Using protein language models for protein interaction hot spot prediction with limited data.

BACKGROUND: Protein language models, inspired by the success of large language models in deciphering human language, have emerged as powerful tools for unraveling the intricate code of life inscribed within protein sequences. They have gained significant attention for their promising applications across various areas, including the sequence-based prediction of secondary and tertiary protein structure, the discovery of new functional protein sequences/folds, and the assessment of mutational impact on protein fitness. However, their utility in learning to predict protein residue properties based on scant datasets, such as protein-protein interaction (PPI)-hotspots whose mutations significantly impair PPIs, remained unclear. Here, we explore the feasibility of using protein language-learned representations as features for machine learning to predict PPI-hotspots using a dataset containing 414 experimentally confirmed PPI-hotspots and 504 PPI-nonhot spots. RESULTS: Our findings showcase the capacity of unsupervised learning with protein language models in capturing critical functional attributes of protein residues derived from the evolutionary information encoded within amino acid sequences. We show that methods relying on protein language models can compete with methods employing sequence and structure-based features to predict PPI-hotspots from the free protein structure. We observed an optimal number of features for model precision, suggesting a balance between information and overfitting. CONCLUSIONS: This study underscores the potential of transformer-based protein language models to extract critical knowledge from sparse datasets, exemplified here by the challenging realm of predicting PPI-hotspots. These models offer a cost-effective and time-efficient alternative to traditional experimental methods for predicting certain residue properties. However, the challenge of explaining why specific features are important for determining certain residue properties remains.

Allosteric coupling between transmembrane segment 4 and the selectivity filter of TALK1 potassium channels regulates their gating by extracellular pH.

Opening of two-pore domain K+ channels (K2Ps) is regulated by various external cues, such as pH, membrane tension, or temperature, which allosterically modulate the selectivity filter (SF) gate. However, how these cues cause conformational changes in the SF of some K2P channels remains unclear. Herein, we investigate the mechanisms by which extracellular pH affects gating in an alkaline-activated K2P channel, TALK1, using electrophysiology and molecular dynamics (MD) simulations. We show that R233, located at the N-terminal end of transmembrane segment 4, is the primary pHo sensor. This residue distally regulates the orientation of the carbonyl group at the S1 potassium-binding site through an interacting network composed of residues on transmembrane segment 4, the pore helix domain 1, and the SF. Moreover, in the presence of divalent cations, we found the acidic pH-activated R233E mutant recapitulates the network interactions of protonated R233. Intriguingly, our data further suggested stochastic coupling between R233 and the SF gate, which can be described by an allosteric gating model. We propose that this allosteric model could predict the hybrid pH sensitivity in heterodimeric channels with alkaline-activated and acidic-activated K2P subunits.

PPI-HotspotDB: Database of Protein-Protein Interaction Hot Spots.

Single-point mutations of certain residues (so-called hot spots) impair/disrupt protein-protein interactions (PPIs), leading to pathogenesis and drug resistance. Conventionally, a PPI-hot spot is identified when its replacement decreased the binding free energy significantly, generally by ≥2 kcal/mol. The relatively few mutations with such a significant binding free energy drop limited the number of distinct PPI-hot spots. By defining PPI-hot spots based on mutations that have been manually curated in UniProtKB to significantly impair/disrupt PPIs in addition to binding free energy changes, we have greatly expanded the number of distinct PPI-hot spots by an order of magnitude. These experimentally determined PPI-hot spots along with available structures have been collected in a database called PPI-HotspotDB. We have applied the PPI-HotspotDB to create a nonredundant benchmark, PPI-Hotspot+PDBBM, for assessing methods to predict PPI-hot spots using the free structure as input. PPI-HotspotDB will benefit the design of mutagenesis experiments and development of PPI-hot spot prediction methods. The database and benchmark are freely available at https://ppihotspot.limlab.dnsalias.org.

Factors allowing small monovalent Li+ to displace Ca2+ in proteins.

