Kinetic Analysis by HPLC-Electrospray Mass Spectrometry of the pH-Dependent Acyl Migration and Solvolysis as the Decomposition Pathways of Ifetroban 1-O-Acyl Glucuronide

Article abstract of DOI:10.1021/ac9708811

The decomposition of both α- and β-anomers of ifetroban 1-O-acyl glucuronide in an aqueous medium was studied at ambient temperature (23 °C). An HPLC/MS technique was used to investigate the decomposition of these anomers via hydrolysis and acyl migration over a pH range of 1.0-13.0. It was found that while no acyl migration occurred at pH ≤4.0, hydrolysis occurred at both acidic and basic pH. The hydrolysis rate was the slowest within the pH range of 3.0-4.0. Outside this range, the rate of hydrolysis increased as the pH was increased or decreased. The acyl migration at pH 5.0 was very slow, but the rate increased as the pH was increased. First-order rate constants (in h^-1), as a function of pH, were estimated according to a simplified scheme for both hydrolysis and acyl migration processes. Methanolysis of the β-anomer was studied in 80% or 100% methanol in water and also in 20% methanol in aqueous buffer solutions with apparent pH values of 3.0, 6.0 and 9.0. A systematic study of varying the pH and ionic strength of the HPLC mobile phase showed that both parameters significantly affected the resolution and retention times of the positional isomers and anomers. The HPLC method developed was capable of separating all the positional isomeric glucuronides and their anomers.

Full text of DOI:10.1021/ac9708811

Anal. Chem. 1998, 70, 1622-1628
Kinetic Analysis by HPLC-Electrospray Mass
Spectrometry of the pH-Dependent Acyl Migration
and Solvolysis as the Decomposition Pathways of
Ifetroban 1-O-Acyl Glucuronide
Sanaullah Khan, Deborah S. Teitz, and Mohammed Jemal*
Department of Metabolism and Pharmacokinetics, Bristol-Myers Squibb Pharmaceutical Research Institute, P.O. Box 191,
New Brunswick, New Jersey 08903-0191
The decomposition of both r- and â-anomers of ifetroban
1 -O-acyl glucuronide in an aqueous medium was studied
at ambient temperature (23 °C). An HP LC/ MS technique
was used to investigate the decomposition of these ano-
mers via hydrolysis and acyl migration over a pH range
of 1 .0 -1 3 .0 . It was found that while no acyl migration
occurred at pH e4 .0 , hydrolysis occurred at both acidic
and basic pH. The hydrolysis rate was the slowest within
the pH range of 3 .0 -4 .0 . Outside this range, the rate of
hydrolysis increased as the pH was increased or de-
creased. The acyl migration at pH 5 .0 was very slow, but
the rate increased as the pH was increased. First-order
rate constants (in h-1), as a function of pH, were estimated
according to a simplified scheme for both hydrolysis and
acyl migration processes. Methanolysis of the â-anomer
was studied in 8 0 % or 1 0 0 % methanol in water and also
in 2 0 % methanol in aqueous buffer solutions with appar-
ent pH values of 3 .0 , 6 .0 and 9 .0 . A systematic study of
varying the pH and ionic strength of the HP LC mobile
phase showed that both parameters significantly affected
the resolution and retention times of the positional
isomers and anomers. The HP LC method developed was
capable of separating all the positional isomeric glucu-
ronides and their anomers.
Figure 1. Structures of the various compounds used as analytes
and internal standards.
Ifetroban (I, Figure 1), a carboxylic acid compound, is a potent
antagonist of thromboxane A2 receptors in smooth muscle and in
Acylglucuronides are known to undergo a variety of spontaneous
platelets.1,2 Ifetroban 1-O-acyl â-
-glucuronide (1â-gluc, II, Figure
D
chemical reactions.5-7 Hydrolysis of the ester bond and intramo-
lecular migration of the acyl group are among the most common
reactions of acyl glucuronides.8 The in vivo biological implications
of the formation of acyl glucuronides have been thoroughly
reviewed.8-11 The in vitro lability of the acyl glucuronides, due
1) was found to be a major metabolite in human plasma.
Glucuronidation is one of the major phase II metabolic pathways
of carboxylic acid compounds and serves to improve the solubility
and hence the elimination of the parent compound.3 The biosyn-
thetic product of the acyl glucuronidation reaction has a 1-O-acyl
â- conformation at the sugar moiety of the resulting conjugate.4
(4) Pigman, W.; Isbell, H. S. Adv. Carbohydr. Chem. 1 9 6 8 , 23, 11-57.
