Gold(I)-catalyzed diastereo- and enantioselective 1,3-dipolar cycloaddition and mannich reactions of azlactones

Article abstract of DOI:10.1021/ja1095045

Azlactones participate in stereoselective reactions with electron-deficient alkenes and N-sulfonyl aldimines to give products of 1,3-dipolar cycloaddition and Mannich addition reactions, respectively. Both of these reactions proceed with good to excellent diastereo- and enantioselectivity using a single class of gold catalysts, namely C₂-symmetric bis(phosphinegold(I) carboxylate) complexes. The development of the azlactone Mannich reaction to provide fully protected anti-α,β-diamino acid derivatives is described. 1,3-Dipolar cycloaddition reactions of several acyclic 1,2-disubstituted alkenes and the chemistry of the resultant cycloadducts are examined to probe the stereochemical course of this reaction. Reaction kinetics and tandem mass spectrometry studies of both the cycloaddition and Mannich reactions are reported. These studies support a mechanism in which the gold complexes catalyze addition reactions through nucleophile activation rather than the more typical activation of the electrophilic reaction component.

Full text of DOI:10.1021/ja1095045

ARTICLE
pubs.acs.org/JACS
Gold(I)-Catalyzed Diastereo- and Enantioselective 1,3-Dipolar
Cycloaddition and Mannich Reactions of Azlactones
Asa D. Melhado,# Giovanni W. Amarante,# Z. Jane Wang, Marco Luparia, and F. Dean Toste*
Department of Chemistry, University of California, Berkeley, California, 94720, United States
S Supporting Information
b
ABSTRACT: Azlactones participate in stereoselective reactions with electron-deficient alkenes and N-sulfonyl aldimines to give
products of 1,3-dipolar cycloaddition and Mannich addition reactions, respectively. Both of these reactions proceed with good to excellent
diastereo- and enantioselectivity using a single class of gold catalysts, namely C2-symmetric bis(phosphinegold(I) carboxylate) complexes.
The development of the azlactone Mannich reaction to provide fully protected anti-R,β-diamino acid derivatives is described. 1,3-Dipolar
cycloaddition reactions of several acyclic 1,2-disubstituted alkenes and the chemistry of the resultant cycloadducts are examined to probe
the stereochemical course of this reaction. Reaction kinetics and tandem mass spectrometry studies of both the cycloaddition and Mannich
reactions are reported. These studies support a mechanism in which the gold complexes catalyze addition reactions through nucleophile
activation rather than the more typical activation of the electrophilic reaction component.
whether phosphinegold(I) carboxylate could be employed as
catalysts for generating nucleophilic reactive intermediates similar
to those proposed by Hayashi and Ito.
1. INTRODUCTION
In their seminal contribution to enantioselective Lewis acid
catalysis, Hayashi and Ito reported an aldol-type reaction between
aldehydes 1 and isocyanate 2 using a bifunctional catalyst containing
both a phosphinegold(I) moiety and an amine (eq 1).1 The authors
proposed transition state A, wherein both the electrophile and
nucleophile simultaneously interact with the catalyst. This early
precedent not withstanding activation of nucleophilic species is not
commonly invoked in gold catalysis. The majority of recent develop-
ments in homogeneous gold catalysis involve reactions which are
proposed to be initiated by coordination of a cationic gold species
with C-C π-bonds to form an activated electrophile.2,3 Therefore
the utility of gold complexes would be significantly extended if they
could be employed as catalysts for transformations that are not
predicated on π-bond activation.
In this context, we were attracted to the potential of our
gold carboxylate complexes to activate azlactones as nucleophiles.
Azlactones participate in a wide variety of transformations allowing
ready access not only to structurally complex amino acid derivatives
but also to highly substituted heterocycles.5 Recent efforts have
focused on the use of azlactones as substrates in catalytic, often
stereoselective, reactions.6 Reactive intermediates derived from
azlactones have recently been employed as nucleophiles in a number
of transformations including Pd-catalyzed arylation7a and allyla-
tion7c-j and organocatalytic conjugate addition6b-d,f,l,k and Man-
nich reactions.8
Additionally, Tepe reported the silver(I) acetate-catalyzed
reaction of azlactones with electron-deficient alkenes to afford
Δ1-pyrrolines and proposed the reaction to be a [3 þ 2] dipolar
cycloaddition proceeding via a metalated m€unchnone intermedi-
ate (eq 2).9 We viewed this as an attractive opportunity to test
our hypothesis that gold carboxylates could activate azlactones as
nucleophiles that would participate in enantioselective addition
We recently developed chiral bis(phosphinegold(I) carboxylate)
complexes as catalysts for the intramolecular hydroamination of
allenes; exploiting the soft, carbophilic nature of cationic gold(I) as a
means of effecting π-bond electrophile activation.4 These carbox-
ylate complexes are reminiscent of the catalyst used by Hayashi and
Ito in that both types of catalyst combine a phosphinegold(I)
moiety with a weak organic base. Thus we were interested to explore
Received: October 21, 2010
Published: February 22, 2011
r
2011 American Chemical Society
3517
dx.doi.org/10.1021/ja1095045 J. Am. Chem. Soc. 2011, 133, 3517–3527
|

Journal of the American Chemical Society
ARTICLE
reactions. Herein we provide a full account of our studies that
resulted in the development of gold-catalyzed enantioselective
[3 þ 2] dipolar cycloaddition10 and Mannich reactions of
azlactones. Furthermore, we also demonstrate kinetic, labeling
and electrospray ionization mass spectrometry (ESI-MS) experi-
ments concerning the reaction mechanisms of both the gold-
catalyzed Mannich and 1,3-dipolar cycloaddition reactions.
Table 1. Enantioselective Gold(I)-Catalyzed 1,3-DCR
of Azlactones with Maleimides/Maleic Anhydride
2. RESULTS AND DISCUSSION
2.1. Development of the Enantioselective Dipolar Cy-
cloaddition Reaction (DCR) of Azlactones with Electron-
Deficient Alkenes. We initially explored the triphenylphosphi-
negold(I) benzoate-catalyzed cycloaddition of azlactone 3a with
maleimide 4 and were pleased to find that cycloadduct 5a was
formed in 86% yield after only 0.5 h (eq 3). Optimization of the
enantioselective reaction revealed two major factors impacting
the enantioselectivity. First, the use of phosphine ligands sub-
stituted with sterically bulky groups was essential to obtaining high
enantioselectivities.11 The improvement from 40% to 83% ee, when
the parent (R)-SEGPHOS(AuOBz)2 and was replaced with (R)-
DTBM-SEGPHOS(AuOBz)2 as a catalyst, is illustrative (eq 3).
Ultimately, (S)-Cy-SEGPHOS(AuOBz)2, (S)-(6), proved to be
the most general catalyst for the enantioselective cycloaddition.
a Isolated yield unless otherwise noted. b Determined by enantiodiscri-
c
minating HPLC. Run at 0.5 M in 3/1 PhF/THF. d Run at 0.5 M in
acetone, using (R)-DTBM-SEGPHOS(AuOBz)2. e Run at 0.25 M in 3/
1 PhF/THF at rt, with 5 mol % (S)-6. f Run at 0.25 M in 3/1 PhF/THF,
at 0 °C with 5 mol % (S)-6. g Run at 0.04 M in PhF, yield by 1H NMR
i
against internal standard. h Using (R)-6. Run at 0.25 M in 3/1 PhF/
THF at -20 °C using (R)-6.
