Miconazole-like Scaffold is a Promising Lead for Naegleria fowleri-Specific CYP51 Inhibitors

Developing drugs for brain infection by Naegleria fowleri is an unmet medical need. We used a combination of cheminformatics, target-, and phenotypic-based drug discovery methods to identify inhibitors that target an essential N. fowleri enzyme, sterol 14-demethylase (NfCYP51). A total of 124 compounds preselected in silico were tested against N. fowleri. Nine primary hits with EC50 ≤ 10 μM were phenotypically identified. Cocrystallization with NfCYP51 focused attention on one primary hit, miconazole-like compound 2a. The S-enantiomer of 2a produced a 1.74 Å cocrystal structure. A set of analogues was then synthesized and evaluated to confirm the superiority of the S-configuration over the R-configuration and the advantage of an ether linkage over an ester linkage. The two compounds, S-8b and S-9b, had an improved EC50 and KD compared to 2a. Importantly, both were readily taken up into the brain. The brain-to-plasma distribution coefficient of S-9b was 1.02 ± 0.12, suggesting further evaluation as a lead for primary amoebic meningoencephalitis.


■ INTRODUCTION
The free-living amoeboflagellate, Naegleria fowleri, is commonly found in bodies of natural water (lakes and rivers) and in swimming pools with inadequate levels of chlorine.In the environment, N. fowleri occurs in three forms�a cyst (dormant form), a trophozoite (ameboid), and a flagellate.N. fowleri trophozoites feed mostly on bacteria but can also act as opportunistic pathogens, causing infection of the central nervous system (CNS) of animals including humans.Primary amoebic meningoencephalitis (PAM) due to N. fowleri is a fulminating brain infection that can result in death within days.PAM has a worldwide distribution, although it occurs most frequently in warmer regions and during hot summer months.−4 In the US, N. fowleri infection is considered "rare" with zero to eight cases reported annually but is likely under-reported. 5Most infected individuals die due to the rapid onset and destructive nature of the disease as well as to the lack of effective treatments. 6PAM cases often go unnoticed in countries with warm climates, poor health infrastructure, and ritual ablution practices that are common in certain religious groups. 7Based on the free-living amoeba registry maintained by the Centers for Disease Control and Prevention (CDC), the fatality rate of PAM is over 97%. 5 Currently, there is no standard regimen for the treatment of Naegleria infection in humans.−12 The CDC-recommended treatment for patients suspected of PAM includes combination therapy, administered intravenously, intrathecally, or orally, consisting of antimycotic drugs AmpB and fluconazole, antibiotics azithromycin and rifampin, an investigational agent miltefosine, and an anti-inflammatory drug dexamethasone.All the documented survivors of PAM received AmpB, but clinical use of AmpB is limited due to its toxicity, including acute infusionrelated reactions and dose-related nephrotoxicity. 4Combination of AmpB with the antileishmaniasis agent, miltefosine, has shown promise, but not all patients who received miltefosine as part of their treatment regimens survived. 13Therefore, development of efficacious and safe drugs for PAM treatment remains an unmet medical need.
