A convenient new synthesis of quaternary ammonium glucuronides of drug molecules

Article abstract of DOI:10.1016/j.tet.2009.10.113

N-Glucuronides, of various structural types, are frequently encountered as drug metabolites. Efficient chemical synthesis of these compounds, both as analytical standards and for toxicological investigation, is therefore an important goal. Earlier syntheses of N⁺- glucuronides of aliphatic tertiary amine drugs involved direct reaction of the drug molecule with a bromosugar, but yields were generally low and of poor reproducibility, with many by-products. In addition the final products were often of low stability, hindering effective isolation and purification. We now report that a stable, readily prepared glucuronic acid hemiacetal is a reliable precursor for metabolites of this type and give three pharmaceutically relevant examples. We report further on the stability of the final metabolites and the conditions required for their isolation and purification.

Full text of DOI:10.1016/j.tet.2009.10.113

Tetrahedron 66 (2010) 537–541
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
A convenient new synthesis of quaternary ammonium glucuronides
of drug molecules
,
Lisa Iddon a, Ryan A. Bragg b, John R. Harding b, Andrew V. Stachulski a
*
a Robert Robinson Laboratories, Department of Chemistry, University of Liverpool, Liverpool, L69 7ZD, UK
b Clinical Pharmacology and DMPK, AstraZeneca Pharmaceuticals, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 27 July 2009
Received in revised form 15 October 2009
Accepted 29 October 2009
Available online 1 November 2009
N-Glucuronides, of various structural types, are frequently encountered as drug metabolites. Efficient
chemical synthesis of these compounds, both as analytical standards and for toxicological investigation, is
therefore an important goal. Earlier syntheses of Nþ- glucuronides of aliphatic tertiary amine drugs involved
direct reaction of the drug molecule with a bromosugar, but yields were generally low and of poor re-
producibility, with many by-products. In addition the final products were often of low stability, hindering
effectiveisolationand purification. Wenowreportthata stable, readily preparedglucuronicacidhemiacetalis
a reliable precursor for metabolites of this type and give three pharmaceutically relevant examples. We report
further on the stability of the final metabolites and the conditions required for their isolation and purification.
Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction
generally specific human metabolites23 and therefore not detected
prior to clinical trials.
Xenobiotic metabolism is a fundamental process by which the
body is detoxified from drug molecules. In phase II metabolism,1,2
a functional group on the drug such as OH, NR2 or SH is conjugated
to a polar unit, yielding a water-soluble metabolite which is ex-
creted. Glucuronides,3,4 including O, N, S, and C-types, are by far the
most important class of phase II metabolites. In continuation of
a long-standing interest in glucuronides and efficient methods for
their synthesis,3b,5 we now report on the synthesis of quaternary
ammonium glucuronides of aliphatic tertiary amine-containing
drugs. At present there are no satisfactory general methods for this
important class.
O
S
N
NMe
2
O
N
Me
NMe
2
NMe
2
1
2
3
4
The stability of all classes of N-glucuronides is strongly pH de-
pendent.24,25 Thus the stability of neutral N-glucuronides is low in
weakly acidic media due to the available nitrogen lone pair.26 In
these conditions, the pyranose ring oxygen becomes protonated
and catalyses the formation of an iminium ion, which in turn
readily hydrolyses (Scheme 1).
With the increasing number of N-heterocyclic and amine-con-
taining drugs now marketed, N-glucuronides6–8 are being reported
9–11
more frequently. Neutral
and quaternary ammonium12,13 N-
glucuronidesresultfrom secondaryandtertiaryamines, respectively.
Heterocyclic N-glucuronides14 are also familiar, e.g. the pyridinium
Nþ-glucuronides15,16 derived from nicotine and analogues.
N-Glucuronidation is common for drugs such as antipsychotics,
antihistamines and tricyclic antidepressants, 17,18 accounting for up
to 25% of the initial dose of ketotifen19 1 and amitriptyline20 2, for
example.21 Further examples are the important anticancer agent
tamoxifen 322 (whose Nþ glucuronide is reported to be equiactive
with the parent) and imipramine 4, another antidepressant. It is
important to note that quaternary ammonium glucuronides are
HO C
H
O
2
HO C
2
HO C
2
+
slow
H
O
+
OH
HO
HO
HO
HO
HO
HO
N
N
:
+
N
HO
HO
HO
Strong
acid
fast
H O fast
2
HO C
2
HO C
2
O
N
O
HO
HO
H
HO C
2
H
O
N
HO
O
+
HO
HO
HO
HO
+ HNMe
2
OH
HO
anomerisation
hydrolysis
* Corresponding author. Tel.: þ44 151 794 3542; fax: þ44 151 7943588.
E-mail address: stachuls@liv.ac.uk (A.V. Stachulski).
Scheme 1. Acid catalysed hydrolysis of neutral N-glucuronides.
0040-4020/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2009.10.113

538
L. Iddon et al. / Tetrahedron 66 (2010) 537–541
By contrast, under strongly acidic conditions N-glucuronides are
probably lead to generation of hemiacetal 9, which could react
independently with 7. Less glucal 6 was formed in this reaction than
with tertiary amines. Reaction of desipramine 8 with 5 gave similar
yields of 8a (15–17%), but transacylation giving amide 10 was then
significant. The use of isobutyrate ester protecting groups31,32 re-
duced transacylation, but yields were not greatly improved.
