Process development and large-scale synthesis of MK-6186, a non-nucleoside reverse transcriptase inhibitor for the treatment of HIV

Article abstract of DOI:10.1021/op200334x

A new synthetic route has been developed to drug candidate 1, a second-generation NNRTI being developed as a potential treatment of HIV. Regiocontrol in a key alkylation step was achieved by selective N-alkylation of hydrazone 13. After a deprotection and cyclisation sequence, 1 was isolated in six steps in 35% overall yield from readily available starting materials.

Full text of DOI:10.1021/op200334x

Article
pubs.acs.org/OPRD
Process Development and Large-Scale Synthesis of MK-6186,
a Non-Nucleoside Reverse Transcriptase Inhibitor for the
Treatment of HIV
Adrian Goodyear,*,† Xin Linghu,*,‡ Brian Bishop,† Cheng Chen,‡ Ed Cleator,† Mark McLaughlin,‡
Faye J. Sheen,† Gavin W. Stewart,† Yingju Xu,‡ and Jingjun Yin‡
†Global Process Chemistry, Merck Sharp and Dohme Research Laboratories, Hertford Road, Hoddesdon, Hertfordshire, EN11 9BU, U.K.
‡Global Process Chemistry, Merck Sharp and Dohme Research Laboratories, Rahway, New Jersey 07065, United States
ABSTRACT: A new synthetic route has been developed to drug candidate 1, a second-generation NNRTI being developed as a
potential treatment of HIV. Regiocontrol in a key alkylation step was achieved by selective N-alkylation of hydrazone 13. After a
deprotection and cyclisation sequence, 1 was isolated in six steps in 35% overall yield from readily available starting materials.
Proposed Process Chemistry Approach to 1. In
designing a second-generation route to 1, the primary goal
was achieving regiocontrol in the N-alkylation. The direct
indazole alkylation had been shown to be nonselective, and
thus, an alternative strategy was required. It was postulated that
the desired outcome could be achieved through alkylation of a
hydrazone such as 13 to give an open hydrazone precursor 14
to the required indazolone (Scheme 2). Once formed, 14 could
then be treated with aqueous acid to invoke a tandem acid
hydrolysis and cyclisation, to afford the desired indazole
product 1 regio-selectively. Bromo compound 8 was readily
available using chemistry previously described at Merck.4
Hydrazone 13 was anticipated to be formed by palladium
mediated cross-coupling of bromide 12 with benzophenone
hydrazone. Bromide 12 itself would be accessed by an
adaptation of the medicinal chemistry route to 5.
INTRODUCTION
■
Inhibition of the reverse transcriptase (RT) enzyme present in
the human immunodeficiency virus (HIV) represents a key
method for the treatment of the HIV infection. The emergence
of non-nucleoside reverse transcriptase inhibitors (NNRTIs)
has provided a welcome boost against the resistance
experienced by the older class of nucleoside reverse tran-
scriptase inhibitors (NRTIs) by viral mutation. However,
recently marketed NNRTI treatments such as efavirenz,
nevirapine, and delavirdine have themselves met with viral
mutation and resistance.1 Second-generation NNRTIs, such as
etravirene2 have recently been developed specifically for
treatment-experienced adult patients with HIV resistance to
an NNRTI. Merck recently reported on the discovery and
development of a novel class of second-generation NNRTIs
containing the pyrazolo[3,4-b]pyridine and biaryl ether key
structural features.3,4 In a further elaboration of this sequence,
compound 1, now with a central indazole moiety was brought
forward for development.5 Herein, the evolution of a robust
synthesis of 1 which is amenable to scale, will be discussed.
Medicinal Chemistry Synthesis of 1. The initial
Medicinal Chemistry route to 1 is outlined in Scheme 1.5 Key
to this synthetic sequence was the formation of the indazole ring
system 6 by cyclisation of the aryl hydrazone formed between
aldehyde 5 and hydrazine hydrate.6 A palladium-mediated
cyanation of 6 gave nitrile 7, and subsequent alkylation with
bromide 8 gave the penultimate 9. Finally, Boc deprotection of 9
afforded the desired compound 1 in 5% overall yield.
