Thermodynamic study of lanthanide(iii) complexes with bifunctional monophosphinic acid analogues of H4dota and comparative kinetic study of yttrium(iii) complexes

View Article Online / Journal Homepage / Table of Contents for this issue

PAPER

www.rsc.org/dalton | Dalton Transactions

Thermodynamic study of lanthanide(III) complexes with bifunctional

monophosphinic acid analogues of H₄dota and comparative kinetic study of

yttrium(^III) complexes†

Michaela Fo¨rsterova´,^aIvona Svobodova´,^b,cPrˇemysl Lubal,*^bPetr Ta´borsky´,^bJan Kotek,^aPetr Hermann*^aand

Ivan Lukesˇ^a

Received 15th September 2006, Accepted 9th November 2006

First published as an Advance Article on the web 29th November 2006

DOI: 10.1039/b613404a

New bifunctional H₄dota-like ligands with three acetic acid and one phosphinic acid pendant arms and

propionate (H₅do3ap^PrA) or 4-aminobenzyl (H₄do3ap^ABn) reactive groups bound to the phosphorus

atom were investigated. Potentiometric studies showed that the ligands have a similar basicity to the

parent H₄dota and the stability constants of their complexes with sodium(I) and selected lanthanide(III)

ions are also similar. Formation and acid-assisted decomplexation kinetics of yttrium(III) complexes

with a series of H₄dota-like ligands (H₄dota and its phosphinic/phosphonic acid analogues) were

studied and the reactions are sensitive to a slight modiﬁcation of the ligand structure. The

(2-carboxyethyl)phosphinic acid derivative H₅do3ap^PrAand the phosphonic acid ligand H₅do3ap form

complexes faster than H₄dota. The most kinetically inert complex is that with H₄do3ap^ABn. Rates of

complexation and decomplexation can depend on the ability to transfer proton(s) outside/inside the

complex cavity and, therefore, on the hydrophobicity of the ligands. The results demonstrate that the

new bifunctional ligands are suitable for labelling biomolecules with yttrium(III) radioisotopes for

utilization in nuclear medicine.

Introduction

Macrocyclic ligands efﬁciently wrap up metal ions and, therefore,

form very stable complexes. The ligands are often derivatives

of cyclen (1,4,7,10-tetraazacyclododecane) and cyclam (1,4,8,11-

tetraazacyclotetradecane). The amines are frequently modiﬁed at

the nitrogen atoms to obtain ligands exhibiting different com-

plexing properties. The most studied compounds of this type are

H₄dota (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid,

Chart 1) and H₄teta (1,4,8,11-tetraazacyclotetradecane-1,4,8,11-

tetraacetic acid) and a number of their derivatives.¹The useful

properties of metal complexes of these ligands have been explored

in several areas, e.g. as contrast agents (CA) in magnetic resonance

imaging (MRI)^2–5or for labelling of biomolecules with metal

radioisotopes for both diagnostic and therapeutic purposes.⁶The

requirements for both utilizations are the same—a harmful metal

ion cannot be deposited anywhere in the body and the complex

must be eliminated from the body unchanged.

To deliver such a complex to a speciﬁc site in an organism,

a metal ion is co-ordinated by, usually, a bifunctional ligand

Chart 1 Structures of ligands mentioned in the text.

^aDepartment of Inorganic Chemistry, Universita Karlova (Charles Uni-

versity), Hlavova 2030, 128 40 Prague 2, Czech Republic. E-mail:

having another reactive group for conjugation of the complex

to a biomolecule. The biomolecule (e.g. monoclonal antibody or

its fragment, oligopeptide or oligosaccharide) ensures the desired

biodistribution of such a conjugate in the organism. As mentioned

above, the metal complex may be useful in medicine (as a CA in

MRI or in nuclear medicine) if such a complex is highly stable

under physiological conditions. Now it is commonly recognized

that kinetic inertness to dissociation of such complexes is more

important than thermodynamic stability and is usually decisive

petrh@natur.cuni.cz; Fax: +420-22195-1253; Tel: +420-22195-1263

^bDepartment of Analytical Chemistry, Masaryk University, Kotla´rˇska´ 2, 611

37 Brno, Czech Republic. E-mail: lubal@chemi.muni.cz; Fax: +420-54949-

2494; Tel: +420-54949-5637

^cInstitute of Analytical Chemistry, Academy of Sciences of the Czech

Republic, Veverˇ´ı 97, 611 42 Brno, Czech Republic

† Electronic supplementary information (ESI) available: Figures of depen-

1

dence of H and ³¹P NMR chemical shift on pH; examples of formation

kinetics data; correlation of k_Hfor dissociation of the complexes with the

ﬁrst protonation constant of the ligands. See DOI: 10.1039/b613404a

This journal is

The Royal Society of Chemistry 2007

Dalton Trans., 2007, 535–549 | 535

©

View Article Online

for stability in vivo. The use of metal radioisotopes brought to

light another requirement—a very fast complexation of short-

lived radioisotopes making it possible to use the isotopes in kit

form in hospitals. The macrocyclic ligands (e.g. H₄dota and H₄teta

derivatives) are mostly more suitable ligands than the acyclic ones

from the point of thermodynamic stability and kinetic inertness of

their complexes. However, they often exhibit a very slow complex

formation in contrast to fast complexation of the acyclic ligands

(e.g. H₅dtpa derivatives; H₅dtpa = diethylenetriaminepentaacetic

acid). Therefore, there is a real need for designing new ligands the

complexes of which would be formed very quickly (advantage of

the acyclic ligands) and which would exhibit high thermodynamic

and kinetic stabilities (typical of macrocyclic ligands). There are

several useful radioisotopes in the lanthanide series (e.g. ¹⁷⁷Lu,

¹⁶⁶Ho, ¹⁵³Tb, ¹⁵³Sm) as well as the most interesting therapeutic

isotope ⁹⁰Y exhibiting a similar chemistry, and hence research

has been focused mainly on ligands able to strongly bind lan-

thanide(III) ions. Among the macrocyclic ligands, derivatives of

H₄dota are the most suitable for complexation of lanthanide(III)

ions.

As mentioned above, complexation of lanthanide(III) ions with

H₄dota derivatives is rather slow⁷and, therefore, it can limit

the utilization of such complexes mostly in the case of labelling

antibodies or their fragments.⁸The property is not critical for

labelling small peptides (e.g. somatostatin analogs) where a short

heating of aqueous solutions to high temperature can be used.⁹

As the most interesting therapeutic metal radioisotope is ⁹⁰Y,

effort was put into the chemistry of trivalent yttrium to ﬁnd

macrocyclic ligands capable of faster complexation. On the basis

of theoretical calculations, Jang et al. suggested that lengthening

of one pendant acetate arm to a propionate arm would lead to

faster complexation kinetics.¹⁰However, the proposal was not fully

conﬁrmed after synthesis of the ligand¹¹as the complexation of Y³⁺

was only twice as fast as with H₄dota under identical experimental

conditions. Another approach was devised by Brechbiel’s group.

They modiﬁed 1,4,7-triazacyclononane with two acetate and

one 2-[bis(carboxymethyl)amino]ethyl pendant arms.¹²The ligand

binds Y³⁺very quickly and the complex formed preserves a high

kinetic inertness due to the macrocyclic framework.

Cyclen has been modiﬁed with phosphonic or phosphinic acid

arms^13–16and such derivatives have been used for the complexation

of lanthanide radioisotopes.¹⁷It was found that the phosphinic

acid derivatives formed complexes relatively quickly. It is also

known that the phosphonic/phosphinic acid derivatives and

their complexes differ in some properties (size and shape of

coordinating group, overall hydrophilicity or acidity/basicity of

the ligands) when compared with their acetate analogs.^16,18–20

In addition, bifunctional ligands can be designed with a reac-

tive functional group on the phosphorus atom substituents.^19,21

Phosphonic/phosphinic acids are deprotonated at lower pH than

carboxylic acids and, therefore, they are able to bind metal ions

in more acidic solutions than carboxylate ligands.²²Therefore, we

assumed that substitution of one acetic acid arm in H₄dota for

a phosphonic/phosphinic acid arm may lead to a faster complex

formation while preserving thermodynamic stability and kinetic

inertness of the complexes formed. Thus, we decided to study the

monophosphonic/phosphinic acid derivatives from the thermo-

dynamic and kinetic points of view. We have recently described

the thermodynamic behaviour of lanthanide(III) complexes with a

monophosphonate analog of H₄dota and kinetic properties of its

Ce³⁺and Gd³⁺complexes.²³

In this paper, we present thermodynamic data on two bifunc-

tional monophosphinic acid derivatives of H₄dota which may

be used for conjugation to biologically active molecules through

distant 4-aminobenzyl or carboxylic groups. As the ligands may

be used for complexation of radioisotopes (e.g. ⁹⁰Y), we were

interested in the kinetic behaviour of their yttrium(III) complexes.

However, there are no reasonable kinetic data for yttrium(III)

complexes of H₄dota-like ligands for comparison. Therefore, we

performed a comparative study of formation/decomplexation

kinetics for a series of such ligands (Chart 1).

Experimental

General

The chemicals for synthesis were purchased from Fluka, Aldrich,

Merck or Lachema and were used as received. Crystalline H₃PO₂

was obtained by evaporation of a commercial 50% aqueous

solution (vacuum, below 40 ^◦C) and vacuum drying over P₂O₅

(4 ^◦C). The following ligands (Chart 1) were synthesized by known

19b

methods: H₄dota,²⁴H₅do3ap,¹⁸H₄do3ap^ABn

,

and H₄dotp^Ph.²⁵

H₃do3a·H₂SO₄(H₃do3a = 1,4,7,10-tetraazacyclododecane-1,4,7-

triacetic acid; Chart 1) was a gift from Bracco SpA (Italy).

Paraformaldehyde was ﬁltered from aged aqueous formaldehyde

and dried over P₂O₅. Elemental analyses were made at the Institute

of Macromolecular Chemistry of the Academy of Sciences of

the Czech Republic. Thin-layer chromatography (TLC) was run

on silica gel aluminium-backed sheets (Silufol (Kavalier, Czech

Republic) or Merck) with propan-2-ol–conc. aq. NH₃–water =

7 : 3 : 3 v/v as an eluent. Spraying with ninhydrin solution or

Dragendorff reagent, iodine vapour or UV irradiation were used

for detection. NMR spectra were recorded on a Varian Unity Plus

1

at 400 MHz for ¹H, 169 MHz for ³¹P and 100 MHz for ¹³C{ H} with

t-BuOH (D₂O) or TMS (CDCl₃) as internal references for ¹H/¹³C

and 85% H₃PO₄as external reference for ³¹P. Full assignment of

¹H/¹³C signals of H₅do3ap^PrAwas done with NOESY, HMQC

and HSQC experiments. ESI/MS spectra were run on a Bruker

ESQUIRE 3000 with ion-trap detection in positive or negative

modes. Water used in the physico-chemical measurements was

puriﬁed using the Milli-Q System (Millipore).

Crystal structure determination

Single crystals of [Na(H₅do3ap^PrA)₂]Cl·10.5H₂O (1) and

[Na(H₄do3ap^PrA)(H₅do3ap^PrA)]·11.5H₂O (2) were obtained by

spontaneous crystallization on standing (one week) of an oily

residue (30 mg) of the last fractions from the Amberlite CG50

chromatography which still contains only the pure ligand (see

below).

