The preparation and utility of bis(sulfinyl)imidoamidine ligands for the copper-catalyzed Diels-Alder reaction

Article abstract of DOI:10.1021/jo020524c

The design and preparation of a novel class of ligands based on the sulfinyl imine functionality is described. In particular, an efficient and modular synthesis of bis(sulfinyl)imidoamidine (siam) ligands is reported. The versatility of the synthetic sequence is demonstrated by the preparation of various analogues to explore the effect of substitution about the ligand framework on catalytic activity. The utility of the siam ligands in asymmetric catalysis is demonstrated in the Cu(II)-catalyzed Diels-Alder reaction where highly enantio- and diastereoselective reactions are reported for a range of N-acyloxazolidinone dienophile and diene substrate combinations. Of particular note is the efficiency of these asymmetric catalysts for reactions involving challenging and relatively unreactive acyclic diene substrates. Finally, structural data are provided for several ligands as well as metal-ligand complexes.

Full text of DOI:10.1021/jo020524c

Th e P r ep a r a tion a n d Utility of Bis(su lfin yl)im id oa m id in e Liga n d s
for th e Cop p er -Ca ta lyzed Diels-Ald er Rea ction
Timothy D. Owens, Andrew J . Souers, and J onathan A. Ellman*
Center for New Directions in Organic Synthesis, Department of Chemistry, University of California,
Berkeley, California 94720
jellman@uclink4.berkeley.edu
Received August 8, 2002
The design and preparation of a novel class of ligands based on the sulfinyl imine functionality is
described. In particular, an efficient and modular synthesis of bis(sulfinyl)imidoamidine (siam)
ligands is reported. The versatility of the synthetic sequence is demonstrated by the preparation
of various analogues to explore the effect of substitution about the ligand framework on catalytic
activity. The utility of the siam ligands in asymmetric catalysis is demonstrated in the Cu(II)-
catalyzed Diels-Alder reaction where highly enantio- and diastereoselective reactions are reported
for a range of N-acyloxazolidinone dienophile and diene substrate combinations. Of particular note
is the efficiency of these asymmetric catalysts for reactions involving challenging and relatively
unreactive acyclic diene substrates. Finally, structural data are provided for several ligands as
well as metal-ligand complexes.
In tr od u ction
The field of asymmetric catalysis has produced re-
markable results, allowing the preparation of chiral,
nonracemic products for a variety of transformations.1
While great strides have been made in the field, the
development of new chiral ligands to accelerate and direct
Lewis acid catalyzed reactions continues to be an area
of intense scrutiny. The most effective ligands are based
upon a small number of successful templates (Figure 1).
The chiral influence of these “privileged” ligand classes
is typically derived from either naturally occurring chiral
sources such as amino acids (1, 2) or tartaric acid (3, 4),
or from synthetically accessible chiral scaffolds (5, 6).
While these ligands afford highly effective catalysts, the
limited number of accessible chiral motifs often restricts
the design of new ligands. Furthermore, even for classical
reactions such as the Diels-Alder cycloaddition, unmet
challenges remain with regard to issues of reactivity and
selectivity for challenging substrates. Herein, we describe
the design of a new class of modular ligands based on
the sulfinyl imine functionality, and the application of
these ligands toward the asymmetric Diels-Alder reac-
tion.
F IGURE 1. Examples of privileged ligand structures.
acteristics. The enantiopure N-tert-butanesulfinyl imine
functionality is readily prepared by the direct condensa-
tion of N-tert-butanesulfinamide 7 with aldehydes or
ketones to afford sulfinyl imines 8.2 Both enantiomers
of 7 are available commercially and are easily prepared
Discu ssion
Su lfin yl Im in e Liga n d s. Our research into the vi-
ability of ligands incorporating the chiral N-sulfinyl imine
functionality was prompted by its many appealing char-
on greater than mole scale.3 Imines 8 are bench stable
compounds that serve as versatile intermediates for the
(1) Comprehensive Asymmetric Catalysis; J acobsen, E. J ., Pfaltz, A.,
Yamamoto, H., Eds.; Springer: New York; 1999.
(2) Liu, G.; Cogan, D. A.; Owens, T. D.; Tang, T. P.; Ellman, J . A. J .
Org. Chem. 1999, 64, 1278-1284.
