Protic media enhanced protodeboronation for a potential H2O2-sensitive ligand system

Article abstract of DOI:10.1016/j.poly.2020.114377

The detection of elevated concentrations of reactive oxygen species (ROS) such as H₂O₂ represents a promising route to identify oxidative misregulation in cells, a hallmark of carcinoma and neurodegeneration. Magnetic resonance imaging (MRI) presents an opportunity to detect ROS over a greater portion of the body compared to strategies such as fluorescence. To sense hydrogen peroxide, potential Fe²⁺ and Co²⁺ MRI sensors were synthesized by appending pinacolborane to two amino-pyridine ligand scaffolds. These were designed to expose phenol upon reaction with H₂O₂, turning on PARACEST (Paramagnetic Chemical Exchange Saturation Transfer) and allowing detection by MRI. Instead of yielding a phenol upon exposure to oxidative conditions, protodeboronation was observed as the major product even under mild conditions. Optimization of oxidation conditions allowed access to the phenolic analogue of the two ligands studied in 11% and 17% conversion under organic conditions. Treatment of the Fe²⁺ and Co²⁺ complexes with H₂O₂ resulted in rapid, partial demetallation with no evidence of the bound or free ligand undergoing an oxidative transformation. These results stress the importance of establishing stability of potential MRI contrast agents in aqueous conditions.

Full text of DOI:10.1016/j.poly.2020.114377

Journal Pre-proofs
Protic Media Enhanced Protodeboronation for a Potential H O -Sensitive Li-
2
2
gand System
Cheryl L. Mathis, Caroline T. Saouma
PII:
S0277-5387(20)30034-6
DOI:
https://doi.org/10.1016/j.poly.2020.114377
POLY 114377
Reference:
To appear in:
Polyhedron
Received Date:
Revised Date:
Accepted Date:
15 September 2019
11 January 2020
17 January 2020
Please cite this article as: C.L. Mathis, C.T. Saouma, Protic Media Enhanced Protodeboronation for a Potential
H O -Sensitive Ligand System, Polyhedron (2020), doi: https://doi.org/10.1016/j.poly.2020.114377
2
2
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will
undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing
this version to give early visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
©
2020 Elsevier Ltd. All rights reserved.

Protic Media Enhanced Protodeboronation for a Potential H O -Sensitive Ligand System
2
2
a
a*
Cheryl L. Mathis and Caroline T. Saouma
aDepartment of Chemistry, University of Utah, 315 S. 1400 E., Salt Lake City, Utah 84112,
U.S.A.
Abstract
The detection of elevated concentrations of reactive oxygen species (ROS) such as H O represents
2
2
a promising route to identify oxidative misregulation in cells, a hallmark of carcinoma and
neurodegeneration. Magnetic resonance imaging (MRI) presents an opportunity to detect ROS
over a greater portion of the body compared to strategies such as fluorescence. To sense hydrogen
2
+
2+
peroxide, potential Fe and Co MRI sensors were synthesized by appending pinacolborane to
two amino-pyridine ligand scaffolds. These were designed to expose phenol upon reaction with
H O , turning on PARACEST (Paramagnetic Chemical Exchange Saturation Transfer) and
2
2
allowing detection by MRI. Instead of yielding a phenol upon exposure to oxidative conditions,
protodeboronation was observed as the major product even under mild conditions. Optimization
of oxidation conditions allowed access to the phenolic analogue of the two ligands studied in 11%
2
+
2+
and 17% conversion under organic conditions. Treatment of the Fe and Co complexes with
H O resulted in rapid, partial demetallation with no evidence of the bound or free ligand
2
2
undergoing an oxidative transformation. These results stress the importance of establishing
stability of potential MRI contrast agents in aqueous conditions.
1
.) Introduction
Reactive oxygen species (ROS) such as hydrogen peroxide (H O ) play crucial roles in
2
2
normal cell signaling pathways.[1] An increase in ROS concentration results in oxidative stress;
this has been linked to cancer, neurodegeneration, and cardiovascular & inflammatory diseases.[2–
4
] It remains critical to understand the role of ROS in healthy tissues and how imbalances result
in malignancies. There is thus a need to develop probes to detect ROS in vivo.
The direct detection of H O in vivo is non-trivial, and sensors that rely on the detection of
2
2
downstream oxidation products have been developed.[5] A chemical moiety that has been
exploited for this purpose is the boronic ester, which reacts with H O to produce alcohols. The
2
2
alcohol may be detected, or may further react, as exemplified in self-immolating pathways in
which phenyl boronic esters ultimately decompose to CO and quinone methide.[6] Regardless of
2
the ultimate product, they can be quantified by a variety of techniques. Most prominent in this

regard is the detection of changes in luminescence or fluorescence upon oxidation.[6–9] Due to
the short wavelengths of irradiation required and the inability to deeply penetrate tissues, this
approach has limitations for in vivo detection despite offering high spatial resolution and
1
8
sensitivity.[10] Recently, an F positron emission tomography (PET) probe was developed that
relies on the oxidation of a boronic ester to a phenol.[11] This approach circumvents the limitations
1
8
of fluorescence but necessitates a source of the radioisotope F, which may limit widespread use.
Magnetic resonance imaging (MRI) is a widely used imaging technique that may be able
to sense ROS. Typically, MRI contrast is generated from water molecules in the body that
3
+
exchange and coordinate to a paramagnetic lanthanide ion such as Gd ; while widely used, this
approach does not allow for selective sensing of analytes.[12–14] However, recent advancements
have led to the development of stimulant-responsive imaging agents that exploit different aspects
of MRI. One example of redox selectivity senses a NAD+/NADH mimic in which the CEST-silent
ligand is reduced and thereby activated.[15] Alternatively, detecting hypoxia in tissues can be
2
+/
3+
achieved by sensitizing a Co Co redox complex to turn on or off, dependent on partial pressures
of O .[16] Similar redox sensitivity is accessed through the use of thiols, which crosslink in
2
oxidizing environments.[17] With regards to H O detection specifically, contrast can be achieved
2
2
by encapsulating MRI-responsive gadolinium oxide nanoparticles that upon exposure to H O
2
2
triggers release.[18] Detection is also achieved by Mn(II) complexes that feature redox-active
phenol and quinol functionalities, whereby oxidation gives dimerized species or quinones,
respectively.[19,20] In this case, a change in the relaxivity of the complexes allows for detection.
1
3
Hyperpolarization of a C MRI signal can be achieved in the H O mediated oxidative
2
2
1
9
decarboxylation of a ketoacid,[21] and F MRI can be used for the detection of fluorophenols,
produced via H O boronic acid oxidation (vide supra).[22]
2
2
O
B
HO
H O
2 2
O
R
R
R
R
1
1
N
2
M
N
N
N
N
2
M
+
N
N
N
N
+
H2O2
2
OH
B
N
N
O
O
N
2
Figure 1. Strategy of a masked PARACEST contrast agent. A paramagnetic ion is ligated by
nitrogen-based chelating ligands featuring “masked” moieties 1 and 2 in the backbone. On the left,
the complex does not serve as a contrast agent. Upon exposure to H O (right), the gray mask 1
2
2
that corresponds to a boronic ester is cleaved, exposing a phenol active for MRI contrast (inset).
Optional inclusion of an orthogonal purple mask, 2, would allow for ratiometric sensing if removal
under non-identical gave a chemically different exchangeable proton.

Given the promise of MRI to detect H O , and the applicability of boronic esters, we aim
2
2
to develop contrast agents by bringing these two approaches together (Figure 1). The oxidation of
a boronic esters to a phenol exposes a proton that readily exchanges with water. If exchange is
slow on the NMR timescale, then selective irradiation of the exchanging resonance will result in a
decrease in the signal attributed to bulk water, producing contrast via Chemical Exchange
Saturation Transfer (CEST).[23] Indeed, H O can directly be detected in this manner, using a
2
2
modified pulse-sequence.[24] The use of paramagnetic metal ions results in PARACEST, whereby
the exchanging proton is paramagnetically shifted from bulk water allowing for faster exchange to
be detected by CEST. Recently, a turn-on approach, whereby a masked contrast agent is activated
for MRI or PARACEST, has been used to sense pH,[25] temperature,[26] and Cu(II) ions.[27]
Moreover, there is potential to append two different moieties to a paramagnetic ion that turn-on
PARACEST in the presence of different analytes. Provided the resulting products have very
distinct chemical shifts, this would allow for ratiometric sensing of analytes (Figure 1).
Herein we describe our efforts to develop PARACEST contrast agents for H O sensing.
2
2
We have prepared two N-based ligands that feature tethered boronic esters, and have installed the
ligands on Fe(II) and Co(II). We studied the fate of these ligands and complexes under oxidative
and biological conditions. Our results indicate that stability under physiological conditions must
be considered when designing boronic ester masked H O sensors. Notably, we found that
2
2
exposure of the ligand or metal complexes to aqueous H O and other protic media resulted in
2
2
protodeboronation and not the desired oxidation to phenol; only oxidation under organic
conditions yielded the desired pathway (Scheme 1).
5
Equiv. H O
2
2
o
THF, 0 C
NR2
HO
NR2
B
O
O
Favored Product
Under all
H
NR2
5
Equiv. H O
2
o
2
Conditions
R-OH, 25 C
Scheme 1: Fate of boronic ester under organic (top) and protic (bottom) conditions. The bottom
pathway gives a protodeboronated product.
2
.) Results and Discussion
2
.1. Synthesis of Ligands.
In order to probe the possible CEST activity of pinacolborane in the presence of H O , two
2
2
3
ligand scaffolds were synthesized (Scheme 2): tridentate N (N,N-bis(pyridin-2-ylmethyl)-4-
6
1
1
2
2
(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)aniline) and hexadentate
N
(N ,N ,N ,N -
tetrakis(pyridin-2-ylmethyl)-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzene-1,2-
diamine). The ligands are outfitted with an aryl pinacolborane which could be cleaved in the
presence of H O to yield a corresponding phenol.[6] While phenol resonances are quite
2
2
deshielded in diamagnetic spectra, we anticipated that the additional influence of a paramagnetic

