FACULTY OF SCIENCE

THESIS

SUBMITTED

TO THE SCHOOL OF GRADUATE STUDIES

OF LAVAL UNIVERSITY

for the

DEGREE OF DOCTOR OF SCIENCE

by

JOSE DOMINGO MEDINA', L.Ch.

ALKALOIDS OF VENEZUELAN APOCYNACEAE

June, 1968

To my wife

Ill

TABLE OF CONTENTS

CONTENTS . .

ACKNOWLEDGEMENTS

LIST OF FIGURES

INTRODUCTION .

Page

iii

iv

v

1

Chapter I : A. Alkaloids of A. excelsum ....

B. Alkaloids of A. cuspa ......

Chapter II : Alkaloids of A. fendleri ......

Chapter III : Alkaloids of T. psychotrifolia . . .

EXPERIMENTAL ..................

General remarks ..............

Chapter I : A. Alkaloids of A. excelsum .

B. Alkaloids of A. cuspa . . .

Chapter II : Alkaloids of A. fendleri . . .

Chapter III : Alkaloids of T. psychotrifolia

9

17

. 46

. 60

. 91

. 92

94

. 98

. Ill

. 118

. 133

REFERENCES 137

IV

■ 'ACKNOWTVBUGEMENTS

I wish to express my profound gratitude to Professor R.H. Burnell for

his continuous encouragement through the duration of this investigation. His

role as scientific tutor was always most important for the achievement of

this work.

I am very much indebted to the Instituto de Investigacion.es Cientificas

CIVIC) of Caracas, Venezuela, who provided the necessary financial help in

the form of an Overseas Scholarship.

I sincerely thank Dr. W.A. Ayer of the Chemistry Department of the

University of Alberta for his cooperation in the performance of HR-100 nuclear

magnetic resonance and mass spectra in the early stages of this work.

I like to express my appreciation to my colleagues of the Organic

Chemistry Laboratories for their cooperation and helpful attitude during the

period spent on this investigation.

I wish to thank Mme M. Veilleux for accepting to undertake the task

of typing this thesis and for the excellent performance of it.

Thanks are due to the National Research Council of Canada for the help

provided in the form of Research Grants.

V

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

1 = "

2.-

3. -

4. -

5. -

6. -

7. -

8. -

9. -

10. -

11.-

12.-

13. -

14. -

15. -

16. -

17. -

18. -

19. -

20. -

21.-

22.-23. -

24. -

LIST OF FIGURES

Page

Nuclear magnetic resonance spectrum of yohimbine ...... 11

Mass spectrum of yohimbine ................ 13

Nuclear magnetic resonance spectrum of 0-acetylyohimbine . . 15

Interconversions of aspidodasycarpine derivatives ..... 24

Mass spectrum of des-O-methylaspidocarpine ......... 31

Nuclear magnetic resonance spectrum of diacetyl-des-O-methyl- asp ldo carp me 33

Mass spectrometric fragmentation of des-O-methylaspido-carp me .......................... 36

Mass spectrum of pyridine XXIX ............... 38

Mass spectrometric fragmentation of pyridine XXIX ..... 44

Nuclear magnetic resonance spectrum of fendleridine .... 48

Mass spectrum of fendleridine ............... 50

Nuclear magnetic resonance spectrum of fendlerine ..... 52

Mass spectrum of fendlerine ................ 54

Nuclear magnetic resonance spectrum of aspidofendlerine . » 58

Nuclear magnetic resonance spectrum and spin decoupling experiments on taberpsychine ................ 63

Mass spectrum of taberpsychine ............... 65

Nuclear magnetic resonance spectrum and decoupling experi­ments on dihydrotaberpsychine-methine ........... 67

Nuclear magnetic resonance spectrum of taberpsychine-methine . 72

Mass spectrum of taberpsychidine .............. 78

Nuclear magnetic resonance spectrum of dihydrotaberpsychidine 81

Mass spectrometric fragmentation of dihydrotaberpsychidine . 82

Mass spectrometric fragmentation of taberpsychidine .... 85

Mass spectrum of 16-epi-vobasinic acid ........... 86

Mass spectrometric fragmentation of 16-epi-vobasinic acid . 88

INTRODUCTION

The study of indole alkaloids offers not only very interesting

chemical problems but also the possibility that the isolated materials

have a certain therapeutical value.

The extensive work done in recent years on Apocynaceae species has

proved them to be an excellent source of this type of alkaloid^.

These considerations and the relatively easy access to the abundant

Venezuelan flora, prompted us to start the examination of the different

species of this family found in Venezuela. To date five species have been

2 3 5examined, namely Aspidosperma vargasi , A. fendleri , A. cuspa , A. excel­

sum ^ and Tabemaemontana psychotrifolia^ „

7Indole alkaloids have been isolated as early as 1841 when Goebel

obtained harmaline (1) from Peganum harmala L. (Rutaceae), but a really

explosive development occured in the last twenty years with the exploitation

of powerful physical methods such as nuclear magnetic resonance and mass

spectrometry.

1

2

The structures of indole alkaloids range from the bare indole itself

8 9(II), isolated from the flowers of many Jasminium and Citrus species, to

very complex polycyclic "dimers" such as Pleiomutine"*"** (III), obtained from

Pleiocarpa mutica"*""*".

IV

To give an idea of the power of nuclear magnetic resonance and mass

spectrometry in the study of the structures of alkaloids, we have chosen

as an example aspidospermine (IV) which embodies a particular pentacyclic

12skeleton characteristic of a well represented group of natural bases

Although nuclear magnetic resonance was only partly responsible for

the structure determination of aspidospermine and the mass spectrum of the

alkaloid was only measured after the structure was known, both physical

methods have had great application in the subsequent investigations of

3

related alkaloids.

The nuclear magnetic resonance spectrum helps to deduce the pattern

of the aromatic substitution from the absorption between 6.6 and 7.35,

especially when the data are compared with those obtained from the ultra­

violet spectrum since the latter changes a great deal depending on the

13substitution on the indolic residue . Of the non-aromatic protons in

aspidospermine, the one found at lowest field is the hydrogen atom at C.2

which appears as a quartet centered at 4.0-4.56. The absence of this peak

is indicative of substitution at C.2 and this provides very useful infor­

mation for the recognition of alkaloids such as aspidofractinine (V) where

C.2 is included in a sixth ring. In alkaloids where the C.2 proton is

absent but a carbomethoxy group is present at C.3, as for example refractine

(VI), a one proton quartet is observed at approximately 3.86 due to the

*hydrogen atom at C.3. The four protons adjacent to the basic nitrogen (at

positions 8 and 10) absorb in the region 2.9-3.36 and in all the alkaloids

As a convention, the nitrogen atom in the indolic residue (which as in aromatic amines has very little or no basicity at all) is referred to as Na and the second nitrogen atom present in the molecule is referred to as the basic nitrogen or N^.

4

of the aspidospermine type form what has been called the "finger print"

of the skeleton"*"^ ’ , a pattern which is not obscured by other absorption

and is profoundly altered in the spectrum of related hexacyclic bases.

The isolated proton at C.19 produces a singlet between 2.2 and 2.56

which in the case of alkaloids bearing an N&-acetyl group is sometimes

hidden under the absorption of the COCH^ but its presence can be deduced

from careful integration.

In addition to the absorptions mentioned, one finds frequently in the

aspidospermine group intense singlets at 3.75-3.906 due to the presence of

aromatic methoxyl groups and at 2.206 for the methyl groups of N-acetyl

residues. In the case of N-propionyl compounds a quartet can be observed

at 2.3-2.86 due to the protons of the methylene group which overlaps a little

with the absorption for the C.19 proton in 60-mc spectra. The methyl of the

propionyl group produces a triplet centered at approximately 1.256. There

is no possibility of confusion with the absorption for the terminal methyl

of the C-ethyl side chain, since the latter produces two principal peaks

centered at 0.656 separated by 5-6 cps. The absence of this absorption is

indicative of a substituent on the terminal carbon of the C-ethyl side chain

and this can be either an oxygenated function at C.21 as in the cases of

cylindrocarpine (VII) and limaspermine (VIII), or the. fact that C.20 and

C.21 are part of a sixth ring as in refractine (VI) or aspidoalbine (IX).

Other groups that can be recognized by means of the nuclear magnetic

resonance spectrum include phenolic hydroxyls whose presence can also be

observed by the change in the ultraviolet spectrum taken in neutral and then

in basic solution. Phenolic hydroxyls appear in this series generally at

5

position 17 where hydrogen bounding with the carbonyls of N-acyl groups is

possible, and in these cases the peaks for the OH protons appear at 10.7-

11.2Ô. When hydrogen bonding is absent, the absorption for the OH proton

is not at such low field, for example the hydroxyl at 0.16 in spegazzin-

idine (X) absorbs at 5.846.

The three proton singlet due to the methyl of a carbomethoxy function

appears in the region 3.55 to 3.706 at slightly higher field than those for

aromatic methoxyls. In some alkaloids of this group one finds a singlet at

about 9.56 due to the proton of an N-formyl residue.

6

The use of nuclear magnetic resonance spectrometry mainly for the

determination of the nature of peripheral groups has as its ideal comple­

ment mass spectrometry, which gives direct information about the skeletal

structure. In some cases complete elucidation of a structure has been

possible by determining the mass spectrum using less than one milligram

of material; thus it was possible to investigate successfully some minor

bases of Aspidosperma which were present in quantities too small to permit

even elementary analysis'^.

In the mass spectrometer, the molecule is cleaved by electron impact

and those fragments which are sufficiently stable and which carry a positive

charge are collected following the order of their molecular weights . The

ions bearing a double charge appear at m/2 where m represents the molecular

weight. The information that can be derived from the pattern of the rela­

tive intensities of peaks plotted against m/e (e representing the charge of

17the ion) can be divided in three parts .

First, the peak at highest molecular weight is generally"*^ that of the

molecular ion, this is the complete molecule with a single charge produced by

extrusion of one electron from the unshared pair of the basic nitrogen by

electron impact. This peak is referred to as M4" and is accompanied by smaller

peaks one and two units higher corresponding to molecular ions containing

heavier isotopes of nitrogen, hydrogen and carbon. Thus with the molecular

weight precisely determined the molecular formula is derived leaving no doubt

as to the correct number of hydrogen atoms present. Although this was also

possible by integration of the nuclear magnetic resonance spectrum, the quan­

tity required for the mass spectral method is minute in comparison.

7

Second, it is possible to classify an unknown alkaloid within a struc­

tural class provided that other alkaloids with the same skeleton are known.

The cleavage of the molecule is independent of many peripheral substituents.

In the aspidospermine group, it has been found that these include hydroxyls

or methoxyls on the aromatic ring, oxygenated functions in the C-ethyl side

chain, either at C.20 or at C.21, carbomethoxy or other one carbon groups

and hydroxyls at position 3, and carbonyl groups at C.4. Thus, the character­

istic fragmentation pattern of aspidospermine can be recognized in all alka­

loids containing the same skeleton, with the only differences being that the

peaks corresponding to the fragments containing extra substituents are dis­

placed to higher molecular weight, in a number of units equal to the molecular

weight of the substituent, while alkaloids not containing the methoxyl group

present in aspidospermine exhibit a pattern for the aromatic fragments at

correspondingly lower molecular weight.

Third and most important, direct structural information can be obtained.

It is often possible, especially when the method is used in conjuction with

NMR and deuterium labelling, to determine the exact location of substituents

and establish the nature of new skeletons without in many cases any extensive

chemical degradation. The importance of this is observed mainly in the case

of alkaloids that do not lend themselves to facile degradation by classical

methods„

The investigation of new plants in search of alkaloids is by nature an

unpredictable venture and the results fall into three different categories.

First of all, the extraction could yield only bases which have been described

previously and which are more or less readily identified by their physical

8

constants and spectral data. This was the outcome of the investigation of

the alkaloids of Aspidosperma excelsum and Aspidosperma cuspa, presented

in Chapter I.

Some investigations present the case where the isolated alkaloids are

previously unknown but their structures can be recognized by non-destructive

spectral methods and the bases may be readily related to known compounds

through minor structural modifications. An illustration of this eventuality

is found in the study of the bases obtained from Aspidosperma fendleri as

described in Chapter II„

Finally, some plants yield new alkaloids which despite the powerful

physical methods available resist structure elucidation until a suitable

chemical degradation is performed. This was the case with the major alka­

loid from Tabemaemontana psychotrifolia, where the presence of an oxide ring

of a type not encountered previously impeded the interpretation of both

spectral and chemical results for a long while. It was again nuclear magne­

tic resonance spectrometry and in particular the use of spin-spin decoupling

techniques on a Hofmann degradation product which finally provided the neces­

sary clue, allowing an un-ambiguous structure to be proposed for the alkaloid

(Chapter III),

-CHAPTER I. -

A. ALKALOIDS OF ASPIDOSPERMA EXCELSUM,-

10

Seeds and bark of the gigantic Venezuelan tree, Aspidosperma excelsum

IQ ^Benth. , were collected near Canaima , and both parts of the plant yielded

two known alkaloids in quantity.

The major constituent of the extracted crude base, a crystalline com­

pound ^22^26^2^3' Presents an ultraviolet spectrum typical of unsubstituted

20indoles . Its infrared spectrum shows the presence of the indolic imino

group as well as that of a hydroxyl, an ester carbonyl and the Bohlmann

21bands typical of trans-quinolizidines.

From the nuclear magnetic resonance spectrum (Figure 1) it was possible

to determine that the carbonyl is that of a methyl ester (3 H singlet 3.786)

and the hydroxyl group is a secondary one (only 1 H adjacent to oxygen, broad

singlet 4.226). Assembling of this data permits an approach to the structure

as presented by (i).

- OH

Cj, H(, N

COOCHj

l

* In the Southern-Venezuelan jungle.

é values 5 2,

Figure 1.- Nuclear magnetic resonance spectrum of yohimbine

12

The mass spectrum of the base shows the molecular ion at 354 m/e (thus

confirming the elemental analysis), as well as an M-59 peak corresponding to

the expulsion of the ester residue and other peaks identical with those ob­

served in the fragmentation of yohimbine-type alkaloids.

That the fragmentation observed in the mass spectrometer (Figure 2)

22was identical with that published for yohimbine must be interpreted with

caution since in this group of alkaloids the stereochemistry only influences

the fragmentation to a negligible extent. However, other characterizing phy-

20sical properties and the constancy of the melting point of the isolated base

when mixed with an authentic sample of yohimbine (XI) confirmed the identity

thus establishing the structure of the base.

OH

XI

The specific rotation and optical rotatory dispersion of the isolated

base showed that it is indeed yohimbine and not one of its several possible

naturally occuring stereoisomers.

relative intensity

N-H

M - CO, CH.

Figure 2.- Mass spectrum of yohimbine

14

The second base isolated from A. excelsum was obtained as the hydro­

chloride by cooling and concentrating the chloroform obtained during the

separation of neutral materials from the crude alkaloid-containing extract.

This solubility of hydrochlorides in chloroform, while at first surprising,

is frequently encountered and provides a direct and efficient method of

separating particular bases from mixtures.

The alkaloid, , is also a member of the indole alkaloids,

as was demonstrated by its ultraviolet spectrum. Its infrared spectrum shows

the presence of the indolic NH, an ester group (1730 cm-1) and an 0-acetyl

(1725 and 1250 cm-1).

The nuclear magnetic resonance spectrum (Figure 3) confirms the presence

of the indolic imino group (1 H singlet 7.876) and of four aromatic protons

(multiplet 7.00-7.608). It also identifies the ester as being a methyl ester

(3 H singlet 3.696) and the 0-acetyl as being secondary (only one H adjacent

to oxygen, multiplet 5.476).

All the spectral data for this base suggested its close relation to

yohimbine, except for the presence of the 0-acetyl residue. Hydrolysis of

the acetate function under acidic conditions known not to affect the methyl

ester grouping, gave a crystalline solid identical in all respects with

yohimbine.

The naturally occuring alkaloid, which was formulated as O-acetyl-

yohimbine, (XII), was prepared by acetylation of yohimbine, thus confirming

the structure.

7

CHjOOC

OAc

I4 6 VALUES

Figure 3Nuclear magnetic resonance spectrum of 0-acetylyohimbine

2, \

16

O COCK

XII

This represents the first recorded isolation of O-acetylyohimbine from

natural sources, and since the base was obtained in the extraction before

any potential acetylating agents had been in contact with the material, the

probability of its being artifactual is minimal.

Later work performed on the mother liquors of the extracts led to the

4 23isolation of excelsinine (10-methoxycorynanthine) (XIII) and ct-yohimbine

(XIV).

CI-LOOC

XIII XIV

-CHAPTER I. -

B. ALKALOIDS OF ASPIDOSPERMA CUSPA.-

18

24Aspidospema cuspa is found in the coastal regions of Venezuela

close to Caracas, as a large shrub in contrast to the enormous trees which

characterize the species A. fendleri and particularly A. excelsum which are

2included in this study and A. vargasi which has also been examined in this

survey of Venezuelan plants. The crude alkaloids were obtained from the

aerial bark by a routine mild extraction procedure and separated by counter-

current distribution. Four relatively stable alkaloids were isolated and

characterized in this manner and all four were previously described but origi­

nating from other species of Aspidosperma. A fifth base was isolated by

preparative paper chromatography on the crude mixture and a structure is pro­

posed even though the lack of sufficient material did not allow a complete

structure proof.

