Towards a total synthesis of (-)-cephalotaxine: Construction of the BCDE-tetracyclic core

Article abstract of DOI:10.1016/S0040-4039(02)01219-4

The synthesis of the BCDE-tetracyclic core of (-)-cephalotaxine has been achieved in eight steps starting from N-Boc-L-proline methyl ester. An alkylidene carbene 1,5-CH insertion reaction was used as a key step in the stereocontrolled synthesis of the DE

Full text of DOI:10.1016/S0040-4039(02)01219-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 6011–6014
Towards a total synthesis of (−)-cephalotaxine: construction of
the BCDE-tetracyclic core
Stephen M. Worden, Renameditswe Mapitse and Christopher J. Hayes*
The School of Chemistry, The University of Nottingham, University Park, Nottingham NG7 2RD, UK
Received 28 May 2002; accepted 24 June 2002
Abstract—The synthesis of the BCDE-tetracyclic core of (−)-cephalotaxine has been achieved in eight steps starting from
N-Boc-L-proline methyl ester. An alkylidene carbene 1,5-CH insertion reaction was used as a key step in the stereocontrolled
synthesis of the DE-spirocyclic fragment and an intramolecular Heck-type cyclisation was used to form the benzazepine ring, and
hence complete the tetracycle formation. © 2002 Published by Elsevier Science Ltd.
The evergreen plum-yews of the genus Cephalotaxus
produce a range of structurally related polycyclic alka-
loids of which (−)-cephalotaxine 1 (Fig. 1) is the most
abundant member.1 Whilst 1 has little known biological
activity, a considerable amount of interest has been
paid to the potential therapeutic uses of a range of
co-occurring cephalotaxine esters. In particular, homo-
harringtonine 22 has attracted considerable attention
due to its promising antileukaemic activity, and this
material is currently undergoing advanced clinical
trials.3
Of the synthetic challenges presented by (−)-cephalotax-
ine 1, the asymmetric construction of the quaternary
stereocentre of the DE-spirocyclic fragment is perhaps
the most demanding. We have an ongoing interest in
the use of alkylidene carbenes for the asymmetric con-
struction of nitrogen-bearing quaternary stereocentres,6
and we wondered whether we could use a 1,5-CH
insertion reaction for the formation of the key spiro-
cyclic ring system. In order to test the suitability of this
strategy for the synthesis of 1, we first chose to examine
the construction of a simpler model of the tetracyclic
skeleton 3. Our retrosynthetic analysis of the BCDE-
tetracyclic core 3 is outlined in Scheme 1. Following
good literature precedent5a we disconnected the C-ring
using a retro-Heck reaction to reveal the spirocyclic
tertiary amine 4 as an advanced intermediate. We were
aware of the fact that a mixture of 3 and its exocyclic
The combination of interesting chemical structure and
profound biological activity have stimulated a large
number of studies directed towards the total synthesis
of these alkaloids. Cephalotaxine itself has attracted
most of the synthetic interest and this has resulted in a
number of rather elegant total syntheses.4 Upon exami-
nation of this work, however, it is perhaps surprising to
find that only three syntheses of enantiomerically pure
(−)-cephalotaxine have been reported to date.5
Heck
N
B
N
C
D
Pd-catalyst
Br
E
H
R=H, (-)-Cephalotaxine 1
O
A
O
3
4
1,5-CH insertion
O
B
N
C
R=
MeO2C
Homoharringtonine 2
D
OH
E
H
OH
steps
N
R
RO
OMe
CO2Me
N
H
Boc
Figure 1.
5
6
* Corresponding author. Tel.: +44-(0)115-951-3045; fax: +44-(0)115-
951-3564; e-mail: chris.hayes@nottingham.ac.uk
Scheme 1. Retrosynthetic analysis of the BCDE core.
0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd.
PII: S0040-4039(02)01219-4

6012
S. M. Worden et al. / Tetrahedron Letters 43 (2002) 6011–6014
double bond isomer 10 were likely to be produced in
the Heck reaction, but we felt that the fully function-
alised cephalotaxine E-ring could be accessed from
either of these materials using standard functional
group interconversions, and this regiochemical issue did
not concern us at this stage of the synthetic planning.