Although successful for the generation of gram quantities of
1, a number of limitations rendered the Medicinal Chemistry
route unsuitable for larger-scale deliveries. First the use of toxic
reagents such as metallic cyanides and hydrazine hydrate are
highly undesirable for a long-term route. Second, poor
regiocontrol in the N-alkylation of 7 coupled with the tendency
of 9 to undergo Boc-deprotection followed by further alkylation
with bromide 8 resulted in the need for a chromatographic
separation of 9. This depressed the yield of this step to a
modest 40%. Finally, the low overall yield of just 5% over the
six steps was not viable for a long-term synthesis.
RESULTS AND DISCUSSION
■
Synthesis of Intermediate 17. The synthesis began with
the preparation of biaryl ether 17 using the commercially
available 4-bromo-1-chloro-2-fluorobenzene 10 as the starting
material (Scheme 3). Using the Medicinal Chemistry
formylation conditions as a starting point, 10 was deprotonated
using LDA at −50 °C in THF. The resulting anion 10′ was
treated with DMF at −50 °C to form the intermediate aminal
15. Quenching the reaction with water and extraction into
MTBE gave a solution of aldehyde 16. Crystallisation of the
aldehyde could then be effected by solvent switch to hexane
followed by cooling to −10 °C to maximise recovery. In the
laboratory, a modest 76% yield was obtained when employing
this protocol. Two key pieces of data provided the insight
required for the optimization of this process. First, examination
of the reaction profile by reverse phase HPLC showed that
>99% conversion to 15 had been achieved. Second, when an
MTBE solution of 16 (post water quench) was left to stand at
20 °C overnight, 30% of 10 was reformed. The inference was
Received: November 22, 2011
Published: January 24, 2012
© 2012 American Chemical Society
605
dx.doi.org/10.1021/op200334x | Org. Process Res. Dev. 2012, 16, 605−611

Organic Process Research & Development
Article
Scheme 1. Medicinal Chemistry synthesis of 1
Scheme 2. Proposed Process Chemistry route to 1
therefore that, after a water quench of 15, the intermediate
aminal could decompose via two competing pathways.
16. This was found to be the case. After a screen of suitable
acids it was found that quenching with 4 M hydrochloric acid
generated 16 exclusively. After further optimization, a protocol
of an inverse addition of the reaction mixture to a biphasic
mixture of 4 M hydrochloric acid and MTBE at 0 °C was
designed. This was not only critical in affording complete
conversion to 16 but also by using an addition controlled
quench, the observed exotherm could be modulated by varying
the addition rate.
Upon quenching of 15 with water or a weak acid (acetic
acid) the aminal can eliminate DMF to reform the stabilized
anion 10′ which is then protonated to give the starting material
10 (Scheme 4). Alternatively the desired outcome of
elimination of dimethylamine can occur to give aldehyde 16.
Working on the hypothesis that these decomposition pathways
could be influenced by the electronic state of the aminal
nitrogen, we reasoned that a quench with strong acid would
lead to protonation of the nitrogen, thus favoring elimination of
the dimethyl amino fragment to generate the desired product
Next, our attention turned to the SNAr coupling of 16 with
phenol 11 to prepare biaryl ether 17. Reviewing the Medicinal
Chemistry procedures showed the reaction could be carried out
606
dx.doi.org/10.1021/op200334x | Org. Process Res. Dev. 2012, 16, 605−611

Organic Process Research & Development
Article
Scheme 3. Preparation of biaryl ether 17
Scheme 4. Decomposition pathways of aminal 15
using K2CO3 in NMP at 60 °C. However, in our hands the
reaction was found to stall at ∼70% conversion, and also the
generation of an unidentified, structurally related impurity at
3% was observed.7 A change of solvent to DMF gave complete
conversion within 2 h at 60 °C but still led to the generation of
this unidentified impurity. After extensive optimisation of the
conditions, we found that the reaction could be reproducibly
performed on gram scale using 2 equiv K2CO3 and 8 volumes
of DMF at 20 °C over 8 h, with less than 1% of the impurity
generated. Biaryl ether 17 could then be isolated in 95% yield
by simply adding water and filtering the resulting slurry.
Identification of a commercial source of 11 obviated the need
to prepare this reagent in-house using zinc cyanide. Once bulk
supplies of 11 had been received, it was found to contain a new
impurity at 3%, assigned as methyl compound 18 by LC−MS.
