Amination of arenes through electron-deficient reaction cascades of aryl epoxyazides

Article abstract of DOI:10.1021/ol035319d

(Matrix presented) Aryl epoxyazides undergo efficient electron-deficient reaction cascades mediated by Lewis acids, leading to regiospecific amination of the aromatic ring.

Full text of DOI:10.1021/ol035319d

ORGANIC
LETTERS
2003
Vol. 5, No. 20
3655-3658
Amination of Arenes through
Electron-Deficient Reaction Cascades of
Aryl Epoxyazides
Stuart Lang,† Alan R. Kennedy,† John A. Murphy,*,† and Andrew H. Payne‡
Department of Pure and Applied Chemistry, UniVersity of Strathclyde,
295 Cathedral Street, Glasgow G1 1XL, U.K., and GlaxoSmithKline Pharmaceuticals,
New Frontiers Science Park, Third AVenue, Harlow, Essex CM 19 5AW, U.K.
john.murphy@strath.ac.uk
Received July 17, 2003
ABSTRACT
Aryl epoxyazides undergo efficient electron-deficient reaction cascades mediated by Lewis acids, leading to regiospecific amination of the
aromatic ring.
Sequencing organic reactions in a single-pot process is a
powerful way of designing complex molecular architectures
from relatively simple starting materials. This is beautifully
exemplified in the use of cascade reactions in free radical1
chemistry and organopalladium chemistry2 for organic
synthesis. Our background in cascade reactions3 includes the
TTF-mediated radical-polar crossover sequence3b and nitra-
tion sequences that highlight cationic chemistry,4 and this
focused our attention on the fact that reaction cascades
involving cations or incipient cations are less well explored.
Cascade reactions of cations are well-known in nature in the
biogenetic [and biomimetic] synthesis of steroids and ter-
penes;5 however, these involve cationic cascades on hydro-
carbon skeletons, and heteroatoms are very poorly repre-
sented. Recent elegant developments of the Schmidt reaction,
particularly by the research groups of Aube´6 and Pearson,7
have shown how useful azide groups can be in intercepting
cations and incipient cations. Our aim in this program is to
bring together Schmidt-type azide interceptions with other
types of electron-deficient rearrangement-epoxide-opening,
aryl, alkyl, and hydride migrations, as well as fragmentations,
to test for compatibility and to come to an understanding of
the ground-rules for sequencing the reactions. The first
epoxide-based Schmidt reactions were very recently reported
by Baskaran and co-workers.8 The very different types of
(5) For a recent example, see: Seeman, M.; Zhai, G.; de Kraker, J.-W.;
Paschall, C. M.; Christianson, D. W.; Cane, D. E. J. Am. Chem. Soc. 2002,
124, 7681.
(6) (a) Gracias, V.; Milligan, G. L.; Aube´, J. J. Am. Chem. Soc. 1995,
117, 8047. (b) Milligan, G. L.; Mossmann, C. J.; Aube´, J. J. Am. Chem.
Soc. 1995, 117, 10449. (c) Wendt, J. A.; Aube´, J. Tetrahedron Lett. 1996
37, 1531. (d) Forsee, J. E.; Aube´, J. J. Org. Chem. 1999, 64, 4381. (e)
Iyengar, R.; Schildknegt, K.; Aube´, J. Org. Lett. 2000, 2, 1625. (f) Desai,
P.; Schildknegt, K.; Agrios, K. A. Mossman, C.; Milligan, G. L.; Aube´, J.
J. Am. Chem. Soc. 2000, 122, 7226.
(7) (a) Pearson, W. H.; Schkeryantz, J. M. Tetrahedron Lett. 1992, 33,
5291. (b) Pearson, W. H.; Walavalkar, R.; Schkeryantz, J. M.; Fang, W.-
K.; Blickensdorf, J. D. J Am. Chem. Soc. 1993, 115, 10183. (c) Pearson,
W. H.; Fang, W.-K. J. Org. Chem. 1995, 60, 4960. (d) Pearson, W. H.;
Gallagher, B. M. Tetrahedron 1996, 52, 12039. (e) Pearson, W. H. J.
Heterocycl. Chem. 1996, 33, 1489. (f) Pearson, W. H.; Fang, W.-K. J. Org.
Chem. 2000, 65, 7158. Pearson, W. H.; Fang, W.-K. J. J. Org. Chem. 2001,
66, 6838. (g) Pearson, W. H.; Hutta, D. A.; Fang, W.-K. J. Org. Chem.
2000, 65, 8326. (h) Pearson, W. H.; Walavalkar, R. Tetrahedron 2001, 57,
5081.
† University of Strathclyde.
‡ GlaxoSmithKline Pharmaceuticals.
(1) For a review, see: Dhimane, A.-L.; Fensterbank, L.; Malacria, M.
Radicals in Organic Synthesis; Renaud, P., Sibi, M., Eds.; Wiley-VCH:
Weinheim, 2001; Vol. 2, Chapter 4.4.
(2) Grigg, R.; Sridharan, V. J. Organomet. Chem. 1999, 576, 65.
(3) (a) Lizos, D. E.; Murphy, J. A. Org. Biomol. Chem. 2003, 1, 117
and references therein. (b) Murphy, J. A. Radicals in Organic Synthesis;
Renaud, P., Sibi, M., Eds.; Wiley-VCH: Weinheim, 2001, Vol. 1; Chapter
2.7.
(4) (a) Murphy, J. A.; Hewlins, S. A.; Lin, J. Tetrahedron Lett. 1995,
36, 3039. (b) Murphy, J. A.; Hewlins, S. A.; Lin, J.; Hibbs, D. E.;
Hursthouse, M. B. J. Chem. Soc., Perkin Trans. 1 1997, 1559.
10.1021/ol035319d CCC: $25.00
© 2003 American Chemical Society
Published on Web 09/13/2003

