Studies on the Synthesis and Biosynthesis Of
STUDIES ON THE SYNTHESIS AND BIOSYNTHESIS OF
INDOLE ALKALOIDS
BY
GEORGE BOHN FULLER
B.A. (cum laude) , Macalester College, 1969 M.Sc, The University of California, Berkeley, 19
A THESIS SUBMITTED IN PARTIAL FULFILMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
in the Department
of
CHEMISTRY
We accept this thesis as conforming to the
required standard /-)
THE UNIVERSITY OF BRITISH COLUMBIA
July, 1974 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study.
I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.
Depa rtment
The University of British Columbia Vancouver 8, Canada ABSTRACT
Part A of this thesis provides a resume1 of the synthesis of various radioactively labelled forms of secodine C76) and provides an evaluation of these compounds, as well as some radioactively labelled forms of tryptophan C25), as precursors in the Biosynthesis of apparicine (81), uleine C83), guatam- buine (90) , and olivacine (88) in Aspidosperma australe.
Only apparicine (81) could be shown to incorporate these precursors to a significant extent. Degradation of apparicine
(81) from Aspidosperma pyricollum provided evidence for the intact incorporation of the secodine system.
Part B discusses the synthesis of 16-epi-stemmadenine (161), which provides an entry into the stemmadenine system with, radioactive labels at key positions in the molecule. The synthesis involved the degradation of strychnine (29) to
Wieland-Gumlich aldehyde (130) by a previously established sequence of reactions. Initial conversion of Wieland-Gumlich aldehyde to nor^fluorocurarine (134) succeeded by a previously described route, although some study was necessary for determin• ing the conditions by which the Oppenauer oxidation of 2B,16a- cur-19-en-17-ol (137) could selectively yield either 23,16a-cur-
19-en-17-al (133) or nor-fluorocurarine (134). When nor-fluoro- curarine (134) could not be converted to the desired stemmadenine system, Wieland^GxunlictL aldehyde was converted to methyl 18- hydroxy^2&,16a-cur-19-en-17^oate (156) by a previously established procedure. Conversion of this compound to methyl 2 6/, 16a-cur-19- en-17-^oate 0.571 was accomplished by successive treatment with, hydrogen bromide and zinc in acetic acid. The ester
157 was converted to its- N Ca I *s£ o rmy-1 derivative 158 by
reaction with methyl formate and sodium hydride. Treatment of this product with dry formaldehyde and sodium hydride in
dimethyl sulfoxide led to the formation of the unexpected but
nevertheless useful tetrahydrooxazine derivative 159. Hydrolysis
of the tetrahydrooxazine moiety was accomplished with methanolic
hydrogen chloride, resulting in the isolation of 2g,16g-carbo- methoxy-cur-19-en-17-ol (160) . Oxidation of compound 160 with
lead tetraacetate followed immediately by treatment with sodium
borohydride in methanolic acetic acid provided 16-epi-stemmaden-
ine C161). Hydride reduction of the C-16 ester function in 161
and authentic stemmadenine (6a) led to the same diol 175 thereby
providing the required interrelationship between the synthetic
and natural compounds. This sequence also established the
previously unknown configuration of stemmadenine (6a) about
C-16 and provided an obvious pathway for the synthesis of
stemmadenine via the saturated aldehyde 133. Also discussed in
Part B is the lead tetraacetate oxidation of the ester 157 to
akuammicine (66), representing the first total synthesis of
that compound.
Part C discusses the synthesis of 16-epi-stemmadenine (161)
labelled with tritium in the aromatic ring. Simultaneous 3
administration of this material and stemmadenine-Car- H) (6a)
to separate portions of A., pyricolluro root sections established
that, while the latter was incorporated into apparicine (81), - iv - no incorporation could be detected in the. case of the former. y —
TABLE OF CONTENTS
Page
TITLE PAGE . . i
ABSTRACT ii
TABLE OF CONTENTS V
LIST OF FIGURES vi
LIST OF TABLES ix
ACKNOWLEDGEMENTS x
INTRODUCTION 1
DISCUSSION 35
PART A 35
PART B 57
PART C 10 7
EXPERIMENTAL 108
SECTION A 112
SECTION B 116
SECTION C 135
BIBLIOGRAPHY 136 - yi -
LIST OF FIGURES
Figure Page
1 Some Representative Indole Alkaloids 2
2 The Biosynthesis of Anthranilic Acid C21) 5
3 The Biosynthesis of Tryptophan (25) from Anthranilic Acid (21) 6
4 The Barger-Hahn-Robinson-Woodward Postulate for Indole Alkaloid Biosynthesis 8
5a The Wenkert Prephenate Postulate for Indole Alkaloid Biosynthesis 9
5b The Thomas Monoterpene Postulate for Indole Alkaloid Biosynthesis 9
6a The Leete Acetate-Malonate Hypothesis for Indole
Alkaloid Biosynthesis 14
6b The Hendrickson Polyketide Modification 14
7 The Early Stages of Indole Alkaloid Biosynthesis as Proven by Experiment 16 8 The Postulated Derivation of Corynanthe Alkaloids from Vincoside (54) 19
9 The Wenkert (A) and Scott (B) Postulates for the Biosynthesis of Strychnos Alkaloids 21
10 The Wenkert Postulate for the Biosynthesis of Iboga and Aspidosperma Alkaloids 22
11 The Postulated Derivation of Aspidosperma and Iboga Alkaloids from Intermediate 75 26
12 The Wenkert Postulate for the Biosynthesis of Uleine (83) ' 29
13 The Djerassi Postulate for the Biosynthesis of Apparicine (81) 29
14 The Results of Incorporation of Various Possible Intermediates into Apparicine (81) and Uleine (83) 33 - yii -
Figure Page
15 The Potier-Janot Postulate for the Biosynthe• sis of Non-tryptamine Alkaloids 34
16 The Conversion of Olivacine (88) to Guatam-
buine (90) 37
17 The Ozonolytic Degradation of Apparicine (81)... 39
18 The Synthesis of Secodine-(Ar-3H) and (14COOCH^)
(76) 7. 40
19 The Synthesis of Secodine-(19-3H) (76) 44
20 The Synthesis of 1-(31-pyridyl)-ethane-(1-3H) (116) 45 21 The Proposed Relationships of Secodine (76) to Stemmadenine (6) in Indole Alkaloid Biosynthe• sis 49 22 The Correlation of Akuammicine (66) and Stemm• adenine (6) via Preakuammicine (2) 58
23 The Attempted Synthesis of 19, 20-Dihydrostemm- adenine 60
24 The Degradation of Strychnine (29) to Wieland- Gumlich Aldehyde 61
25 A Summary of Some Known Reactions in the Wieland-Gumlich Aldehyde Series 64
2 6 The Boekelheide Mechanism for the Oppenauer Oxidation of 133 to 134 66
27 The Routes Considered for the Conversion of Nor-fluorocurarine (134) to Stemmadenine (6).... 68
2 8 The Nuclear Magnetic Resonance Spectrum of Nor- f luorocurarine (134) 72
29 The Reaction of 16-epi-WGA with HCN 73
30 The Reactions of Nor-fluorocurarine (134) with Cyclohexylamine, Pyrrolidine and Morpholine 75
31 The Synthesis of 16-epi-stemmadenine (161) from WGA (130) 78 yiii -
Figure Page
32 The Proposed Mass Spectral Fragmentation of the Methyl Cur-19-en-17-oate System 80
33 Two Possible Routes to Stemmadenine (6) from Methyl 23,16a-cur-19-en-17-oate 0-57) 82
34 The Possible Condensation of Formaldehyde with Methyl 23 ,16a-cur-19-en-17-oate (157) 83
35 The Nuclear Magnetic Resonance Spectrum of the Carbomethoxy Tetrahydrooxazine 159 87
36 The Reversible Formation of the Model Tetrahydro• oxazine (166) 86
37 The Nuclear Magnetic Resonance Spectrum of the Model Tetrahydrooxazine (166) 89
38 The Proposed Mechanism for the Formation of the Carbomethoxy Tetrahydrooxazine 159 93
39 The Edwards and Smith Mechanism for the Zinc and Sulfuric Acid Reduction of Akuammicine (66) 96
4 0 The Nuclear Magnetic Resonance Spectrum of Akuammicine (66) 97
41 The Proposed Mechanism for the Formation of Indole Ester 141a 98
42 The Nuclear Magnetic Resonance Spectrum of the Indole Ester 141a 99
43 The Nuclear Magnetic Resonance Spectrum of the Indole Ester 141b 99
44 The Reduction of Natural and Synthetic Stemmaden• ine Systems to the Diol 175 103
4 5 The Infrared Spectra of Authentic and Synthetic Diol 175 104
46 The Nuclear Magnetic Resonance Spectra of Authentic and Synthetic Diol 175 105
47 The Proposed Route for the Synthesis of Stemm• adenine (6a) 106 - ix -
LIST OF TABLES
Table Page
1 Results of Incorporation of Secodine (76) into Apparicine (81) 47
2 Specific Activities Associated with the Experi• ments in Table 1 48
3 Specific Activities Associated with the Ozonoly- tic Degradation of Apparicine (81) in Experiments 2 and 5 50
4 The Various Compounds Fed to A. australe 52
5 Incorporation Results Associated with Table 3.... 53
6 Effects of Different Reaction Conditions on the Oppenauer Oxidation of the Alcohol 137 66
7 Summary of Results of the Investigations into Anion Formation at C-16 of the N-formyl ester (158) 85
8 A Summary of Pertinent Nuclear Magnetic Reson• ance Chemical Shifts for Compounds 156-159 91
9 Comparison of NMR Data of Natural and Synthetic Stemmadenine Systems 102
10 The Stemmadenine Systems Administered to A. pyricollum 106
11 Incorporation Results Associated with Table 10... 107 - X -
ACKNOWLEDGEMENTS
I would like to express xny appreciation to Professor
James P. Kutney for his guidance and encouragement, and the generous contributions of his time which were provided throughout the course of this work.
I would also like to express my gratitude to my wife, whose unfailing support, confidence, and understanding during
the course of this study were a contribution beyond measure.
I am grateful to The University of British Columbia and to the National Research Council of Canada for the
financial support which they provided. INTRODUCTION
To a chemist, one of the roost fascinating aspects of nature is the abundance of highly complex molecules found in plant systems. Even more fascinating is the fact that these complex molecules are made by the plants from the minerals available in the soil, the gases available from the atmosphere, water and sunlight. Even the most sophisticated of laboratory procedures pales in comparison with the synthetic operations occurring daily in plant systems. It is, therefore, of great importance to chemists to study plant systems in an effort to discover how complex molecules are constructed in vivo in order that in vitro transformations may become more efficient. This thesis concerns itself with such a study, in which an attempt is made to better understand the later stages of indole alkaloid biosynthesis.
At least 10-20% of all plants produce alkaloids,"'" and approximately 25% of these alkaloids contain the indole, dihydro- 2 indole, indolenine, or a-methylene-indoline system. Roughly 3 4
800 indole alkaloids are already known, ' and they can be con• veniently grouped into four main groups: corynanthe, strychnos, aspidosperma and iboga. Examples of these groups are geissoschizine Cl), preakuammicine (2), vindoline (3) and - 2 -
( 7 )
Figure 1. Some Representative Indole Alkaloids. X
- 3 - catharanthine C4), respectively. Differences do occur within various groups in terms of stereochemistry, functionality, etc. One of the most interesting variations in each group is the existence of nine (or ten) membered ring derivatives.
Examples of these are picraphylline (5) (corynanthe), stemmadenine (6) (strychnos), vincadine (7) (aspidosperma) and 16-carbomethoxycleavamine (8) (iboga).
Despite the differences between and within the four main groups of indole alkaloids, there exists an overall unifying
similarity. That is, they can all be formally derived from a
tryptamine residue joined with a C^-C^Q unit. In a formal
sense, the four families can be defined by the arrangement of
carbon atoms in the cg~c^Q residue. Structure (9) indicates
the arrangement found in both corynanthe and strychnos alkaloids, while the arrangements for aspidosperma and iboga are indicated
by structures (10) and (11) respectively.
(9 ) ( 10 ) (11)
There is general agreement that the tryptamine segment is 5—9
derived from tryptophan, in accord with the postulate of
10 Pictet in 1906. However, the origin of the CG-C1Q unit has - 4 -
been the object of speculation since the early thirties when 11 12
Barger and Hahn postulated that the non-tryptamine portion of yohimbine was tyrosine-derived.
In the past decade, the ready availability of radioactive materials for tracer studies has provided sufficient experimental data to elucidate the biosynthetic origins of both the tryptamine residue, and the Cg-C^n segment. The biosynthesis of tryptophan, as shown in Figure 2, has 13 14 been proven ' to occur via the shikimate-chorismate pathway.
Erythrose-4-phosphate (12) undergoes aldol condensation with phosphoenolpyruvic acid (13) to form 3-deoxy-D-arabinoheptulo- sonic acid-7-phosphate (14), which undergoes cyclization and oxidation to form 5-dehydroquinic acid (15). Loss of water results in 5-dehydroshikimic acid (16) which is then reduced by NADPH to shikimic acid (17) and phosphorylated to form shikimic acid-5-phosphate (19). Condensation with phosphoenol• pyruvic acid initially results in 3-enolpyruvyl shikimic acid-5-phosphate (19) which then undergoes dephosphorylation to form chorismic acid (20). Chorismic acid then acquires the amide nitrogen of glutamine to form anthranilic acid"^ (21), 16 17 the major precursor to tryptophan. It has been shown ' in
E. coli that the conversion of anthranilic acid to tryptophan occurs as shown in Figure 3. Formation of N-(51-phosphoribosyl)- anthranilic acid (22) followed by oxidative ring opening to form enol-1-(o-carboxylphenylamino-)-l-desoxyribulose-5-phosphate (23) and subsequent cyclization with loss of carbon dioxide results in the formation of indole-3-glycerylphosphate (24). CHO COO HCOH COPOoH HCOH
ChUOPOoH 1 3 ( 13 ) (12)
HO ?00
0 S-OH \ rl OH HO l\ H ( 15 ) ( 16 ) H OH
ATP / ( -| 7 ) COO COO CH-
H- -O-C COO H ~i *H03P0' ^6HH HO^PO i'-H 3 H OH (19 ) (18 )
-H2PO4 coo" .COO CH- glutamine >r > -o-c . 'NH- '•. H COO ( 21 ) H OH Figure 2. The Biosynthesis of Anthranilic Acid (21). (20) - 6 -
COO
NH CH90P0.H ->
( 21) HO OH
( 22)
H H H 0 0 0
CH 3H c—C 2OP0 H H
( 23 )
-co
H H 0 0 +serxne COO •C •C •CI-^OPO^H -3-phospho- H H glyceraldehyde
pyridoxal H phosphate
( 2U ) ( 25 )
Figure 3. The Biosynthesis of Tryptophan C25) from Anthranilic
Acid (21). - 7 -
Condensation of 24 with serine followed by loss of 3-phospho-
glyceraldehyde completes the biosynthesis of tryptophan (25).
Subsequent to the elucidation of the microbial biosynthesis 18
of tryptophan, it was shov/n that the pathway shown in Figures
2 and 3 is also valid in higher plant systems.
Prior to the first tracer experiments concerned with the
origin of the cg~c^o un;'-t' there were three main postulates in
existence. The original Barger-Hahn postulate had been 19 20-22 elaborated by Robinson and Woodward to the point where
it could accommodate the formation of alkaloids of the strychnos
and corynanthe, as well as the original yohimbe, alkaloid
skeletons as shown in Figure 4. The combined hypothesis
involves the degradation of tyrosine (26) to 3,4-dihydroxy- phenylacetaldehyde (27). Condensation of 27 with tryptamine
(28) could then occur at the £ position of the indole system to yield strychnine (29), or at the a position to yield 30, which contains an aromatic ring E. Fission of the aromatic ring E between the two hydroxyl groups results in 31 which can combine with two C^ units to yield yohimbine (32), or be directly converted to cinchonamine (33).
23 24
The second hypothesis was suggested by Wenkert ' in
1959. As shown in Figure 5a, Wenkert proposed that shikimic acid (17) derived from carbohydrate metabolism would combine with pyruvic acid (34) to form prephenic acid (35). Rearrangement, hydration, retroaldol cleavage and addition of formaldehyde results in 36, the so-called seco-prephenate-formaldehyde (SPF) 2 6 unit. Wenkert commented upon the similarities between the - 8 ~
Figure 4. The Barger-Ilahn--Robinson-Woodward Postulate for
Indole Alkaloid Biosynthesis. - 9 -
COO
0 • II Jr-OH H H3C—C —COO + O^/COO HO' \ H H OH (34)
(17) H ( 35 ) "ooc-
"OOC\^o ^o 0^/COO" o^coo
etro- 4aldo l 'ooc- CHO 'OOC OOC CHO (36) , OH Figure 5a. The Wenkert Prephenate Postulate for Indole Alkaloid
\ Biosynthesis.
"00 c ooc'
H0- -OH CH3C00 "00C-
( 45 )
Figure 5b. The Thomas Monoterpene Postulate for Indole Alkaloid
Biosynthesis. - 10 -
27-30 SPF unit and the recently discovered cyclopentanoid monoterpene glucosides exemplified by verbanalin (37), genipin (38), aucubin (39) and asperuloside (40) , and felt that these monoterpenes could be derived from the SPF unit
(dotted line in Figure 5), or they could be the real pre• cursors to the indole alkaloid C9~cio units- In anY event, both the SPF unit and the monoterpenes possess the carbon skeleton shown in structure 9 to be requisite for entry into the corynanthe-strychnos family of indole alkaloids.
.«0Glu
C00CH-, COOCH.
( 37 ) ( 38 ) TV U -
25
The third postulate was proposed by Thomas who felt that the cyclopentanoid monoterpenes contained the structural features present in the Cg^C]_Q unit of the indole alkaloids, and that their biosyntheses may be related to that of the indole systems. In other words, the monoterpenes were themselves derived from the acetate-mevalonate pathway (Figure 5b) independent of carbohydrate metabolism. This option was later 2 6 chosen by Wenkert in 1962. Shortly after the appearance of the Thomas-Wenkert 31 postulate, Leete began the first experimental evaluation of the biosynthesis of indole alkaloids using radioactive tracer studies. In order to test the Barger-Hahn-Robinson-Woodward 14 postulate (Figure 4), Leete fed tyrosine-2- C (26) to
Rauwolfia serpentina and could not detect any radioactivity in the ajmaline (41) and reserpine (42) which were isolated.
This tentative refutation of the Barger-Hahn-Robinson-Woodward 32 hypothesis was subsequently confirmed by Battersby in his studies on the isoquinoline alkaloids cephaeline (43) and emetine (44). Although these alkaloids lack an indole nucleus, they can be thought of as formally derived from two tyrosine
molecules, and one Cg unit which is structurally related to
Wenkert's proposed SPF unit.2^ When tyrosine-2-14C was fed to
Cephaelis ipecacuanha plants, the alkaloids isolated were shown to be active only at position C-3, with no activity found in the
C-l position as would have been predicted by the Barger-Hahn-
Robinson-Woodward postulate as shown in Figure 4. This experi• ment then provided convincing evidence that the cg~C^n, unit i-s not tyrosine-derived. - 12 -
( U2 )
31 Leete tested WenkertIs prephenic acid hypothesis by feed- 14 33 ing alanine-2- C, a known precursor of pyruvic acid (34) , to
R. serpentina plants. The ajmaline (41) isolated was of low
activity and with only 2% of its activity at the postulated position. The Thomas-Wenkert monoterpene postulate was also 14 tested. Mevalonic acid-2- C (45) was fed to R. serpentina, and no activity could be detected in the ajmaline (41) which 14 was isolated. However, when sodium acetate-1- C was fed to
the plant system, active ajmaline (41) was isolated in which ~ 13 ^
25% of the activity was- located at CT3 and C-r-14. The mono-^ terpene 'postulate requires- activity at C-^,14, (XL 9, C^21 and
C<^16 of ajmaline ' C411.
On the Basis of these experimental findings, Leete chose 34 to expand the Schlittler^-Taylor acetate theory. As shown in
Figure 5a, the Leete precursor 46 which corresponds to Wenkert's
SPF unit is built up from three acetyl-CoA units, a malonyl-CoA 35 . . unit and formaldehyde. Leete was able to show positive incorporation of labelled malonic acid into ajmaline (41) and reserpine (42) after feeding experiments in R. serpentina.
However, the alkaloids possessed 74% of their activity at C-17 and none at C-18 and C-19. Hendrickson subsequently modified the acetate hypothesis in order to incorporate a polyketide- type biogenesis of the Leete precursor. Figure 6b shows the
Hendrickson modification as the linear condensation of five acetate units, followed by loss of the terminal methyl group from the ten carbon chain. Further condensation with a C^ unit followed by aldol cyclization and then a retro-aldol cleavage yields 47, which has the same carbon skeleton as
Leete's precursor 46 and Wenkert's SPF unit (36).
Battersby, working with both R. serpentina and C. ipeca- 32 37 cuanha ' was unable to confirm the results of Leete. His results showed random incorporation of labelled sodium acetate in the C9^c^Q un£t? °£ axx tn-e alkaloids studied. The end of 38
the acetate hypothesis was made official when Leete yerified
Battersby's work and withdrew his precursor from consideration.
