Distance dependence and salt sensitivity of pairwise, coulombic interactions in a protein

NMR structures of small molecules bound to a model of a CUG RNA repeat expansion

Trinucleotide repeat expansions fold into long, stable hairpins and cause a variety of incurable RNA gain-of-function diseases such as Huntington's disease, the myotonic dystrophies, and spinocerebellar ataxias. One approach for treating these diseases is to bind small molecules to these structured RNAs. Both Huntington's disease-like 2 (HDL2) and myotonic dystrophy type 1 (DM1) are caused by a r(CUG) repeat expansion, or r(CUG)exp. The RNA folds into a hairpin structure with a periodic array of 1 × 1 nucleotide UU loops (5'CUG/3'GUC; where the underlined nucleotides indicate the Us in the internal loop) that sequester various RNA-binding proteins (RBPs) and hence the source of its gain-of-function. Here, we report nuclear magnetic resonance (NMR)-refined structures of single 5'CUG/3'GUC motifs in complex with three different small molecules, a di-guandinobenzoate (1), a derivative of 1 where the guanidino groups have been exchanged for imidazole (2), and a quinoline with improved drug-like properties (3). These structures were determined using NMR spectroscopy and simulated annealing with restrained molecular dynamics (MD). Compounds 1, 2, and 3 formed stacking and hydrogen bonding interactions with the 5'CUG/3'GUC motif. Compound 3 also formed van der Waals interactions with the internal loop. The global structure of each RNA-small molecule complexes retains an A-form conformation, while the internal loops are still dynamic but to a lesser extent compared to the unbound form. These results aid our understanding of ligand-RNA interactions and enable structure-based design of small molecules with improved binding affinity for and biological activity against r(CUG)exp. As the first ever reported structures of a r(CUG) repeat bound to ligands, these structures can enable virtual screening campaigns combined with machine learning assisted de novo design.

NMR structures and magnetic force spectroscopy studies of small molecules binding to models of an RNA CAG repeat expansion

RNA repeat expansions fold into stable structures and cause microsatellite diseases such as Huntington's disease (HD), myotonic dystrophy type 1 (DM1), and spinocerebellar ataxias (SCAs). The trinucleotide expansion of r(CAG), or r(CAG)exp, causes both HD and SCA3, and the RNA's toxicity has been traced to its translation into polyglutamine (polyQ; HD) as well as aberrant pre-mRNA alternative splicing (SCA3 and HD). Previously, a small molecule, 1, was discovered that binds to r(CAG)exp and rescues aberrant pre-mRNA splicing in patient-derived fibroblasts by freeing proteins bound to the repeats. Here, we report the structures of single r(CAG) repeat motif (5'CAG/3'GAC where the underlined adenosines form a 1×1 nucleotide internal loop) in complex with 1 and two other small molecules via nuclear magnetic resonance (NMR) spectroscopy combined with simulated annealing. Compound 2 was designed based on the structure of 1 bound to the RNA while 3 was selected as a diverse chemical scaffold. The three complexes, although adopting different 3D binding pockets upon ligand binding, are stabilized by a combination of stacking interactions with the internal loop's closing GC base pairs, hydrogen bonds, and van der Waals interactions. Molecular dynamics (MD) simulations performed with NMR-derived restraints show that the RNA is stretched and bent upon ligand binding with significant changes in propeller-twist and opening. Compound 3 has a distinct mode of binding by insertion into the helix, displacing one of the loop nucleotides into the major groove and affording a rod-like shape binding pocket. In contrast, 1 and 2 are groove binders. A series of single molecule magnetic force spectroscopy studies provide a mechanistic explanation for how bioactive compounds might rescue disease-associated cellular phenotypes.

NMR structures of small molecules bound to a model of an RNA CUG repeat expansion

Trinucleotide repeat expansions fold into long, stable hairpins and cause a variety of incurable RNA gain-of-function diseases such as Huntington's disease, the myotonic dystrophies, and spinocerebellar ataxias. One approach for treating these diseases is to bind small molecules to the structured RNAs. Both Huntington's disease-like 2 (HDL2) and myotonic dystrophy type 1 (DM1) are caused by a r(CUG) repeat expansion, or r(CUG)exp. The RNA folds into a hairpin structure with a periodic array of 1×1 nucleotide UU loops (5'CUG/3'GUC; where the underlined nucleotides indicate the Us in the internal loop) that sequester various RNA-binding proteins (RBP) and hence the source of its gain-of-function. Here, we report NMR-refined structures of single 5'CUG/3'GUC motifs in complex with three different small molecules, a di-guandinobenzoate (1), a derivative of 1 where the guanidino groups have been exchanged for imidazole (2), and a quinoline with improved drug-like properties (3). These structures were determined using nuclear magnetic resonance (NMR) spectroscopy and simulated annealing with restrained molecular dynamics (MD). Compounds 1, 2, and 3 formed stacking and hydrogen bonding interactions with the 5'CUG/3'GUC motif. Compound 3 also formed van der Waals interactions with the internal loop. The global structure of each RNA-small molecule complexes retains an A-form conformation, while the internal loops are still dynamic but to a lesser extent compared to the unbound form. These results aid our understanding of ligand-RNA interactions and enable structure-based design of small molecules with improved binding affinity for and biological activity against r(CUG)exp. As the first ever reported structures of RNA r(CUG) repeats bound to ligands, these structures can enable virtual screening campaigns combined with machine learning assisted de novo design.

Elucidating the Protonation State of the γ-Secretase Catalytic Dyad

γ-Secretase (GS) is an intramembrane aspartyl protease that participates in the sequential cleavage of C99 to generate different isoforms of the amyloid-β (Aβ) peptides that are associated with the development of Alzheimer's disease. Due to its importance in the proteolytic processing of C99 by GS, we performed pH replica exchange molecular dynamics (pH-REMD) simulations of GS in its apo and substrate-bound forms to sample the protonation states of the catalytic dyad. We found that the catalytic dyad is deprotonated at physiological pH in our apo form, but the presence of the substrate at the active site displaces its monoprotonated state toward physiological pH. Our results show that Asp257 acts as the general base and Asp385 as the general acid during the cleavage mechanism. We identified different amino acids such as Lys265, Arg269, and the PAL motif interacting with the catalytic dyad and promoting changes in its acid-base behavior. Finally, we also found a significant pK_a shift of Glu280 related to the internalization of TM6-CT in the GS-apo form. Our study provides critical mechanistic insight into the GS mechanism and the basis for future research on the genesis of Aβ peptides and the development of Alzheimer's disease.