Because Li+ and Ca2+ differ in both charge and size, the possibility that monovalent Li+ could dislodge the bulkier, divalent Ca2+ in Ca2+ proteins had not been considered. However, our recent density functional theory/continuum dielectric calculations predicted that Li+ could displace the native Ca2+ from the C2 domain of cytosolic PKCα/γ. This would reduce electrostatic interactions between the Li+-bound C2 domain and the membrane, consistent with experimental studies showing that Li+ can inhibit the translocation of cytoplasmic PKC to membranes. Besides the trinuclear Ca2+-site in the PKCα/γ C2 domain, it is not known whether other Ca2+-sites in human proteins may be susceptible to Li+ substitution. Furthermore, it is unclear what factors determine the outcome of the competition between divalent Ca2+ and monovalent Li+. Here we show that the net charge of residues in the first and second coordination shell is a key determinant of the selectivity for divalent Ca2+ over monovalent Li+ in proteins: neutral/anionic Ca2+-carboxylate sites are protected against Li+ attack. They are further protected by outer-shell Asp-/Glu- and the protein matrix rigidifying the Ca2+-site or limiting water entry. In contrast, buried, cationic Ca2+-sites surrounded by Arg+/Lys+, which are found in the C2 domains of PKCα/γ, as well as certain synaptotagmins, are prone to Li+ attack.

Influence of solution ionic strength on the stabilities of M20 loop conformations in apo E. coli dihydrofolate reductase.

Interactions among ions and their specific interactions with macromolecular solutes are known to play a central role in biomolecular stability. However, similar effects in the conformational stability of protein loops that play functional roles, such as binding ligands, proteins, and DNA/RNA molecules, remain relatively unexplored. A well-characterized enzyme that has such a functional loop is Escherichia coli dihydrofolate reductase (ecDHFR), whose so-called M20 loop has been observed in three ordered conformations in crystal structures. To explore how solution ionic strengths may affect the M20 loop conformation, we proposed a reaction coordinate that could quantitatively describe the loop conformation and used it to classify the loop conformations in representative ecDHFR x-ray structures crystallized in varying ionic strengths. The Protein Data Bank survey indicates that at ionic strengths (I) below the intracellular ion concentration-derived ionic strength in E. coli (I ≤ 0.237M), the ecDHFR M20 loop tends to adopt open/closed conformations, and rarely an occluded loop state, but when I is >0.237M, the loop tends to adopt closed/occluded conformations. Distance-dependent electrostatic potentials around the most mobile M20 loop region from molecular dynamics simulations of ecDHFR in equilibrated CaCl2 solutions of varying ionic strengths show that high ionic strengths (I = 0.75/1.5M) can preferentially stabilize the loop in closed/occluded conformations. These results nicely correlate with conformations derived from ecDHFR structures crystallized in varying ionic strengths. Altogether, our results suggest caution in linking M20 loop conformations derived from crystal structures solved at ionic strengths beyond that tolerated by E. coli to the ecDHFR function.

How the Local Environment of Functional Sites Regulates Protein Function.

Proteins form complex biological machineries whose functions in the cell are highly regulated at both the cellular and molecular levels. Cellular regulation of protein functions involves differential gene expressions, post-translation modifications, and signaling cascades. Molecular regulation, on the other hand, involves tuning an optimal local protein environment for the functional site. Precisely how a protein achieves such an optimal environment around a given functional site is not well understood. Herein, by surveying the literature, we first summarize the various reported strategies used by certain proteins to ensure their correct functioning. We then formulate three key physicochemical factors for regulating a protein's functional site, namely, (i) its immediate interactions, (ii) its solvent accessibility, and (iii) its conformational flexibility. We illustrate how these factors are applied to regulate the functions of free/metal-bound Cys and Zn sites in proteins.

Free and Bound Therapeutic Lithium in Brain Signaling.

