Protein folding in the post-genomic era

Protein folding is a topic of fundamental interest since it concerns the mechanisms by which the genetic message is translated into the three-dimensional and functional structure of proteins. In these post-genomic times, the knowledge of the fundamental principles are required in the exploitation of the information contained in the increasing number of sequenced genomes. Protein folding also has practical applications in the understanding of different pathologies and the development of novel therapeutics to prevent diseases associated with protein misfolding and aggregation. Significant advances have been made ranging from the Anfinsen postulate to the "new view" which describes the folding process in terms of an energy landscape. These new insights arise from both theoretical and experimental studies. The problem of folding in the cellular environment is briefly discussed. The modern view of misfolding and aggregation processes that are involved in several pathologies such as prion and Alzheimer diseases. Several approaches of structure prediction, which is a very active field of research, are described.     

Directionality in protein fold prediction

Background  

Ever since the ground-breaking work of Anfinsen et al. in which a denatured protein was found to refold to its native state, it has been frequently stated by the protein fold prediction community that all the information required for protein folding lies in the amino acid sequence. Recent in vitro experiments and in silico computational studies, however, have shown that cotranslation may affect the folding pathway of some proteins, especially those of ancient folds. In this paper aspects of cotranslational folding have been incorporated into a protein structure prediction algorithm by adapting the Rosetta program to fold proteins as the nascent chain elongates. This makes it possible to conduct a pairwise comparison of folding accuracy, by comparing folds created sequentially from each end of the protein.     

Protein flexibility and rigidity predicted from sequence

Structural flexibility has been associated with various biological processes such as molecular recognition and catalytic activity. In silico studies of protein flexibility have attempted to characterize and predict flexible regions based on simple principles. B-values derived from experimental data are widely used to measure residue flexibility. Here, we present the most comprehensive large-scale analysis of B-values. We used this analysis to develop a neural network-based method that predicts flexible-rigid residues from amino acid sequence. The system uses both global and local information (i.e., features from the entire protein such as secondary structure composition, protein length, and fraction of surface residues, and features from a local window of sequence-consecutive residues). The most important local feature was the evolutionary exchange profile reflecting sequence conservation in a family of related proteins. To illustrate its potential, we applied our method to 4 different case studies, each of which related our predictions to aspects of function. The first 2 were the prediction of regions that undergo conformational switches upon environmental changes (switch II region in Ras) and the prediction of surface regions, the rigidity of which is crucial for their function (tunnel in propeller folds). Both were correctly captured by our method. The third study established that residues in active sites of enzymes are predicted by our method to have unexpectedly low B-values. The final study demonstrated how well our predictions correlated with NMR order parameters to reflect motion. Our method had not been set up to address any of the tasks in those 4 case studies. Therefore, we expect that this method will assist in many attempts at inferring aspects of function.     

Toward the solution of the protein structure prediction problem

Since Anfinsen demonstrated that the information encoded in a protein’s amino acid sequence determines its structure in 1973, solving the protein structure prediction problem has been the Holy Grail of structural biology. The goal of protein structure prediction approaches is to utilize computational modeling to determine the spatial location of every atom in a protein molecule starting from only its amino acid sequence. Depending on whether homologous structures can be found in the Protein Data Bank (PDB), structure prediction methods have been historically categorized as template-based modeling (TBM) or template-free modeling (FM) approaches. Until recently, TBM has been the most reliable approach to predicting protein structures, and in the absence of reliable templates, the modeling accuracy sharply declines. Nevertheless, the results of the most recent community-wide assessment of protein structure prediction experiment (CASP14) have demonstrated that the protein structure prediction problem can be largely solved through the use of end-to-end deep machine learning techniques, where correct folds could be built for nearly all single-domain proteins without using the PDB templates. Critically, the model quality exhibited little correlation with the quality of available template structures, as well as the number of sequence homologs detected for a given target protein. Thus, the implementation of deep-learning techniques has essentially broken through the 50-year-old modeling border between TBM and FM approaches and has made the success of high-resolution structure prediction significantly less dependent on template availability in the PDB library.     