metal ion would further enhance the chemical shift separation between the phenolic -OH and H O,
2
improving CEST exchange resolution.[23] Chelating nitrogen based amino/pyridine ligands were
chosen given the known robustness of similar ligands under oxidizing conditions;[28] iron and
cobalt complexes are readily prepared,[29,30] and the chelate effect minimizes demetallation.
6
Hexadentate N was designed to maximize stability via the chelate effect. Moreover, we
3
envisioned that a metal can be coordinated by N and another tridentate ligand featuring a masked
imaging agent, allowing for ratiometric sensing (Figure 1).
6
3
Scheme 2: Amino-pyridine ligands, N and N , that feature the aryl boronic esters “mask.”
3
The synthesis of the tridentate amino-pyridine ligand (N ) was accessed in three steps
starting from iodo-aniline as described in the literature.[31] Though the hexadentate analogue,
6
(
N ), is not known in the literature, 1,2-bis[bis(2-pyridylmethyl)amino]benzene, [32] which lacks
B(pin), has been prepared and we followed a modified synthetic protocol (Scheme 3). Briefly, 4-
bromobenzene-1,2-diamine was treated with 4 equiv of 2-chloromethylpyridine hydrochloride
under basic conditions to afford bis-pyridine diamine (1).[33] Using additional equivalents of 2-
chloromethylpyridine did not result in formation of tetra-pyridine diamine (2). Indeed, 1,2-
bis[bis(2-pyridylmethyl)amino]benzene is only prepared as a mixture of the tri- and tetra-pyridine.
As previously reported, a stronger base was needed to deprotonate the disubstituted amine. [34]
Treatment of (1) with 3 equiv of NaH and 4 equiv. of 2-chloromethylpyridine afforded the desired
tetra-pyridine diamine (2) in 79% yield. Installation of pinacolborane was then achieved through
a palladium cross coupling reaction between the aryl bromide and bis(pinacolato)diboron to give
6
ligand (N ) as an orange oil in an overall yield of 21%.
6
Scheme 3: Preparation of N . Reaction conditions: a) H O, 4 equiv 2-chloromethylpyridine
2
hydrochloride, hexadecyl trimethylammonium chloride; b) THF, 4 equiv 2-chloromethylpyridine,
3
equiv NaH; c) 1,4-dioxane, 1.2 equiv bis(pinacolato)diboron, 3 equiv KOAc, 5 mol%
Pd(dppf)∙CH Cl (dppf = 1,1’-Bis(diphenylphosphino)ferrocene).
2
2

2
.2. Synthesis of Metal Complexes.
6
Treatment of N with 1 equiv of Fe(II)Cl results in a color change from beige to yellow,
2
6
1
suggesting coordination and formation of N FeCl (Scheme 4, M = Fe). The H NMR spectrum
2
of the crude reaction mixture shows 12 broad paramagnetically shifted resonances, and none that
6
6
6
+
are attributed to free N . ESI-MS for N FeCl gives an m/z of 689.1, consistent with [N FeCl] .
2
6
While this indicates coordination of N to the Fe, it does not allow us to determine if the chloride
ions are inner or outer sphere; attempts to grow crystals were not successful. To unequivocally
6
2+
access the hexadentate complex, [N Fe] , whereby all of the donor nitrogens coordinate the iron,
this complex was treated with 2 equiv of Tl(I)PF . Crystals suitable for XRD were grown from
6
vapor diffusion of diethyl ether into dichloromethane (Figure 2, left), and shows the fully
6
1
coordinated [N Fe][PF ] . The H NMR spectrum of this species shows 10 paramagnetically
6
2
6
shifted resonances. This spectrum is similar to that of N FeCl in that based on the symmetry of
2
the complex there are fewer than anticipated resonances, suggesting a structural similarity between
the two.
6
6
Scheme 4: Preparation of [N Fe][PF ] and [N Co][PF ] .
6
2
6 2
Fe
Co
N
B
C
N
B
O
C
O
6
6
Figure 2: Solid-state structure (50% thermal ellipsoids) of [N Fe][PF ] (left) and [N Co][PF ]
6
2
6 2
(
right). All calculated hydrogen atoms, solvent, and PF counterions have been removed for
6
clarity.
6
Similar to the reactivity observed between Fe(II)Cl and N , the reaction of Co(II)Cl and
2
2
6
6
N (Scheme 4, M = Co) results in a color change from blue to purple and formation of N CoCl .
Upon treatment with 2 equiv. Tl(I)PF , chloride abstraction ensues to give [N Co][PF ] . Crystals
of [N Co][PF ] suitable for XRD were grown from vapor diffusion of diethyl ether into
2
6
6
6 2
6
6
2
dichloromethane, and the structure is shown in Figure 2 (right). The similarities in the solid-state

6
6
1
structures of [N Co][PF ] and [N Fe][PF ] are reflected in the H NMR spectra with regards to
6
2
6 2
the number of observable resonances.
3
3
To prepare the N -ligated complexes we treated 2 equiv of N with Fe(II)Cl or Co(II)Cl
2
2
1
(
Scheme 5). With both metals we observe residual free ligand by H NMR spectroscopy. This
suggests that only one ligand coordinates the Fe(II) and Co(II). Furthermore, attempts to encourage
1
ligation using triflate metal salts demonstrated broadening of the H NMR ligand resonances
within the diamagnetic region and the absence of paramagnetic peaks, suggesting that stable
binding does not occur (see SI).
N
N
N
M(II)Cl2
N
O
O
Cl
Cl
B
N
M
o
DCM, 25 C, 24h
N
B
O
O
3
N FeCl
2
3
3
(
N )
N CoCl2
3
3
Scheme 5: Preparation N FeCl and N CoCl ; addition of Tl(I)PF does not result in clean
2
2
6
3
conversion to the bis ligated N complexes.
Attempts to grow crystals of either complex were unsuccessful, likely due to the presence of free
3
ligand. In a separate trial whereby a single equivalent of N was added to Fe(II)Cl , we were able
2
to grow crystals from a 1:1 mixture of MeOH and acetone. As seen in Figure 3 the loss of
H
3
pinacolborane suggests an instability to protic solvent, yielding a new product, N FeCl .
2
H
3
Figure 3: Top: Pinacolborane lost to protodeboronation under protic conditions to form N FeCl .
2
H
3
Bottom: Solid-state structure (50% thermal ellipsoids) of N FeCl . All calculated hydrogen
2

atoms have been removed for clarity. While assignments of paramagnetic peaks can be unclear,
3
H
3
the spectra of the as-prepared N FeCl and the crystals of N FeCl showed clear distinctions (see
2
2
SI).
.3. Boronic Ester Instability
The instability of the boronic ester in complex N FeCl under protic conditions raised
2
3
2
3
6
questions about the suitability of the N and N scaffolds for biocompatible chemistry. The
H
3
previously formed N FeCl lost its pinacolborane moiety which was replaced with a para
2
3
substituted hydrogen (Figure 3). When examining the ligand N alone (in the absence of metal),
further evidence for protodeboronation to produce N (Scheme 1) was found by H NMR
spectroscopy. Upon exposure of N to 2 equiv. of water in CD CN, a new triplet (7.1 ppm, J = 7.1
Hz) was observed to grow (Figure 4). This is consistent with the published compound N .[35]
H
3
1
3
3
H
3
8
50
T = 24 hours
T = 4 hours
N
N
N
B
3
2
1
800
N
N
HO
OH
2
equiv. H2O
N
B(OH)3
7
7
50
00
CD3CN, 25 °C
O
O
(HN3)
650
600
T = 0 hours
3
(
N )
5
5
4
4
3
3
2
2
1
1
5
0
50
00
50
00
50
00
50
00
50
00
0
1
.45
1.40
1.35
1.30
1.25
1.20
f1 (ppm)
1.15
1.10
1.05
1.00
0.95
-
50
8
.5
8.0
7.5
7.0
6.5
6.0
5.5
5.0
f1 (ppm)
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1
3
H
3
Figure 4: H NMR spectrum (400 MHz, 298 K, CD CN) of the decomposition of N to N (6.8
3
mM) is observed upon treatment with H O (14 mM) at room temperature, shown at t = 4 hours.
2
The red boxed signal corresponds to para H, whereas the green box corresponds to the shifted
H
3
methylene of N . Insert: Expanded pinacolborane region showing decomposition to pinacol
growth at 1.1 ppm; see SI) with 33% conversion in 24 hours.
(
Furthermore, the appearance of para hydrogen was accompanied by decomposition of the
pinacolborane singlet at 1.2 ppm. Decomposition yielded a new peak at 1.1 ppm over several
hours, which was found to be pinacol (Figure 4). Similar reactivity was observed for identical

