Some recent advances in the chemistry of the opium alkaloids
It is particularly appropriate that recent advances in the chemistry of the opium alkaloids be reviewed in the Bulletin on Narcotics. The standard monograph in this field by Small (1), though excellent for literature until 1931, was hopelessly out of date by the time the relevant chapters appeared in the Manske-Holmes treatise on alkaloids f(2), thus covering the literature to about 1950. Finally Bentley did yeoman service to chemists interested in this field by his contribution of a new monograph, which, however, covers only part of the opium alkaloids, albeit the more complex and therefore the more interesting ones (3).
Author: David Ginsburg
Pages: 18 to 26
Creation Date: 1957/01/01
It is particularly appropriate that recent advances in the chemistry of the opium alkaloids be reviewed in the Bulletin on Narcotics. The standard monograph in this field by Small [(1)] , though excellent for literature until 1931, was hopelessly out of date by the time the relevant chapters appeared in the Manske-Holmes treatise on alkaloids f[(2)] , thus covering the literature to about 1950. Finally Bentley did yeoman service to chemists interested in this field by his contribution of a new monograph, which, however, covers only part of the opium alkaloids, albeit the more complex and therefore the more interesting ones [(3)] .
In this field, as in many others, chapters in treatises, and even in the best of books, become obsolete almost as soon as they appear. Important contributions to the chemistry of the opium alkaloids have appeared in the few years following the appearance of the above-mentioned books, thus necessitating further shorter reviews from time to time. This review, though not comprehensive, will be divided into several sections which appear to be logical ones for the presentation of the material at hand.
It was an honour for the reviewer to be allowed by Mrs. Dorothy Crowfoot Hodgkin, F.R.S., to quote the preliminary results, obtained in collaboration with Maureen Mackay, concerning their crystallographic examination of the structure of morphine, prior to their publication [(4)] . Dr. Cardwell must be congratulated for his success in persuading Mrs. Hodgkin to undertake this difficult crystallographic problem. As an organic chemist whose overtures on behalf of morphine to other X-ray crystallographers were, alas, rebuffed, the reviewer heartily appreciates Dr. Cardwell's catalysis in this respect. Mackay & Hodgkin's results are now at hand [(5)] . These results, based mainly upon the crystallographic examination of morphine hydriodide dihydrate, clearly show that morphine may be represented sterically by formulation I.
The molecule is roughly T-shaped, with two planes nearly at right-angles to one another. One plane includes the aromatic ring A, the five-membered ether ring and ring B, whilst ring C and the piperidine ring form the second plane. The crystallographic work confirms previous chemical knowledge as to the cis-fusion of rings B and C. Indeed, it completely justifies Rapoport's views on the stereochemistry of morphine previously reviewed in the Bulletin on Narcotics[(6)] .
This crystallographic study yielded no evidence, however, on the absolute stereochemistry of the morphine molecule. Several important contributions relating to this point have appeared recently. Bentley & Cardwell [(7)] , after a short review of the existing evidence for the stereochemistry of morphine, proceed to deduce the absolute stereochemistry of this alkaloid. They show that (-) -morphine is I and not the enantiomorph. This paper is a model of what can be accomplished by reasoning based mainly upon an analysis of molecular rotation differences of various morphine derivatives, and upon a minimum of new experimental results. These authors deduce the same formulation (I) for the stereochemistry of morphine as reached by Stork some years previously (2 b), although Stork made no claims at the earlier date as to the absolute stereochemistry of morphine.
The reviewer does not feel competent to judge the niceties of the acrimonious debate between Bentley & Cardwell on the one hand and Stork (8) on the other, nor does he feel justified in presenting his personal views on the matter in this review. Suffice it to say that both parties have reached the same conclusions through reasoning, deductive and otherwise, and the reader is referred to the original papers in order to draw his own conclusions as to the degree of ironclad chemical proof and the line of reasoning - logical, intuitive and other - which appears in the original papers.
Fortunately, an additional paper by Kalvoda, Buchschacher & Jeger (9) on the absolute configuration of morphine appeared on the heels of the above- mentioned paper by Bentley & Cardwell. Completely in the tradition of what one expects from the organic chemistry school at the Eidgenossische Technische Hochschule at Zurich, this is a beautiful addition to the series of papers relating to the complex interrelation of the absolute configuration of complex natural products such as steroids, diterpenes and triterpenes. Also refreshing is the fact that well-founded experimental results are advanced to support the final formulation (again structure I) for the absolute configuration of morphine.
