A total synthesis of (±)-codeine by 1,3-dipolar cycloaddition

Article abstract of DOI:10.1002/anie.201007448

Nitrone cycloaddition on a dearomatized bicyclic phenol enabled the facile construction of the correctly configured phenanthrene skeleton of codeine. Further steps yielded allopseudocodeine in a completely diastereoselective manner and finally (±)-codeine by allylic transposition through the hydrolysis of chlorocodides.

Full text of DOI:10.1002/anie.201007448

Communications
DOI: 10.1002/anie.201007448
Natural Product Synthesis
A Total Synthesis of (ꢀ )-Codeine by 1,3-Dipolar Cycloaddition**
Thomas Erhard, Gunnar Ehrlich, and Peter Metz*
Not only due to its potent analgesic effect but also because of
its complex molecular architecture, (ꢁ)-morphine (1) is one
of the most interesting and most thoroughly investigated
alkaloids. Since millennia humans have used opium, which
contains morphine, to alleviate pain. In the search for
morphine derivatives with reduced side effects, total synthesis
offers an attractive opportunity although an economical route
competitive to isolation is currently not within reach.[1,2]
Several efforts in the recent past, for example, the elegant
synthesis by Magnus,[3] the improved strategy of Fukuyama,[4]
the chemoenzymatic approach of Hudlicky,[5] and the inno-
vative route of Stork,[6] express the permanent interest of the
scientific community. Among the numerous strategies for
construction of the phenanthrene skeleton, an approach
based on intramolecular nitrone cycloaddition starting from a
suitably substituted aldehyde seems powerful enough for this
challenging task.[7] Initial investigations in our group focused
on prochiral p-quinol ether 5 as an appropriate dipolarophile
also offering the option for an enantioselective modification
(Scheme 1). Key intermediate 4 already bears the correct
relative configuration at C9 and C14. Completion of the
Scheme 1. Pentacyclic morphine (1) and retrosynthetic analysis of the
carbocyclic framework should be accomplished by Claisen
applied phenanthrene route.
rearrangement[8] which would also install the quaternary
center. After correction of the oxidation state in the side
chain, morphinane core 3 would be completed by trans-
ylide and methanolysis of the intermediate enol ethers,
dimethyl acetal 7 was readily available on a large scale.[4a,b]
A commutable strategy in terms of ordering the coupling and
oxidation events might be attractive for the construction of
the necessary dienone moiety. Suzuki coupling with A[3]
following the Magnus protocol provided biaryl 8 in excellent
yield after basic hydrolysis of the TBS ether.[14] Phenolic
oxidation[15] with PhI(OAc)2 (PIDA) by inverse addition (8
was added slowly to PIDA in methanol) afforded the p-quinol
ether in good overall yield along with a 2,2-dimethoxy-3,5-
cyclohexadienone[16a] as a byproduct. Alternatively, bromine–
lithium exchange, reaction with p-benzoquinone (pBQ)
monoketal B,[17] and Williamson ether synthesis with bisacetal
9 furnished the corresponding methyl ether. After extensive
experimentation, we succeeded in merging the two pathways
by single and double acetal cleavage; the highly acid-labile
substrates were treated with a catalytic amount of CAN in a
buffered MeCN–water mixture at 608C to give 5 in virtually
quantitative yields.[18] The crystal structure[16b] of 5 unambig-
uously proved the isolation of this crucial intermediate.
Intramolecular nitrone cycloaddition, reduction with L-selec-
tride, and silylation under standard conditions yielded the
protected alcohol 10 with complete diastereoselectivity in
excellent yield over three steps. Boron trichloride induced
allylic rearrangement and subsequent solvolysis of allylic
chloride 11[16c] delivered secondary alcohol 4, which set the
stage for installation of the benzylic quaternary center by
thermal Eschenmoser–Claisen rearrangement.
