Investigation of Ln³⁺ complexation by a DOTA derivative substituted by an imidazothiadiazole: synthesis, solution structure, luminescence and relaxation properties

Abstract

We investigated the coordination properties of original macrocyclic Ln³⁺ complexes comprising an imidazothiadiazole heterocycle. The thermodynamic stability of the Gd³⁺ complex was determined by a combination of potentiometric and photophysical measurements. The kinetic inertness was assessed in highly acidic media. The solution structure of the Ln³⁺ complex was unambiguously determined by a set of photophysical measurements and ¹H, ¹³C, ⁸⁹Y NMR data in combination with DFT calculations, which proved coordination of the heterocycle to Ln³⁺. The ability of the imidazothiadiazole moiety to sensitize Tb³⁺ luminescence was investigated. Finally, the relaxation properties were investigated by recording ¹H nuclear magnetic relaxation dispersion (NMRD) profiles and ¹⁷O measurements. The water exchange rate is similar to that of GdDOTA as the less negative charge of the ligand is compensated for by the presence of a bulky heterocycle. Relaxivity is constant over a large range of pH values, demonstrating the favorable properties of the complex for imaging purposes.

---

Introduction

Lanthanide ions (Ln³⁺) are widely used for a broad range of applications ranging from biomedical applications to telecommunications, catalysis or metallurgy, to name a few.¹ Their unique magnetic and optical properties, combined with similarities in chemical properties, make them unique in the periodic table.² In the last decades, coordination compounds formed with Ln³⁺ have attracted much attention due to their successful biomedical applications, in both diagnosis and therapy.^3–10 For example, several Gd³⁺ complexes have been clinically approved as magnetic resonance imaging (MRI) contrast agents.³ Furthermore, several paramagnetic Ln³⁺ complexes are luminescent and can be used for optical imaging or as probes in bioassays.¹¹ Ln³⁺ complexes used in biological applications need to be sufficiently water soluble and display a high thermodynamic stability and kinetic inertness. Moreover, if they are to be used for their photophysical properties, a chromophore, suitable for transferring energy to the accepting electronic levels of the Ln³⁺ ion, must be present in sufficiently close proximity to ensure an antenna effect. Macrocyclic derivatives based on DOTA (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid) are widely used for Ln³⁺ complexation due to their high thermodynamic stability and exceptional kinetic inertness.¹² DO3A derivatives, for which one acetate arm is absent, have been widely studied as they can be easily functionalized to attach a chromophore,^11,13 or a chemically^14–16 or biologically^17–19 active group.

Heterocycles are versatile moieties as they can potentially coordinate directly to the Ln³⁺ ion and be used as chromophores or as scaffolds for bioactive molecules. However, direct coordination of the Ln³⁺ ion to the heterocycle has been underexplored, and mainly restricted to the use of pyridine derivatives,^20–23 or strictly as chromophores for Ln³⁺ luminescence sensitisation.²⁴

Among the variety of heterocyclic compounds, imidazole or thiadiazole are widely used in coordination chemistry, in particular for the formation of organometallic frameworks and functional materials.^25,26 These two cycles can be combined together in a [5–5] fused ring with a bridgehead nitrogen atom, affording imidazo[2,1-b][1,3,4]thiadiazoles. This bicyclic structure exhibits various biological activities²⁷ such as antifungal, anticancer, antihyperlipidemic, antimicrobial, anti-inflammatory, antituberculosis, anticonvulsant, analgesic, and diuretic properties.^27–36 Therefore, imidazo[2,1-b][1,3,4]thiadiazole and its derivatives have become important structures used in pharmaceutical chemistry.^37–42 The combination of coordinating properties and the biological potential of imidazothiazoles are appealing to generate original Ln³⁺ complexes.

In this context, we recently synthesized L1 and L2 (Scheme 1), which contained an imidazo[2,1-b][1,3,4]thiadiazole heterocycle, linked to a DO3A monoamide macrocycle, and studied the corresponding Ln³⁺ complexes.⁴³ The heterocycles sensitize the Tb³⁺ luminescence with relatively modest quantum yield values, and the corresponding Gd³⁺ complexes show optimised relaxivities compared to the GdDOTA gold standard. In these systems the heterocycle is linked to the macrocycle through an amide function and does not participate in the coordination sphere of Ln³⁺.

Scheme 1 Structures of the ligands discussed in this work.

Here, we describe the synthesis of L3 (Scheme 1), in which the heterocycle is directly attached to the macrocycle. To the best of our knowledge, this is the first time that such a bioactive scaffold has been directly linked to a macrocycle for Ln³⁺ complexation. This should impact the luminescence properties of the Tb³⁺ complex and the relaxation properties through the modification of the Ln³⁺ coordination sphere. To explore the potential of these compounds, it is of prime importance to determine the thermodynamic stability and kinetic inertness of the Ln³⁺ complexes, as well as to probe their structures in solution. The protonation constants of the ligand and the thermodynamic stability of the Gd³⁺ complex were determined by pH-potentiometric titrations combined with UV-visible spectroscopy. The kinetic inertness of the Gd³⁺ complex was evaluated in acidic media. The solution structure of the Ln³⁺ complex was determined using ¹H, ¹³C, and ⁸⁹Y NMR, combined with DFT calculations. The luminescence properties of the Tb³⁺ complex were investigated and the number of water molecules directly coordinated to the Tb³⁺ ion was determined. Finally, the relaxation properties of the Gd³⁺ complex were investigated by relaxometry and ¹⁷O NMR to access the exchange rate of the inner-sphere water molecule.

Results and discussion

Synthesis of ligand L3

Chlorine derivative 4, which can be prepared in three steps from commercially available 2-amino-5-methyl-1,3,4-thiadiazole (1), is a key compound in the synthesis of L3. The condensation of 1 and ethyl bromopyruvate in ethanol under microwave activation led to the formation of ester 2 in moderate yield (30%). In the next step, ester 2 was reduced with DIBAL-H at room temperature in DCM to generate primary alcohol 3 in 89% yield. Chlorination of the alcohol with thionyl chloride led to derivative 4 with a 59% yield. Finally, ligand L3 was obtained in two steps. The N-alkylation of DO3AtBu was performed in the presence of chlorinated derivative 4 and K₂CO₃ as a base in acetonitrile at 60 °C for 3 h, followed by the cleavage of t-butyl esters with a 10 M solution of HCl in dioxane at 50 °C for 16 h. L3 was isolated in 25% yield over two steps (Scheme 2). The ¹H and ¹³C NMR spectra of the various compounds are presented in Fig. S1–4 (ESI†).

Scheme 2 Synthesis pathway to access L3.

Protonation constants of ligand L3 and the stability of the complexes

Potentiometric studies were performed on L3 to determine the species present in solution through the evaluation of their equilibrium constants. The ligand protonation constants were assessed, as defined in eqn (1).

L3 displays seven protonation constants (Table 1). By comparison with L2,⁴³ the first three protonation constants can be attributed to one nitrogen of the heterocycle and two amine groups of the macrocycle. The other protonation constants can be attributed to another nitrogen of the heterocycle, carboxylate functions and/or the remaining nitrogen from the macrocycle. In order to attribute the protonation constants more precisely, UV absorption, excitation and emission spectra of L3 were recorded over the pH range 2–12 (Fig. 1 and S5–6, ESI†). Upon excitation at 300 nm, an emission band with a maximum at 410 nm is observed. The intensity decreases with decreasing pH. By fixing the emission at 400 nm, the excitation spectra show a maximum at 295 nm, which also decreases with decreasing pH. The UV absorption, emission and excitation spectra as a function of pH were fitted to obtain the protonation constants of the system. The results, presented in Table 2, show the presence of two protonation constants for the heterocycle. These protonation constants are in the same range as those previously found for L1 and L2.⁴³ They also match very well with the values of log K_H1 and log K_H5 found by potentiometry. The first protonation constant of the macrocycle (9.25) is slightly lower than that of DOTA, certainly due to the electron-withdrawing effect of the heterocycle. Indeed, it matches well with the values obtained for DO3Apic⁴⁴ and DO3APyNH₂.⁴⁵

Fig. 1 Normalized emission spectra (λ_exc = 300 nm) of L3 (150 μM) in 0.15 M NaCl at 298 K; inset: emission intensities at 400 nm as a function of pH. The line represents the best fit of the data affording the values given in Table 2.

Table 1 Protonation constants measured in NaCl (0.15 M) at 298 K

Log KH	L3	L2	DO3APyNH2 ^b	DO3APic^c	DOTA^d
a From ref. 43. b From ref. 45. c From ref. 44. d From ref. 46, in 0.1 M NaCl.
Log KH1	9.4 (1)	9.09	9.12	9.21	9.37
Log KH2	9.25 (6)	8.98	9.29	8.94	9.14
Log KH3	9.24 (3)	8.51	7.71	4.82	4.63
Log KH4	4.90 (4)	4.77	4.26	3.52	3.91
Log KH5	4 × 10 (8)	3.97	2.11	1.39
Log KH6	3.62 (9)	3.76
Log KH7	1.57 (3)	2.33
Log KH8		2.1

---

Table 2 Protonation constants of L3 and GdL3 determined by UV and fluorescence titrations in 0.15 M NaCl at 298 K

		UV	Emission	Excitation
a Not determined.
L3	Log KH1	9.6 (1)	9.4 (1)	9.0 (2)
	Log KH2	4.1 (1)	4.3 (1)	4.40 (8)
GdL3	Log KH1	nd^a	9.4 (1)	9.1 (2)
	Log KH2	nd^a	3.8 (1)	4.3 (1)

---

Complex stability and protonation constants, log K_ML, log K_MLH and log K_MLOH (eqn (2)–(4)) have been determined for Gd³⁺ as well as for Zn²⁺ and Ca²⁺ (Fig. 2 and S7, ESI†).

