Targeted removal of the FA2 site on human albumin prevents fatty acid-mediated inhibition of Zn2+ binding.

Zinc is required for virtually all biological processes. In plasma, Zn2+ is predominantly transported by human serum albumin (HSA), which possesses two Zn2+-binding sites of differing affinities (sites A and B). Fatty acids (FAs) are also transported by HSA, with seven structurally characterized FA-binding sites (named FA1-FA7) known. FA binding inhibits Zn2+-HSA interactions, in a manner that can impact upon hemostasis and cellular zinc uptake, but the degree to which binding at specific FA sites contributes to this inhibition is unclear. Wild-type HSA and H9A, H67A, H247A, and Y150F/R257A/S287A (FA2-KO) mutant albumins were expressed in Pichia pastoris. Isothermal titration calorimetry studies revealed that the Zn2+-binding capacity at the high-affinity Zn2+ site (site A) was reduced in H67A and H247A mutants, with site B less affected. The H9A mutation decreased Zn2+ binding at the lower-affinity site, establishing His9 as a site B ligand. Zn2+ binding to HSA and H9A was compromised by palmitate, consistent with FA binding affecting site A. 13C-NMR experiments confirmed that the FA2-KO mutations prohibited FA binding at site FA2. Zn2+ binding to the FA2-KO mutant was unaffected by myristate, suggesting binding at FA2 is solely responsible for inhibition. Molecular dynamics studies identified the steric obstruction exerted by bound FA in site FA2, which impedes the conformational change from open (FA-loaded) to closed (FA-free) states, required for Zn2+ to bind at site A. The successful targeting of the FA2 site will aid functional studies exploring the interplay between circulating FA levels and plasma Zn2+ speciation in health and disease.

Structure, Mutagenesis, and QM:MM Modeling of 3-Ketosteroid Δ1-Dehydrogenase from Sterolibacterium denitrificans─The Role of a New Putative Membrane-Associated Domain and Proton-Relay System in Catalysis.

3-Ketosteroid Δ1-dehydrogenases (KstD) are important microbial flavin enzymes that initiate the metabolism of steroid ring A and find application in the synthesis of steroid drugs. We present a structure of the KstD from Sterolibacterium denitrificans (AcmB), which contains a previously uncharacterized putative membrane-associated domain and extended proton-relay system. The experimental and theoretical studies show that the steroid Δ1-dehydrogenation proceeds according to the Ping-Pong bi-bi kinetics and a two-step base-assisted elimination (E2cB) mechanism. The mechanism is validated by evaluating the experimental and theoretical kinetic isotope effect for deuterium-substituted substrates. The role of the active-site residues is quantitatively assessed by point mutations, experimental activity assays, and QM/MM MD modeling of the reductive half-reaction (RHR). The pre-steady-state kinetics also reveals that the low pH (6.5) optimum of AcmB is dictated by the oxidative half-reaction (OHR), while the RHR exhibits a slight optimum at the pH usual for the KstD family of 8.5. The modeling confirms the origin of the enantioselectivity of C2-H activation and substrate specificity for Δ4-3-ketosteroids. Finally, the cholest-4-en-3-one turns out to be the best substrate of AcmB in terms of ΔG of binding and predicted rate of dehydrogenation.

Recognizing and validating ligands with CheckMyBlob.

Structure-guided drug design depends on the correct identification of ligands in crystal structures of protein complexes. However, the interpretation of the electron density maps is challenging and often burdened with confirmation bias. Ligand identification can be aided by automatic methods such as CheckMyBlob, a machine learning algorithm that learns to generalize ligand descriptions from sets of moieties deposited in the Protein Data Bank. Here, we present the CheckMyBlob web server, a platform that can identify ligands in unmodeled fragments of electron density maps or validate ligands in existing models. The server processes PDB/mmCIF and MTZ files and returns a ranking of 10 most likely ligands for each detected electron density blob along with interactive 3D visualizations. Additionally, for each prediction/validation, a plugin script is generated that enables users to conduct a detailed analysis of the server results in Coot. The CheckMyBlob web server is available at https://checkmyblob.bioreproducibility.org.

Synchrotron Radiation as a Tool for Macromolecular X-Ray Crystallography: a XXI Century Perspective.

