Novel double prodrugs of the iron chelator N,N'-bis(2-hydroxybenzyl)ethylenediamine-N,N'-diacetic acid (HBED): Synthesis, characterization, and investigation of activation by chemical hydrolysis and oxidation

European Journal of Medicinal Chemistry 1 18 (2016) 193e207

Contents lists available at ScienceDirect

European Journal of Medicinal Chemistry

journal homepage: h ttp://www.elsevier.com/locate/ejmech

Research paper

Novel double prodrugs of the iron chelator N,N⁰-bis(2-hydroxybenzyl)

ethylenediamine-N,N⁰-diacetic acid (HBED): Synthesis,

characterization, and investigation of activation by chemical

hydrolysis and oxidation

,

Nikki A. Thiele ^{a *}, Khalil A. Abboud ^b, Kenneth B. Sloan ^a

^aDepartment of Medicinal Chemistry, University of Florida, Gainesville, FL 32610, USA

^bDepartment of Chemistry, University of Florida, Gainesville, FL 32611, USA

a r t i c l e i n f o

a b s t r a c t

Article history:

The development of iron chelators suitable for the chronic treatment of diseases where iron accumu-

lation and subsequent oxidative stress are implicated in disease pathogenesis is an active area of

research. The clinical use of the strong chelator N,N⁰-bis(2-hydroxybenzyl)ethylenediamine-N,N⁰-diacetic

acid (HBED) and its alkyl ester prodrugs has been hindered by poor oral bioavailability and lack of

conversion to the parent chelator, respectively. Here, we present novel double prodrugs of HBED that

have the carboxylate and phenolate donors of HBED masked with carboxylate esters and boronic acids/

esters, respectively. These double prodrugs were successfully synthesized as free bases (7a-f) or as

dimesylate salts (8a-c,e), and were characterized by ¹H, ¹³C, and ¹¹B NMR; MP; MS; and elemental

analysis. The crystal structure of 8a was solved. Three of the double prodrugs (8a-c) were selected for

further investigation into their abilities to convert to HBED by stepwise hydrolysis and H₂O₂oxidation.

The serial hydrolysis of the pinacol and methyl esters of N,N⁰-bis(2-boronic acid pinacol ester benzyl)

ethylenediamine-N,N⁰-diacetic acid methyl ester dimesylate (8a) was veriﬁed by LC-MS. The macro half-

lives for the hydrolyses of 8a-c, measured by UV, ranged from 3.8 to 26.3 h at 37 ^ꢀC in pH 7.5 phosphate

buffer containing 50% MeOH. 9, the product of hydrolysis of 8a-c and the intermediate in the conversion

pathway, showed little-to-no afﬁnity for iron or copper in UV competition experiments. 9 underwent a

serial oxidative deboronation by H₂O₂in N-methylmorpholine buffer to generate HBED

(k ¼ 10.3 M^ꢁ1min^ꢁ1). The requirement of this second step, oxidation, before conversion to the active

chelator is complete may confer site speciﬁcity when only localized iron chelation is needed. Overall,

these results provide proof of principle for the activation of the double prodrugs by chemical hydrolysis

and H₂O₂oxidation, and merit further investigation into the protective capabilities of the prodrugs

against H₂O₂-induced cell death.

Received 2 December 2015

Received in revised form

23 March 2016

Accepted 13 April 2016

Available online 14 April 2016

Keywords:

Iron

Double prodrug

Prochelator

Oxidative stress

Boronic acid

Chelation therapy

1. Introduction

synthesis, hydrogen peroxide decomposition, and biotransforma-

tion [1]. Due to the low solubility of Fe(III) hydroxide

Nearly every organism requires iron. Iron is essential as a

cofactor for proteins involved in biological processes such as DNA

synthesis, cellular metabolism, oxygen transport, neurotransmitter

(K_spz 1 ꢂ 10^ꢁ38) [2], the predominant form of the metal in the

biosphere, organisms have developed sophisticated systems to take

up, transport, and store iron. However, in humans, there is no

regulated mechanism for excretion of iron from the body [3,4].

Excess intracellular iron will enter the labile iron pool e a redox-

active and chelatable pool of iron that is purported to be

composed of ferrous (Fe^2þ) and ferric (Fe^3þ) iron loosely bound to

small molecules (e.g. phosphate, citrate, phospholipid head groups,

polypeptides, etc.) [5e7]. Similarly, a pool of redox-active labile

plasma iron has been recognized [8]. Both pools of iron can be

Abbreviations: ROS, reactive oxygen species; DFO, desferrioxamine B; HBED,

N,N⁰-bis(2-hydroxybenzyl)ethylenediamine-N,N⁰-diacetic acid; NMM, N-methyl-

morpholine; EDTA, ethylenediaminetetraacetic acid; FAC, ferric ammonium citrate.

* Corresponding author.

E-mail addresses: nikkithiele@yahoo.com (N.A. Thiele), abboud@chem.u ﬂ .edu

(K.A. Abboud), sloan@cop.u ﬂ .edu (K.B. Sloan).

http://dx.doi.org/10.1016/j.ejmech.2016.04.037

194

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

considered non-transferrin bound iron [9], and are capable of

catalyzing the formation of reactive oxygen species (ROS) via the

well-known Fenton reaction. The Fenton reaction produces highly

reactive hydroxyl radicals from the one-electron reduction of

hydrogen peroxide by ferrous iron [10,11]. Fe^2þcan then be re-

generated by cellular reductants (e.g. glutathione, ascorbic acid) or

superoxide anion, and is available to react once again in the Fenton

reaction to produce more toxic hydroxyl radicals. These radicals can

damage lipids, proteins, and DNA, and may ultimately lead to cell

death [12].

excess iron and oxidative stress.

N,N⁰-bis(2-hydroxybenzyl)ethylenediamine-N,N⁰-diacetic acid

(HBED, Scheme 1) is a hexadentate ligand that strongly chelates

Fe^3þ(log K ¼ 39.68), and, to a lesser extent, Cu^2þand Zn^2þ(log

K ¼ 21.38 and 18.37, respectively) [18]. HBED showed promise in

rodent models of iron overload when orally administered [19]. For

example, when HBED was orally administered to hypertransfused

rats, it was 70% as effective as DFO given by intraperitoneal injec-

tion [20]. However, the chelator performed poorly in iron-

overloaded Cebus apella primates [21] and in

b-thalassemia pa-

Excess iron and oxidative stress have been implicated in the

pathogenesis of a variety of diseases and conditions [13,14]. For

example, in hereditary hemochromatosis and b-thalassemia, sys-

tients [22] due to its poor absorption from the gastrointestinal tract.

With two carboxylic acids (pK_avalues of 2.5 and 1.7) and two

amines (pKa values of 8.3 and 4.6) [18,23], HBED may carry a net

charge of ꢁ1 or ꢁ2 at physiological pH.

temic iron overload is caused by a genetic mutation and obligatory

blood transfusions, respectively. Diseases associated with a more

focal accumulation of labile iron include, but are not limited to,

neurodegenerative diseases such as Parkinson's disease, Alz-

heimer's disease, Friedreich's ataxia, and neurodegeneration with

brain iron accumulation [15]; age-related macular degeneration

[16]; and skin photo-aging [17]. Patients with these conditions may

beneﬁt from iron chelation therapy to scavenge the excess metal,

thereby preventing further production of ROS via Fenton chemistry.

There are three chelators currently approved by the FDA for the

treatment of iron overload: desferrioxamine B (DFO), deferasirox,

and deferiprone. Unfortunately, DFO is poorly absorbed from the

gastrointestinal tract, has a short plasma half-life, and does not

cross the blood brain barrier. It must be administered by continuous

subcutaneous infusion. Deferasirox and deferiprone, while orally

active, have some potential toxicities associated with their use.

Exjade^®(deferasirox) has a boxed warning of possible renal

toxicity, hepatic toxicity, and gastrointestinal hemorrhage. Ferri-

prox^®(deferiprone) has a boxed warning of neutropenia/agranu-

locytosis. All three chelators may also be suboptimal for the

treatment of conditions and diseases that require localized chela-

tion therapy because they are all strong chelators, and thus they

pose a risk of purloining iron and other metals from proteins and

processes dependent upon these metals. Therefore, there remains a

need for the development of iron chelators with more desirable

properties to better treat patients with diseases associated with

In an effort to increase the bioavailability of HBED, its carboxylic

acid groups were masked as simple alkyl esters. These ester pro-

drugs were orally active in hypertransfused rodents [20,24], but

once again fell short in iron-overloaded C. apella [21,25] because the

esters were unable to undergo enzymatic [26] or chemical [24]

hydrolysis to give the active chelator in vivo. Testing in the

C. apella model of iron overload is considered a good predictor of

the behavior of chelators in humans.

Given the limitations of HBED and its simple esters, prodrugs of

HBED with more attractive properties (e.g. ability to convert to the

parent compound chemically and/or enzymatically) are desirable

for treatment of diseases where oxidative stress and iron overload

are implicated. Additionally, for diseases that are associated with

localized iron accumulation or that require only site-speciﬁc che-

lation, it would be beneﬁcial for a prodrug to preferentially convert

to HBED only at those sites where chelation is most needed, thereby

minimizing the risk of sequestering iron required for normal

function.

Here, we introduce a novel series of HBED prodrugs of the

general form 7 (Scheme 1). These prodrugs not only have both

carboxylic acid groups of HBED masked as alkyl esters to potentially

enhance membrane permeability, but also have both phenolic hy-

droxyl groups masked as boronic acids or esters. We postulated that

masking the two phenolate donors of HBED as boronic acids/esters

would 1) minimize off-target chelation by decreasing the afﬁnity of

Scheme 1. Postulated activation pathway for the double prodrugs of HBED. Double prodrugs of the general form 7 have little-to-no afﬁnity for Fe^3þ, and may undergo chemical

hydrolysis at physiological pH to form prodrug 9. In tissues where high levels of ROS (e.g. H₂O₂) are being generated, prodrug 9 may then undergo oxidation to generate the parent

chelator HBED.

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

195

these compounds for iron and other metals, especially when the

two carboxylate donors are also masked as esters; 2) catalyze the

chemical hydrolysis of the carboxylate esters at physiological pH;

and 3) provide a handle for the site-speciﬁc activation of the pro-

drugs by oxidative stress.

9.84. However, the elemental analysis of 3a was unsatisfactory. The

discrepancy between the predicted values (C, 58.59; H, 6.76; N,

8.54) and the observed values (C, 63.13, 63.25; H, 6.49, 6.34; N, 9.21,

9.15) in the elemental analysis most certainly resulted from dehy-

dration of the boronic acid moiety(ies) in 3a to boron oxide(s) when

it was dried in vacuo with moderate heating [46]. Under these

conditions, boronic acids can form cyclotrimeric anhydrides (bor-

oxines) and oligomeric acyclic analogues [47]. Similarly, boronic

acids are subject to gas-phase dehydration and anhydride forma-

tion in the ion source of a mass spectrometer [47], making it

difﬁcult to see the parent ion. ESI mass spectrometric analysis of 3a

in positive mode showed M þ H minus one or two H₂O. However,

M þ H itself was not detected.

Based on these hypotheses, the postulated activation pathway of

the prodrugs is given in Scheme 1. With 4 of the 6 donor atoms in

HBED masked, prodrugs of the general form 7 should have low

afﬁnity for Fe^3þand other physiologically-important metals.

Chemical hydrolysis of the carboxylate and boronic esters of 7 to

their corresponding acids should give 9. With the phenolate donors

still masked as boronic acids, 9 would be expected to exhibit low

afﬁnity for Fe^3þ, and could function as a prodrug in its own right.

When exposed to ROS (e.g. H₂O₂), oxidative deboronation should

unmask the latent phenols [27] to generate the active chelator

HBED, which can then scavenge free iron to prevent further Fenton

chemistry. Although hydrolysis may be tissue nonspeciﬁc, the

boronic acid oxidation step, and thus the complete conversion to

HBED, should occur preferentially in tissues where abundant Fen-

ton chemistry is creating a highly-oxidative environment. While

oxidative stress-activatable “prochelators” are well-known in the

literature [28e41], to our knowledge this is the ﬁrst report of the

use of cascade latentiation (i.e. “double prodrugs”) [42] to deliver a

synthetic chelator after activation by hydrolysis and oxidation.

Described herein are the synthesis and characterization of the

double prodrugs 7, and the results from the investigation into their

activation to HBED.

Determination of accurate melting points of free boronic acids is

also hindered by dehydration [48]. Indeed, we observed that a

sample of 3a sintered at 145e150 ^ꢀC, expanded into a more web-

like form at 165 ^ꢀC, browned gradually thereafter, and ﬁnally

melted at 212e218 ^ꢀC. The same sample re-melted without effer-

vescence at approximately 215e217 ^ꢀC, indicating that no decom-

position had occurred. In contrast, Zhang et al. [45] reported the

melting point of 3a to be 142e148 ^ꢀC. Likely, this instead repre-

sented the dehydration point for 3a, which could have easily been

mistaken for melting.

