Bifunctional Lewis acid-nucleophile-based asymmetric catalysis: Mechanistic evidence for imine activation working in tandem with chiral enolate formation in the synthesis of β-lactams

Article abstract of DOI:10.1021/ja044179f

We report a mechanistically based study of bifunctional catalyst systems in which chiral nucleophiles work in conjunction with Lewis acids to produce β-lactams in high chemical yield, diastereoselectivity, and enantioselectivity. Chiral cinchona alkaloid

Full text of DOI:10.1021/ja044179f

Published on Web 01/07/2005
Bifunctional Lewis Acid-Nucleophile-Based Asymmetric
Catalysis: Mechanistic Evidence for Imine Activation Working
in Tandem with Chiral Enolate Formation in the Synthesis of
â-Lactams
Stefan France, Meha H. Shah, Anthony Weatherwax, Harald Wack,
Justine P. Roth, and Thomas Lectka*
Contribution from the Department of Chemistry, New Chemistry Building,
Johns Hopkins UniVersity, 3400 North Charles Street, Baltimore, Maryland 21218
Received September 24, 2004; E-mail: lectka@jhu.edu
Abstract: We report a mechanistically based study of bifunctional catalyst systems in which chiral
nucleophiles work in conjunction with Lewis acids to produce â-lactams in high chemical yield, diastereo-
selectivity, and enantioselectivity. Chiral cinchona alkaloid derivatives work best when paired with Lewis
acids based on Al(III), Zn(II), Sc(III), and, most notably, In(III). Homogeneous bifunctional catalysts, in which
the catalyst contains both Lewis acidic and Lewis basic sites, were also studied in detail. Mechanistic
evidence allows us to conclude that the chiral nucleophiles form zwitterionic enolates that react with metal-
coordinated imines. Alternative scenarios, which postulated metal-bound enolates, were disfavored on the
basis of our observations.
Scheme 1. Tandem Lewis Acid/Nucleophile Catalyzed Synthesis
of â-Lactams
Introduction
As prime examples of polyfunctional catalyst systems,
enzymes provide precisely arrayed functional groups operating
in concert on a substrate within an active site.1 Metalloenzymes
are an interesting subset of polyfunctional catalysts that utilize
metal ions as Lewis acids and/or redox centers in conjunction
with “organic” functional groups to enhance reaction rates. Zinc
proteases are especially notable examples;2 these enzymes
efficiently chelate zinc hydrates and facilitate nucleophilic attack
on the substrate. Synthetic chemists have long sought to employ
the principles of polyfunctional catalysts (especially bifunctional
catalysts) in designed systems as well.3 It has only been in recent
years that substantial progress has been made in the use of well-
defined bifunctional systems as applied to asymmetric catalysis.4
Several of these systems combine a metal ion (usually as part
of a chiral Lewis acid complex) with a Lewis base in the form
of an organic functional group.5 The precise mechanism by
which synthetic bifunctional systems act is often unknown, due
in part to their inherent complexity. For our part, we made the
recent discovery that a chiral nucleophile working in tandem
with an achiral Lewis acid (a metal salt or derived complex)
could dramatically enhance the rates and yields of the catalytic,
asymmetric â-lactam-forming cycloaddition developed in our
laboratories (Scheme 1).6
We now present a full, mechanistically based account of our
findings, buttressed by substantial kinetic and stereochemical
evidence that allows us to propose a detailed mechanism by
which the reactions occur and to confirm our original design
hypothesis involving imine activation. Our aim is to provide
one of the few instances in which an asymmetric, bifunctionally
catalyzed system has been identified and defined in detail. The
underlying principle of our system is the belief that by
combining an organic catalyst with a metal, one can enhance
reaction rates and yields, particularly in the case of traditionally
slow reactions. Theoretically, the use of a Lewis acid and a
(1) (a) Murray, G. I. J. Pathol. 2001, 195, 135-137. (b) Huang, X.; Holden,
H. M.; Raushel, F. M. Annu. ReV. Biochem. 2001, 70, 149-180.
(2) (a) Kim, D. H.; Mobashery, S. Curr. Med. Chem. 2001, 8, 959-965. (b)
Auld, D. S. Astacins 1997, 119-131. (c) Monetcucco, C.; Schiavo, G.
Trends Biochem. Sci. 1993, 18, 324-327.
(3) Bifunctional catalysis has often been defined as involving a mechanism in
which two functional catalysts (in this case a Lewis base and Lewis acid)
are involved in the rate-determining step, such that rate enhancement is
observed. The term concerted catalysis has also been proposed to describe
the simultaneous action of two catalysts and could be applied to a variant
of our system reported herein. (a) Swain, C. G.; Brown, J. F., Jr. J. Am.
Chem. Soc. 1952, 74, 2538-2543. (b) DiMauro, E. F.; Mamai, A.;
Kozlowski, M. C. Organometallics 2003, 22, 850-855.
(4) (a) Groger, H. Chem.sEur. J. 2001, 7, 5247-5257. (b) Rowlands, G. J.
Tetrahedron 2001, 57, 1865-1882. (c) Steinhagen, H.; Helmchem, G.
Angew. Chem., Int. Ed. Engl. 1996, 35, 2339-2342. (d) Jaeschke, G.;
Seebach, D. J. Org. Chem. 1998, 63, 1190-1197.
(5) (a) Arai, T.; Sasai, H.; Aoe, K.; Okamura, K.; Date, T.; Shibasaki, M.
Angew. Chem., Int. Ed. Engl. 1996, 35, 104-106. (b) Woolley, P. J. Chem.
Soc., Chem. Commun. 1975, 14, 579-580. (c) Dong, S. D.; Breslow, R.
Tetrahedron Lett. 1998, 39, 9343-9346. (d) Noyori, R.; Yamakawa, M.;
Hashiguchi, S. J. Org. Chem. 2001, 66, 7931-7944.
(6) France, S.; Wack, H.; Hafez, A. M.; Taggi, A. E.; Witsil, D. R.; Lectka, T.
Org. Lett. 2002, 4, 1603-1605.
9
1206
J. AM. CHEM. SOC. 2005, 127, 1206-1215
10.1021/ja044179f CCC: $30.25 © 2005 American Chemical Society

Lewis Acid-Nucleophile-Based Asymmetric Catalysis
A R T I C L E S
Lewis base in concert could also lead to a self-quenching
reaction, but when the right pair is combined, for example a
hard metal ion with a soft base (using Pearson’s terminology),
reaction rates may increase significantly.7 The most important
precedent for this is found in the work of Aggarwal on catalysis
of the Baylis-Hillman reaction through the use of an achiral
Lewis base, such as dabco, together with a lanthanide metal
triflate salt, to yield the desired products with enhanced rates.8
efficient bifunctional system for the asymmetric synthesis of
â-lactams.
