Herein we report the development of novel covalent inhibitors of the SARS-CoV-2 main protease (Mpro). The developed compounds VPC285785 and VPC285786 demonstrated moderate inhibition of Mpro (IC50 0.8 µM vs. Nirmatrelvir 0.03 µM), whereas VPC285786 additionally inhibited human cathepsin L (CatL; IC50 4.2 µM vs Nirmatrelvir > 100 µM). In vitro metabolic stability studies in human and mouse microsomes revealed that VPC285786 demonstrated enhanced metabolic stability compared to Nirmatrelvir, with minimal turnover observed during the experimental window. Subsequent mass spectrometry analysis identified putative metabolic products consistent with previously reported oxidation patterns. Pharmacokinetic studies in mice demonstrated that VPC285785 achieved 15% oral bioavailability, supporting the potential for oral administration, whereas VPC285786 showed limited oral exposure despite superior metabolic stability. A side-by-side efficacy study of VPC285785 and Nirmatrelvir in a Murine Hepatitis Virus (MHV) infection model demonstrated that VPC285785 significantly reduced viral load in liver, brain, and spleen tissues compared to vehicle- and Nirmatrelvir-treated controls, while maintaining healthy liver function parameters. These results lay the foundation for further development of VPC285785-series antivirals that could be used as oral, single-agent therapies for SARS-CoV-2 infection, particularly given their dual-targeting mechanism and favorable toxicity profile.

Brain tumors, including primary brain tumors and central nervous system (CNS) metastases, remain among the most challenging malignancies to treat, with therapeutic options often limited by the inability of drugs to penetrate the blood-brain barrier (BBB). Poly(ADP-ribose) polymerase (PARP) is a key enzyme in DNA repair, and inhibition of PARP1 specifically drives synthetic lethality in BRCA-mutated disease. While they have achieved commercial success, first generation PARP inhibitors are limited in their utility as they cannot readily pass through the blood-brain barrier and produce adverse side effects, likely driven by the collateral inhibition of PARP2. Development of a PARP1-selective CNS penetrant inhibitor could reduce toxicity, while providing a new therapeutic option for brain tumors. Traditional drug discovery methods are time-consuming and costly, necessitating innovative approaches. Here, we describe the application of Deep Docking combined with generative artificial intelligence (AI), to discover a novel, PARP1-selective inhibitor for use against brain tumors. Deep Docking utilizes deep learning to accelerate the prediction of the binding affinity of a large number of compounds to target proteins, streamlining the virtual screening process and allowing for rapid docking of billions of compounds against the PARP1 protein. Additionally, state-of-the-art generative algorithms and deep learning techniques for predicting CNS penetrance can be combined with Deep Docking to rapidly perform hit-to-lead optimization of a CNS-penetrating PARP1-selective inhibitor. We will present Deep Docking screening results from billions of compounds. Validating in vitro and in vivo data, including PARP1 inhibition and selectivity, BBB permeability, metabolic stability, pharmacokinetic profile, and CNS penetration will be described. The Deep Docking platform is being used in concert with generative AI approaches to discover and finetune a selective PARP-1 inhibitor for brain tumors. This approach will not only improve the efficiency of drug discovery but also enhance the specificity and efficacy of potential therapeutics. Sarah Truong, Beibei Zhai, Louise Ramos, Mona Marzban, Fariba Ghaidi, Fuqiang Ban, Jason Smith, Mohit Pandey, Ekaterina Manskaia, Kally Singh, Hans Adomat, Xiaoqi Chen, John Langlands, Dennis Brown, Jeffrey Bacha, Colin Collins, Artem Cherkasov, Mads Daugaard. Discovery of novel PARP1-selective inhibitors for treatment of brain tumors using artificial intelligence [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4186.

