Cyclization of the acyl glucuronide metabolite of a neutral endopeptidase inhibitor to an electrophilic glutarimide: Synthesis, reactivity, and mechanistic analysis

Article abstract of DOI:10.1021/jm0706766

The neutral endopeptidase inhibitor (2R)-2-[(1-{[(5-ethyl-1,3,4-thiadiazol- 2-yl)amino]carbonyl}cyclopentyl)-methyl]pentanoic acid 2 is metabolized to acyl glucuronide 3. Unprecedentedly, at pH 7.4, 3 does not undergo the O-acyl migration characteristic o

Full text of DOI:10.1021/jm0706766

J. Med. Chem. 2007, 50, 6165-6176
6165
Cyclization of the Acyl Glucuronide Metabolite of a Neutral Endopeptidase Inhibitor to an
Electrophilic Glutarimide: Synthesis, Reactivity, and Mechanistic Analysis
Xiaoli Meng,† James L. Maggs,‡ David C. Pryde,§ Simon Planken,§ Rosalind E. Jenkins,‡ Torren M. Peakman,§
Kevin Beaumont,# Christopher Kohl,#,| B. Kevin Park,‡ and Andrew V. Stachulski*,†
The Robert Robinson Laboratories, Department of Chemistry, UniVersity of LiVerpool, LiVerpool L69 7ZD, United Kingdom,
Department of Pharmacology and Therapeutics, UniVersity of LiVerpool, LiVerpool L69 3GE, United Kingdom,
Department of DiscoVery Chemistry, and Pharmacokinetics, Dynamics and Metabolism, Pfizer Global Research and
DeVelopment, Ramsgate Road, Sandwich, Kent CT13 9NJ, United Kingdom
ReceiVed June 11, 2007
The neutral endopeptidase inhibitor (2R)-2-[(1-{[(5-ethyl-1,3,4-thiadiazol-2-yl)amino]carbonyl}cyclopentyl)-
methyl]pentanoic acid 2 is metabolized to acyl glucuronide 3. Unprecedentedly, at pH 7.4, 3 does not undergo
the O-acyl migration characteristic of acyl glucuronides but rapid, eliminative cyclization (t1/2 at 37 °C,
10.2 min) to glutarimide 4. Glucuronide 3 was synthesized efficiently via acylation of benzylglucuronate
with N-benzyloxymethyl-protected 2. Glucuronide and imide reacted rapidly in aqueous solution, pH 7.4,
with amino acids and glutathione to form stable amides and unstable thioesters. Imide 4 acylated eight
lysine Nꢀ-amino groups of human serum albumin. Rapid cyclization of 3 was attributed to attack on the
ester linkage by an unusually nucleophilic glutaramide NH (pKa in 2 ) 9.76). N-propyl 3 was refractory to
acyl migration and cyclization. This suggested a synthetic strategy for preparing analogues of 2 that form
chemically stable acyl glucuronides.
Introduction
Recent reports1,2 from Pfizer scientists described a series of
glutaramide analogues 1 (Figure 1) which were potent inhibitors
of neutral endopeptidase (NEPa, EC 3.4.24.11) and therefore
of potential value in the treatment of female sexual arousal
disorder (FSAD).3 From this series, the thiadiazole analogue 2,
(2R)-2-[(1-{[(5-ethyl-1,3,4-thiadiazol-2-yl)amino]carbonyl}-
cyclopentyl)methyl]pentanoic acid (R-21b; R-132), was selected
as a development candidate in view of its combined activity
profile and pharmacokinetic behavior.
Figure 1. Generic structure of glutaramide NEP inhibitors (1) and
structure of 1,3,4-thiadiazole analogue 2.
In common with many carboxylic acid-containing drugs,4 2
is metabolized initially to its O-acyl glucuronide 3.2 While this
property is shared by the series of structure 11b,2, it was soon
realized that 3 had a special reactivity: namely, it cyclized ex
vivo at physiological pH to give the glutarimide 4 (Figure 2).
This very rapid cyclization was attributed to a highly acidic
amide NH and formation of a favorable ring size.5 Subsequent
slow hydrolysis of 4 led to a mixture of 2 and the isomeric
acid 5.
The exceptional nature of this rearrangement and the elec-
trophilic character of imides raises the possibility that O-
glucuronidation of series 1 compounds in vivo will result in
previously unsuspected nonenzymatic pathways, including,
especially, unfamiliar modifications to biological macromol-
Figure 2. Pathways of formation and rearrangement of acyl glucu-
ronide 3 and hydrolysis of glutarimide intermediate 4.
ecules by the glutarimide. The determinants and characteristics
of these reactions are likely to be distinct from those of the
well characterized, and to a certain extent predictable,4,6
transacylation, intramolecular O-acyl migration, and consequen-
tial protein glycation (imine formation) reactions of convention-
ally labile acyl glucuronides.4,7 Potential biological consequences
of the latter metabolites’ reactivity include the idiosyncratic
adverse reactions not infrequently associated with drugs that
form acyl glucuronides.7,8 Unacceptable toxicity has been
associated with 2 in some experimental animals.5 It was
therefore a matter of importance to develop an effective
synthesis of the acyl glucuronide 3 and to characterize the
reactions of 3 and 4 with biological nucleophiles. Despite the
very short half-life of 3 and consequent suppression of acyl
* To whom correspondence should be addressed. Telephone: 44(0)151
794 3542. Fax: 44(0)151 794 3588. E-mail: stachuls@liv.ac.uk
† Department of Chemistry, University of Liverpool.
‡ Department of Pharmacology and Therapeutics, University of Liverpool.
§ Department of Discovery Chemistry, Pfizer Global Research and
Development.
# Pharmacokinetics, Dynamics and Metabolism, Pfizer Global Research
and Development.
| Current address: Preclinical Pharmacokinetics and Metabolism, Actelion
Pharmaceuticals Ltd., CH-4123 Allschwil, Switzerland.
a Abbreviations: Bom, benzyloxymethyl; DIAD, diisopropyl azodicar-
boxylate; FSAD, female sexual arousal disorder; HATU, O-(7-azabenzo-
triazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate; HSA, hu-
man serum albumin; nanoLC-MS/MS, nano liquid chromatography tandem
mass spectrometry; NEP, neutral endopeptidase; RSA, rat serum albumin.
10.1021/jm0706766 CCC: $37.00 © 2007 American Chemical Society
Published on Web 11/07/2007

6166 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Scheme 1. Direct Synthesis of the 1-â-O-Glucuronide from 2a
Meng et al.
N-protecting group to prevent the intramolecular cyclization,
followed by a mild final deprotection step.
First we studied N-allyl protection, as this should allow a
one-step deprotection of both allyl groups when using 6.
Conversion of 2 to the bis-allyl derivative 8 was achieved in
excellent yield under basic conditions (2 equiv of allyl bromide,
2 equiv of K2CO3, DMF, 20 °C, 80%: Scheme 2). Hydrolysis
of the allyl ester was slow but afforded the N-allyl acid 9 in
70% yield and high purity. As we anticipated, the N-protection
completely suppressed the cyclization, and reaction of 9 with 6
(HATU, NMM, MeCN) gave an excellent yield of the protected
acyl glucuronide ester 10 (65%).
Deprotection of 10 proved extremely difficult, however. Using
Pd(PPh3)4 and morpholine14 in THF, the allyl ester, but not the
N-allyl group, was removed. After much experimentation, we
found that 10 was partially deprotected by catalytic transfer
hydrogenolysis (Et3N, HCOOH, Pd-C, 60 °C)15 but that the
N-propyl acyl glucuronide 11 was the product. Incidentally this
N-substituted acyl glucuronide was again highly stable at
physiological pH, in contrast to 3.
a Reagents and conditions: (i) 6, Ph3P, DIAD, THF, separate â-anomer
by prep HPLC, 30%; (ii) 6, HATU, NMM, MeCN, ca. 10% 7 + 50% 4;
(iii) Pd(PPh3)4, pyrrolidine or morpholine, THF, 20 °C, prep HPLC, 70%.
All ) CH2CHdCH2.
migration, and hence attenuation of the Schiff base-Amadori
rearrangement pathway on glycated proteins,9 it was conceivable
that 3 might nevertheless react by rapid acyl transfer to give
protein adducts. On the other hand, the cyclization might prove
to be a ‘safety mechanism’ which lessens the risk of covalent
protein adduct formation.4,7,8 The chemical synthesis of such a
short-lived and potentially highly reactive metabolite was a
challenging task, and its successful achievement is the subject
of this report. We report also a definitive confirmation of the
rearrangement and hydrolytic pathways shown in Figure 2 using
synthetic 3 and 4, a study of the reactivity of imide 4 with amino
acid models of protein nucleophiles, and the covalent modifica-
tion of serum protein by 4 in vitro.
We investigated other protecting groups for the NH moiety
of 2. Introduction of Boc or Z protection was feasible from an
ester of 2, but the resulting acylated amides proved extremely
base-labile. When a benzyl ester of 2 was used, the derived
N-Boc compound could not be hydrogenated without loss of
the Boc group, even in EtOAc as solvent.
Results
Chemistry. Methods for acyl glucuronide synthesis have been
reviewed.4,10 It is feasible to prepare acyl glucuronide 3 as an
anomeric mixture directly from 2 using the Mitsunobu procedure
developed by Juteau et al.11 Thus condensation of allyl glucu-
ronate 611,12 with 2 under typical Mitsunobu conditions (Ph3P,
DIAD, THF) gave an anomeric mixture (â:R ca. 4:1) of
the conjugate 7 in 50% yield along with other products
(Scheme 1).
We therefore resorted to ether protection so that the inter-
mediate would have sufficient base stability. We chose the
N-benzyloxymethyl (Bom) group16 which has been successfully
used in the synthesis of histidine peptides: deprotection by
hydrogenolysis,16 or relatively mild acidolysis,17 is then feasible.
While the Bom group could be introduced via an ester of 2 by
N-alkylation (Bom-Cl, Et3N, CH2Cl2) followed by ester hy-
drolysis, such hydrolyses were sluggish (on the Me or allyl
ester), and we devised instead a very efficient three-step, one-
pot synthesis from 2 (Scheme 3). Thus silylation of 2 using
TBS-Cl and Et3N (2 equiv each) in CH2Cl2 afforded the
presumed N,O-bis-silyl derivative: N-silylation was quite slow,
requiring 20 h at 20 °C. Reaction of this crude product with
Bom-Cl, followed by cleavage of the presumed TBS ester
intermediate using TBAF, then afforded the desired N-Bom acid
12 in 80% overall yield. Reaction with benzyl glucuronate 1318
under similar conditions to the previous (HATU, DABCO,
MeCN) afforded the coupled product 14 in 40% yield (unop-
timized). While this yield was lower than in some other
examples, pure â-product was again obtained: in this case,
NMM gave a very sluggish reaction and DABCO was necessary
for a reasonable rate (see Supporting Information).
Finally we optimized the deprotection of 14 by hydrogenoly-
sis with 10% Pd-C catalyst. Using PriOH, or PriOH-AcOH
mixtures, the benzyl ester was readily removed but the Bom
group was only slowly cleaved. However, by using pure AcOH,
complete deprotection was achieved in 16 h at 20 °C to deliver
highly pure 3, emphasizing again that the acyl glucuronide is
far more stable at mildly acid pH than at physiological pH.13,19
We did not observe (LC-MS) any N-methyl compound which
might have been formed via reduction of an iminium intermedi-
ate. Final purification was achieved simply by evaporation and
trituration of the crude product with acetone to give 3 as a white
solid in >95% purity by LC-MS analysis.
Complete separation of the anomers required preparative
HPLC, and the desired â-conjugate 7 was eventually isolated
in 30% yield. Deprotection of 7 using Pd(PPh3)4 and pyrrolidine
afforded the free acyl glucuronide 3 (70%) without acyl
migration or anomerization, again following preparative HPLC.
Difficulties were encountered in the purification owing to the
extremely short aqueous half-life of 3. It was best to use 0.5%
aq AcOH-acetonitrile, 19:1, as the mobile phase, since 3 is
appreciably more stable at mildly acidic pH than at pH 7.4, in
common with other acyl glucuronides.4,13
We have recently described12 a new, versatile method of acyl
glucuronide synthesis by selective acylation of a glucuronate
monoester such as 6. The great advantage of this procedure is
that the conjugate is formed almost entirely (95:5 or better) as
the desired â-anomer and preparative HPLC at this stage is
unnecessary. Here, however, under our standard conditions
(HATU, NMM, 2 + 6 in MeCN), the major product was the
imide 4 and the desired conjugate 7 was formed in <10% yield,
though entirely as the â-anomer (Scheme 1). Clearly even these
mildly basic conditions cannot effectively prevent the cyclization
reaction. We therefore studied the use of an appropriate
Stability of Acyl Glucuronide 3 and Reactions with
Nucleophiles. Synthetic 3 that was free of imide 4 by LC-MS,
when incubated in 100 mM sodium phosphate, pH 7.4, at
37 °C underwent the reaction sequence shown in Figure 2.

