TEMPO-Mediated Regiospecific Oxidation of Glucosides to Glucuronides

Article abstract of DOI:10.1055/s-2003-42048

A TEMPO/hypochlorite/bromide oxidant has been used for the conversion of aryl and steroidal glucosides to the corresponding glucuronide conjugates in good (48-74%) yield. An isoflavone glucoside failed to undergo this transformation.

Full text of DOI:10.1055/s-2003-42048

LETTER
1981
TEMPO-Mediated Regiospecific Oxidation of Glucosides to Glucuronides
T
EMPO-Mediate
a
d Oxidation
k
of Glucosid
e
es to
G
luc
s
uronides h N. Desai, Len F. Blackwell*
Institute of Fundamental Sciences, Massey University, Palmerston North, New Zealand
Fax +64(6)3505682; E-mail: L.F.Blackwell@massey.ac.nz
Received 21 August 2003
OH
OH
O
OH
O
OH
Abstract: A TEMPO/hypochlorite/bromide oxidant has been used
for the conversion of aryl and steroidal glucosides to the corre-
sponding glucuronide conjugates in good (48–74%) yield. An
isoflavone glucoside failed to undergo this transformation.
R
R
R
+
Key words: oxidation, glucuronide, glucoside, regiospecific,
TEMPO
R = H
R = Me
100%
93%
0%
7%
Equation 1 Selective primary alcohol oxidation using TEMPO
Glucuronide conjugates are important metabolites found
in urine and are used as biomarkers for health and disease.
In our ongoing efforts on the development of the Ovarian
Monitor System, glucuronide conjugate concentrations
are measured to infer the fertile state of the woman during
her menstrual cycle.1 In order to develop such systems ap-
plicable to a range of health issues, we require procedures
for the synthesis of steroidal and phytoestrogenic glucu-
ronides.
The usefulness of dinitrogen tetroxide in preferential oxi-
dation of primary alcohol groups in carbohydrates to
uronic acids was first realised by Maurer and Drefahl,6
and Yackel and Kenyon7 in 1942. Oxygen has also been
used as a selective oxidant, in the presence of a platinum
catalyst; however, the yields are often low.8
In more recent times, nitrosyl radicals have emerged as
useful selective oxidants. Davis9 was the first to investi-
gate the use of TEMPO in the oxidation of glucosides to
glucuronides. Treatment of methyl or octyl glucosides
with a TEMPO/hypochlorite/bromide oxidant afforded
the desired glucuronides in very good yields. A similar
oxidation has been performed on glucosyl azides to afford
glucuronyl azides.10
However, methods for their syntheses are limited and are
not entirely reliable. The Koenigs–Knorr reaction2 is still
the most popular coupling protocol for glycoside and glu-
curonide formation. Reaction between alcohols or phe-
nols and 2-acetoxyglycosyl donors occurs with very high
b-glycoside selectivity through neighbouring group par-
ticipation.3 However, the presence of the methoxycarbon-
yl group in glucuronyl donors reduces their reactivity
relative to glycosyl donors and reaction is not always ob-
served.4 In light of this observation, we wished to investi-
gate the possibility of using glucosides as glucuronide
equivalents, exchangeable through selective primary al-
cohol oxidation.
In this paper, we wish to report our findings regarding the
application of the TEMPO/hypochlorite/bromide oxidant
system for the smooth conversion of aryl and steroidal
glucosides to their respective glucuronides. Scheme 1
shows our retrosynthetic plan.
CO2H
O
CH2OH
O
CH2OAc
O
Though many oxidants have been developed for the
smooth and mild oxidation of primary and secondary al-
cohol groups (such as PCC, Swern reagents, Dess–Martin
periodinane and TPAP), regioselective oxidation of one
alcohol group in the presence of others is still generally
not feasible. In particular, selective oxidation of primary
hydroxyl groups in the presence of secondary ones is a
difficult conversion in organic synthesis.
HO
HO
HO
HO
AcO
AcO
OR
OR
OR
HO
HO
AcO
X = OAc, Br, SOPh etc.
