β-Selective C-Arylation of Diisobutylaluminum Hydride Modified 1,6-Anhydroglucose: Synthesis of Canagliflozin without Recourse to Conventional Protecting Groups

Article abstract of DOI:10.1021/acs.joc.5b00601

The β-selective phenylation of benzyl and boronate protected 1,6-anhydroglucose and the direct phenylation of unprotected 1,6-anhydroglucose (10), pretreated with i-Bu₂AlH, i-Bu₃Al, Et₃Al, Me₃Al, or n-octyl₃Al, with triphenylalane or aryl(chloro)alanes is reported. The utility of the unprotected version of the method is demonstrated by the synthesis of the SGLT2 inhibitor, canagliflozin (1a), from commercially available 10 in one C-C bond-forming step. This approach circumvents the need for conventional protecting groups, and therefore no formal protection and deprotection steps are required. (Chemical Presented).

Full text of DOI:10.1021/acs.joc.5b00601

Article
pubs.acs.org/joc
β‑Selective C‑Arylation of Diisobutylaluminum Hydride Modified 1,6-
Anhydroglucose: Synthesis of Canagliflozin without Recourse to
Conventional Protecting Groups
Julian P. Henschke,*,‡ Chen-Wei Lin,‡ Ping-Yu Wu,‡ Wen-Shing Tsao,‡ Jyh-Hsiung Liao,‡
and Pei-Chen Chiang§
‡Department of Exploratory Research and §Department of Analytical Research and Development, ScinoPharm Taiwan, No. 1, NanKe
8th Road, Tainan Science-Based Industrial Park, ShanHua, Tainan, 74144 Tainan County, Taiwan
S
* Supporting Information
ABSTRACT: The β-selective phenylation of benzyl and
boronate protected 1,6-anhydroglucose and the direct phenyl-
ation of unprotected 1,6-anhydroglucose (10), pretreated with
i-Bu2AlH, i-Bu3Al, Et3Al, Me3Al, or n-octyl3Al, with tripheny-
lalane or aryl(chloro)alanes is reported. The utility of the
unprotected version of the method is demonstrated by the
synthesis of the SGLT2 inhibitor, canagliflozin (1a), from
commercially available 10 in one C−C bond-forming step.
This approach circumvents the need for conventional
protecting groups, and therefore no formal protection and
deprotection steps are required.
More recently several new approaches to β-C-arylglucoside
synthesis have been reported including Lemaire et al.6
transition-metal-free stereoselective (>99:1 β:α) arylation of a
per-O-pivaloyl protected glucosyl bromide substrate with diaryl
zinc reagents and Sakamaki et al.7 palladium-catalyzed Suzuki
cross-coupling of a glucal pinacol boronate with aryl bromides
followed by a hydroboration-oxidation sequence.
INTRODUCTION
The β-C-arylglucosides canagliflozin (1a), dapagliflozin (1b),
ipragliflozin (1c), and empagliflozin (1d) (Figure 1) are
■
Additionally, we recently reported8 that 2,4-di-O-TBDPS
protected 1,6-anhydro-β-D-glucopyranose 2a can be arylated
with high stereoselectivity to produce β-C-aryl-glucosides, such
as 3a (Scheme 1), using triarylalanes, aryl(halo)alanes,
Scheme 1. Arylation of 2b with Ph3Al
Figure 1. β-C-Arylglucosides useful for the treatment of diabetes.
members of a new class of antidiabetes active pharmaceutical
ingredient known as Sodium-coupled GLucose co-Transporter
2 (SGLT2) inhibitors (1).1 Owing to their therapeutic
effectiveness, this class of compound has received much
attention.2 In 1988 and 1989 Kraus et al. and Czernecki et al.
independently3 reported that per-benzyl-protected β-C-arylglu-
cosides can be prepared by the 1,2-addition of aryllithium or
aryl Grignard compounds to 2,3,4,6-tetra-O-benzyl-D-glucono-
lactone followed by silane reduction. Owing to the only
moderate anomeric selectivity (4:1 β:α) and use of benzyl ether
protection in the original method, improvements have since
been made,4 including the use of silyl and acetyl protection and
modification of the reducing agent, that have allowed β:α-
selectivity as high as 65:1 and rendered the method viable for
the preparation of drug candidates on a multikilogram scale.5
aryl(alkyl)haloalanes, or aryl(alkyl)alanes. Demonstrated by
the synthesis of the SGLT2 inhibitors canagliflozin (1a) and
dapagliflozin (1b), the arylalanes were prepared from aryl
lithium compounds or Grignard reagents and AlCl3. Activation
of the triarylalanes with AlCl3 and filtration produced metal
halide-free aryl(chloro)alanes, ArmAlCln, that possessed in-
Received: March 17, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.joc.5b00601
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
creased reactivity. Despite the presence of the unprotected C3
hydroxyl group in 2a, pretreatment with 1 equiv of Me3Al or i-
Bu2AlH to form alkoxy(dialkyl)alanes 2b allowed efficient
phenylation giving 3a in high yield with 1 equiv of Ph3Al. In the
absence of Me3Al or i-Bu2AlH 2 equiv of Ph3Al were required
to achieve the same yield confirming that the dialkylalanyl
moiety behaved as a hydroxyl protecting group. Arylalanes,
vinylalanes, propargylalanes, and/or their derivatives have been
used in the arylation, vinylation and propargylation of hydroxy-
protected 1,2-anhydrofuranose and 1,2- and 1,6-anhydropyr-
anose derivatives.9 Herein we report further work on the
arylation of 1,6-anhydroglucose derivatives and ultimately
demonstrate that 1,6-anhydro-β-D-glucopyranose (10) can be
arylated without the need of conventional protecting groups in
essentially one synthetic step.
Figure 2. Miscellaneous compounds.
and supports our previous supposition that an unprotected, β-
disposed hydroxyl group at C3 is not required to direct the β-
selective arylation of 1,6-anhydro-β-D-glucopyranose deriva-
tives.8a The same degree of efficiency was obtained using
PhOMe as solvent (entry 3). The metal halide-free phenyl-
(chloro)alanes Ph2AlCl (entry 4) or Ph2.5AlCl0.5 (entry 5),12
that were found particularly effective in the arylation of 2b,8a
prepared by combining commercial Ph3Al in n-Bu2O with AlCl3
in THF, were also tested. Surprisingly, HPLC yields of only 7−
12% of 3c were produced after 4 h along with significant levels
of benzyl alcohol. Further heating led to degradation of the
product indicating that the tri-O-benzyl substrate and product
were not compatible with aryl(chloro)alanes.
RESULTS AND DISCUSSION
■
Arylation of Tri-O-protected Substrates. As reported
previously,8 whereas the arylation of 2,4-di-O-TBDPS protected
substrate 2b was efficient, treatment of 2,3,4-tri-O-TBS
substrate 2c in PhOMe with commercial Ph3Al in n-Bu2O
(Table 1, entry 1) only produced a 12% isolated yield of β-C-
a
Table 1. Arylation of Tri-O-protected Analogues
Arylation of 2,4-O-Boronic Ester-Protected Sub-
strates. When the known13 2,4-O-phenylboronic ester 2e,
synthesized by the condensation of 1,6-anhydroglucose 10 and
PhB(OH)2 in a Dean−Stark assembly, was analyzed by TLC or
reverse phase HPLC or by 1H NMR spectroscopy in moist d6-
DMSO, only 1,6-anhydroglucose and PhB(OH)2 were
detected. Wanting to exploit the apparent lability of this
protecting group it was envisioned that direct conversion of 2e
to unprotected β-C-arylglucosides, such as β-3d, should be
possible with only a simple reaction quench and no need for
isolation of the boronate-protected C-glucoside product.
