Capturing time-dependent activation of genes and stress-response pathways using transcriptomics in iPSC-derived renal proximal tubule cells.

Transcriptomic analysis is a powerful method in the utilization of New Approach Methods (NAMs) for identifying mechanisms of toxicity and application to hazard characterization. With this regard, mapping toxicological events to time of exposure would be helpful to characterize early events. Here, we investigated time-dependent changes in gene expression levels in iPSC-derived renal proximal tubular-like cells (PTL) treated with five diverse compounds using TempO-Seq transcriptomics with the aims to evaluate the application of PTL for toxicity prediction and to report on temporal effects for the activation of cellular stress response pathways. PTL were treated with either 50 µM amiodarone, 10 µM sodium arsenate, 5 nM rotenone, or 300 nM tunicamycin over a temporal time course between 1 and 24 h. The TGFß-type I receptor kinase inhibitor GW788388 (1 µM) was used as a negative control. Pathway analysis revealed the induction of key stress-response pathways, including Nrf2 oxidative stress response, unfolding protein response, and metal stress response. Early response genes per pathway were identified much earlier than 24 h and included HMOX1, ATF3, DDIT3, and several MT1 isotypes. GW788388 did not induce any genes within the stress response pathways above, but showed deregulation of genes involved in TGFß inhibition, including downregulation of CYP24A1 and SERPINE1 and upregulation of WT1. This study highlights the application of iPSC-derived renal cells for prediction of cellular toxicity and sheds new light on the temporal and early effects of key genes that are involved in cellular stress response pathways.

Binding of HasA by its transmembrane receptor HasR follows a conformational funnel mechanism.

HasR in the outer membrane of Serratia marcescens binds secreted, heme-loaded HasA and translocates the heme to the periplasm to satisfy the cell's demand for iron. The previously published crystal structure of the wild-type complex showed HasA in a very specific binding arrangement with HasR, apt to relax the grasp on the heme and assure its directed transfer to the HasR-binding site. Here, we present a new crystal structure of the heme-loaded HasA arranged with a mutant of HasR, called double mutant (DM) in the following that seemed to mimic a precursor stage of the abovementioned final arrangement before heme transfer. To test this, we performed first molecular dynamics (MD) simulations starting at the crystal structure of the complex of HasA with the DM mutant and then targeted MD simulations of the entire binding process beginning with heme-loaded HasA in solution. When the simulation starts with the former complex, the two proteins in most simulations do not dissociate. When the mutations are reverted to the wild-type sequence, dissociation and development toward the wild-type complex occur in most simulations. This indicates that the mutations create or enhance a local energy minimum. In the targeted MD simulations, the first protein contacts depend upon the chosen starting position of HasA in solution. Subsequently, heme-loaded HasA slides on the external surface of HasR on paths that converge toward the specific arrangement apt for heme transfer. The targeted simulations end when HasR starts to relax the grasp on the heme, the subsequent events being in a time regime inaccessible to the available computing power. Interestingly, none of the ten independent simulation paths visits exactly the arrangement of HasA with HasR seen in the crystal structure of the mutant. Two factors which do not exclude each other could explain these observations: the double mutation creates a non-physiologic potential energy minimum between the two proteins and /or the target potential in the simulation pushes the system along paths deviating from the low-energy paths of the native binding processes. Our results support the former view, but do not exclude the latter possibility.

The EU-ToxRisk method documentation, data processing and chemical testing pipeline for the regulatory use of new approach methods.

Hazard assessment, based on new approach methods (NAM), requires the use of batteries of assays, where individual tests may be contributed by different laboratories. A unified strategy for such collaborative testing is presented. It details all procedures required to allow test information to be usable for integrated hazard assessment, strategic project decisions and/or for regulatory purposes. The EU-ToxRisk project developed a strategy to provide regulatorily valid data, and exemplified this using a panel of > 20 assays (with > 50 individual endpoints), each exposed to 19 well-known test compounds (e.g. rotenone, colchicine, mercury, paracetamol, rifampicine, paraquat, taxol). Examples of strategy implementation are provided for all aspects required to ensure data validity: (i) documentation of test methods in a publicly accessible database; (ii) deposition of standard operating procedures (SOP) at the European Union DB-ALM repository; (iii) test readiness scoring accoding to defined criteria; (iv) disclosure of the pipeline for data processing; (v) link of uncertainty measures and metadata to the data; (vi) definition of test chemicals, their handling and their behavior in test media; (vii) specification of the test purpose and overall evaluation plans. Moreover, data generation was exemplified by providing results from 25 reporter assays. A complete evaluation of the entire test battery will be described elsewhere. A major learning from the retrospective analysis of this large testing project was the need for thorough definitions of the above strategy aspects, ideally in form of a study pre-registration, to allow adequate interpretation of the data and to ensure overall scientific/toxicological validity.