In the same study, Battersby administered labelled sodium - 14 ^
CoAS^ 0^ HCHO 3 CH3COSC0A 0
OOC COSCoA OOCCH^COSCoA ( 46 )
Figure 6a. The Leete Acetate-Malonate Hypothesis for Indole
Alkaloid Biosynthesis.
5 CH3COSC0A
CoASOC
OHC OH
CoASOC CoASOC CHO
( 47 )
Figure 6D, The. Hendrickson Polyketi.de Modification - 15 -
formate to his plant systems and recovered the label only at
the indolic N-methyl of ajmaline (.41) and the aromatic O-methyl -
39 of cephaelme (43). This finding, later confirmed by Barton,
indicated that formate was merely labelling the C-^ pool in the plant systems and not participating in the buildup of an SPF unit (see Figure 5). More convincing evidence against the
Wenkert proposal of a carbohydrate metabolism derived precursor 40 was obtained when it was shown that radioactive shikimic acid
(17) was incorporated only in the indole nucleus of the alkaloids investigated. This was to be expected on the basis of the known biosynthesis of tryptophan as shown earlier in Figures 2 and 3.
Battersby was able to provide tentative confirmation for the Thomas-Wenkert monoterpene postulate by demonstrating low 14 incorporation of sodium mevalonate-2- C into ajmaline (41) and cephaeline (43) in R. serpentina and C. ipecacuanha.
Further confirmation was forthcoming from the work of Money 41 42 and Scott in Vinca rosea Linn., and the work of Arigoni in both V. rosea Linn, and V. maj or Linn..
More experiments along these lines, coupled with degradative data, demonstrated the intact incorporation of the mevalonate unit according to the pattern required by the Thomas-Wenkert 41-46 hypothesis (Figure 5b). Subsequently, it was shown that • n u v, a. IAO ,42,44,46-52 , , / * ov.\4 9,50 geraniol pyrophosphate (48a) ' ' and nerol (48b) could both serve as precursors to the Cg-C^Q unit. Moving further along the pathway, it was shown that loganin (51), now known to be derived from mevalonate (4 5) via geraniol (48a) 53-64 and/or nerol (48b), was incorporated intact into the - 16 -
( 53 )
Figure 7. The Early Stages of Indole Alkaloid Biosynthesis as
Proven by Experiment. 53 54 65 54 alkaloids of V. rosea, ' ' R. serpentina, and C. ipecac- 6 6 uanha. Moreover, loganin was found to co-occur with the 53 64 64 alkaloids of V. rosea ' and Strychnos nux vomica. It is
noteworthy that the stereochemistry at C-15 of yohimbine (32)
is identical to the stereochemistry at the corresponding
position of loganin (51) .
On either side of loganin (51) on the biosynthetic pathway
are deoxyloganin (50), and secologanin (52). Both have been
isolated from V. rosea, and have been shown to be specifically
incorporated into the alkaloids of V. rosea.^"^
Further, the intermediacy of the hydroxygeraniol deriva•
tive (49) has been demonstrated in V. rosea with randomization 70 71 of the label at the positions marked 2,6, ' indicating that oxidation at both of these carbon atoms is a necessary part of
the sequence to deoxyloganin (50).
With the origin of the Cg-C^Q unit on a sound experimental footing, the next problem was to determine how secologanin (52) was utilized by the plant system in the biosynthesis of an indole alkaloid.
Simultaneously and independently, vincoside (54) and 72 isovincoside (55) were isolated from V. rosea, and 73 74 75 strictosidine (55) from Rhazya stricta. X-ray analysis ' has proven the relative configurations about C-3 to be as shown.
In addition, it has been shown that vincoside (54) is specifi• cally incorporated by V. rosea plants into all three types of 72 76 77 indole alkaloids, ' ' and that it is itself derived from 72 tryptophan (25) and loganin (51). Isovincoside (55) could - 18 - not be incorporated into any of the alkaloids studied.
( 54 ) . ( 55 )
These results indicate the answer to the question: are the different families of indole alkaloids biosynthesized by different convergent pathways, or is there a single unifying sequential pathway, possibly with divergent branches? The fact that vincoside is specifically incorporated into all three families of indole alkaloids indicates that the former is probably not the case. Rather, it appears that vincoside is formed by a convergent pathway involving separate biosynthesis of tryptophan (25) and secologanin (52). The vincoside (54) molecule must then undergo the appropriate rearrangements to the various families of indole alkaloids. Vincoside (54) , then, serves as a convenient dividing line. The biosynthesis of vincoside (54) itself may be thought of as the early stages of indole alkaloid biosynthesis, while the subsequent rearrange• ments make up the later stages of indole alkaloid biosynthesis.
It can be seen that the corynanthe family of alkaloids can be derived directly from vincoside (54) in a straightforward manner involving no rearrangements. Thus, Figure 8 shows the derivation of geissoschizine (1), corynantheine (57), corynan- ^19 r
Figure 8. The Postulated Derivation of Corynanthe Alkaloids
from Vincoside (54). - 20 -
theine aldehyde (56). and ajmalicine C58) .
Proceeding further along the pathway, Wenkert proposed 7 8
in 1965 that, on stereochemical grounds, the strychnos, iboga
and aspidosperma alkaloids could be derived from the corynanthe
skeleton. He proposed a possible mechanism for the conversion
of the corynanthe skeleton to the strychnos system, as shown 79
in pathway A of Figure 9. Scott proposed a different
mechanism (Figure 9, pathway B) for the same conversion, but
so far experimental data is lacking for a decision to be made between the two pathways. 2 6
Earlier, Wenkert had put forth a postulate for the
subsequent rearrangement of the strychnos family of alkaloids
to the iboga and aspidosperma systems. As shown in Figure 10,
the iminium intermediate 60, which is analogous to intermediate
59 in Figure 9A, undergoes retro-Michael reaction to form the
pivotal intermediate 61. Intermediate 61, after conversion
to the corresponding acrylic acid 62 and subsequent ring
closure yields the iboga (63) skeleton. Alternatively,
intermediate 61 can form the corresponding partially reduced
intermediate 64 and undergo Michael addition to the acrylic system followed by transannular cyclization to form the aspidos• perma (65) ring system.
The postulated sequence of biosynthesis, that is, cory• nanthe -*• strychnos aspidosperma -> iboga alkaloids, was given experimental support when it was found that geissoschizine (1) is incorporated intact into the strychnos alkaloid akuammicine 8 0*- 8 2 8283 (66) in V. rosea, and co-occurs with it. ' Furthermore - 21 -
Figure 9. The Wenkert (A) and Scott (B) Postulates for the
Biosynthesis of Strychnos Alkaloids. coo' ( 63 ) ( 65 J
Figure 10. The Wenkert Postulate for the Biosynthesis of Iboga
and Aspidosperma Alkaloids. it was shown that the strychnos alkaloid stemmadenine (6),
analogous to Wenkert's intermediate 60, is a constituent
of V. rosea, ^ 83 and is incorporated intact8^* into the
iboga and aspidosperma alkaloids in that plant system.
(6)
80 82
Finally, Scott's work with V. rosea seedlings ' showed
that sequential analysis of the alkaloidal content of germinated
seedlings reveals the appearance of corynanthe alkaloids first,
followed by strychnos, then aspidosperma, then iboga alkaloids. 14
These results were confirmed by feeding tryptophan-2- C (25)
to germinated seedlings of V. rosea and observing the sequential uptake of radioactivity. Once again, the sequence corynanthe -»•
strychnos -> aspidosperma -»- iboga was observed.
Work in our laboratories concerned itself with the trans• annular cyclization required to form the aspidosperma and iboga alkaloids according to Wenkert's postulate (Figure 10). This 8 3—9 7 process was shown to be a facile one in vitro. For example, oxidation of quebrachamine (68) to its N(b) iminium salt, followed by cyclization and reduction, yields aspidospermidine
(69). Similarly, 16-carbomethoxycleavamine (8) can be converted to catharanthine (4). (8) (A)
Whether or not transannular cyclization was applicable in vivo, however, remained to be shown. Initial feeding experiments resulted in no incorporation of radioactively labelled 16-carbomethoxycleavamine (8), 6,7-dehydrovincadine
(70), quebrachamine (68) and vincaminoreine (71) into the 8 8 corresponding aspidosperma or iboga skeletons in V. rosea.
This preliminary indication of the non-utility of transannular cyclization was confirmed by administering tryptophan-3-^"4C
(25) to V. minor plants over varying time intervals and observing the ratio of activity in the tetracyclic alkaloids vincadine (72) and vincaminoreine (71) to that of the penta- 8 9 cyclic alkaloids vincadifformine (73) and minoyine (74).
This ratio was found to be relatively constant over a time period extending from four hours to fourteen days. The possibility that the constant ratio might be due to an equilibrium between the tetracyclic and pentacyclic systems was eliminated when it was shown that labelled minovine (74) transfers no activity to the tetracyclic alkaloids after a one - 25 -
week feeding period.
COOCH3
( 73 ) R=H
( lh ) R=CH3
The above evidence appeared to indicate the existence of a pivotal intermediate which could be converted in vivo to the tetracyclic and pentacyclic aspidosperma and iboga alkaloids each independently of the other, and without the necessity of a transannular cyclization. Thus, the scheme shown in Figure 2 6
11 appeared to be a viable alternative to the Wenkert postulate. It is noteworthy that the key intermediate, acrylic ester 75 strongly resembles the acrylic ester 62 contained in Wenkert's proposal. Since the dihydropyridinium system contained in intermediate
75, could be expected to be unstable and difficult to obtain synthetically, it was felt that the corresponding piperideine 90 76 named secodine by Smxth or the hydroxy ester 77 could be useful alternatives in biosynthetic investigations. This view 91 received support when Battersby proposed, on the basis of isotopic dilution experiments that 16,17-dihydrosecodin-17-ol - 26 ^
Figure 11. The Postulated Derivation of Aspidosperma and Iboga
Alkaloids from Intermediate 75. - 27 - occurs naturally in V. rosea and Rhazya orientalis.
Subsequently, the isolation of 16,17,15,20-tetrahydrosecodine
(78), 16,17-dihydrosecodine (79) and 16,17,15,20-tetrahydro- 90,92 secodin-17-ol (80) from plant sources was reported
COOCH3 COOCH-
( 78 ) ( 79 )
COOCH3
( 80 ) 3 16,17-Dihydrosecodin-17-ol-(ar- H) (77) was duly synthe- 93 94-97 95 98 sized and administered to V. rosea, V. minor, ' and 95-97 99
Aspidosperma pyricollum. ' In each case, deterioration of the plants was observed, and no incorporation of radioactivity in the alkaloids isolated could be detected. Thus it was felt that the transformation of the hydroxy ester 77 to the corres• ponding acrylic ester was probably not occurring in the plant systems under investigations. Consequently, the necessary 3 dehydration was performed in vitro, and secodine-(ar- H) (76) fed to the same plant systems. In this case, low but definite incorporations of activity were observed in the three alkaloidal families.93"95'98
Further studies using various forms of doubly labelled secodine and involving degradation of the alkaloids isolated gave strong indications that the secodine skeleton is incor- - 28 -
porated intact into vindoline (3), apparicine (81)
100 101 100 catharanthine (4), ' and vincamine (82).
( 83 )
The alkaloid apparicine (81) as well as the alkaloid
uleine (83) represent a novel branch of the indole alkaloid
families. It can be seen that both lack the usual ethylene bridge between the 8 position on the indole nucleus and the b-nitrogen of the alkaloid, and possess only a methylene bridge at that point. Up to this point, it has been assumed that tryptophan is the precursor of the indole nucleus of all indole alkaloids. This assumption appears to be open to question in the cases of apparicine (81), and uleine (83). In fact, 2 6
Wenkert proposed in 1962 that the biosynthetic pathway for uleine, as shown in Figure 12, involves glycosylideneanthranilic acid (84), a precursor of tryptamine analogous to enol-1(o- carboxylphenylamino-)-l-desoxyribulose-5-phosphate (23).
Condensation with an appropriately functionalized C-^Q unit, at that time thought to be the SPF unit (36), followed by conden• sation with methylamine or its equivalent and cyclization would H
( 81 )
Figure 13. The Djerassi Postulate for the Biosynthesis of
Apparicine (81) ^ 30. TV
then yield uleine C831. 102
Later, in 1965, Djerassi proposed that Wenkert's
intermediate 8 5 could also serve as a precursor in the biosynthesis of apparicine (81). Isomerization to the
exocyclic iminium species 8 6 followed by cyclization with
the indole nucleus as shown in Figure 13 would yield apparicine (81).
Neither one of the above postulates allows for apparicine
(81) or uleine (83) to be derived from tryptophan itself, and therefore, both regard these alkaloids as products of a separate convergent biosynthetic pathway rather than as products of some branch along the divergent pathway, which has heretofore been shown to be the case. Experimental evidence concerning the biosynthesis of apparicine (81) and uleine (83) was finally provided by our 3 laboratory when tryptophan- (ar- H) (25) was administered to the
A. pyricollum plant system. Significant incorporation of 99 activity could be detected in apparicine (81). Thus, it appeared that apparicine was probably tryptophan derived, 102 contrary to the postulate of Djerassi.
If tryptophan is the true precursor of the indole portion of apparicine (81), at least one of the two carbons in the tryptophan side chain must be extruded. To study this point, 3 14 s tryptophan labelled with H. in the aromatic ring and C in either C-2 or C-3 was fed to A. pyricollum. It was found that C-3 of tryptophan is incorporated into apparicine (81) 3 14 with retention of the R/ C ratio. On the other hand, over
97% of the label at C-2 was shown to be lost. ^ 31 ^
Stemmadenine (61, which, has already been shown to be an
important intermediate in the biosynthesis of the aspidosperma 8 0
and iboga alkaloids, was then considered as a possible inter•
mediate in the later stages of apparicine biosynthesis because
of its similarity of structure with apparicine, and because it
has been found to co-occur with apparicine in the fruits of A. pyricollum- -i-i . 103 104
A series of experiments (Figure 14) was run in order to
proyide some insight into the mechanism by which C-2 of trypto•
phan is extruded in the biosynthesis of apparicine (81).
Experiment 2 indicates that stemmadenine is a precursor to
apparicine, and therefore, as experiment 4 also suggests, that
extrusion of C-2 of tryptophan occurs well after condensation
of tryptophan with secologanin. Further,, experiment 3 suggests,
but does not prove, that the conversion of the functionality at
C-16 of stemmadenine to the exocyclic methylene of apparicine may be a concerted process, accounting for the low level of
incorporation of vallesamine (87).
At this point, a group of French workers"^5 put forth a postulate for the biosynthesis of apparicine (81), which appears
to be consistent with the above experimental evidence (Figure 15, 10 6 pathway A). Some time later, this same group published a communication postulating a unified system for the biosynthesis of apparicine (81), uleine (83), guatambuine (90), ellipticine
(89)., and olivacine (.88) , as shown in Figure 15, pathways A,B,C,
D and E-, respectively. However, no experimental evidence for this unified hypothesis has yet been published, and apparicine - 32 -
remains the only non-tryptophan type alkaloid for which, biosynthetic information is ayailable. - 33 -
CH. % Incorporation
Experiment No. ( 81 ) ( 83 )
1 DL-Try/ptophan- (Ar- H) 0.02 <0.0001
Figure 14. The Results of Incorporation of Various Possible Intermediates
into Apparicine (81) and Uleine (83). ( 88 )
Figure 15. The Potier-Janot Postulate for the Biosynthesis of
Non-tryptamine Alkaloids - 35 -
DISCUSSION
As was shown in the introduction, the biosynthesis of
indole alkaloids of the corynanthe-strychnos, aspidosperma
and iboga families is relatively well established experimentally.
However, much less is known about those alkaloids which do not
possess the usual two carbon chain between the 8 position on
the indole nucleus and the basic nitrogen.
The work described in Part A of this thesis concerns
itself with preliminary studies of the biosynthesis of
apparicine (81), olivacine (88), uleine (83) and guatambuine
(90). This work includes the isolation of samples of uleine 107
(83) and olivacine (88) from Aspidosperma olivaceum extracts, '
the development of ozonolytic degradation procedures for
apparicine (81) and uleine (83) and the investigation of the possible intermediacy of tryptophan (25) and secodine (76).
Part B concerns itself with the development of a synthesis of stemmadenine (6) which is suitably labelled for more detailed . biosynthetic investigations than had been hitherto possible.
Part C of this thesis discusses the results obtained with the labelled stemmadenine system as a precursor in Aspidosperma pyrico1lum. - 36 -
Part A
The extraction of samples of olivacine (88) and uleine
(83) from A. oiivaceum Mull.-Arg. extracts was accomplished
by following the procedure already- published by Gilbert and 108
Djerassi. This method consisted of thoroughly mixing the
extract with an aqueous solution of acetic acid (10%), filtering
the mixture through celite, and extracting the filtrate with
hexane, benzene and finally with chloroform. The aqueous
portion, after being adjusted to pH 8, was extracted with
chloroform and a second extraction followed at pH 10. It was
found that both extracts were similar, and could be combined
for further purification. Column chromatography through
alumina (activity III), followed by thin layer chromatography
on silica gel allowed pure olivacine (88) and uleine (83) to
be obtained. 109
It has been shown that olivacine (88), when treated
with methyl iodide to yield olivacine methiodide (9) followed
by catalytic hydrogenation yields guatambuine (90), as shown
in Figure 16. Further, work in our laboratory on the plant 95
system A. pyricollum made apparicine available in 100 mg quantities. Thus, samples of all the alkaloids of interest were
relatively easily available.
The reason samples of the alkaloids to be studied were required is threefold. First, they would provide a method of comparison with the alkaloids of A. australe to determine their presence or absence in that plant system. Second, they were - 37 -
needed as starting materials in degradation studies. Third,
they could serve as carriers in isotopic dilution studies.
That is, samples of the inactive alkaloids could be mixed with
the crude extracts obtained from a plant which had been fed
radioactively labelled compounds. The alkaloids which had
been added to the extract could then be re-isolated, purified
and checked for radioactivity. Thus, information could be
obtained on the biosynthesis of alkaloids present in the plant
in very small amounts without the necessity of conducting large
scale plant feedings.
( 90 )
Figure 16. The Conversion of Olivacine (99) to Guatambuine (90).
With the required alkaloids in hand, work was begun on 95 developing degradative procedures. Work in our laboratory
3 14 had already established that secodine (ar- H) and ( COOCH3) is incorporated into apparicine in A. pyricollum. As well, rv 38 -
secodine (19— 3Hl was being made available as a precursor."1"0^
Therefore, an ozonolytic degradation of apparicine, as shown
in Figure 17 appeared to be the method of choice. Ozonolysis
of apparicine would cleave the 15,16 and 18,19 double bonds
and liberate formaldehyde and acetaldehyde, respectively, and
without loss of the hydrogen at position 18. Steam distill•
ation of the ozonolysis reaction mixture into a saturated
solution of dimedone (92) resulted in the formation of
formaldehyde-bisdimedone (93) and acetaldehyde-bisdimedone (94).
The acetaldehyde derivative 94 could be selectively cyclized
to the tricyclic derivative 95, allowing the mixture to be
easily separable by thin layer chromatography on silica gel.
Alternatively, the acetaldehyde tricyclic derivative 95 could be selectively extracted from a basic solution of compounds 93 and 95. Acidification of the aqueous phase followed by extraction would afford the bicyclic formaldehyde derivative."'""'"''''
Thus, an easy degradation was at hand to determine whether or not secodine (76) is incorporated intact into apparicine.
Further, the identical ozonolysis procedure could also be applied to uleine (83). In this case, only the exocyclic methylene was cleaved, and the workup was greatly simplified. Steam distill• ation of the ozonolysis reaction mixture into a saturated aqueous solution of dimedone (92) resulted in formation of formaldehyde-bisdimedone (93). Evaporation of the solvent under reduced pressure followed by recrystallization from ethanol yielded the pure crystalline derivative in excellent overall yield. ( 9U ) ( 93 )
Figure 17. The Ozonolytic Degradation of Apparicine (81).
The conditions for the synthesis of secodine (76), as
shown in Figure 18, had previously been determined in our 93 95
laboratory ' . Nevertheless, xt was necessary to repeat
this synthetic sequence in order to obtain sufficient quantities
of radioactive secodine for the biosynthetic studies described
in this thesis.
The chloroethylindole 96 was condensed with an excess of
3-ethylpyridine in a sealed tube at 120° for 24 hours. The
resulting pyridinium salt was precipitated by pouring the mixture into a large excess of ethyl ether. Conversion of ( 76 )
Figure 18. The Synthesis of Secodine-(Ar-3H) and C14COOCH.J (26). - 41 -
the pyridinium salt 97 to the corresponding piperideine 98 was accomplished with sodium borohydride in methanol at 0°.
The ethyl ester 98 was then reduced with lithium aluminum
hydride in refluxing tetrahydrofuran to the corresponding
alcohol 99, which was treated with benzoyl chloride to form
the corresponding benzoate 100. Column chromatography on
alumina (activity III) afforded the pure benzoate 100 in about
70% overall yield from the starting chloroethylindole 96.
Treatment of the benzoate 100 with a fivefold excess of potassium cyanide at 95° afforded the nitrile 101 in yields which varied from 50% to 70%. For the purpose of introducing a radioactive label into the carbomethoxy group of secodine 14
(76), potassium cyanide- C was used to form the nitrile 101.
