Synthesis and characterization of a biotinylated multivalent targeted contrast agent

Article abstract of DOI:10.1002/cplu.201402329

A new bimodal and multivalent dendritic contrast agent (CA) that targets the protein avidin was prepared and characterized. The tripartite lysine core was used to link the ligand biotin, the fluorescent dye, and the dendron carrying GdDOTA (DOTA=1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid) chelates for amplification of the magnetic resonance imaging (MRI) signal. The longitudinal relaxivity of this dendrimeric CA was greater than those of its GdDOTA chelate and most of the common commercial agents at the investigated high magnetic field (7 T). The capacity of the dendrimeric CA to bind to the target protein was confirmed by fluorescence measurements upon its treatment with NeutrAvidin-agarose gel or NeutrAvidin-coated microspheres and the results were compared with those of its monomeric analogue. The fluorescence intensity of monomer-treated targets was found to be greater than that from those treated with dendrimeric CA; however, a several-fold increase in the MRI signal was observed on the same samples treated with the dendrimeric CA. The inductively coupled plasma mass spectrometry analysis of the digested samples indicated somewhat higher Gd³⁺ content and hence slightly better binding of monomeric versus dendrimeric CA. This bimodal and multivalent targeted probe opens an avenue for the preparation of new nanosized CAs that allow high-resolution MRI of various targets, such as cellular receptors or specific cellular populations. Target and amplify: A nanosized target-specific bimodal and multivalent contrast agent (CA) has been prepared and characterized. This dendrimeric CA bears the ligand biotin and allows dual modal readout by means of magnetic resonance and optical imaging techniques. The demonstrated specific binding to avidin-coated agarose gels and microspheres (see figure) opens new perspectives for high-affinity and high-resolution MRI of various cellular targets.

Full text of DOI:10.1002/cplu.201402329

DOI: 10.1002/cplu.201402329
Full Papers
Synthesis and Characterization of a Biotinylated
Multivalent Targeted Contrast Agent
Serhat Gündüz,[a, b] Anthony Power,[b] Martin E. Maier,[c] Nikos K. Logothetis,[b, d] and
Goran Angelovski*[a]
A new bimodal and multivalent dendritic contrast agent (CA)
that targets the protein avidin was prepared and characterized.
The tripartite lysine core was used to link the ligand biotin, the
fluorescent dye, and the dendron carrying GdDOTA (DOTA=
1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid) che-
lates for amplification of the magnetic resonance imaging
(MRI) signal. The longitudinal relaxivity of this dendrimeric CA
was greater than those of its GdDOTA chelate and most of the
common commercial agents at the investigated high magnetic
field (7 T). The capacity of the dendrimeric CA to bind to the
target protein was confirmed by fluorescence measurements
upon its treatment with NeutrAvidin–agarose gel or NeutrAvi-
din-coated microspheres and the results were compared with
those of its monomeric analogue. The fluorescence intensity of
monomer-treated targets was found to be greater than that
from those treated with dendrimeric CA; however, a several-
fold increase in the MRI signal was observed on the same sam-
ples treated with the dendrimeric CA. The inductively coupled
plasma mass spectrometry analysis of the digested samples in-
dicated somewhat higher Gd3+ content and hence slightly
better binding of monomeric versus dendrimeric CA. This bi-
modal and multivalent targeted probe opens an avenue for
the preparation of new nanosized CAs that allow high-resolu-
tion MRI of various targets, such as cellular receptors or specif-
ic cellular populations.
Introduction
Magnetic resonance imaging (MRI) is a powerful diagnostic
and research tool that allows three-dimensional visualization of
living tissues at sub-millimeter resolutions. It offers excellent
soft-tissue contrast, allowing functionally relevant distinctions
between structures in the brain or other organs. Still, many
other important structures do not possess sufficient intrinsic
contrast to be discriminated from neighboring tissue. Methods
to enhance contrast in the local vicinity of specific structures
are needed.
quently employed are chelates of the paramagnetic lantha-
nide-metal ion gadolinium.[2] These agents shorten the spin–
lattice or longitudinal relaxation time (T1) of water protons in
a concentration-dependent manner, and are widely used in
clinical diagnostics and experimental research. However, the
available monomeric CAs have certain limitations. Typical low-
molecular-weight Gd chelates (e.g., 1,4,7,10-tetraazacyclodode-
cane-1,4,7,10-tetraacetic acid (DOTA), diethylenetriamine penta-
acetic acid (DTPA)) diffuse rapidly through the extracellular
space and distribute in tissues nonspecifically.[1c,3] Furthermore,
relatively high amounts of Gd3+ ions need to be localized to
the area of interest. Various studies suggest a detection limit of
approximately 107 Gd3+ ions per cell for receptor- or cell-label-
ing approaches.[4] The need for high local concentrations of
CA, combined with its rapid diffusion and elimination from the
area of interest, makes it very difficult to visualize specific tar-
gets with standard approaches.
MRI contrast can be enhanced by the introduction of exoge-
nous small-molecule contrast agents (CAs).[1] The most fre-
[a] Dr. S. Gündüz,+ Priv.-Doz. Dr. G. Angelovski
MR Neuroimaging Agents Group
Max Planck Institute for Biological Cybernetics
72076 Tübingen (Germany)
E-mail: goran.angelovski@tuebingen.mpg.de
[b] Dr. S. Gündüz,+ Dr. A. Power,+ Prof. Dr. N. K. Logothetis
Department for Physiology of Cognitive Processes
Max Planck Institute for Biological Cybernetics
72076 Tübingen (Germany)
The development of targeted CAs that recognize and bind
specific biomolecules with high affinity could dramatically in-
crease their local concentration and tissue retention time.[5] To
improve detection, the contrast of a cell-targeting CA could be
amplified by equipping it with a greater number of Gd3+ ions.
Targeted CAs with enhanced contrast-generating potential
would be valuable for diagnostic studies at the cellular and
molecular level. In neuroimaging, they would be of particular
use for imaging specific components of complex brain net-
works undergoing experience-dependent, developmental, or
pathological changes.[6] Increasingly, multimodal imaging
agents that can be detected by both optical and magnetic res-
[c] Prof. Dr. M. E. Maier
Institute for Organic Chemistry
Faculty of Science, University of Tübingen
72076 Tübingen (Germany)
[d] Prof. Dr. N. K. Logothetis
Department of Imaging Science and Biomedical Engineering
University of Manchester
Manchester M13 9PT (UK)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cplu.201402329.
ChemPlusChem 2015, 80, 612 – 622
612
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
onance (MR) methods are
sought after, which extends the
spatial scale of imaging into the
cellular and subcellular range.
Herein, we report the develop-
ment of a multimodal CA that
accomplishes
a twofold goal.
First, it targets a protein-based
receptor with a very high bind-
ing affinity. Second, it follows
a strategy designed to improve
detection of targeted CA by in-
creasing the number of Gd3+
chelates per molecule. We have
therefore prepared two tripar-
tite, multimodal, and targeted
CAs (MT3-CAs). Both consist of
Scheme 1. Synthesis of DOTA-type chelators. Reaction conditions: i) tert-butyl 2-bromo-4-(4-nitrophenyl)butanoate,
K2CO3, DMF, 458C, 16 h, 96%; ii) H2, Pd/C, EtOH, RT, 16 h, 95%; iii) CSCl2, Et3N, RT, 2 h, 71%; iv) formic acid, 608C,
16 h, 94%; v) GdCl3·6H2O, RT, 24 h, 41%.
a tripartite lysine core, which links the high-affinity targeting
ligand biotin to one arm and the optically detectable fluores-
cein isothiocyanate (FITC) moiety to another. The third arm car-
ries either a single, monomeric Gd3+ chelate (mMT3-CA) or
multiple Gd3+ chelates linked through a dendrimeric carrier
(dMT3-CA). The biotin ligand was incorporated as it binds to
the protein avidin with one of the highest affinities of any
known small-molecule–protein interaction.[7] FITC is a fluores-
cent dye commonly used for in vitro characterization of novel
CAs by means of optical imaging methods.[8] Finally, a highly
branched dendrimer capable of carrying up to 32 Gd3+ che-
lates was incorporated, thereby allowing MRI signal amplifica-
tion.[9]
version of 2 with thiophosgene and was used in further syn-
thetic procedures for obtaining monomeric and dendrimeric
CAs.[10] For comparison, a Gd3+ complex of the chelator bear-
ing the nitro group was also prepared. In this case the tert-
butyl esters in 1 were hydrolyzed with formic acid to give the
acid 4, and the complex 5 was prepared by using GdCl3·6H2O
in water (Scheme 1).
