New Multifunctional Agents for Potential Alzheimer’s Disease Treatment Based on Tacrine Conjugates with 2-Arylhydrazinylidene-1,3-Diketones

Alzheimer’s disease (AD) is considered a modern epidemic because of its increasing prevalence worldwide and serious medico-social consequences, including the economic burden of treatment and patient care. The development of new effective therapeutic agents for AD is one of the most urgent and challenging tasks. To address this need, we used an aminoalkylene linker to combine the well-known anticholinesterase drug tacrine with antioxidant 2-tolylhydrazinylidene-1,3-diketones to create 3 groups of hybrid compounds as new multifunctional agents with the potential for AD treatment. Lead compounds of the new conjugates effectively inhibited acetylcholinesterase (AChE, IC50 0.24–0.34 µM) and butyrylcholinesterase (BChE, IC50 0.036–0.0745 µM), with weak inhibition of off-target carboxylesterase. Anti-AChE activity increased with elongation of the alkylene spacer, in agreement with molecular docking, which showed compounds binding to both the catalytic active site and peripheral anionic site (PAS) of AChE, consistent with mixed type reversible inhibition. PAS binding along with effective propidium displacement suggest the potential of the hybrids to block AChE-induced β-amyloid aggregation, a disease-modifying effect. All of the conjugates demonstrated metal chelating ability for Cu2+, Fe2+, and Zn2+, as well as high antiradical activity in the ABTS test. Non-fluorinated hybrid compounds 6 and 7 also showed Fe3+ reducing activity in the FRAP test. Predicted ADMET and physicochemical properties of conjugates indicated good CNS bioavailability and safety parameters acceptable for potential lead compounds at the early stages of anti-AD drug development.


Introduction
Alzheimer's disease (AD) is the most common form of dementia in old age. It is currently considered a modern epidemic because of its increasing prevalence worldwide spectra) and C 6 F 6 (for 19 F NMR spectra, δ -162.9 ppm). The microanalyses (C, H, N) were carried out on a PerkinElmer PE 2400 series II (PerkinElmer, Waltham, MA, USA) elemental analyzer. The column chromatography was performed on Silica gel 60 (0.062-0.2 mm) (Macherey-Nagel GmbH & Co KG, Duren, Germany).
The activity of CES was determined spectrophotometrically by the release of 4nitrophenol at 405 nm in 0.1 M K/Na phosphate buffer pH 8.0, 25 • C. Final enzyme and substrate (4-nitrophenyl acetate) concentrations were 0.02 unit/mL and 1 mM, respectively. Assays were carried out with a blank containing all constituents except porcine CES to assess non-enzymatic hydrolysis. Measurements were performed with a FLU-OStar Optima microplate reader (BMG Labtech, Ortenberg, Germany). Compounds were dissolved in DMSO; the incubation mixture contained 2% (v/v) solvent. The primary evaluation of the inhibitory activity of the compounds was performed by determining the degree of the enzyme inhibition at a compound concentration of 20 µM. For this, a sample of the corresponding enzyme was incubated with the test compound for 5 min; then the enzyme residual activity was determined. Each experiment was performed in triplicate. Compounds inhibiting the enzyme by more than 30% were then selected for determination of IC 50 values (the inhibitor concentration resulting in 50% inhibition of control enzyme activity). Compounds (eight concentrations ranging between 1 × 10 −11 and 1 × 10 −4 M were used to achieve 20 to 80% inhibition) were incubated with each enzyme for 5 min at 25 • C (for temperature equilibration). Substrate was then added and residual enzyme activity relative to an inhibitor-free control was measured using a FLUOStar Optima microplate reader.

Kinetic Study of AChE and BChE Inhibition. Determination of Inhibition Mechanism and Steady-State Inhibition Constants
Mechanisms of human erythrocyte AChE and equine serum BChE inhibition were assessed via a thorough analysis of enzyme kinetics. Residual activity was measured following 5 min incubation at 25 • C with three increasing concentrations of inhibitor and six decreasing substrate concentrations. Inhibition constants K i (competitive component) and αK i (noncompetitive component) were determined by linear regression of 1/V versus 1/[S] double-reciprocal (Lineweaver-Burk) plots.

Propidium Displacement from EeAChE PAS
Propidium iodide, donepezil, and Tris were purchased from Sigma-Aldrich (St. Louis, MO, USA). The ability of the test compounds to competitively displace propidium, a selective ligand of the PAS of AChE, was evaluated by a fluorescence method [85]. Electrophorus electricus (EeAChE) (electric eel, type VI-S, lyophilized powder, Sigma-Aldrich, St. Louis, MO, USA) was used owing to its high degree of purification, high activity, and lower cost than human AChE (hAChE). The applicability of this enzyme has been substantiated earlier [67]. The fluorescence intensity of propidium iodide bound with AChE increases several times; decreasing fluorescence intensity of the bound propidium in the presence of the test compounds shows their ability to bind to the PAS of AChE. To determine the degree of displacement (% displacement) of propidium from the PAS of AChE, EeAChE (final concentration, 7 µM) was incubated with the test compound at a concentration of 20 µM in 1 mM Tris-HCl buffer pH 8.0, 25 • C, for 15 min. Then, propidium iodide solution (final concentration, 8 µM) was added, the samples were incubated for 15 min, and the fluorescence spectrum (530 nm (excitation) and 600 nm (emission)) was taken. Donepezil and tacrine were used as reference compounds. The blank contained propidium iodide of the same concentration in 1 mM Tris-HCl buffer, pH 8.0 at 25 • C. The measurements were carried out in triplicate on a FLUOStar Optima microplate reader (LabTech, Ortenberg, Germany), and the results were calculated by the following formula: where IF AChE + Propidium is the fluorescence intensity of the propidium associated with AChE in the absence of the test compound (taken as 100%), and IF AChE + Propidium + inhibitor is the fluorescence intensity of the propidium associated with AChE in the presence of the test compound.

