Synthesis of a spirocyclic amine related to the marine natural products halichlorine and pinnaic acid

Article abstract of DOI:10.1016/j.tetlet.2005.02.133

The bis-ester 5 was elaborated into tricyclic lactam 8, which has previously been converted into the spirotricyclic amine 9, a compound that represents a large portion of the marine natural product halichlorine.

Full text of DOI:10.1016/j.tetlet.2005.02.133

Tetrahedron
Letters
Tetrahedron Letters 46 (2005) 2853–2855
Synthesis of a spirocyclic amine related to the marine
natural products halichlorine and pinnaic acid
Derrick L. J. Clive,* Jian Wang and Maolin Yu
Chemistry Department, University of Alberta, Edmonton, Alberta, Canada T6G 2G2
Received 9 February 2005;accepted 21 February 2005
Abstract—The bis-ester 5 was elaborated into tricyclic lactam 8, which has previously been converted into the spirotricyclic amine 9,
a compound that represents a large portion of the marine natural product halichlorine.
ꢀ 2005 Published by Elsevier Ltd.
Halichlorine (1) and the pinnaic acids (2, 3) have at-
tracted considerable attention as synthetic targets—in
part, because the substances have biological properties
of interest to those studying various inflammatory pro-
cesses. The Danishefsky school has reported the total
syntheses of both 11 and 22 in optically pure form, while
Christie and Heathcock3 have completed syntheses of
the same compounds as racemates. There is also a large
literature4 on synthetic studies that describe routes to
the spirocyclic core of these compounds. In this regard,
Huxford and Simpkins have recently reported4g a route
from piperidine 4, via allylated derivative 5, to the spiro-
cycle 6 (Scheme 1). Although the latter would require
inversion of stereochemistry of the ester appendage at
C-13 for elaboration into 1 or 2, the synthesis illustrates
the utility of the asymmetric allylation 4 ! 5, devel-
oped6,7 a few years ago by the Simpkins group. We
had previously used this asymmetric allylation in a syn-
thesis of 75 (Scheme 2) but had found that, in our hands,
MeO2C
Bn
MeO2C
Bn
t-BuPh2SiO
N
N
H
N
Bn
13
MeO2C
EtO2C
MeO2C
4
5
6
Scheme 1.
4
5
MOMO
ref. 5
N
9
13
H
7
H
14
MeO2C
HO
4
5
N
N
ref. 11
O
EtO2C
t-BuPh2SiO
present
route
H
H
4
6
9
4
6
9
H
H
3
3
7
8
7
8
10
5
N
5
HN
8
9
O
O
2
2
10
Me
1
1
X
Scheme 2.
HO
11
11
13
14
13
14
O
12
12
H
H
Me
Me
the level of stereocontrol was poor, and the allylated
compound 5 had an ee of only 67% (HPLC). As a result
of correspondence with Professor Simpkins we have
examined the allylation of both small (110 mg) and large
(5–5.5 g) samples of 4 and found that the ee is sensitive
to the scale of operation, and that the process gives poor
results in large scale experiments.8 Notwithstanding this
outcome, we have continued to use the large scale allyl-
ation of 4, and we describe here a synthetic route that is
Cl
Cl
HO
HO
1
2 X = OH
3 X = NHCH2CH2SO3H
Keywords: Halichlorine;Pinnaic acid;Ring closing metathesis;Spiro
compound;Wittig homologation.
*
Corresponding author. Tel.: +1 7804923251;fax: +1 7804928231;
e-mail: derrick.clive@ualberta.ca
0040-4039/$ - see front matter ꢀ 2005 Published by Elsevier Ltd.
doi:10.1016/j.tetlet.2005.02.133

2854
D.L.J.Clive et al./ Tetrahedron Letters 46 (2005) 2853–2855
initially similar to the one just reported4g by Huxford
and Simpkins, but which leads instead to the lactam 8
(Scheme 2). This tricyclic compound (in racemic form)
has been converted11 into the advanced intermediate
( )-9, whose relative stereochemistry is the same as in
halichlorine;the corresponding optically active tert-
butyl ester is an intermediate in the Danishefsky total
synthesis.1 Lactam ( )-8 has also been elaborated11 into
the racemic version of an intermediate in the Danishef-
sky route2 to pinnaic acid.
The hindered nitrogen of 16 did not undergo acylation
at all readily, but it did react with EtO2CCH2COCl in
CH2Cl2 in the presence of Et3N at 0 ꢁC to afford 17a
(72%). Treatment with DBU in refluxing MeCN then ef-
fected both cyclization and dehydration to produce 18a.
