Bronsted acid-promoted olefin aziridination and formal anti-aminohydroxylation

Article abstract of DOI:10.1021/ja045608c

A straightforward synthesis of aziridines is reported from an electron-rich azide (alkyl or aryl azide), electron-deficient olefin, and triflic acid in cold acetonitrile. The only coproduct of the reaction is dinitrogen (N₂). Active ester substrates bearing a nucleophilic carbonyl engage the putative protonated aziridine intermediate to produce the product of olefin aminohydroxylation in which the nitrogen is benzyl protected and the oxygen is acylated. The possibility that a triazoline need not be an intermediate in aziridine production is advanced. Copyright

Full text of DOI:10.1021/ja045608c

Published on Web 01/11/2005
Brønsted Acid-Promoted Olefin Aziridination and Formal
anti-Aminohydroxylation
Joseph M. Mahoney, Colin R. Smith, and Jeffrey N. Johnston*
Department of Chemistry, Indiana UniVersity, 800 East Kirkwood AVenue, Bloomington, Indiana 47405
Received July 21, 2004; E-mail: jnjohnst@indiana.edu
Table 1. Bronsted Acid-Promoted Azirdinations: Azide Scopea
The stereocontrolled functionalization of olefins with heteroatoms
provides direct entry to functionalized carbon chains used as small-
molecule feedstocks in organic synthesis. For example, olefin aziri-
dination and subsequent ring-opening reactions often form the basis
for more complex target synthesis.1 Aminohydroxylation of unsatu-
rated esters provides entry to both R- and â-amino acid derivatives.2
entry
R1
% yield
We envisioned the design of an acid-promoted nucleophilic azide
addition to an electrophilic enone. The use of acid would catalyze
both cycloaddition and nitrogen extrusion steps and would be driven
by the evolution of N2 as a relatively benign coproduct.3 This
approach favors the use of an electron-rich alkyl azide as a
nucleophilic nitrene equivalent,4,5 a behavior synthetically developed
in recent years largely by the groups of Aube´6 (carbonyl electro-
philes) and Pearson7 (carbenium ion electrophiles). Extensive studies
of the [3+2] cycloaddition chemistry of azides with alkenes have
led to protocols for thermal8,9 and photochemical10 decomposition
of the resulting triazolines.11 Triazoline decomposition by acid also
leads to aziridine, but the overall efficiency is typically low and
the scope narrow at present.12 Azide cycloadditions and addition/
rearrangement reactions are more general in an intramolecular
setting (cyclizations) since azides decompose in various ways upon
binding to Lewis acids and transition metals.4
We report here a new Brønsted acid-promoted addition of azides
to activated olefins that delivers the corresponding aziridines in
good yield under relatively mild, non-redox conditions (eq 1).
Furthermore, active ester derivatives bearing a Lewis basic oxygen
readily participate in a tandem process to deliver both R- and
â-substituted threo-R-hydroxy â-amino acids in orthogonally
protected form (eq 2).
1c
2
3
4
5
6
7
8
Bn
Bn
2a
2a
2b
2c
2d
2e
2f
79
92d
88
93
43e
66
76
68
Ph2CH
Ad
p-MeOC6H4
tBuO2CCH2
MeOCCHdCHCH2
Me2CdCHCH2
2g
a All reactions were 0.30 M in substrate and proceeded to complete
conversion. b Isolated yield after chromatography. c The thermal reaction
(no TfOH) produces 4a and 4b in 66% and quantitative yield, respectively.
d Yield of the crystalline 2a triflic acid salt. e Reaction run in CH2Cl2 at
78 °C to minimize azide decomposition.
acid screen was executed in a generally nonparticipating solvent
with low polarity (dichloromethane), as well as in a solvent that is
both a competent Lewis base and polar (acetonitrile).14 The unique
effectiveness of trimethylsilyl triflate and triflic acid, a representative
Brønsted acid, in the transformation emerged. The higher yields
and ease of product purification led to exclusive use of triflic acid
in subsequent experiments. Acetonitrile was identified from a broad
solvent screen as the most effective medium for the aziridination
(CH3CN, 79%; Et2O, 58%; CH2Cl2, 49%; THF, 35%; CHCl3, 33%;
CH3OH and DMF, 0%).
Electron-rich alkyl azides are generally effective donors (Table
1). Benzyl azide engages the olefin to produce the N-benzyl-
protected terminal aziridine in 79% isolated yield after flash
chromatography (Table 1, entry 1). Alternatively, the aziridine may
be isolated in slightly higher yield (92%) by direct crystallization
of the triflic acid salt (Table 1, entry 2). More hindered azides are
also effective, including diphenylmethyl azide (Table 1, entry 3)
and adamantyl azide (Table 1, entry 4). Electron-rich azides derived
from aniline produce the desired aziridine but competitively form
what appears to be the azide dimer or trimer (Table 1, entry 5).
