https://orcid.org/0000-0002-4872-082X
Figures
Abstract
Infections with the pathogenic free-living amoebae Naegleria fowleri can lead to life-threatening illnesses including catastrophic primary amoebic meningoencephalitis (PAM). Efficacious treatment options for these infections are lacking and the mortality rate remains >95% in the US. Glycolysis is very important for the infectious trophozoite lifecycle stage and inhibitors of glucose metabolism have been found to be toxic to the pathogen. Recently, human enolase 2 (ENO2) phosphonate inhibitors have been developed as lead agents to treat glioblastoma multiforme (GBM). These compounds, which cure GBM in a rodent model, are well-tolerated in mammals because enolase 1 (ENO1) is the predominant isoform used systemically. Here, we describe findings that demonstrate these agents are potent inhibitors of N. fowleri ENO (NfENO) and are lethal to amoebae. In particular, (1-hydroxy-2-oxopiperidin-3-yl) phosphonic acid (HEX) was a potent enzyme inhibitor (IC50 = 0.14 ± 0.04 μM) that was toxic to trophozoites (EC50 = 0.21 ± 0.02 μM) while the reported CC50 was >300 μM. Molecular docking simulation revealed that HEX binds strongly to the active site of NfENO with a binding affinity of –8.6 kcal/mol. Metabolomic studies of parasites treated with HEX revealed a 4.5 to 78-fold accumulation of glycolytic intermediates upstream of NfENO. Last, nasal instillation of HEX increased longevity of amoebae-infected rodents. Two days after infection, animals were treated for 10 days with 3 mg/kg HEX, followed by one week of observation. At the end of the one-week observation, eight of 12 HEX-treated animals remained alive (resulting in an indeterminable median survival time) while one of 12 vehicle-treated rodents remained, yielding a median survival time of 10.9 days. However, intranasal HEX delivery was not curative as brains of six of the eight survivors were positive for amoebae. These findings suggest that HEX requires further evaluation to develop as a lead for treatment of PAM.
Author summary
The pathogenic free-living amoeba, Naegleria fowleri, can cause a fatal brain infection if contaminated freshwater is flushed up the nose. Unfortunately, the available treatment options are limited and infected patients have a very low chance of survival. In this work, we have discovered that inhibitors of sugar metabolism designed to treat brain cancer are toxic to the pathogen. The best inhibitor, HEX, was a potent killer of N. fowleri grown in the test tube, leading us to assess the activity of the agent in an animal model of disease. Although HEX extended the life of N. fowleri infected rats compared to rodents that did not receive compound, it was not found to be curative. These findings highlight the promise of the anti-cancer agents in the treatment of brain infections caused by Naegleria fowleri.
Citation: Milanes JE, Yan VC, Pham C-D, Muller F, Kwain S, Rees KC, et al. (2024) Enolase inhibitors as therapeutic leads for Naegleria fowleri infection. PLoS Pathog 20(8): e1012412. https://doi.org/10.1371/journal.ppat.1012412
Editor: Tomoyoshi Nozaki, The University of Tokyo, JAPAN
Received: January 22, 2024; Accepted: July 11, 2024; Published: August 1, 2024
Copyright: © 2024 Milanes et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The NfENO crystal structure is available in the RCSB protein data bank as PDB 7UGH.
Funding: This work was supported by the National Institutes of Health (R21AI175463, P20GM146584 and P20GM113226 to JCM; R01CA231509 to SWM; 75N93022C00036 to IQP. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Human infections by the pathogenic free-living amoebae (pFLA) can lead to life-threatening illnesses, a situation exacerbated by the lack of useful therapeutic drugs for successful interventions. Perhaps the best-known pFLA, Naegleria fowleri establishes infections after trophozoites in contaminated water are introduced into the nasal passages and ultimately penetrate the cribriform plate to infect the brain. This infection, known as primary amoebic meningoencephalitis, is almost invariably fatal due to tissue damage caused by the parasite and concurrent hemorrhagic necrosis of the brain leading to lethal cerebral edema or multiorgan system failure [1]. Due to the high mortality rate, poor therapeutic options, and lack of awareness and detection, this organism is classified as a category B NIAID emerging infectious pathogen.
We have previously found that glucose metabolism is critical to the viability of the infectious trophozoite lifecycle stage [2]. This pathway is anticipated to impinge on multiple critical cellular functions, as it provides the cell with ATP and key metabolic intermediates. For example, glucose 6-phosphate (G6P) generated by the first step in the process is needed for the pentose phosphate pathway (PPP), which generates essential metabolic products including reducing equivalents and sugar nucleotide biosynthetic intermediates.
The role of glucose metabolism during infection is unclear. A comparison of transcriptomes from continuously cultured N. fowleri to amoebae recently isolated from an infected brain revealed that multiple metabolic genes were upregulated including those involved in fatty acid metabolism as well as glycolysis, further clouding the role of glucose metabolism during infection [3]. Growth of N. gruberi, a non-pathogenic Naegleria, requires glucose, even though the organism preferentially metabolizes fatty acids [4]. This suggested that some level of glycolysis or gluconeogenesis is likely also essential for N. fowleri viability, a finding supported by the lack of trophozoite growth when cultured in the absence of glucose [2].
Lin and colleagues have recently described the development of phosphonate inhibitors of human ENO2, the enzyme that interconverts 2-phosphoglycerate (2-PG) to phosphoenolpyruvate (PEP) in glycolysis and gluconeogenesis, for the treatment of ENO1-/- GBM [5]. These inhibitors selectively killed ENO1-deleted glioma cells and eliminated intracranial orthotopic ENO1-deficient tumors in mice. One of the lead compounds, HEX, had an EC50 value of 1.3 μM against ENO1-/- glioma cells but was very well-tolerated by ENO1-rescued glioma and other control cells, with CC50 values >300 μM. HEX also almost completely inhibited intracranial tumor growth and these impacts did not require a breached blood-brain barrier. Importantly, therapeutic doses were at levels well-tolerated in non-human primates (NHPs), with IP delivery of HEX tolerated to 300 mg/kg (body weight) without adverse effects, yielding plasma concentrations of about 750 μM for several hours with a plasma elimination half-life in non-human primates of 61 minutes.
Given the importance of glycolysis to N. fowleri and the known distribution of the phosphonate ENO2 inhibitors to the brain, we have assessed the activity of these compounds as potential anti-amoebic agents. This includes scoring the impact of the inhibitors against the N. fowleri ENO (NfENO) through a series of biochemical assays and against amoebae trophozoites, both in vitro and in a rodent model of disease. The findings of these efforts suggest the agents could be a useful component for treatment of N. fowleri infections.
Materials and methods
Ethics statement
Rats were purchased from Charles River Laboratories and maintained in the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) accredited, PHS assured, USDA registered Godley Snell Research Center at Clemson University. All protocols used in this study were approved by the Clemson University Institutional Animal Care and Use Committee (CU PHS Assurance Number D16-00435 (A3737-01) (Protocol #2023–0121) and performed in accordance with AAALAC International standards.
Chemicals and reagents
β-nicotinamide adenine dinucleotide phosphate (NADP+), adenosine diphosphate (ADP), and dimethyl sulfoxide (DMSO) were obtained from Fisher Scientific (Pittsburgh, PA, USA). Pyruvate kinase/lactate dehydrogenase (PK/LDH) from rabbit muscle and 2-phophoglyceric acid disodium salt hydrate (2-PG) were both purchased from MilliporeSigma (Burlington, MA, USA). Human ENO1, 2, and 3 were graciously provided by Dr. Paul Leonard (MD Anderson Cancer Center Core for Biomolecular Structure and Function).