Fig. 2 Potentiometric titration curves of solutions containing L3 (1.3 mM) with 0 or 1 equivalents of GdCl₃ in 0.15 M NaCl, 298 K.

Due to the slow formation kinetics of the complex, Gd³⁺ stability constants were determined by the out-of-cell batch method, while those for Zn²⁺ and Ca²⁺ could be obtained with an automated titration (Fig. 2 and S7, ESI†). The different species formed and their stability constants are summarised in Table 3. The stability constant found for GdL3 is lower than that of GdDOTA but in the same order of magnitude as that of eight-coordinated GdPyNH₂.⁴⁵ Surprisingly, GdL3 shows two protonation constants of 9.0 and 4.5. The species distribution in a 1/1 molar ratio (Fig. S8, ESI†) indicates that the major species present at physiological pH is GdLH. The absorption spectra of GdL3 as a function of pH do not show important changes (see Fig. S9, ESI†), suggesting that the heterocycle has different features in the complex than in the free ligand. However, emission and excitation spectra of GdL3 vary significantly as a function of pH (Fig. S10 and 11, ESI†), and the fitting of those data gives protonation constants consistent with the potentiometric data (Table 2). Two nitrogen atoms of the heterocycle can be protonated, and the first protonation constant most certainly corresponds to protonation of the nitrogen from the thiazole moiety. The protonation constant of 4.5 could be attributed to protonation of the nitrogen from the imidazole moiety. However, it cannot be excluded that it belongs to a carboxylate function as similar protonations have been evidenced by X-ray structures of a few examples of DOTA-like complexes.^47–49 The presence of the neighbouring heterocycle could stabilize this protonation through hydrogen bonding, which would explain both a higher than expected value for protonation of the carboxylate functions and modification of the UV and fluorescence properties of the complex.

Table 3 Stability constants of the different complexes measured by potentiometric titration in NaCl 0.15 M at 298 K

Log K	L3	DO3APyNH2 ^a	DO3APic	DO3A^d	DOTA
a From ref. 45. b From ref. 44. c In 0.1 M KCl; from ref. 50. d From ref. 51. e In 0.1 M NaCl; from ref. 46. f pGd = −log[Gd]free at pH 7.4 with [Gd] = 1 × 10^−6 M and [L] = 1 × 10^−5 M.
GdL	18.1(2)	18.60	23.31^b	19.06	24.7^e
GdLH	9.0(1)	3.92	2.65
GdLH2	4.5(1)
GdLOH	11.6(1)
ZnL	18.8(1)	16.71	20.25^c	21.57	20.21^d
ZnLH	9.27(3)	6.61	4.42	3.47
ZnLH2	4.4(4)	3.98	3.06	2.07
ZnLH3	3.84(3)	2.96	1.98
ZnLH4	2.95(4)
CaL	13.56(3)	11.88	14.82^c	12.57	16.11^d
CaLH	9.32(2)	5.38	4.59	4.60
CaLH2	4.14(6)		4.32
pGd^f	15.0	15.5	20.96		19.2

---

In order to assess the selectivity of L3 for Ln³⁺versus other important physiological cations, potentiometric titrations were also performed with Zn²⁺ and Ca²⁺. The stability constant value of ZnL3 is slightly higher than that of GdL3, and interestingly two orders of magnitude higher than that of ZnDO3APyNH₂, suggesting an important role of the heterocycle in Zn²⁺ coordination. The lack of selectivity for Gd³⁺vs. Zn²⁺ is compensated for by the high kinetic inertness of the system (vide infra). ZnL3 shows two protonation constants, in the same range as those of GdL3. The presence of the second protonation constant certainly explains the relatively high stability constant of the Zn²⁺ complex compared to Gd³⁺. Finally, CaL3 behaves similarly to ZnL3, with a higher stability constant than CaDO3APyNH2, and two protonation constants.

Kinetic inertness

Kinetic inertness is also an important parameter to assess the toxicity of a given complex. Due to their preorganized structures, macrocyclic ligands are known to form complexes with high kinetic inertness compared to acyclic ligands. Dissociation can occur via the metal-assisted pathway, using physiological cations such as Zn²⁺ or Cu²⁺, or proton-assisted pathways. Metal-assisted dissociation pathways are very often triggered by the formation of dinuclear complexes. In the case of DOTA derivatives, such formation is not commonly observed, leading to very slow dissociation. However, in this case, the stability constant of the Zn²⁺ complex is of the same order of magnitude as that of the Gd³⁺ complex. We therefore decided to investigate the Zn²⁺-catalysed dissociation of GdL3. This reaction was followed by monitoring the increase in the longitudinal proton relaxation rate at pH 5 and 2 in the presence of 10 eq. or 20 eq. of Zn²⁺ (Fig. S12, ESI†). No significant increase in the relaxation rate was observed after a week at pH 2, and a month at pH 5. This means that the metal-assisted pathway is certainly negligible. This also means that Zn²⁺ cannot replace Gd³⁺ within the macrocycle despite the higher stability constant of ZnL3 vs. GdL3, due to the good kinetic inertness of the system. In order to explore the proton-assisted pathway, we used Eu³⁺ as an exchanging ion to accelerate dissociation. We followed dissociation as previously in the presence of 10 eq. of Eu³⁺ at pH 2, 3 and 5.5 (Fig. S13, ESI†). After more than one week, even at pH 2 (where the complex is not thermodynamically stable), no significant dissociation was observed.

Therefore, in order to characterize the kinetic inertness of GdL3 and compare it with those of other complexes, we decided to follow dissociation under highly acidic conditions (1 M HCl). Under these conditions, the complex is not stable and dissociates completely. Given the very high H⁺ concentration, dissociation follows pseudo-first order kinetics and the dissociation rate is directly proportional to the total concentration of the complex [LnL]_t (corresponding to the sum of the concentrations of protonated and non-protonated complexes), where k_obs is the observed pseudo-first order rate constant:

The k_obs value determined for GdL3 is presented in Table 4, together with the calculated dissociation half-life in 1 M HCl. The kinetic inertness of GdL3 is lower than that of GdDOTA, but much higher than that of seven-coordinate GdDO3A (k_obs two orders of magnitude higher). Interestingly, the k_obs value of GdL3 is one order of magnitude lower than those of GdDO3APyNH₂ and EuDO3APic. The lower kinetic inertness of GdDO3APyNH₂ was explained by the presence of a protonable amine function (log K_GdLH = 3.92), which largely accelerated proton assisted dissociation. In the case of GdL3, two protonation constants are observed (log K_GdLH = 9.0 and log K_GdLH2 = 4.5). The first one corresponds to protonation of the nitrogen atom from the thiazole core, which is far away from Gd³⁺ and explains the low impact on complex dissociation. However, the second protonation, of the nitrogen of the imidazole moiety or a neighbouring carboxylate function, is in close proximity to the Gd³⁺ centre, and can accelerate dissociation. For EuDO3APic, the carboxylate function of the picolinate group coordinates at a sterically demanding capping position, and thus protonation of the picolinate moiety facilitates complex dissociation.^52,53

Table 4 Rate constants characterizing dissociation of the complexes determined in 1 M HCl (0.15 M NaCl, unless otherwise stated) at 298 K, and half-life under the same conditions

	k obs (s^−1)	t 1/2 (min)
a From ref. 54. b In 0.1 M KCl from ref. 55. c In 0.1 M KCl from ref. 44. d From ref. 45.
GdL3	1.8 × 10^−4	64
GdDOTA^a	1.8 × 10^−6	6418
GdDO3A^b	2.3 × 10^−2	0.5
EuDO3Apic^c	2.0 × 10^−3	5.7
GdDO3APyNH2 ^d	9.1 × 10^−2	0.13

---

Structural study of the complex

In order to determine the structure of the LnLH complex at physiological pH, we performed NMR studies on various Ln³⁺ complexes. First, we recorded the ¹H NMR spectrum of EuL3H at 700 MHz, 298 K (Fig. S14, ESI†). The spectrum shows two sets of 26 signals, which are indicative of C₁ symmetry and the presence of two isomers. It is well-known that in nine-coordinate Ln³⁺ DOTA-like complexes, there are two possible conformations of the macrocycle, (δδδδ) and (λλλλ), dictated by the gauche conformations of the individual ethylenediamine groups. Moreover, the four pendant arms can have two possible orientations (absolute configurations Δ and Λ). All this results in four possible isomers, existing as two enantiomeric pairs. These isomers adopt either monocapped square antiprismatic (SAP) or monocapped twisted square antiprismatic (TSAP) geometry.^56,57 The chemical shifts of the axial ring protons of the major isomer (29.75, 29.17, 28.59, 27.68 ppm) and minor isomer (11.16, 10.97, 9.42, 9.24 ppm) are similar to those described for the SAP and TSAP isomers of EuDOTA derivatives, respectively.^55,58 Therefore, by comparison of the two sets of signals (Fig. S15, ESI†), it is estimated that 60% of EuL3H exists as the SAP isomer and 40% as the TSAP isomer under these experimental conditions. At the same temperature, the populations of the two isomers in EuDOTA are ca. 80 : 20 (SAP : TSAP),⁵⁸ indicating that the presence of the heterocyclic arm slightly stabilises the TSAP isomer. This is in line with previous studies, which indicate that increasing steric hindrance favours the TSAP isomer.⁵⁹

