Process development and large-scale synthesis of a c-Met kinase inhibitor

Article abstract of DOI:10.1021/op100101q

A highly convergent synthesis of c-Met kinase inhibitor 1 has been demonstrated on a multikilogram scale using three key fragments: dihalotricyclic core 2, chiral sulfamide side chain 3, and pyrazole boronic ester 4. The chirality in sulfamide side chain 3 was installed using the cheap and readily available starting material (S)-epichlorohydrin. A total of 2.71 kg of 1 were isolated in seven steps (the longest linear sequence).

Full text of DOI:10.1021/op100101q

Organic Process Research & Development 2010, 14, 849–858
Process Development and Large-Scale Synthesis of a c-Met Kinase Inhibitor
Gavin W. Stewart,*,† Karel M. J. Brands,† Sarah E. Brewer,† Cameron J. Cowden,† Antony J. Davies,† John S. Edwards,†
Andrew W. Gibson,† Simon E. Hamilton,† Jason D. Katz,‡ Stephen P. Keen,† Peter R. Mullens,† Jeremy P. Scott,†
Debra J. Wallace,† and Christopher S. Wise†
Department of Process Research, Merck Sharp and Dohme Research Laboratories, Hertford Road, Hoddesdon,
Hertfordshire, EN11 9BU, U.K., and Department of Chemistry, Merck and Co., Inc., 33 AVenue Louis Pasteur,
Boston, Massachusetts 02115, U.S.A.
Abstract:
(Scheme 1).7 The challenges of the route for large scale
implementation were the high temperature polyphosphoric acid-
mediated cyclisation to afford the tricyclic core and the synthesis
of the chiral sulfamide side chain. The synthesis of the sulfamide
side-chain 3 (Scheme 2) involved a chiral stationary phase
HPLC separation of the racemic Cbz-protected analogue of
amine 5, which was available in sufficient quantities within our
department for the Medicinal Chemistry delivery. However, the
supplies of 5 were insufficient for a bulk delivery and, in
addition, the 3-step protection-resolution-deprotection sequence
was undesirable. Therefore, an alternative route to sulfamide 3
was required. It was thought that the correct enantiomer of
amine 5 could be obtained from (S)-epichlorohydrin, a cheap
and readily available starting material (Scheme 3).
A highly convergent synthesis of c-Met kinase inhibitor 1 has been
demonstrated on a multikilogram scale using three key fragments:
dihalotricyclic core 2, chiral sulfamide side chain 3, and pyrazole
boronic ester 4. The chirality in sulfamide side chain 3 was installed
using the cheap and readily available starting material (S)-
epichlorohydrin. A total of 2.71 kg of 1 were isolated in seven steps
(the longest linear sequence).
Introduction
The receptor for hepatocyte growth factor (HGF), which is
also known as scatter factor (SF), is c-Met (Mesenchymal
epithelial transition factor). Activation of c-Met by HGF can
induce a variety of cellular responses, including proliferation,
survival, motility, invasion, and changes in morphology.1-3
Aberrant activation of c-Met can increase the tumorigenicity
and metastatic potential of tumor cells.4,5 Such activation can
occur through several mechanisms, including overexpression
of c-Met and increased expression of HGF. Such overexpression
has been reported in cancers of many organs and shown to
increase with tumor progression for several,3 correlating with
shorter patient survival. It is therefore hypothesized that
inhibitors of c-Met could suppress tumor aggressiveness and
increase survival. Compound 1 has undergone clinical inves-
tigation as an orally dosed c-Met kinase inhibitor at Merck.6
We report herein an efficient and practical synthesis of 1 that
is amenable to multikilogram operation.
Results and Discussion
The conversion of (S)-epichlorohydrin to alcohol 6 Via a
low yielding three-step procedure has been reported in the
literature8 and this would provide a good starting point for
further development. Reaction of epichlorohydrin with 2-chlo-
roethanol (Scheme 3) in the presence of BF3 ·OEt2 at 40 °C
gave complete conversion to intermediate alcohol 7, which was
then re-epoxidised to 8 using 8 M aqueous sodium hydroxide
at ambient temperature (67% yield over two steps). However,
significant levels of dimer 9 were present. A screen of alternative
Lewis acids in various solvents was performed but conversions
were poor.
Upon further investigation, it was discovered that inverse
addition of the epichlorohydrin to three equivalents of 2-chlo-
roethanol was optimal, and that formation of dimer 9 could be
minimized by performing the reaction with BF3 ·OEt2 in toluene.
The medicinal chemistry synthesis of 1 was convergent and
proceeded using three key fragments: dihalotricyclic core 2,
chiral sulfamide side chain 3, and the pyrazole boronic ester 4
* To whom correspondence should be addressed. E-mail: gavin_stewart@
merck.com.
† Merck Sharp and Dohme Research Laboratories.
‡ Merck and Co., Inc.
(1) Haddad, R.; Lipson, K. E.; Webb, C. P. Anticancer Res. 2001, 21,
4243–4252.
(2) Maulik, G.; Shrikhande, A.; Kijima, T.; Ma, P. C.; Morrison, P. T.;
Salgi, R. Cytokine Growth Factor ReV. 2002, 13, 41–59.
(3) Longati, P.; Comoglio, P. M.; Bardelli, A. Curr. Drug Targets 2001,
2, 41–55.
Because of the volatile nature of epoxide 8, a concentrated
toluene solution of this compound was used for the next step.
The assay yield for epoxide 8 over 2 steps was 74%.
Laboratory scale reactions of epoxide 8 with aqueous sodium
hydroxide at 90 °C led to ring-opening and subsequent
cyclization to the desired alcohol 6. To complete the synthesis
(4) Jeffers, M.; Rong, S.; Vande Woude, G. F. Mol. Cell. Biol. 1996, 16,
1115–1125.
(5) Matsumoto, K.; Date, K; Shimura, H.; Nakamura, T. Jpn. J. Cancer
Res. 1996, 87, 702–710.
(6) Pan, B.-S.; Chan, G. K.Y.; Chenard, M.; Chi, A.; Davis, L. J.;
Deshmukh, S. V.; Gibbs, J. B.; Gil, S.; Hang, G.; Hatch, H.; Jewell,
J. P.; Kariv, I.; Katz, J. D.; Kunii, K.; Lu, W.; Lutterbach, B. A.;
Paweletz, C. P.; Qu, X.; Reilly, J. F.; Szewczak, A. A.; Zeng, Q.;
Kohl, N. E.; Dinsmore, C. J. Cancer Res. 2010, 70 (4), 1524–1532.
(7) Katz, J. D. et. al.,publication pending.
(8) Kim, H. Y.; Patkar, C; Warrier, R.; Kuhn, R. Bioorg. Med. Chem.
Lett. 2005, 15, 3207–3211.
