Homology modeling using parametric alignment ensemble generation with consensus and energy-based model selection

Development and assessment of novel machine learning models to predict medication non-adherence risks in type 2 diabetics

Background

Medication adherence is the main determinant of effective management of type 2 diabetes, yet there is no gold standard method available to screen patients with high-risk non-adherence. Developing machine learning models to predict high-risk non-adherence in patients with T2D could optimize management.

Methods

This cross-sectional study was carried out on patients with T2D at the Sichuan Provincial People's Hospital from April 2018 to December 2019 who were examined for HbA1c on the day of the survey. Demographic and clinical characteristics were extracted from the questionnaire and electronic medical records. The sample was randomly divided into a training dataset and a test dataset with a radio of 8:2 after data preprocessing. Four imputing methods, five sampling methods, three screening methods, and 18 machine learning algorithms were used to groom data and develop and validate models. Bootstrapping was performed to generate the validation set for external validation and univariate analysis. Models were compared on the basis of predictive performance metrics. Finally, we validated the sample size on the best model.

Results

This study included 980 patients with T2D, of whom 184 (18.8%) were defined as medication non-adherence. The results indicated that the model used modified random forest as the imputation method, random under sampler as the sampling method, Boruta as the feature screening method and the ensemble algorithms and had the best performance. The area under the receiver operating characteristic curve (AUC), F1 score, and area under the precision-recall curve (AUPRC) of the best model, among a total of 1,080 trained models, were 0.8369, 0.7912, and 0.9574, respectively. Age, present fasting blood glucose (FBG) values, present HbA1c values, present random blood glucose (RBG) values, and body mass index (BMI) were the most significant contributors associated with risks of medication adherence.

Conclusion

We found that machine learning methods could be used to predict the risk of non-adherence in patients with T2D. The proposed model was well performed to identify patients with T2D with non-adherence and could help improve individualized T2D management.

Folding Coarse-Grained Oligomer Models with PyRosetta

Non-biological foldamers are a promising class of macromolecules that share similarities to classical biopolymers such as proteins and nucleic acids. Currently, designing novel foldamers is a non-trivial process, often involving many iterations of trial synthesis and characterization until folded structures are observed. In this work, we aim to tackle these foldamer design challenges using computational modeling techniques. We developed CG PyRosetta, an extension to the popular protein folding python package, PyRosetta, which introduces coarse-grained (CG) residues into PyRosetta, enabling the folding of toy CG foldamer models. Although these models are simplified, they can help explore overarching physical hypotheses about how oligomers can form. Through systematic variation of CG parameters in these models, we can investigate various folding hypotheses at the CG scale to inform the design process of new foldamer chemistries. In this study, we demonstrate CG PyRosetta's ability to identify minimum energy structures with a diverse structural search over a range of simple models, as well as two hypothesis-driven parameter scans investigating the effects of side-chain size and internal backbone angle on secondary structures. We are able to identify several types of secondary structures from single- and double-helices to sheet-like and knot-like structures. We show how side-chain size and backbone bond angle both play an important role in the structure and energetics of these toy models. Optimal side-chain sizes promote favorable packing of side chains, while specific backbone bond angles influence the specific helix type found in folded structures.

Novel Allosteric Mechanism of Dual p53/MDM2 and p53/MDM4 Inhibition by a Small Molecule

Restoration of the p53 tumor suppressor for personalised cancer therapy is a promising treatment strategy. However, several high-affinity MDM2 inhibitors have shown substantial side effects in clinical trials. Thus, elucidation of the molecular mechanisms of action of p53 reactivating molecules with alternative functional principle is of the utmost importance. Here, we report a discovery of a novel allosteric mechanism of p53 reactivation through targeting the p53 N-terminus which promotes inhibition of both p53/MDM2 (murine double minute 2) and p53/MDM4 interactions. Using biochemical assays and molecular docking, we identified the binding site of two p53 reactivating molecules, RITA (reactivation of p53 and induction of tumor cell apoptosis) and protoporphyrin IX (PpIX). Ion mobility-mass spectrometry revealed that the binding of RITA to serine 33 and serine 37 is responsible for inducing the allosteric shift in p53, which shields the MDM2 binding residues of p53 and prevents its interactions with MDM2 and MDM4. Our results point to an alternative mechanism of blocking p53 interaction with MDM2 and MDM4 and may pave the way for the development of novel allosteric inhibitors of p53/MDM2 and p53/MDM4 interactions.

Multiepitope Fusion Antigen: MEFA, an Epitope- and Structure-Based Vaccinology Platform for Multivalent Vaccine Development

Vaccines are regarded as the most cost-effective countermeasure against infectious diseases. One challenge often affecting vaccine development is antigenic diversity or pathogen heterogeneity. Different strains produce immunologically heterogeneous virulence factors, therefore an effective vaccine needs to induce broad-spectrum host immunity to provide cross-protection. Recent advances in genomics and proteomics, particularly computational biology and structural biology, establishes structural vaccinology and highlights the feasibility of developing effective and precision vaccines. Here, we introduce the epitope- and structure-based vaccinology platform multiepitope-fusion-antigen (MEFA), and provide instructions to generate polyvalent MEFA immunogens for vaccine development. Conceptually, MEFA combines epitope vaccinology and structural vaccinology to enable a protein immunogen to present heterogeneous antigenic domains (epitopes) and to induce broadly protective immunity against different virulence factors, strains or diseases. Methodologically, the MEFA platform first identifies a safe, structurally stable and strongly immunogenic backbone protein and immunodominant (ideally neutralizing or protective) epitopes from heterogeneous strains or virulence factors of interest. Then, assisted with protein modeling and molecule dynamic simulation, MEFA integrates heterogeneous epitopes into a backbone protein via epitope substitution for a polyvalent MEFA protein and mimics epitope native antigenicity. Finally, the MEFA protein is examined for broad immunogenicity in animal immunization, and assessed for potential application for multivalent vaccine development in preclinical studies.

Prediction of Protein Tertiary Structure via Regularized Template Classification Techniques

We discuss the use of the regularized linear discriminant analysis (LDA) as a model reduction technique combined with particle swarm optimization (PSO) in protein tertiary structure prediction, followed by structure refinement based on singular value decomposition (SVD) and PSO. The algorithm presented in this paper corresponds to the category of template-based modeling. The algorithm performs a preselection of protein templates before constructing a lower dimensional subspace via a regularized LDA. The protein coordinates in the reduced spaced are sampled using a highly explorative optimization algorithm, regressive–regressive PSO (RR-PSO). The obtained structure is then projected onto a reduced space via singular value decomposition and further optimized via RR-PSO to carry out a structure refinement. The final structures are similar to those predicted by best structure prediction tools, such as Rossetta and Zhang servers. The main advantage of our methodology is that alleviates the ill-posed character of protein structure prediction problems related to high dimensional optimization. It is also capable of sampling a wide range of conformational space due to the application of a regularized linear discriminant analysis, which allows us to expand the differences over a reduced basis set.

An Expanded Conformation of an Antibody Fab Region by X-Ray Scattering, Molecular Dynamics, and smFRET Identifies an Aggregation Mechanism

Protein aggregation is the underlying cause of many diseases, and also limits the usefulness of many natural and engineered proteins in biotechnology. Better mechanistic understanding and characterization of aggregation-prone states is needed to guide protein engineering, formulation, and drug-targeting strategies that prevent aggregation. While several final aggregated states-notably amyloids-have been characterized structurally, very little is known about the native structural conformers that initiate aggregation. We used a novel combination of small-angle x-ray scattering (SAXS), atomistic molecular dynamic simulations, single-molecule Förster resonance energy transfer, and aggregation-prone region predictions, to characterize structural changes in a native humanized Fab A33 antibody fragment, that correlated with the experimental aggregation kinetics. SAXS revealed increases in the native state radius of gyration, R_g, of 2.2% to 4.1%, at pH 5.5 and below, concomitant with accelerated aggregation. In a cutting-edge approach, we fitted the SAXS data to full MD simulations from the same conditions and located the conformational changes in the native state to the constant domain of the light chain (C_L). This C_L displacement was independently confirmed using single-molecule Förster resonance energy transfer measurements with two dual-labeled Fabs. These conformational changes were also found to increase the solvent exposure of a predicted APR, suggesting a likely mechanism through which they promote aggregation. Our findings provide a means by which aggregation-prone conformational states can be readily determined experimentally, and thus potentially used to guide protein engineering, or ligand binding strategies, with the aim of stabilizing the protein against aggregation.

