A carborane-derivative "click" reaction under heterogeneous conditions for the synthesis of a promising lipophilic MRI/GdBNCT agent

Article abstract of DOI:10.1002/chem.201201634

In this study, the Huisgen reaction has been used to functionalise a carborane cage with a lipophilic moiety and a 1,4,7,10-tetraazacyclododecane-1, 4,7,10-tetraacetic acid (DOTA) ligand to obtain a new Gd boron neutron-capture therapy (BNCT)/magnetic res

Full text of DOI:10.1002/chem.201201634

FULL PAPER
DOI: 10.1002/chem.201201634
A Carborane-Derivative “Click” Reaction under Heterogeneous Conditions
for the Synthesis of a Promising Lipophilic MRI/GdBNCT Agent**
Antonio Toppino,[a] Maria Elena Bova,[a] Simonetta Geninatti Crich,[a] Diego Alberti,[a]
Eliano Diana,[a] Alessandro Barge,[b] Silvio Aime,[a] Paolo Venturello,[a] and
Annamaria Deagostino*[a]
Abstract: In this study, the Huisgen re-
action has been used to functionalise a
carborane cage with a lipophilic moiety
and a 1,4,7,10-tetraazacyclododecane-
supported ionic-liquid catalyst, and ho-
mogeneous conditions. The ability of
the Gd complex of the synthesised
ligand to form stable adducts with low-
density lipoproteins (LDLs) has been
evaluated and then MRI has been per-
formed on tumour melanoma cells in-
cubated in the presence of a Gd-com-
plex/LDL imaging probe. It has been
concluded that the high amount of in-
tracellular boron necessary to perform
BNCT can be reached even in the pres-
ence of a relatively low-boron-contain-
ing LDL concentration.
1,4,7,10-tetraacetic
acid
(DOTA)
ligand to obtain a new Gd boron neu-
tron-capture therapy (BNCT)/magnetic
resonance imaging (MRI) agent. The
introduction of the triazole units has
been accomplished under both hetero-
geneous conditions, by the use of a Cu-
Keywords: carboranes · click chem-
istry · cycloaddition · heterogeneous
catalysis
imaging
·
magnetic resonance
Introduction
agents, it is easy to consider polynuclear boron derivatives
as potential candidates for BNCT applications. In fact, sever-
al readily functionalised carboranes have been employed to
construct boron delivery vehicles for BNCT, because of
their high content of boron and their stability in vivo.[4–7]
Click chemistry is a modular approach to organic synthe-
sis that uses the most practical and reliable chemical trans-
formations to construct target molecules.[8–10] In particular,
the CuI-catalysed formation of the 1,2,3-triazolic ring, start-
ing from azides and terminal alkynes, represents a powerful
linking reaction owing to its high degree of dependability,
complete specificity and biocompatibility of the reactants.
Moreover, the triazolic moiety does not represent an inert
linker and plays an active role since it readily associates
with biological targets through hydrogen-bonding and
dipole interactions. In recent years, there has been a signifi-
cant interest in developing click reactions that do not re-
quire any metal catalyst yet they exhibit all the beneficial
properties of the copper-catalysed azide–alkyne click reac-
tion.[11]
Boron neutron-capture therapy (BNCT) is a binary radia-
tion therapy for the treatment of cancer that is based on the
capture of thermal neutrons by 10B nuclei that have been se-
lectively delivered to tumour cells. The neutron-capture
event results in the formation of excited 11B nuclei that un-
dergo fission to yield highly energetic He2+ and Li3+ ions.
Cell death is triggered by the release of these charged parti-
cles, which create ionisation tracks along their trajectories,
thereby resulting in cellular damage. It has been estimated
that approximately 10–30 mg of boron per gram of tumour
mass is needed to attain an acceptable therapeutic advant-
age.[1–3] Thus an important aspect relies on the possibility of
delivering high payloads of 10B at the target sites. Although
clinical exploitation of the BNCT strategy is currently being
carried out with lower-molecular-weight boron delivery
4
7
[a] Dr. A. Toppino, M. E. Bova, Dr. S. Geninatti Crich, Dr. D. Alberti,
Prof. E. Diana, Prof. S. Aime, Prof. P. Venturello, Dr. A. Deagostino
Dipartimento di Chimica
The synthesis of a new dual agent for applications in mag-
netic resonance imaging (MRI)/BNCT has been recently de-
veloped in our laboratory, whereby a carborane unit is
linked to a lipophilic chain and to a Gd–DOTA (DOTA=
Universitꢀ degli Studi di Torino
Via Pietro Giuria, 7, 10125 Torino (Italy)
Fax : (+39)0116707642
E-mail: annamaria.deagostino@unito.it
1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic
acid)
complex through amidic bonds (AT101; Figure 1).[12]
[b] Dr. A. Barge
Dipartimento di Scienza e Tecnologia del Farmaco
Universitꢀ degli Studi di Torino, Via Pietro Giuria, 9
10125 Torino (Italy)
According to the synthetic design, the lipophilic probe
AT101 binds to low-density lipoproteins (LDLs) and accu-
mulates at tumour cells characterised by an up-regulation of
LDL transporters. The exploitation of LDL as biological
vectors was indeed the key to obtaining a high concentra-
tion of B atoms per gram at the tumour cells together with a
[**] MRI=magnetic resonance imaging; BNCT=boron neutron-capture
therapy.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201201634.
Chem. Eur. J. 2013, 19, 721 – 728
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
721

tion on the method proposed by Hagiwara et al.[31] which
uses the Cu-supported ionic-liquid catalyst (SILC). Cuprous
bromide is immobilised in the pores of amorphous silica gel
with the aid of an ionic liquid (1-butyl-3-methylimidazolium
hexafluorophosphate, or [bmim]PF6).
Although some examples of Huisgen cyclisation applied
to carboranes have been previously described,[32–35] to the
best of our knowledge, the results reported here represent
the first example of a Huisgen cyclisation that exploits an
immobilised CuI catalyst in carborane chemistry. Then the
ability of the Gd complex of the synthesised ligand to form
stable adducts with LDLs was evaluated and an MRI assess-
ment was carried out on tumour melanoma cells incubated
in the presence of the Gd/B dual probe.
Figure 1. AT101: GdIII--C-(N-DOTAMA-C6-carbamoylmethyl)-C’-palmi-
tamidomethyl-o-carborane complex.
good specificity for tumour cells. In vivo MR images acquisi-
tion showed that the amount of B taken up in the tumour
region was above the threshold for a successful NCT treat-
ment. After neutron irradiation, the treated mouse group
showed a markedly lower tumour growth with respect to the
control group.[13]
Results and Discussion
With the purpose of simplifying the synthetic procedure
but maintaining, and possibly improving, the favourable
properties of AT101, we designed a new structure that con-
tained a triazole linker instead of the amidic function.