Selected crystals were mounted on glass ﬁbres in random

orientation and cooled to 150(1) K. The diffraction data were

collected at 150(1) K (Cryostream Cooler, Oxford Cryosystem)

using a Nonius Kappa CCD diffractometer and Mo-K_aradi-

˚

ation (k = 0.71073 A) and analyzed using the HKL DENZO

program package.²⁶The structures were solved by direct methods

(SIR92),²⁷and reﬁned by full-matrix least-squares techniques

536 | Dalton Trans., 2007, 535–549

This journal is

The Royal Society of Chemistry 2007

©

View Article Online

(SHELXL97).²⁸The scattering factors used for neutral atoms were

included in the SHELXL97 program.

thermal factors. In addition, a high number of residual peaks,

and their close proximity, revealed disorder of some other solvate

water molecules; therefore, the residual peaks were smoothed using

the SQUEEZE command of PLATON.²⁹The total number of

subtracted electrons was 154 per unit cell, i.e. approximately 2

water molecules per formula unit. The propionate side-arm of one

ligand molecule was found to be disordered over two positions

and was reﬁned using the DFIX instruction for C–C bonds. All

non-hydrogen atoms were reﬁned anisotropically. All hydrogen

atoms belonging to carbon and nitrogen atoms and oxygen atoms

of carboxylate pendants were located in the electron difference

map; however, they were ﬁxed in theoretical or original positions

using a riding model (U_eq(H) = 1.2U_eq(C), 1.3U_eq(N or O)). Some

hydrogen atoms belonging to co-crystallized water molecules were

also found, and were treated using the AFIX 3 instruction and

U_eq(H) = 1.3U_eq(O).

In both structures, the sodium ion is octahedrally coordinated

by three oxygen atoms of the carboxylate arms of two independent

ligand molecules, giving the same structural motif in both cases.

In the structure of [Na(H₅do3ap^PrA)₂]Cl·10.5H₂O (1), the highest

maximum on the electron density map was attributed to the

chloride anion. In addition, seven highest residual peaks were

assigned to co-crystallized water molecules. However, a high

number, and close proximity, of other electron maxima point

to huge disorder of the solvate water molecules. Therefore, the

residual peaks were smoothed using the SQUEEZE command

of PLATON.²⁹The total number of subtracted electrons was

74 per unit cell, i.e. approximately 3.5 water molecules per

formula unit. All non-hydrogen atoms were reﬁned anisotropically.

All hydrogen atoms attached to carbon and nitrogen atoms

and oxygen atoms of carboxylate pendants were located in the

electron difference map; however, they were ﬁxed in theoretical or

original positions with thermal parameters U_eq(H) = 1.2U_eq(C) or

1.3U_eq(N or O) as their free reﬁnement dramatically increases the

number of parameters and in some cases leads to unrealistic bond

lengths. Some hydrogen atoms belonging to co-crystallized water

molecules were also found, and were treated using the AFIX 3

instruction and U_eq(H) = 1.3U_eq(O). The quality of the diffraction

data of [Na(H₅do3ap^PrA)(H₄do3ap^PrA)]·11.5H₂O (2) was poor. In

the crystal structure, the most intense residual electronic maxima

were assigned to co-crystallized water molecules, and some of the

water molecules were reﬁned with half occupancy to get reliable

Table 1 gives pertinent crystallographic data.

CCDC reference numbers 621031 and 621032.

For crystallographic data in CIF or other electronic format see

DOI: 10.1039/b613404a

Chemicals and stock solutions for potentiometric titrations

The stock solution of hydrochloric acid (∼0.03 mol dm⁻³) was pre-

pared from a 35% aqueous solution (“puriss”, Fluka). Commercial

Me₄NCl (99%, Fluka) was recrystallized from boiling i-PrOH and

the solid salt was dried under vacuum over P₂O₅until constant

weight (the dried salt is extremely hygroscopic). Carbonate-free

Table 1 Experimental data for crystal structure determination of [Na(H₅do3ap^PrA)₂]Cl·10.5H₂O (1) and [Na(H₄do3ap^PrA)(H₅do3ap^PrA)]·11.5H₂O (2)

Parameter

[Na(H₅do3ap^PrA)₂]Cl·10.5H₂O (1)

[Na(H₅do3ap^PrA)(H₄do3ap^PrA)]·11.5H₂O (2)

Empirical formula

Formula weight

Crystal dimensions/mm

Shape

Colour

Crystal system

C₃₆H₈₇ClN₈NaO_30.5P₂

1240.53

C₃₆H₈₈N₈NaO_31.5P₂

1222.08

0.05 × 0.08 × 0.60

0.20 × 0.28 × 0.38

Rod

Prism

Colourless

Monoclinic

C2/c (No. 15)

29.7688(3)

16.5518(2)

25.3298(3)

90

106.5023(8)

90

11966.6(2)

8

Triclinic

¯

Space group

P1 (No. 2)

˚

a/A

11.9291(4)

16.0112(4)

17.2340(6)

96.163(2)

91.914(1)

108.627(2)

3093.2(2)

2

˚

b/A

˚

c/A

a/^◦

b/^◦

c /^◦

3

˚

V/A

Z

D_c/g cm⁻³

1.332

0.209

1.357

0.173

l/mm⁻¹

F(000)

1322

5224

h range of data collection/^◦

Index ranges

1.81–23.03

−13 ≤ h ≤ 13

−17 ≤ k ≤ 17

−18 ≤ l ≤ 18

8608, 0, 678

1.075

0.0728, 0.0535

0.1577, 0.1467

0.000

1.54–27.10

−37 ≤ h ≤ 38

−21 ≤ k ≤ 21

−32 ≤ l ≤ 32

13163, 2, 727

1.041

0.1047, 0.0908

0.2806, 0.2672

0.001

Data, restraints, parameters

Goodness-of-ﬁt on F²

R, R^ꢀindices [I ≥ 2r(I)]^a

wR, wR^ꢀ[I ≥ 2r(I)]^b

Maximum shift/e.s.d.

−3

˚

Largest diffr. peak and hole/e A

1.076, −0.490

1.504, −0.700

ꢀ

^aR = |F_o− F_c|/ |F_c|; (SHELXL97, ref. 28). ^bwR = [ w(F_o− F_c²)²/ w(F_o²)²]^1/2, w = 1/[r²(F_o²) + (AP)²+ BP]; where P = (F_o+ 2F_c²)/3

2

(SHELXL97, ref. 28).

This journal is

The Royal Society of Chemistry 2007

Dalton Trans., 2007, 535–549 | 537

©

View Article Online

Me₄NOH solution (∼0.2 mol dm⁻³) was prepared from the

Me₄NCl by ion exchange on Dowex 1 (OH⁻-form) under argon

and using CO₂-free water. The hydroxide solution was standard-

ized against potassium hydrogen phthalate and the HCl solution

against a ∼0.2 mol dm⁻³Me₄NOH solution. Stock solutions of

the individual metal cations were prepared by dissolving LnCl₃

hydrates (99.9%; Strem) or dry NaCl (Fluka) in water. The lan-

thanide(III) contents in the solutions were determined by titration

with standard Na₂H₂edta solution. Analytical concentrations of

stock solution of the ligands were determined together with

reﬁnement of protonation constants using the OPIUM software

package³⁰(see below).

dissociation constants as pK(HL) = logb₁and pK(H_nL) = logb_n−

logb_n−1). The (concentration) stability constants are deﬁned by

b_hlm= [H_hL_lM_m]/([H]^h× [L]^l× [M]^m). The water ion product pK_w

(13.81) and stability constants of the Ln³⁺–OH⁻systems included

in the calculation were taken from ref. 31 and 32.

Kinetic measurements

All measurements were carried out on a HP-8453A diode

array spectrophotometer (Hewlett-Packard, USA) at 25.0

0.1 ^◦C. To visualize the formation of Y(Lig) species (Lig =

general macrocyclic ligand) and to eliminate possible hydrolysis

of the Y³⁺ion, a chromogenic reagent Arsenazo III, 2,7-bis[(2-

arsonophenyl)azo]-1,8-dihydroxynaphthalene-3,6-disulfonic acid

(H₈AZ, Fluka), was used. The Arsenazo III is a sensitive and

generally used chromogenic reagent and it has been used for

detection of free yttrium(III)³³or gadolinium(III)³⁴in the course

of formation of macrocyclic ligand complexes. The formation

of yttrium(III) complexes with H₈AZ slows down the formation

of the macrocyclic ligand complexes and, therefore, the rates of

formation can be measured in pH regions where conventional

spectroscopic measurements are sometimes impossible because

of too high rates. In addition, sensitivity of the measurement is

very high and formation of the complexes protects yttrium(III)

against hydrolysis. The formation kinetics of the Y(Lig) species

was followed under pseudo-ﬁrst order conditions with c_Y= 7.8 ×

10⁻⁷mol dm⁻³, c_Lig= 1.7 × 10⁻⁵mol dm⁻³, c_AZ= (5.0–15.0) ×

10⁻⁵mol dm⁻³and I = 0.1 mol dm⁻³(KCl) in the −log[H⁺]

range 5.5–7.1 (0.01 mol dm⁻³solution of 2-morpholinoethane-

1-sulfonic acid (MES), Fluka). Spectral changes were followed at

653 nm. The effect of Arsenazo III bound to yttrium(III) on the

reactivity of the system was eliminated by extrapolation of the

Potentiometric titrations

Titrations were carried out in a vessel thermostatted at 25.0

◦

0.1 C, at ionic strength I = 0.1 mol dm⁻³(Me₄NCl) and

in the presence of excess HCl in the −log[H⁺] range 1.7–11.9

(protonation constants and system with Na⁺) using a PHM 240

pH-meter, a 2-ml ABU 900 automatic piston burette and a GK

2401B combined electrode (all Radiometer). The initial volume

was 5 cm³and the ligand concentration was ∼0.004 mol dm⁻³.

For Na⁺–ligand systems, a Na : L = 10 : 1 molar ratio was

used. Four parallel titrations were carried out; each titration

consists of about 40 points. An inert atmosphere was provided

by the constant passage of argon saturated with the vapour of the

solvent used in the measurements. Titrations with lanthanide(III)

ions were performed at metal-to-ligand molar ratios of 1 : 1. As

the complexation was too slow for conventional titration, the

“out-of-cell” method was used. Each titration consists of 23–

25 points in the −log[H⁺] range from 1.8 to 6.0 (at least two

parallel titrations). Equilibrium was reached after four weeks.

The constants determined by this technique are less precise

and are loaded with high uncertainty due to the less accurate

measurements and a small number of experimental points. The

“kinetic titrations” were run as a normal titration in the Y/La :

H₅do3ap^PrAratio 1 : 1 in the presence of excess HCl in the −log[H⁺]

range 1.7–11.9. Waiting time was ca 1 min per titration point

(for about 50 points). No precipitation of hydroxide was observed

during these titration experiments.

dependence (observed pseudo-ﬁrst order constants k^ꢀ_f,obson c_AZ

)

to zero concentration of the dye (c_AZ= 0; for explanation see

Results and discussion section). Comparison of this method with

a standard one (i.e. measurement of absorbance of a coloured

complex as in the case of Eu³⁺) gave similar results with a

satisfactory precision.³⁵To determine the charges of the species in

the rate-determining reaction, measurements at I → 0.0 mol dm⁻³

(i.e. only with the contributions of dilute reactants and the MES

buffer at −log[H⁺] 6.5) were performed.

The constants (with standard deviations) were calculated with

program OPIUM.³⁰The program minimizes the criterion of the

generalised least-squares method using the calibration function

Dissociation kinetics of the Y(Lig) complexes were measured

in the range of proton concentration 0.05–3.00 mol dm⁻³at an

ionic strength I = 3.0 mol dm⁻³(Na,H)ClO₄with c_Y∼ 2 ×

10⁻⁴mol dm⁻³. The decomplexation was followed via the increase

in absorbance (660 nm) of the Y³⁺–H₈AZ complex formed as

was proposed in the literature.^33,36In order to verify the reliability

of this method, the dissociation of [Y(dotp^Ph)]⁻with the phenyl

chromophores was followed by both methods (direct UV/Vis

spectrophotometry and UV/Vis spectrophotometry with the aux-

iliary H₈AZ ligand). Both methods gave results with a satisfactory

agreement. The method was further validated by measurements

of coloured europium(III) complexes with monophosphorous acid

derivatives of H₄dota.³⁵Agreement between the results obtained

by all the methods was satisfactory. The measured values of

absorbance were corrected for the background signal. All kinetic

data were processed by non-linear regression Excel and HP

software; both gave identical results.