(5) Ding, A.; Ojingwa, J. C.; McDonagh, A. F.; Burlingame, A. L.; Benet, L. Z.
Proc. Natl. Acad. Sci. U.S.A. 1 9 9 3 , 46, 3797-3801.
(6) Grubb, N.; Weil, A.; Caldwell, J. Biochem. Pharmacol. 1 9 9 3 , 46, 357-364.
(7) Dikinson, R. G.; King, A. R. Biochem. Pharmacol. 1 9 9 1 , 42, 2301-2306.
(8) Faed, E. M. Drug Metab. Rev. 1 9 8 4 , 15, 1213-1249.
(9) Spahn-Langguth, H.; Benet, L. Z. Drug Metab. Rev. 1 9 9 2 , 24, 5-48.
(10) Smith, P. C.; Benet, L. Z.; McDonagh, A. F. Drug. Metab. Dispos. 1 9 9 0 , 18,
639-644.
(1) Ogletree, M. L.; Harris, D. N.; Schumacher, W. A.; Hall, S. E.; Brown, B.
R.; Misra, R. N. Circulation 1 9 9 1 , 84, II-79.
(2) Misra, R. N.; Brown, B. R.; Sher, P. M.; Patel, M. M.; Hall, S. E.; Han, W.-
C.; Barrish, J. C.; Flyod, D. M.; Sprague, P. W.; Morrison, R. A.; Ridgewell,
R. E.; White, R. E.; DiDonato, G. C.; Harris, D. N.; Hedberg, A.; Schumacher,
W. A.; Webb, M. L.; Ogletree, M. L. Bioorg. Med. Chem. Lett. 1 9 9 2 , 2,
73-76.
(3) Watt, J. A.; King, A. R.; Dikinson, R. G. Xenobiotica 1 9 9 1 , 21, 403-415.
(11) Hayball, P. J. Chirality 1 9 9 5 , 7, 1-9.
1622 Analytical Chemistry, Vol. 70, No. 8, April 15, 1998
S0003-2700(97)00881-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/11/1998

to hydrolysis, solvolysis, and acyl migration, is a major concern
in quantitative bioanalysis of biofluids for the parent carboxylic
acid compound and/ or its acyl glucuronide. Enzymatic hydrolysis
of glucuronides with â-glucuronidase is a routine method of
sample preparation. Unlike the biosynthetic 1-O-acyl â-glucu-
ronide, the positional isomers are not hydrolyzed by â-glucu-
ronidase,12 and therefore, the analytical results can be potentially
misleading. The susceptibility of acyl glucuronides toward in vitro
hydrolysis will produce results which may not represent the true
concentration profile of the unconjugated drug in a biological
matrix such as plasma and urine. Hence, it is essential to
determine the optimum conditions required to minimize the
degradation of the acyl glucuronide during the various steps of a
bioanalytical assay, including the collection and storage of the
biological samples.
The rate of acyl migration depends on the structure of the
aglycon, pH, and temperature.13 In this paper, we report the
results of an extensive investigation of the hydrolysis and acyl
migration as the decomposition pathways of II in various buffer
solutions over a pH range of 1.0-13.0. The combination of serial
migration of the acyl group over multiple positions across the
glucuronic acid molecule and the associated equilibria makes the
kinetic analysis of such a system a formidable task. This paper
presents an attempt to simplify the kinetic analysis of the acyl
migration in a typical acyl glucuronide.
It is known that acyl glucuronides can undergo reactions with
nucleophilic functions (SH, OH, and NH2) of peptides and proteins
to covalently bond the aglycon moiety of the conjugates by
substituting the glucuronic acid moiety.14,15 Since aqueous
methanol, a potential nucleophile, is one of the most popular
solvent mixtures for solid-phase extraction, reconstitution, and
HPLC, we studied the solvolysis of II in aqueous methanol under
different conditions.
In this paper, we also report the results of a systematic study
of the effects of pH and ionic strength of the HPLC mobile phase
on the chromatographic resolution and the retention times of the
positional isomeric glucuronides and their anomers derived from
the decomposition of II.
Aldrich (Milwaukee, WI). Highly pure (18.2 MΩ) water was
obtained by passing house-distilled deionized water through a
Barnstead Nanopure II system (Syborn/ Barnstead, Boston, MA).
HPLC grade acetonitrile and methanol were obtained from EM
Science (Gibbstown, NJ).
Test Mixture. At ambient temperature (23 °C), unless
otherwise described, a solution of II (4 mg/ 100 mL) was subjected
to a mild hydrolysis condition overnight in 20% methanol in
phosphate buffer, pH 8.0 (10 mM). This allowed the production
of measurable amounts of the positional isomers (VI-VIII), the
aglycon I, and the methyl ester (IX) of the aglycon.
