The role of bent acyclic allene gold complexes in axis-to-center chirality transfers

Article abstract of DOI:10.1002/anie.200802332

(Chemical Equation Presented) Activation of allenes toward nucleophilic attack by gold complexes ([Au], see scheme) often proceeds with axis-to-center chirality transfer. The stereochemical information is shown to be maintained in not only η² a

Full text of DOI:10.1002/anie.200802332

Communications
DOI: 10.1002/anie.200802332
Gold Catalysis
The Role of Bent Acyclic Allene Gold Complexes in Axis-to-Center
Chirality Transfers**
Vincent Gandon,* Gilles Lemi›re, Alexandra Hours, Louis Fensterbank,* and Max Malacria*
Dedicated to Professor Jan Reedjik
Electrophilic activation of readily available enantioenriched
allenes by transition-metal fragments is a maturing strategy to
generate optically active compounds (axis-to-center chirality
transfer).[1] The coordination of an allene to a metal may lead
to several types of structures that can be divided into two
categories (Scheme 1).[2] The h2 complexes I involving one of
allene deviate significantly. Compounds of type II’’ have been
recently isolated as tetraaminoallene rhodium[5] or gold
complexes[6] and characterized by X-ray diffraction studies.[7]
Interestingly, gold is known to activate allenes toward
nucleophilic attack.[8] For instance, a- and b-hydroxyallenes
or -aminoallenes can be cycloisomerized in the presence of
gold(I) or gold(III) catalysts into the corresponding 5- and 6-
membered heterocycles.[9] These transformations have also
been accomplished with axis-to-center transfer of chirality,[10]
even in intermolecular versions.[11] Thus, the question whether
the nucleophile attacks a species of type I–I’’ or II–II’’ is of
paramount importance in accounting for the chirality trans-
fer. Indeed, although the stereochemical information is
maintained in species I–I’’, the axial chirality of the allene
seems to be lost in II or II’. On the other hand, bent allene
complexes should retain the chirality of the starting material.
We ran a set of computations[12] on model chiral allene gold
complexes to understand which factors govern the ground
state of I–II’’, and how these species interconvert.[13] We then
used our findings to predict which structural variations of the
allene could favor a successful chirality transfer. Finally, we
verified our hypothesis on some gold(I)-catalyzed cyclo-
isomerizations.
Scheme 1. Various coordination modes of allenes. M=metal–ligand
fragment.
=
the two orthogonal C C bonds are the most intuitive form.
Depending on the substitution pattern of the allene, the
contribution of the two carbon atoms to the coordination
might not be strictly equivalent, leading to slipped structures
of type I’ or I’’.[3] This case will manifest itself for instance in
the presence of electron-donating groups that will stabilize I’,
or electron-withdrawing groups that favor I’’. The second
category comprises species in which the metal fragment is
coordinated to the central allene carbon only. Along with s-
allylic cations II, there are alternatives, such as zwitterionic
carbenes II’ or h1-coordinated bent allenes II’’.[4] In this type
of complex, the three allene carbons and the metal fragment
lie in the same plane, from which the four substituents of the
We began our study with (R)-1,3-dimethyl allene. Figure 1
shows all the structures that converged as minima in the
presence of AuBr3.[14] We found two diastereomeric com-
plexes of type I, and two C2-coordinated complexes of type II
and II’’.[15] The type II complex is planar, and has C1 C2 and
ꢀ
ꢀ
C2 C3 bond lengths of 1.39 Š, and a C1-C2-C3 angle of 1168.
ꢀ
ꢀ
The latter has shorter C1 C2 and C2 C3 bond lengths
(1.36 Š), and is strongly twisted (Me-C1-C3-Me 65.88) to
minimize the allylic strain. This complex is the less stable
form, the ground state being one of the type I species (DH298
=
3.5 kcalmolꢀ1).
[*]Dr. V. Gandon, G. Lemi›re, Dr. A. Hours, Prof. L. Fensterbank,
Prof. M. Malacria
We investigated a few other metal fragments (Table 1).