The fact that the EuL3H complex provides a well-resolved ¹H NMR spectrum suggests coordination of the heterocyclic arm to the metal ion, as EuDO3A derivatives generally show broad signals in their NMR spectra due to their fluxional behaviour. This is due to a low activation barrier for the arm rotation pathway responsible for the SAP↔TSAP exchange process. In contrast, complexes in which four pendant arms are involved in coordination display rather high activation energies for both rotation of the pendant arms and the (δδδδ) ↔ (λλλλ) inversion of the cyclen unit. The presence of a single water molecule coordinated to the metal centre, evidenced by both luminescence and relaxometric studies (vide infra), is also in line with coordination of the heterocycle, as GdDO3A derivatives contain two water molecules coordinated to the metal ion. However, we cannot exclude an 8-coordinated Ln³⁺ complex due to steric hindrance of the heterocycle. Thus, we performed a DFT study to gain additional insight into the structure of the complex in solution. For this purpose, we modelled the SAP isomer of the GdL3H complex incorporating up to six water molecules. The incorporation of a few explicit second-sphere water molecules is required for a better description of the bond lengths involving inner-sphere water molecules and the number of water molecules directly coordinated to the metal ion. Calculations performed on the GdL3H·6H₂O system using the wB97XD functional provide three SAP energy minima, one with the heterocycle coordinated to the metal ion and a coordinated water molecule (Fig. 3), a second geometry with the arm uncoordinated and two coordinated water molecules, and a third energy minimum with the uncoordinated heterocycle and one inner-sphere water molecule (Fig. 3). Our calculations indicate that the geometry in which the heterocycle is coordinated (N₅O₄ donor set) is much more stable than those lacking heterocycle coordination, either having one (N₄O₄ donor set, ΔG = +9.6 kcal mol⁻¹, Fig. S16, ESI†) or two coordinated water molecules (N₄O₅ donor set, ΔG = +11.78 kcal mol⁻¹, Fig. S17, ESI†). This excludes the possibility of a stable 8-coordinated Ln³⁺ within the complex.

Fig. 3 Geometry of the GdL3H·6H₂O system optimized using DFT calculations (wB97XD functional). Hydrogen atoms bonded to C atoms are omitted for simplicity. Bond lengths of the Gd(III) coordination environment (Å): Gd–N1, 2.697; Gd–N2, 2.674; Gd–N3, 2.713; Gd–N4, 2.783; Gd–N5, 2.549; Gd–O1, 2.387; Gd–O2, 2.352; Gd–O3, 2.384; Gd–O1w, 2.500.

The diamagnetic Y³⁺ complex was also synthesized and its ¹H NMR spectrum was recorded in D₂O (Fig. S18, ESI†). The assignment of the proton signals (Table S4, ESI†) was based on standard 2D experiments (COSY, HSQC, HMBC). Interestingly, the CH₂ between the macrocycle and the heterocycle shows two resonances with a strong geminal coupling, which is characteristic of the absence of free rotation of this CH₂. This is consistent with the two signals observed for this same CH₂ in the corresponding Eu³⁺ complex, confirming heterocycle coordination.

⁸⁹Y NMR shifts are also very informative about the metal environment.⁶⁰ The chemical shift of ⁸⁹Y can be obtained using ¹H,⁸⁹Y heteronuclear shift correlation through scalar coupling (HMQC). As the relaxation time of ⁸⁹Y is very long (200–500 s), this enables much faster acquisition. The spectrum of YL3H, which is presented in Fig. 4, shows clear cross-peaks correlating the ⁸⁹Y NMR signal and the equatorial protons of the pendant arms situated three bonds away from the metal ion (H5, H8, H11 and H14). The ⁸⁹Y NMR chemical shift obtained from this experiment (120.20 ppm) is very similar to those reported for DOTA-tetraamide derivatives including [Y(DOTAM)]³⁺.⁶⁰ Furthermore, the ⁸⁹Y NMR signal is significantly more shielded in [Y(DO3A)] (103 ppm)⁶⁰ than in YL3H, which we attribute to the coordination of the heterocyclic unit in the latter.

Fig. 4 ¹H,⁸⁹Y-HMQC spectrum of YL3H recorded in D₂O, 4.60 mM, pD 7.36.

The ⁸⁹Y NMR chemical shift can be approximated by an empirical expression that adds the contributions of individual donor atoms:⁶⁰

δ^calc(⁸⁹Y) = 863 − (4 S_Nam + S_Ow + 3 S_Oc + S_Nhet) (6)

Taking the tabulated values of the shielding constants of amine groups (S_Nam = 68.1 ppm), carboxylate oxygen atoms (S_Oc = 94.0 ppm) and coordinated water molecules (S_Ow = 107.6 ppm), and estimating the contribution of the heterocycle (S_Nhet) from that reported for pyridine N atoms (S_Npy = 85.7 ppm), we obtain δ^calc(⁸⁹Y) = 115.3 ppm. The good agreement between the experimental shift and that obtained with the aid of the empirical correlation confirms heterocyclic binding and the presence of a water molecule coordinated to the metal ion.

The ⁸⁹Y NMR chemical shifts were also calculated using relativistic DFT, to support the analysis performed using the predictions of eqn (6). For this purpose, we optimized the geometries of the YL3H·6H₂O system and calculated the ⁸⁹Y NMR shielding tensors using the TPSSh functional, which was found to provide good results for this problem. These calculations yielded the shielding constants shown in Table 5, as well as the corresponding diamagnetic and paramagnetic contributions.⁶¹ For chemical shift calculation purposes, we also calculated the shielding constants for the [Y(H₂O)₈]³⁺·16H₂O system. DFT gives calculated shifts in excellent agreement with the experimental results when using the model structure with the heterocyclic arm involved in Ln³⁺ coordination (N₅O₄ donor set). The model structure that contains two inner-sphere water molecules and an uncoordinated heterocyclic unit (N₄O₅ donor set) deviates by 19.5 ppm from the experimental data. Furthermore, the structure with an uncoordinated heterocycle and only one coordinated water molecule (N₄O₄ donor set) deviates even further from the experimental data by up to ∼30 ppm. Thus, our DFT calculations confirm that the complex exists in solution as a mono-hydrated species in which all four pendant arms are coordinated to the metal ion.

Table 5 Isotropic ⁸⁹Y NMR shielding constants (σ_iso), their paramagnetic (σ_para) and diamagnetic (σ_dia) contributions, and chemical shifts obtained with relativistic DFT calculations^a

	Donor set	σ iso	σ dia	σ para	δ calc
a All data in ppm.
YL3H·6H2O	N5O4	2554.1	3627.3	−1073.2	122.4
YL3H·6H2O	N4O5	2573.7	3626.2	−1052.5	102.8
YL3H·6H2O	N4O4	2524.7	3627.6	−1102.8	151.8
[Y(H2O)8]^3+	O8	2676.5	3617.0	−940.5	0.0

---

These calculations give very similar diamagnetic contributions to the isotropic shielding constant σ. This is expected, as the diamagnetic contribution arises from changes in the local magnetic field at the ⁸⁹Y nucleus associated with the charge distribution. The paramagnetic contributions involve contributions from excited states, and it is the main factor affecting the chemical shifts.

Photophysical properties

Photophysical properties were studied for a 0.5 mM solution of TbL3H in HEPES buffer (pH = 7.4). The absorption spectrum presents broad bands in the UV range up to 300 nm (Fig. 5). The excitation spectrum collected upon monitoring the Tb³⁺ emission at 543 nm is dominated by broad ligand-centred bands in the UV range (Fig. 6, top) reflecting the ability of the imidazothiadiazole-based chromophore to sensitize the visible luminescence of Tb³⁺ through the ‘antenna effect’.⁶² In addition, in the excitation spectrum, sharper features corresponding to Tb³⁺ f–f transitions are present in the UV and visible ranges indicating direct excitation of the Tb³⁺ metal ion in TbL3H. Upon excitation of the ligand-centred bands at 270 nm, TbL3H exhibits green emission with sharp features in the range of 480–700 nm due to the ⁵D₄ → ⁷F_J (J = 6–0) transitions of Tb³⁺ in addition to broader ligand-centred bands located in the range of 300–450 nm (Fig. 6, bottom). The measured Tb³⁺-centred quantum yield (Q^L_Tb) value was found to be (0.21 ± 0.02)%. This relatively small value compared to other Tb³⁺ complexes⁶³ can be partially explained by the low efficiency of this chromophore to operate as a sensitizer of Tb³⁺ emission. Nevertheless, the value of Q^L_Tb of TbL3H is 10–13 times higher than the ones reported for comparable Tb³⁺ complexes formed with imidazothiadiazole-based chromophores; this can be explained by the direct coordination of the heterocycle to Tb³⁺.⁴³ In order to estimate the number of water molecules coordinated to Tb³⁺ (q) using phenomenological equations described in the literature,⁶⁴ luminescence decays were recorded for solutions of TbL3H in HEPES buffer (pH = 7.4) and D₂O (pD = 7.4). It was found that the experimental decay curves were best fitted with bi-exponential functions (Table S9, ESI†), indicating the presence of Tb³⁺ cations in different coordination environments or differently affected by non-radiative quenching processes. Average Tb³⁺ luminescence lifetimes were calculated to be (1.5 ± 0.02) ms and (2.4 ± 0.01) ms for solutions in HEPES buffer and D₂O, respectively, giving the value of q equal to 1.1, supporting previous DFT calculations.

