A practical stereoselective synthesis and novel cocrystallizations of an amphiphatic SGLT-2 inhibitor

Article abstract of DOI:10.1021/op200306q

A practical synthesis of the SGLT-2 inhibitor β-C-aryl-d-glucoside (1) has been developed. The route employed 2,3,4,6-tetra-O-trimethlysilyl-d- glucano-1,5-lactone as the key chiral building block, prepared efficiently from the commercially available, inexpensive raw materials, d-gluconolactone and trimethylsilyl chloride. The salient step in the synthesis is the Lewis acid-mediated stereoselective reduction of a methyl C-aryl peracetylated glycoside using a silyl hydride to set the stereochemistry of the crucial anomeric chiral center. Several novel cocrystalline complexes of 1 with l-phenylalanine and l-proline were discovered. Single-crystal structures of these complexes and several synthetic intermediates have been determined. The l-phenylalanine complex was developed and used to purify and isolate the API. All steps were implemented at multikilogram scale.

Full text of DOI:10.1021/op200306q

Article
pubs.acs.org/OPRD
A Practical Stereoselective Synthesis and Novel Cocrystallizations of
an Amphiphatic SGLT-2 Inhibitor
Prashant P. Deshpande,*,† Janak Singh,‡ Annie Pullockaran,‡ Thomas Kissick,‡ Bruce A. Ellsworth,§
Jack Z. Gougoutas,∥ John Dimarco,∥ Michael Fakes,¶ Mayra Reyes,† Chiajen Lai,† Hidegard Lobinger,⊥
Theo Denzel,⊥ Peter Ermann,⊥ Gerard Crispino,‡ Michael Randazzo,† Zenrong Gao,† Renee Randazzo,†
Mark Lindrud,† Victor Rosso,† Frederic Buono,† Wendel W. Doubleday,† Simon Leung,†
Pricilla Richberg,∥ David Hughes,∥ William N. Washburn,§ Wei Meng,§ Kevin J. Volk,†
and Richard H. Mueller†
†Chemical Development, New Brunswick, New Jersey
‡Process Research and Development, Princeton, New Jersey
§Metabolic Disease Discovery Chemistry, Hopewell, New Jersey
⊥Process Research and Development, Regensburg, Germany
∥Analytical Research and Development, Princeton, New Jersey
¶Exploratory Biopharmaceutics and Stability, New Brunswick, New Jersey
Bristol-Myers Squibb Pharmaceutical Co., P.O. Box 191, New Brunswick, New Jersey 08903-0191, United States
ABSTRACT: A practical synthesis of the SGLT-2 inhibitor β-C-aryl-D-glucoside (1) has been developed. The route employed
2,3,4,6-tetra-O-trimethlysilyl-D-glucano-1,5-lactone as the key chiral building block, prepared efficiently from the commercially
available, inexpensive raw materials, D-gluconolactone and trimethylsilyl chloride. The salient step in the synthesis is the Lewis
acid-mediated stereoselective reduction of a methyl C-aryl peracetylated glycoside using a silyl hydride to set the stereochemistry
of the crucial anomeric chiral center. Several novel cocrystalline complexes of 1 with L-phenylalanine and L-proline were
discovered. Single-crystal structures of these complexes and several synthetic intermediates have been determined. The L-
phenylalanine complex was developed and used to purify and isolate the API. All steps were implemented at multikilogram scale.
colorless, neat crystalline compound in 65% yield, >98 A%
purity.
INTRODUCTION
■
The synthetic β-C-arylglucoside (2S,3R,4R,5S,6R)-2-(3-(4-
ethylbenzyl)-(phenyl)-6-hydroxymethyl-tetrahydro-2H-pyran-
3,4,5-triol1 (1) has been identified as a potent and selective
hSGLT2 inhibitor with potential for the treatment of diabetes.
The X-ray single-crystal structure6 of 5 confirmed that a
single isomer was obtained, with the hydroxyl in the α-
configuration.
The key step in this synthesis is the reduction of lactol 5;
typically, this has required sterically hindered silanes7 in order
to obtain high β-selectivity. Thus, reduction of 5 using
triisopropylsilyl hydride (2 equiv) in the presence of BF3·Et2O
(1.05 equiv) at −40 °C in DCM gave an anomeric mixture of
6a (β-aryl)6 and 6b (α-aryl) (>20:1 ratio) in ∼63% yield. The
α-isomer 6b was purged in a subsequent crystallization. Seeking
to further increase the isomer ratio, we screened the reduction
of lactol 5 using other, more readily available silanes. This effort
was unsuccessful since entities such as water-soluble
heptamethylsiloxane, gave good conversion, but poor β-
selectivity (α/β, 1:∼4.5). Hydrogenolysis (Pd/C) of the final
intermediate 6a in THF provided 1. The crude product 1 was
isolated as an amorphous foam with typical yield >95−97% and
with >94 A% purity.8 Chromatography was used during early
development for further purification. The typical laboratory
chromatography yields were >73% and improved 99.3 A%
To generate material in support of phase I clinical trials, the
first-generation Discovery Chemistry synthesis, based on
methodology developed by Kishi2 and Kraus,3 was used to
gain rapid access to compound 1 starting with 2,3,4,5-tetra-O-
benzyl-D-glucose 2 (Scheme 1).
Lactone (3)4 was prepared from compound 2 via TEMPO/
bleach oxidation in >90% yield. On large scale, purification of
the syrupy product 3 was not practical; therefore, it was used
directly in the next step without purification. The aglycon 3-
bromo-(4′-ethyl)-diphenylmethane (4) was synthesized using a
known two-step process.5 The lithio-anion derived from 4 was
coupled with lactone 3 at −78 °C to obtain lactol 5 as a
Received: October 24, 2011
Published: March 28, 2012
© 2012 American Chemical Society
577
dx.doi.org/10.1021/op200306q | Org. Process Res. Dev. 2012, 16, 577−585

Organic Process Research & Development
Article
Scheme 1. First-generation synthesis of aryl glucoside 1
Scheme 2. Process R & D synthesis of tetraacetate 11
purity. All of the above steps, except chromatography, were
optimized for scale-up in our glass plant (200-L scale).
Although this process enabled rapid delivery of supplies to
support early toxicology and preclinical studies, the inability to
crystallize 1 was of paramount concern.
During our process development of the first-generation
synthesis, we identified several disadvantages to using 2,3,4,6-
tetra-O-benzyl-D-glucono-1,5-lactone 3 as a starting material.
For instance, 3 was synthesized from commercially available,
but expensive, 2,3,4,6-tetrabenzylglucose 2. Product 3 was
difficult to isolate and purify at multikilogram scale due to its
syrupy nature. The crude product also gradually decomposed
over several months, thereby limiting its usefulness in a
commercial process. The reduction of intermediate lactol 5
required sterically hindered silanes to give favorable β-
selectivity, and these silanes were expensive and difficult to
obtain on commercial scale. More critically, the synthesis of C-
arylglycosides via 2,3,4,6-tetra-O-benzyl gluconolactone 3
required a hydrogenolytic deprotection in the final step of
the synthesis. Due to the amorphous nature of 1, chromatog-
raphy was required to remove palladium and process impurities.