Therefore we were interested in seeing how the chemical
features of the five-membered ring could affect the biologi-
cal behaviour of the desired dual agent. The 1,2,3-triazole
function is a rigid linking unit that can mimic the atom
placement and the electronic properties of the peptide
bond, thereby avoiding the problem associated with the hy-
drolytic cleavage of the latter. Moreover, alkyne and azide
are convenient and stable starting materials that can be in-
troduced independently and do not react with common or-
ganic reagents or functional groups in biomolecules.[14] In
addition, Huisgen cyclisation simplifies the synthesis of the
substituted carborane, which has to be linked to the DOTA
moiety. Although several authors have recently reported the
synthesis of the Gd–DOTA complex by resorting to Huisgen
cycloaddition,[15–21] this synthetic strategy might present
some problems because of the high affinity of DOTA to-
wards various metals. In particular, when comparing the
equilibrium constant (logKML) of DOTA–Gd and DOTA–
Cu complexes (24.7 and 22.7, respectively),[22–24] one might
surmise that Gd complexation could be partially or totally
hampered by the presence of large amounts of Cu2+ ions.
Since residual Cu coupling was observed in ESI mass spec-
tral analysis, many efforts have been recently made toward
developing “copper-free” methodologies for the introduc-
tion of the triazole scaffold in Gd–DOTA complexes by the
use of very active alkynes.[25,26] Moreover, methods that ex-
ploit, for example, Na2S to precipitate CuII at the end of the
reaction or by means of the thio acid/sulfonyl azide amida-
tion have also been reported.[20]
The synthetic approach started from benzyl but-3-ynyl ether
(2), which was treated with paraformaldehyde to afford 5-
benzyloxypent-2-ynol
(3),
as
previously
reported
(Scheme 1).[12,36] The corresponding disubstituted ortho-car-
Scheme 1. Synthesis of o-carborane 6. Reaction conditions and yields:
a) BnBr, NaH, THF, RT (90%); b) BuLi, (CH2O)n, anhydrous THF,
ꢀ208C!RT (65%); c) B10H14, (bmim)+Clꢀ, anhydrous toluene, 908C
(44%); d) NaH, anhydrous THF, 08C!RT, then propargyl bromide,
reflux overnight (85%); e) THF/H2O 50:50, CuACHTUNRGTENNUG(OAc)2/NaACHTUNGTRENNUNG(C6H7O6)
ꢀ
50:50 or EtOH/H2O 50:50 (4 mL) and Cu SILC (300 mg)/6 (0.5 mmol).
borane (4) was then obtained by reaction of the unprotected
alcohol 3 with decaborane in a mixture of toluene/ACTHNUGRTNEUNG(bmim)Cl
at 908C. The reaction was carried out in the presence of a
slight excess amount of alkyne (1.5 equiv). The product was
isolated with a respectable yield after purification. The peak
at 2585 cmꢀ1 in the IR spectrum of 4 clearly indicated the
presence of the carborane cage, and moreover, the success
of the reaction was also confirmed by the detection of the
characteristic boron isotopic distribution in the ESI mass
spectrum (m/z 309 [M+H]+).
Even though innovative methodologies that utilise immo-
bilised Cu catalysts have been recently described, in some
cases copper hydroxide is used[27,28] or basic conditions are
necessary.[29,30] These procedures could not be utilised in the
case of carboranes that might be converted to nido carbor-
anes in basic medium. Consequently, we focused our atten-
The reaction between C-(2-benzyloxy)ethyl-C’-hydroxy-
methyl-o-carborane (4) and propargyl bromide afforded C-
(2-benzyloxy)ethyl-C’-prop-2-ynyloxy-o-carborane (5) in
ꢀ
good yields. The presence of the triple carbon carbon bond
was assessed by the appearance in the H NMR spectrum of
1
722
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 721 – 728

Carborane-Derivative “Click” Reaction
FULL PAPER
intermediate 5 of the triplet centred at d=2.39 ppm, attrib-
utable to the propargylic proton, and by the signals centred
at d=73.0 and 77.8 ppm in the 13C NMR spectrum. The
crude product was directly used for the subsequent Huisgen
cyclisation reaction. The immobilisation of the copper cata-
lyst on the functionalised silica gel was checked by vibra-
tional spectroscopy before the Huisgen cyclisation was per-
formed. The FT-Raman spectra of mercaptopropyl silica gel,
turned into C-[1-hexadecyl-(1H-1,2,3-triazol-4-yl)-C’-2-hy-
droxyethyl-o-carborACTHNUGRTENUNGane (6) in 16 h. The same reaction car-
ried out at RT (Table 1, entry 2) and at 408C (Table 1,
entry 5) needed 7 days and 40 h, respectively, to finish.
Yields were neither affected by increasing the reaction tem-
perature nor by modifying the ratio between azide 12 and
alkyne 5.
Overall, the yields might be considered reasonable in
light of the generally low reactivity of carboranes. Both the
ꢀ
ꢀ
of [bmim]PF6 and of Cu SILC have been compared. The
ꢀ
spectrum of Cu SILC appears by the superposition of the
reaction catalysed by Cu SILC and by CuACTHNGUETRNNU(G OAc)2 afforded
spectral features of [bmim]PF6 and of mercaptopropyl
only the less hindered regioisomer, as confirmed by the
1
ꢀ
modes, except for the intensity of the n
2582 cm , which shows an evident decrease in the Cu
G
signal at d=7.51 ppm in the H NMR spectrum that corre-
ꢀ1
sponds to the CH group of the triazole ring. 13C NMR spec-
troscopy showed two peaks at d=122.5 and 143.1 ppm,
which are attributable to the triazolic CH and quaternary C,
respectively.
SILC spectrum and is attributable to the formation of Cu S
ꢀ
bonds. In the stretching region of the Cu S modes (250–
450 cmꢀ1) some chain bending modes of [bmim]PF6 units[37]
are also present and this makes it difficult to provide a clear
The functionalised carborane 6 was then deprotected by
Pd/C hydrogenation to afford alcohol 7, which was treated
with propargyl bromide to give derivative 8 in good yields
(Scheme 2). Compound 8 is the desired precursor for the
Huisgen cyclisation with the DOTAMA-C6-azide 15.
