Synthesis and reactions of cyclopentadiene monoaziridine: A concise approach to the core of agelastatin A

Article abstract of DOI:10.1016/S0040-4039(01)02238-9

An improved protocol for the preparation of cyclopentadiene monoaziridine is described (88% yield). The utility of cyclopentadiene monoaziridine is demonstrated by its use in (i) the shortest synthetic entry into 4-amino substituted cyclopentenes and (ii)

Full text of DOI:10.1016/S0040-4039(01)02238-9

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 723–726
Synthesis and reactions of cyclopentadiene monoaziridine:
a concise approach to the core of agelastatin A
Elise Baron, Peter O’Brien* and Timothy D. Towers
Department of Chemistry, University of York, Heslington, York YO10 5DD, UK
Received 15 October 2001; revised 16 November 2001; accepted 22 November 2001
Abstract—An improved protocol for the preparation of cyclopentadiene monoaziridine is described (88% yield). The utility of
cyclopentadiene monoaziridine is demonstrated by its use in (i) the shortest synthetic entry into 4-amino substituted cyclopentenes
and (ii) the preparation of two key intermediates in a concise synthetic approach to the cyclopentane core of agelastatin A. The
agelastatin A model studies make use of a lithium amide-mediated epoxide to allylic alcohol rearrangement reaction. © 2002
Elsevier Science Ltd. All rights reserved.
Acyclic and cyclic 2-vinyl aziridines are useful interme-
1
diates in synthesis. For example, Hudlicky and co-
N
workers have demonstrated the synthetic potential of a
N-p-toluenesulfonyl aziridine derived from a chiral
cyclohexadiene culminating in the total synthesis of
Ts
1
H
1. NH3
2. Boc2O
2
pancratistatin. Such cyclic vinyl aziridines have been
successfully ring opened at the activated allylic position
by a range of nucleophiles including carbon-based
5,6
NHBoc
3
4
nucleophiles, alcohols and, very recently, amines.
We have also previously developed a chiral base route
to a six-ring vinyl aziridine. In addition, a powerful
NHTs
2
NHTs
3
7
and direct entry into cyclopentene-based vinyl aziridi-
nes, utilising irradiation of pyridinium salts, has been
8
developed by Mariano et al. The photochemical step is
In order to prepare two useful synthetic intermediates
cyclopentenes 2 and 3), we wished to study the ring
followed by aziridine ring opening with water, alcohols
and thiols and has recently been applied to the total
(
opening of aziridine 1 with hydride- and amine-based
nucleophiles. Reaction of aziridine 1 with a suitable
source of hydride should furnish sulfonamido cyclopen-
tene 2, considerably improving the usual synthetic entry
9
10
synthesis of mannostatin A and (−)-allosamizoline.
From a synthetic viewpoint, we became interested in
the preparation and subsequent reactivity of the cyclic
14
into 4-amino substituted cyclopentenes. Alternatively,
reaction of aziridine 1 with ammonia followed by Boc
protection should generate diamine 3 which contains
the trans-diamine stereochemistry and appropriate
functionality for conversion into the cyclopentane core
2-vinyl aziridine 1 derived from cyclopentadiene.
Knight and Muldowney have already described a route
to aziridine 1 via the Evans-type direct monoaziridina-
tion of cyclopentadiene using [N-(p-toluenesul-
1
1
fonyl)imino]iodinane (PhIꢀNTs) and Cu(acac)2.
15
1
2
of agelastatin A. Herein we describe our results.
Atkinson and Meades have recently reported a related
mono-aziridination using 3-acetoxyaminoquinazolin-
4(3H)-ones and a three-step route for the synthesis of
Following Knight and Muldowney’s procedure for
11
the benzamide-protected version of 1 has also been
monoaziridination of dienes, cyclopentadiene was
13
described by Olivo and co-workers.
reacted with PhIꢀNTs and 10 mol% of Cu(acac) to
2
1
1
give aziridine 1 in variable yields of 30–45% (lit.,
6
0%). Our optimised procedure for the preparation of 1
Keywords: aziridines; epoxides; rearrangement; allylic alcohols;
cyclopentanes.
involved starting the reaction at 0°C (as the reaction is
initially quite exothermic), use of an excess of diene (2
*
Corresponding author.
0
040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)02238-9