A semi-automated workflow for adverse outcome pathway hypothesis generation: The use case of non-genotoxic induced hepatocellular carcinoma.

The utility of the Adverse Outcome Pathway (AOP) concept has been largely recognized by scientists, however, the AOP generation is still mainly done manually by screening through evidence and extracting probable associations. To accelerate this process and increase the reliability, we have developed an semi-automated workflow for AOP hypothesis generation. In brief, association mining methods were applied to high-throughput screening, gene expression, in vivo and disease data present in ToxCast and Comparative Toxicogenomics Database. This was supplemented by pathway mapping using Reactome to fill in gaps and identify events occurring at the cellular/tissue levels. Furthermore, in vivo data from TG-Gates was integrated to finally derive a gene, pathway, biochemical, histopathological and disease network from which specific disease sub-networks can be queried. To test the workflow, non-genotoxic-induced hepatocellular carcinoma (HCC) was selected as a case study. The implementation resulted in the identification of several non-genotoxic-specific HCC-connected genes belonging to cell proliferation, endoplasmic reticulum stress and early apoptosis. Biochemical findings revealed non-genotoxic-specific alkaline phosphatase increase. The explored non-genotoxic-specific histopathology was associated with early stages of hepatic steatosis, transforming into cirrhosis. This work illustrates the utility of computationally predicted constructs in supporting development by using pre-existing knowledge in a fast and unbiased manner.

Skin sensitization in silico protocol.

The assessment of skin sensitization has evolved over the past few years to include in vitro assessments of key events along the adverse outcome pathway and opportunistically capitalize on the strengths of in silico methods to support a weight of evidence assessment without conducting a test in animals. While in silico methods vary greatly in their purpose and format; there is a need to standardize the underlying principles on which such models are developed and to make transparent the implications for the uncertainty in the overall assessment. In this contribution, the relationship between skin sensitization relevant effects, mechanisms, and endpoints are built into a hazard assessment framework. Based on the relevance of the mechanisms and effects as well as the strengths and limitations of the experimental systems used to identify them, rules and principles are defined for deriving skin sensitization in silico assessments. Further, the assignments of reliability and confidence scores that reflect the overall strength of the assessment are discussed. This skin sensitization protocol supports the implementation and acceptance of in silico approaches for the prediction of skin sensitization.

Predicting the Magnitude of σ-Holes Using VmaxPred, a Fast and Efficient Tool Supporting the Application of Halogen Bonds in Drug Discovery.

Halogen bonding as a modern molecular interaction has received increasing attention not only in materials sciences but also in biological systems and drug discovery. Thus, there is a growing demand for fast, efficient, and easily applicable tailor-made tools supporting the use of halogen bonds in molecular design and medicinal chemistry. The potential strength of a halogen bond is dependent on several properties of the σ-hole donor, e.g., a (hetero)aryl halide, and the σ-hole acceptor, a nucleophile with n or π electron density. Besides the influence of the interaction geometry and the type of acceptor, significant tuning effects on the magnitude of the σ-hole can be observed, caused by different (hetero)aromatic scaffolds and their substitution patterns. The most positive electrostatic potential on the isodensity surface ( Vmax), representing the σ-hole, has been widely used as the standard descriptor for the magnitude of the σ-hole and the strength of the halogen bond. Calculation of Vmax using quantum-mechanical methods at a reasonable level of theory is time-consuming and thus not applicable for larger numbers of compounds in drug discovery projects. Herein we present a tool for the prediction of this descriptor based on a machine-learned model with a speedup of 5 to 6 orders of magnitude relative to MP2 quantum-mechanical calculations. According to the test set, the squared correlation coefficient is greater than 0.94.

Scaffold Effects on Halogen Bonding Strength.

