Article pubs.acs.org/IC
Evaluation of Water Exchange Kinetics on [Ln(AAZTAPh− NO2)(H2O)q]x Complexes Using Proton Nuclear Magnetic Resonance Shima Karimi,† Lorenzo Tei,‡ Mauro Botta,‡ and Lothar Helm*,† †
Laboratoire de Chimie Inorganique et Bioinorganique, Ecole Polytechnique Fédérale de Lausanne (EPFL-BCH), CH-1015 Lausanne, Switzerland ‡ Dipartimento di Scienze e Innovazione Tecnologica, Università del Piemonte Orientale “A. Avogadro”, Viale T. Michel 11, 15121 Alessandria, Italy S Supporting Information *
ABSTRACT: Water exchange kinetics on [Ln(AAZTAPh−NO2)(H2O)q]− (Ln = Gd3+, Dy3+, or Tm3+) were determined by 1H nuclear magnetic resonance (NMR) measurements. The number of inner-sphere water molecules was found to change from two to one when going from Dy3+ to Tm3+. The calculated water exchange rate constants obtained by variable-temperature proton transverse relaxation rates are 3.9 × 106, 0.46 × 106, and 0.014 × 106 s−1 at 298 K for Gd3+, Dy3+, and Tm3+, respectively. Variable-pressure measurements were used to assess the water exchange mechanism. The results indicate an associative and dissociative interchange mechanism for Gd3+ and Dy3+ complexes with ΔV⧧ values of −1.4 and 1.9 cm3 mol−1, respectively. An associative activation mode (Ia or A mechanism) was obtained for the Tm3+ complex (ΔV⧧ = −5.6 cm3 mol−1). Moreover, [Dy(AAZTAPh−NO2)(H2O)2]− with a very high transverse relaxivity value was found as a potential candidate for negative contrast agents for high-field imaging applications.
■
INTRODUCTION The heptadentate polyaminocarboxylate ligand AAZTA (6amino-6-methylperhydro-1,4-diazepine tetraacetic acid) was studied extensively as a potential MRI probe. The Gd3+ complex of AAZTA, having two inner-sphere water molecules, has shown excellent relaxation enhancement properties with high thermodynamic stability in aqueous solution and a nearly complete inertness toward the influence of bidentate endogenous anions (r1 = 7.1 mM−1 s−1 at 20 MHz and 298 K; τm = 90 ns).1,2 In the study presented here, the water exchange properties of lanthanide complexes of an AAZTA derivative, AAZTAPh−NO2 (Scheme 1), [Ln(AAZTAPh− NO2)(H2O)q]− (Ln = Gd3+, Dy3+, or Tm3+), have been determined. Possible variation of the number of coordinated water molecules with different lanthanide complexes of AAZTAPh−NO2 is also discussed. The water exchange rate and its activation parameters can be extracted experimentally by either measuring the temperature dependence of 17O nuclear magnetic resonance (NMR) or 1H NMR relaxation rates of bulk water induced by the presence of a paramagnetic solution. The advantage of 17O NMR measurements is that the outersphere contribution to both transverse and longitudinal 17O relaxation rates can be neglected.3 However, this technique requires relatively high concentrations of Ln3+ (>5 mM),4 sometimes much higher than the limits of the solubility of complexes. The solubility of the lanthanide complexes of AAZTAPh−NO2 (L1) in water at 25 °C is at most 6 mM © XXXX American Chemical Society
because of the nitrophenyl moiety, which greatly reduces the solubility. Such low values preclude accurate 17O NMR measurements for lanthanides other than Gd3+. In this regard, the water exchange rates can be extracted by measuring the temperature dependence of the 1H NMR transverse relaxation rates of bulk water. It has to be noted that the measurement has to be taken at physiological pH, where the proton exchange rate is assumed to be equal to the water exchange rate because with an increase in the acidity or basicity of the solution, the proton exchange may become faster than the water exchange due to the acid- or base-catalyzed pathway.5 Moreover, the outersphere relaxation contribution cannot be neglected during analysis of the 1H NMR data. To assess the accuracy of the water exchange values obtained by proton NMR, water exchange rates of dysprosium(III) complexes of ligands DTTA−Me (L2) and DO3A (L3) (Scheme 1) have been studied. The results were compared with the previously reported values6 obtained by 17O NMR measurements. Moreover, dysprosium(III) complexes of the three ligands (L1−3) are compared and discussed as potential negative contrast agents for very high-field MRI. Furthermore, variable-pressure measurements were taken and used to assign the reaction mechanisms by the determination of the activation volume, ΔV⧧, as a direct Received: April 19, 2016
A
DOI: 10.1021/acs.inorgchem.6b00976 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry Scheme 1. Drawing of Ligands H4AAZTAPh−NO2, H4DTTA−Me, and H3DO3A
⎞ τe2 S(S + 1) ⎛ A ⎞2 ⎛ 1 ⎜ ⎟ ⎜τ + ⎟ = e1 2 2 ⎝ℏ⎠ ⎝ T2,Sc 3 1 + ωs τe2 ⎠
measure of the degree of associativity or dissociativity at the transition state. Again, because of the low solubility of the compounds, the proton NMR method is used. The results were compared with the values obtained by 17O NMR when applicable.