A second notable improvement was achieved by changing the
solvent from THF to fluorobenzene. In this solvent, (S)-(6)
catalyzed the formation of 5a in 95% ee (Table 1, entry 1). The
scope of the gold(I)-catalyzed enantioselective 1,3-dipolar cy-
cloaddition reaction of azlactones with electron-deficient alkenes
is summarized in Tables 1 and 2 in order to facilitate the overall
discussion.
Table 2. Enantioselective Gold(I)-Catalyzed Reactions
of Azlactones and Monosubstituted Alkenesa
The reaction shows excellent scope in terms of the azlactone
substituents (Table 1) and the alkene dipolarophile (Table 2). In
one case where the standard conditions did not afford the cycload-
duct with the desired enantioselectivity, we found that the selectivity
could be improved by either using DTBM-SEGPHOS(AuOBz)2 as
a catalyst (Table 1, entries 7 vs 8). Additionally, while sterically
demanding azlactone substituents resulted in notably decreased
reaction rates in fluorobenzene, running the reaction in a 3:1 mixture
of fluorobenzene:THF restored the reaction rate without significant
impact on the enantioselectivity. The exception was the dialkylsub-
stituted substituted substrate, 2,4-dimethyloxazolin-5-one (3o), which
underwent the desired cycloaddition with N-phenylmaleimide (4)
under the standard conditions with poor enantioselectivity (Table 1,
entry 16). Subsequently, we found that the selectivity of this reaction
could be significantly improved by lowering the reaction temperature.
We obtained cycloadduct 5o in 64% ee by carrying out the reaction at
0°C and in 78% ee at -20 °C (Table 1, entries 17 and 18). The scope
of the reaction was successfully extended to include acrylate esters and
acrylonitrile under very similar conditions; catalyst (S)-6 enabled the
formation of cycloadducts 8a-e with generally high levels of regio-,
diastereo-, and enantioselectivity (Table 2).
entry
product
time (h) yield (%)b ee (%)
99c
95
90
93
76
t
1
2
3
4
5
8a, X = CO2 Bu R = OMe
24
14
14
14
14
56
74
t
8b, X = CO2 Bu R = NHCH2Ph
8c, X = CO2Et R = OMe
8d, X = CO2Me R = OMe
66
89d
8e, X = CN
R = NHCH2Ph
68
a Reactions run at 0.5 M in 3a. b Isolated yield unless otherwise noted.
c Ten equiv. of tert-butyl acrylate used. d Yield determined by 1H NMR
analysis of crude product against an internal standard.
Since 2003, a variety of catalytic systems have been described
for the enantioselective synthesis of substituted pyrrolidines
through the 1,3-dipolar cycloaddition of acyclic R-iminoesters;
generally these methods provide endo-cycloadducts.12,13 In con-
trast, the cycloaddition of azlactone derived dipoles gives rise to
Δ1-pyrrolines and therefore offer the potential for manipulation
of the endocyclic imine.14 For example, palladium-catalyzed
hydrogenation of the imine gave pyrrolidine exo-9 (eq 4).
3518
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527

Journal of the American Chemical Society
ARTICLE
This is complementary to the result generally obtained from the
acyclic dipole precursor.
(Table 2, entry 4). The ester hydrolysis/decarboxylation
sequence is consistent with the facile nature of the H/D exchange
discussed above; however, an attempt to exploit the carbon acid
nature of this system via alkylation (NaH/THF, followed by
addition of allyl bromide) did not proceed smoothly.
When dimethyl fumarate was reacted with 3a and subjected to
the same sequence described above, diastereomers 8d (major)
and epi-8d (minor) were obtained as an inseparable mixture. The
fumarate and maleate diesters were (separately) reacted with 3a
under our standard conditions developed for enantioselective
reactions of acrylate esters. The crude products were then
subjected to the ester hydrolysis/decarboxylation sequence and
finally esterification (or amidation) to allow assessment of the
enantioselectivity of these reactions (Table 3). The products
were obtained with good enantioselectivity, but the endo-/exo-
diastereoselectivity of the reaction with the fumarate ester was
essentially unchanged as compared to that observed with the
achiral catalyst.
In order to avoid the complication of potential C3 epimeriza-
tion, we sought to use alkenes would not produce cycloadducts
with a C3 ester group; however, attempts to react 3a with
cis- and (separately) trans-ethyl crotonate esters suffered from
attenuated reactivity and eventually complex mixtures resulting
from decomposition. While reaction of the p-nitrophenyl sub-
stituted azlactone (3b) with ethyl 4,4,4-trifluoro-2-butenoate
afforded 11 in good yield and stereoselectivity, H/D exchange
at the indicated position was readily observed upon subjection
of 11 to methanol-d4 and the achiral catalyst in acetone-d6
(eq 7).
2.3. Stereospecificity of the Gold(I)-Catalyzed Cycloaddi-
tion Reaction. In the context of cycloaddition chemistry,
the stereospecific conversion of alkene geometry to central
chirality is commonly taken as an indication of a concerted
reaction mechanism, while a nonstereospecific conversion
is regarded as evidence of a stepwise mechanism. Indeed,
the initial impetus for extending the scope of the gold(I)-
catalyzed 1,3-DCR to acyclic 1,2-disubstituted alkenes was
to assess whether the reaction is stereospecific. To this
end, reaction of 3a with dimethyl maleate, catalyzed by
Ph3PAuOBz and monitored by 1H NMR, gave rise to a single
cycloadduct (exo-IV) (eq 5). In contrast, the reaction of 3a
with dimethyl fumarate, similarly monitored, gave rise a 2.5:1
mixture of exo- and endo-IV (eq 6).15 In order to assess
whether the trans-relationship of the diesters arose from a
nonconcerted cycloaddition or a postcycloaddition epimer-
ization, excess methanol-d4 was added to the reaction mixture
after the gold-catalyzed reaction was complete. This addition
led to the rapid and nearly complete disappearance of the
diagnostic doublets due to the C3 methine proton; therefore,
the stereochemical course of these reactions could not be
ascertained.
Ultimately, we decided to evaluate the possibility that appro-
priately deuterated acrylate ester would allow the desired in-
vestigation to proceed. Indeed, we noted that no H/D exchange
was observed when 8a (derived from 3a and tert-butyl acrylate)
was treated with catalytic Ph3PAuOBz and methanol-d4 (eq 8).
Thus, acrylate 12 and its trans-monodeutero analogue (trans-12-
d1) were prepared (see Supporting Information for preparation
of these materials). Monitoring the reaction of 3a with alkene 12
Initial attempts at derivatization and isolation of the products
of these reactions were complicated by the presence excess
maleate and fumarate esters as well as the presence of
byproduct from the cycloaddition and (primarily) derivatiza-
tion reactions. The initial cycloadducts are carboxylic acids; a
fact we sought to exploit in developing a protocol for product
isolation. This lead to the serendipitous discovery that treat-
ment of the crude product obtained from reaction of 3a and
dimethyl maleate with 10% aqueous K2CO3 followed by
reacidification, extraction into organic medium, and finally
by esterification with diazomethane afforded compound rac-
8d. This compound had been previously established as the
product obtained from the gold(I)-catalyzed reaction of 3a
and methyl acrylate followed by treatment with diazomethane
1
by H NMR spectroscopy showed the formation of a single
cycloadduct exhibiting two well-resolved signals (δ = 3.52 ppm:
dd, J2 = 17.5 J3 = 8.0 Hz, 1H) and (δ = 3.46 ppm: dd, J2 = 17.5,
J3 = 8.5 Hz, 1H) due to the C3 methylene group (eq 9). This
compound was isolated and characterized as the N,N-dimethyl
amide. The reaction of 3a and trans-12-d1, similarly monitored,
indicated the formation of a single cycloadduct with deuterium
substitution at only one position (eq 9). Specifically only a simple
doublet was observed in this region of the spectrum (δ = 3.50
ppm: d, J = 7.5 Hz, 1H). With the alkene as the limiting reactant,
3519
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527

Journal of the American Chemical Society
ARTICLE
Table 3. Gold(I)-Catalyzed DCR/Hydrolysis/Decarboxylation Sequence
producta
d.r.
entry
alkene (equiv)
solvent
catalyst
(exo/endo)
(exo/endo)
yield (%)
e.e. (exo/endo)
1
2
3
4
5
maleate (2)
maleate (4)
fumarate (2)
fumarate (2)
fumarate (4)
acetone-d6
2.5 mol% PPh3AuOBz
3.5 mol% (R)-6
8d only
n.a.
n.a.
