Synthesis and characterization of pH-sensitive, biotinylated MRI contrast agents and their conjugates with avidin

Article abstract of DOI:10.1039/c2ob26555a

Responsive or smart contrast agents (SCAs) provide new opportunities in magnetic resonance imaging (MRI) to examine a number of physiological and pathological events. However, their application in vivo remains challenging. Therefore, much research is focused on the optimization of their properties, to enable their use in additional imaging modalities, pre-targeted delivery, or to increase the local concentration of the agent. The key feature in the SCA synthetic modification is the retention of their physicochemical properties related to the specific MR response. Here, we report the preparation and characterization of pH sensitive SCAs appended with a phosphonate pendant arm and either an aliphatic (GdL¹) or aromatic linker (GdL²). The longitudinal relaxivity of GdL¹ and GdL² increases by 146% and 31%, respectively, while the pH decreases from 9 to 5. These two SCAs were converted to the biotinylated systems GdL³ and GdL⁴ and their interaction with avidin was investigated. The binding affinity with avidin was assessed with a fluorescence displacement assay and with MRI phantom experiments in a 3T MRI scanner. The fluorometric assay and MRI E-titrations revealed a 3:1 binding mode of GdL^3-4 to avidin with the binding affinity as high as that of the parent avidin-biotin complex. The high binding affinity was confirmed with MRI by a competitive assay. The avidin-GdL ^3-4 complexes thus obtained exhibit changes in both r₁ and r₂ that are pH dependent. The results reveal a new pathway for the modification and improvement of SCAs to make them more suitable for in vivo application.

Full text of DOI:10.1039/c2ob26555a

Organic&
BiomolecularChemistry
www.rsc.org/obc
Volume 11 | Number 8 | 28 February 2013 | Pages 1261–1424
ISSN 1477-0520
PAPER
Goran Angelovski et al.
Synthesis and characterization of pH-sensitive, biotinylated MRI contrast
agents and their conjugates with avidin
1477-0520(2013)11:8;1-B

Organic &
Biomolecular
Chemistry
View Article Online
View Journal | View Issue
PAPER
Synthesis and characterization of pH-sensitive,
biotinylated MRI contrast agents and their conjugates
Cite this: Org. Biomol. Chem., 2013, 11,
1294
with avidin†
Sandip M. Vibhute,a Jörn Engelmann,b Tatjana Verbić,c Martin E. Maier,d
Nikos K. Logothetisa,e and Goran Angelovski*a
Responsive or smart contrast agents (SCAs) provide new opportunities in magnetic resonance imaging
(MRI) to examine a number of physiological and pathological events. However, their application in vivo
remains challenging. Therefore, much research is focused on the optimization of their properties, to
enable their use in additional imaging modalities, pre-targeted delivery, or to increase the local concen-
tration of the agent. The key feature in the SCA synthetic modification is the retention of their physico-
chemical properties related to the specific MR response. Here, we report the preparation and
characterization of pH sensitive SCAs appended with a phosphonate pendant arm and either an aliphatic
(GdL1) or aromatic linker (GdL2). The longitudinal relaxivity of GdL1 and GdL2 increases by 146% and
31%, respectively, while the pH decreases from 9 to 5. These two SCAs were converted to the biotinylated
systems GdL3 and GdL4 and their interaction with avidin was investigated. The binding affinity with
avidin was assessed with a fluorescence displacement assay and with MRI phantom experiments in a
3T MRI scanner. The fluorometric assay and MRI E-titrations revealed a 3 : 1 binding mode of GdL3–4 to
avidin with the binding affinity as high as that of the parent avidin–biotin complex. The high binding
affinity was confirmed with MRI by a competitive assay. The avidin–GdL3–4 complexes thus obtained
exhibit changes in both r1 and r2 that are pH dependent. The results reveal a new pathway for the
modification and improvement of SCAs to make them more suitable for in vivo application.
Received 6th August 2012,
Accepted 15th November 2012
DOI: 10.1039/c2ob26555a
www.rsc.org/obc
of clinical MRI studies are performed using contrast agents,
employed to enhance image quality.2 The majority of MRI con-
Introduction
Magnetic resonance imaging (MRI) has become an essential trast agents are based on paramagnetic metal ions such as
tool of medical diagnostic imaging and basic research over the Mn2+ and Gd3+, or superparamagnetic iron oxide nanoparticles
past few decades. MRI has numerous advantages over other (SPION).3 Among these, the Gd3+-systems are the most fre-
imaging techniques because of its high spatiotemporal resolu- quently used due to the high magnetic moment, and its long
tion, non-invasiveness and non-ionizing signals.1 Up to 45% electronic spin relaxation time (10−9 s). In the presence of
Gd3+-agents, the relaxation rate of water protons increases
resulting in an increased image contrast. The efficiency of a
contrast agent (termed the longitudinal relaxivity, r1) depends
on many parameters such as number of directly bound water
molecules to metal center i.e. hydration number (q), mean resi-
dence lifetime of the bound water molecules (τm), electron
spin correlation time (T1e), rotational correlation time (τr).4
The tuning of such parameters can enhance the relaxivity of a
given Gd3+-based contrast agent, hence the resulting signal in
an MR image.
Recently there is a growing trend to explore responsive or
smart contrast agents (SCAs) which are able to report physio-
logical changes in the region of interest, allowing noninvasive
imaging of a specific biological process. There are many
aDepartment for Physiology of Cognitive Processes, Max Planck Institute for
Biological Cybernetics, Spemannstr. 38, 72076 Tübingen, Germany.
E-mail: goran.angelovski@tuebingen.mpg.de; Fax: +49 7071 601 919;
Tel: +49 7071 601 916
bHigh-Field Magnetic Resonance Center, Max Planck Institute for Biological
Cybernetics, Tübingen, Germany
cDepartment of Analytical Chemistry, Faculty of Chemistry, University of Belgrade,
Belgrade, Serbia
dInstitute for Organic Chemistry, Faculty of Science, University of Tübingen,
Tübingen, Germany
eImaging Science and Biomedical Engineering, University of Manchester, Manchester,
UK
†Electronic supplementary information (ESI) available: HPLC elution con-
ditions, fluorescence assay calculations and additional MRI phantom data. See
DOI: 10.1039/c2ob26555a
1294 | Org. Biomol. Chem., 2013, 11, 1294–1305
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
different types of SCAs reported so far; those sensitive to a par- center results in a decrease in pH. In this work, we report
ticular enzyme, pH, metal ion concentrations, or partial further synthetic modifications on the propylphosphonate
oxygen pressure.5,6 Significant efforts have been undertaken by containing SCA which involve the introduction of a side linker
researchers to develop pH responsive agents, since changes in to allow the coupling of SCA to a diverse range of functional
pH are associated with a number of physiological and patho- molecules such as dendrimers, nanoparticles, proteins, pep-
physiological processes.7,8 These SCAs contain functional tides or fluorescent tags. Most significantly the physico-chemi-
groups which change their protonation state around neutral cal properties of SCAs, in particular their responsiveness (the
pH, such as phosphonates or arylsulphonamides. The general pH dependent r1 changes), must be retained after coupling to
mechanism reported so far involves alteration of the coordi- the linker. Consequently, the properties of SCAs and the other
nation environment around the metal center at different pH, functional molecules could be used together to further charac-
which affects the number of proximate water molecules result- terize the function of the SCA in vitro and in vivo, broadening
ing in a change in the signal. The first pH responsive agent the scope of their application.
