Synthesis of a series of phenylacetic acid 1-β-O-acyl glucosides and comparison of their acyl migration and hydrolysis kinetics with the corresponding acyl glucuronides

Article abstract of DOI:10.1039/c0ob00820f

We report the synthesis of the 1-β-O-acyl glucoside conjugates of phenylacetic acid (PAA), R- and S-α-methyl-PAA and α,α′- dimethyl-PAA, and measurement of their transacylation and hydrolysis reactivity by NMR methods. These are analogues of acyl glucuronides, the transacylation kinetics of which could be important in adverse drug effects. One aim of this work was to investigate whether, as previously postulated, the free carboxylate group of the acyl glucuronides plays a part in the mechanism of the internal acyl migration. In addition, such acyl glucosides are known to be endogenous biochemicals in their own right and investigation of their acyl migration propensities is novel. Our previously described selective acylation procedure has proved highly successful for 1-β-O-acyl glucuronide synthesis and when subsequently applied to 6-O-trityl glucose, it gave good yields and excellent anomeric selectivity. Mild acidolysis of the O-trityl intermediates gave the desired acyl glucosides in excellent yield with essentially complete β-selectivity. Measurement of the acyl glucoside transacylation kinetics by ¹H NMR spectroscopy, based simply on the disappearance of the 1-β-isomer in aqueous buffer at pH 7.4, showed marked differences depending on the degree of methyl substitution. Further kinetic modelling of the isomerisation and hydrolysis reactions of the acyl glucosides showed considerable differences in kinetics for the various isomeric reactions. Reactions involving the -CH₂OH group, presumably via a 6-membered ring ortho-ester intermediate, are facile and the α-glucoside anomers are significantly more reactive than their β-counterparts. By comparison with degradation rates for the corresponding acyl glucuronides, it can be inferred that substitution of the carboxylate by -CH₂OH in the acyl glucosides has a significant effect on acyl migration for those compounds, especially for rapidly transacylating molecules, and that thus the charged glucuronide carboxylate is a factor in the kinetics.

Full text of DOI:10.1039/c0ob00820f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Synthesis of a series of phenylacetic acid 1-b-O-acyl glucosides and
comparison of their acyl migration and hydrolysis kinetics with the
corresponding acyl glucuronides
Lisa Iddon,†a Selena E. Richards,b Caroline H. Johnson,‡b John R. Harding,c Ian D. Wilson,c
Jeremy K. Nicholson,b John C. Lindonb and Andrew V. Stachulski*§a
Received 4th October 2010, Accepted 11th October 2010
DOI: 10.1039/c0ob00820f
We report the synthesis of the 1-b-O-acyl glucoside conjugates of phenylacetic acid (PAA), R- and
S-a-methyl-PAA and a,a¢-dimethyl-PAA, and measurement of their transacylation and hydrolysis
reactivity by NMR methods. These are analogues of acyl glucuronides, the transacylation kinetics of
which could be important in adverse drug effects. One aim of this work was to investigate whether, as
previously postulated, the free carboxylate group of the acyl glucuronides plays a part in the
mechanism of the internal acyl migration. In addition, such acyl glucosides are known to be
endogenous biochemicals in their own right and investigation of their acyl migration propensities is
novel. Our previously described selective acylation procedure has proved highly successful for
1-b-O-acyl glucuronide synthesis and when subsequently applied to 6-O-trityl glucose, it gave good
yields and excellent anomeric selectivity. Mild acidolysis of the O-trityl intermediates gave the desired
acyl glucosides in excellent yield with essentially complete b-selectivity. Measurement of the acyl
glucoside transacylation kinetics by 1H NMR spectroscopy, based simply on the disappearance of the
1-b-isomer in aqueous buffer at pH 7.4, showed marked differences depending on the degree of methyl
substitution. Further kinetic modelling of the isomerisation and hydrolysis reactions of the acyl
glucosides showed considerable differences in kinetics for the various isomeric reactions. Reactions
involving the –CH2OH group, presumably via a 6-membered ring ortho-ester intermediate, are facile
and the a-glucoside anomers are significantly more reactive than their b-counterparts. By comparison
with degradation rates for the corresponding acyl glucuronides, it can be inferred that substitution of
the carboxylate by –CH2OH in the acyl glucosides has a significant effect on acyl migration for those
compounds, especially for rapidly transacylating molecules, and that thus the charged glucuronide
carboxylate is a factor in the kinetics.
N-glucoside of bromfenac3 2, as well as sulfonamides,4 5-
INTRODUCTION
aminosalicylic acid5 and 4-bromoaniline.6 Interestingly a number
of the N-glucosides found in biological samples exist as mixtures
of both a– and b–anomers6 and may owe their existence to
simple chemical, rather than biochemical, reactions in the body.