6
reaction conditions with N ; the identity of pinacol was confirmed in both cases by addition of
pinacol to the sample, the peak at 1.1 ppm grows upon addition while others are unchanged (see
6
H
6
SI). In 24 hours, 35% conversion from N to N was achieved. Protodeboronation in similar
species where the para substituent is -OR (R = H, Me) is precedented in the literature.[36] We
suspect that the electron donating para amine substituent similarly contributes to the favorability
of the protodeboronation pathway in this system.
6
Additionally, similar reactivity was observed for the two metal complexes, [N Fe][PF ]
6
2
6
6
and [N Co][PF ] , upon treatment with water. When [N Fe][PF ] was treated with 2 equiv H O
in CD CN and monitored by H NMR spectroscopy, a new peak appears at 1.2 ppm concomitant
6
2
6 2
2
1
3
with disappearance at 1.6 ppm, tentatively assigned as the paramagnetic pinacolborane resonance.
6
Within 24 hours, 50% conversion to free pinacol was observed (see SI). [N Co][PF ] likewise is
6
2
not stable, suggesting the boronic ester is compromised in protic media regardless of coordination.
The instability of these complexes and ligands presented above raised concern as normal
CEST conditions necessitate the use of protic solvent (H O) for exchange. Previous reports of acid
2
promoted protodeboronation led us to consider neutral, buffered media.[36] Indeed, numerous
PARACEST reports use buffered media such as phosphate buffered saline (PBS),[37] 4-(2-
hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES),[38] and tromethamine (tris) buffers.[39]
In order to determine if boronic ester instability could be improved by using buffered solutions,
6
1
we monitored the stability of [N Fe][PF ] by H NMR spectroscopy in CD CN and 10 mM PBS
6
2
3
(pH = 7) as outlined by McMahon for CEST imaging.[32] At 24 hours, several new diamagnetic
impurities are observed, including major growth of pinacol at 1.2 ppm (see SI). Importantly, the
free ligands and metal complexes were unstable in the presence of protic species regardless of
whether the environment was organic or not (methanol and water) or if the environment was
buffered (PBS).
2
.5. H O Mediated Oxidation Reactions
2 2 .
In the pursuit of exploring oxidation reactivity to give the desired phenols, N and HON3
HO
6
6
3
(
Table 1), treatment of the two free ligands N and N under various conditions were explored.
Initial oxidation studies utilized the conditions proposed in the seminal work by Brown and Gupta
2 equiv. H O , 5N NaOH, 1:1 THF:MeOH – at both room temperature and 0 °C (entries 1-2).[40]
–
2
2
In both instances, these conditions provided the protodeboronated species described above, as
1
ascertained by H NMR spectroscopy. In the 1:1 THF:MeOH solvent mixture, we observed cloudy
precipitates, therefore reactions were conducted in anhydrous methanol to improve solubility
(entries 3-4). While the instability of the boronic ester in methanol is known (vide supra), we
hypothesized that inclusion of excess H O (20 equiv) would dominate reactivity yielding the
2
2
desired phenol product. When excess of H O was used, protodeboronation was observed (Table
2
2
1
); in contrast, stoichiometric H O yielded no reaction.
2 2

N
N
N
N
N
N
R
1-20 Equiv. H2O2 (20.4%)
Additives, Solvent, °C
R
30 min.
B
O
O
X
X = H, OH
3
N : R = H
6
N : R = N(CH -Py)
2
2
a
Solvent(s)b
Entry
Ligand
Equiv. H O
Temperature
Phenol
Conv. (%)
2
2
c
(Additives)
(°C)
3
d
1
2
3
4
5
6
N
2 (5 equiv. NaOH)
2 (5 equiv. NaOH)
1 (--)
1:1 THF:MeOH 25
1:1 THF:MeOH 0
MeOH
MeOH
THF
-- (21%)
3
3
3
3
6
d
N
N
N
N
N
-- (35%)
e
25
25
0
-- (NR)
f
20 (--)
5 (--)
5 (--)
-- (D)
11 %
THF
0
17 %
3
6
a
Table 1: Oxidative conditions for the conversion of N and N to their phenolic analogues. H O
is a 20.4 % solution in H O. Reactions were run as 0.4 M solutions. Room Temperature is
assumed to be approximately 25 C. Yield of protodeboronation. No Reaction. Degradation to
the starting material and unknown species (Conversion: 25% unknown).
2
2
b
c
2
o
d
e
f
Previous H O sensing reports with successful proof of concept oxidations of boronic esters
2
2
to phenols commonly utilize additives and organic cosolvents (e.g. DMF).[11,22] Synthetic reports
that use H O to cleave pinacolborane often employ harsher conditions and additives, however,
2
2
inclusion of the additives such as NaOH,[40] I ,[41] and NaIO [42] did not promote conversion to
2
4
3
6
phenol (see SI). Optimal reactivity for both N and N was observed at 0°C in THF with no
additives (entries 5-6). These conditions provided the phenol as a minor product, with 11%
3
6
1
conversion for N and 17% conversion for N detected by H NMR spectroscopy (see SI Tables 4-
). An extensive record of all conditions used to oxidize the boronic ester can be found in the SI.
7
6
6
Knowing the best conditions for the ligands, both [N Fe][PF ] and [N Co][PF ] were
6
2
6 2
1
treated analogously with 5 equiv of H O in CD CN at room temperature and observed by H
2
2
3
NMR spectroscopy for changes. Within 5 minutes, we observe appearance of ligand resonances
1
between 1-9 ppm in the H spectra, suggesting rapid demetallation. Notably, at five minutes, the
peaks between 6-9 ppm likely corresponding to the ligand appeared quite cleanly, a representative
6
example for [N Co][PF ] is shown in Figure 5. Additionally, the absence of shifting for the
6
2

2
PF6
2 PF6
N
N
O
2 2
5 Eq H O , (20.4%)
N
N
N
N
N
N
B
X
O
Co
N
Co
N
N
CD3CN
N
3
2
1
*
*
*
*
*
DCM
THF
THF
CD CN
3
2
0
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
f1 (ppm)
1
Figure 5: H NMR spectra (400 MHz, 298 K, CD CN) of demetallation and ligand degradation
3
6
6
for 27.2 mM [N Co][PF ] when treated with 5 equiv. H O . Top: [N Co][PF ] prior to H O
2
6
2
2
2
6 2
2
addition; Middle: treated with H O at 5 min; Bottom: after 30 min. Red asterisks are consistent
2
2
6
6
with the N ligand; the boxed region highlights decomposition N to unknown species.
paramagnetic resonances suggests no reactivity of the ligand while bound at the metal. At 30
minutes however, the released ligand decomposed to several new, unidentified degradation
products. The signals at 30 minutes are not consistent with either the protodeboronated or phenol
6
analogues of N . This is not an unexpected result, as several of the oxidation conditions tested gave
decomposition to unidentified species at room temperature (see SI).
3
. Conclusions
Herein we have presented the synthesis of a new ligand scaffold, N , and coordination
6
3
6
complexes of both N and N with Fe(II) and Co(II). These systems incorporate an aryl boronic
ester as pinacolborane. The presence of protic species are destabilizing in this system, producing
protodeboronated products. The protodeboronation chemistry found under mild conditions in the

