Copper(II), a Peculiar Metal Ion for Complexation With Monensin A Ionophore

ABSTRACT

Monensin A transports Na⁺ ions through the cell membrane, but it is also able to bind divalent metal ions potentially influencing the biological activity of the ionophore. In line with our previous studies, here we explored interaction of the essential Cu(II) with monensinate A and suggested the types of coordination compounds formed. Circular dichroism (VIS‐ and NIR‐CD), VIS light absorption, and electron paramagnetic resonance (EPR) spectroscopies, as well as electrospray ionization mass spectrometry (ESI‐MS) were utilized for identification and characterization of the complexes. Bis or mono Cu(II) complexes prevailed at ligand excess or comparable reactant concentrations, respectively. Excess Cu(II) salt, however, induced the formation of dinuclear coordination species, not observed previously with divalent metal cations as Mg(II), Ca(II), Co(II), Mn(II), Ni(II), Zn(II). The dinuclear complex is EPR‐active being a rare example of non‐EPR‐silent coordination compounds containing two Cu(II) centers and is the first case among the known metal(II) complexes of the antibiotic monensin A. The origin of the counterion did not influence the properties of the bis compound but contributed to the spectral behavior of the newly detected dinuclear Cu(II) species. This study revealed that the interactions of a drug molecule with metal(II) ions during its practical application may be significant and it should be considered in discussions about the mechanism of action.

Copper(II) forms three chiral and EPR‐active complexes with monensinate A in methanol. Beside the bis [CuL₂] and mono [CuL]⁺ species, the new dinuclear construct [Cu₂L]³⁺ exhibits an [ML]⁺‐type and a metal salt‐type copper(II) centers weakly bound to each other through counter ion(s) and thus, being magnetically independent.

1. Introduction

The natural polyether ionophores (PIs) are large group of compounds with pronounced antiparasitic properties against coccidiosis, considered the most costly intestinal disease of poultry. Today, among the over 120 identified structures of PIs, only a few are approved by the U.S. Food and Drug Administration (FDA) as an integral part of an animal health and welfare program: lasalocid, monensin, laidlomycin, salinomycin, narasin, semduramycin, and maduramycin [1]. Known as coccidiostats and antibacterial agents [2, 3, 4, 5, 6, 7], these “old” drugs have recently attracted wide scientific interest due to the discovery of their potential in various fields of medicinal chemistry [8, 9, 10, 11, 12, 13, 14, 15].

The majority of ionophore antibiotics contain a carboxyl moiety (“head”) linked to a polyether chain of various lengths, ending with one or two hydroxyl groups located at the “tail” of the molecule. Except lasalocid (the smallest PI), the H‐bonding between these COOH and OH functions results in the formation of a ring structure with lipophilic exterior (due to the polyether alkyl substituents) and hydrophilic internal space as the ether and hydroxyl oxygen atoms are oriented inward. The shaped cavity can accommodate water molecule or monovalent metal cations. The trapping of M⁺ occurs through complexation with the inner O‐donors [16, 17, 18, 19], thereby allowing PIs to serve as metal ion carriers, which underlies their antimicrobial mode of action.

Monensin is a widely used therapeutic agent in veterinary medicine against coccidiosis and bacterial infections, with sodium monensinate A (MonNa) being its predominant form produced by Streptomyces cinnamonensis during fermentation process. Although belonging to monovalent PIs, some studies revealed that the antibiotic can interact with the divalent cations and influence their bioavailability under in vivo conditions [20, 21]. Cox et al. determined the stability constants of [MgMon]⁺ and [CaMon]⁺ species in methanolic solutions by disproportionative reaction of the cations with the corresponding Ag(I) complex [22]. Later Hamidinia et al. established the selectivity of monensin for binding Pb(II) over Mg(II), Ca(II), and Zn(II) ions and reported the formation of [PbMon]⁺ and [PbMon(OH)] in 80% methanol/water environment [23]. The fragmentary data and the lack of reliable information about the putative structure of the hitherto presented coordination species of monensin with M(II) metal ions led us to initiate an in‐depth study on the ability of the antibiotic to coordinate metal cations in different oxidation states. We found that the monensinate A anion (Mon⁻) derived from monensic acid A (MonH, Scheme 1) reacts with M(II) to form bis complexes of composition [M(Mon)₂(H₂O)₂], which structures have been demonstrated by single crystal X‐ray diffraction for Mn(II), Co(II) [24], Ni(II), Zn(II) [25], Mg(II), Ca(II) [26], and Cd(II) [27] ions. For comparison, we have also collected the crystal structures of monensic acid and monensinate A with monovalent metal ions in the supplementary Table S1. Recently, the systems bearing Mon⁻ and M(II) in methanolic solutions (due to the poor solubility of monensinate in water) were subjected to a combinatorial approach taking the advantages of circular dichroism spectroscopy in the ultraviolet (UV‐CD), visible (VIS‐CD), and near‐infrared (NIR‐CD) wavelength range. It was found that along with known [M(Mon)₂(H₂O)₂], the polyether ionophore is also coordinated as a mono complex [M(Mon)(H₂O)]⁺ depending on the metal‐to‐ligand molar ratio. In a set of experiments, we determined the apparent stability constants of the two types of complexes either from direct titration experiments or from competition reactions between Co(II) and non‐absorbing divalent metal cations [28, 29]. Cu(II) ion readily interacts with molecules exhibiting carboxylic and hydroxyl donor groups [30, 31, 32]. It has also been demonstrated to form a bis complex of [Cu(MonNa)₂Cl₂] composition with MonNa in solid state [33], while the deprotonated antibiotic carrying a water molecule in its cavity could also form a bis complex [Cu(Mon)₂(H₂O)₂] in solution and in solid phase [34].

