Ratiometric pH Imaging with a CoII2 MRI Probe via CEST Effects of Opposing pH Dependences

Article abstract of DOI:10.1021/jacs.7b08574

We report a Co₂-based magnetic resonance (MR) probe that enables the ratiometric quantitation and imaging of pH through chemical exchange saturation transfer (CEST). This approach is illustrated in a series of air- and water-stable Co^II₂ complexes featuring CEST-active tetra(carboxamide) and/or hydroxyl-substituted bisphosphonate ligands. For the complex bearing both ligands, variable-pH CEST and NMR analyses reveal highly shifted carboxamide and hydroxyl peaks with intensities that increase and decrease with increasing pH, respectively. The ratios of CEST peak intensities at 104 and 64 ppm are correlated with solution pH in the physiological range 6.5-7.6 to construct a linear calibration curve of log(CEST_{104 ppm}/CEST_{64 ppm}) versus pH, which exhibits a remarkably high pH sensitivity of 0.99(7) pH unit^-1 at 37 °C. In contrast, the analogous Co^II₂ complex with a CEST-inactive bisphosphonate ligand exhibits no such pH response, confirming that the pH sensitivity stems from the integration of amide and hydroxyl CEST effects that show base- and acid-catalyzed proton exchange, respectively. Importantly, the pH calibration curve is independent of the probe concentration and is identical in aqueous buffer and fetal bovine serum. Furthermore, phantom images reveal analogous linear pH behavior. The Co^II₂ probe is stable toward millimolar concentrations of H₂PO₄^-/HPO₄^2-, CO₃^2-, SO₄^2-, CH₃COO^-, and Ca²⁺ ions, and more than 50% of melanoma cells remain viable in the presence of millimolar concentrations of the complex. The stability of the probe in physiological environments suggests that it may be suitable for in vivo studies. Together, these results highlight the ability of dinuclear transition metal PARACEST probes to provide a concentration-independent measure of pH, and they provide a potential design strategy toward the development of MR probes for ratiometric pH imaging.

Full text of DOI:10.1021/jacs.7b08574

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
II2
Ratiometric pH Imaging with a Co MRI Probe
via CEST Effects of Opposing pH Dependences
Agnes E. Thorarinsdottir, Kang Du, James H. P. Collins, and T. David Harris
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b08574 • Publication Date (Web): 13 Oct 2017
Downloaded from http://pubs.acs.org on October 13, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
Ratiometric pH Imaging with a CoII
Opposing pH Dependences
MRI Probe via CEST Effects of
2
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
†
‡
,†
Agnes E. Thorarinsdottir, Kang Du, James H. P. Collins, and T. David Harris*
†Department of Chemistry, Northwestern University, Evanston, Illinois 60208, USA
‡Advanced Magnetic Resonance Imaging and Spectroscopy Facility, University of Florida, Gainesville, Florida 32611, USA
ABSTRACT: We report a Co -based magnetic resonance (MR) probe that enables the ratiometric quantitation and imaging of pH
through chemical exchange saturation transfer (CEST). This approach is illustrated in a series of air- and water-stable Co com-
2
plexes featuring CEST-active tetra(carboxamide) and/or hydroxyl-substituted bisphosphonate ligands. For the complex bearing
both ligands, variable-pH CEST and NMR analyses reveal highly-shifted carboxamide and hydroxyl peaks with intensities that
increase and decrease with increasing pH, respectively. The ratios of CEST peak intensities at 104 and 64 ppm are correlated with
solution pH in the physiological range 6.5–7.6 to construct a linear calibration curve of log(CEST104 ppm/CEST64 ppm) vs pH, which
2
II
−1
II
exhibits a remarkably high pH sensitivity of 0.99(7) pH unit at 37 °C. In contrast, the analogous Co
2
complex with a CEST-
inactive bisphosphonate ligand exhibits no such pH response, confirming that the pH sensitivity stems from the integration of am-
ide and hydroxyl CEST effects that show base- and acid-catalyzed proton exchange, respectively. Importantly, the pH calibration
curve is independent of the probe concentration, and is identical in aqueous buffer and fetal bovine serum. Furthermore, phantom
images reveal analogous linear pH behavior. The Co
SO , CH COO , and Ca ions, and more than 50% of melanoma cells remain viable in the presence of millimolar concentrations
of the complex. The stability of the probe in physiological environments suggests that it may be suitable for in vivo studies. Togeth-
er, these results highlight the ability of dinuclear transition metal PARACEST probes to provide a concentration-independent meas-
ure of pH, and they provide a potential design strategy toward the development of MR probes for ratiometric pH imaging.
II
−
2−
2−
2
probe is stable toward millimolar concentrations of H
2 4 4 3
PO /HPO , CO ,
2−
−
2+
4 3
quencies.7ac,9 Moreover, the ability of these probes to map
extracellular pH in biological environments has been demon-
strated.
Despite these promising advances, the development of a
single probe that features CEST peaks shifted outside the
tissue magnetization transfer window, is highly responsive in
the physiological pH range, and displays good stability under
physiological conditions remains elusive. Toward this end,
INTRODUCTION
7
c,8,9cegh
Acidic extracellular pH features prominently in a number of
pathological conditions, including cancer, ischemia,
1
1f,2
in-
1
h,2c
1h
flammation,
and infection. As such, the ability to meas-
1
0
ure and spatially map tissue pH would provide valuable in-
formation regarding the role of acidosis in both the initiation
1–3
and the progression of diseases. Toward this end, magnetic
7d,9f,11
resonance imaging (MRI) represents an ideal non-invasive
modality for probing pH, owing to its ability to deeply pene-
trate tissue and generate images with high spatiotemporal
transition metal-based PARACEST probes
offer potential
advantages over their more common lanthanide counterparts.
Specifically, the chemical shifts of transition metal complexes
are primarily governed by through-bond interactions rather
than the dominant through-space interactions of lanthanide
complexes, which renders exchangeable protons extremely
sensitive to the metal coordination environment and thus ame-
4
resolution. Indeed, a number of MR techniques have been
developed to measure pH in vivo, and these methods common-
ly rely on the presence of pH-sensitive exogenous molecular
probes. Among these probes, complexes that exhibit the para-
magnetic chemical exchange saturation transfer (PARACEST)
effect, where exchange of protons on a paramagnetic molecule
with those of bulk H O upon selective irradiation is exploited
2
to generate contrast, are particularly well suited, due to large
hyperfine shifts of their labile protons and the inherent pH
1
2
nable to the design of responsive probes.
In conjunction with the employment of transition metal
ions, one can envision incorporation of two distinct ligand
scaffolds on a single complex, where the two ligands exhibit
CEST effects with opposing pH dependences. For such a
system, the ratio of the two CEST peak intensities should
change dramatically as a function of pH. Along these lines, we
recently reported Fe , Cu , and CuGa complexes supported by
modular dinucleating tetra(carboxamide) ligand and
bisphosphonate ancillary ligands.
ligands, etidronate notably features a pendent hydroxyl group
that can potentially give rise to CEST. Indeed, the presence of
both base-catalyzed exchange of carboxamide protons and
5
6
sensitivity of their exchange rates.
The intrinsic concentration dependence of the CEST effect
intensity requires that the concentration of a PARACEST
probe in the imaged region must be known. A number of
strategies have been reported to overcome this limitation,
including the development of probes with pH-dependent
2
2
a
11lm
Among these ancillary
7
8
changes in the frequency or linewidth of the CEST peak or
in the ratio of CEST intensities from two presaturation fre-
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 13
acid-catalyzed proton exchange of etidronate highlights the
potential of these dinuclear complexes to exhibit pronounced
bubbler filled with concentrated sulfuric acid. All other rea-
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
gents and solvents were purchased from commercial vendors
and used without further purification. Experimental details on
the synthesis of ligands and organic precursors are provided in
the Supporting Information.
II
pH sensitivity. Accordingly, we herein report a Co
2
complex
that displays CEST spectra featuring highly pH-sensitive and
shifted peaks, by virtue of CEST-active carboxamide and
hydroxyl groups with opposing pH dependences. The complex
exhibits excellent chemical stability and retains its CEST
activity in fetal bovine serum, which underscores the potential
suitability of this and related complexes for pH quantitation in
living systems.
Synthesis of Na[LCo
2
(etidronate)]∙0.2NaNO
3
∙2.7H O
2
(1). A pink solution of Co(NO
3
)
2
·6H O (71 mg, 0.24 mmol) in
2
MeOH (2 mL) was added dropwise to a stirring yellow sus-
pension of HL (52 mg, 0.12 mmol) in MeOH (3 mL) to give a
dark orange solution. To this solution, a colorless solution of
etidronic acid monohydrate (27 mg, 0.12 mmol) in MeOH (2
mL) was added dropwise to give a light orange solution. Sub-
sequent addition of sodium methoxide (33 mg, 0.61 mmol) in
MeOH (2 mL) resulted in the formation of a light orange
slurry. After stirring at 25 °C for 3 h, the orange solid was
collected by vacuum filtration, washed with MeOH (5 mL)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
EXPERIMENTAL SECTION
General Considerations. Unless otherwise specified, the
manipulations described below were carried out at ambient
atmosphere and temperature. Air- and water-free manipula-
tions were performed under a dinitrogen atmosphere in a
Vacuum Atmospheres Nexus II glovebox or using standard
Schlenk line techniques. Glassware was oven-dried at 150 °C
for at least 4 h and allowed to cool in an evacuated antecham-
ber prior to use in the glovebox. Acetonitrile (MeCN), diethyl
and Et O (15 mL), and dried under reduced pressure for 16 h
2
to give 1 (44 mg, 44%) as an orange solid. Anal. Calcd. for
C H Co N Na O17.3P2: C, 25.96; H, 3.80; N, 12.11%.
1
8
31.4
2
7.2
1.2
Found: C, 25.96; H, 3.83; N, 12.16%. ICP-OES: Co:P =
1.02:1.00. UV-Vis absorption spectrum (64 M; 50 mM 4-(2-
hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)
2
ether (Et O), diisopropylethylamine (DIPEA), and methanol
(MeOH) were dried using a commercial solvent purification
system from Pure Process Technology and stored over 3 or 4
−1
−1
buffered at pH 7.4, 25 °C): 375 nm (= 13800 M cm ).
ESI-MS (m/z): Calcd. for
([LCo (etidronate)] ): 743.97, found: 743.95; calcd. for
2
Å molecular sieves prior to use. The solvent H O was obtained
18 2 7 14 2
C H26Co N O P
from a purification system from EMD Millipore. Deuterated
solvents were purchased from Cambridge Isotope Laboratories
and Sigma Aldrich. The synthesis of 2,2′-iminobis(acetamide)
was carried out according to a previously reported proce-
−
2
+
18 2 7 14 2 2
C H28Co N O P ([LCo (etidronate)+2H] ): 745.98, found
745.92. FT-IR (ATR, cm ): 3341 (m, broad); 3178 (m,
broad); 2930 (w); 1665 (s); 1595 (m); 1499 (w); 1446 (m);
−1
1
1l
dure. Anhydrous hydrogen chloride gas was generated by
adding concentrated hydrochloric acid to a stirring solution of
concentrated sulfuric acid. The gas was passed through a
1
307 (s); 1097 (s); 1060 (s); 911 (m); 826 (w); 799 (m); 751
(w); 706 (m); 539 (s); 470 (s). Slow diffusion of MeCN vapor
into a concentrated solution of 1 in H
2
O afforded dark orange
−
2+
4−
−
−
Figure 1. Reaction of L , Co , and etidronate (left) or CMDP (center) to form [LCo
1
2
(etidronate)] or [LCo
2
(CMDP)] , as observed in
−
2+
−
2
and 2, respectively. Reaction of L′ , Co , and etidronate to form [L′Co (etidronate)] , as observed in 3 (right).
2
ACS Paragon Plus Environment

