Process Development for a Locally Acting SGLT1 Inhibitor, LX2761, Utilizing sp3-sp2 Suzuki Coupling of a Benzyl Carbonate

Article abstract of DOI:10.1021/acs.oprd.8b00325

Process development for the synthesis of a locally acting sodium-dependent glucose cotransporter 1 (SGLT1) inhibitor, LX2761, a drug candidate for the treatment of diabetes mellitus, is described. A convergent route based on a retrosynthetic disconnection of the central diarylmethane linkage to a benzyl carbonate and a hindered arylboronic ester was designed and developed. The syntheses of these fully elaborated, late-stage crystalline intermediates are discussed, along with optimization of their union via an sp³-sp² Suzuki-Miyaura coupling reaction. The final active pharmaceutical ingredient was isolated in high purity as its l-proline cocrystal. This process was successfully performed on a kilogram scale to support toxicology and preclinical studies.

Full text of DOI:10.1021/acs.oprd.8b00325

Article
pubs.acs.org/OPRD
Cite This: Org. Process Res. Dev. XXXX, XXX, XXX−XXX
Process Development for a Locally Acting SGLT1 Inhibitor, LX2761,
Utilizing sp3−sp2 Suzuki Coupling of a Benzyl Carbonate
Lauren E. Sirois,† Matthew M. Zhao, Ngiap-Kie Lim,† Mark S. Bednarz,‡ Bryce A. Harrison,§
and Wenxue Wu*
Chemical Development, Lexicon Pharmaceuticals, Inc., 110 Allen Road, Basking Ridge, New Jersey 07920, United States
S
* Supporting Information
ABSTRACT: Process development for the synthesis of a locally acting sodium-dependent glucose cotransporter 1 (SGLT1)
inhibitor, LX2761, a drug candidate for the treatment of diabetes mellitus, is described. A convergent route based on a
retrosynthetic disconnection of the central diarylmethane linkage to a benzyl carbonate and a hindered arylboronic ester was
designed and developed. The syntheses of these fully elaborated, late-stage crystalline intermediates are discussed, along with
optimization of their union via an sp3−sp2 Suzuki−Miyaura coupling reaction. The final active pharmaceutical ingredient was
isolated in high purity as its L-proline cocrystal. This process was successfully performed on a kilogram scale to support
toxicology and preclinical studies.
KEYWORDS: dynamic kinetic resolution, anchimeric assistance, alkylzinc, sp3−sp2 Suzuki, cocrystal
causing excess urinary glucose excretion, which is thought to
be responsible for the increases in urinary tract infections seen
with SGLT2 inhibitors.5b Following the promising results with
LX2761 in animal models, kilogram quantities were required
to support additional preclinical and clinical studies.
INTRODUCTION
■
Recently, inhibitors of sodium-dependent glucose cotransport-
ers have attracted considerable attention for the treatment of
diabetes mellitus.1 Several sodium-dependent glucose cotrans-
porter 2 (SGLT2)-selective inhibitors, which lower blood
glucose levels by inhibiting reabsorption of glucose in the
kidneys, have advanced in the clinic and/or received regulatory
approval.2,3 Preclinical and clinical studies of our advanced
drug candidate LX4211 (sotagliflozin) (Figure 1), a dual
LX2761 shares the stereochemically rich thioxyloside motif
with LX4211, and the medicinal chemistry synthetic route was
based on a similar approach. Key intermediate 5 was
synthesized in a straightforward manner from the same
morpholinoamide 1 (Scheme 1).4,5 The subsequent synthetic
steps proved to be more challenging, however. The Heck
coupling approach required expensive methyl 3-butenoate (6)
as the coupling partner, microwave heating, purging of the
undesired branched isomer 7a, and olefin hydrogenation.
Global saponification followed by amidation with amine 8
using coupling reagent HATU gave LX2761. The high affinity
of LX2761 for aqueous media and the challenges of removing
coupling-reagent-related byproducts necessitated preparative
HPLC purification to obtain LX2761 with acceptable purity.
As the project advanced into development phases, a more
convergent and practical synthesis was required. Our revised
retrosynthetic analysis dissected the target at the central
diarylmethane linkage into two fragments of nearly equal
complexity (Scheme 2).
While numerous methods for constructing diarylmethane
motifs have been reported, including several used for the
syntheses of various SGLT2 inhibitors,2c the reaction
conditions for many of them (e.g., Friedel−Crafts acylation
or 1,2-addition followed by reduction)6 were not expected to
be compatible with fully functionalized substrates such as 9
and 10 and would require truncated intermediates or
inefficient protecting group manipulations. With the goal of
joining two fully functionalized fragments of LX2761, a less
Figure 1. SGLT1/2 dual inhibitor LX4211 and locally acting SGLT1
inhibitor LX2761.
inhibitor of SGLT2 (expressed primarily in the kidneys) and
SGLT1 (expressed primarily in the intestine), have shown
potential benefits of reducing or delaying glucose absorption in
the digestive tract.4 By exploiting this latter strategy, Lexicon’s
discovery efforts yielded a potent SGLT1-selective inhibitor,
LX2761 (Figure 1), that is locally acting in the gastrointestinal
system without appreciable systemic exposure.5 As the
pharmacological activity of LX2761 is largely independent of
kidney function, it has potential for the treatment of diabetic
patients with renal impairment. It also has the advantage of not
Received: October 2, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1. Medicinal Chemistry Synthesis of LX2761
based on knowledge gained from the LX4211 process.4,5 As
with the medicinal chemistry route, stereochemically pure
morpholinoamide 1 was used to construct this thioxyloside
fragment of LX2761. 2-Chloro-4-iodotoluene (13) was
selected as a suitable raw material allowing for sequential
functionalization of the aromatic ring. Treatment of 13 with i-
PrMgCl gave the aryl Grignard reagent, which upon addition
to amide 1, whose hydroxy group had been deprotonated with
t-BuMgCl, afforded ketone 14 in 80−90% isolated yield. It was
crucial to accurately charge the appropriate amounts of these
Grignard reagents to achieve high yields. Specifically, over-
charging i-PrMgCl led to its competitive attack on amide 1 to
give the isopropyl ketone side product.
Scheme 2. Retrosynthetic Analysis for the Process Route to
LX2761
The synthesis continued with a substrate-controlled stereo-
selective Luche reduction of ketone 14 in EtOH to afford diol
15 with a 96:4 ratio of diastereomers (Scheme 3). Diol 15 was
directly carried forward to the acid-promoted deprotection of
the acetonide and concomitant ring expansion to give tetraol
16 as a 1:1 anomeric mixture. The highly polar tetraol 16 was
not crystalline and thus telescoped to the acetylation step. The
global acetylation features a notable dynamic kinetic resolution
of the anomeric mixture of 16 to give the desired tetraacetate
anomer 17 with high stereoselectivity (>95:5 in solution,
>99:1 isolated) in high isolated yield (72% over three steps
from 14). The key for achieving the high stereoselectivity was
slow addition of acetic anhydride so that the rate of acetylation
was lower than the rate of anomeric equilibration of 16. The R
configuration of the anomeric center was confirmed by NMR
studies.
Treatment of tetraacetate 17 with thiourea and boron
trifluoride diethyl etherate followed by MeI and i-PrNEt pro-
vided thioether 20 with exceedingly high stereoselectivity
(>99:1).8 The high level of stereochemical control was
attributed to nearly exclusive nucleophilic attack of thiourea
from the bottom face of the pyranose moiety via the postulated
oxocarbenium ion 18 formed by anchimeric assistance from
the neighboring acetate group. The resulting thiourea adduct
conventional approach featuring a late-stage sp3−sp2 Suzuki
coupling of benzylic substrate 9 with arylboronate 10 was
investigated (Scheme 2). Admittedly, the proposed Suzuki
coupling pushed the limits of contemporaneous literature
precedents.7 However, as discussed below, intensive exper-
imentation ultimately yielded a scalable process route to
LX2761.
RESULTS AND DISCUSSION
■
Construction of the Thioxyloside and Elaboration to
the Arylboronic Ester. The synthesis of the “right-hand”
fragment of LX2761, pinacol boronic ester 10 (Scheme 3), was
B
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 3. Preparation of Fully Elaborated Arylboronate 10
a
19 was then converted to the free thiol and methylated with
MeI in situ to give 20. The latter was readily crystallized from
i-PrOH/water in 93% yield with >99% purity.
Table 1. Borylation of o-Methylaryl Chloride 20
The last step in the synthesis of key intermediate 10 entailed
borylation of o-methylaryl chloride 20. While borylation of
hindered aryl chlorides9 has continued to present a challenge
in organic synthesis, Buchwald and co-workers made
significant advances by using palladium catalysts with
biarylphosphine ligands.10 Initial results showed that 2-
dicyclohexylphosphino-2′,6′-dimethoxybiphenyl (SPhos) was
superior to 2-dicyclohexylphosphino-2′,4′,6′-triisopropylbi-
phenyl (XPhos) with regard to both conversion and
minimizing the dehalogenation side product 21 (Table 1).
While the reaction was faster at 80 °C than 60 °C, significantly
more impurity 21 formed. Further refinement identified that a
Pd loading of 2 mol % was required to complete the reaction at
60 °C (Figure 2). Implementing these conditions on a
kilogram scale afforded arylboronate 10 in 91% isolated yield
with ∼99% purity after treatment with activated carbon and
crystallization (Scheme 3).
Synthesis of the Fully Elaborated Benzyl Alcohol. The
first challenge in the synthesis of the “left-hand” fragment of
LX2761, compound 9, was the installation of a carbonyl-
functionalized aliphatic chain to an aryl ring. Although
compound 24a was available via Friedel−Crafts acylation11
of bromobenzene, the subsequent Wolff−Kishner12 (or Zn/
Hg) reduction was undesirable for scale-up because of safety
concerns (Scheme 4a). Other conditions (e.g., Et3SiH/TFA)13
afforded mainly the five-membered-ring lactone side product.
The required one-carbon homologation of aryl bromide 24a
HPLC yield (normalized area %)
a
entry
ligand
T (°C)
20
10
21
1
2
3
4
SPhos
SPhos
XPhos
XPhos
60
80
60
80
0
0
53
0
97
94
36
56
3
6
11
44
a
Conditions: 1.0 equiv of 20, 2.0 equiv of B2pin2, 3.0 equiv of KOAc,
3 mol % Pd(OAc)2, 6 mol % ligand, dioxane (0.2 M), 18 h.
would pose additional chemoselectivity challenges and
decrease the efficiency. Similar arguments precluded an
analogous toluene-based approach via 24b. Therefore,
commercially available 4-bromobenzyl alcohol (26a) was
explored14 as a potential starting material (Scheme 4b).
Although sequential hydroboration/sp3−sp2 coupling15 ap-
peared to be an attractive approach, it was only briefly explored
because of cost considerations and stability/reactivity/
selectivity issues of β,γ-unsaturated substrates 25.
Following reports by Knochel and others16 on Negishi17
couplings using highly active Pd−biarylphosphine systems
(some in the presence of active protons), the coupling of
C
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
bromobutyrate. Indeed, laboratory trials with many of these
reagents gave poor results, as determined by iodine titration
and/or GC (Table 2, entries 1−5).21
Encouragingly, when TMSCl22 (10 mol %) was added to the
reaction mixture with t-BuONa/CuI at ∼60 °C, a high
conversion was observed (Table 2, entry 6). Subsequent
experiments showed that TMSCl alone is sufficient and that t-
BuONa and CuI were not necessary (Table 2, entries 8 and
10−12). Additional experiments showed that high concen-
trations of 29 afforded faster reactions, but the reactions
became less clean at very high concentrations (>50% v/v).
Surprisingly, the alkylzinc reagent 28 prepared in this manner
was ineffective in the Negishi coupling reaction (Table 2, entry
8 vs 9), giving only 20% conversion versus 70% with
commercial material. The latter was presumably prepared
using Rieke zinc. Thus, we postulated that the presence of
lithium chloride, a byproduct in the preparation of Rieke zinc,
might be responsible for the observed differences.23 Indeed,
adding as little as 0.05 equiv of LiCl to our in-house-prepared
zinc reagent 28 greatly accelerated the Negishi coupling
reaction, affording 98% conversion versus 10% without LiCl at
the same time point (Table 2, entry 11 vs 10).24 In the latter
case, the color of the reaction mixture also turned black, an
indication of catalyst deactivation. The optimal charge of LiCl
was 0.5 equiv, allowing the reaction to reach full conversion in
∼2 h (Table 2, entry 12). While the precise role of LiCl is
unclear, its complexation with alkylzinc reagent 28 was quite
apparent: LiCl did not dissolve in pure THF but dissolved
readily in a THF solution of 28. Although TMSCl activation of
zinc dust worked well on a small scale with a magnetic stir bar,
on larger scales with mechanical stirring both TMSCl (5−10
mol %) and LiCl (∼0.5 equiv) were added to provide more
reliable zincation (Table 2, entry 13).25
Figure 2. Effect of catalyst loading on the borylation of aryl chloride
20. Conditions: 1.0 equiv of 20, 1.5 equiv of B2pin2, 3.0 equiv of
KOAc, Pd(OAc)2:SPhos ratio = 1:2, dioxane (8 volumes), 60 °C.
alcohol 26a with alkylzinc reagent 28 was attempted. With
Pd(OAc)2/SPhos, the reaction proceeded reasonably well,
providing 27b in ∼70% yield (Scheme 4c). Other common
catalysts such as Pd(OAc)2/PPh3, PdCl2(PPh3)2, and PdCl2·
dppf gave inferior results. Slow addition of excess alkylzinc
reagent 28 (1.75 equiv) was required to compensate for
competing protonation of 28. Since commercial 28 was
expensive and available in rather low concentration (0.5 M
solution), direct in situ preparation of 28 from ethyl 4-
bromobutyrate was pursued.
