SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
A pre-trained convolutional neural network was fine-tuned for three separate classification tasks, distinguishing 2D images of: 1) single amino acids, 2) protein structural ball and stick images of metalloproteins, and 3) protein structural ball and stick images of metalloenzymes with the metal cofactors removed. Images used in the training, testing and validation are shared here.
Explore the source record for details and available documents.
PARP13/ZAP (zinc-finger antiviral protein) acts against multiple viruses by promoting degradation of viral mRNA. PARP13 has four N-terminal zinc (Zn) fingers that bind CG-rich nucleotide sequences, a C-terminal ADP ribosyltransferase fold, and a central region with a fifth Zn finger and tandem WWE domains. The central PARP13 region, ZnF5-WWE1-WWE2, is implicated in binding poly(ADP-ribose); however, there are limited insights into its structure and function. We present crystal structures of ZnF5-WWE1-WWE2 from mouse PARP13 in complex with ADP-ribose and in complex with ATP. The crystal structures and binding studies demonstrate that WWE2 interacts with ADP-ribose and ATP, whereas WWE1 does not have a functional binding site. Binding studies with poly(ADP-ribose) ligands indicate that WWE2 serves as an anchor for preferential binding to the terminal end of poly(ADP-ribose) chains. The composite ZnF5-WWE1-WWE2 structure forms an extended surface to engage ADP-ribose chains, representing a distinctive mode of recognition that provides a framework for investigating the impact of poly(ADP-ribose) on PARP13 function.
Glycogen is the primary storage polysaccharide in bacteria and animals. It is a glucose polymer linked by α-1,4 glucose linkages and branched via α-1,6-linkages, with the latter reaction catalyzed by branching enzymes. Both the length and dispensation of these branches are critical in defining the structure, density, and relative bioavailability of the storage polysaccharide. Key to this is the specificity of branching enzymes because they define branch length. Herein, we report the crystal structure of the maltooctaose-bound branching enzyme from the enterobacteria E. coli. The structure identifies three new malto-oligosaccharide binding sites and confirms oligosaccharide binding in seven others, bringing the total number of oligosaccharide binding sites to twelve. In addition, the structure shows distinctly different binding in previously identified site I, with a substantially longer glucan chain ordered in the binding site. Using the donor oligosaccharide chain-bound Cyanothece branching enzyme structure as a guide, binding site I was identified as the likely binding surface for the extended donor chains that the E. coli branching enzyme is known to transfer. Furthermore, the structure suggests that analogous loops in branching enzymes from a diversity of organisms are responsible for branch chain length specificity. Together, these results suggest a possible mechanism for transfer chain specificity involving some of these surface binding sites.
Interfacial enzyme reactions require formation of an enzyme–substrate complex at the surface of a heterogeneous substrate, but often multiple modes of enzyme binding and types of binding sites complicate analysis of their kinetics. Excess heterogeneous substrate is often used as a justification to model the substrate as unchanging, but using the study of the enzymatic hydrolysis of insoluble cellulose as an example, we argue that reaction rates are dependent on evolving substrate interfacial properties. We hypothesize that the relative abundance of binding sites on cellulose where hydrolysis can occur (productive binding sites) and binding sites where hydrolysis cannot be initiated or is inhibited (nonproductive binding sites) contribute to rate limitations. We show that the initial total number of productive binding sites (the productive binding capacity) determines the magnitude of the initial burst phase of cellulose hydrolysis, while productive binding site depletion explains overall hydrolysis kinetics. Furthermore, we show that irreversibly bound surface enzymes contribute to the depletion of productive binding sites. Our model shows that increasing the ratio of productive to nonproductive binding sites promotes hydrolysis, while maintaining an elevated productive binding capacity throughout conversion is key to preventing hydrolysis slowdown.