2 2 2 2 13τc2 1 1 ⎛ μ0 ⎞ γI μB μeff ⎛ ⎜ ⎟ ⎜4τc1 + = T2,d 15 ⎝ 4π ⎠ 1 + ωs 2τc2 2 r6 ⎝
■
SYNTHESIS AAZTAPh−NO2 (L1) was synthesized by deprotection of an intermediate used for the preparation of functionalized AAZTA derivatives.2 In particular, the tert-butyl ester-protected derivative of AAZTAPh−NO2 was reacted with a 1:1 mixture of trifluoroacetic acid (TFA) and CH2Cl2, yielding the final ligand L1 that was fully characterized by mass spectrometry and NMR spectroscopy.
+
1 T2,Cu
■
DATA TREATMENT H NMR Relaxation. The measured transverse proton relaxation rate, 1/T2,obs, is the sum of a diamagnetic and a paramagnetic contribution: T2,obs
=
P 1 1 1 + = + m T2,d T2,p T2,d T2r
(1)
where 1/T2r is the reduced transverse relaxation rate and Pm is the mole fraction of the bound water. The paramagnetic relaxation is generally divided into two components, innersphere and outer-sphere: ⎛ 1 ⎞IS ⎛ 1 ⎞OS 1 ⎟⎟ + ⎜⎜ ⎟⎟ = ⎜⎜ T2,p ⎝ T2,p ⎠ ⎝ T2,p ⎠
(5)
2 4 4 2 ⎞ 3τcs 1 ⎛ μ0 ⎞ ωI μB μeff ⎛ ⎜ ⎟ ⎜4τcs + ⎟ 2 2 2 6 5 ⎝ 4π ⎠ (3kT ) r ⎝ 1 + ωI τcs ⎠
(6)
⎛ ⎛ 1 ⎞ZFS 1 2 1 Δ τv[4S(S + 1) − 3]⎜ ⎜ ⎟ = 2 2 25 1 + ω ⎝ T1e ⎠ ⎝ s τv
(2)
The inner-sphere relaxation contribution is obtained from eq 3, where T2,m is the transverse proton relaxation of the bound water, τm is the bound water lifetime (kex = τm−1), and Δωm is the chemical shift difference between the bound water and the bulk water. ⎛ 1 ⎞IS T −2 + τ −1T −1 + Δωm 2 P ⎜⎜ ⎟⎟ = m × 2,m −1 m −2,m τm (τm + T2,m 1)2 + Δωm 2 ⎝ T2,p ⎠
=
⎞ ⎟ 1 + ωI 2τc12 ⎠ 3τc1
where μeff2 = ge2J(J + 1), 1/τci = 1/Tie + 1/τM + 1/τR, and 1/τcs = 1/τM + 1/τR (where i = 1 or 2), γI is the nuclear gyromagnetic ratio, ge is the Landé g factor, μB is the Bohr magneton, μeff is the effective magnetic moment of the Ln3+ ion, J is the total angular quantum number, ωS and ωI are the electron and nucleus Larmor frequencies, respectively, and r is the electron spin−proton distance. τR is the rotational correlation time of the metal−proton vector, and T1e and T2e are the longitudinal and transverse electronic relaxation times of the metal ion, respectively. In the case of Gd3+, the electron spin relaxation rates are governed by the zero-field splitting (ZFS) and the rates can be expressed by the Bloembergen−Morgan theory of paramagnetic electron spin relaxation (eqs 7 and 8),7,8 where Δ2 is the mean square fluctuation of the ZFS and τv is the correlation time for the modulation of the ZFS with activation energy Ev.