Cyclization of the Acyl Glucuronide Metabolite
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 6167
Scheme 2. Acyl Glucuronide Synthesis Using N-Allyl Protectiona
a Reagents and conditions: (i) CH2CHdCH2Br, K2CO3 (2 equiv each), DMF, 20 °C, 80%; (ii) NaOH, aq MeOH, 20 °C, 80%; (iii) 6, HATU, NMM,
MeCN, 0-20 °C, 65%; (iv) Et3N, HCOOH, 10% Pd-C, 60 °C, 70%.
Scheme 3. Synthesis of Acyl Glucuronide 3 Using Benzyloxymethyl (Bom) Protectiona
a Reagent and conditions: (i) TBSCl, CH2Cl2, Et3N, 20 h; (ii) Bom-Cl, Et3N; (iii) TBAF, THF, 80% for steps i to iii; (iv) 13, HATU, DABCO, MeCN,
0-20 °C, 40%; (v) H2-10% Pd-C, AcOH, 80%.
Analysis of serial aliquots by LC-MS (20 to 50 to 70%
acetonitrile in 0.1% formic acid, over 30 and 5 min, respectively;
required to resolve 3 and 2) revealed complete degradation of
3 (tR 18.5 min, m/z 516 [MH+], 340 [MH - 176 (dehydroglu-
curonic acid)]+, 322, 294, 183, 165, 130; m/z 514 [M - H]-,
m/z 338 [M-176]-, 193 [glucuronic acid - 1]-) after 1 h,
yielding 4 (34 min; m/z 322 [MH+], 294, 156, 137, 130), 2 (23
min; m/z 340 [MH+], 322, 183, 165, 137, 130) and 5 (26 min;
m/z 340, 322, 183, 165, 137, 130) that were identified by
chromatographic and mass spectrometric comparisons with
synthetic standards. The two carboxylic acids were formed in
equal proportions (chromatographic peak areas at 254 nm). The
degradation half-life of 3 in phosphate buffer, pH 7.4, at 37 °C
was 10.2 min. Cyclization of 3 proceeded slowly in phosphate
buffer at pH 6.0 over 6 h, but the glucuronide was stable at pH
3.0. Neither LC-MS analysis of these incubations nor 1H NMR
(500 MHz) analysis of incubations performed in deuterated
phosphate buffer (pD ca. 7.2) over 10.5 h gave evidence that 3
underwent acyl migration: none of the multiple peaks of m/z
516 corresponding to positional isomers of 3 was observed,4,13
and the acyl glucuronide decomposed rapidly in the NMR tube
to form only two major products, identified as imide 4 and
D-glucuronic acid (Figure 3). In the latter case, decomposition
was apparent after only 25 min by the appearance of distinctive
doublets attributable to the 1-R, 1-â, and 5-R H of D-glucuronic
acid. The downfield shift of the pyran ring 2′ C-H signal that is
characteristic of acyl group migration from 1â-O to 2 R/â-O19
was not seen.
The proposition that rapid cyclization of 3 at neutral pH
requires the amide NH to be unusually acidic for nucleophilic
attack on the ester linkage5 was tested by determining the pKa
of 2. From pH-metric titration, the carboxylate and amide pKa
were 4.74 and 9.76, respectively.
Despite its short half-life in vitro at physiological pH, 3 might
directly transacylate nucleophilic amino or thiol species in vivo.
This possibility was investigated by incubating 3 with a 40-
fold molar excess of either NR-acetyl L-lysine 15 or N-acetyl
L-cysteine 16 in phosphate buffer, pH 7.4, at 37 °C for up to
18 h. Using LC-MS to monitor the reactions, the products
corresponded to imide 4 detected as an intermediate, and,
respectively, the isomeric amide (17a,b) and thioester (18a,b)
conjugates of 2 (17a and 18a) and 5 (17b and 18b) that were
obtained in similar ratios from 4 itself (Scheme 4; Figure 4;
Table 1). The isomeric conjugates produced by reaction of 3
with glutathione were also identical by LC-MS analysis to those
derived from 4. We conclude that the rearrangement of 3 to 4

6168 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Meng et al.
Figure 3. Sequential 1H NMR spectra (500 MHz) of a solution of acyl glucuronide 3 (39 µM) in deuterated phosphate buffer (pD ca. 7.2; 20 °C)
showing degradation of 3 to imide 4 (see Figure 2) and D-glucuronic acid (DG). Incubation times: A ) 25 min, B ) 42 min, C ) 104 min,
D ) 160 min, E ) 10 h 30 min. 1-â H of acyl glucuronide 3 (δ 5.35, 1 H, d). DG: 1-R H, 1-â H and 5-R H (1 H, d) of D-glucuronic acid. 4-a
(3H, t), 4-b (2H, q), and 4-c (3H, t) correspond to the CH3 of the propyl group, the CH2 of the ethyl group, and the CH3 of the ethyl group of 4,
respectively.
Scheme 4. Reaction of Imide 4 with Amino Acid Nucleophilesa
between 60 min and 4.5 h) showed a slight though persistent
bias in favor of 5 except at 15 min, when the ratio was 1.0.
The ratio was lower (0.65) when 4 (500 µM) was incubated
with human serum albumin (HSA; 15 µM) under the same
conditions but slightly higher (0.9) when it was incubated with
rat serum albumin (RSA). These effects might be attributable
to protein-mediated hydrolysis: in addition to its more familiar
esterase activity,21 HSA is known to possess â-lactamase
activity22 and aryl acylamidase activity which accounts for
approximately 10% of that activity in plasma.23
Imide 4 was reacted with model nucleophiles NR-acetyl
L-lysine 15 and N-acetyl L-cysteine 16 (Scheme 4). The reaction
of 4 with 15 occurs satisfactorily in aqueous ethanol at pH 9,
where 15 exists largely in its free base form and can overcome
the background hydrolysis of 4; a mixture of amides 17a and
17b was obtained in a 3:1 ratio determined by HPLC analysis
with UV detection. At pH 7 only 2 and 5 resulted. The isomers
were separable by preparative HPLC and showed sufficient
differences in their 1H NMR for assignment to amides of 2 and
5, respectively. Reaction of 4 with 16 was effected in acetonitrile
under base catalysis and similarly a mixture of isomeric
thioesters 18a and 18b (2.9:1 ratio by HPLC) resulted, separable
by chromatography. These thioesters were relatively unstable
on standing or in aqueous solution but were convincingly
characterized. It was deduced from preferential formation of
17a and 18a that nucleophilic attack on the imide by 15 and
16, unlike hydrolysis, occurred selectively at the apparently less
hindered C-8 carbonyl. In phosphate buffer, pH 7.4, at 37 °C
rearrangement of purified 18a to 4 proceeded to completion
within 1 h and was extensive even at pH 6.0. However, the
thioester was stable at pH 3.0 for a least 5 h. Incubation of 18a
at pH 7.4 and 37 °C with a 40-fold molar excess of glutathione
for 6 h yielded small amounts of two isomeric glutathione
conjugates. The N-acetyl L-cysteine conjugate was much more
stable in the presence of glutathione, but it was not possible to
attribute the glutathione conjugates’ formation to transacylation
of 18a because of the formation of a small amount of 4.
a Reagents and conditions: (i) NR-acetyl L-lysine 15, EtOH, Na2B4O7/
NaCl buffer, pH 9.22 ( 0.02, (ii) N-acetyl L-cysteine 16, Et3N, DMAP,
acetonitrile. 17a and 18a were the major products under these conditions.
is sufficiently fast that the observed products of nucleophilic
attack may be attributed to the intermediacy of 4.
Stability of Imide 4 and Reactions with Nucleophiles.
Imide 4 is readily available by treatment of acid 2 with DIC
and HOBt in the presence of N-methylmorpholine. In order to
rationalize an earlier observation that isomeric acid 5 is a major
metabolite of 2 in vivo,20 we first showed that opening of 4 in
aqueous conditions at pH 7.4 yielded a mixture of 2 and 5
(Figure 2), identified by comparisons with synthetic standards,
in a roughly 1:1 ratio. Carboxylic acid 2 did not cyclize
spontaneously. The degradation half-life of 4 in phosphate
buffer, pH 7.4, at 37 °C was 8.7 h, but the imide was resistant
to hydrolysis at pH 6.0. When synthetic 4 was hydrolyzed under
these conditions over 4.5 h, the ratio of 2:5 (0.83-0.86, n ) 5,