Here, conventional methods for extracting rare earth metals (REMs) from mined mineral ores are inefficient, expensive, and environmentally damaging. Recent discovery of lanmodulin (LanM), a protein that coordinates REMs with high-affinity and selectivity over competing ions, provides inspiration for new REM refinement methods. Here, we used quantum mechanical (QM) methods to investigate trivalent lanthanide cation (Ln 3+ ) interactions with coordination systems representing bulk solvent water and protein binding sites. Energy decomposition analysis (EDA) showed differences in the energetic components of Ln 3+ interaction with representatives of solvent (water, H 2 O) and protein binding sites (acetate, CH 3 COO – ), highlighting the importance of accurate description of electrostatics and polarization in computational modeling of REM interactions with biological and bioinspired molecules. Relative binding free energies were obtained for Ln 3+ with coordination complexes originating from binding sites in PDB structures of a lanthanum binding peptide (PDB entry 7CCO) and LanM, with explicit consideration of the first hydration shell waters, according to quasi-chemical theory (QCT). Beyond the first shell, the bulk solvent environment was represented with an implicit continuum model. Ln 3+ interactions with (H 2 O) 9 and both binding site models became more favorable, moving down the periodic series. This trend was more pronounced with the protein binding site models than with water, resulting in affinity increasing with periodic number, except for the last REM, Lu 3+ , which bound less favorably than the preceding element, Yb 3+ . Using the truncated 7CCO binding site model, the magnitude and trend of the experimental Ln 3+ relative binding free energies for the whole 7CCO peptide were reproduced. Conversely, the previously reported experimental data for LanM show a preference for the earlier lanthanides; this is likely due to longer-range interactions and cooperative effects, which are not represented by the reduced models. Using the truncated 7CCO binding site model, the magnitude and trend of the experimental Ln 3+ relative binding free energies for the whole 7CCO peptide were reproduced. In contrast to the previously reported experimental data for LanM, the peptide preferentially binds the earlier lanthanides. This difference likely arises due to longer-range interactions and cooperative effects not represented by the peptide. Further investigation of Ln 3+ interactions with whole proteins using polarizable molecular mechanics models with explicit solvent is warranted to understand the influence of longer-ranged interactions, cooperativity, and bulk solvent. Nevertheless, the present work provides new insights into Ln 3+ interactions with biomolecules and presents an effective computational platform for designing specific single-site REM binding peptides more efficiently.
Initial cell attachment of rotavirus (RV) to specific cell surface glycan receptors, which is the essential first step in RV infection, is mediated by the VP8* domain of the spike protein VP4. Recently, human histo-blood group antigens (HBGAs) have been identified as receptors or attachment factors for human RV strains. RV strains in the P[4] and P[8] genotypes of the P[II] genogroup share common recognition of the Lewis b (Le b ) and H type 1 antigens, however, the molecular basis of receptor recognition by the major human P[8] RVs remains unknown due to lack of experimental structural information. Here, we used nuclear magnetic resonance (NMR) spectroscopy-based titration experiments and NMR-derived high ambiguity driven docking (HADDOCK) methods to elucidate the molecular basis for P[8] VP8* recognition of the Leb (LNDFH I) and type 1 HBGAs. We also used X-ray crystallography to determine the molecular details underlying P[6] recognition of H type 1 HBGAs. Unlike P[6]/P[19] VP8*s that recognize H type 1 HBGAs in a binding surface composed of an α-helix and a β-sheet, referred as the “βα binding site”, the P[8] and P[4] VP8*s bind Le b HBGAs in a previously undescribed pocket formed by the edges of two β-sheets, referred to as the “ββ binding site”. Importantly, the P[8] and P[4] VP8*s retain binding capability to non-Le b type 1 HBGAs using the βα binding site. The presence of two distinct binding sites for Le b and non-Le b HBGA glycans in the P[8] and P[4] VP8* domains suggests host-pathogen co-evolution under structural and functional adaptation of RV pathogens to host glycan polymorphisms. Assessment and understanding of the precise impact of this co-evolutionary process in determining RV host ranges and cross-species RV transmission should facilitate improved RV vaccine development and prediction of future RV strain emergence and epidemics.
All chemical forms of energy and oxygen on Earth are generated via photosynthesis where light energy is converted into redox energy by two photosystems (PS I and PS II). There is an increasing number of PS I 3D structures deposited in the Protein Data Bank (PDB). The Triangular Spatial Relationship (TSR)-based algorithm converts 3D structures into integers (TSR keys). A comprehensive study was conducted, by taking advantage of the PS I 3D structures and the TSR-based algorithm, to answer three questions: (i) Are electron cofactors including P700, A -1 and A 0 , which are chemically identical chlorophylls, structurally different? (ii) There are two electron transfer chains (A and B branches) in PS I. Are the cofactors on both branches structurally different? (iii) Are the amino acids in cofactor binding sites structurally different from those not in cofactor binding sites? The key contributions and important findings include: (i) a novel TSR-based method for representing 3D structures of pigments as well as for quantifying pigment structures was developed; (ii) the results revealed that the redox cofactor, P700, are structurally conserved and different from other redox factors. Similar situations were also observed for both A -1 and A 0 ; (iii) the results demonstrated structural differences between A and B branches for the redox cofactors P700, A -1 , A 0 and A 1 as well as their cofactor binding sites; (iv) the tryptophan residues close to A 0 and A 1 are structurally conserved; (v) The TSR-based method outperforms the Root Mean Square Deviation (RMSD) and the Ultrafast Shape Recognition (USR) methods. The structural analyses of redox cofactors and their binding sites provide a foundation for understanding the unique chemical and physical properties of each redox cofactor in PS I, which are essential for modulating the rate and direction of energy and electron transfers.