1
1
(4)
+
⎞ 4 ⎟ 1 + 4ωs 2τv 2 ⎠
(7)
⎛ ⎛ 1 ⎞ZFS 1 2 5 Δ τv[4S(S + 1) − 3]⎜ ⎜ ⎟ = 2 2 50 ⎝ T2e ⎠ ⎝ 1 + ωs τv
(3)
+
1/T2,m is considered to be the sum of scalar (1/T2,sc), dipolar (1/T2,d), and Curie (1/T2,Cu) contributions as given by B
⎞ 2 + 3⎟ 2 2 1 + 4ωs τv ⎠
(8) DOI: 10.1021/acs.inorgchem.6b00976 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry ⎡E ⎛ 1 1 ⎞⎤ ⎟⎥ τv = τv298 exp⎢ v ⎜ − ⎣ R ⎝T 298.15 ⎠⎦
⎛ 1 ⎞OS 48π ⎛ μ ⎞2 N [ln] ωI 2μ 4 μ 4 0⎟ B eff A ⎜ ⎜⎜ ⎟⎟ = [4J(0) + 3J(ωI)] 405 ⎝ 4π ⎠ aLnHDLnH (3kBT )2 ⎝ T2,Cu ⎠ (15)
(9)
For all paramagnetic Ln3+ except Gd3+, a single electronic relaxation time Te may be introduced as the values of T2e are reported to be very small, near those of T1e.9 Moreover, because Te (∼10−13) ≪ τR (∼10−10), τc is dominated by the values of Te. For Curie relaxation, τcs is given by the rotational correlation time as τR ≪ τm. At high magnetic fields (≥7 T), the Curie spin relaxation becomes the dominant contribution to the water 1H relaxation and the scalar contribution can be neglected. Moreover, when τmΔωm2 ≫ 1/T2,m, eq 3 can be simplified to eq 10: ⎛ 1 ⎞IS τmΔωm 2 ⎜⎜ ⎟⎟ = Pm × 1 + τm 2Δωm 2 ⎝ T2,p ⎠
1 ⎡ ⎤ 1 + 4 (iωτd)1/2 ⎥ J(ω) = Re⎢ ⎢ 1 + (iωτ )1/2 + 4 (iωτ ) + 1 (iωτ )3/2 ⎥ d d d ⎣ ⎦ 9 9
The paramagnetic chemical shift (Δωp) is a function of the shift due to the inner-sphere water (Δωm) and the shift of waters in the second coordination sphere (ΔωOS): ⎡ ⎤ Δωm ⎥ Δωp = Pm⎢ + Δ ω OS ⎢⎣ (τmT2,m−1 + 1)2 + τm 2Δωm 2 ⎥⎦
The outer-sphere contribution to the chemical shift is assumed to be linearly related to Δωm, through an empirical constant COS.