Analysis of Protein-Protein Interactions for Intermolecular Bond Prediction

Protein-protein interactions often involve a complex system of intermolecular interactions between residues and atoms at the binding site. A comprehensive exploration of these interactions can help reveal key residues involved in protein-protein recognition that are not obvious using other protein analysis techniques. This paper presents and extends DiffBond, a novel method for identifying and classifying intermolecular bonds while applying standard definitions of bonds in chemical literature to explain protein interactions. DiffBond predicted intermolecular bonds from four protein complexes: Barnase-Barstar, Rap1a-raf, SMAD2-SMAD4, and a subset of complexes formed from three-finger toxins and nAChRs. Based on validation through manual literature search and through comparison of two protein complexes from the SKEMPI dataset, DiffBond was able to identify intermolecular ionic bonds and hydrogen bonds with high precision and recall, and identify salt bridges with high precision. DiffBond predictions on bond existence were also strongly correlated with observations of Gibbs free energy change and electrostatic complementarity in mutational experiments. DiffBond can be a powerful tool for predicting and characterizing influential residues in protein-protein interactions, and its predictions can support research in mutational experiments and drug design.

Inclusion of More Physics Leads to Less Data: Learning the Interaction Energy as a Function of Electron Deformation Density with Limited Training Data

Machine learning (ML) approaches to predicting quantum mechanical (QM) properties have made great strides toward achieving the computational chemist's holy grail of structure-based property prediction. In contrast to direct ML methods, which encode a molecule with only structural information, in this work, we show that QM descriptors improve ML predictions of dimer interaction energy, both in terms of accuracy and data efficiency, by incorporating electronic information into the descriptor. We present the electron deformation density interaction energy machine learning (EDDIE-ML) model, which predicts the interaction energy as a function of Hartree-Fock electron deformation density. We compare its performance with leading direct ML schemes and modern DFT methods for the prediction of interaction energies for dimers of varying charge type, size, and intermolecular separation. Under a low-data regime, EDDIE-ML outperforms other direct ML schemes and is the only model readily transferrable to larger, more complex systems including base pair trimers and porous cages. The underlying physical connection between the density and interaction energy enables EDDIE-ML to reach an accuracy comparable to modern DFT functionals in fewer training data points compared to other ML methods.

Mechanistic Characterization of the Pharmacological Profile of HS-731, a Peripherally Acting Opioid Analgesic, at the µ-, δ-, κ-Opioid and Nociceptin Receptors

Accumulated preclinical and clinical data show that peripheral restricted opioids provide pain relief with reduced side effects. The peripherally acting opioid analgesic HS-731 is a potent dual μ-/δ-opioid receptor (MOR/DOR) full agonist, and a weak, partial agonist at the κ-opioid receptor (KOR). However, its binding mode at the opioid receptors remains elusive. Here, we present a comprehensive in silico evaluation of HS-731 binding at all opioid receptors. We provide insights into dynamic interaction patterns explaining the different binding and activity of HS-731 on the opioid receptors. For this purpose, we conducted docking, performed molecular dynamics (MD) simulations and generated dynamic pharmacophores (dynophores). Our results highlight two residues important for HS-731 recognition at the classical opioid receptors (MOR, DOR and KOR), particular the conserved residue 5.39 (K) and the non-conserved residue 6.58 (MOR: K, DOR: W and KOR: E). Furthermore, we assume a salt bridge between the transmembrane helices (TM) 5 and 6 via K2275.39 and E2976.58 to be responsible for the partial agonism of HS-731 at the KOR. Additionally, we experimentally demonstrated the absence of affinity of HS-731 to the nociceptin/orphanin FQ peptide (NOP) receptor. We consider the morphinan phenol Y1303.33 responsible for this affinity lack. Y1303.33 points deep into the NOP receptor binding pocket preventing HS-731 binding to the orthosteric binding pocket. These findings provide significant structural insights into HS-731 interaction pattern with the opioid receptors that are important for understanding the pharmacology of this peripheral opioid analgesic.

DiffBond: A Method for Predicting Intermolecular Bond Formation

Many tools that explore models of protein complexes are also able to analyze interactions between specific residues and atoms. A comprehensive exploration of these interactions can often uncover aspects of protein-protein recognition that are not obvious using other protein analysis techniques. This paper describes DiffBond, a novel method for searching for intermolecular interactions between protein complexes while differentiating between three different types of interaction: hydrogen bonds, ionic bonds, and salt bridges. DiffBond incorporates textbook definitions of these three interactions while contending with uncertainties that are inherent in computational models of interacting proteins. We used it to examine the barnase-barstar, Rap1a-raf, and Smad2-Smad4 complexes, as well as a subset of protein complexes formed between three-finger toxins and nAChRs. Based on electrostatic interactions established by previous experimental studies, DiffBond was able to identify ionic and hydrogen bonds with high precision and recall, and identify salt bridges with high precision. In combination with other electrostatic analysis methods, DiffBond can be a useful tool in helping predict influential amino acids in protein-protein interactions and characterizing the type of interaction.

Osmolytes Can Destabilize Proteins in Cells by Modulating Electrostatics and Quinary Interactions

Although numerous in vitro studies have shown that osmolytes are capable of stabilizing proteins, their effect on protein folding in vivo has been less understood. In this work, we investigated the effect of osmolytes, including glycerol, sorbitol, betaine, and taurine, on the folding of a protein GB3 variant in E. coli cells using NMR spectroscopy. 400 mM osmolytes were added to E. coli cells; only glycerol stabilizes the folded protein, whereas betaine and taurine considerably destabilize the protein through modulating folding and unfolding rates. Further investigation indicates that betaine and taurine can enhance the quinary interaction between the protein and cellular environment and manifestly weaken the electrostatic attraction in protein salt bridges. The combination of the two factors causes destabilization of the protein in E. coli cells. These factors counteract the preferential exclusion mechanism that is adopted by osmolytes to stabilize proteins.

Measuring how two proteins affect each other's net charge in a crowded environment

Theory predicts that the net charge (Z) of a protein can be altered by the net charge of a neighboring protein as the two approach one another below the Debye length. This type of charge regulation suggests that a protein's charge and perhaps function might be affected by neighboring proteins without direct binding. Charge regulation during protein crowding has never been directly measured due to analytical challenges. Here, we show that lysine specific protein crosslinkers (NHS ester-Staudinger pairs) can be used to mimic crowding by linking two non-interacting proteins at a maximal distance of ~7.9 Å. The net charge of the regioisomeric dimers and preceding monomers can then be determined with lysine-acyl "protein charge ladders" and capillary electrophoresis. As a proof of concept, we covalently linked myoglobin (Z_monomer  = -0.43 ± 0.01) and α-lactalbumin (Z_monomer  = -4.63 ± 0.05). Amide hydrogen/deuterium exchange and circular dichroism spectroscopy demonstrated that crosslinking did not significantly alter the structure of either protein or result in direct binding (thus mimicking crowding). Ultimately, capillary electrophoretic analysis of the dimeric charge ladder detected a change in charge of ΔZ = -0.04 ± 0.09 upon crowding by this pair (Z_dimer  = -5.10 ± 0.07). These small values of ΔZ are not necessarily general to protein crowding (qualitatively or quantitatively) but will vary per protein size, charge, and solvent conditions.