Lithium, a first-line therapy for bipolar disorder, is effective in preventing suicide and new depressive/manic episodes. Yet, how this beguilingly simple monocation with only two electrons could yield such profound therapeutic effects remains unclear. An in-depth understanding of lithium's mechanisms of actions would help one to develop better treatments limiting its adverse side effects and repurpose lithium for treating traumatic brain injury and chronic neurodegenerative diseases. In this Account, we begin with a comparison of the physicochemical properties of Li+ and its key native rivals, Na+ and Mg2+, to provide physical grounds for their competition in protein binding sites. Next, we review the abnormal signaling pathways and proteins found in bipolar patients, who generally have abnormally high intracellular Na+ and Ca2+ concentrations, high G-protein levels, and hyperactive phosphatidylinositol signaling and glycogen synthase kinase-3ß (GSK3ß) activity. We briefly summarize experimental findings on how lithium, at therapeutic doses, modulates these abnormal signaling pathways and proteins. Following this survey, we address the following aspects of lithium's therapeutic actions: (1) Can Li+ displace Na+ from the allosteric Na+-binding sites in neurotransmitter transporters and G-protein coupled receptors (GPCRs); if so, how would this affect the host protein's function? (2) Why are certain Mg2+-dependent enzymes targeted by Li+? (3) How does Li+ binding to Mg2+-bound ATP/GTP (denoted as NTP) in solution affect the cofactor's conformation and subsequent recognition by the host protein? (4) How do NTP-Mg-Li complexes modulate the properties of the respective cellular receptors and signal-transducing proteins? We show that Li+ may displace Na+ from allosteric Na+-binding sites in certain GPCRs and stabilize inactive conformations, preventing these receptors from relaying signal to the respective G-proteins. It may also displace Mg2+ in enzymes containing highly cationic Mg2+-binding sites such as GSK3ß, but not in enzymes containing Mg2+-binding sites with low or zero charge. We further show that Li+ binding to Mg2+-NTP in water does not alter the NTP conformation, which is locked by all three phosphates binding to Mg2+. However, bound lithium in the form of [NTP-Mg-Li]2- dianions can activate or inhibit the host protein depending on the NTP-binding pocket's shape, which determines the metal-binding mode: The ATP-binding pocket's shape in the P2X receptor is complementary to the native ATP-Mg solution conformation and nicely fits [ATP-Mg-Li]2-. However, since the ATP ßγ phosphates bind Li+, bimetallic [ATP-Mg-Li]2- may be more resistant to hydrolysis than the native cofactor, enabling ATP to reside longer in the binding site and elicit a prolonged P2X response. In contrast, the elongated GTP-binding pockets in G-proteins allow only two GTP phosphates to bind Mg2+, so the GTP conformation is no longer "triply-locked". Consequently, Li+ binding to GTP-Mg can significantly alter the native cofactor's structure, lowering the activated G-protein level, thus attenuating hyperactive G-protein-mediated signaling in bipolar patients. In summary, we have presented a larger "connected" picture of lithium's diverse effects based on its competition as a free monocation with native cations or as a phosphate-bound polyanionic complex modulating the host protein function.

Soluble klotho regulates TRPC6 calcium signaling via lipid rafts, independent of the FGFR-FGF23 pathway.

Soluble klotho (sKlotho), the shed ectodomain of α-klotho, protects the heart by down-regulating transient receptor potential canonical isoform 6 (TRPC6)-mediated calcium signaling. Binding to α2-3-sialyllactose moiety of gangliosides in lipid rafts and inhibition of raft-dependent signaling underlies the mechanism. A recent 3-Å X-ray structure of sKlotho in complex with fibroblast growth factor receptor (FGFR) and fibroblast growth factor 23 (FGF23) indicates that its ß6α6 loop might block access to the proposed binding site for α2-3-sialyllactose. It was concluded that sKlotho only functions in complex with FGFR and FGF23 and that sKlotho's pleiotropic effects all depend on FGF23. Here, we report that sKlotho can inhibit TRPC6 channels expressed in cells lacking endogenous FGFRs. Structural modeling and molecular docking show that a repositioned ß6α6 loop allows sKlotho to bind α2-3-sialyllactose. Molecular dynamic simulations further show the α2-3-sialyllactose-bound sKlotho complex to be stable. Domains mimicking sKlotho's sialic acid-recognizing activity inhibit TRPC6. The results strongly support the hypothesis that sKlotho can exert effects independent of FGF23 and FGFR.-Wright, J. D., An, S.-W., Xie, J., Lim, C., Huang, C.-L. Soluble klotho regulates TRPC6 calcium signaling via lipid rafts, independent of the FGFR-FGF23 pathway.

