Modular catalysts for diene cycloisomerization: Rapid and enantioselective variants for bicyclopropane synthesis

Article abstract of DOI:10.1021/ja064335d

Deconstructing the tridentate (triphos)Pt(II) first-generation catalysts into mixed diphosphine/monophosphine combinations (P₂P) has led to new, more active catalysts for the cycloisomerization of 1,6-, and 1,7-dienes into bicyclo-[3.1.0] and -[4.1.0] products. When the diphosphine was the small bite angle dppm, reaction rates were ～20-fold faster than with triphos, although reaction rates and diastereoselectivities were also sensitive to the monophosphine (PMe₃ being optimal for rate, PPh₃ being optimal for selectivity). When the diphosphine was xyl-BINAP or SEGPHOS, the catalysts were enantioselective, and enantio-ratios up to 98:2 were observed. Both sets of catalysts showed enhanced functional group tolerance in comparison to the original (triphos)Pt²⁺ catalyst. X-ray structures for both precatalysts are also reported.

Full text of DOI:10.1021/ja064335d

Published on Web 09/15/2006
Modular Catalysts for Diene Cycloisomerization: Rapid and
Enantioselective Variants for Bicyclopropane Synthesis
Jeremy A. Feducia, Alison N. Campbell, Michael Q. Doherty, and Michel R. Gagne´*
Contribution from the Department of Chemistry, UniVersity of North Carolina at Chapel Hill,
Chapel Hill, North Carolina 27599-3290
Received June 20, 2006; E-mail: mgagne@unc.edu
Abstract: Deconstructing the tridentate (triphos)Pt(II) first-generation catalysts into mixed diphosphine/
monophosphine combinations (P2P) has led to new, more active catalysts for the cycloisomerization of
1,6-, and 1,7-dienes into bicyclo-[3.1.0] and -[4.1.0] products. When the diphosphine was the small bite
angle dppm, reaction rates were ∼20-fold faster than with triphos, although reaction rates and diastereo-
selectivities were also sensitive to the monophosphine (PMe3 being optimal for rate, PPh3 being optimal
for selectivity). When the diphosphine was xyl-BINAP or SEGPHOS, the catalysts were enantioselective,
and enantio-ratios up to 98:2 were observed. Both sets of catalysts showed enhanced functional group
tolerance in comparison to the original (triphos)Pt2+ catalyst. X-ray structures for both precatalysts are
also reported.
Introduction
typifies diene cycloisomerization reactions, and the resulting
products are considerably less structurally diverse (diene f
The consumption of unsaturation as a means to drive the
formation of new rings is a key feature of metal-catalyzed diene
and enyne cycloisomerization reactions.1 In the latter case, the
alkyne’s heightened kinetic and thermodynamic potential can
lead to products wherein all of the unsaturation is consumed
and complex multicyclic products ensue.2,3 Despite utilizing a
diverse set of mechanisms,4 Pt, Pd, Ru, Rh, and Au enjoy special
positions as catalysts combining high activity and controllable
product selectivity,1 although efforts to render them chiral for
asymmetric catalysis have been slower to develop.5,6 In contrast
to enynes, incomplete consumption of the available unsaturation
cycloalkene is typical).4 Cyclopropanes,7 for example, are
common products of enyne cycloisomerization;2 however, the
analogous conversion of a diene into a cyclopropane was, until
recently, unknown.
(1) Reviews: (a) Trost, B. M. Acc. Chem. Res. 1990, 23, 34-42. (b) Trost, B.
M.; Krische, M. J. Synlett 1998, 1-16. (c) Widenhoefer, R. A. Acc. Chem.
Res. 2002, 35, 905-913. (d) Aubert, C.; Buisine, O.; Malacria, M. Chem.
ReV. 2002, 102, 813-834. (e) Echavarren, A. M.; Nevado, C. Chem. Soc.
ReV. 2004, 33, 431-436. (f) Diver, S. T.; Giessert, A. J. Chem. ReV. 2004,
104, 1317-1382. (g) Ma, S.; Yu, S.; Gu, Z. Angew. Chem., Int. Ed. 2006,
45, 200-203.
(2) Mini-review: Bruneau, C. Angew. Chem., Int. Ed. 2005, 44, 2328-2334.
(3) See, for example: (a) Harrak, Y.; Blaszykowski, C.; Bernard, M.; Cariou,
K.; Mainetti, E.; Mouries, V.; Dhimane, A.-L.; Fensterbank, L.; Malacria,
M. J. Am. Chem. Soc. 2004, 126, 8656-8657. (b) Mamane, V.; Gress, T.;
Krause, H.; Fu¨rstner, A. J. Am. Chem. Soc. 2004, 126, 8654-8655. (c)
Fu¨rstner, A.; Hannen, P. Chem. Commun. 2004, 2546-2547. (d) Nieto-
Oberhuber, C.; Munoz, M. P.; Bunuel, E.; Nevado, C.; Ca´rdenas, D. J.;
Echavarren, A. M. Angew. Chem., Int. Ed. 2004, 43, 2402-2406. (e)
Luzung, M. R.; Markham, J. P.; Toste, F. D. J. Am. Chem. Soc. 2004, 126,
10858-10859.
(4) For reviews focusing on the mechanisms of cycloisomerization, see: (a)
Lloyd-Jones, G. C. Org. Biomol. Chem. 2003, 1, 215-236. (b) Me´ndez,
M.; Mamane, V.; Fu¨rstner, A. Chemtracts 2003, 16, 397-425.
(5) Mini-review: Fairlamb, I. J. S. Angew. Chem., Int. Ed. 2004, 43, 1048-
1052.
(6) (a) Corkey, B. K.; Toste, F. D. J. Am. Chem. Soc. 2005, 127, 17168-
17169. (b) Mun˜oz, M. P.; Adrio, J.; Carretero, J. C.; Echavarren, A. M.