Cyclization of the Acyl Glucuronide Metabolite
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 6169
Figure 4. Chromatographic analysis (reversed-phase HPLC) of the products formed during an 18-h incubation of acyl glucuronide 3 with NR-
acetyl L-lysine 15 (40-fold molar excess). (A) UV chromatogram; (B) mass chromatogram ([MH+]) for amides 17a and 17b; (C) mass chromatogram
([MH+]) for acids 2 and 5 formed by hydrolysis of intermediate imide 4 (produced by cyclization of 3).
Table 1. Conjugates Formed by the Reaction of Imide 4 with Amines
NR-acetyl L-serine gave no trace of product at 1 and 20 h,
and Thiols in Aqueous Solutiona
suggested that stable esters of 2 and 5 are formed more slowly
% of
isomeric
conjugate
stabilityd
than NR-amides. Two isomeric conjugates of L-histidine (ratio,
3:1), seen in trace amounts at 1 h, represented 8% of imide
turnover after 20 h. However, the positions(s) of substitution
could not be determined. No derivatives of L-proline were found.
Pairs of conjugates were also obtained from N-acetyl cysteine
16 (18a and 18b) and glutathione, but the thioesters were
unstable under the above conditions and degraded to acids 2
and 5. This may have occurred via 4, which, as noted above, is
created spontaneously from 18b. The susceptibility of thioesters
of 16 to hydrolysis at pH 7.4 has been observed previously.24
Cysteine and 16 appeared to compete with hydrolysis of 4 more
effectively than the other nucleophiles: after 1 h, turnover of 4
in incubations containing cysteine or 16 was only attributable
to conjugate formation (Table 1). Although the formation of
two stable conjugates from L-cysteine over 1 h implies selective
reaction at the amine function, the absence of thioester products
might have been due in part to initial S-acylation followed by
rapid S to N migration as reported for thioesters of γ-lactones25
and hexahydrophthalic anhydride.24 Glutathione disulfide was
unreactive.
nucleophile
cysteine
NR-Ac-cysteine (16)
glutathione
productsb
conjugatesc
100
100
80
47
52
58
53
21
18
2 (1:2.6)
stable
unstable
unstable
stable
stable
stable
stable
stable
stable
2 (1:1.6)e
2 (1:12.3)
4 (2.3:9.5:1:1.3)
2 (9.7:1)e
2 (17.5:1)
2 (8.9:1)
lysine
NR-Ac-lysine (15)
NR-Ac-lysine Me ester
glycine
arginine
2 (3.4:1)
2 (2.2:1)
serinef
a Imide 4 (3 mM) was incubated with nucleophile (120 mM) in phosphate
buffer, pH 7.4, at 37 °C for 1 h. Incubations were analyzed by LC-MS
(MeCN-NH4OAc except for NR-Ac-lysine and NR-Ac-lysine Me ester,
which required MeCN-formic acid). b After 1-h incubation; estimated from
integration of UV (254 nm) chromatogram peaks of the conjugates and
isomeric carboxylic acids 2 and 5 (Figure 2), identified by LC-MS. c After
1-h incubation. The proportions of the isomeric conjugates were estimated
from integration of selected mass chromatogram peaks (m/z for [MH+]);
conjugates are given in order of elution from a reversed-phase column.
d Unstable conjugates were those that hydrolyzed extensively to carboxylic
acids 2 and 5 during the 15-21 h after analysis at 1 h. e Identification of
conjugates of 15 and 16 confirmed by comparison with standards 17a,b
and 18a,b, respectively (Scheme 4). The major amide under these conditions
was 17a (amide of 2); thioethers 18a and 18b were formed in approximately
equal amounts. f Two additional stable conjugates, less polar than those
found after 1 h, obtained after 20 h. Ratio of isomers at 20 h, 2.6:1:2.3:3.1.
Site-Selective Formation of Amide Adducts between Imide
4 and Proteins. HSA and RSA (15 µM) incubated with lower
concentrations of 4 (5 and 50 µM) did not yield sufficient
quantities of modified tryptic peptides for adequate characteriza-
tion by nano liquid chromatography tandem mass spectrometry
(nanoLC-MS/MS). At the end of a 16-h incubation of the highest
concentration of 4 (500 µM) with either HSA or RSA (15 µM),
all but minute traces of imide, detectable by LC-MS but not by
UV absorption, had been hydrolyzed to carboxylic acids 5 and
2 (1.5:1.0 and 1.1:1.0, respectively). The masses of hypothetical
HSA and RSA tryptic peptides modified by addition of 4 (321
amu) were calculated from the amino acid sequences of these
proteins in the Swiss-Prot database (entries P02768 and P02770,
respectively) minus the first 24 N-terminal residues not found
in the circulating proteins. First, it was established that peptides
Although amide conjugates had not been obtained from 4
and a 2-fold molar excess of NR-acetyl L-lysine 15 in aqueous
ethanol at pH 7, isomeric conjugates of 2 and 5 were generally
produced in good yield when the imide was incubated at pH
7.4 and 37 °C with a 40-fold excess of 15 (17a (major) and
17b (minor); Figure 4), NR-acetyl L-lysine methyl ester,
L-arginine, and L-glycine (Tables 1 and 2). L-Lysine was
monoacylated at both the NR- and Nꢀ-amine groups. No
diacylated L-lysine was found. NR-acetyl L-arginine was unre-
active. The resolution of two and four conjugates of L-serine
after 1 and 20 h, respectively, taken with the observation that

6170 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Meng et al.
Table 2. Analytical Characteristics of Conjugates Formed by the Reaction of Imide 4 with Amines and Thiolsa
nucleophile
cysteine
t
R conjugates (min)b
mass spectrumc
27.6, 32.6
m/z 443 ([MH+]), 322([MH - cysteine]+), 314 ([MH - AET]+),
296, 286, 280, 268, 262, 165, 130 ([AET+H]+)
NR-Ac-cysteine (16)
glutathione
lysine
NR-Ac-lysine (15)
NR-Ac-lysine Me ester
glycine
31.9, 34.8d
26.5, 29.2
m/z 485 ([MH+]), 356 ([MH - AET]+), 322, 146, 130
m/z 629 ([MH+]), 500 ([MH - AET]+), 322 ([MH - GSH]+), 130
21.6, 25.2, 26.2, 29.1
m/z 468 ([MH+]), 339 ([MH - AET]+), 322, 321, 293, 183, 165, 130
m/z 510 ([MH+]), 381 ([MH - AET]+), 363, 335, 322, 293, 189, 126
m/z 524 ([MH+]), 395 ([MH - AET]+), 363, 335, 322, 203
m/z 397 ([MH+]), 322, 268 ([MH - AET]+), 250, 240, 222, 194, 165,130
m/z 496 ([MH+]), 367 ([MH - AET]+), 322, 308, 210, 175, 130
m/z 427 ([MH+]), 322, 298 ([MH - AET]+), 280, 130
16.6, 19.0d
19.3, 21.7
24.8, 30.1
23.1, 27.7
23.1, 28.3, 29.5, 31.6
22.1, 27.2
arginine
serinee
histidine
m/z 477 ([MH+]), 348 ([MH - AET]+), 330, 322, 156
a Imide 4 (3 mM) was incubated with nucleophile (120 mM) in sodium phosphate buffer, pH 7.4, at 37 °C for 1-22 h. b Retention times on C-8 column
(MeCN-NH4OAc except for NR-Ac-lysine and NR-Ac-lysine Me ester, which required MeCN-formic acid gradient). c Electropspray MS. AET ) 2-amino-
5-ethyl-1,3,4-thiadiazole moiety (129 amu). Corresponding mass spectra of isomeric carboxylic acids 2 and 5: m/z 340 ([MH+]), 322, 294, 211 ([MH -
AET]+), 183, 165, 137, 130; and 4: m/z 322 ([MH+]), 294, 165, 156, 137, 130. d Conjugates of 15 were 17a (16.6 min; amide of 2) and 17b; conjugates
of 16 were 18a (34.8 min; thioester of 2) and 18b (Scheme 4). e The two less polar serine conjugates were seen after 20 h but not after 1 h.
Table 3. Tryptic Peptides of Modified HSA Identified as the Sites Acylated in Vitro by Imide 4a
modified peptideb
rt (min)c
m/z (predicted)d
m/z (found)d
residues
sequencee
modified lysine
MT1(H)
MT2(H)
MT3(H)
MT4(H)
MT5(H)
MT6(H)
MT7(H)
MT8(H)
45.2
26.3
37.5
48.2
42.1
32.1
45.6
33.1
688.9
534.8
634.8
670.9
981.0
580.2
725.5
569.8
688.8
535.0
635.1
671.1
981.4
580.5
725.7
570.0
137-144
191-197
198-205
210-218
414-428
433-439
525-534
539-545
K(+)YLYEIAR
K-137
K-195
K-199
K-212
K-414
K-436
K-525
K-541
ASSAK(+)QR
LK(+)C(Cm)ASLQK
AFK(+)AWAVAR
K(+)VPQVSTPTLVEVSR
VGSK(+)C(Cm)C(Cm)K
K(+)QTALVELVK
ATK(+)EQLK
a HSA (15 µM) was incubated with imide 4 (500 µM) in phosphate buffer, pH 7.4, at 37 °C for 16 h. Modified protein was reduced, alkylated,and
hydrolyzed with trypsin. Modified peptides were characterized by nanoLC-MS/MS. b MT(H) ) tryptic peptides of HSA modified by addition of 4. c Retention
times of modified peptides on a PepMap 3-µm C-18 Nano Series column. d m/z for [M + 2]2+ of the modified peptide arising from merging of two theoretical
tryptic peptides due to missed cleavage adjacent to the acylated lysine. e C(Cm) ) carboxyamidomethylated cysteine residue. K ) proposed site of acylation
by imide 4. (+) ) site of missed tryptic cleavage. Amino acid sequences are derived from the entry for HSA (ALBU-HUMAN; P02768) in Swiss-Prot
minus the first 24 N-terminal residues not found in the circulating protein. Residues are numbered for this sequence of 585 amino acids.
Table 4. Tryptic Peptides of Modified RSA Identified as the Sites Acylated in Vitro by Imide 4a
modified peptideb
rt (min)
m/z (predicted)d
m/z (found)d
residues
sequencee
modified lysine
MT1(R)
MT2(R)
MT3(R)
MT4(R)
MT5(R)
MT6(R)
MT7(R)
43.5
33.5
38.2
39.9
35.7
46.4
60.7
639.4
674.8
890.9
761.1
920.9
711.4
822.1
639.4
674.8
891.0
761.1
921.0
711.4
822.1
187-195
198-205
275-286
411-428
433-445
525-534
539-557
EK(+)ALVAAVR
K-188
K-199
K-281
K-414
K-436
K-525
K-545
MK(+)C(Cm)SSMQR
LQAC(Cm)C(Cm)DK(+)PVLQK
YTQK(+)APQVSTPTLVEAAR
VGTK(+)C(Cm)C(Cm)TLPEAQR
K(+)QTALAELVK
ATEDQLK(+)TVMGDFAQFVDK
a RSA (15 µM) was incubated with imide 4 (500 µM) in phosphate buffer, pH 7.4, at 37 °C for 16 h. Modified protein was reduced, alkylated, and
hydrolyzed with trypsin. Modified peptides were characterized by nanoLC-MS/MS. b MT(R) ) tryptic peptides of RSA modified by addition of 4. c Retention
times of modified peptides on a PepMap 3-µm C-18 Nano Series column. d m/z for [M + 2]2+ of the modified peptide arising from merging of two theoretical
tryptic peptides due to missed cleavage adjacent to the acylated lysine. e C(Cm) ) carboxyamidomethylated cysteine residue. K ) proposed site of acylation
by imide 4. (+) ) site of missed tryptic cleavage. Amino acid sequences are derived from the entry for RSA (ALBU-RAT; P02770) in Swiss-Prot minus
the first 24 N-terminal residues not found in the circulating protein. Residues are numbered for this sequence of 584 amino acids.
containing the only cysteine residue of HSA and RSA not
forming a disulfide bridge (cysteine-34) had not been modified
by 4. When attempting to predict the sites of proteolysis of
N-acylated proteins, and therefore the sequences of modified
peptides, it was assumed, as confirmed by numerous studies
on HSA modified by acylation or alkylation at many of its lysine
residues,9,26-32 that trypsin would be unable to hydrolyze the
peptide bond C-terminal to a lysine residue adducted by 4. The
simplest and commonest consequence of a missed cleavage at
a trypsin-resistant site is, in effect, to merge two consecutive
peptides by shifting the C-terminal cleavage to the next lysine
or arginine. Thereby the tryptic peptides obtained from a
modified protein either differ substantially from those obtained
from the native protein or have no counterparts.32 By this
method, eight and seven adducted peptides, all products of single
missed cleavages at lysine residues, were found using nanoLC-
MS/MS in digests of the modified HSA and RSA, respectively
(Tables 3 and 4).
For example, modification of lysine-199 in HSA merged the
theoretical dipeptide LK (residues 198-199), which is not
detected in digests of unmodified HSA,32 with CASLQK to give
peptide MT3(H). This particular effect on tryptic hydrolysis has
been obtained previously by acylation of lysine-199 with acyl
glucuronides,9,27 O-acyl ureas,28 a coenzyme A thioester,30
a
carboxylic anhydride,29 and benzylpenicillin.26 An analogous
modified neopeptide, MT2(R), was obtained from RSA. Con-
firmation of the modification sites by collision-induced (MS/
MS) sequencing was not possible because every acylated peptide
gave only [MH+] for the unmodified peptide as a product ion
and/or the corresponding b- and y-series sequence fragments.
Evidently the Nꢀ-lysylamide linkages were invariably weaker
than the peptide bonds. However, except for three peptides
[(MT1(H), MT5(H), and MT5(R)] that yielded only one of the
nonsequence fragments, additional evidence for the proposed
acylations was provided by two nonsequence product ions