We 14−18 and others 19−23 have explored the steroidogenic pathway in free-living amoebae and have pharmacologically validated several steroidogenic enzymes as drug targets.Inhibition of sterol 14-demethylase (CYP51) with a variety of FDA-approved CYP51 inhibitors (conazoles) induced massive autophagocytosis in cultured N. fowleri, leading to cell death after 24 h of drug exposure. 15The amoebicidal effect of CYP51 inhibitors is due to inhibition of 14-demethylation of the endogenous CYP51 substrate in N. fowleri, 31-norlanosterol.14-Demethylation of 31-norlanosterol is a prerequisite for subsequent 4α-demethylation and 24-methylation steps, leading to biosynthesis of essential ergosterol and ergosterollike sterols. 14Depletion of ergosterol, concomitant with the accumulation of intermediates and end-products incompatible with normal permeability and fluidity of the Naegleria cell membrane, leads to the altered morphology and death of N. fowleri cells. 14,15onazoles have been used in combination with AmpB for the treatment of PAM patients.The first case of PAM survival in the United States involved a 9 year-old girl who was treated with both AmpB and miconazole in 1978 (1, Chart 1).These drugs were administered intravenously and intrathecally, in addition to oral rifampin, intravenous dexamethasone, and oral phenytoin. 24In more recent cases, miconazole was replaced with fluconazole, which is preferred when systemic treatment is required.This is because of the improved safety and predictable absorption of fluconazole when administered orally. 25In contrast to fluconazole, miconazole is approved for topical administration only in humans.It is used in the treatment of fungal or yeast infections of the skin or vagina and for the treatment of oropharyngeal candidiasis in patients 16 years and older.Nonetheless, there is a growing interest in miconazole due to its antiinflammatory 26 and neuroprotective effects, which include remyelination of neural progenitor cells in models of multiple sclerosis, 27,28 protection of brain blood vessels from rupture in a hemorrhagic stroke model, 29 and ameliorating memory deficits in mice. 30e have previously reported that the anti-Naegleria activity of conazole analogues increases with an increase in molecular weight and hydrophobicity, while the CNS permeability decreases in the same order. 15Fluconazole, while having the lowest activity among the conazoles against N. fowleri European KUL strain (∼14 μM), is known to rapidly distribute through body tissues, including CNS compartments where it achieves concentrations greater than the MIC 90 of common fungal pathogens. 31,32Posaconazole and itraconazole (EC 50 of ≤10 nM) are superior in their amoebicidal effect to fluconazole and AmpB; however, they have poor blood−brain barrier (BBB) permeability.Slow accumulation in brain tissue upon repeated dosing 33−36 may explain the lack of complete anti-Naegleria efficacy of posaconazole in a mouse model of PAM (two of six mice cured at a 20 mg/kg dose). 27iconazole falls in between of these two extremes in terms of both anti-Naegleria potency (EC 50 of <2 μM) 15 and brain permeability. 37In this present work, the complementarity of the miconazole molecular scaffold to the N. fowleri CYP51 (NfCYP51) active site was demonstrated experimentally via a combination of cheminformatics, biochemistry, X-ray crystallography, and phenotypic cell-based methods.The analogues synthesized based on the miconazole template retained potency against the molecular target and were brain permeable in mice.345.6 ± 43.1 and cLogP of 3.79 ± 0.50 were tested against N. fowleri trophozoites.Nine hits were identified with EC 50 ≤ 10 μM (2a−2i, see Table S1 in the Supporting Information).The top hit (2a, Chart 1) had a miconazole-like scaffold (1, Chart 1) and was singled out following cocrystallization with the NfCYP51 target.New analogues of this hit were then synthesized and characterized to assess (i) the stereo configuration at the chiral carbon center, (ii) the role of the ester linker, (iii) the optimal configuration of substituents at the phenyl and benzoyl moieties, and (iv) BBB permeability potential.By maintaining the azole ring, different substituents were introduced in the o-and/or p-position of the phenyl ring and in the o-, m-, and/or p-position of the benzoyl moiety.We introduced the following: (i) a trifluoromethyl group or two chlorine atoms on the benzoyl residue; (ii) a fluorine atom, a methoxy group, or two chlorine atoms on the aromatic ring of the phenyl moiety (Chart 1).To assess the role of the ester linker, we synthesized the ether counterparts by introducing (i) a trifluoromethyl group or two chlorine atoms on the benzoyl residue and (ii) a fluorine or two chlorine atoms on the aromatic ring of the phenyl moiety (Chart 1).The triazole analogue of hit compound 2a was also synthesized.Notably, among the newly synthesized compounds, derivatives 6b and 10b were designed as the ester and 3,5-dichlorine-substituted miconazole analogues, respectively.
Focused 124-Compound Library.The large, hydrophobic binding site of CYP51 favors molecules with physicochemical properties incompatible with BBB permeability. 38,39Thus, phenotypic and target-based screening alone may not yield quality hits for PAM.In the present work, we used the SwissADME cheminformatic tool 40 to preselect compounds based on a combination of physicochemical parameters (lipophilicity, polar surface area, molecular weight, flexibility, hydrogen bond donor count, and most basic pK a ), reflecting the compound's CNS-multiparameter optimization (MPO) score. 41,42Starting from a library of 7000 compounds originally synthesized to target fungal CYP51, 39,43−46 we identified a panel of 314 molecules endowed with antifungal and antiprotozoal activities.Among them, 124 compounds with an average molecular weight of 345.55 ± 43.13, cLogP of 3.79 ± 0.50, and topological polar surface area (TPSA) of 42.64 ± 15.85 were selected for the phenotypic screening against N. fowleri.All compounds contained an aromatic heterocycle (imidazole or 1,2,4-triazole) capable of coordinating the Fe center of the heme macrocycle in the NfCYP51 active site.