Page 3 of 13
Journal of the American Chemical Society
plate-shaped crystals of Na[LCo
suitable for single-crystal X-ray diffraction analysis.
Synthesis of Na[LCo (CMDP)]∙4.5H O∙MeOH (2). A
pink solution of Co(NO ·6H O (67 mg, 0.23 mmol) in
MeOH (2 mL) was added dropwise to a stirring yellow sus-
pension of HL (49 mg, 0.11 mmol) in MeOH (3 mL) to give a
dark orange solution. A colorless solution of chloro-
2
(etidronate)]∙6.8H
2
O (1′)
APEX2 version 2014.11-0.13 Absorption corrections were
1
4
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
applied using the program SADABS. Space group assign-
ments were determined by examining systematic absences, E-
statistics, and successive refinement of the structures. Struc-
tures were solved using direct methods in SHELXT and re-
2
2
3
)
2
2
1
5
16
fined by SHELXL operated within the OLEX2 interface.
All hydrogen atoms were placed at calculated positions using
suitable riding models and refined using isotropic displace-
ment parameters derived from their parent atoms. In the crys-
methanediphosphonic acid (H
MeOH (2 mL) was then slowly added, resulting in a light
orange solution. Subsequently, colorless solution of
4
CMDP) (27 mg, 0.13 mmol) in
4
−
tal structure of 2′, the Cl atom on the CMDP ligand is posi-
tionally disordered over two positions. The occupancy of the
Cl was freely refined over the two positions. Partially-
a
Na(OMe) (31 mg, 0.58 mmol) in MeOH (2 mL) was added
dropwise to give a light orange suspension. The reaction mix-
ture was stirred at 25 °C for 2.5 h, and then the orange solid
was collected by vacuum filtration, washed with MeOH (10
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
occupied solvent H O molecules not directly bonded to the
sodium ions were modeled isotropically. Thermal parameters
for all other non-hydrogen atoms were refined anisotropically.
Crystallographic data for these compounds at 100 K and the
details of data collection are listed in Table S1.
mL) and Et
2
O (15 mL), and dried under reduced pressure for
1
9 h to give 2 (46 mg, 45%) as an orange solid. Anal. Calcd.
2 7 2
for C18H36ClCo N NaO18.5P : C, 24.43; H, 4.10; N, 11.08%.
Found: C, 24.49; H, 3.65; N, 10.74%. ICP-OES: Co:P =
1
31
1
NMR Spectroscopy. H and P{ H} NMR spectra of lig-
ands and organic precursors were collected at 25 °C at 500 and
202 MHz frequencies, respectively, on Agilent DD2 500 MHz
(11.7 T) or Varian Inova 500 MHz (11.7 T) spectrometers, or
on an automated Agilent DD MR 400 MHz (9.4 T) spectrome-
1
.02:1.00. UV-Vis absorption spectrum (80 M; 50 mM
−1
HEPES buffered at pH 7.4, 25 °C): 375 nm (= 14100 M
cm ). ESI-MS (m/z): Calcd. for
−1
17 2 7 13 2
C H23ClCo N O P
−
2
([LCo (CMDP)] ): 747.92, found 747.93; calcd. for
+
13
1
C
17
H
25ClCo
2
N
7
O
13
P
2
([LCo
2
(CMDP)+2H] ): 749.93, found
ter at 400 and 162 MHz frequencies, respectively. C{ H}
NMR spectra of ligands were obtained at 25 °C on a Bruker
Avance III 500 MHz (11.7 T) system at 126 MHz frequency.
−1
749.89. FT-IR (ATR, cm ): 3339 (w, broad); 3173 (w,
broad); 1663 (s); 1596 (m); 1500 (w); 1447 (m); 1311 (s);
1133 (s); 1097 (s); 911 (m); 878 (w); 799 (w); 752 (m); 688
(m); 662 (m); 529 (s); 475 (s). Slow diffusion of MeCN vapor
1
Variable-temperature H NMR spectra of compounds 1–3
were collected on Agilent DD2 500 MHz (11.7 T) and Agilent
1
into a concentrated solution of 2 in H
block-shaped crystals of Na[LCo (CMDP)]∙8.2H
ble for single-crystal X-ray diffraction analysis.
Synthesis of Na[L′Co (etidronate)]∙1.2NaNO
3). A pink solution of Co(NO ·6H O (54 mg, 0.19 mmol) in
2
O gave dark orange
DD2 400 MHz (9.4 T) spectrometers. H NMR spectra of
O (2′) suita-
samples in aqueous solutions containing 50 mM HEPES and
100 mM NaCl buffered at various pH values were acquired
2
2
2
using D O in an inner capillary to lock the sample. Variable-
2
3
∙1.9H O
2
1
pH H NMR spectra of samples in fetal bovine serum (FBS)
were recorded similarly. The pH of commercially available
FBS (Fisher Scientific, catalog no. MT35010CV) was adjusted
to the desired values by addition of minimal amounts of dilute
aqueous nitric acid and sodium hydroxide solutions. Chemical
shift values () are reported in ppm and referenced to residual
(
3
)
2
2
MeOH (2 mL) was added dropwise to a stirring yellow solu-
tion of HL′ (50 mg, 0.093 mmol) in MeOH (2 mL) to give an
orange solution. After stirring at 25 °C for 5 min, a colorless
solution of etidronic acid monohydrate (21 mg, 0.093 mmol)
in MeOH (2 mL) was added dropwise, followed by addition of
sodium methoxide (25 mg, 0.46 mmol) in MeOH (2 mL). The
resulting orange solution was stirred at 25 °C for 2 h, collected
by vacuum filtration, and dried under reduced pressure. The
resulting red-orange residue was stirred in MeCN (10 mL) for
1
signals from the deuterated solvents ( H NMR spectra: 7.26
ppm for CDCl
3
, 4.79 ppm for D
2
O, and 3.31 ppm for CD
, and 49.00 ppm
O were carried out
with 5% (v/v) MeOH added as a reference ( = 49.50 ppm).
3
OD;
1
3
1
C{ H} NMR spectra: 77.16 ppm for CDCl
3
1
3
3 2
for CD OD). C NMR measurements in D
1
5 min and a small amount of white solid was removed by
3
1
1
P{ H} NMR spectra are referenced to an external standard of
85% phosphoric acid solution in D O (= 0 ppm). For meas-
urements of 1–3 in D O or H O, the chemical shift of the
vacuum filtration. The filtrate was dried under reduced pres-
sure, and the ensuing solid was further dried for 16 h to give 3
(92 mg, 97%) as a red-orange solid. Anal. Calcd. for
C H45.8Co N8.2Na2.2O P : C, 30.74; H, 4.55; N, 11.31%.
26 2 19.5 2
Found: C, 30.88; H, 4.40; N, 11.52%. ICP-OES: Co:P =
2
2
2
solvent signal was set to 0 ppm to simplify comparison be-
1
tween H NMR spectra and the corresponding CEST spectra
(
Z spectra). All coupling constants (J) are reported in Hertz
(Hz). The MestReNova 10.0 NMR data processing software
was used to analyze and process all recorded NMR spectra. T
relaxation times of H O were measured after detuning the
Agilent DD2 400 MHz instrument to 392 MHz to account for
radiation damping, and obtained by fitting H O signal intensi-
1.01:1.00. UV-Vis absorption spectrum (87 M; 50 mM
−
1
HEPES buffered at pH 7.4, 25 °C): 379 nm (= 12300 M
−
1
1
cm ). ESI-MS (m/z): Calcd. for
C
26
H44Co
2
N
7
O
14
P
2
+
2
(
(
[L′Co
2
(etidronate)+2H] ): 858.11, found 858.11. FT-IR
ATR, cm ): 3300 (w, broad); 2930 (w); 1612 (s); 1502 (w);
−1
2
1
8
408 (w); 1297 (s); 1169 (m); 1124 (m); 1061 (s); 896 (m);
09 (w); 751 (w); 688 (w); 645 (w); 542 (s); 462 (s).
ties from experiments with an array of relaxation times im-
plemented in the program vnmr.
X-ray Structure Determination. Single crystals of
by 1H NMR Analysis. The pH-
Estimation of pK
a
Na[LCo
Na[LCo
2
(etidronate)]∙6.8H
2
O
(1′)
and
1
dependent H NMR chemical shifts of the CH
3
resonance from
2
(CMDP)]∙8.2H O (2′) were directly coated with
2
etidronate for compounds 1 and 3, and the CH resonance from
Paratone-N oil, mounted on a MicroMounts rod, and frozen
under a stream of dinitrogen during data collection. The crys-
tallographic data were collected at 100 K on a Bruker Kappa
Apex II diffractometer equipped with an APEX-II detector
and MoKα sealed tube source. Raw data were integrated and
corrected for Lorentz and polarization effects with Bruker
4−
a
CMDP for 2 were used to estimate the pK values of com-
1
pounds 1–3. The change in H NMR chemical shift for these
resonances as a function of pH was fitted to a Boltzmann
1
7
sigmoidal function to model a single ionization event accord-
ing to the following equation:
= A
2
+ (A
1
− A
2
)/(1 + exp((pH – pK
a
)/dx))
(1)
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 13
In this equation,  is the obtained chemical shift, A
theoretical chemical shift of the fully deprotonated species, A
is the theoretical chemical shift of the fully protonated species,
pK is the inflection point of the graph, and dx is a parameter
2
is the
In this equation, is the frequency difference (Hz) between
the tert-butyl resonance of tert-butanol in the sample and
reference solutions, M is the molecular mass of the paramag-
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
w
−
1
a
netic compound (g mol ),  is the operating frequency of the
0
describing the steepness of the curve.
NMR spectrometer (Hz), m is the concentration of the para-
−
3
dia
CEST Experiments. Variable-temperature CEST experi-
ments were carried out on an Agilent DD2 400 MHz (9.4 T)
spectrometer. In a typical CEST experiment, 6–15 mM sam-
ples of 1–3 in either an aqueous buffer solution containing 50
magnetic compound (g cm ), and 
M
is the diamagnetic
3
−1
contribution to the molar susceptibility (cm mol ).
UV-Visible Absorption Spectroscopy. Solution and sol-
id-state UV-Visible spectra were collected in the 200–800 nm
range on an Agilent Cary 5000 UV-Vis-NIR spectrometer
equipped with an integrating sphere for diffuse reflectance
measurements. Solution spectra were collected on 64–87 M
samples of compounds 1–3 in aqueous buffer solutions con-
taining 50 mM HEPES and 100 mM NaCl in the pH range
used for CEST experiments. Diffuse reflectance spectra were
collected on crystalline samples of 1′ and 2′. Samples were
prepared by grinding single crystals of the compounds, fol-
mM HEPES and 100 mM NaCl or FBS at desired pH values
1
(
measured with a pH electrode immediately before H NMR
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
and CEST data collection) were measured. Z-spectra (CEST
1
spectra) were obtained according to the following protocol: H
NMR spectra were acquired from −50 to 130 ppm with a step
increase of 1 ppm using a presaturation pulse applied for 6 s at
a power level (B
lary within the NMR sample tube to lock the sample. The
normalized integrations of the H O signal from the obtained
spectra were plotted against frequency offset to generate a Z-
spectrum. Generally, direct saturation of the H O signal was
1
) of 24 T. D
2
O was placed in an inner capil-
2
4
lowed by mixing with BaSO powder for a two-fold dilution.
Electrochemical Measurements. Cyclic voltammetry
measurements were carried out in a standard one-compartment
cell under dinitrogen using CH Instruments 760c potentiostat.
The cell consisted of a glassy carbon as a working electrode, a
platinum wire as a counter electrode, and a saturated calomel
electrode (SCE) as a reference electrode. Analytes were meas-
ured in aqueous solutions with 100 mM NaCl and 50 mM
HEPES buffered at pH 7.4. All potentials were converted and
referenced to the normal hydrogen electrode (NHE), using a
2
set to 0 ppm, but ±1 ppm shift was observed for several sam-
ples.
Exchange rate constants (kex) were calculated following a
18
2
previously reported method, where the x-intercept (−1/kex
was obtained from a plot of M /(M − M ) (M and M are the
magnetization of the on- and off-resonance, respectively)
)
z
0
z
z
0
2
−1
1
against 1/
various presaturation power levels ranging from 10 to 24 T
applied for 6 s at 37 °C. The B values were calculated based
1
(
1
in rad s ). H NMR spectra were acquired at
23
literature conversion factor.
1
Other Physical Measurements. Preparative reverse-
on the calibrated 90° pulse on a linear amplifier. To correct for
baseline variations, a linear baseline was drawn directly be-
tween the first data point (129–131 ppm) and the data point at
45 ppm frequency offset. Note that due to poor baseline for
the pH 6.62 sample for 2, a linear baseline correction was
applied for each CEST peak by using the data points at 129
and 85 ppm, and at 85 and 45 ppm, respectively. Reported
2
phase HPLC was performed on a Waters 19 × 250 mm
XBridge C18 column, using a Varian Prostar 500 system
equipped with a Varian 363 fluorescence detector and a Vari-
an 335 UV-Vis detector. During HPLC experiments, H O was
2
used as solvent A and MeCN as solvent B. The absorbances at
220 and 285 nm were monitored. The electrode-based pH
measurements were carried out using a Thermo Scientific
Orion 9110DJWP double junction pH electrode connected to a
VWR sympHony B10P pH meter. The pH meter was calibrat-
ed using standardized pH buffer solutions at 4.01, 7.00, and
values of %CEST ((1 − M
z 0
/M ) × 100%) are the differences in
%H
2
O signal reduction between applied on-resonance presatu-
rations (raw data) and the values obtained by inserting the
corresponding frequencies into the linear baseline equations.
To calculate kex, the CEST intensities at the frequency offsets
1
0.00 purchased from LaMotte Company. Elemental analysis
was conducted by Midwest Microlab Inc. Infrared spectra
were recorded for solid samples of 1–3 on a Bruker Alpha
FTIR spectrometer equipped with an attenuated total reflec-
tance accessory. These data are provided in Figure S1. Elec-
trospray ionization mass spectrometry (ESI-MS) measure-
ments were performed on a LC-MS Bruker AmaZon X quad-
rupole ion trap instrument, equipped with a Compass software
version 1.4. All measurements were carried out in MeOH
carrier solvent using positive and/or negative ionization mode.
Inductively-coupled plasma optical emission spectroscopy
(ICP-OES) was performed on a Thermo iCAP 7600 dual view
ICP-OES instrument equipped with a CETAC ASX520 240-
position autosampler. Samples were dissolved in a 3% aque-
ous nitric acid solution and the emissions for Co and P com-
pared to standard solutions.
corresponding to maximum H O signal reductions at 24 T
2
power level were monitored for each pH value. The pH cali-
bration curves were generated by taking the base 10 logarithm
0 z
of the ratios of two CEST signal intensities (reported as M /M
9
c–e,19,20
−
1)
Solution Magnetic Measurements. The solution magnet-
ic moments of compounds 1–3 were determined using the
after a baseline correction was applied.
21
1
Evans method, by collecting variable-pH H NMR spectra at
7 °C (310 K) on an Agilent DD2 500 MHz (11.7 T) spec-
3
trometer. In a typical experiment, the compound (3–7 mM)
was dissolved in a mixture of 2% (v/v) tert-butanol in an
aqueous solution containing 50 mM HEPES and 100 mM
NaCl buffered at a specific pH value. The resulting solution
was placed in an NMR tube containing a sealed capillary with
the same solvent mixture but without the to-be-characterized
paramagnetic compound as a reference solution. Diamagnetic
corrections were carried out based on the empirical formula of
each compound (as determined by elemental analysis) using
Cell Viability Measurement. Melanoma B16F10 cells
were purchased from American Type Culture Collection and
cultured in Dulbecco’s Modified Eagle’s Media (Life Tech-
nologies) with 10% (v/v) FBS (Fisher Scientific), 1 mM sodi-
um pyruvate (Sigma Aldrich), 0.1 mM non-essential amino
acids (Sigma Aldrich), and 4 mM of L-glutamate. Cells were
grown in a humidified incubator operating at 37 °C and 5.0%
2
2
Pascal’s constants. The paramagnetic molar susceptibility
para
3
−1

