AN ABSTRACT OF THE THESIS OF
Brian L. Marguez for the degree of Doctor of Philosophy in Pharmacy presented jj
29, 2001. Title: Structure and Biosynthesis of Marine Cyanobactenal Natural
Products: Development and Application of New NMR Methods.
Abstract approved:
Redacted for Privacy
William H. Gerwick
This thesis is an account of my explorations into the field of natural products
chemistry. These investigations led to the discovery of several novel secondary
metabolites isolated from the marme cyanobacterium Lyngbya majuscula. In addition,
biosynthetic investigations were undertaken using stable isotope-labeled precursors.
The dominant role that NMR spectroscopy plays in the field of natural products
chemistry has led to the development of several novel pulse sequences.
Hectochiorin was discovered during a phytochemical investigation of a
cultured L. majuscula originally collected off the coast of the Caribbean Island,
Jamaica. The absolute stereochemistry was determined by X-ray crystallography.
Through a series of biological evaluations, this compound was found to stimulate actin
polymerization.
The j amaicamide class of compound was isolated from the same organism that
yielded hectochlorin. The structures were elucidated utilizing a variety of NIMR
methods, including a newly developed pulse sequence. Because the producing
organism was in culture, a biosynthetic pathway investigation ensued to elucidate the
carbon framework in jamaicamide A.
The marine natural product barbamide is intriguing due to the incorporation of
a trichioromethyl group into its molecular constitution. Further investigation into the
timing of the chlorination reaction has been pursued. In addition, the isolation of
dechiorobarbamide and the determination of the absolute stereochemistry assignment
of barbamide was accomplished.
A reevaluation of the stereochemistry of antillatoxin necessitated a correction
in the original assignment. Four antillatoxin stereoisomers were obtained from a
collaborator and found to possess differing levels of biological activity. The three
dimensional solution structures of these isomers were evaluated in an effort to
understand the role these stereochemical features play in the observed bioactivity. The
structures were determined utilizing NMR-derived constraints applied to molecular
modeling calculations.
The development of two new pulse sequences for the determination of long-
range heteronuclear coupling constants was also accomplished. The 1,1 ADEQUATE
experiment was modified to yield an ACCORDIAN experiment which can be
optimized to observe of a wide range of '.1cc couplings. This new experiment is
demonstrated for a model compound as well as for the new marine natural product
jamaicamide A.
©Copyright by Brian L. Marquez
June 29, 2001
All Rights Reserved
Structure and Biosynthesis of Marine Cyanobacterial Natural Products:
Development and Application of New NMR Methods
Brian L. Marquez
A THESIS
submitted to
Oregon State University
in partial fulfillment of
the requirements for the degree of
Doctor of Philosophy
Presented June 29, 2001
Commencement June 2002
Doctor of Philosophy thesis of Brian L. Marguez presented on June 29. 2001.
APPROVED:
Redacted for Privacy
Major Professor, representing Pharmacy
Redacted for Privacy
Dean ofthdCollége of
Redacted for Privacy
Dean of the"diadtie School
I understand that my thesis will become part of the permanent collection of Oregon
State University Libraries. My signature below authorizes release of my thesis to any
reader upon request.
Redacted for Privacy
ACKNOWLEDGMENTS
I would like to express my sincere thanks and appreciation to my major advisor,
Dr. William H. Gerwick. Dr. Gerwick has been a tremendous source of inspiration in my
journey through graduate school. I would also like to thank Dr. Philip Proteau for
extremely helpful and insightful conversations, in addition to being a good friend. I thank
Dr. Victor Hsu for his friendship and insights throughout my undergraduate and graduate
years at Oregon State University. I also thank Dr. George Constantine for being a great
inspiration. Additionally, I would like to express my gratitude to Dr. Richard Thies for
taking his time and effort in serving as my Graduate School Representative.
I would like to thank Dr. R. Thomas Williamson for both his extraordinary role as
a mentor and for being a great friend. For her wonderful friendship I thank Ana Carolina
Barrios Sosa.
I would also like the thank Robin Geralds for introducing me to the field of marine
natural products. I would like to thank Lisa Nogle for always being a great friend, and
also her critical reading of this manuscript, several times. In addition, I would also like to
express my appreciation to Ken Milligan for being a good friend from the beginning. I
would also like to thank Dr. T. Mark Zabriskie for his encouragement and support.
I would like to thank Brian Arbogast (Department of Chemistry, OSU) for mass
spectral data, Dr. Alexandre Yokochi (Marine/Freshwater Biomedical Sciences Center)
far assistance in collecting the X-ray diffraction data. In addition, Rodger Kohnert for
many great discussions about NMR spectroscopy and computers.
To my wife, I can never thank you enough for the patience and unconditional
support you have given me. Thank you! Also, thanks to my son Evyn, for just being him.
CONTRIBUTION OF AUTHORS
Chapter II: K. Shawn Watts acquired x-ray diffraction data and solved the
crystal structure of hectochiorin. Drs. Pascal Verclier-Pinard and Ernest Hamel
conducted the studies on the stimulation of actin polymerization.
Chapter III: Dr. R. Thomas Williamson acquired the ACCORD-ADEQAUTE
and the 1H-'5N HMBC data for jamaicamide A. Lisa Nogle assisted in the feeding,
isolation, and acquisition of 13C NMR data for the biosynthesis studies ofjamaicamide
A. In addition she completed the isolation and structure elucidation ofjamaicamide C.
Chapter IV: Dr. Namthip Sitachitta performed the experiments that are noted
as "Review of previous feeding studies" in the chapter. The laboratory of Dr.
Christine L. Willis at the University of Bristol provided all chirally labeled substrates
for the biosynthetic feeding experiments.
Chapter V: Drs. Shioiri and Yokokawa of Nagoyo City University provided
the four stereoisomers of antillatoxin. Dr. Philip S. Magee of the BioSAR Research
Project completed the AMI calculations. Drs. Tatsufumi Okino, Fred Berman, and
Tom Murray provided the bioassay data.
Chapter VI: Collaboration with Dr. Thomas Williamson resulted in the
development of the HSQMBC experiments.
Chapter VII: Collaboration with Dr. Thomas Williamson resulted in the
development of the ACCORD-ADEQUATE experiment.
TABLE OF CONTENTS
CHAPTER I: GENERAL INTRODUCTION
CHAPTER II: STRUCTURE AND ABSOLUTE STEREOCHEMISTRY OF
HECTOCHLOR1N, A POTENT STIMULATOR OF ACT1N ASSEMBLY
Abstract
21
Introduction
22
Results and Discussion
23
Experimental
39
References
43
CHAPTER ifi: ISOLATION, STRUCTURE ELUCIDATION, AND
BIOSYNTHESIS OF THE JAMAICAMIDES
Abstract
45
Introduction
46
Results and Discussion
48
Experimental
70
References
78
CHAPTER IV: THE STRUCTURE ELUCIDATION OF
DECHLOROBARBAMIDE AND BIOSYNTHETIC INVESTIGATIONS OF
BARBAMIDE
Abstract
80
Introduction
81
Results and Discussion
85
Experimental
94
References
100
TABLE OF CONTENTS (CONTINUED)
CHAPTER V: THREE DIMENSIONAL SOLUTION STRUCTURES OF
ANTILLATOX1N AND THREE OF ITS STEREOISOMERS
Abstract
103
Introduction
105
Results and Discussion
108
Experimental
128
References
130
CHAPTER VI: THE HSQMBC EXPERIMENTS AND THEIR APPLICATION
TO THE STEREOCHEMISTRY OF NATURAL PRODUCTS
Abstract
133
Introduction
134
Results and Discussion
139
Experimental
153
References
155
CHAPTER VII: ACCORDIAN OPTIMIZED 1,1-ADEQUATE
Abstract
159
Introduction
160
Results and Discussion
161
Experimental
169
References
170
TABLE OF CONTENTS (CONTINUED)
CHAPTER VIII: CONCLUSIONS
172
BIBLIOGRAPHY
179
LIST OF FIGURES
Page
Figure
1.1
Summary of the fundamental biosynthetic building blocks forming
curacin A (9) as identified from various stable and radioactive isotope
precursor feeding studies (19,20).
9
11.1
ORTEP'7 representation of hectochlorin (1).
28
11.2
Effects of hectochIorin (1), lyngbyabellin B (4), and jasplakinolide (5)
on the actin cytoskeleton of PtK2 cells.
32
11.3
Stimulation of actin polymerization by hectochlorin (1) or
jasplakinolide (5).
35
11.4
Dose-response curves for hectochiorin in the NCI 60-cell line assay
38
ffl.1
Structures of malyngamide Q (1), hectochiorin (2), lyngbyabellin A
(3), and barbamide (4).
47
ffl.2
Partial structures A-G derived from HSQC and HSQC-COSY.
49
ffl.3
Partial structures H and I.
50
ffl.4
Partial Structure ofjamaicamide A including key ACCORD
1,1-ADEQUATE and 'H-'5N HMBC correlations.
51
ffl.5
Structure and 13C NMR spectra of 1 1-bromo-undec-lOynoic acid amide.
52
ffl.6
Structures ofjamaicamides A (5), B (6), and C (7).
53
ffl.7
Two-dimensional plot of the ACCORD-ADEQUATE of
jamaicamide A.
55
ffl.8
Structures of microcolin A (8), ypaoamide (9), and dolastatin 15 (10).
56
ffl.9
'3C NMR spectrum ofjamaicamide A at natural abundance.
61
ffl.io
NMR spectrum ofjamaicamide A isolated from cultures provided
with [1-' 3C]acetate.
62
ifi. 11
'3C NMR spectrum ofjamaicamide A isolated from cultures provided
with [2-'3C}acetate.
62
'3c
LIST OF FIGURES (CONTINUED)
Page
Figure
111.12
Catabolic fate of alamne via transamination and decarboxylation
to acetate.
63
111.13
13C NMR spectrum ofjamaicamide A isolated from cultures
provided with S-[3-13C}alanine.
64
ffl.14 '3C spectrum of isolated jamaicainide A from L. majuscula
supplemented with ['3C3,15N]3-alanine.
65
ffl.15 '3C NMR spectrum ofjamaicamide A isolated from cultures
provided with S-[methyl-'3C]methionine.
66
111.16 Summary of biosynthetic precursors ofjamaicamide A (5).
67
IV.1
Structure of barbamide (1) and dechiorobarbamide (2).
81
P1.2
Chemical structures of dysidin (3) and a trichlorodiketopiperazine (4).
82
P1.3
Biosynthetic hypotheses for the formation of barbamide; pathway
A, chlorination predicted to occur during biosynthesis of leucine from
pyruvate; pathway B, chlorination is believed to occur by novel
mechanisms acting directly on leucine.
83
IV.4
13
NMR spectra of barbamide (1) produced byL. majuscula
culture 19L a) supplemented with [2-13C]-5,5,5-trichloroleucine, and
b) natural abundance control [C-4 of barbamide is indicated (deriving
from C-2 of [2-'3C-5,5,5-trichloro1eucine] (C-4 = major amide isomer;
C-4' = minor amide isomer).
91
P1.5
Summary of biosynthetic precursors of barbamide (1).
92
V.!
Structure of natural antillatoxin with the predicted 4S,5R stereochemistry. 109
V.2
DPFGSE 1D NOE spectrum of natural antillatoxin with selective
irradiation at H5.
110
V.3
Four possible stereoisomers about the C4-05 bond.
ill
V.4
Spacefilling representation of the AM1 minimum for 4R,5R antillatoxin. 113
LIST OF FIGURES (CONTINUED)
Page
Figure
V.5
(a) Twenty overlaid structures taken from the Monte Carlo search of
the constrained energy minimized structure of 4R,5R antillatoxin.
117
V.6
(a) Twenty overlaid structures taken from the Monte Carlo search
of the constrained energy minimized structure 4RSS antillatoxin.
119
V.7
(a) Twenty overlaid structures taken from the Monte Carlo search
of the constrained energy minimized structure 4S,5S antillatoxin.
121
V.8
(a) Twenty overlaid structures taken from the Monte Carlo search
of the constrained energy minimized structure 4S, SR antillatoxin.
123
V.9
All models are displayed looking down the C4-05 bond axis.
124
VI.l
Structures of cyclosponn A (1), okadiac acid (2), strictosidine (3),
antillatoxin (4), scytonemin (5), and strychnine (6).
136
VI.2
The structure of kalkitoxin showing the absolute stereochemistry.
137
VI.3
The 2-dimensional NOESY (800 ms) spectrum of 300 jtg of kalkitoxin. 138
VI.4
The HSQMBC experiment; thin and thick bars represent 90° and
1800 pulses respectively;
140
VI.5
(a) The 2-dimensional HSQMBC spectrum of 353 mM strychnine
CDCI3;
in 500
141
VI.6 The G-BIRDR-HSQMBC; thin and thick bars represent 90° and 180°
pulses respectively;
142
VI.7
(a) The 2-dimensional (i-BIRDR-HSQMBC spectrum of 353 mM
strychnine in 500 j.tL CDC13;
143
V1.8
The 2-dimensional G-BIRDx-HSQMBC spectrum of -300 j.tg of
kalkitoxin.
The 2-dimensional E.COSY spectrum of --300 j.tg of kalkitoxin.
146
VI.9
VI.10 Six possible rotamers for the J-based configuration analysis of the
C7-C8 positions of kalkitoxin (7).
147
148
LIST OF FIGURES (CONTINUED)
Page
Figure
VI.l I Six possible rotamers for the J-based configuration analysis of the C8-C9 149
positions of kalkitoxin (7).
VL12 Six possible rotamers for the f-based configuration analysis of the
C9-Cl0 positions of kalkitoxm (7).
150
VI.13 Representation of rotamers about C7, C8, C9 and ClO with depiction
of all heteronuclear and homonuclear couplings that were used to
define the relative stereochemistry at C7, C8 and ClO using the
f-based configuration approach.
151
VI.14 Differences in '3C NMR shifts between natural kalkitoxin (1)
and four synthetic kalkitoxin stereoisomers.
152
VI.15 CD spectrum of natural kalkitoxin and both (+)- and (-)-synthetic
kalkitoxin (MCOH).
152
VII.1
The pulse sequence for the ACCORD-ADEQUATE; thin and thick
bars represent 90° and 180° pulses respectively;
VII.2 The structures of ethyl trans-crotonate (1) and jamaicamide A
(2).
ADEQUATE and the (b)
utilizing
ethyl
trans-crotonate as a model
ACCORD-ADEQUATE
compound.
161
162
VI1.3 Two-dimensional plots of the (a) 1,1
163
VH.4 Two-dimensional plot of the 1,1 ADEQUATE.
165
VlJ.5
Two-dimensional plot of the ACCORD-ADEQUATE.
166
LIST OF TABLES
Table
Page
II.!
'H and '3C NMR spectral data (in ppm) for hectochiorin (1)
with HMBC correlations.
25
11.2
Space group, unit cell, data collection, and refinement statistics
for hectochiorin (1).
29
11.3
Effects of hectochlorin (1), lyngbyabellin B (4), and jasplakinolide
(5) on cell growth, actin polymerization,, and displacement of
fluorescein isothiocyanate (FITC)-phalloidin from actin polymer.
31
III.!
'H and '3C NMR spectral data (in ppm) for jamaicamide A
(5) with HMBC and ACCORD 1,1-ADEQUATE correlations.
58
111.2
'H and '3C NMR spectral data (in ppm) for jamaicamide B
(6) with HMBC correlations.
59
111.3
'H and '3C NMR spectral data (in ppm) for jamaicamide C (7).
60
111.4
Table of relative enhancement of carbons injamaicamide A
enriched by isotopically labeled feeding experiments (see results
and discussion and experimental sections). The method for
the quantitation is detailed in the experimental section.
68
IV.1
'H NMR (600 MHz, DMSO) and '3C NMR (150 MHz, DMSO)
data for the major conformer of dechlorobarbamide (2).
86
V.1
Biological evaluation of antillatoxin stereoisomers for ichthyotoxicity
and neurotoxicity.
112
V.2
NMR data for the 4R5R-antillatoxin isomer.
116
V.3
NMR data for the 4R5S-antillatoxin isomer.
118
V.4
NMR data for the 4S5S-antillatoxin isomer.
120
V.5
NMR data for the 4S5R-antillatoxin isomer.
122
LIST OF ABBREVIATIONS
ACCORD
Accordian
AMB
Actin Monitor Buffer
ASU
Asymmetric unit
COSY
Correlated Spectroscopy
d
doublet
DCAO
7,7-dichloro-3-acyloxy-2-methyloctanoate
DHIV
Dihydroxyisovaleric acid
DPFGSE
Double Pulsed Field Gradient Spin Echo
DNA
Deoxyribonucleic Acid
FITC
Fluorescein isothiocyanate
HMQC
Heteronuclear Multiple Quantum Coherence
IIMBC
Heteronuclear Multiple Bond Correlation
HETLOC
Heteronuclear Long-range Coupling
HECADE
Heteronuclear Couplings from ASSCI-Domain experiments
with E.COSY
HSQC
Heteronuclear Single-Quantum Correlation
HSQC-COSY
Heteronuclear Single-Quantum Correlation-Correlated
Spectroscopy
HSQC-TOCSY
Heteronuclear Single-Quantum Correlation-Total Correlation
Spectroscopy
HSQMBC
Heteronuclear Single Quantum Multiple Bond Correlation
HRFABMS
High-Resolution Fast Atom Bombardment Mass Spectrometry
IC
Inhibitory Concentration
IR
Infrared
LD
Lethal Dose
m
Multiplet
MS
Mass Spectrometry
LIST OF ABBREVIATIONS (CONTINUED)
NC!
National Cancer Institute
NMR
Nuclear Magnetic Resonance
NOE
Nuclear Overhauser Effect
NOESY
Nuclear Overhauser Effect Spectroscopy
obs
Obscured
Polymerization Inducing Buffer
RPHPLC
Reverse Phase High Pressure Liquid Chromatography
ROESY
Rotating Frame Overhauser Effect Spectroscopy
s
singlet
SAR
Structure Activity Relationship
SAM
S-adenosyl methionine
SPE
Solid Phase Extraction
t
Triplet
TLC
Thin Layer Chromatography
UV
Ultraviolet
VLC
Vacuum Liquid Chromatography
This thesis is dedicated to my best friend, Suzanne.
STRUCTURE AND BIOSYNTHESIS OF MARINE CYANOBACTERIAL
NATURAL PRODUCTS: DEVELOPMENT AND APPLICATION OF NEW
NMR METHODS
CHAPTER I
General Introduction
Chemistry derived from natural sources has played a pivotal role in human
health dating back to ancient cultures around the globe. Traditionally these
compounds have been harvested from teffestrial sources, such as plants and
microbes. However, a prolific number of biologically relevant compounds isolated
from the marine environment are providing a rich source of novel secondary
metabolites. The oceans contain an enormous amount of biodiversity with respect
to the vast numbers of orgarnsms, in addition to a wide range of habitats in which
these organisms thrive. It has been estimated that the number of taxonomically
different organisms that reside in the worlds oceans to be upwards of 10 million.1
These organisms flourish in a wide range of habitats including intertidal reefs and
deep-water environments.
Of the estimated 10 million organisms that flourish in the marine
environment, only a small number have been explored for their chemistry. Of the
more than 10,000 compounds reported, the majority has been isolated and
described from marine algae, sponges, soft corals, sea squirts, and bryozoans.2
Structural classes that have been identified from marine sources range from the
2
daunting polyether compounds such as okadaic acid (1), to terpene-derived
compounds such as halomon (2). These two compounds are structurally diverse,
however, they represent compounds which contain extremely potent biological
activities. Okadaic acid, isolated from two different species of dinoflagellates, is a
potent tumor promoter and has been identified as the major causative agent of
diarrhetic shellfish poisonings.3 Halomon represents a frequently occurring class of
compounds isolated from red
algae.4
The common feature is the monoterpene
backbone coupled with a high degree of halogenation. This compound has very
potent antitumor properties; however, drug development of this compound has been
hampered by difficulty in its synthesis,5 as well as a limited supply of the
compound from the natural source.4
OH
OH
OH
Okadaic Acid (1)
Br
Halomon (2)
Cl
3
An exciting development in the field of marine natural products has been
the emergence of cyanobacteria (blue-green algae) as a source of biologically
potent and structurally novel secondary metabolites. The pioneering work initiated,
and continued, by Dr. Richard Moore of the University of Hawaii has shown these
organisms to contain a plethora of new and diverse chemistiy. An example is the
recently reported compound apratoxin (3)6 The laboratories of Dr. William H.
Gerwick at Oregon State University have also focused on investigating the
secondary metabolites produced by these organisms in an effort to explore their
potential for new pharmaceuticals and agrichemicals. The combined efforts of
these laboratories and others have lead to a recent surge in the number of
compounds reported from these cyanobacterial
organisms.7'8
H3CD1Y'
H3CT
H3
0
H3
N-2
H3C
CH3
Apratoxin A
Of the marine cyanobacteria, Lyngbya
(3)
majuscula
has thus far been shown to
be the most prolific producer of novel secondary metabolites. Examples of the
4
structurally diverse metabolites isolated from this organism include grenadamide
(4),9
a mildly cytotoxic compound, microcolin A
(5),10 a powerful
immunosuppressive agent, and debromoaplysiatoxin (6),1 a potent activator of
protein kinase C.
Grenadarnide (4)
CH3
I
CH3 CH3 CH3
H
0 ,'
Microcolin A (5)
OH
OH
Debromoaplysiatoxin (6)
5
A primary objective of any natural products chemistry program is the search
for new treatments for disease. Of particular importance is the search for
compounds that have the potential to treat, or even cure cancer. Although the last
30 years of research in the field of marine natural products have produced no
marketed drugs, several are either in various stages of clinical trial or in preclinical
trial. An interesting theme that emerges from the mechanism of action of most
cancer chemotherapy agents isolated from cyanobacteria is their antimitotic
activity, arresting cells in mitosis.
Microtubules are highly dynamic tube-shaped
polymers composed of a tubulin heterodimers. Microtubule dynamics are
essential for chromosome movement during mitosis. Two types of interactions
between small molecules and microtubules have been described. The first is the
stabilization of microtubules (taxol, a terrestrial natural product) and the second is
through inhibition of microtubule assembly (the coichicine and ymca alkaloid site).
While their effects on microtubule assembly are opposite, they share the same
outcome, inhibition of cell proliferation at metaphase during mitosis. Two exciting
examples of cyanobacterial derived compounds acting at the ymca or coichicine
sites are the cryptophycins and the curacins, respectively.
A very promising anticancer treatment derives from the fresh water
cyanobacterium Nostoc sp. Cryptophycin 1 (7, originally called cryptophycin A),
was found to block cells at mitosis in the low picomolar range through interaction
with microtubules.12 Interestingly, cryptophycin 1 was originally isolated for its
antifungal properties. The compound imparts its biological properties through
irreversible binding at the ymca site on tubulin. A synthetic derivative,
cryptophycin-52 (8) is currently in clinical trials for its antiproliferative activity.13
It has remarkable potency against cultured human tumor cells and in animal models
(LC50 = 11 pM in HeLa cells). This compound is between 40- and 400-fold more
potent than paclitaxel or the ymca alkaloids in several human cell lines.12 Of
particular note is that tumor cells that are resistant to paclitaxel and the ymca
alkaloids due to overexpression of multidrug resistant proteins are sensitive to
cryptophycin-52.
ILJ
o
HNO -OCH3
oi)
Cryptophycin A (7)
Cryptophycin-52 (8)
7
Bioassay guided fractionation, utilizing a brine shrimp assay, resulted in the
isolation of the unique thiazoline-containing lipid curacin A.'4 Curacin A (9),
isolated from a Caribbean collection of L. majuscula, shows very potent
antiproliferative effects to a variety of cancer cell lines (IC5o's ca. 1-100 nM).
Detailed investigations into its mechanism of action revealed that it inhibited
tubulin polymerization by binding to the same site as coichicine. Extensive
structure-activity relationship (SAR) studies ensued.'5 Through the use
'7
N
OCH3
C
'7,21
Curacin A (9)
OCH3
-Th=
CH3
Curacin B (10)
OCH3
's
H
H3C
Curacin C
NCH3
(11)
N
OCH3
Curacth D
(12)
CH3
of naturally occurring (curacins B, C, D)'6'17 and synthetically derived analogs of
curacin A it was determined that the thiazoline ring and the first carbon atoms of
the lipid side chain, including the C9-C1O olefin and the ClO methyl group, were
crucial to its biological
activity.15
Unfortunately, curacin A was found to be
difficult to work with in vivo because of solubility and stability problems. In an
effort to combat these instabilities, a focused combinatorial library of Curacin A
analogs was produced.18 Several of the synthesized compounds (e.g. 13) showed
inhibition of tubulin polymerization (ca. 1 j.tM), and inhibition of cancer cell
proliferation within an order of magnitude of the natural product (ca. 250
I-s
I8
CH3
OH
Curacm A analog (13)
ÔCH3
Efforts continue in a hope to develop a useful drug based on the Curacin A
backbone. In addition, the biosynthesis of curacin A has been reported based on
incorporation of stable isotopically labeled precursors.'9'2° The biosynthetic
subunits are comprised of 8 intact acetate units, two carbons derived from C2 of
acetate and a cysteine residue.
S-Adenosyl Methionine (SAM)]
[ste]
[35S]cysteine
[methy1.3C]methionine
[2-'3C,N]g1ycine
CH3
Th'N
1
LAcetai
[1
-'3C]acetate
[2-'3C]acetate
[1,2-'3C2]acetate
[1.)3C, '802]acetate
Figure 1.1. Summary of the fundamental biosynthetic building blocks forming curacin A (9)
as identified from various stable and radioactive isotope precursor feeding studies (19,20).
A reoccurring biosynthetic theme seen in the chemistry described from L.
majuscula are the lipopeptides. The generic term "lipopeptide" has been given to
molecules that are hypothesized to derive from a combination of both polyketide
and amino acid derived moieties (see above for curacin A). Examples of these
compounds are carmabin A (14),21 ypaoaniide (15),22 and malyngamide Q (16).
A survey of nitrogen-
containing secondary metabolites isolated from cyanobacteria, by Gerwick
et al., has provided interesting insight into two underlying themes involved in the
biosynthesis of lipopeptides: 1) the condensation of polyketides to
10
LNNN
10
0
Carmabm A (14)
CH3
Ypaoamide (15)
HO
0CH30y
'6
S
OCH3
OCH3
I11
Malyngamide Q (16)
amino acid units via ester or amide linkages, and 2) amino acid moieties as starter
units for polyketide
extension.8
Structures representing the first biosynthetic theme
are typically quite small, containing di- to tetraketide units. However, polyketide
extended amino acids vary from single ketide extensions (the jamaicamides and
barbainide, Chapters III and IV) to as many as 15 acetate or propionate units (e.g.
scytophycin B, 17). Large numbers of pendant methyl groups are observed in the
structures of lipopeptides. Recent reports suggest that these methyl groups arise
11
from S-adenosyl methionine or from C2 of acetate (as shown for virginiamycin and
the jamaicamide class of compounds).24
OCH3
N
0
A
H
OCH3
0
Scytophycin B (17)
A key feature to any structure elucidation or biosynthetic investigation is
the use of nuclear magnetic resonance (NMR) spectroscopy. Advances in NMR
spectroscopy have helped to facilitate the rapid development of the field of natural
products chemistry. These advancements have facilitated rapid structure
elucidation, which in turn allows fast communication of these newly isolated
compounds to the scientific community. The two most important advancements in
NMR spectroscopy, as applied to natural products structure elucidation, are the
development of inverse-detected heteronuclear correlation experiments and pulsed
field gradients.
The first of the inverse-detected heteronuclear correlation experiments to be
described was the HMQC sequence by Bax et al.25 The experiment allowed the
detection of 'JCH correlations by virtue of the more abundant 'H nuclei. Following
12
the HMQC, Bax and coworkers described what could be considered the most useful
experiment to date for small molecule structure elucidation, the HMBC pulse
sequence.26
This experiment allows the detection of protons that are long-range
coupled to '3C, while providing sufficient suppression of protons bound to
achieving a reduction in t1 noise in the two dimensional plot.
While two
dimensional heteronuclear correlation experiments were previously utilized (e.g.
long-range HETCOR) their use of '3C as the directly detected nuclei required
significantly larger amounts of material than their inverse detected counterparts.
The use of pulsed field gradients (PFGs) in conjunction with actively
shielded NEvER probes has provided many improvements in the data acquired for
the structure elucidation of natural products.27 The main advantages of utilizing
PFGs in routine NMR analysis of small molecules are a reduction in the amount of
required phase cycling, which results in shorter acquisition times, and a reduction
mt1 noise, which prior to PFGs, often complicated the analysis of two dimensional
data.'8
An additional advantage includes the efficient suppression of undesirable
signals such as 1H-'2C ('H-'4N) coherences in inverse detected heteronuclear
experiments.27
Progress in NMIR spectroscopy has also allowed the development of
experiments to easily measure long-range heteronuclear coupling constants.
Examples of these include the HETLOC,28 HECADE,29 coupled/decoupled HSQCTOCSY,3° and the recently reported HSQMBC (see Chapter VI).3' These
experiments have facilitated the use of this long-range coupling information for the
13
stereochemical analysis of many natural products. This analysis is accomplished
using the J-based configuration analysis recently reported by Matsumori et al.32
The premise behind this analysis is that through the use of 2'3JCH and 2JHH coupling
constants one can predict the most dominant staggered rotamers in acyclic
molecules with adjacent chiral centers, and apply these to defining the relative
stereochemistry. Examples of the successful implementation of this method
include okadaic acid (1), phormidolide (18), sphinxolide (19), a novel
chloroalkene (2O), and kalkitoxin (21, see Chapter VI).36 In addition, vicinal
heteronuclear coupling constants can be used to measure torsion angles. This is
because the magnitude of the
relationship for dihedral
3JCH
angles.37
coupling constants follow a Karplus-like
Therefore, these coupling constants can be
utilized in the three dimensional structure determination employing NMRconstrained molecular modeling calculations.
As a result of the abundance of unique and bioactive natural products
isolated to date from marine cyanobacteria, as summarized above, I hypothesized
that a continued exploration of these life forms would be productive in the isolation
of additional interesting and useful molecules. In this sense, the unique molecular
diversity of chemistry that has been found in the marine environment provides
great hope for future discoveries. Therefore, continued investigations into the
secondary metabolites of marine organisms will undoubtedly yield exciting new
biological activities and structural challenges for natural products chemists and
NMR spectroscopists alike. In a similar sense, because NIMIR spectroscopy has
14
emerged as one of the most powerful techniques for studying molecular structure, I
reasoned early in my doctoral studies that there was a great potential to develop
additional experiments which could advance the field of organic structure analysis.
The thesis begins with two chapters detailing the phytochemical
investigation of a cultured L. majuscula originally isolated from Hector Bay,
Jamaica. The first chapter outlines the isolation, structure elucidation and absolute
stereochemistry of hectochlorin. The structure elucidation was accomplished via
standard one- and two-dimensional NMR techniques. The absolute
stereochemistry was determined through the use of x-ray crystallography,
incorporating the use of anomalous scattering. Hectochlorin was also determined
to possess the ability to stimulate actin assembly, equipotent to jasplakinolide.
Chapter III details the structure elucidation and biosynthesis of the
jamaicamide class of compounds. The structure ofjamaicamide A was assembled
with a variety of NMR experiments, the most crucial being the 'H-'5N HMBC and
the use of the recently developed ACCORD 1,1 ADEQUATE (Chapter VII). Since
the producing organism was successfully growing in culture, a biosynthetic
investigation was also undertaken. Using isotopically labeled precursors, all of the
biosynthetic units which form jamaicamide A have been elucidated. In addition,
the absolute stereochemistry of one of the two stereocenters was determined.
In a continued effort to elucidate the biosynthetic pathway of barbamide, chapter
IV describes efforts to deduce the substrate for the chlorination reaction leading to
the tnchloromethyl moiety in the natural product. This was done by feeding
15
synthetic [2)3Cj-5,5,5trichloroleucine to the producing organism. High levels of
incorporation indicate that leucine is the probable substrate for chlorination. In
addition, a review of previously published feeding experiments is presented. The
isolation and structure elucidation of a novel barbamide derivative is also given.
Finally, the stereochemistry of C7 of barbamide was detennined.
Phormidolide (IS)
34
0
OCH3
OCH3
OCH3 OCH3
Spinxoiide (19)
CIII CIII CI H
CIH CIH CIII
Chloroalkene (20)
'S
16
3
5
CH3 çH3
CH3
H2sSN/I NtCH3
CH3 CH3
14
15
Kalkitoxin (21)
0
5
CH3
16
Chapter V discusses structural studies on the marine neurotoxin antillatoxm.
A stereochemistry revision was suggested for the C4-05 centers, and proven
correct by comparison to the synthetically derived compound. All four
stereoisomers were synthesized and provided to our laboratory by Professors
Shioiri and Yokokawa at Nagoya City University. Biological testing of all four
isomers showed that the natural compound is the most potent. Based on the
differences in biological activity between these stereoisomers, an investigation into
the solution structures of all four compounds was accomplished by molecular
modeling studies using NMR-derived constraints. Major differences in their three
dimensional solution structures are discussed.
As previously mentioned, the use of heteronuclear coupling constants can
greatly enhance stereochemical investigations of natural products. To further
simplify this process, chapter VI details the development of the HSQMBC
experiment. Discussions of the pulse sequence and validation of the experiment on
the model compound strychnine are presented. As an example, the determination
of the relative stereochemistry of the neurotoxic compound, kalkitoxin is presented.
This relative stereochemistry was confirmed by comparison of the natural product
with synthetic kalkitoxin provided to our laboratory, again through collaboration
with Professors Shioiri and Yokokawa at Nagoya City University.
The penultimate chapter describes the development of the ACCORD 1,1
ADEQUATE sequence. This experiment is a modification of the originally
reported 1,1 ADEQUATE experiment. The experiment utilizes accordion
17
optimization to allow a sampling of a range of 13C-13C coupling constants. The
original experiment is statically optimized and correlations that are smaller or
larger than the optimized value are often missed. Validation of this experiment is
shown for the model trans-ethylcrotonate, and its utility is demonstrated for the
marine natural product jamaicamide A (see chapter III). Chapter VIII will
conclude the thesis.
18
1.
Culotta, E. Science 1994, 263, 918-920.
2. Tan, L. T. Bioactive Natural Products from Marine Algae. Ph.D. Thesis,
Oregon State University, Corvallis, OR, 2001.
3. (a) Yasumoto, T.; Oshima, Y.; Sugawara, W.; Fukuyo, Y.; Oguri, H.; Igarashi,
T.; Fujita, N. Nippon Suisan Gakkaishi 1980, 46, 1405-1411. (b) Murakami,
Y.; Oshima, Y.; Yasumoto, T. Nippon Suisan Gakkaishi 1982, 48, 69-72.
4.
Fuller, R. W.; Cardellina, J. H.; Kato, Y.; Brinen, L. S.; Clardy, J.; Snader, K.;
Boyd, M. R. J. Med. Chem. 1992, 35, 3007-3011.
5. Sotokawa, T.; Noda, T.; Pi, S.; Hirama, M. Angew. Chem. mt. Ed. Engi. 2000,
39, 3430-3432.
6.
Leusch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J.; Corbett, T. H. J. Am.
Chem. Soc. ASAP article.
7. Faulkner, D. J. Nat. Prod. Rep. 2000, 17, 7-55 (and previous articles in this
series).
8. Gerwick, W. H.; Tan, L. T.; Sitachitta, N. Nitrogen-Containing Metabolites
from Marine Cyanobacteria. In Alkaloids, in press.
9. Sitachitta, N.; Gerwick, W. H. J. Nat. Prod. 1998, 61, 68 1-684.
10. Koehn, F. E.; Longley, R. E.; Reed, J. K. .1. Nat. Prod. 1992, 55, 613-619.
11. Moore, R. E. Pure Appi. Chem. 1982, 54, 1919-1934.
12. Panda, D.; Ananthnarayan, V.; Larson, G.; Shih, C.; Jordan, M. A.; Wilson, L.
Biochemistry 2000, 39, 14121-14127.
13. Panda, D.; DeLuca, K.; Williams, D.; Jordan, M. A.; Wilson, L. Proc. Nati.
Acad. Sci. USA 1998, 95, 9313-9318.
14. Gerwick, W. H.; Proteau, P. J.; Nagle, D. G.; Hamel, E. Blokhin, A.; Slate, D.
L. .J. Org. Chem. 1994, 59, 1243-1245.
15. Verdier-Pinard, P.; Lai, J. Y.; Yoo, H. D.; Yu, J.; Marquez, B.; Nagle, D. G.;
Nambu, M.; White, J. D.; Faick, J. R.; Gerwick, W. H.; Day, B. W.; Hamel, E.
Mol. Pharmacol. 1998, 53, 62-76.
19
16. Yoo, H.-D.; Gerwick, W. H. J. Nat. Prod. 1995, 58, 1961 -1965.
17. Marquez, B.; Verdier-Pinard, P.; Hainel, E.; Gerwick, W. H. Phytochemistry
1998, 49, 2387-2389.