Despite extensive investigation, a crystalline form of 1 has
not been discovered, possibly due to the amphiphatic nature of
the molecule. In addition, the amorphous, foamy active
pharmaceutical ingredient (API) 1 becomes a tacky or
gummy solid on exposure to moisture. A syrupy, noncrystalline,
and hygroscopic form presented a formidable challenge to the
development team. The lack of sufficient control over the
isolation and purification of amorphous 1, the high cost of
goods, and the unfavorable physical characteristics of several
intermediates created increased urgency and provided
motivation for the development of a new process and a new
final form for API 1.
Recently, we communicated9 a new methodology for a
general stereoselective synthesis of β-C-arylglucosides in which
the benzyl ethers employed in Scheme 1 were replaced by alkyl
578
dx.doi.org/10.1021/op200306q | Org. Process Res. Dev. 2012, 16, 577−585

Organic Process Research & Development
Article
esters. We also demonstrated that Lewis acid-mediated ionic
reduction of a peracetylated 1-C-arylmethyl-glucoside pro-
ceeded with a high degree of selectivity to give β-C-
arylglucosides, despite initial concerns that a 2-acyl group
might disfavor the desired stereoselective reduction via
intramolecular stabilization of the oxocarbenium ion. In this
publication, we provide an application of the development of
this new approach to synthesize the β-C-arylglucoside drug
candidate 1 in multikilogram quantities. Furthermore, unique
cocrystallizations of API 1 with amino acids as hydrogen-
bonded complexes permitted efficient isolation and purification.
concomitant removal of the TMS protecting groups to produce
a crude solution of methyl glycoside 9. The intermediate 9 was
an amorphous, foamy solid. Extensive investigation failed to
produce a crystalline form of 9. Thus, after an aqueous workup,
the solution was dried via toluene azeotrope and used into the
following step. The toluene solution of 9 was prepared in 83%
assay yield and 87 A% purity.
On the basis of our previous investigations, acetate groups
were chosen to protect the hydroxyls in 9 as the reduction
would provide advantageous crystalline peracetylated 1 (11).
The toluene solution of 9 was fully acetylated by Ac2O/TEA
and catalyzed by DMAP to generate the tetraacetate 10. The
remaining Ac2O was quenched by aqueous phosphoric acid and
an aqueous wash. After phase separation, the toluene solution
was partially concentrated. The dry toluene solution was ready
for the key transformation: reduction.
RESULTS AND DISCUSSION
■
While some derivatives of glucose (such as 2) can be expensive
and, as such, not realistic starting materials, we realized the
advantages of employing materials from the naturally occurring
chiral pool which eliminated the need to control four of the five
stereocenters of compound 1. This suggested we retain the core
of 2 or 3 as a starting material and explore protecting-group
modifications to circumvent the adverse physical characteristics
of the intermediates in the original synthesis. In this manner,
proper protecting-group selection could offer cost advantages
and improvements to the physical form of certain intermediates
for ease of isolation. A final desired change would be the
discovery of a simple deprotection method to avoid the use of
hydrogenolysis which was required in the first-generation
synthesis.
We screened a variety of anomeric reduction conditions,
focusing on the original acid-catalyzed silyl hydride process. We
investigated both Lewis and Brønsted acids as potential
alternatives to BF3·Et2O along with triethylsilyl hydride as an
alternative to the triisopropyl derivative (Table 1). While
Table 1. Study of the reduction of methyl glucoside 10
a
b
entry
catalyst
Lewis acid
BF3·Et2O
BF3·Et2O
BF3·Et2O
equiv
H2O (equiv)
conv
1
2
2
3
4
5
4
5
6
7
4
1.2
1
>99
>99
<20
58
3
A key need was to identify a means to purify compound 1 as
a crystalline API instead of resorting to chromatography. We
discovered that peracetylation of 1 yielded the corresponding
tetraacetate 11 (see Scheme 2) as a stable and pure crystalline
solid that was easy to crystallize and purify. The subsequent
deprotection of this material to 1 was facile, and the resulting
API purity was very high (>99.9 A%). We recognized the
desirable traits of this crystalline derivative for the endgame of
our new process. However, there was a concern that acetates
would not be compatible with the organometallic chemistry
used to generate aglycon 3. As an alternative, we employed the
transient protection of gluconolactone as the per-trimethylsilyl
derivative9 as the trimethylsilyl groups would resist lithioar-
omatics at low temperature but could also be cleaved during the
workup under mildly acidic conditions. Starting with D-
gluconolactone 7, silyl protection was achieved using
TMSCl/N-methyl morpholine in THF to give 7a. The workup
of this intermediate required a carefully controlled process as its
oily nature and sensitivity to moisture precluded easy isolation.
Therefore, the reaction mixture was diluted with toluene prior
to aqueous washes, and the persilylated lactone 7a was obtained
as a solution. The selection of toluene as solvent was critical as
it efficiently extracted the product and enabled azeotropic
drying that was needed in the following step.
Aglycon 4 was transmetalated with n-BuLi to form the
lithiated species 4a which was reacted at −78 °C with the
solution of 7a prepared above to give lactol 8. In this
transformation, the toluene/THF solvent combination (3.5:1)
proved ideal for yield and product purity; in THF solvent, a
competing silyl transfer reaction dominates, forming trimethyl-
silyl-aglycon as a major byproduct.10
Lactol 8b was noncrystalline and problematic to isolate;
therefore, the THF/toluene solution of 8b was quenched into
methanol/methanesulfonic acid. This acidic quench served to:
(1) neutralize any excess organometallic species, (2) promote
in situ trans-ketalization with methanol, and (3) expedite
3
<0.1
1.4
BF3·Et2O
BF3·Et2O
3
3
1.75
1.2
31
BF3(H2O)2/BF3
ZrCl4
0.5/3.5
3.5
4.0
4
98.4
84
0.15
0.2
HBF4·Et2O
BF3.THF
>99
1.0
0
0
ZnCl2
2.5
1.2
Brønsted Acid
CF3SO3H
8
2.5
4.0
2.5
2.5
0.15
2.0
>99
0
9
CH3SO3H
CF3SO3Si(CH3)3
CF3SO3Si(CH3)3
10
11
0.2
93.4
>99
1.2
a
[7a]: 0.04 M, Et3SiH: 4 equiv, 10 °C, order of addition: Lewis acid +
b
Et3SiH, then substrate after 40 min. Conversion of β to α after 110
min determined by HPLC analysis. Typical β to α ratio was observed
>20:1 for all substrates.
several alternative catalysts gave encouraging results, we
decided to retain BF3·Et2O for a combination of cost, product
purity and process robustness. The less expensive and more
readily available triethylsilyl hydride (Et3SiH) replaced
triisopropylsilyl hydride (iPr3SiH) since the resulting isomer
ratio was similar.