Close control over this impurity and subsequent methyl
derivatives of our desired intermediates would be required as
translation of 18 through the synthesis to the final product 1
could be envisaged.
the base would be influenced greatly by the particle size. It was
likely that the stir bar imparted a beneficial grinding effect,
therefore making more base available for reaction. As the scale
of operation was increased, the stoichiometry of base required
to achieve a satisfactory level of conversion also increased. For
plant-scale operation, a protocol of portionwise addition of
5 equiv of K2CO3 was employed which ensured greater than
99% conversion to 17 after an overnight age at 20 °C.
Conversion of 10 to 17 on-scale was implemented in a single
batch campaign. Deprotonation, reaction with DMF, and acid
quench proceeded as expected to give 16 (7.50 kg, 95% assay
yield). The stream was sufficiently pure to allow us to telescope
the sequence. To this end the solution of 16 was solvent
switched from MTBE into DMF and this solution was used
directly in the SNAr to generate a total of 7.82 kg of 17 in an 85%
isolated yield. The main impurity was the methyl compound
generated from reaction of 18 with 16 and was present at 2.8%.
Synthesis of Key Hydrazone Intermediate 13. In
preparation for the coupling of the bromobiaryl ether 17 with
benzophenone hydrazone 19, it was first necessary to protect
17 to prevent any reaction through the aldehyde group. This
was easily achieved by conversion of 17 into the acetal 12 using
trimethylorthoformate in methanol (Scheme 5). Initially 0.1
Scheme 5. Preparation of acetal 12
As the process was scaled up, it became progressively more
difficult to drive the reaction to completion. One contributing
factor was believed to be the change in agitation method from
magnetic stir bar to overhead paddle stirrer. When employing a
heterogeneous system such as K2CO3 in DMF, the solubility of
607
dx.doi.org/10.1021/op200334x | Org. Process Res. Dev. 2012, 16, 605−611

Organic Process Research & Development
Article
Table 1. Catalyst and ligand screening for the cross coupling to 13
a
a
entry
solvent
base
ligand
conversion to 13 (area %)
bis-adduct 21 (area %)
1
2
3
4
5
6
7
2-Me THF
2-Me THF
2-Me THF
2-Me THF
2-Me THF
CPME
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
K3PO4
Xantphos
DPEphos
BINAP
dppf
95
97
93
96
93
91
93
6.5
3.0
2.7
2.0
1.6
1.3
1.4
dppf
Cs2CO3
K3PO4
dppf
CPME
dppf
a
Determined by HPLC analysis.
equiv of PTSA monohydrate catalyst was required, but
subsequently with bulk supplies of 17 exhibiting a higher
water content, an increase to 0.4 equiv was necessary to drive
the reaction to completion. Workup was performed by
neutralisation of the tosic acid with triethylamine, followed by
the addition of water to effect crystallisation. Acetal 12 was
isolated as an off-white solid in 95% yield.
switching from saturated to 2 M K3PO4 resulted in a much
slower reaction, if used in combination with preformed
PdCl2(dppf)·CH2Cl2 complex (2 mol %) a much improved
reaction rate was observed, with complete conversion being
obtained after 18 h at 90 °C. In addition, the reaction profile
was much cleaner when compared to the original catalyst
screening hit.
At this point in the project, evaluation of the downstream
chemistry demonstrated that rejection of methyl impurities
derived from compound 18 would become progressively more
difficult. It was therefore decided to perform an upgrade of 12.
A recrystallisation screen demonstrated that alcoholic solvents,
in particular 2-propanol, gave the best combination of impurity
rejection and recovery. On scale a 92% recovery of 12 was
achieved with a 50% reduction in the level of the methyl
impurity from 2.8 to 1.4%. Evaluation of the upgraded acetal 12
in the downstream chemistry confirmed the levels of impurities
derived from 18 would conform to the desired API
specifications.
