Biomacromolecular quantitative structure–activity relationship (BioQSAR): a proof-of-concept study on the modeling, prediction and interpretation of protein–protein binding affinity

Deep learning-based method for predicting and classifying the binding affinity of protein-protein complexes

Protein-protein interactions (PPIs) play a critical role in various biological processes. Accurately estimating the binding affinity of PPIs is essential for understanding the underlying molecular recognition mechanisms. In this study, we employed a deep learning approach to predict the binding affinity (ΔG) of protein-protein complexes. To this end, we compiled a dataset of 903 protein-protein complexes, each with its corresponding experimental binding affinity, which belong to six functional classes. We extracted 8 to 20 non-redundant features from the sequence information as well as the predicted three-dimensional structures using feature selection methods for each protein functional class. Our method showed an overall mean absolute error of 1.05 kcal/mol and a correlation of 0.79 between experimental and predicted ΔG values. Additionally, we evaluated our model for discriminating high and low affinity protein-protein complexes and it achieved an accuracy of 87% with an F1 score of 0.86 using 10-fold cross-validation on the selected features. Our approach presents an efficient tool for studying PPIs and provides crucial insights into the underlying mechanisms of the molecular recognition process. The web server can be freely accessed at https://web.iitm.ac.in/bioinfo2/DeepPPAPred/index.html.

Cold spots are universal in protein–protein interactions

Proteins interact with each other through binding interfaces that differ greatly in size and physico‐chemical properties. Within the binding interface, a few residues called hot spots contribute the majority of the binding free energy and are hence irreplaceable. In contrast, cold spots are occupied by suboptimal amino acids, providing possibility for affinity enhancement through mutations. In this study, we identify cold spots due to cavities and unfavorable charge interactions in multiple protein–protein interactions (PPIs). For our cold spot analysis, we first use a small affinity database of PPIs with known structures and affinities and then expand our search to nearly 4000 homo‐ and heterodimers in the Protein Data Bank (PDB). We observe that cold spots due to cavities are present in nearly all PPIs unrelated to their binding affinity, while unfavorable charge interactions are relatively rare. We also find that most cold spots are located in the periphery of the binding interface, with high‐affinity complexes showing fewer centrally located colds spots than low‐affinity complexes. A larger number of cold spots is also found in non‐cognate interactions compared to their cognate counterparts. Furthermore, our analysis reveals that cold spots are more frequent in homo‐dimeric complexes compared to hetero‐complexes, likely due to symmetry constraints imposed on sequences of homodimers. Finally, we find that glycines, glutamates, and arginines are the most frequent amino acids appearing at cold spot positions. Our analysis emphasizes the importance of cold spot positions to protein evolution and facilitates protein engineering studies directed at enhancing binding affinity and specificity in a wide range of applications.

Integrated unsupervised-supervised modeling and prediction of protein-peptide affinities at structural level

Cell signal networks are orchestrated directly or indirectly by various peptide-mediated protein-protein interactions, which are normally weak and transient and thus ideal for biological regulation and medicinal intervention. Here, we develop a general-purpose method for modeling and predicting the binding affinities of protein-peptide interactions (PpIs) at the structural level. The method is a hybrid strategy that employs an unsupervised approach to derive a layered PpI atom-residue interaction (ulPpI[a-r]) potential between different protein atom types and peptide residue types from thousands of solved PpI complex structures and then statistically correlates the potential descriptors with experimental affinities (KD values) over hundreds of known PpI samples in a supervised manner to create an integrated unsupervised-supervised PpI affinity (usPpIA) predictor. Although both the ulPpI[a-r] potential and usPpIA predictor can be used to calculate PpI affinities from their complex structures, the latter seems to perform much better than the former, suggesting that the unsupervised potential can be improved substantially with a further correction by supervised statistical learning. We examine the robustness and fault-tolerance of usPpIA predictor when applied to treat the coarse-grained PpI complex structures modeled computationally by sophisticated peptide docking and dynamics simulation. It is revealed that, despite developed solely based on solved structures, the integrated unsupervised-supervised method is also applicable for locally docked structures to reach a quantitative prediction but can only give a qualitative prediction on globally docked structures. The dynamics refinement seems not to change (or improve) the predictive results essentially, although it is computationally expensive and time-consuming relative to peptide docking. We also perform extrapolation of usPpIA predictor to the indirect affinity quantities of HLA-A*0201 binding epitope peptides and NHERF PDZ binding scaffold peptides, consequently resulting in a good and moderate correlation of the predicted KD with experimental IC50 and BLU on the two peptide sets, with Pearson's correlation coefficients Rp = 0.635 and 0.406, respectively.

