Structural modelling pipelines in next generation sequencing projects

Homology Modeling of Transporter Proteins

Membrane transporter proteins are divided into channels/pores and carriers and constitute protein families of physiological and pharmacological importance. Several presently used therapeutic compounds elucidate their effects by targeting membrane transporter proteins, including anti-arrhythmic, anesthetic, antidepressant, anxiolytic and diuretic drugs. The lack of three-dimensional structures of human transporters hampers experimental studies and drug discovery. In this chapter, the use of homology modeling for generating structural models of membrane transporter proteins is reviewed. The increasing number of atomic resolution structures available as templates, together with improvements in methods and algorithms for sequence alignments, secondary structure predictions, and model generation, in addition to the increase in computational power have increased the applicability of homology modeling for generating structural models of transporter proteins. Different pitfalls and hints for template selection, multiple-sequence alignments, generation and optimization, validation of the models, and the use of transporter homology models for structure-based virtual ligand screening are discussed.

Caseinolytic Proteins (Clp) in the Genus Klebsiella: Special Focus on ClpK

Caseinolytic proteins (Clp), which are present in both prokaryotes and eukaryotes, play a major role in cell protein quality control and survival of bacteria in harsh environmental conditions. Recently, a member of this protein family, ClpK was identified in a pathogenic strain of Klebsiella pneumoniae which was responsible for nosocomial infections. ClpK is linked to the thermal stress survival of this pathogen. The genome wide analysis of Clp proteins in Klebsiella spp. indicates that ClpK is present in only 34% of the investigated strains. This suggests that the uptake of the clpk gene is selective and may only be taken up by a pathogen that needs to survive harsh environmental conditions. In silico analyses and molecular dynamic simulations show that ClpK is mainly α-helical and is highly dynamic. ClpK was successfully expressed and purified to homogeneity using affinity and anion exchange chromatography. Biophysical characterization of ClpK showed that it is predominantly alpha-helical, and this is in agreement with in silico analysis of the protein structure. Furthermore, the purified protein is biologically active and hydrolyses ATP in a concentration- dependent manner.

In vivo, in vitro and in silico correlations of four de novo SCN1A missense mutations

Mutations in the SCN1A gene, which encodes for the voltage-gated sodium channel Na_V1.1, cause Dravet syndrome, a severe developmental and epileptic encephalopathy. Genetic testing of this gene is recommended early in life. However, predicting the outcome of de novo missense SCN1A mutations is difficult, since milder epileptic syndromes may also be associated. In this study, we correlated clinical severity with functional in vitro electrophysiological testing of channel activity and bioinformatics prediction of damaging mutational effects. Three patients, bearing the mutations p.Gly177Ala, p.Ser259Arg and p.Glu1923Arg, showed frequent intractable seizures that had started early in life, with cognitive and behavioral deterioration, consistent with classical Dravet phenotypes. These mutations failed to produce measurable sodium currents in a mammalian expression system, indicating complete loss of channel function. A fourth patient, who harbored the mutation p.Met1267Ile, though presenting with seizures early in life, showed lower seizure burden and higher cognitive function, matching borderland Dravet phenotypes. In correlation with this, functional analysis demonstrated the presence of sodium currents, but with partial loss of function. In contrast, six bioinformatics tools for predicting mutational pathogenicity suggested similar impact for all mutations. Likewise, homology modeling of the secondary and tertiary structures failed to reveal misfolding. In conclusion, functional studies using patch clamp are suggested as a prognostic tool, whereby detectable currents imply milder phenotypes and absence of currents indicate an unfavorable prognosis. Future development of automated patch clamp systems will facilitate the inclusion of such functional testing as part of personalized patient diagnostic schemes.

Integrating Molecular Docking and Molecular Dynamics Simulations

Computational methods, applied at the early stages of the drug design process, use current technology to provide valuable insights into the understanding of chemical systems in a virtual manner, complementing experimental analysis. Molecular docking is an in silico method employed to foresee binding modes of small compounds or macromolecules in contact with a receptor and to predict their molecular interactions. Moreover, the methodology opens up the possibility of ranking these compounds according to a hierarchy determined using particular scoring functions. Docking protocols assign many approximations, and most of them lack receptor flexibility. Therefore, the reliability of the resulting protein-ligand complexes is uncertain. The association with the costly but more accurate MD techniques provides significant complementary with docking. MD simulations can be used before docking since a series of "new" and broader protein conformations can be extracted from the processing of the resulting trajectory and employed as targets for docking. They also can be utilized a posteriori to optimize the structures of the final complexes from docking, calculate more detailed interaction energies, and provide information about the ligand binding mechanism. Here, we focus on protocols that offer the docking-MD combination as a logical approach to improving the drug discovery process.

