Citation
Synthesis of rhizoferrin, staphyloferrin A, and polyamine catecholamides and the study of their effects on iron transport in Paracoccus denitrificans

Material Information

Title:
Synthesis of rhizoferrin, staphyloferrin A, and polyamine catecholamides and the study of their effects on iron transport in Paracoccus denitrificans
Alternate title:
Synthesis of rizoferrin and polyamine catecholamides and the study of their effects on iron transport in Paracoccus denitrificans
Creator:
Xin, Mei Guo, 1968-
Publication Date:
Language:
English
Physical Description:
xv, 151 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Amines ( jstor )
Chelates ( jstor )
Citrates ( jstor )
Iron ( jstor )
Kinetics ( jstor )
Ligands ( jstor )
Microorganisms ( jstor )
Paracoccus denitrificans ( jstor )
Polyamines ( jstor )
Siderophores ( jstor )
Department of Medicinal Chemistry thesis Ph. D ( mesh )
Dissertations, Academic -- College of Pharmacy -- Department of Medicinal Chemistry -- UF ( mesh )
Dissertations, Academic -- Medicinal Chemistry -- UF
Iron -- pharmacokinetics ( mesh )
Medicinal Chemistry thesis, Ph. D
Molecular Conformation ( mesh )
Paracoccus denitrificans -- growth & development ( mesh )
Paracoccus denitrificans -- physiology ( mesh )
Research ( mesh )
Siderophores -- biosynthesis ( mesh )
Siderophores -- pharmacology ( mesh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2000.
Bibliography:
Bibliography: leaves 135-150.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Mei Guo Xin.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
025616002 ( ALEPH )
70845296 ( OCLC )

Downloads

This item has the following downloads:


Full Text
SYNTHESIS OF RHIZOFERRIN, STAPHYLOFERRIN A, AND POLYAMINE
CATECHOLAMIDES AND THE STUDY OF THEIR EFFECTS ON
IRON TRANSPORT IN PARACOCCUS DENITRIFICANS
BY
MEl GUO XIN
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA
2000




This dissertation is dedicated to my wife, Hua Yao,
and my daughter, Anni Xin.




ACKNOWLEDGMENTS
Six years ago, when I first entered the graduate program in the College of Pharmacy, University of Florida, I only had the mind for the research, not the background or means to accomplish it. I had not synthesized an organic compound before, let alone tested something in a biological system. Now that I am a Ph.D. candidate in medicinal chemistry, I shall always be grateful to the people who helped me and made all that possible.
I was fortunate to have this opportunity to do research in Dr. Raymond J. Bergeron's laboratory. I owe a great debt to him for his endless patience, daily guidance, supervision, and friendship. For six years, he has taught me as much about life as about science. His farsightedness and dedication to science have always inspired me, even under the most difficult conditions. His quick thinking, vast knowledge, humor and enthusiasm for life and work have affected me as strongly as his annual Fourth of July parties. Dr. Bergeron has a special way to guide a graduate student. With a project seeming quite impossible at first, he would always encourage me to keep on with it and get to the core of the problem. At the key time of a project, he would point out what was wrong and what was right. He was so evenhandedly critical of every single mistake I made in my work. This not only impressed me deeply, but also gave me plenty of chances to understand the research and science. In his lab, I have developed the ability to
III




be an independent scientist learned through my failures and through the chances that were always there if I wanted to try again.
My sincere appreciation is also extended to my supervisory committee, Dr. John Perrin, Dr. Kenneth Sloan, Dr. Richard Streiff, and Dr. Steven Baker, for their fine instruction and effective advice.
I would like further to express my appreciation to Dr. William Weimar. It is with his kind assistance that the iron transport assays were done. In addition, I would like to express my thanks to Dr. James McManis, who spent much time assisting with synthesis.
I would also like to thank my many friends and colleagues in and out of the laboratory, Mr. Richard Smith, Mr. Sam Algee, Dr. K. A. Abboud, Ms. Jan WViegand, and Dr. Christian Ludin for all their assistance.
I will always be indebted to my parents, Jianbin Xin and Yingmei Li, and my parents-in-law, Guowei Yao and Huimin Song, who have always stood behind me and cared for me no matter how far apart we were. I am also very pleased to express my deepest gratitude to my wife, Yao Hua, whose unconditional love and full-hearted support was always there when I needed it most and deserved it least. Without her efforts during these years, this dissertation would not have been possible.
iv