18. Wipf, P.; Reeves, J. T.; Balachandran, R.; Giuliano, K. A.; Hainel, E.; Day, B.
W. J. Am. Chem. Soc. 2000, 122, 9391-9395.
19. Rossi, J. V. M.S. Thesis, Oregon State University, Corvallis, 1997.
20. Sitachitta, N. Ph.D. Thesis, Oregon State University, Corvallis, 2000.
21. Hooper, G. J.; Orjala, J.; Schatzman, R. C.; Gerwick, W. H. J Nat. Prod. 1998,
61, 529-533.
22. Nagle D. G.; Paul, V. J. J. Exp. Mar. Biol. Ecol. 1998, 225, 29-38.
23. Milligan, K. E.; Marquez, B.; Williamson, R. T.; Davies-Coleman, M.;
Gerwick, W. H. J. Nat. Prod. 2000, 63, 965-968.
24. Kingston, D. G. I.; Kolpak, M. X.; LeFevre, J. W.; Borup-Grochtmann, I. J.
Am. Chem. Soc. 1983, 105, 5106-5110.
25. Bax, A.; Griffey, R. H.; Hawkins, B. L. J Magn. Reson. 1983, 55, 301-3 15.
26. Bax, A.; Summers, M. F. J. Am. Chem. Soc. 1986, 108, 2093-2094.
27. Parella, 1. Magn. Reson. Chem. 1998, 36, 467-495.
28. Uhrin, D.; Batta, G.; Hruby, V. J.; Barlow, P. N.; Kover, K. E. J. Magn. Reson.
1998, 130, 155-161.
29. Kozminski, W.; Nanz, D. J. Magn. Reson. 2000, 142, 294-299.
30. Kover, K. E.; Hruby, V. J.; Uhrin, D. J. Magn. Reson. 1997, 129, 125-129.
31. Williamson R. T.; Marquez, B. L.; Gerwick, W. H.; Kover, K. E. Magn. Reson.
Chem. 2000, 38, 265-273.
32. Matsumori, N.; Kaneno, D.; Murata, M.; Nakamura, H.; Tachibana, K. J. Org.
Chem. 1999, 64, 866-876.
20
33. Williamson, R. T. Development and application of NMR spectroscopy to
marine natural products structure and biosynthesis, Ph.D. Thesis, Oregon State
University, Corvallis, OR, 2000.
34. Bassarello, C.; Bifulco, G.; Zampella, A.; D'Auria, M. V.; Riccio, R.; GomezPaloma, L. Eur. J. Org. Chem. 2001, 39-44.
35. Ciminiello, P.; Fattorusso, E.; Forino, M.; Di Rosa, M.; lanaro, A.; Poletti, I
Org. Chem. 2001, 66, 578-582.
36. Wu, M.; Okino, T.; Nogle, L. M.; Marquez, B. L.; Williamson, R. T.; Sitachitta,
N.; Berman, F. W.; Murray, T. F.; McGough, K.; Jacobs, R.; Colsen, K.;
Asano, T.; Yokokawa, F.; Shioiri, T.; Gerwick, W. H. J. Am. Chem. Soc. 2000,
122, 12041-12042.
37. Marshall, J. L. in Carbon-carbon and carbon-proton NMR couplings; Methods
in stereochemical analysis 2; Verlag Chemie International: Deerfield Beach,
FL, 1983.
21
CHAPTER II
STRUCTURE AND ABSOLUTE STEREOCHEMISTRY OF
HECTOCHLORIN, A POTENT STIMULATOR OF ACTIN ASSEMBLY
Abstract
Hectochlorin (1) was isolated from laboratory cultures of a marine isolate of
Lyngbya majuscula collected from Hector Bay, Jamaica. The planar structure was
deduced by one- and two-dimensional NMR spectroscopy. X-ray crystallography
was used to determine the absolute stereochemistry of hectochiorin as
5S, uS, 13S, 14S. Hectochiorin is equipotent to jasplakinolide (5) in its ability to
promote actin polymerization. In addition, hectochlorin shows both a unique
profile of cytotoxicity by the COMPARE algorithm and potent inhibitory activity
towards the fungus Candida albicans. Structurally, hectochlonn resembles
dolabellin and the recently reported lyngbyabellin classes of compounds.
22
Introduction
Cyanobacteria are producers of a wide variety of structurally unique and
biologically active secondary metabolites.' A prevalent structural theme of
metabolites isolated from marine cyanobacteria is lipopeptides.' Examples include
kalkitoxin,2
the curacins,3 and the carmabins.4 Herein we report the isolation,
structure elucidation, absolute stereochemistry and biological properties of
hectochlorin (1) a unique lipopeptide, isolated from a cultured strain of Lyngbya
majuscula collected in Hector Bay, Jamaica. Hectochlorin has potent antifungal
activity against Candida albicans, an intriguing profile of antiproliferative activity
in the NCI 60-cell line assay, and is a strong promoter of actin polymerization.
Structurally, hectochlonn resembles dolabellin (2) and the recently reported
lyngbyabellins A (3) and B (4)6
23
Results and Discussion
Planar Structure. Hectochlorin (1) was isolated from a shallow water
collection of
Lyngbya majuscula
from Hector Bay, Jamaica. In the laboratory,
individual trichomes were isolated utilizing previously described procedures.7 A
unialgal culture was established through repetitive isolation and subculturing, and
was maintained at 28° C with a 16 hour light' 8 hour dark cycle in 10 L SWBG11
media supplemented with filtered air. Although the cultue was unialgal it was not
axenic. To obtain sufficient quantities of algal material for chemical analysis,
subcultures were grown in multiple 15 L sterilized Nalgene pans with 10 L
SWBG1 1 culture media (same light and temperature conditions as above).8
The crude extract (see experimental) was vacuum chromatographed over
silica gel with a gradient of EtOAc and hexanes. 'H NMIR spectra of the
subfractions revealed one fraction in particular which possessed an interesting
series of downfield singlet resonances. This fraction was further purified utilizing
C18
SPE cartridges and RPHPLC to yield hectochlorin as a glassy, pale yellow
solid. HRFABMS established an [M + Hf molecular formula for 1 of
C27H35C12N209S2
(m/z 665.1171, calculated for C27H35C12N209S2, 665.1161). One-
dimensional 'H and 13C NMR showed the presence of four carbonyls and
resonances indicative of two thiazole rings, accounting for 10 of the 11 degrees of
unsaturation implied by the molecular formula. The remaining degree of
unsaturation could be accounted for by an additional ring within the structure of 1.
24
8</
S
0H
26
5
1
0
27
170 0
0
0H
S
25
7S\
24
Dolabellin (2)
Hectochiorin (1)
HN
0
0
0
}134
0_
Lyngbyabellin B (4)
Lyngbyabellin A (3)
GBr
Jasplakinolide (5)
Inspection of the
'NMR spectrum of 1 revealed a series of up field, highly
coupled resonances indicative of an aliphatic chain. A downfield methyl
resonating at 62.09 (H38) showed HMBC correlations to a quatemary carbon at
90.4 ppm (C7) and a methylene carbon at 649.5 (C6). The chemical shift of C7
25
(90.4 ppm) was indicative of a gem-dichioro substituent as observed in dolabellin
(2) and the lyngbyabellins (3 and 4), and was consistent with the molecular formula
of 1. HSQC-COSY was used to further extend this moiety to include an additional
six carbons (C1-05, and C9, see Table 11.1), identifying this unit as 7,7-dichloro-3acyloxy-2 methyloctanoate (DCAO).
Table 11.1. 'H and '3C NMR spectral data (in ppm) for hectochiorin (1) with
HMBC correlations.
}4)a
m
1
-
-
173.0
-
2
3.16(7.4,8.9)
dq
42.6
75.1
1,3,9
2,4,5,9,10
30.9
2, 3, 5, 6
atom no.
S 'H (fin
3
5.33
4a
1.72
m
m
b
1.82
ov
5
1.69
6a
2.13
b
2.25
7
S
13gb
HMBCC
m
m
m
20.8
3,4,6,7
49.3
4, 5, 7, 8
-
90.4
-
8
2.09
s
37.2
9
1.28(7.4)
d
6,7
1,2,3
10
161.1
-
-
147.0
-
s
128.5
10, 11, 13
-
166.2
-
6.83
s
74.7
13, 15, 16, 17, 26
-
81.9
-
1.83
s
24.4
14, 15,17
1.60
s
21.9
14, 15, 16
-
160.4
-
11
12
8.15
13
14
15
16
17
15.0
-
18
19
-
-
147.4
-
20
7.90
s
127.7
18, 19,21
21
-
22
23
24
25
5.64
26
27
165.2
-
s
77.9
21, 23, 24, 25
-
71.6
-
1.31
s
26.7
22,23,25
1.34
s
25.8
22, 23, 24
-
-
168.7
-
2.17
s
20.8
26
aRecorded at 600.04 MHz. bRecorded at 150.14 MHz.
correlation with indicated carbon.
showing long-range
The remaining NMR resonances of hectochiorin (1) contained no 2'3JHH
couplings; therefore, the remaining atoms were assembled solely from 2'3JCH
HMBC data and chemical shift comparisons to known compounds. A 3JCH
coupling was observed between the carbonyl carbon (Cl, 172.9 ppm) of the DCAO
moiety and a proton at 5.64 ppm (H22). HMBC correlations were also observed
from H22 to both geminal methyl groups (C24 and C25) and the downfield
quaternary carbon C23 (71.7 ppm). The chemical shift of C23 indicated the
presence of a tertiary alcohol. Therefore, this five-carbon unit was defined as
dihydroxyisovalerate (DHIV). The pseudo-a proton of the DHIV showed an
additional HMBC correlation to a quaternary carbon at 6 165.5 (C21). Additional
HMBC correlations from proton H20 (7.93 ppm) to this same downfield resonance
(C21) as well as to two additional deshielded quatemary carbon atoms (C18 and
Cl 9) were diagnostic of a thiazole ring.
A second thiazole ring was assembled using 2'3JCH correlations from H12 (6
8.16) to ClO (6 161.1), Cli (6 147.0), and C13 (6 166.4). Connection of this
thiazole to the DCAO moiety was made via an HMBC correlation from H3 to C10.
Additionally, this thiazole unit was connected to a second DHIV unit by a 2JCH
coupling between the C14 methine proton resonating at 6.82 ppm and the
quatemary carbon C13 (6 166.2). The proton at H14 also showed HMBC
correlations to a quatemary carbon at 82.1 ppm (C15) bearing geminal methyl
groups with chemical shifts of 621.9 and 24.4 (C 16 and C 17, respectively), and the
carbonyl of an acetate moiety (168.7 ppm, C26). In contrast to the first DHIV
27
moiety described, the chemical shift of the quaternary carbon bearing the gemdimethyl group was slightly more downfield (Aö = 10.2 ppm between C15 and
C23), indicating that C15 must be attached to a more electronegative substituent.
This chemical shift difference was satisfied by an ester linkage between C15 and
C18. With this ring closure, the 11 degrees of unsaturation required by the
molecular formula were satisfied. Having accounted for all atoms and degrees of
unsaturation in the molecular formula the planar structure of hectochlorin (1) was
complete.
X-Ray Crystallography and Absolute Stereochemistry. Hectochiorin
readily formed large cubic (0.3 mm x 0.3 mm x 0.3 mm) X-ray diffraction quality
crystals from a 1:1 mixture of MeOH and H20. The unit cell dimensions were
determined to be the following: a = 12.266 A, b = 12.684 A, c = 21.415
A, a =
= 90° with a P212121 space group. A total of 5813 reflections were recorded
yielding data that allowed refinement to 0.85 A resolution. At this resolution,
localized density was observed for all non-hydrogen atoms. The structure of 1 was
solved and refined with SHELXS and SHELXL, respectively. The data collection
and refinement statistics are given in Table 11.2. The refined single crystal structure
confirms the structure determined by NIMR and MS techniques, as described above.
An ORTEP drawing representing the asymmetric unit (ASU) is shown in Figure
11.1. The ORTEP representation is drawn with an ellipsoid probability of 50%. As
observed in the ORTEP representation, a single water molecule is confined within
the macrocycle of!. Specifically, this water molecule is located within hydrogen
28
14
H2O
lfr)
18
8
22I
-
\
20
23
Figure 11.1. ORTEP17 representation of hectochiorin (1). Absolute stereochemistry
was defined from x-ray diffraction analysis utilizing anomalous scattering data.
Ellipses are drawn at 50% probability.
bonding distance to both imino nitrogens in the thiazole rings and the hydroxyl
group from a symmetry related molecule (C23 hydroxyl), thus providing three
contacts between hectochlorin and the single water molecule in the ASU.
Anomalous scattering data allowed differentiation of the two possible enantiomeric
forms of hectochiorin to define its absolute structure. Therefore, the absolute
stereochemistry of hectochlorin is 5S, 1 iS, 13S, 145.
29
Table 11.2. Space group, unit cell, data collection, and refinement statistics for
hectochlorin (1).
Space Group
P212121
Unit Cell
a(A)
12.266
b(A)
12.684
c(A)
21.415
a = 3= y (degrees)
90
Refinement Parametersa
R
6.66%(5.59%)
wR2
15.33% (14.33%)
GooF for 407 parameters
1.043
aThe R and wR2 values in parenthesis are for reflections with F2 >4
cr,
the refinement
parameters are defined as follows:
R
l(Fobs)l_I(FcalclI/Fobsl, wR2 = {(w(Ib52 _FQIC2)2)/(w(FObS27)}
GooF (Goodness of Fit) = S =
{(w(PbS2 F,k2 7) I(n
Biological Activity. Luesch et al. found that lyngbyabellin A
(3)6
interferes
with microfilament formation in cultured cells. In our initial studies, we found that
hectochiorin behaves similarly tojasplakinolide (5, also reported asjaspaniide),9
causing hyperpolyinerization of the protein actin. Based on the strong structural
4),ô we compared the effects of
homology between lyngbyabellin A and B (3 and
hectochlorin and lyngbyabellin B with those ofjasplakinolide on cell growth.
30
First, we evaluated the growth inhibitory effects of the drugs on CA46 cells, a
human Burkitt lymphoma line, and found that 1 was as potent as jasplakinolide (5),
with an IC50 value of 20 nM, and was 5 thnes more potent than lyngbyabellin B (4,
Table H.3). Flow cytometry to measure cellular DNA content by propidium iodide
labeling after 24 h treatment at equitoxic concentrations of hectochiorin and
lyngbyabellin B (10 times the IC50 concentrations, 0.2 and 1.0 pM, respectively)
demonstrated a modest accumulation of CA46 cells in the G2IM phase of the cell
cycle (37% with hectochiorin and 28% with lyngbyabellin B vs 16% in the untreated
control). This result was consistent with the conclusion that the pharmacological
target of this group of drugs was the actin component of the cytoskeleton. Additional
observations supporting this conclusion were that hectochlorin and lyngbyabellin B
had no effect on the polymerization of purified tubulin, no effect on the microtubule
component of the cytoskeleton of cultured cells, and did not cause accumulation of
cells arrested in mitosis (cells with condensed chromosomes).
PtK2 cells were used to study the effects of hectochlorin and lyngbyabellin B
on the actin cytoskeleton, in comparison with the effects ofjasplakinolide. In order
to compare the agents at equitoxic concentrations, IC50 values for cell growth were
first determined (Table 11.3). Relative activities of the three drugs were similar to
their relative activities in the CA46 cells, but on average the
IC50
values were 10-fold
higher. For immunofluorescence studies (Figure 11.2) cells were examined at the
IC50
values and at 10-fold higher concentrations. Identical results were obtained with
fluorescein isothiocyanate (FITC)-labeled phalloidin and with a FITC-labeled anti-
31
actin antibody, and images obtained with the latter are presented in Figure 11.2
following 24 h of drug treatment.
Table 11.3. Effects of hectochiorin (1), lyngbyabellin B (4), and jasplakinolide (5) on
cell growth, actin polymerization, and displacement of fluorescein isothiocyanate
(FITC)-phalloidin from actin polymer.
Inhibition of cell
Drug
CA46
Lyngbyabellin B
Jasplakinolide
aCell
ECo (pM) ± SD
EC (pM) ± SD
actin polymerC
PtK2
ICo (i.tM)
Hectochlonn
polymtionb
Displacement of
FITC-phalloidin from
Stimulation of actm
0.02
0.3
20 ± 0.6
0.1
1.0
>50
0.03
0.3
19±0.5
> 60
6.5 ± I
owth was measured afler 24 h at 37°C with the CA46 cells and after 48 h with the PtK2
cells. Actin polymerization was measured by the centrifugation assay,1° The EC50 value
represents the concentration of drug inducing a 50% reduction in the protein content of the
supematant compared with a control without drug. CACth and FITC-phalloidm were incubated at
22°C for I h in AMB with 2 p.L of polymerization inducing buffer (NB) per 100 jtL reaction
mixture. Reaction mixtures were centrifuged and fluorescence of the supematant was measured
as described previously.'0 The EC50 values represent the drug concentration causing an increase
in supematant fluorescence equal to 50% of the maximum increase obtained with phalloidin.
In comparison with the control (Figure lI.2A), an increase in binucleated cells
was detected with all three drugs at both drug concentrations. This is a consequence
of arrest at cytokinesis, as is usually observed with actin-active agents. At the IC50
concentrations, hectochlonn (Figure ll.2B) and lyngbyabellin B (Figure II.2D)
caused an apparent thickening in microfilaments relative to the microfilaments
observed in the untreated control cells after 24 h (Figure ll.2A). This could result
from the bundling of
32
:
Figure 11.2. Effects of hectochlorin (1), lyngbyabellin B (4), and jasplakinolide (5) on the
actin cytoskeleton of PtK2 cells. After 24 h at 37°C cells were processed as described
previously and exposed to an FITC-labeled anti-J3-actin antibody (visualized as green in the
figure) and to the DNA-reactive compound DAPI (visualized as blue)'°. Cells were
examined under a 40x oil objective (N.A. 1.30), and the white bar in panel A indicates 30
j.tm. Asterisks indicate binucleated cells, presumably arrested at cytokinesis. A. No drug. B.
Hectochlorin at 0.3 p.M. C. Hectochiorin at 3.0 p.M. D. Lyngbyabellin B at 1.0 p.M. E.
Lyngbyabellin B at 10 p.M. F. Jasplakinolide at 0.3 p.M. G. Jasplakinolide at 3.0 p.M.
33
actin filaments, as fewer filaments were present in the center of cells as opposed to
the stronger labeling of numerous cortical actm filaments. A similar observation was
reported for A-l0 smooth muscle cells treated with lyngbyabellin
A.6
In contrast, at
its IC value j asplakinolide caused a much more drastic reorganization of the actin
cytoskeleton. F-actin formed clumps distributed throughout the cytoplasm. These
changes have been interpreted as representing numerous patches of short actin
filaments,'° based on the potent hypemucleation of actin assembly caused by the
drug."
At concentrations 10-fold higher than the 1050 values the effects of
hectochiorin (Figure H.2C) and lyngbyabellin B (Figure ll.2E) on the actin
cytoskeleton were more dramatic. Cells presented a hairy appearance due to cellular
protrusions rich in actin filaments. Again, with jasplakinolide there was a different
pattern of actin labeling (Figure II.2G). The cytoplasm of the cells retracted
extensively, and the labeling of F-actin was concentrated near the nucleus and in
small protrusions that gave a spiky appearance to the cells.
As noted above, experiments with purified actin demonstrated that
hectochlorin, like jasplakinolide, induced actin assembly in the absence of exogenous
K ("nonpolymerizing conditions!!, see Figure 11.3, curve 1). In a centrifugal assay
designed for ease of comparison of stimulatory drugs at multiple concentrations,12 we
found that hectochlorin and jasplakinolide had equivalent activity (Table 11.3),
yielding EC50 values of 20 and 19 p.M, respectively. Lyngbyabellin B was minimally
34
active in this assay. It should be noted that values obtained in this assay could be
viewed as equilibrium values because of the relatively long incubation and sample
processing times (total 2.5 h).
Fluorescence studies were conducted to compare the ability of hectochiorin
and jasplakinolide to stimulate actin polymerization. The studies were done by
determining the amount of 90° light scattering over the course of approximately 15
minutes. These measurements were performed despite the limitation that only one
sample could be examined at a time (Figure JJ3))3 In the absence of drug or
exogenously added K no actin polymer was formed during the time frame of the
experiment (Figure 11.3, curve 1), while added K caused the expected rapid
assembly (curve 2) of actin filaments. Higher concentrations of hectochiorin (curves
3-5) caused progressively more extensive assembly reactions. As the amount of
hectochlorin was increased, the lag time became progressively shorter, and the
apparent rates and extents of polymer formation increased. The same general
observation would apply to jasplakinolide (curves 6-8), but at 10 and 25 tM drug the
jasplakinolide-induced assembly reactions were much less robust than the
hectochlorin-induced reactions (compare curve 3 with 6 and curve 4 with 7). In
contrast, the reaction with 50 p.M jasplakinolide (curve 8) had an earlier onset and
was more rapid, if not more extensive, than the reaction with 50 pM hectochlorin
(curve 5).
We were intrigued that the 50 pM concentrations of jasplakinolide or
hectochlorin caused more intense light scattering than was observed with K-mduced
35
assembly (compare curve 2 with curves
5
and 8 in Figure
11.3).
We speculated that
with hectochiorin, in view of the thick filaments present in the PtK2 cells treated at
the IC50 value, this might be due to actin filament bundle formation or, possibly,
formation of polymers of aberrant morphology (previous experiments had confirmed
with jasplakinolide that unbundled actin filaments of normal morphology were
formed under the conditions used here).'2
5
8
Cl)
6
10
Minutes
Figure 11.3. Stimulation of actin polymerization by hectochlorin (1) orjasplakinolide
(5). Actin assembly was followed by 900 light scattering as described in the text.
The figure represents a composite of each reaction mixture followed individually.
Curve I: no addition (actin only). Curve 2: assembly induced with polymerization
inducing buffer (NB). Curve 3: 10 j.tM hectochlorin. Curve 4: 25 jiM hectochiorin.
Curve 5: 50 jiM hectochiorin. Curve 6: 10 jIM jasplakinolide. Curve 7: 25 jiM
jasplakinolide. Curve 8: 50 jiM jasplakinolide.
36
We explored these possibilities using centrifugation and electron microscopy.
We were unable to pellet any significant amount of actin polymer by low speed
centrifugation (20,000 x g for 30 mm at 22 °C). Electron microscopy of samples
containing actin and hectochiorin showed numerous unbundled actin filaments
identical to those induced by the polymerization inducing buffer (PIB) or
jasplakinolide (data not shown).
We observed one further biochemical difference between hectochlorin and
jasplakinolide. As shown previously, jasplakinolide readily displaces FITCphalloidin from actin polymer.12"4 Hectochiorin was unable to do this (Table 11.3).
Hectochiorin was also unable to inhibit FITC-phalloidin binding to polymer when
it was added prior to addition of the fluorescent drug (data not presented). These
results with hectochiorin are similar to our observations with dolastatin 11, which
also promotes actin polymerization.'2
We have shown that hectochiorin is more potent than lyngbyabellin B in its
effects on purified actin and as a cytotoxic compound, but the two agents appear to
have the same basic mode of action on the actin cytoskeleton. Hectochiorin was
quantitatively similar tojasplakinolide, particularly as a cytotoxic agent. These
compounds all promote actin polymerization, but the actin cytoskeleton
rearrangements in cells are different. The major biochemical difference between
hectochlonn and jasplakinolide is the inability of hectochlorin to displace FITC-
phalloidin from actin filaments or even prevent the binding of FITC-phalloidin to the
filaments. Although hectochiorin resembles dolastatin 11 in the inability to interfere
37
with FITC-phalloidin binding to F-actin, dolastatin 11 induces morphological effects
on the cellular actin cytoskeleton that are closer to those ofjasplakinolide than to
those of hectochlorin.12 Thus far, none of these drugs appears to have a significant
effect on actin filament morphology when observed by electron microscopy. Thus,
their different effects on the cellular actin cytoskeleton may result from altered
interactions of actin-associated proteins with actin filaments in drug-treated cells.'5
Hectochiorin was also provided to the National Cancer Institute for
cytotoxicity testing to their panel of 60 different cancer cell lines. These cell lines
are divided into nine tumor types; leukemia, non-small cell lung cancer, colon
cancer, CNS cancer, melanoma, ovarian cancer, renal cancer, prostate cancer and
breast cancer. The very flat shape of the dose-response curves of hectochlorin
against most of these cell lines was quite distinctive (Figure 11.5). This occurred
over four log doses, and is characteristic of compounds which are antiproliferative
but not cytotoxic. Indeed, actual cell killing was not observed in all but a few cell
lines, and then at the highest dose evaluated (10 M). In general, compounds
which inhibit microtubule or actin processes in cells give a similar profile in this
assay, at least under the time regime of the assay. At longer time points antitubulin
and antiactin compounds induce cells to undergo apoptosis. Several cell lines in
the colon, melanoma, ovarian, and renal cancer sub-panels were more strongly
affected by hectochlonn than the remainder of the cell lines.
-
I,..ofl
0f
.--...
- --- --
_____________________________________________
' c.,,1(1
I
0
I
I
3l04h C....o.fl(.. 1(.I,
(flY4fl1
..__.......
--
CloTh: . __
k.0
--
..,...
1.1(14.
....
.:,
LIICITIC
00-1111
_.._.
__...
, *Ii...I. (0olflIoI)
00/I ._. Itt II
KC111I1N
._...
3511(I
......_
I1( 0t4.
taO,g ____
,C,.1I. ......,
11(10171 _..
1(10
., .
101.2114
.._,.. -
(tlC.._ ,..
..o... (1(.hI
IC!
IS
TWO.
......
(Cr41 .o
CI1(st.O(..
Z0/
F _
.1
l.a. _........
fl.tG . - ..._
*11(11
..
.
nI.I -.
Corar
-
-
.040/Ill .
UI
...S. -.
I0Ct.j?..
W.C141 .
.1
a...
1(11.311
.____._.
.11(1101 _.. .
('01311 .
700(1
-
010/3(1.....
Pn131 Co.r
.:
'S.
'0
1
11
- LOW . - - -
(Vt' __. -
ri-I.
......
rAil I
S7I
. .. . -
10
....
t0J1.(
(p.0
1(lI.1()
L111
C
(((IN
(45.1.0 ___._
Figure 11.4. Dose-response curves for hectochiorin in the NCI 60-cell line assay
VCr4i.P.ktl . - - -
((.111 ..__
C01l1.
M-3.:
I:, pm . ......
110/10.1
39
Experimental
General Experimental Procedures. UV and IR spectra were recorded on a
Beckman DU 640B UV spectrophotometer and a Nicolet 510 spectrophotometer,
respectively. NMR spectra were recorded on a 600 MHz (Bruker DRX 600). All
chemical shifts are reported relative to residual CHC13 as internal standard. HRMS
were obtained on a Kratos MS5OTC. X-ray diffraction data was collected on a
Siemens P-4 X-Ray Diffractometer with HiStar area detector (CuK radiation).
Structure solution and refinement was carried out using SHELXS and SHELXL,
respectively. Optical rotations were measured with a Perkin-Elmer model 243
polarimeter. HPLC was performed using Waters 515 pumps and a Waters 996
photodiode array spectrophotometer. TLC grade (10-40 m) Si gel was used for
vacuum chromatography, and Merck aluminum-backed TLC sheets (Si gel 60 F254)
were used for TLC.
Collection and Culture Conditions. The marine cyanobacterium L.
majuscula was collected by hand from shallow water (2 m) on 22 August 1996, at
Hector Bay, Jamaica. The bulk of the cyanobacterial material was preserved in
1PA and a small portion was stored in sea water for culturing. A voucher sample is
available from WHG as collection number JHB-22 Aug 96-01C2. Following
transport to Oregon, the sample reserved for culture was isolated free of
contaminating cyanobactena or other microalgae using standard techniques.7 The
cultures were maintained in a 28 °C controlled temperature room with 16/8
40
light/dark cycle provided by Sylvania 40W cool white fluorescent lights (4.67 j.tmol
photons s1 m2).
The liquid culture medium used for the isolation procedure
consisted of SWBG1 I and ESW. The cyanobacteria were harvested at 6-7 weeks
after initial inoculation. When sufficient cell mass was grown the cells were
harvested according to the technique of Rossi et.
al.8
Extraction and Isolation. Approximately 114 g (wet weight) of the algal
material harvested from culture was repetitively extracted with CH2C12IMeOH
(2:1) to yield 1.1 g of crude extract. The crude extract was then fractionated using
vacuum liquid chromatography (VLC, 9.5 cm x 4 cm) on TLC grade silica gel.
Fractions eluting between 50 % EtOAc in hexanes and 80 % EtOAc in hexanes
were recombined and further purified. The recombined fractions were first
chromatographed over C18 SPE cartridge (gradient elution from 50 % MeOH in
H20 to 100 % MeOH), in which fractions eluting in 70-80% MeOH in H20 were
then subjected to RPHPLC. An isocratic elution profile in 82% MeOH in 1120
(Phenomenex SPHERECLONE ODS, 250 x 10 mm, 5j.t) yielded pure hectochiorin
(1, 35.2 mg, 3.2% of crude extract).
Hectochiorin (1): glassy, pale yellow solid; [a
-8.7 (c 1.04, MeOH); JR
(neat) 3459, 3119, 2983, 2939, 2882, 1756, 1746, 1729, 1713, 1572, 1484, 1244,
1091 cm; 'H and 13C NMR data, see Table 11.1; HRFABMS (in 3-NBA) [M + H]
m/z 665.1171 (calculated for C27H35C12N209S2, 665.1161).
41
X-Ray Crystallography. Five mg of hectochiorin were dissolved in 1 mL
of MeOH, followed by the careful addition of 1 mL of H20, creating two distinct
solvent phases. The vial was then sealed and monitored over the course of several
days for crystal growth. By day three, Hectochlorin had formed large cubic
crystals visible to the naked eye. A 0.3 mm x 0.3 mm x 0.3 mm crystal was
mounted and sealed in a capillary tube. Graphite monochromated CuKa radiation
from a sealed tube (Siemens P4) was used to record 5813 reflections. XSCANS
(Siemens) employed 97 >25c reflections to index the unit cell as: P212121, a =
12.266 A, b = 12.684 A, c = 21.415 A,
a = = = 90°. The structure was solved
and refined with SHELXS and SHELXL respectively (Sheldrick, SHELX-97). A
single molecule of hectochiorin and one solvent molecule (H20) constituted the
ASU (asymmetric unit). Using least squares full matrix, 407 parameters were
refined; the structure refined to an R factor of 6.66% for all reflections, 5.59% for
reflections > 4q; cR2 of 15.33% for all reflections, 14.33% for reflections > 4.
Antimicrobial Assay. The antimicrobial activity of hectochiorin was
evaluated using standard paper sensitivity disk-agar plate methodology (disk
diameter, 6 mm). Hectochiorin gave a 16 mm zone of inhibition at 100 pg/disk and
an 11 mm zone of inhibition at 10 p.g/disk to Candida albicans (ATCC 14053);
however, it was inactive to Pseudomonas aeruginosa (ATCC 10145), Escherichia
co/i (ATCC 11775), Salmonella choleraesuis subsp. choleraesuis (ATCC 14028),
Bacillus subtilis (ATCC 6051), and Staphylococcus aureus (ATCC 12600).
42
Actin Studies. Purified rabbit muscle actin was obtained from Cytoskeleton
(Denver, CO), phalloidin and Antifade Mounting Solution from Molecular Probes
(Eugene, OR), PtK2 cells (normal kidney cells of the kangaroo rat Potorous
tridactylis) and CA46 cells (human Burkitt lymphoma cells) from the American
Type Culture Collection (Manassas, VA), 4',6-diamidino-2-phenylindole (DAPI),
FITC-conjugated phalloidin, and FITC-conjugated f3-anti-actin monoclonal antibody
(clone Ac- 15) from Sigma (St. Louis, MO), and the Chambered Covergiass System
from Nalge Nunc International (Naperville, IL). Jasplakinolide was generously
provided by the Drug Sthesis & Chemistry Branch, National Cancer Institute
(Rockville, MD).
Methodologies for maintenance of PtK2 and CA46 cells in culture,
measurement of drug effects on cell growth, direct immunofluorescence (actin and
DNA), flow cytometry, electron microscopy, measurement of displacement of FITC-
phalloidin from F-actin, and measurement of actin polymerization by centrifugation
were described previously.3d216
Actin assembly was also measured by 90° light scattering (400 nm) in a
fluorometer (Photon Technology International, Lawrenceville NJ) at 22 °C. Each
100 p.L (final volume) reaction mixture contained 25 1iM actin, 5% (v/v) DMSO, and
drug or 2 iL of PIB, as indicated, in actin monomer buffer (AMB). Actin in AIv1B
was added to the cuvette, and light scattering was measured for 3 mm to establish a
background. At this point DMSO, drug in DMSO, or Pffi + DMSO was added to the
cuvette. The cuvette contents were rapidly mixed, and light scattering was measured.
43
References
Gerwick, W. H.; Tan, L. T.; Sitachitta, N. Nitrogen-Containing Metabolites
from Marine Cyanobacteri. In Alkaloids, in press.
2. Wu, M.; Okino, T.; Nogle, L. M.; Marquez, B. L.; Williamson, R. T.; Sitachitta,
N.; Berman, F. W.; Murray, T. F.; McGough, K.; Jacobs, R.; Colsen, K.;
Asano, T.; Yokokawa, F.; Shioiri, T.; Gerwick, W. H. I Am. Chem. Soc. 2000,
122, 12041-12042.
3. (a) Gerwick, W. H.; Proteau, P. J.; Nagle, D. G.; Hamel, E.; Blokhin, A.; Slate,
D. J. Org. Chem. 1994, 59, 1243-1245. (b) Yoo, H-D.; Gerwick, W. H. I Nat.
Prod. 1995, 58, 1961-1965. (c) Marquez, B.; Verdier-Pinard, P.; Hamel, E.;
Gerwick, W. H. Phytochemistry 1998,49,2387-2389. (d) Verdier-Pinard, P.;
Lai, J-Y.; Yoo, H-D.; Yu, J.; Marquez, B.; Nagle, D. A.; Nambu, M.; White, J.
D.; Faick, J. R.; Gerwick, W. H.; Day. B. W.; Hamel, E. Mo!. Pharmacol. 1998,
53, 62-76.
4. Hooper, G. J.; Orjala, J.; Schatzman, R. C.; Gerwick, W. H. J. Nat. Prod. 1998,
61, 529-533.
5. Sone, H.; Kondo, T.; Kiryu, M.; Ishiwata, H.; Ojika, M.; Yamada, K. J. Org.
Chem. 1995, 60, 4774-4781.
6. (a) Luesch, H.; Yoshida, W. Y.; Moore, R. B.; Paul, V. J.; Mooberry, S. L. J.
Nat. Prod. 2000, 63, 611-615. (b) Luesch, H.; Yoshida, W. Y.; Moore, R. E.;
Paul, V. J. I Nat. Prod. 2000, 63, 1437-1439. (c) Milligan, K. E., Marquez, B.
L.; Williamson, R. T.; Gerwick, W. H. J. Nat. Prod. 2000, 63, 1440-1443.
7. Gerwick, W. H.; Roberts, M. A.; Proteau, P. J.; Chen, J-L. I App. Phycol.
1994. 6, 143-149.
8.
9.
Rossi, J. V.; Roberts, M. A.; Yoo, H-D.; Gerwick, W. H. J. App. Phycol. 1997,
9, 195-204.
(a) Crews, P.; Manes, L. V.; Boehler, M. Tetrahedron Lett. 1986, 27, 27972800. (b) Ayscough, K. R.; Stryker, J.; Pokala, N.; Sanders, M.; Crews, P.;
Drubin, D. G. J. Cell Biol. 1997, 137, 399-416. (c) Zabnskie, T. M.; Kiocke, J.
A.; Ireland, C. M.; Marcus, A. H.; Molinski, T. F.; Faulkner, D. J.; Xu, C.;
Clardy, J. I Am. Chem. Soc. 1986, 108, 3123-3124.
10. Spector, I.; Braet, F.; Shochet, N. R.; Bubb, M. R. Microsc. Res. Tech. 1999, 47,
18-37.
44
11. Bubb, M.R.; Spector, I.; Beyer, B. B.; Fosen K. M. I Biol. Chem. 2000, 275,
5163-5 170.
12. Bai, R.; Verdier-Pinard, P.; Gangwar, S.; Stessman, C. C.; McClure, K. J.;
Sausville, E. A.; Pettit, G. R.; Bates, R. B.; Hamel, E. Mo!. Pharmacol. 2001,59,
462-469.
13. Wegner, A.; Engel, J. Biophys. Chem. 1975, 3, 215-225.
14. Bubb, M. R.; Senderowicz, A. M. J.; Sausville, E. A.; Duncan, K. L. K.; Korn,
E.D. I Biol. Chem. 1994,269, 14869-14871.