M
(cm mol ) was calculated using the following equa-
2
1
tion:
para
dia

M
= (3M
w
)/(4π m − 
M
(2)
2
CO , and harvested by incubation with 0.25% TrypLE for 5
0
4
ACS Paragon Plus Environment

Page 5 of 13
Journal of the American Chemical Society
min at 37 °C in a 5.0% CO
ment were sub-cultivated twice after thawing the cell stocks.
2
incubator. Cells for the experi-
tion to produce pH maps of the samples. Gating images were
produced to remove noise from the remainder of the images,
as well as signals from the surrounding doped H O. These
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
B16F10 cells were seeded at a density of 25,000 cells per well
in a 24-well plate and allowed to grow for 24 h before incuba-
tion. Cells were incubated with media containing concentra-
tions of 1 ranging from 0.2–11.3 mM (300 L, 7 concentra-
tions) for 24 h before viability measurements were carried out.
The stock solution of 1 was filtered with a 0.2 m sterile filter
prior to incubation with the cells. Cell viability was measured
using a Guava easyCyte HT flow cytometer equipped with a
2
were generated from binary gating images acquired with no
saturation pulses, and dedicated images suppressing signals
from only the capillaries. The difference between these pro-
duces binary gating images of just the capillary tubes. An
image erosion routine was used to shrink these images to the
central region of the capillary tubes (75.4% of the total cross-
sectional area), to remove unwanted partial volume and sus-
ceptibility effects. These central regions were used for CEST
data analysis and are shown in Figure 6. Values of %CEST are
9
6-well plate/10 tube autosampler (EMD Millipore). Each
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
sample subjected for analysis contained 50 L of a well-mixed
cell suspension in phosphate buffered saline (PBS) and 150
z 0
reported as %CEST = (1 − M /M ) × 100%. Averaged intensi-
ties of the regions shown in Figure 6 (upper) were employed
to calculate the CEST104 ppm/CEST64 ppm ratios and the corre-
sponding log10(CEST104 ppm/CEST64 ppm) values used to gener-
ate the pH calibration curve between pH 6.58 and 7.54. Note
that for both the CEST intensity ratios and the pH calibration