3:1b
2:1c
2:1c
69
55
59
63
60
n.a.
3/1 PhF/THF
acetone-d6
8d only
82/n.a.
n.a.
2.5 mol% PPh3AuOBz
2.5 mol% PPh3AuOBz
3.5 mol% (R)-6
8d/epi-8d
10d/epi-10d
10d/epi-10d
acetone-d6
n.a.
3/1 PhF/THF
94/84
a 8d/epi-8d R = OMe; 10d/epi-10d R = p-O2NC6H4CH2NH. b Isolated d.r., inseparable diastereomers. c Isolated d.r., all four stereolsomers separated by
analytical enantiodiscriminating HPLC.
the cycloaddition reaction was complete before any significant
H/D exchange was observed.
reported organocatalytic approaches to the enantioselective
azlactone Mannich reaction; however, both examples demon-
strated high levels of stereoselectivity, providing the syn-Mannich
addition products with good ee, but each method was limited to a
single class of imines derived from either aliphatic8a or aromatic8b
aldehydes, respectively. Thus, the development of a general,
functional group tolerant method for the enantioselective (and
anti-diastereoselective) Mannich reaction of azlactones would
complement these organocatalytic methods and extend the
realm of gold-catalyzed addition reactions beyond the aldol
reaction.
Toward this goal, we examined a variety of dinuclear gold(I)
benzoate catalysts for the Mannich reaction. Under our
previously developed reaction conditions (fluorobenzene,
rt), the (SEGPHOS)gold(I)-catalyzed reaction of azlactone
3a and N-sulfonylimine 14a provided the desired product in
excellent diastereoselectivity but modest enantioselectivity
(Table 4, entry 1). The use of SEGPHOS ligands with bulky
aryl groups at phosphorus afforded 15a with a modest
improvement in the enantioselectivity (entries 2, 3). On the
other hand, changing the phosphine group from aryl to
cyclohexyl provided a substantial increase in enantioselectiv-
ity (entries 4, 5). The most notable increases in enantioselec-
tivity occurred when the ligand class was changed from the
biaryl to spirocyclic (entries 6 - 8). For example, (R)-xylyl-
SDP(AuOBz)2, (R)-16,19 (see Figure 1 for X-ray structure of
the corresponding dichloride) catalyzed the formation of 15a
with 89% ee at room temperature. Finally, a bulkier group on
the imine enabled the highest level of both enantio- and
diastereocontrol (entries 9 and 10). Further optimization of
the reaction conditions revealed that use of fluorobenzene as
solvent was essential to obtaining the Mannich adducts with
high diastereo- and enantioselectivity.20
Thus it was determined that the gold(I)-catalyzed reaction
of 3a with an acrylate ester is stereospecific; this observation is
most consistent with a concerted process. In all cases, the
gold(I)-catalyzed 1,3-DCR of azlactones with monosubsti-
tuted alkenes has proceeded with excellent regioselectivity.
Moreover that selectivity serves to form a bond between what
had been C4 of the azlactone and the R-position of the R,
β-unsaturated ester (or nitrile); that is to say the observed
regioselectivity is exactly opposite that which might reason-
ably be expected from a stepwise process. This same regios-
electivity was also observed in the reaction of 3b with a
nonsymmetric 1,2-disubstituted alkene (eq 7). Taken to-
gether with the stereospecific nature of the reaction, these
observations on regioselectivity are also suggestive of a con-
certed process.16
2.3. Development of the Enantioselective Mannich Reac-
tion of Azlactones with N-Sulfonyl Aldimines. Investigation
into the mechanism of the gold(I)-catalyzed 1,3-DCR sug-
gested that the reaction proceeds via a concerted process. In
the context of a normal electron demand cycloaddition, an
intermediate such as II or III (eq 2) could be considered as the
nucleophilic component. However we sought to develop a
reaction, predicated on gold(I)-activation of the nucleophile,
which could be more clearly viewed as proceeding through a
traditional nucleophile/electrophile pair of reactants. We thus
posited that such bis(phosphinegold(I) carboxylate) com-
plexes might be competent catalysts for stereoselective reac-
tions of azlactones with electrophiles other than electron-
deficient alkenes, such as imines and/or aldehydes.
Having determined reaction conditions for the highly selective
formation of 15b, we conducted experiments to explore the
scope of the gold(I)-catalyzed enantioselective Mannich reaction
of azlactones with aldimines (Table 5). Addition to both
aromatic and aliphatic imines was achieved with our optimized
catalyst system to give the Mannich products in good yield and
selectivities. Notably, both electron-poor and electron-rich aro-
matic imines, including the furyl-substituted imine (entry 4),
were well tolerated in the reaction. Additionally, both acyclic
and cyclic aliphatic imines were competent electrophiles and
In contrast to the high enantioselectivities obtained in the
Hayashi-Ito aldol reaction,1 the sole previous report of gold(I)-
catalyzed reaction of isocyantes 2 with imines was not
enantioselective.17,18 Moreover, Ooi8a and Wang8b have recently
3520
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527

Journal of the American Chemical Society
ARTICLE
provided the desired products in excellent diastereoselectivity.
For example, the imine derived from 3-benzyloxypropanal
afforded 15l in 96% yield and >20:1 dr (94% ee of the major
diastereomer, entry 11). Furthermore, both electron-poor and
electron-rich azlactones could be used (entries 6-8).
To the best of our knowledge this constitutes the first reported
enantioselective Mannich reaction catalyzed by gold. The
Mannich adducts were obtained in high diastereo- and enantios-
electivities with up to >20:1 dr and 94% ee. Notably, the
gold-catalyzed reaction provides the anti-diastereomer as the
major product, in contrast to previously reported methods which
are highly syn-selective.8a,8b Moreover, ring opening of an
initial Mannich addition product under the action of mineral
acid, followed by esterification with diazomethane provided the
fully and differentially protected R,β-diamino acid22 methyl
ester (17) without loss of enantioenrichment of the material
(eq 10).