annular alkylation of the secondary amine liberated by
[9]
reductive N O bond cleavage. Allylic substitution[10] could
ꢁ
yield allopseudocodeine and another allylic transposition in
which the hydroxy group is shifted from C8 to C6 would
finally yield codeine (2). With allopseudocodeine in hand, a
formal synthesis is already completed.[11] In addition, a second
formal synthesis is possible by preparation of a carbamate
previously used by Fukuyama.[4a,b]
Our synthesis started from commercially available iso-
vanillin (6, Scheme 2). Hydroxy-directed bromination[12]
followed by blocking of the phenol as a methyl ether[13]
afforded 2-bromoveratrylaldehyde. After chain extension
with the methoxymethyl chloride (MOMCl)-derived Wittig
[*] T. Erhard, Dr. G. Ehrlich, Prof. Dr. P. Metz
Fachrichtung Chemie und Lebensmittelchemie
Professur fꢀr Organische Chemie I, Technische Universitꢁt Dresden
Bergstrasse 66, 01069 Dresden (Germany)
Fax: (+49)351-463-33162
E-mail: peter.metz@chemie.tu-dresden.de
Homepage: http://www.chm.tu-dresden.de/oc1/
[**] We thank Dipl.-Chem. Anne Jꢁger and Dipl.-Chem. Dirk Meyer for X-
ray diffraction analysis and the Federal Institute for Drugs and
Medical Devices, especially the Federal Opium Agency for author-
ization to work with law-regulated narcotics and purchase an
authentic sample of (ꢁ)-codeine.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201007448.
3892
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 3892 –3894

Scheme 2. Facile preparation of the highly functionalized phenanthrene 4. a) Br2, cat. Fe powder, NaOAc, HOAc, RT, 92%; b) aq. KOH, Me2SO4,
508C, 95%; c) Ph3PCH2OMeCl, KOtBu, THF, 08C; d) cat. pTsOH·H2O, MeOH, HC(OMe)3, reflux, 95% (2 steps); e) 1. boronic acid A, K2CO3, 1,4-
dioxane/water (7:3), cat. BHT, 2.6 mol% [Pd2(dba)3], 5.2 mol% PCy3, 808C, 2. aq. NaOH, 508C, 95% (2 steps); f) PIDA, MeOH, HC(OMe)3, RT,
66% + 24% o-quinone dimethylacetal; g) cat. CAN, MeCN/borate buffer (1:1), pH 5–6, 608C, 97% crude; h) 1. BuLi, THF, ꢁ788C, 2. pBQ ketal
B, ꢁ788C!ꢁ708C, 70%; i) NaH, MeI, THF, 308C, 94%; j) cat. CAN, MeCN/borate buffer (1:1), pH 5–6, 608C, 104% crude; k) MeNHOH·HCl,
NaHCO3, MgSO4, MeCN, 08C; l) 1. L-selectride, THF, ꢁ788C, 2. MeOH, ꢁ788C!RT, 3. aq. NaOH, H2O2, RT; m) TBSCl, imidazole, cat. DMAP,
CH2Cl2, RT, 80% (3 steps); n) 1. BCl3, CH2Cl2, ꢁ788C, 2. Et3N, MeOH, ꢁ788C!RT; o) ZnO, acetone/water (4:1), 808C, 76% (2 steps).
BHT=2,6-di-tert-butyl-4-methylphenol, CAN=ceric ammonium nitrate, Cy=cyclohexyl, dba=dibenzylideneacetone, DMAP=4-(N,N-dimethylami-
no)pyridine, L-selectride=lithium tri-sec-butylborohydride, TBS=tert-butyldimethylsilyl, pTs =p-toluenesulfonyl.