In this case, use of a fivefold excess of potassium cyanide with sufficient specific activity to be useful for biosynthetic purposes would be prohibitively expensive. This problem was 14 circumvented in the following manner. The potassium cyanide- C to be used in the experiment was diluted with inactive potassium cyanide to one equivalent, and the reaction initiated by adding the benzoate 100 dissolved in N,N-dimethylformamide (DMF) and raising the temperature to 95°. After a short time at this temperature, four more equivalents of inactive potassium cyanide were added to allow the reaction to proceed to completion. In this way, a higher specific activity could be obtained for the nitrile than when the entire dilution took place prior to the start of the reaction. The nitrile 101 was subsequently converted to the corres- - 42 -
ponding methyl ester 102 using the procedure developed by 113 Wenkert. The nitrile was dissolved in methanol whxch
contained 1% water. Saturation of this solution at 0° with
hydrogen chloride gas followed by allowing the mixture to stand
at room temperature for 4 8 hours resulted in formation of the
methyl ester 102 in good yield. At this point in the synthetic
sequence, tritium could be introduced into the aromatic ring
of the molecule by allowing the ester 102 to react with tritiated 3
trifluoroacetic acid (CF^COO H) at room temperature for 48 hours.
The tritiated methyl ester 103 could be recovered from this
reaction mixture in about 80% yield.
Formylation of the methyl ester 102 or 103, was accomplished
by formation of the enolate anion with sodium hydride in benzene
followed by addition of methyl formate to form the enol 104.
The enol 104 was then subjected to reduction with sodium 93 95
borohydride in methanol. It had previously been determined '
that this reduction required low temperatures in order to avoid
formation of the diol 105. This reaction was improved by lower•
ing the temperature to -4 5° and adding the sodium borohydride
in very small amounts over the course of the reaction. Although
the improved method was. somewhat tedious, involving regular
attention over a twelve hour time period, the undesired diol
105 could be almost totally avoided, with a consequent increase
in yield of the desired hydroxy ester 77.
The hydroxy ester 77 could be easily recrystallized, and was relatively stable to long term storage in a vacuum dessi- cator. Thus, when a plant feeding was to be undertaken, only - 43 -
the amount of hydroxy ester 77 needed was dehydrated for
immediate administration of secodine (7 6) to the plant system.
This dehydration was accomplished in the manner previously 93 95
reported. ' Thus, the hydroxy ester was treated with sodium
hydride in benzene, then quickly passed through a short column
of alumina (activity III). In this manner, secodine (76) containing tritium in the 14
aromatic ring and C in the carbonyl of the carbomethoxy group
was obtained. In order to provide more detailed biosynthetic
results, an alternative synthesis of secodine (76) was developed
in our laboratory and this allowed tritium to be introduced at
C-19.^^ This synthesis, shown in Figure 19, allows the intro•
duction of 3-ethylpyridine at a relatively late stage in the
synthetic pathway, making this route more economical than that
previously developed.
The synthesis summarized in Figure 19 has already been
discussed elsewhere"*"^ and will be dealt with only briefly here.
Commercially available ethyl indole-2-carboxylate (106) was
reduced to the corresponding alcohol 107 with lithium aluminum hydride. Formation of the corresponding benzoate 108 with benzoyl chloride followed by displacement with potassium cyanide yielded the nitrile 109. The nitrile was treated with methanolic hydrogen chloride to provide methyl-2-indolyl acetate (110).
Reaction of the methyl ester 110 with ethylene oxide and stannic chloride yielded the hydroxyethyl indole derivative 111. Treat• ment of the hydroxyethylindole 111 with p_-toluenesulfonyl chloride followed by 1-(3'-pyridyl)-ethane-(1- H) (116) provided - 44 -
Figure 19, The Synthesis of Secodine- (19-3H) the corresponding pyridinium salt 112. Reduction of the pyridinium salt 112 with sodium borohydride yielded the desired piperideine 103. The piperideine thus obtained could then be converted to secodine (76) in the manner described in Figure 18. 3
The synthesis of 1-(3'-pyridyl)-ethane- (1- H) has also been described elsewhere1"^ and will be only briefly described here. As shown in Figure 20, 3-acetylpyridine (113) was treated 3 with sodium borotritiide to form 1-(31pyridyl)-ethanol- (1- H)
(114). Treatment with acetic anhydride to form the corresponding acetate 115 followed by catalytic hydrogenolysis resulted in the 3 formation of the desired 1-(31-pyridyl)-ethane-(1- H) (116). 3, . 0 li UH
(113) (114)
3 3H OAc H H
Pd/C
H 2
( 115 ) (116)
Figure 20. The Synthesis of 1-C3'-pyridyl)-Ethane-'Cl- H) C116) . - 46 -
With all the necessary compounds at hand, it was now possible to commence biosynthetic feeding experiments to determine the possible intermediacy of tryptophan and/or
secodine in the biosynthesis of apparicine (81), uleine (83), olivacine (88) and guatambuine (90).
It was decided to continue the work which had already 95 been done on the biosynthesis of apparicine (81) in 3 14
A. pyricollum. To this end, secodine-(19- H, COOCH^) was administered hydroponically to root cuttings of A. pyricollum for five days. Isolation of apparicine by the procedure 95 previously described was followed by ozonolytic degradation and isolation of the dimedone derivatives of acetaldehyde and formaldehyde. The results of this experiment, as well as those 95 114 experiments previously described are shown in Tables 1 and 2.
Experiment 1 provided an indication that the indole nucleus of secodine was being used by the plant system in the biosynthesis of apparicine (81). Furthermore, experiments 2 and 3 showed that the indole nucleus and the carbomethoxy group on C-16 were being incorporated as an intact unit.
Since the original Potier postulate for the biosynthesis of apparicine"^5 (Figure 15, pathway A) called for the loss of the carbomethoxy group and retention of C-17, a modification of the postulate seemed to be necessary. Recognition of this fact 106 was provided by the French group in a subsequent publication.
Experiment 4 provided the first indication that the secodine molecule as a whole was being incorporated into apparicine (81).
If the overall role of secodine in indole alkaloid biosynthesis Table 1.. Results of Incorporation of Secodine (76) into Apparicine (81)
Experiment Compound Fed % Ratio of Activity Ratio of Activity Number Incorporation „^ 3TT ,14„ , . . . 3„ ,14, Fed H/^C Isolated H/ C
95 3 I Secodine-(Ar- H 0.01
2 95 14 Secodine- ( COOCH3) 0.01 _ _
95 3 3 14 Secodine-(Ar- H, COOCH3) 0.015 8.7 8.4
95 4 Secodine-(3,14,15,21-3H, 0.009 4.2 2.2 14
COOCH3)
5 3 14 Secodine-(19- H, COOCH3) 0.024 3. 98 2.05 Specific Activities Associated with the Experiments in Table 1.
Experiment Activity Fed Specific Activity Fed Specific Activity Number Isolated
3H ^"4C 3H dpm/mmol "*"4C dpm/mmol 3H dpm/mmol "^4C dpm/mmol
1 2.57xl08 1.82xl010 7.68xl05
2 7.06xl06 1.19xl09 1.06cl04
3 l.llxlO8 1.28xl07 l.lOxlO10 1.28xl09 1.6xl05 1.94xl04
4 4.42xl07 1.07xl07 9.82xl09 1.27xl09 3.03xl04 1.35xl04
5 6.2xl06 1.56xl06 1.18xl010. 1.98xl09 2.13xl03 1.06xl03 ~ 49 -
C CH2OH°OCH
(6)
ChhOX COOCI-LV CH?C COOCH3
COOCH3 ( 75 ) ( 117 )
aspidosperma apparicine (81) and iboga uleine (83) alkaloids guatambuine (90) (Figure 11) olivacine (88) ellipticine (89) (Figure 15)
Figure 21. The Proposed Relationship of Secodine (76) to
Stemmadenine (6) in Indole Alkaloid Biosynthesis. - 50 - is as shown in Figure 21, then the oxidation of secodine in the iminium species 75 followed by cyclization and rearrangement of the resulting enamine double bond to form stemmadenine 16) would be expected to result in the observed lowering of the
3H/^"4C ratio in experiment 4. 95
With the data from experiments 1-4 already available, it remained to provide degradative evidence for the intact in• corporation of secodine into apparicine. The degradations which were carried out for experiments 2 and 5 are shown in Table 3.
Since the tritium at the C-19 position of secodine in experiment
5 would be expected to be present in equal amounts in both the
R and S configurations, it might be expected that stereospecific enzymic conversion of the enamine 118 to stemmadenine (6) would result in the observed loss of 50% of the tritium label relative to 14C.
Table 3. Specific Activities Associated with the Ozonolytic
Degradation of Apparicine (81) in Experiments 2 and
5 (Table 1).
u i
^ CH20 CH3C0
( 81 ) 4 95 1.06x10 dpm/mmol Cexpt.2) 1.05x10
1.06xl03 dpm/mmol (.expt. 5) (14C) 1. 04x10'
2.13x103 dpm/mmol Cexpt.5) C3H) 2.13x10" - 51 -
Once the data for apparicine had been obtained, it was
decided to investigate the A. australe plant system with the
hope of obtaining biosynthetic information on uleine (83).,
olivacine (88) and guatambuine (90). The precursors to be
studied were fed hydroponically as their acetate salts to whole
A. australe plants for 5 days. The crude alkaloidal extract was
diluted with inactive samples of the desired alkaloids, which
were re-isolated by thin layer chromatography and recrystallized
to constant activity. The results of these experiments are
shown in Tables 4 and 5.
Experiment 1 provided an indication that olivacine and
guatambuine might be derived from tryptophan, but no activity
could be detected in the uleine isolated in this experiment.
Experiments 3-5 involving doubly labelled tryptophan resulted
in either erratic ratios or extremely low levels of radioactivity
in the alkaloids isolated. If the Potier postulate for the
biosynthesis of these alkaloids (Figure 15) is correct, C-2 of
tryptophan should be lost in the cases of apparicine and oli•
vacine, and retained in the cases of uleine and guatambuine.
Although the incorporations are extremely low, the ratios of 3 14
H/ C activity in experiments 3 and 5 do show an increase in
the case of olivacine. The increase observed for apparicine
had already been established in the A. pyricollum plant system."^
As well, it can be seen from experiment 3 that C-2 of tryptophan
appears to be retained in the biosynthesis of uleine. The case
of guatambuine is, however, ambiguous.
Experiment 4 shows that C-3 of tryptophan is incorporated Table 4. The Various Compounds Fed to A. australe
Experiment Compound Fed Weight Wet Plant Activity Fed Ratio Fed Number Fed mg Weight g 14 3 ,14 C dpm H dpm H/ C
3 1 7 Tryptophan-(Ar- H) 6. 01 14. 0 8.28xl0
3 14 2 7 7 Secodine-(Ar- H, COOCH3) 3. 43 15. 7 3.0xl0 8.7xl0 2. 9
3 14 3 7 8 Tryptophan-(Ar- H,2- C) 7. 05 33. 7 6.25xl0 1.90xl0 3. 0 to
3 14 4 7 8 Tryptophan-(Ar- H,3- C) 9. 52 25. 0 8.45xl0 2.97xl0 3. 5
3 14 5 8 8 Tryptophan-(Ar- H,2- C) 17. 8 51. 0 1.33xl0 4.18xl0 3. 15
3 6 Secodine-(3,14,15,21- H, 3. 28 27. 0 1.09xl07 2.19xl07 2. 01 14 COOCH3) Table 5. Incorporation Results Associated with Table 3
Experiment Alkaloid Specific Activity % Incorporation Ratio Number (Dilution mg) Isolated (dpm/mmol)
14 14 14C 3H c 3H V c
* 1 Olivacine 7 (6.6) 2.01X10 241 Apparicine (0) 2.0x10;: 208 Uleine (8.1) 5.19x10^ 001 — _ _ _ Guatambuine (8.0) 4.28x10 • 00534 * 2 Olivacine (4.9) 2. 66x10^ 1.04x10?, 0093* 013 3.90 Apparicine (0) 1. 36x10 2.75x10 0391 # 0274 2.04 Uleine (10.0) not„countable ~ <. 001 <. 001 D b Guatambuine . (10.0) 3. 35xlO 7.55xlO <. 001 <• 001 2.25
3 Olivacine (7.7) 2. 40xl0J? 1.72x10^ 0005^ • 0011 7.18 Apparicine (0) 6. 89xlOJ? 1.06x10^ 0015 078* 154 Uleine (7.9) 7. 42x10;? ' 2.14x10^ m 0011 0011 2. 97 Guatambuine (9.2) 8. 24x10 3.34x10 • 0015 • 0022 4. 06 5* 4 Olivacine (11.0) 1. 63x10:: 9.09x10^ 001 001 5. 57 a Apparicine (0) 9 .0x10 2.64xl0 # 127 105 2. 90 Uleine (4.7) not countable <. 001 <. 001 Guatambuine (5.1) not countable <. 001 <• 001 5* 5* 5 . Olivacine (30.1) 1. 05x10 7.24x10 <. 001 <. 001 6.90 Apparicine (0) (not counted) Uleine (8.2) not countable <. 001 <• 001 Guatambuine (10.1) not countable <. 001 <• 001 * * 7 6 Olivacine (13) 8. 34x10^ 1.85X10 . 146 * • 160 * 2.22 Apparicine (0) 1. 67x10 4.17x10 # 00365 ^ 00508 2.50 Uleine (14) not countable <. 001 <. 001 Guatambuine (10) not countable <. 001 <. 001 * not constant activity 54 -
without loss into apparicine, as had been shown in A. pyri- 104
collum. The Potier postulate (Figure 15) predicts that
C-3 of tryptophan should be lost in the biosynthesis of
olivacine, guatambuine and uleine. While no data could be obtained for uleine and guatambuine, olivacine did show a rise 3 14
in the H/ C ratio in agreement with theory.
Experiments 2 and 6, involving secodine as the compound
fed yielded no definitive data. This is partly due to the low
activities associated with the isotopic dilution technique
required to handle the small amounts of the alkaloids being
studied. As well, Figure 21 shows that in the proposed pathway
from secodine to stemmadenine, the biosynthetic pathway which
has been established for the aspidosperma and iboga alkaloids
must be reversed. Although this reversal appears to take place
in the case of apparicine, it does so with relatively poor
efficiency, as can be seen in Table 1.
Thus, the use of secodine as an intermediate in biosynthetic
investigations concerning olivacine, uleine and guatambuine, can
be shown to have four main disadvantages.0 First, secodine is 115
relatively unstable. It has been shown that secodine under•
goes a Diels-Alder dimerization to form presecamine (119). It 95
has been estimated that this dimerization prevents roughly
30% of secodine fed to the plant from being taken up as the
monomer. Second, secodine itself is not actually on any pro• posed biosynthetic pathway. Rather, a dehydrogenation must take place to form the iminium species 75, which is thought to be a true intermediate. ( 119 )
Third, in order to be incorporated into olivacine, uleine, apparicine and guatambuine, the iminium species 75 must proceed two steps backwards on the established aspidosperma-iboga biosynthetic pathway in order to be inserted into the pathway leading to the desired product. Fourth, it can be seen from
Table 1 that secodine is incorporated very poorly even in apparicine. Whether this is due to a combination of the first three factors, or to some unknown factor, such as difficulty in the transportation of secodine to the site of biosynthesis, has not been determined.
By way of contrast, it can be seen from Figure 14 that in
A. pyricollum, stemmadenine (6) is incorporated into apparicine about 50 times more efficiently than secodine. On further con• sideration, it can be seen that stemmadenine would be an attrac• tive intermediate for the plant feedings involving the "non- tryptamine" type alkaloids for three reasons.
First, stemmadenine is quite stable and can be stored for long periods of time as a crystalline solid without extensive decomposition. Second, stemmadenine is directly on the bio• synthetic pathway proposed by Potier (Figure 15), and need not be altered by the plant system in order to be incorporated.
Third, stemmadenine has already been shown to be relatively - 56 -
efficiently incorporated into apparicine.
Thus, it can be seen that stemmadenine is an.extremely
attractive compound for. biosynthetic investigations. However,
the only source of stemmadenine up to this point has been 116
extraction from plant sources which were unavailable to
our laboratory in significant amounts. Even if stemmadenine
could be obtained, only a limited amount of biosynthetic infor• mation could be obtained, since only the aromatic ring and the
carbomethoxy methyl group could be labelled with tritium, and 14
only the carbomethoxy methyl group could be labelled with C.
As well, the carbomethoxy methyl group is expected to be lost in
the biosynthesis of all the "non-tryptamine" alkaloids.
Thus, the most practical course to take appeared to be to devise a synthesis of stemmadenine which would allow the molecule
to be radioactively labelled in such a way that, with the proper degradative procedure, it could be determined whether or not the stemmadenine skeleton is incorporated intact into apparicine, uleine, olivacine and guatambuine. This synthesis is the subject which is discussed in Part B. •r- 57 -
Part B
Stemmadenine C61 was first isolated from Stemmadenia 116 donnelT-smithii CRose) Woodson in 1958 and was shown to 117 118 possess structure 6 in 19 62. ' At present, only the absolute configurations at C-15 and C-19 have been estab- 79 118 lished, .' while the absolute configuration about C-16 remains to be established.
c^HCOOCH3
(6)
An interesting correlation between the structures of 79 stemmadenine (6) and akuammicine (66) was provided by Scott.
As shown in Figure 22, preakuammicine (2) was isolated from seedlings of Vinca rosea and converted into a separable mixture of akuammicine (66) and stemmadenine (6) by the action of methanolic sodium borohydride. As well, loss of formaldehyde by preakuammicine to form exclusively akuammicine was found to occur upon treatment with methanolic sodium methoxide. These results indicate that stemmadenine and akuammicine possess the same configuration at both. C-15 and C-19. Moreover, the conyersion of preakuammicine to stemmadenine provides the basis for a biogenetic type synthesis of stemmadenine. - 58 -
Figure 22. The Correlation of Akuammicine (66) and Stemmadenine
(6) via Preakuammicine (2).
From a biogenetic standpoint, there are two possible
entries into the nine-membered ring system of stemmadenine. The
reaction of an indolenine such as preakuammicine with sodium
borohydride.via the iminium 120 to form the indole system 6 119 120
(Figure 22, pathway a) has precedent in the literature. '
As well, the a-methylene-indoline minovine (74) has been shown
to react with sodium borohydride in refluxing acetic acid to 121 yield vincaminoreine (71). Thus, either akuammicine or preakuammicine could serve as attractive intermediates in a biogenetic type synthesis of stemmadenine. The use of akuammicine would require the introduction of the hydroxymethylene group at
C-16 after formation of the nine-membered ring.
Originally, work in our laboratory was directed towards the formation of 19,2 0-dihydropreakuammicine (121) as shown in Figure 122 23. Thus 18-,9-dihydrocorynantheine (122) was oxidized with tert-butyl hypochlorite to yield the 7-chloroindolenine 123 which- was then converted to the rhyncophylline enol ether 12 4 with sodium methoxide in methanol. Hydrolysis of 12 4 with aqueous acetic acid to the corresponding lactam 12 5 was followed by further hydrolysis to yield O-desmethylrhyncophy- lline (126). The free aldehyde was then protected as the corresponding ethylene acetal 127 and treated with Meerwein's reagent to yield O-desmethylrhyncophylline ethyl enol ether ethylene acetal (128).
Although the cyclization of an enolate anion in a system 123 such as compound 128 to yield the corresponding indolenine 124 or a-methylene-indoline has been shown to work in some cases, repeated attempts to cyclize enol ether 128 to form the ethylene acetal of the aldehyde corresponding to 19,20- dihydropreakuammicine (12 9) were unsuccessful. This result was 125 confirmed by Winterfeldt and the enol ether cyclization approach was abandoned.
This thesis concerns itself with more recent efforts to synthesize stemmadenine using commercially available strychnine
(29) as starting material, and proceeding by way of its degra• dation product Wieland-Gumlich aldehyde (130).126'127
The degradation of strychnine to Wieland-Gumlich aldehyde 126 127 (WGA) has been known since 1932. ' The sequence was improved 12 8 by Anet and Robinson in 1955, and finally thoroughly studied 12 9 and optimized by Schmid and Karrer in 1969. As shown in
Figure 24, strychnine (29) is treated with sodium ethoxide in the presence of isoamyl nitrite, and the product crystallized as ( 121 )
Figure 23. The Attempted Synthesis of 19,20-Dihydrostemmadenine
(121) . - 61 - isonitrosostrychnine hydrochloride (131). Reaction of isonitro• sostrychnine hydrochloride with thionyl chloride followed by quenching in ice yields N(a)-cyanoformyl-WGA hydrochloride (132) as a crystalline compound. Sodium carbonate hydrolysis and chromatographic purification of the resultant product provides crystalline WGA. Further purification of WGA thus obtained can be carried out by recrystallization in either chloroform or benzene.
Figure 24. The Degradation of Strychnine (29) to Wieland-Gumlich
Aldehyde (130).
130
Once WGA was obtained, four basic synthetic problems . remained. First, the C-18 hydroxy group must be removed in such a way that tritium can be introduced at C-18. Second, in order for the required nine-membered ring system to be formed, either the 1,2- or the 2,16-dehydro species must be obtained. Third, - 62 -
XI7 must either remain as the. aldehyde or be further oxidized to the corresponding methyl ester. Fourth, a way must be found to introduce the remaining carbon atom at C-16 in radioactive form. As an added complication, since the stereochemistry about
C-16 in stemmadenine has not been established, a suitable synthesis must be able to provide either or both configurations at C-16 in the final product.