The synthesis of N,N⁰-bis(2-boronic acid pinacol ester benzyl)

ethylenediamine (5a) from 3a and pinacol (4a) proceeded unre-

markably in toluene (Scheme 2). The chemical shift of free pinacol

in CDCl₃is 1.24 ppm, whereas the chemical shift of the pinacol peak

in the ¹H NMR of 5a was 1.31 ppm. This chemical shift difference,

along with the appropriate integration of 24 H for the latter peak,

suggested that 5a was formed successfully. Given the dehydrated

state of compound 3a, the quantities of 3a and 4a that were needed

for the reaction were somewhat difﬁcult to accurately determine.

Ultimately, reaction quantities were based on the assumption that

the molecular weight of 3a was 309.97 g/mol (minus 1H₂O), rather

than 327.98 g/mol.

The ﬁnal step leading to the synthesis of double prodrugs 7a-c

proceeded smoothly by reﬂuxing the appropriate alkyl bromoace-

tates with compound 5a and DIPEA in dry ACN overnight, followed

by an ether extraction (Scheme 2). These prodrugs were oils. In an

attempt to purify these compounds, silica gel column

2. Results and discussion

2.1. Synthesis and characterization

The synthesis of a series of double prodrugs of HBED (7a-f) and

their corresponding dimesylate salts 8a-c,e is shown in Scheme 2.

N,N⁰-bis(2-boronic acid benzyl)ethylenediamine (3a) was ﬁrst

synthesized via reductive amination using a protocol modiﬁed

from Gray and Houston [43] and Gray et al. [44] (Scheme 2). The ¹H

NMR and ¹³C NMR spectra of 3a in CD₃OD were in agreement with

the spectral data reported by Gray et al. [44] and Zhang et al. [45],

respectively, and the ¹¹B NMR spectrum showed only one peak, at

Scheme 2. Synthesis of HBED double prodrugs 7a-f, their corresponding dimesylate salts 8a-c,e, and 9.

196

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

chromatography was performed on a batch of 7c. However, some

fractions containing 7c were contaminated with free pinacol, sug-

gesting that the boronic acid pinacol esters of 7c were susceptible

to hydrolysis on the column to their corresponding boronic acids.

Facile hydrolysis of aryl boronic pinacol esters has also been

observed by the Franz lab [38,39] and by Achilli et al. [49].

To investigate whether this facile hydrolysis of the boronic es-

ters could be avoided, the pinacol esters of 7a were replaced with

pinanediol esters to give prodrug 7e (N,N⁰-bis(2-boronic acid

pinanediol ester benzyl)ethylenediamine-N,N⁰-diacetic acid methyl

ester) (Scheme 2). This prodrug was synthesized by ﬁrst

condensing 3a with (1R,2R,3S,5R)-(-)-pinanediol (4b) to give 5c

(N,N⁰-bis(2-boronic acid pinanediol ester benzyl)ethylenediamine).

To then make 7e, 5c was reacted with methyl iodoacetate that had

been generated in situ from methyl bromoacetate (6a) using the

Finkelstein reaction. The crude material was successfully puriﬁed

by ﬂash column chromatography to give pure 7e, with no evidence

that hydrolysis of the pinanediol esters was occurring.

In anticipation of examining whether the addition of an

electron-donating substituent to the benzyl rings para to the

boronic acids/esters would increase the rate of oxidation of the

boronic acids/esters to phenols [35] (see Section 2.5), the synthesis

of prodrug 7d (N,N⁰-bis(2-boronic acid pinacol ester-5-methoxy

benzyl)ethylenediamine-N,N⁰-diacetic acid methyl ester) was un-

dertaken (Scheme 2). Prodrug 7d is a derivative of 7a that contains

methoxy substituents para to the boronic pinacol esters on the

benzyl rings. It was prepared by ﬁrst synthesizing 3b (N,N⁰-bis(2-

boronic acid-5-methoxy benzyl)ethylenediamine) from 2-formyl-

4-methoxyphenylboronic acid (1b) and ethylenediamine (2) in the

same manner as described for the preparation of 3a, except the

dichloromethane trituration step was omitted. 3b was condensed

with pinacol (4a) without incident, giving N,N⁰-bis(2-boronic acid

pinacol ester-5-methoxy benzyl)ethylenediamine (5b) as a white

solid. 5b was then reﬂuxed with methyl bromoacetate (6a) and

DIPEA in dry ACN overnight. The ¹H NMR spectrum of the crude

product suggested that two products were present: 7d and an

unidentiﬁable, unsymmetrical side product. Column chromatog-

raphy to purify the crude material was not attempted given the

predilection for pinacol ester hydrolysis (see above). Thus, the

synthesis of 7d must still be optimized before this double prodrug

can be investigated further.

atomic coordinates, bond lengths and angles, anisotropic

displacement parameters, hydrogen coordinates, and hydrogen

bonds are given in the S.I. (Tables S1-6).

The crystal structure of 8a is shown as a stick diagram in Fig. 1.

Two dichloromethane molecules were trapped in the crystal

structure (not shown), one on each side. As expected, no BeN co-

ordination was present since the nitrogens were protonated. Each

protonated nitrogen (N1, N1A) was hydrogen bonded to a meth-

anesulfonate counteranion at O7/O7A. Interestingly, the boronic

pinacol ester appeared to be slightly rotated out of the plane of the

phenyl ring. Additionally, while a trigonal planar boron species

would be expected to show bond angles of approximately 120^ꢀ, the

crystal structure of 8a revealed 3 different bond angles: 113.50^ꢀ

(O1eB1eO2), 126.16^ꢀ(O1eB1eC11), and 120.30^ꢀ(O2eB1eC11)

(Table S3).

Compound 9 was synthesized by hydrolyzing 8a, b, or c in H₂O

for several days and collecting the subsequently-formed precipitate

(Scheme 2). Minor impurities present in the crude precipitate (e.g.

pinacol) were removed by recrystallization from DMSO with H₂O.

Discussed in more detail in the S.I., ¹H, ¹³C, ¹¹B NMR, mass spec-

trometry, and elemental analysis provided evidence that when

precipitated from solution and when dissolved in DMSO-d₆, 9 is

likely in a coordinated form (Scheme 1). However, it is reasonable

to postulate that a hydrated, open form of 9 may also exist in

equilibrium with the coordinated form in aqueous solution

(Scheme 1).

2.2. Aqueous solubilities of 8a-c

Double prodrugs 8a-c were chosen for further investigation.

Their aqueous solubilities were determined in 100 mM pH 7.4

phosphate buffer at room temperature. A high molarity of phos-

phate buffer was necessary to maintain pH stability when the

compounds were added. Due to the inherent instability of the

prodrugs, suspensions were only allowed to stir for 30 min before

they were ﬁltered and analyzed by UV. Solubilities were calculated

using the absorbance values of the ﬁltrates and the molar absorp-

tivities of the prodrugs. The results are given in Table 1. Aqueous

solubilities ranged from 54 mM (8c) to 618 mM (8a), increasing as the

number of carbons in the carboxylate ester moieties of the pro-

drugs decreased.

While esteriﬁcation of free boronic acids allows for more

straightforward compound characterization, double prodrugs of

the general form 7 can be made with free boronic acids. For

example, N,N⁰-bis(2-boronic acid benzyl)ethylenediamine-N,N⁰-

diacetic acid tert-butyl ester (7f, Scheme 2) was prepared by ﬁrst

converting tert-butyl chloroacetate (6d) to the corresponding tert-

butyl iodoacetate using the Finkelstein reaction, followed by reac-

tion with N,N⁰-bis(2-boronic acid benzyl)ethylenediamine (3a) to

give the desired compound as the major product. However, 7f was

not investigated further because of the poor yield of the reaction

(~15%).

To improve the aqueous dissolution and ease of handling of

prodrugs of the general form 7, and as a means to purify 7a-c

without using column chromatography, 7a-c,e were converted to

their corresponding dimesylate salts, 8a-c,e, by precipitation from

ether after addition of methanesulfonic acid. Due to overly-rapid

precipitation from ether upon its formation, 8a required addi-

tional recrystallization. Although successfully synthesized, 8e was

not included in further studies because of its large size, its high

CLogP of 8.8 (ChemBioDraw Ultra 13.0, PerkinElmer), and its

anticipated poor aqueous solubility.

The crystal structure of prodrug 8a, N,N⁰-bis(2-boronic acid

pinacol ester benzyl)ethylenediamine-N,N⁰-diacetic acid methyl

ester dimesylate, was solved. Crystal data and structure reﬁnement,

Fig. 1. Crystal structure of N,N⁰-bis(2-boronic acid pinacol ester benzyl)ethylenedi-

amine-N,N⁰-diacetic acid methyl ester dimesylate (8a), shown as a stick diagram.

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

197

Table 1

Thus, the ions actually observed by the LC-MS method used and

their corresponding structures are shown in Fig. 2A. For clarity, only

one structure is given per ion, but note that for many of the ions,

multiple structures are possible depending on whether one pos-

tulates BeN coordination or B]O formation resulting from the loss

of a molecule of H₂O from a boronic acid (Fig. S5). m/z 637.38, the

parent ion, having both pinacol esters intact, was not observed.

Although m/z 555.3, having one pinacol ester intact, gave a peak

during the initial injection, it was no longer visible in the second

injection an hour later. Therefore, k₁and k₂were not able to be

determined. These ﬁndings support the rapid hydrolysis of the

boronic pinacol esters at physiological pH.

Molar absorptivities (ε), aqueous solubilities (S_AQ), and half-lives for hydrolysis (t_1/2

of prodrugs 8a-c.

)

a

Prodrug

ε₂₇₀(L mol^ꢁ1cm^ꢁ1

)

S_AQ

(

m

M)^a

t_1/2(h)^b

8a

8b

8c

1019 87

1066 80

ND^c

618 78

171 17

54 6.6^d

3.8 0.1, 0.04 0.003^e

5.9 0.3

26.3 1.0

a

Determined in pH 7.4 phosphate buffer at RT.

Determined at 37 ^ꢀC in pH 7.5 phosphate buffer containing 50% MeOH, unless

b

otherwise noted.

c

Not determined.

d

Calculated using ε₂₇₀of 1019 L mol^ꢁ1cm^ꢁ1

.

e

Determined at 23 ^ꢀC in pH 3.1 sodium citrate buffer containing 50% MeOH.

Over the course of 42 h, peaks for m/z 455.2 and 437.2 (Fig. 2A),

both dehydrated species of N,N⁰-bis(2-boronic acid benzyl)ethyl-

enediamine-N,N⁰-diacetic acid methyl ester, decreased with pseudo

ﬁrst-order kinetics (Fig. 2B). These two ions are represented by “A”

in Equation (1) (Section 4.3.1.4), and a plot of the natural log of peak

integration versus time yielded a straight line with the slope of the

linear ﬁt equal to k₃, the rate constant for the hydrolysis of the ﬁrst

carboxylate ester (Fig. 2C). The rate constant obtained from the

decay of m/z 455.2 (1.22 ꢂ 10^ꢁ3min^ꢁ1) was slightly different from

the one obtained from the decay of m/z 437.2 (1.51 ꢂ 10^ꢁ3min^ꢁ1),

so the average was taken (1.37 ꢂ 10^ꢁ3min^ꢁ1), which corresponds to

a half-life of 8.44 h.

Hydrolysis of one of the two carboxylate methyl esters of m/z

455.2 and 437.2 led to the formation of the intermediate “mono

ester” ions m/z 441.2 and 423.2, respectively (Fig. 2A). These ions

are represented by “B” in Equation (2) (Section 4.3.1.4). The

appearance of these ions, and their subsequent disappearance as

the second ester was hydrolyzed, is shown in Fig. 2D. Their t_m

values e the time necessary to reach their maximum peak areas e

were very similar (759 min for m/z 441.2 and 789 min for m/z 423.2)

and were averaged to give 774 min.

2.3. Hydrolysis of prodrugs 8a-c

As noted in Section 2.1, double prodrugs 8a-c all hydrolyzed to

give 9 (Scheme 2), the intermediate in the conversion pathway to

the active chelator HBED (Scheme 1). Experiments were then car-

ried out to determine the rates at which these hydrolyses occurred

under various conditions.

2.3.1. LC-MS-monitored hydrolysis

The hydrolysis of prodrugs 8a-c was anticipated to be more

complex than an A / B reaction because of the presence of two

hydrolyzable boronic pinacol esters and two hydrolyzable carbox-

ylic acid esters. Theoretically, four rate constants may be measured

(Scheme 3).

To observe the intermediates formed during hydrolysis, a pre-

liminary experiment was performed using LC-MS. The hydrolysis of

8a to 9 was carried out in a solution of 25% MeOH and 75% pH 7.41

N-methylmorpholine buffer using Q1MI monitoring. Although

many ions ([M þ H]^þ) were observed during compound optimi-

zation in H₂O/ACN (Fig. S5), not all of them appeared during the

actual LC-MS run, in part because the sodium atoms adducted to

some ions were most likely replaced by protons from the acidic

mobile phase. Additionally, facile dehydration of the boronic acid

moieties probably occurred to a greater extent under the much

higher temperature used during LC-MS than it did during com-

pound optimization.

