Novel Aminoguanidine Hydrazone Analogues: From Potential Antimicrobial Agents to Potent Cholinesterase Inhibitors

A series of thirty-one hydrazones of aminoguanidine, nitroaminoguanidine, 1,3-diaminoguanidine, and (thio)semicarbazide were prepared from various aldehydes, mainly chlorobenzaldehydes, halogenated salicylaldehydes, 5-nitrofurfural, and isatin (yields of 50–99%). They were characterized by spectral methods. Primarily, they were designed and evaluated as potential broad-spectrum antimicrobial agents. The compounds were effective against Gram-positive bacteria including methicillin-resistant Staphylococcus aureus with minimum inhibitory concentrations (MIC) from 7.8 µM, as well as Gram-negative strains with higher MIC. Antifungal evaluation against yeasts and Trichophyton mentagrophytes found MIC from 62.5 µM. We also evaluated inhibition of acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE). The compounds inhibited both enzymes with IC50 values of 17.95–54.93 µM for AChE and ≥1.69 µM for BuChE. Based on the substitution, it is possible to modify selectivity for a particular cholinesterase as we obtained selective inhibitors of either AChE or BuChE, as well as balanced inhibitors. The compounds act via mixed-type inhibition. Their interactions with enzymes were studied by molecular docking. Cytotoxicity was assessed in HepG2 cells. The hydrazones differ in their toxicity (IC50 from 5.27 to >500 µM). Some of the derivatives represent promising hits for further development. Based on the substitution pattern, it is possible to modulate bioactivity to the desired one.


Introduction
Molecular hybridization approach belongs to the most frequent and perspective tools in medicinal chemistry and drug design. Combination of various bioactive scaffolds in one molecular entity offers many advantages [1].
Aminoguanidine (AG; hydrazinecarboximidamide) inhibits selectively inducible nitric oxide synthase and scavenges reactive oxygen species, thus being antioxidant agent protecting various cells and tissues from oxidative stress. Moreover, AG was the first inhibitor of the advanced glycation pathway that plays a crucial role in the pathogenesis of Pharmaceuticals 2021, 14 [2,3]. In addition, AG and its derivatives also showed multiple anti-inflammatory properties [4][5][6]. However, the therapeutical potential is abolished due to nucleophilic reaction with pyridoxal phosphate (active form of vitamin B6) leading to its depletion and in vivo deficiency. It has been suggested that the formation of AG hydrazones does not influence levels of this vital molecule [3]. The condensation of AG with both "simple" aldehydes/ketones and carbonyl compounds with an intrinsic activity has provided conjugates with a significant antibacterial and antifungal activity covering a broad spectrum of Gram-positive, Gram-negative and fungal species such as drug-resistant strains including methicillin-resistant Staphylococcus aureus (MRSA) [5][6][7][8][9]. Replacement of AG by 1,3-diaminoguanidine is tolerated and vice versa [9]. Apart from direct killing of bacteria, hydrazones of AG are capable to inhibit efflux pumps, thus restoring activity of antibiotics and helping overcome acquired resistance, as demonstrated for fluoroquinolone norfloxacin against Staphylococcus aureus [10].
Salicylaldehyde-based hydrazones and imines represent a class of compounds with a great variety of biological activities. For design of this study, antibacterial, antifungal, and cytotoxic actions are of a special interest [16][17][18][19].
Neurotransmitter acetylcholine (ACh) is mainly synthetized from acetyl coenzyme A and choline by choline acetyltransferase. ACh modulates many vital functions and its effects are exerted through muscarinic and nicotinic cholinergic receptors. Action of ACh on synapses is terminated via hydrolytic cleavage catalyzed by either of two structurally similar enzymes-acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE). Drugs boosting ACh neurotransmission have been reported to have a beneficial effect on some types of dementia including Alzheimer's disease, glaucoma, myasthenia gravis, atonia, and tachycardia. They have been used in prophylaxis of nerve paralytic agents. Among various mechanisms of action, AChE and BuChE inhibitors are of a special interest [20].
To our best knowledge, no systematic study evaluating antimicrobial activity of hydrazones obtained from halogenated salicylaldehydes and AG or its isosteres and derivatives has been published. As a part of our ongoing research, we designed, prepared, and screened them as potential antimicrobial agents primarily; then, they were repurposed as inhibitors of acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) and we evaluated toxicity for eukaryotic cell line. In addition, we also investigated hydrazones of 5-nitro five-membered ring heterocycles-2-carbaldehydes systematically in a similar way as a result of a structure-based design strategy.
After biological assessment of this initial series, reflecting structure-activity relationships that were found, we prepared additional derivatives. 5-Nitrofurylidene analogues with modified "amino part" were synthesized with respect to the antibacterial and almost selective AChE inhibition produced by 1n (Figure 2), while me-too approach of 1m initiated by its inhibitory properties for both AChE and BuChE covers both aldehyde (O-acetylation: 2a, 4-OH isomer: 2b, OH group removal: 2c, double bond reduction: 2d) and aminoguanidine substructures (Figure 3). Modification of AG consists in introduction of NO 2 group (nitroaminoguanidines), second NH 2 group enabling attachment of an additional aldehyde, and formally isosteric replacement of =NH group by =O (semicarbazones) and =S (thiosemicarbazones).    Hydrazones were prepared by treatment of amino compound (aminoguanidine hydrochloride, nitroaminoguanidine, 1,3-diaminoguanidine hydrochloride, thiosemicarbazide, or semicarbazide hydrochloride; 1 mmol) with a mild excess of appropriate aldehyde or ketone (1.1 of equivalents or 2.2 of equivalents for disubstituted 1,3-diaminoguanidines 2g and 3c) in methanol under reflux for 3 h (Figures 4 and 5). In some cases, the crystallization was initiated by addition of 5% sodium bicarbonate. Due to high solubility of 3a, its synthesis was realized in water under reflux for 6 h and the solution was basified using solid sodium carbonate. The yields ranged from 65% to 99%. Lower yields were obtained for benzylidene derivative 1a (78%) due to higher water solubility, lipophilic ester 2a (74%), 3,5-diiodobenzylidene analogue 2c (80%), and the only sole ketone (isatin) derivative 1p (65%). The derivatives of 1,3-diaminoguanidine were isolated as hydrochlorides without basification (2e, 2g, and 3c) with yields of 50-93%. For the reaction providing monosubstituted 1,3-diaminoderivative 2e, only 0.95 of equivalents of the aldehyde was used.
The analogue 2d without double bond was prepared from aminoguanidine hydrochloride (1 mmol) dissolved in MeOH that was mixed successively with 3,5-diiodosalicylaldehyde (1.1 mmol) and sodium cyanoborohydride (1.6 mmol) ( Figure 6). After 48 h at rt, the reaction mixture was diluted with 5% sodium bicarbonate and let to stir over night. The product was isolated with yield of 82%.
Regarding compounds 1-3, Gram-positives share a higher susceptibility w from 7.8 µM, while Gram-negative pathogens were significantly more resistant w of ≥62.5 µM. The genus Staphylococcus was the most sensitive, importantly, with nificant differences in susceptibility between MSSA and MRSA strains. Of Gram-n The products were characterized by 1 H and 13 C NMR spectroscopy, IR spectra and melting points, the purity was checked additionally by elemental analysis. From total thirty-one compounds, fifteen were original (1e, 1h, 1i, 1k-1o, 2a-2g). NMR spectra of the representative compounds are illustrated in supplementary material (Figures S1-S10).  Table 1. Piperacillin and parent amino compounds (including AG, 1,3-diaminoguanidine, nitroaminoguanidine, semicarbazide, and thiosemicarbazide) were involved for activity comparison. Two compounds (1c, 1d) were not evaluated due to solubility problems. The parent amino compounds showed all MIC values above 500 µM (not involved in Table 1), so the activity strongly depends on aldehyde/ketone-derived part of the molecules. Ten target compounds avoided antibacterial properties, i.e., their MIC values were ≥500 µM (1d, 1e, 1m, 1p, 2d, 2f, 2g, 3a-3c; not included in Table 1). Beta-lactam drug piperacillin showed higher activity against MSSA, E. faecalis, E. coli as well as P. aeruginosa, but not against S. epidermidis, K. pneumoniae and, of course, MRSA.