Several methods for generating highly active zinc or
activating zinc dust by breaking the oxide surface layer have
been described in the literature.18 The high cost of Rieke zinc
and inconvenience of acid/base washing of zinc dusts
persuaded us to explore other options. Several additives have
also been reported for in situ activation of zinc dusts by
“etching”, such as iodine, 1,2-dibromoethane, alkylmagnesium
halides, alkylaluminum halides, trimethylsilyl halides, diisobu-
tylaluminum hydride,19 and strong bases paired with metal
iodides (e.g., LHMDS/NiI2 or t-BuOK/CuI).20 Unfortunately,
these procedures appeared to be effective only for more active
substrates such as aryl, allyl, or benzyl halides or activated alkyl
halides as opposed to unactivated ones, such as ethyl 4-
Notably, protecting the benzyl alcohol group with a
trimethylsilyl group allowed us to reduce the amount of
alkylzinc 28 from 1.75 to 1.10 equiv and catalyst loading from
2 to 0.5 mol % (Table 2, entries 10−12). While it is generally
preferable to avoid protecting groups, the propensity of ester
27b to oligomerize (via transesterification with the benzyl
alcohol group) and chemoselectivity challenges in the
downstream amidation stage ultimately rendered it necessary
to protect the benzyl alcohol. For improved compatibility with
the downstream chemistry, TMS was replaced by a
Scheme 4. Approaches for Installing the Aliphatic Side Chain
D
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Table 2. Optimization of Zincation/Negishi Coupling
Scheme 5. Negishi Coupling of Alkylzinc to the Aryl
Bromide
zincation to give 28
conv.
Negishi coupling to give 27
conv.
a
b
c
entry
conditions
(%)
R
conditions
(%)
1
2
3
DIBAL-H
31
32
2
−
−
−
−
−
−
−
−
−
LHMDS, NiI2
Mg(OEt)2,
FeCl3
4
BrCH2CH2Br
(DMI solvent)
61
−
−
−
SPhos:Pd ratio from 2:1 to 4:1. Remarkably, when the
SPhos:Pd ratio was increased to 10:1, complete conversion was
achievable in 10−15 h at 65 °C even with 0.01 mol % Pd. A
palladium catalyst loading of 0.10 mol % was chosen to ensure
process robustness.
After quenching of excess alkylzinc 28 with ethanol, the
reaction mixture was concentrated under reduced pressure.
The residue was diluted with toluene, and after aqueous
workup, the crude product 27d solution was concentrated and
flushed with toluene to remove residual ethyl butyrate side
product (bp 121 °C). This avoided the formation of the
corresponding butyramide side product in the amidation step,
which not only consumed valuable amine 8 but also interfered
with the crystallization of the desired product 9a.
Although direct amidation of crude ester 27d would be more
efficient than hydrolysis of the ester followed by coupling with
amine 8, negligible amounts of 9a were observed after aging for
20 h at 70 °C in THF, MeTHF, or ACN or without a solvent.
Evidently, the sterically hindered amine 8 was not competent
in direct amidation of the ethyl ester. Therefore, crude ester
27d was diluted with methanol and treated with 1.5 equiv of
30% NaOH at 30−40 °C for 1−2 h to give carboxylic acid 11
as an oil, which was carried forward to the amidation step
without further purification (Scheme 5).
For the synthesis of amine 8, the initial route investigated
was a peptide coupling of N,N-dimethylethylenediamine (31)
with Boc-protected α-aminoisobutyric acid (AiB-OH) 30
(Scheme 6). Both conventional (DCC, CDI) and process-
friendly “green” propylphosphonic anhydride26 coupling re-
reagents were evaluated,27 ultimately leading to the selection of
CDI. The reaction was facile and clean, but purging the
imidazole byproduct was problematic. Since washing the
organic layer with aqueous acid was not feasible because of the
5
6
NaOtBu, CuI
19
−
−
−
−
−
−
NaOtBu, CuI,
TMSCl
100
7
8
NaOtBu, CuI,
TMSCl, LiCl
77
88
−
−
−
−
20
70
10
98
TMSCl
H
H
Pd(OAc)2 (2 mol %),
SPhos (4 mol %)
9
commercial 28
TMSCl
Pd(OAc)2 (2 mol %),
SPhos (4 mol %)
10
11
88
88
TMS Pd(OAc)2 (0.5 mol %),
SPhos (1.0 mol %)
TMSCl
TMS Pd(OAc)2 (0.5 mol %),
SPhos (1.0 mol %), LiCl
(0.05 equiv)
12
13
TMSCl
88
90
TMS Pd(OAc)2 (0.5 mol %),
SPhos (1.0 mol %), LiCl
(0.5 equiv)
99
TMSCl, LiCl
(0.5 equiv)
−
−
−
a
The reactions were conducted with zinc dust in the presence of
various additives from room temperature to 65 °C in THF, unless
otherwise noted. The conversions for zincation were determined by
iodine titration or GC of the aliquots quenched with acetic acid. The
conversions for Negishi coupling were determined by HPLC.
b
c
tetrahydropyranyl (THP) protecting group. This was readily
accomplished by treatment of 26a with 1.1 equiv of
dihydropyran and 2 mol % pyridinium p-toluenesulfonate
(PPTS) in THF at 50−60 °C. The reaction mixture was then
used directly in the Negishi coupling without quenching,
workup, or isolation (Scheme 5).
For the Negishi coupling, the order of addition proved to be
quite important. The palladium catalyst was readily deactivated
by alkylzinc 28 in the absence of aryl bromide 26c: the
reaction mixture turned dark rapidly with severely degraded
catalyst activity after 2 h at room temperature. Therefore, a
THF solution of alkylzinc 28 was slowly added to a mixture of
the catalyst and the aryl bromide. This controlled addition of
alkylzinc 28 is also an intrinsically safer process because it
moderates the rate of the highly exothermic reaction. The
catalyst loading was further reduced to 0.1 mol % by
scavenging residual oxygen in the reaction system with a
small amount of 28 prior to catalyst charge while increasing
Scheme 6. Initial Synthesis of Amine 8
E
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 7. Alternative Synthesis of Amine 8
Scheme 8. Preparation of Benzyl Alcohol 9b
dimethylamine group in the product, multiple water washes
and back extractions were employed. Removal of the Boc
protecting group by treatment with HCl (4 N in dioxane) in
acetonitrile afforded the bis(HCl) salt of 8 as a hygroscopic
solid. The latter was treated with powdered K2CO3 in
acetonitrile to give free-base 8 as an oil after the inorganic
salts were filtered off and the filtrate was concentrated. This
streamlined process obviated the difficulties of aqueous
workup of highly polar and water-soluble amine 8, affording
the product in 94% yield on a kilogram scale.
A more efficient route was later explored via direct reaction
of methyl ester of α-aminoisobutyric acid hydrochloride (AiB-
OMe·HCl, 32) with 31 (Scheme 7). Although the risks of
dimerization/oligomerization of 32 were expected to be low
because of the steric bulk next to the amino group, an excess of
31 (∼3 equiv) was used. The reaction was carried out at 100
°C without a solvent for 7 h and afforded the desired product 8
cleanly. After direct treatment of the reaction mixture with
solid potassium hydroxide, excess 31 was removed by
codistillation with toluene to afford 8 in nearly quantitative
yield. This new process obviated the use of dichloromethane
and CDI28 as well as the tedious aqueous workup in the initial
process.
and on scale-up, which posed an unacceptable risk for
manufacturing. Fortunately, it was discovered that a high
ratio of mixed anhydride/symmetrical anhydride was not
necessary for high conversion of the amidation. Even with 50%
symmetrical anhydride, the amidation still went to completion
upon aging at slightly higher temperature (35 °C). Reaction
monitoring by HPLC showed that amidation of the sym-
metrical anhydride was faster than that of the mixed anhydride.
Shortly after addition of amine 8 at −25 °C, the levels of
carboxylic acid 11 roughly matched the levels expected from
the symmetrical anhydride. In most other cases, the amidation
would stall at this stage. However, in this case, carboxylic acid
11 was still gradually converted into the desired product 9a.
These results can be rationalized as follows. There were
interconversions among mixed anhydride 33, symmetric
anhydride 34, and pivalic anhydride during the activation of
carboxylic acid 11 with pivaloyl chloride (Figure 3). Mixed
With both carboxylic acid 11 and amine 8 in hand,
amidation29 of 11 via its pivalic acid mixed anhydride was
evaluated. This approach offered easy purging of the pivalic
acid byproduct into the aqueous phase under basic conditions.
The sterically hindered α-amino group in 8 was expected to
enhance the regioselectivity. However, initial results showed
the formation of significant amounts of the undesired
symmetrical anhydride of 11. The ratio of the mixed anhydride
to symmetrical anhydride varied depending on the reaction
conditions such as base, solvent, temperature, stoichiometry,
order of addition, addition rate, and aging time. Adding a
solution of carboxylic acid 11 in THF/toluene to pivaloyl
chloride (1.1 equiv) at low temperature (−25 °C) helped to
improve the ratio of the mixed anhydride to symmetrical
anhydride (Scheme 8). However, the ratio decreased with time
Figure 3. Postulated mechanism for the amidation reaction.
anhydrides 33 and symmetric anhydride 34 rapidly reacted
with amine 8, but the bulky pivalic anhydride did not because
of the combined steric hindrance. On the other hand, carbox-
ylate 11 released from amidation of symmetric anhydride 34
was competent in reacting with pivalic anhydride to regenerate
mixed anhydride 33, which was then converted to the desired
product 9a. The key is the steric hindrance of the amine that
prevents it from reacting with pivalic anhydride.
After quenching with aqueous NaOH followed by workup,
the product (9a) was crystallized from a mixture of MTBE and
n-heptane. Among all of the solvents screened, MTBE was
F
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
uniquely effective in promoting the crystallization. This process
was successfully performed on a 3 kg scale batch, furnishing 9a
in 90% overall yield from 4-bromobenzyl alcohol 26a in four
steps.
Scheme 10. Synthesis of Carbonate 9c via Methyl
Chloroformate
The removal of the THP protecting group by treatment with
35% aqueous HCl (1.2 equiv) in methanol was facile because
of the formation of THP methyl acetal. The resulting
hydroxyamide, 9b, was found to be very water-soluble and
difficult to extract from aqueous media even after saturating it
with NaCl. Therefore, a more practical process was developed
to overcome this challenge. The reaction was quenched with
50% w/w NaOH, and the small amount of water was then
removed by concentration and flushing with methanol
followed by THF. The precipitated sodium chloride was
removed by filtration, and the product was isolated in 95%
yield with 99% purity by crystallization from a mixture of THF
and n-heptane.
With both substrates in hand, the next critical step was their
union via an sp3−sp2 Suzuki−Miyaura reaction.30 While not as
widespread as sp2−sp2 couplings, it has gained in popularity
thanks to advances in boron chemistry (e.g., trifluoroborate
salts) and ligand technology.31 An increasing number of
examples of constructing diarylmethane motifs via Suzuki
couplings of benzyl halides have been described in the
scientific literature.32 However, reports of benzyloxy sub-
strates,33 such as carbonates, acetates, and phosphates, were
more limited at the time of this work.
Strategically, we needed to identify a substrate with a
competent leaving group that is compatible with the amine−
diamide side chain, which precluded halides or similarly active
substrates (Scheme 9). Considering the noticeably greater
reactivity of benzyl carbonates34 over benzyl acetates,
carbonate 9c was selected for investigation.35
Considering the heightened risk in scale-up, a more robust
alternative was required.
N-Heterocyclic carbene (NHC)-catalyzed transesterification
of esters and, in limited examples, carbonates reported by
Nolan and co-workers36 prompted us to investigate milder and
less hazardous dimethyl carbonate (DMC)37 for the carbonate
formation. Surprisingly, with catalytic potassium tert-butoxide
as the base, similar conversions were observed with or without
an NHC catalyst such as IMes·HCl or ICy·HBF4. Additionally,
powdered molecular sieves used to remove generated
methanol were replaced by distillation, a much more practical
option for production (Figure 4). It was also found that
cheaper NaOMe performed similarly to potassium tert-
butoxide.
Scheme 9. Suzuki Coupling Strategy for the Diarylmethane
Linkage
The most straightforward approach to prepare carbonate 9c
was via treatment of benzyl alcohol 9b with methyl
chloroformate (Scheme 10). The reaction was surprisingly
sluggish, often requiring additional charges of methyl
chloroformate and pyridine to push the reaction to completion.
More alarmingly, various levels of quaternary ammonium salt
36 were observed. The culprit was either residual methyl
chloroformate or its degradation product, methyl chloride. In
one experiment, 36 was generated in ∼30% yield, most of
which formed during overnight aging of the organic layer.
Attempts to curtail this side reaction by adding diethylamine or
dipropylamine as sacrificial amines were not entirely effective.
Figure 4. Mechanism for the formation of carbonate 9c.
Mechanistically, the reaction mixture is composed of benzyl
alcohol 9b, benzyl carbonate 9c, dibenzyl carbonate 37,
methanol, DMC, sodium methoxide, and sodium benzyloxide
in a rapid equilibrium (Figure 4). Excess DMC favors desired
product 9c, while removal of methanol and DMC drives the
reaction toward 37. Thus, after the initial distillation, DMC
and THF were added, and the mixture was distilled and diluted
with DMC and THF until the desired conversion was
achieved. Typically, 5 mol % NaOMe was used, affording a
G
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
reaction mixture containing desired product 9c (90−95%)
along with starting material 9b and dibenzyl carbonate 37 (2−
5% each). Higher NaOMe charges were deleterious to the
reaction, converting product 9c and dibenzyl carbonate 37 to
sodium benzyloxide 9b-Na when MeOH and DMC were
removed by distillation (Figure 5).
delivering the product in 96% yield with ∼96% purity on a 3 kg
scale batch. Residual starting material 9b (∼2%) was readily
purged in the downstream Suzuki step, and dibenzyl carbonate
37 (2−3%) reacted to yield one molecule of the desired
product 35 and benzyl alcohol 9b.