Intense X-rays available at powerful synchrotron beamlines provide macromolecular crystallographers with an incomparable tool for investigating biological phenomena on an atomic scale. The resulting insights into the mechanism's underlying biological processes have played an essential role and shaped biomedical sciences during the last 30 years, considered the "golden age" of structural biology. In this review, we analyze selected aspects of the impact of synchrotron radiation on structural biology. Synchrotron beamlines have been used to determine over 70% of all macromolecular structures deposited into the Protein Data Bank (PDB). These structures were deposited by over 13,000 different research groups. Interestingly, despite the impressive advances in synchrotron technologies, the median resolution of macromolecular structures determined using synchrotrons has remained constant throughout the last 30 years, at about 2 Å. Similarly, the median times from the data collection to the deposition and release have not changed significantly. We describe challenges to reproducibility related to recording all relevant data and metadata during the synchrotron experiments, including diffraction images. Finally, we discuss some of the recent opinions suggesting a diminishing importance of X-ray crystallography due to impressive advances in Cryo-EM and theoretical modeling. We believe that synchrotrons of the future will increasingly evolve towards a life science center model, where X-ray crystallography, Cryo-EM, and other experimental and computational resources and knowledge are encompassed within a versatile research facility. The recent response of crystallographers to the COVID-19 pandemic suggests that X-ray crystallography conducted at synchrotron beamlines will continue to play an essential role in structural biology and drug discovery for years to come.

Detecting anomalies in X-ray diffraction images using convolutional neural networks.

Our understanding of life is based upon the interpretation of macromolecular structures and their dynamics. Almost 90% of currently known macromolecular models originated from electron density maps constructed using X-ray diffraction images. Even though diffraction images are critical for structure determination, due to their vast amounts and noisy, non-intuitive nature, their quality is rarely inspected. In this paper, we use recent advances in machine learning to automatically detect seven types of anomalies in X-ray diffraction images. For this purpose, we utilize a novel X-ray beam center detection algorithm, propose three different image representations, and compare the predictive performance of general-purpose classifiers and deep convolutional neural networks (CNNs). In benchmark tests on a set of 6,311 X-ray diffraction images, the proposed CNN achieved between 87% and 99% accuracy depending on the type of anomaly. Experimental results show that the proposed anomaly detection system can be considered suitable for early detection of sub-optimal data collection conditions and malfunctions at X-ray experimental stations.

Structure of the Complex of an Iminopyridinedione Protein Tyrosine Phosphatase 4A3 Phosphatase Inhibitor with Human Serum Albumin.

Protein tyrosine phosphatase (PTP) 4A3 is frequently overexpressed in human solid tumors and hematologic malignancies and is associated with tumor cell invasion, metastasis, and a poor patient prognosis. Several potent, selective, and allosteric small molecule inhibitors of PTP4A3 were recently identified. A lead compound in the series, JMS-053 (7-imino-2-phenylthieno[3,2-c]pyridine-4,6(5H,7H)-dione), has a long plasma half-life (∼ 24 hours) in mice, suggesting possible binding to serum components. We confirmed by isothermal titration calorimetry that JMS-053 binds to human serum albumin. A single JMS-053 binding site was identified by X-ray crystallography in human serum albumin at drug site 3, which is also known as subdomain IB. The binding of JMS-053 to human serum albumin, however, did not markedly alter the overall albumin structure. In the presence of serum albumin, the potency of JMS-053 as an in vitro inhibitor of PTP4A3 and human A2780 ovarian cancer cell growth was reduced. The reversible binding of JMS-053 to serum albumin may serve to increase JMS-053's plasma half-life and thus extend the delivery of the compound to tumors. SIGNIFICANCE STATEMENT: X-ray crystallography revealed that a potent, reversible, first-in-class small molecule inhibitor of the oncogenic phosphatase protein tyrosine phosphatase 4A3 binds to at least one site on human serum albumin, which is likely to extend the compound's plasma half-life and thus assist in drug delivery into tumors.

Automatic recognition of ligands in electron density by machine learning.

Motivation: The correct identification of ligands in crystal structures of protein complexes is the cornerstone of structure-guided drug design. However, cognitive bias can sometimes mislead investigators into modeling fictitious compounds without solid support from the electron density maps. Ligand identification can be aided by automatic methods, but existing approaches are based on time-consuming iterative fitting. Results: Here we report a new machine learning algorithm called CheckMyBlob that identifies ligands from experimental electron density maps. In benchmark tests on portfolios of up to 219 931 ligand binding sites containing the 200 most popular ligands found in the Protein Data Bank, CheckMyBlob markedly outperforms the existing automatic methods for ligand identification, in some cases doubling the recognition rates, while requiring significantly less time. Our work shows that machine learning can improve the automation of structure modeling and significantly accelerate the drug screening process of macromolecule-ligand complexes. Availability and implementation: Code and data are available on GitHub at https://github.com/dabrze/CheckMyBlob. Supplementary information: Supplementary data are available at Bioinformatics online.

