In silico protein design by combinatorial assembly of protein building blocks

Design Glutamate Dehydrogenase for Nonaqueous System by Motifs Reassembly and Interaction Network Analysis

Glutamate dehydrogenases (GDH) serve as the key regulated enzyme that links protein and carbohydrate metabolism. Combined with motif reassembly and mutation, novel GDHs were designed. Motif reassembly of thermophilic GDH and malate dehydrogenase aims to overcome stability and activity tradeoff in nonaqueous systems. Structural compatibility and dynamic cooperation of the designed AaDHs were studied by molecular dynamics simulation. Furthermore, multipoint mutations improved its catalytic activity for unnatural substrates. Amino acid interaction network analysis indicated that the high density of hydrogen-bonded salt bridges is beneficial to the stability. Finally, the experimental verification determines the kinetics of AaDHs in a nonaqueous system. The activity of Aa05 was increased by 1.78-fold with ionic liquid [EMIM]BF4. This study presents the strategy of a combination of rigid motif assembly and mutations of active sites for robust dehydrogenases with high activity in the nonaqueous system, which overcomes the activity-stability tradeoff effect.

Identification of Prospective Ebola Virus VP35 and VP40 Protein Inhibitors from Myxobacterial Natural Products

The Ebola virus (EBOV) is a lethal pathogen causing hemorrhagic fever syndrome which remains a global health challenge. In the EBOV, two multifunctional proteins, VP35 and VP40, have significant roles in replication, virion assembly, and budding from the cell and have been identified as druggable targets. In this study, we employed in silico methods comprising molecular docking, molecular dynamic simulations, and pharmacological properties to identify prospective drugs for inhibiting VP35 and VP40 proteins from the myxobacterial bioactive natural product repertoire. Cystobactamid 934-2, Cystobactamid 919-1, and Cittilin A bound firmly to VP35. Meanwhile, 2-Hydroxysorangiadenosine, Enhypyrazinone B, and Sorangiadenosine showed strong binding to the matrix protein VP40. Molecular dynamic simulations revealed that, among these compounds, Cystobactamid 919-1 and 2-Hydroxysorangiadenosine had stable interactions with their respective targets. Similarly, molecular mechanics Poisson-Boltzmann surface area (MMPBSA) calculations indicated close-fitting receptor binding with VP35 or VP40. These two compounds also exhibited good pharmacological properties. In conclusion, we identified Cystobactamid 919-1 and 2-Hydroxysorangiadenosine as potential ligands for EBOV that target VP35 and VP40 proteins. These findings signify an essential step in vitro and in vivo to validate their potential for EBOV inhibition.

Cheminformatics-Based Study Identifies Potential Ebola VP40 Inhibitors

The Ebola virus (EBOV) is still highly infectious and causes severe hemorrhagic fevers in primates. However, there are no regulatorily approved drugs against the Ebola virus disease (EVD). The highly virulent and lethal nature of EVD highlights the need to develop therapeutic agents. Viral protein 40 kDa (VP40), the most abundantly expressed protein during infection, coordinates the assembly, budding, and release of viral particles into the host cell. It also regulates viral transcription and RNA replication. This study sought to identify small molecules that could potentially inhibit the VP40 protein by targeting the N-terminal domain using an in silico approach. The statistical quality of AutoDock Vina’s capacity to discriminate between inhibitors and decoys was determined, and an area under the curve of the receiver operating characteristic (AUC-ROC) curve of 0.791 was obtained. A total of 29,519 natural-product-derived compounds from Chinese and African sources as well as 2738 approved drugs were successfully screened against VP40. Using a threshold of −8 kcal/mol, a total of 7, 11, 163, and 30 compounds from the AfroDb, Northern African Natural Products Database (NANPDB), traditional Chinese medicine (TCM), and approved drugs libraries, respectively, were obtained after molecular docking. A biological activity prediction of the lead compounds suggested their potential antiviral properties. In addition, random-forest- and support-vector-machine-based algorithms predicted the compounds to be anti-Ebola with IC50 values in the micromolar range (less than 25 μM). A total of 42 natural-product-derived compounds were identified as potential EBOV inhibitors with desirable ADMET profiles, comprising 1, 2, and 39 compounds from NANPDB (2-hydroxyseneganolide), AfroDb (ZINC000034518176 and ZINC000095485942), and TCM, respectively. A total of 23 approved drugs, including doramectin, glecaprevir, velpatasvir, ledipasvir, avermectin B1, nafarelin acetate, danoprevir, eltrombopag, lanatoside C, and glycyrrhizin, among others, were also predicted to have potential anti-EBOV activity and can be further explored so that they may be repurposed for EVD treatment. Molecular dynamics simulations coupled with molecular mechanics Poisson–Boltzmann surface area calculations corroborated the stability and good binding affinities of the complexes (−46.97 to −118.9 kJ/mol). The potential lead compounds may have the potential to be developed as anti-EBOV drugs after experimental testing.