Fig. 5 Absorption spectrum of TbL3H (0.5 mM, HEPES pH 7.4, room temperature).

Fig. 6 Corrected and normalized (top) excitation spectrum collected upon monitoring the Tb³⁺ emission at 543 nm and (bottom) emission spectrum upon excitation at 270 nm of a 0.5 mM solution of TbL3H in HEPES buffer (pH 7.4, room temperature).

Relaxation properties

To determine the relaxivity of the Gd³⁺ complex, paramagnetic relaxation enhancements (PRE) were measured at 60 MHz and 25 °C as a function of concentration (Fig. S20, ESI†). The PREs are linear with respect to the concentration indicating the absence of an aggregation process over the concentration range studied. The relaxivity was found to be 4.11 mM⁻¹ s⁻¹ at 20 MHz, 25 °C, that is, slightly lower than those of GdL1 and GdL2 (4.58 and 4.50 mM⁻¹ s⁻¹, respectively), and higher than that of GdDOTA (r₁ = 3.70 mM⁻¹ s⁻¹).

To characterise the parameters governing proton relaxivity of the complex, nuclear magnetic relaxation dispersion (NMRD) profiles were recorded at three different temperatures (Fig. 7B). Since relaxivity is determined by several physicochemical parameters, including the water exchange rate, electron relaxation parameters and rotational correlation times, it is important to assess the maximum of these parameters independently. In the case of GdL3H, relaxivity decreases with increasing temperature, which is characteristic of small molecular complexes for which relaxivity values are limited by rotation of the system. The water exchange rate is strongly influenced by the Gd³⁺ coordination sphere (type of coordinating function, steric hindrance, coordination number).⁶⁵ The coordination sphere of GdL3H is not classical, therefore it is interesting to determine the water exchange rate of the system independently, even if it does not strongly influence the relaxivity values. To do so, ¹⁷O NMR spectroscopy has been used.

Fig. 7 (A) Temperature dependence of the reduced ¹⁷O transverse relaxation rates (top) of GdL3H (1.8 mM) at 9.4 T. (B) NMRD profiles of GdL3H (1.09 mM) at 25 °C (blue ◆), 37 °C (purple ■), and 50 °C (grey ▲). The curves represent the simultaneous fit to the experimental data points using the SBM theory.

Variable temperature ¹⁷O T₂ measurements give access to the water exchange rate, k_ex. It is also possible to obtain information about the rotational correlation time, τ_R, by recording the ¹⁷O T₁ data, which are determined by dipole–dipole and quadrupolar relaxation mechanisms. Finally, the ¹⁷O chemical shifts give an indication of the number of water molecules directly coordinated to Gd³⁺, q. Longitudinal and transverse ¹⁷O relaxation rates and chemical shifts were measured as a function of temperature for aqueous solutions of GdL3H and a diamagnetic reference. The longitudinal relaxation rates, and chemical shifts measured for GdL3H, were too close to those of the reference and were therefore excluded from the analysis. The reduced ¹⁷O transverse relaxation rates are presented in Fig. 7A. The ¹⁷O-reduced transverse relaxation rates first increase (up to ca. 310 K), then decrease with increasing temperature. It indicates that the complex is in the slow kinetic region at low temperatures and in the fast exchange region at higher temperatures. In the slow kinetic region, 1/T_2r is directly determined by the exchange rate constant k_ex, whereas in the fast exchange region, it is determined by the transverse relaxation rate of the coordinated water oxygen, 1/T_2m, which is in turn influenced by the water exchange rate, k_ex, the longitudinal electronic relaxation rate, 1/T_1e, and the scalar coupling constant, A/ħ. In our case, the slow kinetic region is well-defined and enables a reliable determination of k_ex.

The transverse ¹⁷O relaxation rates, and the NMRD profiles, were simultaneously analysed with Solomon–Bloembergen–Morgan (SBM) theory to yield the microscopic parameters characterising water exchange and rotation (see ESI† for equations). Indeed, if we are not interested in detailed information about electron spin relaxation and if we restrict the analysis of NMRD data to medium and high magnetic fields, the SBM approach gives reliable information about dynamic processes like water exchange and rotational correlation times for small complexes.^66,67 Therefore we decided to include only relaxivity values above 6 MHz in the fitting process.

In the analysis of the data, several parameters have been fixed to common values. The hydration number was fixed to 1, as the complex proved to be monohydrated both by luminescence lifetime measurements on the corresponding Tb³⁺ complex, and by DFT calculations. r_GdO was fixed to 2.5 Å as shown by DFT calculations (vide supra). This is also the distance commonly used for polyaminopolycarboxylate complexes based on available crystal structures and ENDOR results.⁶⁸ The quadrupolar coupling constant, χ(1 + η²/3)^1/2, was set to the value for pure water, 7.58 MHz.⁶⁹ The diffusion coefficients D²⁹⁸_GdH = 26 × 10⁻⁹ m² s⁻¹ were fixed, and the corresponding activation energies E_DGdH were fitted. The Gd–water proton distance was fixed to r_GdH = 3.1 Å, the closest approach between the Gd³⁺ ion and the outer sphere protons was fixed to a_GdH = 3.6 Å, and the scalar coupling constant A/ℏ was fixed to −3.6 × 10⁶ rad s⁻¹. The following parameters have been adjusted: the water exchange rate, k²⁹⁸_ex, the activation enthalpy for water exchange, ΔH^‡, the rotational correlation time, τ²⁹⁸_R, and its activation energy, E_R, and the parameters describing electron spin relaxation, the mean square of the zero field splitting, Δ², and the correlation time for the modulation of the zero field splitting, τ²⁹⁸_V; its activation energy, E_V, was fixed to 1 kJ mol⁻¹. The parameters resulting from the best fit are presented in Tables 6 and S10.†

Table 6 Parameters obtained from the simultaneous fitting of the transverse ¹⁷O NMR relaxation rates as a function of temperature at 9.4 T, and of the NMRD profiles at 298 K, 310 K, and 323 K, using the Solomon–Bloembergen–Morgan theory presented in the ESI†

	GdL3H	GdL1^a	GdDO3APyNH2 ^b	GdDOTA^c
a From ref. 43. b From ref. 45. c From ref. 70.
r 1 (mM^−1 s^−1), 60 MHz, 25 °C	4.11	4.58	4.04	3.7
k ^298ex (10^6 s^−1)	4.8 (6)	2.2	1.3	4.1
ΔH^‡ (kJ mol^−1)	28 (6)	49	52.4	49.8
τ R ^298 (ps)	75 (2)	98	76	77
E R (kJ mol^−1)	9.1 × 10^−2	17	28	16.1

---

The water exchange rate of GdL3H is 4.8 × 10⁶ s⁻¹, which is higher than that of GdL1 or GdL2.⁴³ Interestingly, it is of the same order of magnitude as that of GdDOTA, but higher than that of GdDO3APyNH₂. In the case of dissociative exchange for all DTPA and DOTA derivatives, it was generally observed that the replacement of one negatively charged carboxylate in the complex with a neutral amide decreased the exchange rate to about one third.⁷¹ The same phenomenon was observed with GdDO3APyNH₂, for which a neutral pyridinic ligand replaced a carboxylate. The higher water exchange rate observed for GdL3H can be explained by steric crowding around Gd³⁺ due to the bulky imidazothiadiazole heterocycle compared to pyridine. This is in line with previous results demonstrating an acceleration of the water exchange rate with steric crowding.⁶⁵ Moreover, the presence of several other heteroatoms near the water molecule can also accelerate the water exchange rate through hydrogen bonding networks.

The rotational correlation time is 75 ps, which is similar to that of GdDOTA, and slightly lower than those of GdL1 or GdL2, in accordance with the smaller size of the complex, and this explains the slightly lower relaxivity.

The pH dependence of relaxivity was also investigated at 60 MHz, 25 °C, over the pH range 3.8–12. The results, presented in Fig. 8, show that relaxivity is constant over the pH range studied. The first protonation step of the complex (log K = 9.0) was not expected to alter the relaxivity, as it occurs on the nitrogen atom of the thiazole ring of the heterocycle, but interestingly the second protonation step (log K = 4.5) does not influence the relaxivity either. This second protonation step was expected to happen either on the nitrogen of the imidazole moiety or on a carboxylate function in the vicinity of the heterocycle. When the nitrogen atom of the imidazole moiety is protonated, the heterocycle dissociates from Ln³⁺, and it is clear, from DFT calculations, that Ln³⁺ would be bishydrated in this environment, leading to a substantial increase in relaxivity below pH 6 (see the species distribution in Fig. 8). This is not the case, which means that the second protonation should occur on the carboxylate function, while maintaining a monohydrated complex as previously observed based on the solid state and solution properties of such LnDOTA derivatives.⁴⁷

Fig. 8 pH dependency of the relaxivity of GdL3H at 60 MHz, 25 °C, together with the species distribution calculated from data in Tables 1 and 3.