Microbiology
Regarding compounds 1-3, Gram-positives share a higher susceptibility with MIC from 7.8 µM, while Gram-negative pathogens were significantly more resistant with MIC of ≥62.5 µM. The genus Staphylococcus was the most sensitive, importantly, without significant differences in susceptibility between MSSA and MRSA strains. Of Gramnegative species, P. aeruginosa and A. baumannii were more resistant (MIC from 125 µM).
When focused on fungal species, L. corymbifera and both strains of Aspergillus showed a complete resistance to the investigated molecules (data not shown). Contrarily, T. interdigitale as the most susceptible strain was supressed by fifteen compounds from the concentration of 62.5 µM with 1i, 1l, 2a, 2d, and 2e superiority. Their MIC values are comparable to fluconazole; however, for fluconazole it means 50% inhibition of growth, while for our compounds a complete inhibition. Thus, even with similar numeral values, the derivatives are more potent.

Inhibition of Acetyl-and Butyrylcholinesterase
After outcomes of antimicrobial activity, this follow-up screening was inspired by previous sporadic publication reporting compounds based on AG scaffold as potential AChE [27][28][29] and BuChE [29,30] inhibitors. Additionally, hydrazones [31,32] as well as salicylic compounds [31,33] have shown inhibition of these enzymes. The derivatives 1-3 together with their parent compounds were tested for their ability to suppress the function of AChE from electric eel (EeAChE) and BuChE from equine serum (EqBuChE) using Ellman's method (Table 3). Inhibitor potency is expressed as the concentration resulting in 50% inhibition of enzymatic activity (IC 50 ). Selectivity indexes (SI) were calculated to quantify the preferential affinity for AChE or BuChE. SI is defined as the ratio of IC 50 for AChE/IC 50 for BuChE (Table 3). Values above 1 indicate stronger inhibition of BuChE and vice versa. Clinically used cholinesterases (ChEs) inhibiting drugs galantamine and rivastigmine were included as standards.
Most of the derivatives 1-3 exhibited a dual inhibition of both ChEs, nitro group containing derivatives 2f and 3b-3e being an exception with IC 50 for BuChE over 500 µM. AChE was inhibited more consistently with IC 50 values of 17.95-54.93 µM with isatin derivative 1p superiority. Substitution of benzylidene derivative 1a by chlorine alone or in a combination with 2-OH group did not improve activity generally (1b-1h, 1j). On the other hand, 5-iodosalicylidene 1i and 3,5-dihalogenosalicylidene scaffold, where at least one halogen is bromine or iodine, resulted in improved potency (1m, and especially 1k and 1l). The O-acetylation, OH group removal, insertion of NH 2 or NO 2 group into aminoguanidine part of the molecule as well as its replacement by semicarbazide are tolerated (1m vs. 2a, 2c, 2e, 2f, and 2h). Contrarily, double bond reduction (2d), OH group positional isomer 2b, introduction of second 3,5-diiodosalicylideneamino group (2g with the highest IC 50 of 54.93 µM) or AG replacement by thiosemicarbazide (2i) are detrimental. Among nitro heterocycle-based hydrazones, the best activity is associated with the presence of 5-nitrofurylidene moiety on (nitro)aminoguanidine (1n and 3b); the replacement by thiophene is acceptable (1n vs. 1o), while removal of NO 2 as well as duplication of 5nitrofurylideneamino moiety decreased activity (1n vs. 3a and 3c). The substitution of benzylidene derivatives 1a resulted in more potent inhibitors in the case of iodine (1i, 1l, 1m, 2a, and 2c) and bromine (1k), mainly together with 2-phenolic group. Interestingly, the isomeric chlorobenzaldehyde conjugates 1b-1d exhibited similar inhibition.