Indeed, when 2e was heated with Ph2AlCl in PhMe
overnight followed by a MeOH quench and column
chromatography, a 34% yield of arylglucoside β-3d was
obtained, along with 4% of disaccharide 714 and 23% unreacted
1,6-anhydroglucose (Table 2, entry 1).15 The chromato-
graphically isolated yield of β-3d was improved to 55%
(entry 2) when 2e was prepared and arylated in one pot:
1,6-anhydroglucose was condensed with PhB(OH)2 in PhMe
(Dean−Stark apparatus), mixed with PhCN and 1 equiv of
Ph3Al in n-Bu2O, concentrated under vacuum to remove PhMe
and n-Bu2O, and heated for 3 h. HPLC analysis of the product
mixture prior to purification showed β-3d, PhB(OH)2, PhOH,
and benzene but did not reveal detectable amounts of the α-
anomer, α-3d, indicating that the reaction was highly
stereoselective.16 Arylation was very rapid (≤3.5 h) and
provided good HPLC yields of β-3d when phenyl(chloro)-
alanes were used (entries 4 and 5, Table 2). Although slower
(16 h), the use of 2.4 equiv of Ph2.5AlCl0.5 in PhCN allowed
arylation at as low as 80 °C while providing a pleasing 67%
HPLC yield (entry 6).17
temp (°C)/
yield
b
entry
2
PhmAlCln
solvent
time (h)
product (%)
c
1
2
3
4
5
2c
2d
2d
2d
2d
Ph3Al
PhOMe
n-Bu2O
PhOMe
PhOMe
PhOMe
140/23
140/6
120/6
120/4
120/4
3b
3c
3c
3c
3c
12
64
62
d
Ph3Al
Ph3Al
ef
,
Ph2AlCl
Ph2.5AlCl0.5
7
eg
,
12
a
Reaction conditions: 2 equiv of PhmAlCln unless otherwise stated.
b
c
Isolated yield unless otherwise stated. HPLC showed that no more
3b was formed after about 12 h. 16% of 8a was isolated. See Henschke
et al.8a for experimental details. 11% of 8b was isolated. HPLC yield.
d
e
f
g
63% unreacted 2d and 0.39 equiv of BnOH with respect to 2d. 55%
unreacted 2d and 0.44 equiv of BnOH with respect to 2d.
arylglucoside 3b. In addition to 3b, 16% of a compound found
by mass spectrometry to have a molecular mass of 396 was
isolated by column chromatography. This product, charac-
terized by an unusual double arylation of the 1,6-anhydroglu-
cose backbone, was tentatively assigned the structure 8a
following 1D and 2D NMR spectroscopy and further
characterization as its ketone and 2,4-dinitrophenylhydrazone
derivatives.10a By contrast, treatment of the known11 tri-O-
benzyl-protected analogue 2d with Ph3Al (entry 2) provided
the desired product 3c in good isolated yield (64%) along with
a doubly arylated enol ether coproduct (11%) tentatively
assigned the structure 8b, partly by analogy to 8a, and benzyl
alcohol.10b Benzylation of 3c provided the known4a tetra-O-
benzyl derivative β-4 (Figure 2) that was shown by HPLC
comparison to an authentic mixture of α-4 and β-4 to comprise
a 0.4:99.6 mixture of the α- and β-anomers demonstrating that
the arylation was highly β-stereoselective. This is consistent
with the high stereoselectivity witnessed in the arylation of 2b
In an attempt to find more atom economic boronic ester
protecting groups, several dimers were targeted. While
condensation of 1,6-anhydroglucose with 1,4-benzene dibor-
onic acid failed to give the desired product, condensation of
tetrahydroxydiboron (entry 7) in dioxane under Dean−Stark
B
DOI: 10.1021/acs.joc.5b00601
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
a
a
Table 2. Arylation of 2,4-O-Boronic Esters 5
Table 3. Arylation of 10 Pretreated with Alanes
b
c
entry
R3Al
PhmAlCln
solvent
time (h) yield (%)
time
yield
b
b
d
e
1
Ph3Al
Ph3Al
dioxane
PhOMe
PhOMe
PhOMe
PhOMe
PhOMe
PhOMe
PhOMe
n-Bu2O
PhCl
66
14
22
23
38
43
38
48
61
19
6
73
43
47
50
32
38
55
65
62
52
39
24
47
46
36
58
74
74
74
72
65
72
53
entry
2
PhmAlCln
solvent
(h)
(%)
f
2
Ph3Al
Ph3Al
c
d
1
2
3
4
5
6
7
8
9
2e
1.5 equiv of Ph2AlCl
1 equiv of Ph3Al
3 equiv of Ph3Al
PhMe
20
2.5
6
34
3
Ph3Al
Ph3Al
e
df
,
2e
2e
2e
2e
2e
PhCN
55
63
67
71
67
53
26
65
b
4
Ph3Al
Ph3Al
d
d
C6H4Cl2
5
Me3Al
Ph3Al
2.4 equiv of Ph2.5AlCl0.5 PhCl
3.5
2.5
16
24
6
g
6
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
i-Bu2AlH
Et3Al
Ph3Al
g
2 equiv of Ph2.2AlCl0.8
PhOMe
h
7
Ph3Al
h
2.4 equiv of Ph2.5AlCl0.5 PhCN
I
8
Ph3Al
e
I
h
f
2f
3 equiv of Ph3Al
3 equiv of Ph3Al
3 equiv of Ph3Al
dioxane
9
Ph3Al
j
NA
2e
PhOMe
PhOMe
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Ph3Al
6
Ph3Al
Ph2O
a
Reaction conditions: internal temperature about 140 °C, unless
otherwise stated. HPLC yield calibrated to an internal standard unless
otherwise stated. Heated under reflux. Isolated yield following
column chromatography. 10 was esterified with the boronic acid in a
Dean−Stark apparatus and was used directly without isolation. No α-
anomer was detected by HPLC. Internal temperature 110 °C.
Internal temperature 80 °C. HPLC showed 1.2% of α-3d. 10 was
mixed with 1 equiv of PhB(OH)2 prior to arylation but the boronic
Ph3Al
PhCN
4
b
d
Ph3Al
PhMe
37
72
18
16
24
21
21
26
18
24
48
c
d
d
Ph3Al
dioxane
e
f
Ph1.5AlCl1.5
Ph2AlCl
Ph2.5AlCl0.5
Ph2.5AlCl0.5
Ph2.5AlCl0.5
Ph2.5AlCl0.5
Ph2.5AlCl0.5
Ph2.5AlCl0.5
Ph2.5AlCl0.5
PhOMe
PhOMe
PhOMe
n-Bu2O
PhCl
g
j
h
I
j
k
l
h
ester 2e was not preformed. HPLC showed 1.7% of α-3d.
m
Ph2O
PhOMe
PhOMe
PhOMe
conditions provided a mixture of products (1H NMR
spectroscopy) tentatively containing 2f and/or its isomer(s)
that upon arylation with Ph3Al furnished a 53% HPLC yield of
β-3d and 1.2% of α-3d.
i-Bu3Al
n
n-octyl3Al
a
Reaction conditions: 2 equiv of PhmAlCln unless otherwise stated.