Azides such as tert-butyl glycinyl azide can be smoothly converted
to the aziridine without identifiable ester hydrolysis (Table 1, entry
6). Finally, a variety of allylic azides formed their derived aziridines
in good yield (Table 1, entries 7 and 8).
When benzyl azide and methyl vinyl ketone (1a) are stirred at
room temperature, the 2:1 ketone-azide adduct 4a is formed cleanly
in 66% yield (Table 1).9 Subsequent treatment of 4a with triflic
acid provides only the derived salt. That a Brønsted acid might be
effective, if not uniquely so, in the desired aziridination was based
on the predictions that (1) protonation of the azide would be
reversible, perhaps even disfavored, (2) decomposition of the azide
with solvent-bound proton would be relatively slow,13 (3) the
presence of a protic acid might favor an intermediate aminodia-
zonium ion (A, Vide infra) over the concerted [3+2] cycloaddition
to produce triazoline, and (4) collapse of an intermediate enol to
triazoline would be slow relative to aziridine formation. A Lewis
The ease with which the aziridine triflic acid salt 2a‚HOTf could
be formed and isolated led us to consider the possibility that an
oxygen nucleophile might be internally delivered by an activated
ester heteroatom. The acrylate derivative of N-benzyl methyl
9
1354
J. AM. CHEM. SOC. 2005, 127, 1354-1355
10.1021/ja045608c CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
Table 2. Regiospecific, Stereoselective Formal
We speculate that the triazoline need not be an intermediate and
that rate acceleration and selective formation of the aziridine
intermediate might also be explained by TS. This proposal not only
departs from conventional wisdom but also presents significant
implications for the design of new reactions based on azides, as
well as their stereoselective variants.16 We do not rule out triazoline
intermediacy in all cases, since Brønsted acid clearly promotes the
conversion of triazoline B to aminodiazonium ion A.12
In summary, the addition of electron-rich azides to electron-
deficient olefins is promoted by Brønsted acids. Our present findings
suggest that catalysis is achieved by either providing access to TS
or promoting the formation of the aminodiazonium ion intermediate
A by either direct conjugate addition or triazoline fragmentation.
The aminohydroxylation variant is stereocomplementary to tradi-
tional metal-mediated aminohydroxylations that provide syn-amino
alcohols from E-olefins.
anti-Aminohydroxylation of Activated Olefinsa
entry
R1
R2
drb
% yieldc
1
2
3
4
5
6
7
8
9
H
H
3a
3b
3c
3d
3e
3f
3g
3h
3i
84
88
61
60
46
86
94
82
61
H
Me
Et
H
H
nPr
Bn
Ph
H
H
H
Me
Et
iPr
>20:1
15:1
>20:1
H
H
a All reactions were 0.25 M in substrate and proceeded to complete
conversion. b Diastereomeric ratio determined by 1H NMR spectroscopy.
Relative configuration of 3g assigned by X-ray analysis of a derivative.
c Isolated yield after chromatography.
Acknowledgment. Financial support was provided by the NSF
(CHE-0415811), Indiana University, Boehringer-Ingelheim, Amgen,
Yamanouchi, and Eli Lilly. We thank John Huffman (IU) for X-ray
support.
Supporting Information Available: General experimental proce-
dures and analytical data for all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
Scheme 1. Competition Experiments for Catalyzed
Aminohydroxylation and Thermal [3+2] Cycloaddition
References
(1) (a) Sweeney, J. B. Chem. Soc. ReV. 2002, 31, 247-258. (b) Graham, M.
A.; Wadsworth, A. H.; Thornton-Pett, M.; Rayner, C. M. Chem. Commun.
2001, 966-967.
(2) (a) Andersson, M. A.; Epple, R.; Fokin, V. V.; Sharpless, K. B. Angew.
Chem., Int. Ed. 2002, 41, 472. (b) Juaristi, E.; Lopez-Ruiz, H. Curr. Med.
Chem. 1999, 6, 983-1004.
(3) [3+2] azide/olefin cycloadditions exhibit large, negative activation
entropies: (a) Scheiner, P.; Schomaker, J. H.; Deming, S.; Libbey, W. J.;
Nowack, G. P. J. Am. Chem. Soc. 1965, 87, 306. High pressure has been
shown to accelerate triazoline formation: (b) Anderson, G. T.; Henry, J.
R.; Weinreb, S. M. J. Org. Chem. 1991, 56, 6946.
(4) Fioravanti, S.; Morreale, A.; Pellacani, L.; Tardella, P. A. Synthesis 2001,
1975.
(5) (a) L’Abbe´, G. Chem. ReV. 1969, 69, 345. (b) Scriven, E. F. V.; Turnbull,
K. Chem. ReV. 1988, 88, 297. (c) 1,3-Dipolar Cycloaddition Chemistry;
Padwa, A., Ed.; Wiley: New York, 1984.
