Comparative Studies of Structures and Peroxidase-like Activities of Copper(II) and Iron(III) Complexes with an EDTA-Based Phenylene-Macrocycle and Its Acyclic Analogue

Abstract

With the objective of studying the conformational and macrocyclic effects of selected metal chelates on their peroxidase activities, Cu²⁺ and Fe³⁺ complexes were synthesized with a macrocyclic derivative of ethylenediaminetetraacetic acid and o-phenylenediamine (abbreviated as edtaodH₂) and its new open-chain analogue (edtabzH₂). The Fe³⁺ complex of edtaodH₂ has a peroxidase-like activity, whereas the complex of edtabzH₂ does not. The X-ray study of the former shows the formation of a dimeric molecule {[Fe(edtaod)]₂O} in which each metal with an octahedral coordination is overposed over the macrocyclic cavity, as a result of rigid macrocyclic frame, to form an Fe–O–Fe bridge; the exposure of the central metal to the environment facilitates the capture of oxygen to drive the biomimetic activity. The peroxidase-inactive Fe³⁺ complex consists of a mononuclear complex ion [Fe(edtabz)(H₂O)]⁺, the metal ion of which is suited in a distorted pentagonal bipyramid to be protected from environmental oxygen. The copper(II) complexes, which have mononuclear structures with high thermodynamic stability compared with the iron(III) complexes, show no peroxidase activity. The steric effects play a fundamental role in the biomimetic activity.

1. Introduction

The studies of metal–ligand interaction in coordination compounds have constructed an important research area because the interaction causes and controls the chemical and biological activities of metal centers. From this viewpoint, various types of coordination compounds have been designed and their applications have been exploited in many fields.¹⁻³ In the human organism, metal ions encompass several crucial functions through metalloproteins. Among the metalloenzymes that have attracted attention in relation to the development of biomimetic agents, our special interest is directed to peroxidases (POx), which are heme proteins that catalyze the degradation of H₂O₂ when the human body defends itself against oxidative damage by superoxide anions. Although the active site of POx contains iron(III) coordination centers, the evaluation of similar biomimetic activities has been reported for other transition metal complexes such as copper(II) complexes, to provide interesting results.^4,5 The activity as a biomimetic agent depends, among other factors, on the conformation of the metal–ligand sphere because the oxidation state of the metal is changed during the catalysis.⁶ In our research group, we have developed metal complexes with macrocyclic and acyclic receptors derived from EDTA (ethylenediaminetetraacetic acid) with different topologies.⁷⁻¹⁵ Some of them have coordination modes resembling those found in binuclear metalloproteins. As background of the present work, significant biomimetic capacities of superoxide dismutase and POx (peroxidase) have been found for the binuclear complexes of the EDTA-derived ligands with copper(II)¹⁰ and iron(III),⁸ respectively. Likewise, they present high antioxidant capacity without harmful effects on peripheral blood mononuclear cells in vitro.^8,10 These promising results suggest that the family of the ligands provides metal complexes suitable for deepening the knowledge of the structural and physicochemical factors that control their biomimetic activity; the major conceivable factors include the structural type of the ligand, size of the cavity, and number and nature of the metal in the complexes, all of which can be readily controlled by modifying EDTA with selected groups. In the present work, we have chosen a macrocycle consisting of EDTA and aromatic units (edtaodH₂ in Scheme 1), for the following reasons: the phenylene group integrated in the macrocyclic frame results in the high rigidity which strictly defines the coordination geometry around the central metal; moreover, electronic interaction between the coordinated metal ion and aromatic group favors a change in the oxidation state that may occur in the intermediates of the peroxidase process. To study the effect of the steric effect due to the macrocyclic frame, a new open-chain ligand (edtabzH₂ in Scheme 1) has been synthesized, which consists of a chelating EDTA unit linked with phenyl group in a sterically less-restricted manner. Their Fe³⁺ and Cu²⁺ complexes have been subjected to the tests of peroxidase-like activity; only the iron(III) complex of edtaodH₂ is active. In order to study the correlation of the biomimetic activity with the structural and thermodynamic properties, the metal complexes have been characterized by X-ray diffraction, calorimetric analysis, and UV-spectrometry. On the basis of the results of these studies, we can verify that a greater thermodynamic stability of a complex is not necessarily translated into a better biomimetic activity, but the steric effects may play a more fundamental role in the biomimetic activity.

Scheme 1. EDTA-Derived Ligands and Abbreviation.

2. Results and Discussion

2.1. Characterization of edtabzH₂

Reaction of EDTA dianhydride with aniline gave acyclic ligand edtabzH₂ (Scheme 1), which was characterized by ¹H and ¹³C NMR spectroscopy, Fourier-transform infrared (FTIR) spectroscopy, and electrospray mass spectrometry as described in the Experimental Section. The water solubility is high enough for potentiometric and NMR studies in the pH range 2–12. The logarithmic protonation constants of edtabz^2– were determined by potentiometry in 0.1 M KCl at 25 °C as log[LH]/[L][H] = 6.23 (±0.01) and log[LH₂]/[LH][H] = 3.64 (±0.02); Figure 1 presents the calculated species distribution. These values are close to the corresponding values 6.07 and 3.7 reported for edtaod^2– macrocycle.¹² The pD dependence of the ¹H NMR chemical shifts observed for edtabzH₂ is shown in Figure 1. The protons labeled a, b, and c in Scheme 1 shift downfield with decreasing pD in the range 5.5–8.5, indicating that the amino nitrogen atoms are protonated. Only the proton b shifts downfield with decreasing pD from 5 to 2, indicating protonation at carboxylate oxygen. The NMR δ of proton i is expressed by the following function of pD:

1

where β_n is the overall protonation constant of n-fold protonation in D₂O media, 10^–n(pD) gives D⁺ concentration, δ_i,0 is the chemical shift of proton i in the completely deprotonated species L^2–, and δ_i,n is the chemical shift in the n-fold protonated species. The shifts of all proton signals are well reproduced by eq 1 with the values of β_n, δ_i,0, and δ_i,n presented in the legend of Figure 1. The obtained deuteration constants, log K_D1 = 6.74 (±0.02) and log K_D2 = 3.26 (±0.02), are similar to the corresponding values 6.94 and 3.89 reported for edtaodH₂.

