Development and Scale-Up of a Manufacturing Route for the Non-nucleoside Reverse Transcriptase Inhibitor GSK2248761A (IDX-899): Synthesis of an Advanced Key Chiral Intermediate

Article abstract of DOI:10.1021/acs.oprd.7b00356

A new and improved synthetic route to an intermediate in the synthesis of the phosphinate ester GSK2248761A is described. In the key step, we describe the first process-scale example of a palladium-catalyzed phosphorus-carbon coupling to give the entire backbone of GSK2248761A in one telescoped stage in 65% average yield on a 68 kg scale. This unusual chemistry enabled the route to be reduced from six chemistry stages to four and eliminated a number of environmentally unfriendly reagents and solvents.

Full text of DOI:10.1021/acs.oprd.7b00356

Article
pubs.acs.org/OPRD
Cite This: Org. Process Res. Dev. XXXX, XXX, XXX−XXX
Development and Scale-Up of a Manufacturing Route for the Non-
nucleoside Reverse Transcriptase Inhibitor GSK2248761A (IDX-899):
Synthesis of an Advanced Key Chiral Intermediate
Richard Bellingham,†,§ Gary Borrett,† Guillaume Bret,†,§ Bernadette Choudary,†,§ David Colclough,†,§
Jerome Hayes,† John Hayler,† Neil Hodnett,† Alan Ironmonger,*,† Augustine Ochen,† David Pascoe,†
John Richardson,†,§ Erica Vit,†,§ Francois-Rene Alexandre,‡ Catherine Caillet,‡,∥ Agnes Amador,‡,∥
̧
́ ̀
Stephanie Bot,‡ Severine Bonaric,‡,∥ Daniel da Costa,‡ Marie-Pierre Lioure,‡ Arlene Roland,‡,∥
́
́
̀
Elodie Rosinovsky,‡,∥ Christophe Parsy,‡,∥ and Cyril B. Dousson‡
†GlaxoSmithKline, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, U.K.
‡Merck & Co., Inc.,126 East Lincoln Avenue, Rahway, New Jersey 07065, United States
ABSTRACT: A new and improved synthetic route to an intermediate in the synthesis of the phosphinate ester GSK2248761A is
described. In the key step, we describe the first process-scale example of a palladium-catalyzed phosphorus−carbon coupling to
give the entire backbone of GSK2248761A in one telescoped stage in 65% average yield on a 68 kg scale. This unusual chemistry
enabled the route to be reduced from six chemistry stages to four and eliminated a number of environmentally unfriendly
reagents and solvents.
resulting arylmagnesium species 3 was then quenched with
diethyl chlorophosphite to give intermediate 4, which
rearranged to ethyl H-phosphinate 5 upon workup with dilute
hydrochloric acid. A toluene solution of 5 then underwent
palladium-catalyzed phosphonylation with the iodocinnamoni-
trile 6 starting material to produce ethyl phosphinate ester 7.5
This ethyl phosphinate 7 was then cleaved with bromotrime-
thylsilane and subsequently chlorinated using oxalyl chloride in
the presence of a catalytic amount of N,N-dimethylformamide
(DMF). The resulting phosphinyl chloride was then quenched
in situ with methanol to deliver the backbone of GSK2248761A
containing the required methyl phosphinate moiety. Ester 8
was saponified with lithium hydroxide in aqueous acetonitrile
with the concomitant removal of the N-benzenesulfonyl
protecting group, which provided a carboxylic acid handle for
resolution. A classical resolution was performed using
cinchonidine as the resolving agent, followed by salt
dissociation to give the enantiomerically pure acid 10, a key
intermediate in the synthesis. The final amide bond formation
was performed via activation with 1,1′-carbonyldiimidazole
(CDI) followed by treatment with gaseous ammonia, and the
product was crystallized from aqueous methanol to give
GSK2248761A (1).
INTRODUCTION
GSK2248761A (fosdevirine, IDX-899, 1) (Figure 1) is an
experimental non-nucleoside reverse transcriptase inhibitor
■
Figure 1. Chemical structure of GSK2248761A (IDX-899).
(NNRTI) discovered by Idenix Pharmaceuticals that was in
development for the treatment of HIV.1 The unusual structure
is P-stereogenic and was required as a single S enantiomer.