The selectivity for each particular ChEs can be described on the basis of selectivity indexes. Seven compounds inhibited BuChE significantly more effectively (SI >3: 1d, 1e, 1l, 1m, 2a, 2g, and above all 2e with SI = 16.7). On the other hand, five hydrazones are highly selective AChE inhibitors (SI ≤ 0.1: 2f, and 3b-3e), closely followed by remaining nitro compounds (1n, 1o). Thus, presence of NO 2 group in either arylidene or aminoguanidine portion results in selective AChE inhibition, which is demonstrated illustratively by pair 2e and 2f, where replacement of NO 2 with NH 2 led to more than 239.5fold increased selectivity for BuChE. This modification substantially alters both electronic properties and lipophilicity. Other clear structural features modulating SI are presence of 2-chlorobenzylidene (1b, 1h) leading to higher selectivity for AChE compared to its isomers. Focusing on 3,5-halogenosalicylidenes, heavier halogens (1j-1m) turn selectivity to BuChE. Modification of 3,5-diiodosalicylidene compound 1m provided mainly compounds with reduced BuChE selectivity (phenolic group removal, acetylation, hydrazone reduction, positional isomer, analogous nitroaminoguanidine, semicarbazone and thiosemicarbazone: 2c, 2a, 2d, 2b, 2f, 2h, and 2i, respectively), only the homologue with additional amine group 2e was more selective. Thus, based on the substitution, it is possible to modulate selectivity for both ChEs.
Notably, all precursor amino compounds significantly and selectively inhibited AChE with IC 50 values of 41.71-75.14 µM; AG was least effective. BuChE was inhibited to a lesser extent (from 283.32 µM for 1,3-diaminoguanidine, the most potent precursor ChEs inhibitor), while introduction of nitro group resulted in a complete loss of activity (IC 50 >500 µM). Importantly, their modification to form hydrazones led to substantially better activity against BuChE, by up to two orders of magnitude; nitroaminoguanidines 2f and 3b were exceptions reflecting the potency of this scaffold along with some 5-nitrofurylidene analogues (3c-3e). Conversion of AG, nitroaminoguanidine and semicarbazide to their hydrazones is also advantageous for uniformly improved AChE inhibition (up to 4.2-fold for 1p), whereas results for thiosemicarbazide and 1,3-diaminoguanidine are ambiguous.
These results qualify the presented compounds to be considered as potential agents for treatment of neurodegenerative disorders, myasthenia gravis, and other diseases where boosting of cholinergic neurotransmission is an option.

Type of Inhibition
Based on promising inhibition data, we evaluated type of inhibition of the most active hydrazones. In general, the reversible enzyme inhibitors can be classified as competitive, non-competitive, uncompetitive, or mixed type. The type of inhibition could be distinguished using the Lineweaver-Burk plot [34] and corresponding comparison of two kinetic parameters: maximum velocity (V m ) and Michaelis constant (K M ) of inhibited and uninhibited reactions. Based on changes of these parameters and intercept of lines in Lineweaver-Burk plot, type of inhibition is revealed.
The Lineweaver-Burk plots obtained for the most active dual inhibitors 1p, 2a and 2e and significantly stronger AChE inhibitor 1n are depicted in Figures 8-14. Based on the plots, all these derivatives cause mixed inhibition independently on substitution patterns (5-nitrofurylidene 1n, isatin 1p, 3,5-diiodosalicylidene 2a and 2e). It is supported with changed K M and V m and also their ratios when compared to uninhibited reactions. There are intercepts of the lines in quadrant I, II or III in the Lineweaver-Burk plot, but not on an axis.

Type of Inhibition
Based on promising inhibition data, we evaluated type of inhibition of the most active hydrazones. In general, the reversible enzyme inhibitors can be classified as competitive, non-competitive, uncompetitive, or mixed type. The type of inhibition could be distinguished using the Lineweaver-Burk plot [34] and corresponding comparison of two kinetic parameters: maximum velocity (Vm) and Michaelis constant (KM) of inhibited and uninhibited reactions. Based on changes of these parameters and intercept of lines in Lineweaver-Burk plot, type of inhibition is revealed.
The Lineweaver-Burk plots obtained for the most active dual inhibitors 1p, 2a and 2e and significantly stronger AChE inhibitor 1n are depicted in Figures 8-14. Based on the plots, all these derivatives cause mixed inhibition independently on substitution patterns (5-nitrofurylidene 1n, isatin 1p, 3,5-diiodosalicylidene 2a and 2e). It is supported with changed KM and Vm and also their ratios when compared to uninhibited reactions. There are intercepts of the lines in quadrant I, II or III in the Lineweaver-Burk plot, but not on an axis.

