Synthesis of N-H vinylaziridines: A comparative study

Article abstract of DOI:10.1016/S0040-4020(02)00610-5

Vinylaziridines are useful and versatile synthetic intermediates, as the relief of ring-strain provides a driving force for efficient ring-opening or ring-expansion reactions. Furthermore, the vinyl group can be derivatized into interesting functionalities. The ring-closure of vicinal amino alcohols constitutes a straightforward route to aziridines. Several methods exist for this transformation, although many cannot be applied to vinylaziridines due to their acid lability. This comparative study describes the most effective sequences for the formation of N-H vinylaziridines.

Full text of DOI:10.1016/S0040-4020(02)00610-5

TETRAHEDRON
Pergamon
Tetrahedron 58 (2002) 5979±5982
Synthesis of N±H vinylaziridines: a comparative study
Berit Olofsson, Roel Wijtmans and Peter Somfaip
Organic Chemistry, Department of Chemistry, Royal Institute of Technology, S-100 44 Stockholm, Sweden
Received 15 April 2002; revised 6 June 2002; accepted 13 June 2002
AbstractÐVinylaziridines are useful and versatile synthetic intermediates, as the relief of ring-strain provides a driving force for ef®cient
ring-opening or ring-expansion reactions. Furthermore, the vinyl group can be derivatized into interesting functionalities. The ring-closure of
vicinal amino alcohols constitutes a straightforward route to aziridines. Several methods exist for this transformation, although many cannot
be applied to vinylaziridines due to their acid lability. This comparative study describes the most effective sequences for the formation of
N±H vinylaziridines. q 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction
compounds are rather acid labile, which limits the number
of applicable protocols and makes puri®cation on silica gel
cumbersome. We chose amino alcohol 1 as model substrate,
and the desired transformation to aziridine 2 is shown in
Scheme 1.
Aziridines are useful and versatile synthetic intermediates,
as the relief of ring-strain provides a driving force for
ef®cient ring-opening or ring-expansion reactions.1,2 The
importance of aziridines is also well recognized in asym-
metric synthesis, where the need for chiral auxiliaries and
ligands is continuously increasing.3 Vinylaziridines consti-
tute an important subclass of aziridines. They have proven
to be useful intermediates for various types of natural and
synthetic compounds.4,5 Vinylaziridines can be selectively
ring-opened at the allylic position, take part in conjugate
addition reactions,6 Wittig and Claisen rearrangements,7,8
and furthermore the vinyl group can be derivatized into
interesting functionalities.9
The transformation can be conducted in three general ways:
(1) direct ring-closure of amino alcohol 1 to yield aziridine 2
is the most effective strategy, but suffers from the low reac-
tivity of 1;10 (2) activation of the hydroxy group of 1 into a
better leaving group, which should facilitate ring-closure;
(3) protection of the amino moiety of 1 should also increase
the reactivity towards ring-closure, although a deprotection
step is needed to yield 2. The two latter methods need high-
yielding reaction steps to compete with the direct ring-
closure, as several steps are needed to achieve the desired
transformation.
Existing enantioselective synthetic routes to aziridines
include asymmetric aziridination of alkenes and ring-
closure of vicinal hydroxy azides or amino alcohols.1,10 In
an ongoing project, ef®cient syntheses of both vicinal amino
alcohols and N±H vinylaziridines are of great importance,
and ring-closure of amino alcohols became the most
straightforward route to aziridines. There are several proce-
dures for ring-closure of b-amino alcohols to N-substituted
aziridines, and the plethora of methods encouraged us to
perform a comparative study to ®nd out which is the most
effective in the formation of N±H vinylaziridines. These
2. Results and discussion
2.1. Direct ring-closure
Direct ring-closure of amino alcohols to provide aziridines
is known to be dif®cult, and previous reports show moderate
yields of the corresponding N±H aziridines.1 Initial investi-
gations of the ring-closure of 1 to aziridine 2 using
Mitsunobu conditions11,12 were disappointing, but moderate
yields could be obtained in toluene at re¯ux.13 A carbamate
byproduct was irregularly formed in considerable amounts
due to reaction between the amino alcohol and DEAD.12
This could be prevented by changing the azo compound
ethyl group to the bulkier isopropyl group in DIAD. The
reaction rate could be increased by change of solvent to
THF; this might re¯ect the observation of improved solu-
bility of 1. Unfortunately, puri®cation of 2 demanded
repeated ¯ash chromatography to remove the formed
triphenylphosphine oxide, which decreased the yield
Scheme 1. Ring-closure of amino alcohol 1 to aziridine 2.