Conclusions

In conclusion, we have developed an efficient synthesis pathway to DO3A ligands substituted with an imidazothiadiazole scaffold for Ln³⁺ complexation. Potentiometric studies of the resulting systems show the formation of a stable Gd³⁺ complex with two protonation constants. These protonation steps could be attributed to the nitrogen atom located on the thiazole moiety of the heterocycle and to a carboxylate function in the vicinity of the heterocycle. Protonation of the latter is stabilized by the formation of a H-bond with the heterocycle. This results in a slightly lower stability constant value of the Gd³⁺ complex compared to other Gd³⁺ DO3A complexes with the direct coordination of a pyridinic moiety. The selectivity of the ligand for Gd³⁺vs. Zn²⁺ is ensured by the very good kinetic inertness of the system. The kinetic inertness of the system had to be determined in highly acidic media as the complex is very inert under modestly acidic conditions, and in the presence of competing metal ions, demonstrating its strong inertness. Full characterization of the solution structure of the Ln³⁺ complex was undertaken by a combination of DFT calculations and NMR experiments (¹H, ¹³C and ⁸⁹Y) on the paramagnetic Eu³⁺ and diamagnetic Y³⁺ complexes. The well-resolved Eu^{3+ 1}H NMR spectrum evidenced the presence of two isomers (SAP and TSAP) of C₁ symmetry, with a slightly favoured TSAP isomer compared to GdDOTA. Coordination of the heterocycle to Ln³⁺ was demonstrated by DFT calculations and supported by the experimental ⁸⁹Y chemical shift, which evidenced a N₅O₄ coordination sphere of Ln³⁺ with one coordinated water molecule, as also evidenced by luminescence lifetime measurements on the Tb³⁺ complex. Photophysical studies show the sensitization of Tb³⁺ luminescence by the imidazothiadiazole moiety with relatively modest quantum yield values, which are nevertheless at least ten times higher than those of TbL1 and TbL2 due to the direct coordination of the heterocycle. The presence of the coordinated water molecule enables relaxivities higher than that of GdDOTA to be achieved and this is consistent with the size of the system. Interestingly the water exchange rate, as assessed by ¹⁷O NMR is similar to that of GdDOTA due to the combination of a less negative and more crowded environment around Gd³⁺. Relaxivity is constant over the whole range of studied pH values. Compared to the previous generation of complexes, for which the heterocycle was linked to the macrocycle through an amide function, direct substitution on the macrocycle enables coordination of the heterocycle to Ln³⁺. This leads to an optimized quantum yield for the Tb³⁺ complex, and water exchange rate for the Gd³⁺ complex. Importantly the resulting Ln³⁺ complexes are thermodynamically stable and kinetically inert. The presence of a protonated complex is observed and can be explained by a H-bonding network with the heterocycle. Tuning of the position of the heteroatoms within the heterocycle should allow us to tune the presence of these protonated species and further optimize the kinetic inertness of the system. Altogether, this study of the coordination of Ln³⁺ by an imidazothiazole moiety shows promise in terms of thermodynamic stability, kinetic inertness, and the photophysical and relaxation properties of the system, which should pave the way for the development of efficient systems optimized for imaging properties and with potential therapeutic applications.

Experimental part

General information

¹H NMR and ¹³C NMR spectra were recorded on a Bruker DPX 400 MHz instrument using CDCl₃ and DMSO-d₆ or on a Bruker Avance III HD spectrometer at 298 K using a 5 mm BBFO probe. ¹H and ¹³C spectra were obtained respectively at 600 MHz and 150 MHz. The chemical shifts are reported in parts per million (δ scale), and all coupling constant (J) values are reported in hertz. The following abbreviations were used for the multiplicities: s (singlet), d (doublet), t (triplet), q (quartet), p (pentuplet), m (multiplet), sext (sextuplet), and dd (doublet of doublets). All compounds were characterized by ¹H NMR, and ¹³C NMR, which were consistent with those reported in the literature (ESI†). Melting points are uncorrected. IR absorption spectra were obtained on a PerkinElmer PARAGON 1000 PC instrument, and the values are reported in inverse centimeters. High-resolution mass spectra (HRMS) were recorded on a Bruker Q-TOF MaXis spectrometer. The reactions were monitored by thin-layer chromatography (TLC) using aluminum-backed plates (Kiesel gel 60F254, Merck, Darmstadt, Germany) and visualized using ultraviolet light (λ = 254 nm or 365 nm). Column chromatography was performed using silica gel 60 (0.063–0.200 mm, Merck). Microwave irradiation was carried out in sealed vessels placed in a Biotage Initiator or Biotage Initiator + system (400 W maximum power). The temperatures were measured externally by IR. Pressure was measured by a non-invasive sensor integrated into the cavity lid. All reagents were purchased from commercial suppliers and were used without further purification.

Synthesis and characterization

Ethyl 2-methylimidazo[2,1-b][1,3,4]thiadiazole-6-carboxylate (2). 2-Amino-5-methyl-1,3,4-thiadiazole (1, 0.5 g, 4.34 mmol, 1 eq.) and ethyl bromopyruvate (1 mL, 7.96 mmol, 1.8 eq.) are dissolved in 15 mL of ethanol and stirred under microwave activation for 1.5 h at 100 °C. Then, the solvent is evaporated under reduced pressure and 20 mL of sodium bicarbonate is added to the resulting oil. The solution is extracted three times with 20 mL of DCM and the combined organic phases are washed with 20 mL of brine, dried over MgSO₄ and concentrated under reduced pressure. The crude product is purified by flash chromatography with AcOEt/PE (70/30) as the eluent, to yield compound 2 as a yellow solid (275 mg, 30%). Spectral data correspond to those reported in the literature.

(2-Methylimidazo[2,1-b][1,3,4]thiadiazol-6-yl)methanol (3). Compound 2 (383 mg, 1.81 mmol, 1.0 eq.) is dissolved in 20 mL of anhydrous DCM under inert atmosphere at 0 °C. Then, DIBAL-H (1 M in toluene) (8.16 mL, 8.16 mmol, 4.5 eq.) is added dropwise and the mixture is stirred at room temperature for 24 h. The mixture is diluted with 10 mL of saturated Rochelle's salt solution and stirred for 24 h. The solution is extracted twice with 10 mL of DCM, then twice with 10 mL of EtOAc. The organic phases are collected, dried over MgSO₄ and evaporated under reduced pressure. Compound 3 is obtained (272 mg, 89%) as a light-yellow solid. Mp = 111–113 °C; ¹H NMR (400 MHz, DMSO-d₆) δ (ppm): 6.94 (s, 1H), 4.20 (t, J = 5.7 Hz, 1H), 3.51 (d, J = 5.6 Hz, 2H), 1.77 (s, 3H); ¹³C NMR (101 MHz, DMSO-d₆) δ (ppm): 159.9, 147.4, 143.9, 111.0, 58.2, 17.4; HRMS (EI-MS) m/z calcd for C₆H₈N₃OS: 170.0386 [M + H]⁺, found: 170.0383.

6-(Chloromethyl)-2-methylimidazo[2,1-b][1,3,4]thiadiazole (4). A solution of 3 (0.154 mg, 0.91 mmol, 1.0 eq.) is prepared in 5 mL of dry DCM. The solution is cooled to 0 °C and SOCl₂ (0.1 mL, 1.4 mmol, 1.5 eq.) is added dropwise. The mixture is stirred at 0 °C for 1 h and then stirred at room temperature overnight. 10 mL of water are added to quench the reaction. Next, 10 mL of DCM are added and the solution is neutralized with a solution of saturated NaHCO₃ to reach pH 7–8. The aqueous phase is extracted twice with 10 mL DCM, and the combined organic phases are washed with 10 mL of brine, then dried over MgSO₄, filtered and concentrated under reduced pressure. The product is purified by flash chromatography with the following eluent: PE/EtOAc (70/30). Compound 4 is obtained (101 mg, 59%) as a beige solid. Mp = 101–103 °C; ¹H NMR (400 MHz, CDCl₃) δ (ppm): 7.70 (s, 1H), 4.64 (s, 2H), 2.69 (s, 3H); ¹³C NMR (101 MHz, CDCl₃) δ (ppm): 160.1, 145.6, 142.4, 112.2, 39.9, 17.8; HRMS (EI-MS) m/z calcd for C₆H₇ClN₃S: 188.0045 [M + H]⁺, found: 188.0043.

2,2′,2′′-(10-((2-Methylimidazo[2,1-b][1,3,4]thiadiazol-6-yl)methyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid (L3). Compound 4 (150 mg, 0.80 mmol, 1.0 eq.) and potassium carbonate (792 mg, 5.7 mmol, 7.0 eq.) are added to a 15 mL solution of DO3AtBu (617 mg, 1.2 mmol, 1.5 eq.) in anhydrous acetonitrile. The mixture is heated at 50 °C and stirred for 24 h. Then, potassium carbonate is filtered off and the solution is washed twice with 10 mL of DCM. The solvent is evaporated under reduced pressure. The resulting yellow oil is dissolved in 5 mL of dioxane and a 5 mL solution of hydrochloric acid (HCl 10 M) is added. The mixture is stirred for 16 h at 50 °C. After that, the solvent is evaporated under reduced pressure, and the crude product is dissolved in 2 mL of MilliQ water to adjust its pH to between 4 and 5 with an aq. 2 M NaOH solution. Finally, the crude product is purified by flash chromatography on reversed phase with a gradient of water/MeOH (100 : 0/50 : 50/0 : 100) as the eluent. The resulting product L3 is isolated (98 mg, 25%) as a white solid. ¹H NMR (600 MHz, D₂O) δ (ppm): 8.02 (s, 1H), 3.96 (s, 2H), 3.82–2.96 (m, 25H), 2.71 (s, 3H); ¹³C NMR (151 MHz, D₂O) δ (ppm): 163.8, 145.8, 115.1, 56.1, 55.0–49.5, 48.7, 48.0, 16.8; HRMS (EI-MS) m/z calcd for C₂₀H₃₁N₇O₆S: 498.2131 [M + H]⁺, found: 498.2129.