ABSTRACT Human immunoglobulin preparations contain a diverse range of polyclonal antibodies that reflect past immune responses against pathogens encountered by the blood donor population. In this study, we examined a panel of intravenous immunoglobulins (IGIVs) manufactured over the past two decades (1998–2020) for their capacity to neutralize or enhance Zika virus (ZIKV) infectionin vitro. These IGIVs were selected specifically based on their production dates in relation to the occurrences of two flavivirus outbreaks in the U.S.: the West Nile virus (WNV) outbreak in 1999 and the ZIKV outbreak in 2015. As demonstrated by enzyme-linked immunosorbent assay (ELISA) experiments, IGIVs made before the ZIKV outbreak already harbored antibodies that bind to various peptides across the envelope protein of ZIKV because of the WNV outbreak. Using phage display, the most dominant binding site was mapped precisely to the P2 peptide between residues 211 and 230 within domain II, where BF1176-56, an anti-ZIKV monoclonal antibody, also binds. When tested in permissive Vero E6 cells for ZIKV neutralization, the IGIVs, even after undergoing rigorous enrichment for P2 binding specificity, failed, as did BF1176-56. Meanwhile, BF1176-56 enhanced ZIKV infection in both FcγRII-expressing K562 cells and human peripheral blood mononuclear cells. However, for enhancement by the IGIVs to be detected in these cells, a substantial increase in their P2 binding specificity was required, thus linking the P2 site with ZIKV enhancementin vitro. Our findings warrant further study of the significance of elevated levels of anti-WNV antibodies in IGIVs, considering that various mechanisms operatingin vivomay modulate ZIKV infection outcomes. IMPORTANCE We investigated the capacity of intravenous immunoglobulins manufactured previously over two decades (1998–2020) to neutralize or enhance Zika virus infectionin vitro. West Nile virus antibodies in IGIVs could not neutralize Zika virus initially; however, once the IGIVs were concentrated further, they enhanced its infection. These findings lay the groundwork for exploring how preexisting WNV antibodies in IGIVs could impact Zika infection, bothin vitroandin vivo. Our observations are historically significant, since we tested a panel of IGIV lots that were carefully selected based on their production dates which covered two major flavivirus outbreaks in the U.S.: the WNV outbreak in 1999 and the ZIKV outbreak in 2015. These findings will facilitate our understanding of the interplay among closely related viral pathogens, particularly from a historical perspective regarding large blood donor populations. They should remain relevant for future outbreaks of emerging flaviviruses that may potentially affect vulnerable populations.
In this work, we present a supercomputer-driven pipeline for in silico drug discovery using enhanced sampling molecular dynamics (MD) and ensemble docking. Ensemble docking makes use of MD results by docking compound databases into representative protein binding-site conformations, thus taking into account the dynamic properties of the binding sites. We also describe preliminary results obtained for 24 systems involving eight proteins of the proteome of SARS-CoV-2. The MD involves temperature replica exchange enhanced sampling, making use of massively parallel supercomputing to quickly sample the configurational space of protein drug targets. Using the Summit supercomputer at the Oak Ridge National Laboratory, more than 1 ms of enhanced sampling MD can be generated per day. We have ensemble docked repurposing databases to 10 configurations of each of the 24 SARS-CoV-2 systems using AutoDock Vina. Comparison to experiment demonstrates remarkably high hit rates for the top scoring tranches of compounds identified by our ensemble approach. We also demonstrate that, using Autodock-GPU on Summit, it is possible to perform exhaustive docking of one billion compounds in under 24 h. Finally, we discuss preliminary results and planned improvements to the pipeline, including the use of quantum mechanical (QM), machine learning, and artificial intelligence (AI) methods to cluster MD trajectories and rescore docking poses.