Organometallics 2004, 24, 1293-1300. (c) Charruault, L.; Michelet, V.;
Taras, R.; Gladiali, S.; Geneˆt, J.-P. Chem. Commun. 2004, 850-851. (d)
Liu, C.; Han, X.; Wang, X.; Widenhoefer, R. A. J. Am. Chem. Soc. 2004,
126, 3700-3701. (e) Lei, A.; He, M.; Zhang, X. J. Am. Chem. Soc. 2003,
125, 11472-11473. (f) Mikami, K.; Kataoka, S.; Aikawa, K. Org. Lett.
2005, 7, 5777-5780. (g) Tong, X.; Li, D.; Zhang, Z.; Zhang, X. J. Am.
Chem. Soc. 2004, 126, 7601-7607.
Contrasting most diene cycloisomerization reactions are
electrophilic (triphos)Pt2+ catalysts reported by our group for
the intramolecular conversion of 1,6-, and 1,7-dienes into
bicyclopropanes (e.g., eqs 1 and 2),8 and by Vitagliano’s group
for the intermolecular coupling of two alkenes into a cyclopro-
pane.9 In both cases, two degrees of unsaturation are consumed
to make bicyclic and monocyclic products, respectively.10
Experimental evidence from both groups implicate ionic
mechanisms and the intermediacy of carbocations.11,12 A typical
mechanistic scheme for the reaction in eq 1 is outlined in
Scheme 1, with the key features being the electrophilic activation
of the terminal alkene, the generation of a 3° carbocation, and
a 1,2-hydride shift to establish the R,γ-arrangement of metal
(7) Lebel, H.; Marcoux, J.-F.; Molinaro, C.; Charette, A. B. Chem. ReV. 2003,
103, 977-1050.
(8) (a) Kerber, W. D.; Gagne´, M. R. Org. Lett. 2005, 7, 3379-3381. (b) Kerber,
W. D.; Koh, J. H.; Gagne´, M. R. Org. Lett. 2004, 6, 3013-3015.
(9) Cucciolito, M. E.; D’Amora, A.; Vitagliano, A. Organometallics 2005, 24,
3359-3361.
9
13290
J. AM. CHEM. SOC. 2006, 128, 13290-13297
10.1021/ja064335d CCC: $33.50 © 2006 American Chemical Society

Modular Catalysts for Diene Cycloisomerization
A R T I C L E S
Scheme 1
Scheme 2
generate terpene-like cycloisomerization products.19,20 For ex-
ample, the [3.1.0]-bicyclic core of 4 (R-thujane) constitutes the
carbon skeleton of a large number of monoterpenoids,18b while
the [4.1.0] fragment of 2 is found in thousands of steroid-like
natural products.
The principle that guided the discovery of the (triphos)Pt2+
catalysts was the utilization of ligands that would block the sites
cis to a putative intermediate alkyl (e.g., A in Scheme 1), and
thereby inhibit â-H elimination as a competing decomposition
pathway. This approach, demonstrated by Hahn and Vitagliano
in numerous electrophilic activation systems,21 ultimately led
to the discovery of the catalysts described above. Efforts to
improve the reactivity and selectivity of this first generation
catalyst, however, were hampered by the difficulty of synthesiz-
ing derivatives of the tridentate ligands22 (triphos itself is
commercially available), especially chiral analogues for devel-
oping asymmetric versions of the reactions.
and cation needed for ring-closing cyclopropanation.13 Several
points deserve comment: (1) Square planar Pt(II) complexes
have a strong preference for coordinating/activating the less
substituted alkene,14 in contrast to soft Lewis acids like Hg(II)
or Ag(I).15 This reactivity pattern provides a predictable location
for initiating electrophilic reactions in a poly ene and comple-
ments traditional strategies for initiating cation-olefin reactions,
which rely on reagents like H+,16 Br+, and RSe+ to activate
the more basic (more highly substituted) alkene.17 (2) The 1,2-
migration shown in red was confirmed by D-labeling experi-
ments.8a (3) The cyclopropanation step proceeds via the addition
of a comparatively nucleophilic Pt-C bond to the empty orbital
of the γ-carbocation, we presume in direct analogy to the double
inversion (W-conformation) processes that have been observed
for Sn, Fe, and Ti. The latter three have been shown to be
rigorously stereospecific and proceed via double inversion
stereochemistries.13
To circumvent this impediment, we reasoned that the triden-
tate architecture could be deconstructed into a modular com-
bination of bi- and monodentate phosphine ligands (Scheme 2).
If successful, this approach would enable the effect of ligand
bite angle, cone angle, basicity, etc., to be independently
assessed under catalytic conditions. One potential pitfall of this
approach, however, was the possibility that monophosphine
dissociation at the stage of an intermediate alkyl (e.g., A in
Scheme 1) could provide a pathway for â-H elimination and
catalyst deactivation. The electrophilic nature of the P3PtR+
along with the near rigorous need for associative ligand
substitution at a substitution inert Pt(II)23 were each considered
variables favoring a successful outcome.
Because the individual steps (obviously excluding the Pt) are
so similar to the reactions involved in terpene biosynthesis
(cation generation, cation-olefin cyclization, hydride shifts,
etc.),18 it is not surprising that terpene-like poly ene substrates
(10) For additional examples of Pt(II)-catalyzed alkene activation reactions,
see: (a) Fu¨rstner, A.; A¨ıssa, C. J. Am. Chem. Soc. 2006, 128, 6306-6307.
(b) Qian, H.; Han, X.; Widenhoefer, R. A. J. Am. Chem. Soc. 2004, 126,
9536-9537. (c) Liu, C.; Han, X.; Wang, X.; Widenhoefer, R. A. J. Am.
Chem. Soc. 2004, 126, 3700-3701. (d) Bender, C. F.; Widenhoefer, R. A.
J. Am. Chem. Soc. 2005, 127, 1070-1071. (e) Karshtedt, D.; Bell, A. T.;
Tilley, T. D. J. Am. Chem. Soc. 2005, 127, 12640-12646. (f) Karshtedt,
D.; Bell, A. T.; Tilley, T. D. Organometallics 2004, 23, 4169-4171. (g)
Hahn, C.; Cucciolito, M. E.; Vitagliano, A. J. Am. Chem. Soc. 2002, 124,
9038-9039.