Phenotypic Organism-Based Screen.The 124 selected compounds, all having a CNS-multiparameter optimization (MPO) score >4, 47 were screened against the axenically cultured N. fowleri European KUL strain at 10 μM.Nine hits with ≥50% inhibition were identified (2a−2i, see Table S1 in the Supporting Information).An inhibition of 50% at 10 μM is comparable to the fluconazole reference potency (EC 50 ∼ 14 μM) against the same strain. 15it Cocrystallization with NfCYP51.All nine hits were cocrystallized with recombinant NfCYP51.Only one compound, 2a (used as a racemic mixture) (Chart 1), produced cocrystals with the target.The cocrystal structure of the NfCYP51−2a complex was determined to a resolution of 1.75 Å, which is comparable to the resolution of the NfCYP51− posaconazole complex (1.71 Å, PDB ID 5TL8) determined previously. 15For comparison, the NfCYP51−fluconazole complex diffracted only to a resolution of 2.7 Å (PDB ID 6AY4).This is consistent with the higher affinity of 2a to NfCYP51, suggesting better target engagement.The welldefined electron density in the binding pocket adjacent to heme corresponds to the S-configuration at the chiral carbon center of 2a (Figure 1A,B).Superimposition with fluconazole indicates that both molecules coordinate to the heme iron with the azole moiety, while other aromatic moieties of each compound are engaged in different sets of drug−target interactions (Figure 1C).
2a Binding Pocket.The cocrystal structure of 2a revealed three main groups of drug−target interactions (Figure 2).(a) Imidazole moiety.The aromatic nitrogen atom of the imidazole moiety provided a coordination bond to the heme iron (Figure 2A).There is a small cavity adjacent to C-5 of the imidazolyl moiety that may accommodate a small substituent.(b) The 3-trifluoromethyl-benzoyl moiety is somewhat coplanar to the heme plane (Figure 2B).Void spaces adjacent to the −CF 3 group and C-5 suggest that halogens (i.e., chlorine and fluorine) or a variety of small alkyl groups with a gradual increase in the steric hindrance (such as methyl, ethyl, isopropyl, or sec-butyl groups) may improve drug−target fit and modulate physicochemical properties.(c) Finally, the 4fluorophenyl moiety (Figure 2A) projects toward the long hydrophobic tunnel extending from the heme to the protein surface known to accommodate large hydrophobic moieties in higher molecular weight azole inhibitors (posaconazole and itraconazole). 15The C2−C3 edge of the 4-fluorophenyl moiety faces heme propionate (6.3 Å) and the OH-groups of Tyr107 (5.3 Å) and Tyr120.A void hydrophobic space at C- 6 may accommodate another halogen substituent.For CNS compatibility, we propose staying within the binding envelope of 2a and avoiding extending molecules into the hydrophobic tunnel.
Based on the NfCYP51−2a structure, we synthesized seven new analogues 3a−5a, 3b−5b, and 7 as pure enantiomers and evaluated them for binding affinity and anti-N.fowleri potency.The X-ray structures for the 4b and 5b drug−target complexes were determined at 1.81 and 2.10 Å, respectively (Table 1).Together with 2a, these two structures demonstrated similar binding modes and built a foundation for a further hit-to-lead optimization strategy of the miconazole scaffold (Figure 3).Compounds 6b (ester analog of miconazole) and 8b−10b were synthesized to assess the contribution of the ester linker.