based on phosphonates was reported by Sherry and co-workers
For this purpose we have designed two novel cyclen-based
more than a decade ago.9 In this SCA, hydrogen bonds formed ligands L1 and L2 (Scheme 1). Their synthesis involved incor-
after protonation of the phosphonate play a crucial role cata- porating aliphatic (derivative of 6-aminohexanoic acid) and
lyzing the exchange of protons between inner and outer sphere aromatic linkers (derivative of 4-(p-amino phenyl) butyric acid)
water, resulting in a change in the signal at different pH. onto cyclen derivatives ultimately producing ligands L1 and L2
The sulfonamide based agent operates differently: at acidic pH respectively. These linkers were added on the cyclen nitrogen
the protonation of the nitrogen allows water to approach Gd3+
,
positioned trans to the propylphosphonate pendant arm, thus
whereas at basic pH deprotonation of nitrogen results in its keeping these two pendant moieties as distant as possible to
coordination to the metal center, reducing q and hence relaxi- avoid any interference in the desired activity. Both linkers
vity by hindering the access of water molecules to the inner possess a primary amine as the terminal group, suitable for
sphere.10
coupling to molecules containing free carboxylic acids, or elec-
However there is an upmost need for SCA with a response trophilic groups. Moreover, the aniline group in L2 can be con-
that can be combined with additional properties such as con- verted to an isothiocyanate to enable further coupling to
trolled biodistribution or determination of its exact local con- molecules with free amines. To demonstrate the applicability
centration. Such requirements necessitate additional synthetic of the newly developed SCA, we incorporated biotin and
modifications on SCAs and therefore special attention should studied the properties of the biotinylated SCA and their avidin
be paid towards the changes in the physico-chemical proper- conjugates regarding a pH-response. The biotin–avidin system
ties of SCAs, to ensure their relaxivity changes, along with has a specific interaction with an extraordinary affinity, being
crucial issues such as the kinetic and thermodynamic stability, stable over larger range of pH and temperatures in vivo.17 This
are retained. For instance, the in vivo biodistribution of probes system is thus an excellent tool for the targeted delivery of the
varies considerably with slight changes in their structures.11,12
Proceeding in this direction, Sherry and co-workers
attached a modified version of the previously mentioned phos-
phonate containing SCA to a PAMAM G5 dendrimer to
improve the MR signal by increasing the molecular weight of
the conjugate.13 The relaxivity of the conjugate increased twice
SCA in vivo.
Results and discussion
Synthesis
from pH 6 to pH 9.6 due to the increase in τr. However, due to L1 and L2 were obtained using two similar synthetic routes
its isoelectric point at pH ∼ 6, this dendrimeric SCA precipi- that commenced with the preparation of the two side arms
tates from solution. In another approach, a dual-modal (Scheme 1). Aliphatic brominated linker 2 was prepared by the
MR-PET agent was prepared in which quantification of the diazotization of H-Lys(Z)-OH with sodium nitrite, followed by
local probe concentration was accomplished by the presence the bromination using potassium bromide and tert-butyl pro-
of the PET active 18F nucleus, with the phosphonate-contain- tection of carboxylic acid with 2,2,2-trichloroacetimidate.18 1,7-
ing SCA, enabling pH mapping.14 Other attempts involved Bis(tert-butoxycarbonylmethyl)-1,4,7,10-tetraazacyclododecane
modification of the sulfonamide based pH sensitive agent. (bis-tert-butyl DO2A) 6 was prepared following the procedure
Namely, the accessibility of its phenolic oxygen allowed the reported previously.19 The key reaction of the synthesis was the
coupling of the functional molecule to a poly-β-cyclodextrin selective mono-alkylation of 6 with the newly synthesized
containing 19F reporter. Here an adamantane derivative was bromide 2 to obtain the tris-alkylated cyclen derivative 7,
introduced on the phenolic oxygen to achieve effective non- achieved under high dilution in acetonitrile, using sodium
covalent coupling to the poly-β-cyclodextrin, where detection bicarbonate. The final alkylation of 7 was achieved using
of the 19F signal intensity is used to quantify the amount of diethyl(3-bromopropyl) phosphonate in acetonitrile in the
the SCA.15
presence of potassium carbonate to afford 8 in good yields.
Our group recently reported a DO3A based, pH responsive Hydrogenation of 8 with Pd/C and H2 in ethanol gave primary
probe appended with a phosphonate arm (DO3APP).16 It was amine 9. Treatment of 9 with bromotrimethylsilane and sub-
shown that coordination of the phosphonate arm to the metal sequently formic acid yielded the ligand L1.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 1294–1305 | 1295

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 1 Synthesis of L1 and L2. Reagents and conditions: (i) NaNO2, KBr, 1N HBr, t-butyl-2,2,2-trichloroacetimidate, CHCl3; (ii) SOCl2, Br2, t-butyl-2,2,2-trichloro-
acetimidate, CHCl3; (iii) NaHCO3, CH3CN; (iv) NaHCO3, CH3CN; (v) diethyl (3-bromopropyl) phosphonate, K2CO3, CH3CN; (vi) H2, Pd/C, EtOH; (vii) BrSi(CH3)3, CH2Cl2,
(viii) formic acid (for L1) or TFA (for L2), CH2Cl2.
The aromatic linker 4 was prepared according to a modified
Complexation of ligands L1, L2 with lanthanide chlorides
literature procedure,20 achieving a higher yield (77% vs. 48%) (GdCl3·6H2O, EuCl3·6H2O) was carried out using a standard
than previously reported.21 Selective monoalkylation of 6 was procedure (see the Experimental section) to obtain the desired
achieved following the same procedure as described for the complexes at neutral pH. The successful formation was
aliphatic analogue 7, to afford the nitro derivative 10. The final confirmed by means of ESI-MS and the spectra contained
alkylation of 10 with diethyl(3-bromopropyl) phosphonate appropriate isotope pattern distribution characteristic for Gd3+
resulted in 11. The nitro group was converted into the amine or Eu3+ complexes.
12 by hydrogenation using Pd/C as a catalyst in ethanol. Depro-
tection was carried out as described for L1 to afford L2, except
Relaxometric and luminescence emission experiments
that TFA was used instead of formic acid. Both ligands L1 and A relaxometric study of complexes GdL1–2 was performed at
L2 were characterized by means of 1H, 13C, 31P NMR and
7 T (300 MHz) and 25 °C. The T1 relaxation times were
HR-ESI mass spectrometry.
recorded at pH 9, 7 and 5 using different concentrations of
1296 | Org. Biomol. Chem., 2013, 11, 1294–1305
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
For EuL1 the apparent q value increases from 1.1 to 1.8 as
the pH decreases from 7 to 5. The observed change is in good
agreement with the findings of relaxivity experiments on the
corresponding Gd3+ complex, confirming that the change in r1
is directly influenced by changes in q.
Determination of the q value for EuL2 was not possible due
to the low intensity of emission and low solubility of the
complex at high concentrations (>3–4 mM). One of the reasons
for the weak emission could be the presence of the highly elec-
tron donating aniline derivative in close proximity of the metal
center which quenches the signal, whereas the hydrophobic
aromatic ring of the linker might decrease the solubility.24
The relaxometric and luminescence emission experiments
performed on the modified phosphonate complexes confirm
that the synthetic modifications in close proximity of the
metal centre do not negatively affect the coordination of Ln3+,
hence the pH responsive properties (similar to the DO3APP
parent compound) are retained. This enabled us to pursue
further investigations involving L1–2 towards their potential
MRI application with other biomolecules such as proteins (see
below).
Fig. 1 The pH dependence of the proton relaxivity (r1) of GdL1 and GdL2 (7 T,
25 °C) compared to DO3APP16 (7 T, 21 °C).
Table 1 Emission lifetimes and estimated q values of EuL1 at neutral and
acidic pH/pD
Compound
τ (ms) H2O
τ (ms) D2O
q
pH/pD 7
pH/pD 5
0.56
0.42
1.69
1.51
1.1
1.8
Biotinylation of GdL1–2
complexes. The pH was adjusted using solid lithium hydroxide
and para-toluenesulfonic acid to avoid dilution of GdL1–2 solu-
tions. The exact concentration of Gd3+ was determined by the
bulk magnetic susceptibility shift method.22 The r1 relaxivity at
each pH is obtained as a slope of the linear regression curve
for concentrations of GdL1–2 plotted against the relaxation rate
(R1). The relaxivities of both GdL1–2 at the acidic pH were
slightly higher than for the parent DO3APP compound
(Fig. 1).16 In both cases we observed a continuous increase in
relaxivity from pH 9 to 5. A total increase in r1 of 147% from
pH 9 to pH 5 was recorded for GdL1 (32% from pH 7 to pH 5),
whilst a total increase of 31% in r1 was observed for GdL2 over
the same pH range. The change in relaxivity with a change in
pH is observed as in the parent compound. Protonation of the
phosphonate in acidic pH induces movement of the pendant
arm from the metal center and allows more water molecules to
interact with the metal center resulting in a higher relaxivity.
