Fascinating reactivity in gold catalysis: Synthesis of oxetenes through rare 4-exo-dig allene cyclization and infrequent β-hydride elimination

Article abstract of DOI:10.1039/c1cc13212a

A novel reactivity in gold catalysis, namely the unusual preference for the 4-exo-dig cyclization in allene chemistry as well as the rare β-hydride elimination reaction, was uncovered starting from readily available allenols.

Full text of DOI:10.1039/c1cc13212a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2011, 47, 9054–9056
www.rsc.org/chemcomm
COMMUNICATION
Fascinating reactivity in gold catalysis: synthesis of oxetenes through
rare 4-exo-dig allene cyclization and infrequent b-hydride eliminationw
a
c
Benito Alcaide,*a Pedro Almendros,*b Teresa Martı
´
nez del Campo and Israel Fernandez
´
Received 31st May 2011, Accepted 5th July 2011
DOI: 10.1039/c1cc13212a
A novel reactivity in gold catalysis, namely the unusual preference
for the 4-exo-dig cyclization in allene chemistry as well as the
rare b-hydride elimination reaction, was uncovered starting from
readily available allenols.
pathway.6,8 Owing to our interest in metal-catalyzed processes
employing alkynes and allenes,9 we were attracted to the
possibility that a non-five-membered heterocycle could be
accessed by variation of the allene substitution. We hypothesized
that the product selectivity could be impacted by modulating the
relative stability of the Z2 complexes generated by p-coordination
of the metal to the CQC bonds during the course of the reaction.
Starting allenols were prepared from salicylaldehyde
derivatives via regiocontrolled indium-mediated Barbier-type
carbonyl–allenylation reaction in aqueous media adopting our
methodology.10 In our study on the possibility of selectivity
reversal in Au catalysis, methyl-substituted allenol 1a and
phenyl-substituted allenol 1c were used as the model compounds
and AuCl3 as the catalyst to describe the experiments. The
reaction in dichloromethane proceeded smoothly for both sub-
strates. Allenol 1a afforded the expected cycloisomerization
adduct, dihydrofuran 2a. However, as a first try, we were happy
to notice that the reaction of phenyl-derivative 1c afforded
oxetene 3c as a major component (42% isolated yield), even
dihydrofuran 2c was also isolated as a minor component (17%).
Thus, when the reaction was performed using 1c instead of 1a,
a switch of the regioselectivity was observed. Encouraged by this
preliminary result and with the optimal reaction conditions
established, the scope of this novel oxycyclization reaction was
then studied. To our great delight, the reaction of phenyl-
derivatives 1d and 1e afforded the oxetenes 3d and 3e as the sole
products in good yields (Scheme 1). This intriguing transforma-
tion can be explained invoking an uncommon allene 4-exo-dig
cyclization as well as an infrequent b-hydride elimination reac-
tion in gold catalysis (C–Au bonds prefer protodeauration over
The transition-metal-catalyzed cyclization of allene derivatives
bearing a nucleophilic functionality has led to many synthetically
useful transformations.1,2 However, regioselectivity problems are
significant (endo-trig versus endo-dig versus exo-dig versus exo-trig
cyclization). The metal catalyst can coordinate to either allenic
double bond, depending on the regioselectivity of the attack both
on the structure of the substrate and the nature of the catalyst.
Four different products can be obtained, but the formation of
five- or six-membered rings is favored. The last decade has
witnessed dramatic growth in the number of reactions catalyzed
by gold complexes because of their powerful soft Lewis acidic
nature.3–5 Thus, it is not surprising that gold salts are ranked
among the best catalysts for the selective activation of the allene
moiety.6 An important, yet unexamined, subject in the field
of Au-catalyzed cyclizations is how to control the reaction
selectivity for the substrates carrying different allene substitution
patterns. Importantly, it should be remarked that reports on the
synthesis of strained rings such as four-membered heterocycles
are lacking in the gold chemistry of allenes. Besides, taking into
consideration the chemical and medicinal interest of oxetanes,7
the direct synthesis of oxetane derivatives from allenols emerged
as an attractive transformation to develop.
Despite the efficiency of the recently discovered gold-catalyzed
cycloisomerization reaction of a-hydroxyallenes to yield dihydro-
furans,6,8 the introduction of a wide range of substituents on the
allene group needs to be explored in order to expand its
application in the synthesis of useful chemicals. Traditionally,
metal-catalyzed cyclizations on a-allenols favor a 5-endo-trig
a Grupo de Lactamas y Heterociclos Bioactivos, Departamento de
´
´
´
Quımica Organica, Unidad Asociada al CSIC, Facultad de Quımica,
Universidad Complutense de Madrid, 28040-Madrid, Spain.