Sample Incubation. The buffer systems (10 mM) used to
cover the pH range of the incubation medium were trifluoroacetic
acid/ ammonium trifluoroacetate (pH 1.0), formic acid / ammonium
formate (pH 3.0), acetic acid/ ammonium acetate (pH 4.0 and pH
5.0), and phosphate buffers (pH g6.0). The pH was adjusted with
the corresponding acid and ammonium or potassium hydroxide.
Solutions of II (25 µM, 15.4 µg/ mL) spiked with III (10.0 µM,
4.43 µg/ mL) were incubated at 23 °C in the pH range of 1.0-
13.0. The labeled compound III served as the internal standard
for the quantitation of I produced during incubation. No adjust-
ment of pH over the incubation period was made.
The susceptibility of II toward methanolysis was studied by
incubating II at 23 °C in 20% methanol in aqueous buffers with
apparent pH values of 3.0, 6.0, and 9.0. Methanolysis was also
studied in 100% and 80% methanol in water. Formation of the
methyl ester of I (IX, Figure 1) was monitored to check the
progress of methanolysis. Because no reference standard for IX
was available, the quantitation of the methanolysis product was
done with III as the standard.
HP LC/ MS Analysis. A Hewlett-Packard (Palo Alto, CA) HP
1090 Series II HPLC system equipped with an autoinjector was
used for solvent delivery and programmed sample injection. An
HP 5989B MS Engine quadrupole mass spectrometer equipped
with a pneumatically assisted electrospray interface, rf-only hexa-
pole ion guide (Analytica of Branford, CT) and a Chem-Station
data system, was used for detection. Positive ion mode of
electrospray was used for ionization. The electrospray voltage
was 4.5 kV (measured as ∆V between the needle and the
cylinder), and the quadrupole temperature was set at 110 °C. The
capillary exit (CapEx) voltage was ramped to find the optimum
value. The operating CapEx voltage was set at 130 V. Similarly,
the quadrupole entrance voltage was set at 190 V. High-purity
(99.999%) nitrogen gas was used as the nebulizing gas at an
operating pressure of 80 psi. The drying gas was also high-purity
nitrogen flowing at 1.2 L/ min and was set at 360 °C. The
quadrupole was tuned and calibrated with the electrospray tuning
mixture, supplied by Hewlett-Packard (Part No. 59987-60135), at
unit resolution. HPLC separations were performed on a (150 ×
2 mm i.d.) column packed with 3-µm Hypersil BDS-C-8 packing
material (Keystone, Bellefonte, PA). A flow rate of 0.250 mL/
min was used.
EXPERIMENTAL SECTION
Reagents. The structures of all analytes and internal stan-
dards are shown in Figure 1. Both labeled and unlabeled ifetroban
and the corresponding glucuronide conjugates were obtained from
Bristol-Myers Squibb Pharmaceutical Research Institute (BMS-
PRI). The labeled analogues consisted of [13C3]ifetroban (III) and
[2H5]1â-gluc (IV), which were used as internal standards. The
anomer ifetroban 1-O-acyl R- -glucuronide (1R-gluc, V) was also
D
obtained from BMS-PRI. Chemicals of highest available purity
were used. Ammonium acetate (99.999%), acetic acid (double
distilled), trifluoroacetic acid (TFA), ammonium trifluoroacetate,
formic acid, ammonium formate, potassium hydrogen phosphate,
ammonium hydroxide, and potassium hydroxide were from
(12) Blanckaert, N.; Compernolle, F.; Leroy, P.; Van Hautte, R.; Fevery, J.;
Heirwegh, K. P. M. Biochem. J. 1 9 7 8 , 171, 203-214.
(13) Haines, A. H. In Advances in Carbohydrate Chemistry and Biochemistry;
Tipson, R. S., Horton D., Eds.; Academic Press: New York, 1976; Vol. 33,
pp 1-109.
Initially the test mixture (5 µL) was isocratically chromato-
graphed with a 10 mM ammonium formate/ formic acid buffer in
50% acetonitrile (apparent pH 3.5). From the retention time, as
confirmed by the deuterated internal standard (IV, Figure 1), and
m/ z value of the [M + H]+ ion (617), it was concluded that all
the positional isomers coeluted in one peak at 2.3 min (Figure
(14) Salmon, M., Fenselau, C.; Cukier, LJ. O.; Odell, G. B. Life Sci. 1 9 7 5 , 15,
2069-2078.