Benchmarking polarizable and non-polarizable force fields for Ca2+-peptides against a comprehensive QM dataset.

Explicit description of atomic polarizability is critical for the accurate treatment of inter-molecular interactions by force fields (FFs) in molecular dynamics (MD) simulations aiming to investigate complex electrostatic environments such as metal-binding sites of metalloproteins. Several models exist to describe key monovalent and divalent cations interacting with proteins. Many of these models have been developed from ion-amino-acid interactions and/or aqueous-phase data on cation solvation. The transferability of these models to cation-protein interactions remains uncertain. Herein, we assess the accuracy of existing FFs by their abilities to reproduce hierarchies of thousands of Ca2+-dipeptide interaction energies based on density-functional theory calculations. We find that the Drude polarizable FF, prior to any parameterization, better approximates the QM interaction energies than any of the non-polarizable FFs. Nevertheless, it required improvement in order to address polarization catastrophes where, at short Ca2+-carboxylate distances, the Drude particle of oxygen overlaps with the divalent cation. To ameliorate this, we identified those conformational properties that produced the poorest prediction of interaction energies to reduce the parameter space for optimization. We then optimized the selected cation-peptide parameters using Boltzmann-weighted fitting and evaluated the resulting parameters in MD simulations of the N-lobe of calmodulin. We also parameterized and evaluated the CTPOL FF, which incorporates charge-transfer and polarization effects in additive FFs. This work shows how QM-driven parameter development, followed by testing in condensed-phase simulations, may yield FFs that can accurately capture the structure and dynamics of ion-protein interactions.

Factors Governing the Different Functions of Zn2+-Sites with Identical Ligands in Proteins.

In Zn-proteins, structural Zn-sites are mostly Cys-rich lined by two or more Cys residues, whereas catalytic Zn-sites usually contain His or Asp/Glu residues and a water molecule. Here, we reveal many examples outside this trend with Zn2+ bound to ligands commonly found in both structural and catalytic Zn-sites, namely, Zn-CC(C/H)x (x = D, E, or H2O) sites. We show that these atypical Zn-sites are found in all known life forms (i.e., eukaryotes, bacteria, archaea, and viruses) and can serve structural roles in some proteins but catalytic roles in others. By calculating the physical properties of these atypical Zn-binding sites, we elucidate why Zn-CC(C/H)x sites of the same composition can serve structural and catalytic roles in proteins. Furthermore, we found new sequence/structural motifs characteristic of catalytic Zn-CCHw sites and provide guidelines to predict the structural/catalytic role of atypical Zn-CC(C/H)x sites of unknown function. We discuss how our results could help to design inhibitors targeting catalytic Zn-CC(C/H) H2O sites.

The Zinc Linchpin Motif in the DNA Repair Glycosylase MUTYH: Identifying the Zn2+ Ligands and Roles in Damage Recognition and Repair.

The DNA base excision repair (BER) glycosylase MUTYH prevents DNA mutations by catalyzing adenine (A) excision from inappropriately formed 8-oxoguanine (8-oxoG):A mismatches. The importance of this mutation suppression activity in tumor suppressor genes is underscored by the association of inherited variants of MUTYH with colorectal polyposis in a hereditary colorectal cancer syndrome known as MUTYH-associated polyposis, or MAP. Many of the MAP variants encompass amino acid changes that occur at positions surrounding the two-metal cofactor-binding sites of MUTYH. One of these cofactors, found in nearly all MUTYH orthologs, is a [4Fe-4S]2+ cluster coordinated by four Cys residues located in the N-terminal catalytic domain. We recently uncovered a second functionally relevant metal cofactor site present only in higher eukaryotic MUTYH orthologs: a Zn2+ ion coordinated by three Cys residues located within the extended interdomain connector (IDC) region of MUTYH that connects the N-terminal adenine excision and C-terminal 8-oxoG recognition domains. In this work, we identified a candidate for the fourth Zn2+ coordinating ligand using a combination of bioinformatics and computational modeling. In addition, using in vitro enzyme activity assays, fluorescence polarization DNA binding assays, circular dichroism spectroscopy, and cell-based rifampicin resistance assays, the functional impact of reduced Zn2+ chelation was evaluated. Taken together, these results illustrate the critical role that the "Zn2+ linchpin motif" plays in MUTYH repair activity by providing for proper engagement of the functional domains on the 8-oxoG:A mismatch required for base excision catalysis. The functional importance of the Zn2+ linchpin also suggests that adjacent MAP variants or exposure to environmental chemicals may compromise Zn2+ coordination, and ability of MUTYH to prevent disease.

