Molecular recognition of sialic acid by lanthanide(III) complexes through cooperative two-site binding

Article abstract of DOI:10.1021/ic902461g

Herein we report two new ligands, 1,4,7-tris(carboxymethyl)-10-[2- (dihydroxyboranyl)benzyl]-1,4,7,10-tetraazacyclododecane (L¹) and 1,4,7-tris(carboxymethyl)-10-[3-(dihydroxyboranyl)benzyl]-1,4,7, 10-tetraazacyclododecane (L²), whic

Full text of DOI:10.1021/ic902461g

4212 Inorg. Chem. 2010, 49, 4212–4223
DOI: 10.1021/ic902461g
Molecular Recognition of Sialic Acid by Lanthanide(III) Complexes through
Cooperative Two-Site Binding
†
†
§
Martın Regueiro-Figueroa, Kristina Djanashvili,‡ David Esteban-Gomez, Thomas Chauvin,§ Eva Toth,
ꢀ
ꢀ
,†
ꢀ
´
ꢀ
†
,†
´
Andres de Blas, Teresa Rodrıguez-Blas,* and Carlos Platas-Iglesias*
†
´
~
Departamento de Quımica Fundamental, Universidade da Coruna, Campus da Zapateira, Alejandro de la Sota 1,
‡
~
15008 A Coruna, Spain, Biocatalysis and Organic Chemistry, Department of Biotechnology, Delft University
§
ꢀ
of Technology, Julianalaan 136, 2628 BL Delft, The Netherlands, and Centre de Biophysique Moleculaire,
ꢀ
CNRS, rue Charles Sadron, 45071 Orleans Cedex 2, France
Received December 11, 2009
Herein we report two new ligands, 1,4,7-tris(carboxymethyl)-10-[2-(dihydroxyboranyl)benzyl]-1,4,7,10-tetraazacyclo-
dodecane (L1) and 1,4,7-tris(carboxymethyl)-10-[3-(dihydroxyboranyl)benzyl]-1,4,7,10-tetraazacyclododecane (L2),
which contain a phenylboronic acid (PBA) function and a 1,4,7,10-tetraazacyclododecane-1,4,7-triacetate cage for
complexation of lanthanide ions in an aqueous solution. The pKa of the PBA function amounts to 4.6 in [Gd(L1)] and
8.9 in [Gd(L2)], with the value of the L2 analogue being very similar to that of PBA (8.8). These results are explained by
the coordination of the PBA function of L1 to the GdIII ion, which results in a dramatic lowering of its pKa. As a
consequence, [Gd(L1)] does not bind to saccharides at physiological pH. The nuclear magnetic relaxation dispersion
profiles recorded for [Gd(L1)] and [Gd(L2)] confirm that the phenylboronate function is coordinated to the metal ion in
the L1 derivative, which results in a q=0 complex. The interaction of the [Gd(L2)] complex with 5-acetylneuraminic acid
(Neu5Ac) and 2-R-O-methyl-5-acetylneuraminic acid (MeNeu5Ac) has been investigated by means of spectro-
photometric titrations in an aqueous solution (pH 7.4, 0.1 M 3-(N-morpholino)propanesulfonic acid buffer).
Furthermore, we have also investigated the binding of these receptors with competing monosaccharides such as
D-(þ)-glucose, D-fructose, D-mannose, D-galactose, methyl R-D-galactoside, and methyl R-D-mannoside. The binding
constants obtained indicate an important selectivity of [Gd(L2)] for Neu5Ac (Keq = 151) over D-(þ)-glucose (Keq
=
12.3), D-mannose (Keq=21.9), and D-galactose (Keq=24.5). Furthermore, a very weak binding affinity was observed in
the case of methyl R-D-galactoside and methyl R-D-mannoside. An 8-fold increase of the binding constant of [Gd(L2)]
with Neu5Ac is observed when compared to that of PBA determined under the same conditions (Keq =19). 13C NMR
spectroscopy and density functional theory calculations performed at the B3LYP/6-31G(d) level show that this is due to
a cooperative two-site binding of Neu5Ac through (1) ester formation by interaction on the PBA function of the receptor
and (2) coordination of the carboxylate group of Neu5Ac to the GdIII ion. The emission lifetime of the 5D4 level of TbIII in
[Tb(L2)] increases upon Neu5Ac binding, in line with the displacement of inner-sphere water molecules due to
coordination of Neu5Ac to the metal ion.
Introduction
magnetic resonance imaging1 and as radiopharmaceuticals
for the diagnosis and therapy of tumors.2 An important goal
of the research in this field is to achieve a precise delivery of
these compounds to specific targets in vivo.3 The general
strategy used for this purpose is to attach to these chelates
bioconjugates or recognition moieties that could bind to
specific receptors that are overexpressed in certain tissues.
For instance, sialic acid is considered to be a tumor marker
because it is known to be overexpressed on the surface of
tumor cells.4 Variations in the amount of sialic acid incorpo-
rated have been correlated with colon cancer, brain tumors,
and immune regulation in the human body.5 Moreover,
serum levels of sialic acid can be used as a biological marker
Lanthanide chelates with macrocyclic ligands are of great
interest because they are used both as contrast agents for
*To whom correspondence should be addressed. E-mail: cplatas@udc.es
(C.P.-I.), mayter@udc.es (T.R-B.).
(1) The Chemistry of Contrast Agents in Medical Magnetic Resonance
ꢀ
ꢀ
Imaging; Merbach, A. E., Toth, E., Eds.; Wiley: New York, 2001.
(2) (a) Liu, S. Adv. Drug Delivery Rev. 2008, 60, 1347–1370. (b) Roesch, F.
Radiochim. Acta 2007, 95, 303–311. (c) Volkert, W. A.; Hoffman, T. J. Chem.
Rev. 1999, 99, 2269–2292. (d) De Jong, M.; Breeman, W. A. P.; Kwekkeboom,
E. J.; Valkema, R.; Krenning, E. P. Acc. Chem. Res. 2009, 42, 873–880.
(3) (a) De Leon-Rodriguez, L. M.; Lubag, A. J. M.; Malloy, C. R.;
Martinez, G. V.; Gillies, R. J.; Sherry, A. D. Acc. Chem. Res. 2009, 42, 948–
957. (b) Caravan, P. Acc. Chem. Res. 2009, 42, 851–862. (c) De Leon-Rodriguez,
L. M.; Kovacs, Z. Bioconjugate Chem. 2008, 19, 391–402. (d) Aime, S.; Crich,
S. G.; Gianolio, E.; Giovenzana, G. B.; Tei, L.; Terreno, E. Coord. Chem. Rev.
2006, 250, 1562–1579.
(4) Schauer, R. Glycoconjugate J. 2000, 17, 485–499.
(5) Dall0Olio, F.; Chiricolo, M. Glycoconjugate J. 2001, 18, 841–850.
r
pubs.acs.org/IC
Published on Web 04/07/2010
2010 American Chemical Society