(15) Stogniew, M.; Fenselau, C. Drug. Metab. Dispos. 1 9 8 2 , 10, 609.
Analytical Chemistry, Vol. 70, No. 8, April 15, 1998 1623

620. The release of 2H5-aglycon at m/ z 446 was also monitored
to correct for any hydrolysis of the labeled IV, which was used
as an internal standard. For the data acquisition, the m/ z values
of the conjugates and the m/ z values of aglycons were grouped
together separately as timed events to correspond to the regions
of their respective elution in the chromatogram. The dwell time
corresponded to a scan rate of ∼4.5 scans/ s. Methanolysis of II
was monitored at m/ z 455.
Kinetic Analysis. Kinetic analysis of the key steps of the
decay of II was undertaken with the assumption that the initial
decay followed first-order kinetics and that initially only II was
the source of I. Area ratios (analyte response/ internal standard
response) of chromatographic peaks were used for quantitation.
For the kinetic analysis, area ratios calculated for the peaks of I,
II, and VI were used. Because no conversion of 1â-to 1R-anomer,
or vice versa, was noticed, only the peak corresponding to the
1â-anomer was integrated for the kinetic analysis of the decom-
position of II. For VI, the partially resolved anomers were
summed in the integration process.
Figure 2. Separation of the decomposition products of II: (A) LC/
MS chromatogram of ifetroban 1-O-acyl â-D-glucuronide, incubated
overnight at 4 °C in a pH 8.0 phosphate buffer with 10 mM ammonium
acetate, pH 3.5, 50% acetonitrile as the mobile phase. (B) LC/MS
chromatogram of the test mixture with 10 mM ammonium acetate,
pH 5.0, 30% acetonitrile as the mobile phase.
2A). Chromatographic conditions were optimized to further
resolve this peak and separate all the coeluting positional isomers
and most of the anomers (Figure 2B). This was achieved with
an isocratic elution with a 10 mM ammonium acetate/ acetic acid
buffer in 30% acetonitrile (apparent pH 5.0) at 0.250 mL/ min flow
rate. For the kinetic studies, II, its positional isomers VI-VIII,
the aglycon I, and the ester IX (if produced) were separated by
a gradient elution. The chromatographic conditions were set such
as to achieve maximum resolution of the positional isomers at
the expense of no resolution of some of the anomers. Conse-
quently, the R- and â-anomers of 4-O-acyl glucuronide (VIII)
coeluted. The gradient elution was carried out with 30% aqueous
acetonitrile as solvent A, buffered with ammonium acetate/ acetic
acid (5.0 mM NH4OAc, apparent pH 5.5), and solvent B was 50%
aqueous acetonitrile (5.0 mM NH4OAc). Separation of the
glucuronide isomers was achieved with 100% A during the first 7
min. This was followed by a gradient of 100% A to 100% B in 1.5
min. The aglycon I eluted with 100% B during the next 3.5 min.16
The elution was then brought back to 100% A in 1 min. Finally,
the column was reequilibrated for at least 5 more min. An
isocratic elution with 35% acetonitrile and a flow rate of 0.250 mL/
min was used to study the effect of pH and ionic strength of the
mobile phase on the resolution of the glucuronide isomers.
To minimize the hydrolysis of IV, used as the internal standard
for II, a separate solution of IV (20.0 µM, 12.42 µg/ mL) in the
pH 3.0 buffer was co-injected with the incubated sample. The
autoinjector was programmed such that 5 µL of the incubated
solution, prespiked with III, and 5 µL of the standard solution of
IV were withdrawn sequentially into the needle and then injected
(10 µL) onto the HPLC column at specified intervals. The solution
of IV was replaced twice a day with a refrigerated stock solution.
The mass spectrometer was operated in selected ion monitor-
ing (SIM) mode. The unlabeled II was monitored as its [M +
H]+ ion at m/ z 617 and its 2H5-labeled analogue (IV) was
monitored at m/ z 622. Ifetroban was monitored at m/ z 441 as
unlabeled [M + H]+ ion and at m/ z 444 as its 13C3-labeled
analogue. Occurrence of any exchange reaction between the
unlabeled II and the 13C3-labeled III was also monitored at m/ z
RESULTS AND DISCUSSION
Chromatography. The test mixture, obtained via overnight
incubation, was anticipated to contain I and the positional isomers
of II. In Figure 2A, the peak eluting at 4.5 min is that of I, which
is produced by the hydrolysis of II. The identity of this peak
was confirmed by comparing with the authentic compound. The
late eluting peak at 8.6 min was identified as IX, which was
produced by methanolysis of the glucuronide. Because acyl
glucuronides are known to undergo positional isomerization,8,9,17
the early-eluting peak (2.3 min) was suspected to be a composite
peak of the coeluting glucuronide isomers (II and V-VIII). This
prompted the necessity to establish chromatographic conditions
to further resolve this glucuronide peak.
