A facile synthesis of 6-deoxydapagliflozin

Article abstract of DOI:10.1007/s00706-013-1053-0

A facile synthesis of a potent SGLT2 inhibitor, 6-deoxydapagliflozin, from methyl 2,3,4-tri-O-benzyl-6-deoxy-6-iodo-α-d-glucopyranoside in six steps with an overall yield of 50 % is described. The key steps were the reductive deiodination of the starting

Full text of DOI:10.1007/s00706-013-1053-0

Monatsh Chem (2013) 144:1903–1910
DOI 10.1007/s00706-013-1053-0
ORIGINAL PAPER
A facile synthesis of 6-deoxydapagliflozin
•
Yongheng Shi Huaqiang Xu Bingni Liu Weiling Kong
•
•
•
•
Qunchao Wei Weiren Xu Lida Tang Guilong Zhao
•
•
Received: 15 April 2013 / Accepted: 5 July 2013 / Published online: 1 October 2013
Ó Springer-Verlag Wien 2013
Abstract A facile synthesis of a potent SGLT2 inhibitor,
6-deoxydapagliflozin, from methyl 2,3,4-tri-O-benzyl-6-
deoxy-6-iodo-a-D-glucopyranoside in six steps with an
overall yield of 50 % is described. The key steps were the
reductive deiodination of the starting iodide under hydrog-
enolytic conditions to build the desired 6-deoxyglucose
functionality, the coupling of 2,3,4-tri-O-benzyl-6-deoxy-D-
gluconolactone and (5-bromo-2-chlorophenyl)(4-ethoxy-
phenyl)methane, followed by BF3ÁEt2O-mediated reduction
with Et3SiH in order to construct the desired anomeric
b-configuration. A variety of methods used for the cleavage
of benzyl groups in 2,3,4-tri-O-benzyl-1-[4-chloro-3-(4-eth-
Introduction
Diabetes is a progressive and chronic metabolic disease
characterized by hyperglycemia which, without adequate
treatment, can cause many severe complications. Although
a variety of anti-diabetic drugs are currently available, the
hyperglycemia still cannot be controlled in many cases,
even with combination therapy. New agents with novel
mechanisms of action are thus urgently needed.
More than 99 % of the plasma glucose that is filtered in
the renal glomerulus is reabsorbed into the blood, mainly
by sodium-dependent glucose cotransporter 2 (SGLT2) in
renal proximal tubules [1, 2]. Therefore, inhibition of
SGLT2 leads to the suppression of the reabsorption of
glucose from glomerular filtrate into blood, which lowers
blood glucose levels. SGLT2 inhibitors have become a
promising class of hypoglycemic agents for the treatment
of type 2 diabetes [1, 2], and the most advanced ones—
dapagliflozin (Fig. 1) [3] and canagliflozin [4]—have
recently been approved in the EU and USA.
oxybenzyl)phenyl]-1,6-dideoxy-b-D-glucopyranose
were
intensively screened, leading to the discovery of AlCl3-
mediated cleavage as the optimal method.
Keywords Carbohydrates Á Drug research Á
Hydrogenolysis Á Dapagliflozin Á SGLT2 inhibitor
In our earlier work,deletion of the 6-OH in the sugar
moiety of dapagliflozin led to the discovery of a more
potent SGLT2 inhibitor, 6-deoxydapagliflozin (1, Fig. 1).
In an in vitro assay, 1 was a more active SGLT2 inhibitor,
with IC50 = 0.67 nM against human SGLT2 (hSGLT2), as
compared with 1.1 nM for dapagliflozin, leading to the
identification of 1 as the most active SGLT2 inhibitor
discovered so far in this field. Also in an in vivo assay, 1
also introduced more urinary glucose in a rat urinary glu-
cose excretion test (UGE) and exhibited more potent blood
glucose inhibitory activity in a rat oral glucose tolerance
test (OGTT) than dapagliflozin.
Y. Shi Á L. Tang (&)
School of Pharmacy, Tianjin Medical University,
Tianjin 300070, People’s Republic of China
e-mail: tangld@tjipr.com
B. Liu Á W. Kong Á Q. Wei Á W. Xu Á G. Zhao (&)
Tianjin Key Laboratory of Molecular Design and Drug
Discovery, Tianjin Institute of Pharmaceutical Research,
Tianjin 300193, People’s Republic of China
e-mail: zhao_guilong@126.com
Given the fact that 6-dexoydapagliflozin (1) is a very
promising SGLT2 inhibitor that could be used to treat type
2 diabetes, we advanced it to preclinical trials in our
H. Xu
School of Chemistry and Chemical Engineering, Shandong
University, Jinan 250100, People’s Republic of China
123