15. Carlier, M.-F. Curr. Opin. Cell Biol. 1998, 10, 45-51.
16. Muzaffar, A.; Brossi, A.; Lin, C. M.; Hamel, E. I Med. Chem. 1990, 33, 567571.
17. Farrugia, L. J. I App!. Cryst. 1997,30, 565.
45
CHAPTER III
ISOLATION, STRUCTURE ELUCIDATION, AND BIOSYNTHESIS OF
THE JAMAICAMIDES
Abstract
The presence of covalent halogen atoms has become a prevalent theme in
the secondary metabolites isolated from cyanobacteria over the last decade. The
preponderance of halogenated natural products, to date, contain chlorine atoms,
with very few containing bromine and, to a lesser extent, both halogen atoms
within the same molecule. Herein, the isolation, structure elucidation, and
biosynthetic precursors used to assemble a novel class of halogenated compounds,
jamaicamides A, B, and C, are presented. The absolute stereochemistry of one of
the two stereocenters has been defined. In addition, the biosynthetic precursor
composition, consisting of both polyketide and amino acid origins, includes acetate,
alanine, -alanine, and methylation by S-adenosylmethionine.
46
Introduction
Marine cyanobacteria (blue-green algae) have emerged over the past decade
as prolific producers of novel secondary metabolites as demonstrated by the
preponderance of reports in the recent literature.' A structural feature that has been
seen in greater numbers within these structures isolated is halogenation. These
metabolites can have single or multiple instances of halogen atoms incorporated
into their molecular structures. Examples of such halogenated metabolites include
the malyngamides,2 hectochiorin and the lyngbyabellins,3'4 and the structurally
intriguing metabolite barbamide, containing a rare trichioromethyl moiety.5 An
investigation into the biosynthesis of barbamide is the subject of Chapter IV in this
thesis.
Among the halogenated secondary metabolites defmed from cyanobactena,
the majority contain chlorine atoms, while relatively few incorporate bromine. In
fact, to date, there have been 38 literature reports from cyanobacteria accounting
for 109 different compounds containing halogen atoms, of which only 29 include
bromine
atoms.6
The isolation and structure elucidation of the jamaicamides has
created a new and unique class of halogenated compounds. In particular,
jamaicamide A possesses a bromine atom bound to an acetylemc carbon, the first
example of such a functionality from cyanobacteria and only previously observed
in the marine environment from the sponge Xestospongia muta.7
47
OCH3
2
HN
HN'
NS
\=1
°
4
3
Figure ffl.l. Structures of malyngamide Q (1), hectochiorin (2), lyngbyabellm A
(3), and barbamide (4).
The planar structures of the jamaicamides were determined by standard 1and 2-dimensional NMR methods as well as the use of a new NMR experiment
given the acronym ACCORD-ADEQUATE.8 This latter experiment is elaborated
in detail in chapter VII. The isolation, structure elucidation, and biosynthesis of the
jamaicamides, isolated from a Jamaican collection
discussed herein.
of Lyngbya majuscula is
Results and Discussion
As previously described in chapter II, the algal sample that produces the
jamaicamides, in addition to hectochiorin, was obtained from a Jamaican collection
of Lyngbya majuscula that has been adapted to laboratory culture conditions.9 The
culture samples were isolated free of contaminating cyanobacteria or other
microalgae using standard
techniques.9
The liquid culture medium used for the
isolation procedure consisted of SWBG1 1 and ESW. The cultures were maintained
in a 28° C controlled temperature room with 16/8 light/dark cycle provided by
Sylvania 40W cool white fluorescent lights (4.67 unol photons s1 m2). When
sufficient cell mass was grown (typically 6-7 weeks) the cells were harvested
according to the technique of Rossi et.
al.9
The isolation proceeded via repetitive
extraction of the algal material with a 2:1 mixture of CH2Cl2 and MeOH. The
crude extract was then vacuum chromatographed over a bed of silica gel with a
gradient of EtOAc and hexanes, beginning with 100% hexanes and progressively
increasing percentages of EtOAc. Collected fractions were dried in vacuo and
examined by TLC and l-D 1H NMR spectroscopy. Spectroscopically interesting
fractions were further purified utilizing
C18
SPE cartridges. Pure compounds were
then isolated by RPHPLC from the fractions containing structurally intriguing
elements as discerned by NMR spectroscopy.
HRFABMS for jamaicamide A established a [M + H] molecular formula
of C27H37N2O4C1Br (m/z 567.1625, -0.7 mmu dev.). The isotope peaks at m/z
567/569/571, in an approximately 4:5:1.5 ratio, were consistent with the presence
of a single chlorine atom and bromine atom in the molecule. Using this molecular
formula, 10 degrees of unsaturation were present in this molecule.
Analysis of the HSQC1° and HSQC-COSY" spectra facilitated the
construction of seven (A-G, Figure 111.2) partial structures. Further elaboration of
these structures was accomplished with additional 2D NMR experiments. A single
JCH
correlation was seen between the methine of structure F (6 5.8, H27 in 5) and a
quatemary carbon at 6 141.7 (C6 in 5). Two 3JCH cross peaks between H27 and a
methylene resonating at 6 29.2 (C5 in 5) and a methylene at 6 32.5 (C7 in 5) were
used to connect partial structures E through G (Figure 111.3). The chemical shift of
C27 (6 112.7) indicated the presence of a vinyl chloride as observed in the majority
241
25
'H
21
A
B
C
D
H 27
3
G
Figure 111.2. Partial structures A-G derived from HSQC and HSQC-COSY.
of the malyngamides (1). Connectivity between partial structures E and D was
deduced by an HMBC'2 cross peak between 6 2.13 (H 13 in 5) and a carbonyl
carbon located at 6 172.4 (C14 in 5) and a correlation between the exchangeable
proton of the secondary amide nitrogen (6 6.61) to 6
Partial structures B and D were joined via a
protons
(6 3.7, H325)
and 62.80
(H216).
and 6
175.3 (C17),
3JCH
172.4
and 6 36.6
(C13
in 5).
correlation between the methyl
which in turn was scalar coupled to
In addition, there was a cross peak between 6 6.68
6 2.95
(H18) and
Cl 7, thereby showing the connectivity between structures B and C (Figure
111.3).
At this point in the structure detennination, only two additional connections could
be gleaned from the HMBC data. The first were the 2'3JCH correlations from 66.03
in 5) and 6
(H21
7.10 (1122 in
5) to 6
169.9
resulting in a conjugated enone system
in partial structure H (see Figure ffl.3). The second was a response between 62.16
(H23
in 5) and a quatemary carbon resonating at 6
79.8 (C2
in 5). Inclusion of all
HMBC data reduced the initial seven partial structures to two (H and I, Figure
111.3).
HH = 6.2
3JHR=15.lHz
Hz
H3C.0'
I
6
_
18Hj
I
Figure 111.3. Partial structures H and I. Single-headed arrows represent 3JHH
couplings and double-headed arrows represent important 2'3JCH couplings. The
boxed letters indicate the original partial structure designations (see Figure 111.2).
51
The remaining atoms needed to complete the structure ofjamaicamide A
were C2NOBr. Additionally, four degrees of unsaturation remained to be satisfied.
Acquisition of an ACCORD 1,1 -ADEQUATE showed a 1Jcc coupling between
C18 and C19 ( 165.9). The ACCORD 1,1-ADEQUATE also confirmed all partial
structures (H and I, see Figure 111.4). A 1H-'5N HMBC'3 provided crucial
correlations that satisfied an additional degree of unsaturation by placing the
remaining nitrogen in a pyrrolidone ring.
2'3JNH
correlations between H18, H21,
H22, and H24 to the pyrrolidone nitrogen were observed.
i:
H
CLH
CiI
l:EI
Figure 111.4. Partial Structure ofjamaicamide A including key ACCORD 1,1ADEQUATE and 'H-MN HMBC correlations. The boxed letters indicate the
original partial structure designations (see Figure ffi.2).
Two atoms remained to be assigned to complete the planar structure of
jamaicamide A, specifically a carbon and a bromine atom. Intuitively, with the
chemical shift of C2 (6 79.8) and a remaining quatemary carbon and a bromine, the
placement of a brominated acetylene seemed reasonable. However, there was no
52
indication in the 1D 13C NMR spectrum of an additional carbon atom. In an effort
to gain additional information about the spectral properties of this type of
functionalized acetylene, a 13C NMR spectrum of a model compound, 1 1-bromo-
undec-lO-ynoic acid amide, was acquired. 14 As shown in Figure 111.5, Cli and
dO of this model compound resonate at 37.4 and 80.3, respectively. Upon reinspection of the '3C NMR spectrum for jamaicamide A, a small shoulder on the
C15 ( 38.2) resonance could be observed, which also showed a 3JCH correlation in
the HMBC spectrum to H3 ( 2.16), hence defining the bromoacetylene moiety of
jamaicamide A. With this moiety in place, the planar structure ofjamaicamide A
was complete.
do
80
CII
70
60
ppm
50
I 1-bromo-undec-lO-ynoic acid arnide
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
ppm
Figure 111.5. Structure and '3C NMR spectra of 1 1-bromo-undec-lO-ynoic acid
amide. The boxed region indicates the expansion shown to the upper right of the
full spectrum. The chemical shifts are indicated and show a close match to those
observed for jamaicamide A (see Table 111.1).
53
The absolute stereochemistry of C23 injamaicamide A was determined
through the use of Marfey's analysis.'5 The presence of S-alanine was verified
through ozonolysis and acid hydrolysis ofjamaicamide A, followed by
derivatization with FDAA and HPLC comparison with derivatized S- and R-alanine
standards. The determination of the stereocenter at C9 is currently underway in the
Gerwick laboratory. The geometry of the vinyl chloride moiety was determined by
measuring the
3JCH
coupling constants utilizing the HSQMBC experiment.16
Coupling constant values of
JCS-H27 = 6.7
Hz and 3Jc77 = 4.9 Hz revealed an E
geometry. To investigate the geometry of the C17-C18 double bond, a DPFGSE
1D NOE experiment was utilized.'7 Selective irradiation of the protons at 3.70
ppm (H325) showed a strong NOB enhancement ofHl8
(6.68
ppm), indicating that
this double bond is of E geometry.
27
CI
24
0
20 0
26
Jamaicamide A (5)
R Br
Jamaicamide B (6)
R=H
HN)J1
Jamaicamide C (7)
Figure 111.6. Structures ofjamaicamides A (5), B (6), and C (7).
R
54
Jamaicamide B was also isolated as a pale yellow oil from the lipid extract
of the cultured L. majuscula. Jamaicamide B was of a slightly more polar nature,
eluting before jarnaicamide A under RPHPLC conditions. HRFABMS revealed a
molecular formula of C27H37O4N2C1, again indicating 10 degrees of unsaturation.
The observed ratio of the EM + H] isotope peak cluster in the FABMS was
consistent with the presence of a single chlorine atom (m/z 489/49 1 in an
approximate 1:0.4 ratio). The 1H and '3C resonances for jamaicamide B were
closely related to those ofjamaicamide A. However, the absence of a signal at 37.9
ppm in the 13C spectrum (Cl in 5) and the appearance of a 68.6 ppm '3C signal (Cl
in 6), directly bound to a proton at 1.97 ppm, in addition to the downfield shift of
C2 (79.8 ppm in 5 and 84.1 in 6), indicated that jamaicamide B (6) was the
debromo analog ofjamaicamide A (5).
55
H23-C24
6.
H3-C4,
H9;C26943
H5-C4
H4-05
ppm
g
H11-C12
H15-C16
H1O-C9
I426-C
. 'H7-C8,°
H16-C15
f
30
c.H8-C7
Hl3-CI?b
40
H9-C8 H8-C9
H12-ç13
I
H24-C23
H23C23
H22-C23
60
F
H3-C2
-
70
80
90
100
110
H27-C27
120
H23-C21
H22-C21
H1O-C11
.H12.C11
oH9-C1O
Hil-ClO
H27-C6
H7-C6
H5-C6
140
150
H23-C22
H22-C22
130
H21-C22
160
H18-C19
a H21-C20
H18-C17
7.5 7.0 6.5 6.0 5.5 5.0
H1617
H13-C1 2
4.5 4.0 3.5 3.0 2.5 2.0 1.5
170
1.0 0.5 ppm
Figure 111.7. Two-dimensional plot of the ACCORD-ADEQUATE ofjamaicamide
A. The experiment was performed on a 58 mM sample ofjamaicamide A in CDC13
at 298 K. The data were acquired with 256 scans per increment for a total of 180
F1 increments. The spectral width is 22.6 kHz in F1 and 5 kHz in F2. The data was
processed with an exponential weighting function in F2 (10 Hz) and a t/2 shifted
sine bell in F1 (20 Hz). The F1 dimension was linear predicted to 360 data points
and zero filled to 1K data points. All responses for jamaicamide A (5) are labeled
according to the numbering scheme shown in Figure 111.6.
56
Jamaicamide C was isolated in small quantity (0.5% of crude extract) as a
pale yellow oil from the same extract as the previous two jamaicarnides A and B.
Chromatographically, jamaicamide C was more hydrophobic than either
jamaicamides A and B. The HRFABMS yielded a molecular formula of
C27H3904N2C1, which calculated for 9 degrees of unsaturation. As observed for
jamaicamide B, the isotope pattern of the molecular ion cluster clearly indicated the
presence of a single chlorine atom (m/z 491/493 in an approximate 1:0.4 ratio). 1H
and '3C values (Table ffl.3) clearly demonstrated that jamaicamide C was the
alkene equivalent ofjamaicamide B (Cl and H21 at
respectively and C2 and H2 at
138.4 and
114.6 and ö 5.01, 4.96,
5.82, respectively).
OH
O
o
°
8
9
10
Figure 111.8. Structures of microcolin A (8), ypaoamide (9), and dolastatin 15 (10).
57
Isolation of these metabolites from a cultured L. majuscula provided an
unique opportunity to explore the biosynthetic origins of several structural elements
present in the jamaicamides, such as the methyl pyrrolidone ring and the vinyl
chloride and acetylene moieties. The methyl pyrrolidone ring has also been
observed in the immunosuppressive natural products microcolins A (8), B and C,
and is similar to that observed for ypaoamide (9) and dolastatm 15
(1O).1820
The
vinyl chloride moiety is a predominant theme in the malyngamide class of
compounds. The presence of an acetylene unit in marine natural products is of
infrequent occurrence,7 and the biosynthetic precursor for this unit has never been
established in a marine natural product.
Initial thoughts on the biogenesis of this compound consisted often acetate
units and two amino acid units, specifically alanine and 3-alanine. This initial
hypothesis also included two methyl groups thought to come from methionine.
Therefore, we designed feeding experiments employing isotopically labeled
precursors to investigate these biosynthetic hypotheses.
and '3C NMR spectral data (in ppm) for jamaicamide A (5) with
HMBC and ACCORD 1,1-ADEQUATE correlations.
Table 111.1.
Position
Number
SH(JmHz)
I
-
m
-
2
-
-
3
2.16 (7.2)
4
1.57
5
2.20
dd
m
m
-
-
6
7
8
m
m
obs
5.21(15.1,7.8) dd
5.29(15.1,6.4) dt
2.23
m
2.13 (7.9)
t
1.95
1.30
1.97
HMBL
Accordion 1,1Adequate
39.7
79.8 (s)
19.5 (t)
H3, H4
H4, H5
H3
H4
25.8(t)
H3,H5
H5,H3
29.2 (t)
H3, H4
H4, H5, H7, H8, H27
H5, H8, H9, H10
H7, H9, H10
H4
H5, H7, H27
H8
H7, H9
H7,H8,H10,H11
H8,H9,Hi1,H12
H9,H10,H12,H13
H10,H11,H13
Hil, H12
H12,H13,NH,H15
H8,H10,H26
H9,H11
H12,H10
H11,H13
H16a,b
HiS, H18
H15,H16a,H16b,H18,
H16a,H16b
HiS
H16a,H16b,
1125
H18
S
(mult.)'bc
141.7(d)
32.5 (t)
34.7 (t)
-
-
3.40
2.95, 2.80
m
36.2(d)
136.5(d)
127.5(d)
28.5(t)
36.6(t)
172.4(s)
38.1(t)
m
32.3 (t)
17
-
-
1753 (s,
18
6.68
s
19
-
-
20
-
-
94.9 (d)
165.9(s)
169.9(s)
H25, H16a,b
H22
H21,1122,H23
6.03 (6.2, 1.4)
7.10(6.2, 1.9)
dd
dd
125.7(d)
H22,H23
153.1 (d)
H21, 1123,1124
H22
H21, H23
ddq
58.0(d)
H21,H22,H24
H22,H24
d
s
17.8 (q)
H22, 1123
H23
56.0(q)
1118
-
d
20.7 (q)
H8, 119, H10
H9
s
112.7(d)
H5,H7
-
m
-
-
-
9
10
11
12
13
14
15
16
21
22
23
24
25
26
27
NH
a
a
4.8(6.8,1.8,
1.41 (6.7)
3.71
0.89 (6.8)
5.80
6.66
.
H12
H13
-
H18
1121
Recorded at 500.17 MHz. bRecorded at 125 MHz. cMultiplicity deduced by multiplicity
edited HSQC. dprotons showing long-range correlation with indicated carbon.e Protons
showing 2JCH correlations via 'fcc coupling.
59
Table 111.2. 1H and '3C NMR spectral data (in ppm) for jamaicamide B (6) with
HMBC correlations.
Position
Number
'H (fin Hz)a
m
ö
'3C (mu1t.)Ic
HMBCd
1.97
s
68.6(s)
H3
2
-
-
84.1 (s)
3
2.19
H3,H4
H1,H4,H5
1
4
1.64
5
2.26
m
m
m
6
-
-
7
1.99
8
1.33
9
2.01
m
m
m
10
5.26(15.1,7.8)
5.35(15.1,6.5)
dd
12
2.28
dt
m
13
2.17(7.8)
t
14
-
-
16
3.5
2.98, 2.83
m
m
17
-
-
18
6.72
s
11
15
19
-
20
-
-
21
6.07 (6.2, 1.0)
22
23
7.21(6.2,1.9)
dd
dd
m
24
25
26
27
NH
18.3(t)
26.1 (t)
29.3 (t)
141.9 (d)
32.6 (t)
34.7 (t)
H3, H5
H3, H4
H4, H5, Hi, 118, H27
36.3(d)
136.6(d)
127.5(d)
28.5(t)
36.7(t)
172.4(s)
H7,H8,Hi0,H11
38.2 (t)
32.2 (t)
175.4(s)
94.9(d)
166.0(s)
170.1 (s)
H5, H8, H9, Hi0
H7, H9, H10
H8,H9,Hii,H12
H9,Hi0,H12,H13
H10,Hii,H13
Hii,H12
H12,H13,NH,H15
H16a,b
HiS, 1118
H15,H16a,Hi6b,Hi8,1125
H25,H16a,b
H22
1121,1122,1123
H22, H23
H21,H23,H24
H21,H22,H24
d
125.9(d)
153.1(d)
58.1(d)
17.9(q)
s
56.1 (ci)
1118
d
20.8 (q)
H8, H9, H10
5.81
s
112.7(d)
H5,H7
6.68
m
4.87
1.45 (6.6)
3.74
0.93 (6.7)
H22, H23
-
aRorded at 600.04 MHz. bRecorded at 150.14 MHz. cMuitiplicity deduced by multiplicity
edited HSQC. "Protons showing long-range correlation with indicated carbon.
Table 111.3. 'H and '3C NMR spectral data (in ppm) for jainaicamide C (7).
P
t011
8 'H (fin HZ)a
in
la
5.01 (17.0, 1.8)
dd
b
4.96 (10.2, 1.8)
dd
6 '3C (mu1t)
114.6(t)
3
2.11
m
m
4
1.50
obs
26.3 (t)
5
2.18
m
29.6(t)
-
142.5 (d)
2
5.82
6
138.4(d)
33.5(t)
9
2.01
m
m
m
10
5.26(15.1,7.5)
dd
136.6(d)
11
5.35 (15.3, 6.0)
obs
127.5(d)
12
2.28
m
28.5 (t)
13
2.17
m
36.7(t)
-
172.4(s)
7
1.99
8
1.33
32.6 (t)
34.7 (t)
36.3 (d)
14
-
15
3.51
16
2.98, 2.83
m
m
17
-
-
175.4(s)
18
6.72
s
94.9 (d)
38.2(t)
32.2 (t)
19
-
-
166.0(s)
20
-
-
170.1(s)
21
6.07 (6.2, 1.6)
dd
125.9(d)
22
7.21 (6.2, 2.0)
dd
153.1 (d)
23
4.87
m
58.1 (d)
24
1.45 (6.8)
d
17.9 (q)
25
3.74
s
56.1 (q)
26
0.93 (6.8)
d
20.8 (q)
27
5.76
s
112.0(d)
NH
6.73
obs
-
aRecorded at 400.13 MBz. bRecorded at 100.62 MHz. cMultiplicity deduced by multiplicity
edited HSQC.
61
Acetate Feeding Experiments
To explore the biosynthetic origin of the carbon atoms within the
jamaicamides, various isotopically labeled acetates were fed to the producing
organism. The initial two feeding experiments consisted of [1)3C]acetate and [2'3C]acetate. The 1-D 13C NMIR spectrum ofjamaicamide A isolated from the [1-
'3C]acetate experiment showed that carbons 2, 4, 6, 8, 10, 12, 14, 19, and 20 all
arise from Cl of acetate (Figure 111.10). Analysis of the 1-D '3C NIvIR
jamaicamide A isolated from the L. majuscula supplemented with [2-13C]acetate
showed that carbons 1,3,5,7,9, 11, 13, 18,21 and 27 are derived from C2 of
acetate (Figure ffl.1 1).
26
10
22
1121
18
27
1714
6
160
150
140
130
24
111141
I
''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
170
I
'i
21
H
I
23 25
I
120
110
100
90
80
70
T'''''l'''''''!''''' II''''
60
50
40
30
ppm
Figure 111.9. 13C NMR spectrum ofjamaicamide A at natural abundance.
75
13
10
22
170
160
150
1226
\\II//_i
130
18
27
ii2l
140
It6,
120
110
100
2325
90
80
70
60
I/(
50
40
3
4124
30
ppm
Figure 111.10. '3C NMR spectrum ofjamaicamide A isolated from cultures
provided with [1-'3C]acetate. Bolded/underlined numbers indicate significantly
enriched carbons (see Table ffl.4).
\5
716[ 26
II
111
22
1714
JI
........
6j
I
170
10
160
150
140
23 25
15,1
2
130
120
110
100
90
80
70
60
50
40
30
ppm
Figure 111.11. '3C NMR spectrum ofjamaicamide A isolated from cultures
provided with [2-13CJacetate. Boldedlunderlined numbers indicate significantly
enriched carbons (see Table ffl.4).
63
The next step was to feed doubly labeled [1 ,2-'3C2] acetate to examine the
intact nature of the acetate units in the jamaicamide structures. The observed
coupling patterns in the 1D '3C NMR indicated that acetate was incorporated as
intact units as shown in Figure ffl.16. The results from the above three feeding
experiments show that, starting from Cl, seven intact acetate units are assembled
linearly to form the lipid portion ofjamaicamide A. In addition, C27 was
incorporated from C2 of acetate. Carbons 18 and 19, and 20 and 21 of the
pyrrolidone ring, were also incorporated as intact acetate units.
0
0
HO(&OH
0
a-keglutarate
OOH
40H
alanine
°
°
Ho1AOH
pyruvate
JO2
OH
acetate
NB2
glutamate
Figure 111.12. Catabolic fate of alanine via transamination and decarboxylation to
acetate. The black diamond indicates the location of the '3C labeled carbon in S-[313C]alanine when converted to [2-'3C]acetate.
S-13-'3Clalanine feeding experiment
To examine the origin of C24 of the pyrrolidone ring, S-[3-'3C]alanine was
provided to the jamaicamide producing L. majuscula. Analysis of the '3C NMR
spectrum ofjamaicamide A produced under these conditions revealed a 2.8 fold
64
enhancement of C24, supporting our hypothesis that C22-24 and the tertiary
nitrogen of the jamaicamides arise from S-alanine.
526
139
10
22
I
Ii
23 25
17
15,1
2
6
170
160
150
140
ii
I
130
120
110
100
90
80
70
60
50
40
I
30
ppm
Figure 111.13. '3C NMR spectrum ofjamaicamide A isolated from cultures
provided with S-[3)3Cjalanine. Bolded/underlined numbers indicate significantly
enriched carbons. Italicized/underlined numbers indicate carbon atoms labeled
through catabolism of alanine (see Figure ffl.12 and Table 111.4).
The '3C NMR spectrum ofjamaicamide A isolated from this experiment
showed additional carbon resonances that were enriched in '3C from the S-[3'3C]alanine label (Table 111.4). A pyridoxal phosphate-dependent transamination of
alanine is known to result in the formation of pyruvate, which subsequently
undergoes decarboxylation to produce acetate.21 As shown in Figure 111.12, the
labeled carbon at position 3 of alamne will eventually become C2 of acetate,
thereby providing a source of 3C labeled acetate for incorporation into all acetate
derived subunits ofjamaicamide A.
65
47.9 Hz
...
47.9 Hz
34.5 Hz
10.5 Hz
C17
177
176
175
34.5 Hz
C16
C15
174 ppm
39
38
rEj3F1i
37 ppm
34
33
32
ppm
Figure 111.14. '3C spectrum of isolated jamaicamide A from L. majuscula
supplemented with ['3C3,'5N3-a1anine. Coupling constants are shown to indicate
intact incorporation of the labeled precursor.
['3C3,'5N1 -a1anine feeding experiment
To directly evaluate whether 13-alanine serves as a precursor to C 15-17 and
the secondary amide N in the jamaicamides ['3C3,15N]13-alanine was fed to cultures
of L. majuscula. Analysis of the '3C spectra for isolated jamaicamide A showed
intact corporation of all three isotopically labeled carbons into C15-17. Coupling
constants of 1Jc15..c16 = 34.5 Hz,
Jc16-c17
= 47.9 Hz,
1JC15..N
= 10.5 Hz were
observed. These results clearly indicate that 13-alanine serves as a precursor for this
region ofjamaicamide A.
26
25
139
10
22
14
170
1121
23
18
27
6
160
150
140
\\)6/','2
4
2
130
120
110
100
90
80
24
70
60
50
40
30
ppm
Figure ffl.15. 13C Ntv[R sectrum ofjamaicamide A isolated from cultures
provided with S-[methyl) C]methionine. Bolded/underlined numbers indicate
significantly enriched carbons (see Table 111.4).
S-[methyl-'3Cjmethionine feeding experiment
To establish which carbon atoms in j amaicamide A derived from the S-
methyl of methionine we fed S-[methyl-'3C]methionine. We hypothesized that the
O-CH3 group and C26 in 5, both derived from this source. Analysis of the '3C
NMR spectrum of the isolated jamaicamide A revealed that both the O-CH3 and
C26 carbons arise from the Cl pooi via S-[methyl-13C]methionine, displaying
incorporation enhancements of 2.3 and 2.4 fold, respectively (see Table ffl.4 and
experimental).
Conclusion
The feeding studies described above provide insight into the biosynthetic
precursors responsible for the synthesis of the jamaicamide class of natural
products. The results for the biosynthetic feeding experiments are illustrated in
Figure ffl.16 and tabulated in Table 111.4.
67
H
19
22(4
0
['3C3,'5N]3-alanine
21
S-[3-'3C]alanine
[1-'C]acetate
U [2-3C]acetate
A S-[methyl-13C}methionine
[1,2)3C]acetate
Figure ifi. 16. Summary of biosynthetic precursors ofjamaicamide A (5). These
results are sununarized in Table 111.4 (with the exception of the 1 ,2-'3C2]acetate and
the [13C3,'5N]-alanine feeding results which are described in the experimental
section). Carbons C22, C23 and the tertiary nitrogen are hypothesized to arise from
alanine as suggested by incorporation of S-[3)3C]alanine.
As observed in the '3C NMR spectra ofjamaicamide A isolated from [1'3CJ-acetate and [2-'3C]-acetate and [1,2-' 3C2]acetate feedings, the lipid portion
(Cl-Cl 4) is derived from a heptaketide chain assembled in a linear fashion with
repeating units of acetate. The acetylene moiety arises from an intact unit of
acetate; bromine becomes covalently bound to the C2 position of this acetate unit.
In addition, C27 derives from C2 of acetate. It is hypothesized that this unit may
arise via an HMG CoA synthase like reaction followed by a decarboxylation, which
is then acted upon by a halogenase, to produce the vinyl chloride functionality.
While it is possible to make hypotheses on the sequence of the above described
reactions, molecular genetics will ultimately identify the timing of these events.
Table 111.4. Table of relative enhancement of carbons injamaicamide A enriched
by isotopically labeled feeding experiments (see results and discussion and
experimental sections). The method for the quantitation is detailed in the
experimental section.
Cbona
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
C abundance
[CH3-'3C]-
UC abundance
[1 )3C]-acetate'
'3C abundance
'3C abundance
[2-'3C]-acetatë
[3l3C]alaninec
38.1
1.0
1.6
1.7
0.9
1.1
1.3
0.6
1.1
79.8
19.5
25.8
29.2
141.7
32.5
34.7
36.2
136.5
127.5
28.5
36.6
172.4
1.5
Id
0.9
1.5
0.9
1.0
1.0
1.0
2.0
1.5
1.6
0.9
0.9
1.0
1.2
0.9
1.4
1.9
1.0
1.0
1.3
1.0
2.0
2.0
0.9
1.0
1.0
0.9
1.8
1.5
1.0
2.0
1.0
1.0
1.0
1.1
2.0
1.5
1.0
2.1
1.0
0.9
1.1
1.1
1.1
1.1
1.1
1.3
1.0
1.1
1.1
1.0
1.0
1.2
1.0
1.5
1.1
2.3
1.8
1.7
0.9
0.8
1.3
0.9
0.7
1.1
1.1
1.1
2.4
1.5
1.0
1.0
1.0
0.9
0.9
Chemical
shift (ppm/'
38.1
32.3
175.3
94.9
165.9
169.9
125.7
153.1
17.8
1.0
1.3
56.0
20.7
11.7
1.3
1.1
1.1
58.0
1.0
1.0
1.2
1.2
1.0
1.7
methioninec
1.2
0.9
1.0
0.9
1.0
3.2
1.0
1.1
2.6
1.0
1.5
2.5
1.0
Assignments based on structure ofjaniaicamide A in Figure 111.6. b Referenced to the
CDCI3 centerline at 77.0 ppm. C All enrichments are relative to an average of the
abundances of C16, C22, and C26, with the exception of [CH3J3C]methionine in which
Cl 6, C22, and C24 were used. Bolded/underlined numbers indicate enriched carbon.
Italicized/underlined numbers indicate carbons labeled via alanine catabolism (see Figure
111.12).
The biosynthetic precursor of the Cl 5-C 17 section is unequivocally
established by the observed coupling patterns in the '3C NMR spectrum of
jamaicamide A isolated from cultures provided with exogenous [13C3,'5N]-alanine
('Jc15-c16
= 34.5 Hz,
'Jc16-c17
= 47.9 Hz,
1JC5..N
= 10.5 Hz). The pyrrolidone ring is
comprised of a single unit of acetate condensed with S-alanine as shown by the 13C
enhancements injamaicamide A isolated from L. majuscula provided with [3-'3C}alanine and [1,2-13C2]acetate. The methyl ether carbon (C25) and the C-methyl
group (C26) have been shown to arise from methylation via SAM (2.3 fold and 2.4
fold '3C enhancements observed injamaicamide A isolated from isotope
supplemented cultures).
70
Experimental
General Experimental Procedures. UV and IR spectra were recorded on a
Beckman DU 640B UV spectrophotometer and a Nicolet 510 spectrophotometer,
respectively. NMR spectra were recorded either at a
resonance frequency of
600.04 MHz (Bruker DRX 600), 500.17 MHz (Bruker DRX 500), or 400.13 MHz
(Bruker DPX 400). The Bruker DRX 600 was equipped with a Bruker Q-Switch
TXI probe, the DRX 500 was equipped with a Bruker TXI CryoProbe, and the
DPX 400 was equipped with a Bruker BBO probe. All chemical shifts are reported
relative to residual CHC13 as internal standard. HRMS were obtained on a Kratos
MS 50 TC mass spectrometer. Optical rotations were measured with a PerkinElmer model 243 polarimeter. HPLC was performed using Waters 515 pumps and
a Waters 996 photodiode array spectrophotometer. TLC grade (10-40 m) silica gel
was used for vacuum chromatography, and Merck aluminum-backed TLC sheets
(Silica gel 60 F254) were used for TLC. All solvents were purchased as HPLC
quality. All stable isotope substrates were purchased from Cambridge Isotope, Inc.
Culture Conditions. Following transport to Oregon, the cultures were maintained
in a 28° C controlled temperature room with 16 hr light /8 hr dark cycle provided
by Sylvania 40W cool white fluorescent lights (4.67 pmol photons
m2). The L.
majuscula from the original Hector Bay, Jamaica collection was isolated free of
contaminating cyanobacteria and other microalgae using standard techniques by Dr.
Mary Roberts.5 The liquid culture medium used for the isolation procedure
71
consisted of SWBG1 1 and ESW. Once the cultures were established, scale-up seed
cultures were grown to provide a steady supply of harvestable material. When
sufficient cell mass was grown the cells were harvested according to the techniques
described in Rossi et.
al.6
Extraction and Isolation. A total of 114 g (wet weight) of the harvested alga was
extracted twice with CH2C12/MeOH (2:1) at ambient temperature, followed by
three extractions with heated CH2C12IMeOH (2:1) to yield 1.1 g of crude extract.
The crude extract was then fractionated using vacuum liquid chromatography
(VLC, 9.5 cm x 4 cm) on TLC grade silica gel with a stepwise gradient of
hexane/EtOAc. Eluted material was collected, visualized by 1H NMR, and similar
fractions were recombined. Fractions eluting with a solvent concentration of 5080% EtOAc were further fractionated using a
C18
SPE cartridge with a stepwise
gradient of MeOHIH2O. A fraction eluting in 80 % MeOH in H20 was subjected
to RPHPLC. The final purification was achieved by ODS-HPLC (Phenomenex
250 mm x 10 mm, SPHERECLONE 5m, PDA detection) using 82 % MeOH in
H20 as eluent to give pure jamaicamide A (1, 4% of crude extract), and
jamnaicamide B (2,2.8% of crude extract. Jamaicamide C (3) was isolated from a
subsequent batch of harvested L. majuscula and represents 0.5% of the crude
extract.
72
Jamaicamide A (1). Jamaicamide A (1) was isolated as a pale yellow oil having
272 rim (log c = 3.9); a
the following physical characteristics: UV (MeOH)
=+44°(MeOH,c 1.48); JR (neat) 3314, 2933, 1718, 1666,1599,1543,1431,1395,
1136, 1080, 822 cm '; FABMS (3-NBA) obs. [M + H] cluster at 567/569/57 1
(4:5:1.5); HRFABMS (3-NBA) 567.1625 (-0.7 mmu dev. for C27H37O4N2CIBr); 'H
and '3C NMR data, see Table III.!.
Jamaicamide B. Jamaicamide B (2) was isolated as a pale yellow oil having the
following physical characteristics: UV (MeOH) X
272 nm (log c = 3.9); a
=
+53° (MeOH, c 0.61); IR (neat) 3300, 2931, 2865, 2115, 1718, 1659, 1599, 1544,
1435, 1395, 1136, 1080, 822 cm'; FABMS (3-NBA) obs. [M + H] cluster at
489/49 1 (1:0.4); HRFABMS (3-NBA) 489.2520 (-0.2 mmu dev. for
C27H3804N2C1); 'H and '3C NIvIR data, see Table 111.2.
Jamaicamide C. Jamaicamide C (3) was isolated as a pale yellow oil having the
following physical characteristics: UV (MeOH)
273 rim (log c = 3.8); a =
+49° (MeOH, c 0.39); IR (neat) 3303, 2928, 2857, 1721, 1660, 1601, 1550, 1437,
1397, 1202, 1171, 1082, 823 cm'; FABMS (3-NBA) obs. [M + Hf cluster at
491/493 (1:0.4); HRFABMS (3-NBA) 491.2677 (0.3 mmu dev. for C27H4004N2C1);
'H and '3C NIVIR data, see Table 111.3.
73
Ozonolysis and Acid Hydrolysis of Jamaicamide A. At room temperature, a
slow stream of 03 was bubbled into a 10 mL CH2C12 solution ofjamaicamide A (1,
5 mg, 0.9 mM) which was then sealed in a reaction flask for approximately 5 mm.
The solution was then dried under a stream of N2 and subjected to acid hydrolysis.
Hydrolysis of the jamaicamide A ozomde was carried out in 1 mL of 6N constant
boiling HCI under argon in a threaded Pyrex heavy wall tube sealed with a Teflon
screw cap. The reaction vessel was then placed in a microwave oven (high power
setting, 550W) for 1 min.sa The reaction mixture was dried under a stream of
argon, and derivatized with Marfey's reagent.
Amino Acid Analysis using Marfey's Reagent. To a vial containing 50 tL of a
50 mM solution of pure amino acid standard in H20 was added 100 i.tL of a 36 mM
solution of N-a-(2,4-dinitro-5-fluorophenyl)-L-alanine (FDAA) in (CH3)2C0
followed by 20 p.L of 1 M NaHCO3. The reaction mixture was stirred at room
temperature for 1 h, at which time 10 p.L of 2N HCI was added and let stand for
several minutes.