Computational Design To Reduce Conformational Flexibility and Aggregation Rates of an Antibody Fab Fragment

Computationally guided semirational design has significant potential for improving the aggregation kinetics of protein biopharmaceuticals. While improvement in the global conformational stability can stabilize proteins to aggregation under some conditions, previous studies suggest that such an approach is limited, because thermal transition temperatures ( T_m) and the fraction of protein unfolded ( f_T) tend to only correlate with aggregation kinetics where the protein is incubated at temperatures approaching the T_m. This is because under these conditions, aggregation from globally unfolded protein becomes dominant. However, under native conditions, the aggregation kinetics are presumed to be dependent on local structural fluctuations or partial unfolding of the native state, which reveal regions of high propensity to form protein-protein interactions that lead to aggregation. In this work, we have targeted the design of stabilizing mutations to regions of the A33 Fab surface structure, which were predicted to be more flexible. This Fab already has high global stability, and global unfolding is not the main cause of aggregation under most conditions. Therefore, the aim was to reduce the conformational flexibility and entropy of the native protein at various locations and thus identify which of those regions has the greatest influence on the aggregation kinetics. Highly dynamic regions of structure were identified through both molecular dynamics simulation and B-factor analysis of related X-ray crystal structures. The most flexible residues were mutated into more stable variants, as predicted by Rosetta, which evaluates the ΔΔ G_ND for each potential point mutation. Additional destabilizing variants were prepared as controls to evaluate the prediction accuracy and also to assess the general influence of conformational stability on aggregation kinetics. The thermal conformational stability, and aggregation rates of 18 variants at 65 °C, were each examined at pH 4, 200 mM ionic strength, under which conditions the initial wild-type protein was <5% unfolded. Variants with decreased T_m values led to more rapid aggregation due to an increase in the fraction of protein unfolded under the conditions studied. As expected, no significant improvements were observed in the global conformational stability as measured by T_m. However, 6 of the 12 stable variants led to an increase in the cooperativity of unfolding, consistent with lower conformational flexibility and entropy in the native ensemble. Three of these had 5-11% lower aggregation rates, and their structural clustering indicated that the local dynamics of the C-terminus of the heavy chain had a role in influencing the aggregation rate.

The Yin and Yang of SagS: Distinct Residues in the HmsP Domain of SagS Independently Regulate Biofilm Formation and Biofilm Drug Tolerance

The formation of inherently drug-tolerant biofilms by the opportunistic pathogen Pseudomonas aeruginosa requires the sensor-regulator hybrid SagS, with ΔsagS biofilms being unstructured and exhibiting increased antimicrobial susceptibility. Recent findings indicated SagS to function as a switch to control biofilm formation and drug tolerance independently. Moreover, findings suggested the periplasmic sensory HmsP domain of SagS is likely to be the control point in the regulation of biofilm formation and biofilm cells transitioning to a drug-tolerant state. We thus asked whether specific amino acid residues present in the HmsP domain contribute to the switch function of SagS. HmsP domain residues were therefore subjected to alanine replacement mutagenesis to identify substitutions that block the sensory function(s) of SagS, which is apparent by attached cells being unable to develop mature biofilms and/or prevent transition to an antimicrobial-resistant state. Mutant analyses revealed 32 residues that only contribute to blocking one sensory function. Moreover, amino acid residues affecting attachment and subsequent biofilm formation but not biofilm tolerance also impaired histidine kinase signaling via BfiS. In contrast, residues affecting biofilm drug tolerance but not attachment and subsequent biofilm formation negatively impacted BrlR transcription factor levels. Structure prediction suggested the two sets of residues affecting sensory functions are located in distinct areas that were previously described as being involved in ligand binding interactions. Taken together, these studies identify the molecular basis for the dual regulatory function of SagS.IMPORTANCE The membrane-bound sensory protein SagS plays a pivotal role in P. aeruginosa biofilm formation and biofilm cells gaining their heightened resistance to antimicrobial agents, with SagS being the control point at which both pathways diverge. Here, we demonstrate for the first time that the two distinct pathways leading to biofilm formation and biofilm drug tolerance are under the control of two sets of amino acid residues located within the HmsP sensory domain of SagS. The respective amino acids are likely part of ligand binding interaction sites. Thus, our findings have the potential not only to enable the manipulation of SagS function but also to enable research of biofilm drug tolerance in a manner independent of biofilm formation (and vice versa). Moreover, the manipulation of SagS function represents a promising target/avenue open for biofilm control.

MEFA (multiepitope fusion antigen)-Novel Technology for Structural Vaccinology, Proof from Computational and Empirical Immunogenicity Characterization of an Enterotoxigenic Escherichia coli (ETEC) Adhesin MEFA

Vaccine development often encounters the challenge of virulence heterogeneity. Enterotoxigenic Escherichia coli (ETEC) bacteria producing immunologically heterogeneous virulence factors are a leading cause of children's diarrhea and travelers' diarrhea. Currently, we do not have licensed vaccines against ETEC bacteria. While conventional methods continue to make progress but encounter challenge, new computational and structure-based approaches are explored to accelerate ETEC vaccine development. In this study, we applied a structural vaccinology concept to construct a structure-based multiepitope fusion antigen (MEFA) to carry representing epitopes of the seven most important ETEC adhesins [CFA/I, CFA/II (CS1-CS3), CFA/IV (CS4-CS6)], simulated antigenic structure of the CFA/I/II/IV MEFA with computational atomistic modeling and simulation, characterized immunogenicity in mouse immunization, and examined the potential of structure-informed vaccine design for ETEC vaccine development. A tag-less recombinant MEFA protein (CFA/I/II/IV MEFA) was effectively expressed and extracted. Molecular dynamics simulations indicated that this MEFA immunogen maintained a stable secondary structure and presented epitopes on the protein surface. Empirical data showed that mice immunized with the tagless CFA/I/II/IV MEFA developed strong antigen-specific antibody responses, and mouse serum antibodies significantly inhibited in vitro adherence of bacteria expressing these seven adhesins. These results revealed congruence of antigen immunogenicity between computational simulation and empirical mouse immunization and indicated this tag-less CFA/I/II/IV MEFA potentially an antigen for a broadly protective ETEC vaccine, suggesting a potential application of MEFA-based structural vaccinology for vaccine design against ETEC and likely other pathogens.

Systematic prioritization of functional hotspot in RIG-1 domains using pattern based conventional molecular dynamic simulation

BACKGROUND

Retinoic acid inducible gene 1 (RIG-1), multi-domain protein has a role-play in detecting viral nucleic acids and stimulates the antiviral response. Dysfunction of this protein due to mutations makes the route vulnerable to viral diseases.

AIM

Identification of functional hotspots that maintains conformational stability in RIG-1 domains.

METHODS

In this study, we employed a systematic in silico strategy on RIG-1 protein to understand the mechanism of structural changes upon mutation. We computationally investigated the protein sequence signature for all the three domains of RIG-1 protein that encloses the mutation within the motif. Further, we carried out a structural comparison between RIG-1 domains with their respective distant orthologs which revealed the minimal number of interactions required to maintain its structural fold. This intra-protein network paved the way to infer hotspot residues crucial for the maintenance of the structural architecture and folding pattern.

