Enantioselective synthesis of cis-fused cyclooctanoids via rhodium(I)-catalyzed [4 + 2 + 2] cycloadditions

Article abstract of DOI:10.1021/acs.orglett.5b00267

Catalytic multicomponent [m + n + o]-type cycloadditions offer efficient, atom-economical routes to diverse complex carbocycles. Recently, such transformations have emerged as unique strategies for medium ring carbocycle synthesis. Despite the important d

Full text of DOI:10.1021/acs.orglett.5b00267

Letter
pubs.acs.org/OrgLett
Enantioselective Synthesis of cis-Fused Cyclooctanoids via
Rhodium(I)-Catalyzed [4 + 2 + 2] Cycloadditions
Brendan C. Lainhart and Erik J. Alexanian*
Department of Chemistry, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, United States
S
* Supporting Information
ABSTRACT: Catalytic multicomponent [m + n + o]-type cyclo-
additions offer efficient, atom-economical routes to diverse complex
carbocycles. Recently, such transformations have emerged as unique
strategies for medium ring carbocycle synthesis. Despite the
important developments in this area, however, highly enantioselective
[m + n + o]-type processes accessing medium ring carbocycles have
yet to be developed. Herein, a rhodium-catalyzed [4 + 2 + 2]
cycloaddition of allenedienes with allenes enabling the direct stereoselective synthesis of cis-fused cyclooctanoids is reported.
These cycloadditions are successful with a diverse range of π-components and demonstrate the potential for high levels of
enantioselectivity in a [4 + 2 + 2] process.
atalytic, multicomponent [m + n + o]-type cycloadditions
are applicable to the concise synthesis of diverse carbo-
Table 1. Rh-Catalyzed [4 + 2 + 2] Cycloaddition of
Allenediene 1
C
cycles and heterocycles. Foremost among these cycloadditions
are [2 + 2 + 2] processes, with particularly attractive applications
in the preparation of benzenoid systems.1 Recent work has
extended the use of multicomponent cycloadditions to the
synthesis of medium-ring systems via [3 + 2 + 2] and [4 + 2 + 2]
approaches.2 These transformations offer unique strategies
toward seven- and eight-membered carbocycles that remain a
formidable synthetic challenge.3 However, in contrast to [2 + 2 +
2] processes, enantioselective variants of these cycloadditions are
not known.4 The successful development of reactions accom-
plishing this goal would constitute powerful tools for the direct,
enantioselective synthesis of these challenging carbocycles.
We have previously reported Rh- and Ni-catalyzed [2 + 2 + 2]
cycloadditions of alkenes and allenes for the construction of
stereochemically complex cyclohexanoids, including enantiose-
lective variants.5 Herein, the development of a Rh-catalyzed [4 +
2 + 2] cycloaddition of allenedienes with allenes for the
stereoselective construction of cis-fused cyclooctanoids is
reported. Through the use of chiral phosphoramidites as ligands
in our catalytic system, we have successfully developed the first
examples of highly enantioselective multicomponent cyclo-
additions for the synthesis of medium ring carbocycles.
Our initial efforts targeted the identification of an effective
catalytic system for the [4 + 2 + 2] cycloaddition of allenediene 1
and benzyl allenoate 2 (Table 1). Phosphoramidite MonoPhos
(L1) proved to be an excellent ligand for the reaction and, in
conjunction with 2.5 mol % [Rh(coe)2Cl]2, provided cyclo-
octanoid 3 in 70% yield as a single diastereomer (entry 1). The
simple intramolecular [4 + 2] cycloaddition of the allenediene
substrate was not observed under these conditions.6 Reactions
using either standard monodentate phosphines (PPh3), or
bidentate phosphines (BINAP) in the presence or absence of
AgOTf provided no cycloadduct (entries 2−4). Increasing the
amount of MonoPhos in the catalytic system led to a decrease in
a
Isolated yield.
yield (entry 5). Reactions in the absence of ligand or metal both
provided no product (entries 6 and 7).