Article
Inorganic Chemistry, Vol. 49, No. 9, 2010 4213
Chart 1. 5-Acetylneuraminic Acid (1, Neu5Ac) and 2-R-O-Methyl-5-acetylneuraminic Acid (2, MeNeu5Ac)
of a cancerous state.6 Thus, molecular recognition of sialic
acid is important to developing assays that could enable the
enhanced detection of malignancies and other diseases invol-
ving high sialic acid levels. Sialic acids contain the nine-
carbon amino sugar neuraminic acid, with the most common
member of this family being 5-acetylneuraminic acid (Neu5Ac;
Chart 1).7 Sialic acids contain a carboxy group at the ano-
meric carbon atom that gives the molecule a negative charge
at physiological pH. Neu5Ac and its derivatives generally
occupy the terminal positions of carbohydrate chains of
glycoproteins and glycolipids in biological membranes, and
they appear to play important roles in cellular recognition
processes.8 Neu5Ac is present in an aqueous solution as an
equilibrium between R-pyranose (5-8%) and β-pyranose
forms (92-95%), while the furanose type of Neu5Ac is absent
because of the acetamide moiety present at the C5 position.9
In addition to these cyclic forms, three acyclic forms have
been recently detected at low levels (0.5-1.9% at pH 2).10
However, it must be pointed out that all known glycosides of
Neu5Ac are R-linked, often to a galactose or a N-galactosa-
mine residue through a R(2f3) or R(2f6) linkage.11
phenylboronic acid (PBA)16 and PBA derivatives17 have
been reported in the literature. A noninvasive method for
the detection of sialic acid at the cell membrane by using a
PBA monolayer gold electrode has also been reported re-
cently.18 It has been suggested that more selective artificial
receptors for sialic acid residues in glycoproteins or free sialic
acid should contain both a PBA function and a group
capable of recognizing the negatively charged COO- group
of sialic acid (Chart 1). This has been achieved by the
introduction of positively charged units, such as ammonium
groups, in the framework of the receptor.19-21 A second
strategy uses diboronate receptors, which can bind free sialic
acid through both the glycerol side chain and the R-hydroxy
acid function at the anomeric position (Chart 1).22 A third
approach consists of the use of a metal chelate, typically
a ZnII or CuII complex, to promote a cooperative binding of
boronic acid and metal chelate. Indeed, Shinkai et al. have
developed a fluorescent sensor containing a PBA moiety that is
able to interact with diols and a positively charged 1,10-phen-
anthrolinezinc(II) chelate moiety for carboxylate binding.23
A few conjugates of PBA and lanthanide(III) complexes of
diethylenetriaminepentaacetate (DTPA),20,24 diethylenetria-
minepentaacetate bisamides,21 and 1,4,7,10-tetraazacyclodo-
decane-1,4,7,10-tetraacetate (DOTA) tetraamides25 have
been reported in the literature. These conjugates contain
Boronic acids are a promising class of recognition moie-
ties for the synthesis of artificial receptors and sensors for
R-hydroxycarboxylates12 or molecules containing diol groups13
such as saccharides.14 The use of boronic acids for this pur-
pose is based on their ability to form reversible covalent
complexes with 1,2- and 1,3-diol units in saccharides.15
Detailed investigations on the interaction of Neu5Ac with
(16) Djanashvili, K.; Frullano, L.; Peters, J. A. Chem.;Eur. J. 2005, 11,
4010–4018.
(17) Otsuka, H.; Uchimura, E.; Koshino, H.; Okano, T.; Kataoka, K.
J. Am. Chem. Soc. 2003, 125, 3493–3502.
(18) Matsumoto, A.; Sato, N.; Kataoka, K.; Miyahara, Y. J. Am. Chem.
Soc. 2009, 131, 12022–12023.
(19) Patterson, S.; Smith, B. D.; Taylor, R. A. Tetrahedron Lett. 1998, 39,
3111–3114.
(20) Battistini, E.; Mortillaro, A.; Aime, S.; Peters, J. A. Contrast Media
Mol. Imaging 2007, 2, 163–171.
(6) Krolikowski, F. J.; Reuter, K.; Waalkes, T. P.; Sieber, S. M.;
Adamson, R. H. Pharmacology 1976, 14, 47–51.
(7) Schauer, R. Zoology 2004, 107, 49–64.
(8) Varki, A. Trends Mol. Med. 2008, 14, 351–360.
(9) Friebolin, H.; Kunzelmann, P.; Supp, M.; Brossmer, R.; Keilich, G.;
Ziegler, D. Tetrahedron Lett. 1981, 22, 1383–1386.
(21) (a) Frullano, L.; Rohovec, J.; Aime, S.; Maschmeyer, T.; Prata, M. I.;
Pedroso de Lima, J. J.; Geraldes, C. F. G. C.; Peters, J. A. Chem.;Eur. J.
2004, 10, 5205–5217. (b) Djanashvili, K.; Koning, G. A.; van der Meer,
A. J. G. M.; Wolterbeek, H. T.; Peters, J. A. Contrast Media Mol. Imaging
2007, 2, 35–41.
(22) Levonis, S. M.; Kiefel, M. J.; Houston, T. A. Chem. Commun. 2009,
2278–2280.
(23) (a) Yamamoto, M.; Takeuchi, M.; Shinkai, S. Tetrahedron 1998, 54,
3125–3140. (b) Takeuchi, M.; Yamamoto, M.; Shinkai, S. Chem. Commun. 1997,
1731–1732.
(24) Takahashi, K.; Nakamura, H.; Furumoto, S.; Yamamoto, K.;
Fukuda, H.; Matsumura, A.; Yamamoto, Y. Bioorg. Mol. Chem. 2005, 13,
735–743.
(25) (a) Trokowski, R.; Zhang, S.; Sherry, A. D. Bioconjugate Chem.
2004, 15, 1431–1440. (b) Zhang, S.; Trokowski, R.; Sherry, A. D. J. Am. Chem.
Soc. 2003, 125, 15288–15289. (c) Ren, J.; Trokowski, R.; Zhang, S.; Malloy,
C. R.; Sherry, A. D. Magn. Reson. Med. 2008, 60, 1047–1055.
(10) Klepach, T.; Carmichael, I.; Serianni, A. S. J. Am. Chem. Soc. 2008,
130, 11892–11900.
(11) (a) Varki, A. Nature 2007, 446, 1023–1029. (b) Murrey, H. E.;
Hsieh-Wilson, C. Chem. Rev. 2008, 108, 1708–1731. (c) Angata, T.; Varki, A.
Chem. Rev. 2002, 102, 439–469.
(12) (a) Zhu, L.; Zhong, Z.; Anslyn, E. V. J. Am. Chem. Soc. 2005, 127,
4260–4269. (b) Zhang, X.; Chi, L.; Ji, S.; Wu, Y.; Song, P.; Han, K.; Guo, H.;
James, T. D.; Zhao, J. J. Am. Chem. Soc. 2009, 131, 17452–17463.
(13) Shabbir, S. H.; Joyce, L. A.; da Cruz, G. M.; Lynch, V. M.; Sorey, S.;
Anslyn, E. V. J. Am. Chem. Soc. 2009, 131, 13125–13131.
(14) (a) Mader, H. S.; Wolfbeis, O. S. Microchim. Acta 2008, 162, 1–34.
(b) James, T. D. Top. Curr. Chem. 2007, 277, 107–152. (c) James, T. D.; Shinkai,
S. Top. Curr. Chem. 2002, 218, 159–200. (d) James, T. D.; Tony, D.;
Sandanayake Samankumara, K. R. A.; Shinkai, S. Angew. Chem., Int. Ed.
1996, 35, 1911–1922. (e) Zhang, T.; Anslyn, E. V. Org. Lett. 2007, 9, 1627–1629.
(15) Fujita, N.; Shinkai, S.; James, T. D. Chem. Asian J. 2008, 3, 1076–
1091.

4214 Inorganic Chemistry, Vol. 49, No. 9, 2010
Chart 2. Ligands Reported in This Work
Regueiro-Figueroa et al.
Scheme 1. Synthesis of Ligand H3L1
octadentate binding units for coordination of the LnIII ions,
which therefore are not expected to participate in molecular
recognition processes. Lanthanide(III) complexes with hep-
tadentate ligands such as 1,4,7,10-tetraazacyclododecane-
1,4,7-triacetate (DO3A) derivatives are typically diaqua
systems that are known to bind different oxyanions, which
replace inner-sphere water molecules to give q = 0 com-
plexes.26 A suitable receptor for sialic acid recognition could
therefore be based on a Ln(DO3A) complex, which could
bind the carboxylate function of sialic acid, attached to a
PBA function for recognition of the 1,2- and 1,3-diol groups
of Neu5Ac. Thus, herein we report receptors H3L1 and H3L2,
which contain a PBA function attached to a nitrogen atom of
a DO3A backbone (Chart 2). The presence of these two
recognition units is expected to increase the specificity and
stability of the interaction between the corresponding lantha-
nide complexes and sialic acid. Furthermore, the lanthanide-
(III) complexes of these ligands were designed to interact
reversibly with sialic acid. Interaction of the lanthanide(III)
complexes of L1 and L2 with Neu5Ac and other competing
monosaccharides was investigated in aqueous solutions by
in 82% yield. However, N-alkylation reactions28 of 1,4,7-
tris[(tert-butoxycarbonyl)methyl]-1,4,7,10-tetraazacyclodo-
decane [DO3A(t-BuO)3] with 4 in refluxing acetonitrile in
the presence of Na2CO3 did not provide any results. Most
likely, the presence of the bulky protecting group in 4
causes some steric hindrance that prevents the N-alkyla-
tion. As an alternative, we decided to prepare [2-(bromo-
methyl)phenyl]ethylene borate (2).29 The treatment of
DO3A(t-BuO)3 with 2 in refluxing acetonitrile in the pre-
sence of Na2CO3, followed by purification with column
chromatography, afforded the pure protected precursor 5
in 63% yield. Finally, full deprotection of the tert-butyl
esters and PBA function was achieved with trifluoroacetic
acid (TFA), providing the desired ligand as the trifluoro-
means of spectrophotometric titrations and 1H, 13C, and 11
B
NMR spectroscopies, as well as by density functional theory
(DFT) calculations performed at the B3LYP/6-31G(d) level.
The potential interference of R-hydroxy acid species present
in vivo, such as lactate and citrate, was also analyzed. Fur-
thermore, we also investigated the interaction of these com-
plexes with 2-R-O-methyl-5-acetylneuraminic acid (MeNeu5Ac)
as a model for cell-bound neuraminic residues, which are
R-linked via the -OH group in position 2.9 Finally, the nuc-
lear magnetic relaxation dispersion (NMRD) profiles of the
gadolinium(III) complexes are reported.
acetate salt in 75% yield. The synthesis of H3L2 3CF3-
Results and Discussion
3
COOH was achieved from m-tolylboronic acid following
an analogous synthetic procedure.
Synthesis and Characterization of the Complexes. Li-
Synthesis of the Ligands. The synthetic strategy used for
the preparation of H3L1 is outlined in Scheme 1. In a first
attempt toward L1, we decided to prepare the 2-(bromo-
methyl)phenyl borate ester 4.27 The latter compound was
obtained in two steps from commercially available o-tolyl-
boronic acid, which was reacted with 2,2-dimethyl-1,3-
propanediol to give the protected 2-phenylboronic acid 3.
Free-radical bromination of 3 with N-bromosuccinimide
(NBS) and dibenzoyl peroxide (PDB) afforded compound 4
gands H3L1 and H3L2 were derivatized to form several
charge-neutral lanthanide complexes of formulas [Ln(L1)]
3
2H2O and [Ln(L2)] 4H2O (Ln = La, Eu, Gd, or Tb),
3
which were obtained in 54-79% yield by reaction of the
ligand with equimolar amounts of the corresponding
lanthanide triflate in the presence of triethylamine. The
mass spectra of the complexes (positive ion electrospray
ionization, ESIþ) show intense peaks due to [Ln(HL1)]þ
and [Ln(HL2)]þ entities, which confirms the formation of
the desired complexes.
(26) (a) Dickins, R. S.; Aime, S.; Batsanov, A. S.; Beeby, A.; Botta, M.;
Bruce, J. I.; Howard, J. A. K.; Love, C. S.; Parker, D.; Peacock, R. D.;
Puschmann, H. J. Am. Chem. Soc. 2002, 124, 12697–12705. (b) Aime, S.;
Botta, M.; Bruce, J. I.; Mainero, V.; Parker, D.; Terreno, E. Chem. Commun.
2001, 115–116. (c) Bruce, J. I.; Dickins, R. S.; Govenlock, L. J.; Gunnlaugsson,
T.; Lopinski, S.; Lowe, M. P.; Parker, D.; Peacock, R. D.; Perry, J. J. B.; Aime, S.;
Botta, M. J. Am. Chem. Soc. 2000, 122, 9674–9684. (d) Terreno, E.; Botta, M.;
Boniforte, P.; Bracco, C.; Milone, L.; Mondito, B.; Uggeri, F.; Aime, S. Chem.;
Eur. J. 2005, 11, 5531–5537.
The pKa values of PBAs can be determined following
the change in the UV absorption that occurs when the
boron atom converts from the tetrahedral to the trigonal
(28) Barge, A.; Tei, L.; Upadhyaya, D.; Fedeli, F.; Beltrami, L.; Stefania,
R.; Aime, S.; Cravotto, G. Org. Biomol. Chem. 2008, 6, 1176–1184.
(29) Glennon, J. D.; Walker, A.; Mckervey, M. A. ARKIVOC 2003, 7,
402–411.
(27) (a) James, T. D.; Sandanayake, K. R. A. S.; Iguchi, R.; Shinkai, S.
J. Am. Chem. Soc. 1995, 117, 8982–8987. (b) Stones, D.; Manku, S.; Lu, X.;
Hall, D. G. Chem.;Eur. J. 2004, 10, 92–100.