Investigating the functional role of a buried interchain aromatic cluster in Escherichia coli GrpE dimer

Aromatic clusters in the core of proteins are often involved in imparting structural stability to proteins. However, their functional importance is not always clear. In this study, we investigate the thermosensing role of a phenylalanine cluster present in the GrpE homodimer. GrpE, which acts as a nucleotide exchange factor for the molecular chaperone DnaK, is well known for its thermosensing activity resulting from temperature-dependent structural changes that allow control of chaperone function. Using mutational analysis, we show that an interchain phenylalanine cluster in a four-helix bundle of the GrpE homodimer assists in the thermosensing ability of the co-chaperone. Substitution of aromatic residues with hydrophobic ones in the core of the four-helix bundle reduces the thermal stability of the bundle and that of a connected coiled-coil domain, which impacts thermosensing. Cell growth assays and SEM images of the mutants show filamentous growth of Escherichia coli cells at 42°C, which corroborates with the defect in thermosensing. Our work suggests that the interchain edge-to-face aromatic cluster is important for the propagation of the structural signal from the coiled-coil domain to the four-helical bundle of GrpE, thus facilitating GrpE-mediated thermosensing in bacteria.

Analysis of SLC26A4, FOXI1, and KCNJ10 Gene Variants in Patients with Incomplete Partition of the Cochlea and Enlarged Vestibular Aqueduct (EVA) Anomalies

Pathogenic variants in the SLC26A4, FOXI1, and KCNJ10 genes are associated with hearing loss (HL) and specific inner ear abnormalities (DFNB4). In the present study, phenotype analyses, including clinical data collection, computed tomography (CT), and audiometric examination, were performed on deaf individuals from the Sakha Republic of Russia (Eastern Siberia). In cases with cochleovestibular malformations, molecular genetic analysis of the coding regions of the SLC26A4, FOXI1, and KCNJ10 genes associated with DFNB4 was completed. In six of the 165 patients (3.6%), CT scans revealed an incomplete partition of the cochlea (IP-1 and IP-2), in isolation or combined with an enlarged vestibular aqueduct (EVA) anomaly. Sequencing of the SLC26A4, FOXI1, and KCNJ10 genes was performed in these six patients. In the SLC26A4 gene, we identified four variants, namely c.85G>C p.(Glu29Gln), c.757A>G p.(Ile253Val), c.2027T>A p.(Leu676Gln), and c.2089+1G>A (IVS18+1G>A), which are known as pathogenic, as well as c.441G>A p.(Met147Ile), reported previously as a variant with uncertain significance. Using the AlphaFold algorithm, we found in silico evidence of the pathogenicity of this variant. We did not find any causative variants in the FOXI1 and KCNJ10 genes, nor did we find any evidence of digenic inheritance associated with double heterozygosity for these genes with monoallelic SLC26A4 variants. The contribution of biallelic SLC26A4 variants in patients with IP-1, IP-2, IP-2+EVA, and isolated EVA was 66.7% (DFNB4 in three patients, Pendred syndrome in one patient). Seventy-five percent of SLC26A4-biallelic patients had severe or profound HL. The morphology of the inner ear anomalies demonstrated that, among SLC26A4-biallelic patients, all types of incomplete partition of the cochlea are possible, from IP-1 and IP-2, to a normal cochlea. However, the dominant type of anomaly was IP-2+EVA (50.0%). This finding is very important for cochlear implantation, since the IP-2 anomaly does not have an increased risk of “gushers” and recurrent meningitis.

Protein Function Analysis through Machine Learning

Machine learning (ML) has been an important arsenal in computational biology used to elucidate protein function for decades. With the recent burgeoning of novel ML methods and applications, new ML approaches have been incorporated into many areas of computational biology dealing with protein function. We examine how ML has been integrated into a wide range of computational models to improve prediction accuracy and gain a better understanding of protein function. The applications discussed are protein structure prediction, protein engineering using sequence modifications to achieve stability and druggability characteristics, molecular docking in terms of protein-ligand binding, including allosteric effects, protein-protein interactions and protein-centric drug discovery. To quantify the mechanisms underlying protein function, a holistic approach that takes structure, flexibility, stability, and dynamics into account is required, as these aspects become inseparable through their interdependence. Another key component of protein function is conformational dynamics, which often manifest as protein kinetics. Computational methods that use ML to generate representative conformational ensembles and quantify differences in conformational ensembles important for function are included in this review. Future opportunities are highlighted for each of these topics.

Protein Model Quality Estimation Using Molecular Dynamics Simulation

The estimation of protein model quality remains a challenging task and is important for protein structural model utilization. In the last decade, existing methods that rely on machine learning to deep learning have been developed and shown progressive improvement. Despite utilizing more sophisticated techniques and introducing new features, none of these methods employ explicit protein structure stability information. Hypothetically, protein model quality might be indicated by its structural stability in an in silico system disclosed by the structural difference from its initial structure. One of the possible methods to exploit such information is by implementing molecular dynamics simulations that have shown successful applications in many research fields. We present a novel approach by introducing explicit protein structure stability information using molecular dynamics simulation. Despite using only simple features, small data with no training process required, and a short molecular dynamics simulation time, our method shows comparable performance to the state-of-the-art deep learning-based method.

