Understanding the Coordination Modes of [Cu(acac) 2 (imidazole) n =1,2 ] Adducts by EPR, ENDOR, HYSCORE, and DFT Analysis

.

Thi s v e r sio n is b ei n g m a d e a v ail a bl e in a c c o r d a n c e wit h p u blis h e r p olici e s. S e e h t t p://o r c a . cf. a c. u k/ p olici e s. h t ml fo r u s a g e p olici e s. Co py ri g h t a n d m o r al ri g h t s fo r p u blic a tio n s m a d e a v ail a bl e in ORCA a r e r e t ai n e d by t h e c o py ri g h t h ol d e r s .

■ INTRODUCTION
The interaction of metal ions with bioligands, including proteins, nucleic acids, and their components, forms a central part of medicinal inorganic chemistry. 1,2 These interactions are important from a biological perspective because metals ions play an essential role in many biological processes. 3 Indeed, transition-metal ions and their complexes have become an important area of investigation as potentially new classes of anticancer agents. 2,4 However, problems often arise from the severe toxicity and inherent or acquired resistance to treatment using metal-based drugs. In many cases, a wide variety of metal complexes have been proposed as potential DNA intercalators 5 and, while new metal complexes based on Ru, As, Au, V, and Ti have all shown promise, 6 none are currently in clinical use. By comparison, Cu has several characteristics that offer potential advantages in this role, including redox activity and endogenous presence within the human body. 4 The range of oxidation states, coordination geometries, and ligand structures available to Cu complexes allow for greater flexibility in design. 4 Unlike cisplatin and its derivatives, Cu complexes are believed to noncovalently bind to DNA either through major/minor grooves or via intercalation between base pairs, although other targets such as topoisomerase enzymes have also been proposed.
In these medicinal inorganic chemistry studies, in vivo activity is the ultimate test of the utility of such complexes. However, detailed structural information on how these metal complexes bind to receptors, base pairs, and indeed their mode of action is vital to obtain a complete understanding of the activity and hence enable a program of rational drug design. Many analytical techniques are therefore commonly used to analyze metal interactions and binding, including NMR, UV− vis, circular dichroism, isothermal calorimetry, and X-ray crystallography. For paramagnetic compounds, the resolution offered by NMR is often compromised and certainly considerably diminished. In such cases, the sophistication of information extracted by NMR can be matched by using electron paramagnetic resonance (EPR) spectroscopy. EPR and its related hyperfine techniques, including electron nuclear double resonance (ENDOR) and hyperfine sublevel correlation spectroscopy (HYSCORE), are extremely suitable spectroscopic methods to investigate both the electronic and geometric structure of metal complexes 7 and more generally copper active sites in biology. 8,9 Developing a greater appreciation of how the electronic and structural properties of the transition-metal complex governs the chemical nature of its interaction with biologically relevant substrates such as DNA is fundamentally important.
In this study, we exemplify the detailed information offered by the EPR techniques when investigating the coordination modes of Cu complexes, using imidazole as the choice substrate. Imidazole is common to both proteins and nucleic acids, as it is present in histidine residues and purine bases, respectively, rendering it biologically relevant. Furthermore, imidazole coordination to Cu(II) ions is prevalent in biological systems; coordination of histidine through the imino nitrogen (labeled N 3 ) as opposed to the amino nitrogen (N 1 ) of the imidazole ring is common to all copper proteins, while direct coordination of Cu(II) ions to DNA has been established to occur predominantly via the N 7 nitrogen of guanine (Scheme 1). The cytotoxicity of Cu complexes bearing substituted imidazole such as methyl-and phenyl-benzimidazoles has been demonstrated, 8 while a detailed EPR and ENDOR study of imidazole coordination to a Ru(III) anticancer agent was also reported recently. 9 Therefore, interactions between Cu(II) complexes and imidazole substrates are potentially important as model systems to better understand the nature of the Cu coordination mode with proteins and nucleic acids. Herein, we demonstrate how the coordination mode and structure of the resulting adducts can be investigated using a combination of advanced EPR techniques and density functional theory (DFT).