KEY FINDINGS

Our analysis revealed about 40 hotspot residues that determine the folding pattern of the RIG-1 domains. Also, conventional molecular dynamic simulation coupled with essential dynamics provides conformational transitions of hot spot residues among native and mutant structures. Structural variations owing to hotspot residues in mutants again confirm the significance of these residues in structural characterization of RIG-1 domains. We believe our results will help the researchers to better comprehend towards regulatory regions and target-binding sites for therapeutic design within the pattern recognition receptor proteins.

SIGNIFICANCE

Our protocol employed in this work describes a novel approach in identifying signature residues that would provide structural insights in protein folding.

Analysis of positive and negative allosteric modulation in metabotropic glutamate receptors 4 and 5 with a dual ligand

As class C GPCRs and regulators of synaptic activity, human metabotropic glutamate receptors (mGluRs) 4 and 5 are prime targets for allosteric modulation, with mGlu5 inhibition or mGlu4 stimulation potentially treating conditions like chronic pain and Parkinson’s disease. As an allosteric modulator that can bind both receptors, 2-Methyl-6-(phenylethynyl)pyridine (MPEP) is able to negatively modulate mGlu5 or positively modulate mGlu4. At a structural level, how it elicits these responses and how mGluRs undergo activation is unclear. Here, we employ homology modelling and 30 µs of atomistic molecular dynamics (MD) simulations to probe allosteric conformational change in mGlu4 and mGlu5, with and without docked MPEP. Our results identify several structural differences between mGlu4 and mGlu5, as well as key differences responsible for MPEP-mediated positive and negative allosteric modulation, respectively. A novel mechanism of mGlu4 activation is revealed, which may apply to all mGluRs in general. This involves conformational changes in TM3, TM4 and TM5, separation of intracellular loop 2 (ICL2) from ICL1/ICL3, and destabilization of the ionic-lock. On the other hand, mGlu5 experiences little disturbance when MPEP binds, maintaining its inactive state with reduced conformational fluctuation. In addition, when MPEP is absent, a lipid molecule can enter the mGlu5 allosteric pocket.

Biochemical Characterization of APOBEC3H Variants: Implications for Their HIV-1 Restriction Activity and mC Modification

APOBEC3H (A3H) is the most polymorphic member of the APOBEC3 family. Seven haplotypes (hap I-VII) and four mRNA splicing variants (SV) of A3H have been identified. The various haplotypes differ in anti-HIV activity, which is attributed to differences in protein stability, subcellular distribution, and/or RNA binding and virion packaging. Here, we report the first comparative biochemical studies of all the A3H variants using highly purified proteins. We show that all haplotypes were stably expressed and could be purified to homogeneity by Escherichia coli expression. Surprisingly, four out of the seven haplotypes showed high cytosine (C) deaminase activity, with hap V displaying extremely high activity that was comparable to the highly active A3A. Furthermore, all four haplotypes that were active in C deamination were also highly active on methylated C (mC), with hap II displaying almost equal deamination efficiency on both. The deamination activity of these A3H variants correlates well with their reported anti-HIV activity for the different haplotypes, suggesting that deaminase activity may be an important factor in determining their respective anti-HIV activities. Moreover, mC deamination of A3H displayed a strong preference for the sequence motif of T-mCpG-C/G, which may suggest a potential role in genomic mC modification at the characteristic "CpG" island motif.

SK. Structural exploration of acid sphingomyelinase at different physiological pH through molecular dynamics and docking studies

Acid sphingomyelinase (ASM) hydrolysis the sphingomyelin at physiological pH 5.0 and subsequently leads to ceramide production.

Clustering and percolation in protein loop structures

Background

High precision protein loop modelling remains a challenge, both in template based and template independent approaches to protein structure prediction.

Method

We introduce the concepts of protein loop clustering and percolation, to develop a quantitative approach to systematically classify the modular building blocks of loops in crystallographic folded proteins. These fragments are all different parameterisations of a unique kink solution to a generalised discrete nonlinear Schrödinger (DNLS) equation. Accordingly, the fragments are also local energy minima of the ensuing energy function.

Results

We show how the loop fragments cover practically all ultrahigh resolution crystallographic protein structures in Protein Data Bank (PDB), with a 0.2 Ångström root-mean-square (RMS) precision. We find that no more than 12 different loop fragments are needed, to describe around 38 % of ultrahigh resolution loops in PDB. But there is also a large number of loop fragments that are either unique, or very rare, and examples of unique fragments are found even in the structure of a myoglobin.

Conclusions

Protein loops are built in a modular fashion. The loops are composed of fragments that can be modelled by the kink of the DNLS equation. The majority of loop fragments are also common, which are shared by many proteins. These common fragments are probably important for supporting the overall protein conformation. But there are also several fragments that are either unique to a given protein, or very rare. Such fragments are probably related to the function of the protein. Furthermore, we have found that the amino acid sequence does not determine the structure in a unique fashion. There are many examples of loop fragments with an identical amino acid sequence, but with a very different structure.

Electronic supplementary material

The online version of this article (doi:10.1186/s12900-015-0049-x) contains supplementary material, which is available to authorized users.

Characterization of an Additional Splice Acceptor Site Introduced into CYP4B1 in Hominoidae during Evolution

CYP4B1 belongs to the cytochrome P450 family 4, one of the oldest P450 families whose members have been highly conserved throughout evolution. The CYP4 monooxygenases typically oxidize fatty acids to both inactive and active lipid mediators, although the endogenous ligand(s) is largely unknown. During evolution, at the transition of great apes to humanoids, the CYP4B1 protein acquired a serine instead of a proline at the canonical position 427 in the meander region. Although this alteration impairs P450 function related to the processing of naturally occurring lung toxins, a study in transgenic mice suggested that an additional serine insertion at position 207 in human CYP4B1 can rescue the enzyme stability and activity. Here, we report that the genomic insertion of a CAG triplet at the intron 5–exon 6 boundary in human CYP4B1 introduced an additional splice acceptor site in frame. During evolution, this change occurred presumably at the stage of Hominoidae and leads to two major isoforms of the CYP4B1 enzymes of humans and great apes, either with or without a serine 207 insertion (insSer207). We further demonstrated that the CYP4B1 enzyme with insSer207 is the dominant isoform (76%) in humans. Importantly, this amino acid insertion did not affect the 4-ipomeanol metabolizing activities or stabilities of the native rabbit or human CYP4B1 enzymes, when introduced as transgenes in human primary cells and cell lines. In our 3D modeling, this functional neutrality of insSer207 is compatible with its predicted location on the exterior surface of CYP4B1 in a flexible side chain. Therefore, the Ser207 insertion does not rescue the P450 functional activity of human CYP4B1 that has been lost during evolution.

Factor VIII Interacts with the Endocytic Receptor Low-density Lipoprotein Receptor-related Protein 1 via an Extended Surface Comprising "Hot-Spot" Lysine Residues

Lysine residues are implicated in driving the ligand binding to the LDL receptor family. However, it has remained unclear how specificity is regulated. Using coagulation factor VIII as a model ligand, we now study the contribution of individual lysine residues in the interaction with the largest member of the LDL receptor family, low-density lipoprotein receptor-related protein (LRP1). Using hydrogen-deuterium exchange mass spectrometry (HDX-MS) and SPR interaction analysis on a library of lysine replacement variants as two independent approaches, we demonstrate that the interaction between factor VIII (FVIII) and LRP1 occurs over an extended surface containing multiple lysine residues. None of the individual lysine residues account completely for LRP1 binding, suggesting an additive binding model. Together with structural docking studies, our data suggest that FVIII interacts with LRP1 via an extended surface of multiple lysine residues that starts at the bottom of the C1 domain and winds around the FVIII molecule.