Endogenous mammalian glucoside metabolites, notably O-acyl
glucosides (and galactosides) of bile acids such as compounds 3
(R1, R2 = H, OH) are also known.7,8
Although xenobiotic Phase 2 mammalian metabolic conjugation
generally involves the formation of 1-b-D-glucuronides,1
a
significant number of glucoside metabolites have also been
identified, including both O- and N-linked types. Examples
include the O-acyl glucoside of a selective endothelin ETA
antagonist, [(+)-(5S,6R,7R)-2-isopropylamino-7-[4-methoxy-2-
((2R)-3-methoxy-2-methylpropyl)-5-(3,4-methylenedioxyphenyl)
cyclopenteno [1,2-b] pyridine 6-carboxylic acid]2 1 and the
aDepartment of Chemistry, The Robert Robinson Laboratories, Univer-
sity of Liverpool, Liverpool, UK, L69 7ZD. E-mail: andrew.stachulski@
bioch.ox.ac.uk; Tel: 44(0) 1865-275342
bBiomolecular Medicine, Department of Surgery and Cancer, Faculty of
Medicine, Imperial College London, Sir Alexander Fleming Building, South
Kensington, London, UK, SW7 2AZ
cDMPK, Astra Zeneca Pharmaceuticals, Mereside, Alderley Park, Maccles-
field, Cheshire, UK, SK10 4TG
† Current address: Hammersmith Hospital, Cyclotron Building, DuCane
Road, London, W12 0NN, UK
‡ Current address: Laboratory of Metabolism, National Cancer Institute,
US National Institutes of Health, Building 31, Bethesda, Maryland 20892,
USA
§ Current address: Institute of Glycobiology, Department of Biochemistry,
University of Oxford, South Parks Rd., Oxford, OX1 3QU, UK
As with acyl glucuronides,9,10 acyl glucosides are initially formed
as 1-b-esters as a consequence of the biosynthetic pathway. By
analogy with the corresponding glucuronides, it is anticipated
that transacylation and acyl migration of the acyl group in these
reactive metabolites will also take place (Scheme 1). Hydrolysis
of all of the isomers to the aglycone and glucose is also possi-
ble. Similar effects have also recently been shown using NMR
926 | Org. Biomol. Chem., 2011, 9, 926–934
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
It was therefore of considerable interest to synthesise acyl
glucosides and compare their reaction kinetics with those of
acyl glucuronides. In this way, the aim was both to extend
the understanding of the structure-reactivity relationships of
the compounds, varying here the carbohydrate rather than the
aglycone, and to extend the selective acylation method to other
carbohydrates. In particular, the replacement of the –COO- moiety
of the glucuronides with a neutral -CH2OH moiety could shed
light on the potential involvement of the –COO- group in the
transacylation and hydrolysis reactivity.
We have also extended the kinetic analysis by considering the
full reaction scheme that involves transacylation, mutarotation
and hydrolysis of all isomers. By computer modelling of these re-
actions, a highly detailed view of the reactivity can be determined.
Scheme 1 Reactivity pathways of acyl glucosides by analogy with acyl
glucuronides.
spectroscopy for acyl galactosides with the anomeric position
blocked as a galactoside.11 Clearly, for the normal hexopyranoses,
migration to the –CH2OH group is also feasible, and in this case
the cyclic ortho-ester intermediate will be a 6-membered ring and
not a 5-membered ring as for all other transacylation reactions.9
In their recent studies of acyl glucuronides12 Baba and Yoshioka
included some acyl glucoside examples which were prepared from
fully benzylated glucose precursors. Other groups have reported
syntheses from fully protected intermediates.13
EXPERIMENTAL METHODS
Synthesis
In our earlier studies17 we showed that glucuronate monoesters
were ideal substrates for selective 1-b-acylation, without the
need for global protection. By analogy, we considered that a
6-protected glucose derivative would be a suitable, minimally-
protected substrate and indeed 6-O-trityl glucose proved ideal.
The more reactive primary alcohol is protected and the molecule
is made soluble in organic media. Furthermore, the product
ester is anticipated to have good stability under the mildly acidic
conditions needed for detritylation.21 When phenylacetic acid 9
was reacted with 6-O-trityl glucose 10 under previously defined
The question of possible toxicity of acyl glucuronides has been
much debated and is clearly of importance in the assessment of
the safety of drugs such as ibuprofen,14 mycophenolic acid15 and
diclofenac16 which are extensively metabolised by glucuronidation,
-
followed by acyl migration as in Scheme 1 (with –CO2 replacing
–CH2OH in the case of those drugs). It has been surmised
that the transacylated positional isomers formed from 1-b-acyl
glucuronides can react via Schiff base formation with endogenous
proteins and hence cause toxicological and immunological adverse
effects.9,10
We have probed the reactivity of the acyl glucuronides of
phenylacetic acid and its derivatives with various degrees of a-
methyl substitution (4–7) using NMR analysis of the time course
of acyl migration. Additionally, using DFT-MO calculations
of the transition states of the acyl migration step we inferred
some structure-reactivity correlations and further elucidated the
mechanism.17
◦
conditions22,23 (HATU, NMM, MeCN, 20 C) a steady reaction
occurred andafter chromatography the product11 was obtained as
the b-anomer only in 55% yield. The racemic a–monomethyl acid
12 was likewise coupled and following diastereomer separation,
the products 13 and 14 were obtained in similar yields, again as
single b-anomers. We have assumed that, as in the acyl glucuronide
series,17 the (2R)-diastereoisomer is the less polar and this is
consistent with its faster acyl migration, vide infra. Finally, the
a,a–dimethyl acid 15 was also successfully reacted, giving 16 in
49% yield unoptimised.
These results bear directly on the behaviour of ibuprofen-like
NSAIDs (arylpropionates), where the structure-reactivity of their
acyl glucuronides is more complex than for benzoic acids.18–20
Interestingly, for the arylpropionate series, these results have
already shown that both aryl substitution [even by a remote
isobutyl group, as in ibuprofen (8)] and esterification of the
carboxylate moiety of the glucuronic acid have significant effects
on the degradation rates.14
Deprotection of the formed O-trityl adducts was conveniently
achieved by brief acidolysis (TFA in a 5 : 1 mixture of CH2Cl2
and water: cf. the deprotection of PMB glucuronates17) to deliver
the free acyl glucosides 17 to 20 in >88% yield in each case. All
compounds were obtained in excellent purity although 20 was
rather hygroscopic.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 926–934 | 927
©

View Article Online
Table 1 1H NMR chemical shifts (d, ppm) of selected protons of the
isomers of the acyl glucoside compounds 17–20 obtained from 1H NMR
analysis of the equilibrium reaction mixtures
Positional
isomera
Proton
17b
18b
19b
20b
1-b
1-a
2-b
2-a
3-b
3-a
4-b
4-a
6-b
H1
H1
H1
H1
H3
H3
H1
H1
H6
H6¢
H6
H6¢
5.59 (d)
6.12 (d)
4.79 (d)
5.36 (d)
5.00 (t)
5.20 (t)
5.17 (d)
—
5.56(d)
6.08(d)
—
5.25(d)
4.99(t)
5.18(t)
—
5.57(d)
6.11(d)
4.81(d)
5.36 (d)
4.98 (t)
5.18 (t)
—
5.57 (d)
6.09 (d)
—
5.33 (d)
5.00 (t)
5.19 (t)
—
RESULTS AND DISCUSSION
5.05(d)
5.14 (d)
—
—
Assignment of 1H NMR spectra
4.47 (dd) 4.50 (dd) 4.42 (dd)
4.30 (dd) 4.26 (dd) 4.30 (dd)
4.46 (dd) 4.47 (dd) 4.37 (dd)
4.34 (dd) 4.26 (dd) 4.35 (m)
The assignment of the 1H NMR peaks to the various hydrogens of
each isomer follows from previous studies on acyl glucuronides,
bearing in mind that the aglycone structure has little effect on the
carbohydrate moiety chemical shifts.24 In particular the chemical
shifts of the 1-b isomers can be made by inspection based on
6-a
—
Glucose
Aglycone CH3
methyl
H1 (a-isomer) 5.24 (d)
—
1.42 (d)
—
1.41 (d)
—
—
1.40 (2s)c
a 1-b denotes the 1-b-O-acyl-glucoside, and the other isomers are labeled
by analogy. b d – doublet, dd – doublet of doublets: t – triplet; m – multiplet.
