2.1. Materials and reagents

All the chemicals used in this study were of the analytical grade (AR). DPA was purchased from Sigma Aldrich, UK. The metal salts used (CuCl₂·2H₂O, NiCl₂·6H₂O, CoCl₂·6H₂O, ZnCl₂·2H₂O, and MnCl₂·2H₂O) are obtained from Sigma Aldrich. The dicarboxylic acids (oxalic, malonic, succinic, adipic, and CBDC acids) were provided by Sigma Aldrich with concentration of 0.01 mol dm⁻³. DPA was prepared in three equivalents (0.03 mol dm⁻³) of HNO₃ acid. HNO₃ was prepared by diluting concentrated ampoules (Riedel-de Haën, UK) and standardized against sodium carbonate. Dioxane was provided by Sigma Aldrich. NaOH (titrant) was prepared and standardized against primary standard potassium hydrogen phthalate solution. The absence of carbonate in the titrant was periodically checked by pH potentiometry using the appropriate Gran function [21]. All solutions were prepared shortly before use in deionized water. The metal content of solutions was determined by complexometric EDTA titration [22].

2.2. Synthesis of the ternary metal complexes

The following procedure was used for the synthesis of the 1:1:1 ternary complexes. A minimum quantity of the chloride metal salt (500 mg, 2.10 mmol) dissolved in water was added slowly to the DPA (359 mg, 2.10 mmol) with continuous stirring. The dibasic acids (222 mg, 2.10 mmol for malonic), (302 mg, 2.1 mmol for CBDCA), (265 mg, 2.10 mmol for oxalic), and (248 mg, 2.10 mmol for succinic) were then added to the above mixture. Na₂CO₃ (222 mg, 2.10 mmol) was then added to neutralize the protons released. The mixture is stirred with heating for further 2 h and stored over night. The precipitated solid complexes were isolated by vacuum filtration and washed three times with water, ethanol, and finally with diethyl ether and dried on vacuum over silica gel. The yield ranged from 80% to 90%. The complexes were subjected to elemental and spectroscopic analysis. The elemental analyses were conducted using the Perkin-Elmer 2400 Analyzer and the results were found to be in good agreement (±0.3%) with the calculated values.

2.3. Instruments and measuring techniques

pH-metric measurements were performed by using a Metrohm 686 Titroprocessor equipped with 665 dosimate. The titration cell used for the titration is a double wall cell of 50 mL capacity equipped with a magnetic stirrer. The cell solution was continuously stirred during the titration using magnetic stirring system. The cell is covered with a calibration lid and contains four holes for Metrohm glass electrode, glass tubing for nitrogen injection (degassing), thermometric probe, and a plastic tube for alkaline solution. Before filling of a tube with alkali solution, the tube was washed several times with distilled water and then washed with alkali solution at least three times. Also, the air bubbles were avoided to leak in the tube in order to get accurate results for the measured volumes. The titroprocessor and the glass electrode were calibrated with standard buffer solutions of both phosphate mixture of KH₂PO₄ and Na₂HPO₄ (pH 6.865) and potassium hydrogen phthalate solution (pH 4.008) [23]. The titration reaction was performed using standard solution of 0.05 mol dm⁻³ carbonate free sodium hydroxide solution in presence of purified N₂ atmosphere. The pH meter readings were converted into hydrogen ion concentration as given in the literature using Van Uitert and Hass equation [24]. All potentiometric titrations were carried out at 25 ± 0.05 °C. The temperature of all solutions was maintained at 25 ± 0.05 °C by circulation of thermostated water through the outer jacket of the cell. All titrations were performed in triplicate. Estimated precision was ±0.1 mV and ±0.003 mL for the EMF and titrant readings, respectively.

As is well known, pH-meters read −log₁₀ α_H⁺ (pH) whereas the potentiometric method we used for the calculation of stability constants requires −log₁₀ [H⁺] (p[H]). The pH-meter readings (B) recorded in dioxane-water solutions was converted to hydrogen ion concentrations [H⁺] by using the widely used Van Uitert and Hass equation [24]:

where log₁₀U_H is the correction factor for the solvent composition and ionic strength for which B is determined. Values of pK_w in dioxane-water mixtures were determined as described previously [24,25]. For this purpose, various amounts of standard NaOH solution were added to a solution containing 0.1 mol dm⁻³ NaNO₃. The values of [OH⁻] were calculated from the amount of base added; [OH⁺] was calculated from the pH. The products of [OH⁻] and [H⁺] were used to calculate the pKw values described above. The mean values obtained in this way at 25 °C for log₁₀ [H⁺][OH⁻] are pK_w = 14.17, 14.60, 15.12, and 15.49 for 12.5%, 25%, 37.5% and 50% dioxane-water solutions, respectively. The values were found to be consistent with literature data [26,27,28].