E-mail: alcaideb@quim.ucm.es; Fax: +34-91-3944103
b
c
´
Instituto de Quımica Organica General, IQOG, CSIC,
Juan de la Cierva 3, 28006-Madrid, Spain.
E-mail: Palmendros@iqog.csic.es; Fax: +34-91-5644853
´
´
´
´
Departamento de Quımica Organica, Facultad de Quımica,
Universidad Complutense de Madrid, 28040-Madrid, Spain
w Electronic supplementary information (ESI) available: Experimental
procedures, characterization data of new compounds, and copies of
NMR spectra. See DOI: 10.1039/c1cc13212a
Scheme 1 Divergent oxycyclization reactions of a-allenols 1 under
gold catalysis.
c
9054 Chem. Commun., 2011, 47, 9054–9056
This journal is The Royal Society of Chemistry 2011

View Article Online
b-hydride elimination; indeed, gold-hydrides are rare species
and difficult to access).11–13
migration11 leads to the formation of oxetane species 9.
Uncommon b-hydride elimination rather than protonolysis
of the carbon–gold bond12–14 linked to a reaction of HCl with
the gold hydride would then liberate the oxacycle type 3 with
concomitant regeneration of the catalytic Au(III) salt. In the
conversion of intermediate 9 into alkylidene-oxetenes 3, a
possible alternative step to the b-hydride elimination should
involve a direct transfer of the hydride to the proton of HCl,
thus re-forming AuCl3 and H2.15
The incorporation of different substituents into the salicyl-
aldehyde-derived allenol was tested next. As revealed in
Scheme 1, various aryl-substituted allenols 1f–i were suitable
for such heterocyclization reaction. For example, the p-methoxy-
phenyl substituent at the allene side was tolerated. Similarly,
the ether chain bearing methoxy, (E)-(4-chlorobut-2-enyl)oxy,
or propargyloxy moieties also afforded the desired oxetenes,
but not as the unique product. Starting from allenols 1c, and
1f–h, and performing the reaction at room temperature, a
mixture of two different products arises from competitive
4-exo-dig versus 5-endo-trig cyclizations (Scheme 1). Interestingly,
allenols 1c and 1f–h under AuCl3 catalysis in dichloromethane
at reflux temperature gave as the sole or major product the
oxetene component (Scheme S1, ESIw). Thus, it is possible to
suppress or minimize the formation of the dihydrofuran
ring (cycloisomerization adduct) by performing the reaction
at higher temperature, yielding the oxetene (oxycyclization/
dehydrogenation adduct) as the exclusive or major product. A
general trend can be deduced on the basis of these results: the
four-membered oxetene is the thermodynamic control product
while the dihydrofuran is the kinetic control product.
A coordination to the internal C–C double bond would
similarly lead to intermediate 9 (see Scheme S4, ESIw).
To gain more insights into the reaction mechanisms of the
above gold-catalyzed oxycyclization reactions, a computational
(DFT) study was carried out. The corresponding computed
reaction profiles (PCM-B3LYP/def2-SVP level)16 of methyl-
and phenyl-substituted allenols 1j and 1f are shown in
Scheme 3, which gather the respective free energies in CH2Cl2
solution.
As initially envisaged, two different coordination modes of
the metal fragment to the allenic double bond, i.e. distal vs.
proximal, are possible. Our calculations indicate that although
the proximal approach is slightly more exergonic, the subsequent
4-endo-dig oxycyclization reaction (from complex 10 via
transition state TS4) is clearly kinetically disfavored with
respect to the corresponding 5-endo-trig or 4-exo-dig processes
in view of the much higher activation barrier of the former
transformation. Thus, the oxycyclization reactions start
from the common intermediate complex 4, formed through
coordination of AuCl3 to the distal allenic double bond of 1j, f.
As readily seen in Scheme 3, the 5-endo-trig reaction forms
the zwitterionic complex 5 through an exergonic process
(DGR,298=ꢀ9.5 and ꢀ9.9 kcal molꢀ1 for R = Me and
R = Ph, respectively) via TS1 with a very low activation
A possible pathway for the gold-catalyzed achievement of
dihydrofurans 2 may initially involve the formation of a
complex 4 through coordination of the gold trichloride to the
distal allenic double bond of a-allenols 1. Next, regioselective
5-endo-trig oxyauration forms zwitterionic intermediates 5,
which after loss of HCl generate neutral species 6. Protonolysis
of the carbon–gold bond of 6 liberates adducts 2 with concurrent
regeneration of the Au(III) catalytic species (Scheme 2, left
catalytic cycle).
