Electronic and steric control of regioselectivities in Rh(I)-catalyzed (5 + 2) cycloadditions: Experiment and theory

Article abstract of DOI:10.1021/ja103253d

The first studies on the regioselectivity of Rh(I)-catalyzed (5 + 2) cycloadditions between vinylcyclopropanes (VCPs) and alkynes have been conducted experimentally and analyzed using density functional theory (DFT). The previously unexplored regiochemical consequences for this catalytic, intermolecular cycloaddition were determined by studying the reactions of several substituted VCPs with a range of unsymmetrical alkynes. Experimental trends were identified, and a predictive model was established. VCPs with terminal substitution on the alkene reacted with high regioselectivity (>20:1), as predicted by a theoretical model in which bulkier alkyne substituents prefer to be distal to the forming C-C bond to avoid steric repulsions. VCPs with substitution at the internal position of the alkene reacted with variable regioselectivity (ranging from >20:1 to a reversed 1:2.3), suggesting a refined model in which electron-withdrawing substituents on the alkyne decrease or reverse sterically controlled selectivity by stabilizing the transition state in which the substituent is proximal to the forming C-C bond.

Full text of DOI:10.1021/ja103253d

Published on Web 06/30/2010
Electronic and Steric Control of Regioselectivities
in Rh(I)-Catalyzed (5 + 2) Cycloadditions:
Experiment and Theory
Peng Liu,† Lauren E. Sirois,‡ Paul Ha-Yeon Cheong,†,§ Zhi-Xiang Yu,†,⊥
Ingo V. Hartung,‡,¶ Heiko Rieck,‡,0 Paul A. Wender,*,‡ and K. N. Houk*,†
Department of Chemistry and Biochemistry, UniVersity of California, Los Angeles,
California 90095-1569, and Department of Chemistry, Stanford UniVersity,
Stanford, California 94305-5080
Received April 17, 2010; E-mail: wenderp@stanford.edu; houk@chem.ucla.edu
Abstract: The first studies on the regioselectivity of Rh(I)-catalyzed (5 + 2) cycloadditions between
vinylcyclopropanes (VCPs) and alkynes have been conducted experimentally and analyzed using density
functional theory (DFT). The previously unexplored regiochemical consequences for this catalytic,
intermolecular cycloaddition were determined by studying the reactions of several substituted VCPs with
a range of unsymmetrical alkynes. Experimental trends were identified, and a predictive model was
established. VCPs with terminal substitution on the alkene reacted with high regioselectivity (>20:1), as
predicted by a theoretical model in which bulkier alkyne substituents prefer to be distal to the forming C-C
bond to avoid steric repulsions. VCPs with substitution at the internal position of the alkene reacted with
variable regioselectivity (ranging from >20:1 to a reversed 1:2.3), suggesting a refined model in which
electron-withdrawing substituents on the alkyne decrease or reverse sterically controlled selectivity by
stabilizing the transition state in which the substituent is proximal to the forming C-C bond.
cycloheptene derivatives.4 Figuring as a new strategy-level
reaction in several syntheses,5 the intramolecular version of this
Introduction
A preeminent goal of synthesis is the generation of structural
complexity and functional value with step economy.1 As evident
from the impact of the Diels-Alder (4 + 2) reaction in
synthesis, cycloadditions are especially powerful in this regard.
They proceed in one step with the convergent assembly of
simple components into a new ring or ring system with multiple
stereocenters. Studies on metal-mediated cycloadditions for the
synthesis of medium-sized rings represent a particularly im-
portant area of research, given the number of structurally novel
and biologically potent targets that incorporate such rings.
Although the introduction of methods for the synthesis of seven-
membered rings still lags behind that for smaller rings (par-
ticularly for catalytic, fully intermolecular variants), noteworthy
progress has been made.2 In 1995, we reported the first examples
of metal-catalyzed (5 + 2) cycloadditions between vinylcyclo-
propanes(VCPs)andπ-systems,3 ahomologueoftheDiels-Alder
cycloaddition that provides a general and effective route to
reaction provides a single cycloadduct regioisomer with respect
to insertion of the two-carbon component due to the structural
constraints of the tether between it and the VCP. Thus far,
however, the steric and electronic factors influencing the
regiochemistry of the unconstrained intermolecular process
(which is currently limited to catalysis by rhodium) have not
(2) (a) Yet, L. Chem. ReV. 2000, 100, 2963–3007. (b) Battiste, M. A.;
Pelphrey, P. M.; Wright, D. L. Chem.sEur. J. 2006, 12, 3438–3447.
(c) Wender, P. A.; Croatt, M. P.; Deschamps, N. M. In ComprehensiVe
Organometallic Chemistry III; Crabtree, R. H., Mingos, D. M. P., Eds.;
Elsevier: Oxford, 2007; Vol. 10, pp 603-648. (d) Butenscho¨n, H.
Angew. Chem., Int. Ed. 2008, 47, 5287–5290.
(3) For the first report, see: Wender, P. A.; Takahashi, H.; Witulski, B.
J. Am. Chem. Soc. 1995, 117, 4720–4721.
(4) For intermolecular reactions involving alkynes, see: (a) Wender, P. A.;
Rieck, H.; Fuji, M. J. Am. Chem. Soc. 1998, 120, 10976–10977. (b)
Wender, P. A.; Dyckman, A. J.; Husfeld, C. O.; Scanio, M. J. C. Org.
Lett. 2000, 2, 1609–1611. (c) Wender, P. A.; Barzilay, C. M.;
Dyckman, A. J. J. Am. Chem. Soc. 2001, 123, 179–180. (d) Wender,
P. A.; Gamber, G. G.; Scanio, M. J. C. Angew. Chem., Int. Ed. 2001,
40, 3895–3897. (e) Wender, P. A.; Stemmler, R. T.; Sirois, L. E. J. Am.