Liquid sample preparation. The ligand concentrations were determined by adding an excess of metal to a ligand solution and titrating the excess metal with standardized Na₂H₂EDTA in urotropine buffer (pH 5.6–5.8) in the presence of Xylenol Orange, Eriochrome Black T, or Murexide as indicators. The concentrations of the metal solutions were determined similarly by complexometric titrations. The complexes were prepared by mixing 1 eq. of L, with 1 eq. of Ln³⁺, and the pH was adjusted to 7.4 either with a buffered solution or by adding KOH or HCl to the solution. The absence of free Ln³⁺ was checked by the Xylenol orange test. The concentrations of Ln³⁺-containing solutions were also checked both by ICP-OES and BMS measurements when possible.

GdL3: HRMS (EI-MS) m/z calcd for C₂₀H₃₁N₇O₆SGd: 653.1136 [M + H]⁺, found: 653.1133.

EuL3: ¹H NMR (700 MHz, D₂O): see Fig. S14.† HRMS (EI-MS) m/z calcd for C₂₀H₃₁N₇O₆SEu: 648.1107 [M + H]⁺, found: 648.1111.

YL3: ¹H NMR (600 MHz, D₂O): see Fig. S16.† HRMS (EI-MS) m/z calcd for C₂₀H₃₁N₇O₆SY: 584.0953 [M + H]⁺, found: 584.0953.

TbL3: HRMS (EI-MS) m/z calcd for C₂₀H₃₁N₇O₆STb: 654.1148 [M + H]⁺, found: 654.1144.

Potentiometric studies. Carbonate-free 0.1 M NaOH and 0.1 M HCl were prepared from Fisher Chemicals concentrates. Potentiometric titrations were performed in 0.15 mol L⁻¹ aqueous NaCl under a nitrogen atmosphere and the temperature was controlled to 25.0 ± 0.1 °C with a circulating water bath. The p[H] (p[H] = −log[H⁺], concentration in molarity) was measured in each titration with a combined pH glass electrode (Metrohm) filled with 3 M KCl and the titrant addition was automated by use of a 702 SM titrino system (Metrohm). The electrode was calibrated for the hydrogen ion concentration by titration of HCl with NaOH in 0.15 M electrolyte solution.⁷² A plot of meter reading versus p[H] enables the determination of the electrode standard potential (E°) and the slope factor (f). Continuous potentiometric titrations with 0.1 M HCl and 0.1 M NaOH were conducted on aqueous solutions containing 5 mL of L in 0.15 M NaCl, with 2 minutes waiting time between successive points. The titrations of the metal complexes (Ca²⁺, Zn²⁺) were performed with the same ligand solutions containing 1 equivalent of metal cation, with 2 minutes waiting time between 2 points. To determine the stability constant of GdL3, the out-of-cell batch method was used: 24 batch samples were prepared between pH 2 and 6 at 1 : 1 Gd : L ratio (1.3 mM in NaCl 0.15 M). The samples were kept at 25 °C for 1 week, until equilibrium was reached. Experimental data were refined using the computer program Hyperquad 2008.⁷³ All equilibrium constants are concentration quotients rather than activities and are defined as:

The ionic product of water at 25 °C and 0.15 mol L⁻¹ ionic strength is pK_w = 13.77.⁷⁴ Fixed values were used for pK_w, ligand acidity constants and total concentrations of metal, ligand and acid. All values and errors (one standard deviation) reported are the average of at least three independent experiments.

Photophysical measurements. For the determination of the protonation constants, 0.15 mM samples of L3 and GdL3 were titrated with NaOH or HCl in 0.1 M NaCl. UV-visible absorption spectra were recorded at 298 K on a PerkinElmer UV/Vis/NIR Lambda 19 spectrometer in the range λ = 200–500 nm with data steps of 1 nm, with a 1 cm path length. Luminescence measurements were performed at 298 K on an Agilent Cary Eclipse fluorescence spectrophotometer with a 295–1100 nm emission filter and 5 nm slit widths for excitation and emission wavelengths. Data were treated with HypSpec.^73,75 For photophysical measurements with Tb³⁺ complexes, 0.5 mM solutions were prepared in HEPES buffer (pH = 7.4, 100 mM) or D₂O (pD = 7.4). For collecting photophysical data, samples were placed in 2.4 mm i.d. quartz capillaries or quartz Suprasil cells (Hellma® 115F-QS, bandpass 0.2 or 1 cm). Absorption spectra were measured on a Jasco V-670 UV/visible/NIR spectrophotometer. Emission and excitation spectra were measured on a Horiba–Jobin–Yvon Fluorolog 3 spectrofluorimeter. All spectra were corrected for the instrumental functions. Luminescence lifetimes were determined under excitation at 266 nm provided by a YG 980 Quantel Nd:YAG laser while the Tb³⁺ signal was selected using an iHR320 monochromator (Horiba Scientific) and detected using a R928 photomultiplier tube (185–900 nm). The output signal from the detector was then fed to a Tektronix TDS 754C 500 MHz bandpass digital oscilloscope and then transferred to a PC for treatment with Origin 9®. Luminescence lifetimes are averages of at least three independent measurements. Tb³⁺-centered quantum yields were determined with a Fluorolog 3 spectrofluorimeter using an integration sphere (Model G8, GMP SA, Renens, Switzerland) and a 0.1 mM solution of TbL₁ complex from ref. 20 as a standard (Q^L_Tb = 37 ± 1% in HEPES buffer, pH 7.4). Estimated experimental error for quantum yield determination is 10%.

NMR experiments

The NMR spectra of EuL3H in D₂O (1.98 mM, pD = 6.97) were recorded at 298 K on a Bruker Avance III HD 700 spectrometer equipped with a CPTCI cryoprobe. When necessary, solvent suppression was achieved using an excitation sculpting sequence or a presaturation pulse. ¹H, ¹³C COSY and HSQC spectra were recorded. The NMR spectra of YL3H in D₂O (4.60 mM, pD = 7.36) were recorded at 298 K on a Bruker Avance III HD spectrometer at 298 K using a 5 mm BBFO probe and at 600 MHz for ¹H. ¹H, ¹³C COSY, HMBC and HMQC spectra were recorded. The ⁸⁹Y chemical shifts were obtained using a ¹H–⁸⁹Y HMQC experiment with a delay for evolution of long-range or couplings of 41.66 ms and a 2 s delay for relaxation time. ⁸⁹Y chemical shifts were expressed in ppm relative to Y(EDTA) as an external reference.⁶⁰

DFT calculations. Geometry optimizations were carried out with the Gaussian 16 program package (revision C.01),⁷⁶ using either the hybrid, long-range corrected wB97XD⁷⁷ density functional, which includes atom–atom dispersion corrections, or the hybrid-meta GGA functional TPSSh.⁷⁸ The inner electrons of Gd (46 + 4f⁷) were treated with a large-core quasirelativistic effective core potential, employing the [5s4p3d]-GTO basis set for the outermost 11 electrons.⁷⁹ For yttrium we selected the effective core potential ECP28MDF⁸⁰ (28 electrons in the core) and its associated VTZ basis set. All other atoms were described using the Def2-TZVPP basis set.⁸¹ The effects of bulk water were incorporated into all Gaussian calculations using the integral equation formalism of the polarized continuum model (IEFPCM).⁸² Input geometries were prepared from those of DO3A derivatives investigated previously.⁸³ The ⁸⁹Y NMR shielding tensors were calculated with the GIAO^84,85 method and the TPSSh functional, using the ORCA program package (Version 5.0.4)^86,87 and the ZORA⁸⁸ relativistic method. In these calculations we used the zora-def2-TZVPP basis set for ligand atoms, which contained the exponents of the def2-TZVPP basis set and was recontracted for ZORA calculations by D. A. Pantazis. For yttrium, we selected the SARC-ZORA-TZVPP⁸⁹ basis set. The resolution of identity and chain of spheres (RIJCOSX)^90–92 approximation was used to accelerate the SCF calculations and the calculation of nuclear magnetic resonance chemical shielding constants,⁹³ with the help of the SARC/J basis set for yttrium and the Def2/J⁹⁴ basis set for all other atoms. Bulk solvents were considered using the SMD solvation model.⁹⁵

Relaxometric measurements. Proton NMRD profiles were recorded at 1.09 mM in Hepes buffer (0.1 M, pH 7.4) on a Stelar SMARTracer fast field cycling relaxometer (0.01–10 MHz) and a Bruker WP80 NMR electromagnet adapted to variable field measurements (20–80 MHz) and controlled by a SMARTracer PC-NMR console. The temperature was monitored by a VTC91 temperature control unit and maintained by a gas flow. The temperature was determined by previous calibration with a Pt resistance temperature probe. The longitudinal relaxation rates (1/T₁) were determined in water.

Temperature dependent ¹⁷O NMR measurements. The transverse and longitudinal ¹⁷O NMR relaxation rates (1/T₂, 1/T₁) and the chemical shifts were measured in aqueous solutions of [GdL3] = 1.8 mM in the temperature range 278–348 K, on a Bruker Avance 400 (9.4 T, 54.5 MHz) spectrometer. The temperature was calculated according to previous calibration with ethylene glycol and methanol.⁹⁶ An acidified water solution (HClO₄, pH 3.3) was used as an external reference. Transverse relaxation times (T₂) were obtained by the Carr–Purcell–Meiboom–Gill spin-echo technique.⁹⁷ The technique for the ¹⁷O NMR measurements on Gd³⁺ complexes has been described elsewhere.⁹⁸ The samples were sealed in glass spheres fitted in 10 mm NMR tubes to avoid susceptibility corrections of the chemical shifts.⁹⁹ To improve the sensitivity ¹⁷O-enriched water (10% H¹⁷₂O, CortectNet) was added to reach around 1% enrichment. The least-squares fit of the ¹⁷O NMR data were performed using Visualiseur/Optimiseur¹⁰⁰ running on the MATLAB 8.3.0 (R2014a) platform.