The focus then turned towards the generation of the
hydrazone 13. This could be accomplished by a palladium-
catalysed cross-coupling of the upgraded aryl bromide 12 with
an aryl hydrazone such as benzophenone hydrazone 19.8
Compound 19 is a cheap, readily available commercial material
that is significantly easier and safer to handle on scale than the
hydrazine hydrate which had been previously employed in the
Medicinal Chemistry synthesis. In-house catalyst screening
identified that the use of strong bases such as sodium tert-
butoxide (t-BuONa) promoted hydrazone addition to the
nitrile, giving rise to the amidine 20, and as such, their use was
precluded. Switching to milder bases such as Cs2CO3 or K3PO4
resulted in the formation of the desired product 13 when used
in combination with commonly used amination ligands (such
as Xphos, Davephos, Ruphos, and BINAP).9,10 However, these
highly active ligands were also found to promote the formation
of the undesired bis-addition product 21 through chloride
activation (Table 1, entries 1−3). Less active ligands were then
screened with Pd2(dba)3/dppf in a biphasic mixture of CPME
and saturated K3PO4 (Table 1, entry 7) in particular giving
good conversion whilst minimizing the formation of bis-adduct
21. Optimization of this initial hit highlighted that, although
A screen of alternative reaction solvents showed similar
conversion in CPME, toluene, and THF. Toluene was
ultimately chosen since robust crystallisations of 13 from
toluene/heptane or toluene/2-propanol had been identified.
Further optimisation of the reaction protocol showed that
thorough subsurface nitrogen degassing of the solution of 12
and 19 in toluene and the 2 M aqueous potassium phosphate
solution were required before combination and charging of the
catalyst. Effective mixing was also crucial to ensure constant
suspension of the base, thus maximising mass transfer and
achieving >95% conversion. Control of residual metals (Pd and
Fe) was also important at this stage to ensure compliance with
the specification at the final drug candidate. To this end a 60 wt %
MP-TMT (macroporous polystyrene-2,4,6-trimercaptotriazine)11
resin treatment of the organic phase was employed during workup,
prior to isolation by crystallisation from toluene/2-propanol,
allowing 13 to be isolated in 79% yield from 12.
Alkylation and Indazole Formation: Preparation of
Drug Substance 1. With the issue of regiocontrol during
N-alkylation now solved by the use of intermediate hydrazone
13, there still remained the possibility of Boc-deprotection to
give 22 followed by over-alkylation leading to 23. A review of
the Medicinal Chemistry alkylation conditions highlighted the
use of DBU or Cs2CO3 in DMF or DMSO. An extensive base
screen was carried out; selected results are presented in Table
2. Alkoxide bases in DMF or NMP gave reasonable conversion
but with high levels of 22 and 23 seen. A trend was noted in
that the lithium base performed better than the sodium or
potassium base. Switching to hexamethyl disilazane (HMDS)
bases, the same trend was apparent in DMF or NMP. When the
solvent was changed to THF, both KHMDS and NaHMDS
performed poorly (Table 2, entries 9 and 10). However,
the combination of LiHMDS with THF gave a very clean
reaction (Table 2, entry 11) with very low levels of over-alkylation.
608
dx.doi.org/10.1021/op200334x | Org. Process Res. Dev. 2012, 16, 605−611

Organic Process Research & Development
Article
Table 2. Screening results for alkylation of 13 with 8
a
a
a
entry
base
solvent
product 14 (area %)
de-Boc product 22 (area %)
over-alkylated 23 (area %)
1
2
KOtBu
KOtBu
DMF
THF
DMF
NMP
DMF
NMP
DMF
NMP
THF
THF
THF
NMP
DMAc
71.4
55.8
80.1
79.4
92.4
78.5
83.7
89.6
36.5
58.1
96.2
97.0
88.0
10.9
5.4
5.8
0
10.3
10.6
7.6
3
NaOtBu
NaOtBu
LiOtBu
4
9.8
5
0.6
0
5.7
6
KHMDS
NaHMDS
LiHMDS
KHMDS
NaHMDS
LiHMDS
Cs2CO3
K3PO4
12.3
10.3
5.4
7
5.1
0
8
9
26.0
6.0
3.4
0
9.6
10
11
12
13
3.6
0.5
3.0
1.0
1.0
a
Determined by HPLC analysis.
This combination was therefore chosen to go forward for further
development.
Scheme 6. Preparation of desired compound 1
Following a DOE study of reaction stoichiometry, concen-
tration, and temperature, optimised conditions were imple-
mented, whereby deprotonation of a THF solution of 13 using
1.15 equiv of LiHMDS at −5 °C was followed by addition of
1.15 equiv of 8 at −5 °C. Complete conversion to 14 was
obtained after 1 h with none of the over-alkylation product
23 seen.
An early incarnation of the acid-catalysed hydrolysis/
cyclisation sequence to afford drug candidate 1 was performed
using PTSA monohydrate in ethanol at elevated temperature.