Chem. Soc. 2010, 132, 2532–2533. (f) Wender, P. A.; Sirois, L. E.;
Stemmler, R. T.; Williams, T. J. Org. Lett. 2010, 12, 1604–1607. For
intermolecular reactions involving allenes, see: (g) Wegner, H. A.;
de Meijere, A.; Wender, P. A. J. Am. Chem. Soc. 2005, 127, 6530–
6531.
† University of California.
‡ Stanford University.
§ Present address: Department of Chemistry, Oregon State University,
Corvallis, OR 97331.
⊥ Present address: College of Chemistry, Peking University, Beijing
100871, P.R. China.
¶ Present address: Global Drug Discovery, Lead Generation & Optimiza-
tion, Bayer-Schering Pharma AG, 42096 Wuppertal, Germany.
0 Present address: PM Project Management, Bayer CropScience AG,
40789 Monheim am Rhein, Germany.
(5) For uses of the intramolecular (5 + 2) reaction in synthesis, see: (a)
Wender, P. A.; Fuji, M.; Husfeld, C. O.; Love, J. A. Org. Lett. 1999,
1, 137–139. (b) Wender, P. A.; Zhang, L. Org. Lett. 2000, 2, 2323–
2326. (c) Wender, P. A.; Bi, F. C.; Brodney, M. A.; Gosselin, F. Org.
Lett. 2001, 3, 2105–2108. (d) Ashfeld, B. L.; Martin, S. F. Tetrahedron
2006, 62, 10497–10506. (e) Trost, B. M.; Hu, Y.; Horne, D. B. J. Am.
Chem. Soc. 2007, 129, 11781–11790. (f) Trost, B. M.; Waser, J.;
Meyer, A. J. Am. Chem. Soc. 2008, 130, 16424–16434.
(1) (a) Wender, P. A.; Handy, S. T.; Wright, D. L. Chem. Ind. 1997, 765–
769. (b) Wender, P. A.; Croatt, M. P.; Witulski, B. Tetrahedron 2006,
62, 7505–7511, references therein. (c) Wender, P. A.; Verma, V. A.;
Paxton, T. J.; Pillow, T. H. Acc. Chem. Res. 2008, 41, 40–49. (d)
Wender, P. A.; Miller, B. L. Nature 2009, 460, 197–201.
9
10.1021/ja103253d  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 10127–10135 10127

A R T I C L E S
Liu et al.
Scheme 1. Possible Regioisomers in (5 + 2) Cycloadditions between Substituted VCPs and Terminal Alkynes
Scheme 2. Possible Mechanistic Pathways for the Rhodium-Catalyzed (5 + 2) Cycloaddition
been addressed.6 Substituted VCPs such as 1 and unsymmetri-
cally substituted alkynes could lead to two regioisomeric
products (Scheme 1). Cycloadduct (3,4)-5 would have groups
positioned vicinally as is observed with tethered substituents
in the intramolecular process, whereas cycloadduct (3,5)-5 would
have a substitution pattern complementary to the intramolecular
process. VCP 2 also offers a possible route to regioisomeric
adducts but now having a different substitution pattern relative
to that of VCP 1. In the reaction of VCP 3, a more complex
scenario could unfold as four regioisomeric products are
possible. While critical to the predictable use of the intermo-
lecular process for accessing more substituted, complex synthetic
targets, the regiochemistry of Rh(I)-catalyzed intermolecular (5
+ 2) cycloadditions has not previously been explored experi-
mentally or analyzed theoretically.
There are two general mechanisms proposed for the rhodium-
catalyzed (5 + 2) cycloaddition. One would proceed through
initial formation of a metallacyclohexene followed by alkyne
insertion and reductive elimination (Scheme 2). A second would
involve initial formation of a metallacyclopentene followed by
cleavage of the cyclopropane (ring expansion) and reductive
elimination. We have previously reported on the mechanism
for the reactions of VCPs with simple acetylene, analyzing data
from density functional theory (DFT) calculations.7 According
to the theoretical studies, the metallacyclohexene mechanism
is preferred, and the rate-determining step is the alkyne insertion
to form an eight-membered metallacycle intermediate. The
regioselectivities of oxidative addition or migratory insertion
of alkenes and alkynes to form metallacycles have been
investigated theoretically for a subset of other organometallic
reactions.8 For example, oxidative additions of bis(olefin)metal
complexes in general favor the product in which electron-
withdrawing substituents are positioned R to the metal to
maximize coefficients on the ꢀ-carbon, thereby facilitating C-C
bond formation. However, reactions involving alkynes usually
do not follow this trend, and in many cases the regioselectivities
are dominated by steric effects. At present neither the regiose-
lectivity preference for the metal-catalyzed intermolecular (5
+ 2) cycloaddition nor its magnitude is known. Thus, we sought
to establish the experimental regioselectivities for a standard
set of alkynes and VCPs and to analyze these data using DFT
calculations. Regioselectivity is a key factor in the predictive
use of a new reaction, and its exploration becomes part of the
touchstone for synthetic applications of the method.9 The
widespread use of the Diels-Alder10 reaction in synthesis, for
(7) (a) Yu, Z.-X.; Wender, P. A.; Houk, K. N. J. Am. Chem. Soc. 2004,
126, 9154–9155. (b) Yu, Z.-X.; Cheong, P. H.-Y.; Liu, P.; Legault,
C. Y.; Wender, P. A.; Houk, K. N. J. Am. Chem. Soc. 2008, 130,
2378–2379. (c) Liu, P.; Cheong, P. H.-Y.; Yu, Z.-X.; Wender, P. A.;
Houk, K. N. Angew. Chem., Int. Ed. 2008, 47, 3939–3941.