Figure 1.

(Top) Species distribution diagram of edtabzH₂ determined by potentiometry and (bottom) ¹H NMR δ values observed for edtabzH₂ at different pD values (for labels of protons, see Scheme 1). The solid lines are calculated using eq 1, with the following values of the parameters: log β₁ = 6.74 and log β₂ = 10.00, δ_a,0 = 2.89, δ_a,1 = 3.37, δ_a,2 = 3.42, δ_b, = 3.32, δ_b,1 = 3.66, δ_b,2 = 3.92, δ_c,0 = 3.45, δ_c,1 = 3.98, δ_c,2 = 4.04, δ_p,0 = 7.21, δ_p,1 = 7.22, δ_p,2 = 7.23, δ_m,0 = 7.37, δ_m,1 = 7.32, δ_m,2 = 7.35, δ_o,0 = 7.37, δ_o,1 = 7.36, δ_o,2 = 7.39.

The X-ray structure presented in Figure 2 shows that edtabzH₂ is crystallized as zwitterion dihydrate, in which one of the carboxyl hydrogen atoms is transferred to the adjacent amino group. The aromatic substituents are pointed to the same direction; this orientation is attributed to crystal packing effects because the aromatic groups are ca. 5.587 Å distant. The relevant crystallographic information and final refinement parameters are displayed in Table 1. Selected bond distances, bond angles, and torsional angles are given in Tables S2–S7 (see the Supporting Information).

Figure 2.

Perspective view of the zwitterionic molecular structure of edtabzH₂. Atoms are drawn at the 30% probability level.

Table 1. Selected Crystallographic Data for {[Fe(edtaod)]₂O}·4.4H₂O, edtabzH₂·2H₂O, and [Fe(edtabz)]NO₃·5.5H₂O.

	{[Fe(edtaod)]2O}·4.4H2O	[edtabzH2]·2H2O	[Fe(edtabz)][NO3]·5.5H2O
formula	C32H36Fe2N8O13, 4.4H2O	C22H26N4O6, 2H2O	2(C22H26FeN5O10), 8(O), 3H2O
MW (g mol–1)	931.66	478.50	1334.70
T (K)	100	293	125
space group	Pbca	P212121	P21/c
radiation	Mo Kα	Mo Kα	Cu Kα
a (Å)	18.0203(18)	7.3272(5)	12.1188(8)
b (Å)	18.028(2)	13.6323(10)	18.1908(12)
c (Å)	22.855(3)	23.8535(18)	14.0536(9)
α (deg)	90	90	90
β (deg)	90	90	105.176(2)
γ (deg)	90	90	90
V (Å3)	7424.9(14)	2382.6(3)	2990.1(3)
Z	8	4	2
μ (mm–1)	0.873	0.102	4.759
ρcalcd (g cm–3)	1.667	1.334	1.482
R1 (Fo > 4σFo)	0.0301	0.0778	0.0394
wR2 (all data)	0.0721	0.1468	0.1059
GOF	1.095	1.274	1.061
REFCODE/CCDC	1920517	1920512	1920516

2.2. Cu²⁺ and Fe³⁺ Complexation in Aqueous Solution

Coordination of edtabz^2– and edtaod^2– with Cu²⁺ and Fe³⁺ ions in solution was studied at pH 3.0 by UV–vis spectrometric titrations and ITC. At this pH, the two ligands are in equilibrium between LH^– and LH₂; the amino group is protonated completely, and the carboxylate group is protonated partially (Figure 1). Despite the mixed protonation state, the pH 3.0 was selected to prevent the formation of Fe(OH)₃ so that the complexation of the two metals can be compared. Figure 3 shows the near-UV spectra in the presence of Cu²⁺ or Fe³⁺ ion at different concentrations. All complexes exhibit intense ligand-to-metal charge-transfer transition in the range 280–375 nm. The absorbance of every new band increases almost linearly with the total metal concentration up to the ratio [M]_t/[L]_t ≈ 1. Larger changes in the charge-transfer bands occur in the formation of Fe³⁺ complexes. Every titration curve is interpreted by the metal-to-ligand ratio of 1:1, and it cannot be reproduced by other stoichiometry. The conditional stability constants were determined using the AFFINIMETER program; the results are presented in Table 2. The constants are ranged from 10⁵ and 10⁶ M^–1 (M = mol dm^–3), and higher stability is found for the Cu²⁺ complexes.

Figure 3.

UV–vis spectra of edtabzH₂ and edtaodH₂ in titration with copper(II) and iron(III) nitrates at pH 3.0: (A) Fe³⁺–edtabz^2–, (B) Cu²⁺–edtabz^2– (C) Fe³⁺–edtaod^2–, and (D) Cu²⁺–edtaod^2–. The insets present changes in absorbance at 246 nm for M–edtabz and at 302 nm for M–edtaod: the solid line is the fitting curve generated by program AFFINIMETER.

Table 2. Thermodynamic Parameters Determined by ITC for the 1:1 Complexation of edtaodH₂ and edtabzH₂ with Copper(II) and Iron(III) Nitrates in Potassium Hydrogen Phthalate Buffer (pH 3) and at 298.15 K—Stability Constant (K_ITC), Standard Free Energy (ΔG°), Enthalpy (ΔH°) and Entropy Change (TΔS°), and Conditional Stability Constant (K_UV–vis) Determined by UV–Vis Titration at the Same pH.