Optimization of designed armadillo repeat proteins by molecular dynamics simulations and NMR spectroscopy

A multidisciplinary approach based on molecular dynamics (MD) simulations using homology models, NMR spectroscopy, and a variety of biophysical techniques was used to efficiently improve the thermodynamic stability of armadillo repeat proteins (ArmRPs). ArmRPs can form the basis of modular peptide recognition and the ArmRP version on which synthetic libraries are based must be as stable as possible. The 42-residue internal Arm repeats had been designed previously using a sequence-consensus method. Heteronuclear NMR revealed unfavorable interactions present at neutral but absent at high pH. Two lysines per repeat were involved in repulsive interactions, and stability was increased by mutating both to glutamine. Five point mutations in the capping repeats were suggested by the analysis of positional fluctuations and configurational entropy along multiple MD simulations. The most stabilizing single C-cap mutation Q240L was inferred from explicit solvent MD simulations, in which water penetrated the ArmRP. All mutants were characterized by temperature- and denaturant-unfolding studies and the improved mutants were established as monomeric species with cooperative folding and increased stability against heat and denaturant. Importantly, the mutations tested resulted in a cumulative decrease of flexibility of the folded state in silico and a cumulative increase of thermodynamic stability in vitro. The final construct has a melting temperature of about 85°C, 14.5° higher than the starting sequence. This work indicates that in silico studies in combination with heteronuclear NMR and other biophysical tools may provide a basis for successfully selecting mutations that rapidly improve biophysical properties of the target proteins.     

Promiscuous protein biotinylation by Escherichia coli biotin protein ligase

Biotin protein ligases (BPLs) are enzymes of extraordinary specificity. BirA, the BPL of Escherichia coli biotinylates only a single cellular protein. We report a mutant BirA that attaches biotin to a large number of cellular proteins in vivo and to bovine serum albumin, chloramphenicol acetyltransferase, immunoglobin heavy and light chains, and RNAse A in vitro. The mutant BirA also self biotinylates in vivo and in vitro. The wild type BirA protein is much less active in these reactions. The biotinylation reaction is proximity-dependent in that a greater extent of biotinylation was seen when the mutant ligase was coupled to the acceptor proteins than when the acceptors were free in solution. This approach may permit facile detection and recovery of interacting proteins by existing avidin/streptavidin technology.     

预测蛋白质可结晶性研究进展

解析蛋白质的三维结构具有重要的生物学意义,更是蛋白质功能研究和理性药物设计的基础。目前解析蛋白质结构最重要的方法是X-射线衍射晶体学解析技术。但是运用该技术解析蛋白质结构的关键是获得高质量的蛋白质晶体。然而,据统计仅有42%的可溶纯化蛋白质能够得到晶体,即不同蛋白质的可结晶性表现不同。由于实验方法验证蛋白质的可结晶性耗时耗力,因此,有研究者运用计算机模拟的方法预测蛋白质的可结晶性,从而节省资源与成本并且提高实验的成功率。本文结合我们的研究工作,介绍了几种目前较为成功的蛋白质可结晶性预测方法及其研究途径。     

Protein folding: in vitro studies

Protein folding is a topic of fundamental interest since it concerns the mechanisms by which the genetic information is translated into the three-dimensional and functional structure of proteins. In these post-genomic times, the knowledge of the fundamental principles is required in the exploitation of the information contained in the increasing number of sequenced genomes. Protein folding also has a practical application in the understanding of different pathologies associated with protein misfolding and aggregation. Significant advances have been made ranging from the Anfinsen postulate to the "new view" which describes the folding process in terms of an energy landscape. These insights arise from both theoretical and experimental studies. Unravelling the mechanisms of protein folding represents one of the most challenging problems to day. This is an extremely active field of research involving aspects of biology, chemistry, biochemistry, computer science and physics.     

Accurate computer-based design of a new backbone conformation in the second turn of protein L.

The rational design of loops and turns is a key step towards creating proteins with new functions. We used a computational design procedure to create new backbone conformations in the second turn of protein L. The Protein Data Bank was searched for alternative turn conformations, and sequences optimal for these turns in the context of protein L were identified using a Monte Carlo search procedure and an energy function that favors close packing. Two variants containing 12 and 14 mutations were found to be as stable as wild-type protein L. The crystal structure of one of the variants has been solved at a resolution of 1.9 A, and the backbone conformation in the second turn is remarkably close to that of the in silico model (1.1 A RMSD) while it differs significantly from that of wild-type protein L (the turn residues are displaced by an average of 7.2 A). The folding rates of the redesigned proteins are greater than that of the wild-type protein and in contrast to wild-type protein L the second beta-turn appears to be formed at the rate limiting step in folding.     