Cyclization of the Acyl Glucuronide Metabolite
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 6171
37 °C for 3 h. This in itself indicates 3 had not undergone acyl
migration because C-2, -3, and -4 acyl glucuronides are typically
resistant to â-glucuronidase.33 The rapid degradation of 3 in
rat bile, which has a pH of approximately 7.5,34 was confirmed
by incubating synthetic glucuronide (2 mM) in pre-dose bile at
37 °C: after 1 h, it had degraded completely to 4, 2, and 5
(ratio 2:5, 0.9). Even in fully acidified bile at 5 °C, 3 underwent
extensive degradation to 4, 2, and 5, which was complete by
18 h. The only identified radiolabeled compounds in (unacidi-
fied) urine were the hydroxylated carboxylic acids, which even
in fully acidified bile were never found as glucuronides.
Imide 4 (50 µmol/kg) given iv to cannulated rats persisted
to the extent that 4.5 ( 1.9% and 11.0 ( 4.3% (mean ( SD, n
) 3) of the dose was recovered unchanged in bile over 1 and
5 h, respectively. Acids 2 and 5 and several of the hydroxylated
carboxylic acids that were metabolites of [14C]2 were also found
in bile. Only trace amounts of 4 were detected in bladder urine
by LC-MS at 5 h.
Figure 5. Radiochromatogram (reversed-phase HPLC) of the biliary
metabolites (0-1 h fraction) of [14C]2 (15 µmol/kg; labeled at C-2 of
the thiadiazole ring) in an anesthetized and cannulated adult male rat.
Bile was collected over glacial acetic acid to stabilize acyl glucuronides.
[O]CA ) hydroxylated carboxylic acids (tR 10-17 min). Acyl
glucuronide 3, carboxylic acid 2, and imide 4 were identified by
comparison (LC-MS) with authentic standards.
attributable to a glutaramide moiety: m/z 322 ([MH+] for 4)
and m/z 130 (thiadiazole ring).
Discussion
Metabolism of [14C]2 and Cyclic Imide 4 in Rats. Anaes-
thetized and cannulated adult male rats administered [14C]2 (15
µmol/kg; iv) labeled at C-2 of the thiadiazole ring eliminated
28.7 ( 5.3% and 75.1 ( 8.4% (mean ( SD, n ) 4) of the
radiolabel in bile over 1 and 5 h, respectively. Urine removed
from their bladders after 5 h contained 4.9 ( 3.1% of the dose.
Plasma samples obtained at the same time contained only 0.3-
0.4% of the dose per mL. When bile was collected hourly
without the acidification that stabilizes acyl glucuronide 3, the
radiolabeled material was found to be comprised principally of
carboxylic acids 2 and 5 (50% and 15% of chromatographed
radioactivity, respectively, in 0-1 h bile fraction), identified
by LC-MS, and a mixture of six or seven hydroxycarboxylic
acids (27% collectively, ca. 1-15% individually; tR 10-17 min;
m/z 356 for [MH+]). Except for the least polar isomer, which
yielded m/z 146 (130 + [O]) for the functionalized heterocycle,
all of the hydroxycarboxylic acids yielded m/z 130 and thereby
were modified outside the thiadiazole ring (Table 2). Imide 4
(6%) and traces of dihydroxycarboxylic acids (tR 5.5-8.5 min;
m/z 372) were also found in bile but no trace of acyl glucuronide
3. The hydroxylated and dihydroxylated metabolites were those
found previously in the plasma of rats administered 2.2 Bile
fully acidified by collection onto glacial acetic acid (final acid
concentration in the 1-h fraction, ca. 300 mM) contained
substantial amounts of 3 (65% chromatographed radioactivity;
0-1 h collection) and only small amounts of 4 (7%) and 2 (4%)
(Figure 5). Acid 5 was not seen in the first two hourly bile
collections and never accounted for more than approximately
1% of the radioactivity in later collections, suggesting any 4
produced within hepatocytes was not hydrolyzed extensively,
or if it was hydrolyzed, that 5 was transported selectively into
blood. As the hydroxylated and dihydroxylated carboxylic acids
were found in acidified bile from the first hour, it is likely they
were metabolites of 2 rather than both 2 and 5. In later bile
fractions, a fall in the proportions of 3 was counterbalanced
principally by a rise in the hydroxycarboxylic acids rather than
in 4, 2, or 5. Bile partially acidified by collection onto 10%
(v/v) acetic acid contained substantially less 3 (27-45%; 0-1
h collection) and correspondingly larger amounts of 4 (21-
35%) but still contained only small proportions of acids 2 (6-
9%) and 5 (1-2%); emphasizing the slowness of imide
hydrolysis relative to glucuronide cyclization. Acyl glucuronide
3 was hydrolyzed completely to 2, without formation of 4 or 5,
when fully acidified bile was incubated with H. pomatia
â-glucuronidase (Sigma) in 0.1 M sodium acetate, pH 5.0, at
The unusual instability of acyl glucuronide 3 is emphasized
by comparing the conjugate’s first-order half-life in phosphate
buffer, pH 7.4, at 37 °C (10.2 min) with the half-lives under
similar conditions of 31 structurally diverse acyl glucuronides
(0.26-79 h), 19 of which have half-lives g100 min and only
six-half-lives e30 min.4 None of these glucuronides is known
to degrade in aqueous solution except by acyl migration and
hydrolysis. The most reactive of them, tolmetin glucuronide (t1/2
16 min), irreversibly modifies human plasma protein in vivo35
and in vitro,27,36 but the mechanism of degradation of 3 is so
fundamentally different it is not possible to predict an extent
of protein modification from degradation rates, as it is when
comparing conventionally unstable acyl glucuronides.36 While
it may be presumed that the rapidity of intramolecular nucleo-
philic amidolysis simply precludes a slower intramolecular
migration, the complete stability of N-propyl 3 (acyl glucuronide
11) in phosphate buffer, pH 7.4, at 37 °C over 24 h implies 3
is intrinsically resistant to acyl migration. One potential explana-
tion is the bulky substitution pattern proximate to the aglycone’s
acid moiety: steric hindrance by alkyl substituents at the R-
and â-carbons has been invoked to rationalize the slower
rearrangement of certain acyl glucuronides.4
1
A recent H NMR study of the spontaneous degradation of
several acyl glucuronides at pH 7.4 and 37 °C reported only
very rapid hydrolysis in the case of 3.37 Consequently, it was
concluded that glucuronide metabolites such as 3 are likely to
be relatively unreactive in biological systems and thereby not
significant liabilities in drug development, because a dominant
hydrolysis process will minimize the (acyl migration) pathway
known to produce intermediates that retain the ability to react
covalently with proteins.4,6,7 The present finding that 3 cyclizes
to an electrophilic imide under these conditions emphasizes that
only a detailed deconstruction of an exceptional reactivity will
reveal the full potential for biochemical interactions.
The great rapidity of the cyclization of 3 except under
unphysiologically acidic conditions, which were essential for
definitive identification of the acyl glucuronide in bile, implies
that the conjugate may degrade partially within the liver and is
certain to degrade quantitatively in the alkaline conditions of
the small intestine. Therefore the mixture of carboxylic acids
and imide 4 found in unacidified bile may resemble the
metabolites to which the intestine is exposed in animals
administered carboxylic acid 2.
The only documented rearrangement of acyl glucuronides
bearing any resemblance to cyclization of 3 is a rapid cyclization

6172 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Meng et al.
at pH values >7 that forms δ-lactones from biosynthetic
glucuronides of statin derivatives with 2,4-dihydroxy heptanoic
acid side chains.38 As was the case with 3, the glucuronide of
hydroxy simvastatin does not appear to form acyl migration
products and degrades rapidly to the δ-lactone following its
elimination in rat bile. The biosynthetic acyl-CoA thioester of
hydroxy simvastatin undergoes an analogous lactonization.39
tively, are acylated; four are common to both proteins. Predicting
in vivo protein modifications from in vitro data is currently
highly problematical. Limited studies on benzylpenicillin indi-
cate the same set of HSA lysines is modified in vitro and in
vivo,26 and there is at least a partial overlap in the case of
glucose.45 Protein modification by 4 in vivo remains a possibil-
ity: the occurrence of 5 as the major plasma metabolite of 220,46
in itself implies formation of 4 in vivo; 4 has been found as a
urinary metabolite and minor plasma metabolite of 2.46 The
present studies have shown that the imide can pass from blood
into and through the liver in rats.
An alternative mechanism of cyclization of 3 involving
nucleophilic attack by the amide carbonyl rather than the
nitrogen, with consequential loss of the heterocycle, would
generate initially an imino-ether and then a substituted glutaric
anhydride, and finally, upon hydrolysis, 2-cyclopentyl-4-propyl-
glutaric acid. This mechanism is analogous to the intramolecular
reaction between the carboxamide and epoxide functions of
allylisopropylacetamide epoxide, which produces a stable γ-bu-
tyrolactone, hypothetically via an electrophilic iminolactone.40
Nevertheless, the extensive degradation of 3 in aqueous solution
at pH 7.4 did not yield any 2-cyclopentyl-4-propyl-glutaric acid
detectable by LC-MS.
Conclusion
Although the generic instability of acyl glucuronides has been
analyzed in great detail,4 glucuronide 3 of glutaramide 2 has
now been shown to undergo an unprecedented and rapid
eliminative cyclization (decomposition half-life, 10.2 min) to
the apparent exclusion of the familiar intramolecular acyl
migration. Reactivity of 3 was characterized using glucuronide
prepared efficiently by selective acylation of benzylglucuronate
with N-Bom-protected 2 obtained by a three-step, one-pot
synthesis. The association between exceptional acyl glucuronide
instability and a combination of low amide pKa and formation
of a stable six-membered ring immediately suggested structural
modifications for stabilizing glucuronides of series 1 glutara-
mides by retarding the intramolecular nucleophilic amidolysis.
Pryde et al. 1b,5 have shown that monocarboxylic glutaramides
possessing nonaromatic amide substituents have higher NH pKa
and yield acyl glucuronides with greatly enhanced stability. A
potent NEP inhibitor from this second group, the 4-chlorophen-
propylamide analogue of 2 (S-30), having an estimated pKa of
16,5b was selected as a clinical candidate1b. The decomposition
half-life of S-30 in phosphate buffer, pH 7.4, at 37 °C is 51
h;46 the longest documented half-life of an acyl glucuronide
under comparable conditions is the exceptional 79 h of the
valproic acid conjugate.4 The decomposition half-life of 3 in
rat, dog, and human blood ex vivo at 37 °C is approximately 5
min1b, whereas the acyl glucuronide of S-30, under these
conditions, does not degrade for at least 1 h1b. Additionally,
the cyclic imide of S-30, unlike 4, is resistant to hydrolysis at
pH 7.4 and does not react spontaneously with glutathione in
aqueous solution to form a thioester.46 While it is altogether
premature to draw any conclusions with respect to a possible
causal linkage between the stability/reactivity of these glutara-
mide acyl glucuronides and the parent drugs’ toxicity in vivo,
it is notable that S-30 is free of the acute intestinogastric toxicity,
including especially hemorrhagic necrosis of the mucosal
epithelium, that was associated with 2 in dogs.5b,46
Notwithstanding the rapid cyclization of 3 under physiological
conditions and the reactivity of imide 4, contributory conjugate
formation by direct reactions between 3 and amine or thiol
nucleophiles cannot be excluded. However, this is unlikely to
have a material effect because the amide and thioester conjugates
obtained from 3 were identical to those of 4. Simple carboxylic
acids can be eliminated by rats as thioester conjugates,41,42 but
the instability of thioesters 18a/18b and their glutathione
analogues at pH 7.4 implied that no thioester conjugates of acids
2 or 5 produced in vivo, by either direct reaction or transacy-
lation of the abundant metabolite 3,41 would be sufficiently
persistent to allow detection in biological media ex vivo. In
fact, using the synthetic conjugates of cysteine, lysine, NR-
acetylated cysteine and lysine, and glutathione as standards, no
thioester or amide metabolites of either [14C]2 or 4 were found
in rat bile or urine.
Serum albumins, and particularly HSA, which has 125
nucleophilic side chains, are well-documented targets for low
molecular weight electrophiles. HSA and RSA, when incubated
with the highest concentration of 4 (500 µM), underwent Nꢀ-
acylations of lysine residues that were abundant and stable
enough to be characterized in tryptic peptides of the denatured
and reduced proteins. From the reactions of 4 with NR-acetyl
L-lysine, it is likely the adducts were Nꢀ-lysylamides of both 5
and 2. The absence of an acylated cysteine residue was
predictable from the instability of NR-acetyl L-cysteine conju-
gates (18a and 18b) and glutathione conjugates of 2 and 5 and
conforms with the behavior of hexahydrophthalic anhydride,
which produces stable derivatives of HSA lysines but not
cysteine-34.24,29 All eight of the modified HSA lysines are
known, variously, as sites of acylation by acyl glucuronides,9,27
a coenzyme A thioester,30 a carboxylic anhydride,29 and
benzylpenicillin26 in vitro. However, only lysines-195, -199, and
-541 are commonly encountered as targets of acylation, and
perhaps only lysine-199, which has a pKa estimated to be 7.47,43
is almost invariably modified. It is also the site of acetylation
by aspirin44 and a lesser site of glycation by glucose.45 Albeit
from a very limited database,9,26-32 the set of eight HSA lysines
modified by 4 most closely resembles the six adducted by
tolmetin acyl glucuronide (five residues are modified by both
compounds).27 Selective derivatization of RSA lysines has not
been characterized in comparable detail. The fraction of lysines
modified by 4 is similar: seven (RSA) and eight (HSA) from
53 and 59 residues, respectively. Of the 40 homologous lysines
(Swiss-Prot database) six and five in RSA and HSA, respec-
Experimental Section
Chemical Procedures. Organic extracts were washed finally
with satd aq NaCl and dried over anhydrous Na2SO4 prior to rotary
evaporation at <30 °C. Analytical thin-layer chromatography was
performed using Merck Kieselgel 60 F 254 silica plates. Preparative
column chromatography was performed on Merck 938S silica gel.
Rotations were measured on an Optical Activity polarimeter
operating at the wavelength of the sodium D-line. Infrared spectra
were recorded using a Perkin-Elmer RX1 FTIR instrument, for the
1
physical forms noted. Unless otherwise stated, H and 13C NMR
spectra were recorded on CDCl3 solutions using either Bruker 250
or 400 MHz (100 MHz for 13C) instruments with tetramethylsilane
as internal standard. Both low- and high-resolution mass spectra
were obtained by direct injection of sample solutions into a
Micromass LCT mass spectrometer operated in the electrospray
mode, +ve or -ve ion as indicated. CI mass spectra (NH3) were
obtained on a Fisons Instruments Trio 1000. The pKa of acid 2 were