(8) For examples of related theoretical studies on the regioselectivity of
metal-catalyzed reactions with substituted alkynes, see: (a) Stockis,
A.; Hoffmann, R. J. Am. Chem. Soc. 1980, 102, 2952–2962. (b)
Wakatsuki, Y.; Nomura, O.; Kitaura, K.; Morokuma, K.; Yamazaki,
H. J. Am. Chem. Soc. 1983, 105, 1907–1912. (c) Dahy, A. A.; Koga,
N. Bull. Chem. Soc. Jpn. 2005, 78, 781–791. (d) Yamamoto, Y.;
Kinpara, K.; Saigoku, F.; Takagishi, H.; Okuda, S.; Nishiyama, H.;
Itoh, K. J. Am. Chem. Soc. 2005, 127, 605–613. (e) Liu, P.; Jordan,
R. W.; Kibbee, S. P.; Goddard, J. D.; Tam, W. J. Org. Chem. 2006,
71, 3793–3803. For related insights, see: (f) Dalton, D. M.; Oberg,
K. M.; Yu, R. T.; Lee, E. E.; Perreault, S.; Oinen, M. E.; Pease, M. L.;
Malik, G.; Rovis, T. J. Am. Chem. Soc. 2009, 131, 15717–15728.
(6) For studies involving regiochemical aspects of the rhodium- and
ruthenium-catalyzed intramolecular processes, see: (a) Wender, P. A.;
Dyckman, A. J. Org. Lett. 1999, 1, 2089–2092. (b) Wender, P. A.;
Dyckman, A. J.; Husfeld, C. O.; Kadereit, D.; Love, J. A.; Rieck, H.
J. Am. Chem. Soc. 1999, 121, 10442–10443. (c) Trost, B. M.; Shen,
H. C. Org. Lett. 2000, 2, 2523–2525. (d) Trost, B. M.; Shen, H. C.;
Horne, D. B.; Toste, F. D.; Steinmetz, B. G.; Koradin, C. Chem.sEur.
J. 2005, 11, 2577–2590.
9
10128 J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010

Rh(I)-Catalyzed Intermolecular (5 + 2) Cycloadditions: Regioselectivity
Table 1. (5 + 2) Cycloadditions of VCP 1a and Terminal Alkynesa
A R T I C L E S
Table 2. (5 + 2) Cycloadditions of VCP 2a and Terminal Alkynesa
time
(h)
TFEb
yieldc
time (h) TFEb (%) product yieldc (%) (3,5)-5:(3,4)-5d
entry alkyne
R
(%) product(s) (%) (2,5)-6:(2,4)-6d
entry alkyne
R
1
2
3
4
5
6
4a
4b CH2OH
4c CMe2OH
4d TMS
n-Pr
48
6
5
18
4.25
5
2.75
11.5
8
5
-
-
5
5
-
5
5
-
5
6a
6b
6c
6d
6e
6f
6f
6f
6g
6g
6h
76
77
73
85
84
84
91
84
66
82
78
7.1:1
3.1:1
>20:1
>20:1
3.0:1
1:1.7
1:1.9
1:2.3
2.4:1
2.2:1
7.7:1
1
2
3
4
5
6
4a n-Pr
4c CMe2OH 22
4d TMS 18
8.25
23
4.75
4.25
25
-
5
5
5
5a
5c
5d
5e
5f
78
69
74
90
77
84
>20:1
>20:1
>20:1
>20:1
>20:1
>20:1
4e
4f
4h Ph
CO2Me
COMe
4e
4f
4f
4f
4g
4g
CO2Me
COMe
COMe
COMe
CO-p-CF3-Ph
CO-p-CF3-Ph 4.5
25
5h
7
8e
9
a Conditions: 1.2-1.5 equiv of alkyne 4, 5 mol % [Rh(CO)2Cl]2, DCE
or TFE/DCE (0.1 M), 40 °C; 1% HCl/EtOH. DCE ) 1,2-dichloroethane;
TFE ) 2,2,2-trifluoroethanol. b Volume percent. c Combined isolated
10
11
d
1
4h Ph
7
5
yield. Ratio determined by H NMR.
a Conditions: 1.2-1.5 equiv of alkyne 4, 5 mol % [Rh(CO)2Cl]2,
DCE or TFE/DCE (0.1 M), 40 °C; 1% HCl/EtOH. b Volume percent.
c Combined isolated yield. d Ratio determined by 1H NMR. e 1 mol %
[Rh(CO)2Cl]2.
example, is largely due to the body of knowledge that has been
accumulated regarding predictable regio- and stereoselective
outcomes. We report herein the first experimental study of the
regioselectivities of these synthetically useful (5 + 2) reactions
and a theoretical study that predicts the trends observed
experimentally and explains the origins of the steric and
electronic factors that control regioselectivity.
favored, but in many cases to a lesser extent (Table 2). The use
of larger or electron-rich alkyne substituents resulted in high
selectivities (>20:1) and good yields (entries 3-4). For alkynes
with electron-withdrawing substituents, calculations (vide infra)
predict that the regioselectivity should be attenuated, if not
reversed, by electronic effects. This expectation was borne out
experimentally with electron-poor alkynes affording cycload-
ducts with reduced (entries 5, 9-10) and even reversed
regioselectivities (entries 6-8) in good combined yields.13
Catalyst loading can be lowered to 1 mol %, although a longer
reaction time is required (entry 8). As observed for VCP 1a,
addition of TFE cosolvent improved the reaction times and
yields with negligible change in regioselectivity.
For studies of the (5 + 2) reaction and its regioselectivity,
VCP 3a represents an interesting case, as four regioisomeric
products could be formed with an unsymmetrical alkyne (see
also Scheme 5 and discussion below). When an additional
substituent is present on the cyclopropane ring, oxidative
cyclization with rhodium can proceed with cleavage of either
the more- or less-substituted cyclopropyl bond. According to
our mechanistic proposal (vide infra), cleavage of the more
substituted cyclopropane bond would give rise to (3,5)-5 and
(3,4)-5 products, whereas cleavage of the less substituted bond
would afford (2,5)-6 and (2,4)-6 cycloadducts. If the substituent
on the cyclopropane ring is a methyl group, these are collectively
the same products as for (5 + 2) reactions of VCPs 1a and 2a.