TABLE OF CONTENTS
PAGE
ACKNOW LEDGEMENTS .................................................... iii
LIST OF FIGURES ............................................................ vii
LIST OF ABBREVIATIONS ................................................ x
ABSTRACT ....................................................................... xiii
CHAPTER
I. INTRODUCTION AND BACKGROUND....................... 1
The Properties of Iron in Metabolism.................... 1
Iron Metabolism in Humans............................... 4
Iron Overload in Humans.................................. 6
Iron Utilization and Transport in Microorganisms...... 8
Iron and Bacterial Infection................................ 13
The Implications and Applications of
Siderophores............................................... 15
II. SYNTHESIS OF RHIZOFERRIN............................... 22
Introduction.................... ................................. 22
Synthesis....................................................... 25
Experimental.................................................. 29
Discussion..................................................... 35
III. SYNTHESIS OF STAPHYLOFERRIN A...................... 40
Introduction..................................................... 40
Synthesis....................................................... 41
Experimental.................................................. 44
Discussion..................................................... 47
V




IV. EFFICIENT SYNTHESIS OF POLYAMINE
CATECHOLAMIDES.............................................. 50
Introduction..................................................... 50
Synthesis....................................................... 58
Experimental................................................. 66
Discussion................................................. 74
IV. THE EFFECTS OF RHIZOFERRIN, D- AND LFLUVIABACTIN, L-AGROBACTIN AND LHOMOFLUVIABACTIN ON IRON TRANSPORT IN
PARACOCCUS DENITRIFICANS............................. 86
Introduction..................................................... 86
Experimental................................................. 89
Results......................................................... 93
Discussion.................................................. 110
VI. CONCLUSIONS.................................................. 133
REFERENCES .................................................................. 135
BIOGRAPHICAL SKETCH .................................................. 151
vi




LIST OF FIGURES
Figure Page 1-1 Fenton Reactions and Production of Hydroxyl Radicals ......... 3 1-2 Examples of Microbial Siderophores................................. 10
1-3 Structure of Rhizoferrin................................................... 11
1-4 Structures of L-Parabactin and L-Vibriobactin..................... 17
1-5 Structures of L-Fluviabactin and L-Agrobactin..................... 19
1-6 Structure of Staphyloferrin A............................................ 20
2-1 Retrosynthetic Scheme of Rhizoferrin............................... 24
2-2 X-ray of R-Enantiomer of 1,2-Dimethyl Citrate.................. 26
2-3 Synthesis of Rhizoferrin.................................................. 28
2-4 Hydrolysis of Triesters of Citric Acid with PLE or Subtilisin ..... 37 2-5 The Diastereomeric Salts Formed Between 1,2-Dimethyl
Citrate and Brucine....................................................... 39
3-1 Synthesis of Staphyloferrin A.......................................... 43
3-2 Examples of Chiral Citric Acid-based Siderophores ...... 49 4-1 Structures of Two L-Fluviabactin Analogues....................... 52
4-2 Synthesis of Benzyl-protected Triamine............................. 55
4-3 Synthesis of L-Parabactin Using N4-Benzylspermidine........... 57
4-4 Synthesis of L-Vibriobactin Using Diprotected Triamine ......... 58
4-5 CDI Coupling Reactions.................................................. 61
vii




