loop can be estimated from the structural variability among a number of low energy predictions. The relative value of the present method is gauged by (1) comparing it with one of the most successful previously described methods, and (2) describing its accuracy in recent blind predictions of protein structure. Finally, it is shown that the average accuracy of prediction is limited primarily by the accuracy of the energy function rather than by the extent of conformational sampling.

We present a neural network based method (ChloroP) for identifying chloroplast transit peptides and their cleavage sites. Using cross-validation, 88% of the sequences in our homology reduced training set were correctly classified as transit peptides or nontransit peptides. This performance level is well above that of the publicly available chloroplast localization predictor PSORT. Cleavage sites are predicted using a scoring matrix derived by an automatic motif-finding algorithm. Approximately 60% of the known cleavage sites in our sequence collection were predicted to within ±2 residues from the cleavage sites given in SWISS-PROT. An analysis of 715 Arabidopsis thaliana sequences from SWISS-PROT suggests that the ChloroP method should be useful for the identification of putative transit peptides in genome-wide sequence data. The ChloroP predictor is available as a web-server at http://www.cbs.dtu.dk/services/ChloroP/.

Although it is usually possible to achieve a favorable yield of a recombinant protein in Escherichia coli, obtaining the protein in a soluble, biologically active form continues to be a major challenge. Sometimes this problem can be overcome by fusing an aggregation-prone polypeptide to a highly soluble partner. To study this phenomenon in greater detail, we compared the ability of three soluble fusion partners—maltose-binding protein (MBP), glutathione S-transferase (GST), and thioredoxin (TRX)—to inhibit the aggregation of six diverse proteins that normally accumulate in an insoluble form. Remarkably, we found that MBP is a far more effective solubilizing agent than the other two fusion partners. Moreover, we demonstrated that in some cases fusion to MBP can promote the proper folding of the attached protein into its biologically active conformation. Thus, MBP seems to be capable of functioning as a general molecular chaperone in the context of a fusion protein. A model is proposed to explain how MBP promotes the solubility and influences the folding of its fusion partners.

A simple approach to estimate the number of α-helical and β-strand segments from protein circular dichroism spectra is described. The α-helix and β-sheet conformations in globular protein structures, assigned by DSSP and STRIDE algorithms, were divided into regular and distorted fractions by considering a certain number of terminal residues in a given α-helix or β-strand segment to be distorted. The resulting secondary structure fractions for 29 reference proteins were used in the analyses of circular dichroism spectra by the SELCON method. From the performance indices of the analyses, we determined that, on an average, four residues per α-helix and two residues per β-strand may be considered distorted in proteins. The number of α-helical and β-strand segments and their average length in a given protein were estimated from the fraction of distorted α-helix and β-strand conformations determined from the analysis of circular dichroism spectra. The statistical test for the reference protein set shows the high reliability of such a classification of protein secondary structure. The method was used to analyze the circular dichroism spectra of four additional proteins and the predicted structural characteristics agree with the crystal structure data.

The Escherichia coli biotin holoenzyme synthetase, BirA, catalyzes transfer of biotin to the epsilon amino group of a specific lysine residue of the biotin carboxyl carrier protein (BCCP) subunit of acetyl-CoA carboxylase. Sequences of naturally biotinylated substrates are highly conserved across evolutionary boundaries, and cross-species biotinylation has been demonstrated in several systems. To define the minimal substrate requirements in BirA-catalyzed biotinylation, we have measured the kinetics of modification of a 23-residue peptide previously identified by combinatorial methods. Although the sequence of the peptide bears little resemblance to the biotinylated sequence in BCCP, it is enzymatically biotinylated in vivo. Rates of biotin transfer to the 23-residue peptide are similar to those determined for BCCP. To further elucidate the sequence requirements for biotinylation, transient kinetic measurements were performed on a series of amino- and carboxy-terminal truncations of the 23-mer. The results, determined by stopped-flow fluorescence, allowed identification of a 14-residue peptide as the minimum required sequence. Additional support was obtained using matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometric analysis of peptides that had been incubated with an excess of biotinyl-5′-adenylate intermediate and catalytic amounts of BirA. Results of these measurements indicate that while kinetically inactive truncations showed no significant shift in molecular mass to the values expected for biotinylated species, kinetically active truncations exhibited 100% biotinylation. The specificity constant (kcat/Km) governing BirA-catalyzed biotinylation of the 14-mer minimal substrate is similar to that determined for the natural substrate, BCCP. We conclude that the 14-mer peptide efficiently mimics the biotin acceptor function of the much larger protein domain normally recognized by BirA.