Article
Inorganic Chemistry, Vol. 49, No. 9, 2010 4215
Table 1. Luminescence Emission Lifetimes of [Tb(L1)] and [Tb(L2)] in H2O and
D2O Solutions at 25 °C and pH 7.4 (0.1 M MOPS Buffer) and Calculated
Hydration Numbers (q)a
τ(H2O)/ms
τ(D2O)/ms
qa
qb
[Tb(L1)]
[Tb(L2)]
1.69
1.18
2.05
1.88
0.4
1.3
0.2
1.3
a Equation 1. b Equation 2.
shift of the 11B NMR signal to ca. -20 ppm due to con-
version of the boronic acid function from the tetrahedral
to the trigonal planar form (Figure S2 in the Supporting
Information). Analysis of the 11B NMR titration profile
provides a pKa of 8.9, in nice agreement with the value
obtained from spectrophotometric titrations.
Luminescence lifetime measurements have been widely
used for quantification of the number of inner-sphere
water molecules in lanthanide complexes.33 Although
ions that emit in the near-infrared such as YbIII have also
been used for this purpose, TbIII and EuIII are most com-
monly applied for lifetime measurements because they
emit in the visible region and show longer emission life-
times. Unfortunately, the europium(III) complexes of
L1 and L2 do not show EuIII-centered emission upon exci-
tation through the PBA chromophore at 270 nm. How-
ever, the emission spectrum of the terbium(III) analogues
recorded at pH 7.4 [0.1 M 3-(N-morpholino)propane-
sulfonic acid (MOPS)] shows the typical 5D4
f
7FJ
transitions with maxima at ca. 621 (J = 3), 587 (J = 4),
545 (J = 5), and 490 nm (J = 6) (Figure S3 in the
Supporting Information). The emission lifetimes of the
Tb(5D4) excited-state levels have been measured in D2O
and H2O solutions of the complexes (Table 1) and were
used to calculate the number of coordinated water mole-
cules q according to eqs 1,34 and 2:35
Figure 1. Spectrophotometric titrations of a 10-3 M aqueous solution
of [Gd(L1)] (top) and [Gd(L2)] (bottom) with a standard HCl solution.
Insets: changes in the molar absorbance at selected wavelengths as a
function of the pH. The solid lines represent the best fittings of the
experimental data.
q ¼ 4:2Δkobs
ð1Þ
planar form.30 The UV-vis absorption spectra of the
[Gd(L1)] and [Gd(L2)] complexes recorded at pH 12.0 [I=
0.1 M KCl] show a maximum at ∼290 nm. Upon a
decrease in the pH of the solution, the intensity of this
band decreases, while the formation of a new band with a
maximum at 270 nm is observed (Figure 1). This band can
be assigned as a π* r π singlet-singlet transition cen-
tered on the trigonal-planar PBA unit (1Lb transition
according to the suggestion of Ramsey).31 Analysis of
the titration profiles provides very different pKa values
for the two complexes: 4.6 and 8.9 for the complexes of L1
and L2, respectively. The pKa determined for [Gd(L2)] is
very similar to that determined by the same method for
PBA itself (pKa =9.0; Figure S1 in the Supporting Infor-
mation) as well as to that reported previously for PBA
(pKa = 8.8).32 These results can be only explained by
coordination of the PBA function of L1 to the GdIII ion,
which results in a dramatic lowering of its pKa. The 11B
NMR spectrum of a 10 mM aqueous solution of [La(L2)]
recorded at pH 1.6 shows a single resonance at 10 ppm.
Increasing the pH of the solution provokes an upfield
q ¼ 5:0ðΔkobs - 0:06Þ
where Δkobs =kobs(H2O) - kobs(D2O) and kobs =1/τobs
ð2Þ
.
For [Tb(L2)], we obtain q = 1.3, which suggests the
presence of an equilibrium in an aqueous solution invol-
ving q = 2 and 1 complex species. Average hydration
numbers of 1.3-1.4 have been previously obtained for
different terbium(III) complexes with DO3A-like binding
sites.36 For [Tb(L1)], we obtain clearly longer emission
lifetimes than those for the L2 analogue, in line with the
replacement of inner-sphere water molecules by coordi-
nation of the PBA function to the metal ion. The q values
obtained suggest that only one OH oscillator is coordi-
nating to the lanthanide ion, which can be accounted for
by coordination of a -OH group of the PBA function.
These results suggest that in the terbium(III) complex of
(33) Horrocks, W. D., Jr.; Sudnick, D. R. Acc. Chem. Res. 1981, 14, 384–
392.
(34) Horrocks, W. D., Jr.; Sudnick, D. R. J. Am. Chem. Soc. 1979, 101,
334–340.
(35) Beeby, A.; Clarkson, I. M.; Dickins, R. S.; Faulkner, S.; Parker, D.;
Royle, L.; de Sousa, A. S.; Williams, J. A. G.; Woods, M. J. Chem. Soc.,
Perkin Trans. 2 1999, 493–504.
(36) Faulkner, S.; Pope, S. J. A. J. Am. Chem. Soc. 2003, 125, 10526–
10527.
(30) Soundararajan, S.; Badawi, M.; Kohlrust, C. M.; Hageman, J. H.
Anal. Biochem. 1989, 178, 125–134.
(31) Ramsey, B. G. J. Phys. Chem. 1970, 74, 2464–2469.
(32) Yan, J.; Springsteen, G.; Deeter, S.; Wang, B. Tetrahedron 2004, 60,
11205–11209.

4216 Inorganic Chemistry, Vol. 49, No. 9, 2010
Regueiro-Figueroa et al.
lifetime measurements on terbium(III) complexes some-
times give lower q values than data collected on the cor-
responding europium(III) complex, in full accordance
with the lanthanide contraction.40
Aiming to obtain information on the solution structure
of the lanthanide(III) complexes of L1 and L2, we have
performed DFT calculations at the B3LYP level. On the
grounds of our previous experience,41 in these calcula-
tions, we used the effective core potential (ECP) of Dolg
et al.42 and the related [5s4p3d]-GTO valence basis set
for gadolinium, while the remaining atoms were described
by using the 6-31G(d) basis set. Luminescence lifetime
measurements performed on the terbium(III) complex of
L1 indicate that the PBA function is involved in coordina-
tion to the lanthanide, which results in a q = 0 complex.
Furthermore, the boron atom of the ligand is expected to
be in the tetrahedral form at neutral pH (see above).
However, in the case of L2 complexes, both luminescence
emission lifetimes and relaxivity measurements point to
an equilibrium involving q=1 and 2 complexes, with the
boron atom of the ligand being in the trigonal planar form
at neutral pH. Thus, we performed our calculations on
the [Gd(L1OH)]-, [Gd(L2)(H2O)], and [Gd(L2)(H2O)2]
systems.
Figure 2. 1H NMRD profiles (25 °C) of [Gd(L1)] (0, 0.979 mM),
[Gd(L2)] (O, 1.62 mM), and [Gd(L2)] (b, 1.62 mM) upon the addition
of 10 equiv of Neu5Ac.
L1 coordination of the PBA function does not allow
coordination of water molecules to the metal ion, result-
ing in a q = 0 complex.
NMRD profiles of [Gd(L1)] and [Gd(L2)] (Figure 2; see
also Figure S4 in the Supporting Information) were
recorded at 25 and 37 °C (pH 7.4, N-2-(hydroxyethyl)-
piperazine-N’-2-ethanesulfonic acid buffer). The relaxa-
tion rates of the bulk water protons are enhanced around
a paramagnetic ion because of long-range interactions
(outer-sphere relaxation) and short-range interactions
(inner-sphere relaxation). According to the standard
Solomon-Bloembergen-Morgan model, the latter pro-
cess is governed by four correlation times: the rotational
correlation time of the complex (τR), the residence time of
a water proton in the inner coordination sphere (τm), and
the longitudinal and transverse electronic relaxation
times (T1e and T2e) of the metal center.37 The shape of
the NMRD profiles recorded for [Gd(L1)] and [Gd(L2)]
(Figure 2) is similar to those of other low-molecular-
weight DOTA-type chelates of gadolinium(III). The re-
laxivity values obtained for [Gd(L1)] are clearly lower
than those obtained for the L2 analogue, in line with a
lower hydration number of the former because of coordi-
nation of the phenylboronate pendant arm. The NMRD
profile of [Gd(L2)] recorded at 25 °C is very similar to that
reported for [Gd(DO3A)],38 which suggests that both
complexes present similar hydration numbers. In the case
of [Eu(DO3A)], a hydration equilibrium between q = 2
and 1 complex species with an average hydration number
of q = 1.88 was established on the basis of UV-vis
measurements.39 The relaxivity values therefore suggest
that for [Gd(L2)] the hydration equilibrium is shifted to
the q=2 species and that our luminescence measurements
on the terbium(III) complex underestimate the number of
inner-sphere water molecules. This is in line with previous
investigations that showed that luminescence emission
It is well-known that in LnIII(DOTA)-like complexes
the four ethylenediamine groups adopt gauche confor-
mations, giving rise to two macrocyclic ring conforma-
tions: (δδδδ) and (λλλλ). Furthermore, there are two
possible orientations of the pendant arms (absolute con-
figuration Δ or Λ), resulting in four possible steroisomers,
existing as two enantiomeric pairs. These stereoisomers
differ by the layout of the four acetate arms, adopting
either a square-antiprismatic (SAP) or a twisted square-
antiprismatic (TSAP) geometry.43 The two structures
display a different orientation of the two square planes
formed by the four cyclen nitrogen atoms and the four
binding oxygen atoms, making an angle of ca. 40° in SAP-
type structures, whereas this situation is reversed and
reduced to ca. -30° in TSAP-type derivatives.44 As expec-
ted, our calculations performed in the gas phase provide
the SAP [Δ(λλλλ)] and TSAP [Δ(δδδδ)] isomers as mini-
mum-energy conformations for all complexes. Further-
more, the PBA function of L1 can coordinate to the GdIII
ion by adopting syn or anti conformations.45 In the syn
isomer, the orientation of the PBA function follows the
direction of rotation of the acetate arms, while in the anti
(40) Woods, M.; Aime, S.; Botta, M.; Howard, J. A. K.; Molones, J. M.;
Navet, M.; Parker, D.; Port, M.; Rousseaux, O. J. Am. Chem. Soc. 2000, 122,
9781–9792.
(41) (a) Roca Sabio, A.; Mato-Iglesias, M.; Esteban-Gomez, D.; Toth, E.;
de Blas, A.; Platas-Iglesias, C.; Rodriguez-Blas, T. J. Am. Chem. Soc. 2009,
131, 3331–3341. (b) Gonzalez-Lorenzo, M.; Platas-Iglesias, C.; Avecilla, F.;
Faulkner, S.; Pope, S. J. A.; de Blas, A.; Rodriguez-Blas, T. Inorg. Chem. 2005,
44, 4254–4262. (c) Mato-Iglesias, M.; Roca-Sabio, A.; Palinkas, Z.; Esteban-
Gomez, D.; Platas-Iglesias, C.; Toth, E.; de Blas, A.; Rodriguez-Blas, T. Inorg.
Chem. 2008, 47, 7840–7851.
(42) Dolg, M.; Stoll, H.; Savin, A.; Preuss, H. Theor. Chim. Acta 1989, 75,
173–194.
(43) (a) Aime, S.; Botta, M.; Fasano, M.; Marques, M. P. M.; Geraldes,
C. F. G. C.; Pubanz, D.; Merbach, A. E. Inorg. Chem. 1997, 36, 2059–2068.
(b) Hoeft, S.; Roth, K. Chem. Ber. 1993, 126, 869–873. (c) Aime, S.; Botta, M.;
Ermondi, G. Inorg. Chem. 1992, 31, 4291–4299.
(37) Peters, J. A.; Huskens, J.; Raber, D. J. Prog. Magn. Reson. Spectrosc.
1996, 28, 283–350.
(38) Aime, S.; Botta, M.; Geninatti Crich, S.; Giovenzana, G.; Pagliarin,
R.; Sist., M.; Terreno, E. Magn. Reson. Chem. 1998, 36, S200–S208.
(39) Toth, E.; Ni Dhubhghaill, O. M.; Besson, G.; Helm, L.; Merbach,
A. E. Magn. Reson. Chem. 1999, 37, 701–708.
(44) Parker, D.; Dickins, R. S.; Puschmann, H.; Crossland, C.; Howard,
J. A. K. Chem. Rev. 2002, 102, 1977–2010.
(45) Polasek, M.; Kotek, J.; Hermann, P.; Cisarova, I.; Binnemans, K.;
Lukes, I. Inorg. Chem. 2009, 48, 466–475.