ABTS Radical Cation Scavenging Activity Assay
Radical scavenging activity of the compounds was evaluated by the ABTS radical cation (ABTS •+ ) scavenging assay showing the ability of the compounds to decolorize the ABTS •+ solution [86] with some modifications [87]. Trolox was used as the antioxidant standard; ascorbic acid was used as the comparison compound. All tested compounds and standards were dissolved in DMSO.
The solution of ABTS •+ was produced by mixing 7 mM ABTS aqueous solution with 2.45 mM potassium persulfate aqueous solution in equal quantities and allowing them to react for 12-16 h at room temperature in the dark. Radical scavenging capacity of the compounds was analyzed by mixing 10 µL of compound with 240 µL of ABTS •+ working solution in ethanol (100 µM final concentration), and after 1 h of incubation at 30 • C the decrease in absorbance was measured spectrophotometrically at 734 nm using a Bio-Rad xMark microplate UV/VIS spectrophotometer (Bio-Rad, Hercules, CA, USA). The compounds were tested in the concentration range 5 × 10 −7 -1 × 10 −4 M. Ethanol blanks were run in each assay. Values were obtained from five replicates of each sample and three independent experiments.
Antioxidant activity was reported as Trolox equivalent antioxidant capacity (TEAC values), consisting of the ratio between the slopes obtained from the linear correlation for concentrations of the tested compounds and Trolox with absorbance of ABTS radical. For the compounds, we also determined IC 50 values (compound concentration (µM) required for 50% reduction of the ABTS radical). The calculations were carried out using Origin 6.1 for Windows (OriginLab, Northampton, MA, USA). Results are presented as mean ± SEM calculated using GraphPad Prism version 6.05 for Windows, GraphPad Software (San Diego, CA, USA).
The ferric reducing ability of the compounds was determined by a previously described method [89] as a microplate-adapted version described in [42]. 2,4,6-tris(pyridin-2-yl)-1,3,5-triazine (TPTZ), FeCl 3 ·6H 2 O, Trolox, ascorbic acid and DMSO were obtained from Sigma-Aldrich Chemical Co (St. Louis, MO, USA). The FRAP reagent was prepared by mixing acetate buffer (0.3 M, pH 3.6), TPTZ (10 mM in 40 mM HCl) and FeCl 3 ·6H 2 O (20 mM in distilled water) in a ratio of 10:1:1 immediately before use. Compounds were dissolved in DMSO and tested in the concentration range of 1×10 −6 -1×10 −4 M. The solvent content in the reaction mixture was 4% (v/v). The test compounds (10 µL) were added to the FRAP reagent solution (240 µL) and mixed thoroughly. The reaction was carried out at 37 • C in the dark, the incubation time was 1 h. The absorbance at 600 nm was monitored spectrophotometrically by a FLUOStar OPTIMA microplate reader (BMG Labtech, Germany) at 37 • C. Trolox was used as a standard antioxidant, ascorbic acid as a reference compound. Values were obtained from four replicates of each sample and three independent experiments.
The ferric reducing ability of compounds was expressed as TE units (antioxidant activity in Trolox equivalent) with the values calculated as the ratio of the concentrations of Trolox and the test compound resulting in the same effect.

Metal-chelating Properties of Compounds 6a, 7a and 8a
The complexing studies were made in acetonitrile at 25 • C using a UV-vis spectrophotometer Shimadzu UV-2600 (Shimadzu Corporation, Kyoto, Japan) with wavelength ranging from 190 to 600 nm. Solutions (200 µM in acetonitrile) of the following metals compounds were prepared in volumetric flasks: FeCl 2 ·4H 2 O (99%, Acros Organics by Thermo Fisher Scientific (Kandel, Germany)), CuCl 2 (98%, Alfa Aesar by Thermo Fisher Scientific (Kandel, Germany)), or Zn(NO 3 ) 2 ·6H 2 O (98%, Alfa Aesar by Thermo Fisher Scientific (Kandel, Germany). Solutions of the test compounds were also prepared in acetonitrile at 400 µM concentrations. To a mixture of 0.5 mL test compound solution (40 µM final concentration) and 3.5 mL acetonitrile, 1 mL of the metal solution (CuCl 2 , FeCl 2 ·4H 2 O, or Zn(NO 3 ) 2 ·6H 2 O; 40 µM final concentration) was added. The solution was incubated at 25 • C for 30 min and then the absorption spectra were recorded at 25 • C in a 1 cm quartz cell. The control was prepared by mixing 0.5 mL tested compound solution and 4.5 mL acetonitrile.

Molecular Modeling Studies QM analysis of the Structures
Estimations of pK a values were performed with the Calculator Plugins of MarvinSketch 21.14.0, ChemAxon (http://www.chemaxon.com, accessed on 19 September 2021). Because the pK a values of the tacrine fragment for all considered compounds was estimated as 8.89, all conjugates were used for all further calculations with a protonated endocyclic nitrogen atom of the tacrine fragment.
For all considered compounds an array of possible tautomers was also generated with the Calculator Plugins of MarvinSketch 21.14.0. A conformational search was performed with TorsiFlex v. 2021.3 [91,92], for all torsions of the hydrazone-diketone fragment, with connectivity of the mobile hydrogen skipped. The following main parameters of TorsiFlex were used: 10,000 of steps of a stochastic algorithm (increased to 20,000 and 50,000 when an additional search was needed), HF/3-21G for low level and B3LYP/6-31G* for high level calculations.
QM optimization of the generated structures, as well as proton transfer pathway calculations (optimization, transition state (TS) search and intrinsic reaction path calculations (IRC)) were performed with Gaussian 16 [93] using a DFT method (B3LYP/6-31G*).

Molecular Docking
The X-ray structure of human AChE co-crystallized with donepezil (PDB ID 4EY7 [94]) after removal of water molecules and other molecules, and an optimized X-ray structure of human BChE (PDB ID 1P0I [95,96]) were used for molecular docking. Structures of the ligands in the most stable configuration after QM optimization were used. Partial atomic charges on ligand atoms were assigned from QM data according to the Löwdin scheme [97].

Prediction of ADMET and Physicochemical Profiles
Lipophilicity (LogP ow ) and aqueous solubility (pS) were estimated by the ALogPS 3.0 neural network model implemented in the OCHEM platform [100]. Human intestinal absorption (HIA) [101], blood-brain barrier distribution/permeability (LogBB) [102,103], and hERG-mediated cardiac toxicity risk (channel affinity pK i and inhibitory activity pIC 50 ) [104] were estimated using the integrated online service for the prediction of ADMET properties [105]. This service implements predictive QSAR models based on accurate and representative training sets, fragmental descriptors, and artificial neural networks. The quantitative estimate of drug-likeness (QED) values [106] were calculated using RDKit version 2021.09.2 software [107].