Table 4. Optimization of Reaction Conditions for the En-
antioselective Mannich Reaction
entry
ligand
R
(°C) yielda (%) d.r.b(anti/syn) eec (%)
1
2
3
4
5
6
7
8
9
(R)-SEGPHOS
(R)-DM-SEGPHOS
(R)-DTBM-SEGPHOS Me rt
(S)-Cy-SEGPHOS
(S)-Cy-SEGPHOS
Me rt
Me rt
64
61
20
51
43
62
52
42
61
76d
>20:1
14:1
40
-42
-57
-82
-82
89
88
90
90
94
4.3:1
4.6:1
3.8:1
4.6:1
3.5:1
3.8:1
4.0:1
6.6:1
Me rt
Me -10
(R)-xylyl-SDP, (R)-16 Me rt
(R)-16
(R)-16
(R)-16
Me
5
Me -10
Mes rt
10 (R)-16
Mes -20
Up to this point the gold-catalyzed Mannich reaction has been
optimized for alanine derived azlactones and the reactivity of
substrates bearing larger substituents at C4 is attenuated, parti-
cularly at low temperature. However, appreciable levels of
selectivity may be achieved at room temperature. For example,
allyl substituted substrate 3j was reacted with imine 14b to afford
addition product 15o with 80% ee and 6.7:1 d.r. using catalyst
(R)-6, albeit in modest yield (Table 6, entry 3). Using the more
reactive p-nosyl imine and THF as cosolvent, the corresponding
products (15p/15p0) were obtained in excellent yield but with
poor diastereoselectivity (entry 5). The results in Table 6, which
show the variation of selectivity with respect to reaction condi-
tions including ligand structure, suggest the reaction could be
optimized for other substrates.
a Isolated yield. b Determined by H NMR analysis of crude reaction
mixture. c Determined by enantiodiscriminating HPLC. d Used 5 mol %
of (R)-16.
1
Figure 1. Solid-state structure of (R)-xylyl-SDP(AuCl)2 (50% prob-
ability ellipsoids). Hydrogen atoms omitted for clarity.21
Table 5. Enantioselective Gold(I)-Catalyzed Azlactone Mannich Reactiona
entry
product
Ar1
Ar2
R
yieldb (%)
d.r.c(anti/syn)
eed (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
15b
15c
15d
15e
15f
15g
15h
15i
Ph
Ph
Ph
Ph
Ph
Mes
Mes
Mes
Ph
76
70
58
50e
91
98
89
96
87
88
96
69 f
73
6.6:1
6.9:1
6.9:1
6:1
17.2:1
>20:1
>20:1
>20:1
11:1
94
86
82
92
94
93
87
83
92
93
94
91
93
p-CIC6H4
p-MeOC6H4
3-furyl
Me
Me
Mes
p-Tol
p-Tol
p-Tol
p-Tol
p-Tol
p-Tol
p-Tol
p-Tol
p-Tol
p-MeOC6H4
p-BrC6H4
p-O2NC6H4
Ph
Ph
Ph
Ph
Ph
Me
Me
15j
pent-1-en-5-yl
PhCH2CH2
BnOCH2CH2
iPr
15k
151
15m
15n
8:1
>20:1
>20:1
>20:1
cyclohexyl
a Reactions carried out on 0.15 mmol scale, 1.05 equiv of aromatic imine or 1.2 equiv of aliphatic imine; Mes = 2,4,6-mesityl. b Isolated yield. c Determined by
1H NMR analysis of crude product mixtures. d Determined for major diastereomer by enantiodiscriminating HPLC, absolute stereochemistry of 15b and 15f were
established by X-ray crystallography, and other products assigned by analogy. e Reaction run in 3/2 PhF/CHCl3 at 0.2 M in azlactone. f Reaction run at rt.
3521
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527

Journal of the American Chemical Society
ARTICLE
Table 6. Gold-Catalyzed Mannich Reactions of Azlactone 3ja
entry
catalyst
solvent
product (anti/syn)
Ar
yield (%)
d.r.b(anti/syn)
eec(%)
1
2
3
4
5
PPh3AuOBzd
(R)-16d
THF
PhF
15o/15o0
15oe
15oe
15p/15p0
15p/15p0
Mes
Mes
Mes
70/23
49
2.9:1
4:1
-/-
30
(R)-6
PhF
41
83f
97f
6.7:1
5:1
-80
PPh3AuOBz
THF
p-NO2C6H4-
-/-
(R)-6
PhF/THF 3/1
p-NO2C6H4-
1.7:1
-72/-15
a Reactions carried out on 0.1 mmol scale; Mes = 2,4,6-mesityl. b Determined by H NMR analysis of crude product mixture. c Determined by
1
enantiodiscriminating HPLC. d Using 3 mol % catalyst. e Only the major diastereomer was isolated. f Mixture of diastereomers.
Figure 2. (a) ESI-MS spectrum of the Mannich reaction of azlactone 3a with imine 14b catalyzed by (R)-SEGPHOS(AuOBz)2 at t = 10 min. (b) MS/
MS spectrum resulting from collisionally activated dissociation (CAD) of the singly charged positive ion at m/z = 1465. (c) ESI-MS spectrum of the 1,3-
DCR of azlactone 3a with maleimide 4 catalyzed by (R)-SEGPHOS(AuOBz)2 at time = 40 min. (d) MS/MS spectrum resulting from collisionally
activated dissociation (CAD) of the singly charged positive ion at m/z = 1351.
2.4. Tandem MS Studies. We sought to gain further insight
into the reaction mechanisms of both the gold-catalyzed 1,3-
DCR and Mannich reaction and to obtain experimental data that
would allow some measure of direct comparison to be drawn
between the two reactions. To this end, we investigated the
kinetic behavior of both reactions as well as subjecting both types
of reaction mixture to analysis by ESI-MS(/MS). Tandem mass
spectrometry experiments have previously been used to deter-
mine the nature of reaction intermediates, and we hypothesized
such an experiment would shed some light on the identity of
relevant gold-containing species in solution.23
We began our mechanistic investigations by studying the gold-
catalyzed Mannich reaction. When the reaction of 3a with 14b
catalyzed by (R)-SEGPHOS(AuOBz)2 was monitored by ESI-
MS, after 10 min we were able to detect two ions, each
corresponding to a critical intermediate (Figure 2a): a signal
at m/z 1178.2, which we attribute to the cationic azlactone-
gold(I) complex, and a second signal at m/z 1465.2, consistent
with the cationic intermediate azlactone-gold(I)-imine com-
plex. No signal corresponding to a gold(I)-imine complex was
observed. Both species were characterized by tandem mass
spectrometry experiments. The ESI-MS(/MS) spectrum of
3522
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527

Journal of the American Chemical Society
ARTICLE
Figure 3. Kinetic data for (R)-SEGPHOS(AuOBz)2-catalyzed reaction of 3a with imine 14b to form 15b. (a) Pseudofirst-order kinetics in azlactone 3a
(33.3 nM) was observed when a large excess (10 equiv) of imine 14b (335 nM) was used. (b) Pseudofirst-order behavior in imine 14b (28.5 nM) was
observed when a large excess (10 equiv) of azlactone (290 nM) was used. (c) Average measured kobs at various (R)-SEGPHOS(AuOBz)2 concentrations
(0.33-1.33 nM) for the reaction of 3a (33.3 nM) with 14b (267 nM).
the ion of m/z 1465.2 showed predominantly neutral loss of
the Mannich product (Figure 2b). These data are consistent
with a nucleophile-activation mechanism where the gold is
coordinated to the azlactone.