Normal modes analysis and surface electrostatics of haemagglutinin proteins as fingerprints for high pathogenic type A influenza viruses

Background

Type A influenza viruses circulate and spread among wild birds and mostly consist of low pathogenic strains. However, fast genome variation timely results in the insurgence of high pathogenic strains, which when infecting poultry birds may cause a million deaths and strong commercial damage. More importantly, the host shift may concern these viruses and sustained human-to-human transmission may result in a dangerous pandemic outbreak. Therefore, fingerprints specific to either low or high pathogenic strains may represent a very important tool for global surveillance.

Results

We combined Normal Modes Analysis and surface electrostatic analysis of a mixed strain dataset of influenza A virus haemagglutinins from high and low pathogenic strains in order to infer specific fingerprints. Normal Modes Analysis sorted the strains in two different, homogeneous clusters; sorting was independent of clades and specific instead to high vs low pathogenicity. A deeper analysis of fluctuations and flexibility regions unveiled a special role for the 110-helix region. Specific sorting was confirmed by surface electrostatics analysis, which further allowed to focus on regions and mechanisms possibly crucial to the low-to-high transition.

Conclusions

Evidence from previous work demonstrated that changes in surface electrostatics are associated with the evolution and spreading of avian influenza A virus clades, and seemingly involved also in the avian to mammalian host shift. This work shows that a combination of electrostatics and Normal Modes Analysis can also identify fingerprints specific to high and low pathogenicity. The possibility to predict which specific mutations may result in a shift to high pathogenicity may help in surveillance and vaccine development.

Unique Features of a New Baeyer–Villiger Monooxygenase from a Halophilic Archaeon

Type I Baeyer–Villiger monooxygenases (BVMOs) are flavin-dependent monooxygenases that catalyze the oxidation of ketones to esters or lactones, a reaction otherwise performed in chemical processes by employing hazardous and toxic peracids. Even though various BVMOs are extensively studied for their promising role in industrial biotechnology, there is still a demand for enzymes that are able to retain activity at high saline concentrations. To this aim, and based on comparative in silico analyses, we cloned HtBVMO from the extremely halophilic archaeon Haloterrigena turkmenica DSM 5511. When expressed in standard mesophilic cell factories, proteins adapted to hypersaline environments often behave similarly to intrinsically disordered polypeptides. Nevertheless, we managed to express HtBVMO in Escherichia coli and could purify it as active enzyme. The enzyme was characterized in terms of its salt-dependent activity and resistance to some water–organic-solvent mixtures. Although HtBVMO does not seem suitable for industrial applications, it provides a peculiar example of an alkalophilic and halophilic BVMO characterized by an extremely negative charge. Insights into the behavior and structural properties of such salt-requiring may contribute to more efficient strategies for engineering the tuned stability and solubility of existing BVMOs.

Predicting RNA-Metal Ion Binding with Ion Dehydration Effects

Metal ions play essential roles in nucleic acids folding and stability. The interaction between metal ions and nucleic acids can be highly complicated because of the interplay between various effects such as ion correlation, fluctuation, and dehydration. These effects may be particularly important for multivalent ions such as Mg²⁺ ions. Previous efforts to model ion correlation and fluctuation effects led to the development of the Monte Carlo tightly bound ion model. Here, by incorporating ion hydration/dehydration effects into the Monte Carlo tightly bound ion model, we develop a, to our knowledge, new approach to predict ion binding. The new model enables predictions for not only the number of bound ions but also the three-dimensional spatial distribution of the bound ions. Furthermore, the new model reveals several intriguing features for the bound ions such as the mutual enhancement/inhibition in ion binding between the fully hydrated (diffuse) ions, the outer-shell dehydrated ions, and the inner-shell dehydrated ions and novel features for the monovalent-divalent ion interplay due to the hydration effect.

Linking computation and experiments to study the role of charge-charge interactions in protein folding and stability

Over the past two decades there has been an increase in appreciation for the role of surface charge-charge interactions in protein folding and stability. The perception shifted from the belief that charge-charge interactions are not important for protein folding and stability to the near quantitative understanding of how these interactions shape the folding energy landscape. This led to the ability of computational approaches to rationally redesign surface charge-charge interactions to modulate thermodynamic properties of proteins. Here we summarize our progress in understanding the role of charge-charge interactions for protein stability using examples drawn from my own laboratory and touch upon unanswered questions.

Rational Selection of the 3D Structure of Biomacromolecules for Molecular Docking Studies on the Mechanism of Endocrine Disruptor Action

Molecular modeling has become an essential tool in predicting and simulating endocrine disrupting effects of chemicals. A key prerequisite for successful application of molecular modeling lies in the correctness of 3D structure for biomacromolecules to be simulated. To date, there are several databases that can provide the experimentally-determined 3D structures. However, commonly, there are many challenges or disadvantageous factors, e.g., (a) lots of 3D structures for a given biomacromolecular target in the protein database; (b) the quality variability for those structures; (c) belonging to different species; (d) mutant amino acid residue in key positions, and so on. Once an inappropriate 3D structure of a target biomacromolecule was selected in molecular modeling, the accuracy and scientific nature of the modeling results could be inevitably affected. In this article, based on literature survey and an analysis of the 3D structure characterization of biomacromolecular targets belonging to the endocrine system in protein databases, six principles were proposed to guide the selection of the appropriate 3D structure of biomacromolecules. The principles include considering the species diversity, the mechanism of action, whether there are mutant amino acid residues, whether the number of protein chains is correct, the degree of structural similarity between the ligand in 3D structure and the target compounds, and other factors, e.g., the experimental pH conditions of the structure determined process and resolution.

Determinants of the pKa values of ionizable residues in an intrinsically disordered protein

Intrinsically disordered proteins (IDPs) are prevalent in eukaryotes; in humans, they are often associated with diseases. The protein NUPR1 is a multifunctional IDP involved in the development and progression of pancreatic cancer; therefore, it constitutes a target for drug design. In an effort to contribute to the understanding of the conformational features of NUPR1 and to provide clues on amino acid interactions in disordered states of proteins, we measured the pKa values of all its acidic groups (aspartic and glutamic residues, and backbone C terminus) by using NMR spectroscopy at low (100 mM) and high (500 mM) NaCl concentration. At low ionic strength, the pKa values were similar to those reported for random-coil models, except for Glu18 and Asp19, suggesting electrostatic interactions around these residues. Molecular modelling and simulation indicate an additional, significant role of nearby proline residues in determining the polypeptide conformational features and water accessibility in the region around Glu18, modulating the titration properties of these amino acids. In the other acidic residues of NUPR1, the small deviations of pKa values (compared to those expected for a random-coil) are likely due to electrostatic interactions with charged adjacent residues, which should be reduced at high NaCl concentrations. In fact, at high ionic strength, the pKa values of the aspartic residues were similar to those in a random coil, but there were still small differences for those of glutamic acids, probably due to hydrogen-bond formation. The overall findings suggest that local interactions and hydrophobic effects play a major role in determining the electrostatic features of NUPR1, whereas long-range charge contributions appear to be of lesser importance.