Keywords: amino alcohols; vinylaziridines; ring-closure.
Corresponding author. Tel.: 146-8-7906960; fax: 146-8-7912333;
p
e-mail: somfai@kth.se
0040±4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(02)00610-5

5980
B. Olofsson et al. / Tetrahedron 58 (2002) 5979±5982
Scheme 2. Hydroxy group activation prior to ring-closure.
considerably.14 In our experience, N±H vinylaziridines are
unstable on silica, and careful puri®cation on deactivated
silica improve the yields. Small-scale reactions were
puri®ed to give aziridine 2 in 80% yield, whereas large-
scale reactions gave 2 in 70% yield.15 In an attempt to
avoid the tedious puri®cation, polymer bound triphenyl-
phosphine was used. As expected, this decreased the reac-
tion rate, but gave an easily puri®ed crude product. Due to
byproduct formation, the isolated yield of 2 was slightly
lower.
Scheme 3. Amino group activation prior to ring-closure.
amino alcohols. The lability of N±H vinylaziridines limits
the number of useful activating groups, as the conditions
needed for deprotection of several activating groups are
expected to destroy the aziridine moiety.17 Our choice was
the triphenylmethyl (trityl) group, as it has been success-
fully employed in aziridination reactions and mild con-
ditions are used for deprotection. Activation by tritylation
proceeded almost quantatively,18 and with trityl amino
alcohol 4 in hand, several ring-closing methods were
performed (Scheme 3). The Mitsunobu protocol described
above afforded tritylaziridine 5, easily puri®ed, in 99%
yield.
2.2. Selective activation of the hydroxy group prior to
ring-closure
Selective activation of the hydroxy group prior to ring-
closure is dif®cult, as the amino group is more reactive
towards activating agents such as tosyl chloride. One
solution to this delicate problem is reaction of 1 with
chlorosulfonic acid to form sulfate ester 3 (Scheme 2),
which can be ring-closed to aziridine 2 under basic
conditions.16 Formation of salt 3 was nearly quantitative,
but ring-closure with excess NaOH at re¯ux furnished 2 in
moderate yield (Table 1, entry 1). Various solvents and
bases were screened, and the best result was achieved in
toluene/water with NaOH as base, which gave aziridine 2
in 65% isolated yield (entry 2). Attempts with milder con-
ditions, such as lower temperature, decreased amount of
NaOH or solvent change, decreased the yield (entry 3).
The use of n-BuLi in THF resulted in 50% yield (entry 4);
all other attempts were fruitless (entries 5±7). Ring-closure
of 1 with activation of the hydroxy group accordingly
yielded 2 in 63% over two steps, using the conditions stated
in entry 2.
In a second approach, 4 was mesylated to provide
compound 6, which ring-closed to yield 5 at elevated
temperature (Scheme 4).19 With 1.0 equiv. MsCl, trityl-
aziridine 5 was formed in 88% yield together with 12%
recovered 4. An excess of MsCl (1.25 equiv.) gave a
decreased yield of 5 along with byproducts.
Scheme 4. Ring-closure by in situ mesylation.
Trityl amino alcohols have been reported to react with sul-
furyl chloride to form cyclic sulfamidates, which are
converted in situto aziridines at room temperatrue.
20
2.3. Selective protection of the amino group prior to
ring-closure
When trityl amino alcohol 4 was treated with carefully
distilled surfurylchloride, sulfamidate 7 was indeed formed
but ring-closure to aziridine 5 did not take place at room
temperature. Instead 7 was recovered crude in quantitative
yield (Scheme 5). Attempted puri®cation by ¯ash chroma-
tography converted 7 to the desired aziridine 5 in 60% yield.
The conversion could instead be performed in excellent
yield by heating the reaction mixture to 708C for 1 h.