R = aryl, steroidal
CH2OAc
O
AcO
AcO
+
ROH
X
Recently, Semmelheck5 has demonstrated the selective
oxidation of primary alcohols to aldehydes in the presence
of a secondary alcohol, using TEMPO (2,2,6,6-tetrameth-
ylpiperidin-1-oxyl radical) (Equation 1).
AcO
Scheme 1 Retrosynthetic analysis of glucuronide synthesis
In order to establish the procedure, we wished to be able
to incorporate both steroidal and aryl aglycones into glu-
curonide conjugates. In the first instance, p-cresol was re-
acted with b-D-glucose penta-O-acetate 1 in the presence
of boron trifluoride diethyl etherate. Glucoside 5a11 was
isolated in moderate (53%) yield by this Fischer–Helfer-
ich procedure. Attempts to apply the same procedure to
SYNLETT 2003, No. 13, pp 1981–1984
1
6
.1
0
.2
0
0
3
Advanced online publication: 08.10.2003
DOI: 10.1055/s-2003-42048; Art ID: D21003ST
© Georg Thieme Verlag Stuttgart · New York

1982
R. N. Desai, L. F. Blackwell
LETTER
the sterically hindered and less nucleophilic 2,6-dimeth- the sodium alkoxide (generated in situ) in DMF. How-
ylphenol resulted in failure. The success reported by Kahn ever, the isolated yield was inferior to the above PTC
and co-workers12 in the glycosylation of unreactive sub- procedure. Table 1 summarises these results.
strates with a sulfoxide glycosyl donor, prompted us to in-
Table 1 Syntheses of Protected Glucosides
vestigate such a system for coupling with this substrate.
The required donor was prepared by the coupling of
CH2OAc
O
X
CH2OAc
O
AcO
AcO
AcO
thiophenol with 1-bromo-1-deoxy-2,3,4,6-tetra-O-acetyl-
a-D-glucopyranose (2), followed by mono-oxidation to
the sulfoxide using meta-chloro-peroxobenzoic acid
(Scheme 2).
+
ROH
AcO
OR
AcO
1,2 or 4
AcO
5
X
R
Isolated yield
(%)
O
OAc
OAc
O
Br
AcO
AcO
AcO
AcO
i)
b-OAc
53
OAc
Me
OAc
OAc
5a
1
2
b-SOPh
34
88
Me
ii)
O
S
O
O
SPh
OAc
Me
AcO
AcO
Ph
AcO
AcO
iii)
5b
5c
OAc
b-OAc
O
OAc
OAc
4
3
H
Scheme 2 Glucosyl donor synthesis. Reagents and conditions: i)
HBr, AcOH, dark, r.t., 16 h, 80%; ii) Bu4NHSO4 , Na2CO3, PhSH,
EtOAc, r.t., 30%, 81%; iii) MCPBA, CH2Cl2, –70 °C, 1 h, 81%
H
H
In the event, coupling between sulfoxide donor 4 and 2,6-
dimethylphenol using triflic anhydride and triethyl phos-
phite as an acid scavenger promoted the formation of glu-
coside 5b in moderate (34%) yield. The mass balance was
accounted for by the formation of thioglycoside 3, formed
by deoxygenation with triethyl phosphite.
b-Br
65
H
H
H
O
5d
Estrone was sufficiently reactive to couple efficiently
with the penta-O-acetate 1 and glucoside 5c was iso-
lated in excellent (88%) yield. Attempts to perform the
Fischer–Helferich reaction on testosterone resulted in ex-
tensive decomposition of both the steroid and the sugar.
However, Koenigs–Knorr coupling with bromo sugar 2,
using cadmium carbonate as a promoter successfully
formed the glucoside 5d in good (65%) yield.