The results of representative (5 + 2) reactions of VCP 3a with
ethynyltrimethylsilane, methyl propiolate, and 3-butyn-2-one are
given in Table 3.
Results and Discussion
The (5 + 2) cycloadditions were performed under the
conditions outlined in Table 1. Significantly, the reactions of
terminally substituted VCP 1a with terminal alkynes resulted
in exclusive formation of 3,5-disubstituted cycloheptenones
(3,5)-5.11 This dramatic preference was observed for both
electron-rich and -poor alkynes (entries 1-3 and 4-5, respec-
tively). From a synthetic viewpoint, the resulting (3,5)-substitu-
tion pattern preference complements the regioselectivity of the
intramolecular process, which provides vicinally positioned
substituents with complete selectivity. Despite the increased
steric encumbrance of VCP 1a relative to that of an unsubstituted
VCP, cycloadducts (3,5)-5 were isolated as the ketone products
in good to excellent yields following mildly acidic workup. In
many cases the addition of 2,2,2-trifluoroethanol (TFE) as a
cosolvent in 1,2-dichloroethane (DCE) allowed for shorter
reaction times and improved yields without affecting the
regioselective outcome.12
For reactions of internally substituted VCP 2a, the formation
of 5-substituted adducts (2,5)-6, now having a different substitu-
tion pattern (2,5 for VCP 2a vs 3,5 for VCP 1a), were again
(9) For recent representative examples of regiochemical consequences in
other Rh(I)-catalyzed intermolecular cycloadditions (e.g., (2 + 2 +
2), (3 + 2 + 2), and (4 + 2 + 2) reactions) involving alkynes, see:
(a) Friedman, R. K.; Rovis, T. J. Am. Chem. Soc. 2009, 131, 10775–
10782, and references therein. (b) Evans, P. A.; Inglesby, P. A. J. Am.
Chem. Soc. 2008, 130, 12838–12839. (c) Tanaka, K. Synlett 2007,
1977–1993, and references therein. (d) Evans, P. A.; Lai, K. W.;
Sawyer, J. R. J. Am. Chem. Soc. 2005, 127, 12466–12467. (e)
Gilbertson, S. R.; DeBoef, B. J. Am. Chem. Soc. 2002, 124, 8784–
8785. (f) Murakami, M.; Ubukata, M.; Itami, K.; Ito, Y. Angew. Chem.,
Int. Ed. 1998, 37, 2248–2250.
For the cycloaddition of VCP 3a and ethynyltrimethylsilane
(entry 1), only one observable product was formed, the
4-substituted adduct (2,4)-6d with the methyl group from the
VCP R to the ketone, representing cleavage of the less
(12) For a review of the use of fluorinated alcohols in catalysis, see: (a)
Shuklov, I. A.; Dubrovina, N. V.; Bo¨rner, A. Synthesis 2007, 2925–
2943. For similar effects, see: (b) Reference 3. (c) Reference 4c. (d)
Wender, P. A.; Croatt, M. P.; Deschamps, N. M. Angew. Chem., Int.
Ed. 2006, 45, 2459–2462. (e) Saito, A.; Ono, T.; Hanzawa, Y. J. Org.
Chem. 2006, 71, 6437–6443.
(10) Review: Nicolaou, K. C.; Snyder, S. A.; Montagnon, T.; Vassiliko-
giannakis, G. Angew. Chem., Int. Ed. 2002, 41, 1668–1698, and
references therein.
(11) Minor proximal cycloadduct (3,4)-5f was observed in trace amount
by GC-MS; for all other reactions of VCP 1a, a single cycloadduct
peak was observed in GC-MS analysis of the crude reaction mixture
and was later determined to correspond to the product indicated.
(13) For a related observation, see: Wender, P. A.; Christy, J. P. J. Am.
Chem. Soc. 2006, 128, 5354–5355.
9
J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010 10129

A R T I C L E S
Liu et al.
Scheme 3. VCPs 2a and 3a Provide Access to Regiochemically
Complementary Cycloadducts
Scheme 4. Catalytic Cycle for Rh(I)-Catalyzed (5 + 2)
Cycloadditions
substituted cyclopropane bond.14 The 4-substituted product (2,4)-
6e was also favored for the reaction of methyl propiolate, in a
∼3:1 ratio over 5-substituted adduct (2,5)-6e. These results have
important ramifications for synthetic applications as they suggest
that, for a given alkyne, one could selectively produce either
the (2,5)- or the (2,4)-substituted regioisomer by judicious
selection of the appropriate VCP (2a or 3a, respectively, Scheme
3). VCP 3a provides access to the regiochemically comple-
mentary products of reactions of VCP 2a, with 4-substituted
adducts favored for the former to a similar degree that
5-substituted adducts are favored for the latter. Interestingly,
the (5 + 2) cycloaddition of methyl propiolate also provided a
small amount of product (3,5)-5e, resulting from cleavage of
the more substituted cyclopropyl bond. A significant amount
of (3,5)-5f (similarly resulting from cleavage of the more
substituted cyclopropyl bond) was also produced from reaction
of VCP 3a with 3-butyn-2-one. In accordance with the reversal
of regioselectivity versus reactions of VCP 2a, in this case
5-substituted adduct (2,5)-6f was formed preferentially over
4-substituted adduct (2,4)-6f in a 2.5:1 ratio (Table 3, entry 3
versus Table 2, entry 7). In no case were products (3,4)-5
observed at the completion of the cycloaddition. This, too, is
synthetically significant as that type of substitution is alterna-
tively available with complete selectivity through the intramo-
lecular (5 + 2) reaction.
