Gold and bronsted acid catalyzed hydride shift onto allenes: Divergence in product selectivity

Article abstract of DOI:10.1021/ja202336p

A series of allenyl ethers can be transformed into various fused or spiro tetrahydrofurans and tetrahydropyrans following a hydride shift/cyclization sequence. A divergence in product selectivity, which depends on the nature of the catalyst used (Au(I) complex or Bronsted acid), was observed.

Full text of DOI:10.1021/ja202336p

COMMUNICATION
pubs.acs.org/JACS
Gold and Brønsted Acid Catalyzed Hydride Shift onto Allenes:
Divergence in Product Selectivity
Benoit Bolte and Fabien Gagosz*
Dꢀepartement de Chimie, UMR 7652, CNRS/Ecole Polytechnique, 91128 Palaiseau, France
S Supporting Information
b
resulting vinyl-gold species 4 would then be functionalized in a
ABSTRACT: A series of allenyl ethers can be transformed into
various fused or spiro tetrahydrofurans and tetrahydropyrans
following a hydride shift/cyclization sequence. A divergence in
product selectivity, which depends on the nature of the catalyst
used (Au(I) complex or Brønsted acid), was observed.
final electrophilic demetalation step to furnish compound 5.
Interestingly, this process would be complementary to that
presented in eq 1 since it would allow a functionalization at the
central carbon C(2) of the allene. We decided to study the
feasibility of this process in an intramolecular hydroalkylation of
allenes using allenyl ethers of type 6 as substrates (eq 3). We
indeed conceived that a 6-exo gold activation of the allene in
compound 6 might induce a 1,5-hydride shift proceeding
through a six-membered transition state of type 7.4,5 The
resulting oxonium 8 could then be trapped by the pendant
vinyl-gold species to furnish hydroalkylation products.
uring the past 10 years, an impressive number of studies
D
have highlighted the utility of gold catalysis in organic
synthesis.1 A multitude of structural motifs of various nature
and complexity have thus been assembled using gold-catalyzed
transformations involving the addition of a nucleophile onto a
π-system such as an alkyne, allene, or alkene. Among the various
classes of transformations which have been developed, the
hydrofunctionalization of π-systems has attracted particular
attention.1 In the case of allenes, this process allows a rapid
and efficient access to a large variety of allylic derivatives 3 by a
formal inter- or intramolecular addition of the NuꢀH bond of a
carbon, oxygen, or nitrogen nucleophile across one of the CdC
bonds of the allene (eq 1).2 From a mechanistic point of view,
this transformation involves first a nucleophilic transfer of the
functional group Nu on the gold-activated allene 1 followed by a
protic demetalation of the intermediate vinyl-gold species 2.
The designed transformation was first attempted with model
allene 9 possessing a tetrahydrofuran moiety as the potential
hydride donor group (Table 1). Its treatment with 4 mol % of the
gold complex [(XPhos)Au(NCCH3)SbF6]6 12 in refluxing
CHCl3 for 4 h did not lead to an efficient reaction but validated
our proposal since the formation of spiro compound 10 could be
observed (9%, entry 1). Remarkably, the use of the phoshite gold
complex 13 in CH2Cl2 at 20 °C greatly improved the efficiency of
the reaction (entry 2). Under these conditions, the conversion of
allene 9 was rapid (0.5 h) and a mixture of spiro compound 10
and fused bicyclic compound 11 was obtained in in 30% and 61%
yields respectively.
Interestingly, the selectivity was reversed when the Brønsted
acid HNTf2 was used instead of gold-complex 13 (entry 3).
Under these conditions, the transformation was slower but
furnished exclusively compound 10 in an excellent 95% yield.
The reaction could also be performed using PtCl2 as the catalyst,
even if a longer reaction time and a higher temperature were
required (entry 4). The formation of compound 10 under
gold(I) or Brønsted acid catalysis is remarkable, since it proceeds
under very mild conditions in a stereoselective manner.7 Nota-
bly, two new contiguous asymmetric centers, one of them
quaternary, are formed during the process. The formation of
In conjunction with our recent investigations in the field of
gold-catalyzed hydride transfers,3 we were intrigued by the
possibility of developing a reverse polarization hydrofunctionaliza-
tion of allenes (eq 2). By contrast with a classical hydrofunctio-
nalization, such a process would first involve the transfer of a
hydrogen atom as a hydride onto the gold-activated allene 1. The
Received: March 15, 2011
Published: April 22, 2011
r
2011 American Chemical Society
7696
dx.doi.org/10.1021/ja202336p J. Am. Chem. Soc. 2011, 133, 7696–7699
|