BACKGROUND:The main protease (Mpro) is one of the most attractive targets for antiviral drug discovery against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Mutations in Mpro have been linked to resistance against nirmatrelvir-ritonavir (NIR-RIT), an important therapy for SARS-CoV-2 infection. This study aimed to identify low-frequency antiviral resistance mutations in Mpro from NIR-RIT-treated patients and to analyze the enzymatic properties, inhibitor susceptibility, and structural features of new Mpro clinical variants. METHODS:We screened 1528 SARS-CoV-2-positive patients from 2 hospitals and identified 17 who remained positive after treatment. Whole-genome sequencing of nasopharyngeal specimens was conducted to identify Mpro clinical variants. The impact of these mutations on Mpro activity and inhibitor susceptibility was investigated using a fluorescent enzymatic biosensor in human cells, along with in vitro thermal stability and structure-based analyses of the Mpro mutants and Mpro-NIR complexes. RESULTS:The analysis identified 2 novel Mpro clinical variants: D48D/L58F/P132H (variant 1) and D48D/L67V/K90R/P132H (variant 2). Our data show that the selected clinical mutations are localized in the Mpro N-terminal domain, are far from the catalytic site, and strongly impact NIR resistance without affecting Mpro activity. Structural analysis and thermal denaturation analyses revealed that these mutations may disrupt the substrate binding site's structure and dynamics, reducing protein stability and potentially impacting substrate binding or dimerization without compromising catalytic activity. CONCLUSIONS:Our new Mpro clinical mutations that confer complete resistance to NIR were not identified during previous cell-culture-based studies. More research is needed to explore resistance mechanisms, providing insights into strategies that mitigate resistance and protect therapeutic efficacy.

Lin28 is a key regulator of cancer stem cell gene network that promotes therapy-resistant tumor progression in various tumors. However, no Lin28 inhibitor has been approved to treat cancer patients, urging exploration of novel compounds as candidates to be tested for clinical trials. In this contribution, we applied computer-aided drug design (CADD) in combination with quantitative biochemical and biological assays. These efforts led to the discovery of Ln268 as a drug candidate that can block Lin28 from binding to its RNA substrates and inhibit Lin28 activities. Ln268 suppressed Lin28-mediated cancer cell proliferation and spheroid growth. Results from nuclear magnetic resonance spectroscopy confirmed that Ln268 perturbs the conformation of the zinc knuckle domain of Lin28, validating the rational drug design by CADD. The inhibitory effects of Ln268 are dependent on Lin28 protein expression in cancer cells, highlighting limited off-target effects of Ln268. Moreover, Ln268 synergizes with several chemotherapy drugs to suppress tumor cell growth. In summary, Ln268 is a promising candidate for further development to target Lin28 as a cancer therapy.

AbstractUp to 40% of patients with estrogen receptor (ER)-positive breast cancer will develop resistance against the majority of current ER-directed therapies. Resistance can arise through various mechanisms such as increased expression levels of coregulators, and key mutations acquired in the receptor’s ligand binding domain rendering it constitutively active. To overcome these resistance mechanisms, we explored targeting the ER Activation Function 2 (AF2) site, which is essential for coactivator binding and activation. Using artificial intelligence and the deep docking methodology, we virtually screened > 1 billion small molecules and identified 290 potential AF2 binders that were then characterized and validated through an iterative screening pipeline of cell-based and cell-free assays. We ranked the compounds based on their ability to reduce the transcriptional activity of the estrogen receptor and the viability of ER-positive breast cancer cells. We identified a lead compound, VPC-260724, which inhibits ER activity at low micromolar range. We confirmed its direct binding to the ER-AF2 site through a PGC1α peptide displacement experiment. Using proximity ligation assays, we showed that VPC-260724 disrupts the interaction between ER-AF2 and the coactivator SRC-3 and reduces the expression of ER target genes in various breast cancer models including the tamoxifen resistant cell line TamR3. In conclusion, we developed a novel ER-AF2 binder, VPC-260724, which shows antiproliferative activity in ER-positive breast cancer models. The use of an ER-AF2 inhibitor in combination with current treatments may provide a novel complementary therapeutic approach to target treatment resistance in ER-positive breast cancer.

Searching the vast chemical space for drug-like molecules that bind with a protein pocket is a challenging task in drug discovery. Recently, structure-based generative models have been introduced which promise to be more efficient by learning to generate molecules for any given protein structure. However, since they learn the distribution of a limited protein-ligand complex dataset, structure-based methods do not yet outperform optimization-based methods that generate binding molecules for just one pocket. To overcome limitations on data while leveraging learning across protein targets, we choose to model the reward distribution conditioned on pocket structure, instead of the training data distribution. We design TacoGFN, a novel GFlowNet-based approach for structure-based drug design, which can generate molecules conditioned on any protein pocket structure with probabilities proportional to its affinity and property rewards. In the generative setting for CrossDocked2020 benchmark, TacoGFN attains a state-of-the-art success rate of 56.0% and -8.44 kcal/mol in median Vina Dock score while improving the generation time by multiple orders of magnitude. Fine-tuning TacoGFN further improves the median Vina Dock score to -10.93 kcal/mol and the success rate to 88.8\%, outperforming all optimization-based methods.