The protonation constants of the ligands in the protonated form were determined potentiometrically by titrating the ligand (40 cm³) solution (1.25 × 10⁻³ mol dm⁻³) of constant ionic strength 0.1 mol dm⁻³ (adjusted with NaNO₃). The stability constant of the M(II)(DPA) complexes was determined by titrating 40 cm³ of a solution mixture of M(II) (1.25 × 10⁻³ mol dm⁻³), DPA ligand (2.5 × 10⁻³ mol dm⁻³) in concentration ratio 1:2 and NaNO₃ (0.1 mol dm⁻³). The conditions of measurements of the ternary complexes were the same as for the binary ones, but the solutions contained equivalent amounts of DPA, Co(II), and dicarboxylic acid ligands (L) in concentration ratio 1:1:1.

The microchemical analysis of the separated solid complexes for C, H, and N was performed in the Microanalytical Center, Cairo University. The analyses were performed twice to check the accuracy of the analytical data. IR spectra were measured on a 80486-pc FTIR Shimadzu spectrophotometer using KBr pellets. UV-visible spectra of freshly prepared solutions of the complexes were recorded at room temperature using a Shimadzu UV-1800 recording spectrophotometer. The molar conductance of the complexes was measured in 1.0 × 10⁻³ mol dm⁻³ DMSO solution at 25 °C using a Jenway 4510 conductivity meter.

2.4. Data processing

The calculations were obtained from Ca. 100 data points in each titration using the least-squares computer program MINIQUAD-75 [29]. The stoichiometries and stability constants of the complexes formed were determined by trying various possible composition models. The model selected gave the best statistical fit and was chemically consistent with the titration data without showing any systematic bias in residuals [29]. The fitted model was tested by comparing the experimental titration data point and the theoretical (simulated) curve calculated from the values of the acid dissociation constants of the ligand and the formation constant of the corresponding complex, as given in Table 1. The complexation equilibria were elucidated from the concentration distribution curves obtained with the program SPECIES [30] under the experimental condition used.

2.5. Antimicrobial activity

Disk diffusion susceptibility method [15] in accordance with National Committee for Clinical Laboratory Standards (NCCLS) guidelines was used for initial screening of compounds for antibacterial activity against gram-positive (G⁺) bacteria Staphylococcus aureus, gram-negative (G⁻) bacteria Escherichia coli and also for their antifungal activity against Candida albicans and Aspergillus flavus described in British Pharmacopoeia (2000). Nutrient agar was melted at 45.00 °C, and inoculated with the cell suspension (1:100) bacteria or yeast. The flask was shaken well and poured into a Petri-dish (15 cm in diameter). Filter paper discs (6 mm) Whatman No. 2 were thoroughly moistened with antibiotics (50 g). The treated discs were aseptically transferred and placed on the surface of the inoculated plates and kept in a refrigerator for 1 h to permit diffusion of antimicrobial substances. The plates were incubated at 7.00 °C, for 24 h in the case of bacteria and at 28.00 °C, for 48 h in the case of fungi. The zones of inhibition were measured in mm. The mean values of inhibition were calculated from triple reading in each test [15]. For studying the antimicrobial properties of drug complexes, the following media were used (weights are given in gram per one-liter medium): (1). Nutrient agar medium (pH = 7.4): it consists of beef extract (1 g), yeast extract (2 g), peptone (5 g), sodium chloride (5 g), agar agar (15 g), and distilled water (100 cm³); (2). Sabouraud dextrose agar (SDA) medium (pH = 5.6): it consists of peptone (10 g), dextrose (20 g), agar agar (15 g), and distilled water (100 cm³). Moreover, some commercialized antibiotics were evaluated for their antibacterial activity and a comparison with the synthesized complexes present in this investigation was drawn. The results are expressed as zone of inhibition in millimeters (mm). Experiments were done in triplicate and an average standard deviation of < 2 was observed.