c Methyl groups non-equivalent, two singlets
spin-coupling patterns, and the use of a H-1H COSY spectrum.
1
Selected assignments of anomeric protons and protons adjacent
to the O-acyl group are given in Table 1, and also shown on Fig. 1.
In particular the chemical shifts of the C6 protons in the 6-O-acyl
isomers are determined as these are used to monitor the kinetics
of the 4-O-acyl to 6-O-acyl reversible reactions, and hydrolysis of
the 6-O-acyl isomers. The intensities of these peaks were then used
for the subsequent kinetic analysis.
1-b- acyl glucoside degradation rates
The half-life degradation constants of 17–20 as measured by
1H NMR spectroscopy in phosphate buffer at pH 7.4 are given
in Table 2, where they are compared with the values for the
Fig. 1 600 MHz 1H NMR spectra of an equilibrium mixture following the degradation of the acyl glucosides 17–20. The assignments are as shown.
928 | Org. Biomol. Chem., 2011, 9, 926–934 This journal is The Royal Society of Chemistry 2011
©

View Article Online
Table 2 Degradation rate constants and half-lives of 1-b-O-acyl glu-
curonides (4–7) and 1-b-O-acyl glucosides (17–20)
subtle and is less than that caused by relatively small changes in
the aglycone, as observed by us and others.12,17,20,25
It is clear that, in the case of the glucuronides, carboxylate
is not acting as a general base because the rate is concentration-
independent as we noted previously.17 In the case of the glucosides,
no plausible general acid or base catalysis mechanism could be
proposed involving another molecule. The direct comparison be-
tween glucuronide carboxylate (charged) and glucoside (neutral)
is therefore important. It would appear from our results that it
is the transacylation step that is relatively accelerated for CO2 ,
rather than direct hydrolysis, and this can hardly be attributed to
a steric effect. We return to this point below.
Compound
kd/h-1a
t1/2/h
Acyl glucuronides
Acyl glucosides
4
5
6
7
2.353
0.903
0.405
0.029
0.29
0.78
1.71
23.3
17
18
19
20
1.419
0.782
0.476
0.043
0.49
0.89
1.46
16
-
a The rate constant is a composite value reflecting direct hydrolysis and
acyl migration; under the conditions employed, direct hydrolysis accounts
for <10% of the loss of the initial 1-b-isomer.17 The standard error for the
least squares linear fit to the slope are all 0.003 or less.
Based on the appearance of hydrolysis products in the NMR
spectra, the proportion of the reaction that can be attributed
to hydrolysis varies only slightly with methyl substitution, being
about 15%, 4%, 12% and 5% for 17–20 respectively.
corresponding acyl glucuronides.17 The linearity of the data fits
can be seen in Fig. 2. It will be seen immediately that the sets
of values are overall quite similar, with increasing bulk at the a-
position slowing the degradation and, as found in other series of
compounds, the (2R) compound 18 reacting about twice as fast as
the (2S) compound 19. In addition there are some other interesting
observations to be made. The degradation rate constants in the
two series are not linearly related and thus there is a differential
effect of the structural difference between a glucoside and the
glucuronide as a function of reactivity. For the unsubstituted and
hence fastest reacting compounds, the acyl glucoside has a half life
around 65% longer that the glucuronide, clearly implying that the
presence of the carboxylate function increases the reactivity, and
this is in agreement with our previous finding in the glucuronide
series in cases where the carboxylate function had been esterified.14
Specifically, the ethyl and allyl esters of the1b-acyl glucuronide of
Full kinetic modelling
The average rate constants, obtained through sequential pertur-
bation of the best fit rate constants, with a k value start estimate
of 1 h-1, for each acyl migration reaction for each acyl glucoside
compound, 17–20, are presented in Table 3. A plot of the best-
1
fit calculated intensities overlaid on the experimental H NMR-
determined concentration curves for acyl migration of the acyl
glucoside compound 17 is shown in Fig. 3 for each migratory step.
For certain isomers of some of the compounds, high fitting errors
caused by a low abundance of a given isomer at a particular time
point means that not all rate constants can be determined reliably.
It should be noted that there is good agreement between the
full kinetic analysis and the simple degradation rates in those
cases where both transacylation and hydrolysis can be modeled.
For example, for the combined 1-b to 2-b transacylation and
hydrolysis of the 1-b isomer, the rate constant is 1.42 h-1 from
the simple degradation measurement and 1.46 h-1 by combining
the rate constants from the detailed kinetic modeling (see Tables
3 and 4).