compelling case for such a cycloaddition/fragmentation mech-
anism for the Mannich reaction.27
When the same studies were performed for the cycloaddition
reaction between 3a and N-phenylmaleimide (4), after 40 min
we were able to intercept one ion corresponding to the cat-
ionic azlactone-gold(I)-maleimide intermediate at m/z 1351.2
(Figure 2c). The ESI-MS(/MS) data for this reaction mix-
ture showed that loss of neutral CO2 was the primary mode of
fragmentation for this intermediate. This observation supports
the notion of initial formation of a bridged, bicyclic cycloadduct
followed by extrusion of CO2 (Figure 2d and eq 11).24,25
A cycloaddition/fragmentation mechanism could also be
operative for the Mannich reaction; however, we did not
observe any signals suggesting loss of CO2 when analyzing
these reaction mixtures by ESI-MS(/MS). Analogous inter-
mediates, along this pathway from which fragmentation or
expulsion of CO2 might occur, appear unlikely to produce the
observed stereochemical outcome (cf. eqs 11 and 12). While
VII presumably suffers less severe nonbonding steric interac-
tion between the methyl and phenyl groups as compared to
VIII, it is VIII that would lead to the (experimentally observed)
predominant diastereomer 15a. Additionally in the gold(I)-
catalyzed aldol reaction of 3a with ethyl glyoxalate, fragmenta-
tion of an initial cycloadduct would likely favor formation of
an oxazoline, but this was not observed; instead the reaction
afforded the highly substituted azlactone diastereomers 18a
and 18b (eq 13).26 Reasoning along these lines there is not a
2.5. Kinetic Studies. Kinetic studies performed on the Mannich
reaction showed that the reaction exhibited a first-order depen-
dence on azlactone, imine, and gold(I). Standard solutions of
3523
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527

Journal of the American Chemical Society
ARTICLE
Figure 4. KineticdataforthePh3PAuOBz-catalyzed 1,3-DCR of 3a and 4(a) A plot of -ln(3a/3a0) vs time in the 1,3-DCR of 3a (47.6nmM) and4(476nM)
is nonlinear. (b) Overlays of kinetic data for consecutive additions of azlactone 3a (50.0 nM x 3) to solution of 4(523 nM) catalyzed and Ph3PAuOBz(0.57nM).
Time t = 0 corresponds to the beginning of each cycle, and the left axis represents the concentration of 3a at the beginning of each cycle.
Figure 5. Mechanistic proposals for gold(I)-catalyzed 1,3-DCR and Mannich reactions.
azlactone 3a, N-(mesitylsulfonyl)benzaldimine (14b), and relevant
gold catalysts were prepared individually in deuterated solvent, and
1,3,5-trimethoxybenzene was used as an internal standard.28a A plot
of -ln(3a/3a0) vs time showed that the reaction exhibited pseudo-
first-order behavior in azlactone when a large excess of imine was used
(Figure 3a). When an excess of azlactone was used, the reaction also
showed pseudofirst-order kinetics in imine (Figure 3b).
To determine the kinetic order in gold(I), the reaction was
conducted at 1-4 mol % catalyst loading, under pseudofirst-order
conditions. The pseudofirst-order constants (kobs) are tabulated in
Table S18 (Supporting Information). A plot of the averaged kobs at
each catalyst loading can be subjected to a linear regression with R2 =
0.997 (Figure 3c), suggesting that the reaction is first order in (R)-
SEGPHOS(AuOBz)2. Furthermore, when the same experiments
are conducted with a mononuclear catalyst, Ph3PAuOBz, the same
first-order dependence in gold(I) catalyst is observed. These
experiments suggest that only one gold center is active in the rate-
limiting transition state.
We hypothesized that curvature in the plot could be due to a
less than first-order dependence on azlactone, catalyst deactiva-
tion, or product inhibition. Given that the product is a free
carboxylic acid and the catalyst bears a carboxylate counterion,
product inhibition seemed to be a likely explanation. To test for
product inhibition, 0.5 equiv of product (or of AcOH) was
combined with 3a, 4, and Ph3PAuOBz at the start of the reaction.
However the kinetic profile with either additive present was
effectively unchanged.28c Furthermore, complex (()-19c, com-
posed of [PPh3Au]þ and the conjugate base of cycloadduct (()-
19a, was prepared, isolated, and used to catalyze the reaction
of 3a and 4, again the kinetic profile remained unchanged
(eq 14).28d
Interestingly, the kinetic profiles observed for the reaction of
azlactone 3a with N-phenylmaleimide (4) were much more
complicated. In particular, the logarithmic first-order plot of -
ln(3a/3a0) vs time under flooding conditions was nonlinear at
lower conversions (Figure 4a). However, the reaction exhibited
pseudofirst-order behavior in 3a at higher conversion and longer
reaction times. The same curvature was observed when we
flooded in 3a and monitored the conversion of 4 to product.28b
Furthermore, this behavior was not sensitive to the relative order
of addition of the reagents or prestirring 3a in the presence of
catalyst prior to the addition of 4.
To test for catalyst deactivation,29 the reaction of 3a and 4 was
monitored, and an additional 0.67 equiv of 3a was added each
time the concentration of 3a fell below 15.8 mM (corresponding
to ∼67% conversion). The reaction was “cycled” in this manner
four times, and data for each cycle is presented as a set of overlaid
traces (Figure 4b), where t=0 corresponds to the addition of each
aliquot of starting material. The rate of the reaction in cycle 1 is
clearly faster than that in cycles 2-4, supporting a process in
3524
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527

Journal of the American Chemical Society
ARTICLE
which the concentration of active catalyst decreases at the
beginning of the reaction but stabilizes at longer reaction times.
This observation is consistent with catalyst deactivation at the
early stages of the reaction and explains our earlier observations
of curvature in the pseudofirst-order rate plot. In fact, a plot of -
ln(3a/3a0) vs time generated from cycle 4, (i.e., after the
concentration of catalyst has become constant), shows first-order
dependence in 3a.28e Because the precise concentration of active
catalyst in solution cannot be determined, we were unable to
perform a detailed kinetic analysis of the reaction of 3a with 4.
However, based on the kinetic profiles of the reactions after the
concentration of catalyst has stabilized, we believe that the overall
reaction is positive order in both reactants.
Kleinbeck for previous studies and Dr. Anthony T. Iavarone for
help with the MS studies. G.W.A. thanks the National Council for
Scientific and Technological Development (CNPq)-Brazil for a
postdoctoral fellowship. Z.J.W. thanks the Hertz Foundation,
and M.L. thanks Italian MIUR for graduate fellowships. Solvias
and Takasago are acknowledged for the generous donation of
phosphine ligands and Johnson Matthey for a gift of AuCl3.
’ REFERENCES
(1) (a) Ito, Y.; Sawamura, M.; Hayashi, T. J. Am. Chem. Soc. 1986,
108, 6405. (b) Hayashi, T. Pure Appl. Chem. 1988, 60, 7. (c) Samamura,
M.; Nakayama, Y.; Kato, T.; Ito, Y. J. Org. Chem. 1995, 60, 1727.
(2) For general reviews of gold catalysis see:(a) Shapiro, N. D.;
Toste, F. D. Synlett 2010, 675. (b) F€urstner, A. Chem. Soc. Rev. 2009, 38,
3208. (c) Jimꢀenez-Nꢀu~nez, E.; Echavarren, A. M. Chem. Rev. 2008, 108,
3326. (d) Shen, H. C. Tetrahedron 2008, 64, 7847. (e) Gorin, D. J.;
Sherry, B. D.; Toste, F. D. Chem. Rev. 2008, 108, 3351. (f) Li, Z.;
Brouwer, C.; He., C Chem. Rev. 2008, 108, 3239. (g) Arcadi, A. Chem.
Rev. 2008, 108, 3266. (h) Hashmi, A. S. K. Chem. Rev. 2007, 107, 3180.
(3) For early examples of this reactivity using cationic phos-
phinegold(I) complexes, see:(a) Teles, J. H.; Brode, S.; Chabanas, M.