Sample Preparation. A series of solutions containing [Cu(acac) 2 ] and imidazole in different ratios were prepared such that the ratio of copper to Im was systematically varied (0.5, 1, 2, 5, 10, 30, 40, and 50 equiv) while the concentration of [Cu(acac) 2 ] (0.02 M) and the composition of the CHCl 3 :DMF (1:1) solvent was kept constant. The variable ratio study was monitored by CW X-band EPR in both frozen solution (140 K) and fluid solution (298 K). X-/Q-band EPR, ENDOR, and X-band HYSCORE studies were conducted on samples containing [Cu(acac) 2 ]:Im molar ratios of 1:0 and 1:50 at 10 K using 0.03 M solutions prepared in CDCl 3 :DMF-d 7 (1:1). All samples for EPR, ENDOR, and HYSCORE measurements were prepared on the bench. Dry CHCl 3 and DMF solvents were used to prepare the solutions.

Inorganic Chemistry
Article utilizing an ER4119HS resonator, 100 kHz field modulation at 140 or 298 K, and typically using 10.17 mW MW power. The CW Q-band EPR and ENDOR measurements were recorded on a Bruker Elexsys E500 spectrometer using a Bruker ER5106 QT-E Q-band resonator operating at 10 kHz field modulation and 10 K for ENDOR (and at 100 kHz and 50 K for the EPR). The CW ENDOR spectra were obtained using 5 dB RF attenuation (80 W) from an ENI 3200L RF amplifier at 100 kHz RF modulation depth and 0.5 mW microwave power. Additional X-band Davies ENDOR measurements were also obtained. These Davies−ENDOR experiments 10 were recorded on a Bruker Elexsys E580 spectrometer and carried out using the following pulse sequence: π-T-π/2-τ-π-τ-echo. The experiments were done with mw pulse lengths of t π = 256 ns, t π/2 = 128 ns, and an interpulse time τ of 800 ns. An rf τ pulse of variable frequency and a length of 18 μs were applied during time T of 20 μs.
Hyperfine Sublevel Correlation (HYSCORE) Experiments. The HYSCORE experiments 11 were performed on a Bruker Elexsys E580 spectrometer utilizing a Bruker EN 4118X-MD4 pulsed EPR/ ENDOR resonator and 10 K. The experiments were carried out with the pulse sequence π/2-τ−π/2-t 1 -π-t 2 -π/2-τ-echo. The microwave pulse lengths t π/2 = 16 ns and t π = 16 ns were adopted. The time intervals t 1 and t 1 were varied in steps of 16 ns starting from 100 to 3300 ns. The fixed interpulse delay (τ) values are specified in the figure captions. The adopted shot repetition rate was 1 kHz. A four-step phase cycle was used for eliminating unwanted echoes. Spectra were recorded at two magnetic field positions corresponding to B 0 = 338.6 mT (g ⊥ ) and B 0 = 283.2 mT (corresponding to the single crystal-like position, m I = −3/2, where only molecules with their g z axis aligned along the external magnetic field are selected). The magnetic field was measured by means of a Bruker ER035 M NMR gauss meter.
