ABSTRACT
While research into drug-target interaction (DTI) prediction is fairly mature, generalizability and interpretability are not always addressed in the existing works in this field. In this paper, we propose a deep learning (DL)-based framework, called BindingSite-AugmentedDTA, which improves drug-target affinity (DTA) predictions by reducing the search space of potential-binding sites of the protein, thus making the binding affinity prediction more efficient and accurate. Our BindingSite-AugmentedDTA is highly generalizable as it can be integrated with any DL-based regression model, while it significantly improves their prediction performance. Also, unlike many existing models, our model is highly interpretable due to its architecture and self-attention mechanism, which can provide a deeper understanding of its underlying prediction mechanism by mapping attention weights back to protein-binding sites. The computational results confirm that our framework can enhance the prediction performance of seven state-of-the-art DTA prediction algorithms in terms of four widely used evaluation metrics, including concordance index, mean squared error, modified squared correlation coefficient ($r^2_m$) and the area under the precision curve. We also contribute to three benchmark drug-traget interaction datasets by including additional information on 3D structure of all proteins contained in those datasets, which include the two most commonly used datasets, namely Kiba and Davis, as well as the data from IDG-DREAM drug-kinase binding prediction challenge. Furthermore, we experimentally validate the practical potential of our proposed framework through in-lab experiments. The relatively high agreement between computationally predicted and experimentally observed binding interactions supports the potential of our framework as the next-generation pipeline for prediction models in drug repurposing.
Subject(s)
Algorithms , Drug Repositioning , Drug Development , Proteins/chemistry , Binding SitesABSTRACT
The Ebola virus glycoprotein (GP) gene templates several mRNAs that produce either the virion-associated transmembrane protein or one of two secreted glycoproteins. Soluble glycoprotein (sGP) is the predominant product. GP1 and sGP share an amino terminal sequence of 295 amino acids but differ in quaternary structure, with GP1 being a heterohexamer with GP2 and sGP a homodimer. Two structurally different DNA aptamers were selected against sGP that also bound GP1,2. These DNA aptamers were compared with a 2'FY-RNA aptamer for their interactions with the Ebola GP gene products. The three aptamers have almost identical binding isotherms for sGP and GP1,2 in solution and on the virion. They demonstrated high affinity and selectivity for sGP and GP1,2. Furthermore, one aptamer, used as a sensing element in an electrochemical format, detected GP1,2 on pseudotyped virions and sGP with high sensitivity in the presence of serum, including from an Ebola-virus-infected monkey. Our results suggest that the aptamers interact with sGP across the interface between the monomers, which is different from the sites on the protein bound by most antibodies. The remarkable similarity in functional features of three structurally distinct aptamers suggests that aptamers, like antibodies, have preferred binding sites on proteins.
Subject(s)
Aptamers, Nucleotide , Ebolavirus , Viral Envelope Proteins , Humans , Aptamers, Nucleotide/chemistry , Ebolavirus/chemistry , Viral Envelope Proteins/chemistry , Protein MultimerizationABSTRACT
Experimental results show that the adsorption of the self-assembled monolayers (SAMs) on a gold surface induces surface stress change that causes deformation of the underlying substrate. However, the exact mechanism of stress development is yet to be elucidated. In the present study, multiscale computational models based on molecular dynamics (MD) simulations are applied to study the mechanism governing surface stress change. Distinct mechanisms for adsorption-induced surface deformation, namely, interchain repulsion and thiol-gold interaction-driven gold surface reconstruction, are investigated. Two different interatomic potentials, embedded atom method and surface-embedded atom method (SEAM), are used in the MD simulations to study the reconstruction-induced surface stresses. Comparison of the predicted surface stress changes, resulting from MD and continuum mechanics-based models, with the observed experimental response indicates that a modified SEAM-based multiscale model can better capture the surface stress changes observed during alkanethiol SAM formation, and gold surface reconstruction is the primary factor behind the surface stress change. The interchain repulsions of SAM are found to have a minimal contribution. Also, both the simulations and experiments show that the surface stress change increases with the increase of surface coverage density and larger grain size.