The major base of A. cuspa, ^22^26^2^4’ Presents ultraviolet spectrum

typical of the unsubstituted dihydroindole type of alkaloid. Its infrared

spectrum shows the presence of the aromatic portion and of the indolic imino

group, as well as a hydroxyl, another NH group and a rather complicated car­

bonyl absorption (3 bands at 1750, 1735 and 1720 cm-1) which was the origin

of some confusion but it was demonstrated later as being due to an ester car­

bonyl (1750 cm-1). The two other accompanying bands are probably due to

25Fermi resonance or to hydrogen bonded species.

The nuclear magnetic resonance spectrum identifies the ester grouping

as a methyl ester (3 H singlet 3.766). An important feature in the spectrum

is a quartet for one olefinic proton (5.526, J=6.5 cps) coupled to a three pro­

ton doublet of doublets (1.766, J=6.5 and 2 cps) which is indicative of an

19

ethylidene side chain. The nature of one of the oxygen functions remained

unknown and by analogy with many other Aspidosperma alkaloids it was assumed

to be ethereal.

With this data, the partial structure (ii) could be advanced for the

base.

-COOCH3

xd

H

— ON

H 1%, ~ N H

W

CH%

11

Pyrolysis of the base in vacuum in the presence of zinc powder pro­

duced 3-ethylpyridine as the major volatile product as shown by direct vapor

phase chromatographical comparison and by the nuclear magnetic resonance

spectrum of the principal volatile fraction. This pyridine, or more speci­

fically the corresponding pyridinium ion, is also observed as the base peak

in the mass spectrum of the A. cuspa alkaloid (108 m/e). A much smaller quan­

tity of 3- or 4-methylpyridine was formed in the zinc pyrolysis but the yield

was too low to obtain a clear nuclear magnetic resonance spectrum and unfor­

tunately the retention times of the two methyl-pyridines differ by too little

to be conclusive.

This major alkaloid has now been shown to be aspidodasycarpine

20

which appears in the literature in a preliminary communication^. Original­

ly, this identification was considered and rejected due to an ostensibly

different melting point of the 0,N-diacetyl derivative (XVI) as described in

the literature (m.p. 110°) and that obtained from the plant (irup, 175°). How­

ever, the constancy of the melting point of our aspidodasycarpine when mixed

with an authentic sample*, superposable infrared spectra and identical thin

layer chromatography results conclusively showed the A. cuspa base to be

aspidodasycarpine.

HOCH.

Our original uncertainty led us to prepare other derivatives, thinking

perhaps a structure elucidation would be necessary.

As suggested above, normal acetylation of aspidodasycarpine affords

the 0,N-diacetyl derivative (XVI) but the neutral N-acetyl derivative (XVII)

was prepared by carefully reacting the base with one mole of acetic anhydride

in pyridine at low temperature and the same product (XVII) was obtained by

We thank Dr. C. Djerassi for providing an authentic sample of aspidodasycarpine.

hydrolysis of the 0,N-diacetyl derivative (XVI) in dry methanol containing

a small amount of sulfuric acid.

21

Catalytic reduction of aspidodasycarpine in ethanol containing acetic

acid gave rise to a dihydro compound (XVIII) in which the exocyclic double

bond is reduced. The configuration of the newly formed asymétrie centre is

assumed to be as shown in XVIII with the C-ethyl residue cis to the ester

bearing bridge since examination of models reveals considerable hindrance

to the approach of a hydrogen bearing catalyst from the same side of the

molecule as the bridge. The nuclear magnetic resonance spectrum of the di­

hydro derivative confirms the saturation of the double bond since the peaks

of the ethylidene residue in aspidodasycarpine (1 H quartet 5.526, J=6„5 cps

and 3 H doublet of doublets 1.76, J=6.5 and 2 cps) are no longer observed

and a resonance for a saturated methyl group is now found at 1.06 (doublet

J-6 cps).

Acetylation of the dihydro base (XVIII) afforded the expected 0,N-

diacetyldihydro derivative (XIX) which proved to be a remarkably insoluble

but clearly crystalline substance. The latter could also be prepared by

hydrogenation of 0,N-diacetylaspidodasycarpine (XVI).

HOCH COOCH

XVIII

R<*OCH. COOCH

Hydrolysis of the diacetyldihydro compound (XIX) using the conditions

described above which avoid the hydrolysis of the methyl ester group, led

to N-acetyldihydroaspidodasycarpine (XX) which again was more readily pre­

pared by partial acetylation of the dihydro base (XVIII), Catalytic reduc­

tion of N-acetylaspidodasycarpine (XVII) also afforded the N-acetyldihydro

compound XX.

HOCH COOCH HOCH COOCH

1 mole Aco0

XVIII XX

23

The spectral characteristics of these various acetyl derivatives are in

agreement with their- proposed functionality (see Experimental). These inter­

conversions can be summarized as shown in Figure 4.

The carbinolamine ether structure of aspidodasycarpine suggests that

in acid solution the five-membered ethereal ring should open, affording the

indolenium structure XXI„

HOCH GOOCH

Indeed Djerassi used this protonated form of the alkaloid to explain

the degradation products isolated following reduction with zinc in hydro­

chloric acid^„ We felt it would be of interest to characterize the in-

dolenine, the acetyl derivative of which we found to be readily obtained by

acetylation of dihydroaspidodasycarpine (XVIII), or its acetyl derivatives

XIX and XX, in acetic anhydride containing small quantities of concentrated

sulfuric acid. The product (XXII) absorbs in the ultraviolet spectrum at

222 and 261 my reflecting the expected change in chromophore and the infra­

red and nuclear magnetic resonance spectra now show the absence of the in-

doline NH moiety. All attempts to hydrogenate the indolenine double bond

met with failure, however reduction with sodium borohydride afforded N-acetyl-

24

GOOCH.HOCH.

HOCK COOCH AcOO-L, GOOCH

COOCHHOCH

COOCHHOCH. GOOCH

XVIII

XXII

Figure 4.- a) 1 mole acetic anhydride in pyridine, 5°; b) Acetic

anhydride-pyridine ; c) Sulfuric acid-methanol ; d) PtC^, Hg, EtOH, HOAc

e) Acetic anhydride-sulfuric acid; f) NaBH^ - methanol.

25

dihydroaspidodasycarpine (XX) showing that the ethereal ring is very readily

re-established (Figure 4).

flcOCH COOCH

XXII

The mass spectra of aspidodasycarpine (XV), the dihydro base (XVIII)

and the indolenine (XXII) deserve some comment although no detailed analysis

of the fragmentations will be presented due to the lack of suitable deuter-

ated substrates «

Fragmentation of aspidodasycarpine is characterized by an intense peak

at 108 m/e presumably arising from the ethylpyridinium ion a. The only other

peak of comparable intensity is at 14 mass units lower„ The dihydro deriva­

tive (XVIII) also shows a tendency to produce similar stable ions but as ex­

pected the base peak is now at 110 m/e (b) with a much smaller peak at 108

m/e (a). Relatively pronounced peaks also appear at 130 and 144 m/e (from

the indole moiety).

The indolenine (XXII) also shows the dihydropyridinium ion (b) but as

a relatively weak peak while the major fragments arise from the loss of the

various oxygenated groupings and side chains = The cleavage of the acetyl

26

groups (as ketene) and the ester function is reflected by the M-42 and M-59

peaks respectively but the most intense peak by far arises from the expulsion

of -CHg-O-CO-CH^. The other alkyl chain -CH^-CH^-O-CO-CH^ is also extruded

as judged by the significant peak at M-87. One other intense peak at 385 m/e

(M-113) is postulated as arising from rupture of the piperidine ring and

release of the C-ethyl residue and carbon atoms C.20 and C.21 with the nitro­

gen atom and the accompanying acetyl residue (heavy lines in XXII)„

HI

a (108 m/e) b (110 m/e)

The second alkaloid obtained from A. cuspa analysed for CgiHL.NJO.

and also belongs to the unsubstituted dihydroindole group of alkaloids

(Xmax 233 (10,000) and 283 (3,700) mp) and its ultraviolet spectrum changes

ostensibly when taken using 70% perchloric acid as solvent 274.5

and 302 mp). This is typical for carbinolamine-ethers due to rupture of the

27ether with formation of a quaternary indolenium ion „

The infrared spectrum confirms the presence of the indolic imino group

and the aromatic residue, and reveals the presence of an ester grouping (1747

cm-1) and a hydroxyl (3550 cm-1). In this case, as opposed to aspidodasy-

carpine, the basic nitrogen seems to be tertiary.

27

The nuclear magnetic resonance spectrum provides information which in

time proved to be decisive for the assignment of structure. A three proton

singlet at 3,606 serves to identify the ester residue as a methyl ester and

a four proton multiplet (centered at 7.066) accompanied by a one proton

singlet at 5.166 account for the dihydroindole residue. A quartet (5.386,

J=7.5 cps) ascribed to an olefinic proton coupled to a doublet for three

protons (1,566, J=7.5 cps) indicates the presence of an ethylidene side chain.

A very important feature is a doublet for one proton at 4.756 (J=2.5 cps),

which considering the extent of deshielding should be adjacent to both oxygen

and nitrogen.

This information is summarized in the partial structure (iii) :

IH

Cc, N

COOCH-

OH

KCH„

in

The mass spectrometric fragmentation corroborates the analytical fi­

gures (M+=368). Other peaks are those at 350 (M-18), 337 (M-31) and 239

(M-129) m/e.

All the evidence obtained on this base pointed to a known alkaloid,

des-acetylpicraline (XXIII), first isolated by Britten et al. from Picralima

28

HOCH GOOCH,

N

H

XXIII

Of great help in the characterization of the base is the absorption

in the nuclear magnetic resonance spectrum at 4,755 mentioned earlier but

now attribuable to the lone proton at C.5 which is flanked on one side by

the more basic nitrogen atom and on the other side by the ethereal oxygen

atom. The infrared and ultraviolet spectra are in agreement with those

given in the literature and a mixed melting point with an authentic sample*

showed no depression, thus confirming the identity.

The third alkaloid obtained from A. cuspa belongs to a different group

of Aspidosperma bases and this was evident at an early stage. Absorption in

the ultraviolet spectrum showed the aromatic portion of the molecule to be a

hydroxylated N-acyl indoline. The N-acyl carbonyl group which produces an

intense peak at 1635 .cm-1 in the infrared spectrum is presumably strongly

We thank Dr. W.I. Taylor for kindly providing an authentic sample of des-acetylpicraline.

29

hydrogen bonded and the spectrum also shows an unbonded hydroxyl at 3360 cm-1.

Despite many attempts to dry the sample a peak of variable intensity at 1715

cm-1 persisted in the infrared spectrum. This peak was felt to be due to

acetone of crystallization and this was indeed shown to be the case by analy­

sis and mass spectral determination of the molecular weight. Elemental analy­

sis of the base although somewhat variable had indicated after several deter­

minations a molecular formula of (mol.wt. 414) but the mass spectrum

showed the molecular ion to be at 356 m/e corresponding to

difference between the two formulae being essentially the elements of acetone*.

The possibility of a facile fragmentation in the mass spectrometer could not

be ruled out so the nuclear magnetic resonance spectrum was examined and the

expected peak arising from the methyl groups of a molecule of acetone was

observed at 2.145. The sample was evaporated to dryness and redissolved in

deuterioacetone and then taken to dryness again. This procedure was repeated

twice and the nuclear magnetic resonance spectrum which was then taken re­

vealed the absence of the 2.145 peak. Other features of the nuclear magnetic

resonance spectrum included absorption arising from a saturated C-methyl

grouping which gave a doublet at 0.75 (J=6 cps), an N-acetyl three proton

singlet at 2.255, a complex pattern around 2.805 reminiscent of that con­

sidered diagnostic of the aspidospermine skeleton^, a quartet centered at

4.005 also observed in aspidospermine type bases (C-H adjacent to the in-

do line nitrogen atom), a single peak at 6.365 integrating for two aromatic

protons and finally a low field singlet at 10.855 attributed to the chelated

phenolic hydroxyl group, The latter disappeared from the spectrum on shaking

* The formula was obtained by analysis after several weeks

of drying - see Experimental.

30

with deuterium oxide.

All the evidence accumulated was consistent with the formulation of this

base as a dihydroxy-N-acetyl derivative of aspidospermidine (XXIV) and the

mass spectral fragmentation (Figure 5) was typical for this type of skeleton.

H

XXIV

The loss of a fragment of mass 28 (ethylene) is now well established as the

first step in the fragmentation of aspidospermine derivatives and the peak

at M-43 is in keeping with the presence of an acetyl residue. The principal

indole ions appear at 190, 176 and 162 m/e as opposed to 144 and 130 m/e

in alkaloids bearing no substituents in the indoline moiety which confirms

the placing of the two hydroxyls groups in the aromatic ring. The remainder

of the molecule gives rise to significant peaks at 152 and 138 m/e with by

far the most intense peak at 124 m/e. These three peaks and especially the

29latter are found in the spectra of all bases embodying this skeleton . To

fully describe the structure of the alkaloid the substitution pattern on

the aromatic moiety needed clarification. The confusion arising from the

apparent singlet for the two aromatic protons in the nuclear magnetic reso­

nance spectrum of the base was cleared up by acetylating the two phenolic

relative intensity

Kl-0-9M-28 M

3X

_____L___ 17->

_lL_350

!300 250 m/e

Figure 5„~ Mass spectrum of des-O-methylaspidocarpine

190

200

04

32

hydroxyls with acetic anhydride in pyridine and the spectrum of the acetate

showed not only the signal for two acetate residues but also a clear AB quar­

tet for the aromatic protons (6.99 and 6.886, J=8.5 cps) (Figure 6).

This limits the possible structures for the base to XXV and XXVI,

XXV XXVI

The problem of assignment of structure was solved by methylating the

base by prolonged treatment with diazomethane. The dimethyl derivative

30obtained (XXVII) is known by the name of pyrifolidine and direct compari­

son with an authentic sample* proved the identity. In this way the struc­

ture for the base is established as being as shown in formula XXV which

is in fact O-des-methylaspidocarpine, The latter has been prepared from

31aspidocarpine and isolated later from A. album as one of the very minor

, 32bases

In the light of this structure the mass spectral fragmentation shown

We thank Dr. C. Djerassi for kindly providing an authentic sample of pyrifolidine.

? 4 3 é VALUES %

Figure 6Nuclear magnetic resonance spectrum of diacetyl-des-O-methylaspidocarpine

w

34

in Figure 7 explains all the principal peaks observed.

CI-LO

XXVII

A fourth alkaloid isolated from A. cuspa is a very minor constituent

obtained from a countercurrent distribution of the weaker bases. Its ultra­

violet spectrum (Xmax 230 and 282 my) is that of an unsubstituted dihydro-

indole and suffers a bathochromic shift when measured in concentrated acid

solution 240, 247 and 315 my). This shift is characteristic ofÏÏlcLX.

27carbinolamine ethers ~ where rupture affords indo lenium ions.

The infrared spectrum confirms the presence of the indolic imino group

and shows the presence of a very strong carbonyl peak at 1745 cm-1 accompanied

by a strong band at 1240 cm-1 attributed to an 0-acetyl group. The oversized

carbonyl peak could be due to the presence of another ester-type carbonyl in

the molecule, whose absorption falls in the same region as that of the 0-ace-

tyl group.

The mass spectrum of the base is very similar to that of des-acetyl-

picraline (XXIII) except for the molecular ion peak which appears at 410 m/e

35

(42 units up from that of des-acetylpicraline). On the basis of its mass

spectrometric fragmentation the structure of this alkaloid has been proposed

as XXVIII which is that of picraline, first isolated by Thomas from Picra-

33lima klaineana .

GOOCH.

XXVIII

In effect the fragmentation of our base in the mass spectrometer is

identical to that given in the literature for picraline^. The extremely

small quantity of alkaloid obtained did not permit a more extensive confir­

mation of this speculation.

A fifth product was obtained from the crude mixture of bases by pre­

parative paper chromatography. This base gives a green colour with Vassler's

35reagent (which made its isolation very easy to follow) and is present in

the plant in very small quantities. Whatman paper N° 31 (double thick, 3 mm)

was used and the crude mixture applied in 30 spots per sheet of paper. After

the chromatogram had been developed (solvent: pyridine-ethyl acetate-water,

2.3 : 7.5 : 1.65)^ one strip was cut off the border and sprayed with

Vassler's reagent. The sheet was then divided horizontally following the

36

152, 138

Figure 7Mass spectrométrie fragmentation of des-O-methylaspidocarpine

37

distribution shown by the reagent. The band containing the green spot was

eluted with ethanol and then the solvent evaporated to dryness. The residue

was sublimed and the alkaloid obtained as low melting white crystals.

The ultraviolet spectrum C^mæc 255 and 262 my) is typical of a simple

pyridine and in keeping with this observation the infrared spectrum shows

only typical pyridine peaks (1600 and 1565 cm-1) and aromaticity (725 cm-1).

The mass spectrum of the compound (Figure 8) shows the molecular ion

peak at 149 m/e, which being an odd number indicates that the compound must

contain at least one nitrogen atom. The base peak (very intense) appeared

at 120 m/e.