7
24
E. Baron et al. / Tetrahedron Letters 43 (2002) 723–726
1
. PhI=NTs, Cu(acac)2
MeCN, 0 °C, 15 min
LiEt3BH, THF
0 °C, 1 h
2
. rt, 45 min
N
Ts
NHTs
(2 eq.)
88%
1
79%
2
equiv.) and leaving the reaction for no longer than 1 h
in total. In this way, a consistently high yield (88%) of
aziridine 1 of high purity (after removal of all of the
iodobenzene by-product by high vacuum drying for at
(Boc O, EtOAc) gave diamino cyclopentene 3 in an
2
acceptable 35% yield. The regio- and stereochemistry of
diamino cyclopentene 3 was assigned by analogy with
19
the known ring opening of the corresponding vinyl
epoxide under essentially the same conditions.
16
least 2 h) was obtained.
For the reaction of aziridine 1 with hydride, we utilised
a procedure recently reported by White and Wood for
ring opening of a terminal aziridine in a route to the
NHBoc
1
. 10 eq. NH3
MeOH, rt, 2 h
N
3 2
2. Et N, Boc O
17
Ts
NHTs
kalihinane diterpenoids. Thus, aziridine 1 was treated
EtOAc, rt, 16 h
35%
®
with lithium triethylborohydride (Superhydride ) at
1
3
0
°C in THF to give cyclopentene 2 in 79% yield. This
two step synthesis of 2 is the shortest synthetic entry
into 4-amino substituted cyclopentenes reported to
date.
Next, we proposed to epoxidise the alkene in 3. Reac-
tion of cyclopentene 3 with m-CPBA under standard
conditions gave a 74% isolated yield of epoxide 10 as a
single diastereoisomer. Although the epoxide stereo-
chemistry will ultimately be of no consequence in our
route to agelastatin A (vide infra), we have good evi-
dence that epoxidation proceeds cis to the NHBoc
group and trans to the NHTs group. Thus, two related
1
4
Aziridine 1 has also been utilised in a concise approach
to the cyclopentane core of agelastatin A. Isolated in
1
993 from the deep water marine sponge Agelas den-
dromorpha collected in Coral Sea near New Caledo-
1
5
nia, agelastatin A has an intriguing and unusual
alkaloid structure and exhibits anti-tumour and insecti-
cidal activity. In 1999, Weinreb et al. reported the first
total synthesis of racemic agelastatin A and two of their
late stage synthetic intermediates were enone 4 and
19
systems, amino alcohol 8 and amino acetate 9, were
also epoxidised and they gave the same sense of
diastereoselectivity as each other (shown by conversion
of epoxide 11 into 12 by simple acetylation).
1
8
allylic sulfonamide 5. Using aziridine 1 as the key
building block, we have now completed the preparation
of allylic sulfonamide 6 and enone 7 which contain
comparable functionality to Weinreb’s intermediates.
The stereochemistry of acetoxy epoxide 12 was estab-
lished unequivocally by conversion into chiral (but
racemic) diacetate 13. First of all, epoxide 12 was
−
reacted with sodium phenylselenide: the PhSe attacked
1
2 at the least hindered end of the epoxide (as shown by
Direct ring opening of aziridine 1 with ammonia fol-
lowed by Boc protection was the most reliable way of
preparing diamino cyclopentene 3. Thus, reaction of
aziridine 1 with an excess of ammonia in methanol at
room temperature followed (without isolation) by pro-
tection of the resulting amino sulfonamide intermediate
COSY analysis of the hydroxy selenide intermediate)
and also deprotected the acetate group. Then, after
acetylation of both hydroxyls, selenium oxidation and
subsequent elimination afforded diacetate 13 (31%
overall from 12) which was 1 clearly the chiral
13
diastereoisomer as judged by H and
C NMR
Me
X = SiMe3
O
HO
O
N
O
OTBDMS
NHBoc
O
X
O
Br
H
N
NBoc
NHBoc
NH
NHBoc
N
H
H
H
N
H
N
H
NHTs
NHTs
H
NHSES
O
O
4
5
6
7
Agelastatin A
OAc
O
1
. PhSeSePh, NaBH4
EtOH, reflux, 16 h
NHBoc m-CPBA, NaHCO3
NHBoc
NHBoc
CH Cl , rt, 16 h
2. Ac O, py, rt, 2 h
2
2
2
3
. Py, aq H O , CH Cl
X
X
2 2 2 2
OAc
0
°C → rt, 2 h
for X = OAc
3
8
9
(X = NHTs)
(X = OH)
(X = OAc)
10 (X = NHTs) 74%
11 (X = OH)
63%
12 (X = OAc) 75%
13 (31% from 12)

