On the mechanism of a double ring-closing metathesis reaction

Article abstract of DOI:10.1016/j.tetlet.2004.11.149

A detailed study of the steps involved in the double ring-closing metathesis reaction of 2 to 3 has been carried out. Both the selectivity and mechanism were affected by choice of catalyst. A detailed study of the steps involved in the double ring-closing metathesis reaction of 2 to 3 has been carried out. Both the selectivity and mechanism were affected by choice of catalyst.

Full text of DOI:10.1016/j.tetlet.2004.11.149

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 591–594
On the mechanism of a double ring-closing metathesis reaction
Debra J. Wallace*
Department of Process Research, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
Received 18 October 2004; revised 24 November 2004; accepted 29 November 2004
Abstract—A detailed studyof the steps involved in the double ring-closing metathesis reaction of 2 to 3 has been carried out. Both
the selectivityand mechanism were affected bychoice of catalyst.
Ó 2004 Published by Elsevier Ltd.
We have recentlypublished the synthesis of the NK1
receptor antagonist 1, which proceeds via a diastereo-
selective double ring-closing metathesis (RCM) reaction
of tetraene 2 to afford 3a as the major product in 70%
selectivity( Scheme 1).1,2 As part of our ongoing interest
in these double RCM reactions a detailed studyof the
ring-closing step with a range of commerciallyavailable
metathesis catalysts has been carried out.3 In this letter
the results of these studies are presented along with a
model for when the choice of catalyst can influence the
outcome of a diastereoselective RCM reaction.4
use of the more reactive second generation catalyst 56
would allow for thermodynamic control. However, the
reaction of 2 with both this catalyst and with the
Grubbs–Hoveyda catalyst 67 proceeded in favor of
the other isomer 3b, albeit with low selectivity( Scheme
2). Reaction with catalyst 78 gave a comparable result
to that obtained with 4. As with the first generation cat-
alyst no equilibration of either isomer was seen under
the reaction conditions and we again propose that this
outcome is the result of a kinetic cyclization. Based on
the different stereochemical outcome it seemed likely
that a different mechanism would be operating and we
sought to explore these differences and show the scope
and limitations of trying to influence reaction outcome
based on catalyst selection.
Molecular modeling studies on the diastereomeric com-
pounds 3a and 3b indicated that 3a should be preferred
based on thermodynamic considerations.1a Although a
modest preference for this compound was seen when
employing catalyst 45 the ratio was less than anticipated,
and the non-equilibration of either isolated product iso-
mer under the reaction conditions led us to propose that
the reaction proceeded under kinetic control. In order to
bias the stereochemical outcome in favor of the desired
isomer other catalysts were screened. It was thought that
PCy3
Ru
PCy3
Cl
Cl
4
Ph
OH
O
O
O
3 steps
70% ds
N
H
Ph
N
Ph
3a
N
Ph
Ts
Ts
OCF3
2
1
Scheme 1.
Keywords: Ring-closing metathesis; Spirocycles; Intermediates.
*
Tel.: +1 732 594 0848; fax: +1 732 594 5170; e-mail: debra_wallace@
merck.com
Scheme 2.
0040-4039/$ - see front matter Ó 2004 Published byElsevier Ltd.
doi:10.1016/j.tetlet.2004.11.149