ꢀ
assignment of n
G
attributed to the medium-weak band at 413 cmꢀ1 (see the
ꢀ
Supporting Information). The Cu SILC-catalysed reaction
was accomplished in a 1:1 mixture of EtOH/H2O in which
the catalyst (300 mg of functionalised SiO2/0.5 mmol of
alkyne) was suspended with alkyne 5 and palmitylazide (12)
according to the procedure described in the original paper
of Hagiwara et al.[31] Afterwards the Huisgen cyclisation was
optimised by changing the reaction temperature and the
stoiACHTUNGTRENNUNGchiACHTUNGTRENNUNGometric ratio between the reagents. To compare
ꢀ
yields and results with those obtained in the Cu SILC-cata-
lysed process, the reaction was also carried out by using a
Scheme 2. Synthesis of o-carborane 8. Reaction conditions and yields:
a) 20% Pd/C, H2 atmospheric, CH2Cl2/MeOH, RT (82%); b) NaH, anhy-
drous THF, 08C!RT, then propargyl bromide, reflux overnight (76%).
stoichiometric amount of CuACHTNUGRTENUNG(OAc)2 and sodium ascorbate
as the reducing agent. The Huisgen cyclisation reaction con-
ditions in the latter case were optimised on a model system
by using a mono-substituted ortho carborane. The reactions
carried out in the presence of catalytic amounts of Cu-
ꢀ
The presence of the triple carbon carbon bond was as-
sessed by the appearance of the multiplet at d=2.48 ppm in
1
(OAc)2 afforded the desired product in low yields. The re-
the H NMR spectrum, and by the two signals at d=74.91
sults are reported in Table 1. The yields obtained in the Cu
SILC-catalysed cyclisation were lower than those obtained
using the homogeneous catalysis (Table 1, entry 1). This
finding was not surprising if we consider that in the latter
and 77.43 ppm in the 13C NMR spectrum, which were as-
signed to the quaternary C and CH of the alkynic portion,
respectively.
ꢀ
Compound 15 was obtained by starting from the corre-
sponding alcohol DOTAMA-C6-OH 13,[38] which was previ-
ously transformed into the analogous mesylate 14
(Scheme 3). The formation of mesylate 14 and the subse-
quent conversion into azide 15 was confirmed by the corre-
sponding ESI mass spectra in which the m/z 773 and 720
case a stoiACHTUNGTRENNUNGchiometric amount of catalyst was used. In the
ꢀ
case of Cu SILC-catalysed reactions, the temperature
strongly affected the cyclisation rate, in fact when the reac-
tion was carried out at 608C (Table 1, entry 3) the reactants
1
[M++Na] peaks were detected. Moreover, in the H NMR
Table 1. Huisgen reaction between o-carborane 5 and hexadecyl azide
(12).
spectra of the DOTA-C6 derivatives 13 and 14, the disap-
pearance of the signal centred at d=3.62 ppm, relevant to
the protons of the CH2 bonded to the OH group, and the
appearance of a new signal at d=4.15 ppm that was attribut-
able to the corresponding CH2 group in mesylate 14 were
meaningful. In the 13C NMR spectrum the signal of the
above-mentioned CH2 group moved from d=61.80 ppm in
alcohol 13 to d=70.00 ppm in mesylate 14 and to d=
55.2 ppm in azide 15. The synthesis of C-[(DOTAMA-C6-
(1H-1,2,3-triazol-4-yl)]ethyloxy-C’-[1-hexadecyl(1H-1,2,3-tri-
azol-4-yl)]-o-carborane (9) was accomplished by using the
optimised procedures described for the Huisgen cyclisation
Entry
Catalyst
No. equiv alkyne/
azide
T
t
Yield
[%][a]
[8C] [h]
1
CuN
1
RT
16
53
bate[b]
[c]
ꢀ
2
3
4
5
Cu SILC
1
1
2
1
RT 168
40
36
35
35
[c]
ꢀ
Cu SILC
60
40
40
16
40
40
[c]
ꢀ
Cu SILC
[c]
ꢀ
Cu SILC
[a] Isolated products, purified by column chromatography. [b] Reactions
conditions: THF/H2O 50:50, Cu(OAc)2/Na(C6H7O6) 50:50. [c] Reactions
conditions: EtOH/H2O 50:50 (4 mL) and Cu SILC (300 mg)/6
A
ACHTUNGTRENNUNG
ꢀ
(0.5 mmol).
Chem. Eur. J. 2013, 19, 721 – 728
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
723

A. Deagostino et al.
termined by inductively cou-
pled plasma (ICP)-MS, 14 and
35% of Cu per ligand mole-
cule was found in the two
ligand solutions, respectively,
the ligand titration with GdCl3
in water at neutral pH fol-
lowed by relaxometric meas-
urements showed the equiva-
Scheme 3. Synthesis of DOTAMA-C6-N3 15: a) Et3N, MsCl, anhydrous Et2O, 0 8C!RT (77%); b) NaN3, anhy-
drous DMF, RT (64%).
between C-(2-benzyloxy)ethyl-C’-prop-2-ynyloxy-o-carbor-
lence point exactly at 1:1 ligand/metal concentration. This
indicates the absence of interference by this small amount
of Cu in the Gd complexation process.
ꢀ
ane (5) and hexadecylazide (12). The Cu SILC catalyst was
suspended in a mixture of EtOH/H2O 50:50 in which alkyne
8 and the azide 15 were dissolved and stirred overnight at
608C.
Relaxometric characterisation of the Gd complex and LDL
adduct preparation: The relaxivity (that is, the proton relax-
ation enhancement of water protons in the presence of the
paramagnetic complex at 1 mm concentration) of MEA01
was measured at 21.5 MHz and 298 K and was found to be
16.8 mmꢀ1 sꢀ1, which is in the typical range of slowly moving
supramolecular adducts that involve the neutral GdIII–
DOTA monoamide moiety.
The disappearance of alkyne 8 was checked by TLC by
looking for the disappearance of the spot attributable to 8.
After filtration and evaporation of the solvent the crude
product was purified and the DOTA tert-butyl protecting
groups were removed by using an excess amount of
CF3COOH at RT, as demonstrated by the disappearance of
the tBu signals in the NMR spectra of the product. Huisgen
cyclisation was also carried out by using Cu
presence of Na ascorbate to compare the yields with the
G
The various parameters affecting the observed proton re-
laxivity (r1p) were assessed by analysing the dependence of
the H water relaxation rates as a function of the applied
magnetic field (1/T1 NMRD profile). The best fit of the data
(Table 2) to the values calculated on the basis of the Solo-
1
ꢀ
ones of the Cu SILC-catalysed process. The desired product
9 (Scheme 4) was obtained in 18% yield when Cu SILC
Table 2. Best-fitting parameters obtained from the analysis of the
NMRD profile by considering one inner-sphere water molecule (q=1),
the protons of which are at an average metal distance of 3 ꢄ.