10.1021/jo020524c CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/02/2002
J . Org. Chem. 2003, 68, 3-10
3

Owens et al.
construction of enantiopure amine derivatives.4 Addition-
ally, when employed as ligands, imines 8 have the
potential to coordinate metals through the sulfur, oxygen,
or nitrogen atoms, further enhancing the flexibility of this
ligand class.5 Despite these intriguing attributes, at the
time we began our studies, no research had been pub-
lished with regard to the sulfinyl imine functional group
in asymmetric catalysis.5,6
Liga n d Syn th esis. Our initial efforts to develop
sulfinyl-based ligands were focused primarily on N-tert-
butanesulfinyl imines since they can be prepared by the
Lewis acid-catalyzed condensation of 7 and a variety of
aldehydes and ketones. Several ligands were designed
in direct analogy to highly successful bisoxazoline ligands
(8a ,b,e,h ,i).7 Additionally, we desired to prepare a variety
TABLE 1. Th e P r ep a r a tion of Liga n d s 8
product Lewis acid solvent time (h) temp (°C)
yield (%)
74
58
95
8a
8b
8c
8d
8e
8e
8f
CuSO4
Ti(OEt)4 THF
CuSO4
CuSO4
CuSO4
Ti(OEt)4 THF
CuSO4
Ti(OEt)4 THF
CH2Cl2
4
4
1
18
24
3
24
4
22
75
22
22
40
22
22
22
22
22
22
CH2Cl2
CH2Cl2
CH2Cl2
98
no reaction
34
40
82
98
69
68
CH2Cl2
8f
8g
8h
8i
CuSO4
CuSO4
CuSO4
CH2Cl2
CH2Cl2
CH2Cl2
3
36
1
sulfinyl imines, excess CuSO4 is typically employed as a
Lewis acid catalyst and water scavenger for aldehyde
precursors, while Ti(OEt)4 is the reagent of choice for
ketones.2 However, for the preparation of the sterically
congested compound 8e or the electronically deactivated
imine 8f, CuSO4 was not effective in promoting the
formation of the desired sulfinyl imines. For these less
reactive substrates and for sulfinyl ketimine 8b, Ti(OEt)4
was found to be the optimal catalyst (Table 1). To
investigate the importance of steric and electronic effects
of the sulfinyl substituent, p-toluenesulfinyl imine 8i was
also prepared. Sulfinyl imines of this type are readily
available from the corresponding (S)-p-toluenesulfina-
mide 9.8
The Diels-Alder reaction of cyclopentadiene and N-
acryloyloxazolidinone 10a was chosen for an initial screen
because it has served as a benchmark for the develop-
ment of several Lewis acid catalysts (Table 2).1 While
many metal complexes of ligands 8a and 8e were initially
investigated, including Zn(OTf)2, Zn(SbF6)2, CuOTf, MgI2,
MgBr2, Co(ClO4)2, FeI3, Yb(OTf)3, Co(ClO4), Sn(OTf)2, and
Ni(ClO4)2, all imparted dramatically inferior selectivity
and/or reactivity in comparison to the Cu(OTf)2-catalyzed
reactions. Therefore, Cu(OTf)2 was utilized as the sole
Lewis acid for screening the full set of ligands 8.
A number of ligands catalyzed the desired transforma-
tion with moderate enantioselectivity. In general, C2-
symmetric ligands were preferable to their nonsymmetric
counterparts (entry 1 vs entry 3). Sulfinyl imines derived
from ketones were found to be less selective than those
derived from aldehydes (entry 1 vs entry 2). Ligand 8i,
incorporating the p-tolyl substituent at sulfur, was less
selective than the corresponding tert-butyl derivative 8a
(entry 9 vs 1). Interestingly, although ligand 8i has the
opposite configuration at sulfur, it delivers 11a with the
same absolute configuration. Finally, ligands that do not
incorporate an additional binding element (entries 4 and
of other C2-symmetric and non-C2-symmetric tert-butane-
sulfinyl imine ligands (8c,d ,f,g). For the preparation of
(3) Cogan, D. A.; Liu, G.; Kim, K.; Backes, B. J .; Ellman, J . A. J .
Am. Chem. Soc. 1998, 120, 8011-8019.
(4) (a) Liu, G.; Cogan, D. A.; Ellman, J . A. J . Am. Chem. Soc. 1997,
119, 9913-9914. (b) Tang, T.; Ellman, J . A. J . Org. Chem. 1999, 64,
12-13. (c) Cogan, D. A.; Liu, G.; Ellman, J . A. Tetrahedron 1999, 55,
8883-8904. (d) Borg, G.; Chino, M.; Ellman, J . A. Tetrahedron Lett.
2001, 42, 1433-1436. (e) Dragoli, D. R.; Burdett, M. T.; Ellman, J . A.
J . Am. Chem. Soc. 2001, 123, 10127-10128
(5) (a) Owens, T. D.; Hollander, F. J .; Oliver, A. G.; Ellman, J . A. J .
Am. Chem. Soc. 2001, 123, 1539-1540. (b) A Rh(I) sulfinyl amidine
complex has been isolated and shown to bind through the sulfinyl
sulfur and nitrogen atoms. Souers, A. J .; Owens, T. D.; Oliver, A. G.;
Hollander, F. J .; Ellman, J . A. Inorg. Chem. 2001, 40, 5299-5301.
