Effective synthesis of 1β-acyl glucuronides by selective acylation

Article abstract of DOI:10.1021/ol0507165

(Chemical Equation Presented) Acyl glucuronides are vital metabolites for many carboxylic acid containing drugs. We report an efficient new method for the chemical synthesis of these molecules by selective 1β-acylation of allyl glucuronate with carboxylic

Full text of DOI:10.1021/ol0507165

ORGANIC
LETTERS
2005
Vol. 7, No. 13
2591-2594
Effective Synthesis of 1â-Acyl
Glucuronides by Selective Acylation
|
Jennifer A. Perrie,† John R. Harding,‡ David W. Holt,§ Atholl Johnston,
Paul Meath,† and Andrew V. Stachulski*,†
The Robert Robinson Laboratories, Department of Chemistry, UniVersity of LiVerpool,
LiVerpool L69 7ZD, U.K., AstraZeneca UK Ltd., Drug Metabolism and
Pharmacokinetics Department, Mereside, Alderley Park,
Macclesfield, Cheshire SK10 4TG, U.K., Analytical Unit, St. George’s Hospital
Medical School, London, U.K., and William HarVey Research Institute, Barts and the
London, Queen Mary School of Medicine and Dentistry, Charterhouse Square,
London EC1M 6BQ, U.K.
stachuls@liV.ac.uk
Received April 4, 2005
ABSTRACT
Acyl glucuronides are vital metabolites for many carboxylic acid containing drugs. We report an efficient new method for the chemical synthesis
of these molecules by selective 1 -acylation of allyl glucuronate with carboxylic acids catalyzed by HATU and then mild deprotection through
â
treatment with Pd(PPh3)4 and morpholine. The method is effective for a range of aryl and alkyl carboxylic acids, including important drugs.
Glucuronidation is a vital phase 2 metabolic process1,2
whereby a wide range of drugs and xenobiotics may be
Scheme 1. Reactions and Rearrangements of Acyl
Glucuronidesa
rendered water-soluble, detoxified, and excreted. While
O-glucuronides of alcohols and phenols are well-known and
generally behave as stable organic molecules, acyl (ester)
glucuronides of carboxylic acids have received less attention.
Typically, acyl glucuronides are much less stable than those
of alcohols or phenols, being subject to facile reaction with
nucleophiles and rearrangement at acidic or basic pH
(Scheme 1).3
Many important drug classes, notably the widely used
nonsteroidal antiinflammatory drugs (NSAIDs), contain
carboxylic acids and are metabolized as their acyl glucuron-
ides. Whether acyl glucuronides cause adverse reactionss
a Pathway 1: direct acylation (“glycation”). Pathway 2: Amadori
rearrangement. The nucleophilic residues may be those present in
body proteins.
† University of Liverpool.
‡ AstraZeneca UK Ltd.
§ St. George’s Hospital Medical School.
| Queen Mary School of Medicine and Dentistry.
(1) Kaspersen, F. M.; van Boeckel, C. A. A. Xenobiotica 1987, 17, 1451-
1471.
immune response or direct toxicitysis still a matter of lively
debate.4 For instance, the current consensus is that the acyl
glucuronides of ibuprofen 15 and naproxen 2 are benign,
whereas that of diclofenac 36 is suspect. It has been shown
(2) Stachulski, A. V.; Jenkins, G. J. Nat. Prod. Reports 1998, 15, 173-
186.
(3) Vanderhoeven, S. J.; Lindon, J. C.; Troke, J.; Tranter, G. E.; Wilson,
I. D.; Nicholson, J. K. Xenobiotica 2004, 34, 73-85.
10.1021/ol0507165 CCC: $30.25
© 2005 American Chemical Society
Published on Web 05/27/2005