Optimization of the sp3−sp2 Suzuki Coupling. With
both fully elaborated coupling partners in hand, optimization
of the Suzuki coupling step was performed (see Table 3).
Under conditions from the literature,33a,b the initial reaction
proceeded smoothly in DMF at 80 °C. While the overall
transformation was fairly clean, a major side product (5−10%),
imidazolinone 38, formed via cyclodehydration of the diamide
side chain (Scheme 11). Further screening of the reaction
conditions, including solvent, temperature, palladium precata-
lyst, ligand, and base, was conducted. Since imidazolinone
impurity 38 increased with temperature and aging time, the
reaction temperature range was limited to 50−70 °C. The
liability of the acetate groups to strong bases (e.g., Cs2CO3,
K3PO4) and nucleophiles (e.g., H2O, MeOH, EtOH, n-BuOH)
also limited the scope of reaction conditions. Key screening
results are summarized in Table 3.
Figure 5. Deleterious effect of overcharging NaOMe.
Considering the propensity of the carbonate moiety to
hydrolyze and the high aqueous affinity of 9c, an aqueous
workup was deemed impractical. Quenching the reaction with
acetic acid was convenient and facile, but the isolated product
9c did not perform well in the Suzuki coupling, which was
attributed to residual acetic acid. To overcome this
complication, Et3N·HCl was used to quench the reaction,
affording isolated carbonate 9c that performed well in the
Suzuki coupling reaction. Residual Et3N·HCl and Et3N
byproducts were removed by filtration and distillation,
respectively. Operationally, after the reaction was quenched
with Et3N·HCl, THF was added to dissolve the product. The
resulting slurry was filtered, and the filtrate was concentrated.
The solvent was exchanged to MTBE to crystallize carbonate
9c. n-Heptane was then added to increase the recovery,
As reported in the literature,33 several large-bite-angle
diphosphine ligands (e.g., DPPF, DPEphos) were active in
this transformation. 1,4-bis(diphenylphosphino)pentane
(DPPpentane) and 1,4-bis(diphenylphosphino)butane
(DPPbutane) showed the most consistent reactivity. The
latter was selected for further optimization because of cost
considerations. Precatalyst [Pd(allyl)Cl]2 was the most active
and chemoselective palladium source at 50 °C, while other
common palladium sources (e.g., Pd(OAc)2) required higher
temperatures (>70 °C) to achieve an acceptable conversion.
Reactions in aqueous cosolvent mixtures (∼20 vol % water)
generally gave poor impurity profiles, mostly due to hydrolysis
of the substrates and product. Among the anhydrous systems
tested, alcohols gave the most promising results. Although
Table 3. Screening of Reaction Conditions for the Suzuki Coupling
d
d
entry
solvent
DMF
ligand
time (h)
10 (area %)
35 (area %)
LX2761 (area %)
a
1
DPPpentane
2
20
2
20
2
20
2
20
2
20
2
20
2
20
18
18
18
18
18
50
14
67
55
68
14
90
0
2
0
92
30
45
0
49
84
32
43
31
83
10
98
91
0
−
a
2
MeCN
DPPpentane
DPPpentane
DPPpentane
DPPpentane
DPPpentane
DPPpentane
−
−
−
a
3
DME
a
4
toluene
a
5
EtOH
6
98
−
a
6
dioxane
tert-amyl alcohol
7
70
54
95
98
98
36
a
7
0
<2
−
−
−
b
8
9
10
11
i-PrOH
i-PrOH
i-PrOH
i-PrOH
i-PrOH
DPPbutane
DPPpentane
DPPhexane
DPEphos
DPPF
2
2
64
10
13
b
b
b
c
90
87
−
−
b
c
12
a
Conditions: [Pd(allyl)Cl]2 (5 mol %), DPPpentane (10 mol %), K2CO3 (3.0 equiv), solvent (0.15 M), 70 °C. Up to 5% cyclodehydration side
b
product was observed in all cases at 70 °C. Conditions: [Pd(allyl)Cl]2 (5 mol %), ligand (10 mol %), K2CO3 (3.0 equiv), i-PrOH (0.15 M), 50
c
d
°C. Other impurities were observed (HPLC). HPLC area percent (220 nm), relative distribution of starting material 10 to product 35 (except
where cyclodehydration or API is noted).
H
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 11. Suzuki Coupling of Fully Elaborated Substrates
EtOH performed most efficiently, it also readily cleaved the
acetate groups, giving LX2761 directly (Table 3, entry 5).
Unfortunately, the physicochemical properties of LX2761 were
such that isolation and purification of triacetate 35 was
preferred. Therefore, reaction optimization was continued with
i-PrOH as the solvent since the substrate(s) and the product
were sufficiently stable under the reaction conditions. When
the reaction scale was increased to >50 g, addition of 1−2
equiv of water helped to achieve satisfactory conversion.
Preforming the active catalyst complex by stirring [Pd-
(allyl)Cl]2 and DPPbutane in toluene ([Pd]:ligand = 1:1)
before it was added to the reaction mixture appeared to
enhance the robustness of the reaction. The resulting off-white
precipitate was palladium−allyl−ligand complex 39 (Figure
6)38 on the basis of preliminary HPLC−MS and 31P NMR
data.39 Addition of i-PrOH to the catalyst slurry in toluene
gave a homogeneous solution, which facilitated its air-free
transfer. Isopropanol could also possibly play a role in
nucleophilic activation of the Pd(II)−allyl complex to a
bisphosphine-ligated Pd(0) species.40
Ultimately, only 0.5 mol % [Pd(allyl)Cl]2 and 1.1 mol %
DPPbutane were required under the optimized reaction
conditions. A slight excess (1.1 equiv) of carbonate 9c was
employed to compensate for its hydrolysis (up to ∼5%) during
the reaction. The liability of the product’s acetate groups to
hydrolysis prevented the use of an aqueous workup. Therefore,
after completion of the reaction, acetone was added to dissolve
the product, and the mixture was treated with activated carbon
at 40 °C for 2−4 h to help remove the palladium catalyst.
Acetone was removed by vacuum distillation to crystallize
product 35 in 85% yield with excellent purity (>99 HPLC area
%)41 in our initial scale-up (two batches, 1.5 kg total).
With penultimate intermediate 35 available as a stable,
crystalline solid, the synthesis of LX2761 concluded with
removal of the acetate groups using catalytic sodium
methoxide in ethanol. At this critical juncture of the process,
the rather challenging physicochemical properties of the final
product were addressed. While LX2761 was not very soluble in
pure water, high losses to the aqueous layers were observed
during extraction. Only dichloromethane gave an acceptable
recovery for the workup/extraction. Additionally, it was
challenging to reproducibly obtain crystalline LX2761 directly
from the process stream while providing acceptable impurity
purging. These issues prompted a search for a more suitable
form of LX2761 such as a salt or cocrystal for isolation and
formulation.
Figure 6. Palladium−allyl−DPPbutane complex: (a) reaction
conditions; (b) HPLC (220 nm) after 30 min of stirring at room
temperature (toluene; MeCN diluent); (c) 31P NMR spectrum
overlay (CD2Cl2 diluent) of DPPbutane, DPPbutane monoxide,
DPPbutane dioxide, and 30 min catalyst sample.
A screening of approximately 40 pharmaceutically acceptable
acids yielded several leads; however, follow-up experiments
mostly gave tacky solids that were prone to deliquescence,
discouraging further exploration. Screening to discover a
cocrystal,42 with a particular focus on amino acids and other
common excipients/coformers with hydrogen-bonding capa-
bility, was then conducted. Gratifyingly, a well-defined
LX2761·L-proline43 cocrystal (1:1) was identified that
exhibited acceptable physicochemical properties.
I
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Thus, penultimate intermediate 35 was treated with catalytic
sodium methoxide in EtOH at 40 °C to give LX2761 as the
free base (Scheme 12). Separately, a solution of L-proline in
Acid-catalyzed deprotection of the acetonide and concomitant
ring expansion afforded six-membered-ring tetraol 16 as a 1:1
anomeric mixture. Dynamic kinetic resolution of 16 by
stereoselective acetylation and anomeric equilibration of the
tetraol afforded tetraacetate 17 as a single stereoisomer in high
yield. Boron trifluoride-promoted thiolation gave triacetate 20
with >99:1 stereoselectivity attributed to neighboring group
assistance. Palladium-catalyzed borylation of the hindered aryl
chloride 20 furnished arylboronic ester 10.
Scheme 12. Conversion to LX2761 L-Proline Cocrystal
For the synthesis of benzyl carbonate 9c, a highly efficient
Negishi coupling reaction of aryl bromide 26c with in situ-
prepared alkylzinc 28 was employed to install the aliphatic side
chain. After saponification, carboxylic acid 11 was coupled with
amine 8 via its pivalic mixed anhydride. After removal of the
THP protecting group, benzyl alcohol 9b was converted to
benzyl carbonate 9c using nontoxic dimethyl carbonate.
Joining benzyl carbonate 9c and arylboronic ester 10 via the
key sp3−sp2 Suzuki coupling afforded the final intermediate 35.
Global deprotection followed by cocrystallization with L-
proline delivered high-quality LX2761 active pharmaceutical
ingredient (API). This synthetic process was successfully
performed on a kilogram scale to meet the initial demands of
drug development.
EtOH/water was prepared and charged in portions to the
above reaction mixture. Upon addition of MTBE as an
antisolvent and cooling, the LX2761·L-proline cocrystal
crystallized smoothly. This process was performed successfully
on a kilogram scale in 91% yield with >99% purity (HPLC area
%).
EXPERIMENTAL SECTION
■
General. All reagents and materials were used as received.
Unless additional GC or GC−MS analysis was necessary,
reactions were monitored by reverse-phase HPLC, generally
using a C18 or phenyl-hexyl column with water/MeCN or
water/MeOH as the mobile phase and trifluoroacetic acid
(TFA) as a modifier. Melting point information was collected
using a differential scanning calorimeter (peak temperature).
NMR spectra were acquired in deuterated solvents. Mass
spectrometry data were obtained during LC−MS analysis.
Compound purity was assessed by reverse-phase HPLC and/
CONCLUSIONS
■
A convergent process route to SGLT1 inhibitor LX2761 was
developed (Scheme 13). It features a late stage sp3−sp2 Suzuki
coupling of benzyl carbonate 9c with sterically hindered
arylboronic ester 10. The optimization of this key reaction and
the challenges overcome in assembling the two fully function-
alized Suzuki coupling partners have been discussed.
1
or H NMR analysis.
The synthesis of stereochemically rich arylboronic ester 10
entailed aryl Grignard addition to L-xylose-derived, enantio-
merically pure morpholinoamide 1 to give ketone 14. Luche
reduction of 14 gave diol 15 with 96:4 diastereoselectivity.
(3-Chloro-4-methylphenyl)((3αS,5R,6S,6αS)-6-hy-
droxy-2,2-dimethyltetrahydrofuro[2,3-d][1,3]dioxol-5-
yl)methanone (14). tert-Butylmagnesium chloride (∼1 M in
THF, 5.45 L, 1.15 equiv vs 1) was added to a solution of
Scheme 13. Summary of the Process Route to LX2761
J
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
morpholinoamide 1 (1.25 kg, 4.56 mol assay-corrected) in
anhydrous THF (4.5 L) at −15 to −3 °C, and the resulting
solution was held at −15 °C. In a separate reactor,
isopropylmagnesium chloride in THF (2.65 L, 2.11 M, 5.47
mol, 1.09 equiv vs 13) was added to a solution of aryl iodide
13 (1.29 kg, 5.01 mol, 1.10 equiv vs 1) in anhydrous THF (3.5
L) at −16 to −8 °C, and the mixture was stirred until
transmetalation was complete (IPC by GC after 1−2 h of
aging). The aryl Grignard solution was then added to the
solution of deprotonated morpholinoamide 1 at or below −3
°C. The reaction mixture was maintained at −10 °C until the
reaction was complete and then quenched into a cold (∼5 °C)
solution of citric acid (1.08 kg) in water (6.75 L) at or below
13 °C. The resulting biphasic mixture was agitated for 15 min
at 15−20 °C. The upper organic layer was separated and
washed twice with 25% brine (2 × 2.5 L). The washed organic
phase was concentrated to ∼4 L, and the resulting suspension
was warmed to 50−55 °C. n-Heptane (6 L) was added, and
the mixture was cooled slowly to 0 °C, aged for 1 h, and
filtered. The filter cake was washed with an n-heptane/THF
mixture (4:1, 2.5 L) followed by n-heptane (2 L). The wet
cake was dried under reduced pressure at 40−45 °C with a
gentle nitrogen sweep to afford 1.30 kg (91% yield) of ketone
14 with >99% purity (HPLC area %). mp: 153 °C (DSC peak
temperature). LC−MS: calcd [M + H]+ 313, found m/z 313.
1H NMR (400 MHz, DMSO-d6): δ 1.28 (s, 3H), 1.48 (s, 3H),
2.40 (s, 3H), 4.44 (t, J = 3.8 Hz, 1H), 4.48 (d, J = 3.8 Hz, 1H),
5.52 (d, J = 3.5 Hz, 1H), 5.59 (d, J = 5.0 Hz, 1H), 6.02 (d, J =
3.5 Hz, 1H), 7.51 (d, J = 8.3 Hz, 1H), 7.80 (dd, J = 7.9, 1.6 Hz,
1H), 7.89 (d, J = 1.8 Hz, 1H). 13C NMR (101 MHz, DMSO-
d6): δ 19.8, 26.2, 26.8, 76.1, 84.3, 85.1, 104.6, 111.3, 126.8,
128.1, 131.4, 133.6, 135.6, 140.9, 193.2.
