Louisiana State UniversityLSU Digital Commons

LSU Historical Dissertations and Theses Graduate School

1977

Structure Elucidation by Nuclear MagneticResonance Spectroscopy.Gary Paul JuneauLouisiana State University and Agricultural & Mechanical College

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses

This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion inLSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please contactgradetd@lsu.edu.

Recommended CitationJuneau, Gary Paul, "Structure Elucidation by Nuclear Magnetic Resonance Spectroscopy." (1977). LSU Historical Dissertations andTheses. 3118.https://digitalcommons.lsu.edu/gradschool_disstheses/3118

INFORMATION TO USERS

This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted.

The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.

1. The sign or "target" for pages apparently lacking from the document photographed is "Missing Paga(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.

2. When an imagB on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame.

3. When a map, drawing or chart, etc., pert of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again - beginning below the first row and continuing on until complete.

4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced.

5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received.

University M icrofilm s International300 North Zeeb RoadAnn Arbor Michigan 48106 USASt John s Road. Tyter s GreenHigh Wycombe Bucks England HP10 BHR

77-28,683

JUNEAU, Gary Paul, 1947- STRUCTURE ELUCIDATION BY NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY.

The Louisiana State U nivers ity and A g ricu ltu ra l and Mechanical College, Ph.D., 1977 Chemistry, a n a ly tic a l

Xerox University Microfilms t Ann Arbor. M ichigan 48106

STRUCTURE ELUCIDATION BY NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY

A Dissertation

Submitted to the Graduate Faculty of the Louisiana State University and

Agricultural and Mechanical College in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

in

The Department of Chemistry

byGary Paul Juneau

B.S. Louisiana State University in Nev Orleans, ljYO August, lyYY

ACKNOWLEDGEMENT

The author thanks Dr, Norman S. Bhacca for his guidance during

the first three years of research required for this disseration and

for sharing his experience in the field of Nuclear Magnetic Resonance

Spectroscopy.

The direction of Dr. Nikolaus Fischer is also appreciated. His

help in the field of sesquiterpene lactones made Chapter possible.

Thanks are also given to Dr, Gary Griffin and his student,

Ira Lev, at the University of New Orleans for supplying the nine

furan derivatives studied in Chapter 3*

The alkaloids studied in Chapter 2 were supplied by Dr. B.

Mukherjee working at the Department of Pharmacology, B. C. Roy Post

Graduate Insitute of Basic Medical Sciences, Calcutta, India. His

work is greatly appreciated.

Laboratory assistance in the isolation of repandin was given

by Mr. Stephen Jungk and the author expresses his gratitude for this

aid.

The mass spectra used in Chapter were supplied by Donald

Lee Perry. The author greatly appreciates these.

ii

TABLE OF CONTENTS

Acknowledgement List of Tables List of Figures Abstract

I. Chapter I Introduction

A. General principlesB. Spin TicklingC. Second Order Calculations for Simple Spin SystemsD. Long Range CouplingsE. Nuclear Overhauser EffectsF. Lanthanide Shift ReagentsG. ConclusionsH. Bibliography

II. Chapter II NMR Studies of Bisbenzylisoquinoline Alkaoids

A . Int roduc t ionB. Results

1. Tiliacorine and Acetate2, Tiliacorinine and Acetatei5# Tiliamosines and Norcorinines U. Hoffmann Elimination Products

C. DiscussionD. ConclusionsE. Bibliography

III. Chapter III NMR Studies of Furan Cycloaddition products

A. IntroductionB. ResultsC. DiscussionD. Bibliography

IV. Chapter IV Repandin-A and B, Sesquiterpene Lactonesfrom T. Repanda

A. IntroductionB. Results and Discussion

1. Isolation2. Physical Data

C. Stereochemical ConsiderationsD. Bibliography

V. Vita

LIST OF TABLES

Chapter II PAGE

Table I N-Acetyltiliamosine Shift Reagent Data 8^Table II Tiliacorinine Shift Reagent Data 86Table III Tiliacorinine Shift Reagent Data 87

Chapter IV

Table I NMR Parameters of Repandin-A and B 122Table II Repandin-A and B Coupling Constants 125Table III Repandin-A and B Shift Reagent DataTable IV Repandin-A Mass Spectral Data iy,Table V Repandin-B Mass Spectral Data 1 yj

iv

Chapter I

Fig. 1Fig. 2Fig. 3Fig. i*Fig. 5Fig. 6Fig. 7Fig. 8Fig. 9Fig. 10Fig. 11Fig. 12Fig. 13

Fig. uFig. 18

Chapter II

Fig. 1Fig. 2Fig. 3Fig. hFig. 3Fig. 6Fig. 7Fig. 8Fig. 9Fig. 10Fig. 11Fig. 12Fig. 13Fig. 1/4Fig. 19Fig. 16Fig. 17Fig. 18Fig. 19Fig. 20Fig. 21Fig. 22Fig. 23

LIST OF FIGURES

PAGE

Energy Levels of AX Spectra 3AX Spectrum 5Energy Levels of Tickled AX Spectrum uTickled AX Spectrum 6Line Splitting in Spin Tickling Experiments 8AB Spectrum 12ABX Spectrum 12AA'XX' Spectrum l‘>Bond Angles In Ally lie Coupling I1;Bond Angles in Homoallylic Coupling 15Bond Arrangements in Four Bond Couplings 21Bond Angles in Five Bond Couplings ;.'lBond Arrangements in Long Range Couplings in Multinuclear Aromatics 2lRelaxation processes of Two Interacting Nuclei 2bMaximum Enhancements in Nuclear Overhauser Effects 8b

Tiliacorine - NMR Spectrum J*J*O-Acetyltiliacorine - NMR Spectrum ;7Tiliacorinine - NMR Spectrum <'8O-Acetyltiliacorinine - NMR Spectrum ;1Nortiliacorinine-A and Derivatives - NMR Spectra ,2 Nortiliacorinine-A Acetates - NMR Spectra -iN-Acetyltiliamosine - NMR Spectrum0 ,N-Diacetyltiliamosine - NMR Spectrum 8Methine-A - NMR Spectrum ■*.)Methine-A - NMR Spectrum 1Methine-B - NMR Spectrum < 3Methine-B - NMR Spectrum tJ.Methine-A - Structure t 8

Methine-B - Structure 70Tiliacorinine - Structure "flO-Acetyltiliacorinine - Structure 73Tiliacorine - Structure 77O-Acetyltiliacorine - Structure 79N-Acetyltiliamosine - Structure 800,N-Diacetyltiliamosine - Structure 81Nortiliacorinine-A - Structure ‘.20 ,N-Dlacetylnortiliacorinine-A - Structure 87ORD Curves - Tiliacorine and Tiliacorinine 88

v

PAGE

96

98

100101

105

10010 a10 i

111

11912012112Y1281551971 U17 011 5

2 ,2-Dicyano-3/3,l+ff_*dimethyl-5af-pheny Itetra- hydrofuran - NMR Spectrum2.2-Dlcyano-3jS, 1/^d ime thy 1- phenyl tetra- hydrofuran - NMR Spectrum2 .2-Dlcyano-3a,l/9“dlmethy 1- 5o-phenyltetra- hydrofuran - NMR Spectrum2 .2-Dicyano-3a»l+^-dimethy 1- 5/3-pheny ltetra- hydrofuran - NMR Spectrum2a, 98-dipheny 1-3/3) lj3-dicyanotetrahydrofuran - NMR Spectrum2/3,3o-diphenyl-Ja.lft-dicyanotetrahydrofuran - NMR Spectrum2a , 5or d ipheny 1 - 3/9 j9" d 1 c arbome t hoxy t e t ra-hydrofuran - NMR Spectrum 2a, 5or diphenyl-2 3/9-dlmethyl-30£> l+S^di- cyanotetrahydrofuran - NMR Spectrum 2/3,3/3-diphenyl-2a, 3crdimethy 1-3a, la-di- cyanotetrahydrofuran - NMR Spectrum

Gas Chromatographic Trace of Repandin TMS DerivativesRepandin-A and B - NMR Spectrum - CDClri Repandin-A and B - NMR Spectrum - CD3CN Repandin-A and B Irradiation Experiments Repandin-A and B Irradiation Experiments Repandin-A and B Shift Reagent Experiment Repandin-A and B Mass Spectral Fragments Repandin-A and B - Structure Repandin-A and B - Structure Repandin-A and B - Structure

vi

ABSTRACT

NMR studies, including nuclear Overhauser effects (NOE) , long

range coupling and lanthanide shift reagent experiments were

carried out on several bisbenzylisoquinollne alkaloids from

Tlllacora racemosa, including tiliacorine, tiliacorinine,

N-acetyltiliamosine, and nortiliacorinine-A and their deriva­

tives. The gross structures were determined, showing the

position of the hydroxy, methyl, acetoxy and methoxy groups.

The alkaloids contain two chiral centers which have opposite

configurations in tiliacorine and both of which have the

same configuration in all the other alkaloids. Direct

"proton-proton through space coupling was observed in

N-acetyltllamosine , tiliacorinine and their derivatives. Five

bond "zig-zag" and benzylic couplings were found in the Hoffmann

degraduation products. Several large NOE's were seen in each

of the samples studied.

The structures of nine furan derivatives were determined

by the use of coupling constants and nuclear Overhauser

effects. Some of the coupling constants were determined by

calculations on the AA'XX' patterns and I3C satelite analysis.

The coupling constants for trans-orlented vicinal protons

were found to be 8.0 Hz or larger and those of els oriented

vicinal protons less than 8.0 Hz. Four bond "W" couplings

were observed in three compounds and used to make structural

assignments.

III. A mixture of two sesquiterpene Lactones, repandin A and

B , was Isolated from Tetragonatheca repanda (Composltae). NMR

studies, including NOE, long range coupling and lanthanide

shift reagent experiments were used for the structure elucidation

of the two terpenoids. The two compounds represent novel

1(10) cis, 1( tj)-cis-cyclodecadlene derivatives with different

side chains attached to the medium ring lactone skeleton; an isobu-

tyrate moiety is found in repandin-A and an fy-Tnethylbutyrate in

repandin-B. These side chains are positioned at C - ,■ which was de­

duced by the differences in the chemical shifts of H- 1 in the two

repandins. The presence of a doublet at 2.00 ppm ( >H's ) , a

quartet at L ppm (1 H) and an AB pattern at 1.1 ;

indicate a sarracinate side chain. The NMR data are augmented

by infra-red spectroscopy of the mixture and mass spectroscopy

of the trimethylsilyl (TMS) and nonadeuterotrimethyIsily1

(TMS-d9) derivatives of the two repandins which were separated

by gas chromatography. The major peaks of rapandin-A-TMS

are m/e: 037, 738, 339, 17l, 170, 81; repandin-A-TMS-d'j:

652, 556, 1^7, 368, 180, 179, 81; repandin-B-TMS: u'>l, y,2,

1*62, 5b9, 171, 170, 8 1:, 81; repandin-B-TMS-dd: • , '/{'), J17I,

76B, 180, 179, 85, 81.

viii

CHAPTER I

INTRODUCTION

I. General Principles

There are several very useful techniques and tools in nuclear

magnetic resonance spectroscopy (NMR) besides the use of the commonly

known chemical shifts, coupling constants, decoupling and integration.

It is the purpose of this chapter to show how these techniques are

applied and can be used to elucidate the structures of natural and

synthetic products. Before we can discuss these principles it is help­

ful to know some of the basics of NMR as well as some of the principles

of operation of a spectrometer.

A modern spectrometer operates by the side band principle.

One fixed frequency crystal oscillator provides the carrier wave,

e.g., 100 MHz,and at least two audiofrequency oscillators are required.

In a frequency sweep experiment, one fixed frequency audio oscillator,

the manual oscillator, is used to modulate the carrier wave, e.g.,

2 cj00 Hz. This will generate two sidebands at 100 MHz + 2>00 Hz

and 100 MHz - 25OO Hz. The upper sideband is used for the lock signalto maintain a constant ratio of magnetic flux, H^, to scanning

frequency, Hi, for good resolution by keeping the TMS signal in

resonance. The second oscillator, called the sweep oscillator, varies

as the spectrum is scanned. For instance, for a 10 ppm to 0 ppm scan,

1000 Hz sweep width at 100 MHz, the sweep frequency will range from

35OO Hz to 2500 Hz and thus produce sidebands 100 MHz 4- 33OO to 2‘jOO Hz

here again the upper sideband is used. Experimentally, the band

scanning the spectrum will start at a high frequency and finish at

2

a lower frequency. This is desired since the deshielded nuclei

experience higher magnetic field and require higher radio frequencies

(R.F.) and we scan from the left side of the spectrum to the right.

If a decoupler is to be used, a third audio frequency oscillator

modulates the carrier wave, again using the upper sideband. It is

set at the absorption frequency of the signal to be irradiated or

slightly off resonance in some cases. For instance, if one wanted

to irradiate the chloroform peak one would set this oscillator at

3223 Hz which is the absorption frequency of CHCl-,.

Offsets are accomplished by changing the manual oscillator

frequency. For instance, if one wanted to scan only ‘,00 Hz,

3 ppm at 100 MHz, and one wishes the recorder to scan from 7.00 ppm

to 2.00 ppm, the sweep width would be set to 500 Hz, and the manual

oscillator at 23OO Hz. (the right end of the recorder always

representing 25OO Hz.) The sweep oscillator will now scan from

3000 Hz to 25OO Hz which is the desired range.

In a field sweep experiment the perturbing field, Hi, is

constant and is changed. Experimentally, the sweep frequency

oscillator must be used for the lock and the manual oscillator is

used to scan the spectrum, "the look frequency." The lock signal,

usually the TMS resonance, is varied from 3500 Hz to 25QO Hz and the manual oscillator is set at 2500 Hz again. The magnetic field,

Hq , changes with the lock signal. Now the lower sideband is used

in the field sweep mode because in order for the recorder to scan

the deshielded, low-field region to the high field region, the lock

must vary from a low to a high frequency while the field is changing

from a low level to a high level. As one can see the left side of

5

the spectrum is where the lower levels of are needed to cause

resonance at fixed H^. Offsets are accomplished as before by changing

the manual oscillator frequency.

II. Spin Tickling

Two very powerful double irradiation techniques to gain struc­

tural information are the so-called decoupling and "spin tickling'.'

methods. One problem associated with these methods is that with In­

ternal lock spectrometers irradiation cannot be carried out at a fre­

quency close to the lock signal. Therefore, if the spectrometer is

locked on TMS one cannot irradiate with a powerful signal at fre­

quencies less than the one that corresponds to about 2.00 ppm.This problem can be overcome by using the signal of a solvent absorb­

ing at considerably lower field as a lock signal. Another problem

associated with double resonance is that because of the large audio

frequency (A.F.) power that must be used to saturate a multiplet, a

large "beat signal" results which renders a large area of the spectrum

around that multiplet unobservable. This last problem can be

avoided by the use of the "spin tickling" technique. Since in this

method very low A.F. powers are used, a line can be Irradiated

without disturbing another line as close as 3 or Hz. Spin

tickling gives the same information as decoupling plus it has the

capability of distinguishing between positive and negative couplings

in complex spin systems. In spite of its many advantages this

technique is not used very often. The only limitation of the method

of spin tickling lies in the fact that it cannot be used in cases

where the lines to be observed are very broad.

A quantum mechanical description is useful in understanding

the spin tickling technique.'1' Considering a simple two spin system,

AX, there are four possible quantum states and four possible tran­

sitions where the change in spin quantum number is M ■ + 1,

The energy levels can be represented as shown in Fig. 1. The

arrow to the left represents the spin of the A nucleus and the one

to the right, the X nucleus. If it is assumed that there is

positive coupling between A and X, J > 0, the energy levels of the

paired spins will increase by 3/h and the unpaired spins will show

a decrease of 3fh, J being the coupling constant. Thus transitions

Ap - A a = J and X2 -Xj = J and the resulting spectrum is seen in

Fig. 2. If line Aj is irradiated at low power the elgen function

corresponding to the states Ea and E 4 will mix and two energy levels

for each of these functions will result (see Fig. 5)* Now one

observes four X transitions (E3 , E3 and E 4 , E4) instead of two (Fig. li

From Fig. 1 and Fig. J> and referring to Ai again, Xp is called

a progressive transition and X : is called a regressive transition,

since the common level of Ajl and Xp is an intermediate one and the

common level of Xi and a 2 is either an upper level or a lower one.

This is one way of defining these two terms. In frequency sweep

experiments, progressive transitions will always show line broadening

in tickling experiments and the lines of regressive transitions

will be sharper. The explanation for this will be discussed subse­

quently. This allows assignments which can be very useful in more

complex spin systems. The splitting effect will be seen on any

transition with a level in common with the transition irradiated.

5

Fig, 1 Energy Levels of AX Spectra

ft

2

Fig. 2 AX Spectrum

h 0

a, a2 x, x2

6

Fig. 3 Energy Levels of Tickled Spectrum

2

Fig. Tickled AX Spectrum, A 1 Is irradiated.

In a tickling experiment, a perfect splitting with intensities

of the two new lines being equal and their positions centered

on the original one, will only be observed if the perturbing R.F.

is within .1 Hz of the irradiated transition. If the .1 Hz

tolerance is not kept, the intensities of the two lines will not

be equal and two lines will be off center. An example for

this is given by Freeman involving irradiation near a 13C resonance

in steps of 0.1 Hz and observing an line. The equation

resulting in the position of the split lines was calculated for

the regressive transition.

tui - uurp = + £ (<JJ2-wrs) + v/ (<Jue-ujrg)2 + (2vHeXrs)r’ (1)

The symbols are as follows: oji , the line frequency observed after

irradiation; m . the frequency of the original observed line;

the frequency of the irradiating oscillator; m , the frequency

of the line to be irradiated, yx , the gyromagnetic ratio of the

irradiated nucleus; H2 , the amplitude of the irradiating oscillator;

, the relative intensity of the irradiated line. The amplitudesrsof the lines are given by

L ± * £ I *rpl2 - ^ 2 _tjUrs) / (a>2-'.crs)p + (2vxH.'Ars)3 (2)

the symbols having the same meaning as before. The splittings for

progressive transitions are given by an equation similar to that

of (1) except that the equation becomes negative, and the equation

for line intensity remains the same. Equation (1) is plotted in

Fig. 5 where the solid lines give the splitting for a regressive

transition and the dotted lines give that of a progressive transition.

8

^18* 5 Line Splitting in Spin Tickling Experiments

Progressive

U 2 ~ U r s

In the case of a regressive transition as ujs parts from id to theIT Spositive side, <jui+ parts from «j and shrinks in intensity and (.uj.-

approaches uurp and grows in intensity. The opposite is true if

(Ua departs to the negative side.

It was mentioned before that the split lines which result from

tickling are broader for progressive transitions and sharper for

regressive transitions. In a frequency sweep experiment the half

height width of a line due to imhomogeneity of the magnetic field,

Hq , can be represented in Fig. 5 by a short line of slope v^Vx*

The width of the observed line depends on the range in must be

swept vertically on the plot in order for the bold line to pass

through the curves. The line width decreases as the slope of the

curves approaches v^/v^ at fche crossing point and it can be seen

that the slopes of the curves are close to y */v v in a regressiveiV Atransition.

In progressive transitions the scale of becomes reversed

(the dotted curves in Fig. 5}. Now the slopes of the curves are far

from ya ^Yx ant* retlul-res a ™uch larger range of to sweep the

bold line through the curves. Therefore the tickling lines in a

progressive transition are much broader than those of a regressive

transition.

Spin tickling is very useful in assigning transitions in complex

spin systems. In the case of ABC spectra it is very difficult to

determine the chemical shifts and coupling constants. Frequently,

the values are estimated and the spectra computed with high speed

computers and the computed spectrum of a compound compared with the

10

observed. The couplings and chemical shifts must be adjusted until

the computed spectrum matches the observed. In these cases, good

preliminary Information can be obtained from tickling.

III. Second Order Calculations for Simple Systems

Coupling constants give much Information about molecular structure

and in most cases they are easy to determine. However, when the

differences in chemical shifts of nuclei become small compared to

their coupling constants, difficulties arise due to second order effects.

First, a brief discussion of the nomenclature of spin systems is in

order: Nonequivalent nuclei with chemical shifts comparable with

their couplings are designated A, B, C etc. Another group of nuclei

with one or more of them coupled to the first group is designated

X, Y, Z. If need arises a third group, M, L, K can also be used.

As examples, 2 ,6-dimethylpyridine, I, would have an A r»B system;

2-f luoro - 1+ ,6-dlchlorophenol, II, would represent ABX system. ;

Compounds may have chemically equivalent nuclei which are

not magnetically equivalent,^* a good example being o-dichlorobenzene,

III. Protons 6 and 3 are chemically equivalent and have the same

chemical shifts but since they are coupled differently to protons

^ and 3 they are magnetically nonequivalent. The same is true for

protons U and 5; J 5_ G = 9 Hz and * 3 Hz- We designate

pairs of chemically equivalent but magnetically nonequivalent nuclei

as AA' or XX' etc. So o-dichlorobenzene would represent an AA'BB1

system. These patterns are very common in symmetrical compounds.