The high affinity avidin–biotin conjugate is a widely studied
system used in many therapeutic, diagnostic, and pharmaco-
logical applications.25–27 Avidin is a tetrameric protein with a
high molecular mass (66 k Dalton) and has an extremely high
binding affinity towards biotin (Kd ∼ 1.7 × 10−15) over wide pH
range.17 Once biotinylated, the avidin conjugate of our SCA
can be delivered to specific regions in vivo, or the biotinylated
SCA alone could be used for selective cell-targeting.28 Hence,
we have synthesized two biotinylated molecules L3 and L4 in a
two step procedure commencing with the ester-protected
ligands 9 and 12 (Scheme 2). They were coupled to biotin
using EDCI·HCl and DMAP in DMF at room temperature,
where the reaction was monitored by ESI-MS. The ligands were
obtained following the deprotection of phosphonate and car-
boxylic acids using bromotrimethylsilane and trifluoroacetic
acid respectively. Upon characterization of the ligands, their
Gd3+ complexes were prepared as described previously (see
above and the Experimental section).
q ¼ A′ ðΔ kH O ꢀ Δ kD O ꢀ 0:25Þ
ð1Þ
2
2
Fluorescence displacement assay
The two responsive complexes were characterized further by The binding affinity of the newly synthesized molecules GdL3–4
investigating their Eu3+ analogues. The 31P NMR spectra of to avidin was measured using a fluorometric assay based on the
EuL1 and EuL2 were recorded at pD 5 and 7. A pH-dependent displacement of the fluorescent probe 2-anilinonaphthalene-
low field chemical shift of approximately 2 ppm in both com- 6-sulfonic acid (ANS), employing the method developed by Hor-
plexes indicates that a similar mechanism responsible for the owitz and coworkers.29 Briefly, ANS is strongly fluorescent when
r1 changes is occurring as observed for DO3APP.16 Time attached to avidin whereas the unbound dye has only a low fluore-
resolved luminescence decay experiments were performed on scence intensity. Since the binding affinity of ANS to avidin
complexes EuL1 and EuL2 to evaluate the change in the inner is lower than binding affinity of biotin to avidin, the addition of
sphere hydration at different pH (Table 1). The experiments biotin displaces ANS from the avidin binding pocket resulting
were carried out in 5 mM solutions of EuL1–2 in H2O and D2O. in the quenching of the fluorescence in the sample.
The hydration number was estimated using eqn (1) where A′ is
The fluorescence displacement assay was performed in para-
1.2 ms and the correction factor for the contribution of the llel on biotin, GdL3 and GdL4. Upon addition of biotin or a
second sphere is −0.25 ms−1 23
.
biotinylated probe to the mixture of avidin and ANS, the
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 1294–1305 | 1297

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 2 Synthesis of L3 and L4. Reagents and conditions: (i) biotin, EDCI·HCl, DMAP, DMF; (ii) BrSi(CH3)3, CH2Cl2; (iii) TFA, CH2Cl2.
fluorescence intensity decreased reaching a sharp inflection complexes bind to avidin. The increase in the r1/r2 of conju-
point at 3 equiv. of biotin or GdL3–4 (Fig. S1, ESI†). After this gates is the direct consequence of the τR decrease, which is
point there was no further drop in intensity observed even commonly observed for macromolecular contrast agents at
after the addition of several equivalents of biotin or GdL3–4
.
magnetic fields relevant for clinics.31 The plateaus of r1/r2
The measured stoichiometry is in line with the results pre- values were observed at 0.3–0.4 equiv. of avidin which matches
viously reported by Horowitz and can be associated with the our observations from the fluorometric assay, indicating that
affinity-purified avidin obtained from commercial sources.29 three of the binding sites are occupied in avidin.
This assay shows that only three binding sites are available in
In the subsequent step we studied the pH response of
the studied avidin to interact with ANS and these three mole- GdL3–4 in the absence and presence of avidin. We mixed
cules of dye can be replaced by biotin or biotinylated SCA. GdL3–4 and avidin in a 3 : 1 ratio. The pH was adjusted over the
However, the titration curves of GdL3–4 are comparable to wide range starting from 5.5 to 9.0 using different buffers
biotin indicating that the binding affinities of both com- (25 mM). Longitudinal T1 and transverse T2 relaxation times
pounds are as strong as biotin. The calculated stability con- were again obtained at 25 °C in a 3T MRI scanner.
stants with Ks values 5.1 × 1013, 7.1 × 1013 and 8.0 × 1013 for
A decrease in relaxation times T1 and T2, hence an increase
biotin, GdL3 and GdL4, respectively, suggest that the binding in r1 and r2, was observed for both complexes when the pH
affinities of biotin and biotinylated SCA are not significantly decreased from 9.0 to 5.5 (Fig. 3). A good reproducibility was
different from each other (see ESI†).
observed over five independent experiments performed. The
GdL3–avidin conjugate showed an increase from 10.4 to
12.6 mM−1 s−1 (21%) and from 39.3 to 58.4 mM−1 s−1 (49%) of
r1 and r2 respectively. Within the same pH range, the GdL4–
avidin conjugate showed an increase from 11.2 to 12.6 mM−1 s−1
(12%) and from 40.4 to 52.6 mM−1 s−1 (30%) of r1 and r2
respectively. When comparing these values with those
obtained under same conditions (3T scanner, pH 5.5–9.0) for
GdL3–4 in the absence of avidin, the relative changes in relaxiv-
ities are comparable, except that the absolute r2 values
are much higher in the GdL3–4–avidin conjugates (Fig. 3 and
S3–S4 in ESI†). The results suggest that the investigated SCA
remained responsive not only by the further synthetic modifi-
cations (biotinylation of GdL1–2 to GdL3–4), but GdL3–4 are
capable of a specific and strong binding to the avidin while
retaining their MR activity. Finally, the observed r1/r2 changes
MRI phantom experiments
To confirm the observations from the fluorescence displace-
ment assays, we followed the longitudinal and transverse relax-
ation rates (R1, R2) in samples with a constant concentration of
GdL3–4 in HEPES buffer and varying concentrations of avidin
in a method known as an E-titration.30 These mixtures were
incubated at 37 °C for 2 h and their relaxation times T1 and T2
were measured using a 3T (123 MHz) MRI scanner following
the procedure and parameters as described in the Experimen-
tal section. A linear increase in r1/r2 with addition of avidin
compared to the unbound probe was observed until all biotiny-
lated complexes are bound to avidin. A further increase in
avidin concentration has no further effect on r1/r2 (Fig. 2). The
MRI phantom experiments confirm that biotinylated
1298 | Org. Biomol. Chem., 2013, 11, 1294–1305
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 2 E-titrations of GdL3 (left) and GdL4 (right) with avidin at a 3T MRI scanner (21 °C, HEPES, pH 7.4). The lines correspond to the fit assuming n = 3 and KsII (Av
(GdL3)3) = 7.1 × 1013 or KIsI (Av(GdL4)3) = 8.0 × 1013 obtained from the fluorescence displacement assay. The R2 values were 0.9817 and 0.9924 for r1 and r2 respect-
ively for GdL3, whereas 0.9583 and 0.9438 for r1 and r2 respectively for GdL4. A better fit (R2 = 0.9936 and 0.9963 for r1 and r2 respectively) for GdL4 could be
obtained when n is also fitted (n = 2.4 and 2.3 for r1 and r2 respectively). The fitting equation is described in ESI†.
Fig. 3 pH-dependent r1 (left) and r2 (right) response of GdL3 (top) and GdL4 (bottom) with avidin at a 3T MRI scanner (21 °C). The avidin : GdL3–4 molar ratio was
1 : 3. Values are presented as mean SEM of five independent experiments. The lines represent the result of the sigmoidal fit and are displayed to aid a better visual-
ization of the pH dependent r1/r2 decrease.
are the most pronounced in the pH range 7.0 to 8.5, making mixed with avidin in a fixed 3 : 1 ratio (HEPES buffer, pH 7.4).
GdL3–4 a promising prototype SCA for further investigations Different concentrations of biotin were added and the mix-
in vivo under physiological conditions.
tures were incubated for five hours. If the binding affinity of
the SCA is lower than biotin, GdL3–4 should be displaced by
biotin from the avidin binding pocket, resulting in a decrease
in relaxivity, in particular r2. Following the incubation, the
longitudinal and transverse relaxation times were measured in
a 3T MRI scanner. We observed a minor decrease in relaxivities
Competitive MRI assay
We performed a competitive assay with biotin to further
characterize the binding affinity of the biotinylated SCA
towards avidin and estimate their in vivo potency. GdL3–4 was
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 1294–1305 | 1299

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 4 Competitive MRI assay of avidin : GdL3 (1 : 3, 125 μM GdL3, left) and avidin : GdL4 (1 : 3, 125 μM GdL4, right) with varying concentrations of biotin at a 3T
MRI scanner (21 °C).
only after addition of 100 times excess of biotin whereas the 4-(4-nitro phenyl butyric acid (3) and biotin were purchased
r1/r2 did not change at lower concentrations of biotin (Fig. 4). from Sigma-Aldrich, Germany. Cyclen (5) was purchased from
Such results are expected given the equilibrium exchange of CheMatech, France. Avidin was purchased from Merck,
biotinylated GdL3–4 at high concentrations of biotin. However, Germany. ANS was purchased from Life Technologies GmbH,
these findings confirm that the binding affinity of GdL3 and Germany. Column chromatography was performed using silica
GdL4 to avidin is as high as that of biotin. Furthermore, they gel 60 (70–230 mesh ASTM) or aluminium oxide 90 active
imply the stability of MRI signal changes caused by the SCA basic from Merck (Germany). Reversed-phase high-perform-
either in its premixed solution with biotin or in the presence ance liquid chromatography was performed on a Varian Pre-
of endogenous biotin in vivo.
pStar Instrument (Australia), equipped with PrepStar SD-1
pump heads. Analytical RP-HPLC was performed on the Atlan-
tis C18 column 4.6 mm × 150 mm, particle size 5 μm and
semi-preparative RP-HPLC was performed on the Atlantis C18
column 19 mm × 150 mm, particle size 5 μm (Waters Corpor-
ation, USA) using elution conditions described as in the ESI
(Table S1†).