With regards to the catalytic asymmetric cycloisomerization
of dienes, several catalysts are known, in particular the chiral
Ni-based catalysts of Leitner, which give up to 80% ee for the
cycloisomerization of diethyl diallyl malonate (eq 3), and the
L2Pd2+ catalysts of Heumann, which give ee’s up to 60% for
L2 ) bisoxazoline and sparteine (eq 4).24,25 Because no
enantioselective catalysts for the conversion of dienes into
bicyclopropanes have been reported, the design and synthesis
of chiral metal catalysts for this transformation represent both
an obvious and desirable direction of endeavor, the modular
(11) For early work on metal-induced carbenium ion formation, see: Chisolm,
M. H.; Clark, H. C. Acc. Chem. Res. 1973, 6, 202-209.
(12) For a review of electrophilic approaches to cyclopropanes, see: Taylor, R.
E.; Engelhardt, F. C.; Schmitt, M. J. Tetrahedron 2003, 59, 5623-5634.
(13) For related reactions involving Sn, Fe, and Ti, see: (a) Davis, D. D.;
Johnson, H. T. J. Am. Chem. Soc. 1974, 96, 7576. (b) Fleming, I.; Urch,
C. J. Tetrahedron Lett. 1983, 24, 4591. (c) McWilliam, D. C.; Balasubra-
manian, T. R.; Kuivila, H. G. J. Am. Chem. Soc. 1978, 100, 6407. (d)
Lambert, J. B.; Salvador, L. A.; So, J. H. Organometallics 1993, 12, 697.
(e) Casey, C. P.; Smith Vosejpka, L. J. Organometallics 1992, 11, 738. (f)
Brookhart, M.; Liu, Y. J. Am. Chem. Soc. 1991, 939. (g) Casey, C. P.;
Strotman, N. A. J. Am. Chem. Soc. 2004, 126, 1699 and references therein.
(14) Hegedus, L. S. Transition Metals in the Synthesis of Complex Organic
Molecules; University Science Books: Mill Valley, CA, 1994; pp 199-
236.
(19) (a) Croteau, R. Chem. ReV. 1987, 87, 929-954. (b) Croteau, R. Recent
DeVelopments in FlaVor and Fragrance Chemistry: Proceedings of the
3rd International Harmann & Reimer Symposium; VCH: Weinheim, 1993;
pp 263-273.
(20) For a discussion explicitly comparing Pt- and Au-catalyzed enyne cyclo-
isomerization to terpene biosynthesis, see: Fu¨rstner, A.; Hannen, P. Chem.-
Eur. J. 2006, 12, 3006-3019.
(15) (a) Hegedus, L. S. In ComprehensiVe Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: Elmsford, NY, 1991; Vol. 4, pp 551-569. (b) Bartlett,
P. A. In Asymmetric Synthesis; Morrison, J. D., Ed.; Academic Press: New
York, 1984; Vol. 3, pp 411-454.
(16) (a) Ishibani, H.; Ishihara, K.; Yamamoto, H. J. Am. Chem. Soc. 2004, 126,
11122-11123. (b) Uyanik, M.; Ishihara, K.; Yamamoto, H. Bioorg. Med.
Chem. 2005, 13, 5055-5065 and references therein.
(17) (a) Wendt, K. U.; Schultz, G. E.; Corey, E. J.; Liu, D. R. Angew. Chem.,
Int. Ed. 2000, 39, 2812-2833. (b) Sutherland, J. K. In ComprehensiVe
Organic Synthesis; Trost, B. M., Ed.; Pergamon Press: Elmsford, NY, 1991;
Vol. 1, pp 341-377. (c) Bartlett, P. A. In Asymmetric Synthesis; Morrison,
J. D., Ed.; Academic Press: New York, 1984; Vol. 3, pp 341-409.
(18) Biosynthesis of Isoprenoid Compounds; Porter, J. W., Spurgeon, S. L., Eds.;
John Wiley & Sons: New York, 1981; Vol. 1.
(21) (a) Hahn, C. Chem.-Eur. J. 2004, 10, 5888-5899. (b) Hahn, C.; Morvillo,
P.; Herdtweck, E.; Vitagliano, A. Organometallics 2002, 21, 1807-1818.
(c) Hahn, C.; Morvillo, P.; Vitagliano, A. Eur. J. Inorg. Chem. 2001, 419-
429.
(22) For typical procedures, see: DuBois, D. L.; Miedaner, A.; Haltiwanger,
R. C. J. Am. Chem. Soc. 1991, 113, 8753-8764.
(23) Cross, R. J. AdV. Inorg. Chem. 1989, 34, 219-292.
(24) (a) Bo¨ing, C.; Francio, G.; Leitner, W. Chem. Commun. 2005, 1456-1458.
(b) Bo¨ing, C.; Francio, G.; Leitner, W. AdV. Synth. Catal. 2005, 347, 1537-
1541. (c) Heumann, A.; Moukhliss, M. Synlett 1998, 1211-1212.
(25) For recent examples of chiral Pd(II), Pt(II), and Au(I) catalysts in related
electrophilic activation processes, see ref 6.
9
J. AM. CHEM. SOC. VOL. 128, NO. 40, 2006 13291

A R T I C L E S
Feducia et al.
Table 1. The Effect of Bidentate Ligand on the Cycloisomerization
of 1 (Eq 1) Using [P2(PMePh2)Pt][BF4]2
catalyst design described herein being highly amenable to the
search for selective catalysts.
a
b
P2
(
)
t (h)
dr
yieldc
dppm (72°)
dppbz (83°)
dppe (85°)
dppp (91°)
dppb (98°)
1.5
5
6
15
7
26:1
12:1
19:1
5:1
79%
25%
56%
45%
16%
7:1
a Reaction conditions: 5% (P2)PtI2 (0.06 M in CD3NO2), 5% PMePh2,
11% AgBF4, 23 °C. b See ref 28b. c By GC.
tently provided the best results; nitroethane was usually an
acceptable replacement (although it is a bit slower), while CH2-
Cl2 was considerably slower.