E = E₀+ S × log[H⁺] + j₁× [H⁺] + j₂× K_w/[H⁺]

where the additive term E₀contains the standard potentials of

the electrodes used and contributions of inert ions to the liquid-

junction potential, S corresponds to the Nernstian slope, the value

of which should be close to the theoretical value and j₁× [H⁺]

and j₂× K_w/[H⁺] = j₂× [OH⁻] terms are the contributions

of the H⁺and OH⁻ions to the liquid-junction potential. It is

clear that j₁and j₂cause deviation from a linear dependence

of E on −log[H⁺] only in strongly acidic and strongly alkaline

solutions. The calibration parameters were determined by titration

of standard HCl with standard Me₄NOH before each ligand

or ligand–metal titration to give a pair calibration/titration,

which was used for calculations of the constants. The overall

protonation constants b_nare concentration constants deﬁned by

b_n= [H_nL]/([H]ⁿ× [L]) (they were transformed into stepwise

538 | Dalton Trans., 2007, 535–549

This journal is

The Royal Society of Chemistry 2007

©

View Article Online

NMR titration

was heated at 110 ^◦C under argon for 24 h. After cooling,

volatiles were removed under vacuum. The residue was dissolved in

200 ml of water, partially decolorized with charcoal and the water

evaporated. Non-amine impurities were removed on a Dowex

50 column (3.5 × 20 cm, H⁺-form) by elution with water and

the product was eluted with aqueous (10%) pyridine. Fractions

containing the product were evaporated and the residue was

puriﬁed on an Amberlite CG50 column (6.5 × 20 cm, H⁺-form)

with water elution. Fractions containing the pure product (TLC

and ¹³C and ³¹P NMR check) were combined and water was

evaporated. The glassy residue was dissolved in water (100 ml),

colloids (from the resin) were removed with charcoal and water

was removed under vacuum. The residue was dissolved in water

(5 ml) and the viscous solution was slowly dropped into stirred

anhydrous EtOH (400 ml). The suspension was left overnight to

mature and the solid was then ﬁltered off, washed with anhydrous

EtOH and Et₂O and quickly dried on the ﬁlter. The occluded

EtOH was removed in a vacuum desiccator (P₂O₅, 0.13 kPa, 24 h,

room temperature) and the product was left to equilibrate in air

for 3 d. The yield of white H₅do3ap^PrA·3H₂O was 2.57 g (48%).

Found: C, 39.16; H, 7.32; N, 10.10. C₁₈H₃₃N₄O₁₀P·3H₂O (M =

550.50) requires C, 39.27; H, 7.14; N, 10.18%. NMR (D₂O, 90 ^◦C;

1

The H and ³¹P{ H} NMR titrations over the whole pH region

were performed at 90.0 ^◦C in H₂O at a ligand concentration of

0.05 mol dm⁻³, with presaturation of the water signal and using

a coaxial capillary with D₂O for the lock. The sample solution

pH values (0.3–13.5) were adjusted with NaOH or HCl aqueous

solutions and were measured with a pH-meter calibrated with

standard buffers.³⁷

Synthesis of HO₂CCH₂CH₂P(O)(H)(OH) (Scheme 1)

Crystalline H₃PO₂(10 g, 0.1515 mol) was dissolved in trimethyl

orthoformate (18.2 ml, 0.167 mol) and the mixture was stirred

for 2 h at room temperature to form H₂P(O)(OMe) in situ. In

another ﬂask, a mixture of methyl acrylate (13.7 ml, 0.167 mol)

and i-Pr₂NEt (2.6 ml, 0.015 mol) was cooled to 5 ^◦C. With

vigorous stirring, the solution of H₂P(O)(OMe) was dropped

(1 ml over 5 min) into the acrylate mixture and this was stirred

at room temperature for 3 d. Chloroform (50 ml) was added,

and the mixture was extracted with saturated aqueous NaHCO₃

solution (2 × 50 ml). The aqueous phase was re-extracted with

chloroform (3 × 20 ml), the organic phases were combined, dried

(anhydrous Na₂SO₄) and the volatiles were evaporated in vacuum.

The residual viscous oil was puriﬁed twice on a silica gel column

(6.5 × 25 cm) with petroleum ether (bp 40–60 ^◦C)–CH₂Cl₂–

MeOH 1 : 10 : 2 to give crude MeO₂CCH₂CH₂P(O)(H)(OMe)

2

for numbering see Scheme 1): d_H2.06 (2H, m, J_HP= 12, H20),

2.72 (2H, m, H21), 3.39 (8H, m, H3 + H6 + H8 + H11), 3.41 (2H,

2

d, J_HP= 4, H13), 3.44 (4H, m, H5 + H9), 3.60 (4H, m, H2 +

H12), 3.82 (2H, s, H16), 3.89 (4H, s, H14 + H18); d_C26.5 (1C, d,

2

¹J_CP= 93.8, C20), 27.5 (1C, d, J_CP= 3.0, C21), 50.2 (4C, C3 +

1

(d_P(CDCl₃): 34.0, J_PH550 Hz) contaminated with ∼5% of

C5 + C9 + C11), 50.3 (2C, C6 + C8), 51.1 (2C, C2 + C12), 52.5

(1C, d, ¹J_CP= 85.9, C13), 55.4 (1C, C16), 55.5 (2C, C14 + C18),

172.2 (1C, C17), 172.6 (2C, C15 + C19), 177.8 (1C, d, ³J_CP= 14.4,

(MeO₂CCH₂CH₂)₂P(O)(OMe) (d_P(CDCl₃): 55.5) and a small

amount of MeO₂CCH₂CH₂P(O)(OMe)₂(d_P(CDCl₃) 27.5) The

ester was dissolved in a mixture of water (50 ml), AcOH (50 ml)

and conc. aqueous HCl (50 ml) and the mixture was reﬂuxed

for 12 h. Volatiles were removed under vacuum, the residue was

dissolved in water (100 ml) and decolorized with charcoal. Water

was evaporated under vacuum to give the phosphinopropionic

acid as a very viscous oil (6.6 g, yield 28%, purity ∼93%). There

were no other compounds containing a P–H bond. NMR (D₂O):

d_H1.85 (2H, m, CH₂P), 2.45 (2H, m CH₂CO₂H), 6.91 (1H, td,

¹J_PH558, ³J_HH2, H–P); d_C26.71 (d, ¹J_CP93, CH₂P), 28.25 (d, ²J_CP

8, CH₂CO₂H), 178.80 (d, ³J_CP13, CO₂H); d_P38.3 (m, ¹J_PH559).

1

C22); d_P{ H} 37.0. ESI-MS: m/z 497.2 (M + H)⁺(calc. 497.5),

519.2 (M + Na)⁺, 541.2 (M − H + 2Na)⁺.

Results and discussion

Synthesis

The ligands (except H₅do3ap^PrA) were synthesized using published

methods or were available from our previous studies. To obtain

a ligand with a free carboxylic group in a side chain, we

chose the propionate substituent bound to the phosphorus atom

affording ligand H₅do3ap^PrA. Its carboxylic group can be used

for conjugation to biologically active molecules as it contains

two methylene groups excluding any possible coordination of

the carboxylate group to the central lanthanide(III) ions (as it

would lead to a seven-membered chelate ring). The Mannich

reaction of amine, formaldehyde and P–H precursor is the most

common reaction for formation of the >N–CH₂–P group. First,

it was necessary to get RO₂CCH₂CH₂P(O)(H)(OR) (R = alkyl,

Synthesis of 10-{[(2-carboxyethyl)hydroxyphosphoryl]

methyl}-1,4,7,10-tetraazacyclododecane-1,4,7-triacetic acid

(H₅do3ap^PrA) (Scheme 1)

The crude 3-phosphinopropanoic acid (6.5 g, about 0.043 mol) and

H₃do3a·H₂SO₄(4.76 g, 0.0107 mol) were dissolved in azeotropic

aqueous HCl. The ﬂask was ﬂushed with argon and heated at

90 ^◦C (in oil bath). Paraformaldehyde (3.21 g, 0.107 mol) was

added in small portions over 3.5 h under argon ﬂow. The mixture

Scheme 1 Synthesis of H₅do3ap^PrA

The Royal Society of Chemistry 2007

.

This journal is

Dalton Trans., 2007, 535–549 | 539

©

View Article Online

H) as a P–H precursor for the synthesis of H₅do3ap^PrA. To obtain

the dimethyl ester of the acid we used a published procedure.³⁸

However, we were unable to fully reproduce the results (Scheme 1).

Despite great efforts, the ﬁnal reaction mixture always contained

the desired ester and a lot of by-products. The unreacted methyl

ester of phosphinic acid readily hydrolyses and the resulting

H₃PO₂was easily removed by extraction. A critical impurity was

a ‘protected’ phosphinic acid (MeO)₂CH–P(O)(H)(OMe) (d_P30.8

in CDCl₃) which is ﬁnally back hydrolyzed to the highly reactive

H₃PO₂. The by-product (as well as most of the corresponding

dialkylphosphinate and alkylphosphonate esters) was removed by

chromatography on silica gel. The puriﬁed ester was hydrolyzed to

the acid with purity satisfactory for the next reaction. The possible

impurities do not contain any reactive P–H bonds and can be easily

removed on a Dowex 50 column after the Mannich reaction. The

above puriﬁcation had to be done in the ester forms since H₃PO₂

and HO₂CCH₂CH₂PO₂H₂could not be separated. The Mannich

reaction in aqueous HCl led to the desired product in a moderate

yield. Final puriﬁcation of the macrocyclic ligand was done on

carboxylic cation-exchange resin; the only way to get the product in

the solid state was by dropping the concentrated aqueous solution

of the ligand into stirred anhydrous EtOH. Any other attempts to

isolate the ligand led to oils or to highly hygroscopic solids. The

solid trihydrate of the ligand is stable in air.

Table 2 Coordination geometry of sodium ion found in the struc-

tures of [Na(H₅do3ap^PrA)₂]Cl·10.5H₂O (1) and [Na(H₅do3ap^PrA)-

(H₄do3ap^PrA)]·11.5H₂O (2)

1

2

˚

Distance/A

Na1–O31A

Na1–O31B

Na1–O41A

Na1–O41B

Na1–O51A

2.401(3)

2.353(3)

2.366(3)

2.414(3)

2.621(3)

2.611(3)

2.404(4)

2.391(3)

2.403(3)

2.461(3)

2.685(3)

2.631(3)

Na1–O51B

Angle/^◦

O31A–Na1–O31B

O31A–Na1–O41A

O31A–Na1–O41B

O31A–Na1–O51A

O31A–Na1–O51B

O31B–Na1–O41A

O31B–Na1–O41B

O31B–Na1–O51A

O31B–Na1–O51B

O41A–Na1–O41B

O41A–Na1–O51A

O41A–Na1–O51B

O41B–Na1–O51A

O41B–Na1–O51B

O51A–Na1–O51B

100.04(9)

76.44(9)

176.01(10)

92.83(9)

99.10(9)

172.18(10)

78.32(9)

97.53(9)

93.13(9)

105.63(9)

75.84(8)

94.32(9)

91.00(9)

77.42(9)

162.37(9)

101.87(11)

75.94(9)

172.25(11)

94.09(9)

96.56(10)

175.86(11)

74.97(9)

100.86(9)

92.59(9)

107.65(9)

75.94(8)

91.16(9)

93.46(9)

76.68(9)

160.76(9)

Crystal structures of sodium complexes of H₅do3ap^PrA

The oily residue after evaporation of the last pure (very small)

fraction from Amberlite chromatography was left standing in a

glass ﬂask for some days and several single crystals of two different

shapes appeared. Their solid-state structure was determined by

X-ray crystallography. To our surprise, they were shown to be

compounds containing sodium and the ligand in a 1 : 2 molar ratio.