However, concerns over the generation of tosylate esters as
potential genotoxic impurities meant that a reagent change was
required.12 One set of Medicinal Chemistry conditions involved
the use of concentrated HCl in DMF to effect this
transformation and this was applied to the current reaction
by carefully charging 2 M hydrochloric acid to the completed
alkylation reaction mixture in THF and then heating to 65 °C
(Scheme 6). Complete conversion was achieved after 4 h at this
temperature. After extractive workup, 1 could be crystallised
from ethyl acetate in 75% yield by adding heptane antisolvent.
This however produced a waxy solid of low purity (∼70 wt %)
due to cocrystallisation of benzophenone. Also noted was an
increased propensity of the product to oil out as scale of
operation was increased. These factors meant that a new robust
crystallisation was required. A screen identified methanol as an
ideal solvent for the crystallization of 1 which could overcome
these problems. Benzophenone was found to have relatively
high solubility (>200 mg/mL), whilst the desired product was
found to have low solubility (0.3 mg/mL) in this system. Thus,
after extraction of the product into ethyl acetate and solvent
switching into methanol, 1 could be reproducibly crystallised as
a pale-yellow solid in 79% yield.
Review of the analytical data for the crude API highlighted a
discrepancy between the 90 wt % assay and the 98 area %
profile (both by HPLC). Correcting the wt % assay for solvents
and inorganics only accounted for approximately 2% of the
difference. It was known that the bromide 8 had less than ideal
stability under the reaction conditions, and it was theorised that
8 could polymerise via a de-Boc/alkylation sequence. In order
609
dx.doi.org/10.1021/op200334x | Org. Process Res. Dev. 2012, 16, 605−611

Organic Process Research & Development
Article
to remove the polymeric material thus produced and achieve a
pharmaceutically acceptable purity specification, a further
upgrade was required. An extensive screen of recrystallisations,
carbons, adsorbents, and resin treatments was performed. It was
found that by dissolution in ethyl acetate, filtration through a
silica pad, and crystallisation with heptane, pure 1 could be
obtained in 92% recovery (98 wt %, 99 area % purity).
7.37 (1 H, s), 7.61 (1 H, d, J = 8.4 Hz), 7.67 (1 H, d, J = 8.4
Hz), 10.23 (1 H, s); 13C NMR (100.6 MHz, CDCl3): δ 188.5,
158.2, 149.5, 136.5, 135.8, 132.8, 129.3, 129.2, 126.4, 124.5,
120.7, 116.9, 116.8, 114.6.
3-(3-Bromo-6-chloro-2-dimethoxymethylphenoxy)-5-
chlorobenzonitrile benzhydrylidene-hydrazine (12). A mix-
ture of biaryl ether 17 (7.8 kg, 21.0 mol), PTSA (1.6 kg,
8.4 mol), trimethylorthoformate (6.7 kg, 63.0 mol), and
methanol (30.8 kg) was heated to 65 °C for 30 min. The
reaction was confirmed complete by HPLC and cooled to 25 °C.
Triethylamine (1.1 kg, 10.5 mol) was charged and the mixture
further cooled to 5 °C. Water (39.0 kg) was charged over 1 h,
maintaining the temperature below 25 °C. The resulting slurry
was aged for 1 h and filtered, and the wet cake was washed with
water (20.0 kg). After drying in vacuo at 50 °C for 12 h, 7.66 kg
of acetal 12 was obtained as an off-white solid (92% yield,
99 wt % purity).
CONCLUSION
■
A new and practical synthesis of 1 has been developed and
demonstrated on multikilogram scale. Key to the success of this
route was the formation of the indazole core via a regioselective
N-alkylation of a benzophenone derivative. This strategy
enabled an improvement in the overall yield from 5 to 35%
over six linear steps. In addition, control of impurities generated
by the supply of an impure starting material has been
established. The chemistry employed may form the basis of a
future manufacturing route; however, further work would be
required to resolve the formation of polymeric impurities in the
final stage and thus remove the requirement for a silica
treatment.