1904
Y. Shi et al.
Fig. 1 The structures of
dapagliflozin and
6-deoxydapagliflozin (1)
laboratories. During preclinical development, kilograms of
1 were needed, so a facile synthetic route to 1 that is
characterized by a low cost, scalability, good reproduc-
ibility, and a high yield was required. More importantly,
such a synthetic route will be used for the mass production
of 1 when this drug candidate is forwarded to clinical trials
and the market. The original synthetic route to 1 that was
used during the drug discovery stage utilized dapagliflozin
as a starting material (Scheme 1; details will be published
elsewhere independently). The drawbacks of the original
route were its extremely expensive starting material,
dapagliflozin, and its poor reproducibility, so an alternative
synthetic route with the advantages mentioned above was
needed. Another serious problem with the original syn-
thetic route that had to be taken into consideration was that
dapagliflozin could not be used as the starting material to
synthesize 1 since dapagliflozin is a drug that has already
been fully patented.
glucose building block 2 is first 6-deoxylated to furnish the
6-deoxyglucose building block 6, which in turn is glycos-
idated with 7 to give the desired 6-deoxydapagliflozin (1).
We herein report a synthetic route based on the strategy
involved in pathway II (Scheme 3), which was experi-
mentally found to be superior to that based on pathway I,
as discussed above.
The synthetic route to the target compound 1 is shown in
Scheme 3. The starting material methyl 2,3,4-tri-O-benzyl-
6-deoxy-6-iodo-a-D-glucopyranoside (3) was prepared
from commercially available methyl a-D-glucopyranoside
(2) according to a known method [5, 6]. Iodide 3 was
reductively deiodinated to give 4 in 91 % yield under hy-
drogenolytic conditions using 10 % Pd/C as catalyst in the
presence of Et3N as base in THF/MeOH at room temper-
ature. To our surprise, this hydrogenolysis did not cleave
the benzyl ether functionality present in 3, presumably
because the iodide functionality is more reactive to hy-
drogenolysis than the benzyl ether functionality. Other
reaction conditions were also screened (Table 1). As
shown in Table 1, when the iodide 3 was treated with
Barton–McCombie reagent (n-Bu3SnH/AIBN) [7] in tolu-
ene at room temperature no reaction occurred; however,
when the reaction was carried out at elevated temperatures,
such as reflux, a complex mixture formed with only a trace
amount (3 %, entry 1) of the desired product 4. When the
iodide 3 was treated with LiAlH4 in THF at 0 °C to room
temperature, another complex mixture was produced with
only a trace amount (2 %, entry 2) of 4. When Pd(OH)2
was used as the hydrogenolysis catalyst instead of 10 %
Pd/C, the desired 4 was indeed formed (14 %, entry 4), but
most of the starting material was converted to a few more
polar byproducts, which were believed to result from the
cleavage of at least one of the benzyl groups. The con-
comitant cleavage of the benzyl group in 3 during
deiodination provides evidence that Pd(OH)2 is a more
Results and discussion
In order to find an alternative synthetic route to 1, we
performed retrosynthetic analysis (Scheme 2). As shown in
Scheme 2, target molecule 1 is characterized by two
structural features, 6-deoxyglucose and C-glycoside, so
two key steps, 6-deoxylation and C-glycosidation, should
be involved in retrosynthetic analysis. In pathway I in
Scheme 2, C-glycosidation is performed before 6-deoxy-
lation. Thus, D-gluconolactone (11) is glycosidated with an
aryl bromide 7 to give dapagliflozin, and subsequently
6-deoxylation of dapagliflozin yields the desired 6-de-
oxydapagliflozin (1). Obviously, our original synthetic
route to 1 (as shown in Scheme 1) is based on the strategy
involved in pathway I. In pathway II (see Scheme 2),
6-deoxylation is carried out before C-glycosidation. Thus,
123

A facile synthesis of 6-deoxydapagliflozin
1905
powerful hydrogenolysis catalyst than Pd/C since it can
produce highly active Pd(0) species in situ under hydrog-
enolytic conditions. Note that attempts to control this
reaction by shortening the reaction time failed, since these
polar byproducts were observed to be the major products
once the reaction was initiated.