L of Guava ViaCount reagent. Samples were transferred to a
6-well plate and immediately counted using the Guava
9
ViaCount software module. Viability was measured using the
EasyFit Analysis feature. Attempts to use cells not treated
with 1 as a control to estimate normal cell death remained
unsuccessful. For that reason, cell viability (in %) is reported
without taking normal cell death into account. Therefore, the
reported viability corresponds to the lower limit of cell surviv-
al at each concentration of 1. Note, however, that we observed
a considerable number of dead cells in the control solution not
treated with 1, even though this number could not be accurate-
ly quantified.
9
c–
curve, the CEST signal intensities are reported as M
0
/M
in analogy to the data obtained from NMR measure-
ments.
z
– 1,
e,19,20
RESULTS AND DISCUSSION
Syntheses and Structures. The nitro-substituted tet-
ra(carboxamide) chelating ligand HL was selected as a CEST-
active ligand, and its permethylated analogue HL′ was selected
as a CEST-inactive counterpart. These ligands were synthe-
MRI Phantom Experiments. Samples for phantom ex-
periments contained 17 mM of 1 in aqueous solutions with 50
mM of HEPES and 100 mM of NaCl buffered at selected pH
values ranging from 6.40 to 7.88. All samples were filtered
through a 0.22 m nylon membrane, transferred to borosilicate
glass capillaries (0.2 mm thickness, 1.5–1.8 mm outer diame-
ter), and flame sealed. A bundle of 10 capillaries, each con-
taining a solution of 1 buffered at a specific pH, was placed
within an NMR tube (18 mm outer diameter) filled with an
aqueous solution containing 1 mM gadodiamide (Omniscan)
N
sized through S 2 reactions between 2,2′-iminobis(acetamide)
derivatives and 2,6-bis-(bromomethyl)-4-nitrophenol (see
experimental details and Schemes S1 and S2 in the Supporting
Information). Reaction of the ligands with two equivalents of
Co(NO
methanediphosphonic acid (H
ence of five equivalents of Na(OMe), afforded compounds
3
)
2
·6H
2
O and one equivalent of etidronic or chloro-
4
CMDP) in MeOH, in the pres-
Na[LCo
Na[LCo
Na[L′Co
2
(etidronate)]∙0.2NaNO
(CMDP)]∙4.5H O∙MeOH
2
(etidronate)]∙1.2NaNO ∙1.9H O (3) as orange solids
3
∙2.7H
2
O
(1),
and
2
2
(2),
1
for T matching. CEST experiments were carried out on an 89
2
3
mm vertical bore Bruker Avance III HD 750 MHz (17.6 T)
MRI scanner running ParaVision 6.0.1 (Bruker Biospin,
Billerica, MA, USA). Temperature was maintained at 37 °C
using heated water flowing through the gradient coils. The
probe and samples were allowed to equilibrate at this tempera-
ture for 1 h before acquisition. CEST images were acquired
using a standard spin echo imaging sequence with presatura-
tion pulses (74.3 T, 570 ms total duration) consisting of a
train of 1250 Gaussian pulses, each of 0.44 ms (6.2 kHz
(see Experimental Section and Figure 1). The ancillary ligand
etidronate was selected based on the potential for the hydroxyl
4
−
group to exhibit the CEST effect. The related ligand CMDP
was prepared to serve as an analogous ancillary ligand with no
exchangeable protons, as the two bisphosphonates feature
Table 1. Selected Mean Interatomic Distances (Å) and Angles
(°) in 1′ and 2′ at 100 K
z
bandwidth), applied at 64 and 104 ppm frequency offsets (M ),
respectively. Other imaging acquisition parameters were as
follows: field of view (FOV) = 15 × 15 mm ; matrix = 256 ×
1′
2′
2
Co–Ophenoxo
Co–Oamide
Co–Ophosphonate
Co–N
2.0920(2)
2.1127(1)
2.0618(1)
2.1558(2)
3.6740(3)
122.837(4)
113.882(3)
111.486(6)
170.324(1)
62.27(2)
2.0913(2)
2.1044(1)
2.0684(1)
2.1578(2)
3.6780(3)
123.136(5)
113.993(4)
112.901(6)
170.451(1)
62.83(2)
2
56; repetition time (TR) = 2000 ms; echo time (TE) = 3.75
ms; flip angle = 210°; slice thickness = 2 mm; averages = 3.
Reference unsaturated images were acquired at 0 ppm fre-
quency offset (M
amplitude was set to 0 T. Due to a slight difference in ob-
served chemical shift (ca. 1 ppm) between the H O signals in
the capillaries and the surrounding solution, 0 ppm was de-
fined as the H O signal in the capillary tubes. To reduce chem-
0
) using identical parameters except the pulse
Co···Co
Co–Ophenoxo–Co
O–P–O
2
P–C–P
trans O–Co–Ea
2
ical shift artifacts in the images, 1.2 kHz bandwidth excitation
and refocusing pulses were used in the spin echo sequence. A
sine smoothing filter was applied to the raw k-space data to
remove Gibbs ringing artifacts in the images.
b



sum
mean

5.19(1)
5.24(1)
c

49.120(4)
49.260(4)
All images were produced in MATLAB R2016b version
a
5
E denotes either a N or an O atom from the [CoNO ] coordination
b
9
.1.0 (The MathWorks Inc., Natick, MA, USA). Custom
sphere. Octahedral distortion parameter  = absolute deviation from
90° of each 12 cis angle in [CoNO
O
c
scripts were written in MATLAB to calculate the CEST imag-
es, CEST104 ppm/CEST64 ppm ratios, and to apply the pH calibra-
5
]. Dihedral angle between the Co–
−
phenoxo–Co plane and the plane of the phenolate ring of L .
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 13
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
−
−
Figure 2. Crystal structures of the anionic complexes [LCo
2
(etidronate)] (left) and [LCo
2
(CMDP)] (right), as observed in 1′ and 2′,
respectively. Purple, green, magenta, red, blue, and gray spheres represent Co, Cl, P, O, N, and C atoms, respectively; H atoms are
omitted for clarity.
similar steric and electronic properties.
were collected for samples in aqueous buffer solutions con-
taining 50 mM HEPES and 100 mM NaCl. The spectrum for a
sample of 1 buffered at pH 7.4 features a strong absorption at
75 nm (ε = 13800 M cm ) (see Figure S2). Similarly, a
solution of 2 shows a nearly identical absorption band at this
wavelength (ε = 14100 M cm ) under the same conditions
see Figure S3). Based on these observations and literature
2
precedent for similar phenoxo-bridged Co complexes,
we assign these absorptions to ligand–metal charge transfer
Slow diffusion of MeCN vapor into a concentrated solu-
tion of 1 and 2 in H
2
O afforded plate- and block-shaped crys-
Na[LCo (etidronate)]∙6.8H (1′) and
(CMDP)]∙8.2H O (2′), respectively. Single-crystal X-
−1
−1
3
tals
of
2
2
O
Na[LCo
2
2
−1
−1
ray diffraction analysis at 100 K revealed that 1′ and 2′ are
(
isostructural and crystallize in the monoclinic space group
24c,25a
+
P2
1
/n, with one anionic Co
2
complex and one Na ion consti-
tuting the asymmetric unit (see Table S1). In each complex,
the two nearly identical Co centers reside in distorted octahe-
dral coordination environments, each comprised of a -
phenoxo oxygen atom, a tertiary amine nitrogen atom, and two
II
(LMCT) transitions from the phenolate to Co . The close
similarity of the spectrum for 3 in pH 7.4 buffer, which exhib-
−1
−1
its a single intense band at 379 nm (ε = 12300 M cm ),
further supports the assignments of these spectral features (see
Figures S4 and S5). Notably, both the positions and intensities
of the absorption bands are relatively unaffected by pH be-
tween 5.8 and 8.3 (see Figures S2–S4). Furthermore, the dif-
fuse reflectance spectra collected for crystalline solid-state
samples of 1′ and 2′ feature peaks with maxima at 379 and
76 nm, respectively (see Figures S6 and S7). These data
indicate that the structures of [LCo
LCo (CMDP)] determined from X-ray diffraction analysis
are preserved in aqueous HEPES solutions in the physiologi-
cal pH range.
−
carboxamide oxygen atoms from L . The remaining two coor-
dination sites are occupied by oxygen atoms from the bridging
2
4
bisphosphonate, which coordinates the metal ions in a  –
binding mode (see Figure 2). The hexagonal plane of the
−
aromatic ring of L and the trigonal plane defined by the two
Co centers and the -phenoxo oxygen atom are twisted rela-
tive to another, with dihedral angles of 49.120(4) and
49.260(4)° for 1′ and 2′, respectively (see Table 1).
3
−
2
(etidronate)] and
−
[
2
The mean Co–O bond distances range from 2.0618(1) to
2.1127(1) Å in 1′, and from 2.0684(1) to 2.1044(1) Å in 2′. In
comparison, the slightly longer mean Co–N bond lengths of
2.1558(2) and 2.1578(2) Å for 1′ and 2′, respectively, reflect
weaker coordination of the tertiary amines to the metal centers
due to steric conflicts. These mean bond distances, in conjunc-
tion with the average deviations from 90° observed in the
Solution Magnetic Properties. To assess the magnetic
2
behavior of the three Co complexes, dc magnetic susceptibil-
ity data were obtained at 37 °C for aqueous buffer solutions in
the pH range 5.8–8.4 using the Evans method (see Experi-
21
mental Section). The resulting plots of  T vs pH are shown
in Figures S8–S10. For all compounds,  T varies insignifi-
cantly with pH, affording average values of  T = 6.3(3),
.0(2), and 6.1(2) cm K mol for 1, 2, and 3, respectively
(see Table S2). The mean magnetic moments per Co site
correspond to g values ranging from 2.5(1) to 2.6(1), indica-
tive of a significant contribution from orbital angular momen-
tum to the magnetic moments of 1–3. These data are in accord
with the high magnetic anisotropy of octahedral, S = /
centers,
similar high-spin Co
properties of 1–3 are nearly identical in aqueous solution
within the physiologically relevant pH range at 37 °C.
M
bond angles for the 12 cis angles in the [CoNO
5
] coordination
M
sphere of 5.19(1) and 5.24(1)° for 1′ and 2′, respectively, are
II
24,25
M
consistent with a high-spin Co electronic configuration.
Furthermore, the intramolecular Co···Co distance and Co–
phenoxo–Co angle of 3.6740(3) Å and 122.837(4)° for 1′ and
.6780(3) Å and 123.136(5)° for 2′ are consistent with related
3
−1
6
II
O
3
II
24,25
phenoxo-bridged Co
2
complexes.
Finally, the similar O–
4−
P–O and P–C–P bond angles for the etidronate and CMDP
ancillary ligands in 1′ and 2′, respectively, verify the insignifi-
cant structural changes associated with altering the ancillary
bisphosphonate. Taken together, these comparable structural
metrics for 1′ and 2′ provide validation for the use of CMDP
as a CEST-inactive analogue of etidronate.
3
II
2
Co
1
2a
and agree with values reported for structurally
II
24
2
complexes. In sum, the magnetic
4
−
NMR Spectroscopy. To further examine and compare the
UV-Vis Spectroscopy. To probe the electronic structure
of compounds 1–3 in solution, UV-Visible absorption spectra
1
solution properties of the Co
2
complexes, H NMR spectra
6
ACS Paragon Plus Environment

