Synthesis of the (3R,6S)-3-amino-6-(2,3-difluorophenyl)azepan-2-one of telcagepant (MK-0974), a calcitonin gene-related peptide receptor antagonist for the treatment of migraine headache

Article abstract of DOI:10.1021/ol8011524

(Chemical Equation Presented) Two novel routes have been developed to the (3R,6S)-3-amino-6-(2,3-difluorophenyl)-1-(2,2,2-trifluoroethyl)azepan-2-one 2 of the CGRP receptor antagonist clinical candidate telcagepant (MK-0974, 1). The first employs a ring-closing metathesis of the styrene 7 as the key reaction, while the second makes use of a highly diastereoselective Hayashi-Miyaura Rh-catalyzed arylboronic acid addition to nitroalkene 16. The latter route has been implemented to produce multigram quantities of telcagepant for extensive preclinical evaluation.

Full text of DOI:10.1021/ol8011524

ORGANIC
LETTERS
2008
Vol. 10, No. 15
3235-3238
Synthesis of the (3R,6S)-3-Amino-6-
(2,3-difluorophenyl)azepan-
2-one of Telcagepant (MK-0974), a
Calcitonin Gene-Related Peptide
Receptor Antagonist for the Treatment
of Migraine Headache
Christopher S. Burgey,* Daniel V. Paone, Anthony W. Shaw, James Z. Deng,
Diem N. Nguyen, Craig M. Potteiger, Samuel L. Graham, Joseph P. Vacca, and
Theresa M. Williams
Department of Medicinal Chemistry, Merck Research Laboratories,
West Point, PennsylVania 19486
christopher_burgey@merck.com
Received November 19, 2007
ABSTRACT
Two novel routes have been developed to the (3R,6S)-3-amino-6-(2,3-difluorophenyl)-1-(2,2,2-trifluoroethyl)azepan-2-one 2 of the CGRP receptor
antagonist clinical candidate telcagepant (MK-0974, 1). The first employs a ring-closing metathesis of the styrene 7 as the key reaction, while
the second makes use of a highly diastereoselective Hayashi-Miyaura Rh-catalyzed arylboronic acid addition to nitroalkene 16. The latter
route has been implemented to produce multigram quantities of telcagepant for extensive preclinical evaluation.
Antagonists of the calcitonin gene-related peptide (CGRP)
receptor represents a potential novel therapy for the
treatment of migraine headaches devoid of the cardiovas-
cular liabilities associated with current drugs.1 A recent
report from these laboratories detailed the identification
of a series of potent, small-molecule, azepanone-based
CGRP receptor antagonists potentially suitable for oral
administration during a migraine attack.2 Based on its
potency, pharmacodynamic model efficacy, and pharma-
cokinetics in preclinical species, a member of this class
of compounds was selected as a development candidate
(Figure 1, 1, telcagepant, MK-0974) and has recently
demonstrated efficacy in migraine clinical trials.3 In
support of the preclinical evaluation of this molecule, a
scaleable, stereoselective synthesis was required, and
herein, we disclose our efforts toward this objective.
Among the challenges associated with the development
of a successful synthesis of the (3R,6S)-3-amino-6-(2,3-
difluorophenyl)-1-(2,2,2-trifluoroethyl)-azepan-2-one, 2
(1) (a) Durham, P. L.; Russo, A. F. Pharmacol. Ther. 2002, 94, 77. (b)
Edvinsson, L. Expert Opin. Ther. Targets 2003, 7, 377. (c) Durham, P. L.
N. Engl. J. Med. 2004, 350, 1073. (d) Olesen, J.; Diener, H-C.; Husstedt,
I. W.; Goadsby, P. J.; Hall, D.; Meier, U.; Pollentier, S.; Lesko, L. N. Engl.
J. Med. 2004, 350, 1104–1110. Petersen, K. A.; Birk, S.; Lassen, L. H.;
Krusse, C.; Jonassen, O.; Lesko, L.; Olesen, J. Cephalalgia 2004, 25, 139.
(2) Paone, D. V.; Shaw, A. W.; Nguyen, D. N.; Burgey, C. S.; Deng,
Z. J.; Kane, S. A.; Koblan, K. S.; Salvatore, C. A.; Hershey, J., C.; Wong,
B.; Roller, S. G.; Miller-Stein, C.; Graham, S. L.; Vacca, J. P.; Williams,
T. M. J. Med. Chem. 2007, 50, 5564–5567.
(3) Ho, T. W.; Mannix, L.; Fan, X.; Assaid, C.; Furtek, C.; Jones, C.;
Lines, C.; Rapoport, A. Neurology 2008, 70, 1305–1312.
10.1021/ol8011524 CCC: $40.75
Published on Web 07/01/2008
2008 American Chemical Society