Table 1 Screening of the reaction conditions for reductive deiodin-
ation of 3
Entry
Reagents and condition
Yield/%a
1
2
3
4
a
n-Bu3SnH/AIBN, PhMe, rt to reflux
LiAlH4, THF, 0 °C to rt
10 % Pd/C, Et3N, MeOH, rt
3
2
91
14
Methyl 6-deoxyglucoside 4 was hydrolyzed with 6 M
hydrochloric acid in acetic acid at 80–85 °C to afford lactol
5 as an anomeric mixture, as verified by 1H NMR. It should
Pd(OH)2, Et3N, MeOH, rt
Isolated yield
123

1906
Y. Shi et al.
be noted that the reaction must be quenched within 30 min,
otherwise a lot of byproducts will emerge, complicating the
workup and dramatically decreasing the yield [8]. Lactol 5
was oxidized to the corresponding lactone 6 under Swern
oxidation conditions with Ac2O/DMSO at room tempera-
ture. 7 was treated with n-BuLi at -78 °C in dried THF to
afford the corresponding aryl lithium by bromine–lithium
exchange, which in turn underwent the nucleophilic attack
on lactone 6 to give rise to the adduct after the addition of 6
at -78 °C. The adduct thus obtained was treated with
MsOH/MeOH in a one-pot manner to produce the corre-
sponding methyl glycoside 8 as an anomeric mixture.
Crude 8 obtained after workup was subjected to reduction
with Et3SiH in the presence of BF3ÁEt2O at -30 °C in
dried CH2Cl2 to yield 9 with the desired b-configuration.
Note that the reduction should be carried out at low tem-
peratures, such as -30 to -50 °C, since low temperatures
favor the formation of the thermodynamically stable b-
configuration, as in 9; elevated temperatures, such as 0 °C,
would lead to the formation of a large amount of unwanted
a-anomer.
Finally, the benzyl ether 9 was treated with anhydrous
AlCl3 in anisole at 0 °C to room temperature [9] to afford
the target compound 1. Other reaction conditions to cleave
the benzyl ether functionality in 9 were also attempted; the
results are outlined in Table 2. Initially, we anticipated that
we would acetolyze the benzyl ether functionality in 9 to
give triacetylated 1 by treatment of 9 with an acid in Ac2O,
because the triacetylated 1 can be more readily purified by
column chromatography and/or recrystallization than 1
itself. Triacetylated 1 can also be readily hydrolyzed to give
1. However, when 9 was treated with BF3ÁEt2O in Ac2O at
0 °C to rt [10], the starting 9 was consumed completely
within 1 h, but only a trace amount of triacetylated 1 was
observed and isolated (2 %, entry 1). Likewise, attempts to
use TMSOTf [11] and concentrated H2SO4 as acids in place
of BF3ÁEt2O failed in that, in both cases, the starting 9
disappeared within 30 min, with only a trace amount (3 %,
entry 2) of or no (0 %, entry 3) triacetylated 1 formed,
respectively. Given the unsuccessful results described
above, we turned to the direct cleavage of benzyl ether to
produce 1 instead of the original conversion to its triacetate.
Thus, TMSI was first attempted in the form of neat TMSI
[12] (entry 4) or a solution in MeCN [13] (entry 5), but both
gave poor yields (7 %, entry 4; 6 %, entry 5), along with a
complex mixture of unidentified byproducts. Since Pd-cat-
alyzed hydrogenolysis is commonly used to cleave the
benzyl ether functionality, we systematically studied this
method using 10 % Pd/C or Pd(OH)2 as catalyst and H2,
HCO2H, HCO2NH4, or cyclohexene as the hydrogen source
(entries 6–13). However, all of these attempts were unsuc-
cessful, since the Ar–Cl bond could be concomitantly
cleaved, leading to the formation of a dechlorinated
byproduct 10 (Scheme 4), which could not be separated
from the desired product 1 by TLC and column chroma-
tography because their polarities are extremely similar. To
our surprise, a seemingly complex combination of a number
of reagents, TfOH/TFA/Me2S/m-cresol/(CH2SH)2 [9],
worked very well, giving the desired product 1 in 75 %
yield (entry 14). However, this method is still unsatisfactory
since the odors of the two reagents involved, Me2S and
(CH2SH)2, are extremely unpleasant, making it unsuitable
for the mass production of 1. Finally, two Lewis acids, BCl3
[14] and AlCl3 [9], were attempted, and fortunately both
gave satisfactory results. Thus, the starting 9 was treated
with BCl3 in CH2Cl2 at -50 °C to rt (entry 15) or AlCl3 in
anisole at 0 °C to rt (entry 16), respectively, and both
reactions proceeded rapidly and completed within 30 min.
The desired product 1 was isolated in either 81 or 93 %
yield, respectively. The reaction with AlCl3 possessed
overwhelming advantages over that with BCl3, since the
reaction with BCl3 was not very clean and yielded a few
byproducts, and it involved a much lower reaction tem-
perature (-50 °C to rt vs. 0 °C to rt) and a slightly lower
yield than that with AlCl3 (81 vs. 93 %). So, AlCl3/anisole
was finally selected as the best method for this conversion.