Identification of dual active sites in Caenorhabditis elegans GANA-1 protein: an ortholog of the human α-GAL a and α-NAGA enzymes

Fabry disease (FD) is caused by a defective α-galactosidase A (α-GAL A) enzyme responsible for breaking down globotriaosylceramide (Gb₃). To develop affordable therapeutics, more effort is needed to obtain insights into the underlying mechanism of FD and understanding human α-GAL A structure and function in related animal models. We adopted C. elegans as a model to elucidate the sequence and 3D structure of its GANA-1 enzyme and compared it to human α-GAL A. We constructed GANA-1 3D structure by homology modelling and validated the quality of the predicted GANA-1 structure, followed by computational docking of human ligands. The GANA-1 protein shared sequence similarities up to 42.1% with the human α-GAL A in silico and had dual active sites. GANA-1 homology modelling showed that 11 out of 13 amino acids in the first active site of GANA-1 protein overlapped with the human α-GAL A active site, indicating the prospect for substrate cross-reaction. Computational molecular docking using human ligands like Gb₃ (first pocket)_, 4-nitrophenyl-α-D-galactopyranoside (second pocket), α-galactose (second pocket), and N-acetyl-D-galactosamine (second pocket) showed negative binding energy. This revealed that the ligands were able to bind within both GANA-1 active sites, mimicking the human α-GAL A and α-NAGA enzymes. We identified human compounds with adequate docking scores, predicting robust interactions with the GANA-1 active site. Our data suggested that the C. elegans GANA-1 enzyme may possess structural and functional similarities to human α-GAL A, including an intrinsic capability to metabolize Gb₃ deposits.Communicated by Ramaswamy H. Sarma.

Expression and distribution of synaptotagmin family members in the zebrafish retina

Synaptotagmins belong to a large family of proteins. Although various synaptotagmins have been implicated as Ca²⁺ sensors for vesicle replenishment and release at conventional synapses, their roles at retinal ribbon synapses remain incompletely understood. Zebrafish is a widely used experimental model for retinal research. We therefore investigated the homology between human, rat, mouse, and zebrafish synaptotagmins 1-10 using a bioinformatics approach. We also characterized the expression and distribution of various synaptotagmin (syt) genes in the zebrafish retina using RT-PCR, qPCR, and in situhybridization, focusing on the family members whose products likely underlie Ca²⁺ -dependent exocytosis in the central nervous system (synaptotagmins 1, 2, 5, and 7). Most zebrafish synaptotagmins are well conserved and can be grouped in the same classes as mammalian synaptotagmins, based on crucial amino acid residues needed for coordinating Ca²⁺ binding and determining phospholipid binding affinity. The only exception is synaptotagmin 1b, which lacks 34 amino acid residues in the C2B domain and is therefore unlikely to bind Ca²⁺ there. Additionally, the products of zebrafish syt5a and syt5b genes share identity with mammalian class 1 and 5 synaptotagmins. Zebrafish syt1, syt2, syt5, and syt7 paralogues are found in the zebrafish brain, eye, and retina, excepting syt1b, which is only present in the brain. The complementary expression pattern of the remaining paralogues in the retina suggests that syt1a and syt5a may underlie synchronous release and syt7a and syt7b may mediate asynchronous release or other Ca²⁺ -dependent processes in different retinal neurons.

Bioinformatic Prediction of Signaling Pathways for Apurinic/Apyrimidinic Endodeoxyribonuclease 1 (APEX1) and Its Role in Cholangiocarcinoma Cells

Apurinic/apyrimidinic endodeoxyribonuclease 1 (APEX1) is involved in the DNA damage repair pathways and associates with the metastasis of several human cancers. However, the signaling pathway of APEX1 in cholangiocarcinoma (CCA) has never been reported. In this study, to predict the signaling pathways of APEX1 and related proteins and their functions, the effects of APEX1 gene silencing on APEX1 and related protein expression in CCA cell lines were investigated using mass spectrometry and bioinformatics tools. Bioinformatic analyses predicted that APEX1 might interact with cell division cycle 42 (CDC42) and son of sevenless homolog 1 (SOS1), which are involved in tumor metastasis. RNA and protein expression levels of APEX1 and its related proteins, retrieved from the Gene Expression Profiling Interactive Analysis (GEPIA) and the Human Protein Atlas databases, revealed that their expressions were higher in CCA than in the normal group. Moreover, higher levels of APEX1 expression and its related proteins were correlated with shorter survival time. In conclusion, the signaling pathway of APEX1 in metastasis might be mediated via CDC42 and SOS1. Furthermore, expression of APEX1 and related proteins is able to predict poor survival of CCA patients.

A computational drug repurposing approach in identifying the cephalosporin antibiotic and anti-hepatitis C drug derivatives for COVID-19 treatment

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused over 1.4 million deaths worldwide. Repurposing existing drugs offers the fastest opportunity to identify new indications for existing drugs as a stable solution against coronavirus disease 2019 (COVID-19). The SARS-CoV-2 main protease (Mpro) is a critical target for designing potent antiviral agents against COVID-19. In this study, we identify potential inhibitors against COVID-19, using an amalgam of virtual screening, molecular dynamics (MD) simulations, and binding-free energy approaches from the Korea Chemical Bank drug repurposing (KCB-DR) database. The database screening of KCB-DR resulted in 149 binders. The dynamics of protein-drug complex formation for the seven top scoring drugs were investigated through MD simulations. Six drugs showed stable binding with active site of SARS-CoV-2 Mpro indicated by steady RMSD of protein backbone atoms and potential energy profiles. Furthermore, binding free energy calculations suggested the community-acquired bacterial pneumonia drug ceftaroline fosamil and the hepatitis C virus (HCV) protease inhibitor telaprevir are potent inhibitors against Mpro. Molecular dynamics and interaction analysis revealed that ceftaroline fosamil and telaprevir form hydrogen bonds with important active site residues such as Thr24, Thr25, His41, Thr45, Gly143, Ser144, Cys145, and Glu166 that is supported by crystallographic information of known inhibitors. Telaprevir has potential side effects, but its derivatives have good pharmacokinetic properties and are suggested to bind Mpro. We suggest the telaprevir derivatives and ceftaroline fosamil bind tightly with SARS-CoV-2 Mpro and should be validated through preclinical testing.