Results and Discussion
Metal Screening. A while ago we discovered that ben-
zoylquinine (BQ), in the presence of a stoichiometric base, could
catalyze the formation of â-lactams with high enantioselectivity
and moderate to good chemical yields, employing acid chlorides
as ketene precursors or “equivalents” and R-imino esters as
reaction partners. The mechanistic hypothesis we favor for
â-lactam cycloaddition reactions involves the formation of
zwitterionic enolates that add to the imino ester nucleophilically
to form reactive intermediates leading to the desired â-lactam
products (eq 1).17 Much of the remaining mass balance of the
reaction is attributable to polymerized acid chloride and imino
ester. It occurred to us that an additional mode of activation
may enhance chemical yields; for example, if a Lewis acid
cocatalyst were also present in the reaction to activate the imino
ester further, an enhancement in the yield of desired products
by alteration of the reaction manifold could be expected.
Background. The use of “rationally designed” bifunctional
Lewis acid/Lewis base systems in asymmetric catalysis has only
recently come of age. This new methodology can be divided
into two main classes: two-component systems, in which Lewis
acids and Lewis bases work in tandem; and homogeneous
bifunctional catalysts. However, there are relatively few ex-
amples of the combination of achiral Lewis acids and chiral
Lewis bases. Of note, Shi has reported an efficient aza-Baylis-
Hillman reaction in which Ti(Oi-Pr)4 is used to catalyze imine
formation while a cinchona alkaloid derivative attacks the enone
substrate.9 Similarly, a chiral Lewis acid and an achiral Lewis
base can be utilized for the cyanosilylation of aldehydes,10 as
Chen et al. have shown. In this case the authors employed an
achiral N-oxide in conjunction with a chiral titanium complex
to achieve catalysis. Similarly, Yamamoto has successfully
applied a tandem Lewis acid/Lewis base system to both the
Sakurai-Hosomi allylation and the Mukaiyama aldol reaction.11
His system involves the use of a chiral Ag-phosphine complex
and a catalytic amount of fluoride ion. On the other hand, use
of chiral Lewis acids in tandem with chiral Lewis bases has
been relatively unexplored in organic synthesis.
Bifunctional catalysts, ones that contain both Lewis acidic
and basic sites, can play two possible roles. They can either
activate both reagents and substrates or one reactive center can
bind the substrate while another center performs the transforma-
tion. Shibasaki has designed a series of bifunctional catalysts
that incorporate Lewis basic sites onto a binol template.12 His
catalyst system has been applied to asymmetric hydrophospho-
nylation, as well as the cyanosilylation of aldehydes and related
Michael reactions. Baeza et al. have designed a similar system
for the cyanophosphorylation of aldehydes.13 Hoveyda has
employed a small peptide with a Lewis acidic site for imine
chelation to effect an asymmetric Strecker reaction.14 Kozlow-
ski’s dialkylzinc addition to keto esters utilizes a titanium-
salen complex containing nucleophilic appendages to effect the
transformation.15 Additionally, Annunziata et al.16 have reported
a related bifunctional system in which a Lewis acid and
nucleophile act in sequential reactions but not in tandem. These
examples offered us some background for the design of an
We drew upon our extensive experience with the activation
of R-imino esters by chiral Lewis acids for guidance. Our
previous work in this area detailed the catalytic, asymmetric
alkylation of R-imino esters and N,O-acetals with enol silanes,
silyl ketene acetals, alkenes, and allylsilanes to form a wide
variety of R- and â-amino acid derivatives in high chemical
yield, high anti-diastereoselectivity, and uniformly high enan-
tioselectivity using chiral late transition metal bis(phosphine)
complexes as catalysts (Scheme 2).18
Our catalyst of choice, easily prepared from Cu(I) salts and
(R)- or (S)-tol-binap, has also been exploited by other groups,
demonstrating the wide scope, flexibility, and utility of this chiral
Lewis acid system.19 We began our studies by adding sub-
stoichiometric amounts of metal cocatalysts. We chose the late
transition metal based salts and complexes that worked well in
our Lewis acid-catalyzed imino ester chemistry, including
(7) Pearson, R. G. J. Am. Chem. Soc. 1963, 85, 3533-3539.
(8) (a) Aggarwal, V. J. Org. Chem. 1998, 63, 7183-7189. (b) Aggarwal, V.
J. Org. Chem. 2002, 67, 510-514.
(9) Shi, M.; Jiang, J.-K. Tetrahedron: Asymmetry 2002, 13, 1941-1947.
(10) Chen, F.; Feng, X.; Qin, B.; Zhang, G.; Jiang, Y. Org. Lett. 2003, 5, 949-
952.
(17) Taggi, A. E.; Hafez, A. M.; Wack, H.; Young, B.; Drury, W. J., III; Lectka,
T. J. Am. Chem. Soc. 2000, 122, 7831-7832.
(11) Wadamoto, M.; Ozasa, N.; Yanagisawa, A.; Yamamoto, H. J. Org. Chem.
2003, 68, 5593-5601.
(18) (a) Ferraris, D.; Young, B.; Dudding, T.; Lectka, T. J. Am. Chem. Soc.
1998, 120, 4548-4549. (b) Ferraris, D.; Young, B.; Cox, C.; Drury, W. J.,
III; Dudding, T.; Lectka, T. J. Org. Chem. 1998, 63, 6090-6091. (c) Drury,
W. J., III; Ferraris, D.; Cox, C.; Young, B.; Lectka, T. J. Am. Chem. Soc.
1998, 120, 11006-11007. (d) Ferraris, D.; Dudding, T.; Young, B.; Drury,
W. J., III; Lectka, T. J. Org. Chem. 1999, 64, 2168-2169.
(12) (a) Shibasaki, M.; Sasai, H.; Arai, T. Angew. Chem., Int. Ed. Engl. 1997,
36, 1236-1256. (b) Nogami, H.; Kanai, M.; Shibasaki, M. Chem. Pharm.
Bull. 2003, 51, 702-709.
(13) Baeza, A.; Casas, J.; Najera, C.; Sansano, J. M.; Saa, J. M. Angew. Chem.,
Int. Ed. 2003, 42, 3143-3146.
(19) (a) Yao, S.; Fang, X.; Jørgensen, K. A. J. Chem. Soc., Chem. Commun.