However, sodium (AAS) or chloride (AgNO₃) was not detected in

the bulk material after precipitation (the fraction with crystals was

not mixed with the other ones). Obviously, sodium may have been

released from the borosilicate glass due to long contact with the

concentrated ligand solution.

The main structural features of the crystal structures

of [Na(H₅do3ap^PrA)₂]Cl·10.5H₂O (1) and [Na(H₅do3ap^PrA)-

(H₄do3ap^PrA)]·11.5H₂O (2) are identical. The sodium cation is

octahedrally coordinated by six carboxylate oxygen atoms of two

ligand molecules. Therefore, only the [Na(H₅do3ap^PrA)₂)]⁺cation

found in the structure of 1 is shown in Fig. 1. The geometries

of the sodium coordination spheres are listed in Table 2. In the

structure of 1, both ligands are coordinated in the zwitterionic

form. In each ligand, macrocycle nitrogen atom N1 bearing the

phosphinate pendant group and the “trans” nitrogen atom N7

bearing the acetate arm are protonated. Furthermore, two acetate

groups attached to non-protonated nitrogen atoms N4 and N10

are protonated on uncoordinated oxygen atoms O31 and O51, re-

spectively. The last proton is bound to the carboxylate group of the

propionate side-arm. In the structure of 2, the protonation scheme

of the ligand molecules is identical to the previous case, except for

a missing proton on one (disordered) propionate side-chain.

Fig. 1 Molecular structure of the [Na(H₅do3ap^PrA)₂)]⁺cation in 1 with

atom numbering scheme.

formation of stable complexes with alkali metal ions^39,40(see also

below). The constants are presented in Table 3 together with those

of similar ligands for comparison.

As for other tetraazamacrocyclic ligands, the last two protons

dissociate from the ring nitrogen atoms. Generally, the values

of the dissociation constants of H₄do3ap^ABnand H₅do3ap^PrAare

very similar to those of H₄dota. However, a lower basicity of a-

aminoalkylphosphinic acids than a-aminoalkylphosphonic acid

or a-amino acids is expected due to the electron-withdrawing

Protonation of H₄do3ap^ABnand H₅do3ap^PrA

Overall protonation constants of the two ligands were determined

at 25 ^◦C with Me₄NCl as the background salt due to the expected

540 | Dalton Trans., 2007, 535–549

This journal is

The Royal Society of Chemistry 2007

©

View Article Online

Table 3 Stepwise dissociation and overall protonation constants of H₄do3ap^ABn, H₅do3ap^PrAand similar ligands (I = 0.1 mol dm⁻³, 25 ^◦C)

ꢀ

a

Ligand

pK_A(HL)

pK_A(H₂L)

pK_A(H₃L)

pK_A(H₄L)

pK_A(H₅L)

pK_A(H₆L)

pK_A(H₇L)

pK_A

Ref.

H₅do3ap^PrAb

H₄do3ap^ABnb

12.68

12.68(5)

12.55

12.55(2)

13.83^c

11.9

14.65

13.7

—

9.44

22.12(7)

9.60

22.15(3)

10.35

9.72

12.40

12.2

12.5

12.3

6.83

7.27

8.24

5.04

27.16(8)

5.11

27.26(3)

6.54

4.60

9.28

9.25

8.90

1.97

2.75

3.71

4.34

31.50(8)

4.11

31.37(3)

4.34

4.13

8.09

8.01

3.10

34.60(8)

2.71

34.08(3)

3.09

2.36

6.12

6.17

1.93

36.53(8)

1.54

35.62(4)

1.63

31.50

31.37

This work

H₅do3ap

H₄dota

1.07

35.06

30.35

44.42

43.3

—

41.75

—

23

40

41

42

43

44

45

25

46

H₈dotp^d

5.22

5.17

5.14

1.77

12.8

7.75

5.95

H₄dotp^He

H₄dotp^Phf

H₄dotp^Etg

10.41

11.44

10.94

1.45

22.91

—

ꢀ

pK_A= pK_A(HL) + pK_A(H₂L) + pK_A(H₃L) + pK_A(H₄L). ^bThe determined corresponding overall protonation constants (logb_n) are in italics (25 ^◦C,

a

0.1 M Me₄NCl). ^cDetermined by ³¹P NMR spectroscopy (25 ^◦C, Me₄N(Cl,OH), no ionic strength control). ^dH₈dotp = (1,4,7,10-tetraazacyclododecane-

1,4,7,10-tetraylmethyl)-tetrakis(phosphonic acid); Chart 1. ^eH₄dotp^H= (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraylmethyl)-tetrakis(phosphinic

acid); Chart 1. ^fH₄dotp^Ph= (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraylmethyl)-tetrakis(phenylphosphinic acid); Chart 1. ^gH₄dotp^Et= (1,4,7,10-

tetraazacyclododecane-1,4,7,10-tetraylmethyl)-tetrakis(ethylphosphinic acid); Chart 1.

ability of the phosphinate group.^6,47Surprisingly, the values of

the last dissociation constant of H₄do3ap^ABnand H₅do3ap^PrA

are higher than that of H₄dota under similar conditions. As

anticipated, they are much lower than those of H₅do3ap or H₈dotp.

The increase in the dissociation constants of the phosphinate

ligands compared with that of H₄dota can be caused by differences

in the structure of the HL species having a different scheme of

intramolecular hydrogen bonds. The expected order of pK_A’s

is found for values corresponding to dissociation of a proton

from the H₂L species. These diprotonated species are the most

abundant at physiological pH (Fig. 2a). The next proton should

be mostly bound to the amine/carboxylate group in the side chain.

The other constants correspond to dissociation of the carboxylic

pendant arms similarly to H₄dota and H₅do3ap. The acidity of the

phosphinate group is generally very low¹⁶and its pK_Avalue is out

of the titration range.

To estimate the sites of protonation of different ligand species

in solution, we investigated the dependence of d_Hand d_Pon

solution pH. To get _◦reasonable narrow signals, the spectra had

to be recorded at 90 C. Unfortunately, ¹H NMR signals of ring

protons were overlapping at most pH values and can not be easily

assigned. Therefore, only chemical shifts of the methylene groups

of the pendant arms and the propionate side chain were followed.

The dependences of chemical shifts on pH are shown in Fig. S1.†

The protonation sites can be assigned by analogy with data pub-

lished for aminoalkylphosphonic acids⁴⁸and azamacrocycles with

phosphonicacidpendantarm(s).^13,49–51Ingeneral, thedependences

resemble those observed for H₅do3ap.²³

No changes of chemical shift occur above pH 11. The ﬁrst

changes of chemical shift are seen in the pH range 8–10.5. As

the last two dissociation constants determined by potentiometry

lay in the basic region (pK_A’s 12.68 and 9.44, respectively), two

protons should be attached in this pH range (Scheme 2). In this

case, the values of the pK_A’s determined by NMR spectroscopy

are obviously lower than those determined by potentiometry. Such

a drop in the values of the protonation constants can be explained

by Na⁺-assisted deprotonation (complexation of Na⁺) as NaOH

was used to adjust the pH of the NMR samples. In addition,

the pK_A’s are lowered also by the effect of the high temperature

used during the measurements (90 ^◦C vs. 25 ^◦C). All pendant

arm methylene groups, phosphorus atoms, and even PCH₂groups

of the side chain feel the protonation. Therefore, both protons

are almost statistically distributed over all four nitrogen atoms.

The sum of the chemical shift changes (Dd_H) for all the pendant

methylene groups is almost 2C_N(C_Nis a coefﬁcient corresponding

to the inﬂuence of a-amino group protonation on d_Hof the

Scheme 2 Protonation scheme of H₅do3ap^PrA

The Royal Society of Chemistry 2007

.

This journal is

Dalton Trans., 2007, 535–549 | 541

©

View Article Online

of two protons in the (H₂do3ap)^3–anion, where predominantly

the phosphonate-bearing and the opposite nitrogen atoms are

protonated.²³The next protonation around pH 5 (correspond to

pK_A= 5.04 determined by potentiometry) takes place mainly on

the carboxylate group of the propionate side chain as the highest

Dd_His observed for the –CH₂CH₂CO₂⁻methylene. Below pH 4.5,

two protons should be bound to the acetate pendants (with pK_A’s

close to 4.34 and 3.10 as determined by potentiometry). Again, the

protonations are reﬂected in the chemical shifts of all the pendant

arms. Such behaviour can be explained (similarly to an explanation

suggested for H₅do3ap²³) by a redistribution of protons bound to

the ring nitrogen atoms. A protonation of the acetates bound to

the N4 and N10 atoms causes a clear re-location of ring-bound

protons to the N1 and N7 atoms (Scheme 2). It is supported by a

further decrease of d_P(a higher extent of N1 protonation leads to a

decrease in d_P) and a more pronounce change in d_Hof the N7–CH₂

group; a smaller increase in d_Hof the N4– and N10–CH₂groups

is a consequence of a combination of the effects of deprotonation

of the nitrogen atoms (decreasing d_H) and protonation of the

carboxylate group (increasing d_H). The ﬁnal potentiometrically

determined dissociation constant (pK_A= 1.93) corresponds to

protonation of the N7 acetate group without re-location of the

two ring-bound protons. At low pH, a steep increase in d_Pclearly

points to protonation of the phosphinate moiety.

Stability constants of metal complexes

The stability constants of sodium(I), lanthanide(III) and yt-

trium(III) complexes with H₄do3ap^ABnand H₅do3ap^PrAwere deter-

mined and their values are given in Table 4. Some of the derived

constants and their comparison with those of similar ligands are

presented in Table 5.

Relatively high values of the stability constants of the sodium

complexes with the phosphinate ligands justiﬁed the use of

Me₄NCl as a background electrolyte. The value of the constant

for the Na⁺–H₅do3ap^PrAcomplex (4.72) and those with the

other ligands (4.77 for H₅do3ap²³; 4.2 for H₄dota⁴⁰) are similar.

Lower stability of the Na⁺–H₄do3ap^ABncomplex (2.75) can be a

consequence of the lower overall charge of the complex and/or

the presence of the hydrophobic benzyl substituent.

The values of the stability constants of the Ln³⁺and Y³⁺

complexes with the phosphinate ligands are similar to those with

H₄dota as overall basicities of all the ligands are almost the same

and the complexes are less stable than those of much more basic

H₅do3ap (stability of such complexes depends mainly on the

overall basicity of the ligands¹⁶). As proved by multinuclear NMR

solution^18–20and solid state⁵³studies on monophosphonic/inic

acid H₄dota derivatives, their lanthanide(III) complexes are octa-

coordinated by four amine nitrogen atoms and four oxygen atoms

of the pendant arms with a water molecule in the apical position

above the O4-plane (the apical water molecule is omitted in

formulae throughout the text for the sake of simplicity). Therefore,

the carboxylate/amine group in the phosphorus side chain should

not be directly bound to the central atom. As an indirect proof

of the solution structure, the acidities of the phosphorus side

chain groups are only slightly lower than those in the free ligands.

Fig.

2

The distribution diagrams of the H₅do3ap^PrA(a) and

Y³⁺–H₅do3ap^PrAsystem under equilibrium (b) and non-equilibrium con-

ditions (c).

adjacent CH₂group, the value of this coefﬁcient was determined

to be 0.75 ppm⁵²). It also proves that changes in chemical shift

between pH 8–10.5 are associated with double protonation. The

characteristic changes of d_Pwith protonation of a-amino group

(i.e. in the >N–CH₂–P(O)(R)O⁻moiety) have been known for a

long time;^13,48–51in this case, the Dd_Pis smaller than expected. It

supports the above statement of only a partial protonation. This

rather statistical distribution is in contrast to a good localization

The lanthanide(III) complexes are fully formed at −log[H⁺] > 3

ꢀ

(Fig. 2b). A correlation logb_GdL= (0.85

0.02) × pK_A(ref.