To analyze these trends and to uncover the steric and
electronic effects at play, calculations with density functional
theory (B3LYP/SDD-6-31G* with CPCM solvation model)
were performed. Previous theoretical studies7 revealed that
rhodium-catalyzed intermolecular (5 + 2) cycloadditions likely
occur via the catalytic cycle shown in Scheme 4. The rate-
determining step is the alkyne insertion to form the eight-
membered rhodacycle intermediate 9. Thus, cycloadditions with
unsymmetrically substituted alkynes could lead to two regioi-
someric products based upon the orientation of the alkyne during
migratory insertion. Accordingly, the regioselectivity is deter-
mined by the energy difference between these distal and
proximal alkyne insertion transition states (∆∆Gq(prox-dis), vide
infra). Distal is defined here as alkyne substituent R positioned
away from the forming C-C bond in the 2π insertion transition
state; proximal is defined as R oriented toward the forming C-C
bond (and away from the metal center). The activation free
energies of distal- and proximal-transition states TS1, TS2, and
TS3 for cycloadditions between VCP 1b, 2b, or 3b and terminal
alkynes are given in Tables 4, 5, and 6, respectively. To reduce
computational time, a smaller methyl group (Pg ) Me) was
used as a proxy for the silyl group (Pg ) TBS) on the C-1
oxygen of the VCPs, since 1-methoxy and 1-siloxy VCPs are
predicted to have similar regioselectivities.15 Computational
results reflect experimental trends for the vast majority of
examples studied; indeed, disagreements between the in vitro
and in silico product ratios are generally observed only for cases
in which the energy differences are computed to be quite small
(j1 kcal/mol). Thus, these findings begin to suggest a predictive
model for the regiochemical outcomes of these reactions.
According to the theoretical models, in the rate- and regio-
chemistry-determining alkyne insertion transition states, the
alkyne is in the same plane as rhodium and the two carbons
formerly comprising the vinyl group of the VCP (Figure 1).
Two possible steric interactions might arise in the alkyne
insertion transition state when a bulky group is present on the
alkyne: repulsions between the alkyne substituent R and the
VCP terminus in the proximal-TS and repulsions between R
and the catalyst in the distal-TS. For reactions with all VCPs,
the repulsion between alkyne substituent R and the VCP-derived
fragment is stronger than that between R and the metal center
or ligands, as indicated by spatial and bond length analysis. For
reactions of VCP 1b, significant repulsions are observed between
the alkyne substituent R and the methyl group on the terminal
position of the VCP metallacycle in proximal-TS1. Thus,
formation of (3,5)-substituted products is strongly favored for
the cycloadditions of VCP 1b with all alkynes (Table 4). To
Table 3. (5 + 2) Cycloadditions of VCP 3a and Terminal Alkynesa
entry
alkyne
R
time (h)
product(s)
yieldb (%)
(3,5)-5:(3,4)-5:(2,5)-6:(2,4)-6c
1
2
3
4d
4e
4f
TMS
CO2Me
COMe
18
31
10
6d
5e/6e
5f/6f
87
69
86
0:0:1:>20
1:0:2.1:6.4
1.7:0:2.5:1
a Conditions: 1.2-1.3 equiv of alkyne 4, 5 mol % [Rh(CO)2Cl]2, TFE (5 vol %) in DCE (0.1 M), 40 °C; 1% HCl/EtOH. b Combined isolated yield.
c
1
Ratio determined by H NMR.
9
10130 J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010

Rh(I)-Catalyzed Intermolecular (5 + 2) Cycloadditions: Regioselectivity
A R T I C L E S
Table 4. B3LYP/SDD-6-31G*/CPCM(DCE) Free Energies of
Activation for (5 + 2) Cycloadditions between VCP 1b and
Terminal Alkynes
Table 5. B3LYP/SDD-6-31G*/CPCM(DCE) Free Energies of
Activation for (5 + 2) Cycloadditions between VCP 2b and
Terminal Alkynes
∆Gq(TS2)a,b
predicted ratioc
∆Gq(TS1)a,b
predicted ratioc
entry alkyne
R
distal- proximal- ∆∆Gq(prox-dis)b (2,5)-6:(2,4)-6
entry alkyne
R
distal- proximal- ∆∆Gq(prox-dis)b (3,5)-5:(3,4)-5
1
2
3
4
5
6
7
8
9
4i
4j
H
Me
7.8
7.8
-
-
1
2
3
4
5
6
7
8
4i
4j
4a
H
Me
n-Pr
7.0
10.9
12.1
13.9
6.4
8.0
7.1
9.1
9.1
7.0
16.1
17.5
25.7
13.5
19.4
15.1
11.7
10.6
12.5
8.3
-
-
11.9 12.8
12.8 14.1
14.9 18.8
7.3 10.3
8.5 12.9
14.6 16.9
10.2 10.3
0.9
4.9:1
5.1
>20:1
>20:1d
>20:1
>20:1
>20:1d
>20:1d
>20:1d
11:1d
>20:1
5.3:1
4a n-Pr
4k t-Bu
4b CH2OH
4c CMe2OH
4d TMS
4e
4f
1.3
3.8
2.9
4.3
2.2
8.5:1d
>20:1
>20:1d
>20:1d
>20:1d
1.2:1d
1:>20d
1:1.3d
1:20
5.4
11.8
7.1
11.4
8.0
2.5
1.4
2.1
1.0
4k t-Bu
4b CH2OH
4c CMe2OH
4d TMS
4e
4f
4g
4l
CO2Me
COMe
0.1
CO2Me
COMe
CO-p-CF3-Ph 10.4
CH)O
10.9
8.7
-2.2
-0.2
-1.8
1.1
2.2
3.4
9
10 4g CO-p-CF3-Ph 11.7 11.5
11 4l
12 4m CN
13 4h Ph
14 4n NH2
10
11
12
13
14
CH)O
8.6
8.5
15.4 17.6
8.5 11.9
6.8
9.6
7.3
7.5
14.0
7.6
6.4:1
4m CN
4h Ph
4n NH2
9.8
19.6
13.6
2.3
5.6
6.1
>20:1
>20:1d
>20:1
>20:1d
>20:1
a Relative to the Rh(CO)Cl-VCP π-complex. b Energies in kcal/mol.