Quaternisation of intermediates 7a/8a, as in the glucose series,29
required methyl triflate (MeI was insufficiently reactive), then the
products could be deprotected to give the desired free glucuronides
as described below. However, in view of the poor yields and lack of
generality of this approach we sought an improved synthesis.
generally more stable. Protonation on nitrogen forms a quaternary
ammonium species, rendering the iminium ion inaccessible (Scheme
1). Likewise, quaternary ammonium (zwitterionic) Nþ-glucuronides
are generally more robust in acidic media but may hydrolyse under
basic conditions to release the aglycone.17 N-Glucuronidation is fea-
sible for many, if not most, drugs, which contain an amine group, but
due to the labile nature of their glycosidic linkages, such metabolites
are probably unstable in the excretory media.7 Thus it is likely that
N-glucuronidation has been underestimated.
The recently issued FDA guidelines27 have emphasised the im-
portance of determining the safety of drug metabolites that are
either identified only in humans or are present at disproportion-
ately higher levels (viz those accounting for >10% of parent sys-
temic exposure, as measured by AUC) in humans than in any animal
species used in testing. In some cases it may be necessary to carry
out non-clinical testing of Nþ-glucuronide metabolites: this has
added impetus to the search for effective syntheses. Luo et al.18
reported a two phase system for the synthesis of these metabolites,
but this procedure has proved low-yielding or difficult to re-
peat.19,28 We now report an efficient new approach to Nþ-glucu-
ronide synthesis of some generality, illustrated by three examples.
2.2. Synthesis by oxidation of ND-glucosides
We recently showed the value of (6-O-trityl)glucose 11 in Nþ-
glucoside synthesis29: both the relatively lipophilic amine and the
sugar are made soluble in organic solvent. Reaction of 7 with 11,
followed by acetylation of free OH groups and quaternisation with
MeOTf, afforded the intermediate compound 12 with a free 6-OH in
acceptable yield (30% overall), Scheme 3.
Ph3CO
CH2OH
O
O
+
3 steps
(ref. 29)
AcO
AcO
HO
HO
N
OH
2. Results and discussion
TfO−
HO
AcO
11
12
2.1. Synthesis of ND-glucuronides via the bromosugar
Scheme 3. Nþ-Glucuronide synthesis via the corresponding glucoside.
We first studied the method of Luo et al.18 reacting bromosugar 5
and amitriptyline 2 in a two phase toluene/aq NaHCO3 mixture. The
major organic product from this reactionwas the glucal 6, formed by
E2 elimination owing to the basic action of 2. Direct substitution of 5
by heating with tertiary amines has also been employed, e.g., for the
Nþ-glucuronide of tamoxifen 3.22 The relative success of this re-
action (though still in<10% yield) can be explained by the lower pKa
We hoped to achieve our goal by oxidation of 12. However,
under a variety of conditions (using PDC, RuCl3–NaIO4 or TEMPO in
conjunction with NaBr/NaOCl or trichloroisocyanuric acid) the de-
sired carboxylic acid could not be obtained: generally aglycone
cleavage resulted. The use of TEMPO33 with PhI(OAc)2 as terminal
oxidant in MeCN did give some of the desired glucuronide (these
conditions generate acetic acid as a by-product, which our product
and intermediate will tolerate) but the product was not pure by
LCMS analysis. As well as oxidation of the primary alcohol, oxida-
tion occurred at another site, probably the alkene double bond,
both as well as and in competition with the desired reaction.
Nevertheless, this should be a viable approach for Nþ-glucuronides
of non-functional aglycones, specifically those without other oxi-
disable sites.
of tamoxifen 3, a b-alkoxy amine, compared to amitriptyline 2.
CO Me
2
CO Me
2
CO Me
2
O
AcO
AcO
O
O
AcO
AcO
AcO
AcO
OH
AcO
AcO
AcO
Br
5
6
9
We also employed secondary amines in the direct substitution
reaction, intending to quaternise later: we recently showed this to be
a good approach to Nþ-glucosides.29 Frequently tertiary amino drugs
are commercially available as their des-methyl derivatives, or deme-
thylation can be achieved using the Olofson reaction30: we used
nortryptiline 7 and desipramine 8, derived from 2 and 4 respectively.
The reaction of nortriptyline 7 with 5 (Scheme 2) under basic
conditions led to modest yields of 7a, up to 26%. The basic condi-
tions in this reaction, as well as removing in situ generated HBr,
2.3. Synthesis via a glucuronate ester hemiacetal
Instead a glucuronate hemiacetal (cf. note to Scheme 2, vide
supra) proved to be a robust precursor: in this way no oxidation
step is needed. We illustrate this method with syntheses of the Nþ-
glucuronides of both amitriptyline 2 and imipramine 4, tricyclic
antidepressants in regular use. In another example, we have pre-
pared a model Nþ-glucuronide bearing a different structural fea-
ture, namely a pyrrolidine unit.