Scheme 1a
a Reagents and conditions: (i) Hydrochloric acid (concentrated), 5 min, rt. (ii) NaN3 (5 equiv), DMF, 60 °C, 18 h. (iii) mCPBA (2
equiv), DCM, 16h, 0 °C to rt. (iv) p-TsOH (1 equiv), toluene, ∆, 64 h. (v) LiAlH4 (1 equiv), THF, 0 °C to rt, 2 h, 82%. (vi) DPPA (1.2
equiv), PPh3 (1.2 equiv), DIAD (1.2 equiv), THF, 0 °C to rt, 24 h, 78%, then mCPBA (2 equiv), DCM 16 h, 0 °C, 4 h, 66%.
chemistry seen in their paper compared to this one clearly
illustrate the extensive scope of this area of chemistry for
molecular design.
corresponding amine 11. The alternate process b, would
result in a rare example of electrophilic amination of the
arene.10,11 This route comprises an electron-deficient cascade
featuring epoxide-opening and 1,2-shift of the aryl ring to
give 13 followed by reduction to 14. Either route would
afford attractive applications to synthesis, providing that the
outcome could be tuned so that only one of the processes
predominated and, hence, that complex mixtures were not
formed.
Our plans bring together, in one molecule, three reactive
entities (an arene, an epoxide, and an azide), and so substrates
4a-d and 8 were prepared as shown in Scheme 1. We
predicted that opening of these epoxides in the presence of
Lewis acids would lead to attack by the azide group at the
benzylic position to give a five-membered ring; attack at the
alternate epoxide carbon would lead to a strained four-
membered ring. Compounds 4a,b,d were prepared from the
commercially available cyclopropanes 1a,b,d, while 4c was
prepared from 1c; in turn, this was prepared by reduction of
the corresponding commercially available ketone. Com-
pounds 5-7 were prepared by the route of Ferraz and Silva.9
Turning to the simplest substrate, 4a, we gave some thought
in advance to the possible outcomes of the experiment before
choosing the conditions for the workup of the reaction. Lewis
acid-induced opening of the epoxide should lead to attack
by the azide at the benzylic position to form 9 (Scheme 2),
and then two potentially competing processes come into
play: (a) ring-fragmentation driven by the Lewis acid-linked
alkoxide 9 or (b) aryl-group migration via aziridine 12.
Looking at process a, the imine 10 would be rather sensitive;
therefore, to facilitate isolation, a reduction step was
introduced in situ to trap this compound, if it formed, as the
Epoxyazide 4a was treated with stannic chloride (1.5
equiv) in DCM at 0 °C for 5 min followed by reduction
with NaBH4 (3 equiv) in methanol at 0 °C, affording a single
1
product, 14, cleanly in 88% yield. The H NMR spectrum
of the product showed three upfield protons in the aromatic
region (2H, δ 6.6, and 1H, δ 6.7), suggesting that they
occupied ortho and para positions, respectively, on an aniline
ring. Although the compound was crystalline, the structure
of this solid compound could not be confirmed by X-ray
crystallography due to the poor quality of the crystals.
Moreover, the compound appeared to be rather unstable on
storage.
However, in efforts to transform alcohol 14 into other
compounds, the crystalline and surprisingly stable mesylate
15 was prepared, which on X-ray structure analysis con-
firmed the structure.
Scheme 2
3656
Org. Lett., Vol. 5, No. 20, 2003