Conclusions
In conclusion, we have synthesized and characterized two
novel macrocyclic gadolinium based pH responsive contrast
agents with an aliphatic or aromatic linker. Relaxometric and
luminescence studies at different pH confirmed that for both
agents the main physicochemical properties related to the pH
dependent MRI response were retained after inserting the
linking unit and are comparable to their parent compound.
Furthermore, they were coupled to biotin and the response of
the newly prepared responsive MRI agents was studied in more
detail. The binding affinity of these complexes with avidin was
shown to be as effective as biotin itself through fluorescence
displacement or competitive MRI assays. MRI phantom experi-
ments in the presence of avidin revealed that the relaxivity
was changing along with the change in pH at physiologically
relevant values. These complexes offer a useful pathway
towards target-specific SCAs for in vivo applications. Future
directions will involve applying this strategy in introducing
various functional molecules such as proteins, peptides, recep-
tor ligands or fluorescent tags to SCAs of various kinds
(e.g. pH- or ion-responsive) making them suitable for their
application in vivo.
1H, 13C {1H}, 31P {1H} NMR spectra and relaxometric experi-
ments were recorded on a Bruker Avance III 300 MHz ‘Micro-
bay’ spectrometer (Bruker, Germany). ESI-LRMS were
performed on an ion trap SL 1100 system (Agilent, Germany).
FT-ICR-MS were performed on a Bruker FT-ICR Apex II spec-
trometer (Bruker, Germany). HR-EI-MS were performed on a
MAT Sektorfeld mass spectrometer (Finnigan, Germany).
Luminescence steady-state and time resolved measurements
were performed on a QuantaMasterTM 3 PH fluorescence spec-
trometer from Photon Technology International, Inc. (USA).
Synthetic procedures
tert-Butyl-6-(((benzyloxy)carbonyl)amino)-2-bromohexanoate
(2). Potassium bromide (2.96 g, 12.50 mmol) was added to the
solution of H-Lys(Z)-OH (2.00 g, 7.13 mmol) in HBr (16 mL,
1 N). The resulting mixture was cooled to 0 °C, sodium nitrite
(0.59 g, 8.56 mmol) was added portionwise over 1 h and the
mixture was stirred for a further 2 h. Conc. sulfuric acid (1 mL)
was added slowly at 0 °C and the solution was allowed to reach
room temperature. The mixture was extracted with diethyl
ether (3 × 150 mL). The combined organic layers were dried
over anhydrous sodium sulphate and concentrated to afford a
bromo derivative of acid as a light yellow syrup. This acid
Experimental section
General remarks
All chemicals were purchased from commercial sources (0.50 g, 1.45 mmol) was suspended in chloroform (5 mL) and
and used without further purification. H-Lys(Z)-OH (1), tert-butyl-2,2,2-trichloroacetimidate (0.64 g, 2.90 mmol) in
1300 | Org. Biomol. Chem., 2013, 11, 1294–1305
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
chloroform (5 mL) was added dropwise. After 10 minutes a cata- filtration and the filtrate was concentrated under reduced
lytic amount of trifluoroborane diethyl etherate (10 μL) was pressure. A pure product was afforded by silica gel column
added and the resulting mixture was stirred at room tempera- chromatography (3% methanol/dichloromethane) to give 7 as
ture for 16 h under a nitrogen atmosphere. Sodium bicarbo- an off-white solid (1.65 g, 46%).
nate (1.00 g) was added to the mixture and stirred for another
1H NMR (300 MHz, CDCl3) δ (ppm): 1.43 (m, 27H, (CH3)3C),
10 minutes. The insoluble salts were removed by filtration and 1.56–1.68 (m, 4H, CH2CH2CH2CH2), 2.32–3.53 (br, 21H, CH2
the residue was obtained after concentration of the filtrate. ring, CHCOOtBu and CH2COOtBu), 5.06 (s, 2H, PhCH2CO),
The residue was purified using silica gel column chromato- 5.79 (br s, 1H, CH2NHCOOCH2), 7.19–7.39 (m, 5H, ArH), 9.49
graphy (6% ethyl acetate/hexane) to obtain a light yellow semi- (br s, 1H, NH ring). 13C{1H} NMR (75 MHz, CDCl3) δ (ppm):
solid of 2 (0.30 g, 51%).
23.8, 28.1, 28.2 ((CH3)3C), 29.2, 30.0, 40.4 (NHCH2CH2), 45.6,
1H NMR (300 MHz, CDCl3) δ (ppm): 1.41–1.63 (m, 13H, 48.5, 49.9, 50.9 (CH2 ring), 56.1 (CH2COOtBu), 61.9
CH2CH2CH2CH2, (CH3)3C), 1.87–2.12 (m, 2H, CH2CHBrCO), (CHCOOtBu), 66.1 (PhCH2OC(O)), 81.5, 81.7, ((CH3)3C), 127.7,
3.19 (t, J = 6 Hz, 2H, NHCH2CH2), 4.10 (t, J = 6 Hz, 1H, 127.8, 128.3, 136.9 (ArC), 156.6, 170.1, 171.4 (C(O)). ESI-HRMS:
CH2CHBrCO), 4.84 (br s, 1H, NHCH2), 5.10 (s, 2H, PhCH2OC calculated for C38H66N5O8 [M + H+]+, m/z 720.4911, found
(O)), 7.37–7.32 (m, 5H, ArH). 13C{1H} NMR (75 MHz, CDCl3) 720.4901.
δ
(ppm): 24.4 (NHCH2CH2CH2), 27.7 ((CH3)3C), 29.2
Compound 8. Potassium carbonate (0.27 g, 1.99 mmol) and
(CHBrCH2CH2), 34.4 (NCH2CH2CH2), 40.6 (CHBrCH2CH2), diethyl (3-bromopropyl) phosphonate (0.38 mL, 1.99 mmol)
47.4 (COCHBrCH2), 66.6 (PhCH2OC(O)), 82.4 ((CH3)3C), 128.1, were added to the solution of compound 7 (1.30 g, 1.80 mmol)
128.5, 136.5 (ArC), 156.3, 168.7 (C(O)). ESI-HRMS: calculated in anhydrous acetonitrile (15 mL). The reaction mixture
for C18H26BrNNaO4 [M + Na+]+, m/z 422.0937, found 422.0937.
was heated to 70 °C for 16 h. The insoluble solids were
tert-Butyl-2-bromo-4-(4-nitrophenyl)butanoate (4). 4-(4-Nitro- removed by filtration, the filtrate was concentrated under
phenyl)-butyric acid (3.00 g, 14.34 mmol) was dissolved in reduced pressure and the residue was purified by silica gel
thionyl chloride (4.16 mL, 57.36 mmol) and the mixture was column chromatography (50% methanol/dichloromethane/
stirred at 75 °C for 16 h. Bromine (0.82 mL, 15.78 mmol) was 10% ammonia). Removal of solvents yielded a light yellow oil
added dropwise at the same temperature and the mixture was of 8 (1.20 g, 74%).