The interesting way in which thebaine was degraded to give (-)- cis-2-methyl-2-carboxy-cyclohexylacetic acid is shown in the following reaction scheme. Thebaine (II) was degraded by known methods to yield cis-13-vinyl-6-oxo-5,6,7,8,9,10,13,14-octahydromorphenol methyl ether (III).
The C-6 carbonyl group was protected by forming the ketal (IV) with ethylene glycol under acid-catalyzed conditions. Hydroxylation with osmium tetroxide then yielded the cis-glycol (V) which was cleaved by lead tetraacetate to yield the aldehyde (VI). This was converted to the thioketal VII by treatment with ethanedithiol in the presence of boron trifluoride and the latter was reductively desulfurized by refluxing with Raney nickel. Alternatively the aldehyde could be reduced to the desired angular methyl derivative VIII by the Huang-Minlon procedure. The ketal blocking group was now removed by treatment with acid, and reductive ether ring scission in the resulting free ketone IX was effected by treatment with aluminium amalgam in moist ether. The resulting phenol X was converted into the dimethyl ether through the use of dimethyl sulfate and alkali. The dimethyl ether (XI) was now treated with ethanedithiol and boron trifluoride in order to obtain the 6-thioketal (XII) and this was reductively desulfurized by Raney nickel to yield the dimethyl ether XIII. Chromic acid oxidation attacked the position adjacent to the aromatic nucleus and afforded the ketone XIV which by vigorous ozonolysis followed in turn by oxidative decomposition with performic acid, diazomethane methylation, distillation of the dimethyl ester obtained and alkaline saponification of the latter, finally afforded (-)- cis-2-methyl-2-carboxy-cyclohexylacetic acid (XV).
The following structural formulae illustrate the interrelationship between the ketone XIV (written in a slightly different way) and the diacid XV.
The first important conclusion of the Jeger paper obtained through conversion of thebaine to the diacid XV is that for the first time the junction between rings B and C in 8,14-dihydrothebaine (and in morphine) has been proved to be cis. This appears to be the first experimental proof of this point since the Gulland-Robinson structure for morphine had been advanced in 1925 [(10)] .
The second conclusion reached in this paper is that rings B and C in morphine correspond to rings A and B of the
5β-steroids and that the ethanamine bridge (including carbon atoms [15 ] and  ) has the same spatial position as the β-methyl group at C-10 of the steroids and the cyclic polyterpenes. (See partial formula XVI for steroid which corresponds to the absolute configuration of XV.)
A further paper which touches on the stereochemistry of the halocodides has been published by Stork and Clarke and appeared as a third part in a wider study of the SN2’ reaction [(8)] . In this paper, Stork presents experimental proof of his brilliantly intuitive and mechanistically based arguments (2 b) regarding the structure and stereochemistry of the halocodides and their displacement products as well as the stereochemistry of pseudocodeine and allopseudocodeine. These results, of course, apply analogously to the corresponding morphine derivatives.
α-Chlorocodide is shown to be XVII. β-Chlorocodide is the C-8 isomer with the hindered back side (XVII a). Similarly bromocodide and iodocodide are formulated as XVIII b and XVIII c, respectively, on the basis of infrared comparisons, molecular rotation differences and behaviour of the four halocodides on reduction with lithium aluminium hydride.
( a) R=Cl
( b) R=Br
( c) R=I
Rapoport & Masamune [(11)] have reported on the stereochemistry of 10-hydroxycodeine denivatives. These workers studied the chromic acid oxidation of dihydrodesoxycodeine (XIX) after it had previously been shown by Rapoport & Stevenson [(12)] that the introduction of an oxygen function at C-10 appeared to be a general reaction of codeine derivatives. The choice of XIX was dictated by the absence of other reactive functional groups in the molecule. The reactions carried out by these workers are summarized in the following scheme.
By oxidation of dihydrodesoxycodeine with cold chromic acid in dilute sulfuric acid, the 10-hydroxy-derivative(XX) was obtained. This being an allylic alcohol, it was successfully oxidized by manganese dioxide to yield the 10-ketone(XXI) . Sodium borohydride reduction of the latter gave in quantitative yield an alcohol epimeric with XX. It should be noted in passing that chromic acid oxidation in this case afforded very little of the 10-ketone, whilst Jeger and collaborators ( ; see above) obtained a good yield of 10-oxoderivative in their chromic acid in acetic acid oxidation.