The simplest second order spectrum is given by the two-spin AB

system, an example of which is shown In Fig. 6. The coupling constant

11

M e

O H

C l

(III)

C l

C l

(II)H

D

D

(D)

(IV)

antiH syn

12

Fig. 6 AB Spectrum

Fig. 7 ABX Spectrum

6 4II 9

12 1015 1 14l 1 [_.

is given by J - ^x~XJs “ u3-u4. The chemical shifts are not as obvious,

the difference in shifts, Au, is given by s/(u1-u4) (u^-us). For

very small differences in chemical shift the shifts are not the cen­

ters of each doublet pair. As the two doublets approach each

other the two center lines grow in intensity and the outer lines

decrease. When the two chemical shifts are about equal the outer lines

disappear and the center lines appear as a singlet or as slightly

split singlet. This is called a leaning effect expressed in the fact

that the doublet will always lean towards the doublet to which it

is coupled. This leaning rule can be extended to spin systems of

greater complexity.

Another common spin system is the ABX system. ABX spectra consist7of a maximum of 1^ lines. The AB portion consists of two pseudo­

quartets whose total intensities are equal. These can overlap to

any degree and individual lines may overlap giving fewer than the

expected eight lines. The X portion consists of six lines where the

two outer lines disappear as u.-u_ increases. This portion isA Dsymmetrical about uv but may lean toward the AB part if it is close

X.

to it.

The J* n value can be measured directly and should be found A- ofour times [see Fig. 7, (3-l)» (7“ 5) anc* (8-6); Lines

1> 3» T and 2, L, 6, and 8 form two quartets.]. The separation

between the centers of the quartets give the

two quantities cannot be measured directly and attempted measurements

are seriously in error if J. v - Jn v is large or u.-u. is small.A" A. JJ“ A A 5There is no direct way to obtain u.-u or (JAV - J DV), but one canA o Aa d a

11*

measure D+ and D as and w(7~5) where

D+ - i {C(V uB) + * (J*x-JBX»S + Ja I^

D. - i tC(uA-uB) - * + JAP * <5)

However, if one had to get this Information directly from the spectrum

one would find that two solutions are possible, D+ and D can be

measured but it is impossible to know which is which thus giving

two values of u -u„ and J - J . The X portion of the spectrum A d A “A D“Xhowever, will be quite different for each solution and by calculation

the two solutions can be differentiated. This process is quite in­

volved and impractical. Therefore, some knowledge of the coupling

values must be estimated before the parameters can be determined.

Compounds having AA'XX' systems are frequently encountered

and can be handled by relatively simple calculations. These systems

consist of twelve lines for each of the two identical symmetricalQ

portions, four of which form degenerate pairs. The lines are

labelled in Fig. 8. Certain parameters are used to simplify the

analysis.

K " JAA' + JX X r L * JAX " JAX'

M “ JA A ’ * JX X f N “ JAX + JAX'0 0

Some generalities can be made about AA'XX' spectra. Lines 1,2 and

3, 4 the degenerate pairs, account for half the line intensity

of each portion and the separation of these is N. The separations

(5-6) and (7-8) gives K. Differences (9-10) and (11-12) give M.

Lines 5,6 ,7,8 form a quartet centered on u and also lines 9,10,11,12.A

15

Fig. 8 AA'XX' Spectrum

Fig. 9 Bond Angles in AllyLie Coupling

<p-0

Htrans

Fig. 10 Bond Angles in Homoallylic Coupling

10

Since it is impossible to know the assignment of the outer lines of

each quartet we have the same sort of ambiguity as in the ABX system.

We must again have some estimation of the coupling values involved

before an analysis can be made.

The three spin systems just described are examples of those

that can be relatively easily analyzed. There are many other systems

that are much more difficult and cannot be so readily treated. An ABC

system is one such example. In the past these were calculated by

iterative methods using high speed computers. This involves the use

of a set of chemical shifts and couplings that would approximate ly

fit the spectrum and from calculating the energy levels expected. A

trial spectrum is calculated from these levels and compared to the

observed spectrum. If these spectra do not natch, the parameters•i li)are changed and the process repeated until matching occurs. ’

With the advent of high field spectrometers, this is not often done

today.

IV. Long Range Coupling

Spin-spin coupling in proton NMR gives very useful information.

The most commonly observed is vicinal coupling occurring between

two hydrogen nuclei separated by two carbons or three bonds (H-C-C-H).

In sp ’ hybridized carbons and non-rigid structures these usually

range from + 1. Hz to + lj Hz; the upper and lower limits quoted are

those of exceptional cases. Frequently, getninal couplings between

protons separated by one carbon or two bonds (H-C-H) is observed.

The J-values can either be positive or negative and run the gamut

from +20 Hz to -00 Hz.

Multibond coupling of nuclei through four or more bonds can

also give very useful information about chemical structure. Two

good reviews related to this topic are given by Sternhell and

Jackman . **"*”

The first case to be considered is that of allylic coupling.

Allylic coupling is defined as four bond coupling of a proton

on an allylic carbon to an olefinic proton, which can be c 1soid or

transoid (Fig. 9). Allylic couplings are negative in sign .mJ range

from 0 Hz to Hz and arc similar for both c isold and t ransoid

configurations.*"'1 The couplings are 0 Hz when 4 = O ' and 1 )’ and

are at a maximum when 4 = oo(4 be ing defined in Fig. '“)* the exact

value being dependent of the bond order of the ~ bond, Barfield et

al.*",’*'> have carried out INDO-EPT calculations of the couplings

of these systems and compared their calculations to experimental

values. For transoid couplings a Hz coupling was obtained at

4 = 90°, 0 Hz at 4 = 0°, and -O.b Hz at 4 = 180°. These numbers are

explained by contributions from a 'I electron mechanism acting in

concert with a a electron mechanism. The 9 mechanism predicts that

the coupling will have the highest negative value when the allylic

hydrogen eclipses the 9 electrons of the double bond. A number of

mechanisms have been proposed for the a mechanism which is the one

involved in four bond coupling in saturated systems. In this

mechanism the coupling has the highest positive values when the two

coupled protons are in the same plane and show the "W" configuration.

Homoallylic coupling is a five bond coupling where the center

bond is an olefinic one (H-C-OC-C-H) , Fig. D . lt Homoallylic coupling

18are positive in sign from 0 Hz to ^ Hz and similar for cisotd and

transoid couplings. They also depend on and if one of these

angles equals 0° or 180° the coupling vanishes. The coupling

mechanism involved is probably very similar to that of the allylic

coupling. An example of an extraordinarily large homoallylicITcoupling is given by compound IV . In the cis compound the inter­

proton coupling is y.Gj Hz and in the trans compound it is .08 Hz,

The reasons for large couplings are proton overlap with the ['

orbitals. In addition, there are two routes, double bonds, by which

the nuclear spins can be transmitted. Another example is given by 18compound V . Here, J 4_syn = 0.' Hz and J 4_anti = J -'-Y Hz; here,

the small values are due to the unfavorable bond angles involved.1* *Compounds of type VI offer a case of a similar type coupling.

Albriktsen at al. studied ten compounds of this type and, in general,

the four bond transiod coupling, 11J , ranges from -l.> Hz to >. '■ Hz,

4Jc from -0.7 Hz to -0.8 Hz, the five bond coupling, , between

protons on the same side of the double bonds show + >.Y l!z and J

from +1.Y Hz to +1.8 Hz, The meanings of the terms are made clear

in the figure. Here the overlap of the ’! orbitals between double

bends must play a role as well as a mechanisms.

Benzylic coupling is that between a proton on a benzylic20carbon and a proton on a benzene ring. Coupling between a methyl

group and an ortho proton gives J values in the range betwen - ).o Hz

to -0,9 Hz; to a para proton to -0,w Hz and +),-'■ Hz to +t.b Hz

for a meta proton. Benzylic coupling is postulated to operate

by a p.-<7 cumulative mechanism. When the benzylic proton shows an

(VI)

tH

(VII)

Cl

H

H

Me/ C +

\(VIII)

Me

( IX)

OH

(X)e n

(XI)

angle of 90° to the pLane of the aromatic ring,the H contribution

should be at a maximum and o electron contributions at a minimum,

VII. When the benzylic hydrogen is in the plane of the ring,21the FI contribution vanishes and the <7 mechanism predominates.

This is similar to the mechanism in the allylic coupling. However,

little comparative data is available.

The subject of coupling between four saturated bonds has

received much attention. The largest values of this "propanic

coupling" are observed when the four bonds arranged in the "W"22configuration. Calculations by Barfield jit a K predict a coupling

of +2.0 Hz for protons in the "W" arrangement, + 0.9 Hz for bonds

in a "sickle" configuration, and - 0,5 Hz for bonds in the "fork"

configuration, Fig. 11. Most of the data available in the

literature are given for the "W" arrangement.

Cyclohexanes provide examples of the above couplings. Barfield

has analyzed nineteen compounds and has found that coupling between

equatorial protons range from +1 Hz to +2 Hz, "W" coupling and

between axial and equitorial protons 0.1+ to 0.8 Hz. Between axial

protons, J values of 0.J Hz to 0.9 Hz are observed. Data obtained

on multi-deuterated cyclohexanes by Remijnse ejt al. confirmed the

previous v a l u e s . T h e s e authors found an equatorial-equatorial

coupling through four bonds of 1.7 Hz, "W", and Ji-ax,ueq WGrc

shown to be less than 0.1+ Hz. Ji-ax,3~ax also was smaller than

0.1+ Hz. The work of Hadden and co-workers^ provided further

data: The study of 3,3,5,8“tetradeutero-t-butylcyclobexane and

5 }j J* >5 ,5-hexadeutero-t-butylcyclohexane gave values of 2.70 Hz

21

Fig. 11 Bond Arrangements in Four Bond Coupling

H HH HWH H

llui IIW

H

11^ .11 For k

Fig. 12 Bond Angles in Five Bond Coupling

180Fig. 15 Bond Arrangements in Long Range Couplings in

Multinuclear Aromatics

H H

22and 1.U9 Hz for 4J ; -0.32 Hz and -O.25 Hz for 4J and for' eq-eq ax-ax4J -O.I9 Hz and -0.01 Hz. In all the cyclohexane compoundsax-eqtrue examples of "sickle" and "fork" couplings are not obtained

because the protons involved are not in the plane defined by the

three carbons separating them. More data must be acquired in this

area.

Another form of four bond coupling is that involving

protons on a methyl group. Calculations by Barfield^J indicated

that the coupling is 0 Hz when the methyl group and proton are in

a syiperiplanar orientation, - 0.3 Hz when the C-Me and C-H bonds are 9O0 apart and 0,5 Hz when they are in the anti conformation.

The latter calculated value is much less than what is observed

experimentally, 0.7 Hz and 1.0 Hz. A very interesting case of this

type of coupling is given in the l-chloro-2-methyIpropanic cation,27VIII. Here, the coupling between the methyl and H-l is rj Hz ,

a possible explanation being the delocalization of the chlorine

lone pair electrons into the coupling path.p 8Coupling through five bonds is also very common. Barfield"

and Chakraburti theorize that maximum coupling will be seen when

^ and - 180° and also that the coupling is independent of the

C2-C3 bond angle (see Fig. 12). Most five bond coupling occurs

when all the atoms concerned are arranged in a planar "zig-Zag"

pattern. The atoms in the chain can be heteroatoms such as oxygen

or nitrogen and some of the carbons can be sp hybridized. In

2-hydroxy~3 ,b-dichlorobenzaldehyde , intramolecular hydrogen bonding29exists between the phenolic hydrogens and the aldehydic carbonyl,

23

keeping the OH in the plane of the ring which allows a zig-zag

path between this nucleus and H-^ resulting in a 0.6 Hz coupling.

Six and seven bond coupling have been observed, all through

the planar zig-zag route. An example of a six bond coupling is

given by compound IX. Here, GJ*1 Hz but it must be realized that

there are two paths by which coupling can take place , thereby

increasing the J value.

Multinuclear aromatic hydrocarbons give very interesting',1coupling patterns. In a review, Bartle it jil.^ , classified the

coupling paths as peri , epi , and bay. (See Fig. 13 , phenanthrene

is shown as an example.) Epi couplings range from .3 Hz to 1.0 Hz,

bay from .3 Hz to . b Hz , and peri from .3 Hz to .6 Hz. Epi

coupling probably involves five bond zig-zag coupling augmented

by the H bonds. Peri coupling could very well be an example22of the "fork coupling" forwarded by Barfield. Bay coupling may

be an exception to the zig-zag rule if it takes place by the same

mechanism but most probably it is a case of through space coupling.

Occasionally, coupling through space is seen and this has32been observed in the NMR of several different nuclei. INDO

calculations by several authors indicate that the couplings are

negative when the two nuclei are spacially close. These inter­

actions seem to depend on the van der Waals radii and direct proton-

proton through space coupling is unimportant at distances greater_ 7;2 than 2.2 A. Hilton and Sutcliffe suggest that there exists an

angular dependence but do not specify what type.

Through space coupling has been shown to take place through

non-bonding electrons of heteroatoms and through " clouds of

aromatic rings* One of the earliest examples was given by Winstein

et al. for compound X, with a J ■ 1.1 Hz. The authors3 j D

theorized that the coupling takes place through the unshared electron

pair of the oxygen. They also made the statement that through

space coupling is negligible when two protons are involved directly

but it is much enhanced when the above mechanism is involved.

In XI , it seems that coupling through the chlorine substituent3^is involved. Here, Jp_endo,^-exo = 7*^ Hz which can be considered

very large.

There are several ways to determine the values of long range

couplings. The most obvious route is to get a high enough spectral

resolution allowing the measurement of the small splittings.

However, in most cases when large molecules are studied, the lines

are too broad. Sternhell and Jackman^ describe the "wiggle beat"

method but this approach is not practical for most spectrometers.

In this work, the half height line width at a narrow sweep

range is measured when the coupled proton is irradiated followed

by a measurement without irradiation. The difference between

the two half height widths gives a rough estimate of the coupling.

This technique requires repeated measurements under different

conditions to insure reproducibility.

V. Nuclear Overhauser Effects

A double Irradiation method that gives extremely useful

information is the nuclear Overhauser effect (NOE) experiment.

A very good review on NOE is given by Bell and Saunders,

Experimentally a nucleus is irradiated at relatively low power.

25A second proton nucleus which is spacially close experiences an

increase or decrease of intensity of the signal(s). This effect

is explained graphically in Fig. 1^. We consider nuclei I and II

showing energy levels and transitions as given in Figure lit. Whether

or not the two nuclei are coupled is irrelevant. Normal

relaxation processes occur in transitions, E 4 to E2 and E 4 to E:i

etc. but since the two nuclei are close in space the dipolar

interaction E 4 to Ej is also important. Relaxation is the process

by which nuclei come to their equilibrium spin state populations57after being disturbed by an RF field. When the signal(s) of

nucleus II are irradiated, the spin populations of E 2 and E3

become equal and also those of E? and E 4. However, the E 4 - E 2 transition tends to maintain equilibrium between the two states,

therefore increasing the populations of E2 and E3 relative to Ep

and E 4 because the energy difference between E 4 and E 2 is about

two times that of the other transitions. Since the intensity of the

signal(s) of nucleus I depends on the population differences

E 2-Ep and E3-E4 , an increase of intensity of these signal(s) will

be observed.-t O

Bachers and Schaeffer discuss the maximum theoretical inten­

sities observed in NOE experiments. By their mathematical approach

the maximum signal enhancement is calculated to be

1(1+11 v TJ----------- i±--- / , \2S (S+l) Y]L y }

where I is the spin quantum number of nucleus II and y is its

gyromagnetic ratio, S and are the parameters of nucleus I; nucleus

26

Fig. 1 + Relaxation Processes of Two Interacting Nuclei

E

Fig. lf, Maximum Enhancements in Nuclear Overhauser Effects

% E n h a n c e m e n t

50

40

30

20 H - H M e - H

_____________I______________ I____________ ‘ ■ ■I 2 3 4 5 6O

I n t e r n u c l e a r D is tance A

27

I being the one irradiated. This predicts a maximum enhancement of

50$ for proton-proton NOE's and 2X3$ f°r a 1H-13C interaction.However, in 13C spectroscopy, usually several hydrogen nuclei are

responsible for the relaxation of one 13C nucleus and since all

these 1H nuclei are irradiated, the maximum theoretical enhancement

for 13C nuclei is 299$-

Nuclear Overhauser effects are very dependent on the distance

between the two Interacting nuclei and have been found to be inversely

proportional to the internuclear distance to the sixth power.*>9Bell and Sanders' plot enhancement V£ log of the internuclear dis­

tance and obtained a slope of 6. These data were graphically express­

ed in Fig. 13 and the data for both proton-proton and methyl-proton

interactions are given.

All of the enhancements mentioned are maximum enhancements

and will only be seen if there is no other source of relaxation of

the observed nucleus except the dipole-dipole interaction of the

nucleus being irradiated. In proton NMR, this is nearly the

case In many instances and large effects are seen if precautions

are taken to minimize other sources of relaxation. Dissolved

oxygen must be removed since this paramagnetic material will allow

intermolecular relaxation. Also, the solvent must have a low

concentration of magnetic nuclei, ]H, J , 1DF, and also have a low viscosity to prevent relaxation by the solvent. Deuterated

solvents can be used.

In this dissertation, nuclear Overhauser effects were measured

in the following manner, referring to nucleus I as the one observed

28and nucleus II as the one irradiated: the irradiating oscillator

was set at a frequency at least 1.5 ppni away from II and the signals

of I were scanned. This provided a reference from which to

measure the enhancement. Then the oscillator was set at the exact

frequency of the II absorbance and I was scanned again and compared

to the previous scan. If a narrow sweep width, 100 Hz, was used,

the areas of I in both scans were determined by triangulation or else

the signal for I was Integrated during the scans. The enhancement

is the percentage increase in integrated area one. These

determinations were repeated several times to check for

reproducibility. Sometimes, different solvents or other conditions

were used.

From the preceeding discussion the informational value

of NOE data is obvious. It is a much used technique and the litera­

ture provides many examples.

VI. Lanthanide Shift Reagents

Lanthanide shift reagents can be a further useful tool in

structural studies. An organic compound with a functional group

that can form a complex with a lanthanide Ion will experience NMR

shifts with the addition of such a reagent. These shiftsfooriginate from two effects, one being the contact shift. This

involves direct delocalization of spin polarization of the unpaired

electron through the molecular orbitals of the substrate. The

unpaired electron spin density is spread over proximate atoms

of the substrate causing a change in shielding of the nuclei. This

type of shift Is independent of direction, i.e., is isotropic, and falls

29

off rapidly with distance.

In lanthanide ions used as shift reagents, the hf orbitals,

the ones containing the unpaired electron, are small in size and are

shielded from the substrate by s and p electrons. Thus, contact

terms become unimportant and the pseudocontact mechanism becomes

the important one. The pseudocontact shift is caused by the

magnetic field generated by the unpaired electron. This field

is very directionally dependent and is proportional to (': cos' 8-l)/r1

where 6 is taken to be the angle between the line from the ion to

the nucleus in question and the line from the ion to the lone

pair bearing atom, r is the distance from the ion to that nucleus

mentioned. Two mechanisms have been proposed giving expressions

for the magnitude of the shifts, one by McConnell and Robinsonh2and the other by Bleany. These hypotheses will not be discussed

in detail since it is not necessary for the interpretation of

lanthanide induced shifts (LIS).

Three factors influence the choice of the lanthanide ions

to be used in the reagent: the magnitude of the induced shift ,

the broadening brought about by the reagent, and the direction of

the shift (upfield or downfield). The most widely used lanthanides

are europium and praseodynium which do not induce large shifts

compared to the other lanthanides but give much less line

broadening. Europium shifts the majority of signals downfield

whereas praseodynium causes upfield shifts.

The chelate used in the shift reagent also plays a major

role, the most common reagents used are the lanthanide tris

30

(S diketonates) since they are air stable and soluble in organic

solvents. Furthermore, they also have simpler NMR spectra which

is very desirable. One of the first shift reagents used

was the dipivaloylmethanato complex of europium, Eu(dpm)3 , XII.The addition of fluorine to the chelate increases the solubility

of the reagent and also tends to make the lanthanide more acidic.

The Lewis acidity of the reagent allows its use with less basic

substrates, and gives greater shifting power because of the forma­

tion of a more stable complex. An example of a fluorinated reagent

and the one used in this work is tris- (1 ,1,1,2 ,2 ,3 ,3'heptafluoro-

7 ,7-dimethy loctane-1; ,6-dionato) europium (ill), Eu (fod):i, XIII.