Chemistry
First, the interaction of 2-tolylhydrazinylidene-1,3-diketones 1a-d with hexylamine was studied as a model reaction for the synthesis of conjugates with aminomethylenemodified tacrine. It was found that 2-tolylhydrazinylidene-substituted acetyl-and benzoylacetones 1a,b reacted with hexylamine at the acetyl moiety in refluxing methanol to give products 2a,b in good yields (Scheme 1). Note that the reaction of the benzoylacetone derivative 1b bearing non-equivalent carbonyl centers proceeded chemoselectively to produce the only product 2b. The regioisomeric structure of compound 2b was established by 13 C NMR spectroscopy. The signal of the carbonyl carbon atom in the 13 C NMR spectrum was observed at δ 193 ppm, typical of the benzoyl fragment. The resonating signal of the acetyl group carbon atom in product 2a was observed in the downfield region at δ 198 ppm [108].
The reaction of 1,1,1-trifluoro-3-[2-(4-methylphenyl)hydrazinylidene]pentane-2,4-dione 1c and hexylamine occurred less selectively. This was confirmed by the formation of a mixture of products from which 1,1,1-trifluoro-4-hexylimine-3-(2-[4-methylphenyl)hydrazinylidene]pentan-2-one 2c was isolated in moderate yield (Scheme 1). Its regioisomeric structure was confirmed by 13 C NMR spectroscopy. In the 13 C NMR spectrum of 2c, the signal of the carbonyl atom at the CF 3 group was observed as a quartet in the same region at δ~177 ppm as for the initial diketone 1c [109]. We have shown earlier the preference in the condensation of trifluoromethyl-containing 2-arylhydrazinylidene-1,3diketones at the carbonyl group of the non-fluorinated substituent for reactions with methylamine [110]. However, the mass spectrum of the reaction mixture had molecular ion peaks of 2,2,2-trifluoro-N-hexylacetamide 3 (m/z [C 8  benzoylacetone derivative 1b bearing non-equivalent carbonyl centers proceeded chemoselectively to produce the only product 2b. The regioisomeric structure of compound 2b was established by 13 C NMR spectroscopy. The signal of the carbonyl carbon atom in the 13 C NMR spectrum was observed at δ 193 ppm, typical of the benzoyl fragment. The resonating signal of the acetyl group carbon atom in product 2a was observed in the downfield region at δ 198 ppm [108].
The reaction of 1,1,1-trifluoro-3-[2-(4-methylphenyl)hydrazinylidene]pentane-2,4dione 1c and hexylamine occurred less selectively. This was confirmed by the formation of a mixture of products from which 1,1,1-trifluoro-4-hexylimine-3-(2-[4methylphenyl)hydrazinylidene]pentan-2-one 2c was isolated in moderate yield (Scheme 1). Its regioisomeric structure was confirmed by 13 C NMR spectroscopy. In the 13 C NMR spectrum of 2c, the signal of the carbonyl atom at the CF3 group was observed as a quartet in the same region at δ ~ 177 ppm as for the initial diketone 1c [109]. We have shown earlier the preference in the condensation of trifluoromethyl-containing 2arylhydrazylidene-1,3-diketones at the carbonyl group of the non-fluorinated substituent for reactions with methylamine [110]. However, the mass spectrum of the reaction mixture had molecular ion peaks of 2,2,2-trifluoro- In contrast, the reaction of the 4,4,4-trifluoro-2-[2-(4-methylphenyl)hydrazinylidene]-1-phenylbutane-1,3-dione 1d with hexylamine in refluxing methanol or at room temperature led to a mixture of products. According to the GC/MS, there were predominant peaks corresponding to molecular ions of 2,2,2-trifluoro-N-hexylacetamide 3 and 2-[2-(4-methylphenyl)hydrazinylidene]-1-phenylethan-1-one 4b (m/z [C15H14N2O] + = 239). In this case, the addition of hexylamine at the trifluoroacyl group of 1,3-diketone 1d to form intermediate A and its subsequent cleavage to amide 3 and ketone 4b were apparently to become preferable (Scheme 1). It should be noted that at room temperature incomplete conversion of the initial diketone 1d along with the formation of cleavage products 3 and 4b was observed, and at a temperature of 0-5 °C there were no noticeable changes in the initial reagents. In contrast, the reaction of the 4,4,4-trifluoro-2-[2-(4-methylphenyl)hydrazinylidene]-1-phenylbutane-1,3-dione 1d with hexylamine in refluxing methanol or at room temperature led to a mixture of products. According to the GC/MS, there were predominant peaks corresponding to molecular ions of 2,2,2-trifluoro-N-hexylacetamide 3 and 2-[2-(4methylphenyl)hydrazinylidene]-1-phenylethan-1-one 4b (m/z [C 15 H 14 N 2 O] + = 239). In this case, the addition of hexylamine at the trifluoroacyl group of 1,3-diketone 1d to form intermediate A and its subsequent cleavage to amide 3 and ketone 4b were apparently to become preferable (Scheme 1). It should be noted that at room temperature incomplete conversion of the initial diketone 1d along with the formation of cleavage products 3 and 4b was observed, and at a temperature of 0-5 • C there were no noticeable changes in the initial reagents.
Then, it was found that 4-tolylhydrazinylidene-1,3-diketones 1a-c react with aminomethylene tacrine derivatives 5a-c (synthesized according to the methodology [81]) to give conjugates 6a-c, 7a-c, 8a-c with various substituents in the diketone moiety and the length of the methylene linker (Scheme 2). The best yields of products 6a-c and 7a-c were achieved in the reactions of non-fluorinated 1,3-diketones 1a and 1b in dry methanol under reflux for 8 h. The reaction of the trifluoromethyl-containing analogue 1c was carried out in a mixture of dry methylene chloride and dry methanol (3: 1) under reflux for 5 h. Note that the reaction in dry methanol or ethanol even at room temperature led to acidic cleavage of the starting 2-tolylhydrazinylidene-1,3-diketone 1c and significantly reduced the yields of the target conjugates 8a-c. length of the methylene linker (Scheme 2). The best yields of products 6a-c and 7a-c were achieved in the reactions of non-fluorinated 1,3-diketones 1a and 1b in dry methanol under reflux for 8 h. The reaction of the trifluoromethyl-containing analogue 1c was carried out in a mixture of dry methylene chloride and dry methanol (3: 1) under reflux for 5 h. Note that the reaction in dry methanol or ethanol even at room temperature led to acidic cleavage of the starting 2-tolylhydrazinylidene-1,3-diketone 1c and significantly reduced the yields of the target conjugates 8a-c. Scheme 2. Synthesis of conjugates 6, 7, 8.
The regioisomeric structure of compounds 6a-c, 7a-c, 8a-c was established by 13 C NMR spectroscopy. In the spectra of acetylacetone derivatives 6a-c, a characteristic signal of the carbonyl carbon atom of the acetyl group was observed at δ 198 ppm. In the case of benzoylacetone derivatives 7a-c, the carbonyl carbon atom of the benzoyl fragment resonated at δ 193 ppm, while the spectra of trifluoroacetylacetone derivatives 8a-c were characterized by a quartet signal of the trifluoroacetyl carbon atom at δ 177 ppm. It can be concluded that all compounds 6a-c, 7a-c, 8a-c were formed by condensation of 1,3diketones 1a-c with tacrines 5a-c at the acetyl fragment similarly to the reactions with hexylamine.
It should be noted that the reaction of diketone 1d with tacrines 5a-c was not effective despite attempts to vary the conditions.