Internal temperature 140 °C unless otherwise stated. 3 equiv. HPLC
yield calibrated to internal standard unless otherwise stated. Under
reflux. 48% isolated yield following workup and column chromatog-
raphy. 1 equiv. 100 °C internal temperature. 120 °C internal
temperature. 51% isolated yield following workup and column
chromatography. 4.0% of α-3d and 0.50% of 9 was detected by HPLC
analysis. Isolated yield by column chromatography was 65%. 4.4% of
α-3d and 0.47% of 9 was detected by HPLC analysis. 3.2% of α-3d
and 0.49% of 9 was detected by HPLC analysis. 4.0% of α-3d and
0.53% of 9 was detected by HPLC analysis. The reaction was still
ongoing at 48 h with 40% unreacted 10 detected by HPLC.
b
c
d
To demonstrate the importance of preforming the boronic
ester, direct arylation of a mixture of 1,6-anhydroglucose and 1
equiv of PhB(OH)2 (entry 8) with 3 equiv of Ph3Al was tested.
As compared to the same arylation with boronic ester 2e (entry
9) that provided a 65% HPLC yield of β-3d with no detectable
amounts of the α-anomer, a 26% yield of β-3d was obtained
along with 1.7% of α-anomer α-3d. That any 3d was formed in
this control experiment was, however, unexpected and
suggested that direct arylation of 1,6-anhydroglucose, without
recourse to conventional hydroxyl group protection, might be
possible.
e
f
g
h
I
j
k
l
m
n
Direct Arylation of 1,6-Anhydroglucose Pretreated
with i-Bu2AlH. Having identified above that the arylation of
1,6-anhydroglucose (10) could be accomplished without
preformation of the boronic ester the direct arylation of 10
without recourse to a protecting group installing step was
examined18 using Ph3Al. Pleasingly, heating 10 and 6 equiv of
Ph3Al (1.0 M in n-Bu2O) in dioxane (Table 3, entry 1)
produced over a 66 h period a 73% HPLC yield of 1-C-phenyl-
β-D-glucoside (β-3d), along with 1.8% of the α-anomer α-3d,
providing proof of concept. Following aqueous workup and
column chromatography a somewhat lower isolated yield
(48%) was obtained, owing to difficulties separating the water-
soluble β-3d from the excess of aluminum residues. The
reaction was more rapid in PhOMe, but the yields increased
only slightly as the amounts of the additional Ph3Al were
increased from 1 to 3 equiv (entries 2−4). To reduce the
amount of the arylalane, which would be desirable for the
synthesis of β-C-arylglucoside drugs such as 1 that possess
more complex aryl side chains, and in an analogous strategy to
that which we reported previously for 2a,8a 10 was pretreated
with 3 equiv of Me3Al (entry 5) instead of Ph3Al, followed by
arylation with 2 equiv of Ph3Al. This resulted in a disappointing
32% yield, however. Although 10 was not soluble in PhOMe,
when 3 equiv of i-Bu2AlH were used in the pretreatment step
homogeneous solutions were produced that provided improved
yields of β-3d, that increased with increasing temperature, in
the subsequent arylation step (entries 6−8). In fact a 65%
HPLC yield was achieved along with only trace amounts of the
isobutylated derivative 9 when the arylation was performed at
140 °C for 48 h (entry 8).19 Following column chromatography
a 51% isolated yield was obtained. When a suspension of 10 in
d8-THF was treated with 3 equiv of i-Bu2AlH in PhMe,
vigorous bubbling occurred followed by the generation of a
colorless, homogeneous solution. Analysis by 1H NMR
spectroscopy20a suggested that at least three components had
formed, confirming that in d8-THF, at least, modification of 10
with i-Bu2AlH does not produce a single component that is
directly analogous to 2,3,4-tri-O-protected 1,6-anhydroglucose
C
DOI: 10.1021/acs.joc.5b00601
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Scheme 2. Synthesis of Canagliflozin by the Arylation of 1,6-Anhydroglucose (10) Pretreated with i-Bu2AlH
compounds such as 2c or 2d. Solvent screening (entries 9−14)
showed that while n-Bu2O was comparable to PhOMe in terms
of yield of β-3d, PhCl, Ph2O, PhCN, PhMe, and dioxane
provided less satisfactory results.20b Consistent with the
phenylation of 2a pretreated with Me3Al,8a the rate of arylation
of 10 pretreated with i-Bu2AlH in PhOMe was significantly
increased (16−24 h) by the use of phenyl(chloro)alanes,
PhmAlCln (entries 15−17), prepared by mixing Ph3Al and
AlCl3. Ph2.5AlCl0.5 provided the best HPLC yield (entry 17;
74%) and gave a pleasing 65% isolated yield following careful
aqueous workup to reduce losses of the water-soluble product
and column chromatography. Entry 17 was repeated using n-
Bu2O (entry 18), PhCl (entry 19), or Ph2O (entry 20) as
solvent.20b HPLC yields were significantly improved (72−74%)
upon those using the same solvents with Ph3Al as the arylating
agent (cf. entries 9−11), although around 3−4% α-3d was
formed. The mass balance in these tests was partially accounted
for by unreacted 10 with HPLC typically showing between 10
and 25% remaining.
CONCLUSION
■
In conclusion, the β-selective phenylation of protected 1,6-
anhydroglucose 2 using triphenyl- (Ph3Al) or phenyl(chloro)-
alanes (PhmAlCln) has been demonstrated. The preparation and
use of boronate 2e as a substrate was more convenient than
TBS- and benzyl-protected analogues 2c and 2d, respectively,
because the entire process starting from 1,6-anhydro-β-D-
glucopyranose (10) and finishing at β-3d could be conducted
in one pot. Ultimately, however, it was discovered that 10 itself
could be directly arylated in the absence of conventional
protecting groups by heating 10 with an excess of Ph3Al or by
pretreatment with i-Bu2AlH, i-Bu3Al, Et3Al, Me3Al, or n-
octyl3Al, followed by heating with Ph3Al or PhmAlCln. In
comparison to the conventional method3 that starts with
commercial gluconolactone, which requires global hydroxy
group protection and several synthetic steps following arylation,
the method described herein is concise and should provide a
strong basis for further innovation in β-C-arylglucoside
synthesis. Before this can be fully realized, however, there
remains a need to reduce the amount of arylating agent
(typically 2 equiv with respect to 10) and to lower the reaction
temperature. In exploratory studies8b using 2a both Ti- and Ga-
based arylating agents showed promise, and perhaps these can
replace the Al-reagents reported herein. Finally, the potential of
this method for the preparation of drug molecules was
demonstrated by the synthesis of canagliflozin (1a) from 1,6-
anhydroglucose (10) without the isolation of any hydroxy
group-protected intermediates or product.
Using the entry 17 conditions, Et3Al (entry 21), i-Bu3Al
(entry 22), and n-octyl3Al (entry 23) were tested in the
pretreatment step of 10 in place of i-Bu2AlH.20b Whereas
arylation rates and yields using Et3Al and i-Bu3Al were similar
to those using i-Bu2AlH, pretreatment with n-octyl3Al resulted
in much slower arylation.
