A double ring closing metathesis reaction in the rapid, enantioselective synthesis of NK-1 receptor antagonists.

Article abstract of DOI:10.1021/ol006958g

[structure: see text]. The NK-1 receptor antagonist 1 has been prepared in seven steps from phenylglycine methyl ester. The key steps are a double ring closing metathesis reaction of tetraene 7 to prepare spirocycle 6 and a reductive Heck reaction to introduce the aryl moiety. This latter reaction discriminates the olefins of compound 6 and proceeds in a highly regio- and stereoselective manner.

Full text of DOI:10.1021/ol006958g

ORGANIC
LETTERS
2001
Vol. 3, No. 5
671-674
A Double Ring Closing Metathesis
Reaction in the Rapid, Enantioselective
Synthesis of NK-1 Receptor Antagonists
Debra J. Wallace,* Jonathan M. Goodman,† Derek J. Kennedy,* Antony J. Davies,
Cameron J. Cowden, Michael S. Ashwood, Ian F. Cottrell, Ulf-H. Dolling, and
Paul J. Reider
Department of Process Research, Merck Sharp & Dohme Research Laboratories,
Hertford Road, Hoddesdon, Hertfordshire, EN11 9BU, UK
debra_wallace@merck.com
Received December 5, 2000
ABSTRACT
The NK-1 receptor antagonist 1 has been prepared in seven steps from phenylglycine methyl ester. The key steps are a double ring closing
metathesis reaction of tetraene 7 to prepare spirocycle 6 and a reductive Heck reaction to introduce the aryl moiety. This latter reaction
discriminates the olefins of compound 6 and proceeds in a highly regio- and stereoselective manner.
As part of our ongoing program to prepare selective NK-1
receptor antagonists, we required an efficient enantioselective
synthesis of the 1-oxo-7-azaspirodecane 1. Such a synthesis
should be suitable for generating multigram quantities to
allow for systematic variation of alkyl groups on nitrogen
and oxygen.
The initial route to this compound used a reductive Heck
reaction of spiroalkene 2 with an aryl iodide to set the final
stereocenter (Scheme 1).1 The spiroalkene was obtained in
development purposes, a shorter and more practical synthesis
of compound 1 was required.
We were attracted to the possibility of forming both the
spirocyclic rings in a single step using a double ring closing
metathesis (RCM) reaction.2 We have recently reported the
diastereoselective double RCM reaction of amino acid
derived tetraenes 4a-d.3 This reaction afforded the spiro-
cyclic compounds 5a-d in >90% ds, and we envisaged that
this approach could be extended to spirodiene 6 (R ) Ph)
(Scheme 2).4 In this case, suppression of epimerization at
the labile benzylic stereocenter during the synthesis of the
metathesis precursor 7 would be needed. Selective function-
alization of the two olefins in product 6 would also be a
significant synthetic challenge.
Scheme 1
Scheme 2
three steps from ketone 3; however, this itself required a
seven-step synthesis including a resolution.1c For further
* Questions regarding the NOE studies should be addressed to D. J.
Kennedy (derek_kennedy@merck.com) and those concerning the molecular
modeling to J. M. Goodman (j.m.goodman@ch.cam.ac.uk).
† Present address: University Chemical Laboratory, Lensfield Road,
Cambridge, CB2 1EW, UK.
10.1021/ol006958g CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/16/2001