In researching the vast amount of published literature concerning strychnos alkaloids, it became immediately apparent that appropriate use of nomenclature would be a problem. Many alkaloids were discovered and named independently, only to be found identical in later years. Still others, for example 133 have been named after the parent dimer (nor-hemi-dihydrotoxi- ferine) or after a closely related monomer (18-desoxy-Wieland-
Gumlich aldehyde), while an equally closely related monomer is found as nor-fluorocurarine (134).
(133 ) (134) '
In this thesis, the system of nomenclature proposed by Janot 131 and Le Men as outlined by Edwards and Smith will be used. The curan system 135, with no stereochemistry specified at positions 132 2, 16 and 20, is numbered following Bernauer et al. Thus, — 63 —
Wieland-Gumlich. aldehyde would become 18-hydroxy-2g , 16a-cur-
19-ene-17-al hemiacetal. Due to their pivotal nature in the
work described here, Wieland-Gumlich aldehyde (WGA), nor-
fluorocurarine (134) and akuammicine (66) will be referred to
as such. All other compounds in the curan system will be named
as described above.
Figure 25 represents a summary of some of the known
chemistry of WGA and its derivatives which appeared to be most
useful for our purposes. The reaction of WGA with sodium boro•
hydride forms 23,16a-cur-19-en-17,18-diol (136).128 The diol
136 can then be converted to 23,16a-cur-19-en-17-ol (137) in 132 133
exther of two different ways. Bernauer et al ' were able
to prepare the allylic bromide 138 with hydrogen bromide gas
dissolved in acetic acid. The allylic bromide 138 was then
treated, without purification, with zinc dust in acetic acid
containing a small amount of methanol to increase the solubility
of the bromide. The resultant product was the desired acetate
139 obtained in overall 65% yield. Hydrolysis of the acetate
139 could then be accomplished in excellent yield with methanolic potassium hydroxide. This procedure could be easily modified to introduce a tritium atom at C-18, thus solving one of the basic problems outlined above. Alternatively, a less tedious and higher yielding (85%) 134 procedure was devised by Wieland. Using a specially prepared palladium on charcoal catalyst, the diol 136 could be selectively hydrogenolyzed in an aqueous solution of acetic acid and hydro• chloric acid to yield the desired alcohol 137. This procedure - 64 -
Figure 25. A Summary of Some Known Reactions in the Wieland-
Gumlich Aldehyde Series. ~ 6 5 — was not considered for tritiation purposes for two reasons.
First, to conduct the hydrogenolysis with tritium gas would be technically more difficult than the alternative approach via the bromide 138. Second, when using tritium gas, there is a danger of random hydrogen-tritium exchange as in the Wilzbach 135 . . labelling technique. If even a very small amount of tritium were introduced anywhere but C-18, the labelling procedure would be useless in biosynthetic experiments designed to demon• strate intact incorporation of stemmadenine. In practice, the catalytic hydrogenolysis approach was the method of choice when unlabelled alcohol 137 was desired. As well, the identity of the products obtained by both procedures served as an additional confirmation of structure. 136
In 1959, Wieland reported that Oppenauer oxidation of the alcohol 137 with potassium tert-butoxide and benzophenone in benzene yielded 23,16a-cur-19-en-17-al (134) (nor-fluorocurarine). 134
In a later publication, Wieland reported that the oxidation could be made to proceed directly to nor-fluorocurarine (134) by heating the alcohol 137 with lithium tert-butoxide, benzophenone, 137 and nitrobenzene at 90° in a sealed tube. Boekelheide reported in 1964 that'oppenauer oxidation of the alcohol 137 with potassium tert-butoxide and benzophenone in refluxing benzene with "rigid exclusion of air" gave nor-fluorocurarine (134) as the major product, along with small amounts of the saturated aldehyde 133.
Boekelheide felt that this result, together with the apparent stability of the isolated saturated aldehyde 133 to exposure to air, indicated that formation of the a-methylene-indoline chromo- - 66 - phore of 134 was a direct result of the Oppenauer reaction conditions, probably via the mechanism shown in Figure 26.
\ (HO ) ( 134 ) Figure 26. The Boekelheide Mechanism for the Oppenauer Oxidation
of the Aldehyde 133 to Nor-fluorocurarine (134).
Initial Oppenauer oxidation of the indoline to form the indolen- inium species 140 is followed by loss of a proton and rearrange• ment to form the a-methylene-indoline 134.
It was found that, in fact, either the unsaturated aldehyde
134 or the saturated aldehyde 133, or a mixture of both, could be obtained, depending on the severity of the conditions used. The results of a series of experiments concerning the Oppenauer oxi• dation of the alcohol 137 are shown in Table 6.
Table 6. Effects of Differen t Reaction Conditions on the Oppenauer Oxidation of the Alcohol 137.
Experiment Time % Yi eld Number Solvent Base Temperature (hrs.) (133) (134)
1 benzene t-BuOK reflux 4 10 20 2 benzene t-BuOK room temp. .25 reflux 1.25 39 10 3 toluene t-BuOLi 95° 72 10 47.5 4 toluene t-BuOK 90° * 0 41 3 * 2 equivalents of nitrobenzene were added after 3 hours and the reaction was stopped 5 minutes later. ^ 67 -
Focperiments 1 and 2 demonstrate that, other factors being equal, the proportion of unsaturated aldehyde 134 in the reaction mixture increases as the reaction time increases. In an effort to force the reaction to completion, two methods were tried. First, use of lithium, rather than potassium, tert-butoxide with toluene as solvent required 72 hrs. for completion even at 95° (experi• ment 3). As well, there was still an appreciable amount of saturated aldehyde 133 after the reaction was stopped. However, when potassium tert-butoxide was used as the base in toluene at a temperature of 90° (experiment 4), no starting material remained, and only saturated aldehyde 133 could be detected by thin layer chromatography. When two equivalents of nitrobenzene were added, immediate conversion of the saturated aldehyde 13 3 to the unsat• urated aldehyde 134 was observed, and no saturated aldehyde could be detected in the product mixture. This last result, when con- 134 sidered along with Wieland1s method indicate that nitrobenzene is taking an active part in the conversion of 133 to 134. This 138 is reminiscent of the Skraup quinoline synthesis where sodium nitrobenzene sulfonate is used as a dehydrogenating agent which promotes the formation of an aromatic system.
Once nor-fluorocurarine (134) was obtained, there appeared to be two basic pathways by which it could be converted to stemmadenine (6). As shown in Figure 27, pathway A, oxidation of the aldehyde to its corresponding ester would yield akuammi• cine (66), which could be converted to the indole 141 with sodium borohydride in acetic acid. Introduction of the required hydroxy- methylene group at C-16 would complete the sequence. - 68 -
Figure 27, The Routes Considered for the Conversion of Nor-
fluorocurarine (134) to Stemmadenine (6). - 69 -
Since conyersion of nor-fluorocurarine (134) directly
into its indole derivative with, sodium borohydride in acetic
acid would likely be accompanied by reduction of the aldehyde
to its corresponding alcohol, it would be necessary to protect
the aldehyde as shown in pathway B. Two possible modes of
protection of the aldehyde can be enyisioned. First, the
a-methylene-indoline chromophore could be retained, as in 142.
An example of this would be formation of the ethylene acetal
derivative 142 (X = OCH2CH20). Second, an indolenine enol
derivative could be formed, such as the enol acetate 143 (X =
OAc). Conversion of either of these two compounds to their ring
opened indole derivatives 144 and 145 followed by loss of the
protecting group would yield the indole aldehyde 14 6. Intro•
duction of a carbomethoxy group at C-16 followed by reduction of
the aldehyde would yield the desired stemmadenine system. More•
over, if pathway A were shown to exhibit a predominance of one
isomer at C-16 in the final product, pathway B would be expected
to yield a predominance of the other isomer.
Before describing the attempts to bring the pathways described in Figure 26 to fruition, it would be worthwhile to discuss some of the more interesting properties of nor-fluoro- curarine (134). The N(b) methochloride fluorocurarine (134a) is a natural product first isolated in 1941 from a calabash- curare preparation (used by South American Indians as an arrow 139 140 poison and packed in calabashes or gourds). ' Later, it was identified chromatographically in extracts from the bark of 141 Strychnos mitscherlischil. It is referred to in the literature 7Q .~ both. .as. C~£luorocurarine and as C^.curarine III, and the free 140 base as nor-C-fluorocurarine and nor-C-curarine III.
The ultraviolet spectrum of nor-fluorocurarine (134) is most unusual, having a very intense absorption at 363 nm. It was postulated during the work on structure elucidation of the natural product that this absorption was due to an equilibrium 142 between the two tautomers 134 and 147, which would provide
for conjugation with the aromatic ring and increased stabilization of the system as a whole. . This is in marked contrast with akuammicine (6.6). which would not tautomerize as readily, and displays an equally strong absorption, but at 326 nm. Furthermore,
the absorption band for the carbonyl in the IR spectrum of 134 —T appears at 1650 cm , which is unusually low for an a,g-unsatura- 143 ted aldehyde.
The NMR spectrum of 134 (Figure 28) is particularly worthy of note due to its exceptional resolution of the complex splitting pattern of the protons on the non-indoline carbon atoms. The. predominant features are a 1 proton singlet at 9.396 (CHO), a 1
proton quartet (J - 7 Hz) at 5.436 (~CHCH3) and a .3 proton doublet
CJ - 7Hz) containing further fine splittings at 1.696 (=CH-CH3) . ^ 71 -
With, nor-f luorocurarine (134) in hand, it was possible to
investigate the pathways shown in Figure 27. It was decided to
attempt to oxidize nor-fluorocurarine to an unsaturated ester 144 usxng the method developed by Corey. Treatment of nor- fluorocurarine (134) with sodium cyanide in acidic methanol with 145
a twenty fold excess of activated manganese dioxide resulted
in a complex mixture of seven compounds, none of which displayed
the UV absorption at 326 nm, characteristic of akuammicine (66).
When the same reaction was attempted in neutral methanol, starting
material was recovered along with three compounds which once again
lacked the characteristic UV absorption at 32 6 nm. When the
oxidation was attempted with silver (II) oxide in tetrahydrofuran
and water in the presence of sodium cyanide, followed by treatment with diazomethane, only decomposition of the starting material was
observed.
A variation of the above sodium cyanide assisted oxidations was also attempted. Instead of methanol, diethylamine can be used as solvent. The oxidation then proceeds to the corresponding 146 a,3-unsaturated amide. When nor-fluorocurarine was treated with an excess of active manganese dioxide and sodium cyanide in diethylamine, no products having the expected spectral properties of the desired product could be found in the reaction mixture.
A possible rationalization for the failure of the oxidations involving sodium cyanide can be found in the work of Schmid et 129 al. As shown in Figure 29, the cyanohydrin of 16-epi-WGA (148) under basic aqueous conditions cyclizes and hydrolyzes to the a-hydroxy lactam 149. Figure 28. The Nuclear Magnetic Resonance Spectrum of Nor-fluorocurarine (134). Figure 29. The Reaction of 16-epi-WGA with HCN.
In the case of nor-fluorocurarine (134) the situation is
made even more complex by its known propensity to deformylate 147
under acid conditions, forming the corresponding desformyl
indolenine.
At this point, it was decided to abandon attempts to
oxidize nor-fluorocurarine and try to make use of some of its
unique properties. Specifically, it was decided to try to form
some derivative of the aldehyde moiety which would allow it to
survive the conditions required for formation of the nine- membered ring in compounds 144 and 145, as shown in Figure 27,
pathway B. The easiest derivative to prepare was felt to be the enolate - 148 anion (143, X = 0 ). It has been shown that enolate anions
of aldehydes are not reduced by sodium borohydride. Still
another advantage of enolates is that regeneration of the
aldehyde requires simple reprotonation. Thus nor-fluorocurarine - 74 -
(134) was. treated with, a suspension of sodium hydride in
tetrahydrofuran. Although the vigorous evolution of bubbles
indicated that the enolate had been formed, no reaction occurred
when sodium borohydride was added, even after heating to 50°.
Similarly, when nor-fluorocurarine was treated with lithium
tert-butoxide in tert-butanol followed by sodium borohydride,
starting material was recovered unchanged after stirring for
36 hours at 35°. However, when methanolic sodium methoxide was
used, formation of the nine-membered ring was observed along with
reduction of the aldehyde to form descarbomethoxy stemmadenine
(150) .
CH20H
(15 0 )
The structure of compound 150 was apparent from the disappear•
ance of the carbonyl absorption in the IR spectrum, the presence
of two exchangeable protons in the NMR spectrum, the presence of
an indole absorption pattern in the UV spectrum and a consistent molecular ion and fragmentation pattern in the mass spectrum.
Thus, although the metal enolate of nor-fluorocurarine could be
shown to protect the aldehyde from borohydride reduction with the
use of a sufficiently strong base, the use of such a basic medium
appeared to preclude the formation of the desired nine-membered
ring. Further attempts to form either the enol acetate or the - 75 - enol ether of nor-fluorocurarine were not met with success. 149
It has been shown in model studies that compounds containing an aldehyde conjugated with an a-methylene-indoline can form the corresponding dienamine with relative ease. Thus, refluxing nor-fluorocurarine in benzene containing cyclohexyl- amine and a catalytic amount of acid with azeotropic removal of water resulted in formation of a bright yellow product 151.
Compound 151 showed the expected bathochromic shift in the UV spectrum, having a strong absorption at 387 nm. As well, the
NMR spectrum contained no aldehydic proton signals, although a new singlet was observed at 7.40 6 (NCH = C). No N-H absorptions appeared in the IR spectrum and an a,3-unsaturated imine absorp• tion was found at 1630 cm ^. Finally, the mass spectrum showed the expected molecular ion (m/e = 373).
( 153 )
Figure 30... The Reactions of Nor-f luorocurarine (134) with
Cyclohexylamine, Pyrrolidine, and Morpholine. - 76 -
While reaction of nor-fluorocurarine with morpholine under
similar conditions led only to recovery of starting material
along with some decomposition products, it was found that
reaction with pyrrolidine occurred with relative ease. In fact,
it was found that formation of the pyrrolidine condensation
product occurred without acid catalysis merely by involving a
Soxhlet extractor containing molecular sieves with a refluxing
benzene solution of nor-fluorocurarine and pyrrolidine. While
the UV spectrum of the reaction mixture eventually displayed
the expected shift to 387 nm, attempts to purify the product
resulted in a hypsochromic shift to 372 nm, indicating partial
hydrolysis of the enamine. This result was confirmed by the
NMR spectrum which showed the presence of an aldehydic proton,
and by the mass spectrum which showed a prominent peak at m/e =
292 corresponding to nor-fluorocurarine, as well as a parent
peak at m/e = 345 corresponding to the pyrrolidine enamine 152.
It was hoped that, with the pyrrolidine enamine 152, reaction with methyl chloroformate under equilibrium conditions would
result in the introduction of a carbomethoxy group at C-16.
After hydrolysis of the resultant pyrrolidinium moiety, preaku•
ammicine aldehyde (153) would be formed. However, all attempts
to accomplish this transformation were unsuccessful. At this point it was obvious that the complicated functionality of nor-
fluorocurarine was not amenable to straightforward chemical manipulation, and it was decided to abandon this approach.
During the course of their structure elucidation of 131 akuammrcine (66), Edwards and Smith described the conversion of Wieland-Gumliclx aldoxime (154) to methyl 18-hydroxy-2g, 16a- cur-19-en-17-oate (156). Since the indirect oxidation of WGA to the corresponding methyl ester was one of the basic require• ments for the synthesis of the stemmadenine system, it was
decided to integrate this sequence into the overall scheme shown
in Figure 31. It was found that Wieland-Gumlich aldoxime could conveniently
be prepared by allowing WGA to stir overnight at room temperature
with a large excess of hydroxylamine hydrochloride. The crystal- 127
line product was obtained in pure form, m.p. 242-244° (d) (lit.
m.p. 245 dec), in a final yield of 80-90%. In addition, the IR,
UV, NMR and mass spectral data all were consistent with structure
154. 131
Following the procedure of Edwards and Smith, the Wieland-
Gumlich aldoxime was dehydrated with acetic anhydride in pyridine 131
to form the nitrile diacetate 155 which displayed the reported
absorptions at 2260, 1748 and 1660 cm-1 in the IR spectrum. The
UV spectrum C^max 285, 277, 248 nm) was consistent with the expect•
ed N-acylindoline chromophore, and the NMR spectrum displayed two
sharp methyl singlets at 2.01 and 2.106.
Treatment of the nitrile diacetate 155 with barium hydroxide
in refluxing 35% ethanol for 16 hours, followed by isolation of
the crude product and treatment, without purification, with
methanolic 5% hydrochloric acid for 5 hours yielded methyl 18-
hydroxy-2g,16a-cur-19-en-17-oate (156) which displayed a melting
point (153-154°) in agreement with that reported in the litera- 131 ture (152-154.5°). The IR spectrum was in good agreement ( 160 ) ( 161 )
Figure 31. The Synthesis of 16-epi-SteiTimadenine (161) from
WGA C130). - 79 -
with, that reported by Edwards and Smith./ haying absorption bands
1 at 3615, 3510 and 1730 cm" , and the UV spectrum (Amax 297, 244
nm) was typical of the expected indoline chromophore. The NMR
spectrum contained two 1 proton singlets at 3.40 and 4.226 which
disappeared on exchange with D20, a 3 proton singlet (-COOCH^)
at 3.686 , and a 1 proton triplet (=CH-CH2-OH) at 5.606. The
mass spectrum contained the expected molecular ion at m/e = 340.
With the desired hydroxy-ester 156 in hand, the next problem
was to hydrogenolyze the hydroxy group at C-18. The method of 132 133
Bernauer et al ' whxch has been previously described in
the conversion of the diol 136 to the alcohol 137 was attempted
on the hydroxy ester 156, resulting in a 65% yield of the desired
methyl 2g,16a-cur-19-en-17-oate (157). The IR spectrum contained
peaks at 3420, 2980, 1730 and 1605 cm~x, indicating that the
methyl ester moiety had remained intact. the UV spectrum (A
297, 245 nm) was again characteristic of the expected indoline
chromophore. the NMR spectrum contained a 3 proton doublet of
doublets at 1.586 (J = 2 arid 7 Hz) (=CH-CH_3) , a 3 proton singlet
at 3. 706 (COOCH3) , and a 1 proton quartet (J = 7Hz) at 5.486 (=CH-
CH3). As well, the-molecular ion and fragmentation pattern was 150
found to be identical to that reported in the literature.
As shown in Figure 32, retro-Diels-Alder opening of ring C to
form ion 162 simultaneously relieves the strain of the fused
pentacyclic system and provides for aromatization of the indoline
nucleus. Cleavage at the positions marked a, 8 and y lead to the
observed ions a, b, c and d CR=R'=H) at m/e =130, 194, 144 and
139. In addition, transfer of a hydride from C-14 to C-16, - 80 - during rupture of ring C forms ion 163, which, undergoes cleavage of the C-15, C-16 bond followed by rearrangement of the- double bonds to form the fully aromatic ion e at m/e = 2 51 (R=H).
a 3 y
Figure 32. The Proposed JSlass Spectral Fragmentation .Reactions
of the Methyl Cur-19-en-17-oate System.
131
Sxnce xt has been reported that the allylic hydroxyl
group of the hydroxy ester 156 could not be removed by treatment
with zinc in.acetic acid, the above result was most gratifying. - 81 -
However, when the alternative method of catalytic hydrogenoly- 134
sis via Wieland1s procedure was attempted, the mass spectrum
of the product indicated small, but significant, amounts of
methyl 28,16a-curan-17-oate (156a) . Although this side product
was not evident in the NMR spectrum, it could not be tolerated
due to the importance of mass spectral data in the analysis of
projected intermediates in the synthetic pathway, and it was
therefore decided to use only the bromination-hydrogenolysis
approach even for the preparation of unlabelled material.
At this point, two crucial problems remained to be solved
in order to complete the synthesis. First, it would be essential
to determine the conditions necessary for the introduction of the
hydroxymethylene group at C-16. Second, a procedure would have
to be devised for the conversion of the indoline moiety to the
corresponding indole system with cleavage of ring C at the C-3,
C-7 bond. As shown in Figure 33, two alternative pathways can
be visualized for the completion of the synthetic sequence from
methyl 28,16a-cur-19-en-17-oate (157), depending on the order in which the above two operations are carried out. Rather than
concentrate solely'on pathway A or pathway B of Figure 33, it
seemed to be expedient to initially develop methods for the
conversion of compound 157 to the hydroxy ester 160, and then to
solve the problem of the introduction of the 2,16 double bond to
form the unsaturated ester 66 (akuammicine). Since the ester. 157 was relatively readily available, this would mean that the two most difficult synthetic problems could be investigated with a minimum expenditure of effort and material. - 82 -
(6)
Figure 33. Two Possible Routes to Stemmadenine (6) from Methyl
23 ,16ct-cur-19-en-17-oate (.157) . - 83 -
( 165 )
Figure 34-. The Possible Condensation of Formaldehyde with Methyl
28,16a-cur-19-en-17-oate (157).