In contrast to our Nþ-glucosidation studies,29 here the amine is
reacted successfully with an acylated hemiacetal: to reduce the risk
of transacylation, triisobutyrate 1331 (cf. 9) is preferred. This de-
rivative, a stable crystalline solid at 20 ꢀC, is available in two steps
from glucuronolactone (Scheme 4). Monoesters such as allyl or
benzyl glucuronate, which we used very successfully in acyl glu-
curonide synthesis,34 were ineffective here: on treatment with
primary and secondary amines they readily formed amides.
Nortriptyline 7 and desipramine 8 react readily with 13, giving
N
N
NHMe
NHMe
NMe.COMe
7
8
10
CO Me
2
O
CO Me
2
O
RNHMe 7 or 8, aq. Na CO
2 3
AcO
AcO
AcO
AcO
NMeR
DCM/ acetone
AcO
AcO
Br
selectively b-glycosylamines. This reaction cannot be monitored by
TLC because of the acid instability of the products, but by eva-
poration of aliquots and NMR analysis the conversion to the
5
7a, 8a
Scheme 2. Base-catalysed reaction of a bromosugar with secondary amines.

L. Iddon et al. / Tetrahedron 66 (2010) 537–541
539
_
_
O
CO
O
CO
O
2
2
CO Me
2
CO Me
2
O
O
O
O
i)
i
A
HO
HO
HO
+
Pr CO
2
NR
+
NR
3
i
3
3
i
HO
Pr CO
OH
+
TfO_
i
2
+
O CPr
2
Pr CO
2
i
HO
NR
HO
3
Pr CO
2
i
Pr CO
2
i
Pr CO
2
OH
_
_
2
14a, 14b or 17
B
CO
2
CO
O
O
CO Me
2
+
HO
HO
+ NR
3
CO Me
HO
HO
2
NR
O
+
i
OH
O
Pr CO
2
+
ii)
i
iii), iv), v)
Pr CO
HO
2
HO
i
NMe R
2
Pr CO
i
2
_
Pr CO
2
OH
i
TfO
Pr CO
i
2
Pr CO
2
Scheme 6. Hydrolysis conditions. A: 10% w/v Na2CO3, MeOH; B: 1 M LiOH or NaOH,
MeOH, ꢁ10 ꢀC.
13
14a, 14b
For amitriptyline and imipramine intermediates 14a and 14b,
either 1 M NaOH or 1 M LiOH was satisfactory: some loss of the
aglycone (Scheme 6, pathway B), but no elimination reaction, was
observed under these conditions. When we used 10% w/v Na2CO3
we found that the elimination by-product was significant (Scheme
6, pathway A) and this proved more difficult to remove from the
final product. The elimination reaction is a feature of the iso-
butyrates36b and is negligible when acetates are used; a character-
N
R in 14a
R in 14b
Scheme 4. Nþ-Glucuronide synthesis via the hemiacetal. i) MeOH, NEt3 (0.1 equiv),
trace AcOH, then evap., PriCOCl, py, DCM, 40 ꢀC, 59%; ii) N2H4$H2O, AcOH, DMF, 85%;
iii) 7 or 8, PhMe, 50 ꢀC; iv) pyridine, Ac2O, 0 ꢀC; v) MeOTf, DCM, 15–20% (when iso-
lated, see text).
istic alkene 1H NMR signal (
by-product.
d ca. 6.0, H-4) is observed for this
Conditions B could not be used for 17 owing to the more base-
labile nature of the glycosidic bond: up to 80% loss of aglycone was
observed. Instead, conditions A were satisfactory: up to 20% elim-
ination was observed, but loss of aglycone was completely avoided.
Finally the desired glucuronides 18 (16%), 19 (57%) and 20 (11% in
three steps from 13) were obtained in high purity.
glycosylamines was shown to be 80–90%. Typically these in-
termediates show a peak around
d 4.20 (1H, d, J ca. 8 Hz) for the
anomeric proton and only the -anomer is seen. Prior to quater-
b
nisation with MeOTf, unreacted OH and NH2 groups were acety-
lated using Ac2O/pyridine (otherwise release of TfOH led to
hydrolysis of the adducts); then satisfactory yields of quaternised
products were obtained. The quaternisation is accompanied by
a characteristic downfield shift of 0.5–1 ppm in the anomeric pro-
ton. Purification on silica was usually possible at this stage, though
with significant losses.29 Thus intermediates 14a and 14b were fully
characterised, but sometimes it was simpler to triturate with ether
after the MeOTf step, then after removing the non-polar material,
hydrolyse the residue directly (vide infra).
Hemiacetal 13 also reacts efficiently with cyclic secondary amines
such as pyrrolidine. The resulting glycosylamine 15 (Scheme 5) was
reacted with primary triflate 16 (prepared from the corresponding
alcohol and used immediately),35 giving adduct 17, thus extending
the scope of our quaternisations beyond methylation. The conver-
sion to 17 was good and this compound was fully characterised, but
the product was best hydrolysed directly without purification, cf.
above.
_
_
_
CO2
CO2
CO2
O
O
O
+
+
HO
HO
HO
HO
HO
HO
+
N
N
N
HO
HO
HO
N
OPh
19
18
20
Preparative reverse phase HPLC (aq MeCN, neutral conditions)
was necessary to obtain highly pure product 20, where elimination
occurred:silicachromatographywasacceptablefor 18 and19, though
losses are observed ineithercase. Nevertheless the method described
will conveniently deliver pure product in 50–100 mg amounts.