We report herein two parallel lines of research that have led
to new catalysts for the diastereo- and enantioselective cyclo-
isomerization of 1,6-, and 1,7-dienes into [3.1.0]- and [4.1.0]-
bicyclic products. One catalyst class utilizes a bidentate ligand
with a small bite angle to achieve activities that are at least
20-fold higher than first-generation catalysts along with a higher
functional group compatibility profile and a broader substrate
scope, while the second class utilizes chiral bidentate ligands
to achieve reactions with enantioselectivities reaching 98:2 er.
Because bite angle is often an important determinant of
bidentate ligand activity,28 a systematic screen of this variable
was carried out by varying the length of the carbon spacer.
PMePh2 was chosen as the monodentate phosphine for these
experiments because of its close structural and electronic
resemblance to the triphos deconstruction product (Scheme 2).
As shown in Table 1 for the reaction in eq 1, both the dr and
the rate were sensitive to changes in the bidentate ligand size
and structure. The dppm ligand, with its single methylene bridge
and small bite angle (72°), proved to be the most active,
selective, and high-yielding of the diphosphines.
In addition to the significant observation that cyclopropane
2 was indeed obtained with these modified catalysts, no
dehydrogenated products were detected by GC/MS, indicating
that the feared possibility of PR3 loss and â-H elimination from
an intermediate Pt-R+ was not competitive. In comparison to
the original triphos catalyst, which required elevated tempera-
tures (40 °C) and longer reaction times (10+ h), this catalyst is
much improved, and the yields reflect this.8a
An additional optimization cycle was achieved by testing the
monophosphine’s effect on rate and selectivity with (dppm)-
PtI2. Although there was considerable scatter in the data, Table
2 clearly showed the reaction to be incompatible with mono-
dentate phosphines that were either too large or too electron
deficient. PMePh2 provided the best combination of rate (time
to 100% conversion) and diastereoselectivity (1.5 h, 26:1 dr,
79%), PMe3 (3 h, 12:1 dr, 94%) proved optimal with respect to
yield, and PPh3 (0.8 h, 38:1 dr, 64%) was the most diastereo-
selective. The catalysts with the more bulky monodentate
ligands, however, proved to be a bit capricious and not as
general as those generated from the smaller, more electron-rich
catalysts (PMe3 and PMe2Ph). Because the yield of cyclopropane
was significantly higher for the PMe3 catalyst,29 this catalyst
was chosen for an evaluation of substrate scope (Table 3).
Results and Discussion
Catalyst Activation. In the first-generation catalysts, activa-
tion was most conveniently achieved by protonolysis of the
[(triphos)Pt-Me][BF4] precursor with HNTf2 to give CH4 and
the desired [(triphos)Pt-L][NTf2][BF4] catalyst,8a where L could
be either a weak placeholder ligand like acetone or the terminal
alkene of the substrate itself. This methodology proved to be
key to cleanly generating the active catalyst in a form that
maximized activity while minimizing the production of com-
pounds that tended to either consume the cyclopropane product
or lead to Brønsted acid-derived byproducts. The analogous
activation protocol with the modular [(P2P)PtMe][BF4] precur-
sors was entirely unsuccessful, which ultimately led to a wholly
different direction of research.26 These studies showed that, for
reasons attributable to torsional strain in the triphos coordination
geometry, protonolysis was at least 50 000 times more rapid
for [(triphos)Pt-Me][BF4] than for a large number of [(P2P)-
PtMe][BF4] precursors. Fortunately, however, conditions for
achieving iodide abstraction from intermediate [(P2P)Pt-I][I]
compounds with Ag+ salts could be employed. Iodides were
significantly easier to activate than were the analogous chloride
complexes.
Typical activation procedures for in situ screening of ligand
variants began with isolated P2PtI2 complexes, to which was
added 1 equiv of the monodentate ligand in nitromethane27
(room temperature), followed by activation with 2.2 equiv of
AgBF4 in the presence of substrate (eq 5). Attempts to activate
in the absence of substrate were typically unsuccessful as iodide
associative displacement (or more likely Pt-I‚‚‚Ag) required a
ligand that was better than the nitromethane solvent. Although
a number of solvents were investigated, nitromethane consis-
(28) (a) Freixa, Z.; van Leeuwen, P. W. N. M. J. Chem. Soc., Dalton Trans.
2003, 1890. (b) van Leeuwen, P. W. N. M.; Kamer, P. C. J.; Reek, J. N.
H.; Dierkes, P. Chem. ReV. 2000, 100, 2741. (c) Raebiger, J. W.; Miedaner,
A.; Curtis, C. J.; Miller, S. M.; Anderson, O. P.; DuBois, D. L. J. Am.
Chem. Soc. 2004, 126, 5502.
(26) Feducia, J. A.; Campbell, A. N.; Anthis, J. W.; Gagne´, M. R. Organome-
tallics 2006, 25, 3114-3117.
(27) For exploratory reactions, CD3NO2 was utilized because traces of propi-
onitrile present in reagent grade CH3NO2 poison the catalyst. For preparative
work, nitromethane that had been twice precipitated from a 50:50 solution
with Et2O at -78 °C was sufficiently pure for use, see: Parrett, F. W.;
Sun, M. S. J. Chem. Educ. 1977, 54, 448-449.
(29) The balance of material in these runs is usually a mixture of many
isomerized (presumably cycloisomerized), alkene-containing products that
have proven difficult to separate and identify. The cyclopropanes were most
conveniently separated from the remaining, more highly retained material
using Ag+-impregnated silica gel.
9
13292 J. AM. CHEM. SOC. VOL. 128, NO. 40, 2006

Modular Catalysts for Diene Cycloisomerization
A R T I C L E S
Table 2. Conversion of 1 to 2 (Eq 1) for Various
cycloisomerization pathways existed (entries 5 and 6, vide infra),
they showed an enhanced propensity to generate the cyclopro-
pane product. As demonstrated by entries 1 and 4, the bicyclo-
[4.1.0] structures can be accessed from both 1,6- and 1,7-dienes,
although in the latter case the rates are slightly diminished. Some
enhancements in the functional group compatibility were also
observed (entries 5-9); the original triphos catalyst was unable
to cyclize sulfone 11 and gave poor yields of acetal product
14. The bicyclo-[3.1.0] ketone 16, which is applicable to the
synthesis of a number of thujone and thujanol derivatives,19 is
much slower than the other substrates, a phenomenon that can
likely be traced to the catalyst resting as a nonproductive ketone
adduct instead of an alkene adduct (31P NMR).30 The limit of
activity with the PMe3 catalyst appears to be represented by 17
(entry 9), which formally requires the intermediacy of a
secondary carbocation in the proposed mechanism (vide infra).