Benchmarking pKa Prediction Methods for Residues in Proteins

Methods for estimation of pKa values of residues in proteins were tested on a set of benchmark proteins with experimentally known pKa values. The benchmark set includes 80 different residues (20 each for Asp, Glu, Lys, and His), half of which consists of significantly variant cases (ΔpKa ≥ 1 pKa unit from the amino acid in solution). The method introduced by Case and co-workers [J. Am. Chem. Soc. 2004, 126, 4167-4180], referred to as the molecular dynamics/generalized-Born/thermodynamic integration (MD/GB/TI) technique, gives a root-mean-square deviation (rmsd) of 1.4 pKa units on the benchmark set. The use of explicit waters in the immediate region surrounding the residue was shown to generally reduce high errors for this method. Longer simulation time was also shown to increase the accuracy of this method. The empirical approach developed by Jensen and co-workers [Proteins 2005, 61, 704-721], PROPKA, also gives an overall rmsd of 1.4 pKa units and is more or less accurate based on residue type-the method does very well for Lys and Glu, but less so for Asp and His. Likewise, the absolute deviation is quite similar for the two methods-5.2 for PROPKA and 5.1 for MD/GB/TI. A comparison of these results with several prediction methods from the literature is presented. The error in pKa prediction is analyzed as a function of variation of the pKa from that in water and the solvent accessible surface area (SASA) of the residue. A case study of the catalytic lysine residue in 2-deoxyribose-5-phosphate aldolase (DERA) is also presented.

Characterization of the Dielectric Constant in the Trichoderma reesei Cel7B Active Site

An attempt is made to evaluate the dielectric constant of the Trichoderma reesei Cel7B active site. Through kinetic measurements, the pKa value of the catalytic acid E201 is determined. Mutations (away from E201) with net charge changes are introduced to perturb the E201 pKa. It is shown that the mutation with a +1 charge change (including G225R, G230R, and A335R) decreases the pKa of E201, whereas the mutation with a -1 charge change (including Q149E, A222D, G225D, and G230D) increases the pKa. This effect is consistent with the electrostatic interaction between the changed charge and the E201 side chain. The fitting of the experimental data yields an apparent dielectric constant of 25-80. Molecular dynamics simulations with explicit water molecules indicate that the high solvent accessibility of the active site contributes largely to the high dielectric constant. ONIOM calculations show that high dielectric constant benefits the catalysis through decreasing the energy of the transition state relative to that of the enzyme substrate complex.

Modeling of interaction between cytochrome c and the WD domains of Apaf-1: bifurcated salt bridges underlying apoptosome assembly

BACKGROUND

Binding of cytochrome c, released from the damaged mitochondria, to the apoptotic protease activating factor 1 (Apaf-1) is a key event in the apoptotic signaling cascade. The binding triggers a major domain rearrangement in Apaf-1, which leads to oligomerization of Apaf-1/cytochrome c complexes into an apoptosome. Despite the availability of crystal structures of cytochrome c and Apaf-1 and cryo-electron microscopy models of the entire apoptosome, the binding mode of cytochrome c to Apaf-1, as well as the nature of the amino acid residues of Apaf-1 involved remain obscure.

RESULTS

We investigated the interaction between cytochrome c and Apaf-1 by combining several modeling approaches. We have applied protein-protein docking and energy minimization, evaluated the resulting models of the Apaf-1/cytochrome c complex, and carried out a further analysis by means of molecular dynamics simulations. We ended up with a single model structure where all the lysine residues of cytochrome c that are known as functionally-relevant were involved in forming salt bridges with acidic residues of Apaf-1. This model has revealed three distinctive bifurcated salt bridges, each involving a single lysine residue of cytochrome c and two neighboring acidic resides of Apaf-1. Salt bridge-forming amino acids of Apaf-1 showed a clear evolutionary pattern within Metazoa, with pairs of acidic residues of Apaf-1, involved in bifurcated salt bridges, reaching their highest numbers in the sequences of vertebrates, in which the cytochrome c-mediated mechanism of apoptosome formation seems to be typical.

CONCLUSIONS

The reported model of an Apaf-1/cytochrome c complex provides insights in the nature of protein-protein interactions which are hard to observe in crystallographic or electron microscopy studies. Bifurcated salt bridges can be expected to be stronger than simple salt bridges, and their formation might promote the conformational change of Apaf-1, leading to the formation of an apoptosome. Combination of structural and sequence analyses provides hints on the evolution of the cytochrome c-mediated apoptosis.

Modeling of interaction between cytochrome c and the WD domains of Apaf-1: bifurcated salt bridges underlying apoptosome assembly

Background

Binding of cytochrome c, released from the damaged mitochondria, to the apoptotic protease activating factor 1 (Apaf-1) is a key event in the apoptotic signaling cascade. The binding triggers a major domain rearrangement in Apaf-1, which leads to oligomerization of Apaf-1/cytochrome c complexes into an apoptosome. Despite the availability of crystal structures of cytochrome c and Apaf-1 and cryo-electron microscopy models of the entire apoptosome, the binding mode of cytochrome c to Apaf-1, as well as the nature of the amino acid residues of Apaf-1 involved remain obscure.

Results

We investigated the interaction between cytochrome c and Apaf-1 by combining several modeling approaches. We have applied protein-protein docking and energy minimization, evaluated the resulting models of the Apaf-1/cytochrome c complex, and carried out a further analysis by means of molecular dynamics simulations. We ended up with a single model structure where all the lysine residues of cytochrome c that are known as functionally-relevant were involved in forming salt bridges with acidic residues of Apaf-1. This model has revealed three distinctive bifurcated salt bridges, each involving a single lysine residue of cytochrome c and two neighboring acidic resides of Apaf-1. Salt bridge-forming amino acids of Apaf-1 showed a clear evolutionary pattern within Metazoa, with pairs of acidic residues of Apaf-1, involved in bifurcated salt bridges, reaching their highest numbers in the sequences of vertebrates, in which the cytochrome c-mediated mechanism of apoptosome formation seems to be typical.