Interestingly, the R enantiomer has been shown to be 1800-fold
less active against the wild-type virus and inactive against
NNRTI-resistant viruses.2 Multikilogram quantities were
required to support the clinical development program, and a
more efficient manufacturing route was sought for long-term
commercial production. Herein we describe the development
of an alternative route to compound 10, a key intermediate in
the synthesis of GSK2248761A.
A number of issues were identified within the first-generation
route. Although the synthesis of the Grignard reagent
proceeded smoothly, there were safety and environmental
concerns over the use of superstoichiometric amounts of n-
butyllithium and n-butylmagnesium chloride, which were
needed to generate the active “ate” complex 3 under cryogenic
conditions. A slight excess of n-butyllithium was required for
FIRST-GENERATION SYNTHESIS OF GSK2248761A
■
The first-generation chemistry route is outlined in Scheme 1
and was used by an outsource partner to prepare supplies for
early-phase clinical and preclinical development.3 The synthesis
commenced with Mg−Br exchange on N-benzenesulfonyl-
protected bromoindole 2 using the tributylmagnesate method
described by Oshima under cryogenic conditions.4 The
Received: November 10, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.7b00356
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 1. First-Generation Route to GSK2248761A
this reaction to proceed robustly, as it was observed that only
the “ate” complex was active in the reaction and the
corresponding dibutylmagnesium species was completely
inactive toward Br−Mg exchange. Moreover, when the reaction
was performed under noncryogenic conditions, impurity 11
resulting from addition of a butyl group at the 3-position of the
indole (Figure 2) was poorly controlled. The use of alternative
rather than the required methyl phosphinate 8, which required
later modification in the synthesis. These steps involved the use
of dichloromethane, a nonpreferred solvent, and the combina-
tion of oxalyl chloride and DMF, which is known to generate
dimethylcarbamoyl chloride (DMCC) as a byproduct, as well
as the gaseous byproducts generated during processing, which
presented potential engineering/safety issues. DMCC is
reasonably anticipated to be a human carcinogen,7 and its
generation is strictly governed by COMAH regulations in the
U.K.8 Some of the batches of bromotrimethylsilane had been
shown in the laboratory to partially isomerize the cinnamoni-
trile double bond, forming impurity 12 (Figure 2), which was
difficult to purge in the downstream chemistry. This impurity is
probably formed as a result of the presence of small quantities
of HBr present in the bromotrimethylsilane. Finally, although
the yield of the classical resolution was reasonable, such a late
resolution in the synthesis severely impeded the efficiency of
the overall route. Herein we report our progress in developing
new chemistry to obtain the key late-stage intermediate 10 that
largely circumvents some of the above issues.
Figure 2. Two process impurities formed.
Grignard technologies, such as the isopropylmagnesium
chloride−lithium chloride reagent developed by Knochel,6
was unsuccessful. The diethyl chlorophosphite electrophile
used in the process is highly toxic and moisture-sensitive,
representing potential handling/quality issues, and the resulting
toluene solution of the ethyl H-phosphinate product 5 had to
be stored in a cold room because of concerns over its stability.
The palladium-catalyzed P−C coupling performed well, but the
product 7 was isolated in only 53% yield over the two stages.
However, this chemistry installed the ethyl phosphinate 7
NEW ROUTE DEVELOPMENT
■
We were keen to avoid the necessity of correcting the ethyl
group in 7 to a methyl group in 8 during the synthesis, as this
would eliminate much of the problematic chemistry and greatly
improve the efficiency. The corresponding dimethyl chlor-
ophosphite electrophile had limited commercial availability and
had previously been reported to have limited stability on
B
DOI: 10.1021/acs.oprd.7b00356
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
storage.9 Our limited attempts to prepare a pure sample of
dimethyl chlorophosphite were unsuccessful. We became
interested in the palladium-catalyzed formation of H-
phosphinate esters reported by Montchamp and others
(Scheme 2).5a,10 This chemistry would allow the direct
when we used Pd(dppf)Cl2 and 3-aminopropyltrimethoxysilane
in conjunction with anilinium hypophosphite at reflux in ethyl
acetate (entry 6). Attempts to avoid the use of aniline by
screening other salts of hypophosphorous acid such as
ammonium hypophosphite and sodium hypophosphite resulted
in reduction as the major pathway, with a <5% yield of the
desired product observed and concomitant precipitation of
palladium black. The methylating agent also proved to be
important: when 3-aminopropyltrimethoxysilane was replaced
with tetramethyl orthosilicate or with trimethoxyphenylsilane
and an organic base, the reaction stalled. Interestingly, when the
methyl phosphinate ester of hypophosphoric acid (prepared
either from trimethyl orthoformate5a or tetramethyl orthosili-
cate,10b isolated, and characterized by 31P NMR spectroscopy)
was used as a source of phosphorus, the reaction stalled at less
than 10% conversion. The most efficient coupling partner was
N-benzenesulfonyl bromoindole derivative 2, and we believe
that the protecting group increased the rate of reaction because
of withdrawal of electron density from the indole ring,
facilitating oxidative addition of the palladium species. Our
deconstruction of the reaction suggested that the mechanism
involved the coupling of hypophosphorous acid followed by
methylation of the coupled derivative, releasing the product
from the catalyst.