1998, 2547-2548. (b) Fang, X.; Johannsen, M.; Yao, S.; Gathergood, N.;
Hazell, R. G.; Jørgensen, K. A. J. Org. Chem. 1999, 64, 4844-4849. (c)
Saaby, A.; Fang, X.; Gathergood, N.; Jørgensen, K. A. Angew. Chem., Int.
Ed. 2000, 39, 4114-4115. (d) Yao, S.; Saaby, S.; Hazell, R. G.; Jørgensen,
K. A. Chem.sEur. J. 2000, 6, 2435-2448.
(14) Josephsohn, N. S.; Kuntz, K. W.; Snapper, M. L.; Hoveyda, A. H. J. Am.
Chem. Soc. 2001, 123, 11594-11599.
(15) DiMauro, E. F.; Kozlowski, M. C. Org. Lett. 2002, 4, 3781-3784.
(16) Annunziata, R.; Benaglia, M.; Cinquini, M.; Cozzi, F.; Puglisi, A. Eur. J.
Org. Chem. 2003, 8, 1428-1432.
9
J. AM. CHEM. SOC. VOL. 127, NO. 4, 2005 1207

A R T I C L E S
France et al.
Scheme 2. Catalytic, Asymmetric Alkylation of R-Imino Esters and
N,O-Acetals
Additionally, when Cu(PPh3)2ClO4 was added to a solution of
BQ in toluene, the color rapidly changed from clear to blue,
again implying that the BQ is acting as a ligand and possibly
promoting the oxidation of the metal. We also tried Mg(II),
Sn(II), and Cu(II) salts; however, each afforded poor yield of
desired product (Table 1, entries 3-7). Similarly disappointing
results were observed when Pd(II) and Ni(II) were screened as
cocatalysts. With these other Lewis acids, the poor solubility
of the catalysts may also be at least partially to blame as well.
Table 1. Initial Metal Screen To Form â-Lactam 6a with
Phenylacetyl Chloride and Imino Ester 5a, Catalyzed by BQ 3a in
Toluene
Screening of Triflate Salts and Other Metal Complexes.
In his studies of metals for the Baylis-Hillman reaction,
Aggarwal screened a large number of metal salts and found
that both group III and lanthanide metals gave the best results
in his reactions. These facts, coupled with our own results,
suggested to us that less azaphilic metal salts and complexes
would be desirable to diminish potential binding to the quinu-
clidine nitrogen of the BQ. Although the lanthanides did not
work well for us (Table 1, entries 8 and 10-12), to our
satisfaction, we found that using the metal triflates of Sc(III),
Al(III), Zn(II), and In(III) (10 mol %) resulted in increased rates
of product formation and significantly enhanced chemical yields
(Table 1, entries 16-20). Overall the use of Zn(OTf)2 (85%)
and In(OTf)3 (95%) resulted in the greatest increase in chemical
yield, while Al(OTf)3 (78%) and Sc(OTf)3 (80%) were slightly
less effective but still very promising.
Results with Indium(III). Having established that In(OTf)3
was the best overall Lewis acid cocatalyst for promoting our
reaction,21 we applied this system to various substrates to
determine its applicability and effect on the reaction enantio-
selectivity (ee) and diastereoselectivity (dr) (Figure 1). Generally
speaking, reaction ee’s are maintained in the high 90’s and
chemical yields increased by a factor of 1.5-2. For example,
6a was obtained in 95% yield and 98% ee (dr ) 60:1), which
represents a 1.5-fold yield increase (Figure 1). In the case of
6d we were even able to double the yield without noticeable
decrease in selectivity. An area of minor concern was the
diastereoselectivity of a few of our â-lactam products; in many
cases the dr was clearly maintained with the use of In(III) as a
cocatalyst, but in a minority of cases it eroded somewhat (for
example lactams 6b and 6f) from 50:1 to around 10:1.
Nevertheless, we thought it was a problem worth addressing to
ensure high levels of selectivity for all the reactions.22
entry
Lewis acida
% yieldb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
none
65
27
30
32
35
36
44
45
49
50
52
55
56
63
64
67
78
80
85
95
(Ph3P)3RhOTf
(Ph3P)2PdCl2
Ni(dppe)Cl2
Cu(OTf)2
Mg(OTf)2
Sn(OTf)2
Sm(OTf)3
Cu(Ph3P)2ClO4
Yb(OTf)3
YbCl3
Eu(fod)3
Al(OiPr)3
La(OTf)3
TBDMSOTf
AgOTf
Al(OTf)3
Sc(OTf)3
Zn(OTf)2
In(OTf)3
a 10 mol % of the Lewis acid was added to the standard reaction (entry
1). b Isolated yield after column chromatography.
phosphine-Rh(I) complexes and the precursor salt Cu(PPh3)2-
ClO4. Our rationale was based on the fact that these catalysts
were established to work well at imino ester activation.
However, to our surprise, when we added 10 mol % of these
metal complexes to the standard reaction conditions to form
â-lactam 6a, the overall yield of the reaction decreased in the
presence of the respective metal (Table 1, entries 2 and 9).20
This result is almost certainly due to the preferential binding
of late transition metals such as Cu(I) and Rh(I) to the tertiary
amine catalyst; Cu(I) is known to have a high affinity for amines,
and when we monitored the reaction by UV, a characteristic
shift indicative of metal binding took place, as well as a gradual
shift to characteristic bands of Cu(II), indicating metal oxidation.
The binding of the metal to the quinuclidine nitrogen of benzoyl-
quinine (BQ) naturally reduces its efficacy as a nucleophilic
catalyst and results in low yields. This is an example of the
“self-quenching” conundrum that we had mentioned earlier (eq
2) and an example of why correlations between Lewis acidity
and imino ester binding, for example, cannot easily be drawn.
The coordination chemistry of In(III) is a scantily explored
topic.24 One reason for this may be the fact that In(III) seemingly
binds to many ligands with comparatively low affinity and is
(21) For a recent review on indium-catalyzed reactions, see the following:
Chauhan, K. K.; Frost, C. G. J. Chem. Soc., Perkin Trans. 1 2000, 3015-
3019.
(22) We recently reported the synthesis of a C2-symmetric bis(cyclophane diol)
ligand and its application toward improving the diastereoselectivity of our
bifunctional system: Wack, H.; France, S.; Hafez, A. M.; Drury, W. J.,
III; Weatherwax, A.; Lectka, T. J. Org. Chem. 2004, 69, 4531-4533.