again stirred for 16 h. The reaction mixture was cooled to
1H NMR (300 MHz, CDCl3) δ (ppm): 1.29 (t, J = 6 Hz, 6H,
room temperature and the mixture was poured over ice and OCH2CH3), 1.36–1.80 (m, 37H, CH2CH2P(O), (CH3)3C),
stirred for 2 h. The aqueous layer was extracted with diethyl 2.28–2.94 (br, 18H, CH2 ring), 3.12–3.22 (m, 7H, CHCOOtBu
ether (3 × 200 mL). The combined organic phases were washed and CH2COOtBu), 4.94–4.17 (m, 4H, POCH2), 5.07 (s, 2H,
with brine (200 mL), dried over Na2SO4 and concentrated PhCH2OC(O)), 5.46 (br s, 1H, CH2NHCbz), 7.21–7.42 (m, 5H,
under reduced pressure. The residue was redissolved in chloro- ArH). 13C{1H} NMR (75 MHz, CDCl3) δ (ppm): 16.3 (d, J = 6 Hz,
1
form (30 mL) and a solution of tert-butyl-2,2,2-trichloroaceti- OCH2CH3), 20.2 (d, J = 4.5 Hz, P(O)CH2CH2), 23.3 (d, JPC
=
midate (6.27 g, 28.68 mmol) in chloroform (20 mL) was added 141 Hz, P(O)CH2CH2), 23.3, 28.1, 28.2, 29.3, 29.5, 40.7
slowly followed by boron trifluoride etherate (100 μL). The (CH2CH2NHCOO), 49.5 (NCH2CH2), 52.3, 52.5, 52.6, 56.0 (CH2
resulting mixture was stirred at room temperature for 16 h. ring), 61.2, 61.2 (CH2COOtBu), 64.1 (CHCOOtBu), 66.2
Sodium bicarbonate (2.00 g) was added and stirred for (OCH2CH3), 80.4, 80.5 ((CH3)3C), 127.8, 128.0, 128.3, 136.7
30 minutes. The inorganic impurities were removed by fil- (ArC), 156.3, 170.8, 172.8 (C(O)). 31P{1H} NMR (122 MHz,
tration, the filtrate was concentrated and the residue was puri- CDCl3)
δ
(ppm): 32.9 (s). ESI-HRMS: calculated for
fied by silica gel column chromatography (2% ethyl acetate/ C45H81N5O11P [M + H]+, m/z 898.5670, found 898.5659.
hexane) to afford a light yellow oil of bromide 4 as a pure
product (3.80 g, 77%).
Compound 9. Compound 8 (0.68 g, 0.75 mmol) was dis-
solved in absolute ethanol (20 mL) and 10% Pd–C (70 mg,
1H NMR (300 MHz, CDCl3) δ (ppm): 1.49 (s, 9H, (CH3)3C), 10% w/w) was added. The heterogeneous mixture was shaken
2.21–2.43 (m, 2H, CH2CH2CHBr), 2.77–3.00 (m, 2H, for 16 h under a hydrogen atmosphere (35 psi) in a Parr hydro-
CH2CH2CHBr), 4.07 (dd, J = 9, 6 Hz, 1H, CH2CH2CHBr), 7.38 genator apparatus. The catalyst was removed by filtration
(d, J = 9 Hz, 2H, ArH), 8.18 (d, J = 9 Hz, 2H, ArH). 13C{1H} NMR through the celite and the filtrate was concentrated to obtain a
(75 MHz, CDCl3)
(ArCH2CH2CHBr), 35.7 (ArCH2CH2CHBr), 46.4 (ArCH2CH2CHBr),
82.8 ((CH3)3C), 123.9, 129.3, 146.7, 147.9 (ArC), 168.3 (C(O)). OCH2CH3),
ESI-HRMS for C14H18BrNO4 [M + Na]+, m/z calcd 366.0311, NHCH2CH2CH2CH2, (CH3)3C), 2.15–3.45 (br, 25H), 3.95–4.20
found 366.0307.
(m, 4H, OCH2CH3). 13C{1H} NMR (75 MHz, CDCl3) δ (ppm):
δ
(ppm): 27.7 ((CH3)3C), 33.1 light yellow oil of the amine 9 (0.50 g, 86%).
1H NMR (300 MHz, CDCl3) δ (ppm): 1.31 (t, J = 6 Hz, 6H,
1.37–1.95 (br, 37H, P(O)CH2CH2CH2,
Compound 7. A solution of bromide 2 (2.00 g, 5.00 mmol) 16.5 (d, J = 6 Hz, OCH2–CH3), 20.0 (d, J = 4.5 Hz, P(O)CH2CH2),
in anhydrous acetonitrile (50 mL) was added dropwise over 2 h 23.5 (d, 1JPC = 140 Hz, P(O)CH2CH2), 23.6, 28.2, 28.3 ((CH3)3C),
to the mixture of 619 (2.50 g, 6.25 mmol) and sodium bicarbo- 29.9, 33.5, 42.1 (CH2CH2NH2), 49.6 (NCH2CH2), 52.3, 52.3,
nate (0.42 g, 5.00 mmol) in anhydrous acetonitrile (100 mL) at 52.7, 56.2 (CH2 ring), 61.3, 61.4 (CH2COOtBu, OCH2CH3), 64.2
room temperature. The resulting mixture was stirred at room (CHCOOtBu), 80.5, 80.6 ((CH3)3C), 171.0, 172.9 (C(O)). 31P{1H}
temperature for 2 days. Insoluble salts were removed by NMR (122 MHz, CDCl3) δ (ppm): 31.0 (s), 32.9 (s). ESI-HRMS:
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 1294–1305 | 1301

View Article Online
Paper
Organic & Biomolecular Chemistry
calculated for C21H41N5O9P [M + H]+, m/z 764.5296, found
764.5294.
1H NMR (300 MHz, CDCl3) δ (ppm): 1.30 (t, J = 7.5 Hz, 6H,
P(O)OCH2CH3), 1.37–1.56 (m, 27H, (CH3)3C), 1.61–2.22 (br, 6H,
Ligand L1. Bromotrimethylsilane (0.62 g, 3.90 mmol) was ArCH2CH2, P(O)CH2CH2CH2), 2.31–3.05 (br, 20H, CH2 ring),
added slowly at 0 °C to the solution of the amine 9 (0.30 g, 3.10–3.35 (m, 5H, CHCOOtBu, CH2COOtBu), 3.96–4.18 (m, 4H,
0.39 mmol) in anhydrous dichloromethane (20 mL). The reac- OCH2CH3), 7.38 (d, J = 7.5 Hz, 2H, ArH), 8.13 (d, J = 7.5, 2H,
tion mixture was stirred for 12 h at room temperature. The ArH). 13C{1H} NMR (75 MHz, CDCl3) δ (ppm): 16.5 (d, J = 5.3
solvent was removed under reduced pressure and the residue Hz, P(O)OCH2CH3), 20.1 (d, J = 2.3 Hz, P(O)CH2CH2), 23.5
was dissolved in formic acid (4 mL) and stirred at 60 °C for (d, 1JPC = 140 Hz, P(O)CH2CH2), 28.2, 28.3 ((CH3)3C), 31.4, 32.6
16 h. The reaction mixture was cooled to room temperature (ArCH2CH2CH), 49.8 (NCH2CH2), 52.5, 52.5, 52.7, 56.2 (CH2
and excess volatiles were removed under reduced pressure. The ring), 61.3, 61.4 (CH2COOtBu, OCH2CH3), 64.1 (CHCOOtBu),
residue was dissolved in water (2 mL) and added slowly to 80.7, 81.0 ((CH3)3C), 123.6, 129.3, 146.3, 150.2 (ArC), 170.8,
acetone (200 mL) under vigorous stirring. The solution was 172.3 (C(O)). 31P{1H} NMR (122 MHz, D2O) δ (ppm): 32.8 (s).
cooled to −20 °C for 12 h. The solid product was separated by ESI-HRMS for C41H72N5O11P [M + H]+, m/z calcd 842.5038,
filtration, dried under reduced pressure to afford an off-white found 842.5038.
solid of L1 (0.30 g, quant.).
Compound 12. A mixture of 11 (0.45 mg, 0.53 mmol) and
1H NMR (300 MHz, D2O) δ (ppm): 1.40–2.15 (br, 10H), 10% Pd–C (90 mg, 20% w/w) in absolute ethanol (20 mL) was
2.80–3.35 (br, 23H, CH2 ring, CH2 acid arm), 4.09 (m, 1H, NCH shaken under a hydrogen atmosphere (35 psi) in a Parr hydro-
(COOH)CH2). 13C{1H} NMR (75 MHz, D2O) δ (ppm): 17.2, 23.6, genator. After completion of the reaction, the catalyst was
1
23.6 (d, JPC = 137 Hz, P(O)CH2CH2), 24.8, 26.5, 39.0, 46.9, removed by filtration through celite and the solvent was
47.2, 48.3, 49.6, 49.9, 50.0, 50.1, 53.3, 54.2, 54.4, 63.1 removed under reduced pressure. The residue was purified
(CHCOOH), 170.8, 173.6, 174.58 (C(O)). 31P{1H} NMR through alumina column chromatography (3% methanol/
(122 MHz, D2O) δ (ppm): 27.2 (s). ESI-HRMS for C21H41N5O9P: dichloromethane) to obtain a pure light yellow oil of 12
[M − H+]−, m/z calcd 538.2647, found 538.2648.