The synthesized compounds 2a-c, 6a-c, 7a-c, and 8a-c have a hydrazone-diketone fragment with a mobile hydrogen atom. They can be characterized by prototropic iminoamine, azo-hydrazone and keto-enol tautomerism with the existence of four tautomeric forms AAK, HIK, AIE, and AIK, including Z,E-isomers for the first three forms (Scheme 3). All tautomers except AIK are stabilized by an intramolecular H-bond. The regioisomeric structure of compounds 6a-c, 7a-c, 8a-c was established by 13 C NMR spectroscopy. In the spectra of acetylacetone derivatives 6a-c, a characteristic signal of the carbonyl carbon atom of the acetyl group was observed at δ 198 ppm. In the case of benzoylacetone derivatives 7a-c, the carbonyl carbon atom of the benzoyl fragment resonated at δ 193 ppm, while the spectra of trifluoroacetylacetone derivatives 8a-c were characterized by a quartet signal of the trifluoroacetyl carbon atom at δ 177 ppm. It can be concluded that all compounds 6a-c, 7a-c, 8a-c were formed by condensation of 1,3-diketones 1a-c with tacrines 5a-c at the acetyl fragment similarly to the reactions with hexylamine.
It should be noted that the reaction of diketone 1d with tacrines 5a-c was not effective despite attempts to vary the conditions.
The synthesized compounds 2a-c, 6a-c, 7a-c, and 8a-c have a hydrazone-diketone fragment with a mobile hydrogen atom. They can be characterized by prototropic iminoamine, azo-hydrazone and keto-enol tautomerism with the existence of four tautomeric forms AAK, HIK, AIE, and AIK, including Z,E-isomers for the first three forms (Scheme 3). All tautomers except AIK are stabilized by an intramolecular H-bond. The 1 H NMR spectra of compounds 2a-c, 6a-c, 7a-c, and 8a-c in CDCl3 did not contain the CH proton signal of the AIK form. Instead, a low-field signal of the proton of the HN-or HO-group was observed at δ 14.7-15.4 ppm and, as already mentioned, all 13 C NMR spectra contained low-field signals of carbonyl carbon atoms. These spectral data allowed us to exclude the AIK and AIE tautomers from consideration.
The choice between tautomers AAK and HIK was made using 1 H NMR spectral data and our previous experience. We have earlier shown that 4-(N-methyl)amino-1,1,1trifluoro-3-phenylazopent-3-en-2-one [110] exists in the solid state as an Z-azo-aminoketone tautomer (Z-AAK) according to X-ray diffraction data, and upon dissolution in chloroform transforms into the Z-hydrazo-keto-imine tautomer (Z-HIK), given that in its 1 H NMR spectrum the signal of the N-methyl group was observed as a singlet at δ 2.92 ppm. In contrast, in the 1 H NMR spectra of compounds 2a-c, 6a-c, 7a-c, and 8a-c, the  of compounds 2a-c, 6a-c, 7a-c, and 8a-c. The 1 H NMR spectra of compounds 2a-c, 6a-c, 7a-c, and 8a-c in CDCl 3 did not contain the CH proton signal of the AIK form. Instead, a low-field signal of the proton of the HNor HO-group was observed at δ 14.7-15.4 ppm and, as already mentioned, all 13 C NMR spectra contained low-field signals of carbonyl carbon atoms. These spectral data allowed us to exclude the AIK and AIE tautomers from consideration.
The choice between tautomers AAK and HIK was made using 1 H NMR spectral data and our previous experience. We have earlier shown that 4-(N-methyl)amino-1,1,1-trifluoro-3-phenylazopent-3-en-2-one [110] exists in the solid state as an Z-azo-aminoketone tautomer (Z-AAK) according to X-ray diffraction data, and upon dissolution in chloroform transforms into the Z-hydrazo-keto-imine tautomer (Z-HIK), given that in its 1 H NMR spectrum the signal of the N-methyl group was observed as a singlet at δ 2.92 ppm. In contrast, in the 1 H NMR spectra of compounds 2a-c, 6a-c, 7a-c, and 8a-c, the signals of the N-CH 2 groups were observed as a triplet of doublets at δ 3.4-3.5 ppm due to interaction with the neighboring NH proton.
The 1 H and 13 C NMR spectra of products 2a,c, 6a-c, 7a-c, 8a-c in CDCl 3 contained one set of signals. The spectra of compound 2b containing benzoyl and N-hexylamine substituents were characterized by the presence of a second set of signals, and its 1 H NMR spectrum exhibited two triplets of doublets of a HN-CH 2 group at δ 3.18 and 3.52 ppm, apparently due to the existence of compounds 2b in the form of Z,E-isomers in a ratio of 87:13.
Based on the analysis performed, we believe that in a CDCl 3 solution, compounds 2a-c, 6a-c, 7a-c, and 8a-c predominantly exist in the Z-AAK form in contrast to the N-methyl analog [110] (Scheme 3).
We obtained monocrystals for compound 8a by slow crystallization from acetonitrile. The XRD analysis showed that conjugate 8a exists in the solid state also as Z-AAK (Figure 1) with formation of the intramolecular hydrogen bond between hydrogen H3 of NH group and nitrogen N2 of arylazo moiety (the distance H3···N2 is 1.650 Å) similarly to the N-methyl analog [110]. The molecule of compound 8a has an almost flat structure because 4-amino-1,1,1-trifluoro-3-(tolyldiazenyl)pent-3-en-2-one moiety and tacrine core are located in one plane. However, the butyl spacer C18C17C16C15 is characterized by distortions with 1.293 Å (for C18) and 1.131 Å (for C15) deviations from the common plane. The distance between nitrogens N3···N4 connected with butyl linker is 5.304 Å. To assess computationally the relative stability of tautomers and conformers of the  compounds 2a-c, 6a-c, 7a-c, and 8a-c, instead of manual generation of configurations of interest, as we have formerly done [79,87,111], we resorted to formal generation of tautomers with ChemAxon MarvinSketch. Next, geometries of all generated tautomers (from 34 for compounds 2b and 2c to 86 for compounds 6a-6c), were optimized quantummechanically. The optimized structures of the tautomers were subjected to conformational search with TorsiFlex. Because the conformational search led to a significant decrease of energy of certain tautomers (>20 kcal/mol for example of compound 8a), it proved to be necessary to include in the conformational search all of the generated tautomers, instead of selecting some of them (e.g., top-10, or within 10 kcal/mol from the top tautomer).
The most stable conformation for compound 8a found by TorsiFlex was of the E-AAK isomer. Chemically, E/Z isomerization of the double bond transforms it into a Z-AAK isomer. However, in terms of the conformational search, it is rotation around bond C21- To assess computationally the relative stability of tautomers and conformers of the  compounds 2a-c, 6a-c, 7a-c, and 8a-c, instead of manual generation of configurations of interest, as we have formerly done [79,87,111], we resorted to formal generation of tautomers with ChemAxon MarvinSketch. Next, geometries of all generated tautomers (from 34 for compounds 2b and 2c to 86 for compounds 6a-c), were optimized quantummechanically. The optimized structures of the tautomers were subjected to conformational search with TorsiFlex. Because the conformational search led to a significant decrease of energy of certain tautomers (>20 kcal/mol for example of compound 8a), it proved to be necessary to include in the conformational search all of the generated tautomers, instead of selecting some of them (e.g., top-10, or within 10 kcal/mol from the top tautomer).
The most stable conformation for compound 8a found by TorsiFlex was of the E-AAK isomer. Chemically, E/Z isomerization of the double bond transforms it into a Z-AAK isomer. However, in terms of the conformational search, it is rotation around bond C21-C22 (according to Figure 1