Identification of amino acid determinants in CYP4B1 for optimal catalytic processing of 4-ipomeanol

Mammalian CYP4B1 enzymes are cytochrome P450 mono-oxygenases that are responsible for the bioactivation of several exogenous pro-toxins including 4-ipomeanol (4-IPO). In contrast with the orthologous rabbit enzyme, we show here that native human CYP4B1 with a serine residue at position 427 is unable to bioactivate 4-IPO and does not cause cytotoxicity in HepG2 cells and primary human T-cells that overexpress these enzymes. We also demonstrate that a proline residue in the meander region at position 427 in human CYP4B1 and 422 in rabbit CYP4B1 is important for protein stability and rescues the 4-IPO bioactivation of the human enzyme, but is not essential for the catalytic activity of the rabbit CYP4B1 protein. Systematic substitution of native and p.S427P human CYP4B1 with peptide regions from the highly active rabbit enzyme reveals that 18 amino acids in the wild-type rabbit CYP4B1 protein are key for conferring high 4-IPO metabolizing activity. Introduction of 12 of the 18 amino acids that are also present at corresponding positions in other human CYP4 family members into the p.S427P human CYP4B1 protein results in a mutant human enzyme (P+12) that is as stable and as active as the rabbit wild-type CYP4B1 protein. These 12 mutations cluster in the predicted B-C loop through F-helix regions and reveal new amino acid regions important to P450 enzyme stability. Finally, by minimally re-engineering the human CYP4B1 enzyme for efficient activation of 4-IPO, we have developed a novel human suicide gene system that is a candidate for adoptive cellular therapies in humans.

Efficient representation of uncertainty in multiple sequence alignments using directed acyclic graphs

BACKGROUND

A standard procedure in many areas of bioinformatics is to use a single multiple sequence alignment (MSA) as the basis for various types of analysis. However, downstream results may be highly sensitive to the alignment used, and neglecting the uncertainty in the alignment can lead to significant bias in the resulting inference. In recent years, a number of approaches have been developed for probabilistic sampling of alignments, rather than simply generating a single optimum. However, this type of probabilistic information is currently not widely used in the context of downstream inference, since most existing algorithms are set up to make use of a single alignment.

RESULTS

In this work we present a framework for representing a set of sampled alignments as a directed acyclic graph (DAG) whose nodes are alignment columns; each path through this DAG then represents a valid alignment. Since the probabilities of individual columns can be estimated from empirical frequencies, this approach enables sample-based estimation of posterior alignment probabilities. Moreover, due to conditional independencies between columns, the graph structure encodes a much larger set of alignments than the original set of sampled MSAs, such that the effective sample size is greatly increased.

CONCLUSIONS

The alignment DAG provides a natural way to represent a distribution in the space of MSAs, and allows for existing algorithms to be efficiently scaled up to operate on large sets of alignments. As an example, we show how this can be used to compute marginal probabilities for tree topologies, averaging over a very large number of MSAs. This framework can also be used to generate a statistically meaningful summary alignment; example applications show that this summary alignment is consistently more accurate than the majority of the alignment samples, leading to improvements in downstream tree inference. Implementations of the methods described in this article are available at http://statalign.github.io/WeaveAlign .

The structural basis of substrate recognition by the eukaryotic chaperonin TRiC/CCT

The eukaryotic chaperonin TRiC (also called CCT) is the obligate chaperone for many essential proteins. TRiC is hetero-oligomeric, comprising two stacked rings of eight different subunits each. Subunit diversification from simpler archaeal chaperonins appears linked to proteome expansion. Here, we integrate structural, biophysical, and modeling approaches to identify the hitherto unknown substrate-binding site in TRiC and uncover the basis of substrate recognition. NMR and modeling provided a structural model of a chaperonin-substrate complex. Mutagenesis and crosslinking-mass spectrometry validated the identified substrate-binding interface and demonstrate that TRiC contacts full-length substrates combinatorially in a subunit-specific manner. The binding site of each subunit has a distinct, evolutionarily conserved pattern of polar and hydrophobic residues specifying recognition of discrete substrate motifs. The combinatorial recognition of polypeptides broadens the specificity of TRiC and may direct the topology of bound polypeptides along a productive folding trajectory, contributing to TRiC's unique ability to fold obligate substrates.

Reliable cell cycle commitment in budding yeast is ensured by signal integration

Cell fate decisions are critical for life, yet little is known about how their reliability is achieved when signals are noisy and fluctuating with time. In this study, we show that in budding yeast, the decision of cell cycle commitment (Start) is determined by the time integration of its triggering signal Cln3. We further identify the Start repressor, Whi5, as the integrator. The instantaneous kinase activity of Cln3-Cdk1 is recorded over time on the phosphorylated Whi5, and the decision is made only when phosphorylated Whi5 reaches a threshold. Cells adjust the threshold by modulating Whi5 concentration in different nutrient conditions to coordinate growth and division. Our work shows that the strategy of signal integration, which was previously found in decision-making behaviors of animals, is adopted at the cellular level to reduce noise and minimize uncertainty.

DOI:http://dx.doi.org/10.7554/eLife.03977.001

Budding yeast and other single-celled organisms can reproduce by dividing to produce two daughter cells. The timing of the cell division is critical because if the cell is still small when it divides, the resulting daughter cells may not be big enough to survive.

In budding yeast, the irreversible decision to divide—known as the ‘Start’ checkpoint—is only made once a cell reaches a certain size and is triggered by a protein called Cln3. This protein controls the activity of another protein called Whi5, which normally prevents the cell from dividing by switching off particular genes. Cln3 adds phosphate groups to Whi5 to make ‘phosphorylated Whi5’, which allows the genes involved in cell division to be switched on.

It is commonly believed that the level of Cln3 reflects the size of the cell and the nutrient conditions. Therefore, one model of cell division proposes that the cell passes the Start checkpoint when the level of Cln3 reaches a threshold value. However, levels of the Cln3 protein in cells can naturally fluctuate, and computer simulations based on this model showed that this would not produce reliable decisions on when to divide. So how do cells manage to distinguish noise from the genuine signals that indicate it is the right time to divide?

To address this question, Liu et al. studied yeast cells containing an artificial version of the gene encoding the Cln3 protein whose levels could be adjusted by adding a particular chemical. This revealed that cells with higher levels of Cln3 passed through the Start checkpoint sooner than cells that had lower levels of Cln3.

The observation suggests that cells add up the amount of Cln3 present over a period of time to see if this reaches the threshold needed for the Start checkpoint. This could be possible if, instead of sensing Cln3 levels directly, the cell senses the accumulation of phosphorylated Whi5. To test this idea, Liu et al. carried out additional experiments and found that the decision to pass the Start checkpoint only occurs when the amount of phosphorylated Whi5 reaches a certain threshold.

The cells are able to coordinate their growth and division under different nutrient conditions by altering the threshold of phosphorylated Whi5. When the nutrient supply is poor, more phosphorylated Whi5 needs to be accumulated to allow the cell to pass the Start checkpoint. In this way, cells adjust when they divide according to nutrient conditions. Similar strategies may be found in other signaling or decision-making systems.

DOI:http://dx.doi.org/10.7554/eLife.03977.002

Homology modeling, molecular dynamics, and site-directed mutagenesis study of AlkB human homolog 1 (ALKBH1)

The ability to repair DNA is important for the conservation of genetic information of living organisms. Cells have a number of ways to restore damaged DNA, such as direct DNA repair, base excision repair, and nucleotide excision repair. One of the proteins that can perform direct repair of DNA bases is Escherichia coli AlkB. In humans, there are 9 identified AlkB homologs, including AlkB homolog 1 (ALKBH1). Many of these proteins catalyze the direct oxidative dealkylation of DNA and RNA bases and, as such, have an important role in repairing DNA from damage induced by alkylating agents. In addition to the dealkylase activity, ALKBH1 can also function as an apyrimidinic/apurinic lyase and was proposed to have a distinct lyase active site. To our knowledge, no crystal structure or complete homology model of ALKBH1 protein is available. In this study, we have used homology modeling to predict the structure of ALKBH1 based on AlkB and Duffy-binding-like domain crystal structures as templates. Molecular dynamics simulations were subsequently performed on the predicted structure of ALKBH1. The positions of two disulfide bonds or a zinc-finger motif and a disulfide bond were predicted and the importance of these features was tested by mutagenesis. Possible locations for the lyase active site are proposed based on the analysis of our predicted structures and previous experimental results.