	ITC					UV–vis
complex	n	KITC (M–1)	ΔG° (kcal/mol)	ΔH° (kcal/mol)	–TΔS° (kcal/mol)	KUV–vis (M–1)
[Fe(edtabz)]+	1.04 ± 0.01	1.20 (±0.08) × 107	–9.66 ± 0.11	–2.55 ± 0.01	–7.11	9.52 (±1.12) × 105
[Cu(edtabz)]	1.10 ± 0.02	5.19 (±0.81) × 107	–10.52 ± 0.10	–4.86 ± 0.08	–5.67	3.45 (±0.98) × 106
[Fe(edtaod)]+	1.09 ± 0.03	8.75 (±0.52) × 104	–6.74 ± 0.14	1.49 ± 0.03	–8.23	1.16 (±0.048) × 105
[Cu(edtaod)]	1.10 ± 0.09	2.64 (±0.19) × 106	–8.76 ± 0.06	–1.58 ± 0.01	–7.18	1.34 (±0.26) × 106

The coordination of edtabzH₂ with copper in basic media was also studied at pH 8.5 at which the L^2– species is present at almost 100% level (the observed spectra are included in the Supporting Information). A large hypochromic effect is observed for the 246 nm band, and a hyperchromic effect is observed for the CT band; both changes in absorbance are linear up to [M]_t/[L]_t = 1. This greater spectral change is due to that fact that all donor atoms in L^2– can form coordinate bonds without proton dissociation. For the same reason, the conditional stability constant is as high as the order of 10¹³ M^–1.

The curves obtained for the isothermal titrations of edtabzH₂ with Cu²⁺ and Fe³⁺ ions are presented in the panels A and B of Figure 4, respectively (the curves for edtaodH₂ are given in Supporting Information). The raw data of both (A) Cu²⁺–edtabz^2– and (B) Fe³⁺–edtabz^2– systems present exothermic peaks, which indicate the saturation of the association until small peaks appear due to heat generated by dilution in the last injections. The lower panel (C) of Figure 4 presents the integrated data. The thermal change of copper complexation is steeper with the higher plateau than is the iron complexation, suggesting that the former is accompanied by the higher affinity and greater enthalpic contribution. Thermal parameters determined by curve fittings are summarized in Table 2. The n value of ≈1 is consistent with the 1:1 stoichiometry determined by the UV–vis titrations. It is interesting that, in the acidic environment, both the edtabz^2– complexes have comparable affinities with ΔG around −10 kcal/mol, while [Cu(edtabz)] bears almost double the ΔH value of the Fe complex (i.e., the former is driven by the entropy effect completely). The stability constants determined by the calorimetry are higher than those found by the UV–vis technique. The ITC technique detects all of the thermal processes involved in the formation of the complex. On the other hand, the spectrometric technique traces only the spectral change caused by competition between the complexation and protonation to determine a conditional constant, which is usually smaller than the proper constant. Despite the difference, the relative stabilities determined by the same techniques under identical conditions can be compared and evidently indicate that the Cu²⁺ complexes are more stable than the corresponding Fe³⁺ complexes and the acyclic ligand forms more stable metal complexes.

Figure 4.

Microcalorimetric titrations of edtabzH₂ with Cu(NO₃)₂ and Fe(NO₃)₃ in potassium hydrogen phthalate buffer at 0.1 M, pH 3.0, and T 298.15 K: the raw data of the ITC experiments for the reaction systems (A) Cu²⁺–edtabz^2– and (B) Fe³⁺–edtabz^2–; (C) integrated data and fitting curves with AFFINIMETER (rhomboids for Cu and circles for Fe).

2.3. Characterization in Solid

The copper(II) and iron(III) complexes with the ligands edtaodH₂ and edtabzH₂ were isolated in satisfactory yields. The structures of the metal complexes have been established by various techniques. Some informative FTIR bands of the ligands and their Cu²⁺ and Fe³⁺ complexes are given in Table 3. The characteristic C=O stretching band of carboxylic acid in the free ligand appears around 1700 cm^–1 and moves to a lower wavenumber side by at least 30 cm^–1 upon complex formation. This band shift indicates that the carboxylate groups participate in the formation of the complexes. Other bands that are displaced to smaller wavenumber sides upon complexation are the vibrational bands of amide C=O and C−N bonds, which indicate that the amide groups are also involved in the coordination with the metal ions. The X-ray photoelectron spectra (XPS) are shown in Supporting Information. The Fe³⁺ complexes of edtaod^2– and edtabz^2– exhibit 2p_3/2 core-electron peaks at binding energies of 709.9 and 710.5 eV, respectively; the energies are consistent with the trivalent state. Each peak is accompanied by a broad, intense peak at 4–5 eV higher energy. This additional peak might be due to multiplet splitting caused by five electrons in the 3d level, but they are normally difficult to observe clearly because the net effect would broaden the 2p_3/2 and 2p_1/2 peaks. More viable is a shake-up satellite.¹⁶ In the case of the copper(II) complexes, a 2p_3/2 peak is observed at 932.7 and 932.8 eV for the complexes of edtaod^2– and edtabz^2–, respectively. Both peaks are accompanied by shake-up satellites which appear characteristically of the 3d⁹ open-shell in Cu²⁺ ion.

Table 3. IR and XPS Data of edtabzH₂ and edtaodH₂, and Their Cu²⁺ and Fe³⁺ Complexes^a.