In order for a lanthanide shift reagent to be effective the

substrate must contain a functional group which can act as

a Lewis base and form a complex with the metal. Most often

the donor atoms are oxygen or nitrogen and include alcohols,

phenols, oximes, amines, imines, aza heterocyclics and nitriles.b'5An example for a sulfur donor is the thiocarbonyl. Obviously,

some of these donors are stronger than others and it has been shown

that the order of donor power is as follows: amines, alcohols,lbketones, esters, ethers, thioethers, and nitriles. In poly-

functional molecules the major complexation usually takes place at

the strongest donor but steric interactions may cause a reversal

of major complexation sites.

In the LIS experiments carried out in this dissertation a

solution of the shift reagent, Eu(fod)3 , was added to the NMR-sample.

About I5 to 20 mg. of substrate was weighed and placed In a dry

NMR sample tube and 0.3 ml of solvent, usually CDC1;1 was added.

( X I I )

( X I I I )

A solution of Eu(fod)3 in the same solvent was prepared,

using 1+0 mg of reagent In 0.2 ml. of solvent. The NMR spectrum

of the pure substrate was taken; then 5 u,l of the reagent added and the spectrum scanned again. The reagent was added in ul

to 10 ul increments and spectra were run after each addition.

This gave a series of spectra where the signals were conveniently

followed in the downfield shift. The lanthanide induced shifts,

the difference between the chemical shift of a signal

after addition of reagent and that of the pure substrate, are

proportional to the volume of reagent added.

To determine the distance between nuclei and europium complex-2ing agent the angle factor, 5 cos 0 - 1 can be neglected if

the shift equation is expressed in logarithmic terms

log Au - log rL + log(3cos^0-1) + log K (t,)

where Au is the LIS, K is a constant.

In most organic molecules 0 will vary from (TJ to k0’> which makes

the angular term small in comparison to the distance term. So if

log vs log r^ is plotted a reasonable correlation can be obtained

A number of authors have used this method and obtained acceptable

results, but they were much improved when the angle factor was

included.

Lanthanide shift reagents have two important applications:

(a) to separate overlapping signals and (b) to obtain an approxi­

mation of internuclear distances and relationships in space.

One disadvantage of these reagents is that the signals broaden due

to the increased dipolar relaxation Induced bv the magnetic

dipole of the paramagnetic ion. In some compounds the broadening

effect seems to predominate and the experiment becomes useless.

Examples for this will be discussed later in this dissertation

(e.g., methines A and B). Other compounds give relatively little

broadening and small couplings can be seen even after the

signals are shifted several pptn downfield, as will be shown in

later discussion for repandin - A and B and tiliamosine. In these

cases, Euffod)'. proved to be very useful in separating overlapping

signals. Determining internuclear distances appears to be a

simple matter from the above equation, however,the treatment

applies only to a mono functional substrate. When the substrate is

polyfunctional, contributions due to complexations of all the

groups must be considered which makes analysis impractical

VI. Conclusions

Some of the more useful and more esoteric anc Hilary methods

of NMR have just been described. The ability to obtain chemical

shifts and coupling constants from second order systems has allowed

the assignment of the stereochemistry of the eye loaddi t ion pro-

ducts as will be shown in Chapter -. Spin tickling was used

to link coupled protons and the technique allowed the assignment

of the signals in the repandins in Chapter >■ and in selected

examples of alkaloids in Chapter 1'. Nuclear Overhauser efiects

giving internuclear interactions between spacially close nuclei,

were used extensively in tin* structure elucidations of the alkaloids in

Chapter These structures cound not have been doLormiiU'd

without tills method. bong range couplings arc used t o give infor­

mation of molecular geometry and proved to be useful in assigning

5^signals of both repandin and tiliamosine. Although it is

difficult to obtain internuclear distances from LIS's, some informa­

tion becomes available with their use. For instance, some of

the signals of repandin-A B were separated after the addition

of Fuffod)-.. The application of these methods will be discussed

in greater detail in subsequent chapters.

55Bibliography

1. R. Freeman, W. A, Andersen, J. Chem. Phya., 2L* N o » 9, P. 2055, (1962).

2. R. Freeman, J. Chem. Phys. , j+0, No. 12, p. 557b, (I96I).

5. R. Freeman, W. A. Andersen, op. cit. p. 2055-

1. F. Bovey, Nuclear Magnetic Resonance Spectroscopy, AcademicPress, New York, p. 'jl (1 ,>Uj) .

5. J. A. Pople, W. G. Schneider, H. J. Berstein, High ResolutionNuc lear Magnet 1c Resonance , McGraw Hill, New York , ( 1'.<)).

b. F, Bovey, op. cit. p. >8 .

7. ibid, p. 10S

8 . ibid, p. 117.

9. d. D. Swallen, "Computer Techniques in the Analysis of NMR Spectra", Progress in NMR Spectroscopy, Ed. Elmsly, Feeney,Sutcliffe , Vol. 1 ( i Q b b ) , p . 205-

10. R. Lustig, "Computer Assistance in the Analysis of High Resolution NMR Spectra," NMR Basic Principles and Progress, Vol. <■, Springer Verlag, New YorkT, (1972) .

11. S. Sternheil and L. M, Jackmann, AEEllS. at ions of Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry, Pergamon Press,Oxford , 2nd Ed . (I9G9 ) p^ ;’> 1 *■.

12. S. Sternheil, Quart. Rev. , 25, p. 25s (1909).

I5. S. Stermhell and L. M. Jackmann, loc. cit.

Lb. M. Barfield, A. M. Dean, C. J. Fullick, R. J. Spear, S, Sternheil,P. W. Westerman, J. Am. Chem. Soc.. 97. p. 1jj82 (l:)7s),

15. M. Barfield, 8 . Chakraburti, Chem. Rev. , 69, p. 757, (I969).

16. L. Sternheil, Jackmann, op. cit.

17. E. W. Garbisch, M. C. Griffith, J. Am, Chem. Soc., J 2 , p. 5590, (1968).

18. S. H. Grover, J. B. Stothers, J. Am. Chem. Soc., P* b55l (1969).

19. P. Albriktsen, A. V. Cunliffe, R. K. Harris, J. Mag. Res., 2,p. 150, (1970).

3620. Sternheil and Jackmann, op, cit.

21. Barfield, Chakraburtl, op. cit.

22. M. Barfield, Dean, Falllck, Spear, Stennhell, Westermun, op, cit,

25. Barfield, Chakraburtl, op. cit.

2k. J. Dm Remijnse, B, G. M. Vandegnste, B. M. Wepster , Rec.Trav. Fays Bas, 9 2 , p. 80l, (1973).

25. V. R. Hadden, L. M, Jackmann, Org. Mag. Res. , _S, p. 933, (1973).

26. Barfield, Dean, op. cit,

27. G. A. Olah, J. M. Bollinger, J. Am. Chem. Soc. , 'JO, p. \.)k'{, (I9G8 ).28. Barfield, Chakraburti , op. cit.

29. T. H, Siddall, W. E. Stewart, Chem. Comm. , p. Ill* , (I908).

JO. J. E. Baldwin, R. K, Penschmidt, J. Am, Chem. Soc. , J2, p. r;2ii7 (1970).

31. K. D. Bartle, D. W. Jones, R. S. Mathews, Rev. Pur. Appl. Chem. ,12, P . 191 (1969).

32. J. Hilton, L. H. Sutcliffe, "The Through Space Mechanism in Spin-Spin Coupling," Progress in NMR Spectroscopy, Vol. 10,Part 1, (I975), P- 27-39.

33. S. Winstein, P. Carter, F. A. L. Anet, A. J. R. Bourn, J. A m .Chem. Soc. , 87, ). 32^9, (I963).

3^. K.. Tori, M. Otsunu, Y. Hata, H. Tanida, Chem. Comm. , p. IO96(1968).

39. R. A. Bell, J. K. Saunders, "Some Applications of the Nuclear Overhauser Effect," Topics In Stereochemistry. Vol. 7. Ed. Ailinger, Eliel, Interscience Publishers, (1973) P* 1*

30. "The Measurement and Interpretation of the IntramolecularNuclear Overhauser Effect in Proton NMR Spectra," Perkin Elmer

NMR Quarterly, No. U (June, 1972).

37- See any NMR textbook for an explanation of nuclear relaxation.(Viz. F. Bovey, or J. A. Pople, W. G. Schneider, H. J. Bernstein.)

38. G. Bachers, J. Schaefer, Chem. Rev., 71» No. 6, p. ol f, (1971).

39. Bell and Saunders, op. cit.

37IiO. A. Cockerlll, G. Davies, R. Harden, D. Rackham, Chem, Rev, , 73,

No. 6, p. 553, (1973).

1*1. B. C, Mayo, Chem. Soc. Rev., 2 (1), p. ^9-7^, (1973).

1*2. Reviewed by Cockerlll et al. and Mayo, op, cit.

1*3. Cockerlll et al., op. cit.

1*1*. Ibid, p. Lj7t).

CHAPTER II

NMR STUDIES OF BIS BENZYLISOQUINOLINE ALKALOIDS

I. Introduction

This section is concerned with the structure elucidation of

several bisbenzylisoquinoline alkaloids derived from Tillacora

racemosa/ a woody climber that grows wildly in India. An

isoquinoline alkaloid is one that contains the 1 , 2 , J;- tetrahydro-2isoquinoline moiety, a simple example of which is (+) salosine,

I . The term "nor" when used in connection with these compounds

refers to the absence of a methyl group on a nitrogen atom.

A typical example is nororientallne, II, a norbenzylisoquinoiine

alkaloid, whereas orientaline would contain an N-methyi group. J

The bisbenzylisoquinolines have two benzylisoquinoline moieties

joined in a variety of fashions, two of the twenty-one different

types being represented by the cisampereine type III , and

the berbamine type IV . The alkaloids in Tlliacora racemosa are

of the tiliacorine type, structures of which will be presented later

The alkaloids discussed here are tiliacorine, tiliacorinine,

and their O-acetates; nortiliacorinine-A, its N-acetate and

0 ,N-diacetate; N-acetyltiliamosine and its O-acetate,

which were extracted from air-dried, powdered roots with ethanol

containing 1% acetic acid. This extract was concentrated and

the residue re-extracted with aqueous 1$ acetic acid. This acid

extract was again extracted with ether, cooled and basified with

amnonla providing alkaloid material as a brown gum which was

chromatographed over neutral Alr.OM . One of the fractions yielded ,

'59

HO

H3C0H CH

(I)

HO

HOOCH

(H)

HO

OH

(in)R

ko

after evaporation, a pale brown glass which was dissolved in acetone

and cooled to 0°C, to give tiliacorine as a precipitate. The

mother liquor was evaporated and further fractionated by a JO stage

counter current distribution. The first six fractions contained

a mixture of tiliacorine and tiliacorinine, the next twelve

contained nortiliacorinine-A and B. Tiliacorine and tiliacorinine

were separated by chormatography over Al^O^. The nortiliacorinine-A

and B were separated by counter current distribution using Ethyl

acetate and pH o.b Buffer followed by chormatography over Al 0;,

of the appropriate fractions.

The leaves of the plant contain tiliamosine, its N-acetate,

and N-acetylnortiliacorinine-A. Also other alkaloids , tiliacine,6 7 8corine, mohinine, tiliacoridine, and tiliarine have been extracted

from various parts of the plant. Nortiliacorinine-A and nor-

tiliacorine-A (synonomous with isotiliarine) have been obtained9from Tiliacora funifera.

The structures of tiliacorine and tiliacorinine have been

partially determined by chemical degradation,^ O-Methyltiliacorine

dimethiodide was prepared by adding methyl iodide to the alkaloid

in boiling methanol. Oxidation of this product with alkaline

permanganate gave the diacid V , thus Indicating that the

isoquinoline moieties are linked by a diphenyl system instead of a

diphenyl ether system such as in IV. A tetracarboxylie acid was

also obtained by the above oxidation and this product was shown by

mass Bpectrometry to have the formula C^yHioOji. The alkaloid gave

a blue color when treated with a mixture of concentrated sulfuric

1+1

and nitric acids Indicating a dibenzo-p-dioxin system.^ The

structure of the tetracarboxyllc acid was shown by comparison to

a sample of the known tetracarboxyllc acid, VI. These pieces of

information together with mass spectral evidence

indicated the structure to be VII .

Both tiliacorine and tiliacorinine form the derivatives

’'methine-A” and "methine-B" in ratios 7:2 and 1:3 respectively

when the respective O-methyldimethiodides undergo Hoffmann12eliminations. These "methines" were shown by NMR to be isomeric

and have the gross structure VIII . This indicates that the

two alkaloids are diastereomeric.

The synthesis of d,l-0-methyltiliacorine has been reported12by Anjaneyulu et al, Basically it involves formation of the

dibenzo-p-dioxin system by an unsymmetrical Ullmann reaction. Then

addition of the biphenyl moiety is accomplished by reaction of acetyl

chloride side chains on the biphenyl with ethyl amine chains on

the dioxin system. The tetrahydroisoquinoline ring is formed by12a Bishler-Napieralski cyclization.

The potential medical uses of these compounds prompted this

study. It remained to determine the positions of the side chains and

OH-groups and to obtain information concerning the stereochemistries

of the two asymetric centers of the alkaloids and their acetates.

NMR studies of the Hoffmann elimination products were done to provide

insights into the relationships of chemical shifts, NOE’s, and coup­

lings to the structures of these alkaloids.

HOOC

COOH COOH

MeO OMe

(2H

OMeCOOH

O O H

(YD

COOHOMi

MeO OMe

II. Results

1. NMR Studies of Tiliacorlne and Its Acetate

The NMR spectrum of tiltacorine and accompanying experiments

are shown in Fig. 1. In deuterochloroform one meta-doublet, (a meta­

doublet is the doublet resulting from a coupling of a proton meta

to the proton in question) two ortho-doublets, three singlets,

and a series of overlapping signals containing the CHC1-, signal, two

o,m-doublet of doublets, and one m-doublet appeared in the

Fig. Tiliacorlne\

c.oZI-OM* I Z - O M <

4-n m i t r ■ n m i

H - 7,14. 2 0

rrM24 H l j 1 H27 *1} M 10

H 2 i I iKl - L

ia oo

l i7 00 $ OO 5 00

M-t«a

4 00 3 0 0 2 00

| J.mn.,I tf 7 II Dtcouplmi

Nimtl i fjci I f f 7 31

' < i i!

oo 7 OO e 50

i ill,ii; I iWi .

Normal

If. 2 ■■S t ) l1

T l l i a c o r l n #

2 m"' >0 if I20 f, -V

t r t 3 9 8 | | KO i | r f 3 8 5 » 0£

I M ! I Hi 1

m 1 *

t,,0 T? 2 b ?S \k o )s

- „<v p Q ; i V~»OM. OH

C DC I 5

IOO M M l >11)1 3 0 0 MM)

I k 4 0 4

vvO*

$ l*~ MOC J M -i 1 IF « ib. I • 1

12 OMa i * ' 0M*

! I4 NM«

, w J > “’ ■H 160 , j

•vu;

.0,3H I S ' M I S

M- T, 1 4 , 24 H 23

H II

(*- 8 00

j I H e|HI3 M-

J lS 17

27 H 2 0

:v 'J u-Vv_V i—. ■r I - T-4 10 S 4 0 t4H l(o

1 'J'v .J OO 7 * 0 '

UH 4 0 t

.‘W'"6 do I

A 001

3 OO do

aromatic region. Upon Irradiation of the methoxy signal at 3.85 ppm

and 3.95 PPm nuclear Overhauser, NOE, effects were observed at

the high field aromatic singlet, and the high field o-doublet,

respectively. This identified the singlet as H-20 and the ortho­

doublet as the proton at C-I3 next to the methoxy1 group at C-12

of the biphenyl portion of the molecule. In an expanded spectrum

of the aromatic region, the signal at lowest field, a meta-doublet,

was irradiated and decoupling was observed at the partially visible

higher field o,m-doublet of doublets. Then the lowest field

line of the partially observable lower field o,m-doublet of doublets

was irradiated at low power and a "tickling effect" was observed

at the higher field o-doublet. These experiments established

the chemical shifts of the protons on the biphenyl system.

At this point there still existed a question of whether

the biphenyl moiety Is attached as shown in Figure 1 or whether

the system is reversed, that is OMe being at C-y and -OH at C-12.

Irradiation at 3-39 PPm resulted in a 25$> NOE at aromatic

singlet at 6.99 PPm *> indicating that the signal at 3*39 Ppm

must be due to H-Ua and the aromatic signal represents H-23. A

300 MHz spectrum of tiliacorlne was taken at the University of

Akron, Akron, Ohio and a part of the aliphatic region is presented

in Figure 1. By double irradiation experiments, (not shown

here), H-l6a, H-16, H-16', and H-Ua, H-r;, H - 1)' are readily descerni-

ble. At 100 MHz, irradition of the signal at 2.81 ppm, due to

H-16, showed a long range coupling at the lowest field m-doublet.

This indicated that the lowest field doublet represents H-28,

146

the low field o-doublet H-8, and the high field o-doubiet H-I3 .

The low field methoxyl signal is due to 12-OMe. By a process of

elimination the signal at 6.68 ppm was assigned to H-27, The

N-methyls were assigned by comparison with the other alkaloids

which completes the assignment of all signals.

The irradiation of l+.Oy ppm, H-lGa, resulted In a l'/'y NOE

at H-28 showing the proximity of the two protons. Due to the

overlap of H-29 with other signals and its similar chemical shift

to H-25, the very informative experiment, the irradiation of H-2S

and observation of H-29 could not be done.

The NMR signals of O-acetyltlliacorine, Fig. 2, were assigned

by comparison with the spectrum of tiliacorine. The signals

due to H-27, H-20 and the methoxy groups did not shift. The

remaining aromatic signals were assigned by Irradiating H-26 and

observing decoupling effects at the low o,m-doublet of doublets

and inspection of the leaning patterns. The aliphatic region

of the spectrum is very similar to that of tiliacorine

except for the additional 9~0_acetyl signal at 2.1)i ppm. Upon

irradiation of the H-28 signal at 7.68 ppm a 12$ NOE was observed

for H-25 demonstrating their proximity across the ring,

2. NMR Studies of Ti1iacorinine and Its Acetate

The NMR spectral data of tiliacorinine are shown in Fig. %

All the aromatic signals are reasonable well separated in

deuterochloroform with a slight overlap of the two o,m-doublets

of doublets. The aromatic signals were assigned as follows.

The low field m-doublet at 7.^8 ppm was irradiated and decoupling

hi

Fig. 0-Acetyltiliacorlnei iH 0M» | if OW* 4 NUt *-0*»

I I T- NWl |

M- T, 14. I t M - I T _ H 2J>M IS I H 10

r,

B 00 7 0 0I6 00

H 16 0VWY',V*VVV**V''rf'I4 00 3 0 0

J i ...L Vj r >* L' ' V V V

2 00

0* Ac ttyllll ig c o m n ■out21 ! 50 IT 2

17 ' .6)-A*W1 -5 i -s /* 6tW» Oftc

f r r 7 6 6 D*#<I i

M-9* it. •MI* M 17 M ,* I M ill H I 9 H 1 0I I**•/ M’l

v j r ~iv *-a M V " W W V I6 00 ? 00 6 00

?l OMt 1 12 OM*

H 16a: i. V j v

4 00

4 NUtU

9 0Ac

NWi

■*r / 'v s j ^1 00 ? oo

U8

Fig. Tiliacorinine

lH T

* 23 H21 H'*.lH 20

H ??Mljl *

/J U A v L J >i0 OO I

7 OOI6 OO

C.o,e"«

21 OM*

I*2 OM* <-NM*

H 16*

4 OOI

3 0 0 2 00

N s r n i l

N o f m o l

In 2 0 0, t> e p I

I f t 3 3 7 I

. i * ‘ i nil ! ! I;( . l i III

, ^ A >

Tilfac orinin#

'V 7 ’ 29

‘ ■ J V w w L ■> V

-v6O ' - O ’I jTA ^ B

QftM OH

I t f 3 9 9 I " 5 * 3

U NOE

‘ l|w * A

if* r 40 .....11

T«a ‘In_ i > . V

"of I t Jo/.I Iv v.* ‘■'b Al

Ih aH 23 H 29 M 7 I ►+ 13 ! M 28 ; M 14 j . ' H 27 h 20

' ii 1C 1 1

i8 00 7 OO

M \6o' U3 60

6 00

cocijt O O M M i n i l i 300 M H 1 I n n ’ t

<2 O M * 71 ° M*

I 17- M M *

H - * 0, 3.3 '

iM 16' M 16 I

Vv A /. ■ * >'

13 40 3 OO 2 00

I

* '"N ** \N \p 3 1v \n- ■/ '» ■' L4 00 1 00 2 00

U9occured at the lower field o,m-doublet of doublets at 7*35 ppm.