Halogen bonds have become increasingly popular interactions in molecular design and drug discovery. One of the key features is the strong dependence of the size and magnitude of the halogen's σ-hole on the chemical environment of the ligand. The term σ-hole refers to a region of lower electronic density opposite to a covalent bond, e.g., the C-X bond. It is typically (but not always) associated with a positive electrostatic potential in close proximity to the extension of the covalent bond. Herein, we use a variety of 30 nitrogen-bearing heterocycles, halogenated systematically by chlorine, bromine, or iodine, yielding 468 different ligands that are used to exemplify scaffold effects on halogen bonding strength. As a template interaction partner, we have chosen N-methylacetamide representing the ubiquitously present protein backbone. Adduct formation energies were obtained at a MP2/TZVPP level of theory. We used the local maximum of the electrostatic potential on the molecular surface in close proximity to the σ-hole, V S,max, as a descriptor for the magnitude of the positive electrostatic potential characterizing the tuning of the σ-hole. Free optimization of the complexes gave reasonable correlations with V S,max but was found to be of limited use because considerable numbers of chlorinated and brominated ligands lost their halogen bond or showed significant secondary interactions. Thus, starting from a close to optimal geometry of the halogen bond, we used distance scans to obtain the best adduct formation energy for each complex. This approach provided superior results for all complexes exhibiting correlations with R2 > 0.96 for each individual halogen. We evaluated the dependence of V S,max from the molecular surface onto which the positive electrostatic potential is projected, altering the isodensity values from 0.001 au to 0.050 au. Interestingly, the best overall fit using a third-order polynomial function (R2 = 0.99, RMSE = 0.562 kJ/mol) with rather smooth transitions between all halogens was obtained for V S,max calculated from an isodensity surface at 0.014 au.

Effect of an activated charcoal product (DOAC Stop™) intended for extracting DOACs on various other APTT-prolonging anticoagulants.

Background The aim of the study was to investigate the specificity of an activated charcoal-based product (DOAC Stop™) initially intended for the specific extraction of direct oral anticoagulants (DOACs) from test plasmas on a range of other anticoagulants. Methods Test plasmas were prepared by adding various anticoagulants to pooled normal plasma at concentrations prolonging an activated partial thromboplastin time (APTT) test by a factor of 1.5-3. These plasmas were treated with DOAC Stop™ for 5 and 20 min. Then APTTs were repeated and residual anticoagulant concentrations estimated from dose-response curves. Results The activated charcoal (AC)-based product was found to extract DOACs efficiently. It also bound the intravenous anticoagulants argatroban and lepirudin, but it had no effect on heparin, enoxaparin or danaparoid in plasma. Among other APTT-inhibiting agents that might be present in test plasmas from patients, it extracted protamine, aprotinin and polymyxin. It had no effect on annexin V, thrombomodulin, a typical lupus anticoagulant, a factor VIII antibody, activated protein C or its activator, but it did bind some cationic inhibitors of the APTT with molecular weight below approximately 30 kDa. Conclusions The AC-based product extracted DOACs efficiently with no effect on heparin-type anticoagulants. It did bind argatroban and hirudin-type anticoagulants, which might occur in plasmas from some inpatients, and APTT results obtained after its use should be interpreted after due consideration of patient medications.

Nanodiscs for INPHARMA NMR Characterization of GPCRs: Ligand Binding to the Human A2A Adenosine Receptor.

G-protein-coupled-receptors (GPCRs) are of fundamental importance for signal transduction through cell membranes. This makes them important drug targets, but structure-based drug design (SBDD) is still hampered by the limitations for structure determination of unmodified GPCRs. We show that the interligand NOEs for pharmacophore mapping (INPHARMA) method can provide valuable information on ligand poses inside the binding site of the unmodified human A2A adenosine receptor reconstituted in nanodiscs. By comparing experimental INPHARMA spectra with back-calculated spectra based on ligand poses obtained from molecular dynamics simulations, a complex structure for A2A R with the low-affinity ligand 3-pyrrolidin-1-ylquinoxalin-2-amine was determined based on the X-ray structure of ligand ZM-241,358 in complex with a modified A2A R.

Crystal Structure of the N-Acetylmuramic Acid α-1-Phosphate (MurNAc-α1-P) Uridylyltransferase MurU, a Minimal Sugar Nucleotidyltransferase and Potential Drug Target Enzyme in Gram-negative Pathogens.