Cyclization of the Acyl Glucuronide Metabolite
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 6173
determined at 20 °C by pH-metric titration using a GL pKa titrator
(Sirius Analytical Instruments, Forest Row, East Sussex, UK).
Duplicate measurements were performed with 2 (0.5 mM) in H2O/
MeOH/DMSO (60:30:10, v/v).
t, J ) 7.2 Hz), 1.21-1.65 (13 H, m), 1.95 (1 H, dd, J ) 13.8 and
3.2 Hz), 2.17 (1 H, dd, J ) 13.9 and 9.0 Hz), 2.26 (2 H, m), 2.88
(2 H, m), 4.73 (2 H, m, PhCH2O), 5.75 (2 H, s, NCH2O) and 7.35
(5 H, m, ArH); m/z (CI, NH3) 460 ([MH+], 100%).
(3S,4S,5S,6S)-2-Propenyl 6-[(R)-2-[(1-(5-ethyl-1,3,4-thiadiazol-
2-yl)carbamoyl]cyclopentyl]methyl]pentanoyloxy]-3, 4, 5-trihy-
droxytetrahydro-2H-pyran-2-carboxylate (7). Via the Mitsuno-
bu Reaction. DIAD (0.10 mL, 0.5 mmol) was added over 10 min
to a solution of (2R)-2-[(1-{[(5-ethyl-1,3,4-thiadiazol-2-yl)amino]-
carbonyl}cyclopentyl)methyl]pentanoic acid 2 (0.170 g, 0.5 mmol),
triphenylphosphine (0.135 g, 0.5 mmol), and allyl-R,â-D-glucuronate
6 (0.12 g, 0.5 mmol) in anhydrous THF (3 mL) which was stirred
under N2 at 0 °C. After 1 h the solution was evaporated to dryness
and then purified first by silica chromatography, eluting with 5%
MeOH in CH2Cl2. Appropriate fractions were pooled and evapo-
rated to give the product as an R/â mixture. Further purification
by preparative HPLC (Luna 10 µm C-18 column, 250 × 21.2 mm,
eluting with 30-50% MeCN in H2O over 30 min) afforded on
evaporation of appropriate fractions (tR ) 29.3 min) the desired 7
(3S,4S,5S,6S)-6-[(R)-2-[[1-(5-Ethyl-1,3,4-thiadiazol-2-yl)car-
bamoyl]cyclopentyl]methyl]pentanoyloxy]-3,4,5-trihydroxytet-
rahydro-2H-pyran-2-carboxylic Acid (3). A mixture of compound
14 (0.142 g, 0.2 mmol; see Supporting Information) and 10% Pd-C
(0.106 g) in glacial AcOH (4 mL) was stirred under an H2 balloon
at 20 °C for 18 h. The mixture was filtered through Celite, the
solids were washed with a small quantity of AcOH, then the
combined filtrate and washings were evaporated to give a colorless
oil which was triturated with acetone to afford the product 3 as a
1
white solid (0.070 g, 70%); H NMR (400 MHz, CD3OD): 0.83
(3 H, t, J ) 7.1 Hz), 1.20-1.61 (13 H, m), 2.00 (1 H, dd, J ) 13.8
and 3.2 Hz), 2.17 (1 H, dd, J ) 13.9 and 9.0 Hz), 2.26 (2 H, m),
2.88 (2 H, m), 3.30-3.50 (3 H, 3 m, 3-H, 4-H and 5-H), 3.65 (1
H, brs, 2-H) and 5.43 (1 H, d, J ) 7.5 Hz, 6-H); 13C NMR (100
MHz, CD3OD) 14.4, 21.3, 23.2, 24.9, 25.2, 25.4, 26.7, 35.9, 36.8,
37.7, 38.7, 41.2, 44.1, 56.4, 73.4, 73.9, 77.9, 95.8, 169.1, 176.4,
177.7, 179.8 and 179.8; LC-MS, single peak (20 to 50 to 70%
MeCN in 5 mM NH4OAc, 30 and 5 min, respectively; tR ) 18.5
min); m/z (ES +ve mode) 516 ([MH+], 100%), (ES-ve mode) 514
[M - H]-. High-resolution mass (ES-ve mode) calculated for
C22H32N3O9S [M - H]- 514.1859. Found: 514.1843.
1
as a single â-anomer (foam, 0.081 g, 30%); H NMR (400 MHz,
CD3OD): 0.84 (3 H, t, J ) 7.3 Hz), 1.4-1.7 (13 H, m), 2.00 (1 H,
dd, J ) 13.8 and 3.2 Hz, 1′-H), 2.20-2.40 (3 H, m), 3.09 (2 H, q,
J ) 7.3 Hz), 3.40-3.60 (3 H, m, 3-H, 4-H and 5-H), 4.00 (1 H, d,
J ) 9.7 Hz, 2-H), 4.68 (2 H, d, J ) 5.4 Hz), 5.30 (2 H, m, CH2d
CHCH2), 5.45 (1 H, d, J ) 7.8 Hz, 6-H) and 5.75-6.09 (1 H, m,
CH2dCHCH2); 13C NMR (100 MHz, CD3OD) 13.5, 19.8, 22.8,
23.4, 23.8, 33.6, 34.2, 35.6, 36.9, 39.9, 42.5, 55.1, 65.4, 71.4, 72.1,
75.7, 76.1, 94.3, 117.4, 132.0, 159.4, 167.0, 168.1, 174.8 and 175.7.
LC-MS (30-50% MeCN in 5 mM NH4OAc, 20 min; tR ) 15.5
min); m/z (ES +ve mode) 556 ([MH+], 100%).
(R)-7-(5-Ethyl-1,3,4-thiadiazol-2-yl)-9-propyl-7-azaspiro[4.5]-
decane-6,8-dione (4). N-Methylmorpholine (0.066 mL, 0.6 mmol)
and DIC (0.1 mL, 0.66 mmol) were added successively to a mixture
of compound 2 (0.203 g, 0.6 mmol) and 1-hydroxybenzotriazole
(0.027 g, 0.2 mmol) in CH2Cl2 (3 mL) which was stirred at 0 °C.
The resulting solution was allowed to attain 20 °C, and then after
1 h it was concentrated and directly purified by chromatography,
eluting with 20% EtOAc-hexane. Evaporation of appropriate
fractions afforded the product 4 as a white solid (0.160 g, 83%);
General Procedure for Acyl Glucuronide Formation by
Selective Acylation. A mixture of glucuronate monoester 6 or 13
(0.5 mmol), carboxylic acid (0.5 mmol), HATU (0.19 g, 0.5 mmol),
and 1,4-diazabicyclo[2.2.2]octane (0.112 g, 1 mmol) or NMM (0.11
mL, 1 mmol) in anhydrous MeCN (5 mL) was stirred under nitrogen
at 20 °C. The reaction was monitored by TLC (10% EtOH-CH2-
Cl2), and after 3 h it was quenched by adding Amberlyst A-15 (H+,
1 mmol) and stirring for 30 min. The beads were filtered off, and
the residue plus an MeCN washing was evaporated to dryness
followed by chromatography, eluting with 7% EtOH-CH2Cl2. Ap-
propriate fractions were concentrated to give the product 10 or 14.
(3S,4S,5S,6S)-2-Propenyl 6-[(R)-2-[[1-(2-Propenyl)(5-ethyl-
1,3,4-thiadiazol-2-yl)carbamoyl]cyclopentyl]methyl]pentanoyloxy]-
3,4,5-trihydroxytetrahydro-2H-pyran-2-carboxylate (10). This
was prepared by the general method and obtained as a foam in
65% yield; 1H NMR (400 MHz, CD3OD) 0.79 (3H, t, J ) 7.3 Hz),
1.20-1.60 (13 H, m), 1.96 (1 H, dd, J ) 13.8 and 3.2 Hz), 2.10 (1
H, dd, J ) 13.8 and 9.0 Hz), 2.26 (1 H, m), 2.32 (1 H, m), 2.92 (2
H, m), 3.30-3.70 (3 H, 3 m, 3-H, 4-H and 5-H), 4.02 (1 H, d, J )
9.5 Hz, 2-H), 4.65 (2 H, m), 4.94 (2 H, m), 5.17-5.37 (4 H, m,
CH2dCHCH2), 5.50 (1 H, d, J ) 8.1 Hz, 6-H) and 5.75-6.09 (2
H, m, CH2dCHCH2); 13C NMR (100 MHz, CD3OD) 12.2, 13.4,
20.0, 23.7, 23.8, 24.7, 33.7, 36.1, 39.0, 41.6, 42.9, 52.4, 56.4, 65.1,
71.7, 72.7, 76.0, 76.2, 94.4, 117.1, 117.9, 131.6, 132.2, 160.3, 165.1,
167.9, 174.8 and 186.4. m/z (ES +ve mode) 618 ([MNa+], 100%).
High-resolution mass (ES +ve mode) calculated for C28H41N3O9-
SNa ([MNa+]): 618.2461. Found: 618.2431.
(R)-2-[[1-[Benzyloxymethyl](5-ethyl-1,3,4-thiadiazol-2-yl)car-
bamoyl)]cyclopentyl]methyl]pentanoic Acid (12). A solution of
TBSCl (0.393 g, 2.5 mmol) in CH2Cl2 (2 mL) was added to a
solution of compound 2 (0.339 g, 1 mmol) in CH2Cl2 (5 mL) and
triethylamine (0.35 mL. 2.5 mmol) which had been stirred for 0.5
h at 25 °C. After 20 h, Bom-Cl (0.232 mL, 1 mmol) was added
and stirring was continued for another 6 h. The solvents were
evaporated, and the residue was redissolved in THF (10 mL)
followed by addition of 1 M TBAF in THF (2.5 mL) at 0 °C. After
1 h water (20 mL) was added followed by extraction with EtOAc
(3 × 15 mL). The combined organic extracts were washed with
satd aq NaHCO3, and then standard workup followed by evaporation
afforded the product 12 as an oil (0.367 g, 80%); Anal.
(C24H33N3O4S) C, H, N. 1H NMR (400 MHz, CDCl3): 0.83 (3 H,
1
Anal. (C16H23N3O2S) C, H, N. H NMR (400 MHz, CDCl3) 0.96
(3H, t, J ) 7.2 Hz), 1.45-1.60 (7 H, m), 1.65-2.10 (9H, m), 2.47
(1H, m), 2.77 (1H, m) and 3.16 (2H, q, J ) 7.0 Hz); m/z (CI, NH3)
322 ([MH+]).
Isomeric Acid 5. R-1-(2-tert-Butoxycarbonylpentyl)cyclopen-
tanecarboxylic acid1 (513 mg, 1.80 mmol) was dissolved in
methanol (5 mL), and cesium carbonate (294 mg, 0.9 mmol) added
in one portion with stirring. After 5 min, the mixture was evaporated
to low volume and azeotroped with toluene (2 mL). The residue
was redissolved in DMF (5 mL) under nitrogen, benzyl bromide
(225 µL, 1.83 mmol) added in one portion, and the whole stirred
at room temperature for 3 h. The mixture was poured into EtOAc
(40 mL) and washed with water (40 mL), 1 M HCl (20 mL) and
water again (20 mL). The resulting solution was dried over MgSO4
and evaporated to a thick oil, which was purified by column
chromatography using pentane-EtOAc (2:1, v/v) as eluent to
provide 430 mg of a clear oil. This was dissolved in TFA (2 mL),
left at room temperature for 16 h, and then evaporated to dryness
in vacuo to provide 353 mg of a thick oil. This oil was dissolved
in acetonitrile (5 mL), and 2-amino-5-ethyl-1,3,4-thiadiazole (150
mg, 1.15 mmol), ethyl dimethylaminopropylcarbodiimide HCl (255
mg, 1.23 mmol), 1-hydroxybenzotriazole (173 mg, 1.23 mmol),
and N-methylmorpholine (243 µL) were added in sequence. The
whole was stirred at room temperature for 16 h and at 80 °C for 3
h. It was allowed to cool to room temperature, evaporated to
dryness, redissolved in EtOAc (10 mL), and washed with saturated
NaHCO3 solution. The organics were dried over MgSO4, filtered,
and evaporated to a thick oil, which was purified by column
chromatography using pentane-EtOAc (2:1, v/v) to provide 430
mg of a clear oil. This oil was dissolved in EtOH-MeOH (5:1,
v/v), palladized charcoal (45 mg) added and the whole stirred under
hydrogen (30 psi) for 2 h. The catalyst was removed with a short
plug of celite, and the solution was evaporated to a clear oil. Column
chromatography using a gradient of DCM in MeOH (19:1 to 9:1,
v/v) provided the title compound (120 mg, 20%) as a clear oil.
Anal. (C16H25N3O3S) C, H, N. 1H NMR (400 MHz, CDCl3): 0.89
(3H, t, J ) 6.4 Hz), 1.16-1.86 (13H, m, including 1.38, 3H, t, J