592
D. J. Wallace / Tetrahedron Letters 46(2005) 591–594
Scheme 3.
The double RCM reaction could proceed through the
intermediacyof four monocyclic intermediates 8a, 8b,
9a, and 9b, resulting from the initial formation of a five
membered ring (path A to give 8a and 8b) or a six mem-
bered ring (path B to give 9a and 9b) (Scheme 3). Under
the reaction conditions these monocycles would in turn
be expected to react further to afford the observed spiro-
cyclic products. To explain the selectivity observed in
these reactions, both the relative ratios and the subse-
quent reactions of these intermediates were of interest.
Conversion of the isolated intermediates to products
was then studied byresubjection of the purified isomers
to reaction conditions. The reaction of compounds 8a
and 8b with all four catalysts gave different results
depending on the choice of catalyst. Reaction of 8a with
Ôfirst generationÕ catalysts gave 3a as the sole product
(Scheme 4). However use of the second generation cata-
lysts with 8a gave 36% of the unexpected 3b along with
3a indicating that a competing pathwaywas operating
as well as direct cyclization. In contrast, for 8b the inver-
sion pathwaywas more significant with first generation
catalysts (30% of 3a formed), but onlyaccounted for
13% of the mixture with the second generation catalysts.
While the exact factors underlying the partitioning of
pathways for this particular step are not obvious, it
can be concluded that for formation of these six mem-
bered rings significantlydifferent results are obtained
with first versus second generation catalysts.
The reaction of alkene 2 with the Ôfirst generationÕ type
catalysts 4 and 7 gave no selectivityfor formation of
8a over 8b but led to a high preference for the formation
of 9a over 9b (ca. 90:10 ratio).9 Path A, leading to for-
mation of 8, was seen to be the major pathwayaccount-
ing for over 70% of the reaction intermediates.10 The use
of the Ôsecond generationÕ type catalysts 5 and 6 also
favored path A over path B with five membered rings
accounting for 90% of intermediates.11 As with the pre-
vious catalysts, no selectivity for formation of 8a over 8b
was seen, however in this case 9a and 9b were also
formed in equimolar amounts. Hence, for these cycliza-
tions the two diastereomeric five membered rings are
formed in the same relative ratio independent of cata-
lyst, but different relative ratios of six membered rings
were obtained when switching from first to second gen-
eration catalysts.12 The rate determining step for these
reactions is expected to be the intermolecular reaction
of the catalyst with one of the allyl groups, with the sub-
sequent intramolecular cyclization being rapid.13 As
such, the dominance of path A is due to a preferential
interaction of catalyst with the oxygen bound allyl
group, rather than a kinetic preference for formation
of five membered rings per se. However, the lack of
selectivityor catalsyt dependence in this step (as
opposed to that seen operating in path B) maybe due
to the subsequent intramolecular cyclization being
particularlyrapid for formation of five membered rings.
We propose that the inversion pathwayoccurs via a ring
closing/intramolecular ring opening mechanism (Scheme
5). For example, initial reaction of the catalyst with the
free allyl group in 8b could be followed bydirect cycliza-
tion to give 3b (retention of stereochemistry) or ring
opening in an intramolecular sense to give 10 (inversion
of stereochemistry) followed by rapid ring closing to
give 3a. Initial ring opening of the hindered internal ole-
fin in the five membered ring seemed unlikelyand com-
For all catalysts the relative ratios of the four intermedi-
ates to each other remained essentiallyconstant
throughout the course of the reaction, indicating similar
rates of further conversion for these isomers. However,
the observed ratios of 8a and 9a to 8b and 9b at this
stage of the reaction does not account for the final prod-
uct distribution, leading to the proposal that some
scrambling of the initial stereochemistrymust result dur-
ing the second ring closing step.
Scheme 4.