Synthesis of Canagliflozin from 1,6-Anhydroglucose
Pretreated with i-Bu2AlH. Following on from the results of
the above-described model studies, the synthesis of canagli-
flozin (1a) from 1,6-anhydroglucose (10) was conducted
(Scheme 2). The arylating agent Ar2AlCl (Ar = (3-[[5-(4-
fluorophenyl)-2-thienyl]methyl]-4-methylphenyl) was prepared
as described previously.8a A PhMe−i-Pr2O solution of 3-[[5-(4-
fluorophenyl)-2-thienyl]methyl]-4-methylphenyl bromide was
treated with n-BuLi in n-hexane at 0 °C followed by reaction
with AlCl3 in n-Bu2O at 90 °C. Following dilution with PhMe
and concentration to remove the i-Pr2O, the resulting solution
was filtered to remove the LiCl. An attempt to use this reagent
without filtration resulted in effectively no arylation occurring.
1,6-Anhydroglucose in PhOMe was treated with 3 equiv of i-
Bu2AlH prior to being heated at 140 °C together with 2 equiv
of the above prepared arylating agent in PhMe/n-Bu2O. After
18 h reaction 61% 1a and 33% unreacted 10 were witnessed by
calibrated HPLC (using internal standards); however, no
further 1a was formed after another 2 h, and following aqueous
workup and column chromatography a 50% isolated yield of
97.7%-pure canagliflozin was obtained.
EXPERIMENTAL SECTION
■
Materials and Methods. Experiments were conducted under
anhydrous conditions in a nitrogen atmosphere using Schlenk
techniques. Solvents were dried over 3 or 4 Å molecular sieves, and
oven-dried glassware and gastight syringes were used. High resolution
electrospray ionization (ESI) mass spectrometry was performed using
a QToF tandem mass analyzer. 1H and 13C NMR spectra were
recorded in the specified deuterated solvents at 400 and 100 MHz,
respectively. Commercially obtained or in-house prepared solutions of
organometallic reagents were titrated using standard titration methods,
including Knochel’s method.21 Commercial 1,6-anhydro-β-D-glucopyr-
anose (10), i-Bu2AlH (DIBAL; 1.0 M in PhMe), Ph3Al (1.0 M in n-
Bu2O), Et3Al (1.0 M in n-hexane), i-Bu3Al (1.0 M in n-hexane), and n-
octyl3Al (25% in hexanes), and AlCl3 (0.5 M in THF; titrated by
Eriochrome cyanine R spectrophotometric method22) were used as
supplied. A reference sample of canagliflozin (1a; see spectra below)
was acquired from a commercial source. Authentic samples of α-3d
and β-3d were acquired as described in the Supporting Information
section of ref 8a. Spectra of the known4a compounds 2,3,4,6-tetra-O-
benzyl-1-C-phenyl-α-D-glucopyranoside (α-4) and 2,3,4,6-tetra-O-
benzyl-1-C-phenyl-β-D-glucopyranoside (β-4) were presented in the
D
DOI: 10.1021/acs.joc.5b00601
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Supporting Information section of ref 8a. 1,6-Anhydro-2,3,4-tri-O-
LCMS (ESI) m/z 397 (100, [M + H]+), 398 (22, [M + H + 1]+), 419
(91, [M + Na]+).
benzyl-β-D-glucopyranose (2d) was prepared as reported by Furstner
̈
et al.23 Determination of the stereoselectivity in the arylation of 2d
(see Table 1) with Ph3Al providing 2,3,4-tri-O-benzyl-1-C-phenyl-
glucopyranoside (3c) was achieved by benzylation of the crude 3c (see
below) to give crude 2,3,4,6-tetra-O-benzyl derivative 4. Comparison
of the crude 4 by HPLC analysis (see Figure S5 in the Supporting
Information) to reference samples of known4a 2,3,4,6-tetra-O-benzyl
derivatives α-4 and β-4 (prepared as described in the Supporting
Information of ref 8a) was used to identify the peaks. HPLC analysis
was conducted on an Agilent ZORBAX SB-Phenyl column (5 μm, 4.6
× 250 mm) at 40 °C, monitoring at 220 nm, eluting at a flow rate of
1.0 mL/min with a linear gradient of 50% water/50% MeOH to 20%
water/80% MeOH from 0 to 5 min followed by a linear gradient of
20% water/80% MeOH to 100% MeOH from 5 to 30 min; α-4 eluted
at 21.6 min, while β-4 eluted at 22.6 min (see chromatograms in
Supporting Information). Determination of the stereoselectivity and
quantitation (yield) of 3d and level of 9 formed in the arylation of 2e,
2f, or 10 for Table 2 and Table 3 and reaction monitoring for the
synthesis of canagliflozin was achieved by HPLC analysis using an
Agilent ZORBAX SB-Aq column (3.5 μm, 4.6 × 150 mm) at 35 °C
that was isocratically eluted for 5 min using 100% water followed by a
linear elution gradient from 0% MeCN (containing 0.01% TFA)/
100% water to 95% MeCN (containing 0.01% TFA)/5% water for 27
min using a flow rate of 1 mL/min. Both a UV detector (monitoring at
210 nm; see Figure S6 in the Supporting Information) and ELSD
(Evaporative Light-Scattering Detector; monitoring at 50 °C, 1.8 L/
min gas flow; see Figure S7 in the Supporting Information) were used
in series to cooperatively monitor the HPLC runs. Analysis samples
were prepared by mixing reaction aliquots with 5% TFA in 1:2 water/
MeCN (1 wt % cinchonine was added when the levels of 10 and β-3d
were to be determined by ELSD). The UV detector was used to
monitor and quantify UV-active reaction components including α-3d
(tR 9.4 min), β-3d (tR 8.1 min), and canagliflozin (1a; tR 21.5 min).
The levels of the reaction components were determined with respect
to an internal standard (4-pentylbiphenyl; tR 28.85 min) that was
added (10−20 mol % with respect to 1,6-anhydro-β-D-glucopyranose
(10)) to the reaction prior to initiation. The following calibrated
relative response factors (on a mole: mole basis with respect to 4-
pentylbiphenyl) were used: β-3d 0.23; α-3d 0.23. To monitor and
determine the levels of UV-inactive compounds 10 (tR 2.2 min) and 9
(tR 11.0 min) the ELSD was used. α-3d (tR 9.5 min) and β-3d (tR 8.0
min) could also be detected using the ELSD. The levels of the reaction
components were determined with respect to an external standard
(cinchonine; tR 15.71 min) that was preadded to the sample solvent
(5% TFA in 1:2 water/MeCN) and the following calibrated relative
response factors (on a mole: mole basis with respect to cinchonine):
β-3d 0.43; 10 0.41 were used. Compound 9 was determined by area%
with respect to β-3d. When the ELSD detector was conducted in series
following the UV detector, the tR of the peaks ELSD chromatogram
showed a delay of ca. 0.1 min as compared to when detected using the
UV detector. The peaks corresponding to β-3d and α-3d were
determined by analysis of authentic samples of β-3d and α-3d that
were prepared as reported in the Supporting Information of ref 8a.