(23) Yields in red represent isolated yields from original reaction without the
added Lewis acid. Yields in blue represent isolated yields employing
In(OTf)3 as a cocatalyst.
(20) Standard reaction conditions consisted of 10 mol % BQ 3a, 1 equiv of
proton sponge as a stoichiometric base, 1 equiv of imino ester 5a, and 1
equiv of acid chloride in toluene at -78 °C.
(24) Carty, A. J.; Tuck, D. G. Prog. Inorg. Chem. 1975, 19, 243-337.
9
1208 J. AM. CHEM. SOC. VOL. 127, NO. 4, 2005

Lewis Acid-Nucleophile-Based Asymmetric Catalysis
A R T I C L E S
bidentate ligands and one monodentate ligand.29 This fact may
provide the key to understanding why In(III) works so efficiently
as compared to our other “active” metals (Al, Sc, and Zn).
Ultimately, these fast on/off rates translate into a lower
thermodynamic barrier to ligand exchange. In the Cu(I)-
catalyzed system, the quinuclidine nitrogen of BQ acts in a
deleterious manner by tightly binding the metal ion, while, with
In(III), the binding to the amine is highly reversible and of
perhaps lower affinity. Therefore, regardless of how well the
ligand is activated, it will dissociate from the metal rapidly.
However, this argument raises the question: how does one
account for the increased yields with the other “active” metals
when each is known to have slower on/off exchange rates? The
answer may be fairly straightforward for scandium and alumi-
num. Both are highly oxophilic Lewis acids and therefore are
less likely to bind the quinuclidine nitrogen. Zinc, however, is
a bit more problematic, being an inherently softer Lewis acid
than Sc(III) or Al(III) and is yet somewhat azaphilic. It should,
therefore, bind preferentially to the quinuclidine nitrogen and
thus lower the chemical yield. Since we are conversely observing
an increased yield, a plausible explanation would be that ligand
exchange between the monodentate quinuclidine nitrogen and
the bidentate imino ester favors the latter (see below).
Lowering Metal Catalyst Loading. Although 10 mol % of
the In(OTf)3 was added to the reaction in our standard screening,
we observed that most (if not all) of the salt did not dissolve
completely in the toluene solventsa seemingly “homeopathic”
chemical system! Even when reagents and catalysts were added
(such as BQ), the indium was not well solubilized (in a way,
this was good news, indicating that tight binding of BQ to the
metal was not occurring). This result implies that less cocatalyst
is actually required. A titration experiment indicated that at any
one time only about 10% of the added metal catalyst was
dissolved in the reaction mixture (1 mol % cocatalyst), but it
was enough to enhance the results of the reaction dramatically.
In an attempt to rectify the problem, we employed THF, in
which In(OTf)3 is soluble. However, it led to eroded enantio-
selectivities (85-90% ee’s on average).
Figure 1. â-Lactams synthesized by employing BQ and In(OTf)3 co-
catalyst.23
characterized by fast “on/off” rates.25 However, it binds halide
ions with relatively high affinity. Although its trihalide salts
have been employed as catalysts in numerous reactions,26 few
reports of catalytic systems using In(III) with an organic ligand
have been published, and no effective chiral In(III)-based Lewis
acid catalytic systems are known as of this writing. Ligand
exchange is slow when InX3 is combined with oxo- or aza-
based ligands, such that mixed-ligand species are formed
(InXnLn). Furthermore, In(III) can bind phosphines to form four-
coordinate species and nitrogen- and oxygen-derived ligands
to produce five- and six-coordinate species.27 When bound to
bidentate ligands, In(III) is known to form trigonal bipyramidal
chelates of the type InL3 (where L ) a bidentate ligand).
Whereas most Lewis acids would be deactivated or degraded
in aqueous solutions, In(III) (most often InCl3) binds water
highly reversibly, and/or with low affinity, thus maintaining its
activity. Consequently, we were concerned about the possible
effects of the free chloride ions that are generated during the
course of our reactions. Because of its poor solubility in our
solvent, in situ generation of InCl3 could be detrimental in that
we might not observe the expected increase in yield. Fortunately,
with proton sponge as our base, most chloride ions are
presumably “soaked up” and precipitate as the proton sponge-
hydrochloride salt. Additionally, this alleviates the issue of
coordination of chloride ions to indium.28
Considering that the quantity of fully soluble metal salt
actually needed in the reaction would be considerably less than
10 mol %, we attempted to make the indium salt more soluble
by changing the counterion to something “greasier” and bulkier
than the triflate. Both tetraphenylborate30 and “BArF”31 were
found to be suitable counterions to solubilize the metal, although
Most studies have found that the exchange rate is often so
rapid between the free ligand and InL3 that it has been
hypothesized that the intermediate complex is comprised of two
(25) Chung, H. L.; Tuck, D. G. Can. J. Chem. 1974, 52, 3944-3949.
(26) Ranu, B. C. Eur. J. Org. Chem. 2000, 2347-2356.
(29) Tuck, D. G. Coord. Chem. ReV. 1966, 1, 286-291. (b) Tuck, D. G. In
ComprehensiVe Coordination Chemistry; Wilkinson, G., Ed.; Pergamon
Press: New York, 1987; Vol. 3, Chapter 25.2. (c) Pedrosa de Jesus, J. D.
In ComprehensiVe Coordination Chemistry; Wilkinson, G., Ed.; Pergamon
Press: New York, 1987; Vol. 2, Chapter 15.7.
(27) Paver, M. A.; Russell, C. A.; Wright, D. S. In ComprehensiVe Organo-
metallic Chemistry II; Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.;
Elsevier Science: New York, 1995; Vol. 1, Chapter 11.
(28) No change in the UV spectrum of proton sponge was observed upon addition
of In(OTf)3 in THF (a solvent in which it is fully soluble).
(30) Gillard, R. D.; Vaughan, D. H. Transition Met. Chem. 1978, 3, 44-48.
9
J. AM. CHEM. SOC. VOL. 127, NO. 4, 2005 1209

A R T I C L E S
France et al.
Scheme 3. Metal-Mediated Bifurcated Pathway for the Formation
of â-Lactams
we suspect that in some cases the tetraphenylborate ion
decomposes over the course of the reaction, a common
occurrence.32 Although the salts could be generated in situ
through a phase transfer process, we found that isolating them
led to higher yield and ee (eq 3). We achieved greater success
with the In(BArF)3 salt 8, which, when loaded at 5 mol % gave
â-lactam 6a in 80% yield and 95% ee.