	IR (cm–1)					XPS (eV)
ligand/complex	νN–H est	νCOOH	νCONH(I)	νCONH(II)	νC–N	peak 1	peak 2	peak 3	peak 4
edtabzH2	3273	1700	1611	1548	1219
[Cu(edtabz)]	3318	1670	1598	1544	1103	932.8	940.5	944.4	952.7
[Fe(edtabz)]+	3277	1612	1568	1495	1111	710.5	715.3	723.4	NV
edtaodH2	3261	1692	1659	1509	1160
[Cu(edtaod)]	3229	1643	1593	1499	1090	932.7	939.0	943.8	952.6
[Fe(edtaod)]+	3239	1669	1612	1498	1084	709.9	714.5	723.0	729.3

^aNV = no value obtained for that peak.

2.3.1. X-ray Diffraction Analyses

The most relevant crystallographic data of {[Fe(edtaod)]₂O} are summarized in Table 1. Figure 5 shows the molecular structure in which two Fe(edtaod) chelates labeled A and B are bridged by an oxygen atom. As seen from the perspective views given for each chelate in Figure 6, the iron(III) ions are each embedded in distorted octahedral coordination environments. In the equatorial plane, the Fe1/Fe2 atoms are coordinated by atoms O2A/O2B, N2A/N2B, and N3A/N3B from the ligand molecules and the bridging oxygen atom O9 so that a distorted square plane is generated around each central metal ion. The distortion is seen from the O–Fe–O, O–Fe–N and N–Fe–N bond angles formed between neighboring atoms in the plane; the angles range from 77.50(5) to 104.15(5)° (Table 4). The axial positions of each square plane are occupied by oxygen atoms from the second carboxylate group (O4A/O4B) and the amide group (O1A/O1B) to construct an elongated distorted octahedron. As expected, the Fe–O_carboxylate bond distances [2.0094(12) and 2.0182(12) Å] are shorter than the Fe–O_amide distances [2.0184(12) and 2.0358(12) Å]. The Fe–N distances are in the range of 2.2238(15) to 2.3012(14) Å. Among the Fe–O and Fe–N distances, the shortest is found for Fe–O9 with 1.7919(12) and 1.7930(12) Å, which are comparable to those found in other octahedral complexes of chelating N- and O-donor ligands with Fe–O–Fe bridging. The interatomic distance between the bridged Fe(III) ions is 3.4692(7) Å, which falls well within the normal range observed for μ-oxo monobridged Fe(III) dimers.¹⁷ The Fe(III)–O–Fe(III) angle is 150.80(7)°. The crystallographic structures of the ligand edtaodH₂ and its copper complex [Cu(edtaod)]·H₂O have been reported already.¹¹ Both metal complexes of this ligand crystallize with the same space group (Pbca), but there are major differences in the coordination geometries. In [Cu(edtaod)], the octahedral coordination sphere is accomplished with the oxygen atom from a carboxyl group of an adjacent chelate molecule, whereas in {[Fe(edtaod)]₂O}·4.4H₂O the [Fe(edtaod)]⁺ complex ions are coordinated to an oxide anion forming the μ-oxo bridge between them. The connectivity is strengthened further by two intramolecular N–H···O hydrogen bonds formed between one of the amide hydrogens and a carbonyl oxygen atom from one of the metal-coordinated carboxylate groups (N4A–H4A···O2B and N4B–H4B···O2A; not shown in Figure 6).

Figure 5.

Perspective view of the molecular structure of {[Fe(edtaod)]₂O}. Atoms are drawn at the 30% probability level. For clarity, hydrogen atoms are omitted.

Figure 6.

Perspective view of the coordination environment around each iron atom in the dimer molecule {[Fe(edtaod)]₂O}. Atoms are drawn at the 50% probability level.

Table 4. Selected Bond Lengths (Å) and Bond Angles (deg) for {[Fe(edtaod)]₂O}·4.4H₂O.

bond lengths
Fe1			Fe2
Fe1	O1A	2.0184(12)	Fe2	O1B	2.0358(12)
Fe1	O4A	2.0094(12)	Fe2	O4B	2.0182(12)
Fe1	O2A	1.9974(12)	Fe2	O2B	2.0125(12)
Fe1	O9	1.7919(12)	Fe2	O9	1.7930(12)
Fe1	N2A	2.3012(14)	Fe2	N2B	2.2720(15)
Fe1	N3A	2.2707(14)	Fe2	N3B	2.2238(15)

bond angles
Fe1				Fe2
O1A	Fe1	N3A	90.93(5)	O1B	Fe2	N3B	88.21(5)
O1A	Fe1	N2A	76.41(5)	O1B	Fe2	N2B	76.30(5)
O4A	Fe1	O1A	161.71(5)	O4B	Fe2	O1B	159.20(5)
O4A	Fe1	N3A	78.95(5)	O4B	Fe2	N3B	78.94(5)
O4A	Fe1	N2A	86.52(5)	O4B	Fe2	N2B	85.23(5)
O2A	Fe1	O1A	91.04(5)	O9	Fe2	O1B	96.57(5)
O2A	Fe1	O4A	91.66(5)	O9	Fe2	O4B	101.86(5)
O2A	Fe1	N3A	154.20(5)	O9	Fe2	O2B	102.43(5)
O2A	Fe1	N2A	77.50(5)	O9	Fe2	N3B	101.01(5)
O9	Fe1	O1A	93.25(5)	O9	Fe2	N2B	172.86(5)
O9	Fe1	O4A	103.61(5)	O2B	Fe2	O1B	92.41(5)
O9	Fe1	O2A	104.15(5)	O2B	Fe2	O4B	92.94(5)
O9	Fe1	N3A	101.41(5)	O2B	Fe2	N3B	156.32(5)
O9	Fe1	N2A	169.60(5)	O2B	Fe2	N2B	77.75(5)
N3A	Fe1	N2A	77.99(5)	N3B	Fe2	N2B	79.42(5)

The iron(III) complex of edtabz^2– is isolated in the form of the nitrate salt and crystallized as hydrate with composition Fe(edtabz)NO₃·5.5H₂O. This metal complex is mononuclear with the coordination of a water molecule to form [Fe(edtabz) (H₂O)]⁺, in contrast to μ-oxo bridge formation in {[Fe(edtaod)]₂O}·4.4H₂O. The iron(III) center in [Fe(edtabz)(H₂O)]⁺ exhibits seven-fold coordination with the FeO₅N₂ core embedded in a distorted pentagonal-bipyramidal geometry (Figure 7).