SCHEME 1.

Schematic drawing of monensic acid A (MonH).

We initiated this study to better understand the interaction of Cu(II) with monensinate A. We report the findings of the detailed investigation on Mon^– complexes with Cu(II) perchlorate and chloride in methanolic solution showing the distinct behavior of this metal ion among the studied M(II) cations. The spectral changes in the reaction mixtures containing Cu(II) and Mon⁻ in a molar ratio varying in a broad range (from c.a. 1:10 to 5:1) were followed by means of circular dichroism (VIS‐ and NIR‐CD), VIS light absorption and electron paramagnetic resonance (EPR) spectroscopies, as well as by electrospray ionization mass spectrometry (ESI‐MS) to recognize the formed Cu(II) complexes of Mon^–.

2. Results and Discussion

2.1. CD Spectroscopic Investigation of the Cu(II)—Monensinate A Binary System

Previously, we demonstrated the usefulness of the UV‐, VIS‐, and NIR‐CD spectroscopy in discovering the complex species that may form between Mon^– and divalent metal ions [28, 29]. These results prompted us to utilize this method for the identification and characterization of Mon^– complexes with Cu(II). MonH×H₂O was initially deprotonated by an equivalent amount of tetramethyl ammonium hydroxide (TMAOH) in each experiment, to promote the complex formation. Our preliminary studies revealed that the far UV‐CD is not sensitive against the complex formation with Cu(II) [34]. This suggested that the conformation of the ligand was not significantly affected by the interaction with Cu(II). Therefore, the formation of only one type of complex species could be proposed in solution from these experiments, in contrast to the previous results on other divalent metal ions showing the formation of at least two types of complexes—[M(Mon)₂(H₂O)₂] (bis) and [M(Mon)(H₂O)]⁺ (mono)—in methanol, denoted in the following as [ML₂] and [ML]⁺, respectively [28, 29]. The analysis of the spectral series recorded in the near UV range (270–400 nm) revealed that the system is more complicated than it was implied from the far UV spectra. However, the spectra were unstructured, preventing to obtain more detailed information (Figure S1, Supplementary material.) At the same time, remarkable spectral changes were observed in the reaction mixtures containing Cu(II) perchlorate and Mon⁻ in metal‐to‐ligand molar ratios varying from 1:10 to 5:1 by means of VIS‐ and NIR‐CD spectroscopy, as shown in Figure 1.

FIGURE 1.

Titration of monensinate A with Cu(ClO₄)₂ followed by CD spectroscopy; c_MonH = c_TMAOH = 10.10 mM. Panel (a) shows the spectra recorded at each different metal‐to‐ligand ratio. The spectral intensities were corrected for the dilution. (b) The residual CD intensities obtained by matrix rank analysis of the accumulated data set considering the formation of increasing number of complex species. (c) The change of CD intensities at selected wavelength values depending on the molar ratio of Cu(II) and monensinate A.

The methanolic solution of the metal ion itself would not cause any CD intensity due to the lack of chirality, while monensinate does not absorb the visible light. Thus, only the complexes formed with monensinate A contribute to the CD spectra shown in Figure 1a. The CD titration clearly revealed the formation of more than one complex species in the system. Two isodichroic points were identified at ∼ 860 nm (∼ 17 mdeg) and ∼ 816 nm (∼ 4 mdeg), respectively. The first involved the spectra recorded at metal‐to‐ligand ratios 0.5–1, while the second was observed at metal ion excess. This surprisingly suggests that even more than two chiral complex species may exist in solution during the titration process. To verify this assumption, the matrix rank analysis (using the MRA program [35]) was utilized, calculating also the residual CD intensity values upon considering increasing number of the chiral complexes formed in the reaction mixture. This procedure reveals the minimum number of species necessary for satisfactory description of the experiment data, by achieving the random noise level of the residual intensity on the successive increase of the number of species. Figure 1b undoubtedly demonstrated a systematic and significant deviation from the noise level even after including two complexes in the calculation. Therefore, a third complex species needed to be identified in the Cu(II)—monensinate A system.