The jamaicamide A hydrolysate was derivatized by the addition of 100 j.tL of
H20, followed by 500 .tL of a 36 mM solution of FDAA in (CH3)2C0 followed by
100 jiL of 1M NaHCO3. The reaction mixture was stirred at room temperature for
I h, at which time 50 p.L of 2N HCI was added and let stand for several minutes.
The dried reaction mixture was dissolved in 500 p.L of MeOH and analyzed by
ODS-HPLC (Waters Nova-Pak C18 3.9 mm x 150mm 5p, UV detection at 340 nm)
74
with a linear gradient elution [9:1 triethylammonium phosphate (50 mM, pH
3.0):CH3CN to 1:1 triethylammonium phosphate (50 mM, pH 3.0):CH3CN over 60
mini. The derivative of standard S-alanine showed
tR
= 20.46 mm, standard R-
alanine showed tR = 25.59 mm, and S-alanine liberated from jamaicamide A (1)
showed a tR = 20.46 mm, indicating the stereochemistry at C23 was of S
configuration.
General Culture Conditions and Isolation Procedure for Biosynthetic Feeding
Studies. Approximately 2 g of L. majuscula strain JHB-22 Aug 96-01C2 were
inoculated into a 1 L erlenmeyer flask containing 600 mL of SWBG1 1 medium.
The cultures were grown at 28° C under uniform illumination for three days before
addition of the isotopically labeled precursors on days 3, 6, and 8. Cultures of L.
majuscula were harvested 10 days after inoculation, blotted dry, weighed, and
repetitively extracted with 2:1 CH2C12IMeOH. The filtered lipid extracts were
dried in vacuo, weighed, and applied to silica gel columns (1.5 cm x 15 cm) in 50
% EtOAc in hexanes, and eluted in a stepwise gradient from 50 % EtOAC/hexanes
to 100 % EtOAc. The fraction eluting with 100% EtOAc was collected and
concentrated in vacuo for RPHPLC. Final purification proceeded via ODS-HPLC
(Phenomenex SPHERJS ORB, 85 % MeOH in H20, flow rate of 2 mL/min, PDA
detection) to give pure jamaicamide A. For each feeding experiment compound
purity was determined by I{PLC, 'H and 13C NMR.
75
Quantitative Calculation of the '3C Incorporation from Precursor Feeding
Experiments with Jamaicamide A. The relative '3C incorporation into
jamaicamide A from exogenously supplied isotopically labeled precursors were
calculated as follows. All 13C spectra were recorded using inverse-gated
decoupling and processed with 1.0 Hz line broadening (zgig Bruker pulse
program). The 13C NMR resonance intensities for natural abundance and enriched
samples were listed in a database for jamaicamide A. Average normalization
factors were calculated from three carbon resonances within the jamaicamide A
spectrum that were expected to be unlabelled. Resonances for C16, C22, and C26
were used, with the exception that C24 replaced C27 in the S-[methyl'3C]methionine feeding study based on the hypothesis that C27 was incorporated
via methionine. The average normalization factors were determined by dividing
the intensity of the selected resonances from the natural abundance spectrum by the
intensity of the same selected resonances from the '3C enriched spectrum. These
three values were then used to determine the average normalization value. All
resonances in the '3C enriched spectrum were then multiplied by this averaged
normalization value and listed in Table ffl.4. All values were rounded to the
nearest tenth.
Feeding Sodium [I-'3Clacetate. [1-'3C]acetate (150 mg total) was provided to 3
x 600 mL cultures on days 3, 6, and 8. All three cultures were then harvested on
day 10(3.24 g wet wt., 0.34 g drywt., 63.5mg organic extract). A total of 1.4mg
76
of labeled 1 was isolated from the crude algal extract. The 13C NMR spectrum of 1
showed 1.5 fold enhancement at C2, 1.7 at C4, 1.5 at C6, 2.0 at C8, 1.8 at ClO, 2.1
at C12, 2.1 at C14, 1.8 at C19, and 1.7 at C20 (see Table 111.4).
Feeding Sodium [2-'3C]acetate. [2-13C]acetate (150 mg total) was provided to 3 x
600 mL cultures on days 3, 6, and 8. All three cultures were then harvested on day
10(3.64 g wet wt., 0.29 g dry wt., 58.2 mg organic extract). A total of 0.8 mg of
labeled I was isolated from the crude cyanobacterial extract. The 13C NMR
spectrum of I showed 1.7 fold enhancement at Cl, 1.3 at C3, 2.1 at C5, 1.5 at C7,
1.9 at C9, 1.9 at Cli, 2.1 at C13, 2.1 at C18, 2.3 at C21, and 1.7 at C27 (see Table
IH.4).
Feeding Sodium I1,2-'3C2lacetate. [1,2-'3C2Jacetate (80.4mg total) was diluted
1:2 with unlabeled sodium acetate and was provided to 2 x 1 L cultures on days 3,
6, and 8. All three cultures were then harvested on day 10 (4.24 g wet wt., 0.39 g
dry wt., 81.3 mg organic extract). A total of 1.4mg of labeled 1 was isolated from
the crude cyanobacterial extract. Coupling constants for the intact 13C-'3C units are
as follows:
1Jc1-C2
Hz,
= 40.3 Hz, '.Ji i-cu = 43.3 Hz,
1Jc9.c10
'Jc21-c2o
= 170.7 Hz,
= 63.1 Hz.
JC3-C4
= 33.9 Hz,
1JC5.C6
'Jc13-c14 =
= 43.3 Hz,
49.7 Hz,
'Jc7.c8
.1c18-C19 =
= 33.3
71.9 Hz,
77
Feeding S-[3-'3Cjalanine. S-[3-'3C]alanine (195 mg total) was provided to 3 x
600 mL cultures on days 3, 6, and 8. All three cultures were then harvested on day
10(5.09 g wet wt., 0.64 g dry wt., 49.1 mg organic extract). A total of 2.0 mg of
labeled I was isolated from the crude algal extract. The '3C NMR spectrum of I
showed 2.8 fold enhancement at C24 (see Table 111.4).
Feeding ['3C3,'5NJf3-alanine. ['3C3,'5N1-alanine (150 mg total) was provided to 2
x 1 L cultures on days 3, 6, and 8. All three cultures were then harvested on day 10
(4.48 g wet wt., 0.49 g dry wt., 34.2 mg organic extract). A total of 1.4 mg of
labeled 1 was isolated from the crude algal extract. Coupling constants for the
intact 13C-13C and '3C-15N units are as follows:
Hz,
JC15-N =
'Jc15c16 =
34.5 Hz,
1Jc16.c17
= 47.9
10.5 Hz.
Feeding S-[methyl-'3C]methionine. S-[methyl-13C]methionine (6 mg total) was
provided to 3 x 600 mL cultures on days 3, 6, and 8. All three cultures were then
harvested on day 10(2.76 g wet wt., 0.24 g dry wt., 83.2 mg organic extract). A
total of 0.9 mg of labeled 1 was isolated from the crude algal extract. The
'3C
NtvlR spectrum of 1 showed 2.3 fold enhancement at C25, and 2.4 at C26 (see
Table ffl.4).
References
1.
Faulkner, D. J. Nat. Prod. Rep. 2000, 17, 7-55.
2. For representative see: Milligan, K. E.; Marquez, B.; Williamson, R. T.;
Davies-Coleman, M.; Gerwick, W. H. J. Nat. Prod. 2000, 63, 965-968.
3. Marquez, B. L.; Watts, K. S.; Roberts, M. A.; Verdier-Pinard, P.; Jimenez, J. I.;
Ho, P. S.; Hamel, E.; Scheuer, P. J.; Gerwick, W. H. manuscript in preparation.
4. (a) Leusch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J.; Mooberry, S. L. .1.
Nat. Prod. 2000, 63, 611-615. (b) Leusch, H.; Yoshida, W. Y.; Moore, R. E.;
Paul, V. J. J. Nat. Prod. 2000, 63, 1437-1439. (c) Milligan, K. E.; Marquez, B.
L.; Williamson, R. T.; Gerwick, W. H. J. Nat. Prod. 2000, 63, 1440-1443.
5. (a) Sitachitta, N.; Marquez, B. L.; Williamson, R. T.; Rossi, J. V.; Roberts, M.
A.; Gerwick, W. H.; Nguyen, V-A.; Willis, C. L. Tetrahedron 2000, 56, 91039113. (b) Sitachitta, N.; Rossi, J.; Roberts, M. A.; Gerwick, W. H.; Fletcher,
M. D.; Willis, C. L. J. Am. Chem. Soc. 1998, 129, 7131-7132. (c) Orjala, J. 0.;
Gerwick, W. H. .J. Nat. Prod. 1996, 59, 427-430.
6.
Mann/it, 2000; database of the marine natural products literature; Department
of Chemistry, University of Canterbury: Christchurch, New Zealand, 2000.
7. Patil, A. D.; Kokke, W. C.; Cochran, S.; Francis, T. A.; Tomszek, T.; Westley,
J. W.J. Nat. Prod. 1992, 55, 1170-1177.
8. Williamson, R. T.; Marquez, B. L.; Gerwick, W. H.; Koehn, F. E. Magn. Reson.
Chem. In press.
9. (a) Rossi, J. V.; Roberts, M. A.; Yoo, H-D.; Gerwick, W. H. J. App. Phycol.
1997, 9, 195-204. (b) Sone, H.; Kondo, T.; Kiryu, M.; Ishiwata, H.; Ojika, M.;
Yamada, K. J. Org. Chem. 1995, 60,4774-4781.
10. (a) Kay, L. E.; Keifer, P.; Saarinen, T. .1. Am. Chem. Soc. 1992, 114, 1066310665. (b) Palmer ifi, A. G.; Cavanagh, J.; Write, P. E.; Rance, M. J. Magn.
Reson. 1991, 93, 151-170. (c) Kontaxis, G.; Stonehouse, J.; Laue, E. D.;
Keeler, J. J. Magn. Reson. Ser. A 1994, 111, 70-76.
11. HSQC-TOCSY with 12 Msec mixing time (dipsi)
79
12. (a) Wiliker, W.; Leibfritz, D.; Kerssebaum, R.; Bermel, W. Magn. Reson.
Chem. 1993, 31, 287-292. (b) Bax, A.; Summers, F. J. Am. Chem. Soc. 1986,
108, 2093-2094.
13. (a) Martin, G. B.; Crouch, R. C.; Sharaf, M. H. M.; Schiff, P. L., Jr. 34th Annual
Meetig of the American Society of Pharmacognasy, San Diego, CA, July 18-22,
Abstract P101. (b) Uzawa, J.; Utumi, H.; Koshino, H.; Hinomoto, T.; Anzai, K.
32T NMR Conference, Tokoyo, Japan, November 4-6, 1993; pp 147-150.
14. Purchased from the Sigma-Aldrich Library of Rare Chemicals, Catalog No.
S028879.
15. Marfey, P. CarlsbergRes. Commun. 1984, 49, 591-596.
16. Williamson, R. 1.; Marquez, B. L.; Gerwick, W. H.; Kover, K. E. Magn. Reson.
Chem. 2000, 38, 265-273.
17. (a) Stott, K.; Keeler, J.; Van, Q. N.; Shaka, A. J. J. Magn. Reson. 1997, 125,
302-324. (b) Stott, K.; Stonehouse, J.; Keele, J.; Hwang, T-L.; Shaka, A. J. J.
Am. Chem. Soc. 1995, 117, 4199-4200.
18. (a) Koehn, F. B.; Longley, R. E.; Reed, J. K. J Nat. Prod. 1992, 55, 613-619.
(b) Yoo, H-D.; Gerwick, W. H. unpublished results.
19. Pettit, G. R.; Kamano, Y.; Dufresne, C.; Cerney, R. L.; Herald, D. L.; Schmidt,
J. M. J. Org. Chem. 1989, 54, 6005-6006.
20. Nagle, D. G.; Paul, V. J.; Roberts, M. A. Tetrahedron Lett. 1996, 37, 62636266.
21. Voet, D.; Voet, J. G. In Biochemistry; Wiley & Sons, Inc.: New York, 1990;
Chapter 24.
CHAPTER IV
THE STRUCTURE ELUCIDATION OF DECHLOROBARBAMIDE AND
BIOSYNTHETIC INVESTIGATIONS OF BARBAMIDE
Abstract
Dechlorobarbamide was isolated from a Curacao collection of the marine
cyanobacterium Lyngbya majuscula and its structure determined through
spectroscopic analysis and comparisons with barbamide. The absolute
stereochemistry of the dolaphenine moiety of barbamide was determined to be S,
defining the absolute configuration of barbamide as 2S,7S. Biosynthetic experiments
are described that explore the chemical precursors that constitute barbamide.
Experiments with L-[2H10]leucine demonstrated that chlorination of the pro-R methyl
occurs without detectable activation via the leucine-catabolic pathway. Moreover, an
extremely high level of incorporation of fed [2-' 3C]-5,5 ,5-trichloroleucine into
barbamide indicates that leucine is the probable substrate for the chlorination reaction.
Incorporations of [1,2-'3C2]acetate and [1-13C, l-'80]acetate confirmed the origins of
C-5 and C-6 whereas incorporation of L-[3-'3C]phenylalanine supported the
hypothesis that the phenyl group and its three carbon side-chain in barbamide (C-7, C-
8 and C-lO-C-16) arise from phenylalanine. The thiazole ring (C-17-C-18) of 1 was
shown to likely arise from cysteine through a [2-'3C, '5N]glycine feeding experiment.
E1l
Introduction
Marine organisms are prolific sources of halogenated secondary metabolites.'
A majority of these halogen atoms are incorporated into positions which are
suggestive of their biochemical reaction as electrophilic species. Haloperoxidase
enzymes responsible for the formation of the Xthalogenating species have been found
in many classes of marine organisms and their study has been an area of intense
interest.2
In contrast, a number of sponge-cyanobacterial and cyanobacterial
metabolites possess halogenated functional groups wherein the electronic nature of the
halogenating species is uncertain.' Such an example is the unusual trichloromethyl
group of barbamide (1), a molluscicidal metabolite our group isolated from the marine
cyanobacterium Lyngbya
majuscula.3
CH,
N'S
OCH, CCI3
°
mn
CU3
N"S
OCH3 CUd2
°
\-=/
2
1
Figure IV.1. Structure of barbamide (1) and dechlorobarbamide (2).
Barbamide is an intriguing natural product for several reasons. First, it
possesses the rare trichioromethyl group, a feature which has previously only been
found in a series of sponge-derived metabolites, such as dysidin
diketopiperazine derivative
(4)5
(3)4
and
However, it should be noted that these latter
chlorinated metabolites have recently been localized to the sponge-associated
cyanobacterium Oscillatoria spongeliae.6 Second, barbamide ( and
dechlorobarbamide) contains a dolaphenine moiety, which is a structural feature found
in several biologically active natural products such as dolastatin 1 O'' and symplostatin
1
Last, the discovery of these metabolites (barbamide and dechlorobarbamide)
provides clarification of the metabolic origin of related compounds isolated from
sponge-cyanobacteria assemblages.
H3CO-,,CC13
oJJ
c13cLro
N'CC13
H
OCH3
4
Figure IV.2. Chemical structures of dysidin (3) and a trichlorodiketopiperazine (4).
Since barbamide is produced in laboratory cultures of L. majuscula originally
collected off the coast of Curacao (Ca. 2.4% of the extract), we have been able to
experimentally determine the biosynthetic precursors of barbamide using stable-
isotope labeling methods.9"° We have found that the trichioromethyl group of
barbamide derives from the pro-R methyl group of leucine,'1 and that this chlorination
occurs without detectable activation of the methyl group to facilitate a potential
nucleophilic or electrophilic
process.9
Hence, it is proposed that chlorination of this
83
leucine methyl group occurs through novel biochemical processes, possibly involving
radical
chemistry.9
In this chapter, elaboration of the isolation and structure determination of a
new barbamide derivative, dechlorobarbamide (2), as well as the determination of the
absolute stereochemistry at C-7 of barbamide (1) is discussed. In addition, both a
summary of previously communicated stable isotope feeding experiments, done by Dr.
Namthip
Sitachitta,9°
and an explanation of a new feeding experiment which supports
trichioroleucine as an intermediate in the biosynthetic pathway of barbamide will be
presented. Syntheses of all chiral isotopically labeled leucine compounds were
performed in the laboratory of Dr. Christine L. Willis at the University of Bristol.
H3C
A.
03c
OH
Pyruvate
B.
ci3q .H
7CH3
(H
Chlorination
NH34
Chlorination
0
(4S)-5,5,5-Trichloroleucine
H3q H
7CH3
(H
NH3
0
Leucine
Figure IV.3. Biosynthetic hypotheses for the formation of barbamide; pathway A,
chlorination predicted to occur during biosynthesis of leucine from pyruvate; pathway
B, chlorination is believed to occur by novel mechanisms acting directly on leucine.
Two possibilities are envisioned in the biosynthesis of barbamide, both of
which predict 5,5,5-trichloroleucine as the metabolic precursor to C-1-C-4 and C-9 of
barbamide, but which differ in the timing and biochemical mechanism of the
chlorination reaction (Figure IV.3). In the first, we propose that chlorination occurs
84
during the biosynthesis of leucine, perhaps at the pyruvate stage, during which point
this methyl group is activated to electrophilic mechanisms of chlorination. In the
second, we and others propose that chlorination occurs via a novel mechanism on the
inactivated methyl group of intact leucine.9"2 In either case, the intermediacy of 5,5,5trichioroleucine is envisioned. Transamination and decarboxylation of 5,5 ,5trichloroleucine followed by ketide extension by malonyl CoA could give rise to
carbons 1-6, and 9 of barbamide. It is reasonable to predict that phenylalamne and
cysteine serve as precursors to the phenyl and thiazole rings, respectively. In
agreement with precedents from the biosynthesis of other nonribosomal polypeptides,
N-methylation by S-adenosylmethiomne (SAM) is predicted to occur prior to amide
bond formation between the activated acyl group of the ketide extended
trichloroleucme fragment and the phenylalanine residue.'4 Following amide bond
formation with an activated cysteine residue, heterocyclization of the cysteine side
chain with the carbonyl carbon of phenylalanine followed by oxidative
decarboxylation is predicted to complete formation of the thiazole ring, although the
timing of these reactions relative to other steps in the pathway is uncertain. Finally, at
some point in the pathway, O-methylation of the enol hydroxy group at C-4 also
occurs with involvement of SAM. Currently, efforts to discern the sequence of events
leading to the formation of barbamide and the characterization of the biosynthetic
gene cluster are currently underway in the Gerwick laboratory.
85
Results and Discussion
Isolation and Structure of Dechlorobarbamide (2). The lipid extract from a
1996 Curacao collection of L. majuscula was subjected to silica gel vacuum liquid
chromatography (100% hexanes to 100% ethyl acetate, v/v). A relatively non-polar
fraction (50% ethyl acetate/hexanes) was further fractionated using ODS vacuum
liquid chromatography. A final purification utilizing ODS-HPLC yielded barbamide
(1, Ca. 2.4% of ext.) and dechlorobarbamide (2, ca. 0.1% of ext.).
Comparison of the 'H and '3C NMR data for dechlorobarbamide (2) with those
of barbamide (1) clearly indicated that the two metabolites were closely related.
However, HIR-FABMS of 2 gave an [M+H] at 427.1014 analyzing for
C20H25C12N2S02,
indicating that 2 contained one less chlorine atom than barbamide
(1). Examination of 13C NIvIR data for compounds 1 and 2 revealed that their only
significant difference was the assigned chemical shift at C-9 (6105.6 for 1 and 879.64
for 2). Moreover, an HSQC correlation showed that this new methine carbon in
compound 2 was directly bound to a proton at 6 6.29. 'H-'H COSY showed that this
proton was adjacent to the H-2 methine (6 2.38). Hence, it was clear that
dechiorobarbamide (2) contained a dichioromethyl moiety at C-9 in contrast to the
trichioromethyl group found at this position in barbamide (1). The remaining
structural features of 2 were confirmed as being identical to those found in 1 by 'H-1H
COSy,'5 HSQC,'
and HMBC'7 data (Table IV.1). The i
double bond geometry of
2 was established as E by observation of strong NOE between the OCH3 resonance
(63.57) and H-5 (65.30) upon selective irradiation of the latter signal using the
DPFGSE 1D NOE pulse
sequence.'8
At 50 ppm, dechiorobarbamide (2) was inactive
in the Biomphaleria glabrata molluscicidal assay.
Table IV.1. 1H NMR (600 MHz, DMSO) and 13C NMR (150 MHz, DMSO) data for
the major conformer of dechiorobarbamide (2).
Position
1
2
3a
b
4
5
'H 6 (mult, J in Hz)
0.93(d,6.6)
2.38 (m)
2.81 (dd, 12.1, 8.7)
8a
b
9
13.6(q)
41.6(d)
33.6 (t)
2.54 (dd, 12.1,4.2)
5.30 (s)
6
7
13C & (multb)
6.31 (obscured)
3.51 (dd, 14.8, 9.2)
3.25 (dcl, 14.8, 10.5)
6.29 (obscured)
167.8(s)
93.2 (d)
167.2(s)
54.0(d)
35.7 (t)
79.6(d)
HMBC correlations
33.6,41.6,79.6
13.64, 33.58
13.6,41.6, 79.6, 167.8,
13.6,41.6, 79.6, 167.8
33.6, 167.8
35.7, 167.2, 169.5
54.0, 129.0, 137.8, 54.0,
129.0, 137.8, 169.5
13.6, 33.6
137.8 (s)
10
11
7.35(br. d)
129.0(d)
12
7.27 (m)
128.1 (d)
13
14
7.19(br. d)
126.2(d)
7.27 (m)
128.7 (d)
15
7.35 (br. d)
16
17
18
7.79 (d, 3.2)
7.70(d,3.2)
129.0(d)
169.5(s)
141.9(d)
120.6(d)
NCH3
OCH3
2.86 (s)
3.57 (s)
30.6 (ci)
55.2 (ci)
128.1, 126.2
126.2, 128.7, 137.8
128.7
126.2, 128.7,137.8
126.2, 128.7
141.9
54.0, 167.2
167.8
a
The major conformer was shown to be of E geometry by observation of NOE between H-5
and the OCH3 resonance. b Multiplicity was determined using the DEPT 135 pulse sequence.
Determination of the C-2 and C-7 Stereochemistry in Barb amide (1). The
stereochemistry of C-2 in barbamide (1) was determined through feeding (2S,4R)-[5'3C]Ieucine and (2S,45)-[5-'3C]leucine (see below).9 Selective incorporation of
(2S,4R)-[5-' 3C]leucine into C-9 of barbamide indicates that the stereochemistry of C-2
is S. In order to determine the absolute stereochemistry at C-7 of barbamide (1), N-
87
methyiphenylalanine was liberated through ozonolysis19 and acid hydrolysis for
subsequent derivatization and Marfey's analysis. Of interest, the acid hydrolysis (2
mL, 6 N HC1) was performed in one minute in an ordinary microwave oven (550 W).
This fast and simple procedure should find wide application in the hydrolysis of
amides and esters for microanalysis procedures.2° The released N-
methyiphenylalanine was denvatized with N-a-(2,4-dinitro-5-fluorophenyl)-L-alanine
(FDAA) for Marfey's analysis,2' and determined to be S by comparison of retention
times and co-injections with derivatized (S-), (R-), and (S,R)-N-methylphenylalanine
standards. The stereochemistry of dechlorobarbamide (2) is proposed to be the same
as that in barbamide (1 - 2S,7S) by virtue of a) its co-occurrence in this
cyanobacterium and likely similar biogenesis, and b) its comparable optical rotation
{Iit. for 1
[a]25D -89°
0.046); for 2
[cL]25D
(MeOH, c 1 3),3 re-measurement of 1
[a]25D -82°
(MeOH, c
-67° (MeOH, c 0.046)).
Isolation of Barbamide (1) for Biosynthetic Studies. The marine
cyanobacterium L. majuscula was cultured in our laboratory as described
previously.22
For the precursor incorporation experiments, 50-75 mL of wet-packed cells of L.
majuscula (strain 19L) were inoculated into fresh medium (SWBG1I).22 After three
days of acclimation, isotope-labeled precursors were administered to the cultured
cyanobacterium, incubated for an additional 6 to 7 days and then harvested. The crude
organic extract was subjected to normal phase vacuum liquid chromatography (NPVLC),
C18
solid phase extraction cartridge (SPE), and ODS-HPLC (C18, 80%
MeOHIH2O), respectively, to provide the variously labeled barbamides (1).
Review of Previous Feeding Studies. Previous feeding experiments done by
Dr. Narnthip Sitachitta are detailed below. Determination of the dolaphenine moiety
was deduced through two separate feeding experiments. First, to explore the origin of
the phenyl moiety and adjacent carbon atoms of barbamide (C-7-C-8 and C-10-C-16),
L-[3)3C]phenylalanine was provided to L. majuscula cultures. Analysis of the 1D 13C
NMR spectrum of barbamide isolated from this feeding showed significant
enhancement of C8 (290%) Second, exploration of the origin of the thiazole moiety
was done. Because of cost, and the associated toxicity with feeding cysteine or senne
to L. majuscula an alternative precursor was used. Glycine, an intermediate in the
metabolic pathway of cysteine, was used. Analysis of barbamide from a doubly
labeled [2-13C,'N]glycine experiment showed intact incorporation of the label by
virtue of a splitting of the 2JNH correlation by the
1JCH
coupling (ca. 180 Hz), thereby
showing that cysteine was the precursor to the thiazole ring.
To determine if C-5 and C-6 of barbamide (1) originate from an intact
acetate unit, a [1 ,21 3C2]acetate feeding experiment was conducted. The
13(
NMR
spectrum of 1 produced under these conditions showed an additional doublet structure
for the C-5 and C-6 signals ('J= 71.8 Hz). In addition, an [1-13C,'802]acetate
feeding experiment showed both the orientation of the acetate unit in barbamide, as
well as the origin of the C-6 oxygen atom.
To establish that the N-CH3 and O-CH3 of 1 derive from the C1 pool via SAM,
L-[methyl-13C]methionine was administered to cultures of L. majuscula. Only slight
enrichments were observed. The lack of enrichment is though to be a result of the
small quantities of methionine fed due to its toxic effects to the organism. However,
approximately 2 fold enhancements were observed from the [2)3C,'5N]glycine
feeding, apparently resulting from the contribution of C-2 of glycine to the
C1
pooi, a
carbon source for both N-CH3 and O-CH3 groups.
Many feeding experiments were performed utilizing labeled leucine
precursors. Feeding of L-[2-13C}leucine clearly indicated that leucine or a catabolite
of leucine is the substrate for metabolic
chlorination.9
To examine the fate of Cl of
leucine in the biosynthesis of 1, similar amounts of both L-[1-13C}leucine and L-[2-
'3C]leucine were simultaneously provided to cultures of the cyanobacterium. The '3C
NMR spectrum showed no enhancement of C5, whereas C4 was significantly enriched
(230%) thereby confirming that Cl of leucine is lost in the biosynthesis of barbamide.
To analyze which methyl group of leucine undergoes chlorination, chirally labeled
(2S,4R)-[5-'3C]leucine and (2S,45)-[5)3C]leucine were separately provided to
cultures.2325 Results from the incorporation of the two chirally '3C-labeled leucines
into 1 demonstrated that: 1) the chlorination reaction occurs at the pro-R methyl group
of leucine, and 2) the stereochemistry at C-2 in barbamide is S. To explore the
possible intermediacy of a modified leucine catabolite, L-[2Hio]leucine was fed to
cultures of L. majuscula and the resulting barbamide (1) was analyzed by 2H NMR.
Integration of the 2H NMR spectrum revealed a ratio of 3.00:2.77 for (2H3-1):(2H-2 +
2H2-3), indicating that there was no loss of deuterium from C-3 or C-4 of leucine
during its incorporation into barbamide and therefore the trichioromethyl group of
barbamide (1) is not activated to electrophilic chlorine addition via the leucine
catabolic pathway.
Latest Feeding Studies. The results from the feeding experiments described
above are displayed in Figure P1.4. These feeding experiments provided clear insight
into the origin of most of the carbon atoms and heteroatoms of 1. The findings that
leucine is the substrate for chlorination has been shown indirectly by virtue of its
incorporation into barbamide. Furthermore, leucine undergoes no activation to form a
species suitable for electrophilic addition. While this suggests the possibility of a
novel mechanism for the halogenation process, it remains unknown at what point in
the biosynthesis this halogenation occurrs. Therefore, the intermediacy of
trichioroleucine, as shown in Figure P1.3, was explored with isotopically labeled L-[2'3C]-5,5,5-trichioroleucine.
[2-'3C]-5,5,5-Trichloroleucine Feeding Experiment. To directly evaluate the
possible intermediacy of L-5,5,5-trichloroleucine in the biosynthetic pathway of
barbamide, synthetic [2-'3C]-5,5,5-trichloroleucine was prepared. Routes for the
synthesis of trichioromethylbutanoic acid have been previously reported.29 [113C}Trichioromethylbutanoic acid was converted to [2-13C]trichloroleucine via a
Strecker reaction. A mixture of diastereomers was formed, the major isomer
possessing the required (45)-stereochemistry. This was provided to cultures of L.
majuscula (2 x 1 L, 80 mg each), and after 10 days total, the cells were harvested and
barbamide isolated. '3C NMR analysis of this sample in toluene-d8 showed specific
and very high (ca. 30-fold over natural abundance) incorporation of '3C into C-4
(Figure P1.4).
91
180
160
140
120
100
80
60
40
20
ppm
Figure 1V.4. 13C NMR spectra of barbamide (1) produced by L. majuscula culture 19L
a) supplemented with [2-13C]-5,5,5-trichloroleucine, and b) natural abundance control
[C-4 of barbamide is indicated (deriving from C-2 of [2-'3C]-5,5,5-trichloroleucine]
(C-4 = major amide isomer; C-4' = minor amide isomer). Both 100 MHz '3C NMR
spectra were acquired in toluene-d8 with 24K data points and 3.0 Hz line broadening.
All of the incorporation studies described above have provided insight into the
biosynthetic origin of all carbon atoms and most heteroatoms in barbamide (1). As
depicted in Figure IV.5, incorporations of L-[3-'3 C]phenylalanine and [1,2'3C2]acetate into 1 provided insights into the origins of the phenyl moiety and C-5-C-6
of the molecule, respectively. The intact incorporation of [21 3C, 15N]glycine into the
thiazole ring of 1, detected using a new modified GHNMBC NMR experiment,
strongly supports cysteine as a direct precursor to this part of barbamide. Analysis of
92
the '3C NMR spectrum of! from this latter feeding experiment, as well as from
experiments wherein exogenous [methyl-' 3C]methionine was provided, supplied
convincing evidence that the N-CH3 and O-CH3 groups both derive from the C, pool.
methyl-'3CJmethionine
(2S,4Sj-I23CI-5,S,5-trichloroleucine
(Cl Pool)
/
[3-'3Cjphenylalanine *
\\
CH3
17
12-'3C,'5N1
glycine £
4_[
serine -- cysteine
\\
.
I.
OCH3
Il,2-'3C2lacetate
I1-'3C,'80)acetate o
C13
H
12-'3Clleucine *
(2S,4R)-[53CJteucine V
(2S,4S)-15-'3CJleucine
Figure IV.5. Summary of biosynthetic precursors of barbamide (1). Note,
intermediates in parentheses (serine and cysteine) are hypotheses not yet demonstrated
by direct precursor feeding-incorporation experiments (see text), and several carbon
atoms (C-7 and C-lO-C-16) hypothesized to derive from L-phenylalanine have not yet
been confirmed through specific incorporation experiments.
The leucine feeding experiments have shown that C-I -C-4 plus C-9 of 1
originate from L-leucine. Results from incorporation of the two chirally labeled
leucines established the 2S stereochemistry of barbamide and that the chlorination
reaction occurs at the pro-R methyl of leucine. An incorporation experiment using L[2H,o]leucine showed that the leucinepro-R methyl group is not activated via the
leucine catabolic pathway. The very high level of incorporation from exogenously
applied [2-' 3C] -5,5 ,5-trichloroleucine strongly suggests that trichloroleucine (3) is an
intermediate in the pathway. Taken together, these results indicate that L-leucine is
the substrate for chlorination in this strain of L. majuscula, and that this reaction
occurs without activation of the pro-R methyl group to electrophilic or nucleophilic
mechanisms of chlorine addition. Therefore, we suspect that novel chlorination
reactions, perhaps involving radicals, are
involved.9"2
Moreover, our isolation of
dechiorobarbamide (2) as a minor natural product of L. majuscula suggests that this
chlorination process does not occur by oxidation to a carboxylic acid equivalent
followed by multiple additions of chlorine, but rather, occurs stepwise to form
dichioro- and then trichioromethyl functionalities. Leucine-denved natural products
from cyanobacteria living in association with sponges show a similar spectrum of diand fri-chlorinated methyl groups.5
94
Experimental
General. Nuclear magnetic resonance (NMR) spectra were recorded on
Bruker AM400 and DPX 400 instruments operating at 400.13 MHz for 'H NMR,
61.45 MHz for 2H NMR, and at 100.61 MHz for '3C NMR. Spectroscopic
characterization of synthetic '3C-labeled precursors utilized JEOL 400 MHz ('H) and
270 MHz ('3C) instruments. The GHNMBC experiment performed on 1, and various
NMR experiments in support of the structure elucidation of dechiorobarbamide, were
acquired on a Bruker DRX600 spectrometer operating at a 'H frequency of 600.08
MHz and a '3C frequency of 150.01 MHz. Proton spectra were referenced to 2.50
ppm and 2.09 ppm for DMSO-d6 and toluene-d8, respectively. Carbon spectra were
referenced to 39.51 ppm for DMSO-d6 and 20.4 ppm for toluene-d8. Highperformance liquid chromatography (HPLC) utilized Waters M6000A or Waters 515
pumps, a Rheodyne 7125 injector, and a Waters Lambda-Max 480 LC
spectrophotometer or Photodiode Array Detector model 996. Merck aluminumbacked thin layer chromatography (TLC) sheets (silica gel 60 F254) were used for
TLC. Vacuum liquid chromatography (VLC) was performed with Merck Silica Gel G
for TLC or with Baker Bonded Phase-octadecyl (C 18). All solvents were either
distilled from glass or of HPLC quality. All stable isotope labeled substrates, other
than sodium [l-'3C,'80]acetate which was a gift of S.J. Gould (Chemistry, OSU), were
purchased from Cambridge Isotope, Inc.
Collection. The marine cyanobacterium L. majuscula (voucher specimen
available from WHG as NSB-4 May 96-1) was collected from shallow water (0.1-I m)
on 4 May 1996, at Barbara Beach (Spanish Waters), Curaçao, Netherlands Antilles,
and stored in 2-propanol at reduced temperature until workup.
Bioassay for molluscicidal activity. Evaluation of the molluscicidal activity
of dechlorobarbamide (2) was performed as previously detailed using the test
organism Biomphalaria glabrata.3 Variable amounts of dechiorobarbamide (2) were
dissolved in 20 pL of EtOH and added to 20 mL distilled H20. The snails (2 snails!
assay vessel) were observed after 24 hr and considered dead when no heartbeat could
be detected upon microscopic investigation.
Extraction and isolation of dechlorobarbamide (2). A total of 83.2 g (dry
wt) of the alga was extracted with CH2C12/MeOH (2:1) twice to give 2.29 g of crude
extract. The extract was fractionated using vacuum liquid chromatography (VLC, 9.5
cm x 4 cm) on TLC grade Si gel with a stepwise gradient of hexanes/EtOAc. Eluted
material was collected, visualized by TLC, and similar fractions recombined. A
fraction eluting with 50% EtOAc/hexanes was further fractionated by ODS VLC (3
cm x 5 cm) using a MeOHIH2O gradient (50% MeOH-100%MeOH). Final
purification was achieved by ODS-HPLC (Phenomenex 250 mm x 10 mm
Sphereclone 5
t,
UV detection at 254 nm) using MeOHIH2O (4:1) as eluent to give
pure dechiorobarbamide (2,
Ca.
1.9 mg, 0.1 % of extract) as an oil.
Dechlorobarbamide (2). Dechlorobarbamide (2) was isolated as a pale
yellow oil showing the following: UV (MeOH) 2
238 nm
(E
= 16 000);
[U]25D
-67°
(MeOH, c 0.046); IR Vmax (film) 2927, 1643, 1603, 1453, 1441, 1244, 1167, 1114, 742
cm1; FABMS (3-NBAJ2%TFA) obs. [M+H] cluster at m/z 427/429/431 (100:67:17),
209/211/213 (100:67:17); HRFABMS (3-NBA! 2% TFA) 427.1014 (-0.6 mmu dev.
for C20H25C12N2 02S); for 'H and '3C NIvIR see Table.