In dealing with, the first problem, i.e. introduction of the hydroxymethylene group at C-16, it was felt that treatment of the ester 157 directly with formaldehyde in the presence of base could easily result in initial condensation of formaldehyde with the indoline nitrogen as shown in Figure 34. Subsequent conden• sation of the anion of the carbinolamine 164 thus formed with the ester carbonyl would result in the undesired tetrahydro oxazinone 165. Thus, it was decided to preclude such a reaction pathway by protecting the indoline nitrogen. For this purpose, the N(a) formyl derivative 158 was chosen for the following reasons. First, it appeared easily accessible by condensation of 157 with methyl formate. Second, it would condense selec• tively with the secondary indoline nitrogen, since attack at the tertiary nitrogen would be much more readily reversible. Third, it could be easily removed by refluxing with sodium hydride in - 84 -
151 tetrahydrofuran, a procedure which would not cause any undesirable reactions elsewhere in the molecule.
To this end, the ester 157 was treated with methyl formate and sodium hydride in refluxing benzene, and the desired product methyl 1-formyl-2(3,16a-cur-19-en-17-oate (158) was obtained.
The IR spectrum showed an absence of NH absorption, but did show the presence of two carbonyl groups having absorptions at
1730 and 1668 cm"1. The UV spectrum U 287, 278, 250 nm) was characteristic of the desired N-acylindoline chromophore (cf. compound 155). The NMR spectrum contained a 1 proton singlet at 8.716 (N-CHO) , a 1 proton quartet (J = 7 Hz) at 5.386 (=CH-
CH3), a 3 proton singlet at 3.686 (COOCH^), and a 3 proton
doublet of doublets (J = 2 and 7 Hz) at 1. 536 (=CH-CH_3) . The mass spectrum contained a molecular ion at m/e = 352, and a fragmentation pattern consistent with that shown in Figure 32.
Specifically, ions a, b, c, d and e (R = CHO, R' = H) could be seen at m/e = 158, 194, 172, 139 and 279, respectively.
Initially, it had been hoped that the reaction conditions employed for introducing the N-formyl protecting group might simultaneously result in the introduction of an aldehyde func• tional group on C-16 via the condensation of the enolate anion with methyl formate. That this did not occur, and in fact could not be induced to occur even by means of employing higher temperatures and longer reaction times, seemed to indicate that an investigation of the conditions necessary to form a reactive anion at C-16 was in order.
In order to investigate anion formation at C-16, the N-formyl - 85 - ester 158 was allowed to react with a.variety of base/solvent combinations and then quenched with V^O. The compounds re• covered were submitted for mass spectral analysis. Efficient incorporation of deuterium at C-16, indicating anion formation, could then be determined by a decrease in peak heights at m/e =
352, 194 and 139 (ions M+, b and d, R = CHO, R« = H) with a concomitant increase in peak heights at m/e = 353, 195 and 140
(R = CHO, R' = D). The results of this investigation are summarized in Table 7.
Table 7. Summary of Results of the Investigation into Anion
Formation at C-16 of the N-Formyl Ester 158.
Experiment Tempera• Time % D Incor- ^ Number Solvent Base ture Hrs. poration
153 1 THF J Li NU-Pr)^ -78° . 45 0 2 THF NaH reflux 2.5 0
3 DMSO : NaH 22° .40 >95
4 HMPA NaH 22° .40 >95
5 THF t-BuOK 22° .40 40
6 DMSO t-BuOK 22° .40 * *
7 HMPA t-BuOK 22° .40 **
Approximate, based .on relative peak heights in mass spectra ** Severe decomposition of starting material occurred.
It can be seen from Table 7 that sodium hydride in either dimethyl sulfoxide or hexamethyl phosphoramide (experiments 3 and
4) offered the most hope of success. Of the two solvents, dimethyl sulfoxide was chosen because it appeared to be more easily removed ^ 86 ^ from the reaction mixture, and resulted in better recovery o% material. Accordingly, when the N-formyl ester 158 was allowed to react with dry formaldehyde with, sodium hydride in dimethyl sulfoxide at room temperature, a product was formed which appeared to be less polar than the starting material by thin layer chromatography. The TR spectrum of the product showed no
N-H bands, and only one carbonyl band at 1730 cm 1, while the
UV spectrum CA 299, 246 nml indicated an indoline chromophore. *• max *•
The NMR spectrum (Figure 35) shows a 1 proton quartet (J = 7 Hz)
at 5.466 C=CH-CH3), a typical AB pattern (J =10 Hz) at 5.18
and 4.716 (N-CH2-0) , a 1 proton singlet at 4.946 (N~CH- CCR3) 2) , a 3 proton singlet at 3.706 (COOCH^), and a 3 proton doublet of doublets (J = 2 and 7 Hz) at 1.576 C^CH-CH-). The high reso-
0"22H26°3N2 = 366.1943) with a major fragment at .m/e 366.1815
(loss of CH20). These data are all in accord with the unexpec• ted, but nevertheless useful carbomethoxy tetrahydrooxazine structure 159.
In order to provide further evidence for the proposed structure 159, it was decided to synthesize the analogous descarbomethoxy compound 166 via an independent route.
(137 ) ( 166 )
Figure 36'. The Reversible Formation of the Model Tetrahydro-
oxazine (166). Figure 35. The Nuclear Magnetic Resonance Spectrum of the Carbomethoxy Tetrahydrooxazine 15 88 r
As shown in Figure. 36f 2g, 16a^cur^l9-en-17-ol (1371 previously
prepared by hydrogenolysis of Wieland—Gumlich. diol (136)
(Figure 25) was allowed to react with paraformaldehyde in
methanol in the presence of anhydrous sodium sulfate at room
temperature. This procedure, analogous to that previously 153—155
described in the literature for geissoschizoline (167)
gave in good yield the desired tetrahydrooxazine 166. The IR
(167) spectrum showed the expected disappearance of the 0-H and N-H bands at 3150 and 3395 cm 1, while the uv spectrum (A 297, *• max
249 nm) indicated the expected indoline chromophore. The NMR
spectrum (Figure 37) contains a 1 proton quartet (J = 7 Hz) at
5.416 (=CH-CH3), a typical AB pattern (J = 11 Hz) at 5.23 and
4.676 (N-CH2-0), and a 3 proton doublet (J = 7 Hz) at 1.506
(=CH-CH_3) . The high resolution mass spectrum indicates a mole• cular ion at m/e = 308.1932 (C2oH24ON2 re(3uire 308.1887) with
a major fragment occurring at m/e = 278.1782 (loss of CH2Q).
The most striking similarity in the NMR spectra shown in
Figures 35 and 37 is the presence of the two downfield doublets
caused by the protons on the isolated carbon of the tetrahydro• oxazine system. Because of the rigidity of the hexacyclic system, the a and g protons are held in different magnetic environments and hence have different chemical shifts as well as character- Figure 37. The Nuclear Magnetic Resonance Spectrum of the Model Tetrahydrooxazine 166. - 90 -
istic splitting patterns which, reveal spin coupling between
them. The presence of the 1 proton singlet for the 2(3 proton
at such low field in Figure 36 can be explained by the
differences in configuration at C-16. In the case of the
carbomethoxy tetrahydrooxazine 159, molecular models reveal that
the ether linkage of the tetrahydrooxazine system is projected
into the a plane of the molecule, thus projecting the lone pair
of electrons on the indoline nitrogen into the |3 configuration,
with no facile inversion possible. This situation in turn places
the 2 6 proton into very close proximity to the lone pair of
electrons, resulting in the observed downfield shift. On the
other hand, in the model tetrahydrooxazine 166, exactly the
opposite is the case, with the ether linkage of the tetrahydro•
oxazine projected into the 8 plane of the molecule, and no down-
field shift is observed.
That the carbomethoxy tetrahydrooxazine does in fact possess
the configuration shown at C-16 can be further established by 131
NMR evidence. It has been shown by Edwards and Smith that
the hydroxy ester 156 possesses the 3-carbomethoxy configuration
at C-16 and is not'epimerizable. This is understandable, since
with ring C in a chair conformation, the 3 carbomethoxy group would be in a favorable equatorial orientation. If epimerization were to occur at C-16, this would force the carbomethoxy group
into an extremely unfavourable axial position. With the stereo•
chemistry of the hydroxy ester 156 established, Table 8 correlates
the chemical shift data of that compound with the data for the ester 157, the N-formyl ester 158, and the carbomethoxy tetra• hydrooxazine 159. Table 8. A Summary of Pertinent Nuclear Magnetic Resonance Chemical Shifts for
Compounds 156-159
Chemical Shifts 6 (ppm) - 92 -
The data shown in Table 8 offers convincing evidence that,
in proceeding from the hydroxy ester 156 to the carbomethoxy
tetrahydrooxazine 159, the configuration of the carbomethoxy
group remains unchanged, i.e. 8. In addition, studies of the
molecular model of the N-formyl ester 158 clearly indicate that
approach of an electrophile from the a face would be far more
favorable than from the quite hindered B face.
With the structure of the carbomethoxy tetrahydrooxazine 159
established, it would be interesting to speculate on the mechanism
of its formation. As shown in Figure 38, condensation of the
enolate anion of the N-formyl ester 158 with formaldehyde would
yield an intermediate 168 which could then cyclize reversibly
to the intermediate 169. Although the equilibrium involving
ring opening of 169 to the formate ester 17 0 would not necessarily
favor the formate ester, the anion formed at the indoline nitrogen
could easily condense with another molecule of formaldehyde to
form 171 followed by an irreversible cyclization involving dis•
placement of formate to yield the observed carbomethoxy tetrahydro•
oxazine 159.
Once the carbomethoxy tetrahydrooxazine 159 had been obtained,
the next problem was to hydrolyze the tetrahydrooxazine moiety without affecting the methyl ester. To this end, the model
compound 166 was used as a source of material for study. When
the model tetrahydrooxazine was treated with 10% hydrochloric
acid in methanol at room temperature for 5 hours, quantitative
hydrolysis of the tetrahydrooxazine occurred. This could easily
be shown by comparison of the spectral data of the product of those - 93 -
already available for the alcohol 137.
(171)
Figure 38. The Proposed Mechanism for the Formation of the
Carbomethoxy Tetrahydrooxazine 159.
Similarly, when the carbomethoxy tetrahydrooxazine 159 was treated with refluxing 10% methanolic hydrochloric acid, the hydroxy ester 160 was obtained in good yield. The IR spectrum showed absorption bands at 3350, 2950 and 1725 cm 1 and the
UV spectrum (^max 296, 244 nm) was consistent with a dihydro• indole chromophore. The NMR spectrum contained a 3 proton
quartet (J = 7 Hz) at 5.546 (=CH-CH3), a 1 proton singlet
at 3.736 C00CH3), and a 3 proton doublet of doublets (J = 2 and 7 Hz) at 1.646. That the singlet for the 28 proton has only shifted upfield 0.2 8 ppm may be considered to be due - 94 -
to hydrogen bonding between the proton on the indoline nitrogen
and the oxygen of the alcoholic function. This would have the
effect of maintaining the close proximity between the 23 proton
and the 3 lone pair of electrons on the indoline nitrogen.
Finally, the mass spectrum exhibited the expected molecular ion
at m/e = 354.
Having accomplished one of the twQ remaining synthetic
problems, it remained to devise a method for the oxidation of
the indoline system to either an indolenine or an a-methylene-
indoline system, which could then be easily converted to the
desired indole system. As previously stated, it was decided
to concentrate the initial investigation on the problem of
converting the ester 157 to the corresponding a,3 unsaturated
ester 66 (akuammicine) . It has been shown15*' that lead tetra•
acetate is capable of oxidizing an indoline system to the
corresponding indolenine without serious side reactions involving
the basic nitrogen (N(b)) of the alkaloidal system. Therefore,
lead tetraacetate was the reagent of choice for the investigation
at hand.
The initial investigation consisted of adding a benzene
solution of lead tetraacetate dropwise to a benzene solution of
the ester 157 with frequent monitoring of the progress of the
reaction. This was accomplished in two ways. First, an aliquot would be taken from the reaction mixture, the solvent evaporated
to dryness and a UV spectrum taken of the residue. Second, a small aliquot would be placed on a silica gel thin layer chroma• tography plate and eluted with 20% methanol in benzene. The - 95 - presence of akuammicine in the reaction mixture could then
easily be determined by the appearance of the characteristic
absorption at 325 nm in the UV spectrum, and the charac•
teristic blue color which appears when an a-methylene-indoline 140
is sprayed with eerie sulfate solution.
In this way, it could be seen that the reaction was in fact proceeding as desired, and that the following factors were
important. It was found that optimum results could be obtained when the reaction was carried out in a dilute benzene solution
containing 1-2 equivalents of acetic acid. Further, it was found
that 2 equivalents of lead tetraacetate were required for optimum yields, and must be added very slowly as a benzene solution.
Chromatographic purification of the resultant product mixture and crystallization provided akuammicine (66). The IR, UV and mass spectral data were identical to that already described in 150 157
the literature. ' Further, it can be seen that the NMR
spectrum (Figure 40) is remarkably similar to that of nor-fluoro-
curarine (Figure 28) with the exception of the 3 proton singlet
at 3.82 (COOCH^) and the lack of an aldehydic proton in the former. 158 Since strychnine (29) has been totally synthesized, this then represents the first total synthesis of akuammicine.
With both major synthetic problems solved, the completion of the synthesis of the stemmadenine system involved comparing
the relative merits of pathway A or pathway B shown in Figure 33.
Pathway B was the first to be attempted.
Treatment of akuammicine with an excess of sodium borohydride 121 in refluxing acetic acid led to the desired indole ester 141a NaH. ->
COOCH3
(Ula ) ( U1 b )
and a small amount of the epimer 141b. That the major product would be the a carbomethoxy epimer would be expected from the 131 results of Edwards and Smith who demonstrated that reduction of the 2,16 double bond of akuammicine by zinc and methanolic
sulfuric acid results in methyl 23,16B-cur-19-en-17-oate (172).
As shown in Figure 39, they proposed that the first step would be protonation of C-16 of akuammicine to form the iminium ion 173
COOCH3 " COOCH3 " COOCH
( 66) (1 73) M72)
Figure 39. The Edwards and Smith Mechanism for the Zinc and
Sulfuric Acid Reduction of Akuammicine (66).
Protonation from the 3 face would place the carbomethoxy group in the less strained equatorial position, with ring C having a boat conformation. Similarly, the iminium ion 17 3 in refluxing acetic acid would be expected to undergo opening of ring C as shown in Figure 41. In the presence of sodium boro• hydride, the resultant N(b) iminium ion 173 would be reduced to the corresponding amine 141. The IR spectrum of the a-carbo— Hi
Figure 40. The Nuclear Magnetic Resonance Spectrum of Akuammicine (66). 98 -
methoxy compound had absorption bands at .3450, 2930 and 1725 cm*"1 and the UV spectrum (A 290, 283, 225 nm) was consistent with the desired indole chromophore.
( 173 ) (174) ( U1a ) Figure 41. The Proposed Mechanism for the Formation of Indole
Ester 141a.
The NMR spectrum (Figure 42) contained a 1 proton singlet at
9.086 (indole N-H), a 1 proton quartet (J - 7 Hz) at 5.606
(=CH-CH3), a 1 proton singlet at 4.306 (indole-CH-COOCH3), a 3
proton singlet at 3.886 (COOCH_3) , and a 3 proton doublet (J =
7 Hz) at 1.776 . (=CH-CH3). The mass spectrum showed a molecular ion at m/e = 324, with a base peak at m/e = 123.
When the a-carbomethoxy compound was refluxed in benzene with sodium hydride, clean conversion to the 8-carbomethoxy epimer occurred. the 8-carbomethoxy compound 141b had absorption bands in the IR spectrum at 3460, 2940 and 1725 cm-1. The UV spectrum was identical to that obtained for the epimer 141a, as was the mass spectrum. However, the NMR spectrum (Figure 43) was markedly different, containing a 1 proton singlet at 8.86 (indole
N-H), a 1 proton quartet (J =7 Hz) at 4.466 (=CH-CH3), a 1 proton
doublet (J = 4 Hz) at 4.206 (indole-CH-COOCH3) a 3 proton singlet
at 3.766 COOCH3) and a 3 proton doublet (J = 7 Hz) at 1.586
(=CH-CH3) .
At this point, two factors led to the abandonment of 7 6 5 4 3 2 1
<5 Cppm)
The Nuclear Magnetic Resonance Spectrum of the Indole Ester 141a. 9 8 7 65 4 3 2 1 0
6 (ppm)
Figure 43. The Nuclear Magnetic Resonance Spectrum of the Indole Ester 141b. - 101 -
pathway B as shown in Figure 33. First, the reductiye ring
opening gave, at best, a 30-40% yield of the desired indole
ester 141. This, coupled with the low yield (35%) of
akuammicine from the ester 157 made pathway B unfavorable from
the standpoint of final yield. Second, all attempts to intro•
duce the required hydroxymethylene group at C-16 met with
failure. Thus, the only possibility of success through pathway
B lay in protecting the indole nitrogen prior to formation of
the anion at C-16, making the final yield (if any) of the
stemmadenine system prohibitively low. Thus, it was decided to
attempt to complete the synthesis of the stemmadenine system via pathway A (Figure 33).
Accordingly, the hydroxy ester 160 was treated with lead
tetraacetate under the same conditions used to synthesize akuammicine. The reaction was monitored via thin layer chroma• tography, and the immediate appearance of a less polar product was observed. After 2 equivalents of lead tetraacetate were added, the reaction was terminated by rapidly filtering the crude mixture through a column of alumina (12% water) with methylene chloride as eluent. Since the reaction product was expected to 79 be the preakuammicine system (2) which is known to be unstable, no attempt was made to isolate or purify the product obtained.
The crude oxidation product was dissolved directly in 50% methanolic acetic acid and an excess of sodium borohydride was added to reduce the anticipated iminium ion formed during the ring opening of the preakuammicine system (see Figures 33 and 42).
Purification of the product thus formed was achieved by silica - 102 T.
gel preparatiye layer chromatography plates. The pure product
had absorption bands in the IR spectrum at 3580, 3420, 2920 and -
1725 cm""1. The UV spectrum CA 291, 284, 225 nm) was consis- r max ' '
tent with the desired indole chromophore. The NMR spectrum
contained signals for two exchangeable protons at 10.14 and 2.666
(indole N-H and CH2-OH, respectively), a 1 proton quartet (J =
7 Hz) at 5.426 C=CH-CH3), a 2 proton singlet at 4.366 C-CHj-OH),
a 3 proton singlet at 3.886 (COOCH^), and a 3 proton doublet
CJ = 7 Hz) at 1.526 C=CH-CH3) . The mass spectrum exhibited the
expected molecular ion at m/e = 354, with fragments at M+-17, 18
and 30 corresponding to loss of hydroxide, water and formaldehyde in accord with the fragmentation pattern already described for 117 stemmadenine. That the compound obtained was not in fact stemmadenine can be easily seen from the NMR evidence in Table 9
This comparison of the assigned chemical shift data published for 118a stemmadenine with that described above- clearly indicates that the natural stereochemistry about C-16 in the stemmadenine system has not been obtained.
Table 9. Comparison of NMR Data oh Natural and Synthetic
Stemmadenine Systems.
Compound Indole N-H =CH-CH3 CH2~OH COOCH_3 =CH-CH
Natural 9.4 5.4 4.38 3.79 1.7
Synthetic 10.14 5.4 4.36 3.88 1.5 - 10.3 -
In order to prove that the synthetic material was indeed
16-epi-stemmadenine (161), it was decided to reduce both the 159 natural and synthetic material to the corresponding diol as shown in Figure 44. The two products thus obtained should then be identical in every respect.
( 6a ) (175) ( 161 )
Figure 44. The Reduction of Natural and Synthetic Stemmadenine
Systems to the Diol 175.
Accordingly, both the natural and synthetic stemmadenine compounds were dissolved in tetrahydrofuran and treated with an excess of sodium bis(methoxy^ethylenoxy) aluminum hydride
(commercially available benzene solution). The reaction was rapid and led to essentially only one component in both cases.
After purification via preparative layer chromatography on silica gel, the two products were compared and found to be identical, possessing superimposable IR (Figure 45) and NMR (Figure 46) spectra.
Since the synthetic 16-epi-stemmadenine has known stereo• chemistry about C-16 established earlier on in the synthesis, it now becomes possible to assign the natural stereochemistry about
C-16 in stemmadenine. Thus, the stereochemistry shown in Figure
47 for stemmadenine (6a) is probably correct. 1
Figure 45. The Infrared Spectra of Synthetic CA) and Authentic (B) Diol 175 in chloroform. - 105 -
igure 46. The Nuclear Magnetic Resonance Spectrum of Synthetic
(A), and Authentic (B) Diol 175. ^ 1Q6 «-
Although, the synthesis of the natural stereochemistry
stemmadenine system will not De described in this thesis, it would be interesting to speculate as to how it might be obtained.
As shown in Figure 47, a likely starting point would be the saturated aldehyde 133, the synthesis of which has already been 137 described. Proceeding in a manner strictly analogous to that outlined in Figure 31, protection of the indoline nitrogen as its formamide 176 followed by reaction with methyl chloroformate in base might be expected to yield the N-formyl aldehyde ester 177 which could be deformylated with sodium hydride in tetrahydro• furan to yield the aldehyde ester 178. Oxidation with lead tetraacetate followed by reductive ring opening would then yield the desired stemmadenine (6a).