Structural insights and evaluation of the potential impact of missense variants on the interactions of SLIT2 with ROBO1/4 in cancer progression

The cognate interaction of ROBO1/4 with its ligand SLIT2 is known to be involved in lung cancer progression. However, the precise role of genetic variants, disrupting the molecular interactions is less understood. All cancer-associated missense variants of ROBO1/4 and SLIT2 from COSMIC were screened for their pathogenicity. Homology modelling was done in Modeller 9.17, followed by molecular simulation in GROMACS. Rigid docking was performed for the cognate partners in PatchDock with refinement in HADDOCK server. Post-docking alterations in conformational, stoichiometric, as well as structural parameters, were assessed. The disruptive variants were ranked using a weighted scoring scheme. In silico prioritisation of 825 variants revealed 379 to be potentially pathogenic out of which, about 12% of the variants, i.e. ROBO1 (14), ROBO4 (8), and SLIT2 (23) altered the cognate docking. Six variants of ROBO1 and 5 variants of ROBO4 were identified as "high disruptors" of interactions with SLIT2 wild type. Likewise, 17 and 13 variants of SLIT2 were found to be "high disruptors" of its interaction with ROBO1 and ROBO4, respectively. Our study is the first report on the impact of cancer-associated missense variants on ROBO1/4 and SLIT2 interactions that might be the drivers of lung cancer progression.

ProDCoNN: Protein design using a convolutional neural network

Designing protein sequences that fold to a given three-dimensional (3D) structure has long been a challenging problem in computational structural biology with significant theoretical and practical implications. In this study, we first formulated this problem as predicting the residue type given the 3D structural environment around the C _α atom of a residue, which is repeated for each residue of a protein. We designed a nine-layer 3D deep convolutional neural network (CNN) that takes as input a gridded box with the atomic coordinates and types around a residue. Several CNN layers were designed to capture structure information at different scales, such as bond lengths, bond angles, torsion angles, and secondary structures. Trained on a very large number of protein structures, the method, called ProDCoNN (protein design with CNN), achieved state-of-the-art performance when tested on large numbers of test proteins and benchmark datasets.

An evaluation of different Cripto-1 antibodies and their variable results

Cripto-1 is a protein expressed during embryonal development and has been linked to several malignant processes in cancer. Since the discovery of cripto-1 in the late 1980s, it has become a subject of biomarker investigation in several types of cancer which in many cases relies on immunolocalization of cripto-1 using antibodies. Investigating cripto-1 expression and localization in primary glioblastoma cells, we discovered nonspecific binding of cripto-1 antibody to the extracellular matrix Geltrex. A panel of four cripto-1 antibodies was investigated with respect to their binding to the Geltrex matrix and to the cripto-1 positive control cells NTERA2. The cripto-1 expression was varied for the different antibodies with respect to cellular localization and fixation methods. To further elaborate on these findings, we present a systematic review of cripto-1 antibodies found in the literature and highlight some possible cross reactants with data on sequence alignments and structural comparison of EGF domains.

Genotypic and Mechanistic Characterization of Subtype-Specific HIV Adaptation to Host Cellular Immunity

The extent to which viral genetic context influences HIV adaptation to human leukocyte antigen (HLA) class I-restricted immune pressures remains incompletely understood. The Ugandan HIV epidemic, where major pandemic group M subtypes A1 and D cocirculate in a single host population, provides an opportunity to investigate this question. We characterized plasma HIV RNA gag, pol, and nef sequences, along with host HLA genotypes, in 464 antiretroviral-naive individuals chronically infected with HIV subtype A1 or D. Using phylogenetically informed statistical approaches, we identified HLA-associated polymorphisms and formally compared their strengths of selection between viral subtypes. A substantial number (32%) of HLA-associated polymorphisms identified in subtype A1 and/or D had previously been reported in subtype B, C, and/or circulating recombinant form 01_AE (CRF01_AE), confirming the shared nature of many HLA-driven escape pathways regardless of viral genetic context. Nevertheless, 34% of the identified HLA-associated polymorphisms were significantly differentially selected between subtypes A1 and D. Experimental investigation of select examples of subtype-specific escape revealed distinct underlying mechanisms with important implications for vaccine design: whereas some were attributable to subtype-specific sequence variation that influenced epitope-HLA binding, others were attributable to differential mutational barriers to immune escape. Overall, our results confirm that HIV genetic context is a key modulator of viral adaptation to host cellular immunity and highlight the power of combined bioinformatic and mechanistic studies, paired with knowledge of epitope immunogenicity, to identify appropriate viral regions for inclusion in subtype-specific and universal HIV vaccine strategies.IMPORTANCE The identification of HIV polymorphisms reproducibly selected under pressure by specific HLA alleles and the elucidation of their impact on viral function can help identify immunogenic viral regions where immune escape incurs a fitness cost. However, our knowledge of HLA-driven escape pathways and their functional costs is largely limited to HIV subtype B and, to a lesser extent, subtype C. Our study represents the first characterization of HLA-driven adaptation pathways in HIV subtypes A1 and D, which dominate in East Africa, and the first statistically rigorous characterization of differential HLA-driven escape across viral subtypes. The results support a considerable impact of viral genetic context on HIV adaptation to host HLA, where HIV subtype-specific sequence variation influences both epitope-HLA binding and the fitness costs of escape. Integrated bioinformatic and mechanistic characterization of these and other instances of differential escape could aid rational cytotoxic T-lymphocyte-based vaccine immunogen selection for both subtype-specific and universal HIV vaccines.