Alternatively, an interesting, soluble indium salt could be
made in which the quinine unit was incorporated as part of the
negatively charged counterion. For example, reaction of the
sodium salt of quinine followed by treatment with triphenyl-
borane and metal metathesis with In(OTf)3 produces the mixed
In(III) quinine salt 3b (eq 4). Use of 3b in a standard reaction
results in product with high ee (98%), but only slightly enhanced
yields (ca. 70%).
fluoro-substituted salicylate derivative 3f. As compared to the
spectrum of the metal-free catalyst, studies revealed a slight
shift upfield from 6.0 to 5.7 ppm when 1 equiv of In(OTf)3
was added incrementally.34 We can assume that the appearance
of only one peak in this metal experiment is the result of a
very rapid exchange rate between the free Lewis acid and the
In(III)-salicylate complex, a scenario consistent with the known
coordinating properties of indium. A similar study was per-
formed with zinc as the binding metal. Diethylzinc was heated
with the fluorinated quinine-salicylate to yield Zn-3f, an
uncharged complex. A 19F shift was seen from 6.0 to 4.8 ppm.
Addition of excess ligand 3e produced a separate set of NMR
resonances, indicating a slower metal exchange process, a
marked contrast to the In(III) complex.
Mechanistic Studies on In(III)-Based Catalysis. In our
initial studies, we discovered that the addition of a metal salt
does not increase the rate of acid chloride consumption in the
reaction of 1a and 5a (10 mol % 3a as co-catalyst) with proton
sponge as the stoichiometric basesan observation consistent
with the established rate-determining step of the metal free
reaction with proton sponge involving acylation of the catalyst
by the acid chloride and/or ketene formation.35 These experi-
ments also confirmed that the metal is having its effect after
the rate-determining step of the reaction (Scheme 3). However,
the rate of product formation in the very initial stages of the
reaction is increased by a factor of 3 or more, resulting in a
substantial improvement in reaction chemoselectivity. The fact
that the metal cocatalyst does not participate in the rate-
determining steps (RDS) of the reaction pathway complicates
a potential kinetic analysis. Additionally, we found that the
reproducibility of our kinetic runs was called into question by
the relative insolubility of the In(III)-based salts in the reaction
medium. One way around this problem is to employ a
preemptive strategysto “knock out” the original RDS from the
picture by preformation of the ketene. The best way to do this
is to employ NaH36 as the stoichiometric base. We must also
solubilize the catalyst fully; for this purpose we chose to add
10 vol % propionitrile as a cosolvent. Under these conditions,
we found that the initial rate of reaction (product formation) in
the presence of 10 mol % In(III) was doubled over the rate of
the reaction in the absence of metal.37 The rate of reaction was
found to be dependent on the metal cocatalyst, the ketene, the
Bifunctional Catalysts. Unfortunately, the fact that most of
the metal catalyst did not dissolve in the reaction vessel made
potential mechanistic studies of the system difficult. What we
needed was a completely soluble catalyst to ensure reproduc-
ibility. Therefore, we chose to investigate homogeneous com-
plexes wherein the chiral nucleophile and Lewis acid were both
present in the same entity. Consequently, we prepared the metal
chelating alkaloid derivatives 3c and 3d in two steps from
quinine. Complexation of the sodium alkoxide salt of 3d and
0.5 equiv of the desired metal salt affords a putative bis-
(salicylato)metal complex containing two catalytically active
quinuclidine moieties.33 Complex 3e, derived from In(OTf)3,
yielded â-lactam product in 58% yield (95% ee, dr 1:1) in a
standard reaction involving phenylacetyl chloride, proton sponge
4, and imino ester 5a. The 2:1 ligand:metal stoichiometry may
have attenuated the Lewis acidity of the metal to a large extent
and led to the disappointing result. However, when the complex
derived from a 1:1 metal:ligand stoichiometry was used, it
afforded products in 90% yield, 99% ee, and 10:1 dr. Most
importantly, we now had a homogeneous system in hand with
which we could undertake mechanistic studies.
(31) (a) Brookhart, M.; Grant, B.; Volpe, A. F., Jr. Organometallics 1992, 11,
3920-3922. (b) Chavez, I.; Alvarez-Carena, A.; Molins, E.; Roig, A.;
Maniukiewicz, W.; Arancibia, A.; Arancibia, V.; Brand, H.; Manuel
Manriquez, J. J. Organomet. Chem. 2000, 601, 126-132.
(32) Nishida, H.; Takada, N.; Yoshimura, M. Bull. Chem. Soc. Jpn. 1984, 57,
2600-2604.
(33) See Supporting Information for details.
(34) NMR shifts are reported with respect to a CFCl3 standard.
(35) Taggi, A. E.; Hafez, A. M.; Wack, H.; Young, B.; Ferraris, D.; Lectka, T.
J. Am. Chem. Soc. 2002, 124, 6626-6635.
(36) Taggi, A. E.; Wack, H.; Hafez, A. M.; France, S.; Lectka, T. Org. Lett.
2002, 4, 627-629.
(37) Conversions as close to 20% as possible were measured to negate the effect
of product inhibition and catalyst degradation (which may occur through
acylation of the phenolic oxygen).
To validate the binding characteristics of our salicylate
catalysts, we conducted a series of 19F NMR experiments with
9
1210 J. AM. CHEM. SOC. VOL. 127, NO. 4, 2005

Lewis Acid-Nucleophile-Based Asymmetric Catalysis
A R T I C L E S
Scheme 4. Kinetic Isotope Study To Identify the Rate-Determining
Step
from the outset (our “paper concept”) is the idea that the metal
activates the imine toward nucleophilic attack by the organic
zwitterionic enolate, thereby enhancing the rate of product
formation (structure A), which is also consistent with the kinetic
data that we have amassed.39
The second possibility is that the metal binds to the
zwitterionic enolate, making it more chemoselective in its
reactivity. The problem with this scenario is that the enolate
reactivity should be lowered by metal binding, with concomitant
increase in its thermodynamic stability. The formation of this
chelated intermediate should be favored from a thermodynamic
standpoint, assuming an equilibrium between the ketene and
the enolate (structure B).40 Metal binding is thus expected to
increase the concentration of enolate relative to ketene or acid
chloride, paradoxically accounting for an increase in the rate
of reaction (eq 6)
imine, and the cinchona alkaloid (all on a first-order basis).