Folding funnels have been the focus of considerable attention during the last few years. These have mostly been discussed in the general context of the theory of protein folding. Here we extend the utility of the concept of folding funnels, relating them to biological mechanisms and function. In particular, here we describe the shape of the funnels in light of protein synthesis and folding; flexibility, conformational diversity, and binding mechanisms; and the associated binding funnels, illustrating the multiple routes and the range of complexed conformers. Specifically, the walls of the folding funnels, their crevices, and bumps are related to the complexity of protein folding, and hence to sequential vs. nonsequential folding. Whereas the former is more frequently observed in eukaryotic proteins, where the rate of protein synthesis is slower, the latter is more frequent in prokaryotes, with faster translation rates. The bottoms of the funnels reflect the extent of the flexibility of the proteins. Rugged floors imply a range of conformational isomers, which may be close on the energy landscape. Rather than undergoing an induced fit binding mechanism, the conformational ensembles around the rugged bottoms argue that the conformers, which are most complementary to the ligand, will bind to it with the equilibrium shifting in their favor. Furthermore, depending on the extent of the ruggedness, or of the smoothness with only a few minima, we may infer nonspecific, broad range vs. specific binding. In particular, folding and binding are similar processes, with similar underlying principles. Hence, the shape of the folding funnel of the monomer enables making reasonable guesses regarding the shape of the corresponding binding funnel. Proteins having a broad range of binding, such as proteolytic enzymes or relatively nonspecific endonucleases, may be expected to have not only rugged floors in their folding funnels, but their binding funnels will also behave similarly, with a range of complexed conformations. Hence, knowledge of the shape of the folding funnels is biologically very useful. The converse also holds: If kinetic and thermodynamic data are available, hints regarding the role of the protein and its binding selectivity may be obtained. Thus, the utility of the concept of the funnel carries over to the origin of the protein and to its function.

Whereas previously we have successfully utilized the folding funnels concept to rationalize binding mechanisms (Ma B, Kumar S, Tsai CJ, Nussinov R, 1999, Protein Eng 12:713–720) and to describe binding (Tsai CJ, Kumar S, Ma B, Nussinov R, 1999, Protein Sci 8:1181–1190), here we further extend the concept of folding funnels, illustrating its utility in explaining enzyme pathways, multimolecular associations, and allostery. This extension is based on the recognition that funnels are not stationary; rather, they are dynamic, depending on the physical or binding conditions (Tsai CJ, Ma B, Nussinov R, 1999, Proc Natl Acad Sci USA 96:9970–9972). Different binding states change the surrounding environment of proteins. The changed environment is in turn expressed in shifted energy landscapes, with different shapes and distributions of populations of conformers. Hence, the function of a protein and its properties are not only decided by the static folded three-dimensional structure; they are determined by the distribution of its conformational substates, and in particular, by the redistributions of the populations under different environments. That is, protein function derives from its dynamic energy landscape, caused by changes in its surroundings.

Fourier transform infrared spectroscopy (FTIR), circular dichroism (CD), and electron microscopy (EM) have been used simultaneously to follow the temperature-induced formation of amyloid fibrils by bovine insulin at acidic pH. The FTIR and CD data confirm that, before heating, insulin molecules in solution at pH 2.3 have a predominantly native-like α-helical structure. On heating to 70 °C, partial unfolding occurs and results initially in aggregates that are shown by CD and FT-IR spectra to retain a predominantly helical structure. Following this step, changes in the CD and FTIR spectra occur that are indicative of the extensive conversion of the molecular conformation from α-helical to β-sheet structure. At later stages, EM shows the development of fibrils with well-defined repetitive morphologies including structures with a periodic helical twist of ∼450 Å. The results indicate that formation of fibrils by insulin requires substantial unfolding of the native protein, and that the most highly ordered structures result from a slow evolution of the morphology of the initially formed fibrillar species.