presence of protic solvents and H O was found to dominate for both ligands, and a
2
2
protodeboronated metal complex was isolated and characterized. Protodeboronation for para -OR
R = H, Me) species is precedented in the literature;[36] we hypothesize that the electron donating
(
para amine substituent analogously participates in the protodeboronation pathway seen here.
Furthermore, during the course of this work, pH was found to be implicated in the rate of
protodeboronation. [43]
While several other reports demonstrate that boronic ester oxidation in the presence of
H O can be selectively achieved, conditions typically employ excess H O and organic
2
2
2
2
cosolvents.[44] However, in this system we observe preferential reaction of the aryl boronic ester
to form the protodeboronated species under oxidizing conditions. Reactions conducted in any
mixture of protic solvent yielding a protodeboronated product over the phenol (Scheme 1), and
inclusion of excess H O (20 equiv) cannot shift reactivity toward the desired phenol (Table 1).
2
2
The conditions found to produce the most phenol from the aryl boronic ester ligand require use of
the polar aprotic solvent THF, slight excess of H O (5 equiv.), and 0 °C temperature control.
2
2
6
3
Under the most optimized conditions, 17% (N ) and 11% (N ) conversion was achieved. While it
is quite common to include additives[40–42] and organic cosolvents[22,44] to improve reactivity
and solubility, the loss of pinacolborane in all protic media observed here is not optimal. These
conditions are concerning; in addition to the fundamental instability of the ligands in protic solvent,
the only successful method to produce phenol uncovered was inherently non-biocompatible.
Despite the success of control CEST experiments where no exchange is observed (see SI),
we concluded these systems are incompatible for CEST applications on the basis of instability in
protic media. The goal of this work was to capitalize on the oxidative sensitivity of aryl boranes
to yield phenols in pursuit of transforming a protected non-exchanging – and therefore MRI silent
–
resonance into an activated form. Importantly, CEST systems often utilize H O and buffered
2
solutions which are incompatible with the presented ligands and metal complexes.[23,37] In
conclusion, though the phenolic analogues can be prepared when treated with H O , the instability
2
2
of these species in protic solvents renders them incompatible with biological systems and for
applications as ROS sensors in PARACEST MRI.
4
.) Experimental
4
.1. General
All commercially available organic reagents were purchased from Sigma Aldrich or
Oakwood Chemical and were used as received without further purification unless stated otherwise.
All commercially available inorganic reagents were purchased from Strem Chemicals and were
used as received without further purification unless stated otherwise. All solvents used were
collected from a Pure Process Technology solvent purification system to remove oxygen and
water, stored over activated molecular sieves in a glovebox, and tested with ketyl radical before
use. Anhydrous 1,4-dioxane was purchased from Sigma Aldrich and used as received. Dry
methanol was distilled over CaH , and stored over dried molecular sieves. Acetone was stored
2
over molecular sieves to remove water. Molecular sieves were activated at 200 °C under vacuum
and stored under nitrogen atmosphere in a nitrogen filled MBraun glovebox. NMR solvents were

obtained from Cambridge Isotope Labs, subjected to 3 freeze-pump-thaw cycles, and stored under
nitrogen in the glovebox over sieves. Hydrogen peroxide (aq) solutions were calculated to be 2.6%
and 20.4% (w/w) as determined by titration against sodium thiosulfate. When distilled
chloromethylpyridine was used, the following method was followed: 2-Chloromethyl pyridine
hydrochloride was basicified to pH 8 in H O, extracted 3 times into EtOAc, dried with Na SO ,
2
2
4
1
concentrated in vacuo, and vacuum distilled to give a clear light pink oil. H NMR spectra were
recorded at 300 or 400 MHz (Varian) and chemical shifts are referenced to their respective
deuterated solvent and chemical shifts are reported in ppm. The abbreviations s, d, t, q, and m
stand for the resonance multiplicities singlet, doublet, triplet, quintet, and multiplet. All spectra
were analyzed using Mestrenova. X-ray crystallographic structures were obtained from single
crystals using a Bruker APEX CCD diffractometer. Preparation of ligands was carried out under
nitrogen gas using Schlenk techniques and glass was dried in 120 °C ovens unless otherwise stated.
Metalation reactions and subsequent complex purifications were conducted in the glovebox.
Elemental analyses of metal complexes were performed by Midwest MicroLab under air-free
conditions. Mass Spectrometry data was collected using a Waters Xevo G2S Q-ToF with Acquity
3
I Class UPLC Tandem Mass Spectrometer or an Acquity QDa Mass Detector. The ligand N was
prepared as described in the literature.[27]
4
.5 Preparation of 4-bromo-N1,N2-bis(pyridin-2-ylmethyl)benzene-1,2-diamine (1)
2
-Chloromethyl pyridine hydrochloride (214.46 mmol, 35.18 g) was added to 9 mL H O and
2
sparged under N for 15 minutes. 4-bromobenzene-1,2-diamine (53.75 mmol, 10.05 g) and
2
hexadecyl trimethyl ammonium chloride (1.02 mmol, 0.33 g) were added to the flask which was
sparged again under N for 15 minutes, yielding a reddish brown solution. The mixture stirred at
2
room temperature under N atmosphere for 48 hours. The diphasic, red-brown solution was
2
extracted 3 times with 15 mL dichloromethane (DCM) and 10 mL H O. The organic layer was
2
dried over Na SO , filtered, and concentrated. The red oil was purified by column
2
4
chromatography: EtOAc gradient (10%-50% vs. hexanes) on silica gel; R : 0.6 on 50:50 EtOAc:
f
1
Hexanes. Yield: 15.1 g, 76%. H NMR (400 MHz, CDCl ): δ 4.39 (d, J = 7.0 Hz, 4H, -CH -),
3
2
5
.82 (d, J = 7.9 Hz, 1H, Ar), 6.08 (d, J = 1.8 Hz, 1H, Ar), 6.55 (d, J = 1.8/ 8.1 Hz, 1H, Ar), 7.20
(m, 2H, Py), 7.44 (d, J = 7.8 Hz, 2H, Py), 7.67 (m, 2H, Py), 8.59 (m, 2H, Py).
4
.6 Preparation of 4-bromo-N1,N1,N2,N2-tetrakis(pyridin-2-ylmethyl)benzene-1,2-diamine (2)
In the glovebox, 1 (2.31 mmol, 0.85 g) was added to 5 mL THF and the loaded flask was
transferred to the Schlenk line. The freshly distilled 2-chloromethyl pyridine (8.76 mmol,1.12 g)
was added via syringe. A stirring solution of 5 mL of THF and NaH (6.91 mmol, 0.17 g) were
added by cannula yielding a brown solution. The mixture was refluxed (60 °C) overnight under
N . The reaction was quenched with 15 mL of EtOH, and concentrated. The oil was extracted 3x
2
into 10 mL DCM vs. 10 mL H O. The organic layer was dried over Na SO , filtered, and
2
2
4
concentrated. The resulting brown oil was purified by column chromatography: 70:30 EtOAc:
1
hexanes on alumina; R : 0.35. Yield: 1.0 g, 79%. H NMR (300 MHz, CDCl ): δ 4.71(d, J = 6.4
f
3
Hz, 8H, -CH -), 6.70 (d, J = 8.6 Hz, 1H, Ar), 6.85 (dd, J = 2.9/ 8.5 Hz, 1H, Ar), 6.97 (d, J = 2.3
2
Hz, 1H, Ar), 7.04-14 (m, 8H, Py), 7.40-7.52 (m, 4H, Py), 8.45-8.52 (m, 4H, Py).