Looking at the course of the CD intensity at selected wavelengths upon increasing metal‐to‐ligand molar ratio, two breakpoints can be observed: the first near 0.5 and the second one near 1.0 equivalent Cu(II) content relative to the monensinate A amount in the reaction mixture (Figure 1c). This indicates the formation of bis and mono complexes, identical to those detected with the earlier studied divalent metal ions, but also suggests that the third species formed at metal ion excess should have a higher metal ion content, such as a dinuclear complex with [M₂L]³⁺ composition. Mon^–, however, has no further preferred binding sites for divalent metal ions. One may hypothesize that the second Cu(II) is bound in the cavity of the ligand, but in this case, we would expect a drastic change in the CD spectra during this process, which is not observed. It is also questionable whether the two doubly charged metal ions could exist in close vicinity in this complex. The repulsion would decrease the stability of the complex, while the CD signal did not change significantly at metal ion excess over 2, suggesting the formation of a quite stable complex containing two metal ions.

Another explanation would be the formation of an intermediate product with a composition of [M₂L₃]⁺. By assuming such a species, the two major complexes formed at low and high metal ion content would be the [ML₂] and [ML]⁺ similarly to the previous divalent metal ion—Mon^– systems. Nevertheless, this would not conform with the observations in Figure 1, and the structure of the [M₂L₃]⁺ dinuclear complex is undetermined.

To explore whether the counterion of the applied Cu(II) salt has a role in the complex formation, such as a bridging ligand, we have also carried out CD experiments using CuCl₂ as a titrant. The results are shown in Figure S2a. Similar to the perchlorate‐containing system, at least three optically active complex species were observed with the chloride salt according to the MRA analysis of the CD spectra (Figure S2b). The spectra recorded close to the 1:2 metal‐to‐ligand ratio are very similar to those obtained in the perchlorate‐containing system, suggesting that in this condition the major species is the neutral [ML₂] complex, in which no interaction with the counterions is expected. On increase of the metal salt content, the charged [ML]⁺ mono complex and the possible dinuclear species appear, which need to be neutralized by the counterions in methanol. Indeed, the procession of the curves is somewhat different from the Cu(II)—Mon^–—ClO₄ ^– system, indicating that the interaction with different counterions caused slight changes in the local environment of Cu(II). The stronger interaction of Cu(II) with chloride than with perchlorate ions also affects the equilibria in solution due to the competition of the counterions with Mon^–. It also needs to be mentioned that the signal to noise ratio decreased in parallel with the increasing Cu(II) concentration, due to the high absorbance of the chloride salt in the NIR wavelength region. These facts prevented the measurements above 2 equivalents of CuCl₂ relative to the Mon^– concentration, as well as the quantitative evaluation of the experimental data.

The results of the light absorption measurements, carried out in parallel with CD experiments are shown in Figure S3. The evolution of an absorption peak with a maximum at ∼ 713 nm was observed independently of the counterion up to 0.5 equivalent of the added metal ion, which is then shifted to ∼ 800 nm at up to twofold metal ion content without a large change of the absorbance value (associated with a decrease of the molar absorbance of the complexes from [ML₂] to [ML]⁺ or [M₂L]³⁺). Upon further increase of the metal ion excess, the absorption maximum has gradually shifted to higher wavelengths (∼ 830 nm for ClO₄ ^– and ∼ 900 nm for Cl^– containing systems) and the spectral intensity significantly increased to become more and more similar to those recorded for the methanolic solution of the free metal salts without Mon^–. These observations indicated the dominance of the counterion coordination in the complex formed at metal ion excess. The very high absorption of the chloride salt prevented the titration above 2 equivalents of the metal ion. At the same time, the perchlorate salt exerted much lower absorbance than the chloride salt, but in this system a slight precipitation occurred due to the low solubility of the tetramethyl ammonium perchlorate, which did not disturb the CD measurements, but clearly increased the baseline in the absorption measurements. Furthermore, the absorption spectra were unstructured, and therefore, no detailed analysis was carried out.

2.2. EPR Spectroscopic Measurements in the Cu(II)—Monensinate System

To get a closer insight into this intriguing system, we applied EPR spectroscopy as an independent method. One of the significant differences between the EPR and CD methods is that the methanolic solutions of the inorganic metal salts are detectable by EPR, while they are invisible for CD. As a consequence, the appearance of an extra species (the Cu(II) salt) in comparison to the CD spectral series was expected in EPR experiments. The spectra of the ligand‐free Cu(ClO₄)₂ and CuCl₂ methanolic solutions were also recorded as a reference to distinguish the signals attributed to the pure salts and those derived from different Cu(II) monensinate complexes. Figure 2 shows the EPR spectra recorded at various metal‐to‐ligand molar ratios in the similar concentration range to that applied in CD spectroscopy. The spectra obtained in methanolic solutions at room temperature are unstructured due to the tendency for averaging of spectral features in solutions (Figure 2a). The increased motion of coordinated Cu(II) ions in solutions leads to a signal of rather isotropic shape. The hyperfine splitting is usually observable in the room temperature spectra above A₀ ∼ 40 G (1 G = 10^{– 4} T) depending on the line widths [36, 37, 38, 39]. Here, the unresolved quadruplet pattern is a consequence of the small A₀ coupling constants (see Table 1 below) characteristic for the Cu(II) complexes containing weak oxygen donors [39, 40, 41, 42, 43]. On the other hand, the spectra recorded in frozen solutions, exhibit visible hyperfine structure resulting in four separate lines in the parallel region for each detected Cu(II) complex (Figure 2b).