Other points to note are the generally more facile migration
involving the primary alcohol because of the formation of a 6-
membered, trans-decalin like ring intermediate, and the very fast
formation of the 2-a isomer from the 1-a isomer (whilst noting the
faster rate also for the reverse reaction), but of course only once
the 2-b isomer has been formed and allowed to mutarotate to the
2-a form. In general the 1-a isomers all show rapid reaction rates,
faster than their corresponding 1-b isomers. This can be attributed
to the well-known exo-anomeric effect.26 Additionally, for 1a→2
migration and the reverse reaction the transition state involves cis-
fused five and six membered rings. In the case of 1b→2 migration,
the transition state involves trans fusion and this requires higher
activation energy. In Scheme 3 we summarise the three classes of
transition state alluded to above; rearrangements between the 2
and 3, or 3 and 4 hydroxyls are essentially similar to the 1b→2
case, involving trans-6,5-fused rings.
(S)-ibuprofen had t1/2 = 7.2 and 9.4 h respectively compared to t1/2
=
3.8 h for the parent acyl glucuronide.14 For the slowest reacting
a,a¢-dimethyl compounds where the half lives are many hours,
the reverse was observed; now the glucuronide has the slower
degradation kinetics. The carbohydrate structural effect is thus
It is also instructive to contrast our results with those for the
series of acyl galactosides, referred to above,11 where the anomeric
position was permanently blocked as a b-benzyl galactoside.
Although the authors do not comment in detail on the transition
states involved, it is noticeable that at a pD = 8.0 (their closest
value to our pH of 7.4) the fastest migrations of acetyl and
Fig. 2 The degradation of the 1-b- isomers of the acyl glucosides and
glucuronides as a function of time. The various compounds are identified
by number. Key to compound identification: ᭛ - 4; ꢀ - 5; - 6; ꢀ -7; ᭜
᭺
- 17; ꢁ - 18; - 19; ꢀ - 20.
᭹
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 926–934 | 929
©

View Article Online
Table 3 Average rate constants (h-1) for acyl migration of 1-b-O-acyl glucoside compounds (17–20)and their positional isomers calculated based on the
kinetic model given in the Experimental Methods section and Scheme 2
17
18
19
20
Reaction
k (¥102)
s.d.
k (¥102)
s.d.
k (¥102)
s.d.
k (¥102)
s.d.
1b→2b
2b→3b
3b→2b
3b→4b
4b →3b
4b→6b
6b→4b
1a→2a
2a→1a
2a→3a
3a→2a
3a→4a
4a→3a
4a→6a
6a→4a
141.5
120.2
65.8
94.5
17.4
145.8
133.2
974.4
153.4
64.7
48.8
70.1
58.3
692.7
*
0.0
0.4
75.1
80.0
114.6
62.1
57.3
**
326.1
667.9
173.7
36.2
51.8
65.3
*
4.6
4.3
47.3
40.3
46.4
99.4
853.8
125.9
*
935.0
134.3
12.4
*
1.0
1.7
9.7
3.4
136.9
6.4
*
167.2
11.5
4.9
*
1.5
*
130.8
125.6
*
4.2
*
155.0
*
*
*
0.1
*
88.7
*
*
49.4
*
*
18.9
20.0
16.4
75.8
74.7
88.4
15.4
3.8
17.9
20.3
52.4
248.6
*
18.5
13.9
28.9
**
302.5
107.6
17.0
6.2
12.1
16.9
*
140.0
133.9
106.6
*
*
*
*
*
*
*
*
*
*
*
*
*
22.3
*
936.8
519.7
663.4
747.8
Due to the fast rates of the mutarotation reactions, these were not modelled and fixed fast values were used with appropriate ratios of values based on
the observed proportions of the various anomers. Isomer identification as for Table 2; * - denotes large uncertainty in the determination of k values; ** -
denotes k values approximating zero
Fig. 3 Full kinetic modelling of [17]. The dashed lines are the peak areas (relative concentrations) of positional isomers observed from the NMR profiles.
The solid lines are the peak areas (relative concentrations) predicted from the best-fit kinetic model.
benzoyl groups were to and from the 4-OH group, which is axial
in galactose. Here again, cis-fused 5 and 6-membered rings are
involved. Here, however, migrations to and from the anomeric
position could not be observed and we suggest the particularly
high value of k for the 1a→2 migration in our compounds reflects
both the favourable cis-ring fusion in the transition state and the
expulsion of a better leaving group, the anomeric OH being a
significantly stronger acid.
2→3 and 3→4 rearrangement (forward or reverse) the differences
are minor, but migration from 4→6 again shows a much higher
rate for the a-anomers. Clearly the reaction intermediate’s greater
ring size has an effect on this particular migration but the
stereoelectronic implications are not clear: also the standard
deviations are much higher for these cases compared to those
between O(1) and O(2).
Given the good agreement between the full analysis and the
simple degradation measurement, it is possible to interpret the
latter with some confidence. Thus, comparing compounds 4–7
In cases where the ester is no longer at O(1), the acyl migration
rate differences between a/b hemiacetals are less predictable. For
930 | Org. Biomol. Chem., 2011, 9, 926–934
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Scheme 2 The differential equations describing the various reaction rates (transacylation, mutarotation and hydrolysis) of the acyl glucosides. Here [1b],
for example, denotes the concentration of the b-1-O-acyl glucoside isomer, G denotes free glucose, a hydrolysis product, and k1b-2b is, for example, the
reaction rate constant for the internal transacylation reaction where the acyl group is transferred from the 1-b position to the 2-b position of the glucose
moiety.
Scheme 3 The three classes of transition states appropriate to the acyl migration steps: 1b to 2 [trans-6,5-fused rings] A, 1a to 2 [cis-6,5-fused rings] B,
and 4 to 6 [trans-decalin like] C. In C the anomeric configuration may be a or b.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 926–934 | 931
©

View Article Online
Table 4 Selected average rate constants (h-1) for hydrolysis of acyl
glucoside compounds 17–19 and their positional isomers calculated based
on the kinetic model given in the Experimental Methods and Scheme 2; s.
d. = standard deviation. For compound 20 the rate (not shown) was very
slow
were visualised by UV (254 nm) and anisaldehyde stain where
appropriate. Flash column chromatography was carried out using
ICN (230–400 mesh) silica gel. Optical rotations ([a]D) were
carried out on a PerkinElmer polarimeter 343 at a wavelength
of 598 nm at 20 ◦C. Solvents and concentrations used were as
quoted.