Abstract Brain tumors, including primary brain tumors and central nervous system (CNS) metastases, remain among the most challenging malignancies to treat, with therapeutic options often limited by the inability of drugs to penetrate the blood-brain barrier. Poly(ADP-ribose) polymerase (PARP) is a key enzyme in DNA repair, and inhibition of PARP1 specifically drives synthetic lethality in BRCA-mutated disease. While they’ve achieved commercial success, first generation PARP inhibitors are limited in their utility as they cannot readily pass through the blood-brain barrier and have adverse side effects, likely driven by the collateral inhibition of PARP2. Development of a PARP1-selective, CNS penetrant inhibitor could reduce toxicity, while providing a new therapeutic option for brain tumors. Traditional drug discovery methods are time-consuming and costly, necessitating innovative approaches. Here, we describe the application of Deep Docking, an advanced artificial intelligence (AI) approach, to discover a novel, PARP1-selective inhibitor for use against brain tumors. Deep Docking utilizes deep learning to accelerate the prediction of the binding affinity of a large number of compounds to target proteins, streamlining the virtual screening process and allowing for rapid docking of billions of compounds against the PARP1 protein. Additionally, state-of-the-art generative algorithms and deep learning techniques for predicting drug-like properties such as metabolism, permeability, and safety profiles, can be combined to rapidly perform hit-to-lead optimization. We will present preliminary Deep Docking screening results from billions of compounds, identifying several with predicted high binding affinities for PARP-1. Validating in vitro and in vivo data, including PARP1 selectivity, metabolic stability, pharmacokinetic profile, CNS penetration and safety profile, will be described. Application of the Deep Docking AI platform is being used to significantly accelerate the discovery of a selective PARP-1 inhibitor for brain tumors. This approach will not only improve the efficiency of drug discovery but also enhance the specificity and efficacy of potential therapeutics.

Antivirals with broad coronavirus activity are important for treating high-risk individuals exposed to the constantly evolving SARS-CoV-2 variants of concern (VOCs) as well as emerging drug-resistant variants. We developed and characterized a novel class of active-site-directed 3-chymotrypsin-like protease (3CLpro) inhibitors (C2-C5a). Our lead direct-acting antiviral (DAA), C5a, is a non-covalent, non-peptide with a dissociation constant of 170 nM against recombinant SARS-CoV-2 3CLpro. The compounds C2-C5a exhibit broad-spectrum activity against Omicron subvariants (BA.5, BQ.1.1, and XBB.1.5) and seasonal human coronavirus-229E infection in human cells. Notably, C5a has median effective concentrations of 30-50 nM against BQ.1.1 and XBB.1.5 in two different human cell lines. X-ray crystallography has confirmed the unique binding modes of C2-C5a to the 3CLpro, which can limit virus cross-resistance to emerging Paxlovid-resistant variants. We tested the effect of C5a with two of our newly discovered host-directed antivirals (HDAs): N-0385, a TMPRSS2 inhibitor, and bafilomycin D (BafD), a human vacuolar H+-ATPase [V-ATPase] inhibitor. We demonstrated a synergistic action of C5a in combination with N-0385 and BafD against Omicron BA.5 infection in human Calu-3 lung cells. Our findings underscore that a SARS-CoV-2 multi-targeted treatment for circulating Omicron subvariants based on DAAs (C5a) and HDAs (N-0385 or BafD) can lead to therapeutic benefits by enhancing treatment efficacy. Furthermore, the high-resolution structures of SARS-CoV-2 3CLpro in complex with C2-C5a will facilitate future rational optimization of our novel broad-spectrum active-site-directed 3C-like protease inhibitors.