Scheme 3
A question arises over the origin of alcohol 14; it could
94% yield of the pair of diastereomers 19 (51 and 43% yields
of the separate isomers) (Scheme 3).
arise from direct reduction of the iminium salt 13 or from
reduction of either the adduct 16 (where X- comes from the
Lewis acid) or ketone 17. We also considered that an
epoxide, 18, might be the intermediate, resulting from ring-
closure of 13, although this compound should be very labile.
Detailed information on the nature of the intermediates was
obtained by conducting the reaction without the borohydride
reduction step. Workup established that the ketone, 17, was
definitely not an intermediate. Rather, the isolated product
was a 3:1 mixture of diastereomers of the diol 16 (X ) OH)
with N-CH-O doublets at δ 4.99 (major, J 5.1 Hz) and
5.32 (minor, J 5.1 Hz) (1H) and a CH2-CH-O multiplet at
δ 4.35 (1H).
It is clear from the above examples that neighboring-group
participation by the aryl group occurs very readily; the extent
of participation should depend on the electron-richness of
the arene. The p-fluoro derivative 4c was prepared to see if
the fluorine substituent would inhibit this reaction. However,
the corresponding pyrrolidine mesylate 20 was formed as
the exclusive product (64% yield) (see Scheme 3). Hence,
this substrate behaved exactly as its analogue, 4a, and a
fluorine atom had no apparent effect in diverting the course
of the reaction. Thus, electrophilic amination of the arene
results even with an electron-withdrawing group in the para
position.
The related aryl epoxy-azide substrate 4b was next
subjected to the rearrangement in the presence of stannic
chloride. This reaction proceeded most efficiently in THF
as the solvent. Reduction and mesylation gave a remarkable
Anomalous results occurred, however, when the epoxide
4d reacted in DCM; two products were furnished in a 1.6:1
ratio, namely, the acyclic ketoazide 22 and the pyrrolidinone
25. Compound 22 could arise by 1,2-hydride migration in
cation 21 or by proton loss from 21 to give a stannylenolate
that breaks down to ketone 22 on workup. Similarly, a ketone
25 was formed from the cyclized intermediate 24. The
appearance of two products 22 and 25 on workup led us to
investigate whether changing the solvent might alter the ratio
in the rearrangement reaction. Happily, in THF as the solvent,
essentially only the 3-ketoypyrrolidone 25 is formed (48%),
while in toluene, the acyclic ketone 22 is the dominant
(8) Reddy, P. G.; Varghese, B.; Baskaran, S. Org. Lett. 2003, 5, 583.
(9) Ferraz, H. M. C.; Silva, L. F., Jr. Tetrahedron 2001, 57, 9939.
(10) This complements recent developments in Ar-N bond formation
in palladium-based arylation of amines: (a) Muci, A. R.; Buchwald, S. L.
Top. Curr. Chem. 2002, 219, 131. (b) Littke, A. F.; Fu, G. C. Angew. Chem.,
Int. Ed. 2002, 41, 4176. (c) Stambuli, J. P.; Kuwano, R.; Hartwig, J. F.
Angew. Chem., Int. Ed. 2002, 41, 4746.
(11) For other reactions of azides that lead to migration of aryl groups
to nitrogen, see ref 6f and: Wrobleski, A.; Aube´, J. J. Org. Chem. 2001,
66, 886.
Org. Lett., Vol. 5, No. 20, 2003
3657