Selective protection of the amino group prior to ring-closure
is the most common way to synthesize aziridines from
Table 1. Conditions for ring-closure of 3 to 2
Entry
Solventa
Base
Yield of 2 (%)b
1
2
3
4
5
6
7
Water
NaOHc
NaOHc
NaOHc
n-BuLif
NaOEtf
KOtBuf
Et3Ng
58
76
0
50
,5
0
Acidic detritylation of 5 to N±H aziridine 2 was the most
dif®cult part of the sequence (Scheme 3). Treatment with
TFA and water as a trityl scavenger furnished aziridine 2 in
Toluene/water
THF/waterd
THFe
THF
Toluene
Toluene
,5
a
Reaction performed at re¯ux unless otherwise stated.
Isolated yield except in entry 1.
Excess.
1008C in sealed ¯ask.
2508C to room temperature.
2.5 equiv.
b
c
d
e
f
g
3 equiv.
Scheme 5. Ring-closure via cyclic sulfamidate 7.

B. Olofsson et al. / Tetrahedron 58 (2002) 5979±5982
5981
79% yield.21 Formic acid in methanol worked equally
well,22 whereas the combination of TFA and methanol
gave only decomposition products.23 The optimal reaction
temperature was 2108C; reaction at room temperature gave
2 in moderate yield whereas lower temperature effected no
reaction. Reductive detritylation using TFA/Et3SiH proved
inferior.24 Thus, the three-step transformation of amino
alcohol 1 into aziridine 2 was achieved in 77% via
Mitsunobu cyclization, in 69% via mesylation and in 76%
via cyclic sulfamidate.25
(20.0 mg, 0.074 mmol) was added THF (0.6 mL), the
mixture was cooled to 2508C before addition of n-BuLi
(1.6 M in hexanes, 46 mL, 0.074 mmol). The mixture was
allowed to reach 08C before addition of the remaining
n-BuLi (1.6 M in hexanes, 69 mL, 0.110 mmol) and stirred
at room temperature for 5 h. The solution was diluted with
CH2Cl2 and washed twice with a 2 M NaOH, the water
phase was extracted three times with CH2Cl2. The organic
phase was dried, concentrated and chromatographed
(pentane/Et2O 5:1 to Et2O, deactivated silica) to give
aziridine 2 (6.3 mg, 50%) as a colorless oil.
By detritylation of tritylaziridine 5 with TFA. Tritylaziridine
5 (85.0 mg, 0.205 mmol) was dissolved in CH2Cl2 (10 mL)
and cooled to 2108C before addition of water (100 mL).
TFA (20 mL) was added and the solution was left to reach
08C during 30 min followed by 30 min at 08C. The solution
was stirred for 10 min with 2 M NaOH (5 mL), worked
up and chromatographed (pentane/Et2O 3:2 to Et2O,
deactivated silica) to give aziridine 2 (27.8 mg, 79%) as a
colorless oil.
3. Conclusion
Direct ring-closure under Mitsunobu conditions proved
superior to other employed methods for small-scale
reactions as N±H aziridine 2 was formed in 80% yield.
This should be compared to 74% via sulfate ester 3 and
69±77% via tritylation of the amino group. For large-scale
reactions, the convenience of easy puri®cation could make
the sulfate ester route preferable. Although the same
advantage is achieved by the use of polymer-bound tri-
phenylphosphine in the Mitsunobu reaction, this reagent
does not give complete conversion for all substrates.26
By detritylation of tritylaziridine 5 with formic acid. Trityl-
aziridine 5 (63.2 mg, 0.152 mmol) was dissolved in CHCl3
(7 mL) and cooled to 2158C before addition of methanol
(100 mL). Formic acid (1 mL) in CHCl3 (1 mL) was added
and the solution was allowed to reach 258C during 2 h, then
2 M NaOH (3 mL) was added and the resulting mixture was
stirred for 10 min. The reaction was worked up and puri®ed
as above to give aziridine 2 (20.5 mg, 78%) as a colorless
oil.