Laboratoire de Chimie Organique UMR 7611, Institut de Chimie
MolØculaire FR2769, UniversitØ Pierre et Marie Curie Paris 6
Tour 44-54, 4 place Jussieu, 75252 Paris (France)
Fax: (+33)1-44-27-73-60
With the exception of Au+, no gold(I) complex of type II’’
could be located.[15] Using PtCl2, the C1 C2 and C2 C3 bond
lengths (1.39 Š), the small C1-C2-C3 angle (129.18), and the
moderate Me-C1-C3-Me tilt angle (45.18) compare quite well
with a distorted allylic cation. On the other hand, the allene
ligand is less severely bent in AuX3 complexes (C1-C2-C3
ꢀ
ꢀ
E-mail: vincent.gandon@upmc.fr
louis.fensterbank@upmc.fr
[**]We are grateful to the CNRS, MRES, IUF, and ANR BLAN 06-
2_159258 for financial support of this work, and to the CRIHAN,
Plan InterrØgional du Bassin Parisien (project 2006-013), for the use
of the computing facilities. We thank Prof. F. McDonald (Emory
University) for helpful suggestions on propargyl alcohol resolution,
and Dr. N. Goasdoue (UPMC) for circular dichroism measure-
ments.
=
140.18 and 136.68 with X = Br and Cl, respectively), the C C
bonds are much less perturbed (1.36 Š), and the tilt angles are
much more pronounced (65.88 and 59.78). However, in all
cases (as for AuBr3; see Figure 1), the ground states are
complexes of type I. For instance, with [Au(PMe3)]+, the
allene complexes 1 and 2 were more stable than the allylic
cation 3 (Scheme 2). Whereas 3 was the only C2-coordinated
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200802332.
7534
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 7534 –7538

Angewandte
Chemie
Scheme 2. Interconvertion between gold complexes of 1,3-dimethyl
allene. [Au]={Au(PMe3)}. Relative enthalpies in kcalmolꢀ1. Barrier
heights are indicated in parentheses.
Figure 1. Structures of (R)-1,3-dimethyl allene complexed to AuBr3.
Distances in Š, and relative enthalpies in kcalmolꢀ1
.
observed with some unfunctionalized disubstituted allenes.[10]
However, we reasoned that the presence of an electron-
donating group at the allene could modify this trend. The
compensation of the electron depletion induced by the metal
could balance the steric hindrance at the allylic positions. To
this end, we replaced a methyl group by an acetate or a vinyl
group. The choice of these substituents was dictated by the
numerous applications of gold-catalyzed transformations of
allenyl esters[8] or vinyl allenes.[16]
Scheme 3 shows all possible complexes with an acetate at
the allene. Eight of the twelve candidates converged as
minima.[17] The allylic cation 8 is the ground state of the
system. The participation of the oxygen atom in the con-
Table 1: Structural parameters of some bent allene complexes.
ꢀ
C1 C2 [Š]C1-C2-C3 [
8]Me-C1-C3-Me [
8]
[a]
Au(PR3)+
AuCl
–
–
–
–
–
[a]
–
AuBr3
AuCl3
Au+
1.36
1.36
1.36
1.39
140.1
136.6
130.0
129.1
65.8
59.7
48.7
45.1
PtCl2
[a]The corresponding h2-allene complex was obtained.
structure that could be optimized as a minimum, the other
missing geometries were obtained as transition states. The 908
shift of the gold atom on going from (R)-1 to (R)-2 is achieved
through a bent allene transition state and requires an enthalpy
of activation of 5.7 kcalmolꢀ1. The most hindered transition
state, TS(R)1-(S)1, is planar and allows the inversion of the
allene. However, it lies 18.2 kcalmolꢀ1 above (R)-1. On the
other hand, the stereomutation of (R)-2 occurs at a less
hindered planar transition state lying 9.8 kcalmolꢀ1 above the
complex. Apart from these two racemization processes, the
chirality can also be lost when forming the planar intermedi-
ate 3, through a bent allene type transition state lying
7.9 kcalmolꢀ1 above 2.