(2R,3S,4R,5S,6S)-6-(3-Chloro-4-methylphenyl)-
tetrahydro-2H-pyran-2,3,4,5-tetrayl Tetraacetate (17).
Luche Reduction To Give Diol 15. A solution of sodium
borohydride (93.95 g, 2.48 mol) in 1.0 N sodium hydroxide
(582 mL) was added to a mixture of ketone 14 (2.00 kg, 6.40
mol) and cerium trichloride heptahydrate (1.19 kg, 3.20 mol,
0.50 equiv) in ethanol (14 L) at −20 °C over 1.5 h and aged
(diastereomeric ratio of diol 15 = 96:4). The reaction mixture
was warmed to 0 °C, quenched with water (8 L) at ≤11 °C,
and concentrated to ∼11 L. Ethyl acetate (8 L) was added, and
the pH of the mixture was adjusted to 2.5 with 6 N HCl (∼535
mL). The organic layer was separated, washed sequentially
with 0.25 N sodium hydroxide (2 × 4.5 L) and 25% brine (4
L), distilled to a low volume, and flushed with acetonitrile (12
L) to give a solution of diol 15 (∼9.5 L).
with water (18 L) at ≤30 °C. The resulting suspension was
cooled to 10−15 °C, aged, and filtered. The filter cake was
washed sequentially with 2-propanol (3 × 4 L) and n-heptane
(4 L) and dried under reduced pressure at 40−45 °C to give
2.11 kg (72% yield over three steps from 14, corrected for
97.1% assay) of tetraacetate 17 as an off-white solid with
∼100% purity (HPLC area %). mp: 181 °C (DSC peak
temperature). LC−MS: calcd [M + NH4]+ 460, found m/z
460. 1H NMR (700 MHz, DMSO-d6): δ 1.80 (s, 3H), 1.96 (s,
3H), 2.04 (s, 3H), 2.08 (s, 3H), 2.32 (s, 3H), 4.92 (d, J = 9.7
Hz, 1H), 5.16 (t, J = 9.6 Hz, 1H), 5.21 (dd, J = 9.7, 8.4 Hz,
1H), 5.49 (t, J = 9.6 Hz, 1H), 6.03 (d, J = 8.4 Hz, 1H), 7.23
(dd, J = 7.8, 1.3 Hz, 1H), 7.33 (d, J = 7.8 Hz, 1H), 7.45 (s,
1H). 13C NMR (101 MHz, DMSO-d6): δ 19.4, 20.1, 20.26,
20.31, 20.5, 70.0, 71.82, 71.85, 74.4, 91.1, 126.2, 127.5, 131.0,
133.1, 135.7, 135.9, 168.5, 168.8, 169.1, 169.5.
(2S,3S,4R,5S,6R)-2-(3-Chloro-4-methylphenyl)-6-
(methylthio)tetrahydro-2H-pyran-3,4,5-triyl Triacetate
(20). Boron trifluoride diethyl etherate (1.46 L, 2.50 equiv)
was added to a mixture of tetraacetate 17 (2.09 kg, 4.71 mol)
and thiourea (394 g, 5.18 mol, 1.1 equiv) in ethyl acetate (14
L), and the mixture was aged at 55 °C for 4 h to give thiourea
adduct 19. Methanol (4 L) and methyl iodide (358 mL, 1.22
equiv) were added, and then N,N-diisopropylethylamine (3.59
kg, 5.9 equiv) was slowly added below 15 °C. The reaction
mixture was aged at 20−25 °C until reaction completion,
concentrated to 8−9 L under reduced pressure, and flushed
with isopropyl alcohol (IPA) in portions (8−9 L) to a final
volume of 14.5 L. Water (11 L) was added, and the suspension
was aged at 35 °C for 0.5 h and then at 10 °C for 4 h. The
product was collected by filtration and washed sequentially
with IPA/water (2:1, 6 L), IPA (3 L), and n-heptane (4 L).
Drying under reduced pressure at 40−45 °C gave 1.88 kg
(92.7% yield) of aryl chloride 20 as an off-white solid with
>99% purity (HPLC area %). mp: 170 °C (DSC peak
temperature). LC−MS: calcd [M + NH4]+ 448, found m/z
448. 1H NMR (700 MHz, DMSO-d6): δ 1.79 (s, 3H), 1.96 (s,
3H), 2.05 (s, 3H), 2.13 (s, 3H), 2.32 (s, 3H), 4.73 (d, J = 9.7
Hz, 1H), 4.89 (d, J = 9.9 Hz, 1H), 5.14 (t, J = 9.6 Hz, 1H),
5.18 (t, J = 9.7 Hz, 1H), 5.37 (t, J = 9.4 Hz, 1H), 7.24 (dd, J =
7.8, 1.5 Hz, 1H), 7.33 (d, J = 8.0 Hz, 1H), 7.46 (s, 1H). 13C
NMR (101 MHz, DMSO-d6) δ 10.5, 19.4, 20.1, 20.3, 20.4,
68.8, 72.1, 73.1, 77.4, 81.3, 126.1, 127.5, 131.0, 133.1, 135.7,
136.5, 168.5, 169.1, 169.5.
(2S,3S,4R,5S,6R)-2-(4-Methyl-3-(4,4,5,5-tetramethyl-
1,3,2-dioxaborolan-2-yl)phenyl)-6-(methylthio)-
tetrahydro-2H-pyran-3,4,5-triyl Triacetate (10). A mix-
ture of aryl chloride 20 (2.00 kg, 4.64 mol, 1.0 equiv),
bis(pinacolato)diboron (1.76 kg, 6.95 mol, 1.5 equiv),
potassium acetate (1.37 kg, 14.0 mol, 3.0 equiv), and dioxane
(16 L) was degassed via four vacuum/nitrogen fill cycles.
SPhos (76.2 g, 0.186 mol, 4.0 mol %) and palladium acetate
(20.44 g, 0.091 mol, 2.0 mol %) were added, and the mixture
was degassed (two cycles) and then aged at 60 °C until
reaction completion (28 h). The reaction mixture was cooled
to 20−25 °C, diluted with isopropyl acetate (IPAc) (10 L),
stirred for 15 min, and then filtered through a pad of silica gel
(4 kg). The silica gel pad was washed with IPAc, and the
combined filtrate was concentrated under reduced pressure to
∼2.4 L. The distillation residue was diluted to 20 L with IPAc
and treated with Darco G-60 (100 g) for 3 h at 50 °C. The
mixture was filtered through a pad of diatomite, and the filter
cake was washed with IPAc (2.25 L). The combined filtrate
Deprotection To Give Tetraol 16. Water (3.5 L) and 6 N
HCl (260 mL) were charged to the solution of diol 15, and the
mixture was aged for 3.5 h at 65−70 °C. After complete
conversion, the reaction mixture was cooled to 25 °C and
extracted with 2-MeTHF (10 L). The organic phase was
washed with 25% brine (2 × 4 L), concentrated to ∼5 L, and
then codistilled with acetonitrile in portions (34 L total)
until residual water was <0.5% (as determined by Karl Fischer
analysis). The precipitate was removed by filtration through a
pad of diatomite to give a solution of tetraol 16 in acetonitrile.
Acetylation To Give Tetraacetate 17. The solution of
tetraol 16 in acetonitrile was diluted to ∼11.5 L with
acetonitrile, and then triethylamine (7.15 L, 8.0 equiv) was
added. Acetic anhydride (3.62 L, 6.0 equiv) was added slowly
(5−10 h) at 35 °C. The reaction mixture was stirred until
reaction completion (14 h), cooled to 15 °C, and quenched
K
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
was concentrated to ∼4.5 L to give a pale-yellow suspension. n-
Heptane (18 L) was added slowly at 50 °C, and the resulting
suspension was aged, cooled to 5−15 °C, aged, and filtered.
The filter cake was washed with n-heptane (2 × 4 L) and dried
at 45−50 °C under reduced pressure to give 2.23 kg (91%
bromobutanoate (29) (1.80 kg, 1.32 L, 9.23 mol, 1.0 equiv)
in THF (2.5 L) under nitrogen. The mixture was heated to 55
°C over 0.3 h, aged at 50−65 °C for 2 h (exothermic!) and then
at 60−65 °C for 19 h, cooled to 25 °C, and allowed to settle.
The supernatant (∼4 L) was used for the Negishi coupling
directly.
1
yield corrected for 98.7% H NMR assay) of arylboronic ester
10 as a white solid. mp: 149 °C (DSC peak temperature).
THP Protection of 4-Bromobenzyl Alcohol To Give 26c. A
mixture of 4-bromobenzyl alcohol (26a) (1.16 kg, 6.20 mol,
1.00 equiv), pyridinium p-toluenesulfonate (32.0 g, 0.127 mol,
0.02 equiv), THF (3.7 L), and 3,4-dihydropyran (0.571 kg,
6.79 mol, 1.10 equiv) was aged at 50 °C under nitrogen until
reaction completion. The reaction mixture was cooled and
directly used in the Negishi coupling.
1
LC−MS: calcd [M + NH4]+ 540, found m/z 540. H NMR
(400 MHz, DMSO-d6): δ 1.27−1.34 (br s, 12H), 1.76 (s, 3H),
1.95 (s, 3H), 2.05 (s, 3H), 2.11 (s, 3H), 2.44 (s, 3H), 4.74 (d, J
= 9.8 Hz, 1H), 4.92 (d, J = 10.0 Hz, 1H), 5.04 (t, J = 9.7 Hz,
1H), 5.11 (t, J = 9.7 Hz, 1H), 5.35−5.41 (m, 1H), 7.17 (d, J =
8.0 Hz, 1H), 7.37 (d, J = 7.5 Hz, 1H), 7.57 (d, J = 1.8 Hz, 1H).
13C NMR (101 MHz, DMSO-d6): δ 10.7, 20.1, 20.3, 20.4,
21.6, 24.6, 24.7, 69.1, 72.1, 73.2, 78.0, 81.5, 83.4, 129.3, 129.8,
133.0, 134.7, 144.4, 168.5, 169.1, 169.5. Anal. Calcd: C, 57.48;
H, 6.75; S, 6.14. Found: C, 57.42; H, 6.51; S, 6.13.
Negishi Coupling To Give 27d. To the above solution of
26c was added 0.4−0.5 L of the solution of alkylzinc 28, and
the mixture was stirred for 0.5 h at 35−40 °C under nitrogen.
Meanwhile, a solution of catalyst was prepared by dissolving
Pd(OAc)2 (1.39 g, 6.19 mmol, 0.10 mol %) and SPhos (10.2 g,
24.7 mmol, 0.40 mol %) in THF (62 mL) at 20 °C under
nitrogen. The catalyst solution was charged into the above
reaction mixture (exothermic!). After the initial exotherm
subsided, the remaining alkylzinc 28 was added at 45−60 °C
over 1−2 h. The mixture was aged at 60 °C until reaction
completion, cooled to 20 °C, quenched with EtOH (359 mL,
1.0 equiv), and aged for 0.5 h. The batch was then
concentrated to 3−4 L and diluted with toluene (5 L), water
(3.7 L), and 50% aqueous citric acid (150 mL). The organic
layer was separated, washed with water (2 × 2.3 L),
concentrated to ∼3 L, and flushed with toluene (1.3 L × 3)
to remove residual ethyl butanoate (as monitored by GC) to
give crude 27d (2.20 kg).
2-Amino-N-(2-(dimethylamino)ethyl)-2-methylpropa-
namide (8). Boc-α-aminoisobutyric acid (Boc-AiB-OH) (30)
(3.00 kg, 14.77 mol, 1.00 equiv) was added in portions over
0.5 h to a suspension of 1,1′-carbonyldiimidazole (CDI) (2.56
kg, 15.79 mol, 1.07 equiv) in dichloromethane (21 L) at 20−
30 °C. After 0.5 h of stirring at 20−25 °C, N,N-
dimethylethylenediamine (1.56 kg, 17.72 mol, 1.20 equiv)
was added over 1 h at <30 °C. The mixture was stirred for 0.5
h and then washed with 7.5% aqueous sodium bicarbonate (2
× 15 L) followed by water (2 × 15 L). The aqueous layers
were back-extracted with dichloromethane (15 L). The
combined organic layer was concentrated to ∼6 L, flushed
with acetonitrile (12 L in portions), and then diluted with
acetonitrile (24 L). HCl (4.0 N in dioxane, 9.2 L, 36.8 mol,
2.50 equiv) was added over 2 h at <40 °C (exothermic), and
the resulting thick suspension was stirred at 20−25 °C.
Potassium carbonate (5.10 kg, 36.9 mol, 2.50 equiv) was added
over 0.5 h (CO2 evolution), and the mixture was aged at 30 °C
for 16 h. Upon cooling to 20 °C, the slurry was filtered, and the
filter cake washed with acetonitrile (9 L). The combined
filtrate was concentrated to give 3.04 kg (93.6% yield corrected
for 78.9 wt % assay) of amine 8 as a brown oil (residual
solvents: 10.3 wt % acetonitrile, 9.4 wt % dioxane). LC−MS:
Saponification of 27d to 11. The crude ester 27d (2.20 kg)
was diluted with MeOH (1 L), and then 30% NaOH (0.92 L,
1.50 equiv) was added over 0.3 h (exothermic!). After 0.5 h of
aging at 30−40 °C, toluene (2.30 L) and water (2.30 L) were
added, and the resulting mixture was cooled to 20 °C. The
aqueous layer was separated, mixed with toluene (3.5 L),
cooled to 2 °C, and slowly acidified with 6 N HCl (∼1.45 L)
to pH 4.1 at <5 °C with vigorous stirring. The organic layer
was separated, washed with water (1.2 L) and 1:1 brine/water
(1.2 L), concentrated to a low volume, and then flushed with
toluene (∼3 L in portions) to give crude carboxylic acid 11
1
calcd [M + H]+ 174, found m/z 174. H NMR (400 MHz,
CDCl3): δ 1.35 (s, 6H), 2.24 (s, 6H), 2.42 (t, J = 6.3 Hz, 2H),
3.30 (q, J = 6.0 Hz, 2H), 7.79 (br s, 1H). 13C NMR (101 MHz,
CDCl3): δ 29.2, 36.9, 45.3, 54.7, 58.2, 177.5.