All of the EPR, ENDOR, and HYSCORE simulations were performed using the Easyspin 12 software package running within the MathWorks MatLab environment. The rotational correlation times for the different complexes were computed in Easyspin assuming an isotropic rotational diffusion in the fast motion regime. Once a value of the correlation time is imposed, line widths are computed using the Kivelson formulas 13 within the Redfield limit (motional narrowing). 14 DFT Calculations. Geometries of all species were fully optimized without symmetry constraint using the M06-2X 15 meta-hybrid functional and a basis set of 6-31+G(d,p) 16−18 on light atoms and Stuttgart−Dresden effective core potential and basis set on Cu 19 using the Gaussian09 suite. 20 The resulting geometries were used to estimate EPR parameters in ORCA 21 using the hybrid PBE0 22 functional and basis set of EPRII 23 for light atoms and the Core Properties allelectron basis set for Cu 24 with spin−orbit effects accounted for in a mean field approach. 25 ■ RESULTS AND DISCUSSION CW EPR. T h ee x p e r i m e n t a la n ds i m u l a t e dC WE P R spectrum of the unbound [Cu(acac) 2 ] complex in CHCl 3 :DMF is shown in Figure 1a (and also in the Supporting Information, Figure S1a). Although [Cu(acac) 2 ] is readily soluble in most organic solvents, the Im substrate is not. Hence, for all EPR measurements, the [Cu(acac) 2 ] complex bearing increasing concentrations of Im was dissolved in a CHCl 3 :DMF (1:1) solvent system to ensure complete solubility of Im. The resulting spin Hamiltonian parameters for the unbound [Cu(acac) 2 ] complex in this CHCl 3 :DMF solvent system are listed in Table 1. As discussed previously by us, 26 the g and Cu A parameters observed for unbound [Cu(acac) 2 ]i nf r o z e n solutions depend subtly on the choice of solvent used, and in most cases will produce a well-defined signal with the 63,65 Cu isotope splitting clearly evident on the low field hyperfine component. While very dry noncoordinating solvents such as CHCl 3 :Tol give values of g iso = 2.117 and a iso = 237 MHz, weakly coordinating solvents such as CHCl 3 :DMF (and even slightly wet CHCl 3 :Tol solvents) 26 produce notably different values, as observed in the current case with g iso = 2.135 and a iso = 196 MHz.
To investigate the nature and coordination mode of the adducts formed between [Cu(acac) 2 ] and Im, a speciation study was first performed by increasing the Cu:Im ratios from 1:0 to 1:50 in the CHCl 3 :DMF (1:1) solvent. The complete set of resulting CW EPR spectra for all ratios investigated are shown in the Supporting Information (Figures S1a−i). At 1 equiv of Cu:Im, a mixed EPR spectrum is observed, composed of unbound [Cu(acac) 2 ] bearing no Im coordination along with a second signal readily assigned to a bound [Cu(acac) 2 Im] monoadduct (see Supporting Information, Figure S1b). At 1:5 equiv of Cu:Im, only this [Cu(acac) 2 Im] monoadduct is detected in the EPR spectrum (Supporting Information, Figure  S1d). As the Cu:Im ratio is increased further, a third signal appears in the spectrum which can be readily assigned to a [Cu(acac) 2 Im 2 ] bis-adduct (vide infra). At 1:50 equiv of Cu:Im, only this [Cu(acac) 2 Im 2 ] bis-adduct is observed in the spectrum (Supporting Information, Figure S1i).
The experimental and simulated EPR spectra of the [Cu(acac) 2 Im n=1,2 ] mono-or bis-adducts (obtained at Cu:Im ratios of 1:5 and 1:50) are shown in Figures 1b and c, respectively. The resulting spin Hamiltonian parameters are listed in Table 1. The [Cu(acac) 2 Im] monoadduct reveals a small increase in the g 3 value (Δg z = 0.022) and a concomitant decrease in A 3 (ΔA z = 48 MHz) relative to the unbound [Cu(acac) 2 ] complex in the weakly coordinating CHCl 3 :DMF (1:1) solvent system. These shifts in g 3 /A 3 are even larger when using dry noncoordinating CHCl 3 :Tol (1:1) solvent (cf. Δg z = 0.055 and ΔA z = 100 MHz; 26 see Table 1) and are indicative of axial substrate coordination to the predominantly square planar Cu−O 4 environment in [Cu(acac) 2 ]. 27 The latter deviations observed for Im coordination (relative to the values for [Cu(acac) 2 ] in dry noncoordinating CHCl 3 :Tol) are notably larger compared to those reported for various substituted pyridines (cf. Δg z = 0.043, ΔA z = 74 MHz), 27 indicating a stronger axial coordination of Im with the [Cu(acac) 2 ] complex. Because Im is a stronger base compared to pyridine, this stronger interaction is expected. It should also be briefly mentioned that the axially coordinated [Cu(acac) 2 Py] monoadducts (where Py refers to pyridine, methyl-pyridine, or  Table 1. amino-methyl substituted pyridines) readily form at Cu:Py ratios of 1:1 in dry noncoordinating CHCl 3 :Tol solvents. 27 Even at Cu:Py ratios of 1:50, only [Cu(acac) 2 Py] monoadducts were formed. In the current work, the weakly coordinating CHCl 3 :DMF solvent system contributes to the higher Cu:Im ratios of 1:5 required to form the axial [Cu(acac) 2 Im] monoadduct. As expected, the 14 N superhyperfine interaction in these axially coordinated square planar monoadduct complexes bearing nitrogen-bases (such as [Cu(acac) 2 Im]) is not visible in the EPR spectrum 28 owing to the predominantly d xy ground state for the copper(II) ion.