4-6 Synthesis of Reagents Ethyl 2,3-Dihydroxybenzimidate ....... 63 4-7 Synthesis of L- and D- Fluviabactin, L-Agrobactin, LHomofluviabactin.......................................................... 65
4-8 300 MHz 1H NMR Spectrum of L-Fluviabactin in CD3OD at
5 0 0C ........................................................................... 7 6
4-9 300 MHz 1H NMR NMR Spectrum of D-Fluviabactin in CD3OD
at 500C ....................................................................... 77
4-10 300 MHz 1H NMR NMR Spectrum of L-Agrobactin in CD30D
at 50 0C ....................................................................... 7 8
4-11 300 MHz 'H NMR NMR Spectrum of L-Homofluviabactin in
C D3O D at 50 C ............................................................ 79
4-12 300 MHz 'H NMR NMR Spectrum of y-Methyl Group in LFluviabactin................................................................ 82
4-13 300 MHz 'H NMR NMR Spectrum of y-Methyl Group in LA grobactin.................................................................. 83
4-14 300 MHz 'H NMR NMR Spectrum of y-Methyl Group in LHomofluviabactin.......................................................... 84
4-15 Conformations of L-Agrobactin, L-Fluviabactin, and LH om ofluviabactin.......................................................... 85
5-1 Growth Rate of P. denitrificans in 0.5 gM Minimal Salts
Solution and Induced Siderophore Production.................... 94
5-2 Growth Rate of P. denitrificans in the presence of 2.0 gM
Catecholamide Ligands................................................. 96
5-3 Iron Uptake of [55sFe] from [55Fe]Ferric L-Fluviabactin, DFluviabactin, L-Agrobactin, L-Homofluviabactin, L-Parabactin.. 97 5-4 Kinetic Characteristics of Iron Transport of [55Fe]Ferric LFluviabactin and D-Fluviabactin........................................ 99
5-5 Kinetic of Iron Transport of [55Fe]Ferric L-Fluviabactin........... 101
5-6 Kinetic of Iron Transport of [55Fe]Ferric L-Agrobactin............. 103
viii




5-7 Kinetic of Iron Transport of [55Fe]Ferric L-Homofluviabactin..... 104 5-8 CD Spectra of Ferric L- and D-Fluviabactin.......................... 107
5-9 Configuration of Ferric L-Fluviabactin................................ 108
5-10 Configuration of Ferric D-Fluviabactin................................ 109
5-11 Growth Rate of P. denitrificans in the presence of 1.0 gM
Rhizoferrin and Parabactin.............................................. 111
5-12 Structures of the Catechol Compounds Isolated from P.
denitrificans................................................................. 114
ix




LIST OF ABBREVIATIONS
AcOEt ethyl acetate [ao]25 specific optical rotation at 250C for D (sodium) line in a
1 decimeter cell, neat. Anal. analytic aq aqueous Ar aryl ATCC American Type Culture Collection atm, atmos atmosphere(s), atmospheric ATP adenosine triphosphate BOC tert-butoxycarbonyl BOC-ON 2-(tert-butoxycarbonyloxyimino)-2-phenylacetonitrile bp boiling point anti stereodesciptor Bu butyl Bz benzoyl C6H5COBzOH benzoic acid c concentration (g/100 mL) 0C centigrade degree; Celsius degree ca (circa) about Calc (d) calculate, calculated Cbz carbobenzoxy CD circular dichroism CDCI3 deutero-chloroform CDI 1,1'-carbonyldiimidazole OH2 methylene CH3 methyl CH2CI2 methylene chloride CHCI3 chloroform Chem chemical CH3OH methanol CH2Ph benzyl Ci curie cis stereodescriptor cm centimeter(s) CO2CH3 methyl ester
5 (delta) chemical shift
X




As (delta epsilon) molar ellipticity DCC 1,3-dicyclohexylcarbodiimide DCU 1,3-dicyclohexylurea DFO desferrioxamine B DMAP dimethylaminopyridine DMF dimethylformide DMSO dimethyl sulfoxide DNA deoxyribonucleic acid
(E)- entgegen stereodescriptor P (epsilon) molar extinction coefficient (M-1cm-1) Ed(s) editor(s) EDDA ethylenediamine di-o-hydroxyphenylacetic acid e.g. (exempli gratia) for example Et ethyl Et3N triethylamine Et20 diethyl ether EtOH ethyl alcohol FMN (FMNH2) flavin mononucleotide (reduced form) g grams g/I grams per liter GI gastrointestinal Hb hemoglobin HBr hydrogen bromide HCI hydrogen chloride H20 water HOAc acetic acid HPLC high performance liquid chromatography h hour i.e. (id est) that is I liter X (lambda) wavelength L ligands Me methyl MeOH methanol mg milligram MgSO4 magnesium sulfate min minute(s) ml milliliter (cubic centimeter) mm millimeter mp melting point MS mass spectrometry N normal (equivalent per liter) NaH sodium hydride NaOH sodium hydroxide NH2NH2 hydrazine NH3 ammonia
xi