(10)
Furthermore, under the fast exchange condition, where τm 2Δωm 2 ≪ 1, the transverse relaxation rate will be proportional to the product of water residence time and the square of the chemical shift (eq 11), whereas under the slow exchange condition (where τm2Δωm2 ≫ 1), inner-sphere transverse relaxation rate will be independent of Δωm and will decrease with a decrease in the water exchange rate. In this case, no chemical shift data will be necessary for the assessment of the water exchange rate (eq 12). ⎛ 1 ⎞IS ⎜⎜ ⎟⎟ = PmτmΔωm 2 ⎝ T2,p ⎠
(11)
⎛ 1 ⎞IS P ⎜⎜ ⎟⎟ = m τm ⎝ T2,p ⎠
(12)
ΔωOS = COSΔωm
⎛B B ⎞ (ΔωmT)p = ω0⎜ 1 + 22 ⎟ ⎝T T ⎠
(ΔωmP)T = (Δωm0)T (1 + PP 1 )
298 ⎛ ΔS ⧧ kT ΔH ⧧ ⎞ (kex )p T 1 ⎟= = B exp⎜ − h RT ⎠ 298.15 τm ⎝ R ⎡ ΔH ⧧ ⎛ 1 1 ⎞⎤ ⎜ − ⎟⎥ exp⎢ T ⎠⎦ ⎣ R ⎝ 298.15 (20)
(kexT )p =
The pressure dependence of the water exchange rate is described by eq 21, where ΔV⧧ is the activation volume. Δβ⧧ is the compressibility of activation that is usually very small for aqueous complexes,19 and k0ex refers to the exchange rate constant at zero pressure.
(14)
ln(kexP )T = ln(kex0 )T −
■
Jk (ω)
(
1+ τd Tke
1/2
)
+
1 4
1/2
⎤ ⎥ ⎥, 3/2 ⎥ ⎥⎦
4 9
d
τd Tke
τd Tke
1 9
d
τd Tke
P 2Δβ ⧧ P ΔV ⧧ + RT RT
(21)
RESULTS AND DISCUSSION Water Exchange on [DyL2,3(H2O)2]x Determined by 1H NMR Relaxation. To verify that water exchange on polyaminocarboxylate complexes of lanthanide(III) ions can be studied quantitatively by 1H NMR relaxation, proton
(iωτ + ) (iωτ + ) + (iωτ + ) d
(19)
The temperature dependence of the water exchange rate is described by eq 20,18 where ΔS⧧ and ΔH⧧ are the entropy and enthalpy of activation for the water exchange process, respectively. k298 ex is the water exchange at 298 K. R is the perfect gas constant. h and kB and are the Planck and Boltzmann constants, respectively.
⎛ 1 ⎞OS 16π ⎛ μ ⎞2 N [ln] 0⎟ A ⎜ ⎜⎜ ⎟⎟ = γI 2μB 2 ⎝ ⎠ T 405 4 a D π ⎝ 2,d ⎠ LnH LnH
⎡ ⎢ = Re⎢ ⎢ ⎢⎣ 1 + iωτd +
(18)
where P and T are the experimental pressure and temperature, respectively. Moreover, on the basis of our experimental data, a linear pressure dependence of Δωm is considered17 (eq 19), where (Δω0m)T refers to Δωm at zero pressure and P1 is the proportionality factor.
(13)
μeff 2 [4J1(0) + 3J1(ωI) + 13J2 (ωS)]
(17)
The Δωm is proportional to the magnetic field and its temperature dependence is the sum of contact and pseudocontact terms. Considering B1 and B2 (equal to zero for Gd3+) constants described by Lewis15 and Bleaney16 for Δωm, one can write
The outer-sphere relaxation rates are given as the sum of dipolar and Curie relaxations10 (eq 13). The dipole−dipole relaxation is described by eq 14 as developed by Freed11,12 and Ayant,13 and the Curie relaxation has been described by Fries14 (eq 15), where NA is Avogadro’s constant, aLnH is the distance of the closest approach of a second-coordination-sphere water proton to the metal center, DLnH is the mutual diffusion of bulk water and the complex, τd is the correlation time for the translational diffusion such that τd = aLnH2/DLnH, and J(ω) is the spectral density function. ⎛ 1 ⎞OS ⎛ 1 ⎞OS ⎛ 1 ⎞OS ⎜⎜ ⎟⎟ = ⎜⎜ ⎟⎟ + ⎜⎜ ⎟⎟ ⎝ T2,d ⎠ ⎝ T2,Cu ⎠ ⎝ T2,p ⎠
(16)
k = 1, 2 C
DOI: 10.1021/acs.inorgchem.6b00976 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry Table 1. Water Exchange Propertiesa of [Dy(L2,3)(H2O)2]x Obtained by 1H and 17O NMR
[Dy(DTTA−Me)(H2O)2]−
[Dy(DO3A)(H2O)2] 17
k298 ex k310 ex
6
−1
(×10 s ) (×106 s−1) ⧧ ΔH (kJ mol−1) ΔS⧧ (J mol−1) ΔV⧧ (cm3 mol−1) a
b
1
O
16.6 ± 0.3 27.1 ± 0.5 29 ± 1 −9 ± 3 −0.5 ± 0.3
H
19 ± 4 28 ± 6 20 ± 1 −38 ± 2 −0.2 ± 0.1c
Ob
17
1
H
40.6 ± 0.9 74.1 ± 1.7 36 ± 1 +21 ± 2 +1.8 ± 0.2
45 ± 10 75 ± 17 29 ± 1 0±2 −
For the full list of parameters, see the Supporting Information. bFrom ref 6. cNo outer-sphere contribution for chemical shift is considered.