(dppm)Pt(PR3)-Dications (5 mol %)a
PR3
t (h)
dr
GC yieldb
P(OMe)3
PMe3
PMe2Ph
PEt3
P(2-furyl)3
PMePh2
PPh3
P(4-OMePh)3
P(C6F5)3
P(NMe2)3
1
3
1
5:1
12:1
17:1
20:1
23:1
26:1
38:1
15:1
24:1
78:1
74%
94%
83%
87%
43%
79%
64%
81%
53%
10%
3
3.5
1.5
0.8
0.5
3
48
a Reaction conditions: 5% (dppm)PtI2, 5% PR3, 11% AgBF4, in CD3NO2
(0.06 M in catalyst), 23 °C. b Yield obtained by calibrated GC analysis at
100% conversion.
Table 3. Cycloisomerization of Dienes with 5 mol %
a
[(dppm)Pt(PMe3)][BF4]2
The sulfonamide 19, however, was not compatible with the
catalyst, and immediate decomposition to a myriad of products
occurs, in contrast to chiral catalysts, which efficiently process
this reactant (vide infra). A class of substrates that surprisingly
did not react were the cinnamyl malonates (entry 11). Based
on cation stabilities,31 we expected that the cinnamyl case (21a)
would provide a benzyl cation of stability similar to that of a
tertiary cation, and the trimethoxy cinnamyl (21b) to an even
more stabilized cation. Nevertheless, no reaction was observed
for either substrate, despite seemingly generating cations that
would be more stable than the 2° cation proposed for 9.
It is worth noting that, although Pt(II)-salts (e.g., PtCl2) are
known to react with cyclopropanes,32 we find that the present
electron-deficient complexes are rather inert to cyclopropanes.
If traces of acid are present in the catalyst, however, cyclopro-
pane consumption can be a significant side reaction.33 The
activation protocol utilized herein minimizes the generation of
trace acids and combined with the high cyclopropanation
tendencies of the PMe3-based catalyst provides good yields of
product.
Although a direct relationship between the structure of the
(dppm)(PMe3)Pt2+ catalyst and its superior cyclopropanation
tendencies is not obvious, what is clear from the X-ray structure
of the catalyst precursor [(dppm)(PMe3)PtI][I] (22) is the
reduced bite angle of the dppm ligand (Figure 1).34 The P3-
Pt-P2 bond angle of 72° results in an obvious distortion of the
square plane such that the bond angles around the I (where
alkene coordination occurs) are expanded, which reduces the
(30) Because of extensive splitting and the number of species in solution, in
situ monitoring is uninformative (31P NMR); however, in the triphos
catalysts a resonance for the trans phosphorus is observed at a position
similar to the previously described acetone adduct (δ 79.6 ppm, JPt-P
)
3420 Hz). No evidence for an η2-alkene adduct is observed (see ref 9).
(31) Mayr, H.; Kempf, B.; Ofial, A. R. Acc. Chem. Res. 2003, 36, 66-77.
(32) Puddephatt, R. J. Coord. Chem. ReV. 1980, 33, 149-194.
(33) In several instances, we have identified the isomerization products, and
found them to include the following:
a Reaction conditions: 5% (dppm)PtI2, 5% PMe3, 11% AgBF4, in MeNO2
(0.06 M in catalyst), 23 °C. b Isolated yield; in parentheses is the GC yield
prior to isolation. c By GC. d 10% catalyst loading, 40 °C.
In all cases, the reactions were found to be 10-20 times faster
than the original (triphos)Pt2+ catalyst, they could be carried
out at room temperature instead of 40 °C, and, where different
(34) A suitable crystal was obtained from a solution of CH2Cl2 and pentane at
-26 °C.
9
J. AM. CHEM. SOC. VOL. 128, NO. 40, 2006 13293

A R T I C L E S
Feducia et al.
Scheme 4. Possible Mechanism for Cyclohexene Formation with
4,4-Disubstituted Substrates
bissulfone-derived dienes, 9 and 11, respectively (Scheme 4).
In both cases, a cyclohexene product was obtained in a ratio
that depended on the catalyst, although the selectivity for
cyclopropane using [dppm(PMe3)Pt][BF4]2 was higher than that
obtained with the triphos catalyst (4-5:1 vs ∼3:1). The
mechanism that we presume to be operative follows the
sequence previously proposed by Vitagliano for intermolecular
coupling of ethylene with tri- and tetrasubstituted alkenes.10g
In this case, a sequence of 1,2-hydride shifts migrate the cation
to the position â to the Pt, which can be considered a slipped
form of an η2-alkene complex. Attempts to reoptimize the
catalyst by independently varying the mono- and bidentate
ligands to favor this cycloisomerization pathway were partially
successful, with [(dppe)Pt(PPh3)][BF4]2 being able to invert the
major product of catalysis to the cyclohexene in a 5.4:1 ratio
(eq 6); yields remained low. Unfortunately, attempts to promote
cyclohexene formation with this catalyst on several representa-
tive substrates were unsuccessful, and only traces (<5%) of
cyclohexene products were present, if at all. The role of the
ester and sulfone functional groups in shifting the reaction
pathway is not known, although we note that in situ monitoring
of the catalytic reactions shows that at least some of the catalyst
rests at a coordinated ester/sulfone stage in addition to coordi-
nated alkene.
Figure 1. Chem3D representation of the X-ray structure of the cation of
22. Selected bond lengths (Å): Pt-P1 ) 2.3172(13), Pt-P2 ) 2.2410(13),
Pt-P3 ) 2.3242(12), Pt-I ) 2.6329(4). Selected bond angles (deg): P1-
Pt-P2 ) 99.16(5), P2-Pt-P3 ) 72.13(5), P1-Pt-I ) 90.98(4), P3-Pt-I
) 97.73(3).
