"Orthogonal" Lewis Acids: Catalyzed Ring Opening and Rearrangement of Acylaziridines

Full text of DOI:10.1021/jo980558d

4568
J . Org. Chem. 1998, 63, 4568-4569
Ta ble 1. Rin g Op en in g of Acyla zir id in es 1a -d w ith
“Or th ogon a l” Lew is Acid s: Ca ta lyzed Rin g
Op en in g a n d Rea r r a n gem en t of
Acyla zir id in es
TMSN3
entrya
substrate
metal complexb
product
yieldc (%)
a
b
c
d
e
f
1a
1a
1a
1b
1b
1b
1c
1d
Yb(biphenol)OTf
Ti(O-i-Pr)4
Zr(Cp)2(SbF6)2
Yb(biphenol)OTf
Ti(O-i-Pr)4
Zr(Cp)2(SbF6)2
Yb(biphenol)OTf
Yb(biphenol)OTf
2a
2b
2b
2b
2b
2b
2c
2d
84
64
57
72
67
58
75
80
Dana Ferraris, William J . Drury III,
Christopher Cox, and Thomas Lectka*
Department of Chemistry, J ohns Hopkins University,
Baltimore, Maryland 21218
Received March 26, 1998
g
h
In a recent paper, we demonstrated that select late
transition metals can catalyze the cis-trans conformational
isomerization of amides through coordination of the metal
to Na.1 More oxophilic metals, on the other hand, prefer to
coordinate to the carbonyl O. We were intrigued by the use
of acylaziridines, which can rearrange to oxazolines2 or serve
as ambident electrophiles,3 as probes for Lewis acid-
catalyzed reaction pathway selectivity. It was our hypoth-
esis that coordination of a Lewis acid to the amide nitrogen
of acylaziridines (Na) might be expected to catalyze a
rearrangement to the oxazoline, whereas coordination to the
carbonyl O may be better at activating the substrate toward
external nucleophilic attack. We demonstrate herein that
catalytic quantities of relatively oxophilic metals under
certain conditions activate acylaziridines 1 predominantly
toward external nucleophilic attack, whereas more azaphilic
Lewis acids, which we term “orthogonal,” catalyze the
oxazoline rearrangement (Scheme 1).4
a
All reactions with Yb were performed in CH2Cl2; others were
performed in THF. 10 mol % metal. c Reactions took 48-72 h to
b
full coversion as measured by 1H NMR.
Sch em e 1. Rea r r a n gem en t a n d Rin g Op en in g of
Acyla zir id in es
Several procedures for the Lewis acid-catalyzed asym-
metric synthesis of aziridines by use of transition metal-
based catalysts have emerged in the past few years.5
However, ring-opening and expansion reactions of acylaziri-
dines, employing catalytic quantities of metal-based Lewis
acids, have not been widely reported.6 These reactions are
expected to afford useful synthetic intermediates such as 4
and 5 (Scheme 1). We focused on the reaction of acylated
cyclohexenimine derivatives 1, which are readily available
from either cyclohexene7 or cyclohexene oxide,8 followed by
acylation. Compounds 1a -d were smoothly converted to
ring-opened products 2a -d by TMSN3 in the presence of
10 mol % Yb(2,2′-biphenol)OTf (Table 1, entries a, d, g, h).
The complexes Zr(Cp)2(SbF6)2 and Ti(O-i-Pr)4 were also
found to catalyze nucleophilic attack of TMSN3 (Table 1,
entries b, c, e, f). Of these metals, Yb(III) complexes gave
the fastest rates and most reproducible results. Conversion
to the azide 2 is possible only in the presence of a Lewis
acid, as no nucleophilic attack occurs on the substrates after
1 week without catalyst. To examine the effect of remote
electron-withdrawing and electron-donating groups, we
conducted a series of competition experiments and found
that electron-withdrawing substituents accelerate the reac-
tion, as indicated by a linear correlation of log[k/k0] to
Hammett σ values.9 Electron-withdrawing substituents are
expected to stabilize the leaving group during nucleophilic
attack.