4. Experimental
4.1. General experimental conditions27
Sulfuryl chloride was puri®ed by distillation, collecting the
fraction below 758C. This fraction was washed with crushed
ice and dried over phosphorous pentoxide for 2 h. It was
®nally distilled at atmospheric pressure and the colorless
fraction boiling at 698C was collected. General workup:
the reaction mixture was washed with water, the water
phase was extracted three times with Et2O. The organic
phase was dried (Na2SO4) and concentrated in vacuo.
Flash chromatography of vinylaziridines 2 and 5 was
conducted with deactivated silica, using 10% triethylamine
during packing of the column.
4.1.2. (3S,4R)-4-Amino-1-phenyl-hex-5-en-3-ol sulfate
ester (3). Amino alcohol 1 (49.8 mg, 0.260 mmol) was
dissolved in Et2O (1.5 mL) and cooled to 08C. Chlorosul-
fonic acid (21.2 mL, 0.319 mmol) was slowly added under
vigorous stirring, and a yellow precipitate was formed. The
solution was stirred for 4.5 h, concentrated in vacuo and the
residue was washed twice with ether, twice with isopropanol
and twice with Et2O. After drying under vacuum, salt 3 was
obtained as a yellow solid (66.0 mg, 97%). mp .2508C; 1H
NMR (400 MHz, d6-DMSO): d 8.04 (br s, 3H), 7.31±7.16
(m, 5H), 5.82 (m, 1H), 5.41 (m, 2H), 4.40 (ddd, 1H, Jˆ6.5,
4.5, 2.1 Hz), 3.93 (m, 1H), 2.74 (m, 1H), 2.62 (m, 1H),
1.78±1.62 (m, 2H); 13C NMR (100 MHz, d6-DMSO): d
141.7, 130.5, 128.5, 128.4, 125.9, 121.7, 75.2, 56.1, 32.9,
31.0; [a]Dˆ229.1 (c 0.79, DMF); HRMS (FAB2) Exact
mass calcd for C12H16NO4S (M2H): 270.0800. Found:
270.0794.
4.1.1. (2R,3R)-2-Phenethyl-3-vinyl-aziridine (2).13 Via
Mitsunobu cyclization. To a solution of PPh3 (0.19 mg,
0.73 mmol) in THF (2 mL) at 08C was added DIAD
(0.14 mL, 0.73 mmol). After 20 min, amino alcohol 1
(0.10 g, 0.52 mmol) in THF (1.5 mL) was added, and the
resultant mixture was re¯uxed for 17 h. The solvent was
evaporated at reduced pressure, Et2O was added to the
crude product, and the mixture was stored overnight in the
freezer. Precipitated Ph3PO was removed by ®ltration and
careful ¯ash chromatography (pentane to pentane/EtOH
10:1, deactivated silica) afforded vinylaziridine 2 in 80%
yield as a colorless oil.
4.1.3. (3S,4R)-1-Phenyl-4-(trityl-amino)-hex-5-en-3-ol (4).
Amino alcohol 1 (50 mg, 0.26 mmol) and tritylchloride
(80 mg, 0.26 mmol) were dissolved in CH2Cl2 (0.5 mL),
and cooled to 08C before addition of Et3N (73 mL,
0.52 mmol). The mixture was stirred at 08C for 15 min,
worked up and chromatographed (pentane/EtOAc 20:1 to
5:1) to give tritylated amino alcohol 4 (112 mg, 99%) as a
colorless oil. 1H NMR (400 MHz, CDCl3): d 7.52 (m, 6H),
7.27±7.12 (m, 12H), 7.02 (m, 2H) 5.75 (m, 1H), 5.10 (m,
2H), 3.11 (dd, 1H, Jˆ7.6, 2.8 Hz), 2.53±2.41 (m, 2H), 2.35
(br s, 1H), 2.27 (ddd, 1H, Jˆ13.9, 10.1, 6.5 Hz), 1.92 (d, 1H,
Jˆ5.7 Hz), 1.55±1.43 (m, 1H), 1.39±1.27 (m, 1H). 13C
NMR (100 MHz, CDCl3): d 146.7, 142.0, 136.2, 128.8,
By ring-closure of sulfate ester 3 with NaOH. To sulfate
ester 3 (66.0 mg, 0.243 mmol) was added water (0.3 mL),
toluene (0.2 mL) and NaOH (0.9 g of 33% w/w solution)
and the mixture was re¯uxed for 16 h, worked up and chro-
matographed (pentane/Et2O 5:1 to Et2O, deactivated silica)
to give aziridine 2 (27.0 mg, 76%) as a colorless oil.