Ozonolysis and Acid Hydrolysis of Barb amide. A slow stream of 03 was
bubbled into a 15 mL CH2C12 solution of barbamide (1, 0.73 mM) that was then sealed
in a reaction flask for approximately 8 mm. The solution was then dried under a
stream of argon and subjected to acid hydrolysis. Hydrolysis of the barbamide
ozonide (Ca. 2.5 mg) was carried out in 2 mL of 6 N constant boiling HC1 under argon
in a threaded Pyrex heavy wall tube sealed with a Teflon screw cap. The reaction
vessel was then placed in a microwave oven (high power setting, 550 W) for one
minute.20 The reaction mixture was dried under a stream of argon, and derivatized
with Marfey's reagent.
Amino Acid Analysis using Marfey's Reagent. To a vial containing 50 j.tL
of a 50mM solution of pure amino acid standard in H20 was added 100 tL of a 36
mM solution of N-a-(2,4-dinitro-5-fluorophenyl)-L-alanine (FDAA) in (CH3)2C0
followed by 20 iiL of 1 M NaHCO3. The reaction mixture was stirred at room
temperature for one hour, at which time 10 tL of 2 N HC1 was added and let stand for
several minutes. The barbamide hydrolysate was derivatized by the addition of 100
j.xL of H20, followed by 500 j.iL of a 36 mM solution FDAA in (CH3)2C0 followed by
100 tL of 1 M NaHCO3. The reaction mixture was stirred at room temperature for
one hour, at which time 50 jtL of 2 N HCI was added and let stand for several minutes.
The dry reaction mixture was dissolved in 500 iL of MeOH and analyzed by ODSHPLC (Phenomenex 250 mm x 10 mm Sphereclone 5 p, UV detection at 340 nm)
97
with a linear gradient elution [9:1 triethylammonium phosphate (50 mM, pH
3.0):CH3CN to 1:1 triethylammonium phosphate (50 mM, pH 3.0):CH3CN over 60
mini. The derivative of standard N-methyl-D-phenylalanine showed tR =37.31 mm,
standard N-methyl-L-phenylalanine showed tR = 36.75 mm, and N-methyl-Lphenylalanine obtained from barbamide (1) showed a tR = 36.74 mm.
Analytical data for [2-'3C]-5,5,5-trichloroleucine. 1H NMR (400 MHz, D20,
both diastereomers) 3.8 (1H, din, J= 125 Hz, H-2), 2.98 (1H, m, minor diastereomer,
H-4), 2.87 (1H, m, major diastereomer, H-4), 1.97-2.55 (2H, m, H2-3), 1.40 (3H, d, J
= 6.5 Hz, H3-6); 13C NMIR (67.9 MHz, D20, major diastereomer) E,174.1 (C-i), 104.6
(C-5, -CCI3), 53.2 (C-2, enriched 99%), 51.1 (C-4), 34.8 (C-3), 15.9 (C-6, -CH3); CI
MS (relative abundance) obs. m/z 189.9835 (100%) [(MH
CO2H) C413CH9N35C13
requires 189.9834], 153 (70), 82 (35), 75 (96).
General Culture Conditions and Isolation Procedure. Approximately 3 g of L.
majuscula strain 19L were inoculated into a 2.8 L Fernbach flask containing 1 L of
SWBG1 1 medium. The culture was grown at 2°C under uniform illumination (4.67
tmol photon 5'm2), aerated, and acclimated for 3 days prior to addition of
isotopically labeled precursors. Cultures of L. majuscula were harvested 10 days after
inoculation, blotted dry, weighed, and repetitively extracted with 2:1 CH2C12IMeOH.
The filtered organic extracts were dried in vacuo, weighed, and applied to silica gel
columns (1.5 cm I.D. x 15 cm) in 5% EtOAc/hexanes, and eluted with a stepped
gradient elution of 5% EtOAc to 100% EtOAc. Fractions containing barbamide
(eluted with 50% EtOAc/hexanes) were further fractionated by RP-VLC using a
stepped gradient elution from 60% MeOH/H20 to 100% MeOH. The fractions eluting
with 80% MeOH (barbamide-containing fraction) were subjected to a final
purification by ODS-HPLC [Phenomenex Spherisorb ODS (2), 4:1 MeOHTH2O, flow
rate 3 mL/min, detection at 254 nm] to give pure barbamide (1, 3.86 mg/L). For each
feeding experiment, barbamide identity and purity was established from TLC, PDAHPLC, 'H and '3C NMR spectroscopy.
Calculation of the Results of 13C-Labeled Precursor Feeding Experiments on
Barbamide. The percentage '3C incorporation into barbamide from exogenously
supplied substrates was calculated as follows. The '3C NMR spectral data and
integrations for natural abundance and enriched samples were listed in a database for
both N-methyl amide conformers of barbamide (1). Normalization factors for every
carbon atom in barbamide were calculated by sequentially dividing the integral for
each natural abundance carbon atom into the integration values of all carbon atoms in
the natural abundance spectrum in turn (e.g. in this case, 20 columns of normalization
factors were generated). Multiplication of the normalization factors for each
resonance by the integrated value of the carbon atom being used for normalization in
the '3C enriched sample provided "expected integration values" for each resonance in
the enriched spectrum (20 columns of data). These were used to calculate the
percentage '3C enhancement of each signal by dividing the integrated area of each
carbon peak in enriched barbamide by the above calculated "expected integration
values", and multiplying by 100 (20 columns of calculated percentages). Finally, the
average percentage enhancement for each carbon signal was calculated by considering
all values except those expected to show '3C enrichment, and then rounding to the
nearest 10%.
Feeding [2)3C1-5,5,5-Trichloroleucine HCI. [2-' 3C]-5 ,5,5-trichloroleucine (160
mg total) was supplied to 2 x 1 L cultures on day 3, 6, and 8, and then both flasks were
harvested on day 10 (7.9 g wet wt., 0.63 g dry wt., 67 mg lipid extract). A total of 4.6
mg of labeled I was isolated. The '3C NMR spectrum (toluene-d8} showed a 2,970 %
enrichment in C-4 (in toluene-d8); C-i 90%, C-2 90%, C-3 90%, C-4 2,970%, C-5
70%, C-6 obscured, C-7 obscured, C-8 90%, C-9 80%, C-10
C-15 obscured, C-16
150%, C-17 80%, C-18 80%, 0-methyl obscured, N-methyl 110%. Based on the '3C
integrals for this sample, a T-statistic for C-4 was found equal to 125.45, giving a
>99.95% confidence that its 13C content lies outside of the integral values for the
natural abundance population.37
100
References
1.
Faulkner, D. J. Nat. Prod Rep. 2000, 17, 7-55, and previous articles in this series.
2. Butler, A.; Walker, J. V. Chem. Rev. 1993, 93, 1937-1944.
3. Orjala, J. 0.; Gerwick, W. H. J. Nat. Prod. 1996, 59, 427-430.
4. Hofheinz, von H.; Oberhansli, W. E. Helv. Chim. Acta 1977, 60, 660-669.
5. Dumdei, E. J.; Simpson, J. S.; Garson, M. J.; Byriel, K. A.; Kennard, C. H. L.
Aust. J. Chem. 1997, 50, 139-144.
6. (a) Flowers, A. E.; Garson, M. J.; Webb, R. I.; Dumdei, E. J.; Charan, R. D. Cell
Tissue Res. 1998, 292, 597-607. (b) Bewley, C. A.; Holland, N. D.; Faulkner, D. J.
Experientia 1996, 52, 716-722.
7. Pettit, G. R.; Kamano, Y.; Herald, C. L.; Tuinman, A. A.; Boettner, F. E.; Kizu,
H.; Schmidt, J. M.; Baczynskyj, L.; Tomer, K. B.; Botems, R. J. J. Am. Chem. Soc.
1987, 109, 6883-6895.
8. Harrigan, G. G.; Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Nagle, D. G.; Paul, V.
J.; Mooberry, S. L.; Corbett, T. H.; Valeriote, F. A. J. Nat. Prod. 1998, 61, 10751077.
9. Sitachitta, N.; Rossi, J.; Roberts, M. A.; Gerwick, W. H.; Fletcher, M. D.; Willis,
C. L. .1. Am. Chem. Soc. 1998, 120, 713 1-7132.
10. Williamson, R. T.; Sitachitta, N.; Gerwick, W. H. Tetrahedron Lett. 1999, 40,
5175-5 178.
11. We erroneously assigned the pro-S and pro-R methyl groups in unlabeled leucine.
Before chlorination, C-3 of leucine is the first priority substituent on the C-4 chiral
center; hence, the methyl group which becomes chlorinated is in the pro-R
position.
12. (a)Hartung, J. Angew. Chem., mt. Ed. 1999, 38, 1209-1211. (b)MacMillan, J. B.;
Molinski T. F. J. Nat. Prod. 2000, 63, 155-157.
13. Fu, X.; Zeng, L. -M.; Su, J. -Y.; Pais, M. J. Nat. Prod. 1993, 56, 637-642.
14. Shaw-Reid, C.A.; Kelleher, N.L.; Losey, H.C.; Gebring, A.M.; Berg, C.; Walsh,
C.T. Chem. Biol. 1999, 6, 385-400.
101
15. Hurd, R. E. J. Magn. Reson 1990, 87, 422-428.
16. (a) Palmer III, A. G.; Cavanagh, J.; Write, P. E.; Rance, M. J. Magn. Reson. 1991,
93, 151-170. (b) Kay, L. E.; Keifer, P.; Saarinen, T. J. Am. Chem. Soc. 1992, 114,
10663-10665. (c) Schleucher, J.; Schwendinger, M.; Sattler, M.; Schmidt, P.;
Schedletzky, 0.; Glaser, S. J.; Sorenson, 0. W.; Griesinger, C. I Biomol. NMR
1994, 4, 301-306.
17. (a) Bax, A.; Summers, F. J. Am. Chem. Soc. 1986, 108, 2093-2094. (b) Wilker,
W.; Leibfritz, D.; Kerssebaum, R.; Bennel, W. Mag. Res. Chem. 1993, 31, 287292.
18. Scott, K.; Keeler, J.; Van, Q. V.; Shaka, A. J. J. Magn. Reson. 1997, 125, 302-324.
19. McDonald, L. A.; Ireland, C. M. J. Nat. Prod. 1992, 55, 376-379.
20. Williamson, R. T. Isolation and Synthesis of Bioactive Marine Galactolipids. M.S.
Thesis, University of North Carolina at Wilmington, Wilmington, NC, 1996.
21. Marfey, P. Carlsberg Res. Commun. 1984, 49, 591-596.
22. Rossi, J. V.; Roberts, M. A.; Yoo, H. D.; Gerwick, W. H. J. App!. Phycol. 1997, 9,
195-204.
23. Kelly, N. M.; Reid, R. G.; Willis, C. L. Tetrahedron Lett. 1995, 36, 83 15-8318.
24. Kelly, N. M.; Sutherland, A.; Willis, C. L. Nat. Prod. Rep. 1996, 205-219.
25. Tn naming this isotope-labeled leucine, we erroneously assigned higher priority to
the 13C-containining methyl group than to the C-3 methylene group. However, reevaluation of the priority rules [Hanson, K. .1. Am. Chem. Soc. 1966, 88, 273 12742.] identifies C-3 as the priority group because atomic number is considered a
higher priority than is atomic mass. Atomic mass is considered only if the groups
are otherwise identical (in this case, the C-S and C-6 methyl groups).
26. Fletcher, M. D.; Harding, J. R.; Hughes, R. A.; Kelly, N. M.; Schmalz, H.;
Sutherland, A.; Willis, C. L. J. Chem. Soc., Perkin Trans. 1 2000, 43-51.
27. (a) Kazlauskas, R.; Ligard, R. 0.; Wells, R. J.; Vetter, W. Tetrahedron Lett. 1977,
3183-3186. (b) Kazlauskas, R.; Murphy, P. T.; Wells, R. J. Tetrahedron Lett.
1978,4945-4948. (c) Fu, X.; Zeng, L. -M.; Su, J. -Y.; Pais, M. .1 Nat. Prod. 1997,
60, 695-696. (d) Fu, X.; Ferriera, M. L.; Schmidtz, F. J.; Kelly-Borges, M. I Nat.
Prod. 1998, 61, 1226-1231. (e) Gebreyesus, T.; Yosief, T.; Carmely, S.; Kashman,
Y. Tetrahedron Lett. 1988,29, 3863-3864. (f) Hotheinz, W.; Oberhansli, W. E.
102
Helv. Chim. Acta. 1977, 60, 660-669. (g) Lee, G. M.; Molinski, 1. F. Tetrahedron
Lett. 1992, 33, 7671-7674.
28. Bender, D.A. Amino Acid Metabolism, John Wiley & Sons, London, 1975.
29. (a) See for example: Williard, P. G.; Laszlo, S. E. J. Org. Chem. 1984, 49, 34893493. (b) Helmchen, G.; Wegner, G. Tetrahedron Lett. 1985, 26, 6047-6050. (c)
Brantley, S. E.; Molinski, 1. F. Org. Lett. 1999, 13, 2165-2167.
30. This is within the reported 0.03-0.55 ppm range for '80-labeled carbonyl
compounds [Vederas, J.C. Nat. Prod. Rep. 1987, 277-337].
31. Carmeli, S.; Moore, R. E.; Patterson, G. M. L.; Yoshida, W. Y. Tetrahedron Lett.
1993, 34, 557 1-5574.
32. Seto, H.; Watanabe, H.; Furihata, K. Tetrahedron Lett. 1996, 37, 7979-7982.
33. Martin, G. E.; Crouch, R. S. J. Heterocyclic Chem. 1995, 32, 1665-1669.
34. Buckingham, J.; Donaghy, S. M.; Cadogan, J. I. G.; Raphael, R. A.; Rees, C. W.
Dictionary of Organic Compounds, Chapman and Hall, New York, 1982.
35. Shimojima, Y.; Hayashi, H.; Ooka, T.; Shibuawa, M. Tetrahedron 1984, 40, 25192527.
36. Huffman, W. A.; Ingersoll, A. W. I Am. Chem. Soc. 1951, 73, 3366-3369.
37. Ramsey, F. L.; Schafer, D. W. The Statistical Sleuth; A Course in Methods of Data
Analysis, Duxbury Press, Belmont, CA, 1997, p. 36-42.
103
CHAPTER V
THREE DIMENSIONAL SOLUTION STRUCTURES OF ANTILLATOXIN
AND THREE OF ITS STEREOISOMERS
Abstract
The absolute stereochemistry of antillatoxin has been revised based on a
reinvestigation of the physical data. In particular, the stereochemical revision
utilized newly acquired NOE data in conjunction with both vicinal homonuclear
coupling constants and CD spectral analysis to predict an absolute stereochemistry
of 2'S,4R,
5R, 5'S
for natural antillatoxin. This stereochemistry was subsequently
confirmed by total synthesis. In addition to natural 4R,5R antillatoxin, three
diastereomers about the C4-05 bond of antillatoxin were synthesized. Of the four
isomers, natural 4R,5R antillatoxin possessed potent neurotoxic activity; 4S,5S
antillatoxin displayed modest activity and the 4S,5R and 4R,5S isomers were
essentially inactive. To explore the relationship of these two stereocenters with the
observed variation in biological activity, the three dimensional solution structures
for all four compounds were determined using NIMR derived distance and torsion
angle constraints.
104
Comparison of the three dimensional structures of the four stereoisomers revealed
significant conformational differences, implying that conformation plays an
important role in the biological activity.
105
Introduction
Algae have proven to be prolific producers of structurally novel and
biologically active natural products. The ranges of biological activities are vast,
including the anti-cancer compound such as cryptophycin A
(1),1
and potent
neurotoxins such as the structurally daunting polyether maitotoxin
(2),2
immense
in both size and complexity. These neurotoxins can also take a much less complex
form, however. Examples of these structurally modest but significantly toxic
compounds include domoic (3) and kainic acids
(4),34
as well as the recently
described metabolite kalkitoxin (5, see chapter VI).5
First reported in 1995 by Orj ala et al., antillatoxin (6) was isolated as a
potent ichthyotoxin (LD50 = 0.5 g/mL) from the lipid extract of a Lyngbya
majuscula collected from the island of Curacao, in the southern Caribbean
Ocean.6'7
The structure of antillatoxin was assembled via standard spectroscopic
techniques. Biosynthetically, it could be envisioned that tert-butanoic acid serves
as the starter unit for polyketide synthase (PKS) extension by four acetate units
with subsequent methylation from the Cl pooi (or two propionate units and two
acetates), providing the initial 15 carbon acid. The exo-olefin could arise from C2
of acetate in a similar fashion as observed in the jamaicamide series of compounds
(see Chapter III). This polyketide portion of antillatoxin could then undergo nonnbosomal polypeptide synthetase (NRPS) extension by alanine, valine (which
undergoes N-methylation), and glycine, completing the linear structure. Ring
106
closure by nucleophilic attack of the C5 hydroxyl group on the glycine carbonyl
would lead to the final structure of antillatoxin.
In the initial 1995 report elucidating the structure of antillatoxm, the
stereochemistry of the two optically active amino acids was accomplished by chiral
TLC analysis, revealing an S stereochemistry at both a-carbons. The
stereochemistry of C4 and CS was characterized by analyzing a combination of
homonuclear dipolar and scalar coupling interactions, circular dichroism (CD)
spectroscopy, and molecular modeling.7 Interpretation of these spectral data
initially indicated that the absolute stereochemistry of antillatoxin was
2'S,4S,SR,5'S. However, spectral data obtained from isomer, as observed from total
synthesis, were not in agreement with that of the natural product.
Herein, the revised stereochemistry, as predicted through the use of newly
acquired NIvIR data and the original CD data is presented. This prediction has been
confirmed by NIv[R spectral comparison to the four C4-05 stereoisomers
synthesized by Drs. Shioiri and Yokokawa. Additionally, biological data is
presented showing that a vast difference in biological activity exists among these
four stereoisomers. Therefore, the three dimensional structures of all four
stereoisomers were determined by NMR-restrained molecular modeling in an effort
to gain insight
into
structure-activity relationship in this class of neurotoxin.
Herein, the results of these modeling studies are presented which demonstrate that
conformational variations exist between the four stereoisomers in solution and may
contribute to their differences in neurotoxicity.
107
id(3)
OH
L,_.4L0
HN'O
HOK
ocH3
OJ1J
H0H
Ciyptophycin A (1)
CH3
CH3 CH3
Kalkitoxin (5)
Kainic Acid (4)
CH3 CH3
0
H3C0
CH3
o
CH3 CH3 CHH
lfxCH:)<
H3C-N
CH3
0
H
H
Antillatoxin (6)
Results and Discussion
In the original report of antillatoxin, chiral TLC methods were used to
determine an S configuration for two of the four stereocenters in 6
(2'S,5'S).
The
remaining two stereocenters were assessed by NMR data, molecular modeling, and
analysis of the CD spectrum of i. Through, observance of a strong dipolar
coupling correlation between H4 and H5 in the NOESY spectrum, it was deduced
that these two protons were on the same face of the molecule, and thus a cis
relationship was assigned. This correlation, together with 22 additional NOE
correlations and a single torsion angle (determined by an 11 Hz homonuclear
coupling between H4 and 115) were used to restrain molecular modeling
calculations. Four different structure files were created (45,
4R,5R)
5R, 4S, 5S, 4R, 5S,
and molecular dynamics using a simulated annealing protocol were used for
the structure calculations. Based on the NItvfR derived distance restraints and
torsion angle restraint, the
4R, 5S
stereoisomer consistently violated the structure
parameters, however, the remaining isomers satisfied the NMR derived constraints.
The rms deviations for the three isomers (4S,5S,
4S,5R, 4R,5R)
were
2.5, 1.4,
and
1.9 angstroms, respectively.7 Furthermore, analysis of the CD spectrum indicated a
right handed helicity, deduced by a heterochromic exiton coupling at 227 nm.
Using all of these data, the assignment of 4S,5R was made.7
Subsequent to this original structure report, the total synthesis was
completed by two independent
groups.8'9
A 111 NMR comparison of the synthetic
109
material and natural antillatoxin revealed inconsistencies between the two
structures (Figure V.1). Careful examination of the two spectra revealed that the
'H resonances displaying the largest chemical shifts discrepancies were clustered
around the C4-05 region. The disagreement in 'H and '3C chemical shifts led to a
reevaluation of the spectral data.
0041
I
Figure V.1. Structure of natural antillatoxin with the predicted 4S, 5R
stereochemistry. Values in bold represent & values for 4S,5R antillatoxin '3C
NMIR chemical shift synthetic antillatoxin '3C NMR shift. Values in italics
indicate the &5 values for 4S,5R antillatoxin 'H NMR chemical shift - synthetic
antillatoxin 'H NMR shift.
Reevaluation of the Original Stereochemical Assignment. Having
determined that the difference between synthetic
4S,5R
antillatoxin and natural
antillatoxin was centered about the C4-05 stereocenters, a reevaluation of the
original NMR data was conducted. Analysis of the NOESY spectrum revealed that
this data was presented as the magnitude calculated spectrum, and thus, it was
determined that the NOE correlation between H4 and H5 was the result of a COSY
110
(scalar coupling) artifact. Therefore, new dipolar coupling data was acquired on
natural antillatoxin. A DPFGSE 1D NOE experiment was performed using
selective excitation of the resonance of interest, resulting in a spectrum where the
irradiated peak is inverted (depending on the molecular correlation time) with
respect to the protons that are dipolar coupled.'° The results of this experiment are
shown in Figure V.2. When H5 was selectively irradiated with a 60 msec Gaussian
8
7
H7
H12b
6
5
4
3
2
ppm
Figure V.2. DPFGSE 1D NOE spectrum of natural antillatoxin with selective
irradiation at H5. The arrow indicates where in the 'H spectrum that H4 resonates.
shaped pulse, only H7, H12b, H13, and H14 showed enhancement. No
enhancement of H4 was observed, indicating that these protons were on opposite
faces of the molecule. With this new information, the CD data could be analyzed
to predict the correct stereochemistry. Two of the possible rotamers about the C4C5 {4R5S, 4S5R (original assignment)] bond could be eliminated based on the trans
111
relationship determined by these new NOE data. Using the diene to olefin right
handed helicity indicated by the heterochromic coupling in the CD spectrum, the
4S,5S structure could also be eliminated. Therefore, the only remaining
stereoisomer, 4R, 5R, was predicted to be the correct stereoisomer (see Figure V.3).
14
13'
15
11
12' N
Hi
4R,5R
Antillatoxin
4S, 5R Antillatoxin
L10
"ss
4R,5S Antillatoxin
4S,5S
Antillatoxin
Figure V.3. Four possible stereoisomers about the C4-05 bond. Elimination of
three of these can be accomplished by analysis of the physical data (see text).
Coincident with this stereochemical prediction, Drs. Yokokawa and Shioiri of
Nagoya City University provided synthetic samples of all four possible
stereoisomers about the C4-05 bond. Comparison of the 'H NMR spectra of the
synthetic compounds and natural antillatoxin revealed that the 4R, 5R
stereochemistry was correct.
112
With the unique opportunity of possessing all four stereoisomers of
antillatoxin, a series of biological assays were performed in our laboratory and
Professor Tom Murray's laboratory at the University of Georgia.6 The data for
these assays are presented in Table V.1. The natural isomer (4R, 5R) of antillatoxin
was undoubtedly the most potent of the four compounds in all of the bioassays
performed.
Table V.1. Biological evaluation of antillatoxin stereoisomers for ichthyotoxicity
and neurotoxicity.
Stereoisomer
Ichthyotoxicitya
Microphysiometryb
4R,SR
Ca.O.1tM
4mM
42nM
0.18iM
4S,5S
slight
5-10 iM
1.4 p.M
5-10 jiM
4R,5S
inactive
5-10 p.M
8.1 jiM
Ca. 10 p.M
4S,5R
slight
inactive
inactive
Ca. 10 p.M
LDH
Assayb
Neuro 2a Assaya
3Assays performed at the College of Pharmacy, OSU by Dr. Tatsufumi Okino. bAssays
performed at the Department of Physiology and Pharmacology, College of Veterinary
Medicine, University of Georgia by Drs. Fred Berman and Tom Murray.
Based on the differences in the biological activities observed for these four
stereoisomers, a hypothesis was developed that conformational differences between
the stereoisomers was responsible. Having all four isomers in sufficient quantities
to allow for NMR analysis provided a unique opportunity to explore the
significance of the three dimensional structures on the observed differences in
toxicity.
AMI Semi-Empirical Molecular Modeling Calculations. In collaboration
with Dr. Philip S. Magee of the BioSAR Research Project, the semi-empirical
113
AM 1" minimum was determined for the
4R, 5R
antillatoxin (natural) isomer. The
structure is illustrated in Figure V.3 below.
Figure V.4. Spacefilling representation of the AM1 minimum for 4R,5R
antillatoxin.
Using HyperChem,'2 an estimated logP value of 3.98 was calculated by Dr.
Magee for antillatoxin. This value indicated that natural antillatoxin exists outside
of the systemic range, and suggests the possibility that antillatoxin was a CNS
toxin, exerting toxicity via some sort of binding and blocking action. Interestingly,
the mechanism of antillatoxin's neurotoxicity (Nat channel blocker) was not known
at the time of these calculations.
Molecular Modeling Calculations. The solution structure studies were
initiated with the assignment of the 'H and '3C NMR resonances in each of the four
isomers. This was accomplished via a combination of 'H NMR, E.COSY,'3 and
HSQMBC'4 experiments. Assignment of the diastereotopic protons at p
and C8' was done through simultaneous interpretation of the dipolar couplings and
the distance restrained model. In addition, the geminal protons at C12 (H12a and
114
H12b) were assigned based on a combination of ROE data and measurement of the
3JCH
couplings. Once all the 'H and 13C NMR resonances were assigned, analysis
of dipolar coupling interactions, homonuclear couplings, and vicinal heteronuclear
couplings ensued. T-ROESY was used to analyze the through space
interactions.'5
The T-ROESY experiment was chosen to reduce the possibility of any strong scalar
coupling artifacts that sometimes arise in a standard ROESY
experiment.'5
To
deduce the distance restraints, the volumes of all correlations in the T-ROESY
spectrum were integrated. These integration values were then normalized and
assigned values of strong (s), medium (m), and weak (w) as follows: s =2± 1 A,
m = 2.5 ± 1.5 A, and w =3.0±2
JHJI
A.
'H NMR and E.COSY were used to measure
couplings, and the HSQMBC experiment was used to measure
constants.
3JCH
coupling
The 4) angles of the alanine residue and the diastereotopic protons of
glycine were detenrnned through measurement of 3JNH..CLH. This coupling constant
was then used to calculate the bond angle using 3JNH..CIH = 6.7cos29 - 1.3cosO +
1.5.16 The 4) angle was also calculated by measuring the 3JHa-CO(i-1) coupling
constant and calculating a respective bond angle with the following Karplus
equation: 3Jco(..l) = 4.Ocos2O - 1.8cos9 + 0.l8.' To determine the x angle of the
N-Me valine residue the 3JHJI..HU and
3JH-CO coupling
constants were by measured in
the 1H NMR spectra. These values were then used to derive the bond angle values
by using 3JHHa
0.47,
6.7cos2O
respectively.18"9
1 .3cos9 + 1.5 and 3JHCO = 8.O6cos2O - 0.87cosO +
These data have been tabulated in Tables V.2-5.
115
Molecular modeling calculations were performed using the Macromodel 7.0
software package utilizing the MM2* forcefleld.20'2' The molecular modeling was
initiated by assembling the Dreiding model within the build menu of Macromodel
7.0. Once the structure was constructed, the stereochemistry was set at each of the
four stereocenters. To derive the correct starting bond geometries, a steepest
descent (SD) minimization with 100 iterations was performed using the MM2
force field. At this point, the distance constraints were entered into a constraint file
within the minimization menu. Once these distance constraints were in place, a
1000 iteration Polak-Ribier Conjugate Gradient (PRCG)22 minimization using the
MMf force field was performed. The distance constrained model was then
subjected to torsion angle constraints derived from
3JHH
homonuclear and 3JCH
heteronuclear coupling constants converted to their corresponding angles as
calculated with an appropriate Karplus equation. Once all constraints were in place
(distance and torsion angle) an additional PRCG minimization using the MM2*
forcefleld was done. To further probe the available conformations within this
minimum, a 1,000 step Monte Carlo search was performed using the MM2 force
field. Twenty of the lowest energy structures were overlaid for each of the four
isomers and the rms deviations calculated (excluding all hydrogen atoms). For
each antillatoxin isomer, the structures with the lowest energy minimum
(determined several times throughout the calculation) from the constrained
conformational searches are shown in Figures V.5b-8b. The data tables and
structure figures are delineated below for each of the four isomers.
116
4R,5R-Antillatoxin.
Table V.2 tabulates the NMR data used to generate the
torsion angle and distance constraints for the molecular modeling calculations. A
total of 22 distance constraints and 9 torsion angle constraints were used in the
Table V.2. NMR data for the 4R,5R-antillatoxin isomer.
Am
'H (ppm)
(Hz)
'J
'C (ppm)
HSQMBC (Hz)'
ROESYb
-
-
169.8
-
-
2a
2.73 (d)
12.2
44.6
1,3,4(7.0), 12 (5.5)
2b (s)
b
3.13 (d)
12.2
1,3,4(5.1), 12(6.1)
2a (s),4 (m), 13(m)
3
-
146.7
-
-
2, 3, 5,6(4.2), 12,13
14(m), 13(s)
3 (2.5), 4,6,7(4.3), 9'
2a (m), 4(w), 13 (m), 14(w)
4
2.22(m)
-
37.6
5
5.07 (d)
11.1
82.8
(<1), 13(1.2), 14(2.5)
6
-
-
129.7
-
7
5.90(s)
-
135.6
618,5(7.9),9(5.1),14
4 (w), 5 (s), 9 (s), 13 (m), 14(w),
(8.1), 15(3.0)
15(m)
8
-
-
130.0
-
9
5.28 (s)
-
140.7
7 (7.7), 8, 10, 11(3.7), 15,
13(w), 14(s), 15(w), 11,16,17
(9.7), 16(3.7), 17 (3.7)
(m)
10
-
-
32.1
11
1.11 (s)
-
30.5
9,10
12a
4.87(s)
-
111.7
2(11.4),4(4.1)
b
4.91 (s)
-
13
0.83 (d)
6.9
14
1.50(s)
15
5(m), 12b (s), 13(m)
2 (8.2),4 (10.8)
2b (m), 12a (s)
18.3
3,4,5
2b (m), 4 (s), 12a (w), 12b (m),
-
12.0
5,6,7
5(w), 9(s), 12' (m), 13 (w), 15
1.78(s)
-
17.2
13(w), 14 (w), 7(m) 11,16,17 (m)
16
1.11(s)
-
30.5
17
1.11 (s)
-
30.5
7,8,9
9,10
9,10
I'
9.24(d)
8.9
-
1,2', 11'
2(m), 2b (s), 4 (w), 11' (w)
2'
5.34(dq)
6.5,8.2
42.0
1 (<1),3', 11'
5'(m), 11'(s)
3'
-
-
172.6
-
14(w),
(w)
4'
-
-
-
-
5'
4.41 (d)
10.9
65.2
3'(< 1),6', 12'(5.4), 15'
2' (m), 13' (w), 12' (w), 14' (m),
(2.5), 13'
15' (m)
6'
-
-
167.6
-
7'
8.34(bd)
9.5
-
1,2
2(w), 5' (m), 8'a (m)
8'a
3.47 (dd)
1.5,18.3
40.2
6'/9'
4(w)S'b(s)
b
4.35 (dd)
9.6, 183
6'/9'
8'a(s)
-
-
10'
1c'7R
-
-
-
-
-
1.25 (d)
6.6
18.3
2', 3'
2' (s), 1' (w)
12'
2.70(s)
-
27.8
3',5',
13'(m), 14'(m), 15'(m)
13'
2.24(m)
-
25.5
6'(2.5)
14'
0.88 (d)
6.5
18.5
5', 13', 15'
14(m), 15'(m)
12(m), 13(m), 15'(m)
15'
0.80 (d)
6.7
17.8
5', 13', 14'
II' (w), 13' (m), 14' (m)
Ii
-
All data acquired in DMSO-d6 at 298 K. aCoupling constants measured using the
HSQMBC experiment, utilizing 'JCH values for peak fitting protocol (see Chapter VI).
bThe abbreviations are as follows: s = strong, in = medium, w = weak. The numeric values
assigned to these letters are described in the text.
117
structure calculation. An error of± 200 was tolerated for the torsion angle
parameters. The constrained model, using the MM2* force field, had a minimum
energy of 258.73 kJ/mol. The MC search generated a new minimum of 255.45 kJ,
which was found four times (Figure V.5b). The 20 lowest energy structures were
overlaid (Figure V.5 a) and an rms deviation was calculated using all non-hydrogen
atoms. Of the 20 structures used for the calculation, the energy difference between
the minimum and twentieth structure was 1.12 kJ/mol. The rms deviation for these
structures were 0.57
A.
Figure V.5. (a) Twenty overlaid structures taken from the Monte Carlo search of
the constrained energy minimized structure of 4R, 5R antillatoxin. (b) The lowest
energy conformation of 4R, 5R antillatoxin.
118
Table V.3. NMR data for the 4R,5S-antillatoxin isomer.
nr
'H (ppm)
3Jwi (Hz)
'3C (ppm)
HSQMBC (Hz)
ROESYb
-
-
169.4
-
-
2a
2.90(d)
16.1
40.6
1,3,4,12
2b(s), 1'(w),4 (w),
b
3.04(d)
16.1
1,3,4,12
2a(s), 12a(m)
1
5(w), 12a (w)
3
-
-
144.5
-
-
4
2.49 (obs)
-
42.5
2(6.0), 3,5,6(2.8),
4(s), 12b (m), 13 (m),
12(6.9), 13
14(m)
5
4.77 (bs)
-
81.0
3 (4.3), 4,6, 13 (3.7),
2a (w), 4(m), 7 (m),
14(1.7), 7(4.3), 9' (1.2)
13(w), 14(m)
6
-
-
130.4
-
-
7
5.61 (s)
-
130.4
5 (6.6), 6/8, 14 (8.6),
5 (m), 9(m), 8b (w),
9(6.0), 15(2.5)
13 (m), 15(m)
8
-
-
130.4
-
-
9
5.19 (s)
-
139.0
7(7.9), 10, 15, (10.4),
7 (m), 14(m),
11(4.6), 16(4.6), 17(4.6)
11,16,17(m)
10
-
-
32.1
-
-
11
1.11 (s)
30.5
9,10
15(m)
12a
4.94 (s)
-
117.3
1,2(7.0), 3,4(10.1)
2a (w), 2b (m), 12b (s)
b
5.01 (s)
-
13
0.90(d)
7.3
14
1.63 (s)
15
2(11.0),3,4(6.4)
4(s), 12a(s), 13(m)
13.9
3,4,5
7 (m), 12b (m), 4(m),
-
15.4
5,6/7
13 (w), 4 (m), 5 (m), 9 (w)
1.72 (s)
-
17.7
7/8,9, 10
7(w), 13(w), 11,16,
16
17
1.11(s)
-
30.5
1.11 (s)
-
30.5
9,10
9,10
17(m)
15(m)
15(m)
1'
8.32 (d)
8.2
-
2'
2a (w)
2'
4.65 (dq)
6.5, 8.2
42.8
1', 3'
5' (m), 11' (m) 14' (w)
-
-
5(w)
11,16,17 (w)
4'
5'
-
3.92 (d)
9.7
171
-
-
-
-
-
64.7
3' (2.1), 6', 12' (4.3), 13',
2' (m), 13' (w), 7(m),
15' (2.0)
14' (w), 15' (m)
6'
-
-
168.9
-
-
7'
7.96 (dd)
5.4, 7.3
-
-
5' (m), 8'a (s), 8'b (w)
8'a
3.68(dd)
4.9,16.7
41.4
9',6'
8'b(s), 13(m)
b
4.16(dd)
7.1,16.7
9',6'
8'a(s)
9'
-
-
167.7
-
-
Iv-
-
-
-
-
1.18(d)
6.5
17.5
2',3'
2'(s)
12'
2.65(s)
-
29.9
3',5'
5'(w), 13'(m), 14'(s)
13'
2.21 (m)
26.8
-
14'(m), 15'(m)
14'
0.76(d)
6.9
18.6
5', 13', 15'
5'(w), 12'(w), 13'(s),
15'
1.01 (d)
6.5
20.9
5', 13', 14'
13'(m), 14'(m)
11'
15' (m)
All data acquired in DMSO-d6 at 298 K. aco,1mg constants measured using the
HSQMBC experiment, utilizing 1JCH values for peak fitting protocol (see Chapter 6). bThe
abbreviations are as follows: s = strong, m = medium, w = weak. The numeric values
assigned to these letters are described in the text.
119
4R,SS-Antillatoxin. Table V.3 tabulates the NMR data used to generate
the torsion angle and distance constraints for the molecular modeling calculations.
A total of 16 distance constraints and 7 torsion angle constraints were used in the
structure calculation. An error of± 200 was tolerated for the torsion angle
parameters. The constrained model, using the MM2* force field, had a minimum
energy of 284.32 kJ/mol. The MC search generated a new minimum of 259.36 kJ,
which was found 16 times (Figure V.6b). The 20 lowest energy structures were
overlaid (Figure V.6a) and an rms deviation was calculated using all non-hydrogen
atoms. Of the 20 structures used for the calculation, the energy difference between
the minimum and structure twenty was 0.72 Id/mo!. The rms deviation for the
twenty isomers was 0.77 A.