The plot of the appearance of 9 (m/z 409.3) over time was

slightly s-shaped (Fig. 2E), which is expected if k₃is not consider-

ably different from k₄[50]. 9 is represented by “C” in Equation (3)

(Section 4.3.1.4). By inputting k₃(k_AB, 1.37 ꢂ 10^ꢁ3min^ꢁ1) and t_m

(774 min) into Equation (4) (Section 4.3.1.4), k₄(k_BC) was solved

using the Solver function in Microsoft Excel. A rate constant of

1.22 ꢂ 10^ꢁ3min^ꢁ1(t_1/2¼ 9.5 h) was obtained, which is very close to

Scheme 3. The stepwise hydrolysis of double prodrugs 8a-c, shown in free base form (7a-c), to generate 9. The boronic pinacol esters are expected to hydrolyze rapidly, followed

more slowly by the two carboxylate esters. Four rate constants may be measured.

198

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

Fig. 2. The hydrolysis of double prodrug 8a to 9 in a 25:75 solution of MeOH:pH 7.4 N-methylmorpholine buffer, monitored by LC-MS at room temperature. A) The ions observed

during the hydrolysis and their corresponding structures. B) Over the course of 42 h, m/z 455.2 and 437.2 decreased with pseudo ﬁrst-order kinetics. C) Plots of the natural log of

peak area vs. time for both ions yielded straight lines with the slopes of the linear ﬁts equal to k₃, the rate constant for the hydrolysis of the ﬁrst carboxylate ester. The average of

these two values (1.37 ꢂ 10^ꢁ3min^ꢁ1) corresponds to a half-life of 8.4 h. D) During the hydrolysis, the intermediate “mono ester” ions, m/z 441.2 and 423.2, were observed to form

from the hydrolysis of the ﬁrst carboxylate ester and then subsequently decay as the second carboxylate ester was hydrolyzed. E) Hydrolysis of the second carboxylate ester yielded

m/z 409.3 (9). The plot of its appearance was slightly s-shaped.

the averaged value of k₃. Therefore, it can be concluded that there is

little difference between the rates of hydrolysis of the ﬁrst and

second carboxylate esters of 8a.

HBED undergo chemical hydrolysis near physiological pH,

providing support for the ﬁrst step in the proposed activation

pathway of these prodrugs (Scheme 1). Further investigation into

the mechanism of hydrolysis of the carboxylate esters is needed.

Hydrolysis of the carboxylate esters of the double prodrugs may be

intramolecularly catalyzed by the boronic acids generated after

pinacol ester hydrolysis. While catalysis of hydrolysis of carboxylate

esters by boronic acids built in to a compound has not been pre-

viously reported in the literature, several studies have shown that

ester hydrolysis can be catalyzed by the addition of boric acid to the

system [51e53].

2.3.2. UV-monitored hydrolysis

The rates of hydrolysis of the double prodrugs 8a-c were

investigated further by UV in pH 7.5 phosphate buffer at 37 ^ꢀC. Since

8b and 8c have lower aqueous solubilities than 8a (Section 2.2), 50%

MeOH was added as a co-solvent to the phosphate buffer to ensure

rapid dissolution of all prodrugs. Spectra obtained during a typical

hydrolysis, shown for 8a, are given in Fig. 3. Over time, absorbance

at 270 and 277 nm decreased with pseudo ﬁrst-order kinetics, and

the ﬁnal spectrum matched that of an authentic sample of 9 (Fig. 3).

No distinction between the hydrolysis of the ﬁrst and second

boronic esters and carboxylate esters of 8a-c was apparent, so the

half-lives of the prodrugs (Table 1) are macro half-lives. While 8a

hydrolyzed slightly faster than 8b (t_1/2¼ 3.8 0.1 h vs. 5.9 0.3 h,

respectively), 8c was considerably slower to hydrolyze (t_1/

2.4. Relative binding afﬁnity of 9 for iron and copper

Only the phenolic hydroxyl groups of HBED remain masked, as

boronic acids, in compound 9 (Scheme 1), the intermediate formed

after hydrolysis of the bis boronic pinacol esters and bis carboxylate

esters of double prodrugs 8a-c. As such, 9 could still potentially

function as a tetradentate chelator due to the presence of its two

amino nitrogens and two carboxylate oxygens. Additionally, Swamy

et al. [54] found that the boronic acid moieties of a bis phenyl-

boronic acid-conjugated probe participated in the selective chela-

tion of copper ions. Therefore, if 9 can be present in an open

(hydrated) form in solution (Scheme 1), its boronic acids could also

serve as chelating ligands. To ascertain whether 9 was active as a

chelator, the relative afﬁnity of 9 for iron and copper was

investigated.

¼ 26.3 1.0 h), probably because its branched isopropyl esters

2

sterically hindered the attack of water. Hydrolysis was not depen-

dent upon the presence of MeOH in the solution: 8a was also found

to hydrolyze in pH 7.4 phosphate buffer without MeOH with a half-

life of 4.2 0.2 h at 23 ^ꢀC (not shown; n ¼ 2). Interestingly, when

the hydrolysis of 8a was measured at 23 ^ꢀC in pH 3.1 citrate buffer

containing 50% MeOH, it was found to be very rapid (t_1/

¼ 2.4 0.2 min). The faster rate of hydrolysis was perhaps due to

2

general acid catalysis.

Overall, these studies demonstrate that the double prodrugs of

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

199

Fig. 3. UV spectra acquired during the hydrolysis of double prodrug 8a at 37 ^ꢀC in pH

7.4 phosphate buffer containing 50% MeOH. Over the course of the hydrolysis, absor-

bance at 270/277 nm decreased (solid lines). The ﬁnal spectrum acquired matched the

spectrum of an authentic sample of 9 (dotted line).

Fig. 4. Afﬁnity of 9 and EDTA for iron in the presence of HBED. Shown are UV/vis

spectra of solutions of HBED, 9, EDTA, HBED þ 9, and HBED þ EDTA all containing

30 m

M ferric ammonium citrate (FAC) in pH 7.5 phosphate buffer at 23 ^ꢀC. An absor-

bance at 481 nm (A₄₈₁) was characteristic of HBED chelated to iron. 9 (300

mM) was

unable to compete with HBED (90 M) for binding to iron. A₄₈₁was the same

m

2.4.1. Afﬁnity for Fe^3þ

regardless of whether 9 was present in solution with HBED þ FAC. In contrast, 300

mM

EDTA decreased the amount of Fe-HBED chelate by 83%, as evidenced by a decrease in

absorbance at 481 nm in the spectrum of HBED þ EDTA þ FAC compared to the

spectrum of HBED þ FAC without EDTA.

The ability of 9, compared to that of ethylenediaminetetraacetic

acid (EDTA), to bind Fe^3þ(as ferric ammonium citrate, FAC) in the

presence of HBED was determined by UV/vis in a competition

experiment modiﬁed from Charkoudian et al. [35]. The experiment

was performed at 23 ^ꢀC in pH 7.5 phosphate buffer. Under these

conditions, a solution of Fe-HBED chelate was light pink in color

and had an absorbance at 481 nm (A₄₈₁), whereas solutions of 9 or

EDTA in the presence of iron were colorless and did not absorb at

this wavelength (Fig. 4). Therefore, A₄₈₁was monitored to detect

any change in the amount of iron chelated by HBED upon addition

of 9 or EDTA. After a 22 h equilibration period, 10 equiv of 9

decreased the absorbance at 410 nm by ~17%, 57%, and 94%,

respectively (Fig. 5). This indicates that the EDTA was chelating the

copper present, because Cu-EDTA has no absorbance at 410 nm

(Fig. 5). Additionally, the solution grew lighter in color with

increasing EDTA concentration. The solutions of 100

mM

copper þ 100

m

M EDTA with or without 300 M 9 were both faintly

m

blue in color. These results suggest that excess 9 cannot compete

with EDTA (log K ¼ 18.7 [55]) for copper chelation. Thus, the

interaction between 9 and copper, potentially involving 9's boronic

acid moieties, must be a relatively weak one.

(300

mM) was unable to compete with 3 equiv of HBED (90 mM) for

1 equiv of iron (30

m

M): A₄₈₁in the spectrum of HBED þ 9 þ FAC

was unchanged from that of HBED þ FAC without 9 (Fig. 4), and

both solutions were light pink. In contrast, 10 equiv of EDTA

Collectively, the results of these afﬁnity studies indicate that 9

has, at best, a weak afﬁnity for iron and copper, and is therefore

expected to pose little risk for off-target chelation of these metals

(Scheme 1). Thus, 9 is still a prodrug of HBED and requires further

activation before the strong chelator HBED is liberated. This ﬁnal

activation step e oxidation by H₂O₂e is discussed next.

(300 mM) was able to reduce the amount of Fe-HBED chelate by

83%, as evidenced by the decrease in A₄₈₁from 0.088 in the spec-

trum of HBED þ FAC without EDTA to 0.015 in the spectrum of

HBED þ FAC with EDTA (Fig. 4), and by the change in color of the

solution from light pink to almost colorless. Since 9, even in excess,

failed to compete with HBED for iron, 9 either has little-to-no af-

ﬁnity for iron or at least is a much weaker chelator than EDTA (log

K ¼ 25.0 [55]).

2.5. LC-MS monitored oxidation of 9 by H₂O₂

The reaction scheme for the oxidation of prodrug 9 to HBED

under pseudo ﬁrst-order conditions of excess H₂O₂at ambient

temperature is shown in Fig. 6A. Reactions were monitored by LC-

MS, which allowed for the formation and disappearance of the

“mono” intermediate, N,N⁰-bis(2-boronic acid, 2⁰-hydroxy benzyl)

ethylenediamine-N,N⁰-diacetic acid, to be observed in this

A / B / C series reaction. The rate of oxidation was determined in

pH 7.4 NMM buffer using 5e59ꢂ excess H₂O₂.

2.4.2. Afﬁnity for Cu^2þ

The relative afﬁnity of 9 for copper was determined by a UV/vis

competition experiment with EDTA at 23 ^ꢀC in pH 7.4 phosphate

buffer containing 25% MeOH. When

a solution of 300 mM

9 þ 100

m

M copper (as copper(II) sulfate pentahydrate) was pre-

pared, it was light blue-green in color. The spectrum of this solution

had an absorbance at 410 nm (Fig. 5). These observations suggest

that 9 has some afﬁnity for or interaction with copper. However, the

A representative plot of peak area of m/z 409.3 (9), 399.4 (mono

intermediate), and 389.3 (HBED) vs. time for an oxidation reaction

addition of 20, 60, and 100

m

M EDTA to the solution of 9 þ copper

200

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

Fig. 5. Afﬁnity of 9 for copper in the presence of various concentrations of EDTA. Shown are UV/vis spectra of 300

in pH 7.5 phosphate buffer containing 25% MeOH at 23 ^ꢀC with 100

M copper (II) sulfate pentahydrate. The absorbance at 410 nm in the spectrum of 9 þ Cu suggests that 9 weakly

interacts with this metal. However, addition of 20, 60, and 100 M EDTA to the solution decreased the absorbance at 410 nm by ~17%, 57%, and 94%, respectively, which indicates

m

M 9, 100

m

M EDTA, or 300

m

M 9 þ EDTA (various concentrations)

m

that 9, even in excess, cannot compete with EDTA for copper chelation.

using 25ꢂ excess H₂O₂is given in Fig. 6B. One boron of m/z 409.3 (9)

was quickly oxidized to the corresponding phenol to generate the

mono intermediate m/z 399.4. The intermediate m/z 399.4 then

slowly disappeared as the second boron was oxidized to generate

m/z 389.3, HBED. Plots of ln peak area vs. time for m/z 409.3 (9) and

389.3 (HBED) using various concentrations of H₂O₂were all linear.

The observed pseudo ﬁrst-order rate constants for the oxidation of

9 to the mono intermediate (k_AB) and for the formation of HBED

from the mono intermediate (k_BC)were linearly dependent on the

concentration of H₂O₂, and the second-order rate constants (k)

obtained from the slopes of these straight lines were

30.8 M^ꢁ1min^ꢁ1(r²¼ 0.99) and 10.3 M^ꢁ1min^ꢁ1(r²¼ 0.98),

respectively.

Additionally, k_BCwas measured from the decay of the mono

intermediate (rather than from the formation of HBED) at different

H₂O₂concentrations. This was done by plotting ln peak area of m/z

399.4 vs. time, beginning at the time point where the peak of m/z

409.3 (9) had disappeared to the extent that it was no longer able to

be integrated. Interestingly, k obtained from the slope of this linear

ﬁt of k_BCvs. H₂O₂concentration was 6.3 M^ꢁ1min^ꢁ1(r²¼ 0.95). In

theory, the rate constant for the disappearance of the mono inter-

mediate (B) should equal the rate constant for the formation of

HBED (C) in an A / B / C reaction. The fact that HBED was formed

faster than the mono intermediate was oxidized suggests that

either some of A (m/z 409.3, 9) was oxidizing directly to C (HBED),

or that there was another intermediate formed from the oxidation

of 9 that was not noticed during LC-MS optimization.

The above set of experiments to determine the second-order

rate constant for the oxidation of 9 in pH 7.4 NMM buffer were

confounded by the fact that H₂O₂can oxidize NMM to the corre-

sponding N-oxide [56]. NMM N-oxide may then, in turn, oxidize 9

[57], possibly at a different rate than oxidation by H₂O₂. As noted in

Section 4.5.2, the stock solution of H₂O₂used in the experiments

was prepared in NMM buffer.

