Andrey Krokhotin

RNA-Puzzles Round II: assessment of RNA structure prediction programs applied to three large RNA structures

This paper is a report of a second round of RNA-Puzzles, a collective and blind experiment in three-dimensional (3D) RNA structure prediction. Three puzzles, Puzzles 5, 6, and 10, represented sequences of three large RNA structures with limited or no homology with previously solved RNA molecules. A lariat-capping ribozyme, as well as riboswitches complexed to adenosylcobalamin and tRNA, were predicted by seven groups using RNAComposer, ModeRNA/SimRNA, Vfold, Rosetta, DMD, MC-Fold, 3dRNA, and AMBER refinement. Some groups derived models using data from state-of-the-art chemical-mapping methods (SHAPE, DMS, CMCT, and mutate-and-map). The comparisons between the predictions and the three subsequently released crystallographic structures, solved at diffraction resolutions of 2.5-3.2 Å, were carried out automatically using various sets of quality indicators. The comparisons clearly demonstrate the state of present-day de novo prediction abilities as well as the limitations of these state-of-the-art methods. All of the best prediction models have similar topologies to the native structures, which suggests that computational methods for RNA structure prediction can already provide useful structural information for biological problems. However, the prediction accuracy for non-Watson-Crick interactions, key to proper folding of RNAs, is low and some predicted models had high Clash Scores. These two difficulties point to some of the continuing bottlenecks in RNA structure prediction. All submitted models are available for download at http://ahsoka.u-strasbg.fr/rnapuzzles/.

iFoldRNA v2: Folding RNA with Constraints

UNLABELLED A key to understanding RNA function is to uncover its complex 3D structure. Experimental methods used for determining RNA 3D structures are technologically challenging and laborious, which makes the development of computational prediction methods of substantial interest. Previously, we developed the iFoldRNA server that allows accurate prediction of short (<50 nt) tertiary RNA structures starting from primary sequences. Here, we present a new version of the iFoldRNA server that permits the prediction of tertiary structure of RNAs as long as a few hundred nucleotides. This substantial increase in the server capacity is achieved by utilization of experimental information such as base-pairing and hydroxyl-radical probing. We demonstrate a significant benefit provided by integration of experimental data and computational methods. AVAILABILITY AND IMPLEMENTATION http://ifoldrna.dokhlab.org CONTACT dokh@unc.eu.

Coexistence of Phases in a Protein Heterodimer

A heterodimer consisting of two or more different kinds of proteins can display an enormous number of distinct molecular architectures. The conformational entropy is an essential ingredient in the Helmholtz free energy and, consequently, these heterodimers can have a very complex phase structure. Here, it is proposed that there is a state of proteins, in which the different components of a heterodimer exist in different phases. For this purpose, the structures in the protein data bank (PDB) have been analyzed, with radius of gyration as the order parameter. Two major classes of heterodimers with their protein components coexisting in different phases have been identified. An example is the PDB structure 3DXC. This is a transcriptionally active dimer. One of the components is an isoform of the intra-cellular domain of the Alzheimer-disease related amyloid precursor protein (AICD), and the other is a nuclear multidomain adaptor protein in the Fe65 family. It is concluded from the radius of gyration that neither of the two components in this dimer is in its own collapsed phase, corresponding to a biologically active protein. The UNRES energy function has been utilized to confirm that, if the two components are separated from each other, each of them collapses. The results presented in this work show that heterodimers whose protein components coexist in different phases, can have intriguing physical properties with potentially important biological consequences.

Kinks, loops, and protein folding, with protein A as an example.

The dynamics and energetics of formation of loops in the 46-residue N-terminal fragment of the B-domain of staphylococcal protein A has been studied. Numerical simulations have been performed using coarse-grained molecular dynamics with the united-residue (UNRES) force field. The results have been analyzed in terms of a kink (heteroclinic standing wave solution) of a generalized discrete nonlinear Schrödinger (DNLS) equation. In the case of proteins, the DNLS equation arises from a C(α)-trace-based energy function. Three individual kink profiles were identified in the experimental three-α-helix structure of protein A, in the range of the Glu16-Asn29, Leu20-Asn29, and Gln33-Asn44 residues, respectively; these correspond to two loops in the native structure. UNRES simulations were started from the full right-handed α-helix to obtain a clear picture of kink formation, which would otherwise be blurred by helix formation. All three kinks emerged during coarse-grained simulations. It was found that the formation of each is accompanied by a local free energy increase; this is expressed as the change of UNRES energy which has the physical sense of the potential of mean force of a polypeptide chain. The increase is about 7 kcal/mol. This value can thus be considered as the free energy barrier to kink formation in full α-helical segments of polypeptide chains. During the simulations, the kinks emerge, disappear, propagate, and annihilate each other many times. It was found that the formation of a kink is initiated by an abrupt change in the orientation of a pair of consecutive side chains in the loop region. This resembles the formation of a Bloch wall along a spin chain, where the C(α) backbone corresponds to the chain, and the amino acid side chains are interpreted as the spin variables. This observation suggests that nearest-neighbor side chain-side chain interactions are responsible for initiation of loop formation. It was also found that the individual kinks are reflected as clear peaks in the principal modes of the analyzed trajectory of protein A, the shapes of which resemble the directional derivatives of the kinks along the chain. These observations suggest that the kinks of the DNLS equation determine the functionally important motions of proteins.