The corresponding transformation in the methyl ester
series (16 ! 17b ! 18b) could be accomplished without
isolation of any intermediates, and in better overall yield
(83% vs 51%), by allowing 16 to react with MeO2CCH2-
COCl at 0 ꢁC in MeCN in the presence of Et3N, fol-
lowed by addition of DBU, and a 10 h reflux period.
Compound 18a was convertible into the advanced inter-
mediate 21 by three simple steps. Catalytic hydrogen-
ation of the double bond of 18a gave 19a (72%), in
which epimerization had occurred at C-14, but the other
newly-created stereocenter (C-13) was generated in the
desired manner. Desilylation also took place during
the reduction, and so this protecting group was rein-
stalled (19a ! 20a, 96%) under standard conditions.
In the same way as described by Huxford and Simpkins,
but using material of lower ee, the bis-ester 5 was reduced
(LiAlH4, 71%) and the less hindered primary hydroxyl in
the product was protected (87%) as its t-BuMe2Si ether
(5 ! 10 ! 11). Interestingly, silylation occurs at the C-
4 hydroxyl, whereas pivaloylation of the same alcohol re-
sulted in the opposite regioselectivity.5 Swern oxidation
afforded aldehyde 12, and reaction of the crude material
with vinylmagnesium bromide gave the single alcohol 13
in 81% overall yield. (The whole process was much less
efficient if an attempt was made to purify the aldehyde.)
We did not establish the stereochemistry at the hydroxyl-
bearing carbon, as that center is subsequently converted
to sp2 hybridization, but Huxford and Simpkins did
make a stereochemical assignment (S) for the corre-
sponding compound in which the C-4 alcohol is pro-
tected as a t-BuPh2Si ether. Ring closing metathesis
with 2 mol% Grubbs II catalyst in refluxing CH2Cl2 then
gave the cyclopentenol 14 (85–95%), which was subjected
to Swern oxidation (14 ! 15, 72%). From this point our
route differs from that of Huxford and Simpkins. The
double bond in enone 15 was readily saturated by hydro-
genation over Pd–C, with the N-benzyl group being
removed at the same time (15 ! 16, 86%) (Scheme 3).
c
Si*O
Si*O
Si*O
O
a
b
N
N
N
O
O
H
O
RO2C
RO2C
16
17a, R = Et
17b, R = Me
18a, R = Et
18b, R = Me
g
f
d
Si*O
HO
Si*O
e
N
N
N
O
O
O
14
14
14
13
EtO2C
H
HO
EtO2C
H
H
21
20a
19a
4
h
MeO2C
t-BuMe2SiO
b
HO
a
HO
O
Si*O
Si*O
N
N
N
j
i
Bn
Bn
Bn
N
N
N
O
O
MeO2C
HO
c
HO
14
14
14
I
5
10
11
12
H
H
H
24
k
22
23
Si*O
Si*O
d
Si*O
Bn
e
O
HO
N
N
N
O
O
Bn
Bn
5
m
5
N
5
N
l
N
HO
O
O
O
OH
13
14
14
f
H
H
H
8
26
25
t-BuMe2SiO
t-BuMe2SiO
Si* = t-BuMe2Si
g
N
N
Bn
H
Scheme 4. Reagents and conditions: (a) Et3N, EtO2CCH2COCl,
MeCN, 0 ꢁC, then 4 h at room temperature;(b) DBU, reflux, 5 h,
51% from 16;(c) Et 3N, MeO2CCH2COCl, MeCN, 0 ꢁC, then removed
ice bath, 4 h;DBU, reflux, 10 h, 83%;(d) Pd–C, EtOAc, H 2, 50 psi,
48 h, 72% ;(e) t-BuMe2SiCl, imidazole, DMAP, CH2Cl2, 7 h, 96%;(f)
DIBAL, PhMe, À78 ꢁC, 15 min, then NaBH4, MeOH, 5 h, 50%;(g)
LiBH4, THF–MeOH, 18 h, 87% ;(h) I 2, Ph3P, imidazole, CH2Cl2, 3 h,
96%;(i) Bu 3SnH, AIBN, PhH, reflux, 4 h;(j) Bu 4NF, THF, 10 h, 99%
from 22;(k) Dess–Martin periodinane, CH 2Cl2, 1 h, 93%;(l)
(Me3Si)2NK, PhMe, Ph3PCH2(OMe)Cl, 0 ꢁC, then 3 N HCl, aqueous
THF, 3 h, ca. 75%;(m) LiBH 4, THF, 0 ꢁC to room temperature (ice
bath left in place but not recharged), 4 h, 87%.
