Synthesis, thermal, spectral, antimicrobial and cytotoxicity profile of the Schiff bases bearing pyrazolone moiety and their Cu(II) complexes

A series of Schiff bases resulted in the [1 + 1] condensation of 8-alkyl-2-hydroxy-tricyclo[7.3.1.0^2.7]-tridecan-13-one with the 4-amino-2,3-dimethyl-1-phenyl-3-pyrazolin-5-one, and their complexes CuL(CH₃COO)₂·nH₂O were synthesized. The compounds were characterized by microanalytical, ESI-MS, NMR, IR, electronic and EPR spectra. Based on the ESI-MS and IR spectra, a mononuclear structure with both Schiff base and acetate as chelate was proposed for complexes. An elongated rhombic stereochemistry was assigned considering electronic and EPR data. The thermal analyses have evidenced processes as water elimination, acetate decomposition as well as oxidative degradation of the Schiff base moiety. The bioevaluation of compounds in relation to planktonic and biofilm-embedded microbial cells, as well as to human cells, indicates an improved activity of complexes over ligands. The same tendency was observed for antioxidant activity.

Keywords: 4-Aminoantipyrine; Copper complex; Biofilm; Cytotoxicity; Schiff base; Thermal behaviour

Due to their rich coordinative chemistry and various biological applications, Schiff bases are intensively studied. The coordination ability of multidentate ligands, especially those with different donor atoms, was exploited in order to improve target biological properties and to develop species with a different mechanism of action in comparison with organic bioactive species.

Among these derivatives, the multifunctional species based on nitrogen heterocycles as pyrazolone derivatives are known for their antipyretic, analgesic, antiinflammatory, antitumour and antidiabetes effects [1-10]. The mechanism of these effects involves nonselective inhibition of cyclooxygenase isozymes or the reactive oxygen and nitrogen species scavenging [10-12]. From this species, 4-amino-2,3-dimethyl-1-phenyl-3-pyrazolin-5-one (4-aminoantipyrine) is well known for its antiinflammatory effects [1, 8-10], but the assays evidenced several side effects such as leukopenia, gastrointestinal irritation, cardiovascular problems and renal injury [1-3].

In order to overcome these limitations and to extend the range of its biological applications, 4-aminoantipyrine was used in the synthesis of Schiff bases and complexes designed with such ligands demonstrated useful biological activities. Thus, Cu(II) complexes with Schiff bases obtained by the condensation of this scaffold with 2-hydroxybenzaldehyde/terephthalic aldehyde [13], Cu(II), VO(II), Ni(II) and Mn(II) compounds with the derivative resulted in condensation with 3-formyl-6-methyl-chromone [14], and Cu(II) and VO(IV) species with Schiff base obtained by condensation with 3-hydroxy-5-(hydroxymethyl)-2-methyl-4-pyridinecarboxaldehyde (pyridoxal) [15] exhibited an improved antibacterial activity in comparison with free ligands. Besides antimicrobial activity, a series of Co(II), Cu(II), Ni(II) and Zn(II) complexes with Schiff bases obtained by 4-aminoantipyrine condensation with cinnamaldehyde/benzaldehyde [16] or 3-(cinnamyl)-pentane-2,4-dione [17] exhibited the ability to intercalate DNA bases, proving an antiproliferative potential.

As for 2-hydroxy-8-alkyl-tricyclo[7.3.1.02.7]tridecyl scaffold, this was involved in synthesis of a series of hydrazones with antimicrobial and antitumour activities [18], while Ca(II) and Cu(II) complexes with crowned Schiff base (2-hydroxy-8-methyl-tricyclo[7.3.1.02.7]tridec-13-N-4′(benzo-15-crown-5-ether)-imine) were characterized as species with potent antimicrobial and antibiofilm activity [19]. For the last species, the thermal behaviour was also reported [19].

These results encouraged us to design new Schiff bases by assembling 4-aminoantipyrine with 2-hydroxy-8-alkyl-tricyclo[7.3.1.02.7] tridecyl moiety as well as synthesizing their Cu(II) complexes. Copper(II) was chosen for this purpose having in view its stereochemical versatility, redox ability [20] and involvement in physiological processes both for humans and for microorganisms [21], including pathogenic ones. Both free ligands and their Cu(II) complexes were characterized by a variety of physico-chemical techniques. Moreover, their biological properties were assayed on prokaryotic and eukaryotic, both microbial and human cells.