NMR nuclear magnetic resonance OD optical density OSu N-hydroxysuccinimide p. page(s) PAGE polyacrylamide gel electrophoresis Pd/C palladium carbon Ph phenyl pH acid-base scale: log of reciprocal of hydrogen ion concentrtion
ppm parts per million Pr propyl (normal)
(R) rectus (right) stereodescriptor RNA ribonucleic acid mRNA messenger RNA
(S)- sinister (left) stereodescriptor satd saturated Sec- secondary soln solution tert tertiary TFA trifluoroacetic acid THF tetrahydrofuran Thr threonine TLC thin-layer chromatography Tris tris(hydroxymethyl)-aminomethane, 2-amino-2hydroxymethylpropane-1,3-diol UV ultraviolet wt weight
xii




Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
SYNTHESIS OF RHIZOFERRIN AND POLYAMINE CATECHOLAMIDES
AND THE STUDY OF THEIR EFFECTS ON IRON TRANSPORT IN PARACOCCUS DENITRIFICANS
By
Mei Guo Xin
December 2000
Chairman: Raymond Bergeron
Major Department: Medicinal Chemistry
The synthesis of rhizoferrin, a novel siderophore isolated from Rhizopus microsporus var. rhizopodiformis, is presented. The absolute configurations of two chiral citric acid moieties in this natural product were confirmed to be (R,R) by X-ray crystallography. Also, synthesis of staphyloferrin A, a new siderophore isolated from Staphylococcus hyicus DSM 20459, was completed. The absolute configurations of two chiral citrates in staphyloferrin A were also verified as (R,R). Furthermore, fluviabactin, a polyamine catecholamide isolated from Vibrio fluvials, and its analogs and homologues (L-agrobactin, L-homofluviabactin and D-fluviabactin) were synthesized.
The effects of rhizoferrin and the polyamine catecholamides on iron
transport and growth of Paracoccus denitrificans were studied. Rhizoferrin was unable to promote the bacterial growth when the iron concentration in the medium was low. On the other hand, L-fluviabactin, L-agrobactin, and Lxiii




homofluviabactin derived from L-threonine were able to stimulate the growth rate of Paracoccus denitrificans, whereas D-fluviabactin did not show any promoting effects on bacterial growth. In iron accumulation studies with 55Fe radiolabeled materials, Paracoccus denitrificans acquired 5sFe from sFe ferric L-fluviabactin, ferric L-agrobactin, and ferric L-homofluviabactin much more efficiently than from 5Fe ferric D-fluviabactin. Furthermore, in kinetic studies, all ferric polyamine catecholamide chelators derived from L-threonine displayed both high-affinity and low-affinity components on the Lineweaver-Burk double reciprocal plots, while ferric D-fluviabactin lacked a high-affinity iron transport. The CD spectra of the two chelates, ferric L-fluviabactin and ferric D-fluviabactin, were essentially mirror images. This suggests that it is the change of stereochemistry of oxazoline ring and not the polyamine backbone of the polyamine catecholamides that most affects iron transport in Paracoccus denitrificans, and that the molecular asymmetries of these chelates may contribute to their distinctive kinetic features.
xiv