D2 [sꢀ2/1019]
tV [ps]
24ꢁ3
tR [ns]
66ꢁ10
tM [ms]
1ꢁ0.3
MEA01
1.2ꢁ0.5
mon–Bloembergen–Morgan equations (for the inner sphere
contribution) and of Freedꢂs equation (for the outer sphere
contribution) indicates that the increase in relaxivity with
respect to the parent Gd–DOTA complex is a result of sig-
nificant lengthening of tr as a consequence of its self-assem-
bling in water in micellar aggregates.
To pursue the formation of supramolecular adducts with
LDLs, MEA01 micelles were disaggregated by using an
excess amount of b-cyclodextrin (b-CD).[39] The transfer of
the MEA01 units from the micelles to the supramolecular
“host/guest” complex with b-CD was followed by measuring
the relaxation rate of the solvent water protons. Then the
stepwise addition of the b-CD/MEA01 adduct to the aque-
ous solution of LDL particles allowed the translocation of
the amphiphilic molecules into the LDL particles to occur.
This step was again assessed by measuring the changes in
the relaxation enhancement of the solvent water protons
(Figure 2). The analysis of the titration curve showed a rea-
sonably high MEA01-to-LDL binding affinity (Ka =(1.4ꢁ
0.5)ꢃ104 mꢀ1) and a number of binding sites on the surface
of the LDLs equal to 300ꢁ50. The dynamic light scattering
(DLS) measurements indicated that the size of the formed
LDL adducts ((21.8ꢁ2) nm) was almost the same as that
Scheme 4. Synthesis of GdIII–MEA01. Reaction conditions and yields:
a) DOTAMA-C6N3 15, THF/H2O 50:50, CuACHTNUTGRNENUG(OAc)2/NaACHUTGNTREN(NUGN C6H7O6) 1:1, RT
ꢀ
(35%) or EtOH/H2O 50:50 (4 mL), Cu SILC, 608C (18%); b) CH2Cl2,
CF3COOH, RT (99%), c) GdCl3, H2O, RT, pH 7.
was utilised, and in 35% yield under homogeneous condi-
tions. These low yields were not surprising because of the
low reactivity of DOTA and carborane moieties. It should
be stressed that in the former case a catalytic amount of im-
mobilised recyclable Cu+ was used, whereas in the latter the
product was obtained by using a stoichiometric amount of
catalyst. The mass spectrum of the final ligand 9 did not
show any evidence of the Cu–MEA01 adduct, nor did it
ꢀ
when Cu SILC was used to introduce the triazole moiety
when the Huisgen reactions were carried out under homoge-
neous conditions. Although, when using the Cu/B ratio de-
724
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 721 – 728

Carborane-Derivative “Click” Reaction
FULL PAPER
Figure 2. Titration of native LDL into the Gd/B/L–b-CD adduct. The
samples were incubated for 2 h at 378C prior to the R1 measurements.
found in the native LDL particles ((22.5ꢁ1) nm). This indi-
cates the absence of any change in the overall protein struc-
ture upon the binding of B/Gd-containing units. On the
basis of these results, one might conclude that, as far as the
formation of inclusion complexes with LDL is concerned,
the behaviour of MEA01 is very similar to what has previ-
ously been reported for AT101.
Cell uptake experiments: Using the method reported above,
an MEA01/LDL adduct made up of 295ꢁ15 Gd complex
per protein was prepared by incubating LDL and b-CD/
MEA01 340:1 for 2 h at 378C. After purification, about
76% of the total protein was recovered as MEA01 adduct
and the MEA01 loading efficiency into LDLs was 70%.
These values are similar to those obtained for the parent
compound AT101.[40] The stability of MEA01/LDL in
plasma was evaluated by measuring the relaxation rate of a
0.07 mm complex solution dissolved in human serum (sero-
norm) at 378C. The R1obsd remained unchanged during the
monitored period of 24 h. Furthermore, an increasing
amount (0–150 mgmLꢀ1) of human albumin was added to a
solution of MEA01 (0.07 mm) in phosphate buffer solution
(PBS) without observing any increase of the paramagnetic
contribution to the relaxation rate, thus excluding any inter-
action with this protein. We can conclude that the stability
of the ME01/LDL adduct is high enough to prevent its dis-
sociation when injected into the bloodstream.
Figure 3. In vitro uptake experiments in cultured B16 cells. A) T1-weight-
ed spin–echo MR image of an agar phantom containing unlabelled cells
(1) and cells incubated with 5, 10, 20 and 30 mgmLꢀ1 of MEA01/LDL (2–
4, respectively) for 16 h at 378C. B) Relaxation rates measured at 7 T in
B16 cells incubated for 16 h at 378C with MEA01/LDL as a function of
the concentration of LDL particles. The vertical lines indicate the stand-
ard deviations of the relaxation rates [sꢀ1].
16 h of cell incubation in the presence of the adduct
(20 mgmLꢀ1), the uptake by B16 cells decreased by 65%
when the concentration of native LDL particles added to
the culture medium was 150 mgmLꢀ1.
Gd/B-loaded LDL particles were added to the B16 mela-
noma cell growth medium (previously incubated for 24 h
with a lipoprotein deficient serum (LPDS) to increase LDL
receptor expression).
Cellular labelling experiments proved that after 16 h of in-
cubation in the presence of 10–50 mgmLꢀ1 MEA01/LDL,
the amount of internalised Gd was sufficient to generate hy-
perintense signals in MR images (Figure 3). The correspond-
ing relaxation rates measured in the labelled cells showed
saturation-like behaviour. Competition assays with free
LDL were carried out to demonstrate that the uptake of the
MEA01/LDL adduct from B16 cells involves LDLRs. After
Table 3 shows the amount of Gd and B atoms internalised
by B16, which was measured by ICP-MS after 16 h incuba-
tion in the presence of increasing amounts of MEA01/LDL
(calculated under the assumption that 1 g of tissue contains
1ꢃ109 cells).
Table 3. Amount of Gd and B atoms internalised by B16 after 16 h incu-
bation in the presence of increasing amounts of MEA01/LDL.
Entry Adduct [mg per mL LDL] Boron [ppm] (MEA01/LDL adduct)
1
2
3
10
20
30
16
22
28
Chem. Eur. J. 2013, 19, 721 – 728
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
725

A. Deagostino et al.
We can conclude that, by using MEA01/LDL adduct, the
minimum amount of intracellular boron necessary to per-
form BNCT (about 20 ppm) is reached at a relatively low
boron-containing labelled LDL concentration.
Experimental Section
General: Flasks and all equipment used for the generation and reaction
of moisture-sensitive compounds were dried by using an electric heater
under Ar. THF was distilled from benzophenone ketyl, anhydrous Et2O
was distilled from LiAlH4 and anhydrous CH2Cl2 from CaH2 prior to use.