Conclusions

The reported model of an Apaf-1/cytochrome c complex provides insights in the nature of protein-protein interactions which are hard to observe in crystallographic or electron microscopy studies. Bifurcated salt bridges can be expected to be stronger than simple salt bridges, and their formation might promote the conformational change of Apaf-1, leading to the formation of an apoptosome. Combination of structural and sequence analyses provides hints on the evolution of the cytochrome c-mediated apoptosis.

Reviewers

This article was reviewed by Andrei L. Osterman, Narayanaswamy Srinivasan, Igor N. Berezovsky, and Gerrit Vriend (nominated by Martijn Huynen).

Electronic supplementary material

The online version of this article (doi:10.1186/s13062-015-0059-4) contains supplementary material, which is available to authorized users.

Buried ionizable networks are an ancient hallmark of G protein-coupled receptor activation

Seven-transmembrane receptors (7TMRs) have evolved in prokaryotes and eukaryotes over hundreds of millions of years. Comparative structural analysis suggests that these receptors may share a remote evolutionary origin, despite their lack of sequence similarity. Here we used structure-based computations to compare 221 7TMRs from all domains of life. Unexpectedly, we discovered that these receptors contain spatially conserved networks of buried ionizable groups. In microbial 7TMRs these networks are used to pump ions across the cell membrane in response to light. In animal 7TMRs, which include light- and ligand-activated G protein-coupled receptors (GPCRs), homologous networks were found to be characteristic of activated receptor conformations. These networks are likely relevant to receptor function because they connect the ligand-binding pocket of the receptor to the nucleotide-binding pocket of the G protein. We propose that agonist and G protein binding facilitate the formation of these electrostatic networks and promote important structural rearrangements such as the displacement of transmembrane helix-6. We anticipate that robust classification of activated GPCR structures will aid the identification of ligands that target activated GPCR structural states.

Mode of binding of the cyclic agonist peptide TC14012 to CXCR7: identification of receptor and compound determinants

The chemokine receptor CXCR7 is an atypical CXCL12 receptor that, as opposed to the classical CXCL12 receptor CXCR4, signals preferentially via the β-arrestin pathway and does not mediate chemotaxis. We previously reported that the cyclic peptide TC14012, a potent CXCR4 antagonist, also engaged CXCR7, albeit with lower potency. Surprisingly, the compound activated the CXCR7-arrestin pathway. The reason underlying the opposite effects of TC14012 on CXCR4 and CXCR7, and the mode of binding of TC14012 to CXCR7, remained unclear. The mode of binding of TC14012 to CXCR4 is known from cocrystallization of its analogue CVX15 with CXCR4. We here report the the mode of binding of TC14012 to CXCR7 by combining the use of compound analogues, receptor mutants, and molecular modeling. We find that the mode of binding of TC14012 to CXCR7 is indeed similar to that of CVX15 to CXCR4, with compound positions Arg2 and Arg14 engaging CXCR7 key residues D179(4.60) (on the tip of transmembrane domain 4) and D275(6.58) (on the tip of transmembrane domain 6), respectively. Interestingly, the TC14012 parent compound T140 is not a CXCR7 agonist, because of conformational constraints in its pharmacophore, which in TC14012 are relieved through C-terminal amidation. However, an engineered salt bridge between the CXCR7 ECL2 substitution R197D and compound residue Arg1 permitted T140 agonism by repositioning the compound in the binding pocket. In conclusion, our results show that the opposite effect of TC14012 on CXCR4 and CXCR7 is not explained by different binding modes. Rather, engagement of the interface between transmembrane domains and extracellular loops readily triggers CXCR7, but not CXCR4, activation.

Comparative structural analysis of haemagglutinin proteins from type A influenza viruses: conserved and variable features

Background

Genome variation is very high in influenza A viruses. However, viral evolution and spreading is strongly influenced by immunogenic features and capacity to bind host cells, depending in turn on the two major capsidic proteins. Therefore, such viruses are classified based on haemagglutinin and neuraminidase types, e.g. H5N1. Current analyses of viral evolution are based on serological and primary sequence comparison; however, comparative structural analysis of capsidic proteins can provide functional insights on surface regions possibly crucial to antigenicity and cell binding.

Results

We performed extensive structural comparison of influenza virus haemagglutinins and of their domains and subregions to investigate type- and/or domain-specific variation. We found that structural closeness and primary sequence similarity are not always tightly related; moreover, type-specific features could be inferred when comparing surface properties of haemagglutinin subregions, monomers and trimers, in terms of electrostatics and hydropathy. Focusing on H5N1, we found that variation at the receptor binding domain surface intriguingly relates to branching of still circulating clades from those ones that are no longer circulating.

Conclusions

Evidence from this work suggests that integrating phylogenetic and serological analyses by extensive structural comparison can help in understanding the ‘functional evolution’ of viral surface determinants. In particular, variation in electrostatic and hydropathy patches can provide molecular evolution markers: intriguing surface charge redistribution characterizing the haemagglutinin receptor binding domains from circulating H5N1 clades 2 and 7 might have contributed to antigenic escape hence to their evolutionary success and spreading.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-014-0363-5) contains supplementary material, which is available to authorized users.

Compensatory mutations occur within the electrostatic interaction range of deleterious mutations in protein structure

A compensatory mutation (CM) counter balances lethal effects of a deleterious mutation (DM), ensuring the persistence of both through natural selection. However, little is known about the biological aspects of CMs those restore the structural alterations of proteins caused by slightly DMs. Here, by analyzing the evolution of the UDP-glycosyltransferase 73B4 protein among monocot-dicot plants, we investigate the occurrence of CMs around slightly DMs in 3D space. Our results illustrate that CMs exhibit significantly higher tendency to occur within the range of electrostatic interaction around the slightly DMs, compared to occurring randomly in the protein.

Molecular interactions and residues involved in force generation in the T4 viral DNA packaging motor

Many viruses utilize molecular motors to package their genomes into preformed capsids. A striking feature of these motors is their ability to generate large forces to drive DNA translocation against entropic, electrostatic, and bending forces resisting DNA confinement. A model based on recently resolved structures of the bacteriophage T4 motor protein gp17 suggests that this motor generates large forces by undergoing a conformational change from an extended to a compact state. This transition is proposed to be driven by electrostatic interactions between complementarily charged residues across the interface between the N- and C-terminal domains of gp17. Here we use atomistic molecular dynamics simulations to investigate in detail the molecular interactions and residues involved in such a compaction transition of gp17. We find that although electrostatic interactions between charged residues contribute significantly to the overall free energy change of compaction, interactions mediated by the uncharged residues are equally if not more important. We identify five charged residues and six uncharged residues at the interface that play a dominant role in the compaction transition and also reveal salt bridging, van der Waals, and solvent hydrogen-bonding interactions mediated by these residues in stabilizing the compact form of gp17. The formation of a salt bridge between Glu309 and Arg494 is found to be particularly crucial, consistent with experiments showing complete abrogation in packaging upon Glu309Lys mutation. The computed contributions of several other residues are also found to correlate well with single-molecule measurements of impairments in DNA translocation activity caused by site-directed mutations.