1
(∼3 L). H NMR (700 MHz, CDCl3): δ 1.53−1.69 (m, 4H),
1.72−1.79 (m, 1H), 1.85−1.92 (m, 1H), 1.97 (quin, J = 7.5
Hz, 2H), 2.38 (t, J = 7.4 Hz, 2H), 2.68 (t, J = 7.6 Hz, 2H),
3.56−3.60 (m, 1H), 3.95 (ddd, J = 11.4, 8.9, 3.0 Hz, 1H), 4.49
(d, J = 12.0 Hz, 1H), 4.73−4.75 (m, 1H), 4.78 (d, J = 7.0 Hz,
1H), 7.18 (d, J = 8.0 Hz, 2H), 7.31 (d, J = 8.0 Hz, 2H), 9−12
(br, 1H). 13C NMR (101 MHz, CDCl3): δ 19.2, 25.4, 26.2,
30.5, 33.2, 34.6, 62.0, 68.6, 97.6, 128.0, 128.4, 135.8, 140.5,
179.4.
Alternative Preparation of Amine 8. A mixture of N,N-
dimethylethylenediamine (4.0 mL, 37.2 mmol, 2.9 equiv) and
methyl α-amino isobutyrate hydrochloride (AiB-OMe·HCl)
(32) (2.00 g, 13.0 mmol, 1.0 equiv) was aged at 100 °C for 7 h
and then cooled to room temperature. Toluene (6 mL) and
potassium hydroxide pellets (1.0 g, 15.2 mmol) were added
and the mixture was aged at 50 °C for 6 h, cooled to room
temperature, and filtered through a pad of Celite. The Celite
was rinsed with toluene (3 × 6 mL) and the combined filtrate
concentrated at 50 °C under reduced pressure to give amine 8
as a yellow oil (2.25 g, 97% yield corrected for assay) (∼5 mol
N-(1-((2-(Dimethylamino)ethyl)amino)-2-methyl-1-
oxopropan-2-yl)-4-(4-(((tetrahydro-2H-pyran-2-yl)oxy)-
methyl)phenyl)butanamide (9a). Crude carboxylic acid 11
in toluene (2.64 kg, 6.20 mol theoretical) was diluted with
toluene (0.6 L) and THF (0.6 L), and then Et3N (1.05 L, 7.53
mol, 1.22 equiv) was added. The mixture was charged slowly
(2 h) to a solution of pivaloyl chloride (823 g, 6.83 mol, 1.10
equiv) in toluene (2.6 L) and THF (0.86 L) at −20 to −30 °C.
The reaction mixture was aged for 0.3 h at −25 °C, and then
Et3N (0.822 kg, 8.12 mol, 1.32 equiv) was added, followed by
amine 8 (1.48 kg, 87 wt %, 1.28 kg active, 7.43 mol, 1.20
equiv) at −20 to −25 °C. The mixture was aged for 0.5 h at
1
% residual toluene by NMR); H NMR (700 MHz, DMSO-
d6): δ 1.17 (s, 6H), 1.87 (br s, 2H), 2.15 (s, 6H), 2.29 (t, J =
6.6 Hz, 2H), 3.13 (q, J = 6.5 Hz, 2H), 7.88 (br s, 1H).
4-(4-(((Tetrahydro-2H-pyran-2-yl)oxy)methyl)-
phenyl)butanoic Acid (11). Preparation of Alkylzinc 28.
TMSCl (50.1 g, 58.6 mL, 0.462 mol, 0.05 equiv) was added to
a mixture of anhydrous LiCl (196 g, 4.61 mol, 0.50 equiv), zinc
dust (726 g, 11.1 mol, 1.20 equiv), and ethyl 4-
L
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
−25 °C and then at 35 °C until reaction completion (16 h).
The reaction mixture was cooled to 15 °C, quenched with
water (1.2 L), basified with 30% NaOH (1.26 L) to pH 12.5 at
15−25 °C, and then diluted with water (3.5 L). The organic
layer was separated and washed with 1:1 brine/water (3.5 L ×
2), and the aqueous layers were back-extracted with toluene/
THF (1.0/0.3 L). The combined organic phase was
concentrated to 3−4 L and flushed with toluene (2 L × 3)
to give 3.18 kg of crude amide 9a (maintained at ≥45 °C to
prevent premature crystallization). MTBE (6.5 L) was added,
and the mixture was aged at 40 °C for 0.5 h, cooled to 30 °C,
and aged for 1 h. After addition of n-heptane (10 L) and aging
at 45 °C, the suspension was cooled to 15 °C over 2−3 h and
aged for 4 h. The product was filtered, washed with 1:1
MTBE/heptane (12 L), and dried at 45 °C under reduced
pressure to give 2.41 kg (89.8% overall yield from 26a) of 9a as
a white solid with 98.6% purity (HPLC area %). mp: 88 °C
(DSC peak temperature). LC−MS: calcd [M + H]+ 434, found
slowly distilled under slightly reduced pressure. After 2−3 L
had been distilled (1.5 h), THF (2 L) and dimethyl carbonate
(2 L) were added, and the distillation was continued, followed
by dilution with DMC (∼4 L; an excess of DMC is important
to convert 37 to 9c). The distillation and dilution operations
were repeated until satisfactory conversion was achieved: 1.2%
9b, 95.7% 9c, 3.2% 37. The reaction mixture was quenched
with Et3N·HCl (96 g, 0.697 mol, 0.097 equiv), aged at 20−25
°C, and then concentrated to ∼5 L. THF (5 L) was added to
dissolve the product, and the resulting thin slurry was filtered
through a pad of diatomite. The filter cake was washed with
THF (2 L), and the combined filtrate (∼13 L) was
concentrated to ∼6 L. MTBE (8 L) and seed crystals (∼10
g) were added, and the mixture was aged at 40 °C for 0.5 h. n-
Heptane (12 L) was slowly added, and the mixture was aged at
40 °C for 2−4 h and then at 10 °C for 2 h. The crystallized
product was filtered, washed with 2:1 heptane/MTBE (12 L),
and dried at 40 °C under reduced pressure to give 2.82 kg
(96.5% yield) of benzyl carbonate 9c as a white solid with
95.9% purity (HPLC area %). mp: 92 °C (DSC peak
temperature). LC−MS: calcd [M + H]+ 408, found m/z
408. 1H NMR (400 MHz, CDCl3): δ 1.55 (s, 6H), 1.94 (quin,
J = 7.5 Hz, 2H), 2.18 (t, J = 7.5 Hz, 2H), 2.22 (s, 6H), 2.43 (t,
J = 6.0 Hz, 2H), 2.64 (t, J = 7.5 Hz, 2H), 3.32 (q, J = 5 Hz,
2H), 3.78 (s, 3H), 5.12 (s, 2H), 6.28 (br s, 1H), 6.79 (br s,
1H), 7.18 (d, J = 7.6 Hz, 2H), 7.30 (d, J = 8.0 Hz, 2H). 13C
NMR (101 MHz, CDCl3): δ 25.1, 26.9, 34.8, 36.3, 37.0, 45.0,
54.8, 56.9, 57.6, 69.5, 128.5, 128.7, 132.8, 142.0, 155.7, 172.2,
174.6.
(2S,3S,4R,5S,6R)-2-(3-(4-(4-((1-((2-(Dimethylamino)-
ethyl)amino)-2-methyl-1-oxopropan-2-yl)amino)-4-
oxobutyl)benzyl)-4-methylphenyl)-6-(methylthio)-
tetrahydro-2H-pyran-3,4,5-triyl Triacetate (35). A mix-
ture of arylboronate 10 (651 g, 1.25 mol, 1.00 equiv), benzyl
carbonate 9c (560 g, 1.37 mol, 1.10 equiv), and potassium
carbonate (325 mesh, 518 g, 3.75 mol, 3.0 equiv) in IPA (6.5
L) was degassed by three vacuum/nitrogen fill cycles. Water
(22.4 mL, 1.24 mol, 1.0 equiv) was slowly added with stirring,
and the suspension was heated to 50 °C. The catalyst solution
(prepared separately) was then added, and the reaction
mixture was aged at 50 °C until reaction completion (10−20
h).
Preparation of the Catalyst Solution. Degassed toluene
(130 mL) and IPA (65 mL) were added to [Pd(allyl)Cl]2
(2.28 g, 6.22 mmol, 1.0 mol % Pd) and 1,4-diphenylphosphi-
nobutane (DPPbutane) (5.84 g, 13.7 mmol, 1.1 mol %) under
nitrogen, and the mixture was stirred at 20−25 °C for 10−20
min to give a pale-yellow-orange solution (which had to be
used within 1 h for best results).
Improved Catalyst Preparation. [Pd(allyl)Cl]2 and
DPPbutane were combined in degassed toluene at ambient
temperature and stirred for 0.5 h to give a suspension.
Immediately before transfer to the reaction mixture, degassed
IPA was added to form a yellow-orange solution.
After the above Suzuki coupling reaction was complete,
acetone (8.0 L) and Darco G-60 (32.5 g) were added, and the
mixture was stirred at 40 °C for 2−4 h. The slurry was filtered
through a thin pad of diatomite powder, and the filter cake was
rinsed with a warm 3:2 acetone/IPA solution (4 L). The
combined filtrate and wash was concentrated to 9−10 L at 40
°C under reduced pressure, seeded with 35 (3.03 g), stirred for
0.5−1 h, concentrated to ∼10 L, flushed with IPA (3 L), aged
at 40 °C for 2 h, room temperature for 2 h, and then filtered.
1
m/z 434. H NMR (700 MHz, CD3OD): δ 1.42 (s, 6H),
1.51−1.62 (m, 4H), 1.68−1.75 (m, 1H), 1.82−1.92 (m, 3H),
2.21−2.24 (m, 2H), 2.23 (s, 6H), 2.42 (t, J = 6.9 Hz, 2H), 2.64
(t, J = 7.7 Hz, 2H), 3.28 (t, J = 6.9 Hz, 2H), 3.52−3.56 (m,
1H), 3.91 (ddd, J = 11.4, 8.6, 3.2 Hz, 1H), 4.45 (d, J = 11.6
Hz, 1H), 4.68−4.72 (m, 2H), 7.19 (d, J = 8.0 Hz, 2H), 7.27
(d, J = 8.0 Hz, 2H). 13C NMR (176 MHz, CD3OD): δ 20.6,
25.7, 26.7, 28.7, 31.9, 36.2, 36.6, 38.4, 45.6, 57.8, 59.2, 63.5,
70.1, 99.4, 129.3, 129.7, 137.3, 142.7, 175.5, 177.3.
N-(1-((2-(Dimethylamino)ethyl)amino)-2-methyl-1-
oxopropan-2-yl)-4-(4-(hydroxymethyl)phenyl)-
butanamide (9b). Concentrated HCl (37%, 544 mL, 6.53
mol, 1.2 equiv) was slowly added to a cold (0−5 °C) solution
of THP-protected amide 9a (2.36 kg, 5.44 mol) in MeOH (7.1
L) at <25 °C (caution: very exothermic!). The mixture was
warmed to 35 °C, aged until reaction completion (1 h), and
then cooled to 0−5 °C. Aqueous NaOH (50 wt %, 365 mL,
1.21 equiv) was slowly added at 5−10 °C until the pH was
10−12. The reaction mixture was concentrated to near dryness
(∼3 L) and flushed with MeOH (1.5 L × 4) and then THF (2
L × 4) at 45−50 °C under reduced pressure. The precipitated
NaCl was filtered warm (40−45 °C to prevent premature
crystallization of the product), and the filter cake was rinsed
with warm THF (40 °C, ∼2 L). n-Heptane (2 L) was added to
the combined filtrate, and the mixture aged at 30−35 °C for 45
min to give a white slurry. Additional n-heptane (7 L) was
charged over 0.5−1 h, and the mixture aged at 30 °C for 0.5
h, 10−15 °C for 3 h, and then filtered. The filter cake was
washed with 2:1 n-heptane/THF (6 L) and dried at 45 °C
under reduced pressure to give 1.83 kg (96.2% yield) of 9b as a
white solid with 99.2% purity (HPLC area %). mp: 92 °C
(DSC peak temperature). LC−MS: calcd [M + H]+ 350, found
1
m/z 350. H NMR (700 MHz, DMSO-d6): δ 1.29 (s, 6H),
1.76 (quin, J = 7.6 Hz, 2H), 2.08 (s, 6H), 2.09 (t, J = 7.3 Hz,
2H), 2.21 (t, J = 6.9 Hz, 2H), 2.53 (t, J = 7.6 Hz, 2H), 3.07 (t,
J = 6.9 Hz, 2H), 4.45 (s, 2H), 7.13 (d, J = 8.0 Hz, 2H), 7.22
(d, J = 8.0 Hz, 2H). 13C NMR (176 MHz, DMSO-d6): δ 25.2,
27.0, 34.4, 34.9, 36.8, 45.1, 55.6, 57.9, 62.7, 126.5, 128.0, 139.9,
140.2, 171.5, 173.9.