Prior to our studies of enantioselective variants of the [4 + 2 +
2], we elected to perform an initial survey of the substrate scope
(Table 2). The cycloaddition was successful using a variety of
allenediene tether units (entries 1−3). The identity of the cis-
fused cyclooctanoid 7 was confirmed by X-ray crystallography
(Figure 1). Importantly, substrates benefiting from the Thorpe−
Ingold effect were not required in the process, as demonstrated
by the reaction of methylene-tethered substrate 8, which
proceeded in 75% yield (entry 4). Substitution of the diene at
the 2-position in allenediene 10 provided cycloadduct 11 in
moderate yield (49%); we also observed allenediene [4 + 2]
Received: January 27, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.5b00267
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Table 2. Scope of Rhodium-Catalyzed [4 + 2 + 2] Allenediene-
a
Allene Cycloaddition
Figure 1. ORTEP diagram of cycloaddition product 7 (thermal
ellipsoids set at 50% probability).
products in moderate yield as single diastereomers. The reaction
of dienoate 16 demonstrated the viability of electron-poor dienes
in the [4 + 2 + 2] process. We next sought to investigate the
scope of the cycloaddition with respect to substitution of the
allene components. Cycloaddition of 1,1-disubstituted allene-
diene 18 produced cyclooctanoid 19 containing a quaternary
stereogenic center as a single diastereomer (entry 10). The
exogenous allene is not limited to benzyl allenoate, as
demonstrated by the reactions of entries 11−13 involving allenes
containing aliphatic, aromatic, and heteroatom substitution,
respectively. Furthermore, cycloaddition of 1 with 1,1-disub-
stituted allenoate 26 delivered cyclooctanoid 27 containing a
tetrasubstituted alkene as a single stereoisomer.
We next turned our focus to the development of a highly
enantioselective variant of the [4 + 2 + 2] process (Table 3). Our
initial studies involved substrate 28, as 2-substitution of the diene
component had been shown to increase levels of undesired [4 +
Table 3. Optimization Studies of Enantioselective Rh-
Catalyzed [4 + 2 + 2] Allenediene-Allene Cycloaddition
a
Reactions performed using 1 equiv of allenediene and 2 equiv of
allene in the presence of 2.5 mol % [Rh(coe)2Cl]2 and 6 mol %
P(OEt)3 in PhMe at 100 °C for 72 h. X = C(CO2Et)2 unless otherwise
b
c
d
noted. Isolated yield. 6 mol % MonoPhos used as ligand. Reaction
temperature 130 °C.
cycloaddition (28%), however. This selectivity issue was easily
addressed by substituting P(OEt)3 for MonoPhos as ligand. The
intramolecular allenediene [4 + 2] process was not observed in
any cases using P(OEt)3 as ligand.
Our studies continued with allenediene substrates containing
substitution at the 1- and 3-positions of the diene (entries 7−9),
which were also well tolerated, providing the cyclooctanoid
B
DOI: 10.1021/acs.orglett.5b00267
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
2] cycloaddition (e.g., Table 1, entry 5). Our rationale was that
this choice would allow for the simultaneous determinination of
reaction enantioselectivity and chemoselectivity with a challeng-
ing allenediene substrate. We were pleased to discover that the
reaction of 28 with allenoate 2 using (R)-MonoPhos as ligand
provided [4 + 2 + 2] product 29 in a highly enantioselective
fashion (97.5:2.5 er, entry 1). The cyclooctanoid product was
formed in moderate yield (39%), however, with a significant
amount of 30 produced. Increasing the reaction temperature led
to a predominance of the [4 + 2] pathway (entry 2), and lowering
the reaction temperature further (e.g., 40 °C) resulted in low
conversions. Encouraged by these initial results, we next
performed experiments involving modified catalytic systems.
The addition of silver salts as halide abstractors increased the
reaction chemoselectivity at the expense of enantioselectivity
(entries 3 and 4). A survey of phosphoramidite ligands was
performed, and altering substitution on the amine moiety with
the bulkier piperidine (L2) and methylbenzylamine (L3)
moieties maintained reaction enantioselectivity, but decreased
product selectivity (entries 5 and 6). A large increase in the diol
dihedral angle (L4) provided no improvement over L1 (entry 7);
however, a modest increase in dihedral angle with (S)-H8-
MonoPhos (L5) led to a somewhat improved level of product
selectivity (47% yield of cyclooctanoid), while maintaining high
enantioselectivity (entry 8). Solvent modifications had minimal
effect on the reaction outcome (entries 9 and 10).