10.1021/op100101q  2010 American Chemical Society
Published on Web 06/02/2010
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
849

Scheme 1. Medicinal Chemistry Synthesis of 1
Scheme 2. Medicinal Chemistry Side Chain Synthesis
Scheme 3. New Route to Amine
of this fragment, activation of alcohol 6 with a suitable leaving
group and subsequent displacement with methylamine was
required. Initially, the mesylate (10) was investigated, however,
on treatment with methylamine in methanol at 80 °C (sealed
tube), a mixture of desired amine 5 and tertiary amine 11 was
formed and further optimization did not lead to significant
improvements. The corresponding tosylate (12) was found to
be a crystalline solid that could be cleanly converted into desired
amine 5 upon reaction with methylamine under the conditions
described above. Subsequent development efforts therefore
focused on tosylate 12. Due to the fact that alcohol 6 is highly
water-soluble, a through process from epoxide 8 to tosylate 12
was developed. The equivalents and concentration of sodium
hydroxide used were optimized to reduce formation of ether
13 and maximize conversion of alcohol 6 to tosylate 12,
respectively. After workup and a solvent switch to toluene,
850
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

Scheme 4. Preparation of Sulfamide 3
tosylate 12 could be crystallized by simply adding heptane at
low temperature (64% corrected yield over 2 steps).
which was known to be crystalline (unlike the TFA salt, which
was an oil). Thus, Cbz-sulfamide 16 was prepared from 17 by
repeating the conditions described above to prepare 15, except
After examining various combinations of reaction solvent,
concentration, and temperature for the reaction of tosylate 12
with methylamine, it was found that 4 M methylamine in
ethanol at 60 °C gave the highest yield of amine 5 with minimal
formation of tertiary amine 11 (<3 GCAP [gas chromatography
area percent]). Once the reaction was complete, the p-toluene-
sulfonic acid, which is generated in this reaction, could be
simply removed by adding sodium ethoxide and filtering off
the sodium tosylate formed (using MTBE as an antisolvent).
The resulting solution of amine 5 was then solvent-switched to
propan-2-ol (which also served to remove excess methylamine)
and desired product 5 was then isolated as its HCl salt in 68%
corrected yield.
Conversion of Amine 5 to Sulfamide 3. The most direct
method of converting amine 5 to sulfamide 3 (Scheme 4) would
be to displace ammonia from sulfamide (H2NSO2NH2). A
number of examples of this reaction using secondary dialky-
lamines have been reported in the literature,9 typically with a
large excess of sulfamide in polar solvents at elevated temper-
atures. Initial attempts at reacting amine 5 with 5 equiv of
sulfamide in either diglyme or dioxane (at 110 °C) looked
promising; however, the lack of a good analytical method and
pressing deadlines meant that this approach was put on hold.
As an alternative strategy, it was decided to modify the
Medicinal Chemistry procedure. Thus, the HCl salt of amine 5
was free-based in DCM by addition of 3.0 equiv of Hu¨nig’s
base (Et3N or Cy2NMe gave insoluble HCl salts, which would
have required an additional filtration). The resulting solution
of 5 was then added to a solution of (chlorosulfonyl)carbamate
14 in DCM (prepared by treatment of tBuOH in DCM at ∼0
°C with chlorosulfonylisocyanate) to afford Boc-protected
sulfamide 15, which was isolated by crystallization in 93% yield.
The Boc group was then removed, using the Medicinal
Chemistry conditions (22 equiv of TFA in DCM), to afford
sulfamide 3-TFA as a stable mono-TFA salt. Prolonged
attempts to flush out the residual TFA by either regular or
azeotropic distillation were unsuccessful.
t
BnOH was substituted for BuOH. Sulfamide 16 was then
isolated in 83% yield by crystallization from IPAc. However,
this crystallization gave very fine hair-like needles, which
resulted in a gelatinous mixture that was very slow to filter.
The same morphology was observed from toluene/heptane. A
through-process from amine salt 5-HCl to sulfamide 3 was
therefore pursued. Hydrogenolysis of Cbz-sulfamide 16 to
sulfamide 3 was found to proceed smoothly, using as little as
1 mol % of 10% Pd/C (50 wt % water) at 1 atm hydrogen in
methanol, to afford an essentially quantitative yield of 3 after
6 h. Sulfamide 3 was readily isolated by crystallization from
IPA, and recovery could be maximized when heptane was used
as an antisolvent. On scale, 3 was isolated in 89% yield over 2
steps. In summary, 3.13 kg of 3 were isolated over seven steps
in 29% overall yield
Preparation of the Tricyclic Core 2. The tricyclic core 2
was prepared using a modified version of the Medicinal
Chemistry procedure. In the first step, nicotinate 18 was
synthesized from vinamidinium salt 19 (Scheme 5) using
chemistry previously reported by Merck Process Research.10
Thus, vinamidinium salt 19 (1.2-1.5 equiv) was reacted with
the potassium enolate of methyl acetoacetate (formed using 1.0
equiv of KOtBu) in THF at 45 °C, in the presence of 1.0 equiv
of DABCO. Ammonium acetate (2.4 equiv) was then added to
the resulting reaction mixture, and it was aged overnight at g60
°C. In the Medicinal Chemistry procedure, crude nicotinate 18
(∼65% assay yield in our hands) was then isolated as a dark
brown oil via an extractive workup and purified by column
chromatography. With the aim of running this chemistry in the
large scale preparative laboratory on 50 kg of vinamidinium
salt 19, the need to remove this silica treatment was imperative.
It was quickly discovered that the nicotinate 18 was soluble in
heptane, or at least heptane containing low levels of ethyl
acetate, and this provided the basis for a more practical means
of purifying the crude product of this reaction. Initially,
nicotinate 18 was obtained by simply extracting the concentrated
dark-brown oil with heptane; however, it was subsequently
shown that it could also be extracted into heptane directly from
the crude final reaction mixture. Nicotinate 18 (13.2 kg; 96.4
LCWP [liquid chromatography weight percent]; 63% corrected
yield) was then isolated by solvent-switching to methanol and
adding water to crystallize.
It was anticipated that using a Cbz group in place of the
Boc group for this chemistry would provide two major benefits:
Cbz-protected sulfamide 16 would have a chromophore,
facilitating HPLC analysis and quantification, and the depro-
tection, via hydrogenolysis, would afford the free sulfamide (3),
Nicotinate 18 was then condensed with 4-bromobenzalde-
hyde upon dissolving in THF and adding 2 equiv of KOtBu.
Following the Medicinal Chemistry procedure, Friedel-Crafts
substrate 20 could then be isolated in 76-81% yield (uncor-
rected) by crystallization from 3 M HCl-EtOH. However, using
(9) For example Alker, D.; Campbell, S. F.; Cross, P. E.; Burges, R. A.;
(10) Marcoux, J.-F.; Marcotte, F.-A.; Wu, J.; Dormer, P. G.; Davies, I. W.;
Carter, A. J.; Gardiner, D. G. J. Med. Chem. 1990, 33, 585–591.
Hughes, D.; Reider, P. J. J. Org. Chem. 2001, 66, 4194–4199.
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
851

Scheme 5. Preparation of Tricyclic Core 2
this material (or material provided by the Medicinal Chemistry
group) in the subsequent Friedel-Crafts cyclization led to
significant foaming upon mixing with hot polyphosphoric acid
(PPA). The same problem was also observed when substrate
20 and PPA were mixed cold and then heated to reaction
temperature. This was eventually attributed to the large amounts
of KCl and HCl present in the isolated Friedel-Crafts substrate
(>40 000 ppm K; >86 000 ppm Cl). To avoid this problem,
the isolation of Friedel-Crafts substrate 20 was changed to a
combination of several aqueous washes followed by a solvent
switch to methanol, from which the desired product crystallized
in 67% corrected yield.