4
.7 Preparation of N1,N1,N2,N2-tetrakis(pyridin-2-ylmethyl)-4-(4,4,5,5-tetramethyl-1,3,2-
6
dioxaborolan-2-yl)benzene-1,2-diamine (N )
2
(6.16 mmol, 3.39 g) was dissolved in 16 mL of anhydrous 1,4-dioxane, and sparged under N2
for 15 minutes. Bis(pinacolato)diboron (7.39 mmol, 1.87 g), KOAc (18.5 mmol, 1.82 g), and
Pd(dppf)-CH Cl (0.31 mmol, 0.25 g) were added to an oven dried flask, and dissolved in 2 mL
2
2
dry 1,4-dioxane and sparged under N for 15 minutes. The solution of 5 was added to the second
2
flask via syringe, and the reaction was refluxed overnight in a 90 °C oil bath. The clear orange
solution was filtered over celite, washed with EtOAc, dried over Na SO , and concentrated. The
2
4
red oil was purified by column chromatography: gradient EtOAc (10-50% vs. hexanes) on
+
1
alumina; R : 0.3 in EtOAc. Yield: 0.99 g, 27%. Mass Spectrometry (ESI ) (m/z): 598.2. H NMR
f
(
300 MHz, CD CN): δ 1.26 (s, 12H, -CH ), 4.66 (s, 4H, -CH -), 4.79 (s, 4H, -CH -), 6.88 (d, J =
3 3 2 2
7
7
8
.9 Hz, 1H, Ar), 7.06 (d, J = 9.5 Hz, 1H, Ar), 7.19-7.09 (m, 5H, Ar, Py), 7.20 (m, 4H, Py), 7.57-
.46 (m, 4H, Py), 8.43 (d, J = 6.6 Hz, 4H, Py). C NMR (400 MHz, CD3CN): δ 24.2, 58.4, 59.1,
3.4, 115.7, 121.3, 121.9, 122.2, 122.8, 129.4, 134.4, 136.3, 136.6, 149.0, 165.6.
1
3
4
.8 Preparation of N1,N1,N2,N2-tetrakis(pyridin-2-ylmethyl)-4-(4,4,5,5-tetramethyl-1,3,2-
6
dioxaborolan-2-yl)benzene-1,2-diamine iron(II) chloride (N FeCl )
2
6
N (0.18 mmol, 0.11 g) was dissolved in 4 mL DCM in the glovebox. 0.19 mmol (0.02 g)
anhydrous Fe(II)Cl was stirred in 1 mL DCM and the beige slurry was added dropwise to the
2
6
solution of N via pipette. The solution stirred for 24 hours at room temperature and became
yellow-brown. The solution was concentrated until saturated and 5 mL diethyl ether (DEE) was
added, a light yellow solid precipitated at 0 °C. Yield: 7.8 mg, 6 %. The filtered and dried material
6
was used directly in the preparation of ([N Fe][PF ] ) without further purification. Mass
6
2
+
1
Spectrometry (ESI ) (m/z): 689.1. H NMR (400 MHz, CD Cl ): δ 1.1 (s), 1.5 (s), 1.6 (s), 3.4
2
2
(s), 14.4 (s), 16.3 (s), 16.5 (s), 19.7 (s), 33.9 (s), 34.9 (s), 42.5 (s), 44.5 (s).
4
.9 Preparation of N1,N1,N2,N2-tetrakis(pyridin-2-ylmethyl)-4-(4,4,5,5-tetramethyl-1,3,2-
6
dioxaborolan-2-yl)benzene-1,2-diamine iron(II) hexafluorophosphate ([N Fe][PF ] )
6
2
6
TlPF (0.36 mmol, 0.13 g) was added to a rapidly stirring solution of (N FeCl ) (0.19 mmol) in
6
2
DCM (0.5 mL) in the glovebox. Immediately a light precipitate was formed, and the reaction was
stirred overnight to give a chartreuse green cloudy solution. The solution was filtered over celite
to provide a dark green clear solution, which was concentrated to a minimum of solvent in vacuo.
6
Crystals of ([N Fe][PF ] ) suitable for x-ray crystallography were grown at 0 °C by vapor
6
2
diffusion of DEE into DCM. Yield: 107 mg, 60%. Evan’s Method (DCM): 5.1 B.M. Mass
+
-1
-1
Spectrometry (ESI ) (m/z): 327.1. UV-vis (CH Cl ) λ , nm (ε, M cm ): 215 (19430), 230
2
2
max
1
(
16560), 265 (8470), 245 (6850). H NMR (400 MHz, CD Cl ): δ -1.2 (s), 8.3-8.6 (m) 10.3 (s)
2 2
5
3.5 (s), 60.0 (s), 71.7 (s), 94.7 (s), 97.1 (s), 116.5 (s).

4
.10 Preparation of N1,N1,N2,N2-tetrakis(pyridin-2-ylmethyl)-4-(4,4,5,5-tetramethyl-1,3,2-
6
dioxaborolan-2-yl)benzene-1,2-diamine cobalt(II) chloride (N CoCl ).
2
6
N (0.085 mmol, 0.05 g) was dissolved in 0.5 mL DCM in the glovebox. Anhydrous Co(II)Cl
2
,
(
0.085 mmol, 0.01 g), was stirred with 0.5 mL DCM, the clear blue solution was added dropwise
6
to N via pipette. The solution stirred for 24 hours at room temperature and became burgundy. The
solution was concentrated to a minimum and 2 mL DEE was added, a dark red solid precipitated
6
at 0 °C. The filtered and dried material was used directly in the preparation of [N Co][PF ] without
6
2
+
1
further purification. Yield: 2.5 mg, 4%. Mass Spectrometry (ESI )(m/z): 692.1. H NMR (400
MHz, CD Cl ): δ 0.2 (s), 2.6 (s), 3.4 (s), 8.4 (s), 8.8 (s), 9.2 (s), 16.4 (s), 18.0 (s), 23.3 (s), 26.8
2
2
(s), 29.8 (s), 41.3 (s), 47.3 (s).
4
.11 Preparation of N1,N1,N2,N2-tetrakis(pyridin-2-ylmethyl)-4-(4,4,5,5-tetramethyl-1,3,2-
6
dioxaborolan-2-yl)benzene-1,2-diamine cobalt(II) hexafluorophosphate ([N Co][PF ] )
6
2
6
TlPF (0.17 mmol, 0.06 g) was added to a rapidly stirring solution of (N CoCl ) (0.085 mmol,
6
2
0
.062 g) in DCM (0.5 mL) in the glovebox. Immediately a light precipitate was formed, and the
reaction was stirred overnight to give a dark mauve cloudy solution. The solution was filtered over
celite to provide a dark purple clear solution, which was concentrated until super saturated in
6
vacuo. Crystals of [N Co][PF ] suitable for x-ray crystallography were grown with DCM and
6
2
DEE as solvents at 0 °C by vapor diffusion. Yield: 25.8 mg, 16%. Evan’s Method (DCM): 3.8
+
-1
-1
B.M. Mass Spectrometry (ESI ) (m/z): 328.7. UV-vis (CH Cl ) λ , nm (ε, M cm ): 220 (11790),
2
1
2
max
2
30 (12860), 240 (6740), 260 (12340), 305 (1520), 365 (1740). H NMR (400 MHz, CD Cl ): δ -0.6
2
2
(s), 10.0 (s), 10.8 (s), 13.3 (s), 16.6-16.9 (m), 44.5 (m), 63.1 (m), 73.1 (s), 190.8 (s), 194.0 (s).
4
.12 Preparation of N,N-bis(pyridin-2-ylmethyl)-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-
3
yl)aniline iron(II) chloride (N FeCl )
2
3
N (2.5 mmol, 1.01 g) was dissolved in 2.5 mL DCM in the glovebox. Anhydrous Fe(II)Cl (1.26
2
mmol, 0.16 g) was stirred with 2.5 mL DCM, the beige slurry was added dropwise to the solution
3
of N via pipette. The solution stirred for 24 hours at room temperature and became a vibrant, clear
1
yellow. When no further changes were observed by H NMR, the material was filtered over celite,
1
and dried in vacuo. Yield: 0.79g, 60%. H NMR (400 MHz, CD Cl ): δ -17.7 (s), 0.3 (s), 3.1 (s),
2
2
4
.1 (s), 9.0 (s), 52.7 (s), 54.1 (s), 123.5 (s).
4
.14 Preparation of N,N-bis(pyridin-2-ylmethyl)-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-
3
yl)aniline cobalt(II) chloride (N CoCl )
2
3
N (2.5 mmol, 1.0 g) was dissolved in 2.5 mL DCM in the glovebox. Anhydrous Co(II)Cl (1.25
2
3
mmol, 0.16 g) was stirred with 2.5 mL DCM and was added dropwise to the solution of N via
pipette. The dark blue solution stirred for 24 hours at room temperature and became a vibrant,
1
clear blue. When no further changes were observed by H NMR, the material was filtered over
1
celite, and dried in vacuo. Yield: 0.66 g, 50 %. H NMR (400 MHz, (CD ) SO): δ -52.0 (s), -51.2
3
2
(s), -22.1 (s), -1.7 (s), 15.8 (s), 48.6 (s), 65.3 (s), 66.2 (s), 86.9 (s).