Figure 7.

Perspective view of the molecular structure of [Fe(edtabz) (H₂O)]⁺. Atoms are drawn at the 30% probability level.

In the coordination polyhedron, the amino nitrogen atoms and the amide oxygen atoms of the edtabz^2– ligand occupy four positions in the equatorial plane, while the carboxylate oxygen atoms reside at the axial sites. The metal-coordinated water molecule comprises the remaining equatorial site (Figure 7). The O–Fe–O, O–Fe–N and N–Fe–N bond angles formed between neighboring atoms in the equatorial plane range from 70.13(5) to 74.03(5)° (Table 5). The O–Fe–O bond angle between the axial substituents is 160.71(6)°. The Fe–N_amine, Fe–O_amide, Fe–O_carboxylate, and Fe–O_w distances are 2.3276(16)–2.3534(15), 2.1892(13)–2.1477(13), 1.9435(13)–1.9609(13), and 2.0282(14) Å, respectively (Table 5); these distances are similar to the values established for {[Fe(edtaod)]₂O}·4.4H₂O and other structurally related iron(III) complexes reported previously.^18,19 In the crystal structure, adjacent [Fe(edtabz)(H₂O)]⁺ molecules are linked together through strong O–H···O hydrogen-bonding interactions between the Fe–OH₂ water molecules and one of the carboxylate oxygen atoms [OW1···O6 = 2.687(2) Å], yielding dimeric aggregates with a Fe···Fe distance of 5.3181(5) Å. There are no direct intermolecular contacts between the metal chelate and nitrate ions.

Table 5. Selected Bond Lengths (Å) and Bond Angles (deg) for [Fe(edtabz)]NO₃·5.5H₂O.

bond lengths
Fe1	O1	2.1892(13)	Fe1	OW1	2.0282(14)
Fe1	O2	2.1477(13)	Fe1	N1	2.3534(15)
Fe1	O3	1.9609(13)	Fe1	N2	2.3276(16)
Fe1	O5	1.9435(13)

bond angles
O1	Fe1	N1	70.13(5)	O5	Fe1	O2	95.68(6)
O1	Fe1	N2	141.89(5)	O5	Fe1	O3	160.71(6)
O2	Fe1	O1	146.26(5)	O5	Fe1	OW1	104.40(6)
O2	Fe1	N1	143.56(5)	O5	Fe1	N1	85.67(6)
O2	Fe1	N2	70.95(5)	O5	Fe1	N2	78.35(5)
O3	Fe1	O1	96.12(5)	OW1	Fe1	O1	72.71(5)
O3	Fe1	O2	92.06(5)	OW1	Fe1	O2	74.03(5)
O3	Fe1	OW1	94.70(6)	OW1	Fe1	N1	140.89(6)
O3	Fe1	N1	77.68(5)	OW1	Fe1	N2	144.96(6)
O3	Fe1	N2	87.59(6)	N2	Fe1	N1	73.73(5)
O5	Fe1	O1	87.19(5)

2.4. Peroxidase-like Activity

The capability of acting as peroxidase mimetics was determined for the ligands and their metal complexes in phosphate buffer solution (pH 7.4), as the assay kit was supplied. At pH 3 which ITC was studied, the reference peroxidase is not stable enough for use. The thermodynamic data, however, are supposed to be informative of the relation between the relative stability and peroxidase-like activity at pH 7.4. As shown in Table 6, the Fe³⁺–edtaod^2– system shows the most significant peroxidase-like capability with a specific activity equivalent to 0.24 ± 0.04 mU/mL horseradish peroxidase (HRP).

Table 6. Specific Peroxidase-like Activities (mU/mL HRP) of Fe³⁺ and Cu²⁺ Complexes at pH 7.4 (0.25 M Sodium Phosphate), Compared with Free Metal Ions and Uncoordinated Ligands.

	Fe3+–L	Cu2+–L	L
edtaod2–	0.24 ± 0.04	undetectable	undetectable
edtabz2–	undetectable	0.02 ± 0.009	undetectable
bis(edtapo)4–a	0.13 ± 0.03	n/a	undetectable
free metal	not active	not active

^aMacrocycle derived from two edta and two bis(4-aminophenyl) ether [ref (8)].

The activity of the Fe³⁺–edtaod^2– complex is higher than the values 0.13 mU/mL HRP (at a final concentration of 1 × 10^–4 M) reported for some binuclear Fe³⁺ complexes of macrocycles derived from EDTA and aromatic diamines,⁸ in contrast to the inactiveness of the Fe³⁺–edtabz^2– complex. The catalytic cycle of HRP in cells is initiated by hydrogen peroxide through an electron oxidation of the ferric ground state of the enzyme, and the resulting transient oxo-Fe⁴⁺ complex is then reduced by an electron transfer from the substrate. This process can be repeated by a second hydrogen peroxide molecule reacts to return the iron to its state ferric ground state.²⁰ The Fe³⁺–edtaod^2– complex is supposed to form a dimeric structure with a μ-oxo bridge in solution as found by the X-ray study because the mass spectrum exhibits the dimeric-molecular species with well-defined isotope peaks—without showing any peak attributable to monomeric species (the spectrum is given in the Supporting Information). The metal center of the [Fe(edtaod)]⁺ unit is displaced from the plane constructed by ligand atoms to be exposed to environment, owing to the steric effect of the rigid macrocyclic framework (Figure 6). As a result, the Fe³⁺ center can capture an O^2– ion and can stabilize the transient Fe⁴⁺ state during peroxidase action. The inactiveness of the Fe³⁺–edtabz^2– complex is attributable to the coordination geometry of [Fe(edtabz)(H₂O)]⁺ in which the metal ion is so well suited in the coordination sphere formed by the donor atoms of the ligand as to be protected against an O^2– ion (Figure 7). Thus, the structural features of the two Fe³⁺ chelates cause the difference in their ability to mimic peroxidase, as pointed out for pyridyl based macrocyclic Fe³⁺ complexes.²¹