Arylation of 2c with Ph3Al for Table 1, Entry 1 and Isolation of
Side Product 8a. In addition to the column chromatographic isolation
of 2,3,4-tri-O-tert-butyldimethylsilyl-1-C-phenyl-β-D-glucopyranoside
(3b; 69 mg, 12%) from the arylation of 1,6-anhydro-2,3,4-tri-O-tert-
butyldimethylsilyl-β-D-glucopyranose (2c; 0.51 g, 1.0 mmol) reported
in the Experimental Section of ref 8a, 63 mg (16%) of a reaction side
product that herein is proposed (see Supporting Information for
support of this structure) to be (4R*)-2-O-tert-butyldimethylsilyl-3-
deoxy-2,3-dehydro-4-phenyl-1-C-phenyl-β-D-glucopyranoside (8a) was
also isolated by column chromatography. 1H NMR (400 MHz,
CDCl3) δ 7.45−742 (m, 2H), 7.40−7.33 (m, 5H), 7.33−7.28 (m,
3H), 5.13 (dd, J = 3.2, 1.6 Hz, 1H), 5.01 (dd, J = 1.6, 1.6 Hz, 1H), 3.79
(ddd, J = 9.8, 3.0, 1.8 Hz, 1H), 3.69 (ddd, J = 9.5, 5.5, 3.5 Hz, 1H),
3.57 (br, 2H), 0.70 (s, 9H), 0.16 (s, 3H), −0.04 (s, 3H); 13C NMR
(100 MHz, CDCl3) δ 150.7, 141.8, 139.6, 128.7, 128.6, 128.4, 128.2,
128.1, 127.0, 105.4, 81.0, 79.4, 62.8, 42.8, 25.3, 17.7, −4.6, −5.0;
General Procedure for the Synthesis of 2,3,4-Tri-O-benzyl-1-C-
phenyl-β-D-glucopyranoside (β-3c) by the Arylation of 2d with
Arylalanes PhmAlCln for Table 1, Entries 2−5. These compounds
were prepared by heating an n-Bu2O (4.0 mL; Table 1, entry 2) or
PhOMe (4.0 mL; Table 1, entries 3, 4, or 5) solution of 1,6-anhydro-
2,3,4-tri-O-benzyl-β-D-glucopyranose (2d; 0.43 g, 1.0 mmol) and
Ph3Al (2.0 mL, 2.0 mmol, 1.0 M in n-Bu2O), or Ph2AlCl (2.0 mmol; a
mixture of Ph3Al (1.3 mL, 1.3 mmol, 1.0 M in n-Bu2O) and AlCl3 (1.3
mL, 0.67 mmol, 0.5 M in THF)), or Ph2.5AlCl0.5 (2.0 mmol; Ph3Al
(1.67 mL, 1.67 mmol, 1.0 M in n-Bu2O) and AlCl3 (0.67 mL, 0.33
mmol, 0.5 M in THF)) at 140 or 120 °C for 4−6 h. After cooling to
ambient temperature, THF (10 mL), then diatomaceous earth (1 g),
then 15% aqueous NaOH (1 mL), and then Na2SO4 (2 g) were added
sequentially to the product mixture, and the resulting suspension was
stirred and then filtered. The filtrate was concentrated to give a yellow
1
oil (HPLC, LCMS, and H NMR spectroscopic analysis showed the
crude product to be mostly composed of β-3c, 8b, BnOH, and
biphenyl) which was purified by silica gel column chromatography
(eluting with 1:20 EtOAc/n-heptane) affording the product 2,3,4-tri-
O-benzyl-1-C-phenyl-β-D-glucopyranoside (β-3c; entry 2:0.32 g, 64%
or entry 3:0.31 g, 62%) as a white solid. 1H NMR (400 MHz, CDCl3)
δ 7.48−7.31 (m, 15H), 7.24−7.19 (m, 3H), 6.95−6.92 (m, 2H), 5.00
(d, J = 11.2 Hz, 1H), 4.95 (d, J = 10.8 Hz, 1H), 4.94 (d, J = 11.2 Hz,
1H), 4.74 (d, J = 10.8 Hz, 1H), 4.41 (d, J = 10.0 Hz, 1H), 4.31 (d, J =
9.6 Hz, 1H), 3.93 (ddd, J = 11.8, 6.1, 2.6 Hz, 1H), 3.87 (dd, J = 9.0, 9.0
Hz, 1H), 3.81−3.70 (m, 3H), 3.59−3.53 (m, 2H), 1.97 (dd, J = 6.8,
6.8, 1H); 13C NMR (100 MHz, CDCl3) δ 139.0, 138.6, 138.0, 137.6,
128.54, 128.52, 128.47, 128.26, 128.23, 128.1, 128.0, 127.744, 127.735,
127.69, 127.67, 86.6, 84.3, 81.7, 79.4, 78.3, 75.7, 75.2, 74.9, 62.4; FTIR
(neat) 3447, 3088, 3063, 3031, 2871, 1497, 1454, 1398, 1359, 1211,
1151, 1067, 1028, 751, 736, 697 cm−1; [α]D20 = −6.6 (c 1.0, MeOH);
LCMS (ESI) m/z 528 (100, [M + NH4]+), 529 (35, [M + NH4 + 1]+),
533 (5, [M + Na]+); ESI QTof calculated for [C33H34O5 + NH4]+ =
528.2745, found 528.2771; mp 111.1 °C. Additionally, a side product
tentatively assigned the structure 8b was also isolated by column
1
chromatography (46 mg (contaminated with 65 mol % (by H NMR
with respect to 8b) BnOH which is equivalent to 7.2 mg), equivalent
to 38.8 mg of 8b, 11%). The proposed structure for 8b was supported
by LCMS and 1H and 13C NMR analysis (see the Supporting
1
Information) and by analogy to 8a (see above). H NMR (400 MHz,
CDCl3) δ 7.1−7.5 (m, 15H), 5.26 (d, J = 1.2 Hz, 1H), 4.88 (s, 1H),
4.80 (d, J = 12.8 Hz, 1H), 4.76 (d, J = 12.8 Hz, 1H), 4.61 (t, J = 5.8
Hz, 1H), 3.66 (m, 1H), 3.52 (m, 1H), 3.34 (m, 2H); 13C NMR (100
MHz, CDCl3) δ 153.8, 143.1, 140.3, 137.6, 129.0, 128.9, 128.8, 128.7,
128.5, 128.4, 127.9, 127.1, 126.9, 100.1, 82.3, 77.8, 68.6, 63.4, 61.7,
42.2; LCMS (ESI) m/z 355 (13, [M + H − H2O]+), 373 (57, [M +
H]+), 390 (100, [M + NH4]+), 395 (44, [M + Na]+). Further
confirmation of the structure of β-3c was obtained by 1) conversion of
β-3c to its known per-benzyl derivative, 2,3,4,6-tetra-O-benzyl-1-C-
phenyl-β-D-glucopyranoside (β-4) (benzylation procedure: crude 3c
(18 mg, 0.03 mmol) in THF (2 mL) was treated with NaH (60%
dispersion in mineral oil; 1.8 mg, 0.045 mmol) for 1 h at room
temperature, then benzyl bromide (10 μL, 0.08 mmol) was added, and
the mixture was stirred at 50 °C overnight (TLC shows complete
conversion) before being diluted with EtOAc, extracted with water,
and washed with brine giving crude β-4), which conformed with that
reported by Ellsworth et al.4a (see spectra of an authentic sample of β-
4 in the Supporting Information section of ref 8a), and 2)
hydrogenolysis of β-3c affording 1-C-phenyl-β-D-glucopyranoside (β-
3d) as confirmed by 1H NMR spectroscopy (spectroscopic
comparison of this material to an authentic sample of β-3d is shown
in the Supporting Information).