The compounds (1-hydroxy-2-oxopyrrolidin-3-yl) phosphonic acid (deoxy-SF2312), (1-hydroxy-2-oxopiperidin-3-yl) phosphonic acid (HEX), and (1-hydroxy-2-oxoazepan-3-yl) phosphonic acid (HEPTA), were synthesized following published procedures reported in the literature [5,6]. Please see the synthetic scheme (S1 Fig) and spectroscopic characterization (S2 Fig).
Protein purification, crystallization, and structure determination
Protein purification, crystallization, and structure determination were conducted by the Seattle Structural Genomics Center for Infectious Disease (SSGCID) [7,8] following standard protocols described previously [9–11]. (See S1 Text for further information.) Briefly, E.coli codon optimized sequence encoding residues 2–512 of the full length protein (NF0118810, AmoebaDB, https://amoebadb.org/amoeba/app/) was used to develop five constructs composed of residues 14–512, 28–512, 44–512, 62–512, 80–512 that were predicted by Xtalpred (https://xtalpred.godziklab.org/XtalPred-cgi/xtal.pl) to be optimal for crystallization. Following cloning into pBG1861 [10], which appends a non-cleavable 6xHis N-terminal tag to fusion proteins, constructs were transformed into E. coli BL21(DE3)R3 Rosetta cells. Two constructs, NfENO(44–512) and NfENO(62–512), yielded soluble protein. Please note that an alternate N-terminal translation start site has been annotated in Uniprot as A0A6A5BXC3, yielding a predicted 485 residue polypeptide that includes residues equivalent to 28–512 of NF0118810 (see S3 Fig and S1 Table).
Following large scale growth in auto-induction media [12], bacteria were lysed by sonication and clarified lysate was subject to sequential Ni2+-affinity and size-exclusion chromatography (SEC). The SEC peak fractions eluted as a single large peak consistent with a molecular mass of about 50 kDa, suggesting a monomeric enzyme, which was consistent with the size of the denatured purified protein as determined by SDS-PAGE. Peak fractions were pooled and concentrated.
Purified NfENO(44–512) and NfENO(62–512) were screened for crystallization in 96-well sitting-drop plates against the JCSG++ HTS (Jena Bioscience) and PACT premier HT96 (Molecular Dimensions) crystal screens. In addition to protein-only crystallizations, co-crystallization trials with 2-PG (5 mM) and HEX (5 mM) were also pursued. Only NfENO(44–512) in complex with 2-PG crystallized to form diffraction quality crystals, which were flash-frozen by plunging directly into liquid nitrogen without cryoprotectant exchange.
X-ray data were collected at 100 K on beamline 21-ID-F, LS-CAT, at the Advanced Photon Source, Argonne National Laboratory. Data were processed with XDS reduced with XSCALE [13] and the structure was solved by molecular replacement using the structure of B. subtilus ENO (PDB 4A3R) as a search model [14]. Structures were refined using iterative cycles of Phenix [15] followed by manual rebuilding of the structure using Coot [16] to yield a resolution of 1.95 Å, with the quality of all structures checked using MolProbity [17]. All data-reduction and refinement statistics are shown in S2 and S3 Tables. Structure figures were prepared and analyzed using PyMOL (v.1.5; Schrodinger) and coordinates and structure factors have been deposited in the Protein Data Bank www.rcsb.org [18] with accession number 7UGH.
ENO Assays
ENO was assayed in triplicate using a coupled reaction to measure enzyme activity [19]. Briefly, enzyme (50 nM) in assay buffer (100 mM HEPES, pH 8.0, 3.3 mM MgSO4) was loaded in black 96-well plates with or without inhibitors. To start the reaction, substrate buffer (3.75 mM 2-PG, 1.75 mM ADP, 1 U of pyruvate kinase/lactate dehydrogenase, and 0.4 mM NADH in assay buffer) was added to each well. The rate of reduction in fluorescence (excitation: 360 nm; emission: 460 nm), a measure of NAD production, was then scored on a Biotek Synergy H1 microplate reader every 20 seconds for 4 minutes. Kinetic analyses were performed with Prism 9.0 (GraphPad Software, San Diego, CA) using the Michaelis-Menten model. For assays involving inhibitors, compounds (in DMSO) were incubated at RT with enzyme in assay buffer (pH 7) for 15 minutes prior to initiation of the reaction by addition of the substrate. All assays were performed in triplicate with vehicle control (DMSO) included.
Ligand preparation, optimization, and molecular docking
The chemical structures of HEX, (1,5-dihydroxy-2-oxopyrrolidin-3-yl) phosphonic acid (SF2312), deoxy-SF2312, and HEPTA were drawn with ChemOffice professional 19 suite (PerkinElmer, Waltham, MA), and three-dimensional (3D) structures were generated with VeraChem Vconf (VerChem LLC, Germantown, MD). The 3D structures were optimized by Gaussian 16 suite (Gaussian Inc., Wallingford, CT) with Density Functional Theory (DFT), employing the B3LYP/6-311G (d,p) level of theory [20]. The phosphonate-bearing stereocenter was modeled with the (R)-configuration by analogy to the enantiomer of HEX that crystallized in human ENO (PDB 5IDZ). Since SF2312 likely exists as a mixture of cis and trans diastereomers at physiological pH, we modeled both diastereomers independently while maintaining the phosphonate-bearing stereocenter in the (R)-configuration [5, 6, 21]. The NfENO 3D crystal structure (PDB 7UGH) was retrieved from the RCSB protein data bank. The protein was then prepared for the docking analysis by first removing the co-crystallized ligand, heteroatoms, and water molecules using Pymol Molecular Graphics 2.0 (Schrödinger LLC, New York, NY). The optimized ligands and the protein were further organized using AutoDock Tools (The Scripps Research Institute, La Jolla, CA) to convert all structures into pdbqt formats. The grid box was placed around the region of the active site of the protein. The size of the grid box was kept at 48, 52, and 52 for X, Y, and Z respectively with the center of the grid box maintained at –18.910, –1.626, and –15.343 respectively for X, Y, and Z. The molecular docking studies were carried out in vacuo with AutoDock vina using specific docking parameters and scoring functions [22]. The binding affinities of ligands were measured in kcal/mol as a unit for a negative score [22]. The binding conformation with the highest negative value was taken as the best pose for the corresponding protein-ligand complex. Subsequently, the best binding pose of each complex was analyzed using Pymol and Discovery Studio (Dassault Systèmes, Waltham, MA) to reveal the protein-ligand interactions.