BuLi (1.6m in hexanes) was obtained from Aldrich. (Bmim)+Clꢀ was
purchased from Solvent Innovation GmbH. Decaborane was bought
from KATCHEM spol. sro. All commercially obtained reagents and sol-
vents were used as received. Products were purified by preparative
column chromatography on Macherey–Nagel silica gel for flash chroma-
tography, 0.04–0.063 mm/230–400 mesh using mixtures of petroleum
ether 40–60 (EP) and diethyl ether (EE) or CHCl2 and MeOH as eluants.
o-Carborane 9 was deprotected by following the procedure already re-
ported.[12]
Conclusion
In this study, it has been shown that the use of the Huisgen
reaction greatly simplifies the synthesis of lipophilic
GdBNCT/MRI agents. Moreover, a heterogeneous catalytic
methodology has been applied to the insertion of carboranes
for the first time. As a matter of fact, the proposed synthetic
procedure allows the new ligand structure (MEA01) to be
obtained in relatively good yields with significantly fewer
steps than the previously reported AT101 system. In particu-
lar, the new synthetic strategy allows for the synthetic steps
to be reduced from fourteen to nine and the overall yield to
be improved from 3 to 8%, as evidenced in Table 4.
Reactions were monitored by TLC using silica gel on TLC-PET foils
Fluka (2–25 mm, layer thickness 0.2 mm, medium pore diameter 60 ꢄ).
Carboranes and their derivatives were visualised on TLC plates by using
a 5% PdCl2 aqueous solution in HCl. 1H NMR spectra were recorded at
200 MHz, 13C NMR spectra at 50.2 MHz. Data were reported as follows:
chemical shifts [ppm] from tetramethylsilane as internal standard, multi-
plicity (s=singlet, d=doublet, t=triplet, q=quartet, dd=double-dou-
blet, m=multiplet, br=broad), coupling constants [Hz], integration and
assignment. 13C NMR spectra were measured with complete proton de-
coupling. Chemical shifts were reported [ppm] from the residual solvent
as an internal standard. GC-MS spectra were obtained using a mass se-
lective detector HP 5970 B instrument operating at an ionising voltage of
70 eV connected to an HP 5890 GC with a cross-linked methyl silicone
capillary column (25 mꢃ0.2 mmꢃ0.33 mm film thickness). ESI-MS spec-
tra were obtained using a Waters micromass ZQ spectrometer equipped
with an ESI ion source. IR spectra were recorded using a Perkin–Elmer
Table 4. Comparison of some properties of AT101 and MEA01.
AT101
MEA01
synthetic steps
14
3
17.3
9
8.1
16.8
overall yield [%][a]
relaxivity [mmꢀ1 sꢀ1
]
[a] Isolated products, purified by column chromatography.
ꢀ
BX FTIR. The Cu SILC catalyst was prepared by following the litera-
ture. Benzyl but-3-ynyl ether l (2) and 5-benzyloxypent-2-ynol (3) were
synthesised as described in the literature and their spectroscopic data cor-
responded with those reported.[36]
To take full advantage of the Gd–DOTA moiety as an
MRI probe and allow the Gd–DOTA complex to be
formed, the introduction of the triazoles as the linker be-
tween a carborane cage, a long alkylic chain and DOTA
ligand, respectively, was previously accomplished by the use
of Cu–SILC catalyst, whereby Cu+ ions are immobilised on
a silica gel matrix. The MRI/BNCT agent MEA01 was pre-
pared by following a procedure that also exploits the homo-
geneous catalysis to compare the relaxivity properties and
the biological behaviour in both cases. The complex ob-
tained was fully characterised from the relaxometric point
of view and, after disaggregation with b-cyclodextrin,
MEA01/LDL adducts were prepared and the results showed
a reasonably high MEA01-to-LDL binding affinity, very
close to that of AT101 (Table 4). Finally, the MEA/01 ad-
ducts were tested in vitro on B16 melanoma cells and a suit-
able intracellular Gd/B concentration was detected after
24 h of incubation. This result is particularly useful in light
of future in vivo applications in which higher concentrations
of LDL adducts at the target site are usually difficult to
reach. Furthermore, we can conclude that the substitution of
amidic groups with triazolic rings does not significantly
change the relaxometric properties of the Gd complex and
its affinity for the biological carrier LDLs.
The 1/T1 nuclear magnetic relaxation dispersion profiles of water protons
were measured over
a continuum of magnetic field strengths from
0.00024 to 2.4 T (which correspond to 0.01–80 MHz in proton Larmor
frequency) using a fast field-cycling Stelar Spinmaster FFC relaxometer
(from 0.01 to 20 MHz) and using a Stelar Spinmaster variable magnetic
field instrument (from 20 to 80 MHz). The Gd, B and Cu content was de-
termined by using inductively coupled plasma mass spectrometry (ICP-
MS; element-2; Thermo-Finnigan, Rodano (MI), Italy). Sample digestion
was performed with concentrated HNO3 (70%, 2 mL) under microwave
heating (Milestone MicroSYNTH Microwave labstation). Dynamic light-
scattering measurements, which were made to determine the size of the
micellar system, were performed using a Malvern Zetasizer SZ appara-
tus. A laser was used as the light source to illuminate the sample particles
within the cell.