Inorganic Chemistry
At higher Cu:Im ratios (1:50), a second copper(II) signal is clearly visible in the EPR spectrum ( Figure 1c and Supporting Information, Figures S1g−i). This appears at the expense of the [Cu(acac) 2 Im] monoadduct signal. The predominantly axial EPR signal (Table 1) of this new adduct is characterized by a   14 N hyperfine values used in the EPR simulation were extracted from the ENDOR spectra (vide infra) and are given later in Table 4.
The structural nature of the [Cu(acac) 2 Im 2 ] bis-adduct cannot be reliably extracted from the frozen solution EPR spectrum alone. The observed g values (g ⊥ = 2.059 and g ∥ = 2.288) are consistent with a Cu(II) center in an axially elongated tetragonal ligand field (g ∥ > g ⊥ ) environment. 29,34 The 14 Ns u p e r h y p e r fine pattern indicates that both Im substrates must coordinate in the equatorial (xy) plane, limiting the plausible coordination geometries to trans-equatorial or cisequatorial (Scheme 2). Representative g and Cu A parameters for copper(II) complexes possessing a Cu−O 4 N 2 environment with cis-equatorial and trans-equatorial geometries (with respect to the nitrogen ligands) are listed in Table 1. As can be seen from the Table, there is a reasonable variation in the g and Cu A values for both coordination isomers, suggesting that the cisand trans-equatorial conformations of [Cu(acac) 2 Im 2 ] cannot be easily differentiated by EPR alone. While the point group symmetries of MA 4 B 2 -type complexes should in principle be different for cis-versus trans-isomers producing axial or rhombic g tensors, the distortion that occurs in the complexes coupled with the broad line-widths in frozen solution will mean that any difference in g and Cu A anisotropy will not be visible at X-or Qband frequencies. It should also be mentioned that, despite the strength of the base, Im itself does not displace the acac ligand in the complex because the spectral features observed in Figure  1b are not consistent with those arising from [Cu(Im) 4 ]. 36,37 Formation of the mono-and bis-adducts are also expected to be temperature dependent. At 140 K, the X-band EPR spectra showed a distribution of copper species, including those arising from unbound [Cu(acac) 2 ], [Cu(acac) 2 Im], and [Cu-(acac) 2 Im 2 ], depending on the ratio of Cu to Im used (Supporting Information, Figure S1). At a Cu:Im ratio of 1:10, the 140 K spectrum was almost exclusively composed of [Cu(acac) 2 Im] adducts (Figure 1b), whereas at a Cu:Im ratio of 1:50, the spectrum was dominated by [Cu(acac) 2 Im 2 ] (Figure 1c). However, a wider distribution of copper adducts was observed in the X-band EPR spectra recorded at 298 K, depending on the Cu:Im ratios used ( Figure 2). Using the anisotropic spin Hamiltonian parameters listed in Table 1, the isotropic EPR spectra were simulated in the fast motional regime, and the resulting rotational correlational times (τ R ) were obtained by simulation (using Easyspin 12−14 )( Table 2). Representative examples of the simulated isotropic EPR spectra for the Cu:Im ratios of 1:0 and 1:50 are shown in Figure 3 (the corresponding simulation for the Cu:Im ratio of 1:10 is given in the Supporting Information, Figure S2). According to the analysis of the isotropic simulations, the EPR spectrum recorded using a Cu:Im ratio of 1:10 contains a contribution from both unbound [Cu(acac) 2 ] (45.5%) and [Cu(acac) 2 Im] (54.5%). However, for a Cu:Im ratio of 1:50, the room temperature spectrum (Figure 3) contains a contribution from both [Cu(acac) 2 Im] (79.0%) and [Cu(acac) 2 Im 2 ] (21.0%), unlike the 140 K equivalent spectrum (Figure 1c), which revealed only the presence of [Cu(acac) 2 Im 2 ] species. These results highlight the expected temperature dependency of Im binding.