6174 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Meng et al.
) 7.5 Hz), 1.91-2.03 (1H, m), 2.22-2.40 (1H, m), 2.42-2.73
(2H, m), 2.95 (2H, dq, J ) 7.5, 2.4 Hz); m/z (ES+ve mode) 340
([MH+]).
µM) in deuterated phosphate buffer (100 mM), pD ca. 7.2, at
20 °C was confirmed by NMR spectroscopy using 1H and 1D and
2D TOCSY methods. Data were acquired on a Varian Unity Inova
500 MHz spectrometer equipped with a Narolac 3-mm inverse
detection probe.
Reactions of Acyl Glucuronide 3 and Cyclic Imide 4 with
Amino Acids and Glutathione. Acyl glucuronide 3 (2 mM) and
either NR-acetyl L-lysine 15, N-acetyl L-cysteine 16, or glutathione
(80 mM) in 0.5 mL of 100 mM sodium phosphate, pH 7.4, were
incubated at 37 °C for 1-20 h in 1.5 mL polypropylene microtubes.
Imide 4 (3 mM; 1.0 mL) and either glutathione, glutathione
disulfide, or an amino acid (120 mM; Table 1) were incubated under
the same conditions. Aliquots (50-100 µL) were removed at
intervals and analyzed by LC-MS.
Nr-Acetyl L-Lysine Adducts (17a, 17b). A mixture of imide 4
(0.128 g, 0.4 mmol) and NR-acetyl L-lysine 15 (0.150 g, 0.8 mmol)
was stirred in EtOH (2 mL) and sodium tetraborate buffer, pH 9.22
( 0.02, (BDH buffer tablet, 25 mM Na2B4O7 with 1.7 mM NaCl;
2 mL) at 37 °C for 24 h. HCl (1 M) was added to give a pH of 3,
and then the product was extracted with EtOAc (3 × 15 mL) and
worked up in the usual manner. The crude product was separated
by preparative HPLC (Luna 10 µm C-18 column, 250 × 21.2 mm,
eluting with 30 to 50 to 70% MeCN in 0.1% aq formic acid over
20 and 10 min, respectively) to afford 17a and 17b (combined yield
60%) having tR 29.9 and 31.0 min, respectively. For 17a (major
isomer; amide of 2): Anal. (C24H39N5O5S‚2H2O) C, H, N. 1H NMR
(400 MHz, CD3OD) 0.74 (3H, t, J ) 7.1 Hz), 1.05-1.42 (11H,
m), 1.45-1.62 (7H, m), 1.72 (2H, m), 1.89 (3H, s, CH3CO), 2.02-
2.22 (4H, m), 2.81-2.98 (4H, m) and 4.24 (1H, m, NHCHCO2H);
LC-MS (20 to 50 to 70% MeCN in 0.1% aq formic acid, 30 and 5
min, respectively) tR 16.6 min; m/z (ES+ve mode) 510 ([MH+],
100%). For 17b (minor isomer; amide of 5): 1H NMR (400 MHz,
CD3OD) 0.80 (3H, t, J ) 7.3 Hz), 1.12-1.60 (18H, m), 1.70 (2H,
m), 1.87 (4H, m), 1.99 (1H, m), 2.10 (1H, m), 2.45 (1H, m) 2.90-
3.15 (4H, m) and 4.24 (1H, m, NHCHCO2H); LC-MS (20 to 50 to
70% MeCN in 0.1% aq formic acid, 30 and 5 min, respectively) tR
19 min; m/z (ES+ve mode) 510 ([MH+], 100%). Ratio 17a to 17b
(254 nm): 3.0.
Nr-Acetyl L-Cysteine Adducts (18a, 18b). Triethylamine (0.17
mL, 1.2 mmol) was added to a mixture of imide 4 (0.096 g, 0.3
mmol), N-acetyl L-cysteine 16 (0.098 g, 0.6 mmol) and a trace of
4-dimethylaminopyridine in anhydrous MeCN (3 mL) with stirring
at 20 °C. After 4 h, when reaction appeared complete by TLC (60%
EtOAc-hexane), the mixture was evaporated to dryness and
purified by preparative HPLC (Luna 10 µm C-18 column, 250 ×
21.2 mm, eluting with 50-70% MeCN in 0.1% aq formic acid
over 20 min) to afford 18a and 18b (combined yield 42%, ratio
3:1, w/w) having tR 13.5 and 10.3 min, respectively. For 18a (major
isomer; thioester of 2): 1H NMR [400 MHz, (CD3)2CO] 0.82 (3H,
t, J ) 7.4 Hz), 1.26-1.72 (13H, m), 1.96 (3H, s, CH3CO), 2.03
(1H, s), 2.32-2.43 (3H, m), 2.60 (1H, m), 3.01-3.18 (3H, m),
3.43 (1H, m) and 4.65 (1H, m, NHCHCO2H); LC-MS (20 to 50 to
70% MeCN in 0.1% aq formic acid, 30 and 5 min, respectively),
tR 27.5 min; m/z (ES+ve mode) 485 ([MH+], 100%). High-
resolution mass (ES-ve mode) calculated for C21H31N4O5S2 [M -
H]-: 483.1736. Found: 483.1749. For 18b (minor isomer; thioester
of 5): 1H NMR [400 MHz, (CD3)2CO] 0.90 (3H, t, J ) 7.2 Hz),
1.22-1.68 (13H, m), 1.98 (2H, m), 2.10-2.20 (2H, m), 2.30
(3H, s, CH3CO), 2.70 (1H, m), 3.04 (2H, q, J ) 7.0 Hz), 3.19 (1H,
dd, J ) 13.9 and 8.0 Hz), 3.46 (1H, dd, J ) 13.9 and 4.6 Hz) and
4.82 (1H, m, NHCHCO2H); LC-MS (20 to 50 to 70% MeCN in
0.1% aq formic acid, 30 and 5 min, respectively), tR 23.4 min; m/z
(ES+ve mode) 485 ([MH+], 100%). Ratio 18a to 18b (254 nm):
2.9.
Covalent Modification of Serum Albumin by Cyclic Imide
4. Imide 4 (final concentration, 5, 50, or 500 µM) dissolved freshly
in either DMSO or acetonitrile (5 µL) was added to a solution (1
mg/mL; 15 µM; 0.5 mL) of either HSA or RSA (approx 99% pure,
essentially globin free and fatty acid free; Sigma, UK) in 100 mM
sodium phosphate, pH 7.4. The mixture was incubated in polypro-
pylene microtubes at 37 °C for 16 h. An aliquot (50 µL) was taken
for LC-MS analysis. Disulfide bridges in a second aliquot of protein
were reduced with dithiothreitol (final concentration, 1 mM; room
temperature, 15 min) and the free thiols carboxyamidomethylated
with iodoacetamide (final concentration, 55 mM; room temperature,
15 min). The solution of protein thus modified (1 µL) was diluted
with 50 mM ammonium hydrogen carbonate, pH 7.6, (1:10, v/v)
and digested with 50 ng of sequencing grade modified trypsin
(Promega UK, Southampton, Hampshire, UK) at 37 °C for 16 h.
The tryptic digest (3 µL injection volume; equivalent to 0.3 µg of
protein) was chromatographed on a PepMap 3 µm C-18 Nano Series
column (15 cm × 75 µm; LC Packings, Amsterdam, Netherlands)
preceded by a 5 mm C18 µ-precolumn cartridge. A gradient from
5% acetonitrile/0.05% TFA (v/v) to 48% acetonitrile/0.05% TFA
(v/v) in 60 min was applied at a demanded flow rate of 200 nL/
min using an UltiMate capillary LC system (Dionex Corp,
Sunnyvale, CA). Eluted peptides were delivered to a QSTAR Pulsar
i mass spectrometer (Applied Biosystems, Foster City, CA) via a
10 µm i.d. PicoTip nanoelectrospray emitter (New Objective,
Woburn, MS), and spectra were acquired in positive-ion mode
through information-dependent acquisitions controlled by Analyst
software as described previously.47
Metabolism of [14C]2 and Cyclic Imide 4 in Rats. Adult male
Wistar rats were anesthetized with urethane (1.4 mg/mL in isotonic
saline; 1 mL/kg, ip). Cannulae were inserted into the trachea, jugular
vein, and common bile duct, and the penis was ligated. Either [14C]-
2 (15 µmol/kg, 10 µCi; in 200 µL of DMSO) radiolabeled at C-2
of the thiadiazole ring (18 mCi/mmol; radiochemical purity, 98%
as determined by HPLC) or 4 (50 µmol/kg; in 300 µL of DMSO),
both dissolved in DMSO immediately before administration, was
injected over 20 min via the jugular vein. Bile was collected into
polypropylene microtubes hourly for 5 h. When rats were given
[14C]2, the bile fractions were in some cases acidified as they
collected in order to stabilize acyl glucuronide metabolites for
characterization. For acidification, the tubes contained either 15
µL of glacial acetic acid or 10 µL of water-acetic acid (9:1, v/v).
Urine was removed from the bladder at the end of each 5-h
experiment without acidification, and a blood sample was taken at
the same time. Bile (20 µL), blood plasma (50 µL), and urine (20
µL) were sampled for radioactivity and stored at - 20 °C.
HPLC Assay of Cyclic Imide 4. Bile (30-100 µL) from rats
administered 4 (50 µmol/kg) was chromatographed at room
temperature on a Symmetry 5 µm C-8 column (3.9 mm × 150
mm; Waters Corp, Milford, MS) using a gradient of acetonitrile
(10-30% over 30 min, 30-70% over 5 min) in 10 mM ammonium
acetate. The flow rate was 0.9 mL/min. The LC system consisted
of two Jasco PU980 pumps (Jasco UK, Great Dunmow, Essex, UK)
and a Jasco HG-980-30 mixing module. Eluted 4 (38 min) was
identified by cochromatography with authentic compound and
matching of electrospray mass spectra, and its absorbance measured
at 254 nm with a Jasco UV-975 spectrophotometer. A linear (r2 )
Stability of Acyl Glucuronide 3 and Cyclic Imide 4 in
Aqueous Solution. Degradation half-lives of 3 (39 µM) and 4 (62
µM) in 100 mM sodium dihydrogen phosphate buffer, pH 7.4, were
determined at 37 °C. Aliquots (100 µL) were removed at 10-min
intervals for 60 min and thereafter at 90 min, 3, 5, 7, and 27 h, and
mixed immediately with 1% trifluoroacetic acid (20 µL). Acidified
solution (80 µL) was eluted from a Hypersil HS C-18 column (50
mm × 4.4 mm; Thermo Fisher Scientific, Waltham, MA) with a
gradient of acetonitrile (25% for 3 min, 25% to 90% over 6 min,
90% for 5 min) in 4 mM ammonium formate, pH 3.5, delivered
by a Shimadzu 10A HPLC system (Shimadzu UK, Milton Keynes,
Buckinghamshire, UK). 3 (3.8 min), 2 (8.8 min), 5 (9.4 min), and
4 (11.5 min) were detected at 254 nm and identified by cochro-
matography with authentic compounds and matching of electrospray
product-ion mass spectra. Peak areas of 3 and 4 (λ ) 254 nm) in
sequential aliquots were integrated and subjected to linear regression
analysis (r2 ) 0.99) to obtain degradation half-lives (t1/2 ) ln2/
slope). The absence of acyl migration during degradation of 3 (39

