New route to azaspirocycles via the organolithium-mediated conversion of β-alkoxy aziridines into cyclopentenyl amines

Article abstract of DOI:10.1021/ol062073e

(Chemical Equation Presented) A new three-step route to azaspirocycles involving the organolithium-mediated conversion of β-alkoxy aziridines into substituted cyclopentenyl amines, hydroboration, and cyclization has been developed. The methodology is util

Full text of DOI:10.1021/ol062073e

ORGANIC
LETTERS
2006
Vol. 8, No. 22
5145-5148
New Route to Azaspirocycles via the
Organolithium-Mediated Conversion of
â
-Alkoxy Aziridines into Cyclopentenyl
Amines
Stephen P. Moore,† Susannah C. Coote,† Peter O’Brien,*,† and John Gilday‡
Department of Chemistry, UniVersity of York, Heslington, York YO10 5DD, U.K.,
and Process R & D, AstraZeneca, AVlon Works, SeVern Road, Hallen, Bristol BS10
7ZE, U.K.
paob1@york.ac.uk
Received August 22, 2006
ABSTRACT
A new three-step route to azaspirocycles involving the organolithium-mediated conversion of â-alkoxy aziridines into substituted cyclopentenyl
amines, hydroboration, and cyclization has been developed. The methodology is utilized in the construction of the pentacyclic ring system of
cephalotaxine.
The azaspirocyclic substructure is present in a number of
naturally occurring alkaloids including the cylindricines,
fasicularin, lepadiformine, halichlorine, histrionicotoxin,
cephalotaxine, and the pinnaic acids. As a result of the
biological activity associated with these natural products and
their derivatives, they have attracted much recent synthetic
activity.1-4 In this paper, we describe a new route to
cyclopentenyl azaspirocycles 1 (n ) 1, 2), which are
potentially useful intermediates for the synthesis of cepha-
lotaxine and halichlorine (Figure 1). The usefulness of our
approach, which utilizes lithiated aziridines, is shown by the
synthesis of a fully elaborated analogue of cephalotaxine.
Our group has an ongoing interest in the chemistry of
lithiated aziridines5,6 and we have previously reported the
Figure 1.
conversion of cyclic â-methoxy aziridines into substituted
allylic sulfonamides.7,8 As a typical example, the transforma-
† University of York.
‡ AstraZeneca.
(6) For reviews containing information on the chemistry of lithiated
aziridines, see: (a) Satoh, T. Chem. ReV. 1996, 96, 3303. (b) Hodgson, D.
M.; Bray, C. D.; Humphreys, P. G. Synlett 2006, 1. (c) For a special issue
on oxiranyl and aziridinyl anions, see: Tetrahedron 2003, 9693-9847.
(7) Rosser, C. M.; Coote, S. C.; Kirby, J. P.; O’Brien, P.; Caine, D.
Org. Lett. 2004, 6, 4817.
(8) For the conversion of â-methoxy epoxides into substituted allylic
alcohols, see: (a) Dechoux, L.; Doris, E.; Mioskowski, C. Chem. Commun.
1996, 549. (b) Doris, E.; Dechoux, L.; Mioskowski, C. Synlett 1998, 337.
(1) El Bialy, S. A.; Braun, H.; Tietze, L. F. Synthesis 2004, 2249.
(2) Dake, G. R. Tetrahedron 2006, 62 3467.
(3) Clive, D. L. J.; Yu, M.; Wang, J.; Yeh, V. S. C.; Kang, S. Chem.
ReV. 2005, 105, 4483.
(4) Weinreb, S. M. Chem. ReV. 2006, 106, 2531.
(5) (a) O’Brien, P.; Rosser, C. M.; Caine, D. Tetrahedron Lett. 2003,
44, 6613. (b) O’Brien, P.; Rosser, C. M.; Caine, D. Tetrahedron 2003, 59,
9779. (c) Huang, J.; O’Brien, P. Chem. Commun. 2005, 5696.
10.1021/ol062073e CCC: $33.50
© 2006 American Chemical Society
Published on Web 10/04/2006