that some acyl glucuronides, e.g., that of diflunisal 4, can
cause binding of the parent drug to plasma protein (Figure
1).7
the free acyl glucuronide. We have employed this method
to synthesize up to 200 mg of the acyl glucuronide of
diclofenac 3, but our experience12 emphasized some short-
comings of the method. In particular, â/R mixtures from 5:1
to 2:1 were obtained in all cases11 (we observed 4:112).
Satisfactory purification required both column chromatog-
raphy and preparative HPLC: our best yield was 20%.
We considered that instead of the Mitsunobu reaction,
proceeding ultimately by an SN2 reaction at the anomeric
center,13 selective acylation might be a superior method,
exploiting the kinetic anomeric effect14 to favor 1â-acylation,
Scheme 2. There was some precedent in a report15 of
Figure 1. NSAIDs metabolizing as their acyl glucuronides.
Scheme 2. Acyl Glucuronide Synthesis via Mitsunobu (M) or
Acylation (A) Routes
As shown (Scheme 1), both direct acyl transfer (hydrolysis
or reaction with other nucleophiles) and acyl migration
followed by imine formation, then tautomerism (Amadori
rearrangement) can occur. Recently, there has been an
attempt to quantify the acyl donor ability of acyl glucuronides
compared to other bioconjugates such as thioesters.8
It is therefore important to synthesize acyl glucuronides
pure and in quantity as single 1â-anomers, as they occur in
vivo, for thorough evaluation. Previously, two main ap-
proaches have been employed: either a fully protected
derivative such as 5 (Figure 2) has been prepared, followed
selective 1â-acylation of glucose derivatives using activated
esters of the carboxy component. We now report the
successful realization of this concept in an effective synthesis
of several 1â-acyl glucuronides.
We simplified the synthesis of 7 by using a resin-bound
fluoride base16 instead of DBU.11 This greatly eased the
workup (we had found it very difficult to remove DBU traces
completely) and afforded crystalline 7 as an R/â mixture in
70% yield, containing variable amounts (from 5 to 20%) of
a coeluting impurity by comparison with previous NMR
data.11 The impurity does not affect the efficiency of the
following steps or the purity of the final products; recrys-
tallization from MeOH affords essentially pure 7, whose
preparation we are continuing to optimize.
Figure 2. Glucuronic acid intermediates. All ) CH2CHdCH2.
We then studied various combinations of carbodiimides,
active ester-forming reagents, and base catalysts to activate
and couple the carboxylic acid component, using 4-bro-
mobenzoic acid 8 as the model substrate (Figure 3).
1-Hydroxybenzotriazole (HOBt) in conjunction with DIC
(superior to DCC) was effective: later we found that the
uronium reagent HATU17 was the reagent of choice.
In early experiments, the active ester was preformed and
then reacted with 7 and a base catalyst in acetonitrile. We
later found that combining all reagents from the start gave
by conjugation to the carboxylic acid (e.g., via the imidate9
method), or unprotected glucuronic acid 6 has been used.
Intermediates such as 5 take many steps to prepare, and the
use of unprotected glucuronic acid has been restricted to
some special cases, e.g., retinoic acids.10
A promising alternative is to employ a monoester of 6,
namely allyl glucuronate 7. Owing to the greater reactivity
of the anomeric hydroxy group, 7 will react with carboxylic
acids in a Mitsunobu reaction11 to deliver fair yields of the
desired conjugates. Deprotection with Pd(PPh3)4 then releases
(11) Juteau, H.; Gareau, Y.; Labelle, M. Tetrahedron Lett. 1997, 38,
1481-1484. The original report of the use of the Mitsunobu procedure for
glycosyl ester synthesis was by: Smith, A. B.; Hale, K. J.; Rivero, R. A.
Tetrahedron Lett. 1986, 27, 5813-5816. Smith, A. B.; Hale, K. J.; Vaccaro,
H. A.; Rivero, R. A. J. Am. Chem. Soc. 1991, 113, 2092-2112.
(12) Kenny, J. R.; Maggs, J. L.; Meng, X.; Sinnott, D.; Clarke, S. E.;
Park, B. K.; Stachulski, A. V. J. Med. Chem. 2004, 47, 2816-2825.
(13) Recent investigations have shown that, for some hindered secondary
alcohols, the Mitsunobu reaction may proceed via an acyloxyphosphonium
intermediate: Ahn, C.; Correia, R.; De Shong, P. J. Org. Chem. 2002, 67,
1751-1753. Scheme 2 posits the “normal” alkoxyphosphonium intermedi-
ate; we consider the anomeric position is not particularly hindered. See
also: Hughes, D. L.; Reamer, R. A. J. Org. Chem. 1996, 61, 2967-2971.
(14) Schmidt, R. R. Angew. Chem., Int. Ed. Engl. 1986, 25, 212-235.
(15) Plusquellec, D.; Roulleau, F.; Bertho, F.; Lefeuvre, M. Tetrahedron
1986, 42, 2457-2467.
(4) Bailey, M. J.; Dickinson, R. G. Chemico-Biological Interactions 2003,
145, 117-137. Shipkova, M.; Armstrong, V. W.; Oellerich, M.; Wieland,
T. H. Ther. Drug Monit. 2003, 25, 1-16.
(5) Tan, S. C.; Patel, B. K.; Jackson, S. H. D.; Swift, C. G.; Hutt, A. J.
Xenobiotica 2002, 32, 683-697.
(6) Grillo, M. P.; Knutson, C. G.; Sanders, P. E.; Walden, D. J.; Hua, E.
M.; Ware, J. A. Drug Metab. Dispos. 2003, 31, 1327-1336.
(7) Wang, M.; Dickinson, R. G. Drug Metab. Dispos. 1998, 26, 98-
104.
(8) Li, C.; Benet, L. Z.; Grillo, M. P. Chem. Res. Toxicol. 2002, 15,
1309-1317.
(9) Schmidt, R. R.; Grundler, G. Synthesis 1981, 885-887.
(10) Barua, A. B.; Huselton, C. A.; Olson, J. A. Synth. Commun. 1996,
26, 1355-1361.
2592
Org. Lett., Vol. 7, No. 13, 2005