It should also be noted that the integrated EPR signal intensity in the room temperature spectra was found to decrease by ca. 45% as the Im concentrations in solution increased ( Figure 2). This observation was also detected in the 140 K frozen solution EPR spectra (Supporting Information, Figure S1). Anderson et al. 37 also reported a reduced Cu(II) EPR signal intensity in solutions of Cu(II) salts bearing high Im concentrations and ascribed the observation to the possible precipitation of coordination polymers such as [Cu-(Im) 2 (Im − )] n m+ , whereby anionic Im − acted as a bridging ligand. Certainly, under basic conditions, deprotonation at the N 1 atom of imidazole produces an imidazolate anion (Im − ), which is well-documented to bridge transition-metal ions through N 1 and N 3 coordination. 38 In fact imidazolate bridges are evident in some multimetal enzymes, including SOD, 39 and can mediate magnetic couplings between the metal centers. 40 Therefore, although no precipitate was evident in the current study, it seems reasonable that the loss in Cu(II) signal intensity may at least in part be attributed to the spin−spin interactions occurring in Im − bridged [Cu(acac) 2 (Im)-(Im − )] n m+ -type polymers, similar to those described by Anderson et al. 37 DFT Analysis of the Adducts. The geometry optimized structure of the monoaxial adduct was calculated and found to be square pyramidal, bearing an axially coordinated Im ligand (Figure 4a  It should be noted that the trans-equatorial structure was predicted to lie 23.0 kJ mol −1 higher in energy compared to the trans-axial structure and 17.3 kJ mol −1 higher than the cis-mixed plane at the M06-2X (PBE0) level, and we ascribe these energy differences to the choice of functional and basis set used (see Supporting Information).
1 H ENDOR. While EPR cannot reliably distinguish between the different structural isomers of the cis-mixed plane or transequatorial [Cu(acac) 2 Im 2 ] adducts, 1 H ENDOR experiments can aid the discrimination between possible structures formed in frozen solution. We previously showed how important structural information on the coordination geometry of [Cu(acac) 2 ] adducts can be revealed through a complete angular selective 1 H ENDOR analysis, including the tilt angle and orientation of coordinated pyridine substrates with respect to the Cu−O4 plane of [Cu(acac) 2 ]. 27 Therefore, the Q-band 1 H ENDOR spectra of the bis-adduct was recorded using both protic and deuterated Im (i.e., [Cu(acac) 2 Im 2 ] and [Cu(acac) 2 (Im-d 4 ) 2 ]) using fully deuterated CHCl 3 :DMF solvents (see Supporting Information, Figure  S3). This experiment enabled the Im derived proton couplings to be readily identified in the spectra because the remaining signals in the ENDOR spectra must arise from the protons of the bis(acetylacetonate) ligand itself. The 1 H ENDOR spectra of this ligand in the unbound [Cu(acac) 2 ]c o m p l e xi s deceptively complex, bearing couplings that arise from the methine protons, the fully averaged methyl group protons, and a subset of methyl group protons undergoing hindered rotation on the EPR time scale such that a very anisotropic hyperfine tensor is produced (as revealed by variable temperature X-band Mims ENDOR). 26 This hindered rotation was found to occur in 120°jumps such that a large A dipolar and a iso component (greater than the fully averaged methyl group tensor) is always observed in the spectra. 26 In the current system, these bis(acetylacetonate) derived protons are still visible in the angular selective 1 H ENDOR spectra of [Cu(acac) 2 Im 2 ] bearing protic Im ( Figure 5). However, for clarity, only the 1 H ENDOR signals originating from the coordinated Im substrates are shown in the simulation (Figure 5), and for this reason, not all of the experimental lines are reproduced by the simulation. The resulting experimentally derived principal hyperfine values for the Im protons are listed in Table 3.