Functional interaction with filamin A and intracellular Ca2+ enhance the surface membrane expression of a small-conductance Ca2+-activated K+ (SK2) channel

For an excitable cell to function properly, a precise number of ion channel proteins need to be trafficked to distinct locations on the cell surface membrane, through a network and anchoring activity of cytoskeletal proteins. Not surprisingly, mutations in anchoring proteins have profound effects on membrane excitability. Ca(2+)-activated K(+) channels (KCa2 or SK) have been shown to play critical roles in shaping the cardiac atrial action potential profile. Here, we demonstrate that filamin A, a cytoskeletal protein, augments the trafficking of SK2 channels in cardiac myocytes. The trafficking of SK2 channel is Ca(2+)-dependent. Further, the Ca(2+) dependence relies on another channel-interacting protein, α-actinin2, revealing a tight, yet intriguing, assembly of cytoskeletal proteins that orchestrate membrane expression of SK2 channels in cardiac myocytes. We assert that changes in SK channel trafficking would significantly alter atrial action potential and consequently atrial excitability. Identification of therapeutic targets to manipulate the subcellular localization of SK channels is likely to be clinically efficacious. The findings here may transcend the area of SK2 channel studies and may have implications not only in cardiac myocytes but in other types of excitable cells.

Crystal structure of the nucleotide-binding domain of mortalin, the mitochondrial Hsp70 chaperone

Mortalin, a member of the Hsp70-family of molecular chaperones, functions in a variety of processes including mitochondrial protein import and quality control, Fe-S cluster protein biogenesis, mitochondrial homeostasis, and regulation of p53. Mortalin is implicated in regulation of apoptosis, cell stress response, neurodegeneration, and cancer and is a target of the antitumor compound MKT-077. Like other Hsp70-family members, Mortalin consists of a nucleotide-binding domain (NBD) and a substrate-binding domain. We determined the crystal structure of the NBD of human Mortalin at 2.8 Å resolution. Although the Mortalin nucleotide-binding pocket is highly conserved relative to other Hsp70 family members, we find that its nucleotide affinity is weaker than that of Hsc70. A Parkinson's disease-associated mutation is located on the Mortalin-NBD surface and may contribute to Mortalin aggregation. We present structure-based models for how the Mortalin-NBD may interact with the nucleotide exchange factor GrpEL1, with p53, and with MKT-077. Our structure may contribute to the understanding of disease-associated Mortalin mutations and to improved Mortalin-targeting antitumor compounds.

The PS1 hairpin of Mcm3 is essential for viability and for DNA unwinding in vitro

The pre-sensor 1 (PS1) hairpin is found in ring-shaped helicases of the AAA+ family (ATPases associated with a variety of cellular activities) of proteins and is implicated in DNA translocation during DNA unwinding of archaeal mini-chromosome maintenance (MCM) and superfamily 3 viral replicative helicases. To determine whether the PS1 hairpin is required for the function of the eukaryotic replicative helicase, Mcm2-7 (also comprised of AAA+ proteins), we mutated the conserved lysine residue in the putative PS1 hairpin motif in each of the Saccharomyces cerevisiae Mcm2-7 subunits to alanine. Interestingly, only the PS1 hairpin of Mcm3 was essential for viability. While mutation of the PS1 hairpin in the remaining MCM subunits resulted in minimal phenotypes, with the exception of Mcm7 which showed slow growth under all conditions examined, the viable alleles were synthetic lethal with each other. Reconstituted Mcm2-7 containing Mcm3 with the PS1 mutation (Mcm3(K499A)) had severely decreased helicase activity. The lack of helicase activity provides a probable explanation for the inviability of the mcm3(K499A) strain. The ATPase activity of Mcm2-7(3K499A) was similar to the wild type complex, but its interaction with single-stranded DNA in an electrophoretic mobility shift assay and its associations in cells were subtly altered. Together, these findings indicate that the PS1 hairpins in the Mcm2-7 subunits have important and distinct functions, most evident by the essential nature of the Mcm3 PS1 hairpin in DNA unwinding.

Cold adaptation, ca2+ dependency and autolytic stability are related features in a highly active cold-adapted trypsin resistant to autoproteolysis engineered for biotechnological applications

Pig trypsin is routinely used as a biotechnological tool, due to its high specificity and ability to be stored as an inactive stable zymogen. However, it is not an optimum enzyme for conditions found in wound debriding for medical uses and trypsinization processes for protein analysis and animal cell culturing, where low Ca(2+) dependency, high activity in mild conditions and easy inactivation are crucial. We isolated and thermodynamically characterized a highly active cold-adapted trypsin for medical and laboratory use that is four times more active than pig trypsin at 10(°) C and at least 50% more active than pig trypsin up to 50(°) C. Contrary to pig trypsin, this enzyme has a broad optimum pH between 7 and 10 and is very insensitive to Ca(2+) concentration. The enzyme is only distantly related to previously described cryophilic trypsins. We built and studied molecular structure models of this trypsin and performed molecular dynamic calculations. Key residues and structures associated with calcium dependency and cryophilicity were identified. Experiments indicated that the protein is unstable and susceptible to autoproteolysis. Correlating experimental results and structural predictions, we designed mutations to improve the resistance to autoproteolysis and conserve activity for longer periods after activation. One single mutation provided around 25 times more proteolytic stability. Due to its cryophilic nature, this trypsin is easily inactivated by mild denaturation conditions, which is ideal for controlled proteolysis processes without requiring inhibitors or dilution. We clearly show that cold adaptation, Ca(2+) dependency and autolytic stability in trypsins are related phenomena that are linked to shared structural features and evolve in a concerted fashion. Hence, both structurally and evolutionarily they cannot be interpreted and studied separately as previously done.

Accurate structure prediction of peptide-MHC complexes for identifying highly immunogenic antigens

Designing an optimal HIV-1 vaccine faces the challenge of identifying antigens that induce a broad immune capacity. One factor to control the breadth of T cell responses is the surface morphology of a peptide-MHC complex. Here, we present an in silico protocol for predicting peptide-MHC structure. A robust signature of a conformational transition was identified during all-atom molecular dynamics, which results in a model with high accuracy. A large test set was used in constructing our protocol and we went another step further using a blind test with a wild-type peptide and two highly immunogenic mutants, which predicted substantial conformational changes in both mutants. The center residues at position five of the analogs were configured to be accessible to solvent, forming a prominent surface, while the residue of the wild-type peptide was to point laterally toward the side of the binding cleft. We then experimentally determined the structures of the blind test set, using high resolution of X-ray crystallography, which verified predicted conformational changes. Our observation strongly supports a positive association of the surface morphology of a peptide-MHC complex to its immunogenicity. Our study offers the prospect of enhancing immunogenicity of vaccines by identifying MHC binding immunogens.

Quantifying variances in comparative RNA secondary structure prediction

Background

With the advancement of next-generation sequencing and transcriptomics technologies, regulatory effects involving RNA, in particular RNA structural changes are being detected. These results often rely on RNA secondary structure predictions. However, current approaches to RNA secondary structure modelling produce predictions with a high variance in predictive accuracy, and we have little quantifiable knowledge about the reasons for these variances.

Results

In this paper we explore a number of factors which can contribute to poor RNA secondary structure prediction quality. We establish a quantified relationship between alignment quality and loss of accuracy. Furthermore, we define two new measures to quantify uncertainty in alignment-based structure predictions. One of the measures improves on the “reliability score” reported by PPfold, and considers alignment uncertainty as well as base-pair probabilities. The other measure considers the information entropy for SCFGs over a space of input alignments.