Further disconnection of 4 via a retro-1,5-CH insertion
reaction revealed the alkylidene carbene-substituted
pyrrolidine 6 as a key reactive intermediate. We envis-
aged 6 as being accessed from readily available N-Boc
proline methyl ester in only a few synthetic steps.
Guided by our retrosynthetic analysis (Scheme 1), the
desired alkylidene carbene cyclisation precursor 8 was
synthesised in six steps from N-Boc proline methyl ester
5 in good overall yield (Scheme 2). Dibal-H reduction
of 5 and homologation of the resulting aldehyde first
produced the corresponding enone. Catalytic hydro-
genation and a second olefination using bromomethyl
triphenylphosphorane then provided the 1-bromo-1-
alkene 7 as a 2.5:1 mixture of E:Z isomers. Acid-
catalysed cleavage of the Boc-protecting group and
reductive amination finally afforded the desired 1,5-CH
insertion precursor 8.
room temperature,7 and to our pleasure the desired
spirocyclic product 4 was produced as the major
product in good isolated yield (Scheme 2).8 The success
of this reaction meant that we were now in a position to
effect the desired Heck-cyclisation reaction. After con-
sidering a number of alternatives, the palladacycle 99
seemed like the most appropriate choice of catalyst for
this transformation. Thus, after heating a solution of 4
(MeCN/DMF/H2O (5:5:1))5a and 9 (10 mol%) at 120°C
for 2 h, we were able to isolate the desired tetracyclic
product 3 and its exocyclic double bond isomer 10 (6:1
ratio, respectively) in a combined yield of 25–30%.
Although we were initially pleased with this result, we
were a little concerned to find that the enamine 11
(41%) was formed as the major new product under the
reaction conditions (Scheme 2).10
We believe that the formation of 11 can be explained by
invoking an intramolecular palladium-catalysed redox
process (Scheme 3). The process begins with an initial
oxidative addition to produce the expected
organometallic species, but instead of undergoing the
usual carbo-palladation of the adjacent olefin, the oxi-
dised palladium species forms a chelate 12 with the
basic nitrogen atom of the spirocycle. A tautomerisa-
tion of this complex results in the production of an
iminium ion 14, which can lose HBr to form 13. A final
reductive elimination affords the observed enamine 11
and regenerates the palladium catalyst, thus completing
the catalytic cycle.11 A number of other palladium
catalysts (e.g. Pd(PtBu3)2/MeNCy2,12a Pd(OAc)2/dppp/
K2CO3)12b were tested for their ability to perform the
desired Heck reaction, but in general these were found
to result in lower conversions and the enamine 11 was
produced as the major new compound in all cases.
With the desired cyclisation precursor 8 in hand, we
were now ready to examine the two remaining CꢀC
bond formations necessary for tetracycle construction.
Firstly, the alkylidene carbene 1,5-CH insertion reac-
tion was examined. Thus, a diethyl ether solution of 8
was treated with an excess (2.0 equiv.) of KHMDS at
a-d
CO2Me
N
N
Boc
Boc
Br
Br
7
5
e,f
From these results, it became clear that we needed to
modify our synthetic plan to take into account these
new findings. The most obvious way of preventing the
formation of the enamine 11 would be to use an amide
18, rather than an amine-derived cyclisation precursor,
and the synthesis of this material became our next goal.
g
N
N
Br
4
Br
Ar
O
P
Pd
Ar
8
O
Pd
O
h
P
O
Ar
Ar
H
9
oxidative
N
N
N
N
N
addn
Br
Pd
+
+
Br
H
H
12
4
tautomerisation
3
25%
10
4%
11
41%
H
N
Scheme 2. Reagents and conditions: (a) Dibal-H, CH2Cl2,
−78°C (72%); (b) Ph3PCHC(O)Me, CH2Cl2, rt (80%); (c) H2,
Pd/C, EtOAc (92%); (d) KHMDS, (Ph3PCH2Br)Br, THF
(70%); (e) TFA (30 equiv.), DCM, rt (100%); (f) o-Br-
PhCH2CHO, THF, HOAc, NaBH(OAc)3 (70%); (g)
KHMDS, Et2O, rt (64%); (h) MeCN:DMF:H2O (5:5:1), cata-
lyst 9 (10 mol%), 120°C.
