Adrian Stecuła

The introduction of AlphaFold 2 has spurred a revolution in modelling the structure of proteins and their interactions, enabling a huge range of applications in protein modelling and design. In this paper, we describe our AlphaFold 3 model with a substantially updated diffusion-based architecture, which is capable of joint structure prediction of complexes including proteins, nucleic acids, small molecules, ions, and modified residues. The new AlphaFold model demonstrates significantly improved… Show more

The introduction of AlphaFold 2 has spurred a revolution in modelling the structure of proteins and their interactions, enabling a huge range of applications in protein modelling and design. In this paper, we describe our AlphaFold 3 model with a substantially updated diffusion-based architecture, which is capable of joint structure prediction of complexes including proteins, nucleic acids, small molecules, ions, and modified residues. The new AlphaFold model demonstrates significantly improved accuracy over many previous specialised tools: far greater accuracy on protein-ligand interactions than state of the art docking tools, much higher accuracy on protein-nucleic acid interactions than nucleic-acid-specific predictors, and significantly higher antibody-antigen prediction accuracy than AlphaFold-Multimer v2.3. Together these results show that high accuracy modelling across biomolecular space is possible within a single unified deep learning framework. Show less

High throughput screening (HTS) is routinely used to identify bioactive small molecules. This requires physical compounds, which limits coverage of accessible chemical space. Computational approaches combined with vast on-demand chemical libraries can access far greater chemical space, provided that the predictive accuracy is sufficient to identify useful molecules. Through the largest and most diverse virtual HTS campaign reported to date, comprising 318 individual projects, we demonstrate… Show more

High throughput screening (HTS) is routinely used to identify bioactive small molecules. This requires physical compounds, which limits coverage of accessible chemical space. Computational approaches combined with vast on-demand chemical libraries can access far greater chemical space, provided that the predictive accuracy is sufficient to identify useful molecules. Through the largest and most diverse virtual HTS campaign reported to date, comprising 318 individual projects, we demonstrate that our AtomNet® convolutional neural network successfully finds novel hits across every major therapeutic area and protein class. We address historical limitations of computational screening by demonstrating success for target proteins without known binders, high-quality X-ray crystal structures, or manual cherry-picking of compounds. We show that the molecules selected by the AtomNet® model are novel drug-like scaffolds rather than minor modifications to known bioactive compounds. Our empirical results suggest that computational methods can substantially replace HTS as the first step of small-molecule drug discovery. Show less

The human solute carrier 22A (SLC22A) family consists of 23 members, representing one of the largest families in the human SLC superfamily. Despite their pharmacological and physiological importance in the absorption and disposition of a range of solutes, eight SLC22A family members remain classified as orphans. In this study, we used a multifaceted approach to identify ligands of orphan SLC22A15. Ligands of SLC22A15 were proposed based on phylogenetic analysis and comparative modeling. The… Show more

The human solute carrier 22A (SLC22A) family consists of 23 members, representing one of the largest families in the human SLC superfamily. Despite their pharmacological and physiological importance in the absorption and disposition of a range of solutes, eight SLC22A family members remain classified as orphans. In this study, we used a multifaceted approach to identify ligands of orphan SLC22A15. Ligands of SLC22A15 were proposed based on phylogenetic analysis and comparative modeling. The putative ligands were then confirmed by metabolomic screening and uptake assays in SLC22A15 transfected HEK293 cells. Metabolomic studies and transporter assays revealed that SLC22A15 prefers zwitterionic compounds over cations and anions. We identified eight zwitterions, including ergothioneine, carnitine, carnosine, gabapentin, as well as four cations, including MPP+, thiamine, and cimetidine, as substrates of SLC22A15. Carnosine was a specific substrate of SLC22A15 among the transporters in the SLC22A family. SLC22A15 transport of several substrates was sodium‐dependent and exhibited a higher Km for ergothioneine, carnitine, and carnosine compared to previously identified transporters for these ligands. This is the first study to characterize the function of SLC22A15. Our studies demonstrate that SLC22A15 may play an important role in determining the systemic and tissue levels of ergothioneine, carnosine, and other zwitterions. Show less