Angew. Chem., Int. Ed. 1998, 37, 1415. (b) E. Mizushima, E.; Sato, K.;
Hayashi, T.; Tanaka, M. Angew. Chem., Int. Ed. 2002, 41, 4563. (c) Reetz,
M. T.; Sommer, K. Eur. J. Org. Chem. 2003, 3485. (d) Mizushima, E.;
Hayashi, T.; Tanaka, M. Org. Lett. 2003, 5, 3349. (e) Nieto-Oberhuber,
C.; Munuz, M. P.; Bunuel, E.; Nevado, C.; Cardenas, D. J.; Echavarren,
A. M. Angew. Chem., Int. Ed. 2004, 43, 2403. (f) Kennedy-Smith, J. J.;
Staben, S. T.; Toste, F. D. J. Am. Chem. Soc. 2004, 126, 4526. (g)
Mamane, V.; Gress, T.; Krause, H.; F€urstner, A. J. Am. Chem. Soc. 2004,
126, 8654. (h) Luzung, M. R.; Markham, J. P.; Toste, F. D. J. Am. Chem.
Soc. 2004, 126, 10858.
(4) (a) Lalonde, R. L.; Wang, Z. J.; Mba, M.; Lackner, A. D.; Toste,
F. D. Angew. Chem., Int. Ed. 2010, 49, 598. (b) LaLonde, R. L.; Sherry,
B. D.; Kang, E. J.; Toste, F. D. J. Am. Chem. Soc. 2007, 129, 2452.
(5) For seminal reports see:(a) Pl€ochl, J. Ber. Dtsch. Chem. Ges. 1883,
16, 2815. (b) Elrenmeyer, E. Ber. Dtsch. Chem. Ges. 1900, 33, 2036. For a
classic mid-20th century text see. (c) Oxazoles and Oxazolones; Clarke,
H. T., Johnson, J. R., Robinson, R., Eds.; Princeton University Press:
Princeton, NJ, 1949. For recent reviews see: (d) Mosey, R. A.; Fisk, J. S.;
Tepe, J. J. Tetrahedron: Asymmetry 2008, 19, 2755. (e) Fisk, J. S.; Mosey,
R. A.; Tepe, J. J. Chem. Soc. Rev. 2007, 36, 1432.
(6) For selected examples of azlactone (and azlactone derived)
nucleophiles in enantioselective catalysis see: (a) Dong, S.; Liu, X.;
Chen, X.; Mei, F.; Zhang, Y.; Gao, B.; Lin, L.; Feng, X. J. Am. Chem. Soc.
2010, 132, 10650. (b) Jiang, H.; Paix~ao, M. V.; Monge, D.; Jørgensen,
K. A. J. Am. Chem. Soc. 2010, 132, 2775. (c) Alba, A.-N. R.; Companyꢀo,
X.; Valero, G.; Moyano, A.; Rios, R. Chem.—Eur. J. 2010, 16, 5354. (d)
Alba, A.-N. R.; Valero, G.; Calbet, T.; Font-Bardía, M.; Moyano, A.; Rios,
R. Chem.—Eur. J. 2010, 16, 9884. (e) Terada, M.; Tanaka, H.;
Sorimachi, K. J. Am. Chem. Soc. 2009, 131, 3430. (f) Uraguchi, D.; Ueki,
Y.; Ooi, T. Science 2009, 326, 120. (g) Uraguchi, D.; Asai, Y.; Seto, Y.;
Ooi, T. Synlett 2009, 658. (h) Uraguchi, D.; Asai, Y.; Ooi, T. Angew.
Chem., Int. Ed. 2009, 48, 733. (i) Jiang, J.; Qing, J.; Gong, L.-Z. Chem.—
Eur. J. 2009, 15, 7031. (j) Dietz, F. R.; Gr€oger, H. Synthesis 2009, 4208.
(k) Cabrera, S.; Reyes, E.; Alemꢀan, J.; Milleli, A.; Kobbelgaard, S.;
Jørgensen, K. A. J. Am. Chem. Soc. 2008, 130, 12031. (l) Alemꢀan, J.;
Milelli, A.; Cabrera, S.; Reyes, E.; Jørgensen, K. A. Chem.—Eur. J. 2008,
14, 10958. (m) Shaw, S. A.; Aleman, P.; Christy, J.; Kampf, J. W.; Va, P.;
Vedejs, E. J. Am. Chem. Soc. 2006, 128, 925. (n) Shaw, S. A.; Aleman, P.;
Vedejs, E. J. Am. Chem. Soc. 2003, 125, 13368. (o) Liang, J.; Ruble, J. C.;
Fu, G. C. J. Org. Chem. 1998, 63, 3154. (p) Ruble, J. C.; Fu, G. C. J. Am.
Chem. Soc. 1998, 120, 11532.
3. CONCLUSIONS
The majority of recently reported gold-catalyzed transforma-
tions fall into the reactivity manifold in which gold serves to
activate carbon-carbon π-bonds as electrophiles toward nu-
cleophilic additions. In this report we provide evidence that
phosphinegold(I)-carboxylate complexes can serve to activate
pro-nucleophiles toward deprotonation by the counterion. On
the basis of the data presented in the preceding sections, we
propose two related catalytic cycles for the 1,3-DCR and
Mannich reactions. In both cases initial activation of the azlac-
tone as a nucleophile is followed by C-C bond forming reaction
with the electrophile (Figure 5).
We have proposed a divergent set of intermediates for the 1,3-
DCR (II/III f IX) and Mannich reactions (II/III f X). With
electron-deficient alkenes, the activated azlactones react through
a concerted 1,3-dipolar cycloaddition process.30 In the presence
of chiral biarylphosphine ligands, the chemistry provides a highly
diastereo- and enantioselective entry into complex Δ1-pyrrolines. In
contrast, the gold-catalyzed reaction of azlactones with activated
imines proceeds through an addition mechanism. Employing spiro-
bisphosphines as ligands, the first gold(I)-catalyzed enantioselective
Mannich reaction was developed. Thus, the bisphosphinegold(I)-
catalyzed Mannich reaction provides direct access to a variety of
aliphatic and aromatic R,β-diamino acid derivatives in high diastereo-
and enantioselectivities. Moreover, these studies introduce spiro-
bisphosphines19 as an alternative class of ligands to the biaryl-derived
bisphosphine11 ligands typically applied in enantioselective catalysis
with gold.31 Further studies regarding this mode of gold(I) activation
are ongoing in our laboratories and will be reported in due course.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures, ana-
b
lytic and spectroscopic data for new compounds, mass spectro-
metric, kinetic, and X-ray crystallographic data. This material is
available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
fdtoste@berkeley.edu
Author Contributions
#These authors contributed equally.
(7) (a) Liu, X.; Hartwig, J. F. Org. Lett. 2003, 5, 1915. (b) Frꢀebault,
F.; Luparia, M.; Oliveira, Goddard, R.; Maulide, N. Angew. Chem., Int. Ed.
2010, 49, 5672. (c) Trost, B. M.; J€akel, C.; Plietker, B. J. Am. Chem. Soc.
2003, 125, 4438. (d) Trost, B. M.; Lee, C. J. Am. Chem. Soc. 2001, 123,
’ ACKNOWLEDGMENT
We gratefully acknowledge NIHGMS (RO1 GM073932),
Amgen and Novartis for financial support. We thank Dr. Florian
3525
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527

Journal of the American Chemical Society
ARTICLE
12191. (e) Trost, B. M.; Ariza, X. J. Am. Chem. Soc. 1999, 121, 10727. (f)
Trost, B. M.; Heinemann, C.; Ariza, X.; Weigand, S. J. Am. Chem. Soc.
1999, 121, 8667. (g) Trost, B. M.; Lee, C. B. J. Am. Chem. Soc. 1998, 120,
6818. (h) Trost, B. M.; Ariza, X. Angew. Chem., Int. Ed. 1997, 36, 2635.
For Mo-catalyzed allylation see: (i) Trost, B. M.; Dogra, K.; Franzini, M.