The optimized process starts with dilution of the toluene
solution of 10 with acetonitrile followed by reduction with
Et3SiH, BF3·Et2O, and one equivalent of water. Water is critical
for the reaction to progress to completion, water and BF3·OEt2
generate a strong Brønsted acid, such as H+BF3OH−, which
accelerates11 the formation of an oxa-carbenium ion inter-
mediate. These strongly acidic conditions are essential to
compensate for the deactivating effect of the carbohydrate
acetoxy-substituents. On the basis of the literature12 and studies
conducted,9a we believe that predominant axial hydride attack
from the α-face on the oxacarbenium ion intermediate provides
579
dx.doi.org/10.1021/op200306q | Org. Process Res. Dev. 2012, 16, 577−585

Organic Process Research & Development
Article
Scheme 3. Synthesis of complex 13
Figure 1. X-ray structure of complexes 12 (1-F2) and 13 (1-F1).
equatorially substituted β-C-arylglucoside 116 as the major
product (>20:1 β:α).
mol of water) (see Figure 1). Both forms were fine needles and
relatively stable at room temperature (22−25 °C), but when
suspended in water, the (1-F2) complex 12 dissolved and
crystallized as the (1-F1) complex 13. The structures of these
hydrogen-bonded complexes could only be solved from
synchrotron X-ray diffraction data of the very fine, hair-like
crystals.6 Our studies showed that the 1:1 complex is more
stable than the 1:2 complex in water and water-rich
environment (e.g., EtOH/H2O = 17/83). When 1:2 complex
was slurried in water it converted to 1:1 complex within hours;
similarly, 50/50 mixture of 1:1 complex and 1:2 complexes that
were slurried in water (40 °C) for overnight converted to pure
1:1 complex. Crystals of 1:2 complex appear to be long and
thin rods, melt at 140−145 °C and lose ∼2.5% weight at 100
°C when observed by DSC. This weight loss corresponds to the
loss of 1 mol equiv of H2O. Crystals of the 1:1 complex appear
to be fine, hair-like materials or short needles/rods, melt at
100−105 °C, and lose ∼3.3% weight at 125 °C. In addition, the
1-F1 complex provided a significant weight advantage over the
1-F2 complex. On the basis of these observations form 1-F1
(13) was chosen as the final form for further development.
In addition to L-phenylalanine, several stable crystalline
complexes of 1 with D-phenylalanine and L-proline were
prepared and their stuctures6 determined by single-crystal X-
ray analysis. The power of such foreign agents as cocrystalliza-
tion partners in assembling crystal structures of pharmaceutical
productswhich alone crystallize with difficulty, if at allhas
been increasingly recognized in recent years.15
The use of Et3SiH presented safety considerations on
multikilogram scale that required attention. Unreacted silyl
hydride generated hydrogen gas in spent reaction streams and
rendered waste hazardous to store and transport. To mitigate
this risk, 2,2-dimethoxypropane was charged to the mature
reaction mixture to quench the excess Et3SiH. After complete
disappearance of triethylsilyl hydride was shown by GC
analysis, aqueous workup and crystallization of the product
from ethanol produced a 74% yield of the final intermediate 11
(98.6 HPLC A% purity). Saponification of tetraacetate 11 (aq
NaOH/EtOH) produced API 1 as an amorphous compound
(Scheme 3).
To solve the isolation and final crystal form issues of the
amorphous, amphiphatic drug substance 1, we focused on a
cocrystallization approach.13 Although in principle, virtually any
pair of molecules may cocrystallize,14 we reasoned that another
amphiphatic cocrystallization component might favorably pack
and interact with both the polar hydroxyls and the hydrophobic
hydrocarbon regions of 1, thus enhancing the probability of
cocrystallization. We identified several natural amino acids as
possible partners for cocrystallization with 1 but concentrated
on L-phenylalanine which has a polar and hydrophobic nature
similar to that of the drug molecule. These efforts provided two
crystal forms (complexes): 12 = form 1-F2 (composition: 1
mol of 1 + 2 mol of L-phenylalanine +1 mol of water), and 13 =
1-F1 (composition: 1 mol of 1 + 1 mol of L-phenylalanine +1
580
dx.doi.org/10.1021/op200306q | Org. Process Res. Dev. 2012, 16, 577−585

Organic Process Research & Development
Article
As there was a defined form and a route to a pure API, the
development of the multikilogram synthesis only remained. In
contrast to the hydrogenation required to cleave the benzyl
protecting groups of the final intermediate in the first-
generation synthesis (Scheme 1), the final step now required
only aqueous base to remove the acetates. The saponification is
rapid and irreversible and can be performed under a wide range
of reagent concentrations and temperature conditions.
From all our experimental data we have observed that the
stereochemical integrity of the product in the final step was
very effectively preserved. Essentially no process changes were
required to scale up the final intermediate 11 to 64 kg scale.
Deacetylation using EtOH/aq NaOH (Scheme 3) followed by
neutralization with aq HCl to pH 7 generated a solution of free
compound 1 which was converted to the L-phenylalanine
complex 13 (1-F1) and crystallized as a monohydrate. After
filtration and drying, 13 (1-F1) was produced with 99.9 A%
purity and 80−88% yield (58 kg). This material has been used
for clinical investigations.
100% B; flow 2.5 mL/min, Inj Vol 5 μL, UV absorption 220
nm, runtime 10 min. Moisture determinations were performed
with a Karl Fisher titrator. For persilylated TMS lactone (7a),
the gas chromatographic analyses were performed using HP
5890 with FID (He carrier) with a Restek Corp. column: Rtx-5
lot # 7992B, 30 m × 0.32 mm I.D. Initial temperature 200 °C,
final temperature 260 °C, rate 5 °C/min, injection volume 1
μL, total run time 15 min. For triethylsilyl hydride, gas
chromatographic analysis was performed using instrument HP
GC5890 II with FID or equivalent, column: Rtx-5, 30 m × 32
mm I.D., 0.25 μm film thickness; Initial temp. 40 °C, hold for 1
min, rate 1:5 °C/min, final temp. 60 °C, no hold, rate 2:50 °C/
min, final temp.: 200 °C, hold for 0.2 min, run time 8 min.
Injection port temperature 260 °C, detector temperature 280
°C with He gas carrier. This GC method determines the
residual triethylsilyl hydride (TESH) at low ppm level. An
external standard calibration was used for quantitation of
triethylsilyl hydride in process samples. Detection limit for this
GC method is 0.2 ppm triethylsilyl hydride. Initial development
scale-up of first-generation reactions (Scheme 1) were
conducted in kilo lab with 100- or 200-L size reactors, and
new synthesis (Scheme 2) was demonstrated at the pilot plant
in typical 2000-L reactors. Purified (deionized) water was used
for all plant operations.