We also investigated the coupling in the alternative synthetic
direction starting from cinnamonitrile derivative 6 and found
that methyl phosphinate 27 could be synthesized in 80% yield
but was unstable and readily underwent demethylation to form
28, followed by oxidation to the phosphonate 29 in air upon
storage (Scheme 3). Our optimized reaction conditions for the
reaction giving 30 involved performing the reaction in N-
methyl-2-pyrrolidinone (NMP) spiked with a small quantity of
water, and it was necessary to add trimethyl orthoformate to
retard demethylation during coupling. These issues could be
partially circumvented by telescoping the second coupling to
iodoindole derivative 16, but we chose to pursue the other
reaction direction, as it seemed more amenable to process
development.
Scheme 2. Palladium-Catalyzed P−C Coupling Utilizing
Anilinium Hypophosphite
formation of the H-phosphinate methyl ester, which could
then be followed by a second palladium-catalyzed reaction to
construct the complete framework of GSK2248761A.
Idenix also noted this chemistry before the project was in-
licensed to GSK and had instructed Johnson-Matthey to
perform a catalyst screen, which demonstrated that in principle
this transformation should be possible.11 Building on the
Pd(dppf)Cl2 catalyst conditions identified by Johnson-Matthey,
we investigated many alternative coupling conditions/sub-
strates to establish the most feasible direction and reaction
conditions for the coupling (Table 1).
The coupling was investigated starting initially with a
haloindole derivative to produce a phosphinic ester that
could then be coupled onward. Three major impurity pathways
were apparent in the reaction profile depending on the coupling
conditions and the reaction partner. Direct reduction of the
haloindole can occur to give the parent indole (e.g., 25 and 26),
hydrolysis of the methyl phosphinate intermediate can occur
upon prolonged heating (e.g., 21, 22, and 24), and the
protecting group could also be cleaved during the reaction (e.g.,
23, 24, and 26). We found that the reductive dehalogenation
could be controlled by judicious choice of the catalyst, methyl
source, and phosphinate salt. Our best results were obtained
Table 1. Summary of P−C Coupling Reactions
a
b
entry
substrate
conditions
product distribution
1
2
3
4
5
6
17
16
16
19
18
2
A
A
B
17, 59%; 23, 0%; 26, 36%
2, 26%; 23, 35%; 26, 34%
2, 4%; 23, 55%; 26, 24%
19, 8%; 22, 34%; 23, 13%
18, 10%; 22, 64%; 23, 8%
2, 5%; 20, 82%; 23, 2%
C
C
C
a
Conditions A: 2 mol % Pd(OAc)2, 2 mol % dppp, 1.2 equiv of anilinium hypophosphite, 1.2 equiv of (3-aminopropyl)trimethoxysilane, MeCN, 82
°C. Conditions B: 2 mol % Pd(dppf)Cl2, 1.2 equiv of anilinium hypophosphite, 1.2 equiv of (3-aminopropyl)trimethoxysilane, MeCN, 82 °C.
Conditions C: 2 mol % Pd(dppf)Cl2, 1.2 equiv of anilinium hypophosphite, 1.2 equiv of (3-aminopropyl)trimethoxysilane, EtOAc, 77 °C.
b
Determined as HPLC % a/a of species detected at 220 nm.