Complicating any analysis is the fact that the rate equation must
have a term for the base (metal-free) reaction, which proceeds
in parallel with the catalyzed reaction (eq 5). Additionally, we
note that the propionitrile cosolvent (1) erodes the reaction
diastereoselectivity, (2) increases the initial rate of the base
reaction, and (3) lowers the rate of the metal catalyzed process
(presumably by reducing the Lewis acidity of the metal salt).
Thus, its use is not ideal from either a synthetic or a mechanistic
standpoint.
Naturally the potential solution involves the use of the
quinine-salicylate ligand 3c (10 mol %) as a basis for
mechanistic studies. Ligand 3c also serves as the “shuttle” base
in the preliminary ketene synthesis. In a standard experiment,
we “preform” phenylketene in a reaction vessel from phenyl-
acetyl chloride over enough time to ensure complete acid
chloride consumption. In the control experiment, imino ester
5a is added to the reaction mixture containing 3c and the
“preformed” ketene. The rate of product formation is then
measured over the course of time. Concurrently, imino ester
5a is added to a ketene solution containing the metal complex
In(III)-3c. Under these conditions, the data show that the rate
of product formation is linearly dependent on both the concen-
tration of metal complex and the imino ester 5a.
Which step is now rate determining? Is it the C-C lactam
bond formation or a “downstream” transacylation step leading
to ring formation? A kinetic isotope study addresses this question
(Scheme 4). Once again, preformation of phenylketene removes
acylation of catalyst as the RDS, whereas carbon-carbon bond
formation and transacylation remain as potential RDS’s. Phe-
nylketene-d1 is expected to give rise to a negligible KIE if the
latter is the RDS, but an inverse secondary effect would be
observed if C-C bond formation is the RDS, as a change in
hybridization from sp2 to sp3 takes place. In the actual
experiment, we found a KIE of 0.8, indicating that C-C bond
formation is likely the RDS.38
The third possibility is that both alternatives could be
operating, with the metal organizing both enolate and imine in
a termolecular activated complex (structure C).
Addressing Scenario A. Metal Binding to the Imino Ester.
In addition to kinetic data, we have amassed physical and
chemical evidence that metal binding to the imino ester plays
a pivotal role in the reaction. For example, as potentially
chelating substrates, the coordination of metals and metal
complexes to R-imino esters is well precedented in both our
work and the work of others.41 Given the fact that metals will
bind the catalyst as well as the imine, it is often difficult to
correlate metal-binding capacity with catalytic efficacy. Nev-
ertheless, we sought to examine, through IR and NMR studies,
the binding of catalytically active metals to imino ester substrate
5a. For example, we found that In(III) induces a small shift of
the imine proton in the 1H NMR spectrum (8.28 to 8.04 ppm),
along with a correspondingly small shift of the ester carbonyl
Role of the Lewis Acid. In our original communication on
this topic, we postulated that one of three plausible mechanistic
alternatives could account for the observed effect of the Lewis
acid in this bifunctional system. The one that we had conceived
(39) We have previously found that imine 5a can be activated by late transition
metal chiral bis(phosphine) complexes to catalyze the enantio- and
diastereoselective addition of various nucleophiles: Ferraris, D.; Young,
B.; Cox, C.; Dudding, T.; Drury, W. J., III; Ryzhkov, L.; Taggi, A. E.;
Lectka, T. J. Am. Chem. Soc. 2002, 124, 67-77.
(38) Due to the number of potential manifolds in our reaction pathway, no one
step can be defined as the only rate-determining step. The calculated KIE
should be considered as an aggregate resulting from an ensemble of RDS’s,
which is in fact a common but underappreciated occurrence in complex
chemical reactions.
(40) It is possible that the metal salt rather than catalyzing the formation of
product suppresses the formation of undesired side products.
(41) (a) Yamamoto, Y.; Ito, W. Tetrahedron 1988, 44, 5415-5423. (b) Juhl,
K.; Hazell, R. G.; Jørgensen, K. A. J. Chem. Soc., Perkin Trans. 1 1999,
16, 2293-2297.
9
J. AM. CHEM. SOC. VOL. 127, NO. 4, 2005 1211

A R T I C L E S
France et al.
stretch (from 1751 to 1748 cm-1), and a larger shift of the imine
stretch (1630 to 1625 cm-1) in the IR spectrum. Zn(II), another
1
efficacious metal, produces stronger changes in H chemical
shifts (8.28 to 7.92 ppm) and IR frequencies (from 1751 to 1741
cm-1 and from 1630 to 1597 cm-1, respectively). Thus, it looks
like Zn(II) is a better Lewis acid than In(III) towards the imino
ester, but it has lower on/off rates and perhaps a higher affinity
for the quinuclidine nitrogen.
Another IR study was performed to understand the role of
In(III) in our reaction. We had previously seen by NMR that
ligand 3c and In(III) form a complex when they are present in
equimolar concentrations, but what happens when the imino
ester is also present, as it is in our reaction? When the complex
is in solution along with the imino ester in a 1:1 ratio, we
observe a change in the IR spectra, more dramatic than with
just the metal. The imino ester carbonyl stretch now shifts from
1751 to 1746 cm-1, and the imine stretch now appears at 1621
cm-1 (from 1630 cm-1). The carbonyl stretch for ligand 3c also
shifts in the presence of In(III) and the imine (from 1684 to
1694 cm-1).
However, we have designed what we believe to be a more
elegant test for productive metal binding to imino ester
substrates. Imino esters 5b and c, derived from ethyl glyoxylate
and p- and o-methoxyaniline, respectively, are well-behaved
substrates for the â-lactam forming reaction in the absence of
metal cocatalysts. As expected, in the absence of metal, substrate
5b reacts slightly faster than imino ester 5c, which is more
hindered due to the proximity of the methoxy group to the imine
nitrogen (1.5:1 relative initial rate of formation of lactams 9a
to 9b, eqs 7 and 8). On the other hand, in the presence of In-
(OTf)3 and Zn(OTf)2 cocatalysts (10 mol %), the reactivity is
reversedssubstrate 5c now reacts significantly faster (5:20 ratio
for the relative rate of formation of 9a to 9b by NMR with
In(OTf)3). By virtue of its o-methoxy group, imino ester 5c is
tridentate in nature and, therefore, a better ligand for metals in
general. Thus, if the metal activates the imine, all other things
being equal, the better metal binder should be more reactive.
Furthermore, since we have established that the metal cocatalyst
increases the initial reaction rate, the chelated imine system
should result in much faster product formation. (It should also
be noted that imine 5c can solubilize In(OTf)3 fairly well in
toluene conditions, whereas 5b does not.)