product (83%). Although the isolated yield of 25 is not very
high, this reflects difficulties in handling the compound; a
notable feature of the ketone 25 was its instability in air.12
All of the examples so far presented feature epoxides with
C-O bonds to a (secondary or tertiary) benzylic carbon and
to an aliphatic (secondary) carbon. The attack of azide was
seen exclusively at the benzylic site. In substrate 8, a
secondary benzylic carbon competes with a tertiary aliphatic
carbon for attack by the azide group. Reaction with SnCl4
led to tricyclic amine 29 in 74% yield, and no other product
was isolated, implying that the benzylic carbon is still the
site of interception by the azide group. Despite the strain
inherent in 27, neighboring group participation by the aryl
ring still occurs. Cleavage of the aziridine gives the bridged
intermediate 28. Unlike in previous examples, hydride shift/
deprotonation is not possible here; so instead, a regiospecific
ring-contraction occurs with assistance from the oxygen atom
to afford ketone 29. This substrate was then used to analyze
the effect of changing the acid used for the reaction.
As seen in Table 1, the reaction works fairly well for all
of the Lewis acids except TiCl4, which led to decomposition.
Importantly, protic acid also works fairly well, but the best
conditions (94% yield) are seen when BF3‚OEt2 is used.
The structure of the final product was confirmed by in
situ conversion of the product ketone to its dithiane, 30,
which was subjected to crystal structure determination.
In summary, in the presence of acid stimulus, aryl
epoxyazides afford high yields of products deriving from
Table 1. Effect of Variation of Acid on Yield of 29
entry
conditions
yield 29 (%)
1
2
3
4
5
SnCl4, DCM, 0 °C, 40 s
BF3‚OEt2, DCM, 0 °C, 3 min
BF3‚OEt2, DCM, rt, 2 min
TiCl4, DCM, 0 °C, 10 s
TfOH, DCM, 0 °C, 5 s
74
92
94
0
53
cascades of electron-deficient reactions such as epoxide-
opening, electrophilic amination of an arene, and hydride or
alkyl shifts. The reactions are generally highly efficient and
can be tuned by varying the acid used or the reaction solvent.
We are currently applying these reactions to the synthesis
of more complex natural compounds.
Acknowledgment. We thank the EPSRC and Glaxo-
SmithKline for a studentship to S.L. and EPSRC National
Mass Spectrometry Service, Swansea, for high-resolution
mass spectra.
Supporting Information Available: X-ray crystal struc-
tures for 15 and 30, as well as spectroscopic data for
compounds 2-8, 14, 15, 19, 20, 22, 25, 29, and 30. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(12) Leonard, N. J.; Fuller, G.; Dryden, H. L., Jr. J. Am. Chem. Soc.
1953, 75, 3727.
OL035319D
3658
Org. Lett., Vol. 5, No. 20, 2003