D. J. Wallace / Tetrahedron Letters 46(2005) 591–594
593
Reaction
14a : 14b
O
O
O
with internal
alkene ?
O
4
5
6
7
40 : 60
40 : 60
40 : 60
40 : 60
O
Ru
TsHN
Ph
TsHN
Ph
N
Ph
TsHN
Ph
N
Ph
11
14a
14b
Ts
Ts
8b
12
Reactionwith
most accessible
alkene ?
15a : 15b
OH
"
"
OH
OH
Ru
4
5
6
7
36 : 64
63 : 37
67 : 33
40 : 60
O
O
Retention
pathway
N
Ph
N
Ph
15a
N
Ph
15b
Ts
Ts
Ts
Ru
13
N
Ph
N
Ph
3b
Ts
Ts
Scheme 7.
Inversion
pathway
The major isomer from reactions with the first genera-
tion catalysts had the opposite chirality at the newly
formed stereocenter to that obtained when the hydroxyl
group was protected as the allyl ether as in 2.
O
Ph
Ru
O
Ph
N
N
Ts
10
Ts
3a
Scheme 5.
In conclusion, a detailed studyof different RCM reac-
tions in the phenyl glycine derived system has been car-
ried out. In these substrates the formation of five
membered rings was shown to be independent of the
choice of catalyst. Where this cyclization defined a ste-
reocenter (i.e., 2 to 8a and 8b, and 12 to 14a and 14b)
reactions proceed with low selectivity. In the cases where
the stereocenter has been previouslyset (i.e cyclization
of 9a and 9b) no equilibration was seen and rapid for-
mation of a single product was observed. In contrast,
the formation of six membered rings can be influenced
bycatalyst choice, with different selectivities for forma-
tion of 9a and 15a being achieved with first and second
generation catalysts. Where the stereocenter has previ-
ouslybeen set as in 8a and 8b inversion mechanisms
can operate during the formation of the six membered
rings, with different product ratios resulting from differ-
ent catalysts. These results can probably be attributed to
the kinetic ease of formation of five membered rings, in
these fast cyclizations selectivity is less likely and inver-
sion does not playa role. The formation of six mem-
bered rings is expected to be slower allowing for some
selectivityin cyclizations or for inversion mechanisms
to complete. A knowledge of the intermediates formed
during the double RCM reactions along with their fur-
ther reactions accounts for the stereochemical outcome
with all catalysts studied to date.
pound 11 would be expected to lead to other reaction
intermediates, which were not observed during these
reactions.
The formation of products from 9a and 9b was also
studied. With all four catalysts 9a afforded solely 3a,
and 9b afforded solely 3b, no equilibration or inversion
was seen, indicating that the choice of catalyst does
not affect formation of these five membered rings
(Scheme 6). Notablythe relative ratios of the four inter-
mediates combined with a knowledge of how each com-
pound is then converted to one or both products
accounts for the stereochemical outcome of each of
the reactions.
The related cyclization of compounds 12 and 13 have
also been studied and results are in line with expectation
based on the intermediates obtained from 2 (Scheme 7).
The reaction of 12—to form a five membered ring—was
independent of catalyst; in all cases a near non-selective
cyclization to give a ca. 40:60 mixture of 14a and 14b
was obtained.14 However reaction of 13 to from the
six membered rings 15a and 15b did depend on the
choice of catalyst. Although the reactions showed only
modest selectivity, a different major product was seen
when switching from first to second generation catalysts.
Acknowledgements
O
We thank Tom Novak for the high resolution mass
spectral data and Dr. Cheng-yi Chen of support of this
work.
O
4, 5, 6 or 7
N
Ph
Ph
N
Ph
9a
Ts
Ts
3a
References and notes
O
O
4, 5, 6 or 7
1. (a) Wallace, D. J.; Goodman, J. M.; Kennedy, D. J.;
Davies, A. J.; Cowden, C. J.; Ashwood, M. S.; Cottrell, I.
F.; Dolling, U.-H.; Reider, P. J. Org. Lett. 2001, 3, 671; (b)
Wallace, D. J.; Cowden, C. J.; Kennedy, D. J.; Ashwood,
M. S.; Cottrell, I. F.; Dolling, U.-H. Tetrahedron Lett.
2000, 41, 2027.
N
N
Ph
9b
Ts
Ts
3b
Scheme 6.