Synthesis of 1,6-anhydro-β-D-glucopyranose 2,4-O-phenylboronate
(2e; see spectra below) was prepared as reported by Spring et al.13b
and following filtration was washed with PhMe and was then used
without purification, whereas in some cases crude boronic ester 2e was
used directly in arylation reactions without isolation from its PhMe
1
solution. H NMR (400 MHz, CDCl3) δ 7.83−7.87 (m, 2H), 7.46−
E
DOI: 10.1021/acs.joc.5b00601
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
7.51 (m, 1H), 7.37−7.42 (m, 2H), 5.65 (t, J = 2.4 Hz, 1H), 4.63−4.67
(m, 1H), 4.58 (d, J = 8.0 Hz, 1H), 4.19−4.22 (m, 1H), 4.12−4.16 (m,
1H), 4.08−4.10 (m, 1H), 3.94 (dd, J = 7.6 Hz, 4.8 Hz, 1H), 3.44 (d, J
= 8.8 Hz, 1H); 13C NMR (100 MHz, CDCl3) δ 134.6, 131.5, 128.0,
101.9, 76.8, 70.6, 70.5, 69.3, 66.5 (the ipso carbon (attached to B)
signal was not observed due to low intensity). Also see the spectra in
the Supporting Information. Boronic ester 2e is water sensitive.
Addition of 1 drop of water to 2e in d6-DMSO in an NMR tube
showed complete hydrolysis back to 1,6-anhydro-β-D-glucopyranose
(10) and PhB(OH)2 (see the Supporting Information).
Attempted synthesis of the 2,4,2′,4′-O-diboron dimer 2f of 1,6-
anhydro-β-D-glucopyranose was conducted by heating a mixture of
1,6-anhydro-β-D-glucopyranose (10; 324 mg, 2 mmol) and tetrahy-
droxydiboron (90 mg, 1 mmol) in dioxane (40 mL) under reflux in a
Dean−Stark apparatus (with molecular sieves installed in the side arm)
for 15 h. 1H NMR analysis (see 1H NMR spectrum in the Supporting
Information) of an aliquot of the product mixture that was evaporated
to dryness indicated that it was composed of a mixture of products
tentatively containing 2f.
an HPLC yield was required; see Figures S1, S2, and S3 in the
Supporting Information that show reaction plots for Table 3, entries
8−14, entries 18−20, entries 21−23, respectively) in PhOMe (5 mL)
was added i-Bu2AlH (3.7 mL, 3.7 mmol, 1.0 M in PhMe) over a 2 min
period at ambient temperature. Vigorous evolution of gas was
observed, and a homogeneous solution was formed. After having
been stirred for 5 min (gas formation had ceased) Ph3Al (2.1 mL, 2.1
mmol, 1.0 M in n-Bu2O) and AlCl3 (0.81 mL, 0.41 mmol, 0.5 M in
THF) were added to the colorless solution resulting in the formation
of a dark green to black solution. The solution was heated at 140 °C
for 24 h at which time HPLC assay analysis indicated a 74% yield
based on the internal standard (4-pentylbiphenyl). After cooling to
ambient temperature, THF (100 mL), diatomaceous earth (10 g), and
10% aq. NaOH (3 mL) were sequentially added to the product
mixture, and the resulting suspension was filtered. The filtrate was
concentrated to give a yellow oil that was purified by silica gel column
chromatography (eluting with 1:20 to 1:10 MeOH/DCM) to give β-
3d (198 mg, 65%).
Reference Sample of 1-C-Isobutyl-D-glucopyranoside (9), a Minor
Side Product in the Arylation of 10 Treated with i-Bu2AlH. To a
suspension of 1,6-anhydro-β-D-glucopyranose (10; 324 mg, 2.00
mmol) in PhOMe (5 mL) in an oven-dried, two-neck flask was added
over a 2 min period i-Bu3Al (10 mL, 10 mmol, 1.0 M in hexanes).
After having been stirred at ambient temperature for 5 min the
colorless and clear solution was heated at 140 °C for 19 h. After the
solution cooled to ambient temperature, THF (20 mL), diatomaceous
earth (2 g), and 15% aq. NaOH (2 mL) were added sequentially, and
the resulting suspension was stirred at ambient temperature for 2 h.
Anhydrous Na2SO4 (2 g) was added, and the mixture was stirred for
30 min and was then filtered. The filtrate was concentrated and
purified by silica gel column chromatography (eluting with 1:40 to 1:5
MeOH−DCM) to give 9 (30 mg, 7%). LCMS (ESI) m/z 238 (100,
[M + NH4]+), 458 (28, [2 M + NH4]+). Although the configuration of
C1 was not determined, it was presumed that compound 9 was 1-C-
isobutyl-β-D-glucopyranoside (β-9) (see 1H NMR spectrum in the
Supporting Information).
Synthesis of Canagliflozin (1a) by the Arylation of i-Bu2AlH-
Pretreated 10. To 1,6-anhydro-β-D-glucopyranose (10; 124 mg, 0.76
mmol) and 4-pentylbiphenyl (51 mg; used as an internal standard so
that the reaction could be monitored by HPLC assay) in PhOMe (4
mL) was added i-Bu2AlH (2.29 mL, 2.29 mmol, 1.0 M in PhMe) over
a 2 min period at ambient temperature. After having been stirred for 5
min (gas formation had ceased) the colorless, homogeneous solution
was added by syringe into a solution of bis-(3-[[5-(4-fluorophenyl)-2-
thienyl]methyl]-4-methylphenyl)chloroalane in n-Bu2O/PhMe (1.52
mmol; prepared exactly as described in ref 8a). The flask was washed
with PhOMe (1 mL), and this was also added into the reaction
mixture. The mixture was then heated at 140 °C for 20 h (see the
reaction monitoring plot (Figure S4) in the Supporting Information;
showed 61% 1a at 18 h by UV analysis, calibrated to the internal
standard, 4-pentylbiphenyl, and 33% 10 (upon ELSD analysis,
calibrated to the internal standard, cinchonine)), then the mixture
was cooled to ambient temperature, diluted with THF (15 mL), and
cooled to 0 °C. Celite 545 (2.0 g) and 10% aq. NaOH (3.0 mL) were
added slowly, and the suspension was stirred at ambient temperature
for 1 h. Anhydrous Na2SO4 (2.0 g) was added, and the mixture was
filtered through a pad of Celite 545. The flask and the Celite pad were
washed with THF (10 mL) giving a yellow clear filtrate. The Celite
was slurried with MeOH (10 mL) for 0.5 h and was filtered again. The
filtrates were combined and concentrated to afford an oil which was
chromatographed over a column of silica gel eluting with MeOH/
DCM (1:10) to give 97.7% HPLC-pure (see the UV and ELSD
chromatograms in the Supporting Information) canagliflozin (1a) as a
white power (170 mg, 50%). 1H and 13C NMR spectra (see the