The inactiveness of the ligands is predictable because of the lack of transition metal centers. The negligible activity of the Cu²⁺ complexes may be related to the relative stability of the transient complexes [Cu³⁺(edtaod)]⁺ and [Cu³⁺(edtabz)]⁺ formed in the course of peroxidase action because the lower stability of the transient complexes the lower ability of metal center to mimic the catalytic cycle described by a peroxidase.^20,22

3. Conclusions

We have synthesized and characterized a new open-chain ligand edtabzH₂ derived from EDTA and carried out the comparative study of the Cu²⁺ and Fe³⁺ complexes and the corresponding complexes of an EDTA-derived macrocycle edtaodH₂. The stability constants determined by the ITC technique for the Cu²⁺ and Fe³⁺ complexes of edtabzH₂ ligand are higher than those obtained for the corresponding metal complexes of edtaodH₂. The stability constants determined by UV–vis also have found that the Cu²⁺ complexes are more stable than the Fe³⁺ complexes, and the acyclic ligand forms more stable metal complexes than does the macrocyclic ligand. No evidence was found for the macrocyclic effect on the stability of the complexes. The X-ray diffraction study of the acyclic ligand revealed that two aromatic groups are oriented to the same direction. In solution, a closer interannular contact may be induced by the hydrophobic effect to preorganize a pseudo-macrocyclic cavity. The macrocyclic ligand edtaodH₂ is naturally preorganized for complexation, but the small rigid frame generates a high steric hindrance, decreasing the stability even for copper, which normally has the highest affinity with O, N-ligands among transition metals. The X-ray studies indicate that edtabz^2– forms a larger number of coordinate bonds as a result of its flexible frame that can adapt to the ionic radii of iron and copper. The Fe–edtaodH₂ complex presents a significant POx biomimetic activity, whereas the Fe–edtabzH₂ complex is almost inactive despite a larger number of coordinate bonds. This fact indicates that the activity is controlled not only by stability but also by other factors such as the geometry of the complex. The X-ray structure of {[Fe(edtaod)]₂O} involves oxygen of the external origin within the first coordination sphere. Probably, the POx biomimetic associates with this oxygen, which exchanges with the oxygen of superoxide radical and makes Fe³⁺ ion responsible for the process of dismutation to create hydrogen peroxide from molecular oxygen. In the Fe–edtabzH₂ complex, the coordination sphere of the metal ion is totally occupied by the donor atoms of the ligand. Thus, the acyclic ligand has higher affinity to Fe³⁺ and Cu²⁺ ions so that they may be useful for chelating agents in metal intoxication rather than artificial POx mimics.

4. Experimental Section

All reagents used were of analytical grade. N,N-Dimethylformamide (DMF, Fisher Scientific) was dried with molecular sieves for 24 h, and other solvents were used without further purification. The solutions of Cu(NO₃)₂·2.5H₂O and Fe(NO₃)₃·9H₂O were standardized by EDTA complexometry.

4.1. Synthesis of the Ligands

The macrocycle edtaodH₂ (Scheme 1) was synthesized by the method reported previously, and the purity was checked by measurement of the decomposition point as well as IR and ¹H NMR spectra.¹²

The acyclic ligand edtabzH₂ (Scheme 1) was obtained by a reaction of ethylenediaminetetraacetic dianhydride with aniline. To 2.4 g (9.3 mmol) of EDTA dianhydride dissolved in 8 mL of dry DMF was added 2 mL (20 mmol) of aniline distilled in advance. The resulting reaction mixture was left to stand overnight. Any solids were filtered out. The filtrate was concentrated by a rotary evaporator to 5 mL, into which acetone was added. Precipitates formed were filtered off, washed with acetone several times until a colorless solid was obtained. It was dried in vacuum for 8 h at room temperature. Yield: 3.21 g, 90%. mp 182–183 °C (dec). Found: C, 55.70; H, 5.84; N, 11.58%. Calcd for C₂₂H₂₆O₆N₄·2H₂O: C, 55.46; H, 5.92; N, 11.76%. ¹H NMR (400 MHz, D₂O–Na₂CO₃, pD = 9.0, DSS): δ 2.86 (s, 4H, H_a), 3.29 (s, 4H, H_b), 3.42 (s, 4H, H_c), 7.20 (t, 2H, J = 8 Hz, H_p), 7.36 (t, d, 4H, H_m, H_o). ¹³C NMR (100 MHz, D₂O–Li₂CO₃, pD = 9.0, DSS): 175.24 (COOH), 168.75 (CONH), 136.02 (Ar, hipso), 129.02 (Ar, m), 125.58 (Ar, p), 121.48 (Ar, o), 57.81(C_c), 57.56 (C_b), 52.29 (C_a). Mass spectrum (ESI^–) m/z: 441.1 (100%), [(M – H)⁻].