Protein apparent dielectric constant and its temperature dependence from remote chemical shift effects

A NMR protocol is introduced that permits accurate measurement of minute, remote chemical shift perturbations (CSPs), caused by a mutation-induced change in the electric field. Using protein GB3 as a model system, ¹H^N CSPs in K19A and K19E mutants can be fitted to small changes in the electric field at distal sites in the protein using the Buckingham equation, yielding an apparent dielectric constant ε_a of 8.6 ± 0.8 at 298 K. These CSPs, and their derived ε_a value, scale strongly with temperature. For example, CSPs at 313 K are about ∼30% smaller than those at 278 K, corresponding to an effective ε_a value of about 7.3 at 278 K and 10.5 at 313 K. Molecular dynamics simulations in explicit solvent indicate that solvent water makes a significant contribution to ε_a.

Design of net-charged abc-type collagen heterotrimers

Net-negatively-charged heterospecific A:B:C collagen peptide heterotrimers were designed using an automated computational approach. The design algorithm considers both target stability and the energy gap between the target states and misfolded competing states. Structural characterization indicates the net-negative charge balance on the new designs enhances the specificity of the target state at the expense of its stability.

Role of electrostatic potential in the in silico prediction of molecular bioactivation and mutagenesis

Electrostatic potential (ESP) is a useful physicochemical property of a molecule that provides insights into inter- and intramolecular associations, as well as prediction of likely sites of electrophilic and nucleophilic metabolic attack. Knowledge of sites of metabolic attack is of paramount importance in DMPK research since drugs frequently fail in clinical trials due to the formation of bioactivated metabolites which are often difficult to measure experimentally due to their reactive nature and relatively short half-lives. Computational chemistry methods have proven invaluable in recent years as a means to predict and study bioactivated metabolites without the need for chemical syntheses, or testing on experimental animals. Additional molecular properties (heat of formation, heat of solvation and E(LUMO) - E(HOMO)) are discussed in this paper as complementary indicators of the behavior of metabolites in vivo. Five diverse examples are presented (acetaminophen, aniline/phenylamines, imidacloprid, nefazodone and vinyl chloride) which illustrate the utility of this multidimensional approach in predicting bioactivation, and in each case the predicted data agreed with experimental data described in the scientific literature. A further example of the usefulness of calculating ESP, in combination with the molecular properties mentioned above, is provided by an examination of the use of these parameters in providing an explanation for the sites of nucleophilic attack of the nucleic acid cytosine. Exploration of sites of nucleophilic attack of nucleic acids is important as adducts of DNA have the potential to result in mutagenesis.

Rapid calculation of protein pKa values using Rosetta

We developed a Rosetta-based Monte Carlo method to calculate the pK(a) values of protein residues that commonly exhibit variable protonation states (Asp, Glu, Lys, His, and Tyr). We tested the technique by calculating pK(a) values for 264 residues from 34 proteins. The standard Rosetta score function, which is independent of any environmental conditions, failed to capture pK(a) shifts. After incorporating a Coulomb electrostatic potential and optimizing the solvation reference energies for pK(a) calculations, we employed a method that allowed side-chain flexibility and achieved a root mean-square deviation (RMSD) of 0.83 from experimental values (0.68 after discounting 11 predictions with an error over 2 pH units). Additional degrees of side-chain conformational freedom for the proximal residues facilitated the capture of charge-charge interactions in a few cases, resulting in an overall RMSD of 0.85 pH units. The addition of backbone flexibility increased the overall RMSD to 0.93 pH units but improved relative pK(a) predictions for proximal catalytic residues. The method also captures large pK(a) shifts of lysine and some glutamate point mutations in staphylococcal nuclease. Thus, a simple and fast method based on the Rosetta score function and limited conformational sampling produces pK(a) values that will be useful when rapid estimation is essential, such as in docking, design, and folding.

Using DelPhi capabilities to mimic protein's conformational reorganization with amino acid specific dielectric constants

Many molecular events are associated with small or large conformational changes occurring in the corresponding proteins. Modeling such changes is a challenge and requires significant amount of computing time. From point of view of electrostatics, these changes can be viewed as a reorganization of local charges and dipoles in response to the changes of the electrostatic field, if the cause is insertion or deletion of a charged amino acid. Here we report a large scale investigation of modeling the changes of the folding energy due to single mutations involving charged group. This allows the changes of the folding energy to be considered mostly electrostatics in origin and to be calculated with DelPhi assigning residue-specific value of the internal dielectric constant of protein. The predicted energy changes are benchmarked against experimentally measured changes of the folding energy on a set of 257 single mutations. The best fit between experimental values and predicted changes is used to find out the effective value of the internal dielectric constant for each type of amino acid. The predicted folding free energy changes with the optimal, amino acid specific, dielectric constants are within RMSD=0.86 kcal/mol from experimentally measured changes.

Sequence recombination improves target specificity in a redesigned collagen peptide abc-type heterotrimer

Stability of the collagen triple helix is largely governed by its imino acid content, namely the occurrence of proline and 4R-hydroxyproline at the X and Y positions, respectively, of the periodic (Gly-X-Y)(n) sequence. Although other amino acids at these positions reduce stability of the triple helix, this can be partially compensated by introducing intermolecular side-chain salt bridges. This approach was previously used to design an abc-type heterotrimer composed of one basic, one acidic, and one neutral imino acid rich chain (Gauba and Hartgerink, J Am Chem Soc 2007;129:15034-15041). In this study, an abc-type heterotrimer was designed to be the most stable species using a sequence recombination strategy that preserved both the amino acid composition and the network of interchain salt bridges of the original design. The target heterotrimer had the highest T(m) of 50 °C, 7 °C greater than the next most stable species. Stability of the heterotrimer decreased with increasing ionic strength, consistent with the role of intermolecular salt bridges in promoting stability. Quantitative meta-analysis of these results and published stability measurements on closely related peptides was used to discriminate the contributions of backbone propensity and side-chain electrostatics to collagen stability.