In vitro growth inhibition and selection for resistance
N. fowleri strain Nf69 (ATCC 30215) or strain TY (ATCC 30107) trophozoites and N. gruberi strain NEG-M (ATCC 30224) were cultured axenically as previously described [2]. Briefly, parasites were seeded at 1 x 104 cells/mL (N. fowleri) or 5 x 103 cells/mL (N. gruberi) in 100 μL of the indicated medium in white TC-treated 96-well plates (Thermo Scientific Nunc Microwell 136101). For assays under standard growth conditions ("+ glc"), N. fowleri cells were seeded into Nelson’s Complete Media (NCM, 0.17% liver infusion broth (BD Difco, Franklin Lakes, NJ), 0.17% glucose, 0.012% sodium chloride, 0.0136% potassium phosphate monobasic, 0.0142% sodium phosphate dibasic, 0.0004% calcium chloride, 0.0002% magnesium sulfate, 10% heat-inactivated fetal bovine serum (FBS), 1% penicillin-streptomycin) at 37°C. For assays in modified medium ("-glc/+gly"), cells were seeded into the media without glucose but supplemented with 10 mM glycerol. For standard growth of N. gruberi, cells were seeded into PYNFH-complete media (1.0% peptone, 1.0% yeast extract, 0.1% yeast nucleic acid, 0.0015% folic acid, 0.0001% hemin, 0.0362% potassium phosphate monobasic, 0.05% disodium phosphate, 10% heat-inactivated fetal bovine serum (FBS), 1% penicillin-streptomycin) at 25°C. For determination of EC50 value, concentrations of drug ranging from 0.07826–5 μM were used. After culturing for 48 hr (37°C, 5% CO2 for N. fowleri, 25°C, 5% CO2 for N. gruberi) in the presence of compound or vehicle, plates were equilibrated to RT for 15 min and CellTiter Glo reagent (Promega, Madison, WI, USA) was added followed by orbital shaking for 2 min. After an additional incubation (RT, 10 min), luminescence was scored on a BioTek Synergy H1 mircoplate reader. Parasite density was scored and kinetic analyses were performed with Prism 9.0 (GraphPad Software, San Diego, CA) using the Michaelis-Menten model. All viability assays were performed in technical triplicates with at least three biological replicates.
To generate resistance in N. fowleri, cells were seeded in 75 cm2 TC-treated flasks (Falcon 353135) at 1 x 104 cells/mL in the presence of HEX, starting at the calculated (using GraphPad software) EC25 concentration. After one week, HEX levels were increased to the EC50 concentration for an additional two weeks. At this point, the compound was increased weekly (2 x EC50 value, 5 x EC50 value, and 10 x EC50 value), with cells passaged as needed throughout in the presence of the compound. Amoebae were then cultured for one month in 10 x EC50 levels of HEX before performing viability assays.
HEK-293 (ATCC CRL-1573) cells were cultured axenically in Dulbecco’s Modified Eagle Medium with 4.5g/L D-glucose, L-glutamine (Gibco 11965) and 10% fetal bovine serum (FBS) at 37°C, 5% CO2. For determination of CC50 values, cells were seeded at 1 x 104 /mL in the presence of compound at concentrations ranging from 3.125–200 μM. After a 48 hr incubation (37°C, 5% CO2), CellTiter Blue reagent (Promega, Madison, WI, USA) was added to each well. After an additional 3-hour incubation (37°C, 5% CO2) with reagent, fluorescence was scored on a a BioTek Synergy H1 mircoplate reader and percent growth inhibition values were solved.
Analysis of metabolites after HEX treatment
For unbiased metabolomics analyses, N. fowleri trophozoites were grown to near 100% confluency in 6-well TC-treated plates (Thermo Scientific Nunclon Delta Surface 140685). HEX (25 μM) or vehicle control was added to appropriate wells for an additional 6 hours. Samples were then washed three times in RT PBS prior to the addition of ice-cold CH3CN. Plates were then incubated at –20°C for 20 min, followed by the addition of ice-cold nanopure H2O and removal of the solution to sterile microcentrifuge tubes. This process was repeated twice, and the total volume of extraction material was adjusted to 2.5 x 106 cell equivalents per sample. Tubes were stored at –80°C prior to analysis.
Metabolomic analysis
Each cell sample was vortexed with glass beads for 5 min in a MM200 Mix Grinder (Retsch USA, Newtown, PA, USA). The mixture was centrifuged at 14,000 rpm for 20 min at 4°C. The supernatant was transferred into a new tube and lyophilized overnight. The dried sample was reconstituted in 60 μL 50% CH3CN and analyzed in random order on a parallel two-dimensional liquid chromatography-mass spectrometry (2DLC-MS) system, a Thermo DIONEX UltiMate 3000 HPLC system (Thermo Fisher Scientific, Waltham, MA, USA) coupled with a Thermo Q Exactive HF Hybrid Quadrupole-Orbitrap Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). The UltiMate 3000 HPLC system was equipped with a hydrophilic interaction chromatography (HILIC) column and a reversed-phase chromatography (RPC) column configured in parallel mode. Each sample was analyzed by the 2DLC-MS under both positive (+) and negative (–) modes to obtain full MS data for metabolite quantification. A pooled sample was also prepared for each sample group and analyzed by 2DLC-MS/MS to acquire MS/MS spectra at three collision energies (20, 40, and 60 eV) for metabolite identification. Detailed 2DLC-MS operation parameters and data analysis methods were reported in our previous publications [23–25].
Rodent model of infection
To determine dosing regimen, nine CD rats (males, aged 42–48 days, Charles River) were treated (once daily for five days) by intranasal instillation with 1.6, 16, or 120 mg/kg HEX in PBS and observed for adverse effects. While none of the rats had any noticeable reaction to treatment (and weight gain was similar amongst all animals), the high viscosity of the highest concentration raised concerns about the agent reaching the depths of the nasal cavity, leading us to continue tests using 16 mg/kg HEX.
Twenty-four CD rats (12 females, 12 males, aged 42–48 days, Charles River) housed in pairs were infected by nasal instillation with 2.5 x 105 N. fowleri strain 69 trophozoites that had been isolated after passage through mice to ensure virulence. Rats were provided unlimited food and water throughout the duration of the experiment. The two groups consisted of equal numbers of males and females. Groups received either vehicle (PBS) or HEX (3 mg/kg) delivered by nasal instillation (25 μL/nare) for 10 consecutive days starting 48hr after infection. Rodents were monitored carefully and euthanized when symptoms of infection, including failure to respond to stimuli or seizures, were evident. After 17 days, surviving animals were sacrificed and brains were removed to assess cure. Rats were considered cured if amoebae from isolated brains failed to grow in culture after two weeks.
Results
N. fowleri trophozoites are reliant on glycolysis for normal growth in vitro, suggesting that inhibitors of enzymes in the pathway could serve as useful anti-amoebic compounds. Supporting this supposition, inhibitors of glucokinase, the first enzyme in the pathway, have been identified that have modest amoebicidal activity [2]. Recently, a series of phosphonate inhibitors of human ENO, which catalyzes the conversion of 2-PG to PEP in glycolysis and the reverse reaction in gluconeogenesis, have been developed as potential treatments for cancer in the brain [5]. These compounds have been shown to cure a rodent model of GBM through inhibition of mammalian ENO2 while being well-tolerated in non-human primate models.
Phosphonohydroxamates inhibit N. fowleri ENO in vitro
N. fowleri harbors a single ENO gene (NfENO, NF0118810) that is 47% identical to human ENO2 (Table 1), being predicted to be a 512 amino acid protein (molecular weight, 56 kDa). (Please note, the ENO gene from the Ty strain of N. fowleri, NfTy_054390, encodes a protein predicted to be 662 residues, due to an extension at the N-termini but otherwise identical). In both genes, the N-terminal region is predicted to be disordered; a truncation of NF0118810 (residues 44–512) yielded soluble protein, while full-length protein expression was not successful. This truncation, NafoA.00379.a.B5.PW38978, had an apparent KM for 2-PG of 923 ± 0.336 μM, within the ranges reported for human skeletal muscle ENO (0.2 mM) and Leuconostoc mesenteroides (2.61 mM) [26,27]; a similar truncation used in crystallography studies yielded PDB 7UGH.