Page 7 of 13
Journal of the American Chemical Society
were collected for aqueous solutions of 1–3 buffered at select-
ed pH values. All compounds gave sharp, well-resolved NMR
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
II
spectra, consistent with high-spin Co ions in pseudo-
1
2
octahedral geometry. The spectrum for 1 at pH 7.18 features
22 paramagnetically shifted resonances that range in chemical
shift from −110 to 185 ppm vs H O (see Figure S11, upper).
2
The resonances at 9.5, 13, 64, 68, 102, 104, and 105 ppm are
assigned to exchangeable protons on the carboxamide groups
and the etidronate hydroxyl group, as evidenced by their dis-
2
appearance in the analogous spectrum recorded in neutral D O
(see Figure S11, lower). The anticipated two additional amide
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
resonances are most likely concealed by the broad H O signal.
The appearance of the 22 NMR signals as 10 pairs of closely
spaced peaks is consistent with the pseudo-C symmetry of
(etidronate)] in 1, where a slight lowering from C
2
2
−
[
LCo
2
results from the asymmetry of the etidronate ligand. The two
remaining peaks correspond to the hydroxyl and methyl sub-
stituents on etidronate. Upon raising the pH from 6.18 to 8.14,
the exchangeable proton resonances become significantly
more broad, indicative of faster proton exchange (see Figure
S12).
Figure 3. CEST spectra collected at 37 °C for 12.8 mM aqueous
solutions of 1 with 50 mM HEPES and 100 mM NaCl buffered at
pH 6.50–8.14 (red to blue). The legend gives the pH and corre-
sponding color of each sample. Inset: Expanded view of the CEST
peaks of interest.
In comparison, the spectrum for 2 at pH 7.18 exhibits 23
paramagnetically shifted peaks in the range −105–180 ppm vs
mM of 1 with 50 mM HEPES and 100 mM NaCl buffered at
pH values ranging from 6.50 to 8.14 (see Figure 3). The spec-
trum at pH 6.50 exhibits two peaks, centered at 66 and 102
H
2
O, with exchangeable carboxamide signals at 4.7, 6.8, 9.5,
11, 68, 70, 102, and 104 ppm, verified by comparison of the
spectra recorded in H O and D O (see Figure S13). Notably,
2
2
ppm, with 2.0 and 14% H
Note that CEST signals from the labile protons below 13 ppm
are masked by direct saturation of the H O solvent. As the pH
2
O signal reduction, respectively.
the presence of four highly-shifted amide resonances that are
well separated from the four remaining amide peaks confirms
the inequivalency of the two amide NH protons due to re-
2
2
6
is raised to 8.14, the CEST peak at 66 ppm shifts to 64–65
ppm, and the intensity increases monotonically to 27%. This
increase in CEST intensity with pH is consistent with base-
catalyzed proton exchange that is typical for carbox-
stricted C–N bond rotation. This inequivalence is a common
observation for amide-appended transition metal complex-
9
f,11a–fhj,26ac
es.
The close similarity between the spectra for 1
and 2 (see Figure S14) suggests that these compounds are
structurally similar in solution, as observed in the solid-state.
The replacement of the intense peak at 66 ppm in the spectrum
for 1 with a peak at 138 ppm in the spectrum for 2 indicates
9
af,11ad–fh,20
amides.
In stark contrast, the CEST effect of the downfield-shifted
peak shows a very different pH profile. First, the frequency
corresponding to maximum CEST intensity is markedly af-
fected by pH variations and shifts from 102 to 106 ppm in the
pH range 6.50–8.14. Surprisingly, the CEST intensity remains
relatively constant between pH 6.50 and 7.60, but then under-
goes a significant increase when the pH is raised further. The
dramatically different pH dependences of the two CEST fea-
tures for 1 are evident from a plot of the CEST intensities at
3
that these signals correspond to the CH and CH resonances
4
−
from the etidronate and CMDP ancillary ligands, respective-
ly. Moreover, the linewidths of the carboxamide peaks for 2
show similar pH dependence between pH 6.62 and 8.34, as
observed for 1 (see Figure S15).
1
In analogy to the H NMR features of the Co
2
complexes
−
of L , the resonances in the spectrum for 3 at pH 7.47 span
from −110 to 190 ppm vs H O and display a similar spectral
profile (see Figure S16). The intense peaks at −9.8, −7.5, −3.7,
.6, 2.0, 22, and 26 ppm are assigned to methyl groups on L′ ,
and the CH resonance from etidronate is observed at 62 ppm.
Furthermore, comparison of the spectra recorded in pH 8.08
buffer and slightly basic D O reveals the disappearance of the
6
4 and 104 ppm vs pH (see Figure S19). We hypothesize that
2
the unusual CEST behavior at 104 ppm stems from contribu-
tions of overlapping carboxamide and hydroxyl resonances to
−
1
1
the observed CEST effect, as suggested by H NMR analysis.
3
To better understand the causes for the unusual CEST
properties of 1, analogous variable-pH CEST spectra were
collected for aqueous solutions containing 12 mM of 2 or 13
mM of 3. The spectra for 2 in the pH range 6.62–8.34 show
two peaks at 68 and 102 ppm with CEST intensities that in-
crease significantly when the pH is raised (see Figure S20),
similar to that observed for the peak at ca. 64 ppm for 1. Im-
portantly, the nearly identical pH dependences of the two
CEST effects for 2 (see Figure S21) supports the hypothesis
that the unique CEST behavior of 1 can be attributed to the
etidronate hydroxyl group. Indeed, CEST spectra for 3 ob-
tained between pH 5.80 and 8.08 confirm the PARACEST
activity of the ancillary etidronate, as a single peak that shifts
from 94 to 103 ppm is observed (see Figure S22). The nature
of this pH-induced shift in CEST frequency is discussed be-
low. In conjunction with this frequency shift, the hydroxyl
CEST signal undergoes a significant decrease in intensity with
2
peak at 103 ppm (see Figure S17). This observation indicates
that the etidronate hydroxyl group provides a well-resolved
NMR signal under basic conditions, and further corroborates
the presence of three exchangeable proton resonances for 1 in
the 102–105 ppm range. Upon lowering the pH to 5.80, the
hydroxyl peak for 3 undergoes significant line broadening,
suggesting an increase in the proton exchange rate (see Figure
S18). Importantly, inspection of the NMR linewidths of the
carboxamide peaks for 1 and 2 and the hydroxyl resonance for
3 implies opposing pH dependences of the proton exchange
rates for these two functional groups, and therefore highlights
the potential utility of 1 for ratiometric pH imaging.
CEST Properties. In order to investigate the feasibility of
employing 1 as a pH-responsive PARACEST probe, CEST
spectra were collected for aqueous solutions containing 12.8
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
increasing pH, after reaching a maximum intensity of 20% at
Page 8 of 13
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
pH 6.11 (see Figure S23). The observation of optimal CEST
under slightly acidic conditions is consistent with PARACEST
1
1f,27
agents bearing alcohol donors.
Interestingly, all previously
reported PARACEST agents with CEST-active hydroxyl
protons feature OH groups directly bonded to the metal cen-
9
bg,11f,27,28
ter.
This further demonstrates the remarkably high
CEST peak shift and intensity of the ancillary hydroxyl group
in 1 and 3.
The proton exchange rates at 37 °C were estimated by em-
1
8
ploying the Omega plot method. The rate constants (kex) for
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
the amide protons in 2 increase from 2.7(2) × 10 (68–69 ppm)
2
−1
3
and 3.5(3) × 10 (102 ppm) s at pH 6.62 to 1.0(1) × 10 (68–
6
2
−1
9 ppm) and 8.0(3) × 10 (102 ppm) s at pH 8.34 (see Fig-
ures S24–S26 and Table S3). These values are consistent with
9
f,11ad–fhj
11lm
rates reported for mono-
and dinuclear
transition
metal PARACEST agents bearing pendent carboxamides. In
A B ppm
Figure 4. Ratios of CEST intensities (CEST ppm/CEST )
contrast, the hydroxyl proton exchange in 3 is fastest at pH
from presaturation at 104 (A) and 64 (B) ppm for 12.8 mM aque-
ous buffer solutions of 1 (blue), and at 102 (A) and 68 (B) ppm
for 12 mM solutions of 2 (red) vs pH. Inset: Semilog form of the
plot for 1. Circles denote experimental data, and the black line
corresponds to a linear fit of the data.
3
−1
5
.80 (kex = 1.5(1) × 10 s ), and then decreases sharply as the
2
−1
pH is raised to 7.47 (kex = 2.5(2) × 10 s ) (see Figures S27
and S28, and Table S4). The opposite pH trends for exchange
1
rates in 2 and 3 are in accord with H NMR and CEST data,
and reflect the base- and acid-catalyzed exchange of the NH
2
and OH protons, respectively, in these Co complexes. To
cal applications. Toward this end, CEST spectra for aqueous
buffer solutions containing 6.4 and 8.5 mM of 1 were recorded
analogously to the 12.8 mM sample (see Figures S32 and
S33). The CEST effects at 64 and 104 ppm did not vary signif-
icantly with different probe concentrations (see Figures S34–
S37). This observation suggests that the spin-lattice relaxation
compare, the rate constants for the two CEST features of 1 are
2
similar to those for 2 and 3, with values of 2.1(3)–7.3(3) × 10
(
2
−1
64–66 ppm) and 2.8(2)–7.6(3) × 10 (101–106 ppm) s in the
pH range 6.18–8.14 (see Figures S29–S31 and Table S5).
Note that the pH-dependent changes of the rate constants for 1
are less obvious than those observed for 2 and 3. Most likely,
this difference results from asymmetric CEST peaks for 1 and
the contribution of both NH and OH protons to the CEST
effect at 101–106 ppm. Therefore, more elaborate methods are
needed to accurately determine the exchange rate for each
CEST effect of 1.
rate of H
concentration range,
2
O is close to the proton exchange rates within this
6
ef,9dh,30
1
consistent with T analysis (see
Table S6). Most importantly, these experiments show that
plots of CEST104 ppm/CEST64 ppm vs pH are nearly identical for
the three concentrations in the pH range 6.50–8.15 (see Fig-
ures S38–S40). Indeed, linear fits of the corresponding
log(CEST104 ppm/CEST64 ppm) values as a function of pH afford-
ed the following equations (see Figures S38 and S39, insets):
Ratiometric CEST Analysis. To assess the potential of
compound 1 to enable ratiometric pH quantitation, the pH
dependence of the ratio of CEST intensities at 104 and 64 ppm
(CEST104 ppm/CEST64 ppm) was investigated. Remarkably, the
6.4 mM: log(CEST104 ppm/CEST64 ppm) = −1.05 × pH + 7.9 (4)
data reveal a substantial decrease in the intensity ratio from a
value of 8.35 to 0.82 in the pH range 6.50–7.60, while no
significant change is observed at higher pH (see Figure 4).
Moreover, the logarithm (log) of this ratio was found to vary
linearly with pH in this range (see Figure 4, inset), according
to the following equation:
8
.5 mM: log(CEST104 ppm/CEST64 ppm) = −1.03 × pH + 7.7 (5)
The pH calibration curves obtained for various concentrations
of 1 (Equations 3–5) are summarized in Figure S41. For a
given log(CEST104 ppm/CEST64 ppm) value, the deviation in pH
was found to be ca. 0.02–0.09 pH units over the pH range
2
9
6
.50–7.60. This observation demonstrates the ability of 1 to
log(CEST104 ppm/CEST64 ppm) = −0.99 × pH + 7.4
(3)
quantitate solution pH in a concentration-independent manner
within the error of 0.1 pH unit.
Conversely, the analogous ratio of CEST intensities at 102 and
6
8 ppm (CEST102 ppm/CEST68 ppm) for 2 is relatively unaffected
by pH changes, with values of 0.57–0.98 between pH 6.62 and
.34 (see Figure 4). This comparison highlights the essential
Temperature Effects. An important challenge facing pH-
responsive MR probes is the ability to deconvolute pH re-
sponses from temperature effects of the CEST peak frequency
and intensity, owing to the temperature dependences of hyper-
8
role of the etidronate hydroxyl group to enable ratiometric
quantitation of pH in the physiological range with 1, using
Equation 3 as a calibration curve. The slope of a linear calibra-
tion curve provides a useful measure of probe sensitivity.
12
9
fine shifts and proton exchange rates. To examine how
temperature variation affects the pH calibration curve, varia-
1
ble-pH H NMR and CEST spectra were collected at the addi-
−1
Notably, the absolute value of 0.99(7) pH unit obtained for 1
tional temperatures 35 and 39 °C on 12.8 mM solutions of 1
buffered at pH 6.50–8.14. The data show very similar pH-
dependent behavior as observed at 37 °C (see Figures S42–
S45), albeit with nearly all resonances shifted by ca. 1 ppm
is ca. 2–4-fold greater than those reported for related rati-
ometric PARACEST pH probes at 37 °C, even when com-
pared to instances where the CEST intensity ratios are em-
7
ac,9c–f
ployed directly.
away and toward the H
2
O signal at 35 and 39 °C, respectively
To further evaluate the efficacy of 1 as a ratiometric pH
probe, we first sought to determine whether this pH calibration
curve is affected by the concentration of the probe, since a
concentration-independent measure is critical for physiologi-
(
see Figure S46), consistent with Curie behavior of high-spin
II 12
Co . Upon increasing the temperature from 35 to 39 °C, a
moderate increase in CEST intensities at 64 and 104 ppm was
observed (see Figures S47–S50). Importantly, temperature
8
ACS Paragon Plus Environment