Compound 10. Sodium bicarbonate (0.37 mg, 4.40 mmol)
(0.37 g, 78%).
1H NMR (300 MHz, CDCl3) δ (ppm): 1.29 (t, J = 7.5 Hz, 6H,
and a solution of bromide 4 (1.51 g, 4.40 mmol) in anhydrous P(O)OCH2CH3), 1.35–1.47 (br, 27H, (CH3)3C), 1.60–1.87 (br,
acetonitrile (50 mL) were added slowly to a solution of bis-tert- 6H, ArCH2CH2CHBr, P(O)CH2CH2CH2), 2.36–2.95 (br, 20H,
butyl DO2A 6 (2.20 g, 5.50 mmol) in anhydrous acetonitrile CH2 ring), 3.10–3.33 (m, 5H, CHCOOtBu, CH2COOtBu) 3.55
(100 mL) over 3 h at room temperature. The resulting mixture (br s, 2H, ArNH2), 3.93–4.16 (m, 4H, OCH2CH3), 6.60 (d, J = 7.5
was stirred at room temperature for 2 days. The inorganic Hz, 2H, ArH), 6.97 (d, J = 7.5, 2H, ArH). 13C{1H} NMR (75 MHz,
impurities were removed by filtration, the filtrate was concen- CDCl3) δ (ppm): 16.4 (d, J = 6 Hz, –P(O)–OCH2CH3), 20.0 (d, J =
1
trated under reduced pressure and the residue was purified 4.5 Hz, P(O)CH2CH2CH2), 23.4 (d, JPC = 140 Hz, P(O)
by silica gel column chromatography (3% methanol/dichloro- CH2CH2CH2), 28.1, 28.3 (CH3)3C), 29.6 (P(O)CH2CH2CH2), 31.8
methane). The product 10 was afforded as an off- (ArCH2CH2), 32.3 (ArCH2CH2), 49.7 (NCH2CH2), 52.2, 52.3,
white solid after recrystallization from cold diethyl ether 52.8, 53.3, 55.6, 55.9 (CH2 ring), 56.13 (CH2COOtBu), 61.3
(1.90 g, 65%).
(d, J = 3 Hz, P(O)OCH2CH3), 61.3, 64.5 (CHCOOtBu), 80.5, 80.5
1H NMR (300 MHz, CDCl3) δ (ppm): 1.40 (s, 9H, (CH3)3C) ((CH3)3C), 115.1, 129.1, 132.0, 144.2 (ArC), 171.0, 172.8 (C(O)).
1.49 (s, 18H, (CH3)3C), 1.93–2.25 (br, 2H, CHCH2CH2), 2.38 31P{1H} NMR (122 MHz, CDCl3) δ (ppm): 32.9 (s), 31.0 (s).
(d, J = 12 Hz, 2H, ArCH2CH2), 2.53–3.37 (br, 21H, CHCOOtBu ESI-HRMS for C41H74N5O9P [M + H]+, m/z calcd 812.5296,
and CH2COOtBu, CH2 ring), 7.63 (d, J = 9 Hz, 2H, ArH), 8.17 found 812.5296.
(d, J = 9 Hz, 2H, ArH). 13C{1H} NMR (75 MHz, CDCl3) δ (ppm):
Ligand L2. Bromotrimethylsilane (0.57 mL, 4.31 mmol) was
28.0, 28.3 ((CH3)3C), 31.6 (ArCH2CH2), 32.3 (ArCH2CH2), 45.8, added to a solution of 12 (0.35 g, 0.431 mmol) in anhydrous
48.2, 49.5, 50.8 (CH2 ring), 55.5 (CH2COOtBu), 60.3 dichloromethane (15 mL) cooled at 0 °C. The reaction mixture
(CHCOOtBu), 81.6, 82.2 ((CH3)3C), 123.7, 130.1, 146.6, 148.9 was allowed to warm up to room temperature and was stirred
(ArC), 170.04, 170.9 (C(O)). ESI-HRMS for C34H57N5O8 [M + H]+, for 16 h. The reaction mixture was concentrated under reduced
m/z calcd 664.4279, found 664.4282.
pressure to remove the excess reagent and solvent. The residue
Compound 11. Compound 10 (0.93 g, 1.40 mmol) was dis- was suspended in anhydrous dichloromethane (15 mL), tri-
solved in anhydrous acetonitrile (20 mL) and potassium car- fluoroacetic acid (10 mL) was added at 0 °C and the solution
bonate (0.23 g, 1.68 mmol), diethyl (3-bromopropyl) was stirred at room temperature for 16 h. The reaction mixture
phosphonate (0.32 mL, 1.68 mmol) and catalytic amount of was concentrated under reduced pressure, the residue was dis-
potassium iodide were added at room temperature. The result- solved in methanol and added to diethyl ether under vigorous
ing mixture was stirred at 70 °C for 16 h. After cooling to room stirring. After filtration and drying, L2 was obtained as an off-
temperature, the insoluble materials were removed by fil- white solid (0.21 g, 84%).
tration, the filtrate was concentrated under reduced pressure
to afford residue which was purified by silica gel ArCH2CH2, P(O)CH2CH2CH2), 2.50–4.0 (br, 23H, CH2 ring,
1H NMR (300 MHz, D2O) δ (ppm): 1.53–2.50 (br, 8H,
a
column chromatography (50% methanol/dichloromethane/ CH2COOH and CHCOOH, NCH2CH2CH2P(O)), 7.34 (d, J =
2% ammonia) to obtain 11 as a light yellow oil (0.75 g, 63%).
7.5 Hz, 2H, ArH), 7.42 (d, J = 7.5 Hz, 2H, ArH). 13C{1H} NMR
1302 | Org. Biomol. Chem., 2013, 11, 1294–1305
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
1
(75 MHz, D2O) δ (ppm): 17.4, 23.9 (d, JPC = 135 Hz), 26.4,
1H NMR (300 MHz, D2O) δ (ppm): 1.01–2.43(br, 18H),
31.7, 47.0, 47.3, 47.4, 48.1, 49.4, 49.6, 49.8, 50.2, 53.6, 54.2, 2.49–4.06 (br, 28H), 4.35 (s, 1H), 4.53 (s, 1H). 13C{1H} NMR
1
54.5, 61.3, 123.4, 128.3, 130.6, 141.3 (ArC), 171.3, 173.3, 174.5 (75 MHz, D2O) δ (ppm): 17.4, 24.1, 25.0 (d, Jpc = 134 Hz, P(O)
(CvO). 31P{1H} NMR (122 MHz, D2O) δ (ppm): 26.6 (s). CH2CH2), 25.2, 25.4, 27.7, 28.0, 28.3, 35.5, 38.6, 39.8, 47.9,
ESI-HRMS for C25H42N5O9P: [M − H]−, m/z calcd 586.2647, 48.2, 49.5, 52.5, 53.1, 53.8, 54.0, 54.3, 54.5, 55.4, 60.2, 62.1,
found 586.2643.