Antigenic Determinants of Der p 1: Specificity and Cross-Reactivity Associated with IgE Antibody Recognition.

Der p 1 and Der f 1 are major allergens from Dermatophagoides pteronyssinus and D. farinae, respectively. An analysis of antigenic determinants on both allergens was performed by site-directed mutagenesis. The analysis was based on the x-ray crystal structures of the allergens in complex with Fab fragments of three murine mAbs that interfere with IgE Ab binding: the two Der p 1-specific mAbs 5H8 and 10B9, and the cross-reactive mAb 4C1. On one hand, selected residues in the epitopes for mAb 5H8 and mAb 4C1 were substituted with amino acids that resulted in impaired Ab binding to Der p 1. On the other hand, an epitope for the Der p 1-specific mAb 10B9, which partially overlaps with mAb 4C1, was created in Der f 1. The mutation of 1-3 aa residues in Der f 1 was sufficient to bind mAb 10B9. These residues form hydrogen bonds with CDRs of the Ab other than H CDR3. This observation unveils an exception to the dominant role of H CDR3 commonly observed in Ag recognition. Overall, this study resulted in the identification of important residues for mAb and IgE Ab recognition in group 1 mite allergens. This information can be used to engineer allergen mutants with reduced IgE Ab binding for immunotherapy.

Correction to: Classification, substrate specificity and structural features of D-2-hydroxyacid dehydrogenases: 2HADH knowledgebase.

Following publication of the original article [1], we have been notified that some important information was omitted by the authors from the Competing interests section. The declaration should read as below.

A close look onto structural models and primary ligands of metallo-ß-lactamases.

ß-Lactamases are hydrolytic enzymes capable of opening the ß-lactam ring of antibiotics such as penicillin, thus endowing the bacteria that produce them with antibiotic resistance. Of particular medical concern are metallo-ß-lactamases (MBLs), with an active site built around coordinated Zn cations. MBLs are pan-reactive enzymes that can break down almost all classes of ß-lactams, including such last-resort antibiotics as carbapenems. They are not only broad-spectrum-reactive but are often plasmid-borne (e.g., the New Delhi enzyme, NDM), and can spread horizontally even among unrelated bacteria. Acquired MBLs are encoded by mobile genetic elements, which often include other resistance genes, making the microbiological situation particularly alarming. There is an urgent need to develop MBL inhibitors in order to rescue our antibiotic armory. A number of such efforts have been undertaken, most notably using the 3D structures of various MBLs as drug-design targets. Structure-guided drug discovery depends on the quality of the structures that are collected in the Protein Data Bank (PDB) and on the consistency of the information in dedicated ß-lactamase databases. We conducted a careful review of the crystal structures of class B ß-lactamases, concluding that the quality of these structures varies widely, especially in the regions where small molecules interact with the macromolecules. In a number of examples the interpretation of the bound ligands (e.g., inhibitors, substrate/product analogs) is doubtful or even incorrect, and it appears that in some cases the modeling of ligands was not supported by electron density. For ten MBL structures, alternative interpretations of the original diffraction data could be proposed and the new models have been deposited in the PDB. In four cases, these models, prepared jointly with the authors of the original depositions, superseded the previous deposits. This review emphasizes the importance of critical assessment of structural models describing key drug design targets at the level of the raw experimental data. Since the structures reviewed here are the basis for ongoing design of new MBL inhibitors, it is important to identify and correct the problems with ambiguous crystallographic interpretations, thus enhancing reproducibility in this highly medically relevant area.

A Gcn5-Related N-Acetyltransferase (GNAT) Capable of Acetylating Polymyxin B and Colistin Antibiotics in Vitro.