Conclusions

Our predictive accuracy improves on the PPfold reliability score. We can successfully characterize many of the underlying reasons for and variances in poor prediction. However, there is still variability unaccounted for, which we therefore suggest comes from the RNA secondary structure predictive model itself.

The Protein Model Portal--a comprehensive resource for protein structure and model information

The Protein Model Portal (PMP) has been developed to foster effective use of 3D molecular models in biomedical research by providing convenient and comprehensive access to structural information for proteins. Both experimental structures and theoretical models for a given protein can be searched simultaneously and analyzed for structural variability. By providing a comprehensive view on structural information, PMP offers the opportunity to apply consistent assessment and validation criteria to the complete set of structural models available for proteins. PMP is an open project so that new methods developed by the community can contribute to PMP, for example, new modeling servers for creating homology models and model quality estimation servers for model validation. The accuracy of participating modeling servers is continuously evaluated by the Continuous Automated Model EvaluatiOn (CAMEO) project. The PMP offers a unique interface to visualize structural coverage of a protein combining both theoretical models and experimental structures, allowing straightforward assessment of the model quality and hence their utility. The portal is updated regularly and actively developed to include latest methods in the field of computational structural biology.

Database URL:http://www.proteinmodelportal.org

Convergence of IRBIT, phosphatidylinositol (4,5) bisphosphate, and WNK/SPAK kinases in regulation of the Na+-HCO3- cotransporters family

Fluid and HCO3(-) secretion is a vital function of secretory epithelia, involving basolateral HCO3(-) entry through the Na(+)-HCO3(-) cotransporter (NBC) NBCe1-B, and luminal HCO3(-) exit mediated by cystic fibrosis transmembrane conductance regulator (CFTR) and solute carrier family 26 (SLC26) Cl(-)/HCO3(-) exchangers. HCO3(-) secretion is highly regulated, with the WNK/SPAK kinase pathway setting the resting state and the IRBIT/PP1 pathway setting the stimulated state. However, we know little about the relationships between the WNK/SPAK and IRBIT/PP1 sites in the regulation of the transporters. The first 85 N-terminal amino acids of NBCe1-B function as an autoinhibitory domain. Here we have identified a positively charged module within NBCe1-B(37-65) that is conserved in NBCn1-A and all 20 members of the NBC superfamily except NBCe1-A. This module is required for the interaction and activation of NBCe1-B and NBCn1-A by IRBIT and their regulation by phosphatidylinositol 4,5-bisphosphate (PIP2). Activation of the transporters by IRBIT and PIP2 is nonadditive but complementary. Phosphorylation of Ser65 mediates regulation of NBCe1-B by SPAK, and phosphorylation of Thr49 is required for regulation by IRBIT and SPAK. Sequence searches using the NBCe1-B regulatory module as a template identified a homologous sequence in the CFTR R domain and Slc26a6 sulfat transporter and antisigma factor antagonist (STAS) domain. Accordingly, the R and STAS domains bind IRBIT, and the R domain is required for activation of CFTR by IRBIT. These findings reveal convergence of regulatory modalities in a conserved domain of the NBC that may be present in other HCO3(-) transporters and thus in the regulation of epithelial fluid and HCO3(-) secretion.

Structural basis for dsRNA recognition, filament formation, and antiviral signal activation by MDA5

MDA5, a viral double-stranded RNA (dsRNA) receptor, shares sequence similarity and signaling pathways with RIG-I yet plays essential functions in antiviral immunity through distinct specificity for viral RNA. Revealing the molecular basis for the functional divergence, we report here the crystal structure of MDA5 bound to dsRNA, which shows how, using the same domain architecture, MDA5 recognizes the internal duplex structure, whereas RIG-I recognizes the terminus of dsRNA. We further show that MDA5 uses direct protein-protein contacts to stack along dsRNA in a head-to-tail arrangement, and that the signaling domain (tandem CARD), which decorates the outside of the core MDA5 filament, also has an intrinsic propensity to oligomerize into an elongated structure that activates the signaling adaptor, MAVS. These data support a model in which MDA5 uses long dsRNA as a signaling platform to cooperatively assemble the core filament, which in turn promotes stochastic assembly of the tandem CARD oligomers for signaling.

Defining and predicting structurally conserved regions in protein superfamilies

MOTIVATION

The structures of homologous proteins are generally better conserved than their sequences. This phenomenon is demonstrated by the prevalence of structurally conserved regions (SCRs) even in highly divergent protein families. Defining SCRs requires the comparison of two or more homologous structures and is affected by their availability and divergence, and our ability to deduce structurally equivalent positions among them. In the absence of multiple homologous structures, it is necessary to predict SCRs of a protein using information from only a set of homologous sequences and (if available) a single structure. Accurate SCR predictions can benefit homology modelling and sequence alignment.

RESULTS

Using pairwise DaliLite alignments among a set of homologous structures, we devised a simple measure of structural conservation, termed structural conservation index (SCI). SCI was used to distinguish SCRs from non-SCRs. A database of SCRs was compiled from 386 SCOP superfamilies containing 6489 protein domains. Artificial neural networks were then trained to predict SCRs with various features deduced from a single structure and homologous sequences. Assessment of the predictions via a 5-fold cross-validation method revealed that predictions based on features derived from a single structure perform similarly to ones based on homologous sequences, while combining sequence and structural features was optimal in terms of accuracy (0.755) and Matthews correlation coefficient (0.476). These results suggest that even without information from multiple structures, it is still possible to effectively predict SCRs for a protein. Finally, inspection of the structures with the worst predictions pinpoints difficulties in SCR definitions.

AVAILABILITY

The SCR database and the prediction server can be found at http://prodata.swmed.edu/SCR.

CONTACT

91huangi@gmail.com or grishin@chop.swmed.edu

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics Online.

A gradient of ATP affinities generates an asymmetric power stroke driving the chaperonin TRIC/CCT folding cycle

The eukaryotic chaperonin TRiC/CCT uses ATP cycling to fold many essential proteins that other chaperones cannot fold. This 1 MDa hetero-oligomer consists of two identical stacked rings assembled from eight paralogous subunits, each containing a conserved ATP-binding domain. Here, we report a dramatic asymmetry in the ATP utilization cycle of this ring-shaped chaperonin, despite its apparently symmetric architecture. Only four of the eight different subunits bind ATP at physiological concentrations. ATP binding and hydrolysis by the low-affinity subunits is fully dispensable for TRiC function in vivo. The conserved nucleotide-binding hierarchy among TRiC subunits is evolutionarily modulated through differential nucleoside contacts. Strikingly, high- and low-affinity subunits are spatially segregated within two contiguous hemispheres in the ring, generating an asymmetric power stroke that drives the folding cycle. This unusual mode of ATP utilization likely serves to orchestrate a directional mechanism underlying TRiC/CCT's unique ability to fold complex eukaryotic proteins.

Modeling of open, closed, and open-inactivated states of the hERG1 channel: structural mechanisms of the state-dependent drug binding

The human ether-a-go-go related gene 1 (hERG1) K ion channel is a key element for the rapid component of the delayed rectified potassium current in cardiac myocytes. Since there are no crystal structures for hERG channels, creation and validation of its reliable atomistic models have been key targets in molecular cardiology for the past decade. In this study, we developed and vigorously validated models for open, closed, and open-inactivated states of hERG1 using a multistep protocol. The conserved elements were derived using multiple-template homology modeling utilizing available structures for Kv1.2, Kv1.2/2.1 chimera, and KcsA channels. Then missing elements were modeled with the ROSETTA De Novo protein-designing suite and further refined with all-atom molecular dynamics simulations. The final ensemble of models was evaluated for consistency to the reported experimental data from biochemical, biophysical, and electrophysiological studies. The closed state models were cross-validated against available experimental data on toxin footprinting with protein-protein docking using hERG state-selective toxin BeKm-1. Poisson-Boltzmann calculations were performed to determine gating charge and compare it to electrophysiological measurements. The validated structures offered us a unique chance to assess molecular mechanisms of state-dependent drug binding in three different states of the channel.