High-purity reagents were purchased from Sigma-Aldrich (Cu(CH3COO)2·H2O) and Merck (4-amino-2,3-dimethyl-1-phenyl-3-pyrazolin-5-one) and were used without further purification. TLC plates and solvents were achieved from Sigma-Aldrich, Merck and Chimopar and used as received.

Chemical analysis of carbon, nitrogen and hydrogen has been performed using a PerkinElmer PE 2400 analyser.

The molar conductance was determined for 10−3 mol dm−3 solutions of complexes in DMSO with a multi-parameter analyser CONSORT C861.

Mass spectra were recorded with Maxis Bruker 4G mass spectrometer with an electrospray ionization (ESI) source operating in positive mode. Samples were dissolved in DMSO at 1 mg mL−1, and then, the solution was diluted with methanol up to a final concentration of 1 μg mL−1. Molecular ions scanning range (m/z) was 50-1250.

1H-NMR and 13C-NMR spectra were recorded on a Varian Inova-300 spectrometer at selected temperatures, in deuterated solvent CDCl3, isotopic purity 99.9%. (The annotations of atoms in NMR spectra are presented in Table 1S.) Chemical shifts were measured in parts per million from internal standard TMS.

IR spectra were recorded in KBr pellets with a Bruker Tensor 37 spectrometer in the range 400-4000 cm−1.

Electronic spectra by diffuse reflectance technique, with spectralon as standard, were recorded in the range 200-1500 nm on a Jasco V670 spectrophotometer.

The X-band EPR measurements were recorded at 20°C for solid samples and at − 173°C for DMSO solutions on a JEOL FA100 spectrometer. The general settings used were as follows: sweep field 1000 G, frequency 100 kHz, gain in the range 100-200, sweep time 1800s, time constant 1s, modulation width 2 G, microwave power 1 mW. The magnetic field calibration was performed with a DPPH (2,2-diphenylpicrylhydrazyl) standard marker, exhibiting a narrow EPR line at g = 2.0036.

The heating curves (TG, DTG and DTA) were recorded using a Labsys 1200 SETARAM instrument with a sample mass of 6-17 mg over the temperature range of 20-900 °C, using a heating rate of 10°C min−1. The measurements were taken in synthetic air atmosphere (flow rate 17 cm3 min−1) by using alumina crucibles.

The X-ray powder diffraction patterns were collected on a DRON-3 diffractometer with a nickel-filtered Cu Kα radiation (λ = 1.5418 Å) in 2θ range of 5-70°, a step width of 0.05° and an acquisition time of 2 s/step.

The antimicrobial activity of the obtained complexes was tested against Gram-negative and Gram-positive bacterial and fungal reference strains (Escherichia coli ATCC 8739, Pseudomonas aeruginosa ATCC 27853, Staphylococcus aureus ATCC 6538, Bacillus subtilis ATCC 6633, Candida albicans ATCC 10231). The respective strains were streaked on tryptic soy agar and incubated for 15-18 h at 37 °C. From the obtained fresh cultures, microbial suspensions of 1.5 × 108 CFU mL−1 density (corresponding to 0.5 McFarland turbidimetric standard) were prepared in sterile saline. The tested compounds were suspended in DMSO to prepare a stock solution of 10 mg mL−1 concentration. The quantitative assay of the antimicrobial activity was performed by a liquid medium microdilution method in 96-multiwell plates. Briefly, binary dilutions ranging from 500 to 15.6 µg mL−1 were performed in a total volume of 200 µL Mueller-Hinton broth. The wells were then seeded with a fix volume of 50 µL of the microbial suspensions. Positive and negative controls (culture medium seeded with the microbial suspensions and sterile culture medium, respectively) were used in the experiment. The plates were incubated at 37 °C for 24 h, and then, the absorbance of the wells liquid content was measured at 600 nm (Apollo LB 911 ELISA plate reader). The minimal inhibitory concentration (MIC) was considered as the lowest concentration of the tested compound that induced an inhibition of the microbial growth corresponding to an absorbance value inferior to that of the positive control used in the experiment [22-25]. The efficiency of the obtained complexes was compared to that of standard antimicrobial substances, respectively, carbenicillin and fluconazole. For the evaluation of the influence of the tested suspensions on the ability of microbial strains to colonize the inert substratum, a microtiter plate method was used. In this purpose, the microplates used for the MIC assay were emptied and washed three times by phosphate-buffered saline. The biofilm formed on the plastic wells wall was fixed for 5 min with cold methanol, coloured for 15 min by violet crystal solution and then re-suspended in a 33% acetic acid solution. The optical density of the coloured suspension was quantified by measuring the absorbance at 490 nm. The minimal biofilm eradication concentration (MBEC) values were considered as the lowest concentration of the tested compound that induced an inhibition of microbial adherence and biofilm development on the plastic wells, revealed by a decrease in the absorbance value in comparison with the positive control used in the experiment [26-28].