594
D. J. Wallace / Tetrahedron Letters 46(2005) 591–594
6. Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett.
2. For other synthetic approaches to this molecule see: (a)
Kulagowski, J. J.; Curtis, N. R.; Swain, C. J.; Williams, B.
J. Org. Lett. 2001, 3, 667; (b) Maligres, P. E.; Waters, M.
M.; Lee, J.; Reamer, R. A.; Askin, D.; Ashwood, M. S.;
Cameron, M. J. Org. Chem. 2002, 67, 1093; (c) Raab, C.
E.; Dean, D. C.; Melillo, D. G. J. Labelled Compd.
Radiopharm. 2001, 44, 815.
1999, 1, 953.
7. (a) Garber, S. B.; Kingsbury, J. S.; Gray, B. L.; Hoveyda,
A. H. J. Am. Chem. Soc. 2000, 122, 8168; (b) Gessler, S.;
Randl, S.; Blechert, S. Tetrahedron Lett. 2000, 41,
9973.
8. Kingsbury, J. S.; Harrity, J. P. A.; Bonitatebus, P. J., Jr.;
Hoveyda, A. H. J. Am. Chem. Soc. 1999, 121, 791.
9. Reactions were run in 0.1 M dichloromethane at room
temperature using 5 mol % catalyst, additional catalyst
was sometimes added.
3. For recent reviews of RCM reactions see: (a) Snapper, M.
L., Hoveyda, A. H., Eds. Tetrahedron symposium in print
1999, 55, 8141; (b) Alkene Metathesis in Organic Synthesis,
Furstner, A., Ed.; Top. Organomet. Chem. 1998, 1; (c)
Blechert, S. Pure Appl. Chem. 1999, 8, 1393; (d) Furstner,
A. Angew. Chem., Int. Ed. 2000, 39, 3012; (e) Trnka, T.
M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18; (f)
Schrock, R. E.; Hoveyda, A. H. Angew. Chem., Int. Ed.
2003, 42, 4592; (g) Walters, M. A. Prog. Heterocycl. Chem.
2003, 15, 1; (h) Deiters, A.; Martin, S. F. Chem. Rev. 2004,
104, 2199.
4. For formation of spirocyclic compounds via multiple
RCM reactions see: (a) Bassindale, M. J.; Hamley, P.;
Leitner, A.; Harrity, J. P. A. Tetrahedron Lett. 1999, 40,
3247; (b) Schmidt, B.; Westhus, M. Tetrahedron 2000, 56,
2421; (c) Schmidt, B.; Wildemann, H. J. Org. Chem. 2000,
65, 5817; (d) Schmidt, B.; Wildemann, H. J. Chem. Soc.,
Perkin Trans. 1 2000, 2916; (e) Wallace, D. J.; Bulger, P.
G.; Kennedy, D. J.; Ashwood, M. S.; Cottrell, I. F.;
Dolling, U.-H. Synlett 2001, 357; (f) Heck, M.-P.; Baylon,
C.; Nolan, S. P.; Mioskowski, C. Org. Lett. 2001, 3, 1989;
(g) Wybrow, R. A. J.; Johnson, L. A.; Auffray, B.; Moran,
W. J.; Adams, H.; Harrity, J. P. A. Tetrahedron Lett. 2002,
43, 7851; (h) Edwards, A. S.; Wybrow, R. A.; Johnson, C.;
Adams, H.; Harrity, J. P. A. Chem. Commun. 2002, 1542;
(i) Wallace, D. J. Tetrahedron Lett. 2003, 44, 2145; (j)
Wybrow, R. A. J.; Stevenson, N. G.; Harrity, J. P. A.
Synlett 2004, 140; (k) Wybrow, R. A. J.; Edwards, A. S.;
Stevenson, N. G.; Adams, H.; Johnstone, C.; Harrity, J. P.
A. Tetrahedron 2004, 60, 8869.
10. A different ratio of intermediates was observed when the
reaction was carried out using the original solvent of
chloroform.
11. Compounds 8a, 8b, and 9a have been previouslycharac-
terized, see Ref. 1a. Keydata for 9b: Rotation aD À148 (c
1
1.6, CHCl3): H (400 MHz, CDCl3) d 7.64 (d, J = 8.1 Hz,
2H), 7.19 (m, 5H), 7.10 (d, J = 8.1 Hz, 2H), 6.20–6.16 (m,
1H), 6.10–6.06 (m, 1H), 5.80 (ddt, J = 15.7, 10.5, 5.3 Hz,
1H), 5.46 (dd, J = 10.8, 17.5 Hz, 1H), 5.30 (s, 1H), 5.23
(dq, J = 17.2, 1.8 Hz, 1H), 5.21 (dd, J = 17.4, 1.0 Hz, 1H),
5.11 (dq, J = 10.4, 1.6 Hz, 1H), 5.08 (dd, J = 10.8, 1.0 Hz,
1H), 4.00 (ddd, J = 18.4, 3.5, 2.5 Hz, 1H), 3.88 (ddt,
J = 12.3, 5.2, 1.5 Hz, 1H), 3.84 (ddt, J = 12.3, 5.4, 1.5 Hz,
1H), 3.60 (dt, J = 18.4, 2.3 Hz, 1H), 2.34 (s, 3H): 13C
(100.6 MHz, CDCl3) d 142.7, 138.0, 137.3, 136.8, 135.6,
129.1, 129.0, 128.1, 127.7, 127.6, 126.2, 116.8, 115.6, 75.3,
64.2, 64.1, 41.7, 21.4: LCMS 418.1 (M = Na, 10), 338
(MH+ÀAllOH, 10), 182 (100): HRMS calcd for
C23H26NO3S (M+H) 396.1633, found 396.1635.
12. In this letter we use the term Ôfirst generation catalystÕ to
cover those laking the dihydroimidazole ligand, that is, 4
and 7 and Ôsecond generation catalystÕ to cover those with
the dihydroimidazole ligand, that is, 5 and 6.
13. (a) Ulman, M.; Grubbs, R. H. Organometallics 1998, 17,
2484; (b) Hoye, T. R.; Zhao, H. Org. Lett. 1999, 1, 1123.
14. The stereochemical outcome of this reaction was deter-
mined byconversion of the mixtures of 14a and b, or 15a
and b to the known compounds 8 and 9.
5. Schwab, P.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc.
1996, 118, 100.