Broadly neutralizing antibodies (bNAbs) have remarkable breadth and potency against most HIV-1 subtypes and are able to prevent HIV-1 infection in animal models. However, bNAbs are extremely difficult to induce by vaccination. Defining the developmental pathways towards neutralization breadth can assist in the design of strategies to elicit protective bNAb responses by vaccination. Here, HIV-1 envelope glycoproteins (Env)-specific IgG + B cells were isolated at various time points post infection from an HIV-1 infected elite neutralizer to obtain monoclonal antibodies (mAbs). Multiple antibody lineages were isolated targeting distinct epitopes on Env, including the gp120-gp41 interface, CD4-binding site, silent face and V3 region. The mAbs each neutralized a diverse set of HIV-1 strains from different clades indicating that the patient’s remarkable serum breadth and potency might have been the result of a polyclonal mixture rather than a single bNAb lineage. High-resolution cryo-electron microscopy structures of the neutralizing mAbs (NAbs) in complex with an Env trimer generated from the same individual revealed that the NAbs used multiple strategies to neutralize the virus; blocking the receptor binding site, binding to HIV-1 Env N-linked glycans, and disassembly of the trimer. These results show that diverse NAbs can complement each other to achieve a broad and potent neutralizing serum response in HIV-1 infected individuals. Hence, the induction of combinations of moderately broad NAbs might be a viable vaccine strategy to protect against a wide range of circulating HIV-1 viruses.
Photosystem II (PSII) is the water-splitting enzyme central to oxygenic photosynthesis. To drive water oxidation, light is harvested by accessory pigments, mostly chlorophyll (Chl) a molecules, which absorb visible light (400–700 nm). Some cyanobacteria facultatively acclimate to shaded environments by altering their photosynthetic machinery to additionally absorb far-red light (FRL, 700–800 nm), a process termed far-red light photoacclimation or FaRLiP. During far-red light photoacclimation, FRL-PSII is assembled with FRL-specific isoforms of the subunits PsbA, PsbB, PsbC, PsbD, and PsbH, and some Chl-binding sites contain Chls d or f instead of the usual Chl a. The structure of an apo-FRL-PSII monomer lacking the FRL-specific PsbH subunit has previously been determined, but visualization of the dimeric complex has remained elusive. Here, we report the cryo-EM structure of a dimeric FRL–PSII complex. The site assignments for Chls d and f are consistent with those assigned in the previous apo-FRL-PSII monomeric structure. All sites that bind Chl d or Chl f at high occupancy exhibit a FRL-specific interaction of the formyl moiety of the Chl d or Chl f with the protein environment, which in some cases involves a phenylalanine sidechain. The structure retains the FRL-specific PsbH2 subunit, which appears to alter the energetic landscape of FRL-PSII, redirecting energy transfer from the phycobiliprotein complex to a Chl f molecule bound by PsbB2 that acts as a bridge for energy transfer to the electron transfer chain. Collectively, these observations extend our previous understanding of the structure-function relationship that allows PSII to function using lower energy FRL.
Molecular understanding of neutralizing antibody responses to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) could accelerate vaccine design and drug discovery. We analyzed 294 anti–SARS-CoV-2 antibodies and found that immunoglobulin G heavy-chain variable region 3-53 (IGHV3-53) is the most frequently used IGHV gene for targeting the receptor-binding domain (RBD) of the spike protein. Co-crystal structures of two IGHV3-53–neutralizing antibodies with RBD, with or without Fab CR3022, at 2.33- to 3.20-angstrom resolution revealed that the germline-encoded residues dominate recognition of the angiotensin I converting enzyme 2 (ACE2)–binding site. This binding mode limits the IGHV3-53 antibodies to short complementarity-determining region H3 loops but accommodates light-chain diversity. These IGHV3-53 antibodies show minimal affinity maturation and high potency, which is promising for vaccine design. Knowledge of these structural motifs and binding mode should facilitate the design of antigens that elicit this type of neutralizing response.
The proline catabolic pathway consisting of proline dehydrogenase (PRODH) and L-glutamate-γ-semialdehyde (GSAL) dehydrogenase (GSALDH) catalyzes the four-electron oxidation of L-proline to L-glutamate. Chemical probes to these enzymes are of interest for their role in cancer and inherited metabolic disease. Here, we report the results of a crystallographic fragment-screening campaign targeting both enzymes. A unique aspect of our approach is the screening of both enzymes simultaneously using crystals of the bifunctional PRODH-GSALDH enzyme, proline utilization A (PutA). A 288-fragment library from Zenobia was screened in crystallo in cocktails of six fragments. Validation X-ray crystallography with individual fragments identified seven crystal hits distributed in the PRODH active site, GSALDH aldehyde substrate-binding site, and GSALDH NAD+ adenine-binding site. The fragment bound in the PRODH active site, 4-methoxybenzyl alcohol, is structurally distinct from all known PRODH inhibitors as it lacks an anionic anchor and stabilizes open conformations of the active site, motivating the study of eighteen analogs. In total, thirteen crystal structures with resolutions ranging from 1.32 Å to 1.80 Å were determined, resolving the poses and interactions of seven fragments from the Zenobia library and five analogs of 4-methoxybenzyl alcohol. These results expand the chemical space of probes targeting proline catabolic enzymes and provide new structural information for further inhibitor development.