2,3,4,6-Tetra-O-(phenylmethyl)-1-C-[3-[(4-
ethylphenyl)methyl]phenyl]glucopyranose (5). A reactor
was purged with argon gas and charged sequentially with
bromide 4 (4.5 kg), TMEDA (7.1 L), and anhydrous THF
(40.5 L). The solution was cooled to −70 to −75 °C, and n-
BuLi (6.5 L, 2.5 M in hexane) was added while maintaining
temperature at −65 to −70 °C range. The orange-red solution
became yellow at the end of the addition of n-BuLi. HPLC after
stirring for 30 min at −70 °C showed disappearance of bromide
4. A second reactor with a solution of lactone 3 (7.5 kg) in
THF (12.5 L) was prepared under argon gas. This solution was
transferred in five equal portions to the organometallic solution
prepared above at ≤−65 °C. After addition of each portion of
the lactone solution over 10 min while maintaining temperature
≤−65 °C, the batch was agitated for 3 min. After the fourth
charge, the batch was analyzed for lactone content by HPLC. If
the lactone content was >5%, the further charge of the fifth
portion of lactone solution was not performed. Otherwise, after
charging the fifth portion of lactone solution the batch was
agitated further for 30 min at −70 to −75 °C or until the
lactone 3 was <5% by HPLC. The cooling was stopped, and a
solution of 10% citric acid in water (59.3 L) was added. The
temperature of the mixture was allowed warm up to 18 °C over
3−5 h. MTBE (20.2 L) was added, and the mixture was
agitated for 10 min. The layers were separated, and the aqueous
layer was extracted with MTBE (20.2 L). The organic layers
were combined and washed with aqueous saturated NaHCO3
solution (20.2 L) and then with brine (20.2 L). The organic
layer was evaporated to give 11.125 kg of oil. The product was
dissolved in MeOH (44.5 L), the clear solution was agitated,
and water (8.9 L) was added to produce an oily suspension.
The temperature of the mixture was kept above 27 °C during
this process. Heptanes (133.5 L) was added, and the mixture
was stirred for 10 min. The turbid heptanes layer was separated,
and the MeOH/water layer was extracted with heptanes (2 ×
66.7 L). The heptanes layers were passed through a Celite filter
and then concentrated under vacuum to ∼60 L volume. Some
oil separated during concentration. The mixture was warmed to
65 °C to dissolve some separated oil and then cooled to 40 °C
CONCLUSIONS
■
We have demonstrated a robust and practical synthesis of drug
substance 1 on manufacturing scale. The salient features of the
synthesis are (a) the replacement of the conventional benzyl
protecting group of the chiral sugar synthon by a combination
of transitory silyl protection followed by ester protection, (b)
demonstration on multikilo scale of the feasibility of ester
protection and stereoselective ionic reduction using a less
hindered triethylsilyl hydride (Et3SiH) for the reduction of
methyl glycoside 10 to obtain β-C-arylglucoside 11, (c) the use
of peracetylated β-C-arylglucoside 11 to control quality for the
overall synthesis as it was highly crystalline with good impurity-
purging capability, (d) replacement of the hydrogenation
operation by a mild ester hydrolysis in the final chemical
step, and (e) demonstration of an unusual means for
purification via cocrystallization with L-phenylalanine. Problems
with oily intermediates were circumvented by conducting the
entire synthesis in two telescoped processes. Anti-diabetic
activity and other structure−activity relationship data of the
SGLT-2 inhibitor 1 will be reported in due course.
EXPERIMENTAL SECTION
■
All new compounds described in the Experimental Section
were fully characterized. The experimental outlined below
provides representative procedures for what was run in the
kilo/pilot-plant facilities. 1H NMR spectra were obtained at 400
MHz, 13C NMR spectra were obtained at 100 MHz at room
temperature (22−25 °C) in CDCl3 or CD3OD, except as
indicated. Chemical shifts are reported in ppm downfield from
an internal tetramethylsilane standard or relative to residual
solvent signal. Mass spectroscopy was performed using a
Finnigan Navigator single quadrupole electrospray mass
spectrometer or a Finnigan aQa APCI mass spectrometer.
HPLC analysis results are described as area percent (A%).
HPLC analysis was performed with a YMC Pack FL-ODS 4.6
mm × 50 mm, 5 μm; The solvent A was 0.2% H3PO4 solution,
and solvent B was 0.2% acetonitrile; start 100% A and 0% B
flow 4 mL/min, Inj Vol 10 μL, UV absorption 210 nm, runtime
20 min 70% acetonitrile containing 0.1% TFA. Also used was
YMC S3 ODS-A 44.6 mm × 50 mm at room temperature (22−
25 °C); the solvent A was 0.2% H3PO4 solution, and solvent B
was 90% acetonitrile and 10% water; start 95% A and 5% B to
581
dx.doi.org/10.1021/op200306q | Org. Process Res. Dev. 2012, 16, 577−585

Organic Process Research & Development
Article
13C NMR (100 MHz, CDCl3) δ 15.54, 28.37, 41.48, 69.05,
73.43, 74.86, 75.10, 75.63, 76.84, 78.31, 79.38, 81.76, 84.4,
86.64, 125.33, 127.49, 127.55, 127.59, 127.69, 127.91, 127.99,
128.15, 128.19, 128.31, 128.40, 128.50, 128.76, 128.88, 137.69,
138.16, 138.20, 138.36, 138.70, 138.31, 141.37, 141.89.
and seeded with 5 (∼1%). After the start of crystallization, the
batch was cooled to 20 °C and held for 12 h. The mixture was
further cooled to 0−5 °C, and then the solids were collected by
filtration or by centrifuge. The cake was washed with 0 °C
heptanes (3 × 4.5 L) and dried under vacuum at 30 °C to a
constant weight to give 7.12 kg (59.4%, >99.00 A%) of
compound 5: Anal. Calcd for C49H50O6: C, 81.86; H, 7.01. Fd.
C, 81.66; H, 6.88. HRMS: Calcd for M + H = 717.3580 Da,
Measured 717.3590 Da. 1H NMR (400 MHz, CDCl3) δ 1.20 (t,
3H, J = 7.5 Hz), 1.62 (s, 1H), 2.59 (q, 2H, J = 7.5 Hz), 3.56 (d,
1H, J = 9.2 Hz), 3.72−3.75 (dd, 1H, J = 1.8 and 9.2 Hz), 3.81−
3.88 (m, 3H), 3.95 (s, 2H), 4.08 (t, 1H, J = 9.2 Hz), 4.16−4.20
(m, 1H), 4.37 (d, 1H, J = 10.5 Hz), 4.55 (d, 1H, J = 12 Hz),
4.65 (d, 1H, J = 6.2 Hz), 4.68 (d, 1H, J = 3.5 Hz), 4.87−4.93
(m, 3H), 6.96 (d, 2H, J = 7.5 Hz), 7.19 (d, 2H, J = 7.5 Hz),
7.21−7.52 (m, 24H, Ar-H). 13C NMR (100 MHz, CDCl3) δ
15.54, 28.37, 41.60, 68.91, 72.16, 73.33, 75.02, 75.41, 75.67,
77.20, 78.33, 83.44, 85.17, 97.92, 123.92, 126.72, 127.45,
127.55, 127.63, 127.69, 127.93, 128.13, 128.25, 128.29, 128.36,
128.72, 129.14, 137.45, 138.12, 138.26, 138.50, 138.66, 141.28,
141.91, 142.41.