NfENO shares key residues required for interactions with phosphonate inhibitors found to inhibit human ENO2, including Arg372 (in ENO2) required to form a salt bridge with the compound [5]. Additionally, the limited cytotoxicity and promising distribution to the brain suggested that the phosphonate compounds could be ideal early lead chemotypes for anti-N. fowleri development. As a first step to assess the possible use of the agents against amoebae, we tested a panel of analogs against NfENO and the human enolases (Fig 1). The phosphonate compounds were potent inhibitors of recombinant NfENO. SF2312, a phosphonate antibiotic produced by the actinomycete Micromonospora with pan-enolase inhibition activity [6], had an IC50 value of 0.31 ± 0.07 μM. HEX (1-hydroxy-2-oxopiperidin-3-yl) phosphonic acid, a six-membered ring structure, had improved specificity for ENO2 [5] and similarly was the most potent of the compounds tested against NfENO with an IC50 value of 0.14 ± 0.04 μM. As it is against ENO2, HEX is a competitive inhibitor of NfENO. Five- and seven-membered analogs of HEX were active against NfENO, including deoxy-SF2312 and HEPTA (Fig 1). Modification of the phosphonates to generate prodrugs largely eliminated enzyme inhibition, likely due to steric interference of the added pivaloyloxymethyl (POM) groups. BenzylHEX, which was designed as a control that cannot bind (and therefore cannot inhibit) human ENO2 [5], also had no impact on NfENO at 10 μM, indicating the hydroxamate portion of HEX is important for activity.
To assess the selectivity of the compounds, inhibitors were tested against human ENO2 (gamma-enolase, ENO2). Generally, the sensitivities of ENO2 to inhibitors paralleled those of NfENO, with HEX being the most potent ENO2 inhibitor tested (Fig 1). The one notable difference was deoxy-SF2312, which was more than 10-fold more potent against ENO2 than NfENO.
Compounds were tested against NfENO, HsENO2, N. fowleri trophozoites, and HEK-293 cells. (NA) Not applicable. Compounds were not active against N. fowleri trophozoites. (ND) Not determined.
Crystal structure of NfENO
We determined the X-ray crystal structure of homodimeric NfENO at 1.95 Å resolution, with 2-PG bound at the active site (PDB 7UGH) (Fig 2A). A comparison with the human HEX-bound ENO2 (PBD 51DZ [5]) shows that the overall fold is well conserved, with a root-mean-square deviation (RMSD) of 1.98 Å across the 417 Cα atom pairs (Fig 2B). The phosphate group of the ligands occupies the same position, and the binding pocket is highly conserved; on the superimposed structures, all residues within 5 Å of HEX are identical except for Lys243 that aligns with Ser157 in human ENOs 1, 2, and 3 (see alignment, S3 Fig). In the NfENO structure, the amine group of the Lys243 side-chain lies close to the phosphate group of the ligand, in contrast to Ser157 in the human enzyme (see Fig 2B, superposition of the active sites). According to the b-factor, Lys243 is in a flexible loop. While the predicted additional binding of NfENO-Lys243 and HEX occurs uniquely in NfENO, human ENO2 might employ alternative binding modes to achieve a similar level of inhibition potency. Indeed, structural analysis through superposition of the active sites of the HEX-bound ENO2 and HEX-docked NfENO (S4 Fig) revealed that the phosphate group in HEX interacts with the side chain of Arg-372 in ENO2, a residue that is absent in NfENO. Altogether, these observations suggest that Lys243 may be a good target for developing inhibitors specific to the N. fowleri enzyme.
(A) Crystal structure showing 2-PG bound at the active site of NfENO (PDB 7UGH, 1.95 Å). (B) The active site of NfENO (PDB 7UGH, grey) overlayed on the human HEX-bound ENO structure (PDB 5IDZ, pink), showing that all residues of the binding pocket are conserved with the sole exception of Lys243, which corresponds to Ser157 in the human enzyme (all other residue numberings refer to NfENO). Carbon atoms of the NfENO 2-phosphoglyceric acid ligand are colored green, while those of HEX are shown in black. The phosphate groups of the two ligands overlap and are both positioned close to NfENO Lys243; the yellow dotted line shows the closest interatomic distance between the Lys243 amine group and the HEX phosphate.
Molecular modeling of HEX compounds with NfENO
To gain insight into how HEX, SF2312, deoxy-SF2312, and HEPTA interact with the NfENO enzyme at the atomic level, we performed a molecular docking simulation of the compounds with NfENO (PDB 7UGH) using Autodock Vina (Fig 3). The docking analysis revealed that HEX binds strongly to the active site of NfENO with a binding affinity of –8.9 kcal/mol, a slightly higher affinity than 2-PG (–8.2 kcal/mol). The binding pose of the NfENO-HEX complex showed that the hydroxamate and carbonyl moieties chelated a Na+ cation while the phosphonate moiety formed polar interactions with Arg460, His459, Ser461, and Ala121 residues. Additionally, Lys243, a residue found in NfENO but not human ENO2, appears to engage in productive interactions with both the hydroxamate and phosphonate through side chain interactions. The NfENO-HEX interactions from the docking simulation were similar to the interactions observed in the co-crystallization of HEX with human ENO2 (PDB 5IDZ) with the exception of the additional interactions with Lys243 noted above [5]. Further, SF2312, deoxy-SF2312, and HEPTA also bind to the active site of the NfENO enzyme with comparatively weaker binding affinities of –7.6, –6.5, and –6.1 kcal/mol, respectively. Interestingly, Lys243 remains unengaged with bound compound in the docking models of these structural analogs, in contrast to the predicted binding interaction with HEX. These results further underscore the potential importance of leveraging additional interactions with this residue–unique to NfENO–as a design strategy in future drug development campaigns. The predicted poorer binding affinity of the HEX analogs is consistent with the blunted efficacy of these compounds against NfENO in vitro as compared to HEX. These results also indicate that the six-membered ring of HEX offers a rather tuned binding conformation to fit into the binding pocket of NfENO compared with the smaller-sized SF2312 and deoxy-SF2312 and the larger size of HEPTA.
Binding poses of (A) 2-PG, (B) HEX (note the unique engagement of Lys243 with drug), (C) SF2312, (D) deoxy-SF2312, and (E) HEPTA at the active site of NfENO.
The phosphonate inhibitors are potent anti-amoebic compounds
To assess the impact of these compounds on amoebae viability, trophozoites were cultured in the presence of compound or vehicle (1% DMSO) for 48 hr, and viability was scored with CellTiter-Glo reagent. SF2312 was a potent amoebicide, with an EC50 value of 1.6 ± 0.11 μM (Fig 1). Modification of SF2312 to an esterase-labile prodrug by the addition of pivaloyloxymethyl (POM) moieties to yield POMSF modestly (about 2.4-fold) increased the potency of the compound.
HEX was almost 7-fold more potent as an amoebicide than SF2312, with an EC50 value of 0.21 ± 0.02 μM. Unlike SF2312, modification of HEX with POM moieties to yield POMHEX did not improve the inhibitory potency of HEX. Other phosphonate prodrug analogs, including CDP25, were active in the cell toxicity assays even though they lacked activity against enzyme in vitro, suggesting that cellular esterases had processed the compounds into their active form. Nevertheless, none of the prodrugs were as potent as HEX, which was at least six-fold more active against N. fowleri than it was against human ENO1-deficient cells that it was designed to kill [5]. Even a short exposure to HEX was sufficient to suppress amoebae viability, with a one-hour treatment of HEX followed by 48 hr of culture in HEX-free media resulting in an EC50 value of 8.06 ± 2.6 μM.
Phosphonate toxicity was not limited to Nf69 strain N. fowleri (ATCC 30215) trophozoites, as TY strain N. fowleri (ATCC, 30107; EC50 value of 0.87 ± 0.23 μM) was also sensitive to the compound. However, the agent was not broadly anti-amoebic, as HEX had no impact on the growth of the non-pathogenic amoebae N. gruberi, the pathogenic free-living amoeba Balamuthia mandrillaris, or the obligate parasite Entamoeba histolytica.