Figure 1. 1H NMR temperature dependence of ln(1/T2r) of (A) [Dy(DO3A)(H2O)2] and (B) [Dy(DTTA−Me)(H2O)2]− at pH 5.7, B0 = 9.4 T (■) and 18.8 T (●), and ∼100 mM Ln3+. Dashed lines correspond to the best fits of the experimental data.
transverse relaxation rates and chemical shifts of [Dy(DTTA− Me)(H2O)2]− and [Dy(DO3A)(H2O)2] were measured. Experimental data have been fitted using the full Swift and Connick equation (eq 3) for the inner-sphere relaxation contribution. The following parameters have been fixed: q = 2, −10 m2 s−1. The aLnH = 3.6 × 10−10 m, and D298 LnH = 25 × 10 diffusion constant is assumed to obey the Arrhenius law. The electronic relaxation time (Te) and rotational correlation time (τR,H) were fixed to the values obtained from 17O NMR measurements.6 Moreover, because COS which indicates the outer-sphere contribution to the chemical shift is ill-defined, we have fitted the data two times, fixing COS to the two limiting values of 0 and 0.25. In this regard, the errors reported in Table 1 represent the effect of COS variation on the water exchange parameters. The experimental and calculated results are shown in Figures 1 and 2, and the corresponding parameters are listed in Table 1. It is apparent from Figure 1 that both [DyL2,3(H2O)2]x compounds are in the fast exchange region as defined by kex ≫ 1/T2m, Δωm over the whole temperature range studied. The obtained exchange rate constants and activation parameters, including ΔV⧧, are close to the values previously reported from 17 O NMR measurements,6 supporting the hypothesis that reliable kex values can be obtained by proton NMR measurements. It has to be noted that the outer-sphere contribution to
Figure 2. 1H NMR pressure dependence of ln(1/T2r) of [Dy(DO3A)(H2O)2] at 20 °C, pH 5.7, B0 = 9.4 T, and 86.32 mM Dy3+. Dashed lines correspond to the best fitting of the experimental data.
the total proton relaxation rates is at most 3% and 9% at 18.8 T for Dy3+ complexes of DO3A and DTTA−Me, respectively. Although 1H NMR measurements provide reliable results, because of uncertainties in outer-sphere contributions to the chemical shifts, 17O NMR measurements should be considered as the first choice for the assessment of water exchange rate constants if the solubility limits of the samples permit. D
DOI: 10.1021/acs.inorgchem.6b00976 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
Figure 3. 1H NMRD profiles of [Ln(AAZTAPh−NO2)(H2O)q]− (red diamonds) at ∼5 mM Ln3+ and of [Ln(DO3A)(H2O)2] (blue circles) and [Ln(DTTA−Me)(H2O)2]− (■) at ∼100 mM Ln3+ reported in ref 6 at 25 °C (Ln = Dy3+ or Tm3+).
Figure 4. 1H NMR temperature dependence of ln(1/T2r) of [Ln(AAZTAPh−NO2)(H2O)q]− (Ln = Gd3+, Dy3+, and Tm3+) at pH 5.7, B0 = 9.4 T (■) and 18.8 T (●), and ∼5 mM Ln3+. Dashed lines correspond to the best fits of the experimental data.