(6) Other catalysts based on chirality solely at sulfur have been
developed. (a) Khiar, N.; Fernandez, I.; Alcudia, F. Tetrahedron Lett.
1993, 34, 123-126. (b) Tokunoh, R.; Sodeoka, M.; Aoe, K.; Shibasaki,
M. Tetrahedron Lett. 1995, 44, 8035-8038. (c) Bolm, C.; Kaugmann,
D.; Zehnder, M.; Neuburger, M. A. Tetrahedron Lett. 1996, 37, 3985-
3988. (d) Hiroi, K.; Suzuki, Y.; Kawagishi, R. Tetrahedron Lett. 1999,
40, 715-718. (e) Bolm, C.; Simic, O. J . Am. Chem. Soc. 2001, 123,
3830-3831. (f) Harmata, M.; Ghosh, S. K. Org. Lett. 2001, 3, 3321-
3323.
(7) (a) J ohnson, J . S.; Evans, D. A. Acc. Chem. Res. 2000, 33, 325-
335 and references therein. (b) Nishiyama, H.; Sakagucki, H.; Naka-
mura, T.; Horihata, M.; Kondo, M.; Itoh, K. Organometallics 1989, 8,
846-848. (c) Kanemasa, S.; Oderatoshi, Y.; Sakaguchi, S.; Yamamoto,
H.; Tanaka, J .; Wada, E.; Curran, D. P. J . Am. Chem. Soc. 1998, 120,
3074-3088.
(8) The (S)-p-toluenesulfinamide can be prepared in a single step
from Andersen’s reagent. Davis, F. A.; Zhang, Y.; Andemichael, Y.;
Fang, T.; Fanelli, T. L.; Zhang, H. J . Org. Chem. 1999, 64, 1403-1406.
4
J . Org. Chem., Vol. 68, No. 1, 2003

Preparation of Bis(sulfinyl)imidoamidine Ligands
TABLE 2. Diels-Ald er Rea ction w ith Su lfin yl Im in e
Liga n d s
entry ligand temp (°C)/time (h) yield (%) ee (%)a
dr
1
2
3
4
5
6
7
8
9
8a
8b
8c
8d
8e
8f
8g
8h
8i
-78/6
-78/6
-78/6
-78/6
-78/6
-78/6
-78/6
-78/6
-78/6
69
24
70
68
91
13
72
46
39
72 (R) 94/6
9 (R) 94/6
32 (S) 90/10
53 (S) 93/7
<5 (S) 93/7
40 (S) 95/5
12 (R) 90/10
37 (R) 97/3
37 (R) 95/5
F IGURE 3. Geometry of ligand 8b in the solid state.
a
The absolute configuration refers to the stereochemistry at
the 2-position of 11a . The ee was determined by chiral HPLC.
F IGURE 4. Pfaltz’ aza-semicorrin 12 and a general sulfinyl
imidoamidine ligand.
SCHEME 1a
F IGURE 2. Geometry of ligand 8a in the solid state.
5) afforded 11a in lower selectivity than ligand 8a , in
which the sulfinyl imine units are bridged by a pyridine
ring. Presumably, in the absence of an additional binding
element, coordination of both sulfinyl imine substituents
to produce a highly organized complex is disfavored.
X-r a y An a lysis of Liga n d s 8a a n d 8b. Analysis of
the X-ray crystal structures of ligands 8a and 8b reveals
that significant conformational changes are required for
coordination of both sulfinyl imine groups (Figure 2).
Ligand 8a is found to adopt an approximately syn-
periplanar (s-cis) configuration for the CdN-SdO unit
in the solid state, with the SdO bond nearly planar with
the CdN bond (dihedral angle)17°). While crystal pack-
ing forces are relevant, coplanar geometry is also ex-
pected to be an energetic minimum based on theoretical
calculations of related sulfinyl imines.9 Rigidity about the
N-S bond of the N-sulfinyl imine moiety is presumably
important for successful chirality transfer using this
directing group and for ligand preorganization in asym-
metric catalysis.
a
Conditions: (a) HC(OMe)3, cat. p-TsOH, 100 °C, 3 h; (b) BuLi,
THF, -78 °C, 1 h, then 13, rt, 1 h.
Secon d -Gen er a tion Liga n d s: Su lfin yl Im id oa m -
id in es (sia m ). In comparison to bisoxazoline ligands, the
sulfinyl imine nitrogen is less Lewis basic and presum-
ably coordinates more weakly to copper. Given that
conformational changes are necessary for tridentate
binding of ligands 8a and 8b through the sulfinyl
nitrogen, we sought to design a catalyst system that
would provide a more strongly chelating N-sulfinyl
binding element. Pfaltz and co-workers have shown that
imidoamidines of the general structure 12 (aza-semicor-
rins) are valuable ligands for asymmetric catalysis.10
Sulfinyl imine ligands with this architecture might
provide a more rigid scaffold while increasing the coor-
dinating ability of the sulfinyl imine moiety (Figure 4).