In order to determine whether the potential reaction between

NMM and H₂O₂signiﬁcantly affected rate constant determination

for the oxidation of 9, a brief LC-MS experiment was conducted in a

manner similar to that described above, but replacing NMM buffer

with pH 7.4 phosphate buffer for both the reaction buffer and the

H₂O₂stock buffer. Only single or duplicate samples were run for

each H₂O₂concentration. Due to ion suppression, m/z 409.3 (9)

peaks were difﬁcult to integrate past 2e4 time points, so the

second-order rate constant could not be accurately determined for

the oxidation of 9 to the mono intermediate. The second-order rate

constants for the disappearance of m/z 399.4 (mono intermediate)

and formation of m/z 389.3 (HBED), respectively, were

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

201

needed to curtail the Fenton reaction. Off-target chelation should

be minimized by this site speciﬁcity. Additional studies are un-

derway to investigate whether the prodrugs can provide protection

to cells exposed to H₂O₂.

4. Experimental

All solvents were purchased from Fisher Scientiﬁc or Sigma

Aldrich. Reaction solvents were ACS certiﬁed and LC-MS solvents

were Optima grade. Solvents noted as “dry” were obtained

following storage over 3 Å molecular sieves (Acros Organics). LC-

MS solvents were ﬁltered using Magna 0.22 micron nylon mem-

brane ﬁlters (GVS Maine). Syringe ﬁltration was performed using

Magna 0.45 micron nylon membrane ﬁlters (GVS Maine). Ammo-

nium iron(III) citrate, copper(II) sulfate pentahydrate (99þ%),

hydrogen peroxide (50 wt% solution in water, stabilized), methyl

bromoacetate (99%), pinacol (99%), and tert-butyl chloroacetate

(97%) were purchased from Acros Organics. Diisopropylethylamine

(peptide synthesis grade), ethylenediamine (anhydrous), formic

acid (Optima LC/MS), and hydrochloric acid (TraceMetal grade)

were purchased from Fisher Scientiﬁc. Ethyl bromoacetate (98%),

ethylenediaminetetraacetic acid disodium dihydrate, 2-formyl-4-

methoxyphenylboronic acid (97%), isopropyl bromoacetate (99%),

methanesulfonic acid (99%), N-methylmorpholine (99.5%),

(1R,2R,3S,5R)-(-)-pinanediol (99%), sodium borohydride (ꢃ98%),

and sodium iodide (ReagentPlus, ꢃ99%) were purchased from

Sigma Aldrich. 2-formylphenylboronic acid (97%) was purchased

from Matrix Scientiﬁc. The mono-hydrochloride salt of HBED was

generously donated by PPC ADOB (Poland) and recrystallized from

boiling 85% EtOH. Reactions were monitored by thin-layer chro-

matography (Whatman UV₂₅₄aluminum-backed silica gel). NMR

spectra were recorded on a Varian Mercury-Plus 400 MHz spec-

trometer at ambient temperature, a Bruker Avance III 500 MHz

spectrometer at 27 ^ꢀC, and a Bruker Avance III 600 MHz spec-

trometer at 27 ^ꢀC. The chemical shifts of ¹H and ¹³C spectra are

expressed in ppm and J values in Hz. CDCl₃and DMSO-d₆spectra

were referenced to the TMS internal standard, while CD₃OD spectra

were referenced to the solvent residual peak at 3.31 ppm. The

splitting of proton resonances in the reported ¹H spectra is deﬁned

as: s ¼ singlet, d ¼ doublet, t ¼ triplet, sep ¼ septet, dd ¼ doublet of

doublets, m ¼ multiplet, br ¼ broad, app ¼ apparent. The ¹¹B NMR

spectra were referenced to an external standard of boron triﬂuoride

diethyl etherate (0.0 ppm) in CDCl₃and chemical shifts were re-

ported in ppm. Background from the borosilicate glass in the probe

and NMR tube was minimized by subtracting the spectrum of the

empty NMR tube from its corresponding spectrum with sample

added. Melting points (MP) were determined on a Mel-temp

melting point apparatus. Elemental analysis (EA) was performed

by Atlantic Microlab, Inc. (Norcross, GA). UV/Visible spectra were

recorded on two Shimadzu spectrophotometers, models 2550

(equipped with a CPS controller) and 2700 (equipped with a TCC

controller), using 1-cm quartz cuvettes. Mass spectra were recor-

ded on a 3200 Qtrap LC/MS/MS system (AB Sciex, USA) in electro-

spray ionization (ESI) mode. High-resolution mass spectra were

recorded on an Orbitrap XL in positive ESI mode (Thermo Finnigan,

USA). LC-MS was performed using a Prominence SIL-20 autosam-

pler (Shimadzu, Japan), LC-20AD pumps (Shimadzu, Japan), and a

Prominence SPD-M20A photo diode array detector (Shimadzu,

Japan), connected to the 3200 Qtrap LC/MS/MS system in ESI mode

with positive polarity.

Fig. 6. LC-MS-monitored oxidation of prodrug 9 by 25ꢂ excess H₂O₂at room tem-

perature in pH 7.4 NMM buffer. A) The reaction scheme for the oxidation of 9 (A) to

HBED (C) via a mono-oxidized intermediate (B). Structures are shown as [MþH]^þions.

B) A representative plot of peak area of m/z 409.3, 399.4, and 389.3 vs. time. As m/z

409.3 (9) disappears, a mono-oxidized intermediate (m/z 399.4) forms and then is

further oxidized to m/z 389.3 (HBED).

7.6 M^ꢁ1min^ꢁ1(r²¼ 0.95) and 9.7 M^ꢁ1min^ꢁ1(r²¼ 0.92), which

were similar to the respective rate constants of 6.3 M^ꢁ1min^ꢁ1and

10.3 M^ꢁ1min^ꢁ1that were obtained when oxidation was carried out

in NMM buffer. Therefore, it was concluded that any reaction be-

tween NMM and H₂O₂minimally affected the oxidation rates.

These experiments indicate that after the double prodrugs of the

general form 7 hydrolyze to prodrug 9, 9 can then undergo oxida-

tion by H₂O₂to give HBED (Scheme 1). This required second step in

the activation pathway of the double prodrugs before an active

chelator is released should not only prevent indiscriminant metal

sequestration until full activation is achieved, but may also provide

the prodrugs with site-speciﬁcity. In a highly-oxidative environ-

ment, such as in tissues where excess labile iron is contributing to

free radical production, 9 may be preferentially converted to HBED.

HBED, in turn, should prevent further Fenton chemistry by scav-

enging the harmful iron.

3. Conclusions

Novel double prodrugs have been synthesized and were found

to convert by stepwise chemical hydrolysis and oxidation to the

strong iron chelator HBED. Hence, proof of principle was estab-

lished for their activation. The ability of the carboxylate esters of

these double prodrugs to chemically hydrolyze may be attributed

to the presence of hydrolyzable boronic esters that mask the

phenolic hydroxyl groups of the parent compound. This is an

advantage over previously-reported HBED prodrugs containing

carboxylate ester promoieties that were difﬁcult to hydrolyze. The

necessity for a second step e oxidation e to occur in the activation

pathway of the double prodrugs before a strong chelator is released

is also advantageous. The prodrugs should fully unmask preferen-

tially at sites of high oxidative stress, where chelation is most

4.1. Synthesis and characterization

4.1.1. N,N⁰-bis(2-boronic acid benzyl)ethylenediamine (3a)

Procedure modiﬁed from Gray and Houston [43] and Gray et al.

202

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

[44]. To a stirred solution of ethylenediamine (2, 2.31 g, 0.038 mol)

in 250 mL dry MeOH at room temperature was added 2-

formylphenylboronic acid (1, 11.51 g, 0.077 mol), which gave a

bright-yellow solution containing minimal light-colored precipi-

tate. After 2 h, 3 equiv sodium borohydride (4.36 g, 0.12 mol) was

added in 3 batches, 15 min apart, to the reaction ﬂask immersed in

an ice bath. The reaction ﬂask was removed from the ice bath

10 min after the ﬁnal sodium borohydride addition and was stirred

for an additional 50 min at room temperature. The faintly-yellow

reaction mixture containing ﬁne, white precipitate was concen-

trated on the rotary evaporator (35 ^ꢀC) and then further dried in

vacuo to a cream-colored solid.

To this solid was added 450 mL DCM, and the mixture was

stirred vigorously for 1 h before the boron salts were removed by

vacuum ﬁltration. The ﬁltered white solid was washed with 100 mL

DCM and discarded after ¹H NMR in CD₃OD conﬁrmed that it did

not contain a signiﬁcant amount of 3a. The slightly yellow ﬁltrate

was concentrated via rotary evaporator (35 ^ꢀC) and in vacuo to a

cream-colored solid.

used in the next step without further puriﬁcation. ¹H (400 MHz,

CDCl₃):

d

7.63 (d, 2 H, J ¼ 8.16, PheH), 6.80 (dd, 2 H, J ¼ 8.1, 2.4,

PheH), 6.70 (d, 2 H, J ¼ 2.40, PheH), 3.93 (s, 4 H, PheCH₂eNHR),

3.80 (s, 6 H, PheOCH₃), 2.77 (s, 4 H, NHReCH₂eCH₂), 1.29 (s, 24 H,

pinacol CH₃). ¹³C (APT, 100 MHz, CDCl₃):

d 160.95 (Ph quaternary),

145.45 (Ph quaternary), 135.41 (Ph), 112.53 (Ph), 112.24 (Ph), 81.84

(BOeCeR₃), 55.29 (OCH₃), 52.70 (CH₂), 45.72 (CH₂), 25.35 (pinacol

CH₃). MP: 167e170 ^ꢀC.

4.1.5. N,N⁰-bis(2-boronic acid pinanediol ester benzyl)

ethylenediamine (5c)

A similar procedure to the one described above for 5a was fol-

lowed, but replacing pinacol (4a) with (1R,2R,3S,5R)-(-)-pinanediol

(4b). A white, crystalline solid was isolated and used in the next

step without further puriﬁcation. ¹H (400 MHz, CDCl₃):

d 7.71 (d,

2 H, J ¼ 7.1), 7.14e7.31 (m, with toluene impurity), 4.30 (d, 2 H,

J ¼ 8.1), 4.01 (d, 2 H, J ¼ 13.7), 3.87 (d, 2 H, J ¼ 13.7), 2.69e2.84 (m,

4 H), 2.38 (m, with toluene impurity), 2.22 (m, 2 H), 2.09 (t, 2 H,

J ¼ 5.5), 1.95 (m, 4 H), 1.51 (d, 2 H, J ¼ 10.5), 1.45 (s, 6 H), 1.30 (s, 6 H),

0.88 (s, 6 H).

To this solid was added 195 mL 1 M NaOH. After vigorous

swirling, most of the solid dissolved. Further precipitation of white

solid occurred when the ice-cooled mixture was adjusted to a ﬁnal

pH of 8.29 with concentrated HCl. The suspension was vacuum

ﬁltered and washed with 4 ꢂ 20 mL cold H₂O. The white, powdery

solid was dried in vacuo at room temperature (2 days) and at 70 ^ꢀC

(12 h) to give 10.34 g of the title compound 3a in 87% yield. ¹H

4.1.6. General procedure for preparation of N,N⁰-bis(2-boronic acid

pinacol ester benzyl)ethylenediamine-N,N⁰-diacetic acid esters (7a-

c)

Into a round-bottom ﬂask equipped with a stir bar was added 5a

from the previous step (0.0032 mol, 1.0 equiv) and 100 mL ACN,

followed by 1.16 g (0.0090 mol, 2.79 equiv) DIPEA and another

50 mL dry ACN. The ﬂask containing an off-white suspension was

then equipped with a condenser and drying tube, and heated to

82 ^ꢀC. Next, 2.03 equiv. of methyl bromoacetate (6a), ethyl bro-

moacetate (6b), or isopropyl bromoacetate (6c) was quickly added,

and the off-white suspension was reﬂuxed for 20.5 h. After cooling

for 1 h, the goldehued mixture containing ﬁne, light-colored pre-

cipitate was concentrated at 40 ^ꢀC on the rotary evaporator to a soft

solid, which was further dried in vacuo. This crude product was

then extracted with 250 mL dry diethyl ether for 2 h, followed by

vacuum ﬁltration. The faintly peach-colored, crystalline solid was

washed with 50 mL dry ether and then discarded. The pale-yellow

ﬁltrate was concentrated at 35 ^ꢀC on the rotary evaporator and then

further dried in vacuo to yield an oil, which was used in the next

step without further puriﬁcation.

(500 MHz, CD₃OD):

PheH), 4.10 (s, 4 H, PheCH₂eNHR), 3.22 (s, 4 H, RNHeCH₂eCH₂).