All these data suggest that there is a continuum between
these species which depends on the substituents of the allene
and on the metal fragment. A subtle balance between the
delocalization of the positive charge and the inherent allylic
strain seems to govern the possible geometries. With a simple
unfunctionalized allene, coordination to the central carbon
atom seems to be disfavored. This could prevent the
racemization, and may account for the good chirality transfers
Scheme 3. Various coordination modes of (R)-buta-1,2-dienyl acetate
to gold(I). [Au]={Au(PMe3)}. Relative enthalpies in kcalmolꢀ1
.
ꢀ
ꢀ
ꢀ
Selected structural parameters for 15: C1 C2 1.35, C2 C3 1.38, C3 O
1.33 Š, C1-C2-C3 1338. [a]These species did not converge and col-
lapsed to the corresponding planar or twisted C2-coordinated com-
pound.
Angew. Chem. Int. Ed. 2008, 47, 7534 –7538
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7535

Communications
ꢀ
jugation is evident from the short C3 O distance of 1.31 Š.
All other isomers are at least 3.0 kcalmolꢀ1 less stable. It is
worthy of note that two other C2-coordinated allenes are
more stable (10) or of similar energy (15) than the most stable
h2-allene complex 4.
trisubstituted allenes (Scheme 6). Nevertheless, the ground
state still is a C2-coordinated compound, that is, 21 or 24.
From this set of results, it seems obvious that the allylic
strain can prevent the formation of planar structures, and may
All the species are connected by low-lying transition
states; two examples are shown in Scheme 4. The 908 rotation
of the CHOAc group to transform 5 into 8 has an enthalpy of
Scheme 6. Selected examples of gold complexes of hexa-1,3,4-triene.
[Au]={Au(PMe3)}. Relative enthalpies in kcalmolꢀ1
.
retard the racemization of allenes (see for instance (R)-1!
(S)-1 in Scheme 2). The formation of C2-coordinated allenes,
either as allylic cations, bent allenes, or any other resonance
form (such as pentadienyl cations in the case of vinyl allenes),
is favored by the presence of electron-donating groups.
Although cationic in nature, these species retain the stereo-
chemical information of the substrate. Therefore, a reaction
that requires the formation of C2-coordinated allenes may
give a better chirality transfer as the substitution of the allene
increases. One such reaction could be the gold(I)-catalyzed
cycloisomerization of ene–vinyl allenes.[13b] This transforma-
tion is known to give polycyclic compounds diastereoselec-
tively (Scheme 7). The mechanism has been previously
studied by DFT computations: the first step is a metalla-
Nazarov cyclization which gives a cyclopentenylidene inter-
mediate; this gold carbene is then trapped by the pendant
Scheme 4. Interconvertion between gold complexes of (R)-buta-1,2-
dienyl acetate. [Au]={Au(PMe3)}. Relative enthalpies in kcalmolꢀ1
,
barrier heights indicated in parentheses.
activation of 2.4 kcalmolꢀ1. Interestingly, the 458 shift of the
metal from the h2 allene complex 7 to the bent allene complex
15 is straightforward (DH°298 = 0.1 kcalmolꢀ1). Importantly,
the stereochemistry of the bent allene complex 15 may also
invert via a planar transition state.
The same feature was observed with vinyl allenes
(Scheme 5). The allylic cation 17 was obtained as the
ground state. Again, the conjugation was obvious from the
C3 C4 bond length (1.43 Š). In contrast with the 1,3-
C C bond to furnish the cyclopropyl ring.[13b,18] Interestingly,
ꢀ
=
dimethyl allene system, the bent structures 18 and 19 could
be modeled, although they are slightly less stable than the h2-
coordinated allene. The geometrical parameters of these
species, and in particular the quite acute C1-C2-C3 angle of
123.58 (18) and 126.38 (19), suggest that they are close to
regular allylic cations, although they are not planar at all. No
more planar structures can be modeled in the case of
the calculations showed that the metalla-Nazarov transition
state connects the gold-carbene to a species complexed at C2
Scheme 5. Selected examples of gold complexes of hexa-1,3,4-triene.
[Au]={Au(PMe3)}. Relative enthalpies in kcalmolꢀ1. Selected structural
ꢀ
ꢀ
ꢀ
ꢀ
parameters for 18: C1 C2 1.37, C2 C3 1.39, C3 C4 1.44, C4 C5
1.36 Š, C1-C2-C3 123.5, H-C1-C3-C4 48.78.