When the lowest field line of the lower field o,m-doublet of

doublets, J.hO ppm, was irradiated and tickling was observed at

the high field o-doublet at 6.95 ppm* The methoxy signals

at 3*98 and 3.85 ppm were irradiated with nuclear Overhauser effects observed at the high field m-doublet, 7 .78 ppm, and high

field aromatic singlet, 6 .3O ppm, respectively, thus identifying

the H-20 signal and the biphenyl proton signal next to the methoxy

group.

The arrangement of the biphenyl moiety in tiliacorinine was

determined in a similar way as described before for tiliacorine.

At 300 MHz, H-16a , H-16, H-161 and H-i+a, H-5 , H- 3' could be

identified by inspection and decoupling experiments. At 100 MHz

the position of H-^+a was determined by irradiation at 3.;)7 ppm

resulting in a 15$ NOE at the lowest field aromatic singlet, H-25.

Irradiation at 2.80 ppm, H-16, resulted in a long range decoupling

at the high field m-doublet, 7.58 ppm, identifying it as H-28.

Now all the aromatic signals can be assigned. The location of the

9-OH was deduced by integrating the aromatic region followed by

addition of DpO and reintegration.

When H-25 at 8.05 ppm was irradiated, a 10$ NOE was observed

for H-28, and a 20$ NOE for H-29, 7*58 ppm, but most surprisingly

a large decoupling effect was seen at H-29 (see Fig. 3)*

This appears to be a coupling through space, since it would be

highly unlikely that coupling could take place through seven

bonds arranged in the shape of an unclosed heptagon which has

very little likelyhood of being planar. No nuclear Overhauser

effect was observed for the H-28 signal when H-16a was irradiated

suggesting no close proximity of the two protons. The two rings

of the biphenyl system are peri planar.

The spectrum of O-acetyltiliacorinine in CgDg and CDCl3 are

presented In Figure 1. All aromatic signals are clearly visible

except H- lb and H-7, the two o ,m doublet of doublets which are

completely overlapped. Irradiation of the methoxy signals around

3.83 ppm resulted in nuclear Overhauser effects at the high

field singlet at 6.29 ppm and the higher field m-doublet at

6.85 ppm allowing the assignments of H-20 and H-I3 , respectively.

The broadened m-doublet was assigned to H-29, by analogy, to the

parent compound. Irradiation of the low field singlet at 7.29 ppm

caused similar Overhauser effects on the H-28 and H-29 signals

and also resulted In the decoupling (sharpening) of the

H-29 absorption at 7.61 ppm. When the lower fields m-doublet,

7,65 ppm, was irradiated an Overhauser effect was observed at the

H-25 signal. Some difference between the aliphatic region in

the NMR spectrum of tiliacorinine and its acetate is probably

due to differences in the biphenyl ring conformations. This

is probably due to increased steric interaction of the C-9

and C-12 substituents.

3 . NMR Studies of Tiliamosines and Norcorinlnes

Due to the low solubility of nortiliacorinine-A in deutero-

chloroform, a mixture of 20% DMS0-d6 In deuterochloroform

was used to obtain the spectrum shown in Fig. 5- There is only

one N-methyl present In nortiliacorinlne A signal unlike

51

Fig. O-Acetyltiliacorinine

H-llH-t! wr

y u ^i3 OO 2 OO4 OO6 OOT O O• OO

0.Acatviiiii4cof*n*n*

Ifr. 1 §3M O tt t

■ at

iff.r to

13 ,lXa„oZ«

CDCt j (00 rn*l 300 MH.i hoi o*t' 4 MM* i r ' MW*

I,I r r T. • 1 ■at , ■ at

S «0

I« OO

17 00e oo

52

Fig. Nortlllacorlnlne-A and Derivatives

O . N ' S l a c a t y l n o r t K i d t o r i n l n S ' - A

corfJi

M-20H ISl *24 I H 27

i t l l t - O I A C* D «

I T - N t i O

H-401

tH4«

a**rAV > v ' Wi v j W i,

S O A C • 4 ~NAc

e oo 7 0 0 6 5 0 5 0 0 4 00

2I-0M*

t3 0 0

CDCI,2 00

iz-om# 9 0 A t

N - 3 ' M-3 I ] H -16 1 M - i « '

3 0 0 H H t

4 - N Ac' 1 7 - K M *

t " w 1 v V !

H 24*•*•*« < n . M-ir” mi °

H- T 1 \ J H -14 1\ jH- 4 a

3 20 1 2 10

6 00

I rr 3 01 N O I

I7 0 0

N o r m a l

6 00 5 00

I r r 3 4 2, HOE

I I4 0 0 3 O O

N or ti I i oc or Ini n * - A

ON*

2 00

A jJ X x L L AiIrr 4lt

N O t

« A fior m0 I

"Ti'v %-{0U< OHCOC lj A 20 % OMSO-dS

I N O t

M 20M|3 *-2?IG"

X y \A L v0 00

I I7 00 6 50

2E0MiI 7- NM«

4 50 4 00 3 00 2 00

55the preceedlng two compounds. Many of the aromatic signals overlap

making assignments difficult and allowing only limited nuclear

Overhauser effect studies. The positions of the H-I3 and H-20

signals were determined by Irradiations of the methoxy signals

at 3.82 ppm and 5*91 ppm. H-la and H-25 were assigned by irradiating

1+.19 PPm and observing the NOE at 8 .O5 ppm, H-25. Also nuclear

Overhauser effects are observed at 7.^8 pptn and 7.39 PPm when H-25

was irradiated thus identifying H-28 and H-29 as a pair. H-27

and H-8 , the only remaining clear signals, were assigned by a

process of elimination.

The spectra of 0 ,N-diacetylnortiliacorinine-A are also

shown In Fig. 5, The signals in this case were assigned by compari­

son with the spectrum of N-acetyltiliamosine, a very well studied

compound of great spectral similarity. In Fig. 6 are shown 0,N-

diacetyltiliacorinlne-A again and also N-acetyltiliacorinine-A

for comparison. The later compound is almost identical to N-

acetyltiliamosine.

The spectrum of N-acetyltiliamosine is shown in Fig. 7.

All of the aromatic signals are well separated. When the high

field m-doublet was irradiated, 7.^5 pp™) decoupling was observed

at the lower field o,m-doublet of doublets at 7.92 ppm. From

the leaning pattern of the two o-doublets it is obvious that

the high field o-doublet is coupled to the high field o,m-doublet

of doublets. When the methoxy signal at 3*99 PP™ was irradiated

a large NOE was observed at the high field o-doublet at (>.90 ppm.

So this assigns the protons of the biphenyl system. The position

5U

Fig. 6 Nortlllacorlnlne-A Acetates

1 790*o

C DC I j

k■ i i

4 0 0 J 0 0 Z 0 0

N'octfylnorl Miocorlntn* - A

_/v

8 00I

7 0 0 6 00

Ol i *2> 10 i7 2ior ---> VA. . A j!SV -4 ■''*•>'*> *V” Y* '> ‘T„ ^

,Th ->1 ? *» i>*' n{7 )>■_■*-■ Mr\ v - ' V. . ' 'S > * «jm • f h

COCI,

! f « . ' ".'■y 4 i ■ 1 V *l ij **, jVJ . ^ - J Lr W. V

I8 00 7 OO 6 OO

-A\~■S 00 4 on *1 0 0

I2 OO

55

Fig. N-AcetyltiliamosineN' AeityltHiam«irA#

c«°«I*. !!

. V w V V 4 i ^i |

I |l'

IS 0M«

S O t t t - O M *

I T - K M *

4- WAt

M S S H S *nsa H-ISM-r I **• | 'i *13 *-*s M-*«

aJ O c ljU -J .! iy ^ / v V > ' \A/ L _

tS 00

I I7 00 6 50 5 00 4 OO 3 00 2 OO

i * i t a s

JiXJi>U ,W I L<

I1.* sat !U

I i

J KO I

i K « r w « l

I AI

N - A c a t f l t m o m c t i n t

OM*HI 50 t r

,‘j** jx yl :*

UViV^l

litOK* OH

I*lit 3 60

I'rSOB f fl

! *N #r m 11

‘ sMiu’v

a tii 4I {

in M l t k ,, [ill ! | 1 i A 14a |J «J1J ^ Vu v

| P i n *91 \

M 'IS* 1

C D C i 3*00 MHt -111 300 WHi Imtrt

Ni MS’ IS | H I S '

) 1 K

J I 'wv-v**, V A ^ v '"A \ A3 60 L 3 SO

rtf 3 S5 I S O M . T- s o a si - o m *

’ I n 2 1 0

M Z S l ! H I4 rL r M |3 |

i- t i sj 111/; IAa ;. /iaM.'JU jit. . V Jai ' I

S 8 04 NAc

a IT - N M *I

TJH 40

8 00 7 00 6 00 5 00 a 00i

3 OO 2 00

of this system on the rest of the molecule was determined similarly

as previously demonstrated in tiliacorine and tiliacorinine.

Again, this involved initial finding of the H~5 H-16 signals

by irradiation of the H-Ua and H-l6a signals which are clearly

visible at 300 MHz and at 100 MHz* When the spectrum was irradiated

at 3 .O5 ppm, H-5, decoupling was observed at the low field o ,m-

doublet of doublets at 7.92 ppm and at the high field m-doublet

at 7 * ^ ppm. Irradiation of H-16 at 2.93 PP1*1 > the l°w field

m-doublet at 7-70 ppm sharpened.

There are three methoxy signals in the NMR spectrum of

N-acetyltiliamosine, one of which is due to the C-12 methoxy group.

The others could be situated on any two of three positions

C-27, C-20 or C-21. These signals were assigned by comparison

with NMR spectra of the previous compounds. H-27 resonates

always around 6.6 ppm, H-20 around 6.2 ppm. So because there is

no signal present around 6.2 ppm and there is one at O.G'i ppm it

can be assumed that the third methoxy group is attached to

C-20. This was corroborated by the fact that irradiation of

higher field methoxy signals resulted in no NOE.

WhenH-16a was irradiated a 15$ NOE was observed for the H-28

signal at 7-70 ppm. Irradiation of H-25 resulted in 15$ and 20$

NOE's at the signals due to H-28 and H-29, respectively. Further­

more, decoupling was observed for H-29, again showing the trans-

annular proximity of these protons. The H-23 signal also showed

a 20$ enhancement when H-^+a at 5-1& PP™ was irradiated showing that

the two protons lie in a nearly peri orientation.

57

N-acetyltiliamosine has one less N-methyl group than tilia­

corine or tillacorinine. Since H-7 Is shifted nearly 0.75 PPm

downfleld It can be assumed that the acetate Is attached to

N-1+. The anisotropic effect of a carbonyl group would account

for the large downfleld shift. This same effect is also seen

in 0 ,N-diacetylnortiliacorinine-A as well as in 0,N-diacetyltilia-

mosine. It should be noted that the chemical shift of H-7 in

nortiliacorlnlne-A is 0.5 ppm upfield from the N-acetate.

Therefore, the N-methyl signal at about 2.25 ppm is due to the

17-NMe. In tiliacorine and tillacorinine and their acetates

the two N-methyls resonate at about 2.25 Ppm an^ 2.05 PP™* By

comparison with the spectra of the latter compounds the former

resonances are due to 17-N-methyls and the later are due to ii-N-

methyIs.

The spectrum of 0,N-diacetyltiliamosine is shown in Fig. 8 .

The two m-doublets due to H-28 and H-29 overlap in duetero-

chloroform solution, but can be separated by adding deuterobenzene.

In this mixed solvent system the o,m-doublet of doublets at 8.O5 ppm

was irradiated causing decoupling of the high field m-doublet at

7.^5 ppm. The other signals due to the biphenyl system were

assigned by the leaning patterns as done in earlier examples.

In deuterochloroform, irradiation of the methoxy signals at

5.85 ppm caused NOE at the high field o-doublet at O.5O ppm, H-lJ.When the overlapping H-28 and H-29 signals were irradiated,

7.55 ppm, a 15^ enhancement was observed at low field singlet

at 7.90 ppm, H-25. The remaining aromatic singlet was assigned

Fig.

r,8

8 0 ,N-Diacetyitiliainosine

I2 0MI

III T N M *

' 4 NAc* 0 * t

H-14h z a Ih zs, *za i . *\ i rV H 8 ! M15

•»» * iff ^ I ' H

» J *!a oo

l -ZT

I H 40

I I I '7 00 6 50 5 50 4 00 I3 00 2 00

OM*J !ll.

I ff t TB

Iff lot

HOC * • • •« '

I" T 60 f

j VIrr 9 0 5 A

a j

8 00

h - za « tail

!jCl7 00

in 3 §3

n ot J.L J jjv U a

[ ,vV Viff,: h - z a a z »

n ?3 *■ L« H 22i: V'4; *v* i

'vA\U

8 COi i

7 00 6 50

OM* C

O.NDiocatyHiliomoalna

c d c i 5 a c t o,

!L f f

H 3 M 16

^ / V5 00 4 0 0

; j(|. . »> ' A

v ^i oo

I2 00

C D C I

zoazt-OMiI? DM*

i T N M * f t 4 N A c

ft-OAC

, - A ^ -5 OO

M 16s ft \ u ■' M

4 00 3 on 2 00

59

to H-27. In the mixed solvent system H-ba and H-l6a were Irradiated

and the decoupling observed in the aliphatic region allowing

assignment of H-5 and H-16. These frequencies were then irradiated:

H-16, 2.78 ppm and H-5, 3-03 ppm and decouplings were observed

at the signals of H-28, H-lb, H-20, H-7 respectively. This is

more evidence for the previous assignments,

b . NMR Studies of Hoffmann Elimination Products

Spectra of methine-A, the product of Hoffmann eliminations on

tiliacorine and tillacorinine dimethiodides, are shown in Fig, 9

and Fig. 10. The NMR studies on this sample were done in deutero-

chloroform as well as in a deuterobenzene-deuterochloroform

mixture because of better separation of signals particularly the

H-13 from H-16 signals. In a mixed solvent system, the three

methoxy signals at 3* ^ ppm, 3*60 ppm and 3 . 70 ppm apon irradiation

gave NOE's at the high field o-doublet at 6.b9 ppm, the low

field o-doublet at 0,87 ppm and the aromatic singlet at ( .71 ppm

identifying the latter as H-20. In CDC1:3 solution, Fig. 9 , the

o,m-doublet of doublets at 7*10 ppm was irradiated causing the

o-doublet and high field m-doublet to decouple. After irradiating

most of the aliphatic region, H-lia was found to absorb at about

2.55 ppm. Since irradiation of H-ba at 2.53 PPm gave a 2ty,j NOE

and line sharpening of 0.2 Hz ("fork coupling") was found at

the high field aromatic singlet, H-25, as well as a 0,1 Hz line

sharpening at the low field aromatic singlet, H-27 (five bond

zig-zag coupling), it was assumed that this was the position of H-ba.

Also a decoupling was found at the high field m-doublet at .75 ppm,

60

Fig. 9 Methine-AM *ihlnt'A

. A. A*** a v JI"* ,> 72* ? 3 ?3 4*•;

v H r\ 7* 7% /■V \«/ VA . y \'1 ? \ 9 P

W»" 6 Mr

Normal

(V f ,,M V 1 . ■, tj

j W ^ v ’ Vw°M ^ ^

I r r 3 4 0

NO tIrr HO

NO F

irr } 9 7j N O E 1

I '■. i

j 'm. v ■* t \/ - ,1 * "sj i' r r ?\ * 4

N OF

I i

Jl ' * ,

j NO* NCF

Vi ^><WJ+ * J ' i W l ^ ' V ■' -1 V 1 * “V St* V**, •■'v- Nr ,

\ t t 7 7 6

0 c P 1

I r r ft 0 9

Nqr m o 1 a ?! f > » P ' ) » # rT I r k ' ■ hi g

4 A 1 7 NM|

i j n tli'f

61

Fig. 10 Methine-A

'2 0M«9 S 2 I O M i 4 S t 7 N M *

M l * M -2

K 20 M T

H - « a 14 1 H 25 H - f 8 t m n t

M * 7 i - . h ^ c ■;•>! i M l®' ;*’* Hie° . S ' 1 .■* - r W i ■ V t - i | j | J ) |k ' r 1 1 l| - . I " / ' I ^

*vAA.-------------- ^ ' v ^I I ! I I8 0 0 7 0 0 6 0 0 5 00 4 0 0 5 00 ? OO

N©f frt a<H-2 7

i

1

irr 2 61O t p l

\

^rJ lV/2 ■ W ^ V v v V \ ^ / * V

N 0 f f* 0 1

t h i n« - A

;wt '0 ;i'JO J ? j? 6 1 - I • •1•| ' ’■ Yr"

V 17

Irr 2 6 <

' J Ms

NO t ft f>cp‘

V vv W ' v '* '*' v ^ ^ '■ -.■* V *.v! V.,___- v ■ v > \ 5/I r r 7 1 0

Ot p >

liD c p<

I

j j ii u11 * i \ i

1 C M *

fc2 OMf i 7 N M l

H - J9 M- 2

H OO

m • i >a ►4

7 i)0

S.' ',> ^ w

f■ JO 4 00 ;’ oo

6 2

probably due to H-5. In the mixed solvent system, what looks like

a doublet of doublets at 3*97 ppm was Irradiated to give a 20$ NOE

at the lower field m-doublet at 6,89 ppm identifying the former as the H-I6a and the latter as the H-28 absorptions. Now all

of the aromatic and methoxy proton signals can be systematically

assigned.

The methine A spectra also exhibited two twelve line vinyl

patterns. Working in mixed solvent, when the lowest field

line at 8.21 ppm was irradiated, spin tickling was observed for the

signals near 5,1; and 1^.1 ppm as shown in Fig. 10. From

this experiment together with the coupling constants the vinyl

signals could be assigned but it still was necessary to determine

which one was due to the Ol-vtnyl and the C-l9a vinyl.

Setting the decoupling oscillator at 7*76 ppm, a line sharpening

of 0.2 Hz was observed at the H-27 signal ("sickle coupling").

Also when 8.09 ppm was irradiated, the H-20 signal sharpened

by about the same amount allowing the low field signals at around

8.07 ppm to be assigned to H-19 and hence the remaining olefinic

signals can be assigned.

When the H-18cis and H-3cis signals (cjs meaning cis to the

alkene band attached to the benzene ring) were irradiated, S.VO

ppm, NOE's of 12$ and 23$ were observed for the H-20 and H-27

signals, respectively. Irradiation of the H-2s absorption resulted

in no NOE's for the signals due to H-28 or H-2y.

Methine-B, the other Hoffmann elimination product of tiliacorine

and tiliacorinine, has its spectra presented in Figs. 11 and 12.

65

Fig 11 Methine-B

•M2)OM*4 • I T N M *

l 7 OM*

H 2 7 - j

H 23M 20i 2 7

n 2ftH'3 C i •

H T | H ft ^ (j K-)| trqni jI

H-fft

I8 00 I f r 3 B ft

HOEI

' i f r A - .H - 2 , ' T " J h -|4

2 H ‘ 3 t r o n > M Ifco' 1 1 . l ,

tT 00 f> 00

In 3 6 9 NOE

{ AV T*

Irr 5 98Drill

j i V & aNOE

f >li '/> V-v. 4

00 t ff 3 3ft

4 00

N O E

,y. .i1 0 0

M *t h im - B

L.i

2 00

h !t w v A

?[>,i. jnX'V**

"<“■/ j,2B *? /•5>- A *

V ' . T '<0 /

fl* HN1N*)|h

■ 1 0 •Me OM*

Normal I f r 3 3 7

NOE NOE

X'

f r r 7 9 6 t

i i+H - I 9

I8 OO

V/V.' * ‘ i . ‘ -

h i j a i*H ?fl H Eli. M 20| J M - 3 c I *

M - I 0 [ i |M 7 I N B | 11 » '

H 29

* i * f 14 »,__, n'lv N 1H i7 00 6 00 5 00

9 O M t

71 OM* i 17 OMt

H - J i i o n M 'fin

4 00

C DC fa

4 ft I 7 N M «

5 'f ? oo

6k

Fig. 12 Methine-BM t t h i r>t * 6

Nor m « 1 Irr 355N O E

20V2 'L

y; "y 'v yIV /;**■ >' j -j^ v?V

2 8 2 9■V* <•* VM I ^ , V

i 3~ ?\ » 8Mi DMr

1 v% ' VK* y V

■ 11 i

i \ J. ' *(* * ' W ■ '

\ V V'^ V *

No r moi I r r 4 0 0

N OE

U(

I i

it »*' >■ *' ' , ' I

^ ' ‘ w ' - V v - v . ,

o 't I , 1 <

.. *1 } eI ,■

W ua/V*U S V ; , i /'■, v vl

I r r 6 0 3

N 0 I ■ N Of mO I I r r 6 7 ?

m .j 9 N U E & O c p i

■S

W■ • ■, A , ' \ . '■/■* V 1 I J \ -\y V W- *^vw/ Vv ^ ^

Irr 6 0 4i D t pi

9 OM#

2r OM# i <t> r»M4 A I ? N M

*** * A“■ - -- ' V.

h 23H 2 0

J Mm i4 H'3cn 'n-Jir7 i ! *<?9

. 1 l ' 1 ■ *v *■ w' . . ,_"V \

H 4 r > ,

H P

H O n no

#

O f 1

65In deuterochioroform solution, the high field m-doublet, 5-97

ppm, was irradiated causing decoupling of the o,m-doublet of doublets

at 7.17 ppm. Furthermore, a 10$ NOE was observed at the high field

aromatic singlet at 6.69 ppm, H-25. When the methoxy signals

at 5*88 PPm > 5*69 ppm, and 3.56 ppm were irradiated, NOE's were observed for aromatic singlet at 6.71 ppm, H-20, the o-doublet

at 6.87 ppm, and the ABX system at 6.5O PPm , respectively. The

above data together with the leaning patterns and the comparison

with the methine-A spectra was considered sufficient to assign

the aromatic protons and the methoxy signals of methine B.