C
DOI: 10.1021/acs.oprd.7b00356
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 3. Palladium-Catalyzed Coupling Starting with Aryl Iodide 6
Scheme 4. Telescoped Pd-Catalyzed P−C Coupling
Table 2. Summary of Resolution Experiments
a
b
b
b
entry
conditions
first % d.r. of salt
second % d.r. of salt
% e.r. of 10
% yield
1
2
3
A
B
C
97.0
96.5
96.2
99.3
99.8
99.8
99.3
99.8
99.8
42
40
40
a
Conditions A: 1 equiv of cinchonidine in acetone at 50 °C, then cool to 20 °C, solvent swap to EtOAc, filter, wash, reslurry in EtOAc at 50 °C, then
cool to 20 °C, filter. Suspend in EtOAc, wash with 0.1 M HCl(aq), water, and brine, solvent swap to n-heptane to crystallize. Conditions B: 1 equiv
of cinchonidine in EtOAc at 50 °C, then cool to 20 °C, filter, wash. Suspend in EtOAc, wash with 0.1 M HCl(aq), water, and brine, azeotrope dry,
charge with 1 equiv of cinchonidine in EtOAc at 50 °C, then cool to 20 °C, filter, wash. Suspend in EtOAc, wash with 0.1 M HCl(aq), water, and
brine, solvent swap to n-heptane to crystallize. Conditions C: 1 equiv of cinchonidine in EtOAc at 50 °C, then cool to 20 °C, filter, wash, reslurry in
acetone at 50 °C, then cool to 20 °C, filter. Suspend in EtOAc, wash with 0.1 M HCl(aq), water, and brine, solvent swap to methylcyclohexane to
b
crystallize. ChiralPak AD-H column (0.46 cm × 25 cm), isocratic elution with 75:25:0.04 isopropanol/methanol/TFA, detection at 270 nm.
We attempted to perform a one-pot reaction where a
solution of intermediate 20 was treated with 1 equiv of aryl
iodide 6 and triethylamine in ethyl acetate to form methyl
phosphinate ester 8 directly (Scheme 4). A large amount of
demethylation of 8 occurred during the reaction, and we
hypothesized that the iodide liberated as the reaction
proceeded was responsible via SN2 attack onto the methyl
group. The demethylation pathway was shut down when
triethylamine was replaced with dicyclohexylamine in order to
generate an insoluble iodide salt that would not participate in
this side reaction. More detailed studies revealed that ethyl
acetate was not completely inert in the coupling reaction and
caused methyl to ethyl phosphinate ester exchange to give the
undesired ethyl phosphinate ester 7 at low but unacceptable
D
DOI: 10.1021/acs.oprd.7b00356
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 3. Representative PXRD traces of cinchonidine salt 31 slurried in (a) ethyl acetate (red trace) and acetone (green trace).
levels that were not purged downstream. The solvent was
switched to isopropyl acetate to retard this pathway, and the
reaction was performed in a 2 L controlled lab reactor (CLR)
to evaluate the potential scale-up. The first coupling performed
as desired, but the dicyclohexylamine hydrogen iodide
produced as the second reaction progressed, in combination
with polymeric silane byproducts, produced large hard balls
that we anticipated would cause damage to our vessels as well
as engineering challenges upon scale-up. We elected to perform
a weakly acidic workup of 20 followed by a filtration to remove
the problematic silane byproducts, and we changed the solvent
to toluene to facilitate this. The toluene solution of methyl H-
phosphinate ester 20 was azeotropically dried, and we found
that solid potassium carbonate was an excellent base in the
second coupling, as no demethylation occurred and the rate of
reaction was significantly increased. Indeed, only an additional
charge of 0.5 mol % Pd(dppf)Cl2 catalyst was required to
robustly perform the reaction within 2 h. After hot aqueous
workup and azeotropic distillation, methyl phosphinate ester 8
crystallized as a toluene solvate, which provided excellent
purification to give >98% a/a HPLC purity. The new
telescoped process was successfully performed in the pilot
plant in four batches, affording the product in 65% average
yield, with some initial minor processing issues associated with
the filtration of the silane polymer that were overcome with the
addition of Celite to the vessel.
cinchonidine had previously been shown to be a very efficient
resolving agent.3 The addition of 1 equiv of cinchonidine
(based on assay) to an acetone solution caused the desired
diastereomeric salt 31 to crystallize out. The solvent was
exchanged to ethyl acetate to increase the yield, and 31 was
filtered to provide material with 96−97% d.r., which was
immediately reslurried in ethyl acetate at 50 °C, followed by a
cooling ramp to room temperature and filtration. The salt was
washed with ethyl acetate to give material with >99.1% d.r. that
was subsequently converted into chirally enriched S enantiomer
10 in an excellent yield of 42% for the whole process (Table 2).