Article
Inorganic Chemistry, Vol. 49, No. 9, 2010 4217
form, the aromatic ring is bent backward. Our calcula-
tions indicate that the syn conformation is the most stable
one for both the SAP and TSAP forms. Furthermore,
they provide relative free energies of the SAP isomer with
respect to the TSAP one of -26.6 ([Gd(L1OH)]-), -1.63
([Gd(L2)(H2O)]), and -6.92 kJ mol-1 ([Gd(L2)(H2O)2]),
3
which suggests that the complexes adopt in solution a
SAP geometry. The 1H NMR spectrum of the europium
complex of L1 recorded in a D2O solution supports this
hypothesis. In DOTA-like complexes, the signals due to
axial protons in the macrocyclic backbone are usually
found between 24 and 36 ppm in the SAP isomer and
between 5 and 12 ppm in the TSAP isomer.46 In the case of
[Eu(L1OH)]-, these signals are observed between 28 and
35 ppm (Figure S5 in the Supporting Information), which
suggests that the major species in solution corresponds to
the SAP form. The 1H NMR spectrum of the L2 analogue
shows very broad signals, as is usually observed for
N-alkylated lanthanide(III) complexes of DO3A. As a
result, the positions of the signals due to axial protons of
the macrocycle could not be established.
The minimum-energy conformations obtained for the
[Gd(L1OH)]- and [Gd(L2)(H2O)2] complexes are shown
in Figure 3. For [Gd(L2)(H2O)], the mean distance be-
tween the metal ion and the donor atoms of the carboxy-
˚
late groups is 2.30 A in both isomers, while the average
˚
˚
Gd-N distances amount to 2.73 A (SAP) and 2.74 A
(TSAP). The distance between the metal ion and the
oxygen atom of the coordinated water molecule is slightly
˚
longer for the TSAP isomer (2.52 A) than for the SAP one
(2.50 A), in line with a higher steric compression around
˚
the inner-sphere water molecule in the TSAP form.47 As
expected because of the increase in the coordination
number, for [Gd(L2)(H2O)2], the bond distances of the
metal coordination environment are slightly longer than
those for the analogue with q = 1: the mean distance be-
tween the metal ion and the donor atoms of the carboxy-
Figure 3. Minimum-energy conformations obtained from DFT calcu-
lations for the [Gd(L1OH)]- (top) and [Gd(L2)(H2O)2] (bottom) systems
[Δ(λλλλ) forms]. Hydrogen atoms are omitted for simplicity.
D-fructose, D-galactose, and D-mannose has been follo-
wed by means of spectrophotometric titrations on 10-3 M
solutions of the complexes (pH 7.4, MOPS buffer).
Furthermore, we have also investigated the interaction
of these complexes with lactate and citrate, which exist in
relatively high concentration in blood and are known to
bind PBAs.49 For comparative purposes, we have also
carried out spectrophotometric titrations to investigate
the interaction of these saccharides and R-hydroxy acids
with PBA under the same experimental conditions. In the
case of [Gd(L1)], saccharide addition did not provoke any
change on the absorption spectrum of the complex. Thus,
we conclude that the coordination of the PBA function to
the GdIII ion at physiological pH prevents saccharide
binding to this complex. However, the situation is very
different for the L2 analogue.
˚
late groups is 2.34 A in both isomers, while the average
˚
˚
Gd-N distances amount to 2.75 A (SAP) and 2.78 A
(TSAP). For [Gd(L1OH)]-, the average Gd-N distances
are 2.76 A (SAP) and 2.78 A (TSAP), while the average
˚
˚
˚
Gd-O distance amounts to 2.33 A for both isomers. The
calculated Gd-O bond distances are very close to those
observed in the solid state for gadolinium(III) complexes
with DO3A and DOTA derivatives, while the calculated
˚
Gd-N distances are typically 0.05-0.10 A longer than
those observed in the solid state.48
Study of the Interaction of [Gd(L2)] with Carbohydrates
and r-Hydroxy Acids. The interaction of [Gd(L1)] and
[Gd(L2)] with several carbohydrates such as D-(þ)-glucose,
The absorption spectrum of [Gd(L2)] recorded at pH
7.4 presents a weak absorption with a maximum at ca. 270 nm
typical of the PBA chromophore.31 The intensity of this
band progressively decreases upon saccharide addition
(Figure 4; see also Figures S6 and S7 in the Supporting
Information). The addition of lactate or citrate results in
similar spectral changes (Figures S8 and S9 in the Sup-
porting Information). Nonlinear least-squares fits of the
UV-vis titration profiles allowed us to determine the
(46) (a) Main, M.; Snaith, J. S.; Meloni, M. M.; Jauregui, M.; Sykes, D.;
Faulkner, S.; Kenwright, A. M. Chem. Commun. 2008, 5212–5214. (b) Adair,
C.; Woods, M.; Zhao, P.; Pasha, A.; Winter, P. M.; Lanza, G. M.; Athey, P.; Sherry,
A. D.; Kiefer, G. E. Contrast Media Mol. Imaging 2007, 2, 55–58.
(47) Woods, M.; Kovacs, Z.; Zhang, S.; Sherry, A. D. Angew. Chem., Int.
Ed. 2003, 42, 5889–5892.
(48) (a) Aime, S.;Batsanov, A. S.; Botta, M.; Howard, J. A. K.; Lowe, M. P.;
Parker, D. New J. Chem. 1999, 23, 669–670. (b) Chang, C. A.; Francesconi, L. C.;
Malley, M. F.; Kumar, K.; Gougoutas, J. Z.; Tweedle, M. F.; Lee, D. W.; Wilson., L. J.
Inorg. Chem. 1993, 32, 3501–3508. (c) Aime, S.; Anelli, P. L.; Botta, M.; Fedeli, F.;
Grandi, M.; Paoli, P.; Uggeri, F. Inorg. Chem. 1992, 31, 2422–2428. (d) Platzek, J.;
Blaszkiewicz, P.; Gries, H.; Luger, P.; Michl, G.; Muller-Fahrnow, A.; Raduchel, B.;
Sulzle, D. Inorg. Chem. 1997, 36, 6086–6093. (e) Kumar, K.; Chang, C. A.;
Francesconi, L. C.; Dischino, D. D.; Malley, M. F.; Gougoutas, J. Z.; Tweedle, M. F.
Inorg. Chem. 1994, 33, 3567–3575.
(49) (a) Friedman, S.; Pace, B.; Pizer, R. J. Am. Chem. Soc. 1974, 96,
5381–5384. (b) Sartain, F. K.; Yang, X.; Lowe, C. R. Chem.;Eur. J. 2008, 14,
4060–4067.