Experimental
Melting points were determined with an X-4 (Biocotek,
1
Zheijiang, China) microscopic melting point apparatus. H
NMR and 13C NMR spectra were taken on a Bruker
(Rheinstetten, Germany) AV400 nuclear magnetic reso-
nance spectrometer, using DMSO-d6 or CDCl3 as solvents
and TMS as internal standard. IR spectra were recorded on
a Bruker (Ettlingen, Germany) Vector 22 FT-IR spectro-
photometer as KBr discs or thin films. High-resolution
mass spectra were obtained with an Agilent (Santa Clara,
CA, USA) 6510 Q-TOF mass spectrometer with the elec-
trospray ionization (ESI) technique and direct injection.
Methyl 2,3,4-tri-O-benzyl-6-deoxy-6-iodo-a-D-glucopy-
ranoside (3) was prepared from commercially available
methyl a-D-glucopyranoside (2) according to a known
method [5, 6] (m.p.: 57 °C, literature value [15]: 61–62 °C).
(5-Bromo-2-chlorophenyl)(4-ethoxyphenyl)methane
(7)
was prepared by a known procedure [3] (m.p.: 40–41 °C,
literature value [3]: 37 °C). Dried THF and CH2Cl2 were
distilled from sodium and CaH2, respectively.
Methyl 2,3,4-tri-O-benzyl-6-deoxy-a-D-glucopyranoside
(4, C28H32O5)
A 250-cm3 round-bottomed flask was charged with 17.23 g
3 (30 mmol), 9.11 g Et3N (90 mmol), 1.00 g 10 % Pd/C,
100 cm3 MeOH, and 50 cm3 THF, and the resulting
mixture was subjected to hydrogenolysis with a hydrogen
123

A facile synthesis of 6-deoxydapagliflozin
1907
Table 2 Screening of reaction
Entry
Reagents and conditions
Yield/%a
conditions for the cleavage of
the benzyl ether functionality in
1
BF3ÁEt2O, Ac2O, 0 °C to rt
TMSOTf, Ac2O, -50 °C to rt
cH2SO4, Ac2O, -20 °C to rt
TMSI (neat), 0–50 °C
TMSI, MeCN, 0–50 °C
H2, 10 % Pd/C, THF/MeOH, rt
2b
3b
0b
7
9
2
3
4
5
6
7
8
6
17c
1c
0d
0d
4c
1c
0d
0d
75
81
93
HCO2H, 10 % Pd/C, THF/MeOH, rt to 50 °C
HCO2NH4, 10 % Pd/C, THF/MeOH, rt to 50 °C
Cyclohexene, 10 % Pd/C, THF/MeOH, rt to 50 °C
H2, Pd(OH)2, THF/MeOH, rt
9
a
Isolated yield
10
b
The product was 2,3,4-tri-O-
11
HCO2H, Pd(OH)2, THF/MeOH, rt to 50 °C
HCO2NH4, Pd(OH)2, THF/MeOH, rt to 50 °C
Cyclohexene, Pd(OH)2, THF/MeOH, rt to 50 °C
TfOH, TFA, Me2S, m-cresol, (CH2SH)2, -15 °C to rt
BCl3, CH2Cl2, -50 °C to rt
acetyl-1
c
Contaminated with
12
inseparable dechlorinated
13
byproduct 10
14
d
Only dechlorinated byproduct
10 was separated out, with none
of the desired 1
15
16
AlCl3, anisole, 0 °C to rt
balloon at room temperature until 3 was consumed
completely, as indicated by TLC (typically within 10 h).
The reaction mixture was filtered off and the filtrate was
concentrated on a rotary evaporator to half its original
volume and poured into 300 cm3 water. The mixture thus
obtained was extracted with CH2Cl2 (100 cm3 9 3). The
combined extracts were washed successively with 100 cm3
1 % diluted hydrochloric acid and saturated brine, dried
over Na2SO4, and evaporated on a rotary evaporator to
afford the crude product 4 as a pale yellow oil, which was
purified by column chromatography to yield the pure
product 4 as a colorless oil (12.25 g, 91 %). Rf = 0.60
(EtOAc/hexane = 1/6 by v/v); 1H NMR (400 MHz,
DMSO-d6): d = 7.26–7.34 (m, 15H), 4.83 (d, 1H,
J = 11.2 Hz), 4.77 (d, 1H, J = 11.2 Hz), 4.74 (d, 1H,
J = 3.2 Hz, H-1), 4.69 (d, 1H, J = 11.2 Hz), 4.60–4.67
(m, 2H), 4.59 (d, 1H, J = 11.6 Hz), 3.71 (t, 1H,
J = 9.2 Hz), 3.54–3.58 (m, 1H), 3.47 (dd, 1H,
J = 3.6 Hz, 9.6 Hz), 3.28 (s, 3H, OCH3), 3.10 (t, 1H,
J = 9.2 Hz), 1.16 (d, 3H, J = 6.4 Hz, CHCH3) ppm; 13C
NMR (100 MHz, DMSO-d6): d = 138.76, 138.47, 138.45,
128.14, 128.08, 127.58, 127.54, 127.42, 127.28, 96.77,
83.11, 80.89, 79.83, 74.35, 74.10, 71.37, 65.89, 54.43,
17.71 ppm; IR (thin film): v = 3,088 (w), 3,063 (w), 3,031
(w), 2,905 (s), 1,496 (s), 1,453 (s), 1,360 (s), 1,074
(s) cm-1; HR–MS: calcd for C28H36NO5 ([M ? NH4]?)