17
18
19
k (*102)
s.d.
k (*102)
s.d.
k (*102)
s.d.
Acylation of 6-O-trityl glucose 10: General Method
1b
1a
6a
4.6
—*
0.6
3.1
—*
0.5
3.9
53.2
—*
0.4
10.8
—*
6.0
88.9
—*
0.1
6.8
—*
To a solution of 6-O-trityl glucose 1027 (0.71 mmol) in MeCN
(8 ml) was added HATU (0.78 mmol), and the respective
carboxylic acid (0.78 mmol). The reaction was then flushed with
N2, then NMM (1.42 mmol) was added while stirring at 20 ◦C.
After 4 h, the reaction was quenched by addition of Amberlite H+
resin (1.42 mmol) and stirred for 30 min. The resin was filtered
off and washed with MeCN, then the solvent was removed in
vacuo and the residue purified by column chromatography using
5% EtOH–DCM. The a-dimethyl substituted carboxylic acid 15
required a longer reaction time (16 h).
* denotes large uncertainty in the determination of k values.
with 17–20, removal of the carboxylate function causes the overall
reaction to slow considerably for the fast reacting compounds
but to speed up for the slower reacting compounds. This can be
interpreted as the loss of the carboxylate causing a slowing of the
transacylation but having no effect on hydrolysis. A similar effect
had been observed previously for the acyl glucuronide of ibuprofen
where, if the free carboxylate is esterified, then the transacylation
also slows.14 However, in that case, it was not possible to exclude a
possible steric as opposed to electronic effect on the reaction rate.
In the acyl glucoside series, it is highly unlikely that a steric effect
could be responsible.
Some selected hydrolysis rates could be extracted from the
kinetic analyses, but in many cases the slow reactions meant that
there was no significant variation in the concentration of the
hydrolysis products over time. The determined values are given
in Table 4. It can be seen that the hydrolysis rates for the various
1-b isomers are comparable (except for compound 20, where the
reaction was very slow and a value could not be measured). The
hydrolysis rates of the 1-a isomers are considerably faster.
In summary, therefore, this study has shown that acyl glucoside
conjugates of a series of phenylacetic acids with varying a-
methylation can undergo acyl migration at rates comparable
to their related acyl glucuronides. The reaction rate constants
for a glucoside and a glucuronide of a given aglycone show
marked differences and this implicates the carboxylate group of
the glucuronic acid in the transacylation mechanism.
6-O-Trityl 1-(2-phenyl)-acetyl-b-D-glucopyranose 11
Yield: 55%. Found: C, 72.9; H, 6.1; [M+Na]+, m/z 563.205.
C33H32O7 requires C, 73.3; H, 6.0%; C33H32O7Na requires m/z
563.204; 1H NMR dH [CD3CN, 400 MHz]: 3.12–3.16 (1H, dd, 6-
H, J = 10.2, 5.7 Hz), 3.22–3.25 (1H, dd, 6-H¢, J = 10.2, 2.0 Hz),
3.34–3.38 (2H, m, 2-H, 4-H), 3.53-3.55 (2H, m, 3-H, 5-H), 3.77
(2H, s, CH2Ph), 5.49–5.51 (1H, d, 1-H, J = 7.8 Hz), 7.22–7.34
(14H, m, ArH), 7.41–7.44 (6H, m, ArH); 13C NMR dC [CD3CN,
100 MHz]: 17.4, 40.1, 56.6, 62.8, 69.5, 72.2, 75.8, 76.2, 85.8, 95.3,
126.7, 127.4, 128.2, 129.1, 133.6, 143.7, 170.1; m/z (ES +ve ion
mode) 563 ([M+Na]+, 100%).
6-O-Trityl 1-([(2R)-phenyl]propanoyl)-b-D-glucopyranose 13
Yield: 53%. Found: C, 70.6; H, 6.6; [M+Na]+, m/z 577.220.
C34H34O7·H2O requires C, 71.3; H, 6.3%; C34H34O7Na requires
m/z 577.220; 1H NMR dH [(CD3)2CO, 400 MHz]: 1.54–1.56 (3H,
d, CH(CH3), J = 7.2 Hz), 3.24–3.28 (1H, dd, 6-H¢, J = 9.9, 4.2 Hz),
3.38–3.41 (1H, dd, 6-H, J = 9.9, 2.0 Hz), 3.42–3.50 (2H, m, 2-H,
4-H), 3.57–3.64 (2H, m, 3-H, 5-H), 3.87–3.93 (1H, q, CH(CH3)),
4.23 (1H, m, OH), 4.46 (1H, d, OH), 4.49 (1H, m, OH), 5.58–5.59
(1H, d, 1-H, J = 7.8 Hz), 7.21–7.39 (12H, m, ArH), 7.40–7.41
(2H, m, ArH), 7.50–7.52 (6H, m, ArH); 13C NMR dC [(CD3)2CO,
100 MHz]: 19.9, 46.5, 64.7, 71.7, 74.2, 77.7, 78.7, 87.4, 96.2, 128.2,
128.2, 128.9, 128.9, 129.8, 130.1, 141.8, 145.6, 174.2; m/z (ES +ve
ion mode) 577 ([M+Na]+,100%).
Experimental
Synthetic intermediates were purchased from Sigma–Aldrich
group and used as supplied; the preparation of the known
compound 10 is referenced. All organic solvents were anhydrous
and of AR grade. Moisture sensitive reactions were carried out
in anhydrous organic solvents, under a nitrogen atmosphere.