IDeAS: automated design tool for hetero-chiral protein folds

Incorporating D amino acids in the protein design alphabet can in principle multiply the design space by many orders of magnitude. All native proteins are polymers composed of L chiral amino acids. Practically limitless in diversity over amino acid sequences, protein structure is limited in folds and thus shapes, principally due to the poly L stereochemistry of their backbone. To diversify shapes, we introduced both L- and D α-amino acids as design alphabets to explore the possibility of generating novel folds, varied in chemical as well as stereo-chemical sequence. Now, to have stereochemically-defined proteins tuned chemically, we present the Inverse Design and Automation Software, IDeAS. Retro-fitting side chains on a backbone with L and D stereochemistry, the software demonstrate functional fits over stereo-chemically diverse folds in a range of applications of interest in protein design.

Fragger: a protein fragment picker for structural queries

Protein modeling and design activities often require querying the Protein Data Bank (PDB) with a structural fragment, possibly containing gaps. For some applications, it is preferable to work on a specific subset of the PDB or with unpublished structures. These requirements, along with specific user needs, motivated the creation of a new software to manage and

query 3D protein fragments. Fragger is a protein fragment picker that allows protein fragment databases to be created and queried. All fragment lengths are supported and any set of PDB files can be used to create a database. Fragger can efficiently search a fragment database with a query fragment and a distance threshold. Matching fragments are ranked by distance to the query. The query fragment can have structural gaps and the allowed amino acid sequences matching a query can be constrained via a regular expression of one-letter amino acid codes. Fragger also incorporates a tool to compute the backbone RMSD of one versus many fragments in high throughput. Fragger should be useful for protein design, loop grafting and related structural

bioinformatics tasks.

A Protocol for the Design of Protein and Peptide Nanostructure Self-Assemblies Exploiting Synthetic Amino Acids

In recent years there has been increasing interest in nanostructure design based on the self-assembly properties of proteins and polymers. Nanodesign requires the ability to predictably manipulate the properties of the self-assembly of autonomous building blocks, which can fold or aggregate into preferred conformational states. The design includes functional synthetic materials and biological macromolecules. Autonomous biological building blocks with available 3D structures provide an extremely rich and useful resource. Structural databases contain large libraries of protein molecules and their building blocks with a range of sizes, shapes, surfaces, and chemical properties. The introduction of engineered synthetic residues or short peptides into these building blocks can greatly expand the available chemical space and enhance the desired properties. Herein, we summarize a protocol for designing nanostructures consisting of self-assembling building blocks, based on our recent works. We focus on the principles of nanostructure design with naturally occurring proteins and synthetic amino acids, as well as hybrid materials made of amyloids and synthetic polymers.

Energy functions in de novo protein design: current challenges and future prospects

In the past decade, a concerted effort to successfully capture specific tertiary packing interactions produced specific three-dimensional structures for many de novo designed proteins that are validated by nuclear magnetic resonance and/or X-ray crystallographic techniques. However, the success rate of computational design remains low. In this review, we provide an overview of experimentally validated, de novo designed proteins and compare four available programs, RosettaDesign, EGAD, Liang-Grishin, and RosettaDesign-SR, by assessing designed sequences computationally. Computational assessment includes the recovery of native sequences, the calculation of sizes of hydrophobic patches and total solvent-accessible surface area, and the prediction of structural properties such as intrinsic disorder, secondary structures, and three-dimensional structures. This computational assessment, together with a recent community-wide experiment in assessing scoring functions for interface design, suggests that the next-generation protein-design scoring function will come from the right balance of complementary interaction terms. Such balance may be found when more negative experimental data become available as part of a training set.

A polypeptide "building block" for the β-trefoil fold identified by "top-down symmetric deconstruction"

Fibroblast growth factor-1, a member of the 3-fold symmetric β-trefoil fold, was subjected to a series of symmetric constraint mutations in a process termed "top-down symmetric deconstruction." The mutations enforced a cumulative exact 3-fold symmetry upon symmetrically equivalent positions within the protein and were combined with a stability screen. This process culminated in a β-trefoil protein with exact 3-fold primary-structure symmetry that exhibited excellent folding and stability properties. Subsequent fragmentation of the repeating primary-structure motif yielded a 42-residue polypeptide capable of spontaneous assembly as a homotrimer, producing a thermostable β-trefoil architecture. The results show that despite pronounced reduction in sequence complexity, pure symmetry in the design of a foldable, thermostable β-trefoil fold is possible. The top-down symmetric deconstruction approach provides a novel alternative means to successfully identify a useful polypeptide "building block" for subsequent "bottom-up" de novo design of target protein architecture.