The pair of epimeric alcohols at C-10 was exhaustively studied by Rapoport & Masamune and, on the basis of physical constants and chemical reactions, the hydroxyl group in XX was assigned the axial and that in XXII was assigned the equatorial confirmation.
Finally, a paper by Elad & Ginsburg has appeared [(13)] , in which the rates of saponification of epimeric pairs of various morphine derivatives were determined and the conformations of the hydroxyl groups in these pairs of componds were assigned. Thus in dihydrocodeine the C-6 hydroxyl group was assigned the axial, whilst in dihydro isocodeine it was assigned the equatorial conformation. Similarly, in dihydropseudocodeine the C-8 hydroxyl group was assigned the equatorial, and in dihydro allopseudocodeine it was assigned the axial conformation. Finally, in dihydrothebainol A, the C-6 hydroxyl group was assigned the equatorial and in its epimer, dihydrothebainol B, it was assigned the axial conformation.
Synthesis and partial synthesis
The full account ([14 ] a) of the first synthesis of morphine, previously disclosed as a short communication ([14 ] b), has been published by Gates & Tschudi. This synthesis has been reviewed in the Bulletin on Narcotics, but there are several improvements in the synthesis now fully reported, which are outlined below.
Previously, the lactam XXIII had been reduced directly with lithium aluminium hydride, and the resulting secondary amine was methylated with formaldehyde-formic acid to yield rac-β-Δ 6-dihydrodesoxycodeine methyl ether (XXV). The lactam XXIII has now been converted to the N-methyl lactam (XXIV) by means of sodium hydride and methyl iodide, and this was then reduced with lithium aluminium hydride to give XXV in 88 % yield fromXXIII .
XXIII XXIV XXV
This procedure affords a superior yield and quality of product, since the reduction step is much shorter and the reaction mixture is homogeneous throughout.
Another improvement effected was the use of the Goto modification (15; see below) for the conversion of
1-bromodihydrothebainone to 1-bromocodeinone(XXVIII) . The improved yield results from refluxing the crude 2, 4-dinitrophenylhydrazone (XXVII) with pyridine. Apparently, the bromine at C-7 is hydrolysed during the process.
Sinomenine (XXIX) was reported by Goto at the Zurich International Congress of Pure and Applied Chemistry
( cf. 16) to have been converted into (+)-morphine through a series of reactions essentially utilizing the steps of the Gates synthesis of morphine (14) through (+)-dihydrothebainone,
the enantiomorph of Gates’ (-)-dihydrothebainone. It is a pity that the over-all yield of (+)-morphine from sinomenine was at best 0.3%, so that insufficient quantities of material have been available, and pharmacological testing has thus not been feasible.
Another interesting partial synthetic scheme disclosed very recently is the transformation of a morphine derivative into thebaine. Whilst the reverse change has been known for decades, Rapoport, Reist & Lovell [(17)] have reported the conversion of codeinone dimethyl ketal (XXXIV) into thebaine (II). The reactions are summarized in the accompanying scheme.
Ketalization of codeinone(XXX) with trimethyl orthoformate, methanol and sulfuric acid unexpectedly afforded 8-methoxy-Δ 6-dihydrothebaine (XXXI), whose structure was conclusively established.
The desired dimethyl ketal of codeinone (XXXIV) was obtained by potassium t-amyloxide dehydrobromination of 7-bromodihydrocodeinone dimethyl ketal (XXIII), itself prepared by methyl hypobromite addition to Δ6-dihydrothebaine (XXXII), according to a procedure due to Dr. Lyndon F. Small.
Treatment of the dimethylketal XXXIV with dry p-toluenesulfonic acid in chloroform gave thebaine (II) in 40% yield.
Since the series of transformations codeine → dihydrocodeinone → Δ 6-dihydrothebaine has been effected, and since codeine itself has been prepared by total synthesis, the conversion reported by Rapoport and co-workers may be considered as constituting a formal synthesis of thebaine. Since thebaine has recently, been converted into neopine (see below, section 4), this work also constitutes a formal synthesis of neopine.
3. Synthetic morphinans
A large number of morphinan derivatives has been described by Hellerbach, Grüssner & Schnider [(18)] . Since the extraordinary effectiveness of N-Allyl normorphine(XXXV) has been established in counteracting the respiratory depressant analgesic properties of morphine and of other analgesics [(19)] , this recent report is concerned with various (-)-3-hydroxy- N-allylmorphinan derivatives.