Journal of the American Chemical Society
COMMUNICATION
Table 1. First Attempt of Hydroalkylation with Allene 9
Table 2. Substrate Scope: Cyclic Ethers as Hydride Donor
Groupsa
a Isolated yield. b NMR yield. c Ratio determined by 1H NMR spectros-
copy of the crude reaction mixture.
compound 11 is also of interest since it corresponds to the formal
insertion of the CꢀO bond of the tetrahydrofuran ring into the
C
C
(1)dC(2) bond of the allene combined with a migration of the
(2)dC(3) bond.
Given the novelty of this transformation and its synthetic
potential,8 we explored its scope using either 13 or HNTf2 as the
catalyst. As seen from the results compiled in Table 2, a series of
other allenes 14aꢀf could be used as substrates. In all cases, the
transformation proved to be efficient with overall yields in
hydroalkylated products ranging from 69% to 96%. The reaction
could be performed with substrates possessing another type of
linker (entries 1 and 2), a substituted tetrahydrofuran (entries
3ꢀ8), a differently substituted allene (entries 9 and 10),9 or a
tetrahydropyran as the hydride donor group (entries 11 and 12).
The same trend in selectivity was observed: the use of HNTf2 as
the catalyst led to the exclusive formation of spiro compounds 15
while the use of gold complex 13 generally furnished mixtures of
compounds 15 and 16, the latter being the major component.
Exceptionally, when allenes 14a and 14f were used as the
substrates, the gold catalysis led to the exclusive formation of
compounds 16a and 15f, which were isolated in respectively 91%
and 80% yields (entries 1 and 11). Notably, while the conversion
of 2-substituted tetrahydrofuran 14b into spiro compound 15b
was moderately selective (entries 3 and 4), the transformations of
3-substituted tetrahydrofurans 14c and 14d, used as diastereoi-
someric mixtures, were conversely extremely selective (entries
5ꢀ8).10 The corresponding spiro compounds 15c and 15d,
containing three asymmetric centers, were indeed obtained with
diasteroisomeric ratios greater than 24:1. In the gold-catalyzed
hydroalkylation of allene 14e, the conjugated diene 16e was the
major product instead of the regular fused bicyclic compound.11
A mechanistic proposal for the gold and Brønsted acid
catalyzed formations of spiro compounds 10 and/or fused
bicyclic compounds 11 from allene 9 is presented in Scheme 1.
The acid or gold activation of the allene moiety in substrate 9
induces a 1,5-hydride shift that produces oxonium 17, a common
intermediate in the formation of compounds 10 and 11. In the
gold-catalyzed process (A = AuLþ), intermediate 17 evolves into
spiro compound 18 by the direct reaction of the oxonium with
the CꢀAu bond.12,13 Compound 18 may also be produced in a
two-step sequence from oxonium 17 by elimination of the
a Unless otherwise noted, X = C(CO2Me)2. b Isolated yield. c Ratio
1
determined by H NMR spectroscopy of the crude reaction mixture.
d X = C(CH2OBn)2. e d.r. determined by 1H NMR spectroscopy.
cationic gold fragment in an intermediate carbocation 20. This
latter species could be formed by the alternative trapping of
oxonium 17 by the CdC bond of the vinyl-gold species. A gold-
catalyzed tetrahydrofuran ring-opening in spiro compound 18
produces an allylic carbocation that subsequently ring-closes to
produce oxepane 11. A proton loss/protodeauration sequence
on carbocation 20 via gold compound 21 would account for the
competitive formation of spiro compound 10.14
In the Brønsted acid catalyzed process (A = Hþ), oxonium 17
would be trapped by the pendant alkene to produce carbocation
20. A subsequent proton loss would furnish spiro compound
10.15 It is interesting to note that intermediate 18 (leading to
product 11) and product 10 possess isomeric structures that only
differ in the position of the CdC double bond. The generally
observed divergence in product selectivity might be explained by
the preferential direct and/or stepwise formation of the isopro-
pylidene intermediate 18 under gold catalysis, while, under
7697
dx.doi.org/10.1021/ja202336p |J. Am. Chem. Soc. 2011, 133, 7696–7699