3.1. Protonation constant of DPA

Before calculation of the stability constants of metal complexes, the acid dissociation constant of the ligand should be first determined. The cumulative (overall) formation constant β_pqr of a species is defined on the basis of the following equilibrium (where charges on individual species are omitted for simplicity):

If there is no metal complex formation (p = 0), then only protonation of the ligand occurs. The ligand is prepared in three equivalents HNO₃ solution. Analysis of the potentiometric data of DPA in the triprotonated form yields two pK_a values corresponding to the protonated amino groups (Table 1). These results are in agreement with previous investigation [15]:

3.2. Binary complexes involving DPA with the metal ions Cu(II), Ni(II), Co(II), Zn(II), and Mn(II)

Each metal ion with the ligand was titrated with NaOH. Comparing the titration curve of the free ligand with those of the complexed ligand indicated that addition of the metal ions to the free ligand shifts the curve to lower pHs. In other words, the curves of the complexes are situated at lower pHs than the free ligand curve as shown in Figure 2. This can be explained simply as a result of proton release from the coordinated ligand, which implies formation of complexes through release of hydrogen ion upon complex formation. The complex formation equilibria of the metal complexes are characterized by fitting the potentiometric data to various models. All feasible theoretical models were compared with those experimentally obtained. The equilibrium patterns were chosen so as to lie between the observed and the calculated data applying accurate statistical analysis involving the sum of squares of residuals. The whole titration data obtained was fitted with different composition models. The best model was found to be consistent with the formation of the complexes with stoichiometric coefficient 110 together with the protonated form in case of Mn(II) ions (Table 1). The validity of the complex formation model was tested by comparing the potentiometric titration data with the theoretically simulated curve obtained from the protonation constants of the ligand and the formation constants of the corresponding complex species.

Figure 2: Potentiometric titration curve of the [Co-DPA] system.

Estimation of the concentration distribution of the various species in solution provides a useful picture of metal ion binding. The main features observed in the species distribution plots in these systems are shown in the speciation diagram obtained for Mn(II)-, Cu(II)-, and Co(II)-DPA complexes; as shown in Figure 3. The protonated complex for Mn(II) prevails with formation degree of 90% at pH ≈ 4.5. The 110 complex species starts to form at pH ≈ 4 and its concentration increases with increase of pH. Cu(II) and Co(II) also form 1:1 complexes with DPA with the stoichiometry 110, with formation degree of 55% and 45% at pH ≈ 4 and 6.5, respectively.

3.3. Correlation of the properties of investigated central metal ions with the formation constants of their complexes

It is necessary to correlate the stability constants of metal complexes with the characteristic properties of the metal ions, in an effort to explain why a given ligand prefers binding to one metal rather than another, and to give information about the nature of chemical bonding in complexes and make possible the estimation of unknown stability constants. Here we have discussed the stability constants of the binary metal complexes (M(II)-L) and the ionic radius, the ionization energy, the atomic number, and the electronegativity of the selected metal ions. The formation constants of M^II-complexes of bivalent 3D transition metal ions with DPA, as given in Table 1, are in the following order: Mn²⁺(logβ₁₁₀ = 4.01) < Co²⁺(logβ₁₁₀ = 4.55) < Ni²⁺(logβ₁₁₀ = 6.88) < Cu²⁺(logβ₁₁₀ = 7.65) > Zn²⁺(logβ₁₁₀ = 5.87), which is in accordance with Irving-Williams order [16,32,33]. The correlation between the logKML and the reciprocal ionic radii (1/r) of the studied bivalent transition metal ions was found to be almost linear (Figure 4(a)). Also, a good linear correlation has been obtained between logK_ML and the electronegativities of the metal ions under study (Figure 4(b)). This in accordance with the fact that increasing electronegativity of the metal ions (Zn²⁺(1.65) < Co²⁺(1.88) < Ni²⁺(1.91) < Cu²⁺(2.0)) (Table S1) will decrease the electronegativity difference between the metal atom and the donor atom of the ligand. Thus, the metal ligand bond would have more covalent character, which may lead to greater stability of the metal chelates. A good linear relationship has been obtained between logK_ML and the second ionization potential of the bivalent metal ions under study (Figure 4(c)). In general, it is noted that the stability constant of the Cu²⁺ complex is quite large compared to the other metals. The ligand field will give Cu²⁺ some extra stabilization due to tetragonal distortion of the octahedral symmetry. The Cu(II) complex will be further stabilized due to the Jahn-Teller effect [34,35]. Thus, logK value for the Cu²⁺-complex deviates significantly when logK values of metal chelates are plotted against properties of the metal ions.