Deeper exploration of uncharacterized Gcn5-related N-acetyltransferases has the potential to expand our knowledge of the types of molecules that can be acylated by this important superfamily of enzymes and may offer new opportunities for biotechnological applications. While determining native or biologically relevant in vivo functions of uncharacterized proteins is ideal, their alternative or promiscuous in vitro capabilities provide insight into key active site interactions. Additionally, this knowledge can be exploited to selectively modify complex molecules and reduce byproducts when synthetic routes become challenging. During our exploration of uncharacterized Gcn5-related N-acetyltransferases from Pseudomonas aeruginosa, we identified such an example. We found that the PA3944 enzyme acetylates both polymyxin B and colistin on a single diaminobutyric acid residue closest to the macrocyclic ring of the antimicrobial peptide and determined the PA3944 crystal structure. This finding is important for several reasons. (1) To the best of our knowledge, this is the first report of enzymatic acylation of polymyxins and thus reveals a new type of substrate that this enzyme family can use. (2) The enzymatic acetylation offers a controlled method for antibiotic modification compared to classical promiscuous chemical methods. (3) The site of acetylation would reduce the overall positive charge of the molecule, which is important for reducing nephrotoxic effects and may be a salvage strategy for this important class of antibiotics. While the physiological substrate for this enzyme remains unknown, our structural and functional characterization of PA3944 offers insight into its unique noncanonical substrate specificity.

Structural, Biochemical, and Evolutionary Characterizations of Glyoxylate/Hydroxypyruvate Reductases Show Their Division into Two Distinct Subfamilies.

The d-2-hydroxyacid dehydrogenase (2HADH) family illustrates a complex evolutionary history with multiple lateral gene transfers and gene duplications and losses. As a result, the exact functional annotation of individual members can be extrapolated to a very limited extent. Here, we revise the previous simplified view on the classification of the 2HADH family; specifically, we show that the previously delineated glyoxylate/hydroxypyruvate reductase (GHPR) subfamily consists of two evolutionary separated GHRA and GHRB subfamilies. We compare two representatives of these subfamilies from Sinorhizobium meliloti (SmGhrA and SmGhrB), employing a combination of biochemical, structural, and bioinformatics approaches. Our kinetic results show that both enzymes reduce several 2-ketocarboxylic acids with overlapping, but not equivalent, substrate preferences. SmGhrA and SmGhrB show highest activity with glyoxylate and hydroxypyruvate, respectively; in addition, only SmGhrB reduces 2-keto-d-gluconate, and only SmGhrA reduces pyruvate (with low efficiency). We present nine crystal structures of both enzymes in apo forms and in complexes with cofactors and substrates/substrate analogues. In particular, we determined a crystal structure of SmGhrB with 2-keto-d-gluconate, which is the biggest substrate cocrystallized with a 2HADH member. The structures reveal significant differences between SmGhrA and SmGhrB, both in the overall structure and within the substrate-binding pocket, offering insight into the molecular basis for the observed substrate preferences and subfamily differences. In addition, we provide an overview of all GHRA and GHRB structures complexed with a ligand in the active site.

Classification, substrate specificity and structural features of D-2-hydroxyacid dehydrogenases: 2HADH knowledgebase.

BACKGROUND: The family of D-isomer specific 2-hydroxyacid dehydrogenases (2HADHs) contains a wide range of oxidoreductases with various metabolic roles as well as biotechnological applications. Despite a vast amount of biochemical and structural data for various representatives of the family, the long and complex evolution and broad sequence diversity hinder functional annotations for uncharacterized members. RESULTS: We report an in-depth phylogenetic analysis, followed by mapping of available biochemical and structural data on the reconstructed phylogenetic tree. The analysis suggests that some subfamilies comprising enzymes with similar yet broad substrate specificity profiles diverged early in the evolution of 2HADHs. Based on the phylogenetic tree, we present a revised classification of the family that comprises 22 subfamilies, including 13 new subfamilies not studied biochemically. We summarize characteristics of the nine biochemically studied subfamilies by aggregating all available sequence, biochemical, and structural data, providing comprehensive descriptions of the active site, cofactor-binding residues, and potential roles of specific structural regions in substrate recognition. In addition, we concisely present our analysis as an online 2HADH enzymes knowledgebase. CONCLUSIONS: The knowledgebase enables navigation over the 2HADHs classification, search through collected data, and functional predictions of uncharacterized 2HADHs. Future characterization of the new subfamilies may result in discoveries of enzymes with novel metabolic roles and with properties beneficial for biotechnological applications.

A transient post-translational modification of active site cysteine alters binding properties of the parkinsonism protein DJ-1.