Addressing Metal Binding to the Enolate and Metal Ion
Catalysis of Ketene Dimerization. We also attempted to
examine the interaction of metal ions (especially In(III) and
Zn(II)) with ketenes and their zwitterionic enolates, thus coming
at the problem from another angle. However, we were con-
strained by an unanticipated discovery; we found that when
preformed ketene solutions are treated with 10 mol % In(OTf)3
in toluene, a fast dimerization reaction ensues to afford high
yields of the corresponding unsymmetrical ketene dimers in
racemic form. Surprisingly enough, Lewis acid catalyzed ketene
dimerization appears not to be a mechanistically well-studied
reaction, although a few examples are known in the litera-
ture.42 In our case, it seems that if free ketene is generated
during our lactam-forming process, the corresponding reac-
tion with the imino ester is much faster than dimerization,
which does not appear to be a problem under our reaction
conditions.
Ultimately, we found that by moving to a more stable ketene,
we were able to examine the interaction of this species with
metal ions in the presence of BQ. In previous work, we
characterized the zwitterionic enolate complex of ketene 2b with
BQ by IR and UV-Vis spectroscopy. We found that ketene
2b is in equilibrium with complex 10 and that the equilibrium
constant, even for such an electrophilic ketene, was still <1.
An important question arises: how would a metal ion affect
the equilibrium of eq 9? Although a metal-bound zwitterionic
enolate should not be chemically more reactive than the
corresponding unbound form, it may be more selective. Ad-
ditionally, as discussed earlier, the presence of a metal such as
In(III) may increase the amount of enolate that forms, thus
increasing the rate of reaction through a perturbation of
equilibrium (eq 10). We addressed this question by an examina-
tion of the effect that In(III) has on the relative concentrations
of enolate and ketene. We discovered that the perturbation using
ketene 2b, BQ 3a, and In(OTf)3 is small (as measured by UV-
Vis spectroscopy). Even with 1 equiv of In(III) in soluble form
(THF solvent), the amount of ketene did not significantly
diminish. As we would expect enolate 10, substituted with
carboethoxy groups, to be a good metal binder, the effect of
metal on the equilibrium constant of simple ketenes in enolate
formation may be negligible in comparison.
(42) Wynberg, H.; Staring, E. G. J. Am. Chem. Soc. 1982, 104, 166-168.
9
1212 J. AM. CHEM. SOC. VOL. 127, NO. 4, 2005

Lewis Acid-Nucleophile-Based Asymmetric Catalysis
A R T I C L E S
Reaction Mechanism: Addressing the Role of “Free”
Metal. A plausible reaction mechanism involving catalyst 3c
must consider the possibility that both free and complexed
indium takes part in the reaction. In an attempt to elucidate the
roles of both ligand and metal, we performed standard UV-
Vis experiments in which their proportions were varied. We
determined that when the ligand to metal stoichiometry in THF
was 1:1, substantial complexation occurred but that a small
amount of free ligand remained. This fact implies that some
“free” or very loosely bound metal is also in solution and plays
a role in our overall reaction rate. Free metal is almost certainly
in fast exchange with bound metal. In toluene, the quantity of
free metal is considerably lower as the ligand binding constant
increases, but the rate of reaction is approximately the same as
it is in THF (namely, the initial rate is ∼5 times the rate without
metal)! We can conclude that in toluene, the metal complex
plays a larger kinetic role than it does in THF.
Additionally, from the kinetic studies we performed in toluene
with 10% propionitrile cosolvent (a system in which the activity
of the metal is fairly attenuated due to solvation by the nitrile
and the rate of the non-metal-catalyzed reaction increases as a
result of increased polarity of the reaction medium) we estimate
that ∼20% of the product formation can be accounted for by
the non-metal-catalyzed pathway. In toluene, using ligand 3c
and In(III) as catalysts, the non-metal pathway drops consider-
ably, and therefore, the majority of the yield is attributed to the
other two components. Therefore, our overall rate equation must
be modified to include both the free metal and free ligand terms
(eq 11). The rate constants (k1-k3) in this equation include the
concentrations of all other substrates. An abbreviated proposed
mechanism for the reaction involving free In(III) is shown in
eq 12. Enolate formed from ligand (or complex In(III)-3c)
reacts with In(III)-bound imino ester 5a to produce the lactam
product.
Figure 2. Expected facial discrimination of indium complex 12 for
intramolecular bifunctional activation.
Figure 3. PM3-based model of metal enolate binding in complex 12a
showing expected si-face selectivity.
this prediction, we performed MMFF calculations on In(III)-
based enolate complexes 12a and b which should lead to re-
and si-face selectivity, respectively (Figure 2). The si-face
structure is almost 6 kcal/mol lower in energy, whereas the
corresponding re-face structure is disadvantaged by virtue of a
“flipped up” salicylate moiety, which seemingly adds strain to
the complex. Thus, our prediction is that if the metal binds to
the enolate oxygen, the opposite enantiomer should be formed
preferentially.
Despite running the reaction under varied conditions, we have
never observed this stereochemical reversal using complex 12
or even an attenuation of ee that would accompany such a
pathway. We also know that the association constants of a
variety of metals (including In(III)) with the salicylate group
would preclude the presence of much free metal salt in
solutionscertainly too little would be present to account for
the observed results. Thus, it appears as though intramolecular
binding of the metal to the enolate oxygen can once again be
disfavored at this point. This conclusion can also be rationalized
by considering that a fair amount of strain energy is imparted
in an activated complex such as 12a. Figure 3 shows a model
of a PM3 calculation that we performed on complex 12a,
showing how binding of the metal to the enolate oxygen would
enforce si-face selectivity.
Reaction Mechanism: Addressing Scenarios B and C. An
Alternative Test for Metal Binding to the Enolate. Upon
further consideration we realized that we also had a potential
test in hand for the diagnosis of intramolecular metal binding
of In(III) to the enolate oxygen in complex 12. Consider
complex In(III)-3d, in which a metal binding unit is attached
to quinine. If the metal does not bind to the oxygen atom of
the enolate simultaneously, we would expect re-face selectivity
(using our molecular mechanics calculations as a guide) as we
have invariably seen in BQ-catalyzed reactions. On the other
hand, if the metal binds to the oxygen atom of the enolate, for
geometric reasons, the ketene faces should flip as the oxygen
moves to bind the metal. This would lead to si-face exposure
and a clear reversal of the observed stereoselectivity. To make
What then about intermolecular binding of the enolate oxygen
through the metal center of another complex molecule? This is
9
J. AM. CHEM. SOC. VOL. 127, NO. 4, 2005 1213

A R T I C L E S
France et al.