Modeled structural basis for the recognition of α2-3-sialyllactose by soluble Klotho.

Soluble Klotho (sKlotho) is the shed ectodomain of antiaging membrane Klotho that contains 2 extracellular domains KL1 and KL2, each of which shares sequence homology to glycosyl hydrolases. sKlotho elicits pleiotropic cellular responses with a poorly understood mechanism of action. Notably, in injury settings, sKlotho confers cardiac and renal protection by down-regulating calcium-permeable transient receptor potential canonical type isoform 6 (TRPC6) channels in cardiomyocytes and glomerular podocytes. Inhibition of PI3K-dependent exocytosis of TRPC6 is thought to be the underlying mechanism, and recent studies showed that sKlotho interacts with α2-3-sialyllactose-containing gangliosides enriched in lipid rafts to inhibit raft-dependent PI3K signaling. However, the structural basis for binding and recognition of α2-3-sialyllactose by sKlotho is unknown. Using homology modeling followed by docking, we identified key protein residues in the KL1 domain that are likely involved in binding sialyllactose. Functional experiments based on the ability of Klotho to down-regulate TRPC6 channel activity confirm the importance of these residues. Furthermore, KL1 domain binds α2-3-sialyllactose, down-regulates TRPC6 channels, and exerts protection against stress-induced cardiac hypertrophy in mice. Our results support the notion that sialogangliosides and lipid rafts are membrane receptors for sKlotho and that the KL1 domain is sufficient for the tested biologic activities. These findings can help guide the design of a simpler Klotho mimetic.-Wright, J. D., An, S.-W., Xie, J., Yoon, J., Nischan, N., Kohler, J. J., Oliver, N., Lim, C., Huang, C.-L. Modeled structural basis for the recognition of α2-3-sialyllactose by soluble Klotho.

How Pb2+ Binds and Modulates Properties of Ca2+-Signaling Proteins.

Abiogenic lead (Pb2+), present in the environment in elevated levels due to human activities, has detrimental effects on human health. Metal-binding sites in proteins have been identified as primary targets for lead substitution resulting in malfunction of the host protein. Although Pb2+ is known to be a potent competitor of Ca2+ in protein binding sites, why/how Pb2+ can compete with Ca2+ in proteins remains unclear, raising multiple outstanding questions, including the following: (1) What are the physicochemical factors governing the competition between Pb2+ and Ca2+? (2) Which Ca2+-binding sites in terms of the structure, composition, overall charge, flexibility, and solvent exposure are the most likely targets for Pb2+ attack? Using density functional theory combined with polarizable continuum model calculations, we address these questions by studying the thermodynamic outcome of the competition between Pb2+ and Ca2+ in various model Ca2+-binding sites, including those modeling voltage-gated calcium channel selectivity filters and EF-hand and non-EF-hand Ca2+-binding sites. The results, which are in good agreement with experiment, reveal that the metal site's flexibility and number of amino acid ligands dictate the outcome of the competition between Pb2+ and Ca2+: If the Ca2+-binding site is relatively rigid and crowded with protein ligands, then Pb2+, upon binding, preserves the native metal-binding site geometry and at low concentrations, can act as an activator of the host protein. If the Ca2+-binding site is flexible and consists of only a few protein ligands, then Pb2+ can displace Ca2+ and deform the native metal-binding site geometry, resulting in protein malfunction.

How First Shell-Second Shell Interactions and Metal Substitution Modulate Protein Function.