Structure-based discovery and redesign of TGF-β1 Elbow epitope recognition by its type-II receptor in hypertrophic scarring biotherapy

Transforming growth factor-β1 (TGF-β1) signaling pathway has been implicated in the fibroblast activation of hypertrophic scarring (HS). Previously, we proposed a new biotherapeutic strategy to combat HS by disrupting the intermolecular interaction of TGF-β1 with its cognate type-II receptor (TβR-II). Here, we further demonstrate that the binding site of TGF-β1 to TβR-II is not overlapped with the conformational wrist epitope and linear knuckle epitope that are traditionally recognized as the functional binding sites of bone morphogenetic protein-2 (BMP-2) to its type-II receptor (BMPR-II), which can thus be regarded as a new functional site we called elbow epitope. Structural, energetic, and dynamic investigations reveal that the elbow epitope consists of two sequentially discontinuous, spatially vicinal segments Loop^30-34 and Turn^90-95 ; they cannot work effectively to independently interact with TβR-II. Rational redesign of the epitope is performed using an integrated in silio-in vitro method based on crystal and modeled structure data. In the procedure, the two epitope segments are split from the interface of TGF-β1-TβR-II complex and then connected with each other in a head-to-tail manner by adding a flexible poly-(Gly)n linker between them, thus resulting in a series of combined peptides. We found that the peptide affinity reaches maximum at n = 2, which shares a consistent binding mode with the elbow epitope at native complex interface. The linker of either too long (n > 2) or too short (n < 2) cannot properly place the gap space between the two segments, thus impairing the binding compatibility of designed peptides with TβR-II active site.

Integrated In Silico-In Vitro Identification and Optimization of Bone Morphogenic Protein-2 Armpit Epitope as Its Antagonist Binding Site

Bone morphogenic protein-2 (BMP-2) is the most documented member of BMP family and plays a crucial role in bone formation and growth. In this study, we systematically analyze and compare the complex crystal structures and interaction properties of BMP-2 with its cognate receptors BMPR-I/BMPR-II and with its natural antagonist crossveinless-2 (CV-2) using an integrated in silico-in vitro strategy. It is found that the antagonist-binding site is not fully overlapped with the two receptor-binding sites on BMP-2 surface; the antagonist can competitively disrupt BMP-2-BMPR-II interaction using a blocking-out-of-site manner, but has no substantial influence on BMP-2-BMPR-I interaction. Here, the antagonist-binding site is assigned as a new functional epitope armpit to differ from the traditional conformational epitope wrist and linear epitope knuckle at receptor-binding sites. Structural analysis reveals that the armpit comprises three sequentially discontinuous, structurally vicinal peptide segments, separately corresponding to a loop region and two β-strands crawling on the protein surface. The three segments cannot work independently when splitting from the protein context, but can restore binding capability to CV-2 if they are connected to a single peptide. A systematic combination of different-length polyglycine linkers between these segments obtains a series of designed single peptides, from which several peptides that can potently interact with the armpit-recognition site of CV-2 with high affinity and specificity are identified using energetic analysis and fluorescence assay; they are expected to target BMP-2-CV-2 interaction in a self-inhibitory manner.

Rational Design and Intramolecular Cyclization of Hotspot Peptide Segments at YAP–TEAD4 Complex Interface

Background:

The Yes-Associated Protein (YAP) is a central regulator of Hippo pathway involved in carcinogenesis, which functions through interaction with TEA Domain (TEAD) transcription factors. Pharmacological disruption of YAP–TEAD4 complexes has been recognized as a potential therapeutic strategy against diverse cancers by suppressing the oncogenic activity of YAP.

Objective:

Two peptides, termed PS-1 and PS-2 are split from the interfacial context of YAP protein. Dynamics simulations, energetics analyses and fluorescence polarizations are employed to characterize the intrinsic disorder as well as binding energy/affinity of the two YAP peptides to TEAD4 protein.

Methods:

Two peptides, termed PS-1 and PS-2 are split from the interfacial context of YAP protein. Dynamics simulations, energetics analyses and fluorescence polarizations are employed to characterize the intrinsic disorder as well as binding energy/affinity of the two YAP peptides to TEAD4 protein.

Result:

The native conformation of PS-2 peptide is a cyclic loop, which is supposed to be constrained by adding a disulfide bond across the spatially vicinal residue pair Arg87-Phe96 or Met86- Phe95 at the peptide’s two ends, consequently resulting in two intramolecular cyclized counterparts of linear PS-2 peptide, namely PS-2(cyc87,96) and PS-2(cyc86,95). The linear PS-2 peptide is determined as a weak binder of TEAD4 (Kd = 190 μM), while the two cyclic PS-2(cyc87,96) and PS-2(cyc86,95) peptides are measured to have moderate or high affinity towards TEAD4 (Kd = 21 and 45 μM, respectively).

Conclusion:

PS-1 and PS-2 peptides are highly flexible and cannot maintain in native active conformation when splitting from the interfacial context, and thus would incur a considerable entropy penalty upon rebinding to the interface. Cyclization does not influence the direct interaction between PS-2 peptide and TEAD4 protein, but can largely reduce the intrinsic disorder of PS-2 peptide in free state and considerably minimize indirect entropy effect upon the peptide binding.