64.6, 165.2, 172.0, 173.7, 174.6, 176.5 (C(O)). 31P{1H} NMR
Compound 13. 1-Ethyl-3-(3-dimethylaminopropyl) carbodii- (122 MHz, D2O) δ (ppm): 23.6 (s), 24.0 (s). ESI-HRMS for
mide (58 mg, 0.301 mmol) was added to a suspension of C31H56N7O11PS: [M − H]−, m/z calcd 764.3423, found 764.3422.
biotin (88 mg, 0.36 mmol) and 4-dimethylamino pyridine
Ligand L4. The compound L4 was synthesized from 14 fol-
(76 mg, 0.60 mmol) in anhydrous dimethylformamide (8 mL) lowing the same procedure as described for L2. The purifi-
at room temperature. The solution of amine 9 (0.23 g, cation was accomplished by RP-HPLC, following the method
0.30 mmol) in anhydrous dimethylformamide (6 mL) was as described in the Experimental section (68 mg, 58%).
added slowly after 10 minutes, and the mixture was stirred for
1H NMR (300 MHz, D2O) δ (ppm): 1.00–2.00 (br, 12H),
another 16 h. The reaction mixture was concentrated under 2.00–3.85 (br, 30H), 4.31 (s, 1H), 4.51 (s, 1H), 7.29 (m, 4H).
reduced pressure and the resulting residue was dissolved in 13C{1H} NMR (75 MHz, D2O) δ (ppm): 16.8, 24.8 (d, J = 134 Hz,
dichloromethane (100 mL) and water (100 mL). The organic P(O)CH2CH2), 25.1, 27.7, 28.0, 29.1, 31.6, 36.2, 39.7, 46.8, 48.7,
layer was separated and the aqueous layer was extracted with 48.9, 49.3, 49.5, 50.3, 53.5, 53.8, 55.3, 56.4, 57.5, 60.2, 62.0,
dichloromethane (2 × 75 mL). The combined organic layers 63.5, 122.2, 129.4, 135.2, 137.6 (ArC), 165.2, 173.1, 175.2,
were washed with brine, dried over Na2SO4 and concentrated 175.2, 175.4 (C(O)). 31P{1H} NMR (122 MHz, D2O) δ (ppm):
under reduced pressure. The residue was purified by silica gel 23.0 (s), 23.7(s). ESI-HRMS for C35H56N7O11PS: [M − H]−, m/z
column chromatography (10% methanol/dichloromethane) to calcd 812.3423, found 812.3426.
yield product 13 as an off-white solid (0.16 g, 54%).
Preparation of Ln3+ complexes. GdL1–4 and EuL1–2 were pre-
1H NMR (300 MHz, CDCl3) δ (ppm): 1.20 (t, J = 6 Hz, 6H, pared from the respective solution of the ligand (1.0 equiv.)
(P(O)OCH2CH3), 1.26–1.44 (br, 33H), 1.44–1.84 (br, 12H), and the aqueous solution of GdCl3·6H2O for L1–4 or
1.87–2.60 (br, 13H), 2.61–3.40 (br, 16H,), 3.81–4.06 (m, 4H, EuCl3·6H2O (1.0 equiv.) for L1–2. For each complex the lantha-
OCH2CH3), 4.24 (m, 1H, CH), 4.39 (m, 1H, CH), 5.91 (br s, 1H, nide salt solution was added portionwise over 6 h. Once no
NH), 6.33 (br s, 1H, NH), 8.48 (br s, 1H, NH). 13C{1H} NMR further ligand could be observed by ESI-MS, the reaction
(75 MHz, CDCl3) δ (ppm): 16.2 (d, J = 5.3 Hz, P(O)OCH2CH3), mixture was stirred at 60 °C for an additional 16 h. The pH
18.9, 22.5, 24.4, 25.6, 26.2, 27.6, 27.6, 27.8, 29.0, 35.44, 4.34, value of the solution was periodically checked and adjusted to
40.4, 44.5, 47.0, 48.1, 49.0, 50.6, 52.3, 53.2, 54.6, 54.8, 55.3, 6.5–7.0 by using an aqueous solution of sodium hydroxide
55.8, 57.0, 59.9, 61.2, 61.3, 61.5, 82.2, 82.2, 83.0, 172.80, 173.6, (0.5 M). The solutions were then treated with Chelex® 100
173.8, 174.9. 31P NMR (75 MHz, CDCl3) δ (ppm): 26. 6 (s). sodium form for 4 h at room temperature to remove any excess
ESI-HRMS for C47H89N7O11PS: [M − H]−, m/z calcd 990.6072, lanthanide ion. The absence of free Gd3+/Eu3+ was confirmed
found 990.6080.
with the xylenol orange test.32 The complexes were character-
Compound 14. Following the same procedure as described ized by ESI-LRMS in negative mode and the appropriate
for compound 13, compound 14 was isolated as a light yellow isotope pattern distribution for Gd3+ and Eu3+ was obtained.
solid (98 mg, 56%).
Relaxometric experiments
1H NMR (300 MHz, CDCl3) δ (ppm): 1.28 (t, J = 6 Hz, 6H,
OCH2CH3), 1.40–1.53 (br, 27H, (CH3)3C), 1.54–3.73 (br, 39H), The relaxivities of the complexes GdL1–2 are an average of
3.91–4.18 (br, 5 H, CHCOOtBu, OCH2CH3), 4.30 (s, 1H, NHCH three measurements at concentrations ranging from 1–3 mM
(CH)CH2), 4.48 (s, 1H, NHCH(CH)CH2), 5.42–5.88(m, 2H, in H2O. The pH was adjusted with solid LiOH and p-TsOH. For
SCH2CH), 7.01 (d, J = 6 Hz, 2H, ArH), 7.53 9d, J = 6 Hz, 2H, each measurement the exact concentration of the Gd3+
ArH), 8.34 (br, 1H, NH). 13C{1H} NMR (75 MHz, CDCl3) complex was determined using the bulk magnetic suscepti-
δ (ppm): 14.1, 16.3 (d, J = 6 Hz, P(O)OCH2CH3), 18.9, 23.4 bility shift.
1
(d, JPC = 135 Hz, P(O)CH2CH2CH2), 25.5, 27.7, 27.8, 28.0
((CH3)3C), 29.6, 33.4, 36.6, 40.2, 43.9, 47.0, 47.8, 48.5, 50.7,
Luminescence steady-state and time resolved experiments
52.5, 54.6, 54.9, 55.4, 55.9, 56.5, 58.9, 60.4, 61.47, 61.5, 61.9, All experiments were performed with 5 mM EuL1–2 at 25 °C.
82.2, 82.5, 82.9 ((CH3)3C), 120.3, 129.0, 135.9, 137.0 (ArC), The steady-state measurements were performed in H2O with
164.5, 172.5, 173.0, 174.3, 175.0 (C(O)). 31P{1H} NMR the excitation and emission slits set to 1 nm bandpass. The
(122 MHz, CDCl3)
C51H89N7O11PS: [M
1038.6073.
δ
(ppm): 31.1 (s). ESI-HRMS for time resolved measurements were performed in H2O and D2O.
H]−, m/z calcd 1038.6072, found Excitation and emission slits were set to 15 and 5 nm band-
pass respectively. Datasets are recorded with a 100 μs delay
−
Ligand L3. The compound L3 was synthesized from 13 fol- and 10 μs resolution, and are an average of 15 scans. Each
lowing the same procedure as described for L2. The purifi- reported value is the mean of three independent measure-
cation was accomplished by RP-HPLC, following the method ments and obtained curves are fitted to a first order exponen-
as described in the Experimental section (0.16 g, 49%).
tial decay with R2 > 0.99.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 1294–1305 | 1303

View Article Online
Paper
Organic & Biomolecular Chemistry
Spectrophotometric assay
residuals. The obtained T1/T2 values of the samples were
converted to R1 (= 1/T1) and R2 (= 1/T2).
Stock solutions of avidin and ANS were prepared at concen-
trations 4 μM and 40 μM, respectively, in 0.15 M sodium phos-
phate buffer at pH 7. Buffered solutions of avidin (1 mL, 4 μM)
and ANS (1 mL, 40 μM) were mixed 1 : 1 (v/v) and the resulting
conjugate was titrated with biotin/GdL3/GdL4 (320 μM) in 5 μL
aliquots. All measurements were performed at 25 °C with
the excitation and emission wavelengths of 328 and 408 nm
respectively, and excitation and emission slits of 2 nm
bandpass.
r1;2 ¼ ðR1;2obs ꢀ R1;2diaÞ=½SCAꢁ
ð4Þ
Relaxivities r1/r2 were calculated by using eqn (4), where
R1,2obs – observed relaxation rate, R1,2dia – diamagnetic contri-
bution to the relaxation rate and [SCA] – applied concentration
of SCA.
Competitive assay
Stock solutions of avidin and GdL3/GdL4 were prepared at
0.225 mM and 0.250 mM respectively in HEPES buffer
(25 mM, pH 7). A series of different concentrations of biotin
were prepared from 250 mM to 1 mM also in HEPES buffer.
Avidin (66.7 μL) was mixed with GdL3/GdL4 (200 μL) and
samples with 0 to 500 times excess of biotin were obtained by
adding different volumes of biotin stock solutions 10 min after
the preparation of the avidin–GdL3/GdL4 mixture. The volume
was adjusted to 400 μL by adding HEPES buffer. Mixtures were
incubated at 37 °C for 5 h. Longitudinal and transverse relax-
ation times were recorded on the 3T MRI scanner.