On the role of thermal backbone fluctuations in myoglobin ligand gate dynamics

We construct an energy function that describes the crystallographic structure of sperm whale myoglobin backbone. As a model in our construction, we use the Protein Data Bank entry 1ABS that has been measured at liquid helium temperature. Consequently, the thermal B-factor fluctuations are very small, which is an advantage in our construction. The energy function that we utilize resembles that of the discrete nonlinear Schrödinger equation. Likewise, ours supports topological solitons as local minimum energy configurations. We describe the 1ABS backbone in terms of topological solitons with a precision that deviates from 1ABS by an average root-mean-square distance, which is less than the experimentally observed Debye-Waller B-factor fluctuation distance. We then subject the topological multi-soliton solution to extensive numerical heating and cooling experiments, over a very wide range of temperatures. We concentrate in particular to temperatures above 300 K and below the Θ-point unfolding temperature, which is around 348 K. We confirm that the behavior of the topological multi-soliton is fully consistent with Anfinsens thermodynamic principle, up to very high temperatures. We observe that the structure responds to an increase of temperature consistently in a very similar manner. This enables us to characterize the onset of thermally induced conformational changes in terms of three distinct backbone ligand gates. One of the gates is made of the helix F and the helix E. The two other gates are chosen similarly, when open they provide a direct access route for a ligand to reach the heme. We find that out of the three gates we investigate, the one which is formed by helices B and G is the most sensitive to thermally induced conformational changes. Our approach provides a novel perspective to the important problem of ligand entry and exit.

Solitons and collapse in the λ-repressor protein.

The enterobacteria lambda phage is a paradigm temperate bacteriophage. Its lysogenic and lytic life cycles echo competition between the DNA binding λ-repressor (CI) and CRO proteins. Here we scrutinize the structure, stability, and folding pathways of the λ-repressor protein, which controls the transition from the lysogenic to the lytic state. We first investigate the supersecondary helix-loop helix composition of its backbone. We use a discrete Frenet framing to resolve the backbone spectrum in terms of bond and torsion angles. Instead of four, there appears to be seven individual loops. We model the putative loops using an explicit soliton Ansatz. It is based on the standard soliton profile of the continuum nonlinear Schrödinger equation. The accuracy of the Ansatz far exceeds the B-factor fluctuation distance accuracy of the experimentally determined protein configuration. We then investigate the folding pathways and dynamics of the λ-repressor protein. We introduce a coarse-grained energy function to model the backbone in terms of the C(α) atoms and the side chains in terms of the relative orientation of the C(β) atoms. We describe the folding dynamics in terms of relaxation dynamics and find that the folded configuration can be reached from a very generic initial configuration. We conclude that folding is dominated by the temporal ordering of soliton formation. In particular, the third soliton should appear before the first and second. Otherwise, the DNA binding turn does not acquire its correct structure. We confirm the stability of the folded configuration by repeated heating and cooling simulations.

Predicting the functional consequences of non-synonymous single nucleotide polymorphisms in IL8 gene

Here we report an in-silico approach for identification, characterization and validation of deleterious non-synonymous SNPs (nsSNPs) in the interleukin-8 gene using three steps. In first step, sequence homology-based genetic analysis of a set of 50 coding SNPs associated with 41 rsIDs using SIFT (Sorting Intolerant from Tolerant) and PROVEAN (Protein Variation Effect Analyzer) identified 23 nsSNPs to be putatively damaging/deleterious in at least one of the two tools used. Subsequently, structure-homology based PolyPhen-2 (Polymorphism Phenotyping) analysis predicted 9 of 23 nsSNPs (K4T, E31A, E31K, S41Y, I55N, P59L, P59S, L70P and V88D) to be damaging. According to the conditional hypothesis for the study, only nsSNPs that score damaging/deleterious prediction in both sequence and structural homology-based approach will be considered as ‘high-confidence’ nsSNPs. In step 2, based on conservation of amino acid residues, stability analysis, structural superimposition, RSMD and docking analysis, the possible structural-functional relationship was ascertained for high-confidence nsSNPs. Finally, in a separate analysis (step 3), the IL-8 deregulation has also appeared to be an important prognostic marker for detection of patients with gastric and lung cancer. This study, for the first time, provided in-depth insights on the effects of amino acid substitutions on IL-8 protein structure, function and disease association.