(6) Milligan, G. L.; Mossman, C. J.; Aube´, J. J. Am. Chem. Soc. 1995, 117,
10449.
(7) For the alkylation of azides by carbenium ions often formed by Lewis or
Brønsted acid, see: (a) Pearson, W. H. J. Heterocycl. Chem. 1996, 33,
1489. (b) Pearson, W. H.; Walavalkar, R.; Schkeryantz, J. M.; Fang, W.;
Blickensdorf, J. D. J. Am. Chem. Soc. 1993, 115, 10183. (c) Pearson, W.
H.; Schkeryantz, J. M. Tetrahedron Lett. 1992, 33, 5291.
(8) For an excellent example in which azide-olefin [3+2] cycloaddition
provides a triazoline that thermally forms aziridinesbut via diazo insertion
to an iminessee: Sha, C.-K.; Ouyang, S.-L.; Hsieh, D.-Y.; Chang, R.-
C.; Chang, S.-C. J. Org. Chem. 1986, 51, 1490.
(9) (a) Huisgen, R.; Szeimies, G.; Mobius, L. Chem. Ber. 1966, 99, 475. (b)
Broeckx, W.; Overbergh, N.; Samyn, C.; Smets, G.; L’Abbe´, G.
Tetrahedron 1971, 27, 3527.
(10) Leading references: (a) Dahl, R. S.; Finney, N. S. J. Am. Chem. Soc.
2004, 126, 8356-8357. (b) Allemann, S.; Vogel, P. Synthesis 1991, 923-
928.
(11) For additional examples of triazoline decomposition pathways, see: (a)
Wohl, R. A. J. Org. Chem. 1973, 38, 3862 (ring expansions). (b)
Goldsmith, D. J.; Soria, J. J. Tetrahedron Lett. 1986, 39, 4701. (c) Hermes,
M. E.; Marsh, F. D. J. Org. Chem. 1972, 37, 2969.
(12) Acid-induced decomposition of triazolines via N1-N2 cleavage; leading
references: (a) Smith, R. H., Jr.; Wladkowski, B. D.; Taylor, J. E.;
Thompson, E. J.; Pruski, B.; Klose, J. R.; Andrews, A. W.; Michejda, C.
J. J. Org. Chem. 1993, 58, 2097-2103. (b) Hu¨nig, S.; Schmitt, M. Liebigs
Ann. 1996, 559-573.
(13) However, acid-promoted decomposition of benzyl azide to the corre-
sponding Mannich base can be efficient: Wrobleski, A.; Aube, J. J. Org.
Chem. 2001, 66, 886.
carbamate (1a) was prepared and subjected to the standard
aziridination conditions to deliver oxazolidine dione 3a as a single
regioisomer (>20:1, 1H NMR). Use of R-substituted acrylates led
to generally high yields of the expected oxazolidine diones 3b-f
(Table 2, entries 2-6).15 Of these substrates, the R-benzyl-
substituted acrylate was noticeably more sluggish and required
warming to achieve complete conversion (Table 2, entry 5). A series
of â-substituted imides were then subjected to the aminohydroxy-
lation conditions with very similar results. Crotonyl imide 1g formed
the desired â-amino R-acyloxy acid in protected form in 94% yield
(Table 2, entry 7). Spectroscopic analysis of the crude reaction
mixture (1H NMR) suggested that a single regio- and stereoisomer
was formed in this transformation. â-Substituents with increasing
size were generally tolerated (Table 2, entries 7-9).
Throughout the course of these studies, substantial effort was
applied toward the identification and isolation of the potential
triazoline intermediate (e.g., B in Scheme 1), since the literature
implicates it without exception as the intermediate to aziridine. It
was not until imide 1g was stirred in neat benzyl azide that the
triazoline could be formed, isolated, and characterized. This
triazoline failed to thermally convert to aziridine or oxazolidinedi-
one, but its exposure to triflic acid led to clean low-temperature
(-20 °C) conversion to 3g. However, a competition experiment
between 1b and 1g revealed a pronounced substituent effect in the
cycloaddition not observed in the aminohydroxylation (Scheme 1).
(14) As expected, most Lewis acids resulted only in unchanged starting
materials, regardless of the hard/soft character of the Lewis acid.
(15) For the hydrolysis of oxazolidine diones, see: (a) Shapiro, S. L.; Rose, I.
M.; Testa, F. C.; Roskin, E.; Freedman, L. J. Am. Chem. Soc. 1959, 81,
6498. (b) Finkbeiner, H. J. Am. Chem. Soc. 1965, 87, 4588.
(16) Interestingly, cyclic enones are transformed to their ring-contracted
vinylogous amides under similar conditions: Reddy, D. S.; Judd, W. R.;
Aube´, J. Org. Lett. 2003, 5, 3899.
JA045608C
9
J. AM. CHEM. SOC. VOL. 127, NO. 5, 2005 1355