Synthesis of monomeric contrast agent
To compare the properties of the dendrimeric CA, we also syn-
thesized an analogous monomeric version mMT3-CA, which
possesses an identical targeting moiety (d-biotin) and fluores-
cent dye (FITC). The synthetic route was established starting
from lysine as the core molecule, which enables branching in
three directions and coupling of three different moieties under
various coupling conditions (Scheme 2). Ne-Boc-l-lysine methyl
ester hydrochloride (H-Lys(Boc)-OMe; Boc=tert-butoxycarbon-
yl) was biotinylated using N’-(3-dimethylaminopropyl)-N-ethyl-
carbodiimide (EDC) as a coupling reagent to give 6. To enable
FITC coupling through its isothiocyanate group, the methyl
ester in 6 was converted to an amide with an excess of ethyl-
enediamine in methanol to result in the primary amine 7. FITC
coupling afforded 8, whereas the primary amine 9 was pre-
pared from 8 by removing the Boc group using concentrated
hydrogen chloride solution in methanol.
Two different avidin-displaying targets, agarose beads and
polystyrene microspheres, were used to assess specific binding
of MT3-CAs. Relaxometric, fluorescence, MRI, and inductively
coupled plasma mass spectrometry (ICP-MS) measurements
were performed to compare the potential of mMT3-CA and
dMT3-CA as targeted imaging agents.
Results
Synthesis of monomeric and dendrimeric targeted contrast
agents
Synthesis of DOTA-type chelators and their Gd3+ complexes
To ensure a high stability of the Gd3+ complex for application
of successive monomeric and dendrimeric CAs, we prepared
a DOTA-type chelator linked to an isothiocyanate group that
easily reacts with primary amines. The chosen chelator con-
tains four carboxylic groups that form a strong complex with
Gd3+, which results in a negatively charged complex, an ad-
vantageous property in terms of water solubility and the over-
all charge of the dendrimeric targeted CA.
The biotin- and FITC-containing intermediate 9 was then
used as a starting material in the synthesis of both the mono-
meric and dendrimeric CAs (see below). For the monomeric
CA, the macrocyclic isothiocyanate 3 was directly coupled to 9
in basic medium to give 10, whereas the tert-butyl esters were
hydrolyzed with formic acid to afford 11. Its complexation with
Gd3+ at neutral pH resulted in mMT3-CA, the monomeric CA
that was used in the subsequent comparison studies with the
dendrimeric analogue.
Preparation of the desired chelator started from tri-tert-butyl
1,4,7,10-tetraazacyclododecane-1,4,7-triacetate
(DO3A-tBu
ester), which was alkylated with tert-butyl 2-bromo-4-(4-nitro-
phenyl)butanoate to yield 1 (Scheme 1). The nitro group in
1 was reduced by palladium-catalyzed hydrogenation to give
the aniline 2. The isothiocyanate 3 was finally obtained by con-
Synthesis of dendrimeric contrast agent
For the synthesis of a target-specific dendrimeric CA we used
the biotinylated intermediate 9 (see above) and a modified
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
613
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Scheme 2. Synthesis of biotinylated and fluorescently labeled monomeric targeted CA (mMT3-CA) and the reactive intermediate 12. Reaction conditions:
i) EDC·HCl, biotin, DMAP, DMF, RT, 48 h, 64%; ii) ethylenediamine, MeOH, RT, 24 h, 77%; iii) FITC, Et3N, DMF, RT, 16 h, 88%; iv) concd HCl, MeOH, RT, 1 h, 97%;
v) 3, Et3N, DMF, 458C, 16 h, 68%; vi) formic acid, 608C, 24 h, 95%; vii) GdCl3·6H2O, pH 7.0, RT, 24 h, 51%; viii) SMCC, H2O/DMF, pH 7.5–9.0. DMAP=N,N-dimethyl-
aminopyridine.
generation 4 (G4) poly(amidoamine) (PAMAM) dendrimer with
the cystamine core.[11] The latter allows the controlled coupling
of two different functionalities at the dendrimer, the first one
through the amino groups on the dendrimer surface, and the
second through the sulfhydryl group that can be formed by
the reduction of cystamine. This type of dendrimer has already
been used for the preparation of biotin–dendrimer conjugates,
however with less stable acyclic chelator for Gd3+ and a lower
dendrimer generation (G2), which ultimately produces lower
signal enhancement.[12]
PAMAM G4 dendrimers.[14] The calculated masses correspond
to an average of 37 DOTA chelates per dendrimer molecule,
which corresponds to 55–60% conversion of the amino surface
groups into the thiourea product with the DOTA-NCS units.
1
The appearance of aromatic protons in the H NMR spectra is
clear evidence of the DOTA–dendrimer conjugate formation
1
(these peaks do not exist in H NMR spectra of the commercial
PAMAM G4 dendrimer); however, integration of spectra was
less successful than expected,[14] usually indicating a slightly
higher amount of DOTA units (when comparing regions of aro-
matic versus aliphatic protons) than amounts determined by
MALDI-TOF.
Intermediate 9 was first converted into the sulfhydryl-reac-
tive molecule 12 by coupling it with the succinimidyl-4-(N-
maleimidomethyl)cyclohexane-1-carboxylate (SMCC) linker.[13]
For this purpose, the N-hydroxysuccinimide (NHS) ester group
in the SMCC linker was reacted with the primary amine in 9
under controlled conditions (pH 7.5 and room temperature) to
give 12 (Scheme 2). In parallel, the macrocyclic Gd3+ chelators
(DOTA) were attached to the amino surface groups of the den-
drimer by using previously prepared DOTA-NCS ligand 3, thus
affording dendrimer 13 (Scheme 3). Tert-butyl esters in 13 were
hydrolyzed using formic acid to give water-soluble dendri-
mer 14. Both dendrimer products 13 and 14 were analyzed by
matrix-assisted laser desorption/ionization time-of-flight
The disulfide bond in dendrimer 14 was cleaved at pH 7.0
by
using
tris(2-carboxyethyl)phosphine
hydrochloride
(TCEP·HCl) to obtain dendron 15 (Scheme 3). To prevent recou-
pling of the thiol compound, maleimide 12 was immediately
added to the reaction mixture.[11] The maleimide group of the
SMCC linker reacts with the thiol group of 15 in a slightly
acidic medium to give 16,[13] therefore the pH was maintained
at 6.5–7.0 during the reaction. The unreacted dendron 15 was
reoxidized by bubbling oxygen gas through the reaction mix-
ture to give dendrimer 14. The excess amounts of 12 and 14
were removed by using a G-15 Sephadex column and water as
eluent. After the purification, the complexation of 16 with
GdCl3·6H2O at neutral pH resulted in the dendrimeric CA
dMT3-CA. The excess of Gd3+ ions was removed using ethyle-
1
(MALDI-TOF) mass spectrometry (MS) and H NMR spectrosco-
py. MALDI-TOF spectra exhibited broad signals with patterns
similar to those previously reported for DTPA-functionalized
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
614
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Scheme 3. Synthesis of the dendrimeric MRI CA dMT3-CA. Reaction conditions: i) 3, Et3N, DMF, 458C, 48 h, 81%; ii) formic acid, 608C, 24 h, 84%; iii) TCEP·HCl,
pH 7.0, RT, 1 h; iv) 12, pH 6.5–7.0, RT, 16 h, 87%; v) GdCl3·6H2O, pH 7.0, RT, 24 h, 79%.
nediaminetetraacetic acid (EDTA), a common Gd3+ chelator.
The GdEDTA complex, excess of EDTA, and other impurities
were finally removed by using a centrifugal filter unit with
3 kDa molecular weight cutoff (MWCO) and the dendrimeric
complex dMT3-CA was used in subsequent studies. The den-
drons 16 and dMT3-CA were also analyzed by means of
MALDI-TOF MS showing again broad peaks with an average
number of 19 DOTA chelators attached to the PAMAM G4 den-
dron, which is completely in line with results observed for the
initial dendrimers 13 and 14 that have double the number of
DOTA units (see above).
(PBS) at different concentrations of Gd3+, prepared by diluting
the stock solutions of 5, mMT3-CA, and dMT3-CA with appro-
priate amounts of PBS. The exact concentration of Gd3+ in
stock solutions was determined by the bulk magnetic suscepti-
bility shift method.[15] The r1 relaxivity of each CA was obtained
as a slope of the linear regression curve for concentrations of
Gd3+ plotted against relaxation rate (R1). The r1 values of
(4.78Æ0.10), (6.95Æ0.10), and (6.08Æ0.06) mmÀ1
s
À1 (calculated
per mm of Gd3+) were determined for 5, mMT3-CA, and
dMT3-CA, respectively.