typically δ 5.8 (1H, d, J ) 7.8 Hz) for the â-anomer; any
traces of R-anomer show δ 6.3 (d, J ) 3.5 Hz).3,18 As noted
above, essentially pure â-product was obtained by a simple
silica column; preparative HPLC was unnecessary.
Mycophenolic acid 1319 affords a most interesting ex-
ample: this Penicillium metabolite has valuable antibiotic
and immunosuppressant activity. We believe this is the first
chemical synthesis of its acyl glucuronide, which had
previously been obtained by preparative HPLC separation
of an enzymatically synthesized mixture of its aryl and acyl
glucuronides.20,21 It was not necessary to protect the phenolic
group in 13, though the yield was slightly lower here than
in most other examples.
Figure 3. Carboxylic acids used in this study (with 1).
an equally good yield. The pKa of the base used proved
critical: we found the optimum value was around 7.5 to 8.
DABCO and N-methylmorpholine (NMM) were particularly
effective, the latter being marginally superior. Weaker bases,
e.g., pyridine, gave very slow reaction, while stronger ones,
e.g., DMAP or triethylamine, led to overreaction.
The base was removed by neutralization with Amberlyst
A-15 (H+); the product was then isolated by filtration,
evaporation, and silica gel chromatography. For the reaction
of 8, we obtained the desired product 9 in 52% yield using
HATU (1 molar equiv) and NMM (2 molar equiv), with 1
equiv each of 7 and 8, and then employed the method for a
series of carboxylic acids 1 and 10-15. Table 1 summarizes
In the cases of ibuprofen, (S)-1, and (R)-O-methylmandelic
acid 12 the acyl glucuronides were formed as separable
1
epimeric mixtures22 in a 4:1 ratio. The most diagnostic H
NMR signals for the conjugates are those of the CH3CH(Ar)-
CO-unit for (S)-1 and of the CH(OMe) unit for 12. It is well-
known that (S)-1 racemizes at physiological pH. Zomepirac
15 is another significant clinical example: this antiinflam-
matory agent was withdrawn following serious allergic
reactions23 which, it was postulated, could be due to its acyl
glucuronide.
Finally, we confirmed the deprotection procedure using
Pd(PPh3)4 in conjunction with pyrrolidine11 or morpholine24
in THF: morpholine gave cleaner and higher yielding
products. For example, using the Pd(PPh3)4-morpholine
combination, the acyl glucuronide allyl esters of 11 and 13
were deprotected to afford the free acyl glucuronides 16a
and 16b in 80% yield following brief silica chromatography
(Figure 4).25
Table 1. Yields and R/â Ratios of Acyl Glucuronides by the
HATU-NMM Procedurea,b
Mitsunobu
method11
carboxylic
acid (equiv) HATU, NMM
equiv of
yield
(%)
yieldc
(%)
â/R
ratio
In summary, we have demonstrated a convenient synthesis
of 1â-O-acyl glucuronides by selective acylation that admits
a wide variety of structural types. Conditions are mild and
â/R ratio
8 (1)
8 (2)
1 (2)
10 (1)
11 (1)
12 (1)
13 (1)
14 (1)
15 (1)
1.2
2.4
2.4
1.3
1.3
1.3
2.4
1.2
1.3
52
59
65
52
66
65
44
43
52
19:1
48
5:1
â only
â only
â only
â only
â only
19:1
(16) Available from Sigma-Aldrich Co. Glucuronic acid in DMF was
stirred with allyl bromide (1.1 equiv) and the resin-bound base at 40 °C for
16 h followed by filtration, evaporation and chromatography.
(17) Carpino, L. A. J. Am. Chem. Soc. 1993, 115, 4397-4398.
(18) Ruhl, R.; Thiel, R.; Lacker, T. S. J. Chromatogr. 2001, B757, 101-
109.
(19) Florey, H. W.; Gilliver, K.; Jennings, M. A.; Sanders, A. G. Lancet
1946, (i), 46. Shipkova, M.; Voland, A.; Grone, H. J. Ther. Drug Monit.
2003, 25, 206. Holt, D. W.; Johnston, A. Ther. Drug Monit. 2004, 26, 244-
247.
30
31
38
5:1
3:1
2:1
â only
â only
a All reactions were carried out at 20 °×b0C in acetonitrile. b Reaction
time of 2 h usually sufficient (monitored by TLC). c In the Mitsunobu
procedure, 2 equiv of carboxylic acid was always used.
(20) Kittleman, M.; Rheinegger, U.; Espigat, A.; Oberer, L.; Aichholz,
R.; Francotte, E.; Ghisalba, O. AdV. Synth. Catal. 2003, 345, 825-829.
(21) Covalent binding of the deprotected product to monophosphate
dehydrogenase has been demonstrated, with superior response to enzymati-
cally prepared material. Further details will be published elsewhere.
(22) Ikegawa, S.; Murao, N.; Ooashi, J.; Goto, J. Biomed. Chromatogr.
1998, 12, 317-321.
the yields obtained by the HATU procedure: using the
HOBt-DIC method, yields were generally around 20% less.
In many cases a very satisfactory yield of conjugate was
obtained using just 1 equiv of HATU, but as noted in Table
1, the use of extra base or 2 equiv of HATU was beneficial
in some cases, e.g., 8 (59%). It is also noteworthy that the
presence of an ortho substituent, as in 2-bromobenzoic acid
10, does not appreciably lower the yield, in contrast to the
Mitsunobu procedure.
(23) Levy, D. B.; Vasilomanolakis, E. C. Drug Intell. Clin. Pharm. 1984,
18, 983-984.
(24) Kunz, H.; Dombo, B. Angew. Chem., Int. Ed. Engl. 1988, 27, 711-
713.
(25) It was difficult to exclude completely small traces of Pd(PPh3)4 from
the final product, even after chromatography. We have found very recently
that a polystyrene-bound form of the Pd reagent (available from Argonaut
1
Technologies Ltd.) affords reagent-free product at the H NMR detection
limit. The commercial resin was stirred in THF for 16 h, then filtered off
and added to a solution of the allyl ester in THF containing morpholine (1
equiv) with stirring at 20 °C until complete deprotection was observed by
TLC. Filtration, evaporation and chromatography (10-50% EtOH in CH2-
Cl2, silica) then afforded the free acyl glucuronide, e.g., 16a.
In all cases, the â/R ratio was excellent, 95:5 or better,
1
the key H NMR signal being that of the anomeric proton,
Org. Lett., Vol. 7, No. 13, 2005
2593