The rapid global spread of the SARS-CoV-2 virus facilitated the development of novel direct-acting antiviral agents (DAAs). The papain-like protease (PLpro) has been proposed as one of the major SARS-CoV-2 targets for DAAs due to its dual role in processing viral proteins and facilitating the host's immune suppression. This dual role makes identifying small molecules that can effectively neutralize SARS-CoV-2 PLpro activity a high-priority task. However, PLpro drug discovery faces a significant challenge due to the high mobility and induced-fit effects in the protease's active site. Herein, we virtually screened the ZINC20 database with Deep Docking (DD) to identify prospective noncovalent PLpro binders and combined ultra-large consensus docking with two pharmacophore (ph4)-filtering strategies. The analysis of active compounds revealed their somewhat-limited diversity, likely attributed to the induced-fit nature of PLpro's active site in the crystal structures, and therefore, the use of rigid docking protocols poses inherited limitations. The top hits were assessed against recombinant viral proteins and live viruses, demonstrating desirable inhibitory activities. The best compound VPC-300195 (IC50: 15 μM) ranks among the top noncovalent PLpro inhibitors discovered through in silico methodologies. In the search for novel SARS-CoV-2 PLpro-specific chemotypes, the identified inhibitors could serve as diverse templates for the development of effective noncovalent PLpro inhibitors.

The p53 protein, known as the 'guardian of the genome', plays an important role in cancer prevention. Unfortunately, p53 mutations result in compromised activity with over 50% of cancers resulting from point mutations to p53. There is considerable interest in mutant p53 reactivation, with the development of small-molecule reactivators showing promise. We have focused our efforts on the common p53 mutation Y220C, which causes protein unfolding, aggregation, and can result in the loss of a structural Zn from the DNA-binding domain. In addition, the Y220C mutant creates a surface pocket that can be stabilized using small molecules. We previously reported the bifunctional ligand L5 as a Zn metallochaperone and reactivator of the p53-Y220C mutant. Herein we report two new ligands L5-P and L5-O that are designed to act as Zn metallochaperones and non-covalent binders in the Y220C mutant pocket. For L5-P the distance between the Zn-binding di-(2-picolyl)amine function and the pocket-binding diiodophenol was extended in comparison to L5, while for L5-O we extended the pocket-binding moiety via attachment of an alkyne function. While both new ligands displayed similar Zn-binding affinity to L5, neither acted as efficient Zn-metallochaperones. However, the new ligands exhibited significant cytotoxicity in the NCI-60 cell line screen as well as in the NUGC3 Y220C mutant cell line. We identified that the primary mode of cytotoxicity is likely reactive oxygen species (ROS) generation for L5-P and L5-O, in comparison to mutant p53 reactivation for L5, demonstrating that subtle changes to the ligand scaffold can change the toxicity pathway.

Up to 80% of breast cancers (BCa) rely on the estrogen receptor (ER) for their growth and progression. This dependence on ER has led to many hormonal therapies that target this receptor. However, almost 40% of these cancers will acquire resistance over the course of the treatment period. One potential cause of resistance is mutations in the estrogen binding site (EBS) of ER. As such, there is an increasing need for novel inhibitors that targets ER at a site separate from the EBS. Here, we propose targeting the activation-function (AF2) pocket of ER that is important for cofactor binding and transcription activation. Billions of compounds were screened through an in-silico deep docking method, and potential AF2 inhibitors were then validated in cell-based and biophysical assays. We tested the effect of potential AF2 inhibitors on ER transcriptional activity using luciferase reporter assay in ER-positive T47D-kbluc cells. We then evaluated the effect of molecules on cell viability of ER-positive T47D and ER-negative MDA-MB-231 cells using PrestoBlue assays in order to exclude off-target effects. From these cell-based assays, we identified several inhibitors that effectively reduced transcriptional activity and viability in ER-positive T47D cells at low micromolar concentrations. We conducted PGC-1α peptide displacement assay to confirm their AF2 binding and estradiol displacement assays to exclude any binding to the EBS. Proximity ligation assay (PLA) showed disruption of the interaction between ER and coactivator SRC-3 upon treatment with ER-AF2 inhibitors in T47D cells. Current work focuses on confirming the direct binding between the compounds and recombinant ER-ligand binding domain by various biophysical assays (MST, BLI, and ITC). Future work aims to solve the structure of ER-LBD in a complex with our lead compound by X-ray crystallography. We predict that the use of potent ER-AF2 inhibitors along with current treatments, will provide a novel tactic that can act as a complementary therapeutic to target treatment resistance in ER+ BCa. Citation Format: Jane Foo, Francesco Gentile, Joseph Lee, Helene Morin, Shabnam Massah, Maria Guo, Jason Smith, Fuqiang Ban, Artem Cherkasov, Nada Lallous. Characterization of ER-AF2 inhibitors in breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3079.