On the basis of our previous work, the tosyl group was
chosen as a suitable N-protecting group, and hence, com-
mercially available phenylglycine ester 8 was converted to
its N-tosyl derivative 9 by reaction with tosyl chloride and
triethylamine (Scheme 3). Cerium-mediated addition of
vinylmagnesium bromide to ester 9 afforded the tertiary
alcohol 10 without loss of optical purity as determined by
chiral HPLC methods (Diacel Chiracel OD-R 0.46 × 25 cm
column). As expected, in the absence of cerium chloride,
1,4-addition to the enone intermediate (structure not shown)
was a significant side reaction. N,O-Diallylation of 10 was
achieved in a single step by reaction with excess sodium
hydride and allyl bromide to give tetraene 7 in excellent
yield.
of the major isomer to known intermediates (vide infra)
confirmed this assignment. The pseudoaxial orientation of
the phenyl group was also ascertained by NOE studies, and
this may be significant to the stereochemical outcome of the
RCM reaction. When employing benzyl protecting groups
on nitrogen (Bn or PMB), the selectivity of the related double
RCM reaction was reversed, i.e., 5-epi-6 was favored over
6 in Scheme 4 (Bn protecting group instead of Ts, 65% ds).
NMR studies of these spirocyclic compounds then indicated
that the 6-phenyl group was equatorial.7
Scheme 4
Scheme 3
To complete the synthesis of target 1, selective function-
alization of the two olefins of 6 was now required. Initial
studies looked at the selective removal of the piperidine
olefin; however, this led to mixtures of reduced products.
An ambitious approach was to perform the reductive Heck
reaction on the diene and hope for a regio- and stereoselective
reaction. In theory, provided only single addition occurs,
eight different isomeric products could result. We were
gratified to find that reaction of spirodiene 6 with aryl iodide
12, in the presence of catalytic palladium acetate, afforded
the desired compound 13 in 60% yield with 90% ds (Scheme
5). Notably the addition of water (5%) to the reaction solvent
The key double RCM reaction of 7 proceeded smoothly
under our previously optimized conditions employing the
Grubbs’ catalyst 115 (Scheme 4). With only 4 mol % of
catalyst, this gave the two chromatographically distinct and
crystalline diastereoisomers 6 and 5-epi-6 in 86% yield and
70% ds, favoring the desired compound.6 The stereochem-
istry of the products was determined by NOE studies. In
particular, the minor isomer 5-epi-6 gave an NOE enhance-
ment as shown that was absent in the major isomer 6. A
single-crystal X-ray determination and subsequent conversion
Scheme 5
(1) (a) Baker, R.; Curtis, N. R.; Elliott, J. M.; Harrison, T.; Hollingworth,
G. J.; Jackson, P. S.; Kulagowski, J. J.; Rupniak, N. M. WO 97/49710. (b)
Kulagowski, J. J.; Curtis, N. R.; Swain, C. J.; Williams, B. J. Org. Lett
2001, 3, 667. (c) Harrison, T.; Williams, B. J.; Swain, C. J.; Ball, R. G.
Bioorg. Med. Chem. Lett. 1994, 4, 2545.
(2) For recent reviews of the metathesis reaction in synthesis, see: (a)
Grubbs, R. H.; Chang, S. Tetrahedron 1998, 54, 4413. (b) Armstrong, S.
K. J. Chem. Soc., Perkin Trans. 1 1998, 371. (c) Schuster, M.; Blechert, S.
Angew. Chem., Int. Ed. Engl. 1997, 36, 2036. (d) Tetrahedron symposium
in print; Snapper, M. L., Hoveyda, A. H., Eds.; 1999; Vol. 55, p 8141. (e)
Alkene Metathesis in Organic Synthesis; Furstner, A., Ed. Top. Organomet.
Chem. 1998, 1. (f) Furstner, A. Angew. Chem., Int. Ed. 2000, 39, 3012.
(3) Wallace, D. J.; Cowden, C. J.; Kennedy, D. J.; Ashwood, M. S.;
Cottrell, I. F.; Dolling, U.-H. Tetrahedron Lett. 2000, 41, 2027.
(4) For other examples of diastereoselective double RCM reactions,
see: (a) Lautens, M.; Hughes, G. Angew. Chem., Int. Ed. 1999, 38, 129.
(b) Bassindale, M. J.; Hamley, P.; Leitner, A.; Harrity, J. P. A. Tetrahedron
Lett. 1999, 40, 3247. (c) Schmidt, B.; Westhus, M. Tetrahedron 2000, 56,
2421. (d) Bassindale, M. J.; Edwards A. S.; Hamley, P.; Adams, A.; Harrity,
J. P. A. J. Chem. Soc., Chem. Commun. 2000, 1035. (e) Lautens, M.;
Hughes, G.; Zunic, V. Can. J. Chem. 2000, 78, 868. (f) Schmidt, B.;
Wildemann, H. J. Org. Chem. 2000, 65, 5817. (g) Wallace, D. J.; Bulger,
P. G.; Kennedy, D. J.; Ashwood, M. S.; Cottrell, I. F.; Dolling, U.-H. Synlett.
In press.
was found to be key for obtaining high selectivities for this
transformation.8 While the exact role of the water is unclear,
672
Org. Lett., Vol. 3, No. 5, 2001