Hydrogen bonds to metal-ligands in proteins play a vital role in biological function. They help to stabilize/protect the metal complex and enhance metal-binding affinity/specificity, enzyme-substrate recognition, and enzyme activation. Yet, knowledge of the preferred hydrogen-bonding partners of metal ligands in different metalloproteins is lacking. Using well-calibrated methods, we have determined the preferred hydrogen-bonding partners of Cys- bound to native Zn2+ or xenobiotic Cd2+ in Zn-fingers of varying net charge and solvent accessibility as well as the key factors underlying the observed preference. We show how secondary hydrogen-bonding interactions with metal-bound thiolates might exert a significant impact on Zn2+→Cd2+ substitution and thus protein function. Knowing which Zn-fingers may be vulnerable to structural deformation by Cd2+ is important since this would lead to their inactivation, which might impair cell growth, differentiation, cell-cycle control, and DNA repair.

Factors governing when a metal-bound water is deprotonated in proteins.

Understanding when a metal-bound water molecule in a protein is deprotonated is important as this affects the charge distribution in the metal-binding/enzyme active site and thus their interactions, the enzyme mechanism, and inhibitor design. The protonation state of the metal-bound water molecule at a given pH depends on its pKa value, which in turn depends on the properties of the cation, its ligands, and the protein environment. Here, we reveal how and the extent to which (i) the first-shell composition (type, charge, and number of ligands), (ii) the metal site's immediate surroundings (first-shellsecond-shell hydrogen-bonding interactions, metal-ligand distance constraints, and ligand-binding mode) and (iii) the protein architecture and coupled solvent interactions (long-range electrostatic interactions and solvent exposure of the site) affect the Zn2+-bound water pKa. The results, which are consistent with available experimental pKa values of Zn2+-bound water, provide guidelines to predict when Zn2+-bound water would likely be deprotonated at physiological pH.

An efficient protocol for computing the pKa of Zn-bound water.

At a given pH, whether a metal-bound water molecule is deprotonated or not can be determined if the pKa of the metal-bound water molecule (denoted pKw) is known. Although protocols/tools to predict the protonation states of titratable amino acid residues and small molecules have been developed, an efficient and accurate method to predict the absolute pKw values of metal complexes is lacking. Here, we present calibrated methods for optimizing the geometries and computing the absolute pKw values of a wide range of Zn2+-complexes containing protein-like ligating groups. We tested 18 different geometry-optimization methods on 19 ultra high-resolution structures of Zn2+ complexes of varying coordination numbers and ligating atoms and 98 methods in reproducing 36 experimental pKw values of diverse Zn2+ complexes in the absence and presence of explicit water molecules. The results underscore the importance of estimating the Zn2+-bound water/hydroxide solvation properly, whereas correcting for the basis set superposition error was not found to be important. The protocol presented can be used to (i) evaluate the geometries of the different Zn2+-sites found in proteins and (ii) to dissect the individual contributions of the various factors modulating the pKw in Zn2+-sites found in proteins. Predicting absolute pKw values in various environments with efficiency and accuracy will indicate when a Zn2+-bound water molecule is deprotonated, thus providing physical insight into the mechanisms of enzyme-catalyzed reactions and the design of drug candidates that can displace a metal-bound water molecule.

Influence of the Selectivity Filter Properties on Proton Selectivity in the Influenza A M2 Channel.

The homotetrameric M2 proton channel of influenza A plays a crucial role in the viral life cycle and is thus an important therapeutic target. It selectively conducts protons against a background of other competing cations whose concentrations are up to a million times greater than the proton concentration. Its selectivity is largely determined by a constricted region of its open pore known as the selectivity filter, which is lined by four absolutely conserved histidines. While the mechanism of proton transport through the channel has been studied, the physical principles underlying the selectivity for protons over other cations in the channel's His4 selectivity filter remain elusive. Furthermore, it is not known if proton selectivity absolutely requires all four histidines with two of the four histidines protonated and if other titratable amino acid residues in lieu of the histidines could bind protons and how they affect proton selectivity. Here, we elucidate how the competition between protons and rival cations such as Na+ depends on the selectivity filter's (1) histidine protonation state, (2) solvent exposure, (3) oligomeric state (the number of protein chains and thus the number of His ligands), and (4) ligand composition by evaluating the free energies for replacing monovalent Na+ with H3O+ in various model selectivity filters. We show that tetrameric His4 filters are more proton-selective than their trimeric His3 counterparts, and a dicationic His4 filter where two of the four histidines are protonated is more proton-selective than tetrameric filters with other charge states/composition (different combinations of His protonation states or different metal-ligating ligands). The [His4]2+ filter achieves proton selectivity by providing suboptimal binding conditions for rival cations such as Na+, which prefers a neutral or negatively charged filter instead of a dicationic one, and three rather than four ligands with oxygen-ligating atoms.