The Medicinal Chemistry group carried out the Friedel-Crafts
cyclization of acid 20 to tricyclic ketone 2 by heating a solution
of 20 in PPA to 200-220 °C overnight (63% yield). Such high
temperatures would restrict operations to 20 L glassware, with
a heating mantle, and hence running the reaction at lower
temperatures was initially investigated. However, repeating this
reaction at a range of temperatures confirmed that formation
of 2 was only observed at temperatures of g180 °C, and 200
°C was required for a reasonable reaction ratesat 215 °C,
complete conversion was observed after ∼4 h. Interestingly,
under acidic conditions, a significant amount of what is believed
to be lactone 21 is rapidly formed. It appears, however, that
this compound is in equilibrium with acid 20 as both are
consumed as the Friedel-Crafts reaction progresses and the
ratio of the two remains unchanged. Other conditions (Eaton’s
Reagent; MeSO3H; H2SO4; TfOH; Tf2O; P2O5; TFA-TfOH;
PhSH-MeSO3H) were screened for this Friedel-Crafts reaction
without success. Similarly, attempts to carry-out the cyclization
on the analogous acyl chloride (prepared from acid 20 and
SOCl2) were unsuccessful using a variety of conditions (TfOH
in DCE; MeSO3H; AlCl3 in nitrobenzene; TiCl4 in nitrobenzene;
AgOTf in nitrobenzene).
and then combine these for a single workup. Thus, after
quenching the reaction with water, the resulting slurry could
be readily transferred out of the reaction vessel at ambient
temperature. Diluting this slurry with acetonitrile and then
filtering rejected essentially all of the UV active impurities
(increased LCAP [liquid chromatography area percent] from
∼70% to >95%) and afforded desired product 2 as a crude dark-
brown solid. Although the LCAP purity of this material was
very high, the LCWP purity was only 60-70%, and it is
believed that 2 is isolated here as a phosphate salt. Therefore,
triethylamine was added to a suspension of this crude solid in
DCM and the resulting solution treated with 20 wt % EcoSorb
C-941. After removal of the EcoSorb, a solvent-switch to
acetonitrile resulted in tricyclic ketone 2 (97 LCWP; >99 LCAP)
crystallizing from solution in up to 76% yield.
The Friedel-Crafts cyclization was run in six batches (1.2
kg of 93 wt % 20 per batch) at ∼210 °C. Reactions were
complete after 4 h (average assay yield was 80%), and after
quenching with water, the six batches were combined. Work-
up was carried out as described above to afford 5.4 kg of a
black solid (70% yield of 2 corrected for purity). During this
workup, 4% of tricyclic ketone 2 was lost to the initial isolation,
and 5% was lost to the second. After a carbon treatment to
remove triethylamine salts, 2 was isolated in 67% yield from
acid 20 (corrected for purities). In summary, 4.28 kg of 2 were
isolated over 3 steps in 28% overall yield.
Coupling of the Tricyclic Core with the Sulfamide Side
Chain. The Medicinal Chemistry coupling of tricyclic ketone
2 with the TFA salt of sulfamide 3 to give penultimate 22
(Scheme 6) was carried out using 5 mol % Pd2(dba)3, 15 mol
% Xantphos, and 4.0 equiv of Cs2CO3 in dioxane (40 mL/g).
Dioxane was replaced with THF or toluene, and various
permutations of catalyst and base were then screened for this
reaction (Table 1), thus confirming that Pd2(dba)3, Xantphos,
and Cs2CO3 was the optimum catalyst-base combination.
Xantphos consistently gave higher yields than BINAP, and
Pd2(dba)3 was found to be superior to Pd(OAc)2 (entries 1-8).
The reaction failed when using Pd(dppf)Cl2.CH2Cl2 as a catalyst
(entries 9-10). Cs2CO3 gave higher assay yields than the other
bases tested (K2CO3, K3PO4, iPr2NEt).
Since reactions run in toluene became thick slurries, requiring
higher dilution than those in THF and necessitating multiple
batches, THF was chosen for scale-up. At the end of the
On the basis of these results, it was decided to run multiple
batches of the Friedel-Crafts reaction in PPA at 210-220 °C
852
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

Scheme 6. Pd-Mediated Coupling of Bromide 2 and Sulfamide 3
Table 1. Screen of Pd Catalyst/Ligand/Base/Solvent
entry
Pd source
ligand
base
solvent
assay yield of 22a
1
2
Pd(OAc)2
Xantphos
Xantphos
BINAP
BINAP
Xantphos
Xantphos
BINAP
BINAP
-
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
K2CO3
K3PO4
Tol.
THF
Tol.
THF
Tol.
THF
Tol.
THF
Tol.
THF
Tol.
Tol.
61%
0%
Pd(OAc)2
3
Pd(OAc)2
0%
4
Pd(OAc)2
11%
96%
87%
34%
24%
0%
5
Pd2(dba)3
6
Pd2(dba)3
7
Pd2(dba)3
8
Pd2(dba)3
9
Pd(dppf)Cl2.CH2Cl2
Pd(dppf)Cl2.CH2Cl2
Pd2(dba)3
10
11
12
13
14
15
16
17
18
-
0%
Xantphos
Xantphos
Xantphos
Xantphos
Xantphos
Xantphos
Xantphos
Xantphos
64%
64%
40%
1%
Pd2(dba)3
Pd2(dba)3
K3PO4
Tol. + 0.5 eq. H2O
Pd2(dba)3
DIPEA
K2CO3
K3PO4
K3PO4
DIPEA
Tol.
Pd2(dba)3
THF
76%
49%
57%
0%
Pd2(dba)3
THF
Pd2(dba)3
THF + 0.5 eq. H2O
Pd2(dba)3
Tol.
a All reactions were run on 100 mg scale at 80 °C using 0.20 equiv of Pd, 0.30 equiv of ligand, 1.50 equiv of base, and 1.1-1.2 equiv of 3 in 20 mL/g solvent for 18 h.
Reactions were homogenised using DMAc prior to assaying.
Scheme 7. Preparation of Pyrazole Boronate 4
Scheme 8. Preparation of Bromopyrazole 23
the methylation was successfully demonstrated on 200 g scale,
for the bulk campaign it was decided to buy in commercially
available 1-methylpyrazole (24) and to perform only the
bromination step in-house. It was found that 1.5 equivalents of
bromine was required to achieve complete consumption of
1-methylpyrazole and, after a standard extractive workup and
solvent-switch, the desired product (23) could be obtained as a
concentrated solution in THF that was suitable for direct use
in the following step. This bromination was run on scale to
afford a 47 wt % THF solution of 23 (11.6 kg) in 90% assay
yield.
reaction, the resulting slurry was simply filtered to give a solid
which was ∼40 wt % desired product 22, with the remainder
being cesium salts. Interestingly, although the solubility of 22
in THF is 116 mg/mL, at a reaction volume of 18 mL/g only
3% remains in the liquors after this filtrationspresumably this
is due to 22 existing as its less soluble cesium salt under these
conditions. Although the product from this reaction had very
low wt% purity, it could be successfully used in the subsequent
Suzuki coupling. In addition, this isolation procedure removed
nearly all of the organic byproduct, including unreacted starting
materials and des-bromo tricyclic core. Coupled product 22 was
isolated in 69% corrected yield.
Synthesis of the Pyrazole Boronate. Another compound
which we were unable to source in sufficient amounts, and
within the necessary timelines, was boronate 4 (Scheme 7).
Thus, a process was required in order to support the preparation
of 5 kg of this Suzuki coupling reagent. The synthesis of 4
from 4-bromo-1-methyl-1H-pyrazole (23) has previously been
reported,11 and it was hoped that this chemistry could be used
as the basis for developing a viable process.
Conversion of bromopyrazole 23 to the corresponding lithio
species, using BuLi under the literature conditions, was found
to be poor. In particular, alkylation of the lithiated pyrazole with
1-bromobutane generated a significant byproduct during the
lithium-halogen exchange. Despite these problems, quenching
with trimethylborate allowed isolation of 25 as a complex
mixture of oligomers and hydrates in approximately 20% yield.