In summary, two new multimodal, targeted CAs were syn-
thesized by coupling either a single (mMT3-CA) or dendrimer-
linked (dMT3-CA) Gd3+ chelate, FITC, and biotin around a tri-
partite core. The loading efficiency of the dendrimer carrier
was 60%, thus generating a CA with 19 Gd3+ chelates per mol-
ecule.
Binding of targeted contrast agents to NeutrAvidin–agarose
gel
To test the performance of the synthesized CAs, NeutrAvi-
dinꢁ–agarose gel was used as an in vitro model of target bio-
logical tissue. This material consists of NeutrAvidin, a variant of
the avidin protein, immobilized onto beaded agarose. Solu-
tions containing 1 mm of the targeted CA, as well as a commer-
cially available nontargeted CA Dotaremꢂ, were incubated with
the beads. A fourth sample was incubated with PBS containing
no CA, to control for any background signal resulting from the
beads alone. The beads were washed to remove excess CA
Longitudinal relaxivity of monomeric and dendrimeric CAs
Longitudinal r1 relaxivity values of 5, mMT3-CA, and dMT3-CA
were determined at 7 T (300 MHz) and 258C. The T1 relaxation
times were recorded at pH 7 in phosphate-buffered saline
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
615
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
and then analyzed by fluorescence measurements, MRI, and
ICP-MS. The experiments were repeated in triplicate.
beads or beads treated with Dotarem. Quantification of fluo-
rescence signals in a multiplate reader revealed a similar trend.
The intensity of beads treated with monomeric CA (mMT3-CA)
was 3.6-fold greater than that for beads treated with the den-
Fluorescence measurements and images
Contrast agent binding was verified by checking the beads for
fluorescence of the FITC moiety (lex =485 nm, lem =520 nm),
which is present on both of the targeted CAs but not Dotarem.
Beads treated with the targeted CA were fluorescent when
placed on a blue light source and viewed through an amber
filter (Figure 1, top). The beads that had been treated with the
monomeric CA appeared brighter than those treated with the
dendrimer. No fluorescence was observed with untreated
Table 1. Analysis of CA binding to NeutrAvidin–agarose gel.[a]
Sample
Fluorscence
intensity
[a.u.][a]
T1w MRI
intensity
(norm.)[b]
Relative Gd3+
content[a,b]
Ratio Gd3+[a,c]
PBS
–
–
1.0
1.1
1.9
9.6
–
1.0
11.2
151.5
–
–
–
13.5
Dotarem
mMT3-CA
dMT3-CA
52965
14891
[a] Fluorescence intensity; reported results are means of three independ-
ent experiments. [b] As determined by ICP-MS. [c] Ratio of Gd3+ content
of dendrimer/monomer (dMT3-CA/mMT3-CA).
drimer (dMT3-CA; Figure 1, bottom, and Table 1). As expected,
fluorescence intensities of both PBS- and Dotarem-treated
beads were below background level. Overall, the fluorescence
studies indicate that both targeted CAs bind efficiently and
specifically to the target protein avidin. The quantitative data
suggest a possible greater binding efficiency for the monomer-
ic agent in comparison to the dendrimeric CA.
T1-weighted magnetic resonance imaging
Eppendorf tubes containing PBS-, Dotarem-, monomer-, or
dendrimer-treated NeutrAvidin–agarose beads were imaged in
the MRI scanner operating at 7 T. A set of T1-weighted (T1w)
images was generated using a series of 30 repetition times
(TRs) ranging from 21 to 8000 ms. For each set of parameters
we acquired transverse and sagittal plane images at an in-
plane resolution of 200 mm and slice thickness of 2 mm (Fig-
ure 2a,b). We visually inspected the set of transverse images,
and performed region of interest (ROI) analysis (for details see
Figure 1. Fluorescence intensities of CA-treated NeutrAvidin–agarose gel
measured in a microplate reader with excitation/emission filters appropriate
for the FITC fluorophore (lex =485 nm/lem =520 nm). Bar plots show the
mean intensities of beads of three experiments Æ standard deviation (SD).
Top inset: fluorescence images of beads treated with PBS, Dotarem, mMT3-
CA, and dMT3-CA.
Figure 2. Representative transverse and sagittal plane T1w MR images of NeutrAvidin–agarose gel treated with PBS, Dotarem, mMT3-CA (Monomer), or dMT3-
CA (Dendrimer). Acquisition parameters: B0 =7 T, field of view=4”4 cm (a) or 4”6 cm (b), matrix 200”200 (a) or 200”300 (b), slice thickness=2 mm, echo
time TE=6 ms, TR=58.5 ms. c) Plot showing the quantified T1w signal intensity for each CA sample calculated from a dataset of images similar to (a), aver-
aged from three replicate experiments and normalized to the PBS control. Bars indicate Æ1 SD.
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
616
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
the Experimental Section) to
quantify the average pixel inten-
sity of each sample. The signal
intensities of the CA-treated
samples varied depending on
the TR used for image acquisi-
tion. The greatest signal change
was found with the dendrimeric
CA at TR=59 ms; therefore, fur-
ther analysis was performed on
this dataset. Binding of the den-
drimeric agent caused a nearly
tenfold increase in bead signal
intensity (Figure 2c). Binding of
the monomer also led to a no-
ticeable enhancement (about
twofold). Dotarem treatment
caused only a marginal increase
Figure 3. Microscopic analysis of 2 mm NeutrAvidin-coated polystyrene microspheres treated with PBS, FITC,
mMT3-CA (Monomer), or dMT3-CA (Dendrimer). Left image shows the beads under bright-field illumination; right
image with excitation/emission filters appropriate for the detection of the FITC fluorophore (lex =485 nm/
lem =520 nm).
(1.1-fold), which suggests a small degree of nonspecific bead
retention of this nontargeted CA. In summary, binding of both
monomeric and dendrimeric CA increased target contrast in
T1w images. Comparatively, the dendrimeric CA showed an
overall increase in performance of approximately fivefold over
the monomer.
geted MRI CAs (mMT3-CA and dMT3-CA) can additionally be
used for imaging at the microscopic scale.
Discussion
The effectiveness of the developed synthetic strategy and the
properties of the obtained products can be scrutinized
through several aspects studied in the experiments described
above. We synthesized bimodal, multivalent targeted CAs that
can be detected through both optical fluorescence and MRI
contrast signals. The synthetic route is based on the building
block model, which allows easy replacement of the individual
moieties. This strategy would allow easy modifications and po-
tentially further improvements of the function of our targeted
CAs. In this study we have initially coupled the core lysine mol-
ecule to biotin as a prototype ligand that binds specifically to
the protein avidin. Depending on the desired target receptor,
our approach allows implementation of versatile ligands that
ensure the binding specificity.
Inductively coupled plasma mass spectrometry analysis
The samples containing NeutrAvidin–agarose gel treated with
PBS, Dotarem, mMT3-CA, or dMT3-CA were digested in con-
centrated HNO3 to obtain the Gd3+ in the ionic form and its
amount was determined by ICP-MS. The ICP-MS data were
consistent with observations derived from the fluorescence
and MRI studies (Table 1). Gadolinium was detectable in the
three bead samples treated with Dotarem, monomer (mMT3-
CA), or dendrimer (dMT3-CA). The nonspecific nature of the
Dotarem interaction was indicated by much lower bead-associ-
ated Gd3+ in comparison to either of the targeted agents
mMT3-CA and dMT3-CA. The beads treated with dMT3-CA
were associated with approximately 14 times more Gd3+ rela-
tive to monomeric CA (Table 1). This is consistent with the den-
drimer’s determined approximately 19:1 content of GdDOTA/
biotin but slightly lower binding to the NeutrAvidin–agarose
gel (indicated by fluorescence experiments, see above).
Furthermore, we used a standard fluorescent dye FITC,
which allows assessment of the targeted CA also by means of
optical methods. Replacement of the dye with ones having
more convenient optical properties that allow more thorough
in vitro studies with cell cultures can now be performed in
a straightforward and selective fashion. This is an additional
advantage of our synthetic approach: common procedures
with the multivalent carrier (dendrimers, nanoparticles) often
involve nonspecific coupling of fluorescent dyes to the carrier
surface without stoichiometric control. Herein, we report a pro-
cedure that ensures coupling of a single fluorescent unit to
the desired targeted molecule, which can be very advanta-
geous for various quantitative fluorescence assays.
