Total synthesis of the cytotoxic guaipyridine sesquiterpene alkaloid (+)-cananodine

Article abstract of DOI:10.1002/ejoc.200600414

The enantiospecific total synthesis of the cytotoxic guaipyridine sesquiterpene alkaloid (+)-cananodine (1) is described. A chiral pool/chiral auxiliary based approach is described for the synthesis of a key oxazolidinone intermediate. Subsequent key step

Full text of DOI:10.1002/ejoc.200600414

SHORT COMMUNICATION
DOI: 10.1002/ejoc.200600414
Total Synthesis of the Cytotoxic Guaipyridine Sesquiterpene Alkaloid
(+)-Cananodine
Donald Craig*[a] and Gavin D. Henry[a]
Keywords: C–C coupling reactions / Catalysis / Claisen rearrangement / Microwave-assisted synthesis / Pyridine synthesis /
Ring-closing metathesis
The enantiospecific total synthesis of the cytotoxic guaipyrid-
ine sesquiterpene alkaloid (+)-cananodine (1) is described. A
chiral pool/chiral auxiliary based approach is described for
the synthesis of a key oxazolidinone intermediate. Subse-
quent key steps involve diastereoselective oxazolidinone al-
lylation, cycloheptenylmethanol formation using ring-closing
olefin metathesis, microwave-assisted decarboxylative
Claisen rearrangement reaction, and use of a novel pyridine-
forming method.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
Cananodine is a naturally occurring guaipyridine sesqui- of the requisite allylic electrophile (Scheme 1). In light of
terpene alkaloid which shows sub-micromolar activity this analysis, we chose instead to use the decarboxylative
against Hep G2 human hepatocarcinoma cell lines. It was Claisen rearrangement (dCr) reaction recently uncovered in
isolated from the perfumery tree Cananga odorata (ylang our laboratory.[5] This is a new, catalytic variant of the clas-
ylang) in 2001 by workers at the Kaohsiung Medical Uni- sical Ireland–Claisen rearrangement, in which silyl ketene
versity in Taiwan, who assigned the structure 1 using NMR acetals generated in situ from allylic α-tosyl esters under
and MS techniques.[1] Cananodine is structurally related to weakly basic conditions undergo [3,3]-sigmatropic re-
β-patchoulene (2),[2] and patchouli pyridine (3).[3]
arrangement followed by acetate-induced decarboxylation
We have recently initiated a major new programme for to give γ,δ-unsaturated sulfones directly.[6] Unlike most (π-
the development of methods for the de novo synthesis of allyl)metal-based methods, this sigmatropic rearrangement
polysubstituted arenes from unsaturated, acyclic precursors. mediated approach enables regiospecific allylation α to the
Part of this strategy is based on the recognition that 1,5- arylsulfonyl moiety. This communication describes the ap-
dicarbonyl compounds bearing a C-3 leaving group are plication of the dCr reaction in conjunction with our new
suitably disposed to enter into redox-neutral condensation pyridine chemistry in the first total synthesis of (+)-canano-
reactions with ammonia, to provide pyridines. We have dine.
demonstrated the effectiveness of this approach for the syn-
The retrosynthetic analysis of 1 is depicted in Scheme 2.
thesis of 2,4,6-trisubstituted pyridines, in which the requi- Diene 4 (R = CO2Me) was envisaged to be the product
site 1,5-dicarbonyl substrates possessing a 4-tolylsulfonyl of dCr reaction of the substituted (4-tolylsulfonyl)acetate 5,
substituent at C-3 were synthesised by oxidative cleavage of which indicated the substituted cycloheptenylmethanol 6 as
1,6-dienes, using (allyl)Pd0 chemistry (Scheme 1).[4] We were a key intermediate. This would be synthesised by ring-clos-
keen to apply the new pyridine-forming method to the syn- ing olefin metathesis of the diene 7, the product of dia-
thesis of cananodine, because it would demonstrate the stereoselective allylation of the N-acyloxazolidinone 8.
compatibility of the method with a multi-functional sub-
Compound 8 could be prepared in quantities viable for
strate. However, we were mindful that the 2,3,6-trisubstitu- the planned total synthesis by the sequence depicted in
tion pattern present in the natural product was incompat- Scheme 3. Straightforward regioselective oxidative cleav-
ible with the palladium(0)-based approach employed during age[7] of (R)-(–)-citronellene with reductive workup, fol-
method development,[4] because it would require access to a lowed by 4-tolylsulfonylation and cyanide displacement
1,6-diene 4 in which one of the required C–C bond-forming gave the nitrile 9, which upon hydrolysis, acid chloride for-
reactions had taken place at the more substituted terminus mation and condensation with the lithio salt of the (S)-vali-
nol-derived oxazolidinone gave 8.[8]
With the N-heptenoyloxazolidinone 8 in hand, attention
[a] Department of Chemistry, Imperial College London,
was turned to its diastereoselective allylation (Scheme 4).
The requisite allylic bromide 12 was straightforwardly pre-
pared from commercially available 2-methylenepropane-1,3-
diol by NaH-mediated monosilylation[9] using TBDPSCl,
South Kensington Campus, London SW7 2AZ, UK
Fax: +44-20-7594-5868
E-mail: d.craig@imperial.ac.uk
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
3558
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 3558–3561