Table 4. Enantioselective Rh-Catalyzed [4 + 2 + 2]
Cycloadditions of Allenedienes with Allene 2
With our optimized conditions for the enantioselective [4 + 2
+ 2] in hand, we studied a range of allenediene substrates (Table
4). Increasing the steric bulk of the 2-position of the diene from
isopropyl (entry 1) to trimethylsilyl (entry 2) did not reduce the
reaction enantioselectivity, but did somewhat decrease chemo-
selectivity. Cycloaddition of −OTBS diene 33 delivered silyl enol
ether product 34 in modest yield (and high enantioselectivity); in
this case the cyclooctanoid was the major product (entry 3).
Decreasing the size of the 2-substituent (entry 4) noticeably
improved product selectivity, but at the expense of enantiose-
lectivity (85:15 er). While there are limitations with respect to
the diene substitution in these enantioselective reactions, the
vinyl silane and silyl enol ether functionality of products 32 and
34 are expected to enable access to a variety of derivatives via
standard synthetic manipulations.7
a
b
Isolated yield. Calculated by 1H NMR spectroscopy of crude
c
reaction mixtures using an internal standard. Reaction temperature
100 °C.
The reaction of allenediene 18, containing a 1,1-disubstituted
allene, yielded [4 + 2 + 2] product 19 with high chemoselectivity
but modest enantioselectivity (entry 5). Interestingly, a
synergistic effect was observed when this allene substitution
pattern was combined with 2-substitution on the diene. In the
event, reaction of 35 delivered cyclooctanoid 36, bearing a
quaternary center at the ring fusion, in high enantioselectivity
(97.5:2.5). The enantioselectivity of this process was higher than
the reactions of either substituent in isolation (entries 4 and 5).
In order to gain insight into a potential reaction mechanism,
we sought to verify whether the undesired [4 + 2] cycloadduct
was formed with an identical level of enantioselectivity as the [4 +
2 + 2] product. This would be consistent with a common
enantiodetermining step. However, the [4 + 2] pathway
proceeded with lower enantioselectivity than the [4 + 2 + 2]
pathway (77.5:22.5 vs 96.5:3.5, Figure 2). The removal of
exogenous allene from the system provided a high yield of [4 + 2]
product 39 in nearly racemic fashion. These results suggest that
the enantiodetermining steps of the two cycloadditions are not
identical and that the added allene component is a factor in
determining the enantioselectivity of both pathways.
Figure 2. Enantioselectivity of [4 + 2] pathway.
Given the complete cis-diastereoselectivity of these cyclo-
additions, we propose that both the [4 + 2 + 2] and [4 + 2]
pathways proceed via a cis-fused metallacyclopentane inter-
mediate (B and B′, Figure 3). The cis-diastereoselectivity of the
allenediene oxidative coupling has previously been attributed to
the higher strain energy of a trans-fused metallacycle.6 We
C
DOI: 10.1021/acs.orglett.5b00267
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(2) (a) Inglesby, P. A.; Evans, P. A. Chem. Soc. Rev. 2010, 39, 2791.
(b) Gilbertson, S. R.; DeBoef, B. J. Am. Chem. Soc. 2002, 124, 8784.
(c) Evans, P. A.; Baum, E. W. J. Am. Chem. Soc. 2004, 126, 11150.
(d) Evans, P. A.; Baum, E. W.; Fazal, A. N.; Pink, M. Chem. Commun.
2005, 63. (e) Wender, P. A.; Christy, J. P. J. Am. Chem. Soc. 2006, 128,
5354. (f) Hilt, G.; Janikowski, J. Angew. Chem., Int. Ed. 2008, 47, 5243.
(g) Yamasaki, R.; Terashima, N.; Sotome, I.; Komagawa, S.; Saito, S. J.
Org. Chem. 2010, 75, 480. (h) Saya, L.; Bhargava, G.; Navarro, M. A.;
Gulías, M.; Lop
Angew. Chem., Int. Ed. 2010, 49, 9886. (i) Canlas, G. M. R.; Gilbertson, S.
R. Chem. Commun. 2014, 50, 5007. (j) Araya, M.; Gulías, M.; Fernandez,
I.; Bhargava, G.; Castedo, L.; Mascarenas, J. L.; Lop
ez, F. Chem.Eur. J.
́
ez, F.; Fernandez, I.; Castedo, L.; Mascarenas, J. L.
́
̃
́
́
̃
2014, 20, 10255.