Upgrade of 12. Acetal 12 (7.6 kg) was charged to a vessel
followed by 2-propanol (60.0 kg). The slurry was heated to
82 °C to dissolve the solids, the batch was cooled to 5 °C, and
the resulting slurry aged for 1 h. The batch was filtered and the
cake washed with 2-propanol (20.0 kg). The solid was dried in
vacuo at 50 °C with a nitrogen sweep for 12 h to afford 6.96 kg
1
EXPERIMENTAL SECTION
of 12 as an off-white solid (92% yield). Mp 108−110 °C; H
■
NMR (400 MHz, CDCl3): δ 3.36 (6 H, s), 5.62 (1 H, s), 6.93
(1 H, m), 7.07 (1 H, m), 7.31 (1 H, m), 7.37 (1 H, d, J = 8.4
Hz), 7.54 (1 H, d, J = 8.4 Hz); 13C NMR (100.6 MHz, CDCl3):
δ 158.8, 148.6, 135.9, 133.3, 131,9, 131.8, 128.7, 125.5, 122.0,
120.9, 117.4, 117.2, 114.0, 105.3, 56.0.
General. Starting materials were obtained from commercial
suppliers and were used without further purification. HPLC
analyses were performed on an Agilent Series 1100 liquid
chromatograph equipped with a UV detector (wt % and area %
purity). NMR spectra were obtained at 400 MHz for 1H
and 100 MHz for 13C. All coupling constants are reported in
hertz (Hz).
3-[3-(N′-Benzhydrylidene-hydrazino)-6-chloro-2-di-
methoxymethylphenoxy]-5-chlorobenzonitrile (13). To a 400 L
vessel was charged acetal 12 (6.96 kg, 16.7 mol),
benzophenone hydrazone 19 (3.43 kg, 17.5 mol), and toluene
(33.0 kg). The resulting solution was degassed with subsurface
nitrogen for 1 h. Potassium phosphate tribasic (14.2 kg) was
dissolved in water (32.0 kg) and the resulting solution degassed
with subsurface nitrogen for 1 h. PdCl2(dppf)·CH2Cl2 (272 g)
was charged to the toluene solution of 12 and 19 and the mixture
degassed with subsurface nitrogen for a further 30 min. The
potassium phosphate solution was then charged to the 400 L
vessel and the vigorously agitated biphasic mixture heated to
reflux for 19 h. The reaction was confirmed complete by HPLC
analysis. The mixture was cooled to 20 °C, diluted with toluene
(33 kg), and stirred for 10 min. The phases were separated, and
the toluene solution was washed with water (25 kg) and
saturated aqueous brine (25 L). MP-TMT resin (5.0 kg) was
charged and the mixture agitated for 17 h at 20 °C. The resin
was removed by filtration and the cake washed with toluene
(20.0 kg). The toluene solution was concentrated by distillation
under reduced pressure at <45 °C to a volume of ∼20 L and
then diluted by addition of 2-propanol (76 kg) over 30 min.
The resulting slurry was aged at 20 °C for 2 h, cooled to 0 °C
for 1 h ,and then filtered. The cake was washed with 2-propanol
(20 kg) and then dried at 55 °C under vacuum with a nitrogen
sweep for 16 h to afford hydrazone 13 as a pale-yellow solid
(7.0 kg, 79% yield, 94.0 area% purity, 11 ppm Pd, 12 ppm Fe).
6-Bromo-3-chloro-2-fluorobenzaldehyde (16). LDA was
prepared by charging n-butyllithium (23% in hexane, 10.7 kg,
38.4 mol) to a mixture of diisopropylamine (4.0 kg, 40.1 mol)
and THF (35 L) at 0−10 °C in a 400 L vessel. The LDA
solution was then charged to a solution of 10 (7.0 kg, 33.4 mol)
in THF (35 L) at −60 °C over 1 h and aged at −60 °C for a
further hour. DMF (40 kg, 100.0 mol) was then added at a rate
sufficient to maintain the temperature at −50 to −55 °C and
aged for 1 h. HPLC analysis was used to confirm the reaction
was complete. The reaction was quenched by transferring into a
vigorously stirred mixture of water (45 kg), concentrated HCl
(33 kg), and MTBE (70 L), maintaining the temperature at
0−5 °C. The batch was warmed to 15 °C, and the phases were
separated. The aqueous was re-extracted with MTBE (35 L),
and the organics were combined and concentrated under
reduced pressure to a volume of 10−15 L. DMF (35 L) was
charged and the solution concentrated under reduced pressure to
remove the remaining MTBE. Aldehyde 16 was obtained as a
solution in DMF (54 kg, 7.50 kg of 16 by assay, 95% assay yield).