Improving computational protein design by using structure-derived sequence profile

Designing a protein sequence that will fold into a predefined structure is of both practical and fundamental interest. Many successful, computational designs in the last decade resulted from improved understanding of hydrophobic and polar interactions between side chains of amino acid residues in stabilizing protein tertiary structures. However, the coupling between main-chain backbone structure and local sequence has yet to be fully addressed. Here, we attempt to account for such coupling by using a sequence profile derived from the sequences of five residue fragments in a fragment library that are structurally matched to the five-residue segments contained in a target structure. We further introduced a term to reduce low complexity regions of designed sequences. These two terms together with optimized reference states for amino-acid residues were implemented in the RosettaDesign program. The new method, called RosettaDesign-SR, makes a 12% increase (from 34 to 46%) in fraction of proteins whose designed sequences are more than 35% identical to wild-type sequences. Meanwhile, it reduces 8% (from 22% to 14%) to the number of designed sequences that are not homologous to any known protein sequences according to psi-blast. More importantly, the sequences designed by RosettaDesign-SR have 2-3% more polar residues at the surface and core regions of proteins and these surface and core polar residues have about 4% higher sequence identity to wild-type sequences than by RosettaDesign. Thus, the proteins designed by RosettaDesign-SR should be less likely to aggregate and more likely to have unique structures due to more specific polar interactions.

A trade-off relationship between energetic cost and entropic cost for in vitro evolution

In this paper, we consider two complementary cost functions for the landscape exploring processes to obtain the global optimum sequence through in vitro evolution protocol: one is the entropic cost C(etp), which is based on the deviation from the uniformity of a mutant distribution in sequence space, and the other is the energetic cost C(eng), which is based on the total number of sequences to be generated and evaluated. Based on a prior knowledge about the structure of a given fitness landscapes, the conductor of the experiment can think up the efficient search algorithm (ESA), which requires the minimum number of points (=sequences) to be searched up to the global optimum. For five typical fitness landscapes, we considered their respective (putative) ESA, C(etp)(*) and C(eng)(*) based on the ESA. As a result, we found a trade-off relationship between C(etp)(*) and C(eng)(*) for every case, that is, C(eng)(*)+C(etp)(*) is approximately equal to the logarithm of the volume of the sequence space. C(etp)(*) and C(eng)(*) are interpreted in terms of the information-theoretic concepts.

Exploring protein structural dissimilarity to facilitate structure classification

BACKGROUND

Classification of newly resolved protein structures is important in understanding their architectural, evolutionary and functional relatedness to known protein structures. Among various efforts to improve the database of Structural Classification of Proteins (SCOP), automation has received particular attention. Herein, we predict the deepest SCOP structural level that an unclassified protein shares with classified proteins with an equal number of secondary structure elements (SSEs).

RESULTS

We compute a coefficient of dissimilarity (Omega) between proteins, based on structural and sequence-based descriptors characterising the respective constituent SSEs. For a set of 1,661 pairs of proteins with sequence identity up to 35%, the performance of Omega in predicting shared Class, Fold and Super-family levels is comparable to that of DaliLite Z score and shows a greater than four-fold increase in the true positive rate (TPR) for proteins sharing the Family level. On a larger set of 600 domains representing 200 families, the performance of Z score improves in predicting a shared Family, but still only achieves about half of the TPR of Omega. The TPR for structures sharing a Super-family is lower than in the first dataset, but Omega performs slightly better than Z score. Overall, the sensitivity of Omega in predicting common Fold level is higher than that of the DaliLite Z score.

CONCLUSION

Classification to a deeper level in the hierarchy is specific and difficult. So the efficiency of Omega may be attractive to the curators and the end-users of SCOP. We suggest Omega may be a better measure for structure classification than the DaliLite Z score, with the caveat that currently we are restricted to comparing structures with equal number of SSEs.

Structural analysis of a beta-helical protein motif stabilized by targeted replacements with conformationally constrained amino acids

Here we study conformational stabilization induced in a beta-helical nanostructure by position-specific mutations. The nanostructure is constructed through the self-assembly of the beta-helical building block excised from E. coli galactoside acetyltransferase (PDB code 1krr , chain A; residues 131-165). The mutations involve substitutions by cyclic, conformationally constrained amino acids. Specifically, a complete structural analysis of the Pro-Xaa-Val sequence [with Xaa being Gly, Ac 3c (1-aminocyclopropane-1-carboxylic acid) and Ac 5c (1-aminocyclopentane-1-carboxylic acid)], corresponding to the 148-150 loop region in the wild-type (Gly) and mutated (Ac 3c and Ac 5c) 1krr , has been performed using Molecular Dynamics simulations and X-ray crystallography. Simulations have been performed for the wild-type and mutants of three different systems, namely the building block, the nanoconstruct and the isolated Pro-Xaa-Val tripeptide. Furthermore, the crystalline structures of five peptides of Pro-Xaa-Val or Xaa-Val sequences have been solved by X-ray diffraction analysis and compared with theoretical predictions. Both the theoretical and crystallographic studies indicate that the Pro-Ac n c-Val sequences exhibit a high propensity to adopt turn-like conformations, and this propensity is little affected by the chemical environment. Overall, the results indicate that replacement of Gly149 by Ac 3c or Ac 5c significantly reduce the conformational flexibility of the target site enhancing the structural specificity of the building block and the nanoconstruct derived from the 1krr beta-helical motif.

Principles of nanostructure design with protein building blocks

Currently there is increasing interest in nanostructures and their design. Nanostructure design involves the ability to predictably manipulate the properties of the self-assembly of autonomous units. Autonomous units have preferred conformational states. The units can be synthetic material science-based or derived from functional biological macromolecules. Autonomous biological building blocks with available structures provide an extremely rich and useful resource for design. For proteins, the structural databases contain large libraries of protein molecules and their building blocks with a range of shapes, surfaces, and chemical properties. The introduction of engineered synthetic residues or short peptides into these can expand the available chemical space and enhance the desired properties. Here we focus on the principles of nanostructure design with protein building blocks.