Human OAT1 and OAT3 play major roles in renal drug elimination and drug-drug interactions. However, there is little information on the interactions of drug metabolites with transporters. The goal of this study was to characterize the interactions of drug metabolites with OAT1 and OAT3 and compare their potencies of inhibition with those of their corresponding parent drugs. Using HEK293 cells stably transfected with OAT1 and OAT3, 25 drug metabolites and their corresponding parent drugs were… Show more

Human OAT1 and OAT3 play major roles in renal drug elimination and drug-drug interactions. However, there is little information on the interactions of drug metabolites with transporters. The goal of this study was to characterize the interactions of drug metabolites with OAT1 and OAT3 and compare their potencies of inhibition with those of their corresponding parent drugs. Using HEK293 cells stably transfected with OAT1 and OAT3, 25 drug metabolites and their corresponding parent drugs were screened for inhibitory effects on OAT1- and OAT3-mediated 6-carboxyfluorescein uptake at a screening concentration of 200 μM for all but 3 compounds. 20 and 24 drug metabolites were identified as inhibitors (inhibition > 50%) of OAT1 and OAT3, respectively. Seven drug metabolites were potent inhibitors of either or both OAT1 and OAT3 with Ki values less than 1 μM. 22 metabolites were more potent inhibitors of OAT3 than OAT1. Importantly, one drug and four metabolites were predicted to inhibit OAT3 at unbound plasma concentrations achieved clinically (Cmax,u/Ki values ≥ 0.1). In conclusion, our study highlights the potential interactions of drug metabolites with OAT1 and OAT3 at clinically relevant concentrations, suggesting that drug metabolites may modulate therapeutic and adverse drug response by inhibiting renal drug transporters. Show less

Rare neglected diseases may be neglected but are hardly rare, affecting hundreds of millions of people around the world. Here, we present a hit identification approach using AtomNet, the world's first deep convolutional neural network for structure-based drug discovery, to identify inhibitors targeting aspartate N-acetyltransferase (ANAT), a promising target for the treatment of patients suffering from Canavan disease. Despite the lack of a protein structure or high sequence identity homologous… Show more

Rare neglected diseases may be neglected but are hardly rare, affecting hundreds of millions of people around the world. Here, we present a hit identification approach using AtomNet, the world's first deep convolutional neural network for structure-based drug discovery, to identify inhibitors targeting aspartate N-acetyltransferase (ANAT), a promising target for the treatment of patients suffering from Canavan disease. Despite the lack of a protein structure or high sequence identity homologous templates, the approach successfully identified five low-micromolar inhibitors with drug-like properties. Show less

Variation in steroid hormone levels has wide implications for health and disease. The genes encoding the proteins involved in steroid disposition represent key determinants of interindividual variation in steroid levels and ultimately, their effects. Beginning with metabolomic data from genome-wide association studies (GWAS), we observed that genetic variants in the orphan transporter, SLC22A24 were significantly associated with levels of androsterone glucuronide and etiocholanolone glucuronide… Show more

Variation in steroid hormone levels has wide implications for health and disease. The genes encoding the proteins involved in steroid disposition represent key determinants of interindividual variation in steroid levels and ultimately, their effects. Beginning with metabolomic data from genome-wide association studies (GWAS), we observed that genetic variants in the orphan transporter, SLC22A24 were significantly associated with levels of androsterone glucuronide and etiocholanolone glucuronide (sentinel SNPs p-value <1x10-30). In cells over-expressing human or various mammalian orthologs of SLC22A24, we showed that steroid conjugates and bile acids were substrates of the transporter. Phylogenetic, genomic, and transcriptomic analyses suggested that SLC22A24 has a specialized role in the kidney and appears to function in the reabsorption of organic anions, and in particular, anionic steroids. Phenome-wide analysis showed that functional variants of SLC22A24 are associated with human disease such as cardiovascular diseases and acne, which have been linked to dysregulated steroid metabolism. Collectively, these functional genomic studies reveal a previously uncharacterized protein involved in steroid homeostasis, opening up new possibilities for SLC22A24 as a pharmacological target for regulating steroid levels. Show less