5
. References
[
1] S.G. Rhee, H O , a Necessary Evil for Cell Signaling, Science, 312 (2006) 1882–1883.
2
2
doi:10.1126/science.1130481.
[2] J.P. Fruehauf, F.L. Meyskens, Reactive Oxygen Species: A Breath of Life or Death?, Am.
Assoc. Cancer. Res., 13 (2007) 789–794. doi:10.1158/1078-0432.ccr-06-2082.
[3] M.P. Mattson, Pathways towards and away from Alzheimer’s disease, Nature, 430 (2004)
6
31-639. doi:10.1038/nature02621.
[4] D. Harman, The aging process, Proc. Natl. Acad. Sci. U.S.A., 78 (1981) 7124–7128.
doi:10.1073/pnas.78.11.7124.
[5] R.S. Balaban, S. Nemoto, T. Finkel, Mitochondria, Oxidants, and Aging, Cell, 120 (2005)
4
83–495. doi:10.1016/j.cell.2005.02.001.
[6] A.R. Lippert, G.C. Van de Bittner, C.J. Chang, Boronate Oxidation as a Bioorthogonal
Reaction Approach for Studying the Chemistry of Hydrogen Peroxide in Living Systems, Acc.
Chem. Res., 44 (2011) 793–804. doi:10.1021/ar200126t.
[
7] L-F. Lu, Y-Y. Li, M. Zhang, G. Shi, Visual fluorescence detection of H O and glucose based
2 2
on “molecular beacon”-hosted Hoechst dyes, Analyst, 140 (2015) 3642–3647.
doi:10.1039/c5an00499c.
[8] F. Wang, X. Liu, C-H. Lu, I. Willner, Cysteine-Mediated Aggregation of Au Nanoparticles:
The Development of a H O Sensor and Oxidase-Based Biosensors, ACS Nano., 7 (2013) 7278–
2
2
7
286. doi:10.1021/nn402810x.
[9] A.E. Albers, V.S. Okreglak, C.J. Chang, A FRET-Based Approach to Ratiometric
Fluorescence Detection of Hydrogen Peroxide, J. Am. Chem. Soc., 128 (2006) 9640–9641.
doi:10.1021/ja063308k.
[10] J.P. Miller, D. Maji, J. Lam, B.J. Tromberg, S. Achilefu, Noninvasive depth estimation
using tissue optical properties and a dual-wavelength fluorescent molecular probe in vivo,
Biomed. Opt. Express., 8 (2017) 3095–3109. doi:10.1364/boe.8.003095.
[
11] V. Carroll, B.W. Michel, J. Blecha, H. VanBrocklin, K. Keshari, D. Wilson, C.J. Chang, A
1
8
Boronate-Caged [ F]FLT Probe for Hydrogen Peroxide Detection Using Positron Emission
Tomography, J. Am. Chem. Soc., 136 (2014) 14742–14745. doi:10.1021/ja509198w.

[12] P. Caravan, J.J. Ellison, T.J. McMurry, R.B. Lauffer, Gadolinium(III) Chelates as MRI
Contrast Agents: Structure, Dynamics, and Applications, Chem. Rev., 99 (1999) 2293–2352.
doi:10.1021/cr980440x.
[13] S. Zhang, P. Winter, K. Wu, A.D. Sherry, A Novel Europium(III)-Based MRI Contrast
Agent, J. Am. Chem. Soc., 123 (2001) 1517–1518. doi:10.1021/ja005820q.
[14] K. Mason, N.J. Rogers, E.A. Suturina, I. Kuprov, J.A. Aguilar, A.S. Batsanov, D.S. Yufit,
D. Parker, PARASHIFT Probes: Solution NMR and X-ray Structural Studies of Macrocyclic
Ytterbium and Yttrium Complexes, Inorg. Chem., 56 (2017) 4028-4038.
doi:10.1021/acs.inorgchem.6b02291.
[
15] S.J. Ratnakar, S. Viswanathan, Z. Kovacs, A.K. Jindal, K.N. Green, A.D. Sherry,
Europium(III) DOTA-tetraamide Complexes as Redox-Active MRI Sensors, J. Am. Chem. Soc.,
34 (2012) 5798-5800. doi: 10.1021/ja211601k
1
[
16] P.B. Tsitovich, J.A. Spernyak, J.R. Morrow, A Redox-Activated MRI Contrast Agent that
Switches Between Paramagnetic and Diamagnetic States, Angew. Chem. Int. Ed., 52 (2013)
3379-14000. doi: 0.1002/anie.201306394.
1
[17] P.B. Tsitovich, P.J. Burns, A.M. McKay, J.R. Morrow, Redox-activated MRI contrast
agents based on lanthanide and transition metal ions, J. Inorg. Biochem., 133 (2014), 143-154.
[
18] M.L. Viger, J. Sankaranarayanan, C. de Gracia Lux, M. Chan, A. Almutairi, Collective
Activation of MRI Agents via Encapsulation and Disease-Triggered Release, J. Am. Chem. Soc.,
35 (2013) 7847–7850. doi:10.1021/ja403167p.
1
[19] M. Yu, M.B. Ward, A. Franke, S.L. Ambrose, Z.L. Whaley, T.M. Bradford, J.D. Gorden,
R.J. Beyers, R.C. Cattley, I. Ivanović-Burmazović, D.D. Schwartz, C.R. Goldsmith, Adding a
Second Quinol to a Redox-Responsive MRI Contrast Agent Improves Its Relaxivity Response to
H O , Inorg. Chem., 56 (2017) 2812-2826. doi:10.1021/acs.inorgchem.6b02964.
2
2
[20] M. Yu, S.L. Ambrose, Z.L. Whaley, S. Fan, J.D. Gorden R. J. Beyers, D.D. Schwartz, C.R.
Goldsmith, A Mononuclear Manganese(II) Complex Demonstrates a Strategy To Simultaneously
Image and Treat Oxidative Stress, J. Am. Chem. Soc., 136 (2014) 12836-12839.
[
21] A.R. Lippert, K.R. Keshari, J. Kurhanewicz, C.J. Chang, A Hydrogen Peroxide-Responsive
1
3
Hyperpolarized C MRI Contrast Agent, J. Am. Chem. Soc., 133 (2011) 3776–3779.
doi:10.1021/ja111589a.
[
22] H. Nonaka, Q. An, F. Sugihara, T. Doura, A. Tsuchiya, Y. Yoshioka, S. Sando,
1
9
Phenylboronic Acid-based F MRI Probe for the Detection and Imaging of Hydrogen Peroxide
Utilizing Its Large Chemical-Shift Change, Anal. Sci., 31 (2015) 331–335.
doi:10.2116/analsci.31.331.

[23] P.C.M. van Zijl, N.N. Yadav, Chemical exchange saturation transfer (CEST): What is in a
name and what isn’t?, Magn. Reson. Med., 65 (2011) 927–948. doi:10.1002/mrm.22761.
[24] D. Ryoo, X. Xu, Y. Li, J.A. Tang, J. Zhang, P.C.M. van Zijl, G. Liu, Detection and
Quantification of Hydrogen Peroxide in Aqueous Solutions Using Chemical Exchange
Saturation Transfer (CEST), Anal. Chem., 89 (2017) 7758–7764.
doi:10.1021/acs.analchem.7b01763.
[
25] A.E. Thorarinsdottir, K. Du, J.H.P. Collins, D.T. Harris, Ratiometric pH Imaging with a
II
Co MRI Probe via CEST Effects of Opposing pH Dependences, J. Am. Chem. Soc., 139
2
(2017) 15836-15847. doi:10.1021/jacs.7b08574.
[26] S. Zhang, C.R. Malloy, A.D. Sherry, MRI Thermometry Based on PARACEST Agents, J.
Am. Chem. Soc., 127 (2005) 17572–17573. doi:10.1021/ja053799t.
[27] N.N. Paranawithana, A.F. Martins, V.C. Jordan, P. Zhao, S. Chirayil, G. Meloni, A.D.
Sherry, A Responsive Magnetic Resonance Imaging Contrast Agent for Detection of Excess
Copper(II) in the Liver In Vivo, J. Am. Chem. Soc., 141 (2019) 11009–11018.
doi:10.1021/jacs.8b13493.
[28] Z. Zhu, H. Chen, S. Li, X. Yang, E. Bittner, C. Cai, Tripodal amine ligands for accelerating
Cu-catalyzed azide–alkyne cycloaddition: efficiency and stability against oxidation and
dissociation, Catal. Sci. Technol., 7 (2017) 2474–2485. doi:10.1039/c7cy00587c.
[
29] J.L. Fillol, Z. Codolà, I. Garcia-Bosch, L. Gómez, J.J. Pla, M. Costas, Efficient water
oxidation catalysts based on readily available iron coordination complexes, Nat. Chem., 3 (2011)
07-813. doi:10.1038/nchem.1140.
8
[30] L. Lewis, N. Fogel, Thermochemistry of cobalt(II)chloride pyridine complexes, J. Chem.
Soc. A., (1970) 1141–1146. doi:10.1039/j19700001141.
[31] C. Gwizdala, C.V. Singh, T.R. Friss, J.C. MacDonald, S.C. Burdette, Quantifying factors
that influence metal ion release in photocaged complexes using ZinCast derivatives, Dalton
Trans., 41 (2012) 8162–8174. doi: 10.1039/c2dt30135k.
[32] H. Wu, X. Zhang, X. Yin, Y. Inaba, H. Miki, K. Takeshita, Selective separation of
cadmium(II) from zinc(II) by a novel hydrophobic ionic liquid including an N, N, N′, N′-
tetrakis(2-methylpyridyl)-1,2-phenylenediamine-4-amido structure: a hard–soft donor combined
method, Dalton Trans., 47 (2018) 10063–10070. doi:10.1039/c8dt02228c.
[33] S. Turba, O. Walter, S. Schindler, L.P. Nielsen, A. Hazell, C.J. McKenzie, F. Lloret, J.
Cano, M. Julve, Syntheses, Structures, and Properties of Copper(II) Complexes of Bis(2-