protecting group occludes reduction of the re face of the
olefin that would produce the required trans-azepanone;
accordingly, deprotection of the 2,4-dimethoxybenzyl group
with trifluoroacetic acid afforded the alternative hydrogena-
tion substrate 9. Screening of a range of metal catalysts and
solvents revealed that hydrogenation using Pd/C in toluene,
with in situ Boc reprotection, afforded the desired product
10 in a 2: 1 trans/cis ratio (51% isolated trans). Treatment
of the lactam 10 with 1 equivalent of sodium hydride
followed by addition of 2,2,2-trifluoroethyltrichloromethane
sulfonate at low temperature (-30 °C) afforded 11. Depro-
tection to the targeted primary amine 2 followed by urea
formation with piperidine 126 gave telcagepant (1) in high
yield.
The preceding route could be implemented to deliver
limited quantities of 1 for preliminary characterization,
but it suffered from several drawbacks. The key metathesis
reaction was inefficient, requiring large amounts of the
Ru catalyst (30 mol %) which would render scale-up
prohibitive; additionally, the hydrogenation proceeded
with only modest levels of diastereoselectivity to produce
a cis/trans mixture. As a consequence, we sought to
develop a more efficient, scaleable, and diastereoselective
route.
Figure 1. Telcagepant.
(Figure 2), of telcagepant is the remote C6 stereocenter.
To address this problem, a route was envisioned in which
olefin 3 would serve as a substrate to investigate diaste-
reoselective reductions; implicit in this strategy is the
potential utilization of the C3 amino group to establish
the remote C6 stereocenter. A ring-closing metathesis
carbon-carbon bond forming reaction of the R-styrene 4
would assemble the 7-member lactam ring, with a D-
amino acid starting material serving to establish the
required 3R-amino stereochemistry.
A second-generation route was proposed which would
employ an unprecedented diastereoselective Hayashi-Miyaura
Rh-catalyzed arylboronic acid addition7 to nitroalkene 5 to
install the key C6 stereocenter (Figure 2, 5 to 2). The
efficiency of this transformation was difficult to predict due
to the conspicuous lack of examples employing R-unsubsti-
tuted nitroalkenes; additionally, as the use of these Rh-
catalyzed additions has been limited to achiral substrates, it
was uncertain if the stereochemical outcome would be under
catalyst control or be subject to substrate control due to the
preexisting C3 stereocenter residing in 5.
The commercially available D-glutamic acid derivative
13 was elaborated, using a modification of the literature
conditions,8 to the orthogonally protected diester 14 in
81% yield over two steps (Scheme 2). DIBAL-H effected
a selective half-reduction of the γ-ester to afford aldehyde
15, which was used without purification in a Henry
reaction with nitromethane (toluene, catalytic tetrameth-
ylguanidine). Subsequent addition of methanesulfonyl
chloride and triethylamine to the intermediate nitro alcohol
effected elimination to give the nitro olefin 16 in good
overall yield.
Figure 2. Two routes to azepanone 2.
A Suzuki cross-coupling reaction of vinyl bromide 6
(Scheme 1), prepared in two steps from L-allylglycine,2 with
2,3-difluorophenylboronic acid proceeded in good yield to
afford the metathesis reaction precursor 7. Cyclization of this
challenging substrate was accomplished using 30 mol % of
the Grubbs second-generation ruthenium catalyst4 under
moderately dilute conditions to deliver the azepenone 8 in
58% yield.5 The metathesis product 8 served as the substrate
to study diastereoselective olefin reduction conditions. Reac-
tion of olefins such as 8 with an array of hydrogenation
catalysts and solvents afforded the saturated caprolactams
with trans/cis ratios ranging from 1:5 to 1:10. Molecular
modeling studies indicated that the methylene of the DMB
With an established synthesis of the alkene 16 in place,
experimentstowardtheexecutionofthekeyHayashi-Miyaura
Rh-catalyzed arylboronic acid addition could be under-
(6) Burgey, C. S.; Stump, C. A.; Nguyen, D. N.; Deng, Z. J.; Quigley,
A. G.; Norton, B. R.; Bell, I. M.; Mosser, S. D.; Salvatore, C. A.; Rutledge,
R. Z.; Kane, S. A.; Koblan, K; Vacca, J. P.; Graham, S. L.; Williams, T. M.
Bioorg. Med. Chem. Lett. 2006, 16, 5052–5056.
(7) (a) Hayashi, T.; Yamasaki, K. Chem. ReV. 2003, 103, 2829–2844.
Hayashi, T.; Senda, T.; Ogasawara, M. J. Am. Chem. Soc. 2000, 122, 10716–
10717.
(4) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18–29.
(5) Both the Grubbs first-generation and the Grubbs-Hoveyda catalysts
were ineffective for this transformation. The major byproducts detected in
this reaction were dimerization of the substrate and an adduct incorporating
the styrene from the Grubbs precatalyst.
(8) Padron, J. M.; Kokotos, G.; Martin, T.; Markidis, T.; Gibbons, W. A.;
Martin, V. S. Tetrahedron: Asymmetry 1998, 9, 3381–3394.
(9) Senda, T.; Ogasawara, M.; Hayashi, T. J. Org. Chem. 2001, 66,
6852–6856.
3236
Org. Lett., Vol. 10, No. 15, 2008