In order to evaluate the cytotoxic effects of the complexes in correlation with their ligands, the Promega CellTiter-Glo® Luminescent Cell Viability Assay was used. Briefly, each well of a 96-white-well plate was seeded 1x 104 cells. Different concentrations of compounds were added in triplicate on cells. After 24 h, the effects induced by compounds were evaluated with the CellTiter-Glo® Reagent, which quantifies ATP as an indicator of metabolically active cells generating a luminescent signal proportional to the number of viable cells.

The human immortalized keratinocyte (HaCaT) was seeded in 24-well plates (1 × 105 cells/well) in DMEM/F 12 supplemented with 10% foetal bovine serum and 5 µg mL−1 compound solution. The cells were harvested after 24 h and fixed in cold ethanol 70%. After ethanol removal by centrifugation, the cells were stained with 50 µL FxCycle PI/RNAse staining solution (Applied Biosystems) for 1 h at 37 °C. Flow cytometer acquisitioned events were analysed using the FlowJo software.

TAC was measured using the DPPH method [29]. The absorbance of a mixture of DPPH and the compound in methanol was measured initially and after 30 min at 517 nm. Thus, 1.8 mL methanolic solution of DPPH (10−4 mol dm−3) was mixed with 0.2 mL methanolic solution (2 mg mL−1) of each compound [30, 31]. A solution of ascorbic acid (AA) was used as control. The TAC values were determined using the following equation:TAC=Abs0-Abs30minAbs0×100.

PHOTO (COLOR)

To a solution of 1 mmol of 2-hydroxy-8-alkyl-tricyclo[7.3.1.02.7]-tridecan-13-one in 10 mL ethanol or methanol was added 1 mmol of 4-amino-2,3-dimethyl-1-phenyl-3-pyrazolin-5-one. The mixture was stirred and refluxed for 2 h and afterword was stirred at room temperature overnight. The reaction was stopped and the solvent was removed under reduced pressure. A solid was formed with 35-55% yield. The new Schiff bases were purified by chromatography using AcOEt 100% or DCM/MeOH = 9:1 as solvent. The purity of the obtained product was checked by thin-layer chromatography using AcOEt/MeOH (7:1, v/v) as eluent. The β-cycloketol derivatives were obtained by a Michael reaction according to a modified Tilicenko-Barbulescu method [32, 33].

To a solution containing 1 mmol copper (II) acetate hydrated in 25 mL ethanol, 1 mmol Schiff base was added. The reaction mixture was magnetically stirred at 50 °C for 4-5 h, until a sparingly soluble species was formed. The brown-coloured precipitate was filtered off, washed with ethanol and air-dried.

The [1 + 1] condensation of 4-amino-2,3-dimethyl-1-phenyl-3-pyrazolin-5-one and 2-hydroxy-8-alkyl-tricyclo[7.3.1.02.7]-tridecane-13-one produced new Schiff bases that in reaction with copper(II) acetate generated the species [CuL(CH3COO)2] as shown in Scheme 1. The Schiff bases and complexes have been characterized by microanalytical, ESI-MS, IR, UV-Vis-NIR and EPR spectroscopy and thermal analysis. Chemical analyses are consistent with both Schiff bases and complexes formation (Table 1). The complexes behave as nonelectrolyte as their molar conductance values (in DMSO) are in the range of 7-9 Ω−1 cm2 mol−1 [34].Synthetic route for Schiff bases and complexes preparation together with proposed coordination

PHOTO (COLOR)

Analytical data of Schiff bases and complexes

aAcOEt/MeOH = 7:1

The most relevant IR bands and assignments for the Schiff bases and complexes are summarized in Table 2. The formation of Schiff bases was confirmed by appearance of a strong band in the range 1590-1650 cm−1 assigned to ν(C=N) vibration. This band is shifted by 25-35 cm−1 to lower wave numbers in comparison with the ligands, thus indicating their coordination through the nitrogen atom of azomethine group [16, 17]. A similar shift by 10-16 cm−1 can be noticed for ν(C=O) vibration.