Ionic strength and pH of the mobile phase are major factors
in the separation of carboxylic acids. Solvent strength and flow
rate may also be utilized to improve the resolution. Thus the
variables that were manipulated to achieve the intended resolution
were (i) pH of the mobile phase (3.5-6.5); (ii) the buffer
concentration (1.0-10.0 mM ammonium acetate); (iii) the organic
content of the mobile phase (20-50% acetonitrile); and (iv) flow
rate of the mobile phase (0.2-0.4 mL/ min). Consequently not
only all the four positional isomers but also most of the anomers
were completely or partially resolved (Figure 2B).
Effect of pH of the Mobile P hases. At a constant ionic
strength of 1.0-10.0 mM NH4OAc, a pH range of 3.5-6.5 of the
mobile phase was investigated to accomplish an optimum resolu-
tion of the glucuronide isomers. The chromatographic peaks
corresponding to the aglycon and the different glucuronide
isomers progressively shifted to longer retention times with
decreasing pH. This was expected due to the protonation of the
carboxylic acid groups of the aglycon and the glucuronic acid
moiety of the different positional isomers.
It was anticipated that the longer retention times achieved by
the lower pH would result in enhanced resolution of the isomeric
mixture. However, it was found that the resolution was progres-
sively lost with decreasing pH. Consequently, the longer the
retention time, the worse the resolution. The effect of pH on the
(16) In the methanolysis experiments, 6 min more was added to the elution time
with 100% B to elute the methyl ester IX.
(17) Bradow, G.; Kan, L.; Fenselau, C. Chem. Res. Toxicol. 1 9 8 9 , 2, 316-324.
1624 Analytical Chemistry, Vol. 70, No. 8, April 15, 1998

Figure 4. Resolution vs [NH4OAc] at pH 5.5 with 35% aqueous
acetonitrile and a flow rate of 0.250 mL min: (A) 1.0 mM, bandwidth
at the baseline 0.6 min; (B) 2.5 mM, bandwidth at the baseline 0.8
min; (C) 5.0 mM, bandwidth at the baseline 1.0 min; (D) 7.5 mM,
bandwidth at the baseline 1.1 min; (E) 10.0 mM, bandwidth at the
baseline 1.2 min.
Figure 3. Resolution vs pH at 7.5 mM [NH4OAc] with 35% aqueous
acetonitrile and a flow rate of 0.250 mL min: (A) pH 3.5, bandwidth
at the baseline 0.8 min; (B) pH 4.0, bandwidth at the baseline 0.5
min; (C) pH 4.5, bandwidth at the baseline 0.6 min; (D) pH 5.0,
bandwidth at the baseline 1.0 min; (E) pH 5.5, bandwidth at the
baseline 1.1 min; (F) pH 6.0, bandwidth at the baseline 1.1 min; (G)
pH 6.5, bandwidth at the baseline 1.0 min.
Table 1. Induction Times of the Positional Isomers
VI-VIII as a Function of pH over a One-Week-Long
Incubation
separation of the isomers, at a constant ionic strength of 7.5 mM
ammonium salt, is shown in Figure 3. At pH 3.5, all the isomers
coeluted in one broad band (Figure 3A). This band started to
compress at the baseline as the pH of the mobile phase was
increased from 3.5 to 4.5 (Figure 3B,C). Further increase in pH
resulted in the expansion of the bandwidth to a constant value of
1.1 min (Figure 3D-G), notwithstanding the shorter retention
times. This led to the resolution of the isomers. The data indicate
that the extent to which the retention of an isomer is affected by
pH is different for different isomers. Thus, the order of elution
of the isomers could be subject to change with pH. This is in
contrast with the effect of pH on the retention behavior of the
pH 1.0 pH 3.0 pH 4.0 pH 5.0 pH 6.0 pH 7.4 pH 9.0
VI
VII
VIII none
nonea
none
none
none
none
none
none
none
72 h
none
none
5 h
75 h
none
0.5 h
3 h
7.5 h
0.2 h
1.5 h
4.5 h
a
None, represents no formation of the respective isomer.
flow rate of 0.250 mL/ min. Under these isocratic elution condi-
tions, the aglycon eluted at ∼18 min, which was deemed to be
too long. A gradient elution, described in the Experimental
Section, was thus used in order to elute the aglycon in less than
12 min.