N
reductive
elimn
-HBr
11
Pd
H
Pd
Br
H
13
14
Scheme 3.

S. M. Worden et al. / Tetrahedron Letters 43 (2002) 6011–6014
6013
We first tried to access 18 from the amide 15 via a
1,5-CH insertion reaction, and although 15 proved
relatively easy to prepare, we were unable to effect the
BCDE-tetracyclic core structure 19 in 62% yield. In
addition, we have shown that a related tertiary amine-
derived Heck reaction precursor 4 leads to the forma-
tion of the BCDE-tetracyclic core as a mixture of
regioisomers (3 and 10) in only modest yield. During
this reaction the enamine 11 was formed as the major
reaction product via an interesting palladium-catalysed
pathway. From this initial proof of concept study, we
can see that our proposed route to (−)-cephalotaxine 1
should be feasible from an amide-derived substrate
similar to 18, although some optimisation of its synthe-
sis will be necessary. These and further studies in this
area will be published in due course.
desired KHMDS-mediated cyclisation reaction.13
A
successful synthesis of 18 was however completed from
7 via the three-step sequence shown in Scheme 4.
Firstly, KHMDS induced 1,5-CH insertion of 7 pro-
duced the spirocycle 16 (>93% e.e.) and deprotection of
the Boc group afforded the corresponding secondary
amine product 17. Purification of this material proved
difficult and it was used in crude form in subsequent
reactions. Finally, the synthesis of 18 was completed,
albeit in modest overall yield, using standard amide
forming conditions.14 Pleasingly, cyclisation of 18 medi-
ated by 10 mol% of the catalyst 9 afforded the desired
tetracyclic amides 19 and 20 as the major products of
the reaction as a 3:1 mixture in 54% yield. The isolated
yield of tetracyclic product could be increased to 62% if
Pd(PtBu3)2 was used as catalyst in the presence of
MeNCy2, but in this case only the endocyclic olefin
isomer 19 was observed.
Acknowledgements
The authors thank Dr. Geo Adam for his interest in
this work and Roche Products (Welwyn), F. Hoff-
mann-LaRoche AG (Basel), the EPSRC (studentship
for S.W.), the University of Botswana (studentship for
R.M.) and the School of Chemistry, University of
Nottingham, for financial support.
In summary, we have shown that the tetracyclic core of
the cephalotaxus alkaloids can be accessed from proline
using an alkylidene carbene 1,5-CH insertion to con-
struct the DE spirocyclic ring system. We have also
shown that a Heck-type cyclisation of an amide-derived
precursor 18 leads to the formation of the desired
References
1. (a) Wall, M. E. J. Am. Pharm. Assoc. 1954, 43, 505; (b)
Paudler, W. W.; Kerley, G. I.; McKay, J. J. Org. Chem.
1963, 28, 2194; (c) Powell, R. G.; Weisleder, D.; Smith,
C. R., Jr.; Wolff, I. A. Tetrahedron Lett. 1969, 9, 4081;
(d) Abraham, D. J.; Rosenstein, R. D.; McGandy, E. L.
Tetrahedron Lett. 1969, 9, 4081; (e) Paudler, W. W.;
McKay, J. J. Org. Chem. 1973, 38, 2110.
N
a, b
N
Br
O
Boc
Br
Br
7
15
2. Powell, R. G.; Weisleder, D.; Smith, C. R., Jr.; Rohwed-
der, W. K. Tetrahedron Lett. 1970, 10, 815.
c
c
3. O’Brien, S.; Kantarjian, H.; Koller, C.; Feldman, E.;
Beran, M.; Andreeff, M.; Giralt, S.; Cheson, B.; Keating,
M.; Freireich, E.; Rios, M. B.; Talpaz, M. Blood 1999,
93, 4149–4153.
4. (a) Ikeda, M.; El Bialy, S. A. A.; Hirose, K.; Kotake, M.;
Sato, T.; Bayomi, S. M. M.; Shehata, I. A.; Abdelal, A.
M.; Gad, L. M.; Yakura, T. Chem. Pharm. Bull. 1999, 47,
983; (b) Koseki, Y.; Sato, H.; Watanabe, Y.; Nagasaka,
T. Org. Lett. 2002, 4, 885; (c) Ikeda, M.; Hirose, K.-I.; El
Bialy, S. A. A.; Sato, T.; Yakura, T.; Bayomi, S. M. M.