Molecular Docking Study
The molecular modelling study was performed in order to estimate possible binding modes of the prepared inhibitors with ChE. It has proved to be a useful tool to elucidate the interactions between inhibitors and enzymes.
The active site of AChE was studied intensively in the past and is currently very thoroughly described. It consists of several sub-sites playing role in the stabilization of the substrate in the enzyme cavity (anionic site, Trp86, Tyr337, Phe338), contributing to the substrate specificity (oxyanion hole, acyl pocket) and securing the hydrolysis of the sub-

Molecular Docking Study
The molecular modelling study was performed in order to estimate possible binding modes of the prepared inhibitors with ChE. It has proved to be a useful tool to elucidate the interactions between inhibitors and enzymes.
The active site of AChE was studied intensively in the past and is currently very thoroughly described. It consists of several sub-sites playing role in the stabilization of the substrate in the enzyme cavity (anionic site, Trp86, Tyr337, Phe338), contributing to the substrate specificity (oxyanion hole, acyl pocket) and securing the hydrolysis of the substrate in the first place (esteratic site, consisting of catalytic triad Ser203, His447 and Glu334) [35]. The active site is located at the bottom of strikingly narrow and deep cavity which penetrates halfway into the enzyme. At the rim of the gorge several amino acid residues (Tyr72, Asp74, Tyr124, Trp286 and Tyr341) form the so-called peripheral anionic site (PAS). Binding of the ligand to PAS can modulate the activity of AChE through conformational changes in the active site [36]. The overall structure of BuChE is very similar to that of AChE and similarly to AChE, the active site is located at the bottom of the gorge. However, a considerable amount of aromatic residues present in active center of AChE are replaced by hydrophobic residues in BuChE, thus allowing more voluminous substrates to enter [37].
The binding energies of particular ligands and enzymes are summarized in Table 4. Compound 1p, being the most potent inhibitor of AChE in the series, was chosen as the representative for the binding mode analyses. The top-scored docking pose was observed stretched between the intrinsic anionic site and PAS. The conformation of ligand was stabilized by a number of hydrogen bond interactions (with residues Tyr72, Asp74, Asn87, Tyr124, Ser125, Tyr37) and π-π stacking (Trp86) (Figure 15). Compounds 1m, 2a, 2e, and 3b showed very similar orientation in the cavity interacting preferably with the amino acid residues of the PAS (namely Asp74, Tyr124, Trp286, Tyr341 and Tyr337; illustrated for 3b- Figure 16). amino acid residues of the PAS (namely Asp74, Tyr124, Trp286, Tyr341 and Tyr337; illustrated for 3b- Figure 16).    Concerning the interaction with BuChE, compounds 1m, 1p, 2a, and 2e show similar conformation near the entrance of the active gorge. Their position was s by several hydrogen bonds and π-π stacking (with residues Gly121, Thr120, Gln6 and Trp82). These interactions are demonstrated for 2e ( Figure 17).  Concerning the interaction with BuChE, compounds 1m, 1p, 2a, and 2e showed very similar conformation near the entrance of the active gorge. Their position was stabilized by several hydrogen bonds and π-π stacking (with residues Gly121, Thr120, Gln67, Asn68 and Trp82). These interactions are demonstrated for 2e ( Figure 17).  Concerning the interaction with BuChE, compounds 1m, 1p, 2a, and 2e show similar conformation near the entrance of the active gorge. Their position was st by several hydrogen bonds and π-π stacking (with residues Gly121, Thr120, Gln67 and Trp82). These interactions are demonstrated for 2e ( Figure 17).  Compound 3b differs significantly from the other ligands in its position in the enzyme (Figure 18), which corresponds with poor in vitro results. This ligand was located deep within the gorge near the active site Ser198. Few possible hydrogen bonds stabilized the ligand in its pose (Gly115, Gly116, Tyr128, Ser198).
Compound 3b differs significantly from the other ligands in its position in zyme (Figure 18), which corresponds with poor in vitro results. This ligand was deep within the gorge near the active site Ser198. Few possible hydrogen bonds st the ligand in its pose (Gly115, Gly116, Tyr128, Ser198).

Cytotoxicity
Cytotoxicity of the tested compounds was measured using the standard hep line HepG2. The used CellTiter 96 ® assay is based on the reduction of tetrazoli (MTS) in living cells to formazan, which is then determined colorimetrically. Th tion of the reagent is related to availability of NADH or NADPH. The decline in these metabolically important compounds in the cell causes that the production azan is reduced.
The parameter IC50 was used as a measure of cytotoxicity, which allows the tative comparison of the toxicity among tested compounds (Table 5). For toxicity tion, we chose parent guanidine derivatives and the most promising antibacterial 2c, 2e, and 3e) and antifungal agents (2e), the most potent inhibitors of both ChE 2a, and 2e) or only AChE (1n, and 2c).
Parent amino compounds (aminoguanidine, 1,3-diaminoguanidine, nitr guanidine, semicarbazide and thiosemicarbazide; data not shown) as well as th derivative 1p showed no significant cytotoxicity in the range of tested concentratio (up to 500 µM). Other compounds can be divided into four subgroups accordin determined IC50 values. Three compounds showed significant toxicity with IC50 b µM, thus being more toxic than anticancer drug tamoxifen: 1n, 1o, and 2c. The group exhibited moderate toxicity with IC50 higher than 29 µM and those obta tamoxifen: 3e, 2a, and 1m. The third group containing the hydrazone 2e showed tively low toxicity with IC50 above 100 µM. These values are comparable or sup term of selectivity to previously reported aminoguanidine-based aromatic hyd [22]. The most toxic compounds exhibited similar IC50 values as AG-based benzo hydrazones reported as apoptosis-promoting anticancer agents [15].