(3) (a) Lop
́
ez, F.; Mascarenas, J. L. Chem.Eur. J. 2007, 13, 2172.
̃
(b) Mehta, G.; Singh, V. Chem. Rev. 1999, 99, 881. (c) Butenschon, H.
Angew. Chem., Int. Ed. 2008, 47, 5287. (d) Yu, Z.-X.; Wang, Y.; Wang, Y.
̈
Chem.Asian J. 2010, 5, 1072.
Figure 3. Mechanistic rationale for inequivalent enantioselectivities in
[4 + 2 + 2] and [4 + 2] pathways.
(4) Using an alkyne-alkyne-diene [4 + 2 + 2], Gilbertson has reported a
moderately enantioselective reaction (70.5:29.5 er) yielding an eight-
membered carbocycle: DeBoef, B.; Counts, W. R.; Gilbertson, S. R. J.
Org. Chem. 2007, 72, 799. Rovis has reported a highly enantioselective
approach to azocine heterocycles via a [4 + 2 + 2] of dienyl isocyanates
and alkynes: Yu, R. T.; Friedman, R. K.; Rovis, T. J. Am. Chem. Soc. 2009,
131, 13250.
propose that two pathways with inequivalent enantioselectivities
lead to the cycloadducts formed. The first route involves an
enantioselective oxidative coupling of allene-bound coordination
complex A to form metallacycle B. This pathway subsequently
produces the [4 + 2 + 2] and [4 + 2] products with high
enantioselectivity. Alternatively, a pathway solely involving the
allenediene can simultaneously deliver the [4 + 2] product with
poor enantioselectivity. Consistent with this hypothesis is the
observed increase in enantioselectivity of the [4 + 2] process with
added allene and the near racemic product formed in its absence
(Figure 2).
In conclusion, a Rh-catalyzed [4 + 2 + 2] cycloaddition of
allenedienes and allenes for the construction of fused cyclo-
octanoids is reported. This catalytic cycloaddition is successful
with a diverse set of diene and allene components and provides
unique 5−8 cis-fused carbocycles in a highly stereoselective
fashion. The use of phosphoramidite ligands has enabled the
development of asymmetric variants of this process, which
represent the first examples of highly enantioselective multi-
component cycloadditions in medium ring carbocycle synthesis.
(5) (a) Noucti, N. N.; Alexanian, E. J. Angew. Chem., Int. Ed. 2013, 52,
8424. (b) Brusoe, A. T.; Edwankar, R. V.; Alexanian, E. J. Org. Lett. 2012,
14, 6096. (c) Brusoe, A. T.; Alexanian, E. J. Angew. Chem., Int. Ed. 2011,
50, 6596.
(6) Wender, P. A.; Jenkins, T. E.; Suzuki, S. J. Am. Chem. Soc. 1995, 117,
1843.
̀
(7) (a) Fleming, I.; Dunogues, J.; Smithers, R. The Electrophilic
Substitution of Allylsilanes and Vinylsilanes. In Organic Reactions; John
Wiley & Sons, Inc.: New York, 2004; pp 57−575. (b) Oshima, K.
Vinylsilanes. In Science of Synthesis; Fleming, I., Ed.; Thieme: Stuttgart,
Germany, 2002; Vol. 4, pp 713−754. (c) Kobayashi, S.; Manabe, K.;
Ishitani, H.; Matsuo, J.-I. Silyl Enol Ethers. In Science of Synthesis;
Fleming, I., Ed.; Thieme: Stuttgart, Germany, 2002; Vol. 4, pp 317−369.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures and spectral data for all new
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: eja@email.unc.edu.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support was provided by the NSF (CAREER award
CHE-1054540). Acknowledgement is also made to the donors of
the ACS PRF (50245-DNI) for partial support.
REFERENCES
■
(1) (a) Tanaka, K., Ed. Transition-Metal-Mediated Aromatic Ring
Construction; John Wiley & Sons: Hoboken, NJ, 2013. (b) Broere, D. L.
́
J.; Ruitjer, E. Synthesis 2012, 44, 2639. (c) Domínguez, G.; Perez-
Castells, J. Chem. Soc. Rev. 2011, 40, 3430. (d) Nakamura, I.; Yamamoto,
Y. Chem. Rev. 2004, 104, 2127.
D
DOI: 10.1021/acs.orglett.5b00267
Org. Lett. XXXX, XXX, XXX−XXX