Scheme 7. Cycloisomerization of ene–vinyl allenes. R=H, Me, OAc;
Alk=alkenyl group.
7536
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 7534 –7538

Angewandte
Chemie
of the starting allene framework. That species is not planar
and should keep the chiral information.[19]
Of the four possible bent allene structures, only two are
reactive with respect to the Nazarov cyclization, that is, those
with the vinyl group anti to the metal fragment. Interestingly,
any of these two diastereomers should lead to the same
enantiomer after the cyclization, therefore an efficient trans-
fer of the stereochemical information is expected (see
Figure 2 for an example with R = OAc). However, to
decrease the rate of racemization of the allene, a third
substituent at the allene might be necessary (R ¼
6
H).
Scheme 8. Gold(I)-catalyzed cycloisomerization of enantioenriched
ene-vinyl allenes. [a]The enantiomeric excess could not be determined.
necessitate the presence of a vinyl group. Perhaps the idea of
the formation of a chiral h1-complex as reactive intermediate
which is cationic in nature could be extended to other
reactions, such as the cycloisomerization of allyl allenes.[23]
This reaction has been achieved with very good retention of
the stereochemical information. In this case, it is no longer
possible to invoke a pentadienyl cation. The cyclization would
occur on a chiral type I-I’’ or II’’ intermediate, which will be
the object of future study. At this stage, we can say that the
formation of C2-coordinated allene complexes is always
possible, that these species are not necessarily planar allylic
cation but can be twisted and therefore chiral, and that they
can be considered as reactive species as good as h2-allene
complexes.
Figure 2. Cycloisomerizations of diastereomeric allene gold complexes
leading to the same enantiomer. Distances in Š, relative enthalpies in
kcalmolꢀ1. M=[AuPMe3]+.
To verify this hypothesis, we attempted the gold(I)-
catalyzed cycloisomerization of enantioenriched ene–vinyl
allenes. We synthesized the enantioenriched compounds
depicted in Scheme 8. In (S)-26, the allene moiety is
disubstituted. Gold(I)-catalyzed cycloisomerization furnished
the tetracyclic compound 27 in 87% yield and 0% enantio-
meric excess. Chiral GC analysis after 10 minutes revealed
that the starting compound was completely racemized. A
similar loss of stereochemical information was made during
the gold(I)-catalyzed cycloisomerization of vinyl allenes into
cyclopentadienes.[16c] On the other hand, the reaction of the
trisubstituted allenes (R)-28 and (R)-30 proceeded with
perfect transfer of chirality. Trisubstituted allenes (allenyl
acetates) can also be generated in situ by the well-established
gold-catalyzed 3,3-rearrangement of propargyl acetates.[20]
Under the same experimental conditions, efficient chirality
transfers were observed during the cycloisomerization of the
esters (R)-32 and (R)-34.[21]
Received: May 19, 2008
Published online: August 1, 2008
Keywords: allenes · cycloisomerization ·
.
density functional calculations · gold · homogeneous catalysis
[1] For recent reviews on stereoselective gold catalysis involving
allenes, see: a) N. Bongers, N. Krause, Angew. Chem. 2008, 120,
2208; Angew. Chem. Int. Ed. 2008, 47, 2178; b) R. A. Widen-
hoefer, Chem. Eur. J. 2008, 14, 5382.
[2] For reviews on transition metal–allene complexes, see: a) T. L.
Jacobs in The Chemistry of the Allenes, Vol. 2 (Ed.: S. R.
Landor), Academic Press, London, 1982, pp. 277; b) B. L. Shaw,
A. J. Stringer, Inorg. Chim. Acta Rev. 1973, 7, 1.
[3] a) O. Eisenstein, R. Hoffmann, J. Am. Chem. Soc. 1981, 103,
4308; b) H. M. Senn, P. E. Blöchl, A. Togni, J. Am. Chem. Soc.
2000, 122, 4098.
[4] For experimental evidence of the coordination of copper, silver,
and gold to the allene central carbon, see: J. H. B. Chenier, J. A.