4-(4-((1-((2-(Dimethylamino)ethyl)amino)-2-methyl-
1-oxopropan-2-yl)amino)-4-oxobutyl)benzyl Methyl
Carbonate (9c). To a suspension of benzyl alcohol 9b
(2.50 kg, 7.16 mol) in dimethyl carbonate (5.0 L) and THF (2
L) was added NaOMe (25 wt % in MeOH, 4.36 M, 82 mL, 5
mol %). The mixture was aged at 40 °C for 10 min and then
M
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
The filter cake was washed with IPA (2.0 L + 1.0 L) and n-
heptane (1.0 L) and dried at 40−45 °C under reduced
pressure to afford 819 g (84.5% yield corrected for 93.5 wt %
assay) of triacetate 35 as a white solid with 99.5% purity
(HPLC area %). mp: 143 °C (DSC peak temperature). LC−
AUTHOR INFORMATION
Corresponding Author
*E-mail: wwu@lexpharma.com.
ORCID
Lauren E. Sirois: 0000-0002-1948-3749
Matthew M. Zhao: 0000-0002-4163-3555
Ngiap-Kie Lim: 0000-0002-3760-4874
Wenxue Wu: 0000-0002-5979-9350
■
1
MS: calcd [M + H]+ 728, found m/z 728. H NMR (400
MHz, DMSO-d6): δ 1.29 (s, 6H), 1.70 (s, 3H), 1.68−1.77 (m,
2H), 1.94 (s, 3H), 2.04 (s, 3H), 2.06 (s, 6H), 2.09 (s, 3H),
2.05−2.10 (m, 2H), 2.17 (s, 3H), 2.20 (t, J = 6.9 Hz, 2H), 3.06
(q, J = 6.4 Hz, 2H), 3.88 (d, J = 15.5 Hz, 1H), 3.92 (d, J = 15.4
Hz, 1H), 4.65 (d, J = 9.8 Hz, 1H), 4.89 (d, J = 10.0 Hz, 1H),
5.06 (t, J = 9.0 Hz, 1H), 5.11 (t, J = 9.8 Hz, 1H), 5.36 (t, J =
9.5 Hz, 1H), 7.01 (d, J = 8.3 Hz, 2H), 7.08 (d, J = 8.0 Hz, 2H),
Present Addresses
†L.E.S. and N.-K.L.: Small Molecule Process Chemistry,
Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080.
‡M.S.B.: Abzena Biotechnology, 360 George Patterson
Boulevard, Suite 101E, Bristol, PA 19007.
7.10−7.16 (m, 4H), 7.19 (t, J = 5.4 Hz, 1H), 7.83 (s, 1H). 13
C
§B.A.H.: Morphic Rock Therapeutic, 35 Gatehouse Drive,
Waltham, MA 02451.
NMR (101 MHz, DMSO-d6): δ 10.3, 19.0, 20.1, 20.3, 20.5,
25.2, 27.0, 34.3, 35.0, 36.9, 38.1, 45.1, 55.7, 57.9, 68.9, 72.4,
73.2, 78.2, 81.0, 124.9, 128.3, 128.4, 128.8, 130.1, 134.3, 136.4,
137.5, 138.9, 139.4, 168.4, 169.2, 169.5, 171.5, 173.9.
Notes
The authors declare no competing financial interest.
N-(1-((2-(Dimethylamino)ethyl)amino)-2-methyl-1-
oxopropan-2-yl)-4-(4-(2-methyl-5-((2S,3R,4R,5S,6R)-
3,4,5-trihydroxy-6-(methylthio)tetrahydro-2H-pyran-2-
yl)benzyl)phenyl)butanamide·(S)-Pyrrolidine-2-carbox-
ylate (LX2761·L-Proline). A solution of sodium methoxide in
MeOH (25 wt %, 19.4 mL, 85 mmol, 0.05 equiv) was charged
into a suspension of triacetate 35 (1.30 kg, 1.21 kg active, 1.66
mol, 1.00 equiv) in EtOH (6.0 L). The reaction mixture was
aged at 45 °C until reaction completion (3−5 h). Meanwhile, a
solution of L-proline (220 g, 1.91 mol, 1.15 equiv) in EtOH/
H2O (485 mL/102 mL) was prepared. Approximately 50%
(∼380 mL) of the L-proline solution was added, followed by
seeds of LX2761·L-proline cocrystal (11.9 g). After 1 h of
stirring at 40 °C, the remaining L-proline solution (355 mL)
was added slowly (1 h), and the resulting thick suspension was
aged at 40 °C for 1 h before the addition of MTBE (10 L) at
30 °C over 1 h. The mixture was aged for 1 h, cooled to 20 °C,
and aged for 2−5 h. The product was filtered, washed
sequentially with 2:1 MTBE/EtOH (3.6 L) and MTBE (7 L),
and then dried at 30−45 °C under reduced pressure to furnish
LX2761·L-proline cocrystal in 91.0% yield with 99.7% purity
(HPLC area %). mp: 150 °C (DSC peak temperature);
crystalline by XRPD. LC−MS: calcd [M + H]+ 602, found m/z
ACKNOWLEDGMENTS
■
The authors thank Leonard Hargiss for analytical support,
particularly for assistance with HPLC method development
and LC−MS analysis. We also thank Philip Keyes for
supporting NMR structural assignments and Vincent Caruso
for chromatography support. We gratefully acknowledge the
helpful collaboration with discovery chemists Kenneth Carson,
David Rawlins, Nicole Goodwin, and Eric Strobel; general
project input from Weiguo Liu, Boguslawa Wilk, and Evan
Bekos; and Jeffrey Eckert and Steven Weissman for assistance
in manuscript preparation.
REFERENCES
■
(1) (a) Tahrani, A. A.; Barnett, A. H.; Bailey, C. J. SGLT Inhibitors
in Management of Diabetes. Lancet Diabetes Endocrinol. 2013, 1,
140−151. (b) Sabino-Silva, R.; Mori, R. C.; David-Silva, A.; Okamoto,
M. M.; Freitas, H. S.; Machado, U. F. The Na+/Glucose
Cotransporters: From Genes to Therapy. Braz. J. Med. Biol. Res.
2010, 43, 1019−1026. (c) DeFronzo, R. A.; Norton, L.; Abdul-Ghani,
M. Renal, Metabolic and Cardiovascular Considerations of SGLT2
Inhibition. Nat. Rev. Nephrol. 2017, 13, 11−26. (d) Mudaliar, S.;
Polidori, D.; Zambrowicz, B.; Henry, R. R. Sodium-Glucose
Cotransporter Inhibitors: Effects on Renal and Intestinal Glucose
Transport: From Bench to Bedside. Diabetes Care 2015, 38, 2344−
2353.
(2) For reviews and syntheses, see: (a) He, Z.-X.; Zhou, Z.-W.;
Yang, Y.; Yang, T.; Pan, S.-Y.; Qiu, J.-X.; Zhou, S.-F. Overview of
Clinically Approved Oral Antidiabetic Agents for the Treatment of
Type 2 Diabetes Mellitus. Clin. Exp. Pharmacol. Physiol. 2015, 42,
125−138. (b) Cai, W.; Jiang, L.; Xie, Y.; Liu, Y.; Liu, W.; Zhao, G.
Design of SGLT2 Inhibitors for the Treatment of Type 2 Diabetes: A
History Driven by Biology to Chemistry. Med. Chem. 2015, 11, 317−
328. (c) Aguillon, A. R.; Mascarello, A.; Segretti, N. D.; de Azevedo,
H. F. Z.; Guimaraes, C. R. W.; Miranda, L. S. M.; de Souza, R. O. M.
A. Synthetic Strategies toward SGLT2 Inhibitors. Org. Process Res.
Dev. 2018, 22, 467−488. and references therein.
(3) For chemistry reports of approved SGLT-2 inhibitors
contemporaneous to this work, see: (a) 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. Discovery of Dapagliflozin: A Potent,
Selective Renal Sodium-Dependent Glucose Cotransporter 2
(SGLT2) Inhibitor for the Treatment of Type 2 Diabetes. J. Med.
1
602. LX2761/L-proline molar ratio: 1.0/1.0 (1H NMR). H
NMR (700 MHz, CD3OD): δ 1.41 (s, 6H), 1.83−1.90 (m,
2H), 1.92−2.01 (m, 2H), 2.07−2.13 (m, 1H), 2.14 (s, 3H),
2.18−2.22 (m, 5H), 2.25 (s, 6H), 2.26−2.33 (m, 1H), 2.44 (t,
J = 6.9 Hz, 2H), 2.59 (t, J = 7.6 Hz, 2H), 3.18−3.23 (m, 1H),
3.29 (t, J = 6.9 Hz, 2H), 3.35−3.39 (m, 2H), 3.41 (t, J = 9.3
Hz, 1H), 3.46 (t, J = 8.8 Hz, 1H), 3.92−3.99 (m, 3H), 4.13 (d,
J = 9.4 Hz, 1H), 4.39 (d, J = 9.5 Hz, 1H), 7.06 (d, J = 8.2 Hz,
2H), 7.09 (d, J = 8.2 Hz, 2H), 7.14 (d, J = 8.2 Hz, 1H), 7.16−
7.18 (m, 2H). 13C NMR (101 MHz, CD3OD): δ 12.0, 19.7,
25.3, 25.7, 28.7, 30.6, 36.1, 36.6, 38.2, 40.2, 45.6, 47.2, 57.7,
59.2, 62.8, 73.9, 76.2, 79.6, 83.7, 87.5, 127.0, 129.6, 130.0,
130.8, 131.2, 137.7, 138.0, 139.5, 140.4, 140.7, 174.4, 175.6,
177.4.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.oprd.8b00325.
Copies of 1H and 13C NMR spectra for new compounds
(PDF)
N
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Chem. 2008, 51, 1145−1149. (b) Nomura, S.; Sakamaki, S.; Hongu,
M.; Kawanishi, E.; Koga, Y.; Sakamoto, T.; Yamamoto, Y.; Ueta, K.;
Kimata, H.; Nakayama, K.; Tsuda-Tsukimoto, M. Discovery of
Canagliflozin, a Novel C-Glucoside with Thiophene Ring, as Sodium-
Dependent Glucose Cotransporter 2 Inhibitor for the Treatment of
Type 2 Diabetes Mellitus. J. Med. Chem. 2010, 53, 6355−6360.
(c) Wang, X.-J.; Zhang, L.; Byrne, D.; Nummy, L.; Weber, D.;
Krishnamurthy, D.; Yee, N.; Senanayake, C. H. Efficient Synthesis of
Empagliflozin, an Inhibitor of SGLT-2, Utilizing an AlCl3-Promoted
Silane Reduction of a β-Glycopyranoside. Org. Lett. 2014, 16, 4090−
4093. (d) Bernhardson, D.; Brandt, T. A.; Hulford, C. A.; Lehner, R.
S.; Preston, B. R.; Price, K.; Sagal, J. F.; St. Pierre, M. J.; Thompson, P.
H.; Thuma, B. Development of an Early-Phase Bulk Enabling Route
to Sodium-Dependent Glucose Cotransporter 2 Inhibitor Ertugli-
flozin. Org. Process Res. Dev. 2014, 18, 57−65. (e) Bowles, P.; Brenek,
Harris, A. L.; Smith, M.; Thompson, A. Y.; Xiong, W.; Mseeh, F.;
Bruce, D. J.; Diaz, D.; Gopinathan, S.; Li, L.; O’Neill, E.; Thiel, M.;
Wilson, A. G. E.; Carson, K. G.; Powell, D. R.; Rawlins, D. B.
Discovery of LX2761, a Sodium-Dependent Glucose Cotransporter 1
(SGLT1) Inhibitor Restricted to the Intestinal Lumen, for the
Treatment of Diabetes. J. Med. Chem. 2017, 60, 710−721.
(6) Some milder approaches have been reported. For Friedel−Crafts
benzylations, see: (a) Mertins, K.; Iovel, I.; Kischel, J.; Zapf, A.; Beller,
M. Gold-Catalyzed Benzylation of Arenes and Heteroarenes. Adv.
̈
Synth. Catal. 2006, 348, 691−695. (b) Schafer, G.; Bode, J. W.
Friedel-Crafts Benzylation of Activated and Deactivated Arenes.
Angew. Chem., Int. Ed. 2011, 50, 10913−10916. For Barbier−
Grignard-type benzylations, see: (c) Zhou, F.; Li, C.-J. The Barbier−
Grignard-Type Arylation of Aldehydes using Unactivated Aryl Iodides
in Water. Nat. Commun. 2014, 5, 4254. (d) Bering, L.; Jeyakumar, K.;
Antonchick, A. P. Metal-Free C−O Bond Functionalization: Catalytic
Intramolecular and Intermolecular Benzylation of Arenes. Org. Lett.
2018, 20, 3911−3914.
́
S. J.; Caron, S.; Do, N. M.; Drexler, M. T.; Duan, S.; Dube, P.;
Hansen, E. C.; Jones, B. P.; Jones, K. N.; Ljubicic, T. A.; Makowski, T.
W.; Mustakis, J.; Nelson, J. D.; Olivier, M.; Peng, Z.; Perfect, H. H.;
Place, D. W.; Ragan, J. A.; Salisbury, J. J.; Stanchina, C. L.;
Vanderplas, B. C.; Webster, M. E.; Weekly, R. M. Commercial Route
Research and Development for SGLT2 Inhibitor Candidate
Ertugliflozin. Org. Process Res. Dev. 2014, 18, 66−81. (f) Ohtake, Y.;
Sato, T.; Kobayashi, T.; Nishimoto, M.; Taka, N.; Takano, K.;
Yamamoto, K.; Ohmori, M.; Yamaguchi, M.; Takami, K.; Yeu, S.-Y.;
Ahn, K.-H.; Matsuoka, H.; Morikawa, K.; Suzuki, M.; Hagita, H.;
Ozawa, K.; Yamaguchi, K.; Kato, M.; Ikeda, S. Discovery of
Tofogliflozin, a Novel C-Arylglucoside with an O-Spiroketal Ring
System, as a Highly Selective Sodium Glucose Cotransporter 2
(SGLT2) Inhibitor for the Treatment of Type 2 Diabetes. J. Med.