Scheme 1
Scheme 3
satisfactory yields of aziridines cis-11 (48%) and cis-12
(53%) after chromatography (∼70:30 cis:trans diastereose-
tion of aziridine cis-2 into allylic sulfonamide 3 (54% yield)
was achieved using 2.5 equiv of n-BuLi and proceeded via
aziridine R-lithiation, carbenoid insertion into the organo-
lithium, and subsequent elimination of lithium methoxide
(Scheme 1). In these reactions, aziridines derived from
cyclopentenes gave higher yields of allylic sulfonamides
compared with those of the corresponding cyclohexene-
derived aziridines. Thus, our intitial efforts focused on
developing a route to functionalized versions of cyclopen-
tenyl sulfonamides 3 suitable for manipulation into halichlo-
rine- and cephalotaxine-like azaspirocycles.
Our proposed strategy for the synthesis of the azaspiro-
cyclic substructure present in cephalotaxine and halichlorine
via lithiated aziridines is presented in Scheme 2. We
anticipated preparing alkene-substituted aziridines cis-6 (from
cyclopentenones 4 and 5) and then using excess organo-
lithium reagents to convert them into allylic amines 7 using
our established methodology.7 Subsequent hydroboration of
the less sterically hindered alkene to give 8 and then hydroxyl
activation and cyclization would afford azaspirocycles 1.
1
lectivity from the H NMR spectra of the crude reactions).
Of note, these reactions also showed a high level of
chemoselectivity as products from reaction of the terminal
alkene in 9 and 10 were not detected. Low temperature (-78
°C) methylation using KHMDS and methyl iodide then
completed the synthesis of the required â-methoxy aziridines
cis-13 and cis-14 (Scheme 3).
The organolithium-mediated transformation of â-methoxy
aziridines cis-13 and cis-14 into substituted allylic amines
was next investigated (Table 1). These reactions were best
carried out using 2.5 equiv of organolithium in Et2O or
tBuOMe (TBME) at -78 °C for 5 min before warming to
room temperature. Pleasingly, the reactions proceeded
Table 1. Organolithium-Mediated Conversion of â-Methoxy
Aziridines into Substituted Allylic Sulfonamides
Scheme 2
entry
SM
R
solvent
prod
% yielda
1
2
3
4
5
6
7
8
cis-13
cis-13
cis-13
cis-13
cis-13
cis-13
cis-13
cis-13
cis-14
cis-14
cis-14
nBu
nBu
Me3SiCH2
Me3SiCH2
Ph
Et2O
TBME
Et2O
TBME
Et2O
TBME
Et2O
TBME
Et2O
Et2O
15a
15a
15b
15b
15c
15c
15db
15db
16a
16b
16c
71
81
63
77
58
68
85
82
80
82
63
Ph
sBu
sBu
To start with, known cyclopentenones 49 and 510 (see
Supporting Information for their preparation) were reduced
to allylic alcohols 9 and 10, respectively, in good yield
(Scheme 3). Then, aziridination11 using Chloramine-T (TsN-
ClNa) and phenyltrimethylammonium tribromide (PTAB)
proceeded with good cis-selectivity (cis stereochemistry
assigned by analogy with our earlier work7,12) providing
9
10
11
nBu
Me3SiCH2
Ph
Et2O
a Isolated yield after chromatography. b Obtained as a 1:1 mixture of
diastereoisomers.
1
(12) We have identified some useful, diagnostic trends in the H NMR
spectra of aziridines 11 and 12 that enable assignment of their relative
stereochemistry. For the CHN proton: δH 3.45 (d, J ) 2.5 Hz) for cis-11
and δH 3.35 (s) for trans-11; δH 3.42 (d, J ) 2.5 Hz) for cis-12 and δH
3.31 (s) for trans-12. Thus, for the cis aziridines, the CHN proton appears
more downfield and is a doublet (J ) 2.5 Hz) compared to the trans
aziridines.
(9) Piers, E.; Banville, J.; Lau, C. K.; Nagakura, I. Can. J. Chem. 1982,
60, 2965.
(10) Wolff, S.; Agosta, W. C. J. Org. Chem. 1981, 46, 4821.
(11) Jeong, J. U.; Tao, B.; Sagasser, I.; Henniges, H.; Sharpless, K. B.
J. Am. Chem. Soc. 1998, 120, 6844.
5146
Org. Lett., Vol. 8, No. 22, 2006