3. Conclusions
CO Me
2
CO Me
2
O
O
N
i
i
In summary, we have described a promising new method for the
synthesis of quaternary ammonium glucuronides from a stable,
readily available hemiacetal sugar. The inherently low stability of
this class of metabolites poses difficulties in both synthesis and pure
product isolation. Glycosylamine formation using secondary
amines, followed by quaternisation with a primary triflate, affords
the protected precursors in good yields although there are stability
issues with the intermediates. Successful hydrolysis can be achieved
for both linear and cyclic quaternary esters: optimum conditions
depend on the aglycone. The final products can be obtained in high
purity using preparative HPLC. We are continuing to refine the hy-
drolysis and purification steps, but we believe the method described
will be of great value in the pharmaceutical industry as a practical
synthesis of quaternary ammonium glucuronide metabolites.
i)
Pr CO
Pr CO
2
2
i
i
Pr CO
Pr CO
2
2
OH
i
i
Pr CO
Pr CO
2
2
13
15
CO Me
2
O
i
ii)
OPh
16
TfO
_
Pr CO
i
2
+
TfO
N
Pr CO
2
i
Pr CO
2
17
OPh
Scheme 5. Synthesis of a pyrrolidinium Nþ-glucuronide. i) Pyrrolidine (1.4 equiv),
toluene, 50 ꢀC, 100% yield; ii) 16, DCE, 50 ꢀC, 19% (chromatographed).
2.4. Hydrolysis
Finally, careful base hydrolysis released the zwitterionic Nþ-
glucuronides (Scheme 6). The final products are sufficiently base-
stable to tolerate these conditions for a short time, but the optimum
conditions depend on the aglycone. Two side reactions are possi-
ble: elimination of the 4-isobutyrate can occur (pathway A) or the
aglycone can be liberated (pathway B). The elimination reaction
was known from early work on degradation of glucuronic acid
polymers and may also be observed in O-glucuronide synthesis.36,37
4. Experimental
4.1. General experimental methods
All organic solvents were anhydrous and of AR grade. Vacuum
rotary evaporation was carried out at <30 ꢀC. Analytical thin-layer
chromatography was performed using Merck Kieselgel 60 F 254

540
L. Iddon et al. / Tetrahedron 66 (2010) 537–541
silica plates; preparative column chromatography was performed
on Merck 9385 silica gel. Optical rotations were measured on
a Perkin Elmer model 343 polarimeter at 589 nm and 20 ꢀC. Infra-
red spectra were obtained using a Jasco FT/IR-4200 instrument.
Both 1H and 13C NMR spectra were recorded for the solvents noted
using either Bruker 250 MHz or 400 MHz instruments (the latter
operating at 100 MHz for 13C observation) with tetramethylsilane
as internal standard. Mass spectra in the chemical ionisation (CI)
mode were obtained using a VG7070E mass spectrometer. Both low
and high resolution electrospray mode (ES) mass spectra were
obtained using a Micromass LCT mass spectrometer operating in
the þve or –ve ion mode as indicated. Elemental microanalysis was
performed by Mr. Steve Apter (Liverpool). 3-[2,2,3,3-2H4] Trime-
thylsilyl propionate sodium salt (TSP), sodium dihydrogen phos-
phate and disodium hydrogen phosphate, were purchased from
Sigma–Aldrich Company, Ltd (Gillingham, Dorset, UK). HPLC–NMR
grade deuterium oxide (2H2O) was obtained from Goss Scientific
Instruments (Essex, UK).
3ꢂCH(CH3)2], 2.95 (3H, s, NþCH3), 3.08 (3H, s, NþCH3), 3.18 (4H, br s,
CH2CH2), 3.4–3.5 (1H, m, NCH2CH2CH2), 3.55–3.65 (1H, m,
NCH2CH2CH2), 3.76 (3H, s, CO2CH3), 3.8–3.98 (2H, m, NCH2CH2CH2
Nþ), 4.5–4.53 (1H, d, J¼9.5 Hz, 5-H), 5.09–5.18 (1H, t, 4-H), 5.30–5.50
(2H, 2ꢂt, 2-Hþ3-H), 5.60–5.65 (1H, d, J¼8.6 Hz, 1-H) and 6.90–7.23
(8H, m, Ar); 13C NMR (100 MHz, CDCl3): 18.5, 18.8, 19.0, 21.1, 32.5,
34.1, 34.2, 34.3, 46.8, 48.3, 49.0, 53.4, 60.9, 67.4 (4-C), 68.2 (3-C), 72.8
(2-C), 74.1 (5-C), 93.6 (1-C), 120.0, 123.9, 127.3, 130.7, 134.7, 147.3,
166.3, 175.1 and 175.9 (ꢂ2); m/z (ES þve ion mode) 681 [M]þ 100%.
4.3. Methyl [N-(2,3,4-tri-O-isobutyryl)-b-D-glucopy-
ranuronato]-pyrrolidine 15
To a solution of 1331 (1.09 g, 2.6 mmol) in toluene (8 mL) under
a nitrogen atmosphere was added pyrrolidine (0.26 mL, 0.22 g,
3.12 mmol). The reaction was heated to 50 ꢀC for 6 h, then solvent
and excess reagent were removed in vacuo to give the product 15 as
an oil, which was sufficiently pure to use directly (1.23 g, 100%).