466.2588, found 466.2595.
2,3,4-Tri-O-benzyl-6-deoxy-a/b-D-glucopyranose
(5, C27H30O5)
A 250-cm3 round-bottomed flask was charged with 12.03 g
4 (26.8 mmol) and 80 cm3 glacial acetic acid, and the
mixture was stirred at room temperature before 8 cm3 6 M
hydrochloric acid were added. The resulting mixture was
123

1908
Y. Shi et al.
stirred at 80–85 °C for 30 min, when TLC analysis
indicated that 4 was almost consumed. The reaction
mixture was poured into 300 cm3 ice water, and the
mixture thus obtained was extracted with CH2Cl2
(100 cm3 9 3). The combined extracts were washed
successively with saturated aqueous NaHCO3 and saturated
brine, dried over Na2SO4, and evaporated on a rotary
evaporator to afford the crude product 5 as a brown oil,
which was purified by column chromatography to furnish
the pure product 5 as a white solid (10.01 g, 86 %).
Rf = 0.30 (EtOAc/hexane = 1/5 by v/v); m.p.: 92–94 °C;
1H NMR (400 MHz, DMSO-d6): d = 7.25–7.35 (m, 15H),
6.92 (d, 0.55H, J = 6.0 Hz), 6.45 (d, 0.45H, J = 4.8 Hz),
5.13 (t, 0.43H, J = 3.6 Hz), 4.54–4.90 (m, 5.57H), 3.76–
3.82 (m, 0.97H), 3.53 (t, 0.74H, J = 9.0 Hz), 3.37–3.42
(m, 1.19H), 3.19 (t, 0.86H, J = 8.6 Hz), 3.03–3.13 (m,
1.24H), 1.18 (d, 1.74H, J = 6.0 Hz), 1.13 (d, 1.26H,
J = 6.4 Hz) ppm; 13C NMR (100 MHz, DMSO-d6):
d = 138.89, 138.84, 138.72, 138.61, 138.55, 138.43,
128.15, 128.11, 128.08, 128.04, 127.65, 127.60, 127.51,
127.48, 127.40, 127.33, 127.27, 127.24, 127.22, 96.51,
89.40, 83.75, 83.54, 83.33, 83.17, 80.74, 80.32, 74.36,
74.27, 74.05, 74.01, 73.43, 71.13, 69.87, 65.26, 17.87,
17.81 ppm; IR (KBr): v = 3,407 (s), 3,064 (w), 3,031 (w),
2,919 (m), 2,874 (m), 1,497 (w), 1,453 (s), 1,360 (s), 1,080
(s) cm-1; HR–MS: calcd for C27H34NO5 ([M ? NH4]?)
452.2431, found 452.2438.
J = 5.6 Hz, 8.8 Hz), 1.39 (d, 3H, J = 6.4 Hz, CHCH3)
ppm; 13C NMR (100 MHz, CDCl3): d = 168.99, 137.36,
137.30, 136.77, 128.48, 128.45, 128.43, 128.34, 128.14,
127.99, 127.93, 81.36, 81.15, 77.29, 74.58, 73.48, 73.21,
72.97, 18.27 ppm; IR (KBr): v = 3,086 (w), 3,060 (w),
3,029 (m), 2,944 (m), 2,877 (s), 1,746 (s), 1,496 (s), 1,454
(s), 1,364 (s), 1,226 (s), 1,193 (s), 1,108 (s), 1,070 (s),
1,017 (s) cm-1
;
HR–MS: calcd for C27H32NO5
([M ? NH4]?) 450.2275, found 450.2272.
2,3,4-Tri-O-benzyl-1-[4-chloro-3-(4-ethoxybenzyl)phenyl]-
1,6-dideoxy-b-D-glucopyranose (9, C42H43ClO5)
A dried 250-cm3 round-bottomed flask was charged with
5.86 g 7 (18 mmol) [3], 60 cm3 dried THF, and a magnetic
stirring bar, flushed with nitrogen, and sealed with a rubber
septum. The flask was cooled to -78 °C with liquid
nitrogen–EtOH and magnetic stirring was initiated.