c Calculated ratio at 298 K. d See Table 2 for ratio observed experi-
mentally.
a Relative to the Rh(CO)Cl-VCP π-complex. b Energies in kcal/mol.
c Calculated ratio at 298 K. d See Table 1 for ratio observed experi-
mentally.
contrast, electron-withdrawing groups on the alkyne stabilize
proximal-TS1 and -TS2, the pathways leading to 4-substituted
products. This orientation allows the largest coefficient of the
alkyne π* orbital to interact with the metal, thus enhancing back-
donation of electron density from the filled rhodium dxy orbital
(Figure 2). The resultant stability manifests itself in lower
activation barriers for proximal-TS for electron-withdrawing
substituents as compared to alkyl or phenyl alkyne substituents.
For the cycloadditions of VCP 2b and alkynes with small and
strongly electron-withdrawing groups, such as ketones or
aldehydes (entries 9 and 11, Table 516), these electronic effects
dominate over steric repulsion and reverse the regioselectivity
to favor the 4-substituted product (Figure 2).
As briefly discussed above, when additional substitution is
present on the cyclopropane (as in VCP 3), the ring can open
in two ways during oxidative cyclization, with cleavage of either
the more- or less-substituted bond. For each of the resulting
metallacyclohexenes, the subsequent alkyne insertion step can
then proceed with the alkyne in either of two different
orientations (as for VCPs 1 and 2), leading to 4-substituted and
5-substituted products for each (Scheme 5). If the cyclopropyl
substituent is a methyl group, the four possible regioisomeric
cycloadducts are the same as the possible products from (5 +
2) reactions of VCPs 1 and 2 with terminal alkynes, i.e., (3,4)-
5, (3,5)-5, (2,4)-6, and (2,5)-6. However, different trends of
regioselectivity are observed, since the alkyne insertion orienta-
tions leading to either 4-substituted or 5-substituted products
are reversed relative to those in TS1 and TS2. Distal-TS3-a
illustrate this effect, the transition state structures for the reaction
between VCP 1b and propyne 4j (R ) Me) are shown in Figure
1. The distal transition state, distal-TS1j, is strongly favored
by 5.1 kcal/mol. Due to larger steric repulsions, the forming
C-C bond length in proximal-TS1j is 0.06 Å longer than in
distal-TS1j. For cycloadditions involving other alkyl-substituted
alkynes, the distal pathway preferences correlate with the steric
bulk of the alkyl groups (entries 2-4, Table 4). Cycloadditions
of internally substituted VCP 2b exhibit a slightly diminished
preference for the distal pathway, due to the reduced steric bulk
of the VCP terminus. For example, the reaction between pentyne
4a (R ) n-Pr) and VCP 1b is highly distal-selective (∆∆Gq )
5.4 kcal/mol), whereas the analogous reaction involving VCP
2b is less so (∆∆Gq ) 1.3 kcal/mol, experimentally 1.2 kcal/
mol or 7.1:1).
While steric effects appear to govern the regioselectivity of
many of these reactions, the combined experimental and
theoretical results indicate that electronic effects related to the
alkyne substituent also have a significant influence on the degree
and even the sense of regioselectivity. Electron-donating sub-
stituents reinforce the steric preference for distal alkyne orienta-
tion in TS1 and TS2. For example, the small, but powerfully
electron-donating amino substituent of alkyne 4n (R ) NH2) is
predicted to yield the 5-substituted product exclusively for both
VCPs 1b and 2b (entry 14, Table 4 and entry 14, Table 5). In
(14) For examples of intramolecular (5 + 2) reactions featuring cyclopropyl
substitution and studies of catalyst and stereochemical influence on
cyclopropyl ring cleavage, see references 5 and 6.
(15) 1-Methoxy VCPs (Pg ) Me) were surrogates for 1-silyloxy VCPs
(Pg ) TBS) in the calculations due to their similar regioselectivities.
Reaction of 1-TMSO-VCP 2 and propynal (4l) is computed to favor
the proximal pathway by 1.6 kcal/mol, similar to the 1.8 kcal/mol
proximal preference for the 1-MeO-VCP 2b (Table 5, entry 11).
(16) A very small proximal preference is predicted for alkyne 4g (Table 5,
entry 10). This corresponds to the low regioselectivity observed
experimentally for alkyne 4g (Table 2, entries 9 and 10), in which,
however, the distal product (2,5)-6g is slightly favored.
9
J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010 10131

A R T I C L E S
Liu et al.
Figure 1. Optimized structures of the alkyne insertion transition states distal- and proximal-TS1j for the cycloaddition of VCP 1b with propyne (4j).
Scheme 5. Possible Regioisomeric Pathways for Rh(I)-Catalyzed (5 + 2) Cycloadditions of VCP 3a
a Single enantiomers of 15-a, 15-b, TS3-a, and TS3-b are depicted for clarity.
this scenario, the cleavage of cyclopropyl C-C bonds is
reversible and the two metallacyclic intermediates 15-a and
15-b are in equilibrium with each other. Previous theoretical
and experimental studies on rhodium(I)-catalyzed (5 + 2)
reactions also suggest that steps prior to alkyne insertion in
the catalytic cycle are reversible.17 Insertion of alkyne to
intermediates 15-a and 15-b leads to stable eight-membered
metallacycles (see intermediate 9, for example, Scheme 4).