The elegant synthetic approach is concerned with the synthesis of l-( p-hydroxybenzyl)-l,2,3,4,5,6,7,8-octahydro iso-quinoline (XXXVI), which was resolved into its optical antipodes; these could then be cyclized by the now famous Grewe acid cyclization procedure to yield optically active 3-hydroxymorphinan (XXXVII a). Ethers and esters were prepared at the 3-position, and N-alkylation was effected to introduce the allyl and other unsaturated groups attached to nitrogen.
An improved synthesis of (-)-3-hydroxymorphinan was developed. The optically active (+)-XXXVI was N-benzylated, and the product (XXXVIII) was cyclized to give (-)-1-( p-hydroxybenzyl)-2-benzyl-octahydro isoquinoline
(XXXIX) . The latter was cyclized to give (-)-3-hydroxy- N-benzylmorphinan (XXXlX). Hydrogenolysis of the N-benzyl group then afforded (-)-3-hydroxymorphinan in excellent yield.
In a further communication from the Hoffman-La Roche group, Grüssner, Hellerbach & Schnider [(20)] have elucidated the structure of by-products obtained during the cyclization of octahydro isoquinolines to morphinans. They show through exhaustive methylations and degradations and comparisons with authentic compounds that these by-products are compounds of apomorphine-like structure.
A solution of the practical problem of racemization of the optically active 1-( p-hydroxybenzyl)-l,2,3,4,5,6,7,8-octahydro isoquinoline,which is not required for further cyclization as it would give the pharmacologically inactive enantiomorph of 3-hydroxymorphinan, has been offered by Brossi & Schnider [(21)] . The following reaction scheme is followed:
The undesired active form of XXXVI a or of its methyl ether (XXXVI b) is dehydrogenated with palladium-charcoal in boiling tetralin. Reduction of the resulting Bz-tetrahy droisoquinoline(XL) with sodium in isoamyl alcohol affords the racemic form of XXXVI a or b, respectively, and this can be resolved optically by recycling of the racemate with additional batches of synthetic racemic material. The dehydrogenation, of course, destroys the asymmetric carbon atom marked with an asterisk in structure XXXVI a or b. Conversion of XL to XLI can also be achieved by catalytic reduc- tion of the quaternary bromobenzylates formed from XL by way of the tertiary N-benzyl- 1,2,3,4,5,6,7,8-octahydro isoquinoline s(racemic XXXVIII).
Schnider and co-workers [(22)] have shown that photooxidation of (+)-3-methoxy- N-methyl-morphinan (XLIII) hydrobromide yields (-)-3-methoxy-10-oxo-N-methyl-morphinan(XLIII) . The same compound is formed by chromic acid oxidation in dilute sulfuric acid. (See discussion in section 1 regarding differences in behaviour in oxidative introduction of oxygen function at C-10.)
Lithium aluminium hydride reduction of XLIII is appa-rently not stereospecific in this case. Pharmacological testing showed that the C-10 oxo- or hydroxy-derivatives of laevo-XLII had lost almost all of their analgesic activity. Similarly, the corresponding derivatives of the dextro-series showed marked decrease in antitussive activity.
An important discovery has been made by Gates & Hughes [(23)] in connexion with the closure of the oxide bridge in the morphine series. It has long been known from Schopf’s work that the 4,5-oxide bridge in the morphine series is readily formed when rings B and C are cis-fused. The cyclization reaction involving dibromination followed by treatment with alkali fails, however, when rings B and C are trans-fused .Indeed, Bentley & Cardwell[(7)] have even questioned the capability of the existence of a trans-fused B-C derivative containing the oxide ring.
Gates & Hughes have found that although alkali does not effect closure of the oxide bridge in dibromo- trans-dihydrothebainone (XLIV), the action of boiling collidine produces in moderate yield a saturated non-phenolic monobromoketone, which upon reduction by zinc and ammonia yields the known l-bromo- trans-dihydrothebainone (XLVI). The product is therefore formulated as the hitherto unknown l-bromo- trans-dihydrocodeinone (XLV).
If this very interesting observation and the explanation given by these workers are further developed, the availability of a substance of this type (XLV) should permit, as the authors themselves point out, the preparation of further pentacyclic trans-derivatives, including trans-morphine itself!