Defining the boundaries of species specificity for the Saccharomyces cerevisiae glycosylphosphatidylinositol transamidase using a quantitative in vivo assay

In eukaryotes, GPI (glycosylphosphatidylinositol) lipid anchoring of proteins is an abundant post-translational modification. The attachment of the GPI anchor is mediated by GPI-T (GPI transamidase), a multimeric, membrane-bound enzyme located in the ER (endoplasmic reticulum). Upon modification, GPI-anchored proteins enter the secretory pathway and ultimately become tethered to the cell surface by association with the plasma membrane and, in yeast, by covalent attachment to the outer glucan layer. This work demonstrates a novel in vivo assay for GPI-T. Saccharomyces cerevisiae INV (invertase), a soluble secreted protein, was converted into a substrate for GPI-T by appending the C-terminal 21 amino acid GPI-T signal sequence from the S. cerevisiae Yapsin 2 [Mkc7p (Y21)] on to the C-terminus of INV. Using a colorimetric assay and biochemical partitioning, extracellular presentation of GPI-anchored INV was shown. Two human GPI-T signal sequences were also tested and each showed diminished extracellular INV activity, consistent with lower levels of GPI anchoring and species specificity. Human/fungal chimaeric signal sequences identified a small region of five amino acids that was predominantly responsible for this species specificity.

An all-atom model of the structure of human copper transporter 1

Human copper transporter 1 (hCTR1) is the major high affinity copper influx transporter in mammalian cells that also mediates uptake of the cancer chemotherapeutic agent cisplatin. A low resolution structure of hCTR1 determined by cryoelectron microscopy was recently published. Several protein structure simulation techniques were used to create an all-atom model of this important transporter using the low resolution structure as a starting point. The all-atom model provides new insights into the roles of specific residues of the N-terminal extracellular domain, the intracellular loop, and C-terminal region in metal ion transport. In particular, the model demonstrates that the central region of the pore contains four sets of methionine triads in the intramembranous region. The structure confirms that two triads of methionine residues delineate the intramembranous region of the transporter, and further identifies two additional methionine triads that are located in the extracellular N-terminal part of the transporter. Together, the four triads create a structure that promotes stepwise transport of metal ions into and then through the intramembranous channel of the transporter via transient thioether bonds to methionine residues. Putative copper-binding sites in the hCTR1 trimer were identified by a program developed by us for prediction of metal-binding sites. These sites correspond well with the known effects of mutations on the ability of the protein to transport copper and cisplatin.

Tracing determinants of dual substrate specificity in glycoside hydrolase family 5

Enzymes are traditionally viewed as having exquisite substrate specificity; however, recent evidence supports the notion that many enzymes have evolved activities against a range of substrates. The diversity of activities across glycoside hydrolase family 5 (GH5) suggests that this family of enzymes may contain numerous members with activities on multiple substrates. In this study, we combined structure- and sequence-based phylogenetic analysis with biochemical characterization to survey the prevalence of dual specificity for glucan- and mannan-based substrates in the GH5 family. Examination of amino acid profile differences between the subfamilies led to the identification and subsequent experimental confirmation of an active site motif indicative of dual specificity. The motif enabled us to successfully discover several new dually specific members of GH5, and this pattern is present in over 70 other enzymes, strongly suggesting that dual endoglucanase-mannanase activity is widespread in this family. In addition, reinstatement of the conserved motif in a wild type member of GH5 enhanced its catalytic efficiency on glucan and mannan substrates by 175 and 1,600%, respectively. Phylogenetic examination of other GH families further indicates that the prevalence of enzyme multispecificity in GHs may be greater than has been experimentally characterized. Single domain multispecific GHs may be exploited for developing improved enzyme cocktails or facile engineering of microbial hosts for consolidated bioprocessing of lignocellulose.

Template-based protein structure modeling using TASSER(VMT.)

Template-based protein structure modeling is commonly used for protein structure prediction. Based on the observation that multiple template-based methods often perform better than single template-based methods, we further explore the use of a variable number of multiple templates for a given target in the latest variant of TASSER, TASSER(VMT) . We first develop an algorithm that improves the target-template alignment for a given template. The improved alignment, called the SP(3) alternative alignment, is generated by a parametric alignment method coupled with short TASSER refinement on models selected using knowledge-based scores. The refined top model is then structurally aligned to the template to produce the SP(3) alternative alignment. Templates identified using SP(3) threading are combined with the SP(3) alternative and HHEARCH alignments to provide target alignments to each template. These template models are then grouped into sets containing a variable number of template/alignment combinations. For each set, we run short TASSER simulations to build full-length models. Then, the models from all sets of templates are pooled, and the top 20-50 models selected using FTCOM ranking method. These models are then subjected to a single longer TASSER refinement run for final prediction. We benchmarked our method by comparison with our previously developed approach, pro-sp(3) -TASSER, on a set with 874 easy and 318 hard targets. The average GDT-TS score improvements for the first model are 3.5 and 4.3% for easy and hard targets, respectively. When tested on the 112 CASP9 targets, our method improves the average GDT-TS scores as compared to pro-sp3-TASSER by 8.2 and 9.3% for the 80 easy and 32 hard targets, respectively. It also shows slightly better results than the top ranked CASP9 Zhang-Server, QUARK and HHpredA methods. The program is available for download at http://cssb.biology.gatech.edu/.

Using structure to explore the sequence alignment space of remote homologs

Protein structure modeling by homology requires an accurate sequence alignment between the query protein and its structural template. However, sequence alignment methods based on dynamic programming (DP) are typically unable to generate accurate alignments for remote sequence homologs, thus limiting the applicability of modeling methods. A central problem is that the alignment that is “optimal” in terms of the DP score does not necessarily correspond to the alignment that produces the most accurate structural model. That is, the correct alignment based on structural superposition will generally have a lower score than the optimal alignment obtained from sequence. Variations of the DP algorithm have been developed that generate alternative alignments that are “suboptimal” in terms of the DP score, but these still encounter difficulties in detecting the correct structural alignment. We present here a new alternative sequence alignment method that relies heavily on the structure of the template. By initially aligning the query sequence to individual fragments in secondary structure elements and combining high-scoring fragments that pass basic tests for “modelability”, we can generate accurate alignments within a small ensemble. Our results suggest that the set of sequences that can currently be modeled by homology can be greatly extended.

It has been suggested that, for nearly every protein sequence, there is already a protein with a similar structure in current protein structure databases. However, with poor or undetectable sequence relationships, it is expected that accurate alignments and models cannot be generated. Here we show that this is not the case, and that whenever structural relationship exists, there are usually local sequence relationships that can be used to generate an accurate alignment, no matter what the global sequence identity. However, this requires an alternative to the traditional dynamic programming algorithm and the consideration of a small ensemble of alignments. We present an algorithm, S4, and demonstrate that it is capable of generating accurate alignments in nearly all cases where a structural relationship exists between two proteins. Our results thus constitute an important advance in the full exploitation of the information in structural databases. That is, the expectation of an accurate alignment suggests that a meaningful model can be generated for nearly every sequence for which a suitable template exists.

Finding homes for orphan cytochrome P450s: CYP4V2 and CYP4F22 in disease states

The cytochrome P450 (CYP) 4 family of enzymes contains several recently identified membersthat are referred to as “orphan P450s” because their endogenous substrates are unknown.Human CYP4V2 and CYP4F22 are two such orphan P450s that are strongly linked to ocular andskin disease, respectively. Genetic analyses have identified a wide spectrum of mutations in the CYP4V2gene from patients suffering from Bietti’s crystalline corneoretinal dystrophy, and mutations in theCYP4F22 gene have been linked to lamellar ichthyosis. The strong gene–disease associations provideunique opportunities for elucidating the substrate specificity of these orphan P450s and unraveling thebiochemical pathways that may be impacted in patients with CYP4V2 and CYP4F22 functional deficits.