Testing targeted CA for microscopic analysis of NeutrAvidin-
coated microspheres
The functionality of the targeted CA was further assessed by
fluorescence microscopy after binding to 2 mm spherical poly-
styrene microspheres displaying the target protein NeutrAvi-
din. The microspheres could be observed with a FITC-appropri-
ate filter set following treatment with either of the targeted
CAs (mMT3-CA or dMT3-CA) but not when treated with the
nontargeted FITC fluorophore (Figure 3). Similar to other ex-
periments described above, the fluorescence intensity of the
monomer-treated targets was greater than that for those treat-
ed with the dendrimer. These data indicate that these two tar-
Finally, in this approach we used G4 PAMAM dendron with
maximally 32 surface coupling groups. Increasing the initial
dendrimer generation can certainly lead to further MRI signal
amplification. Targeted nanosized products with two different
types of dendrons could be also prepared by following this
strategy;[16] for instance, although a higher-generation dendron
is used for MRI signal amplification, a lower-generation den-
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
617
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
dron could carry several ligands to improve the binding effica-
cy of the CA to its target.
Based on considerations discussed above, both CAs are ex-
pected to increase the MRI contrast of materials displaying the
target protein. Greater gains are predicted for the dendrimeric
agent as it carries multiple Gd3+ chelates per molecule and
has a higher molecular relaxivity at relevant, high field
strengths used in experimental research.
The longitudinal r1 relaxivities of all the studied Gd com-
plexes (5, mMT3-CA, and dMT3-CA) are in line with previously
studied low-, middle-, and high-molecular-weight complexes at
high magnetic fields. As expected, a low-molecular-weight
complex 5 has the lowest r1 relaxivity. Despite the increase in
their size, monomeric and dendrimeric CAs do not exhibit
a dramatic increase in the r1, which is probably a result of the
magnetic field applied in this experiment (7 T).[17] Nevertheless,
the r1 ꢀ114 mmÀ1 sÀ1 for dMT3-CA when calculated per dendri-
mer molecule (approximately 19 monomeric units with each r1
Extending the applications of the CAs described here, studies
with 2 mm polystyrene microspheres suggest that they can also
be used to image the target structure by using fluorescence
microscopy. Therefore they could be used as true multimodal
agents with improved range of spatial resolution. This is a valua-
ble property as the features of MRI imaging (whole-body pene-
tration, micrometer to centimeter spatial resolution, noninva-
sive) are highly complementary to those of fluorescence micro-
scopic imaging (limited tissue penetration, nanometer to micro-
meter resolution, performed on fixed histological specimens).
The use of a single multimodal agent combining both optical
and MRI signaling capabilities allows for experiments employ-
ing both imaging approaches simultaneously.
ꢀ6 mmÀ1
s
À1) suggests this CA could indeed be useful and uti-
lized for signal amplification in targeted MRI approaches.
The targeting properties of the synthesized monomeric and
dendrimeric CAs were studied by using various methods and
types of targets. The agents exploit the high-affinity biotin–
avidin interaction and were found to stably label avidin-coated
microspheres and NeutrAvidin–agarose gel beads. The fluores-
cence microscopy experiments qualitatively exhibit stronger
fluorescence intensity by monomeric mMT3-CA than dendri-
meric dMT3-CA. Similarly, the fluorescence spectroscopy meas-
urements resulted in 3.6-fold stronger fluorescence intensity of
monomeric CA. Although a direct quantitative comparison of
the binding cannot be performed because the emission prop-
erties of dye in the targeted molecule may change upon inter-
action with the protein, we conclude that mMT3-CA binds
more efficiently than dMT3-CA most likely owing to the small-
er size of the former.
The NeutrAvidin-coated polystyrene microspheres can also
be considered as a good model of cells with a higher content
of avidin expressed on their surface. Consequently, the specific
binding of our targeted CAs envisages a number of diverse ap-
plications for these probes to investigate living systems. In one
possible implementation, the polypeptide target avidin could
be encoded as a transgene, manipulated using molecular biol-
ogy, and introduced into living cells of model organisms. As
shown previously, a fusion between avidin and a membrane-
associated protein can be generated for expression on the
outer surface of the cell as a receptor for targeted agents.[18] By
deploying genetically engineered target receptors with the
available strategies for spatiotemporal control of transgene ex-
pression (e.g., tissue-specific or inducible promoters), the cor-
responding CA could be used to study the development, phys-
iology, or pathology of specific cell types and tissues in multi-
modal high-field MRI/optical experiments.
On the other hand, we were able to show that MRI efficacy
of the targeted CA can be substantially enhanced by generat-
ing a multivalent dendrimeric structure decorated with almost
20 Gd3+ ions per molecule. Despite weaker binding of dMT3-
CA owing to its size, we observed an approximately tenfold
MRI signal enhancement relative to a nonspecific and commer-
cially available CA, Dotarem, and an approximately fivefold
signal increase compared with mMT3-CA. These results can
potentially be improved by increasing the dendron generation,
incorporation of a longer spacer between the ligand and the
rest of the dendrimeric molecule, or another dendron with in-
creased valency for ligands. However, we note that increase in
the MRI signal is not directly proportional to the local increase
of Gd3+ concentration (achieved through higher dendrimer va-
lency or binding efficiency), since the background MR signal
also plays an important role and contributes to the recorded
signal.
Conclusion
A new synthetic strategy for the preparation of target-specific
multivalent contrast agents (CAs) has been developed. A bimo-
dal biotinylated dendrimeric CA was prepared and its proper-
ties were compared with those of its monomeric analogue.
Fluorescence-based studies proved the efficient binding of
both monomeric and dendrimeric CAs to avidin-coated micro-
spheres and agarose gel beads. Magnetic resonance imaging
(MRI) experiments with avidin-coated agarose gel beads show
remarkable signal enhancement of dendrimeric CAs, thus pro-
viding a promising application outlook for targeted CAs based
on this approach. Used alone or in combination with appropri-
ate nanosized carriers, we foresee new attempts in MRI of vari-
ous cellular receptors and consequently specific cellular popu-
lations. This could lead to our better understanding of essen-
tial physiological and pathological processes and would be
a great asset to the current clinical diagnostics or neuroimag-
ing for understanding brain function.
The results obtained from ICP-MS experiments could actually
be the most directly linked to binding efficiency of our target-
ed CAs, thereby resolving the potential confusion generated
by the discrepancy in signal quantifications obtained from
fluorescence spectroscopy and MRI experiments. The Gd3+
amounts determined in various samples undoubtedly show
a specific binding of mMT3-CA and dMT3-CA to NeutrAvidin–
agarose gel, over Dotarem. Furthermore, the 13.5:1 ratio of
Gd3+ content of dendrimer/monomer (dMT3-CA/mMT3-CA)
suggests a less than two times weaker binding of dendrimer,
given its 19 times higher Gd3+ content over the monomer.
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
618
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
catalyst (300 mg) was added to the solution. The heterogeneous
mixture was shaken for 16 h under a hydrogen atmosphere (30 psi)
in a Parr hydrogenator apparatus. The catalyst was removed by fil-
tration through Celite. The solvent was evaporated to obtain
Experimental Section
General remarks
All commercially available chemicals were used without further pu-
rification. DO3A-tBu ester was purchased from Click Chemistry
Technology, Beijing, China. d-Biotin and Ne-Boc-l-lysine methyl
ester hydrochloride were purchased from Sigma–Aldrich, Germany.
FITC was purchased from Biomol, Germany. SMCC was purchased
from Shangai YFan Chemistry Co., Ltd., Shanghai, China. Cys-
PAMAM-G4 was purchased from Dendritic Nanotechnologies Inc.,
USA. Tert-Butyl 2-bromo-4-(4-nitrophenyl)butanoate was prepared
according to the literature procedure.[19] FluoSpheres NeutrAvidin-
labeled microspheres, 1.0 mm, nonfluorescent, were obtained from
Molecular Probes Inc., Eugene, Oregon, USA. NeutrAvidin high-ca-
pacity agarose resin was purchased from Pierce Biotechnology,
Rockford, USA.
a
brown amorphous solid compound (2.74 g, 95%). 1H NMR
(300 MHz, CDCl3): d=6.86 (d, J=8.2 Hz, 2H; ArH), 6.59 (d, J=
8.3 Hz, 2H; ArH), 3.55–1.65 (overlapping m, 27H; CH2 cyclen ring,
ArCH2CH2CHCOOtBu and CH2COOtBu), 1.63–1.25 ppm (br, 36H;
(CH3)3C); 13C NMR (75 MHz, CDCl3): d=175.6, 172.8, 172.6 (C(O)),
145.1, 120.0, 129.6, 115.1 (ArC), 81.8, 81.7, 81.6 ((CH3)3C), 58.2
(CHCOOtBu), 55.7, 55.4, 52.5, 52.4, 51.6, 48.5, 48.3, 48.1, 47.0, 44.4
(CH2 cyclen ring), 33.1 (ArCH2CH2), 28.2 (ArCH2CH2), 27.9, 27.8,
+
26.1 ppm ((CH3)3C); FT-ICR/HRMS: m/z calcd for C40H70N5O8
748.5219 [M+H]+; found: 748.5224.