labeling), that transforms the E-AAK isomer into Z-AAK.
Multiple increases in the number of cycles of the conformational search led to the same E-AAK leading conformation, while the Z-AAK conformation was not found among the solutions. However, the presence in the results of other conformations requiring rotation around this bond but much higher in energy, proves that this torsion was included in the conformational search.
A manual flip around bond C21-C22 of the best conformational search solution transforming E-AAK into Z-AAK, followed by QM energy minimization, led to a 2.12 kcal/mol energy decrease, which is in agreement with the crystallographic results. Thus, formal generation of all possible tautomers followed by a conformational search required considerable computational resources but did not lead to the most stable configuration, while manual generation of possible isomers, tautomers, and major conformers proved to be more efficient.
The Z-AAK tautomer can be also transformed into Z-HIK via short-distance proton transfer. The energy profile of this process in vacuo was calculated for compound 8a (Figure 2), yielding a 4.5 kcal/mol energy barrier and a 1.05 kcal/mol energy loss. Thus, the system predominantly exists in the Z-AAK form.

Esterase Profile Assessment
The method for evaluation of the esterase profile developed by our group includes the determination of the inhibitory activity of the synthesized compounds against enzymes of the cholinesterase family-AChE and BChE-as well as a structurally close enzyme-carboxylesterase (CES). Inhibition of AChE and BChE in the brain increases acetylcholine levels and improves cognitive functions in AD. CES is responsible for hydrolysis of numerous therapeutically important ester-containing drugs, and therefore inhibition of CES by anticholinesterase compounds used in AD therapy could lead to adverse drug-drug interactions.
Human erythrocyte AChE, equine serum BChE, and porcine liver CES were used to assess the esterase profile. The applicability of this set of enzymes for this purpose has been shown earlier [112][113][114][115]. The results on the esterase profile of model compounds 2 and conjugates 6-8 are displayed in Table 1.  of compounds 2,6,7,8 and their ability to displace propidium iodide from the peripheral