A satisfactory fit to the experimental spectra was obtained using three sets of hyperfine tensors (labeled sets I, II, and III in Table 3). The deconvoluted simulation is shown in the Supporting Information ( Figure S5). The DFT calculated hyperfine tensors were used as the starting point in the simulation, and the parameters adjusted and modified only slightly to obtain the best visual fit with the experimental spectra. Owing to the different proton environments in the bound Im substrates, the observation of these different tensors is not unexpected. The hyperfine values were compared to the theoretical 1 H hyperfine tensors calculated by DFT (Table 3). In all cases, the Im derived protons from the [Cu(acac) 2 Im 2 ] bis-adducts, including the cis-mixed plane, trans-axial, or transequatorial structures, produced different theoretical couplings. However, the hyperfine tensors predicted for the structures bearing one or two equatorially bound Im substrates (i.e., cismixed plane and trans-equatorial; Scheme 2 and Figures 4b and d) most closely matched the experimental values so that sets I, II, and III can be assigned to the H 2 /H 4 protons, the H 5 protons, and the amine H 1 proton of the Im ligand, respectively.
The higher symmetry trans-equatorial structure, containing two equivalent Im substrates, would be expected to produce an ENDOR spectrum less complex compared to that of the cismixed plane structure where both Im substrates are inequivalent, leading to more hyperfine couplings. Therefore, although the angular selective 1 H ENDOR data appear to be more consistent with the trans-equatorial structure; nevertheless, we cannot confidently discriminate between the cis-and trans-structures (Figures 4b and d)  14 N ENDOR and HYSCORE. The nitrogen superhyperfine pattern observed in the CW X-band EPR spectrum (Figure 1c)   Table 2.  14 N hyperfine and structure may be of limited diagnostic value and must be treated cautiously.

Inorganic Chemistry
The angular selective CW Q-band 14 N ENDOR spectra were therefore recorded, and the resulting experimental and simulated spectra are shown in Figure 6a. The spectra were successfully simulated using a single 14 N tensor, indicative of an equivalent nitrogen environment (with no evidence of a second strongly coupled nitrogen), and the resulting parameters are listed in Table 4. The hyperfine tensor was found to deviate

Inorganic Chemistry
Article slightly from axial symmetry with the largest principal axes approximately directed to the copper ion. According to the DFT geometry optimized structures, the 14 N hyperfine tensor calculated for the imino N 3 nitrogen in the trans-equatorial adduct most closely matched the experimental values (Table 4).
To corroborate the Q-band ENDOR experiments, additional angular selective X-band Davies ENDOR measurements were also performed ( Figure 6b). As expected at this frequency, the spectra contain overlapping signals from both 1 H and 14 N nuclei in the region between 10−25 MHz. The broad line width of the 14 N signals prevented the accurate determination of N A and N Q. Hyperfine selective Davies ENDOR measurements were also performed to suppress the 1 H signals without reducing or distorting the 14 N signals. 41 While proton suppression was successful, the 14 N signal remained broad and poorly resolved, as commonly observed for strongly coupled nitrogens in several copper proteins. 42,43 Nevertheless, an excellent fit to the experimental X-band Davies ENDOR was achieved using the 1 H and 14 N hyperfine tensors extracted from the Q-band spectra. The deconvoluted simulation of the Davies ENDOR is shown in the Supporting Information ( Figure S4). The low anisotropy in the nitrogen hyperfine coupling is characteristic of Im coordination 36,44 and is typical of σ dominant bonding as expected for Cu(II)−Im coordination 45 (unlike for example cis-o rtrans-coordination of Cu−N2O2 as in Cu(II)-Salen complexes 46 ). Combined, these observations are consistent with equatorial coordination of Im to [Cu-(acac) 2 ].