2,3,4,6-Tetra-O-(phenylmethyl)-1,5-anhydro-1-C-[3-
[(4-ethylphenyl)methyl]phenyl]glucopyranose (6a). To a
solution of glucopyranose 5 (1.22 kg) in CH3CN (21.89 L) at
−25 °C was added triisopropylsilane (0.93 L), followed by
BF3·Et2O (0.28 L) at such rate that temperature does not
exceed ≤−20 °C. (Caution: The reaction is exothermic.) After
stirring for 2.5 h at −25 °C, HPLC analysis indicated 5 was
≤0.6 HPLC A%. A saturated aqueous solution of NaHCO3
(9.73 L) was added over 10 min. (Caution: Slight foaming due
to the evolution of H2 and CO2.) The temperature of the
mixture increased to −5 °C. MTBE (12.16 L) was added, and
the agitation was continued for 10 min at 5 °C. Water (4.86 L)
was added to dissolve the precipitated solid, and the layers were
separated. The organic layer was washed with aqueous
saturated NaHCO3 (9.73 L), and this aqueous wash was
extracted with MTBE (4.86 L). The combined organic extracts
were washed with brine (9.73 L). The organic layer was
evaporated under vacuum to give 1.34 kg of product 6 as
mixture of isomers (β:α ∼20:1) as an oil. The material was
dissolved in MeOH (6.69 L) at room temperature (22−25 °C),
and seeds of 6a (∼1%) were added. The suspension was stirred
for 2 h at room temperature (22−25 °C) and for 1 h at 0 °C.
The crystals were filtered, washed with MeOH (2 × 0.97 L) at
room temperature (22−25 °C), and dried under vacuum at 30
°C to a constant weight to give 0.83 kg of product 6a. The
compound was redissolved in heptanes (8.3 L) at 60 °C. The
turbid solution was polish filtered at 60 °C, and the filter was
washed with heptanes (0.83 L). The filtrate was cooled slowly
to 38 °C, and seeds of compound 6a were added.
Crystallization started at 35 °C. The mixture was stirred at
room temperature (22−25 °C) for 12 h and then was cooled to
0 °C. After 2 h the crystals were filtered and washed with cold
heptanes (2 × 0.83 L). The cake was dried under vacuum to
constant weight to yield 0.76 kg (63%, 98.0 A%) of compound
6a: Anal. Calcd for C49H50O5: C, 80.08; H, 6.85. Fd. C, 79.89;
H, 6.61. HRMS: Calcd for M − H2O + H = 719.3737 Da,
Measured 719.3729 Da. 1H NMR (400 MHz, CDCl3) δ 1.21 (t,
3H, J = 7.5 Hz), 2.57−2.63 (q, 2H, J = 7.5 Hz), 3.53 (t, 1H, J =
4.4 Hz), 3.57−3.62 (m, 1H), 3.74−3.83 (m, 5H), 3.90 (s, 2H),
4.23 (d, 1H, J = 9.3 Hz), 4.35 (d, 1H, J = 10.2 Hz), 4.57 (d, 1H,
J = 12.3 Hz), 4.64−4.68 (dd, 2H, J = 2.6 and 10.9 Hz), 4.87−
4.97 (m, 3H), 6.89 (d, 2H, J = 7.4 Hz), 7.17−7.37 (m, 26H).
(2S,3R,4R,5S,6R)-2-(3-(4-Ethylbenzyl)-(phenyl)-6-hy-
droxymethyl-tetrahydro-2H-pyran-3,4,5-triol (1). Hy-
dogenolysis of 6a. A hydrogenation vessel was charged with
peroxide-free THF (200 mL) and purged with N2 gas for 20
min. The catalyst 10% Pd/C (4.0 g, 50% wet with water,
Heraeus-type K-02105) and compound 6a (20.0 g) were
added, and the mixture was purged with N2 gas for 10 min.
Hydrogen gas was sparged through the mixture at 20−25 °C
(cooling required) until the H2 uptake ceased. After 4 h
completion of reduction was confirmed by HPLC (IPC: 6a
should be ≤0.1 HPLC A%). The mixture was sparged with N2
for 10 min and then stirred with charcoal (6.0 g) for 18 h to
remove color. The mixture was filtered over a pad of filter aid
(1.0 g), and the filter was washed with THF (40 mL). The
filtrate was evaporated in vacuum to give 10.3 g (103%, >97 A
%) of crude compound 1 as a colorless foam . The product was
slightly hygroscopic and therefore was stored under a N2
atmosphere: HRMS Calcd for C21H27O5: M + H = 359.1859
Da, Measured = 359.1863 Da. 1H NMR (400 MHz, DMSO-d6)
δ 1.13 (t, 3H, J = 7.4 Hz), 2.51−2.56 (q, 2H, J = 7.4 Hz), 3.11−
3.29 (m, 4H), 3.30 (s, 1H), 3.33 (br s, 6H), 3.40−3.45 (m,
1H), 3.66−3.70 (m, 1H), 3.87 (s, 2H), 3.95 (d, 1H, J = 9.3
Hz), 7.07−7.22 (m, 8H). 13C NMR (100 MHz, DMSO-d6) δ
15.58, 28.35, 41.32, 61.87, 70.02, 74.68, 77.2, 77.97, 79.44,
81.82, 125.07, 127.93, 128.58, 128.82, 129.0, 137.88, 138.40,
141.53, 141.97.
(2S,3R,4R,5S,6R)-2-(3-(4-Ethylbenzyl)-(phenyl)-6-hy-
droxymethyl-tetrahydro-2H-pyran-3,4,5-triol:2 L-Phe-
nylalanine:H2O Complex (12). A slurry of C-arylglucoside
1 (5.10 g) in EtOH (30.0 mL), water (32.0 mL) and L-
phenylalanine (4.2 g) was heated at 80 °C to obtain a clear
solution and then cooled to 20 °C over 2 h. The complex starts
to crystallize ∼40−45 °C. The temperature was held at 20 °C
for 6 h. The crystals were filtered and washed with 0 °C water
(20 mL) followed by MTBE (20 mL). The product was dried
under vacuum at 40 °C to constant weight to give 7.1 g (70.5%,
99 A%) of the complex 12: Anal. Calcd for C39H50N2O10: C,
1
66.27; H, 7.13; N, 3.96. Fd. C, 66.22; H, 7.15; N, 3.93. H
NMR (400 MHz, CD3OD/D2O) δ 1.16 (t, 3H, J = 7.6 Hz,
CH3), 2.56 (q, 4H, J = 7.5 Hz, CH2), 3.06 (dd, 1H, J = 8.5 Hz, J
= 14.4 Hz, PhAl-CH2), 3.30 (m, 1H, PhAl-CH2), 3.46 (m, 3H,
H-2, H-3, H-4), 3.53 (q, 1H, J = 8.8 Hz, H-5), 3.69 (dd, 1H,
11.8 Hz, 5.6 Hz, H-6a), 3.85 (m, 2H, H-6b, PhAl-HCOOH),
3.92 (s, 2H, Ph-CH2-Ph), 4.15 (d, 1H, 9.4 Hz, H-1), 4.71
(MeOH), 5.0 (1H, HOH), 7.10 (m, 4H, Ar-H), 7.3 (m, 9H,
Ar-H). 13C NMR (100 MHz, CD3OD/D2O) δ 16.37, 29.34,
37.93, 42.27, 57.48, 62.77, 71.60, 75.98, 79.31, 81.81, 83.50,
126.75, 128.65, 129.01, 129.59, 129.92, 130.16, 130.54, 136.84,
139.82, 140.26, 143.09, 143.39, 174.29.