Syn t h esis of Su lfin yl Im id oa m id in e (sia m )
Liga n d s. Heating 7 in refluxing trimethyl orthoformate
with catalytic p-TsOH cleanly afforded O-methyl-N-tert-
butanesulfinyl imidate 13 in 92% yield. From this
intermediate, the nonsymmetric ligand 14 incorporating
a pyridyl-binding element was prepared in a single step
from 2-methylaminopyridine (Scheme 1). When this
ligand was employed in the Cu(OTf)2-catalyzed Diels-
Alder reaction, promising selectivity was observed (Table
3). We were hopeful that a C2-symmetric variant would
Sulfinyl ketimine 8b is found to adopt a similar,
extended geometry in the solid state (Figure 3). Interest-
ingly, while 8b also adopts a roughly s-cis configuration,
the greater deviation from planarity (OdS-NdC dihe-
dral angle ) 34°) coincides with diminished enantio-
selectivity relative to 8a . Most importantly, neither
ligand 8a nor 8b adopts the proper geometry in the solid
state required for tridentate binding.
(9) (a) Tietze, L. F.; Schuffenhuer, A. Eur. J . Org. Chem. 1998, 1998,
1629-1637. (b) Bharatam, P. V.; Uppal, P.; Kaur, A.; Kaur, D. J . Chem.
Soc., Perkin Trans. 2 2000, 43-50.
(10) (a) Pfaltz, A. Synlett 1999, 835-842 and references therein. (b)
More recently Reiser has prepared analogous aza-bis(oxazolines). Glos,
M.; Reiser, R. Org. Lett. 2000, 2, 2045-2048.
J . Org. Chem, Vol. 68, No. 1, 2003
5

Owens et al.
SCHEME 3a
TABLE 3. Im id oa m id in e Ca ta lyzed Diels-Ald er
Rea ction s
entry ligand
metal
time (h) yield (%) ee (%)a
drb
1
2
3
4
14
16a
14
Cu(OTf)2
Cu(OTf)2
Cu(SbF6)2
Cu(SbF6)2
6
6
6
0.1
45
nr
92
96
62
93:7
56
98
95:5
>99:1
16a
a
b
Determined by chiral HPLC. Determined by 1H NMR.
a
Conditions: (a) RNH2 (10 equiv), THF, rt, 1 h; (b) BuLi, THF,
-78 °C, then 13, rt, 1 h. (c) CF3CH2NH2 (15 equiv), 60 °C, 2 h; (d)
K2CO3, 13, THF, 65 °C, 24 h.
SCHEME 2a
SCHEME 4a
a
Conditions: (a) MeNH2/MeOH, rt, 12 h; (b) KH, THF, 0 °C,
30 min, then 13, 1 h.
improve the selectivity. Ligand 16a was therefore pre-
pared in two steps from 7 (Scheme 2). Addition of
methylamine to 13 afforded the N-sulfinyl amidine 15a
in quantitative yield. Following deprotonation, 15a adds
to a second equivalent of 13 to afford the symmetric bis-
(sulfinyl)imidoamidine (siam) 16a (3 steps, 58% overall
yield). In contrast to results for ligand 14, the catalyst
prepared from Cu(OTf)2 and ligand 16a was completely
unreactive at -78 °C. However, the use of the less
coordinating SbF6 counterion produced an extremely
reactive and selective catalyst (entry 4).11
a
Conditions: (a) NH3/EtOH, rt, 12 h; (b) CH3C(OMe)3, cat.
p-TsOH, 110 °C, 3 h; (c) KH, THF; (d) KH, THF, rt, 15 min, then
MeI, rt, 12 h.
SCHEME 5a
Syn th esis of Mod ified sia m Liga n d s. The impor-
tance of different substituents on the sulfinyl imidoami-
dine (siam) framework was investigated next. Ligands
containing various substituents at the internal nitrogen
were prepared by first adding amines into sulfinyl
imidate 13 to afford the corresponding sulfinyl amidines
15b-d (Scheme 3). While most additions were complete
within minutes at room temperature, the electron-poor
trifluoroethylamine added more slowly. Compound 13
was heated in neat trifluoroethylamine at 60 °C in a
pressure tube to obtain good conversion to 15d in a
reasonable amount of time. Amidines 15b and 15c were
deprotonated with BuLi and then 13 was added to afford
the bis(sulfinyl)imidoamidines 16b and 16c in 58% and
81% yields, respectively. The preparation of the electron-
poor ligand 16d required modified conditions. Heating
amidine 15d with 13 in refluxing THF in the presence
of a slight excess of K2CO3 (1.2 equiv) afforded the desired
ligand 16d in 30% yield after 24 h.