The X-band HYSCORE spectrum of the [Cu(acac) 2 Im 2 ] adduct complex (Figure 7) is characterized by cross peaks in both the (+,+) and (−,+) quadrants, arising from transitions associated with the remote amine 14 N nucleus of the imidazole ring (N 1 ). At X-band, the hyperfine coupling term for this specific 14 N interaction is approximately twice the nuclear Zeeman term, leading to the so-called cancellation condition such that the two terms cancel out in one of the two M S spin manifolds. The nuclear frequencies of this particular M S manifold corresponds to the nuclear quadrupolar resonance a For comparison, the DFT calculated 1 H hyperfine tensors for the geometry optimized adducts are also listed. Euler rotation of hyperfine tensor A to g tensor is given as a set of three Euler angles based on the zyz′ convention. Euler angles are in degrees and their uncertainties are listed in footnotes a and b. b ±10°: hyperfine tensor principal values are in MHz with uncertainty. c ±0.4 MHz: for the bis-adducts, the protons from one Im unit are labeled H 1−5 , and for the second Im unit, they are labeled H 1 ′ −5 ′ (vide infra Figure 9).

Inorganic Chemistry
Article (NQR) frequencies ν − , ν + , and ν 0 , which appear in the ESEEM spectra at 0.7 MHz (ν + ) and 1.4 MHz (ν − ≈ ν 0 ). A feature appearing at about 4 MHz in the ESEEM spectrum is due to the nuclear double-quantum transition frequency, ν DQ of the other M S manifold. 47−49 The HYSCORE spectrum recorded at the maximum echo intensity (Figure 7a) is dominated by elongated cross peaks appearing at (±0.65, +4), (±4, +0.65), (±1.4, +4), and (±4, +1.4) MHz, which correspond to (ν + ,ν DQ ) and (ν − (ν 0 ), ν DQ ) frequencies of the remote Im nitrogen nucleus, consistent with the ESEEM results. As reported by Mims and Peisach, 47 these frequencies correspond to a Fermi contact interaction term, a iso , of 1.5−2.0 MHz, a nuclear quadrupole coupling e 2 qQ/h ≈ 1.4 MHz, and an asymmetry parameter, η, of 0.9−1. In addition to these signals, two cross peaks at about (2.2, 3.9) and (3.9, 2.2) MHz are present in the (+,+) quadrant, associated with the combination frequencies due to the presence of at least two remote nitrogen nuclei coupled to the same electron spin. 50 The HYSCORE spectra were thus simulated considering a three-spin system (S = 1 / 2 , I a = 1, and I b = 1) with two equivalent nitrogen nuclei with spin Hamiltonian parameters typical for remote 14 N nuclei of Im, as listed in Table 6.3. The simulations are displayed in red in Figure 7 and provide a convincing fit at both magnetic field settings. HYSCORE experiments thus indicate the presence of magnetically equivalent remote N 1 nitrogen atoms of Cu coordinated Im rings. EPR Spectra of [Cu(acac) 2 ] with Imidazole Derivatives (Im-2−4). We were also interested to explore whether the formation of the [Cu(acac) 2 Im 2 ] bis-adduct was limited to Im only. Hence, a number of other Im derivatives, including 2methyl-imidazole (Im-2), 4(5)-methyl-imidazole (Im-3), and benzimidazole (Im-4), were also investigated (see Scheme 1).
The absence of any bis-adducts for Im-2 and Im-4 must be attributed to steric effects because the basicity for all of the Imderivatives is relatively large. Im-4 is clearly too bulky to form the trans-equatorial conformation, whereas the presence of a methyl group in position 2 of Im-2 also prevents formation of the trans-equatorial coordination mode. In contrast, in the case of Im-3, tautomerization will effectively result in the methyl group occurring at position 5 (see Scheme 1), therefore pointing away from the ligand methyl groups of the acetylacetonate units and thus enabling the formation of the trans-equatorial structure.