Cytotoxicity
Cytotoxicity of the tested compounds was measured using the standard hepatic cell line HepG2. The used CellTiter 96 ® assay is based on the reduction of tetrazolium dye (MTS) in living cells to formazan, which is then determined colorimetrically. The reduction of the reagent is related to availability of NADH or NADPH. The decline in levels of these metabolically important compounds in the cell causes that the production of formazan is reduced.
The parameter IC 50 was used as a measure of cytotoxicity, which allows the quantitative comparison of the toxicity among tested compounds (Table 5). For toxicity evaluation, we chose parent guanidine derivatives and the most promising antibacterial (1n, 1o, 2c, 2e, and 3e) and antifungal agents (2e), the most potent inhibitors of both ChE (1m, 1p, 2a, and 2e) or only AChE (1n, and 2c). Parent amino compounds (aminoguanidine, 1,3-diaminoguanidine, nitroaminoguanidine, semicarbazide and thiosemicarbazide; data not shown) as well as the isatin derivative 1p showed no significant cytotoxicity in the range of tested concentrations used (up to 500 µM). Other compounds can be divided into four subgroups according to the determined IC 50 values. Three compounds showed significant toxicity with IC 50 below 11 µM, thus being more toxic than anticancer drug tamoxifen: 1n, 1o, and 2c. The second group exhibited moderate toxicity with IC 50 higher than 29 µM and those obtained for tamoxifen: 3e, 2a, and 1m. The third group containing the hydrazone 2e showed a relatively low toxicity with IC 50 above 100 µM. These values are comparable or superior in term of selectivity to previously reported aminoguanidine-based aromatic hydrazones [22]. The most toxic compounds exhibited similar IC 50 values as AG-based benzothiazole hydrazones reported as apoptosis-promoting anticancer agents [15].
Thus, cytotoxicity is not a uniform class effect and based on aldehyde/ketone and amino compound, it is possible to modify toxic action.
To evaluate potential effect on cell membranes, we also examined identical derivatives (1m, 1n, 1o, 1p, 2a, 2c, 2e, and 3e) using lactate dehydrogenase (LDH) release assay (CytoTox 96 ® ) in HepG2 cells. It is a highly sensitive test capable of detecting early, low level damage of cell membranes. LDH is a stable cytosolic enzyme released into medium upon cell lysis. The release time from the cells to the surrounding medium is approximately 9 h. The amount of LDH released in the culture supernatants after 24 h of incubation is quantified by a combined enzyme assay that results in the conversion of the tetrazolium salt (iodonitrotetrazolium violet) to a red formazan product. This color product is determined spectrophotometrically, and its amount is proportional to the number of cells lysed, i.e., with damaged membrane integrity. EC 50 (half-maximal effective concentration) was used as a measure of cytotoxicity, which allows a quantitative comparison between the tested compounds. The EC 50 of the graduated dose response curve therefore represents the concentration of compound where 50% of its maximal effect is observed (Table 6). For all tested substances, the applied toxicity parameter was found in the concentration range from 0 to 500 µM. The compounds can be divided into subgroups according to their EC 50 values. Four compounds showed significant toxicity/membrane damaging effect with EC 50 below 6 µM: 1n, 1o, 2c, and 3e. The second group showed milder toxicity with EC 50 greater than 20 µM (1m and 2a). The third group containing two substances (1p, 2e) showed comparatively lower toxicity with EC 50 above 100 µM. As well as toxicity in the CellTiter 96 assay, LDH release assay also showed specific structure-dependent effect on membrane integrity that was more significant for compounds bearing 5-nitro group (1n, 1o, 3e).

General
All reagents and solvents were purchased from Merck (Darmstadt, Germany), Penta Chemicals (Prague, Czech Republic), Avantor (Stříbrná Skalice, Czech Republic), and Lach-Ner (Neratovice, Czech Republic). They were used as received without any purification. The reactions and the purity of the products were monitored by thin-layer chromatography (TLC) using a mixture with a ratio of dichloromethane to methanol (MeOH) and n-hexane of 4:1:1 (v/v/v) as the eluent. Plates were coated with 0.2 mm Merck 60 F254 silica gel (Merck Millipore, Darmstadt, Germany) and were visualized by UV irradiation (254 nm). The melting points were determined on a Melting Point B-560 apparatus (BÜCHI, Flawil, Switzerland) using open capillaries. The reported values are uncorrected. Infrared spectra were recorded on a FT-IR spectrometer using ATR-Ge method (Nicolet 6700 FT-IR, Thermo Fisher Scientific, Waltham, MA, USA) in the range 600-4000 cm −1 . The NMR spectra were measured in dimethyl sulfoxide (DMSO)-d 6 or deuterated chloroform (CDCl 3 ) using a Varian V NMR S500 instrument (500 MHz for 1 H and 126 MHz for 13 C; Varian Corp., Palo Alto, CA, USA) or a JNM-ECZ 600R (600 MHz for 1 H and 151 MHz for 13 C; JEOL, Tokio, Japan). The chemical shifts δ are given in ppm and were referred indirectly to tetramethylsilane via signals of the solvents. The coupling constants (J) are reported in Hz. Elemental analysis (C, H, N) was performed on an automatic microanalyser Vario MICRO Cube Element Analyzer (Elementar Analysensysteme, Hanau, Germany). Both calculated and found values are given as percentages.
The identity of the known compounds was established using NMR ( 1 H and 13 C) and IR spectroscopy. The characterizations are consistent with the literature data. The purity was checked by melting points measurement and elemental analysis. The compounds were considered pure if they agree within ± 0.4% with theoretical values.