Figure 2. Stabilizing back-donation from the rhodium dxy orbital to π*y of
This step is irreversible (see Supporting Information for
the alkyne is largest in the proximal transition state (illustrated here for
details). Thus, alkyne insertion (TS3) is the regioselectivity-
TS2 involving VCP 2b and alkyne 4l).
determining step.18 Under these Curtin-Hammett-like condi-
tions,19 the distribution of cycloadduct formation is deter-
mined by the energy differences among the alkyne insertion
transition states in the four regioisomeric pathways. The
calculated activation energies for the four pathways of the
reactions of VCP 3b and several alkynes are given in
Table 6.
and -TS3-b, in which the alkyne substituent R is distal to the
forming C-C bond, ultimately lead to 4-substituted products,
while proximal-TS3-a and -TS3-b, in which R is proximal to
the forming C-C bond, ultimately lead to 5-substituted products
(Scheme 5).
Cleavage of either the more- or less-substituted cyclopropyl
C-C bond of VCP 3b (Pg ) Me) is predicted to be very
fast. The activation barriers for cleavage of the more
substituted (bond a cleaved) and less substituted (bond b
cleaved) cyclopropyl C-C bonds are only 2.1 and 0.4 kcal/
mol, respectively, with respect to the Rh(CO)Cl-VCP π
complex. These barriers are much lower than those of any
of the subsequent alkyne insertion steps possible. Thus, in
(17) See references 6b and 7, for example.
(18) For all pathways except distal-a, alkyne insertion is also the rate-
determining step. For pathway distal-a, subsequent reductive elimina-
tion requires slightly higher activation energies than alkyne insertion.
Thus, both alkyne insertion and reductive elimination are the slow
steps in the catalytic cycle for pathway distal-a. See Supporting
Information for details.
(19) (a) Seeman, J. I. Chem. ReV. 1983, 83, 83–134. (b) Seeman, J. I.
J. Chem. Educ. 1986, 63, 42–48.
9
10132 J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010

Rh(I)-Catalyzed Intermolecular (5 + 2) Cycloadditions: Regioselectivity
A R T I C L E S
Table 6. B3LYP/SDD-6-31G*/CPCM(DCE) Free Energies of Activation of the Alkyne Insertion Transition States for (5 + 2) Cycloadditions
between VCP 3b and Terminal Alkynes
∆Gq(TS3)a b
,
entry
alkyne
R
proximal-TS3-a
distal-TS3-a
proximal-TS3-b
distal-TS3-b
1
2
3
4
5
6
4i
H
Me
TMS
CO2Me
COMe
CHdO
6.3
11.8
15.1
9.0
7.5
5.4
6.3
10.9
13.9
8.9
9.6
7.3
5.5
10.9
14.3
8.8
7.8
5.8
5.5
10.0
12.1
8.1
9.0
6.5
4j
4d
4e
4f
4l
a Relative to the Rh(CO)Cl-VCP π-complex. b Energies in kcal/mol.
Figure 3. Optimized structures of the alkyne insertion transition states for the cycloaddition of VCP 3b and alkyne 4d.
The relative activation energies among the four possible
alkyne insertion transition states of the reactions of VCP 3b
are influenced by several factors. As explained in the analysis
for VCPs 1b and 2b, one can imagine steric repulsions related
to the alkyne substituent and the forming C-C bond in the
proximal transition states for VCP 3b, as well as the energy-
lowering electronic influence of strong backbonding from
rhodium to the π* orbital of the alkyne in these same
proximal transition states when the substituent thereon is
electron-withdrawing (see Figure 2). Moreover, for VCP 3b,
the distal transition states involving cleavage of the more
substituted cyclopropyl bond (distal-TS3-a) are destabilized
by potential repulsions between the methyl group on the
cyclopropyl ring and the R substituent on the alkyne. To
exemplify these points graphically, the computed transition
states for the reaction of VCP 3b and alkyne 4d (R ) TMS)
are illustrated in Figure 3.
For example, due to steric repulsions associated with the
forming C-C bond, bulky alkyne substituents such as TMS
orient distal to the forming C-C bond in the lower-energy
alkyne insertion transition states (distal-TS3-a and -TS3-b).
Furthermore, the steric interaction of an especially bulky
alkyne substituent such as TMS with the methyl group from
cyclopropane results in a lower activation barrier for distal-
TS3-b as compared to distal-TS3-a. Thus, in these cases,
distal-TS3-b is the most favored pathway, leading to cy-
9
J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010 10133

A R T I C L E S
Liu et al.
Figure 4. Summary of regiochemical outcomes for intermolecular (5 + 2) reactions of VCPs 1, 2, and 3 with terminal alkynes.
cloadducts (2,4)-6 preferentially (entries 2 and 3, Table 6).
Conversely, for alkynes with small and strongly electron-
withdrawing groups such as formyl or acyl, stabilizing d f
π* backbonding interactions are maximized when R is
proximal to the forming C-C bond (proximal-TS3). Thus,
when this R group is small, both pathways leading to
5-substituted products (proximal-TS3-a and -TS3-b) are
computed to be lower in energy (entries 5 and 6, Table 6),
and because cleavage of either the more- or less-substituted
cyclopropyl bonds is reversible, several products are ex-
pected. For alkyne 4e (R ) CO2Me), all four pathways are
computed to have similar activation energies due to compet-
ing steric and electronic effects.
selective access to (3,4)-, (3,5)-, (2,5)-, and (2,4)-substituted
cycloheptenones.