Chemistry.The synthesis of hit compounds 2a−2i (see Table S1 in the Supporting Information) was carried out as previously reported. 44,45,48,49The R-enantiomers 3a−5a were synthesized in accordance with the procedures outlined in Scheme 1.A mixture of the appropriate phenacyl bromide with 1H-imidazole in dimethylformamide (DMF) afforded the desired ketones 14−16, which were reduced via catalytic asymmetric hydrogenation in the presence of a chiral diamine ligand complexed with ruthenium ((S,S)-TsDPEN Ru-(pcymene)Cl) in dichloromethane (DCM), Et 3 N, and HCOOH, as described in the literature. 50,51The obtained R-configured alcohols 11a−13a were treated with sodium hydride in anhydrous acetonitrile, and subsequently, the proper acyl chloride was added to give compounds 3a−5a, as previously described. 43he synthesis of S-enantiomers 3b−6b and 7 was performed, as reported in Scheme 2. The synthetic approach resembles that described above for compounds 2a−4a.Noteworthily, the synthetic pathway of compound 7 starts with the reaction of 2-bromo-1-(4-fluorophenyl)ethan-1-one (commercially available) with 1H-1,2,4-triazole in the presence of Et 3 N in acetone, yielding a ketone derivative, 18.This compound underwent catalytic asymmetric hydrogenation 50 in a similar fashion to that described in Scheme 1, giving the Sconfigured alcohol 17 that was acylated 43 with 3-(trifluoromethyl)benzoyl chloride to furnish derivative 7.
Evaluation of Binding and Biological Activities of the Newly Synthesized Compounds.We assessed the anti-Naegleria potency, binding affinity to the NfCYP51 target, and cocrystallization propensity of the newly synthesized analogues 3a−5a, 3b−6b, 7, and 8b−10b.Compounds 4b and 5b were the only two new analogues that produced cocrystals with NfCYP51 (Figure 3).Two chlorine atoms introduced on the aromatic ring of the benzoyl moiety (4b) or phenyl moiety (5b) notably improved the K D and EC 50 , compared to the parental hit, 2a; however, the potency of both analogues was inferior to miconazole (Table 2).The methoxy-substituent was inferior to fluorine in the 4-fluorophenyl moiety (3a and 3b).Overall, the EC 50 correlated with K D and cocrystallization propensity of the analogues and the S-configuration was superior to the R-configuration for all tested enantiomer pairs.The hydrolytically unstable ester moiety featured in 2a and the first seven analogues distinguished them from miconazole.
The hydrolytic instability of the ester moiety is consistent with the high clearance and short plasma half-life of 2a observed in a single-dose PK experiment.2a administered intraperitoneally at 20 mg/kg could be detected at low levels, in both plasma (4 ng/mL) and brain (0.1 ng/mg), only 30 min post injection, while the hydrolysis product, 3-trifluoro benzoic acid, was one of the main circulating metabolites for up to 24 h.The X-ray structures for the 2a, 4b, and 5b drug−target complexes showed that the ester moiety is not directly involved in the interactions with NfCYP51.Thus, another set of compounds, 8b−10b, was synthesized, where the ester moiety was replaced with a hydrolytically more stable ether.Specifically, 8b and 9b were the ether counterparts of 4b and 5b, respectively, while 10b was the 3,5-chlorine-substituted miconazole analogue.Indeed, in miconazole, two chlorine substituents are in the para/ortho configuration, while the meta/meta configuration was supported by the X-ray cocrystal structure of 2a that predicted the potential interference of V132 with a para-substituent.Finally, miconazole analogues 6b carrying the ester moiety were also synthesized.
Most of the analogues from this latest set showed linear binding to NfCYP51 with an increase in compound concentration from 0.025 to 0.5 μM, until the target saturation is reached (Figure 4).Linear dependence does not allow for accurate calculation of K D .Reducing the target concentration below 0.5 μM is prohibited by the sensitivity of this UV−vis spectroscopy measurements using available instrumentation.We can only estimate K D to be ≤5 nM based on the transformation of the parabolic isotherms into linear saturation curves that gradually occurs at target concentrations exceeding K D by ∼100-fold.Compared to their ester counterparts, analogues 8b−10b had an improved EC 50 , higher binding affinity (Table 2), and better target saturation (Figure 4).Against N. fowleri, 9b was equipotent to miconazole, while 8b and 10b were slightly inferior.The higher binding affinity/ target saturation and higher biological activity of the ether congeners are likely due to a better drug−target engagement achieved due to the higher flexibility of the ether linker.Also, an obvious advantage of the ether moiety over the ester moiety is its high plasma stability and the reported miconazole elimination half-life is 24 h. 52he binding curves plateau at ∼0.3 μM (instead of 0.5 μM).This observation indicates that not all binding sites participate in inhibitor binding.A plausible explanation of this phenomenon is irreversible deterioration of the heme Fe thiolate bond upon binding of azole inhibitors, as we demonstrated elsewhere by UV−vis spectroscopy for CYP51 of N. fowleri and Acanthamoeba castellanii. 53Depending on experimental conditions, deterioration rates are on a minuteto-hours time scale that is consistent with protein conformational motions.The broad Soret band indicates an enormous structural heterogeneity and flexibility of the heme pocket; close inspection of the spectra suggests that multiple species are present. 53ood−Brain Barrier Permeability in Mice.To experimentally evaluate brain penetration of the best acting compounds 8b and 9b, BBB permeability was determined in vivo.Compounds 8b and 9b were administered at 40 mg/kg i.p. dose, using miconazole (1) as a reference.Miconazole was used as a reference throughout this present work because

Journal of Medicinal Chemistry
among existing azole drugs, miconazole molecular scaffolds were singled out as the most complementary to the NfCYP51 active site.After 1 h of drug exposure, blood and brain samples were collected and analyzed by means of HPLC-ESI-MS/MS.As shown in Table 3, both 8b and 9b confirmed their capability of crossing the BBB in mice.8b was equally permeable to miconazole, with an average brain distribution coefficient (brain/plasma ratio) of 0.43 ± 0.08, while 9b distributed to the brain with a B/P coefficient of 1.02 ± 0.12.