Author contributions

E. Caillet, L. Nunes, M. Ndiaye: investigation and visualisation. S. V. Eliseeva: investigation and writing-original draft. A. Pallier, M. Isaac and J-F. Morfin: supervision, writing-review and editing. H. Meudal: investigation. S. Petoud and S. Routier: supervision, writing, review and editing. C. Platas-Iglesias: conceptualization, investigation, writing – original draft. F. Buron, C. S. Bonnet: conceptualization, project management, funding acquisition, writing-original draft.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors gratefully acknowledge the Région Centre Val de Loire (RTR Motivhealth), the Labex programs SYNORG (ANR-11-LABX-0029) and IRON (ANR-11-LABX-0018-01), the Ligue contre le Cancer du Grand Ouest (comités des Deux Sèvres, du Finistère, de l'Ile et Villaine, du Loir et Cher, de Loire Atlantique, du Loiret, de la Vienne, du Morbihan, de l'Eure et Loire), ITMO Cancer of Aviesan within the framework of the 2021-2030 cancer control strategy, funds administered by Inserm, SFR FED4226 Neuro, the fédération de recherche "Physique et chimie du vivant" FR CNRS 2708 and the Réseau ‘Molécules Marines, Métabolisme & Cancer’ of the Cancéropôle Grand Ouest for their financial support. This project received funding from the European Union's Horizon 2020 research and innovation program under the Marie Sklodowska-Curie agreement No. 898850. F. B. and S. R. thank the projects CHemBio (FEDER-FSE 2014-2020-EX003677), Valbiocosm (FEDER-FSE 2014-2020-EX003202), Techsab (FEDER-FSE 2014-2020-EX011313), and QUALICHIM (APR-IA-PF 2021-00149467) for their financial support of ICOA, UMR 7311, University of Orléans, CNRS. We also thank the SALSA and MO2VING platforms for spectroscopic measurements and chromatographic analyses (NMR, HPTLC, HPLC, MS, HRMS). C. P.-I. thanks the Centro de Supercomputación de Galicia (CESGA) for providing access to supercomputing facilities. S. P. acknowledges support from the Institut National de la Santé et de la Recherche Médicale (INSERM).