Figure V.6. (a) Twenty overlaid structures taken from the Monte Carlo search of the
constrained energy minimized structure 4R, 5S antillatoxin. (b) The lowest energy
conformation of 4R, 5S antillatoxin.
120
Table V.4. NIvIR data for the 4S,5S-antillatoxin isomer.
Ai
!
(ppm)
3JHH (Hz)
'3C (ppm)
ROESYb
HSQMBC (Hz)'
-
-
169.8
-
-
2a
2.87(d)
14.6
41.9
1,3, 4(3.6), 12(5.4)
2b (s)
b
3.07 (d)
14.6
1,3,4 (4.9), 12 (4.2)
2a (s), 1' (m)
3
-
-
4
2.97 (dd)
5
5.02 (d)
146.7
-
-
7.0, 10.4
39.1
2 (3.0)c, 3,5,6, 13
14(m), 13 (w)
10.5
83.4
3 (2.2), 4,7 (4.7), 13(0.8),
7(s), 4(w), 12a (m), 14
14(2.4), 9' (1.5)
(w)
6
-
-
128.8
-
-
7
5.93 (s)
-
135.8
6/8,5(7.8), 14(7.5), 9
4(w), 5(s), 9(w)
8
-
-
130.2
-
-
9
5.29(s)
-
140.5
7(7.0), 10,11(4.3), 15,
7(w), 14(w)
10
-
-
32.0
-
-
11
1.12(s)
-
30.5
9,10
-
12a
4.84 (s)
-
111.1
b
4.80 (s)
-
13
0.82(d)
7.0
17.3
14
1.66(s)
-
15
1.79(s)
16
1.12 (s)
17
(6.0), 15 (3.0)
(9.6), 16 (4.3), 17 (4.3)
2(6.0), 3,4(9.1), 13
5(m), 12b (s)
2 (12.0), 3,4 (5.0), 13
2a (m), 2b (m), 12b (s)
12.2
3,4,5
5,6,7
-
17.4
7, 8, 9, 10
4(w)
4(m),9(w)
9(w)
-
30.5
9, 10
-
1.12(s)
-
30.5
9,10
-
1'
8.64(d)
8.5
-
-
2b (m), 2' (w)
2'
4.97 (dq)
6.5, 8.6
42.8
3'
1' (w), 11' (w), 5' (m)
3'
-
-
171.8
-
-
4'
-
-
-
-
5'
4.42 (d)
10.4
64.2
6'
-
-
168.2
-
-
7'
7.76 (t)
4.7
-
-
5' (m), 8'a (w), 81, (w)
8'a
3.98 (dd)
4.39, 17.9
42.2
6', 9'
8
b
3.73 (dd)
5.1, 17.8
6', 9'
8'a (s), 7' (w)
-
-
3' (2.9), 6', 12' (4.8), 14', 15'
2' (m), 7' (m), 14' (w), 15'
(2.3), 13'
(m)
(s), 7' (w)
9'
-
-
166.5
-
10'
-
-
-
-
-
11'
1.21(d)
6.5
17.8
2',3'
2'(w)
12'
2.65 (s)
-
28.7
3', 5'
13' (w), 14' (s)
13'
2.22(m)
-
26.1
-
14'(s),15'(m)
14'
0.78 (d)
6.9
18.1
5', 13', 15'
5'(w), 13'(s)
15'
0.94 (d)
6.5
19.2
5', 13' 14'
5' (m), 13' (m), 15' (m)
All data acquired in DMSO-d6 at 298 K. acoupling constants measured using the
HSQMBC experiment, utilizing 1JCH values for peak fitting protocol (see Chapter 6). bThe
abbreviations are as follows: s = strong, m = medium, w = weak. The numeric values
assigned to these letters are described in the text. cThis value is a good estimate of the
heteronuclear coupling constant. This value was not accurately obtained due to peak shape
complications.
121
4S,5S-Antillatoxin. Table V.4 tabulates the NMR data used to generate the
torsion angle and distance constraints for the molecular modeling calculations. A
total of 18 distance constraints and 12 torsion angle constraints were used in the
structure calculation. An error of ± 200 was tolerated for the torsion angle
parameters. The constrained model, using the MM2* force field, had a minimum
energy of 254.89 kJ/mol. The MC search generated a new minimum of 252.64 Id,
which was found 3 times (Figure V.7b). The 20 lowest energy structures were
overlaid (Figure V.7a) and an rms deviation was calculated using all non-hydrogen
atoms. Of the 20 structures used for the calculation, the energy difference between
the minimum and structure twenty was 0.89 kJ/mol. The rms deviation for the
twenty isomers was 0.49 A.
Figure V.7. (a) Twenty overlaid structures taken from the Monte Carlo search of the
constrained energy minimized structure 4S, 5S antillatoxin. (b) The lowest energy
conformation of 4S,5S antillatoxin.
122
Table V.5. NMR data for the 4S,5R-antillatoxin isomer.
'H (ppm)
3J,114 (Hz)
'3C (ppm)
HSQMBC (Hz)'
ROESYb
-
-
170.2
-
-
2a
2.86(d)
14.9
42.9
1,3,4(3.8), 12 (1.8)
2b (s), 5(w)
b
3.31 (obs)
-
1,3,4(1.0), 12(5.2)
2a(s),
3
-
-
143.9
4
2.64 (obs)
-
38.5
2,3,5,6, 12, 13
5(m), 14(m), 13(m)
5
5.09 (bs)
-
81.2
3 (4.4), 4,6,7(3.4), 9',
2a (w), 4(s), 13 (w),
13(3.2), 14(2.0)
14(m)
-
6
-
-
128.4
-
-
7
5.48 (s)
-
130.0
6,8,5(6.0), 9(5.0),
2b (w), 4 (w), 5(s), 9(s),
14(8.1), 15(2.6)
13(m), 15(m)
8
-
-
130.1
-
-
9
5.18 (s)
-
139.5
7/8,10, 11(3.8), 15, (9.2),
4(w), 7(s), 13 (w),14 (m),
16(3.8), 17(3.8)
15(w), 11,16,17(m)
10
-
-
31.9
-
-
11
1.09(s)
-
30.5
9,10
-
12a
4.98(s)
-
116.0
2(10.9),4(4.9)
4(w), 14(m), 13(m)
b
5.02 (s)
-
2 (6.6), 4 (9.4)
2a (s), 4(w), 14 (w)
13
0.86 (d)
7.4
12.7
3,4, 5
2b (w), 4 (m), 5 (w),
14
1.59 (s)
-
14.8
15
1.72 (s)
-
17.7
5,6,7
7,8,9
2b(w), 4(m), 5(w), 9(m)
4(w), 7(m), 9(w),
11,16,17(w)
16
1.01 (s)
-
30.5
9, 10
-
17
1.01(s)
-
30.5
9,10
-
1'
8.8
-
1,2', 11'
2' (w), 2b (m), 11' (w)
2'
8.56(d)
5.10(dq)
6.8,8.8
42.6
1 (3.1),3', 11'
1(w), 5'(m), 11'(m)
3'
-
-
172.0
-
-
-
-
-
-
-
5'
4.23 (d)
10.5
64.7
3' (3.1), 6', 12' (6.0), 13',
2' (m), 13' (w), 7' (w), 14'
14'(4.9), 15(4.5)
(w), 15' (w)
6'
-
-
168.3
-
-
7'
8.48 (dd)
3.0,8.1
-
1,2
2' (w), 5'(m), 8'a (m)
8'a
3.65 (dd)
3.1, 17.4
41.4
6'/9'
8'b (s)
b
4.24 (dd)
8.2, 17.6
6'/9'
8'a (s)
9'
-
-
167.7
-
-
lv
-
-
-
-
-
11'
1.20(d)
6.5
17.4
2', 3'
2' (m)
12'
2.65 (s)
-
28.7
3', 5',
13' (w)
13'
2.23 (m)
-
26.4
6' (3.6)
12 (m), 14' (s), 15' (s)
14'
0.75(d)
6.7
18.3
5', 13', 15'
5(w), 12' (m), 13'(s), 15'
15'
0.94(d)
6.5
19.2
5', 13', 14'
5' (m), 13' (s), 14' (m)
7 (w), 12b (w)
4'
(m)
All data acquired in DMSO-d6 at 298 K. aco,1thg constants measured using the
HSQMBC experiment, utilizing JCH values for peak fitting protocol (see Chapter 6). "The
abbreviations are as follows: s = strong, m = medium, w = weak. The numeric values
assigned to these letters are described in the text.
123
4S,5R-Antillatoxin. Table V.5 tabulates the NMR data used to generate the
torsion angle and distance constraints for the molecular modeling calculations. A
total of 29 distance constraints and 7 torsion angle constraints were used in the
structure calculation. An error of± 200 was tolerated for the torsion angle
parameters. The constrained model, using the
force field, had a minimum
energy of 359.69 kJ/mol. The MC search generated a new minimum of 346.25 kJ,
which was found 3 times (Figure V.8a). The 20 lowest energy structures were
overlaid (Figure V.8a) and an rms deviation was calculated using all non-hydrogen
atoms. Of the 20 structures used for the calculation, the energy difference between
the minimum and structure twenty was 5.23 kJ/mol. The rms deviation for the
twenty isomers was 0.64 A.
Figure V.8. (a) Twenty overlaid structures taken from the Monte Carlo search of the
constrained energy minimized structure 4S, 5R antillatoxin. (b) The lowest energy
conformation of 4S,5R antillatoxin.
124
Figure V.9. All models are displayed looking down the C4-05 bond axis. (a) The
lowest energy confonnation of 4R,5R antillatoxin. (b) The lowest energy
conformation of 4R,5S antillatoxin. (c) The lowest energy conformation of 4S,5S
antillatoxin. (d) The lowest energy conformation of 4S, 5R antillatoxin.
125
Conclusions. As observed in Figure V.9, the most active isomer (4R,5R),
appears as an "L" shape with a hydrophobic interior and a hydrophilic exterior
surface which contains most of the electronegative substituents. The calculated
model suggests that a hydrogen bond (2.0 A) possibly exists between the glycine
NH (H7') and the carboxyl terminus (Cl) of the polyketide chain (see introduction).
In comparing the NMR-constrained with the semi-empirically derived models the
same general shape is observed. However, the N-Me group is pointing away from
the aliphatic chain, whereas the NMR constrained model indicates that this group is
pointing towards the chain.
The
4S,5S compound
also has an "L" shaped structure. However, it is
inverted with respect to the natural isomer. The lipid tail is pointing into the plane
of the page in Figure V.6. Most of the electronegative substituents are facing the
interior of the molecule, in sharp contrast to the 4R,5R isomer. There's a potential
hydrogen bond of 2.0 A between HT and the carbonyl oxygen of Cl. In contrast,
the 4R,
5S
isomer has an extended conformation with no apparent hydrogen bond
interactions present. Interestingly, this isomer has a cis amide bond between the
alanine and N-Me valine residues. Likewise, the 4S,5R structure has an extended
structure, however, with a slight twist in the macrocycle. The observed twist in the
ring allows H7' and the carbonyl oxygen of Cl to come within 1.9 A of each other,
thereby indicating a potential hydrogen bond interaction.
The aim of this study was to compare the solution structures of the four
stereoisomers. While no definitive conclusions can be drawn as to what molecular
126
features of antillatoxin are responsible for the dramatic variation in biological
activities, the overall structures must play a role in these differences. This
conclusion is drawn because to the only differences between the molecules under
study are the stereocenters at positions C4 and C5. Comparison of the solution
confonnations shows large differences exist between the four stereoisomers.
Therefore, concluding that these differences in conformation play a definitive role
in the observed toxicity for antillatoxin and the three stereoisomers is reasonable.
Two potential caveats that are that 1) a minor conformation may be
responsible for the biological activity, and 2) organic solvents were used for these
NMR studies. Because these four structures (Figure V.9) represent a time-averaged
picture of the major conformation present in solution for each stereoisomer; it is
conceivable that a minor conformation is responsible for the observed biological
activity and is not represented by the models derived in this study. In recent work
by Nevins et
al.
it is shown that the presumption of a single or strongly preferred
conformation for a small molecule characterized by easily rotated bonds involves
risk and should be done with great
Snyder et
al.
caution.23
This point was exemplified by
in their work with the solution conformation of taxol.24 They found
that the major conformation observed in chloroform was inactive, and that a minor
conformer was responsible for the observed biological activity. While the
molecular modeling studies were not carried out in aqueous solution, the use of
DMSO for modeling cyclic peptides is common in the Iiterature.253° In addition, it
127
has been argued that the use of organic solvents mimics lipophilic sites in
biological systems (e.g.
membranes).3032
The results from this work will not be fully realized until a more detailed
understanding of the exact interaction between antillatoxin and the sodium channel
is determined. Once their interaction is further elaborated, these four solution
structures can be used to explore likely points of interaction which activate the
sodium channel. Furthermore, a more complete understanding could provide
knowledge of a new drug binding site on the sodium channel, and antillatoxin
could act as a molecular probe to better understand the pharmacology of this site.
These types of studies will also give insight into the role that these major
conformations (Figure V.9) have on the biological activity. Additionally,
determination of the structural entities of antillatoxin responsible for its biological
activity can be used to design chemical modifications of antillatoxin (or one of the
three isomers) to further elucidate structure-activity relationships in this chemical
class.
128
Experimental
NMR Measurements. All NIvIR data were recorded on a Bruker DRX
spectrometer operating at 600.03 MHZ. Spectra were acquired at 298 K in 99.96%
DMSO-d6. Two mg of each of the four stereoisomers were individually dissolved
in 0.4 mL DMSO-d6 (9.9 mM) and transferred to a Shegemi microcell matched to
deuterated DMSO. The spectrometer was equipped with a 5 mm Bruker Q-Switch
TXI probe. The 'H and '3C 90° pulse widths, at a 0 db power level (-6 db
maximum), were 9.42 p.sec and 14.2 tsec, respectively. For the DPFGSE 1D NOE
experiment, a 60 msec Gaussian shaped pulse at 69 db was used for selective
irradiation. The experiment was acquired with 1028 scans with 32 k data points
and processed with 3.0 Hz line broadening with zero-filling to 64 k. The
HSQMBC experiment was optimized for an 8 Hz (31.2 msec) long range
heteronuclear coupling. A total of 188 scans per 256 increments were acquired
with 2 k data points in F1. The data was processed to 4 k data points in F2 and
linear predicted to 512 followed by zero-filling to 1 kin the F, dimension.
The
spin-lock pulse in the TROESY experiment was set at 400 msec. A 90o shifted
cosine function was applied to both dimensions. The data matrix was zero-filled to
1024 x 1024 data points.
Molecular Modeling. All calculations were either performed on an SGI
1NDY running LRJX 6.3, or a NEC 266 MHz Pentium running SuSE Linux (2.2.14
kernel). The software used in the calculations was Macromodel 7.0. The MM2
129
force field was used for all minimizations and the Monte Carlo conformation
search. The steepest descent (SD) minimization method was used for obtaining the
correct starting geometries of the bonds. Distance restraints were then applied to
the model from within the minimization menu of the Macromodel software. These
distances were applied based on the determination of whether the correlation was
strong, medium, or weak, and the following numerical values were assigned: s =2
± 1 A, m = 2.5 ± 1.5 A, andw
3.0 ± 2
A.
The determination ofa "strong" ROE
correlation was determined by the integration of the dipolar coupling between a
pair of geminal protons. This constrained structure was then energy minimized
using the PRCG method with 1,000 iterations. Torsion angle restraints were added
based on the bond angles derived from applying the
3JCH
coupling constants to the
appropriate Karplus equation. An error of± 200 was tolerated for these restraints.
This structure was then minimized using a 1,000 iteration PRCG method. The
resulting structure was then subjected to a Monte Carlo conformation search to
more thoroughly explore the conformational space allowed by the distance and
torsion angle restraints. A 1,000 step conformation search was done. Of the
structures generated, the twenty lowest energy structures were overlaid and an rmsd
was calculated. The new minimum generated during the search was found multiple
times for each isomer.
130
References
1.
Smith, C. D.; Zhang, X.; Moobeny, S. L.; Patterson, G. M.; Moore, R. E.
Cancer Res. 1994, 14, 3779-3784.
2. (a) Murata, M.; Iwashita, T.; Yokoyama, A.; Sasaki, M.; Yasumoto, 1. J. Am.
Chem. Soc. 1992, 114, 6594-6596. (b) Murata, M.; Naoki, H.; Iwashita, T.;
Matsunaga, S.; Sasaki, M.; Yokoyama, A.; Yasumoto, T. J. Am. Chem. Soc.
1993, 115, 2060-2062.
3. Wright, J. L. C.; Boyd, R. K.; de Freitas, A. S. W.; Falk, M.; Foxall, R. A.;
Jamieson, W. D.; Laycock, M. V.; McColloch, A. W.; Mclnnes, A.G. Can. J.
Chem. 1989, 67, 481-490.
4. Nitta, I.; Watase, H. Tomiie, Y. Nature 1958, 181, 761-762.
5. Wit, M.; Okino, T.; Nogle, L. M.; Marquez, B. L.; Williamson, R. T.; Sitachitta,
N.; Berman, F. W.; Murray, T. F.; McGough, K.; Jacobs, R.; Colsen, K.;
Asano, T.; Yokokawa, F.; Shioiri, T.; Gerwick, W. H. I Am. Chem. Soc. 2000,
122, 12041-12042.
6. Berman, F. W.; Murray, T. F.; Gerwick, W. H. Toxicon 1999, 37, 1645-1648.
7. Orjala, J.; Nagle, D. G.; Hsu, V. L.; Gerwick, W. H. J. Am. Chem. Soc. 1995,
117, 8281-8282.
8. Yokokawa, F.; Shioiri, T. J. Org. Chem. 1998, 63, 8638-8639.
9. White, J. D.; Hanselmann, it.; Wardrop, D. J. J. Am. Chem. Soc. 1999, 121,
1106-1107.
10. Stott, K.; Stonehouse, J.; Keeler, J.; Hwang, T.-L.; Shaka, A. J. J. Am. Chem.
Soc. 1995, 117, 4199-4200.
11. Dewar, M. J. S.; Zoebisch, E. V.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem.
Soc. 1985, 107, 3902-3909.
12. HyperChem, Hypercube Inc. Gainesville, FL.
13. Wiliker, W.; Liebfritz, D.; Kerssebaum, R.; Lobman, J. J. Magn. Reson. Ser. A
1993, 102, 348-350.
131
14. Williamson, R. T.; Marquez, B. L.; Kover, K. E.; Gerwick, W. H. Magn.
Reson. Chem. 2000, 38, 265-273.
15. (a) Hwang, T.-L.; Shaka, A. J. J. Am. Chem. Soc. 1992, 114, 3157-3159. (b)
Hwang, T.-L.; Shaka, A. J. .1. Magn. Reson. Ser. B 1993, 102, 155-165.
16. Ludvigsen, S.; Andersen, K. V.; Poulsen, F. M. J. Mol. Biol. 1991,217, 731736.
17. Kao. L.-F.; Barfield, M.; J. Am. Chem. Soc. 1985, 107, 2323-2330.
18. Haasnoot, C. A. G.; DeLeeuw, F. A. A. M.; Altona, C. Tetrahedron 1980, 36,
2783-2792.
19. Aydin, R.; Gunther, H. Magn. Reson. Chem. 1990,28, 448-457.
20. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton, M.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. I Comput. Chem. 1990,
11,440-462.
21. Allinger, N. L. J. Am. Chem. Soc. 1977, 99, 8127-8134.
22. Polak, E.; Ribiere, G. Revenue Francaise Informat. Recherche Operationelle,
1969, v. 35.
23. Nevins, N.; Cicero, D.; Snyder, J. P. .1. Org. Chem. 1999, 64, 3979-3986.
24. Snyder, J. P.; Nevins, N.; Cicero, D. 0.; Jansen, J. J. Am. Chem. Soc. 2000,
122, 724-725.
25. Inman, W.; Crews, P. J. Am. Chem. Soc. 1989, 111, 2822-2829.
26. Morita, H.; Yoshida, N.; Takeya, K.; Itokawa, H.; Shirota, 0. Tetrahedron
1996, 52, 2795-2802.
27. Fujikawa, A.; Yasuko, I.; Inoue, M.; Ishida, T.; Nemoto, N.; Kobayashi, Y.;
Kataoka, R.; Ikai, K.; Takesako, K.; Kato, I. J. Org. Chem. 1994, 59, 570-578.
28. Ishida, T.; Ohishi, H.; Inoue, M.; Kamigauchi, M.; Sugiura, M.; Takao, N.;
Kato, S, Hamada, Y.; Shioiri, T. I Org. Chem. 1989, 54, 5337-5343.
29. Morita, H.; Kayashita, T.; Takeya, K.; Itokawa, H.; Shiro, M. Tetrahedron
1997, 53, 1607-1616.
132
30. Kock, M.; Kessler, H.; Seebach, D.; Thaler, A. J. Am. Chem. Soc. 1992, 114,
2676-2686.
31. Seebach, D.; Ko, S-Y.; Kessler, H.; Kock, M.; Reggelin, M.; Schmieder, P.;
Walkinshaw, M. D.; Boelsterli, J. J.; Bevek, D. Helv. Chim. Acta. 1991, 74,
1953-1990.
32. Behrens, S.; Matha, B.; Bitan, G.; Gilon, C.; Kessler, H. mt. i Pept. Protein.
Res. 1996,48, 569-579.
133
CHAPTER VI
THE HSQMBC EXPERIMENTS AND THEIR APPLICATION TO THE
STEREOCHEMISTRY OF NATURAL PRODUCTS
Abstract
Two experiments are presented that facilitate the measurement of long-range
heteronuclear coupling constants. These experiments, the HSQMBC and the GBIRD j,x-HSQMBC, are based on the evolution of single quantum coherence (SQC).
Utilization of SQC alleviates complications that arise from the evolution of
homonuclear couplings as in the HMBC-type experiments. Utilizing the G-BIRDHSQMBC and the E.COSY the relative stereochemistry of the marine natural product
kalkitoxin is presented. The relative stereochemistry of this neurotoxic natural
product predicted by this analysis was confirmed by total synthesis.
134
Introduction
The use of long-range heteronuclear coupling information to solve structure
problems has been occasionally reported in the literature.' The infrequent use of these
couplings is almost certainly due to the inherent problems associated with their
measurement. These difficulties arise from the fact that the coupling constants of
interest are usually quite small, 1-10 Hz (same magnitude as 3JHH coupling constants),
and are associated with low sensitivity nuclei such as
'3C
or 'SN. In order to address
these problems, there have been several recent reports of new experiments designed to
alleviate the complications associated with the measurement of these valuable
Jc11
(n
=2,3) couplings. These experiments are the sensitivity-improved hetero-(co,)-halffiltered TOCSY (HETLOC),2 sensitivity-improved HSQC-HECADE,3 HSQCTOCSY,4 GSQMBC,5 J-Resolved HMBC-2,6 J-IMPEACH-MBC,7 and phase-
sensitive HMBC (psHMBC). These experiments, as well as the two described in this
chapter, the HSQMBC and the G-BIRD-HSQMBC, have recently been reviewed.'
Because the magnitude of 3JCH coupling constants follows a Karplus-type
relationship similar to that which defines
3JHH
coupling constants, the use of long-
range heteronuclear coupling constants can greatly enhance the computer-aided
determination of three-dimensional solution
structures.810
To date, the NMR-derived
molecular constraints used in the modeling of natural products have been
predominantly NOE distance restraints and geometrical constraints derived from
homonuclear 3J-j coupling constants. 8,12,13 While a reasonably accurate model of
small molecule solution structures can usually be obtained using homonuclear dipolar
135
and scalar coupling information alone, a more accurate and precise model can be
realized with the additional torsion angle restraints provided by long-range
heteronuclear coupling constants.1319 This concept has been demonstrated in the
three-dimensional structure determination of cyclosporin A
(2),1422
(1),2021
okadaic acid
strictosidine (3) and related vincoside alkaloids,23 and a cyanobactenal
metabolite known as antillatoxin (4, Chapter V).
While the determination of the three-dimensional structure of a natural product
is important for studying its possible chemical reactivity and insight into its possible
biological role, there are also other uses for long-range heteronuclear coupling
information. For example, situations arise where there are no adjacent (spin-spin or
dipolar coupled) protons to gain insight into structural and conformational
ambiguities. For example, the geometry of scytonemin (5) could not be determined
except through the use of heteronuclear coupling information.25 The stereochemistry
about this double bond was readily determined to be E using 3JCH coupling constants
determined from an HSQMBC
experiment.2527
Heteronuclear coupling constants are also invaluable to solving a wide variety
of problems in organic and inorganic chemistry.27 For example, a recent paper
reported a method for the determination of relative stereochemistry in acyclic systems
from NMR analysis alone. Matsumori et al., described what is termed the "J-based
136
Okadaic Acid (2)
Cyclosporin A (1)
Strictosidine (3)
Antillatoxin (4)
17
12
23
Scytonemin (5)
Strychnine (6)
Figure V1.1. Structures of cyclosporin A (1), okadaic acid (2), strictosidine (3),
antillatoxin (4), scytonemin (5), and strychnine (6).
137
configuration analysis", a method which concisely articulates the use of long-range
heteronuclear coupling constants in the configurational analysis of organic
compounds.28
Their method utilizes a combination °fJcH (n =2, 3) and 3JHH
coupling constants to establish the relative stereochemistry between any two
asymmetric centers so long as each carbon center between them is a chiral methine
center or contains resolvable diastereotopic methylene protons. Recently, there have
been several reports in the literature utilizing this f-coupling approach to solve several
very complex stereochemical prob1ems.'4'29
Herein, the application of the "J-based configuration analysis" will be
presented on the marine natural product kalkitoxin (7). This compound, first isolated
in 1995 by Mm Wu in the Gerwick laboratory, is a potent neurotoxin isolated from
Lyngbya majuscula.30'35
The compound contains five stereocenters, of which three are
very well suited for this type of configuration analysis. The highly flexible nature of
kallcitoxin did not allow the use of dipolar coupling information (Figure VI.3),
therefore the long-range heteronuclear, and vicinal homonuclear couplings were used
for the analysis. The analysis was carried out using the E.COSY and G-BIRDR,XHSQMBC experiments.
S CH3
CH3 CH3
NjCH3
CH3 CH3
14
0
15
Kalkitoxrn (7)
Figure VI.2. The structure of kalkitoxin showing the absolute stereochemistry.
138
ppm
0
0
:
:.
r
1
7.
I
I
2
3
0
5
5
7
8
7
6
5
4
3
2
ppm
2.
Figure VI.3. The 2-dimensional NOESY (800 ms) spectrum of -300
jtg
of kalkitoxin.
139
Results and Discussion
The HSQMBC experiment (Figure VI.4) was designed to overcome the
deleterious effects of homonuclear coupling through the use of the HSQC pulse
sequence as the fundamental building block. The initial INEPT transfer allows one to
place SQC magnetization onto heteronuclei. This approach effectively removes
homonuclear coupling evolution during t,. A high-powered trim pulse is used (prior
to transfer of magnetization to the X nuclei) to help dephase any unwanted
magnetization. A 'H 180° pulse in the middle of ti effectively refocuses heteronuclear
chemical shift and heteronuclear coupling evolution. After t1, a phase-encoding
gradient is applied, followed by a '3C 1800 pulse to refocus chemical shift and
coupling evolution during the initial selection gradient. The experiment then
incorporates a gradient z,z-filter3637 to destroy any unwanted magnetization
(dispersive contributions to the lineshape) before coherence transfer back to 'H for
detection with a decoding gradient. The lack of a low-pass filter allows the retention
of the
1JCH
couplings. In cases of well resolved correlations, the
1JCH
couplings can be
used to facilitate the measurement of the coupling constants of interest (Figure VI.5).38
However, if moderate overlap is observed these couplings can severely complicate
analysis. To circumvent this complication, a new experiment was developed to
alleviate these
1JCH
couplings.
In order to increase sensitivity and efficiently remove
'JCH
correlations from
the spectra, a G-BIRDR,X filter was incorporated into the initial INEPT transfer (Figure
VI.6). This operator is very efficient at removing
1JCH
couplings from the spectrum.
140
The filtering efficiency of this GBIRDR,X is very good for a range of 'JCH'S
130-190 Hz). Therefore, if the filter (delay) is tuned for
1JCH=
(< 10%
for
145 Hz, good
suppression can be obtained for both aliphatic and aromatic regions. In addition, by
virtue of the phase cycle, effective decoupling of remote protons is achieved during
the INEPT preparation thereby increasing the amplitude and providing a more uniform
phase during the generally inefficient iNEPT magnetization transfer (Figure VI.5).
Removal of the remote homonuclear couplings from the data by this experiment
generally allows one to accurately measure coupling constants directly from an F2
slice through the G-BWDp,x-HSQMBC correlation of interest (Figure VI.7).
Utilization of the G-BIRDR,x-HSQMBC allowed the measurement of the
'3JCH
coupling constants needed to assign the relative stereochemistry of kalkitoxin.
Y
4R
Figure VI.4. The HSQMBC experiment; thin and thick bars represent 90° and 180°
pulses respectively; the gray bar represents a 2.0 msec high powered trim pulse; t = 1/
(2JxH);t/2=l/(4JxJ);I =0,2; 4)2=O,O,2,2;4)3=O,O,O,O,2,2,2,2;4)R=O,2,O,
2, 2,0,2, 0; gradient values for the HSQMBC are Gl: G2: G3 = 80: 10: ±20 (gradient
values are expressed as a percentage of a maximum value of 72 Gcm' and were 1 ms
in duration).
141
a)
±JLLJJIIILIWJJL
ppm
ppm
."
...*.
40
s'a
60
-
.
,
.:
.a
80 -
g
48
+
.:i. .'.,ç ;
H15aC13
a.
a
1.45
ppm
c)
100
120
1.40
,,
I
.-
1
I..
ppm
..
140
78
160
HllbC12
a
4
3
ppm
2
2.65
3.8 Hz
ppm
7.6Hz
e)
J
31
\
\
\
/
k) ______,.A
g)
1.45
1.40
1.35 ppm
2.65
(J V\
2.60
ppm
Figure VI.5. (a) The 2-dimensional HSQMBC spectrum of 350 mM strychnine in 500
jtL CDC13; (b) expansion of the 2D HSQMBC spectrum showing the H15a to C13
correlation; (c) expansion of the 2D HSQMBC spectrum showing the Hi lb to C12
correlation; (d) slice through the H15a 'JCH correlation from the HSQMBC spectrum;
(e) spectrum d that is 180° phase shifted and horizontally shifted by the 3JCH coupling
constant; (f) the summation of spectra d and e; (g) slice through the H15a-C13
correlation taken from the HSQMBC spectrum; (h) slice through the Hi lb 1JCH
correlation from the HSQMBC spectrum; (i) spectrum h that is 180° phase shifted and
horizontally shifted by the 2JCH coupling constant; (j) the summation of spectra h and
i; (k) slice through the Hi lb-C12 correlation taken from the HSQMBC spectrum.
The data were acquired in approximately 1.5 hours with 4 K data points in F2, 512
increments with 8 scans per increment, and F2 x F1 spectral window of 6,000 x 27,000
with the carrier frequency set at 2,700 and 12,800 respectively. The gradient ratios
were set to (Gi: G2: G3) 8: 1: ±2. The delay t/2 is set to 1/4J. The trim pulse was set
to 2 msec duration.
142
Figure VI.6. The G-BIRDR-HSQMBC; thin and thick bars represent 90° and 1800
pulses respectively; the gray bar represents a 2.0 msec high powered trim pulse;
t= lI(2Jxii); t12= lI(4'Jxji); 4 =0,2; 2=0,0,2,2;43=0,0, 0,0,2,2,2,2; 4R=
0,2,0,2,2,0,2,0; gradient values for the G-BIRDR-HSQMBC are(G1: G2: G3: G4:
G5) 2.5: -2.5: 8: 1: ±2 (gradient values are expressed as a percentage of a maximum
value of 72 Gcm1 and were 1 ms in duration).
Kalkitoxin possesses a number of potent biological activities. Kalkitoxin (7)
was strongly ichthyotoxic to the common goldfish (Carassius auratus, LC50 700 nM),
potently brine shrimp toxic (Artemia sauna, LC50 170 nM), and potently inhibited cell
division in a fertilized sea urchin embryo assay (1050
Ca.
25 nM).3° in a primary cell
culture of rat neurons, natural kailcitoxin displayed an exceptional level of
neurotoxicity (LC50 3.86 nM), and its effects were inhibitable with NMDA receptor
antagonists.30'39
Additionally, kailcitoxin is highly active in an inflammatory disease
model which measures IL-lB-induced sPLA2 secretion from HepG2 cells (IC50 27
nM)30'4°
Finally, preliminary evidence suggests that kalkitoxin is an exquisitely
potent blocker of the voltage sensitive Na channel in mouse neuro-2a cells (EC50 of 7
= I nM; EC50 of saxitoxin = 8
This unique biological activity coupled with
its structurally simple constitution make kalkitoxin an appealing lead compound for
143
JL11LJr
a)
ppm
1
40
60
b)
ppm
.
d
48
H
-..
-I
.
.
HI 5aC1 3
80
1.40
1.45
ppm
c)
ppm
120
140
78{
160
m
H11b-C12
8
7
6
5
4
3
2
ppm
2.65
ppm
7.3 Hz
3.9 Hz
d)
1.45
1.40
1.35
ppm
2.70
2.65
2.60
2.55 ppm
Figure VI.7. (a) The 2-dimensional G-BIRDR-HSQMBC spectrum of 350 mM
strychnine in 500 i.tL CDC13; (b) expansion of the 2D G-BIRDR-HSQMBC spectrum
showing the H15a to C13 correlation; (c) expansion of the 2D HSQMBC spectrum
showing the Hi lb to Cl2 correlation; (d) slice through the Hl5a-Cl3 correlation from
a G-BIRDR-HSQMBC spectrum strychnine, showing the measurement of the 3JCH
coupling constant; (e) slice through the Hi lb-C12 correlation from a G-BIRDRHSQMBC spectrum of strychnine, showing the measurement of the JCH coupling
constant. The data were acquired in approximately 1.5 hours with 4 K data points in
F2, 512 increments with 8 scans per increment, and F2 x F1 spectral window of 6,000 x
27,000 with the carrier frequency set at 2,700 and 12,800 respectfully. Once acquired
the data were processed with 4 K data points zero-filled to 8K in F2 and 512 data
points zero-filled to I K and linear predicted to 768 data points in F1. A cosine
window function was applied to both dimensions before Fourier transformation with I
Hz and 0.3 Hz line broadening in F2 and F1 respectively. The gradient ratios for the
G-BIRDR-HSQMBC are (Gi: G2: G3: (34: G5)2.5: -2.5: 8: 1: ±2. The delay t/2 is set
to 1/4J. The trim pulse was set to 2 msec duration.
144
drug development. However, to pursue these endeavors, the stereochemistry needed
to be determined.
The absolute stereochemistry of C3 of 7 was readily deduced by Dr. Tatsufumi
Okino.3°
Kalkitoxin was subjected to ozonolysis and acid hydrolysis to yield cysteic
acid. The ozonized hydrolysate and both R and S cysteic acid standards were then
derivatized with Marfey's reagent. Analysis of these derivatives by RPHPLC defined
C3 of 7 as R. The limited amount of remaining kalkitoxin (- 300 .tg of 7 remained
because of chemical instability) precluded determination of the C2' stereochemistry.
However, the remaining stereocenters at positions C7, C8, and ClO were arranged in
such a manner as to allow exploration by the J-based configuration analysis.
To measure the 2'3JCH coupling constants needed to define the relative
stereochemistry of kalkitoxin several obstacles needed to be overcome. First, the
amount of available kalkitoxin was only 300 p.g. Second, the similarity of chemical
shifts in the area of interest and the complexity of the correlations precluded a
straightforward analysis. To overcome the limited sample size of natural kalkitoxin
all data used in this analysis were recorded on a Bruker® 500 MHz DRX spectrometer
equipped with a Bruker® 5 mm TXI CryoProbe (benzene-d6). Utilization of this probe
allowed the acquisition of heteronuclear correlation data in reasonable times with
more than sufficient signal to noise (S/N). In the initial analysis, the HSQMBC
experiment (figure VI.4) was used to measure these long-range heteronuclear
couplings constants, however, residual
1JCH
couplings severely complicated the
correlations that were needed. To overcome this inherent complexity, the G-BIRDR,XHSQMBC' was used for the measurement of these coupling constants.3°
145
The
JCH
values were measured by the G-BIRDR,x-HSQMBC (Figure VL8),'
and the 3JHH values were determined utilizing the E.COSY experiment (Figure VI.9).42
Of the six possible rotamers possible for each of the three stereocenters, each
possesses a single rotamer that is consistent with the homo- and heteronuclear scalar
coupling constants (Figures VI.7-9). Not all possible long-range heteronuclear
correlations could be measured due to overlap of the C13 and C14 methyl groups, and
because in some instances, no correlation was observed. The reason for the lack of
correlations is presumably due to the very small
2'3JCH
coupling constants between
these atoms. This rational is supported by the expected "small" coupling constants
predicted through the J-based configuration analysis (Figures VI.1O-12).