Supporting Information) conformed to that reported in the
Synthesis of 1-C-phenyl-β-D-glucopyranoside (β-3d) by the arylation of
2e and 2f with arylalanes PhmAlCln for Table 2 and characterization of
side product 7. To a solution of 1,6-anhydro-β-D-glucopyranose 2,4-O-
phenylboronate (2e; 248 mg, 1.0 mmol; alternatively, the crude 2e
product mixture was directly used in arylations (e.g., entry 2) without
isolation of 2e) in PhOMe (5 mL), or another solvent reported in
Table 2, at ambient temperature was added Ph3Al (3.0 mL, 3.0 mmol,
1.0 M in Bu2O), or Ph2AlCl (for entry 1: a mixture of Ph3Al (1.0 mL,
1.0 mmol, 1.0 M in n-Bu2O) and AlCl3 (1.0 mL, 0.50 mmol, 0.5 M in
THF) was used), or Ph2.5AlCl0.5 (for entries 4 and 6: a mixture of
Ph3Al (2.0 mL, 2.0 mmol, 1.0 M in n-Bu2O) and AlCl3 (0.8 mL, 0.4
mmol, 0.5 M in THF or neat AlCl3)), or Ph2.2AlCl0.8 (for entry 5: a
mixture of Ph3Al (1.5 mL, 1.50 mmol, 1.0 M in n-Bu2O) and AlCl3
(1.1 mL, 0.55 mmol, 0.5 M in THF)). The mixture was heated at 140
°C (internal temperature), or as otherwise stated in Table 2, for 2.5−
24 h. The product mixture was quenched by treatment with MeOH (3
mL) or 5% TFA in MeCN (3 mL) followed by concentration under
reduced pressure to remove the volatiles followed by chromatographic
purification of the residue over silica gel (eluting with 1:10 MeOH/
DCM) providing β-3d. The spectroscopic data for β-3d conformed
with that reported in ref 8a and is also presented in the Supporting
Information. In addition to β-3d and unreacted starting material that
was recovered as 1,6-anhydro-β-D-glucopyranose (10), a polar side
product was also isolated by column chromatography that was
tentatively characterized as 1,6-anhydro-3-O-β-D-glucopyranosyl-β-D-
glucopyranose (7) using a combination of LCMS analysis and 1H, 13C,
DEPT-135, HMQC, COSY, HMBC (a correlation between H1′ and
C3 supports the proposed (1 → 3)-disaccharide linkage), and NOESY
1
NMR experiments (see spectra in the Supporting Information). H
NMR (400 MHz, d4-MeOH) δ 5.41 (1H, s), 4.59 (1H, m), 4.40 (1H,
d, J = 7.6 Hz), 4.23 (1H, d, J = 7.2 Hz), 3.94 (1H, m), 3.94 (1H, m),
3.73 (1H, m), 3.73 (1H, m), 3.66 (1H, m), 3.66 (1H, m), 3.40 (1H,
m), 3.33 (1H, m), 3.29 (1H, m), 3.21 (1H, t, J = 8.4 Hz); 13C NMR
(100 MHz, d4-MeOH) δ 103.9, 101.6, 80.4, 76.4, 76.8, 76.2, 73.6, 70.2,
69.1, 68.1, 64.8, 61.4; LCMS (ESI) m/z 342 (100, [M + NH4]+), 347
(15, [M + Na]+), 363 (10, [M + K]+).
Synthesis of 1-C-Phenyl-β-D-Glucopyranoside (β-3d) by the
Arylation of i-Bu2AlH-Pretreated 10 with Arylalanes PhmAlCln for
Table 3. The following example from Table 3, entry 17 is
representative of the general method used for the other entries in
Table 3 with appropriate changes to the solvent used and the arylating
agent (for Ph2AlCl, Ph3Al (1.6 mL, 1.6 mmol, 1.0 M in n-Bu2O) and
AlCl3 (1.7 mL, 0.85 mmol, 0.5 M in THF) were used, and for
Ph1.5AlCl1.5, Ph3Al (1.2 mL, 1.2 mmol, 1.0 M in n-Bu2O) and AlCl3
(2.5 mL, 1.3 mmol, 0.5 M in THF) were used). Reactions were
terminated when HPLC analysis showed no further formation of β-3d.
For entries 21−23, commercially acquired 1.0 M Et3Al in n-hexane, 1.0
M i-Bu3Al in n-hexane, and 25% n-octyl3Al in hexanes were used,
respectively, in place of 1.0 M i-Bu2AlH in PhMe. To a suspension of
1,6-anhydro-β-D-glucopyranose (10; 204 mg, 1.26 mmol) and 4-
pentylbiphenyl (47 mg, 0.21 mmol; used as an internal standard when
Supporting Information of ref 6 in ref 8a and were identical to
25
those of a commercially acquired sample of canagliflozin. [α]D 25
=
=
+15.3 (c 1.0, MeOH) (the commercial canagliflozin had an [α]D
+16.1 (c 1.0, MeOH); HRMS ESI QTof calculated for [C24H25FO5S +
F
DOI: 10.1021/acs.joc.5b00601
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
NH4]+ = 462.1750, found 462.1745; mp 99.2 °C (the commercial
canagliflozin had a mp of 100.6 °C).
(8) (a) Henschke, J. P.; Wu, P.-Y.; Lin, C.-W.; Chen, S.-F.; Chiang,
P.-C.; Hsiao, C.-N. J. Org. Chem. 2015, 80, 2295−2309. (b) Henschke,
J. P.; Lin, C.-W.; Wu, P.-Y.; Hsiao, C.-N.; Liao, J.-H.; Hsiao, T.-Y. U.S.
Patent 8,952,139 B2, 2015.
(9) (a) Singh, I.; Seitz, O. Org. Lett. 2006, 8, 4319−4322. (b) Allwein,
S. P.; Cox, J. M.; Howard, B. E.; Johnson, H. W. B.; Rainier, J. D.
Tetrahedron 2002, 58, 1997−2009. (c) Alzeer, J.; Cai, C.; Vasella, A.
Helv. Chim. Acta 1995, 78, 242−264. (d) Vasella, A. Pure Appl. Chem.
1998, 70, 425−430. (e) Bohner, T. V.; Beaudegnies, R.; Vasella, A.
Helv. Chim. Acta 1999, 82, 143−160. (f) Stichler-Bonaparte, J.; Bernet,
B.; Vasella, A. Helv. Chim. Acta 2002, 85, 2235−2257.
(10) Formal elimination of the C3−O group was also seen in the
arylation of 2a (see ref 8a), but this was not accompanied by arylation
of the sugar ring. (a) See the Supporting Information for experimental
details and spectra of 8a as well as spectra of its ketone derivative,
(4R*)-2-keto-3-deoxy-4-phenyl-1-C-phenyl-β-D-glucopyranoside
(8aa), and its hydrazone derivative, (4R*)-2,3-dideoxy-4-phenyl-1-C-
phenyl-β-D-glucopyranoside 2-(2,4-dinitrophenyl)hydrazone (8ab).
(b) See the Supporting Information for experimental details and
spectra of 8b.
ASSOCIATED CONTENT
* Supporting Information
■
S
1D and 2D NMR spectra of 10, 10 in d8-THF pretreated with i-
Bu2AlH, 8a, 8aa, 8ab, 2d, crude and isolated β-3c, 8b, β-3d,
crude 2e, 2e treated with water, tentative 2f, 9, tentative 7, 1a;
Figures S1, S2, S3, and S4 provide reaction plots for Table 3,
entries 8−14, entries 18−20, entries 21−23, and the synthesis
of canagliflozin, respectively. Figures S5, S6, and S7 show
HPLC chromatograms for the arylation of 2d and the arylation
of 10 pretreated i-Bu2AlH. The Supporting Information is
available free of charge on the ACS Publications website at
DOI: 10.1021/acs.joc.5b00601.
AUTHOR INFORMATION
Corresponding Author
*E-mail: julian.henschke@scinopharm.com.