Vacuum rotary evaporation was carried out at <35 ◦C. 1H
NMR and 13C NMR spectra were recorded on a Bruker AMX
400 spectrometer (at 400 MHz for 1H and 100 MHz for 13C
spectroscopy). Chemical shifts (d) are given in ppm from internal
tetramethylsilane (TMS) and coupling constants (J) are quoted in
Hertz (Hz). Mass spectrometry was carried out on a VG analytical
7070E machine, using electron ionisation (EI). A Micromass LCT
spectrometer was used for HRMS in the electrospray mode, and
elemental analysis was performed using a Thermo Flash EA1112
analyzer, configured for automated elemental analysis. Thin layer
chromatography was carried out using Merck 5 ¥ 2 cm aluminium
backed plates with a 0.2 mm layer of Kieselgel 60 F254. They
6-O-Trityl 1-([(2S)-phenyl]propanoyl)-b-D-glucopyranose 14
Yield: 53%. 1H NMR dH [(CD3)2CO, 400 MHz]: 1.49–1.50 (3H, d,
CH(CH3) J = 7.1 Hz), 3.16–3.20 (1H, dd, 6-H¢, J = 9.9, 5.9 Hz),
3.34–3.37 (1H, dd, 6-H, J = 9.9, 2.0 Hz), 3.38–3.50 (2H, m, 2-H,
4-H), 3.55–3.62 (2H, m, 3-H, 5-H), 3.88–3.94 (1H, q, CH(CH3)),
5.60–5.62 (1H, d, 1-H, J = 7.8 Hz), 7.20–7.38 (12H, m, ArH),
7.40–7.41 (2H, m, ArH), 7.50–7.52 (6H, m, ArH); 13C NMR dC
[(CD3)2CO, 100 MHz]: 20.1, 46.4, 64.6, 71.6, 74.2, 77.7, 78.6, 87.3,
96.2, 128.1, 128.2, 128.8, 128.9, 129.9, 130.0, 142.0, 145.6, 174.1.
932 | Org. Biomol. Chem., 2011, 9, 926–934
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
◦
1
349.1263; [a]D293 K = +9.2 (c = 0.006 gml-1, MeOH); H NMR dH
[(CD3)2CO, 400 MHz]: 1.55 (3H, s, C(CH3)), 1.59 (3H, s, C(CH3)),
3.29–3.34 (1H, m, 2-H), 3.38–3.42 (2H, m, 4-H, 5-H), 3.48–3.52
(1H, m, 3-H), 3.67–3.72 (1H, m, 6-H), 3.80–3.83 (1H, m, 6¢-H),
5.54–5.56 (1H, d, 1-H, J = 8.1 Hz), 7.21–7.25 (1H, m, ArH),
7.30–7.35 (2H, m, ArH), 7.39–7.42 (2H, m, ArH); 13C NMR dC
[(CD3)2CO, 100 MHz]: 27.2, 27.8, 47.9, 63.0, 71.7, 74.2, 78.4, 78.8,
96.3, 127.1, 127.8, 129.5, 146.0, 176.2; m/z (ES +ve ion mode) 349
([M+Na]+, 100%).
6-O-Trityl 1-(2,2-dimethylphenyl)acetyl-b-D-glucopyranose 16
Yield: 49%. Found: [M+Na]+, m/z 591.283; C35H36O7Na requires
m/z 591.236; 1H NMR dH (CD3CN, 400 MHz): 1.58 (3H, s,
C(CH3)), 1.64 (3H, s, C(CH3)), 3.0–3.14 (1H, dd, 6-H), 3.21–3.38
(3H, m, 6-H¢, 2-H, 4-H), 3.51-3.56 (2H, m, 3-H, 5-H), 5.51–5.53
(1H, d, 1-H, J = 8.1 Hz), 7.22–7.32 (14H, m ArH), 7.40–7.45 (6H,
m, ArH); 13C NMR dC (CD3CN, 100 MHz): 25.1, 26.3, 46.3, 62.7,
69.7, 72.1, 75.7, 76.3, 85.8, 94.3, 125.4, 126.5, 126.7, 127.5, 128.1,
128.3, 143.7, 144.1, 174.9; m/z (ES +ve ion mode) 591 ([M+Na]+,
100%).
Reactivity monitored by 1H NMR spectroscopy
Deprotection of the 6-O-trityl group by acid hydrolysis: General
method
1H NMR spectroscopy was carried out on the acyl glucosides at
310 K using a standard water peak presaturation pulse sequence.28
A Bruker AVANCE600 spectrometer (Bruker Biospin, Rheinstet-
ten, Germany) was used with a 5 mm broadband inverse (BBI)
To a solution of the trityl protected compound (0.4 mmol) in DCM
(4 ml) was added water (0.8 mmol) followed by TFA (0.4 mmol)
with vigorous stirring at RT. The reactions were monitored by
TLC and showed completion in 30 min. The solvent was removed
in vacuo and the residue purified by column chromatography using
10–20% EtOH–DCM.
1
probe operating at a H observation frequency of 600.13 MHz.
Once the acquisition parameters were optimised using a standard
sample, 550 mL of 100 mM sodium phosphate buffer (pH 7.4) was
added quickly to one of the samples containing 3.5, 3.3, 3.0 and
3.3 mg of the acyl glucoside compounds 17–20 respectively. The
sample was transferred to a 5 mm NMR tube containing 50 mL
sodium trimethylsilyl[2,2,3,3-2H4]propionate (TSP) (0.5 mg mL-1
in 2H2O) and 16 scans were acquired with a 1H spectral width of
20.02 ppm. The spectrum was phased manually, and an automated
program was run by which 159 spectra were acquired sequentially.
The first 61 experiments were acquired with a time delay of
37.8 s between experiments; the subsequent 95 experiments were
acquired with a time delay of 812.9 s between each experiment.
The total reaction times for the two segments are 1.92 and 23.45
h respectively. For 20, the automated experiment was repeated as
the reaction products had not appeared to any significant amount
after 25 h. An exponential apodization function corresponding to
a line-broadening factor of 0.3 Hz was applied to the FIDs. Spectra
were chosen at different time points, and manually phase- and
baseline-corrected. The anomeric proton doublet of the various
acyl glucosides was chosen to monitor the reactions. Assignment
of the NMR spectra of the positional isomers has been given
previously for acyl glucuronides24 and those for acyl glucosides
follow closely except for the 5- and 6-position protons. The
assignments for the C6 CH2 protons and the C5-H proton were
made by inspection, based on the known greater proportion of the
b-anomer and were confirmed by measurement of 2-dimensional
COSY spectra. Peak integration was used to measure the intensity
of the NMR resonances which were referenced to the TSP peak
intensity set at unity. The logarithms of the intensities were plotted
against time, and the slope of the graph gave the degradation rate
constant assuming first-order kinetics.