3.4. Effect of solvent

Mostly, the physical constants of bioligands are studied in aqueous media. Despite this, little is known about the chemistry of biomolecules in mixtures of organic solvents and water. Determination of the protonation constants of bioactive ligands and the stability constants of their metal complexes in various media is important for complete understanding of the physicochemical behavior of such molecules. It is well established that the “effective” or “equivalent solution” dielectric constants in protein or active site cavities of enzymes are small compared to those in bulk water [14,36]. Estimates for the dielectric constants in such locations range from 30 to 70 [36]. Hence by using aqueous solutions containing ≈ 10–50% dioxane, one may expect to simulate to some degree the situation in active site cavities. This allows extrapolating the data into physiological conditions.

Careful examination of media effects on the equilibrium constants (Table 2) reveals the following features.

(1) pK_a of DPA increases linearly with increasing the percentage of organic solvent in the medium (Figure 5). This may be correlated with the ability of a solvent of relatively low dielectric constant to increase the electrostatic forces between the proton and the ligand and, consequently, the pK_a value increases.

(2) The stability constant (log₁₀K₁) of the Co(II)-DPA complex increases with the increase of dioxane concentration (Figure 5). This can be interpreted in terms of an electrostatic model. In general, lowering of the dielectric constant of a medium (by increasing dioxane content) favors the interaction between the Co(II) ion and DPA and, consequently, the stability constant of the complex increases.

3.5. Ternary complexes involving DPA with Co(II) and some selected dibasic acids

In the Co(DPA)-dicarboxylic acid system, computer analysis of the pH titration data showed the presence of the 1:1:1 species and its protonated form. The results in Table 3 show that the formation constant of the 1:1:1 complexes involving the formation of five- and six-membered chelate rings may be attributed more to the higher stability of the five- and six-membered rings, known to be energetically favored, than the seven- and eight-membered rings. It is interesting to note that CBDC acid has a higher stability constant than malonic acid, although they both form six-membered chelate rings. This may be due to the higher pK_a values of CBDC acid in comparison with the pK_a values of malonic acid. The concentration distribution diagram of the [Co-DPA-CBDC acid] system (Figure 6) shows that the main species in the physiological pH range is the ring-closed form, 1110, which reaches a maximum concentration of ≈ 78% around pH 7. The interest in this high stability of the [Co-DPA-CBDC acid] complex at the physiological pH range is that the sugar group of DNA may undergo stacking interactions with the cyclobutane ring. Such interactions could favor reactions with DNA. The latter effect is similar to that reported for carboplatin, where stacking interactions between the cyclobutane ring of CBDC acid and the sugar group of DNA occur and form part of the increased antitumor activity [37].

3.6. Characterization of the isolated solid complexes

All the synthesized complexes are air-stable, nonhygroscopic, high melting, and insoluble in H₂O, but soluble in coordinating solvents such as DMF and DMSO. The formulation of these complexes is based on the elemental analysis, IR, UV-vis, magnetic and conductance measurements. The elemental analysis of the complexes (Table 4) revealed the presence of water molecules. All have the general composition [Co(DPA)L]nH₂O (where L = Ox., Mal., Suc. or CBDC acid; n = 2 or 1). The complexes are nonionic as revealed by their low conductance values. The study of the spectral and magnetic properties of the solid complexes provides information that can throw considerable light on their geometry. Careful examination of the recorded spectra of the different synthesized ternary complexes reveals the following features.

Table 4: Microanalytical data, magnetic moment, molar conductance and electronic spectral data of the prepared ternary complexes.