■ CONCLUSIONS
Using a combination of cheminformatics, target-based, and phenotypic drug discovery methods, we have identified a lead scaffold suitable for further development of drug candidates to treat brain infection with N. fowleri.Only compounds with a CNS-multiparameter optimization (MPO) score >4 were selected for primary phenotypic screening against N. fowleri followed by cocrystallization with the molecular target, NfCYP51.We identified a promising hit, 2a, which resembled miconazole, the first member of the "conazole" pedigree that in the 1970s was used in combination with AmpB and rifampin to treat PAM.It was later replaced with fluconazole because of improved safety and better absorption.Fluconazole remains the only CYP51 inhibitor in a combination of drugs recommended by the CDC for the treatment of PAM.In laboratory assays, fluconazole kills N. fowleri at concentrations much higher than other azole drugs.We reason that the success of an inhibitor in treating N. fowleri infection in humans would sum from complementarity to a molecular target combined with an adequate BBB permeability, allowing a drug to enter the central nervous system from systemic circulation.
Compound 2a showed better potency and drug−target complementarity than fluconazole; the follow up ester analogues 4b and 5b retained the binding mode and improved the binding and potency.The hydrolytically more stable ether analogues, 8b and 9b, had further improved binding and drug potency; 9b with a 3-trifluoromethyl substituent was equipotent to miconazole.The ester analogue of miconazole, 6b, and the meta/meta chlorine-configured 10b were comparable to miconazole in binding affinity and slightly inferior to miconazole in anti-Naegleria potency.
The in vivo assessment of the BBB permeability of active compounds, 8b and 9b, confirmed that both were detected in brain tissue.Plasma and brain concentrations of 8b were similar to those of miconazole; 9b demonstrated better brain permeability than miconazole and 8b.The B/P distribution coefficient of ∼1 suggests free distribution of 9b to the brain.
This study demonstrated for the first time complementarity between miconazole congeners and the NfCYP51 target and predicted medicinal chemistry modifications within the binding envelope that would retain potency yet gain BBB permeability.These results now warrant further development of miconazole analagos into anti-Naegleria drug candidates.

■ EXPERIMENTAL SECTION
Chemistry: General.Solvents and reagents were of analytical grade and, when necessary, purified and dried by standard methods.Merck silica-gel 60 F 254 plates were used for analytical TLC.Column chromatography was performed on silica gel (Merck, 70−230 mesh).Melting points were determined with a Buchi 530 capillary apparatus and are uncorrected.Infrared (IR) spectra were recorded on a PerkinElmer Spectrum-one spectrophotometer.NMR spectra were recorded on a Bruker AC 400 spectrometer at 400 MHz for 1 H and 100 MHz for 13 C; the following abbreviations were used: s for singlet, bs for broad singlet, d for doublet, t for triplet, dd for double doublet, m for multiplet; chemical shifts are given in d with respect to the residual solvent signal; coupling constants are given in Hz.DMSO-d 6, CD 3 CN, acetone-d 6 , and CD 3 OD of 99.9% isotopic purity (Aldrich) were used.Mass spectra were recorded on a ThermoFinnigan LCQ Classic LC/MS/MS ion trap equipped with an ESI source and a syringe pump; samples (10 −4 −10 −5 M in methanol (MeOH)/H 2 O 80:20) were infused in an electrospray system at a flow rate of 5−10 μL min −1 ; when necessary, 50 μL of 10 −2 M HCOOH was added to the sample solutions to promote the analyte ionization; the ESI-MS data are given as m/z, with mass expressed in amu.