4218 Inorganic Chemistry, Vol. 49, No. 9, 2010
Regueiro-Figueroa et al.
those determined previously in a 0.1 M phosphate buffer
using the fluorescent reporter Alizarin Red S (Table 2).50
These results indicate that (i) the spectral changes obser-
ved in the absorption spectrum can be used to investigate
the interaction of PBAs with saccharides and (ii) the
nature of the buffer does not have an important effect
on the binding constants. The Keq values obtained for
[Gd(L2)] are somewhat higher than those obtained for
PBA, with a ca. 2-fold increase being observed for
glucose, galactose, and mannose. In the case of fructose,
a 4-fold increase of the association constant is observed.
The energy difference between the two binding processes
(3.5 kJ mol-1) suggests that this could be due to differ-
3
ences in hydration of the host to which fructose is binding
or to a favorable difference in hydration of the adduct in
the case of the lanthanide(III) complex. The binding
constants obtained for the interaction of [Gd(L2)] with
lactate and citrate are slightly lower than those obtained
for PBA. The affinities of [Gd(L2)] for lactate and citrate
are similar to those observed for glucose and galactose,
respectively.
Study of the Interaction of [Gd(L2)] with Neu5Ac and
MeNeu5Ac. The addition of Neu5Ac to solutions of PBA
or [Gd(L2)] resulted in poor changes in the absorption
spectra. Furthermore, analysis of the titration curves is
complicated by the absorption of saccharide in the UV
region of the spectrum. Thus, we have performed compe-
titive spectrophotometric titrations to investigate the
interaction of Neu5Ac with these systems. In a typical
experiment, a 10-3 M solution of the PBA derivative con-
taining a given amount of Neu5Ac was prepared, and
aliquots of a standard solution containing D-fructose and
Neu5Ac were added. The concentration of Neu5Ac in the
latter solution was identical with that used in the titration
cell to keep its concentration constant during the experi-
ment. The changes in the absorption spectrum as a func-
tion of the equivalents of D-fructose added (Figure 5; see
also Figure S16 in the Supporting Information) were used
to determine the binding constants for the interaction of
PBA and [Gd(L2)] with Neu5Ac because the binding
constants for the interaction with D-fructose are known
(Table 2). The binding constant obtained for the interac-
tion of PBA with Neu5Ac (Keq=19) is in good agreement
with that obtained in a 0.1 M phosphate buffer using the
fluorescent reporter Alizarin Red S (Keq = 21).50 This
result gave us confidence that the competitive assay
provides accurate binding constants for the interaction
of PBAs with Neu5Ac. A comparison of the titration
profiles obtained upon fructose addition in the absence
(Figure 4) and in the presence (Figure 5) of Neu5Ac
clearly shows that in the latter case a larger excess of
fructose is required to reach the plateau in which the
gadolinium(III) complex is fully bound to fructose. This
indicates that Neu5Ac is efficiently competing with fruc-
tose for binding to the gadolinium(III) complex. The bind-
ing constant obtained for the interaction of [Gd(L2)] with
Neu5Ac (Keq=151) is much higher than that obtained for
PBA by using the same method.
Figure 4. Spectrophotometric titrations of 10-3 M solutions of [Gd(L2)]
with a standard solution of D-fructose (top) or D-glucose (bottom) in an
aqueous solution at pH 7.4 (0.1 M MOPS buffer). Insets: changes in the
molar absorbance at 276 nm upon saccharide addition.
Table 2. Binding Constants (Keq Values, M-1) Determined for the Interaction of
[Gd(L2)] and PBA with Different Saccharides and R-Hydroxy Acids in an
Aqueous Solution at 25 °C and pH 7.4 (0.1 M MOPS Buffer)a
[Gd(L2)]
PBAb
PBAc
D-fructose
D-(þ)-glucose
646 ( 14
12.3 ( 0.6
24.5 ( 0.6
21.9 ( 0.5
151 ( 4
158 ( 4
5.3 ( 0.1
15.8 ( 0.4
11.2 ( 0.3
19.1 ( 0.4
d
160
4.6
15
13
21
d
D-galactose
D-mannose
5-acetylneuraminic acid
2-R-O-methyl-5-acetylneuraminic
acid
79 ( 2
methyl R-D-galactoside
methyl R-D-mannoside
lactate
1.48 ( 0.03
1.38 ( 0.03
14.8 ( 0.3
24.0 ( 0.5
d
d
d
d
d
d
25.7 ( 1.1
35.5 ( 0.8
citrate
a The errors given correspond to 1 standard deviation. b This work.
c Reference 50, 0.1 phosphate buffer. d Not determined.
binding constants listed in Table 2. Our results indicate
the following binding affinity trend of these saccharides
to [Gd(L2)]: fructose > galactose ∼ mannose > glucose.
The binding constants obtained for the interaction with
PBA itself under the same conditions follow the same
trend. Furthermore, the binding constants obtained for
PBA by using spectrophotometric titrations (Figures S10-
S15 in the Supporting Information) are very similar to
Neu5Ac is present in solution as an equilibrium be-
tween R-pyranose (5-8%) and β-pyranose forms (92-
95%), while the furanose type of Neu5Ac is absent
because of the acetamide moiety present at the C5 posi-
tion.8 Thus, in solution, the major form of free Neu5Ac is
(50) Springsteen, G.; Wang, B. Tetrahedron 2002, 58, 5291–5300.

Article
Inorganic Chemistry, Vol. 49, No. 9, 2010 4219
Figure 5. Spectrophotometric titration of a 10-3 M solution of [Gd(L2)]
containing Neu5Ac (0.01 M) with a standard solution of D-fructose in an
aqueous solution at pH 7.4 (0.1 M MOPS buffer).
expected to be the β-pyranose form, while all known
glycosides of Neu5Ac are R-linked.9 Therefore, we have
investigated the interaction of [Gd(L2)] with MeNeu5Ac
as a model for neuraminic residues present in the terminal
positions of the carbohydrate chains (Figure S17 in the
Supporting Information). A competitive assay analogous
to that employed for Neu5Ac provided a binding con-
stant for the interaction of [Gd(L2)] with MeNeu5Ac
(Keq=79) somewhat lower than that determined for Neu5Ac
but still substantially larger than that obtained for the
interaction of PBA with Neu5Ac. These results sug-
gest that the sialic acids bind more strongly to [Gd(L2)]
when they adopt the β-pyranose form. The interaction of
[Gd(L2)] with methyl R-D-galactoside and methyl R-D-
mannoside, which are models of units commonly present
in glycan chains, was also investigated by means of spectro-
photometric titrations (Figures S18 and S19 in the Sup-
porting Information). Our results show that [Gd(L2)] pre-
sents a good selectivity for both Neu5Ac and MeNeu5Ac
over these saccharides (Table 2).
It has been shown that the optimal pH for the binding
of boronic acids to diols is often above the pKa value of
the boronic acid species because the binding of cis-diols is
normally stronger when PBA is in the tetrahedral form.32
However, both PBA and [Gd(L2)] possess very similar
pKa values, and therefore the high binding constant
observed for the interaction of [Gd(L2)] with Neu5Ac
and MeNeu5Ac in comparison to PBA cannot be ex-
plained in terms of the respective pKa values of the PBAs
(8.8 and 8.9, respectively). It has been shown that ester
formation with sugars lowers the pKa value of the boron
functionality by about 2-3 pH units and converts the
boron atom from a neutral sp2 to an anionic sp3 moiety.50
Indeed, a UV-vis pH titration of a 10-3 M solution of
[Gd(L2)] containing 0.05 M Neu5Ac allowed us to deter-
mine an apparent pKa of 7.2 (Figure S20 in the Supporting
Information). From these data, together with the bind-
ing constant obtained for the interaction of [Gd(L2)] and
Neu5Ac, we estimate a pKa for the ester form of ∼7.0.
Thus, upon binding to Neu5Ac at pH 7.4, the boron atom
of [Gd(L2)] is expected to be largely in the tetrahedral form.
11B NMR spectroscopy confirms the change of boron
hybridization upon Neu5Ac binding from trigonal (δ ∼
10 ppm) to tetrahedral (δ ∼ -15 ppm) form (see Figure 6).16
Figure 6. 11B NMR spectra of a 10 mM D2O solution of [La(L2)] recor-
ded in the absence and presence of 1 equiv of Neu5Ac.
The emission lifetime of the Tb(5D4) excited-state level
of a solution containing 10-3 M [Tb(L2)] and 0.05 M
Neu5Ac determined at pH 7.4 (0.1 M MOPS) amounts to
1.42 ms. This lifetime is substantially longer than that
determined in the absence of Neu5Ac (1.18 ms), in agree-
ment with the displacement of inner-sphere water mole-
cules due to the binding of the COO- group of Neu5Ac to
the lanthanide ion. It must be pointed out that according
to the binding constant reported in Table 2 under the
experimental conditions used for lifetime measurements
ca. 88% of the terbium(III) complex is expected to be
bound to Neu5Ac. Relaxivity measurements provide
additional evidence for the binding of Neu5Ac through
the carboxylate group. Indeed, the addition of 10 equiv of
Neu5Ac to a 1.62 mM solution of [Gd(L2)] at 25 °C results
in a decrease of relaxivity over the entire range of proton
Larmor frequencies (Figure 2). In the high-field region,
where relaxivity is essentially controlled by the factor
qτR/rH6, relaxivity decreases from ca. 7.0 mM-1 s-1 to
3
ca. 6.5 mM-1 s-1 (20 MHz, 25 °C). Although this
3
relaxivity decrease is relatively limited, even if one con-
siders that under the experimental conditions used only
60% of the gadolinium(III) complex is expected to be
bound to Neu5Ac, it clearly evidences the binding of
Neu5Ac. A similar effect is observed at 37 °C (Figure S4
in the Supporting Information).
The binding of Neu5Ac to [La(L2)] was investigated by
13C NMR spectroscopy. The spectrum obtained for a
solution containing 6 mM Neu5Ac and 18 mM [La(L2)] at
pD 8.6 is compared to that of free Neu5Ac in Figure 7.
Under these conditions, most of Neu5Ac is bound to the
lanthanum(III) complex (>95%) because only very small
peaks due to free Neu5Ac are observed. PBA is known to
bind covalently and reversibly to 1,2- and 1,3-diols to give
five- and six-membered cyclic esters.21 The spectrum
shown in Figure 7 is consistent with the presence of a
single bound species, pointing to a very specific interac-
tion of Neu5Ac with the lanthanum(III) complex. The
resonances due to carbon nuclei of the [La(L2)] complex