TASSER_low-zsc: an approach to improve structure prediction using low z-score-ranked templates

In a variety of threading methods, often poorly ranked (low z-score) templates have good alignments. Here, a new method, TASSER_low-zsc that identifies these low z-score-ranked templates to improve protein structure prediction accuracy, is described. The approach consists of clustering of threading templates by affinity propagation on the basis of structural similarity (thread_cluster) followed by TASSER modeling, with final models selected by using a TASSER_QA variant. To establish the generality of the approach, templates provided by two threading methods, SP(3) and SPARKS(2), are examined. The SP(3) and SPARKS(2) benchmark datasets consist of 351 and 357 medium/hard proteins (those with moderate to poor quality templates and/or alignments) of length < or =250 residues, respectively. For SP(3) medium and hard targets, using thread_cluster, the TM-scores of the best template improve by approximately 4 and 9% over the original set (without low z-score templates) respectively; after TASSER modeling/refinement and ranking, the best model improves by approximately 7 and 9% over the best model generated with the original template set. Moreover, TASSER_low-zsc generates 22% (43%) more foldable medium (hard) targets. Similar improvements are observed with low-ranked templates from SPARKS(2). The template clustering approach could be applied to other modeling methods that utilize multiple templates to improve structure prediction.

Automated protein structure modeling with SWISS-MODEL Workspace and the Protein Model Portal

Comparative protein structure modeling is a computational approach to build three-dimensional structural models for proteins using experimental structures of related protein family members as templates. Regular blind assessments of modeling accuracy have demonstrated that comparative protein structure modeling is currently the most reliable technique to model protein structures. Homology models are often sufficiently accurate to substitute for experimental structures in a wide variety of applications. Since the usefulness of a model for specific application is determined by its accuracy, model quality estimation is an essential component of protein structure prediction. Comparative protein modeling has become a routine approach in many areas of life science research since fully automated modeling systems allow also nonexperts to build reliable models. In this chapter, we describe practical approaches for automated protein structure modeling with SWISS-MODEL Workspace and the Protein Model Portal.

Computational modeling of laminin N-terminal domains using sparse distance constraints from disulfide bonds and chemical cross-linking

Basement membranes are thin extracellular protein layers, which separate endothelial and epithelial cells from the underlying connecting tissue. The main noncollagenous components of basement membranes are laminins, trimeric glycoproteins, which form polymeric networks by interactions of their N-terminal (LN) domains; however, no high-resolution structure of laminin LN domains exists so far. To construct models for laminin β(1) and γ(1) LN domains, 14 potentially suited template structures were determined using fold recognition methods. For each target/template-combination comparative models were created with Rosetta. Final models were selected based on their agreement with experimentally obtained distance constraints from natural cross-links, that is, disulfide bonds as well as chemical cross-links obtained from reactions with two amine-reactive cross-linkers. We predict that laminin β(1) and γ(1) LN domains share the galactose-binding domain-like fold.

A chaperone cascade sorts proteins for posttranslational membrane insertion into the endoplasmic reticulum

Tail-anchored (TA) proteins are posttranslationally inserted into either the endoplasmic reticulum (ER) or the mitochondrial outer membrane. The C-terminal transmembrane domains (TMDs) of TA proteins enable their many essential cellular functions by specifying the membrane target, but how cells process these targeting signals is poorly understood. Here, we reveal the composition of a conserved multiprotein TMD recognition complex (TRC) and show that distinct TRC subunits recognize the two types of TMD signals. By engineering mutations in a mitochondrial TMD, we switch over its TRC subunit recognition, thus leading to its misinsertion into the ER. Biochemical reconstitution with purified components demonstrates that TRC tethers and enzymatically activates Get3 to selectively hand off ER-bound TA proteins to Get3. Thus, ER-bound TA proteins are sorted at the top of a TMD chaperone cascade that culminates with the formation of Get3-TA protein complexes, which are recruited to the ER membrane for insertion.

A score of the ability of a three-dimensional protein model to retrieve its own sequence as a quantitative measure of its quality and appropriateness

BACKGROUND

Despite the remarkable progress of bioinformatics, how the primary structure of a protein leads to a three-dimensional fold, and in turn determines its function remains an elusive question. Alignments of sequences with known function can be used to identify proteins with the same or similar function with high success. However, identification of function-related and structure-related amino acid positions is only possible after a detailed study of every protein. Folding pattern diversity seems to be much narrower than sequence diversity, and the amino acid sequences of natural proteins have evolved under a selective pressure comprising structural and functional requirements acting in parallel.

PRINCIPAL FINDINGS

The approach described in this work begins by generating a large number of amino acid sequences using ROSETTA [Dantas G et al. (2003) J Mol Biol 332:449-460], a program with notable robustness in the assignment of amino acids to a known three-dimensional structure. The resulting sequence-sets showed no conservation of amino acids at active sites, or protein-protein interfaces. Hidden Markov models built from the resulting sequence sets were used to search sequence databases. Surprisingly, the models retrieved from the database sequences belonged to proteins with the same or a very similar function. Given an appropriate cutoff, the rate of false positives was zero. According to our results, this protocol, here referred to as Rd.HMM, detects fine structural details on the folding patterns, that seem to be tightly linked to the fitness of a structural framework for a specific biological function.

CONCLUSION

Because the sequence of the native protein used to create the Rd.HMM model was always amongst the top hits, the procedure is a reliable tool to score, very accurately, the quality and appropriateness of computer-modeled 3D-structures, without the need for spectroscopy data. However, Rd.HMM is very sensitive to the conformational features of the models' backbone.

BCL::contact-low confidence fold recognition hits boost protein contact prediction and de novo structure determination

Knowledge of all residue-residue contacts within a protein allows determination of the protein fold. Accurate prediction of even a subset of long-range contacts (contacts between amino acids far apart in sequence) can be instrumental for determining tertiary structure. Here we present BCL::Contact, a novel contact prediction method that utilizes artificial neural networks (ANNs) and specializes in the prediction of medium to long-range contacts. BCL::Contact comes in two modes: sequence-based and structure-based. The sequence-based mode uses only sequence information and has individual ANNs specialized for helix-helix, helix-strand, strand-helix, strand-strand, and sheet-sheet contacts. The structure-based mode combines results from 32-fold recognition methods with sequence information to a consensus prediction. The two methods were presented in the 6(th) and 7(th) Critical Assessment of Techniques for Protein Structure Prediction (CASP) experiments. The present work focuses on elucidating the impact of fold recognition results onto contact prediction via a direct comparison of both methods on a joined benchmark set of proteins. The sequence-based mode predicted contacts with 42% accuracy (7% false positive rate), while the structure-based mode achieved 45% accuracy (2% false positive rate). Predictions by both modes of BCL::Contact were supplied as input to the protein tertiary structure prediction program Rosetta for a benchmark of 17 proteins with no close sequence homologs in the protein data bank (PDB). Rosetta created higher accuracy models, signified by an improvement of 1.3 A on average root mean square deviation (RMSD), when driven by the predicted contacts. Further, filtering Rosetta models by agreement with the predicted contacts enriches for native-like fold topologies.

Performance of the Pro-sp3-TASSER server in CASP8

The performance of the protein structure prediction server pro-sp3-TASSER in CASP8 is described. Compared to CASP7, the major improvement in prediction is in the quality of input models to TASSER. These improvements are due to the PRO-SP(3) threading method, the improved quality of contact predictions provided by TASSER_2.0, multiple short TASSER simulations for building the full-length model, and the accuracy of model selection using the TASSER-QA quality assessment method. Finally, we analyze the overall performance and highlight some successful predictions of the pro-sp3-TASSER server.