Rational design and chemical modification of TEAD coactivator peptides to target hippo signaling pathway against gastrointestinal cancers

Human Hippo signaling pathway has been recognized as a new and promising therapeutic target of gastrointestinal cancers, which is regulated by the intermolecular recognition between the TEA domain (TEAD) transcription factor and its prime coactivators. The coactivator proteins adopt two hotspot sites, namely α-helix and Ω-loop, to interact with TEAD. Here, we demonstrate that both the α-helix and Ω-loop peptides cannot maintain in structured state when splitting from the full-length coactivator proteins; they exhibit a large intrinsic disorder in free state that prevents the coactivator peptide recognition by TEAD. Rational design is used to optimize the interfacial residues of coactivator α-helix peptides, which can effectively improve the favorable direct readout effect upon the peptide binding to TEAD. Chemical modification is employed to constrain the free α-helix peptide into native ordered conformation. The method introduces an all-hydrocarbon bridge across i and i + 4 residues to stabilize the helical structure of a free coactivator peptide, which can considerably reduce the unfavorable indirect readout effect upon the peptide binding to TEAD. The all-hydrocarbon bridge is designed to point out of the TEAD-peptide complex interface, which would not disrupt the direct intermolecular interaction between the TEAD and peptide. Therefore, the stapling only improves peptide affinity, but does not alter peptide specificity, to TEAD. Affinity assay confirms that the binding potency of coactivator α-helix peptides is improved substantially by >5-fold upon the rational design and chemical modification. Structural analysis reveals that the optimized/stapled peptides can form diverse nonbonded interactions such as hydrogen bonds and hydrophobic contacts with TEAD, thus conferring stability and specificity to the TEAD-peptide complex systems.

Inverse in silico-in vitro fishing of unexpected paroxetine kinase targets from tumor druggable kinome

The serotonin selective reuptake inhibitor paroxetine has been clinically observed to reposition a significant suppressing potency on human tumors by unexpectedly targeting diverse kinase pathways involved in tumorigenesis. Here, we describe an inverse in silico-in vitro strategy to fish potential kinase targets using the paroxetine as bait. This is different (inverse) to the traditional drug discovery process that commonly screens small-molecule inhibitors for a specific kinase target. In the procedure, cell viability assays demonstrate that paroxetine has strong cytotoxicity on human tumor cell lines. Various protooncogene protein kinases are ontologically/manually enriched to define a druggable kinome, and a systematic interaction profile of paroxetine with the kinome is created, which indicates that paroxetine can potentially bind to some known targets or key regulators of human tumors. Kinase assays determine that paroxetine can effectively inhibit c-Src family kinases at nanomolar or micromolar levels. It is observed that the paroxetine ligand forms a tightly packed interface against the active site of these unexpected kinase targets to constitute several specific hydrogen bonds/π-π/cation-π stackings and a number of nonspecific hydrophobic/vdW contacts, while exposing a portion of molecular surface to solvent. More significantly, the ligand adopts two distinct binding modes (i.e., class I and class II) to interact with different kinases; the class-I mode has a higher stability and inhibitory activity than class-II mode. Steric clash seems to cause the ligand flipping from class I to class II. Graphical abstract.

Molecular selectivity design of mitogen-inducible gene-derived phosphopeptides between oncogenic HER kinases

Mitogen-inducible gene (MIG) is a natural negative regulator of the oncogenic HER kinase signaling by binding at the activation interface of kinase domain to disrupt the kinase dimerization. In this study, we systematically examine the binding structures, dynamics and energetics of MIG region 2 to four HER kinases based on their crystal or modeled complex structures, and identify an 8-mer phosphopeptide segment pYpY from the core strand sequence of MIG region 2 as the binding hotspot of MIG protein to HER kinases. We demonstrate that the small pYpY phosphopeptide can partially restore the binding affinity of full-length MIG protein, but exhibit a moderate selectivity over different HER kinases (S = 2.3-fold). In addition, the two phosphotyrosine residues pTyr394 and pTyr395 play an essential role in MIG-HER binding; dephosphorylation of them would fully eliminate the binding capability. A machine evolution algorithm is used to optimize the wild-type pYpY phosphopeptide, aiming to simultaneously improve affinity for these kinases and to maximize the affinity gap between different kinases. Consequently, a population is computationally evolved as selective phosphopeptide candidates; the dissociation constants of four representatives with HER kinases are systematically determined using binding affinity analysis, from which their selectivity is derived. The designed pYpYp3 phosphopeptide possesses a high selectivity over different HER kinases (S = 4.8-fold) and satisfactory affinity profile to these kinase (K_D = 140-1000 μM). Structural analysis observes that the global binding modes of pYpYp3 to different kinases are roughly consistent, but its local conformation may vary considerably, thus conferring specificity to the phosphopeptide.