Found: [MþH]þ m/z, 472.256; C23H38O9N requires m/z, 472.255; 1H
NMR (400 MHz, CDCl3): 0.98–1.2 (18H, m, 3ꢂCH(CH3)2), 1.60–1.61
(4H, m, CH2CH2), 2.28–2.42 (3H, m, 3ꢂCH(CH3)2), 2.68–2.84 (4H, m,
CH2NCH2), 3.67 (3H, s, CO2CH3), 3.90–3.93 (1H, d, J¼10.0 Hz, 5-H),
4.28–4.31 (1H, d, J¼9.3 Hz, 1-H), 5.03–5.14 (2H, 2ꢂt, 2-Hþ4-H) and
5.20–5.26 (1H, t, 3-H); 13C NMR (100 MHz, CDCl3): 17.7, 17.9, 17.9,
23.5, 32.8, 32.9, 32.9, 45.7, 51.5, 67.9, 68.7, 71.7, 73.0, 89.2, 174.3,
174.5, 174.9 and 181.6; m/z (ES þve ion mode) 472 [MþH]þ 100%.
4.2. General method for the synthesis of protected
quaternary ammonium glucuronides
To a solution of 1331 (0.50 g, 1.19 mmol) in toluene (5 mL) at
50 ꢀC was added nortriptyline 7 or desipramine 8 (1.19 mmol) as
a solution in toluene (5 mL). The reaction was heated to 50 ꢀC under
a nitrogen atmosphere for 24 h. The toluene was then removed in
vacuo and the residue dissolved in pyridine (15 mL). The solution
was cooled to 0 ꢀC and placed under nitrogen. Acetic anhydride
(0.24 g, 2.39 mmol) was then added slowly to the reaction mixture,
which was left stirring at 0 ꢀC for 4 h. The reaction was then
quenched with satd aq NaHCO3 until pH 7–8 was achieved: DCM
(20 mL) was added and the organic layer was separated. The
aqueous phase was then extracted with further DCM (3ꢂ50 mL)
and the organic layers combined, washed with brine (50 mL) and
water (2ꢂ50 mL) and dried over Na2SO4. Solvent was then removed
in vacuo; pyridine residues were removed under high vacuum. The
residue was then dissolved in dry DCM (8 mL) and evacuated with
nitrogen: methyl triflate (0.39 g, 2.39 mmol) was then added to the
stirred solution, which was left stirring for 17 h at 20 ꢀC. The DCM
was then removed in vacuo and the residue chromatographed us-
ing EtOAc, then 10:90 PriOH:DCM.
4.4. (3-Phenoxypropyl)trifluoromethane sulfonate 1635
To a solution of 3-phenoxypropan-1-ol (0.21 g, 1.4 mmol) in
DCM (2 mL) at 0 ꢀC under N2 was added 2,3,6-collidine (0.254 g,
2.1 mmol) followed by triflic anhydride (0.258 mL, 0.434 g,
1.54 mmol). The reaction was left at 0 ꢀC for 2 h and then the sol-
vent removed in vacuo. The residue was partitioned between ether
(50 mL) and water (30 mL) and the organic layer separated. The
organic layer was then successively washed with 1 M HCl
(2ꢂ20 mL), satd NaHCO3 (20 mL), water (30 mL), dried over MgSO4
and evaporated to afford crude product 16 (0.119 g, 30%) as a foam,
which was used directly: 1H NMR (250 MHz, CDCl3): 2.22–2.40 (2H,
q, CH2CH2CH2), 4.0–4.18 (2H, t, CH2CH2OAr), 4.68–4.90 (2H, t,
CF3SO2OCH2), 6.82–7.10 (3H, m, ArH), 7.26–7.41 (2H, m, ArH).
4.2.1. Methyl(N-(2,3,4-tri-O-isobutyryl)-
b
-
D
-glucopyranuronato)-
4.5. [b-D-Glucopyranuronato-N-1-(3-phenoxypropyl)]-
N,N-dimethyl-3-(10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5-yli-
dene)-1-propanammonium trifluoromethanesulfonate 14a. Non-
crystalline foam, yield 0.177 g (18%). Found: C, 57.8; H, 6.4; N, 1.4;
[M]þ m/z 678.3660; C40H52NO12SF3 requires C, 58.0; H, 6.3; N, 1.7%;
[C39H52NO9]þ requires m/z, 678.3642; 1H NMR [400 MHz,
(CD3)2SO]: 1.00–1.06 [18H, m, 3ꢂCH(CH3)2], 2.40–2.49 (3H, m,
3ꢂCH(CH3)2), 2.50–2.58 (2H, br s, CHCH2CH2Nþ), 2.73–2.92 (2H,
br s, CH2CH2), 2.98–3.09 [6H, 4 s, Nþ(CH3)2], 3.20–3.31 (4H, br s,
CH2CH2), 3.34 (3H, s, CO2CH3), 3.62–3.71 (2H, m, CHCH2CH2Nþ),
4.50–4.52 (1H, m, 5-H), 5.14–5.30 (1H, m, 3-H), 5.30–5.40 (1H, m,
4-H), 5.50–5.51 (1H, t, 2-H), 5.7–5.84 (2H, t, ¼CHCH2CH2Nþ and m,
1-H) and 7.08–7.30 (8H, m, Ar); 13C NMR [100 MHz, (CD3)2SO]: 14.3,
18.3, 18.6, 18.7, 18.8, 18.9, 22.4, 23.1, 31.6, 33.4, 33.5, 33.6, 33.6, 48.9,
48.9, 53.2, 67.1, 67.8, 72.3, 72.8, 90.5, 126.3, 126.5, 127.6, 128.0, 128.4,
128.5, 128.6, 128.7, 130.5, 137.1, 139.1, 139.3, 140.4, 174.8, 175.1 and
175.4 (ꢂ2); m/z (ES þve ion mode) 678 [M]þ 100%.