Scheme 5. Proposed Mechanism for the Synthesis of â-Lactams
Catalyzed by Metal Complex In(III)-3d
where the labor involved in our kinetics experiments finally
pays off. For example, we would anticipate that if another
molecule of complex In(III)-3d were involved in the RDS, the
rate dependence should vary as the square of the catalyst’s
concentration. That it does not argues against this enolate metal
binding scenario.
We know from kinetics experiments that C-C bond forma-
tion occurs in the RDS, the stereochemistry-determining step
in the cyclization reaction. The RDS must also involve the
catalyst and the enolate. The same conclusions apply if we
consider that the imino ester is activated by another complex
molecule as well.
Reaction Mechanism: An Alternative Way of Ruling Out
Metal/Enolate Binding. Catalytic Halogenation Reactions.
A reaction that provides evidence to support our metal binding
hypothesis is the catalytic, enantioselective chlorination of acid
chlorides using benzoylquinine as the catalyst and the electro-
philic, perchlorinated orthoquinone 14 as a halogenating agent
(eq 13).43 When we added a large variety of metal salts and
complexes in an attempt to increase yields, no effect whatsoever
was observed. If enolate stabilization were playing a role, we
would, of course, expect the same effect as is observed in the
â-lactam forming reactions.
the RDS. Finally, transacylation followed by catalyst regenera-
tion allow for additional catalytic cycles.
Conclusion
In summary, we have characterized an efficient bifunctionally
catalyzed process for the synthesis of â-lactams that offers high
yields as well as high diastereo- and enantioselectivities without
deleterious byproduct formation by using metal salts in com-
bination with chiral cinchona alkaloid nucleophiles. In(III) was
found to be the most effective metal cocatalyst. A plausible
mechanism for the activity of free metal and ligand-bound metal
in our reaction is offered on the basis of the following
mechanistic evidence: (1) Although In(III) solubilization ex-
periments demonstrated activity at lower catalyst loadings, we
were unable to overcome the deleterious solvent effects (lowered
diastereoselectivity). (2) NMR and IR experiments offer sub-
stantive evidence for metal binding to the imino ester, strength-
ening our argument for the formation of an activated imine
complex. (3) UV-Vis data disfavor the possibility of enolate-
In(III) binding, as well as the formation of a ternary complex
with catalyst 3a. (4) Substantial binding and catalysis are
observed when complex In(III)-3c is employed in our reaction,
although there is a component of free metal catalysis in our
rate equation. (5) Comparison of the rates of lactam formation
with imines 5b and c showed that In(III) can activate the imino
ester and drastically change the relative rate of reaction. (6)
Stereochemical evidence is supported by molecular modeling
studies that disfavor intramolecular metal-enolate binding
employing In(III)-3c. These results have led to an understand-
ing of the role of In(III) as well as further clarifying our overall
rate equation and providing an example of a clearly elucidated
bifunctional catalyst system. Future expansion of this concept
will include the effects of the combination of chiral nucleophiles
Thus, we can conclude that the most plausible scenario for
bifunctional activation involves a bifunctional complex that is
consistent with scenario A, in which the metal binds to the imino
ester following formation of the zwitterionic enolate.
Mechanistic Hypothesis for Bound Metal. All of these data
finally allow us to formulate a mechanism for the bifunctional
reaction of catalyst system 3c with ketenes and imino esters to
produce â-lactams (Scheme 5). The first step involves the
formation of a non-metal-coordinated zwitterionic enolate by
the reaction of complex In(III)-3d with phenylketene. Imino
ester 5a then coordinates to the salicylate bound indium to form
a ternary complex 16 in which C-C bond formation occurs as
(43) (a) France, S.; Wack, H.; Taggi, A. E.; Hafez, A. M.; Wagerle, T. R.; Shah,
M. H.; Dusich, C. L.; Lectka, T. J. Am. Chem. Soc. 2004, 126, 4245-
4255. (b) Wack, H.; Taggi, A. E.; Hafez, A. M.; Drury, W. J., III; Lectka,
T. J. Am. Chem. Soc. 2001, 123, 1531-1532.
9
1214 J. AM. CHEM. SOC. VOL. 127, NO. 4, 2005

Lewis Acid-Nucleophile-Based Asymmetric Catalysis
A R T I C L E S
and filtered through Celite. Absorption onto silica gel followed by
column chromatography (10% EtOAc/hexanes) afforded product 6a
in 95% yield (46 mg) and 98% ee (dr 60:1).
and chiral Lewis acids, the effects of Lewis acids on other imine
substrates, and applications to bifunctional solid-phase synthesis
procedures.
Acknowledgment. T.L. thanks the NIH (Grant GM 064559),
the Sloan, Guggenheim and Dreyfus Foundations, and Merck
& Co. for generous support. S.F. and T.L. thank the Ford
Foundation, GlaxoSmithKline, and NOBCChE for support.
H.W. thanks Johns Hopkins for a Sonneborn Fellowship. The
authors would like to thank Professor John Toscano for his
insight and suggestions.
Experimental Section
General Procedure for the Tandem Nucleophile/Lewis Acid
Promoted Synthesis of â-Lactams. To a suspension of In(OTf)3 (3
mg, 0.013 mmol), benzoylquinine 3a (5.6 mg, 0.013 mmol), and proton
sponge (28 mg, 0.13 mmol) in toluene (7.5 mL) at -78 °C was added,
dropwise, phenylacetyl chloride (1a) (20 mg, 0.13 mmol) in toluene
(0.5 mL).44 A solution of imine 5a (32 mg, 0.13 mmol) in toluene (1
mL) was then added via syringe pump over 1 h. The reaction was
allowed to warm to room temperature over 16 h before it was quenched
with 1 M HCl (3 mL). The aqueous layer was then extracted twice
with CH2Cl2, and the combined organic layers were dried over MgSO4
Supporting Information Available: General procedures for
the bifunctional, catalytic synthesis of â-lactams, synthesis of
bifunctional catalysts and their complexes, and compound
characterization, kinetic data, and UV-Vis data. This material
is available free of charge via the Internet at http://pubs.acs.org.
(44) For 6d,f, the acid chloride was added to a solution of proton sponge and
BQ in 7.5 mL of toluene and stirred at 0 °C for 30 min and then cooled to
-78 °C. A suspension of In(OTf)3 and the imine (1 mL of toluene) was
added via syringe pump over 1 h.
JA044179F
9
J. AM. CHEM. SOC. VOL. 127, NO. 4, 2005 1215