Mutations in the human protein DJ-1 cause early onset of Parkinson's disease. A reactive cysteine residue (Cys106) of DJ-1 is crucial for its protective function, although the underlying mechanisms are unclear. Here we show that a fraction of bacterially expressed polyhistidine-tagged human DJ-1 could not be eluted from a Ni-nitrilotriacetate (Ni-NTA) column with 150 mM imidazole. This unusually tight binding was accompanied by the appearance of blue violet color on the Ni-NTA column. We demonstrate by X-ray crystallography that Cys106 is carboxymethylated in a fraction of DJ-1 tightly bound to Ni-NTA and that the replacement of Cys106 by serine abrogates the tight binding and the appearance of blue violet color. However, carboxymethylation of purified DJ-1 is insufficient to confer the tight binding to Ni-NTA. Moreover, when eluted protein was re-applied to the Ni-NTA column, no tight binding was observed, indicating that the formation of high affinity complex with Ni-NTA depends on a transient modification of Cys106 that transforms into a Cys106-carboxymethyl adduct upon elution from Ni-NTA. We conclude that an unknown metabolite reacts with Cys106 of DJ-1 to result in a transient post-translational modification. This modification is distinct from simple oxidation to sulfinic or sulfenic acids and confers altered binding properties to DJ-1 suggesting that it could serve as a signal for sensing oxidant stress.

On the Structure and Reaction Mechanism of Human Acireductone Dioxygenase.

Acireductone dioxygenase (ARD) is an intriguing enzyme from the methionine salvage pathway that is capable of catalysing two different oxidation reactions with the same substrate depending on the type of the metal ion in the active site. To date, the structural information regarding the ARD-acireductone complex is limited and possible reaction mechanisms are still under debate. The results of joint experimental and computational studies undertaken to advance knowledge about ARD are reported. The crystal structure of an ARD from Homo sapiens was determined with selenomethionine. EPR spectroscopy suggested that binding acireductone triggers one protein residue to dissociate from Fe2+ , which allows NO (and presumably O2 ) to bind directly to the metal. Mössbauer spectroscopic data (interpreted with the aid of DFT calculations) was consistent with bidentate binding of acireductone to Fe2+ and concomitant dissociation of His88 from the metal. Major features of Fe vibrational spectra obtained for the native enzyme and upon addition of acireductone were reproduced by QM/MM calculations for the proposed models. A computational (QM/MM) study of the reaction mechanisms suggests that Fe2+ promotes O-O bond homolysis, which elicits cleavage of the C1-C2 bond of the substrate. Higher M3+ /M2+ redox potentials of other divalent metals do not support this pathway, and instead the reaction proceeds similarly to the key reaction step in the quercetin 2,3-dioxygenase mechanism.

Refining the macromolecular model - achieving the best agreement with the data from X-ray diffraction experiment.

Refinement of macromolecular X-ray crystal structures involves using complex software with hundreds of different settings. The complexity of underlying concepts and the sheer amount sof instructions may make it difficult for less experienced crystallographers to achieve optimal results in their refinements. This tutorial review offers guidelines for choosing the best settings for the reciprocal-space refinement of macromolecular models and provides practical tips for manual model correction. To help aspiring crystallographers navigate the process, some of the most practically important concepts of protein structure refinement are described. Among the topics covered are the use and purpose of R-free, geometrical restraints, restraints on atomic displacement parameters (ADPs), refinement weights, various parametrizations of ADPs (full anisotropic refinement and TLS), and omit maps. We also give practical tips for manual model correction in Coot, modelling of side-chains with poor or missing density, and ligand identification, fitting, and refinement.

Insight into the 3D structure and substrate specificity of previously uncharacterized GNAT superfamily acetyltransferases from pathogenic bacteria.

Members of the Gcn5-related N-acetyltransferase (GNAT) superfamily catalyze the acetylation of a wide range of small molecule and protein substrates. Due to their abundance in all kingdoms of life and diversity of their functions, they are implicated in many aspects of eukaryotic and prokaryotic physiology. Although numerous GNATs have been identified thus far, many remain structurally and functionally uncharacterized. The elucidation of their structures and functions is critical for broadening our knowledge of this diverse and important superfamily. In this work, we present the structural and kinetic analyses of two previously uncharacterized bacterial acetyltransferases - SACOL1063 from Staphylococcus aureus strain COL and CD1211 from Clostridium difficile strain 630. Our structures of SACOL1063 show substantial flexibility of a loop that is likely responsible for substrate recognition and binding compared to structures of other homologs. In the CoA complex structure, we found two CoA molecules bound in both the canonical AcCoA/CoA-binding site and the acceptor-substrate-binding site. Our work also provides initial clues regarding the substrate specificity of these two enzymes; however, their native function(s) remain unknown. We found both proteins act as N- rather than O-acetyltransferases and preferentially acetylate l-threonine. The combination of structural and kinetic analyses of these two previously uncharacterized GNATs provides fundamental knowledge and a framework on which future studies can be built to elucidate their native functions.