Subsequent treatment with an excess of pinacol in THF lead to
formation of the desired 4.
Related research from within our department12 had previ-
ously demonstrated the superiority of an in situ quench with
triiso-propylborate (B(OiPr)3) over a stepwise lithiation-quench
procedure using trimethylborate. Applying these conditions,
Initially, bromopyrazole 23 was prepared from pyrazole in
75% yield via a two-step N-methylation (MeI, KOH, water)
and bromination (Br2, DCM) sequence (Scheme 8). Although
(11) Ivachtchenko, A. V.; Kravchenko, D. V.; Zheludeva, V. I.; Pershin,
(12) Li, W.; Nelson, D. P.; Jenson, M. S.; Hoerrner, R. S.; Cai, D.; Larsen,
D. G. J. Heterocyclic Chem. 2004, 41, 931–939.
R. D.; Reider, P. J. J. Org. Chem. 2002, 67, 5394–5397.
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
853

Scheme 9. Suzuki Coupling of Chloride 22 and -ate Complex 28
namely, adding 1.2 equiv of BuLi to a mixture of 23 and 1.2
equiv of B(OiPr)3 in THF/toluene at -70 °C to bromopyrazole
23, resulted in a 90% conversion to boronate 26 (observed as
boronic acid 25 on reverse-phase HPLC). Increasing the BuLi
and tri-iso-propylborate charges, to 1.5 and 1.3 equiv respec-
tively, gave essentially complete conversion. The solvent
composition was also adjusted from 1:4 to 1:1 THF/toluene to
reduce the viscosity of the reaction mixture at low temperature.
Final Coupling. Using -ate complex 28 (∼ 1.70 equiv), the
Suzuki coupling with chloride 22 (Scheme 9) consistently ran
to completion within 1 h using 5 mol % Pd(PtBu3)2 in DMF at
100 °C. Furthermore, no additional base was required. Initially,
1 was crystallized directly from the reaction mixture by simply
adding hydrochloric acid. Not surprisingly, material isolated in
this way was found to have very high residual palladium levels
(up to 14 000 ppm). After screening various resins/carbons, it
was found that addition of 2 M aqueous sodium hydroxide to
the final reaction mixture followed by treating with EcoSorb
C-941 reduced residual palladium in isolated 1 considerably.
Residual palladium in the isolated product was reduced further
when it was found that only 0.5 mol % Pd(PtBu3)2 was actually
needed for the reaction to reach completion in 1 h at 100 °C.
This chemistry was run on scale (using 0.6 mol % catalyst) to
afford an 80% yield of 1 (2.81 kg) with the material containing
150 ppm residual palladium.
Carbon-treating the aqueous solution of 1 sodium salt, during
the final salt formation, was briefly examined as a means of
reducing the residual palladium content down to acceptable
levels (<50 ppm). However, in the time frame required, we were
unable to identify a practical procedure that reliably gave
sufficient palladium reduction as well as high recovery and,
therefore, it was decided to include a separate palladium removal
step. We immediately focused on use of the MP-TMT
(macroporous polystyrene-2,4,6-trimercaptotriazine) resin, which
had proven to be very successful for palladium reduction in
other projects within the department. Since the solubility of 1
is highest in polar aprotic solvents, resin treatments were
assessed using solvents such as DMF, DMSO, DMAc, and
NMP. It was found that an overnight age with 20 wt % MP-
TMT resin in any of these solvents (10 mL/g) gave an order of
magnitude reduction in the Pd level (from 330 ppm to ∼10-30
ppm), with the treatment in DMF proving the most effective.
After removal of the resin, 1 (>95% recovery) was crystallized
by slow addition of water. This process was run using 17 wt
% resin, to afford 2.71 kg of 1 (97% recovery) with only 7
ppm residual palladium. Residual Cs levels were also measured
at this point and found to be 9 ppm.
Despite these improvements in conversion, the isolation of
acid 25 still proved difficult. Forming boronate 4 directly, by
quenching the reaction with 2-propanol followed by pinacol,
was therefore investigated as a means of avoiding the prob-
lematic isolation of 25. Unfortunately, despite obtaining an
encouraging 75% assay yield of boronate 4 (from bromopyra-
zole 23) via this method, the isolation of this compound by an
extractive workup procedure proved equally unsuccessful. All
attempts were hampered by the high aqueous solubility of
boronate 4 and/or decomposition of 4 to boronic acid 25 and
1-methylpyrazole (24). However, a significant breakthrough was
realised when a reaction to form boronic acid 25 was quenched
with 5 equiv of water rather than the large excess usually
employed. In this case, a white solid crystallized from solution
which, after isolation by a simple filtration, was tentatively
assigned as trihydroxyborate complex 27. Quenching instead
with pinacol, and then warming to room temperature before
adding the water, resulted in -ate complex 28 crystallizing from
solution. A simple filtration provided 81.5 wt % 28 in up to
88% yield (on 100 g scale). The balance of the wt% was a
combination of water (10.4 wt %), LiOH (4.6 wt %), and
residual solvents (3.9 wt %). Whilst it was possible to convert
-ate complex 28 to boronate 4 (in ∼80% recovery), this material
was found to be much more active than 4 in Suzuki couplings
and was therefore used as is.13 Using this procedure, 28 was
isolated in 75% yield. In summary, 16.6 kg of 27 were isolated
over 3 steps in 71% overall yield.
Conclusion
A highly convergent synthesis of c-Met kinase inhibitor 1
has been demonstrated on scale. The key feature of the synthesis
was the installation of chirality in sulfamide side chain 3 using
the cheap and readily available starting material (S)-epichloro-
hydrin. A total of 2.71 kg of 1 were isolated in 29% yield over
seven steps (the longest linear sequence) with the average yield
per step being 85%.
(13) Mullens, P. R. Tetrahedron Lett. 2009, 50, 6783–6786.
854
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

Experimental Section
(1 H, dd, J ) 5.4, 10.6 Hz), 7.37 (2 H, d, J ) 8.1 Hz), 7.81
(2H, d, J ) 8.3 Hz); 13C NMR (100.6 MHz, CDCl3): δ 145.1,
132.6, 129.9, 127.9, 72.3, 68.8, 67.5, 66.3, 66.2, 21.6; HRMS
(ESI+) calcd for C12H17O5S 273.0797, found 273.0797. Anal.
Calcd for C12H17O5S: C, 52.93; H, 5.92. Found: C, 52.53; H,
5.83.
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. NMR spectra
were obtained at 400 MHz for 1H and 100.6 MHz for 13C. All
coupling constants are reported in hertz (Hz).
(2R)-1,4-Dioxan-2-yl-N-methylmethanaminium Chloride
(5). A solution of tosylate 12 (7.76 kg of 82 LCWP material,
23.4 mol), methylamine (46.9 kg of a 33 wt % solution in
ethanol, 499 mol), and ethanol (49.0 kg) was heated to 65 °C
overnight in a sealed standard Pfaudler glass-lined vessel fitted
with a 5 barg bursting disk. The resulting solution was
concentrated, by atmospheric distillation, to a volume of ∼15
L. Sodium ethoxide (8 kg of a 21 wt % solution in ethanol,
24.7 mol) and MTBE (34.5 kg) were then added at 50 °C in
the following manner: Half the NaOEt was added over 30 min,
then half the MTBE, followed by the rest of the NaOEt, and
finally the remaining MTBE. The resulting slurry was cooled
to 20 °C, aged for 30-60 min, and then filtered, washing the
wet-cake with MTBE (11.5 kg). The combined filtrate and
washes were solvent-switched to IPA (using 30.5 kg IPA), by
atmospheric distillation, to a final volume of ∼30 L. Concen-
trated hydrochloric acid (2.46 kg) was then added, keeping the
batch temperature <60 °C. IPA (91.4 kg) was added, and the
batch was concentrated, by atmospheric distillation, to a volume
of ∼30 L. It was then aged at 50 °C until a seedbed formed.