Experimental Section
General Procedure for (5 + 2) Cycloadditions. In an oven-
dried, argon-purged glass test tube fitted with a rubber septum,
[Rh(CO)2Cl]2 catalyst (0.025 mmol, 9.7 mg) was dissolved in
freshly distilled DCE (2.5 mL) at room temperature to form a
homogeneous yellow solution. In a separate oven-dried, nitrogen-
flushed glass vial fitted with a rubber septum, vinylcyclopropane
(0.500 mmol, 106 mg) was dissolved in DCE (1.0 mL). Under
an inert atmosphere, this colorless solution was transferred with
rinsing (1.5 mL DCE) via syringe to the vessel containing the
stirring rhodium catalyst solution. Argon was bubbled through
a needle into the resulting yellow solution for five minutes before
addition of an alkyne (0.60-0.75 mmol) in a single portion via
syringe. Solid alkynes were dissolved in 1.0 mL of the total
DCE volume to be used and transferred under an inert
atmosphere. The reaction vessel was then immediately lowered
into a preheated 40 °C oil bath and allowed to stir under positive
pressure of argon for the indicated time. Upon consumption of
the VCP (as indicated by TLC analysis), the reaction solution
was cooled to room temperature and treated with 0.1 mL of a
1% HCl/EtOH solution. After hydrolysis of enol ether cycload-
duct(s) (generally 5-20 min), the reaction solution was passed
through a short pad of silica gel with Et2O eluent and
concentrated in Vacuo to a brown oil. Following initial GC-MS
Conclusions
The sense and magnitude of the regioselectivity preferences
for Rh(I)-catalyzed (5 + 2) cycloadditions of substituted
VCPs and substituted alkynes have been established experi-
mentally and analyzed theoretically (see Figure 4 for a
summary of outcomes). The regioselectivity is controlled by
the steric encumbrance of the alkyne substituent, moderated
or reversed in select cases by electronic effects. Bulky or
electron-rich alkyne substituents are postulated to orient distal
to the forming C-C bond during alkyne insertion to avoid
steric repulsions, leading to preferential formation of (3,5)-
and (2,5)-disubstituted cycloadducts 5 and 6, respectively,
for VCPs of type 1 and 2, over the (3,4)- and (2,4)-disubsti-
tuted regioisomers. Electron-withdrawing substituents on the
alkyne decrease or reverse this selectivity by stabilizing
the transition state in which the substituent is proximal to
the forming C-C bond in the alkyne insertion transition
states. The regiochemistry-determining alkyne insertion
transition states involving VCPs such as 3 are subject to these
same steric and electronic influences while providing syn-
thetically complementary products. These studies provide the
experimental and mechanistic foundation for understanding
and predicting regioselectivities in the Rh-catalyzed inter-
molecular (5 + 2) cycloaddition and related processes in
synthesis. Additionally, when considered with the intramo-
lecular (5 + 2) reactions, this work provides the basis for
1
and H NMR analysis, the crude product(s) were purified via
silica gel flash column chromatography (EtOAc/pentane or Et2O/
petroleum ether eluent). For reactions producing multiple
regioisomers, (5 + 2) cycloadducts were collected together
following chromatography, and a combined isolated yield was
reported. When feasible, further chromatography furnished
analytically pure samples of each regioisomer for spectroscopic
analysis. When TFE was used, it was added to the stirring
mixture of VCP and [Rh(CO)2Cl]2 in DCE (reduced volume)
prior to bubbling of argon through the reaction solution. Full
experimental details and characterization data can be found in
the accompanying Supporting Information.
Theoretical Calculations. Geometry optimizations, frequen-
cies, and solvation energy calculations were performed with the
B3LYP functional implemented in Gaussian 03.20 The Stuttgart/
Dresden effective core potential21 was used on rhodium, and
the 6-31G(d) basis set was employed for other atoms. All
(20) Frisch, M. J.; et al. Gaussian 03, Rev. D.01; Gaussian, Inc.:
Wallingford, CT, 2004.
9
10134 J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010

Rh(I)-Catalyzed Intermolecular (5 + 2) Cycloadditions: Regioselectivity
A R T I C L E S
reported free energies involve zero-point vibrational energy
corrections, thermal corrections to Gibbs free energy at 298 K,
and solvation free energy corrections computed by singlet point
CPCM22 calculations on gas-phase optimized geometries. Pa-
rameters for the solvent used in experiments, 1,2-dichloroethane
(ε ) 10.125), were used in the CPCM calculations. The
molecular cavities were built up using the United Atom
Topological Model (UAHF), which was reported to be more
accurate than cavity models such as UFF, Pauling, or Bondi in
calculating solvation free energies using CPCM.23 Figures 1 and
3 were prepared using CYLView.24
Acknowledgment. We thank the National Science Foundation
(CHE-0548209, CHE-0450638, and CHE-0131944), the NSF
Graduate Research Fellowship (L.E.S.), Deutsche Forschungsge-
meinschaft (H.R.), and Deutscher Akademischer Austausch Dienst
(I.V.H.) for financial support. The Vincent Coates Foundation Mass
Spectrometry Laboratory at Stanford University and TeraGrid
resources provided by NCSA are acknowledged for mass spectra
and computational facilities.
(21) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor.
Chim. Acta 1990, 77, 123–141.
(22) (a) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995–2001.
(b) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem.
2003, 24, 669–681.
(23) (a) Barone, V.; Cossi, M.; Tomasi, J. J. Chem. Phys. 1997, 107, 3210–
3221. (b) Houk, K. N.; Tahano, Y. J. Chem. Theory Comput. 2005,
1, 70–77. (c) Gao, D. Q.; Svoronos, P.; Wong, P. K.; Maddalena, D.;
Hwang, J.; Walker, H. J. Phys. Chem. A 2005, 109, 10776–10785.
(d) Yu, A; Liu, Y; Wang, Y. Chem. Phys. Lett. 2007, 436, 276–279.
(24) Legault, C. Y. CYLView, 1.0b; Universite´ de Sherbrooke: Que´bec,
Montreal, Canada, 2009; http://www.cylview.org.
Supporting Information Available: Complete experimental
details and characterization data for cycloadducts and prepara-
tion of VCPs. Complete reference 20 and optimized geometries
and energies of all computed species. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA103253D
9
J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010 10135