pyrrolidinium-1-yl 20
To a solution of 15 (0.74 g, 1.57 mmol) in dry DCE (4 mL) under
nitrogen was added 16 (0.49 g, 1.72 mmol) in DCE (4 mL). The re-
action was heated to 50 ꢀC for 24 h. The DCE was then removed in
vacuo and the residue triturated with diethyl ether. The residue
containing intermediate 17 (0.47 mmol) was without purification
dissolved in methanol (8 mL), cooled to ꢁ10 ꢀC, stirred and treated
with 10% aq Na2CO3 (2 mL, 1.89 mmol). After addition the reaction
was stirred at ꢁ10 ꢀC for 2 h, then Amberlite Hþ resin was added to
give a pH of 6.5. Methanol was removed in vacuo and the residue
partitioned between water (50 mL) and DCM (50 mL). The aqueous
phase was further extracted with DCM (3ꢂ50 mL), then evaporated
to dryness to give an amorphous solid (0.365 g, 61%); by NMR
analysis this material was 4:1 product: eliminated by-product (see
Scheme 6). HPLC using a gradient of MeCN in H2O under neutral
conditions was used to purify the material further: 20 was thus
obtained as a colourless foam (0.066 g, 11%); [
293K
4.2.2. Methyl(N-(2,3,4-tri-O-isobutyryl)-
b
-
D
-glucopyranuronato)-
a]
¼ꢁ11.47
D
N,N-dimethyl-3-(10,11-dihydro-5H-dibenzo[b,f]azepin-5-yl)-1-prop-
anammonium trifluoromethanesulfonate 14b. Non-crystalline foam,
yield 0.158 g (16%). Found: [M]þ m/z 681.376, C38H53O9N2 requires
m/z 681.375; 1H NMR (400 MHz, CDCl3): 1.05–1.08 [18H, m,
3xCH(CH3)2], 2.1–2.2 (2H, m, NCH2CH2CH2), 2.35–2.6 [3H, m,
(0.8 g mLꢁ1 in MeOH); Found: [MH]þ m/z 382.1871; C19H28O7N
requires m/z 382.1866; nmax cmꢁ1 2500–3500 (br s), 2973, 2916,
1611, 1486, 1443, 1412 (m),1092 (vs), 883 (m) and 742; 1H NMR
(600 MHz, CD3OD): 2.08–2.27 (4H, m, CH2CH2CH2CH2), 2.28–2.46
(2H, m, NþCH2CH2CH2O), 3.46–3.53 (2H, m, 3-H þ 4-H), 3.61–3.69

L. Iddon et al. / Tetrahedron 66 (2010) 537–541
541
(1H, dt, CH2 NþCH2), 3.71–3.79 (4H, m, CH2 NþCH2, NþCH2CH2CH2O
and 5-H), 3.79–3.83 (1 H, t, 2-H), 3.92–4.11 (4H, m, NþCH2CH2CH2O,
CH2 NþCH2), 4.67–4.71 (1H, d, J¼8.9 Hz, 1-H), 6.89–6.93 (3H, m, Ar)
and 7.22–7.28 (2H, m, Ar); 13C NMR (150 MHz, CD3OD): 23.6, 24.8,
25.6, 59.4, 62.4, 64.7, 66.2, 72.2, 73.3, 79.1, 79.8, 96.3, 116.0, 122.6,
131.0, 160.4 and 175.5; m/z (ES þve ion mode) 382 [MH]þ 100%. For
the NMR of intermediate 17, see Supplementary data.
78.9, 79.5, 95.8, 121.1, 124.5, 128.1, 131.5, 136.0, 149.5 and 175.3; m/z
(ES þve ion mode) 457 [MH]þ 100%.
Acknowledgements
We are grateful to Astra Zeneca plc and the EPSRC for funding
(CASE award to LI) and to Olaf Beckonert (Imperial College, London)
for 600 MHz NMR spectra.
4.6. [b-D-Glucopyranuronato-(N,N-dimethyl-3-(10,11-
dihydro-5H-dibenzo[a,d]cyclohepten-5-ylidene))-1-
propanammonium 18
Supplementary data
The supplementary data associated with this article can be
found in the on-line version at doi:10.1016/j.tet.2009.10.113.
To a solution of 14a (0.114 g, 0.14 mmol) in MeOH (2 mL) was
added 1 M NaOH (0.42 mL) at ꢁ10 ꢀC with stirring. After 3 h at this
temperature, TLC (5:3:2 EtOAc/IPA/H2O) showed complete re-
action: Amberlite Hþ resin (0.42 mmol) was added to give a pH of 7.
