Total synthesis of (-)-raumacline

Article abstract of DOI:10.1039/b310274b

The total synthesis of (-)-raumacline from L-tryptophan was achieved, featuring a cis-specific Pictet-Spengler reaction, a stereoselective Dieckmann cyclization, and an epimerization step that allowed complete stereocontrol of five chiral centres.

Full text of DOI:10.1039/b310274b

Total synthesis of (2)-raumacline
Patrick D. Bailey, Paul D. Clingan, Timothy J. Mills, Richard A. Price and Robin G. Pritchard
Department of Chemistry, UMIST, PO Box 88, Manchester, UK M60 1QD. E-mail: p.bailey@umist.ac.uk;
Fax: (0)161-200-4541; Tel: (0)161-200-4448
Received (in Cambridge, UK) 26th August 2003, Accepted 19th September 2003
First published as an Advance Article on the web 14th October 2003
The total synthesis of (2)-raumacline from
L
-tryptophan
electron-withdrawing nitrile groups of 5 in order to control the
stereochemistry, but early stages in the synthesis proceeded
with disappointingly poor cis stereoselectivity of only 3 : 1
(Scheme 3), and required extensive chromatography at an early
stage.
was achieved, featuring a cis-specific Pictet–Spengler reac-
tion, a stereoselective Dieckmann cyclization, and an
epimerization step that allowed complete stereocontrol of
five chiral centres.
We therefore returned to our original route,6h improving the
procedures for the preparation of the advanced intermediates
14a/b (Scheme 4). Two features of the improved synthesis are
Raumacline 1 was first identified in 1990;1 it is a bridged indole
alkaloid possessing six chiral centres, and is a member of the
ajmaline family of alkaloids (see Scheme 1). The important
medicinal properties of the indole alkaloids were highlighted in
a review in 1988 that contained 359 references,2 and the
ajmaline family have long been used for the treatment of
cardiovascular irregularities.3,4 We have developed stereose-
lective routes to cis 1,3-disubstituted tetrahydro-b-carbolines
especially noteworthy. Firstly, the initial three steps from -
L
tryptophan 9 yielded crystalline cyano-sulfonamide,9 and this
could be prepared in pure form without the need for any
chromatography; however, we found that the sodium/liquid
ammonia deprotection was capricious unless we ran the reaction
in THF,10 which provided the amino-nitrile 10 in pure form.
Secondly, the kinetically controlled Pictet–Spengler reaction5
yielded only the cis stereoisomer, and we initially assumed that
we had simply failed to observe the minor trans isomer. Careful
examination of the crude NMR spectra revealed that the
reaction was virtually 100% stereoselective; this almost unique
observation prompted us to undertake a study of related
reactions, and we have now been able to extend the range of cis
specific Pictet–Spengler reactions.11 Routine alkylations of 11,
followed by removal of the silyl protection and Swern oxidation
provided the cyano-aldehyde 13, from which the Michael
acceptors 14a and 14b were prepared using Wittig or Horner–
Wadsworth–Emmons methodology respectively.
starting from
-tryptophan,5 and leading to a number of natural
L
product syntheses,6a–h although more complex targets such as
raumacline had eluded us until now.
Cook’s group has completed the only total synthesis of
(2)-raumacline,7 but their approach required the use of more
expensive -tryptophan as the starting material (using a key
D
epimerization/cyclization step to construct the 1,3-bridge), and
also involve recycling of material in an oxy-Cope rearrange-
ment step that generates several isomers. In this communica-
tion, we describe a different approach that provides access to the
ajmaline family of alkaloids from
-tryptophan,8 with complete
L
stereochemical control of five chiral centres, and we report the
total synthesis of (2)-raumacline.
Our approach was to identify a protected form of the
dialdehyde 2, which would not only be an advanced inter-
mediate in the synthesis of (2)-raumacline, but could also
potentially be elaborated to targets such as ajmaline 3 and
suaveoline 4 (Scheme 1). In earlier work towards suaveoline,6f,h
we had found that the dinitrile 5 underwent base induced ring-
closure without any significant stereocontrol (Scheme 2), so this
route was only of limited use. We had planned to modify the
Scheme 3 Formation of 8 via the Pictet–Spengler reaction of 7.
Scheme 1 RSA reveals 2 as an advanced intermediate for 1b, 3 and 4.
Scheme 4 a) LiAlH4 (98%); b) TsCl, py (78%); c) KCN (86%); d) Na (3
eq.), NH3(l), 30 min, THF (88%); e) TBSOCH2CH2CHO, 3 Å mol. sieves,
24 h; f) CH2Cl2, TFA (2 eq.), 278 °C ? RT, 6 h (80%); g) BnBr, 70 °C
(75%); h) MeI, NaH, 0 °C (87%); i) TBAF (96%); j) Swern (100%); k)
Ph3PNCHCO2Me (82%, 14a, R = H and RA = Me, E/Z = 4 : 1) or
(EtO)2POCHEtCO2Et, NaH (65%, 14b, R = Et and RA = Et, E/Z = 5 :
3).
Scheme 2 Dieckmann cyclization of dinitrile 5.
CHEM. COMMUN., 2003, 2800–2801
2800
This journal is © The Royal Society of Chemistry 2003