TPO1, a Member of a Novel Protein Family, Is Developmentally Regulated in Cultured Oligodendrocytes

Abstract: Although the myelin membrane contains only a small set of major proteins, more sensitive assays indicate the presence of a plethora of uncharacterized proteins. We have used an antibody perturbation approach to reversibly block the differentiation of prooligodendroblasts into myelinating cells, and, in combination with a differential screening procedure, identified novel mRNAs that are activated during this period. One cDNA, TPO1, recognizes a 5.5-kb mRNA that is strongly up-regulated in oligodendrocytes after release of the differentiation block and that is expressed at high levels in brain tissue during active myelination. This cDNA represents at least two mRNAs differing from each other in their 5'-termini. The TPO1 cDNA contains an open reading frame of 1,380 bp, encoding a protein of 51.8 kDa with a predicted pl of 9.1 that contains two regions homologous to nonclassic zinc finger motifs. Subcellular localization studies suggest the enriched presence of TPO1 in spherical structures along the major cytoplasmic processes of oligodendrocytes. TPO1, along with homologues expressed in testis, placenta, and PC12 cells, form a novel family of proteins with multiple hydrophobic domains possibly serving as membrane spanning regions. We postulate that in oligodendrocytes, TPO1 encodes a protein factor involved in myelin biogenesis.     

Exploring the kinetic requirements for enhancement of protein folding rates in the GroEL cavity

The chaperonin system, GroEL and GroES of Escherichia coli enable certain proteins to fold under conditions when spontaneous folding is prohibitively slow as to compete with other non-productive channels such as aggregation. We investigated the plausible mechanisms of GroEL-mediated folding using simple lattice models. In particular, we have investigated protein folding in a confined environment, such as those offered by the GroEL, to decipher whether rate and yield enhancement can occur when the substrate protein is allowed to fold within the cavity of the chaperonins. The GroEL cavity is modeled as a cubic box and a simple bead model is used to represent the substrate chain. We consider three distinct characteristic of the confining environment. First, the cavity is taken to be a passive Anfinsen cage in which the walls merely reduce the available conformation space. We find that at temperatures when the native conformation is stable, the folding rate is retarded in the Anfinsen cage. We then assumed that the interior of the wall is hydrophobic. In this case the folding times exhibit a complex behavior. When the strength of the interaction between the polypeptide chain and the cavity is too strong or too weak we find that the rates of folding are retarded compared to spontaneous folding. There is an optimum range of the interaction strength that enhances the rates. Thus, above this value there is an inverse correlation between the folding rates and the strength of the substrate-cavity interactions. The optimal hydrophobic walls essentially pull the kinetically trapped states which leads to a smoother the energy landscape. It is known that upon addition of ATP and GroES the interior cavity of GroEL offers a hydrophilic-like environment to the substrate protein. In order to mimic this within the context of the dynamic Anfinsen cage model, we allow for changes in the hydrophobicity of the walls of the cavity. The duration for which the walls remain hydrophobic during one cycle of ATP hydrolysis is allowed to vary. These calculations show that frequent cycling of the wall hydrophobicity can dramatically reduce the folding times and increase the yield as well under non-permissive conditions. Examination of the structures of the substrate proteins before and after the change in hydrophobicity indicates that there is global unfolding involved. In addition, it is found that a fraction of the molecules kinetically partition to the native state in accordabce with the iterative annealing mechanism. Thus, frequent "unfoldase" activity of chaperonins leading to global unfolding of the polypeptide chain results in enhancement of the folding rates and yield of the folded protein. We suggest that chaperonin efficiency can be greatly enhanced if the cycling time is reduced. The calculations are used to interpret a few experiments on chaperonin-mediated protein folding.     

Inhibitory effects of RraA and RraB on RNAse E-related enzymes imply conserved functions in the regulated enzymatic cleavage of RNA

RraA and RraB are recently discovered protein inhibitors of RNAse E, which forms a large protein complex termed the degradosome that catalyzes the initial step in the decay and processing of numerous RNAs in Escherichia coli . Here, we report that these E. coli protein inhibitors physically interact with RNAse ES, a Streptomyces coelicolor functional ortholog of RNAse E, and inhibit its action in vivo as well as in vitro ; however, unlike their ability to differentially modulate E. coli RNAse E action in a substrate-dependent manner by altering the composition of the degradosome, both proteins appear to have a general inhibitory effect on the ribonucleolytic activity of RNAse ES, which does not interact with E. coli polynucleotide phosphorylase, a major component of the degradosome. Our findings suggest that these regulators of RNAse activity have a conserved intrinsic property enabling them to directly act on RNAse E-related enzymes and inhibit their general ribonucleolytic activity.     