(2S,3R,4R,5S,6R)-2-(3-(4-Ethylbenzyl)-(phenyl)-6-hy-
droxymethyl-tetrahydro-2H-pyran-3,4,5-triol:L-Phenyla-
lanine Complex (13). A slurry of C-arylglucoside 1 (10.0 g) in
EtOH (14.0 mL) was heated to 60 °C to obtain a clear
solution. This solution was added to a suspension of L-
phenylalanine (4.6 g) in water (180 mL). EtOH (3.6 mL) was
used as a rinse to transfer the residual solution of compound 1.
The mixture was heated to 83 °C and then cooled slowly to 52
− 54 °C over 10 to 15 min and seeds (∼1%) of complex 13
582
dx.doi.org/10.1021/op200306q | Org. Process Res. Dev. 2012, 16, 577−585

Organic Process Research & Development
Article
were added. (Note: The seeding process was repeated at this
temperature if the original seed crystals had dissolved). The
mixture was cooled to 40−42 °C and stirred for 4 h. The
temperature was lowered to 22−25 °C over 2−4 h, and stirring
was continued for 2 h. Finally, the mixture was cooled to 18 °C
and stirred for 2 h. The crystals were filtered and washed with
ice-cold water (25 mL) and MTBE (2 × 5 mL). The product
was dried under vacuum at 40 °C to constant weight to give
12.0 g (82%, 99 A%) of the complex 13. (Full characterization
data for 13 is given at the end of final process.)
Typical reaction times were 4−7 h. A solution of H3PO4 (48.44
kg) in water (528 kg) was charged. If necessary, additional
H3PO4 was added until at pH ≤3. The mixture was stirred for
10 min and then the bottom aqueous phase was separated. The
organic phase was washed by water (245.54 kg). The organic
solution containing the acetylated product 10 was concentrated
at atmospheric pressure to a volume of 4−6 L/kg of methyl-1-
C-(4′-ethyldiphenylmethane-3-yl)-α-D-glucopyranoside (9),
this solution was directly used for the next telescopic process.
α-D-Glucopyranoside, methyl 1-C-[3-[(4-ethylphenyl)-
1
a). 2,3,4,6-Tetra-O-acetyl-β-D-glucopyranosyl-1-C-(4′-
ethyldiphenyl-methane-3-yl) (11). Synthesis of 7a. D-
Glucono-1,5-lactone (7) (35 kg) and THF (344.8 kg) was
treated with 4-methylmorpholine (146.5 kg, ∼8 equiv), and the
slurry was cooled to 5 °C. Trimethylsilylchloride (110.9 kg, 6
equiv) was introduced, the slurry was stirred for 15 min and
then warmed to 30−35 °C over 0.5 h. After 5 h the reaction
was complete by GC analysis. The reaction mixture was cooled
to 0−5 °C. Toluene (454.1 kg) was charged and then the
reaction was quenched with water (700 kg). The mixture was
stirred for 10−15 min, and then the phases were separated. The
bottom aqueous layer was removed, and a solution of
NaH2PO4 (13.0 kg) in water (260 kg) was added. The mixture
was agitated for 10 min, and then the phases were separated.
The bottom aqueous layer was removed, and water (273 kg)
was charged. The mixture was agitated for 10 min, and the
aqueous layer was separated. The organic solution was distilled
at 40−60 °C under reduced pressure (23−25 in. Hg), while
adding toluene until the solution was <0.05% water by KF
analysis. The concentration of the silylated product solution
was adjusted to ∼0.1 g/mL by adding anhydrous toluene or
distillation, as necessary. The solution yield of 2,3,4,6-tetra-O-
trimethylsilyl-D-gluconolactone (7a)9,10 ranged from 91 to 98%,
which is measured by a GC assay based on the standard GC
curve of the product or by weighing after removal of the
solvent.
Conversion of 7a to 11. 5-Bromo-4′ethyldiphenylmethane
(4) (44.8 kg), anhydrous THF (78.8 kg) and toluene (281 kg)
at < −70 °C was treated with n-BuLi (48.9 kg, 2.5 M in
hexanes) at < −70 °C. The mixture was stirred until the
lithiation was complete by HPLC analysis (when compound 4
is ≤1% relative to the hydrocarbon). This lithiated anion
solution derived was added to the cooled 2,3,4,6-tetra-O-
trimethylsilyl-D-gluconolactone solution prepared above at ←
70 °C. The mixture was agitated at ←70 °C for one to two
hours.. A solution of CH3SO3H (22.2 kg, 1.4 equiv) in 334 kg
MeOH was charged while maintaining the temperature of the
reaction mixture ←60 °C. After completion of the methyl
glycosidation reaction (by HPLC) an aqueous sodium
bicarbonate solution (11 kg of NaHCO3 in 220 kg water)
was added. The organic layer was washed with water (220 kg),
and the aqueous phase was removed. The combined aqueous
layers were extracted with ethyl acetate (223 kg). The organic
layers were combined, and the solvent was distilled at 35−60
°C under reduced pressure (∼25 in. Hg) until the KF of the
solution was <0.07% and the amount of EtOAc was <1%
relative to toluene by GC analysis. The solution yield of the
product from the coupling ranged from 72 to 89%.
methyl]phenyl] (9). H NMR (400 MHz, CDCl3) δ 1.17 (t,
3H, J = 7.5 Hz), 2.52−2.58 (q, 2H, J = 7.5 Hz), 2.84 (m, 5H),
3.01 (br s, 1H), 3.24 (d, 1H, J = 9.3 Hz), 3.55−3.65 (m, 2H),
3.81−3.95 (m, 6H), 4.01 (br s, 1H), 5.01−5.22 (m, 1H), 7.01−
7.06 (m, 4H), 7.15 (t, 1H, J = 7.9 Hz), 7.28 (d, 2H, J = 7.9 Hz),
7.36 (s, 1H). 13C NMR (100 MHz, CDCl3) δ 15.58, 28.37,
41.46, 49.16, 62.11, 70.22, 74.44, 74.68, 77.22, 101.33, 125.03,
127.71, 127.87, 128.11, 128.76, 128.88, 137.92, 138.16, 141.04,
141.89.