Page 9 of 13
Journal of the American Chemical Society
changes do not affect the CEST104 ppm/CEST64 ppm values above
pH 7.0. In contrast, temperature changes cause significant
deviations in the pH profile of CEST104 ppm/CEST64 ppm below
pH 7.0 (see Figures S51–S53). Here, fits of the log(CEST104
ppm/CEST64 ppm) vs pH plots using data from the pH range
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
.50–7.60 gave the following linear equations (see Figures
S51 and S52, insets):
3
3
5 °C: log(CEST104 ppm/CEST64 ppm) = −1.19 × pH + 8.9 (6)
9 °C: log(CEST104 ppm/CEST64 ppm) = −0.79 × pH + 5.9 (7)
The significant effect of temperature variations on the pH
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
calibration curve is primarily due to the temperature-induced
shifts in CEST frequencies (see Figure S54). Indeed, fits of the
log(CEST105 ppm/CEST65 ppm) and log(CEST103 ppm/CEST63 ppm
)
vs pH plots for the data obtained from pH 6.50–7.60 at 35 and
9 °C, respectively, provided excellent linear correlations
following the equations:
3
1
Figure 5. H NMR chemical shift (frequency offset) of the CH
3
3
3
5 °C: log(CEST105 ppm/CEST65 ppm) = −1.06 × pH + 7.8 (8)
9 °C: log(CEST103 ppm/CEST63 ppm) = −0.95 × pH + 7.1 (9)
resonance from etidronate vs pH for aqueous buffer solutions of
1. Red circles denote experimental data, and the black line corre-
sponds to a sigmoidal fit of the data (Equation 1). Inset: Schemat-
ics of the anion from 1, highlighting the protonation state of
etidronate.
The calibration curves represented by Equations 8 and 9 close-
ly resemble that obtained at 37 °C (see Figure S55), demon-
strating that the %CEST at 105 and 65 ppm, and at 103 and 63
ppm, should be employed for pH measurements at 35 and 39
°C, respectively. One potential route to address temperature
To further investigate the biocompatibility of 1, prelimi-
nary cell viability experiments were carried out using mela-
noma B16F10 cells as a model. The study revealed that >50%
of the cells are viable after incubation with millimolar concen-
trations of 1 for 24 h (see Figure S64). Note that the %viability
values are reported without taking normal cell death into ac-
count, which can be appreciable, and thus only correspond to
the lower limits of cell survivals at given probe concentrations
heterogeneity in physiological environments with this Co
2
probe could involve constructing multiple pH calibration
curves, one at each temperature, and then determine the sur-
1
rounding temperature independently by exploiting the
H
NMR chemical shift of a resonance that shifts insignificantly
with pH. Such simultaneous quantitation of pH and tempera-
8,9bgh
ture using PARACEST probes has been reported.
(see Experimental Section).
Complex Stability and Biocompatibility Studies. The
cyclic voltammogram collected for an aqueous solution of 1 in
HEPES buffer at pH 7.4 exhibits an irreversible oxidation
process at ca. 560 mV vs NHE (see Figure S56). We assign
NMR Studies of pH-Induced Structural Changes. In
addition to changes in CEST peak intensities with pH, varia-
tions in the frequency of CEST peaks may also be employed
7
ac
for ratiometric pH sensing. Such CEST frequency changes
are typically caused by a pH-dependent interconversion be-
II
II
III
2
this event to the Co /Co Co oxidation, which verifies that 1
31
7
is inert towards reaction with oxygen in solution.
tween species of different protonation states. Indeed, the
In order to further assess the stability of 1 under physio-
CEST peaks for 1–3 show slight shifts with pH, which sug-
gests modest structural changes in solution. To gain further
logical conditions, 10 mM aqueous solutions of the Co
2
com-
plex buffered at pH 7.3 were incubated with 10 mM solutions
insight into potential pH-induced structural changes in the Co
2
−
2−
2−
2−
−
2+
1
of the ions H
2
PO
4
/HPO
4
, CO
3
, SO
4
, CH
3
COO , and Ca
complexes, H NMR spectra were collected for samples of 1–3
in aqueous buffer solutions over a broad pH range. The car-
boxamide peaks for 1 show moderate changes in chemical
shifts between pH 2.69 and 8.87 (see Figure S65), while the
1
for 16 h at 25 °C. The H NMR spectra of these solutions
collected at 37 °C appear identical to the spectrum obtained
previously at the same pH, albeit showing the additional ions
(see Figures S57 and S58). Furthermore, compound 1 exhibits
analogous NMR and CEST properties in fetal bovine serum
(FBS) as in HEPES buffer in the pH range 6.6–7.6 (see Fig-
CH
ppm, following a sigmoidal pH profile.
A fit of the CH chemical shift vs pH data to Equation 1
gave a pK value of 5.01(3) (see Figure 5). Similarly, the eti-
dronate CH resonance for 3 shifts from 44.00 to 62.68 ppm in
the pH range 1.56–8.82, and a corresponding sigmoidal fit of
the data afforded a value of pK = 5.28(5) (see Figure S66). In
3
resonance from etidronate shifts dramatically, by 19.55
3
2
ures S59–S62). The observation of a slightly broader H O
a
resonance in FBS compared to buffer is presumably due to
contributions from labile protons of proteins in the serum.
Importantly, the highly-shifted CEST peaks for 1 are unaffect-
ed by this broadness near the diamagnetic region, and the pH
calibration curves obtained in FBS and buffer are essentially
identical (see Figure S63). It is important to note that the addi-
tional feature at ca. 88 ppm in the CEST spectra does not
impact the CEST analysis of 1. The exact nature of this feature
is currently unknown but likely stems from a miniscule
amount of an OH-containing impurity, as it is most prominent
at acidic pH and no signals are observed in this regime in the
3
a
addition, the changes in resonance frequencies of the carbox-
amides for 2 resemble those for 1, albeit less pronounced (see
Figure S67). Comparably, a fit of the chemical shift vs pH
4
−
data for the CH resonance from CMDP to Equation 1 yield-
ed a pK of 4.40(2) (see Figure S68). These dramatic pH-
dependent chemical shift changes of the CH and CH reso-
a
3
nances from the bisphosphonates strongly suggest that the
ancillary ligands become protonated at low pH. The similar
trends observed for all complexes and the excellent agreement
of the data to a model for a single ionization event, together
1
corresponding H NMR spectra. Taken together, these results
demonstrate the high stability of 1 in physiological environ-
ments and suggest its potential for in vivo studies.
32
with the pKa values of the free bisphosphonic acids, are most
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 13
consistent with protonation/deprotonation of one of the cobalt- presaturation power was required to saturate the labile protons
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
coordinated P–O oxygen atoms as the source of peak shifts in
this pH range (see Figure 5, inset). Indeed, protonated phos-
phate oxygen donors have been observed in the solid-state in
of 1 owing to the larger Zeeman splitting on the 17.6 T MRI
scanner. The %CEST at 64 ppm increased from 1.1 to 24%
upon moving from pH 6.40 to 7.88, while presaturation at 104
ppm afforded values of 5.4 to 15% within this range (see
Figures S69 and 6, upper). These pH-dependent trends in
CEST intensity are consistent with those observed in the NMR
study. Moreover, the ratio of CEST intensities at 104 and 64
ppm (CEST104 ppm/CEST64 ppm) decreased substantially from pH
3
3
transition metal bisphosphonate complexes.
The observation of a considerably lower pK
in accord with the insignificant variation in CEST frequencies
a
for 2 than 1 is
4
−
of the amide peaks for 2, as CMDP is nearly completely
deprotonated above pH 6.5. Furthermore, the value of pK
.28(5) for 3 is in line with the observed pH dependence of the
a
=
6
.58 to 7.54 (see Figures S70 and 6, lower left), and a plot of
5
log of the ratios between averaged phantom intensities at these
frequencies (log(CEST104 ppm/CEST64 ppm)) vs pH gave an
excellent linear fit in analogy to Equations 3–9 (see Figure
S70, inset). Using this calibration curve and the CEST104
ppm/CEST64 ppm values per pixel, a pH map was generated (see
Figure 6, lower right). This result highlights that the pH-
dependent changes in CEST intensity ratios can be clearly
visualized by MRI. Furthermore, the pH values calculated
from the calibration curve are in good agreement with those
independently measured by a pH electrode (see Table S7). In
sum, phantom imaging experiments further demonstrate the
ability of 1 to ratiometrically quantitate solution pH in the
physiological pH range 6.5–7.6. Future efforts will be geared
toward improving the homogeneity and overall quality of
CEST images through pulse sequence optimization, as well as
to investigate the feasibility of pH imaging with 1 on lower
hydroxyl CEST frequency. This behavior stems from transi-
tioning from a state with considerable contributions from both
protonation states of etidronate at pH 5.8, to a state with near
exclusively the fully deprotonated ligand above pH 7.1. Final-
ly, these NMR studies establish the integrity of 1–3 in aqueous
solutions over a wide pH range.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MR Phantom Imaging. To further examine the practicali-
ty of 1 for ratiometric pH imaging through PARACEST,
CEST images of phantoms containing a series of 17 mM
aqueous solutions of 1 buffered at selected pH values from
6
.40 to 7.88 were collected on a 17.6 T MRI scanner. For each
pH value, two images were acquired at 37 °C after irradiation
at 64 and 104 ppm vs H
2
O, respectively, using 74.3 T presat-
uration pulses. Corresponding control images were collected
at 0 ppm frequency offset with 0 T power. Note that the high
Figure 6. CEST images of phantoms containing 17 mM aqueous buffer solutions of 1 in the pH range 6.40–7.88, collected at 37 °C on a
7.6 T MRI scanner. Upper: Images constructed from CEST effect upon presaturation at 64 ppm (left) and 104 ppm (right), respectively.
Lower: Ratiometric CEST104 ppm/CEST64 ppm map obtained by taking the pixel-wise ratios of CEST signal intensities at 104 and 64 ppm
left), and a pixel-wise pH map calculated from the corresponding log(CEST104 ppm/CEST64 ppm) values by using the calibration curve dis-
1
(
played in Figure S70, obtained from averaged phantom intensities at 64 and 104 ppm between pH 6.58 and 7.54 (right). White numbers
next to each phantom sample denote the pH of the corresponding solution measured by a pH electrode.
1
0
ACS Paragon Plus Environment