Computational systematic selectivity of the Fasalog inhibitors between ROCK-I and ROCK-II kinase isoforms in Alzheimer's disease

Human Rho-associated coiled-coil forming kinase (ROCK) is a class of essential neurokinases that consists of two structurally conserved isoforms ROCK-I and ROCK-II; they have been revealed to play distinct roles in the pathogenesis of Alzheimer's disease (AD) and other neurological disorders. Selective targeting of the two kinase isoforms with small-molecule inhibitors is a great challenge due to the surprisingly high homology in kinase domain (92 %) and the full identity in kinase active site (100 %). Here, we describe a computational protocol to systematically profile the selectivity of Fasudil and its 25 analogs (termed as Fasalogs) between the two kinase isoforms. It is suggested that the substitution of Fasudil's 1,4-diazepane moiety with rigid ring such as Ripasudil and Dimehtylfasudil would render the resulting inhibitors of ROCK-II over ROCK-I (II-o-I) selectivity, while the substitution with long, flexible group such as H-89 and BDBM92607 tends to have I-o-II selectivity. Structural analysis reveals that the inhibitor affinity is not only determined by the identical active site, but also contributed from the non-identical first and second shells of the site as well as other non-conserved kinase regions, which can indirectly influence the active site and inhibitor binding through allosteric effect. A further kinase assay basically confirms the computational findings, which also exhibits a good consistence with theoretical selectivity over 10 tested samples (Rp = 0.89). In particular, the Fasalog compounds Dimehtylfasudil and H-89 are identified as II-o-I and I-o-II selective inhibitors. They can be considered as promising lead molecular entities to develop new specific ROCK isoform-selective Fasalog inhibitors.

Rational molecular targeting of the inter-subunit interaction between human cardiac troponin hcTnC and hcTnI using switch peptide-competitive biogenic medicines

The human cardiac troponin (hcTn) has been implicated in diverse cardiovascular diseases (CDs). The protein function is regulated by the inter-subunit interaction between the N-terminal domain of hcTnC and the C-terminal switch peptide of hcTnI; disruption of the interaction has been recognized as a potential therapeutic strategy for CDs. Here, we report use of biogenic medicines as small-molecule competitors to directly disrupt the protein-protein interaction by competitively targeting the core binding site (CBS) of hcTnC NTD domain. A multistep virtual screening protocol is performed against a biogenic compound library to identify competitor candidates and competition assay is employed to verify the screening results. Consequently, two compounds Collismycin and Compound e are identified as strong competitors (CC50 < 10 μM) with hcTnI for hcTnC CBS site, while other tested compounds are found to have moderate (CC50 = 10-100 μM), low (CC50 > 100 μM) or no (CC50 = N.D.) potency. The competitor ligands are anchored at the core groove of hcTnC CBS site through aromatic and hydrophobic interactions, while few peripheral hydrogen bonds are formed to further confer specificity for domain-compound recognition. These molecular-level findings would benefit from further in vitro and in vivo studies at cellular and animal levels, which can help to practice the ultimate therapeutic purpose.

Systematic response of staurosporine scaffold-based inhibitors to drug-resistant cancer kinase mutations

Human protein kinases have been established as promising druggable targets in cancer therapy. However, a large number of acquired drug-resistant kinase mutations are observed after first- and second-line kinase inhibitor treatments, largely limiting the application of small-molecule inhibitors in the targeted cancer therapy. Previously, the pan-kinase inhibitor staurosporine and its derivatives have been reported to selectively inhibit gatekeeper mutants over wild-type kinases, suggesting that the staurosporine scaffold is potentially helpful in developing wild-type-sparing inhibitors of drug-resistant kinase mutants. Here, a systematic response profile of 32 staurosporine scaffold-based inhibitors (SSBIs) for 61 ontology-enriched drug-resistant cancer kinase mutations is created using a combination of in silico analysis and in vitro assay, from which it is possible to identify those mutations that have the potential to cause resistance or confer sensitivity to SSBIs. The profile reveals that SSBIs exhibit distinct responses to kinase gatekeeper and nongatekeeper mutations, and SSBIs bearing p7 substituents can considerably influence their response to kinase gatekeeper mutations, particularly for the mutations of the Ile residue, which possesses a Cβ methyl group that tends to cause steric clash with bound SSBIs. Nongatekeeper mutations generally have a moderate and unfavorable effect on SSBI activity, as most of them are outside the kinase active site and do not directly contact inhibitor ligands. In addition, it is found that resistance is commonly caused by mutation-induced hindrance effects, whereas sensitivity is primarily conferred by mutation-established additional interactions.