4.2. Syntheses of Copper(II) and Iron(III) Complexes

The metal complexes were synthesized by reactions between appropriate metal nitrates and ligands. The ligands were in 10% excess to ensure complete coordination of metals in the 1:1 complexation. In 15 mL of water, 0.46 mmol of an appropriate ligand was suspended and solubilized by adding solid Li₂CO₃. To the solution was added a 10 mL aqueous solution of 0.42 mmol of an appropriated metal nitrate with stirring. Pale blue color was developed in the reaction of copper, and brilliant orange color was in the iron complexation. After stirring was continued for 30 min, the solution was heated at a temperature between 40 and 50 °C. When it was concentrated to half the initial volume, the product started to precipitate. Crystalline powder was filtered off and dried in vacuum for about 8 h at room temperature. The crystals of the metal complexes are stable in air without efflorescing. They were characterized by IR, ESI-MS spectra, X-ray diffraction analyses, and thermogravimetric analysis (TGA). [Fe₂(edtaod)₂O]·4.4H₂O (based on X-ray structure). Anal. Fe, 12.1%; calcd 12.0%. MS (ESI^–) m/z: 850.6 (100%), [(⁵⁶FeL)₂O – H]⁻ (calcd 851.1). mp, 244 °C (decomposed). Fe(edtabz)NO₃·5.5H₂O (based on X-ray structure). Anal. Fe, 8.6%; calcd 8.5%. MS (ESI⁺) m/z: 496.1 (100%), [⁵⁶FeL]⁺ (calcd 496.1). mp, 226 °C (decomposed). Cu(edtaod)·H₂O (based on X-ray structure). Anal. Cu, 14.0%; calcd 14.3%. MS (ESI^–) m/z: 424.0 (100%), [⁶³CuL – H]⁻ (calcd 424.1). mp, 240 °C (decomposed). Cu(edtabz). Anal. Cu, 12.1%; calcd for Cu(edtabz)·H₂O, 12.2%. MS (ESI^–) m/z: 250.5 (100%) [⁶³CuL – 2H]^2– (calcd 250.5). mp, 286 °C (decomposed).

4.3. Crystal Growth

Crystals of edtabzH₂ were grown by diffusing acetone into an aqueous solution placed at the bottom of a small-bore tube. After 2 days, crystals suitable for SCXRD analysis were obtained. Crystals of the iron(III) complexes were obtained by the saturation method. An aqueous solution of the metal nitrate was added into a concentrated solution of edtabzH₂ or edtaodH₂ at ∼60 °C under stirring. Heating and stirring were ceased after 30 min, and the solutions were left to evaporate at ambient temperature. Red crystals suitable for SCXRD analysis were formed in 3 or 4 days.

4.4. Potentiometric Titration

The potentiometric titration of edtabzH₂ was carried out using Dosimat plus Metrohm semiautomatic titrator consisting of a 20 mL burette and a jacketed titration cell at a controlled temperature of 25.0 ± 0.1 °C under a nitrogen atmosphere. The pH was determined by a Metrohm model 827 pH meter equipped with a combination electrode model 6.00228.010. The electrode was calibrated with 0.01 M HCl and 0.01 M CO₂-free KOH in 0.1 M KCl by assuming complete dissociation of the strong acid and base, so as to determine molar hydrogen-ion concentration. Based on Gran’s plots, pK_w was obtained to be 13.78. Precisely weighed 44.2 mg of edtabzH₂ was dissolved in a mixture of deaerated 0.1 M KCl solution (100 mL) and 0.1 M HCl (1.5 mL); the initial concentration of edtabzH₂ was ca. 1 × 10^–3 M. Nitrogen gas humidified by 0.1 KCl was bubbled for 1 h in the sample solution before titration, to minimize the effects of O₂ and CO₂. Titration was performed by adding a standardized 0.10 M KOH in 0.05 mL increments under a nitrogen atmosphere. The computer program HYPERQUAD was used to calculate the protonation constants. The titration curves of the metal–ligand systems were unable to be interpreted probably because of the complexity especially in the iron complexation.

4.5. Spectroscopic Measurements

The ¹H and ¹³C NMR spectra were obtained with a Bruker AVANCE 400 spectrometer for D₂O solutions at a probe temperature of approximately 23 °C. The internal reference was sodium 2,2-dimethyl-2-silanpentane-5-sulfonate (DSS). For the study of pD dependence, a minimum quantity of diluted KOD or DCl solution was used to adjust the pD of the sample solutions. The pH value of each sample solution was measured with a long-stem combination electrode inserted into the NMR tube after NMR experiments. The electrode was calibrated with standard aqueous buffers, and the measured pH values were converted to the pD values by the relation pD = pH_meas + 0.44.²³

IR spectra were recorded on a PerkinElmer FT-IR Spectrometer Model Frontier equipped with ATR accessory. UV–visible spectroscopy was carried out using PerkinElmer LAMBDA 45. For the studies of complexation at a constant pH, the stock solution of the ligand was made at a concentration of 5 × 10^–4 M, and the stock solutions of the metal ions were prepared form the appropriate nitrates so that the concentrations were about 30 times those of the ligand stock solutions. Titrations were carried out at pH 3 adjusted with 0.1 M HCl/NaCl. A quartz cuvette of the spectrometer was loaded with 3 mL of a stock solution of a ligand studied, to which 10 μL aliquots of an appropriate metal stock solution were added successively with a calibrated micropipette so that the ratios of the total concentrations [M]_t/[L]_t were in a desired range. Each spectrum was recorded after the sample solution was stirred thoroughly for 3 min. The titration of the Cu–edtabzH₂ system was carried out in 0.1 M CAPSO buffer at pH 8.5 as well.

The high-resolution XPS of Fe and Cu 2p electrons were taken using PerkinElmer PHI 5100 with a Mg Kα source under a vacuum of 4 × 10^–9 Torr. Each peak was referenced to C 1s (285.0 eV). Curve fitting of the peaks was performed assuming a Gaussian–Lorentzian function with baseline correction.