The medium band in the range 3340-3450 cm−1 that appears in both ligands and complexes spectra indicates that the hydroxyl group remains protonated.

As for acetate, the difference (Δ) between νas(COO) and νs(COO) stretching vibrations is often used as a criterion for assessing the nature of acetate in complexes, respectively, free ion and unidentate or bidentate (bridge or chelate) ligand [35, 36]. As a result, a Δ value in the range 120-140 cm−1 can be considered an indicative of the acetate coordination as bidentate for all complexes. According to several authors, the shifting of the band assigned to νas(COO) in comparison with CH3COONa can be used as criterion to differentiate the chelate for bridge coordination. For all complexes, the shift of this band to lower wavenumbers suggests the acetate presence as chelate [37, 38].

The new bands in the range 400-510 cm−1 were assigned to stretching vibrations ν(Cu-O) and ν(Cu-N), respectively [39].

The values identified in the 1H NMR and 13C NMR spectra of ligands are presented in Table 1S. Regarding the NMR spectra, the major change that appears in the newly synthesized compounds can be noticed in 13C-NMR. Consequently, the original peak at about 220 ppm (due to the C=O moiety in the starting materials tricycloketone) is replaced by the new one at lower values, around 159-162 ppm, corresponding to the formation of the C=N bonds in the synthesized Schiff bases.

The m/z values identified in the MS spectra confirm the structures proposed in Scheme 1 for both Schiff bases and complexes (Table 2S). Thus, in the positive mode the pseudomolecular ions of complexes were not found but fragments such [Cu(CH3COO)2(R1-H) + H2O]+and those provided by ligands splitting as R1 and R2 appear in all spectra (Scheme 2).Schiff base main fragmentation

PHOTO (COLOR)

In the absence of structure characterization obtained with single-crystal X-ray diffraction, the correlation between electronic and EPR data provides valuable information concerning stereochemistry of Cu(II) in complexes. The absorption maxima values from solid-state spectra of ligands and complexes are presented in Table 3S. In the 24,000-28,000 cm−1 range, two intense bands appear in the spectra of both Schiff bases and complexes assigned to n → π* and π → π* intraligand transitions, respectively. These bands are shifted in the complexes spectra as a result of coordination. The band that appears with low intensity in the ligands spectra in the visible domain at about 18,000 cm−1 is obscured in the complexes spectra by the d-d one. The complexes spectra exhibit also a broadband in the visible range due the d-d transitions which can be assigned to a distorted octahedral symmetry [40]. This band is assigned to dz2→dx2-y2 transition for a rhombic elongated stereochemistry.

The powder EPR spectra of all complexes exhibit a pattern characteristic for a rhombic elongated octahedral stereochemistry (Fig. 1). All three g factors can be observed as an indicative of a rhombic local environments with all molecular axes parallel aligned (Table 4S). Moreover, the spectra exhibit a resolved hyperfine structure that accounts for a structure with isolated Cu(II) centres [20].Powder EPR spectra of complexes (1)-(4) recorded at room temperature

PHOTO (COLOR)

As a general observation, the ligands decompose without melting. Thermal behaviour of ligand L1 (Fig. 2, Table 3) seems to follow the same pattern of fragmentation as that from mass spectroscopy. Thus, the first exothermic decomposition step corresponds to R2 fragment degradation. The remaining fragment (R1) undergoes several oxidative degradation processes as all thermoanalytical curves (TG, DTG and DTA) indicate. The other ligands, L2-L4, decompose in several not well-delimited steps. The DTG and DTA curves' shape indicates an overlap of two or three exothermic processes.TG, DTG and DTA curves for L1

PHOTO (COLOR)