Before following the discussion, we must point out that since the

amount of material obtained was very small, neither nuclear magnetic reso­

nance spectrum nor elemental analysis were performed on this compound and

the speculation below is based on the limited information available, es­

pecially on the mass spectrometric fragmentation very particular of pyri-

dines.

The data from the ultraviolet and infrared spectra point to a pyridine

and the mass spectrum definitely confirms this with the appearance of a peak

accounting for the loss of HCN (27 units; see Figure 8 : from 106 to 79 m/e)

which is only found in nitrogen containing hetero-aromatic compounds, Since

the molecular ion peak appears at 149 m/e and all the evidence excludes the

presence of oxygenated groups in the molecule (the presence of an ethereal

moiety would lead to structures that could not possibly account for the frag­

mentation observed), this pyridine must have substituents accounting for five

100-

Figure 8,- Mass spectrum of pyridine XXIX

04CO

extra carbon atoms. These conditions could be met by several pyridines but,

on the basis of the mass spectrum, structures XXIX and XXX are favored.

39

XXIX

The placing of a side chain of at least three carbon atoms (in line)

37in the 3-position is in keeping with the view that in pyridine the electron

density is relatively high at the 3(5)-position but low at the 2(6)- and 4-

positions, thus making more stable the carbonium ions formed adjacent to the

ring at the 3(5)-position. In our case, by far the most favored fragment is

that at 120 m/e (M-29) which can be obtained from either one of the possible

structures by loss of and formation of carbonium ions such as (iv) and

(vi) which are stabilized as explained above and by ring expansion to give

the respective aza-tropilium ions (v) and (vii). \

Separate observations in the mass spectra of 4-propylpyridine and 2-

ethylpyridine show clearly that the ions formed in those cases by loss of

.C2H5 and .OHL respectively (affording viii and ix) are very much less stable

(as judged by the peak intensity) than that obtained in the case of 3-ethyl-

pyridine by loss of .CH_ (x).

40

For the location of the other substituent on the ring, several consi­

derations have been taken in account. First of all, the already mentioned

loss of HCN from the molecule is a clear indication that at least one of the

positions adjacent to the nitrogen must be free. Furthermore, if one consi­

ders the case with propyl and ethyl substituents (pyridine XXIX), the ethyl

41

chain could not be located in a position adjacent to the nitrogen since the

mass spectrum of a model compound (i.e. 2-ethylpyridine) shows that the M-15

peak is insignificant, while in our mass spectrum the M-15 (134 m/e) is a

fair-sized peak. Neither could it be in the 3(5)-position, in which case

the M-15 would be bigger than it appears in the spectrum of the unknown ma­

terial .

Considering a sec-butyl and methyl substitution (as in pyridine XXX)

the case is not so clear cut since the M-15 peak could be produced by loss

of .CHj from the sec-butyl chain producing a well stabilized ion.

However, pyridine XXIX appears to be the more favoured if one considers

the biogenetic point of view. It is now well accepted that the Woodward

38fission unit (xi) is the biogenetic precursor (together with tryptamine)

of a great number of indole alkaloids. This fission unit has been explained

as derived from a polyacetate chain and the ringed carbon atoms in (xi) repre­

sent those derived from the carbonyl groups.

©

I

I I

^© C

xi

©

i i

©"^© c

XI1

Later developments led to the discovery that the Woodward fission unit

is probably derived from a condensed chain of three acetate units and one

39malonate unit (xii) . This new aspect made it applicable to an even greater

number of indole alkaloids. For illustration we can use two apparently very

different alkaloids, vobasine (XXXI) and yohimbine (XXXII) where the strong

lines indicate the acetate-malonate chain.

42

COOCH

XXXI

CI-LOOC

XXXII

In the case of vobasine, one of the carboxyl groups of the malonate

residue has been lost leaving the fragment as the original tetraacetate unit.

In both structures the extra atom marked (B) is the so called berberine bridge,

present in all the alkaloids showing the "cryptic" unit and explained by

Robinson^ as originating from formaldehyde or its biochemical equivalent.

Comparing the tri-acetate-malonate unit (xii) with the fragment from

which pyridine XXIX (xiii) might originate one observes that in fragment

(xiii) the acetate and malonate units are arranged in a different fashion.

xiiiXll

43

However, this does not invalidate the hypothesis since, for example in

vincadifformine (XXXIII), a fragment (xiv) with a different arrangement is

observed ^1.

XXXIII xiv

Regarding pyridine XXIX as the possible structure, the mass spectro-

42metric fragmentation shown in Figure 9 accounts for all the important peaks

in the mass spectrum.

Here then, we have presented a series of facts consistent with struc­

ture XXIX for the pyridine in discussion but, by no means intended to be pre­

sented as categorical proof of the structure of this minor alkaloid of A.

cuspa.

XXIX

44

Figure 9-- Mass spectrométrie fragmentation of pyridine XXIX

45

The possible biogenetical importance of this compound has rendered it

more interesting than originally anticipated but further research is neces­

sarily delayed until fresh plant material becomes available and a new quan­

tity of the pyridine is obtained.

-.CHAPTER "II.-

ALKALOIDS OF ASPIDOSPERMA FENDLERI

47

Extraction of the seeds of A. fendleri Woodson^ yielded four alkaloids,

of which fendlerine, is the most abundant ; this alkaloid is also

found in large quantities in the trunk and root bark.

Fendleridine, is the simplest of the bases present and, on

the basis of its ultraviolet spectrum, belongs to the unsubstituted dihydro­

indole group of alkaloids.

The infrared spectrum reveals peaks characterizing the dihydroindole

imino group and the aromatic ring but gives no information concerning the

nature of the oxygen atom, which is probably ethereal.

The nuclear magnetic resonance spectrum (Figure 10) confirms the pre­

sence of four aromatic protons (multiplet 6.55 to 7.505), but absorption typi­

cal of olefinic protons and of methyl groupings of any sort is absent. A two

proton quartet (centered at 3.985), which has also been observed in other di­

hydro indo lie Aspidosperma alkaloids, has been shown to be characteristic of

the -O-CHg- grouping as found in aspidoalbine^ (XXXIVA) and aspidolimidine^

(XXXV).

CH,0

XXXIV A R=H, RicCOCHgCHg

XXXIV B R=CH3, Ri=H XXXV

IH

Figure 10.- Nuclear magnetic resonance spectrum' of fendleridine

00

49

This region of the nuclear magnetic resonance spectrum of fendleri-

dine is strikingly similar to that of O-methyl-N-despropionylaspidoaTbine

(XXXIVB), except for the methoxyl bands in the latter^. A plausible struc­

ture for fendleridine is that shown by XXXVI ; this is supported by the mass

spectrometric fragmentation pattern (Figure 11).

XXXVI

In the mass spectrum, confirmation of the analytical figures was shown

by the molecular ion peak at 296 m/e; the typical aspidospermine type M-28

47 48peak ’ (a in Figure 11) arising from the loss of ethylene is also present.

Rupture of the C.10-C.11 bond (at x in a) gives rise to fragment b, which is

responsible for the most intense peak at 138 m/e. This fragment has also

been observed in the fragmentation of aspidoalbine^ (XXXIV) and aspidoli-

midine^ (XXXV). The indolic portion of the molecule gives rise to weaker

peaks at 144 m/e (rupture at y) and 130 m/e (rupture at x). The other sig­

nificant peaks in the mass spectrum arise from expulsion of the oxide bridge

(intense peak at 252 m/e) and subsequent fragmentation as described above.

The analogy with aspidoalbine (XXXIV), aspidolimidine (XXXV) and haplo-

cine^g (XXXVII), and the other spectral evidence confirms the structure of

fendleridine (XXXVI) which is then the simplest representative of this hexa-

Figure 11.- Mass spectrum of fendleridine

180 140 IOO

51

cyclic type of Aspidosperma alkaloid.

CH3

XXXVII

The major base present in A. fendleri which we named fendlerine,

^23^30^2^4’ s^ows ultraviolet absorption attributable to a phenolic N-acyl-

dihydroindole and, as expected, the spectrum changes in basic solution.

The infrared spectrum confirms the presence of the N-acyl residue (in­

tense band 1635 cm-1), which is presumably strongly hydrogen bonded to the

phenolic hydroxyl since the amide peak is at longer wavelength than expected

for the simple N-acyl dihydroindole system^5^ and virtually no absorption

is observed in the hydroxyl stretching region.

From the nuclear magnetic resonance spectrum (Figure 12) the N-acyl

grouping was shown to be a propionyl residue (triplet and quartet centered

at 1.246 and 2.536 respectively); this was confirmed by acidic hydrolysis

of fendlerine and vapor phase chromatography of the volatile acid. Only

propionic acid was detected. A one proton peak at 12.426 confirms the sus­

picion that the phenolic hydroxyl is hydrogen bonded with the N-acyl residue.

The partial structure (xv) can then be written for the alkaloid.

4 3 <f VALUES

Figure 12,- Nuclear magnetic resonance spectrum of fendlerine

Ln

53

NOH

xv

Other peaks in the nuclear magnetic resonance spectrum include a quartet

for two aromatic protons (centered at 6.896) and one methoxyl residue (singlet

3.846). Since from spectral evidence the nature of one of the oxygen func­

tions remained obscure, the probability that fendlerine contained an ethereal

moiety (other than the methoxyl) was a tempting assumption. This hypothesis

was to some extent strengthened by the appearance of a peak in the hydroxyl

region of the infrared spectrum of the product obtained by hydrogenating the

base over platinum on charcoal. The partial structure can then be extended

to (xvi).

Under the conditions employed, most of the peaks in the mass spectrum

of fendlerine (Figure 13) were of low intensity, but the most important peak

was at 138 m/e (fragment b, Figure 11) as in the case of fendleridine (XXXVI).

The analytical figures are confirmed by the molecular ion peak at 398

m/e; other peaks of importance are those associated with the loss of the N-

propionyl residue. The most prominent indolic fragment gives rise to a peak

at 161 m/e which can only be related to a dihydroxy derivative of fragment c

(Figure 11), and this serves to localize the methoxyl residue on the aromatic

relative intensity

COCHCH

%

3

325

50 -

3?0

338342

~T—350

ilL300400 m/e

Figure 15Mass spectrum of fendlerine

/38

â

no

too

Ul4*»

55

ring of the indole moiety.

OH 'O^CH^CHg

~ H (arom.)

N— OCH^

xvi

At this point, two structures (XXXVIII and XXXIX] are plausible for

fendlerine since the nuclear magnetic resonance spectrum indicates that the

two aromatic protons must be ortho to one another (quartet 6.896, J=8 cps).

Both structures are in all respects compatible with the spectral data of

fendlerine.

The location of the methoxyl residue was determined by conversion of

the base into aspidolimidine (XXXV). Acid hydrolysis of fendlerine gave the

des-N-propionyl derivative, which could not be crystallized but which was

acetylated with acetyl chloride in pyridine. The product, a mixture of the

N-acetyl and 0,N-diacetyl derivatives, was hydrolyzed in methanolic sodium

hydroxide under nitrogen to give the pure aspidolimidine (XXXV). That the

structure XXXVIII represents fendlerine is thus un-equivocally established.

Since the number of bases with this same hexacyclic skeleton is in­

creasing, it would seem logical to name them as derivatives of the simplest

member. Preference is given to the name aspidoalbidine, which has already

52been used to describe this system . Thus fendlerine is N-propionyl-16-

methoxy-17-hydroxyaspidoalbidine (XXXVIII).

56

Ct-LO

CH, O

ch3 ch3

XXXVIII XXIX

The small amount of a high melting base, aspidofendlerine (m.p. 278°

(decomp.) was easily purified by fractional sublimation; analysis indicated

the molecular formula ^21^26^04* The ultraviolet spectrum is similar to

that of fendlerine (XXVIII), and changes in ethanol containing alkali and in

the presence of buffered boric acid, the latter suggesting the presence of

31a catechol residue

The infrared spectrum shows the presence of hydroxyl and N-acyl func­

tions.

Peaks in the nuclear magnetic resonance spectrum (Figure 14) are inter­

preted as showing that aspidofendlerine contains both bonded and unbonded

phenolic hydroxyl groups (singlets 11.06 and 5.795 respectively), two adja­

cent aromatic protons (quartet centered at 6.845), and an N-acetyl residue

(singlet 2.305).

In this case, there was again one of the oxygen functions whose nature

remained unknown and as in the other examples it was assumed to be ethereal.

This supposition was justified since the mass spectrum of the base gave peaks

57

of extremely low intensity with the exception of the dominant peak at 138 m/e

(fragment b in Figure 11), showing that this base is also of the fendleridine

(XXXVI) skeletal type.

On a small scale, aspidofendlerine, assumed to be N-acetyl-16,17-di-

hydroxyaspidoalbidine (XL), was treated with diazomethane ; aspidol imidine

(XXXV) was isolated as the major product, thus confirming the structure.

XL

The fourth base obtained from the extraction appears to be unrelated

to the three already described. Absorption in the ultraviolet is similar to

that of a dihydroindole, and the infrared spectrum suggests the presence of

an ester residue. The mass spectrum is unlike that of a member of the hexa-

cyclic aspidoalbidine series.

Alkaloids of the trunk bark of Aspidosperma fendleri Woodson.-

Extraction and isolation of the alkaloids of the trunk bark yielded

almost exclusively fendlerine accompanied by smaller amounts of aspido-

limidine and fendleridine.

Alkaloids of the root bark of Aspidosperma fendleri Woodson.-

Extraction of the root bark has yielded, apart from large quantities

T“ / 3à values

Figure l4e- Nuclear magnetic resonance spectrum of aspi^ofendlerine

inCO

59

of fendlerine and aspidolimidine, three minor alkaloids, none of which seems

to be related to the aspidoalbidine type of alkaloids and, to date, their

structures remain unknown and will be discussed elsewhere.

-.CHAPTER III.-

ALKÀLOIDS OF TABERNAEMONTANA PSYCHOTRIFOLIA

61

The extraction of the alkaloids of Tabernaemontana (Ervatamia) Psycho-

trifolia H.B.K.* yielded four principal alkaloids but the abundance of the

major base taberpsychine, and the fact that it could not be readi­

ly placed in a "skeletal" group by mass spectrometry made it the principal

area of interest.

The ultraviolet spectrum of taberpsychine shows absorption typical of

20the indole moiety , with no substituents other than the "normal" ones at

position two and three.

The infrared spectrum confirms the presence of the aromatic nucleus and

of the indolic NH, but the absence of absorption due to hydroxyl or carbonyl

groups fails to provide information concerning the nature of the oxygen atom

present. Since neither acetylation nor mild oxidation reactions produced

any change in the molecule the oxygen atom was, by exclusion assumed to be

ethereal. With this information, the partial structure (xvii) can be written

where the C^ residue could contain up to four rings.

H|q N O

xvii

* Collection of G. Agostini at the Botanical Garden of the Universidad Central de Venezuela, Caracas.

62

The nuclear magnetic resonance spectrum gives useful evidence about the

other half of the molecule, including the functionality of the second nitro­

gen atom. A broad quartet at 5.385 (J=6.5 cps) integrating for one proton,

accompanied by a doublet of doublets for three protons at 1.685 (J=6.5 and

2.0 cps respectively) is clear evidence for an ethylidene side chain. De­

coupling experiments (figure 15) left no doubt about this and at the same time

localised the origin of the small coupling observed in the methyl signal (and

also of the broadening of the quartet for the olefinic proton). In effect,

while irradiating the methyl signal, one finds that the broad lower field part

(H in Figure 15) of an AB quartet (3.605 and 2.935, J=14 cps) becomes a sharp

doublet ; while irradiating this broad doublet (3.605) the disappearance of

the small splitting of the methyl group signal is observed (this also sharpens

the quartet for the olefinic proton). This AB quartet is found in a region

typical for protons at carbon atoms adjacent to nitrogen, permitting the

placing of the ethylidene side chain in the sequence (xviii), with the methyl­

ene group adjacent to the basic nitrogen atom responsible for the AB quartet

discussed above.

H

CH3

The sharp singlet at 2.535 integrating for three protons must be

assigned to a methyl group on the basic nitrogen (since the indolic nitrogen

j values

Figure 15 »- Nuclear magnetic resonance spectrum and spin decoupling experiments on t ab e rpsychine

64

does not bear a substituent). The spectrum also confirms the presence of

four aromatic protons (4H multiplet 7.07-7.666) and the indolic NH (singlet

8.426).

A very important feature in the nuclear magnetic resonance spectrum

is a pair of doublets at 5.136 (J=10 and 2 cps) integrating for one proton.

This signal is only ascribable to a proton adjacent both to an aromatic

moiety and an oxygen atom. The splitting pattern suggests that it is also

adjacent to a methylene. This information helps to reduce the possible lo­

cation of the oxygen atom to one of the two carbons linked to the indole re­

sidue. However, as yet, there is not enough evidence to decide which.