Unlike their acyclic amide counterparts, acylaziridines are
highly pyramidalized at nitrogen, which makes the acyl-
aziridine nitrogen more basic. Experimental10 as well as
theoretical11 evidence indicates that acylaziridines may
undergo N-protonation. This behavior stands in sharp
contrast to that of simple amides, which undergo preponder-
ant O-protonation under all circumstances.12 We obtained
a crystal structure of 1d that clearly indicates pyramidal-
ization of the aziridine nitrogen (Figure 1);13 aziridine 1d
exhibits an out-of-plane angle of 68°, as defined by Ohwa-
da,14 and a C-N bond length of 1.388 Å. These data differ
significantly from those of normal amides, which are es-
sentially planar (out-of-plane angles < 10°) and have C-N
bond lengths less than 1.34 Å.14
(1) Cox, C.; Ferraris, D.; Murthy, N. N.; Lectka, T. J . Am. Chem. Soc.
1996, 118, 5332.
(2) Nishiguchi, T.; Tochio, H.; Nabeya, A.; Iwakura, Y. J . Am. Chem. Soc.
1969, 91, 5835.
(3) (a) Lygo, B. Tetrahedron Lett. 1994, 35, 5073. (b) Legters, J .; Willem,
J . G. H.; Thijs, L.; Zwanenburg, B. Recl. Trav. Chim. Pays-Bas 1992, 111,
59.
(4) We define azaphilicity in regard to Na using Pearson’s HSAB theory
as a useful guideline. See: Huheey, J . E. Inorganic Chemistry; Harper and
Row: New York, 1983; pp 312-315.
(5) (a) Li, Z.; Quan, R. W.; J acobsen, E. N. J . Am. Chem. Soc. 1995, 117,
5889. (b) Evans, D. A.; Faul, M. M.; Bilodeau, M. T. J . Am. Chem. Soc.
1994, 116, 2743. (c) Li, Z.; Conser, K. R.; J acobsen, E. N. J . Am. Chem.
Soc. 1993, 115, 5326. (d) Evans, D. A.; Faul, M. M.; Bilodeau, M. T.;
Anderson, B. A.; Barnes, D. M. J . Am. Chem. Soc. 1993, 115, 5328. (e) Evans,
D. A.; Margaret, M. F.; Bilodeau, M. T. J . Org. Chem. 1991, 56, 6744.
(6) For a general review of aziridine chemistry, see: Pearson, W. H.; Lian,
B. W.; Bergmeier, S. C. In Comprehensive Heterocyclic Chemistry II;
Katritzky, A. R., Rees, C. W., Scriven, E. F. V., Eds.; Pergamon: New York,
1996; Vol 1a.
We decided to investigate metals classically thought of as
more azaphilic, specifically the salts Zn(OTf)2, Cu(OTf)2, and
Sn(OTf)2, to enhance the possibility of N-coordination. These
metal salts did not catalyze the addition of nucleophiles to
(9) The Hammett correlation gave a F value of 2.4; see the Supporting
Information for the plot. For a discussion of Hammett correlations and
tabular σ values, see: Advanced Organic Chemistry; March, J ., Ed.; Wiley:
New York; 1992; pp 278-286.
(10) Olah, G. A.; Szilagyi, P. J . J . Am. Chem. Soc. 1969, 91, 2949.
(11) Cho, S. J .; Cui, C.; Lee, J . Y.; Park, J . K.; Suh, S. B.; Park, J .; Kim,
B. H.; Kim, K. S. J . Org. Chem. 1997, 62, 4068.
(12) Olah, G. A.; Calin, M. J . Am. Chem. Soc. 1968, 90, 401.