O
O
15
Si* = t-BuMe2Si
16
Scheme 3. Reagents and conditions: (a) LiAlH4, THF, 0 ꢁC, 5 h, 71%;
(b) t-BuMe2SiCl (0.80 equiv), imidazole, DMAP, CH2Cl2, 10 min, then
t-BuMe2SiCl (0.35 equiv), 30 min, 87%;(c) Swern oxidation, 12 is not
very stable to flash chromatography over silica and is best used crude;
(d) vinylmagnesium bromide, THF, À20 ꢁC, 30 min, 81% from 11;(e)
2 mol% Grubbs II catalyst, CH2Cl2, 35 ꢁC, 4 h, 85–95%;(f) Swern
oxidation, 72%;(g) Pd–C, EtOAc, H 2, 46 psi, 9.5 h, 86%.

D.L.J.Clive et al./ Tetrahedron Letters 46 (2005) 2853–2855
2855
Hughes, D. L.;Stockman, R. A. Tetrahedron Lett. 2004,
45, 8371–8374;(f) Feldman, K. S.;Perkins, A. L.;Masters,
K. M. J.Org.Chem. 2004, 69, 7928–7932;(g) Huxford, T.;
Simpkins, N. S. Synlett 2004, 2295–2298.
Reduction of the ester group (DIBAL-H, followed by
NaBH4, 50%) then gave alcohol 21. The same com-
pound could be generated directly from enone 18b by
treatment with LiBH4 (87%). We had initially intended
to oxidize the resulting primary alcohol to the corre-
sponding aldehyde, but our attempts to do so led only
to the removal of the hydroxymethyl group and form-
ation of a carbonyl at C-14. Accordingly, the hydroxyl
was replaced by iodine (96%) using Ph3P, I2, and imid-
azole12 (21 ! 22), from which point stannane reduction
gave the a-methyl lactam 23 (99%). Deprotection
(Bu4NF, THF, 99% overall from 22) and oxidation with
the Dess–Martin reagent afforded aldehyde 25 (93%),
which was homologated (ca. 75% yield) to 26 by Wittig
reaction with Ph3P@CH(OMe) (generated in PhMe with
(Me3Si)2NK as base), followed by mild acid hydrolysis
(3 N HCl in aq THF, 3 h). The conditions for the Wittig
reaction are critical, and use of a low (0 ꢁC) temperature
is essential. If the reaction is done at room temperature
epimerization at C-5 occurs. Simple LiBH4 reduction of
crude aldehyde 26 gave alcohol 8 (87%);when the previ-
ous Wittig reaction was run at room temperature, the
same sequence led to the C-5 epimer of 8. Alcohol 8
has been converted into the spirotricyclic amine 9
(Scheme 4).11
5. Yu, M.;Clive, D. L. J.;Yeh, V. S. C.;Kang, S.;Wang, J.
Tetrahedron Lett. 2004, 45, 2879–2881.
6. Goldspink, N. J.;Simpkins, N. S.;Beckmann, M. Synlett
1999, 1292–1294.
7. Bambridge, K.;Begley, M. J.;Simpkins, N. S. Tetrahedron
Lett. 1994, 35, 3391–3394.
8. Experimental details are given in Ref. 6 for asymmetric
benzylation of 5. By following that procedure (including
the scale of operation) we obtained the benzylated product
with [a]D +32 (c 1.7, CHCl3);the reported value (Ref. 6) is
[a]D +27 (c 1.7, CHCl3). Our allylation experiments were
modeled on the benzylation procedure;our chiral bis-
25
amine had ½aꢀD +192 (c 0.97, CHCl3);lit. (Ref. 7) +205 (c
0.7, CHCl3). (a) With precooling of bis-ester solution: in a
small scale allylation, the dilithium diamide was 0.0979 M,
and the bis-ester solution [precooled to À78 ꢁC (see Ref. 9
for description of the apparatus)], that was added over
5 min, was 0.0925 M (110 mg of bis-ester);neat allyl
bromide (15 equiv, not precooled) was added over 2 min.