Clinical and Functional Characterization of the Recurrent TUBA1A p.(Arg2His) Mutation

The TUBA1A gene encodes tubulin alpha-1A, a protein that is highly expressed in the fetal brain. Alpha- and beta-tubulin subunits form dimers, which then co-assemble into microtubule polymers: dynamic, scaffold-like structures that perform key functions during neurogenesis, neuronal migration, and cortical organisation. Mutations in TUBA1A have been reported to cause a range of brain malformations. We describe four unrelated patients with the same de novo missense mutation in TUBA1A, c.5G>A, p.(Arg2His), as found by next generation sequencing. Detailed comparison revealed similar brain phenotypes with mild variability. Shared features included developmental delay, microcephaly, hypoplasia of the cerebellar vermis, dysplasia or thinning of the corpus callosum, small pons, and dysmorphic basal ganglia. Two of the patients had bilateral perisylvian polymicrogyria. We examined the effects of the p.(Arg2His) mutation by computer-based protein structure modelling and heterologous expression in HEK-293 cells. The results suggest the mutation subtly impairs microtubule function, potentially by affecting inter-dimer interaction. Based on its sequence context, c.5G>A is likely to be a common recurrent mutation. We propose that the subtle functional effects of p.(Arg2His) may allow for other factors (such as genetic background or environmental conditions) to influence phenotypic outcome, thus explaining the mild variability in clinical manifestations.

De novo mutations in GRIN1 cause extensive bilateral polymicrogyria

Polymicrogyria is a malformation of cortical development. The aetiology of polymicrogyria remains poorly understood. Using whole-exome sequencing we found de novo heterozygous missense GRIN1 mutations in 2 of 57 parent-offspring trios with polymicrogyria. We found nine further de novo missense GRIN1 mutations in additional cortical malformation patients. Shared features in the patients were extensive bilateral polymicrogyria associated with severe developmental delay, postnatal microcephaly, cortical visual impairment and intractable epilepsy. GRIN1 encodes GluN1, the essential subunit of the N-methyl-d-aspartate receptor. The polymicrogyria-associated GRIN1 mutations tended to cluster in the S2 region (part of the ligand-binding domain of GluN1) or the adjacent M3 helix. These regions are rarely mutated in the normal population or in GRIN1 patients without polymicrogyria. Using two-electrode and whole-cell voltage-clamp analysis, we showed that the polymicrogyria-associated GRIN1 mutations significantly alter the in vitro activity of the receptor. Three of the mutations increased agonist potency while one reduced proton inhibition of the receptor. These results are striking because previous GRIN1 mutations have generally caused loss of function, and because N-methyl-d-aspartate receptor agonists have been used for many years to generate animal models of polymicrogyria. Overall, our results expand the phenotypic spectrum associated with GRIN1 mutations and highlight the important role of N-methyl-d-aspartate receptor signalling in the pathogenesis of polymicrogyria.

Coenzyme Q Biosynthesis: Evidence for a Substrate Access Channel in the FAD-Dependent Monooxygenase Coq6

Coq6 is an enzyme involved in the biosynthesis of coenzyme Q, a polyisoprenylated benzoquinone lipid essential to the function of the mitochondrial respiratory chain. In the yeast Saccharomyces cerevisiae, this putative flavin-dependent monooxygenase is proposed to hydroxylate the benzene ring of coenzyme Q (ubiquinone) precursor at position C5. We show here through biochemical studies that Coq6 is a flavoprotein using FAD as a cofactor. Homology models of the Coq6-FAD complex are constructed and studied through molecular dynamics and substrate docking calculations of 3-hexaprenyl-4-hydroxyphenol (4-HP6), a bulky hydrophobic model substrate. We identify a putative access channel for Coq6 in a wild type model and propose in silico mutations positioned at its entrance capable of partially (G248R and L382E single mutations) or completely (a G248R-L382E double-mutation) blocking access to the channel for the substrate. Further in vivo assays support the computational predictions, thus explaining the decreased activities or inactivation of the mutated enzymes. This work provides the first detailed structural information of an important and highly conserved enzyme of ubiquinone biosynthesis.

Recognizable cerebellar dysplasia associated with mutations in multiple tubulin genes

Mutations in alpha- and beta-tubulins are increasingly recognized as a major cause of malformations of cortical development (MCD), typically lissencephaly, pachygyria and polymicrogyria; however, sequencing tubulin genes in large cohorts of MCD patients has detected tubulin mutations in only 1-13%. We identified patients with a highly characteristic cerebellar dysplasia but without lissencephaly, pachygyria and polymicrogyria typically associated with tubulin mutations. Remarkably, in seven of nine patients (78%), targeted sequencing revealed mutations in three different tubulin genes (TUBA1A, TUBB2B and TUBB3), occurring de novo or inherited from a mosaic parent. Careful re-review of the cortical phenotype on brain imaging revealed only an irregular pattern of gyri and sulci, for which we propose the term tubulinopathy-related dysgyria. Basal ganglia (100%) and brainstem dysplasia (80%) were common features. On the basis of in silico structural predictions, the mutations affect amino acids in diverse regions of the alpha-/beta-tubulin heterodimer, including the nucleotide binding pocket. Cell-based assays of tubulin dynamics reveal various effects of the mutations on incorporation into microtubules: TUBB3 p.Glu288Lys and p.Pro357Leu do not incorporate into microtubules at all, whereas TUBB2B p.Gly13Ala shows reduced incorporation and TUBA1A p.Arg214His incorporates fully, but at a slower rate than wild-type. The broad range of effects on microtubule incorporation is at odds with the highly stereotypical clinical phenotype, supporting differential roles for the three tubulin genes involved. Identifying this highly characteristic phenotype is important due to the low recurrence risk compared with the other (recessive) cerebellar dysplasias and the apparent lack of non-neurological medical issues.