"Pinning strategy": a novel approach for predicting the backbone structure in terms of protein blocks from sequence

The description of protein 3D structures can be performed through a library of 3D fragments, named a structural alphabet. Our structural alphabet is composed of 16 small protein fragments of 5 C alpha in length, called protein blocks (PBs). It allows an efficient approximation of the 3D protein structures and a correct prediction of the local structure. The 72 most frequent series of 5 consecutive PBs, called structural words (SWs)are able to cover more than 90% of the 3D structures. PBs are highly conditioned by the presence of a limited number of transitions between them. In this study, we propose a new method called "pinning strategy" that used this specific feature to predict long protein fragments. Its goal is to define highly probable successions of PBs. It starts from the most probable SW and is then extended with overlapping SWs. Starting from an initial prediction rate of 34.4%, the use of the SWs instead of the PBs allows a gain of 4.5%. The pinning strategy simply applied to the SWs increases the prediction accuracy to 39.9%. In a second step, the sequence-structure relationship is optimized, the prediction accuracy reaches 43.6%.

Stability of tubular structures based on beta-helical proteins: self-assembled versus polymerized nanoconstructs and wild-type versus mutated sequences

In this work we used atomistic molecular dynamics simulations to examine different aspects of tubular nanostructures constructed using protein building blocks with a beta-helical conformation. Initially, we considered two different natural protein building blocks, which were extracted from the protein data base, to compare the relative stabilities of the nanotubes obtained made of self-assembled and covalently linked repeats. Results show nanotubes constructed by linking building blocks through covalent bonds are very stable suggesting that the basic principles of polymer physics are valid when the repeating units are made of large fragments of proteins. In contrast, the stability of self-assembled nanostructures strongly depends on the attractive nonbonding interactions associated to building blocks aligned in a complementary manner. On the other hand, we investigated the ability of a conformationally constrained synthetic amino acid to enhance the stability of both self-assembled and polymerized nanotubes when it is used to substitute natural residues. Specifically, we considered 1-aminocyclopentane-1-caboxylic acid, which involves strong stereochemical constraints produced by the cyclopentane side chain. We found that the incorporation of this amino acid within the more flexible regions of the beta-helical building blocks is an excellent strategy to enhance the stability of the nanotubes. Thus, when a single mutation is performed in the loop region of the beta-helix, the bend architecture of the whole loop is stabilized since the conformational mobility is reduced not only at the mutated position but also at the adjacent positions.

Coarse-grained representation of beta-helical protein building blocks

A general strategy to develop coarse-grained models of beta-helical protein fragments is presented. The procedure has been applied to a building block formed by a two-turn repeat motif from E. coli galactoside acetyltransferase, which is able to provide a very stable self-assembled tubular nanoconstruct upon stacking of its replicas. For this purpose, first, we have developed a computational scheme to sample very efficiently the configurational space of the building block. This method, which is inspired by a strategy recently designed to study amorphous polymers and by an advanced Monte Carlo algorithm, provides a large ensemble of uncorrelated configurations at a very reasonable computational cost. The atomistic configurations provided by this method have been used to obtain a coarse-grained model that describes the amino acids with fewer particles than those required for full atomistic detail, i.e., two, three, or four depending on the chemical nature of the amino acid. Coarse-grained potentials have been developed considering the following types of interactions: (i) electrostatic and van der Waals interactions between residues i and i + n with n >/= 2; (ii) interactions between residues i and i + 1; and (c) intra-residue interactions. The reliability of the proposed model has been tested by comparing the atomistic and coarse-grained energies calculated for a large number of independent configurations of the beta-helical building block.

Molecular diversity in engineered protein libraries

Engineered protein libraries, defined here as a collection of different mutant variants of a single specific protein, are intentionally designed to be rich in molecular diversity and can span ranges from as little as 400 different variants to greater than 10(12) members per library. The goal of engineering libraries is to generate new protein variants, identified upon screening, that possess desired novel properties. Exploitation of the natural organization of the genetic code has led to 'focused' libraries that are lower in overall complexity yet biased towards variants with preferred biophysical properties. An emerging trend, in which computational algorithms are blended with in vivo screens, is also leading towards greater and more rapid success in the field of protein design.

Structure by design: from single proteins and their building blocks to nanostructures

Nanotechnology realizes the advantages of naturally occurring biological macromolecules and their building-block nature for design. Frequently, assembly starts with the choice of a "good" molecule that is synthetically optimized towards the desired shape. By contrast, we propose starting with a pre-specified nanostructure shape, selecting candidate protein building blocks from a library and mapping them onto the shape and, finally, testing the stability of the construct. Such a shape-based, part-assembly strategy is conceptually similar to protein design through the combinatorial assembly of building blocks. If the conformational preferences of the building blocks are retained and their interactions are favorable, the nanostructure will be stable. The richness of the conformations, shapes and chemistries of the protein building blocks suggests a broad range of potential applications; at the same time, it also highlights their complexity. In this Opinion article, we focus on the first step: validating such a strategy against experimental data.