Esterase Profile Assessment
The method for evaluation of the esterase profile developed by our group includes the determination of the inhibitory activity of the synthesized compounds against enzymes of the cholinesterase family-AChE and BChE-as well as a structurally close enzymecarboxylesterase (CES). Inhibition of AChE and BChE in the brain increases acetylcholine levels and improves cognitive functions in AD. CES is responsible for hydrolysis of numerous therapeutically important ester-containing drugs, and therefore inhibition of CES by anticholinesterase compounds used in AD therapy could lead to adverse drug-drug interactions.
Human erythrocyte AChE, equine serum BChE, and porcine liver CES were used to assess the esterase profile. The applicability of this set of enzymes for this purpose has been shown earlier [112][113][114][115]. The results on the esterase profile of model compounds 2 and conjugates 6-8 are displayed in Table 1. enzyme-carboxylesterase (CES). Inhibition of AChE and BChE in the brain increases acetylcholine levels and improves cognitive functions in AD. CES is responsible for hydrolysis of numerous therapeutically important ester-containing drugs, and therefore inhibition of CES by anticholinesterase compounds used in AD therapy could lead to adverse drug-drug interactions. Human erythrocyte AChE, equine serum BChE, and porcine liver CES were used to assess the esterase profile. The applicability of this set of enzymes for this purpose has been shown earlier [112][113][114][115]. The results on the esterase profile of model compounds 2 and conjugates 6-8 are displayed in Table 1. The results presented in Table 1 indicate that compounds 2a-c -N-hexylamine derivatives of 2-tolylhydrazinylidene-1,3-diketones 1a-c inhibit cholinesterases very weakly and do not inhibit CES. At the same time, all 3 groups of conjugates of 2tolylhydrazinylidene-1,3-diketones 1a-c with tacrine 6, 7, and 8 exhibit high inhibitory activity against cholinesterases -at the level and above the parent pharmacophore tacrine with predominant inhibition of BChE and weak inhibition of СES.
Prior to the research, we feared that the synthesized conjugates 6, 7, and 8 might exhibit anti-CES activity undesirable for AD therapy agents, because we have previously found that one of the starting compounds, 2-tolylhydrazinylidene-1,3-diketone 1c, exhibits moderate inhibitory activity against this enzyme [79]. However, it turned out that the isomerization of the hydrazone tautomer to the azo form upon replacement of one of the carbonyl groups by the alkylamine functional group led to a significant decrease in the anti-CES activity.
AChE inhibition. The structure of the R substituent at the carbonyl carbon atom (Me, Ph, or CF3) had practically no effect on the anti-AChE activity of conjugates. At the same time, for all series of conjugates 6, 7, and 8, an increase in inhibitory activity against AChE was observed with an increase in the spacer length (from 4 to 8 CH2 groups). The anti- AChE activity of conjugates 6b,c, 7b,c, and 8b,c was higher than the activity of the parent pharmacophore tacrine . Compounds 6c, 7c, 8b, and 8c  The results presented in Table 1 indicate that compounds 2a-c-N-hexylamine derivatives of 2-tolylhydrazinylidene-1,3-diketones 1a-c inhibit cholinesterases very weakly and do not inhibit CES. At the same time, all 3 groups of conjugates of 2-tolylhydrazinylidene-1,3-diketones 1a-c with tacrine 6, 7, and 8 exhibit high inhibitory activity against cholinesterases-at the level and above the parent pharmacophore tacrine with predominant inhibition of BChE and weak inhibition of CES.
Prior to the research, we feared that the synthesized conjugates 6, 7, and 8 might exhibit anti-CES activity undesirable for AD therapy agents, because we have previously found that one of the starting compounds, 2-tolylhydrazinylidene-1,3-diketone 1c, exhibits moderate inhibitory activity against this enzyme [79]. However, it turned out that the isomerization of the hydrazone tautomer to the azo form upon replacement of one of the carbonyl groups by the alkylamine functional group led to a significant decrease in the anti-CES activity.
AChE inhibition. The structure of the R substituent at the carbonyl carbon atom (Me, Ph, or CF 3 ) had practically no effect on the anti-AChE activity of conjugates. At the same time, for all series of conjugates 6, 7, and 8, an increase in inhibitory activity against AChE was observed with an increase in the spacer length (from 4 to 8 CH 2 groups). The anti- AChE activity of conjugates 6b,c, 7b,c, and 8b,c was higher than the activity of the parent pharmacophore tacrine . Compounds 6c, 7c, 8b, and 8c exhibited the highest anti-AChE activity (IC 50 = 0.27, 0.25, 0.28 and 0.24 µM, respectively).
BChE inhibition. All conjugates were active against BChE. There was no pronounced effect of the spacer length on the inhibitory activity. However, anti-BChE activity depended on the structure of the R substituent at the carbonyl carbon atom: conjugates 7 with R = Ph were most active and reached the level of tacrine. The most effective BChE inhibitors were compounds 6c, 7a, 7b, 7c (IC 50 = 0.054, 0.036, 0.047, and 0.0745 µM, respectively).
The selectivity toward BChE compared to AChE is maximum for compounds with a short spacer (CH 2 ) 4 ; it is especially pronounced for 7a (R = Ph).

Kinetic Studies of AChE and BChE Inhibition
The mechanism of inhibitory action of the conjugates toward AChE and BChE was studied using compound 6b as a sample. The graphical analysis of the kinetic data on AChE ( Figure 3A) and BChE ( Figure 3B) inhibition by 6b in the Lineweaver-Burk doublereciprocal plot demonstrates the changes in both K m and V max that attest to a mixed type of inhibition. The inhibition constants are as follows: K i = 0.254 ± 0.018 µM (competitive component) and αK i = 0.473 ± 0.041 µM (noncompetitive component) for AChE and K i = 0.095 ± 0.008 µM (competitive component) and αK i = 0.224 ± 0.022 µM (noncompetitive component) for BChE. The mechanism of inhibitory action of the conjugates toward AChE and BChE was studied using compound 6b as a sample. The graphical analysis of the kinetic data on AChE ( Figure 3A) and BChE ( Figure 3B

Molecular Modeling Studies
The most stable Z-AAK conformer of the considered compounds was taken for molecular docking studies. It was found that all compounds bind to the hAChE in a uniform way, occupying both the CAS and PAS. In the CAS, the protonated tacrine fragment binds forming π-π stacking interactions with the Trp86 side chain and a hydrogen bond with its main chain hydrogen oxygen (Figure 4, A-C), as was observed previously for tacrine-containing conjugates [116,117]. Compounds 6a-c also form a few hydrogen bonds between Tyr124 phenolic hydroxyl group and the linker nitrogen atoms or carbonyl atom of the acetyl group (Figure 4, A). With increasing linker length, the tolyl substituent occupies more of the PAS, forming π-π or T-stacking interactions with Trp286. For compounds 7a-c these interactions are with the phenyl group, which is advancing into the PAS and interacting with Trp286, instead of with the tolyl substituent, while the