Coordination Mode of [Cu(acac) 2 Im 2 ]. Over the years, numerous Cu(II) complexes have been studied as model systems to explore the structure, coordination, and binding in Cu(II) proteins. Invariably, studies of imidazole, substituted imidazoles, and histidine interactions with Cu(II) ions or complexes have been undertaken using EPR, ENDOR, and  Cu hyperfine component. Owing to the strong basicity of the Im substrate, the in-plane bis-(acetylacetonato) ligands must rearrange to facilitate the inplane equatorial coordination of the two Im units. The resulting tetragonally distorted octahedral complex contains a Cu− N2O2 plane with the two coordinating nitrogens coming from the N 3 Im substrate. The g x component of the g tensor is positioned along this equatorial N−Cu−N direction with the g y component lying almost along the equatorial O−Cu−O direction and with g z almost along the axial O−Cu−O direction ( Figure 9). Orientation selective 1 H ENDOR revealed the hyperfine couplings to three sets of protons on the Im ring. The two protons adjacent to the imino N 3 nitrogen (H 2 and H 4 ) gave similar hyperfine tensors according to the DFT calculations, which were indistinguishable in the experimental ENDOR spectra. The remaining two protons (labeled H 5 and H 1 in Scheme 1) produced sufficiently different hyperfine tensors so that all proton sets could be distinguished in the angular selective 1 H ENDOR simulations.
Angular selective 14 N ENDOR spectra were also recorded at X-band (Davies ENDOR) and Q-band (CW ENDOR) frequencies, yielding information on the hyperfine coupling and nuclear quadrupole coupling to the coordinating N 3 nitrogen ( Table 4). The 14 N hyperfine tensor of the imino N 3 nitrogen was found to be nearly axially symmetric with the largest principal axis, N A y , oriented almost directly along the Cu−N bond direction and the g x -component ( Figure 9) Table 4). The experimental e 2 qQ/h and η values were also typical for strongly coordinating nitrogens in Cu− N2O2-type complexes. The experimental N A values for the remote N 1 Im nitrogen, as determined by HYSCORE, were smaller compared to the theoretical values (Table 4) because DFT often overestimates these parameters for remote nitrogens. Nevertheless, the observed values were in the region expected for equatorially bound Im and were certainly larger than those predicted for the cis-mixed plane and trans-axial structures. Taken together, it is clear that the 14 N ENDOR and   Table 1.

Inorganic Chemistry
Article HYSCORE analysis is entirely consistent with the formation of the trans-equatorial [Cu(acac) 2 Im 2 ] adduct.

■ CONCLUSION
An experimental (EPR, ENDOR, and HYSCORE) and computational study of imidazole interactions with a simple [Cu(acac) 2 ] complex was undertaken. A growing number of cytoxic Cu(II)-based complexes contain the acetylacetonate ligand; therefore, a better understanding of how such complexes interact with imidazole, representing the side chain of the amino acid histidine, for example, is experimentally important. At a relatively low ratio of Cu to Im, a [Cu(acac) 2 Im n=1 ] monoadduct is formed. The Im was found to coordinate in the axial position, as confirmed by the small shift in the g 3 value (Δg z = 0.022) and the concomitant decrease in the Cu A 3 value (ΔA z = 48 MHz) relative to the unbound [Cu(acac) 2 ] complex. At higher ratios of Cu to Im, a [Cu(acac) 2 Im 2 ] bis-adduct is formed, as revealed by the superhyperfine pattern detected in the CW EPR spectra, w h i c hc a nb ei n t e r p r e t e do n l yb a s e do nt w os t r o n g l y coordinating and largely equivalent nitrogens. Different structural isomers of this bis-adduct are possible, and detailed 1 H and 14 N hyperfine analysis reveals that the trans-equatorial conformer is formed. Three individual sets of 1 H tensors were detected in the ENDOR spectra and assigned to the H 2 /H 4 ,H 5 , and H 1 protons of Im. These values were consistent with either a cis-mixed plane or trans-equatorial structure for [Cu-(acac) 2 Im 2 ]. However, angular selective 14 N ENDOR (both CW and pulsed) provided more detailed insights into the hyperfine and quadrupole values for the coordinating imino N 3 nitrogen; these parameters were in excellent agreement with the geometry optimized structure for the trans-equatorial [Cu(acac) 2 Im 2 ] structure only. Equally, the hyperfine and quadrupole values for the remote amine 14 N were determined by simulation of the X-band HYSCORE spectra, and a reasonably good agreement was achieved between theory and experiment. The ability of the coordinating ligand in the Cubased complexes to flip between cis-and trans-conformations (from unbound to Im-bound adducts) must therefore be considered when designing novel cytotoxic Cu(II)-based complexes for target interactions with proteins bearing imidazole residues.