Scheme 3. Possible Cyclization Pathways to the Thujane Ring
Skeleton
steric congestion around the site of activation. The structural
parameters are otherwise similar to previously reported
structures.28b
Previous mechanistic studies on the triphos-based catalyst
have demonstrated that alkene activation and cation generation
are rapid and reversible on the time scale of cyclopropanation.8a
For example, studies have shown that several trappable cations
are in equilibrium prior to hydride transfer (Scheme 3). Although
the choice of pathway(s) (initial 6-endo or 5-exo cyclization)
to the products is not known,35,36 it seems likely that the dppm
ligand plays a subtle role in enhancing the rate of the reaction,
perhaps by somehow facilitating a rate-determining hydride shift
followed by a fast cyclopropanation step or via a scenario where
the hydride shift is rapid and it is the C-C bond forming
cyclopropanation step that is somehow facilitated.
Regardless of the source of the selectivity for cyclopropanes
over other cycloisomerization products, the (dppm)(PMe3)Pt2+
catalyst is the most efficient, most functional group tolerant
catalyst yet discovered for the cycloisomerization of 1,6- and
1,7-dienes into bicyclopropanes.
Chiral Catalysts. The test reaction for our search for
enantioselective variants of the (P2P)Pt2+ catalysts was the
conversion of 1 to the [4.1.0] bicycloheptane 2, proceeding via
the putative mechanism in Scheme 1. As before, the catalysts
were generated by first adding the monodentate phosphine to
the chiral P2PtI2 to generate the (P2P)Pt-I+ precatalyst (30 min),
followed by activation with 2.5 equiv of AgBF4 in the presence
of substrate. This protocol enabled an efficient screening of
various combinations of chiral bi- and achiral monodentate
ligands.
Early in the study, it was noted that most chiral bidentate
ligands (e.g., BINAP) required a small monodentate ligand (i.e.,
PMe3) to efficiently generate the requisite P2P-Pt2+ catalyst
structure. As shown in Table 4, yields of 2 were good to
excellent, although the diastereoselectivities were modest. The
enantioselectivities varied considerably with diphosphine struc-
ture, and optimal selectivities were observed with 3,5-xyl-
The only substrates that generated significant quantities of a
second cycloisomerization product were the malonate- and
(35) The pathways differ only in the timing of ring construction, cyclohexyl
followed by pinching to the 5-3 in the 6-endo pathway, or beginning with
the cyclopentane (5-exo path) and annulating the cyclopropane in step 2.
(36) The biosynthetic pathway to the thujanes goes via a conceptually related
6-endo pathway; see ref 19b.
9
13294 J. AM. CHEM. SOC. VOL. 128, NO. 40, 2006

Modular Catalysts for Diene Cycloisomerization
A R T I C L E S
Table 4. The Effect of Diphosphine on the Cycloisomerization of 1
Table 5. Cycloisomerization Reactions with R-25 and R-26
a
(Eq 1) for in situ Generated [(P2)(PMe3)Pt2+][BF4]2
(5 mol %)a
P2
% 2b
drb
% ee
3,5-xyl-BINAP
Tol-BINAP
BINAP
92%
60%
86%
84%
77%
83%
69%
36%
77%
34%
3.5:1
3.3:1
3.5:1
2.7:1
5:1
3.6:1
3.8:1
6.4:1
16:1
9.7:1
91%
63%
62%
75%
57%
53%
55%
11%
72%
0%
SEGPHOS
C1-TUNEPHOS
C2-TUNEPHOS
C3-TUNEPHOS
CHIRAPHOS
QUINAP
BINAP(S)
a Reaction conditions: 10% (P2)PtI2 (0.02 M in CD3NO2), 10% PMe3,
25% AgBF4, 23 °C. b By calibrated GC analysis.
Chart 1. Optimum Catalysts for Diene Cycloisomerization
a Reaction conditions: 5% (P2)PtI2, 5% PMe3, 12.5% AgBF4, in MeNO2
(0.05 M in catalyst), 23 °C. b The absolute stereochemistries of entries 1-5
are in analogy to entry 6 (vide infra). c Isolated yields of analytically pure
product; yields in parentheses are calibrated GC yields. d Enantioselectivities
were obtained by chiral GC (â-cyclosil column). e 0.02 M in catalyst. f By
GC. g 10% catalyst loading. h 40 °C.
BINAP (xyl-BINAP). Interestingly, the series of C1-, C2-, and
C3-TUNEPHOS ligands indicated that the bite angle37 was not
a significant factor, in contrast to our expectations based on
the achiral work (vide supra). A moderately enantio- but highly
diastereoselective catalyst was found with the P,N ligand
QUINAP;38 however, the P,S ligand BINAP(S)39 was poorly
behaved.40,41
The catalysts in Chart 1 proved to be the most useful of the
tested combinations generating 2 with ee’s of 91% and 75%,
and with GC yields of 92% and 84%, respectively. The reactions
were found to be sensitive to solvent (yield and % ee for the
major isomer increased from CH2Cl2 (38%, 91% ee), to EtNO2
(77%, 91% ee), to MeNO2 (85%, 88% ee)) and concentration.42
Under the optimized conditions, these catalysts were highly
chemoselective and could be run with catalyst loadings as low
as 2 mol %, although our general protocol utilized 5 mol %
loadings.
byproducts were identified and shown to result from further
reactions on 2,33 but in the other entries they were present in
trace quantities, and not identified. The susceptibility of 2 to
additional reactions could be partially compensated for by
concentration and catalyst loading variations; lower concentra-
tions (0.02 M) provided improved yields (82%) and selectivities
(95% ee for the major), as did higher catalyst loadings (10 mol
%, 0.02 M, 92%).
In general, the reaction rates are slightly slower than those
with the dppm/PMe3 catalyst, although they are still quicker
than those with the triphos-based catalyst. As reflected in the
yields, these chiral catalysts are also more prone to generating
non-cyclopropane byproducts than is the dppm/PMe3 catalyst.