Chem. 2012, 55, 7828−7840. (g) Imamura, M.; Nakanishi, K.; Suzuki,
T.; Ikegai, K.; Shiraki, R.; Ogiyama, T.; Murakami, T.; Kurosaki, E.;
Noda, A.; Kobayashi, Y.; Yokota, M.; Koide, T.; Kosakai, K.; Ohkura,
Y.; Takeuchi, M.; Tomiyama, H.; Ohta, M. Discovery of Ipragliflozin
(ASP1941): A Novel C-Glucoside with Benzothiophene Structure as
a Potent and Selective Sodium Glucose Co-transporter 2 (SGLT2)
Inhibitor for the Treatment of Type 2 Diabetes Mellitus. Bioorg. Med.
Chem. 2012, 20, 3263−3279. (h) Kakinuma, H.; Oi, T.; Hashimoto-
Tsuchiya, Y.; Arai, M.; Kawakita, Y.; Fukasawa, Y.; Iida, I.; Hagima,
N.; Takeuchi, H.; Chino, Y.; Asami, J.; Okumura-Kitajima, L.; Io, F.;
Yamamoto, D.; Miyata, N.; Takahashi, T.; Uchida, S.; Yamamoto, K.
(1S)-1,5-Anhydro-1-[5-(4-ethoxybenzyl)-2-methoxy-4-methylphen-
yl]-1-thio-D-glucitol (TS-071) is a Potent, Selective Sodium-Depend-
ent Glucose Cotransporter 2 (SGLT2) Inhibitor for Type 2 Diabetes
Treatment. J. Med. Chem. 2010, 53, 3247−3261.
(4) (a) Goodwin, N. C.; Mabon, R.; Harrison, B. A.; Shadoan, M.
K.; Almstead, Z. Y.; Xie, Y.; Healy, J.; Buhring, L. M.; DaCosta, C. M.;
Bardenhagen, J.; Mseeh, F.; Liu, Q.; Nouraldeen, A.; Wilson, A. G. E.;
Kimball, S. D.; Powell, D. R.; Rawlins, D. B. Novel L-Xylose
Derivatives as Selective Sodium-Dependent Glucose Cotransporter 2
(SGLT2) Inhibitors for the Treatment of Type 2 Diabetes. J. Med.
Chem. 2009, 52, 6201−6204. (b) Powell, D. R.; Smith, M.; Greer, J.;
Harris, A.; Zhao, S.; DaCosta, C.; Mseeh, F.; Shadoan, M. K.; Sands,
A.; Zambrowicz, B.; Ding, Z. M. LX4211 Increases Serum Glucagon-
Like Peptide 1 and Peptide YY Levels by Reducing Sodium/Glucose
Cotransporter 1 (SGLT1)−Mediated Absorption of Intestinal
Glucose. J. Pharmacol. Exp. Ther. 2013, 345, 250−259. (c) Lapuerta,
P.; Zambrowicz, B.; Strumph, P.; Sands, A. Development of
Sotagliflozin, a Dual Sodium-dependent Glucose Transporter 1/2
Inhibitor. Diabetes Vasc. Dis. Res. 2015, 12, 101−110. (d) Chen, J.;
Nyamweya, N. N.; Ong, K. K. H. Compositions Comprising and
Methods of Using Inhibitors of Sodium-Glucose Co-Transporters 1
and 2, US 2016/0296547 A1, Oct 13, 2016. (e) Zhao, M. M.; De
Paul, S. M.; Perlberg, A. Solid Forms of (2S,3R,4R,5S,6R)-2-(4-
Chloro-3-(4-ethoxybenzyl)phenyl)-6-(methylthio)tetrahydro-2H-
pyran-3,4,5-triol and Methods of Their Use. US8217156B2, 2009.
(5) (a) Carson, K. G.; Goodwin, N. C.; Harrison, B. A.; Rawlins, D.
B.; Strobel, E.; Zambrowicz, B. Inhibitors of Sodium Glucose
Cotransporter 1. U.S. Patent 9,688,710 B2, June 27, 2017.
(b) Goodwin, N. C.; Ding, Z.-M.; Harrison, B. A.; Strobel, E. D.;
(7) For selected examples of diarylmethane synthesis using other
cross-coupling strategies, see: Negishi-type: (a) Duplais, C.;
Krasovskiy, A.; Wattenberg, A.; Lipshutz, B. H. Cross-couplings
Between Benzylic and Aryl Halides “On Water”: Synthesis of
Diarylmethanes. Chem. Commun. 2010, 46, 562−564. (b) Bedford,
R. B.; Huwe, M.; Wilkinson, M. C. Iron-catalysed Negishi Coupling of
Benzylhalides and Phosphates. Chem. Commun. 2009, 600−602.
Hiyama-type: (c) Srimani, D.; Bej, A.; Sarkar, A. Palladium
Nanoparticle Catalyzed Hiyama Coupling Reaction of Benzyl Halides.
J. Org. Chem. 2010, 75, 4296−4299. Kumada-related: (d) Kofink, C.
C.; Knochel, P. Synthesis of Functionalized Diarylmethanes via a
Copper-Catalyzed Cross-Coupling of Arylmagnesium Reagents with
Benzylic Phosphates. Org. Lett. 2006, 8, 4121−4124. (e) Yu, D.-G.;
Wang, X.; Zhu, R.-Y.; Luo, S.; Zhang, X.-B.; Wang, B.-Q.; Wang, L.;
Shi, Z.-J. Direct Arylation/Alkylation/Magnesiation of Benzyl
Alcohols in the Presence of Grignard Reagents via Ni-, Fe-, or Co-
Catalyzed sp3 C−O Bond Activation. J. Am. Chem. Soc. 2012, 134,
14638−14641. Other: (f) Chen, C.-R.; Zhou, S.; Biradar, D. B.; Gau,
H.-M. Extremely Efficient Cross-Coupling of Benzylic Halides with
Aryltitanium Tris(isopropoxide) Catalyzed by Low Loadings of a
Simple Palladium(II) Acetate/Tris(p-tolyl)phosphine System. Adv.
Synth. Catal. 2010, 352, 1718−1727. (g) Zhao, F.; Tan, Q.; Xiao, F.;
Zhang, S.; Deng, G.-J. Palladium-Catalyzed Desulfitative Cross-
Coupling Reaction of Sodium Sulfinates with Benzyl Chlorides.
Org. Lett. 2013, 15, 1520−1523. (h) Tellis, J. C.; Primer, D. N.;
Molander, G. A. Single-electron Transmetalation in Organoboron
Cross-coupling by Photoredox/Nickel Dual Catalysis. Science 2014,
345, 433−436.
(8) For a similar thioalkylation of anomeric acetates, see: Ibatullin, F.
̈
M.; Shabalin, K. A.; Janis, J. V.; Shavva, A. G. Reaction of 1,2-trans-
Glycosyl Acetates with Thiourea: A New Entry to 1-Thiosugars.
Tetrahedron Lett. 2003, 44, 7961−7964.
(9) For a review of palladium activation of aryl chlorides, see:
(a) Littke, A. F.; Fu, G. C. Palladium-Catalyzed Coupling Reactions
of Aryl Chlorides. Angew. Chem., Int. Ed. 2002, 41, 4176−4211. For
representative borylations of hindered substrates, see: (b) Ishiyama,
T.; Ishida, K.; Miyaura, N. Synthesis of Pinacol Arylboronates via
Cross-coupling Reaction of Bis(pinacolato)diboron with Chloroar-
enes Catalyzed by Palladium(0)-Tricyclohexylphosphine Complexes.
Tetrahedron 2001, 57, 9813−9816. (c) Kawamorita, S.; Ohmiya, H.;
Iwai, T.; Sawamura, M. Palladium-Catalyzed Borylation of Sterically
Demanding Aryl Halides with a Silica-Supported Compact Phosphane
Ligand. Angew. Chem., Int. Ed. 2011, 50, 8363−8366. (d) Molander,
G. A.; Trice, S. L. J.; Dreher, S. D. Palladium-Catalyzed, Direct
Boronic Acid Synthesis from Aryl Chlorides: A Simplified Route to
Diverse Boronate Ester Derivatives. J. Am. Chem. Soc. 2010, 132,
17701−17703. (e) Iwai, T.; Harada, T.; Tanaka, R.; Sawamura, M.
Silica-supported Tripod Triarylphosphines: Application to Palladium-
catalyzed Borylation of Chloroarenes. Chem. Lett. 2014, 43, 584−586.
(10) (a) Billingsley, K. L.; Barder, T. E.; Buchwald, S. L. Palladium-
Catalyzed Borylation of Aryl Chlorides: Scope, Applications, and
O
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Computational Studies. Angew. Chem., Int. Ed. 2007, 46, 5359−5363.
(b) Billingsley, K. L.; Buchwald, S. L. An Improved System for the
Palladium-Catalyzed Borylation of Aryl Halides with Pinacol Borane.
J. Org. Chem. 2008, 73, 5589−5591.
A.; Knochel, P. Efficient Synthesis of Functionalized Organozinc
Compounds by the Direct Insertion of Zinc into Organic Iodides and
Bromides. Angew. Chem., Int. Ed. 2006, 45, 6040−6044. (b) Sase, S.;
Jaric, M.; Metzger, A.; Malakhov, V.; Knochel, P. One-Pot Negishi
Cross-Coupling Reactions of in-situ Generated Zinc Reagents with
Aryl Chlorides, Bromides, and Triflates. J. Org. Chem. 2008, 73,
7380−7382.
(24) For a recent discussion of the complicated role of such salts in
Negishi couplings, see: McCann, L. C.; Organ, M. G. On the
Remarkably Different Role of Salt in the Cross-Coupling of Arylzincs
from that Seen with Alkylzincs. Angew. Chem., Int. Ed. 2014, 53,
4386−4389 and references therein .
(25) For safety, it is recommended that the reaction mixture be held
at 50 °C before heating to reflux (approximately 65−70 °C).
(26) Dunetz, J. R.; Xiang, Y.; Baldwin, A.; Ringling, J. General and
Scalable Amide Bond Formation with Epimerization-Prone Substrates
Using T3P and Pyridine. Org. Lett. 2011, 13, 5048−5051. and
references therein.
(27) Other strategies explored for amide couplings include boric and
arylboronic acids. See: (a) Gernigon, N.; Al-Zoubi, R. M.; Hall, D. G.
Direct Amidation of Carboxylic Acids Catalyzed by ortho-Iodo
Arylboronic Acids: Catalyst Optimization, Scope, and Preliminary
Mechanistic Study Supporting a Peculiar Halogen Acceleration Effect.
J. Org. Chem. 2012, 77, 8386−8400. and references therein.
(b) Mylavarapu, R. K.; GCM, K.; Kolla, N.; Veeramalla, R.;
Koilkonda, P.; Bhattacharya, A.; Bandichhor, R. Boric Acid Catalyzed
Amidation in the Synthesis of Active Pharmaceutical Ingredients. Org.
Process Res. Dev. 2007, 11, 1065−1068.
(28) For a discussion of CDI degradation, see: Engstrom, K. M.;
Sheikh, A.; Ho, R.; Miller, R. W. The Stability of N,N-Carbon-
yldiimidazole Toward Atmospheric Moisture. Org. Process Res. Dev.
2014, 18, 488−494.
(29) For a review of amide coupling strategies, see: Valeur, E.;
Bradley, M. Amide Bond Formation: Beyond the Myth of Coupling
Reagents. Chem. Soc. Rev. 2009, 38, 606−631.
(30) For a review, see: Miyaura, N.; Suzuki, A. Palladium-Catalyzed
Cross-Coupling Reactions of Organoboron Compounds. Chem. Rev.
1995, 95, 2457−2483.
(31) For example, see the reviews in ref 14 (alkyl couplings) and:
Molander, G. A.; Ellis, N. Organotrifluoroborates: Protected Boronic
Acids that Expand the Versatility of the Suzuki Coupling Reaction.
Acc. Chem. Res. 2007, 40, 275−286.
(11) Seed, A. J.; Sonpatki, V.; Herbert, M. R. 3-(4-Bromobenzoyl)-
propanoic Acid. Org. Synth. 2002, 79, 204−208.
(12) For a discussion of the process hazards of hydrazine, see:
(a) Niemeier, J. K.; Kjell, D. P. Hydrazine and Aqueous Hydrazine
Solutions: Evaluating Safety in Chemical Processes. Org. Process Res.
Dev. 2013, 17, 1580−1590. For an example of a kilogram-scale
reduction, see: (b) Kuethe, J. T.; Childers, K. G.; Peng, Z.; Journet,
M.; Humphrey, G. R.; Vickery, T.; Bachert, D.; Lam, T. T. A Practical,
Kilogram-Scale Implementation of the Wolff−Kishner Reduction.
Org. Process Res. Dev. 2009, 13, 576−580.
(13) For example, see: Mirza-Aghayan, M.; Boukherroub, R.;
Rahimifard, M. A Simple and Efficient Hydrogenation of Benzyl
Alcohols to Methylene Compounds using Triethylsilane and a
Palladium Catalyst. Tetrahedron Lett. 2009, 50, 5930−5932.