< 6a ) (178)
Figure 47. The Proposed £oute for the Synthesis of Stemmadenine
(6a) ^ 107 ^
Part c
Although, the original goal of the synthesis of stemmadenine having the natural configuration about C-16 was not met, it was nevertheless of interest to determine the usefulness of synthetic
16-epi-stemmadenine as a biosynthetic precursor. To that end, a sample of 16-epi-stemmadenine (161) was labelled with tritium in the aromatic ring by exchange with tritiated trifluoroacetic 121 acid, and the radioactive compound administered to root cuttings of six year old Aspidosperma pyricollum plants. At the same time, a similarly labelled radioactive sample of authentic stemmadenine"''"'"^ was also administered to a separate portion of root sections. In this way, the efficiency of incorporation of the C-16 epimer could be directly compared with the natural isomer.
The results of this study are shown in Tables 10 and 11.
Table 10. The Stemmadenine Systems Administered to A. pyricollum
Experiment Compound Fed Weight Wet Plant Activity Number Fed mg Weight g Fed dpm
3 49 3.29x10 1 stemmadenine-(Ar- H) 1.3 29 2.85xl010 2 16-epi-stemmadenine- 1.4
(Ar-3H - 108 T
Table 11, Incorporation Results Associated with. Table 10.
Experiment Compound Isolated Specific Activity % Incorporation Number Cwt, mg) (dpm/mmol)
1 apparicine (37) 3.52xl06- 0.109
2 apparicine (15) , <0.0001
The data shown in Tables 10 and 11 clearly indicate that while authentic stemmadenine, as expected from the results shown in
Figure 14 is incorporated into apparicine to a significant extent, the C-16 epimer is not. Although results in biosynthetic experi• ments must always be interpreted with great care, it may be the case that either the conversion of the stemmadenine system to the indolenine 117 or the subsequent loss of formaldehyde as shown in
Figure 15 (pathway A) or both is an enzymatically controlled process for which the natural stereochemistry about C-16 is vital.
This of course assumes that the Potier postulate is correct, something which has not yet been proven. In any case, the results shown in Tables 10 and 11 indicate, but do not prove', that the stereochemistry about C-16 of stemmadenine is an important factor in its role as precursor in the plant system. Further information on that point await the successful completion of the synthesis of stemmadenine, a project currently underway in our laboratory. - 109 -
EXPERIMENTAL
Melting points were determined on a Kofler block and are uncorrected. The ultraviolet (UV) spectra were recorded in methanol using a Cary 15 recording spectrometer. The infrared
(IR) spectra were recorded with a Perkin-Elmer Model 457 spectrometer in chloroform solution with a cell path of 0.2 mm , using a matched reference cell filled with chloroform (unless otherwise noted). Calibration was achieved vising the -16-01 • cm ~ absorption band of polystyrene. Nuclear magnetic resonance spectra (NMR) were obtained with deuteriochloroform solutions
(unless otherwise indicated) at 100 MHz on a Varian HA-100 or a Varian XL-100 nuclear magnetic resonance spectrometer. All
NMR spectra obtained via the Fourier Transform technique (FT) will be so noted and were obtained with the Varian XL-100 instrument. Chemical shifts are given in 6(ppm) with reference to tetramethylsilane as the internal standard. The multipli• city, integrated peak areas, and proton assignments are given in parentheses. Mass spectra were determined on an AEI-MS-902 or an Atlas CH-4B mass spectrometer,with high resolution mass spectra determined by the former. Woelm neutral alumina and
EM Reagents GF254 silica gel \i?ere used for thin and preparative - 110
layer chromatography. Woelm neutral alumina (activity III)
was used for column chromatography in section A. In sections
B and C, Woelm neutral alumina (12% water) was used for column
chromatography. In section B, two solvent systems were used for
the development of thin and preparative layer chromatography
plates. System A consisted of 20% triethylamine in chloroform,
and system B consisted of 20% methanol in benzene.
Radioactivity was measured with Nuclear-Chicago Mark 1 or
Mark XI liquid scintillation counters in counts per minute (cpm).
The radioactivity of a sample in disintegrations per minute (dpm)
was subsequently determined using the external standard techni•
que16^'1*''1' using a built-in Barium-133 source of gamma radiation.
A typical supply of the liquid scintillator solution which was
used was made up of the following components: toluene (1 liter),
2,5-diphenyloxazole (4 g) and 1,4-bisI2-(5-phenyloxazoly)Jbenzene
(.0.05 g) . In practice a sample was dissolved in benzene (1 ml)
or in methanol (1 ml) in a counting vial. The volume was then made up to 15 ml with the above scintillator solution. For each
sample counted the background activity was determined for the
counting vial to be used by filling the vial with the appropriate
solvent and scintillator solution and counting-the background activity in cpm. The vial was then emptied, refilled with the sample to be counted and the solvent/scintillator solution mixture and its activity determined. The difference in actiyity (cpm) be• tween the sample and the previously determined background was then used for subsequent calculations. Each vial was counted for a time period long enough for the total counts for the sample, less the - Ill -
total counts for the background, to exceed one thousand counts.
The A. pyricollum and A. australe plants used in this study were grown in the Horticulture Department greenhouse, the
University of British Columbia. - 112 -
Section A
Degradation of apparicine (81)
A solution of apparicine C81) CIO.1 mg) in acetic acid CIO ml)
was treated with ozone gas until a blue color appeared (15 min)^
Water (10 ml) was then added, and the mixture allowed to stand
for 30 min. The mixture was then steam distilled and the dis•
tillate (30 ml) collected. The distillate was then treated with
a saturated solution of dimedone (92) in water (10 ml) and allowed
v to stand at room temperature for 16 hr. Glacial acetic acid
(1 ml) was then added and the mixture refluxed for 6 hr. The
solvent was then evaporated under reduced pressure and the residue
purified via silica gel preparative layer chromatography plates
developed with a solution of 2 0% ethyl acetate in chloroform.
Formaldehyde bisdimedone (93) was obtained (0.5 mg, 5%) as white
crystals which could be recrystallized from ethanol, m.p. 190-192° 112
(lit. m.p. 191-191.5°), and which was identical with an authentic
sample of formaldehyde bisdimedone (93) by thin layer chromatography
and mixed melting point comparison. Cyclized acetaldehyde bisdime•
done (95) was obtained (0.8 mg, 8%) as white crystals which could be purified by recrystallization from ethanol, m.p. 175-176° 112
(lit. m.p. 176-177°), and which was identical with an authentic
sample by thin layer chromatography and mixed melting point compari•
son.
The administration of labelled secodine (76) to A. pyricollum
The particular labelled secodine (76) to be administered was 113 - prepared from 16,17-dihydrosecodin-17-ol C77) as previously 93 95 described. ' The pale yellow gum was then dissolved in ethanol CO.5 ml} and to this solution was added 0.1 N acetic acid CO.5 ml) and distilled water C1.0 ml). This solution was administered to root cuttings (usually 30-6 0 g) of 2-3 year old
A. pyricollum plants in a large test tube for 5 days. During that time, the root cuttings were kept moist by periodic additions of distilled water.
The extraction of apparicine (81) from A. pyricollum.
The root cuttings to which labelled secodine (76) had been administered were mascerated with methanol in a Waring blender, filtered, and washed with methanol. The solvent was evaporated under reduced pressure and the residue taken up in 2 N hydrochloric acid (150 ml). This mixture was first extracted with benzene
( 3 x 75 ml), then made basic with 15 N ammonium hydroxide and extracted with chloroform (3 x 10 0 ml). The combined chloroform extracts were dried over anhydrous sodium sulfate, filtered, and evaporated to dryness under reduced pressure.
The residue thus obtained was purified via preparative layer chromatography on silica gel plates developed with 30% ethyl acetate in ethanol, affording pure apparicine (81), which could be recrystallized from acetone. The apparicine thus obtained was degraded by ozonolysis as described above. Further data pertaining to experiment 5, Tables 1 and 2 are as follows. Weight of seco• dine C76) fed: 2.1 mg; wet plant weight: 64 g; weight apparicine isolated: 33.2 mg; weight formaldehyde bisdimedone isolated - 114 - after degradation: 2.1 mg; weight acetaldehyde bisdimedone derivative isolated after degradation of apparicine: 4 mg.
Administration of labelled compounds to A. australe
The compound to be fed was dissolved in ethanol (5-10 drops).
To this solution was added 0.1 N acetic acid (5 drops) and distilled water (0.5 ml). This solution was administered hydro- ponically to whole 1-2 year old A. australe plants whose root systems were contained in test tubes. Absorption of the solution by the plants was observed in 2-3 hr, after which the roots were kept moist with additional quantities of distilled water for the 5 day feeding period.
Extraction of alkaloids from A. australe
The A. australe plants were mascerated in a Waring blender with methanol, filtered, and remascerated until the filtrate was colorless. The solvent was evaporated under reduced pressure and the residue taken up in 2 N hydrochloric acid (150 ml).
This mixture was first extracted with benzene (3 x 75 ml), then made basic with 15 N ammonium hydroxide and extracted with chloro• form (3 x 100 ml). The combined chloroform extracts were dried over anhydrous sodium sulfate, filtered and evaporated to dryness under reduced pressure. The residue was diluted with a small amount of inactive olivacine (88) (4-13 mg) and chromatographed on a preparative layer silica gel plate developed with a solution of 30% ethanol in ethyl acetate, yielding pure olivacine, which could be further purified by recrystallization in a mixture of T~ 115 -
chloroform and methanol. The remainder of the preparative layer
plate was extracted with methanol and the solvent remoyed under
reduced pressure. The resulting residue was then chromatograph•
ed on a preparative layer alumina plate developed with a solution
of 10% methanol in benzene, which afforded pure apparicine (81)
which could then be recrystallized from acetone. The remainder
of the plate was extracted as before, and the residue diluted
with inactive uleine C83) and guatambuine C90) C8-14 mg). This mixture was then chromatographed on a preparative layer silica
gel plate developed with 50% acetic acid in ethyl acetate, yield•
ing the acetate salts of uleine (83) and guatambuine (90).
These salts were made basic with 15 N ammonium hydroxide and
extracted with methylene chloride. The methylene chloride
extracts were then dried over anhydrous sodium sulfate, filtered,
and evaporated to dryness under reduced pressure, yielding
uleine (83) and guatambuine (90) as the free bases. Uleine was recrystallized from methanol and guatambuine was recrystallized
from a mixture of methanol and chloroform. The data obtained
from the administration of various labelled compounds to A. australe can be found in Tables 4 and 5. r- 116 Section B Degradation of strychnine C2 9) to 23,16q-cur-19-en-17-ol (137) The degradation of strychnine to Wieland-Gumlich aldehyde has been exhaustively studied and described in great detail by 129 Schmid and Karrer. As well, the subsequent transformation of Wieland-Gumlich aldehyde to the alcohol 137 has been described 132 133 both via the allylic bromide 138, ' and via the direct hydro- 134 genolysis of the diol 126. Details of quantities used and yields obtained in this work are given below. Wherever modifi• cations in the experimental procedure were employed or additional data were obtained for characterization of the intermediates in the degradation, this is included. As well, this practice will be followed wherever else a particular reaction or sequence of reactions has been previously described. Isonitrosostrychnine hydrochloride (.131) A suspension of strychnine (134 g) in dry ethanol (800 ml) was treated with isoamyl nitrite (211 g) and sodium ethoxide (37.2 g sodium in 1 1 ethanol). . Isolation of the product by crystallization from aqueous hydrochloric acid yielded the desired hydrochloride 131 )129 g, 80%). Wieland-Gumlich aldehyde (1301 Isonitrosostrychnine hydrochloride (11 g) was treated with thionyl chloride (20 ml) followed by quenching in ice (100 g). - 117 - Filtration of the aqueous slurry followed by treatment with steam at pH 3-3.5 (methyl orange indicator) yielded the desired Wieland-Gumlich aldehyde (130) (5 g, 58%) as colorless crystals which could be recrystallized from chloroform or benzene m.p. 213-215° (lit.129 m.p. 214-216). NMR signals: 6.6-7.2 (2 multiplets, 4H, aromatic C-H), 5.82 (broad singlet, IH, C-19 H), 5.0 (singlet, IH, C-17 H) . 2 8,16a-cur-19-en-17,18-diol (136) Wieland-Gumlich aldehyde (5 g) was treated with sodium boro- 133 hydride (1.25 g) in methanol (200 ml) as outlined by Bernauer to yield the desired diol 136 as colorless crystals (4.1 g, 82%) 12 which could be recrystallized from methanol m.p. 250-251° (lit. m.p. 251°). NMR signals (CD^OD): 6.6-7.2 (2 multiplets, 4H, aro• matic C-H), 5.72 (triplet, J = 7 Hz, IH, C-19 H), 4.13 (doublet + of doublets, J = 7 and 2 Hz, 2H, C-18 H_2) . Mass spectrum: M , m/e = 312; main peaks = 294, 281, 182, 144 (base peak). High resolution mass spectrum: calc. for CinH„.N„0o: 312.1837. Found: 312.1784. 2g-16a-17-acetoxy-18-bromo-cur-19-en (138) The diol 136 (2.3 g) was dissolved in glacial acetic acid (300 ml), and treated with a saturated (at 0°) solution of hydrogen bromide in acetic acid (6 ml) as outlined by Bernauer. The allylic bromide thus obtained was used in the next reaction without isolation or purification. ^ 118 - 2g,16q-17-acetoxy~cur-19-en (139! The crude allylic bromide obtained in the preceding reaction was dissolved in acetic acid C200 ml) and treated with zinc powder (15 g) according to the method described by Bernauer. The crude product thus obtained was purified by elution with benzene through a column of alumina, yielding a colorless oil (1.5 g, 63%). NMR signals: 6.6-7.2 (2 multiplets, 4H, aromatic C-H), 5.58 (quartet, J = 7 Hz, IH, C-19 H), 2.04 (singlet, 3H, OOCCH3), 1.59 (doublet of doublets, J =7 and 2 Hz, 3H, C-18 H^). Mass spectrum: M+, m/e = 338; main peaks: 279, 144 (base peak). High resolution mass spectrum: calc. for C2iH26N2°2: 338.1993. Found: 338.1978. 2g,16a-cur-19-en-17-ol (137) The acetate 139 (1.5 g) obtained in the preceding reaction was treated with IN methanolic potassium hydroxide (75 ml) at room temperature for 30 min. The solvent was evaporated under reduced pressure, and the residue treated with water (50 ml) and methylene chloride (50 ml). Extraction of this mixture with methylene chloride (3 x 50 ml) followed by drying of the combined extracts over anhydrous sodium sulfate, filtration, and evapora• tion of the solvent under reduced pressure yielded the crude alcohol. Purification was effected by elution through a column of alumina with benzene having an ether gradient, resulting in isolation of the pure alcohol 137 as colorless crystals (1.0 g, 77%) which could be further purified by recrystallization from benzene m.p. 168-171° (lit.134 m.p. 170-174°). NMR signals: 6.5-7.2 (2 multiplets, 4H., aromatic C-H.)., 5.47 (quartet, J - 7 Hz, IK, C-18 tt), 4.70 Cbroad singlet, IH, OH), 1.58 (doublet of . doublets, J = 7 and 2 Hz, 3H, C-19 H^). Mass spectrum: M+, m/e = 296, main peaks: m/e = 294, 279, 266, 166 (base peak). High resolution mass spectrum: calc. for C]_gH24N20: 296.1887. Found: 296.1883. Catalytic hydrogenolysis of 23,16a-cur-19-en-17,18-diol (136) The previously obtained diol 136 (3.5 g) was dissolved in a mixture of water (175 ml), acetic acid (105 ml), and concentrated hydrochloric acid (1.75 ml). Palladium on charcoal (10%, 543 mg) was added, and the mixture hydrogenolyzed as outlined by Wieland, yielding the desired alcohol 137 (3.0 g, 91%) which could be recrystallized directly from the crude product without the necess ity of chromatographic purification. The alcohol thus obtained was identical in every respect with that described above. Oppenauer oxidation of 23,16a-cur-19-en-17-ol (137); experiment 1 A solution of the alcohol 137 (50 mg) in benzene (10 ml) was treated with benzophenone (150 mg) and potassium tert-butoxide (84 mg) in benzene (1.7 ml) according to the procedure outlined 137 by Boekelheide. The crude reaction product thus obtained was purified via preparative layer chromatography on silica gel (solvent system A), resulting in the isolation of nor-fluoro• curarine (134) (10 mg, 20%), and 28-cur-19-en-17-al (133) (5 mg, 10%), which were characterized as follows: Nor-fluorocurarine m.p. 180-185° (lit.134 m.p. 185-186°). NMR signals: 9.39 12 0 - Csinglet, IE, CHOI, 6.8-7.4 (multiplet, 4E, aromatic C-R) , 5.43 (quartet, J - 7 Hz, 1H, C-19 H) , 1.60 (doublet with further fine splitting, J = 7 Hz, 3K, C-18 H3). Mass spectrum: M , m/e = 292; main peaks: 263, 249, 167, 121 (base peak). High resolu• tion mass spectrum: calc. for C^gH^^O: 292.1575. Found: 292.1550. 