CHAPTER I
INTRODUCTION AND BACKGROUND The Properties of Iron in Metabolism
Virtually all living cells require iron as a nutrient. It plays critical roles in many biological redox processes. Although iron has a variety of oxidation states, from -2 to +6, it is most simply characterized by the Fe(ll)/Fe(IIl) redox couple. When appropriately coordinated, Fe(II) can easily and reversibly be converted to Fe(lll), and act as an electron donor or acceptor. The main advantage of iron, however, is that in its interaction with coordinating ligands the properties of iron can be modified. Consequently, the ease of electron acceptance can vary over a wide range. There are six molecular orbitals available for a bonding, and these can form bonds along the three axes of symmetry, thus giving the octahedral geometry commonly found in iron complexes (1). The equilibrium between Fe(ll) and Fe(lll), however, is very sensitive to both pH and the nature of the interacting ligands (2). Because the metal serves in a variety of biological redox systems such as cytochromes, oxidases, peroxidases, ribonucleotide reductase (3,4), and various other enzymes as essential cofactors, this sensitivity has been exploited at a cellular level.
Despite its prevalence in biological redox systems, there are numerous problems associated with iron in metabolism. One is related to its solubility
1




2
(5). Iron in early biomolecules was essentially in the Fe(ll) state due to a reducing atmosphere. With the evolution of photosynthesizing blue-green algae, a surplus of oxygen was produced, which resulted in the conversion of Fe(ll) to Fe(lll). This shift in favor of a higher oxidation state had profound effects on the solubility of iron. While both Fe(II) and Fe(Ill) exhibit exceedingly high affinities for hydroxyl ions in aqueous solution, the two oxidation states
2+ 3+
exist as Fe(H20)6 and Fe(H20)6, which readily hydrolyze and polymerize at physiological pH values, forming insoluble complexes. In the case of Fe(lll),
-38
the solubility product may be less than 1038 (6). This translates to a maximum
-18 3+ 2+ solubility of 10- M for Fe(H20)6 In contrast, the solubility of Fe(H20)62+
-1
permits solutions of 10 M. Thus, living systems have had to develop rather sophisticated methods for acquiring iron from ferric hydroxides, the predominant oxidation state prevailing in an aerobic environment. While mammalian cells use proteins, for example, transferrin (7) or ferritin (8), to solve this problem, microorganisms produce a group of relatively low molecular weight, virtually iron-specific ligands, termed siderophores, for the purpose of accessing and utilizing this transition metal (9,10,11).
The main problem with iron metabolism, however, may be related to its toxicity (12). The reaction between iron and oxygen produces hydroxyl radicals and this highly reactive species reacts with biological membranes and combines with a number of natural products to produce carcinogens (13,14). The origins of hydroxyl radical generation lie in the existence of redox cycling iron and the reaction driven by the superoxide anion. The superoxide anion can




3
either be transformed to hydrogen peroxide and oxygen by superoxide dismutase or react with Fe(ll), reducing Fe(ll) to Fe(ll). The problematic species in this chemistry is hydrogen peroxide. Hydrogen peroxide is generally converted to water by glutathione peroxidase or decomposed to water and oxygen by catalase. However, when the level of hydrogen peroxide is above the Km of these enzymes and some Fe(ll) is available, hydrogen peroxide will be converted to the hydroxyl radical. Furthermore, this event can be cyclical, as shown in the following equations:
02" + LFe(III) -4 02 + LFe(ll),
2 02 + 2 H+ H202 + 02,
H+ + LFe(ll) + H202 -- HO' + LFe(lll) + H20.
Figure 1-1. Fenton Reactions and Production of Hydroxyl Radicals.
Hydroxyl radical-related cell damage is seen in inflammatory bowel disease
(15) as well as in certain neurodegenerative disorders, including Alzheimer and Parkinson's (16,17).
While iron is the second most common metal in the earth's crust (5% of total elements), it is not readily extracted from its insoluble oxides by living cells, thus its bioavailability is generally limited and higher species often exhibit