b-Br
–
O
O
OMe
5e
With the peracetylated glucosides 5a–d in hand, we
turned our attention to their hydrolyses to the free glu-
copyranosides. Deprotection of the sugar moiety was car-
ried out using a standard Zemplen deacetylation protocol
(NaOMe–MeOH) to afford the free glucosides. However,
it was found that the deprotection could also be carried out
using methanolic Na2CO3 at ambient temperature. In the
case of formononetin tetra-O-acetyl-b-D-glucopyranoside
5e, it was safer to use this procedure so as to avoid the for-
mation of the 4,5-didehydro glucoside. The milder car-
bonate salt also reduces the possibility of destructive base-
mediated chromene ring opening. Table 2 shows the
glucopyranosides 6a–e isolated in moderate to good
(57–68%) yield.14
Several attempts were also made to synthesise formonon-
etin glucoside 5e using either the Fischer–Helferich or
Koenigs–Knorr protocols. In all cases, none of the desired
product was formed. The insolubility of the substrate in
the reaction media was probably the cause. However, the
phase-transfer-catalysed (PTC) glycosylation procedure
of Wahala13 worked well and the deprotected glucoside
was isolated in moderate (32%, 2 steps, average 57% per
step, see Table 2) yield. Unreacted formononetin was
present in all runs, but attempts to drive the reaction to
completion by extending the reaction time resulted in sig-
nificant decomposition of the sugar (via hydrolysis and
glucal formation) and base-induced ring opening of the
chromene ring in formononetin. It was also possible to
form the desired product by reacting the bromosugar with
Synlett 2003, No. 13, 1981–1984 © Thieme Stuttgart · New York

LETTER
TEMPO-Mediated Oxidation of Glucosides to Glucuronides
1983
While there have been a few reports on the use of TEMPO
in the selective oxidation of pyranosides, a literature
survey revealed only one method involving the use of
TEMPO-mediated selective oxidation on various phenyl
b-D-glucosides.15 The simple p-cresol glucuronide was
conveniently prepared using tert-butyl hypochlorite–
TEMPO in water in good yield, as reported by Melvin et
al. The advantage of this method was the absence of inor-
ganic salts and hence resulted in easy isolation and puri-
fication of the glucuronide. However, t-BuOCl is
extremely unstable at ambient temperature and has only a
limited lifetime at 0 °C. It was decided to investigate an al-
ternative TEMPO/NaOCl method as reported by Nooy16
with some modifications to account for the various gluco-
sides at hand. Table 3 summarises the results.
Table 2 Formation of Deacetylated Glucosides
CH2OAc
O
CH2OH
Na2CO3
AcO
AcO
HO
HO
O
OR
OR
MeOH, rt
AcO
HO
5
6
R
Yield (%)
57
Me
6a
68
67
Me
Me
6b
Table 3 Oxidation of Glucosides to Glucuronides with TEMPO
CH2OH
O
CO2M
O
O
TEMPO, NaOCl, NaBr
H2O, pH 10.5
HO
HO
HO
HO
OR
OR
H
HO
HO
H
H
6
7
R
M
Yield (%)
74
6c
K
65
Me
H
7a
H
H
Na
74
62
Me
O
6d
32 (2 steps)
Me
O
O
7b
Na
Na
Na
O
OMe
H
6e
H
H
7c
Purification of the crude p-cresol, 2,6-dimethylphenol, es-
trone and testosterone glucosides was conveniently
achieved by neutralising the reaction mixture with 1 M
HCl or by bringing the pH of the reaction mixture to pH
ca 6. The compounds were then purified by reverse phase
chromatography on a Waters® C18 Sep-Pak column, elut-
ing with water to remove water-soluble inorganic salts
followed by 50% aq MeOH to recover the desired gluco-
side in almost pure form. The purification of the for-
mononetin glucoside required special attention due to the
acid and base sensitive nature of the chromene ring. In this
case, the crude reaction mixture was carefully neutralised
with Amberlite 120, resulting in a pH ca 7. The glycosidic
linkage is stable at pH 7 for long periods. The formonon-
etin glucoside was purified initially by eluting with a 50–
70% aq MeOH gradient using a non-ionic, neutral, Am-
berlite XAD-2 stationary phase. MeOH was removed and
the product further purified by reverse phase chromato-
graphy on a Waters® C18 Sep-Pak column eluting with
50% aq MeOH and MeOH.
48
H
H
H
O
7d
0
O
O
OMe
7e
Initially, methyl b-D-glucoside was oxidised using
TEMPO–NaBr–NaOCl in water at 0 °C and pH 10.5, con-
trolled by the addition of 0.5 M NaOH using a pH-stat.