Interestingly, amoebae could be adapted to the presence of HEX by continuous culture in gradually increasing amounts of the agent. Cells grown in the presence of increasing concentration of HEX over a 2.5-month period became more tolerant of the agent, with near-normal growth rates at concentrations ten-fold higher than the EC50 value for naïve amoebae. In viability assays, these amoebae were more resistant to the agent, with an EC50 value of 0.57 ± 0.30 μM, a value 2.7-fold higher than unacclimated cells in the same medium (0.21 ± 0.03 μM). Removal of selection for three weeks resulted in reversion to HEX sensitivity (EC50 value of 0.11 ± 0.01), suggesting the change was dynamic and not permanent.
Amoebae grown on alternative carbons sources remain sensitive to HEX
Based on transcriptomics analyses, amoebae from the brains of infected rodents upregulate a suite of metabolic genes, including several involved in lipid metabolism [3]. We anticipate that NfENO inhibition would impact both glycolysis (from glucose to pyruvate) and gluconeogenesis (from pyruvate to G6P) (Fig 4A). However, glycerol from lipid metabolism could circumvent NfENO inhibition and feed gluconeogenesis through conversion of glycerol 3-phosphate (Gly3P) to dihydroxyacetone phosphate (DHAP) to generate G6P, potentially ameliorating the toxicity of NfENO inhibition (Fig 4A).
(A) Schematic of glycolysis and gluconeogenesis in N. fowleri. (B) Comparative metabolite intensities from amoebae cultured with two different carbon sources (glucose, +glc, or glycerol (-glc,+gly)). Statistical analysis was performed using pairwise two-tail tests with the Benjamini Hochberg method for multiple testing correction with the false discovery rate set at 25%. * indicates significance after the t test and the Benjamini Hochberg correction.
To assess this, amoebae grown in standard medium, which contains abundant glucose (9 mM, “+glc”) were transferred to medium with glycerol (10 mM) and no glucose (“-glc/+gly”), which supported normal trophozoite proliferation with no change in doubling time. While amoebae in glucose-bearing media were highly sensitive to the inhibitor, (EC50 values of 0.21 ± 0.02 μM), amoebae cultured in the absence of glucose (with glycerol as an alternative carbon source) were less sensitive to the agent, with an EC50 value about 10-fold higher (2.5 ± 0.11 μM).
HEX treatment in vitro leads to accumulation of glycolytic intermediates
Metabolomic profiling of amoebae grown in glucose and treated with HEX revealed a marked accumulation of glycolytic intermediates upstream of NfENO (Fig 4B and S4 Table). The increased intermediates include G6P, F6P, and 2-/3-PG, with 7.4-, 4.5- and 78-fold increases, respectively (see S5 Table). Pyruvate levels were essentially unchanged. Gly3P abundance, presumably from increased levels of DHAP, also increased 17.9-fold from what was observed in untreated cells.
Amoebae grown in glycerol instead of glucose had an increased abundance of metabolites downstream of the carbon source when compared to cells grown in glucose. Elevated intermediates included Gly3P (150-fold, when compared to cells in glucose), and 2-/3-PG (11-fold). Increased flux from these intermediates may explain the 5.8-fold increase in pyruvate over levels seen in cells grown in glucose. While HEX treatment of these cells led to further accumulation of 2-/3-PG (7.2-fold higher levels than untreated cells grown in glycerol), HEX caused a nearly 50% decrease in pyruvate levels (see S6 Table).
HEX extends the life of infected rodents, but is not curative
The promising in vitro anti-amoebic activity of HEX and its performance in non-human primates, where IP delivery at 300 mg/kg (body weight) was well tolerated and yielded plasma concentrations of about 750 μM for several hours [5], encouraged us to consider the value of the agent in a rat disease model. Rats have a plasma elimination half-life that is very similar to that of the non-human primates (about 60 minutes, [5]), suggesting that parenteral delivery could prove useful. However, disease progression can be rapid, leading us to consider using intranasal instillation as a means of delivery of HEX to the site of infection, including the olfactory bulb.
To assess the impact of HEX on amoebae in an authentic setting, rats were infected by introduction of 2.5 x 105 trophozoites into the nares (Day 0, Fig 5) and then treatment was initiated two days later by nasal instillation of either HEX or PBS vehicle once a day for 10 days. At the conclusion of treatment on day 11, most (11 of 12) HEX-treated rodents remained. This contrasted with the vehicle treated group, which had five of 12 remaining. By the end of the experiment on the eighteenth day post-infection, 8 of twelve HEX-treated animals remained alive, resulting in an inability to score median survival time, while 1 of twelve vehicle-treated animals survived (median survival time of 10.9 days). Sex did not influence survival after HEX treatment, as there was no difference in median survival time between HEX-treated males and females.
Survival could have resulted from either elimination of the pathogen or suppression of the infection. To assess this, brains were extracted from all survivors and scored for infection by culture of the pathogen. Of the eight HEX treated animals that survived the experiment, six had on-going infections as detected by culture of amoebae from extracted brains, suggesting that HEX treatment suppressed disease but was not curative. The lone PBS treated control survivor was negative for amoebae, likely because an infection was never established.
(A) Schematic of experimental set-up. Rats were infected with 2.5 x 105 N. fowleri trophozoites on D0 by nasal instillation and then treated with vehicle or HEX for 10 days starting on D2 by nasal instillation. Twelve rats were in each group (6 males, 6 females). Created with BioRender.com. (B) Kaplan-Meier survival curves for PBS- and HEX-treated groups. Median survival time for vehicle-treated animals was 10.88 days post-infection, while the median survival time for HEX-treated rats was undefined (since >50% of the population survived). A log-rank (Mantel-Cox) test determined a significant difference between the median survival time for each group with a p-value of 0.0006 (***).
Discussion
Current treatment regimens for N. fowleri infection include agents that have been empirically identified because of their use in one of the few cases of successful treatment. More recently, miltefosine and azithromycin have been added to treatment cocktails, but infection outcomes remain poor. With the goal being the development of improved treatments for N. fowleri infections, we have explored the value of the amoeba ENO enzyme, a key component in both glycolysis and gluconeogenesis, as a potential therapeutic target. In glycolysis, ENO catalyzes the conversion of 2-PG to PEP for use by pyruvate kinase in the payoff stage that yields ATP. ENO also plays a role in gluconeogenesis, with PEP generated from gluconeogenic substrates like lactate and amino acids being converted by ENO to 2-PG which ultimately yields G6P that is required for the pentose phosphate pathway.
Here, we have found that phosphonate ENO inhibitors designed to be selective for human ENO2 are active against amoeba enzyme and toxic to the trophozoite stage. N. fowleri is remarkably sensitive to one such inhibitor, HEX, with EC50 values about 6-fold lower than those required to kill ENO-1-deficient GBM cells, the cancer type that was the target of the initial phosphonate design campaign [5]. The reasons for the hypersensitivity of the amoebae are unclear but could be related to differences in the uptake of the compound between the protist and human cells. Generally, the cell permeability of phosphate-bearing compounds, including phosphonate-based drugs, is poor given that they are charged at physiological pH [28,29]. Some bacteria use an ABC transporter system encoded in the gene cluster phnCDE for phosphonate uptake [30]. The phnC gene, which shares similarity with nucleotide-dependent permeases, has been identified in eukaryotic phytoplankton that can utilize phosphonates as a phosphorus source [31], and N. fowleri harbors a possible homolog (E-value = 2 x 10−30, KAF0981242.1, FDP41_013030, AmoebaDB). The amoebae have homologs of additional known phosphonate transporters, although these proteins have other described or predicted roles indicating that phosphonate uptake may be a moonlighting function. These additional candidates include potential glycerol-3-phosphate (Gly3P) and hexose-6-phosphate transporters that have been implicated in the uptake of the phosphonate drugs fosmidomycin and fosfomycin in some bacteria [32], candidates that will be assessed in future loss of function studies as genetic tools for gene manipulation in N. fowleri become available.