A total of 5−10 mg of each compound was synthesized for phenotypic organism-based screening and target-based evaluation assays.All compounds are >95% pure, as determined by combustion analysis.Analytical results agreed to within ±0.40% of the theoretical values.

General Procedure B (GP-B) to Obtain R-Configured Alcohols (11a−13a).
A nitrogen atmosphere was established in a three-neck flask containing the appropriate imidazolyl-ethanone (1 mmol) and [(S,S)-TsDPEN]Ru-(p-cymene)Cl (0.001 mmol) before the addition of dichloromethane (7 mL) and Et 3 N (5 mmol).Formic acid (5 mmol) was added over a period of an hour.The mixture was stirred at room temperature for 26 h.NaHCO 3 saturated solution was added cautiously and the organic layer with water and brine and dried over anhydrous Na 2 SO 4. 47 The solvent was removed under vacuum, affording the desired product.For each compound, imidazolylethanone, melting point (°C), yield (%), IR, 1 H NMR, and elemental analysis are reported.

General Procedure C (GP-C) to Obtain S-Configured Alcohols (11b−13b and 17).
A nitrogen atmosphere was established in a three-neck flask containing the appropriate imidazolyl-ethanone (1 mmol) and [(R,R)-TsDPEN]Ru-(p-cymene) Cl (0.001 mmol) before the addition of dichloromethane (7 mL) and  Amounts of 1, 8b, and 9b (40 mg/kg ip doses) quantified in duplicate in brain and blood.Reported as mean values ± standard deviation of triplicate analysis and calculated brain/blood ratio.Et 3 N (5 mmol).Formic acid (5 mmol) was added over a period of an hour.The mixture was stirred at room temperature for 26 h.NaHCO 3 saturated solution was added cautiously and the organic layer with water and brine and dried over anhydrous Na 2 SO 4 .The solvent was removed under vacuum, affording the desired product.For each compound, imidazolyl-ethanone, melting point (°C), yield (%), IR, Compound 12b was prepared from 15 according to the literature 50,51 by means of GP-C.173−175 °C; 58% as pink solid; IR ν OH 3116 cm −1 ; 1 H NMR (CD 3 CN, δ) 3.90 (bs, 1H, OH), 4.10 (dd, J = 7.3 Hz, J = 14.1 Hz, 1H, CH a H b ), 4.15 (dd, J = 4.5, J = 14. 1  Compound Evaluation in the Naegleria Cell-Based Assay.The compounds were screened against N. fowleri European KUL strain axenically cultured in Nelson's medium supplemented with 10% fetal bovine serum at 37 °C. 59The screening assay in 96-and 384-well formats (Z'-value of 0.95 ± 0.1) was performed as described elsewhere. 9All the experiments were performed using trophozoites harvested during the logarithmic phase of growth. 60The primary screen was performed in duplicate at 10 μM in a 384-well format (2,500 amoebae/well).The dose−response curves were generated in triplicate for the follow up compounds by serial dilution of compounds from 50 to 0.39 μM in a 96-well plate with 10,000 amoebae/well.Assay plates were incubated for 48 h at 37 °C, and cell viability was determined by a CellTiter-Glo Luminescent Cell Viability Assay. 9,58The experiments were conducted in a biosafety cabinet following BSL-2 procedures as specified in the UCSD Biosafety Practices Guidelines.