Incompatibility of outer membrane proteins OmpA and OmpF of Escherichia coli with secretion in Bacillus subtilis: Fusions with secretable peptides

The secretion of the outer membrane proteins OmpA and OmpF of Escherichia coli has previously been found to be blocked at an early intracellular step, when these proteins were fused to a bacillar signal sequence and expressed in Bacillus subtilis. We have now fused these proteins to long secretable polypeptides, the amino-terminal portions of alpha-amylase or beta-lactamase. In spite of this, no secretion of the fusion proteins was detected in B. subtilis. With the exception of a small fraction of the beta-lactamase fusion, the proteins were cell-bound with uncleaved signal sequences. Protease accessibility indicated that the fusion proteins were not even partially exposed on the outer surface of the cytoplasmic membrane. Thus there was no change of the location compared to the OmpA or OmpF fused to the signal sequence only. We conclude that, like OmpA and OmpF, the fusion proteins fold into an export-incompatible conformation in B. subtilis before the start of translocation, which we postulate to be a late post-translational event.     

Intracellular signalling proteins as ‘smart’ agents in parallel distributed processes

In eucaryotic organisms, responses to external signals are mediated by a repertoire of intracellular signalling pathways that ultimately bring about the activation/inactivation of protein kinases and/or protein phosphatases. Until relatively recently, little thought had been given to the intracellular distribution of the components of these signalling pathways. However, experimental evidence from a diverse range of organisms indicates that rather than being freely distributed, many of the protein components of signalling cascades show a significant degree of spatial organisation. Here, we briefly review the roles of ‘anchor’, ‘scaffold’ and ‘adaptor’ proteins in the organisation and functioning of intracellular signalling pathways. We then consider some of the parallel distributed processing capacities of these adaptive systems. We focus on signalling proteins-both as individual ‘devices’ (agents) and as ‘networks’ (ecologies) of parallel processes. Signalling proteins are described as ‘smart thermodynamic machines’ which satisfy ‘gluing’ (functorial) roles in the information economy of the cell. This combines two information-processing views of signalling proteins. Individually, they show ‘cognitive’ capacities and collectively they integrate (cohere) cellular processes. We exploit these views by drawing comparisons between signalling proteins and verbs. This text/dialogical metaphor also helps refine our view of signalling proteins as context-sensitive information processing agents.     

Folding properties of the hepatitis B core as a carrier protein for vaccination research

The hepatitis B core (HBc) protein has been used successfully in numerous experiments as a carrier for heterologous peptides. Folding and capsid formation of the chimeric proteins is not always achieved easily. In silico analyses were performed to provide further comprehension of the feasibility for predicting successful capsid formation. In contrast to previous work, we show that common in silico predictions do not ensure assembly into particles. We included new considerations regarding capsid formation of HBc fusion proteins. Not only the primary sequence and the length of the inserts seem important, also the rigidity, the distance between the N and the C-terminus and the presence of cysteines, which could form disulphide bonds, could influence proper capsid formation. Furthermore, new conformational insights were formulated when linkers were added to create extra flexibility of the chimeric particles. Different hypotheses were suggested to clarify the obtained results. To this extent, the addition of glycine-rich linkers could lower high rigidity of the insert, removal of the strain of the core protein or ease interaction between the HBc and the insert. Finally, we observed specific changes in capsid formation properties when longer linkers were used. These findings have not been reported before in this and other virus-like particle carriers. In this study, we also propose a new high-yield purification protocol for fusion proteins to be used in vaccination experiments with the carrier protein or in comparative studies of particulate or non-particulate HBc fusion proteins.     

Multiple pore lining residues modulate water permeability of GlpF

The water permeability of aquaporins (AQPs) varies by more than an order of magnitude even though the pore structure, geometry, as well as the channel lining residues are highly conserved. However, channel gating by pH, divalent ions or phosphorylation was only shown for a minority of AQPs. Structural and in silico indications of water flux modulation by flexible side chains of channel lining residues have not been experimentally confirmed yet. Hence, the aquaporin “open state” is still considered to be a continuously open pore with water molecules permeating in a single‐file fashion. Using protein mutations outside the selectivity filter in the aqua(glycerol)facilitator GlpF of Escherichia coli we, to the best of our knowledge, for the first time, modulate the position of the highly conserved Arg in the selectivity filter. This in turn enhances or reduces the unitary water permeability of GlpF as shown in silico by molecular dynamics (MD) simulations and in vitro with purified and reconstituted GlpF. This finding suggests that AQP water permeability can indeed be regulated by lipid bilayer asymmetry and the transmembrane potential. Strikingly, our long‐term MD simulations reveal that not only the conserved Arg in the selectivity filter, but the position and dynamics of multiple other pore lining residues modulate water passage through GlpF. This finding is expected to trigger a wealth of future investigations on permeability and regulation of AQPs among others with the aim to tune water permeability for biotechnological applications.     