aglycon-based R- and â-anomers of an ether-linked â-D-glucu-
ronide. The retention of these anomers is reported to be equally
affected by a change in pH of the mobile phase causing them to
move in unison.18
Identification of the P ositional Isomers and Their Respec-
tive Anomers. In Figure 2, the chromatographic peaks obtained
from the test mixture, an incubated solution of II, have been
assigned to their respective glucuronide isomers. The 1-O-acyl
â- and R-glucuronide anomers (II and V, respectively) were
identified by matching their retention times with those of the
authentic compounds. The 1-O-acyl R-anomer (V) was present
as a synthetic impurity in II and was not a product of mutarotation.
Interconversion between II and V is not expected because the
anomeric hydroxyl in II or V is already locked by the acyl
substituent. The assignment of 2-O-, 3-O-, and 4-O-acyl positional
isomers (VI, VII, and VIII, respectively) was based on the
respective induction time, which is the time elapsed between the
onset of the incubation of II and the emergence of the corre-
sponding peak in the chromatograms (Table 1). The partially
resolved doublet which eluted at 7.5 min was identified as the
2-O-acyl R/ â-glucuronide (VI) because it was the first new
glucuronide peak to appear in the course of the incubation of II.
The two peaks of the doublet emerged simultaneously; however,
Effect of Ionic Strength of the Mobile P hase. Increasing
the concentration of an ammonium salt (1.0-10.0 mM) at a given
pH (in the pH range 3.5-6.5) increased the retention times of
the glucuronide isomers, but the increase was not as dramatic as
with the lowering of pH. Increasing ionic strength also improved
the resolution of the late-eluting isomers. Figure 4 depicts the
effects of changing the ammonium acetate concentration, at a fixed
pH of 5.5. The retention times increased and the resolution
improved with the increase in ammonium acetate concentration.
The bandwidth gradually increased with increasing ammonium
acetate concentration. The order of elution of the isomers did
not appear to change.
A pH value of 5.0-5.5 and ammonium acetate concentration
of 5.0-10.0 mM provided acceptable resolution for a mobile phase
of 30-35% aqueous acetonitrile, with an isocratic elution and a
(18) Bowers, L. D.; Sanaullah. J. Chromatogr., B 1 9 9 6 , 687, 61-68.
Analytical Chemistry, Vol. 70, No. 8, April 15, 1998 1625

Figure 6. Methanolysis of a 25 µM (15.4 µg/mL) solution of 1-O-
acyl â-D-glucuronide incubated at 23 °C in (A) 100% methanol (O)
and 80% aqueous methanol (0); (B) 20% methanol in aqueous buffer
solutions at pH 3.0 (0), 6.0 (O), and 9.0 (4, inset).
six-membered ring structure of â-D-glucuronic acid, having a chair
Figure 5. Isomeric transformation of the 1-O-acyl â- and R-D-
glucuronide at pH 11.0, chromatographed with 30% acetonitrile (5.0
mM NH4OAc, pH 5.5): 1-O-acyl â-D-glucuronide incubated for (A)
0.5 and (B) 5.0 min; 1-O-acyl R-D-glucuronide incubated for (C) 0.5
and (D) 5.0 min.
conformation with all the hydroxyl groups in the equatorial
position, can be easily anticipated as the most stable conformation.
Thus, vicinal equatorial hydroxyl groups of glucuronic acid will
have staggered conformation. In an undistorted chair conforma-
tion, the torsion angle between vicinal equatorial trans groups is
the same as that between the axial and equatorial cis groups.
Consequently, the migration of an acyl substituent on the sugar
moiety to an adjacent trans hydroxyl, via the formation of the cyclic
orthoester intermediate,14 is likely, although the energy required
for the intermediate cyclization process involving the acyl group
and an adjacent hydroxyl group is lower for cis-(eq,ax) than the
trans-(eq, eq).19 Therefore, a sequential migration of the acyl group
from position 1 toward position 4 can be anticipated. Unlike II,
where all the ring hydroxyl groups and the acyl substituent are
trans, the acyl substituent in V at C-1 is axial and has a cis
conformation relative to the C-2 hydroxyl function. Therefore, a
relatively fast 1-O-acyl R f 2-O-acyl R-conversion is expected.20
This is in agreement with our experimental observation (data not
reported).