The vinyl proton were assigned by irradiating the four-line

pattern at 8.00 ppm and observing the decoupling in the high

field vinyl region. In nixed solvent decoupling was observed for

H-20 when the above pattern was irradiated ("sickle coupling"),

allowing the signals at 8.00 ppm to be assigned to H-1CJ. The H-2

signals are partially overlapping with the aromatic signals

around 7.00 ppm. In a mixture solvent, irradiation of the

vinyl "cis" protons gave and 13$ NOE's for the respective

H-27 and H -20 signals, which were also observed for methine-A.

When the four line pattern at it.00 ppm, H-16a, (mixed solvent)

was irradiated, a NOE was observed for the H-28 signal

demonstrating their proximity as in previous examples. When

the spectrum was irradiated at 3.55 PPW* H-5a, a 15$ NOE was observed at H-2, and a 5$ NOE as well as a 0,2 Hz sharpening

of the H-25 signal was also seen. This Indicated that H-Jia is in

an equitorial position and the decoupling at H-2,- must represent

66

a "sickle coupling." When the H-25 signal was irradiated an 18^

NOE and decoupling was observed for the H-29 absorption, Indicating

a close proximity and peri relationship between the two protons.

This concludes the assignments of the NMR spectra of the

tiliacorine type alkaloids and their derivatives The gross

structures of all the compounds discussed are in agreement. (it

is Interesting to note that Shamma and Fog's work agrees with

the proposed structure of tiliacorinine based on independentI7;chemical and spectral data. )

III. Discussion

Blsbenzylisoquinolines have two chiral centers, C-ita and

C-I6a, with four stereoisomers possible, that is, two pairs

of enantiomers( with either both methylene bridges a. or ft, or one

bridge at and the other ft. There are two diastereomers distinguisha­

ble by NMR since it is impossible to see a difference in mirror

images.

Tiliacorine and tiliacorinine have the same gross

structure and must be diasteromeric, Nortiliacorinine-A and

derivatives are chemically related to Tiliacorinine and the NMR

spectra of tiliamosine derivatives are very similar to that of

the nortiliacorinine-A derivatives. For instance, 0,N-diacety1-

tiliamosine has exactly the same NMR spectrum as 0,N-diacetyl-

nortiliacorinine-A (see Figs. 6 and M) except for the appearance of

the C-20-OMe and absence of H-20 in the former. All compounds

mentioned in this paragraph have the same stereochemical

configuration (the existence of enantiomers in the same plant

67

can generally be excluded on biogenetic grounds) with the exception

of tiliacorine which must represent the other diastereomer.

The most informative NMR data obtained were those of the

two Hoffmam elimination products, methine-A and methine-B which

have the same gross structure but are diastereomeric. Tiliacorine

dimethiodide forms mostly methine-A and tiliacorinine dlmethio-

dideforms mainly produces methine-B. Therefore, the determination

of the stereochemistry of these two products would result in the

complete structural assignments of the original compounds if the

ratios of products from the starting materials remain constant and

no rearrangement of the chiral centers takes place except for

partial racemization.

The stereochemistry of the two products was deduced by using

Dreiding stereo models. In the case of methine-A, Fig. 1,5, the

26$ NOE plus 0.2 Hz decoupling between H-2‘5 and H-^a, in addition

to the five bond zig-zag coupling 0.1 Hz, between H-27 and H-ba

demands that H-^ia is nearly co-planar with H-2^ and H-27 and

must have a spatial arrangement as shown in Fig. 13. Also,

the large downfield shift of H-2 indicates it is deshielded by

the proximate N-l*, hence lending more credence to the above

conclusion. N-17 must also deshield its neighboring vinylic proton

H-19, as Indicated by its large downfield shift. In addition,

the large NOE's between H-5 and H-27 and also between H-17 and

H-20 is more indication that the vinyl groups are oriented away

from the nitrogen atoms as shown in the figure. The strong NOE

(20$) between H-16a and H-28 indicates that these two nuclei are

68

Fig. 13 Methine-A, Structure and Chemical Shifts in CDCl3

<1 in h-

a> • *c m t-

_c — s.*— *” \a> in \

in >N . -/(&O /GO

• 0

/

IO CM

■ ' /

69positioned closely. An extremely important piece of negative

information that must be explained is that there is no NOE between

H-25 and H-29.

First, the stereomodels were arranged showing C-ta and C-16a

in the S configurations. Keeping in mind the points made in

the last paragraph, it is extremely difficult to avoid a NOE

between H-2^ and H-29 i-n this stereochemical arrangement which

would be true also for the mirror image, Ea-R, loa-R. In contrast,

if C-16a is changed to the R configuration and C-J.a retains

the S configuration and again the above experimental criteria are

considered, H-29 can now be moved nearly four Angstroms away

from H-25, an internuclear distance which should make a NOE

unobservable. On the basis of these arguments, the structure

of methine-A is either ^a-S , 16a-R or ^a-R, l6a-S.

Similar arguments can be used to deduce the structure of

methine**Bf Fig. Ui. In this compound no large NOE is observed

between H-2^ and H-^a but a 0.2 Hz-sickle coupling exists between

the two protons. Also a l'>5& NOE was found between H-^-a and H-2.

These facts indicate that H-lia must be coplanar with H-21. and

also spacially close to H-2 as shown in Fig. 1^. Also, the

H-2 signal does not show the large downfield shift normally

caused by the anisotropic effect of the nitrogen lone pair.

However, H-19 is positioned far downfield as in methine-A and must

be stereochemically close to the N-17. There is also the 20 .

NOE between H-lcj and H-28, again showing close proximity of these

two protons. The most important clue to the structure of niethlne B

70

Fig. Methine-B, Structure and Chemical Shifts in CDC13

CDI0)

0>

CM CM

CM CM

71

Is the 18$ NOE and the unexpected coupling between H-25 and H-29-

Keeping the above experimental facts in mind stereomodels

were constructed with C-4a and C-16a in the S configurations. In

this arrangement the model can be bent so that the biphenyl moiety

is far below the plane of p-dioxin plane, thus minimizing steric

interaction. Also H-29 and H-29 can be easily situated in close

proximity and at the same time their C-H bonds can be arranged

almost parallel which must account for the decoupling and NOE.

When C-la and C-lOa are both put in the R configuration the same

reasoning holds.

Alternatively, if C-^a is set in the R configuration and C-10

in the S configuration, H-21; and H-29 are stereochemically

close but there would be much steric interaction between the

two. It would seem likely that H-iia would flip from exo to the

endo position to allow the biphenyl system to get farther below

the p-dioxin plane giving methine-A. So considering the two

possibilities the former arrangement, lia-S, 16a-S or ^a-R, I6a-R

appear to be the much more likely stereochemical arrangement.

In both, methine-A and methine-B, the H-29 resonances appear

extremely upfield for aromatic protons, 5-79 ppm for methine-A

and 5.98 ppm for methlne-B. In the other alkaloids H-29

resonates around 7.9 ppm so in the elimination products H-29 is

at least I.9 ppm more shielded. Also, in the elimination

products the H-29 absorptions are at relatively high field,

6 .I5 ppm for methine-A and 6.69 ppm for methine-B. In most other

alkaloids H-25 resonates at about 7.99 ppm- However, tiliacorine

72and its acetate are exceptions, here H-25 resonates at 6.95 PPm

and 6 .5O ppm respectively and H-29 around 7-1 PP™ and 7-2 ppm

respectively. All these anomalies are most likely due to the

anisotropic shielding effects of the aromatic rings to which H- 29

and H-25 are attached. In the compounds that show these shieldings

H-29 is situated directly above or below the aromatic ring

skeleton containing H-25 and is positioned directly below the

ring skeleton containing H-29*

Hoffmann eliminations generally proceed stereoselectlvely

and should take place without change in the configuration at C-J4a

and C-lba during the formation of methine-A and methine-B.

One would therefore expect that each alkaloid yield only one

elimination product and not two diasteriomers in unequal pro­

portions. A possible explanation for the partial racemization

at the chiral centers is that the base used in the reaction, KOH, is

strong enough to abstract the relatively acidic benzylic

hydrogen, H-la, and H-16a leading to a partial proton exchange at

those positions. Proton exchange has to be slower than elimina­

tion of H-2 or H-19, otherwise total racemization would have

occurred. Alternatively, asymmetric reprotonation could also

lead to unequal amounts of enantiomers. If one could be

certain of this explanation the structures of the other samples

could be related to the two elimination products. Using the

above arguments, it is highly likely that methine-A has the same

stereochemistry as tiliacorine and methine-B the same as

tiliacorinine. Since nothing is known about the mechanistic

73

details of the reaction, It would be advisable to have the stereo­

chemistry of each alkaloid determined separately.

In tiliacorinine, Fig. I5 , a 25$ NOE Is observed between

H-2a and H-25 indicating that H-2a is nearly periplanar and endo

to the neighboring aromatic ring. A 10$ NOE between the H-21;

and H-28 signals indicates the proximity of the nuclei. In

addition, the 20$ NOE and the coupling between H-25 and H-29

indicate their spatial closeness probably in a parallel arrange­

ment. So taken together this must mean that the two rings of

the biphenyl system are periplanar. The absence of NOE

between H-lOa and H-28 lends additional support for this.

When the stereomodels were examined and arranged such that

both C-2a and C-lOa have the S configuration, it could be seen

that H-29 can get into the close parallel position to H-25 needed

for the NOE and coupling. In this arrangement, H-28 can

approach H-25 closely also. Both H-28 and H-29 are near H-21; to

give the NOE in a variety of conformations, such as whether or

not the p-dioxin ring is bent upward or downward. Based on

these arguments, the 2a-S, iCa-S or 2a-R, l6a-R configurations

could be tentatively suggested for tiliacorinine.

Arranging the stereo model in configuration la-S , loa-R,

again H-2a approaches coplanarity with H-25 and H-28 as well

as H-29 can also approach H-25 in the same fashion as above.

But it seems more likely that H-28 would approach H-lOa to Lessen

steric hinderance. Also, the required relationships between

H-25, H-28, and H-2y to explain the experimental data can only take

T i l i a c o r i n i n e

6.636.30

3.378.05

N 3.6 8 ^ 2 , 7 7 3.40

7.587.672.31 Me

7.2 67.35

6 . 9 9

M e O3. 98

O H6. 9 5

M e2 .65

Fig. I5

Tiliacorinine, Structure

and Chemical

Shifts in

CDCl^

-Ac

efy

ltil

iac

ori

nin

e

lb

Fig. 16 O-AcetyIti1iacorinine, Structure and ChemicalShifts in CDCl^

4> <0

CJfO

IO

O

tn

CJ

CJ

roco

•ro

<0

76place if the p-dioxln system (shown as In Fig. I5) is pointed

upward, which from model considerations is not a stable conformation.

Alternatively, in a conformation with much less steric hlnderance,

in which the p-dioxin ring is pointed downward, H-28 would be

too remote to give any NOE with H-25. In addition, H-29 would

be too far away from H-25 to give a 20% NOE. Taking all

this into consideration, the la-S, 16a-S configuration or the

enantlomer is the likely structure for tiliacorinine. By analogy

its major Hoffmann elimination product, methine-B, should have

the same configurations if the elimination reaction proceeded with

predominant retention of configuration and the two asymmetric

centers.

The data on tiliacorine, Fig, If, is sparse due to inopportune

signal overlaps. A NOE between H-16a and H-28 (15%) indicated

the nearness of these two protons. Also there exists the same

relationship between H-la and H-2'. as found in tiliacorinine.

However, the signal due H-29 appears relatively sharp; suggesting

that no coupling to H-25 occurs, which had been observed in

tiliacorinine. When in a Dreiding model C-la is chosen as S

and C-lOa as the R configuration, it can be seen that the biphenyl

system and H-29 get about 5 ^ away from H-25 and discussed above.

Tiliacorine as mentioned before gives a large upfield shift for

the H-2',t signal and a lesser upfield shift for the H-21. signal.

As discussed earlier, this is probably due to the shielding

effect of the two aromatic rings containing these protons. There­

fore, H-29 must be situated below the H-2'. containing aromatic ring.

Til

iac

or

ine

77

Fig, 17 Tiliacorine, Structure and Chemical Shiftsin CDC13

ir>10

X00

CO OkCM<7>

IO ^" 00 O .^ C M

00to4)

kO

CM

- A

cety

l ti

lia

c or

i ne

78

Fig. 18 O-Acetyltiliacorine, Structure and ChemicalShifts tn CDCIt

«■CDroo>

CNi

in

m

in

a>

CVJ

79This could happen only if a ia-S, l6a-R configuration or its

mirror image were present. Steric hinderance would not allow

tiliacorine to have the same configurations (R,R orS,s) at

the two chiral carbons.

The data for N-acetyltiliamosine Is very similar to that of

tiliacorinine and the same couplings patterns as well as NOE's

are observed. In addition, there is found a large downfield

shift of H-T due to the anisotropic effect of the J.-N-acetate

carbonyl group. Unfortunately, stereomodeIs indicated that

the carbonyl can approach H-7 in either diastereomer thus providing

little structural information. However, the same line of thought

that led to the structure of tiliacorinine can be applied to

N-acetyltiliamosine (Fig, iy) and 0 ,N-diacetyItiliamosine

(Fig. 20), suggesting 7a-S, 16a-S configurations or their enanti-

meric structure for the compounds.

Lanthanide shift reagent experiments were carried out on

N-acetyltiliamosine, tiliacorine, tiliacorinine and their

O-acetates, and methlnes A and B, The shift reagent used was

Eu(fod):3. Because of the complexity of the data, due to too many

complexation sites, useful Information could not be obtained.

However, the data are presented for tiliamosine, tiliacorine,

and tiliacorinine since on these compounds some interesting results

were obtained. In many Instances where signals broadened

quickly and others where there were overlapping, so many protons

signals did not permit interpretations.

In N-acetyltiliamosine (Table i), it is interesting to see

-Ac

ety

ltil

iam

os

ine

Fig. 19 N-Acetyltiliamosine, Structure and ChemicalShifts in CDCl-^

80

CJ CJ

\

00)20000 0010 ro

w0

CJ

0,N"Diacetylti(iamosineOMe 3 . 9 2 , 3 . 8 3 or3. 80 I 6.69

7. 9 0

2.27 M e 7.9 87 .36

6.86M e O OAc 2.153.80 or

A c

2 .2 7

N o r t i l i a c o r i n i n e ~A OMe 3 . 8O ( CDC l 3 2 0 % D M S O - d g )

6 . 6 5

4 . 1 48 . 0 5/ \. 7 . 4 0 or 7. 4 9 o r' 7. 4 9 7. 4 0

7.35 7.5 5or 7.55 or 7.35

6 . 9 4

3. 91

6 . 80OH6. 9 8 OSro

Fig. 21

Nortiliacorinine-A, Structure

and Chemical

Shifts

0 , N - D i a c e t y I n o r t i l i a c o r i nin e -AOMe 3 .84

6.7 46.36

5.21 Ac2.23

\ m < V > ^ 2 . 8 4

N 4 0 4 \2.7 83.21

7.5 5 7.582.23 M e

7.40 8.026.9

* M e3.78

O A c2.14

Ftg. 22

0jN-Diacetylnortiliacorinine-A, Structure

and Chemtcal

Shifts

Table I N-Acetyltiliamosine Shift Reagent Data(ppm after each addition of Eu(fod)^)

8 ^

E u :'mosine Mole Ratio

0.00 0.0491 0.136 0.245 Di f ferenceat 0.245

Protons

20-0Me 3.88 3.88 3.90 3-93 .05H-27 6.65 6.6 6 6.74 6.83 . 18H -25 7.97 8.01 8.16 8.35 .38

H-7 7.92 8.18 8.77 9.35 1.63h -8 7.09 7.06 6.94 6.79 -.27

H-13 6 ,94 6.94 6.94 6.94 0.00H-l4 7.35 7.35 7.35 7.35 0.00

H-28 7.70 7.71 7.75 7.80 . 10H-29 7.44 7.49 7-60 7.77 .35H-4a 5.18 5.58 6.43 6.93 1.75H-I6a 3 .6b 3.62 3.65 3-67 .0321-OMe 3.84 3.82 3.83 3.84 0.0012-OMe 3.99 3-97 3.96 3.96 0.00

17-NAc 2.25 2.41 2.82 3.46 1.214 -NMe 2,25 2.25 2.27 0

*

CO .05

85that H-I3 , H-lb, H-28 and 12-OfMe show very little L.I.S. whereas

H-7, H-8 , and H-29 have very large L.I.S.'s. H-8 shows a very

unusual upfield shift due to the angular effect of the 9~0H - Eu

complexation, H-7 undergoes a dramatic L.I.S. which is probably

due to the combined effect of 9~0H and N— L amide complexation.

It should be noted that the H-^a shift is very large also. H-I6a

does not shift very much so it seems that the complexation

of europium occurs much more strongly at amides than amines ,

which is the opposite of what a person would think, since amines1 10are stronger bases.

In tiliacorine (Table II) it is interesting to note that H-16

shows a greater L.I.S. than H- Ua yet i+-NMe shifts more strongly\

than 16-NMe. This indicates that the greater complexation takes

probable place at 4-N and the greater shift of H-16a is due to

the angular contribution to its L.I.S. In tiliacorine, Table III,

17“NMe shows a greater shift than 1-NMe. This can probably be

explained by steric hinderance of the biphenyl group.

Both tiliacorine and tiliacorinine show Cotton Effects and

their O.R.D. spectra are presented in Fig. 23. In the region

below 250 urn, due to instrumental limitations, the ORD data in

this region is less reliable but the general trend of the curves

can be used. The only information that can be obtained from the

spectra is that the two compounds have two different stereochemis­

tries. In these compounds there are four aromatic rings (four

chromophores) contributing to the spectra. This makes the

spectra too complex to analyze without compounds of known similar

86

Table II Tiliacorine Shift Reagent Data(ppm at each concentration of Eu(fod)3 )

Eu:'corine Mole Ratio

0.000 0.073 0.127 0.220 Differenceat 0.220

Proton

H-20 6.26 6.32 6.35 6.38 .12

H-27 (-■.69 6.73 6.77 LI. 81 .12H-2, o.99 7.00 7.11 7.17 .18

H-7 - - - - -

h -8 - - - - -

H-l3 (broadens) 6.90 6.95 6.95 6.98 .08

H-12 - - - - -

H-28 7.95 8.00 8.07 8.12 .21

H-29 7.2 s 7.31 7.57 7.21 .10

H-2a 5.59 5.29 3.60 3.70 • 31H- 16a 2.08 2.22 2.20 2. 58 .50

21-OMe 7.86 3.37 3.89 5.91 .0912-OMe 5-96 5.99 2.0 2 (broadens)

17-NMe 2.69 2.80 2.95 3-07 .382-NMe 2.35 2.27 2.65 2.80 .27

87Table III Tiliacorinine Shift Reagent Data

(ppm at each concentration of Eu(fod) ,)

Eu:1corinine Mole Ratio

0.000 0.003 0 .I9I 0.319 Difference at 0.319

P roton

H-20 6. 80 6. 31 0.33 0.32 .12

H-27 0.08 u . 6 8 0.73 0.87 .23

H-2'j 8.0 0 8.10 8.23 8.32 .37

H-7 7.28 7.28 7.80 7.32 7.70 .32

H-8 (broadens) o.yo 7.01 7.05 7.0 - .10

H- 13 ■ *.■■.96 7.00 7.0 3 .10

H - U 7.8;* 7.87 7.32 7.99 .21

H-28 7.07 7.08 7.79 7-92 .29

H-20 7.38 7.00 7.73 7.87 .19

H-9a - - - - -

H-loa - - - - -

21-OMe A . 2.1, 8' .87 ■ v « 3.92 .07

12-OMe 3-98 7.00 i *»01) 3,ly .21

17-NMe 2.03 2.79 2.95 3.10 .39

3-NMe 2.31 2.33 2.39 2.08 • 37

8 8

Fig. 23 ORD Curves of Tiliacorine and Tiliacorinine

O R D C u r v e sT i l i a c o r i n i n eT i l i a c o r i n e4 0 , 0 0 0

3 0 , 0 0 0

2 0,000

10,000

3 0 0 n m 4 0 0 nm X

- 10,000

- 20,000

- 3 0 , 0 0 0

- 4 0 , 0 0 0 J L

structure for comparison.