¹³C (APT, 100 MHz, CD₃OD):

142.77 (Ph quaternary), 131.47 (Ph),

128.49 (Ph), 127.65 (Ph), 123.86 (Ph), 55.10 (PheCH₂eNHR), 45.94

d 7.42e7.44 (m, 2 H, PheH), 7.17e7.24 (m, 6 H,

d

(RNHeCH₂eCH₂). ¹¹B (161 MHz, CD₃OD):

d

9.84. MP: 145e150 ^ꢀC

(dehydration), 212e218 ^ꢀC (melted). EA Found: C, 63.13, 63.25; H,

6.49, 6.34; N, 9.21, 9.15. Calc. for C₁₆H₂₂B₂N₂O₄: C, 58.59; H, 6.76; N,

8.54. MS (m/z): 311.2 (M þ HeH₂O; Calc: 311.17).

4.1.2. N,N⁰-bis(2-boronic acid-5-methoxy benzyl)ethylenediamine

(3b)

A similar procedure to the one described above for 3a was fol-

lowed, but replacing 2-formylphenylboronic acid (1a) with

2-formyl-4-methoxyphenylboronic acid (1b) and omitting the

dichloromethane trituration step. A pale-tan powder was isolated in

59% yield. ¹H (400 MHz, CD₃OD):

d

7.32 (d, 2 H, J ¼ 8.0, PheH), 6.81

(dd, 2 H, J ¼ 8.0, 2.2, PheH), 6.77 (d, 2 H, J ¼ 2.2, PheH), 4.07 (s, 4 H,

PheCH₂eNHR), 3.78 (s, 6 H, OCH₃), 3.20 (s, 4 H, RNHeCH₂eCH₂).

4.1.6.1. N,N⁰-bis(2-boronic acid pinacol ester benzyl)ethylenedi-

amine-N,N⁰-diacetic acid methyl ester (7a). Yellow-orange oil con-

taining a small amount of light-colored precipitate. ¹H (400 MHz,

4.1.3. N,N⁰-bis(2-boronic acid pinacol ester benzyl)ethylenediamine

(5a)

CDCl₃):

d

7.68 (d, 2 H, J ¼ 7.2, PheH), 7.30e7.38 (m, 4 H, PheH), 7.20

In a round-bottom ﬂask equipped with a stir bar was added 3a

(1.00 g, 0.0032 mol), pinacol (4a, 0.77 g, 0.0065 mol), and 80 mL

toluene. A Dean-Stark trap was attached and ﬁlled with toluene.

The white suspension was reﬂuxed for 22 h, cooled for 1 h, and then

concentrated at 55 ^ꢀC on the rotary evaporator. The off-white solid

(5a) was further dried in vacuo and used in the next step without

(dt, 2 H, J ¼ 7.3, 1.2, PheH), 3.94 (s, 4 H, PheCH₂eNR₂), 3.61 (s, 6 H,

OCH₃), 3.36 (s, 4H, R₂NeCH₂eCOOR), 2.72 (s, 4 H, R₂NeCH₂e-

CH₂eNR₂), 1.31 (s, 24 H, CH₃pinacol). ¹³C (APT, 100 MHz, CDCl₃):

d

172.33 (C]O), 145.44 (Ph quaternary), 135.32 (Ph), 130.43 (Ph),

129.36 (Ph), 126.33 (Ph), 83.58 (BOeCeR₃), 58.15 (PheCH₂eNR₂),

54.37 (R₂NeCH₂eCOOR), 51.18 (R₂NeCH₂eCH₂eNR₂), 51.12

(COOeCH₃), 25.05 (CH₃pinacol). ¹¹B NMR (193 MHz, CDCl₃):

further puriﬁcation. ¹H (400 MHz, CDCl₃):

PheH), 7.23e7.30 (m, PheH), 7.13 (d, 2 H, J ¼ 7.2, PheH), 3.96 (s, 4 H,

PheCH₂eNHR), 2.77 (s, 4 H, NHReCH₂eCH₂), 1.31 (s, 24 H, CH₃). ¹³

(APT,100 MHz, CDCl₃): 144.48 (Ph quaternary),134.90 (Ph),130.51

(Ph), 128.39 (Ph), 127.32 (Ph), 83.15 (BOeCeR₃),

d

7.685 (br d, 2 H, J ¼ 5.8,

d

31.48. MS (m/z): 637.3 (M þ H; Calc: 637.38).

C

d

4.1.6.2. N,N⁰-bis(2-boronic acid pinacol ester benzyl)ethylenedi-

amine-N,N⁰-diacetic acid ethyl ester (7b). Yellow-orange oil con-

taining a small amount of light-colored precipitate. ¹H (400 MHz,

53.97(PheCH₂eNHR), 46.85 (RNHeCH₂eCH₂), 26.70 (CH₃). MP:

211.5e217.5 ^ꢀC. MS (m/z): 493.6 (M þ H; Calc: 492.34).

CDCl₃):

d

7.68 (d, 2 H, J ¼ 7.3, PheH), 7.36 (d, 2 H, J ¼ 7.5, PheH), 7.31

4.1.4. N,N⁰-bis(2-boronic acid pinacol ester-5-methoxy benzyl)

ethylenediamine (5b)

A similar procedure to the one described above for 5a was fol-

lowed, but replacing 3a with 3b. A white solid was isolated and

(t, 2 H, J ¼ 7.4 Hz, PheH), 7.20 (t, 2 H, J ¼ 7.2, PheH), 4.08 (q, 4 H,

J ¼ 7.1, OeCH₂eCH₃), 3.95 (s, 4 H, PheCH₂eNR₂), 3.34 (s, 4 H,

R₂NeCH₂eCOOR), 2.73 (s, 4 H, R₂NeCH₂eCH₂eNR₂), 1.31 (s, 24 H,

CH₃pinacol), 1.22 (t, 6 H, J ¼ 7.1, OeCH₂eCH₃).

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

203

4.1.6.3. N,N⁰-bis(2-boronic acid pinacol ester benzyl)ethylenedi-

amine-N,N⁰-diacetic acid isopropyl ester (7c). Yellow-orange oil. ¹H

(400 MHz, CDCl₃):

3.26 (s, 4 H), 3.01 (s, 4 H), 1.46 (s, 18 H).

d

7.68 (d, 2 H, J ¼ 7.1, PheH), 7.37 (d, 2 H, J ¼ 7.4,

4.1.10. General procedure for conversion of compounds 7a-c,e to

dimesylate salts (8a-c,e)

To a stirring solution of 7a-c,e (1.0 equiv) in dry diethyl ether at

room temperature was added methanesulfonic acid (1.9 equiv)

dropwise over ~5 min. After 2 h, the precipitate was collected via

vacuum ﬁltration, washed with dry ether, and dried in vacuo.

PheH), 7.31 (t, 2 H, J ¼ 7.3, PheH), 7.19 (t, 2 H, J ¼ 7.1, PheH), 4.98

(sep, 2 H, J ¼ 6.2, OeCHe(CH₃)₂), 3.96 (s, 4 H, PheCH₂eNR₂), 3.31 (s,

4 H, R₂NeCH₂eCOOR), 2.73 (s, 4 H, R₂NeCH₂eCH₂eNR₂), 1.30 (s,

24 H, CH₃pinacol), 1.20 (d, 12 H, J ¼ 6.2, OeCH₂e(CH₃)₂). ¹³C (APT,

100 MHz, CDCl₃):

d 171.40 (C]O), 145.65 (Ph quaternary), 135.30

(Ph), 130.37 (Ph), 129.19 (Ph), 126.19 (Ph), 83.50 (BOeCeR₃), 67.43

(COOeCH), 58.09 (PheCH₂eNR₂), 54.85 (R₂NeCH₂eCOOR), 51.18

(R₂NeCH₂eCH₂eNR₂), 25.01 (CH₃pinacol), 22.04 (OeCH₂e(CH₃)₂).

4.1.10.1. N,N⁰-bis(2-boronic acid pinacol ester benzyl)ethylenedi-

amine-N,N⁰-diacetic

acid

Recrystallization with dry diethyl ether from dry DCM gave a white

solid (59% yield over 3 steps). ¹H NMR (500 MHz, CDCl₃):

7.92 (dd,

methyl

ester

dimesylate

(8a).

¹¹B NMR (193 MHz, CDCl₃):

693.45).

d

31.58. MS (m/z): 693.3 (M þ H; Calc:

d

2 H, J ¼ 7.4, 1.2, PheH), 7.77 (d, 2 H, J ¼ 7.6, PheH), 7.52 (dt, 2 H,

J ¼ 7.6,1.5, PheH), 7.43 (dt, 2 H, J ¼ 7.4, 0.9, PheH), 4.90 (s, 4 H, CH₂),

4.22 (s, 4 H, CH₂), 4.11 (s, 4 H, CH₂), 3.66 (s, 6 H, COOCH₃), 2.79

(s, 6H, SO₃CH₃), 1.41 (s, 24 H, CH₃pinacol). ¹³C NMR (APT, 100 MHz,

4.1.7. N,N⁰-bis(2-boronic acid pinacol ester-5-methoxy benzyl)

ethylenediamine-N,N⁰-diacetic acid methyl ester (7d)

A similar procedure to the one described above for 7a was fol-

lowed, but replacing N,N⁰-bis(2-boronic acid pinacol ester benzyl)

ethylenediamine (5a) with N,N⁰-di-(2-boronic acid pinacol ester-5-

methoxy benzyl)ethylenediamine (5b). A yellow-orange oil con-

taining light-colored precipitate was obtained. ¹H NMR suggested it

contained two products, one of which was the title compound. ¹H

CDCl₃): d 166.24 (C]O), 137.08 (Ph), 134.44 (Ph quaternary), 132.69

(Ph), 132.25 (Ph), 129.53 (Ph), 84.96 (BOeCeR₃), 59.42 (CH₂), 53.08

(COOeCH₃), 52.54 (CH₂), 50.33 (CH₂), 39.30 (SO₃CH₃), 24.83 (CH₃

pinacol). ¹¹B NMR (161 MHz, CDCl₃):

d

30.39. MP: 158.5e160.5 ^ꢀC.

EA Found: C, 51.88; H, 6.90; N, 3.48. Calc for C₃₆H₅₈B₂N₂O₁₄S₂: C,

52.18, H, 7.06, N, 3.38. MS (m/z): 637.4 (M þ H, minus 2 SO₃CH₃;

Calc: 637.38.

(400 MHz, CDCl₃):

d

7.65 (d, 2 H, J ¼ 8.2, PheH), 7.04 (d, 2 H, J ¼ 2.6,

PheH), 6.72 (dd, 2 H, J ¼ 8.2, 2.6), 3.96 (s, 4 H, CH₂), 3.77 (s, 6 H,

OCH₃), 3.63 (s, 6 H, OCH₃), 3.39 (s, 4 H, CH₂), 2.77 (s, 4 H, CH₂), 1.30

(s, 24 H, pinacol CH₃).

4.1.10.2. N,N⁰-bis(2-boronic acid pinacol ester benzyl)ethylenedi-

amine-N,N⁰-diacetic acid ethyl ester dimesylate (8b). Off-white solid

4.1.8. N,N⁰-bis(2-boronic acid pinanediol ester benzyl)

ethylenediamine-N,N⁰-diacetic acid methyl ester (7e)

(77% yield over 3 steps). ¹H NMR (500 MHz, CDCl₃):

d 7.92 (dd, 2 H,

J ¼ 7.4, 1.1, PheH), 7.79 (d, 2 H, J ¼ 7.6, PheH), 7.52 (dt, 2 H, J ¼ 7.6,

1.4, PheH), 7.43 (dt, 2 H, J ¼ 7.4, 1.0, PheH), 4.92 (s, 4 H, CH₂), 4.22

(s, 4 H, CH₂), 4.15 (s, 4 H, CH₂), 4.08 (q, 4 H, J ¼ 7.1, OeCH₂eCH₃),

2.79 (s, 6 H, SO₃CH₃), 1.41 (s, 24 H, CH₃pinacol), 1.19 (t, 6 H, J ¼ 7.1,

A similar procedure to the one described above for 7a was fol-

lowed, with some modiﬁcations. Methyl bromoacetate (6a) was

ﬁrst converted to methyl iodoacetate by reacting 6a (0.0037 mol,

2.2 equiv) with sodium iodide (0.0036 mol, 2.11 equiv) overnight at

room temperature in 45 mL dry ACN. DIPEA (0.0049 mol, 2.91

equiv) and N,N⁰-di-(2-boronic acid pinanediol ester benzyl)ethyl-

enediamine (5c, 0.0017 mol, 1.0 equiv) were then added, and the

mixture was reﬂuxed overnight. The reaction mixture was

concentrated and extracted with diethyl ether to give a yellow-

brown oil. This material was puriﬁed by ﬂash column chromatog-

raphy (silica gel, 0e30% gradient of ethyl acetate in hexanes) to give

the title compound as a yellow oil in 66% yield. ¹H (400 MHz,

OeCH₂eCH₃). ¹³C NMR (APT, 100 MHz, CDCl₃):

d 165.77 (C]O),

137.07 (Ph), 134.46 (Ph quaternary), 132.81 (Ph), 132.22 (Ph), 129.51

(Ph), 84.96 (BOeCeR₃), 62.63 (COOeCH₂eCH₃), 59.43 (CH₂), 52.66

(CH₂), 50.40 (CH₂), 39.32 (SO₃CH₃), 24.83 (CH₃pinacol), 13.87

(COOeCH₂eCH₃). ¹¹B NMR (161 MHz, CDCl₃):

d

31.35. MP:

140e143 ^ꢀC. EA Found: C, 53.08; H, 7.40; N, 3.26. Calc. for

C

₃₈H₆₂B₂N₂O₁₄S₂þ 0.14C₄H₁₀O (ether impurity): C, 53.42; H, 7.37;

N, 3.23. MS (m/z): 665.5 (M þ H, minus 2 SO₃CH₃; Calc: 665.41.