N. fowleri responds to different environments with a dynamic metabolic approach to exploit available resources. In vitro, the modest lipid amounts in the 10% fetal bovine serum included in typical media (yielding 2–3% of the fatty acids and cholesterol normally available to cells in a vertebrate [33]) is insufficient to support growth [2]. Nevertheless, lipid metabolism may be important to amoebae in the brain, based on observed increases in fatty acid metabolism gene expression in isolates from infected rodents [3]. Additionally, lipids are the preferred carbon source of N. gruberi. The likely importance of lipid metabolism raises the question of why HEX is toxic to amoebae? First, glucose metabolism remains critical, even in the presence of lipids with glycolytic genes being upregulated in the brain-isolated amoebae. In N. gruberi, small levels of glucose are consumed even in the presence of abundant lipids, likely for the critical role of replenishing TCA intermediates through anaplerosis [4].
We anticipate that NfENO inhibition would impact both glycolysis (from glucose to pyruvate) and gluconeogenesis (from pyruvate to G6P). Given that, glycerol from lipid metabolism could circumvent NfENO inhibition and feed gluconeogenesis to generate G6P, ameliorating the toxicity of NfENO inhibition. However, cells grown in either glucose or glycerol remained sensitive to the agent. In both cases, the accumulation of glycolytic intermediates, in particular 2-/3-PG, could explain toxicity. Glycolytic intermediates serve as important allosteric regulators of other pathways, with their accumulation being dysregulating. If this were true, HEX treatment of cells grown in glycerol (which had increased levels of Gly3P, 2-/3-PG, and pyruvate relative to HEX-treated cells in glucose), should have been more sensitive to the agent. But they are not, being about 10-fold less sensitive, suggesting that the causes of HEX toxicity are more complex. One possibility is that 2-PG accumulation disrupts amoebae osmoregulation through an impact on trehalose, an osmoprotectant found in other systems, including N. gruberi [34]. Trehalose-6-phosphate synthase (TPS, which N. fowleri harbors as NfTy_062120) is required for trehalose synthesis and could be inhibited by 2-PG [35]. While we anticipate that HEX-triggered accumulation of 2-PG would inhibit TPS similarly in cells grown in either glycerol or glucose, the glycerol in the medium may partially substitute for trehalose as an osmoprotectant.
While HEX did not eliminate the amoebae in most of the treated animals, it did have a profound impact on the development of disease, with only one rodent succumbing to infection during treatment. This suggests that higher concentrations of agent or increased frequency of treatment could prove useful for eradication of the infection. Of note, HEX resistance was not a likely explanation for the lone HEX treatment failure, as amoebae isolated from the HEX treated rat remained as sensitive to HEX as the amoebae from a vehicle treated rodent (S5 Fig).
While most of the treated animals remained infected and HEX was not curative, disease symptoms were largely suppressed while the agent was being provided. It is unclear how the amoebae were avoiding the lethal impact of HEX, but it is possible that in the brain environment the organisms encyst, lying quiescent until treatment was removed. Nevertheless, the delay in disease progression may prove useful in treatment regimens, as HEX may allow other promising, but slow acting, anti-amoebic agents time to exert their lethal effect. Further, continued treatment beyond the 10-day period could sufficiently suppress the amoebae to allow the host immune system to respond to eliminate the invader.
Intranasal instillation has proven to be a fast-acting route of drug administration for several brain-active agents, including anticonvulsive compounds [36,37]. The success of the HEX treatment may be in part due to direct transport of the agent to the brain, which would reduce the impact of dilution by distribution [38]. Direct transport can result from endocytosis by olfactory sensory cells followed by exocytosis into the synaptic clefts in the olfactory bulb, which would introduce HEX into the CNS in a position to interact with invading amoebae at the site of brain infection [39].
It is unlikely that HEX will serve as a component of a pan-species anti-amoebic cocktail as the compound is inactive against other pathogenic free-living amoebae, including Balamuthia mandrillaris. B. mandrillaris harbors an ENO (BmENO, 000586F| quiver) that is 56% identical to human ENO2 and is predicted to have the residues required for HEX binding (Table 1). The lack of sensitivity of B. mandrillaris suggests that BmENO may be less important for the metabolic success of the amoebae. Alternatively (or additionally), the observed difference in toxicity may result from differences between the amoebae in HEX transport or metabolism.
In this study, we have found that inhibition of central carbon metabolism can impact amoebae viability, both in vitro and in an animal model. The favorable pharmacological profiles of HEX and related phosphonates suggests that they may be useful in the treatment of N. fowleri brain infections [40], and their promising activity in a rodent model supports this possibility. Ultimately, the ENO inhibitors may prove useful for future therapeutics, either alone or in concert with other agents (like lipidolysis inhibitors) that would in tandem block multiple key pathways required for both ATP and reducing equivalent biosynthesis with catastrophic consequences.
Supporting information
S1 Text. Methods for NfENO protein purification, crystallization, and structure determination.
https://doi.org/10.1371/journal.ppat.1012412.s001
(DOCX)
S1 Fig. Synthetic route for deoxy-SF2312, HEX, and HEPTA.
The compounds (1-hydroxy-2-oxopyrrolidin-3-yl) phosphonic acid (deoxy-SF2312) 1, (1-hydroxy-2-oxopiperidin-3-yl) phosphonic acid (HEX) 2, and (1-hydroxy-2-oxoazepan-3-yl) phosphonic acid (HEPTA) 3, were synthesized following published procedures reported in the literature (5, 8). Briefly, a mixture of O-benzylhydroxylamine hydrochloride and chloroformate in pyridine was stirred at room temperature under N2 for 4 h to yield ethyl benzyloxycarbamate (4). Triethyl phosphite in excess of the appropriate dibromoalkane (i.e. n = 1 for Deoxy-SF2312, n = 2 for HEX, and n = 3 for HEPTA) was stirred at 90°C overnight to yield the appropriate alkylbromophoshonate (5) as light yellow oil after purification by silica gel column chromatography (0–8% MeOH in DCM). Potassium carbonate was added to a solution of 4 and 5 in MeCN and the mixture was stirred at 90°C overnight to yield the linear diethyl phosphonate-carbamate (6) as a yellow oil after purification by silica gel column (0–8% MeOH in DCM). Treatment of 6 in THF with LiHMDS at 0°C for 3 h under N2 afforded the cyclized diethyl phosphonate (7) as a yellow oil after purification by silica gel column chromatography (0–8% MeOH in DCM). Hydrolysis of 7 with iodotrimethylsilane in DCM at 0–25°C for 4 h under N2 yielded the benzyloxy phosphonic acid (8) as a yellow solid after reverse phase column chromatography (C18 silica gel; 0–25% MeOH in H2O). Finally, the benzyl group was removed by hydrogenolysis of (1.8 mmol, 8) dissolved in 10 mL of H2O/MeOH (1:1 v/v) containing 30 mmol of acetic acid and palladium on carbon (10%, 50 mg) stirred at 65°C for 18–36 h at 1 atm H2 to give the desired products as an off-white oil (1) or light-yellow oil (2 and 3).
https://doi.org/10.1371/journal.ppat.1012412.s002
(DOCX)
S2 Fig. Spectroscopic characterization of deoxy-SF2312, HEX and HEPTA.