Zinc dust distillation of taberpsychine, following a small scale method

53described by Biemann coupled with gas-chromatographic analysis of the vola­

tile distillate, furnished very important data. The major product was iden­

tified, by direct comparison in the gas chromatograph with an authentic sample

and by nuclear magnetic resonance spectrometry of the material recovered, as

being 3-ethylpyridine. The product could only be formed from the part of the

molecule containing the basic nitrogen. This evidence, coupled to the fact

that the analogous pyridinium ion is observed in the fragmentation in the mass

spectrum (107 m/e) as one of the major peaks (Figure 16), permitted us to con­

clude that the basic nitrogen atom is located in a six-membered ring.

Now the partial structure can be extended to (xix).

Hydrogenation of taberpsychine afforded a dihydro derivative,

in which the ethylidene side chain was no longer present as shown by the

nuclear magnetic resonance spectrum, while a signal due to a methyl group on

relative intensity

100-

122

50—

I0>

154

130

ISO 200 m/e

Figure 16„- Mass spectrum*ef talerpsychine

M* (x.308

5)

(H-15) 233

250

Ü

300

CT\U1

66

a saturated carbon appeared as a triplet at 0.986 (J=7 cps).

ÇH3

xix

The methiodide of the dihydro derivative undergoes Hofmann degradation

with potassium tert—>but oxide to give a major product, whose ultra­

violet spectrum indicates the presence of a double bond conjugated to the in­

dole chromophore^. The presence of this new double bond is corroborated

by the nuclear magnetic resonance spectrum of the dihydrotaberpsychine-methine,

in which one finds the signals for two olefinic protons at 6.696 (doublet

J=12 cps) and 5.556 (doublet of doublets J=12 and 8 cps) as the low field AB

part of an ABX system (see Figure 17). The fact that the double bond intro­

duced during this reaction is conjugated with the aromatic moiety and the

assumption of a very likely tryptamine biogenesis allows the linkage of the

two units presented in partial structure (xix) and the placing of the oxygen

atom on the carbon attached to position 2 of the indolic portion as shown in

(xx) .

The methylene group adjacent to the carbon bearing the oxygen atom is

explained if one remembers the pattern for the proton adjacent to both the

aromatic residue and oxygen (doublet of doublets J=10 and 2 cps),

values

Figure IT.- Nuclear magnetic resonance spectrum and decoupling experiments on dihydrotaberpsychine-methine

o\

68

Spin decoupling experiments (Figure 17) on dihydrotaberpsychine-methine

provided precious information concerning the missing linkage of the oxygen

atom. In effect, it demonstrated that the X part (1 H) of the ABX system in­

volving the olefinic protons formed by the degradation, is also coupled to an

A'B' system (doublet 4.216, J=ll cps and triplet 3.796, J=ll cps) appearing

in a region typical for protons adjacent to oxygen.

The Hofmann product must therefore contain the sequence:

ha hb h* ha, hI I I I I

INDOLE ---C---- C--C ----c — 0 -— c —- IND0LEi I I

C H», CHgxxi

This observation is only compatible with the partial structure shown

by (xxii) .

69

H H N

XXII

From the molecular formula of the base, which necessitates a pentacyclic

structure, we need now to close another ring and have only one possible point

of attachment (at the starred carbon) since the nature and number of protons

at all other positions has been demonstrated in the discussion above, thus

the complete structure for the major Hofmann product is given as XLI in which

the configuration of the C-ethyl residue must be as shown, arising from hydro­

gen addition to the more exposed face of the molecule, and hence that for taber-

psychine as XLII which is compatible with all spectral data.

i i

H H

XLI XLII

70

A minor product from the Hofmann degradation of dihydrotaberpsychine

methiodide was also isolated although it could not be induced to crystallize.

However, the nuclear magnetic resonance spectrum shows apart from the expected

absorptions, a pair of doublets (4.98 and 5.636, J=1 cps) typical of a ter­

minal methylene group, which together with the preceeding evidence permitted

the formulation of the material as structure XLIII.

H

XLIII

Once the structure of taberpsychine was elucidated, the result obtained

from the Hofmann degradation performed on the methiodide of taberpsychine

itself could be rationalized, whereas before a rather complex nuclear magne­

tic resonance spectrum had rendered it difficult.

This degradation was effected using the same conditions as for the di­

violet spectrum shows the presence of a conjugated diene in addition to the

original indole residue, was obtained. The nuclear magnetic resonance spec­

trum of this taberpsychine-methine shows a very complex pattern in the region

where the absorption for olefinic protons is expected (see Figure 18). The

71

pattern is further complicated by the presence in the same area of the absorp­

tion due to the C.3 proton adjacent to the aromatic ring and the oxygen atom.

It was analyzed to show that the olefinic protons are those of a conjugated

diene which could only be formed by elimination of one of the protons in the

methyl group and migration of the double bond to open the piperidinic ring,

leaving the nitrogen attached to the eight-membered ring. The structure of

this degradation product was assigned as shown in XLIV.

XL IV

A minor product of this Hofmann degradation was observed by thin layer

chromatography but could not be isolated. This is probably the isomer of the

diene as depicted in XLV.

H

XLV

<$ values

Igure l8.- Nuclear magnetic resonance spectrum of taterpsyehine-methine

OO

73

With structure XLII established as that of taberpsychine, the problem

of the stereochemistry of the molecule is almost reduced to a question of

absolute stereochemistry, since once the configuration of one of the asymétrie

centers is fixed the closing of the rings determines the configuration at the

other centers. Thus it is the choice of either the stereochemistry shown in

structure XLVI or its complete mirror image.

XL VI

The product distribution in the Hofmann degradation of taberpsychine

methiodide, about 80% of the diene, can be readily explained by the study

of models of the molecule. In effect, the tri-dimensional structure XLVII

shows taberpsychine methiodide and below are shown Newman projections of

the two systems where the elimination could take place. It is readily seen

XLVI I

74

that the more favored is that involving the double bond, which is quasi-anti­

parallel to the bond to be broken, with the elimination taking place as shown

in partial structures (xxii) and (xxiii).

xxii XXlll

The small amount of the other isomer could not be produced by a normal

trans-elimination since the two activated "benzylic" protons are not suitably

placed to allow this, and probably the reaction proceeds through a carbanion

. intermediate.

75

Examination of molecular models of taberpsychine suggests that the hydro­

genation of the exocyclic double bond would be specifically from that face of

the molecule which would lead to'an equatorial ethyl group (A->B). Hydrogena­

tion from the other face, which should give the axial ethyl residue (as in C)

is seriously hindered by the eight-membered ring and the aromatic system.

The results of the Hofmann degradation of the methiodide of the dihydro com­

pound show that the predominant reaction involves the elimination of one of

the protons adjacent to the indole ring. It would be difficult to rationalise

this observation on the basis of structure C where the equatorial hydrogen sub­

stituent, introduced by hydrogenation, is stereochemically well situated for

a trans-elimination.

A

C

76

Another new alkaloid obtained from T. psychotrifolia was named taber-

psychidine, C ^qH24^2®2’

Its particular ultraviolet spectrum [^max 224 (10,000) and 318 (12,000)

20my] permitted its classification as a derivative of an a-acyl indole , as

shown by (xxiv).

iH

R

R'

O

XXIV

From the infrared spectrum one obtains information for the presence of

the carbonyl group (strong band 1650 cm-1) as well as that of the aromatic

moiety and the indolic NH. This spectrum also shows the presence of a hy­

droxyl group.

The nuclear magnetic resonance spectrum was difficult to perform due

to the only slight solubility of the base in most organic solvents, but the

problem was partly solved by the use of a hot saturated chloroform solution.

The spectrum showed the presence of a methyl group attached to the basic nitro­

gen (singlet 2.536) and of a methyl group on a tri-substituted double bond

(1 H quartet 5.406, J=6 cps and 3 H doublet 1.686, J=6 cps) in addition to

the information already known.

With this evidence in hand, the partial formula (xxv) is advanced.

77

Acetylation of the base with acetic anhydride in pyridine at room temper­

ature gives a monoacetate (molecular weight by mass spectrometry 366) and while

this product is rather unstable it could be characterized spectrally.

- OH

XXV

The infrared spectrum proves that it is an O-acetyl derivative (peaks

at 1745 and 1235 cm-1) and the nuclear magnetic resonance spectrum reveals

that the acetyl methyl is unusually shielded (3 H singlet 1.776). The most

useful information, however, is given by the mass spectrum (see Figure 19).

In effect, the spectrum of the base itself shows two major peaks at 152 and

122 m/e while in the acetate one finds a new peak at 194 m/e (42 mass units

from 152 m/e). Analyzing this data, the peak at 122 m/e strongly suggests a

pyridinium ion of the type depicted by (xxvi), coming from the part of the

molecule which contains the basic nitrogen since the indolic residue could

not possibly give such a peak. This pyridinium ion has been proposed as oc­

curring in the mass spectrometric degradation of alkaloids of this type^.

Following this reasoning, the peak at 152 m/e (30 mass units from 122

m/e) must be due to pyridinium (xxvi) plus a hydroxymethyl group, a fact which

seems to be confirmed by the appearance of the 194 m/e peak in the spectrum

of the acetate. The sequence of fragmentation would be then as given by

(xxvii).

relative intensity

H O CH

Figure 19»- Mass spectrum of tàberpsychidine

79

CH3i

xxvi (122 m/e)

194 m/e 152 m/e 122 m/e

xxvii

At this stage, the similarity to alkaloids related to vobasine (XLVII)

was evident and a derivative was prepared to help prove the structure chemi­

cally.

Reduction of taberpsychidine with sodium borohydride yielded a compound,

^20^26^2^2’ wh°se ultraviolet spectrum no longer shows the presence of the

keto group conjugated to the aromatic residue but the absorption is typical

of an unsubstituted indole. The infrared spectrum confirms the absence of

80

the carbonyl at 1650 cm""1, while the original OH band is now considerably

broadened.

R = COOCH3

XLVII

The nuclear magnetic resonance spectrum (Figure 20) although exhibiting

very little change, presents some new absorption which is both indication

that the keto group has been reduced and support for an important statement

made earlier concerning the structure of taberpsychine. This absorption is

a broad doublet appearing at 5.276 (1 H, J=6 cps) ascribable to the proton

on the carbon atom, previously part of the carbonyl group, which is now ad­

jacent to both oxygen and the aromatic moiety.

The mass spectrum of this derivative shows the appearance of a rather

strange peak at 154 m/e, which could be thought to arise from the half of the

molecule where the reduced center is found since in the base itself the peak

is insignificant. However, the relation with that center is only an indirect

one and it is produced by a different mode of fragmentation (Figure 21) in­

volving the new hydroxyl groupé.

Dihydrotaberpsychidine was compared with the product obtained by Renner

x-z

HOCK

OH

<$ VALUES

Figure 20.- Nuclear magnetic resonance spectrum of dihydrata"berpsychidine

82

HOCH.

OH

32 fa

V

CHINs

I2X

ch3

Figure 21.- Mass spectrometric fragmentation of dihydrotaberpsychidine

83

et al. from vobasine (XLVII) by reduction first with sodium borohydride (to

give vobasinol) and then with lithium aluminum hydride (to produce vobasinediol)

or reduction of vobasine with lithium aluminum hydride, All the physical data

given in the literature for vobasinediol are identical with those measured for

our dihydrotaberpsychidine, thus allowing us to write structure XLVIII for the

reduction product and XLIX for the base itself.

57

CH,

<N,

HOCMjT

o

XLVII I XLIX

The structure proposed, which is justified by the mass spectrometric

fragmentation shown in Figure 22, is identical with that proposed for affi-

58nine and although the physical constants given in the literature for the

latter were misleading,taberpsychidine is in fact affinine.

That the stereochemistry at C.16 is as shown in structure V (the same

as in vobasine (XLVII)), can be readily seen from the extreme shielding of

the acetyl methyl (3 H singlet 1.776) observed in the nuclear magnetic reso­

nance spectrum of O-acetyl taberpsychidine. In vobasine, the same shielding

is observed for the methyl of the ester group, this appearing at 2.636.

The third alkaloid whose structure was elucidated is a compound which

84

57 ■has been previously prepared by Renner et al. but never before reported as

a naturally occuring compound. It is known as 16-epi-vobasinic acid, and was

prepared from vobasine by hydrolysis with 20% methanolic potassium hydroxide.

The alkaloid, CgnHggNgO.,, is a high melting compound (m.p. 295° decomp.)

only sparingly soluble in most organic solvents which suggested it to be a

salt. However, no anion was detectable pointing to an internal salt or amino

acid.

The ultraviolet spectrum of the base C^max 283 (13,100) and 316 (20,000)

my] is that of an a-acyl indole and the infrared spectrum presents two very

typical bands in the region for carbonyl groups. One of them at 1650 cm-1

is assigned, in agreement with the ultraviolet spectrum, to a keto group on

carbon 3 conjugated to the indole moiety and the other at 1610 cm-1, a very

strong band which is only compatible with a carboxylate carbonyl.

The nuclear magnetic resonance spectrum (performed in 2% D^SO. in D2O

solution) shows the presence of an N-methyl group (3 H singlet at 3.005, dis­

placed to lower field due to the quaternisation of the basic nitrogen) and of

an exocyclic ethylidene chain (1 H quartet 5.935 and 3 H doublet 1.685, J=7

cps) in addition to the absorption expected from the indolic residue.

The mass spectrum (Figure 23) of the base again presents its principal

peak at 122 m/e, very typical fragmentation in the mass spectrometric degra­

dation of vobasine-like alkaloids, produced by the pyridinium ion of the type

shown by xxvi (page 79). Another important peak is that at 166 m/e (44 mass

units = -COO from 122) which coupled to the evidence for the presence of an

acidic residue in the molecule, leads to the conclusion that the latter is

Figure 22,- Mass spectrométrie fragmentation of taberpsychidine

CO -en

+ z

relative intensity

Figure 2$Mass spectrum of l6-epi-vobasinic acid

ooOx

attached to the six-membered ring producing the pyridinium ion (xxviii) in

the fragmentation.

87

ch3t

ch3

N + M-t-r ^ r ^

k--------- >HOOC

xxviii (166 m/e)

Based on this evidence, the structure for the alkaloid was proposed

as shown by L, which is that of 16 epi-vobasinic acid.

All spectral data are in agreement with this structure and a fragmen­

tation pattern, as given in Figure 24, accounts for all the principal peaks

observed in the mass spectrum.

R = COOH

L

Further proof for the structure is provided by méthylation of the base

by prolonged treatment with diazomethane. The infrared spectrum of the

Z-I

Figure 24 0- Mass spectraaietric Tranent at sen of l6-epi-voilas inic acid

oo00

89

product obtained shows the absorption typical of esters (1740 cm-1), while

the nuclear magnetic resonance spectrum proves it is a methyl ester (3 H

singlet 3.536). The position of the absorption for the methyl group of the

ester in the spectrum determines the stereochemistry at carbon 16, as being

epimeric to that in vobasine (XLVII) where the same absorption shows the methyl

group of the ester to be very shielded (3 H singlet 2.636) presumably because

it lies immediately above the electronic cloud of the aromatic ring.

The methylated derivative presents the same physical properties, and

infrared, nuclear magnetic resonance and mass spectra identical with those of

5716-epi-vobasine given in the literature . Direct comparison however was not

possible due to the lack of an authentic sample.

A fourth alkaloid which was called base M for purpose of identification,

^21^26^2^3’ obtained from a countercurrent distribution, still remains with an

unknown structure.

Its ultraviolet spectrum is that of an unsubstituted indole [X 227 r max

(44,700), 278 (sh. 8,800), 286 (9,600) and 294 (8,600) mp]. The infrared

spectrum shows the absorption for the indolic NH and the aromatic protons, as

well as a hydroxyl group and an ester carbonyl (at 1730 cm-1) accompanied by

another slightly weaker band at 1710 cm-1 probably due to the presence of

hydrogen bonded species.

The nuclear magnetic resonance spectrum demonstrates the presence of

the indolic residue and identifies the ester grouping as being a methyl ester.

It also indicates the existence of an N-methyl residue (3 H singlet 2.26) and

of a saturated methyl group with only one adjacent proton (3 H doublet 1.266,

90

J=7 cps). The latter should be adjacent to oxygen or nitrogen since decoupling

experiments showed its appearance as a doublet of doublets at 3.906 (J=7 and

3 cps).

The mass spectrum confirms the analytical figures (M+= 354,1956; ^21^26

NgO? requires: 354.1943) but the fragmentation pattern does not fit a structure

related to that of the other alkaloids obtained from this plant.

The information obtained on this base can be resumed as in partial struc­

ture (xxix) where the C-8 fragment can involve up to three rings.

— OH

Ce HS Ml0

:N- CH,

XCOOCH

H

CH,

xxix

With the

at a plausible

amount of material available it was not possible to arrive

structure for the alkaloid and further results will be re­

ported elsewhere.

-.EXPERIMENTAL.-

-.GENERAL REMARKS.-

Melting points are uncorrected. Unless otherwise stated optical rota­

tions and ultraviolet spectra (e values in parentheses) were measured in

ethanol. Ultraviolet spectra were registered on a Beckmann spectrophotometer

model DK-1A and optical rotations in a Carl Zeiss polarimeter with circular

scale (c = 1.0 unless otherwise stated). Infrared spectra were performed on

nujol mulls or potassium bromide pellets using a Beckmann spectrophotometer

model IR-4. Nuclear magnetic resonance (NMR) spectra were measured on 5-10%

solutions with Varian Associates spectrometers models HR-100 or A-60. Tetra-

methylsilane protons taken as 0 p.p.m. Mass spectra were registered using

either an AEI MS-2H or a Varian Associates M-66 spectrometer.