By ring-closure of salt 3 with BuLi. To sulfate ester 3

5982
B. Olofsson et al. / Tetrahedron 58 (2002) 5979±5982
3. Tanner, D.; Birgersson, C.; Gogoll, A. Tetrahedron 1994, 50,
9797±9824.
128.3, 127.9, 127.8, 126.5, 125.7, 117.1, 72.5, 71.6, 59.7,
36.0, 32.4; IR (neat): 3452, 3059, 3027, 2929 cm21
;
4. Ibuka, T.; Nakai, K.; Habashita, H.; Hotta, Y.; Fujii, N.;
Nimura, N.; Miwa, Y.; Taga, T.; Yamamoto, Y. Angew.
Chem. Int. Ed. Engl. 1994, 33, 652±654.
[a]Dˆ228.9 (c 0.18, CH2Cl2); HRMS (EI1) Exact mass
calcd for C31H31NO (M): 433.2406. Found: 433.2391.
Ê
5. Somfai, P.; Ahman, J. Synthesis andRing-expansion of
4.1.4. (2R,3R)-2-Phenethyl-1-trityl-3-vinyl-aziridine (5).
Via Mitsunobu cyclization. Triphenylphosphine (42.2 mg,
0.161 mmol) was dissolved in THF (1 mL) and cooled to
08C. DIAD (31.7 mL, 0.161 mmol) was added and the solu-
tion stirred for 15 min. Trityl amino alcohol 4 (50.0 mg,
0.115 mmol) was added in THF (1 mL), the mixture was
heated at re¯ux for 16 h and concentrated in vacuo. Flash
chromatography (pentane/EtOAc 50:1 to 8:1, deactivated
silica) gave aziridine 5 (42.7 mg, 99%) as a colorless oil.
1H NMR (400 MHz, CDCl3): d 7.44 (d, 6H, Jˆ7.1 Hz),
7.25±7.12 (m, 12H), 7.02 (d, 2H, Jˆ7.1 Hz), 5.08 (m,1H),
4.70 (m, 1H), 4.50±4.37 (m, 1H), 2.70±2.53 (m, 3H), 2.06±
1.93 (m, 2H), 1.77±1.65 (m, 1H); 13C NMR (100 MHz,
CDCl3):d 146.3, 141.9, 136.1, 130.0, 128.4, 128.3, 127.3,
126.6, 125.7, 117.1, 72.7, 46.2, 40.9, 34.3, 33.7; IR (neat):
3059, 3027, 2928, 2856 cm21; [a]Dˆ190.1 (c 0.96,
CH2Cl2); HRMS (CI1) exact mass calcd for C31H30N
(M1H): 416.2378. Found: 416.2378.
Vinylaziridines; Targets in Heterocyclic Systems; Italian
Society of Chemistry, 1999; pp 341±367.
6. Aoyama, H.; Mimura, N.; Ohno, H.; Ishii, K.; Toda, A.;
Tamamura, H.; Otaka, A.; Fujii, N.; Ibuka, T. Tetrahedron
Lett. 1997, 38, 7383±7386.
Ê
7. Ahman, J.; Jarevang, T.; Somfai, P. J. Org. Chem. 1996, 61,
Ê
8148±8159.
È
8. Lindstrom, U. M.; Somfai, P. Chem. Eur. J. 2001, 7, 94±98.
9. Patai, S. The Chemistry of Alkenes; Vol. 1; Wiley: New York,
1964.
10. Osborn, H. M. I.; Sweeney, J. Tetrahedron: Asymmetry 1997,
8, 1693±1715.
11. Hughes, D. L. Org. Prep. Proc. Int. 1996, 28, 127±164.
12. Mitsunobu, O. Synthesis 1981, 1±28.
È
13. Lindstrom, U. M.; Somfai, P. Synthesis 1998, 109±117.
14. The puri®cation problem is similar for all tested substrates to
date.
15. Olofsson, B.; Khamrai, U.; Somfai, P. Org. Lett. 2000, 2,
4087±4089.