The reaction was complete within 45 minutes and the
Synlett 2003, No. 13, 1981–1984 © Thieme Stuttgart · New York

1984
R. N. Desai, L. F. Blackwell
LETTER
crystals; Rf 0.70 (2:1 EtOAc–hexane), mp 110–111 °C.
glucuronide was isolated as the sodium salt by filtration
through a Waters® Sep-Pak C18 column using water,
then 67% aqueous MeOH. Glucuronides 7a–d were syn-
thesised in a similar manner and were isolated as their
sodium or potassium salts in good (48–74%) yield,17
except for formononetin glucuronide 7e. All attempts to
prepare this material resulted in failure. Both the highly
alkaline conditions required, which are incompatible with
the base sensitive chromene ring, and the low solubility
of formononetin in the reaction medium, are probable
factors.
1H NMR (400 MHz, CDCl3): d = 7.11 (d, J = 8.6 Hz, 2 H),
6.90 (d, J = 8.6 Hz, 2 H), 5.31–5.15 (envelope, 3 H), 5.03 (d,
J = 7.3 Hz, 1 H), 4.27 (dd, J = 11.9, 5.0 Hz, 1 H), 4.18 (dd,
J = 11.9, 2.3 Hz, 1 H), 3.90–3.79 (m, 1 H), 2.31 (s, 3 H), 2.09
(s, 3 H), 2.07 (s, 3 H), 2.06 (s, 3 H), 2.04 (s, 3 H). 13C NMR
(100 MHz, CDCl3): d = 170.4, 170.1, 169.2, 169.1, 154.6,
132.7, 129.8, 116.9, 99.4, 72.4, 71.9, 71.1, 68.2, 61.9, 20.6,
20.5.
(12) Kahn, D.; Walker, S.; Cheng, Y.; Vanengen, D. J. Am.
Chem. Soc. 1989, 111, 6881.
(13) Lewis, P.; Katalia, S.; Wahala, K. J. Chem. Soc., Perkin
Trans. 1 1998, 2481.
(14) Preparation of 6a: Compound 5a (0.25 g, 0.57 mmol) was
dissolved in MeOH (10 mL) and aq Na2CO3 (2 M, 3.5 mL)
was added. The reaction mixture was stirred at ambient
temperature for 5 h with monitoring by TLC. The mixture
was neutralised with HCl (1 M) and the solvent removed
under reduced pressure. The compound was then purified by
reverse phase chromatography on a Waters® C18 Sep-Pak
column by eluting with water and then with 50% aq MeOH.
Appropriate fractions were pooled and the solvent removed
under reduced pressure to afford 6a (88 mg, 57%) as a white
solid. 1H NMR (400 MHz, D2O): d = 7.44 (d, J = 8.6 Hz, 2
H), 7.35 (d, J = 8.6 Hz, 2 H), 4.25 (d, J = 9.8 Hz, 1 H), 4.08
(d, J = 3.9 Hz, 1 H), 3.86–3.80 (envelope, 5 H), 2.63 (s, 3 H),
13C NMR (100 MHz, D2O): d = 155.8, 131.7, 129.7, 116.6,
101.4, 76.9, 73.8, 70.3, 61.4, 19.7. HRMS–FAB+: m/z [M]
calcd for C13H18O6: 270.1103; found, 270.1090.
In conclusion, we have been able to show that TEMPO
can be used as a selective oxidant for the conversion of
glucosides to glucuronides in good yield. This procedure
should compete well with the more traditional Koenigs–
Knorr protocol.
Acknowledgement
We would like to thank Dr Jeremy J. Fullbrook for the preparation
of this manuscript.
References
(1) Blackwell, L. F.; Brown, J. B.; Vigil, P.; Gross, B.; Sufi, S.;
d’Arcangues, C. Steroids 2003, 68, 465.
(15) Herebert, R. B.; Melvin, F.; McNeill, A.; Henderson, P. J. F.
Tetrahedron Lett. 1994, 35, 4763.
(16) Nooy, A. E. J.; Besemer, A. C.; van Bekkum, H. Carbohydr.
Res. 1995, 269, 89.
(2) (a) Hadd, H. E. Steroids 1994, 59, 608. (b) Conrow, R. B.;
Bernstein, S. J. Org. Chem. 1971, 36, 863. (c)Bernstein, S.;
Solomon, S., Eds.; Chemical and Biological Aspects of
Steroid Conjugates; Springer-Verlag: New York Inc., 1970,
3–5.