Through a syringe, 11.25 cm3 1.6 M n-BuLi in hexane
(18 mmol) were added dropwise. After this addition, the
reaction mixture was stirred for another 30 min at this
temperature, before a solution of 7.79 g 6 (18 mmol) in
20 cm3 dried THF was added dropwise. After this addition,
the stirring was continued for 1 h at this temperature. A
solution of 3.84 g MsOH (40 mmol) in 15 cm3 MeOH was
added dropwise via a syringe at this temperature. After this
addition, the reaction mixture was stirred at room temper-
ature overnight. The reaction mixture was poured into
300 cm3 water and extracted with CH2Cl2 (100 cm3 9 3).
The combined extracts were washed with saturated brine,
dried over Na2SO4, and evaporated on a rotary evaporator
to afford the crude product 8 as a yellow oil (Rf = 0.20,
EtOAc/hexane = 1/10 by v/v), which was used directly in
the next step without further purification.
2,3,4-Tri-O-benzyl-6-deoxy-D-gluconolactone
(6, C27H28O5)
A dried 250-cm3 round-bottomed flask was charged with
60 cm3 of DMSO and, while stirring, 40 cm3 of acetic
anhydride was added dropwise. The flask was cooled in an
ice water bath once the addition had commenced. After this
addition, the mixture thus obtained was stirred for another
30 min before a solution of 9.91 g 5 (22.81 mmol) in
20 cm3 DMSO was added dropwise. After this addition,
the reaction mixture was stirred at room temperature
overnight until TLC analysis showed that the reaction was
completed. The reaction mixture was poured into 400 cm3
ice water and stirred for another 30 min. The mixture thus
obtained was extracted with CH2Cl2 (100 cm3 9 3). The
combined extracts were washed successively with saturated
aqueous NaHCO3 and saturated brine, dried over Na2SO4,
and evaporated on a rotary evaporator to afford the crude
product 6 as a pale yellow oil, which was purified by
column chromatography to furnish the pure product 6 as a
white solid (8.09 g, 82 %). Rf = 0.30 (EtOAc/hexane =
1/10); m.p.: 66–67 °C; 1H NMR (400 MHz, CDCl3):
d = 7.22–7.38 (m, 15H), 4.92 (d, 1H, J = 11.2 Hz), 4.68
(d, 1H, J = 11.2 Hz), 4.64 (d, 1H, J = 11.6 Hz), 4.62 (d,
1H, J = 11.6 Hz), 4.50–4.56 (m, 3H), 4.10 (d, 1H,
J = 4.8 Hz), 3.88 (t, 1H, J = 5.4 Hz), 3.44 (dd, 1H,
The crude 8 obtained above and 4.19 g Et3SiH
(36 mmol) were dissolved in 60 cm3 dried CH2Cl2 and
stirred at -30 °C. This was followed by the dropwise
addition of 2.55 g BF3ÁEt2O (18 mmol) dissolved in 5 cm3
dried CH2Cl2. After this addition, the stirring was contin-
ued for 1 h at -30 °C and then overnight at room
temperature until 8 was consumed completely, as indicated
by TLC analysis. To the reaction mixture, 20 cm3 saturated
aqueous NaHCO3 were carefully added, and the mixture
thus obtained was stirred for 30 min before being poured
into 200 cm3 of ice water. The mixture thus obtained was
extracted with CH2Cl2 (100 cm3 9 3). The combined
extracts were washed with saturated brine, dried over
Na2SO4, and evaporated on a rotary evaporator to afford
the crude product 9 as a brown oil, which was purified by
column chromatography to furnish the pure product 9 as a
white solid (10.03 g, 84 % from 6). Rf = 0.30 (EtOAc/
hexane = 1/10 by v/v); m.p.: 97–98 °C; 1H NMR
(400 MHz, DMSO-d6): d = 7.41 (d, 1H, J = 8.4 Hz),
7.22–7.35 (m, 12H), 7.14–7.20 (m, 3H), 7.03 (d, 2H,
123

A facile synthesis of 6-deoxydapagliflozin
1909
J = 8.4 Hz), 6.83–6.85 (m, 3H), 6.73 (d, 2H, J = 8.4 Hz),
4.76–4.82 (m, 3H), 4.66 (d, 1H, J = 11.2 Hz), 4.35 (d, 1H,
J = 10.8 Hz), 4.23 (d, 1H, J = 9.6 Hz, H-1), 3.89–4.01
(m, 4H), 3.78 (d, 1H, J = 10.8 Hz), 3.69 (t, 1H,
J = 8.8 Hz), 3.47–3.55 (m, 2H), 3.27 (t, 1H,
J = 9.2 Hz), 1.27 (t, 3H, J = 7.0 Hz, CH2CH3), 1.20 (d,
3H, J = 6.0 Hz, CHCH3) ppm; 13C NMR (100 MHz,
DMSO-d6): d = 156.83, 138.58, 138.35, 138.33, 137.78,
132.45, 130.95, 130.71, 129.46, 129.08, 128.17, 128.12,
127.91, 127.71, 127.49, 127.43, 127.33, 127.29, 127.24,
127.10, 114.21, 85.57, 83.44, 83.38, 79.66, 74.67, 74.46,
74.19, 73.68, 62.78, 37.47, 17.98, 14.57 ppm; IR (KBr):
v = 3,030 (m), 2,976 (m), 2,901 (s), 1,611 (s), 1,511 (s),
1,475 (s), 1,452 (s), 1,395 (m), 1,360 (s), 1,245 (s), 1,174
(m), 1,137 (s), 1,109 (s), 104 (s) cm-1; HR–MS: calcd for
C42H47ClNO5 ([M ? NH4]?) 680.3137, found 680.3125.