Binding Titrations of NfCYP51.Spectral binding titrations of NfCYP51 with the test compounds were performed at 25 °C with 0.5 μM NfCYP51.The concentration of NfCYP51 was determined for the ferrous (reduced) carbon monoxide (CO)-bound species at 450 nm (ε = 91,000 M −1 cm −1 ) that represents an active protein fraction with the intact heme Fe thiolate bond.All the inhibitors were dissolved in 100% DMSO as 100 and 200 μM stock solutions.Miconazole and fluconazole were used as references.For each titration, 4 mL of NfCYP51 (in 50 mM potassium phosphate, pH 8.0, and 10% glycerol) was split equally between the reference and sample plastic UV-cuvettes (Cat.#67.758;Sarstedt, Germany).Inhibitor aliquots of 0.5 μL were successively added to the sample cuvette in 25 nM (first eight aliquots) and 50 nM (last six aliquots) increments in a concentration range of 0.025−0.5 μM; DMSO alone was added to the reference cuvette, with the total added volume being less than 1% of the sample volume.Spectra were recorded from 350 to 500 nm.A binding isotherm was generated by plotting a difference between the absorbance maximum at 430 nm and the absorbance minimum at 410 nm as a function of inhibitor concentration.The data were analyzed in GraphPad Prism 9 with the rearrangement of the Morrison binding equation 61 to determine the dissociation constants ), where ΔA is the difference between the absorbance maximum and minimum, ΔA max is the extrapolated maximum absorption difference, [L] is the ligand (inhibitor) concentration, and [E] is the enzyme concentration.
X-ray Crystallography.Prior to crystallization, NfCYP51 at 0.5 mM concentration in storage buffer (50 mM potassium phosphate, pH 8.0) was incubated with 1.2 molar excess of the respective ligand for 30 min on ice.Crystals were then set up in 96-well plates using a hanging drop crystallization protocol and Mosquito liquid pipetting robot (STP LabTech, Boston, MA).We have used crystallization conditions similar to our report for NfCYP51 previously: 30 mM CaCl 2 ; 33% v/v PEG MME 550; 100 mM bis-Tris propane, pH 7.0; concentration of Jefframine M-600 varied from 3 to 4.6%. 15The plates were configured by using the Dragonfly liquid pipetting robot equipped with the Designer software (STP LabTech, Boston, MA).All the crystals were obtained from 23 °C.Diffraction data were collected remotely at beamline 8.3.1,Advanced Light Source, Lawrence Berkeley National Laboratory.Data indexing, integration, and scaling were conducted using XDS. 62The structures were determined by molecular replacement using the NfCYP51− posaconazole complex (PDB 5TL8) as a molecular replacement model.The final models were built and refined using the BUCCANEER and REFMAC5 modules of the CCP4 software suite 63 and COOT software. 64Data collection and refinement statistics are listed in Table 2.
In Vivo BBB Permeability.All mice were maintained on a 12 h light/dark cycle in a temperature-controlled environment with access to food and water ad libitum.Mice were randomly divided into the following three groups (N = 2 each): (1) miconazole, mice were treated intraperitoneally (i.p.) with 40 mg/kg miconazole; (2) 8b, mice were treated i.p. with 40 mg/kg 8b; (3) 9b, mice were treated i.p. with 40 mg/kg 9b.Miconazole, 8b, and 9b were dissolved in 150 μL of DMSO.Mice were sacrificed 1 h following the drug administration, and blood and brain samples were collected.Blood (100 μL) was quickly added with 5-fold H 2 O/CH 3 CN 8:2, centrifugated to 2500 rpm for 15 min, and plasma was collected.Brain was homogenized in 0.5 mL of distilled water and added with 2 mL of CH 3 CN.Thus, homogenates were centrifugated to 2500 rpm for 15 min and the supernatants were collected.Plasma and brain supernatants were stored at −20 °C until analysis.
Ethics Statements.All experiments involving animals were carried out according to Sapienza University's Ethics Committee, approval code 890/2021-PR, approved on 17 November 2021.Animal care followed the IASP and European Community (EC L358/1, 18/12/ 86) guidelines on using and protecting animals in experimental research.Eight week-old female C57BL/6 mice were used for the experiments (Charles River, Lecco, Italy).
HPLC-ESI-MS/MS Analysis.HPLC-ESI-MS/MS Instrumental Conditions.The targeted analysis was performed by a Waters system composed of a 1525 μ HPLC (Milford, MA, USA), coupled with a Quattro Micro Tandem MS/MS with an ESI source (Micromass, Manchester UK), using a Supelco Ascentis Express C18 (15 cm × 2.1 mm) 2.7 μm analytical column, A (deionized water/formic acid 0.02%) and B (acetonitrile/formic acid 0.02%) as mobile phase, with the following optimized elution binary gradient with linear interpolation: 0−1 min, 30% B; Calibration Standard Solution Preparation and Curve Calculation.Three stock solutions of 1, 8b, and 9b were prepared by dissolving 1 mg mL −1 of each compound in methanol and stored at 4 °C.Each stock solution was appropriately diluted with the mobile phase (A/B, 70:30, v:v) and used to optimize the ESI source parameters and the MRM method.A working solution containing 1, 8b, and 9b was then prepared by diluting 1:100 with the mobile phase (A/B, 70:30, v:v) equal aliquots of each compound stock solution, and it was used to optimize the chromatographic conditions.