CDCl₃):

d

7.67 (d, 2 H, J ¼ 7.3), 7.32 (m, 4 H), 7.21 (m, 2 H), 4.34 (d,

4.1.10.3. N,N⁰-bis(2-boronic acid pinacol ester benzyl)ethylenedi-

amine-N,N⁰-diacetic acid isopropyl ester dimesylate (8c). White solid

2 H, J ¼ 8.3), 3.98 (d, 2 H, J ¼ 13.5), 3.91 (d, 2 H, J ¼ 13.5), 3.62 (s, 6 H),

3.35 (s, 4 H), 2.75 (m, 4 H), 2.37 (m, 2 H), 2.21 (m, 2 H), 2.09 (m, 2 H),

1.91 (m, 4 H), 1.42 (s, 6 H), 1.30 (m, 8 H), 0.88 (s, 6 H).

(81% yield over 3 steps). ¹H (400 MHz, CDCl₃):

d

7.93 (dd, 2 H, J ¼ 7.3,

0.9, PheH), 7.81 (d, 2 H, J ¼ 7.6, PheH), 7.52 (dt, 2 H, J ¼ 7.5, 1.1,

PheH), 7.43 (t, 2 H, J ¼ 7.2, PheH), 4.96 (s, 4 H, CH₂), 4.89 (sep, 2 H,

J ¼ 6.2, OeCHe(CH₃)₂), 4.20 (s, 8 H, CH₂), 2.81 (s, 6 H, SO₃CH₃), 1.42

(s, 24 H, CH₃pinacol),1.16 (d,12 H, J ¼ 6.2, OeCH₂e(CH₃)₂). ¹³C NMR

4.1.9. N,N⁰-bis(2-boronic acid benzyl)ethylenediamine-N,N⁰-

diacetic acid tert-butyl ester (7f)

Into a round-bottom ﬂask equipped with a stir bar was added

tert-butyl chloroacetate (6d, 0.002 mol, 2 equiv) in 10 mL dry ACN,

followed by NaI (0.002 mol, 2 equiv). The pale-yellow suspension

containing white precipitate was protected from light and stirred at

room temperature for 3.5 h. Next, to this suspension was added

DIPEA (0.0029 mol, 2.86 equiv), followed 10 min later by N,N⁰-

bis(2-boronic acid benzyl)ethylenediamine (3a, 0.00097 mol, 0.95

equiv). The ﬂask containing an off-white suspension was then

equipped with a condenser and drying tube, and reﬂuxed for 18 h.

After cooling for 1 h, the suspension was concentrated at 40 ^ꢀC on

the rotary evaporator to a cream solid, which was further dried in

vacuo. This crude product was then extracted with 50 mL dry

diethyl ether for 2 h, followed by vacuum ﬁltration. The ﬁltrate was

concentrated at 35 ^ꢀC on the rotary evaporator and then further

dried in vacuo to give the title compound as a pale-yellow solid

(lightweight crystalline shards) in approximately 15% yield. ¹H

(APT, 100 MHz, CDCl₃): d 165.30 (C]O), 137.11 (Ph), 134.50 (Ph

quaternary),132.86 (Ph),132.18 (Ph),129.48 (Ph), 84.95 (BOeCeR₃),

71.04 (COOeCH), 59.37 (CH₂), 52.80 (CH₂), 50.46 (CH₂), 39.34

SO3CH3), 24.83 (CH₃pinacol), 21.50 (OeCH₂-(CH₃)₂).l), 21.50

(OeCH₂e(CH₃)₂). ¹¹B NMR (161 MHz, CDCl₃):

d

31.6. MP: 83e94 ^ꢀC.

EA Found: C, 54.11; H, 7.66; N, 3.18. Calc. for C₄₀H₆₆B₂N₂O₁₄S₂: C,

54.30; H, 7.52; N, 3.17. HRMS (m/z): 693.3946 (M þ H, minus 2

SO₃CH₃; Calc: 693.4452).

4.1.10.4. N,N⁰-bis(2-boronic acid pinanediol ester benzyl)ethylenedi-

amine-N,N⁰-diacetic acid methyl ester dimesylate (8e). White solid

(72% yield). ¹H (400 MHz, CDCl₃):

d

7.94 (d, 2 H, J ¼ 7.4), 7.68 (d, 2 H,

J ¼ 7.6), 7.52 (t, 2 H, J ¼ 7.6), 7.44 (t, 2 H, J ¼ 7.6), 4.98 (d, 2 H,

J ¼ 12.6), 4.86 (d, 2 H, J ¼ 12.6), 4.63 (d, 2 H, J ¼ 8.4), 4.09e4.24 (m,

8 H), 3.69 (s, 6 H), 2.78 (s, 6 H), 2.42 (m, 2 H), 2.14e2.30 (m, 4 H),

2.04 (d, 2 H, J ¼ 15), 1.96 (br s, 2 H), 1.56 (s, 6 H), 1.32 (s, 6 H), 1.18 (d,

2 H, J ¼ 11.1), 0.90 (s, 6 H).

(400 MHz, MeOD, major product):

d 7.1e7.46 (m, 8 H), 3.94 (s, 4 H),

204

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

4.1.11. General procedure for preparation of 4,4'-(ethane-1,2-diyl)

bis(10-hydroxy-3,4,5,10-tetrahydro-2H-4

⁴,10 ⁴-benzo[3,4][1,2]

azaborolo[2,1-b][1,3,2]oxazaborol-2-one) (9)

To 8a-c was added H₂O, and the resulting acidic solution (pH

3e4 by litmus) was allowed to stir at room temperature. White

solid began to precipitate within a few hours. After several days, the

precipitate was collected, washed with cold water, dried in vacuo,

and then recrystallized from DMSO with water to yield a snow-

white solid (dried in vacuo). The yield for the synthesis of the title

compound from 8c was 46% after subtraction of DMSO impurity. ¹H

added to 2970 mL phosphate buffer and stirred for 30 min. The

l

ensuing solution was analyzed by UV spectrophotometry. With

absorbance (A), concentration (c) and path length known, ε₂₇₀

(L mol^ꢁ1cm^ꢁ1) was calculated using Beer's law:

A₂₇₀¼ c*ε₂₇₀ꢅ[

The aqueous solubilities of double prodrugs 8a-c were deter-

mined in triplicate in 100 mM pH 7.4 phosphate buffer. Double

prodrug (10e11 mg) was stirred at room temperature in 5 mL

phosphate buffer for 30 min. Each suspension was ﬁltered through

(500 MHz, DMSO-d₆, mixture of diastereomers):

d 7.35e7.40

a

0.2 micron polyethersulfone membrane ﬁlter. An aliquot

(m, 4 H, PheH), 7.18e7.28 (m, 8 H, PheH), 7.12 (app d, 4 H, J ¼ 7.12,

PheH), 5.18 (s, 2 OH), 5.16 (s, 2 OH), 4.455 (d, 2 H, J ¼ 15.4, CH₂),

4.385 (d, 2 H, J ¼ 15.4, CH₂), 4.14e4.17 (overlapping doublets,

2 H þ 2 H, J ¼ 15.35, CH₂), 4.00 (d, 2 H, J ¼ 16.9, CH₂), 3.935 (d, 2 H,

J ¼ 16.9, CH₂), 3.62 (overlapping doublets, 4 H, J ¼ 16.9, CH₂), 3.32

(m, hidden by H₂O peak, presumed 4 H, NeCH₂eCH₂eN), 3.24 (m,

(2970 L) of the ﬁltrate (pH 7.3) was immediately diluted with 30

m

mL

DMSO and analyzed by UV spectrophotometry. The concentration

of each compound was determined using Beer's law (above) and

the previously-measured molar absorptivities, except for com-

pound 8c, for which the concentration was calculated using the

molar absorptivity of 8a.

4 H, NeCH₂eCH₂eN). ¹³C (100 MHz, DMSO-d₆):

d 169.65, 169.63

(C]O); 139.86, 139.79 (quaternary Ph); 129.55, 129.52 (Ph); 127.80

4.3. Hydrolysis of prodrugs 8a-c

(Ph); 126.94 (Ph); 122.78 (Ph); 61.90, 61.76 (CH₂); 57.51, 57.37

(CH₂); 48.69 (NeCH₂eCH₂eN). ¹¹B (161 MHz, DMSO-d₆):

d 14.33.

4.3.1. LC-MS-monitored hydrolysis

MP: >325 ^ꢀC (sintered and began browning at 285 ^ꢀC). EA Found: C,

57.98; H, 5.60; N, 6.69. Calc. for C₂₀H₂₂B₂N₂O₆þ 0.149C₂H₆OS

(DMSO impurity): C, 58.09; H, 5.50; N, 6.68. HRMS (m/z): 408.1788

(M^ꢄþ; Calc: 408.1664). MS (m/z): 408.4 (M^ꢄþ), 409.3 (M þ H; Calc:

409.17).

4.3.1.1. Compound optimization. Since postulated hydrolysis in-

termediates could not be synthesized and optimized, compound 8a

was dissolved in 1:1H₂O:ACN and directly infused into the mass

spectrometer. Over the course of a day, the mass spectrum was

monitored for ions with the distinctive ~50%:100%:20% isotopic

distribution pattern that was calculated for this type of compound

containing 2 boron atoms. Declustering potential (DP), entrance

potential (EP), and collision cell entrance potential (CEP) were

optimized for each ion. Because of the rapid hydrolysis of the

pinacol ester, the parent ion (m/z 637.4) and the parent ion minus

one pinacol (m/z 555.3) had to be optimized in 100% ACN. Fig. S5

shows the monitored ions ([M þ H]^þ) and their corresponding

structures. Some ions may correspond to more than one structure.

4.1.12. Crystallization and X-Ray data collection

Crystals suitable for x-ray crystallography of compound 8a, the

dimesylate salt of N,N⁰-bis(2-boronic acid pinacol ester benzyl)

ethylenediamine-N,N⁰-diacetic acid methyl ester, were successfully

grown from hot dichloromethane upon cooling.

X-ray crystallography was performed by Dr. Khalil A. Abboud,

Department of Chemistry, University of Florida. X-ray intensity data

were collected at 100 K on a Bruker DUO diffractometer using

MoK

a

radiation (

l

¼ 0.71073 Å) and an APEXII CCD area detector.

4.3.1.2. Hydrolysis conditions. The rate of hydrolysis of 8a to 9 was

determined at ambient temperature. A 3.2 mM stock solution of 8a

was prepared in dry ACN. A 50 mM stock solution of N-methyl-

morpholine (NMM) buffer was prepared, ﬁltered, and adjusted to

pH 7.41 using formic acid. To ensure all species remained dissolved

throughout the reaction, the hydrolysis was carried out in 25:75

MeOH:NMM buffer (v/v), which gave a solution with an apparent

pH of ~7.1. The hydrolysis (n ¼ 1) was initiated by adding a portion

Raw data frames were read by the program SAINT and inte-

grated using 3D proﬁling algorithms. The resulting data were

reduced to produce hkl reﬂections and their intensities and esti-

mated standard deviations. The data were corrected for Lorentz and

polarization effects, and numerical absorption corrections were

applied based on indexed and measured faces.

The structure was solved and reﬁned in SHELXTL6.1 (Bruker-

AXS, Wisconsin, USA), using full-matrix least-squares reﬁnement.

The non-H atoms were reﬁned with anisotropic thermal parame-

ters, and all of the H atoms were calculated in idealized positions

and reﬁned riding on their parent atoms. The asymmetric unit

consists of a half cation and one counterion and a dichloromethane

solvent molecule. The Cl and H atoms of the solvent are disordered

and reﬁned in two parts. The amine proton was obtained from a

Difference Fourier map and reﬁned freely. In the ﬁnal cycle of

reﬁnement, 5529 reﬂections (of which 4832 are observed with

of 8a stock solution (39.80 mM, 0.53% ACN by volume) to the buffer.

Aliquots were removed from the reaction solution via autosampler

every 60 min for 42 h.

4.3.1.3. LC-MS conditions. Some separation of reaction species was

achieved by liquid chromatography using a Luna C₈(50 ꢂ 4.6 mm,

5 mm) column (Phenomenex, USA), connected to a guard cartridge.

The column was at ambient temperature during the chromato-

graphic runs. The sample volume was 25 L. The mobile phase

I > 2

s

(I)) were used to reﬁne 309 parameters and the resulting R₁,

m

wR₂and S (goodness of ﬁt) were 2.85%, 7.48% and 1.056, respec-

tively. The reﬁnement was carried out by minimizing the wR₂

function using F²rather than F values. R₁is calculated to provide a

reference to the conventional R value but its function is not

minimized.

consisted of 0.1% formic acid in H₂O (solvent A, pH 2.7) and 0.1%

formic acid in MeOH (solvent B) at a ﬂow rate of 0.5 mL/min. The

one-hour program was as follows: 40% B (0e5 min), 40%e100% B

(linear gradient, 5e20 min), 100% B (20e45 min), 100%e40% B

(linear gradient, 45e47 min), and 40% B (47e60 min).