The resulting slurry was cooled to 20 °C and then filtered,
washing the wet-cake with 1:1 heptane/IPA (15.5 L). The filter-
cake was dried to afford HCl salt 5 (2.83 kg, 95 GCWP, 68%
corrected yield) as a white solid. Mp 218-219 °C; [R]D +28.9
(c ) 0.1, CH3OH); 1H NMR (400 MHz, DMSO-d6): δ 2.51 (3
H, s), 2.84 (1 H, dd, J ) 8.8, 13.1 Hz), 2.95 (1 H, dd, J ) 3.5,
12.9 Hz), 3.25 (1 H, dd, J ) 9.8, 11.3 Hz), 3.47 (1 H, m,),
3.59 (1 H, m,), 3.66 (1 H, d, J ) 11.7 Hz), 3.74 (1 H, dd, J )
2.7, 11.3 Hz), 3.78 (1 H, d, J ) 11.7 Hz), 3.89 (1 H, m,); 9.12
(2 H, s, br); 13C NMR (100.6 MHz, DMSO-d6): δ 71.0, 68.1,
66.1, 66.0, 48.6, 33.3; HRMS (ESI+) calcd for C6H14NO2
132.1025, found 132.1022. Anal. Calcd for C6H14ClNO2: C,
42.99; H, 8.42; N, 8.36. Found: C, 42.71; H, 8.44; N, 8.32.
Benzyl{[[(2R)-1,4-Dioxan-2-ylmethyl](methyl)amino]-
sulfonyl}carbamate (16). Dichloromethane (4.5 L/kg based
on HCl salt 5 [2.80 kg, 16.7 mol], 12.6 L, 16.7 kg,) was charged
to a vessel, and then cooled to -20 °C. Chlorosulfonylisocy-
anate (2.95 kg, 20.9 mol) was added in a single portion, and
benzyl alcohol was then added over approximately 20 min,
maintaining the internal temperature at <3 °C. The resulting
batch was aged for 15 min whilst recooling to -20 °C. In a
second vessel, N,N-diisopropylethylamine (8.63 kg, 66.8 mol)
was added over 5 min to a stirred mixture of HCl salt 5 in
dichloromethane (12 L/kg based on HCl salt 5, 33.5 L, 44.4
kg). Once the contents of this second vessel were homogenized,
they were added to original vessel over ∼45 min, such that the
internal temperature remained <3 °C. The second vessel and
addition line were rinsed through with dichloromethane (0.3
L/kg based on salt 5, 1.1 kg). The resulting reaction mixture
was then aged for 20 min at -5 to 0 °C. The reaction was
carefully quenched with 4 M hydrochloric acid (10.57 L;
prepared from 1.93 kg of 38% hydrochloric acid and 9.32 kg
(2R)-2-[(2-Chloroethoxy)Methyl]Oxirane (8). A solution
of 2-chloroethanol (10.9 L, 162.5 mol) and BF3 ·OEt2 (0.34 L,
2.7 mol) in toluene (20 L) was heated to an internal temperature
of 33 °C. The jacket temperature was decreased to 20 °C and
epichlorohydrin (4.2 L, 53.6 mol) was added dropwise at a rate
sufficient to maintain the reaction temperature at 36-38 °C.
The resulting mixture was then aged at 36 °C for 20 min. The
reaction was then cooled to 16 °C and aqueous NaOH (12.5 L
of 46/48% aqueous NaOH and 12.5 L of water) was added
over 1 h, with vigorous stirring, maintaining the reaction
temperature below 20 °C. The batch was aged for 1 h at ambient
temperature, and GC analysis was then used to confirm that
cyclization was complete. The remaining water (10 L) was then
added. The two layers were separated, combining the rag with
the toluene phase. The aqueous phase was extracted with toluene
(13 L), and the combined organic layers were washed with water
(10 L). The resulting organic layer was distilled to low volume,
monitoring the distillate for loss of product. The final crude
mass was 9.10 kg (60 GCWP ≡ 5.43 kg of epoxide 8; 74%
yield). The NMR data obtained were identical to those reported
in the literature.14
(2S)-1,4-Dioxan-2-ylmethyl 4-Methylbenzenesulfonate (12).
Water (15.7 kg) was charged to a 50 L vessel, followed by
46/48% aqueous NaOH (19.6 kg). The resulting solution was
heated to 90 °C, and epoxide 8 (9.10 kg of 60 GCWP solution
≡ 5.43 kg of epoxide, 39.8 mol) was then added. The reaction
mixture was aged at 90 °C for 1 h. The reaction mixture was
cooled to ambient temperature, and DCM (13.9 kg) was then
added followed by solid TsCl (3.73 kg, 19.2 mol), in a single
portion. The biphasic mixture was aged overnight at ambient
temperature. Water (10.5 kg) was added, and the lower DCM
layer was separated off. The aqueous phase was then extracted
twice with DCM (2 × 13.9 kg). The combined organic layers
were washed with 5% brine (0.52 kg NaCl in 10.5 kg water),
and then solvent-switched to toluene to a final concentration
of 2 L/kg (based on assay). A second batch was run on an
identical scale, and the two batches were combined before
proceeding. Heptane (7 L) was added, followed by 1 wt % of
seed. The resulting batch was then cooled to 0 °C, another 1
wt % seed was added, and it was left to age over the weekend
at ambient temperature. The batch was then recooled to 4 °C
and aged for 2 h. The resulting solid was collected by filtration,
washing the wet-cake with 4 L of 8:1 heptane/toluene. After
drying at 25 °C under vacuum overnight (with a nitrogen
sweep), tosylate 12 was obtained as a white solid (8.05 kg, 82.4
LCWP, 64% corrected yield, 88.6 LCAP). Mp 55-57 °C; [R]D
-13.0 (c ) 0.1, CHCl3); 99.3% ee by chiral HPLC; 1H NMR
(400 MHz, CDCl3) δ 2.47 (3 H, s), 3.37 (1 H, dd, J ) 9.6,
11.2 Hz), 3.70 (6 H, m), 3.97 (1 H, dd, J ) 4.8, 10.4 Hz), 4.03
(14) Halil, H.; Colak, N.; Oeztuerkmen, N. Collect. Czech. Chem. Commun.
1994, 59, 239–242.
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
855

of water), maintaining the internal temperature <3 °C (typically
takes 10-15 min). The two layers were separated, and the
organic phase was washed with water (18.6 kg). The resulting
organic layer was then solvent-switched to MeOH (using ∼165
L MeOH) to a final concentration of 20 L/kg (based on assay
of sulfamide 16).
stirred volume. Methanol (39.6 kg) was added, the batch was
concentrated to minimum stirred volume, as described above,
and a second portion of methanol (39.6 kg) was then added.
The batch was cooled to 20 °C, and water (50 kg) was slowly
added over a period of 30 min. Once this addition was complete,
the batch was seeded and then cooled to 0 °C. Once a slurry
had formed (45 min), a second portion of water (37.5 kg) was
added and the batch was aged at 0 °C for another 1 h. The
slurry was then filtered, washing the wet-cake with 5:9
methanol/water (14 L). The resulting filter-cake was dried under
vacuum at ambient temperature to afford nicotinate 18 (13.22
kg, 96.4 LCWP, 63% corrected yield) as a light brown solid.