:
4-(4-Isothiocyanatophenyl)-2-(4,7,10-tris-tert-butoxycarbonylmethyl-
1,4,7,10-tetraazacyclododec-1-yl)butyric acid tert-butyl ester (3):
Thiophosgene (7.83 mmol, 0.40 mL) was added to a mixture of 2
(2.67 mmol, 2.00 g) and triethylamine (1 mL) in dichloromethane
(20 mL). The reaction mixture was stirred at room temperature for
24 h. The solvent was removed under reduced pressure, the crude
product was purified by column chromatography (silica gel, 10%
methanol/dichloromethane), and the product 3 was obtained as
a light brown amorphous solid (1.49 g, 71%). 1H NMR (300 MHz,
CDCl3): d=7.07 (s, 2H; ArH), 3.51–1.56 (overlapping m, 27H; CH2
cyclen ring, ArCH2CH2CHCOOtBu and CH2COOtBu), 1.55–1.28 ppm
(br, 36H; (CH3)3C); 13C NMR (75 MHz, CDCl3): d=175.1, 172.8, 172.7
(C(O)), 140.4, 137.8, 129.9, 129.3, 125.6 (ArC), 82.3, 81.9, 81.7, 81.6
((CH3)3C), 58.8 (CHCOOtBu), 55.8, 55.5, 55.4, 52.5, 52.4, 51.9, 48.5,
48.2, 47.1, 44.6 (CH2 cyclen ring), 34.0 (ArCH2CH2), 27.9, 27.8,
25.8 ppm; FT-ICR/HRMS: m/z calcd for C41H68N5O8S+: 790.4788
[M+H]+; found: 790.4777.
1H and 13C NMR spectroscopy and relaxometric experiments were
performed on a Bruker Avance III 300 MHz spectrometer, Germany.
Fourier transform ion cyclotron resonance (FT-ICR) MS analysis was
performed by a Bruker FT-ICR Apex II spectrometer, Germany. Elec-
trospray ionization (ESI)-TOF-MS experiments were performed on
MAXIS 3G spectrometer, Bruker Daltonics Inc., Germany. MALDI-
TOF MS analysis was performed by The Scripps Center for Mass
Spectrometry, La Jolla, CA, USA. ICP-MS experiments were per-
formed by Currenta GmbH & Co. OHG, Leverkusen, Germany. Fluo-
rescence microscopic imaging of microspheres was performed
with an AxioImager.Z1 upright microscope (Zeiss, Germany)
equipped with an AxioCamHR3 camera. Fluorescence imaging of
agarose gel was performed using a SafeImager 2.0 blue (470 nm)
LED source transilluminator (Life Technologies) and the supplied
amber emission filter. Fluorescence intensity measurements were
performed on a Fluostar Optima Plate Reader (BMG Labtech, Ger-
many). The MR images were acquired at 218C using a 7 T Bruker
4-(4-Nitrophenyl)-2-(4,7,10-tris-carboxymethyl-1,4,7,10-tetraaza-
cyclododec-1-yl)butyric acid (4): The nitrobenzene derivative 1
(0.13 mmol, 100 mg) was dissolved in formic acid and the mixture
was stirred at 608C for 16 h. Formic acid was evaporated under re-
duced pressure and the product was dried under high vacuum to
give 4 as a light brown amorphous solid (67 mg, 94%). 1H NMR
(300 MHz, CDCl3): d=8.16 (d, J=8.4 Hz, 2H; ArH), 7.48 (d, J=
8.4 Hz, 2H; ArH), 4.04–2.64 (overlapping m, 25H; CH2 cyclen ring,
ArCH2CH2CHCOOtBu and CH2COOtBu), 2.18–1.82 ppm (br, 2H;
ArCH2CH2CHCOOtBu); ESI-LRMS: m/z calcd for C24H36N5O10+: 554.2
[M+H]+; found: 554.2.
1
BioSpec 70/30 USR Preclinical MRI system with a H/volume trans-
mit–receive coil tuned to the proton resonance (300.3 MHz).
Synthetic procedures
4-(4-Nitrophenyl)-2-(4,7,10-tris-tert-butoxycarbonylmethyl-1,4,7,10-
tetraazacyclododec-1-yl)butyric acid tert-butyl ester (1): A suspen-
sion of tri-tert-butyl 2,2’,2’’-(1,4,7,10-tetraazacyclododecane-1,4,7-
triyl)triacetate (1.54 g, 3.02 mmol), tert-butyl 2-bromo-4-(4-nitro-
phenyl)butanoate (1.35 g, 3.92 mmol), and potassium carbonate
(1.04 g, 7.54 mmol) in dimethylformamide (DMF, 10 mL) was pre-
pared and stirred at 458C for 16 h. Dimethylformamide was re-
moved under reduced pressure. The residue was dissolved in di-
chloromethane (200 mL) and extracted with water (2”200 mL). Di-
chloromethane was evaporated under reduced pressure and
a brown amorphous solid product (2.25 g, 96%) was obtained by
column chromatography (silica gel, 10% methanol/dichlorome-
Complex 5: Macrocycle 4 (40 mg, 0.072 mmol) was dissolved in
water and the pH was adjusted to 7.0 with aqueous sodium hy-
droxide (1m). GdCl3·6H2O was added to the solution and the pH
was maintained at 7.0 with the sodium hydroxide solution. The
mixture was stirred at room temperature for 24 h. Chelex was
added to the reaction mixture to remove excess Gd3+ ions. After
filtration to remove Chelex, water was removed under reduced
1
pressure to give a light brown solid product 5 (21 mg, 41%). ESI-
thane). H NMR (300 MHz, CDCl3): d=8.12 (d, J=8.3 Hz, 2H; ArH),
À
TOF/HRMS: m/z calcd for C24H31GdN5O10
found: 707.1325.
:
707.1317 [MÀH]À;
7.34 (d, J=8.5 Hz, 2H; ArH), 3.61–1.59 (overlapping m, 27H; CH2
cyclen ring, ArCH2CH2CHCOOtBu and CH2COOtBu), 1.53 (s, 9H;
(CH3)3C), 1.43 ppm (br, 27H; (CH3)3C); 13C NMR (75 MHz, CDCl3): d=
174.9 172.8, 170.4 (C(O)), 149.1, 146.5, 129.6, 123.6 (ArC), 82.4, 81.8
((CH3)3C), 59.3 (CHCOOtBu), 55.8, 55.5, 52.6, 52.4, 52.0 (CH2 cyclen
ring), 34.5 (ArCH2CH2), 28.1 (ArCH2CH2), 27.9, 27.8, 27.8 ppm
((CH3)3C); FT-ICR/HRMS: m/z calcd for C40H68N5O10
[M+H]+; found: 778.4957.