A mechanistic rationale for the gold-catalyzed conversion of
aryl-substituted allenols 1 into oxetenes 3 is more intricate. It
is worth noting that the cyclization affords cycloadducts 3
from a 4-exo-dig cyclization/dehydrogenation process instead
of that from the usually preferred 5-endo-trig cycloisomerization
reaction. The pathway proposed in Scheme 2 (right catalytic
cycle) looks valid for the formation of products type 3. It
could be presumed that the initially formed gold complex 4,
through coordination of the AuCl3 to the distal allenic double
bond, undergoes an intramolecular attack (rare 4-exo-dig
versus normal 5-endo-trig oxyauration) by the hydroxy group,
giving rise to the oxetene intermediate 7. Loss of HCl in
intermediate 7 generates neutral species 8, which after 1,3-gold
barrier (DGa = +3.3 and +4.1 kcal molꢀ1). The sequential
298
loss of HCl, which forms neutral species 6, and protonolysis
of the C–Au bond via TS3 produce the final dihydrofuran 2
regenerating the catalyst. Interestingly, the barrier energy of
this step is clearly more favorable when R = Me than when
R = Ph (DDGa298=6.0 kcal molꢀ1), which is in nice agreement
with the preferential formation of dihydrofurans 2 at expenses
of oxetenes 3 for methyl-substituted allenols 1a, b (see Scheme 1).
The 4-exo-dig process, which transforms complex 4 into 7
via TS2, also occurs with a low activation barrier (DGa
=
298
+7.6 and +7.8 kcal molꢀ1 for R = Me and R = Ph,
respectively) in an exergonic transformation as well. Despite
that, the barrier energies are higher than those computed for
the 5-endo-trig process, thus indicating that the latter reaction
is kinetically favored. Nevertheless, the low barriers of both
processes should be translated into a mixture of reaction
products, as experimentally found (see Scheme 1). Strikingly,
the computed reaction energy difference between both nucleo-
philic additions indicates that whereas the 5-endo-trig reaction
is thermodynamically favored when R = Me (DDGR,298
=
+5.1 kcal molꢀ1), the 4-exo-dig reaction is favored when
R = Ph (DDGR,298 = ꢀ2.2 kcal molꢀ1). This justifies the
formation of oxetenes 3 as the major reaction product when
the reaction is conducted under reflux conditions (see Scheme S1,
ESIw). Loss of HCl in intermediate 7 generates neutral
species 8. This step is much easier than the corresponding loss
Scheme 2 Mechanistic explanation for the gold-catalyzed oxycyclization
of methyl- and aryl-substituted a-allenols 1.
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 9054–9056 9055

View Article Online
Scheme 3 Computed reaction profile for the gold-catalyzed oxycyclization reactions of allenols 1j (R = Me, plain values) and 1f (R = Ph, values
in parentheses). Free energies (DG298, in kcal molꢀ1) have been computed at the PCM(CH2Cl2)-B3LYP/def2-SVP level.
of HCl which transforms 5 into 6 due to close proximity of the
hydrogen and chlorine atoms in complex 7 (computed Hꢁ ꢁ ꢁCl
distance of 1.822 and 1.830 A for R = Me and R = Ph,
2 (a) B. Alcaide, P. Almendros and C. Aragoncillo, Chem. Soc. Rev.,
2010, 39, 783; (b) M. Brasholz, H.-U. Reissig and R. Zimmer, Acc.
Chem. Res., 2009, 42, 45; (c) B. Alcaide and P. Almendros, Eur. J.
Org. Chem., 2004, 3377; (d) A. S. K. Hashmi, Angew. Chem., Int.
Ed., 2000, 39, 3590.
respectively). Complex
8
isomerizes into 80-bis after
3 For selected examples, see: (a) A. S. K. Hashmi, Angew. Chem., Int.
Ed., 2010, 49, 5232; (b) A. S. K. Hashmi, Chem. Rev., 2007, 107, 3180.
4 (a) B. Lipshutz and Y. Yamamoto (ed.), Chem. Rev., 2008, 108(8);
(b) G. J. Hutchings, M. Brust and H. Schmidbaur (ed.), Chem. Soc.
Rev., 2008, 37(9).
re-coordination of the metal fragment and is converted into
complex 9 through TS5 (a saddle point associated with the
1,3-migration of the metal moiety).11 Finally, a b-hydride
elimination reaction occurs (via TS6)16 to produce the final
oxetene 3 and AuHCl2 (which would regenerate the initial
catalyst upon reaction with HCl). From the data in Scheme 3,
it becomes obvious that this step constitutes the bottle-neck of
the process when R = Ph (DGa298 = +21.1 kcal molꢀ1) but is
kinetically favored over the corresponding Au–C protonolysis
involving TS3 and leading to dihydrofuran 2f.