Pd/C, 1 cm3 THF, and 5 cm3 MeOH, and the reaction
mixture was stirred at 50 °C under N2 until 9 was
consumed completely (typically within 12 h). The reaction
mixture was filtered off, and the filtrate was evaporated on
a rotary evaporator to afford the crude product as a residue,
which was purified by column chromatography to afford
the pure 10 as a white solid after trituration with EtOAc/
hexane (1/2 by v/v) (0.23 g, 84 %). Rf = 0.35 (EtOAc);
m.p.: 117–118 °C; 1H NMR (400 MHz, DMSO-d6):
d = 7.06–7.22 (m, 6H), 6.81 (d, 2H, J = 8.8 Hz), 4.93
(d, 1H, J = 5.2 Hz), 4.87 (d, 1H, J = 4.4 Hz), 4.70 (d, 1H,
J = 5.2 Hz), 3.93–3.99 (m, 3H), 3.84 (s, 2H), 3.27 (dd, 1H,
J = 6.0 Hz, 9.2 Hz), 3.15–3.24 (m, 2H), 2.90–2.96 (m,
1H), 1.29 (t, 1H, J = 7.0 Hz), 1.15 (d, 3H, J = 6.4 Hz)
ppm; 13C NMR (100 MHz, DMSO-d6): d = 156.75,
140.97, 140.43, 133.03, 129.62, 128.04, 127.70, 127.60,
125.30, 114.26, 81.58, 78.18, 75.76, 75.58, 74.71, 62.83,
40.21, 18.22, 14.64 ppm; IR (KBr): v = 3,386 (s), 3,031
(w), 2,977 (m), 2,904 (m), 1,613 (m), 1,511 (s), 1,478 (m),
1,443 (m), 1,394 (m), 1,299 (m), 1,247 (s), 1,101 (s), 1,051
(s), 1,013 (s) cm-1; HR–MS: calcd for C21H30NO5
([M ? NH4]?) 376.2118, found 376.2124.
1-[4-Chloro-3-(4-ethoxybenzyl)phenyl]-1,6-dideoxy-b-D-
glucopyranose (1, C21H25ClO5)
A dried 100-cm3 round-bottomed flask was charged with
6.63 g 9 (10 mmol) and 60 cm3 anisole and cooled in an
ice water bath. While stirring, 13.33 g anhydrous AlCl3
(100 mmol) were added portionwise. After this addition,
the reaction mixture was stirred at room temperature for
30 min, when TLC analysis indicated that 9 was consumed
completely. The reaction mixture was poured into 300 cm3
ice water and extracted with EtOAc (100 cm3 9 3). The
combined extracts were washed with saturated brine, dried
over Na2SO4, and evaporated on a rotary evaporator to
afford the crude product 1 as a yellow oil. The crude 1 was
purified by column chromatography to furnish the pure
product 1 as a white foam, which was further purified by
trituration with EtOAc/hexane (1/2 by v/v) to afford the
pure product 1 as a white solid (3.65 g, 93 %). Rf = 0.35
Acknowledgments The authors thank the Key Project of Tianjin
Science and Technology Support Plan (10ZCKFSH01300) and
Natural Science Foundation of China (21302141) for financial
support.