While vaccines remain at the forefront of global healthcare responses, pioneering therapeutics against SARS-CoV-2 are expected to fill the gaps for waning immunity. Rapid development and approval of orally available direct-acting antivirals targeting crucial SARS-CoV-2 proteins marked the beginning of the era of small-molecule drugs for COVID-19. In that regard, the papain-like protease (PLpro) can be considered a major SARS-CoV-2 therapeutic target due to its dual biological role in suppressing host innate immune responses and in ensuring viral replication. Here, we summarize the challenges of targeting PLpro and innovative early-stage PLpro-specific small molecules. We propose that state-of-the-art computer-aided drug design (CADD) methodologies will play a critical role in the discovery of PLpro compounds as a novel class of COVID-19 drugs.

Alzheimer's disease (AD) is a chronic neurodegenerative disorder characterized by progressive and irreversible damage to the brain. One of the hallmarks of the disease is the presence of both soluble and insoluble aggregates of the amyloid beta (Aβ) peptide in the brain, and these aggregates are considered central to disease progression. Thus, the development of small molecules capable of modulating Aβ peptide aggregation may provide critical insight into the pathophysiology of AD. In this work we investigate how photoactivation of three distorted Ru(ii) polypyridyl complexes (Ru1-3) alters the aggregation profile of the Aβ peptide. Photoactivation of Ru1-3 results in the loss of a 6,6'-dimethyl-2,2'-bipyridyl (6,6'-dmb) ligand, affording cis-exchangeable coordination sites for binding to the Aβ peptide. Both Ru1 and Ru2 contain an extended planar imidazo[4,5-f][1,10]phenanthroline ligand, as compared to a 2,2'-bipyridine ligand for Ru3, and we show that the presence of the phenanthroline ligand promotes covalent binding to Aβ peptide His residues, and in addition, leads to a pronounced effect on peptide aggregation immediately after photoactivation. Interestingly, all three complexes resulted in a similar aggregate size distribution at 24 h, forming insoluble amorphous aggregates as compared to significant fibril formation for peptide alone. Photoactivation of Ru1-3 in the presence of pre-formed Aβ1-42 fibrils results in a change to amorphous aggregate morphology, with Ru1 and Ru2 forming large amorphous aggregates immediately after activation. Our results show that photoactivation of Ru1-3 in the presence of either monomeric or fibrillar Aβ1-42 results in the formation of large amorphous aggregates as a common endpoint, with Ru complexes incorporating the extended phenanthroline ligand accelerating this process and thereby limiting the formation of oligomeric species in the initial stages of the aggregation process that are reported to show considerable toxicity.

In Alzheimer's disease (AD), excessive amounts of quinolinic acid (QUIN) accumulate within the brain parenchyma and dystrophic neurons. QUIN also regulates glutamate uptake into neurons, which may be due to modulation of Na+-dependent excitatory amino acid transporters (EAATs). To determine the biological relationships between QUIN and glutamate dysfunction, we first quantified the functionality and kinetics of [3H]QUIN uptake in primary human neurons using liquid scintillation. We then measured changes in the protein expression of the glutamate transporter EAAT3 and EAAT1b in primary neurons treated with QUIN and the EAAT inhibitor L-trans-pyrrolidine-2,4-dicarboxylic acid (2,4-PDC) using western blotting and immunohistochemistry. Immunohistochemistry was further used to elucidate intracellular transport of exogenous QUIN and the lysosomal-associated membrane protein 2 (LAMP2). Structural insights into the binding between QUIN and EAAT3 were further investigated using molecular docking techniques. We report significant temperature-dependent high-affinity transport leading to neuronal uptake of [3H]QUIN with a Km of 42.2 μM, and a Vmax of 9.492 pmol/2 min/mg protein, comparable with the uptake of glutamate. We also found that QUIN increases expression of the EAAT3 monomer while decreasing the functional trimer. QUIN uptake into primary neurons was shown to involve EAAT3 as uptake was significantly attenuated following EAAT inhibition. We also demonstrated that QUIN increases the expression of aberrant EAAT1b protein in neurons further implicating QUIN-induced glutamate dysfunction. Furthermore, we demonstrated that QUIN is metabolised exclusively in lysosomes. The involvement of EAAT3 as a modulator for QUIN uptake was further confirmed using molecular docking. This study is the first to characterise a mechanism for QUIN uptake into primary human neurons involving EAAT3, opening potential targets to attenuate QUIN-induced excitotoxicity in neuroinflammatory diseases.