6-tert-Butoxycarbonylamino-2-[5-(2-oxohexahydrothieno[3,4-d]imid-
azol-6-yl)pentanoylamino]hexanoic acid methyl ester (6): d-Biotin
(3.37 mmol, 0.82 g), EDC·HCl (5.05 mmol, 0.97 g), and DMAP
(8.42 mmol, 1.03 g) were dissolved in dimethylformamide (25 mL);
H-Lys(Boc)-OMe (3.37 mmol, 1.00 g) was added to the mixture. The
reaction mixture was stirred under nitrogen at room temperature
for 48 h. The solvent was evaporated and the residue was poured
into water (100 mL) and extracted with dichloromethane (3”
100 mL). The collected organic phases were dried over sodium sul-
+
: 778.4961
4-(4-Aminophenyl)-2-(4,7,10-tris-tert-butoxycarbonylmethyl-1,4,7,10-
tetraazacyclododec-1-yl)butyric acid tert-butyl ester (2): The nitro-
benzene derivative 1 was dissolved in ethanol (10 mL) and Pd/C
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
619
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
fate and evaporated under reduced pressure. The crude product
was purified by column chromatography (silica gel, 10% metha-
nol/dichloromethane) to give 6 as a colorless amorphous solid
(1.05 g, 64%). 1H NMR (300 MHz, CDCl3): d=4.86–4.68 (br, 1H;
SCH2CHNH), 4.57–4.40 (br, 1H), 4.37–4.22 (br, 1H), 3.71 (s, 3H;
NHCHCOOCH3), 3.19–2.97 (br, 2H; CH2NHCOOC(CH3)3), 2.90 (dd, J=
12.8, 4.8 Hz, 1H; SCH2CHNH), 2.71 (d, J=12.8 Hz, 1H; SCH2CHNH),
(br, 3H; SCHCH and CH2NHCOOC(CH3)3), 3.07–2.85 (br, 1H; SCH),
2.81–2.63 (br, 2H; SCH2), 2.45–2.23 (br, 2H; CH2CONHCHCONH),
2.09–1.38 ppm (br, 12H; -CH2-); 13C NMR (75 MHz, CD3OD): d=
183.2 (C(S)), 176.6, 175.1, 172.7, 167.8 (C(O)), 160.7, 143.7, 134.4,
132.7, 131.6, 127.9, 126.7, 121.2, 118.4, 103.5 (ArC), 63.9, 62.4, 57.1,
55.2, 44.9, 41.1, 40.6, 40.5, 36.5, 34.7, 32.4, 29.8, 29.7, 29.5, 28.2,
26.8, 26.0, 24.2 ppm (-CH- and -CH2-); ESI-TOF-MS: m/z calcd for
C39H46N7O8S2+: 804.2844 [M+H]+; found: 804.2839.
2.24
(br,
2H;
CH2C(O)NH),
1.86–1.56
(br,
6H;
CH2CH2CH2C(O)NHCHCH2), 1.54–1.25 ppm (br, 15H; SCHCH2 and
CH2CH2CH2NHCOOC(CH3)3); 13C NMR (75 MHz, CDCl3): d=173.9,
173.5, 164.4, 156.1 (C(O)), 79.0 ((CH3)3C), 61.9, 60.2, 55.9 (CH cyclen
ring), 52.2 (NHCHCOOCH3), 52.0 (NHCHCOOCH3), 40.4, 40.1, 35.4,
31.4, 29.5 (-CH2-), 28.4 ((CH3)3C), 28.2, 27.9, 25.4, 22.7 ppm (-CH2-);
ESI-TOF/HRMS: m/z calcd for C22H38N4NaO6S+: 509.2404 [M+Na]+;
found: 509.2398.
Compound 10: The amine 9 (0.12 mmol, 0.10 mg), 3 (0.19 mmol,
0.15 mg), and triethylamine (0.36 mmol, 0.05 mL) were dissolved in
dimethylformamide (5 mL); the reaction mixture was stirred at
458C for 16 h. The solvent was evaporated and the crude product
was purified by column chromatography (silica gel, 20% metha-
nol/dichloromethane) to give the orange product 10 (0.20 g, 68%).
1H NMR (300 MHz, CD3OD): d=8.27 (br, 1H; ArH), 7.55–6.94 (br,
6H; ArH), 6.92–6.38 (br, 6H; ArH), 4.51–4.12 (br, 2H; SCH2CHCH),
3.94–3.33 (br, 5H), 3.28–3.03 (br, 9H), 3.38–3.02 ppm (br, 20H);
13C NMR (75 MHz, CD3OD): d=183.0, 182.4 (C(S)), 177.1, 176.4,
175.2, 174.6, 171.2, 166.1, 162.0 (C(O)), 154.4, 142.3, 139.2, 138.7,
130.6, 129.5, 126.0, 125.7, 114.2, 111.8, 103.7 (ArC), 82.3, 82.9, 82.8
((CH3)3C), 63.3, 61.7, 59.5, 57.0, 56.9, 55.6, 55.5, 56.5, 55.2, 54.0, 53.9,
53.3, 45.4, 45.2, 41.2, 40.2, 40.1, 36.5, 32.7, 30.8, 29.8, 29.6, 29.4
(-CH- and -CH2-), 28.6, 28.5, 28.4 (C(CH3)3), 26.9, 24.4, 24.1 ppm
(-CH2-); ESI-TOF/HRMS: m/z calcd for C80H113N12NaO16S32+: 808.3723
[M+H+Na]2+; found: 808.3722.
{5-(2-Aminoethylcarbamoyl)-5-[5-(2-oxohexahydrothieno[3,4-d]imid-
azol-6-yl)pentanoylamino]pentyl}carbamic acid tert-butyl ester (7):
Biotinylated lysine 6 (1.00 g, 2.06 mmol) was dissolved in methanol
(10 mL); ethylenediamine (1 mL, 14.98 mmol) was added to the so-
lution. The reaction mixture was kept at room temperature for
24 h. The crude product was washed with ethyl acetate to remove
excess ethylenediamine and yield 7 as a colorless amorphous solid
(0.81 g, 77%). 1H NMR (300 MHz, CDCl3): d=4.69–4.06 (br, 3H;
SCH2CHCH and NHCHC(O)NH), 3.85–2.68 (br, 9H; CH2SCH,
NH2CH2CH2NH and CH2NHCOO), 2.45–2.00 (br, 2H; CH2(CO)NH),
1.97–1.13 ppm (br, 21H; -CH2- and (CH3)3C); 13C NMR (75 MHz,
CDCl3): d=176.3, 174.9, 174.4, 165.1 (C(O)), 62.1, 60.7, 55.1 (CH
cyclen ring), 54.5 (HNCHC(O)NH), 40.2, 40.0, 39.9, 37.2, 34.9, 30.7
(-CH2-), 29.1 ((CH3)3C), 28.3, 27.5, 25.2, 22.6 ppm (-CH2-); ESI-TOF/
HRMS: m/z calcd for C23H43N6O5S+: 515.3010 [M++H]+; found:
515.3008.
Compound 11: Protected monomeric chelator 10 (0.085 mmol,
135 mg) was dissolved in formic acid (4 mL). The reaction mixture
was stirred at 608C for 24 h. Formic acid was removed under re-
duced pressure to yield the orange product 11 (110 mg, 95%).
1H NMR (300 MHz, D2O): d=1H NMR (300 MHz, D2O): d=7.58–6.86
(br, 9H; ArH), 6.74–6.45 (br, 4H; ArH), 4.64–4.05 (br, 2H;
SCH2CHCH), 3.82–0.70 ppm (br, 51H; -CH- and -CH2-); ESI-TOF/
HRMS: m/z calcd for C64H81N12O16S3+: 1369.50501 [M+H]+; found:
Compound 8: Biotinylated amine 7 (0.51 mmol, 0.26 g), fluorescein
isothiocyanate (0.61 mmol, 0.24 g), and triethylamine (0.72 mmol,
0.10 mL) were dissolved in dimethylformamide (5 mL). The reaction
mixture was stirred for 16 h at room temperature. The solvent was
evaporated. The residue was purified by flash column chromatog-
raphy (silica gel, gradient 10 to 20% methanol/dichloromethane)
to yield 8 as an orange amorphous solid (0.40 g, 88%). 1H NMR
(300 MHz, CD3OD): d=7.78 (d, J=7.8 Hz, 1H; ArH), 7.54 (s, 1H;
ArH), 7.15 (d, J=8.2 Hz, 1H; ArH), 6.92 (d, J=8.6 Hz, 2H; ArH), 6.66
(d, J=2.4 Hz, 2H; ArH), 6.57 (dd, J=9.0, 2.5 Hz, 2H; ArH), 4.50–4.40
(br, 1H; SCH2CHCH), 4.32–4.16 (br, 2H; SCH2CHCH and
NHCHC(O)NH), 3.89–3.63 (br, 2H; NHCH2CH2NHC(S)NH), 3.56–3.32
(br, 3H; SCHCH and CH2NHCOOC(CH3)3), 3.18–3.07 (br, 1H; SCH),
2.87 (dd, J=12.9, 5.0 Hz, 1H; SCH2), 2.66 (d, J=12.9 Hz, 1H; SCH2),
2.35–2.12 (br, 2H; CH2C(O)NHCHC(O)NH), 1.85–1.51 (br, 6H; -CH2-
and C(CH3)3), 1.50–1.27 ppm (br, 16H); 13C NMR (75 MHz, CD3OD):
d=181.0 (C(S)), 174.4, 173.3, 169.7, 162.9 (C(O)), 159.9, 156.5, 152.6,
147.1, 140.3, 130.2, 128.9, 127.4, 124.3, 118.9, 112.5, 109.9, 102.4
(ArC), 78.8 (C(CH3)3), 61.5, 61.4, 59.8, 55.1, 53.32, 43.6, 41.9, 39.8,
39.5, 38.4, 36.2, 34.8, 31.0, 28.9 (-CH- and -CH2-), 27.8, 27.6, 27.5
((CH3)3C), 25.0, 22.5 ppm (-CH2-); ESI-TOF/HRMS: m/z calcd for
C44H54N7O10S2+: 904.3368 [M+H]+; found: 904.3378.