Number of Inner-Sphere Water Molecules on [Ln(AAZTAPh−NO2)(H2O)q]−. The Gd3+ complex of AAZTAPh− NO2 was considered to have two water molecules in the first coordination sphere, as found for the parent Gd−AAZTA complex.1 However, the number of water molecules coordinated to the other lanthanide complexes of AAZTAPh−NO2 can possibly change when moving from Gd3+ to Tm3+. The hydration state of lanthanide chelates can be assessed by different techniques such as luminescence lifetime measurements for Eu3+ and Tb3+ chelates or analysis of the 17O NMR lanthanide chemical shift data.20 As mentioned before, the latter cannot be used in this study because of low chelate concentrations. However, because inner-sphere proton relaxivity is linearly proportional to the number of coordinated water molecules, 1H nuclear magnetic relaxation dispersion (NMRD) profiles can be used for the assessment of hydration number variation. In this regard, longitudinal water proton relaxation rate measurements were taken for both Dy3+ and Tm3+ complexes. The results are compared with the previously reported NMRD profiles6 for the q = 2 chelates [LnDO3A(H2O)2] and [Ln(DTTA−Me)(H2O)2]− (Ln = Dy3+ or Tm3+) (Figure 3). As shown in Figure 3, the decrease in the relaxivity of [Tm(AAZTAPh−NO2)(H2O)q]− compared to the those of the other [TmL2,3(H2O)2]x complexes indicates a change in the number q of inner-sphere water molecules. However, the relaxivity of [Dy(AAZTAPh−NO2)(H2O)q]− is in good agreement with those observed previously for [DyL2,3(H2O)2]x complexes. In this regard, hydration numbers of 2 and 1 for
dysprosium(III) and thulium(III) complexes of AAZTAPh− NO2, respectively, were assumed for the further studies. Water Exchange on [Ln(AAZTAPh−NO2)(H2O)q]−. Water exchange on [Ln(AAZTAPh−NO2)(H2O)q]− has been studied by 1H NMR relaxation for Ln = Gd3+, Dy3+, and Tm3+. Because of the low lanthanide concentration (∼5 mM), very small chemical shift differences (Δωp/2π < 40 Hz) were observed for Dy3+ and Tm3+ chelates. However, as mentioned above, in the slow exchange regime, the inner-sphere transverse relaxation rate will be independent of Δωm and no chemical shift data will be necessary for the assessment of the water exchange rate constants. In the case of [Gd(AAZTAPh− NO2)(H2O)2]−, because a concentration of ∼5 mM is sufficient for accurate 17O NMR measurements, variable-temperature and -pressure 17O NMR relaxation rates and chemical shift measurements were performed. Hence, a simultaneous fit of 17 O and 1H NMR relaxation rates and chemical shifts was performed to assess the water exchange kinetic parameters. Again, the following parameters were fixed: aLnH = 3.6 × 10−10 −10 m, and D298 m2s−1. The diffusion constant was LnH = 25 × 10 assumed to obey the Arrhenius law. τR,O was calculated from the 17O NMR longitudinal relaxation rates as the quadrupolar contribution to the longitudinal relaxation is dominant. τR,H was thought to be equal to τR,O. When Ln = Dy3+ or Tm3+, because τmΔωm2 ≫ 1/T2,m, the simplified Swift and Connick equation (eq 10) was used; hence, no further information about rotational correlation time was needed. Moreover, the chemical shift at zero pressure (Δω0m)T in eq 19 was fixed to the value obtained from the fit of 1H NMR variable-temperature transverse relaxation rates. The experimental data and E
DOI: 10.1021/acs.inorgchem.6b00976 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry calculated curves are shown in Figures 4−7, and the corresponding fitted parameters are listed in Table 2.
Figure 5. 1H NMR pressure dependence of ln(1/T2r) of [Ln(AAZTAPh−NO2)(H2O)2]− (Ln = Gd3+, Dy3+, or Tm3+) at 20 °C, pH 5.7, and B0 = 9.4 T. Dashed lines correspond to the best fits of the experimental data.