Mass spectra of electrospray ionization (ESI/MS) were obtained on 6130 Quadrupole LC/MS of Agilent Technologies, in the negative and positive ionization modes.

TGA was carried out on a thermogravimetric analyzer PerkinElmer Pyris 1 TGA for studying the thermal stability and compositions of the complexes. A sample with an initial weight of 5 mg was set in a ceramic pan and analyzed in a temperature ranging of 25–800 °C under N₂ or O₂ atmosphere.

4.6. Calorimetric Measurements

Microcalorimetry was performed using a MicroCal VP-ITC microcalorimeter at a temperature of 298.15 K, stirrer speed of 307 rpm, filter period of 2 s, and reference power of 10 μJ s^–1. A ligand solution was titrated with an appropriate metal nitrate solution for four reaction systems, Fe–edtabzH₂, Cu–edtabzH₂, Fe–edtaodH₂, and Cu–edtaodH₂. The stock solutions 0.02 M of the metal nitrates were standardized by EDTA titration prior to each experiment. To prevent the hydrolysis of Fe(NO₃)₃, the stock solution was acidified with a 2% HNO₃. All measurements were carried out at pH 3.0 in 0.05 M potassium hydrogen phthalate buffer. The influence of the buffer on the complexation was negligible, as confirmed by blank tests: the injection of the metal solution into the buffer alone generated small peaks that were due to dilution unaccompanied by complexation. The raw data were corrected by subtracting the heat of dilution. The titration curves were analyzed with AFFINIMETER application software, which provides thermodynamic and kinetic parameters.²⁴ The conditions of each titration system are reported in the Supporting Information.

4.7. Single-Crystal X-ray Diffraction Analyses

Diffraction data were obtained on Bruker APEX (for edtabzH₂·2H₂O), APEX II DUO (for {[Fe(edtaod)]₂O}·4.4H₂O), and D8 QUEST (for [Fe(edtabz)]NO₃·5.5H₂O) diffractometer systems equipped with a monochromator and a CDD area detector, using either Cu Kα (λ = 1.54178 Å) or Mo Kα (λ = 0.71073 Å) radiation. The measured intensities were reduced to F² and corrected for absorption using the multiscan method SADABS.²⁵ Corrections were made for Lorentz and polarization effects. Structure solution, refinement, and data output were performed with the OLEX2²⁶ program package using SHEXLT²⁷ for the structure solution and SHELXL²⁸ for the refinement. Nonhydrogen atoms were refined anisotropically. C–H hydrogen atoms were placed in geometrically calculated positions based on the riding model. N–H and O–H hydrogen atoms were located from difference Fourier maps and refined with distance restraints.

The ligand edtabzH₂·2H₂O crystallizes in the form of a zwitterion hydrate with two crystallographically independent water molecules in the asymmetric unit. In the crystal structure of [Fe(edtabz)]NO₃·5.5H₂O, nitrate counterion is disordered over two positions. For the refinement of the disorder, U_ij restraints (DELU) and geometric restrictions (FLAT) were employed. Aside from the water molecule (OW1) coordinated to the iron atom, this crystal structure comprises six crystallographically independent sites with uncoordinated water molecules (OW2–OW7) having occupancies of 1.00 (OW2, OW4–OW7) and 0.50 (OW3). Among these, four molecules labeled OW2, OW4, OW5, and OW7 that are disordered over two positions were refined without hydrogen atoms. The asymmetric unit of {[Fe(edtaod)]₂O}·4.4H₂O is constituted by the binuclear iron(III) complex and five crystallographically independent uncoordinated water molecules (OW1–OW5), two of which (OW4 and OW5) were refined with an occupancy of 0.70. Crystal structure analyses were performed with Mercury,²⁹ and Diamond³⁰ was used for the creation of figures.

Crystallographic data for the structures reported in this paper have been deposited with the Cambridge Crystallographic Data Centre as supplementary publications no. CCDC—1920517, 1920512, and 1920516. Copies of the data can be obtained free of charge on application to CCDC, 12 Union Road, Cambridge CB2 1EZ, UK (fax: (+44)1223-336-033; e-mail: deposit@ccdc.cam.ac.uk, www: http://www.ccdc.cam.ac.uk).

4.8. Peroxidase-like Activity

Peroxidase-like activity was determined using the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen cat. no. A22188) which is provided in the form of phosphate buffer solution (pH 7.4). The assay was performed pH in a total volume of 100 μL per microplate well. A standard curve was prepared to determine the activity of HRP in a concentration range of 0–2 mU/mL. The wells were loaded individually with 50 μL of the standard material and test materials (edtaodH₂, edtabzH₂, and their copper(II) and iron(III) complexes), and every sample solution was adjusted to a final concentration of 1 × 10^–3 M. To each well, 50 μL of a working solution was added to start the reaction; the working solution was composed of a 1 × 10^–4 M Amplex Red reagent (10-acetyl-3,7-dihydroxyphenoxazine) and 2 × 10^–3 M H₂O₂ in 0.25 M sodium phosphate, pH 7.4. After incubation at room temperature for 30 min under protection from light, the absorbance of each solution in the wells was measured with a microplate reader at 550 nm. The specific activities were determined by correlating the absorbance readings to a peroxidase standard curve prepared with HRP and were reported as mU/mL of HRP activity.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.9b03164.

• Mass spectra of ligands and complexes, IR spectra of ligands and complexes, UV–vis spectrum of edtabzCu at different pH values, ITC edtaod with Cu and Fe, XPS, and X-ray tables of edtabz, edtabzFe, and edtaodFe (PDF)

• Crystallographic data (CIF)

• Crystallographic data (CIF)

• Crystallographic data (CIF)

The authors declare no competing financial interest.