1369.50390; m/z calcd for C64H82N12O16S32+: 685.2561 [M+2H]2+
;
found: 685.2561.
mMT3-CA: The monomeric chelator 11 (82 mg, 60 mmol) was dis-
solved in water and the pH was adjusted to 7.0 with 0.1m sodium
hydroxide. GdCl3·6H2O (26 mg, 70 mmol) was added to the solution
and the pH was maintained at 7.0 with the sodium hydroxide solu-
tion. The mixture was stirred at room temperature for 24 h. Chelex
was added to the reaction mixture to remove excess Gd3+ ions.
After filtration to remove Chelex, water was removed under re-
duced pressure to give the solid orange product mMT3-CÀA
(46 mg, 51%). ESI-TOF/HRMS: m/z calcd for C64H76GdN12O16S3
1522.3920 [M]À; found: 1522.3940.
:
Compound 12: The amine 9 (100 mg, 0.12 mmol) was dissolved in
PBS (pH 7.5). SMCC (20 mg, 0.60 mmol) was dissolved in dimethyl-
formamide (0.5 mL) and added to the solution. The reaction mix-
ture was stirred at room temperature for 4 h to give 12. The prod-
uct was used in the next step without further purification. ESI-
LRMS: m/z calcd for C51H57N8O11S2À : 1021.4 [MÀH]À; found: 1021.4.
Compound 9: Compound 8 (0.28 mmol, 0.25 g) was dissolved in
methanol (5 mL) and concd HCl (5 mL) was added to this solution.
The reaction mixture was kept at room temperature for 1 h. The
solvent was evaporated under reduced pressure to give the prod-
uct 9 (0.22 g, 97%). 1H NMR (300 MHz, CD3OD): d=8.15 (d, J=
6.4 Hz, 1H; ArH), 7.64 (d, J=9.2 Hz, 2H; ArH), 7.44 (d, J=8.3 Hz,
1H; ArH), 7.38 (d, J=2.1 Hz, 2H; ArH), 7.23 (dd, J=9.3, 2.1 Hz, 2H;
ArH), 4.63–4.49 (br, 1H; SCH2CHCH), 4.45–4.17 (br, 2H; SCH2CHCH
and NHCHCONH), 4.08–3.67 (br, 2H; NHCH2CH2NHCSNH), 3.63–3.42
Compound 13: Cys-PAMAM-G4 (45.2 mg, 3.12 mmol) was dissolved
in dimethylformamide. Macrocycle 3 (240 mg, 0.30 mmol, 1.5 equiv
according to dendrimer surface groups) and triethylamine (56 mL,
0.40 mmol) were added to the dendrimer. The reaction mixture
was stirred at 458C for 48 h. The solvent was evaporated under re-
duced pressure and the residue was purified by lipophilic Sepha-
dex column chromatography using methanol as eluent. The dark
1
brown product 13 was obtained (165 mg, 81%). H NMR (300 MHz,
CDCl3): d=7.49 (br, ArH), 6.92 (br, ArH), 4.38–1.56 (br), 1.56–
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
620
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
1.21 ppm (br, (CH3)3C); MALDI-TOF-MS: m/z calcd for
C2100H3664N430Na45O412S392+: 21880 [M+45Na]2+; found: 21876.
echo sequence was used for imaging. T1-weighted saturation re-
covery MSME images were obtained with a series of varying repeti-
tion times and contrast, through the transverse and longitudinal
planes of the sample tubes. Acquisition parameters were: field of
view 4”4 cm2, matrix 200”200, one axial 2 mm slice, TE=6 ms,
and TR=21–8000 ms (logarithmic steps, 30 images). For longitudi-
nal images the field of view was 4”6 cm2, matrix 200”300, and
two slices.
Compound 14: The protected dendrimeric chelator 13 (180 mg,
3.45 mmol) was dissolved in formic acid (5 mL) and the mixture was
stirred at 608C for 16 h. Formic acid was evaporated under re-
duced pressure and the product was freeze-dried to give 14
(120 mg, 84%). 1H NMR (300 MHz, D2O): d=7.23 (br, ArH), 4.45–
0.40 ppm
C1524H2512N430Na21O412S38(H2O)17
found: 17717.
(br);
MALDI-TOF-MS:
m/z
calcd
for
;
2+
:
17717 [M+21Na+17H2O]2+
Processing of T1-weighted MR images for display purposes: The 32-
bit grayscale image files (2Dseq format) generated by the Bruker
scanner Paravision software were opened using the ImageJ
(v. 1.47d, Fiji package) image processing software. The images
were converted to 8-bit files using the “Image-Type-8-bit” com-
mand and saved as TIFF format. The TIFF images were opened
with Adobe Photoshop CS5 and rotated 908 clockwise.
Compound 15: The dendrimeric chelator 14 (120 mg, 5.8 mmol)
was dissolved in water. TCEP·HCl (16.63 mg, 58 mmol) was added
and the pH was adjusted to 7.0 with 0.1m sodium hydroxide solu-
tion. The reaction mixture was stirred at room temperature for 1 h.
The reaction was monitored by TLC using Ellman’s reagent. The ap-
pearance of a yellow spot confirmed the presence of thiol groups
in 15. The pH was then adjusted to 6.5–7.0 with hydrochloric acid.
A solution of 12 was added to the reaction mixture and the pH
was maintained at 6.5–7.0. The mixture was stirred at room tem-
perature for 16 h. The unreacted dendron 15 was reoxidized by
bubbling oxygen gas through the reaction mixture to give dendri-
mer 14. Subsequently, the mixture was concentrated under re-
duced pressure and the excess amounts of 12 and dendrimer 14
were removed using a G-15 Sephadex column and water as eluent.
Water was evaporated under reduced pressure and the product 16
was freeze-dried (110 mg, 87%). 1H NMR (300 MHz, D2O): d=7.21
(br, ArH), 4.09–0.51 ppm (br); MALDI-TOF-MS: m/z calcd for
C838H1350N228Na16O225S22(H2O)82+: 19347 [M+16Na+8H2O]2+; found:
19347.
Quantitative analysis of T1-weighted MR images: The original T1w
images generated by the Paravision software (32-bit, 2Dseq format
image files) were opened in the ImageJ (v. 1.47d, Fiji package) anal-
ysis software. A circular ROI was placed manually around each
bead/CA sample outlining the largest possible diameter fitting
within the walls of the sample tube, to maximize the number of
pixels sampled for calculation (1000–1500 pixels). The ImageJ func-
tion “Analyze-Measure” was used to calculate the mean pixel inten-
sity and total number of pixels within the ROI. Mean pixel intensi-
ties were exported into spreadsheets for further analysis.
Fluorescence intensity measurements
Equal volumes of CA-treated NeutrAvidin–agarose beads were
placed in a 96-well plate and fluorescence intensities were mea-
sured in a FluoStar Optima microplate reader (BMG Labtech, Ger-
many).
dMT3-CA: The dendronic chelator 16 (100 mg, 5.06 mmol) was dis-
solved in water and the pH was adjusted to 7.0 with 0.1m sodium
hydroxide. GdCl3·6H2O (90 mg, 243 mmol) was added to the solu-
tion and the pH was maintained at 7.0 with the sodium hydroxide
solution. The mixture was stirred at room temperature for 24 h.
EDTA (120 mg, 324 mmol) was added to the solution to remove
excess Gd3+ while maintaining the pH at 7.0 with the aqueous
sodium hydroxide solution. The mixture was stirred at room tem-
perature for 24 h and the major part of GdEDTA and free EDTA was
removed using water and a G-15 Sephadex column. The residues
of GdEDTA and EDTA were removed by centrifugation with a 3 kDa
centrifugal filter unit (102 mg, 79%). MALDI-TOF-MS: m/z calcd for
Fluorescence images of CA-treated NeutrAvidin–agarose beads:
Fluorescence imaging was performed by placing the beads on
a
SafeImager 2.0 blue (470 nm) LED source transilluminator
equipped with the supplied amber emission filter. Images were
captured with a digital camera and cropped around a region con-
taining the samples by using Photoshop CS5 software.