FIGURE 2.

EPR spectral series of Cu(ClO₄)₂ monensinate A system at different metal‐to‐ligand molar ratios recorded in methanolic solution at (a) room temperature (298 K) and (b) in frozen solution at 100 K. (c) The parallel region of the EPR spectra shown in panel (b) with tentative assignment of the quadruplets related to each single EPR‐active Cu(II) center.

TABLE 1.

The EPR parameters of the Cu(II) complexes calculated from the spectra recorded under various conditions. The hyperfine coupling constants refer to the ⁶³Cu isotope. The coupling constants are in gauss (G) units (1 G = 10^{– 4} T), while they are provided in MHz as well (values in parenthesis). The estimated error for g is 0.001, while for A it falls between 0.1 and 0.5 G.

	g⊥	A⊥	g∥	A∥	g0	A0
Cu(ClO4)2
[ML2]	2.080	11.5 (33.5)	2.384	128.4 (428.4)	2.185	40.6 (124.2)
[ML]+	2.081	12.3 (35.8)	2.392	125.5 (420.2)	2.192	35.0 (107.4)
M(II)	2.087	12.8 (37.4)	2.426	109.0 (370.1)	2.212	23.4 (72.4)
CuCl2
[ML2]	2.080	10.3 (30.0)	2.380	129.4 (431.0)	2.178	41.2 (125.6)
[ML]+	2.082	13.5 (39.3)	2.392	125.0 (418.5)	2.183	34.2 (104.5)
M(II)	2.096	17.7 (51.9)	2.427	103.8 (352.6)	2.187	34.2 (104.7)

As seen in Figure 2c, the parallel regions of the spectra of both Cu(II):Mon^– 1:2 and 1:1 mixtures show one dominant quadruplet due to the bis ([CuL₂]—blue dashed line) and mono ([CuL]⁺—red dashed line) coordination species, respectively. In addition, accompanying lines of low intensity are also visible—the spectrum of 1:2 mixture contains signals assigned to the [CuL]⁺ species, while that of the 1:1 mixture possesses the signal owing to the Cu(ClO₄)₂ salt (black dashed line). It should be noted that their intensity is about 10 times lower than the intensity of the main signals. Unlike the above observations, the spectrum of 2:1 mixture shows two quadruplets with comparable intensities, seemingly attributed to the [CuL]⁺ mono complex and the Cu(II) salt.

By a qualitative assignment of each quadruplet belonging to the various Cu(II) centers, we could thus identify three different species, with Cu(ClO₄)₂ being one of these, that is, only two Cu(II) complexes of Mon^– were found by EPR spectroscopy in contrast to the CD results. This contradiction may be explained by considering the formation of an EPR‐silent complex, in which two Cu(II) ions interact antiferromagnetically with each other. There are, however, several arguments against this suggestion:

• The EPR signal intensity shall either completely disappear or at least decrease significantly in case of the formation of an EPR‐silent complex in comparison to the systems containing mainly mononuclear complexes. The contribution of an intermediate [M₂L₃]⁺ dinuclear species or an eventually formed dimer with [M₂L₂]²⁺ composition is expected to affect mainly the spectrum recorded at 1:1 metal‐to‐ligand ratio (an intermediate state within the spectral series shown in Figure 2), but no drastic intensity decrease was observed in this spectrum.

• It is also worthwhile mentioning that the room temperature EPR spectrum of the [ML₂] complex including two bulky ligand molecules has lower intensity than the other recorded spectra at the same total Cu(II) concentrations. This complex seems to be in the size range that is large enough to cause the increase of the rotational correlation time (being proportional to the molecular radius of the equivalent rotating sphere) and as a consequence, the broadening of the lines in the spectrum [44]. As no intensity decrease was observed at 1:1 metal‐to‐ligand ratio, the formation of similarly large [M₂L₂]²⁺ or [M₂L₃]⁺ dinuclear species—even if these were not EPR‐silent—is unlikely.

• If an EPR silent [M₂L]³⁺ dinuclear species was formed at metal ion excess, we would expect a significant intensity decrease in the spectrum recorded at increased metal‐to‐ligand ratio, that is, in the 2:1 spectrum, which could not be observed, as well. As mentioned above, this spectrum, obtained in frozen solution, exhibits two signals, which can be qualitatively assigned to Cu(ClO₄)₂ with a nearly 50%‐50% ratio related to that of the signal of [ML]⁺ complex (being a major species at 1:1 metal‐to‐ligand ratio, see Figure 2c). This would imply that the metal ion excess does not participate in further complex formation process, which is a contradiction with the CD spectroscopic results.