DIRECT IDENTIFICATION OF BASE-PAIRED RNA NUCLEOTIDES BY CORRELATED CHEMICAL PROBING

Many RNA molecules fold into complex secondary and tertiary structures that play critical roles in biological function. Among the best-established methods for examining RNA structure are chemical probing experiments, which can report on local nucleotide structure in a concise and extensible manner. While probing data are highly useful for inferring overall RNA secondary structure, these data do not directly measure through-space base-pairing interactions. We recently introduced an approach for single-molecule correlated chemical probing with dimethyl sulfate (DMS) that measures RNA interaction groups by mutational profiling (RING-MaP). RING-MaP experiments reveal diverse through-space interactions corresponding to both secondary and tertiary structure. Here we develop a framework for using RING-MaP data to directly and robustly identify canonical base pairs in RNA. When applied to three representative RNAs, this framework identified 20%-50% of accepted base pairs with a <10% false discovery rate, allowing detection of 88% of duplexes containing four or more base pairs, including pseudoknotted pairs. We further show that base pairs determined from RING-MaP analysis significantly improve secondary structure modeling. RING-MaP-based correlated chemical probing represents a direct, experimentally concise, and accurate approach for detection of individual base pairs and helices and should greatly facilitate structure modeling for complex RNAs.

Computational Methods Toward Accurate RNA Structure Prediction Using Coarse-Grained and All-Atom Models

Computational methods can provide significant insights into RNA structure and dynamics, bridging the gap in our understanding of the relationship between structure and biological function. Simulations enrich and enhance our understanding of data derived on the bench, as well as provide feasible alternatives to costly or technically challenging experiments. Coarse-grained computational models of RNA are especially important in this regard, as they allow analysis of events occurring in timescales relevant to RNA biological function, which are inaccessible through experimental methods alone. We have developed a three-bead coarse-grained model of RNA for discrete molecular dynamics simulations. This model is efficient in de novo prediction of short RNA tertiary structure, starting from RNA primary sequences of less than 50 nucleotides. To complement this model, we have incorporated additional base-pairing constraints and have developed a bias potential reliant on data obtained from hydroxyl probing experiments that guide RNA folding to its correct state. By introducing experimentally derived constraints to our computer simulations, we are able to make reliable predictions of RNA tertiary structures up to a few hundred nucleotides. Our refined model exemplifies a valuable benefit achieved through integration of computation and experimental methods.

Protein folding: Over half a century lasting quest: Comment on “There and back again: Two views on the protein folding puzzle” by Alexei V. Finkelstein et al.

Most proteins fold into unique three-dimensional (3D) structures that determine their biological functions, such as catalytic activity or macromolecular binding. Misfolded proteins can pose a threat through aberrant interactions with other proteins leading to a number of diseases including Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis [1,2]. What does determine 3D structure of proteins? The first clue to this question came more than fifty years ago when Anfinsen demonstrated that unfolded proteins can spontaneously fold to their native 3D structures [3,4]. Anfinsen’s experiments lead to the conclusion that proteins fold to unique native structure corresponding to the stable and kinetically accessible free energy minimum, and protein native structure is solely determined by its amino acid sequence. The question of how exactly proteins find their free energy minimum proved to be a difficult problem. One of the puzzles, initially pointed out by Levinthal, was an inconsistency between observed protein folding times and theoretical estimates. A self-avoiding polymer model of a globular protein of 100-residues length on a cubic lattice can sample at least 1047 states. Based on the assumption that conformational sampling occurs at the highest vibrational mode of proteins (∼picoseconds), predicted folding time by searching among all the possible conformations leads to ∼ 1027 years (much larger than the age of the universe) [5]. In contrast, observed protein folding time range from microseconds to minutes. Due to tremendous theoretical progress in protein folding field that has been achieved in past decades, the source of this inconsistency is currently understood that is thoroughly described in the review by Finkelstein et al. [6]. Finkelstein and colleagues provide a summary of existing protein folding theories with a special emphasis on the theoretical evaluation of the rates to overcome the free-energy barrier separating the natively folded (N) and unfolded states (U). The authors first make estimates of the free-energy barrier height from the perspective of the folded state by looking into protein unfolding (N-to-U transition). They argue that due to the principle of detailed balance (i.e. the rates of direct and reverse reactions should be equal), the protein folding time can be derived from the time the protein takes to unfold. Then the authors look into the free-energy barrier height from the perspective of the unfolded state (U-to-N transition) and provide rough estimates of the size of the conformational space that the protein samples en route to its native structure. This size is significantly lower than the initial estimate suggested by Levintal, hence