Total Synthesis of (+)-Cananodine
SHORT COMMUNICATION
Scheme 1.
Scheme 2.
alcohol 7 with Grubbs’ catalyst gave the cycloheptenylme-
thanol 6 in 88% yield over the two steps from 10. Removal
of the oxazolidinone auxiliary was accomplished at this
stage by methanolysis under basic conditions, after which
esterification of the primary alcohol with 4-methyl-2-(4-tol-
ylsulfonyl)-4-pentenoic acid[10] completed in 77% over two
steps the assembly of dCr substrate 5 (Scheme 4).
Scheme 3.
The stage was now set to assess the effectiveness of the
followed by 4-tolylsulfonylation and SN2 reaction with lith- dCr reactions of substrate 5. Initial studies carried out using
ium bromide in diethyl ether. Addition of this bromide to conventional thermal conditions were discouraging; treat-
the lithium enolate of 8 in THF/DMPU at –78 °C to –20 °C ment of 5 with N,O-bis(trimethylsilyl)acetamide (BSA) and
resulted in diastereoselective allylation, providing the 1,7- potassium acetate in toluene under reflux failed to yield any
diene 10 in 90% yield. Ring-closing olefin metathesis of the desired rearrangement/decarboxylation product. In-
(RCM) could be carried out at this stage; cycloheptene 11 stead, 3-methyl-1-(4-tolylsulfonyl)-3-butene was formed as
was formed in virtually quantitative yield upon reaction of the sole identifiable product in 52% yield, presumably by
10 in the presence of 5 mol-% of Grubbs’ second-generation hydrolysis of the 4-tolylsulfonyl-substituted ester followed
catalyst in dichloromethane. However, problems were en- by decarboxylation of the product 4-methyl-2-(4-tolylsulfo-
countered with the subsequent desilylation: exposure of 11 nyl)-4-pentenoic acid. In marked contrast, substrate 5
both to acidic (methanolic HCl) and to basic (TBAF) con- underwent highly stereoselective dCr reaction under micro-
ditions resulted in significant decomposition, and we had wave-irradiation conditions to give the 1,6-diene 4 in 44–
been concerned beforehand that basic reagents would lead 71% yield as a mixture of two of the four possible dia-
to erosion of de by epimerisation of the stereocentre α to stereoisomers. Although we have not determined the stereo-
the acyl moiety. These problems led to the adoption of a chemistry of the two products, structures 4 are proposed
sequence modified simply by reversing the order of the based on the assumption that formation of the new C–C
RCM and deprotection steps. Thus, treatment of 10 with bond takes place anti to the vicinal methyl substituent.
methanolic HCl, and reaction of the product acyclic allylic Ozonolysis of the dienes 4 with sequential addition of tri-
Eur. J. Org. Chem. 2006, 3558–3561
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3559