We have evaluated the EPR spectra using the EPR program developed by Rockenbauer et al., to obtain quantitative results [45]. The calculations provided an acceptable fit of the experimental curves, as shown in Figure 3. The derived individual EPR spectra are shown in Figure S4, and the EPR parameters for each mononuclear complex species obtained from the fit of the experimental data are collected in Table 1. Based on these values, the spectra of the metal salts can clearly be distinguished from the Mon^– complexes in frozen solutions by their significantly smaller A and larger g parameters. The obtained values are in agreement with those published in the literature for Cu(II) solvates in various solutions. The published g₀ and A₀ values for Cu(ClO₄)₂ e.g., vary in the 2.160–2.199 and 25–39 G ranges, while the g_|| and A_|| in the 2.370–2.428 and 95.5–118.3 G, respectively [39, 40, 42, 46]. The Cu(II) centers in the solvates are generally assumed to possess an axially elongated octahedral geometry, although there is a debate on the number of the coordinated solvent molecules [47, 48].

FIGURE 3.

Fit of the experimental EPR spectra of Cu(II)—monensinate A systems at different metal‐to‐ligand molar ratios recorded in methanolic solution at (a),(c) room temperature (298 K) and (b),(d) in frozen solution at 100 K. The series of the spectra were obtained with Cu(ClO₄)₂ (a),(b) and with CuCl₂ (c),(d). The colored symbols represent the experimental points, while the fit is shown by the black dashed lines.

No significant difference was detected between the parameters of Mon^– complexes formed with either perchlorate or chloride Cu(II) salts. The EPR parameters assigned to the [CuL₂] and [CuL]⁺ complexes are similar to each other due to the oxygen donor atom surroundings provided by Mon^– and methanol. The similarity of the g _av values, calculated from anisotropic low temperature spectra also shows similar type of coordination in both the ClO₄ ^– and the Cl^– containing systems: g _av: 2.182 [CuL₂], 2.184 [CuL]⁺, and g _av: 2.180 [CuL₂], 2.185 [CuL]⁺, respectively. The closeness of the g _av parameters from the spectra recorded in the frozen state and the g₀ from the spectra of the solutions at room temperature revealed that the Cu(II) ions preserved their coordination mode upon the temperature change.

Fitting of the spectra recorded at Cu(II):Mon^– 2:1 molar ratio assuming the presence of [ML]⁺ and the free metal salt as explained above, provided parameters being close to those determined for the [ML]⁺ from the Cu(II):Mon^– 1:1 system and for the pure copper(II) salts, but combining these data slightly increased the estimated error of EPR parameters (e.g., from ∼ 0.5 G to ∼ 3 G for the A_||). The ratio of the two components ([ML]⁺ and the metal salts) was ∼ 1:1, as expected from the composition of the system.

We have recorded a new series of EPR spectra extending the range of the titration toward copper(II) excess to better understand the reason for the increased error in parameters and to find the missing species identified in the CD experiments. If an EPR silent [M₂L]³⁺ complex was formed with moderate stability, the gradual disappearance of the signal related to [ML]⁺ would be expected while the intensity of the signal of the free copper(II) salt would increase as the metal ion excess—not bound to monensinate A—is increasing. The series of these spectra is shown in Figure S5. It can clearly be seen that the signal closely related to the [ML]⁺ complex did not disappear even at fourfold excess of the metal ion. The integral of the EPR signals also supported that no spins have been lost in the complex formation process. Furthermore, the quantitative evaluation of the spectra considering the presence of [ML]⁺ and the Cu(II) salt in varying molar ratios revealed the calculated ratios of the two components which agreed well with the initial composition of the system, that is, 1:1 at Cu(II):Mon^– 2:1 concentration ratio (suggesting one copper(II) bound in an [ML]⁺‐type complex and one being unbound) and 1:3 at Cu(II):Mon^– 4:1 concentration ratio (suggesting one copper(II) bound in an [ML]⁺‐type complex and three being unbound). By inspecting the obtained parameters, we have noticed that there is a clear jump in their values from the parameters of the pure species (see e.g., the A_|| values with filled symbols in Figure 4a, b) compared to those parameters calculated in the systems, where the substantial formation of [M₂L]³⁺ complex may occur (empty symbols in Figure 4a, b).

FIGURE 4.

The A_|| parameters calculated from the EPR spectra for the Cu(ClO₄)₂ salt (a) and the [CuL]⁺ complex (b) at increasing metal‐to‐ligand molar ratio. We have considered the presence of these two species, allowing their ratio and parameters to vary in the fitting procedure of the spectra recorded in the systems with metal ion excess. The point at the nominally “0 ratio” in panel (a) denoted by a filled red triangle was obtained from the experiment containing only the free metal salt.

Figure 4 clearly shows that the difference between the two groups of the parameters, that is, those obtained in the “pure” systems containing mainly the free salt or the [ML]⁺ complex and those obtained for the systems containing metal ion excess is larger than the expected error shown in Table 1. Nevertheless, the parameters calculated for the systems with metal ion excess do not differ to an extent which would necessitate the assumption of a substantially different environment. Therefore, we have reevaluated the EPR spectra of the new series by including two new environments for both Cu(ClO₄)₂ and the [CuL]⁺ complex, representing the “independent type” and the “[M₂L]³⁺‐type” environments in the hypothetical dinuclear complex, the structure of which is suggested in Scheme 2.