Figure 6. 17O NMR temperature dependence of ln(1/T2r) (top), ln(1/T1r) (middle), and Δωm (bottom) of [Gd(AAZTAPh−NO2)(H2O)2]− at pH 5.7, B0 = 9.4 T, and 4.32 mM Gd3+. Dashed lines correspond to the best fits of the experimental data.
As shown in Figure 4, AAZTAPh−NO2 complexes with the three lanthanides studied appear to be in different NMR exchange regimes in the temperature range studied: fast exchange is found for Gd3+, intermediate for Dy3+, and slow for Tm3+ complexes. The results show a 2 order of magnitude decrease in the exchange rate constant on going from Gd3+ to 298 Tm3+ complexes. A similar strong decrease in kex was 21 previously reported by Graeppi et al. for PDTA complexes. The very slow exchange rates observed for [Tm(AAZTAPh− NO2)(H2O)]− reduce the inner-sphere contribution to proton relaxivity by inefficient transfer of the bound water relaxation to the bulk.22 The outer-sphere relaxivity of this complex at 20 °C is ∼30% of the total proton relaxivity. In the case of [Dy(AAZTAPh−NO2)(H2O)2]−, outer-sphere contributions are only at most 8% (see the Supporting Information). − Furthermore, k298 ex on [Gd(AAZTAPh−NO2)(H2O)2] is 2.8 times slower than on [Gd(AAZTA)(H2O)2]− (k298 ∼ 11 × ex 106).1 The remarkable decrease in kex may find its explanation in structural differences between the AAZTAPh−NO2 and AAZTA complexes or in the occurrence of a different population of diastereoisomers.23 The activation volumes ΔV⧧ measured for the q = 2 chelates [Gd(AAZTAPh−NO2)(H2O)2]− and [Dy(AAZTAPh−NO2)(H2O)2]− are slightly negative for the first one (−1.4 cm3 mol−1) and slightly positive (1.9 cm3 mol−1) for the second (Table 2). Interchange type mechanisms for the water exchange reactions can therefore be assigned to both of them, an associatively activated Ia mechanism to the exchange on the Gd3+ chelate and a dissociatively activated Id mechanism to the
Figure 7. 17O NMR pressure dependence of ln(1/T2r) of [Gd(AAZTAPh−NO2)(H2O)2]− at 20 °C, pH 5.7, B0 = 9.4 T, and 4.32 mM Gd3+. Dashed lines correspond to the best fits of the experimental data.
exchange on the Dy3+ chelate. The change in activation mode is probably due to the decrease in the ionic radius from 1.107 Å for nine-coordinated Gd3+ to 1.083 Å for nine-coordinated Dy3+.24 The activation volume measured for the q = 1 chelate [Tm(AAZTAPh−NO2)(H2O)]− is clearly negative (−5.6 cm3 mol−1), which is indicative of an associative mode of activation (Ia or A). The decrease in the coordination number from nine to eight from Dy3+ to Tm3+ is responsible for the change in mechanism. In the eight-coordinated chelate, there is enough space for an incoming water molecule and therefore for an increase in the coordination number to nine in the transition F
DOI: 10.1021/acs.inorgchem.6b00976 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry Table 2. Water Exchange Propertiesa of [Ln(AAZTAPh− NO2)(H2O)q]− [Gd(AAZTAPh− NO2)(H2O)2]−b q k298 ex (×106 s−1) 310 kex (×106 s−1) ΔH⧧ (kJ mol−1) ΔS⧧ (J mol−1) ΔV⧧ (cm3 mol−1)
[Dy(AAZTAPh− NO2)(H2O)2]−
nuclear magnetic relaxation. 1H NMR has been found to represent a valuable and reliable method for assessing the exchange rate constant of metal-coordinated water molecules when 17O NMR is not applicable because of the low chelate concentration. A change in the number of inner-sphere water molecules from two to one was observed upon replacement of Dy3+ by the smaller Tm3+. Moreover, water exchange rate constants were found to decrease more than 2 orders of magnitude from [Gd(AAZTAPh−NO2)(H2O)2]− to [Tm(AAZTAPh−NO2)H2O]−. Water exchange reactions on bisaqua Gd3+ and Dy3+ complexes follow an associative and dissociative interchange mechanism, respectively, as evidenced by the activation volumes of −1.4 and 1.9 cm3 mol−1, respectively. The water exchange on mono-aqua complex [Tm(AAZTAPh−NO2)H2O]− follows an associative activation mode (Ia or A mechanism) with a ΔV⧧ of −5.6 cm3 mol−1. Moreover, [Dy(AAZTAPh−NO2)(H2O)2]− is characterized by a nearly optimal rate of water exchange to maximize its transverse relaxivity. These properties make this complex a potentially very effective negative contrast agent for highmagnetic field imaging applications. AAZTAPh−NO2 complexes can therefore be used as positive, T1 contrast agents using Gd3+ as the paramagnetic ion and as negative, T2 contrast agents using Dy3+ as the lanthanide ion.