D. Craig, G. D. Henry
SHORT COMMUNICATION
Scheme 4.
Scheme 5.
phenylphosphane and ethanolic ammonia gave the pyridine the possibility that optically impure (–)-(R)-citronellene
13 in 48% yield. Finally, exposure of 13 to excess methyl- gave rise to the discrepancy in optical behaviour, because
magnesium bromide gave (+)-cananodine (1) in almost (S)-valine was the ultimate source of the second of the two
quantitative yield. The completion of the synthesis of 1 is stereocentres present in the target structure; any erosion of
depicted in Scheme 5.
Compound 1 prepared as described above gave H and resulted in the formation of mixtures of diastereomers.
13C NMR spectra which are identical to ones kindly pro-
In summary, the first total synthesis of (+)-cananodine
the optical purity of either chiral pool source would have
1
vided by Professor Yang-Chang Wu of Kaohsiung Medical has been achieved using new Claisen rearrangement and
University.[11] Professor Wu and co-workers obtained their pyridine chemistry developed within our laboratory. Cur-
data from the isolated, natural sample; we have been unable rent studies are underway to apply this powerful combina-
to locate a source of the required plant material. However, tion of methods to other natural and non-natural pyridine-
the optical rotation value obtained by us {[α]2D1 = +17.9 (c containing structures. The results of these investigations will
= 1.34, CHCl3)} differs significantly from that obtained by be reported in due course.
Professor Wu’s team {[α]2D5 = –76.2 (c = 0.06, CHCl3)}; of
Supporting Information (see footnote on the first page of this arti-
particular concern to us was the different sign of the two
specific rotation values. From the quoted specific rotation,
concentration (0.6 mgmL–1) and path length (0.1 dm)[12] of
the solution of natural 1 on which the optical activity was
measured by Professor Wu’s team, a measured rotation of
–0.0046° may be inferred. Our rotation, measured on a
solution having a concentration of 13.4 mgmL–1 with a
path length of 0.25 dm was +0.06°, and we suggest that the
greater order of magnitude of our value makes it signifi-
cantly less susceptible to experimental error.
cle): Full experimental details and spectroscopic information are
provided for compounds 1 and 4–13.
Acknowledgments
We thank EPSRC and Aventis CropScience (supported Quota Stu-
dentship to G. D. H.) for financial support of this research. We
extend our gratitude to Professor Yang-Chang Wu, Kaohsiung
Medical University, Taiwan for valuable discussions and for provid-
ing copies of NMR spectra of authentic cananodine.
In spite of the conflicting optical rotation values de-
scribed above, the results described herein demonstrate that
an unambiguous synthesis of structure 1 has been achieved:
the spectroscopic homogeneity of late-stage material gener-
ated using the (–)-(R)-citronellene-based route shows that
intermediates 4, 5, 6, 7, 10 and 13 together with cananodine
1 are formed as single stereoisomers. This clearly rules out
[1] T.-J. Hsieh, F.-R. Chang, Y.-C. Chia, C.-Y. Chen, H.-F. Chiu,
Y.-C. Wu, J. Nat. Prod. 2001, 64, 616–619.
[2] G. Büchi, R. E. Erickson, N. Wakabayashi, J. Am. Chem. Soc.
1961, 83, 927–938.
[3] G. Büchi, I. M. Goldman, D. W. Mayol, J. Am. Chem. Soc.
1966, 88, 3109–3113; for previous synthetic studies of guaipyri-
dines, see A. Van der Gen, L. M. Van der Linde, J. G. Witte-
3560
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 3558–3561