MCCE analysis of the pKas of introduced buried acids and bases in staphylococcal nuclease

The pK(a)s of 96 acids and bases introduced into buried sites in the staphylococcal nuclease protein (SNase) were calculated using the multiconformation continuum electrostatics (MCCE) program and the results compared with experimental values. The pK(a)s are obtained by Monte Carlo sampling of coupled side chain protonation and position as a function of pH. The dependence of the results on the protein dielectric constant (ε(prot)) in the continuum electrostatics analysis and on the Lennard-Jones non-electrostatics parameters was evaluated. The pK(a)s of the introduced residues have a clear dependence on ε(prot,) whereas native ionizable residues do not. The native residues have electrostatic interactions with other residues in the protein favoring ionization, which are larger than the desolvation penalty favoring the neutral state. Increasing ε(prot) scales both terms, which for these residues leads to small changes in pK(a). The introduced residues have a larger desolvation penalty and negligible interactions with residues in the protein. For these residues, changing ε(prot) has a large influence on the calculated pK(a). An ε(prot) of 8-10 and a Lennard-Jones scaling of 0.25 is best here. The X-ray crystal structures of the mutated proteins are found to provide somewhat better results than calculations carried out on mutations made in silico. Initial relaxation of the in silico mutations by Gromacs and extensive side chain rotamer sampling within MCCE can significantly improve the match with experiment.

Deciphering the role of glucosamine-6-phosphate in the riboswitch action of glmS ribozyme

The GlmS ribozyme is believed to exploit a general acid-base catalytic mechanism in the presence of glucosamine-6-phosphate (GlcN6P) to accelerate self-cleavage by approximately six orders of magnitude. The general acid and general base are not known, and the role of the GlcN6P cofactor is even less well understood. The amine group of GlcN6P has the ability to either accept or donate a proton and could therefore potentially act as an acid or a base. In order to decipher the role of GlcN6P in the self-cleavage of glmS, we have determined the preferred protonation state of the amine group in the wild-type and an inactive G40A mutant using molecular dynamics simulations and free energy calculations. Here we show that, upon binding of GlcN6P to wild-type glmS, the pK(a) of the amine moiety is altered by the active site environment, decreasing by about 2.2 from a solution pK(a) of about 8.2. On the other hand, we show that the pK(a) of the amine group slightly increases to about 8.4 upon binding to the G40A inactive mutant of glmS. These results suggest that GlcN6P acts as a general acid in the self-cleavage of glmS. Upon binding to glmS, GlcN6P can easily release a proton to the 5'-oxygen of G1 during self-cleavage of the backbone phosphodiester bond. However, in the G40A inactive mutant of glmS, the results suggest that the ability of GlcN6P to easily release its proton is diminished, in addition to the possible lack of G40 as an effective base.

Dynamic regulation of fibrinogen: integrin αIIbβ3 binding

This study demonstrates that two orthogonal events regulate integrin αIIbβ3's interactions with fibrinogen, its primary physiological ligand: (1) conformational changes at the αIIb-β3 interface and (2) flexibility in the carboxy terminus of fibrinogen's γ-module. The first postulate was tested by capturing αIIbβ3 on a biosensor and measuring binding by surface plasmon resonance. Binding of fibrinogen to eptifibatide-primed αIIbβ3 was characterized by a k(on) of ~2 × 10(4) L mol(-1) s(-1) and a k(off) of ~8 × 10(-5) s(-1) at 37 °C. In contrast, even at 150 nM fibrinogen, no binding was detected with resting αIIbβ3. Eptifibatide competitively inhibited fibrinogen's interactions with primed αIIbβ3 (K(i) ~0.4 nM), while a synthetic γ-module peptide (HHLGGAKQAGDV) was only weakly inhibitory (K(i) > 10 μM). The second postulate was tested by measuring αIIbβ3's interactions with recombinant fibrinogen, both normal (rFgn) and a deletion mutant lacking the γ-chain AGDV sites (rFgn γΔ408-411). Normal rFgn bound rapidly, tightly, and specifically to primed αIIbβ3; no interaction was detected with rFgn γΔ408-411. Equilibrium and transition-state thermodynamic data indicated that binding of fibrinogen to primed αIIbβ3, while enthalpy-favorable, must overcome an entropy-dominated activation energy barrier. The hypothesis that fibrinogen binding is enthalpy-driven fits with structural data showing that its γ-C peptide and eptifibatide exhibit comparable electrostatic contacts with αIIbβ3's ectodomain. The concept that fibrinogen's αIIbβ3 targeting sequence is intrinsically disordered may explain the entropy penalty that limits its binding rate. In the hemostatic milieu, platelet-platelet interactions may be localized to vascular injury sites because integrins must be activated before they can bind their most abundant ligand.

Predicting protein pK(a) by environment similarity

A statistical method to predict protein pK(a) has been developed by using the 3D structure of a protein and a database of 434 experimental protein pK(a) values. Each pK(a) in the database is associated with a fingerprint that describes the chemical environment around an ionizable residue. A computational tool, MoKaBio, has been developed to identify automatically ionizable residues in a protein, generate fingerprints that describe the chemical environment around such residues, and predict pK(a) from the experimental pK(a) values in the database by using a similarity metric. The method, which retrieved the pK(a) of 429 of the 434 ionizable sites in the database correctly, was crossvalidated by leave-one-out and yielded root mean square error (RMSE) = 0.95, a result that is superior to that obtained by using the Null Model (RMSE 1.07) and other well-established protein pK(a) prediction tools. This novel approach is suitable to rationalize protein pK(a) by comparing the region around the ionizable site with similar regions whose ionizable site pK(a) is known. The pK(a) of residues that have a unique environment not represented in the training set cannot be predicted accurately, however, the method offers the advantage of being trainable to increase its predictive power.

Effective approach for calculations of absolute stability of proteins using focused dielectric constants

The ability to predict the absolute stability of proteins based on their corresponding sequence and structure is a problem of great fundamental and practical importance. In this work, we report an extensive, refinement and validation of our recent approach (Roca et al., FEBS Lett 2007;581:2065-2071) for predicting absolute values of protein stability DeltaG(fold). This approach employs the semimacroscopic protein dipole Langevin dipole method in its linear response approximation version (PDLD/S-LRA) while using the best fitted values of the dielectric constants epsilon'(p) and epsilon'(eff) for the self energy and charge-charge interactions, respectively. The method is validated on a diverse set of 45 proteins. It is found that the best fitted values of both dielectric constants are around 40. However, the self energy of internal residues and the charge-charge interactions of Lys have to be treated with care, using a somewhat lower values of epsilon'(p) and epsilon'(eff). The predictions of DeltaG(fold) reported here, have an average error of only 1.8 kcal/mole compared to the observed values, making our method very promising for estimating protein stability. It also provides valuable insight into the complex electrostatic phenomena taking place in folded proteins.