Countercurrent distribution fractions were numbered from the first sta­

tionary phase (fraction N° 1) to the furthest advanced buffer phase (highest

numbered fraction). Aqueous phase was acetate buffer.

Elementary analyses were performed by Dr. Franz Pascher, Bonn, Germany.

The extractions of the alkaloids were performed in the following manner

(unless otherwise stated):

The plant material was dried in a current of warm air as soon as possible

after collection and then reduced to small pieces, about an inch long, with a

.chaff or tobacco cutting machine. A Wiley type grinding mill was then used

to further reduce the material to a coarse powder which was then extracted

by percolation with methanol or ethanol at room temperature. The volume of

solvent necessary varies with the plant but normally involves 10-20 litres

92

93

per kilo of ground material. The solvent is removed by distillation, or more

rapidly in a climbing film evaporator but care is taken not to heat the solu­

tion excessively. At this stage the extract contains every conceivable type

of natural material and is a black viscous tar.

The crude extract is suspended or dissolved in dilute acid and the in­

soluble portion removed by filtration where possible or by decantation. Ex­

traction with chloroform and suitably changing the pH of the aqueous solution

enables one to obtain a crude base fraction which is predominantly alkaloidal.

To separate the individual alkaloids the most rewarding first step is

to distribute the crude base between an organic solvent and an aqueous buffer.

The pH of the buffer is chosen, in the first instance to give a distribution

coefficient of about 0.5 but depending on the complexity of the mixture it

may be changed. This first countercurrent, which could involve over 100 gms.

of base, is best performed in large separatory funnels. Depending on the

results of this first distribution, further separation is achieved by normal

manipulations. The experimental details of the individual separations have

been limited to those steps which provided the pure alkaloids discussed in

the thesis.

-.CHAPTER I.-

A. ALKALOIDS OF ASPIDOSPERMA EXCELSUM.

95

EXTRACTION AND ISOLATION OF THE ALKALOIDS.-

The A. excelsum bark (6.5 kg) was milled and extracted by methanol per­

colation at room temperature„ Evaporation of the methanol was greatly com­

plicated by foaming, and the tar like residue certainly contained a large

volume of water. The crude tar was mixed with 2% hydrochloric acid and the

acid solution continuously extracted with chloroform to remove non basic im­

purities. Paper chromatography showed the presence of at least two alkaloids

in quantity in the resulting aqueous solution, and one of the bases had been

extracted, as the hydrochloride, into the chloroform.

Yohimbine (XI)

The hydrochloric acid solution above was basified with ammonia, and

the crude base (149 g) was obtained by chloroform extraction (entrainement of

some of the aqueous phase as an emulsion exaggerated the yield of basic ma­

terial) . A portion of the crude base (49 g) was distributed in a counter-

current distribution apparatus between a stationary chloroform phase and ace­

tate buffer (pH 3.72) for a total of 50 transfers. The tubes in the center

of the apparatus contained the major base, which crystallized from acetone

(6.39 g) and which was recrystallized for analysis from the same solvent,

m.p. 241-242°, [<x]D + 45°.

Anal. Found :: C, 71.2; H, 7.5; N,

OO ; o, 13.6. C21^26^2^3 squires

c, 71.2; H, 7.4; N, 7.9 ; o, 13.6%.

- UV spectrum : Amax 226 (36,500), 278 (7,400) and 286 (sh. 6,100) my.

IR spectrum : peaks at 3510 (OH), 3330 (NH), 2820 and 2780 (Bohlmann trans-

quinolizidine bands), 1735 (ester), 765 and 750 (aromatic)

cm-1.

96

NMR spectrum : 1 H broad singlet 7.806 (NH);4 H multiplet 7.00-7.626

(aromatic); 1 H broad singlet 4,226 (C.17 H); 3 H singlet

3.786 (methyl ester).

Mass spectrum : 354 (M+), 295 (M+ - COOCHj), 184, 170, 169, 156, m/e.

These spectral results agree with those given in the literature for

22yohimbine , and the identity was confirmed by direct comparison (infrared

and mixture melting point).

Q-acetylyohimbine (XII)

The chloroform containing the non-basic materials was concentrated to

about half volume and cooled. A crystalline substance, collected by filtra­

tion (4.58 g), was shown to be an alkaloid hydrochloride and was recrystal­

lized frapi methanol-acetone, m.p. 274-275° (decomp.).

Anal. Found

UV spectrum

C, 63.7; H, 6.9; Cl, 8.5. C^HggNgO^Cl requires :

C, 63.8; H, 6.8; Cl, 8.2%.

\nax 222 (39,500), 274 (7,500), 281 (7,500) and 290 (6,000)

my.

IR spectrum

NMR spectrum

Mass spectrum :

peaks at 3370 (NH), 1730 (ester), 1725 and 1250 (O-acetyl),

and 740 (aromatic) cm"1.

1 H broad singlet 7.876 (NH); 4 H multiplet 7.00-7.606

(aromatic); 1 H multiplet 5.476 (C.17 H); 3 H singlet 3.696

(methyl ester); 3 H singlet 2.046 (O-acetyl).

396 (M+), 395, 184, 170, 169, 156, m/e.

For direct comparison, a sample of yohimbine was acetylated in acetic

anhydride-pyridine at room temperature. The two samples were identical in all

97

respects and showed no depression of the melting point when mixed. A sample

of the natural 0-acetylyohimbine (300 mg) was hydrolyzed in anhydrous methanol

containing a small amount of concentrated sulfuric acid (conditions known not

to affect the methyl ester). The product was shown by direct comparison to

be yohimbine.

0-acetylyohimbine was also obtained from the countercurrent distribution.

A total of 6.31 g was isolated.

Further experiments with the mother liquors have yielded the minor bases

4 23excelsinine (10-methoxycorynanthine) and a-yohimbine .

-.CHAPTER I.-

B. ALKALOIDS OF ASPIDOSPERMA CUSPA.

99

EXTRACTION AND ISOLATION OF THE ALKALOIDS.-

The A. cuspa bark was air dried and finely divided (12.6 kg) and extrac­

ted by percolation with ethanol until the percolate no longer gave positive

tests with the usual alkaloid reagents. The ethanolic solution was concentra­

ted under reduced pressure to give a tar which was triturated several times

with 2% aqueous hydrochloric acid. The aqueous extracts were combined and

extracted continuously with chloroform to remove non-basic material and then

rendered alkaline with ammonia and the crude basic fraction (80 g) obtained by

chloroform extraction. The crude bases were redissolved in dilute hydrochloric

acid and after removing neutrals (as above), the crude base (54 g) was again

obtained by chloroform extraction.

The crude basic material (54 g) was distributed between a stationary

chloroform phase and acetate buffer (pH 3.72) in a fifty tube countercurrent

distribution apparatus. The most advanced aqueous phases (tubes 37-50) showed

by paper chromatography one principal alkaloid in quantity and aspidodasycar-

pine (12.5 g) was obtained from these tubes by normal manipulation. The crude

base from the remaining tubes was combined and distributed again between chloro­

form (stationary) and acetate buffer (pH 3.42) for fifty transfers and tubes

40-48 afforded more aspidodasycarpine (1.5 g). Tubes 6-8 contained essential­

ly one base, which was shown later to be des-O-methylaspidocarpine (978 mg)

and tubes 9-15 gave a crystalline alkaloid (1.089 g) shown to be bumamine

(des-acetylpicraline). The buffer phase in tubes 1-5 afforded picraline

(5 mg). Further countercurrent distribution of the mother liquors afforded

more burnamine (371 mg) and des-O-methylaspidocarpine (210 mg) .

From the crude mixture, pyridine XXIX (5 mg) was obtained by preparative

paper chromatography using Whatman paper N° 31 (double thick, 3 mm) and

pyridine-ethyl acetate-water (7.5:2.6:1.65) as solvent.

100

Burnamine (des-acetylpicraline) (XXIII)

The material obtained after recrystallizing from acetone gave m.p,

190-191° (decomp.), [a]D - 151° .

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

C, 68.4; H, 6.7; N, 7.5; 0, 17.5. requires :

C, 68.5; H, 6.6; N, 7.6; 0, 17.4%.

X 233 (8,800) and 283 (3,660) my; ?0% 274.5 and

max ’ max

302 my,

peaks at 3550 (OH), 3050 (NH), 1747 (ester), 1615 (aromatic)

cm-1.

complex 4 H multiplet centered at 7.066 (aromatic); 1 H

singlet 5.166 (NH) ; 1 H quartet 5.386, J=7.5 cps (1 olefinic

proton); 3 H doublet 1.566, 0*7.5 cps (C-methyl); 1 H doublet

4.756, J=2.5 cps (C.5 H); 3 H singlet 3.606 (methyl ester).

Mass spectrum : 368 (M+), 350 (M - H^O), 337 (M - .CH^OH), 320, 309 (M -

.COOŒL), 239, 194, 180, 168, 157, 144, 130, m/e.

A mixed melting point with an authentic sample showed no depression and

the infrared spectrum was superposable with that of burnamine*.

Des-O-methylaspidocarpine (XXV)

The base crystallized from acetone to give a substance containing acetone

We thank Dr. W.I. Taylor for kindly providing an authentic sampleof burnamine.

101

of crystallization, m.p. 139-140°, [al^ + 97°. For analysis the sample was

dried in vacuo for several weeks during which time the analysis changed from

that of the hydrate to that of the base alone.

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

C, 70.4; H, 8.0; Mol. Wt. (by mass spectrometry) 356.1978.

^21^28^2^3 reclui-res• C, 70.7; H, 8.0; Mol. Wt. 356.2022.

X 224.5 (18,800) and 259.5 (7,300) my; 236.5

(15,800) and 300 (3,000) my.

peaks at 3250 (OH), 1635 (N-acyl), 1580 (aromatic) cm-1.

1 H singlet 10.856 (chelated phenolic hydroxyl); 2 H singlet

6.366 (aromatic); 1 H quartet 4.006, J=5 cps (C.2 H); 3 H

singlet 2.256 (N-acetyl); 3 H doublet 0.706, J=4 cps (satu­

rated C-methyl).

Mass spectrum : 356 (M^), 328 (M - CH.CH^), 327 (M - .C^Hg), 190, 162, 152,

124, m/e.

The di-O-acetyl derivative was prepared dissolving the base (100 mg) in

pyridine (5 ml) and adding an excess of acetic anhydride (0.5 ml). The mix­

ture was left standing over night. Methanol was then added to hydrolyze the

excess acetic anhydride and the solution taken to dryness under reduced

pressure. The residue was redissolved in chloroform, backwashed twice with

ammonia-water, the chloroform dried over sodium sulfate and evaporated to

dryness. The acetate (120 mg) was crystallized from ethanol, m.p. 138-140°

(decomp.), [a]D - 19°.

IR spectrum : peaks at 1780 and 1210 (0-acetate), 1680 (N-acetyl), 1610

(aromatic) cm"1.

102

NMR spectrum : 2 H AB quartet 6.99 and 6.885, J=4.8 cps (aromatic); 6 H

singlet 2.215 (O-acetyl); 3 H singlet 2.155 (N-acetyl);

3 H triplet 0.615, J=3.6 cps (methyl on saturated ethyl

side chain).

Mass spectrum : nominal mass 440.

(-) Pyrifolidine (from des-O-methylaspidocarpine) (XXVII)

Des-O-methylaspidocarpine (130 mg) was dissolved in dry acetone (25 ml)

and anhydrous potassium carbonate (1 g) was added together with dimethyl sul­

fate (0.5 ml). The mixture was refluxed for 17 hours and then another portion

of dimethyl sulfate (0.5 ml) added and the procedure was repeated after 30

hours. The reflux was continued for a total of 40 hours. The mixture was

then diluted with cold water, the acetone removed under reduced pressure and

the aqueous solution extracted four times with chloroform. The chloroform was

evaporated to dryness and the residue (100 mg) crystallized and recrystallized

from acetone-ether, m.p. 150-151°, [a]^ - 83.5°, [a]^ - 95° (CHCl^).

UV spectrum

IR spectrum

NMR spectrum

Mass spectrum

\nax 224 (28,900), 252 (11,700) and 288 (2,800) mp.

peaks at 1680 (N-acetyl), 1610 (aromatic) cm-1.

2 H AB quartet 6.49 and 6.685, J=8 cps (aromatic); 3 H

singlet 3.675 and 3 H singlet 3.596 (methoxyls); 3 H singlet

2.056 (N-acetyl); 3 H triplet 0,736, J=6 cps (methyl on sa­

turated ethyl side chain).

384 (M+), 369 (M - 15), 356 (M - CH^CH^), 355 (M - .C^Hg),

353 (M - OCH3), 341 (M - 43), 190, 171, 152, 124, m/e.

Comparison of this product with an authentic sample* by spectral and

We thank Dr. Carl Djerassi for kindly providing an authentic sampleof pyrifolidine.

103

chromatographie methods confirmed their identity.

Aspidodasycarpine (XV)

The base was recrystallized several times from acetone, m.p. 207-209°

(decomp.), [ot]D - 130°, [a]D - 114° (CHClj).

Anal. Found

ÜV spectrum

IR spectrum

NMR spectrum

C, 68.2; H, 7.0; N, 7.8; 0, 17.5. requires :

C, 68.1; H, 7.1; N, 7.6; 0, 17.3%.

X 239 (10,000) and 292 (4,300) my.

peaks at 3360 (OH), 3580 (indoline NH), 3300 (N.-H), 1610

(aromatic) and 1720 (ester) cm-1.

4 H multiplet 7.005 (aromatic); 1 H quartet 5.525, J=6.5 cps

(1 olefinic proton); 3 H doublet of doublets 1.75, J=6.5 and

2 cps (unsaturated C-methyl); 3 H singlet 3.765 (methyl ester).

Mass spectrum : 370 (M+), 368, 339, 325; 267, 263, 232, 204, 172, 156, 144,

130, 108, m/e.

A mixed melting point with aspidodasycarpine*

the infrared spectra were identical.

showed no depression and

N,Q-diacetylaspidodasycarpine (XVI)

a) To aspidodasycarpine (3.02 g) in pyridine (20 ml) was added acetic

anhydride (10 ml). After 48 hours excess methanol was added with cooling and

the solution then evaporated to dryness under reduced pressure. The residue

was recrystallized from acetone (2.85 g), m.p. 175°, [a]^ - 174° (CHCl^),

* We thank Dr. Carl Djerassi for kindly providing an authentic sampleof aspidodasycarpine.

104

[alD - 161°.

Anal. Found : C, 66.2; H, 6.8; N, 6.0; 0, 21.0. CggH^NgOg requires:

C, 66.1; H, 6.7; N, 6.2; 0, 21.1%.

UV spectrum : A 241.5 (8,900) and 297.5 (3,100) mp.

IR spectrum : peaks at 3450 (indoline NH), 1750 and 1230 (0-acetyl),

1715 (methyl ester), 1630 (N-acetyl) cm-1.

NMR spectrum 4 H multiplet 7.086 (aromatic); 1 H quartet 5.676, J=7 cps

(olefinic); 1 H singlet 4.856 (indoline NH); 3 H singlet

3.786 (methyl ester); 3 H singlet 2.176 (N-acetyl); 3 H

singlet 1.926 (0-acetate); 3 H doublet 1.736, J=7 cps (un­

saturated C-methyl).

b) The same N,O-diacetyl derivative (XVI) was obtained by similar

acetylation of N-acetylaspidodasycarpine (XVII).

N-acetylaspidodasycarpine (XVII)

a) To aspidodasycarpine (500 mg) dissolved in pyridine (10 ml) was

slowly added a cold solution of acetic anhydride (100 pi) in pyridine. After

6 hours at 5°C the solution was allowed to warm up to room temperature and

then evaporated to dryness under reduced pressure without heating. The resi­

due, upon dissolving in acetone, afforded crystals of the N-acetyl deriva­

tive (302 mg) and from the mother liquors a further quantity (103 mg) was

obtained by extraction of the neutral components in the usual manner. A

small amount of unchanged aspidodasycarpine (47 mg) was obtained from the

basic impurities separated by washing with hydrochloric acid.

105

The pure N-acetyl derivative showed m.p. 250-253°, [a]p - 147°.

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

C, 66.8; H, 7.1; N, 7.0; 0, 19.2. ^23^28^2^5 re(lu^-res:

C, 67.0; H, 6.8; N, 6.8; 0, 19.4%.

X 241.5 (8,000) and 298.5 (2,800) my.

peaks at 3250 (indoline NH), 1760 (methyl ester), 1620

(N-acetyl) cm-1.

4 H multiplet 6.756 (aromatic); 1 H quartet 5.386, J=7 cps

(olefinic proton); 3 H singlet 2.056 (N-acetyl); 3 H doublet

1.656, J=7 cps (unsaturated C-methyl).

b) N,O-diacetylaspidodasycarpine (200 mg) was dissolved in dry methanol

(10 ml) and concentrated sulfuric acid (0.25 ml) was added. 18 hours of

reflux afforded a neutral product (80 mg) shown to be the N-acetyl derivative

(XVII) and a basic fraction (65 mg) identified as aspidodasycarpine.

c) N,O-diacetylaspidodasycarpine (200 mg) was reacted with excess

sodium borohydride in methanol (15 ml) at room temperature. Water was added

to the reaction after 18 hours and the foam obtained by chloroform extraction

(185 mg) crystallized upon solution in acetone. The product was identical

in all respects with the N-acetyl derivative (XVII) obtained previously.