Identification, characterization, and comparison of n-alkanols and anesthetics binding to the C1b subdomain of protein kinase cα: similar function with different binding sites

Protein kinase C (PKC) is a family of lipid-activated enzymes involved in anesthetic preconditioning signaling pathways. Previously, n-alkanols and general anesthetics have been found to activate PKC by binding to the kinase C1B subdomain. In the present study, we attempt to ascertain the molecular mechanism and interaction mode of human PKCα C1B subdomain with a variety of exogenous n-alkanols and volatile general anesthetics as well as endogenous activator phorbol ester (PE) and co-activator diacylglycerol (DG). Systematic bioinformatics analysis identifies three spatially vicinal sites on the subdomain surface to potentially accommodate small-molecule ligands, where the site 1 is a narrow, amphipathic pocket, the site 2 is a wide, flat and hydrophobic pocket, and the site 3 is a rugged, polar pocket. Further interaction modeling reveals that site 1 is the cognate binding region of natural PE activator, which can moderately simulate the kinase activity in an independent manner. The short-chain n-alkanols are speculated to also bind at the site to competitively inhibit PE-induced kinase activation. The long-chain n-alkanols and co-activator DG are found to target site 2 in a nonspecific manner, while the volatile anesthetics prefer to interact with site 3 in a specific manner. Since the site 1 is composed of two protein loops that are also shared by sites 2 and 3, binding of n-alkanols, DG and anesthetics to sites 2 and 3 can trigger a conformational displacement on the two loops, which enlarges the pocket size and changes the pocket configuration of site 1 through an allosteric mechanism, consequently enhancing kinase activation by improving PE affinity to the site.

Systematic profiling of staralog response to acquired drug resistant kinase gatekeeper mutations in targeted cancer therapy

Kinase-targeted therapy has been widely used as a lifesaving strategy for cancer patients. However, many patients treated with targeted cancer drugs are clinically observed to rapidly develop acquired resistance. Kinase gatekeeper mutation is one of the most chief factors contributing to the resistance, which modulates the accessibility of kinase's ATP-binding pocket. Previously, the pan-kinase inhibitor Staurosporine and its analogs (termed as Staralogs) have been reported to exhibit wild-type sparing selectivity for some kinase gatekeeper mutants, such as EGFR T790M, Her2 T798M and cSrc T338M. Here, we describe an integrative approach to systematically profile the molecular response of 15 representative Staralogs to 17 kinase gatekeeper mutations in targeted cancer therapy. With the profile we are able to divide gatekeeper mutations into three classes (i.e. classes I, II and III) and to divide Staralogs into two groups (i.e. groups 1 and 2) using heuristic clustering. The class I and II mutations confer consistent sensitivity and resistance for all Staralogs, respectively, while the class III mutations address divergent effects on different Staralogs. The mutations to Ile residue can generally reduce Staralog affinity by inducing unfavorable steric hindrance, whereas the mutations to Met and Leu residues would improve Staralog affinity by establishing favorable S···π interaction, van der Waals packing and/or hydrophobic contact. The group 1 and 2 Staralogs are primarily determined by carbonyl or hydroxyl substitution state at the position 7 of Staralog core, where points to kinase gatekeeper residue and can thus be directly influenced by gatekeeper mutation.

Inverse screening of Simvastatin kinase targets from glioblastoma druggable kinome

The statin drug Simvastatin is a HMG-CoA reductase inhibitor that has been widely used to lower blood lipid. However, the drug is clinically observed to reposition a significant suppressing potency on glioblastoma (GBM) by unexpectedly targeting diverse kinase pathways involved in GBM tumorigensis. Here, an inverse screening strategy is described to discover potential kinase targets of Simvastatin. Various human protein kinases implicated in GBM are enriched to define a druggable kinome; the binding behavior of Simvastatin to the kinome is profiled systematically via an integrative computational approach, from which most kinases have only low or moderate binding potency to Simvastatin, while only few are identified as promising kinase hits. It is revealed that Simvastatin can potentially interact with certain known targets or key regulators of GBM such as ErbB, c-Src and FGFR signaling pathways, but exhibit low affinity to the well-established GBM target of PI3K/Akt/mTOR pathway. Further assays determine that Simvastatin can inhibit kinase hits EGFR, MET, SRC and HER2 at nanomolar level, which are comparable with those of cognate kinase inhibitors. Structural analyses reveal that the sophisticated T790 M gatekeeper mutation can considerably reduce Simvastatin sensitivity to EGFR by inducing the ligand change between different binding modes.