5 (a) A. Corma, A. Leyva-Perez and M. J. Sabater, Chem. Rev.,
´
2011, 111, 1657; (b) J. Muzart, Tetrahedron, 2008, 6, 5815.
6 (a) N. Krause and C. Winter, Chem. Rev., 2011, 111, 1994;
(b) N. Bongers and N. Krause, Angew. Chem., Int. Ed., 2008, 47, 2178.
7 J. A. Burkhard, G. Wuitschik, M. Rogers-Evans, K. Muller and
¨
E. M. Carreira, Angew. Chem., Int. Ed., 2010, 49, 9052.
8 (a) A. Hoffmann-Roder and N. Krause, Org. Lett., 2001, 3, 2537;
¨
(b) A. S. K. Hashmi, M. C. Blanco, D. Fischer and J. W. Bats, Eur.
J. Org. Chem., 2006, 1387; (c) M. Asikainen and N. Krause, Adv.
Synth. Catal., 2009, 351, 2305; (d) D. Eom, D. Kang and P. H. Lee,
J. Org. Chem., 2010, 75, 7447.
9 See, for instance: (a) B. Alcaide, P. Almendros and R. Carrascosa,
Chem.–Eur. J., 2011, 17, 4968; (b) B. Alcaide, P. Almendros and
Therefore, it can be concluded that whereas the 5-endo-
trig - loss of HCl - protonolysis sequence (which produces
dihydrofurans 2) is followed for methyl-substituted
allenols, the 4-exo-dig - loss of HCl - 1,3-Au-migration -
b-elimination pathway (which leads to oxetenes 3) is preferred
for phenyl-substituted allenols. Both processes share the same
common intermediate 4, formed through distal coordination
of the allenic double bond, and are strongly favored over the
corresponding 4-endo-dig pathway involving the proximal
coordination.
M. T. Quiro
P. Almendros and T. Martı
2007, 46, 6684.
10 B. Alcaide, P. Almendros and T. Martı
Chem., Int. Ed., 2006, 45, 4501.
´
s, Adv. Synth. Catal., 2011, 353, 585; (c) B. Alcaide,
´
nez del Campo, Angew. Chem., Int. Ed.,
´
nez del Campo, Angew.
11 A. S. K. Hashmi, A. M. Schuster, S. Litters, F. Rominger and
M. Pernpointner, Chem.–Eur. J., 2011, 17, 5661.
12 A. S. K. Hashmi and G. J. Hutchings, Angew. Chem., Int. Ed.,
2006, 45, 7896.
Support for this work by the MICINN [CTQ2009-09318,
CTQ2010-20714-C02-01, and Consolider-Ingenio 2010
(CSD2007-00006)], CAM (Projects S2009/PPQ-1752 and
S2009/PPQ-1634), UCM-Santander (Grant GR35/10-A) is
13 B. Alcaide, P. Almendros, T. Martınez del Campo, E. Soriano and
´
J. L. Marco-Contelles, Chem.–Eur. J., 2009, 15, 9127.
14 H. Ito, T. Saito, T. Miyahara, C. Zhong and M. Sawamura,
Organometallics, 2009, 28, 4829.
gratefully acknowledged. I.F. is a Ramon y Cajal fellow.
´
15 Despite that this mechanistic scenario cannot be completely ruled
out, it seems less unlikely taking into account the computed high
activation barrier of this particular step (see Scheme S2, ESIw).
Besides, the preparation of oxetene 3d from allenol 1d wasalso
accomplished using Pt(II) catalysis (see Scheme S3, ESIw), probably
involving the formation of well known platinum hydride species. It
can be suggested that the hydride transfer to HCl occurs from
AuHCl2 to form H2 regenerating AuCl3.
Notes and references
1 For selected reviews, see: (a) Cumulenes and Allenes, , Science of
Synthesis, Houben-Weyl Method of Molecular Transformation, ed.
N. Krause, George Thieme, Sttutgart, 2007, vol. 44; (b) S. Ma,
Chem. Rev., 2005, 105, 2829; (c) Modern Allene Chemistry, ed.
N. Krause and A. S. K. Hashmi, Wiley-VCH, Weinheim, 2004.
16 See computational details in the ESIw.
c
9056 Chem. Commun., 2011, 47, 9054–9056
This journal is The Royal Society of Chemistry 2011