1H and 13C NMR spectra were obtained using Bruker Advance 500 MHz spectrometers. Chemical shifts are reported in parts per million (ppm). Spectra are referenced to residual solvent peaks. Flash column chromatography was carried out using ZEOCHEM silica gel (40–63 μm). Analytical and preparative thin-layer chromatography (TLC) were performed on Sorbtech silica G TLC plates. All non-aqueous reactions were performed under an inert atmosphere of nitrogen in flame-dried glassware containing a stir bar unless otherwise noted. Acetonitrile (ACN), tetrahydrofuran (THF), dichloromethane (DCM), methanol (MeOH), and pyridine were obtained from commercial sources and dried following standard distillation procedures. All other solvents were obtained from commercial sources and used without drying unless otherwise noted. All water and aqueous solutions were made using deionized (DI) water.
https://doi.org/10.1371/journal.ppat.1012412.s003
(DOCX)
S3 Fig. ClustalOmega alignment of NfENO with human ENOs colored by percent equivalent score.
[https://doi.org/10.1093/nar/gku316]. The top lines show the secondary structure elements of NfENO. The scale above the alignment corresponds to the NfENO sequence. Active site residues are highlighted with a blue triangle; the black star marks the position of NfENO Lys243.
https://doi.org/10.1371/journal.ppat.1012412.s004
(DOCX)
S4 Fig. Comparison of the active sites of HEX-bound ENO2 and HEX-docked NfENO.
(A) Superposition of the active sites of HEX-bound ENO2 and HEX-docked NfENO. (B) NfENO lacks Arg-372 and likely engages with HEX through Lys-243, which differs from (C) ENO2.
https://doi.org/10.1371/journal.ppat.1012412.s005
(DOCX)
S5 Fig. Amoeba resistance to HEX does not explain the therapeutic failure observed in the one HEX-treated rodent that succumb to infection.
Trophozoites were cultured from brains of a HEX-treated (HEX) or PBS-treated (Vehicle) rodent for two weeks in media followed by testing against HEX in a standard viability assay. Both cultures responded similarly, with EC50 values of 0.09 ± 0.04 and 0.1 ± 0.08 μM for HEX- or Vehicle-treated rodents, respectively. Drug concentrations were tested in triplicate, with some error bars smaller than the graphing symbols.
https://doi.org/10.1371/journal.ppat.1012412.s006
(DOCX)
S1 Table. NfENO sequences from Uniprot, AmoebaDB, and the truncation used for crystallography that is deposited in the PDB.
https://doi.org/10.1371/journal.ppat.1012412.s007
(DOCX)
S2 Table. X-ray Data collection and processing.
Values for the outer resolution shell are given in parentheses.
https://doi.org/10.1371/journal.ppat.1012412.s008
(DOCX)
S3 Table. Crystallographic structure solution and refinement quality statistics.
Values for the outer shell are given in parentheses.
https://doi.org/10.1371/journal.ppat.1012412.s009
(DOCX)
S4 Table. Unbiased metabolomics results for amoebae grown in the presence of HEX.
Fold-change (FC) is the mean ratio of peak area of a metabolite detected in the first group divided by the corresponding value in the second group. One-way ANOVA with the Tukey method for post-test was used for metabolite abundance significance test, and the Benjamini and Hochberg method was used for multiple test correction. ND stands for not detected. The metabolite was either not present in the sample or its concentration in sample is smaller than the sensitivity of the analytical platform.
https://doi.org/10.1371/journal.ppat.1012412.s010
(DOCX)
S5 Table. Fold-change in abundance1 of metabolites in response to carbon source and HEX treatment, relative to growth in glucose alone.
1Relative to average intensities of metabolites detected in cells grown under standard (+glc) conditions.
https://doi.org/10.1371/journal.ppat.1012412.s011
(DOCX)
S6 Table. Fold-change in abundance of metabolites of cells grown in glycerol after HEX treatment, 1relative to growth in glycerol alone.
1Relative to average intensities of metabolites detected in cells grown in -glc/+gly conditions.
https://doi.org/10.1371/journal.ppat.1012412.s012
(DOCX)
Acknowledgments
The authors would like to thank Drs. Dennis Kyle and Christopher Rice for their support and Drs. Heather Walters and Lesly Temesvari for their assistance in testing HEX against Entamoeba. We thank Dr. Paul Leonard (MD Anderson Cancer Center Core for Biomolecular Structure and Function) for providing purified recombinant human enolases. The authors would also like to thank Colm Roster and Caroline Palmentiero for their technical support in these projects. We acknowledge all the people at Seattle Structural Genomics of Infectious Diseases (SSGCID) that contributed, including Justin K. Craig (performed cloning), Dr. Wesley C. Van Voorhis (supervised protein cloning and expression group), Sandhya Subramanian (Bioinformatics), Robin Stacey (Project Director), and Dr. Peter Myler (SSGCID PI).
The abbreviations used are: 2-PG, 2-phosphoglycerate; 2-/3-PG, a mixture of 2-PG and 3-phosphoglycerate; DHAP, dihydroxyacetone phosphate; ENO2, human enolase 2; F6P, fructose 6-phosphate; G6P, glucose 6-phosphate; G6PDH, glucose 6-phosphate dehydrogenase; GBM, glioblastoma multiforme; Gly3P, glycerol 3-phosphate; HEPTA, (1-hydroxy-2-oxoazepan-3-yl) phosphonic acid; HEX, (1-hydroxy-2-oxopiperidin-3-yl) phosphonic acid; NfENO, N. fowleri enolase; SF2312, (1,5-dihydroxy-2-oxopyrrolidin-3-yl) phosphonic acid
References
- 1. Yoder JS, Eddy BA, Visvesvara GS, Capewell L, Beach MJ. The epidemiology of primary amoebic meningoencephalitis in the USA, 1962–2008. Epidemiol Infect. 2010;138(7):968–75. pmid:19845995
- 2. Milanes JE, Suryadi J, Abendroth J, Van Voorhis WC, Barrett KF, Dranow DM, et al. Enzymatic and structural characterization of the Naegleria fowleri glucokinase. Antimicrob Agents Chemother. 2019;63(5).
- 3. Herman EK, Greninger A, van der Giezen M, Ginger ML, Ramirez-Macias I, Miller HC, et al. Genomics and transcriptomics yields a system-level view of the biology of the pathogen Naegleria fowleri. BMC Biol. 2021;19(1):142.
- 4. Bexkens ML, Zimorski V, Sarink MJ, Wienk H, Brouwers JF, De Jonckheere JF, et al. Lipids are the preferred substrate of the protist Naegleria gruberi, relative of a human brain pathogen. Cell Rep. 2018;25(3):537–43 e3.