(13) Crystals of 1d were obtained by slow evaporation of CH2Cl2. Crystal
data for 1d : triclinic, P1; a ) 6.6677(7) Å, b ) 8.2722(9) Å, c ) 12.1198(13)
Å, V ) 628.29(12) Å3; Z ) 2; dcalcd ) 1.423 Mg/m3; F(000) ) 280; µ(Mo KR)
) 0.120 mm-1; λ(Mo KR) ) 0.710 73 Å; 3567 reflections measured, 2133
observed (I > 2σ(I)); 172 variables; R ) 0.0705, Rw ) 0.1811, GOF ) 1.034.
(14) Ohwada, T.; Achiwa, T.; Okamoto, I.; Shudo, K.; Yamaguchi, K.
Tetrahedron Lett. 1998, 865.
(7) (a) Hassner, A.; Matthews, G. J .; Fowler, F. W. J . Am. Chem. Soc.
1969, 91, 5046. (b) Fowler, F. W.; Hassner, A.; Levy, L. A. J . Am. Chem.
Soc. 1967, 89, 2077.
(8) Zhang, Z.; Scheffold, R. Helv. Chim. Acta 1993, 76, 2602.
S0022-3263(98)00558-1 CCC: $15.00 © 1998 American Chemical Society
Published on Web 06/17/1998

Communications
J . Org. Chem., Vol. 63, No. 14, 1998 4569
of 6 to 7 (eq 1) is 57 kcal/mol. When a Li+ ion is coordinated
to Na this energy is marginally lowered to a still prohibitively
high 53 kcal/mol,18 which suggests that the isomerization
is not concerted in organic media.19
Alternatively, the reaction can proceed by a heterolytic
pathway in a stepwise fashion. This alternative is in accord
with the rate accelerations we observed in better cation-
solvating organic solvents. Substituent effects for the ox-
azoline rearrangement are also understandable in terms of
enhanced nucleophilicity of the carbonyl oxygen in interme-
diate 8, where metal-Na-coordination is expected to leave
the carbonyl group free to attack the cis face of the incipient
cationic center (eq 2). The fact that oxophilic metals do not
catalyze the rearrangement implies that metal coordination
to Na is an essential feature of the rearrangement.
F igu r e 1. Crystal structure of 1d (50% ellipsoids). Selected bond
distances (Å): O(1)-C(1) 1.225(4); C(1)-N(1) 1.388(4). Selected
torsion angles (deg): C(1)-N(1)-C(2)-C(7) 112.9; O(1)-C(1)-
N(1)-C(2) 39.1.
Ta ble 2. Rea r r a n gem en ts of Acyla zir id in es 1a -d
entrya substrate
metalb
timec (h) product yield (%)
a
b
c
d
e
f
g
h
i
j
k
l
m
n
1a
1a
1a
1b
1b
1b
1b
1b
1c
1c
1c
1d
1d
1d
Cu(OTf)2
Sn(OTf)2
Zn(OTf)2
Cu(OTf)2
Sn(OTf)2
Cu(OTf)2
Sn(OTf)2
Zn(OTf)2
Cu(OTf)2
Sn(OTf)2
Zn(OTf)2
Cu(OTf)2
Sn(OTf)2
Zn(OTf)2
6
10
7
48
48
7
12
12
7
15
12
16
24
24
3a
3a
3a
3b
3b
3b
3b
3b
3c
3c
3c
3d
3d
3d
89
83
63
80
30
84
87
74
80
79
60
76
67
62
a
All reactions were run in THF/DME (20/1) except those in
entries d and e, which were in CH2Cl2. 10 mol % metal. c Time
b
to full conversion by 1H NMR.
acylaziridines, but instead promoted the rearrangement of
1a -d to 2-aryloxazolines 3a -d (Table 2).15 In fact, the
rearrangement proceeded smoothly in the presence of a
variety of nucleophiles (enol silanes, TMSCN, TMSN3, and
water) and also in their absence. Although the oxazoline
rearrangement proceeds at room temperature in CH2Cl2, it
does so only very slowly (months). However, 1b was
converted to 3b by 10 mol % Cu(OTf)2 in 80% yield after 2
days, with the mass balance accounted for by unconverted
starting material. Under the same conditions, Sn(OTf)2
showed only 30% conversion (Table 2, entries d, e). The
reaction rate can be greatly accelerated by the use of THF/
DME, which promoted full conversion of 1b to 3b in 7 h by
Cu(II) in 84% yield (Table 2, entry f). Whereas Sn(OTf)2 is
an inefficient catalyst in CH2Cl2, in THF/DME it isomerizes
1b to 3b in under 12 h in 87% yield (Table 2, entry g).
Likewise, this isomerization proceeds in the presence of 10
mol % Zn(OTf)2 with minimal loss of yield (entries c, h, k,
n).
To determine the regio- and stereochemistry of rearrange-
ment, optically active substrate (R)-9 was synthesized by
literature methods.20 The reaction of (R)-9 with 10 mol %
Cu(OTf)2 produced the enantiomerically pure natural prod-
uct (R)-oxytriphine 10 (eq 3), whose stereo- and regiochem-
istry was determined by comparison to authentic material.21
The exclusive formation of enantiomerically pure 10 indi-
cates that a carbocation, solvated from the backside or in
the form of a tight ion pair to preserve stereochemistry, could
be an intermediate in this pathway.
Ack n ow led gm en t. The authors thank Dr. Victor G.
Young, J r., director of the X-ray Crystallographic Labora-
tory at the University of Minnesota, for obtaining the
crystal structure of 1d. T.L. thanks the NIH (R29 GM54348)
for support of this work. D.F. thanks J HU for a Marks
Fellowship, and C.C. thanks the Organic Division of the
American Chemical Society for a Graduate Fellowship
(sponsored by Organic Reactions, Inc.).
Competition experiments on the rearrangement led us to
conclude that electron-donating substituents increase the
rate of reaction.16 This trend is opposite to that of the Lewis
acid-catalyzed additions (1 to 2) analyzed above. A literature
report proposed the uncatalyzed oxazoline rearrangement
to be a concerted process.17 We have calculated transition
states at 6-31G* for the Lewis acid-catalyzed and the
uncatalyzed versions, both of which can be viewed as orbital
symmetry disallowed “frontside” nucleophilic attacks. The
calculated energy of activation (∆Gq) for the transformation
Su p p or tin g In for m a tion Ava ila ble: Experimental proce-
dures, spectroscopic data for all new compounds, protocols for
mechanistic experiments, and tables of crystal data (15 pages).
J O980558D
(16) The Hammett plot gave a F value of -2.0. See the Supporting
Information for the plot.
(17) Heine, H. W.; Proctor, Z. J . Org. Chem. 1958, 23, 1554.
(18) Interestingly, the low energy structure at 6-31G* shows simulta-
neous coordination of Li+ to Na and O.
(19) Lithium cations are known to catalyze the N-acylaziridine to
oxazoline rearrangment: Tanner, D.; Tedenborg, L.; Almario, A.; Pettersson,
I.; Cso¨regh, I.; Kelly, N. M.; Andersson, P. G.; Ho¨gberg, T. Tetrahedron 1997,
53, 4857.
(20) Nabeya, A.; Shigemoto, T.; Iwakura, Y. J . Org. Chem. 1975, 40, 3536.
(21) (a) Akhmedzhanova, V. I.; Batsuren, D.; Shakirov, R. S. Chem. Nat.
Cmpds. 1993, 29, 778. (b) Myers, A. I.; Hanagan, M. A.; Trefonas, L. M.;
Baker, R. J . Tetrahedron 1983, 39, 1991.
(15) These compounds are assigned cis stereochemistry on the basis of
spectral similarity between 3b and material generated by known literature
methods. All new oxazolines have been assigned a similar stereochemistry
on the basis of analogy: Winstein, S.; Boschan, R. J . Am. Chem. Soc. 1950,
72, 2, 4669.