However, in our hands we were unable to robustly obtain the
chiral upgrade upon reslurrying in ethyl acetate under a variety
of conditions, including various seeding and temperature
cycling regimes. In order to meet our timelines, we chose to
regenerate the free acid and reform the salt with cinchonidine
to generate the required d.r. Upon scale-up to 93 kg, the first
salt formation was performed in ethyl acetate at 50 °C, followed
by cooling to 20 °C, filtration, and washing to give 96.5% d.r.
The second salt formation, filtration, and washing provided
material with 99.8% d.r., which was dissociated again, and the
product 10 was obtained upon crystallization from n-heptane in
an overall yield of 40% with 99.8%e.r. As the use of two
cinchonidine salt formations was wasteful and potentially
unnecessary, we revisited the reslurry conditions to upgrade the
chiral purity.
Ester 8 was saponified and deprotected using lithium
hydroxide in aqueous acetonitrile at 40 °C on a 125 kg scale.
The product 9 crystallized directly from the reaction mixture as
the hydrate upon adjustment of the pH using aqueous
hydrochloric acid in 90% yield with excellent purity. The acid
provided a handle for a classical resolution in which the alkaloid
During our investigations, comparison of the powder X-ray
diffraction (PXRD) traces of cinchonidine salt 31 formed in
different solvents (Figure 3) revealed that a form change upon
reslurrying was a potential source of the irreproducibility. The
initial salt formation performed in ethyl acetate gave form I
(red trace), which when dried and reslurried in acetone at 50
E
DOI: 10.1021/acs.oprd.7b00356
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
°C and then aged at 20 °C gave form II (green trace).
Interestingly, when the reslurrying was performed in this order
a chiral upgrade from 96.2% d.r. to >99.5% d.r. was observed.
Conversely, when the crystallization was performed in acetone
and the salt was subsequently reslurried in either fresh acetone
or ethyl acetate, there was no change in form or upgrade in d.r.,
suggesting that the change in form was in part responsible for
the chiral upgrade upon reslurrying. The new process was
scaled up to a 5 L CLR scale to demonstrate that this new
reslurrying process was not scale-dependent. Gratifyingly, the
reslurry protocol performed as expected to give acid 10 with
99.8% e.r. in 39% overall yield with excellent purity.
In summary, we have demonstrated the first pilot-plant-scale
palladium-catalyzed H-phosphinate ester coupling to produce
the entire backbone of GSK2248761A. Compared with the
previous route, the new route is two stages shorter, eliminates
most of the materials of concern for production, controls the
impurities identified as problematic, and improves the overall
yield from 15% to 23%. The key intermediate produced in this
synthesis was successfully converted into GSK2248761A active
pharmaceutical ingredient that met clinical specifications, and
the procedure will be reported in a subsequent publication.
were separated, and the organic phase was washed with
aqueous sodium chloride (153 kg of a 20% w/w solution) and
distilled down to ∼200 L under reduced pressure. Toluene
(153 L) was charged into the vessel, and the crystallization
mixture was heated to 70 °C, cooled to 20 °C over ∼2.5 h, and
stirred for 4 h. The protected indole phosphinate 8 (56.3 kg,
72%) was isolated by filtration, washed with toluene (3 × 103
1
L), and dried at 60 °C under vacuum in an agitated drier. H
NMR (500 MHz, CDCl3): δ 1.45 (t, J = 7.1 Hz, 3H), 2.35 (s,
3H), 2.40 (s, 3H), 3.79 (d, J = 11.7 Hz, 3H), 4.53 (q, J = 7.1
Hz, 2H), 5.92 (d, J = 16.7 Hz, 1H), 7.13−7.19 (m, 3H), 7.23−
7.27 (m, 2H), 7.32−7.39 (m, 3H), 7.52 (t, J = 7.9 Hz, 2H),
7.64 (t, J = 7.4 Hz, 1H), 7.75 (dd, J = 13.4, 8.0 Hz, 2H), 7.85
(d, J = 1.9 Hz, 1H), 7.94 (dd, J = 8.8, 1.3 Hz, 1H), 8.08 (d, J =
7.6 Hz, 2H). 13C NMR (126 MHz, CDCl3): δ 12.1, 19.5, 19.7,
50.3, 50.4, 61.8, 96.2, 113.5, 115.9, 120.0, 123.5, 125.3, 125.4,
125.9, 126.4, 127.3, 127.7, 127.8, 129.2, 130.3, 132.2, 132.3,
133.2, 135.3, 136.1, 138.0, 147.5, 159.9.