SCHEME 2.

Suggested structural formulae of the [ML₂], [ML]ClO₄, and [M₂L](ClO₄)₃ complexes of Cu(II) and monensinate A.

The obtained EPR parameters shown in Table 2 support the hypothetic structure of the [Cu₂L]³⁺ species. The two copper(II) ions within the new dinuclear species form a construct in which the positive charges of the metal ions are neutralized by the complexed perchlorate/chloride anions. One or two of these counter ions may also serve as bridging ligand(s) between the metal ions or eventually the solvent can also participate as the second bridging ligand. This association might be promoted by the solvent properties, which clearly does not favor the presence of free solvated ions. However, the interaction between the two copper(II) centers must be weak enough to allow them to behave magnetically separately in the EPR experiments.

TABLE 2.

The EPR parameters of Cu(ClO₄)₂ and the [CuL]⁺ complex in two different environments. The hyperfine coupling constants refer to the ⁶³Cu isotope. The coupling constants are in gauss (G) units (1 G = 10^{− 4} T), while they are provided in MHz as well (values in parenthesis). The estimated error for g is 0.001, while for A it falls between 0.1 and 0.5 G.

	g⊥	A⊥	g∥	A∥
Cu(ClO4)2
free salt	2.087	12.3 (35.9)	2.427	108.3 (367.9)
inside [M2L]3+	2.100	22.5 (66.1)	2.427	104.7 (355.7)
[ML]+
[ML]+	2.082	13.6 (39.6)	2.393	124.9 (418.3)
inside [M2L]3+	2.082	11.7 (34.1)	2.397	119.1 (399.6)

We have initiated geometry optimization of the dinuclear complex but failed to reach convergence in calculations (the details are described in the Electronic Supplementary Information section as an Appendix). The preliminary data indicated a Cu···Cu distance of ∼3.25 Å for the Cl⁻—bridged Cu(II) centers. Although this is not a final result, this distance is in the expected range based on the crystal structures of various di‐ or oligonuclear copper(II) complexes in which perchlorate [49, 50, 51] or chloride ions [52, 53, 54, 55, 56] serve as bridges. The Cu···Cu distance in these complexes vary between ∼2–4 Å, which would imply that these species are EPR‐inactive. At the same time, upon dissolving these crystals and recording their EPR spectra, the dissociation of the complexes was assumed, since the spectra reflected magnetically independent copper(II) centers in most of the experiments. Carboxylate bridged dimeric species behave similarly: the EPR spectra recorded in frozen solutions still manifested dimeric signals, while independent copper(II) centers were observed at room temperature [57]. Recently, a dinuclear copper(II) complex with either a mono‐chloride or a mono‐perchlorate bridge was shown to keep its dimeric nature in solution by the careful investigation of the rotational correlation time while retaining the independent nature of the copper(II) centers, similarly to the [Cu₂L]³⁺ complex in our system [58]. Based on these data, we suggest that the presence of the magnetically independent Cu(II) centers is most probably attributed to a particular local structure/geometry of the metal ions preventing the antiferromagnetic coupling. Comprehensive theoretical calculations are needed to explore this phenomenon in the future. Nevertheless, due to the unusual behavior of the dinuclear species based on the present state‐of‐the‐art of EPR spectroscopy of copper(II) complexes, we cannot exclude the formation of a complex, in which the two metal ions are bound to the two termini of the monensinate ligand (see Scheme 2).

2.3. Mass Spectrometric Experiments Indicate the Existence of the [M₂L]³⁺ Complex Species

ESI‐MS experiments were carried out in diluted solutions of the Cu(II)—Mon^– systems at various metal‐to‐ligand ratios to check the feasibility of the above considerations by a further independent method. All the suggested major complexes in Table 1 could be identified for both the Cu(ClO₄)₂ (Table 3) and the CuCl₂ (Table S2) containing reaction mixtures (see in addition Figure S6).

TABLE 3.

The calculated exact monoisotopic masses for various compositions of the Cu(ClO₄)₂ coordination species of monensinate A compared to those masses found in the experiments. (The monoisotopic mass is the exact mass calculated using the mass of the most abundant isotope of each element in that molecule.).