133 Saturated aldehyde 133 IR bands: 3400, 2920, 1715, 1605. UV absorptionsr : A max 298,/ 245;/ lo^g e 3.40,/ 3.83. NMR signalsv : 9.78 (singlet, IH, CHO), 6.5-7.2 (2 multiplets, 4H, aromatic C-H), 5.46 (quartet, J = 7 Hz, IH, C-19 H), 1.60 (doublet of doublets, + J = 7 and 2 Hz, 3H, C-18 H_3) . Mass spectrum: M , m/e = 294; main peaks: 251, 218, 164 (base peak), 145. High resolution mass spectrum: calc. for C19H22N20: 294.1731. Found: 294.1762. Oppenauer oxidation of 2g, 16ot-cur-19-en-17-ol (137); experiment 2 To a solution of the alcohol 137 (300 mg) in dry benzene (50 ml) was added benzophenone (900 mg) and a saturated solution of potassium tert-butoxide in benzene (10 ml). The mixture was stirred at room temperature for 15 min, and another portion of the saturated butoxide solution (5 ml) was added and the mixture heated to reflux. After 15 min refluxing, a further portion (5 ml) of the butoxide solution was added, and the reaction allowed to continue until thin layer chromatography (silica gel, solvent system A) indicated that all starting material had been consumed (total reaction time: 1.5 hr). The reaction was then quenched with water (25 ml) and extracted with benzene. The benzene extracts were combined, dried over anhydrous sodium sulfate, filtered, and the solvent evaporated under reduced - 121 - pressure. The. residue thus obtained was purified by elution through a column of alumina with, benzene having an ether gradient, yielding the saturated aldehyde 133 (115 mg, 39%) and a small amount of nor-fluorocurarine (134) (30 mg, 10%). Oppenauer oxidation of 28,16a-cur-19-en-17-ol (137); experiment 3 To a mixture of benzene (10 ml) and tert-butan.ol (10 ml) was added lithium metal (1.48 g), and the mixture allowed to stir, at 40° until all the metal had been consumed. The excess solvent was then evaporated under reduced pressure, and a solution of the alcohol 137 (600 mg) and benzophenone (3.6 g) in toluene (150 ml) was added to the reaction flask. The mixture was allowed to stir at 90° for 48 hr. At this point, potassium tert-butoxide (2.0 g) was added to the reaction mixture, followed by stirring at 95° for 24 hr. Extraction of the crude reaction product and purification as described above yielded nor-fluorocurarine (134) (280 mg, 47.5%) as well as a small amount of the saturated aldehyde 133 (60 mg, 10%). Oppenauer oxidation of 2g,16a-cur-19-en-17-ol (137); experiment 4 To a solution of the alcohol 137 (1.75 g) in toluene (450 ml) was added benzophenone (21.8 g), and the mixture heated to 90°. Potassium tert-butoxide (7.34 g) was then added and the mixture stirred at 90° for 3 hr. Nitrobenzene (20 ml) was then added, and an instantaneous color change from dark orange to dark green was noted. Extraction of the crude reaction product and purifi• cation as described above yielded nor-fluorocurarine (709 mg, 41%). ^ 122 - No saturated aldehyde 133 could be detected In the reaction mixture. Des-carbomethoxy stemmadenine (150) Nor-fluorocurarine (134) (50 mg) was dissolved in methanol (5 ml), and sodium borohydride (55 mg) added in three equal portions at 15 min intervals at room temperature. After a total reaction time of 4 5 min, the solvent was removed under reduced pressure, and the residue treated with a mixture of water (5 ml) and methylene chloride (5 ml). Extraction of this mixture with methylene chloride (3 x 5 ml) was followed by drying the combined methylene chloride extracts over anhydrous sodium sulfate. After filtration, the solvent was evaporated under reduced pressure, and the residue purified by elution through a column of alumina with benzene having an ether gradient. The indole alcohol 150 (24 mg, 48%) was obtained as a colorless oil which could not be induced to crystallize. IR bands: 3350, 3250 (sh), 2900, 1460. UV absorptionsc : A 290, 283 nm; lo3 g e 3.74, 3.78. NMR signals: max ^ 9.08 (singlet, IH, indole N-H), 7.0-7.6 (multiplet, 4H, aromatic C-H), 5.47 (quartet, J = 7 Hz, IH, C-19 H) , 4.19 (doublet, J = 5 Hz, 2H, CH2OH), 4.00 (doublet, J = 5 Hz, IH, C-16 H), 1.66 + (doublet, J = 7 Hz, 3H, C-18 H_3) . Mass spectrum: M , m/e =• 296; main peaks: 166, 123 (base peak). High resolution mass spectrum: calc. for C19H24N20: 296.1887. Found: 296.1906. Nor-fluorocurarine-cyclohexylamine enamine (151) Nor-fluorocurarine (50 mg) was refluxed with cyclohexylamine (2 ml) and p_-toluenesulfonic acid (5 mg) in benzene (15 ml) with - 123 - azeotropic remoyal of water for 4 hr. The reaction mixture was then cooled and shaken quickly with a saturated aqueous solution of sodium carbonate CIO ml). The benzene layer was dried over anhydrous sodium sulfate, filtered, and the solvent evaporated under reduced pressure. The residue thus obtained was purified via preparative layer chromatography (silica gel, solvent system B), resulting in isolation of the desired enamine 151 as a bright yellow oil (23 mg, 36%). IR bands: 2925, 1630. UV absorptions: ^max 387, 247 nm: log e 4.28, 3.91. NMR signals: 6.9-7.6 Cmultiplet, 5H, aromatic C-H and NCH=C), 5.44 (quartet, J = 7 Hz, IH, C-19 H), 4.06 (singlet, IH, N-H), 1.64 (doublet, J = 7 Hz, 3H, C-18 H^). Mass spectrum: M , m/e = 373. Nor-fluorocurarine-pyrrolidine enamine (152) Nor-fluorocurarine (134) (50 mg) was refluxed with pyrroli• dine (2 ml) in benzene (10 ml) with azeotropic removal of water for 27 hr. At this time, the UV spectrum of the crude reaction mixture indicated the expected shift to A 387, and the solvent ^ max ' evaporated under reduced pressure. However, any attempt to purify the crude product (65 mg) resulted in partial hydrolysis of the enamine, as shown by the reappearance of a singlet at 9.39 in the NMR spectrum, therefore, full characterization of the pyrrolidine enamine 152 was not achieved. Wieland-Gumlich aldoxime C154) Wieland-Gumlich aldehyde C130) (5.0 g) was dissolved in pyridine (75 ml), and hydroxylamine hydrochloride (5.55 g) added. The mixture was allowed to stir at room temperature for 16 hr. The solvent was then evaporated under reduced pressure, and the residue dissolved in water (150 ml). Concentrated ammonium hydroxide was then added dropwise with vigorous stirring, resulting in the precipitation of the desired oxime as white crystals. When the pH of the aqueous suspension reached 10, addition was stopped, and the mixture filtered and washed with water. The residue was then recrystallized from methanol, yielding the desired oxime 154 (4.75 g, 91%) m.p. 243°(d) (lit.127 m.p. 245° (d)). IR absorptions (KBr): 3350, 2900, 1600. NMR signals: 6.6-7.4 (multiplet, 4H, aromatic C-H), 5.67 (triplet J = 7 Hz, IH, C-19 H) . Mass spectrum: M+, m/e = 325; main peaks: 308, 267, 144 (base peak). High resolution mass spectrum: calc. C H N : for 19 23 3°2 325.1789. Found: 325.1817. 18-acetoxy-l-aoetyl-16-cyano-17-nor-2(B, 16a-cur-19-ene (155) Wieland-Gumlich aldoxime (4.75 g) was treated with acetic anhydride (50 ml) and pyridine (8.25 ml) according to the procedure previously described by Smith.131 The desired nitrile 155 was obtained as a light brown oil (5.35 g, 93%). NMR signals 7.0-7.3 (multiplet, 4H, aromatic C-H), 5.64 (triplet, J = 7 Hz, IH, C-19 H), 2.10 (singlet, 3H, CHACON), 2.00 (singlet, 3H, CH^COO). Mass spectrum: M , m/e = 391; main peaks; m/e = 351, 331, 219, 144 (base peak). High resolution mass spectrum: calc. for C23R25N3°3: 291-1895- Found: 391.1801. - 125 - Methyl 18-hydroxy-2g , 16q-cur-19-en-17-oate (156) The nitrile 155 (5.35 g). was treated with barium hydroxide (21.0 g) in ethanol (110 ml) and water (220 ral) exactly accord- 131 mg to the procedure outlined by Smith. The crude reaction product was then treated with 5% methanolic hydrogen chloride (330 ml) according to Smith's procedure, resulting in the isolation of the desired hydroxy ester 156 as a colorless crystalline solid (2.5 g, 53.5%) which could be further purified 131 by recrystallization from methanol m.p. 153-154° (lit. m.p. 154-156). NMR signals-: 6.6-7.2 (multiplet, 4H, aromatic C-H), 5.62 (triplet, J = 7 Hz, IH, C-19 H), 4.22 (disappears after addition of D20) (singlet, IH, O-H) , 3.68 (singlet, 3H, COOCH_3) , 3.39 (disappears after addition of D20) (singlet, IH, N-H). Mass spectrum: M+, m/e = 340; main peaks: 308, 267, 178 (base peak), 144. High resolution mass spectrum: calc. for C^H^N-jO^: 340.1786. Found: 340.1745. Methyl-2g,16ct-cur-19-en-17-oate (157) The hydroxy ester 156 (200 mg) was dissolved in glacial acetic acid (60 ml), and a saturated (at 0°) solution of hydrogen bromide in acetic acid (1.8 ml) was added. The mixture was allowed to stir at room temperature in the dark for 4 8 hr, after which the solvent was evaporated under reduced pressure. The crude product thus obtained was dissolved in glacial acetic acid (100 ml) and methanol (10 ml), and powdered zinc (10 g) was added. After stirring at room temperature for 3 hr, the mixture was filtered, the residue washed with methanol and the filtrate evaporated under reduced pressure. The residue was taken up in ethyl acetate (50 mil and a saturated aqueous solution of sodium carbonate (50 ml). The mixture was extracted with ethyl acetate (3 x 50 ml), and the combined organic extracts dried over anhydrous sodium sulfate, filtered, and evaporated under reduced pressure. The residue was purified on a column of alumina with elution by benzene, resulting in the desired ester 157 as a colorless oil (130 mg, 65%) which could not be induced to crystallize. IR bands: 3420, 2950, 1730, 1605, 1480, 1440. UV absorptions: Amax 297' 244; log e 3-47' 3-81- NMR signals: 6.5-7.2 (multiplet, 4H, aromatic C-H), 5.48 (quartet, J = 7 Hz, IH, C-19 H), 3.70 (singlet, 3H, COOCH_3) , 1.58 (doublet of doublets, J = 7 and 2 Hz, 3H, C-18 H^). Mass spectrum: M , m/e = 324; main peaks: m/e = 251, 194 (base peak)r 144, 139 and 130. High resolution mass spectrum: calc for C2QH24N202: 324.1837. Found: 324.1766. Methyl-l-formyl-26,16a-cur-19-en-17-oate (158) The ester 157 (250 mg) was dissolved in benzene (2 ml) and added to sodium hydride (100 mg) (obtained by washing a 50% oil dispersion (200 mg) with benzene (3x1 ml)), followed by the addition of methyl formate (3 ml). The mixture was heated to reflux with stirring under an atmosphere of nitrogen gas for 40 min. The excess sodium hydride was then destroyed by the drop- wise addition of glacial acetic acid until ebullition was no longer observed. The mixture was then eluted through a column of alumina with benzene. The first fraction (25 ml) yielded pure N-formyl ester 158 (178 mg), and the succeeding four 127 - fractions C10Q ml) yielded less pure material which, was purified via preparative layer chromatography on silica gel (solvent system B), the combined processes yielding a white crystalline solid (216 mg, 80%) m.p. 162-170°. IR bands: 2960, 1730, 1668, 1600. UV absorptionsr : X 287, 278, 250; ' ' ' max ' ' log e 3.44, 3.46, 3.98. NMR signals: 8.71 (singlet,.IH, NCHO), 5.38 (quartet, J = 7 Hz, IH, C-19 H) , 3.68 (singlet, 3H, COOCH_3) , 1.53 (doublet of doublets, J = 7 and 2 Hz, C-18 H ) Mass spectrum: H+, m/e = 352; main peaks: 279, 194, 172 (base peak), 144. High resolution mass spectrum: calc. for C2iH24N2°3: 352.1785; Found: 352.1829. Carbomethoxy tetrahydrooxazine 159 Paraformaldehyde (400 mg) was mixed with phosphorous pen- toxide (400 mg) and heated under vacuum (1 mm). The formaldehyde which was released was condensed as a slurry by passing the vapor into a flask cooled by liquid nitrogen. This slurry was then similarly distilled into a flask containing sodium hydride (25 mg, obtained by washing a 50% oil dispersion (50 mg) with benzene (3x1 ml)) and dimethyl sulfoxide (2 ml). After the distillation was complete, the reaction flask was warmed to room temperature and the N-formyl ester 158 (50 mg) was dissolved in dimethyl sulfoxide (1 ml) and added to the reaction mixture. After stirring at room temperature for 4 hr, the reaction was quenched by adding glacial acetic acid until ebullition was no longer observed. The mixture was then made basic with a - . 128 - saturated aqueous solution of sodium carbonate (10 ml) and extracted with, ethyl acetate (4 :x 15 ml}. The ethyl acetate extracts were then combined and washed with water (-4 x 10 ml) , dried over anhydrous sodium sulfate, filtered and the solvent evaporated under reduced pressure. Purification was achieved by eluting the crude reaction product through a column of alumina with benzene as eluent. The desired product was obtained as a colorless oil (40 mg, 78%). IR bands: 2960, 1730, 1602. UV absorptions: Amax 299, 246; log e 3.51, 3.97. NMR (FT) signals (Figure 36): 6.6-7.2 (2 multiplets, 4H, aromatic C-H), 5.46 (quartet, J = 7 Hz, IH, C-19 H), 5.18 (doublet, J = 10 Hz, IH, N-CHH-O), 4.71 (doublet, J = 10 Hz, IH, N-CHH-O), 4.94 (singlet, IH, C-2 H) , 3.70 (singlet, 3H, COOCH_3) , 1.57 (doublet of doublets, J = 7 and 2 Hz, 3H, C-18 H). Mass spectrum: M+, m/e = 366; main peaks: 336, 280, 279, 250, 206, 143 (base peak). High resolution mass spectrum: calc. for C22H26N2°3: 366-1943- Found: 366.1905. Model tetrahydrooxazine 166 The alcohol 137 (500 mg) was dissolved in methanol (15 ml), and paraformaldehyde (500 mg) and anhydrous sodium sulfate (1.0 g) were added. The mixture was allowed to stir at room temperature for 18 hr, then filtered and the solvent evaporated under reduced pressure. The residue was taken up in a saturated aqueous solu• tion of sodium carbonate (25 ml) and methylene chloride (25 ml). The mixture was extracted with methylene chloride (3 x 25 ml) and the organic extracts combined, dried over anhydrous sodium sulfate, filtered, and the solvent evaporated under reduced pressure. The - 129 - residue was purified by elution through, a column of alumina with methylene chloride as eluent, yielding the desired tetrahydro• oxazine 166 as a colorless oil (520 mg, 100%). IR bands: 2940, 2870, 1601.4. UV absorptions: \ 297, 249; log e 3.52, 3.98. c max NMR signals (Figure 37}: 6.76-7.3 (2 multiplets, 4H, aromatic C-H), 5.41 (quartet, J = 7 Hz, IH, C-19 H), 5.25 (doublet, J = 11 Hz, IH, N-CHH-O), 4.67 (doublet, J = 11 Hz, IH, N=CHH-0), 1.50 (doublet with some fine splittings, J = 7 Hz, 3H, C-18 H^). Mass spectrum: M+, m/e = 308; main peaks: 278, 149, 148, 144, 143 (base peak). High resolution mass spectrum: calc. for C20H24N2O: 308-1887- Found: 308.1932. Hydrolysis of model tetrahydrooxazine 166 The tetrahydrooxazine 166 (80 mg) was dissolved in 10% methanolic hydrogen chloride (10 ml) and stirred at room tempera• ture for 8 hr. The methanol was then evaporated under reduced pressure and the residue taken up in a saturated aqueous solution of sodium carbonate (5 ml) and extracted with methylene chloride (3x5 ml). The combined methylene chloride extracts were dried over anhydrous sodium sulfate, filtered, and the solvent evaporated under reduced pressure. The alcohol 137 was obtained as color• less crystals (77 mg, 100%) which were identical with a previously prepared standard sample in every respect. 2g,16g-carbomethoxy-cur-19-en-17-ol (16 0) The carbomethoxy tetrahydrooxazine 159 (71 mg) was dissolved in 10% methanolic hydrogen chloride (15 ml) and the solution r> 130 - refluxed for 2,5 hx. Isolation of the product as. described aboye for the alcohol 137 followed by; elution through a short column of aiumijia with methylene chloride yielded the desired hydroxy ester 160 as a colorless oil (63.5 mg, 92%). IR bands: 3360, 2960, 1728, 1605. W absorptions: ^max 296, 244: log e 3.38, 3.76. NMR signals (FT): 6.7-7.2 (multiplet, 4H, aromatic C-H), 5.54 (quartet, J = 7 Hz, IH, C-19 H), 4.66 (singlet, IH, C-2 H) , 3.73 (singlet, 3H, COOCH_3) , 1.64 (doublet of doublets, + J = 7 and 2 Hz, 3H, C-18 H_3) . Mass spectrum: M , m/e = 352; main peaks: 324, 251, 224, 199, 144 (base peak). High resolu• tion mass spectrum: calc. for C21H26N2°3: 354.1942. Found: 354.1915. Akuammicine (66) . < The ester 157 (200 mg) was dissolved in benzene (25 ml), and acetic acid (35 pi) was added. A solution of lead tetra• acetate (550 mg) in benzene (40 ml) was added dropwise with stirring at room temperature over a period of 1.75 hr. The reaction mixture was allowed to stir for an additional 15 min, after which a 1 M aqueous sodium carbonate solution (10 ml) was added, and extracted with ethyl acetate (3 x 10 ml). The combined organic extracts were dried over anhydrous sodium sulfate, filtered, and the solvent evaporated under reduced pressure. The residue was then purified by elution through a column of alumina with, benzene, yielding the desired akuammicine (661 as colorless crystals which could be recrystallized from ethyl acetate (70 mg, 35%) m.p. 178-180° (lit.158 m.p. 180-181.5°). - 131 - IR bands; 3495, 2960, 1670, 1601. UV absorptions; A 326, 296, 226; log e 4.17, 4.03, 4.05. NMR signals (FT) (Figure 40): 9.01 (singlet, IH, N-H), 6.8-7.3 (multiplet, 4H, aromatic C-H), 5.34 (quartet, J = 7 Hz, IH, C-19 H), 3.82 (singlet, 3H, COOCH3), + 1.62 (.doublet, J = 7 Hz, 3H, C-18 H_3) . Mass spectrum: M , m/e = . 322; main peaks: 121 (base peak). High resolution mass spectrum: calc. for C20H22N2O2: 322.1680. Found: 322.1657. Des-hydroxymethylene stemmadenine (141a) To a solution of akuammicine (66) (94 mg) in refluxing glacial acetic acid (10 ml) was added sodium borohydride (1 g) as rapidly as possible. The solution was allowed to stir without heating for 15 min, then cooled in an ice bath. The mixture was slowly made basic with 7 N ammonium hydroxide and then extracted with ethyl acetate (3x5 ml). The ethyl acetate extracts were combined, dried over anhydrous sodium sulfate, filtered, and the solvent evaporated under reduced pressure. The residue was purified by preparative layer chromatography on silica gel (solvent system B), yielding the desired indole ester 141a as a colorless oil (50 mg, 50%). IR bands: 3450, 2930, 1720, 1670 (w), 1601 (w). UV absorptions: A 290, 283, 225; log e 3.73, HI 3.X 3.74, 4.37. NMR signals (FT) (Figure 42): 9.08 (singlet, IH, indole N-H), 7.1-7.6 (multiplet, 4H, aromatic C-H), 5.60 (quartet, J = 7 Hz, IH, C-19 H), 4.30 (singlet, IH, C-16 H), 3.88 (singlet, 3H, COOCH3, 1.77 (doublet, J = 7 Hz, 3H, C-18 H_3) . Mass spectrum: M , m/e ~ 324; main peaks: 194, 123 (base peak). High resolution mass spectrum: calc. for Cor.H_.N„0„: 324.1837. Found: 324.1857. 20 24 2 2 - 132 - Epimeric Indole ester 141b A solution of indole ester 141a (10 mg) in benzene (1 ml) was added to sodium hydride (10 mg, obtained by washing a 50% oil dispersion (20 mg) with benzene), and methyl formate (1 ml) was distilled (over phosphorous pentoxide) directly onto the mixture. The reaction mixture was heated to reflux for 2.5 hr, then cooled and passed through a short column of alumina, followed by additional benzene (50 ml). The solvent was evapor• ated under reduced pressure, and the residue purified on a silica gel preparative layer chromatography plate (solvent system B) yielding the epimeric indole ester 141b as a colorless oil (3.6 mg, 36%). IR bands: 3460, 2940, 1720, 1670 (w), 1601 (w). UV absorptions: ^max 290, 283, 225; log e 3.73, 3.74, 4.37. NMR signals (FT): 8.8 (singlet, IH, indole N-H), 7.0-7.5 (multiplet, 4H, aromatic C-H), 4.64 (quartet, J = 7 Hz, IH, C-19 H), 4.20 (doublet, J = 4 Hz, IH, C-16 H) , 3.76 (singlet, 3H, COOCH-j) , 1.58 + (doublet, J = 7 Hz, 3H, C-18 H_3) . Mass spectrum: M , m/e = 324; main peaks: 194, 123 (base peak). 