57b) can be applied and the following values for the stability

542 | Dalton Trans., 2007, 535–549

This journal is

The Royal Society of Chemistry 2007

©

View Article Online

Table 4 Stability constants of metal ion complexes with H₄do3ap^ABnand H₅do3ap^PrA(I = 0.1 mol dm⁻³Me₄NCl, 25 ^◦C)

a

logb_hlm

Equilibrium^b

hlm

Na⁺

La³⁺

Ce³⁺

Eu³⁺

Gd³⁺

Lu³⁺

Y³⁺

H₅do3ap^PrA

M + L ꢀ [ML]

011

111

211

4.72(6)

23.1(1)

27.45(6)

30.58(3)

24.4(1)

28.86(4)

31.24(5)

25.3(1)

29.77(3)

31.0(3)

25.04(9)

29.64(2)

31.49(6)

25.5(1)

29.87(2)

31.1(2)

24.63(9)

29.23(2)

31.16(6)

H + M + L ꢀ [M(HL)]

2H + M + L ꢀ [M(H₂L)]

H₄do3ap^ABn

M + L ꢀ [ML]

011

111

211

2.75(5)

22.4(3)

27.11(5)

23.4(2)

28.09(2)

29.98(8)

25.0(1)

29.23(2)

24.04(8)

28.80(1)

24.0(2)

29.14(2)

23.9(1)

28.40(6)

H + M + L ꢀ [M(HL)]

2H + M + L ꢀ [M(H₂L)]

^ab_hlm= [H_hL_lM_m]/([H]^h× [L]^l× [M]^m). ^bCharges are omitted for clarity.

Table 5 Stability (logK_ML) or dissociation (pK_A) constants of lanthanide(III) and yttrium(III) complexes with H₄do3ap^ABn, H₄do3ap^PrA, H₄dota and

H₅do3ap

logK_MLor pK_A

Equilibrium^a

La³⁺

Ce³⁺

Eu³⁺

Gd³⁺

Lu³⁺

Y³⁺

H₅do3ap^PrA

M + L ꢀ [ML]

23.1

4.4

3.13

24.4

4.5

2.38

25.3

4.5

1.2

25.04

4.60

1.85

25.5

4.4

1.2

24.63

4.60

1.93

[M(HL)] ꢀ H + [ML]

[M(H₂L)] ꢀ H + [M(HL)]

H₄do3ap^ABn

M + L ꢀ [ML]

22.4

4.7

23.4

4.7

1.89

25.0

4.2

24.04

4.76

24.0

5.1

23.9

4.3

[M(HL)] ꢀ H + [ML]

[M(H₂L)] ꢀ H + [M(HL)]

H₄dota

M + L ꢀ [ML]

22.9^b

24.6^c

1.9^c

24.7^b

25.4^b

24.3^f

24.9^g

24.4^h

26.48^d

24.67^e

25.58^d

[M(HL)] ꢀ H + [ML]

H₅do3apⁱ

M + L ꢀ [ML]

25.5

5.3

26.5

5.5

27.8

5.55

27.5

5.56

28.6

5.2

26.0

5.7

[M(HL)] ꢀ H + [ML]

^aCharges are omitted for clarity. ^bRef. 55. ^cRef. 56. ^dRef. 54. ^eRef. 39. ^fRef. 57. ^gRef. 58. ^hRef. 59. ⁱRef. 23.

Table 6 Overall stability^aand corresponding consecutive dissocia-

tion constants^bfor the H₅do3ap^PrAcomplexes obtained from “non-

equilibrium” titrations

constants of gadolinium(III) complexes were calculated: 26.8 0.6

(H₅do3ap^PrA) and 26.7 0.6 (H₄do3ap^ABn). The estimated values

are in good agreement with those determined experimentally.

To facilitate interpretation of the formation kinetics data (see

below) we ran titrations under non-equilibrium conditions to

estimate the distribution of species during the formation of the

complexes. This type of experiment has already been used for the

determination of the stability constants of “kinetic” intermediates

in Ln³⁺–H₄dota systems.^39,54We used H₅do3ap^PrAas the ligand

binds lanthanide(III) ions much faster than the other ligands

(see below). In the experiments, we titrated a mixture of a

lanthanide(III) and the ligand starting at acidic pH but the waiting

time was not long enough to establish full equilibrium. However,

the waiting time will be probably insufﬁcient mainly at −log[H⁺] <

6; above this value, complexation should be fast enough to reach

the thermodynamic equilibrium in 1 min (waiting time per titration

point). The approach cannot provide real stability constants for the

intermediates in the La³⁺/Y³⁺–H₅do3ap^PrAsystems (as those which

have been published for the H₄dota complexes^39,54) but it may give,

through the pK_Avalues, an idea of which sites are protonated

during the complexation process. The values of the constants

for the La³⁺–H₅do3ap^PrAand Y³⁺–H₅do3ap^PrAsystems are given

in Table 6. The values of the last two dissociation constants are

Constant

La³⁺

Y³⁺

logb_H3LM

logb_H2LM

logb_HLM

logb_LM

logb_H–1LM

pK(H₃LM)

pK(H₂LM)

pK(HLM)

33.28(2)

30.90(1)

27.20(2)

21.92(4)

9.48(5)

2.38

33.22(2)

30.74(1)

26.84(2)

21.50(4)

8.2(1)

2.48

3.90

3.70

5.28

5.34

^ab_hlm= [H_hL_lM_m]/([H]^h× [L]^l× [M]^m). ^bpK(H_nLM) = logb_HnLM

−

logb_Hn−1LM

.

rather high, higher than or comparable with dissociation constants

corresponding to the deprotonation of the acetic pendants of the

free ligand. Therefore, such values indirectly prove the mechanism

of complexation (see below) with kinetic intermediates protonated

on one or two nitrogen atoms. The structure of the ﬁnal complex

species above −log[H⁺] > 7 must be identical with the ﬁnal species

obtained under equilibrium conditions (octadentate coordination

of the ligand with one water molecule).^18,19,35aThe water molecule

This journal is

The Royal Society of Chemistry 2007

Dalton Trans., 2007, 535–549 | 543

©

View Article Online

in the apical position is deprotonated only at very high −log[H⁺].

The distribution diagrams under the non-equilibrium conditions

(more reliable for interpretation of kinetics of formation ) are

different from that at the full equilibrium. An example is shown

in Fig. 2.

rearrangement of eqn (2) and (3).

k_f× K^∗× [ligand]_tot

k ^ꢀ

=

f,obs

1 + K^∗× [ligand]_tot

1

A

×

(4)

2

A

1 + K₁× [AZ] + K₁× K₂× [AZ]

Kinetic study of yttrium(III) complexes with H₄dota-like ligands

A similar mathematical model was recently derived to describe

the inhibiting effect of zinc(II) complexes on phosphodiester

cleavage.⁶⁵If the measured pseudo-ﬁrst order rate constant k^ꢀ_f,obs

is extrapolated to zero concentration of the competing ligand

(Arsenazo III), eqn (4) is simpliﬁed to eqn (5). In addition, if

the conditional stability constant of the intermediate K* is high

and the employed experimental conditions are pseudo-ﬁrst order

(c_Lig≥ 10c_Y), one can assume that K* × [ligand]_totꢁ 1 and then

the formation rate is given by eqn (5).

The most suitable ligands for the yttrium(III) ion are derivatives

of H₄dota as their complexes are thermodynamically very stable

(logK_[Y(dota)]= 24–25; for stabilities of complexes with the other

ligands see Table 5) and kinetically inert. The structure of the

[Y(dota)]⁻complex in solution^58b,60as well as in the solid state⁶¹is,

as noted above, identical to the other lanthanide(III) complexes of

H₄dota with an octacoordinated central ion sandwiched between

parallel N4- and O4-planes. The same structural motif was also

found in [Y(Hdo3ap)]⁻(ref. 53) and [Y(Hdo3ap^ABn)] (ref. 62)

complexes inthesolidstate. Therefore, such structures are expected

for yttrium(III) complexes with all the ligands studied here. Despite

an extensive use of the ⁹⁰Y complexes in nuclear medicine only a

few quantitative kinetic studies on the yttrium(III) complexes of

H₄dota-like ligands have been published.^{11,33,36,58b,60}However, the

results are not comparable as they were obtained under different

or undeﬁned experimental conditions. Therefore, we decided to

perform a kinetic investigation into the behaviour of yttrium(III)

complexes under our conditions.

k_f× K^∗× [ligand]_tot

k_f,obs

=

≈

= k_f

(5)

1 + K^∗× [ligand]_tot

K^∗× [ligand]_tot

The great beneﬁt of the proposed experimental method is the

fact that the rate of formation reaction can be measured in

a pH region where measurement by conventional spectroscopy

is often impossible due to it being a very fast reaction. In

addition, the sensitivity of the measurement is very high due to

the strongly absorbing reagent, which forms very stable complexes

with lanthanide(III) ions, thus protecting them against hydrolysis

at higher pH.

Formation kinetics

Application of Arsenazo III to the measurement of the for-

mation kinetics of the Y³⁺–ligand complexes is shown in Fig.

S1 and S2.† The dependences of the observed pseudo-ﬁrst order

rate constant k^ꢀ_f,obson the concentration of Arsenazo III (Fig. S1

and S2†) are strictly linear. This proves that only one relatively

stable Y³⁺–Arsenazo III complex (probably of the Y : AZ = 1 :

1 stoichiometry) is formed. The measured values of k^ꢀ_f,obswere

extrapolated to zero concentration of Arsenazo III and these

values were used for calculation of the rate constant k_OHas the

rate-determining step of the reactions is catalyzed by hydroxide

ions (see below).⁷It is seen in Fig. 3 that the dependences k_f,obson

The formation reactions of the yttrium(III) ion with the studied

macrocyclic ligands were followed under pseudo-ﬁrst order condi-

tions in the −log[H⁺] range 5.5–7.1. The rate law for the formation

of the [Y(Lig)] species under pseudo-ﬁrst order conditions (c_Lig

c_Y) can be postulated as in eqn (1) and (2)

ꢁ

d[Y(L)]

= k_f,obs× [Y^{3 +}

]

(1)

tot

dt

k_f× K^∗× [ligand]_tot

1 + K^∗× [ligand]_tot

k_f,obs

=

(2)

where K* is the conditional equilibrium constant for the formation

of a kinetic intermediate and [ligand]_totis total concentration of

ligand.

The presence of an additional ligand (Arsenazo III) inhibits

the formation reaction of the complex as the concentration of the

free metal ion decreases due to formation of the Y³⁺–Arsenazo III

complex (assuming that complexes of Ln³⁺with Arsenazo III do

not react with the macrocylic ligand as transchelation is usually

a relatively slow reaction).^63,64Thus, the conditional pseudo-ﬁrst

order rate constant k^ꢀ_f,obsis lower than k_f,obsby a factor of d_Y

correcting the total concentration of Y³⁺to the concentration of

free Y³⁺ion (eqn (3)).

k_f,obs

A

k^ꢀ_f,obs= k_f,obs× d_Y

=

(3)

2

A

1 + K₁× [AZ] + K₁× K₂× [AZ]

Fig. 3 Hydroxide-assisted formation of yttrium(III) complexes with

the studied macrocyclic ligands; H₅do3ap (open squares ꢁ, full line),

H₅do3ap^PrA(open circles ᭺, full line), H₄dota (full circles ᭹, dashed line),

H₄dotp^Ph(full triangles ꢂ,full line) and H₄do3ap^ABn(open triangles , full

line).