Rational design and improvement of the dimerization-disrupting peptide selectivity between ROCK-I and ROCK-II kinase isoforms in cerebrovascular diseases

Human rho-associated coiled-coil forming kinases (ROCKs) ROCK-I and ROCK-II have been documented as attractive therapeutic targets for cerebrovascular diseases. Although ROCK-I and ROCK-II share a high degree of structural conservation and are both present in classic rho/ROCK signaling pathway, their downstream substrates and pathological functions may be quite different. Selective targeting of the two kinase isoforms with traditional small-molecule inhibitors is a great challenge due to their surprisingly high homology in kinase domain (~90%) and the full identity in kinase active site (100%). Here, instead of developing small-molecule drugs to selectively target the adenosine triphosphate (ATP) site of two isoforms, we attempt to design peptide agents to selectively disrupt the homo-dimerization event of ROCK kinases through their dimerization domains which have a relatively low conservation (~60%). Three helical peptides H1, H2, and H3 are split from the kinase dimerization domain, from which the isolated H2 peptide is found to have the best capability to rebind at the dimerization interface. A simulated annealing (SA) iteration method is used to improve the H2 peptide selectivity between ROCK-I and ROCK-II. The method accepts moderate degradation in peptide affinity in order to maximize the affinity difference between peptide binding to the two isoforms. Consequently, hundreds of parallel SA runs yielded six promising peptide candidates with ROCK-I over ROCK-II (I over II [IoII]) calculated selectivity and four promising peptide candidates with ROCK-II over ROCK-I (II over I [IIoI]) calculated selectivity. Subsequent anisotropy assays confirm that the selectivity values range between 13.2-fold and 83.9-fold for IoII peptides, and between 5.8-fold and 21.2-fold for IIoI peptides, which are considerably increased relative to wild-type H2 peptide (2.6-fold for IoII and 2.0-fold for IIoI). The molecular origin of the designed peptide selectivity is also analyzed at structural level; it is revealed that the peptide residues can be classified into conserved, non-conserved, and others, in which the non-conserved residues play a crucial role in defining peptide selectivity, while conserved residues confer stability to kinase-peptide binding.

PreDBA: A heterogeneous ensemble approach for predicting protein-DNA binding affinity

The interaction between protein and DNA plays an essential function in various critical natural processes, like DNA replication, transcription, splicing, and repair. Studying the binding affinity of proteins to DNA helps to understand the recognition mechanism of protein-DNA complexes. Since there are still many limitations on the protein-DNA binding affinity data measured by experiments, accurate and reliable calculation methods are necessarily required. So we put forward a computational approach in this paper, called PreDBA, that can forecast protein-DNA binding affinity effectively by using heterogeneous ensemble models. One hundred protein-DNA complexes are manually collected from the related literature as a data set for protein-DNA binding affinity. Then, 52 sequence and structural features are obtained. Based on this, the correlation between these 52 characteristics and protein-DNA binding affinity is calculated. Furthermore, we found that the protein-DNA binding affinity is affected by the DNA molecule structure of the compound. We classify all protein-DNA compounds into five classifications based on the DNA structure related to the proteins that make up the protein-DNA complexes. In each group, a stacked heterogeneous ensemble model is constructed based on the obtained features. In the end, based on the binding affinity data set, we used the leave-one-out cross-validation to evaluate the proposed method comprehensively. In the five categories, the Pearson correlation coefficient values of our recommended method range from 0.735 to 0.926. We have demonstrated the advantages of the proposed method compared to other machine learning methods and currently existing protein-DNA binding affinity prediction approach.

Molecular modeling and rational design of hydrocarbon-stapled/halogenated helical peptides targeting CETP self-binding site: Therapeutic implication for atherosclerosis

The human plasma cholesteryl ester transfer protein (CETP) collects triglycerides from very-/low-density lipoproteins (V/LDL) and exchanges them for cholesteryl esters from high-density lipoproteins (HDL), which has recognized as an important therapeutic target for atherosclerosis. The protein has a C-terminal amphipathic α-helix that serves as self-binding peptide to fulfill biological function by dynamically binding to/unbinding from its cognate site (termed self-binding site) in the same protein. Previously, we successfully derived and halogenated the helical peptide to competitively disrupt the self-binding behavior of CETP C-terminal tail. However, the halogenated peptides have only a limited affinity increase as compared to native helical peptide (∼3-fold), thus exhibiting only a moderate competitive potency. Here, instead of optimizing the direct intermolecular interaction of peptide with CETP self-binding site we attempt to further improve the peptide competitive potency by reducing its conformational flexibility with hydrocarbon-stapling technique. Computational analysis reveals that the helical peptide has large intrinsic disorder in unbound free state, which would incur a considerable entropy penalty upon rebinding to the self-binding site. All-hydrocarbon bridge is designed and optimized on native and halogenated peptides in terms of the helical pattern and binding mode of self-binding peptide. Dynamics simulation and circular dichroism indicate that the stapling can considerably reduce peptide disorder in free state. Energetics calculation and fluorescence assay conform that the binding affinity of stapled/halogenated peptides is improved substantially (by > 5-fold), thus exhibiting an effective competition potency with native peptide for the self-binding site. Structural examination suggests that the binding modes and nonbonded interactions of native and halogenated peptides are not influenced essentially due to the stapling.