The N-acetylmuramic acid α-1-phosphate (MurNAc-α1-P) uridylyltransferase MurU catalyzes the synthesis of uridine diphosphate (UDP)-MurNAc, a crucial precursor of the bacterial peptidoglycan cell wall. MurU is part of a recently identified cell wall recycling pathway in Gram-negative bacteria that bypasses the general de novo biosynthesis of UDP-MurNAc and contributes to high intrinsic resistance to the antibiotic fosfomycin, which targets UDP-MurNAc de novo biosynthesis. To provide insights into substrate binding and specificity, we solved crystal structures of MurU of Pseudomonas putida in native and ligand-bound states at high resolution. With the help of these structures, critical enzyme-substrate interactions were identified that enable tight binding of MurNAc-α1-P to the active site of MurU. The MurU structures define a "minimal domain" required for general nucleotidyltransferase activity. They furthermore provide a structural basis for the chemical design of inhibitors of MurU that could serve as novel drugs in combination therapy against multidrug-resistant Gram-negative pathogens.

Automated assignment of NMR chemical shifts based on a known structure and 4D spectra.

Apart from their central role during 3D structure determination of proteins the backbone chemical shift assignment is the basis for a number of applications, like chemical shift perturbation mapping and studies on the dynamics of proteins. This assignment is not a trivial task even if a 3D protein structure is known and needs almost as much effort as the assignment for structure prediction if performed manually. We present here a new algorithm based solely on 4D [(1)H,(15)N]-HSQC-NOESY-[(1)H,(15)N]-HSQC spectra which is able to assign a large percentage of chemical shifts (73-82 %) unambiguously, demonstrated with proteins up to a size of 250 residues. For the remaining residues, a small number of possible assignments is filtered out. This is done by comparing distances in the 3D structure to restraints obtained from the peak volumes in the 4D spectrum. Using dead-end elimination, assignments are removed in which at least one of the restraints is violated. Including additional information from chemical shift predictions, a complete unambiguous assignment was obtained for Ubiquitin and 95 % of the residues were correctly assigned in the 251 residue-long N-terminal domain of enzyme I. The program including source code is available at https://github.com/thomasexner/4Dassign .

Using Surface Scans for the Evaluation of Halogen Bonds toward the Side Chains of Aspartate, Asparagine, Glutamate, and Glutamine.

Using halogen-specific Connolly type molecular surfaces, we herein invented a new type of surface-based interaction analysis employed for the study of halogen bonding toward model systems of biologically relevant carboxylates (ASP/GLU) and carboxamides (ASN/GLN). Database mining and statistical assessment of the PDB revealed that such interactions are widely underrepresented at the moment. We observed important distance-dependent adaptions of the binding modes of halobenzenes from a preferential oxygen-directed to a bifurcated interaction geometry of the carboxylate. In addition, halogen···π contacts perpendicular to the nitrogen atom of the carboxamide become increasingly important for the lighter halogens. Our analysis on a MP2/TZVPP level of theory is backed by CCSD(T)/CBS reference calculations. To put the vast interaction energies into perspective, we also performed COSMO-RS calculations of the solvation free energy. Facilitating the visualization of our results mapped onto any binding site of choice, we aim to inspire more design studies showcasing these underrepresented interactions.

Accurate ab initio prediction of NMR chemical shifts of nucleic acids and nucleic acids/protein complexes.

NMR chemical shift predictions based on empirical methods are nowadays indispensable tools during resonance assignment and 3D structure calculation of proteins. However, owing to the very limited statistical data basis, such methods are still in their infancy in the field of nucleic acids, especially when non-canonical structures and nucleic acid complexes are considered. Here, we present an ab initio approach for predicting proton chemical shifts of arbitrary nucleic acid structures based on state-of-the-art fragment-based quantum chemical calculations. We tested our prediction method on a diverse set of nucleic acid structures including double-stranded DNA, hairpins, DNA/protein complexes and chemically-modified DNA. Overall, our quantum chemical calculations yield highly/very accurate predictions with mean absolute deviations of 0.3-0.6 ppm and correlation coefficients (r(2)) usually above 0.9. This will allow for identifying misassignments and validating 3D structures. Furthermore, our calculations reveal that chemical shifts of protons involved in hydrogen bonding are predicted significantly less accurately. This is in part caused by insufficient inclusion of solvation effects. However, it also points toward shortcomings of current force fields used for structure determination of nucleic acids. Our quantum chemical calculations could therefore provide input for force field optimization.