Synthesis of Precursors
Nitroaminoguanidine (N-nitrohydrazinecarboximidamide) was synthesized according to previously reported procedure [39]. Briefly, nitroguanidine (10 mmol) was added into 25 mL of hot water and then, hydrazine hydrate (12 mmol) was added dropwise under vigorous stirring. The reaction mixture was heated at 60 • C. After 2 h at this temperature, the reaction mixture was cooled down, acidified with concentrated hydrochloric acid and stored for 24 h at +4 • C. The resulted precipitate was filtered off, washed thoroughly by cold water and dried. 2-Formyl-4,6-diiodophenyl acetate was synthesized as follows. 3,5-Diiodosalicylaldehyde (1 mmol) was dissolved in dry dichloromethane (7 mL) followed by addition of triethylamine (1.5 mmol) and after cooling to 0 • C, acetyl chloride (1.2 mol) was added in one portion under vigorous stirring. The mixture was to let stir without external cooling for one hour (based on TLC indication of a complete reaction; mobile phase: n-hexane: ethyl acetate 4:1, v/v). The solvent was evaporated under reduced pressure and using ethyl acetate and water, the residue was transferred into a separation funnel. The organic layer was washed with water, 10% aqueous sodium carbonate, 0.1 M aqueous hydrochloric acid, and saturated brine. The organic phase was dried over anhydrous sodium sulphate. The filtrate was concentrated under reduced pressure and n-hexane was added to start precipitation. After 24 h at +4 • C, the suspension was filtered off to give the ester.

Synthesis of Targeted Derivatives 1-3
Hydrazones: Aminoguanidine hydrochloride (or nitroaminoguanidine, 1,3-diaminoguanidine hydrochloride, thiosemicarbazide, or semicarbazide hydrochloride; 1 mmol) was dissolved in 8-10 mL of MeOH and then an appropriate aldehyde or ketone (1.1 mmol or 2.2 mmol for the synthesis of disubstituted 1,3-diaminoguanidines 2g and 3c) was added in one portion. The reaction mixture was heated under reflux for 3 h and then let to stir to cool down to room temperature overnight. Then, the reaction mixture was stored at −20 • C for 2 h. If there was present a precipitate, it was collected by filtration. The isolated crystals were suspended in 5% sodium bicarbonate and stirred for 30 min. Then, the suspension was filtered off, the crystals were washed by cold water, diethyl ether and dried. When no spontaneous precipitation was observed, the methanolic solution was diluted by 5% sodium bicarbonate, stored for 1 h at +4 • C. In some cases, the reaction mixture was partly evaporated after basification (e. g., 2a and 3a). The resulting precipitate was collected by filtration, washed by cold water, diethyl ether and dried. The products were crystallized from methanol if necessary.
The imine 3a was prepared in water under reflux for 6 h and the solution was basified using solid sodium carbonate.
Synthesis of the reduced analogue 2d: Aminoguanidine hydrochloride (1 mmol) was dissolved in 10 mL of methanol and then 3,5-diiodosalicylaldehyde (1.1 mmol, 411.3 mg) was added in one portion. After a complete dissolution, sodium cyanoborohydride (100 mg; 1.6 mmol) was added. The mixture was stirred for 48 h at room temperature. Then, the reaction mixture was diluted with 5% sodium bicarbonate and let to stir over night at room temperature. The precipitate was filtered off, washed by water, and dried to obtain pure 2d.
(E)-2-Benzylidenehydrazine-1-carboximidamide 1a. White solid; yield 78%; mp 178-181 • C (180-182 • C [13] The microdilution broth method was performed according to EUCAST (The European Committee on Antimicrobial Susceptibility Testing) instructions [48] with minor modifications at the concentration range of investigated compounds being 0.49-500 µM depending on their solubility; for several derivatives, the highest possible concentration was 125 µM. Broad-spectrum beta-lactam antibiotic piperacillin (PIP) was employed as a reference compound. The results of PIP were read after 24 h. Standard antibiotics (ciprofloxacin and gentamicin; data not shown) and internal quality control/reference strains are routinely included in parallel testing-basic screening of antibacterial activity.