Complex	Composition	Formula	Calc. exact mass	Measured mass
[ML2]	[Cu(Mon)2] + H+ (+ CH3OH)	C73H123O23Cu	1430.7746	1430.7726
[ML]+	[CuMon]+ + ClO4 − + H+ (+ CH3OH)	C37H64O16CuCl	862.3173	862.3150
	[CuMon]+ (+ CH3OH)	C37H63O12Cu	762.3610	762.3612
[M2L]3+	[Cu2Mon]3+ + 2 ClO4 − (+ CH3OH)	C37H63O20Cu2Cl2	1023.1876	1023.1857
[M2L2]2+	[Cu2Mon2]2+ + ClO4 − (+ CH3OH)	C73H122O27Cu2Cl	1591.6449	1591.6446
	[Cu2Mon2]2+ + ClO4 − (+ 2CH3OH)	C74H126O28Cu2Cl	1623.6711	1623.6684

Comparison of the simulated isotopic pattern of the [M₂L]³⁺ complex revealed good agreement with the experimental spectrum (Figure 5). A side reaction was also observed under the conditions of the mass spectrometric experiments. The Cu(II) ions seem to promote the oxidation of the organic ligand in diluted solutions (see more details in Figure S7). Therefore, better agreement was observed with the complex formed with Mon^– lacking two hydrogen atoms (Figure 5a) than with that of the original ligand in Figure 5b. The mass 1025.1832 obviously cannot originate from the original molecule, because the expected mass from it should be 1025.2033. The difference between the two is about 20 ppm, while the measurement accuracy in this experiment is approximately 2 ppm.

FIGURE 5.

Fit of the experimental and simulated ESI mass spectra of Cu(ClO₄)2—monensinate A dinuclear complex recorded in methanolic solution in positive ion mode. The simulation was carried out for the complex including oxidized ligand lacking two hydrogen atoms (a) while in the simulation shown in panel (b), the original ligand molecule was applied.

2.4. Evaluation of the CD Spectra Using the New Model

Based on the above considerations, we have applied a new model in the calculation of the complex stability constants from the CD spectroscopic titrations using the PSEQUAD program [59], including the formation of three complex species with [ML₂], [ML]⁺, and [M₂L]³⁺ compositions. The species distribution diagram was constructed based on the calculated complex formation constants, and it is shown in Figure 6 together with the calculated molar CD spectra of the individual complexes. The similarity of the CD spectra obtained for the [ML]⁺ and [M₂L]³⁺ complexes is in agreement with the suggested structure of the dinuclear species, as the chirality is mostly propagated through its [ML]⁺‐type centre, while the other metal ion is not bound directly within a chiral environment.

FIGURE 6.

(a) Species distribution diagram of the Cu(ClO₄)₂—monensinate A system in methanolic solution. logβ’([ML]⁺) = 8.28 (0.06); logβ’([ML₂] = 12.03 (0.07) and logβ’([M₂L]³⁺) = 12.82 (0.14)—the standard deviations in parenthesis; c(Mon^–) = 10.10 mM. (b) The calculated molar ellipticity spectra for the individual complexes.

It is worth mentioning that the conditional stability constant determined for the [CuL]⁺ complex is much higher than those obtained for the Ni(II), Co(II), and Zn(II) ions studied previously with logβ’([ML]⁺) ∼ 6 [29], while the affinity of Cu(II) toward the second ligand is similar to all these metal ions: logK ₂ ^L’ ∼ 4 (K ₂ ^L’ refers to the ML + L = ML₂ equation). This suggests that the formation of the [ML]⁺ complex of Cu(II) is more preferred than for the above listed divalent transitional metal ions, most probably due to geometrical requirements of the metal ions versus bulkiness of the ligand molecule. On the other hand, the reaction of [CuL]⁺ with the second Cu(II) ion yielded a logK₂ ^M’ value of 4.54 (K ₂ ^M’ refers to the ML + M = M₂L equation) indicating the preference of Cu(II) to form dinuclear complex at metal ion excess in methanolic solution.

3. Experimental Section

3.1. Materials

Monensic acid A (MonH×H₂O) was synthetized as previously reported [16] from sodium monensinate A (MonNa), which was obtained from Biovet Ltd. (Bulgaria). High‐purity analytical grade methanol (VWR, Avantor, Hungary) was applied as a solvent in the experiments. The pure analytical‐grade metal(II) salts (CuCl₂, Cu(ClO₄)₂×6H₂O) were ordered from Sigma‐Aldrich Chemie GmbH (Schnelldorf, Germany). The concentration of MonH was checked by UV‐CD spectroscopic measurement, normalized to a previous standard sample spectrum. To deprotonate the carboxylic group of the ligand, to each MonH sample one equivalent of tetramethylammonium hydroxide (TMAOH) was added, the concentration of which was determined by acid–base titrations. The concentrations of the methanolic metal(II) stock solutions were determined by the complexometric method using ethylenediamine tetraacetic acid.

3.2. UV‐Vis Absorption and CD Spectroscopy

Electronic absorption spectra were recorded on a Shimadzu UV 1800 spectrophotometer in 1 cm quartz cuvette. Circular dichroism spectroscopic measurements were performed using a Jasco J‐1500 spectrophotometer (JASCO Corporation, Tokyo, Japan). During the experiments, a steady flow of nitrogen was maintained. For determining the concentration of monensinate A, a stepwise scanning mode was carried out within the 180–330 nm wavelength range, whereas continuous scanning mode at a speed of 200 nm/min was used between 300 and 1200 nm. Each measurement was performed with a 2‐second response time and a resolution of 1 nm. A 0.2 mm cylindrical quartz cuvette (Hellma, Müllheim, Germany) served for the UV range measurements, while the samples were analyzed using 1 cm standard cuvette in the visible range. Depending on the spectral intensity, 3 to 10 scans were averaged. Calculation of the stability constants, species distribution diagrams, and molar CD spectra of the complexes was carried out by fitting the experimental CD spectra with the PSEQUAD software [59].