smoothly and allylic sulfonamides 15a-d and 16a-c were
generated in 58-85% yield. For aziridine cis-13, TBME was
the preferred solvent as it generally gave a ∼10% improve-
ment in yield (Table 1, compare entries 1/2, 3/4, and 5/6).
In general, reactions using PhLi gave lower yields compared
with those using primary and secondary alkyllithium reagents
(Table 1, compare entries 1/3/5/7 and 9-11). In all of the
thus been the subject of several recent synthetic studies.1,16
As outlined in Scheme 4, we envisaged that azaspirocycle
21 could be obtained from allylic sulfonamide 22 by two
cyclization processes (including N-Ts deprotection). To
prepare allylic sulfonamide 22, reaction of aziridine cis-13
with excess functionalized aryllithium 23 (obtained by
lithium-bromine exchange) was to be explored.
1
examples shown in Table 1, we saw no evidence in the H
NMR spectra of the crude products of intramolecular
cyclopropanation of the aziridines, even though such a
process is well established for certain lithiated epoxides and
aziridines equipped with a pendent alkene.13
Scheme 4
With allylic sulfonamides 15a-d and 16a-c in hand, we
were ready to investigate a two-step conversion into the
desired azaspirocycles. First of all, chemo- and regioselective
hydroboration of the terminal alkene was achieved using
9-BBN14 to give alcohols 17a-d and 18a-c in 35-86%
yield (Table 2). Then, cyclization under Mitsunobu condi-
tions15 using DEAD or DIAD and PPh3 gave azaspirocycles
19a-d and 20a-c in 61-87% yield (Table 2). Thus, a
simple, three-step route from â-methoxy aziridines cis-13
and cis-14 to azaspirocycles 19a-d and 20a-c has been
established. Azaspirocycles 20a-c correspond to the spiro-
cyclic substucture of halichlorine, and Me3SiCH2-substituted
allylic sulfonamide 20b could be elaborated toward this
natural product.
To investigate the viability of using functionalized aryl-
lithium reagents in combination with â-methoxy aziridines,
reactions between the aryllithiums derived from either
bromobenzene or bromides 25-2717 and aziridines cis-27 and
cis-13 were studied (Table 3). Lithium-bromine exchange18
of bromobenzene or 24 was accomplished using n-BuLi, and
Table 3. Transformation of â-Methoxy Aziridines into
Substituted Allylic Sulfonamides Using Functionalized
Aryllithium Reagents
Table 2. Hydroboration-Mitsunobu Cyclization of Allylic
Sulfonamides 15 and 16
entry
SM
R
prod, % yielda
prod, % yielda
1
2
3
4
5
6
7
15a
15b
15c
15d
16a
16b
16c
nBu
17a, 35
17b, 86
17c, 79
17d, 60b
18a, 64
18b, 71
18c, 69
19a, 65
19b, 76
19c, 61
19d, 71b
20a, 79
20b, 87
20c, 76
Me3SiCH2
Ph
sBu
nBu
Me3SiCH2
Ph
entry
SM
ArBr
solvent
product
% yielda
1
2
3
4
5
6
7
cis-2
cis-2
PhBr
25
26
26
26
Et2O
Et2O
Et2O
TBME
Et2O
Et2O
TBME
24ab
24bb
15e
15e
15e
15f
46c
46
64
54
44d
43
37
a Isolated yield after chromatography. b Obtained as a 1:1 mixture of
diastereoisomers.
cis-13
cis-13
cis-13
cis-13
cis-13
27
27
15f
Our next aim was to utilize this azaspirocycle synthetic
methodology for the preparation of azaspirocycle 21, a
potential intermediate for cephalotaxine synthesis. Naturally
ocurring esters derived from cephalotaxine (e.g., deoxyhar-
ringtonine) show pronounced antileukemic activity and have
a Isolated yield after chromatography. b Reaction carried out using 2.5
equiv of ArLi. c 69% yield of 24a using a commercial solution of PhLi in
nBu2O. d Reaction carried out using 1.5 equiv of 26 and 3 equiv of nBuLi.
(13) (a) Dechoux, L.; Agami, C.; Doris, E.; Mioskowski, C. Tetrahedron
2003, 59, 9701. (b) Hodgson, D. M.; Chung, Y. K.; Paris, J.-M. J. Am.
Chem. Soc. 2004, 126, 8664. (c) Hodgson, D. M.; Humphreys, P. G.; Ward,
J. G. Org. Lett. 2006, 8, 995.
(14) Riber, D.; Hazell, R.; Skrydstrup, T. J. Org. Chem. 2000, 65, 5382.
(15) Henry, J. R.; Marcin, L. R.; McIntosh, M. C.; Scola, P. M.; Davis
Harris, G.; Weinreb, S. M. Tetrahedron Lett. 1989, 30, 5709.
(16) For selected recent examples, see: (a) Eckelbarger, J. D.; Wilmot,
J. T.; Gin, D. Y. J. Am. Chem. Soc. 2006, 128, 10370. (b) Zhao, Z.; Mariano,
P. S. Tetrahedron 2006, 62, 7266. (c) Li, W.-D. Z.; Ma, B.-C. J. Org. Chem.
2005, 70, 3277. (d) Planas, L.; Perard-Viret, J.; Royer, J. J. Org. Chem.
2004, 69, 3087. (e) Tietze, L. F.; Schirok, H. J. Am. Chem. Soc. 1999, 121,
10264.
Org. Lett., Vol. 8, No. 22, 2006
5147