Scheme 6
in the absence of this additive significant quantities of
Me), 5b (R ) i-Pr), and their C-5 epimers using the MM2*
force field10 as implemented in Macromodel 511 showed that
in each case there was little energetic difference between
the two diastereoisomers.12 This suggests that these reactions
do not proceed under thermodynamic control as good
selectivity was obtained in RCM reactions to produce 5a
and 5b (>90% ds). In contrast, when comparing 6 (R )
Ph) with 5-epi-6, a larger energy difference (ca. 10 kJ/mol)
in favor of 6 was calculated. Hence, it would have been
expected, on the basis of thermodynamic considerations, that
a higher selectivity would be obtained for this compound
rather than the observed lower selectivity.
We have also resubjected purified 6 and 5-epi-6 to the
metathesis reaction and found that they do not equilibrate
under these conditions, even under an ethylene atmosphere.
In view of this it is worth considering that as two rings are
formed, the order of ring formation is important, i.e., the
reaction could proceed by the initial formation of either a
five- or a six-membered ring (Path A or Path B in Scheme
6). The relative energy levels of monocyclized intermediates
may be more significant in determining the stereochemical
outcome than those of the final product.
In an attempt to detect these intermediates and elucidate
the mechanistic pathway, a reaction was carried out with
reduced catalyst load. Hence, treatment of the starting
material 7 with catalyst 11 (1 mol %) and quenching after
30 min led to a mixture containing 20% spirodienes (70%
ds), 20% residual starting material, and four intermediates.13
Purification of this mixture by preparative HPLC allowed
for isolation and identification of these four compounds. The
major intermediates observed were the dihydrofurans 14a
regioisomeric products were observed. In line with previous
studies, the arylpalladium species approached from the less
hindered face of the dihydrofuran to give the desired epimer.
This stereochemical outcome was deduced by a single NOE
interaction as shown and confirmed by later comparison with
known intermediates.
Hydrogenation of 13 using Pearlman’s catalyst cleaved
the phenolic protecting group and reduced the remaining
double bond. Removal of the tosyl protecting group was
achieved by reaction with sodium naphthalide to give 1 in
78% yield over the final two steps. This completed the
synthesis of an important NK-1 receptor antagonist in
optically pure form from a commercially available starting
material in just seven steps, without the need for a resolution.
Formation of the N-Boc derivative of 1 gave material
identical in all respects (NMR, HPLC, RD) to that synthesized
previously.1
With the synthesis of compound 1 completed, we turned
our attention to the mechanistic aspects of the double RCM
reaction. We were surprised by the lower selectivity obtained
for the RCM of tetraene 79 compared with those of related
tetraenes 4a-d (70% ds vs >90% ds).3 To obtain an insight
into this differing selectivity and the origin of the stereo-
control, molecular modeling studies were carried out.
Analysis of the previously synthesized structures 5a (R )
(5) (a) Schwab, P.; France, M. B.; Ziller, J. W.; Grubbs, R. H. Angew.
Chem., Int. Ed. Engl. 1995, 34, 2039. (b) Schwab, P.; Grubbs, R. H.; Ziller,
J. W. J. Am. Chem. Soc. 1996, 118, 100.
(6) Isomer ratios quoted were determined by hplc analysis (Zorbax
Eclipse XDB-C8 0.46 × 25 cm column). Yields are for the isolated
compounds after column chromatography.
(7) (a) Scott, J. W.; Durham, L. J.; deJongh, H. A. P.; Burckhardt, U.;
Johnson, W. S. Tetrahedron Lett. 1967, 2381. (b) Johnson, F. Chem. ReV.
1968, 68, 375.
(8) For examples of the use of water in Heck reactions, see: (a) Genet,
J. P.; Blart, E.; Savignac, M. Synlett 1992, 715. (b) Lemaire-Audoire, S.;
Savignac, M.; Dupuis, C.; Genet, J. P. Tetrahedron Lett. 1996, 37, 2003.
(c) Jeffery, T. Tetrahedron 1996, 52, 10113.
(10) Allinger, N. L. J. Am. Chem. Soc. 1977, 99, 8127.
(11) Mohamedi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440.
(12) Chang, G.; Guida, W. C.; Still, W. C. Monte Carlo searches were
performed to ensure that all low-energy conformations had been included
in the calculations. J. Am. Chem. Soc. 1989, 111, 4379.
(13) Relative ratios were determined by HPLC analysis.
(9) Despite screening a range of reaction conditions (temperature, solvent,
additives and other catalysts), this selectivity could not be improved.
Org. Lett., Vol. 3, No. 5, 2001
673