financial support to J.A.P. and to P.M. and to Roche
Pharmaceuticals for a generous gift of mycophenolic acid.
Finally, we are grateful to Xiaoli Meng for experimental
assistance.
Supporting Information Available: Full characterization
of new compounds described and copies of 1H and 13C NMR
spectra of all compounds made. This material is available
free of charge via the Internet at http://pubs.acs.org.
Figure 4. Acylation product and free acyl glucuronides after
deprotection.
OL0507165
(26) General Acylation Procedure. 4-Bromobenzoic acid 8 (0.101 g,
0.5 mmol), allyl glucuronate 7 (0.117 g, 0.5 mmol), and HATU (0.190 g,
0.5 mmol) were stirred in dry acetonitrile (5 mL) with N-methylmorpholine
(0.110 mL, 0.101 g, 2 equiv) under nitrogen at 20 °C. The reaction was
monitored by TLC (10% EtOH-CH2Cl2, Merck Kieselgel analytical plates),
and after 2 h it was quenched by addition of Amberlyst A-15 (H+ form, 2
equiv). After evaporation at <30 °C (Buchi rotavapor), the residue was
chromatographed on Merck 938 S silica, eluting with 7% EtOH-CH2Cl2.
Appropriate fractions were pooled and evaporated, eventually under high
vacuum, to afford the product 911 as a foam (0.108 g, 52%).
a satisfactory rate is obtained at 20 °C.26 We believe this
method should be extremely useful for the preparation of
acyl glucuronide metabolites of many marketed and prospec-
tive carboxylic acid containing drugs.
Acknowledgment. We are grateful to the EPSRC, As-
traZeneca PLC, and Analytical Services International for
2594
Org. Lett., Vol. 7, No. 13, 2005