Howard, B. Mile, J. Am. Chem. Soc. 1985, 107, 4190.
Thus, with a reaction that transits through central allene
carbon complex, the chirality can still be transferred. The
reactive species could be described as pentadienyl cations of
helical chirality,[22] and argued that the representation as a
bent allene complex is convoluted. However, we have shown
that the formation of chiral bent allene complexes does not
Angew. Chem. Int. Ed. 2008, 47, 7534 –7538
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7537

Communications
[5] a) C. A. Dyker, V. Lavallo, B. Donnadieu, G. Bertrand, Angew.
[12] M. J. Frisch et al., Gaussian 03, Gaussian, Inc., Pittsburgh, PA,
2003. DFT/B3LYP/LACVP(d,p) calculations. This methodology
gave a very good agreement between computed and X-ray
structures of the bent allene complexes reported in Refs. [5] and
[6] (see Supporting Information).
Chem. 2008, 120, 3250; Angew. Chem. Int. Ed. 2008, 47, 3206;
b) O. Kaufhold, F. E. Hahn, Angew. Chem. 2008, 120, 4122;
Angew. Chem. Int. Ed. 2008, 47, 4057.
[6] A. Fürstner, M. Alcarazo, R. Goddard, C. W. Lehmann, Angew.
Chem. 2008, 120, 3254; Angew. Chem. Int. Ed. 2008, 47, 3210.
[7] Bent allenes can also exist as such: see Ref. [5] and references
therein. For theoretical evidence, see: R. Tonner, G. Frenking,
Angew. Chem. 2007, 119, 8850; Angew. Chem. Int. Ed. 2007, 46,
8695. For another recent example, see: T. Yamaguchi, Y.
Yamamoto, D. Kinoshita, K.-a. Akiba, Y. Zhang, C. A. Reed,
D. Hashizume, F. Iwasaki, J. Am. Chem. Soc., DOI: 10.1021/
ja710423d.
[8] For recent reviews on gold-mediated homogeneous catalysis,
see: a) V. Michelet, P. Y. Toullec, J.-P. GenÞt, Angew. Chem.
2008, 120, 4338; Angew. Chem. Int. Ed. 2008, 47, 4268; b) S. A.
Hashmi, Chem. Rev. 2007, 107, 3180; c) D. J. Gorin, F. D. Toste,
Nature 2007, 446, 395; d) A. Fürstner, P. W. Davies, Angew.
Chem. 2007, 119, 3478; Angew. Chem. Int. Ed. 2007, 46, 3410;
e) C. H. M. Amijs, C. Ferrer, A. M. Echavarren, Chem.
Commun. 2007, 698; f) E. JimØnez-Nfflæez, A. M. Echavarren,
Chem. Commun. 2007, 333; g) A. R. Chianese, S. J. Lee, M. R.
GagnØ, Angew. Chem. 2007, 119, 4118; Angew. Chem. Int. Ed.
2007, 46, 4042; h) L. Zhang, J. Sun, S. A. Kozmin, Adv. Synth.
Catal. 2006, 348, 2271; i) M. Malacria, J.-P. Goddard, L.
Fensterbank in Comprehensive Organometallic Chemistry,
3rd ed. (COMC-III), Vol. 10 (Eds.: R. Crabtree, M. Mingos),
Elsevier, Amsterdam, 2006, chap. 10.07, pp. 299.
[13] For recent computational studies involving h2 allene-gold com-
plexes, see: a) A. Correa, N. Marion, L. Fensterbank, M.
Malacria, S. P. Nolan, L. Cavallo, Angew. Chem. 2008, 120,
623; Angew. Chem. Int. Ed. 2008, 47, 613; b) G. Lemi›re, V.
Gandon, K. Cariou, T. Fukuyama, A.-L. Dhimane, L. Fenster-
bank, M. Malacria, Org. Lett. 2007, 9, 2207; c) G. Lemi›re, V.
Gandon, N. Agenet, J.-P. Goddard, A. de Kozak, C. Aubert, L.
Fensterbank, M. Malacria, Angew. Chem. 2006, 118, 7758;
Angew. Chem. Int. Ed. 2006, 45, 7596.