1-(2-phenyl)acetyl-b-D-glucopyranose 17
Yield: 92%. Found: C, 56.3; H, 6.1; [M+Na],+ m/z 321.095.
C14H18O7 requires C, 56.4; H, 6.1%; C14H18O7Na requires m/z
321.095; [a]D293 K = +1.8◦ (c = 0.011 gml-1, MeOH); H NMR dH
1
(CD3CN, 400 MHz): 3.26–3.31 (2H, m, 2-H, 4-H), 3.33–3.41 (2H,
m, 3-H, 5-H), 3.56–3.60 (1H, dd, 6-H, J = 12.0, 5.2 Hz), 3.70–3.74
(1H, dd, 6-H¢, J = 12.0, 2.5 Hz) 3.74 (2H, s, CH2Ph), 5.45–4.47
(1H, d, 1-H, J = 8.1 Hz), 7.29–7.37 (5H, m, ArH); 13C NMR
dC (CD3CN, 100 MHz): 40.0, 61.0, 69.5, 72.2, 75.9, 76.8, 94.1,
126.8, 128.2, 129.2, 133.6, 170.2; m/z (ES +ve ion mode) 321
([M+Na]+,100%).
1-([(2R)-phenyl]propanoyl)-b-D-glucopyranose 18
Yield: 90%. Found: [M+Na]+, m/z 335.1095; C15H20O7Na requires
m/z, 335.1107; [a]D293 K = -7.2◦ (c = 0.017 gml-1, MeOH); 1H NMR
dH [(CD3)2CO, 400 MHz]: 1.47–1.49 (3H, d, CH(CH3), J = 7.2 Hz),
3.32–3.36 (1H, t, 2-H, J = 8.6 Hz), 3.44–3.53 (3H, m, 3-, 4-, 5-H),
3.69–3.73 (1H, dd, 6-H, J = 11.8, 4.4 Hz), 3.80–3.86 (2H, m, 6-H¢,
CH(CH3)), 5.51–5.53 (1H, d, 1-H, J = 8.2 Hz), 7.25–7.30 (5H, m,
ArH); 13C NMR dC [(CD3)2CO, 100 MHz]: 19.7, 46.4, 62.8, 71.5,
74.2, 78.4, 78.8, 96.0, 128.2, 128.9, 129.7, 141.8, 174.1; m/z (ES
+ve ion mode) 335 ([M+Na]+, 100%).
1-([(2S)-Phenyl]propanoyl)-b-D-glucopyranose 19
Yield: 88%. Found: [M+Na]+, m/z 335.111; C15H20O7Na requires
m/z 335.1107; 1H NMR dH [(CD3)2CO, 400 MHz]: 1.48–1.49 (3H,
d, CH(CH3), J = 7.15 Hz), 3.36–3.41 (2H, m, 2-H, 3-H), 3.43-3.48
(1H, m, 5-H), 3.52–3.57 (1H, t, 4-H, J = 8.95 Hz), 3.62–3.66
(1H, dd, 6-H, J = 12.1, 5.2 Hz), 3.78–3.81 (1H, dd, 6-H¢, J =
12.1, 2.3 Hz), 3.88–3.92 (1H, q, CH(CH3), J = 7.1 Hz), 5.51–5.53
(1H, d, 1-H, J = 8.1 Hz), 7.27–7.39 (5H, m, ArH); 13C NMR dC
[(CD3)2CO, 100 MHz]: 19.7, 46.3, 62.3, 70.9, 73.7, 77.6, 78.6, 96.0,
128.4, 128.9, 129.9, 141.6, 174.9.
The % of the reaction due to hydrolysis was obtained by
integration of the aglycone methyl doublets of 18 and 19 at ~
d 1.41, and the singlet resonance from the two methyl groups of
20 at d 1.47. For 17, the free a-glucose anomeric proton resonance
at d 5.25 was integrated.
Detailed kinetic modelling of all reactions
1-(2,2-dimethylphenyl)acetyl-b-D-glucopyranose 20
The kinetic models used to describe the acyl migration, anomer-
ization and hydrolysis reactions of the acyl glucoside compounds
17–20 are given in Scheme 2. First order irreversible reactions were
Yield: 89%. Found: C, 58.71; H, 6.9; m/z, [M+Na]+ 349.1266;
C16H22O7 requires C, 58.9; H, 6.8%; C16H22O7Na requires m/z,
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 926–934 | 933
©

View Article Online
used to describe the 1-b to 2-b acyl migration and hydrolysis, and
first order reversible reaction models were used to describe the
remainder of the acyl migration and all anomerization reactions.
Owing to the fast rate of the mutarotation reactions, these were
not modelled and fixed values based on the observed b- to a-
anomer ratios were used. Thus the calculated concentrations of all
species at all time points were obtained by solving the differential
equations given in Scheme 2.
The software GEPASI v.329 was used to solve the differen-
tial equations and estimate the rate constants by minimizing
the residual sum of squares between observed and calculated
peak intensities with the Levenberg–Marquardt gradient descent
method. A 15-point Savitsky–Golay filtering30 with a second-order
polynomial was applied to the 1H NMR reaction profiles of 17–20
to increase the signal-to-noise ratio prior to kinetic modeling. For
each of the 1H NMR spectra of 17–20 at each time point the water
peak region between d ~4.5–4.7 was removed to avoid the effects of
imperfect water suppression. A weighted least squares 10th order
polynomial baseline correction was applied between d ~0.5–4.5
and d ~4.7–7 respectively. The peak areas of selected resonances
for each positional isomer were measured relative to that of the
TSP peak set to unity, which was assumed to be constant through
the time period followed.