Alcohol 4 was converted to amide 12 in very good yield
(Scheme 3). Amide reduction by LiH2N·BH3 furnished
primary alcohol 13 nearly quantitatively.[19] The more com-
monly used superhydride[20] failed, probably because of
concomitant reduction of the isoxazolidine moiety. Tosylation
of the resulting alcohol and hydrogenolysis of the isoxazoli-
dine with Raney nickel[9] induced spontaneous intramolecular
alkylation affording the properly bridged isoquinoline 3.[16d]
Desilylation with TBAF and cleavage of the aryl methyl ether
by boron tribromide caused, after alkaline workup, cycliza-
tion to a mixture of allopseudocodeine (16) and allopseudo-
morphine (15) along with some not demethylated epoxide
17.[10] Allylic substitution could be performed by BBr3 starting
from 17 too, allowing the recycling of this material. Chemo-
selective methylation of phenol 15 succeeded by a reported
procedure with sodium ethoxide and Me3NPhCl as the
methylating agent.[21] At this point, the first formal synthesis
had been successfully achieved. Oxidation of 16 with DMP to
Scheme 3. Completion of both formal and total syntheses of (ꢀ)-codeine. a) MeC(OMe)2NMe2, toluene, Dean–Stark trap, reflux, 87%; b) LDA,
BH3·NH3, 08C!RT, 97%; c) pTsCl, Et3N, cat. DMAP, CH2Cl2, 08C!RT; d) EtOH/EtOAc (3:1), Et3N, Raney Ni, 1 atm H2, RT, 86% (2 steps);
e) TBAF, THF, 08C, 100% crude; f) 1. BBr3, CH2Cl2, ꢁ658C, 2. aq. NaHCO3, RT, 15% 16 + 19% 15 + 16% 17 (2 steps); g) Me3NPhCl, NaOEt,
toluene, reflux, 80%; h) 1. BBr3, CH2Cl2, ꢁ658C, 2. aq. NaHCO3, RT, 37% 16 + 16% 15; i) DMP, NaHCO3, CH2Cl2, RT, 91%; j) ClC(O)OMe,
NaHCO3, CHCl3, reflux, 93%; k) SOCl2, 08C!RT, 77% crude; l) 1,4-dioxane/water (1:1), 1008C, microwave irradiation, 72%; m) DMP, NaHCO3,
CH2Cl2, RT, 29% 19 + 30% pseudocodeinone; n) NaBH4, MeOH, RT, 99%. DMP=Dess–Martin periodinane, LDA=lithium diisopropylamide,
TBAF=tetrabutylammonium fluoride.
Angew. Chem. Int. Ed. 2011, 50, 3892 –3894
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3893

Communications
give pseudocodeinone and N-demethylation with methyl
Keywords: allylic transposition · Claisen rearrangement ·
dearomatization · nitrone cycloaddition · total synthesis
.
chloroformate[22] afforded the known codeine precursor
18,[4a,b] the spectral data of which were identical to that
reported by Fukuyama et al.
[1] J. Frackenpohl, Chem. Unserer Zeit 2000, 34, 99 – 112.
[2] T. Hudlicky, J. Zezula, Synlett 2005, 388 – 405.
[3] P. Magnus, N. Sane, B. P. Fauber, V. Lynch, J. Am. Chem. Soc.
2009, 131, 16045 – 16047.
For the critical allylic transposition required for the
conversion of 16 to 2, hydrolysis of b-chlorocodide (21,
Scheme 4) was envisioned.[11a,b] In contrast to the literature
[4] a) T. Fukuyama, S. Yokoshima, T. Kan, K. Uchida, Org. Lett.
2006, 8, 5311 – 5313; b) T. Fukuyama, S. Yokoshima, T. Kan, K.
Uchida, Heteocycles 2009, 77, 1219 – 1234; c) T. Fukuyama, H.
Koizumi, S. Yokoshima, Chem. Asian J. 2010, 5, 2192 – 2198.
[5] T. Hudlicky, H. Leisch, A. T. Omori, K. J. Finn, J. Gilmet, T.
Bissett, D. Ilceski, Tetrahedron 2009, 65, 9862 – 9875.
[6] G. Stork, A. Yamashita, J. Adams, G. R. Schulte, R. Chesworth,
Y. Miyazaki, J. J. Farmer, J. Am. Chem. Soc. 2009, 131, 11402 –
11406.
[7] a) P. J. Parsons, M. Chandler, J. Chem. Soc. Chem. Commun.
1984, 322 – 323; b) P. J. Parsons, C. S. Penkett, Chem. Rev. 1996,
96, 195 – 206.
[8] a) A. M. M. Castro, Chem. Rev. 2004, 104, 2939 – 3002; b) J.
Mulzer, J. W. Bats, B. List, T. Opatz, D. Trauner, Synlett 1997,
441 – 444; c) J. Mulzer, D. Trauner, J. W. Bats, A. Werner, J. Org.