With a competent and robust catalyst in hand, we turned our
attention to an evaluation of scope. As shown in Table 5, a
number of 1,6- and 1,7-dienes could be converted to the
expected [3.1.0] and [4.1.0] products while maintaining a high
level of enantioselectivity. In most cases, R-25 was best,
although R-26 did reproducibly surpass it for the aza-bicycle
in entry 6. This latter entry is noteworthy as the dppm/PMe3
catalyst completely destroyed this substrate within 10 min.
Yields for the isolated products are typically fair to good, as
higher levels of alkene byproducts are obtained relative to the
dppm/PMe3 catalyst. For the reaction of 1, several of these
As part of a sequence of experiments to identify the role of
possible catalyst impurities, a cycloisomerization reaction on 1
was carried out in the absence of PMe3. Activating (P2)PtI2 with
2.5 equiv of AgBF4 in the presence of 10 equiv of 1 led to a
catalyst that provided enantiomeric 2 (ent-2)43 in moderate yields
and enantioselectivity. The absence of PMe3 thus inverts the
sense of selectivity, leading to a situation where both enanti-
omers of product can be obtained from a single enantiomer of
BINAP. This same trend was observed when the chiral
diphosphine was modified (Table 6), although none of these
catalysts proved as enantioselective or high yielding as the P2P
(37) Zhang, Z.; Qian, H.; Longmire, J.; Zhang, X. J. Org. Chem. 2000, 65,
6223-6226.
(38) In situ analysis (31P NMR) suggests that the major species has the PMe3
coordinated cis to the P-ligand of QUINAP (JP-P is 20 Hz, cf. ∼200 Hz
for a trans coupling), that is, a PPN ligand array.
(39) Faller, J. W.; Wilt, J. C.; Parr, J. Org. Lett. 2004, 6, 1301-1304.
(40) The PHOX ligands give a stable (PN)PtI2 complex; however, PR3 addition
displaces the nitrogen from the metal. Catalysts derived from this ligand
do not form cyclopropane products.
(41) The Josiphos and Walphos ligands in the Solvias ligand kit were also
screened and found to not provide any cyclopropane products.
(42) Yields at higher concentrations were compromised by the formation of
tetrasubstituted alkene (see ref 33). This was especially problematic for 2.
(43) Because the absolute stereochemistry of 2 was assigned by analogy to 20,
we caution against any overinterpretation of the absolute assignment.
9
J. AM. CHEM. SOC. VOL. 128, NO. 40, 2006 13295

A R T I C L E S
Feducia et al.
Table 6. The Effect of Diphosphine on the Cycloisomerization of 1
a
(Eq 7) for in situ Generated [(P2)Pt2+][BF4]2
P2
% 2b
drb
eec
xyl-BINAP
BINAP
SEGPHOS
C1-TUNEPHOS
CHIRAPHOS
QUINAP
50%
64%
54%
55%
79%
47%
52%
12:1
15:1
21:1
15:1
27:1
17:1
15:1
25%
58%
65%
56%
39%
59%d
0%
BINAP(S)
a Reaction conditions: 10 mol % (P2)PtI2 (0.02 M in CD3NO2), 25%
AgBF4, 23 °C. b By calibrated GC analysis. c Opposite absolute stereo-
chemistry than (P2P)Pt2+ catalysts. d Same absolute stereochemistry as
(P2P)Pt2+ catalyst.
analogues (yet they were significantly more diastereoselective).
Interestingly enough, the QUINAP catalyst did not change its
sense of enantioselectivity.
Attempts to improve the (P2)Pt+2 catalyst through counterion
Figure 2. Chem3D representation of the X-ray structure of the R-28 cation.
Selected bond lengths (Å): Pt-P1 ) 2.3321(12), Pt-P2 ) 2.2653(12), Pt-
P3 ) 2.3450(12), Pt-I ) 2.6469(4). Selected bond angles (deg): P1-Pt-
P2 ) 91.66(4), P2-Pt-P3 ) 98.04(5), I-Pt-P1 ) 88.74(3), I-Pt-P3 )
86.91(3).
modifications (SbF6 or OTf-) led to species that rapidly
-
decomposed the transiently observed cyclopropanes. When
applied to a different diene arrangement, the reactivity of these
catalysts also did not translate and multiple isomeric products
were obtained in the cycloisomerization of 3 (GC/MS). In the
end, the chiral P2Pt2+ catalysts do enantioselectively provide
Scheme 5. Double Stereodifferentiating Reactions on
â-Citronellene (3)
-
ent-2, but only with 1 and only as the BF4 salt. We suspect
that this is a consequence of their overly electrophilic nature.
Perhaps the most surprising aspect of these results was that
no products corresponding to loss of H2 from a putative â-H
elimination pathway were observed in the GC/MS of the
reactions. The original guiding notion that the cis coordination
sites need to be rigorously blocked to prevent â-H elimination
may need to be revisited, especially for Pt(II) where the data
indicate that â-H elimination does not compete with cyclo-
isomerization.
The absolute stereochemistry of the bicyclopropane products
was ascertained directly for the aza-bicycle in entry 6 of Table
5. As shown in eq 8, the product was detosylated with Mg in
MeOH44 to yield the volatile free base, which was directly
converted to the mono-oxalate salt and analyzed by optical
rotation. The sign of the rotation matched that of the previously
reported compound,45 and established that (R)-xyl-BINAP and
(R)-SEGPHOS generated the 1R,5R-azabicyclohexane (20). The
remainder of the [3.1.0] bicyclic structures in Table 5 were
assigned by analogy, as was the [4.1.0] in entry 3. Compounds
2 and 8 we assign for the sake of convenience by analogy,
although the consequences of a change in diene arrangement
are unknown (compare Schemes 1 and 3).46
the initial cyclization rate of 3 to the natural product cis-thujane
(4) was determined (Scheme 5). In the event, S-25 cyclizes 3
∼3.7 times faster than does R-25, suggesting that the (S)-catalyst
is matched to the 1R,5R stereochemistry of 4. Conversely, the
(R)-catalyst is better matched to the 2S,5S enantiomer of 4 and
by analogy to the 2R,5R stereochemistry of 20 (eq 8).47 For
cis-thujane 4, the matched catalyst is also more diastereoselec-
tive, although the existing stereogenic center is clearly stereo-
dominant.48
To help generate a model for the sense of asymmetric
induction in the cyclopropanation reactions, an X-ray structure
of the iodide precursor to R-25 was obtained, R-28. As shown
in Figure 2, the structure shows the expected distortion that
rotates the P3-Pt-I plane 25° counter clockwise from the P2-
Pt-P1 plane. As indicated by the forward projecting green 3,5-
xylyl substituents, the coordination site for alkene activation
(after I- removal) is highly asymmetric and should efficiently
promote diastereoselective alkene binding/activation (vide infra).