References
1. Washburn WN (2009) J Med Chem 52:1785
2. Hardman TC, Dubrey SW (2011) Diabetes Ther 2:133
3. Meng M, Ellsworth BA, Nirschl AA, McCann PJ, Patel M,
Girotra RN, Wu G, Sher PM, Morrison EP, Biller SA, Zahler
R, Deshpande PP, Pullockaran A, Hagan DL, Morgan N,
Taylor JR, Obermeier MT, Humphreys WG, Khanna A, Dis-
cenza L, Robertson JM, Wang A, Han S, Wetterau JR, Janovitz
EB, Flint OP, Whaley JM, Washburn WN (2008) J Med Chem
51:1145
4. Nomura S, Sakamaki S, Hongu M, Kawanishi E, Koga Y, Sa-
kamoto T, Yamamoto Y, Ueta K, Kimata H, Nakayama K,
Tsuda-Tsukimoto M (2010) J Med Chem 53:6355
5. Lee JC, Francis S, Georg GI, Dutta D, Gupta V, Tash JS, Yang Y,
Zhu J-Y, Schoenbrunn E (2012) J Org Chem 77:3082
6. Bartoszewicz A, Kalek M, Stawinski J (2008) Tetrahedron
64:8843
7. Zhao GL, Lou YY, Zhang LS, Shao H, Xu WR, Tang LD, Zou
MX (2012) Synth Commun 42:2885
8. Koto S, Morishima N, Miyata Y, Zen S (1976) Bull Chem Soc
Jpn 49:2639
9. Kakinuma H, Oi T, Hashimoto-Tsuchiya Y, Arai M, Kawakita Y,
Fukasawa Y, Iida I, Hagima N, Takeuchi H, Chino Y, Asami J,
Okumura-Kitajima L, Io F, Yamamoto D, Miyata N, Takahashi
T, Uchida S, Yamamoto K (2010) J Med Chem 53:3247
10. Xu J, Egger A, Bernet B, Vasella A (1996) Helv Chim Acta
79:2004
1
(EtOAc); m.p.: 148–149 °C; H NMR (400 MHz, DMSO-
d6): d = 7.35 (d, 1H, J = 8.4 Hz), 7.25 (s, 1H), 7.18 (d,
1H, J = 8.0 Hz), 7.08 (d, 2H, J = 8.4 Hz), 6.81 (d, 2H,
J = 8.4 Hz), 4.95 (d, 1H, J = 5.2 Hz, OH), 4.90 (d, 1H,
J = 4.4 Hz, OH), 4.79 (d, 1H, J = 5.6 Hz, OH), 3.92–4.01
(m, 5H), 3.24–3.29 (m, 1H), 3.18–3.22 (m, 1H), 3.09–3.15
(m, 1H), 2.89–2.95 (m, 1H), 1.29 (t, 3H, J = 7.0 Hz,
CH2CH3), 1.15 (d, 3H, J = 6.0 Hz, CHCH3) ppm; 13C
NMR (100 MHz, DMSO-d6): d = 156.85, 139.65, 137.82,
131.83, 131.16, 130.58, 129.52, 128.65, 127.14, 114.26,
80.71, 77.98, 75.77, 75.51, 74.81, 62.84, 37.55, 18.19,
14.62 ppm; IR (KBr): v = 3,564 (w), 3,385 (s), 2,981 (s),
2,899 (s), 2,861 (s), 1,613 (m), 1,512 (s), 1,477 (m), 1,247
(s), 1,102 (s), 1,045 (s), 1,012 (s) cm-1; HR–MS: calcd for
C21H29ClNO5 ([M ? NH4]?) 410.1729, found 410.1724.
1,6-Dideoxy-1-[3-(4-ethoxybenzyl)phenyl]-
b-D-glucopyranose (10, C21H26O5)
A 100-cm3 round-bottomed flask was charged with 0.50 g
9 (0.75 mmol), 0.88 g HCO2NH4 (14 mmol), 0.10 g 10 %
11. Lee SH, Song K-S, Kim JY, Kang M, Lee JS, Cho S-H, Park H-J,
Kim J, Lee J (2011) Bioorg Med Chem 19:5813
123

1910
Y. Shi et al.
12. Song K-S, Lee SH, Kim MJ, Seo HJ, Lee J, Lee S-H, Jung ME,
Son E-J, Lee MW, Kim J, Lee J (2011) ACS Med Chem Lett
2:182
13. Lee J, Lee S-H, Seo HJ, Son E-J, Lee SH, Jung ME, Lee MW,
Han HK, Kim J, Kang J, Lee J (2010) Bioorg Med Chem 18:2178
14. Kim MJ, Lee SH, Park SO, Kang H, Lee JS, Lee KN, Jung ME,
Kim J, Lee J (2011) Bioorg Med Chem 19:5468
15. Skaanderup PR, Poulsen CS, Hyldtoft L, Jørgensen MR, Madsen
R (2002) Synthesis 1721
123