The product (5) had [a]D +40 (c 2.7, CHCl3);the value
observed by Simpkins et al. (private communication from
Professor Simpkins) was [a]D +33 (c 2.7, CHCl3). In a
large scale allylation, the dilithium diamide was 0.0818 M,
and the bis-ester solution (precooled to À78 ꢁC), that was
added over 90 min, was 0.343 M (5.0 g of bis-ester);neat
allyl bromide (3.3 equiv, not precooled) was added over
2 min. The product (5) had [a]D +26 (c 2.7, CHCl3). (b)
Without precooling of bis-ester solution: in a second large
scale allylation, the dilithium diamide was 0.0926 M, and
the bis-ester solution, that was added over 90 min, but was
not precooled, was 0.945 M (5.5 g of bis-ester);neat allyl
bromide (3 equiv, not precooled) was added over 3 min.
The product (5) had [a]D +26 (c 2.7, CHCl3). While we did
not examine our sample of SimpkinsÕ base used in the
above experiments by HPLC, the [a]D measurements we
made on the allylation products show that there is
significant erosion of ee in scaling up the allylation. (c)
With precooling of bis-ester solution: two further experi-
ments were performed, using concentrations close to
those for our benzylation, but again, the ee (HPLC) for
the large scale work was lower: the dilithium diamide was
0.11 M and the bis-ester solution (precooled to À78 ꢁC),
that was added over 10 min, was 0.09 M (110 mg of bis-
ester);neat allyl bromide (15 equiv, not precooled) was
added over 3 min. The product (5) had [a]D +31.5 (c 1.0,
CHCl3) and had an ee of 66.5% (HPLC, Chiralcel OD,
25 cm · 0.46 cm, 1% i-PrOH–hexane, detection at 440 nm,
flow1.2 mL/min). On a larger scale: the dilithium diamide
was 0.09 M and the bis-ester solution (precooled to
À78 ꢁC), that was added over 2 h, was 0.10 M (5.0 g of
bis-ester);neat allyl bromide (3 equiv, not precooled) was
added over 5 min. The product (5) had [a]D +24 (c 3.2,
CHCl3) and had an ee of 55.5% (HPLC). Cf. Ref. 10.
9. Suzuki, M.;Yanagisawa, A.;Noyori, R. J.Am.Chem.
Soc. 1988, 110, 4718–4726.
The above experiments provide a new route to com-
pounds relevant to the synthesis of halichlorine and
the pinnaic acids. Our work also shows that proper tem-
perature control must be exercised during homologation
of aldehydes of type 25 by Wittig reaction, in order to
suppress epimerization.
Acknowledgments
We thank the Natural Sciences and Engineering
Research Council of Canada for financial support,
and Dr. C. Diaper for HPLC analyses.
References and notes
1. Trauner, D.;Schwarz, J. B.;Danishefsky, S. J. Angew.
Chem., Int. Ed. 1999, 38, 3542–3545.
2. (a) Carson, M. W.;Kim, G.;Hentemann, M. F.;Trauner,
D.;Danishefsky, S. J. Angew.Chem,. Int.Ed.
2001, 40,
4450–4452;(b) Carson, M. W.;Kim, G.;Danishefsky,
S. J. Angew.Chem,. Int.Ed. 2001, 40, 4453–4456.
3. (a) Christie, H. S.;Heathcock, C. H. Proc.Natl.Acad.Sci.
2004, 101, 12079–12084;(b) Cf. Hayakawa, I.;Arimoto,
H.;Uemura, D. Heterocycles 2003, 59, 441–444.
4. (a) For model studies, see Ref. 5 and Yokota, W.;Shindo,
M.;Shishido, K. Heterocycles 2001, 54, 871–885;(b)
Keck, G.;Dalton, S. A. Abstract 187, 226th ACS
National Meeting, New York, September 7–11, 2003;(c)
Hayakawa, I.;Arimoto, H.;Uemura, D. J.Chem.Soc,.
Chem.Commun. 2004, 1222–1223;(d) Dake, G. R.;
Fenster, M. D. B.;Hurley, P. B.;Patrick, B. O. J.Org.
Chem. 2004, 69, 5668–5675;(e) Arini, L. G.;Szeto, P.;
10. For effects of concentration on ee, see: Majewski, M.;
Wang, F. Tetrahedron 2002, 58, 4567–4571.
11. Matsumura, Y.;Aoyagi, S.;Kibayashi, C. Org.Lett. 2004,
6, 965–968.
12. Garegg, P. J.;Samuelsson, B. J.Chem.Soc., Perkin Trans.
1 1980, 2866–2869.