[14] AuBr3 has been used as catalyst for intermolecular hydro-
amination of allenes resulting in chirality transfer: see
Ref. [11a].
[15] Failed attempts led to allene complexes of type I, and no
complex of type II’ could be found.
[16] see Ref. [13b], and: a) L. Zhang, S. Wang, J. Am. Chem. Soc.
2006, 128, 1442; b) H. Funami, H. Kusama, N. Iwasawa, Angew.
Chem. 2007, 119, 927; Angew. Chem. Int. Ed. 2007, 46, 909;
c) J. H. Lee, F. D. Toste, Angew. Chem. 2007, 119, 930; Angew.
Chem. Int. Ed. 2007, 46, 912.
[17] 9 and 10 are not strictly planar, but very close to planarity. Their
“racemizations” require an enthalpy of activation of only 0.1 and
0.01 kcalmolꢀ1, respectively.
[18] F.-Q. Shi, X. Li, Y. Xia, L. Zhang, Z.-X. Yu, J. Am. Chem. Soc.
2007, 129, 15503.
[19] As no enantioenriched allenes were used in Ref. [13b], the
discussion therein was potentially misleading, suggesting a
planar nature of the allene complexes.
[20] For recent examples, see Ref. [13b] and: a) C. H. Oh, A. Kim, W.
Park, D. I. Park, N. Kim, Synlett 2006, 2781; b) S. Wang, L.
Zhang, J. Am. Chem. Soc. 2006, 128, 14274; c) H.-S. Yeom, S.-J.
Yoon, S. Shin, Tetrahedron Lett. 2007, 48, 4817; d) K. Cariou, E.
Mainetti, L. Fensterbank, M. Malacria, Tetrahedron 2004, 60,
9745.
[21] The relative arrangement was determined by NOE experi-
ments[13b] and on the basis of an X-ray analysis (see Supporting
information). The absolute configuration of 35 was confirmed by
circular dichroism after hydrolysis into the corresponding ketone
(see Supporting information).
[9] See, for example: R. Zriba, V. Gandon, C. Aubert, L. Fenster-
bank, M. Malacria, Chem. Eur. J. 2008, 14, 1482, and references
therein.
[10] For selected examples, see: a) A. Hoffmann-Röder, N. Krause,
Org. Lett. 2001, 3, 2537; b) N. Morita, N. Krause, Org. Lett. 2004,
6, 4121; c) B. Gockel, N. Krause, Org. Lett. 2006, 8, 4485; d) A.
Buzas, F. Istrate, F. Gagosz, Org. Lett. 2006, 8, 1957; e) C. T. J.
Hyland, L. S. Hegedus, J. Org. Chem. 2006, 71, 8658; f) N.
Morita, N. Krause, Eur. J. Org. Chem. 2006, 4634; g) N. T. Patil,
L. M. Lutete, N. Nishina, Y. Yamamoto, Tetrahedron Lett. 2006,
47, 4749; h) Z. Zhang, C. Liu, R. E. Kinder, X. Han, H. Qian,
R. A. Widenhoefer, J. Am. Chem. Soc. 2006, 128, 9066; i) N.
Morita, N. Krause, Angew. Chem. 2006, 118, 1930; Angew. Chem.
Int. Ed. 2006, 45, 1897; j) Z. Liu, A. S. Wasmuth, S. G. Nelson, J.
Am. Chem. Soc. 2006, 128, 10352; k) C. Deutsch, B. Gockel, A.
Hoffmann-Röder, N. Krause, Synlett 2007, 1790.
[22] O. N. Faza, C. S. Lꢀpez, R. ꢁlvarez, A. R. de Lera, J. Am. Chem.
Soc. 2006, 128, 2434.
[23] A. Buzas, F. Gagosz, J. Am. Chem. Soc. 2006, 128, 12614.
[11] a) N. Nishina, Y. Yamamoto, Angew. Chem. 2006, 118, 3392;
Angew. Chem. Int. Ed. 2006, 45, 3314; b) Z. Zhang, R. A.
Widenhoefer, Org. Lett. 2008, DOI: 10.1021/ol800646h.
7538
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 7534 –7538