(E)-5-Chloro-3-((3-(2-cyanovinyl)-5-methylphenyl)-
(methoxy)phosphoryl)-1H-indole-2-carboxylic Acid Hydrate
(9). Lithium hydroxide monohydrate (5.3 kg, 126 mol) was
added to a stirred solution of protected indole phosphinate 8
(42.6 kg, 63.1 mol) in acetonitrile (202 kg) and water (128 L),
and the resulting mixture was stirred at 22 °C for 2 h. The
reaction mixture was heated to 40 °C for 12 h and then cooled
back to 25 °C, and the pH was adjusted to pH 1−1.5 using
aqueous hydrochloric acid (2 M solution). Water (149 L) was
added over 30 min and the resulting slurry was aged at 40 °C
for 30 min, cooled to 25 °C over 2 h, and then stirred at this
temperature for a further 2 h. Acid 9 (23.1 kg, 85%) was
isolated by filtration, washed with 1:2 acetonitrile/water (86 L),
water (85 L), and 9:1 acetone/water (86 L), and dried at 70 °C
under vacuum in an agitated drier. 1H NMR (500 MHz,
DMSO-d6): δ 2.35 (s, 3H), 3.72 (d, J = 11.7 Hz, 3H), 6.51 (d, J
= 16.7 Hz, 1H), 7.38 (dd, J = 8.9, 1.3 Hz, 1H), 7.65−7.73 (m,
3H), 7.85 (d, J = 13.6, 1H), 7.99 (d, J = 1.9 Hz, 1H), 12.97 (br
s, 1H), 14.38 (br s, 1H). 13C NMR (125 MHz, DMSO-d6): δ
20.7, 51.7, 98.1, 115.0, 118.5, 120.9, 125.4, 126.7, 127.8, 127.9,
131.2, 133.5, 133.9, 134.0, 134.6, 134.7, 138.8, 138.9, 149.6,
160.4.
EXPERIMENTAL SECTION
■
General Experimental. Reactions were monitored using
HPLC on a Luna C18(2) column (50 mm × 2 mm, 3 μm),
eluting with a gradient of 0 to 95% (0.05% v/v TFA in water to
0.05% v/v TFA in acetonitrile) over 8 min at 40 °C with
detection at 220 nm. Chiral HPLC was performed on a
ChiralPac AD-H column (250 mm × 4.6 mm) with isocratic
elution (70:30 n-heptane/ethanol) at 1 mL/min over 20 min
with detection at 270 nm. NMR spectra were recorded on a
Bruker 400 MHz Ultrashield or Bruker AV500 spectrometer.
The following abbreviations are used to explain the multi-
plicities: s = singlet, d = doublet, t = triplet, q = quartet, m =
multiplet, b = broad, dd = doublet of doublets, dt = doublet of
triplets. Mass spectra were recorded on a Waters ZQ mass
spectrometer, and high-resolution mass spectrometry (HRMS)
was performed on an LQT Orbitrap spectrometer.