16-epi-stemmadenine (161) To a solution of the hydroxy ester 160 (27 mg) in benzene (1 ml) was added acetic acid (10 u1). A solution of lead tetra• acetate (67 mg) in benzene (1 ml) was then added dropwise with stirring over a period of 30 min. The reaction was allowed to stir for 15 min, then was eluted through a small column of alumina with methylene chloride (25 ml). The solvent was evaporated under reduced pressure, and the residue dissolved in 133 r- a 50%. solution of methanol in acetic acid (.4 ml). Sodium borohydride CO.5 g) was then added to the stirred solution as rapidly as possible, and the reaction allowed to stir for 15 min. The mixture was then cooled in an ice bath, made basic with 7 N ammonium hydroxide, and extracted with methylene chloride (3 x 5 ml). The combined methylene chloride extracts were dried over anhydrous sodium sulfate, filtered, and the solvent evaporated under reduced pressure. Purification of the residue by preparative layer chromatography on silica gel Csolvent system B) yielded 16-epi-stemmadenine (161) (7.0 mg, 26%) as a colorless oil which could not be induced to crystall• ize. IR bands: 3580, 3420, 2920, 1725, 1605 (w). UV absorp• tions: A 291, 284, 225; log e 3.68, 3.69, 4.33. NMR signals max J (FT): 10.14 (singlet, IH, indole N-H), 7.0-7.7 (multiplet, 4H, aromatic C-H), 5.42 (quartet, J = 7 Hz, IH, C-19 H), 4.36 (singlet, 2H, CH2OH), 3.88 (singlet, 3H, COOCH_3) , 1.52 (doublet, + J = 7 Hz, 3H, C-18 H3). Mass spectrum: M , m/e = 354; main peaks: 336, 324, 123 (base peak). High resolution mass spectrum; calc. for C21H26N2°3: 354-1942- Found: 354.1926. Stemmadenine diol (175) . 159 To a stirred solution of authentic stemmadenine (10 mg) in tetrahydrofuran (11 ml) was added excess sodium bis (methoxy- methylenoxy) aluminum hydride as the commercially available 70% benzene solution CRed-Al) CO.2 ml). The reaction was allowed to stir at room temperature for 30 min, then made basic with 5 N ammonium hydroxide and extracted with methylene chloride (3 x TS 134 - 5 mil. The methylene chloride extracts were combined, dried over anhydrous sodium sulfate, filtered, and the solvent evaporated under reduced pressure. The residue was purified by preparative layer chromatography on silica gel (solvent system B), yielding the desired diol 175 (7 mg, 76%) as a colorless oil which could not be induced to crystallize. In an identical manner, the synthetic 16-epi-stemmadenine (161) was reduced to diol 17 5. The two products were found to be identical in every respect, each providing the data given below. IR bands: 3400, 2920, 1470. UV absorptions: A 290, c • max 283, 266; log e 3.73, 3.74, 4.41. NMR signals: 9.68 (singlet, IH, indole N-H), 7.0-7.6 (multiplet, 4H, aromatic C-H), 5.40 (quartet, J = 7 Hz, IH, C-19 H), 4.0-4.5 (multiplet, 4H, 2CCH_2OH), 1.64 (doublet, J = 7 Hz, 3H, C-18 H_3) . Mass spectrum: M+, m/e = 326; main peaks: 324, 276, 197, 123 (base peak). High resolution mass spectrum: calc. for C„rtH„-No0o: 326.1994. 20 26 2 2 Found: 326.1983. f - 135 - Section C . 16-epl— stemmadenine (Ar- Hi (161) Tritiated trifluoroacetic acid-^-1- (1.57 g, 1 Ci) was combined with. 16-epi-stemmadenine (18 mg) by means of a vacuum transfer system. The acid solution was maintained under a dry nitrogen atmosphere at room temperature for 48 hr. The tritiated trifluoroacetic acid was then removed via the vacuum transfer system, and the residue taken up in 7 N ammonium hydroxide and extracted with methylene chloride (3 x 5 ml). The methylene chloride extracts were combined, dried over anhydrous sodium sulfate, filtered, and the solvent evaporated under reduced pressure. The residue was then purified via preparative layer chromatography on silica gel (solvent system B) yielding the tritiated 16-epi-stemmadenine (7.0 mg, 39%) as a colorless oil. Activity: 2.04xl010 dpm/mg; 7.23xl012 dpm/mmol. Biosynthetic experiments involving the administration of labelled stemmadenine (6)110 and 16-epi-stemmadenine (161) to A. pyricollum were carried out exactly as described in Section A. The results of these experiments are shown in Tables 10 and 11. -> 136 ~ BIBLIOGRAPHY K. Mothes, The Alkaloids, 6, R.F. Manske, ed, , Academic Press, New York C1960]. T. Robinson, "The Biochemistry of Alkaloids", Molecular Biochemistry and Biophysics, 3_, A. Kleinzeller, G.F. Springer, H.G. Wittmann, ed., Springer-Verlag, New York C1968) . M. Hesse, Indolalkaloide in Tabellen, Springer, Berlin, 1 (1964) , 2 (1968) . V. Snieckus, The Alkaloids, 11, R.F. Manske, ed., Academi Press, New York (1960). D. Groger, K. Stolle, K. Mothes, Tetrahedron Letters, 257 (1964) . E. Leete, A. Ahmad, I. Kompis, J. Amer. Chem. Soc., 87, 4168 (1965). E. Leete, Acc. Chem. Res., 2, 59 (1969). I.D. Spenser, Compr. Biochem., 20, 231 (1968). A.R. Battersby, A.R. Bunnett, P.G. Parsons, Chem. Comm., 137 (1968). A. Pictet, Arch. Pharm., 224, 389 (1906) G. Barger, C. Scholz, Helv. Chim. Acta, 16, 1343 (1933). G. Hahn, H. Ludwig, Chem. Ber., 67, 203 (1934). L. Fowden, Plant Biochemistry, p. 361, ed., J. Bonner, Academic press, New York (1965). J.R. Mattoon, Biogenesis of Natural Compounds, ed., P. Bernfeld, p. 34, Pergamon, Oxford (1963). - 137 - 15. M.I, Gibson, F, Gibson, J. Biochem, 90, 16Q (1964) 16. V.H. Cheldelin, A. Bauch, Biogenesis of Natural Compounds, ed., P. Bernfeld, p. 685, Pergamon, Oxford (1963). 17. J.R. Mattoon, op. cit, p. 36. 18. R.G.H. Cotton, F. Gibson, Biochim. Biophys. Acta, 156, 187 (1968). 19. R. Robinson, The Structural Relation of Natural Products, Clarendon Press, Oxford (1955). 20. R.B. Woodward, Angew Chem., 68, 13 (1956). 21. R.B. Woodward, Nature, 162, 155 (1948). 22. R.B. Turner, R.B. Woodward, The Alkaloids, 3_, R.F. Manske, ed., Academic Press, New York (1953). 23. E. Wenkert, N.V. Bringi, J. Amer. Chem. Soc, 81, 1474 (1959). 24. E. Wenkert, N.V. Bringi, Experientia, 15, 165 (1959). 25. R. Thomas, Tetrahedron Letters, 544 (1961). 26. E. Wenkert, J. Amer. Chem. Soc, 84, 98 (1962). 27. G. Buchi, R.E. Manning, Tetrahedron Letters, 5 (1960). 28. C. Djerassi et al., J. Org. Chem., 26, 1192 (1961). 29. W. Haegele, F. Kaplan, H. Schmid, Tetrahedron Letters, 110 (1961). 30. J. Grimshaw, Chem. and Ind., 403 (1961). 31. E. Leete, S. Ghosal, P.N. Edwards, J. Amer. Chem. Soc, 8j4, 1068 (1962). 32. A.R. Battersby, R. Binks, W. Lawrie, G.V. Parry, B.R. Webster, J. Chem. Soc, 7459 (1965). - 138 - 33. J.G. Leyin, D.B. Sprlnson, Biochem. and Biophys. Res. Commun., 3, 157 (1960). 34. E. Schlittler, W.I. Taylor, Experientia/ 16, 224 (I960). 35. E. Leete, S. Ghosal, Tetrahedron Letters, 1179 (1962). 36. J.B. Hendrickson, The Biosynthesis of Steroids, Terpenoids and Acetogenins, J.G. Richards and J.B. Hendrickson, ed., W.A. Benjamin, New York (1969). 37. A.R. Battersby, R. Binks, W. Lawrie, G.V. Parry, B.R. Webster, J. Chem. Soc, 369 (1963). 38 E. Leete, A. Ahmad, I. Kompis, J. Amer. Chem. Soc, 87, 4168 (1965). 39 D.H.R. Barton, G.W. Kirby, R.H. Prager, E.M. Wilson, J. Chem. Soc, 3990 (1965). 40 K. Stolle, D. Groger, K. Mothes, Chem. and Ind., London, 2065 (1965). 41. F. McCapra, T. Money, A.I. Scott, I.G. Wright, Chem. Comm., 537 (1965). 42. H. Goeggel, D. Arigoni, Chem. Comm., 538 (1968) 43. A.R. Battersby, R.T. Brown, R.S. Kapil, A.O. Plunkett, J.B. Taylor, Chem. Comm., 46 (1966). 44 A.R. Battersby, R.T. Brown, J.A. Knight, J..A. Martin, A.O. Plunkett, Chem. Comm., 346 (1966). 45 A.I. Scott, Karl Folkers Lecture Series, University of Wisconsin, Dec. 1964. 46 T. Money, I.G. Wright, F. McCapra, A.I. Scott, Prea. Nat. Acad. Sci. U.S., 53, 9Q1 (1965). 139 - E,S, Hall, F, McCapra, T, Money, K, Fukumoto, J.R. Hanson B,S, Mootoo, G.T. Phillips, A,I. Scott, Chem.. Comm. , 348 (1966). T. Money, I.G. Wright, F. McCapra, E.S. Hall, A.I. Scott, J. Amer. Chem. Soc, 90, 4144 C1968) . A.R. Battersby, R.T. Brown, R.S. Kapil, J.A. Knight, J.A. Martin, A.O. Plunkett, Chem. Comm., 888 (1966). A.R. Battersby, Pure Appl. Chem., 14, 117 (1967). P. Loew, H. Goeggel, D. Arigoni, Chem. Comm., 347 (1966). E. Leete, S. Ueda, Tetrahedron Letters, 4915 (1966). A.R. Battersby, R.T. Brown, R.S. Kapil, A.O. Plunkett, Chem. Comm., 890 (1966). A.R. Battersby, R.S. Kapil, J.A. Martin, Mrs. Lucy Mo, Chem. Comm., 890 (1966). A.R. Battersby, R.S. Kapil, R. Southgate, Chem. Comm., 131 (1968). A.R. Battersby, E.S. Hall, R. Southgate, J. Chem. Soc., (c), 721 (1969). S. Brechbuhler-Bader, C.J. Coscia, P. Loew, C. von Szczepanski, D. Arigoni, Chem. Comm., 136 (1968). C.J. Coscia, R. Guarnaccia, Chem. Comm., 139 (1968). A.R. Battersby, J.C. Byrne, R.S. Kapil, J.A. Martin, T.G. Payne, D. Arigoni, P. Loew, Chem. Comm., 951 (1968). R, Guarnaccia, L. Botta, C.J. Corcia, J. Amer. Chem. Soc, 91_, 204 (1969). P.J. Lentz, M.G. Rossman, Chem. Comm., 1269 (1969). - 140 62. H_, Inouye, T, Yoshida, S, Tobita, M. Okigawa, Tetrahedron , 16, 39.05 C1970L. 63. H. Inouye, T. Yoshida, S. Tobita, Tetrahedron Letters, 2945 C1968). 64. R. Guarnaccia, L. Botta, C.J. Corcia, J. Amer. Chem. Soc, 92_, 6098 (1970). 65. P. Loew, D. Arigoni, Chem. Comm., 137 (1968). 66. A.R. Battersby, B. Gregory, Chem. Comm., 134 (1968). 67. A.R. Battersby, A.R. Burnett, P.G. Parsons, Chem. Comm. , 826 (1970) . 68. H. Inouye, S. Ueda, Y. Aoki, Y. Takeda, Tetrahedron Letters, 2351 (1969) . 69. A.R. Battersby, A.R. Burnett, P.G. Parsons, J. Chem. Soc., (c) 1187 (1969). 70. S. Escher, P. Loew, D. Arigoni, Chem. Comm., 823 (1970). 71. A.R. Battersby, S.H. Brown, T.G. Payne, Chem. Comm., 827 (1970) . 72. A.R. Battersby, A.R. Burnett, P.G. Parsons, J. Chem. Soc., (c), 1193 (1969). 73. G.N. Smith, Chem. Comm., 912 (1968). 74. W.P. Blackstock, R.T. Brown, G.K. Lee, Chem. Comm., 910 (1971) . 75. K.T.D. DeSilva, G.N. Smith, K.E.H. Warren, Chem. Comm., 905 (1971) . 76. A.R. Battersby, A.R. Burnett, E.S. Hall, P.G. Parsons, Chem. Comm., 1582 (1968). - 141 - 77. R.T. Brown, G.N, Smith., S.J, Stapleford, Tetrahedron Letters, 4349 0.9.69!. 78. E. Wenkert, B. Wickberg, J. Amer. Chem. Soc, 87, 1580 C1965) . 79. A.I. Scott, A.A. Qureshi, J. Amer. Chem. Soc, 91, 5874 C1969). 80. A.A. Qureshi, A.I. Scott, Chem. Comm., 949 (1968). 81. A.R. Battersby, E.S, Hall, Chem. Comm., 793 (1969). 82. A.I. Scott, P.C. Cherry, A.A. Qureshi, J. Amer. Chem. Soc, 91, 4932 (1969). . 83. J.P. Kutney, E. Piers, R.T. Brown, J.Amer. Chem. Soc., 92^, 1700 (1970). 84. J.P. Kutney, W.J. Cretney, J.R. Hadfield, E.S. Hall, V.R. Nelson, J. Amer. Chem. Soc, 92, 1704 (1970). 85. J.P. Kutney, R.T. Brown, E. Piers, J.R. Hadfield, J. Amer. Chem. Soc, 92, 1708 (1970). 86. J.P. Kutney, W.J. Cretney, P. LeQuesne, B. McKague, E. Piers, J. Amer. Chem. Soc, 92, 1712 (1970). 87. J.P. Kutney, N. Abdurahman, C. Gletsos, P. LeQuesne, E. Piers, I. Vlattas, J". Amer. Chem. Soc. , 92, 1727 (1970). 88. J.P. Kutney, W.J. Cretney, J.R. Hadfield, E.S. Hall, V.R. Nelson, D.C. Wigfield, J. Amer. Chem. Soc, 90, 3566 (1968). 89. J.P. Kutney, C. Ehret, V.R. Nelson, W.C. Wigfield, J. Amer. Chem. Soc, 90, 5929 (1968). 90. G.A. Cordell, G.E. Smith, G.N. Smith, Chem. Comm., 193 (1970). 91. A.R. Battersby, A.K. Bhatnager, Chem. Comm., 189 (1970). R.T. Brown, G.F. Smith., K.§. Stapleford, D.A. Taylor, Chero. Comm., 19.0 (1970) . R.S. Sood, Studies Related to the Biosynthesis of Indole Alkaloids, Ph.D. Thesis, The University of British Columbia C1970) . J.P. Kutney, J.F. Beck, N.J. Eggers, H.W. Hanssen, R.S. Sood, N.D. Westcott, J. Amer. Chem. Soc, 93, 7322 (1971). J.F. Beck, Studies on the Biosynthesis of Indole Alkaloids, Ph.D. Thesis, The University of British Columbia (1971). J.P. Kutney, C. Ehret, G. Poulton, R.S. Sood, N.D. Westcott, Bioorg. Chem., 1, 194 (1971). J.P. Kutney, J. Heterocyclic Chem. , 9_, Supplementary Issue, S-l (1972). J.P. Kutney, J.F. Beck, V.R. Nelson, R.S. Sood, J. Amer. Chem. Soc, 93_, 255 (1971). J.P. Kutney, V.R. Nelson, D.C. Wigfield, J. Amer. Chem. Soc, 91, 4278 (1969) . N.J. Eggers, Studies on the Biosynthesis of Indole Alkaloids, Ph.D. Thesis, The University of British Columbia (1973). J.P. Kutney, N.J. Eggers, Heterocycles, 1, 11 (1973). J.A. Joale, H. Monteiro, L.J. Durham, B. Gilbert, C. Djerassi, J. Chem. Soc, 4773 (1965). R.R. Arndt, S.H. Brown, N.C. Ling, P. Roller, C. Djerassi, J.M. Ferreira, B. Gilbert, E.C. Miranda, S.E. Flores, A.P. Duarte, E.P. Carrazzoni, Phytochemistry, 6, 1652 (1967). J.P. Kutney, V.R. Nelson, D.C. Wigfield, J. Amer. Chem. Soc., 4279 (1969). - 143 - 10 5. A. Ahond, A. Cava, C. Kan-Fan, Y. Langlois, P. Potier, Chem. Comm., 517 C1970). 106. P. Potier, M.M. Janot, CR. Acad. Sc. Paris, 276, 1727 (1973). 107. Kindly provided by Professor B. Gilbert, Research Center for Natural Products, Faculty of Pharmacy, Rio de Janeiro, Brazil. 108. B. Gilbert, A.P. Duarte, Y. Nakagawa, J.A. Joule, S.E. Flores, J.A. Brissolese, J. Campello, E.P. Carrazzoni, R.J. Owellen, E.C. Blossey, K.S. Brown, Jr., C. Djerassi, Tetrahedron, 21, 1144 (1965). 109. M.A. Ondetti, V. Deulofew, Tetrahedron, 15, 160 (1961). 110. In our laboratory, this conversion was carried out by Mr. D. Grierson, whose assistance is gratefully acknow• ledged. 111. G. Kunesch, J. Polonsky, Phytochemistry, 8, 1221 (1969). 112. Horning and Horning, J. Org. Chem., 11, 95 (1946). 113. E. Wenkert, K.G. Dave, CT. Gnewuch, P.W. Sprague, J.Amer. Chem. Soc, 90, 5251 (1968). 114. J.P. Kutney, J.F. Beck, G.B. Fuller, Heterocycles, 1, 5 (1973). 115. G.A. Cordell, G.F. Smith, G.N. Smith, Chem. Comm., 191 (1970). 116. F. Walls, 0. Collera, A. Sandoval L., Tetrahedron, 2, 173 (1958). 117. A. Sandoval, F. Walls, J.N. Shoolery, J.M. Wilson, H. Budzikiewicz, C Djerassi, Tetrahedron Letters, 409 (1962). - 144 - 118. al K. Biemann, A.L, Burlijigame, D, Stauffacher, Tetrahedron Letters, 527 (1962). b) D. Schumann, H. Schmid, Helv. Chimica Acta, 46, 1996 C1963) . c) M.-M. Janot, Pure Appl.Chem., 6, 635 (1963). 119. G.F. Smith, J.T. Wrobel, J. Chem. Soc, 792 (1960). 120. K. Biemann, G. Spiteller, J. Amer. Chem. Soc, 84, 4578 (1962) . 121. V.R. Nelson, Studies Related to Synthesis and Biosynthesis of Indole Alkaloids, Ph.D. Thesis, The University of British Columbia (1969). 122. M. Hearne, unpublished results, The University of British Columbia (1971). 123. T. Oishi, M. Magai, Y. Ban, Tetrahedron Letters, 491 (1968). 124. J.Y. Laronze, J. Laronze-Fontaine, J. Levy, J. LeMen, Tetrahedron Letters, 491 (1974). 125. A.J. Gaskell, H.E. Radung, E. Winterfeldt, Tetrahedron, 26_, 5353 (1970) . 126. H. Wieland, W. Gumlich, Liebigs Ann. Chem., 494, 191 (1932). 127. H. Wieland, K. Kaziro, Liebigs Ann. Chem., 506, 60 (1933). 128. F.A.L. Anet, Sir R. Robinson, J. Chem. Soc, 2253 (1955). 129. J.R. Hymon, H. Schmid, P. Karrer, A. Boiler, H. Els, P. Fabrui, A. Furst, Helv. Chim. Acta, 52, 1564 (1969). 130. The early work in the synthetic investigations described here were carried out on a very generous gift of WGA supplied by Professor H. Schmid, The University of Zurich, Switzerland. 131. P.N. Edwards, G.F. Smith, J. Chem. Soc, 152 (1961). ^ 145 - 132. K. Berna,uer, F. Berlage, W. von Philipsborn, H. Schmid, P. Karrer, Helv. Chim. Acta, 41, 2293 (1958). 133. K. Bernauer, H. Schmid, (to Hoffmann-LaRoche, Inc.) U.S. Patent 3, 073, 832 (Cl. 260 = 293), Jan. 15, 1963, CA. , 58, 9160 b (1963). 134. H. Fritz, E. Besch, T. Wieland, Ann., 663, 150 (1963). 135. K.E. Wilzbach, J. Amer. Chem. Soc, 79, 1013 (1957). 136. H. Fritz, E. Besch, T. Wieland, Angew. Chem., 71, 126 (1959). 137. M. Grdinic, D.A. Nelson, V. Boekelheide, J. Amer. Chem. Soc, 86, 3357 (1964). 138. L. Rosenthaler, Folia. Pharm., 4_, 523 (1962). 139. H. Wieland, H.J. Pistor, K. Bahr, Ann., 547, 140 (1941). 140. For a review on calabash-curare alkaloids, see The Alkaloids., 11, R.F. Manske, ed., Academic Press, New York (1968). 141. J. Kebrle, H. Schmid, P. Waser, P. Karrer, Helv. Chim. Acta, 363, 345 (1953). 142. H. Fritz, E. Besch, T. Wieland, Ann., 617, 166 (1958). 143. K. Nakanishi, Infrared Absorption Sepctroscopy, Holden Day Inc., San Francisco. (1962). 144. E.J. Corey, N.W. Gilman, B.E. Ganem,.J. Amer. Chem. Soc, 90, 5616 (1968). 145. J. Attenburrow et al. , J. Chem. Soc, 1094 (1952). 146. N.W. Gilman, Chem. Comm., 733 (1971). 147. K. Bernauer, W. Arnold, Ch. Weissmann, H. Schmid, P. Karrer, Helv. Chim. Acta, 43, 717 (1960). 148. D.I-L.R. Barton, R.H. Hesse, M.M. Pechet, C. Wiltshire, Chem. Comm ., 1017 (1972). - 146 - 149. BL. Fritz, A. Krekel, K. .Meyer, Ann. , 664, 188 C19 63) . 150. B. Budzikiewicz, J.M. Wilson, C. Djerassi, J. Levy, J. LeMen, M.-M. Janot, Tetrahedron, 19, 1265 (1963). 151. J.C. Powers, R. Seidner, T.G. Parsons, H.J. Berwin, J. Org. Chem., 31, 2623 (1966) . 152. R.J. Cregge, J.L. Hermann, CS. Lee, J.E. Richman, R.H. Schlesinger, Tetrahedron Letters, 2425 (1973). 153. F. Puisieux, A. LeHir, M. Delepine, Compt. Rendus. 252, 902 (1961). 154. F. Puisieux, A. LeHir, R. Goutarel, M.-M. Janot, J. LeMen, Ann. Pharm. Franc, 17, 626 (1959). 155. F. Puisieux, R. Goutarel, M.-M. Janot, J. LeMen, A. LeHir, Compt. Rendus., 250, 1285 (1960). 156. M.F. Bartlett, B.F. Lambert, W.I. Taylor, J. Amer. Chem. Soc., 88, 729 (1964). 157. Physical Data of Indole and Dihydroindole Alkaloids, 1_, Lilly Research Laboratory, Eli Lilly and Co., Fifth Edition (1961). 158. R.B. Woodward, M.P. Cava, W.D. Olles, A. Hunger, H.U. Daeniker, K. Schenker, J. Amer. Chem. Soc, 76, 4749 (1954). 159. Samples of natural stemmadenine kindly provided by U. Renner, Ciba-Geigy Research Laboratories, Basel, Switzerland. 160. R.W. Hendler, Ana. Biochem. , 1_, 110 (1964). 161. Mark I Liquid Scintillation Systems Instruction Manual, Section 1, Nuclear Chicago Corporation, Des Plaines, Illinois (1966) . 147 - "Nevertheless, scholarship, as was true for art in the olden days, must indeed have far-flung grazing grounds, and in pursuit of a subject which interests no one but himself a scholar can accumulate knowledge which provides colleagues with information as valuable as that stored in a dictionary or an archive." -Hermann Hesse, The Glass Bead Game