α-D-Glucopyranoside, Methyl 1-C-[3-[(4-ethylphenyl)-
methyl]phenyl]-2,3,4,6-tetraacetate (10). Anal. Calcd for
C30H36O10: C, 64.73; H, 6.51. Fd. C, 64.55; H, 6.46. HRMS:
Calcd for M − MeOH + H = 525.2125 Da, Measured 525.2124
1
Da. H NMR (400 MHz, CDCl3) δ 1.19 (t, 3H, J = 7.5 Hz),
1.71 (s, 3H), 1.94 (s, 3H), 2.05 (bt, 3H), 2.10 (s, 3H), 2.05 (s,
3H), 2.6 (q, 2H, J = 7.5 Hz), 3.1 (s, 3H), 3.9 (s, 2H), 4.02−
4.06 (m, 1H), 4.21−4.24 (dd, 1H, J = 2.2 and 12.3 Hz), 4.33−
4.37 (dd, 1H, J = 4.8 and 12.3 Hz), 4.94 (d, 1H, J = 10.1 Hz),
5.22 (t, 1H, J = 9.7 Hz), 5.57 (t, 1H, J = 9.7 Hz), 7.04 (d, 2H, J
= 7.4 Hz), 7.10 (d, 2H, J = 7.4 Hz), 7.16 (m, 1H), 7.23−7.34
(m, 3H). 13C NMR (100 MHz, CDCl3) δ 15.60, 20.20, 20.60,
20.64, 20.74, 28.39, 41.50, 49.49, 62.22, 68.73, 68.85, 71.41,
73.99, 100.36, 124.44, 127.19, 127.99, 128.62, 128.74, 129.68,
135.96, 138.02, 141.22, 142.05, 168.94, 169.56, 170.13, 170.71.
The reactor containing the acetylated intermediate 10 was
charged with acetonitrile (174.8 kg) and water (0.6 kg, 1 equiv
with respect to acetylated intermediate). The mixture was
cooled to ≤15 °C, and Et3SiH (30.3 kg) was added. BF3·Et2O
(24.5 kg, 2.1 equiv) was charged over >20 min, maintaining
temperature at <15 °C. The reaction mixture was stirred for 4−
7 h until complete by HPLC analysis. The mixture was cooled
to ∼15 °C and then charged the required amount of 2,2-
dimethoxypropane based on GC assay to determine the
remaining amount of Et3SiH, and KF to determine the
remaining amount of H2O. The reaction mixture was stirred
until the complete disappearance of Et3SiH was established by
GC analysis. The Et3SiH quench was judged to be complete
when a neat sample was assayed to have an undetectable
amount of Et3SiH in the GC assay. The minimum quantifiable
amount for this GC assay is ∼0.2 ppm. An aqueous solution of
NaHCO3 (120.02 kg NaHCO3 in 120.17 kg water) was
charged at room temperature (22−25 °C) until the pH of the
aqueous phase was ≥6. The mixture was agitated for ≥10 min,
and the phases were split. The lower aqueous phase was
removed, and a 20% solution of NaCl (24.03 kg salt in 120.17
kg water) was used to wash the organic layer. The aq phase was
separated, and the product-rich organic phase was concentrated
at atmospheric pressure until most of the acetonitrile was
removed (∼230 L of distillate was collected). Toluene (223.49
kg) was added, and the distillation was continued until the pot
temperature reached at least 112 °C and a volume of 4−6 L/kg
of methyl-1-C-(4′-ethyldiphenylmethane-3-yl)-α-D-glucopyra-
noside (9) input was reached. Heptane (240 kg) was charged
DMAP (0.11 kg), diisopropylethyl amine (64.33 kg) and
acetic anhydride (45.56 kg) were charged to the solution of
methylglucoside 9 above. The solution was stirred at ≤35 °C
until the combined area percent of the intermediate acetylated
species was ≤2% of the area of the tetra-acetylated product 10.
583
dx.doi.org/10.1021/op200306q | Org. Process Res. Dev. 2012, 16, 577−585

Organic Process Research & Development
Article
while maintaining the temperature at >70 °C. The solution was
cooled to ∼60 °C over ≥1 h, and the slurry was held at 60 10
°C for ≥1 h. The slurry was cooled to 20 10 °C over ≥1 h.
The slurry was filtered in a centrifuge, and the cake was washed
with ≥2 cake volumes of heptanes (58 kg). The wet cake was
dried under vacuum at ≤60 °C to an LOD of ≤0.2% to give
≥40 kg of the tetraacetate 11 (overall yield 86%, 99 A% purity):
Anal. Calcd for C29H34O9: C, 66.14; H, 6.50. Fd. C, 66.26; H,
6.50. HRMS: Calcd for M + H = 527.2281 Da, Measured
H), 7.3 (m, 9H, Ar-H). 13C NMR (100 MHz, CD3OD/D2O) δ
16.1, 29.2, 37.8, 42.1, 57.4, 62.7, 71.5, 75.8, 79.2, 81.7, 83.3,
126.5, 128.5, 128.8, 129.4, 129.5, 129.7, 130.0, 130.4 136.8,
139.6, 140.1, 142.9, 143.2, 174.1.
AUTHOR INFORMATION
■
Corresponding Author
*prashant.deshpande@bms.com
1
Notes
527.2278 Da. H NMR (400 MHz, CDCl3) δ 1.20 (t, 3H, J =
The authors declare no competing financial interest.
7.5 Hz), 1.65 (s, 3H), 1.98 (s, 3H), 2.05 (s, 3H), 2.07 (s, 3H),
2.57−2.62 (q, 2H, J = 7.5 Hz), 3.79−3.83 (m, 1H), 3.93 (s,
1H), 4.13−4.17 (dd, 1H, J = 2.3 and 12.3 Hz), 4.25−4.29 (dd,
1H, J = 4.8 and 12.3 Hz), 4.35 (d, 1H, J = 10.1 Hz), 5.09 (t, 1H,
J = 9.6 Hz), 5.22 (t, 1H, J = 9.7 Hz), 5.30 (t, 1H, J = 9.2 Hz),
7.05 (d, 2H, J = 6.2 Hz), 7.10 (d, 2H, J = 7.9 Hz), 7.13−7.29
(m, 4H). 13C NMR (100 MHz, CDCl3) δ 15.62, 20.12, 20.52,
20.76, 28.38, 41.40, 62.35, 68.59, 72.64, 74.25, 76.09, 80.18,
124.56, 127.95, 128.75, 129.48, 136.40, 141.32, 142.03, 150.12,
168.81, 169.50, 170.39, 170.75.
ACKNOWLEDGMENTS
■
We thank Dr. David Kronenthal, Dr. Rodney Parsons, and Dr.
Jaan Pesti for thoughtful comments during the preparation of
this manuscript.
DEDICATION
■
This manuscript is dedicated to the memory of John Dimarco, a
colleague and friend.
(2S,3R,4R,5S,6R)-2-(3-(4-Ethylbenzyl)-(phenyl)-6-hy-
droxymethyl-tetrahydro-2H-pyran-3,4,5-triol/L-phenyla-
lanine Complex/Water (13). Water (68.70 kg), 2,3,4,6-tetra-
O-acetyl-1-C- (4′-ethyldiphenyl-methane-3-yl)-β-D-glucopyra-
nose (11, 64.40 kg), and EtOH (62.44 kg, SDA-3A, 190
proof) were stirred with minimum agitation at 20 °C. NaOH
(57.02 kg, 0.1 N) was introduced, and the mixture was heated
slowly to 40−50 °C. After 1−2 h, the in-process HPLC
(excluding the solvent front and system peaks) of the
deacetylated product 1 in the solution was >97%. The solution
was cooled to 20 °C, water (154.56 kg) was charged, and the
solution temperature was adjusted to 18−25 °C. The mixture
was stirred for 1 h. Concentrated HCl (37%, ∼123.44 kg) was
added to adjust the pH to 6.3. L-Phenylalanine (20.20 kg) and
water (141.68 kg) were added, the slurry was heated to 75 °C,
and the clear solution was passed through a polish filter.