Mp 45-47 °C. 1H NMR (400 MHz, DMSO-d6): δ 2.68 (3 H,
s), 3.87 (3 H, s), 8.21 (1 H, d, J ) 2.4 Hz), 8.69 (1 H, d, J )
2.4 Hz); 13C NMR (100.6 MHz, DMSO-d6): δ 165.5, 157.5,
150.6, 137.4, 128.8, 126.2, 52.9, 24.0; HRMS (ESI+) calcd
for C8H9NO2Cl 186.0322, found 186.0319. Anal. Calcd for
C8H9NO2Cl: C, 51.77; H, 4.34; N, 7.55. Found: C, 51.55; H,
4.25; N, 7.54.
N-[(2R)-1,4-Dioxan-2-ylmethyl]-N-methylsulfamide (3).
To the methanol solution of Cbz sulfamide 16 (20 L/kg in
MeOH, ∼5.17 kg (15.0 mol) of 16 by assay, ∼106 L) that
was prepared above was added the palladium on carbon
catalyst (801 g, 0.38 mol). The mixture was degassed and
then stirred under hydrogen (1 atm, 15 psi) until the reaction
was complete. The resulting reaction mixture was filtered,
washing the vessel and catalyst residue through with MeOH
(1-2 L/kg, approximately 5-10 L). The combined filtrate
and wash was then solvent-switched to isopropanol, at 35-40
°C under vacuum, to a final volume of 10 L/kg (based on
assay of product). The resulting mixture was aged, allowing
a seedbed to form. Heptane (15 L/kg, ∼47.3 L, 32.4 kg)
was then added, and the resulting slurry was filtered, washing
the wet-cake with 1:3 isopropanol/heptane (2.7 and 7.2 kg,
respectively). The filter-cake was dried under vacuum at 50
°C (18 h) to afford sulfamide 3 (3.13 kg, 95.2 LCWP, 89%
corrected yield from amine HCl salt 5) as a white solid. Mp
2-[(E)-2-(4-Bromophenyl)vinyl]-5-chloronicotinic Acid (20).
Nicotinate 18 (7.0 kg, 37.7 mol), 4-bromobenzaldehyde (7.0
kg, 37.7 mol), and THF (99.5 kg) were charged to a vessel.
The resulting stirred solution was put under a nitrogen atmo-
sphere, and then cooled to -5 °C. To a second vessel were
charged potassium tert-butoxide (8.46 kg, 75.4 mol) and THF
(47.3 kg). Once the KOtBu had dissolved, this solution was
added via a 1 µm in-line filter to the contents of original vessel,
whilst maintaining the internal temperature in this vessel at <0
°C. The second vessel and transfer lines were rinsed through
with THF (7.1 kg) into the batch. The resulting reaction mixture
was then warmed to 25 °C over a period of 1 h. After aging at
this temperature for 1.5 h, the reaction was quenched by addition
of water (53.8 kg), ensuring the batch temperature remained
<30 °C. The biphasic mixture was stirred for 10 min, and the
lower aqueous layer was then separated off. To the remaining
THF layer was added ethyl acetate (48.6 kg) and 2 M
hydrochloric acid (5.3 kg of conc. HCl diluted with 22.4 kg
water). The resulting biphasic mixture was stirred for 10 min,
and the lower aqueous layer was then separated off. The organic
layer was washed with half-saturated brine (2.0 kg NaCl in 19.5
kg water) and then concentrated, using partial vacuum at <40
°C, to ∼27 L. The distillation was then continued whilst slowly
adding ethyl acetate to maintain a constant batch volume of
∼27 L, until 48.5 kg of ethyl acetate had been added. This
procedure was then repeated with methanol (42.6 kg). Once
this was complete, the batch volume was adjusted to 56 L by
adding additional methanol (15.0 kg), and the batch temperature
was adjusted to 25 °C. The slurry was then filtered, washing
the wet-cake with methanol (8.5 kg). The resulting filter-cake
was dried under vacuum at 50 °C to afford acid 20 (9.21 kg,
93.1 LCWP, 67% corrected yield) as a yellow solid. Mp
1
233 °C dec.; [R]D +8.25 (c ) 0.08, DMF); H NMR (400
MHz, DMSO-d6): δ 2.68 (3 H, s), 2.91 (1 H, s), 2.92 (1 H,
s), 3.21 (1 H, m), 3.44 (1 H, m), 3.56 (1 H, m), 3.65 (4 H,
m), 6.71 (2 H, s); 13C NMR (100.6 MHz, DMSO-d6): δ 73.5,
69.1, 66.3, 66.2, 51.6, 37.1; HRMS (ESI+) calcd for
C6H15N2O4S 211.0753, found 211.0744. Anal. Calcd for
C6H15N2O4S: C, 34.28; H, 6.71; N, 13.32. Found: C, 34.28;
H, 6.71; N, 13.19.
Methyl 5-Chloro-2-methylnicotinate (18). Methyl acetoac-
etate (12.6 kg, 108.5 mol) and THF (247.3 kg) were charged
to a vessel, put under a nitrogen atmosphere, and then cooled
to -10 °C. Solid potassium tert-butoxide (12.9 kg, 114.5 mol)
was then carefully added in four equal portions, ensuring that
the internal temperature remained <0 °C. Once this addition
was complete, the batch was warmed to 25 °C over a period of
35 min, and then aged at this temperature for another 70 min.
Solid vinamidinium salt 19 (50.0 kg, 163.1 mol) was then added,
followed by DABCO (12.2 kg, 108.7 mol). The resulting
reaction mixture was warmed to 45 °C, and then left to age at
this temperature for 3 h. At this point, solid ammonium acetate
(20.1 kg, 260.7 mol) was added, the temperature of the batch
was increased to 60 °C, and it was aged at this temperature
overnight (16 h). The resulting reaction mixture was cooled to
20 °C, and water (120 kg) was then added. After stirring for 5
min, the lower aqueous phase was removed. Brine (14.4 kg
NaCl in 40 kg water) was then added, the biphasic mixture
was stirred for 5 min, and the lower aqueous layer was separated
off. The resulting THF layer was extracted three times with
heptane. The combined heptane extracts were then washed twice
with 3:1 v/v water/acetonitrile (2 × 37.6 and 9.8 kg, respec-
tively). The resulting heptane layer was concentrated to
minimum stirred volume, using partial vacuum and maintaining
the batch temperature between 30 and 40 °C. Heptane (34.1
kg) was added, and the batch was reconcentrated to minimum
1
238-240 °C. H NMR (400 MHz, DMSO-d6): δ 7.60 (4 H,
m), 7.81 (1 H, d, J ) 16.2 Hz), 8.05 (1 H, d, J ) 16.2 Hz),
8.25 (1 H, d, J ) 2.5 Hz), 8.78 (1 H, d, J ) 2.5 Hz), 13.82 (1
H, s, br); 13C NMR (100.6 MHz, DMSO-d6): δ 166.8, 152.6,
150.7, 138.1, 135.9, 134.4, 132.3, 129.6, 129.3, 126.3, 125.4,
122.4; HRMS (ESI+) calcd for C14H10NO2ClBr 337.9583,
found 337.9574.