Syntheses. 1-Benzenesulfonyl-5-chloro-3-[methyl-3-((E)-
2-cyanovinyl)-5-methylphenyl]phosphinoyl-1H-indole-2-car-
boxylic Acid Ethyl Ester (8). Anilinium hypophosphite (27.5 kg,
173 mol) was suspended in toluene (232 L), and 3-
aminopropyltrimethoxysilane (31 kg, 173 mol) was added
while the temperature was maintained at 25 5 °C. After 30
min, the resulting solution was transferred to another vessel
containing bromoindole 2 (51 kg, 115 mol) and Pd(dppf)Cl2·
acetone (3.6 kg, 4.6 mol) in toluene (385 L) over
approximately 1.5 h while the temperature was maintained at
77−82 °C. After 30 min, the reaction mixture was cooled to 20
°C, and aqueous hydrochloric acid (0.1 M, 306 kg) was added,
followed by the addition of Celite (6 kg). The suspension was
stirred for 30 min and filtered, and the Celite was washed with
toluene (103 L). The phases were separated, and the organic
phase was washed with aqueous sodium chloride (153 kg of a
20% w/w solution), after which ∼95 L was distilled off under
reduced pressure. The reaction solution was transferred to
another vessel containing iodocinnamonitrile 6 (29.2 kg, 109
mol), potassium carbonate (20.9 kg, 151 mol), and Pd(dppf)-
Cl2·acetone (0.5 kg, 0.63 mol). The reaction mixture was
heated to 75 °C for 2 h and then cooled to 35 °C, and a
solution of sodium hydrogen carbonate (153 kg of a 7% w/w
solution) was added, followed by THF (409 L). The phases
(R,E)-5-Chloro-3-((3-(2-cyanovinyl)-5-methylphenyl)-
(methoxy)phosphoryl)-1H-indole-2-carboxylic Acid (10). Cin-
chonidine (64.7 kg, 220 mol) was added to a stirred solution of
carboxylic acid 9 (93.2 kg, 225 mol) in ethyl acetate (1126 L) at
50 °C. The suspension was stirred at 50 °C for 2 h, cooled to
22 °C for 2 h, and then filtered, washing with ethyl acetate (300
L). The damp cake was reloaded into the reaction vessel and
suspended in ethyl acetate (1180 L) and aqueous hydrochloric
acid (0.5 M, 465 kg) until dissolved. The solution was then
filtered through an R55SP Cuno cartridge, and the layers were
separated. The organic phase washed with water (298 L) and
aqueous sodium chloride solution (10% w/w, 2 × 199 kg) and
then azeotroped dry, adjusting to a final volume of
approximately 930 L. The solution was heated to 50 °C, and
cinchonidine (32.4 kg, 110 mol) was added. The resulting
slurry was stirred for 2 h, cooled to 22 °C for 2 h, filtered, and
washed with ethyl acetate (300 L). The damp cake was
reloaded into the reaction vessel and suspended in ethyl acetate
(940 L) and aqueous hydrochloric acid (0.5 M, 466 kg) until
dissolved. The layers were separated, and the organic phase was
washed with water (279 L) and aqueous sodium chloride
solution (10% w/w, 2 × 199 kg), after which the solvent was
swapped to n-heptane (745 L) via distillation. The crystal-
F
DOI: 10.1021/acs.oprd.7b00356
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(4) (a) Kitagawa, K.; Inoue, A.; Shinokubo, H.; Oshima, K. Angew.
Chem., Int. Ed. 2000, 39, 2481. (b) Inoue, A.; Kitagawa, K.; Shinokubo,
H.; Oshima, K. J. Org. Chem. 2001, 66, 4333.
(5) (a) Lei, H.; Stoakes, M. S.; Schwabacher, A. W. Synthesis 1992,
1992, 1255. (b) Xu, Y.; Li, Z.; Xia, J.; Guo, H.; Huang, Y. Synthesis
1983, 1983, 377.
(6) Krasovskiy, A.; Knochel, P. Angew. Chem., Int. Ed. 2004, 43, 3333.
(7) National Toxicology Program. Dimethylcarbamoyl Chloride,
CAS No. 79-44-7. In 12th Report on Carcinogens, 12th Addition; U.S.
Department of Health and Human Services: Research Triangle Park,
NC, 2011.
(8) Control of Major Accident Hazards (COMAH). http://www.hse.
gov.uk/comah/.
(9) Lippman, A. E. J. Org. Chem. 1965, 30, 3217.
(10) (a) Montchamp, J.-L.; Dumond, Y. R. J. Am. Chem. Soc. 2001,
123, 510. (b) Dumond, Y. R.; Baker, R. L.; Montchamp, J.-L. Org. Lett.
lization mixture was cooled to 10 °C and stirred for 2 h, and the
crystals were filtered off, washed with n-heptane (185 L), and
dried in a vacuum oven at 58 °C to give chiral acid 10 (34.4 kg,
39.7%, 99.8% e.r.).
Alternative Process for 10. Cinchonidine (204.1 g, 693
mmol) was added to a stirred solution of carboxylic acid 9 (300
g, 693 mmol) in ethyl acetate (3.6 L) at 50 °C. The suspension
was stirred at 50 °C for 2 h, cooled to 22 °C over 1 h, held for 2
h, and then filtered off, washing with ethyl acetate (900 mL).