J. Am. Chem. Soc. 2004, 126, 1944. (j) Trost, B. M.; Dogra, K. J. Am.
Chem. Soc. 2002, 124, 7256.
Chem. 2009, 15, 2401. (c) Verkade, J. M. M.; vanHemert, L. J. C.;
Quaedflieg, P. J. L. M.; Rutjes, F. P. J. T. Chem. Soc. Rev. 2008, 37, 29. (d)
Ting, A.; Schaus, S. E. Eur. J. Org. Chem. 2007, 35, 5797. (e) Marques,
M. M. B. Angew. Chem., Int. Ed. 2006, 45, 348.
(19) We have previously reported a single example of a xylyl-SDP-
(AuCl)2/AgSBF6 catalyzed enantioselective reaction. Sethofer, S. G.;
Staben, S. T.; Hung, O. Y.; Toste, F. D. Org. Lett. 2008, 10, 4315.
(20) Reaction of 3a with 14b catalyzed by 3 mol% (R)-xylyl-SDP-
(AuOBz)2 at rt produced 15b with lower diastereo- and enantioselec-
tivity in the following solvents: PhMe: 1.6:1 dr and 84% ee (major); 1,2-
DCE: 2:1 dr and 86% ee (major); PhMe/THF (1:1): 1.7:1 dr and 75%
ee (major), 42% ee (minor); CH2Cl2: 2:1 dr and 86% ee (major);
CHCl3: 2.1:1 dr and 85% ee (major).
(21) The assignment of stereochemical designators (R or S) for
axially chiral spiranes follows a unique historical convention. Eliel, E. L.;
Wilen, S. H.; Mander, L. N. Stereochemistry of Organic Compounds;
Wiley: New York, 1994; p 1138.
(22) For recent reviews of stereoselective methods for the prepara-
tion of R,R-and R,β-diamino acids see: (a) Wang, J.; Zhang, L.; Jiang, H.;
Liu, H. Curr. Pharm. Des. 2010, 16, 1252. (b) Arrayas, R. G.; Carretero,
J. C. Chem. Soc. Rev. 2009, 38, 1940. (c) Viso, A.; de la Pradilla, R. F.;
García, A.; Flores, A. Chem. Rev. 2005, 105, 3167. (d) Westermann, B.
Angew. Chem., Int. Ed. 2003, 42, 151. For R,R-diamino acids see: (e)
Tanaka, M. Chem. Pharm. Bull. 2007, 55, 349. For β-amino acids see: (f)
Weiner, B.; Szymanski, W.; Janssen, D. B.; Minnaard, A. J.; Feringa, B. L.
Chem. Soc. Rev. 2010, 39, 1656.
(23) For reviews of tandem mass spectrometry in the context of
identification of reactive intermediates see: (a) Eberlin, M. N. Eur. J.
Mass Spectrom. 2007, 13, 19. (b) Plattner, D. A. Int. J. Mass Spectrom.
2001, 207, 125. (c) Gross, M. L. Int. J. Mass Spectrom. 2000, 200, 611. (d)
Futrell, J. H. Int. J. Mass Spectrom. 2000, 200, 495. For recent examples of
using mass spectrometry to study reaction mechanisms see: (e) Garver,
J. M.; Fang, Y.-R.; Eyet, N.; Villano, S. M.; Bierbaum, V. M.; Westaway,
K. C. J. Am. Chem. Soc. 2010, 132, 3808. (f) Amarante, G. W.; Milagre,
H. M. S.; Vaz, B. G.; Ferreira, B. R. V.; Eberlin, M. N.; Coelho, F. J. Org.
Chem. 2009, 74, 3031. (g) Schrader, W.; Handayani, P. P.; Zhou, J.; List,
B. Angew. Chem., Int. Ed. 2009, 48, 1463. (h) Nolin, K. A.; Krumper, J. R.;
Pluth, M. D.; Bergman, R. G.; Toste, F. D. J. Am. Chem. Soc. 2007, 129,
14684. (i) Takats, Z.; Koch, K. J.; Cooks, R. G. Anal. Chem. 2001, 73,
4522. For an example involving the mechanism of gold-catalyzed
cyclopropanation see: (j) Fedorov, A.; Moret, M.-E.; Chen, P. J. Am.
Chem. Soc. 2008, 130, 8880.
(24) The thermal decarboxylation of triphenylphosphinegold(I)
carboxylates has been reported; see: Fackler, J. P., Jr.; Khan, Md. N. I.;
King, C.; Staples, R. J.; Winpenny, R. E. P. Organometallics 1991, 10,
2178.
(25) Pd-catalyzed enantioselective Michael addition of azlactones to
enones was recently reported. (a) Weber, M.; Jautze, S.; Frey, W.; Peters,
R. J. Am. Chem. Soc. 2010, 132, 12222. Organocatalytic azlactone
Michael reaction, see: (b) Cabrera, S.; Reyes, E.; Alemꢀan, J.; Milelli,
A.; Kobbelgaard, S.; Jørgensen, K. A. J. Am. Chem. Soc. 2008, 130, 12031.
(26) For examples of azlactone in enantioselective aldol reactions,
see: (a) Misaki, T.; Takinoto, G.; Sugimura, T. J. Am. Chem. Soc. 2010,
132, 6268. (b) Terada, M.; Tanaka, H.; Sorimachi, K. J. Am. Chem. Soc.
2009, 131, 3430.
(8) (a) Uraguchi, D.; Ueki, Y.; Ooi, T. J. Am. Chem. Soc. 2008, 130,
14088. (b) Liu, X.; Deng, L.; Jiang, X.; Yan, W.; Liu, C.; Wang, R. Org.
Lett. 2010, 12, 876. For closely related heterocycles as nucleophiles see:
(c) Uraguchi, D.; Koshimoto, K.; Ooi, T. Chem. Commun. 2010, 46, 300.
(9) Peddibhotla, S.; Tepe, J. J. J. Am. Chem. Soc. 2004, 126, 12776.
(10) The studies on the gold-catalyzed [3 þ 2]-dipolar cycloaddi-
tion of azlactones were partially communicated in: Melhado, A. D.;
Luparia, M.; Toste, F. D. J. Am. Chem. Soc. 2007, 129, 12638.
(11) Chiral ligands substituted with bulky phosphine substituents
have proved essential for high enantioselectivity in a number of gold-
catalyzed enantioselective transformations: (a) Mu~niz, M. P.; Adrio, J.;
Carretero, J. C.; Echavarren, A. M. Organometallics 2005, 24, 1293. (b)
Johansson, M. J.; Gorin, D. J.; Staben, S. T.; Toste, F. D. J. Am. Chem. Soc.
2005, 127, 18002. (c) Zhang, Z.; Widenhoefer, R. A. Angew. Chem., Int.
Ed. 2007, 46, 283. (d) Luzung, M. R.; Mauleꢀon, P.; Toste, F. D. J. Am.
Chem. Soc. 2007, 129, 12402. (e) Uemura, M.; Watson, I. D. G.;
Katsukawa, M.; Toste, F. D. J. Am. Chem. Soc. 2009, 131, 3464. (f)
Zhang, Z.; Lee, S. D.; Widenhoefer, R. A. J. Am. Chem. Soc. 2009, 131,
5372. (g) Chao, C.-M.; Vitale, M. R.; Toullec, P. Y.; Gen^et, J.-P.;
Michelet, V. Chem.—Eur. J. 2009, 15, 1319. (h) Chao, C.-M.; Genin, E.;
Toullec, P. Y.; Gen^et, J.-P.; Michelet, V J. Organomet. Chem. 2009, 694,
538. (i) Chao, C.-M.; Beltrami, D.; Toullec, P. Y.; Michelet, V. Chem.