The relative stereochemistry at Cl was suggested by observation of a small
(1.3 Hz gauche) 3J
between H7-H8, a large (6.1 Hz, anti) 3JCH between H8-C6, and
a small (< 1 Hz, gauche)
3JCH
between H7-C9 (see Figure VI.13). Stereochemistry at
positions C8 and ClO was related through the intervening C9 diastereotopic methylene
protons. The low field proton at C9
(H9a) showed a large (8.2 Hz, anti) 3JHH to H8
whereas the high field proton (H9b) showed a small (4.4 Hz, gauche) JHH to H8.
Additionally, small (< 1 Hz) 3JCH were observed from
119a
and H9b to C7, and a large
(8.2 Hz, anti) 3JCH from H9b to Cl4. The relative stereochemistry at C9 and ClO was
determined by a large (9.4 Hz, anti)
gauche) 3JHH for
3JHH
between H9b and H1O and a small (3.1Hz,
H9aH1 0. Finally, a large (7.4 Hz, anti)
3JCH was
measured for
C15. In summary, these data strongly supported a 7R*, 8S', 1OS* relative
H9a-
146
ppm
20
:
40
.
60
-
-
100
. .4
120
0-
140
p.
6
4
3
2
1
ppm
Figure VI.8. The 2-dimensional G-BIRDR,x-HSQMBC spectrum of 300 j.tg of
kalkitoxm.
147
ppm
1
2
t
::"
I
3
4
a,
5
I
I
I
6
6
5
4
3
2
Figure VL9 The 2-dimensional E.COSY spectrum of -'300
1
tg
ppm
of kalkitoxin.
148
H7
C14
H7
C9
C9
C13
C6
C7
C6
H8Vt\. C14
C13
C6
C7
H8
3i
1-17
H8
C7
C14
C13
C9
JH7-H8
LG
ISMJ
SM
3
JH7-C9
SM
ISMJ
LG
3
JH7-C14
SM
JHS-C6
SM
JH8-C13
SM
3
3
SM
SM
ILG!
LG
H7
H7
H7
C91t\. C14
C14../±\ H8
C7
C7
C6
C13
H8
C6
C9
C7
C13
C9
H8
C6
C13
C14
3JH7H8
LG
SM
SM
3JH7.c9
SM
LG
SM
3JH7.C14
SM
SM
LG
3JH8.C6
SM
LG
SM
3JH8-C13
SM
SM
LG
Figure VI. 10. Six possible rotainers for the J-based configuration analysis of the C7C8 positions of kalkitoxm (7). Italicized values indicate matches of the experimental
values with expected values. Boxed values indicate the correct rotamer. Circled
values were not present in the G-BIRDR,x-HSQMEC spectrum.
149
H9b
H9a
ClO
H91
ClO
H8
H8
C14
ClO
H9a
C8
C8
C8
C14
C7
C7
Cl
H8
C14
H91
H9a
JH8-J-19a
I LG I
SM
SM
3JH8..H9b
ISM I
LG
SM
3.JH9a-C7
I SM I
LG
SM
SM
LG
JH9a-C14
3JH967
I SM I
SM
LG
3JH914
I LG I
SM
SM
SM
LG
3JH8.CIO
Cl
Cl
ClO
H9a
ClO
119b
H8
C14
H9a
C8
C8
C8
C14
C7
H91,
118
H8
C14
ClO
H9b
3JH8H
LG
SM
SM
3JHS-H9b
SM
SM
LG
JH9a-C7
SM
LG
SM
JH9a-CI4
SM
SM
LG
3JH967
LG
SM
SM
3JH9b-C14
SM
LG
SM
3JH8-Clo
SM
LG
SM
Figure VI. 11. Six possible rotamers for the J-based configuration analysis of the C8C9 positions of kalkitoxin (7). Italicized values indicate matches of the experimental
values with expected values. Boxed values indicate the correct rotamer. Circled
values were not present in the G-BIRDR,x-HSQMBC spectrum.
150
H9b
(
CII
HIO
C15
H9a
C9 )
HIO".4/'CI I
C8
)(L
H9a
H9a
( C9 )
ci i-..l-J cis
CI5'4" HIO
C8
C8
3JH9a-HIO
SM
SM
LG
3JH9b..HIO
LG
SM
SM
3Jj-i9a-CI1
LG
SM
SM
JH9a-Cl5
SM
LG
SM
3JH9lI
SM
LG
SM
3JH9I5
SM
SM
LG
3JH10..C8
SM
LG
SM
HIO
CII
H9a,(H9b
H9aH9b
CI5.4./CII
HIO"4"C15
(c9)
C8
C9)
H9b
C9 )
C15
H9a)CçH9b
C9)
CI1+-"HIO
C8
C8
3JH9a-HIO
SM
ISM I
LG
3JH9H1O
SM
ILG I
SM
3JH9a-cII
LG
3JH-CI5
SM
3JH9CII
SM
LG
3JH91,-CI5
LG
SM
JHIO-C8
LG
SM
SM
ILG I
SM
Figure VI.12. Six possible rotamers for the J-based configuration analysis of the C9ClO positions of kalkitoxin (7). Italicized values indicate matches of the experimental
values with expected values. Boxed values indicate the correct rotamer. Circled
values were not present in the G-BIRDR,x-HSQMBC spectrum.
151
stereochemistry for 1 (Figure VI.13). In combination with the above determined 3R
absolute stereochemistry, the total number of stereochemical possibilities was reduced
to four (3R,7R,8S,
3JcH<1HZ
1 0S,2'R, 3R,7S,8R, 1 OR,2'R, 3R,7R,8S, 1 OS,2'S, 3R,7S,8R, 1 OR,2'S).
JHK1.3Hz
3
JCH7.4HZ
3J<1I-Iz
(___
iø H9 .7t. H91
(18
C7.
4C6 '-.,_.' C13
HH = 4.4 Hz
('H9a
H91*
3HH =82 Hz t
C14
do
\K
C14
3JcH=6.1 Hz
3JcH=8.21-lz
3HH = 3.1
O
C8
i
3JHH=9.41-Iz
Figure VI.13. Representation of rotamers about C7, C8, C9 and ClO with depiction of
all heteronuclear and homonuclear couplings that were used to define the relative
stereochemistry at C7, C8 and ClO using the J-based configuration approach.
To determine the absolute stereochemistry of natural kalkitoxin, kalkitoxins
having all possible configurations were synthesized by Drs. Shioiri and Yokokawa;
(3R,7R,8S,1OS,2'R)-kalkitoxin was found to be identical with the natural substance
73O43
Comparison of '3C NMR chemical shifts between four synthesized
diastereoisomers and natural kalkitoxin showed very small differences of less than
0.2
ppm (Figure VI.14). However, both the 3S,7S,8R,1OR,2'S and 3R,7R,8S,1OS,2'R
isomers showed maximal '3C NMR differences of 0.026 ppm. The CD spectrum of
the
3S,7S,8R,1OR,2'S
isomer was of equal intensity but opposite sign to natural
kalkitoxin. Correspondingly, the CD of the 3R,7R,8S,1OS,2'R isomer was essentially
identical to natural compound
7 (Figure
VI. 15). Hence, by a combination of J-based
configuration, chemical degradation and Marfey's analysis, total synthesis, and chiral
152
optical measurements, natural kalkitoxin was deduced to have
3R,7R,8S,1OS,2'R
absolute stereochemistry.
3S,7R,8S,1OS,2'S
3R,7R,8S,IOS,2'S
0.1
0
ppm
I
-0.1
I
-.021
-0.2
3R,7S,SR, 1OR,2'S
3R,7R,8S, IOS,2'R
0.1
ppm
--
-
0
0.1
0
-0-I
-.0.2
1
6
-02
16
11
1
1'
6
11
16
1'
5'
Carbon number
5'
Carbon number
Figure VI.14. Differences in '3C NMR shifts between natural kalkitoxin (1) and four
synthetic kallutoxm stereolsomers.
8
6
4
2
LE 0
-2
-4
-6
Natural Kalkitoxin
3S,7S,8R,IOR,2'S
3R,7R,8S,IOS,2'R
*
N
N
*
N
0
N
*
*
N
0
0
? (nm)
Figure VI. 15. CD spectrum of natural kalkitoxin and both (+)- and (-)-synthetic
kalkitoxin (MeOH).
153
Experimental
General. All experiments utilizing strychnine were performed on a Bruker
DRX600 NMR instrument with a Bruker Q-switch 5 mm triple resonance probe with
shielded triple axis gradients. The 90° pulse widths were 8.9 p.s for 'H and 14.4 p.s for
l3
at power levels of 0 db (-6 db max). The gradient percentages given in Figure
VI.5-7 correspond to a maximum strength of 72 G/cm. Experiments on kalkitoxin
were performed on a Bruker DRX500 NMR instrument equipped with a Bruker 5mm
TXI CryoProbe by Dr. Kimberly Colson of Bruker Instruments. The 90° pulse widths
were 6.9 p.s for 'H and 13 p.s for 13C at power levels of 3 db and -2 db (-6 db max),
respectively.
Data Collected on Strychnine. Both the HSQMBC and the G-BIRDpHSQMBC were collected using a 353 mM sample of strychnine in 500 p.L of CDC13.
For the HSQMBC, the data were acquired in approximately 1.5 hours with 4 K data
points in F2, 512 increments with 8 scans per increment, and F2
x F,
spectral window
of 6,000 Hz x 27,000 Hz with the carrier frequency set at 2,700 Hz and 12,800 Hz
respectively. Once acquired the data were processed with 4 K data points zero-filled
to 8K in F2 and 512 data points zero-filled to 1 K and linear predicted to 768 data
points in F1. A cosine window function was applied to both dimensions before
Fourier transformation with 1 Hz and 0.3 Hz line broadening in F2 and F, respectfully.
The gradient ratios were set to (Gi: G2: G3) 8: 1: ±2. The delay t/2 is set to 1/4J.
The trim pulse was set to 2 msec duration. The G-BIRDR,x-HSQMBC data were
acquired in approximately 1.5 hours with 4K data points in F2, 512 increments with 8
154
scans per increment, and F2
x F1
spectral window of 6,000 Hz x 27,000 Hz with the
carrier frequency set at 2,700 Hz and 12,800 Hz respectfully. Once acquired the data
were processed with 4 K data points zero-filled to 8K in F2 and 512 data points zerofilled to 1 K and linear predicted to 768 data points in F1. A cosine window function
was applied to both dimensions before Fourier transformation with 1 Hz and 0.3 Hz
line broadening in F2 and F1 respectively. The gradient ratios for the G-BIRDR,x-
HSQMBC are(G1: G2: G3: G4: G5)2.5: -2.5:8: 1: ±2. The delay 'r/2 is set to 1/4J.
The trim pulse was set to 2 msec duration.
Data Collected on Kalkitoxin. All data acquired for the f-based
configuration analysis were performed on a - 300 j.tg sample of kalkitoxin in 350 .tL
of benzene-d6 placed in a Shegemi tube paramagnetically matched to CDC13. The data
for the G-BIRDR,x-HSQMBC were acquired with 2 K data points in F2, 256 increments
with 600 scans per increment, and F2 x F1 spectral window of 5,500 x 25,000 with the carrier
frequency set at 1,900 and 12,575 respectively. Once acquired the data were processed with 2
K data points zero-filled to 4K in F2 and 256 data points zero-filled to 1 K and linear predicted
to 384 data points in F1. A shifted cosine window function was applied in F1 and a sine
window function in
F2,
with 1 Hz and 0.3 Hz line broadening in F2 and
F1
respectively. The
data for the E.COSY were acquired with 4 K data points in F2, 512 increments with 63 scans
per increment, and F2
x F1
spectral window of 4,800 x 4,800 with the carrier frequency set at
2,400 in both dimensions. Once acquired the data were processed with 4 K data points in F2
and 512 data points zero-filled to 1 K and linear predicted to 1 K data points in F1. A shifted
sine window function was applied in F1 and
and F, respectively
F2,
with 0.3 Hz and
1 Hz
line broadening in F2
155
References
1.
Marquez, B. L.; Gerwick, W. H.; Williamson, R. T. Magn. Reson. Chem.
Accepted.
2. Uhrin, D.; Batta, G.; Hruby, V. J.; Barlow, P. N.; Kover, K. E. J. Magn. Reson.
1998, 130, 155-161.
3. Kozminski, W.; Nanz, D. J. Magn. Reson. 2000, 142, 294-299.
4. Kover, K. E.; Hruby, V. J.; Uhrin, D. .1. Magn. Reson. 1997, 129, 125-129.
5. Marek, R.; Kralik, L.; Skienar, V. Tetrahedron Lett. 1997, 38, 665-668.
6. Furihata, K.; Seto, H. Tetrahedron Lett. 1999,40, 6271-6275.
7. Williamson, R. T.; Marquez, B. L.; Gerwick, W. H.; Martin, G. E.;
Krishnamurthy, V. V. Magn. Reson. Chem. 2001, 39, 127-132.
8. Eberstadt, M.; Mierke, D. F.; Kock, M.; Kessler, H. He/v. Chim. Acta 1992, 75,
2583-2592.
9. Williamson, R. T.; Boulanger, A.; Vulpanovici, A.; Roberts, M. A.; Gerwick, W.
H. in preparation.
10. Marquez, B. L.; Gerwick, W. H. (SMASH) Small Molecule NMR Conference: Oral
Presentation; Chicago, IL 2000.
11. Clore, G. M.; Gronenborn, A. M.; Marius, C. G. Curr. Opin. Chem. Biol. 1998, 2,
564-570.
12. Matsumori, N.; Kaneno, D.; Murata, M.; Nakamura, H.; Tachibana, K. J. Org.
Chem. 1999, 64, 866-876.
13. Kock, M.; Junker, J. J. Org. Chem. 1997, 62, 8614-8615.
14. Murata, M.; Yasumoto, T. Nat. Prod. Rep. 2000, 17, 293-3 14.
15. Kobayashi, M.; Aoki, S.; Kitagawa, I.; Kobayashi, Y.; Nemoto, N.; Fujikawa, A.
Symposium papers 0f38th Symposium on the Chemistry of Natural Products,
Organizing Committee of the Symposium: Sendi, 1996; 61-66.
16. Kobayashi, M.; Aoki, S.; Kitigawa, I. Tetrahedron Lett. 1994, 35, 1243-1246.
156
17. Fujita, K.; Fujiwara, M.; Yamasaki, C.; Matsuura, T.; Furihata, K.; Seto, H.
Symposium papers of 38!' Symposium on the Chemistry of Natural Products,
Organizing Committee of the Symposium: Sendi, 1996; 379-384.
18. Hayakawa, Y.; Kim, J-W.; Adachi, H.; Shin-ya, K.; Fujita, K.; Seto, H. J. Am.
Chem. Soc. 1998, 120, 3524-3525.
19. Falk, M.; Spierenburg, P. F.; Walter, J. A. J. Comp. Chem. 1996, 17, 409-417.
20. Kock, M.; Junker, J. Bioorganic Chemistry; Diederichsen U, Ed.; Wiley-VCH
Verlag GmbH: Weinheim, Germany, 1999; 365-378.
21. Kock, M.; Kessler, H.; Seebach, D.; Thaler, A. J. Am. Chem. Soc. 1992, 114,
2676-2686.
22. Seebach, D.; Ko, S-Y.; Kessler, H.; Kock, M.; Reggelm, M.; Schmieder, P.;
Walkinshaw, M. D.; Boelsterli, J. J.; Bevek, D. Helv. Chim. Acta. 1991, 74, 19531990.
23. Matsumori, N.; Murata, M.; Tachibana, K.; Tetrahedron 1995, 51, 12229-12238.
24. (a) Patthy-Lukats, A.; Karolyhazy, L.; Szabo, L. F.; Podanyi, B. J. Nat. Prod.
1997, 60, 69-75 (b) Patthy-Lukats, A.; Kocsis, A.; Szabo, L.; Podanyi, B. J. Nat.
Prod. 1999, 62, 1492-1499.
25. Proteau, P. J.; Gerwick, W. H.; Garcia-Pichel, F.; Castenholz, R. Experientia 1993,
49, 825-829.
26. Marquez, B. L.; Williamson, R. T.; Gerwick, W. H. Experimental Nuclear
Magnetic Resonance Conftrence: Poster number 021; Orlando, 1999.
27. Williamson, R. T.; Marquez, B. L.; Gerwick, W. H. (SMASH) Small Molecule
NMR Conference: Oral Presentation; Chicago, IL 2000.
28. Eberstadt, M.; Gemmecker, G.; Mierke, D. F.; Kessler, H. Angew. Chem. mt. Ed.
Engl. 1995, 34, 167 1-1695.
29. Matsumori, M.; Kaneno, D.; Murata, M.; Nakamura, H.; Tachibana, K. J. Org.
Chem. 1999, 64, 866-876.
30. Wu, M.; Okino, T.; Nogle, L. M.; Marquez, B. L.; Williamson, R. T.; Sitachitta,
N.; Berman, F. W.; Murray, T. F.; McGough, K.; Jacobs, R.; Colsen, K.; Asano,
T.; Yokokawa, F.; Shioiri, T.; Gerwick, W. H. J. Am. Chem. Soc. 2000, 122,
12041-12042.
157
31. Murata, M.; Matsuoka, S.; Matsumori, N.; Paul, G. K.; Tachibana, K. J. Am.
Chem. Soc. 1999, 121, 870-871.
32. Rundlof, T.; Kjellberg, A.; Damberg, C.; Nishida, T.; Widmaim, G. Magn. Reson.
Chem. 1998, 36, 839-847.
33. Tvaroska, I.; Taravel, F. R. Adv. Carb. Chem. Biochem. 1995, 51, 15-61.
34. Milligan, K. E.; Marquez, B. L.; Williamson, R. T.; Davies-Coleman, M.;
Gerwick, W. H. .1. Nat. Prod. 2000, 63, 965-968.
35. Wu, M. M.S. Thesis, Oregon State University, 1997
36. Bruchwiler, D.; Wagner, G. J. Magn. Reson. 1986, 69, 546-551.
37. John, B. K.; Plant, D.; Hurd, R. E. J. Magn. Reson. 1993, 101, 113-117.
38. Sheng, S.; van Halbeek, H. I Magn. Reson. 1998, 130, 296-299.
39. Berman, F.W.; Gerwick, W.H.; Murray, T.F. Toxicon 1999, 37, 1645-1648.
40. Tan, L.T.; Williamson, R.T.; Watts, K.S.; Gerwick, W.H.; McGougli, K.; Jacobs,
R. J. Org. Chem. 2000, 65, 419-425.
41. Manger, R.L.; Leja, L.S.; Lee, S.Y.; Hungerford, J.M.; Hokama, Y.; Dickey,
R.W.; Granade, H.R.; Lewis, R.; Yasumoto, T.; Wekell, M.M. J. AOAC Intern.
1995, 78, 521-527.
42. Griesinger, C; Sørenson, O.W.; Ernst, R.R. J. Am. Chem. Soc. 1985, 107, 63946396.
43. Pure kalkitoxin showed the following: [aID25 = +16° (c = 0.07, CHC13); CD c
0.022, EtOH XX 226 nm (AE +4.75), 207.8 (0.0); IR (CHC13) 2961, 2928, 2880,
1643, 1464, 1086, 1410, 1380 cm': UV (MeOH)
250 nm ( = 2600): HR
ElMS (70 eV) m/z obs. [Mf 366.2696 (15.9, 0.9 nimu dev. for C21H38N20S); 1H
NMR (berizene-d6, 500 MHz) 65.85 (IH, ddd, 1=17.2, 10.3, 6.1 Hz), 5.24 (1H,
ddd, 1=17.2, 1.6, 1.6 Hz), 5.01 (lH, d, J=10.3 Hz), 4.75 (1H, dd, 1=7.8, 7.5 Hz),
3.35 (2H, m), 2.94 (1H, dd, J=10.5, 8.8 Hz), 2.72 (1H, dd, 1=10.7, 8.4 Hz), 2.55
(1H, m), 2.43 (3H, s), 2.31 (1H, m), 2.28 (1H, m), 2.05 (1H, m), 1.87 (1H, m),
1.54 (1H, m), 1.39 (1H, m), 1.38 (1H, m), 1.34 (1H, m), 1.24 (1H, m), 1.10 (1H,
m), 1.10 (3H, d,J=6.7 Hz), 1.02 (IH, m), 0.95 (d, 3H,J6.8 Hz), 0.88 (3H, d,
.1=7.5 Hz), 0.85 (3H, d, J=6. 1 Hz), 0.76 (3H, d, J=6.8 Hz); '3C NMR (DMSO-d6,
100 MHz) 6174.79 (C-i'), 169.18 (C-5), 137.96 (C-2), 115.24 (C-i), 77.91 (C-3),
44.88 (C-12), 39.4 (C-9), 37.85 (C-4), 37.49 (C-6), 36.68 (C-7), 36.20 (C-2'),
34.92 (C-li), 34.70 (C-16), 33.40 (C-8), 27.47 (C-b), 26.64 (C-3'), 19.07 (C-iS),
158
17.13 (C-5'), 16.0 (C-13), 16.0 (C-14), 11.69 (C-4'); '3C NMR (benzene-d6, 125
MHz, from HSQC and HSQMBC data sets) 5175.5 (C-i'), 170.2 (C-5), 138.3 (C2), 115.3 (C-i), 79.2 (C-3), 46.0 (C-12), 40.3 (C-9), 38.9 (C-4), 38.6 (C-6), 37.6
(C-2'), 37.5 (C-7), 36.0 (C-i 1), 34.5 (C-16), 34.4 (C-8), 28.3 (C-b), 27.8 (C-3'),
19.5 (C-iS), 17.8 (C-5'), 16.4 (C-i3), 16.4 (C-14), 12.4 (C-4').
159
CHAPTER VII
ACCORDIAN OPTIMIZED 1,1-ADEQUATE
Abstract
A new approach for the acquisition of 'H-detected INADEQUATE-type data is
presented. This method uses a decremented variable delay for the evolution of 1.Jcc
couplings. It is demonstrated that this variation of the 1,1 ADEQUATE experiment
provides superior performance when compared to the normal static-optimized 1,1-
ADEQUATE experiment. This improvement is manifested in both the number and
quality of correlations present in the processed data. The technique is validated using
the model compound ethyl trans-crotonate and its utility is demonstrated utilizing the
structurally unique marine natural product jamaicamide A.
160
Introduction
One of the most powerful experiments for small molecule structure elucidation
is the well-known 13C-13C INADEQUATE (incredible natural abundance double
quantum transfer experiment).1 Unfortunately, this experiment, which displays
information on the connectivity of adjacent 13C atoms, is also one of the most
insensitive NMR experiments available for small molecule structural analysis. The
reason for the low sensitivity associated with the INADEQUATE experiments arises
from the low natural abundance of molecules with intact '3C-'3C bonds (1O). In
order to circumvent this problem, a number of experiments based on the detection of
'3C-13C coupling information through the more sensitive 1H nucleus have been
reported (inverse detection). These experiments include the original INEPTiNADEQUATE experiment proposed by Otting,2 the series of ADEQUATE (adequate
sensitivity double-quantum spectroscopy) experiments proposed by
Griesenger,3
and a
number of experiments based on the evolution of '3C-13C multiple quantum
coherence.4
With the ADEQUATE experiment, large enhancement gains are
demonstrated with the use of an accordion-type delay for the evolution of '3C-'3C
couplings within a refocused 1,1-ADEQUATE experiment. Through collaboration
with Dr. R. Thomas Williamson at Wyeth-Ayerst Research a modified ADEQUATE
pulse sequence has been developed and given the acronym ACCORD-ADEQUATE
(Figure VII.!).
161
Results and Discussion
A number of long-range
JCH
heteronuclear correlation experiments have been
reported that utilize the ACCORDION principle to optimize for a wide range of 'JCH
coupling constants.5 These pulse sequences include the original ACCORD-HMBC,6'7
IMPEACH-MBC,8 CIGAR-HMBC,9 and the 2J,3J-HMBC.1° These experiments have
been shown to provide a higher overall number ofJcH correlations, albeit with a
sacrifice in intensity of the strongest correlations relative to data obtained from a
static-optimized experiment (e.g. }[MBC). One inherent problem with these
ACCORDION optimized experiments is that they generally increase the overall length
of an HMBC-type pulse sequence to 300-500 msec. In these cases,
T2
relaxation
effects can seriously reduce sensitivity. This caveat is especially important in
molecules with a molecular weight over approximately 400-500 amu.
03
Figure VII. 1. The pulse sequence for the ACCORD-ADEQUATE; thin and thick bars
represent 90° and 180° pulses respectively; the hashed bar represents a 120° pulse; A =
l/4(1JcH); c = l/4('Jmjn) and is decremented according to [l/((4*hJccmj
1 /4CJccmin) and is decremented according to [1/((4* 1Jccmin
'Jccmax)/ni)];
1fccmax)Ini) + t1jncrt]; t
1/2(1Jcc); 4i
= x, -x; 4)2 = x, x, x, x, -x, -x, -x, -x; 4) x, x, x,
-y' -y; 4)R = x, -x, -x, x, x, x, x, -x -x, x, x, -x, x, -x, -x, x; the gradient ratios were set to (Gi: G2: G3) 78.4:
x, x, x, x, x, -x, -x, -x, -x, -x, -x, -x, -x; 4)4 x, x, -x, -x; 4)s y' y'
77.4: -59 (gradient values are expressed as a percentage of a maximum value of 56
Gcm1)
162
The reason for these lengthy pulse sequences is that the corresponding range of 1/2
JCH
delays can range from 50 msec (10 Hz) to 250 msec (2 Hz). In addition, all of
these experiments utilize a refocusing period to prevent the evolution of heteronuclear
JCH
couplings in the F1 dimension. Fortunately, the one-bond 13C-'3C couplings are
larger in magnitude (30-75 Hz) than
JCH
couplings and require a proportionally
smaller range of optimization. As a result, the use of accordion-optimization in the
1,1-ADEQUATE or 1NEPT-]NADEQUATE experiments generally adds little time to
the overall length of the experiment. This feature makes the inverse-detected
INADEQUATE experiments very amenable to accordion-optimization of 'fcc
couplings.
CI
O
CH
0
27
0
4
2
O'6
24
22/'
ri
19
2
20
Figure VH.2. The structures of ethyl trans-crotonate (1) and jamaicamide A (2). The
crucial connectivity correlation between C18 and C19 is indicated by arrows which
indicate the initial magnetization transfer via 1JCH then the evolution of the '.Jcc
coupling. This correlation was absent in several statically optimized GHMBC
experiments and the static-optimized 1,1 ADEQUATE. However, this connectivity
was readily observed in the ACCORD-ADEQUATE experiment (see text).
A 1,1 ADEQUATE experiment is normally optimized for a compromise fcc
value of approximately 40-50 Hz for adjacent 13C atoms. However, with a 1/2 'Jc
value optimized for 50 Hz (10 msec), '3C-'3C couplings may be missed for aliphatic
ppm
40
60
80
100
120
H4-C2
140
H2-C3
'H4-C3
160 -
H2-C1
P7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0 2.5 2.0
ppm
7.0
6.5
6,0
5.5
5.0
4.5
4.0
3.5
3.0
2.5
2.0
ppm
Figure Vll.3. Two-dimensional plots of the (a) 1,1 ADEQUATE and the (b) ACCORD-ADEQUATE utilizing ethyl trans-crotonate as a
model compound. Both experiments were performed on a 75 ji.L sample of ethyl trans-crotonate in deuterated DMSO (525 p.L) at 340 K. (a)
The data were acquired with 64 scans per 180 F, increments. (b) The data were acquired with 64 scans per increment for a total of 180 F,
increments (90 echo and 90 anti-echo increments). Spectral sweep widths were 22.6 kHz in F, and 5 kHz in F2. The data from both
experiments were processed with an exponential weighting function in F, (10 Hz) and in F2 (5 Hz). The F, dimension was linear predicted to
360 data points and zero filled to 1K data points. The F2 dimemsion was acquired and processed with 2K data points. As seen for the H2-C2,
and H3-C3 correlations, strong coupling artifacts appear as HSQC correlations and should be analyzed with caution.'5 The correlation for the
114-C2 correlation is observed by taking a lower cut into the two dimensional spectrum.
164
systems with smaller magnitudes (30-35 Hz) or for olefinic and aromatic systems of
larger magnitude (65-75 Hz). At the very least, signals originating from these smaller
or larger than average couplings will suffer from diminished sensitivity. By extending
the range of optimization from 30-70 Hz, one retains the overall average length of the
pulse sequence relative to the previously reported 1,1 -ADEQUATE experiment. For
example, even for the smallest coupling constant of interest (- 30 Hz), the largest
increment is extended by a mere 2-3 msec.
To validate the ACCORD-ADEQUATE experiment, a model compound was
needed that contained a broad range of 'Jc coupling constants. The compound ethyl
trans-crotonate (Figure VII.2, 1) was chosen based on its wide range of2Jcc coupling
constants (38 Hz to 75
Hz).'2"3
The 1,1-ADEQUATE and ACCORD-ADEQUATE
data matrices are shown in Figure VII.3. The data for both experiments were acquired
utilizing the same number of scans per F, increment for direct comparison. The
ACCORD-ADEQUATE spectrum shows all possible correlations with a substantial
increase in signal to noise (S/N) when compared to the 1,1-ADEQUATE data that
shows marginal S/N and no correlations from H3 to C2 and C3 (Figure VII.4).
165
H23-C24
H3-C4,
H4-C3
H5-C4
H4-05
ppm
H8-C7
H15-C16
-.
iH2
17-C8
H9-C8' H8-C9
H1O-C9
40
H121-C13
H2423
H2223
60
80
H3-C2
100
120
H12-C1F
140
H5-C6
° H23-C22
H22-C22
H21-C22
160
H21-C20
7
6
5
4
3
2
1
ppm
Figure Vll.4. Two-dimensional plot of the 1,1 ADEQUATE. The experiment
was performed on a 58 mM sample ofjamaicamide A in CDC13 at 298 K. The
data were acquired with 256 scans per 180 F1 increments. Spectral sweep
widths were 22.6 kHz in F1 and 5 kHz in F2. The data was processed with an
exponential weighting function in F2 (10 Hz) and a ir/2 shifted sine bell in F1
(20 Hz). The F1 dimension was linear predicted to 360 data points and zero
filled to 1K data points. All responses for jamaicamide A (2) are labeled
according to the numbering scheme shown in Figure VII.2. Pulsed field
gradients and phase cycling are used to suppress protons bound to 12C and
protons bound to a single 'SC, however, incomplete suppression can be
observed in the form of axial peaks in the data matrix.
166
H3-C4,
H23-C24
H5-C4
6.
H11C12
114-05
H13-C1
H15-C16
H10-C9
..
H16-C15
ppm
H9C26,n4..c3
t
30
aH8-C7
°
911
H7-C8
H9-C8' H8-C9
40
H121-ç13
50
1124-C23
H22-C23
60
H23-C23
70
80
'H3-C2
90
100
110
H27-C27
120
H23-C21
1122-C21
H12-C11
H10-C1I
[
130
0 H9-C1O
H7-C6
H5-C6
hi-dO
H27-C6
140
iso
6 H23-C22
H22-C22
E
H21-C22
H18-C19
' H2i-C20
° H13 -C12
H16-C17
H18-C17
7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0
3.5
L'6°
170
3.0 2.5 2.0 1.5 1.0 0.5 ppm
Figure Vll.5. Two-dimensional plot of the ACCORD-ADEQUATE. The
experiment was performed on a 58 mM sample ofjamaicamide A in CDC13 at
298 K. The data were acquired with 256 scans per 180 F1 increments.
Spectral sweep widths were 22.6 kHz in F1 and 5 kElz in F2. The data was
processed with an exponential weighting function in F2 (10 Hz) and a it/2
shifted sine bell in F1 (20 Hz). The F1 dimension was linear predicted to 360
data points and zero filled to 1K data points. All responses for jamaicamide A
(2) are labeled according to the numbering scheme shown in Figure Vll.2.
Pulsed field gradients and phase cycling are used to suppress protons bound to
'2C and protons bound to a single 13C, however, incomplete suppression can be
observed in the form of axial peaks in the data matrix.
167
As an additional example of the benefits associated with using this method, we
show a comparison of a static-optimized refocused 1,1-ADEQUATE experiment and
an ACCORD-ADEQUATE experiment acquired on the marine cyanobacterial (blue-
green algae) metabolitejamaicamide A (Figure VII.2, 2). This structurally intriguing
compound was isolated from a Jamaican collection of the blue-green alga, Lyngbya
majuscula (Figure Vll.2).'1 The vast majority of the planar structure ofjamaicamide
A was easily assembled with the normal battery of 2D NMR experiments utilized for
small molecule structure elucidation, including HSQC, HSQC-COSY, GHMBC, 1H-
'5N GHMBC, and DPFGSE 1D NOE (see Chapter HI). However, evidence
supporting the connectivity between C-18 and C-19 was not observed in the GHMIBC
experiments, even though a series of HMBC experiments with a variety of delays for
the evolution of long-range
2'3JCH
couplings (4 Hz to 12 Hz) were acquired. In order
to visualize this essential molecular connectivity and to confirm our structural
proposal (Figure VII.2), we acquired a static-optimized 1,1 -ADEQUATE experiment
(optimized for 50 Hz) and an ACCORD-ADEQUATE (optimized over a 40-70 Hz
range). In addition, utilization ofjamaicamide A allowed a direct comparison of the
two experiments (1,1-ADEQUATE and ACCORD-1,1-ADEQUATE) on a "real life"
structural ambiguity. As shown in Figure Vll.4, the ACCORD-ADEQUATE shows all
possible
1JHCC
correlations whereas the static-optimized experiment only shows 58%
of the desired responses.
As can be seen with this example, the refocused ACCORD-ADEQUATE pulse
sequence provides superior performance when compared to the static-optimized 1,1 -
ADEQUATE. Validation of the ACCORD-ADEQUATE pulse sequence was
demonstrated on the model compound ethyl trans-crotonate, in which all possible 1Jcc
correlations were observed. All possible correlations were also observed for
jamaicamide A, providing straightforward structure elucidation of the planar structure
as well as confirmation for the partial structures determined by standard 2-dimensional
NMR methods (see above). To aid in combating the inherently low sensitivity
associated with ADEQUATE-type experiments, a further increase in S/N can be
envisioned by the incorporation of broadband inversion or refocusing pulses such as
the CHIRP pulses originally reported for the INADEQUATE experiment)4
169
Experimental
All NMR experiments were performed on a Bruker DRX 500 NMR spectrometer
operating at a 'H resonance frequency of 500.15 MHz and a '3C resonance frequency
of 125.77 MHz. All data were referenced to residual DMSO (2.50, 39.51 ppm) or
CHC13 (7.26, 77.0 ppm) solvent. A 5mm 'H-detected triple resonance Bruker
CryoProbeTM with the 1H receiver coil held at 27 K and the 'H preamp at 70 K was
utilized for all data acquisition. All other probe hardware was operated at ambient
temperature.
The data for trans-ethyl crotonate (Aldrich Chemical) were acquired on a 67 mg
sample (75 pL) diluted in 525 j.tL deuterated DMSO (100%-d6, Cambridge Isotopes,
Inc.). The data for jamaicamide A were acquired on a 20 mg sample dissolved in 600
i.tL CDC13 (99.99% CDC13; Cambridge Isotopes, Inc.).
I
The optimization values were calculated as follows: [1/((4* Jcc1nin
where
1fccmin
is the smallest '3C '3C coupling of interest, 'Jcc
I
Jccmax)/ni)1
is the largest '3C-' 3C
coupling of interest, and ni is the number of echo and antiecho increments in F,.
170
References
1.
Bax, A.; Freeman, R.; Frenkiel, T. A.;J. Am. Chem. Soc. 1981,103,2102-2104.
2. (a) Weigelt, J.; Ofting, G. J. Magn. Reson. Ser. A 1995, 113, 128-130. (b)
Meissner, A.; Moskau, D.; Nielsen, C.; Sorensen, 0. W.; .1. Magn. Reson. 1997,
124,245-249.
3. (a) Kock, M.; Reif, B.; Fenical, W.; Griesinger, C. Tetrahedron Lett. 1996, 37,
363-366. (b) Reif, B.; Kock, M.; Kerssebaum, R.; Kang, W.; Fenical, W.;
Griesinger, C. J. Magn. Reson. Ser. A 1996, 118, 282-285. (c) Reif, B.; Kock, M.;
Kerssebaum, R.; Schleucher, J.; Griesinger, C.; I Magn. Reson. Ser. B. 1996, 112,
295-301.
4. (a) Pratum, T. K.; Moore, B. S. J. Magn. Reson. Ser. B, 1993, 102, 91-97. (b)
Pratum, T. K. J. Magn. Reson. Ser. A, 1995, 117, 132-135.
5. Kogler, H.; Sorensen, 0. W.; Bodenhausen, G.; Ernst, R. R. J. Magn. Reson. 1983,
55, 157-163.