IR spectra of the complexes exhibit strong NH absorption bands, being unchanged compared to their position in the ligand, indicating that the NH group of DPA is not involved in coordination. The IR spectra of the formed ternary cobalt complexes reveal the presence of water molecules, as indicated by a broad band of ν(O−H) at 3,420–3,450 cm⁻¹ and two weaker bands at approximately 899–911 cm⁻¹ and 685–690 cm⁻¹, respectively [6,38], which is in accordance with the results of elemental analyses. It was reported that [39] coordinated water exhibit frequencies at 825 cm⁻¹, 575 cm⁻¹, and 500 cm⁻¹. The absence of spectral bands in these regions indicates that the water molecules are not coordinated but are present as lattice water. The band in the region 1,625–1,610 cm⁻¹, that is characteristic of the (C=N) stretching mode [40], is shifted towards lower frequencies in the spectra of the cobalt complexes 1,609–1,590 cm⁻¹ indicating the involvement of the nitrogen in chelation with the cobalt ion [41]. The pyridyl ring vibration bands at 1,578–1,608 cm⁻¹ and 765–779 cm⁻¹, respectively, indicate that the nitrogen atoms of the pyridyl ring of the ligands donate a pair of electrons each to the central metal forming coordinate covalent bond [42]. This is further supported by the appearance of a new medium intensity band in the region 480–520 cm⁻¹ assignable to ν(M−N) vibration [43]. The bands in the 2,923–2,530 cm⁻¹ region (medium to weak) can be assigned to C−H stretching vibrations in both the primary and secondary ligands. The bands in the 1,619–1,603 cm⁻¹ region in all the prepared complexes are typical of coordinated carboxylate group stretching [44]. This is based on the fact that the unionized and uncoordinated COO stretching band occurs at 1,750–1,700 cm⁻¹ whereas the ionized and coordinated COO stretching band appears at 1,650–1,590 cm⁻¹. It was reported previously that the magnitude of the Δν[Δν = ν_asy(COO⁻) − ν_sym(COO⁻)] can be correlated with the coordination modes of the carboxylate anion [45,46]. The calculated values of Δν for each of the studied complexes were found to be > 200 cm⁻¹, reflecting the monodentate coordination of each carboxylate group. Therefore, the dicarboxylic acids act as bidentate O⁻, O⁻ ligands and coordinate to the Co(II) metal via two monodentate carboxylate groups. The band appearing in the 1,400–1,391 cm⁻¹ region could be attributed to the symmetric stretching vibration of the coordinated carboxylates [44].

3.7. Electronic spectra, magnetic studies and conductivity measurements

Further insight into the structures of the isolated solid complexes is proposed on the basis of the spectral and magnetic studies. The spectra of all of the complexes contain an absorption band in the 15,822–16,260 cm⁻¹ range that may be assigned to a d–d transition of metal ions [47]. The Co(II) complexes show bands at 16,168 cm⁻¹, 16,210 cm⁻¹, 15,844 cm⁻¹, and 15,723 cm⁻¹ for the CBDC, malonate, oxalate, and succinate complexes, respectively, (Table 4). These bands are probably due to the ⁴A₂(F) → ⁴T₁(P) transition of tetrahedral geometry [48]. Charge transfer transition bands are also appearing for all the studied Co(II) complexes at the range of ca. 23,070–23,350 cm⁻¹ as reported previously for similar systems [49,50].

The observed values of magnetic moments for complexes are generally diagnostic of the coordination geometry about the metal ion. The calculated magnetic moment of the synthesized complexes is given in Table 4. All the values are in the range of 4.1–4.5 B.M, which are in consistence with a tetrahedral geometry for the Co(II) ion as reported previously [51,52].

The synthesized ternary complexes were dissolved in DMSO and the molar conductivities of 10 mol dm⁻³ of their solutions were measured at 25 °C. The values obtained are given in Table 4. The molar conductivity values (Λ M) of the Co(II) complexes were found to be between 15.9 Ω⁻¹ cm² mol⁻¹ and 20.8 Ω⁻¹ cm² mol⁻¹. The low values indicate that all Co(II) complexes are nonelectrolytes. This is in accordance with the fact that the conductivity values for a nonelectrolyte are below 50 Ω⁻¹ cm² mol⁻¹ in DMSO solution [53,54].

3.8. Structural interpretation

From all of the above observations and according to the results reported in this paper based on the analytical data, IR, molar conductivity, spectral (solid reflectance) and magnetic measurements, the structure of these mixed-ligand complexes is tetrahedral. This indicates that the DPA ligand behaves as a bidentate chelating ligand through the two pyridine rings while the dicarboxylic acids are coordinated via the two oxygen atoms of carboxylate groups.