An Fe corrole is shown to bind to the amyloid-beta peptide and limit reactive oxygen species generation and peptide aggregation of relevance to Alzheimer's disease.

Alzheimer's disease (AD) is the most common form of dementia, leading to loss of cognition, and eventually death. The disease is characterized by the formation of extracellular aggregates of the amyloid-beta (Aβ) peptide and neurofibrillary tangles of tau protein inside cells, and oxidative stress. In this study, we investigate a series of Ru(III) complexes (Ru-N) derived from NAMI-A in which the imidazole ligand has been substituted for pyridine derivatives, as potential therapeutics for AD. The ability of the Ru-N series to bind to Aβ was evaluated by NMR and ESI-MS, and their influence on the Aβ peptide aggregation process was investigated via electrophoresis gel/western blot, TEM, turbidity, and Bradford assays. The complexes were shown to bind covalently to the Aβ peptide, likely via a His residue. Upon binding, the complexes promote the formation of soluble high molecular weight aggregates, in comparison to peptide precipitation for peptide alone. In addition, TEM analysis supports both amorphous and fibrillar aggregate morphology for Ru-N treatments, while only large amorphous aggregates are observed for peptide alone. Overall, our results show that the Ru-N complexes modulate Aβ peptide aggregation, however, the change in the size of the pyridine ligand does not substantially alter the Aβ aggregation process.

The p53 protein plays a major role in cancer prevention, and over 50 % of cancer diagnoses can be attributed to p53 malfunction. The common p53 mutation Y220C causes local protein unfolding, aggregation, and can result in a loss of Zn in the DNA-binding domain. Structural analysis has shown that this mutant creates a surface site that can be stabilized using small molecules, and herein a multifunctional approach to restore function to p53-Y220C is reported. A series of compounds has been designed that contain iodinated phenols aimed for interaction and stabilization of the p53-Y220C surface cavity, and Zn-binding fragments for metallochaperone activity. Their Zn-binding affinity was characterized using spectroscopic methods and demonstrate the ability of compounds L4 and L5 to increase intracellular levels of Zn2+ in a p53-Y220C-mutant cell line. The in vitro cytotoxicity of our compounds was initially screened by the National Cancer Institute (NCI-60), followed by testing in three stomach cancer cell lines with varying p53 status', including AGS (WTp53), MKN1 (V143A), and NUGC3 (Y220C). Our most promising ligand, L5, is nearly 3-fold more cytotoxic than cisplatin in a large number of cell lines. The impressive cytotoxicity of L5 is further maintained in a NUGC3 3D spheroid model. L5 also induces Y220C-specific apoptosis in a cleaved caspase-3 assay, reduces levels of unfolded mutant p53, and recovers p53 transcriptional function in the NUGC3 cell line. These results show that these multifunctional scaffolds have the potential to restore wild-type function in mutant p53-Y220C.

An Fe corrole is shown to bind to the amyloid-beta peptide and limit reactive oxygen species generation and peptide aggregation of relevance to Alzheimer's disease.

Recently, crystalized mouse ketimine reductase/CRYM complexed with NADPH was found to have pyruvate bound in its active site. We demonstrate that the enzyme binds α-keto acids, such as pyruvate, in solution, and catalyzes the formation of N -alkyl-amino acids from alkylamines and α-keto acids (via reduction of imine intermediates), but at concentrations of these compounds not expected to be encountered in vivo. These findings confirm that, mechanistically, ketimine reductase/CRYM acts as a classical imine reductase and may explain the finding of bound pyruvate in the crystallized protein.