E. Baron et al. / Tetrahedron Letters 43 (2002) 723–726
725
2
0
spectroscopy (a meso diacetate would have been gen-
erated from the other diastereomeric epoxide). Based
on this, it seems reasonable to suggest that cyclopente-
nes 3, 8 and 9 (X=NHTs, OH and OAc, respectively)
are all epoxidised cis to the NHBoc group irrespective
of the nature of X. It is not clear whether this is due to
a simple hydrogen bond NHBoc-directed epoxidation
or due to an inherent facial bias in such trans-disubsti-
tuted cyclopentenes.
ates for agelastatin A synthesis. Our model studies on
the new route to the core of agelastatin A have led to
concise syntheses of allylic sulfonamide 6 and of enone
7 in four and five steps respectively from cyclopentadi-
ene. Furthermore, by running the lithium amide-medi-
ated rearrangement of epoxide 10 to allylic alcohol 15
under kinetic resolution conditions, it should be possi-
ble to generate enantiomerically enriched intermediates
for use in the first asymmetric synthesis of agelastatin
A.
24
In order to set up the required distribution of hydroxyl
(
or keto) and diamino functionality present in Wein-
reb’s intermediates 4 and 5 for agelastatin A synthesis,
we first needed to rearrange the epoxide in 10 into the
corresponding allylic alcohol. We elected to use a
lithium amide method for this rearrangement because
Acknowledgements
We thank the EPSRC and Lancaster Synthesis for a
CASE award (to T.D.T.) and the EU for a grant (to
E.B.).
21,22
of our interest in chiral base chemistry.
Thus, epox-
ide 10 was treated with four equivalents of the racemic
diamine-derived base 14 (due to the presence of two
acidic NH protons in 10 and our preference for two
equivalents of base per epoxide) in THF at room
temperature. The reaction was generally clean and typi-
cally afforded an 80% crude yield of an 85:15 mixture
of allylic alcohol 15 and starting epoxide 10 (by H
NMR spectroscopy). Our attempts at isolating pure
allylic alcohol 15 were thwarted as it was inseparable
from the epoxide. Instead, this crude epoxide and
allylic alcohol mixture was either acetylated or silylated
to give 16 or 6, respectively, and in this way it was
possible to generate allylic acetate 16 in 65% yield and
silyl protected allylic sulfonamide 6 in 39% yield (yields
are over the two steps of rearrangement and protec-
tion). Finally, the allylic acetate was deprotected using
potassium carbonate in methanol and the crude allylic
alcohol was oxidised using PDC in DMF (conditions
used by Weinreb et al. in their agelastatin A total
References
c
1. For a recent review, see: M Coull, W.; Davis, F. A.
1
Synthesis 2000, 1347–1365. For leading references on
acyclic 2-vinyl aziridines, see: Toda, A.; Aoyama, H.;
Mimura, N.; Ohno, H.; Fujii, N.; Ibuka, T. J. Org.
Chem. 1998, 63, 7053–7061.
2
3
2. (a) Tian, X. R.; Hudlicky, T.; K o¨ nigsberger, K. J. Am.
Chem. Soc. 1995, 117, 3643–3644; (b) Hudlicky, T.; Tian,
X. R.; K o¨ nigsberger, K.; Maurya, R.; Rouden, J.; Fan,
B. J. Am. Chem. Soc. 1996, 118, 10752–10765.
3. Hudlicky, T.; Tian, X. R.; K o¨ nigsberger, K.; Rouden, J.
J. Org. Chem. 1994, 59, 4037–4039.
4. Hudlicky, T.; Abboud, K. A.; Entwistle, D. A.; Fan, R.;
Maurya, R.; Thorpe, A. J.; Bolonick, J.; Myers, B.
Synthesis 1996, 897–911.
1
8
synthesis ) to give enone 7 in 43% yield over the two
steps.
5. Paul, B. J.; Martinot, T. A.; Willis, J.; Hudlicky, T.
Synthesis 2001, 952–956.
. Paul, B. J.; Hobbs, E.; Buccino, P.; Hudlicky, T. Tetra-
6
In summary, the readily prepared cyclopentadiene
hedron Lett. 2001, 42, 6433–6435.
7. O’Brien, P.; Pilgram, C. D. Tetrahedron Lett. 1999, 40,
8427–8430.
monoaziridine 1 has been used in a new approach to
4-amino substituted cyclopentenes and some intermedi-