Subject(s)
Molecular Dynamics Simulation , Sulfhydryl Compounds , Adsorption , Gold , Surface PropertiesABSTRACT
Severe acute respiratory syndrome coronavirus 2 (SARS CoV-2) has been the primary reason behind the COVID-19 global pandemic which has affected millions of lives worldwide. The fundamental cause of the infection is the molecular binding of the viral spike protein receptor binding domain (SP-RBD) with the human cell angiotensin-converting enzyme 2 (ACE2) receptor. The infection can be prevented if the binding of RBD-ACE2 is resisted by utilizing certain inhibitors or drugs that demonstrate strong binding affinity towards the SP RBD. Sialic acid based glycans found widely in human cells and tissues have notable propensity of binding to viral proteins of the coronaviridae family. Recent experimental literature have used N-acetyl neuraminic acid (Sialic acid) to create diagnostic sensors for SARS-CoV-2, but a detailed interrogation of the underlying molecular mechanisms is warranted. Here, we perform all atom molecular dynamics (MD) simulations for the complexes of certain Sialic acid-based molecules with that of SP RBD of SARS CoV-2. Our results indicate that Sialic acid not only reproduces a binding affinity comparable to the RBD-ACE2 interactions, it also assumes the longest time to dissociate completely from the protein binding pocket of SP RBD. Our predictions corroborate that a combination of electrostatic and van der Waals energies as well the polar hydrogen bond interactions between the RBD residues and the inhibitors influence free energy of binding.Communicated by Ramaswamy H. Sarma.
ABSTRACT
Biosensors hold great potential for revolutionizing personalized medicine and environmental monitoring. Their construction is the key factor which depends on either manufacturing techniques or robust sensing materials to improve efficacy of the device. Functional graphene is an attractive choice for transducing material due to its various advantages in interfacing with biorecognition elements. Graphene and its derivatives such as graphene oxide (GO) are thus being used extensively for biosensors for monitoring of diseases. In addition, graphene can be patterned to a variety of structures and is incorporated into biosensor devices such as microfluidic devices and electrochemical and plasmonic sensors. Among biosensing materials, GO is gaining much attention due to its easy synthesis process and patternable features, high functionality, and high electron transfer properties with a large surface area leading to sensitive point-of-use applications. Considering demand and recent challenges, this perspective review is an attempt to describe state-of-the-art biosensors based on functional graphene. Special emphasis is given to elucidating the mechanism of sensing while discussing different applications. Further, we describe the future prospects of functional GO-based biosensors for health care and environmental monitoring with a focus on additive manufacturing such as 3D printing.
Subject(s)
Biosensing Techniques , Graphite , Printing, Three-Dimensional , Electrochemical TechniquesABSTRACT
Experimental and computational approaches are utilized to investigate the influence of electrostatic fields on the binding force between human coagulation protein thrombin and its DNA aptamer. The thiolated aptamer was deposited onto gold substrate located in a liquid cell filled with binding buffer, then the thrombin-functionalized atomic force microscopy (AFM) probe was repeatedly brought into contact with the aptamer-coated surface under applied electrical potentials of -100, 0, and 100â¯mV respectively. Force drops during the pull-off process were measured to determine the unbinding forces between thrombin and aptamer in a range of loading rates spanning from ~3â¯×â¯102 to ~1â¯×â¯104â¯pN/s. The results from experiments showed that both of the binding strength and propensity of the complex are drastically diminished under positive electrode potential, whereas there is no influence on the molecular binding from negative electrode potential. We also used a theoretical analysis to explain the nature of electrostatic potential and field inside the aptamer-thrombin layer, which in turn could quantify the influence of the electrostatically repulsive force on a thrombin molecule that promotes dissociation from the aptamer due to positive electrode potential, and achieve good agreement with the experimental results. The study confirms the feasibility of electrostatic modulation upon the binding interaction between thrombin and aptamer, and implicates an underlying application perspective upon nanoscale manipulation of the stimuli responsive biointerface.