Designing a nanotube using naturally occurring protein building blocks

Here our goal is to carry out nanotube design using naturally occurring protein building blocks. Inspection of the protein structural database reveals the richness of the conformations of proteins, their parts, and their chemistry. Given target functional protein nanotube geometry, our strategy involves scanning a library of candidate building blocks, combinatorially assembling them into the shape and testing its stability. Since self-assembly takes place on time scales not affordable for computations, here we propose a strategy for the very first step in protein nanotube design: we map the candidate building blocks onto a planar sheet and wrap the sheet around a cylinder with the target dimensions. We provide examples of three nanotubes, two peptide and one protein, in atomistic model detail for which there are experimental data. The nanotube models can be used to verify a nanostructure observed by low-resolution experiments, and to study the mechanism of tube formation.

A permissive secondary structure-guided superposition tool for clustering of protein fragments toward protein structure prediction via fragment assembly

MOTIVATION

Secondary-Structure Guided Superposition tool (SSGS) is a permissive secondary structure-based algorithm for matching of protein structures and in particular their fragments. The algorithm was developed towards protein structure prediction via fragment assembly.

RESULTS

In a fragment-based structural prediction scheme, a protein sequence is cut into building blocks (BBs). The BBs are assembled to predict their relative 3D arrangement. Finally, the assemblies are refined. To implement this prediction scheme, a clustered structural library representing sequence patterns for protein fragments is essential. To create a library, BBs generated by cutting proteins from the PDB are compared and structurally similar BBs are clustered. To allow structural comparison and clustering of the BBs, which are often relatively short with flexible loops, we have devised SSGS. SSGS maintains high similarity between cluster members and is highly efficient. When it comes to comparing BBs for clustering purposes, the algorithm obtains better results than other, non-secondary structure guided protein superimposition algorithms.

A structural alphabet for local protein structures: improved prediction methods

Three-dimensional protein structures can be described with a library of 3D fragments that define a structural alphabet. We have previously proposed such an alphabet, composed of 16 patterns of five consecutive amino acids, called Protein Blocks (PBs). These PBs have been used to describe protein backbones and to predict local structures from protein sequences. The Q16 prediction rate reaches 40.7% with an optimization procedure. This article examines two aspects of PBs. First, we determine the effect of the enlargement of databanks on their definition. The results show that the geometrical features of the different PBs are preserved (local RMSD value equal to 0.41 A on average) and sequence-structure specificities reinforced when databanks are enlarged. Second, we improve the methods for optimizing PB predictions from sequences, revisiting the optimization procedure and exploring different local prediction strategies. Use of a statistical optimization procedure for the sequence-local structure relation improves prediction accuracy by 8% (Q16 = 48.7%). Better recognition of repetitive structures occurs without losing the prediction efficiency of the other local folds. Adding secondary structure prediction improved the accuracy of Q16 by only 1%. An entropy index (Neq), strongly related to the RMSD value of the difference between predicted PBs and true local structures, is proposed to estimate prediction quality. The Neq is linearly correlated with the Q16 prediction rate distributions, computed for a large set of proteins. An "expected" prediction rate QE16 is deduced with a mean error of 5%.

Concepts and schemes for the re-engineering of physical protein modules: generating nanodevices via targeted replacements with constrained amino acids

Physically building complex multi-molecular structures from naturally occurring biological macromolecules has aroused a great deal of interest. Here we focus on nanostructures composed of re-engineered, natural 'foldamer' building blocks. Our aim is to provide some of the underlying concepts and schemes for crafting structures utilizing such conformationally relatively stable molecular components. We describe how, via chemical biology strategies, it is further possible to chemically manipulate the foldamer building blocks toward specific shape-driven structures, which in turn could be used toward potential-designed functions. We outline the criteria in choosing candidate foldamers from the vast biological repertoire, and how to enhance their stability through selected targeted replacements by non-proteinogenic conformationally constrained amino acids. These approaches combine bioinformatics, high performance computations and mathematics with synthetic organic chemistry. The resulting artificially engineered self-organizing molecular scale structures take advantage of nature's nanobiology toolkit and at the same time improve on it, since their new targeted function differs from that optimized by evolution. The major challenge facing nanobiology is to be able to exercise fine control over the performance of these target-specific molecular machines.

Correlated evolutionary pressure at interacting transcription factors and DNA response elements can guide the rational engineering of DNA binding specificity

Understanding the molecular mechanisms of the specific interaction between transcription factor proteins and DNA is key to comprehend the regulation of gene expression and to develop technologies to engineer transcription factors. Thus far, although there have been several attempts to elucidate protein-DNA interaction through amino acid-base recognition codes, sequence based profiles, or physical models of interaction, the greatest successes in engineering DNA binding specificity remain experimental. Here we present the first systematic evidence of correlated evolutionary pressure at interacting amino acid residues and DNA base-pairs in transcription factors, and show that it can be used to rationally engineer DNA binding specificity. The correlation is between the relative evolutionary importance of protein residues and DNA bases, measured, respectively, in terms of the Evolutionary Trace (ET) rank and information entropy. The evolutionarily most important residues interact with the most conserved base-pairs within the response element while residues of least importance interact with the most variable base-pairs. The correlation averages 0.74 over 12 unrelated families of transcriptional regulators, including nuclear hormone receptors, basic helix-loop-helix, ETS- and homeo-domain family. To test the predictive power of this correlation, we targeted a mutational swap of top-ranked ET residues in a transcription factor, LRH-1. This redirects LRH-1 binding as predicted and showed that, in this case, evolutionary importance and binding specificity are coupled sufficiently strongly for the Evolutionary Trace to guide the computational design of DNA binding specificity. This establishes the existence of evolutionary importance correlation at protein-DNA interfaces, and demonstrates that it is a useful principle for the rational engineering of binding specificity.