Antifungal Activity
Antifungal activity was evaluated against four yeasts, namely: Candida albicans ATCC 24443, CCM 8320, Candida krusei ATCC 6258, CCM 8271, Candida parapsilosis ATCC 22019, CCM 8260, Candida tropicalis ATCC 750, CCM 8264; and four strains of filamentous fungi, namely: Aspergillus fumigatus ATCC 204305, Aspergillus flavus CCM 8363, Lichtheimia corymbifera CCM 8077, and Trichophyton interdigitale ATCC 9533, CCM 8377. A microdilution broth method was performed according to EUCAST instructions [49,50] with minor modification. Briefly, tested compounds were dissolved in DMSO and diluted in a twofold manner with RPMI-1640 medium with L-glutamine, supplemented with 2% glucose (w/v) and buffered to pH 7.0 with 3-(N-morpholino)propane-1-sulfonic acid (all components were purchased from Sigma-Aldrich, St. Louis, MO, USA). The final concentration of DMSO in the tested medium did not exceed 1% (v/v) of the total solution composition, and it was confirmed that this concentration did not inhibit the fungal growth. Static incubation was performed in dark and humidified atmosphere, at 35 ± 2 • C for 24 and 48 h (72 and 120 h for Trichophyton interdigitale). Positive controls consisted of test fungus solely, while negative controls consisted of medium and DMSO. Internal quality control was included too. Standard antimycotic drugs (amphotericin B and voriconazole) such as internal quality controls were included in assays, too (data not shown). Visual inspection was used for MIC endpoints evaluation. The experiments were conducted in duplicates. For the results to be valid, the difference in MIC determined from two parallel measurements must not be greater than one step on the dilution scale. MIC was determined by naked eye in the well with the lowest drug concentration, where no visible growth of microbial agent was detected. MIC determination scale was identical as for antibacterial activity.
Parent five hydrazine derivatives and antifungal triazole drug fluconazole were involved as reference compounds. Antifungal activities of fluconazole are presented as mean 50% growth inhibition. Results were read after 24 h (yeasts) or 48 h (moulds) microdilution plates cultivation without agitation at 35 ± 2 • C in humidified atmosphere. The results were read with a microdilution plate reader (SynergyTM HTX, BioTek Instruments, Inc., VT, USA) at wavelength 530 nm.

Evaluation of Acetyl-and Butyrylcholinesterase Inhibition
The activities of the compounds 1-3 were measured via the Ellman's spectrophotometric method that was modified according to Zdražilová et al. [51]. The activity of cholinesterase enzymes is assessed indirectly by quantifying the amount of 5-mercapto-2-nitrobenzoic acid ion formed in the reaction of the thiol reagent 5,5 -disulfanediylbis(2nitrobenzoic acid) (DTNB) with thiocholine, a product of the hydrolysis of acetylthiocholine (ATCh) or butyrylthiocholine (BTCh) catalysed by AChE and BuChE, respectively.
AChE was obtained from electric eel (Electrophorus electricus L.) and BuChE from equine serum. The benzazepine alkaloid galantamine and carbamate drug rivastigmine were used as reference enzyme inhibitors. These compounds are approved for treatment of Alzheimer's disease. All the enzymes, the thiol reagent, ATCh, BTCh, rivastigmine and galantamine were purchased from Sigma-Aldrich.
The enzyme activity in final reaction mixture (2 mL) was 0.2 U/mL, the concentration of ATCh or BTCh 40 µM and the concentration of DTNB 0.1 mM for all reactions. The investigated hydrazide-hydrazones were dissolved in DMSO and then diluted by demineralized water (conductivity 3 µS) to the concentration of 1 mM. For all of them as well as galantamine and rivastigmine five different concentrations of inhibitor in final reaction mixture were used. Final DMSO concentration in the reaction mixture was 0.3%. All measurements were carried in triplicates and the average values of reaction rate (v 0 -uninhibited reaction, v i -inhibited reaction) were used for construction of the dependence v 0 /v i vs. concentration of inhibitor. From obtained equation of regression curve, IC 50 values were calculated.
For the determination of the type of inhibition of the selected inhibitors (1n, 1m, 1p, 2a, and 2e), Lineweaver-Burk plot was used, and the measuring procedure was analogous to the determination of IC 50 . The enzyme activity in the final reaction mixture (2 mL) was 0.2 U/mL, a concentration of ATCh and BTCh 20-80 µM and a concentration of DTNB was 0.1 mM. For each of the substrate concentrations, four different concentrations of the inhibitor were used (0.5, 1, 1.5, and 2 µM for AChE, and predominantly 1, 2, 3 and 4 µM for BuChE). The dependence absorbance vs. time was observed and the reaction rate was calculated. All measurements were carried in duplicates and the average values of reaction rate were used for the construction of the Lineweaver-Burk plot. From obtained equations of regression curves, the values of Michaelis constant (K M ) and maximum velocity (V m ) were calculated, and the type of inhibition was identified.

Molecular Docking
Crystallographic structures of human AChE and human BChE were obtained from protein data bank (www.rcsb.org, accessed on 26 January 2021; pdb codes 4PQE and 1POI, respectively). The 3D structures of ligands were prepared in Chem3D Pro 19.1 (ChemBioOffice 2019 Package, CambridgeSoft, Cambridge, MA, USA). In the preparation process all water molecules were removed from the enzymes and structures of enzymes and ligands were optimized using UCSF Chimera software package (Amber force field) [52]. Docking was performed using Autodock Vina [53] and Autodock 4.