3-(3-Bromo-6-chloro-2-formylphenoxy)-5-chlorobenzoni-
trile (17). Phenol 11 (3.7 kg, 24.1 mol), K2CO3 (6.7 kg, 19.2 mol),
and the DMF solution of aldehyde 16 (54 kg, 13.9 wt %,
26.5 mol) were charged to a 160 L vessel and aged at 21 °C for
20 h. Further charges of K2CO3 (3.35 kg) were made at 2, 5,
and 8 h (total charge 16.8 kg, 48.2 mol). The reaction was
confirmed complete by HPLC analysis and then cooled to
5 °C. Water (29.6 kg) was charged at such a rate as to maintain
the temperature below 25 °C. The resulting slurry was aged for
1 h at 25 °C and filtered, and the wet cake was washed with
water (10.0 kg). The filter cake was dried in vacuo to afford
biaryl ether 17 (7.82 kg, 85% yield from 11). Mp 176−179 °C;
1H NMR (400 MHz, CDCl3): δ 6.93 (1 H, m), 7.09 (1 H, m),
1
Mp 160−164 °C; H NMR (400 MHz, CDCl3): δ 3.01 (6 H,
s), 5.29 (1 H, s), 6.98 (1 H, m), 7.07 (1 H, m), 7.36 (7 H, m),
7.54 (1 H, m), 7.62 (4 H, m), 7.80 (1 H, d, J = 9.0 Hz); 13C
NMR (100.6 MHz, CDCl3): δ 158.5, 147.1, 146,6, 144.2, 138.0,
136.3, 133.7, 131.7, 129.4, 129.1, 129.0, 128.4, 128.3, 126.7,
125.8, 120.6, 117.1, 117.0, 115.8, 114.6, 114.4, 112.2, 101.3,
54.2.
610
dx.doi.org/10.1021/op200334x | Org. Process Res. Dev. 2012, 16, 605−611

Organic Process Research & Development
Article
Poehnelt, R.; Prasad, S.; Sanchez, R.; Torrent, M.; Vacca, J. P.; Wan, B.
L.; Yan, Y. Bioorg. Med. Chem. Lett. 2008, 18, 2959−2966. (b) Tucker,
T. J.; Sisko, J. T.; Tynebor, R. M.; Williams, T. M.; Felock, P. J.; Flynn,
J. A.; Lai, M. T.; Liang, Y.; McGaughey, G.; Liu, M.; Miller, M.; Moyer,
G.; Munshi, V.; Perlow-Poehnelt, R.; Prasad, S.; Reid, J. C.; Sanchez,
R.; Torrent, M.; Vacca, J. P.; Wan, B. L.; Yan, Y. J. Med. Chem. 2008,
51, 6503−6511.
(4) Kuethe, J. T.; Zhong, Y. L.; Alam, M.; Alorati, A. D.; Beutner, G.
L.; Cai, D.; Fleitz, F. J.; Gibb, A. D.; Kassim, A.; Linn, K.; Mancheno,
D.; Marcune, B.; Pye, P. J.; Scott, J. P.; Tellers, D. M.; Xiang, B.;
Yasuda, N.; Yin, J.; Davies, I. W. Tetrahedron 2009, 65, 5013−5023.
(5) Gomez, R.; Jolly, S. J.; Williams, T.; Vacca, J. P.; Torrent, M.;
McGaughey, G.; Lai, M. T.; Felock, P.; Munshi, V.; DiStefano, D.;
Flynn, J.; Miller, M.; Yan, Y.; Reid, J.; Sanchez, R.; Liang, Y.; Paton, B.;
Wan, B. L.; Anthony, N. J. Med. Chem. 2011, 54, 7920−7933.
(6) Lokhande, P. D.; Raheem, A.; Sabale, S. T.; Chabukswar, A. R.;
Jagdale, S. C. Tettrahedron Lett. 2007, 48, 6890−6892.
(7) Not identified. No ionization by LC−MS or GC−MS. This was
not the methyl impurity related to 18.