References

1. J.-C. G. Bunzli, Acc. Chem. Res., 2006, 39, 53–61.
2. D. Parker, R. S. Dickins, H. Puschmann, C. Crossland and J. A. K. Howard, Chem. Rev., 2001, 102, 1977–2010.
3. A. E. Merbach, L. Helm and E. Toth, The Chemistry of Contrast Agents in Medical Magnetic Resonance Imaging, John Wiley & Sons, Chichester, 2nd edn, 2013.
4. J. Wahsner, E. M. Gale, A. Rodríguez-Rodríguez and P. Caravan, Chem. Rev., 2019, 119, 957–1057.
5. S. V. Eliseeva and J.-C. G. Bunzli, Chem. Soc. Rev., 2010, 39, 189.
6. S. Sun, Y. Zhao, J. Wang and R. Pei, J. Mater. Chem. B, 2022, 10, 9535–9564.
7. S. Lacerda, K. Djanashvili and C. S. Bonnet, in Supramolecular Chemistry in Biomedical Imaging, ed. S. Faulkner, T. Gunnlaugsson and G. O. Maille, The Royal Society of Chemistry, 2022, pp. 163–206.  10.1039/9781782624028-00163.
8. C. Herlan and S. Bräse, Dalton Trans., 2020, 49, 2397–2402.
9. M. Van de Voorde, K. Van Hecke, T. Cardinaels and K. Binnemans, Coord. Chem. Rev., 2019, 382, 103–125.
10. N. S. Chundawat, S. Jadoun, P. Zarrintaj and N. P. S. Chauhan, Polyhedron, 2021, 207, 115387.
11. J.-C. G. Bunzli, Chem. Rev., 2010, 110, 2729–2755.
12. E. Brucher, G. Tircso, Z. Baranayai, Z. Kovacs and A. D. Sherry, in The Chemistry of Contrast Agents in Medical Magnetic Resonance Imaging, ed. A. E. Merbach, L. Helm and E. Toth, John Woley & Sons, Chichester, 2nd edn, 2013, ch. 4, pp. 157–208.
13. S. Shuvaev, M. Starck and D. Parker, Chem. – Eur. J., 2017, 23, 9974–9989.
14. K. P. Malikidogo, H. Martin and C. S. Bonnet, Pharmaceuticals, 2020, 13, 436.
15. J. Lux and A. D. Sherry, Curr. Opin. Chem. Biol., 2018, 45, 121–130.
16. A. D. Sherry, D. D. Castelli and S. Aime, NMR Biomed., 2023, 36, e4698.
17. S. Lacerda, Inorganics, 2018, 6, 129.
18. P. Désogère, S. B. Montesi and P. Caravan, Chem. – Eur. J., 2019, 25, 1128–1141.
19. E. Gianolio, R. Stefania, E. Di Gregorio and S. Aime, Eur. J. Inorg. Chem., 2012, 2012, 1934–1944.
20. J. He, C. S. Bonnet, S. V. Eliseeva, S. Lacerda, T. Chauvin, P. Retailleau, F. Szeremeta, B. Badet, S. Petoud, E. Toth and P. Durand, J. Am. Chem. Soc., 2016, 138, 2913–2916.
21. N. Hamon, L. Bridou, M. Roux, O. Maury, R. Tripier and M. Beyler, J. Org. Chem., 2023, 88, 8286–8299.
22. G. Nizou, C. Favaretto, F. Borgna, P. V. Grundler, N. Saffon-Merceron, C. Platas-Iglesias, O. Fougère, O. Rousseaux, N. P. van der Meulen, C. Müller, M. Beyler and R. Tripier, Inorg. Chem., 2020, 59, 11736–11748.
23. M. Le Fur, E. Molnár, M. Beyler, O. Fougère, D. Esteban-Gómez, O. Rousseaux, R. Tripier, G. Tircsó and C. Platas-Iglesias, Inorg. Chem., 2018, 57, 6932–6945.
24. C. P. Montgomery, B. S. Murray, E. J. New, R. Pal and D. Parker, Acc. Chem. Res., 2009, 42, 925–937.
25. L. M. T. Frija, A. J. L. Pombeiro and M. N. Kopylovich, Coord. Chem. Rev., 2016, 308, 32–55.
26. G. Mercuri, G. Giambastiani and A. Rossin, Inorganics, 2019, 7, 144.
27. B. A. Bhongade, S. Talath, R. A. Gadad and A. K. Gadad, J. Saudi Chem. Soc., 2016, 20, S463–S475.
28. H. A. Mohamed and B. F. Abdel-Wahab, J. Sulfur Chem., 2012, 33, 589–604.
29. S. Saba, in Comprehensive Heterocyclic Chemistry II, ed. A. R. Katritzky, C. W. Rees and E. F. V. Scriven, Pergamon, Oxford, 1996, pp. 191–198.  DOI:10.1016/B978-008096518-5.00173-8.
30. A. Cristina, D. Leonte, L. Vlase, L. C. Bencze, S. Imre, G. Marc, B. Apan, C. Mogoşan and V. Zaharia, Molecules, 2018, 23, 2425.
31. N. Terzioglu and A. Gürsoy, Eur. J. Med. Chem., 2003, 38, 781–786.
32. R. Romagnoli, P. G. Baraldi, F. Prencipe, J. Balzarini, S. Liekens and F. Estévez, Eur. J. Med. Chem., 2015, 101, 205–217.
33. W. S. Alwan, R. Karpoormath, M. B. Palkar, H. M. Patel, R. A. Rane, M. S. Shaikh, A. Kajee and K. P. Mlisana, Eur. J. Med. Chem., 2015, 95, 514–525.
34. J. Marco-Contelles, E. Pérez-Mayoral and P. Ballesteros, in Comprehensive Heterocyclic Chemistry III, ed. A. R. Katritzky, C. A. Ramsden, E. F. V. Scriven and R. J. K. Taylor, Elsevier, Oxford, 2008, pp. 199–305.  DOI:10.1016/B978-008044992-0.01005-1.
35. H. M. Patel, B. Sing, V. Bhardwaj, M. Palkar, M. S. Shaikh, R. Rane, W. S. Alwan, A. K. Gadad, M. N. Noolvi and R. Karpoormath, Eur. J. Med. Chem., 2015, 93, 599–613.
36. S. Cascioferro, G. Li Petri, B. Parrino, B. El Hassouni, D. Carbone, V. Arizza, U. Perricone, A. Padova, N. Funel, G. J. Peters, G. Cirrincione, E. Giovannetti and P. Diana, Molecules, 2020, 25, 329.
37. C. Pescheteau, M. Place, A. Sava, L. Nunes, L. Profire, S. Routier and F. Buron, RSC Adv., 2022, 12, 6303–6313.
38. M. Place, C. Copin, M. Apotrosoaei, S. Constantin, I. M. Vasincu, L. Profire, F. Buron and S. Routier, J. Org. Chem., 2017, 82, 13700–13707.
39. C. Copin, N. Henry, F. Buron and S. Routier, Synlett, 2016, 1091–1095.
40. C. Copin, F. Buron and S. Routier, Eur. J. Org. Chem., 2016, 1958–1962.
41. C. Copin, S. Massip, J.-M. Léger, C. Jarry, F. Buron and S. Routier, Eur. J. Org. Chem., 2015, 6932–6942.
42. C. Copin, N. Henry, F. Buron and S. Routier, Eur. J. Org. Chem., 2012, 3079–3083.
43. M. Ndiaye, E. Caillet, S. V. Eliseeva, T. Masson, L. A. Marchetti, A. Pallier, J.-F. Morfin, S. Petoud, S. Routier, F. Buron and C. S. Bonnet, Eur. J. Inorg. Chem., 2024, 27, e202300648.
44. M. Regueiro-Figueroa, B. Bensenane, E. Ruscsák, D. Esteban-Gómez, L. J. Charbonnière, G. Tircsó, I. Tóth, A. d. Blas, T. Rodríguez-Blas and C. Platas-Iglesias, Inorg. Chem., 2011, 50, 4125–4141.
45. S. Laine, R. Jouclas, C. S. Bonnet, P. Retailleau, V. Steinmetz, A. Pallier, Z. Garda, G. Tircsó, P. Durand and É. Tóth, Eur. J. Inorg. Chem., 2024, e202300784.
46. K. Kumar, C. A. Chang, L. C. Francesconi, D. D. Dischino, M. F. Malley, J. Z. Gougoutas and M. F. Tweedle, Inorg. Chem., 1994, 33, 3567–3575.
47. C. Harriswangler, J. C. Frías, M. T. Albelda, L. Valencia, E. García-España, D. Esteban-Gómez and C. Platas-Iglesias, Inorg. Chem., 2023, 62, 17030–17040.
48. J. A. K. Howard, A. M. Kenwright, J. M. Moloney, D. Parker, M. Woods, J. A. K. Howard, M. Port, M. Navet and O. Rousseau, Chem. Commun., 1998, 1381–1382,  10.1039/A802847H.
49. M. Storm Thomsen, H. O. B. Andersen and T. J. Sørensen, Dalton Trans., 2022, 51, 14118–14124.
50. M. Regueiro-Figueroa, E. Ruscsák, L. Fra, G. Tircsó, I. Tóth, A. de Blas, T. Rodríguez-Blas, C. Platas-Iglesias and D. Esteban-Gómez, Eur. J. Inorg. Chem., 2014, 2014, 6165–6173.
51. S. Gündüz, S. Vibhute, R. Botár, F. K. Kálmán, I. Tóth, G. Tircsó, M. Regueiro-Figueroa, D. Esteban-Gómez, C. Platas-Iglesias and G. Angelovski, Inorg. Chem., 2018, 57, 5973–5986.
52. A. Rodríguez-Rodríguez, M. Regueiro-Figueroa, D. Esteban-Gómez, T. Rodríguez-Blas, V. Patinec, R. Tripier, G. Tircsó, F. Carniato, M. Botta and C. Platas-Iglesias, Chem. – Eur. J., 2017, 23, 1110–1117.
53. Z. Garda, V. Nagy, A. Rodríguez-Rodríguez, R. Pujales-Paradela, V. Patinec, G. Angelovski, É. Tóth, F. K. Kálmán, D. Esteban-Gómez, R. Tripier, C. Platas-Iglesias and G. Tircsó, Inorg. Chem., 2020, 59, 8184–8195.
54. Z. Baranyai, Z. Pálinkás, F. Uggeri, A. Maiocchi, S. Aime and E. Brücher, Chem. – Eur. J., 2012, 18, 16426–16435.
55. A. Takács, R. Napolitano, M. Purgel, A. C. Bényei, L. Zékány, E. Brücher, I. Tóth, Z. Baranyai and S. Aime, Inorg. Chem., 2014, 53, 2858–2872.
56. E. J. Corey and J. C. Bailar Jr., J. Am. Chem. Soc., 1959, 81, 2620–2629.
57. J. K. Beattie, Acc. Chem. Res., 1971, 4, 253–259.
58. S. Aime, M. Botta and G. Ermondi, Inorg. Chem., 1992, 31, 4291–4299.
59. C. Platas-Iglesias, Eur. J. Inorg. Chem., 2012, 2012, 2023–2033.
60. Y. Xing, A. K. Jindal, M. Regueiro-Figueroa, M. Le Fur, N. Kervarec, P. Zhao, Z. Kovacs, L. Valencia, P. Pérez-Lourido, R. Tripier, D. Esteban-Gómez, C. Platas-Iglesias and A. D. Sherry, Chem. – Eur. J., 2016, 22, 16657–16667.
61. J. C. Facelli, Prog. Nucl. Magn. Reson. Spectrosc., 2011, 58, 176–201.
62. J.-C. G. Bünzli, Coord. Chem. Rev., 2015, 293–294, 19–47.
63. D. Parker, J. D. Fradgley and K.-L. Wong, Chem. Soc. Rev., 2021, 50, 8193–8213.
64. A. Beeby, D. Parker and J. A. G. Williams, J. Chem. Soc., Perkin Trans. 2, 1996, 1565–1579.
65. E. Toth, L. Helm and A. E. Merbach, in The Chemistry of Contrast Agents in Medical Magnetic Resonance Imaging, ed. A. E. Merbach, L. Helm and E. Toth, John Wiley & Sons, Chichester, 2nd edn, 2013, ch. 2, pp. 25–81.
66. P. H. Fries and E. Belorizky, J. Chem. Phys., 2005, 123, 124510.
67. P. Mieville, H. Jaccard, F. Reviriego, R. Tripier and L. Helm, Dalton Trans., 2011, 40, 4260.
68. P. Caravan, J. J. Ellison, T. J. McMurry and R. B. Lauffer, Chem. Rev., 1999, 99, 2293–2352.
69. B. Halle and H. Wennerstrom, J. Chem. Phys., 1981, 75, 1928–1943.
70. D. H. Powell, O. M. NiDhubhghaill, D. Pubanz, L. Helm, Y. S. Lebedev, W. Schlaepfer and A. E. Merbach, J. Am. Chem. Soc., 1996, 118, 9333–9346.
71. E. Toth, L. Helm and A. Merbach, in The Chemistry of Contrast Agents in Medical Magnetic resonance Imaging, ed. A. Merbach, L. Helm and E. Toth, John Wiley & Sons, Chichester, 2nd edn, 2013, pp. 25–81.
72. A. E. Martell and R. J. Motekaitis, Determination and use of stability constants, VCH, 1992.
73. P. Gans, A. Sabatini and A. Vacca, Talanta, 1996, 43, 1739–1753.
74. R. M. Smith, R. J. Motekaitis and A. E. Martell, NIST Standard Reference Database.
75. P. Gans, A. Sabatini and A. Vacca, Ann. Chim., 1999, 89, 45–49.
76. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato, A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman and D. J. Fox, Gaussian 16, Revision B.01, Gaussian, Inc., Wallingford CT, 2016.
77. J.-D. Chai and M. Head-Gordon, Phys. Chem. Chem. Phys., 2008, 10, 6615–6620.
78. J. Tao, J. P. Perdew, V. N. Staroverov and G. E. Scuseria, Phys. Rev. Lett., 2003, 91, 146401.
79. M. Dolg, H. Stoll, A. Savin and H. Preuss, Theor. Chim. Acta, 1989, 75, 173–194.
80. K. A. Peterson, D. Figgen, M. Dolg and H. Stoll, J. Chem. Phys., 2007, 126, 124101.
81. F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305.
82. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–3094.
83. O. V. Yazyev, L. Helm, V. G. Malkin and O. L. Malkina, J. Phys. Chem. A, 2005, 109, 10997–11005.
84. R. Ditchfield, J. Chem. Phys., 2003, 56, 5688–5691.
85. T. Helgaker, M. Jaszuński and K. Ruud, Chem. Rev., 1999, 99, 293–352.
86. F. Neese, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2012, 2, 73–78.
87. F. Neese, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2018, 8, e1327.
88. E. v. Lenthe, E. J. Baerends and J. G. Snijders, J. Chem. Phys., 1993, 99, 4597–4610.
89. J. D. Rolfes, F. Neese and D. A. Pantazis, J. Comput. Chem., 2020, 41, 1842–1849.
90. F. Neese, J. Comput. Chem., 2003, 24, 1740–1747.
91. S. Kossmann and F. Neese, Chem. Phys. Lett., 2009, 481, 240–243.
92. R. Izsák and F. Neese, J. Chem. Phys., 2011, 135, 144105.
93. G. L. Stoychev, A. A. Auer, R. Izsák and F. Neese, J. Chem. Theory Comput., 2018, 14, 619–637.
94. F. Weigend, Phys. Chem. Chem. Phys., 2006, 8, 1057–1065.
95. A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys. Chem. B, 2009, 113, 6378–6396.
96. D. S. Raiford, C. L. Fisk and E. D. Becker, Anal. Chem., 1979, 51, 2050–2051.
97. S. Meiboom and D. Gill, Rev. Sci. Instrum., 1958, 29, 688–691.
98. K. Micskei, L. Helm, E. Brucher and A. E. Merbach, Inorg. Chem., 1993, 32, 3844–3850.
99. A. D. Hugi, L. Helm and A. E. Merbach, Helv. Chim. Acta, 1985, 68, 508–521.
100. F. Yerly, VISUALISEUR 2.3.5 and OPTIMISEUR 2.3.5, Lausanne, Switzerland, 1999.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4dt00533c

---