a
Conditions: (a) cat. p-TsOH, THF, HC(OMe)3, rt, 3 h; (b)
MeNH2/MeOH; (c) K2CO3, 22, THF, 60 °C, 12 h.
imidate 13. Deprotonation with KH was followed by the
addition of imidate 19 to afford the nonsymmetric imi-
doamidine 20 in 78% yield. Deprotonation of 20 was
followed by the addition of 5.0 equiv of MeI to afford 17
in 48% yield.12
Finally, to investigate the importance of the substitu-
ent on sulfur, ligand 21 was prepared beginning with
p-toluenesulfinamide 9 (Scheme 5). In contrast to the
preparation of tert-butanesulfinyl imidate 14a , heating
Preparation of ligand 17 incorporating substitution at
the sp2-carbon was next explored (Scheme 4). Amidine
18 was prepared by addition of ammonia in EtOH to
(11) Evans, D. A.; Murry, J . A.; von Matt, P.; Norcross, R. D.; Miller,
S. J . Angew. Chem., Int. Ed. Engl. 1995, 34, 798-800.
(12) Addition of methylamine to 18 followed by deprotonation and
addition to imidate 13 did not provide the desired product 17.
6
J . Org. Chem., Vol. 68, No. 1, 2003

Preparation of Bis(sulfinyl)imidoamidine Ligands
TABLE 4. In flu en ce of Su bstitu tion on sia m Liga n d s
TABLE 5. Su bstr a te Gen er a lity in th e Diels-Ald er
Rea ction Usin g Liga n d 16a a
substrate
(R ))
%
time temp yield ee
entry
n
adduct cat.a (h) (°C) (%) (%)b drc
entry
liganda
R1
R2
i-Bu
Bn
CH2CF3
Me
R3
drb
eec
1
2
3
1
1
1
1
1
2
H (10a )
11a
11b
11c
11d
11a
11e
10
10
10 16
10
1
10 16
0.1 -78
96
76
58
85
95
50
98 99:1
97 98:2
94 95:5
96 97:3
98 99:1
90 98:2
1
2
3
4
5
16b
16c
16d
17
t-Bu
t-Bu
t-Bu
t-Bu
p-tol
H
H
H
Me
H
99:1
99:1
99:1
99:1
94:6
97
96
98
60
32
CH3 (10b)
Ph (10c)
CO2Et (10d )
H (10a )
8
-40
0
4
2
8
-78
-78
0
5d
6
21
Me
H (10a )
a
a
All reactions were run with 10 mol % CuCl2, 20 mol % AgSbF6,
Reactions were run with 10 mol % CuCl2, 20 mol % AgSbF6,
11 mol % ligand, and 3 Å mol sieves. Determined by 1H NMR.
b
b
and 11 mol % 16a . Determined by chiral HPLC or GC (see
c Determined by HPLC.
experimental data, Supporting Inforamtion). c Determined by 1H
d
NMR. Reaction run with 1.0 mol % CuCl2, 2.0 mol % AgSbF6,
and 1.1 mol % 16a .
9 in neat trimethyl orthoformate with catalytic p-TsOH
afforded only a 17% yield of 22. The major product of the
reaction was racemic methyl p-tolylsulfinate.13 More
importantly, the sulfinyl imidate was obtained with only
70% ee. The increased electrophilicity of the sulfur center
of 9 relative to 7 parallels the reactivity trend observed
for sulfinyl imines 8.14 Fortunately, running the reaction
in THF at a 0.2 M concentration suppressed the forma-
tion of the methyl sulfinate byproduct, affording the
desired imidate in 78% yield and with >99% ee. From
imidate 22, the preparation of the sulfinyl amidine 23
and bis(sulfinyl)imidoamidine 21 was straightforward.
Ca ta lysis w ith Mod ified sia m Liga n d s. With the
new ligands in hand, the effect of permutations of the
siam framework was studied (Table 4). Substitution on
the internal nitrogen (R2) is well tolerated and produces
catalysts that maintain high selectivities (entries 1 and
2). The ability to incorporate substitution at this position
has important implications with regard to preparation
of a support-bound ligand.15 Importantly, the ability to
modify the electronic properties of the ligand by substitu-
tion at this position (16d , entry 3) would prove beneficial
for less reactive systems (vide infra). In contrast, ligand
17 was found to provide very poor selectivity for the
Diels-Alder reaction. The use of the p-toluenesulfinyl-
derived ligand 21 afforded 11a with dramatically reduced
enantioselectivity in comparison to the corresponding
tert-butyl derivative 16a .
reactions with cyclopentadiene (entry 5). Furthermore,
the high reactivity of the system allows for the use of
much less reactive dienes such as cyclohexadiene (entry
6). The importance of the bidentate activation of the
dienophile is highlighted by the failure of acrolein to
undergo the Diels-Alder reaction with high selectivity
(eq 2).