Performance of protein-ligand docking with simulated chemical shift perturbations.

Protein chemical shift perturbations (CSPs) that result from the binding of a ligand to the protein contain structural information about the complex. Therefore, the CSP data, typically obtained during library screening from two-dimensional (2D) nuclear magnetic resonance (NMR) spectra, are often available before attempts to solve the experimental structure of the complex are started, and can be used to solve the complex structure with CSP-based docking. Here, we compare the performance of the post-docking filter and the guided-docking approaches using either amide or α-proton CSPs with 10 protein-ligand complexes. We show that the comparison of experimental CSPs with CSPs simulated for virtual ligand positions can be used to evidence protein conformational change upon binding and possibly improve the CSP-based docking.

On-the-Fly Integration of Data from a Spin-Diffusion-Based NMR Experiment into Protein-Ligand Docking.

INPHARMA (interligand nuclear Overhauser enhancement for pharmacophore mapping) determines the relative orientation of two competitive ligands in the protein binding pocket. It is based on the observation of interligand transferred NOEs mediated by spin diffusion through protons of the protein and is, therefore, sensitive to the specific interactions of each of the two ligands with the protein. We show how this information can be directly included into a protein-ligand docking program to guide the prediction of the complex structures. Agreement between the experimental and back-calculated spectra based on the full relaxation matrix approach is translated into a score contribution that is combined with the scoring function ChemPLP of our docking tool PLANTS. This combined score is then used to predict the poses of five weakly bound cAMP-dependent protein kinase (PKA) ligands. After optimizing the setup, which finally also included trNOE data and optimized protonation states, very good success rates were obtained for all combinations of three ligands. For one additional ligand, no conclusive results could be obtained due to the ambiguous electron density of the ligand in the X-ray structure, which does not disprove alternative ligand poses. The failures of the remaining ligand are caused by suboptimal locations of specific protein side chains. Therefore, side-chain flexibility should be included in an improved INPHARMA-PLANTS version. This will reduce the strong dependence on the used protein input structure leading to improved scores overall, not only for this last ligand.

Data quality in drug discovery: the role of analytical performance in ligand binding assays.

Despite its importance and all the considerable efforts made, the progress in drug discovery is limited. One main reason for this is the partly questionable data quality. Models relating biological activity and structures and in silico predictions rely on precisely and accurately measured binding data. However, these data vary so strongly, such that only variations by orders of magnitude are considered as unreliable. This can certainly be improved considering the high analytical performance in pharmaceutical quality control. Thus the principles, properties and performances of biochemical and cell-based assays are revisited and evaluated. In the part of biochemical assays immunoassays, fluorescence assays, surface plasmon resonance, isothermal calorimetry, nuclear magnetic resonance and affinity capillary electrophoresis are discussed in details, in addition radiation-based ligand binding assays, mass spectrometry, atomic force microscopy and microscale thermophoresis are briefly evaluated. In addition, general sources of error, such as solvent, dilution, sample pretreatment and the quality of reagents and reference materials are discussed. Biochemical assays can be optimized to provide good accuracy and precision (e.g. percental relative standard deviation <10 %). Cell-based assays are often considered superior related to the biological significance, however, typically they cannot still be considered as really quantitative, in particular when results are compared over longer periods of time or between laboratories. A very careful choice of assays is therefore recommended. Strategies to further optimize assays are outlined, considering the evaluation and the decrease of the relevant error sources. Analytical performance and data quality are still advancing and will further advance the progress in drug development.

Simplified Method for Removing Direct Oral Anticoagulant Interference in Mechanical Coagulation Test Systems-A Proof of Concept.