Species differences in renal drug transporters continue to plague drug development with animal models failing to adequately predict renal drug toxicity. For example, adefovir, a renally excreted antiviral drug, failed clinical studies for HIV due to pronounced nephrontoxicity in humans. In this study, we demonstrated that there are large species differences in the kinetics of interactions of antiviral drugs with OAT1 (SLC22A6) and identified a key amino acid residue responsible for these… Show more

Species differences in renal drug transporters continue to plague drug development with animal models failing to adequately predict renal drug toxicity. For example, adefovir, a renally excreted antiviral drug, failed clinical studies for HIV due to pronounced nephrontoxicity in humans. In this study, we demonstrated that there are large species differences in the kinetics of interactions of antiviral drugs with OAT1 (SLC22A6) and identified a key amino acid residue responsible for these differences. In OAT1 stably transfected HEK293 cells, the Km of tenofovir for human OAT1 was significantly lower than for OAT1 orthologs from common preclinical animals, including cynomolgus monkey, mouse, rat and dog. Chimeric and site-directed mutagenesis studies along with comparative structure modeling identified serine at position 203 (S203) in hOAT1 as a determinant of its lower Km value. Further, S203 is conserved in apes and in contrast, alanine at the equivalent position is conserved in preclinical animals and Old World monkeys, the most related primates to apes. Intriguingly, transport efficiencies are significantly higher for OAT1 orthologs from apes with high serum uric acid levels than the orthologs from species with low serum uric acid levels. In conclusion, our data provide a molecular mechanism underlying species differences in renal accumulation of nephrotoxic antiviral drugs and a novel insight into OAT1 transport function in primate evolution. Show less

Many anticancer and antiviral drugs are purine or pyrimidine analogs, which use membrane transporters to cross cellular membranes. Concentrative nucleoside transporters (CNTs) mediate the salvage of nucleosides and the transport of therapeutic nucleoside analogs across plasma membranes by coupling the transport of ligands to the sodium gradient. Of the three members of the human CNT family, CNT3 has the broadest selectivity and the widest expression profile. However, the molecular mechanisms… Show more

Many anticancer and antiviral drugs are purine or pyrimidine analogs, which use membrane transporters to cross cellular membranes. Concentrative nucleoside transporters (CNTs) mediate the salvage of nucleosides and the transport of therapeutic nucleoside analogs across plasma membranes by coupling the transport of ligands to the sodium gradient. Of the three members of the human CNT family, CNT3 has the broadest selectivity and the widest expression profile. However, the molecular mechanisms of the transporter, including how it interacts with and translocates structurally diverse nucleosides and nucleoside analogs are unclear. Recently, the crystal structure of vcCNT showed that the prokaryotic homolog of CNT3 forms a homo-trimer. In this study, we successfully expressed and purified the wild type human homolog, hCNT3, demonstrating the homo-trimer by size exclusion profiles and glutaraldehyde cross-linking. Further, by creating a series of cysteine mutants at highly conserved positions guided by comparative structure models, we cross-linked hCNT3 protomers in a cell-based assay, thus showing the existence of hCNT3 homo-trimers in human cells. The presence and absence of cross-links at specific locations along TM9 informs us of important structural differences between vcCNT and hCNT3. Comparative modeling of the trimerization domain and sequence co-evolution analysis both indicate that oligomerization is critical to the stability and function of hCNT3. In particular, trimerization appears to shorten the translocation path for nucleosides across the plasma membrane and may allow modulation of the transport function via allostery. Show less

Organic cation transporter 1 (OCT1) plays a critical role in the hepatocellular uptake of structurally diverse endogenous compounds and xenobiotics. Here, we identified competitive and noncompetitive OCT1 interacting ligands in a library of 1,780 prescription drugs by combining in silico and in vitro methods. Ligands were predicted by docking against a comparative model based a eukaryotic homolog. In parallel, highthroughput screening (HTS) was conducted using the fluorescent probe substrate… Show more