Cyclization of the Acyl Glucuronide Metabolite
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24 6175
(4) Stachulski, A. V.; Harding, J. R.; Lindon, J. C.; Maggs, J. L.; Park,
B. K.; Wilson, I. D. Acyl Glucuronides: Biological Activity,
Chemical Reactivity and Chemical Synthesis. J. Med. Chem. 2006,
49, 6931-6945.
0.99) calibration graph (10 - 75 nmol) was constructed from an
acetonitrile solution of authentic imide (1 mM). Peak areas of 4
were computed using MassLynx 3.5 integration software (Waters
Corp, Manchester, UK). The intraassay coefficient of variation was
10% (n ) 6).
(5) (a) See Pryde, D. in Armer, R.; Warne, P.; Witherington, J. Recent
Disclosures of Clinical Drug Candidates. Small-Molecule NEP
Inhibitors for the Treatment of Sexual Dysfunction. Drug News
Perspect. 2006, 19, 67. (b) See Middleton, D. in Heap, C. R.;
Pechulis, A. D.; Kirova-Snover, M.; Yang, Z. Designing Small
Molecular NEP Inhibitors for the Treatment of FSAD. Presented at
the 30th American Chemical Society National Medicinal Chemistry
Symposium, June 25-29, 2006, Seattle, WA. Technical Reports
2006, 11 (10), 10-11, Albany Molecular Research Inc., Albany, NY.
(6) Wang, J.; Davis, M.; Li, F.; Azam, F.; Scatina, J.; Talaat, R. A Novel
Approach for Predicting Acyl Glucuronide Reactivity via Schiff Base
Formation: Development of Rapidly Formed Peptide Adducts for
LC/MS/MS Measurements. Chem. Res. Toxicol. 2004, 17, 1206-
1216.
(7) Bailey, M. J.; Dickinson, R. G. Acyl Glucuronide Reactivity in
Perspective: Biological Consequences. Chem.-Biol. Interact. 2003,
145, 117-137.
(8) Boelsterli, U. A. Xenobiotic Acyl Glucuronides and Acyl CoA
Thioesters as Protein-reactive Metabolites with the Potential to Cause
Idiosyncratic Drug Reactions. Curr. Drug Metab. 2002, 3, 439-
450.
(9) Qiu, Y.; Burlingame, A. L.; Benet, L. Z. Mechanisms for Covalent
Binding of Benoxaprofen Glucuronide to Human Serum Albumin.
Studies by Tandem Mass Spectrometry. Drug Metab. Dispos. 1998,
26, 246-256.
(10) Stachulski, A. V.; Jenkins, G. N. The Synthesis of O-Glucuronides.
Nat. Prod. Rep. 1998, 15, 173-188.
(11) Juteau, H.; Gareau, Y.; Labelle, M. A Convenient Synthesis of â-Acyl
Glucuronides. Tetrahedron Lett. 1997, 38, 1481-1484.
(12) Perrie, J. A.; Harding, J. R.; Holt, D. W.; Johnston, A.; Meath, P.;
Stachulski, A. V. Effective Synthesis of 1â-Acyl Glucuronides by
Selective Acylation. Org. Lett. 2005, 7, 2591-2594.
LC-MS and Radiochromatographic Analyses. Aliquots (10-
100 µL) of aqueous chemical reaction mixtures and solutions of
synthetic compounds (1-2 mg/mL) were chromatographed on a
Symmetry 5 µm C-8 column as described above unless otherwise
indicated. Synthetic compounds were dissolved in acetonitrile except
for glucuronides and the adducts of NR-acetyl L-lysine and NR-
acetyl L-cysteine, which were dissolved in acetonitrile-1.0%
aqueous acetic acid (2:1, v/v). Acyl glucuronide 3 was susceptible
to transacylation by methanol at room temperature. Relative
quantities of reactants and products were estimated from peak areas
of absorbance at 254 nm. Radiolabeled compounds in bile (20-80
µL) and urine (100 µL) were eluted with a gradient of acetonitrile
(20-50% over 30 min, 50-70% over 5 min) in formic acid (0.1%,
v/v). The eluate was monitored at 254 nm. Eluate split-flow to the
LC-MS interface was ca. 40 µL/min. A Quattro II mass spectrom-
eter (Waters Corp, Manchester, UK) fitted with the standard coaxial
electrospray source was operated using nitrogen as the nebulizing
and drying gas. The interface temperature was 80 °C; the electro-
spray capillary voltage was 3.8 kV. Full scan spectra were acquired
in the positive-ion mode between m/z 50-1050 over a scan duration
of 5 s. Fragmentation of analyte ions was achieved at cone voltages
between 25 and 50 V. Selected mass chromatogram peaks were
integrated using MassLynx 3.5 software. Radiolabeled analytes were
quantified with a Radiomatic A250 flow scintillation analyzer
(Perkin-Elmer, Pangbourne, Berkshire, UK) as described previ-
ously.48
(13) Kenny, J. R.; Maggs, J. L.; Meng, X.; Sinnott, D.; Clarke, S. E.;
Park, B. K.; Stachulski, A. V. Syntheses and Characterizations of
the Acyl Glucuronide and Hydroxy Metabolites of Diclofenac. J.
Med. Chem. 2004, 47, 2816-2825.
(14) Kunz, H.; Dombo, B. Solid Phase Synthesis of Peptides and
Glycopeptides on Polymeric Supports with Allylic Anchor Groups.
Angew. Chem., Int. Ed. Engl. 1988, 27, 711-713.
(15) Koch, T.; Hesse, M. Synthesis of Macrocyclic N-Unsubstituted
Imides. Synthesis 1992, 931-933.
Acknowledgment. This work was funded by Pfizer Global
Research and Development through a studentship awarded to
X.M. The work was additionally supported by a grant awarded
under the MRC Proteomics Initiative. The synthesis of acyl
glucuronide 3 using Mitsunobu chemistry was investigated
initially at Pfizer. [14C]2 was prepared by Amersham Bio-
sciences and supplied by Pfizer Global Research and Develop-
ment. The QSTAR and Quattro II mass spectrometers in the
Department of Pharmacology and Therapeutics (Liverpool) were
purchased and maintained with grants from the Wellcome Trust.
We are indebted to Philip J. Roberts, Department of Pharmacol-
ogy and Therapeutics (Liverpool), for performing the animal
experiments.
(16) (a) Brown, T.; Jones, J. H. Protection of Histidine Side-Chains with
π-Benzyloxymethyl- or π-Bromobenzyloxymethyl-Groups. J. Chem.
Soc., Chem. Commun. 1981, 648-649. (b) Brown, T.; Jones, J. H.;
Richards, J. D. Further Studies on the Protection of Histidine Side-
Chains in Peptide Synthesis - the Use of the π-Benzyloxymethyl
group. J. Chem. Soc., Perkin Trans. 1 1982, 1553-1561.
(17) Side-reactions are known with this protecting group on acidolytic
deprotection: Gesquie`re, J.-C.; Diesis, E.; Tartar, A. Conversion of
N-Terminal Cysteine to Thiazolidine Carboxylic Acid During
Hydrogen Fluoride Deprotection of Peptides Containing Nπ-Bom
Protected Histidine. J. Chem. Soc., Chem. Commun. 1990, 1402-
1403.
(18) (a) Stachulski, A. V.; Bowkett, E. R.; Iddon, L.; Maggs, J. L.; Meng,
X.; Park, B. K. Reactive Metabolites: New Developments in the
Synthesis of 1-â-O-Acyl Glucuronides by Selective Acylation. Drug
Metab. ReV. 2006, 38, 135-136. (b) Bowkett, E. R.; Harding, J. R.;
Maggs, J. L.; Park, B. K.; Perrie, J. A.; Stachulski, A. V. Efficient
Synthesis of 1â-O-Acyl Glucuronides via Selective Acylation of Allyl
or Benzyl D-Glucuronate. Tetrahedron 2007, 63, 7596-7605.
(19) Akira, K.; Uchijima, T.; Hashimoto, T. Rapid Internal Acyl Migration
and Protein Binding of Synthetic Probenecid Glucuronides. Chem.
Res. Toxicol. 2002, 15, 765-772.
(20) Barber, C. G.; Cook, A. S.; Maw, G. N.; Pryde, D. C.; Stobie, A.
Cyclopentyl-Substituted Glutaramide Derivatives as Inhibitors of
Neutral Endopeptidase. U.S. Pat. Appl. 0052370, 2002.
(21) Sakurai, Y.; Ma, S.-F.; Watanabe, H.; Yamaotsu, N.; Hirono, S.;
Kurono, Y.; Kragh-Hansen, U.; Otagiri, M. Esterase-Like Activity
of Serum Albumin: Characterization of its Structural Chemistry
Using p-Nitrophenyl Esters as Substrates. Pharm. Res. 2004, 21,
285-292.
(22) Tsuda, Y.; Tsunoi, T.; Watanabe, N.; Ishida, M.; Yamada, H.; Itoh,
T. Stereoselective Binding and Degradation of Sulbenicillin in the
Presence of Human Serum Albumin. Chirality 2001, 13, 236-243.
(23) Manoharan, I.; Boopathy, R. Diisopropylfluorophosphate-Sensitive
Aryl Acylamidase Activity of Fatty Acid Free Human Serum
Albumin. Arch Biochem. Biophys. 2006, 452, 186-188.
Supporting Information Available: Microanalytical and high-
resolution mass spectrometric data for 3, 4, 8-10, 12, 17a, and
18a, syntheses of 8, 9, and 11, and spectroscopic and analytical
data for 14 prepared by a published method. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
(1) (a) Maw, G. N.; Stobie, A.; Planken, S.; Pryde, D. C.; Sanderson,
V.; Platts, M. Y.; Corless, M.; Stacey, P.; Wayman, C.; Van Der
Graaf, P.; Kohl, C.; Coggon, S.; Beaumont, K. The Discovery of
Small Molecule Inhibitors of Neutral Endopeptidase. Structure-
Activity Studies on Functionalised Glutaramides. Chem. Biol. Drug
Des. 2006, 67, 74-77. (b) Pryde, D. C.; Cook, A. S.; Burring, D. J.;
Jones, L. H.; Foll, S.; Platts, M. Y.; Sanderson, V.; Corless, M.;
Stobie, A.; Middleton, D. S.; Foster, L.; Barker, L.; Van Der Graaf,
P.; Stacey, P.; Kohl, C.; Coggon, S.; Beaumont, K. Novel Selective
Inhibitors of Neutral Endopeptidase for the Treatment of Female
Sexual Arousal Disorder. Bioorg. Med. Chem. 2007, 15, 142-159.
(2) Pryde, D. C.; Maw, G. N.; Planken, S.; Platts, M. Y.; Sanderson,
V.; Corless, M.; Stobie, A.; Russell, R.; Foster, L.; Barker, L.;
Wayman, C.; Van der Graaf, P.; Stacey, P.; Morren, D.; Kohl, C.;
Beaumont, K.; Coggon, S.; Tute, M. Novel Selective Inhibitors of
Neutral Endopeptidase for the Treatment of Female Sexual Arousal
Disorder. Synthesis and Activity of Functionalized Glutaramides. J.
Med. Chem. 2006, 49, 4409-4424.
(3) Berman, J. R. Physiology of Female Sexual Function and Dysfunc-
tion. Int. J. Impotence Res. 2005, 17, S44-S51.