Procedure for the azide–alkyne cycloaddition in hetereogeneous and ho-
mogeneous conditions:
o-Carborane 6: Method A: o-Carborane 5 (0.72 mmol, 0.25 g) and palmi-
tilazide 12 (0.72 mmol, 0.19 g) in a 25 mL round-bottomed flask were sus-
ꢀ
pended in a solution of H2O/MeOH (4 mL, 1:1) with Cu SILC (450 mg)
and stirred at 608C overnight under Ar. The solution was then filtered
and the solvent was evaporated. The crude was purified by column chro-
matography (eluent: CH2Cl2/MeOH 99:1) to afford a pale yellow oil
(0.18 g, 40%). Method B: o-Carborane 5 (1.50 mmol, 0.53 g) and palmi-
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
0.30 g). The THF was then removed by evaporation and CH2Cl2 was
added. The aqueous phase was then extracted twice with CH2Cl2 and the
organic phases were washed with brine (2ꢃ10 mL). The crude was puri-
fied by column chromatography (eluent: CH2Cl2/MeOH 99:1) to afford a
pale yellow oil (0.23 g, 53%). 1H NMR (200 MHz, CDCl3, Me4Si): d=
0.89 (brt, J=8.3 Hz, 3H; CH3), 1.26 (brs, 26H; CH
3ACHTUNGTNERN(UNG CH2)13), 1.87 (m,
2H; CH3A(CH2)12CH2), 2.47 (t, J=6.8 Hz, 2H; PhCH2OCH2CH2), 3.59 (t,
CTHUNGTRENNUNG
726
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 721 – 728

Carborane-Derivative “Click” Reaction
FULL PAPER
J=6.8 Hz, 2H; PhCH2OCH2CH2), 4.08 (s, 2H; B10H10C2CH2O), 4.29 (t,
J=7.2 Hz, 2H; (CH2)13CH2CH2N3C2), 4.47 (s, 2H; OCH2N3C2), 4.65 (s,
2H; PhCH2O), 7.34 (m, 5H; PhCH2OCH2CH2), 7.48 ppm (s, 1H;
HN3C2); 13C NMR (50.2 MHz, CDCl3, Me4Si): d=14.0 (q), 22.5 (t), 26.3
(t), 28.8 (t), 29.2 (5ꢃt), 29.4 (2ꢃt), 29.5 (2ꢃt), 30.1 (t), 31.8 (t), 34.6 (t),
50.2 (t), 64.5 (t), 68.2 (t), 69.7 (t), 73.0 (t), 75.5 (s), 76.8 (s), 122.3 (d),
127.5 (2ꢃd), 127.7 (d), 128.3 (2ꢃd), 137.4 (s), 143.4 ppm (s); IR (neat):
n˜max =2925, 2854, 2585, 1456, 1103 cmꢀ1; MS (ESI+): m/z: 614 [M+H]+;
elemental analysis calcd (%) for C31H59B10N3O2: C 60.65, H 9.69, N 6.84;
found C 60.68, H 9.67, N 6.85.
CONHCH2ACHTUGNTRENUN(G CH2)4, 1.89 (m, 2H; CH3CATHNUGTRENN(UNG CH2)13CH2), 2.10–3.00 (m, 18H;
NCH2CH2N, B10H10C2CH2CH2), 3.10–3.40 (m, 10H; NCH2COOtBu,
NHCH2CH2), 3.59 (m, 2H; B10H10C2CH2CH2O), 4.05 (s, 2H;
B10H10C2CH2O), 4.27 (m, 4H; N3C2CH2O), 4.65 (m, 4H; CH2N3C2), 7.60
(m, 1H; HN3C2), 7.70 (m, 1H; HN3C2); 8.00 ppm (m, 1H; NH);
13C NMR (50.2 MHz, CDCl3, Me4Si): d=13.9 (q), 22.5 (t), 25.6 (t), 25.8
(t), 26.0 (t), 26.3 (t), 26.4 (t), 27.7 (9ꢃq), 28.5 (t), 29.1 (t), 29.5 (9ꢃt),
32.1 (t), 38.9 (t), 44.9 (t), 50.0–60.0 (14ꢃt), 64.0 (t), 64.4 (t), 68.0 (t), 69.7
(t), 75.5 (s), 80.7 (s), 81.5 (2ꢃs), 81.6 (s), 122.3 (d), 123.0 (d), 143.2 (s),
143.7 (s), 171.6 (s), 171.3 (s), 171.8 (s), 172.1 ppm (s); IR (neat): v˜max
=
3420, 2927, 2574, 1732, 1674, 1539 cmꢀ1; MS (ESI+): m/z: 1260 [M+H]+;
elemental analysis calcd (%) for C61H119B10N11O9: C 58.20, H 9.53, N
12.24; found: C 58.22, H 9.52, N 12.25.
o-Carborane 7: o-Carborane 6 (2.8 mmol, 0.17 g) in a 25 mL round-bot-
tomed flask was dissolved in a solution of CH2Cl2/MeOH (15 mL, 50:50),
then Pd/C was added (20% w/w, 34 mg), the solution was stirred at RT
overnight under H2 atmosphere. The solution was then filtered and the
solvent was removed by evaporation. The crude was purified by column
chromatography (eluent: CH2Cl2/MeOH 99:1) to afford a pale yellow oil
(0.12 g, 82%). 1H NMR (200 MHz, CDCl3, Me4Si): d=0.90 (brt, J=
GdIII–MEA01: The ligand was dissolved in water at low concentration
(0.1 mm) and the pH was adjusted to 7 by adding 1m NaOH. An equimo-
lar amount of GdCl3·6H2O was dissolved in water (0.5 mL) and slowly
added to the first solution while maintaining the pH value at 6.5 with
NaOH. The mixture was then stirred at room temperature for 16 h. Xyle-
nol Orange UV spectrophotometry was used to check for the absence of
free GdIII ions.
6.6 Hz, 3H; CH3), 1.25 (brs, 26H; CH
3ACHTUNGTRENUN(GN CH2)13), 1.91 (m, 2H; CH3-
ACHTUNGTRENNUNG
3.74 (t, J=7.2 Hz, 2H; HOCH2CH2), 4.12 (s, 2H; B10H10C2CH2O), 4.35
(t, J=7.2 Hz, 2H; (CH2)12CH2CH2N3C2), 4.66 (s, 2H; HOCH2N3C2),
7.51 ppm (s, 1H; HN3C2); 13C NMR (50.2 MHz, CDCl3, Me4Si): d=13.9
(q), 22.5 (t), 26.3 (t), 28.8 (t), 29.1 (t), 29.2 (4ꢃt), 29.3 (2ꢃt), 29.5 (2ꢃt),
30.0 (t), 31.7 (t), 37.1 (t), 50.4 (t), 60.5 (t), 64.0 (t), 70.0 (t), 75.4 (s), 77.1
(s), 122.7 (d), 143.0 ppm (s); IR (neat): n˜max =3368, 2928, 2585, 1467,
1052 cmꢀ1; MS (ESI+): m/z: 524 [M+H]+; elemental analysis calcd (%)
for C24H53B10N3O2: C 55.03, H 10.20, N 6.11; found: C 55.00, H 9.87, N
6.11.