The resin was then filtered and washed with methanol, and the
methanol removed in vacuo: the residue was partitioned between
DCM (30 mL) and water (30 mL), the aqueous phase was separated
and extracted with more DCM (3ꢂ30 mL) and water was removed
in vacuo to yield the crude product. The residue was then purified
using reverse phase C18 silica gel, eluting with 100% H2O then
References and notes
1. Gibson, G. G.; Skett, P. Introduction to Drug Metabolism, 2nd ed.; Blackie/
Chapman and Hall: London, 1994; Chapters 1 and 2.
2. Zamek- Gliszczynski, M. J.; Hoffmaster, K. A.; Nezasa, K.; Tallman, M. N.;
Brouwer, K. L. R. Eur. J. Pharm. Sci. 2006, 27, 447–486.
3. (a) Kaspersen, F. M.; van Boeckel, C. A. A. Xenobiotica 1987, 17, 1451–1471;
(b) Stachulski, A. V.; Jenkins, G. N. Nat. Prod. Rep. 1998, 15, 173–186.
4. Ritter, J. K. Chem. Biol. Interact. 2000, 129, 171–193.
5. Stachulski, A. V.; Harding, J. R.; Lindon, J. C.; Maggs, J. L.; Park, B. K.; Wilson, I. D.
J. Med. Chem. 2006, 49, 6931–6945.
6. Shaffer, C. L.; Gunduz, M.; Scialis, R. J.; Fang, A. F. Drug Metab. Dispos. 2007, 35,
1188–1195.
10–25% MeCN in H2O to afford the product 18 (0.064 g, 16%) as an
293K
amorphous glass; [
a
]
¼ꢁ5.80ꢀ (0.5 mg/mL in MeOH). Found:
D
[MH]þ m/z, 454.2227; C26H32O6N [MH]þ requires m/z, 454.2230;
nmax cmꢁ1 2500–3500 (br s), 2970, 2916,1611, 1485, 1447,1407,1094
(vs), 885 (m) and 743; 1H NMR (600 MHz, CD3OD at 323 K): 2.57–
2.72 (2H, m, NCH2CH2CH), 2.78–3.0 (2H, br m, CH2CH2), 3.15 (3H, s,
NCH3), 3.19 (3H, s, NCH3), 3.40–3.47 (3H, m, 3-H, NCH2, CH2CH2),
3.51–3.62 (2H, br s, NCH2, CH2CH2), 3.69–3.74 (2H, m, 2-H, 4-H),
3.88–3.98 (1H, br s, 5-H), 4.3–4.51 (1H, br s, 1-H), 5.73–5.82 (1H, t,
¼CH CH2) and 7.04–7.31 (8H, m, Ar); 13C NMR [150 MHz, CD3OD at
323 K]: 24.4, 33.2, 34.9, 64.2, 64.8, 71.5, 71.6, 77.8, 78.7, 79.0, 79.1,
94.7, 95.2, 124.9, 125.2, 127.3, 127.7, 128.8, 131.1, 131.4, 138.5, 140.7,
140.8, 141.6, 148.5 and 175.1 m/z (ES þve ion mode) 454 [MH]þ,
100%.
7. Franklin, R. B. Drug Metab. Dispos. 1998, 26, 829.
8. Green, M. D.; Tephly, T. R. Drug Metab. Dispos. 1998, 26, 860–867.
9. Borlak, J.; Gasparic, A.; Locher, M.; Schupke, H.; Hermann, R. Metab. Clin. Exp.
2006, 55, 711–721.
10. Nakazawa, T.; Miyata, K.; Omura, K. Drug Metab. Dispos. 2006, 34, 1880–1886.
11. Yan, Z.; Caldwell, G. W.; Gauthier, D. Drug Metab. Dispos. 2006, 34, 748–755.
12. Dalgaard, L.; Larsen, C. Xenobiotica 1999, 29, 1033–1041.
13. Zeng, S.; Qian, M. R. Acta Pharmacologica Sinica 2006, 27, 623–628.
14. Araya, I.; Tsubuki, T.; Saito, T.; Numata, M.; Akita, H. Chem. Pharm. Bull. 2007, 55,
1039–1043.
15. Upadhyaya, P.; McIntee, E. J.; Hecht, S. S. Chem. Res. Toxicol. 2001, 14, 555–561.
16. Caldwell, W. S.; Greene, J. M.; Byrd, G. D.; Chang, K. M.; Uhrig, M. S.; deBethizy,
J. D.; Crooks, P. A.; Bhatti, B. S.; Riggs, R. M. Chem. Res. Toxicol. 1992, 5, 280–285.
17. Hawes, E. M. Drug Metab. Dispos. 1998, 26, 830–837.
18. Luo, H.; Hawes, E. M.; McKay, G.; Midha, K. K. J. Pharm. Sci. 1992, 81, 1079–1083.
19. Mey, U.; Wachsmuth, H.; Breyer-Pfaff, U. Drug Metab. Dispos.1999, 27,1281–1292.
20. Fischer, D.; Breyer-Pfaff, U. J. Pharm. Pharmacol. 1995, 47, 534–538.
21. de Leon, J. Int. J. Neuropsychopharmacol. 2003, 6, 57–72.