Cytotoxicity Determination
The human hepatocellular liver carcinoma cell line HepG2 (passage 6-8 for HepG2 cells, 20-23 for LDH assay) purchased from Health Protection Agency Culture Collections (ECACC, Salisbury, UK) was cultured in minimum essentials eagle medium (MEM) supplemented with 10% fetal bovine serum, 1% L-glutamine solution and non-essential amino acid solution in a humidified atmosphere containing 5% CO 2 at 37 • C. For subculturing, the cells were harvested after trypsin/EDTA treatment at 37 • C. All used chemicals were obtained from Sigma-Aldrich (St. Louis, MO, USA). To evaluate cytotoxicity of the compounds, the cells treated with the tested substances were used as experimental groups whereas untreated cells served as control groups. The range of concentrations tested were 1-500 µM (1m, 1o, 1p, 2c, 2e, 3e, and tamoxifen) or 1-250 µM for 1n and 2a.
The cells were seeded in a density of 15,000 cells per well in a 96-well plate. The next day the cells were treated with each of the tested substances at a broad range of concentrations (based on their solubility) in triplicates. The compounds were dissolved in DMSO (maximum incubation concentration of DMSO was 1% v/v). The controls representing 100% cell viability, 0% cell viability (treated with 10% DMSO), no cell control and vehiculum control were incubated in parallel also in triplicates. After 24 h of incubation in a humidified atmosphere containing 5% CO 2 at 37 • C, the reagent from the kit CellTiter 96 AQueous One Solution Cell Proliferation Assay ® (CellTiter 96; PROMEGA, Fitchburg, WI, USA) was added. After 2 h of incubation at 37 • C, absorbance of samples was recorded at 490 nm (TECAN, Infinita M200, Grödig, Austria). A standard toxicological parameter IC 50 was calculated by nonlinear regression from a semilogarithmic plot of incubation concentration versus percentage of viability relative to untreated controls using GraphPad Prism 8 software (GraphPad Software, Inc., La Jolla, CA, USA).
Results of the experiments are presented as inhibitory concentrations that reduce viability of the cell population to 50% from the maximal viability (IC 50 ). Known cytotoxic drug tamoxifen (as citrate salt) was involved as a reference compound.
For evaluating effect on membrane integrity (LDH release), the CytoTox 96 ® Non-Radioactive Cytotoxicity Assay (PROMEGA, Fitchburg, WI, USA) was performed. The range of concentrations tested was 1-500 µM. After 24 h of incubation at 37 • C, 50 µL of medium was removed and 50 µL of reagent was added, and then after 30 min the reaction was stopped and the absorbance of the samples at 490 nm was recorded (TECAN). The standard toxicological parameter EC 50 was calculated by non-linear regression from a semi-logarithmic plot of incubation concentration versus percent absorbance relative to untreated controls using GraphPad Prism 9 software (GraphPad Software, Inc., La Jolla, CA, USA).

Conclusions
Based on the previous reports of especially robenidine analogues [9,22] and our experience with halogenated salicylidene derivatives, we designed and synthesized their analogues derived from chlorobenzaldehydes, halogenated salicylaldehydes, 5-nitrofurfural, isatin, plus related aldehydes and AG, 1,3-diaminoguanidine, nitroaminoguanidine, and (thio)semicarbazone as potential bioactive compounds. In summary, we have prepared thirty-one compounds, in most cases hydrazones, fifteen of which have not yet been reported previously. Some of the known derivatives have not yet been biologically evaluated.
Initially, we were focused on antimicrobial evaluation against Gram-positive and Gram-negative bacteria, yeasts, and filamentous fungi. The compounds showed broadspectrum of antibacterial activity with MIC from 7.81 µM; Gram-positive strains including MRSA were more susceptible. The best activity was associated with 3,5-diiodobenzylidene, 5-nitrofurylidene, and (5-nitrothiophen-2-yl)methylidene scaffolds. However, results were beyond our expectations based on previous reports. Antifungal activity was rather mild with MIC from 62.5 µM; 3,5-diiodosalicylidene scaffold was the optimal choice.
Then, we evaluated inhibition of AChE and BuChE. The compounds have been identified as potent to moderate inhibitors of both AChE and BuChE with a variable selectivity depending on substitution pattern. 5-Nitrofurylidene led to selective inhibition of AChE, while N'-(2-hydroxy-3,5-diiodobenzylidene)hydrazinecarboximidhydrazide was the most potent and also selective BuChE inhibitor. Isatin-AG condensate was the most active against AChE along with identical IC 50 for of BuChE. Having these results in hand, we evaluated type of inhibition of the most active compounds with the conclusion that they are mixed-type inhibitors. Molecular docking provided insight into their interactions with the enzymes. The hydrazones bind closely to intrinsic anionic site and PAS of AChE, while in BuChE they are bound near the entrance to the active gorge.
Finally, the most promising hydrazones, along with their synthetic precursors, were screened for their toxicity in eukaryotic cell line. The compounds exhibited variable toxicity with IC 50 ranging from low micromolar concentration (lower than obtained for tamoxifen) to non-toxic derivatives.
Taken together, several derivatives can be considered as promising hits especially for inhibition of cholinesterases and antibacterial properties. The coincidence of antimicrobial and cytotoxic actions of some compounds may also be perspective. The biological profile is clearly dependent on both parent amino and carbonyl compounds; it is possible to achieve target activity and to separate them. However, prior to extensive hit-to-lead optimization campaign, a caution of addressing potential membrane-damaging effect indicated for some compounds by LDH release assay and pan-assay interference (PAINS)-like behavior derived from several structural motifs should be excluded.