the resulting aryllithiums were then reacted with model
aziridine cis-2 to give allylic sulfonamides 24a and 24b in
46% yield for each reaction (Table 3, entries 1/2). By way
of comparison, the equivalent reaction with a commercially
available solution of PhLi delivered allylic sulfonamide 24a
in 69% yield.
Scheme 5
Next, reactions between aziridine cis-13 and the aryllithi-
ums obtained from bromides 26 and 27 were investigated
(Table 3, entries 3-7). The desired allylic sulfonamides 15e
and 15f were generated in 37-64% yield, and Et2O was
preferred to TBME as solvent in these cases (Table 3,
compare entries 3/4 and 6/7). Allylic sulfonamide 15f has
been further elaborated toward cephalotaxine (vide infra),
and it was best generated (43% yield) using 3 equiv of the
aryllithium derived from bromide 27 in Et2O (Table 3, entry
6). We also briefly attempted to reduce the quantity of aryl
bromide 26 required in the reaction. Thus, 3 equiv of n-BuLi
was combined with only 1.5 equiv of aryl bromide 26 before
reaction with aziridine cis-13. It was anticipated that the
excess n-BuLi would carry out the required aziridine
lithiation and then the resulting lithium carbenoid would
insert into the aryllithium to produce 15e. Indeed, this turned
out to be the case as allylic sulfonamide 15e was produced
in 44% yield (Table 3, entry 5) accompanied by small
amounts of 15a (not isolated).19
Having demonstrated that functionalized aryllithium re-
agents could be used in lithiation-carbenoid insertion pro-
cesses with aziridines, we set about constructing the penta-
cyclic ring structure present in cephalotaxine (Scheme 5).
Thus, the tri-isopropylsilyl group in allylic sulfonamide 15f
was deprotected using TBAF, and the resulting alcohol was
cyclized under Mitsunobu conditions to give sulfonamide
28 in quantitative yield. Next, hydroboration followed by
N-Ts deprotection using Na/naphthalene20 gave amino al-
cohol 29 (56% yield over two steps). Finally, 29 was cyclized
21
to pentacycle 21 using PPh3/CBr4 (71% yield), thus
completing a connective approach to the cephalotaxine core.22
In summary, a three-step route to azaspirocycles via the
organolithium-mediated conversion of â-alkoxy aziridines
into substituted cyclopentenyl amines, hydroboration, and
Mitsunobu cyclization has been developed. In a useful
extension of our lithiated aziridine methodology, function-
alized aryllithium reagents have been successfully employed,
and this enabled a concise and connective approach to the
pentacyclic ring system of cephalotaxine (nine steps from
cyclopentenone 4) to be developed.
Acknowledgment. We thank the EPSRC and AstraZen-
eca for a CASE award (to S.P.M.) and the EPSRC for a
DTA award (to S.C.C.).
Supporting Information Available: Full experimental
procedures, characterisation data, and copies of 1H NMR and
13C NMR spectra. This material is available free of charge
via the Internet at http://pubs.acs.org.
(17) Bromides 25 and 26 are commercially available. Bromide 27 was
prepared by silylation of the alcohol (see Supporting Information). For
synthesis of the alcohol, see: Tietze, L. F.; Schirok, H.; Wo¨hrmann, M.;
Schrader, K. Eur. J. Org. Chem. 2000, 2433.
(18) (a) Bailey, W. F.; Luderer, M. R.; Jordan, K. P. J. Org. Chem. 2006,
71, 2825. (b) Lin, X.; Kavash, R. W.; Mariano, P. S. J. Org. Chem. 1996,
61, 7335.
OL062073E
(20) Williams, J. T.; Bahia, P. S.; Kariuki, B. M.; Spencer, N.; Philp,
D.; Snaith, J. S. J. Org. Chem. 2006, 71, 2460.
(21) For a related cyclisation, see: Kavash, R. W.; Mariano, P. S.
Tetrahedron Lett. 1989, 30, 5709.
(19) For an example of the use of mixtures of two organolithium reagents
in a related reaction of an epoxide, see: Hodgson, D. M.; Paruch, E.
Tetrahedron 2004, 60, 5185.
(22) For a different route to a similar intermediate for cephalotaxine
synthesis, see: Sha, C.-K.; Young, J.-J.; Yeh, C.-P.; Chang, S.-C.; Wang,
S.-L. J. Org. Chem. 1991, 56, 2694.
5148
Org. Lett., Vol. 8, No. 22, 2006