ICP-MS determination of Gd3+ associated with CA-treated
NeutrAvidin–agarose beads
C838H1274Gd19N228Na37O225S22(H2O)372+: 11676 [M+37Na+37H2O]2+
;
found: 11676.
Equal volumes of the CA-treated NeutrAvidin–agarose samples
were digested with 26% nitric acid at 508C for 72 h to release che-
lated Gd3+ into solution. The Gd3+ concentration was quantified
by ICP-MS at Currenta GmbH & Co. (Leverkusen, Germany).
Binding of contrast agents to NeutrAvidin–agarose beads
Packed beads were incubated with 1 mm CA in PBS (0.01m,
pH 7.4) for 30 min at 48C with rotary shaking. Excess CA was re-
moved and the beads were washed by repeated centrifugation
and resuspension in PBS. After the final wash step, supernatant
was removed and packed beads were kept for MRI scanning and
subsequent analyses as described below.
Fluorescence microscopy
Microscopic images were acquired using the AxioImager.Z1 upright
microscope (Zeiss) equipped with an AxioCamHR3 (Zeiss) camera.
Images were acquired over identical exposure times across all sam-
ples and exported as 16-bit grayscale, 1388”1040 pixel TIFF
format files.
MRI experiments
For MRI, 2 mL Eppendorf tubes containing NeutrAvidin–agarose
bead/CA sample were placed in a 1% agar solution to avoid sus-
ceptibility distortions. Imaging was performed on a 7 T Bruker Bio-
Spec 70/30 scanner, equipped with a BGA-12S gradient insert
using a commercial quadrature volume resonator, Bruker 1H 112/
086 QSN. For data acquisition a multislice–multiecho (MSME) spin–
Acknowledgements
We thank Dipl. Ing. Michael Beyerlein for help in performing
MRI experiments. Financial support from the Max-Planck Soci-
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
621
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
2007, 31, 1152–1160; f) M. A. Bruckman, X. Yu, N. F. Steinmetz, Nano-
technology 2013, 24, 462001.
ety, the Turkish Ministry of National Education (PhD fellowship
to S.G.), and European COST CM1006 and TD1004 Actions is
gratefully acknowledged.
[10] a) S. Venkataraman, Z. A. Chowdhury, A. L. Lee, Y. W. Tong, I. Akiba, Y. Y.
Yang, Macromol. Rapid Commun. 2013, 34, 652–658; b) C. A. Boswell,
P. K. Eck, C. A. S. Regino, M. Bernardo, K. J. Wong, D. E. Milenic, P. L.
Choyke, M. W. Brechbiel, Mol. Pharmaceut. 2008, 5, 527–539.
[11] D. A. Tomalia, B. Huang, D. R. Swanson, H. M. Brothers, J. W. Klimash, Tet-
rahedron 2003, 59, 3799–3813.
Keywords: chelates
·
contrast agents
·
dendrimers
·
gadolinium · imaging agents
[12] H. Xu, C. A. S. Regino, Y. Koyama, Y. Hama, A. J. Gunn, M. Bernardo, H.
Kobayashi, P. L. Choyke, M. W. Brechbiel, Bioconjugate Chem. 2007, 18,
1474–1482.
[13] G. Mattson, E. Conklin, S. Desai, G. Nielander, M. D. Savage, S. Morgen-
sen, Mol. Biol. Rep. 1993, 17, 167–183.
[14] H. Xu, C. A. S. Regino, M. Bernardo, Y. Koyama, H. Kobayashi, P. L.
Choyke, M. W. Brechbiel, J. Med. Chem. 2007, 50, 3185–3193.
[15] D. M. Corsi, C. Platas-Iglesias, H. van Bekkum, J. A. Peters, Magn. Reson.
Chem. 2001, 39, 723–726.
[1] a) C. T. Yang, K. H. Chuang, MedChemComm 2012, 3, 552–565; b) G.-P.
Yan, L. Robinson, P. Hogg, Radiography 2007, 13, e5–e19; c) C. F. G. C.
Geraldes, S. Laurent, Contrast Media Mol. Imaging 2009, 4, 1–23.
[2] a) A. E. Merbach, E. Toth, The Chemistry of Contrast Agents in Medical
Magnetic Resonance Imaging, Wiley, Chichester, 2001; b) R. B. Lauffer,
Chem. Rev. 1987, 87, 901–927.
[3] G. E. Hagberg, I. Mamedov, A. Power, M. Beyerlein, H. Merkle, V. G. Kise-
ˇ
lev, K. Dhingra, V. Kubìcek, G. Angelovski, N. K. Logothetis, Contrast
Media Mol. Imaging 2014, 9, 71–82.
[16] C. R. DeMattei, B. H. Huang, D. A. Tomalia, Nano Lett. 2004, 4, 771–777.
[17] J. B. Livramento, C. Weidensteiner, M. I. M. Prata, P. R. Allegrini, C. Ger-
aldes, L. Helm, R. Kneuer, A. E. Merbach, A. C. Santos, P. Schmidt, E. Toth,
Contrast Media Mol. Imaging 2006, 1, 30–39.
[18] a) S. Lehmann, E. G. Garayoa, A. Blanc, R. Keist, R. Schibli, M. Rudin, J.
Nucl. Med. 2011, 52, 445–452; b) P. Lehtolainen, T. Wirth, A. K. Taskinen,
P. Lehenkari, O. Leppanen, M. Lappalainen, K. Pulkkanen, A. Marttila, V.
Marjomaki, K. J. Airenne, M. Horton, M. S. Kulomaa, S. Yla-Herttuala,
Gene Ther. 2003, 10, 2090–2097; c) T. Mäntylä, J. M. Hakumaki, T. Huhta-
la, A. Narvanen, S. Yla-Herttuala, Magn. Reson. Med. 2006, 55, 800–804;
d) F. Pinaud, X. Michalet, G. Iyer, E. Margeat, H. P. Moore, S. Weiss, Traffic
2009, 10, 691–712; e) F. Pinaud, D. King, H. P. Moore, S. Weiss, J. Am.
Chem. Soc. 2004, 126, 6115–6123.
[4] a) S. Aime, A. Barge, C. Cabella, S. G. Crich, E. Gianolio, Curr. Pharm. Bio-
technol. 2004, 5, 509–518; b) K. Hanaoka, A. J. M. Lubag, A. Castillo-
Muzquiz, T. Kodadek, A. D. Sherry, Magn. Reson. Imaging 2008, 26, 608–
617.
[5] H. Kobayashi, S. Kawamoto, T. Saga, N. Sato, T. Ishimori, J. Konishi, K.
Ono, K. Togashi, M. W. Brechbiel, Bioconjugate Chem. 2001, 12, 587–
593.
[6] G. Gambino, J. Engelmann, L. Tei, M. Botta, N. K. Logothetis, I. Mamedov,
Biomaterials 2013, 34, 7135–7142.
[7] M. D. Melamed, N. M. Green, Biochem. J. 1963, 89, 591–599.
[8] R. Mishra, W. Su, R. Pohmann, J. Pfeuffer, M. G. Sauer, K. Ugurbil, J. En-
gelmann, Bioconjugate Chem. 2009, 20, 1860–1868.
[9] a) M. Longmire, P. L. Choyke, H. Kobayashi, Curr. Top. Med. Chem. 2008,
8, 1180–1186; b) E. C. Wiener, M. W. Brechbiel, H. Brothers, R. L. Magin,
O. A. Gansow, D. A. Tomalia, P. C. Lauterbur, Magn. Reson. Med. 1994, 31,
1–8; c) G. Yan, C. Ai, L. Li, R. Zong, F. Liu, Chin. Sci. Bull. 2010, 55, 3085–
3093; d) M. Polasek, P. Hermann, J. A. Peters, C. F. G. C. Geraldes, I.
Lukes, Bioconjugate Chem. 2009, 20, 2142–2153; e) S. Langereis, A. Dirk-
sen, T. M. Hackeng, M. H. P. van Genderen, E. W. Meijer, New J. Chem.
´
[19] S. M. Vibhute, J. Engelmann, T. Verbic, M. E. Maier, N. K. Logothetis, G.
Angelovski, Org. Biomol. Chem. 2013, 11, 1294–1305.
Received: September 22, 2014
Published online on December 12, 2014
ChemPlusChem 2015, 80, 612 – 622
www.chempluschem.org
622
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim