Double stereodifferentiation in the catalytic asymmetric aziridination of imines prepared from α-chiral amines

Article abstract of DOI:10.1002/chem.201102520

The catalytic asymmetric aziridination of imines and diazo compounds (AZ reaction) mediated by boroxinate catalysts derived from the VANOL and VAPOL ligands was investigated with chiral imines derived from five different chiral, disubstituted, methyl amines. The strongest matched and mismatched reactions with the two enantiomers of the catalyst were noted with disubstituted methyl amines that had one aromatic and one aliphatic substituent. The synthetic scope for the AZ reaction was examined in detail for α-methylbenzyl amine for cis-aziridines from α-diazo esters and for trans-aziridines from α-diazo acetamides. Optically pure aziridines could be routinely obtained in good yields and with high diastereoselectivity and the minor diastereomer (if any) could be easily separated. The matched case for cis-aziridines involved the (R)-amine with the (S)-ligand, but curiously, for trans-aziridines the matched case involved the (R)-amine with the (R)-ligand for imines derived from benzaldehyde and n-butanal, and the (R)-amine with the (S)-ligand for imines derived from the bulkier aliphatic aldehydes pivaldehyde and cyclohexane carboxaldehyde.

Full text of DOI:10.1002/chem.201102520

DOI: 10.1002/chem.201102520
Double Stereodifferentiation in the Catalytic Asymmetric Aziridination of
Imines Prepared from a-Chiral Amines
Li Huang, Yu Zhang, Richard J. Staples, Rui H. Huang, and William D. Wulff*[a]
Abstract: The catalytic asymmetric
aziridination of imines and diazo com-
pounds (AZ reaction) mediated by
boroxinate catalysts derived from the
VANOL and VAPOL ligands was in-
vestigated with chiral imines derived
from five different chiral, disubstituted,
methyl amines. The strongest matched
and mismatched reactions with the two
enantiomers of the catalyst were noted
with disubstituted methyl amines that
had one aromatic and one aliphatic
substituent. The synthetic scope for the
AZ reaction was examined in detail for
a-methylbenzyl amine for cis-aziridines
from a-diazo esters and for trans-aziri-
dines from a-diazo acetamides. Opti-
cally pure aziridines could be routinely
obtained in good yields and with high
diastereoselectivity and the minor dia-
stereomer (if any) could be easily sepa-
rated. The matched case for cis-aziri-
dines involved the (R)-amine with the
(S)-ligand, but curiously, for trans-aziri-
dines the matched case involved the
(R)-amine with the (R)-ligand for
imines derived from benzaldehyde and
n-butanal, and the (R)-amine with the
(S)-ligand for imines derived from the
bulkier aliphatic aldehydes pivaldehyde
and cyclohexane carboxaldehyde.
Keywords: asymmetric synthesis ·
aziridination · aziridines · boroxi-
nate catalyst · homogeneous cataly-
sis
Introduction
Catalysts prepared from either the VANOL or VAPOL
ligand and B((OPh)3 have provided a general method for
ACHTUNGTRENNUNG
the asymmetric catalytic synthesis of aziridines that involves
the reaction of imines and diazo compounds.[1,2] Neither the
VANOL or VAPOL ligand will react with BACTHNURTGNENG(U OPh)3 at room
temperature, but upon addition of the imine substrate there
is an immediate assembly of the catalyst that has been
shown to be a chiral boroxinate of the type 9 (Scheme 1).[3]
The catalyst is an ion pair consisting of a boroxinate anion
and an iminium cation that results from the protonation of
the imine. The efficiency and scope of the catalytic asym-
metric aziridination reaction (AZ) has evolved over the
years mainly with the identification of the optimal aryl sub-
stituents for the diarylmethyl group on the nitrogen in imine
1.[4] Over the entire range of imines that have been exam-
ined, the optical purities of the cis-aziridines 3 range from
77 to 99% ee; which includes imines derived from electron-
rich and electron-poor aromatic aldehydes as well as 18, 28,
and 38 aliphatic aldehydes.[1c,4] While many reactions give
aziridines with 98–99% ee, those substrates that give less
than ideal asymmetric inductions usually require an optical
upgrade of the product by any number of procedures. In an
Scheme 1.
effort to palliate the less than perfect induction for these
substrates, we considered investigating the aziridination of
imines of the type 4 prepared from chiral amines. Less than
complete asymmetric induction imparted by the catalyst in
this case would result in the formation of the diastereomers
5 and 6, which should normally be easier to separate than
the two enantiomers of 3. As long as the double-stereo dif-
ferentiation in the reactions of imines of the type 4 is signifi-
cant, facile access to useful quantities of either antipode of
cis-3-substituted aziridine-2-carboxylate esters should be
possible.
[a] L. Huang, Dr. Y. Zhang, Dr. R. J. Staples, Dr. R. H. Huang,
Prof. W. D. Wulff
Department of Chemistry, Michigan State University
East Lansing, MI 48824 (USA)
E-mail: wulff@chemistry.msu.edu
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201102520.
5302
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5302 – 5313

FULL PAPER
efficient aziridination, but we do not know if one aryl group
and one alkyl group will give reasonable rates and selectivi-
ties, since unsymmetrically disubstituted methyl imines of
the type 4 (Scheme 1) have not been previously investigat-
ed.
The aziridinations of chiral imines of the type 4 with
diazo compounds have not been previously investigated
with chiral catalysts; however, three reports have appeared
that describe these reactions with nonchiral catalysts.[5] Two
of the more thorough studies are summarized in Scheme 3.
Scheme 2.
Background: The most successful nitrogen substituents on
the imine 1 for the catalytic asymmetric aziridination reac-
tion are those that can be generated from the amines 10–13
(Scheme 2).[1c,4] A snapshot of the comparison of these
amines in the AZ reaction is provided in Scheme 2 for
imines 15 and 17, prepared from benzaldehyde and cyclo-
hexane carboxaldehyde, respectively. These data are
a proper reflection of the observed trends. The MEDAM
and BUDAM imines give very high asymmetric inductions
for imines derived from aryl aldehydes, but they give lower
inductions with imines from aliphatic aldehydes.[4b,c] They
also give higher asymmetric inductions than benzhydryl
imines derived from all classes of aldehydes. However, since
the MEDAM and BUDAM amines 12 and 13 are not com-
mercially available, in many cases the most practical ap-
proach to aziridines of high optical purity with this method
is to perform the aziridination with benzhydryl imines fol-
lowed by an upgrade of the optical purity of the aziridine
product by crystallization, extraction, or chemical conver-
sion.
The set of chiral amines that we chose to examine in the
double stereodifferentiation study outlined in Scheme 1 are
the four amines 19–22 shown in Scheme 2. These were
chosen to examine the effect of competition between aryl
and alkyl groups, between aryl groups of different electron
density, and between two alkyl groups of different size.
While two aryl groups on the methyl amine is ideal for the
aziridination reaction (10–13, Scheme 2), two alkyl groups
proved to be detrimental. The imine prepared from the bis-
cyclohexylmethyl amine 14 and benzaldehyde reacts 20
times slower than the corresponding benzhydryl imine and
the asymmetric induction drops from 89 to 74% ee (in
CH2Cl2).[4b] Thus we know that two aryl groups provide for
Scheme 3.
Ha and Lee and co-workers generated imines of the type 4
in-situ during the Lewis acid mediated aziridination of a-
amino nitriles 23 derived from (R)-a-methylbenzyl amine
(R)-19 with ethyl diazoacetate.[5b] The optimal conditions in-
volved reaction with one equivalent of tin tetrachloride and
gave a mixture of diastereomeric aziridines 24 and 25,
whereby the former predominated in ratios ranging from
58:42 to 71:29. Both aryl and alkyl substituents could be in-
troduced into the aziridine in the 3-position and the yields
of the major diastereomer 24 ranged from 18–52%. Lee and
co-workers found that the imines (S)-26 could be directly
induce to react with ethyl diazoacetate by a combination of
20 mol% cobalt chloride and 40 mol% silver tetrafluorobo-
rate to give a mixture of the diastereomers 27 and 28 in
ratios favoring 28 ranging from 57:43 to 73:27 with yields of
28 ranging from 0–65%.[5c] This study only examined imines
generated from aryl aldehydes. To summarize previous stud-
ies, imines generated from a-methylbenzyl amine give
a slight diastereomeric preference for a particular diastereo-
mer, the relative stereochemistry of which is independent of
the nature of the nonchiral Lewis acid. The degree of the
stereoselection is low and the yields of aziridines produced
are not generally useful.
Results and Discussion
Double stereodifferentiation with amines 19–22: The initial
screen of chiral imines was carried out with the imine 26a
Chem. Eur. J. 2012, 18, 5302 – 5313
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5303

W. D. Wulff et al.
Table 1. Matched and mismatched aziridinations of the phenethyl imine (R)-26a.[a]
protons for a trans-aziridine. A
cis/trans selectivity of >50:1
can be assigned in all cases. The
assignment of the relative ste-
reochemistry in aziridine 24a
was made after comparison of
its rotation with that reported
in the literature for this com-
pound[5b] and by its conversion
to the (R)-enantiomer of phe-
nylalanine ethyl ester 49
(Scheme 6, see below). The
standard reaction time for these
reactions was 24 h, but no
effort was made to determine
the minimum reaction times,
which are undoubtedly signifi-
cantly less than 24 h since reac-
tions stopped after 1 h are 82–
87% complete (Table 1, en-
tries 6 and 7).
Ligand
Solvent
t
Conv.
[%]
24a:25a[b]
Yield
Yield
Yield
[h]
24a [%][c]
25a [%][c]
29/30 [%][d]
1
2
3
4
5
6
7
8
9
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
CCl4
CCl4
CCl4
CCl4
24
24
24
24
24
1
100
100
100
100
100
87
>98:2
33:67
>98:2
31:69
86:14
97:3
41:59
96:4
33:67
>97:3
31:69
75:25
79
20[d]
86
<2
40
<2
26
2/2
27/17
4/3
33/19
nd
4/6
3/8
1/7
7/11
3/3
18[d]
nd
52
BN
CCl4
nd
(S)-VAPOL
(R)-VAPOL
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
toluene
toluene
toluene
toluene
toluene
toluene
toluene
2[d]
1
82
18
74
17
69
22
31
26[d]
3[d]
24
24
24
24
24
100
100
100
100
66
33[e]
<2[d]
48[d]
10[d]
10
11
12
9/15
7/8
BN
[a] Unless otherwise specified all reactions were run at 0.5m in imine with 1.1–1.2 equivalents of 2 and
10 mol% catalyst. The cis:trans selectivity was >50:1 in all cases. The catalyst for the reactions in CCl4 was
prepared from 1 equivalent of the ligand and 3 equivalent of B
mental Section given in the Supporting Information (entries 1–4), and the catalyst for the reactions in toluene
was prepared from 1 equivalent of the ligand, 4 equivalent of B(OPh)3 and 1 equivalent H2O according to
ACHTUNGRTENN(UNG OPh)3 according to Method A in the Experi-
The catalytic aziridination of
the cyclohexylethyl imine 31a
pits the effects of a cyclohexyl
group versus a methyl group in
vying for the diastereoselection
AHCTUNGTRENNUNG
Method B in the Experimental Section given in the Supporting Information (entries 6–12). Imine (R)-26a was
prepared from an imine of>99% ee and is an oil and was purified by distillation prior to use. [b] Determined
from the 1H NMR spectrum of the crude reaction mixture. [c] Isolated yield after chromatography on silica
1
gel. [d] Yield from H NMR spectrum of the crude reaction mixture based on the isolated yield of 24a or 25a.
in
competition
with
the
1
[e] Yield by H NMR is 34%.
VAPOL and VANOL catalysts.
It turns out that these effects
derived from the (R)-enantiomer of a-methylbenzyl amine
19. The reactions of this imine with ethyl diazoacetate 2
were carried out with catalysts derived from both the
VANOL and VAPOL ligands and in both CCl4 and toluene
as solvents with the results presented in Table 1. There is an
inherent preference for this chiral imine to give diastereo-
appear to be small as there is very little difference in the se-
lectivities seen with the (R)- or (S)-enantiomers of either
ligand (Table 2). This is also true with the nonchiral catalyst
Table 2. Matched and mismatched aziridinations of the cyclohexylethyl
imine (R)-31a.[a]
mer 24a with the nonchiral catalyst BACTHNUTRGENUG(N OPh)3, which pro-
vides an 86:14 selectivity in CCl4 and a 75:25 selectivity in
toluene (Table 1, entries 5 and 10). The results from Table 1
definitely show that there is a synergism between the chiral
centers in the imine and in the catalyst with the matched
case resulting from the reaction of the (R)-enantiomer of
imine 26a and the (S)-enantiomer of the ligand. In the
matched case, the stereoselectivity is ꢀ25:1 in favor of 24a
over 25a for both ligands and in both solvents (Table 1, en-
tries 1, 3, 8, and 10). The mismatched cases result from the
reaction of the (R)-enantiomer of the imine 26a with the
(R)-enantiomer of the ligand. The catalyst derived from the
(R)-ligand is able to flip the diastereoselection in favor of
diastereomer 25a but only by a factor of approximately two
and again this is true for both ligands and for both solvents
(Table 1, entries 2, 4, 9, and 11). The total yields of aziridine
products drops in the mismatched cases and the mass bal-
ance is largely made up for by the formation of the enamine
products 29 and 30. In all cases, no trans isomers of 24a or
Ligand
Conv. 32a:33a[b] Yield
[%]
Yield
Yield
32a [%][c] 33a [%][d] 34/35 [%][d]
1
2
3
4
5
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
90
66
64
83:17
20:80
83:17
25:75
56:44
35
7
32
15
nd
6
28
5
45
nd
3/7
3/5
2/4
4/9
nd
(R)-VANOL 100
(OPh)3 only 38
BACHTUNGTRENNUNG
[a] Unless otherwise specified all reactions were run at 0.5m in imine
with 1.2 equivalents of 2 and 10 mol% catalyst at room temperature for
24 h. The cis:trans selectivity was>50:1 in all cases. The catalyst for the
reactions was prepared from 1 equivalent of the ligand, 4 equivalents of
BACHTNUGTRNEUNG(OPh)3 and 1 equivalent H2O according to Method B in the Experimen-
tal Section in the Supporting Information. nd=not determined. Imine
(R)-31a was an oil and used without purification after being prepared
from the corresponding amine with>99% ee. [b] Determined from the
1H NMR spectrum of the crude reaction mixture. [c] Isolated yield after
chromatography on silica gel. [d] Yield from 1H NMR spectrum of the
crude reaction mixture based on the isolated yield of 32a.
1
25a could be detected in the H NMR spectra of crude reac-
tion mixtures, as indicated by the absence of doublets with
the proper coupling constants (Jꢁ3 Hz) for the methine
5304
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5302 – 5313

Catalytic Asymmetric Aziridination of Imines
FULL PAPER
B
G
ligand (R with S and S with R). No detectable amount of
the diastereomer 37a was formed in the reaction of (S)-36a
with either the (R)-VAPOL or (R)-VANOL catalysts, both
of which give the aziridine product as a single diastereomer
38a in 80–85% yield. This is a slightly stronger matched/
mismatched pair than seen in the reactions of the a-methyl-
benzyl imine 26a. The selectivity flips over for the mis-
matched case with the (S)-VAPOL catalyst giving a 69:31
preference for the 37a over 38a, but the (S)-VANOL gives
an essentially equal mixture of the two. This is also reflected
32a and 33a (Table 2, entry 5). The selectivity imparted by
the ligands is 5:1 in favor of 32a with the (S)-enantiomers of
VANOL and VAPOL (Table 2, entries 1 and 3) and 3–4:1 in
favor of 33a with the (R)-enantiomers of the ligands (en-
tries 2 and 4). The relative stereochemistry of 32a and 33a
were not assigned, but rather assumed to correlate with 24a
and 25a. The total mass balance of these reactions for the
two diastereomers together is only 35–60%, which translates
to 45–60% if the percent conversion for the reaction is
taken into account. The remainder of the material balance
was not determined, except for the small amounts of the en-
in the fact that the nonchiral catalyst BACTHNUTRGENUG(N OPh)3 gives a strong
selectivity in favor of the matched diastereomer (92:8). The
major product 37a from the mismatched reaction of the
imine (S)-36a with (S)-VAPOL was crystalline and the rela-
tive stereochemistry of 37a was determined by X-ray dif-
fraction.
amine products that were detected in the H NMR spectrum
of the crude reaction mixture. It was noted that the reaction
of the imine 31a was not complete in 24 h with either the
(S)-VAPOL- or (R)-VAPOL-derived catalysts. This is in
contrast to the corresponding aromatic analogue 26a in
which the reaction reaches the same degree of completion
in 1 h for both the matched and mismatched cases (Table 1,
entries 6 and 7). This is consistent with the fact that the azir-
idination with the imine derived from the all aliphatic amine
14 and benzaldehyde is twenty five times slower than the
corresponding imine derived from the aromatic benzhydryl
amine 10.[4b]
The aziridination of the imine 36a puts a tert-butyl group
up against a phenyl group in the control of diastereoselectiv-
ity in reaction with ethyl diazoacetate. As can be seen from
the data in Table 3, there are significantly disparate effects
between the (R)- and (S)-enantiomers for both the VAPOL
and VANOL ligands. The matched case is the same as with
the a-methylbenzyl imine 26a in which the (S)-enantiomer
of the imine is matched with the (R)-enantiomer of the
With the finding that both the a-methylbenzyl imine 26a
and the a-tert-butylbenzyl imine 36a have strong matched
and mismatched relationships with the VANOL and
VAPOL ligands, it was decided to investigate how the rates
of these reactions compare with the corresponding benzhy-
dryl imine 41a. These experiments were conducted by per-
forming pair-wise reactions between equimolar amounts of
two substrates with a deficiency of ethyl diazoacetate, such
that the reactions stops at 20% conversion at which point
the ratio of the two products is determined. By this measure,
the reaction of the a-tert-butylbenzyl imine 36a in the
matched case is a factor of three slower than the benzhydryl
imine 41a (Scheme 4). Interestingly, the a-methylbenzyl
imine 26a is three times faster than the benzhydryl imine
41a in the matched case. A similar experiment reveals that
the a-methylbenzyl imine 26a is 1.3 times faster than the
benzhydryl imine 41a in the mismatched case. Thus, the re-
Table 3. Matched and mismatched aziridinations of the phenylneopentyl
imine (S)-36a.[a]
Ligand
37a:38a[b]
Yield
Yield
Yield
37a [%][c]
38a [%][d]
39/40 [%][d]
1
2
3
4
5
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
69:31
<2:98
52:48
<2:98
8:92
56
25
nd
<2
11/2
<2
<2[d]
36
80[c]
33
<2[d]
nd
85[c]
nd
B
G
24/11
[a] Unless otherwise specified all reactions were run at 0.5m in imine
with 1.1 equivalents of 2 and 10 mol% catalyst at room temperature for
24 h. The cis:trans selectivity was >50:1 in all cases. The catalyst for the
reactions was prepared from 1 equivalent of the ligand, 3 equivalents of
BACHTUNGTRENNUNG(OPh)3 according to Method A in the Experimental Section in the Sup-
porting Information. nd=not determined. Imine (S)-36a was purified by
crystallization and was prepared from an amine of greater than 99.5% ee.
[b] Determined from the 1H NMR spectrum of the crude reaction mix-
ture. [c] Isolated yield after chromatography on silica gel. [d] Yield from
1H NMR spectrum of the crude reaction mixture based on the isolated
yield of 37a or 38a.
Scheme 4.
Chem. Eur. J. 2012, 18, 5302 – 5313
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5305

W. D. Wulff et al.
action of the a-methylbenzyl imine 26a in the matched case
is 2.3 times faster than the same reaction in the mismatched
case. This is consistent with the slightly greater reaction
progress noted for the matched reaction when the reaction
of the imine (R)-26a was stopped after 1 h (Table 1, en-
tries 6 and 7).
the stereoselectivity in the aziridination reaction. This result
was a little unexpected, since it was observed by X-ray anal-
ysis that the binding of the imine substrate in the VAPOL
catalyst 9 results from different types of interactions for the
two phenyl groups.[3b] However, it should be noted that in
computational studies on the transition states for the aziridi-
nation reaction that these types of interactions of the phenyl
groups with the catalyst do not exist.[6]
As indicated by the data in Scheme 5, chiral benzhydryl
substituents provide absolutely no resistance to the catalyst
Substrate scope for cis-aziridines with a-methylbenzyl imine
26: With regard to the selection of a chiral auxiliary, it was
decided to move forward with the a-methylbenzyl amine 19
(Scheme 2). The a-tert-butylbenzyl amine 21 was shown to
give slightly higher diastereoselectivities in reactions of its
imine from benzaldehyde (Table 3) than the corresponding
imine from amine 19 (Table 1); however, the a-tert-butyl-
benzyl amine 21 is not commercially available. Both enan-
tiomers of the a-methylbenzyl amine 19 are commercially
available and, furthermore, both enantiomers are roughly
the same price as the benzhydryl amine 10. The diastereose-
lectivity for the aziridination of the imine 26a are very high
and the two diastereomers 24a and 25a are easy to separate
by column chromatography (Table 1). The isolated yields of
the major diastereomer 24a are also very high (86%). Even
though the yields and diastereoselectivities are slightly
higher in CCl4 than in toluene, it was decided to examine
the scope of the reaction with imines from other aldehydes
in toluene as solvent for reasons of cost and safety. With the
VANOL catalyst in toluene the diastereomer 24a was the
exclusive product with no detectable amount of the diaste-
reomer 25a (Table 1, entry 10).
Scheme 5.
that dominates these reactions with complete catalyst con-
trol. The reaction of the imine (S)-43a (>99% ee) with
a phenyl and a p-bromophenyl substituent gave a 97:3 mix-
ture of 44a to 45a in favor of 44a with the (S)-VAPOL cata-
lyst and a 3:97 mixture with the (R)-VAPOL catalyst. The
fact that the p-bromo substituent has no effect at all and
that the matched and mismatched reactions both give a 97:3
selectivity is consistent with the fact that the benzhydryl
imine 41a will react with ethyl diazoacetate to give aziridine
42a in 93% ee (96.5:3.5 e.r.) in carbon tetrachloride as sol-
vent.[1c] On this basis the relative stereochemistry of 44a and
45a were assigned such that the 2R,3R isomer is produced
by the S enantiomer of the ligand, as is the case for the reac-
tion of imine 41a. An alternative method for probing for
matched and mismatched effects is to look for their manifes-
tation in kinetic resolution experiments. This was examined
for the racemic benzhydryl imines 43a and 43b, bearing p-
bromo and p-methoxy substituents, respectively. The reac-
tion of the racemic p-bromophenyl imine 43a with
0.5 equivalents of ethyl diazoacetate gave a 50:50 mixture of
the diastereomers 44 and 45. This is consistent with the reac-
tions of the optically pure imine 43a with catalysts from (S)-
and (R)-VAPOL. Exactly the same results were obtained
with the racemic p-methoxy imine 43b and thus it is clear
that electronic differences in the two phenyl groups of the
benzhydryl unit on the imine do not have any influence on
The scope of the aziridination of imines from a-methyl-
benzyl amine 19 was explored with five additional aromatic
aldehydes and with three aliphatic aldehydes and the results
are summarized in Table 4. The imines 26c and 26d are
solids and were purified by crystallization prior to use. All
of the rest of the imines in Table 4 are oils and were used
without purification directly after they were prepared. The
optimal results for the para-nitrophenyl imine (R)-26b is
with (S)-VANOL, which gives the diastereomeric aziridine
24b in 90% isolated yield along with 4% of the diastereo-
mer 25b, which could be easily be separated by chromatog-
raphy on silica gel (Table 4, entry 3). A similar situation was
found with the para-bromophenyl imine (R)-26c for which
the strongest matched case was found with the (S)-VANOL
ligand giving a 97:3 mixture of aziridines 24c and 25c, which
were easily separated to give 24c in 77% yield (Table 4,
entry 8). Both ligands were equally effective in the aziridina-
tion of the para-methylphenyl imine (R)-26d with the isola-
tion of the matched imine 24d in 70–71% yield with no de-
tectable amount of the diastereomer 25d formed in either
reaction (Table 4, entries 11 and 13). A dramatic difference
between the two ligands was observed with the para-me-
thoxyphenyl imine (R)-26e, for which even for the matched
case the (S)-VAPOL catalyst only goes to 46% completion,
while the (S)-VANOL catalyst goes to completion and pro-
vides a 63% isolated yield of the pure aziridine 24e
5306
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5302 – 5313

Catalytic Asymmetric Aziridination of Imines
FULL PAPER
Table 4. Matched and mismatched aziridinations of the phenethyl imines (R)-26.[a]
for the imine (R)-26g derived
from cyclohexane carboxalde-
hyde. In fact the diastereoselec-
tivity is nearly equal with the
(S)- and (R)-ligands (83:17 vs
23:77) and thus this is more
a case of catalyst control rather
than substrate control. The dia-
stereomers 24g and 25g are
easy to separate and a greater
than 70% yield of either pure
diastereomer can be obtained
with the proper choice of the
chirality of the VANOL ligand
(Table 4, entries 28 and 29).
The imine (R)-26h derived
from tert-butyl carboxaldehyde
R
Ligand
Conv.
[%]
24:25[b]
Yield
Yield
Yield
24 [%][c]
25 [%][d]
29/30 [%][d]
1
2
3
4
5
6
7
8
9
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
100
100
100
100
100
100
100
100
100
94
100
100
100
100
61
94:6
33:67
96:4
33:67
71:29
94:6
38:62
97:3
31:69
77:23
>98:2
33:67
>98:2
38:62
80:20
95:5
74
19
90
19
31
82
24
77
21
35
71
19
70
21
28
35
nd
63
nd
nd
62
23
52
16
<21
66
14
72
23
17
61
13
79
22
18
28[h]
29[i]
4
38
3
4/5
12/14
4/3
12/14
5/7
38
12
5
BACTHNGUTREN(UNG OPh)3 only
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
4/4
40[e]
2
13/13
11/9
9/9
11/8
0/1
3/7
3/13
6/8
4/3
3/5
nd
7/23
nd
46
10
<1
38[f]
<1
35
7
2
nd
1
nd
nd
<1
16
<1
13
<7
13
47
14
76
21
6
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
B
N
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
gives
a
matched and mis-
matched relationship with both
VANOL and VAPOL that is
more closely related to aryl al-
dehydes than to cyclohexane
carboxaldehyde. A 93:7 selec-
tivity in favor of aziridine 24h
is observed for the matched
case with the (S)-VANOL cata-
lyst resulting in a 79% isolated
yield of the pure aziridine. The
imine (R)-26i prepared from
the a-unbranched aldehyde n-
butanal, reacted to give low
yields of the aziridines 24i and
25i. Even in the matched case
with a 80:20 ratio of 24i to 25i,
the aziridine 24i could only be
isolated in 28% yield. It is in-
teresting to note that the same
imine will react with the diazo-
acetamide 51 to give a trans-
aziridines 53i (Table 7, see
below) in much greater yields.
The causes of the low yields for
imine 26i may be related to
those responsible for low yields
in the aziridination of imines
derived from the DAM amine
11 and n-butanal.[4c] Finally, the
B
N
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
46
21
100
29
8
47:53
98:2
44:56
nd
B
N
nd
1/0
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
100
100
100
100
28
100
100
100
100
53
100
31
100
100
43
>98:2
41:59
>98:2
44:56
75:25
83:17
23:77
83:17
23:77
55:45
91:9
38:62
93:7
41:59
58:42
80:20
47:53
15/0
9/6
11/0
<15/<27
0/0
BACTHNUTRGEN(UGN OPh)3 only
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
0/0
0/0
0/0
0/0
0/0
0/0
0/0
0/0
B
N
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
22[g]
6
31
13
7
B
N
0/0
0/0
0/0
(S)-VANOL
(R)-VANOL
100
100
33
[a] Unless otherwise stated all reactions were run at 0.5m in imine with 1.2 equivalents of 2 and 10 mol% cata-
lyst in toluene at room temperature for 24 h. The catalyst was prepared from 1.0 equivalent of the ligand,
4.0 equivalents of BACHTUNGTRENNUNG(OPh)3 and 1.0 equivalent H2O according to Method B in the Experimental Section in the
Supporting Information. nd=not determined. Imine (R)-26c and (R)-26d were purified by crystallization. All
1
other imines were oils and used without purification. [b] Determined from the H NMR spectrum of the crude
1
reaction mixture [c] Isolated yield after chromatography on silica gel. [d] Yield from H NMR spectrum of the
crude reaction mixture based on the isolated yield of 24. [e] Isolated yield is 27%. [f] Isolated yield is 32%.
[g] Isolated yield is 20%. [h] Yield from 1H NMR spectrum is 41% with Ph3CH as internal standard. [i] Yield
1
determined from H NMR spectrum with Ph3CH as internal standard.
reactions catalyzed by B
ACHTUNGTRENNUNG(OPh)3
tend to give low yields and se-
(Table 4, entry 18). Note that both catalysts are very slow in
the mismatched cases giving 21 and 29% conversion under
the same conditions (Table 4, entries 17 and 19). A single
diastereomer of the ortho-methylphenyl aziridine 24 f is pro-
duced in the reactions of the corresponding imine (R)-26 f
with catalysts from both ligands with a higher yield regis-
tered with the (S)-VAPOL catalyst (Table 4, entry 21). The
weakest matched and mismatched relationship was found
lectivities in the range of 55:45–80:20 which is very similar
to those observed for the nonchiral Lewis acids SnCl4 and
CoCl2 (Scheme 3).
In all of the aziridination reactions of imines 26 investigat-
ed up to this point (Tables 1 and 4) the products have been
exclusively the cis isomers of the diastereomeric aziridines
24 and 25. Deviant from these observations are the reactions
of imines 26 derived from ortho-halo-benzaldehydes
Chem. Eur. J. 2012, 18, 5302 – 5313
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5307

W. D. Wulff et al.
Table 5. Matched and mismatched aziridinations of the o-bromo- and iodophenyl imines (R)-26.[a]
the same as that determined for
the phenyl aziridine 24a
(Table 1). The relative stereo-
chemistry of the cyclohexyl
aziridine 24g was assigned by
a chemical correlation with the
known aziridine 62g[4a] as out-
lined in Scheme 9 (see below).
The assignment of the relative
stereochemistry for the tert-
butyl-substituted aziridine 24h
was made by conversion to the
p-bromobenzoate 48, which was
a solid that gave crystals suita-
X
Ligand
Yield
Yield
Yield
Yield
cis/
24:25 46:47 Yield
29/30 [%][c]
24 [%][b] 25 [%][c] 46 [%][d] 47 [%][c] trans
ble
for
X-ray
analysis
(Scheme 6). The stereochemis-
try of the n-propyl aziridine 24i
was assumed to be the same as
the phenyl, cyclohexyl and tert-
butyl analogues. The relative
stereochemistry of the cis- and
trans-isomers 24j and 46j was
determined as outlined in
Scheme 6. Tin hydride reduc-
tion of the 24j gives an 89%
yield of aziridine 24a, a com-
pound that was found to be
identical with the aziridine
formed from the reaction of
imine (R)-26a in the matched
case with (S)-VAPOL (Table 1,
entry 1). The assignment of the
relative stereochemistry for the
trans-aziridine 46j was deter-
1
2
3
4
5
6
7
8
Br (S)-VAPOL
Br (R)-VAPOL
Br (S)-VANOL
Br (R)-VANOL
48
10
43
10
5
10
1
8
4
1
8
1
4
14
33
8
30
50
15
40
15
30
51
7
3
11
2
7
7
4
7
3
9
71:29 91:9
67:33 10/16
35:65 50:50 93:7 15/21
70:30 97:3 44:56 10/12
36:64 55:45 92:8 10/13
23:77 75:25 88:12 13/13
50:50 94:6 67:33 11/17
26:74 50:50 91:9 15/21
52:48 96:4 67:33 10/12
25:75 63:37 91:9 10/13
23:77 33:67 85:15 11/11
Br
I
I
I
I
B
(OPh)3 only 13
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
(R)-VANOL
21
8
23
7
9
10[e]
I
B
N
6
12
[a] Unless otherwise specified all reactions were run at 0.5m in imine with 1.2 equivalents of 2 and 10 mol%
catalyst at room temperature for 24 h and went to 100% completion. The catalyst for the reactions was pre-
pared from 1 equivalent of the ligand, 4 equivalents of BACTHNUTRGNEU(NG OPh)3 and 1 equivalent H2O according to Method B
in the Experimental Section in the Supporting Information. Imine (R)-26j and (R)-26k were oils and used
without purification. [b] Isolated yield of 24j after chromatography on silica gel. The yields of 24k were deter-
1
mined from H NMR spectrum of the crude reaction mixture and based on the isolated yield of 46k. [c] Yield
from 1H NMR spectrum of the crude mixture and based on the isolated yield of 24j or 46k. [d] Isolated yield
of 46k after column chromatography on silica gel. The yield of 46j were determined from the 1H NMR spec-
trum of the crude mixture and are based on the isolated yield of 24j. [e] The reaction went to 80% comple-
tion.
(Table 5). The ortho-bromo and ortho-iodo derivatives 26j
and 26k give mixtures of all four possible diastereomers. As
was seen with all of the other reactions of imines 26 produc-
ing cis-aziridines, the matched case for the (R)-enantiomers
of 26j and 26k is with the (S)-enantiomers of VANOL and
VAPOL catalysts. For example, the reaction of (R)-26j with
the (S)-VANOL catalyst gives a 97:3 selectivity for the cis
diastereomer 24j over the cis diastereomer 25j with a total
cis:trans selectivity of 70:30. It was interesting to observe
that the matched cases for the trans-aziridines 46 and 47 are
the mismatched cases for the cis-aziridines 24 and 25. The
reaction of the (R)-enantiomer of 26k with the (R)-VANOL
catalyst gives a 92:8 mixture of the trans diastereomers 46k
and 47k with a total trans:cis selectivity of 64:36. In the
matched reactions of 26j for the formation of both the cis
and trans diastereomers (24 and 46), facial selectivity for the
reaction of the imine is the same, but that for the diazo com-
pound is changed. Very similar reaction patterns were seen
for the reactions of the ortho-iodo imine 26k with the ex-
ception that the yields tended to be a little lower for the cis-
aziridines in the matched case.
mined by hydrogenation with Pearlmanꢁs catalyst, which re-
sulted in the reduction of the bromide, reductive ring-open-
ing of the aziridine, and reductive cleavage of the a-methyl-
benzyl group on the nitrogen. The resulting ethyl ester of
phenyl alanine 49 was found to have the S configuration by
comparing its rotation to that of the known compound.[1a]
Similar treatment of 24a obtained as the major diastereo-
mer from the reaction of imine (R)-26a in the matched case
with (S)-VAPOL (Table 1) gave the (R)-enantiomer of phe-
nylalanine ethyl ester, which reveals that the cis-aziridine
24a (and thus cis-aziridine 24j) and the trans-aziridine 46j
differ in the configuration at C2. It was assumed that the rel-
ative stereochemistry of the ortho-iodophenyl aziridines 24k
and 46k is the same as that for the corresponding ortho-bro-
mophenyl aziridines. The aziridination of the ortho-bromo-
phenyl imine 41j derived from benzhydryl amine has been
previously shown to give a 1.9:1 mixture of the cis- and
trans-aziridines (Scheme 6).[1c] It was later shown that the
absolute stereochemistry of the cis-isomer 42j is (2S,3S) and
that of the trans-isomer 50j is (2R,3S).[7] It is of interest to
note that the while the matched cis-aziridine 24 differs from
the matched trans-aziridine 46 by the configuration at C2
(Table 5), the cis- and trans-aziridines 42j and 50j from the
The relative stereochemistry of the cis-aziridines 24 with
an aryl substituent R shown in Table 4 was assumed to be
5308
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5302 – 5313

Catalytic Asymmetric Aziridination of Imines
FULL PAPER
Scheme 7.
tyl tin hydride without opening the aziridine ring to give the
trans-aziridine 46a in 68% yield (Scheme 7), an aziridine
which is not observed in the reaction of imine (R)-26a
(Table 1). However, the usefulness of this approach is limit-
ed not only by the fact that other reducible functional
groups cannot be present, but also by the fact that the trans-
a-halophenyl aziridines can only be obtained in low yields
(30–50%) even in the matched case (Table 5).
The most practical approach to trans-aziridines using the
double stereodifferentiation approach would be if the reac-
tions of a-methylbenzyl imines with diazoacetamides were
themselves trans-selective as are the reactions of nonchiral
imines. That this is the case, is shown by the data in Table 6,
which presents the results of the aziridination of the imine
(R)-26a with the diazoacetamide 51. The VAPOL catalyst
gives a 88:12 ratio of the trans to cis isomers with a 91:9
mixture of the two trans diastereomers 53a and 54a. The
relative stereochemistry of major diastereomer 53a pro-
duced from the reaction of (R)-26a with the (R)-VANOL or
(R)-VAPOL catalyst was established by the reductive ring-
opening to give the phenyl alanine derivative 59a of known
optical rotation (Scheme 8).[8] As is the case for the forma-
tion of the trans isomers from the a-halophenyl imines 23j
and 23k, the matched reaction is that of the (R)-imine with
the (R)-VAPOL ligand. The VANOL ligand gives a slightly
better profile with an 89:11 trans to cis ratio and a 95:5 ratio
of the two trans diastereomers, although both ligands give
the same isolated yield of the trans diastereomer 53a (70–
71%). The profile for the VANOL catalyst could be im-
proved by lowering the reaction temperature to 08C
(Table 6, entries 4 versus 5). Here the trans:cis selectivity is
now 96:4 and the selectivity for the trans diastereomer 53a
is complete with no detectable amount of 54a formed. As
a consequence of these increased selectivities, the isolated
amount of the trans-aziridine 53a is increased from 70 to
78% yield. The relative stereochemistry of the cis-aziridine
55a was determined by its conversion to 24a (see Support-
ing Information).
Scheme 6.
nonchiral imine 41j differ by the configuration at C-3
(Scheme 6).
trans-Aziridines from a-methylbenzyl imines 26 and diazoa-
cetamides: The origin of the formation of substantial pro-
portions of trans-aziridines from imines 26j and 26k that
bear an ortho-halogen substituent (Table 5) is not under-
stood at this time, but it is clear that this is not due to just
the presence of an ortho-substituent, since the ortho-methyl-
phenyl imine 26 f does not give an detectable amount of the
trans-aziridine (Table 4, entries 21–25). What is perhaps
better understood is the reason that sec-diazoacetamides of
the type 51 give trans-aziridines (Scheme 7).[6,7] This switch
in diastereoselectivity from cis for the diazo ester
(Scheme 1) to trans for the diazoacetamide 51 has been at-
tributed to the ability of the N H bond of the 51 to partici-
2
À
pate in a hydrogen bond to the anionic boroxinate core of
the VANOL catalyst 9 (Scheme 1).[6] The trans-aziridines 52
can be obtained in high yields with excellent enantioselectiv-
ities for imines derived from both aliphatic and aromatic al-
dehydes.[7]
Access to diastereomerically and enantioselectively pure
trans-aziridines via a-methylbenzyl imines can be expanded
beyond ortho-halophenyl imines, since the ortho-bromo sub-
stituent in the trans-aziridine 46j can be reduced with tribu-
Chem. Eur. J. 2012, 18, 5302 – 5313
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5309

W. D. Wulff et al.
Table 6. Matched and mismatched trans-aziridinations of imine (R)-26 with diazoace-
tamide 51.[a]
surprising, since the same imine with the diazo
ester 2 shows an extremely weak matched/mis-
matched relationship (Table 4). The amount of cis
diastereomers could not be determined in the reac-
tions in Table 7, because the aliphatic regions of the
1H NMR spectrum of the crude reaction mixtures
consisted of many overlapping peaks that prevented
the unequivocal identification of their presence. It
was of great interest to learn that the matched reac-
tion with the cyclohexyl imine 26g is of the (R)-
imine with the (S)-ligand, which is opposite to that
observed for the phenyl imine 26a also in the trans-
aziridination (Table 6). The tert-butyl-substituted
imine (R)-26h also gave high selectivities (ꢀ97:3)
for the trans diastereomer 53h over the trans diaste-
reomer 54h and its matched case is the same as the
cyclohexyl imine 26g, for which the (R)-imine gives
the highest selectivity with the (S)-ligands. The tert-
butyl-substituted trans diastereomer 53h was a crys-
talline compound that provided single crystals that
were analyzed by X-ray diffraction to show that the
phenyl and tert-butyl imines (R)-26a and (R)-26h
give the same relative stereochemistry in the major
diastereomer in the matched cases, even though the
matched case with (R)-26a is with the (R)- ligands
and for (R)-26h is with the (S)-ligands. The aziri-
dines 53g and 53i were assumed to have the same
relative stereochemistry as 53a and 53h in the
matched case. The unbranched aliphatic imine (R)-
26i also gives high stereoselectivity (95:5) for the
Ligand
Yield Yield Yield Yield trans/ 53a:54a 55a:56a Yield
53a
54a
55a
56a
cis
57/58
[%][b] [%][c] [%][c] [%][c]
[%][c]
1
2
3
4
(S)-VAPOL
(R)-VAPOL
(S)-VANOL
25
71
28
17
7
18
4
<4
3
14
9
5
16
6
2
4
2
5
6
3
1
3
2
79:21
88:12
68:32
89:11
96:4
60:40 83:17
91:9 50:50
62:38 75:25
11/19
6/6
18/16
8/10
3/4
(R)-VANOL 70
95:5
>95:5
95:5
67:33
67:33
60:40
5[d] (R)-VANOL 78
6[e] (R)-VANOL 69
7
91:9
86:14
6/9
4/4
B
N
8
75:25 75:25
[a] Unless otherwise specified all reactions were run at 0.5m in imine with 1.2 equiva-
lents of 51 and 10 mol% catalyst at room temperature for 20 h and went to 100%
completion. The catalyst for the reactions was prepared from 1 equivalent of the
ligand, 3 equivalents of BH3·SMe2, 2 equivalents PhOH and 3 equivalents H2O accord-
ing to Method C in the Experimental Section in the Supporting Information. Imine
(R)-26a (>99% ee) is an oil and was purified by distillation prior to use. [b] Isolated
yield after chromatography on silica gel. [c] Yield from 1H NMR spectrum of the
crude reaction mixture and based on the isolated yield of 53a. [d] The reaction was
run at 08C for 24 h. [e] The catalyst was prepared from 1 equivalent of (R)-VANOL, trans diastereomer 53 and surprisingly, its matched
4 equivalents B(OPh)3 and 1 equivalent H2O according to Method B in the experi-
mental section. [f] This reaction went to 64% completion.
ACHTUNGTRENNUNG
case is the same as the phenyl imine 26a and oppo-
site to cyclohexyl and tert-butyl imines 26g and
26h. This variation in the matched/mismatched pair
with the structure of the imine is unexpected and
its meaning is not clear at this time. It is also not clear why
the reaction of the n-propyl imine (R)-26i with the diazoa-
cetamide 51 gives the trans-aziridine 53i in excellent yields
(79–80%) while reaction of the same imine with ethyl diazo-
acetate only gives low yields of the cis-n-propyl aziridine 24i
(Table 4). Nonetheless, it is important that at least one of
the diazo compounds gives good yields, since not only are
aziridines with unbranched aliphatic groups at the 3- posi-
tion accessible, but the ring-opening of these aziridines
allows access to straight chain a- and b-amino acids deriva-
tives (Scheme 10, see below).
Scheme 8.
Since the asymmetric inductions for the trans-aziridines
52 (Scheme 7) derived from aliphatic aldehydes were not
generally as high as they were for those from aromatic alde-
hydes, focus was drawn to the evaluation of a-methylbenzyl
imines derived from aliphatic aldehydes and, in particular,
the unbranched, a-branched, and a,a-branched examples
presented in Table 7. Beginning with the cyclohexyl imine
(R)-26g it was found that very high selectivities (>96:4)
were observed for the formation of the trans diastereomer
53g over the trans diastereomer 54g even at 258C to pro-
vide a 78% isolated yield of 53g with the VAPOL catalyst
and an 80% yield with the VANOL catalyst. This high
degree of matched/mismatched stereo double differentiation
with the cyclohexyl imine 26g and the diazo acetamide 51 is
Synthesis of a- and b-amino acid derivatives: The aryl-sub-
stituted cis-aziridines 24a, 24d, and 24 f could be ring-
opened with Pearlmanꢁs catalyst under one atmosphere of
hydrogen in methanol, which also resulted in the cleavage
of the a- methylbenzyl group on the nitrogen (Scheme 9). It
was found convenient for isolation purposes to perform the
reduction in the presence of Boc2O (Boc=tert-butyloxycar-
bonyl), which resulted in the isolation of the N-Boc alanine
derivatives 60 in excellent yields. The reduction of the para-
nitrophenyl-substituted aziridine 24b occurred with simulta-
5310
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5302 – 5313

Catalytic Asymmetric Aziridination of Imines
FULL PAPER
Table 7. Matched and mismatched trans-aziridinations of alkyl imines
(R)-26 with diazoacetamide 51.[a]
R
Ligand
Conv.
[%]
53:54 Yield
Yield
53 [%][c] 54 [%][d]
1
(S)-VAPOL
100
>96:4
78
<3
2
3
4
5
(R)-VAPOL
(S)-VANOL
(R)-VANOL
100
100
100
67:33
>96:4
75:25
91:9
59
80
30
<3
20
4
60
B
G
42[e]
6
(S)-VAPOL
87
97:3
69
2
7
8
9
(R)-VAPOL
(S)-VANOL
(R)-VANOL
70
81
63
64
50:50
>97:3
67:33
30
61
19
30
<2
10
8
10
B
G
83:17 42[e]
11
(S)-VAPOL
100
52:48
15
14
12
13
14
15
(R)-VAPOL
(S)-VANOL
(R)-VANOL
100
100
100
95:5
67:33
95:5
79
40
80
43
4
20
4
B
(OPh)3 only 100
83:17
9
Scheme 9.
[a] Unless otherwise stated all reactions were run at 0.2m in imine with
1.2 equivalents diazoacetamide 51 and 10 mol% catalyst in toluene at
room temperature for 20–24 h. The catalyst was prepared from 1.0 equiv-
alent of the ligand, 3.0 equivalents of BH3·SMe2, 2.0 equivalents of PhOH
and 3.0 equivalents H2O according to Method C in the Experimental Sec-
tion in the Supporting Information. All imines were oil and used without
is quite curious given that the corresponding aliphatic-sub-
stituted trans-aziridines 46 undergo nearly complete ring-
opening under the same conditions (Scheme 10). The hydro-
genation of aziridine-2-carboxylates is generally observed to
be faster with aryl versus alkyl groups in the 3-position and
the switch in regioselectivity with alkyl groups to opening at
the 3-position to give b-amino esters has been previously ob-
served.[10–12]
1
purification. [b] Determined from the H NMR spectrum of the crude re-
action mixture. [c] Isolated yield after chromatography on silica gel.
1
[d] Yield from H NMR spectrum of the crude reaction mixture based on
the isolated yield of 53. [e] Yield from 1H NMR spectrum of the crude
reaction mixture with Ph3CH as internal standard.
neous reduction of the nitro group to an aniline and final
isolation provided the bis-Boc-protected 4-aminophenylala-
nine derivative 61. It was anticipated that reduction of the
para-bromophenyl aziridine 24c under the same conditions
would lead to reduction of the bromide substituent concomi-
tant with ring-opening based on our previous experience in
the synthesis of the cell adhesion inhibitor BIRT-377.[9] In
this synthesis we found that it was possible to achieve reduc-
tive ring-opening of the benzhydryl aziridine corresponding
to 24c with BH3·NMe3 and trifluoroacetic acid with the
para-bromide substituent being left untouched. The aliphat-
ic-substituted cis-aziridines 24g and 24h do not undergo
ring-opening under the same conditions as the aryl aziri-
dines, but rather simply undergo reductive cleavage of the
a-methylbenzyl group to yield the N-Boc aziridines 62g and
62h. If the catalyst loading and reaction time are increased,
a small amount (24%) of the b-amino ester (R)-63g can be
observed for the cyclohexyl aziridine 24g (Scheme 9). This
Scheme 10.
Chem. Eur. J. 2012, 18, 5302 – 5313
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5311

W. D. Wulff et al.
Table 8. Averaged ligand and N-substituent effects over nine imines with
aromatic and aliphatic substituents R1.
The direct reductive ring-opening of the trans-aziridine-2-
carboxamide 53a was relatively ineffective giving only
a 40% isolated yield of the corresponding phenylalanine de-
rivative 59a (Scheme 8). In an effort to increase the overall
efficiency of the conversion of trans-aziridine-2-carboxa-
mides to a- and b-amino acid derivatives, an alternative ap-
proach was examined that begins with the initial conversion
of the trans-aziridine-2-carboxamides 53 to the trans-aziri-
dine-2-carboxylates 46 (Scheme 10). This was accomplished
by the in-situ activation of the amide by the introduction of
a Boc group and then nucleophilic displacement of the re-
sulting carbamate with sodium ethoxide to generate the cor-
responding ethyl esters 46. Reductive ring-opening of these
trans-aziridine esters proceeded smoothly to give N-Boc
protected a-amino esters 60 and the b-amino esters 63 in
good to excellent yields. Note that the 3-phenylaziridine-2-
carboxamide 53a could be converted to the phenylalanine
derivative 60a in 83% overall yield, a dramatic improve-
ment of the 40% yield for the direct reductive ring-opening
(Scheme 8 vs. 10). We were quite pleased to find that unlike
the cis-aziridine ethyl esters 24 (Scheme 9) the trans-aziri-
dine ethyl esters 46 with aliphatic substituents underwent
ring-opening much faster and gave good-to-excellent yields
of b-amino esters. The cyclohexyl aziridine 46g gave 63g in
90% yield and the n-propyl aziridine 46i gave 63i in 77%
yield along with a 12% yield of the unopened N-Boc aziri-
dine 64i under these conditions. The reductive ring-opening
of the tert-butyl aziridine 46h is slower and gives only
a 55% yield of 63h and required a doubling of the reaction
time. The faster rates for the reductive ring-opening of
trans-aziridine-2-carboxylate ethyl esters with aliphatic sub-
stituents is not understood at this time and may be a surface
phenomenon of the heterogeneous catalyst involving coordi-
nation of the ester group during reduction which may be
hindered by a neighboring cis-aliphatic substituent.
N-substituents
(P)[a]
Aziri- Ligand
dine
Average Average
Source
yield [%] ee [%][b]
(R)-a-methylbenzyl cis
VAPOL 70[c,d]
VANOL 72[c,e]
100 (ꢀ87) this work
100 (ꢀ90)
cis
trans[f] VAPOL 74
100 (ꢀ90)
trans[f] VANOL 75
100 (ꢀ90)
benzhydryl
DAM
cis
cis
cis
cis
cis
cis
cis
cis
VAPOL 70
VANOL 77
VAPOL 73
VANOL 78
VAPOL 88
VANOL 90
VAPOL 92
VANOL 91
88
88
88
85
95
94
97
96
[1c]
[4c]
[4b]
[4c]
BUDAM
MEDAM
[a] See Scheme 2. [b] % de is indicated in parentheses. [c] Isolated yield
of purified major diastereomers. [d] Eight substrates. [e] Seven substrates.
[f] Four substrates.
crystallization.[1c] The best substituent in terms of yields and
asymmetric induction is the MEDAM group and this is the
substrate of choice for many applications, especially when
the product aziridines are not crystalline.[4c] In the present
work, it has been demonstrated that imines derived from a-
methylbenzyl amine 19 will undergo the AZ reaction to give
good yields of aziridines for which there is a strong matched
case between the (R)-imine and (S)-VANOL- or (S)-
VAPOL-derived catalyst for cis-aziridines (the matched case
depends on the substituent for trans-aziridines). The advant-
age of this method is that no final enhancement of the opti-
cal purity of the aziridine products is needed and this ad-
vantage is enabled by the fact that the minor diastereomer
that is formed (if any) can be easily separated. This aziridine
synthesis with a-methylbenzyl amine 19 has the additional
advantage that this amine is commercially available in both
antipodes, each of which is about the same cost as benzhy-
dryl amine. Furthermore, the aziridination with a-methyl-
benzyl amine 19 may well prove superior to those with
MEDAM or BUDAM amines 12 and 13 in providing trans-
aziridines with high optical purity.[7] Averages for trans-aziri-
dines are not given for other N-substituents, since the ali-
phatic and aromatic imines were individually optimized for
a particular N-substituent and the data for a particular N-
substituent over a range imines is not yet available.[7] Thus,
there are a variety of options for obtaining aziridines with
high optical purity from the AZ reaction and the aziridina-
tion with a-methylbenzyl amine 19 is one that has particular
features that may prove to be ideal for many situations.
Conclusion
The catalytic asymmetric aziridination reaction catalyzed by
VANOL and VAPOL boroxinate catalysts was first reported
in 2000 from the reaction of ethyl diazoacetate with imines
bearing a benzhydryl substituent.[1b,c] Improvements in asym-
metric inductions and yields have been subsequently realiz-
ed by employing other diarylmethyl substituents on the
imine, such as DAM, BUDAM and MEDAM (Scheme 2).
Table 8 presents the average yield and average asymmetric
induction with imines derived from nine different aromatic
and aliphatic aldehydes with five different N-substituents
(including the a-methylbenzyl substituent from the present
work) and with both the VANOL and VAPOL ligands. It is
interesting and still quite astonishing to consider that the
VANOL and VAPOL ligands give essentially the same
yields and inductions in all cases. Benzhydryl amine 10 is
commercially available and although it gives lower induc-
tions, the aziridine products tend to be crystalline and often
the optical purity can be increased to>99% ee with one
Experimental Section
For experimental details and characterization of the compounds please
see the Supporting Information. CCDC-838600 (37a), 838602 (48), and
838601 (53h) contain the supplementary crystallographic data for this
5312
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 5302 – 5313

Catalytic Asymmetric Aziridination of Imines
FULL PAPER
paper. These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[7] A. Desai, W. D. Wulff, J. Am. Chem. Soc. 2010, 132, 13100–13103.
[8] P. Pelagatti, M. Carcelli, F. Calbiani, C. Cassi, C. P. Elviri, U. Rizzot-
ti, D. Rogolino, Organmetallic 2005, 24, 5836–5844.
[9] A. P. Patwardhan, V. R. Pulgam, Y. Zhang, W. D. Wulff, Angew.
Chem. 2005, 117, 6325–6328; Angew. Chem. Int. Ed. 2005, 44, 6169–
6172.
[10] F. A. Davis, J. Deng, Y. Zhang, R. C. Haltiwanger, Tetrahedron
2002, 58, 7135–7143, and references therein.
Acknowledgements
This work was supported by the National Science Foundation (CHE-
0750319) and the National Institute of General Medical Sciences
(GM094478).
[11] a) D. Tanner, Angew. Chem. 1994, 106, 625; Angew. Chem. Int. Ed.
Engl. 1994, 33, 599; b) H. N. I. Osborn, J. B. Sweeney, Tetrahedron:
Asymmetry 1997, 8, 1693–1715; c) W. McCoull, F. A. Davies, Syn-
thesis 2000, 1347–1365; d) B. Zwanenburg, P. T. Holte, Top. Curr.
Chem. 2001, 216, 93–124; e) J. B. Sweeney, Chem. Soc. Rev. 2002,
31, 247–258; f) P. Dauban, R. H. Dodd, Synlett 2003, 1571; g) X. E.
Hu, Tetrahedron 2004, 60, 2701–2743; h) C. Mçsser, C. Bolm in
Transition Metals for Organic Synthesis (Eds.: M. Beller, C. Bolm),
Wiley-VCH, Weinheim, 2006; i) I. D. G. Watson, L. Yu, A. K. Yudin,
Acc. Chem. Res. 2006, 39, 194–206; j) M. Pineschi, Eur. J. Org.
Chem. 2006, 4979–4988; k) G. S. Singh, M. D’hooghe, N. De Kimpe,
Chem. Rev. 2007, 107, 2080–2135; l) C. A. Olsen, H. Franzyk, J. W.
Jaroszewski, Eur. J. Org. Chem. 2007, 1717–1724; m) P. Lu, Tetrahe-
dron 2010, 66, 2549.
[1] a) J. C. Antilla, W. D. Wulff, J. Am. Chem. Soc. 1999, 121, 5099–
5100; b) J. C. Antilla, W. D. Wulff, Angew. Chem. 2000, 112, 4692–
4695; Angew. Chem. Int. Ed. 2000, 39, 4518–4521; c) Y. Zhang, A.
Desai, Z. Lu, G. Hu, Z. Ding, W. D. Wulff, Chem. Eur. J. 2008, 14,
3785–3803. For a review, see: d) Y. Zhang, Z. Lu, W. D. Wulff, Syn-
lett 2009, 2715–2739.
[2] For a recent review on asymmetric aziridination, see: H. Pellissier,
Tetrahedron 2010, 66, 1509–1555.
[3] a) G. Hu, L. Huang, R. H. Huang, W. D. Wulff, J. Am. Chem. Soc.
2009, 131, 15615–15617; b) G. Hu, A. K. Gupta, R. H. Huang, M.
Mukherjee, W. D. Wulff, J. Am. Chem. Soc. 2010, 132, 14669–14675.
[4] a) Z. Lu, Y. Zhang, W. D. Wulff, J. Am. Chem. Soc. 2007, 129, 7185–
7194; b) Y. Zhang, Z. Lu, A. Desai, W. D. Wulff, Org. Lett. 2008, 10,
5429–5432; c) M. Mukherjee, A. K. Gupta, Z. Lu, Y. Zhang, W. D.
Wulff, J. Org. Chem. 2010, 75, 5643.
[5] a) H.-J. Ha, K.-H. Kang, J.-M. Suh, Y.-G. Ahn, Tetrahedron Lett.
1996, 37, 7069–7070; b) K.-D. Lee, J.-M. Suh, J.-H. Park, H.-J. Ha,
H. G. Choi, C. S. Park, J. W. Chang, W. K. Lee, Y. Dong, H. Yun,
Tetrahedron 2001, 57, 8267–8276; c) S.-H. Lee, I.-W. Song, Bull.
Korean Chem. Soc. 2005, 26, 223–224.
[12] A. P. Patwardhan, Z. Lu, V. R. Pulgam, W. D. Wulff, Org. Lett. 2005,
7, 2201–2204.
[13] a) J. M. Bao, W. D. Wulff, J. B. Dominy, M. J. Fumo, E. B. Grant,
A. C. Rob, M. C. Whitcomb, S. M. Yeung, R. L. Ostrander, A. L.
Rheingold, J. Am. Chem. Soc. 1996, 118, 3392–3405; b) A. Ding,
W. E. G. Osminski, H. Ren, W. D. Wulff, Org. Process Res. Dev.
2011, 15, 1089–1107.
[14] R. Imwinkelried, L. S. Hegedus, Organometallics 1988, 7, 702–706.
[15] G. R. Clemo, C. Gardner, R. Raper, J. Chem. Soc. 1939, 1958–1960.
[6] M. Vetticatt, A. Desai, W. D. Wulff, J. Am. Chem. Soc. 2010, 132,
13104–13107.
Received: August 14, 2011
Published online: March 20, 2012
Chem. Eur. J. 2012, 18, 5302 – 5313
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5313