Optimization of overexpression of a chaperone protein of steroid C25 dehydrogenase for biochemical and biophysical characterization.

Molybdenum is an essential nutrient for metabolism in plant, bacteria, and animals. Molybdoenzymes are involved in nitrogen assimilation and oxidoreductive detoxification, and bioconversion reactions of environmental, industrial, and pharmaceutical interest. Molybdoenzymes contain a molybdenum cofactor (Moco), which is a pyranopterin heterocyclic compound that binds a molybdenum atom via a dithiolene group. Because Moco is a large and complex compound deeply buried within the protein, molybdoenzymes are accompanied by private chaperone proteins responsible for the cofactor's insertion into the enzyme and the enzyme's maturation. An efficient recombinant expression and purification of both Moco-free and Moco-containing molybdoenzymes and their chaperones is of paramount importance for fundamental and applied research related to molybdoenzymes. In this work, we focused on a D1 protein annotated as a chaperone of steroid C25 dehydrogenase (S25DH) from Sterolibacterium denitrificans Chol-1S. The D1 protein is presumably involved in the maturation of S25DH engaged in oxygen-independent oxidation of sterols. As this chaperone is thought to be a crucial element that ensures the insertion of Moco into the enzyme and consequently, proper folding of S25DH optimization of the chaperon's expression is the first step toward the development of recombinant expression and purification methods for S25DH. We have identified common E. coli strains and conditions for both expression and purification that allow us to selectively produce Moco-containing and Moco-free chaperones. We have also characterized the Moco-containing chaperone by EXAFS and HPLC analysis and identified conditions that stabilize both forms of the protein. The protocols presented here are efficient and result in protein quantities sufficient for biochemical studies.

Structural Analysis of Der p 1-Antibody Complexes and Comparison with Complexes of Proteins or Peptides with Monoclonal Antibodies.

Der p 1 is a major allergen from the house dust mite, Dermatophagoides pteronyssinus, that belongs to the papain-like cysteine protease family. To investigate the antigenic determinants of Der p 1, we determined two crystal structures of Der p 1 in complex with the Fab fragments of mAbs 5H8 or 10B9. Epitopes for these two Der p 1-specific Abs are located in different, nonoverlapping parts of the Der p 1 molecule. Nevertheless, surface area and identity of the amino acid residues involved in hydrogen bonds between allergen and Ab are similar. The epitope for mAb 10B9 only showed a partial overlap with the previously reported epitope for mAb 4C1, a cross-reactive mAb that binds Der p 1 and its homolog Der f 1 from Dermatophagoides farinae. Upon binding to Der p 1, the Fab fragment of mAb 10B9 was found to form a very rare α helix in its third CDR of the H chain. To provide an overview of the surface properties of the interfaces formed by the complexes of Der p 1-10B9 and Der p 1-5H8, along with the complexes of 4C1 with Der p 1 and Der f 1, a broad analysis of the surfaces and hydrogen bonds of all complexes of Fab-protein or Fab-peptide was performed. This work provides detailed insight into the cross-reactive and specific allergen-Ab interactions in group 1 mite allergens. The surface data of Fab-protein and Fab-peptide interfaces can be used in the design of conformational epitopes with reduced Ab binding for immunotherapy.

Magnesium-binding architectures in RNA crystal structures: validation, binding preferences, classification and motif detection.

The ubiquitous presence of magnesium ions in RNA has long been recognized as a key factor governing RNA folding, and is crucial for many diverse functions of RNA molecules. In this work, Mg(2+)-binding architectures in RNA were systematically studied using a database of RNA crystal structures from the Protein Data Bank (PDB). Due to the abundance of poorly modeled or incorrectly identified Mg(2+) ions, the set of all sites was comprehensively validated and filtered to identify a benchmark dataset of 15 334 'reliable' RNA-bound Mg(2+) sites. The normalized frequencies by which specific RNA atoms coordinate Mg(2+) were derived for both the inner and outer coordination spheres. A hierarchical classification system of Mg(2+) sites in RNA structures was designed and applied to the benchmark dataset, yielding a set of 41 types of inner-sphere and 95 types of outer-sphere coordinating patterns. This classification system has also been applied to describe six previously reported Mg(2+)-binding motifs and detect them in new RNA structures. Investigation of the most populous site types resulted in the identification of seven novel Mg(2+)-binding motifs, and all RNA structures in the PDB were screened for the presence of these motifs.