Distant homologies between proteins are often discovered only after three-dimensional structures of both proteins are solved. The sequence divergence for such proteins can be so large that simple comparison of their sequences fails to identify any similarity. New generation of sensitive alignment tools use averaged sequences of entire homologous families (profiles) to detect such homologies. Several algorithms, including the newest generation of BLAST algorithms and BASIC, an algorithm used in our group to assign fold predictions for proteins from several genomes, are compared to each other on the large set of structurally similar proteins with little sequence similarity. Proteins in the benchmark are classified according to the level of their similarity, which allows us to demonstrate that most of the improvement of the new algorithms is achieved for proteins with strong functional similarities, with almost no progress in recognizing distant fold similarities.

It is also shown that details of profile calculation strongly influence its sensitivity in recognizing distant homologies. The most important choice is how to include information from diverging members of the family, avoiding generating false predictions, while accounting for entire sequence divergence within a family. PSI-BLAST takes a conservative approach, deriving a profile from core members of the family, providing a solid improvement without almost any false predictions. BASIC strives for better sensitivity by increasing the weight of divergent family members and paying the price in lower reliability. A new FFAS algorithm introduced here uses a new procedure for profile generation that takes into account all the relations within the family and matches BASIC sensitivity with PSI-BLAST like reliability.

This paper surveys the emerging role of statistical mechanics and polymer theory in protein folding. In the polymer perspective, the folding code is more a solvation code than a code of local φψ propensities. The polymer perspective resolves two classic puzzles: (1) the Blind Watchmaker's Paradox that biological proteins could not have originated from random sequences, and (2) Levinthal's Paradox that the folded state of a protein cannot be found by random search. Both paradoxes are traditionally framed in terms of random unguided searches through vast spaces, and vastness is equated with impossibility. But both processes are partly guided. The searches are more akin to balls rolling down funnels than balls rolling aimlessly on flat surfaces. In both cases, the vastness of the search is largely irrelevant to the search time and success. These ideas are captured by energy and fitness landscapes. Energy landscapes give a language for bridging between microscopics and macroscopics, for relating folding kinetics to equilibrium fluctuations, and for developing new and faster computational search strategies.

A novel superfamily designated MAPEG (Membrane Associated Proteins in Eicosanoid and Glutathione metabolism), including members of widespread origin with diversified biological functions is defined according to enzymatic activities, sequence motifs, and structural properties. Two of the members are crucial for leukotriene biosynthesis, and three are cytoprotective exhibiting glutathione S-transferase and peroxidase activities. Expression of the most recently recognized member is strongly induced by p53, and may therefore play a role in apoptosis or cancer development. In spite of the different biological functions, all six proteins demonstrate common structural characteristics typical of membrane proteins. In addition, homologues are identified in plants, fungi, and bacteria, demonstrating this superfamily to be generally occurring.

We describe a new classifier for protein secondary structure prediction that is formed by cascading together different types of classifiers using neural networks and linear discrimination. The new classifier achieves an accuracy of 76.7% (assessed by a rigorous full Jack-knife procedure) on a new nonredundant dataset of 496 nonhomologous sequences (obtained from G.J. Barton and J.A. Cuff). This database was especially designed to train and test protein secondary structure prediction methods, and it uses a more stringent definition of homologous sequence than in previous studies. We show that it is possible to design classifiers that can highly discriminate the three classes (H, E, C) with an accuracy of up to 78% for β-strands, using only a local window and resampling techniques. This indicates that the importance of long-range interactions for the prediction of β-strands has been probably previously overestimated.

We examine the similarities and differences between two widely used knowledge-based potentials, which are expressed as contact matrices (consisting of 210 elements) that gives a scale for interaction energies between the naturally occurring amino acid residues. These are the Miyazawa–Jernigan contact interaction matrix M and the potential matrix S derived by Skolnick J et al., 1997, Protein Sci 6:676–688. Although the correlation between the two matrices is good, there is a relatively large dispersion between the elements. We show that when Thr is chosen as a reference solvent within the Miyazawa and Jernigan scheme, the dispersion between the M and S matrices is reduced. The resulting interaction matrix B gives hydrophobicities that are in very good agreement with experiment. The small dispersion between the S and B matrices, which arises due to differing reference states, is shown to have dramatic effect on the predicted native states of lattice models of proteins. These findings and other arguments are used to suggest that for reliable predictions of protein structures, pairwise additive potentials are not sufficient. We also establish that optimized protein sequences can tolerate relatively large random errors in the pair potentials. We conjecture that three body interaction may be needed to predict the folds of proteins in a reliable manner.