Scheme 1. Ring-Closing Metathesis Route
taken. Application of the standard literature conditions (3
mol % Rh(acac)(C2H4)2/BINAP, 10:1 dioxane/water,
100 °C)7 led to rapid hydrolysis of the 2,3-difluorophe-
nylboronic acid to 1,2-difluorobenzene. Lowering the
reaction temperature to 45 °C allowed olefin addition of
the boronic acid to proceed competitively with decom-
position;9 however, addition of further quantities of metal/
ligand (10-20 mol %, cumulative) and boronic acid were
typically required for the reaction to reach completion.10
These conditions could be implemented to deliver the
desired adduct 17 with high levels of diastereoselection
(93:7) and in good yield (80%; Table 1, entry 1).
Constrained by the propensity of the highly electron-
deficient 2,3-difluorophenylboronic acid to decompose at
elevated temperatures, the generation of more active
catalysts was explored as a means to accelerate this
reaction. The addition of inorganic bases to these Rh-
catalyzed addition reactions has been shown to have a
pronounced rate-accelerating effect,11 presumably due to
Scheme 2. Hayashi-Miyaura Route
Org. Lett., Vol. 10, No. 15, 2008
3237

to EDC-mediated lactamization to afford the diBoc capro-
lactam 19 in 65% average yield for the two steps (Scheme
2). Selective deprotection with trifluoroacetic acid in dichlo-
romethane effected conversion of 19 to azepanone 10, which
was identical to the material from the ring-closing metathesis
route and served to confirm the stereochemical outcome of
the Hayashi-Miyaura reaction.
Table 1. Optimization of Hayashi-Miyaura Reaction
The production of the 6S stereocenter via the S-BINAP-
Rh(L)n complex is consistent with the stereochemical model
proposed by Hayashi for arylboronic acid additions to
enones.7 Alkene 16 contains a residing stereocenter that, in
principle, could exert stereocontrol in this transformation.
In order to confirm that the observed diastereoselectivity is
the result of catalyst control, the reaction was performed with
racemic BINAP. This experiment revealed a modest prefer-
ence for the undesired 6R stereocenter (Table 1, entry 3);
furthermore, use of the enantiomeric R-BINAP afforded the
diastereomeric 6R product with excellent levels of selectivity
(Table 1, entry 4). Thus, the slight inherent diastereofacial
preference of the substrate can be easily overcome by the
S-BINAP-Rh(L)n complex to afford the required 6S product.
entry BINAP T (°C) mol % catalyst yield (%) 6S:6R
1
S
S
rac
R
45
35
35
35
10-20
2.5
3
80
93:7b
93:7c
2a
3a
4a
96
50d
76d
41:59c
3:97c
3
a 50 mol % of NaHCO3 was added. b Ratio determined by NMR. c Ratio
determined by HPLC. d Unoptimized.
the formation of a highly active BINAP-Rh-OH species.12
Reaction optimization studies revealed that NaHCO3 (0.5
equiv) was an efficacious additive, enabling the realization
of an improved procedure that typically proceeded to
completion within 6-12 h at 35 °C with practical amounts
of catalyst (2.5 mol % Rh(acac)(C2H4)2/BINAP) and
boronic acid (2.5 equiv).13 These improved reaction
conditions reproducibly delivered the desired product in
high yield and diastereoselectivity (96%, 93: 7, Table 1,
entry 2) on a scale of up to 2 kg.
In summary, we have developed two novel routes to the
(3R,6S)-3-amino-6-(2,3-difluorophenyl)-1-(2,2,2-trifluoroet-
hyl)azepan-2-one, 2, of the CGRP receptor antagonist
telcagepant (1). Ultimately, the highly stereoselective syn-
thesis employing a key Hayashi-Miyaura nitro olefin
addition was implemented to produce multigram quantities
of telcagepant to support initial preclinical and clinical
studies.
Exhaustive Pd/C-catalyzed hydrogenolysis of nitroalkane
ester 17 gave amino acid 18 which was directly subjected
Acknowledgment. We thank the analytical chemistry,
mass spectroscopy, and NMR analysis groups for their
assistance.
(10) A characteristic deep red color served as a visual indicator of the
catalyst activity, and transition to a pale orange signaled that the reaction
had stalled.
(11) Itooka, R.; Iguchi, Y.; Miyaura, N. J. Org. Chem. 2003, 68, 6000–
6004.
Supporting Information Available: Experimental pro-
cedure and characterization data are available. This material
is available free of charge via the Internet at http://pubs.acs.org.
(12) Hayashi, T.; Takahashi, M.; Takaya, Y.; Ogasawara, M. J. Am.
Chem. Soc. 2002, 124, 5052–5058.
(13) Notably, the characteristic deep red color of the active catalyst is
maintained throughout the course of the reaction.
OL8011524
3238
Org. Lett., Vol. 10, No. 15, 2008