Molecular Modeling Studies
The most stable Z-AAK conformer of the considered compounds was taken for molecular docking studies. It was found that all compounds bind to the hAChE in a uniform way, occupying both the CAS and PAS. In the CAS, the protonated tacrine fragment binds forming π-π stacking interactions with the Trp86 side chain and a hydrogen bond with its main chain hydrogen oxygen ( Figure 4A-C), as was observed previously for tacrine-containing conjugates [116,117]. Compounds 6a-c also form a few hydrogen bonds between Tyr124 phenolic hydroxyl group and the linker nitrogen atoms or carbonyl atom of the acetyl group ( Figure 4A). With increasing linker length, the tolyl substituent occupies more of the PAS, forming π-π or T-stacking interactions with Trp286. For compounds 7a-c these interactions are with the phenyl group, which is advancing into the PAS and interacting with Trp286, instead of with the tolyl substituent, while the latter protrudes out of the gorge entrance. Additionally, the carbonyl oxygen atom of the keto-group forms hydrogen bonds with the Phe295 and Arg296 main chain nitrogen atoms ( Figure 4B).  In the case of BChE, binding is much more uniform with the carbonyl oxygen atom of the keto group in the oxyanion hole ( Figure 5, A-C). While ligands with increasing linker length occupy similar positions in the BChE active site, the tacrine fragment could form π-cation and π-π stacking interactions with Tyr332 or Trp82, supported by ionic interactions and hydrogen bonds with the Asp70 or His438 main chain oxygen atom, respectively. In the case of compounds 7a-c, there are additional π-π or T-stacking Regarding compounds 8a-c with a trifluoromethylketone (TFK) group, specific interactions change with elongation of the linker. For compound 8a, one of the fluorine atoms was found in the oxyanion hole. For compound 8b, it forms hydrogen bonds with the Phe295 and Arg296 main chain nitrogen atoms. For compound 8c, it interacts with Trp286 side chain, forming an F-π contact [118] (Figure 4C). Such modes of interactions were previously observed for a TFK compound, sliding down the gorge [119]. However, in the present case, no position with the carbonyl oxygen atom in the oxyanion site was found, which excludes the possibility of a subsequent covalent reaction between the enzyme and inhibitor.
For all the groups of inhibitors, increasing occupancy of the PAS with elongation of the linker can be associated with an increase in the propidium displacement ability.
In the case of BChE, binding is much more uniform with the carbonyl oxygen atom of the keto group in the oxyanion hole ( Figure 5A-C). While ligands with increasing linker length occupy similar positions in the BChE active site, the tacrine fragment could form π-cation and π-π stacking interactions with Tyr332 or Trp82, supported by ionic interactions and hydrogen bonds with the Asp70 or His438 main chain oxygen atom, respectively. In the case of compounds 7a-c, there are additional π-π or T-stacking interactions of the tolyl substituent with the Trp231 side chain ( Figure 5B). In the case of compounds 8a-c, the trifluoromethyl group forms F-π interactions with the Trp231 side chain ( Figure 5C). Overall, the position of the TFK group of compounds 8a-c in the catalytic site of BChE suggests the subsequent formation of a labile covalent bond and an adduct, called a tetrahedral intermediate analog [119]. The study of compounds as potential inhibitors of the proaggregant activity of AChE was carried out by assessment of the degree of displacement of the selective PAS ligand propidium iodide from the EeAChE PAS. As mentioned above, AChE PAS interacts with

Displacement of Propidium from the EeAChE PAS
The study of compounds as potential inhibitors of the proaggregant activity of AChE was carried out by assessment of the degree of displacement of the selective PAS ligand propidium iodide from the EeAChE PAS. As mentioned above, AChE PAS interacts with soluble Aβ peptides, promoting their aggregation. In this regard, compounds blocking the AChE PAS are potential antiaggregant agents.
As seen from Table 1, conjugates of tacrine with 2-tolylhydrazinylidene-1,3-diketones at a concentration of 20 µM reduce the fluorescence intensity of the propidium iodide bound to EeAChE and displace propidium at or above the level of the control compound donepezil (11-19%). These data indicate the potential ability of conjugates to block AChEinduced aggregation of β-amyloid, which agrees with the mixed type inhibition of AChE by the conjugates ( Figure 3A) and the results of molecular docking ( Figure 4A-C). On the whole, the ability to displace propidium in all groups of conjugates 6, 7, and 8 increases with spacer elongation in agreement with the molecular docking results. The conjugates 6a-c (R = Me) exhibit the highest activity, displacing propidium from EeAChE PAS in the range of 15.6-19%, while the optimal spacer length is n = 6.
The antioxidant activity of the compounds 2a-c, 6a-c, 7a-c, 8a-c was assessed by spectrophotometric ABTS and FRAP tests. The ABTS assay evaluates the binding of a model ABTS radical cation (ABTS •+ ), which is realized by the mechanism of single electron transfer (SET) and/or hydrogen atom transfer (HAT). The FRAP assay (Ferric Reducing Antioxidant Power) measures the ability of compounds to reduce the ferric 2,4,6-tripyridyls-triazine complex [Fe(TPTZ) 2 ] 3+ to [Fe(TPTZ) 2 ] 2+ , which occurs exclusively by the SET mechanism. The results are presented in Table 2.
It was found that model compounds aminoenketones 2a-c, which are the reaction products of the interaction of 2-arylhydrazinylidene-1,3-diketones 1a-c (Scheme 1) with hexylamine, exhibit high radical-scavenging activity in the ABTS assay at the level of the standard antioxidant Trolox. The activity of the compounds did not depend on the substituent R at the carbonyl carbon atom (CH 3 , Ph or CF 3 ). Conjugates 6a-c, 7a-c, 8a-c, which are a combination of 2-arylhydrazinylidene-1,3-diketones 1a-c and the anticholinesterase pharmacophore tacrine retained high ABTS •+ -scavenging activity of diketones at the Trolox level. The variation of the substituent R at the carbonyl carbon atom, as well as the spacer length, practically did not affect the radical-scavenging activity of the conjugates.
In the FRAP assay, the model compounds 2a,b and conjugates 6a-c, 7a-c, also demonstrated fairly good activity. However, their activity in the FRAP test was twice lower than in the ABTS test.
The most notable finding was that there was no influence of the structure of the R group in the carbonyl fragment on the activity in the ABTS test: both aminoenketone 2 and conjugates of 1,3-diketones with tacrine 6-8 were highly active. At the same time, in the FRAP test, compound 2c with R=CF 3 was not active and a weak ferric reducing ability was observed for conjugates 8a-c with the CF 3 substituent in the carbonyl moiety. Taking into account that the FRAP assay measures the Fe 3+ reducing ability occurring exclusively by the SET mechanism, whereas the binding of the ABTS radical cation can be realized by SET and/or HAT mechanisms, the observed different structure-activity relationships (in particular, no decrease in activity in the ABTS test for compounds 8a-c with a CF 3 substituent) may indicate different mechanisms of the antioxidant action of the studied compounds in the ABTS and FRAP tests. That is, the SET mechanism may be involved in the ferric reducing test and the HAT may be operating in the scavenging of the ABTS radical cation. model ABTS radical cation (ABTS •+ ), which is realized by the mechanism of single electron transfer (SET) and/or hydrogen atom transfer (HAT). The FRAP assay (Ferric Reducing Antioxidant Power) measures the ability of compounds to reduce the ferric 2,4,6-tripyridyl-s-triazine complex [Fe(TPTZ)2] 3+ to [Fe(TPTZ)2] 2+ , which occurs exclusively by the SET mechanism. The results are presented in Table 2. model ABTS radical cation (ABTS •+ ), which is realized by the mechanism of single electro transfer (SET) and/or hydrogen atom transfer (HAT). The FRAP assay (Ferric Reducin Antioxidant Power) measures the ability of compounds to reduce the ferric 2,4,6-tr pyridyl-s-triazine complex [Fe(TPTZ)2] 3+ to [Fe(TPTZ)2] 2+ , which occurs exclusively by th SET mechanism. The results are presented in Table 2.