Su bstr a te Scop e: Acyclic Su bstr a tes. Although few
asymmetric catalysts are effective for less reactive dienes,
the success of the Diels-Alder reaction with cyclohexa-
diene encouraged us to investigate acyclic substrates
(Table 6). Utilizing the Cu(II)-siam catalyst, isoprene
and 2,3-dimethylbutadiene both react with 10a to afford
the Diels-Alder product in high yield and with excellent
enantioselectivity (entries 1 and 2). It is worth noting that
for these particular substrates, the bisoxazolines are
much less selective (∼60-65% ee).16 Incorporation of
substituents at the terminal position of the diene, how-
ever, resulted in greatly diminished selectivity and
reactivity as seen for the reaction of piperylene (entry
3). Interestingly, substitution at the terminal position of
the diene is tolerated when used in conjunction with
different bidentate dienophiles (vide infra). The selectiv-
ity of the catalyst system is sensitive to the size of
substituents on the 2-position of the diene. For example,
while high reactivity is maintained for 2-phenylbutadi-
ene, selectivity is poor (entry 4). The presence of the bulky
TBDPS ether results in both lower reactivity and lower
selectivity (entry 5).
Su bstr a te Scop e: Dien op h iles. The Cu(SbF6)2-
siam catalyst system was then tested with a series of
substituted dienophiles (Table 5). Selectivities remained
high for imides derived from crotonic acid (10b), cinnamic
acid (10c), and fumaric acid (10d ) (entries 2-4), even at
the higher temperatures required for the less reactive
substrates. Importantly, the efficiency of the catalyst
system permits catalyst loading levels as low as 1% for
(13) Evans, D. A.; Faul, M. M.; Colombo, L.; Bisaha, J . J .; Clardy,
J .; Cherry, D. J . Am. Chem. Soc. 1992, 114, 5977-5985.
(14) For the addition of Grignard reagents into p-toluenesulfinyl
imines, attack at sulfur competes with attack at the carbon-nitrogen
double bond. Moreau, P.; Essiz, M.; Merour, J .; Bouzard, D. Tetrahe-
dron: Asymmetry 1997, 8, 591-594. Attack at sulfur is not observed
for tert-butanesulfinyl imines. See ref 4a.
Another challenging set of Diels-Alder substrates
involves the combination of acyclic dienes and substituted
dienophiles (eq 3).17 As expected for this substrate
(15) For
a
review of support bound catalysts for asymmetric
(16) Evans, D. A.; Barnes, D. M.; J ohnson, J . S.; Lectka, T.; von Matt,
P.; Miller, S. J .; Murry, J . A.; Norcross, R. D.; Shaughnessy, E. A.;
Campos, K. R. J . Am. Chem. Soc. 1999, 121, 7582-7594.
synthesis see: Clapham, B.; Reger, T. S.; J anda, K. D. Tetrahedron
2001, 57, 4637-4662. See also ref 10b.
J . Org. Chem, Vol. 68, No. 1, 2003
7

Owens et al.
TABLE 6. Cu (SbF 6)2-16a -Ca ta lyzed Diels-Ald er Rea ction s of 10a w ith Acyclic Dien es
a
b
Refers to ratio of 1,2-endo; other regioisomers. Determined by NMR. ee of major isomer.
F IGURE 5. Crystal structure for siam ligand 16a .
combination, the reaction was extremely sluggish with
ligand 16a and only the activated diene derived from
ethyl fumarate (10d ) was found to be sufficiently reactive,
producing 26 with 53% ee and in less than 65% yield.
For this particular reaction, the increased reactivity of
the siam ligand derived from trifluoroethylamine (16d )
was necessary to obtain good selectivity and complete
conversion to product.
Su bstr a te Scop e: Com p lex Su bstr a te. Perhaps the
most challenging test for asymmetric catalysts lies in
their application to complex molecule synthesis.18 Re-
cently, Murai, Ishihara, and co-workers utilized the
Cu(SbF6)2-siam catalyst system for Diels-Alder reac-
tions of CBZ-protected-R-methylene lactam dienophiles
in elegant studies toward the synthesis of the spirocyclic
core of gymnodimine (eq 4).19 In this transformation only
a single Diels-Alder product was observed and was
isolated in 58% yield. Although bidentate dienophiles are
necessary (eq 2), this example demonstrates that asym-
metric catalysis is not restricted to N-acyl-oxazolidinone-
based dienophiles. Furthermore, for this transformation,
substitution at the terminal position of the diene does
not negatively impact selectivity (Table 6, entry 3).
Mech a n istic Con sid er a tion s. While Cu(II) is known
to coordinate well to nitrogen, there remained the pos-
sibility of alternative binding modes with the sulfinyl
group (via sulfur or oxygen). Additionally, a crystal
structure of 16a was obtained that revealed a nearly
linear arrangement along the axis of the imidoamidine
with the sulfinyl groups adopting an s-cis conformation
(Figure 5). In comparison to the structures of sulfinyl
imines 8a and 8b, the CdN-SdO dihedral angle is even
closer to planarity.20 Once again, for chelation via the
imido-nitrogens, a conformational change would be nec-
essary.