- 5. Lin YH, Satani N, Hammoudi N, Yan VC, Barekatain Y, Khadka S, et al. An enolase inhibitor for the targeted treatment of ENO1-deleted cancers. Nat Metab. 2020;2(12):1413–26. pmid:33230295
- 6. Leonard PG, Satani N, Maxwell D, Lin YH, Hammoudi N, Peng Z, et al. SF2312 is a natural phosphonate inhibitor of enolase. Nat Chem Biol. 2016;12(12):1053–8. pmid:27723749
- 7. Myler PJ, Stacy R, Stewart L, Staker BL, Van Voorhis WC, Varani G, Buchko GW. The Seattle Structural Genomics Center for Infectious Disease (SSGCID). Infect Disord Drug Targets. 2009;9(5):493–506. pmid:19594426
- 8. Stacy R, Begley DW, Phan I, Staker BL, Van Voorhis WC, Varani G, et al. Structural genomics of infectious disease drug targets: the SSGCID. Acta Crystallogr Sect F Struct Biol Cryst Commun. 2011;67(Pt 9):979–84. pmid:21904037
- 9. Bryan CM, Bhandari J, Napuli AJ, Leibly DJ, Choi R, Kelley A, et al. High-throughput protein production and purification at the Seattle Structural Genomics Center for Infectious Disease. Acta Crystallogr Sect F Struct Biol Cryst Commun. 2011;67(Pt 9):1010–4. pmid:21904042
- 10. Choi R, Kelley A, Leibly D, Hewitt SN, Napuli A, Van Voorhis W. Immobilized metal-affinity chromatography protein-recovery screening is predictive of crystallographic structure success. Acta Crystallogr Sect F Struct Biol Cryst Commun. 2011;67(Pt 9):998–1005. pmid:21904040
- 11. Serbzhinskiy DA, Clifton MC, Sankaran B, Staker BL, Edwards TE, Myler PJ. Structure of an ADP-ribosylation factor, ARF1, from Entamoeba histolytica bound to Mg(2+)-GDP. Acta Crystallogr F Struct Biol Commun. 2015;71(Pt 5):594–9.
- 12. Studier FW. Protein production by auto-induction in high density shaking cultures. Protein Expr Purif. 2005;41(1):207–34. pmid:15915565
- 13. Xds Kabsch W. Acta Crystallogr D Biol Crystallogr. 2010;66(Pt 2):125–32.
- 14. Newman JA, Hewitt L, Rodrigues C, Solovyova AS, Harwood CR, Lewis RJ. Dissection of the network of interactions that links RNA processing with glycolysis in the Bacillus subtilis degradosome. J Mol Biol. 2012;416(1):121–36.
- 15. Adams PD, Afonine PV, Bunkoczi G, Chen VB, Davis IW, Echols N, et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr. 2010;66(Pt 2):213–21. pmid:20124702
- 16. Emsley P, Lohkamp B, Scott WG, Cowtan K. Features and development of Coot. Acta Crystallogr D Biol Crystallogr. 2010;66(Pt 4):486–501. pmid:20383002
- 17. Chen VB, Arendall WB 3rd, Headd JJ, Keedy DA, Immormino RM, Kapral GJ, et al. MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr D Biol Crystallogr. 2010;66(Pt 1):12–21. pmid:20057044
- 18. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, et al. The Protein Data Bank. Nucleic Acids Res. 2000;28(1):235–42. pmid:10592235
- 19. Pancholi V, Fischetti VA. alpha-enolase, a novel strong plasmin(ogen) binding protein on the surface of pathogenic streptococci. J Biol Chem. 1998;273(23):14503–15. pmid:9603964
- 20. Lee C, Yang W, Parr RG. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys Rev B Condens Matter. 1988;37(2):785–9. pmid:9944570
- 21. Pisaneschi F, Lin YH, Leonard PG, Satani N, Yan VC, Hammoudi N, et al. The 3S enantiomer drives enolase inhibitory activity in SF2312 and its analogues. Molecules. 2019;24(13). pmid:31324042
- 22. Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem. 2010;31(2):455–61. pmid:19499576
- 23. Prodhan MAI, Shi B, Song M, He L, Yuan F, Yin X, et al. Integrating comprehensive two-dimensional gas chromatography mass spectrometry and parallel two-dimensional liquid chromatography mass spectrometry for untargeted metabolomics. Analyst. 2019;144(14):4331–41. pmid:31192319
- 24. Wei X, Shi X, Kim S, Patrick JS, Binkley J, Kong M, et al. Data dependent peak model based spectrum deconvolution for analysis of high resolution LC-MS data. Anal Chem. 2014;86(4):2156–65. pmid:24533635
- 25. Wei X, Sun W, Shi X, Koo I, Wang B, Zhang J, et al. MetSign: a computational platform for high-resolution mass spectrometry-based metabolomics. Anal Chem. 2011;83(20):7668–75. pmid:21932828
- 26. Pietkiewicz J, Gamian A, Staniszewska M, Danielewicz R. Inhibition of human muscle-specific enolase by methylglyoxal and irreversible formation of advanced glycation end products. J Enzyme Inhib Med Chem. 2009;24(2):356–64. pmid:18830874
- 27. Lee JH, Kang HK, Moon YH, Cho DL, Kim D, Choe JY, et al. Cloning, expression and characterization of an extracellular enolase from Leuconostoc mesenteroides. FEMS Microbiol Lett. 2006;259(2):240–8.
- 28. Yan VC, Pham CD, Ballato ES, Yang KL, Arthur K, Khadka S, et al. Prodrugs of a 1-hydroxy-2-oxopiperidin-3-yl phosphonate enolase inhibitor for the treatment of ENO1-deleted cancers. J Med Chem. 2022;65(20):13813–32. pmid:36251833
- 29. Westheimer FH. Why nature chose phosphates. Science. 1987;235(4793):1173–8. pmid:2434996
- 30. Martin JF, Liras P. Molecular Mechanisms of phosphate sensing, transport and signalling in Streptomyces and related Actinobacteria. Int J Mol Sci. 2021;22(3). pmid:33498785
- 31. Whitney LP, Lomas MW. Phosphonate utilization by eukaryotic phytoplankton. Limnol Oceanogr Lett. 2018;4(1):18–24.
- 32. Sanders S, Bartee D, Harrison MJ, Phillips PD, Koppisch AT, Freel Meyers CL. Growth medium-dependent antimicrobial activity of early stage MEP pathway inhibitors. PLoS One. 2018;13(5):e0197638. pmid:29771999
- 33. Else PL. The highly unnatural fatty acid profile of cells in culture. Prog Lipid Res. 2020;77:101017. pmid:31809755
- 34. Opperdoes FR, De Jonckheere JF, Tielens AG. Naegleria gruberi metabolism. Int J Parasitol. 2011;41(9):915–24.
- 35. Mendes V, Acebron-Garcia-de-Eulate M, Verma N, Blaszczyk M, Dias MVB, Blundell TL. Mycobacterial OtsA structures unveil substrate preference mechanism and allosteric regulation by 2-oxoglutarate and 2-phosphoglycerate. mBio. 2019;10(6). pmid:31772052
- 36. Glauser T, Shinnar S, Gloss D, Alldredge B, Arya R, Bainbridge J, et al. Evidence-Based Guideline: Treatment of convulsive status epilepticus in children and adults: Report of the Guideline Committee of the American Epilepsy Society. Epilepsy Curr. 2016;16(1):48–61. pmid:26900382
- 37. Kapoor M, Cloyd JC, Siegel RA. A review of intranasal formulations for the treatment of seizure emergencies. J Control Release. 2016;237:147–59. pmid:27397490
- 38. Erdo F, Bors LA, Farkas D, Bajza A, Gizurarson S. Evaluation of intranasal delivery route of drug administration for brain targeting. Brain Res Bull. 2018;143:155–70. pmid:30449731
- 39. Shah P, Lalan M, Barve K. Intranasal delivery: An attractive route for the administration of nucleic acid based therapeutics for CNS disorders. Front Pharmacol. 2022;13:974666. pmid:36110526
- 40. Barekatain Y, Khadka S, Harris K, Delacerda J, Yan VC, Chen KC, et al. Quantification of phosphonate drugs by (1)H-(31)P HSQC shows that rats are better models of primate drug exposure than mice. Anal Chem. 2022;94(28):10045–53. pmid:35792073