We have crystallized Drosophila melanogaster acetylcholinesterase and solved the structure of the native enzyme and of its complexes with two potent reversible inhibitors, 1,2,3,4-tetrahydro-N-(phenylmethyl)-9-acridinamine and 1,2,3,4-tetrahydro-N-(3-iodophenyl-methyl)-9-acridinamine—all three at 2.7 Å resolution. The refined structure of D. melanogaster acetylcholinesterase is similar to that of vertebrate acetylcholinesterases, for example, human, mouse, and fish, in its overall fold, charge distribution, and deep active-site gorge, but some of the surface loops deviate by up to 8 Å from their position in the vertebrate structures, and the C-terminal helix is shifted substantially. The active-site gorge of the insect enzyme is significantly narrower than that of Torpedo californica AChE, and its trajectory is shifted several angstroms. The volume of the lower part of the gorge of the insect enzyme is ∼50% of that of the vertebrate enzyme. Upon binding of either of the two inhibitors, nine aromatic side chains within the active-site gorge change their conformation so as to interact with the inhibitors. Some differences in activity and specificity between the insect and vertebrate enzymes can be explained by comparison of their three-dimensional structures.

The in vitro refolding of hen egg-white lysozyme is studied in the presence of various osmolytes. Proline is found to prevent aggregation during protein refolding. However, other osmolytes used in this study fail to exhibit a similar property. Experimental evidence suggests that proline inhibits protein aggregation by binding to folding intermediate(s) and trapping the folding intermediate(s) into enzymatically inactive, “aggregation-insensitive” state(s). However, elimination of proline from the refolded protein mixture results in significant recovery of the bacteriolytic activity. At higher concentrations (>1.5 M), proline is shown to form loose, higher-order molecular aggregate(s). The supramolecular assembly of proline is found to possess an amphipathic character. Formation of higher-order aggregates is believed to be crucial for proline to function as a protein folding aid. In addition to its role in osmoregulation under water stress conditions, the results of this study hint at the possibility of proline behaving as a protein folding chaperone.

The three-dimensional structure of the anti-apoptotic protein Bcl-xL complexed to a 25-residue peptide from the death promoting region of Bad was determined using NMR spectroscopy. Although the overall structure is similar to Bcl-xL bound to a 16-residue peptide from the Bak protein (Sattler et al., 1997), the Bad peptide forms additional interactions with Bcl-xL. However, based upon site-directed mutagenesis experiments, these additional contacts do not account for the increased affinity of the Bad 25-mer for Bcl-xL compared to the Bad 16-mer. Rather, the increased helix propensity of the Bad 25-mer is primarily responsible for its greater affinity for Bcl-xL. Based on this observation, a pair of 16-residue peptides were designed and synthesized that were predicted to have a high helix propensity while maintaining the interactions important for complexation with Bcl-xL. Both peptides showed an increase in helix propensity compared to the wild-type and exhibited an enhanced affinity for Bcl-xL.

The solution structure and stability of N-terminally truncated β2-microglobulin (ΔN6β2-m), the major modification in ex vivo fibrils, have been investigated by a variety of biophysical techniques. The results show that ΔN6β2-m has a free energy of stabilization that is reduced by 2.5 kcal/mol compared to the intact protein. Hydrogen exchange of a mixture of the truncated and full-length proteins at μM concentrations at pH 6.5 monitored by electrospray mass spectrometry reveals that ΔN6β2-m is significantly less protected than its wild-type counterpart. Analysis of ΔN6β2-m by NMR shows that this loss of protection occurs in β strands I, III, and part of II. At mM concentration gel filtration analysis shows that ΔN6β2-m forms a series of oligomers, including trimers and tetramers, and NMR analysis indicates that strand V is involved in intermolecular interactions that stabilize this association. The truncated species of β2-microglobulin was found to have a higher tendency to self-associate than the intact molecule, and unlike wild-type protein, is able to form amyloid fibrils at physiological pH. Limited proteolysis experiments and analysis by mass spectrometry support the conformational modifications identified by NMR and suggest that ΔN6β2-m could be a key intermediate of a proteolytic pathway of β2-microglobulin. Overall, the data suggest that removal of the six residues from the N-terminus of β2-microglobulin has a major effect on the stability of the overall fold. Part of the tertiary structure is preserved substantially by the disulfide bridge between Cys25 and Cys80, but the pairing between β-strands far removed from this constrain is greatly perturbed.