3.3. Mathematical Treatment of the CD Experimental Data

The results of the calculations of the stability constants, and molar spectra are largely determined by the appropriate species matrix. Assuming too many species, and as a consequence, an increased number of fitted parameters tends to improve the overall fit, even if these species are present in negligible amounts, but may result in distorted stability constants and molar intensities (e.g., absorbance, ellipticity) if they are included by mistake. To facilitate a more accurate determination of the minimal number of independent species, required to fit the data, we applied a matrix rank analysis (MRA) procedure to the set of the CD spectra [35] recorded in systems containing the same basic components. The principle of the residual curve calculations within MRA procedure was described previously [29]. The essence of the procedure is that the intensity values related to the subsequently included increasing number of species are eliminated in an iterative way, until the residual intensity curve exhibits random behavior, with no systematic deviation from the zero line, meaning that no further species need to be considered.

3.4. EPR Spectroscopy

The registration of the X‐band EPR spectra (100 and 298 K) was performed on a Bruker BioSpin EMXplus10/12 EPR spectrometer (Karlsruhe, Germany). All measurements were carried out at 9.4 GHz frequency of microwave electromagnetic radiation. The EPR simulations were performed using Aniso‐Spin program, Xenon software (Version 1.2.1.iso). The isotropic spectra (recorded at 298 K) were evaluated by the EPR program [45]. The spectra of the detected complexes were characterized by the following parameters: g₀, the hyperfine coupling constant A₀ and the relaxation parameters α, β, and γ referring to the line widths as: W_MI = α + β M_I + γ M_I ², where M_I is the magnetic quantum number of Cu(II) nuclei. The EPR program was also applied for the analysis of the anisotropic EPR spectra, handling the superposition of up to three component curves. We assumed axial or rhombic symmetry of the complexes. Anisotropy of the relaxation parameters α, β, and γ was also taken into consideration. The spectra were calculated as the weighted sum of the curves of complexes of ⁶³Cu(II) and ⁶⁵Cu(II) considering their natural abundances. The noise‐corrected regression parameter R _j computed from the average square deviation between the respective experimental and calculated curves characterized the quality of the fit.

3.5. Electrospray Ionization Mass Spectrometry

HRMS spectra were recorded on a Q Exactive Plus hybrid quadrupole‐orbitrap mass spectrometer (Thermo Scientific, Waltham, MA, USA) equipped with a heated electrospray ionization (HESI‐II) probe that was used in both negative and positive ion mode. Samples were introduced with FIA (flow injection analysis) method, eluent stream (methanol) was provided by an Applied Biosystems 140C syringe pump.

4. Conclusion

The coordination ability of the polyether ionophorous antibiotic monensin A to bind Cu(II) ions was evaluated in methanolic solutions by a set of complementary spectroscopic methods and mass‐spectrometry. The VIS‐ and NIR‐CD spectroscopy evidenced the distinctive character of Cu(II) forming three different chiral species which depend on metal‐to‐ligand molar ratio. Thus, the monensinate A anion is bound in a neutral bis [CuL₂] complex at copper(II) deficit, while at comparable reactant concentrations the mono [CuL]⁺ species prevailed, which proved to be at least two orders of magnitude more stable than those formed with Ni(II), Co(II) or Zn(II). Due to the increased stability of the [CuL]⁺ complex Cu(II) may become a good candidate for a competing agent with the native sodium ion. The observed species at these reaction conditions corroborate well our previous solution studies using monensinate A and colored Co(II), Ni(II) or colorless Mg(II), Ca(II), Zn(II) metal cations. Notably, further increase of Cu(II) in the system led to the formation of dinuclear [Cu₂L]³⁺ complex not observed before. The two metal centers are most likely linked to each other by at least one counter ion originating from the added copper(II) salt (Cu(ClO₄)₂ or CuCl₂). The existence of the dimer in solution was confirmed by EPR spectroscopy. A peculiar property of this complex is that it contains two magnetically independent copper(II) centers, one with a similar coordination environment to [CuL]⁺ and the second—to the free metal salt. The presence of the dinuclear species was further checked by ESI‐MS, becoming the first example of dinuclear monensinate A construct bearing divalent metal cations. The outcome of the present study confirms the potential of monensin to form diverse coordination species, the description of which aids our understanding of the biological activity of the antibiotic. Furthermore, taking into account that PIs are also supposed to become potential anticancer agents [8, 9, 10], it is worth considering to apply them to transport Cu(II) across the cell membrane of cancer cells ultimately leading to cuproptosis, recently identified as a new therapeutic strategy [60, 61].

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting File 1: chem70718‐sup‐0001‐SuppMat.pdf.