6176 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 24
Meng et al.
(24) Ahlfors, S. R.; Kristiansson, M. H.; Lindh, C. H.; Jo¨nsson, B. A.;
Hansson, C. Adducts Between Nucleophilic Amino Acids and
Hexahydrophthalic Anhydride, a Structure Inducing Both Types I
and IV Allergy. Biomarkers 2005, 10, 321-335.
(37) Prueksaritanont, T.; Subramanian, R.; Fang, X.; Ma, B.; Qiu, Y.;
Lin, J. H.; Pearson, P. G.; Baillie, T. A. Glucuronidation of Statins
in Animals and Humans: a Novel Mechanism of Statin Lactonization.
Drug Metab. Dispos. 2002, 30, 505-512.
(38) Li, C.; Subramanian, R.; Yu, S.; Prueksaritanont, T. Acyl-Coenzyme
A Formation of Simvastatin in Mouse Liver Preparations. Drug
Metab. Dispos. 2006, 34, 102-110.
(39) Walker, G. S.; Atherton, J.; Bauman, J.; Kohl, C.; Lam, W.; Reily,
M.; Lou, Z.; Mutlib, A. Determination of Degradation Pathways and
Kinetics of Acyl Glucuronides by NMR Spectroscopy. Chem. Res.
Toxicol. 2007, 20, 876-886.
(40) Prickett, K. S.; Baillie, T. A. Evidence for the in Vitro Metabolism
of Allylisopropylacetamide to Reactive Intermediates. Mechanistic
Studies with Oxygen-18. Biomed. Mass Spectrom. 1984, 11, 320-
331.
(41) Grillo, M. P.; Knutson, C. G.; Sanders, P. E.; Waldon, D. J.; Hua,
F.; Ware, J. A. Studies on the Chemical Reactivity of Diclofenac
Acyl Glucuronide with Glutathione: Identification of Diclofenac-
S-Acyl-Glutathione in Rat Bile. Drug Metab. Dispos. 2003, 31,
1327-1336.
(42) Olsen, J.; Li, C.; Skonberg, C.; Bjornsdottir, I.; Sidenius, U.; Benet,
L.; Hansen, S. H. Studies on the Metabolism of Tolmetin to the
Chemically Reactive Acyl-Coenzyme A Thioester Intermediate in
Rats. Drug Metab. Dispos. 2007, 35, 758-764.
(43) Wa, C.; Cerny, R. L.; Hage, D. S. Identification and Quantitative
Studies of Protein Immobilization Sites by Stable Isotope Labeling
and Mass Spectrometry. Anal. Chem. 2006, 78, 7967-7977.
(44) Walker, J. E. Lysine Residue 199 of Human Serum Albumin is
Modified by Acetylsalicylic acid. FEBS Lett. 1976, 66, 173-175.
(45) Lapolla, A.; Fedele, D.; Reitano, R.; Arico, N. C.; Seraglia, R.; Traldi,
P.; Marotta, E.; Tonani, R. Enzymatic Digestion and Mass Spec-
trometry in the Study of Advanced Glycation End Products/Peptides.
J. Am. Soc. Mass. Spectrom. 2004, 15, 496-509.
(46) Pfizer Global Research and Development, unpublished observations.
(47) Jenkins, R. E.; Kitteringham, N. R.; Hunter, C. L.; Webb, S.; Hunt,
T. J.; Elsby, R.; Watson, R. B.; Williams, D.; Pennington, S. R.;
Park, B. K. Relative and Absolute Quantitative Expression Profiling
of Cytochromes P450 Using Isotope-Coded Affinity Tags. Proteomics
2006, 6, 1934-1947.
(25) Jones, J. B.; Young, J. M. Carcinogenicity of Lactones III. The
Reactions of Unsaturated γ-Lactones with L-Cysteine. J. Med. Chem.
1968, 11, 1176-1182.
(26) Yvon, M.; Anglade, P.; Wal, J. M. Identification of the Binding Sites
of Benzyl Penicilloyl, the Allergenic Metabolite of Penicillin, on the
Serum Albumin Molecule. FEBS Lett. 1990, 263, 237-240.
(27) Ding, A.; Zia-Amirhosseini, P.; McDonagh, A. F.; Burlingame, A.
L.; Benet, L. Z. Reactivity of Tolmetin Glucuronide with Human
Serum Albumin. Identification of Binding Sites and Mechanisms of
Reaction by Tandem Mass Spectrometry. Drug Metab. Dispos. 1995,
23, 369-376.
(28) Zia-Amirhosseini, P.; Ding, A.; Burlingame, A. L.; McDonagh, A.
F.; Benet, L. Z. Synthesis and Mass-Spectrometric Characterization
of Human Serum Albumins Modified by Covalent Binding of Two
Non-Steroidal Anti-Inflammatory Drugs: Tolmetin and Zomepirac.
Biochem. J. 1995, 311, 431-435.
(29) Kristiansson, M. H.; Lindh, C. H.; Jo¨nsson, B. A. G. Determination
of Hexahydrophthalic Anhydride Adducts to Human Serum Albumin.
Biomarkers 2003, 5, 343-359.
(30) Olsen, J.; Bjornsdottir, I.; Tjornelund, J.; Honore Hansen, S.
Identification of the Amino Acids of Human Serum Albumin
Involved in the Reaction with the Naproxen Acyl Coenzyme A
Thioester Using Liquid Chromatography Combined with Fluores-
cence and Mass Spectrometric Detection. Anal. Biochem. 2003, 312,
148-156.
(31) Lindh, C. H.; Kristiansson, M. H.; Berg-Andersson, U. A.; Cohen,
A. S. Characterization of Adducts Formed Between Human Serum
Albumin and the Butadiene Metabolite Epoxybutanediol. Rapid
Commun. Mass Spectrom. 2005, 19, 2488-2496.
(32) Aldini, G.; Gamberoni, L.; Orioli, M.; Beretta, G.; Regazzoni, L.;
Maffei Facino, R.; Carini, M. Mass Spectrometric Characterization
of Covalent Modification of Human Serum Albumin by 4-Hydroxy-
trans-2-Nonenal. J. Mass Spectrom. 2006, 41, 1149-1161.
(33) Sekikawa, H.; Yagi, N.; Lin, E. T.; Benet, L. Z. Apparent Intramo-
lecular Acyl Migration and Hydrolysis of Furosemide Glucuronide
in Aqueous Solution. Biol. Pharm. Bull. 1995, 18, 134-139.
(34) Cho, W. K.; Boyer, J. L. Vasoactive Intestinal Polypeptide is a Potent
Regulator of Bile Secretion from Rat Cholangiocytes. Gastroenterol-
ogy 1999, 117, 420-428.
(35) Munafo, A.; Hyneck, M. L.; Benet, L. Z. Pharmacokinetics and
Irreversible Binding of Tolmetin and its Glucuronic acid Esters in
the Elderly. Pharmacology 1993, 47, 309-317.
(36) Bolze, S.; Bromet, N.; Gay-Feutry, C.; Massiere, F.; Boulieu, R.;
Hulot, T. Development of an In Vitro Screening Model for the
Biosynthesis of Acyl Glucuronide Metabolites and the Assessment
of their Reactivity toward Human Serum Albumin. Drug Metab.
Dispos. 2002, 30, 404-413.
(48) Maggs, J. L.; Bishop, L. P. D.; Edwards, G.; O’Neill, P. M.; Ward,
S. A.; Winstanley, P. A.; Park, B. K. Biliary Metabolites of
â-Artemether in Rats: Biotransformations of an Antimalarial En-
doperoxide. Drug Metab. Dispos. 2000, 28, 209-217.
JM0706766