Preparation of LDL adducts: MEA01 micelle disaggregation was carried
out by mixing an excess amount of b-CD with a 0.07 mm complex solu-
tion (30’, 208C, 50:1). MEA01-loaded LDL particles were prepared by
incubating LDL (0.2 mm; Biomedical Technologies Inc., Stoughton, MA,
USA) and MEA01/b-CD adducts (0.07 mm) for 2 h at 378C. The b-CD
and the unbound complex were eliminated by dialysis. MEA01-LDL
loaded with 295 complexes per LDL particle was used for cell experi-
ments. The final Gd concentration was determined by T1 measurements
in the 1H NMR spectra of the mineralised complex solution (at 20 MHz,
2588C, in 6m HCl at 1208C for 16 h) and the protein concentration was
determined by means of a commercial Bradford assay (Biorad, Hercules,
CA, USA).
o-Carborane 8: o-Carborane 7 (0.28 mmol, 0.15 g) in a 50 mL three-
necked round-bottomed flask was dissolved in anhydrous THF (20 mL)
and cooled to 08C, then NaH (2.5 equiv, 0.7 mmol, 0.02 g) was slowly
added. The reaction mixture was stirred at 08C for 20 min and at RT for
1 h, then 80% propargyl bromide solution was added (5 equiv, 1.4 mmol,
0.13 mL). The reaction was stirred at reflux overnight and was then
quenched with a saturated aqueous NH4Cl and extracted with Et2O (3ꢃ
10 mL). The combined extracts were washed with brine (1ꢃ10 mL),
dried and evaporated under reduced pressure. The crude was purified by
column chromatography (eluent: EP/EE 99:1) to afford a pale yellow oil
(0.12 g, 76%). 1H NMR (200 MHz, CDCl3, Me4Si): d=0.88 (brt, J=
Cell lines: Mouse melanoma (B16-F10) were obtained from the Ameri-
can Type Culture Corporation and the cells were cultured in Dulbeccoꢂs
modified eagle medium (DMEM; Lonza) supplemented with glucose
(4.5 gLꢀ1), 10% (v/v) fetal bovine serum (FBS), 2 mm glutamine,
100 UmLꢀ1 penicillin, and 100 UmLꢀ1 streptomycin. For uptake experi-
ments, 3ꢃ105 B16 were seeded in 6 cm dishes. The cells were incubated
the following day for 24 h with culture media added with 6% lipopro-
tein-deficient serum (Biomedical Technologies Inc., Stoughton, MA,
USA) to increase low-density lipoprotein receptor (LDLR) expression.
Finally, the cells were incubated for 16 h with MEA01-loaded LDL parti-
cles at different concentrations (10–50 mgmLꢀ1). At the end of the incu-
bation, the cells were washed with ice-cold PBS (3ꢃ10 mL), detached
with trypsin/ethylenediaminetetraacetic acid (EDTA), and transferred
into glass capillaries for MRI analysis (see below). The B and Gd content
in each cell line was determined by using inductively coupled plasma
mass spectrometry (element-2; Thermo-Finnigan, Rodano (MI), Italy).
6.4 Hz, 3H; CH3), 1.26 (brs, 26H; CH
3ACHTUNGTRENUN(GN CH2)13), 1.95 (m, 2H; CH3-
ACHTUNGTRENNUNG
7.0 Hz, 2H; B10H10C2CH2CH2O), 4.09 (s, 2H; OCH2B10H10), 4.12 (d, J=
2.4 Hz, 2H; CH2CꢂCH), 4.36 (t, J=7.2 Hz, 2H; (CH2)12CH2CH2N3C2),
4.70 (s, 2H; OCH2N3C2), 7.27 ppm (s, 1H; HN3C2); 13C NMR (50.2 MHz,
CDCl3, Me4Si): d=13.9 (q), 22.4 (t), 26.2 (t), 28.7 (t), 29.1 (2ꢃt), 29.2 (5ꢃ
t), 29.4 (2ꢃt), 30.0 (t), 31.6 (t), 34.2 (t), 50.1 (t), 58.0 (t), 64.2 (t), 67.5 (t),
69.5 (t), 75.0 (s), 75.4 (s), 77.4 (s), 78.7 (d), 122.5 (d), 143.1 ppm (s); IR
(neat): v˜max =3308, 2926, 2586, 2102, 1736 cmꢀ1; MS (ESI+): m/z: 562
[M+H]+; elemental analysis calcd (%) for C27H55B10N3O2: C 57.72, H
9.87, N 7.48; found: C 57.75, H 9.99, N 7.47.
Magnetic resonance imaging: MR images were acquired using a Bruker
Avance 300 spectrometer (7T) equipped with a Micro 2.5 microimaging
probe (Bruker BioSpin, Ettlingen, Germany). Glass capillaries that con-
tained about 2ꢃ106 cells were placed in an agar phantom and MR imag-
ing was performed by using a standard T1-weighted multislice spin–echo
sequence (TR/TE/NEX=200:3.3:8, FOV=1.2 cm, NEX=number of exi-
tations; FOV=field of view). The T1 relaxation times were calculated by
using a standard saturation recovery spin echo.
o-Carborane 9: Method A: o-Carborane 8 (0.18 mmol, 0.10 g) and tert-
butyl-DOTAMA-C6 azide 15 (0.18 mmol, 0.12 g) in a 25 mL round-bot-
tomed flask were suspended in H2O/MeOH (4 mL, 50:50) solution with
ꢀ
Cu SILC (150 mg) and stirred at 608C overnight under Ar. The solution
was then filtered and the solvent was evaporated. The crude was purified
by column chromatography (eluent: gradient from CH2Cl2/MeOH 98:2
to CH2Cl2/MeOH 80:20) to afford a pale yellow oil (0.04 g, 18%). Meth-
od B: o-Carborane 8 (0.33 mmol, 0.18 g) and tert-butyl-DOTAMA-C6
azide 15 (0.33 mmol, 0.23 g) in a 50 mL round-bottomed flask were dis-
solved in H2O/THF (10 mL, 50:50) and stirred overnight at RT in the
Acknowledgements
presence of 20 mol% CuACTHNUTRGNEUNG(OAc)2 (0.007 mmol, 0.30 g) and 40 mol% Na
The authors gratefully acknowledge financial and scientific support from
Regione Piemonte, Nano-IGT project (Converging Technologies),
ENCITE project (FP7-HEALTH-2007A), Fondazione Compagnia di San
Paolo (Torino), and EU Action COST D38.
ascorbate (0.13 mmol, 0.03 g). The THF was then evaporated and CH2Cl2
was added. The aqueous phase was then extracted twice with CH2Cl2 and
the organic phases were washed with brine (2ꢃ10 mL). The crude was
purified by column chromatography (eluent: gradient from CH2Cl2/
MeOH 98:2 to CH2Cl2/MeOH 80:20) to afford a pale yellow oil (0.15 g,
35%). 1H NMR (200 MHz, CDCl3, Me4Si): d=0.87 (brt, J=6.8 Hz, 3H;
CH3), 1.24 (brs, 26H; CH
G
[1] R. F. Barth, J. Neuro-Oncol. 2003, 62, 1.
Chem. Eur. J. 2013, 19, 721 – 728
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
727

A. Deagostino et al.
[2] R. F. Barth, J. A. Coderre, M. G. H. Vicente, T. E. Blue, Clin.
Cancer Res. 2005, 11, 3987.
[22] A. E. Martell, R. D. Hancock, Metal Complexes in Aqueous Solu-
tion, Plenum Press, New York, 1996.