Organic cation transporter 1 (OCT1) plays a critical role in the hepatocellular uptake of structurally diverse endogenous compounds and xenobiotics. Here, we identified competitive and noncompetitive OCT1 interacting ligands in a library of 1,780 prescription drugs by combining in silico and in vitro methods. Ligands were predicted by docking against a comparative model based a eukaryotic homolog. In parallel, highthroughput screening (HTS) was conducted using the fluorescent probe substrate ASP+ in cells overexpressing human OCT1. Thirty OCT1 competitive ligands, defined as ligands predicted in silico as well as found by HTS, were identified. Of the 167 ligands identified by HTS, five were predicted to potentially cause clinical drug interactions. Finally, virtual screening of 29,332 metabolites predicted 146 competitive OCT1 ligands, of which an endogenous neurotoxin, 1-benzyl-1,2,3,4-tetrahydroisoquinoline, was experimentally validated. In conclusion, by combining docking and in vitro HTS competitive and noncompetitive ligands of OCT1 can be predicted. Show less

Membrane transporters are key determinants of therapeutic outcomes. They regulate systemic and cellular drug levels influencing efficacy as well as toxicities. Here we report a unique phosphorylation-dependent interaction between drug transporters and tyrosine kinase inhibitors (TKIs), which has uncovered widespread phosphotyrosine-mediated regulation of drug transporters. We initially found that organic cation transporters (OCTs), uptake carriers of metformin and oxaliplatin, were inhibited by… Show more

Membrane transporters are key determinants of therapeutic outcomes. They regulate systemic and cellular drug levels influencing efficacy as well as toxicities. Here we report a unique phosphorylation-dependent interaction between drug transporters and tyrosine kinase inhibitors (TKIs), which has uncovered widespread phosphotyrosine-mediated regulation of drug transporters. We initially found that organic cation transporters (OCTs), uptake carriers of metformin and oxaliplatin, were inhibited by several clinically used TKIs. Mechanistic studies showed that these TKIs inhibit the Src family kinase Yes1, which was found to be essential for OCT2 tyrosine phosphorylation and function. Yes1 inhibition in vivo diminished OCT2 activity, significantly mitigating oxaliplatin-induced acute sensory neuropathy. Along with OCT2, other SLC-family drug transporters are potentially part of an extensive ‘transporter-phosphoproteome’ with unique susceptibility to TKIs. On the basis of these findings we propose that TKIs, an important and rapidly expanding class of therapeutics, can functionally modulate pharmacologically important proteins by inhibiting protein kinases essential for their post-translational regulation. Show less

Tyrosine kinase inhibitor (TKI) resistance due to acquired secondary kinase domain (KD) mutations in BCR-ABL1 represents a common clinically encountered problem in patients with chronic myeloid leukemia (CML). The prototypic ABL1 TKI imatinib is vulnerable to a large number of resistance-conferring secondary KD mutations. Clinical management of imatinib-resistant disease has been successfully guided by in vitro studies of mutant sensitivities to alternative TKIs, which are largely predictive of… Show more

Tyrosine kinase inhibitor (TKI) resistance due to acquired secondary kinase domain (KD) mutations in BCR-ABL1 represents a common clinically encountered problem in patients with chronic myeloid leukemia (CML). The prototypic ABL1 TKI imatinib is vulnerable to a large number of resistance-conferring secondary KD mutations. Clinical management of imatinib-resistant disease has been successfully guided by in vitro studies of mutant sensitivities to alternative TKIs, which are largely predictive of clinical responsiveness. Show less

Activating mutations in FLT3 occur in ~30% of adult acute myeloid leukemia, primarily consisting of internal tandem duplication (ITD) mutations (~25%) and point mutations in the tyrosine kinase domain (~5%), commonly at the activation loop residue D835. Secondary kinase domain mutations in FLT3-ITD, particularly at the D835 residue are frequently associated with acquired clinical resistance to effective FLT3 tyrosine kinase inhibitors (TKIs). Molecular docking studies have suggested that D835… Show more