Exploring calcium ion-dependent effect on the intermolecular interaction between human secreted phospholipase A2 and its peptide inhibitors in coronary artery disease

Human secreted phospholipase A₂ (hsPLA₂) is a small calcium ion (Ca²⁺)-regulatory protein secreting from platelets, eosinophils and T-lymphocytes, which has been established as an important biomarker and potential target for the diagnosis and therapy of coronary artery disease. Short peptide inhibitors are used to competitively suppress the enzymatic activity of hsPLA₂. Here, Ca²⁺ effect on the intermolecular recognition and interaction between hsPLA₂ and its peptide inhibitors is investigated systematically by using molecular modeling and bioinformatics analysis. Dynamics simulations reveal that the hsPLA₂ structure bound with Ca²⁺ is rather stable and has low thermal motion; removal of Ca²⁺ considerably increases structural flexibility and intrinsic disorder of the protein. Energetics calculations suggest that presence of Ca²⁺ can effectively promote the interaction of hsPLA₂ with peptide inhibitors. In particular, the local substructures of hsPLA₂ such as helix H1, loop L2 and double-stranded β-sheet DS that participate in peptide recognition are involved in or nearby Ca²⁺-coordinating site and can be directly stabilized by the Ca²⁺. In addition, a significant concentration-dependent effect of Ca²⁺ on peptide-hsPLA₂ binding is observed in vitro, that is, a little of Ca²⁺ can largely improve peptide binding affinity, but high Ca²⁺ concentration does not increase the affinity substantially. The correlation between calculated free energy and experimental binding affinity over different peptide inhibitors is improved considerably by adding Ca²⁺ to hsPLA₂. Specifically, the FLSYK peptide can generally bind to Ca²⁺-bound hsPLA₂ with a moderate or high affinity (K_d ranges between 56 and 210 μM), but have only a modest affinity or even nonbinding to Ca²⁺-free hsPLA₂ (K_d > 400 μM or = n.d.).

Molecular design and optimization of hepatic cancer SLP76-derived PLCγ1 SH3-binding peptide with the systematic N-substitution of peptide PXXP motif

The phospholipase Cγ1 (PLCγ1) is essential for T-cell signaling and activation in hepatic cancer immune response, which has a regulatory Src homology 3 (SH3) domain that can specifically recognize and interact with the PXXP-containing decapeptide segment (¹⁸⁵ QPPVPPQRPM¹⁹⁴ , termed as SLP76^185-194 peptide) of adaptor protein SLP76 following T-cell receptor ligation. The isolated peptide can only bind to the PLCγ1 SH3 domain with a moderate affinity due to lack of protein context support. Instead of the traditional natural residue mutagenesis that is limited by low structural diversity and shifted target specificity, we herein attempt to improve the peptide affinity by replacing the two key proline residues Pro187 and Pro190 of SLP76^185-194 PXXP motif with nonnatural N-substituted amino acids, as the proline is the only endogenous N-substituted amino acid. The replacement would increase peptide flexibility but can restore peptide activity by establishing additional interactions with the domain. Structural analysis reveals that the domain pocket can be divided into a large amphipathic region and a small negatively charged region; they accommodate hydrophobic, aromatic, polar, and moderate-sized N-substituted amino acid types. A systematic replacement combination profile between the peptide residues Pro187 and Pro190 is created by structural modeling, dynamics simulation, and energetics analysis, from which six improved and two reduced N-substituted peptides as well as native SLP76^185-194 peptide are identified and tested for their binding affinity to the recombinant protein of the human PLCγ1 SH3 domain using fluorescence-based assays. Two N-substituted peptides, SLP76^185-194 (N-Leu187/N-Gln190) and SLP76^185-194 (N-Thr187/N-Gln190), are designed to have high potency (K_d  = 0.67 ± 0.18 and 1.7 ± 0.3 μM, respectively), with affinity improvement by, respectively, 8.5-fold and 3.4-fold relative to native peptide (K_d  = 5.7 ± 1.2 μM).

Systematic profiling of clinical missence mutation effects on the intermolecular interaction between human growth hormone and its receptor in isolated growth hormone deficiency

Isolated growth hormone deficiency (IGHD) is the most common pituitary hormone deficiency and can result from congenital or acquired causes. Among the known factors, genetic mutations in human growth hormone (hGH) remain the most frequent cause of IGHD, which influence the binding of hGH to its cognate receptor (hGHbp). Although previous studies have systematically investigated the residue importance at hGH-hGHbp complex interface, the molecular role of IGHD-associated residue mutations in the complex function still remains largely unexplored. Here, a total of 21 known hGH naturally-occurring missence mutations that have been clinically observed to be involved in IGHD disorder are collected and confirmed by original literature; they effects on the conformation, energetics and dynamics of hGH-hGHbp recognition and interaction are dissected at molecular level by using atomistic dynamics simulations, binding energy calculations and fluorescence spectroscopy assays. A systematic profile of hGH-hGHbp binding response to these clinical missence mutations is created, based on which it is revealed that (i) most mutations have appreciably unfavorable effect on the binding, which potentially destabilize the complex interaction, while only very few are predicted as moderate stabilizers for the complex system, and (ii) these disease-related mutations can locate either at complex interface or in hGH protein interior far away from the interface; both can influence the complex binding through either direct interaction or indirect allostericity. Two mutations, E100K (non-interface) and G146R (interface), are identified to address potent destabilization effect on hGH-hGHbp complex system; they can reduce the complex binding affinity by 8-fold (K_d changes from 0.76 to 5.9 nM) and 46-fold (K_d changes from 0.76 to 34.7 nM), respectively.