Working solutions containing all compounds, 1, 8b, and 9b, at the final concentrations of 2.5, 5.0, 10.0, 25.0, 35.0, and 50.0 ng mL −1 were prepared by appropriately diluting each compound stock solution with the mobile phase (A:B, 70:30 v:v) and analyzed in triplicate (25 μL injected) to construct the calibration curves.Calibration curves were calculated with equal weighted least-squares linear regression analysis of the MRM peak area against the standard nominal concentration.A very good linearity was found for all compounds in the analyzed range, with R 2 values of 0.99 for all compounds, as shown in Table S2.The matrix effect (ME) for each compound, in brain (B) and in plasma (P), was evaluated by comparing the matrix-matching calibration curve (2.5, 10, 25, and 50 ng mL −1 ) with the corresponding calibration curve. 65ME values, reported in Table S2, were found in agreement with a generally weak matrix effect. 66Chromatographic (elution time t R ) and mass spectral (MRM transition, dwell time) data are reported in Table S2 too.The calibration curves were used for the quantitation of each compound 1, 8b, and 9b identified in the analyzed biological matrices.
Biological Sample Preparation.The biological samples were appropriately diluted with the mobile phase (A:B, 70:30, v:v), filtered at 0.22 μm, and analyzed in triplicate (25 μL injected).The amounts of compounds 1, 8b, and 9b, quantitated in duplicate in brain and in blood samples are reported in Table 3 as mean values ± standard deviation of triplicate analysis.

■ ASSOCIATED CONTENT
* sı Supporting Information

Figure 1 .
Figure 1.Drug−target interactions of 2a.(A) Electron density map (blue mesh) at 1.75 Å delineates the binding pose of 2a (pale yellow sticks).Amino acid residues within <4 Å of the inhibitor are shown in green sticks.Heme is pink.(B) Binding pocket (volume of 5178 Å 3 ) accommodating 2a (pink) is shown in shades of green.Amino acid side chains at the site boundaries are in blue sticks.(C) Superimposition of 2a (pale yellow) and fluconazole (dark red, from PDB 6AY4) is shown.(B, C) Heme is yellow.

Figure 2 .
Figure 2. 2a binding pocket.(A) Imidazole and 4-fluorophenyl moieties binding.(B) 3-Trifluoromethyl-benzoyl moiety binding.In (A) and (B), 2a is in pink, heme (HEM) is in yellow, and amino acid residues at the pocket boundaries are in blue.Pocket boundaries are green.Blue arrows point at the adjacent void spaces.

Figure 3 .
Figure 3. Binding mode of compounds (A) 2a, (B) 4b, and (C) 5b to NfCYP51.In (A)−(C), the inhibitors are colored pink, heme is in yellow, and amino acid residues at the pocket boundaries are in blue.Heteroatoms are colored oxygen in red, nitrogen in blue, chlorine in green, and fluorine in olive.

Figure 4 .
Figure 4. Compounds bind to NfCYP51.Binding isotherms of (A) miconazole (blue), fluconazole (cyan), and 2a (black), (B) 4b and 8b, (C) 6b and 10b, and (D) 5b and 9b.Solid and hollow circles represent ester and ether derivatives, respectively.A max is the absorbance at 430 nm, and A min is the absorbance at 410 nm.The NfCYP51 concentration is 0.5 μM.Solid curves represent the Morrison binding equation fit, whereas dotted lines demonstrate the linear dependence of the data points on inhibitor concentrations.Binding were performed in triplicate, and standard deviations in three independent titrations are shown in red; for most of the data points, deviations are smaller than the size of symbols and thus are invisible.

Table 1 .
Data Collection and Refinement Statistics a Data for the highest resolution shell are shown in parentheses.b PEG MME: polyethylene glycol monomethyl ether.