856
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development

1
7-Bromo-3-chloro-5H-benzo[4,5]cyclohepta[1,2-b]pyridin-
5-one (2). Polyphosphoric acid (12.0 kg) was charged to a 20
L round-bottom flask, fitted with an overhead stirrer, N2 inlet,
N2 bubbler, and temperature probe. The flask was put under a
nitrogen atmosphere, and then heated to ∼180 °C (internal
temperature). Solid acid 20 (1.12 kg, 3.3 mol) was added
portion-wise, and the reaction mixture was then heated to 215
°C (internal temperature). The resulting dark brown mixture
was left to age at this temperature (210-220 °C) until the
reaction was complete. After 5 h, the heating mantle was
removed and the reaction mixture was allowed to cool to 90
°C. Water (10.8 kg) was then VERY CAREFULLY added to
the stirred batch at a rate sufficient to maintain the internal
temperature at 80-100 °C. Once the addition was complete,
the batch was allowed to cool to <50 °C, and then poured into
a workup vessel, rinsing the 20 L flask into the vessel with 4.8
kg of water. In total six batches of this Friedel-Crafts
cyclization was run, using the above procedure, and these were
then combined at this stage in a 400 L vessel for a single
workup. To the combined aqueous slurry was added acetonitrile
(113.8 kg) over a period of 1 h at 20-25 °C. Once the addition
was complete, the resulting mixture was stirred for 15 min and
then filtered, washing the wet-cake with acetonitrile (28.4 kg).
The filter-cake was dried on the filter overnight under a stream
of nitrogen. The dried dark green-brown solid (9.10 kg) was
charged back to the empty 400 L vessel. Dichloromethane (280
kg) and triethylamine (3.4 kg, 33.6 mol) were added, and the
resulting mixture was stirred for 30 min at ambient temperature.
A slurry of EcoSorb C-941 (0.9 kg) in dichloromethane (13.3
kg) was then added, rinsing in with a second portion of
dichloromethane (13.3 kg), and the batch was left to age for a
further 1 h. The resulting mixture was filtered through a pad of
Hyflo Super-Cel (followed by a 1 µm in-line filter), washing
the filter pad twice with dichloromethane (31 and 28 kg). The
combined filtrate and washes were concentrated to a volume
of 50 L, at atmospheric pressure, and then solvent-switched to
acetonitrile to a final volume of 120 L, also at atmospheric
pressure. The resulting slurry was cooled to 20 °C, via a linear
ramp over a period of 6 h, and then aged at this temperature
overnight. The slurry was then filtered, washing the wet-cake
three times with acetonitrile (3 × 23 kg). The resulting filter-
cake was dried under vacuum at 60 °C to afford 5.44 kg of
crude tricyclic ketone 2 (82.1 LCWP, 94.2 LCAP) as a slightly
tacky black solid. This crude solid was charged to a vessel,
and dichloromethane (224.7 kg) and EcoSorb C-941 (0.81 kg)
were added. The resulting stirred mixture was heated to a gentle
reflux, aged at this temperature for 1 h, and then cooled to
ambient temperature. The batch was then filtered through a pad
of Solka-Floc (followed by a 1 µm in-line filter), washing the
filter pad with dichloromethane (11.4 kg). The combined filtrate
and wash were concentrated to a volume of 55 L, at atmospheric
pressure, and then solvent-switched to heptane to a final volume
of 47 L (used 88 kg heptane), also at atmospheric pressure.
The resulting slurry was allowed to cool to ambient temperature
overnight and then filtered, washing the wet-cake twice with
heptane (2 × 11.0 kg). The resulting filter-cake was dried under
vacuum at 40 °C to afford tricyclic ketone 2 (4.28 kg, >100
LCWP relative to standard, 99.6 LCAP, 67% corrected yield)
as a beige-coloured solid. Mp 185-186 °C. H NMR (400
MHz, DMSO-d6): δ 7.36 (1 H, d, J ) 12.7 Hz), 7.49 (1 H, d,
J ) 12.7 Hz), 7.81 (1 H, d, J ) 8.6 Hz), 8.05 (1 H, dd, J )
2.2, 8.6 Hz), 8.24 (1 H, d, J ) 2.2 Hz), 8.47 (1 H, d, J ) 2.6
Hz), 9.03 (1 H, d, J ) 2.6 Hz); 13C NMR (100.6 MHz, DMSO-
d6): δ 188.0, 152.6, 150.4, 139.1, 137.9, 136.5, 134.9, 134.2,
134.0, 133.9, 132.9, 132.8, 131.3, 123.8; HRMS (ESI+) calcd
for C14H8NOClBr 319.9478, found 319.9469. Anal. Calcd for
C14H8NOClBr: C, 52.45; H, 2.20; N, 4.37. Found: C, 52.57;
H, 2.20; N, 4.26.
N′-(3-Chloro-5-oxo-5H-benzo[4,5]cyclohepta[1,2-b]pyri-
din-7-yl)-N-[(2R)-1,4-dioxan-2-ylmethyl]-N-methylsulfa-
mide (22). Tricyclic ketone 2 (3.4 kg, 10.6 mol), sulfamide 3
(2.88 kg, 13.7 mol), Xantphos (937 g, 1.6 mol), Pd2(dba)3 (674
g, 0.7 mol), cesium carbonate (5.5 kg, 16.8 mol), and THF (53.7
kg) were charged to a vessel. The resulting mixture was
degassed, by three vacuum/nitrogen purge cycles followed by
subsurface degassing for 30 min, heated to 60 °C, and then
aged at this temperature for 24 h. The bright yellow slurry
obtained was cooled to 20 °C and filtered, washing the wet-
cake with toluene (2 × 15 kg). The resulting filter-cake was
dried, over the weekend, under vacuum at 60 °C to afford crude
sulfamide 22 (9.6 kg, 34 LCWP, 96 LCAP, 69% corrected
yield) as a yellow solid. Mp 167-169 °C; [R]D -35.0 (c )
0.002, DMF); 1H NMR (400 MHz, DMSO-d6): δ 2.80 (3 H s),
3.15 (3 H, m), 3.42 (2 H, m), 3.62 (4 H, m), 7.24 (1 H, d, J )
12.7 Hz), 7.43 (1 H, d, J ) 12.7 Hz), 7.60 (1 H, dd, J ) 2.4,
8.6 Hz), 7.81 (1 H, d, J ) 8.6 Hz), 7.99 (1 H, d, J ) 2.4 Hz),
8.51 (1 H, d, J ) 2.4 Hz), 8.98 (1 H, d, J ) 2.4 Hz), 10.53 (1
H, s); 13C NMR (100.6 MHz, DMSO-d6): δ 188.5, 152.3, 150.6,
140.7, 138.6, 137.9, 134.6, 134.5, 133.5, 130.8, 130.6, 130.0,
123.5, 118.7, 73.5, 68.7, 66.2, 66.1, 51.4, 36.8; HRMS (ESI+)
calcd for C20H21N3O5SCl 450.0890, found 450.0897; Anal.
Calcd for C20H21N3O5SCl: C, 53.39; H, 4.48; N, 9.34. Found:
C, 53.33; H, 4.39; N, 9.29.
4-Bromo-1-methyl-1H-pyrazole (23). A stirred solution of
1-methyl-1H-pyrazole (24, 6.6 kg, 80.4 mol) in dichloromethane
(105.3 kg) was cooled to 10 °C under an atmosphere of nitrogen.