4220 Inorganic Chemistry, Vol. 49, No. 9, 2010
Regueiro-Figueroa et al.
formation of a five-membered ester at C7 and C8 is limited
by the unfavorable erythro configuration of the glycerol
tail,53 while the formations of a five-membered cyclic
ester at C8 and C9 and a six-membered cyclic ester at
C7 and C9 are both possible. The very large shift observed
for C8 suggests that the binding to the glycerol tail occurs
through C8 and C9. The 1H NMR spectral data further
support a binding at positions 8 and 9 because the signal
due to H8 experiences an important upfield shift upon
binding (0.45 ppm), while the chemical shift of H7
remains nearly unaffected. The H3-H5 1H NMR signals
of Neu5Ac also undergo important downfield shifts upon
binding to [La(L2)] (0.19-0.78 ppm). Thus, the lantha-
nide(III) complexes of L2 appear to bind to Neu5Ac
through a cooperative two-site binding through (i) ester
formation by the interaction on the PBA function of the
receptor and (ii) coordination of the carboxylate function
of Neu5Ac to the LnIII ion.
Figure 7. 13C NMR spectra of Neu5Ac (6 mM) and Neu5Ac (6 mM) þ
[La(L2)] (18 mM) in a D2O solution (pD = 8.6, 25 °C). The broad reso-
nancesobservedin the range of50-70 ppm are due to carbonnucleiof the
[La(L2)] complex.
In order to rationalize the selectivity that [Gd(L2)]
shows for Neu5Ac, we have characterized the
Table 3. 1H and 13C NMR Chemical Shifts (ppm) for Neu5Ac and Its Complexed
[Gd(L2OH) Neu5Ac]2- system by means of DFT calcula-
3
Form with [La(L2)] in D2O at pD 8.6a
tions in the gas phase (B3LYP model). Different compu-
tational studies have shown that this computational
approach provides molecular geometries in good agree-
ment with the solution structures of the complexes.41 The
main difference between the gas-phase geometries and the
actual solution structures is the Ln-N distances, which
are overestimated with calculations performed in the gas
free
complex
free
complex
C1
C2
C3
C4
C5
C6
C7
C8
C9
C10
C11
178.3
98.0
41.0
68.9
53.9
71.8
71.4
71.9
64.9
176.3
23.7
170.8
b
H3eq
H3ax
H4
H5
H6
2.16
1.76
3.97
3.85
3.92
3.46
3.70
3.56
3.79
2.00
2.94
2.79
4.16
4.31
3.83
3.47
3.34
3.72
3.79
2.04
42.1
74.4
54.4
74.7
68.4
81.8
62.1
176.8
23.5
54
H7
H8
˚
phase by ca. 0.1 A. For the reasons given above, our
calculations were performed for both the R-pyranose and
β-pyranose forms of Neu5Ac. PBA is known to bind
covalently and reversibly to 1,2- and 1,3-diols to give five-
and six-membered cyclic esters.16 It has been previously
shown that the formation of a five-membered ester at C7
and C8 is limited by the unfavorable erythro configura-
tion of the glycerol tail.53 Thus, we have only considered
in our calculations the five-membered cyclic ester formed
at C8 and C9 and the six-membered cyclic ester formed at
C7 and C9 (Chart 1). Geometry optimizations provided
up to eight energy minima that correspond to combina-
H9
H90
H11
a See Chart 1 for the labeling scheme. b Not observed.
are extremely broad. Broad NMR signals are often obser-
ved for N-alkylated LnIII(DO3A) complexes because of
exchange phenomena.51 However, the 13C NMR signals
due to the bound form of Neu5Ac could be assigned with
the aid of 2D COSY, HSQC, and HMBC experiments.
The chemical shifts observed for the bound form of Neu5Ac
are compared to those of free Neu5Ac in Table 3. The
assignments of the NMR signals of free Neu5Ac are in
agreement with those reported in the literature.52 The
interaction with [La(L2)] causes very important shifts of
the 13C NMR signals of Neu5Ac. In particular, the signal
due to the carboxylate group of Neu5Ac (C1) shifts upfield
by 7.6 ppm, while the signal due to C10 is little affected.
These results are consistent with the binding of the carboxy-
late function of Neu5Ac to the lanthanide. The carbon
nuclei of the glycerol tail of Neu5Ac also experience
important shifts upon binding to [La(L2)]. In particular,
a downfield shift of 9.9 ppm is observed for C8, while C7
and C9 undergo upfield shifts of 1.8 and 2.7 ppm,
respectively. It has been previously shown that the
tions of the SAP and TSAP conformations of the GdIII
-
(DO3A) cage forming esters with the R-pyranose and
β-pyranose forms of Neu5Ac at the C8-C9 and C7-C9
positions. The relative free energies of the different con-
formations obtained as a result of geometry optimizati-
ons are given in Table 4. Our results show that the adduct
formed at C8 and C9 is more stable than that formed at
C7 and C9 for both the R-pyranose and β-pyranose forms
of Neu5Ac, in nice agreement with the NMR data. The
adducts formed by the SAP form of the complex are more
stable than those formed by the TSAP one. Furthermore,
our results suggest that the binding to the β-pyranose
form of Neu5Ac is slightly more favorable than that to the
R-pyranose conformation. The two lowest-energy con-
formations calculated for the [Gd(L2OH) Neu5Ac]2-
3
system are shown in Figure 8. The carboxylate unit of
Neu5Ac is predicted to bind to the GdIII ion in a slightly
asymmetrical bidentate fashion, with the GdIII-ONeu5Ac
(51) Lowe, M. P.; Parker, D.; Reany, O.; Aime, S.; Botta, M.; Castellano,
G.; Gianolio, E.; Pagliarin, R. J. Am. Chem. Soc. 2001, 123, 7601–7609.
(52) (a) Czarniecki, M. F.; Thornton, E. R. J. Am. Chem. Soc. 1977, 99,
8273–8279. (b) Gervay, J.; Batta, G. Tetrahedron Lett. 1994, 35, 3009–3012.
(c) Klepach, T.; Zhang, W.; Carmichael, I.; Serianni, A. S. J. Org. Chem. 2008,
73, 4376–4387.
(54) (a) Cosentino, U.; Villa, A.; Pitea, D.; Moro, G.; Barone, V.;
Maiocchi, A. J. Am. Chem. Soc. 2002, 124, 4901–4909. (b) Platas-Iglesias,
C.; Mato-Iglesias, M.; Djanashvili, K.; Muller, R. N.; Vander Elst, L.; Peters, J. A.;
de Blas, A.; Rodriguez-Blas, T. Chem.;Eur. J. 2004, 10, 3579–3590.
(53) Chai, W.; Yuen, C.-T.; Feizi, T.; Lawson, A. M. Anal. Biochem. 1999,
270, 314–322.

Article
Inorganic Chemistry, Vol. 49, No. 9, 2010 4221
Table 4. Relative Free Energies (kJ mol-1) of the Different Conformations
two-site binding through (i) ester formation by the interac-
tion on the PBA function of the receptor and (2) coordination
of the carboxylate function of Neu5Ac to the LnIII ion.
Compared with simple PBA, complex [Gd(L2)] represents
an improvement of about 8-fold inaffinity and about 3.4-fold
in selectivity for Neu5Ac over glucose, the most common
saccharide present in vivo. Furthermore, [Gd(L2)] also shows
a good selectivity for Neu5Ac over lactate and citrate, which
occur in relatively high concentration in blood. On the
contrary, PBA forms stronger complexes with both lactate
and citrate than with Neu5Ac.
3
Obtained for the [Gd(L2OH) Neu5Ac]2- System by Means of DFT Calculations
3
(B3LYP Model) Compared to the Complex with the Lowest Energya
R-Neu5Ac
β-Neu5Ac
SAP(C8-C9)
SAP(C7-C9)
TSAP(C8-C9)
TSAP(C7-C9)
5.8
77.9
9.7
0.0
42.0
7.9
70.7
33.7
a SAP and TSAP indicate square-antiprismatic and twisted square-
antiprismatic geometries, respectively; relative energies of esters formed
at both the C8-C9 and C7-C9 positions are given.
The binding affinity for MeNeu5Ac, a model for sialic acid
residues on cell surfaces, is somewhat lower than that for
free Neu5Ac. However, a much weaker binding affinity was
observed with methyl R-D-galactoside and methyl R-D-man-
noside, which are models of units commonly present in
glycan chains. Thus, this receptor shows promise for the
preparation of new sensors for sialic acid recognition on cell
surfaces and radiopharmaceuticals for the diagnosis and
therapy of tumors. Furthermore, we believe that the general
approach reported here can be used to design more effective
receptors. For instance, lowering the pKa of the PBA function
should increase the binding affinity of the receptor to the
glycerol tail of sialic acids, while increasing the positive
charge of the lanthanide chelate will strengthen the binding
to the carboxylate group of Neu5Ac. By contrast to the com-
plexes of L2, the L1 analogues do not bind to sugars because
of an unprecedented coordination of the phenylboronate
function of the ligand to the LnIII ion.
Experimental Section
Physical Methods. Elemental analyses were carried out on a
ThermoQuest Flash 1112 elemental analyzer. Electrospray ioni-
zation time-of-flight (ESI-TOF) mass spectra were recorded
using a LC-Q-q-TOF Applied Biosystems QSTAR Elite spec-
trometer in both positive and negative modes. Fast atom bom-
bardment (FAB) mass spectra were recorded using a FISONS
QUATRO mass spectrometer with a Cs-ion gun and 3-nitro-
benzyl alcohol as the matrix. IR spectra were recorded using a
Bruker Vector 22 spectrophotometer equipped with a Golden
Gate attenuated total reflectance (ATR) accessory (Specac). 1H
and 13C NMR spectra were run on a Bruker Avance 500 spec-
trometer. Chemical shifts are reported in δ values. 11B NMR
spectra was measured on a Bruker Avance 400 spectrometer
using a solid-state probe and 4 mm zirconia rotors, spinning at
the magic angle with a rate of 20 Hz. Peak positions were
measured with respect to the resonance of a 0.1 M solution of
boric acid (0 ppm). Electronic spectra in the UV-vis range were
recorded at 25 °C on a Perkin-Elmer Lambda 900 UV-vis
spectrophotometer using 1.0 cm quartz cells. Sugar binding
studies were performed on 10-3 M solutions of PBA or
[Gd(L2)] (10 mL) at pH 7.4 (MOPS buffer). Typically, aliquots
of a fresh standard solution of the envisaged saccharide (ca. 0.5 M;
containing 10-3 M PBA or [Gd(L2)] to avoid dilution) were
added, and the UV-vis spectra of the samples were recorded.
Glucose solutions were allowed to stand at least 24 h to allow
them to reach anomeric equilibrium. Competitive UV-vis
titrations were performed in a similar way by using 0.05 or
0.01 M Neu5Ac in the titration cell for PBA and [Gd(L2)],
respectively. An analogous competitive assay was used to
investigate the binding of MeNeu5Ac to [Gd(L2)]. In order to
keep the Neu5Ac concentration constant during the titration
aliquots of a standard solution containing D-fructose (0.5 M),
Neu5Ac (0.05 or 0.01 M) and the PBA derivative (10-3 M) were
added, and the UV-vis spectra of the samples were recorded.
All spectrophotometric titration curves were fitted with the
Figure 8. Optimized structures of the [Gd(L2OH) Neu5Ac]2- system
obtained from DFT calculations at the B3LYP/6-31G(d) level. Top:
R-pyranose form of Neu5Ac. Bottom: β-pyranose form of Neu5Ac.
Hydrogen atoms are omitted for simplicity.
3
˚
distances amounting to 2.49 and 2.53 A (R-pyranose) and
˚
2.53 and 2.56 A (β-pyranose). Thus, our DFT calculations
confirm that a two-site cooperative binding of Neu5Ac to
[Gd(L2)] is possible for both the R- and β-pyranose
conformations.
Conclusions
In conclusion, we have described a straightforward app-
roach to preparing the first example of a new family of lantha-
nide complexes for sialic acid recognition due to a cooperative