Page 11 of 13
Journal of the American Chemical Society
helpful discussions, and Mr. G. Yamankurt and Mr. S. M.
field MRI scanners. Eventually, the actual potential of the
II
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Co
probe for ratiometric mapping of pH will be evaluated in
Tatro for experimental assistance.
2
small animal imaging studies.
REFERENCES
CONCLUSIONS AND OUTLOOK
2
The foregoing results demonstrate the ability of Co complex-
es to provide a concentration-independent measure of solution
pH over a range relevant for detecting physiological abnor-
malities through ratiometric PARACEST imaging. In particu-
lar, the systematic study of 1–3 illustrates the opposing pH-
dependent CEST properties of carboxamide NH and etidronate
OH protons. The potential of Co
pH probes is further highlighted by the stability of 1 in physio-
logical environments and good agreement between pH from
phantom images of 1 and those measured by an electrode.
Considering the excellent tunability of the phenoxo-bridged
dinuclear platform, ongoing work is focused on investigating
the CEST behavior of related ancillary bisphosphonate lig-
ands, and on incorporating other CEST-active functional
groups on the dinucleating ligand scaffold, in efforts to opti-
mize the pH-dependent CEST properties of this family of
molecules for imaging pH in vivo. We anticipate that this
broadly generalizable platform will aid in developing pH-
responsive probes with higher sensitivity and stability, in
particular those suitable for in vivo applications.
(1) (a) Kallinowski, F.; Schlenger, K. H.; Runkel, S.; Kloes,
M.; Stohrer, M.; Okunieff, P.; Vaupel, P. Cancer Res. 1989,
49, 3759. (b) Tannock, I. F.; Rotin, D. Cancer Res. 1989,
49, 4373. (c) Vaupel, P.; Kallinowski, F.; Okunieff, P. Can-
cer Res. 1989, 49, 6449. (d) Gillies, R. J.; Raghunand, N.;
Karczmar, G. S.; Bhujwalla, Z. M. J. Magn. Reson. Imaging
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
II
2
complexes as ratiometric
2
002, 16, 430. (e) Gatenby, R. A.; Gillies, R. J. Nat. Rev.
Cancer 2004, 4, 891. (f) Gillies, R. J.; Raghunand, N.; Gar-
cia-Martin, M. L.; Gatenby, R. A. IEEE Eng. Med. Biol.
Mag. 2004, 23, 57. (g) Hashim, A. I.; Zhang, X.; Wojtkowi-
ak, J. W.; Martinez, G. V.; Gillies, R. J. NMR Biomed.
2011, 24, 582. (h) Kato, Y.; Ozawa, S.; Miyamoto, C.;
Maehata, Y.; Suzuki, A.; Maeda, T.; Baba, Y. Cancer Cell
Int. 2013, 13, 89.
(
2) (a) Katsura, K.; Ekholm, A.; Asplund, B.; Siesjö, B. K. J.
Cereb. Blood Flow Metab. 1991, 11, 597. (b) Sun, P. Z.;
Wang, E.; Cheung, J. S. Neuroimage 2012, 60, 1. (c) Ra-
jamäki, K.; Nordström, T.; Nurmi, K.; Åkerman, K. E. O.;
Kovanen, P. T.; Öörni, K.; Eklund, K. K. J. Biol. Chem.
2013, 288, 13410.
(
3) (a) Wike-Hooley, J. L.; Haveman, J.; Reinhold, H. S. Radi-
other. Oncol. 1984, 2, 343. (b) Gerweck, L. E.; Seetha-
raman, K. Cancer Res. 1996, 56, 1194. (c) Wojtkowiak, J.
W.; Verduzco, D.; Schramm, K. J.; Gillies, R. J. Mol.
Pharm. 2011, 8, 2032. (d) Pilon-Thomas, S.; Kodumudi, K.
N.; El-Kenawi, A. E.; Russell, S.; Weber, A. M.; Luddy, K.;
Damaghi, M.; Wojtkowiak, J. W.; Mulé, J. J.; Ibrahim-
Hashim, A.; Gillies, R. J. Cancer Res. 2016, 76, 1381.
ASSOCIATED CONTENT
Supporting Information
Additional experimental details and characterization data for
1, 2, and 3, including crystallographic data for 1′ and 2′. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(4) (a) Lauffer, R. B. Chem. Rev. 1987, 87, 901. (b) Caravan,
P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B. Chem. Rev.
AUTHOR INFORMATION
1
999, 99, 2293. (c) Rieke, V.; Pauly, K. B. J. Magn. Reson.
Corresponding Author
*dharris@northwestern.edu
Imaging 2008, 27, 376.
(
(
5) Ward, K. M.; Aletras, A. H.; Balaban, R. S. J. Magn. Reson.
2000, 143, 79.
ORCID
6) Selected references: (a) Zhang, S.; Winter, P.; Wu, K.;
Sherry, A. D. J. Am. Chem. Soc. 2001, 123, 1517. (b)
Zhang, S.; Merritt, M.; Woessner, D. E.; Lenkinski, R. E.;
Sherry, A. D. Acc. Chem. Res. 2003, 36, 783. (c) Zhou, J.;
van Zijl, P. C. M. Prog. Nucl. Magn. Reson. Spectrosc.
Agnes E. Thorarinsdottir: 0000-0001-9378-4454
Kang Du: 0000-0001-9947-2320
T. David Harris: 0000-0003-4144-900X
Notes
2
006, 48, 109. (d) Woods, M.; Woessner, D. E.; Sherry, A.
The authors declare no competing financial interests.
D. Chem. Soc. Rev. 2006, 35, 500. (e) Ali, M. M.; Liu, G.;
Shah, T.; Flask, C. A.; Pagel, M. D. Acc. Chem. Res. 2009,
ACKNOWLEDGMENT
4
2, 915. (f) Viswanathan, S.; Kovacs, Z.; Green, K. N.;
Ratnakar, S. J.; Sherry, A. D. Chem. Rev. 2010, 110, 2960.
g) Terreno, E.; Castelli, D. D.; Viale, A.; Aime, S. Chem.
This research was funded by the Air Force Research Laborato-
ry under agreement no. FA8650-15-5518 and Northwestern
University. T.D.H. thanks the Alfred P. Sloan Foundation, and
A.E.T. gratefully acknowledges the Leifur Eiriksson Founda-
tion for funding support. X-ray crystallography and NMR
spectroscopy was performed at the Integrated Molecular
Structure Education and Research Center (IMSERC) at
Northwestern University, with support from the Soft and
Hybrid Nanotechnology Experimental (SHyNE) Resource
(
Rev. 2010, 110, 3019. (h) Soesbe, T. C.; Wu, Y.; Sherry, A.
D. NMR Biomed. 2013, 26, 829.
(7) (a) Wu, Y.; Soesbe, T. C.; Kiefer, G. E.; Zhao, P.; Sherry,
A. D. J. Am. Chem. Soc. 2010, 132, 14002. (b) Wang, X.;
Wu, Y.; Soesbe, T. C.; Yu, J.; Zhao, P.; Kiefer, G. E.;
Sherry, A. D. Angew. Chem., Int. Ed. 2015, 54, 8662. (c)
Wu, Y.; Zhang, S.; Soesbe, T. C.; Yu, J.; Vinogradov, E.;
Lenkinski, R. E.; Sherry, A. D. Magn. Reson. Med. 2016,
(NSF NNCI-1542205), the State of Illinois, and International
7
5, 2432. (d) Tsitovich, P. B.; Cox, J. M.; Spernyak, J. A.;
Institute for Nanotechnology (IIN). Quantitative analysis of
Co and P was performed at the Northwestern University
Quantitative Bio-element Imaging Center. We thank Prof. T.
J. Meade for generous donation of B16F10 cell supplies, fetal
bovine serum, and for use of his HPLC system, lyophilizer,
and laminar flow hood, Prof. C. A. Mirkin for use of his flow
cytometer, Dr. L. M. Lilley for experimental assistance and
Morrow, J. R. Inorg. Chem. 2016, 55, 12001.
(8) McVicar, N.; Li, A. X.; Suchý, M.; Hudson, R. H. E.;
Menon, R. S.; Bartha, R. Magn. Reson. Med. 2013, 70,
1
016.
1
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 13
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
9) (a) Aime, S.; Castelli, D. D.; Terreno, E. Angew. Chem., Int.
(18) Dixon, W. T.; Ren, J.; Lubag, A. J. M.; Ratnakar, J.;
Ed. 2002, 41, 4334. (b) Castelli, D. D.; Terreno, E.; Aime,
S. Angew. Chem., Int. Ed. 2011, 50, 1798. (c) Liu, G.; Li,