3-Chloro-5-[5-chloro-1-(1H-pyrazolo[3,4-b]pyridin-3-yl-
methyl)-1H-indazol-4-yloxy]benzonitrile (1). Hydrazone 13
(7.0 kg, 13.1 mol) was dissolved in THF (50 kg) in a 400 L
vessel and cooled to −5 °C. LiHMDS (13.5 kg, 1.0 M solution
in THF, 15.2 mol) was charged over 15 min, maintaining the
temperature at −5 °C. The dark-red solution was aged at
−5 °C for 10 min, and then a solution of bromide 8 (4.72 kg,
15.2 mol) in THF (13 kg) was charged over 15−30 min
(maintaining the reaction temperature at −5 °C). The reaction
was warmed to 20 °C over ∼1 h. The reaction was confirmed
complete by HPLC analysis. Hydrochloric acid (2 M, 26.5 kg,
53 mol) was charged slowly to the vessel and the mixture
heated to 65 °C and aged for 4 h. The reaction was confirmed
complete by HPLC analysis. The phases were separated, and
the aqueous layer was extracted with ethyl acetate (27 kg). The
organic phases were combined, diluted with ethyl acetate
(54 kg), and then washed with 1 M NaOH (23 kg) followed by
two water washes (39 kg each). The solution was concentrated
to 30 L at reduced pressure, and then methanol was (100 kg)
charged. The solution was concentrated to 30 L under reduced
pressure. The resulting slurry was cooled to −5 °C and aged for
2 h. The product was filtered and the cake washed with
methanol (15 kg). After drying in vacuo at 50 °C for 16 h,
crude 1 was obtained as a yellow solid (4.5 kg, 90 wt % purity,
70% yield corrected for assay).
(8) Szmant, H. H.; McGinnis, C. J. Am. Chem. Soc. 1950, 72, 2890−
2892.
(9) Surry, D. S.; Buchwald, S. L. Chem. Sci. 2011, 2, 27−50.
(10) Hartwig, J. F. Synlett 1997, 4, 329−340.
(11) Purchased from Biotage Ltd.
(12) Robinson, D. I. Org. Process Res. Dev. 2010, 14, 946−959.
Purification of 1. To a 400 L vessel were charged 1 (4.5 kg)
and ethyl acetate (81 kg). The contents were heated to 60 °C
until all the solid had dissolved. The solution was cooled to
20 °C and filtered through a bed of silica gel (22.5 kg). The silica
was washed with ethyl acetate (203 kg) until all product had
been recovered. The solution was concentrated under reduced
pressure to a volume of 40 L (maintaining the batch temperature
below 40 °C). The resulting slurry was cooled to 20 °C and
heptane (92 kg) charged over 30 min. The slurry was cooled to
−5 °C and aged for 2 h, and the product was filtered. The wet
cake was washed with ethyl acetate/heptane (1:3, 20 L) followed
by heptane (20 L). After drying in vacuo at 50 °C for 16 h,
pure 1 was obtained as a pale-yellow solid (3.75 kg, 98 wt %
1
purity, 92% yield corrected for assay). Mp 190−192 °C; H
NMR (400 MHz, CDCl3): δ 6.01 (2 H, s), 6.95 (1 H, m), 7.16
(2 H, m), 7.35 (1 H, t), 7.41 (1 H, d, J = 9.0 Hz), 7.51 (1 H, d,
J = 9.0 Hz), 7.80 (1 H, s), 8.03 (1 H, d, J = 9.0 Hz), 8.63 (1H, d,
J = 9.0 Hz), 12.88 (1 H, s); 13C NMR (100.6 MHz, CDCl3): δ
158.1, 152.7, 149.3, 141.7, 140.7, 140.3, 136.5, 130.3, 130.2,
129.1, 126.3, 121.3, 119.3, 118.6, 117.6, 117.4, 116.9, 114.6,
113.8, 108.7, 47.6; HRMS (ESI+) calcd for C21H12Cl2N6O
434.0450, found 434.0528.
AUTHOR INFORMATION
Corresponding Author
adrian_goodyear@merck.com
■
ACKNOWLEDGMENTS
We thank Simon Hamilton, Thorsten Platz, and Sophie
Strickfuss for analytical support throughout this work.
■
REFERENCES
■
(1) Tarby, C. M. Curr. Top. Med. Chem. 2004, 4, 1045−1057.
(2) http://www.tibotec.com/bgdisplay.jhtml?itemname1/4HIV_
tmc125.
(3) (a) Tucker, T. J.; Saggar, S.; Sisko, J. T.; Tynebor, R. M.;
Williams, T. M.; Felock, P. J.; Flynn, J. A.; Lai, M. T.; Liang, Y.;
McGaughey, G.; Liu, M.; Miller, M.; Moyer, G.; Munshi, V.; Perlow-
611
dx.doi.org/10.1021/op200334x | Org. Process Res. Dev. 2012, 16, 605−611