Journal of the American Chemical Society
COMMUNICATION
Scheme 1. Mechanistic Proposal (X = C(CO2Me)2, A = Hþ
or LAuþ)
Scheme 3. Hydride Shift from Benzyl Ether 22
Table 3. Substrate Scope: Benzyl Ethers as Hydride Donor
Groups
Scheme 2. Hydroalkylation with Deuterium-Labeled Allene
9(D)
Brønsted acid catalysis, a regiospecific proton loss proceeding on
intermediate 20 would lead exclusively to the isopropenyl
compound 10. This mechanistic proposal is supported by the
reactions of deuterium-labeled allene 9(D) with gold catalyst 13
or HNTf2, which furnished compounds 10(D) and 11(D) with
the deuterium next to the oxygen in the tetrahydrofuran ring
being transferred to position C(3) of the newly formed six-
membered cycle (Scheme 2).16
We next attempted the hydroalkylation process on substrates
possessing a benzylether moiety as the hydride donor group.3a
We were delighted to see that a complete divergence in product
selectivity was observed when allene 22 was treated with either
gold complex 13 or HNTf2 (Scheme 2). Under gold catalysis,
tetrahydropyran 24 was obtained in 94% yield, while tetrahydro-
pyran 23 was produced in 84% yield when HNTf2 was used as the
catalyst.17 A mechanism analogous to that presented in Scheme 1,
involving a common oxonium intermediate 25, accounts for the
selective production of tetrahydropyrans 23 and 24 (Scheme 3).
The stereoselective formation of compound 23,18 containing
four asymmetric centers, two being formed during the hydro-
alkylation process, may be explained by considering the highly
ordered chairlike transition state 26 leading to carbocation 27
from oxonium 25. The relative trans relationship between the
a Isolated yield. b d.r. determined by 1H NMR spectroscopy.
phenyl and isopropenyl substituents in product 23 results from
the pseudoequatorial positions of the phenyl and isopropylidene
group in transition state 26. An analogous disposition accounts
7698
dx.doi.org/10.1021/ja202336p |J. Am. Chem. Soc. 2011, 133, 7696–7699