2 C. Y. Tang, J. H. Hochman, B. Ma, R. Subramanian and K. P. Vyas,
Drug Metab. Dispos., 2003, 31, 37–45.
3 S. K. Kirkman, M. Y. Zhang, P. M. Horwatt and J. Scatina, Drug
Metab. Dispos., 1998, 26, 720–723.
4 G. D. Paulson, J. M. Griddings, C. H. Lamoreux, E. R. Mansager and
C. B. Struble, Drug Metab. Dispos., 1981, 9, 142–146.
5 J. Tjornelund, S. H. Hansen and C. Cornett, Xenobiotica, 1989, 19,
891–899.
6 H. Major, J. Castro-Perez, J. K. Nicholson and I. D. Wilson, Xenobiot-
ica, 2003, 33, 855–869.
7 T. Iida, R. Nakamori, R. Yabuta, S. Yada, Y. Takagi, N. Mano, S.
Ikegawa, J. Goto and T. Nambara, Lipids, 2002, 37, 101–110.
8 G. Kakiyama, A. Hosoda, T. Iida, Y. Fujimoto, T. Goto, N. Mano, J.
Goto and T. Nambara, J. Chromatogr., A, 2006, 1125, 112–116.
9 A. V. Stachulski, J. R. Harding, J. C. Lindon, J. L. Maggs, B. K. Park
and I. D. Wilson, J. Med. Chem., 2006, 49, 6931–6945.
10 H. Spahn-Langguth and L. Z. Benet, Drug Metab. Rev., 1992, 24, 5–48.
11 M. U. Roslund, O. Aitio, J. Warna, H. Maaheimo, D. Y. Murzin and
R. Leino, J. Am. Chem. Soc., 2008, 130, 8769–8772.
12 A. Baba and T. Yoshioka, Chem. Res. Toxicol., 2009, 22, 158–172.
13 G. Kakiyama, S. Sadakiyo, T. Iida, K. Mushiake, T. Goto, N. Mano, J.
Goto and T. Nambara, Chem. Phys. Lipids, 2005, 134, 141–150.
14 C. H. Johnson, I. D. Wilson, J. R. Harding, A. V. Stachulski, L. Iddon,
J. K. Nicholson and J. C. Lindon, Anal. Chem., 2007, 79, 8720–8727.
15 M. Shipkova, V. W. Armstrong, E. Wieland, P. D. Niedmann, E.
Schutz, G. Brenner-Weiss, M. Voihsel, F. Braun and M. Oellerich, Br.
J. Pharmacol., 1999, 126, 1075–1082.
16 J. R. Kenny, J. L. Maggs, X. Meng, D. Sinnott, S. E. Clarke, B. K. Park
and A. V. Stachulski, J. Med. Chem., 2004, 47, 2816–2825.
17 L. Iddon, M. Iqbal, X. Meng, P. Jayapal, N. G. Berry, A. V. Stachulski,
C. H. Johnson, J. C. Lindon, J. R. Harding and I. D. Wilson, Org.
Biomol. Chem., 2009, 7, 2525–2533.
Abbreviations
18 U. G. Sidelmann, S. H. Hansen, C. Cavaghan, H. A. J. Carless, J. C.
Lindon, R. D. Farrant, I. D. Wilson and J. K. Nicholson, Anal. Chem.,
1996, 68, 2564–2572.
19 S. J. Vanderhoeven, J. C. Lindon, J. Troke, G. E. Tranter, I. D. Wilson
and J. K. Nicholson, Xenobiotica, 2004, 34, 73–85.
20 S. J. Vanderhoeven, J. Troke, G. E. Tranter, I. D. Wilson, J. K. Nicholson
and J. C. Lindon, Xenobiotica, 2004, 34, 889–900.
HATU
O-(7-azabenzotriazol-1-yl)-N,N,N¢,N¢-
tetramethyluronium hexafluorophosphate
N-methylmorpholine
NMM
PMB
4-methoxybenzyl.
21 Greene’s Protective Groups in Organic Synthesis (4th ed.), P. G. M. Wuts
and T. W. Greene, Wiley-Interscience , 2007, pp. 152-156.
22 E. R. Bowkett, J. R. Harding, J. L. Maggs, B. K. Park, J. A. Perrie and
A. V. Stachulski, Tetrahedron, 2007, 63, 7596–7605.
23 J. A. Perrie, J. R. Harding, D. W. Holt, A. Johnston, P. Meath and A.
V. Stachulski, Org. Lett., 2005, 7, 2591–2594.
24 R. D. Farrant, M. Spraul, I. D. Wilson, A. W. Nicholls, J. K. Nicholson
and J. C. Lindon, J. Pharm. Biomed. Anal., 1995, 13, 971–977.
25 A. Baba and T. Yoshioka, Chem. Res. Toxicol., 2009, 22, 1998–2008.
26 I. Tvaroska and T. Bleha, Adv. Carbohydr. Chem. Biochem., 1989, 47,
45–123.
Acknowledgements
We are grateful to the EPSRC and BBSRC, and AstraZeneca plc,
for funding (DTA awards to LI and CHJ respectively). SER is
funded by Nestle´ through an Imperial College London/Nestle
Alliance. We thank Mr. Alan Mills and Mr. Steve Apter of
The University of Liverpool for the MS and elemental analyses
respectively.
27 S. K. Chaudhary and O. Hernandez, Tetrahedron Lett., 1979, 20, 95–98.
28 O. Beckonert, H. C. Keun, T. M. D. Ebbels, J. Bundy, E. Holmes, J. C.
Lindon and J. K. Nicholson, Nat. Protoc., 2007, 2, 2692–2703.
29 P. Mendes, Computer Applications in the Biosciences, 1993, 9, 563–571.
30 A. Savitzky, Anal. Chem., 1964, 36, 1627–1639.
References
1 G. J. Dutton, Glucuronidation of Drugs and Other Compounds, CRC
Press, Boca Raton, FL, USA, 1980.
934 | Org. Biomol. Chem., 2011, 9, 926–934
This journal is
The Royal Society of Chemistry 2011
©