[3] M. F. Hawthorne, M. W. Lee, J. Neuro-Oncol. 2003, 62, 33.
[4] A. H. Soloway, W. Tjarks, B. A. Barnum, F. G. Rong, R. F. Barth,
I. M. Codogni, J. G. Wilson, Chem. Rev. 1998, 98, 1515.
[5] J. F. Valliant, K. J. Guenther, A. S. King, P. Morel, P. Schaffer, O. O.
Sogbein, K. A. Stephenson, Coord. Chem. Rev. 2002, 232, 173.
[6] G. Wu, R. F. Barth, W. Yang, R. J. Lee, W. Tjarks, M. V. Backer,
J. M. Backer, Anti-Cancer Agents Med. Chem. 2006, 6, 167.
[7] E. L. Crossley, E. J. Ziolkowski, J. A. Coderre, L. M. Rendina, Mini-
Rev. Med. Chem. 2007, 7, 303.
[23] W. P. Cacheris, S. K. Nickle, A. D. Sherry, Inorg. Chem. 1987, 26,
958.
[24] R. E. Mewis, S. J. Archibald, Coord. Chem. Rev. 2010, 254, 1686.
[25] M. K. Schultz, S. G. Parameswarappa, F. C. Pigge, Org. Lett. 2010,
12, 2398.
[26] M. E. Martin, S. G. Parameswarappa, M. S. OꢂDorisio, F. C. Pigge,
M. K. Schultz, Bioorg. Med. Chem. Lett. 2010, 20, 4805.
[27] T. Katayama, K. Kamata, K. Yamaguchi, N. Mizuno, Chemsuschem
2009, 2, 59.
[8] H. C. Kolb, M. G. Finn, K. B. Sharpless, Angew. Chem. Int. Ed. 2001,
40, 2004.
[28] K. Yamaguchi, T. Oishi, T. Katayama, N. Mizuno, Chem. Eur. J.
2009, 15, 10464.
[9] H. C. Kolb, K. B. Sharpless, Drug Discovery Today 2003, 8, 1128.
[10] C. R. Becer, R. Hoogenboom, U. S. Schubert, Angew. Chem. Int.
Ed. 2009, 48, 4900.
[11] J.-F. Lutz, Angew. Chem. Int. Ed. 2008, 47, 2182.
[12] S. Aime, A. Barge, A. Crivello, A. Deagostino, R. Gobetto, C.
Nervi, C. Prandi, A. Toppino, P. Venturello, Org. Biomol. Chem.
2008, 6, 4460.
[29] C. Girard, E. Onen, M. Aufort, S. Beauviere, E. Samson, J. Hersco-
vici, Org. Lett. 2006, 8, 1689.
[30] U. Sirion, Y. J. Bae, B. S. Lee, D. Y. Chi, Synlett 2008, 2326.
[31] H. Hagiwara, H. Sasaki, T. Hoshi, T. Suzuki, Synlett 2009, 643.
[32] C. Di Meo, L. Panza, F. Campo, D. Capitani, L. Mannina, A. Banza-
to, M. Rondina, A. Rosato, V. Crescenzi, Macromol. Biosci. 2008, 8,
670.
[13] S. Geninatti-Crich, D. Alberti, I. Szabo, A. Deagostino, A. Toppino,
A. Barge, F. Ballarini, S. Bortolussi, P. Bruschi, N. Protti, S. Stella, S.
Altieri, P. Venturello, S. Aime, Chem. Eur. J. 2011, 17, 8479.
[14] J. C. Jewett, C. R. Bertozzi, Chem. Soc. Rev. 2010, 39, 1272.
[15] J. M. Bryson, W. J. Chu, J. H. Lee, T. M. Reineke, Bioconjugate
Chem. 2008, 19, 1505.
[33] B. A. Wojtczak, A. Andrysiak, B. Gruner, Z. J. Lesnikowski, Chem.
Eur. J. 2008, 14, 10675.
[34] Z. J. Lesnikowski, J. Organomet. Chem. 2009, 694, 1771.
[35] R. Djeda, J. Ruiz, D. Astruc, R. Satapathy, B. P. Dash, N. S. Hos-
mane, Inorg. Chem. 2010, 49, 10702.
[36] Y. Wu, P. J. Carroll, W. Quintana, Polyhedron 1998, 17, 3391.
[37] R. W. Berg, M. Deetlef, K. R. Seddon, I. Shim, J. M. Thompson, J.
Phys. Chem. B 2005, 109, 19018.
[16] S. Knçr, A. Modlinger, T. Poethko, M. Schottelius, H.-J. Wester, H.
Kessler, Chem. Eur. J. 2007, 13, 6082.
[17] I. Dijkgraaf, A. Y. Rijnders, A. Soede, A. C. Dechesne, G. W.
van Esse, A. J. Brouwer, F. H. M. Corstens, O. C. Boerman, D. T. S.
Rijkers, R. M. J. Liskamp, Org. Biomol. Chem. 2007, 5, 935.
[18] Y. Song, E. K. Kohlmeir, T. J. Meade, J. Am. Chem. Soc. 2008, 130,
6662.
[19] C.-B. Yim, O. C. Boerman, M. de Visser, M. de Jong, A. C. De-
chesne, D. T. S. Rijkers, R. M. J. Liskamp, Bioconjugate Chem. 2009,
20, 1323.
[38] A. Barge, L. Tei, D. Upadhyaya, F. Fedeli, L. Beltrami, R. Stefania,
S. Aime, G. Cravotto, Org. Biomol. Chem. 2008, 6, 1176.
[39] K. C. Briley-Saebo, S. Geninatti-Crich, D. P. Cormode, A. Barazza,
W. J. M. Mulder, W. Chen, G. B. Giovenzana, E. A. Fisher, S. Aime,
Z. A. Fayad, The Journal of Physical Chemistry B 2009, 113, 6283.
[40] S. Geninatti-Crich, D. Alberti, I. Szabo, A. Deagostino, A. Toppino,
A. Barge, F. Ballarini, S. Bortolussi, P. Bruschi, N. Protti, S. Stella, S.
Altieri, P. Venturello, S. Aime, Chem. Eur. J. 2011, 17, 8479.
[20] C. B. Yim, I. Dijkgraaf, R. Merkx, C. Versluis, A. Eek, G. E.
Mulder, D. T. S. Rijkers, O. C. Boerman, R. M. J. Liskamp, J. Med.
Chem. 2010, 53, 3944.
[21] M. Q. Tan, X. M. Wu, E. K. Jeong, Q. J. Chen, Z. R. Lu, Biomacro-
molecules 2010, 11, 754.
Received: May 9, 2012
Revised: September 14, 2012
Published online: November 14, 2012
728
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 721 – 728