(17) (a) Narasaka, K.; Iwasawa, N.; Inoue, M.; Yamada, T.; Na-
kashima, M.; Sugimori, J . J . Am. Chem. Soc. 1989, 111, 5340-5345.
(b) Ghosh, A. K.; Cho, H.; Cappiello, J . Tetrahedron: Asymmetry 1998,
39, 3687-3691.
(18) Nicolau, K. C.; Snyder, S. A.; Montagnon, T.; Vassilikogianna-
kis, G. Angew. Chem., Int. Ed. 2002, 41, 1668-1698.
(19) (a) Tsujimoto, T.; Ishihara, J .; Horie, M.; Murai, A. Synlett 2002,
399-402. (b) Ishihara, J .; Horie, M.; Shimada, Y.; Tojo, S.; Murai, A.
Synlett 2002, 403-406.
8
J . Org. Chem., Vol. 68, No. 1, 2003

Preparation of Bis(sulfinyl)imidoamidine Ligands
F IGURE 6. X-ray structure of Cu2(16a )4 M2L4-helicate 27 (2
views). The Cu2Cl62- counterion was omitted for clarity. Color
legend: sulfur, yellow; oxygen, red; chlorine, green.
F IGURE 8. Crystal structure of Zn2I4(16a )2 complex 28. Color
legend: sulfur, yellow; oxygen, red; iodine, purple.
suggesting that the primary metal species in solution
remains oxygen bound even in the presence of the
coordinating dienophile.
The propensity for ligand 16a to form dimeric com-
plexes via the sulfinyl oxygen coordination is not limited
to Cu(II) complexes. When ligand 16a is stirred with ZnI2,
complex 28 is obtained (Figure 8). The geometry about
Zn is tetrahedral with the iodide counterions occupying
the other coordination sites. A similar helical twist is also
observed when viewed along the axis of the two zinc
atoms.
F IGURE 7. IR overlay of 16a in solution (A), Cu(SbF6)2‚(16a )
in the solid state (B), and Cu(SbF6)2‚(16a ) in solution (C).
While we do not currently have a mechanistic ratio-
nalization of the observed stereochemical induction for
siam-catalyzed reactions, several observations are rel-
evant. First of all, in contrast to the bisoxazoline catalyst
systems, the presence of a 2-fold excess of ligand relative
to copper does not inhibit the reaction rate.23 Second,
while not dramatic, the reaction of cyclopentadiene and
imide 10a (eq 5) exhibits nonlinearity24 with regard to
ligand enantiopurity (Figure 9), proving that other spe-
cies besides a simple monomeric species are present
under the reaction conditions. Furthermore, analysis of
the catalyst mixture by mass spectrometry after 1 h of
stirring shows the presence of several monomeric and
dimeric species, including CuL2, Cu2L2, CuL3, and Cu2L3.25
Careful catalyst preparation is critical to the success
of the Cu(II)-siam catalyst system. Typically, 11 mol %
of 16a is stirred together with 10 mol % of CuCl2 and 20
mol % of AgSbF6 in the presence of 3 Å molecular sieves
at room temperature for 1 h prior to addition of the imide
and diene.26 Shorter catalyst aging times (10 min) are
acceptable, but if the catalyst is aged longer (8h) the
reaction rate drops dramatically and nearly racemic
products are obtained! Furthermore, when imide 10a was
present during the catalyst aging period, no catalytic
activity was observed for the reaction of 10a and cyclo-
pentadiene.
X-r a y Str u ctu r e of a Cu (II)-sia m Com p lex. After
much effort, a crystalline CuCl2-siam complex (27) was
obtained that was suitable for X-ray crystallographic
studies. Somewhat surprisingly, the complex crystallized
as a dimeric species in which each of four siam ligands
coordinate two copper atoms via the sulfinyl oxygen.
Interestingly, complex 27 (Figure 6) exists as an M2L4
helicate, with each ligand maintaining a linear geometry
and making a quarter turn between the two Cu atoms.21
This represents only the second reported M2L4 quadruple-
stranded helicate and the first example in which the
chiral information is introduced by the ligand. An unco-
ordinated chloride counterion exists within the cage of
the helicate. A second chloride ion and a Cu2Cl6 counter-
ion occupy the axial coordination sites of the Cu(II)
complex. Notably, a crystal structure of a Cu(SbF)2-siam
complex was obtained and showed the same binding
mode as 27, however the resolution was poor.