View Online
key stereoselective ring-closure step, but also retains an
epimerizable centre that allows us to exploit the thermodynamic
stability of the raumacline structure (see Fig. 2). We have
therefore not only completed a total asymmetric synthesis of
(2)-raumacline, but we have also developed a new ster-
eospecific route to the tetracyclic core of the ajmaline family of
indole alkaloids.
We thank Dr Ian Fleet & Mr Alan Morissey for assistance
with mass spectra and NMR, and the EPSRC for financial
support.
Notes and references
1 L. Polz, J. Stöckigt, H. Takayama, N. Uchida, N. Aimi and S. Sakai,
Tetrahedron Lett., 1990, 31, 6693.
2 Y. Ban, Y. Murakami, Y. Iwasawa, M. Tsuchiya and N. Takano, Med.
Chem. Rev., 1988, 8, 231.
Scheme 5 a) LiNEt2,THF, 278 °C (15a, 63%; 15b, 99%); b) LiBH4 (16a,
75%, 16b, 84%); c) TsOH·H2O, THF, reflux (17a, 63%; 17b, 88%); d)
DIBAL-H (54%, R = H; 50%, R = Et); e) H2/Pd–C (100%, R = H or R
= Et). [a, R = H, RA = Me; b, R = RA = Et].
3 J. Stöckigt, GIT Fachz. Lab., 1988, 32, 608.
4 A.-U.-H. Gilani, Phytother. Res., 1998, 12, S66.
We next focused on the deethyl analogue 1a of raumacline,
starting from 14a. Base induced Michael reaction generated the
bridged compound 15a, with total control of the C-15 chiral
centre, but only 2 : 1 selectivity for the a-stereochemistry at C-
16; the stereochemistry of the a-isomer 15a (shown in Scheme
5) was assigned from NMR data,12 whilst the X-ray crystal
structure of the minor b-isomer is shown in Fig. 1.13
5 P. D. Bailey, S. P. Hollinshead, N. R. McLay, K. M. Morgan, S. J.
Palmer, S. N. Prince, C. D. Reynolds and S. D. Wood, J. Chem. Soc.,
Perkin Trans. 1, 1993, 431.
6 (a) P. D. Bailey, P. J. Cochrane, K. Lorenz, I. D. Collier, D. P. J. Pearson
and G. M. Rosair, Tetrahedron Lett., 2001, 42, 113; (b) P. D. Bailey and
N. R. McLay, Tetrahedron Lett., 1991, 32, 3895; (c) P. D. Bailey and N.
R. McLay, J. Chem. Soc., Perkin Trans. 1, 1993, 441; (d) P. D. Bailey,
S. P. Hollinshead, N. R. McLay, J. H. Everett, C. D. Reynolds, S. D.
Wood and F. Giordano, J. Chem. Soc., Perkin Trans. 1, 1993, 451; (e)
P. D. Bailey, I. D. Collier, S. P. Hollinshead, M. H. Moore, K. M.
Morgan, D. I. Smith and J. M. Vernon, J. Chem. Soc., Chem. Commun.,
1994, 1559; (f) P. D. Bailey and K. M. Morgan, Chem. Commun., 1996,
1479; (g) P. D. Bailey, I. D. Collier, S. P. Hollinshead, M. H. Moore, K.
M. Morgan, D. I. Smith and J. M. Vernon, J. Chem. Soc., Perkin Trans.
1, 1997, 1209; (h) P. D. Bailey and K. M. Morgan, J. Chem. Soc., Perkin
Trans. 1, 2000, 3578.
However, it transpired that the a/b stereochemical problem
could be resolved later in the synthesis. Lithium borohydride
reduction of either of the epimers gave the cyano-alcohols 16a,
for which derivatisation of the a-isomer as the Mosher’s ester14
allowed us to confirm the optical integrity.15 Pleasingly, by
refluxing EITHER the a- or b-epimer of 16a with tosic acid
hydrate in THF, trans decalin pentacycle 17a was isolated as the
ONLY stereoisomer; the b-epimer presumably initally cyclized
to the cis decalin, which epimerised to the more stable trans
isomer via tautomerisation of the lactone or imidate (cf.
raumacline in Fig. 2). Reduction with DIBAL, followed by
catalytic hydrogenolysis,7 generated deethyl raumacline 1b,
with complete control of all five chiral centres.
Finally, we repeated the synthetic sequence in Scheme 5, but
starting from 14b. In this case, cyclization via Michael reaction