The constants ^AK₁and ^AK₂are equilibrium constants corre-

sponding to the formation of Y³⁺–Arsenazo III containing-species

[Y(AZ)] and [Y(AZ)₂]; such species were assumed to be present in

Ln³⁺–Arzenazo III systems.^63,64Eqn (4) can be easily obtained by

544 | Dalton Trans., 2007, 535–549

This journal is

The Royal Society of Chemistry 2007

©

View Article Online

Table 7 The kinetic parameters of hydroxide-assisted formation of yttrium(III) complexes with H₄dota-like ligands (I = 0.1 mol dm⁻³KCl, 25 ^◦C)

Ligand

k_OH× 10⁻⁵/dm³mol⁻¹s⁻¹

k_H2O× 10²/s⁻¹

s^a(99%, pH 5.5)/min

s^a(99%, pH 6.5)/s

H₄dota

6.4(6)

34(3)

19(1)

1.93(8)

6.0(2)

3.0(8)

—

2.3

4.3

7.7

75.8

24.4

72

26

46

455

146

H₅do3ap

H₅do3ap^PrA

H₄do3ap^ABn

H₄dotp^Ph

^a99% complex formation; c_L= 1.67 × 10⁻⁵mol dm⁻³, c_M= 1.67 × 10⁻⁷mol dm⁻³, 25 ^◦C.

[OH⁻] are linear; the dependence was ﬁtted by a straight line using

eqn (6)

partners with charges of the same sign;⁶⁸in our case, between

[Y(HL)]*⁽⁻⁾and OH⁻species. In addition, the distribution diagram

calculated from “kinetic titration” data for the Y³⁺–H₅do3ap^PrA

system (Fig. 2c) shows that the most abundant protonated species

in the −log[H⁺] range of the kinetic experiments (5.5–7.1) is the

monoprotonated [Y(Hdo3ap^PrA)]⁻complex. The hydroxide non-

catalyzed pathway (k_H2O) was detected only for H₄dota. This may

be due to a slightly different solution structure of a reactive

ligand species and/or the intermediate complex which supports

the reaction non-catalyzed by hydroxide ion.

k_f,obs= k_OH× [OH⁻] + k_H

(6)

O

2

where k_OHand k_H2Oare rate constants for the pathways catalyzed

and non-catalyzed by hydroxide ion, respectively. All experimental

results are presented in Table 7.

In accordance with the above data, a reaction mechanism was

proposed (Scheme 3) and it is similar to a commonly accepted

mechanism for the formation of trivalent lanthanide ion complexes

with H₄dota-like ligands.⁷On the basis of literature data,^7,23,54

the employed −log[H⁺] range (5.5–7.1) and with regard to the

pK_Avalues (and to corresponding distribution diagrams) of the

ligands, we suggest that the reactive ligand species are ligand forms

protonated on two ring nitrogen atoms. Recently, such double

nitrogen protonation has been observed in crystal structures of the

[Er₂(H₂O)₂(H₂do3ap^Ph)₂]²⁺complex crystallized from aqueous so-

lutions at pH < 4.5 (for structure of H₄do3ap^Ph, see Chart 1)⁶⁶or for

the protonated gadolinium(III) complex of H₄dota tetraamide.⁶⁷

These (H₂L)^2–nligand species (n = 4 for H₄dota, H₄dotp^Phand

H₄do3ap^ABn; n = 5 for H₅do3ap and H₅do3ap^PrA) react with the

metal ion to form intermediate complexes [Y(HL)]*^(4–n)where the

metal ion is bound to four oxygen atoms on the pendant arms.

In the rate-determining step, the metal ion is transferred into the

ligand cavity (with a simultaneous removal of the last proton from

The obtained value of the pseudo-ﬁrst order rate constant k_f,obs

◦

for H₄dota (0.17 s⁻¹, pH 7.1, 25 C) differs from that published

◦

(k_obs7.7 × 10⁻⁵s⁻¹, pH 7.8, 25 C)³⁶and is much closer to value

determined by ¹H NMR spectroscopy (k_obs2.4 × 10⁻³s⁻¹, pH 6.7,

◦

37 C).⁶⁰However, Kimura et al.³⁶determined only a relative

order of reactivity during yttrium(III) complex formation for

compounds having a different size of macrocyclic rim. Therefore,

the experimental conditions employed in both studies are different

and the comparison of results from both studies is only qualitative.

From the practical point of view, it is very useful to compare

the time necessary for full complexation (Table 7). At pH 5.5,

the rate of complexation of H₄dota (2.3 min) is comparable to

that of H₃do3a (2.4 min)³³despite the fact that one pendant arm

stabilizing the kinetic intermediate is missing in H₃do3a. Because

of the availability of the hydroxide-independent pathway, the

complexation of H₄dota is relatively faster in more acidic solutions

when compared with the phosphorus ligands. The complexa-

tion rates with monophosphonic/inic acid H₄dota derivatives

highly depend on the phosphorus substituent. Ligands with an

acidic group (OH or CH₂CH₂CO₂H) form complexes as fast as

H₄dota and much faster than H₄do3ap^ABnwith a hydrophobic

benzyl substituent. Similarly, the presence of hydrophobic phenyl

substituents may be responsible for the rather slow reaction of

H₄dotp^Phin comparison with H₄dota. It is commonly accepted

that the formation reaction is driven by a proton transfer from

the ring nitrogen atoms to the oxygen atoms of the pendant

arms in the reaction intermediate.¹¹Consequently, the proton is

transferred to the bulk water. Therefore, the observed differences

in the formation kinetics of the yttrium(III)–macrocycle systems

can be explained by changes in the proton transfer ability of the

pendant arms (ligands with pendant arms with sterically accessible

ionisable group such as H₅do3ap or H₅do3ap^PrAtransfer the proton

more efﬁciently) or in accessibility of the ligand interior of the

intermediate (hydrophobic substituents retard the complexation

reaction) for the bulk water molecules in the rate-determining

step. It was also suggested that the value of the last dissociation

constant of such macrocyclic ligands may be responsible for the

differences in the rate of complex formation.⁶⁰Data presented in

the cavity) to form the ﬁnal complex and the step is catalyzed (k_OH

or uncatalyzed (k_H2O) by hydroxide ions.

)

Scheme 3 Mechanism of formation of yttrium(III) complexes with the

H₄dota-like ligands.

The mechanism is also supported by measurement of k_f,obsvalues

at ionic strength I → 0.0 mol dm⁻³(i.e. only with the contribution

of the dilute reactants and the MES buffer at −log[H⁺] 6.5) and

at I = 0.1 mol dm⁻³. The values of k_f,obsfor chosen macrocyclic

ligands (H₅do3ap, H₅do3ap^PrA) increased several (5–10) times with

an increase of ionic strength. This means that the rate-limiting step

deﬁned by rate constant k_OHcorresponds to a reaction between

This journal is

The Royal Society of Chemistry 2007

Dalton Trans., 2007, 535–549 | 545

©

View Article Online

this paper show that such an explanation is not correct for this

ligand series because the most basic ligand, H₅do3ap, forms the

yttrium(III) complex quickly but the opposite is true for the least

basic ligand, H₄dotp^Ph.

Assuming fast protonation of the yttrium(III) complexes (de-

scribed by constant K₁), this mechanism leads to eqn (11)

k₁+ k₂× K₂× [H⁺]

k_d,obs

=

(11)

1 + K₂× [H⁺]

where the equilibrium constant K₂is deﬁned by eqn (10). If k₁ꢂ

Dissociation kinetics

k₂× K₂(see ﬁtting in Fig. 4) then eqn (11) is simpliﬁed to eqn (12)

It is commonly accepted that kinetic inertness is more important

for stability of metal complexes in vivo than thermodynamic

stability.^7,69Mostly, kinetic inertness is proved through an acid-

catalyzed complex dissociation as it is the most important

decomposition pathway of such complexes.⁷The decomplexation

can be also assisted by other metal ions such as Zn²⁺and/or

Cu²⁺which are also present in bioﬂuids.^7,69–72However, these

metal-catalyzed pathways are not important for complexes of

macrocyclic H₄dota-like ligands.^69,71Therefore, to determine and

compare the kinetic inertness of the yttrium(III) complexes, acid-

assisted decomplexation was performed. The kinetics of the acid-

assisted decomplexation of the complexes was studied in the region

0.05–3.0 mol dm⁻³HClO₄at I = 3.0 mol dm⁻³(H,Na)ClO₄and

25 ^◦C.

k₁+ k₂× K₂× [H⁺]

1 + K₂× [H⁺]

k₂× K₂× [H⁺]

1 + K₂× [H⁺]

k_d,obs

=

≈

(12)

In the case of a low value of equilibrium constant K₂(K₂can be

neglected in the denominator), eqn (12) is further simpliﬁed to eqn

(13)

k_d,obs= k₂× K₂× [H⁺] = k_H× [H⁺]

(13)

This means that the rate-determining step is associated with

an equilibrium reaction in which the reaction intermediate is

formed.

In the literature, the mechanism of dissociation of lanthanide(III)

complexes with H₄dota has been established.^7,34,73,74The oxygen

atom(s) of the pendant arm(s) are protonated and a proton is

transferred onto a ring nitrogen atom with subsequent decompo-

sition of the complex. The mechanism is commonly described by

the reactions shown in Scheme 4. The pre-equilibrium reaction

(characterized by constant K₁) leads to relatively kinetically inert

complexes; the value of the constant K₁was determined by

potentiometric titrations for the [Y(dota)]⁻complex (logK₁

=

3.08)^58band for several lanthanide(III) complexes of H₅do3p

(logK₁= 1.4–1.7).²³

Fig.

4

Acid-assisted dissociation of yttrium(III) complexes with

H₄dota-like ligands (I = 3.0 mol dm⁻³(H,Na)ClO₄, 25 ^◦C); H₄dota (full

circles ᭹, dashed line), H₄dotp^Ph(full triangles ꢂ,full line), H₅do3ap^PrA

(open circles ᭺, dashed line), H₅do3ap (open squares ꢁ, full line) and

H₄do3ap^ABn(open triangles , dashed line).

The dependences of the pseudo-ﬁrst-order constant k_d,obson

solution acidity are presented in Fig. 4. The experimental data were

ﬁtted by eqn (12) and the values of the parameters are presented

in Table 8. For the H₅do3ap and H₄dotp^Phcomplexes, some

difﬁculties in the non-linear ﬁtting appeared (mainly associated

with a low value of K₂) leading to high errors of parameters k₂and

K₂. For H₄do3ap^ABn, only eqn (13) could be used as the dependence

of k_d,obson [H⁺] is almost linear; therefore, the value of K₂× [H⁺]

has to be low, respectively K₂× [H⁺] ꢂ 1. If we suppose that K₂×

[H⁺] reaches not more than 5% of unity, one obtains (for the used

range of [H⁺]) an estimate of K₂as ∼0.05. The estimates are given

in Table 8.

Scheme 4 Mechanism of dissociation of yttrium(III) complexes (charges

are omitted) with the H₄dota-like ligands; n = 2 for H₅do3ap, H₅do3ap^PrA

and H₄do3ap^ABn; n = 1 for H₄dota and H₄dotp^Ph

.