Via mesylate 6. Trityl amino alcohol 4 (50.1 mg, 0.115
mmol) was dissolved in THF (1 mL). Et3N (48 mL,
0.346 mmol) and mesyl chloride (9.0 mL, 0.117 mmol)
were added and the solution was stirred at room temperature
for 1 h, followed by heating at re¯ux for 36 h. The mixture
was diluted with CH2Cl2 and washed twice with sat. aq.
NaHCO3 solution, dried, concentrated and chromato-
graphed (pentane/EtOAc 50:1 to 20:1, deactivated silica)
to give aziridine 5 (42.7 mg, 88%) as a colorless oil. Starting
material 4 was recovered in 12% yield.
16. Stogryn, E. L.; Brois, S. J. J. Am. Chem. Soc. 1967, 89, 605±
609.
17. Greene, T. W.; Wuts, P. G. M. Protective Groups in Organic
Synthesis; 3rd ed; Wiley: New York, 1999.
18. Evans, P. A.; Holmes, A. B.; Russell, K. J. Chem. Soc., Perkin
Trans. 1 1994, 3397±3409.
19. Willems, J. G. H.; Hersmis, M. C.; de Gelder, R.; Smits,
J. M. M.; Hammink, J. B.; Dommerholt, F. J.; Thijs, L.;
Zwanenburg, B. J. Chem. Soc., Perkin Trans. 1 1997, 963±
967.
Via cyclic sulfamidate 7. Trityl amino alcohol 4 (137.6 mg,
0.318 mmol) was dissolved in toluene (6 mL). Et3N
(133 mL, 0.954 mmol) was added and the mixture cooled
to 2508C. Freshly distilled sulfurylchloride (32 mL,
0.398 mmol) was slowly added as a solution in toluene
(1.5 mL) after which the solution became white. After 1 h
stirring, the temperature was raised to room temperature, the
¯ask was sealed and the solution was heated to 708C for 1 h,
at which time the color had become yellow. Workup and
puri®cation as described above gave aziridine 5 (127.4 mg,
97%) as a colorless oil.
20. Kuyl-Yeheskiely, E.; Lodder, M.; van der Marel, G. A.; van
Boom, J. H. Tetrahedron Lett. 1992, 33, 3013±3016.
21. Alsina, J.; Giralt, E.; Albericio, F. Tetrahedron Lett. 1996, 37,
4195±4198.
22. Bosche, U.; Nubbemeyer, U. Tetrahedron 1999, 55, 6883±
6904.
23. Church, N. J.; Young, D. W. J. Chem. Soc., Perkin Trans. 1
1998, 1475±1482.
24. Vedejs, E.; Klapars, A.; Warner, D. L.; Weiss, A. H. J. Org.
Chem. 2001, 66, 7542±7546.
25. In a similar reaction sequence, the trityl group was replaced by
the 2,4-dinitrobenzenesulfonyl group (Ns0), which proved
unsuccessful.28,29 Although both the protection and the ring-
closure were fast reactions, the yields were poorer than in
the tritylation sequence. Furthermore, deprotection of the
Ns0-aziridine to 2 was unsuccessful, instead affording the
corresponding ring-opened diamine.
Acknowledgements
This work was supported ®nancially by the Swedish Natural
Science Research Council and the Royal Institute of
Technology.
26. Olofsson, B.; Somfai, P. Submitted for publication.
Ê
27. For general experimental procedures, see: Ahman, J.; Somfai,
P. Tetrahedron 1995, 51, 9747±9756.
28. Fukuyama, T.; Cheung, M.; Jow, C. K.; Hidai, Y.; Kan, T.
Tetrahedron Lett. 1997, 38, 5831±5834.
References
Á
29. Farras, J.; Ginesta, X.; Sutton, P. W.; Taltavull, J.; Egeler, F.;
1. McCoull, W.; Davis, F. A. Synthesis 2000, 1347±1365.
2. Rai, K. M. L.; Hassner, A. Adv Strain Interest Org Mol 2000,
8, 187±257.
Â
Romea, P.; Urpõ, F.; Vilarrasa, J. Tetrahedron 2001, 57, 7665±
7674.