The solid-state structure proved invaluable for deter-
mining the predominant binding mode of ligand 16a in
solution. In both the crystalline Cu(SbF6)2-siam complex
and a freshly prepared solution of the Cu(SbF6)2-siam
catalyst in CH2Cl2, the sulfinyl SdO stretch shifts to
longer wavelengths (969 cm-1 vs 1077 cm-1 in the free
ligand, Figure 7). A shift to longer wavelengths also
occurs with oxygen binding of sulfoxide ligands to Cu(II)
species.22 The addition of imide 10a after preparation of
the catalyst does not alter the sulfinyl SdO stretch,
Con str a in ed Liga n d . While IR and X-ray studies
demonstrate that the predominant ligand species in the
(23) Evans, D. A.; Miller, S. J .; Lectka, T. J . Am. Chem. Soc. 1993,
115, 6460-661.
(20) The ligand crystallizes as a pair of molecules in the asymmetric
unit. The dihedral angles are 9.0° and 9.7° for one molecule and 4.4°
and 15.6° in the other. See Supporting Information for crystallographic
data.
(21) (a) Albrecht, M. Chem. Rev. 2001, 101, 3457-3497. (b) One
other M2L4-helicate has been characterized. McMorran, D. A.; Steel,
P. J . Angew. Chem., Int. Ed. 1998, 37, 3295-3297.
(22) Huang, Z.; Liao, D.; Zhang, R.; Zhang, X.; Huang, T.; Wang, H.
Polyhedron 1996, 15, 981-984.
(24) Kagan, H. B. Synlett 2001, 888-899.
(25) Electrospray ionization in MeCN showed diagnostic isotopic
distributions consistent with the presence of the above species (see
Supporting Information).
(26) While not strictly necessary, better and more reproducible
results were found when molecular sieves were added to the reaction
mixture. For example, compound 25b is produced in only 69% ee in
the absence of molecular sieves.
J . Org. Chem, Vol. 68, No. 1, 2003
9

Owens et al.
SCHEME 6a
a
Conditions: (a) THF, MgSO4; (b) NaN3, DMF, NaI; (c) Ph3P,
THF, H2O; (d) K2CO3, THF, 13, reflux.
Con clu sion
The sulfinylimidoamidine (siam) ligands coordinate to
Cu(SbF6)2 to afford active and selective catalysts for the
Diels-Alder reaction employing a variety of substrates.
These ligands are readily synthesized in only three steps
and present a highly modular class of ligands. Studies
of the crystallographic and solution structures of these
compounds point to a unique coordination of the sulfinyl
oxygen to both Cu and Zn species. Importantly, ligands
16 impart high selectivity for the Diels-Alder reaction
of a variety of challenging substrates including the
relatively unreactive acyclic dienes.
F IGURE 9. Graph of product enantioselectivity vs ligand 16a
enantioselectivity.
reaction mixture are oxygen-bound Cu complexes, there
remains the possibility that nitrogen-bound complexes
are responsible for catalysis. If this is the case, con-
strained ligands such as 29 should be effective because
the ligand geometry favors coordination through nitro-
gen. Furthermore, the ligand constraints would prohibit
the dimeric bonding pattern found in Cu(II)-siam com-
plexes. Ortho ester 30 was reacted with 7 and MgSO4 to
afford an inseparable mixture of imidate 31 and ester
32 (Scheme 6). Azide displacement of the product mixture
with NaN3 afforded the desired azide 33 in 55% overall
yield from 7. Upon heating 33 with triphenylphosphine
and Et3N, cyclic amidine 34 was obtained in 78% yield.
Finally, 34 was heated with 13 and K2CO3 in THF to
afford ligand 29 in four overall steps. The constrained
ligand 29 efficiently catalyzed the reaction of cyclopen-
tadiene and imide 10a but afforded the product with low
selectivity (96:04 dr, 45% ee). Interestingly, use of ligand
17, which incorporates the same substitution pattern
minus any cyclic constraints, provided similarly poor
levels of induction and reactivity (Table 4, entry 4). While
a mechanistic model is not currently available to explain
this, it appears that any substitution at the amidino-sp2-
carbon is deleterious to the selectivity of the reaction.
Ack n ow led gm en t. The authors gratefully acknowl-
edge the assistance of Laurie Schenkel in the final
preparation of the manuscript. The authors also ac-
knowledge support from the NSF (Grant No. CHE-
0139468). Dr. Fred J . Hollander and Allen G. Oliver are
gratefully acknowledged for solving all of the X-ray
crystal structures reported in this paper. The Center
for New Directions in Organic Synthesis is supported
by Bristol-Myers Squibb as a Sponsoring Member and
Novartis as a Supporting Member.
Su p p or tin g In for m a tion Ava ila ble: Experimental Pro-
cedures, spectral data for all new compounds, and mass
spectrometric and crystallographic data. This material is
available free of charge via the Internet at http://pubs.acs.org.
J O020524C
10 J . Org. Chem., Vol. 68, No. 1, 2003