Molecular Design of Peptide-Fc Fusion Drugs

Background:

Peptide-Fc fusion drugs, also known as peptibodies, are a category of biological therapeutics in which the Fc region of an antibody is genetically fused to a peptide of interest. However, to develop such kind of drugs is laborious and expensive. Rational design is urgently needed.

Methods:

We summarized the key steps in peptide-Fc fusion technology and stressed the main computational resources, tools, and methods that had been used in the rational design of peptide-Fc fusion drugs. We also raised open questions about the computer-aided molecular design of peptide-Fc.

Results:

The design of peptibody consists of four steps. First, identify peptide leads from native ligands, biopanning, and computational design or prediction. Second, select the proper Fc region from different classes or subclasses of immunoglobulin. Third, fuse the peptide leads and Fc together properly. At last, evaluate the immunogenicity of the constructs. At each step, there are quite a few useful resources and computational tools.

Conclusion:

Reviewing the molecular design of peptibody will certainly help make the transition from peptide leads to drugs on the market quicker and cheaper.

Systematic profiling of SH3-mediated Tau-Partner interaction network in Alzheimer's disease by integrating in silico analysis and in vitro assay

The aberrant assembly of microtubule-associated protein Tau (τ) into insoluble aggregates is closely related to Alzheimer's disease (AD), which is elicited from Tau phosphorylation events and regulated by the specific intermolecular recognition between the proline-rich PxxP motifs of Tau and the SH3 domains of its diverse partner proteins/kinases. Here, we attempt to create a systematic interaction profile for the 10 SH3 domains of previously reported Tau partners across all the 18 Tau PxxP peptides. A number of biologically functional SH3-PxxP interaction events are identified from the profile and then tested using fluorescence spectroscopy. It is revealed that (i) the region (residues 520-560) precedent to the tubulin-binding partial repeats of Tau protein is an important target of SH3 domains, where contains the three PxxP peptides τp^527-536, τp^530-539 and τp^547-556 that exhibit different binding profiles towards the investigated SH3 domains, (ii) as compared to τp^527-536 and τp^547-556, the τp^530-539 peptide located between them has only a modest binding potency to most SH3 domains, suggesting that the three peptides contribute unevenly to Tau-SH3 interactions, and (iii) some other Tau PxxP peptides, particularly those within the residue range 490-510 that is neighboring to the region 520-560, can also interact effectively with several SH3 domains. The SH3 domain of the well known Tau partner kinase Fyn is determined to have high or moderate affinity for an array of Tau PxxP peptides, including τp^137-146, τp^493-502, τp^527-536 and τp^547-556 (K_d ranges 15.7-85.6 μM).

Systematic analysis and identification of unexpected interactions from the neuroprotein drug interactome in hydrocephalus pharmacological intervention

Hydrocephalus is a neurological condition caused by an abnormal accumulation of cerebrospinal fluid; pharmacological intervention of the disease has been found to elicit a variety of adverse drug reactions (ADRs) in central nervous system (CNS) by unexpectedly targeting certain functional neuroproteins. Here, a systematic neuroprotein drug interactome (SNDI) is created for 11 hydrocephalus drugs/metabolites plus 20 control drugs across 518 druggable pockets on the surface of 472 CNS neuroproteins via a large-scale molecular docking approach. Heuristic clustering analysis of the SNDI profile divides the 31 investigated drug ligands into a distinct panel and a background panel; the former consists of two hydrocephalus drugs (Furosemide and Triamterene) and their respective metabolites (Furosemide glucuronide and Hydroxytriamterene) that are inferred to have generally high affinity towards the whole array of neuroprotein pockets. A total of 13 neuroproteins are enriched in gene ontology semantic mining as putative unexpected targets of the distinct panel, and their intermolecular interactions with hydrocephalus drugs/metabolites are investigated in detail using dynamics simulation and energetics analysis. We also perform kinase assay and viability test to substantiate the interactome analysis. It is found that the Furosemide and Triamterene have significant cytotoxic effects on normal human astrocytes, in which the Triamterene can inhibit the neurokinase ROCK2, a representative of putative unexpected targets, with a high activity, which is comparable with the sophisticated ROCK2 inhibitor Fasudil.