■
(11) (a) Bourke, D. G.; Collins, D. J.; Hibberd, A. I.; McLeod, M. D.
Aust. J. Chem. 1996, 49, 425−434. (b) Furstner, A.; Albert, M.;
̈
Notes
Mlynarski, J.; Matheu, M.; DeClercq, E. J. Am. Chem. Soc. 2003, 125,
13132−13142.
The authors declare no competing financial interest.
(12) Organoalanes can exist in solution as equilibrium mixtures (see:
Zhou, S.; Wu, K.-H.; Chen, C.-A.; Gau, H.-M. J. Org. Chem. 2009, 74,
3500−3505 and Zhou, S.; Chuang, D.-W.; Chang, S.-J.; Gau, H.-M.
Tetrahedron: Asymmetry 2009, 20, 1407−1412 ). Fractional
stoichiometry represents a mixture of species; for example, Ph2.5AlCl0.5
represents a mixture of Ph3Al and Ph2AlCl but may include other
permutations that might exist in equilibrium.
ACKNOWLEDGMENTS
■
We thank Drs. C.-N. Hsiao and T.-Y. Hsiao for discussions
during the course of this project, Y.-H. Chen, H.-C. Su, and T.-
Y. Wang for NMR and mass spectroscopy support, and M.-F.
Ho and C.-C. Hung for supporting experiments.
(13) (a) Shafizadeh, F.; McGinnis, G. D.; Chin, P. S. Carbohydr. Res.
1971, 18, 357−361. (b) Su, X.; Thomas, G. L.; Galloway, W. R. J. D.;
Surry, D. S.; Spandl, R. J.; Spring, D. R. Synthesis 2009, 2009, 3880−
3896.
(14) Structure supported by LCMS and 1D and 2D NMR
spectroscopy (see the Supporting Information).
(15) In a control experiment, heating 2e in PhCN under reflux did
not produce β-3d; however, after 3 days an ca. 1:1 mixture of 10 and
disaccharide 7 were detected by TLC analysis.
(16) This was further confirmed by analysis of an extracted-ion
chromatogram (XIC) constructed from the LCMS data; no ions
corresponding to an isomer of β-3d were seen at the retention time
that the α-anomer was known to elute.
(17) PhOMe, n-Bu2O, PhMe, 1,4-dioxane, PhCN, and even MeCN
could be used as reaction solvent; however, PhOMe and PhCN
appeared better media to conduct the reaction in.
(18) A quantitative HPLC method using ELSD and UV detection in
series was developed so that the consumption of UV-inactive 10 and
the formation of UV-active β-3d could be monitored throughout the
reaction.
(19) After aqueous workup to separate β-3d from the aluminum
residues the 1H NMR spectrum (see the Supporting Information)
showed essentially only β-3d, PhOMe, n-Bu2O, and sodium
phenoxide.
(20) See (a) spectra or, (b) reaction plots, in the Supporting
Information.
REFERENCES
■
(1) (a) Sheridan, C. Nat. Biotechnol. 2012, 30, 899−900. (b) Hard-
man, T. C.; Dubrey, S. W. Diabetes Ther. 2011, 2, 133−145. (c) Bays,
H. Curr. Med. Res. Opin. 2009, 25, 671−681. (d) Washburn, W. N.
Expert Opin. Ther. Pat. 2009, 19, 1485−1499. (e) Washburn, W. N. J.
Med. Chem. 2009, 52, 1785−1794.
(2) (a) Nomura, S.; Kawanishi, E.; Ueta, K. U.S. Patent 7,943,788,
May 17, 2011. (b) Meng, W.; Ellsworth, B. A.; Nirschl, A. A.; McCann,
P. J.; Patel, M.; Girotra, R. N.; Wu, G.; Sher, P. M.; Morrison, E. P.;
Biller, S. A.; Zahler, R.; Deshpande, P. P.; Pullockaran, A.; Hagan, D.
L.; Morgan, N.; Taylor, J. R.; Obermeier, M. T.; Humphreys, W. G.;
Khanna, A.; Discenza, L.; Robertson, J. G.; Wang, A.; Han, S.;
Wetterau, J. R.; Janovitz, E. B.; Flint, O. P.; Whaley, J. M.; Washburn,
W. N. J. Med. Chem. 2008, 51, 1145−1149. (c) Kadokura, T.; Zhang,
W.; Krauwinkel, W.; Leeflang, S.; Keirns, J.; Taniuchi, Y.; Nakajo, I.;
Smulders, R. Clin. Pharmacokinet. 2014, 53, 975−988. (d) Grempler,
R.; Thomas, L.; Eckhardt, M.; Himmelsbach, F.; Sauer, A.; Sharp, D.
E.; Bakker, R. A.; Mark, M.; Klein, T.; Eickelmann, P. Diabetes, Obes.
Metab. 2012, 14, 83−90.
(3) (a) Kraus, G. A.; Molina, M. T. J. Org. Chem. 1988, 53, 752−753.
(b) Czernecki, S.; Ville, G. J. Org. Chem. 1989, 54, 610−612.
(4) (a) Ellsworth, B. A.; Doyle, A. G.; Patel, M.; Caceres-Cortes, J.;
Meng, W.; Deshpande, P. P.; Pullockaran, A.; Washburn, W. N.
Tetrahedron: Asymmetry 2003, 14, 3243−3247. (b) Deshpande, P. P.;
Singh, J.; Pullockaran, A.; Kissick, T.; Ellsworth, B. A. J. Org. Chem.
2007, 72, 9746−9749.
(21) Krasovskiy, A.; Knochel, P. Synthesis 2006, 890−891.
(22) See: Standard Methods for the Examination of Water and
Wastewater. http://www.standardmethods.org/store/ProductView.
cfm?ProductID=492 (accessed Apr 13, 2015).
(5) Deshpande, P. P.; Singh, J.; Pullockaran, A.; Kissick, T.;
Ellsworth, B. A.; Gougoutas, J. Z.; Dimarco, J.; Fakes, M.; Reyes,
M.; Lai, C.; Lobinger, H.; Denzel, T.; Ermann, P.; Crispino, G.;
Randazzo, M.; Gao, Z.; Randazzo, R.; Lindrud, M.; Rosso, V.; Buono,
F.; Doubleday, W. W.; Leung, S.; Richberg, P.; Hughes, D.; Washburn,
W. N.; Meng, W.; Volk, K. J.; Mueller, R. H. Org. Process Res. Dev.
2012, 16, 577−585.
(23) (a) Furstner, A.; Albert, M.; Mlynarski, J.; Matheu, M.;
̈
DeClercq, E. J. Am. Chem. Soc. 2003, 125, 13132−13142. Also see:
Bourke, D. G.; Collins, D. J.; Hibberd, A. I.; McLeod, M. D. Aust. J.
Chem. 1996, 49, 425−434.
(6) Lemaire, S.; Houpis, I. N.; Xiao, T.; Li, J.; Digard, E.; Gozlan, C.;
̀
Liu, R.; Gavryushin, A.; Diene, C.; Wang, Y.; Farina, V.; Knochel, P.
Org. Lett. 2012, 14, 1480−1483.
(7) Sakamaki, S.; Kawanishi, E.; Nomura, S.; Ishikawa, T. Tetrahedron
2012, 68, 5744−5753.
G
DOI: 10.1021/acs.joc.5b00601
J. Org. Chem. XXXX, XXX, XXX−XXX