Y.; Sheth, V. R.; Pagel, M. D. Mol. Imaging 2012, 11, 47.
Vinogradov, E.; Hancu, I.; Lenkinski, R. E.; Sherry, A. D.
Magn. Reson. Med. 2010, 63, 625.
(19) (a) Chen, L. Q.; Howison, C. M.; Jeffery, J. J.; Robey, I. F.;
Kuo, P. H.; Pagel, M. D. Magn. Reson. Med. 2014, 72,
1408. (b) Moon, B. F.; Jones, K. M.; Chen, L. Q.; Liu, P.;
Randtke, E. A.; Howison, C. M.; Pagel, M. D. Contrast
Media Mol. Imaging 2015, 10, 446.
(d) Sheth, V. R.; Liu, G.; Li, Y.; Pagel, M. D. Contrast Me-
dia Mol. Imaging 2012, 7, 26. (e) Sheth, V. R.; Li, Y.;
Chen, L. Q.; Howison, C. M.; Flask, C. A.; Pagel, M. D.
Magn. Reson. Med. 2012, 67, 760. (f) Dorazio, S. J.;
Olatunde, A. O.; Spernyak, J. A.; Morrow, J. R. Chem.
Commun. 2013, 49, 10025. (g) Castelli, D. D.; Ferrauto, G.;
Cutrin, J. C.; Terreno, E.; Aime, S. Magn. Reson. Med.
2014, 71, 326. (h) Rancan, G.; Castelli, D. D.; Aime, S.
Magn. Reson. Med. 2016, 75, 329. (i) Krchová, T.; Gáli-
sová, A.; Jirák, D.; Hermann, P.; Kotek, J. Dalton Trans.
(
20) Aime, S.; Barge, A.; Castelli, D. D.; Fedeli, F.; Mortillaro,
A.; Nielsen, F. U.; Terreno, E. Magn. Reson. Med. 2002, 47,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
39.
(
21) (a) Evans, D. F. J. Chem. Soc. 1959, 2003. (b) Schubert, E.
M. J. Chem. Educ. 1992, 69, 62.
(22) Bain, G. A.; Berry, J. F. J. Chem. Educ. 2008, 85, 532.
(23) Sawyer, D. T.; Sobkowiak, A. J.; Roberts, J. Electrochemis-
try for Chemists, 2nd ed.; John Wiley & Sons: New York,
2
016, 45, 3486.
(
(
10) (a) Wolff, S. D.; Balaban, R. S. Magn. Reson. Med. 1989,
1
5
0, 135. (b) Zaiss, M.; Bachert, P. Phys. Med. Biol. 2013,
8, R221.
1
995.
(
24) Selected references: (a) Sakiyama, H.; Ito, R.; Kumagai, H.;
Inoue, K.; Sakamoto, M.; Nishida, Y.; Yamasaki, M. Eur. J.
Inorg. Chem. 2001, 2027. (b) Hossain, M. J.; Yamasaki, M.;
Mikuriya, M.; Kuribayashi, A.; Sakiyama, H. Inorg. Chem.
11) Selected references: (a) Dorazio, S. J.; Tsitovich, P. B.;
Siters, K. E.; Spernyak, J. A.; Morrow, J. R. J. Am. Chem.
Soc. 2011, 133, 14154. (b) Dorazio, S. J.; Morrow, J. R.
Eur. J. Inorg. Chem. 2012, 2006. (c) Dorazio, S. J.;
Tsitovich, P. B.; Gardina, S. A.; Morrow, J. R. J. Inorg.
Biochem. 2012, 117, 212. (d) Tsitovich, P. B.; Morrow, J.
R. Inorg. Chim. Acta 2012, 393, 3. (e) Olatunde, A. O.;
Dorazio, S. J.; Spernyak, J. A.; Morrow, J. R. J. Am. Chem.
Soc. 2012, 134, 18503. (f) Dorazio, S. J.; Morrow, J. R.
Inorg. Chem. 2012, 51, 7448. (g) Tsitovich, P. B.;
Spernyak, J. A.; Morrow, J. R. Angew. Chem., Int. Ed.
2013, 52, 13997. (h) Dorazio, S. J.; Olatunde, A. O.;
Tsitovich, P. B.; Morrow, J. R. J. Biol. Inorg. Chem. 2014,
2
002, 41, 4058. (c) Tian, J.-L.; Gu, W.; Yan, S.-P.; Liao,
D.-Z.; Jiang, Z.-H. Z. Anorg. Allg. Chem. 2008, 634, 1775.
(d) Daumann, L. J.; Comba, P.; Larrabee, J. A.; Schenk, G.;
Stranger, R.; Cavigliasso, G.; Gahan, L. R. Inorg. Chem.
2
013, 52, 2029.
(
25) (a) Johansson, F. B.; Bond, A. D.; Nielsen, U. G.; Mou-
baraki, B.; Murray, K. S.; Berry, K. J.; Larrabee, J. A.;
McKenzie, C. J. Inorg. Chem. 2008, 47, 5079. (b) Seidler-
Egdal, R. K.; Johansson, F. B.; Veltzé, S.; Skou, E. M.;
Bond, A. D.; McKenzie, C. J. Dalton Trans. 2011, 40, 3336.
1
9, 191. (i) Jeon, I.-R.; Park, J. G.; Haney, C. R.; Harris, T.
(26) (a) Wayland, B. B.; Drago, R. S.; Henneike, H. F. J. Am.
Chem. Soc. 1966, 88, 2455. (b) Stewart, W. E.; Siddall, T.
H. Chem. Rev. 1970, 70, 517. (c) Ming, L.-J.; Lauffer, R.
B.; Que, Jr., L. Inorg. Chem. 1990, 29, 3060.
D. Chem. Sci. 2014, 5, 2461. (j) Olatunde, A. O.; Bond, C.
J.; Dorazio, S. J.; Cox, J. M.; Benedict, J. B.; Daddario, M.
D.; Spernyak, J. A.; Morrow, J. R. Chem. Eur. J. 2015, 21,
1
8290. (k) Tsitovich, P. B.; Cox, J. M.; Benedict, J. B.;
(
(
27) Huang, C.-H.; Morrow, J. R. Inorg. Chem. 2009, 48, 7237.
28) (a) Woods, M.; Woessner, D. E.; Zhao, P.; Pasha, A.; Yang,
M.-Y.; Huang, C.-H.; Vasalitiy, O.; Morrow, J. R.; Sherry,
A. D. J. Am. Chem. Soc. 2006, 128, 10155. (b) Huang, C.-
H.; Morrow, J. R. J. Am. Chem. Soc. 2009, 131, 4206. (c)
Huang, C.-H.; Hammell, J.; Ratnakar, S. J.; Sherry, A. D.;
Morrow, J. R. Inorg. Chem. 2010, 49, 5963. (d) Hammell,
J.; Buttarazzi, L.; Huang, C.-H.; Morrow, J. R. Inorg. Chem.
Morrow, J. R. Inorg. Chem. 2016, 55, 700. (l) Du, K.;
Harris, T. D. J. Am. Chem. Soc. 2016, 138, 7804. (m) Du,
K.; Waters, E. A.; Harris, T. D. Chem. Sci. 2017, 8, 4424.
(
2
n) Burns, P. J.; Cox, J. M.; Morrow, J. R. Inorg. Chem.
017, 56, 4545.
(
12) (a) James, T. L. Nuclear Magnetic Resonance in Biochemis-
try: Principles and Applications; Academic Press: New
York, 1975. (b) Bertini, I.; Luchinat, C. NMR of
Paramagnetic Molecules in Biological Systems; The
Benjamin/Cummings Publishing Company, Inc.: Menlo
Park, 1986. (c) Bertini, I.; Luchinat, C. Coord. Chem. Rev.
1996, 150, 1. (d) Bertini, I.; Luchinat, C.; Parigi, G.
Solution NMR of Paramagnetic Molecules: Applications to
Metallobiomolecules and Models; Elsevier Science B.V.:
Amsterdam, 2001.
2
011, 50, 4857. (e) Ferrauto, G.; Castelli, D. D.; Terreno,
E.; Aime, S. Magn. Reson. Med. 2013, 69, 1703.
29) Note that the common logarithm with base 10 was used
throughout this study. For brevity, the notation “log” is em-
ployed in the pH calibration equations rather than the pre-
cise notation “log10.”
(
(30) (a) Ward, K. M.; Balaban, R. S. Magn. Reson. Med. 2000,
44, 799. (b) Terreno, E.; Castelli, D. D.; Cravotto, G.; Mi-
lone, L.; Aime, S. Invest. Radiol. 2004, 39, 235.
(
(
13) APEX2, version 2014.11-0; Bruker Analytical X-ray Sys-
tems, Inc.: Madison, WI, 2014.
14) Sheldrick, G. M. SADABS, version 2.03; Bruker Analytical
X-ray Systems, Inc.: Madison, WI, 2000.
(
(
31) Wood, P. M. Biochem. J. 1988, 253, 287.
32) (a) Raymond, G. G.; Born, J. L. Drug Intell. Clin. Pharm.
1
2
986, 20, 683. (b) Vepsäläinen, J. J. Curr. Med. Chem.
002, 9, 1201. (c) Alanne, A.-L.; Hyvönen, H.; Lahtinen,
(15) (a) Sheldrick, G. M. SHELXTL, version 6.12; Bruker Ana-
lytical X-ray Systems, Inc.: Madison, WI, 2000. (b) Shel-
drick, G. M. Acta Crystallogr., Sect. A: Found. Adv. 2015,
M.; Ylisirniö, M.; Turhanen, P.; Kolehmainen, E.; Peränie-
mi, S.; Vepsäläinen, J. J. Molecules 2012, 17, 10928.
7
1, 3.
(33) (a) Hu, J.; Zhao, J.; Hou, H.; Fan, Y. Inorg. Chem. Com-
mun. 2008, 11, 1110. (b) Guo, Z.-F.; Li, B.; Guo, J.-Z.;
Yang, P.; Shi, L.-F.; Liu, L. Transition Met. Chem. 2014,
(
16) Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J.
A. K.; Puschmann, H. J. Appl. Crystallogr. 2009, 42, 339.
17) OriginPro, version 9.0; OriginLab, Corp.: Northampton,
MA, 2003.
(
3
9, 353.
1
2
ACS Paragon Plus Environment

Page 13 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
3
ACS Paragon Plus Environment