IV. ConclusionsA wealth of NMR data concerning these alkaloids has been

acquired. The positions of the methoxy, hydroxy and acetoxy

groups were determined for all of the compounds concerned and

the stereochemical configuration of each was shown to be one of

two possible enantlomers. In addition to the elucidation of

the structures new information concerning long range couplings

and anisotropic effects was obtained. A small, 0.2 Hz, benzylic

coupling for protons in the "fork configuration" and a 0.2 Hz

benzylic coupling for protons in the "sickle configuration" were

observed. The most surprising observation was the H-2'; to H-29

through space coupling which was seen In several of the alkaloids.

As mentioned in the introduction, coupling through space is supposed

to be negligible for direct proton-proton coupling but here the

coupling can be estimated to be Q.f Hz T .2 Hz which is quite

signifleant,

In the Hoffman elimination products, three of the four vinylic

protons in these compounds resonate at around 9.00 ppm which is

unusually downfield. The cause was shown to be the anisotropy

in the deshielding by the nitrogen atoms. Aromatic absorptions

of about 0.00 ppm which are unusually upfield have also been

mentioned and are shown to be due to the shielding effects of

diamagnetic currents in the aromatic rings.

In spite of this plethora of data It still remains to determine

the exact stereochemistry of these compounds. If some degradation

90

scheme could be devised which would remove several benzene

chromophores while preserving the stereochemistries of C-l+a and C-16a

and still leave a rigid structure, O.R.D. spectroscopy might

provide the answer. Or else a compound similar to the alkaloids

with known stereochemistry could be synthesized and then compared

to the original samples by N.M.R. or O.R.D.-C.D. AnjaneyuluITet al. used a Bischler-Napieralski cyciization to form the rings

containing the chiral centers but the reaction is not stereospecific

so it cannot be used. So it seems that it will be up to an X-ray

spectroscopist to determine the chirality of the compounds.

Bibliography

91

1. B. Anjaneyulu, T. R, Govindachari, N. Viswanathan, K. W. G.Gopinath, B. R. Pal, Tetrahedron, 2 5 . 3 0 9 I , ( I 9 6 9 ) .

2. M. Shamma, The Isoqulnoline Alkaloids-Chemlstry and Pharmacology , P. XVII, Academic Press, New York, (1972

J. Ibid, p. ^5.

k. Ibid, p . I I 5 .

9. B. Anjaneyulu, T. R. Govindachani, N. Viswanathan, K. W. G.Gopinath, B. R. Pal, Op. Cit.

6. T. K. Palit , M. P. Khare, phytochemistry, 8, 599* (I969)*

7. A. Barua, P, Chakrabarti, A. S. Dutta Gupta, Journal of theIndian Chemical Society, (9) > 920 (1970).

8. K. V. J. Row, L. Ramachandra Row, Journal of Scientific and Industrial Research, India, 16B, I 5 6 (1997?-

3. A. N, Tackle, D. Dwuma-Badu, Photochemistry 12, 203 (1975).

10. B. Anjaneyulu, T. R. Govindachani, N. Viswanathan, K. W. G.Gopinath, B, R. Pai, Op. Cit.

11. I. R. C. Bick, A. R. Todd, J, Chem. Soc. , p. 1606 (I95O).

12. B. Anjaneyulu, T. R. Govindachani, N. Viswanathan, K. W» G.Gopinath, B. R. Pai, Op. Cit.

lj. B. Anjaneyulu, T. R. Govindachari, N. Wiswanathan, Tetrahedron 27, i+39 (1971).

lit. M. Shamma, James Fog, J. Org. Chem. il, (7), 1295 (l'J70).

15. B. Anjaneyulu, T. R. Govindachani, N, Viswanathan, K. W G.Gopinath, B. R. Pai, Op. Cit.

16. A. Cockerill, G. Davies, R. Hardin, D. Rackham, Chem. Rev. 73,(b) p. y>3 (1973).

17* B. Anjaneyulu, T. R. Govindachani, N. Wiswanathan, Tetrahedron 2 Y , op. cit.

CHAPTER III

NMR STUDIES OF FURAN CYCLOADDITION PRODUCTS

I. Introduction

lhis chapter is concerned with the determination of the stereo­

chemistries of tetrahydrofuran derivatives by the use of NMR.

Nine compounds were provided by Dr. Gary Griffin, Ira Lev and

John Wong at the University of New Orleans. Compound I was made

by the photochemical eyeloadditlon of 1,l-dicyano-2-phenylethyiene

oxide with cis 2-butene. Compound II was prepared by thermal

cycloaddition of the same starting materials. Ill and IV were

prepared by photochemical and thermal cycloaddition of 1,1-dicyano-

2-phenylethylene oxide with trans 2-butene, respectively.

The addition of trans 1 ,2-diphenylethylene oxide to trans

dicyanoethylene yielded products V and VI. The addition of the same

epoxide with trans dicarbomethoxyethylene yielded VII. In addition,

we analyzed the product of the cycloaddition of trans or c i s 1 ,2-diemthyl-

1 ,2-diphenylethylene oxide with trans cyanoethylene, VIII. The

product of the same oxide with cis dicyanoethylene was designated

as IX.

The products of these reactions are tertahydrofuran derivatives

tives and the preparation of I is shown schematically. Compound I

has four possible structures, both the C-3 a°d C-Jj methyl groups

can either be cis or trans to the C-*> phenyl group. Similar

possibilities exist for compounds II, III, and IV. Eight possible

structures are possible for the other five compounds. For example, in

Compound V, whose preparation is shown, the C-I and C-;> cyano groups

02

95

CN

C N /M e M e

CN

'CN(I)

Rj and Rg = H and M e R g and R4 e H and M e

C N

+C N

(V)

R | and R2= H a n d C NR 3 a n d 84= H a n d C N R g and Rg*H o n d $ R 7 a nd Rg= M and f)

9b

can be either cis or trans to the C-5 phenyl group and the C-2

phenyl can be cis or trans to the C-5 phenyl moiety.

We were Interested In the determination of the exact configura­

tion of the above compounds. From the stereochemistry of the

different products it was expected to obtain information on the

mechanisms of these reactions. Retention of configuration from

substrates to products would indicate concerted cycloaddition reactions,

and from the configurational assignments the type of cycloaddition

could be determined. For instance, in a concerted process, if a cis

ethylene derivative and a cis epoxide were reacted,the 2 and 5

substituents of the furan type product should be cis to each other

and the substituents 5 and J-4 would also have a cis orientation."^

If these reactions were i+-+g cycloadditions the favored reactions

would be suprafacial-suprafacial thermally, and suprafacial-

antarafacial photochemically. Antarafacial additions are not

as likely for steric reasons and will lead to a change of confi­

guration of the substituents of the antara-reacting addend in the

final product.

Radical and zwitterionic mechanisms would only have various

degrees of concertedness, A breaking of the C - C bond of the

epoxide ring may lead to a zwitterion with a positive charge+ _

on one carbon and a negative on the other q-,-C-0-C(CN)p . The

positions of the charges and the stability of the ion would be

determined by the inductive effects of the substituents on the

oxirane system. Whether or not rotation occurs after the ion is

formed would depend on the size of these substituents. A rotation

95would lead to a change of configuration from starting materials to

products; for Instance, a 2,5~cls furan compound will be derived

from a trans epoxide. The zwitterlon can then add to the olefin

either in a stepwise fashion or by a completely concerted process.

In the concerted process the pair of electrons responsible for the

negative charge forms a new bond between the epoxide carbon and

what was the olefin and simultaneously the electron pair from the

II bond of the olefin forms the other bond completing the five

membered ring. In the stepwise mechanism a new zwitterlon would

be formed after the electron pair responsible for the negative

charge forms the new bond. This new ion would also be stabilized

by substituents involving rotations with obvious consequences.

Therefore, the stereochemistry of the reactants and products

gives a clue to mechanism. In general, zwitterionic mechanisms will

be more likely in polar solvents^ and when there are electron

withdrawing groups on one side of the epoxide and releasing

groups on the other. Using molecular orbital symmetry considera­

tions, concerted cycloadditions are favored when electron

withdrawing groups are present In one starting compound and‘3releasing groups on the other.

II, Results

The spectrum of compound I is shown in Fig. 1. The doublet

at It.85 ppm must be due to the proton adjacent to the phenyl

group on the basis of its chemical shift, splitting patterns

and slight broadening. The tall peak at about T.J ppm Is obviously

the phenyl signal. The multiplet at 2.02 ppm was assigned to

Fig, I Compound I

(4.83) H CNLong Range Couplings

H-5, 4Me = Q4Hz {7.25-7.40)Ph■*,':"'’,"", K*3,4Me : 0.2Hz ,

lMe(L37)(109)

Irr at 1.37H-5

Irr. ot 4 83

H-3

7.45 7.20 4 83 3 0 2 2.62 (37 1.09

4-Me

H-b because of its large number of splittings (16 lines theoreti­

cally) and the five-line pattern at 3.02 ppm was ascribed to

H-3* The assignments of the two methyl signals were accomplished

by double irradiation. When the doublet at 1.37 ppm is irradiated,

decoupling takes place at H-3; therefore, this signal is the

C-3-Me and the doublet at 1.09 ppm has to be due to the C-l-Me,

The coupling constants are shown in the figure. Nuclear Overhauser

effect experiments were performed but irradiation of the C-l-Me,

C-3~Me and li— 5 gave no effects. The configuration of compound 1

was determined using the values of the coupling constants.

The J3 f4 value was determined to be 7*3 Hz, indicating that

H-3 and H-l are cis to each other. The fact that H-3 is attached

to a carbon with an oxygen and still shows a fairly large

coupling (J4)5 = 8.5 Hz) means that H-l and H-3 are trans oriented.

Comparison of the coupling constants with those of the other

compounds corraborate the above assignments. In addition, 1-bond

couplings were observed between H-3 and the C-l-Mej )*' H

and H-3 and C-l-Me * 0.2 Hz).

The spectrum of compound II is shown in Fig. 2 and the signals

were assigned as outlined before. It should be noticed in

this case that the phenyl signal is very much broadened, A

probable cause for this is steric interaction of other substituent

on the furan ring viz. C-’.-Me. The value is 3.7 Hz

which indicates a cis relationship between the respective protons

and the J value (1.8 Hz) is evidence that these two protons are cis

to each other. This defines the configuration of compound II as

F i g . 2 C ompound 11

(7.1-7 5) P

7.50

(3.29)eT-

(0.69) ^

(5.46) ....*CN i

H 2.98

Me(t.44) I 1

Long RGnge Coup! r.gs H-5, 4 Me = 0.2Hz H-3, 4Me 0.3Hz H-3, H-5 = 0.3Hz

—10 NOE - H-4.H-5

H-5

1: L '* W Irr. at 0.69

3 -Me

Irr. at 5 46

4 - Me

7.10 5.46 3 5 0 3 2 9 2.63 144 0.69

CD

99

shown in the Figure 2. There are also lt-bond couplings observed

(jr-_4Me “ Hz and " °*5 Hz). In addition, a long range

"W-coupling", between H-J and H-5, of 0,5 Hz is found which

corraborates the structure proposed. A 10$ NOE between H-5

and H-5 is further evidence that these two protons have a c is

orientation. Upfield from H-5 and downfield from each

methyl doublet extra sets of doublets of low intensity appear.

These are due to an impurity which could be another cycloaddition

product or else the ethylene oxide starting material.

The spectrum of compound III is shown in Fig. *'■. The signals

were assigned as before. J 4_r- is 9*75 Hz indicating a trans

relationship and J^_4, which is 11.5 Hz also indicates a trans relationship between the two protons. So the configuration is as

shown in Figure 5 , There was also a long range coupling between

H-5 and H-J indicating that the two nuclei are on the same side

of the furan ring. A 1‘5$ NOE at H-5 and a 10$ NOE at H-5 were

observed when i -Me was irradiated which means that all three of

these are on the same side of the plane of the furan ring. The

long range coupling and the NOE's are additional evidence for the

structure shown in the figure.

The spectrum of IV is shown in Fig, Ji. Determination

of the couplings in this case isn’t so obvious since H-5 and

H-k overlap. The 500 MHz spectrum of the H-5 and H-5 region is

shown in the insert. The two quartets of H-5 which are due to

coupling from the C-5 Me and H-f, were found and the distance in

Hz between the centers of the two quartets was taken to be J

Compound III

45 % NOE - 4 -Me 8 H-5 40% NOE - 4-Me 8 H-3

(4.7 3) H

(725-7.40) Ph

9.75

(1.99)6£

(i.OSitfe

fthlMIlK"*'

H(2.62)

H-5

i

\ # J

Me(L44)

Irr, at 4.73

trr. at 1.44H-3

C-3-Me

H-4

7 00 4 7 3 2.62 1.99 1.44

Fig. ^ Compound IV

(6-9-7.2) P ,CN

(5,43)

Me (1.37)(2.48) Ks”""""":6.5 y

(0.69) Me

3-Me3 0 0 MHz INSERT

10% NOE - H-3,4Me ~ ! 0 \ NOE - H-4.H-5

H-5H -3 H -4

.372.503.005 4 3 3.50

|4-Meii

0 69

102

which was 11.0 Hz. The four quartets of H-b were assigned and the

centers of the two quartets were taken to be J3_4 which was 11.0 Hz.

The 11.0 Hz J3_4 was found between two alternating quartet centers

and J 4_5 was taken to be the distance between each of the end

quartet centers and J4_.5 was taken to be the distance between each

of the end quartet centers and their neighboring center which

was 8.25 Hz. A person could not simply measure the splitting at

H-5 at 100 MHz because of the second order effects involved. The

coupling of 11 Hz is obviously a trans coupling and the .2S Hz

coupling at first glance seems to be a trans coupling but other

evidence indicates otherwise. The phenyl signals are spread

due to the b-Me being cis to the benzene ring. Also, there was

a 10% NOE observed between H-b and H-5* In addition to this

it can be seen that when 1-Me is cis to a phenyl group it is very

much shielded, 0.6y ppm* and when trans b-Me comes at about

1.05 ppm. So this would indicate that H-b and H-5 are almost

periplanar to each other and that the 5-phenyl and 1— Me are not

bending down toward the plane of the five membered ring.

A 10% NOE between b-Me and H-jj was also seen. This indicates

that the structure of the compound is as shown in the Fig. b.

The spectrum of V is shown in Fig. 5* Here the signals are

perfectly visible and the couplings can be measured directly

which makes the assignment of the signals easy. J3-4 is Hz

which indicates a cis relationship between the two protons.

J 4_s is 6.5 Hz which is a cis coupling and is Y. 5 Hz which

is a borderline value, it could be either a cis or tcans coupling.

Fig. 5 Compound V

Normal

JI

7 . 4 5 5 . 3 2

C0Ci 3 100 MHz

I rr . 7 . 4 5NOE

3 . 3 8tmpuri ty

H-4

ii *■>t+HWrkHM+rf

3 . 8 3 3 . 3 8

lOiv

Since the molecule is not symmetrical and the two nitrile groups

are cis. one of the protons H-5 or H-2 must be cis to H-5 and H-^ and the other must be trans. Since J 4.5 is the smaller coupling

H-h and H-5 are probably the ones that are cis to each other. From

looking at the other spectra in this series it is seen that if H-li

or H-5 is cis to a phenyl group that signal is more upfield

than one that is trans to a benzene ring due to the anisotropy

of the benzene ring. This also means that H-2 and H-5 must be

trans to each other and H-E and H-5 must be cis. The signals

due to the phenyl groups appear to have a fairly sharp signal

superimposed on a complex pattern again indicating that one ring

is freely rotating while the other is restricted one trans to

a nitrile and the other c is to a nitrile.

When the phenyl signals were irradiated, 7-^*5 PPm , a l'/f< NOE

was seen at the H-5 and H-2 signals along with some decoupling,

(interesting but not illuminating.) Also there was an NOE at

the H-5 signal and none was observed at the H-l signal when

the phenyl signals were irradiated corraborating the proposed struc­

ture, However, the absence on a NOE between a phenyl and a

proton two carbons away in these compounds does not necessarily

mean a trans relationship. When H-5 was irradiated, 5*5^ PPm »

a h% NOE was measured at the H-5 signal again corraborating the

structure. However, a J./, NOE is very small and there is a

chance that the observation is not authentic. So the NOE's give

additional evidence for the structure but can't be taken by

themselves as proof. But there is ample evidence that the structure

105

shown in Fig. 5 Is correct.

The spectrum for VI is shown In Fig. 6. Here we have a

symmetrical molecule and the AA'XX* pattern is seen. The H-5,

H-U system appears to be a six line pattern. The tallest peaks

in the system are lines 1 and 2 and 3 and h and make up half

of the line intensity of the system.^* Their separation in Hz

is N » J v + J.v t* (We designate H-3 and H-3 as A and A' and H-2AX AXand H-5 as X and X'.) We can assume that J.Y , is very smallAaand approaches zero from structural considerations. So N,

which is 8.5 Hz equals J 4_ r, and Jp_3 . The inner two lines are

lines 6,9 and 7*12 which overlap and the two small outer lines are

5,10 and 8,11. The separation of lines 5 and 6, and 7 and 7 is

K = J A , + J-.,; the separation of 9,10 and 11,12 gives AA AAM ■* J. A i - J w f Both of these quartets are centered on the H-5 AA XXand H-3 resonance and from the above formulae and the fact that

J , « = 0, it can be seen why the lines overlap. So allXXthat is necessary is to measure the separations of one of the

two inner lines to its nearest outer line, e.g., 0-3, and this

is J3_ 4 which is 10.5 Hz. Now it is known that H-3 and H-3

are trans to each other and H-3 and H-5, and H-2 and H-5 are also

trans to each other. This is enough information to write the

structure as it is shown in Fig. 6. Additional evidence for the

structure is supplied by the relative sharpness of the phenyl

peak and the 7$ NOE measured at the H-3 and H-3 signals when

the aromatic signal at 7.35 PPm was Irradiated. There was also

the large 18^ NOE at the H-2 and H-3 signals when this same signal

Fig. 6 Contpound VI

NormalC D C U I 0 0 MHz Irr. 7. 4 5

NOEH Normal I rr . 7 . 4 5

NOE a Dcpl

6 . 9 ---- ------

1 , 2 - I ------ 3 , 4

5.10 I N S, ft

" C N

H - 3 8 4

y n »» I I|»I

5 . 4 47 . 45 3 . 4 9

106

107w a s Irradiated.

The spectrum of VII is shown in Fig. 7- Here again an AA'XX'

spectrum is seen and the couplings were calculated as before here

J 4-5 is 8 Hz and J3_ 4 is 6.5 Hz. So H-3 and H-2 are cis to each other and H-2 and H-5 also H-2 and H-3 are trans. The phenyl

signal is fairly sharp also which verifies the structure as shown

in the figure. When the phenyl signal was irradiated the ll/j

NOE and decoupling at the H-2 and H-5 signals are seen but only

a very slight 1$ or 2$ NOE at the H-5 and H- 5 signals which could

very well be statistical error. No NOE was seen when the

carbomethoxy peak was irradiated; in these cases NOE is never seen.

The spectrum of VIII is shown in Fig. 8. J3_ A is 12 Hz which

means that the two protons are in a trans orientation. The

aromatic region of the spectrum consists of one fairly sharp signal

and a complex set of signals which means that one phenyl group

is cis to a nitrile and the other is trans to its neighboring

nitrile. Also H-2 is very much upfield from H-5 which means

that H— It is cis to a benzene ring.

When the methyl signal at I.80 ppm was irradiated decoupling

is seen at 5.57 ppm so this Identifies the methyl signal and

doublet as those due to 5-Me and H— h. When the methyl signal

at I.96 ppm was irradiated an 18$ NOE was seen at the doublet

at 5*81 ppm, so the Me signal is that of 2-Me and the doublet

is due to H-3 and also a cis relationship is indicated between

H-3 and 2-Me. When the decoupler was set at 7*22 ppm which is the

sharper aromatic signal a 7$ NOE was seen at 3* i ppm, H-2, and

Fig. 7 Compound ?II

H

MeOOC COOMe

H

C DC U I 00 MHz

Nor ma l Irr. 7 . 3 3

NOE a Dcpi

COOMa's

H - 2 a 5H - 3 a 4

3. 9 8 3 . 2 85 . 9 47.33 o00

Fig. 8 Compound VIII

N o r m a l

CNCDCI3 100 Muz

i

<b- 5

I rr. 1.80

* D cpI

4

' t

Normal

H - 3 H - 4

NOE

rr. 1.97 Me- 2NOE Me “5

Uul.