Dihydroaspidodasycarpine (XVIII)

Aspidodasycarpine (500 mg) was dissolved in methanol (50 ml) containing

concentrated hydrochloric acid (3 ml) and platinum oxide (90 mg) was added.

The mixture was shaken with hydrogen at 50 p.s.i. for 22 hours, filtered and

part of the methanol was removed under reduced pressure. After diluting with

106

water and basi frying with ammonia, the product (500 mg) was obtained by chloro­

form extraction. Only one component was present as shown by thin layer chroma­

tography and this crystallized slowly from acetone, m.p. 209-211°, [a]^ - 212°.

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

Mass spectrum

C, 67.6; H, 7.7; N, 7.7; 0, 17.4. CgiH^gN^O. requires:

C, 67.7; H, 7.6; N, 7.5; 0, 17.2%.

\nax 241 (9,500) and 297 (3,700) my.

peaks at 3550 (indoline NH), 3400 (OH), 3170 (N.-H), 1723

(methyl ester), 1610 (aromatic) cm-1.

4 H multiplet 7.435 (aromatic); 3 H singlet 3.966 (methyl

ester); 3 H doublet 1.005, J=6 cps (saturated C-methyl).

372.2054 (M+; calculated for C^H.gN^O.: 372.2049), 342

(M - .CHgO), 329, 282, 254, 253, 232, 226, 211, 194, 181,

144, 130, 110, m/e.

N,Q-diacetyldihydroaspidodasycarpine (XIX)

a) Dihydroaspidodasycarpine (350 mg) was acetylated using acetic

anhydride (1 ml) in pyridine (5 ml) for 20 hours. Crystals which appeared

in the solution during the reaction were filtered (129 mg) and washed with,

and then recrystallized from, ethanol, m.p. 321-324° (decomp.), [a]^ - 168°.

Anal. Found : C, 64.6; H, 6.8; N, 6.2; 0, 22.6. ^gH^NgOg,. 1/2 H^O requires :

C, 64.5; H, 7.1; N, 6.0; 0, 22.4%.

UV spectrum : X^^ 240 (6,000) and 295 (1,800) my.

IR spectrum : peaks at 3200 (indoline NH), 1755 (0-acetyl), 1735 (methyl

ester), 1645 (N-acetyl), 1230 (0-acetyl) cm-1,

due to the limited solubility of the compound, only the more

intense peaks were visible: 3 H singlet 2.255 (N-acetyl) and

NMR spectrum

107

3 H singlet 2,125 (O-acetyl).

Mass spectrum : 456 (M+), 425 (M - 31), 414 (M - 42), 397, 384, 237, 225, 172,

161, 152, 144, 130, 110, 108, m/e.

b) N,O-diacetylaspidodasycarpine (500 mg) was hydrogenated over Adams'

catalyst (100 mg) in ethanol (100 ml) containing acetic acid (50 ml). After

24 hours the catalyst was removed by filtration and the solution evaporated

to dryness under reduced pressure. The residue crystallized from methanol

(414 mg), m.p. 321-324° (decomp.) and was readily identified as the same N,0-

diacetyldihydro derivative obtained above.

N-acetyldihydroaspidodasycarpine (XX)

a) By partial acetylation of dihydroaspidodasycarpine.

To a cooled solution of dihydroaspidodasycarpine (134 mg) in pyri­

dine (5 ml) was added acetic anhydride (25 pi). After 2 hours at 5°C and

standing at room temperature overnight, the solution was evaporated to dry­

ness under reduced pressure and the white residue redissolved in chloroform.

Washing the chloroform with dilute hydrochloric acid afforded some unreacted

dihydroaspidodasycarpine (28 mg) and the residue obtained by evaporating the

washed (ammonia-water) chloroform solution crystallized from acetone, m.p.

262-264°, [a]D - 193°.

Anal. Found : C, 66.8; H, 7.1; N, 7.0; 0, 19.2. C^H^^Og requires:

C, 66.6; H, 7.3; N, 6.8; 0, 19.3%.

UV spectrum : 241 (8,000) and 298 (2,800) mp.

IR spectrum : peaks at 3550 (OH), 3200 (indoline NH), 1740 (methyl ester),

1640 (N-acetyl) cm-1.

108

NMR spectrum : 4 H multiplet 7.046 (aromatic) ; 3 H singlet 3.756 (methyl

ester); 3 H singlet 2.156 (N-acetyl); 3 H doublet 0.996,

J=5 cps (saturated C-methyl).

Mass spectrum : 414 (M+), 384 (M - CH^O), 352, 226, 172, 161, 152, 144,

143, 130, 110, m/e.

b) By hydrogenation of N-acetylaspidodasycarpine.

N-acetylaspidodasycarpine (200 mg) was dissolved in methanol (50 ml)

containing hydrochloric acid (3.5 ml). The solution was shaken under hydrogen

(50 p.s.i.) with Adams' catalyst (90 mg) for 48 hours and then filtered. Some

methanol was removed on a rotatory evaporator and the solution then diluted

with water (200 ml). Chloroform extraction gave a quantitative yield of a

pale yellow foam which crystallized on contact with acetone, m.p. 262-264°

(identical with the material obtained in (a) above).

c) By hydrolysis of N,O-diacetyldihydroaspidodasycarpine.

N,O-diacetyldihydroaspidodasycarpine (100 mg) dissolved in dry

methanol (10 ml) containing concentrated sulfuric acid (0.25 ml) was refluxed

for 20 hours. The foam (87 mg) obtained by chloroform extraction afforded a

crystalline product identical with that obtained above.

A1>2 dehydro-4,5,17-triacetyldihydroaspidodasycarpine (XXII)

To N,O-diacetyldihydroaspidodasycarpine (254 mg) suspended in acetic

anhydride (15 ml) was added concentrated sulfuric acid (9 drops) which suf­

ficed to dissolve the suspended material. After 5 hours at room temperature

the mixture was poured onto crushed ice and dilute ammonia and then extracted

109

with chloroform. The latter was washed with water and dried over sodium sulfate,

and on filtration and evaporation to dryness yielded a foam (227 mg) extremely

soluble in acetone but which was crystallized from acetone-ether (148 mg), m.p.

166-168°, Co3d - 31°.

Anal. Found : C, 64.9; H, 6.8; N, 5.9; 0, 22.6. requires:

C, 65.0; H, 6.9; N, 5.5; 0, 22.5%.

UV spectrum : A 222 (24,200) and 261 (6,800) mp.

IR spectrum : peaks at 1755 (methyl ester), 1755 and 1230 (0-acetyl),

1655 (N-acetyl) cm-1.

NMR spectrum : 4 H multiplet 7.555 (aromatic); 3 H singlet 3.775 (methyl

ester); 3 H singlet 2.275 (N-acetyl); 6 H singlet 1.735

(2 0-acetyl); 3 H doublet 1.085, J=8 cps (saturated C-methyl)

Mass spectrum : 498.2368 (M+; calculated for ^ykL^F^Oy : 498.2366), 456

(M - CHg=C0), 439 (M - .C00CH_), 425, 411, 397, 385, 252,

152, 136, 128, 110, 97, m/e.

Sodium borohydride reduction of the indolenine XXII

The indolenine (102 mg) was reacted with sodium borohydride (25 mg) in

methanol (10 ml) for 24 hours. A crystalline product, which had separated

from the solution, was collected by filtration (15 mg) and shown by the normal

methods to be N,0-diacetyldihydroaspidodasycarpine, m.p. 325-327° (decomp.).

The mother liquors were diluted with water and the product isolated by chloro­

form extraction. The resulting foam gave a further quantity of the diacetyl-

dihydroderivative (8 mg) and then the more soluble N-acetyldihydroaspidodasy-

carpine (54 mg) identified in the usual manner.

110

Picraline (XXVIII)

The base (5 mg) obtained by normal procedures from fractions 1-5 (buffer)

from a countercurrent distribution (chloroform-buffer acetate pH 3.42) showed

m.p. 170-175° (impure).

UV spectrum : 230 and 282 mu.

IR spectrum : peaks at 3340 (indoline NH), 1740 and 1240 (0-acetyl),

1740 (methyl ester), 1610 and 740 (aromatic) cm-1.

Mass spectrum : 410 (M+), 395, 382 (M - 28), 379 (M - .0CH3), 367 (M -

CH3C0) 351 (M - .COOCH3 or M - .OCOCHg), 337 (M - .CH^-

OCOCH3), 239, 194, 180, 157, 144, 130, 124, m/e.

The mass spectrometric fragmentation is identical with that of picraline

34reported in the literature

Pyridine XXIX

The base obtained by preparative paper chromatography was purified by

sublimation, m.p. 120° (probably a salt).

UV spectrum : X_ 256 and 262 my.r max

IR spectrum : peaks at 1600 and 1565 (pyridine), 725 (aromatic) cm-1.

Mass spectrum : 149 (M+), 134, 120, 106, 79, 77, 51, m/e (see figure 8,

page 38).

-.CHAPTER II.-

ALKALOIDS OF ASPIDOSPERMA FENDLERI.

112

ISOLATION OF THE ALKALOIDS FROM THE SEEDS.-

The dried seeds (480 g) of Aspidosperma fendleri Woodson‘S were conti­

nuously extracted with methanol in a Soxhlet extractor until the extract showed

no further base with the common alkaloid reagents.

Concentration of the extract gave a brown tar which was dissolved in 2%

hydrochloric acid; after filtration the non-basic material was removed by con­

tinuous extraction with chloroform. The aqueous solution was then basified

with ammonia and the crude mixture of bases (14.8 g) was obtained by chloro­

form extraction. Shaking the chloroform containing the non-basic material with

hydrochloric acid, basification and extraction yielded a further quantity of

crude base (12.7 g; total: 27.5 g).

The crude base (27 g) was distributed between chloroform (stationary

phase) and acetate buffer (pH 4.45) in a countercurrent distribution apparatus

for a total of 20 transfers. The buffer entering the first tube of the appa­

ratus was then replaced by acetate buffer of pH 4.00, and a further 20 trans­

fers were completed. The mixture of stronger bases obtained in the leading

aqueous fractions (tubes 33 to 40; 917 mg) was chromatographed over type H

alumina in benzene. The first eluate which contained alkaloid (benzene-chloro­

form, 3:1) gave white crystals of fendleridine (209 mg).

Fractions 15 to 32 (obtained from the countercurrent distribution)

yielded aspidofendlerine (135 mg) which was recrystallized from methanol.

Chromatography of countercurrent fractions 1 to 12 over alumina in

benzene gave the major base, fendlerine (12.6 g), as a thick, pale yellow oil

which crystallized on contact with acetone.

113

Fractions 13 and 14 from the countercurrent contained, in addition to

fendlerine, a small amount of another base which was purified by repeated

preparative chromatography on Whatman 3 mm paper. In this manner a total of

64 mg (m.p. 176-178°) was obtained but, apart from the UV spectrum [X 240

(8,000) and 290 (3,100) mpl and the IR spectrum (no OH or NH bands, an ester

carbonyl peak at 1735 cm-1), this base has resisted more positive charac­

terization.

Fendleridine (Aspidoalbidine) (XXXVI)

The base was recrystallized from acetone.and then sublimed for analysis,

m.p. 185-186°.

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

C, 77.0; H, 8.0; N, 9.4; 0, 5.8. C^gH^^N^O requires:

C, 77.0; H, 8.2; N, 9.4; 0, 5.4%.

\nax 242 (7,300) and 293 (3,050) mp.

peaks at 3300 (indoline NH) and 1600 (aromatic) cm-1,

nothing outstanding except for a 4 H complex multiplet from

6.55 to 7.506 (aromatic) and a 2 H AB quartet 3.93 and

4.036, J=3 cps (methylene adjacent to oxygen in the oxide

ring).

Mass spectrum : 296 (M+), 268 (M - OLrCHg), 252 (M - CH.CH0), 144, 138,

130, m/e.

Fendlerine (N-propionyl-16 methoxy-17 hydroxyaspidoalbidine) (XXXVIII)

After several recrystallizations from acetone the base melted at

179-181°.

114

Anal. Found

UV spectrum

G, 69.1; H, 7.5; N, 7.2; 0, 16.2; OCH^, 7.7 .

requires : C, 69.3; H, 7.6; N, 7.0; 0, 16.1; OCH^(one),

7.8%.

225 (19,200) and 258 (3,560) mp; 220

(17,200) and 300 (2,440) mp.

IR spectrum

NMR spectrum

Mass spectrum :

peaks at 1635 (N-acyl), and 1602 and 1570 (aromatic) cm-1.

1 H singlet 12.426 (chelated phenolic OH); 1 H singlet 10.756

(NH); 2 H AB quartet 6.71 and 7.056, J=8 cps (aromatic);

3 H singlet 3.856 (O-methyl); 2 H quartet 2.42 and 2.646,

J=7 cps, and 3 H triplet 1.256, J=7 cps (N-propionyl residue).

398 (M+), 370 (M - CH^CH^), 356, 355, 342 (M - propionyl),

325 (M - propionyl - CH^O), 176, 174, 161, 138, m/e.

To confirm the presence of the N-propionyl group, the base (50 mg) was

hydrolyzed in 20% sulfuric acid and a sample of the aqueous distillate was

injected into a vapor phase chromatograph, yielding only propionic acid (by

comparison with an authentic sample).

The perchlorate, m.p. 245-248°, was prepared by dissolving fendlerine in

dilute aqueous perchloric acid and extracting with chloroform. The chloroform

solution was dried over sodium sulfate and evaporated to give the crystalline

salt, which was recrystallized from ethanol.

Anal, Found : N, 5.0; 0, 26.6; Cl, 6.5 . C23^3. HCIO^. C2H ,-OH requires:

N, 5.1; 0, 26.4; Cl, 6.5%.

The characteristic peaks of ethanol (of crystallization) were clearly

visible in the NMR spectrum of this salt, determined as a solution in deuterium

oxide.

115

Aspidofendlerine (N-acetyl-16,17 dihydroxyaspidoalbidine) (XL)

The small amount of base was sublimed twice and then recrystallized from

ethanol, m.p. 278° (decomp.)»

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

Mass spectrum :

C, 67.9; H, 7.2; N, 7.5; 0, 17.5 . C21H26N2°4 requires :

C, 68.1; H, 7.1; N, 7.6; 0, 17.3%.

\nax (H,500) and 259 (5,000) my; boric acid

250 (broad plateau) and 290 (shoulder) my.

peaks at 3240 (OH), 1635 (N-acyl), and 1615 and 1580 (aroma­

tic) cm"1.

1 H singlet 11.066 (chelated phenolic OH); 2 H AB quartet

6.75 and 7.026, J=8.5 cps (aromatic); 3 H singlet 2.306

(N-acetyl).

384 (M^, 356 (M - CH^CH^), 341, 294, 242, 227, 138, m/e.

Aspidolimidine (from fendlerine) (XXXV)

Fendlerine (330 mg) was hydrolyzed by refluxing with 10% sulfuric acid

for four hours. The solution was then cooled and, in an atmosphere of nitro­

gen, made basic with ammonia and extracted with chloroform. The dried solu­

tion gave a glass (260 mg) which could not be induced to crystallize. The

UV spectrum showed peaks at 215, 240 (shoulder) and 296 my; N-acyl absorption

was absent in the infrared spectrum. The hydrolysis product (230 mg) was

dissolved in pyridine (10 ml), acetyl chloride was added and the mixture was

shaken well for 10 minutes, allowed to stand at room temperature for 2 hours

and then poured into ice water and basified with ammonia. Extraction with

chloroform gave a brown foam (240 mg), which showed both N-acetyl and 0-acetyl

absorption in the infrared spectrum. Under nitrogen, the crude product was

116

hydrolyzed with dilute methanelie sodium hydroxide (3 hours) and then poured

into water ; a yellow foam was obtained by extraction with chloroform and

evaporation of the chloroform solution to dryness . Alumina chromatography of

the foam gave a crystalline product (150 mg), which was recrystallized from

ethanol and sublimed for analysis, m.p. 194-197°.

The ultraviolet, infrared and NMR spectra were identical with those of

aspidolimidine, and a mixed melting point with a sample* of the latter gave no

depression.

Anal. Found : C, 68.6; H, 7.3; N, 7.3; 0, 16.6 . C22H28N2°4 recluires:

C, 68.7; H, 7.3; N, 7.3; 0, 16.6%.

NMR spectrum : 1 H singlet 10.746 (chelated phenolic OH); 2 H AB quartet

6.71 and 7.056, J=4.8 cps (aromatic); 3 H singlet 3.876

(0-methyl); 2 H multiplet centered at 4.086 (methylene ad­

jacent to oxygen in the oxyde ring); 3 H singlet 2.306 (N-

acetyl).