Activating mutations in FLT3 occur in ~30% of adult acute myeloid leukemia, primarily consisting of internal tandem duplication (ITD) mutations (~25%) and point mutations in the tyrosine kinase domain (~5%), commonly at the activation loop residue D835. Secondary kinase domain mutations in FLT3-ITD, particularly at the D835 residue are frequently associated with acquired clinical resistance to effective FLT3 tyrosine kinase inhibitors (TKIs). Molecular docking studies have suggested that D835 mutations primarily confer resistance by stabilizing an active Asp-Phe-Gly in ('DFG-in') kinase conformation unfavorable to the binding of type II FLT3 TKIs, which target a 'DFG-out' inactive conformation. We profiled the activity of active type II FLT3 TKIs against D835 kinase domain mutants that have been clinically detected to date. We found that type II inhibitors (quizartinib, sorafenib, ponatinib and PLX3397) retain activity against specific D835 substitutions. Modeling studies suggest that bulky hydrophobic substitutions (D835Y/V/I/F) at this residue are particularly resistant, whereas mutations that preserve interactions between D835 and S838 are relatively sensitive (D835E/N). All mutants retain sensitivity to the type I inhibitor crenolanib. These results suggest that patients with relatively sensitive D835 mutations should be included in clinical trials of type II FLT3 TKIs. Show less

Capping is a mechanical defect in tablet formulation and manufacture. Understanding what influences tablet capping in terms of process variables and formulation properties and developing specialized techniques to correlate these variables with mechanical failures are practical interests of the pharmaceutical industry. Tablet compression emulator was used to rank order capping tendencies of a diverse sample set. The compression forces of 5–25 kN were used to compress round, beveled edge, and… Show more

Capping is a mechanical defect in tablet formulation and manufacture. Understanding what influences tablet capping in terms of process variables and formulation properties and developing specialized techniques to correlate these variables with mechanical failures are practical interests of the pharmaceutical industry. Tablet compression emulator was used to rank order capping tendencies of a diverse sample set. The compression forces of 5–25 kN were used to compress round, beveled edge, and oval shape tablets. Compression speeds of 25, 40, and 80 rpm were chosen as representative ranges for bench-to-manufacturing-scale processing. Tablets were tested by in-house developed nondestructive ultrasonic method. The measurements revealed that elastic modulus vary with different testing orientations that indicated elastic modulus anisotropy in tablets. It was shown that altering process conditions such as tooling, compression force, and compression speed significantly impact the capping tendencies of formulations. A systematic approach has been applied to develop a predictive tool to assess capping tendencies of formulations. The developed tool is fast, material sparing, and has potential to flag the risk of manufacturability issues and provide insight into the performance of formulations during early development. Show less

Acute myeloid leukemia (AML) is a clinically heterogeneous disease, with a 5-year disease-free survival (DFS) ranging from under 10% to over 70% for distinct groups of patients. At our institution, cytarabine, etoposide and busulfan are used in first or second remission patients treated with a two-step approach to autologous stem cell transplantation (ASCT). In this study,
we tested the hypothesis that polymorphisms in the pharmacokinetic and pharmacodynamic pathway genes of these drugs are… Show more

Acute myeloid leukemia (AML) is a clinically heterogeneous disease, with a 5-year disease-free survival (DFS) ranging from under 10% to over 70% for distinct groups of patients. At our institution, cytarabine, etoposide and busulfan are used in first or second remission patients treated with a two-step approach to autologous stem cell transplantation (ASCT). In this study,
we tested the hypothesis that polymorphisms in the pharmacokinetic and pharmacodynamic pathway genes of these drugs are associated with DFS in AML patients. A total of 1659 variants in 42 genes were analyzed for their association with DFS using a Cox-proportional hazards model. One hundred and fifty-four genetically European patients were used for the primary analysis.
An intronic single nucleotide polymorphism (SNP) in ABCC3 (rs4148405) was associated with a significantly shorter DFS (hazard ratios (HR) 1⁄4 3.2, P 1⁄4 5.6 10 6) in our primary cohort. In addition, a SNP in the GSTM1-GSTM5 locus, rs3754446, was significantly associated with a shorter DFS in all patients (HR 1⁄4 1.8, P 1⁄4 0.001 for 154 European ancestry; HR 1⁄4 1.7,
P 1⁄4 0.028 for 125 non-European patients). Thus, for the first time, genetic variants in drug pathway genes are shown to be associated with DFS in AML patients treated with chemotherapy-based autologous ASCT. Show less