Computational diagnosis of protein conformational diseases: short molecular dynamics simulations reveal a fast unfolding of r-LDL mutants that cause familial hypercholesterolemia

The molecular basis of conformational diseases frequently resides in mutant proteins constituting a subset of the vast mutational space. While the subtleties of protein structure point to molecular dynamics (MD) techniques as promising tools for an efficient exploration of such a space, the average size of proteins and the time scale of unfolding events make this goal difficult with present computational capabilities. We show here, nevertheless, that an efficient approach is already feasible for modular proteins. Familial hypercholesterolemia (FH) is a conformational disease linked to mutations in the gene encoding the low density lipoprotein receptor. A high percentage of these mutations has been found in the seven small modular binding repeats of the receptor. Taking advantage of its small size, we have performed an in depth MD study of the fifth binding repeat. Fast unfolding dynamics have been observed in the absence of a structural bound calcium ion, which agrees with its reported essential role in the stability of the module. In addition, several mutations detected in FH patients have been analyzed, starting from the native conformation. Our results indicate that in contrast with the wild type protein and an innocuous control mutant, disease-related mutants experience, in short simulation times (2-8 ns), gross departures from the native state that lead to unfolded conformations and, in some cases, to binding site desorganization deriving in calcium release. Computational diagnosis of mutations leading to conformational diseases seems thus feasible, at least for small or modular pathogenic proteins.     

On the evolutionary origin of the chaperonins

An analysis of the apical domain of the Group-I and Group-II chaperonins shows that they have structural similarities to two different protein folds: a "swivel-domain" phosphotransferase and a thioredoxin-like peroxiredoxin. There is no significant sequence similarity that supports either similarity and the degree of similarity based on structure is comparable but weak for both relationships. Based on possible evolutionary transitions, we deduced that a phosphotransferase origin would require both a large insertion and deletion of structure whereas a peroxiredoxin origin requires only a peripheral rearrangement, similar to an internal domain-swap. We postulate that this change could have been triggered by the insertion of a peroxiredoxin into the ATPase domain that led to the modern chaperonin domain arrangement. The peroxidoxin fold is the most highly embellished member of the thioredoxin super-family and the insertion event may have "overloaded" the core, leading to a rearrangement. A peroxiredoxin origin for the domain also provides a functional explanation, as the peroxiredoxins can act as chaperones when they adopt a multimeric ring complex, similar to the chaperonin subunit configuration. In addition, several of the GroEL apical domain hydrophobic residues which interact with the unfolded protein are located in a position that corresponds to the protein substrate binding region of the peroxiredoxin fold. We suggest that the origin of the ur-chaperonin from a thioredoxin/peroxiredoxin fold might also account for the number of thioredoxin-fold containing proteins that interact with chaperonins, such as tubulin and phosducin-like proteins.     

Molecular dynamics simulations highlight mobile regions in proteins: A novel suggestion for converting a murine V(H) domain into a more tractable species

The V(H) region of the murine antibody 1F7 has been identified as a single-domain chorismate mutase, but a tendency to denature and aggregate has hampered its biochemical characterization. Standard mutagenesis approaches targeting antibody chain dimerization areas have been exhausted. We describe a new approach to the problem, where we use molecular dynamics (MD) simulations to find the differences between the untractable protein and the known soluble V(H) domain from a llama antibody. MD simulations of proteins yield information on the relative stability and fluctuations of parts of the proteins. By comparing simulation results of two related proteins their differences in stability and fluctuations can be analyzed and may suggest mutations aimed at (de)stabilization of one of the two proteins. For the mouse versus llama simulations, this approach highlights an untried area in the protein which shows increased fluctuations. The replacement of this eight-residue segment with the corresponding llama sequence gave a chimeric mutant which shows significantly decreased fluctuations. We see this as a general scheme to generate suggestions for mutagenesis experiments, not only obviously generalizable to other immunoglobulin domains, but to other protein systems as well.