3.9. Antimicrobial activity

Metal chelates were well known to have an enhanced antimicrobial and biological activity. It has been found that the Cu(II) complex of DPA catalysis the hydrolysis of methyl acetate, simple amides, and triesters of phosphoric acids [55]. These systems are considered as very bioactive models for metalloenzyme substrate interaction. In this study, we had screened the biological potential of the synthesized complexes in vivo against different species of bacteria and fungi. In the process of testing the antimicrobial activity of these compounds, we have used more than one test organism in order to increase the chance of detecting antibiotic principles in the tested materials. The organisms used in the present investigations include one G⁺ (S. aureus) as well as one G⁻ (E. coli) bacteria. All of the tested compounds show a remarkable pharmacological activity against different bacterial strains, that is, G⁺ and G⁻ bacteria. The disc diffusion method was used to evaluate the antibacterial activity of the synthesized ternary complexes. The medium used for growing the culture was nutrient agar. The data are listed in Table 5. The increase in biological activity of the metal chelates may be attributed to the effect of the metal ion on the normal cell process. The cell toxicity is increased and this could be explained in the light of Tweedy’s chelation theory [56]. The polarity of the metal ion is considerably reduced upon chelation because of partial sharing of metal positive charge with the donor ligand, and the possible π-electron delocalization that occurs within the whole chelate ring upon coordination. This chelation could enhance the lipophilic character of the central metal atom and accordingly the hydrophobic character and liposolubility of the complex would increase, favoring the process of permeation through the lipid layers of the cell membrane. This would increase the rate of uptake and entrance of the complexes into the cell membrane, thus the antimicrobial activity of the testing compounds. Accordingly, the antimicrobial activity of the complexes present in this investigation can be referred to the increase of their lipophilic character which in turn deactivates enzymes responsible for respiration processes and probably other cellular enzymes, which play a vital role in various metabolic pathways of the tested microorganisms. On comparing the biological activity of the synthesized complexes with the standards tetracycline (antibacterial agent) and amphotericin B (antifungal agent), the following results are obtained:

(a) using G⁻ E. coli and G⁺ S. aureus, we discovered that the biological activity of the CBDC acid complex is higher than that of the malonate, oxalate, and succinate complexes and slightly lower than that of tetracycline standard;

(b) all complexes were found to be more active against G⁺ than G⁻ bacteria.

Table 5: The antibacterial and antifungal activity of the synthesized metal complexes.

The antifungal activities of the synthesized complexes were tested against two fungi (A. flavus and C. albicans) test organisms in vitro using disc-diffusion method. The amphotericin B was used as a standard drug. The susceptibility of the selected strains of fungi towards the synthesized complexes was judged by measuring the diameter of the formed zone of inhibition. The results are given in Table 5. Metal ions are adsorbed on the cell walls of the microorganisms, disturbing the respiration processes of the cells and thus blocking the protein synthesis that is required for further growth of the organisms. Hence, metal ions are essential for the growth-inhibitory effects. According to Overtone’s concept of cell permeability, the lipid membrane that surrounds the cell favors the passage of only lipid-soluble materials, so lipophilicity is an important factor controlling the antifungal activity. Upon chelation, the polarity of the metal ion will be reduced due to the overlap of the ligand orbitals and partial sharing of the positive charge of the metal ion with donor groups. This increased lipophilicity facilitates the penetration of the complexes into lipid membranes, further restricting proliferation of the microorganisms. The data in Table 5 shows the following:

(c) using A. flavus fungus: the antifungal activity of the CBDC acid, the oxalate and the succinate complexes are higher than that of the standard antifungal agent amphotericin B; whereas, the antifungal activity of the malonate complex is slightly lower than that of the standard;

(d) using C. albicans fungus: the biological activity of malonate complex is found to be higher than the other complexes, and all have lower values than that of the standard. The importance of this lies in the fact that these complexes could be applied fairly in the treatment of some common diseases caused by E. coli, for example, Septicaemia, Gastroenteritis, Urinary tract infections, and hospital acquired infections [49] or any other infections caused by any of these particular strains.