Generally, the rate of dissociation of the complexes is given by

eqn (7).

d[complex]

−

= k_d,obs× [complex]_tot

(7)

dt

where the total concentration of a complex is obtained from mass

balance (eqn (8)) and by corresponding equilibrium constants (eqn

(9) and (10)) (for n see Scheme 4)

Our experimental data (k_d,obs= 2.00 × 10⁻⁴s⁻¹(0.01 M H⁺), see

Table 8) for the [Y(dota)]⁻complex differs from those published

by Parker et al.^17,58(k_d,obs= 3.3 × 10⁻⁷s⁻¹(0.01 M H⁺), k_d,obs= 1.5 ×

10⁻⁵s⁻¹(0.1 M H⁺)) obtained with different experimental methods

(application of ⁹⁰Y radionuclide using HPLC with radiometric de-

tection or ¹³C NMR) and under different experimental conditions

(a very low concentration of the complex with ⁹⁰Y or c_complex∼1 ×

10⁻³mol dm⁻³for NMR, I = 0.1 mol dm⁻³, glycine–HCl buffer,

[Y(L)]_tot= [Y(H_nL)] + [Y(H_n+1L)]

(8)

[H_nYL]

K₁=

(9)

[H_n−1YL][H⁺]

[H_{n + 1}YL]

K₂=

(10)

[H_nYL][H⁺]

546 | Dalton Trans., 2007, 535–549

This journal is

The Royal Society of Chemistry 2007

©

View Article Online

Table 8 Parameters of acid-assisted dissociation of yttrium(III) complexes with H₄dota-like ligands (I = 3.0 mol dm⁻³(H,Na)ClO₄, 25 ^◦C)

Decomposition in 0.01 M H⁺, 25 ^◦C

k₂× 10³/s⁻¹

K₂/m³mol⁻¹

logK₂

k_H× 10³/m³mol⁻¹s⁻¹

k

_d,obs/s⁻¹

s_1/2/h

a

Ligand

H₄dota^b

3.7(3)

7.0^d

5.7(2.2)

0.2^d

0.76

−0.70^d

<−1.3^d

0.11

21.1^c

2.00 × 10⁻⁴

1.09 × 10⁻⁵

1.02 × 10⁻⁵

4.36 × 10⁻⁵

2.98 × 10^−5f

3.24 × 10^−5g

0.96

17.7

18.9

4.4

H₅do3ap

1.09(5)^e

1.02(3)^e

4.4^c

H₄do3ap^ABn

H₅do3ap^PrAb

H₄dotp^Phb

>25^d

<0.05^d

1.3(4)

3.4(5)

8.8(2.5)^f

8.9(1.8)^g

0.34(13)^f

0.37(17)^g

−0.47^f

3.0^c,f

6.5^f

−0.44^g

3.3^c,g

5.9^g

^alogK₂= pK_A([H_n+1YL]). ^bFitted according to eqn (12). ^ck_H= k₂× K₂. ^dEstimated values according to eqn (12). ^eFitted according to eqn (13). ^fVIS

measurement with auxiliary ligand (Arsenazo III). ^gDirect UV measurement.

∼37 ^◦C). The differences between the results can be assigned

to different experimental conditions, mostly solution acidity,

temperature, metal concentration and ionic strength; it can be seen

also from Parker’s data where the ten-times increase of solution

acidity can lead to 45-times acceleration of complex dissociation.

We used NaClO₄to maintain a constant ionic strength and this

high concentration of Na⁺ions can also somewhat perturb our

kinetic data through a Na⁺–complex ion-paring similar to that

suggested for the metal-assisted dissociation pathway of complexes

of acyclic ligands.⁷⁰All these effects can have a high impact on the

rate of complex dissociation. It is seen from Table 8 that values

of the rate constant k₂are not dramatically different for various

yttrium(III) complexes except the [Y(do3ap^ABn)]⁻complex where

k₂is much higher than those for the other ligands. A much higher

variation is observed for values of equilibrium constant K₂of the

complexes. Complexes of phosphonic/phosphoric acid derivatives

have a lower proton afﬁnity in the rate-determining step than

the complexes of H₄dota or H₃do3a.³³The lowest afﬁnity was

found for the yttrium(III) complex of H₄do3ap^ABnwhich is already

a positively charged (as [Y(H₂do3ap^ABn)]⁺species due to additional

protonation of amino group in the pre-equilibrium step, Scheme 4)

before binding of another proton in the proton-catalyzed pathway.

A combination of the different proton afﬁnities and various

rate constants leads to differences in half-lives of the complexes

in acid solution. A practical comparison of the rates of acid-

assisted decomposition (half-lives) is shown in Table 8. Complexes

of phosphonic/phosphoric acid derivatives are more inert than

that of H₄dota and their inertness varies with phosphorus atom

substituents, mostly as a consequence of the differences in proton

afﬁnity. This is most evident in the case of the positively charged

complex of H₄do3ap^ABn. Such a deceleration of the decomplexa-

tion reaction due to the formation of a stable positively charged

intermediate has been observed for copper(II)⁷⁵and nickel(II)⁷⁶

complexes of the bis(methylphosphonic acid) derivative of cyclam

(1,8-H₄te2p, Chart 1). However, as for the formation kinetics,

an inﬂuence of hydrophobic substituent(s) preventing access of

water molecules cannot be avoided. As expected, the complex of

heptadentate H₃do3a³³is less inert than the complex of octaden-

tate H₄dota. A rough correlation of k_Hvalues (k_H= k₂× K₂)

with the ﬁrst protonation constant of the ligands is obvious from

Fig. S3 (such correlations have been used to explain differences

in reactivity/inertness of other macrocyclic complexes³⁴). This

points out that more basic ligands form more inert complexes.

It can be explained by the idea that the metal–nitrogen bond

is harder to split as bonding energy is higher for bonds with

more basic nitrogen atoms. However, more data on the extended

series of complexes are necessary to conﬁrm such considera-

tions.

Conclusions

The bifunctional monomethylphosphinic acid derivatives of

H₄dota form lanthanide(III) complexes of the same thermody-

namic stability as those with H₄dota itself. Kinetic studies on

yttrium(III) complexes with several macrocyclic ligands showed

that the formation rate of complexes of the monophosphonic/inic

ligands is comparable to that of the H₄dota complex. The

formation rate of the complexes depends highly on ligand structure

and minor changes such as substitution of phosphorus atom lead

to a big difference in the formation kinetics. The proton transfer

ability imposed by the structure of a kinetic intermediate seems

to be responsible for the variations in rates of formation. Kinetic

inertness of complexes of the new ligands is higher than that of

the H₄dota complex and the data suggest that substitution with

phosphonic/inic acid pendant arms is a suitable modiﬁcation

of the H₄dota skeleton in the design of new ligands for ⁹⁰Y

immunotherapy. Rates of the acid-assisted decomplexation of the

complexes depend again on the substituents on the phosphorus

atom.

Finally, rates of complexation and decomplexation depend on

the ability to transfer proton(s) outside/inside the complex cavity

and/or its water accessibility and, therefore, on the hydrophobicity

of the ligands. Yttrium(III) complexes of the monophospho-

nic/inic acid derivatives of H₄dota are highly thermodynamically

stable and kinetically inert and, therefore, they are suitable for

possible medicinal uses.

Acknowledgements

We thank Ms M. Mal´ıkova´ for help with titration experiments. We

also thank Dr J. Rudovsky´ and T. Vitha for NMR measurements

and Dr I. C´ısarˇova´ for X-ray data collections. The work was sup-

ported by the Grant Agency of the Czech Republic (203/06/0467),

Grant Agency of Universita Karlova (423/2004/B-CH/PrF) and

Academy of Science of the Czech Republic (KAN 201110651).

The work was performed in the frame of COST D38 and EMIL

(No. LSHC-2004-503569) and DiMI (No. LSHB-2005-512146)

“Network of Excellence” projects.

This journal is

The Royal Society of Chemistry 2007

Dalton Trans., 2007, 535–549 | 547

©

View Article Online

Chemometrics ‘95, Pardubice, Czech Republic, 1995, p. 63;

full version of “OPIUM” is available (free of charge) on

http://www.natur.cuni.cz/∼kyvala/opium.html; (b) M. Kyvala, P.

References

1 (a) M. Meyer, V. Dahaoui-Ginderey, C. Lecomte and R. Guilard,

Coord. Chem. Rev., 1998, 178–180, 1313; (b) X. Liang and P. J. Sadler,

Chem. Soc. Rev., 2004, 33, 246.

2 P. Caravan, J. J. Ellison, T. J. Mc Murry and R. B. Lauffer, Chem. Rev.,

1999, 99, 2293.

3 (a) The Chemistry of Contrast Agents in Medical Magnetic Resonance

´

Lubal and I. Lukes, IX. Spanish-Italian, and Mediterranean Congress

on Thermodynamics of Metal Complexes (SIMEC 98), Girona, Spain,

1998.

ˇ

31 (a) A. E. Martell and R. M. Smith, Critical Stability Constants, Plenum

Press, New York, 1974–1989, vol. 1–6; (b) A. E. Martell, R. M. Smith

and R. J. Motekaitis,NIST Standard Reference Database 46 (Critically

Selected Stability Constants of Metal Complexes), Version 7.0, NIST

Standard Reference Data, Gaithersburg, MD, 2003.

32 C. F. Baes, Jr. and R. E. Mesmer, The Hydrolysis of Cations, Wiley,

New York, 1976.

33 H.-Z. Cai and T. A. Kaden, Helv. Chim. Acta, 1994, 77, 383.

34 (a) K. Kumar, C. A. Chang and M. F. Tweedle, Inorg. Chem., 1993, 32,

587; (b) K. Kumar and M. F. Tweedle, Inorg. Chem., 1993, 32, 4193.

35 (a) P. Ta´borsky´, I. Svobodova´, Z. Hnatejko, P. Lubal, S. Lis, M.

Fo¨rsterova´, P. Hermann, I. Lukesˇ and J. Havel, J. Fluoresc., 2005, 15,

507; (b) P. Ta´borsky´, I. Svobodova´, P. Lubal, Z. Hnatejko, S. Lis, P.

Hermann and I. Lukesˇ, unpublished results.

´

Imaging, ed. A. E. Merbach and E. To´th, Wiley, Chichester, 2001;

(b) Top. Curr. Chem., 2002, 221.

4 (a) S. G. Crich, A. Barge, E. Battistini, C. Cabella, S. Coluccia, D.

Longo, V. Mainero, G. Tarone and S. Aime, J. Biol. Inorg. Chem., 2005,

10, 78; (b) S. Aime, C. Cabella, S. Colombatto, S. G. Crich, E. Gianolio

and F. Maggioni, J. Magn. Reson. Imaging, 2002, 16, 394.

5 (a) M. Bottrill, L. Kwok and N. J. Long, Chem. Soc. Rev., 2006, 35,

557; (b) V. J. Venditto, C. A. S. Regino and M. W. Brechbiel, Mol.

Pharmacol., 2005, 2, 302.

6 Handbook of Radiopharmaceuticals. Radiochemistry and Applications,

ed. M. J. Welch and C. S. Redvanly, Wiley, Chichester, 2003.

7 E. Bru¨cher, Top. Curr. Chem., 2002, 221, 123.

8 W. A. P. Breeman, M. de Jong, T. J. Visser, J. L. Erion and E. P. Krenning,

Eur. J. Nucl. Med. Mol. Imaging, 2003, 30, 917.

9 M. V. Deshmukh, G. Voll, A. Ku¨hlewein, H. Ma¨cke, J. Schmitt, H.

Kessler and G. Gemmecker, J. Med. Chem., 2005, 18, 1506.

10 Y.-H. Jang, M. Blanco, S. Dasgupta, D. A. Keire, J. E. Shively and

W. A. Goddard, J. Am. Chem. Soc., 1999, 121, 6142.

11 D. A. Keire, Y. H. Jang, L. Li, S. Dasgupta, W. A. Goddard and J. E.

Shively, Inorg. Chem., 2001, 40, 4310.

12 H.-S. Chong, K. Garmestani, D. Ma, D. E. Milenic, T. Overstreet and

M. W. Brechbiel, J. Med. Chem., 2002, 45, 3464.

13 C. F. G. C. Geraldes, A. D. Sherry and W. P. Cacheris, Inorg. Chem.,

1989, 28, 3336.

14 C. J. Broan, K. J. Jankowski, R. Kataky, D. Parker, A. M. Randall and

A. Harrison, J. Chem. Soc., Chem. Commun., 1990, 1739.

15 (a) I. Belskii, Y. M. Polikarpov and M. I. Kabachnik, Usp. Khim., 1992,

61, 415; (b) D. Parker and J. A. G. Williams, J. Chem. Soc., Dalton

Trans., 1996, 3613.