Journal of the American Chemical Society
COMMUNICATION
for the cis relationship between the phenyl group and the alkyl
(3) (a) Bolte, B.; Odabachian, Y.; Gagosz, F. J. Am. Chem. Soc. 2010,
132, 7294. (b) Odabachian, Y.; Dias-Jurberg, I.; Gagosz, F. J. Am. Chem.
Soc. 2010, 132, 3543.
substituent at carbon C(6)
.
The hydroalkylation process was further extended to substi-
tuted allenes 28aꢀe (Table 3). The transformations proved to be
efficient, and a range of tetrahydropyrans were obtained in yields
ranging from 69% to 95%. The gold catalysis produced exclu-
sively compounds of type 30, while the Brønsted acid catalysis
delivered only compounds of type 29 in a stereoselective manner.
Notably, the transformation could be performed with an aryl-
substituted benzyl ether substrate (entries 9 and 10).
In summary, we have shown that a range of allenyl ethers can
be transformed into various spiro tetrahydrofurans and tetrahy-
dropyrans following a hydride shift/cyclization sequence
catalyzed by a gold(I) complex or a Brønsted acid. This
transformation, which corresponds to a formal hydroalkylation
of an allene, proceeds under mild experimental conditions and is
applicable to substrates possessing various hydride donor groups.
It also represents a powerful method to stereoselectively convert a
secondary or tertiary sp3 CꢀH bond into a new CꢀC bond.19
Importantly, a clear-cut divergence in product selectivity was ob-
served when the reaction was catalyzed either by the gold complex or
by the Brønsted acid. Further studies related to gold and Brønsted
acid catalyzed hydride transfers onto π-systems are underway.
(4) No example of a direct gold-catalyzed hydride transfer onto an
allene has been reported. For examples of 1,5-hydride shifts observed in
gold(I)-catalyzed processes, see: (a) Harrak, Y.; Simonneau, A.; Malacria,
M.; Gandon, V.; Fensterbank, L. Chem. Commun. 2010, 46, 865. (b)
Jimꢀenez-Nꢀu~nez, E.; Raducan, M.; Lautenbach, T.; Molawi, K.; Solorio,
C. R.; Echavarren, A. M. Angew. Chem., Int. Ed. 2009, 48, 6152. (c) Cui,
Li; Peng, Y.; Zhang, L. J. Am. Chem. Soc. 2009, 131, 8394. (d) Bhunia, S.;
Liu, R.-S. J. Am. Chem. Soc. 2008, 130, 16488. (e) Zhou, G.; Zhang, J.
Chem. Commun. 2010, 46, 6593.
(5) For examples of Lewis acid catalyzed sequences of a hydride
transfer from an ether to an activated alkene followed by a
cyclization, see: (a) McQuaid, K. M.; Long, J. Z.; Sames, D. Org. Lett.
2009, 11, 2972. (b) McQuaid, K. M.; Sames, D. J. Am. Chem. Soc. 2009,
131, 402. (c) Pastine, S. J.; Sames, D. J. Am. Chem. Soc. 2005, 127, 12180.
(6) For the synthesis of gold complexes 12 and 13, see: Amijs,
C. H. M.; Lꢀopez-Carrillo; Raducan, V. M.; Pꢀerez-Galꢀan, P.; Ferrer, C.;
Echavarren, A. M. J. Org. Chem. 2008, 73, 7721.
(7) The stereochemistry of 10 was determined by 1H NMR
spectroscopy.
(8) The spiroether structural unit is found in a number of natural
products.
(9) The reaction could not be performed with monosubstituted
allenes.
(10) This selectivity may be explained by considering the possible
steric interactions between the substituent on the THF ring at C(3) and
the pendant alkene group in the chairlike transition state leading to
intermediate 20 (see Scheme 1).
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures and
b
spectral data for new compounds. This material is available free
of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
gagosz@dcso.polytechnique.fr
(11) Compound 16e corresponds to an open form of the bicyclic
compound generally obtained. Its formation probably results from a
gold-catalyzed oxepane ring opening.
(12) (a) Luzung, M. R.; Mauleꢀon, P.; Toste, F. D. J. Am. Chem. Soc.
2007, 129, 12402. (b) Zhang, L. J. Am. Chem. Soc. 2005, 127, 16804.
(13) The formation of intermediate 18 could not be observed when the
reaction of 9with gold catalyst 13 was monitored by 1H NMR spectroscopy.
(14) Compounds 10 and 11 remain unchanged when they were
separately treated with gold catalyst 13 thus precluding the formation of
11 from 10 and 10 from 11.
’ ACKNOWLEDGMENT
This article is dedicated with affection to the memory of our
friend and colleague Prof. Pascal Le Floch (Ecole Polytechnique).
The authors thank Prof. Samir Z. Zard for helpful discussions and
Rhodia Chimie Fine (Dr. F. Metz) for a generous gift of HNTf2.
(15) Under HNTf2 catalysis, and even after prolonged reaction
times, compound 10 was not transformed into compound 11 thus
precluding an isomerization of 10 into 18 via 20.
(16) The deuteration at carbon C(3) in products 10(D) and 11(D)
was equally shared between the two available positions. This specific
pattern tends to show that no memory of chirality is involved in the
hydroalkylation of allene 9. The nucleophilic trapping of oxonium 17
should proceed in this case with no facial selectivity.
’ REFERENCES
(1) For recent reviews on gold catalysis, see: (a) F€urstner, A. Chem.
Soc. Rev. 2009, 38, 3208. (b) Michelet, V.; Toullec, P. Y.; Gen^et, J. P.
Angew. Chem., Int. Ed. 2008, 47, 4268. (c) Jimꢀenez-Nꢀu~nez, E.;
Echavarren, A. M. Chem. Rev. 2008, 108, 3326. (d) Li, Z.; Brower, C.;
He, C. Chem. Rev. 2008, 108, 3239. (e) Arcadi, A. Chem. Rev. 2008,
108, 3266. (f) Gorin, D. J.; Toste, F. D. Chem. Rev. 2008, 108, 3351.
(g) Hashmi, A. S. K. Chem. Rev. 2007, 107, 3180. (h) F€urstner, A.;
Davies, P. W. Angew. Chem., Int. Ed. 2007, 46, 3410.
(2) For selected examples of gold-catalyzed hydrofunctionalization
of allenes, see: (a) Lavallo, V.; Frey, G.; Donnadieu, B.; Soleilhavoup, M.;
Bertrand, G. Angew. Chem., Int. Ed. 2008, 47, 5224. (b) Zhang, Z.;
Bender, C. F.; Widenhoefer, R. A. J. Am. Chem. Soc. 2007, 129, 14148.
(c) Zhang, Z.; Widenhoefer, R. Angew. Chem., Int. Ed. 2007, 46, 283. (d)
LaLonde, R. L.; Sherry, B. D.; Kang, E. J.; Toste, F. D. J. Am. Chem. Soc.
2007, 129, 2452. (e) Tarselli, M.; Chianese, A.; Lee, S.; Gagnꢀe, M.
Angew. Chem., Int. Ed. 2007, 46, 6670. (f) Zhang, Z.; Liu, C.; Kinder,
R. E.; Han, X.; Qian, H.; Widenhoefer, R. A. J. Am. Chem. Soc. 2006,
128, 9066. (g) Gockel, B.; Krause, N. Org. Lett. 2006, 8, 4485. (h)
Nishina, N.; Yamamoto, Y. Angew. Chem., Int. Ed. 2006, 45, 3314. (i)
Buzas, A.; Istrate, F.; Gagosz, F. Org. Lett. 2006, 8, 1957.
(17) For a (3,5) oxonium-ene type cyclization leading to tetrahy-
dropyrans of types 23 and 24, see: Loh, T.-P.; Hu, Q.-Y.; Tan, K.-T.;
Cheng, H.-S. Org. Lett. 2001, 3, 2669.
(18) No other diastereoisomer could be detected by 1H NMR
spectroscopy of the crude reaction mixture.
(19) For a recent review on gold-mediated CꢀH bond functiona-
lization, see: Boorman, T. C.; Larrosa, I. Chem. Soc. Rev. 2011, 40, 1910.
7699
dx.doi.org/10.1021/ja202336p |J. Am. Chem. Soc. 2011, 133, 7696–7699