generated 15b in essentially quantitative yield, with a/b
selectivity of up to 4 : 1, but composed of a 1 : 1 mixture of C-18
epimers. The latter were separable once the lactone 17b had
been formed, allowing us to complete the synthesis of
(2)-raumacline 1b. Our approach not only exploits a cis
specific Pictet–Spengler reaction early in the synthesis, and a
7 X. Fu and J. M. Cook, J. Am. Chem. Soc., 1992, 114, 6910.
8
L
-Tryptophan is about one third the price of
D-tryptophan, and the L-
isomer is the biosynthetic building block for the indole alkaloids.
9 Based on early steps in the synthesis of dregamine: J. P. Kutney, G. K.
Eigenforf, H. Matsue, A. Murai, K. Tanaka, W. L. Sung, K. Wada and
B. R. Worth, J. Am. Chem. Soc., 1978, 100, 938.
10 G. T. Bourne, D. Crich, J. W. Davies and D. C. Horwell, J. Chem. Soc.,
Perkin Trans. 1, 1991, 1693.
11 In general, tryptophan allyl ester reacts with aromatic aldehydes to yield
only the cis diastereoisomer, when the Pictet–Spengler reaction is
carried out under conditions of kinetic control (loc. cit.); manuscript in
preparation. As far as we are aware, the only other example of a cis
specific Pictet–Spengler reaction was reported by: G. Massiot and T.
Mulamba, J. Chem. Soc., Chem. Commun., 1984, 715.
12 NMR data for 15a: the a-epimer showed 12.1 Hz coupling between H-
15 and H-16, and the only NOE from H-16 was to one of the H-14
protons; cf. for the b-epimer, H-16 showed an NOE to H-15, but there
was only a small coupling between these protons.
13 X-Ray data for 15a. 0.25 3 0.15 3 0.07 mm3, monoclinic, space group
P21, a
= 10.3050(4), b = 7.2730(3), c = 14.2530(8) Å, b =
90.388(2)°, V = 1068.21(9) Å3, Z = 2, rcalcd = 1.286 Mg m23, 2qmax
= 54.98°, molybdenum radiation, l = 0.71073 Å, k CCD f and w scans
to full Ewald sphere, temperature = 150(2) K, number of measured/
independent reflections 9756/4320 [R(int)
= 0.0465], number of
reflections included in the refinement 4320 [3530I > 2s(I)], Lorentz
and polarization corrections were performed, absorption correction
semi-empirical from equivalents (m
=
0.082 mm21, max./min.
transmission 0.9943/0.9797). The structure was solved by direct
methods using SHELXS-97 (G. M. Sheldrick, SHELXS-97, Program
for crystal structure solution, 1997, University of Göttingen, Germany),
refinement method full matrix least squares on F2 using SHELXL-97
(G. M. Sheldrick, SHELXL-97, Program for crystal structure refine-
ment , 1997, University of Göttingen, Germany), no. of parameters =
361, H atoms were subjected to isotopic refinement, final residuals
refined against ¡F2¡ were wR2 = 0.1084 (all data), R1 = 0.0509 [I >
Fig. 1 X-Ray crystal structure of the b-epimer of 15a.
2s(I)], max. and min. residual electron density 0.20 and 20.20 e Å23
,
and the absolute configuration of 15a was not determined. CCDC
207148. See http://www.rsc.org/suppdata/cc/b3/b310274b/ for crys-
tallographic data in CIF or other electronic format.
14 J. A. Dale and H. S. Mosher, J. Am. Chem. Soc., 1973, 95, 512.
15 The Mosher’s esters were prepared as described in the literature,14 with
distinctive methoxy singlets at d 3.03 for the derivative with R-acid, and
d 3.25 with the S-acid; in neither case could we observe any signal due
to the other diastereoisomer.
Fig. 2 The predicted 3D structure of (2)-raumacline 1b, showing the all-
equatorial cis decalin conformation.
CHEM. COMMUN., 2003, 2800–2801
2801