MRI phantom experiments
MR imaging of the samples was performed at 3T (123 MHz,
21 °C) on a clinical human MR scanner (MAGNETOM Tim
Trio, Siemens Healthcare, Germany). Stock solutions of GdL3–4
(0.25 mM) and avidin (0.225 mM) were prepared in HEPES
buffer (25 mM, pH 7.4) and water, respectively. The concen-
tration of GdL3–4 was kept constant (0.125 mM) while the con-
centration of avidin varied up to a ratio of 0.9 avidin– GdL3–4
.
The mixtures were kept for 3 h at 37 °C for incubation,
followed by the measurement in a 3T MRI scanner. For
the subsequent pH-dependent MRI experiments of
GdL3–4 with avidin, following buffers were used to adjust
pH:MES buffer for pH 5.5 and 6.3, HEPES buffer for
pH 7.0, 7.5 and 8.0, and CHES buffer for pH 8.6 and 9.0. The
Acknowledgements
The authors thank Dr Sven Gottschalk for help in performing
the MRI experiments. We also thank Dr Matteo Placidi,
Dr Ilgar Mamedov and Dr Pascal Kadjane for the helpful dis-
cussions. The financial support of the Max-Planck Society,
German Research Foundation (DFG, grant AN 716/2-1) and
German Ministry for Education and Research (BMBF, FKZ:
01EZ0813) is gratefully acknowledged.
obtained results are mean
experiments.
SEM of five independent
Longitudinal relaxation times (T1) were measured using an
inversion recovery sequence to obtain images from an axial
slice of 1 mm thickness through the samples. The inversion
time (Ti) was varied from 23 ms to 3000 ms in 12 steps. Images
were read out with a turbo spin echo technique, acquiring
5 echoes per scan. The repetition time (TR) was 10 000 ms to
ensure complete relaxation. Six averages per Ti were possible
within 18 min. For T2, a home-written spin-echo sequence was
used with echo times varying from 25 ms to 275 ms in 10 steps
and a repetition time of 8 s. Diffusion sensitivity was reduced
by minimizing the crusher gradients surrounding the refocus-
ing pulse. A matrix of 256 × 256 voxels was used over a field-of-
view of 110 × 110 mm2 resulting in a voxel volume of 0.43 ×
0.43 × 1 mm3.
Notes and references
1 R. B. Lauffer, Chem. Rev., 1987, 87, 901–927.
2 J. S. Ananta, B. Godin, R. Sethi, L. Moriggi, X. Liu,
R. E. Serda, R. Krishnamurthy, R. Muthupillai,
R. D. Bolskar, L. Helm, M. Ferrari, L. J. Wilson and
P. Decuzzi, Nat. Nanotechnol., 2010, 5, 815–821.
3 A. E. Merbach and E. Toth, The Chemistry of Contrast Agents
in Medical Magnetic Resonance Imaging, Wiley, Chichester,
2001.
4 J. L. Major and T. J. Meade, Acc. Chem. Res., 2009, 42,
893–903.
5 B. Yoo and M. D. Pagel, Front. Biosci., 2008, 13, 1733–1752.
6 G. Angelovski and I. Mamedov, Curr. Inorg. Chem., 2011, 1,
76–90.
Data analysis was performed by fitting of relaxation curves
with self-written routines under MATLAB 7.1 R14 (The Math-
works Inc., United States). The series of T1 (with varying t = TI)
and T2 (with varying t = TE) relaxation data were fitted to the
eqn (2) and (3).
S ¼ S0 ð1 ꢀ exp ðꢀt=T1ÞÞ þ S
exp ðꢀt=T1Þ
ð2Þ
ð3Þ
ðTI¼0Þ
7 N. Raghunand, C. Howison, A. D. Sherry, S. R. Zhang and
R. J. Gillies, Magn. Reson. Med., 2003, 49, 249–257.
8 L. M. De Leon-Rodriguez, A. J. M. Lubag, C. R. Malloy,
G. V. Martinez, R. J. Gillies and A. D. Sherry, Acc. Chem.
Res., 2009, 42, 948–957.
S ¼ S0 exp ðꢀt=T2Þ
Nonlinear least-squares fitting of three parameters S0,
S(TI=0), and T1/T2 was done for manually selected regions of
interest with the Trust-Region Reflective Newton algorithm
implemented in MATLAB. The quality of the fit was controlled
by visual inspection and by calculating the mean errors and
9 S. R. Zhang, K. C. Wu and A. D. Sherry, Angew. Chem., Int.
Ed., 1999, 38, 3192–3194.
1304 | Org. Biomol. Chem., 2013, 11, 1294–1305
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
10 M. P. Lowe, D. Parker, O. Reany, S. Aime, M. Botta, 23 A. Beeby, I. M. Clarkson, R. S. Dickins, S. Faulkner,
G. Castellano, E. Gianolio and R. Pagliarin, J. Am. Chem.
Soc., 2001, 123, 7601–7609.
11 P. Caravan, Chem. Soc. Rev., 2006, 35, 512–523.
12 S. Aime and P. Caravan, J. Magn. Reson. Imaging, 2009, 30,
1259–1267.
D. Parker, L. Royle, A. S. de Sousa, J. A. G. Williams and
M. Woods, J. Chem. Soc., Perkin Trans. 2, 1999, 493–503.
24 W. D. Horrocks, J. P. Bolender, W. D. Smith and
R. M. Supkowski, J. Am. Chem. Soc., 1997, 119, 5972–5973.
25 M. Wilchek and E. A. Bayer, Methods Enzymol., 1990, 184,
5–13.
13 M. M. Ali, M. Woods, P. Caravan, A. C. L. Opina, M. Spiller,
J. C. Fettinger and A. D. Sherry, Chem.–Eur. J., 2008, 14, 26 P. Lehtolainen, T. Wirth, A. K. Taskinen, P. Lehenkari,
7250–7258.
O. Leppanen, M. Lappalainen, K. Pulkkanen, A. Marttila,
V. Marjomaki, K. J. Airenne, M. Horton, M. S. Kulomaa and
S. Yla-Herttuala, Gene Ther., 2003, 10, 2090–2097.
14 L. Frullano, C. Catana, T. Benner, A. D. Sherry and
P. Caravan, Angew. Chem., Int. Ed., 2010, 49, 2382–2384.
15 E. Gianolio, R. Napolitano, F. Fedeli, F. Arena and S. Aime, 27 P. L. Weiden, H. B. Breitz, O. Press, J. W. Appelbaum,
Chem. Commun., 2009, 6044–6046.
16 I. Mamedov, A. Mishra, G. Angelovski, H. A. Mayer,
J. K. Bryan, S. Gaffigan, D. Stone, D. Axworthy, D. Fisher
and J. Reno, Cancer Biother. Radiopharm., 2000, 15, 15–29.
L. O. Palsson, D. Parker and N. K. Logothetis, Dalton 28 J. Su, F. Chen, V. L. Cryns and P. B. Messersmith, J. Am.
Trans., 2007, 5260–5267.
Chem. Soc., 2011, 133, 11850–11853.
17 N. M. Green, Methods Enzymol., 1990, 184, 51–67.
18 K. P. Eisenwiener, P. Powell and H. R. Macke, Bioorg. Med.
Chem. Lett., 2000, 10, 2133–2135.
19 Z. Kovacs and A. D. Sherry, Synthesis, 1997, 759–763.
20 M. Giardiello, M. P. Lowe and M. Botta, Chem. Commun.,
2007, 4044–4046.
21 N. J. Wardle, A. H. Herlihy, P.-W. So, J. D. Bell
and S. W. A. Bligh, Bioorg. Med. Chem., 2007, 15, 4714–
4721.
29 D. M. Mock, G. Langford, D. Dubois, N. Criscimagna and
P. Horowitz, Anal. Biochem., 1985, 151, 178–181.
30 R. A. Dwek, Nuclear Magnetic Resonance (N.M.R.) in Bio-
chemistry: Applications to Enzyme Systems, Clarendon Press,
Oxford, 1973.
31 J. B. Livramento, C. Weidensteiner, M. I. M. Prata,
P. R. Allegrini, C. Geraldes, L. Helm, R. Kneuer,
A. E. Merbach, A. C. Santos, P. Schmidt and E. Toth, Con-
trast Media Mol. Imaging, 2006, 1, 30–39.
22 D. M. Corsi, C. Platas-Iglesias, H. van Bekkum and 32 B. Paull and P. R. Haddad, TrAC, Trends Anal. Chem., 1999,
J. A. Peters, Magn. Reson. Chem., 2001, 39, 723–726.
18, 107–114.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 1294–1305 | 1305