Chem. Pharm. Bull. 1998, 46, 1084; (d) For an excellent
review, see: Miah, M. A. J.; Hudlicky, T.; Reed, J. W.
The Alkaloids 1998, 51, 199.
O
e
N
(from 17)
N
Br
R
16, R=Boc
17, R=H
18
d
f or g
O
N
O
N
5. (a) Tietze, L. F.; Schirok, H. J. Am. Chem. Soc. 1999,
121, 10264; (b) Nagasaka, T.; Sato, H.; Saeki, S.-I.
Tetrahedron: Asymmetry 1997, 8, 191; (c) Isono, N.;
Mori, M. J. Org. Chem. 1995, 60, 115.
+
H
H
19
20
6. (a) Gabaitsekgosi, R.; Hayes, C. J. Tetrahedron Lett.
1999, 40, 7713; (b) Green, M. P.; Prodger, J. C.; Sher-
lock, A. E.; Hayes, C. J. Org. Lett. 2001, 3, 3377.
7. (a) Taber, D. F.; Christos, T. E.; Neubert, T. D.; Batra,
D. J. Org. Chem. 1999, 64, 9673; (b) Taber, D. F.;
Neubert, T. D. J. Org. Chem. 2001, 66, 143; (c) Taber, D.
F.; Meagley, R. P.; Doren, D. J. J. Org. Chem. 1996, 61,
5723.
Scheme 4. Reagents and conditions: (a) MeOH, AcCl, rt; (b)
o-Br-PhCH2CO2H, EDCI, HOBt, NMM, DCM (80%, two
steps); (c) KHMDS, Et2O, rt (60–65%); (d) MeOH, AcCl, rt
or TFA, CH2Cl2; (e) o-Br-PhCH2CO2H, EDCI, HOBt,
NMM, DCM (29%, two steps); (f) MeCN:DMF:H2O (5:5:1),
catalyst 9 (10 mol%), 120°C (41% 19 and 13% 20); (g)
Pd(PtBu3)2, NMeCy2, dioxane, 100°C (62% 19 only).

6014
S. M. Worden et al. / Tetrahedron Letters 43 (2002) 6011–6014
8. The spirocyclic amine 4 has an [h]D=−29 (c 1.06,
Buchwald, S. L. Angew. Chem., Int. Ed. Engl. 1995, 34,
1348 and references cited therein.
12. (a) Littke, A. F.; Fu, G. C. J. Am. Chem. Soc. 2001,
123, 6989; (b) Vallain, K. S. A.; Larhed, M.; Hallberg,
A. J. Org. Chem. 2001, 66, 4340.
13. Compound 15 contains a particularly acidic methylene
moiety, and we believe that initial deprotonation at this
site by KHMDS results in the subsequent 1,1-elimina-
tion reaction being retarded.
CHCl3) but we have been unable to find a suitable assay
to determine the % enantiomeric excess. Previous studies
from our research group, however, have shown that
alkylidene carbene 1,5-CH insertions of this type pro-
ceed with retention of stereochemistry (see Ref. 6).
8
9. Herrmann, W. A.; Brossmer, C.; Ofele, K.; Reisinger,
C.-P.; Priermeier, T.; Beller, M.; Fischer, H. Angew.
Chem., Int. Ed. Engl. 1995, 34, 1844.
14. We believe that the modest overall yield observed for
the formation of 18 from 16 is due to decomposition of
the intermediate spirocyclic secondary amine 17 under
the acidic deprotection conditions. Control experiments
show that samples of the ammonium salt of 17 decom-
pose over time, even at −20°C. We are currently examin-
ing the use of different protecting groups for the
secondary amine that can be removed under neutral or
basic conditions.
10. Tietze et al. showed that a similar precursor to 4 could
be cyclised with Herrmann’s catalyst to produce a tetra-
cyclic product in 81% yield (see Ref. 5a). We believe
that steric interaction between the palladium catalyst
and the methyl substituent on the olefin in 4 facilitates
preferential coordination to the tertiary amine rather
than the trisubstituted olefin.
11. b-Hydride elimination of palladium–amine complexes is
a known process. See: Guram, A. S.; Rennels, R. A.;