22. Kaku, T.; Ogura, T.; Nishiyama, T.; Ohnuma, T.; Muro, K.; Hiratsuka, A. Biochem.
Pharmacol. 2004, 67, 2093–2102.
4.7.
b-D-Glucopyranuronato-(N,N-dimethyl-3-(10,11-dihydro-
5H-dibenzo[b,f]azepin-5-yl)-1-propanammonium) 19
To a solution of 14b (0.115 g, 0.14 mmol) in MeOH (2 mL) was
added 1 M LiOH (0.42 mL) with stirring at ꢁ10 ꢀC. After 2 h at this
temperature, TLC (100% EtOAc) showed completion. The reaction
was quenched with Amberlite Hþ resin (0.42 mmol) to give a pH of
7. The resin was then filtered and washed with methanol, and the
methanol removed in vacuo: the residue was partitioned between
DCM (30 mL) and water (30 mL). The aqueous phase was separated
and extracted with more DCM (3ꢂ30 mL), then the water was
removed in vacuo to yield crude product, which was purified by
23. Chiu, S. H. L.; Huskey, S. H. W. Drug Metab. Dispos. 1998, 26, 838–847.
24. Dalgaard, L. Acta Chem. Scand. B 1983, 37, 923–928.
25. Isbell, H. S.; Frush, H. J. J. Org. Chem. 1958, 23, 1309–1319.
26. Baker, J. W. J. Chem. Soc. 1928, 1583; J. Chem. Soc. 1929, 1205–1210.
27. ‘Guidance for Industry: Safety Testing of Drug Metabolites’, Food and Drug Au-
thority of the USA, Feb. 2008, http://www.fda.gov/CDER/guidance/6897fnl.pdf.
28. Calligaro, D. O.; Fairhurst, J.; Hotten, T. M.; Moore, N. A.; Tupper, D. E. Bioorg.
Med. Chem. Lett. 1997, 7, 25–30.
29. Iddon, L.; Bragg, R. A.; Harding, J. R.; Pidathala, C.; Bacsa, J.; Kirby, A. J.;
Stachulski, A. V. Tetrahedron 2009, 65, 6396–6402.
30. Olofson, R. A.; Martz, J. T.; Senet, J. P.; Piteau, M.; Malfroot, T. J. Org. Chem. 1984,
49, 2081–2082.
31. Brown, R. T.; Carter, N. E.; Mayalarp, S. P.; Scheinmann, F. Tetrahedron 2000, 56,
7591–7594. The -tetraisobutyrate is obtained as a crystalline solid in 60% yield
a-Chloroethyl chloroformate is the recommended reagent.
preparative HPLC using MeCN/H2O affording 19 (0.064 g, 57%) as an
293K
amorphous glass; [
a]
¼ꢁ18.25 (2 g mLꢁ1 in MeOH); Found: C,
b
D
61.0; H, 7.0; N, 5.4; [MH]þ m/z, 457.2343. C25H32O6N2.2H2O requires
C, 61.0; H, 7.4; N, 5.7%; C25H33O6N2 requires [MH]þ m/z 457.2339;
nmax cmꢁ1 2500–3500 (br s), 2920, 2850,1610, 1486, 1447, 1403 (m),
1094 (vs), 883 (w) and 755; 1H NMR (400 MHz, CD3OD): 2.0–2.16
(2H, m, NþCH2CH2CH2), 3.05 (3H, s, NþCH3), 3.09 (3H, s, NþCH3),
3.18 (4H, s, CH2CH2), 3.37–3.54 (3H, m, 3-H and NþCH2CH2CH2),
3.69–3.71 (2H, m, 2-Hþ4-H), 3.73–3.84 (2H, m, 5-H and
NþCH2CH2CH2), 3.88–3.98 (1H, m, NþCH2CH2CH2), 4.51–4.55 (1H, d,
J¼8.9 Hz, 1-H), 6.92–6.97 (2H, m, Ar) and 7.11–7.20 (6H, m, Ar); 13C
NMR (100 MHz, CD3OD): 22.4, 33.6, 48.6, 49,6, 49.7, 63.8, 71.9, 73.1,
from glucuronolactone. Subsequent anomeric deacylation is best achieved with
hydrazine/acetic acid.
32. Iddon, L. PhD thesis, University of Liverpool, 2009.
33. A most useful recent review of TEMPO oxidations is Vogler, T.; Studer, A.
Synthesis 2008, 1979–1993.
34. (a) Perrie, J. A.; Harding, J. R.; Holt, D. W.; Johnston, A.; Meath, P.; Stachulski,
A. V. Org. Lett. 2005, 7, 2591–2594; (b) Bowkett, E. R.; Harding, J. R.; Maggs, J.
L.; Park, B. K.; Perrie, J. A.; Stachulski, A. V. Tetrahedron 2007, 63, 7596–7605.
35. Shi, Z.; He, C. J. Am. Chem. Soc. 2004, 126, 13596–13597.
36. (a) McCleary, C. W.; Rees, D. A.; Samuel, J. W. B.; Steele, I. W. Carbohydr. Res.
1967, 5, 492–495; (b) Stanford, D.; Stachulski, A. V. Tetrahedron Lett. 2007, 48,
2361–2364.
37. Lin, T.-H.; Kovacs, P.; Glaudemans, C. P. J. Carbohydr. Res. 1989, 188, 228–238.