7
26
E. Baron et al. / Tetrahedron Letters 43 (2002) 723–726
(1.88 g, 88%) as a brown solid, mp 93–95°C; R (Et O)
8
. Ling, R.; Yoshida, M.; Mariano, P. S. J. Org. Chem.
996, 61, 4439–4449.
. (a) Ling, R.; Mariano, P. S. J. Org. Chem. 1998, 63,
072–6076; (b) Cho, S. J.; Ling, R.; Kim, A.; Mariano, P.
S. J. Org. Chem. 2000, 65, 1574–1577.
0. Lu, H.; Mariano, P. S.; Lam, Y.-f. Tetrahedron Lett.
001, 42, 4755–4757.
f
2
−1 1
3 3
1
0.8; IR (CHCl ) 1321, 1159, 719 cm ; H NMR (CDCl ,
9
270 MHz): l 7.83 (d, 2H, J=8.5), 7.34 (d, 2H, J=8.5),
5.95 (br s, 2H), 3.86–3.83 (m, 1H), 3.72–3.71 (m, 1H),
6
13
2.59 (br s, 2H), 2.45 (s, 3H); C NMR (67.9 MHz,
1
CDCl ): l 144.3, 138.1, 135.3, 129.6, 127.9, 127.3, 50.7,
3
+
4
2
44.5, 35.6, 21.6; MS (CI, NH ) m/z 236 M+NH , 82;
HRMS (CI, NH ) m/z calcd for C H NO S M+H
3
+
1
1
1. Knight, J. G.; Muldowney, M. P. Synlett 1995, 949–951.
2. Atkinson, R. S.; Meades, C. K. J. Chem. Soc., Perkin
Trans. 1 2001, 1518–1527.
3
12 14
2
236.0745, found 236.0744.
17. White, R. D.; Wood, J. L. Org. Lett. 2001, 3, 1825–1827.
18. Stien, D.; Anderson, G. T.; Chase, C. E.; Koh, Y.-h.;
Weinreb, S. M. J. Am. Chem. Soc. 1999, 121, 9574–9579.
19. (a) Sundram, H.; Golebiowski, A.; Johnson, C. R. Tetra-
hedron Lett. 1994, 35, 6975–6976; (b) Kelly, R. C.; Schlet-
ter, I.; Stein, S. J.; Wierenga, W. J. Am. Chem. Soc. 1979,
1
3. Olivo, H. F.; Hemenway, M. S.; Hartwig, A. C.; Chan,
R. Synlett 1998, 247–248.
1
4. For other routes to 4-amino substituted cyclopentenes
(
e.g. 2), see: Barrett, S.; O’Brien, P.; Steffens, H. C.;
Towers, T. D.; Voith, M. Tetrahedron 2000, 56, 9633–
640 and references cited therein.
9
101, 1054–1056.
20. Spectroscopic data for 13: H NMR (CDCl
1
1
5. (a) D’Ambrosio, M.; Debitus, C.; Ribes, O.; Pusset, J.;
, 270 MHz):
3
Leroy, S.; Pietra, F. Chem. Commun. 1993, 1305–1306;
l 6.14 (br d, 1H, J=6.0), 6.06 (dd, 1H, J=2.0, 6.0), 5.74
(br d, 1H, J=5.0), 5.57 (br d, 1H, J=5.0), 5.01 (d, 1H,
J=5.0), 4.33–4.29 (m, 1H), 2.09 (s, 3H), 2.07 (s, 3H), 1.45
(
b) D’Ambrosio, M.; Guerriero, A.; Chiasera, G.; Pietra,
F. Helv. Chim. Acta 1994, 77, 1895–1902; (c) D’Ambro-
sio, M.; Guerriero, A.; Ripamonti, M.; Debitus, C.;
Waikedre, J.; Pietra, F.. Helv. Chim. Acta 1996, 79,
13
(s, 9H); C NMR (67.9 MHz, CDCl ): l 171.5, 170.4,
3
155.8, 137.2, 132.7, 81.2, 80.4, 75.9, 57.5, 28.7, 21.42,
21.36.
7
27–735.
1
6. Preparation of 1: PhIꢀNTs (3.4 g, 9.0 mmol) was added
portionwise to a stirred solution of freshly distilled
cyclopentadiene (1.2 g, 18.2 mmol) and Cu(acac)2 (240
21. For a review, see: O’Brien, P. J. Chem. Soc., Perkin
Trans. 1 1998, 1439–1457. For leading references, see: de
Sousa, S. E.; O’Brien, P.; Pilgram, C. D. Tetrahedron
Lett. 2001, 42, 8081–8083; de Sousa, S. E.; O’Brien, P.;
Steffens, H. C. Tetrahedron Lett. 1999, 40, 8423–8425.
22. de Sousa, S. E.; O’Brien, P.; Poumellec, P. J. Chem. Soc.,
Perkin Trans. 1 1998, 1483–1492.
mg, 0.9 mmol) in MeCN (10 mL) at 0°C under N . After
2
stirring for 15 min, the reaction was allowed to warm to
rt and stirred for a further 45 min. Then, the reaction
mixture was poured into NaOH(aq.) (1 M, 200 mL). Et O
2
(
50 mL) was added and the layers were separated. The
23. The phenylselenide route used to generate diacetate 13
(from 12) was low yielding when applied to the conver-
sion of epoxide 10 into allylic alcohol 15.
aqueous layer was extracted with Et O (2×50 mL) and
the combined organic extracts were dried (Na SO ) and
2
2
4
evaporated under reduced pressure. The crude product
24. Kee, A.; O’Brien, P.; Pilgram, C. D.; Watson, S. T.
was placed under high vacuum for 2 h to give aziridine 1
Chem. Commun. 2000, 1521–1522.