pyridylmethyl) Derivatives of o-, m-, and p-Phenylenediamine and Aniline, Inorg. Chem., 47
(2008) 9612–9623. doi:10.1021/ic800244g.
[34] N. Zhao, Y-H. Wu, H-M. Wen, X. Zhang, Z-N. Chen, Conversion from ILCT to
LLCT/MLCT Excited State by Heavy Metal Ion Binding in Iridium(III) Complexes with
Functionalized 2,2′-Bipyridyl Ligands, Organometallics, 28 (2009) 5603–5611.
doi:10.1021/om900334f.
[35] T. Maruyama, Y. Fujie, N. Oya, E. Hosaka, A. Kanazawa, D. Tanaka, Y. Hattori, J.
Motoyoshiya, Control of peroxyoxalate chemiluminescence by nitrogen-containing ligand
quenching: turning off and on by ligand–metal ion host–guest interactions, Tetrahedron, 67
(2011) 6927–6933. doi:10.1016/j.tet.2011.06.078.
[36] G. Zhang, Y. Li, J. Liu, Acid-promoted metal-free protodeboronation of arylboronic acids,
RSC Adv., 7 (2017) 34959–34962. doi:10.1039/c7ra05979e.
[37] X. Yang, X. Song, Y. Li, G. Liu, S.R. Banerjee, M.G. Pomper, M.T. McMahon, Salicylic
Acid and Analogues as diaCEST MRI Contrast Agents with Highly Shifted Exchangeable Proton
Frequencies , Angew. Chem. Int. Ed., 52 (2013) 8116–8119. doi:10.1002/anie.201302764.
1
9
[
38] M. Yu, B.S. Bouley, D. Xie, E.L. Que, Highly fluorinated metal complexes as dual F and
PARACEST imaging agents, Dalton Trans., 48 (2019) 9337–9341. doi:10.1039/c9dt01852b.
[39] C-Y Lin, N.N. Yadav, J. Ratnakar, A.D. Sherry, P.C.M. Zijl, In vivo imaging of paraCEST
agents using frequency labeled exchange transfer MRI, Magn. Reson. Med., 71 (2014) 286–293.
doi:10.1002/mrm.24603.
[40] H.C. Brown, S.K. Gupta, Hydroboration. XXXIX. 1,3,2-Benzodioxaborole (catecholborane)
as a new hydroboration reagent for alkenes and alkynes. General synthesis of alkane- and
alkeneboronic acids and esters via hydroboration. Directive effects in the hydroboration of
alkenes and alkynes with catecholborane, J. Am. Chem. Soc., 97 (1975) 5249–5255.
doi:10.1021/ja00851a038.
[41] A. Gogoi, U. Bora, An Iodine-Promoted, Mild and Efficient Method for the Synthesis of
Phenols from Arylboronic Acids, Synlett., 23 (2012) 1079–1081. doi:10.1055/s-0031-1290654.
[42] A.J. Neel, A. Milo, M.S. Sigman, F.D. Toste, Enantiodivergent Fluorination of Allylic
Alcohols: Data Set Design Reveals Structural Interplay between Achiral Directing Group and
Chiral Anion, J. Am. Chem. Soc., 138 (2016) 3863–3875. doi:10.1021/jacs.6b00356.
[43] P.A. Cox, A.G. Leach, A.D. Campbell, G.C. Lloyd-Jones, Protodeboronation of
Heteroaromatic, Vinyl, and Cyclopropyl Boronic Acids: pH-Rate Profiles, Autocatalysis, and
Disproportionation, J. Am. Chem. Soc., 138 (2016) 9145-9157. doi: 10.1021/jacs.6b03283.

[44] V.S. Lin, B.C. Dickinson, C.J. Chang, Boronate-Based Fluorescent Probes Imaging
Hydrogen Peroxide in Living Systems, Methods Enzymol., 526 (2013) 19–43.
doi:10.1016/b978-0-12-405883-5.00002-8.

Protic Media Enhanced Protodeboronation for a Potential H O -Sensitive Ligand System
2
2
a
a*
Cheryl L. Mathis and Caroline T. Saouma
aDepartment of Chemistry, University of Utah, 315 S. 1400 E., Salt Lake City, Utah 84112,
U.S.A.
Abstract
The detection of elevated concentrations of reactive oxygen species (ROS) such as H O represents
2
2
a promising route to identify oxidative misregulation in cells, a hallmark of carcinoma and
neurodegeneration. Magnetic resonance imaging (MRI) presents an opportunity to detect ROS
over a greater portion of the body compared to strategies such as fluorescence. To sense hydrogen
2
+
2+
peroxide, potential Fe and Co MRI sensors were synthesized by appending pinacolborane to
two amino-pyridine ligand scaffolds. These were designed to expose phenol upon reaction with
H O , turning on PARACEST (Paramagnetic Chemical Exchange Saturation Transfer) and
2
2
allowing detection by MRI. Instead of yielding a phenol upon exposure to oxidative conditions,
protodeboronation was observed as the major product even under mild conditions. Optimization
of oxidation conditions allowed access to the phenolic analogue of the two ligands studied in 11%
2
+
2+
and 17% conversion under organic conditions. Treatment of the Fe and Co complexes with
H O resulted in rapid, partial demetallation with no evidence of the bound or free ligand
2
2
undergoing an oxidative transformation. These results stress the importance of establishing
stability of potential MRI contrast agents in aqueous conditions.
1
.) Introduction

Reactive oxygen species (ROS) such as hydrogen peroxide (H O ) play crucial roles in
2
2
normal cell signaling pathways.[1] An increase in ROS concentration results in oxidative stress;
this has been linked to cancer, neurodegeneration, and cardiovascular & inflammatory diseases.[2–
4
] It remains critical to understand the role of ROS in healthy tissues and how imbalances result
in malignancies. There is thus a need to develop probes to detect ROS in vivo.
The direct detection of H O in vivo is non-trivial, and sensors that rely on the detection of
2
2
downstream oxidation products have been developed.[5] A chemical moiety that has been
exploited for this purpose is the boronic ester, which reacts with H O to produce alcohols. The
2
2
alcohol may be detected, or may further react, as exemplified in self-immolating pathways in
which phenyl boronic esters ultimately decompose to CO and quinone methide.[6] Regardless of
2
the ultimate product, they can be quantified by a variety of techniques. Most prominent in this
regard is the detection of changes in luminescence or fluorescence upon oxidation.[6–9] Due to
the short wavelengths of irradiation required and the inability to deeply penetrate tissues, this
approach has limitations for in vivo detection despite offering high spatial resolution and
1
8
sensitivity.[10] Recently, an F positron emission tomography (PET) probe was developed that
relies on the oxidation of a boronic ester to a phenol.[11] This approach circumvents the limitations
1
8
of fluorescence but necessitates a source of the radioisotope F, which may limit widespread use.
Magnetic resonance imaging (MRI) is a widely used imaging technique that may be able
to sense ROS. Typically, MRI contrast is generated from water molecules in the body that
3
+
exchange and coordinate to a paramagnetic lanthanide ion such as Gd ; while widely used, this
approach does not allow for selective sensing of analytes.[12–14] However, recent advancements
have led to the development of stimulant-responsive imaging agents that exploit different aspects
of MRI. One example of redox selectivity senses a NAD+/NADH mimic in which the CEST-silent
ligand is reduced and thereby activated.[15] Alternatively, detecting hypoxia in tissues can be
2
+/
3+
achieved by sensitizing a Co Co redox complex to turn on or off, dependent on partial pressures
of O .[16] Similar redox sensitivity is accessed through the use of thiols, which crosslink in
2
oxidizing environments.[17] With regards to H O detection specifically, contrast can be achieved
2
2
by encapsulating MRI-responsive gadolinium oxide nanoparticles that upon exposure to H O
2
2
triggers release.[18] Detection is also achieved by Mn(II) complexes that feature redox-active
phenol and quinol functionalities, whereby oxidation gives dimerized species or quinones,
respectively.[19,20] In this case, a change in the relaxivity of the complexes allows for detection.
1
3
Hyperpolarization of a C MRI signal can be achieved in the H O mediated oxidative
2
2
1
9
decarboxylation of a ketoacid,[21] and F MRI can be used for the detection of fluorophenols,
produced via H O boronic acid oxidation (vide supra).[22]
2
2