Scheme 1
in the beginning the late-eluting one was bigger than the earlier
one. The emergence of the peak eluting at 6.8 min followed that
of the 7.5-min doublet. Thus, the peak at 6.8 min was identified
as 3-O-acyl glucuronide (VII). Attempts to further resolve this
peak entailed the coelution and broadening of other peaks. The
doublet at 5.1 min was the last one to emerge. Thus it was
identified as the 4-O-acyl glucuronide (VIII).
The distinction between the 2-O-acyl R- and â-anomers was
based on the assumption that the 1-O-acyl â-glucuronide (II)
first transforms to 2-O-acyl â positional isomer which then
converts reversibly to the corresponding 2-O-acyl R- anomer.
The sequence of events related to the 1-O-acyl â- to 2-O-acyl
â- and 1-O-acyl R- to 2-O-acyl R-conversion is depicted in Scheme
1. This is supported by the isomeric transformation profile
obtained from the incubation of II and V in a pH 11.0 buffer
(Figure 5). In Figure 5A, where the starting glucuronide is II,
the last peak (6.3 min) of the partially resolved doublet decayed
immediately to yield the early peak (5.8 min) of the doublet, as
shown in Figure 5B. Thus, the peak at 6.3 min represents the
initial 1-O-acyl â- f 2-O-acyl â- T 2-O-acyl R-conversion. However,-
when V was the starting glucuronide, the sequence of the
anomeric conversions reversed (Figure 5C and D), as now the
5.8-min peak of the doublet represents the initial 1-O-acyl R- f
2-O-acyl R- T 2-O-acyl â-conversion. The assignment of R- and
â-anomers of VII and VIII could be achieved based on the same
reasoning.
Methanolysis of II. Figure 6A represents methanolysis of a
25 µM (15.4 µg/ mL) solution of II in 100% methanol and 80%
methanol in water. As shown in Figure 6A, the rate of metha-
nolysis increases with increasing concentration of methanol.
Analysis of the data (Figure 6B) for the methanolysis of II in 20%
methanol in aqueous buffer solutions with apparent pH values of
3.0, 6.0, and 9.0 revealed that the rate of methanolysis increased
with increasing pH. While at pH 3.0 and pH 6.0 the methanolysis
was not very extensive, at pH 9.0 the original glucuronide
predominantly converted to IX in 1 h. No direct esterification of
free ifetroban (I) to the ester IX was detected under the conditions
of methanolysis of II.
Decomposition Kinetics. In Figure 7, the concentrations of
different positional isomers are plotted as a function of incubation
time of II in a pH 9.0 buffer. Scheme 1 depicts the steps involved
in the hydrolysis and transacylation of the acyl glucuronides. The
kinetic model for such a scheme will comprise a complex
The sequential migration of an acyl substituent on a glucuronic
acid can be rationalized by considering the structural character-
istics of the closed ring pyranosyl structure of glucuronic acid. A
(19) Shvets, V. I. Russ. Chem. Rev. 1 9 7 4 , 43, 488-502.
(20) Chung, S.-K.; Chang, Y.-T. J. Chem. Soc., Chem. Commun. 1 9 9 5 , 13-14.
1626 Analytical Chemistry, Vol. 70, No. 8, April 15, 1998

Table 2. Variation of the First-Order Rate Constants
(h-1) of the Decay of Ifetroban 1-O-Acyl
â-D-Glucuronide as a Function of pH
pH 1.0
pH 3.0
pH 5.0
pH 7.4
pH 9.0
k
k1
k2
4.1 × 10-2 7.1 × 10-4 2.7 × 10-3 8.8 × 10-2 1.5 × 10-1
1.2 × 10-2 2.6 × 10-2
7.6 × 10-2 1.3 × 10-1
To calculate the individual rate constants, k1 and k2 at a pH
>5, the following expressions can be derived.22
d[ifetroban]/ dt ) k1[1â-gluc]
(5)
Figure 7. Typical decay curves of 25 µM (15.4 µg/mL) solution of
1-O-acyl â-D-glucuronide incubated in pH 9.0 phosphate buffer at 23
°C: 1â-gluc (*); 2-gluc (0); 3-gluc (O); 4-gluc (4); ifetroban (3).
Substituting for [1â-gluc] from eq 3, eq 5 can be rewritten as
d[ifetroban]/ dt ) k1[1â-gluc]0e-kt
(6)
(7)
Scheme 2
[ifetroban]/ [1â-gluc]0 ) (k1/ k)(1 - e-kt
)
Similarly the initial decay of II via the k2 route can be
represented by
[2-gluc]/ [1â-gluc]0 ) (k2/ k)(1 - e-kt
)
(8)
(9)
combination of consecutive, opposing, and parallel reactions.21
Because the acyl migration proceeds in a sequential fashion,
Scheme 1 can be simplified to Scheme 2 to represent the initial
decomposition of II. Even with this simplified scheme, the kinetic
analysis is still complicated by the concurrent hydrolysis of VI.