Complex (1) decomposes in three well-delimited steps (Fig. 3). The first one corresponds to ligand decomposition. Several thermal effects were evidenced on DTA curve indicating a complex process. The residue obtained after this step was isolated and investigated by both IR spectroscopy and chemical analysis which confirm the composition of copper acetate. The IR spectrum of isolated residue is a very simple one (Fig. 1S up) containing only bands characteristic for acetate ion (νas(COO), 1530 s; δas(CH), 1464 vs; νs(COO), 1435 vs; δs(CH), 1376 m; δ(COO), 660 s; π(COO), 570 s). Considering the Δ value of 95 cm−1, one can consider the chelate nature of acetate ion in this residue [35].TG, DTG and DTA curves for [CuL1(CH3COO)2] (1)

PHOTO (COLOR)

In the second step, an endothermic one, copper acetate decomposes to copper carbonate. The IR spectrum recorded for residue isolated at 515 °C contains bands at 1455 and 838 cm−1 which are characteristic for the free carbonate ion (ν3 and ν2 respectively) (Fig. 1S down). Finally, copper carbonate decomposes at copper (II) oxide as XRD diffractogram (PDF No.: 8000076) on final residue indicates (Fig. 2S).

According to TG curve, complex (2) is stable up to 145 °C when it starts to decompose in several exothermic steps. The mass loss in the 145-650 °C range corresponds to L2 ligand oxidative degradation and acetate decomposition. The intermediate residue isolated at 650 °C was identified as copper carbonate according to IR spectrum. In the last step, copper carbonate decomposes into copper oxide. The nature of final residue was confirmed by XDR diffractogram pattern (Fig. 4). TG, DTG and DTA curves for [CuL2(CH3COO)2] (2)

PHOTO (COLOR)

Since complex (3) has the same composition with complex (2) [it differs only by the nature of radical R from ligand L: n-propyl for (2) and i-propyl for (3)], a similar thermal behaviour is expected. Indeed, from the point of view of mass loss the same steps were identified on TG curve. However, some differences have been observed. First, complex (3) is less thermal stable. It starts to decompose at 115 °C with an endothermic effect followed by some exothermic ones. The intermediate residue isolated at 560 °C is copper carbonate as both chemical analysis and IR spectrum indicate. The final residue at 900 °C is copper (II) oxide (Fig. 5).TG, DTG and DTA curves for [CuL3(CH3COO)2] (3)

PHOTO (COLOR)

Complex (4) is the only one which is hydrated and consequently the least thermal stable. The first, endothermic, decomposition step corresponds to water molecule elimination. The low temperature range observed indicates its hydration nature [41-44]. According to the mass loss, the next thermal event is ligand oxidative decomposition, which is a complex process as both DTA and DTG curves indicate. The last decomposition step corresponds to copper acetate decomposition into copper (II) oxide (Fig. 6).TG, DTG and DTA curves for [CuL4(CH3COO)2]·H2O (4)

PHOTO (COLOR)

The antimicrobial activity of the obtained complexes and of their ligands was assessed using quantitative assays against bacterial and fungal strains, in planktonic (Table 4) and biofilm (Table 5) growth state.

This activity was evaluated using as standard antimicrobial substances carbenicillin, in case of the bacterial strains, and fluconazole, in case of the fungal strains. These reference antimicrobials have been chosen in accordance with CLSI recommendations [45, 46]. Regarding the ligands, the most intensive antimicrobial activity against planktonic cells was exhibited by L4, which exhibited moderate antimicrobial activity with MIC of 0.25 mg mL−1 against E. coli and of 0.125 mg mL−1 against C. albicans strains, followed by L3 and L2 which exhibited moderate antimicrobial activity with MIC of 0.25 mg mL−1 against the two Gram-negative bacterial strains.

The complex (4), followed by complex (1), was the most active against all tested microbial strains, exhibiting the same MIC value of 0.125 and 0.25 mg mL−1, respectively, towards all five tested microbial strains. For these two complexes, the antimicrobial activity was superior to that of their ligands for all five tested microbial strains, as revealed by the MIC values which were two to four times lower than those of the ligands. Complex (3) exhibited a moderate effect (MIC of 0.25 mg mL−1) only against S. aureus strain. The complexes (2) and (3) exhibited similar or lower antimicrobial activity in comparison with their ligands (Table 4).