Challenging a paradigm: theoretical calculations of the protonation state of the Cys25-His159 catalytic diad in free papain

A central mechanistic paradigm of cysteine proteases is that the His-Cys catalytic diad forms an ion-pair NH(+)/S(-) already in the catalytically active free enzyme. Most molecular modeling studies of cysteine proteases refer to this paradigm as their starting point. Nevertheless, several recent kinetics and X-ray crystallography studies of viral and bacterial cysteine proteases depart from the ion-pair mechanism, suggesting general base catalysis. We challenge the postulate of the ion-pair formation in free papain. Applying our QM/SCRF(VS) molecular modeling approach, we analyzed all protonation states of the catalytic diad in free papain and its SMe derivative, comparing the predicted and experimental pK(a) data. We conclude that the His-Cys catalytic diad in free papain is fully protonated, NH(+)/SH. The experimental pK(a) = 8.62 of His159 imidazole in free papain, obtained by NMR-controlled titration and originally interpreted as the NH(+)/S(-) <==> N/S(-) NH(+)/S(-) <==> N/S(-) equilibrium, is now assigned to the NH(+)/SH <==> N/SH NH(+)/SH <==> N/SH equilibrium.

Comparative effects of H+ and Ca2+ on large-conductance Ca2+- and voltage-gated Slo1 K+ channels

Large-conductance Ca(2+)- and voltage-gated Slo1 BK channels are allosterically activated by depolarization and intracellular ligands such as Ca(2+). Of the two high-affinity Ca(2+) sensors present in the channel, the RCK1 sensor also mediates H(+)-dependent activation of the channel. In this study, we examined the comparative mechanisms of the channel activation by Ca(2+) and H(+). Steady-state macroscopic conductance-voltage measurements as well as single-channel openings at negative voltages where voltage-sensor activation is negligible showed that at respective saturating concentrations Ca(2+) is more effective in relative stabilization of the open conformation than H(+). Calculations using the Debye-Hückel formalism suggest that small structural changes in the RCK1 sensor, on the order of few angstroms, may accompany the H(+)-mediated opening of the channel. While the efficacy of H(+) in activation of the channel is less than that of Ca(2+), H(+) more effectively accelerates the activation kinetics when examined at the concentrations equipotent on macroscopic voltage-dependent activation. The RCK1 sensor therefore is capable of transducing the nature of the bound ligand and transmits qualitatively different information to the channel's permeation gate.

The pK(a) values of acidic and basic residues buried at the same internal location in a protein are governed by different factors

The pK(a) values of internal ionizable groups are usually very different from the normal pK(a) values of ionizable groups in water. To examine the molecular determinants of pK(a) values of internal groups, we compared the properties of Lys, Asp, and Glu at internal position 38 in staphylococcal nuclease. Lys38 titrates with a normal or elevated pK(a), whereas Asp38 and Glu38 titrate with elevated pK(a) values of 7.0 and 7.2, respectively. In the structure of the L38K variant, the buried amino group of the Lys38 side chain makes an ion pair with Glu122, whereas in the structure of the L38E variant, the buried carboxyl group of Glu38 interacts with two backbone amides and has several nearby carboxyl oxygen atoms. Previously, we showed that the pK(a) of Lys38 is normal owing to structural reorganization and water penetration concomitant with ionization of the Lys side chain. In contrast, the pK(a) values of Asp38 and Glu38 are perturbed significantly owing to an imbalance between favorable polar interactions and unfavorable contributions from dehydration and from Coulomb interactions with surface carboxylic groups. Their ionization is also coupled to subtle structural reorganization. These results illustrate the complex interplay between local polarity, Coulomb interactions, and structural reorganization as determinants of pK(a) values of internal groups in proteins. This study suggests that improvements to computational methods for pK(a) calculations will require explicit treatment of the conformational reorganization that can occur when internal groups ionize.

A computational approach for the rational design of stable proteins and enzymes: optimization of surface charge-charge interactions

The design of stable proteins and enzymes is not only of particular biotechnological importance, but also addresses some important fundamental questions. While there are a number of different options available for designing or engineering stable proteins, the field of computational design provides fast and universal methods for stabilizing proteins of interest. One of the successful computational design strategies focuses on stabilizing proteins through the optimization of charge-charge interactions on the protein surface. By optimizing surface interactions, it is possible to alleviate some of the challenges that accompany efforts to redesign the protein core. The rational design of surface charge-charge interactions also allows one to optimize only the interactions that are distant from binding sites or active sites, making it possible to increase stability without adversely affecting activity. The optimization of surface charge-charge interactions is discussed in detail along with the experimental evidence to demonstrate that this is a robust and universal approach to designing proteins with enhanced stability.

Characteristics of mutants designed to incorporate a new ion pair into the structure of a cold adapted subtilisin-like serine proteinase

Structural comparisons of VPR, a subtilisin-like serine proteinase from a psychrotrophic Vibrio species and a thermophilic homologue, aqualysin I, have led us to hypothesize about the roles of different residues in the temperature adaptation of the enzymes. Some of these hypotheses are now being examined by analysis of mutants of the enzymes. The selected substitutions are believed to increase the stability of the cold adapted enzyme based on structural analysis of the thermostable structure. We report here on mutants, which were designed to incorporate an ion pair into the structure of VPR. The residues Asp17 and Arg259 are assumed to form an ion pair in aqualysin I. The cold adapted VPR contains Asn (Asn15) and Lys (Lys257) at corresponding sites in its structure. In VPR, Asn 15 is located on a surface loop with its side group pointing towards the side chain of Lys257. By substituting Asn15 by Asp (N15D) it was considered feasible that a salt bridge would form between the oppositely charged groups. To mimic further the putative salt bridge from the thermophile enzyme the corresponding double mutant (N15D/K257R) was also produced. The N15D mutation increased the thermal stability of VPR by approximately 3 degrees C, both in T(50%) and T(m). Addition of the K257R mutation did not however, increase the stability of the double mutant any further. Despite this stabilization of the VPR mutants the catalytic activity (k(cat)) against the substrate Suc-AAPF-NH-Np was increased in the mutants. Molecular dynamics simulations on wild type and the two mutant proteins suggested that indeed a salt bridge was formed in both cases. Furthermore, a truncated form of the N15D mutant (N15DDeltaC) was produced, lacking a 15 residue long C-terminal extended sequence not present in the thermophilic enzyme. In wild type VPR this supposedly moveable, negatively charged arm on the protein molecule might interfere with the new salt bridge introduced as a result of the N15D mutation. Removal of the C-terminal arm improved the thermal stability (T(m) approximately +1.5 degrees C) of the truncated enzyme (VPRDeltaC) as compared to the wild type VPR. Introduction of the N15D substitution into VPRDeltaC improved the thermal stability further by about 3 degrees C, or to about the same extent as in the wild type. However, contrary to what was observed for the wild type, the introduction of the putative salt bridge did not affect the catalytic properties (k(cat)) of the C-terminal truncated enzyme.