4
deficiency states (18). Paradoxically, iron overload conditions also occur (19). Iron overload disorders have been widely associated with many diseases, including Cooley's anemia (20,21), and often need special chelation therapy (22,23). Considering its relative insolubility, the question has become why does iron overload occur in living cells?
Iron Metabolism in Humans
Eukaryotes have developed large proteinaceous molecules to solve the access, storage, and utilization problems associated with iron. The iron proteins in humans may be broadly grouped as heme proteins, iron flavoproteins, and a heterogeneous group of proteins (24). Iron plays important roles in many biological processes in humans. For example, iron is found at the active centers of iron-binding proteins responsible for oxygen storage and transport (24). Iron is important for electron transport in the heme-containing protein, cytochrome (25). This metal is also found in a large number of nonheme-containing enzymes, including various oxidases, reductases, and dehydrogenases (25).
Iron metabolism in humans is generally divided into different functional compartments. The largest iron compartment in humans is hemoglobin iron, normally containing approximately 67% of the 4 g of iron present in a 70-kg man (25). This heme-containing protein found in erythrocytes is responsible for reversibly binding oxygen in the lung, as well as carrying oxygen to tissues via the circulatory system (25).




5
The iron storage compartment in humans has two distinct forms of
proteins: ferritin and hemosiderin. Ferritin is a water-soluble complex of ferric hydroxide and a protein, apoferritin. Apoferritin forms a shell within which ferric ions, hydroxyl ions, and oxygen are dispersed in a lattice-like relationship (26). Hemosiderin, a water-insoluble iron-storage protein, is found predominantly in cells of the monocyte-macrophage system, which contains approximately 25 to 30 percent iron by weight (8). Under pathologic conditions, it may accumulate in large quantities in almost every tissue of the body (8).
Myoglobin is structurally similar to hemoglobin, but it is monomeric: each myoglobin molecule consists of a heme group surrounded by loops of a long polypeptide chain. Myoglobin is present in small amounts in all skeletal and cardiac muscle cells, and may serve as an oxygen reservoir to protect against cellular injury during periods of oxygen deprivation (25).
Based on its total iron content, normally 3.0 mg (25), the iron transport
compartment in humans is the smallest of all human iron compartments, but it is kinetically the most active, since its iron is replaced at least 10 times every 24 hours (25). This iron is bound to the specific protein transferrin, a somewhat elongated glycoprotein. Transferrin has a molecular weight of 80 kDa, and binds two molecules of ferric ion tightly but reversibly. This iron shuttle protein is found in the serum and is also present in various extracellular fluids in the body (7,27).
In addition to the iron storage protein, ferritin, humans have evolved a highly efficient iron transport and delivery system using transferrin. The




6
regulation of iron mobilization and storage proteins occurs at the translational level. Transferrin carries iron into the cell via the transferrin receptor. The receptor message RNAs of both ferritin and transferrin contain an iron regulatory element in untranslated regions of the messager RNA (28,29). When the iron concentration in cells is low, an iron regulatory protein (30) binds to the ferritin iron regulatory element, and the translation is blocked. However, when an iron regulatory protein binds to the iron regulatory element of transferrin receptor message RNAs, the half-life of the messenger is increased, thus there is a decrease in ferritin synthesis and an increase in the transferrin receptors on the cell surface (28,31).
Iron Overload in Humans
Iron metabolism in primates is characterized by a highly efficient recycling process with no specific mechanism for eliminating and absorbing this transition metal. A healthy adult male has a total of 4 g of iron in his body. Iron absorption is most efficient in the duodenum and becomes progressively less so further along the alimentary canal (32,33). On the other hand, iron excretion is relatively fixed, less than one one-thousandth of the iron absorbed is lost each day (34). Thus, iron balance is actually maintained by the extremely efficient recycling of the iron already present in the body.
Iron requirements are largely determined by the amount lost from the
body. Normally, the small obligatory loss due to desquamation from epithelial surfaces, gastrointestinal blood loss, and biliary excretion amounts to about 1 mg daily or less, and menstrual loss averages about 15-20 mg monthly (19). It