4222 Inorganic Chemistry, Vol. 49, No. 9, 2010
Regueiro-Figueroa et al.
1
HYPERQUAD program.55 Binding constants were obtained
from a simultaneous fit of the UV-vis absorption spectral
changes at 6-8 selected wavelengths. A minimum of 25 absor-
bance data points at each of these wavelengths was used. Exci-
tation and emission spectra were recorded on a Perkin-Elmer
LS-50B spectrometer. Luminescence lifetimes were calculated
from the monoexponential fitting of the average decay data, and
they are averages of at least 3-5 independent determinations.
The 1/T1 NMRD profiles of the gadolinium(III) complexes
were recorded on a Stelar SMARtracer FFC fast-field-cycling
relaxometer covering magnetic fields from 2.35 ꢀ 10-4 to 0.25 T,
which correspond to a proton Larmor frequency range of 0.01-
10 MHz. The relaxivity at higher fields was recorded using a
Bruker WP80 adapted to variable field measurements and con-
trolled by the SMARtracer PC-NMR console. The temperature
was controlled by a VTC90 temperature control unit and fixed
by a gas flow. The temperature was determined according to
previous calibration with a platinum resistance temperature probe.
Computational Methods. All calculations were performed by
employing hybrid DFT with the B3LYP exchange-correlation
functional,56,57 and the Gaussian 03 package (revision C.01).58
Relativistic effects were considered through the use of relativis-
tic ECPs (RECPs). Different computational studies on lantha-
nide(III) complexes have shown that the 4f orbitals do not parti-
cipate in bonding because of their contraction into the core.59
Thus, full geometry optimizations of the [Gd(L1OH)]-, [Gd(L2)-
1597 (CdC), 1325 (C-B). H NMR (CDCl3, 500 MHz, 25 °C,
TMS): δ 7.86 (d, 1H, ArH), 7.45 (m, 2H, ArH), 7.31 (m, 1H, ArH),
4.92 (s, 2H, CH2Br), 4.42 (s, 4H, OCH2). 13C NMR (CDCl3, 125.8
MHz, 25 °C, TMS): δ 144.3, 136.7, 131.6, 130.3, 127.7, 66.0, 33.8.
2-[2-(Bromomethyl)phenyl]-5,5-dimethyl-1,3-dioxa-2-borinane (4).
This compound was prepared by using a slight modification of
the literature procedure.27 A mixture of 3 (1.22 g, 5.98 mmol),
NBS (1.12 g, 6.28 mmol), and PDB (0.090 g, 0.370 mmol) in
CCl4 (20 mL) was refluxed for 16 h. The reaction mixture was
allowed to cool to room temperature, the precipitated succini-
mide was removed by filtration, and the filtrate was concen-
trated under reduced pressure. The crude product was purified
by column chromatography on SiO2 with CH2Cl2 as the eluent
to give 1.39 g of 4 as a yellow oil. Yield: 82%. Elem anal. Calcd
for C12H16BBrO2: C, 50.93; H, 5.70. Found: C, 50.47; H, 5.57.
MS (FAB): m/z(% BPI) 203(100) ([C12H16BO2]þ), 283(10)
([C12H17BBrO2]þ). IR (ATR, cm-1): ν 2959, 2935, 2871 (C-H),
1599 (CdC), 1321 (C-B). 1H NMR (CDCl3, 500 MHz, 25 °C,
TMS): δ 7.81 (d, 1H, ArH), 7.37 (m, 1H, ArH), 7.28 (m, 2H,
ArH), 4.93 (s, 2H, CH2Br), 3.82 (s, 4H, OCH2), 1.07 (s, 6H,
-CH3). 13C NMR (CDCl3, 125.8 MHz, 25 °C, TMS): δ 143.5,
135.5, 130.5, 130.2, 127.5, 72.4, 34.5, 31.8, 21.9.
1,4,7-Tris[(tert-butoxycarbonyl)methyl]-10-[2-(2-methylphenyl)-
1,3-dioxa-2-borolanyl]-1,4,7,10-tetraazacyclododecane (5). Com-
pound 2 (0.140 g, 0.576 mmol) andNa2CO3 (0.250 g, 2.359mmol)
were added to a solution of DO3A(t-BuO)3 (0.300 g, 0.583 mmol)
in acetonitrile (25 mL). The mixture was heated to reflux with
stirring under an inert atmosphere (Ar) for a period of 24 h, and
then the excess of Na2CO3 was filtered off and the filtrate
concentrated to dryness. The residue was partitioned between
H2O and CH3Cl (20:10), and the organic phase was separated,
dried over MgSO4, filtered, and evaporated to dryness. The crude
product was purified by columnchromatography onAl2O3 witha
CH2Cl2/MeOH 5% mixture as the eluent to give 0.310 g of 5 as a
(H2O)], [Gd(L2)(H2O)2], and [Gd(L2OH) Neu5Ac]2- systems
3
were performed in vacuo by using the RECP of Dolg et al. and
the related [5s4p3d]-GTO valence basis set for the gadolinium
atom42 and the 6-31G(d) basis set for carbon, hydrogen, boron,
nitrogen, and oxygen atoms. The RECP of Dolg et al. includes
46 þ 4f7 electrons in the core for the gadolinium atom, leaving
the outermost 11 electrons to be treated explicitly, in line with
the nonparticipation of 4f electrons in bonding. Thus, this RECP
treats [Kr]4d104f7 as a fixed core, while only the 5s25p66s25d16p0
shell is explicitly taken into account. The stationary points
found on the potential energy surfaces as a result of the geo-
metry optimizations on the [Gd(L1OH)]-, [Gd(L2)(H2O)], [Gd-
brown oil. Yield: 63%. Elem anal. Calcd for C35H59BN4O8
3
2CH2Cl2: C, 52.62; H, 7.52; N, 6.63. Found: C, 51.80; H, 7.73; N,
6.79. MS (ESIþ): m/z 649 ([C33H58BN4O8]þ). IR (ATR, cm-1):
ν2976, 2931, 2846 (C-H), 1724 (CdO), 1597 (CdC), 1366 (C-B).
1H NMR (CDCl3, 500 MHz, 25 °C, TMS): δ 7.73 (d, 1H, ArH),
7.58 (d, 1H, ArH), 7.21 (m, 1H, ArH), 7.13 (m, 1H, ArH), 4.25 (s,
4H, OCH2), 3.71-2.05 (m, 24H, -NCH2-), 1.33 (s, 27H,
-OtBu). 13C NMR (CDCl3, 125.8 MHz, 25 °C, TMS): δ 172.8,
172.5, 171.9, 143.6, 135.9, 131.0, 129.2, 126.1, 116.2, 82.2, 82.0,
81.9, 77.2, 65.4, 56.9, 56.1, 55.3, 55.2, 55.1, 49.6, 27.5, 27.4.
(3-Methylphenyl)ethyleneboronate (6). A mixture of (3-methyl-
phenyl)boronic acid (1.00 g, 7.36 mmol) and ethylene glycol (0.59 g,
9.56 mmol) in dry toluene (25 mL) was refluxed for 20 h using a
Dean-Stark apparatus for water removal. The reaction mixture
was concentrated at reduced pressure, and the product was puri-
fied by column chromatography on SiO2 with CH2Cl2 as the
eluent to give 1.01 g of 6 as a colorless oil. Yield: 85%. Elem anal.
Calcd for C9H11BO2: C, 66.73; H, 6.84. Found: C, 66.56; H, 6.81.
MS (FAB): m/z 162 (100%) ([C9H11BO2]þ). IR (ATR, cm-1):
ν 3024, 2975, 2908 (C-H), 1607 (CdC), 1336 (C-B). 1H NMR
(CDCl3, 500 MHz, 25 °C, TMS): δ 7.65 (m, 2H, ArH), 7.31 (m,
2H, ArH), 4.39 (s, 4H, OCH2), 2.39 (s, 3H, CH3). 13C NMR
(CDCl3, 125.8 MHz, 25 °C, TMS): δ 137.2, 135.5, 132.2, 131.8,
127.8, 66.0, 21.3.
(L2)(H2O)2], and [Gd(L2OH) Neu5Ac]2- systems have been
3
tested to represent energy minima rather than saddle points
via frequency analysis. The relative free energies of the different
conformations calculated for these systems include non-poten-
tial-energy contributions (that is, zero point energy and thermal
terms) obtained by frequency analysis.
Chemicals and Starting Materials. (2-Methylphenyl)ethylene-
boronate (1)29 and 5,5-dimethyl-2-o-tolyl-1,3-dioxa-2-borinane
(3)27 were prepared according to literature procedures. All other
chemicals were purchased from commercial sources and used
without further purification. Solvents were of reagent grade,
purified by the usual methods.
[2-(Bromomethyl)phenyl]ethyleneboronate (2). This compound
was prepared by using a slight modification of the literature
procedure.29 A mixture of 1 (0.96 g, 5.93 mmol), NBS (1.11 g,
6.22 mmol), and PDB (0.090 g, 0.370 mmol) in CCl4 (20 mL) was
refluxed for 16 h. The reaction mixture was allowed to cool to room
temperature, the precipitated succinimide was removed by filtra-
tion, and the filtrate was concentrated under reduced pressure. The
crude product was purified by column chromatography on SiO2
with CH2Cl2 as the eluent to give 0.570 g of 2 as a yellow
hygroscopic solid. Yield: 40%. Elem anal. Calcd for C9H10BBrO2:
C, 44.87; H, 4.18. Found: C, 45.01; H, 4.25. MS (ESI-): m/z 239
([C9H9BBrO2]-). IR (ATR, cm-1): ν 3064, 2987, 2913 (C-H),
[3-(Bromomethyl)phenyl]ethyleneboronate (7). A mixture of 6
(1.01 g, 6.23 mmol), NBS (1.17 g, 6.55 mmol), and PDB (0.095 g,
0.380 mmol) in CCl4 (20 mL) was refluxed for 16 h. The reaction
mixture was allowed to cool to room temperature, the precipi-
tated succinimide was removed by filtration, and the filtrate was
concentrated under reduced pressure. The crude product was
purified by column chromatography on SiO2 with CH2Cl2 as the
eluent to give 0.81 g of 7 as a yellow oil. Yield: 54%. Elem anal.
Calcd for C9H10BBrO2: C, 44.87; H, 4.18. Found: C, 45.13; H,
4.21. MS (ESIþ): m/z 161 ([C9H10BO2]þ). IR (ATR, cm-1):
ν 3025, 2976, 2908 (C-H), 1605 (CdC), 1338 (C-B). 1H NMR
(55) Gans, P.; Sabatini, A.; Vacca, A. Talanta 1996, 43, 1739–1753.
(56) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(57) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789.
(58) Frisch, M. J.; et al. Gaussian03, revision C.01; Gaussian, Inc.: Wallingford,
CT, 2004.
(59) (a) Maron, L.; Eisenstein, O. J. Phys. Chem. A 2000, 104, 7140–7143.
(b) Boehme, C.; Coupez, B.; Wipff, G. J. Phys. Chem. A 2002, 106, 6487–6498.