(3) For the mechanism, see: Garegg, P. J.; Konradsson, P.;
Kvanstrom, I.; Norberg, T.; Svensson, S. C. T.; Wigilius, B.
Acta Chem. Scand. 1985, 39, 569.
(4) Mueller, T.; Schneider, R.; Schmidt, R. Tetrahedron Lett.
1994, 35, 4763.
(5) Semmelheck, M. F.; Schmidt, C. R.; van Bekkum, H.
Recl.Trav. Chim. Pays-Bas. 1994, 113, 1665.
(6) Maurer, K.; Drefahl, G. Ber. 1942, 75, 1489.
(7) Yackel, D. F. C.; Kenyon, W. O. J. Am. Chem. Soc. 1942, 64,
121.
(8) (a) Heyns, K.; Paulsen, H. Angew. Chem., Int. Ed. Engl.
1957, 69, 600. (b) Kiss, J.; Noack, K.; D’souza, R. Helv.
Chim. Acta 1975, 301.
(9) Davis, N. J.; Flitsch, S. L. Tetrahedron Lett. 1993, 34, 1181.
(10) Gyorgydeak, Z.; Thiem, J. Carbohydrate Res. 1995, 268, 85.
(11) Preparation of 5a: b-D-Glucose penta-O-acetate (0.36 g,
0.92 mmol) and boron trifluoride diethyl etherate (0.13 ml,
0.92 mmol) were added to a solution of p-cresol (0.2 g, 1.84
mmol) in CH2Cl2 (5 mL) containing molecular sieves (4 Å,
activated). The resulting reaction mixture was protected
from moisture and stirred at 25–30 °C overnight (16 h). The
reaction mixture was diluted with CH2Cl2 (40 mL) and
washed with aq KOH (2 M, 4 × 25 mL), water (2 × 25 mL)
and brine (25 mL). The organic extracts were dried (MgSO4)
and concentrated under reduced pressure to afford a crude
off-white solid. The crude product was recrystallised from
anhyd EtOH to afford 5a (214 mg, 53%) as colourless
(17) Preparation of 7a: Glucoside 6a (200 mg, 0.74 mmol) was
dissolved in distilled water (10 mL) and TEMPO (0.005
equiv) and NaBr (0.15 equiv) were added. The solution was
cooled to 0 °C and a cold solution of 12–15% hypochlorite
in water (previously brought to pH = 10 by addition of 4 M
HCl) was added. The pH was controlled at ca 10–10.5 by
dropwise addition of KOH (0.5 M) with a syringe. The
reaction was complete within 30 min (TLC) during which
time the pH was generally stable. The reaction was quenched
by addition of EtOH (5 mL) and the mixture neutralised with
1 M HCl. The organic solvent was removed under reduced
pressure and the remaining solution was freeze-dried. The
crude product was purified using either XAD-2 column or
Waters® Sep-Pak column chromatography by eluting with
water, 50% aq MeOH and MeOH. Appropriate fractions
were pooled and the solvent removed under reduced
pressure and the crude product freeze-dried to afford pure 7a
(176 mg, 74%) as a white solid. 1H NMR (400 MHz, D2O):
d = 7.14 (d, J = 8.7 Hz, 2 H), 6.98 (d, J = 8.7 Hz, 2 H), 4.96
(d, J = 9.2 Hz, 1 H), 3.77 (d, J = 4.0 Hz, 1 H), 3.57–3.53
(envelope, 3 H), 2.13 (s, 3 H). 13C NMR (100 MHz, D2O):
d = 175.9, 154.9, 133.7, 130.7, 117.1, 100.9, 76.5, 73.2, 72.2,
20.1. HRMS–FAB+: m/z [M + H] calcd for C13H16O7K,
323.0533; found, 323.0521. HRMS–FAB+: m/z [M + K]
calcd for C13H15O7K2: 361.0092; found, 361.0059. MS
(FAB+): m/z (%) = 361 (10) [M + K], 323 (15) [M + H], 315
(10), 223(40), 131 (100), 39 (27) [K].
Synlett 2003, No. 13, 1981–1984 © Thieme Stuttgart · New York