(14) For reviews of alkyl halide/alkyl organometallic couplings, see:
(a) Jana, R.; Pathak, T. P.; Sigman, M. S. Advances in Transition
Metal (Pd,Ni,Fe)-Catalyzed Cross-Coupling Reactions Using Alkyl-
organometallics as Reaction Partners. Chem. Rev. 2011, 111, 1417−
1492. (b) Frisch, A. C.; Beller, M. Catalysts for Cross-Coupling
Reactions with Non-activated Alkyl Halides. Angew. Chem., Int. Ed.
2005, 44, 674−688.
(15) For a review, see: Chemler, S. R.; Trauner, D.; Danishefsky, S.
J. The B-Alkyl Suzuki−Miyaura Cross-Coupling Reaction: Develop-
ment, Mechanistic Study, and Applications in Natural Product
Synthesis. Angew. Chem., Int. Ed. 2001, 40, 4544−4568.
(16) (a) Manolikakes, G.; Munoz Hernandez, C.; Schade, M. A.;
̃
Metzger, A.; Knochel, P. Palladium- and Nickel-Catalyzed Cross-
Couplings of Unsaturated Halides Bearing Relatively Acidic Protons
with Organozinc Reagents. J. Org. Chem. 2008, 73, 8422−8436. and
references therein. (b) Yang, Y.; Oldenhuis, N. J.; Buchwald, S. L.
Mild and General Conditions for Negishi Cross-Coupling Enabled by
the Use of Palladacycle Precatalysts. Angew. Chem., Int. Ed. 2013, 52,
615−619.
(17) For a review, see: Negishi, E.-I.; Zeng, X.; Tan, Z.; Qian, M.;
Hu, Q.; Huang, Z. Palladium- or Nickel-Catalyzed Cross-Coupling
with Organometals Containing Zinc, Aluminum, and Zirconium: The
Negishi Coupling. In Metal-Catalyzed Cross-Coupling Reactions, 2nd
ed.; de Meijere, A., Diederich, F., Eds.; Wiley-VCH: Weinheim,
Germany, 2008; Chapter 15.
(32) For representative examples, see: (a) Nobre, S. M.; Monteiro,
A. L. Synthesis of Diarylmethane Derivatives from Pd-catalyzed
Cross-coupling Reactions of Benzylic Halides with Arylboronic Acids.
Tetrahedron Lett. 2004, 45, 8225−8228. and references therein.
(b) Molander, G. A.; Elia, M. D. Suzuki−Miyaura Cross-Coupling
Reactions of Benzyl Halides with Potassium Aryltrifluoroborates. J.
Org. Chem. 2006, 71, 9198−9202. and references therein.
(33) Carbonates: (a) Kuwano, R.; Yokogi, M. Suzuki−Miyaura
Cross-Coupling of Benzylic Carbonates with Arylboronic Acids. Org.
Lett. 2005, 7, 945−947. (b) Yu, J.-Y.; Kuwano, R. Suzuki−Miyaura
Coupling of Diarylmethyl Carbonates with Arylboronic Acids: A New
Access to Triarylmethanes. Org. Lett. 2008, 10, 973−976. Acetates:
(c) Kuwano, R.; Yokogi, M. Cross-coupling of Benzylic Acetates with
Arylboronic Acids: One-pot Transformation of Benzylic Alcohols to
Diarylmethanes. Chem. Commun. 2005, 5899−5901. (d) Stewart, G.
W.; Maligres, P. E.; Baxter, C. A.; Junker, E. M.; Krska, S. W.; Scott, J.
P. An Approach to Heterodiarylmethanes via sp2-sp3 Suzuki−Miyaura
Cross-coupling. Tetrahedron 2016, 72, 3701−3706. Pivalates:
(e) Chen, Q.; Fan, X.-H.; Zhang, L.-P.; Yang, L.-M. Ni(II) Source
as a Pre-catalyst for the Cross-coupling of Benzylic Pivalates with
Arylboronic Acids: Facile Access to Tri- and Diarylmethanes. RSC
Adv. 2015, 5, 15338−15340. Phosphates: (f) McLaughlin, M.
Suzuki−Miyaura Cross-Coupling of Benzylic Phosphates with
Arylboronic Acids. Org. Lett. 2005, 7, 4875−4878. Carbamates:
(g) Ohtsuki, A.; Yanagisawa, K.; Furukawa, T.; Tobisu, M.; Chatani,
N. Nickel/N-Heterocyclic Carbene-Catalyzed Suzuki−Miyaura Type
Cross-Coupling of Aryl Carbamates. J. Org. Chem. 2016, 81, 9409−
(18) For reviews, see: (a) Knochel, P.; Singer, R. D. Preparation and
Reactions of Polyfunctional Organozinc Reagents in Organic
Synthesis. Chem. Rev. 1993, 93, 2117−2188. (b) Klatt, T.;
̈
Markiewicz, J. T.; Samann, C.; Knochel, P. Strategies to Prepare
and Use Functionalized Organometallic Reagents. J. Org. Chem. 2014,
79, 4253−4269. Metal powders: (c) Rieke, R. D. Preparation of
Highly Reactive Metal Powders and Their Use in Organic and
Organometallic Synthesis. Acc. Chem. Res. 1977, 10, 301−306.
(19) Girgis, M. J.; Liang, J. K.; Du, Z.; Slade, J.; Prasad, K. A Scalable
Zinc Activation Procedure Using DIBAL-H in a Reformatsky
Reaction. Org. Process Res. Dev. 2009, 13, 1094−1099.
(20) Lang, S.; Murso, A.; Wietelmann, U. Highly Reactive Zinc
Form, Method for the Production Thereof, and Use of the Same. U.S.
Patent 7,759,509, 2010.
(21) In the lab, GC monitoring was found to be more reproducible.
Reaction aliquots were quenched into acetic acid, extracted to an
organic layer, and analyzed. The resulting ethyl butyrate was
quantified against ethyl 4-bromobutyrate. In some cases (with
chloride additives), small amounts of ethyl 4-chlorobutyrate were
also observed.
(22) Also see: Picotin, G.; Miginiac, P. Activation of Zinc by
Trimethylchlorosilane. An Improved Procedure for the Preparation of
beta-Hydroxy Esters from Ethyl Bromoacetate and Aldehydes or
Ketones (Reformatsky Reaction). J. Org. Chem. 1987, 52, 4796−4798.
(23) For examples of the use of LiCl in zinc insertions/one-pot
Negishi reactions, see: (a) Krasovskiy, A.; Malakhov, V.; Gavryushin,
P
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
9414. (h) Wang, X.-X.; Luo, M.-J.; Lu, J.-M. N-Heterocyclic Carbene-
Palladium(II)-1-Methylimidazole Complex Catalyzed Suzuki−
Miyaura Coupling of Benzyl Carbamates with Arylboronic Acids.
Org. Biomol. Chem. 2015, 13, 11438−11444. (i) Harris, M. R.; Hanna,
L. E.; Greene, M. A.; Moore, C. E.; Jarvo, E. R. Retention or Inversion
in Stereospecific Nickel-Catalyzed Cross-Coupling of Benzylic
Carbamates with Arylboronic Esters: Control of Absolute Stereo-
chemistry with an Achiral Catalyst. J. Am. Chem. Soc. 2013, 135,
3303−3306. Sulfonates: (j) Wang, X.-X.; Xu, B.-B.; Song, W.-T.; Sun,
K.-X.; Lu, J.-M. N-heterocyclic Carbene-palladium(II)-1-methylimi-
dazole Complex Catalyzed Suzuki−Miyaura Coupling of Benzyl
Sulfonates with Arylboronic Acids. Org. Biomol. Chem. 2015, 13,
4925−4930. Ethers: (k) Tobisu, M.; Yasutome, A.; Kinuta, H.;
Nakamura, K.; Chatani, N. 1,3-Dicyclohexylimidazol-2-ylidene as a
Superior Ligand for the Nickel-Catalyzed Cross-Couplings of Aryl and
Benzyl Methyl Ethers with Organoboron Reagents. Org. Lett. 2014,
16, 5572−5575. Alcohols: (l) Cao, Z.-C.; Yu, D.-G.; Zhu, R.-Y.; Wei,
J.-B.; Shi, Z.-J. Direct Cross-coupling of Benzyl Alcohols to Construct
Diarylmethanes via Palladium Catalysis. Chem. Commun. 2015, 51,
2683−2686.
(c) Hruszkewycz, D. P.; Balcells, D.; Guard, L. M.; Hazari, N.; Tilset,
M. Insight into the Efficiency of Cinnamyl-Supported Precatalysts for
the Suzuki−Miyaura Reaction: Observation of Pd(I) Dimers with
Bridging Allyl Ligands During Catalysis. J. Am. Chem. Soc. 2014, 136,
7300−7316. and references therein.
(41) A small amount of residual pinacol borate salt in the product
was present but readily purged in the downstream step.
(42) For a review of cocrystal API development, see: (a) Schultheiss,
N.; Newman, A. Pharmaceutical Cocrystals and Their Physicochem-
ical Properties. Cryst. Growth Des. 2009, 9, 2950−2967. For
sonication screening with minimal solvent, see: (b) Morrison, H.;
Mrozek-Morrison, M.; Toschi, J.; Luu, V.; Tan, H.; Daurio, D. High
Throughput Bench-Top Cocrystal Screening via a Floating Foam
Rack/Sonic Bath Method. Org. Process Res. Dev. 2013, 17, 533−539.
(43) See ref 2c and: Imamura, M.; Nakanishi, K.; Shiraki, R.; Onda,
K.; Sasuga, D.; Yuda, M. Cocrystal of C-Glycoside Derivative and L-
Proline. WO2007114475A1, 2007.
(34) For examples of other reactions of benzyl carbonates, see:
Displacement with carbon nucleophiles: (a) Kuwano, R.; Kondo, Y.;
Matsuyama, Y. Palladium-Catalyzed Nucleophilic Benzylic Substitu-
tions of Benzylic Esters. J. Am. Chem. Soc. 2003, 125, 12104−12105.
(b) Trost, B. M.; Czabaniuk, L. C. Palladium-Catalyzed Asymmetric
Benzylation of 3-Aryl Oxindoles. J. Am. Chem. Soc. 2010, 132, 15534−
15536. Coupling with C−H activation: (c) Mukai, T.; Hirano, K.;
Satoh, T.; Miura, M. Palladium-Catalyzed Direct Benzylation of
Azoles with Benzyl Carbonates. Org. Lett. 2010, 12, 1360−1363.
(35) A similar strategy was reported for a revised synthesis of
tofogliflozin. See: Ohtake, Y.; Emura, T.; Nishimoto, M.; Takano, K.;
Yamamoto, K.; Tsuchiya, S.; Yeu, S.-Y.; Kito, Y.; Kimura, N.; Takeda,
S.; Tsukazaki, M.; Murakata, M.; Sato, T. Development of a Scalable
Synthesis of Tofogliflozin. J. Org. Chem. 2016, 81, 2148−2153.
̈
(36) Grasa, G. A.; Guveli, T.; Singh, R.; Nolan, S. P. Efficient
Transesterification/Acylation Reactions Mediated by N-Heterocyclic
Carbene Catalysts. J. Org. Chem. 2003, 68, 2812−2819.
(37) For a review of DMC chemistry, see: (a) Tundo, P.; Selva, M.
The Chemistry of Dimethyl Carbonate. Acc. Chem. Res. 2002, 35,
706−716. For esterfication of carboxylic acids, see: (b) Ji, Y.;
Sweeney, J.; Zoglio, J.; Gorin, D. J. Catalytic Methyl Transfer from
Dimethylcarbonate to Carboxylic Acids. J. Org. Chem. 2013, 78,
11606−11611.
(38) For a report on the isolation of similar compounds (including
bidentate phosphine ligands and complexes with arenes), see: Johns,
A. M.; Utsunomiya, M.; Incarvito, C. D.; Hartwig, J. F. A Highly
Active Palladium Catalyst for Intermolecular Hydroamination.
Factors that Control Reactivity and Additions of Functionalized
Anilines to Dienes and Vinylarenes. J. Am. Chem. Soc. 2006, 128,
1828−1839.
(39) For a related discussion of NMR monitoring of bidentate
phosphines, their oxides, and Pd complexes, see: Merritt, J. M.; Buser,
J. Y.; Campbell, A. N.; Fennell, J. W.; Kallman, N. J.; Koenig, T. M.;
Moursy, H.; Pietz, M. A.; Scully, N.; Singh, U. K. Use of Modeling
and Process Analytical Technologies in the Design of a Catalytic
Amination Reaction: Understanding Oxygen Sensitivity at the Lab
and Manufacturing Scales. Org. Process Res. Dev. 2014, 18, 246−256.
(40) For proposed catalyst activation modes of Pd−allyl−NHC
complexes, see ref 38 and: (a) Marion, N.; Navarro, O.; Mei, J.;
Stevens, E. D.; Scott, N. M.; Nolan, S. P. Modified (NHC)Pd(allyl)Cl
(NHC = N-Heterocyclic Carbene) Complexes for Room-Temper-
ature Suzuki−Miyaura and Buchwald-Hartwig Reactions. J. Am. Chem.
Soc. 2006, 128, 4101−4111. and references therein. For a discussion
of reduction of Pd(OAc)2 by bisphosphine ligands in the presence of
additives, see: (b) Amatore, C.; Jutand, A.; Thuilliez, A. Formation of
Palladium(0) Complexes from Pd(OAc)2 and a Bidentate Phosphine
Ligand (dppp) and Their Reactivity in Oxidative Addition. Organo-
metallics 2001, 20, 3241−3249. The involvement of Pd(I) dimer
equilibria cannot be discounted at this time. For a discussion, see:
Q
DOI: 10.1021/acs.oprd.8b00325
Org. Process Res. Dev. XXXX, XXX, XXX−XXX