BACKGROUND: Direct oral anticoagulants (DOACs) cause unwanted interference in various hemostasis assays, including lupus anticoagulant (LA) testing, where false positive and false negative identification may occur. DOAC Stop (DS) is an activated charcoal (AC) product used to specifically and effectively adsorb DOACs from test plasma. This process normally requires plasma treatment, centrifugation and plasma separation prior to tests, but inexperienced operators may also inadvertently transfer residual AC particles, thereby potentially adversely affecting clot detection. METHODS: We hypothesized that residual DS might not be problematic for mechanical clot detection. We therefore investigated the potential impact of DS and a new DS liquid (DS-L) product on clotting tests using a mechanical clot detection system. Varying concentrations of DS were added to normal and abnormal plasmas with and without DOAC presence. Clotting tests including PT, APTT and dRVVT were performed directly in the analyzer without plasma/DS centrifugation. RESULTS: DS up to double the recommended treatment level had only minor effects on all test results, despite completely obscuring visibility in the plasma/reagent mix. This confirms that the centrifugation step may be able to be omitted when using mechanical detection systems. CONCLUSIONS: Should DS carryover into treated plasmas occur, this should not cause issues with testing performed on mechanical clot-sensing devices. Moreover, we hypothesize that DS can be used directly in these systems, without the need for centrifugation, thereby simplifying its many potential applications.

Evaluation of the impact of iPSC differentiation protocols on transcriptomic signatures.

Human induced pluripotent stem cells (iPSC) have the potential to produce desired target cell types in vitro and allow for the high-throughput screening of drugs/chemicals at population level thereby minimising the cost of drug discovery and drug withdrawals after clinical trials. There is a substantial need for the characterisation of the iPSC derived models to better understand and utilise them for toxicological relevant applications. In our study, iPSC (SBAD2 or SBAD3 lines obtained from StemBANCC project) were differentiated towards toxicologically relevant cell types: alveolar macrophages, brain capillary endothelial cells, brain cells, endothelial cells, hepatocytes, lung airway epithelium, monocytes, podocytes and renal proximal tubular cells. A targeted transcriptomic approach was employed to understand the effects of differentiation protocols on these cell types. Pearson correlation and principal component analysis (PCA) separated most of the intended target cell types and undifferentiated iPSC models as distinct groups with a high correlation among replicates from the same model. Based on PCA, the intended target cell types could also be separated into the three germ layer groups (ectoderm, endoderm and mesoderm). Differential expression analysis (DESeq2) presented the upregulated genes in each intended target cell types that allowed the evaluation of the differentiation to certain degree and the selection of key differentiation markers. In conclusion, these data confirm the versatile use of iPSC differentiated cell types as standardizable and relevant model systems for in vitro toxicology.

Simultaneous occurrence of three different valence tautomers in meso-vinylruthenium-modified zinc porphyrin radical cations.

The mixed-valent radical cation of a styrylruthenium-modified meso-tetraarylzinc porphyrin forms a mixture of three different valence tautomers (VTs) in CH2Cl2 or 1,2-C2H4Cl2 solutions. One of these VTs has the charge and spin delocalized over the porphyrin and the styrylruthenium moieties, while the other two display charge and spin localization on just one of the different redox sites. The relative amounts of the three different VTs were determined by EPR and IR spectroscopies at variable temperatures, while delocalization in the ground state was confirmed by DFT calculations.

Are predefined decoy sets of ligand poses able to quantify scoring function accuracy?

Due to the large number of different docking programs and scoring functions available, researchers are faced with the problem of selecting the most suitable one when starting a structure-based drug discovery project. To guide the decision process, several studies comparing different docking and scoring approaches have been published. In the context of comparing scoring function performance, it is common practice to use a predefined, computer-generated set of ligand poses (decoys) and to reevaluate their score using the set of scoring functions to be compared. But are predefined decoy sets able to unambiguously evaluate and rank different scoring functions with respect to pose prediction performance? This question arose when the pose prediction performance of our piecewise linear potential derived scoring functions (Korb et al. in J Chem Inf Model 49:84-96, 2009) was assessed on a standard decoy set (Cheng et al. in J Chem Inf Model 49:1079-1093, 2009). While they showed excellent pose identification performance when they were used for rescoring of the predefined decoy conformations, a pronounced degradation in performance could be observed when they were directly applied in docking calculations using the same test set. This implies that on a discrete set of ligand poses only the rescoring performance can be evaluated. For comparing the pose prediction performance in a more rigorous manner, the search space of each scoring function has to be sampled extensively as done in the docking calculations performed here. We were able to identify relative strengths and weaknesses of three scoring functions (ChemPLP, GoldScore, and Astex Statistical Potential) by analyzing the performance for subsets of the complexes grouped by different properties of the active site. However, reasons for the overall poor performance of all three functions on this test set compared to other test sets of similar size could not be identified.