O
B
HO
H O
2 2
O
R
R
R
R
1
1
N
2
M
N
N
N
N
N
2
M
+
N
N
N
N
+
H2O2
2
OH
B
N
O
O
N
2
Figure 1. Strategy of a masked PARACEST contrast agent. A paramagnetic ion is ligated by
nitrogen-based chelating ligands featuring “masked” moieties 1 and 2 in the backbone. On the left,
the complex does not serve as a contrast agent. Upon exposure to H O (right), the gray mask 1
2
2
that corresponds to a boronic ester is cleaved, exposing a phenol active for MRI contrast (inset).
Optional inclusion of an orthogonal purple mask, 2, would allow for ratiometric sensing if removal
under non-identical gave a chemically different exchangeable proton.
Given the promise of MRI to detect H O , and the applicability of boronic esters, we aim
2
2
to develop contrast agents by bringing these two approaches together (Figure 1). The oxidation of
a boronic esters to a phenol exposes a proton that readily exchanges with water. If exchange is
slow on the NMR timescale, then selective irradiation of the exchanging resonance will result in a
decrease in the signal attributed to bulk water, producing contrast via Chemical Exchange
Saturation Transfer (CEST).[23] Indeed, H O can directly be detected in this manner, using a
2
2
modified pulse-sequence.[24] The use of paramagnetic metal ions results in PARACEST, whereby
the exchanging proton is paramagnetically shifted from bulk water allowing for faster exchange to
be detected by CEST. Recently, a turn-on approach, whereby a masked contrast agent is activated
for MRI or PARACEST, has been used to sense pH,[25] temperature,[26] and Cu(II) ions.[27]
Moreover, there is potential to append two different moieties to a paramagnetic ion that turn-on
PARACEST in the presence of different analytes. Provided the resulting products have very
distinct chemical shifts, this would allow for ratiometric sensing of analytes (Figure 1).
Herein we describe our efforts to develop PARACEST contrast agents for H O sensing.
2
2
We have prepared two N-based ligands that feature tethered boronic esters, and have installed the
ligands on Fe(II) and Co(II). We studied the fate of these ligands and complexes under oxidative
and biological conditions. Our results indicate that stability under physiological conditions must
be considered when designing boronic ester masked H O sensors. Notably, we found that
2
2
exposure of the ligand or metal complexes to aqueous H O and other protic media resulted in
2
2
protodeboronation and not the desired oxidation to phenol; only oxidation under organic
conditions yielded the desired pathway (Scheme 1).

5
Equiv. H O
2
2
o
THF, 0 C
NR2
HO
H
NR2
B
O
O
Favored Product
Under all
NR2
5
Equiv. H O
2
o
2
Conditions
R-OH, 25 C
Scheme 1: Fate of boronic ester under organic (top) and protic (bottom) conditions. The bottom
pathway gives a protodeboronated product.
2
.) Results and Discussion
2
.1. Synthesis of Ligands.
In order to probe the possible CEST activity of pinacolborane in the presence of H O , two
2
2
3
ligand scaffolds were synthesized (Scheme 2): tridentate N (N,N-bis(pyridin-2-ylmethyl)-4-
6
1
1
2
2
(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)aniline) and hexadentate
N
(N ,N ,N ,N -
tetrakis(pyridin-2-ylmethyl)-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzene-1,2-
diamine). The ligands are outfitted with an aryl pinacolborane which could be cleaved in the
presence of H O to yield a corresponding phenol.[6] While phenol resonances are quite
2
2
deshielded in diamagnetic spectra, we anticipated that the additional influence of a paramagnetic
metal ion would further enhance the chemical shift separation between the phenolic -OH and H O,
2
improving CEST exchange resolution.[23] Chelating nitrogen based amino/pyridine ligands were
chosen given the known robustness of similar ligands under oxidizing conditions;[28] iron and
cobalt complexes are readily prepared,[29,30] and the chelate effect minimizes demetallation.
6
Hexadentate N was designed to maximize stability via the chelate effect. Moreover, we
3
envisioned that a metal can be coordinated by N and another tridentate ligand featuring a masked
imaging agent, allowing for ratiometric sensing (Figure 1).
6
3
Scheme 2: Amino-pyridine ligands, N and N , that feature the aryl boronic esters “mask.”
3
The synthesis of the tridentate amino-pyridine ligand (N ) was accessed in three steps
starting from iodo-aniline as described in the literature.[31] Though the hexadentate analogue,
6
(
N ), is not known in the literature, 1,2-bis[bis(2-pyridylmethyl)amino]benzene, [32] which lacks
B(pin), has been prepared and we followed a modified synthetic protocol (Scheme 3). Briefly, 4-
bromobenzene-1,2-diamine was treated with 4 equiv of 2-chloromethylpyridine hydrochloride
under basic conditions to afford bis-pyridine diamine (1).[33] Using additional equivalents of 2-

chloromethylpyridine did not result in formation of tetra-pyridine diamine (2). Indeed, 1,2-
bis[bis(2-pyridylmethyl)amino]benzene is only prepared as a mixture of the tri- and tetra-pyridine.
As previously reported, a stronger base was needed to deprotonate the disubstituted amine. [34]
Treatment of (1) with 3 equiv of NaH and 4 equiv. of 2-chloromethylpyridine afforded the desired
tetra-pyridine diamine (2) in 79% yield. Installation of pinacolborane was then achieved through
a palladium cross coupling reaction between the aryl bromide and bis(pinacolato)diboron to give
6
ligand (N ) as an orange oil in an overall yield of 21%.
6
Scheme 3: Preparation of N . Reaction conditions: a) H O, 4 equiv 2-chloromethylpyridine
2
hydrochloride, hexadecyl trimethylammonium chloride; b) THF, 4 equiv 2-chloromethylpyridine,
3
equiv NaH; c) 1,4-dioxane, 1.2 equiv bis(pinacolato)diboron, 3 equiv KOAc, 5 mol%
Pd(dppf)∙CH Cl (dppf = 1,1’-Bis(diphenylphosphino)ferrocene).
2
2
2
.2. Synthesis of Metal Complexes.
6
Treatment of N with 1 equiv of Fe(II)Cl results in a color change from beige to yellow,
2
6
1
suggesting coordination and formation of N FeCl (Scheme 4, M = Fe). The H NMR spectrum
2
of the crude reaction mixture shows 12 broad paramagnetically shifted resonances, and none that
6
6
6
+
are attributed to free N . ESI-MS for N FeCl gives an m/z of 689.1, consistent with [N FeCl] .
2
6
While this indicates coordination of N to the Fe, it does not allow us to determine if the chloride
ions are inner or outer sphere; attempts to grow crystals were not successful. To unequivocally
6
2+
access the hexadentate complex, [N Fe] , whereby all of the donor nitrogens coordinate the iron,
this complex was treated with 2 equiv of Tl(I)PF . Crystals suitable for XRD were grown from
6
vapor diffusion of diethyl ether into dichloromethane (Figure 2, left), and shows the fully
6
1
coordinated [N Fe][PF ] . The H NMR spectrum of this species shows 10 paramagnetically
6
2
6
shifted resonances. This spectrum is similar to that of N FeCl in that based on the symmetry of
2
the complex there are fewer than anticipated resonances, suggesting a structural similarity between
the two.

6
6
Scheme 4: Preparation of [N Fe][PF ] and [N Co][PF ] .
6
2
6 2
Fe
Co
N
B
C
N
B
O
C
O
6
6
Figure 2: Solid-state structure (50% thermal ellipsoids) of [N Fe][PF ] (left) and [N Co][PF ]
6
2
6 2
(
right). All calculated hydrogen atoms, solvent, and PF counterions have been removed for
6
clarity.
6
Similar to the reactivity observed between Fe(II)Cl and N , the reaction of Co(II)Cl and
2
2
6
6
N (Scheme 4, M = Co) results in a color change from blue to purple and formation of N CoCl .
Upon treatment with 2 equiv. Tl(I)PF , chloride abstraction ensues to give [N Co][PF ] . Crystals
of [N Co][PF ] suitable for XRD were grown from vapor diffusion of diethyl ether into
2
6
6
6 2
6
6
2
dichloromethane, and the structure is shown in Figure 2 (right). The similarities in the solid-state
6
6
1
structures of [N Co][PF ] and [N Fe][PF ] are reflected in the H NMR spectra with regards to
6
2
6 2
the number of observable resonances.
3
3
To prepare the N -ligated complexes we treated 2 equiv of N with Fe(II)Cl or Co(II)Cl
2
2
1
(
Scheme 5). With both metals we observe residual free ligand by H NMR spectroscopy. This
suggests that only one ligand coordinates the Fe(II) and Co(II). Furthermore, attempts to encourage
1
ligation using triflate metal salts demonstrated broadening of the H NMR ligand resonances
within the diamagnetic region and the absence of paramagnetic peaks, suggesting that stable
binding does not occur (see SI).