110

none was seen at 3,81 ppm, H-3. When the complex aromatic pattern

centered at J.k6 ppm was irradiated no NOE's are seen. So all

this information is consistent with VIII being that shown in

Figure 8.

Compound IX is a very symmetrical molecule and its data

is shown in Fig. 9* Only one signal is seen for H-3 and H-l and

one signal is responsible for both 2-Me and 3-Me. The phenyl peak

is somewhat broadened probably due to a restricted rotation of the

phenyl groups by either the methyl groups or the cyano groups

in some bent conformation. The coupling between H~3 and H- 3

was determined by sateiite analysis.^ The region around H-3

and H-*4, 3.73 ppm, was run 200 times with a time average computer

and the resulting spectrum is shown in the Fig. 9 . Exactly 73 Hz

away and on both sides of the large H-3 “ H— 1* signal a doublet

with a splitting of 7.3 Hz can be seen. These are the i;,C sateiite

peaks of H~3 and H-h and the splitting of the doublets is J 4

which means that they are cis oriented.

When the phenyl signal at 7.28 ppm is irradiated an NOE

is observed for the singlet due to H-3 and H-^t. When the methyl

singlet at I.96 ppm is irradiated decoupling is seen at ppm

but no NOE is observed. Therefore, the structure of IX is as

shown in Fig. 9. The broadened phenyl pattern is exceptional

and can be explained by a bent conformation which allows the cyano

groups to hinder rotation of the phenyl ring.

III. Discussion

Several generalizations can be made from the NMR spectra of

MeMe

NC'''

0 ‘s

Figure 9 Compound IX C0CI3 t o o MHzN o r m a l I rr. 7 . 2 8

NOE

l 3 C S a t t l i t e tC S a t « h t * i3 . 7 3

7 . 2 8

H - 3 a 4

M e ' s

1 9 63 . 7 3

111

112

the previously discussed derivatives. Coupling constants between

H~3 and H-I4 of less than 8 indicate cls-orientation of the respective

protons and values above 8 Hz indicate trans relationships.

Similarly, J 4-3 or of 7.5 Hz or larger indicate a trans

relationship between the respective protons, lower values indicate

cis configurations. When or H-^ are cis to a phenyl at

C-l and/or C-5, they will be about 0.5 ppm more upfield than a

proton trans to a phenyl due to the shielding effects of the

benzene ring. A phenyl group attached to carbon experiencing no

steric or anisotropic effects will appear as a singlet. However,

if a phenyl group on a furan ring is cis to a more bulky group

such as cyano or methyl the phenyl protons will have different

chemical shifts due to hindered rotation and steric and anisotropic

effects of other substituents and the phenyl signal(s) will be

split to a more complex pattern. These coupling and chemical

shift effects can be conveniently used to identify the

structure of compounds of this type.

The following generalizations can be made about nuclear

Overhauser effects: A proton attached to a carbon that also holds

a phenyl group will show a NOE of about 16% when the phenyl signals

are irradiated. If a proton is separated by 2 carbons and cis

to the phenyl group usually there will be an NOE of about 7%

upon irradiation of the phenyl signals. However, in compound VII

we see only a very minor NOE between a phenyl and a cis proton

due to a conformational factor. NOE's betwen a methyl group and

a proton cis to it on a neighboring carbon range from 10% to 16%,

Conclusions concerning the chemistry involved are much more

difficult to be had. The reactions leading to products I, II, III,

and IV occur with retention of stereochemistry of the butadiene

systems, i.e. trans reactants give trans products and cis reactants

give cis products. The addition of trans-1 ,2-dicyanoethylene to

trans-1,^-diphenyloxirane gave a major product with cis nitrile and

trans phenyl groups, V, and trans phenyl and nitrile groups as a

minor product, VI. The addition of trans-dlphenyloxirane to trans-

dicarboxymethoxyethylene yielded VII which has c is phenyl groups

and cis carbomethoxy groups. The reactions leading to products VIII,

and IX occured with retention of steroechemistry giving cis and cia

phenyl groups and trans and cis cyano groups, respectively.

Concerted reactions occur with retention of stereochemistry.

Reactions that involve dipolar mechanisms may or may not take place

with retention of stereochemistry and radical mechanisms result inua loss of configuration. Another factor to be considered is the

photochemical and thermal Isomerization that results in a change

of configuration of the oxiranes.^0 These complications make the

postulation of mechanism from stereochemical data alone impossible.

11/4

Bibliography

1. J . March, Advanced Organic Chemistry Reactions Mechanisms and Structures. McGraw Hill, New York, i960, p. 627-

2. K. Houk, "Pericyciic Reactions and Orbital Symmetry", in Survey of Progress in Chemistry, Academic Press, New York, 1975 > Vol. 5 , p. 150.

5. Ibid.

Z4. Ibid. p. I55.

‘j. J- March, loc. cit.

0. F. Bovey, Nuclear Magnetic Resonance Spectroscopy, AcademicPress, New York, 1969, p. llo.

7. ibid. p. Ho.

8 . J. W. Emsley, J. Feeney, L. H. Suttcliff, High Resolution Nuclear Magnetic Resonance, Permagon Press, New York, 19(-'j > Vol. I. p.

9 . K. N. Houk, op. cit. p. I5I.10. K. Ishikawa, G. Griffin, I. J. Lev, J . Org. Chem., it 1 , 57^7, (1976).

CHAPTER It

REPANDIN-A AND B, SESQUITERPENE LACTONES FROM TETRAGONATHECA REPANDA

(COMPOSITAE, HELIANTHEA).

I. Introduction

This chapter describes the structure elucidation of repandin-

A and B, two sesquiterpene lactones, from Tetragonatheca repanda

(Compositae, Heliantheae) for the purpose of a chemotaxonomic study

of this genus within the tribe Heliantheae, The repandtns

belong to the ten-membered ring sesquiterpenes, the gerwacrano-1 2lides which are divided into four sub-groups: * the germacrolides

represent 1(10), b( b )-trans , trans-cyclodecadienes , I; the

melampolides have a 1( 10)-cls (rj) - trans-eye lodecadlenes skeleton,

II; the heliangolides , a 1( 10)-trans , If 5)-cls-dlene mediuin

ring, III, and the 1(10), Ij( 1}) -cis , cls-cyclodecadienes, IV.

II, Results and Discussion

A. Isolat ion

Dried leaves were extracted three times with chloroform in a

Waring blender; the solvent was removed _in vacuo leaving a brown

syrup which was dissolved in ethanol. An aqueous solution of yfi

lead acetate was added and the gummy precipitate was filtered

over celite. The filtrate was evaporated and the resulting

aqueous phase was extracted with chloroform several times. These

extracts were combined and evaporated leaving a brown syrup of

terpenoid containing material. Silica gel chromatography, using

as an elutant i'$> methanol in chloroform, provided several fractions,

U',

116

Germocrolide

( I )M e l omp o l i d e

(II)

Hefiongolide

( III)

U

12

(IV)

117one of which was further purified by thick layer chromatography,

which provided crystalline material recrystallized from diethylether.

B . Physical Data

The white crystals had a melting point of 128° -12^'’ C and a uv

maximum at 210 nm, e * 23,000 indicating 2 or 3 Of|9_nnsaturated carbonyl

systems {e 8 ,000/chromophore). The CD spectrum showed a mlmitnum at

21c3 n m , [43 = -IbOxLO'1 and a maximum at 2 cj.) mm, “ lsxlJ'5 ’/mole cm.

The mass spectrum showed a highest mass of 507 and a series of lines dif­

fering by 11 amu. Since the gas chromatograph of the TMS derivatives

(Fig. l) indicated a mixture of two compounds, no attempt was made

to analyze the mass spectrum of the mixture.

The infra-red spectrum contained a strong band at 8. Mg ; due

to 0-H stretching vibrations. C-H stretching bands are

observed at 3.2lu (weak) due to olefinic groups and two strong

bands at 3.38 and 3.I811 due to methyl and methylene groups.

The absence of a band at 3*31., typical of symmetrical stretching

of methylene groups indicates the presence of a strained ring

system."* Bands at 'j.Yl ., ;,'..79u, 3.82u. and ‘;.y0 ■ , suggested

the presence of four carbonyl groups and olefinic stretching

bands were seen at and G.03u. Two strong C-0 stretching

absorptions are observed at j.Olu and 8.83,-. which must be due

either to a lactone or a conjugated ester. Furthermore, there

were numerous bands between Y.OOkI arid lo.JO l which were not

assigned. The IR spectrum did not indicate or preclude the presence

of an epoxide moiety, absorptions for which are generally observed/

at 8.00.1, 10.‘O n - 12. and near 12. bu. Many absorption

118

Fig. 1 GC Trace of the TMS-Derivitive of "Repandln"

Column: 9 f e e t ; 1% SE-30

Co n d i t i o n : In l e t : 170°; 5 ° / m i n , t e m p , p r o g r a m m e d ,A i r

ESI IA

0 5T i m e , m i n .

119signals in this range were present in the repandin spectrum.

The NMR spectrum of the repandln mixture is shown in Fig. 2

with signals typical of a sesquiterpene lactone. Since it was

not possible to separate the two compounds the crystalline mixture

was used for all NMR studies. The chemical shifts in several

solvents are given tn Table 1 and the coupling data are presented

in Table II.

The signals were assigned as follows: In CDC1M solution,

the broadened signal at 5.08 ppm, H-7, was irradiated resulting

in the collapse of the doublets at 6.1(7 ppm, H-1'5 , and . it ppm,tlH- l?jj • ln addition, collapse of the doublet of doublets, H-8,

at 6.25 Ppm into a sharp doublet was observed and a sharpening

of about 1.2 Hz at the doublet at 1.9l ppm, H-u, was found

upon irradiation. The experiment was repeated in CD ,CN solution

giving the same results. Irradiation of the doublet at i. 't ppm,

H-5, in a CDCiM-C(lD c solvent mixture caused the doublet at ' . 'I

ppm, H-6, to decouple.

In CD:,CN solution, (see Fig. 5), in the region of ■■.') ppm,

signals consisting of a broadened doublet at ppm

and a sharp doublet at f.9l ppm, ^-CHpOH, overlapped with the

lower field doublet, H-b. When the signal at 2.80 ppm, H-5, was

irradiated, the doublet at 4.96 ppm sharpened and irradiated at

3.85ppm, the H-5 doublet, the signal at f.‘,<l ppm, sharpened by

about 1.0 Hz and H-6 decoupled.

In CD-^CN solution, the highest field line of the doublet

at r;.95 ppm, H~9, was irradiated at low A.F. level and no effect

H - 9 " A “

s i 1

j K t A y J * w v V V \

R epandin

- COOM*0 2‘ j‘

1 O - ^ C H C H j R t p o n d i n - A

C H j9 2' 3‘ 4’

O-C —CMCHgCHj R«pondin-B1' C H 3 0

3yCH3

CHgON

C D C I 3 100 M H l

n n»2 -Me B

2 ' - M «

'I^ iI■ ■;, tjf- . j s J \

' v ' w J v ' y ^

Spin T ic k l in g

Irr. 3 12

V 'A ^ A A J

h - 6

H - 3 - X H - »

Irr. 9 . 8 7

Irr 3 .18

Irr. 6 . 7 3

Repondtn A Q B

C D j C N100 MHz

Normal Irr. 2 . 6 0

JNormal

S

J

10’ COQMa4 - C H2OH

I H* 6 H-5

5.00

3"-Ma 8 CO^HCN

I H-130 I7 . 0 0 6 . 0 0 5 . 0 0 4 . 0 0 3 . 0 0 2.00

S'-Ma’a”2 M a

It

1.00 ro

Fig. 3

Repandin-A and

B CD3CN

TABLE I 1H-NMR Parameters* of Repandin A and B in Various Solvents Chemical Shifts

Proton Assignment CDC13 C^Dg Acetone-de CD,CN Dioxane-da

H-l 6.72 6. 56 t.23 6,73 6.62H-2 2.65 i 2.3 -- -- -- --H-5 ~ 2.3 -- -- -- --C- -CHsOH 2.97 -.52 ^*91» 2.93 b.9l, b.Ot b.37H-5 3-9b 5 • 66 3-92 3-37 3.33h-6 *•.91 6.02 u • 39 ! m 1: • b.80H-T 5. OS 6.05 3.1b ;. n 3.03h -6 r P v • C. j 6.62 c.23 6.17 6.17h-9(a ) 6.06 - 6.31 6.0b 5.35 5.96h-L'(b ) 6. Ob 6.0b ; *) 5 5.9bH-l3a 6.27 C .23 6.17 6.20 6.20H-lJb 5.36 5.26 5*79 5.80 5.68C-lO-COOMe 2.3^ 5.27 3.76 3.6c 3.75H-2'(A) 2.bj 2.53 -- -- --2'a-Me(a ) I.05 0.9b 0.99 0.97 0.992 'b-Me(A) l.ob 0.97 0.99 0.93 0.99H-l '(b ) 0 .76 0.71 0.75 } ■ 75 0.7b£ '-Me(B) 1.01 O .96 0.97 o.9c 3.962"-CH. OH '-.15, b.17 2.26, U M b.02, b.13 b.oi b.01H-5" " 6.07 6.19 6.27 0.07 b.08; "-Me 2. jO 1, /9 — £.->3 r ■.; 0 6.27

122

TABLE II Repandtn A and B: Coupling Constants125

Assignments Coupling Constants Hz

P 12 Hz , 5 Hzj 4-ch^oh gem 1.7 Hz [CD;,CN]J 5-6 9-5 HzJ r- 7 ~ 1.2 HzJ ~r- e 1.8 Hz [CDmCNJJfl-9 9.r; Hz

T o Hzj ^"CH£0H,4” ~ 0.8 Hz■V'CH^OH GEM 11.0 HzJ ?"CHsOH> i" - 1.5 Hz

■ V ( a ),3 '(a ) 7.0 HzJ?'(b ) , p’Me( B) 7.0 HzJ3 t(B)f4i(B) 7.') HzJy.lja 1.7 Hz

2.0 Hz

J ■ »^CHpOH - 1.0 Hz

I2it

could be detected. However, when the low-field signal was

Irradiated at the same power level a distinct tickling effect, two

lines splitting to four, on the high field signal of the doublet

of doublets centered at 6.17 ppm, H-8, was observed. When

the doublet of doublets at 6.75 ppm, H-l, was irradiated the doublet

at 5.9 ) ppm, H-: ), sharpened due to the ally lie coupling

between the two protons. Unfortunately, an accurate measurement

of the coupling parameters could not be obtained. This experiment

was also done in CDCl) with the same results. In CDCl-,, when the

decoupler was set at ppm and 2.80 ppm, H-2 and H-2', H-l

collapsed into doublets.

From correlation with other samples, it was derived that

the doublet of doublets at 0.76 ppm (CDCl3 ) is due to H-l and

the chemical shift indicated an olefinic proton els to a carbonyl

group. The two doublets at 6.67 ppm and 'j.86 ppm were assigned

the exocyclic methylene proton absorptions common to sesquiterpene

lactones. With this information, the resonances of a ten-

carbon ring skeleton with an a,^-unsaturated -y-lactone ring

were assigned.

It now remained to assign the other NMR signals which must be

due to chains attached to the medium ring. In CDCl3 , a doublet

at 2,00 ppm, typical of a methyl group attached to an

olefin carbon adjacent to a carbon with one proton, was irradiated.

Decoupling was observed at the complex multiplet at 6. it 7 ppm,

and the AB-pattern at ii.l5 ppm (2''-CH^OH) sharpened.

The experiment was repeated in a CDC13-C(?D,', solvent mixture with

125the same results (Fig. 5)> When the signals at I+.I5 ppm were

irradiated, a 15$ NOE was found at H-3" and when the decoupler was

set at 6 .3O ppm, a small NOE could be observed for the 2"-CHpOH

signals, indicating a cIb-relationship between H-31' and 2"-CH20H suggesting the presence of a side chain representing the sarractnic

acid moiety. This was verified by the mass spectral data,

as outlined later.

As mentioned before, the repandin sample represented a mixture

of two compounds, repandin-A and repandin-B which showed

almost identical NMR spectra. The difference between the two

compounds appeared to rest in the composition of one of the side

chains. At about 1.03 ppm, two overlapping doublets were seen

giving integrations for two methyl groups. When these were irradiated,

decoupling was observed at about 2.5 ppm; the doublets decoupled

to singlets when the 2.3 ppm region was Irradiated. From this

together with the mass spectral data it was concluded that these

signals are due to an isobutyrate side chain, A smaller doublet

was observed at about 1.00 ppm together with a triplet at

slightly higher field. This pattern was consistant with a a-methyl-

butyrate side chain which agreed with the mass spectral data for

the TMS-derivative of repandin-B. This also explained the

difference of 1U mass units in the lines in the mass spectrum of

the mixture sample.

At this point, the positions of the side chains are still

unknown. The difference in side chains, isobutyrate or 2-methyl-

butyrate in the respective repandins, was used to determine their

126positions. It could be expected that the proton on the carbon

at which these are attached, should have slightly different chemical

shifts. It was found that H-9 consisted of a set of two doublets.

This splitting did not decouple upon irradiation of any other

proton signals in the spectrum, except H-8, and remained different

when the spectrum was taken in different solvents. It can be

assumed that, the more intense H-9 signals observed at lower

field is due to the H-‘J doublet of repandin-A and the other doublet

Is due to H-9 of repandin-B (see Fig. 2). In contrast, the H-0

resonances represent a broadened doublet which sharpens when

the H-7 resonances are irradiated. Also, the signals for the

carbomethyoxy group at C-10 appear to be slightly separated.

From the preceeding discussion one may conclude that the

Isobutyrate and a-methylbutyrate groups of the two repandins are

attached to C~9 and the sarracinates are attached to C-u or C-8.

More evidence for this conclusion was obtained from shift

reagent studies.

In C^De, the signal due to the ^-CH^jOH was irradiated

resulting in a 15$ NOE of the H- 9 resonance. This value was

obtained by integrating the H-9 signal while irradiating the

^-CH^jOH signals and 70 Hz off-resonance irradiation, with

subsequent subtraction of the integrals. The value reported was

the result of ten determinations; statistical considerations

Indicate an error range of + NOE. The experiment was

repeated by observing the H-9 signals at a 100 Hz sweep width

before and after irradiation of the ii-CH^OH signal and subsequent

Rtpo

ndin

Af

iB

100M

Hz

Irro

diot

ion

Ex

p*r

im«n

t«

Fig. b Repandin-A and B Irradiation Experiments

127

4 r

o u£ + o -P z «oo

0lito

i1 u +O) «"S —b <->

°cw

<n

aEeoWu

O _* o oo

0fO1Xa

o*> uo

oo0

4 -

o

. ^ = _ Ir

Rop

ondi

nA

ftB

1

00

MH

z Ir

radi

atio

n E

xp

eri

men

ts

128

Fig, 5 Repandin-A and B Irradiation Experiments

129superposition of the spectra after four scans, a definite

Increase in area in every case of these determinations was observed.

Therefore, the observed NOE is definitely genuine and repandin-A

and B in all probability contain a L( 5)~cls epoxide systems.

The shift reagent data is presented in Table III. The

differences between the chemical shifts of each proton in pure

CDCI3 and the chemical shifts after each addition of Eu(fod)., are

listed. It can be seen from the L.I.S. data of 2"-CH;30H that

the strongest complexation takes place at this moiety and that very

little interaction is found at the i-CHpOH. The large L.I.S.

of H-I3a , H - a n d H-8 indicate that there exists some complexation

at the lactonic carbonyl, A large L.I.S. at H-9 indicates

interaction at the isobutyrate or cr-methylbutyrate carbonyl group.

These numerous effects make internuclear distance estimations

impossible.

In several spectra of the shift reagent series it was observed

that some of the signals split into two sets of signals due to the

two-compound mixture. As shown in Fig, 6 , the two signals

due to the C-10 COQMe group are readily discernable and have the

same area ratio as the G.C. trace. Other protons showing this

effect are: H-9, H-13 , and H-13*, while H-3 , H-6 , H-7 and H-8d dshow very little or no splitting. The most probable explanation

for this effect is complexation taking place between the shift

reagent and the l'-carbonyl groups with the complexation rates

of the two different side chains being slightly different. One

might speculate that the difference in the complexation rates may

TABLE III3

Repandin A and B; Shift Reagent Data

Eu{fod)^/^, . dRepandins H-l C-9-CH20H H-5 h-6 H-7 H-8 h ~9(a )

.113 .11 108 .23 .32 .29 .90 .29

.226 .27 .12,.19 .25 .65 .58 .79 .61

.3-0 •37 .31 ■ t 9 I.03 .90 1.18 .89-^3 . .27,.20 ■91 1.35 1.20 1.56 1.29. 3O6 . cc 7 ri . ** -* > * 2 V 1.16 1.79 1.52 1-99 1.59.'■79 .81 * J * I.38 2.07 1.82 2.37 1.89.799 • 99 .55, .71 1.60 2.37 2.10 2.75 2.18.7O 5 1.3 5 .31, .51 1.51 2.70 2.10 3.13 2.99

1.019 1.13 0 5 ,.90 2.02 5-07 2.67 3.^5 2.771.139 1.28 • y • 9 5 2.21 3.36 2.99 3.97 3.021.359 1.91 .cl,1.0c 2-55 3.68 3.35 9.32 3-1+9L.ySk 1.5 c- • C y , 1 . 1 i 2.76 3-99 3.80 9.89,9.91 3.961.509 1.66 t; 1 2* 3.13 “ ™ 1.17 5.26,5.30 ~ 9.99

Values reported are chemical shifts, 5, after each addition of Euffod).. minus chemical shift of pure sample

^Molar Ratio cThe shifts are difficult to follow; some speculation is involved.dLetters A and B in parentheses refer to repandin A and B, respectively.