Metal-chelating Properties
It is known that decreasing excess concentrations of metals in the brain by chelating agents is one of the approaches to AD treatment [51,52]. The complexation abilities of compounds 6a, 7a, 8a for biometals such as Cu 2+ , Fe 2+ and Zn 2+ in acetonitrile were studied by UV-Vis spectrometry [81] and the results for conjugates 6a, 7a, 8a are shown in Figure 6A-C. In the UV-Vis spectra of 6a, 7a, 8a, very broad bands from 340 to 470 nm and from 280 to 340 nm are ascribed to π-π* transitions, and the peak around 240 nm is due to intramolecular charge transfer.
The spectra of the individual ions and compounds were subtracted from the spectra of the mixtures of each compound with each ion. A red shift of absorbance calculated in this way indicates the formation of a ligand-ion complex in the mixture. For example (Figure 7), it can be seen that there are two red shifts of absorbance from 240 to 253 nm and from 328 to 339 nm for the interaction of conjugate 8a with Zn 2+ . For additional graphics, please see Supplementary, Figures S1-S8. The results obtained demonstrate metal-chelating ability for conjugates 6a, 7a, and 8a to all three biometal ions (Cu 2+ , Fe 2+ and Zn 2+ ).

Prediction of ADMET and Physicochemical Profiles
The results of our computational estimates of selected ADMET and physicochemical properties for compounds 2, 6-8 are shown in Table 3. All of the compounds had high predicted values for intestinal absorption, enabling their oral administration. Moreover, we could expect reasonable CNS activity in view of rather high predicted blood-brain barrier permeability (brain concentration exceeds the plasma concentration), although some optimization of this parameter might be desirable. The cardiac toxicity risk parameters (hERG pK i and pIC 50 ) fell within 4.2-7.6 log units for all the analyzed compounds, which was within the lower or medium part of their possible range (3-9 log units). According to the commonly accepted drug-likeness guidelines, the predicted lipophilicities and aqueous solubilities, as well as the molecular weights of the compounds, were within or close to the desirable range for potential drug compounds, although the LogP values in some cases violated the original Rule-of-5 limits (however, given that some of the compounds were outside of the model applicability domain, the predicted values were not fully reliable). The integral quantitative estimates of drug-likeness (QED) were in the 0.1-0.4 range. Consequently, the predicted ADMET and physicochemical properties of the compounds were acceptable for potential lead compounds in the discovery phase. Nevertheless, additional studies and structure optimization would be desirable to help maximize safety and improve the pharmacokinetic profile.

Conclusions
In summary, we synthesized the new conjugates of tacrine modified with 2-tolylazo-1,3-aminoenketone moieties with various acyl substituents through alkylene spacers of different lengths. According to NMR spectroscopy, the compounds were characterized by a predominant existence in solutions as a (Z)-2-azo-1,3-aminoenketone tautomer. According to XRD, compound 8a in crystals also exists in (Z)-2-azo-1,3-aminoenketone form that agrees with QM calculations.
In general, conjugates 6a-c, 7a-c, 8a-c demonstrated high inhibitory activity against both cholinesterases AChE and BChE, with selectivity for BChE, being mixed type inhibitors of both cholinesterases, and with weak inhibition of CES.
The structure of the R substituent at the carbonyl carbon atom (Me, Ph, or CF 3 ) had practically no effect on the anti-AChE activity of the conjugates. However, an increase in inhibitory activity was observed for all conjugates upon elongation of the alkylene spacer, in agreement with the results of molecular docking. Thus, a considerable group of compounds 6b,c, 7b,c, 8b,c showed high anti-AChE activity exceeding the activity of tacrine: IC 50 = 0.24-0.34 µM.
The synthesized conjugates demonstrated high inhibitory activity against BChE. Their binding in the BChE active site was more uniform than seen with AChE, with the carbonyl oxygen atom of the keto group in the oxyanion hole. The most active inhibitors of BChE were compounds 6c and 7a-c (IC 50 = 0.036-0.0745 µM), which reached the activity of tacrine.
All conjugates 6-8 showed a good ability to displace propidium from EeAChE PAS that along with the mixed type of inhibition revealed by the kinetics study and the molecular docking results indicates their potential ability to block AChE-induced aggregation of Aβ.
The synthesized conjugates have good antioxidant potency. All compounds 6-8 demonstrated the ability to effectively scavenge ABTS radical cation at the Trolox level; compounds 6, 7 also showed rather high Fe 3+ reducing activity in the FRAP test. Moreover, metalchelating ability for biometals such as Cu 2+ , Fe 2+ and Zn 2+ was demonstrated for conjugates 6a, 7a, and 8a.
Finally, computational predictions of physico-chemical and ADMET properties of the compounds were consistent with drug-like characteristics and low toxicity.
The obtained results allowed us to consider the synthesized conjugates as new multifunctional agents for the potential treatment of AD. In particular, the non-fluorinated conjugates 6b,c, 7b,c with (CH 2 ) 6 and (CH 2 ) 8 linkers exhibited the most promising set of properties and are recommended for further in-depth studies, e.g., lipid peroxidation in brain homogenates, neuroprotective activity in cell cultures, cognition enhancing efficacy in animal models, and safety and bioavailability assessments.