Bromine (19.3 kg, 120.8 mol) was then added over a period of
40 min, whilst maintaining the internal reaction temperature
below 28 °C. The addition line was rinsed through with
dichloromethane (5 kg) into the batch. The resulting reaction
mixture was aged at 20-25 °C for 1.5 h. The reaction was
then carefully quenched by addition of aqueous sodium sulfite
(5 kg Na2SO3 in 20 kg water) over a period of 20 min, whilst
maintaining the internal temperature below 25 °C. Once the
addition was complete, the batch was aged for a further 45 min.
Brine (5 kg NaCl in 25 kg water) was then added, the biphasic
mixture was stirred for 5 min, and the two layers were separated.
The aqueous layer was extracted twice with dichloromethane
(2 × 43.9 kg), and the combined organic layers were then
washed with half-saturated brine (2.5 kg NaCl in 25 kg water).
The resulting dichloromethane solution was concentrated to
minimum stirred volume by distillation under partial vacuum
(>500 mbar) at <35 °C. THF (44 kg) was added and the solution
was reconcentrated to minimum stirred volume, as described
above. A second charge of THF (44 kg) was then added and
the solution was concentrated to a volume of ∼20 L. The
Vol. 14, No. 4, 2010 / Organic Process Research & Development
•
857

resulting THF solution (24.7 kg) of bromopyrazole 23 (11.6
kg by assay, 90% assay yield) was transferred to a 25 L
container, and stored under nitrogen at 5 °C.
added, which led to the product crystallizing from solution. The
resulting slurry was heated to 30 °C, cooled to 20 °C, reheated
to 30 °C, and then left to age overnight. After the overnight
age, water (4 L) was added. The slurry was then filtered,
washing the wet-cake with 2:1 water/DMF (3 L). The resulting
filter-cake was dried under vacuum at 50 °C to afford 1 (2.81
kg, 100 LCWP relative to standard, 99.4 LCAP, 80% corrected
yield) as a yellow solid. Mp 192-193 °C; [R]D +18.3 (c )
0.04, DMF); 1H NMR (400 MHz, DMSO-d6): δ 2.81 (3 H, s),
3.16 (3 H, m), 3.42 (2 H, m), 3.60 (4 H, m), 3.91 (3 H, s), 7.25
(1 H, d, J ) 12.5 Hz), 7.35 (1 H, d, J ) 12.5 Hz), 7.57 (1 H,
m), 7.77 (1 H, d, J ) 8.6 Hz), 7.99 (1 H, m), 8.15 (1 H, s),
8.48 (1 H, s), 8.58 (1 H, m), 9.21 (1 H, m), 10.55 (1 H, s); 13C
NMR (100.6 MHz, DMSO-d6): δ 190.2, 150.6, 149.7, 140.3,
138.8, 137.2, 134.0, 133.3, 133.2, 133.0, 131.6, 130.3, 129.5,
128.8, 123.4, 118.7, 118.1, 73.5, 68.7, 66.2, 66.1, 51.4, 39.3,
36.8; HRMS (ESI+) calcd for C24H26N5O5S 496.1655, found
496.1666. Anal. Calcd for C24H26N5O5S: C, 58.17; H, 5.08; N,
14.13. Found: C, 57.98; H, 5.06; N, 14.25.
Reduction of Palladium Level in 1. 1 (2.8 kg, 5.7 mol)
and DMF (28 L) were charged to a vessel, under a nitrogen
atmosphere. The contents were then warmed to 40 °C and
stirred at this temperature until solid had fully dissolved. The
resulting solution was cooled to 20 °C, the MP-TMT resin (480
g) was added, and the mixture was then left to stir at ambient
temperature overnight. After this age, the batch was filtered
through a pad of Solka-Floc (2.0 kg), washing the pad twice
with DMF (2 × 2.8 L). The combined filtrate and washes were
charged to a second vessel. Water (11 L) was added and the
batch was then seeded (3 g of 1). After stirring for 1 h, in order
to form a seed-bed, a second portion of water (22 L) was slowly
added over a period of 1 h. Once this addition was complete,
the slurry was aged for 2 h at ambient temperature and then
filtered, washing the wet-cake with 1:1 DMF/water (10 L)
followed by water (10 L). The resulting filter-cake was dried
under vacuum at 60 °C to afford 1 (2.71 kg, 7 ppm Pd, >100
LCWP relative to standard, 99.7 LCAP, 97% recovery) as a
yellow solid.
1-Methyl-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-
1H-pyrazole Lithium Hydroxy-ate Complex (28). The THF
solution of 4-bromo-1-methyl-1H-pyrazole (23) (24.3 kg total,
11.4 kg of 23 by assay, 70.8 mol) prepared above was diluted
with THF (47.3 kg) and toluene (56.9 kg), and then put under
a nitrogen atmosphere. Triisopropylborate (16.6 kg, 88.2 mol)
was added, rinsing the addition line into the batch with toluene
(2 kg), and the resulting stirred reaction mixture was cooled to
-70 °C. Hexyllithium (33 wt % in hexane, 30.0 kg, 108 mol)
was then slowly added whilst maintaining the internal reaction
temperature below -63 °C and, once the addition was complete,
the batch was aged for 30 min at this temperature. Pinacol (12.0
kg, 101 mol) was then added, and the resulting reaction mixture
was warmed to 25 °C over 40 min. After aging at this
temperature for 1 h, water (6.1 kg) was added over a period of
10 min. The slurry obtained was aged for ∼2.5 h and then
filtered, washing the wet-cake with MTBE (20.9 kg). The
resulting filter-cake was dried on the filter for 20 min, and then
under vacuum at 35 °C for >12 h, to afford lithium -ate complex
28 (16.56 kg, 78 wt %, 79% corrected yield) as a white solid.
Mp 136-139 °C. 1H NMR (400 MHz, DMSO-d6): δ 0.90 (6
H, m), 1.05 (6 H, m), 1.46 (1 H, s, br), 3.68 (3 H, s), 7.00 (1
H, s), 7.05 (1 H s); 13C NMR (100.6 MHz, DMSO-d6): δ 143.1,
132.7, 77.5, 76.5, 74.0, 37.9, 26.7, 26.5, 26.0, 25.6; HRMS
(ESI+) calcd for C10H18N2O2B (- LiO) 209.1461, found
209.1456.
N-[(2R)-1,4-Dioxan-2-ylmethyl]-N-methyl-N′-[3-(1-meth-
yl-1H-pyrazol-4-yl)-5-oxo-5H-benzo[4,5]cyclohepta[1,2-b]py-
ridin-7-yl]sulfamide (1). Penultimate 22 (3.2 kg, 7.1 mol),
pyrazole -ate complex 28 (2.7 kg, 11.7 mol) and Pd(PtBu3)2
(21.8 g, 0.04 mol) were charged to a vessel. DMF (27 L) was
added, and the mixture was degassed by vacuum/nitrogen purge
cycles followed by subsurface degassing. The reaction mixture
was then heated to 100 °C and left to age at this temperature
for 1 h. The resulting batch was cooled to 25 °C and 2 M
aqueous NaOH (14 L) was then added, whilst maintaining the
internal temperature below 27 °C. Ecosorb C-941 (350 g) was
added and the batch was aged at 22 °C for 1.5 h before filtering
through a pad of Solka-Floc (3.0 kg). The filter pad was washed
with 1:1 DMF/2 M NaOH(aq) (3 L) followed by 1:1 DMF/water
(6 L), and the combined filtrate and washes were returned to
the (cleaned) vessel. Five M Hydrochloric acid (4.8 L) was then
Acknowledgment
We thank Dean Thompson for mass spectrometry work.
Received for review April 14, 2010.
OP100101Q
858
•
Vol. 14, No. 4, 2010 / Organic Process Research & Development