The damp cake was reloaded into the reaction vessel, stirred in
acetone (3 L) at 50 °C for 1.5 h, cooled to 22 °C, stirred for 1
h, filtered off, and washed with acetone (600 mL) and ethyl
acetate (600 mL). The damp cake was reloaded into the
reaction vessel and stirred in ethyl acetate (3.3 L). Aqueous
hydrochloric acid (0.5 M, 1 L) was added, and the were layers
separated. The organic phase washed with water (600 mL) and
aqueous sodium chloride solution (10% w/w, 2 × 400 mL) and
then azeotroped dry, adjusting to a final volume of 900 mL.
The solvent was swapped to dry methylcyclohexane (900 mL)
via distillation under reduced pressure. The crystallization
mixture was cooled to 20 °C and held overnight. The crystals
were filtered off, washed with methylcyclohexane (600 mL),
and dried in a vacuum oven at 60 °C to give 10 (113 g, 39%,
99.8% e.r.).
̀
2000, 2, 3341. (c) Deprele, S.; Montchamp, J.-L. J. Organomet. Chem.
2002, 643−644, 154. (d) Xu, Y.; Wei, H.; Xia, J. Liebigs Ann. Chem.
1988, 1988, 1139. (e) Holt, D. A.; Erb, J. M. Tetrahedron Lett. 1989,
30, 5393.
(11) We acknowledge Johnson-Matthey Catalysts, Catalysis and
Chiral Technologies for initial reaction screening and coupling
experiments.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: Alan.2.Ironmonger@gsk.com.
ORCID
Jerome Hayes: 0000-0001-9875-0464
John Hayler: 0000-0003-3685-3139
Franco̧ is-Rene Alexandre: 0000-0003-1396-3757
́
Notes
The authors declare no competing financial interest.
§R.B., G.B., B.C., D.C., J.R., and E.V. are former GSK
employees.
∥C.C., A.A., S.B., A.R., E.R., and C.P. are former Idenix
employees.
REFERENCES
■
(1) (a) Zhou, X.-J.; Pietropaolo, K.; Damphousse, D.; Belanger, B.;
Chen, J.; Sullivan-Bolyai, J.; Mayers, D. Antimicrob. Agents Chemother.
2009, 53, 1739. (b) Storer, R.; Dousson, C.; Alexandre, F.-R.; Roland,
A. US20060074054 A1, 2006. (c) Castellino, S.; Groseclose, M. R.;
Sigafoos, J.; Wagner, D.; de Serres, M.; Polli, J. W.; Romach, E.; Myer,
J.; Hamilton, B. Chem. Res. Toxicol. 2013, 26, 241. (d) Klibanov, O.
M.; Kaczor, R. L. Curr. Opin. Invest. Drugs 2010, 11, 237.
(2) (a) Alexandre, F.-R.; Amador, A.; Bot, S.; Caillet, C.; Convard, T.;
Jakubik, J.; Musiu, C.; Poddesu, B.; Vargiu, L.; Liuzzi, M.; Roland, A.;
Seifer, M.; Standring, D.; Storer, R.; Dousson, C. B. J. Med. Chem.
2011, 54, 392. (b) Dousson, C.; Alexandre, F.-R.; Amador, A.; Bonaric,
S.; Bot, S.; Caillet, C.; Convard, T.; da Costa, D.; Lioure, M.-P.;
Roland, A.; Rosinovsky, E.; Maldonado, S.; Parsy, C.; Trochet, C.;
Storer, R.; Stewart, A.; Wang, J.; Mayes, B. A.; Musiu, C.; Poddesu, B.;
Vargiu, L.; Liuzzi, M.; Moussa, A.; Jakubik, J.; Hubbard, L.; Seifer, M.;
Standring, D. J. Med. Chem. 2016, 59, 1891.
(3) (a) Storer, R.; Alexandre, F.-R.; Dousson, C.; Moussa, A. M.;
Bridges, E. PCT Int. Appl. WO2008/042240 A2, 2008. (b) Alexandre,
F.-R.; Amador, A.; Bonaric, S.; Bot, S.; Caillet, C.; Da Costa, D.;
Lioure, M. P.; Mayes, B. A.; Parsy, C.; Roland, A.; Rosinovsky, E.;
Stewart, A.; Wang, J.; Moussa, A.; Dousson, C. Presented at the 236th
ACS National Meeting, Philadelphia, PA, Aug 17−21, 2008.
G
DOI: 10.1021/acs.oprd.7b00356
Org. Process Res. Dev. XXXX, XXX, XXX−XXX