Chem. 1998, 63, 5908 – 5918; d) N. Chida, H. Tanimoto, R. Saito,
Tetrahedron Lett. 2008, 49, 358 – 362.
[9] M. Bols, O. L. Lopez, J. G. Fernꢁndez-Bolaꢂos, V. H. Lillelund,
Org. Biomol. Chem. 2003, 1, 478 – 482.
[10] D. F. Taber, T. D. Neubert, A. L. Rheingold, J. Am. Chem. Soc.
2002, 124, 12416 – 12417.
[11] a) E. Speyer, H. Rosenfeld, Ber. Dtsch. Chem. Ges. 1925, 58,
1113 – 1116; b) L. Knorr, H. Hꢃrlein, Ber. Dtsch. Chem. Ges.
1908, 41, 969 – 975; c) L. Knorr, H. Hꢃrlein, Ber. Dtsch. Chem.
Ges. 1907, 40, 4889 – 4892; d) M. Gates, J. Am. Chem. Soc. 1953,
75, 4340 – 4341.
[12] B. M. Trost, W. Tang, J. Am. Chem. Soc. 2002, 124, 14542 – 14543.
[13] A. K. Sinhababu, R. T. Borchardt, J. Org. Chem. 1983, 48, 2356 –
2360.
[14] G. Fenteany, S. V. Ankala, Tetrahedron Lett. 2002, 43, 4729 –
4732.
[15] S. Quideau, L. Pouysꢄgu, D. Deffieux, Tetrahedron 2010, 66,
2235 – 2261.
[16] a) CCDC 801503, b) 801502 (5), c) 801501, and d) 801500contain
the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
Scheme 4. A closer look at the allylic transposition.
report,[11a] treatment of 16 with thionyl chloride did not afford
pure 21 but led to a mixture of the isomeric chlorocodides 21/
23 along with 6-demethoxythebaine (24)[23] as an elimination
product. Hydrolysis of this crude mixture under microwave
irradiation in dioxane–water yielded a chromatographically
inseparable mixture of isocodeine (22) along with pseudoco-
deine (20) and 16. However, after oxidation of this mixture
with DMP, codeinone (19) could be readily separated from
pseudocodeinone and was finally reduced by treatment with
sodium borohydride[11d] in methanol following Gatesꢀ proce-
dure. The synthetic (ꢀ )-codeine (2) obtained in this way
showed spectral data identical to that of an authentic sample.
In conclusion, we have developed a straightforward route
from isovanillin (6) to allopseudocodeine (16) as part of one
total and two formal syntheses of the target alkaloid 2.
Several crucial steps such as the challenging acetal cleavage,
intramolecular nitrone cycloaddition, and Claisen rearrange-
ment allowed the facile construction of the morphine core.
Further optimization of the final allylic transposition is the
subject of ongoing work.
[17] M.-E. Trꢅn-Huu-Dꢅu, R. Wartchow, E. Winterfeldt, Y.-S. Wong,
Chem. Eur. J. 2001, 7, 2349 – 2369.
[18] I. E. Markꢆ, A. Ates, A. Gautier, B. Leroy, J.-M. Plancher, Y.
Quesnel, J.-C. Vanherck, Tetrahedron 2003, 59, 8989 – 8999.
[19] A. G. Myers, B. H. Yang, H. Chen, L. McKinstry, D. J. Kopecky,
J. L. Gleason, J. Am. Chem. Soc. 1997, 119, 6496 – 6511.
[20] M. G. Banwell, A. D. Findlay, Org. Lett. 2009, 11, 3160 – 3162.
[21] V. M. Rodionov, Bull. Soc. Chim. Fr. 1926, 39, 305 – 325.
[22] P. S. Portoghese, M. M. Abdel-Monem, J. Med. Chem. 1972, 15,
208 – 210.
[23] a) H. Rapoport, C. W. Hutchins, G. K. Cooper, S. Pꢇrro, J. Med.
Chem. 1981, 24, 773 – 777; b) V. N. Kalinin, I. L. Belyakova, V. V.
Derunov, J. K. Park, H. Schmidhammer, Mendeleev Commun.
1995, 5, 22 – 23.
Received: November 26, 2010
Published online: March 23, 2011
3894 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 3892 –3894