[Tm(AAZTAPh− NO2)(H2O)2]−
2 3.9 ± 0.1
2 0.46 ± 0.01
1 0.014 ± 0.0007
7.8 ± 0.3
0.84 ± 0.02
0.029 ± 0.001
42 ± 1
36 ± 0.9
46 ± 1
+21 ± 2 −1.4 ± 0.8 (Ia)
−15.5 ± 1 +1.9 ± 0.1 (Id)
−9.8 ± 2 −5.6 ± 0.2 (Ia/A)
a
For the full list of parameters, see the Supporting Information. bFrom a simultaneous fit of 17O and 1H NMR data.
state or an intermediate in the case of a limiting associative mechanism A. [Dy(AAZTAPh−NO2)(H2O)2]− Complexes as T2 Exchange Agents. Dysprosium complexes have been proposed as T2 relaxation agents for MRI mainly because of the high magnetic moment of Dy3+.25−28 To be considered as a good T2 agent, the compounds should have one or more inner-sphere water molecules and a water exchange rate falling within a narrow range of values:26 the maximal transverse relaxivity, r2, is reached for τm = Δωm−1 or kex = Δωm, and its value is 1.8 × 10−5qΔωm/2. If we assume that the paramagnetic chemical shift Δωm is approximately the same (1 × 106 rad s−1) for the three bis-aqua dysprosium complexes with L = DO3A, DTTA−Me, and AAZTAPh−NO2, we can calculate r2 at 9.4 T and 37 °C (Figure 8). The transverse relaxivity of [Dy(AAZTAPh−
■
EXPERIMENTAL PROCEDURES
Materials. LnCl3·xH2O species [Ln = Gd3+, Dy3+, or Tm3+ (99.9%)] were obtained from Aldrich and were used without further purification. The H4DTTA−Me was provided by EPFL through the synthetic platform; the synthesis of the ligand has been described previously,29 and H3DO3A was provided by CheMatech Co. (Dijon, France). Other reagents were obtained from Sigma-Aldrich Chemical Co. and commercial sources and used without further purification. Synthesis of L1. TFA (5 mL) was added dropwise to a solution of AAZTAPh−NO2(tBu)4, synthesized as described in ref 2 (1.0 g, 1.3 mmol) in CH2Cl2 (5 mL). The mixture was stirred at room temperature overnight and then evaporated in vacuo. The solid residue was washed twice with Et2O (10 mL), isolated by centrifugation, and dried in vacuo, yielding L1 as its trifluoroacetate salt: 1H NMR (500 MHz, D2O) δ 8.20 (d, 2H, J = 9.3 Hz, CH), 7.56 (d, 2H, J = 9.3 Hz, CH), 4.19 (s, 2H, CH2O), 3.59 (s, 4H, CH2C O), 3.58 (s, 4H, CH2CO), 3.55−3.43 (m, 8H, NCH2); 13C NMR (125 MHz, D2O) δ 179.9, 179.5 (COOH), 154.1 (NCO), 144.5, 142.7 (CAr), 125.2, 118.8 (CHAr), 63.4 (CH2O), 61.3 (CH2CO), 59.9 (CqCy), 57.7, 52.9 (CH2Cy); ESI+ MS (m/z) 542.2 (M + H+), calcd for C21H28N5O12 542.5. Preparation of [Ln(L)(H2O)q]x Complexes. Ln3+ solutions (1 M, pH