Reaction species were detected using quadrupole 1 multiple ion

(Q1MI) monitoring in positive electrospray ionization mode. The

optimized source parameters were as follows: 35 psi (curtain gas),

5500 (IonSpray voltage), 650 ^ꢀC (temperature), 60 psi (nebulizer

gas), 55 psi (turbo gas), 5 (vertical probe). The peaks on the

extracted ion chromatogram corresponding to the molecular ions

observed during the hydrolysis were integrated using Analyst

4.2. Aqueous solubilities of 8a-c

The molar absorptivities of prodrugs 8a and 8b were deter-

mined at 270 nm in 100 mM pH 7.4 sodium phosphate buffer. For

each prodrug, three stock solutions of known concentrations were

prepared in dry DMSO. A 30 mL aliquot of each stock solution was

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

205

v1.4.2 software (AB Sciex, USA).

of all samples was measured upon completion of hydrolysis to

conﬁrm buffer stability.

4.3.1.4. Reaction kinetics. The integrated rate equations for a series

ﬁrst-order reaction are given in Equations (1)e(3), where [A] rep-

resents the concentration of the reactant, [B] the concentration of

the intermediate, and [C] the concentration of the ﬁnal product.

4.4. Relative binding afﬁnity of 9 for iron and copper

4.4.1. Afﬁnity for Fe^3þ

The relative afﬁnity of 9 for iron was determined by a compe-

tition experiment between 9 and HBED in 20 mM pH 7.5 sodium

phosphate buffer, and these results were compared to a competi-

tion experiment between ethylenediaminetetraacetic acid (EDTA)

and HBED. Stock solutions of 9 (9.4 mM) and HBED-HCl (37 mM)

were prepared in DMSO and MeOH, respectively. Stock solutions of

ferric ammonium citrate (FAC, 15 mM) and EDTA (60 mM) were

prepared in H₂O. Five solutions were made containing the

½Aꢆ ¼ ½Aꢆ₀expðꢁk_ABtÞ

k_AB½Aꢆ₀

(1)

½Bꢆ ¼

ðexpðꢁk_ABtÞ ꢁ expðꢁk_BCtÞÞ

(2)

k_BCꢁ k_AB

ꢀ

ꢁ

1

½Cꢆ ¼ ½Aꢆ₀1 ꢁ

ðk_BCexpðꢁk_ABtÞ ꢁ k_ABexpðꢁk_BCtÞÞ

k_BCꢁ k_AB

following: HBED only (90

(300 M), HBED (90 and 300

HBED þ EDTA (90 and 300 M, respectively). All solutions con-

tained 30

mM), 9 only (300

mM), EDTA only

(3)

m

þ

9

mM, respectively), and

m

Thus, a plot of the natural log of the peak area obtained for A

ions (m/z 455.2 and 437.2, Fig. S5) versus time yielded a straight

line, where the negative slope of the linear ﬁt was the observed rate

constant (k_AB) for the decay of A to B. Using k_ABand t_m, the time

necessary to reach the maximum concentration of B (m/z 441.2 and

423.2, Fig. S5), Equation (4) was solved for k_BC, the observed rate

constant for the conversion of B to C, by employing the Solver

function in Microsoft Excel.

m

M FAC. The solutions were allowed to equilibrate for

22 h at 23 ^ꢀC before spectra were acquired. The differences between

spectra in absorbance at 481 nm, indicative of an HBED-Fe chelate,

were noted.

4.4.2. Afﬁnity for Cu^2þ

The relative afﬁnity of 9 for copper was determined by a

competition experiment between 9 and EDTA in a solution of

20 mM pH 7.5 phosphate buffer containing 25% MeOH. This buffer

system was used to ensure all species stayed in solution throughout

the experiment, and had an apparent pH of 7.95. Stock solutions of

9 and EDTA were prepared as described in Section 4.4.1, and a stock

solution of copper(II) sulfate pentahydrate was made in H₂O

ꢂ

ꢃ

k_AB

k_BC

ln

t_m

¼

(4)

k_ABꢁ k_BC

(15 mM). Cuvettes were prepared containing solutions of 100

copper, 300 M 9, and either 0, 20, 60, or 100 M EDTA. A solution

containing only 100 M each of copper and EDTA was also pre-

mM

4.3.2. UV-monitored hydrolysis

m

Hydrolysis studies were performed at 23 ^ꢀC in 30 mM pH 3.1

sodium citrate buffer containing 50% MeOH (8a only), and at 37 ^ꢀC

in 30 mM pH 7.5 phosphate buffer containing 50% MeOH (8a-c).

Methanol was used as a cosolvent to ensure that all species

remained in solution during the hydrolysis. With methanol added,

the apparent pH of the citrate and phosphate systems were 3.8 and

8.5, respectively. Fresh 13e18 mM solutions of 8a-c were prepared

in dry ACN every 6 days as necessary, and were diluted to

m

pared. These solutions were allowed to equilibrate for 24 h at 23 ^ꢀC

before their spectra were acquired. The differences between spectra

in absorbance at 410 nm, indicative of an interaction between 9 and

copper, were noted.

4.5. LC-MS-monitored oxidation of 9 by H₂O₂

150e200

mM (1.2% ACN by volume) in the appropriate buffer sys-

4.5.1. Compound optimization

tem. Spectra were acquired every 30 (8a,b) or 90 (8c) minutes, or

every 30 s for the citrate buffer system, immediately after com-

pound addition. Data was collected until very limited or no spectral

changes were detected. While the l_maxof 8a-c was 272 nm in the

citrate system and 270 nm in the phosphate system, the decrease in

absorbance at 277 nm was used instead to follow the hydrolyses.

Total decrease in absorbance during each reaction was larger at the

latter wavelength, so monitoring this instead of the l_maxwave-

length reduced the inﬂuence of instrument variation toward the

end of the reactions, when the change in absorbance was nearing

the limit of the spectrophotometer's detection.

Once a hydrolysis reaction was near completion, absorbance

values at the wavelength monitored began to repeat over a signif-

icant number of time points. To prevent these repeat values from

affecting the accuracy of the rate constant determination, they

were handled in the following manner: all time points where the

absorbance reading was the same were averaged, with the excep-

tion of the ﬁrst 7 time points, which were never averaged for repeat

absorbance values. Thus, the negative slope of the linear ﬁt of

ln(A_tꢁA_∞) vs. mean time in minutes, where A_tis absorbance at time

t and A_∞is the absorbance upon reaction completion, gave the

pseudo ﬁrst-order rate constant (k_obs). All studies were performed

in triplicate, and the averaged k_obswas reported 1 standard de-

viation. Half-lives were calculated by dividing k into 0.693. The pH

Stock solutions of 9 and HBED monohydrochloride were pre-

pared in DMSO and diluted to 18

mM with 100 mM pH 7.4 N-

methylmorpholine (NMM) buffer. Each solution was then directly

infused into the mass spectrometer, and parameters were opti-

mized for m/z 409.3 (9) and m/z 389.3 (HBED). Ion structures are

given in Fig. 6.

In order to detect and optimize for any intermediates formed

during the oxidation reaction, a 16 mM solution of 9 was prepared in

100 mM pH 7.4 NMM buffer with 127ꢂ excess H₂O₂. Over 60 min of

direct infusion of this reaction solution, m/z 409.3 (9) disappeared,

m/z 389.3 (HBED) appeared, and a third species with an m/z of 399.4

appeared and then subsequently disappeared. This latter ion, with

an isotopic pattern of ~25%:100%:22% that indicated only one boron

atom was present in the molecule, was the “mono” oxidation in-

termediate N,N⁰-(2-boronic acid, 2⁰-hydroxy benzyl)ethylenedi-

amine-N,N⁰-diacetic acid þ H^þe 1 H₂O (Fig. 6).

4.5.2. Reaction conditions

The rate of oxidation of 9 to HBED was determined under

pseudo ﬁrst-order conditions of excess H₂O₂at ambient tempera-

ture. A 5.88 mM stock solution of 9 was prepared in DMSO. A

100 mM NMM buffer was prepared, ﬁltered, and adjusted to pH 7.4

using formic acid. A 62.6 mM stock solution of H₂O₂was prepared

in this NMM buffer using 50 wt% solution H₂O₂in H₂O, and was

206

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

refrigerated when not in use. Samples were prepared by diluting an

aliquot of 9 to a ﬁnal concentration of 59 M with NMM buffer, with

Bruker spectrometers for ¹¹B NMR acquisition; to Laura Faux for

providing LC-MS training; and to Scott Wasdo for his willingness to

share his knowledge, offer advice, and help with mass

spectrometry.

m

samples containing 1% DMSO. Oxidation was then initiated by

addition of an aliquot of H₂O₂stock (0.29e3.5 mM, equaling

~5e59ꢂ molar excess of H₂O₂). Aliquots were removed from each

sample via autosampler every 30e33 min (for samples with 5ꢂ or

12ꢂ excess H₂O₂) or 11 min (for samples with 25ꢂ, 43ꢂ, or 59ꢂ

excess H₂O₂) until the peak area of m/z 389.3 (HBED) no longer

increased. The oxidation reaction was repeated 3e6 times for each

concentration of H₂O₂.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://

dx.doi.org/10.1016/j.ejmech.2016.04.037.

The rate of oxidation of 9 to HBED was also determined in

10 mM phosphate buffer. The buffer was prepared using monobasic

sodium phosphate monohydrate and dibasic sodium phosphate

heptahydrate, and was titrated to pH 7.4 using formic acid. A

13.4 mM stock solution of H₂O₂was prepared in this buffer using

50 wt% solution H₂O₂in H₂O. Oxidation samples were prepared and

analyzed as described in the preceding paragraph. The oxidation

reaction was run 1e2 times for each concentration of H₂O₂.

References

[1] R.R. Crichton, The importance of iron for biological systems, in: Inorganic

Biochemistry of Iron Metabolism: from Molecular Mechanisms to Clinical

Consequences, John Wiley and Sons, Inc, Chichester, West Sussex, 2001, pp.

17e48.

[2] K.N. Raymond, C.J. Carrano, Coordination chemistry and microbial iron

transport, Acc. Chem. Res. 12 (1979) 183e190.

[3] G.J. Anderson and G.D. McLaren. (Eds.), Iron Physiology and Pathophysiology

in Humans, Humana Press, 2012, 704

[4] S.L. Byrne, D. Krishnamurthy, M. Wessling-Resnick, Pharmacology of iron

transport, Annu. Rev. Pharmacol. Toxicol. 53 (2013) 17e36.

[5] A. Jacobs, An intracellular transit iron pool, Ciba Found. Symp. (1976) 91e106.

[6] O. Kakhlon, Z.I. Cabantchik, The labile iron pool: characterization, measure-

ment, and participation in cellular processes, Free Radic. Biol. Med. 33 (2002)

1037e1046.

4.5.3. LC-MS conditions

Very minor separation of reaction species was achieved by

liquid chromatography using a Luna C₈(50 ꢂ 4.6 mm, 5

mm) col-

umn (Phenomenex, USA), connected to a guard cartridge. The col-

umn was at ambient temperature during the chromatographic

[7] M. Kruszewski, Labile iron pool: the main determinant of cellular response to

oxidative stress, Mutat. Res. 531 (2003) 81e92.

runs. The sample volume was 25 mL. The isocratic mobile phase,

ﬂowing at a rate of 0.5 mL/min, contained 40:60 v/v MeOH:H₂O,

with 0.1% formic acid. The time program was 11 min.

[8] B.P. Esposito, W. Breuer, P. Sirankapracha, P. Pootrakul, C. Hershko,

Z.I. Cabantchik, Labile plasma iron in iron overload: redox activity and sus-

ceptibility to chelation, Blood 102 (2003) 2670e2677.

ꢀ

[9] P. Brissot, M. Ropert, C. Le Lan, O. Loreal, Non-transferrin bound iron: a key

Reaction species were detected using quadrupole 1 multiple ion

(Q1MI) in electrospray ionization mode with positive polarity, since

this technique does not require separation of compounds on the

column for peak integration. The optimized source parameters

were as follows: 30 psi (curtain gas), 5500 (IonSpray voltage),

700 ^ꢀC (temperature), 60 psi (nebulizer gas), 70 psi (turbo gas), 3

(vertical probe). The peaks on the extracted ion chromatogram

corresponding to the molecular ions observed during the oxidation

were integrated using Analyst v1.4.2 software (AB Sciex, USA).

role in iron overload and iron toxicity, Biochim. Biophys. Acta 1820 (2012)

403e410.

[10] H.J.H. Fenton, LXXIII. Oxidation of tartaric acid in presence of iron, J. Chem.

Soc. Trans. 65 (1894) 899e910.

[11] H.J.H. Fenton, XLI. The constitution of a new dibasic acid, resulting from the

oxidation of tartaric acid, J. Chem. Soc. Trans. 69 (1896) 546e562.

[12] B. Halliwell, J.M.C. Gutteridge, Oxidative stress: adaptation, damage, or

stimulation?, in: Free Radicals in Biology and Medicine, third ed., Oxford

University Press Inc., New York, 1999, pp. 246e350.

[13] K. Jomova, M. Valko, Advances in metal-induced oxidative stress and human

disease, Toxicology 283 (2011) 65e87.

[14] R. Fleming, P. Ponka, Iron overload in human disease, N. Engl. J. Med. 366

(2012) 348e359.

[15] L. Zecca, M.B.H. Youdim, P. Riederer, J.R. Connor, R.R. Crichton, Iron, brain

4.5.4. Reaction kinetics

ageing and neurodegenerative disorders, Nat. Rev. Neurosci.