6. Berger, S.; Wagner, R. Magn. Reson. Chem. 1998, 36, S44.
7. Martin, G. E.; Hadden, C. E.; Crouch, R. C.; Krishnamurthy, V. V. Magn. Reson.
Chem. 1999, 37, 5 17-528.
8. Hadden, C. B.; Martin, G. E.; Krishnamurthy, V. V. J. Magn. Reson. 1999, 140,
274-280.
9. Hadden, C. E.; Martin, G. E.; Krishnamurthy, V. V. Magn. Reson. Chem. 2000,
38, 143-147.
10. Krishnamurthy, V. V.; Russell, D. J.; Hadden, C. E.; Martin, G. E.; J. Magn.
Reson. 2000, 146, 232-239.
11. Marquez, B. L.; Nogle, L. N.; Williamson, R. 1.; Gerwick, W. H.; J. Nat. Prod.
Manuscript in preparation.
12. Bax, A.; Freeman, R.; Kempsell, S. P. .1. Am. Chem. Soc. 1980, 102, 4849-485 1.
13. Homonuclear 'Jcc coupling constants for ethyl trans-crotonate were measured as
follows: (C1-C2 = 75.2 Hz; C2-C3 = 70.1 Hz; C3-C4 = 41.3 Hz; C5-C6 37.9
Hz).
171
14. (a) Boehien, J. M.; Rey, M.; Bodenhausen, G. J. Magn. Reson. 1989, 84, 191-197.
(b) Boehien, J. M.; Burghardt, I.; Rey, M.; Bodenhausen, G.; J. Magn. Reson.
1990, 90, 183-191.
172
CHAPTER VIII
CONCLUSIONS
The first three chapters of this thesis deal with the isolation and structure
elucidation of four new secondary metabolites. Additionally, biosynthetic
investigations of two of these compounds were accomplished. All of these
compounds were isolated from the marine cyanobacteria Lyngbya majuscula. A
single collection of L. majuscula adapted to laboratory culture conditions has
yielded three of these new metabolites, namely hectochlorin and jamaicamides A
and B. These compounds were isolated via a phytochemical guided fractionation.
The planar structure of hectochlorin was deduced through standard two-
dimensional NMR techniques. To investigate the absolute stereochemistry, x-ray
diffraction studies were employed. Under the appropriate solvent conditions,
hectochiorin readily formed large diffraction quality crystals. Collection of high
resolution data allowed the refinement of the structure parameters to 0.85
A.
Incorporating anomalous scattering, the absolute stereochemistry was determined.
Hectochlorin also stimulates the assembly of actin filaments similar to that
observed for jasplakinolide.
173
8
13
CX)\O
16))27
0
I
o
24
Hectochiorin
The planar structures ofjamaicamides A and B were also put together by
two-dimensional NMR methods. However the standard set of experiments (e.g.
COSY, HSQC, HMBC) was not sufficient to completely elucidate their structures.
Therefore a new experiment was used, namely the ACCORD-ADEQUATE. This
experiment allowed the assignment of intact '3C-13C atoms. In addition, a
HMBC experiment provided an additional key correlation. The remaining
challenge associated with this structure elucidation was the brominated acetylene
moiety. To elucidate this portion of the structure the use of a model compound was
required. Acquisition of a '3C NMR of the model compound allowed the
assignment of the remaining two atoms (C and Br).
27
Cl
24
0
20 0
26
Jamaicamide A (5)
R = Br
Jamaicamide B (6)
R=H
R
174
Since the producer of these compounds thrived under laboratory culture
conditions, biosynthetic studies were done. Through feeding isotope labeled
substrates, the biosynthetic units ofjamaicamide were determined. Shown below is
a summary of the feeding studies.
22(4
0
['3C3,'5N}f3-alanine
21
{13C}acetate
U 12-'3Clacetate
£ S-[methyl-'3C}methiornne
[1,2.3C2}acetate
S-[3)3C]alarnne
The last of the four new compounds described in this thesis is a simple
derivative of the molluscicidal natural product barbamide. This compound,
dechlorobarbamide, was assembled by one- and two-dimensional NMR methods in
tandem with chemical shift comparisons to the "parent" compound barbamide.
CH,
L)
N'NS
\=1
OCH3 CCI3
°
Barbaniide
CH,
OCH, CHCl2
N'S
Dechiorobarbamide
175
The stereochemistry at C7 of barbamide was deduced through ozonolysis, acid
hydrolysis and treatment with Marfey's reagent. This derivatized hydrolysate was
compared to standards by HPLC retention time.
A biosynthetic feeding study was done to learn more about the halogenation
process that creates the trichioromethyl group in barbamide. A large '3C NIMIR
enhancement of C4 was observed in the 1D '3C NMR spectrum of barbamide
isolated from cultures supplemented with [2-' 3C]-5,5 ,5-trichloroleucine.
Observation of this enhancement provided additional evidence that leucine serves
as the substrate for the chlorination reaction in barbamide.
The ichthyotoxic metabolite antillatoxin was investigated. During the
course of a total synthesis of antillatoxin, it was discovered that the original
stereochemical assignment was incorrect. A detailed analysis of newly acquired
NMR data of natural antillatoxin allowed the prediction of different
stereochemistry at the two centers thought to be responsible for the discrepancy
between natural and synthetic antillatoxin. This new stereochemical prediction was
subsequently proven correct. All four isomers centered about these two centers
were provided to our laboratory. Analysis of the biological properties of these four
isomers showed varying levels of activity, with the natural compound being the
most active. In an effort to understand the role that conformation might have on
the bioactivity, the solution structures were determined. These structures were
generated through NMIR constrained molecular modeling calculations. Analysis of
the four solution structures revealed substantial conformational differences exist,
176
which is hypothesized to play a major role in the biological activity of antillatoxin
and three of its synthetic stereoisomers.
14
1 3
15
11
16
TTh10
5
17
H
HI
4R,5R
Antillatoxin (natural)
4S, 5R Antillatoxin
*"*
's"
H
4R,5S Antillatoxin
4S,5S Antillatoxin
A further application of NMR to the stereochemical analysis of natural
products is shown through though use of long-range heteronuclear coupling
constants applied to the J-based configuration analysis. The use of this analysis
requires the measurement of 2'3JCH coupling constants. In the past, these coupling
constants have been notoriously difficult to measure. However, two experiments
are presented that allow the facile measurement of these long-range scalar coupling
constants. These experiments were given the acronyms HSQMBC and G-BIRDR,XHSQMBC. Both of these experiments provide excellent line-shape, and in the case
of the G-BIRDR,x-HSQMBC, typically allow direct measurement of the coupling
177
constants of interest. The G-BffiDjc-HSQMBC experiment, in addition to the
E.COSY experiment, were used to determine the relative stereochemistry of the
neurotoxic marine natural product kalkitoxin. Three of the five stereocenters in
kalkitoxin were predicted correctly, as shown by a comparison of the natural
product with synthetic compound provided to our laboratory.
"S
H3CH34.
CH
0
CH3 CH3
14
15
Kalkitoxin
3JcH<1Hz
JHH13Hz
JCH-7.4Hz
JCH<1Hz
i
(f
H9
H%
3HH = 4.4 Hz ('H9a
L
3JHH = 8.2 Hz
CH = 6.1 Hz
Cl 5
3HH = 3.1
Cl 3
C14
kC14
3J8.2 Hz
H9b*
C8
KH = 9.4 HZ
The final chapter in this thesis details the development of a new
ADEQUATE experiment, namely the ACCORD-ADEQUATE. This experiment
incorporates an accordion delay period that allows a sampling of a range of
coupling constants. The experiment is extremely valuable to the natural products
chemist because it gives information that directly correlates intact 13C-'3C atoms at
natural abundance. While the experiment is quite insensitive, it will yield high
178
quality data on 5-10 mg of compound in a 24 hour period. As an example of its
utility, data was acquired for jamaicamide A. Of particular note is the appearance
of a key correlation that was missing from all other heteronuclear correlation data.
This thesis has provided compelling evidence that marine cyanobacteria are
producers of diverse, and potent biologically active metabolites. Of the molecules
presented, hectochiorin possesses strong activator of actin assembly, barbamide is a
potent molluscicidal compound, and both antillatoxin and kalkitoxin are powerful
neurotoxins. It is therefore hypothesized that the worlds oceans are valid places to
search for new pharmaceutical agents, or molecules to use as tools in better
understanding biochemical phenomenon. In addition to the large diversities of
biological activities, it has been shown that the use of NIvIIR is an effective method
for investigating complex stereochemical issues. Of primary importance in these
types of analyses is the use of long-range heteronuclear coupling constants.
179
BIBLIOGRAPHY
Allinger, N. L. J. Am. Chem. Soc. 1977, 99, 8127.
Aydin, R.; Gunther, H. Magn. Reson. Chem. 1990, 28, 448-457.
Ayscough, K. R.; Stryker, J.; Pokala, N.; Sanders, M.; Crews, P.; Drubin, D. G. J.
Cell Biol. 1997, 137, 399-416.
Bai, R.; Verdier-Pinard, P.; Gangwar, S.; Stessman, C. C.; McClure, K. J.; Sausville,
E. A.; Pettit, G. R.; Bates, R. B.; Hamel, E. Mo!. Pharmacol. 2001, 59, 462-469.
Bassarello, C.; Bifulco, G.; Zampella, A.; D'Auria, M. V.; Riccio, R.; GomezPaloma, L. Eur. J. Org. Chem. 2001, 39-44.
Bax, A.; Freeman, R.; Frenkiel, T. A.; .1. Am. Chem. Soc. 1981, 103, 2102-2104.
Bax, A.; Freeman, R.; Kempsell, S. P. J. Am. Chem. Soc. 1980, 102, 4849-4851.
Bax, A.; Griffey, R. H.; Hawkins, B. L. J. Magn. Reson. 1983, 55, 301-3 15.
Bax, A.; Summers, M. F. .1. Am. Chem. Soc. 1986, 108, 2093-2094.
Behrens, S.; Matha, B.; Bitan, G.; Gilon, C.; Kessler, H. mt. .1. Pept. Protein. Res.
1996, 48, 569-579.
Bender, D.A. Amino Acid Metabolism, John Wiley & Sons, London, 1975.
Berger, S.; Wagner, R. Magn. Reson. Chem. 1998, 36, S44.
Berman, F. W.; Murray, T. F.; Gerwick, W. H. Toxicon 1999, 37, 1645-1648.
Bewley, C. A.; Holland, N. D.; Faulkner, D. J. Experientia 1996, 52, 7 16-722.
Boehien, J. M.; Burghardt, I.; Rey, M.; Bodenhausen, G.; J. Magn. Reson. 1990,
90, 183-191.
Boehien, J. M.; Rey, M.; Bodenhausen, G. J. Magn. Reson. 1989, 84, 191-197.
Brantley, S. E.; Molinski, T. F. Org. Lett. 1999, 13, 2165-2 167.
Bruechwiler, D.; Wagner, G. .1. Magn. Reson. 1986, 69, 546-551.
180
Bubb, M. R.; Senderowicz, A. M. J.; Sausville, E. A.; Duncan, K. L. K.; Kom, E.D.
J. Biol. Chem. 1994,269, 14869-14871.
Bubb, M.R.; Spector, I.; Beyer, B. B.; Fosen K. M. J. Biol. Chem. 2000,275, 5 1635170.
Buckingham, J.; Donaghy, S. M.; Cadogan, J. I. G.; Raphael, R. A.; Rees, C. W.
Dictionaiy of Organic Compounds, Chapman and Hall, New York, 1982.
Butler, A.; Walker, J. V. Chem. Rev. 1993, 93, 1937-1944.
Carlier, M.-F. Curr. Opin. Cell Biol. 1998, 10, 45-51.
Canneli, S.; Moore, R. E.; Patterson, G. M. L.; Yoshida, W. Y. Tetrahedron Lett.
1993, 34, 557 1-5574.
Cimimello, P.; Fattorusso, E.; Forino, M.; Di Rosa, M.; lanaro, A.; Poletti, J. Org.
Chem. 2001, 66, 578-582.
Clore, G. M.; Gronenbom, A. M.; Marius, C. G. Curr. Opin. Chem. Biol. 1998,2,
564-570.
Crews, P.; Manes, L. V.; Boehier, M. Tetrahedron Leti. 1986, 27, 2797-2800.
Culotta, E. Science 1994, 263, 918-920.
Dewar, M. J. S.; Zoebisch, E. V.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc.
1985, 107, 3902-3909.
Dumdei, E. J.; Simpson, J. S.; Garson, M. J.; Byriel, K. A.; Kennard, C. H. L. Aust.
J. Chem. 1997, 50, 139-144.
Eberstadt, M.; Gemmecker, G.; Mierke, D. F.; Kessler, H. Angew. Chem. mt. Ed.
Engi. 1995, 34, 1671-1695.
Eberstadt, M.; Mierke, D. F.; Kock, M.; Kessler, H. Helv. Chim. Acta 1992, 75,
2583-2592.
Falk, M.; Spierenburg, P. F.; Walter, J. A. J. Comp. Chem. 1996, 17, 409-4 17.
Famigia, L. J. .1. Appi. Cryst. 1997,30,565.
Faulkner, D. J. Nat. Prod. Rep. 2000, 17, 7-55.
181
Fletcher, M. D.; Harding, J. R.; Hughes, R. A.; Kelly, N. M.; Schmalz, H.;
Sutherland, A.; Willis, C. L. J. Chem. Soc., Perkin Trans. 1 2000, 43-5 1.
Flowers, A. E.; Garson, M. J.; Webb, R. I.; Dumdei, E. J.; Charan, R. D. Cell
Tissue Res. 1998, 292, 597-607.
Fujikawa, A.; Yasuko, I.; Inoue, M.; Ishida, T.; Nemoto, N.; Kobayashi, Y.;
Kataoka, R.; llcai, K.; Takesako, K.; Kato, I. .1. Org. Chem. 1994, 59, 570-578.
Fujita, K.; Fujiwara, M.; Yamasaki, C.; Matsuura, T.; Furihata, K.; Seto, H.
Symposium papers of 38t12 Symposium on the Chemistry of Natural Products,
Organizing Committee of the Symposium: Sendi, 1996; 379-384.
Fuller, R. W.; Cardellina, J. H.; Kato, Y.; Brinen, L. S.; Clardy, J.; Snader, K.;
Boyd, M. R. J. Med. Chem. 1992, 35, 3007-3011.
Furihata, K.; Seto, H. Tetrahedron Lett. 1999, 40, 627 1-6275.
Fu, X.; Ferriera, M. L.; Schmidtz, F. J.; Kelly-Borges, M. J. Nat. Prod. 1998, 61,
1226-1231.
Fu, X.; Zeng, L. -M.; Su, J. -Y.; Pais, M. J. Nat. Prod. 1997, 60, 695-696.
Fu, X.; Zeng, L. -M.; Su, J. -Y.; Pais, M. J. Nat. Prod. 1993, 56, 637-642.
Gebreyesus, T.; Yosief, T.; Carmely, S.; Kashman, Y. Tetrahedron Lett. 1988, 29,
3863-3864.
Gerwick, W. H.; Proteau, P. J.; Nagle, D. G.; Hand, E. Blokhin, A.; Slate, D. L. J.
Org. Chem. 1994, 59, 1243-1245.
Gerwick, W. H.; Roberts, M. A.; Proteau, P. J.; Chen, J-L. .1. Appi. Phycol. 1994. 6,
143-149.
Gerwick, W. H.; Tan, L. T.; Sitachitta, N. Nitrogen-Containing Metabolites from
Marine Cyanobacteria. In Alkaloids, in press.
Griesinger, C; Sørenson, O.W.; Ernst, R.R. J. Am. Chem. Soc. 1985, 107, 639463 96.
Haasnoot, C. A. G.; DeLeeuw, F. A. A. M.; Altona, C. Tetrahedron 1980, 36,
2783-2792.
182
Hadden, C. E.; Martin, G. E.; Krishnamurthy, V. V. J. Magn. Reson. 1999, 140,
274-280.
Hadden, C. E.; Martin, G. E.; Krishnamurthy, V. V. Magn. Reson. Chem. 2000,38,
143-147.
Hanson, K. J. Am. Chem. Soc. 1966, 88, 273 1-2742.
Harrigan, G. G.; Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Nagle, D. G.; Paul, V.
J.; Mooberry, S. L.; Corbett, T. H.; Valeriote, F. A. J. Nat. Prod. 1998, 61, 10751077.
Hartung, J. Angew. Chem., mt. Ed. 1999, 38, 1209-1211.
Hayakawa, Y.; Kim, J-W.; Adachi, H.; Shin-ya, K.; Fujita, K.; Seto, H. J. Am.
Chem. Soc. 1998, 120, 3524-3525.
Helmchen, G.; Wegner, G. Tetrahedron Lett. 1985,26, 6047-6050.
Hotheinz, von H.; Oberhansli, W. E. Helv. Chem. Acta 1977, 60, 660-669.
Hooper, G. J.; Orjala, J.; Schatzman, R. C.; Gerwick, W. H. .1 Nat. Prod. 1998, 61,
529-533.
Huffinan, W. A.; Ingersoll, A. W. .1. Am. Chem. Soc. 1951, 73, 3366-3369.
Hurd, R. E. J Magn. Reson 1990, 87, 422-428.
Hwang, T.-L.; Shaka, A. J. I Am. Chem. Soc. 1992, 114, 3157-3159.
Hwang, T.-L.; Shaka, A. J. .1. Magn. Reson. Ser. B 1993, 102, 155-165.
HyperChem, Hypercube Inc. Gainesville, FL.
hunan, W.; Crews, P. J. Am. Chem. Soc. 1989, 111, 2822-2829.
Ishida, T.; Ohishi, H.; lnoue, M.; Kamigauchi, M.; Sugiura, M.; Takao, N.; Kato, S,
Hamada, Y.; Shioiri, T. .1. Org. Chem. 1989, 54, 5337-5343.
John, B. K.; Plant, D.; Hurd, R. E. .1. Magn. Reson. 1993, 101, 113-117.
Kao, L.-F.; Barfield, M.; J. Am. Chem. Soc. 1985, 107, 2323-2330.
Kay, L. E.; Keifer, P.; Saarinen, T. J. Am. Chem. Soc. 1992, 114, 10663-10665.
183
Kazlauskas, R.; Ligard, R. 0.; Wells, R. J.; Vetter, W. Tetrahedron Lett. 1977,
3183-3 186.
Kazlauskas, R.; Murphy, P. T.; Wells, R. J. Tetrahedron Lett. 1978, 4945-4948.
Kelly, N. M.; Reid, R. G.; Willis, C. L. Tetrahedron Lett. 1995, 36, 8315-8318.
Kelly, N. M.; Sutherland, A.; Willis, C. L. Nat. Prod. Rep. 1996, 205-2 19.
Kingston, D. G. I.; Kolpak, M. X.; LeFevre, J. W.; Borup-Grochtmann, I. .1. Am.
Chem. Soc. 1983,105,5106-5110.
Kobayashi, M.; Aoki, S.; Kitagawa, I.; Kobayashi, Y.; Nemoto, N.; Fujikawa, A.
Symposium papers 0f38th Symposium on the Chemistry of Natural Products,
Organizing Committee of the Symposium: Sendi, 1996; 6 1-66.
Kobayashi, M.; Aoki, S.; Kitigawa, I. Tetrahedron Lett. 1994,35, 1243-1246.
Kock, M.; Junker, J. Bioorganic Chemistry; Diederichsen U, Ed.; Wiley-VCH
Verlag GmbH: Weinheim, Germany, 1999; 365-378.
Kock, M.; Junker, J. J. Org. Chem. 1997, 62, 8614-8615.
Kock, M.; Kessler, H.; Seebach, D.; Thaler, A. J. Am. Chem. Soc. 1992, 114, 26762686.
Kock, M.; Reif, B.; Fenical, W.; Griesinger, C. Tetrahedron Lett. 1996, 37, 363366.
Koehn, F. E.; Longley, R. E.; Reed, J. K. J. Nat. Prod. 1992, 55, 613-619.
Kogler, H.; Sorensen, 0. W.; Bodenhausen, G.; Ernst, R. R. J. Magn. Reson. 1983,
55, 157-163.
Kontaxis, G.; Stonehouse. J.; Laue, E. D.; Keeler, J. J. Magn. Reson. Ser. A 1994,
111, 70-76.
Kover, K. E.; Hruby, V. J.; Uhrin, D. J. Magn. Reson. 1997, 129, 125-129.
Kozminski, W.; Nanz, D. .1. Magn. Reson. 2000, 142, 294-299.
Krislmamurthy, V. V.; Russell, D. J.; Hadden, C. E.; Martin, G. E.; J. Magn. Reson.
2000, 146, 232-239.
184
Lee, G. M.; Molinski, T. F. Tetrahedron Lett. 1992, 33, 767 1-7674.
Ludvigsen, S.; Andersen, K. V.; Poulsen, F. M. I Mol. Biol. 1991, 217, 73 1-736.
Leusch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J.; Corbett, T. H. I. Am. Chem.
Soc. ASAP article.
Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J. J. Nat. Prod. 2000, 63, 14371439.
Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J.; Moobeny, S. L. I. Nat.
Prod. 2000, 63,611-615.
MacMillan, J. B.; Molmski T. F. J. Nat. Prod. 2000, 63, 155-157.
Manger, R.L.; Leja, L.S.; Lee, S.Y.; Hungerford, J.M.; Hokama, Y.; Dickey, R.W.;
Granade, H.R.; Lewis, R.; Yasumoto, T.; Wekell, M.M. I. AOAC Intern. 1995, 78,
521-527.
Marek, R.; Kralik, L.; Skienar, V. Tetrahedron Lett. 1997, 38, 665-668.
Marfey, P. CarlsbergRes. Commun. 1984, 49, 591-596.
Marinlit, 2000; database of the marine natural products literature; Department of
Chemistry, University of Canterbury: Christchurch, New Zealand, 2000.
Marquez, B. L.; Gerwick, W. H. (SMASH) Small Molecule NMR Conference: Oral
Presentation; Chicago, IL 2000.
Marquez, B. L.; Gerwick, W. H.; Williamson, R. T. Magn. Reson. Chem. Accepted.
Marquez, B. L.; Nogle, L. M.; Williamson, R. 1.; Gerwick, W. H. J. Nat. Prod.
Manuscript in preparation.
Marquez, B. L.; Watts, K. S.; Roberts, M. A.; Verdier-Pinard, P.; Jimenez, J. I.; Ho,
P. S.; Hamel, E.; Scheuer, P. J.; Gerwick, W. H. manuscript in preparation.
Marquez, B. L.; Williamson, R. T.; Gerwick, W. H. Experimental Nuclear
Magnetic Resonance Conference: Poster number 021; Orlando, 1999.
Marquez, B.; Verdier-Pinard, P.; Hamel, E.; Gerwick, W. H. Phytochemistry 1998,
49, 2387-2389.
185
Marshall, J. L. in Carbon-carbon and carbon-proton NMR couplings; Methods in
stereochemical analysis 2; Verlag Chemie International: Deerfield Beach, FL,
1983.
Martin, G. E.; Crouch, R. S. J. Heterocyclic Chem. 1995, 32, 1665-1669.
34th
Annual
Martin, G. E.; Crouch, R. C.; Sharaf, M. H. M.; Schiff, P. L., Jr.
Meeting of the American Society of Pharmacognasy, San Diego, CA, July 18-22,
Abstract P101.
Martin, G. E.; Hadden, C. E.; Crouch, R. C.; Krishnamurthy, V. V. Magn. Reson.
Chem. 1999, 37, 5 17-528.
Matsumori, N.; Kaneno, D.; Murata, M.; Nakamura, H.; Tachibana, K. J. Org.
Chem. 1999, 64, 866-876.
Matsumori, N.; Murata, M.; Tachibana, K.; Tetrahedron 1995, 51, 12229-1223 8.
McDonald, L. A.; Ireland, C. M. .1. Nat. Prod. 1992, 55, 376-379.
Meissner, A.; Moskau, D.; Nielsen, C.; Sorensen, 0. W.; J. Magn. Reson. 1997,
124,245-249.
Milligan, K. E.; Marquez, B. L.; Williamson, R. T.; Gerwick, W. H. J. Nat. Prod.
2000, 63, 1440-1443.
Milligan, K. E.; Marquez, B.; Williamson, R. 1.; Davies-Coleman, M.; Gerwick,
W. H. J. Nat. Prod. 2000, 63, 965-968.
Mohamade, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton, M.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput. Chem. 1990, 11,
440.
Moore, R. E. Pure App!. Chem. 1982, 54, 1919-1934.
Morita, H.; Kayashita, T.; Takeya, K.; Itokawa, H.; Shiro, M. Tetrahedron 1997,
53, 1607-1616.
Monta, H.; Yoshida, N.; Takeya, K.; Itokawa, H.; Shirota, 0. Tetrahedron 1996,
52, 2795-2802.
Murakami, Y.; Oshima, Y.; Yasumoto, T. Nippon Suisan Gakkaishi 1982, 48, 6972.
Murata, M.; Iwashita, T.; Yokoyama, A.; Sasaki, M.; Yasumoto, T. J. Am. Chem.
Soc. 1992, 114, 6594-6596.
Murata, M.; Matsuoka, S.; Matsumon, N.; Paul, G. K.; Tachibana, K. J. Am. Chem.
Soc. 1999, 121, 870-871.
Murata, M.; Naoki, H.; Iwashita, T.; Matsunaga, S.; Sasaki, M.; Yokoyama, A.;
Yasumoto, T. I Am. Chem. Soc. 1993, 115, 2060-2062.
Murata, M.; Yasumoto, T. Nat. Prod. Rep. 2000, 17, 293-3 14.
Muzaffar, A.; Brossi, A.; Lin, C. M.; Hamel, E. .1. Med. Chem. 1990,33, 567-571.
Nagle, D. G.; Paul, V. J. J. Exp. Mar. Biol. EcoL 1998, 225, 29-38.
Nagle, D. G.; Paul, V. J.; Roberts, M. A. Tetrahedron Lett. 1996, 37, 6263-6266.
Nitta, I.; Watase, H.; Tomiie, Y. Nature 1958, 181, 76 1-762.
Orjala, J.; Nagle, D. G.; Hsu, V. L.; Gerwick, W. H. .1. Am. Chem. Soc. 1995, 117,
8281-8282.
Orjala, J. 0.; Gerwick, W. H. J. Nat. Prod. 1996,59,427-430.
Palmer III, A. G.; Cavanagh, J.; Write, P. E.; Rance, M. J. Magn. Reson. 1991, 93,
151-170.
Panda, D.; Ananthnarayan, V.; Larson, G.; Shib, C.; Jordan, M. A.; Wilson, L.
Biochemistry 2000, 39, 14121-14127.
Panda, D.; DeLuca, K.; Williams, D.; Jordan, M. A.; Wilson, L. Porc. Nat!. Acad.
Sci. 1998, 95, 9313-9318.
Parella, T. Magn. Reson. Chem. 1998, 36, 467-495.
Patil, A. D.; Kokke, W. C.; Cochran, S.; Francis, T. A.; Tomszek, 1.; Westley, J.
W.J. Nat. Prod. 1992, 55, 1170-1177.
Patthy-Lukats, A.; Karolyhazy, L.; Szabo, L. F.; Podanyi, B. .1. Nat. Prod. 1997,
60, 69-75.
Patthy-Lukats, A.; Kocsis, A.; Szabo, L.; Podanyi, B. .1. Nat. Prod. 1999, 62, 14921499.
187
Pettit, G. R.; Kamano, Y.; Dufresne, C.; Cemey, R. L.; Herald, D. L.; Schmidt, J.
M. .J. Org. Chem. 1989, 54, 6005-6006.
Pettit, G. R.; Kamano, Y.; Herald, C. L.; Tuinman, A. A.; Boettner, F. E.; Kizu, H.;
Schmidt, J. M.; Baczynskyj, L.; Tomer, K. B.; Botems, R. J. J. Am. Chem. Soc.
1987, 109, 6883-6895.
Polak, E.; Ribiere, G. Revenue Francaise Informat. Recherche Operationelle, 1969,
v. 35.
Pratum, T. K. J. Magn. Reson. Ser. A, 1995, 117, 132-135.
Pratum, T. K.; Moore, B. S. J. Magn. Reson. Ser. B, 1993, 102, 9 1-97.
Proteau, P. J.; Gerwick, W. H.; Garcia-Pichel, F.; Castenholz, R. Experientia 1993,
49, 825-829.
Ramsey, F. L.; Schafer, D. W. The Statistical Sleuth; A Course in Methods of Data
Analysis, Duxbury Press, Belmont, CA, 1997, p. 36-42.
Reif, B.; Kock, M.; Kerssebaum, R.; Kang, W.; Fenical, W.; Griesinger, C. J
Magn. Reson. Ser. A 1996, 118, 282-285.
Reif, B.; Kock, M.; Kerssebaum, R.; Schleucher, J.; Griesinger, C.; .1. Magn.
Reson. Ser. B. 1996, 112, 295-301.
Rossi, J. V. M.S. Thesis, Oregon State University, Corvallis, 1997.
Rossi, J. V.; Roberts, M. A.; Yoo, H-D.; Gerwick, W. H. J. App!. Phycol. 1997, 9,
195-204.
Rundlof, T.; Kjellberg, A.; Damberg, C.; Nishida, T.; Widmaim, G. Magn. Reson.
Chem. 1998, 36, 839-847.
Schleucher, J.; Schwendinger, M.; Sattler, M.; Schmidt, P.; Schedletzky, 0.;
Glaser, S. J.; Sorenson, 0. W.; Griesinger, C. .1. Biomol. NMR 1994, 4, 301-306.
Scott, K.; Keeler, J.; Van, Q. V.; Shaka, A. J. .1. Magn. Reson. 1997, 125, 302-324.
Seebach, D.; Ko, S-Y.; Kessler, H.; Kock, M.; Reggeim, M.; Schmieder, P.;
Walkinshaw, M. D.; Boelsterli, J. J.; Bevek, D. Helv. Chim. Acta. 1991, 74, 19531990.
Seto, H.; Watanabe, H.; Furihata, K. Tetrahedron Lett. 1996, 37, 7979-7982.
188
Shaw-Reid, C.A.; Kelleher, N.L.; Losey, H.C.; Gebring, A.M.; Berg, C.; Walsh,
C.T. Chem. Biol. 1999, 6, 385-400.
Sheng, S.; van Halbeek, H. J. Magn. Reson. 1998, 130, 296-299.
Shimojima, Y.; Hayashi, H.; Ooka, T.; Shibuawa, M. Tetrahedron 1984, 40, 25192527.
Sitachitta, N.; Gerwick, W. H. J. Nat. Prod. 1998, 61, 681-684.
Sitachitta, N.; Marquez, B. L.; Williamson, R. T.; Rossi, I. V.; Roberts, M. A.;
Gerwick, W. H.; Nguyen, V-A.; Willis, C. L. Tetrahedron 2000, 56, 9103-9113.
Sitachitta, N. Ph.D. Thesis, Oregon State University, Corvallis, 2000.
Sitachitta, N.; Rossi, J.; Roberts, M. A.; Gerwick, W. H.; Fletcher, M. D.; Willis, C.
L. J. Am. Chem. Soc. 1998, 120, 7131-7132.
Smith, C. D.; Zhang, X.; Mooberry, S. L.; Patterson, G. M.; Moore, R. E. Cancer
Res. 1994, 14, 3779-3784.
Sone, H.; Kondo, T.; Kiryu, M.; Ishiwata, H.; Ojika, M.; Yamada, K. J. Org. Chem.
1995, 60, 4774-4781.
Sotokawa, T.; Noda, T.; Pi, S.; Hirama, M. Angew. Chem. mt. Ed. Engi. 2000, 39,
3430-3432.
Spector, I.; Bract, F.; Shochet, N. R.; Bubb, M. R. Microsc. Res. Tech. 1999, 47, 1837.
Stott, K.; Keeler, J.; Van, Q. N.; Shaka, A. J. J Magn. Reson. 1997, 125, 302-324.
Stott, K.; Stonehouse, J.; Keele, J.; Hwang, T-L.; Shaka, A. J. J. Am. Chem. Soc.
1995, 117, 4199-4200.
Tan, L. T. Bioactive natural products from marine algae. Ph.D. Thesis, Oregon
State University, Corvallis, OR, 2001.
Tan, L.T.; Williamson, R.T.; Watts, K.S.; Gerwick, W.H.; McGough, K.; Jacobs,
R. J. Org. Chem. 2000, 65, 419-425.
Tvaroska, I.; Taravel, F. R. Adv. Carb. Chem. Biochem. 1995,51, 15-61.
Uhrin, D.; Batta, G.; Hruby, V. J.; Barlow, P. N.; Kover, K. E. J. Magn. Reson.
1998, 130, 155-161.
Uzawa, J.; Utumi, H.; Koshino, H.; Hinomoto, T.; Anzai, K. 32 NMR
Conference, Tokoyo, Japan, November 4-6, 1993; pp 147-150.
Vederas, J.C. Nat. Prod. Rep. 1987,277-337.
Verdier-Pinard, P.; Lai, J. Y.; Yoo, H. D.; Yu, J.; Marquez, B.; Nagle, D. G.;
Nambu, M.; White, J. D.; Faick, J. R.; Gerwick, W. H.; Day, B. W.; Hamel, E. Mo!.
Pharmacol. 1998, 53, 62-76.
Voet, D.; Voet, J. G. In Biochemistry; Wiley & Sons, Inc.: New York, 1990;
Chapter 24.
Wegner, A.; Engel, J. Biophys. Chem. 1975, 3, 215-225.
Weigelt, J.; Otting, G. J. Magn. Reson. Ser. A 1995, 113, 128-130.
White, J. D.; Hanselmann, R.; Wardrop, D. J. J. Am. Chem. Soc. 1999, 121, 11061107.
Williamson, R. 1.; Boulanger, A.; Vulpanovici, A.; Roberts, M. A.; Gerwick, W.
H. in preparation.
Williamson, R. T. Development and application of NMR spectroscopy to marine
natural products structure and biosynthesis, Ph.D. Thesis, Oregon State University,
Corvallis, OR, 2000.
Williamson, R. T. Isolation and Synthesis of Bioactive Marine Galactolipids. M.S.
Thesis, University of North Carolina at Wilmington, Wilmington, NC, 1996.
Williamson, R. T.; Marquez, B. L.; Gerwick, W. H.; Koehn, F. E. Magn. Reson.
Chem. In press.
Williamson, R. T.; Marquez, B. L.; Gerwick, W. H.; Martin, G. E.; Krishnamurthy,
V. V. Magn. Reson. Chem. 2001, 39, 127-132.
Williamson, R. T.; Marquez, B. L.; Gerwick, W. H. (SMASH) Smal! Mo!ecu!e
NMR Conference: Oral Presentation; Chicago, IL 2000.
Williamson, R. T.; Marquez, B. L.; Kover, K. E.; Gerwick, W. H. Magn. Reson.
Chem. 2000,38,265-273.
190
Williamson, R. T.; Sitachitta, N.; Gerwick, W. H. Tetrahedron Lett. 1999, 40,
5175-5178.
Williard, P. G.; Laszlo, S. E. J. Org. Chem. 1984, 49, 3489-3493.
Wiliker, W.; Leibfritz, D.; Kerssebaum, R.; Bermel, W. Magn. Reson. Chem. 1993,
31,287-292.
Wiliker, W.; Liebfritz, D.; Kerssebaum, R.; Lobman, J. J. Magn. Reson. Ser. A
1993, 102, 348-350.
Wipf, P.; Reeves, J. 1.; Balachandran, R.; Giuliano, K. A.; Hamel, E.; Day, B. W.
J. Am. Chem. Soc. 2000, 122, 9391-9395.
Wright, J. L. C.; Boyd, R. K.; de Freitas, A. S. W.; Falk, M.; Foxall, R. A.;
Jamieson, W. D.; Laycock, M. V.; McColloch, A. W.; Mclnnes, A.G. Can. J.
Chem. 1989, 67, 481-490.
Wu, M. Novel Bioactive Secondary Metabolites from the Marine Cyanobacterium
Lyngbya majuscula, M.S. Thesis, Oregon State University, 1997.
Wu, M.; Okino, T.; Nogel, L. M.; Marquez, B. L.; Williamson, R. T.; Sitachitta, N.;
Berman, F. W.; Murray, T. F.; McGough, K.; Jacobs, R.; Colsen, K.; Asano, T.;
Yokokawa, F.; Shioiri, T.; Gerwick, W. H. .J. Am. Chem. Soc. 2000, 122, 1204112042.
Yasumoto, T.; Oshima, Y.; Sugawara, W.; Fukuyo, Y.; Oguri, H.; Igarashi, T.;
Fujita, N. Nippon Suisan Gakkaishi 1980, 46, 1405-1411.
Yokokawa, F.; Shioiri, T. J. Org. Chem. 1998, 63, 8638-8639.
Yoo, H.-D.; Gerwick, W. H. J. Nat. Prod. 1995, 58, 1961 -1965.
Zabriskie, T. M.; Kiocke, J. A.; Ireland, C. M.; Marcus, A. H.; Molinski, T. F.;
Faulkner, D. J.; Xu, C.; Clardy, J. J. Am. Chem. Soc. 1986, 108, 3 123-3124.