Commun. 2009, 6988. (j) Bandini, M.; Eichholzer, A. Angew. Chem., Int.
Ed. 2009, 48, 9533. (k) Sethofer, S. G.; Mayer, T.; Toste, F. D. J. Am.
Chem. Soc. 2010, 132, 8278.
(12) For reviews of recent advances in the catalytic enantioselective
1,3-DCR see: (a) Kissane, M.; Maguire, A. R. Chem. Soc. Rev. 2010, 39, 845.
(b) Nꢀajera, C.; Sansano, J. M.; Yus, M. J. Braz. Chem. Soc. 2010, 21, 377.
(c) Stanley, L. M.; Sibi, M. P. Chem. Rev. 2008, 108, 2887. (d) Nair, V.
Tetrahedron 2007, 63, 12247. (e) Pellissier, H. Tetrahedron 2007, 63, 3235.
(13) For selected reports of metal-catalyzed enantioselective reac-
tions of acyclic azomethine ylides see: (a) Longmire, J. M.; Wang, B.;
Zhang, X. J. Am. Chem. Soc. 2002, 124, 13400. (b) Chuo, C.; Li, X.;
Schreiber, S. L. J. Am. Chem. Soc. 2003, 125, 10174. (c) Cabrera, S.;
Arrayꢀas, R. G.; Carretero, J. C. J. Am. Chem. Soc. 2005, 127, 16394. (d)
Zeng, W.; Chen, G.-Y.; Zhou, Y.-G.; Li, Y.-X. J. Am. Chem. Soc. 2007,
129, 750. (f) Yan, X.-X.; Peng, Q.; Zhang, Y.; Zhang, K.; Hong, W.; Hou,
X.-L.; Wu, Y.-D. Angew. Chem., Int. Ed. 2006, 45, 1979. (g) Lꢀopez-Pꢀerez,
A.; Adrio, J.; Carretero, J. C. J. Am. Chem. Soc. 2008, 130, 10084. (h)
Wang, C.-J.; Liang, G.; Xue, Z.-Y.; Gau, F. J. Am. Chem. Soc. 2008, 130,
17250. (i) Filippone, S.; Maroto, E. E.; Martín-Domenech, A.; Suarez,
M.; Martin, N. Nat. Chem. 2009, 1, 578. (j) Lꢀopez-Pꢀerez, A.; Adrio, J.;
Carretero, J. C. Angew. Chem., Int. Ed. 2009, 48, 340. (k) Robles-Machín,
R.; Alonso, I.; Adrio, J.; Carretero, J. C. Chem.—Eur. J. 2010, 16,
5286and references therein.
(14) Various transformations of the Δ1-pyrroline imine group have
been described. The majority of cases concern systems bearing either H,
an heteroatom, or a carbonyl/carboxyl group at the C2 position. See:
Shevkhgeimer, M.-G. A. Chem. Heterocycl. Compd. 2003, 39, 405.
(15) These results are essentially identical to those obtained by Tepe
in the silver-catalyzed cycloadditions of maleate and fumarate (see ref 9).
(16) Steric control of regioselectivity cannot be explicitly ruled out.
See: (a) Coppola, B. P.; Noe, M. C.; Schwartz, D. J.; Abdon, R. L., II;
Trost, B. M. Tetrahedron 1994, 50, 93. (b) Steglich, W.; Gruber, P.;
H€ofle, G.; K€onig, W. Angew. Chem., Int. Ed. 1971, 10, 653. (c) Gotthardt,
H.; Huisgen, R.; Bayer, H. O. J. Am. Chem. Soc. 1970, 92, 4340 and ref 5.
(17) Hayashi, T.; Kishi, E.; Soloshonok, V. A.; Uozumi, Y. Tetra-
hedron Lett. 1996, 37, 4969.
(27) Fragmentation of the initial cycloadduct of an azlactone and an
imine has been proposed to result in the formation of imidazolines, see:
(a) Peddibhotla, S.; Tepe, J. J. Synthesis 2003, 9, 1433. (b) Peddibhotla,
S.; Jayakumar, S.; Tepe, J. Org. Lett. 2002, 4, 3533.
(28) (a) For details see Supporting Information. (b) See Supporting
Information, Figure S15. (c) See Supporting Information, Figure S16.
(d) See Supporting Information for preparation of (()-19c and Figure
S17 for kinetics. (e) See Supporting Information Figures S18 and S19.
(29) For similar experiments, see: Shekhar, S.; Ryberg, P.; Hartwig,
J. F.; Mathew, J. S.; Blackmond, D. G.; Strieter, E. R.; Buchwald, S. L.
J. Am. Chem. Soc. 2006, 128, 3584.
(18) For recent reviews of stereoselective Mannich-type chemistry
see (a) Kobayashi, S.; Matsubara, R. Chem.—Eur. J. 2009, 15, 10694. (b)
Marques-Lopez, E.; Merino, P.; Tejero, T.; Herrera, R. P. Eur. J. Org.
(30) For selected recent reports of gold-catalyzed cycloaddition
reactions see: (a) Teng, T.-M.; Liu, R.-S. J. Am. Chem. Soc. 2010, 132,
3526
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527

Journal of the American Chemical Society
ARTICLE
9298. (b) Lꢀopez-Carrillo, V.; Echavarren, A. M. J. Am. Chem. Soc. 2010,
132, 27. (c) Partyka, D. V.; Gao, L.; Teets, T. S.; Updegraff, J. B., III;
Deligonul, N.; Gray, T. G. Organometallics 2009, 28, 6171. (d) Benitez,
D.; Tkatchouk, E.; Gonzalez, A. Z.; Goddard, W. A., III; Toste, F. D. Org.
Lett. 2009, 11, 4798. (e) Harmata, M.; Huang, C. Tetrahedron Lett. 2009,
50, 5701. (f) Shapiro, N. D.; Shi, Y.; Toste, F. D. J. Am. Chem. Soc. 2009,
131, 11654. (g) Liu, F.; Yu, Yihua, Y.; Zhang, J. Angew. Chem., Int. Ed.
2009, 48, 5505. (h) Mauleꢀon, P.; Zeldin, R. M.; Gonzalez, A. Z.; Toste,
F. D. J. Am. Chem. Soc. 2009, 131, 6348. (i) Trillo, B.; Lꢀopez, F.;
Montserrat, S.; Ujaque, G.; Castedo, L.; Lledꢀos, A.; Mascare~nas, J. L.
Chem.—Eur. J. 2009, 15, 3336. (j) Yeom, H.-S.; Lee, J.-E.; Shin, S. Angew.
Chem., Int. Ed. 2008, 47, 7040. (p) Zhang, G.; Zhang, L. J. Am. Chem. Soc.
2008, 130, 12598. (k) Shapiro, N. D.; Toste, F. D. J. Am. Chem. Soc.
2008, 130, 9244.
(31) Recently chiral phosphoramidite and phosphite ligands have
also been employed in enantioselective gold-catalyzed reactions. (a)
Alonso, I.; Trillo, B.; Lꢀopez, F.; Montserrat, S.; Ujaque, G.; Castedo, L.;
Lledꢀos, A.; Mascare~nas, J. L. J. Am. Chem. Soc. 2009, 131, 13020. (b)
Gonzalez, A. Z.; Toste, F. D. Org. Lett. 2010, 12, 200. (c) Teler, H.;
Fl€ugge, S.; Goddard, R.; F€urstner, A. Angew. Chem., Int. Ed. 2010, 49,
1949.
3527
dx.doi.org/10.1021/ja1095045 |J. Am. Chem. Soc. 2011, 133, 3517–3527