Using an Old Drug to Target a New Drug Site: Application of Disulfiram to Target the Zn-Site in HCV NS5A Protein.

In viral proteins, labile Zn-sites, where Zn(2+) is crucial for maintaining the native protein structure but the Zn-bound cysteines are reactive, are promising drug targets. Here, we aim to (i) identify labile Zn-sites in viral proteins using guidelines established from our previous work and (ii) assess if clinically safe Zn-ejecting agents could eject Zn(2+) from the predicted target site and thus inhibit viral replication. As proof-of-concept, we identified a labile Zn-site in the hepatitis C virus (HCV) NS5A protein and showed that the antialcoholism drug, disulfiram, could inhibit HCV replication to a similar extent as the clinically used antiviral agent, ribavirin. The discovery of a novel viral target and a new role for disulfiram in inhibiting HCV replication will enhance the therapeutic armamentarium against HCV. The strategy presented can also be applied to identify labile sites in other bacterial or viral proteins that can be targeted by disulfiram or other clinically safe Zn-ejectors.

Factors controlling the selectivity for Na(+) over Mg(2+) in sodium transporters and enzymes.

Na(+) and Mg(2+) play different crucial roles in biological systems. Both cations are present in comparable amounts in the cytosol, but how monovalent Na(+) can compete with the divalent Mg(2+), which can better accept charge from negatively charged ligands, in sodium transporters/enzymes has not been investigated. Hence, it is not clear how Na(+) and Mg(2+)-binding sites have evolved to discriminate the "right" cation among non-cognate ones from the surrounding milieu and the physical basis governing the selectivity for Na(+) over Mg(2+). The results, which are consistent with available experimental data, reveal that in proteins, the selectivity for Na(+) over Mg(2+) in sodium-binding sites stem mainly from the size, charge, and charge-accepting ability differences between Na(+) and Mg(2+). A protein could achieve Na(+) selectivity by (i) reducing the number of metal-ligating ligands, (ii) maintaining an optimal balance of different ligating-strength ligands whose interactions in the metal-binding site would favor Na(+) over rival mono/divalent cations, (iii) increasing the solvent exposure of the metal-binding site, or (iv) increasing binding site rigidity forcing Mg(2+) to adopt the coordination distances/geometry of Na(+). Sodium-binding proteins use one or more of these factors to achieve Na(+) selectivity.

Identifying RNA-binding residues based on evolutionary conserved structural and energetic features.

Increasing numbers of protein structures are solved each year, but many of these structures belong to proteins whose sequences are homologous to sequences in the Protein Data Bank. Nevertheless, the structures of homologous proteins belonging to the same family contain useful information because functionally important residues are expected to preserve physico-chemical, structural and energetic features. This information forms the basis of our method, which detects RNA-binding residues of a given RNA-binding protein as those residues that preserve physico-chemical, structural and energetic features in its homologs. Tests on 81 RNA-bound and 35 RNA-free protein structures showed that our method yields a higher fraction of true RNA-binding residues (higher precision) than two structure-based and two sequence-based machine-learning methods. Because the method requires no training data set and has no parameters, its precision does not degrade when applied to 'novel' protein sequences unlike methods that are parameterized for a given training data set. It was used to predict the 'unknown' RNA-binding residues in the C-terminal RNA-binding domain of human CPEB3. The two predicted residues, F430 and F474, were experimentally verified to bind RNA, in particular F430, whose mutation to alanine or asparagine nearly abolished RNA binding. The method has been implemented in a webserver called DR_bind1, which is freely available with no login requirement at http://drbind.limlab.ibms.sinica.edu.tw.