Because small bite angles were of importance in the dppm
catalysts, it is worth noting that the two I-Pt-P bond angles
are significantly smaller, 87° and 89°, than those observed in
the dppm/PMe3 structure (98° and 91°, respectively), consistent
Consistent with this stereochemical analysis were double
stereodifferentiating reactions carried out with enantiomeric pairs
of catalyst and (3S)-â-citronellene (3). In separate experiments,
(47) A change in priority due to N and branching inverts the Cahn-Ingold-Prelog
designators.
(48) In a traditional kinetic resolution mode using racemic 3 and R-25 (5 mol%),
an S-factor of 3.5 ( 0.3 was obtained. Martin, V. S.; Woodard, S. S.;
Katsuki, T.; Yamada, Y.; Ikeda, M.; Sharpless, K. B. J. Am. Chem. Soc.
1981, 103, 6237-6240.
(44) Alonso, D. A.; Andersson, P. G. J. Org. Chem. 1998, 63, 9455-9461.
(45) Polon˜ski, T.; Milewska, M. J.; Katrusiak, A. J. Am. Chem. Soc. 1993, 115,
11410-11417.
(46) We note that each of the substrates in Table 5 gives -ve rotations (MeOH,
546 nm), except for 2 which gives a +ve rotation.
9
13296 J. AM. CHEM. SOC. VOL. 128, NO. 40, 2006

Modular Catalysts for Diene Cycloisomerization
A R T I C L E S
Scheme 6. Diastereofacial Binding Energetics from PM3
Calculations on a Pd Analogue
Summary
We describe herein how deconstructing the tridentate archi-
tecture of the triphos ligand into a modular combination of
bidentate and monodentate ligands provides an efficient ap-
proach for catalyst discovery and optimization. Two significant
catalyst types were successfully discovered and optimized. The
first class utilized the small bite angle diphosphine dppm in
combination with PMe3 to generate the most efficient catalyst
yet discovered for the cycloisomerization of 1,6- and 1,7-dienes
into bicyclopropanes. These catalysts efficiently operate at the
5 mol % level, at ambient temperatures, and show tolerance to
functional groups that include sulfonamides, acetals, esters,
sulfones, and ketones (somewhat).
A class of catalysts for the enantioselective cycloisomerization
of dienes was also discovered and optimized. The optimum
catalyst was based on the xyl-BINAP ligand in combination
with PMe3, although in one case the SEGPHOS ligand was
preferable. Although more limited in reaction scope, these chiral
catalysts provided bicyclopropane products with enantioselec-
tivities into the mid-90s; they represent the most enantioselective
catalysts for diene cycloisomerization yet discovered (cf., eqs
3 and 4).
Scheme 7. Model for Stereochemical Transfer Invoking a
Stereochemistry-Defining Alkene Activation Step
During the course of catalyst discovery, several unexpected
leads presented themselves. For example, in the case of allyl
prenyl malonate substrate, a minor cyclohexene byproduct was
isolated. Subsequent experiments to optimize the formation of
this alternative cycloisomerization product were undertaken, and
conditions (alternative diphosphine and monophosphine) for
generating the cyclohexene as the major product were discov-
ered. Unfortunately, this new catalyst was not similarly suc-
cessful on other substrate classes, and so the endeavor was only
partially successful. A similarly surprising result was observed
in the asymmetric cycloisomerization chemistry, where leaving
the PMe3 out of the catalyst formulation actually inverts the
absolute sense of the stereoselection. Efforts to optimize this
lead were only partially successful, but again showed that
breaking the original tridentate architecture into a modular
combination of ligands is not only efficient with respect to
catalyst discovery and optimization, but ultimately also leads
to catalysts that can be easily constructed from commercially
available components.
with the need for a small ligand at the P1 site for these catalysts;
the dppm catalyst, on the other hand, could tolerate monodentate
ligands as large as PPh3.
To address the question of diastereofacial preference in this
coordination environment, the coordinates from the X-ray
structure were imported in MacSpartan and the iodide ligand
was replaced with propylene and the metal changed to Pd.
Because there are two diastereofaces to the alkene and two
rotamers for each diastereomer, four possible η2-alkene starting
structures were generated. Each structure was geometry opti-
mized at the PM3 level, and the energies were compared
(Scheme 6). For both diastereomers, the lowest energy rotamers
project the methyl substituent into the top quadrants of the chiral
environment (A and C), with the lowest energy structure being
that which points it toward the PMe3, C.
Acknowledgment. We thank Mr. Aujin Kim for preparing
the substrate in entries 9 and 11 of Table 3. The National
Institute of General Medicine (GM-60578) is gratefully ac-
knowledged for support. We thank Dr. Peter White for X-ray
structural determinations; correspondence regarding them should
be directed to his attention at UNC Chapel Hill. We also wish
to thank the Takasago Corp. for a generous gift of SEGPHOS,
and Chiral Quest for providing samples of the TUNEPHOS
ligands.
Given the significant proviso that alkene activation/C-C bond
formation is not likely turnover limiting, and that this process
is quite likely reversible (it is with triphos), this model for alkene
activation does correctly predict that antara-facial attack of the
trisubstituted alkene onto the activated alkene of the lowest
energy isomer will afford the observed 2R,5R isomer of 20
(Scheme 7). The model also holds for the 6-endo pathway
(Scheme 3), which first creates a stereogenic platinated carbon
center and then transfers this information into the cyclopropane
by an invertive cyclopropanation.13
Supporting Information Available: Full experimental pro-
cedures, X-ray data tables, and sample NMR spectra. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA064335D
9
J. AM. CHEM. SOC. VOL. 128, NO. 40, 2006 13297