Aldehyde dehydrogenases are versatile enzymes that serve a range of biochemical functions. Although traditionally considered metabolic housekeeping enzymes because of their ability to detoxify reactive aldehydes, like those generated from lipid peroxidation damage, the contributions of these enzymes to other biological processes are widespread. For example, the plant pathogen Pseudomonas syringae strain PtoDC3000 uses an indole-3-acetaldehyde dehydrogenase to synthesize the phytohormone indole-3-acetic acid to elude host responses. In this work we investigate the biochemical function of AldC from PtoDC3000. Analysis of the substrate profile of AldC suggests that this enzyme functions as a long-chain aliphatic aldehyde dehydrogenase. The 2.5 Å resolution X-ray crystal of the AldC C291A mutant in a dead-end complex with octanal and NAD + reveals an apolar binding site primed for aliphatic aldehyde substrate recognition. Functional characterization of site-directed mutants targeting the substrate- and NAD(H)-binding sites identifies key residues in the active site for ligand interactions, including those in the “aromatic box” that define the aldehyde-binding site. Overall, this study provides molecular insight for understanding the evolution of the prokaryotic aldehyde dehydrogenase superfamily and their diversity of function.
Effective gene regulation by the tetrahydrofolate riboswitch depends not only on ligand affinity but also on the kinetics of ligand association, which involves two cooperative binding sites. Here, we have determined a 1.9-Å resolution crystal structure of the ligand-free THF riboswitch aptamer. The pseudoknot binding site ‘unwinds’ in the absence of ligand, whereby the adjacent helical domains (P1, P2, and P3) become disjointed, resulting in rotation and misalignment of the gene-regulatory P1 helix with respect to P3. In contrast, the second binding site at the three-way junction, which is the first to fold, is structurally conserved between apo and holo forms. This suggests a kinetic role for this site, in which binding of the first ligand molecule to the stably folded three-way junction promotes formation of the regulatory pseudoknot site and subsequent binding of the second molecule. As such, these findings provide a molecular basis for both conformational switching and kinetic control.
Abstract New drug production, from target identification to marketing approval, takes over 12 years and can cost around $2.6 billion. Furthermore, the COVID-19 pandemic has unveiled the urgent need for more powerful computational methods for drug discovery. Here, we review the computational approaches to predicting protein–ligand interactions in the context of drug discovery, focusing on methods using artificial intelligence (AI). We begin with a brief introduction to proteins (targets), ligands (e.g. drugs) and their interactions for nonexperts. Next, we review databases that are commonly used in the domain of protein–ligand interactions. Finally, we survey and analyze the machine learning (ML) approaches implemented to predict protein–ligand binding sites, ligand-binding affinity and binding pose (conformation) including both classical ML algorithms and recent deep learning methods. After exploring the correlation between these three aspects of protein–ligand interaction, it has been proposed that they should be studied in unison. We anticipate that our review will aid exploration and development of more accurate ML-based prediction strategies for studying protein–ligand interactions.
Family 1 Coenzyme A transferases (CtfAB) from the extremely thermophilic bacterium, Thermosipho melanesiensis, has been used for in vivo acetone production up to 70°C. This enzyme has tentatively been identified as the rate-limiting step, due to its relatively low-binding affinity for acetate. However, existing kinetic and mechanistic studies on this enzyme are insufficient to evaluate this hypothesis. Here, kinetic analysis of purified recombinant T. melanesiensis CtfAB showed that it has a ping-pong bi-bi mechanism typical of Coenzyme A (CoA) transferases with Km values for acetate and acetoacetyl-CoA of 85 mM and 135 μM, respectively. Product inhibition by acetyl-CoA was competitive with respect to acetoacetyl-CoA and non-competitive with respect to acetate. Crystal structures of wild-type and mutant T. melanesiensis CtfAB were solved in the presence of acetate and in the presence or absence of acetyl-CoA. These structures led to a proposed structural basis for the competitive and non-competitive inhibition of acetyl-CoA: acetate binds independently of acetyl-CoA in an apparent low-affinity binding pocket in CtfA that is directly adjacent to a catalytic glutamate in CtfB. Similar to other CoA transferases, acetyl-CoA is bound in an apparent high-affinity binding site in CtfB with most interactions occurring between the phospho-ADP of CoA and CtfB residues far from the acetate binding pocket. This structural-based mechanism also explains the organic acid promiscuity of CtfAB. High-affinity interactions are predominantly between the conserved phospho-ADP of CoA, and the variable organic acid binding site is a low-affinity binding site with few specific interactions.