Aspidolimidine (from aspidofendlerine) (XXXV)

Aspidofendlerine (35 mg) was dissolved in methanol and treated with

ethereal diazomethane for 4 hours. The volatile materials were removed at

the pump and the residue triturated with acetone. The insoluble part was

shown to be unchanged aspidofendlerine (15 mg) and the soluble portion was

repeatedly sublimed to give a crystalline product (17 mg), identical (mel­

ting point, mixed melting point and infrared spectrum) with aspidolimidine.

We thank Dr. B. Gilbert for kindly providing an authentic sample of aspidolimidine.

117

Extraction and isolation of the alkaloids of the trunk bark.-

The dried trunk bark (10 kg) was ground and percolated with methanol

until the extract no longer gave a colour with Vassler's reagent.

Isolation of the alkaloids, following the procedure used for the seeds,

yielded almost exclusively fendlerine and small amounts of aspidolimidine and

fendleridine.

Alkaloids of the root bark.-

Extraction of the root bark yielded, apart from large quantities of

fendlerine and aspidolimidine, three minor alkaloids, none of which seems

to be related to the aspidoalbidine type of alkaloids and, to date, their

structures remain unknown.

-.CHAPTER III.-

ALKALOIDS OF TABERNAEMONTAM PSYCHOTRIFOLIA.

119

EXTRACTION AND ISOLATION OF THE ALKALOIDS.-

The ground trunk bark (6.5 kg) was extracted with methanol, until the

alcoholic extract gave a negative reaction with Vassler's reagent.

Evaporated to dryness this extract yielded a brown tar, which was

triturated five times with 200 ml fractions of 2% hydrochloric acid. The

acidic solution was then continuously extracted with chloroform to separate

non-basic materials. From this chloroform, after concentration and cooling,

was obtained the soluble hydrochloride of taberpsychine (6.627 g).

The aqueous solution was then basified with ammonia and again conti­

nuously extracted with chloroform. The chloroform containing the basic ma­

terial yielded, on cooling, a black powder (1.505 g) which was filtered,

redissolved in 2% hydrochloric acid and extracted once with chloroform. This

chloroform was discarded and the aqueous solution neutralized with ammonia

then extracted with chloroform four times. The chloroform was taken to dry­

ness and the white amorphous 16-epi-vobasinic acid (454 mg) was obtained.

The chloroform solution containing the bases was evaporated to dryness

to give a semi-solid tar (61 g; total crude bases : 68.08 g)„

Part of the crude bases (10 g) was dissolved in acetone-ethanol and on

slow evaporation a fine precipitate was obtained. Filtered and washed with

cold acetone it resulted in a mixture of 16-epi-vobasinic acid and taberpsy-

chidine (300 mg). Both materials are virtually insoluble in all solvents,

but taberpsychidine is slightly soluble in hot methanol and so the mixture was

separated by boiling the solid in methanol and filtering while hot. Taberpsy­

chidine (180 mg) precipitated from the methanol solution as an amorphous solid.

120

Another portion of the crude base (25 g) was submitted to a counter-

current distribution with acetate buffer, pH 3.4 (25 ml), using chloroform as

stationary phase (25 ml), for a total of fifty transfers.

The fractions 18 to 40 yielded taberpsychine (12.7 g).

Fractions 41 to 50 yielded taberpsychidine (1.22 g).

Fractions 8 to 17 yielded a minor base whose structure remains unknown

(60 mg) (Base M).

Fractions 1 to 7 (10 g) contained only neutral material and were not

investigated further.

In all, four bases were isolated and characterized. Two of them are

completely new, one whose structure remains unknown and although the fourth

57has been prepared previously by Renner et al , this is the first time it is

obtained as a naturally occurring compound.

Taberpsychine (XLII)

The base was recrystallized three times from acetone and sublimed for

analysis, m.p. 208° (decomp.), [a]^ - 243°.

Anal. Found : C, 77.9; H, 8.0; N, 9.0; 0, 5.2 . requires:

C, 77.9; H, 7.8; N, 9.1; 0, 5.2%.

UV spectrum : Xmax 222 (33,100), 272 (sh. 6,500), 280 (7,500) and 286

(6,650) my.

peaks at 3241 (NH), 2747 (N-methyl), 744 and 728 (aromatic)

cm-1.

IR spectrum

121

NMR spectrum : 1 H singlet 8.426 (NH); 4 H multiplet 7.07-7.666 (aromatic);

1 H broad quartet 5.386, J=6.5 cps, and 3 H doublet of

doublets 1.686, J=6„5 and 2.0 cps (ethylidene side chain);

3 H singlet 2.536 (N-methyl); 1 H doublet of doublets 5.136,

J=10 and 2 cps (C.3 H); 2 H AB quartet 2.93 and 3.606, J=14

cps (C.21 methylene).

Mass spectrum : 308.1849 (M*1-; ^20^24^2^ recluires: 308.1887), 293, 279, 154,

130, 122, 121, 108, 107, m/e.

Taberpsychine methiodide.-

Taberpsychine (409 mg) was dissolved in acetone (10 ml) and methyl

iodide (0.5 ml) added while heating. The solution was boiled for a few mir

nutes and then allowed to cool. The white needles of taberpsychine methiodide

obtained were filtered by suction, washed twice with cold acetone and dried

(574 mg). The product was recrystallized for analysis from hot acetone, m.p.

272-274° (decomp.).

Anal. Found : C, 55.7 ; H, 5.8 ; n, 6.4; 0, 3.9 ; i, 28.1 . Cg^HgyNgOI requires:

c, 56.0 ; H, 6.0 ; n, 6.2; 0, 3.6 ; i, 28.2%.

UV spectrum : X^^ 222 (34,400), 276 (sh. 7,200) and 284 (7,800) my.

IR spectrum : peaks at 3170 (NH), 2830 (N-methyl) and 780 (aromatic) cm-1.

NMR spectrum : 1 H singlet 8.356 (NH); 4 H multiplet 7.00-7.876 (aromatic);

1 H broad quartet 5.436, J-7 cps, and 3 H broad doublet 1.656,

J=7 cps (ethylidene side chain); 3 H singlet 3.286, and 3H

singlet 3.036 (2 N-methyls); 1 H doublet of doublets 5.236,

J=10 and 2 cps (C.3 H).

122

3-Ethyl pyridine from taberpsychine.-

Sublimed taberpsychine (194 mg) was mixed with dried zinc dust (1 g),

introduced in a glass tube (7 mm internal diameter) and an additional amount

of zinc dust (0.5 g) placed on top of the mixture. All this was covered with

glass wool and the tube sealed under vacuum. The tube was heated in the su­

blimation block at 280°C for three hours. A yellow distillate was obtained in

the cooler part of the tube protuding from the block. The tube was allowed

to cool and then cut open. The distillate was dissolved in methanol and

analyzed in the gas chromatograph with a 4 ft. column of Apiezon L 10%. The

spectrum shows two peaks (3:1 ratio) with retention times of 6 and 12.5 mi­

nutes respectively at a column temperature of 130°C and increasing at the

rate of l°/min., which were identified by direct comparison as corresponding

to 3-ethyl pyridine and 3-methyl-5-ethyl pyridine respectively. The major

fraction was recovered in a small U tube cooled by acetone-dry ice, dissolved

in deuterated chloroform and introduced into an NMR tube.

NMR spectrum : 2 H multiplet 8.27-8.476 (H.2 and H.6); 1 H broad doublet

7.406, J=8 cps (H.4); 1 H quartet 7.03 and 7.136, J=8 cps

(H.5); 2 H quartet 2.626, J=7 cps, and 3 H triplet 1.226,

J=7 cps (ethyl chain).

Dihydrotaberpsychine.-

Taberpsychine (1.210 g) was dissolved in ethanol (100 ml) and platinum

oxide (500 mg) added together with glacial acetic acid (10 ml). This mix­

ture was left shaking for 50 hours in an atmosphere of hydrogen (50 p.s.i.).

The catalyst was then filtered off and the alcoholic solution evaporated to

dryness under reduced pressure. The resulting brown foam was redissolved in

123

2% hydrochloric acid, basified with ammonia and extracted four times with

chloroform. The chloroform was dried over sodium sulfate and evaporated to

dryness. The product (1.150 g) was dissolved in acetone and the dihydrotaber-

psychine obtained as white crystals, m.p. 191-193°.

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

C, 77.2; H, 8.4; N, 8.9; 0, 5.4 . CggHg^NgO requires :

C, 77.4; H, 8.4; N, 9.0; 0, 5.2%.

X 223 (34,100), 277 (sh. 9,000), 284 (9,800) and 293

(8,300) my.

peaks at 3271 (indole NH), 2788 (N-methyl), 745 and 726

(aromatic) cm-1.

4 H multiplet 7.00-7.806 (aromatic); 1 H doublet of doublets

5.126, J=10 and 2 cps (C.3 H); 3 H singlet 2.536 (N-methyl);

3 H triplet 0.986, J=7 cps (methyl in saturated ethyl side

chain).

Mass spectrum : 310 #), 295 (M - 15), 279 (M - 31), 265, 251, 195, 180,

170, 168, 155 (M++), 144, 138, 130, 124, 122, 108, m/e.

Dihydrotaberpsychine methiodide.-

A sample of recrystallized dihydrotaberpsychine (650 mg) was dissolved

in acetone (15 ml) and methyl iodide (2.5 ml) was added to the warm solution.

The mixture was boiled for five minutes and then allowed to cool. The amor­

phous dihydrotaberpsychine methiodide obtained was recrystallized for analy­

sis from hot acetone, m.p. 255-258° (decomp.).

Anal. Found : C, 55.7; H, 6.5; N, 6.3; 0, 3.6; I, 28.3 . re“

quires: C, 55.8; H, 6.5; N, 6.2; 0, 3.5; I, 28.1%.

124

UV spectrum : Amax 221 (55,200), 275 (sh. 10,900), 284 (11,700) and 292

(10,400) my.

Dihydrotaberpsychine-methine (XLI)

Dihydrotaberpsychine methiodide (550 mg) was suspended in tert-butyl

alcohol (15 ml) and a solution (25 ml) of potassium (1 g) in tert-butyl al­

cohol added slowly. The mixture was refluxed for 24 hours, then evaporated to

dryness under reduced pressure, the residue was redissolved in water and the

aqueous solution extracted four times with chloroform. After evaporation of

the chloroform to dryness, dihydrotaberpsychine-methine (418 mg) was obtained

as a pale yellow foam which was crystallized from acetone and sublimed for

analysis, m.p. 153-155°.

Molecular weight

UV spectrum

IR spectrum

NMR spectrum

by mass spectrometry : 324.2175 . £21^28^2^ reclu^Lres*

324.2202 .

A _ 229-5 (38,500), 271 (12,000), 284 (11,400) and 294

(sh. 9,000) my.

peaks at 3250 (indole NH), 2820 and 2760 (N-methyl), 745

and 730 (aromatic) cm”1.

1 H singlet 8.846 (indole NH); 4 H multiplet 7.00-7.646

(aromatic); 1 H doublet 6.696, J=12 cps (C.6 H) and 1H

doublet of doublets 5.556, J=12 and 8 cps (C.5 H), both

being the low field portion of an ABX system; 1 H doublet

of doublets 4.846, J=12 and 2 cps (C.3 H); AB part of a

second ABX system; 1 H doublet 4.216, J=ll cps and 1 H

triplet 3.796, J=ll cps (methylene adjacent to oxygen in

the oxyde ring); 6 H singlet 2.206 (N-dimethyl); 3 H triplet

125

0.875, J=7 cps (methyl on saturated ethyl side chain).

Mass spectrum : 324 (M+), 310, 295, 280, 266, 194, 185, 180, 169, 168, 162,

(M++), 156, 137, 130, 124, 58, m/e.

A minor product which could not be induced to crystallize was identified by

means of its NMR spectrum as the other isomer of dihydrotaberpsychine-methine

(structure XLIII).

NMR spectrum : principal feature was a pair of doublets at 4.97 and 5.635,

J=2 cps due to a terminal methylene.

Hydrogenation of dihydrotaberpsychine-methine.-

Dihydrotaberpsychine-methine (400 mg) was dissolved in ethanol (50 ml),

platinum oxide (150 mg) and glacial acetic acid (15 ml) were added and the

mixture left hydrogenating for 24 hours under hydrogen at a pressure of 52

p.s.i. The reaction mixture was then filtered and evaporated to dryness un­

der reduced pressure. The residue was dissolved in water, basified with

ammonia and extracted four times with chloroform. The chloroform solution,

after drying over sodium sulfate and evaporating to dryness, yielded a yellow

foam (340 mg). This foam was dissolved in acetone and the crystalline deri­

vative was obtained and recrystallized for analysis from the same solvent,

m.p. 184-186°.

Anal. Found : C, 77.5; H, 9.2; N, 8.4; 0, 5.1 . C^l^SO^^ recïuires:

C, 77.3; H, 9.3; N, 8.6; 0, 4.9%.

UV spectrum : Xjnax 224.5 (46,000), 279 (sh. 11,050), 285 (12,000) and

293 (10,300) mp.

peaks at 3250 (indole NH), 2750 (N-methyl) and 740 (aromatic)

cm-1.

IR spectrum

126

NMR spectrum : 1 H broad singlet 9.236 (indole NH); 4 H multiplet 6.92-

7.736 (aromatic); 1 H broad doublet 5.106, J=10 cps (C.3 H) ;

6 H singlet 2.206 (N-methyls); 3 H triplet 0.856, J=7 cps

(methyl in saturated ethyl side chain).

Mass spectrum : 326.2362 (M4"; Cg^HggNgO requires : 326.2358), 282, 281, 268,

226, 225, 180, 168, 156, 144, 130, m/e.

Taberpsychine-methine (XLIV)

Recrystallized taberpsychine (1.687 g) was suspended in tert-butyl

alcohol and a solution (25 ml) of potassium (1 g) in tert-butyl alcohol added.

The suspension was refluxed for 18 hours. The reaction mixture was then ta­

ken to dryness under reduced pressure, redissolved in water and extracted

five times with ether. The ether was dried over sodium sulfate and evaporated

to dryness to give a brown foam (1.200 g). After redissolving the foam in

ether and reducing the solution to a small volume, taberpsychine-methine

crystallized selectively (589 mg). Recrystallization for analysis was from

ether, m.p, 194-196° (decomp.).

Anal. Found

UV spectrum

IR spectrum

MMR spectrum

C, 78.1; H, 8,0; N, 8.5; 0, 5.0 . ^21^26^0 requires:

C, 78.2; H, 8.1; N, 8.7; 0, 5.0%.

X 224 (46,300), 278 (sh. 8,800), 284 (9,200) and 292.5

(8,100) my.

peaks at 3270 (indole NH), 3075 (aromatic), 2766 (N-methyl),

1590, 740 and 725 (aromatic) cm-1.

I H broad singlet 8.186 (NH) ; 4 H multiplet 6.98-7.646

(aromatic); 1 H doublet of doublets 6.40 and 6.516, J=18 and

II cps (H ); 1 H doublet 5.286, J=18 cps (H^); 1 H doublet

127

5.216, Jrll cps (Hc); 2 H doublet 5.226, J=8 cps (2 Hd) ;

1 H doublet 5.096, J=ll cps (C.3 H); 6 H singlet 2.276 (2 N-

methyl).

Mass spectrum : 322 (M+), 278 (M - .N(CH3)2), 215, 194, 183, 180, 168, 156,

136, 130, m/e.

A minor product was observed by thin layer chromatography but could not be

isolated. This is probably the other isomer formed in small quantity.

Hydrogenation of taberpsychine-methine.-

Taberpsychine-methine (429 mg) was dissolved in ethanol (50 ml) and

introduced into a hydrogenation bottle together with acetic acid and plati­

num oxide (200 mg). This mixture was shaken for 18 hours under hydrogen

(50 p.s.i.). The solution was then filtered, evaporated to dryness under

reduced pressure and the residue redissolved in water, basified with ammonia

and extracted five times with ether. The ether was evaporated to dryness

and a white foam (400 mg) was obtained. The tetrahydro compound was crys­

tallized from acetone-ether and sublimed for analysis, m.p. 153-155°.

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

C, 77.2; H, 9.3; N, 8.3; 0, 5.0 . C2jH3qN20 requires :

C, 77.3, H, 9.3; N, 8.6; 0, 4.9%.

^max 225 (29,300), 278 (7,500), 284.5 (8,100) and 293

(7,050) my.

peaks at 3250 (indole NH), 2766 (N-methyl), 740 and 730

(qromatic) cm"1,

1 H broad singlet 8.826 (indole NH); 4 H multiplet 6.97-

7.736 (aromatic); 1 H broad doublet 5.126, J=10 cps (C.3 H) ;

6 H singlet 2.356 (N-methyIs); 3 H triplet 1.256, J=9 cps,

and 3 H doublet 0.936, J=6 cps (saturated methyls).

128

Taberpsychidine (XLIX) ( = Affinine).-

The very insoluble base was recrystallized from a large volume of

ethanol and sublimed for analysis, m.p. 273-275° (decomp.).

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

C, 73.8; H, 7.5; N, 8.5; 0, 10.0 . ^2{Fi2^'p2 acquires:

C, 74.0; H, 7.5; N, 8.6; 0, 9.9%.

X 236 (sh. 13,200) and 320 (13,400) mp.

peaks at 3150 (broad, NH and OH), 2800 (N-methyl), 1650

(3-keto), 740 (aromatic) cm-1.