Microbial biofilms are defined as structurally and functionally organized polyspecific communities embedded in an extracellular matrix of exopolymers developed on tissue or inert substrata [27], functionally organized for increased metabolic efficiency, pathogenic synergism, greater resistance to environmental conditions and stress factors, including host defence mechanisms and antibiotics as well as for enhanced virulence [28, 47]. This resistance or tolerance, also called phenotypic or behavioural, is mediated by multiple and incompletely elucidated yet mechanisms, such as biofilm matrix impermeability, entrance in a slow-growing or starvation state, selection of phenotypic variants or persisters cells and activation of stress response genes. The microbial strains used in this study belong to the panel of microorganisms involved in biofilm-associated diseases [48], reason for which the compounds were tested for their antibiofilm activity.

Regarding the antibiofilm activity of the obtained complexes and of their ligands, all ligands exhibited a very good inhibitory activity against the biofilms formed by all tested strains, superior to that of their complexes, the MBEC values being much lower (up to six times) than the corresponding MIC ones, ranging from 2.5 to 0156 mg mL−1. This behaviour could be explained by the fact that the smaller molecules of the ligands penetrate more efficiently the microbial biofilm protective matrix.

From the four complexes, the complex (4) probed to be also the most active against all tested microbial strains, exhibiting a very low MBEC value of 0.0625 mL−1 against C. albicans biofilm. Complexes (1) and (2) exhibited moderate antibiofilm activity (MBEC of 0.125-0.25 mg mL−1) only against all tested microbial strains. Complex (3) exhibited a moderate inhibitory activity against S. aureus and C. albicans biofilms with an MBEC value of 0.125 mg mL−1 (Table 5).

It is remarkable that the activity of the tested compounds on adherent microbial cells was much superior to that of the reference antimicrobials, suggesting their promising potential to develop novel antibiofilm agents.

The compounds proved to exhibit a relatively high toxicity on HaCaT cells at concentrations higher than 3 µg mL−1 (Fig. 7). The viability assay results were also confirmed by the microscopic examination, revealing morphological changes in the cell monolayers (round, small cells, detached monolayer) (Fig. 3S), and by the cell cycle, assays revealing the occurrence of the apoptosis/necrosis peaks in left of G0/G1 phase (Fig. 4S). However, the ligands exhibited a lower toxicity compared with their complexes, especially L1.Cellular viability of HaCaT in compounds presence

PHOTO (COLOR)

The new Schiff bases and their complexes exhibited good antioxidant activity (Table 6) as indicated both TAC and AAEq (ascorbic acid equivalent) values [49]. The highest antioxidant activity was proved by compound (2), followed by compound (3). A weak antioxidant activity was observed for compounds (4) and Schiff base L4 containing the phenyl substituent in the tricyclo[7.3.1.02.7]tridec-13-ylidene backbone. It seems that the higher antioxidant activity for complexes (1)-(3) can be correlated with the presence of an aliphatic side chain, while the presence of an aromatic ring decreases this activity, as shown for complex (4). All complexes exhibit an increased antioxidant activity than corresponding Schiff bases as was observed for other copper (II) complexes with similar ligands [50].

Four new 4-aminoantipyrine-derived Schiff bases and their copper (II) complexes were synthesized and characterized for their physico-chemical properties and biological activity. Both ligands, Schiff bases and acetate, act as chelate leading to octahedral compounds.

Thermal decomposition trend is maintained for this class of complexes, several steps being observed like Schiff base oxidative degradation, copper acetate conversion into carbonate and copper carbonate decomposition to oxide, which was the final residue for all four complexes. The intermediates and final residues were isolated and identified through chemical analyses, IR spectra and powder XRD.

Taken together, the biological activity evaluation results reveal that complex (4) represents a promising candidate for developing novel antimicrobial agents, particularly for combating the biofilm-associated fungal infections, while complexes (2) > (3) > (1) could be further investigated for their antiproliferative effects, taken into account the high cytotoxic effect induced at low concentrations on the spontaneously transformed aneuploid immortal keratinocyte HaCat cell line.

The complexes proved better antioxidant activity than corresponding Schiff base ligands.

The online version of this article (10.1007/s10973-018-7681-1) contains supplementary material, which is available to authorized users.

The authors thank to researcher Drăghici Constantin from Romanian Academy C. D. Neniţescu Organic Chemistry Institute for the help in NMR data interpretation.