Article
Inorganic Chemistry, Vol. 49, No. 9, 2010 4223
(CDCl3, 500 MHz, 25 °C, TMS): δ 7.85 (s, 1H, ArH), 7.61 (d,
1H, ArH), 7.52 (d, 1H, ArH), 7.38 (m, 1H, ArH), 4.51 (s, 2H,
CH2Br), 4.39 (s, 4H, OCH2). 13C NMR (CDCl3, 125.8 MHz,
25 °C, TMS): δ 136.2, 135.4, 134.9, 132.1, 128.4, 66.1, 33.4.
1,4,7-Tris[(tert-butoxycarbonyl)methyl]-10-[2-(3-methylphenyl)-
1,3-dioxa-2-borolanyl]-1,4,7,10-tetraazacyclododecane (8). Com-
pound 7 (0.140g, 0.576mmol) andNa2CO3 (0.250 g, 2.359 mmol)
were added to a solution of DO3A(t-BuO)3 (0.300 g, 0.583 mmol)
in acetonitrile (25 mL). The mixture was heated to reflux with
stirring under an inert atmosphere (Ar) for a period of 24 h, and
then the excess of Na2CO3 was filtered off and the filtrate
concentrated to dryness. The residue was partitioned between
H2O and CH3Cl (20:10), and the organic phase was separated,
dried over MgSO4, filtered, and evaporated to dryness. The crude
productwas purifiedbycolumnchromatography onAl2O3 witha
CH2Cl2/MeOH 5% mixture as the eluent to give 0.330 g of 8 as a
temperature for 24 h. The solid was isolated by filtration,
washed with acetonitrile and diethyl ether, and dried under
vacuum.
[La(L1)] 2H2O. Yield: 0.063 g, 79%. Elem anal. Calcd for
3
C21H34BLaN4O10: C, 38.67; H, 5.25; N, 8.59. Found: C, 38.29;
H, 4.83; N, 8.25. MS (ESIþ): m/z 617 [C21H31BLaN4O8]þ. IR
(ATR, cm-1): ν 1581 (CdO), 1395 (C-B), 1254 (B-O).
[Eu(L1)] 2H2O. Yield: 0.058 g, 72%. Elem anal. Calcd for
3
C21H34BEuN4O10: C, 37.91; H, 5.15; N, 8.42. Found: C, 37.73;
H, 4.73; N, 8.25. MS (ESIþ): m/z 631 [C21H31BEuN4O8]þ. IR
(ATR, cm-1): ν 1587 (CdO), 1378 (C-B), 1253 (B-O).
[Gd(L1)] 2H2O. Yield: 0.049 g, 60%. Elem anal. Calcd for
3
C21H34BGdN4O10: C, 37.61; H, 5.11; N, 8.36. Found: C, 37.22;
H, 4.82; N, 8.50. MS (ESIþ): m/z 636 [C21H31BGdN4O8]þ. IR
(ATR, cm-1): ν 1576 (CdO), 1385 (C-B), 1247 (B-O).
[Tb(L1)] 2H2O. Yield: 0.053 g, 65%. Elem anal. Calcd for
3
brown oil. Yield: 67%. Elem anal. Calcd for C35H59BN4O8
3
C21H34BN4O10Tb: C, 37.52; H, 5.10; N, 8.33. Found: C, 37.03;
H, 4.77; N, 8.27. MS (ESIþ): m/z 637 [C21H31BN4O8Tb]þ. IR
(ATR, cm-1): ν 1583 (CdO), 1377 (C-B), 1246 (B-O).
2CH2Cl2: C, 52.62; H, 7.52; N, 6.63. Found: C, 52.90; H, 7.75; N,
6.78. MS (ESIþ): m/z 649 ([C33H58BN4O8]þ). IR (ATR, cm-1):
ν 2970, 2929, 2849 (C-H), 1720 (CdO), 1601 (CdC), 1309
[La(L2)] 4H2O. Yield: 0.045 g, 54%. Elem anal. Calcd for
1
3
(C-B). H NMR (CDCl3, 500 MHz, 25 °C, TMS): δ 7.52 (s,
C21H38BLaN4O12: C, 36.65; H, 5.56; N, 8.14. Found: C, 36.43;
H, 5.46; N, 8.12. MS (ESIþ): m/z 617 [C21H31BLaN4O8]þ, 639
[C21H30BLaN4NaO8]þ. IR (ATR, cm-1): ν 1575 (CdO), 1381
(C-B), 1252 (B-O).
1H, ArH), 7.30 (d, 1H, ArH), 7.13 (d, 1H, ArH), 6.90 (m, 1H,
ArH), 3.32 (s, 4H, OCH2), 2.63-1.79 (m, 24H, -NCH2-), 1.08
(s, 27H, -OtBu). 13C NMR (CDCl3, 125.8 MHz, 25 °C, TMS): δ
172.1, 171.3, 135.1, 134.7, 132.0, 131.2, 126.6, 115.6, 81.4, 81.1,
77.2, 70.9, 58.2, 54.6, 54.3, 49.2, 48.4, 30.6, 26.7, 26.6.
[Eu(L2)] 4H2O. Yield: 0.049 g, 57%. Elem anal. Calcd for
3
C21H38BEuN4O12: C, 35.96; H, 5.46; N, 7.99. Found: C, 35.88;
H, 4.94; N, 8.22. MS (ESIþ): m/z 631 [C21H31BEuN4O8]þ. IR
(ATR, cm-1): ν 1586 (CdO), 1376 (C-B), 1245 (B-O).
1,4,7-Tris(carboxymethyl)-10-[2-(dihydroxyboranyl)benzyl]-
1,4,7,10-tetraazacyclododecane Trifluoroacetate Salt (H3L1 3CF3-
3
COOH). A solution of compound 5 (0.310 g, 0.368 mmol) in a
1:1 CH2Cl2/TFA mixture (5 mL) was stirred at room tempera-
ture for 24 h. The mixture was concentrated to dryness, and the
resulting oil was dissolved in methanol (∼1 mL). The addition of
diethyl ether resulted in the formation of a white solid, which
was collected by filtration and dried under vacuum. Yield: 0.230 g,
[Gd(L2)] 4H2O. Yield: 0.054 g, 64%. Elem anal. Calcd for
3
C21H38BGdN4O12: C, 35.70; H, 5.42; N, 7.93. Found: C, 35.32;
H, 5.19; N, 7.68. MS (ESIþ): m/z 636 [C21H31BGdN4O8]þ. IR
(ATR, cm-1): ν 1587 (CdO), 1378 (C-B), 1248 (B-O).
[Tb(L2)] 4H2O. Yield: 0.047 g, 55%. Elem anal. Calcd for
3
C21H38BN4O12Tb: C, 35.61; H, 5.41; N, 7.91. Found: C, 35.13;
H, 4.97; N, 7.81. MS (ESIþ): m/z 637 [C21H31BN4O8Tb]þ, 659
[C21H30BN4NaO8Tb]þ. IR (ATR, cm-1): ν 1589 (CdO), 1378
(C-B), 1247 (B-O).
75%. Elem anal. Calcd for C21H33BN4O8 3CF3COOH: C,
3
39.43; H, 4.41; N, 6.81. Found: C, 39.04; H, 4.56; N, 7.15. MS
(ESI-): m/z 461 ([C21H30BN4O7]-). IR (ATR, cm-1): ν 3085
(C-H), 1726, 1659 (CdO), 1353 (C-B). 1H NMR (D2O, pD=
0.6, 500 MHz, 25 °C, tBuOH): δ 7.80 (m, 1H, ArH), 7.54-7.46
(m, 3H, ArH), 4.62-2.97 (m, 24H, CH2).
Acknowledgment. The authors thank Ministerio de Edu-
ꢀ
cacion y Ciencia and FEDER (Grant CTQ2006-07875/PPQ)
1,4,7-Tris(carboxymethyl)-10-[3-(dihydroxyboranyl)benzyl]-
1,4,7,10-tetraazacyclododecane Trifluoroacetate Salt (H3L2 3CF3-
and Xunta de Galicia (Grant INCITE08ENA103005ES) for
financial support. This research was performed in the frame-
work of the EU COST Action D38 “Metal-Based Systems for
Molecular Imaging Applications”. The authors are indebted to
3
COOH). A solution of compound 8 (0.330 g, 0.391 mmol) in a
1:1 CH2Cl2/TFA mixture (5 mL) was stirred at room tempera-
ture for 24 h. The mixture was concentrated to dryness, and the
resulting oil was dissolved in methanol (∼1 mL). The addition of
diethyl ether resulted in the formation of a white solid that was
collected by filtration and dried under vacuum. Yield: 0.265 g,
ꢀ
Centro de Supercomputacion de Galicia (CESGA) for providing
the computer facilities. M.R.-F. thanks Ministerio de Educacion
ꢀ
y Ciencia (FPU program) for a predoctoral fellowship. The
authors gratefully acknowledge Dr. Joop A. Peters for fruitful
scientific discussions.
82%. Elem anal. Calcd for C21H33BN4O8 3CF3COOH: C,
3
39.43; H, 4.41; N, 6.81. Found: C, 39.13; H, 4.48; N, 7.20. MS
(ESI-): m/z 461 ([C21H30BN4O7]-). IR (ATR, cm-1): ν 3092
(C-H), 1722, 1661 (CdO), 1357 (C-B). 1H NMR (D2O, pD=
0.6, 500 MHz, 25 °C, tBuOH): δ 7.83 (m, 2H, ArH), 7.62 (d, 1H,
ArH), 7.48 (m, 2H, ArH), 4.48-2.82 (m, 24H, CH2).
Supporting Information Available: UV-vis spectra of PBA as
a function of the pH, 11B NMR spectra of [La(L2)] as a function
of the pH, excitation and emission spectra of [Tb(L1)] and [Tb(L2)],
spectrophotometric titrations of [Gd(L2)] with D-galactose and
D-mannose, spectrophotometric titrations of PBA with different
saccharides, spectrophotometric titrations of [Gd(L2)] and Me-
Neu5Ac with D-fructose, spectrophotometric titrations of
[Gd(L2)] with methyl R-D-galactoside and methyl R-D-manno-
side, UV-vis pH titrations of [Gd(L2)] in the presence of Neu5Ac,
Cartesian coordinates of the [Gd(L1OH)]-, [Gd(L2)(H2O)],
General Procedure for Preparation of the [Ln(L1)] 2H2O and
3
[Ln(L2)] 4H2O Complexes. A mixture of H3L1 3CF3COOH or
3
3
H3L2 3CF3COOH (0.100 g, 0.122 mmol), triethylamine (0.074 g,
3
0.732 mmol), and Ln(OTf)3 (0.122 mmol, Ln= La, Eu, Gd, or
Tb) in 2-propanol (10 mL) was heated to reflux for 24 h. The
reaction was allowed to cool to room temperature and then
concentrated to dryness. The addition of 5 mL of THF (L1) or
acetonitrile (L2) resulted in the formation of a white precipitate,
which was isolated by filtration. The solid was then suspended in
10 mL of THF (L1) or acetonitrile (L2) and stirred at room
[Gd(L2)(H2O)2], and [Gd(L2OH) Neu5Ac]2- systems opti-
3
mized in vacuo (B3LYP), and complete ref 51. This material is
available free of charge via the Internet at http://pubs.acs.org.