and 14b formed in approximately equal amounts. Addition-
ally, small quantities of the six-membered ring 15a and the
eight-membered ring 16 were produced. Moreover, the
relative ratios of 14a, 14b, and 15a to each other remain
constant throughout the course of the reaction. On the basis
of these observations, the predominant pathway operating
is the initial formation of the five-membered ring (Path A).14
When re-subjecting the monocyclic compounds to the
reaction conditions, the isolated intermediates 14a and 15a
both reacted to afford the expected metathesis product 6 as
the sole spirocycle. Surprisingly, when 14b was re-subjected
to the reaction conditions, a 65:35 mixture of 5-epi-6 and 6
was obtained. We believe this mixture results from the
expected formation of 5-epi-6 as well as cyclization of the
pendant allyl group onto the preformed dihydrofuran ring
(to give 15a).15 The exclusive conversion of 14a and 15a to
give 6, together with the nonequilibration of the final
spirocycles, provides evidence that the reaction selectivity
is usually based on the first cyclization. However, the
conversion of 14b into a mixture of 5-epi-6 and 6 indicates
that competing pathways can operate.16
To summarize, three intermediates in the double ring
closing metathesis reaction of tetraene 7 have been isolated.
This has allowed us to propose that the predominant pathway
operating is the initial formation of the dihydrofurans. Re-
subjection of these intermediates to the reaction conditions
has provided further insight into the reaction mechanism. In
conjunction with the regio- and stereoselective reductive
Heck reaction, this chemistry has allowed for a new seven-
step synthesis of a key NK-1 receptor antagonist.
Acknowledgment. We thank Richard Ball and Nancy
Tsou for the crystallographic studies, Andrew Gibson for
preliminary experiments, and Paul Byway for mass spec-
trometry.
Supporting Information Available: Experimental pro-
cedures and spectroscopic data for compounds 1, N-Boc-1,
6, 5-epi-6, 7, 9, 10, 13 14a 14b 15b, and 16. Tables of
molecular modeling data. This material is available free of
charge via the Internet at http://pubs.acs.org.
OL006958G
(14) It is possible that 15a and 15b are formed in significant quantities
but that their further cyclization is faster than that for the dihydrofurans
14a and 14b. However, the consistent ratio of the two product isomers
throughout the reaction (70:30 at all conVersions) does not support this.
(15) No acyclic compounds were observed during this reaction.
(16) Details of modeling studies on the intermediate compounds are
provided in the Supporting Information.
674
Org. Lett., Vol. 3, No. 5, 2001