Total Synthesis of (+)-Cananodine
SHORT COMMUNICATION
veen, Recl. Trav. Chim. Pays-Bas 1972, 91, 1433–1440; T. Okat-
ani, J. Koyama, K. Tagahara, Y. Suzuta, Heterocycles 1987, 26,
595–597; J. Koyama, T. Okatani, T. Teruyo, K. Tagahara, S.
Kiyoshi, Y. Suzuta, H. Irie, Heterocycles 1987, 26, 925–927.
[4] D. Craig, G. D. Henry, Tetrahedron Lett. 2005, 46, 2559–2562.
[5] a) D. Bourgeois, D. Craig, N. P. King, D. M. Mountford, An-
gew. Chem. Int. Ed. 2005, 44, 618–621; b) D. Craig, F. Grelle-
pois, Org. Lett. 2005, 7, 463–465; c) D. Craig, F. Grellepois,
A. J. P. White, J. Org. Chem. 2005, 70, 6827–6832; d) D. Craig,
N. P. King, J. T. Kley, D. M. Mountford, Synthesis 2005, 3279–
3282; e) D. Bourgeois, D. Craig, F. Grellepois, D. M. Mount-
ford, A. J. W. Stewart, Tetrahedron 2006, 62, 483–495; for a
two-step, two-pot sequence involving classical LDA/Me3SiCl-
mediated Ireland–Claisen rearrangement followed by decar-
boxylation, see: f) A. H. Davidson, N. Eggleton, I. H. Wallace,
J. Chem. Soc., Chem. Commun. 1991, 378–380.
7.5 Hz, 1 H, CH alkene), [4.96 (ddd, J = 17.0, 2.0, 1.0 Hz, 1
H), and 4.93 (ddd, J = 11.5, 2.0, 1.0 Hz, 1 H), CH2 alkene×2],
4.43 (ddd, J = 8.5, 3.5, 3.5 Hz, 1 H, CHN), [4.26 (dd, J = 9.0,
8.5 Hz, 1 H), and 4.20 (dd, J = 9.0, 3.0 Hz, 1 H), CH2O×2],
2.43–2.33 [m, 1 H, CH(CH3)2], 2.20–2.11 (1 H, m, δ amide),
1.67–1.62 (m, 2 H, β amide), 1.40–1.32 (q, J = 8.0 Hz, 2 H, γ
amide), 1.00 (d, J = 7.0 Hz, 3 H, CH3), [0.92 (d, J = 7.0 Hz, 3
H), and 0.88 (d, J = 7.0 Hz, 3 H), Me of isopropyl×2]. 13C
NMR (75 MHz): δ = 173.3 (C amide), 154.1 (C carbamate),
144.3 (CH alkene), 112.9 (CH2 alkene), 63.3 (OCH2), 58.4
(NCH), 37.7 (CH δ amide), [35.9, 35.6 (CH2 α, γ amide×2)],
28.4 [CH(CH3)2], 22.2 (CH2 β amide), 20.2 (CH3 Me), [18.0
and 14.7 (CH3 of isopropyl×2)]. MS(CI): m/z = 271 [M +
NH4]+, 254 [MH]+, 184, 171, 130, 124, 84, 49; calcd. for
C14H23NO3 [MH]+ 254.1756, found 254.1750.
[9] H.-Y. Lee, H. S. Tae, B. G. Kim, H.-M. Choi, Tetrahedron Lett.
2003, 44, 5803–5806.
[10] 4-Methyl-2-(4-tolylsulfonyl)-4-pentenoic acid was prepared by
hydrolysis (LiOH, H2O) of the corresponding ethyl ester, which
was synthesised by methallylation of ethyl (4-tolylsulfonyl)ace-
tate prepared as described in ref.[4]
[6] For a recent report of the mechanistically distinct Carroll re-
arrangement of an allylic α-phenylsulfonyl ester under highly
basic conditions in the context of vitamin D3 analogue synthe-
sis, see: M. A. Hatcher, G. H. Posner, Tetrahedron Lett. 2002,
43, 5009–5012.
[7] A. G. M. Barrett, R. A. E. Carr, S. V. Attwood, G. Richardson,
N. D. A. Walshe, J. Org. Chem. 1986, 51, 4840–4856.
[11] See the Supporting Information for spectroscopic data for nat-
ural and synthetic 1, and full experimental details and charac-
terisation data for all synthetic intermediates.
[12] Personal communication from Professor Yang-Chang Wu.
Received: May 11, 2006
[8] Spectroscopic data for 8: Rf = 0.29 (20% EtOAc/petroleum
ether). [α]2D4 = +73.9 (c = 6.0, CHCl ); IR (film): ν
2929, 2885, 1734, 1471, 1464, 1363, 1255, 1178, 1097, 837, 779,
= 2954,
˜
3
max
1
669 cm–1. H NMR (400 MHz): δ = 5.68 (ddd, J = 17.5, 10.5,
Published Online: July 3, 2006
Eur. J. Org. Chem. 2006, 3558–3561
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3561