At pH e4.0, only ifetroban was produced. Thus, at pH e4.0,
there was only hydrolysis and no acyl migration occurred. At
pH 5.0, the transformation of 1-O-acyl to the 2-O-acyl positional
isomer was noticeable only after 72 h (Table 1). Therefore, only
the deconjugation kinetics due to hydrolysis is reported here in
the pH range 1.0-5.0. Above pH 5.0, II was consumed in two
parallel processes: (i) hydrolysis, and (ii) 1-O-acyl f 2-O-acyl
migration. Therefore, at pH >5.0, both hydrolysis and transacy-
lation are considered to account for the loss of II.
Dividing eq 7 by eq 8 gives
[ifetroban]/ [2-gluc] ) (k1/ k2) ) S
where S is the slope of the plot of [ifetroban (I)] vs [2-gluc (VI)].
Equation 2 and eq 9 can be combined to obtain
k1 ) kS/ (S + 1)
k2 ) k/ (S + 1)
(10)
(11)
The rate law governing the decomposition processes depicted
in Scheme 2, is represented by
The rate constants for the decay of II by hydrolysis (pH range
1.0-5.0) and by the combined mechanism (pH 7.4 and 9.0) are
listed in Table 2. The plot of the concentration of I vs the
concentration of VI, to calculate k1 and k2 from the decomposition
data of pH 7.4 and pH 9.0, showed small upward curvature. This
small positive deviation from the expected straight line, even in
the initial part of the decay process, indicated that the hydrolysis
product I at pH >5.0 is not exclusively produced by the k1 route
of Scheme 2. The hydrolysis of VI, produced via the k2 route,
also contributed to the concentration of I. Thus, the biased
concentration of I would result in the overestimation of k1. This
is a clear demonstration of the complexity of kinetic analysis of
the decomposition of acyl glucuronides as a consequence of the
hydrolyzability of each of the positional isomers.
-d[1â-gluc]/ dt ) k[1â-gluc]
k ) (k1 + k2)
(1)
where
(2)
[1â-gluc] ) [1â-gluc]0e-kt
(3)
(4)
ln([1â-gluc]/ [1â-gluc]0) ) -kt
Eq 4 represents a simple first-order reaction and can be used to
determine the apparent overall decomposition rate constant, k.
In the pH range 1-5, k2 ) 0 and thus k ) k1.
CONCLUSION
Depending on pH of the solution, 1-O-acyl â-D-glucuronides in
aqueous solutions can undergo hydrolysis or hydrolysis and
(21) Sidelmann, U. G.; Hansen, S. H.; Gavaghan, C.; Carless, H. A. J.; Lindon, J.
C.; Farrant, R. D.; Wilson, I. D.; Nicholson, J. K. Anal. Chem. 1 9 9 6 , 68,
2564-2572.
(22) Moore, J. W.; Pearson, R. G. Kinetics and Mechanism, 3rd ed.; John Wiley:
New York, 1981; Chapter 8.
Analytical Chemistry, Vol. 70, No. 8, April 15, 1998 1627

intramolecular acyl migration. Adjusting the pH to 3.0-4.0
drastically reduces the rate of both hydrolysis and acyl migration.
At pH e4.0, only the hydrolysis occurs. At pH g5.0, both
hydrolysis and acyl migration occur, the rate for both increasing
with increase in pH. At high pH values (pH >11.0), not only the
rate of acyl migration but also the rate of hydrolysis of 1-O-acyl
glucuronide and its positional isomers is greatly enhanced.
Hence, the production of the parent acid compound is fast since
the half-lives of all isomers of the glucuronide are relatively short.
The chromatographic resolution of the positional isomers is highly
affected by the pH and ionic strength of the HPLC mobile phase.
Thus, it is important to select the appropriate chromatographic
conditions depending on whether resolving the positional isomers
is desired. Finally, it should be stressed that for the accurate
determination of an 1-O acyl-â- -glucuronide and/ or the parent
D
acid compound in biological matrixes, it is essential to minimize
the in vitro hydrolysis and acyl migration reactions. The choice
of a solvent for preparing the stock solutions, reconstitution, and
chromatographic mobile phase has to be made such as to avoid
the solvolysis of the glucuronide.
Received for review August 14, 1997. Accepted January
28, 1998.
AC9708811
1628 Analytical Chemistry, Vol. 70, No. 8, April 15, 1998