1 H broad singlet 9.806 (indole NH); 4 H multiplet 7.00-

7.836 (aromatic); 1 H broad quartet 5.406, J=6 cps, and

3 H broad doublet 1.686, J=6 cps (ethylidene side chain);

3 H sharp singlet 2.536 (N-methyl).

Mass spectrum : 324 (M^, 293 (M - .CH^OH), 158, 152, 122, 108, m/e.

Taberpsychidine acetate.-

Taberpsychidine (150 mg) was dissolved in pyridine (5 ml) and an

excess of acetic anhydride (1 ml) added. The mixture was left standing

overnight at room temperature. Methanol was then added to hydrolyze the

excess anhydride and the solution evaporated to dryness under reduced pres­

sure. The residue was redissolved in water, basified with ammonia and the

aqueous solution extracted with chloroform. The product is very unstable

and decomposes while in solution, so crystallization was not possible.

However IR, NMR and mass spectra indicate that it is indeed the mono-acetate

129

of the base.

UV spectrum (qualitative) : 223 (shoulder) 237 (shoulder) and 319 my.

IR spectrum : peaks at 3300 (indole NH), 1745 and 1235 (0-acetyl), 1650

(3-keto), 1575 and 740 (aromatic) cm-1.

NMR spectrum : 1 H singlet 9,806 (indole NH); 4 H multiplet 7.00-7.846

(aromatic); 1 H broad quartet 5.406, J=6 cps and 3 H

doublet 1.676, J=6 cps (ethylidene side chain); 3 H singlet

2.536 (N-methyl); 3 H singlet 1.776 (0-acetyl).

Mass spectrum : 366 (M+), 306 (M - acetic acid), 293 (M - CH^COOCH^.),

263, 194, 158, 122, m/e.

Vobasinediol (from taberpsychidine) (XLVIII)

Taberpsychidine (110 mg) was suspended in absolute methanol (25 ml)

and an excess of sodium borohydride (100 mg) added. A drying tube (CaCl^)

was placed on top of the flask and the mixture left standing overnight at

room temperature. Water (200 ml) was then added and the aqueous solution

extracted four times with ether. After evaporation to a small volume, colour

less needles of vobasinediol (100 mg) were obtained. The material was su­

blimed for Analysis, m.p. 244-245° (decomp.), [al^ - 60° (Methanol).

Anal. Found : C, 73.6; H, 8.0; N, 8.4; 0, 9.7 . £‘2^2$'p2 requires :

C, 73.6; H, 8.0; N, 8.6; 0, 9.8%.

UV spectrum : X^x 224 (21,600), 283 (6,100) and 291 (5,400) my.

IR spectrum : peaks at 3300 (NH and OH), 2880 (N-methyl), 740 (aromatic)

cm-1.

130

NMR spectrum : 1 H singlet 9.756 (indole NH); 4 H multiplet 6.90-7.626

(aromatic); 1 H broad quartet 5.476, J=7 cps and 3 H

broad doublet 1.686, J=7 cps (ethylidene side chain);

1 H broad doublet 5.276, J=6 cps (C.3 H); 3 H singlet

2.436 (N-methyl).

Mass spectrum : 326 (M+), 308 (M - H^O), 293 (M - H^O - 15), 277 (M - H^O -

.CH^OH), 183, 180, 154, 152, 136, 130, 122, m/e.

16-epi-Vobasinic acid (L)

The base was recrystallized several times from large volumes of hot

ethanol, m.p. 295° (decomp.).

Anal. Found

UV spectrum

IR spectrum

NMR spectrum

Mass spectrum :

C, 70.8; H, 6.6 ; n,

oCO ; 0, 14.4 . requires

C, 71.0; H, 6.6 ; n,

to

00 ; 0, 14.2%.

A 238 (13,100) and 316 (20,000) mp.

peaks at 3105 (indole NH), 1650 (3-keto), 1610 (carboxyl-

ate) and 753 (aromatic) cm-1.

(in 2% DgSO^/DgO solution) 4 H multiplet 6.88-7.776 (aroma­

tic) ; 1 H broad quartet 5.936, J=7 cps and 3 H broad

doublet 1.686, J=7 cps (exocyclic ethylidene); 3 H singlet

3.006 (N-methyl on quaternary nitrogen).

338 (M+), 293 (M - .COOH), 180, 166, 158, 122, m/e.

16-epi-Vobasine (from 16-epi-vobasinic acid).-

16-epi-Vobasinic acid (200 mg) was suspended in dry methanol (20 ml)

and ethereal diazomethane solution (20 ml) added. The mixture was refluxed

overnight on a water bath. The solution was then evaporated to dryness and

131

redissolved in ether. 16-epi-Vobasine (190 mg) crystallized on slow evapora

tion and was recrystallized from ether, m.p. 185-187°.

Anal. Found :: C, 71.5; H, 6.8; N, 8.0; 0, 13.8 . requires:

C, 71.6; H, 6.9; N, 8.0; 0, 13.6%.

UV spectrum :: X 237 (10,550) and 320 (12,000) my.

IR spectrum :: peaks at 3300 (indole NH), 2750 (N-methyl), 1740 (methyl

ester), 1645 (3-keto) and 740 (aromatic) cm-1.

NMR spectrum : 1 H broad singlet 9.575 (indole NH); 4 H multiplet 7.00-

7.835 (aromatic); 1 H broad quartet 5.515, J=7 cps and

3 H broad doublet 1.735, J=7 cps (exocyclic ethylidene);

3 H singlet 3.535 (methyl ester); 3 H singlet 2.525 (N-

methyl).

Mass spectrum :: 352 (M+), 293 (M - .COOCHJ, 180, 166, 158, 122, m/e.

Base M.-

The small amount of alkaloid (60 mg) obtained was recrystallized

twice from acetone, m.p. 162-164°, [alp - 129°.

Anal. Found :: C, 71.0; H, 7.3; N, 8.1; 0, 13.5 . requires:

C, 71.2; H, 7.4; N, 7.9; 0, 13.5%.

UV spectrum :: X^ax 22? (44,700), 278 (sh. 8,800), 286 (9,600) and 294

(8,600) my.

IR spectrum :: peaks at 3410 (OH), 3258 (NH), 1730 (methyl ester), 745

(aromatic) cm"1.

NMR spectrum :: 1 H broad singlet 7.905 (indole NH); 4 H multiplet 7.00-

7.525 (aromatic); 1 H singlet 4.095 (l^CHOH ); 3 H singlet

132

3.686 (methyl ester); 3 H singlet 2.126 (N-methyl); 3 H

doublet 1.266, J=7 cps (saturated methyl); 1 H doublet of

doublets 3.906, J=7 and 3 cps (H adjacent to methyl group).

Mass spectrum : 354.1956 (M+; requires: 354.1943), 339 (M - 15),

336 (M - 18), 308, 295 (M - 59), 213, 205, 194, 180, 168,

167, 154, 152, 140, 130, 122, 108, m/e.

133

-.SUMMARY.-

With the object of studying the alkaloids present in Venezuelan plants

we undertook the examination of several Apocynaceae species, of which a few

proved to contain little or no alkaloid at all. Only the plants yielding

significant amounts of natural bases are reported here, namely Aspidosperma

excelsum, A. cuspa, A. fendleri, and Tabernaemontana psychotrifolia. The

alkaloids were isolated and purified in most cases by countercurrent distribu­

tion and their structures elucidated utilising non-destructive spectral methods

wherever possible. Correlation or comparison with known alkaloids afforded a

final proof for the majority of the assigned structures.

The alkaloids encountered can be roughly divided into five structural

types, apart from a simple pyridine which was isolated as a very minor product

from A. cuspa. While evidence is presented towards the elucidation of its

structure, lack of material precludes a definite proof at this time.

Pyridine XXIX R = H yohimbine R = COŒL 0-acetyl yohimbine

In the first group are found both principal alkaloids obtained from

A. excelsum which are of the yohimbine skeletal type. The two alkaloids are

yohimbine itself and O-acetylyohimbine. Both were previously reported in the

literature, but this is the first time O-acetylyohimbine is found as a natural

ly occurring base.

Representing a second structural group are fendlerine, fendleridine,

and aspidofendlerine, all obtained from A. fendleri and des-O-methylaspidocar-

pine, obtained from A. cuspa. All of these contain the aspidospermine type

skeleton or present small but important modifications.

R = R' = R'' = H fendleridine

R = OCHg; R' = OH; R'' = CXX^Hg fendlerine des-O-methylaspidocarpine

R = R' = OH; R'' = COCHL aspidofendlerine

A third group involves three alkaloids related to picraline, all ob­

tained from A. cuspa. These are picraline itself, des-acetylpicraline and

aspidodasycarpine. The latter containing a skeleton biogenetically very close

ly related to that of picraline.

R = H desacetyl picraline

R = COCHg picralineaspidodasycarpine

135

The fourth group includes two alkaloids obtained from T. psychotrifolia

which are related to vobasine. These are affinine and 16-epi-vobasinic acid.

The latter had been prepared previously but never before reported as a natu­

rally occurring compound. Another base, which we called base "M" for purpose

of identification, is believed to be of the same skeletal group, but its struc­

ture remains unknown.

R = CH2OH; R' = H affinine

R = H; R' = CH^OH 16-epi-vobasinic acid

Finally, in the fifth group is the major base from T. psychotrifolia,

taberpsychine, which represents a new structural class derived from the

vobasine-like skeleton.

taberpsychine dihydrotaberpsychine-methine

136

It is noteworthy to mention the major product from the Hofmann degra­

dation of dihydrotaberpsychine, which was principally responsible for the

structure elucidation of taberpsychine. In fact spin decoupling experiments

in the nuclear magnetic resonance spectrum of this methine revealed the size

of, and the substitution on, the oxide ring,

137

REFERENCES

1. R.H.F. Manske (editor)

"The alkaloids", Vol. VIII. Academie Press, New York and London, 1965.

2. R.H. Burnell, and D. Della-Casa

Can. J. Chem., 45, 89 (1967).

3. R.H. Burnell, J.D. Medina, and W.A. Ayer

Can. J. Chem., 44, 28 (1966).

4. P.R. Benoin, R.H. Burnell, and J.D. Medina

Can. J. Chem., 45, 725 (1967).

5. R.H. Burnell, and J.D. Medina

Phytochemistry, 1968 (in press).

6. P.R. Benoin, R.H. Burnell, and J.D. Medina

Tetrahedron Letters, 807 (1968).

7. F. Goebel

Ann., 38, 363 (1841).

8. A. Hesse

Ber., 32, 2611 (1899).

9. A. Hesse, and 0. Zeitschel

J. Prakt. Chem., 66, 481 (1902).

10. a) M. Hesse, F. Bodmer, and H. Schmid

Helv. Chim. Acta, 49, 964 (1966).

b) D.W. Thomas, H. Achenbach, and K. Biemann

J, Am. Chem. Soc., 88, 1537 (1966).

11. W.G. Kump, and H. Schmid

Helv. Chun. Acta, 44, 1503 (1961).

138

12. B. Gilbert (ed. R.H.F, Manske)

"The Alkaloids", Vol. VIII, p. 367. Academie Press, New York and London,

1965.

13. A. Scott

"Interpretation of the Ultraviolet Spectra of Natural Products", p. 297.

Pergamon Press, New York, 1964.

14. C. Djerassi, A.A.P.G. Archer, T. George, B. Gilbert, J.N. Schoolery, and

L.F. Johnson

Experientia, 16, 532 (1960).

15. C. Djerassi, B. Gilbert, J.N. Schoolery, L.F. Johnson, and K. Biemann

Experientia, 17, 162 (1961).

16. H.K. Schnoes, A.L. Burlingame, and K. Biemann

Tetrahedron Letters, 993 (1962).

17. H. Budzikiewicz, C. Djerassi, and D.H. Williams

"Structure Elucidation of Natural Products by Mass Spectrometry", Vol. I:

Alkaloids. Holden-Day, Inc., San Francisco, 1964.

18. P. Bommer, W. McMurray, and K. Biemann

J. Am. Chem. Soc., 8(5, 1439 (1964).

19. R.E. Woodson

Ann. Missouri Bo tan. Garden, 38_, 171 (1951).

20. N. Neuss

"Physical Data of Indole and Dihydroindole Alkaloids". Eli Lilly, Indian­

apolis, Indiana, 1960.

21. F. Bohlmann

Ber., 91, 2157 (1957).

139

22. C. Djerassi

"Chemistry of Natural Products", Vol. II, p. 593 (IUPAC). Butterworths,

London, 1963.

23. N. Thi-Sen

B.Sc. Thesis. Université Laval, 1967.

24. R.E. Woodson

Ann. Missouri Botan. Garden, 38_, 162 (1951).

25. K. Nakanishi

"Infrared Absorption Spectroscopy-Practical", p. 19. Holden-Day, Inc.,

San Francisco, 1962.

26. M. Ohashi, J.A, Joule, and C. Djerassi

Tetrahedron Letters, 3899 (1964).

27. J. Lévy, J. LeMen, and M.-M. Janot

Bull. Soc. Chim. France, 1658 (1961).

28. A.Z. Britten, and G.F. Smith

J. Chem. Soc., 3850 (1963).

29. K. Biemann, M. Friedmann-Spiteller, and G. Spiteller

J. Am. Chem. Soc., £5, 631 (1963).

30. C. Djerassi, A.A.P.G. Archer, T. George, B. Gilbert, and L.D. Antonaccio

Tetrahedron, 16_, 212 (1961).

31. S. McLean, K. Palmer, and L. Marion

Can. J. Chem., 38., 1547 (i960).

32. C. Ferrari, S. McLean, L. Marion, and K. Palmer

Can. J. Chem., 41_, 1531 (1963).

33. A.F. Thomas

D. Phil. Thesis, Oxford University, 1954.

140

34. A.Z. Britten, G.F. Smith, and G. Spiteller

Chem, and Ind. (London), 1492 (1963).

35. R. Munier, and M. Macheboeuf

Bull. Soc. Chim. Biol., 31, 1144 (1949).

36. H. Schmid, J. Kerble, and P. Karrer

Helv. Chim. Acta, 32, 1864 (1952).

37. J. Ridd (ed. A.R. Katritzky)

"Physical Methods in Heterocyclic Chemistry", p. 130. Academic Press,

London, 1963.

38. R.B. Woodward

Nature, 162, 155 (1948).

39. E. Leete, S. Ghosal, and P.N. Edwards

J. Am. Chem. Soc., 84, 1068 (1962).

40. R. Robinson

"The Structural Relations of Natural Products", p. 86. Oxford University

Press, London and New York, 1955.

41. R. Robinson

"Chemistry of Natural Products", Vol. II, p. 616 (IUPAC). Butterworths,

London, 1963.

42. H. Budzikiewicz, C. Djerassi, and D.H. Williams

"Interpretation of Mass Spectra of Organic Compounds", p. 256. Holden-Day,

Inc., San Francisco, 1964.

43. R.E. Woodson

Ann. Missouri Botan. Garden, 38_, 185 (1951) .

44. C. Djerassi, L.D. Antonaccio, H. Budzikiewicz, J.M. Wilson, and B. Gilbert

Tetrahedron Letters, 1001 (1962).

141

45. B. Gilbert, J.A. Brissolese, J.M. Wilson, H. Budzikiewicz, L.J. Durham,

and C. Djerassi

Chem, and Ind. (London), 1949 (1962).

46. J.N. Schoolery

Discussions Faraday Soc., 3£, 104 (1962)„

47. C. Djerassi, H. Budzikiewicz, R.J. Owellen, J.M. Wilson, W.G. Kump, D.J.

LeCount, A.R. Battersby, and H.S. Schmid

Helv. Chim. Acta, 46, 742 (1963).

48. K. Biemann, M. Friedmann-Spiteller, and G. Spiteller

Tetrahedron Letters, 485 (1961).

49. M.P. Cava, S.K. Talapatra, K. Nomura, J.A. Weisbach, B, Douglas, and E.C.

Schoop

Chem, and Ind. (London), 1242 (1963).

50. 0.0. Orazi, R.A. Corral, J.S.E. Holker, and C. Djerassi

J. Org. Chem., 21, 979 (1956).

51. B. Witkop, and J.B. Patrick

J. Am. Chem. Soc., 76_, 5603 (1954).

52. K.S. Brown, H. Budzikiewicz, and C. Djerassi

Tetrahedron Letters, 1731 (1963).

53. K. Biemann, and G. Spiteller

J. Am. Chem. Soc., 84, 4578 (1962).

54. J. Szmuszkovicz

J. Am. Chem. Soc., 82, 1180 (1960).

55. J.M. Bruce

J. Chem. Soc., 2366 (1959).

56. H. Budzikiewicz, C. Djerassi, and D.H. Williams

"Structure Elucidation of Natural Products by Mass Spectrometry", Vol.

p. 72. Holden-Day, Inc., San Francisco, 1964.

57. U. Renner, D.A. Prins, A.L. Burlingame, and K. Biemann

Helv. Chim. Acta, 46, 2186 (1963).

58. J.A. Weisbach, R.F, Raffauf, 0. Ribeiro, E. Macki, and B. Douglas

J. Pharm. Sci., 52, 350 (1963).