One hundred-forty-five full-length aldehyde dehydrogenase-related sequences were aligned to determine relationships within the aldehyde dehydrogenase (ALDH) extended family. The alignment reveals only four invariant residues: two glycines, a phenylalanine involved in NAD binding, and a glutamic acid that coordinates the nicotinamide ribose in certain E-NAD binary complex crystal structures, but which may also serve as a general base for the catalytic reaction. The cysteine that provides the catalytic thiol and its closest neighbor in space, an asparagine residue, are conserved in all ALDHs with demonstrated dehydrogenase activity. Sixteen residues are conserved in at least 95% of the sequences; 12 of these cluster into seven sequence motifs conserved in almost all ALDHs. These motifs cluster around the active site of the enzyme. Phylogenetic analysis of these ALDHs indicates at least 13 ALDH families, most of which have previously been identified but not grouped separately by alignment. ALDHs cluster into two main trunks of the phylogenetic tree. The largest, the “Class 3” trunk, contains mostly substrate-specific ALDH families, as well as the class 3 ALDH family itself. The other trunk, the “Class 1/2” trunk, contains mostly variable substrate ALDH families, including the class 1 and 2 ALDH families. Divergence of the substrate-specific ALDHs occurred earlier than the division between ALDHs with broad substrate specificities. A site on the World Wide Web has also been devoted to this alignment project.

The room-temperature liquid salt, ethylammonium nitrate (EAN), has been used to enhance the recovery of denatured-reduced hen egg white lysozyme (HEWL). Our results show that EAN has the ability to prevent aggregation of the denatured protein. The use of EAN as a refolding additive is advantageous because the renaturation is a one-step process. When HEWL was denatured reduced using routine procedures and renatured using EAN as an additive, HEWL was found to regain 75% of its activity. When HEWL was denatured and reduced in neat EAN, dilution resulted in over 90% recovery of active protein. An important aspect of this process is that renaturation of HEWL occurs at concentrations of 1.6 mg/mL, whereas other renaturation processes occur at significantly lower protein concentrations. Additionally, the refolded-active protein can be separated from the molten salt by simple desalting methods. Although the use of a low-temperature molten salt in protein renaturation is unconventional, the power of this approach lies in its simplicity and utility.

Left-handed polyproline II helices (PPII) are contiguous elements of protein secondary structure in which the φ and ψ angles of constituent residues are restricted to around −75° and 145°, respectively. They are important in structural proteins, in unfolded states and as ligands for signaling proteins. Here, we present a survey of 274 nonhomologous polypeptide chains from proteins of known structure for regions that form these structures. Such regions are rare, but the majority of proteins contain at least one PPII helix. Most PPII helices are shorter than five residues, although the longest found contained 12 amino acids. Proline predominates in PPII, but Gln and positively charged residues are also favored. The basis of Gln's prevalence is its ability to form an i, i + 1 side-chain to main-chain hydrogen bond with the backbone carbonyl oxygen of the proceeding residue; this helps to fix the ψ angle of the Gln and the φ and ψ of the proceeding residue in PPII conformations and explains why Gln is favored at the first position in a PPII helix. PPII helices are highly solvent exposed, which explains why apolar amino acids are disfavored despite preferring this region of φ/ψ space when in isolation. PPII helices have perfect threefold rotational symmetry and within these structures we find significant correlation between the hydrophobicity of residues at i and i + 3; thus, PPII helices in globular proteins can be considered to be amphipathic.