Subject(s)
Aptamers, Nucleotide/chemistry , Thrombin/chemistry , Biosensing Techniques/methods , Electricity , Electrochemical Techniques/methods , Electrodes , Gold/chemistry , Mechanical Phenomena , Models, Molecular , Protein Binding , Static Electricity , Surface PropertiesABSTRACT
Nanoporous alumina membranes have become a ubiquitous biosensing platform for a variety of applications and aptamers are being increasingly utilized as recognition elements in protein sensing devices. Combining the advantages of the two, we report label-free sensitive detection of human α-thrombin by an aptamer-functionalized nanoporous alumina membrane using a four-electrode electrochemical cell. The sensor response to α-thrombin was determined in the presence of a high concentration (500⯵M) of human serum albumin (HSA) as an interfering protein in the background. The sensor sensitivity was also characterized against γ-thrombin, which is a modified α-thrombin lacking the aptamer binding epitope. The detection limit, within an appreciable signal/noise ratio, was 10â¯pM of α-thrombin in presence of 500⯵M HSA. The proposed scheme involves the use of minimum reagents/sample preparation steps, has appreciable response in presence of high concentrations of interfering molecules and is readily amenable to miniaturization by association with existing-chip based electrical systems for application in point-of-care diagnostic devices.
Subject(s)
Aptamers, Nucleotide/chemistry , Biosensing Techniques/instrumentation , Nanopores , Thrombin/analysis , Electrodes , Equipment Design , Humans , Limit of Detection , Membranes, Artificial , Nanopores/ultrastructure , Serum Albumin/analysisABSTRACT
We report an aptamer functionalized stimuli responsive surface that can controllably switch between binding and releasing its specific ligand under application of electrical stimuli. The high affinity of the aptamer for thrombin makes the surface undergo specific binding while electrostatic field induced actuation of the aptamer is utilized to release the ligand from the surface. Atomic force microscopy (AFM) was utilized to determine the characteristic height change, associated with the specific binding of thrombin, on anti-thrombin aptamer coated surfaces. Subsequently, the thrombin/aptamer complex covered surfaces were subjected to different magnitudes of electrostatic field and height changes on the surface were measured to investigate the influence of an electrical field. Application of positive electrical potential led to the removal of thrombin from the aptamer-covered surface. While under moderate magnitude of negative electrical potential the binding complexes were maintained, increasing the magnitude led to the removal of both molecules from the surface. Molecular dynamics (MD) simulations of the thrombin/aptamer complex under electrostatic fields show that thrombin dissociates from the aptamers in the presence of a positive electric field. These results demonstrate that aptamer covered surfaces undergo specific binding to the ligand and an electrostatic field may be used to disrupt the binding and on-demand release of the ligand from the surface.
ABSTRACT
The binding/unbinding of the human thrombin and its 15-mer single stranded DNA aptamer, under the application of external stimulus in the form of electrostatic potential/electric field, is investigated by a combination of continuum analysis and atomistic molecular dynamics simulation. In agreement with the experiments that demonstrate the influence of electrostatic potential on the thrombin/aptamer complex, our computations show that the application of positive electric field successfully unbinds the thrombin from the aptamer. Results from umbrella sampling simulations reveal that there is a decrease in the free energy of binding between the thrombin and aptamer in presence of positive electric fields. Hydrogen bonding and non-bonded interaction energies, and hence the free energy of binding, between the thrombin and its aptamer reduce as the applied electric field is shifted from negative to positive values. Our analyses demonstrate that application of electrical stimulus modifies the molecular interactions within the complex and consequently, electrical field can be used to modulate the association between the thrombin and its aptamer.