Cohen [(24)] has published a short note on the rationalization of a biogenetic course for the formation of morphine with phenolic oxidative coupling as applied in the case of usnic acid ( ; see last paragraph of discussion on p. 532). All chemists interested in the structural relations of natural products await exciting developments in this field emanating from Prof. Barton's laboratory.
Bentley and Ball[(26) ] have published a short contribution reporting two new Diels-Alder adducts of thebaine, in connexion with their investigation of the complex flavo-thebaone rearrangement. These adducts were obtained with methyl-vinyl-ketone and with phenyl-vinyl-ketone. A review of this complex subject must await further results from the Aberdeen laboratories.
Bentley & Cardwell [(7)] have confirmed through Kuhn-Roth determination that methebenol (XLVII) has the struc- ture shown, since one molecular equivalent of acetic acid was produced. However, 6-methoxy-thebentriene (XLVIII), for example, gave no acetic acid on Kuhn-Roth oxidation and is therefore formulated as shown.
The utilization of thebaine extracted from opium - e.g., its conversion to codeine - is a matter of practical interest. Conroy [(27) ] therefore investigated the potential conversion of thebaine to codeinone (LI). Though the original aim was not realized, the properties of a new ketone, neopinone(L) , were investigated.
Bromination of thebaine (II) yielded 14-bromocodeinone (XLIX), and reduction with palladium in chloroformmethanol afforded the new neopinone (L). This ketone is related to the rare opium alkaloid, neopine (LII), in the same way as codeinone (LI) is related to codeine(LIII).
SMALL, Chemistry of the Opium Alkaloids, U.S. Govt. Printing Office, Washington, D.C., 1932.002
(a) HOLMES, in Manske-Holmes, The Alkaloids, Vol. 2, pp. 1-171, 203-217, Academic Press, New York, 1952.
(b) STORK, ibid., pp. 171-203.003
BENTLEY, "The Chemistry of the Morphine Alkaloids ", Clarendon Press, Oxford, 1954.004
ELAD & GINSBURG, J. Chem. Soc., 1954, 3052.005
MACKAY & HODGKIN, ibid., 1955, 3261.006
GINSBURG, Bulletin on Narcotics, Vol. V, No. 4.007
BENTLEY & CARDWELL, J. Chem. Soc., 1955, 3252.008
STORK & CLARKE, J. Amer. Chem. Soc., 78, 4619 (1956).009
KALVODA, BUCHSCHACHER & JEGER, Helv. Chim. Acta, 38, 1847 (1955).010
GULLAND & ROBINSON, Mem. Proc. Manchester Lit. Phil. Soc., 69, 79 (1925).011
RAPOPORT & MASAMUNE, J. Amer. Chem. Soc., 77, 4330 (1955).012
RAPOPORT & STEVENSON, ibid., 76, 1796 (1954).013
ELAD & GINSBURG, ibid., 78 , 3691 (1956).014
( a) GATES & TSCHUDI, ibid., 78, 1380 (1956).
( b) GATES & TSCHUDI, ibid., 74, 1109 (1952).015
GOTO & YAMAMOTO, Proc. Jap. Acad., 30, 769 (1954).016
GOTO, Bull. Agr. Chem. Soc. Japan, 19, No. 2, 1 (1955).017
RAPOPORT, REIST & LOVELL, J. Amer. Chem. Soc., 78, 5128 (1956).018
HELLERBACH, GRÜSSNER & SCHNIDER, Helv. Chim. Acta, 39, 429 (1956).019
CLARK, PESSOLANO, WEIJLARD & PFISTER, J. Amer. Chem. Soc., 75 , 4963 (1953) and references given in footnotes 1 and 2 in their paper.020
GRÜSSNER, HELLERBACH, BROSSI & SCHNIDER, Helv. Chim. Acta, 39 , 1371 (1956).021
BROSSI & SCHNIDER, ibid., 39, 1376 (1956).022
HAFLIGER, BROSSI, CHOPARD-DIT-JEAN, WALTER & SCHNIDER, ibid., 39, 2053 (1956).023
GATES & HUGHES, Chemistry and Industry, 1956, 1506.024
COHEN, ibid., 1956, 1392.025
BARTON, DEFLORIN & EDWARDS, J. Chem. Soc., 1956, 530.026
BENTLEY & BALL, Chemistry and Industry, 1956, 1428.027
CONROY, J. Amer. Chem. Soc., 77, 5960 (1955).