16 I. Lukesˇ, J. Kotek, P. Vojt´ısˇek and P. Hermann, Coord. Chem. Rev.,

2001, 216–217, 287.

17 K. P. Pulukkody, T. M. Norman, D. Parker, L. Royle and C. J. Broan,

J. Chem. Soc., Perkin Trans. 2, 1993, 605.

18 J. Rudovsky´, P. C´ıgler, J. Kotek, P. Hermann, P. Vojt´ısˇek, I. Lukesˇ, J. A.

Peters, L. V. Elst and R. N. Muller, Chem.–Eur. J., 2005, 11, 2373.

19 (a) J. Rudovsky´, M. Botta, P. Hermann, K. I. Hardcastle, I. Lukesˇ and

S. Aime, Bioconjugate Chem., 2006, 17, 975; (b) J. Rudovsky´, J. Kotek,

P. Hermann, I. Lukesˇ, V. Mainero and S. Aime, Org. Biomol. Chem.,

2005, 3, 112; (c) J. Rudovsky´, P. Hermann, M. Botta, S. Aime and I.

Lukesˇ, Chem. Commun., 2005, 2390.

36 M. Kodama, T. Koike, A. B. Mahatma and E. Kimura, Inorg. Chem.,

1991, 30, 1270.

37 D. J. Alner, J. J. Greczek and A. G. Smeeth, J. Chem. Soc. A, 1967, 1205.

38 A. E. Wro´blewski and J. G. Verkade, J. Am. Chem. Soc., 1996, 118,

10168.

39 A. Bianchi, L. Calabi, C. Giorgi, P. Losi, M. Palma, P. Paoli, P. Rossi,

B. Valtancoli and M. Virtuani, J. Chem. Soc., Dalton Trans., 2000, 697.

40 G. Anderegg, F. Arnaud-Neu, R. Delgado, J. Felcman and K. Popov,

Pure Appl. Chem., 2005, 77, 1445.

41 F. Marques, L. Gano, M. P. Campello, S. Lacerda, I. Santos, L. M. P.

Lima, J. Costa, P. Antunes and R. Delgado, J. Inorg. Biochem., 2006,

100, 270.

42 R. Delgado, L. C. Siegfried and T. A. Kaden, Helv. Chim. Acta, 1990,

73, 140.

´

43 A. D. Sherry, J. Ren, J. Huskens, E. Bru¨cher, E. To´th, C. F. C. G.

Geraldes, M. M. C. A. Castro and W. P. Cacheris, Inorg. Chem., 1996,

35, 4604.

44 X. Sun, M. Wuest, Z. Kovacs, A. D. Sherry, R. Motekaitis, Z. Wang,

A. E. Martell, M. J. Welch and C. J. Anderson, J. Biol. Inorg. Chem.,

2003, 8, 217.

45 K. Bazakas and I. Lukesˇ, J. Chem. Soc., Dalton Trans., 1995, 1133.

46 I. La´za´r, A. D. Sherry, R. Ramasamy, E. Bru¨cher and R. Kira´ly, Inorg.

Chem., 1991, 30, 5016.

47 T. Kiss and I. La´za´r, in Aminophosphonic and Aminophosphinic Acids.

Chemistry and Biological Activity, ed. V. P. Kukhar and H. R. Hudson,

Wiley, Chichester, 2000, ch. IX, pp. 285–326.

48 T. G. Appleton, J. R. Hall, A. D. Harris, M. A. Kimlin and I. J. Mc

Mahon, Aust. J. Chem., 1984, 37, 1833.

49 L. Burai, J. Ren, Z. Kovacs, E. Bru¨cher and A. D. Sherry, Inorg. Chem.,

1998, 37, 69.

20 J. Rudovsky´, M. Botta, P. Hermann, A. Koridze and S. Aime, Dalton

Trans., 2006, 2323.

50 J. Kotek, P. Vojt´ısˇek, I. C´ısarˇova´, P. Hermann, P. Jurecˇka, J. Rohovec

and I. Lukesˇ, Collect. Czech. Chem. Commun., 2000, 65, 1289.

51 K. P. Guerra, R. Delgado, L. M. P. Lima, M. G. B. Drew and V. Fe´lix,

Dalton Trans., 2004, 1812.

52 J. L. Sudmeier and C. N. Reilley, Anal. Chem., 1964, 9, 1698.

53 P. Vojt´ısˇek, P. C´ıgler, J. Kotek, J. Rudovsky´, P. Hermann and I. Lukesˇ,

Inorg. Chem., 2005, 44, 5591.

54 J. Moreau, E. Guillon, J.-C. Pierrard, J. Rimbault, M. Port and M.

Aplincourt, Chem.–Eur. J., 2004, 10, 5218.

55 W. P. Cacheris, S. K. Nickle and A. D. Sherry, Inorg. Chem., 1987, 26,

958.

21 C. J. Broan, E. Cole, K. J. Jankowski, D. Parker, K. Pulukkody, B. A.

Boyce, N. R. A. Beeley, K. Millar and A. T. Millican, Synthesis, 1992,

63.

22 (a) L. M. Shkolnikova and M. A. Porai-Koshits, Usp. Khim., 1990, 59,

1111; (b) K. L. Nash, R. D. Rogers, J. Ferraro and J. Zhang, Inorg.

Chim. Acta, 1998, 269, 211.

23 P. Ta´borsky´, P. Lubal, J. Havel, J. Kotek, P. Hermann and I. Lukesˇ,

Collect. Czech. Chem. Commun., 2005, 70, 1909.

24 J. F. Desreux, Inorg. Chem., 1980, 19, 1319.

25 J. Rohovec, M. Ky´vala, P. Vojt´ısˇek, P. Hermann and I. Lukesˇ,

Eur. J. Inorg. Chem., 2000, 195.

26 (a) Z. Otwinowski and W. Minor, HKL Denzo and Scalepack Program

Package, Nonius BV, Delft, Netherlands, 1997; (b) Z. Otwinowski and

W. Minor, Methods Enzymol., 1997, 276, 307.

27 A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C.

Burla, G. Polidori and M. Camalli, SIR92—A Program for Automatic

Solution of Crystal Structures by Direct Methods, J. Appl. Crystallogr.,

1994, 27, 435.

56 L. Burai, I. Fa´bia´n, R. Kira´ly, E. Szilgyi and E. Bru¨cher, J. Chem. Soc.,

Dalton Trans., 1998, 243.

57 (a) K. Kumar, C. A. Chang, L. C. Franncesconi, D. D. Dischino, M. F.

Malley, J. Z. Gougoutas and M. F. Tweedle, Inorg. Chem., 1994, 33,

3567; (b) K. Kumar, M. F. Tweedle, M. F. Malley and J. Z. Gougoutas,

Inorg. Chem., 1995, 34, 6472.

58 (a) J. P. L. Cox, K. J. Jankowski, R. Kataky, D. Parker, N. R. A. Beeley,

B. A. Boyce, M. A. W. Eaton, K. Millar, A. T. Millican, A. Harrison

and C. Walker, J. Chem. Soc., Chem. Commun., 1989, 797; (b) C. J.

Broan, J. P. L. Cox, A. S. Craig, R. Kataky, D. Parker, A. Harrison,

A. M. Randall and G. Ferguson, J. Chem. Soc., Perkin Trans. 2, 1991,

87.

28 (a) G. M. Sheldrick, SHELXL97—A Computer Program for Reﬁnement

of Crystal Structures, University of Go¨ttingen, Go¨ttingen, Germany,

1997; (b) G. M. Sheldrick and T. R. Schneider, Methods Enzymol.,

1997, 277, 319.

29 A. L. Spek, PLATON—A Multipurpose Crystallographic Tool, Utrecht

59 M. Koudelkova´, H. Vinsˇova´ and V. Jedina´kova´-Krˇ´ızˇova´, J. Chro-

University, Utrecht, The Netherlands, 2005.

matogr., A, 2003, 990, 311.

30 (a) M. Ky´vala and I. Lukesˇ, International Conference on

548 | Dalton Trans., 2007, 535–549

This journal is

The Royal Society of Chemistry 2007

©

View Article Online

60 A. Keire and M. Kobayashi, Bioconjugate Chem., 1999, 10, 454.

61 (a) D. Parker, K. Pulukkody, F. C. Smith, A. Batsanov and J. A. K.

Howard, J. Chem. Soc., Dalton Trans., 1994, 689; (b) C. A. Chang,

L. C. Franncesconi, M. F. Malley, K. Kumar, J. Z. Gougoutas and

M. F. Tweedle, Inorg. Chem., 1993, 32, 3501.

62 J. Rudovsky´, J. Kotek, P. Hermann and I. Lukesˇ, Inorg. Chem., 2006,

45, 3097.

68 (a) J. H. Espenson, Chemical Kinetics and Reaction Mechanisms,

McGraw Hill, New York, 1981; (b) R. G. Wilkins, Kinetics and Mech-

anism of Reactions of Transition Metal Complexes, VCH, Weinheim,

1991; (c) R. W. Hay, Reaction Mechanisms of Metal Complexes, Ellis

Horwood, Chichester, 2000.

69 M. F. Tweedle, J. J. Hagan, K. Kumar, S. Mentha and C. A. Chang,

Magn. Reson. Imaging, 1991, 9, 409.

63 (a) Y.-W. Lu, G. Laurent and H. Pereira, Talanta, 2004, 62, 959; (b) E.

Hosten and H. Rohwer, Anal. Chim. Acta, 1997, 345, 227; (c) H. Rohwer

and E. Hosten, Anal. Chim. Acta, 1997, 339, 271; (d) H. Rohwer, N.

Collier and E. Hosten, Anal. Chim. Acta, 1995, 314, 219.

64 Y. Shi, Q. Ji, E. M. Eyring and R. van Eldik, J. Chem. Soc., Dalton

Trans., 1996, 2127.

65 O. Iranzo, A. Y. Kovalevsky, J. R. Morrow and J. P. Richard, J. Am.

Chem. Soc., 2003, 125, 1988.

66 J. Rohovec and P. Vojt´ısˇek, Collect. Czech. Chem. Commun., 2006, 71,

264.

70 L. Sarka, L. Burai and E. Bru¨cher, Chem.–Eur. J., 2000, 6, 719.

71 S. Laurent, L. V. Elst, F. Copoix and R. N. Muller, Invest. Radiol.,

2001, 36, 115.

72 J. Kotek, F. K. Ka´lma´n, P. Hermann, E. Bru¨cher, K. Binnemans and I.

Lukesˇ, Eur. J. Inorg. Chem., 2006, 1976.

´

73 E. To´th, E. Bru¨cher, I. La´za´r and I. To´th, Inorg. Chem., 1994, 33, 4070.

74 S.-L. Wu and W. deW. Horrocks, Inorg. Chem., 1995, 34, 3724.

75 J. Kotek, P. Lubal, P. Hermann, I. C´ısarˇova´, I. Lukesˇ, T. Godula,

I. Svobodova´, P. Ta´borsky´ and J. Havel, Chem.–Eur. J., 2003, 9,

233.

67 P. A. Stenson, A. L. Thompson and D. Parker, Dalton Trans., 2006,

76 J. Kotek, P. Vojt´ısˇek, I. C´ısarˇova´, P. Hermann and I. Lukesˇ, Collect.

3291.

Czech. Chem. Commun., 2001, 66, 363.

This journal is

The Royal Society of Chemistry 2007

Dalton Trans., 2007, 535–549 | 549

©

Article Doi

Thermodynamic study of lanthanide(iii) complexes with bifunctional monophosphinic acid analogues of H4dota and comparative kinetic study of yttrium(iii) complexes

DOI: 10.1039/b613404a

Source and publish data:

Authors:

Article abstract of DOI:10.1039/b613404a

Full text of DOI:10.1039/b613404a

Products guided by the article

R&D Labs maybe for 57384-29-9

Relevant to this article

Hot Product