The arabidopsis DNA polymerase δ has a role in the deposition of transcriptionally active epigenetic marks, development and flowering

DNA replication is a key process in living organisms. DNA polymerase α (Polα) initiates strand synthesis, which is performed by Polε and Polδ in leading and lagging strands, respectively. Whereas loss of DNA polymerase activity is incompatible with life, viable mutants of Polα and Polε were isolated, allowing the identification of their functions beyond DNA replication. In contrast, no viable mutants in the Polδ polymerase-domain were reported in multicellular organisms. Here we identify such a mutant which is also thermosensitive. Mutant plants were unable to complete development at 28°C, looked normal at 18°C, but displayed increased expression of DNA replication-stress marker genes, homologous recombination and lysine 4 histone 3 trimethylation at the SEPALLATA3 (SEP3) locus at 24°C, which correlated with ectopic expression of SEP3. Surprisingly, high expression of SEP3 in vascular tissue promoted FLOWERING LOCUS T (FT) expression, forming a positive feedback loop with SEP3 and leading to early flowering and curly leaves phenotypes. These results strongly suggest that the DNA polymerase δ is required for the proper establishment of transcriptionally active epigenetic marks and that its failure might affect development by affecting the epigenetic control of master genes.

Three DNA polymerases replicate DNA in Eukaryotes. DNA polymerase α (Polα) initiates strand synthesis, which is performed by Polε and Polδ in leading and lagging strands, respectively. Not only the information encoded in the DNA, but also the inheritance of chromatin states is essential during development. Loss of function mutants in DNA polymerases lead to lethal phenotypes. Hence, hypomorphic alleles are necessary to study their roles beyond DNA replication. Here we identify a thermosensitive mutant of the Polδ in the model plant Arabidopsis thaliana, which bears an aminoacid substitution in the polymerase-domain. The mutants were essentially normal at 18°C but arrested development at 28°C. Interestingly, at 24°C we were able to study the roles of Polδ in epigenetic inheritance and plant development. We observed a tight connection between DNA replication stress and an increase the deposition of transcriptionally active chromatin marks in the SEPALLATA3 (SEP3) locus. Finally, we tested by genetic means that the ectopic expression of SEP3 was indeed the cause of early flowering and the leaf phenotypes by promoting the expression of FLOWERING LOCUS T (FT). These results link Polδ activity to the proper establishment of transcriptionally active epigenetic marks, which then impact the development of multicellular organisms.

Modelling three-dimensional protein structures for applications in drug design

A structural perspective of drug target and anti-target proteins, and their molecular interactions with biologically active molecules, largely advances many areas of drug discovery, including target validation, hit and lead finding and lead optimisation. In the absence of experimental 3D structures, protein structure prediction often offers a suitable alternative to facilitate structure-based studies. This review outlines recent methodical advances in homology modelling, with a focus on those techniques that necessitate consideration of ligand binding. In this context, model quality estimation deserves special attention because the accuracy and reliability of different structure prediction techniques vary considerably, and the quality of a model ultimately determines its usefulness for structure-based drug discovery. Examples of G-protein-coupled receptors (GPCRs) and ADMET-related proteins were selected to illustrate recent progress and current limitations of protein structure prediction. Basic guidelines for good modelling practice are also provided.

Formulation of probabilistic models of protein structure in atomic detail using the reference ratio method

We propose a method to formulate probabilistic models of protein structure in atomic detail, for a given amino acid sequence, based on Bayesian principles, while retaining a close link to physics. We start from two previously developed probabilistic models of protein structure on a local length scale, which concern the dihedral angles in main chain and side chains, respectively. Conceptually, this constitutes a probabilistic and continuous alternative to the use of discrete fragment and rotamer libraries. The local model is combined with a nonlocal model that involves a small number of energy terms according to a physical force field, and some information on the overall secondary structure content. In this initial study we focus on the formulation of the joint model and the evaluation of the use of an energy vector as a descriptor of a protein's nonlocal structure; hence, we derive the parameters of the nonlocal model from the native structure without loss of generality. The local and nonlocal models are combined using the reference ratio method, which is a well-justified probabilistic construction. For evaluation, we use the resulting joint models to predict the structure of four proteins. The results indicate that the proposed method and the probabilistic models show considerable promise for probabilistic protein structure prediction and related applications.