Deionized water at 75 °C (51.52 kg) was used to rinse the
filtering flask and then added to the mixture to adjust the
composition of the solvent to ∼2 vol % EtOH in solution. The
slurry was heated at 75 °C, and then the clear solution was
cooled to 55−60 °C, typically maintained at ∼57 °C. Seed
crystals of the complex 13 (322 g) were added, the slurry was
cooled to 40 °C over 1 h, and the pot temperature was
maintained at 40 °C for 4 h. The slurry was cooled further to
18−25 °C over 2 h and stirred at this temperature for 12−16 h.
The slurry was collected on a Robatel centrifuge filter. The
filter cake was washed with 10 °C water (322 kg) to remove the
inorganics, NaCl and NaOAc. The water wash was continued
until the conductivity of the wash was <0.001 σ siemens. The
filter cake was washed further with EtOAc (290 kg) to remove
any excess compound 1. The wet cake was dried under vacuum
at 18−25 °C for at least 4 h and then at 40 °C for at least 12 h.
The drying was stopped at KF reading = 2.8−3.6% (water that
corresponds to monohydrate). Complex 13 was obtained as a
white solid (54−58 kg, yield 80−88%, >99.90 A% purity): Anal.
Calcd for C30H39NO8: C, 66.52; H, 7.25; N, 2.58. Fd. C, 66.37;
REFERENCES
■
(1) (a) Ellsworth, B.; Washburn, W. N.; Sher, P. M.; Wu, G.; Meng,
W. Preparation of C-aryl glucoside SGLT-2 inhibitors. PCT Int. Appl.
WO 2001027128 A1 20010419, 2001 (b) Ellsworth, B. A.; Meng, W.;
Patel, M.; Girotra, R. N.; Wu, G.; Sher, P. M.; Hagan, D. L.;
Obermeier, M. T.; Humphreys, W. G.; Robertson, J. G.; Wang, A.;
Han, S.; Waldron, T. L.; Morgan, N. N.; Whaley, J. M.; Washburn, W.
N. Bioorg. Med. Chem. Lett. 2008, 18 (17), 4770.
(2) (a) Lewis, M. D.; Cha, K.; Kishi, Y. J. Am. Chem. Soc. 1982, 104,
4976. (b) Babirad, S. A.; Wang, Y.; Kishi, Y. J. Org. Chem. 1987, 52,
1370.
(3) Kraus, G. A.; Molina, M. T. J. Org. Chem. 1988, 53, 752.
(4) (a) Rajanikanth, B.; Seshadri, R. Tetrahedron Lett. 1989, 30 (6),
755. (b) Benhaddou, R; Czernecki, S; Farid, W; Ville, G; Xie, J; Zegar,
A. Carbohydr. Res. 1994, 260 (2), 243.
(5) (a) Gibson, M. S.; Vines, S. M.; Walthew, J. M. J. Chem. Soc.,
Perkin Trans. 1 1975, 155. (b) Olah, G. A.; Wang, Q.; Prakash, G. K. S.
Synlett 1992, 647.
(6) Crystallographic coordinates and unit cell data (excluding
structure factors) for 5, 6a, 11, 12, 13, and L-proline forms 1-P1, 1-
P2, 1-P3, 1-TP have been deposited with the Cambridge Crystallo-
graphic Data Centre as supplementary publication numbers CCDC
756030−756038. These data can be obtained free of charge via www.
ccdc.cam.ac.uk/const/retrieving.html (or from the CCDC, 12 Union
Road, Cambridge CB21EZ,U.K. Fax + 44 1223 336033. E-mail
deposit@ccdc.cam.ac.uk).
(7) Ellsworth, B. A.; Doyle, A. G.; Patel, M.; Carceres-Cortes, J.;
Meng., W.; Deshpande, P. P.; Pullockaran, A.; Washburn, W. N.
Tetrahedron: Asymmetry 2003, 14, 3243.
(8) The impurities were identified by LC/MS as a monobenzyl
derivative, over-reduced ring-opened product and traces of the α-
isomer. The crude product also showed oxidative instability over time.
(9) (a) Deshpande, P. P.; Ellsworth, B. A.; Buono, F. G.; Pullockaran,
A.; Singh, J; Kissick, T. P.; Huang, M.-H.; Lobinger, H.; Denzel, T;
Mueller, R. H. J. Org. Chem. 2007, 72 (25), 9746. (b) Horton, D;
Priebe, W. Carbohydr. Res. 1981, 94 (1), 27. (c) Deshpande, P. P.;
Ellsworth, B. A.; Singh, J.; Denzel, T. W.; Lai, C; Crispino, G;
Randazzo, M. E.; Gougoutas, J. Z. Methods of producing C-aryl
glucoside SGLT2 inhibitors. U.S. 20040138439 A1 20040715, 2004.
(10) (a) Barrett, A. G. M.; Pena, M.; Willardsen, A. J. J. Chem. Soc.,
Chem. Commun. 1995, 11, 1147. (b) Barrett, A. G. M.; Pena, M.;
Willardsen, A. J. J. Org. Chem. 1996, 61, 1082.
1
H, 7.37; N, 2.66. H NMR (400 MHz, CD3OD/D2O) δ 1.16
(t, 3H, J = 7.6 Hz, CH3), 2.57 (q, 4H, J = 7.5 Hz, CH2), 3.06
(dd, 1H, J = 8.5 Hz, J = 14.4 Hz, PhAl-CH2), 3.30 (m, 1H,
PhAl-CH2), 3.46 (m, 3H, H-2, H-3, H-4), 3.53 (q, 1H, J = 8.8
Hz, H-5), 3.69 (dd, 1H, 11.8 Hz, 5.6 Hz, H-6a), 3.85 (m, 2H,
H-6b, PhAl-HCOOH), 3.92 (s, 2H, Ph-CH2-Ph), 4.15 (d, 1H,
9.4 Hz, H-1), 4.71 (MeOH), 5.0 (1H, HOH), 7.10 (m, 4H, Ar-
(11) (a) Akiyama, T.; Takaya, J.; Kagoshima, H. Chem. Lett. 1999,
947−948. (b) Ribe, S.; Wipf, P. Chem. Commun. 2001, 299.
(12) Deslongchamps, P. Pure Appl. Chem. 1993, 65, 1161.
584
dx.doi.org/10.1021/op200306q | Org. Process Res. Dev. 2012, 16, 577−585

Organic Process Research & Development
Article
(13) Gougoutas, J. Z. Amino acid complexes of C-aryl glucosides for
treatment of diabetes. PCT Int. Appl. WO 2002083066 2002.
(14) Mutig, T; Kemnitz, E.; Troyanov, S. I. J. Flourine Chem. 2010,
131, 861.
(15) Dedicated issue on Pharmaceutical Co-crystal; please see Mol.
Pharm. 2007, 4, 299.
585
dx.doi.org/10.1021/op200306q | Org. Process Res. Dev. 2012, 16, 577−585