5 (2004)

The integrated rate equations for a series ﬁrst-order reaction are

given by Equations (1)e(3) in Section 4.3.1.4, where [A] represents

the concentration of the reactant (9), [B] the concentration of the

intermediate (N,N⁰-(2-boronic acid, 2⁰-hydroxy benzyl)ethylenedi-

amine-N,N⁰-diacetic acid), and [C] the concentration of the ﬁnal

product (HBED).

The natural log of the peak area obtained for A (m/z 409.3,

compound 9) versus time yielded a straight line, where the nega-

tive slope of the linear ﬁt was the observed rate constant (k_AB) for

the decay of A to B. In a case where k_AB>> k_BC(i.e. the second step is

rate determining), Equation (3) reduces to Equation (5):

863e873.

[16] J.L. Dunaief, Iron induced oxidative damage as a potential factor in age-related

macular degeneration: the Cogan lecture, Invest. Ophthalmol. Vis. Sci. 47

(2006) 4660e4664.

[17] C. Pourzand, R.D. Watkin, J.E. Brown, R.M. Tyrrell, Ultraviolet A radiation in-

duces immediate release of iron in human primary skin ﬁ broblasts: the role of

ferritin, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 6751e6756.

[18] F. L'Eplattenier, I. Murase, A.E. Martell, New multidentate ligands. VI.

Chelating tendencies of N,N'-(di(2-hydroxybenzyl)ethylenediamine-N,N'-

diacetic acid, J. Am. Chem. Soc. 89 (1967) 837e843.

[19] R.W. Grady, C. Hershko, HBED: a potential oral iron chelator, Ann. N.Y. Acad.

Sci. 612 (1990) 361e368.

[20] R.W. Grady, A. Jacobs, The screening of potential iron chelating drugs, in:

A.E. Martell, W.F. Anderson, D.G. Badman (Eds.), Development of Iron Chela-

tors for Clinical Use : Proceedings of the Second Symposium on the Devel-

opment of Iron Chelators for Clinical Use, Elsevier/North Holland, Inc., New

York, NY, 1981, pp. 133e164.

½Cꢆ ¼ f1 ꢁ expðꢁk₂tÞg½Aꢆ₀

(5)

Therefore, ln(A_∞ꢁA_t) for C (HBED, m/z 389.3), where A_∞is the

peak area at the ﬁnal time point and A_tis the peak area at time t,

was plotted against time. The negative slope of the linear ﬁt was the

observed rate constant (k_BC) for the decay of B to C.

[21] H.H. Peter, R.J. Bergeron, R.R. Streiff, J. Wiegand, A comparative evaluation of

iron chelators in a primate model, in: R.J. Bergeron, G.M. Brittenham (Eds.),

The Development of Iron Chelators for Clinical Use, CRC Press, Boca Raton,

Florida, 1994, pp. 373e394.

[22] R.W. Grady, A.D. Salbe, M.W. Hilgartner, P.J. Giardina, Results from a phase 1

clinical trial of HBED, in: C. Hershko, A.M. Konain, P. Aisen (Eds.), Progress in

Iron Research, Plenum Press, New York, 1994, pp. 351e359.

[23] R.J. Motekaitis, A.E. Martell, M.J. Welch, Stabilities of trivalent metal com-

plexes of phenolic ligands related to N,N'-bis(2-hydroxybenzyl)ethylenedi-

amine-N,N'-diacetic acid (HBED), Inorg. Chem. 29 (1990) 1463e1467.

[24] C.G. Pitt, Y. Bao, J. Thompson, M.C. Wani, H. Rosenkrantz, J. Metterville, Esters

and lactones of phenolic amino carboxylic acids: prodrugs for iron chelation,

J. Med. Chem. 29 (1986) 1231e1237.

The slopes of the lines through the plots of k_ABand k_BCvs. H₂O₂

concentration provided the second-order rate constants

k

(M^ꢁ1min^ꢁ1) for the oxidation of 9 (m/z 409.3) to the mono inter-

mediate (m/z 399.4) and for the oxidation of the mono intermediate

to HBED (m/z 389.3), respectively.

[25] R.J. Bergeron, J. Wiegand, G.M. Brittenham, HBED: a potential alternative to

deferoxamine for iron-chelating therapy, Blood 91 (1998) 1446e1452.

[26] B. Faller, C. Spanka, T. Sergejew, V. Tschinke, Improving the oral bioavailability

of the iron chelator HBED by breaking the symmetry of the intramolecular H-

Acknowledgements

A special thanks to Jim Rocca for providing training on the

N.A. Thiele et al. / European Journal of Medicinal Chemistry 118 (2016) 193e207

207

bond network, J. Med. Chem. 43 (2000) 1467e1475.

393 (2012) 294e303.

ꢀꢁ

ꢀ

ꢁ

ꢀ

ꢂꢁ

ꢀ

[27] H.G. Kuivila, A.G. Armour, Electrophilic displacement reactions. IX. Effects of

substituents on rates of reactions between hydrogen peroxide and benzene-

boronic acid, J. Am. Chem. Soc. 79 (1957) 5659e5662.

[28] J.B. Galey, J. Dumats, I. Beck, B. Fernandez, M. Hocquaux, N,N'-bis-dibenzyl

ethylenediaminediacetic acid (DBED) - a site-speci ﬁ c hydroxyl radical scav-

enger acting as an oxidative stress activatable iron chelator in-vitro, Free

Radic. Res. 22 (1995) 67e86.

[29] J.B. Galey, J. Dumats, S. Genard, O. Destree, P. Pichaud, P. Cctroux, L. Marrot,

I. Beck, B. Fernandez, G. Barre, N,N'-bis-(3,4,5-trimethoxybenzyl) ethylenedi-

amine N,N'-diacetic acid as a new iron chelator with potential medicinal

applications against oxidative stress, Biochem. Pharmacol. 51 (1996)

103e115.

[41] H. Jansova, M. Machacek, Q. Wang, P. Haskova, A. Jirkovska, E. Potuckova,

ꢁ

ꢂ

F. Kielar, K.J. Franz, T. Simunek, Comparison of various iron chelators and

prochelators as protective agents against cardiomyocyte oxidative injury, Free

Radic. Biol. Med. 74 (2014) 210e221.

[42] H. Bundgaard, The double prodrug concept and its applications, Adv. Drug

Deliv. Rev. 3 (1989) 39e65.

[43] C.W. Gray, T.A. Houston, Boronic acid receptors for alpha-

hydroxycarboxylates: high af ﬁ nity of Shinkai's glucose receptor for tartrate,

J. Org. Chem. 67 (2002) 5426e5428.

[44] C.W. Gray, B.T. Walker, R.A. Foley, T.A. Houston, Boronate derivatives of

bioactive amines: potential neutral receptors for anionic oligosaccharides,

Tetrahedron Lett. 44 (2003) 3309e3312.

[45] X. Zhang, L. You, E.V. Anslyn, X. Qian, Discrimination and classi ﬁ cation of

ginsenosides and ginsengs using bis-boronic acid receptors in dynamic

multicomponent indicator displacement sensor arrays, Chem. Eur. J. 18 (2012)

1102e1110.

[46] H.R. Snyder, J.A. Kuck, J.R. Johnson, Organoboron compounds, and the study of

reaction mechanisms. Primary aliphatic boronic acids, J. Am. Chem. Soc. 60

(1938) 105e111.

[47] D.G. Hall, Structure, properties, and preparation of boronic acid derivatives:

overview of their reactions and applications, in: D.G. Hall (Ed.), Boronic Acids:

Preparation and Applications in Organic Synthesis, Medicine and Materials,

Wiley-VCH, Weinheim, Germany, 2011, pp. 1e133.

ꢀ

[30] J.B. Galey, O. Destree, J. Dumats, P. Pichaud, J. Marche, S. Genard, G. Bracciolli,

L. Le Capitaine, H. Plessix, L. Brambilla, O. Cantoni, Protection of U937 cells

against oxidative injury by a novel series of iron chelators, Free Radic. Biol.

Med. 25 (1998) 881e890.

ꢀ

[31] J. Galey, O. Destree, J. Dumats, S. Genard, P. Tachon, Protection against

oxidative damage by iron chelators: effect of lipophilic analogues and pro-

drugs of N,N ‘-bis(3,4,5-trimethoxybenzyl)ethylenediamine-N,N ‘-diacetic acid

(OR10141), J. Med. Chem. 43 (2000) 1418e1421.

[32] G. Serratrice, J. Galey, E. Saint Aman, J. Dumats, Iron(III) complexation by new

aminocarboxylate chelators e thermodynamic and kinetic studies, Eur. J.

Inorg. Chem. 2001 (2001) 471e479.

[33] M. Pygmalion, L. Ruiz, E. Popovic, J. Gizard, P. Portes, X. Marat, K. Lucet-

Levannier, B. Muller, J. Galey, Skin cell protection against UVA by Sideroxyl, a

new antioxidant complementary to sunscreens, Free Radic. Biol. Med. 49

(2010) 1629e1637.

[34] L.K. Charkoudian, D.M. Pham, K.J. Franz, A pro-chelator triggered by hydrogen

peroxide inhibits iron-promoted hydroxyl radical formation, J. Am. Chem. Soc.

128 (2006) 12424e12425.

[48] L.S. Santucci, H. Gilman, Some bromine-containing and sulfur-containing ar-

omatic boronic acids, J. Am. Chem. Soc. 80 (1958) 193e196.

[49] C. Achilli, A. Ciana, M. Fagnoni, C. Balduini, G. Minetti, Susceptibility to hy-

drolysis of phenylboronic pinacol esters at physiological pH, Cent. Eur. J.

Chem. 11 (2013) 137e139.

[50] J.E. House, Series ﬁ rst-order reactions, in: Principles of Chemical Kinetics, Wm.

C. Brown, Dubuque, IA, 1997, pp. 40e45.

[35] L.K. Charkoudian, D.M. Pham, A.M. Kwon, A.D. Vangeloff, K.J. Franz, Modi ﬁ -

cations of boronic ester pro-chelators triggered by hydrogen peroxide tune

reactivity to inhibit metal-promoted oxidative stress, Dalton Trans. (2007)

5031e5042.

[36] L.K. Charkoudian, T. Dentchev, N. Lukinova, N. Wolkow, J.L. Dunaief, K.J. Franz,

Iron prochelator BSIH protects retinal pigment epithelial cells against cell

death induced by hydrogen peroxide, J. Inorg. Biochem. 102 (2008)

2130e2135.

[37] M.G. Dickens, K.J. Franz, A prochelator activated by hydrogen peroxide pre-

vents metal-induced amyloid beta aggregation, ChemBioChem. 11 (2010)

59e62.

[38] M.G.D. Leed, N. Wolkow, D.M. Pham, C.L. Daniel, J.L. Dunaief, K.J. Franz, Pro-

chelators triggered by hydrogen peroxide provide hexadentate iron coordi-

nation to impede oxidative stress, J. Inorg. Biochem. 105 (2011) 1161e1172.

[39] F. Kielar, M.E. Helsel, Q. Wang, K.J. Franz, Prochelator BHAPI protects cells

against paraquat-induced damage by ROS-triggered iron chelation, Metal-

lomics 4 (2012) 899e909.

[51] B. Capon, B.C. Ghosh, The mechanism of the hydrolysis of phenyl salicylate

and catechol monobenzoate in the presence and absence of borate ions,

J. Chem. Soc. B (1966) 472e476.

[52] D.W. Tanner, T.C. Bruice, Boric acid esters. I. A general survey of aromatic li-

gands and the kinetics and mechanism of the formation and hydrolysis of

boric acid esters of salicylamide, N-phenylsalicylamide, and disalicylimide,

J. Am. Chem. Soc. 89 (1967) 6954e6971.

[53] T. Okuyama, H. Nagamatsu, T. Fueno, Mechanism of hydrolysis of hydroxy

thiolesters in the presence of boric acid, J. Org. Chem. 46 (1981) 1336e1342.

[54] K. Swamy, S. Ko, S. Kwon, H. Lee, C. Mao, J. Kim, K. Lee, J. Kim, I. Shin, J. Yoon,

Boronic acid-linked ﬂ uorescent and colorimetric probes for copper ions,

Chem. Commun. (2008) 5915e5917.

[55] A.E. Martell, R.M. Smith, Critical Stability Constants, Plenum Press, New York,

1974.

[56] X. Ma, S. Yu, F. Liu, Synthesis of N-methylmorpholine, J. Qingdao Univ. Sci.

Technol. 29 (2008) 283e286.

[57] G.W. Kabalka, H.C. Hedgecock Jr., A mild and convenient oxidation procedure

for the conversion of organoboranes to the corresponding alcohols, J. Org.

Chem. 40 (1975) 1776e1779.

[40] F. Kielar, Q. Wang, P. Boyle, K. Franz, A boronate prochelator built on a triazole

framework for peroxide-triggered tridentate metal binding, Inorg. Chim. Acta

Article Doi

Novel double prodrugs of the iron chelator N,N'-bis(2-hydroxybenzyl)ethylenediamine-N,N'-diacetic acid (HBED): Synthesis, characterization, and investigation of activation by chemical hydrolysis and oxidation

DOI: 10.1016/j.ejmech.2016.04.037

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.ejmech.2016.04.037

Full text of DOI:10.1016/j.ejmech.2016.04.037

Products guided by the article

R&D Labs maybe for 35998-29-9

Relevant to this article

Hot Product