TABLE III (continued) Repandin A and B; Shift Reagent Data

Eu(fod)^/bRepandins H-9(B)d ClO-COOMe(A) C-lO-COOMe(B) H-13a(A) H-l3a(B) H-l3b(A) H-13M b)

.UJ .29 .08 .36 .22 .22 .23 .23

.226 .61 .It .16 .55 -55 .18 .1*8

.31+0 .39 .25 .25 .90 • 90 .76 .76

.1*53 1.2- -35 • 35 1.2t 1.26 .98 1.00* C O 1.63 . It .1*6 1.56 l.cit 1.29 1.32.'■■79 1.91+ . 36 • 55 1.91 1.97 1.53 1.58

2.23 * C 'V . 66 2.21 2.28 1.76 1.81.905 2.55 . 7b .76 2.52 2.59 2.00 2.03

I.0 I9 £.32 .90 .36 2.78 2.35 2.19 2.21+I.13I* 3.09 1.33 • 99 3.01+ 3 - U 2.39 2.1+21-359 3.53 1.23 1.18 3.13 * c 1 J * y A 2.72 2.76l.;3l* -.00 1.51 1.1*5 3.81 3.38 —1.309 as :.,U 1.71. 1.66 1..07 1 1* *■* • --

aValues reported are chemical shifts, *, after each addition of Eu(fod);; minus chemical shift of pure sample

^Volar RatioQThe shifts are difficult to foLlow; some speculation is involved.^Letters A and B in parentheses refer to repandin A and B, respectively.

TABLE III (continued) Repandln A and B; Shift Reagent Data

Fu(fod) Repandins C-2'-Me(A) C-2'Me(B) C-3 ’-Me(B) C-2"-CH^0H C-3,,-Me H-3"c

.115 .07 .O'7 .01 •CT, .75 .11 .11

.£26 .10 .15 .13 1.1T,1.1*5 .22 .26• Vo ,2c .26 .21 1.62,2.13 -33 .3^' 1; * 1. 1 * " .56 .27 2.16, — -63 —

* . r- r- . u K .lo * 9 2.82,5.75 .56 —

. "9 . 05 - 3.28,6.15 .t6 —■ '91* .■ 6 .'_:6 - J. 70 ,5,95 .76.905 . ’6 ,’76 • 6.12,5.67 .66 —

I.0I9 .02 .52 .61 6.65,6.16 -91 1.^31.156 • 9l .91 • r ' .'0 .99 1.511.^9 1.9 6 i .o i . 7 5 9.19,7.23 1.12 I . c71.566 1.21 1.21 .59 5. 53,7.60 1.26 1.801.6:9 1 * j j 1.33 0 t* y 5.86,6.2^ 1.33 1.90

‘'Values reported are chemical shifts, after each addition of Eu(fod). minus chemical shift of pure sampleMolar Ratio

QThe shifts are difficult to follow; some speculation is involved,^Letters A and B in parentheses refer to repandin A and B, respectively.

Rep

andi

n A

88

+Eu

(f

od

)2

Fig. 6 Repandln-A and B Shift Reagent Experiment133

]

)

1 -

13*

be due to differences in steric hinderance of the most strongly

complexed functional group 2"-0H. If this were the case, every

signal in the spectrum should be expected to show splitting upon

addition of Eu(fod)3 . This provides additional evidence that

the isobutyrate and 2-methylbutyrate are located at C-'j, and

H-I3 , H-I3, , and the COOMe group at C-lO are situated in space& D

closely to the 1'-carbonyl group, whereas H-3 , H-o, H-7 and H -M

are not.

The mass spectral data of the trimethyIsily1 (TMS) and

nonadeuterotrimethyIsily1 (TMS-d ,) derivatives of repandin-A and B

are shown in Table IV and Table V and the fragmentation patterns

of the sidechains are shown in Fig. f>. The parent peaks are not

visible but strong M-lB are seen, which is quite common in

TMS-derivatives. With the help of the NMR data molecular

weights of O32 and 600 were derived. If two TMS groups are taken

into consideration, the molecular weights of the original compounds

would be 3O8 and 322, respectively. In the TMS derivatives

of both compounds fragments of 1^3 mass units are lost.

A base peak at m/e lyi indicates the loss of the acylium ion of

sarracinic acid. In the TMS-d^ derivatives, M-I32 and M-l‘/(' fragments are observed corraborating the presence of a

sarracinate side chain in both repandln-A and B,

The mass spectrum of repandin-A-TMS shows strong peaks at

m/e 71 and h3 and a peak at m/e 3181 (M-88) as does the spectrum

of the TMS-d0 derivative. These peaks are due to the isobutyrate

acylium ion, the Isopropyl cation and the ion resulting from

TABLE IV Repandin-A Mass Spectral Data

m/e TMS IntensIty^ m/e TMS(d-y) Asslgnments

652 0 670 M+637 I 632 M-19 (c h 3 )621 8 639 M-31 (OCH-J

2 382 m - h h

969 3 998 M-109 (a-9)939 18 837 M-ll;)9.36 60 390 M-116926 7 962 M- 129'923 20 961 M-12'i909 2 --- M- 1 , >669 20 698 M-20 -’i14 38 83 697 M-20 6373 11 386 M-277360 28 369 M-202399 30 508 M-291301 11 810 M- ')1 > 1270 19 270 M- 032173 23 179 From D171 100 180 D-l170 to 179 D-l - H168 33 177 C,,H -A183 21 161 From D163 32 192 D-2103 9 112 A-81 61 81 From Lactone Ring73 10 82 A- 171 1‘, 71 B-l63 10 ;‘3 B-2

TABLE V Repandin-B Mass Spectral Data

136

m/e TMS Intens ity$ m/e TMS{d-9) Assignments

666 0 681; M+65I 2 666 M-15635 7 653 M- 51

3 582 M-102 (C)563 2 581 M-105 (a-3)553 18 371 M-ll'j552 39 370 M- 115

COLf\ 11 5 50 M- 12“'537 lj 55'' M- 127509 1 527 M-157

16 1+72 M-20 51+62 52 J«71 M-20 6375 17 386 M-2,'1360 3!+ 569 M- 30 0359 39 568 M- 507301 11 310 M- 565270 lb 270 M- ;V »62b9 15 269 M- 5 J7173 21 179 From D171 100 180 D-l170 1+0 179 D-l - H168 35 177 C. H -A155 20 161 From DIL3 30 152 D-285 33 85 C-l81 30 81 From Lactone

Ring73 20 82 A- I71 20 71 B-l57 18 57 C-2^3 2 Jv5 --

157

Fig. 7 Mass Spectral Fragments of Side Chains, "Repandins*

0 or C

A -3 A -2103 89

C H2- | o | s i ( C H3)3

B B-l B-2 88 71 43

H 0 f { j ^ H - ° H 3

o c h 3

C c-l C-2 102 85 57

H O TM0

f HC H.

D D-l 188 171 143 D -2

HO CH-

( C H 3)3 S I 0 cti2

1?8

the loss of isobutyric acid, respectively. The mass spectrum of

repandin-B-TMS shows strong peaks at m/e 85 > 57 > and 56*1 (M-102)

as does the TMS-da derivative which represents the a-methylbutyrate

analogs. The mass spectrum of repandin-A-TMS does not contain

peaks at m/e 85, and 57 and that of repandin-B-TMS shows no signi­

ficant peaks at m/e 71 and h1}.With the exception of the above signals the mass spectra of

the derivatives of the two repandins give identical fragmentation

patterns as can be seen from Tables I V and V . The fragmentation

of the ten-membered ring occurs with rearrangments thus making

assignments difficult and proposing mechanisms and structures would

be purely speculative even with the aid of the TMS-d., derivatives.

However, assignments of some signals were attempted.

The M-3I-fragments are most probably due to the loss of

a methoxy group from the C-10 carbomethyoxy moiety. The loss

of 105 mass units is due to the loss of the C-7 CH;OTMS.

The loss of m/e 115, ll t, and 128, 129 could be due to clevages

between the an^ carbon bonds leading to ions V and V I .

The signal at m/e 375 is present in the spectra of both repandin

derivatives indicating the loss of the butyrate groups; the TMS-d;j

spectra indicate that one T M S group is lost during this fragmen­

tation. This fragment could not be assigned but rearrangements

must have taken place. Peaks at 50Q , yj) and 301 atnu represent

a similar problem. The signal at 270 amu is present in the spectra

of both repandln-TMS and T M S - d n derivatives, which indicates

that both T M S groups and the butyrate groups are lost during

o*

A

H B

( V I )

C H 2 0 Si M e 2

- H

H(V I I I )

H O Me

( V I I )

11+0fragmentation.

The signals at m/e 173 and I55 could be formed by McLafferty

rearrangements of the sarraclnate side chain leading to the

ions VII and VIII. Mass number 81 could be due to rearrangement

of the lactone ring providing ion IX.

C. Stereochemical Considerations

It was assumed that both repandin A and B have the same

stereochemistries around the medium ring since they have such

highly similar NMR spectra. The mass spectral and NMR data are

consistent with the molecular formulae and C;,rHMll0 , j

for repandin A and B, respectively, and a gross structure as

shown in Figure 2. It was attempted to determine the stereochemis­

tries of several different structural possibilities by the use of

Dreiding models by matching the stereochemical relations with

experimentally observed NMR parameters. First the 1)-epoxide

function was set in a cis-orientation and the lactone ring arranged

in the specified position. Changes in the stereochemical positions

of the side chains were correlated with the NMR data. The orienta­

tion of the epoxide ring was adjusted to match the J ■ coupling.

Since one cannot differentiate between enantiomers only one cIs

and one trans configuration of each gross structure of the Y,f' and 7,8 lactones must be considered.

The possibility of repandin-A and B being 7-8 trans lactonized

was dismissed. The structure with a 78, lactone ring with 6of,9tt

sldechalns would result in large and J 7>(, values. In addition

H-I3 , cannot approach the butyrate side chain at C-() closely, a b

n iTherefore this structure is probably not the correct one.

Other 7>8 tranglactone stereomodels do not correlate with the

observed data. In all of the other three possible structures,

several dihedral angles of protons do not agree with the observed

couplings. It is seen that the 1'-carbonyl approaches H-9, the

C-10 COOMe, H-l7) and H-l5u hut also H-7 which, however, showsa bno sign of the splitting effect upon addition of the shift reagent.

The L,I.S.'s expected for these structures are very different from

the L. I. S.’s observed so these three orientations, Gfl - !>0, <a - 9a

and 6ft - are less likely to represent the structure of repandin

A and B.

Furthermore, 7,8-cls lactonizations do not seem to be likely

configurations for the repandins. Stipulating a 7“Of,8-a lactone,

the 6ft - 9a structure would give a large J G 7 and a small J sl ^

which do not agree with the experimental data. In addition, the

1'-carbonyl would not approach H-15 and the C-10 COOMe. The

6a,9^ stereochemistry would necessitate a small J ;, and the same9

with the 1' carbonyl group. A structure with Ca~9a suffers from

the same shortcomings as the other models mentioned.

The repandins are much more likely to represent 7~6 lactones.

The structure with a 7ft,6a lactone and 8a,9/9 sidechains is a very

likely structure. This configuration would result in small JG>7

and J7>Q values because the respective dihedral angles would be

approximately 90° in the favored conformation. The J G ■.> coupling

would be large and if the 6(5) epoxide is a to the medium ring J.,j6

would also be large. In addition the stereomodels indicate that

the butyrate side chain can approach H-lJ> . t H-9 and C-10 COOMe

U 2quite closely while remaining distant from H-8 and H-7) thus

accounting for the shift reagent data.

7,6-cia-lactones are sterically more rigid in the ten-

membered ringB. The repandins-A and B have very small values

for J y jR and J v 8 requiring the involved protons to have dihedral

angles of approximately yOu . If the carbon skeleton were not

rigid the protons could adopt dihedral angles other than the K)0

resulting in larger coupling constants. If a ofl - 78 lactone

is stipulated and the C-8 substituent is set in the a position

and the C~9 substituent in the position, the predicted couplings

agree quite well with the observed data. H-8 cai be deshielded

by the C-l(10)-double bond and possibly by the hf^epoxide function.

The 1'-carbonyl can come within 1 or 2 k from H-1 ^ ant! the 10-COOMe

without approaching H-8 or H-7. If the *4(9)-epoxide adopts the

a-orlentat ion, the J . coupling will match the observed values.

A structure even more rigid and sterically hindered is one with

the Bfl - 9# configurations. The expected couplings match the data

and the 1 '-carbonyl approaches H-l3 and the C-10 COOMe quite

closely but remains distant from H-7 and H-8 . An ql epoxide will

also allow a large J 8 G agreeing with the observed value.

The 8a,9a configuration would give a small value

but a large J7(B. In addition, the 1 '-carbonyl would approach H-7

disfavoring this configuration. The 8or9a configuration would also

give a large J7 jB an^ a small JB u and have the l'-carbonyl

approaching H-7. So this structure is also very unlikely.

According to Samek's rule an allylie coupling smaller than 3 Hz

Indicates a els-lactone fusion, ^ However, Samek considered only

seventeen samples and no cls-fused germacranolides are reported

In the literature to this date. A cursory survey of the literature

revealed that, in the germacranolides with trans-fused lactones,

the allyllc couplings, J 7 13, are 3.0 Hz or larger and lactonic

couplings, J o r J7j6 range from 2.0 Hz to 11.0

Schafezadeh et al.^ reported allylic couplings, J7 ly, of 2.5 Hz,

2.5 Hz, and 2.8 Hz for germacranolides with trans-lactone fusions.

However, some of the author's comments lead one to believe that the

structures are still uncertain and it is possible that these

authors are dealing with cis-fused lactones.

Examples of both trans- and cis-fused lactones can be found

among the trans-decalIn type eudesmanolides. Allylic couplings for

trans-fused lactones range from 3 .O Hz to 3*5 and for cis

fused 1actones 1.0 Hz to 2.5 Hz."^"*^ The lactonic couplings for the tranB and cis fused lactones in eudesmanolides range

from 10 Hz to 11 Hz and from 7.0 H2 to 8.3 Hz, respectively. The

guaianolides and pseudoguaianolides have allylic couplings of over

3.0 Hz except in rare cases and it is impossible to distinguish

trans and cis fused lactones from the J7 couplings. The lactonic

couplings are also large and have no correlation to the lactone

fusions. ^

The preceeding discussion of allylic and lactonic couplings

will lead one to conclude that in germacranolides both the allylic

and lactonic couplings should be smaller for a cls-fused lactone

than for a trans-fused lactone suggesting 7/ - cis lactones in

Fig. 6 Repandin-A and B, Structure of 7 6 trans Lactone, 83, 9cy

OR

CCLCH

HO-CH

0

R1 * I s obut y r a t e or 2 ~Methy I b u t y r a t e R" = Sar r ac ina t e

Fig. 9

Fig.

lit'

Re pandin-A and B, Structure of 7 £ cis Lactone,

H Hi

) Repandtn-A and B, Structure of 7.6 cis Lactone, 'Ja

OMe0

H

Lk6the repandins. However, the effects of ring strain and the fact

that the angles between vicinal protons and allylic and exocyclic

methylene protons are most probably quite different for the trans-

decalin skeletons (eudesmane) and cyclodecadiene skeletons

(germacrane) make a rigorous comparison of the two groups most

difficult.

From the use of stereomodels it seems that repandin A and B

have one of the following structures: (a) ■ /S sarracinate,

1ft, GO"lactone , OQr_isobutyrate or 2-methy lbuty rate , and a 0 epoxide,

Fig. 8; (b) a IR, C$~lactone, 8a sarracinate, 9/9 butyrate ,

Ct epoxide. Fig. 9, a 7 / 3 , lactone, 8ft saracinate, (JQ butyrate,

a epoxide, Fig. 10. However, one is easily led astray by

stereomodels and in spite of sincere attempts to be objective

eroneous conclusions must be expected due to the unfamiliarity of a

novel skeletal system. Although one of the three structures mentioned

above should be the most likely for the repandins A and B, X-ray

crystallography will be necessary for a final and unambiguous

determination of the structure.

lit?

Bibliography and Comments

1. A. Yoshioka, T. Mabry, B. Tlmmerraann, Sesquiterpene Lactones Chemistry and Plant Distribution. University of Tokyo Press,1973, P- 7 *

2. N. Fischer, R, Wiley, D. Perry, J. Org. Chem. Ul, 5936 (1976).

5. R. Silverstein, G. C. Bassler, Spectrometrie Identification ofOrganic Compounds, 2nd ed., John Wiley and Sons, New York , (1967)P . 79.

it. Z. Samek, Tet. Let. (9 ), 67I (1970).

5. N. Fischer, R. Wiley, J. Wander, J. C. Chem. Comm. , I59 (1972).6. N. Fischer, R. Wiley, H. N. Lin, K. Karimian, S. Politz ,

Phytoehem. U , 225 (1975).

7. D. Perry, N. Fischer, J. Org. Chem. , hO, 31,80 (1975).

8 . K. Tori, I. Horibe, K. Kuriyama, H. Tada, K. Takeda, Chem. Comm.1395 (1971).

9. W. Herz, P. S. Kalyanaraman, J. Org. Chem. 5 0 , 3286, (1*979)-

10. N. Fischer, R. Wiley, D. Perry, K. Haegele , J. Org. Chem. 51,3956,(l97b).

11. B. Drozydz , Z. Samek, M. Holub, S. Herout, Col1. Czech. Chem. Comm.1 8 , 727, (1973).

12. M. A. Irwin, T. A. Ge1smann. Phy tochem. 12. 871, (1973).

13- F. Shafizadeh, N. Bhadane, Phytochem. 12, 897, (1973)*

lb. S. Matseuda, T. A. Geismann, Tet. Let., 2159, (1967).

15- R. W. Doskotch, F. S. El Feralay, J. Org. Chem. 35, 1928, (1970).

16. T. A. Gelssman, T. S. Griffin, M. A. Irwin, Phytochem. 8 , 1297(1969).

17. M. A. Irwin, T. A. Geissman, Phytochem. 12, 871, (1975).

18. M. A. Irwin, T. A. Geissman, Phytochem. 10, 037, (I07I).

U 819. W. Herz, N. Vlawanathan, J. Org. Chem. 29. 1022, {I969).

20. J. A .Marshal, N. Cohen, J. Org. Chem. 29, 3727, (1969).

21. W. Herz, Y. Sumi, V. Sudharsaman, D. Raulals, J. Org. Chem. 32,3658, (1969).

22. H. Yoshioka, T, Mabry, W. Herz, Org. Chem. 33, 627 (1970),

23. Concluded from the spectra given In reference I.

VITA

Gary Paul Juneau was born on December 22, Lyk'f in New Orleans,

Louisiana the son of Mr, and Mrs, Morris Juneau. He was educated in

New Orleans at Holy Rosary Elementary School and Cor Jesu High School

which were operated by the Sisters of Saint Joseph and the Brothers

of the Sacred Heart, respectively. In ly68 he entered Louisiana State

University in New Orleans and graduated with a Bachelor oi Science

degree in August of l'/fO. From September of T/fO to April of l'/{2

the author served in the second batallion of the sixteenth artillery

and was stationed in Nurnberg, Germany. From June of l'/{2 to August

of 1'jTj he worked in the field of artheriosclerosis at Louisiana State

University School of Medicine in New Orleans under Dr. Gerald

Berensen. Shortly thereafter, he entered the Graduate School at

Louisiana State University in Baton Rouge and is currently a candidate

for a degree of Doctor of Philosophy in the Department, of Chemistry.

EXAMINATION AND THESIS REPORT

Candidate:

M ajo r Field:

I itle of Thesis:

Gary Paul Juneau

Chemis try

Structure Elucidation by Nuclear Magnetic Resonance Spectroscopy

A pproved

E X A M IN IN G C O M M IT T E E :

S'S ts&r** M a jo r P ro fessor and C h a irm a n

D ean of the G ra d u a te School

Date of Exam ination:

June 10, 1977