7
is possible to compensate for all physiological losses with a dietary iron intake of 11 mg daily. Iron balance is maintained by physiological adjustments of the absorption mechanism so that the amount of iron crossing the epithelial barrier of small intestines is related to, and regulated by, internal iron status
(35).
Iron overload may be caused by increased iron absorption. For example, there may be an increase in the amount of available dietary iron or increased absorptive activity by the gut epithelium. An increase in iron content of the diet alone may not result in iron loading. Iron overload is often associated with dietary siderosis (18) or a genetic metabolic disorder, for example, idiopathic haemochromatosis (36). Iron overload also may be caused by transfusional iron loading, as seen in patients with severe anemia requiring repeated blood transfusions (37). The iron overload disorders are often classified as primary and secondary hemochromatosis. Primary hemochromatosis refers to an inappropriately increased mucosal absorption and is easily treated by periodic venesection (38).
Secondary hemochromatosis, however, presents a much more
complicated problem. It refers to patients with iron overload secondary to anemia, who require repeated blood transfusions (39). Since the normal lifespan of a red blood cell is only about 120 days, the iron from the transfused red cells is eventually removed by the reticuloendothelial cells of the spleen and stored in the reticuloendothelium or in the parenchyma cells of the liver
(19). However, the efficient iron-recycling system in humans ensures that little




8
of the iron introduced by transfused red cells is ever excreted. The result is an increase in the total amount of iron stored in the body, and eventually iron can reach toxic levels.
The most common anemia of the secondary hemochromatoses is 3thalassemia (39). Patients with this disease require repeated blood transfusions until the iron in their body is overloaded. Most of them die from the toxic effects of iron overload in their second or third decade. Obviously, this excess iron cannot be removed by phlebotomy, as in the case of primary hemochromatosis, because the origin of the excess iron is the transfused red blood cells. Consequently, the only alternative is chelation therapy (39). Desferrioxamine B (40), a siderophore from Streptomyces pilosus, has been used for the treatment of 1-thalassemia for the last 30 years. This drug has a number of shortcomings, such as its marginal oral activity and its poor to moderate efficiency, as well as its high cost. Thus, considerable effort has been invested in the search for alternative iron chelators from siderophores and other synthetic ligands.
Iron Utilization and Transport in Microorganisms
Except for certain members of the lactobacilli, iron is a nutrient required by all microorganisms. Iron plays important roles in many microbial biological processes, such as electron transport, tricarboxylic acid cycle, and DNA biosynthesis. For example, iron is involved in the electron transport metabolism of microorganisms because of the presence of cytochromes and non-heme iron in the respiratory chains of aerobic and anaerobic species. Iron




9
sulfur protein has been considered as fully equivalent to that of the cytochromes in microbial metabolism (41). Iron also plays a critical role in ribonucleotide reductase, the enzyme responsible for synthesis of deoxyribonucleotides required for DNA formation (41,42).
Although iron in its hydroxide polymer is profoundly insoluble, most microbes have developed clever methods to deal with this problem. It is generally accepted that microorganisms utilize two kinds of iron uptake systems, low-affinity systems and high-affinity systems (41), to acquire iron from the environment. Low-affinity systems are only seen when the levels of iron are high. No specific solublizing and transporting compounds or membrane receptors seem to be required. More importantly, microbes have developed high-affinity systems for the assimilation of iron for use when the concentrations of iron are low. High-affinity systems consist of two important parts: siderophores and the transport apparatus.
Siderophores are low molecular weight, iron-specific ligands produced by microorganisms under iron-starved conditions. They form tight but soluble complexes with Fe(lll). More than 100 sideophores have been isolated from bacteria and fungi. Most of these siderophores fall into two basic families: hydroxamates and catecholamides (43). The catecholamides include both cyclic tricatechols as exemplified by enterobactin and many linear polyaminecontaining catecholamides. The hydroxamates consist of a number of structurally different compounds including desferrioxamines, rhodotorulic acid, and citrate-containing ligands. Examples of these two families include




10
hydroxamate desferrioxamine B (40) and catecholamide enterobactin (44) (Figure 1-2).
0 0
OHO
y N N O H NH2
0 0 0 Desferrioxamine B (Streptomyces pilosus)
0 OH
HN :P& OH
OH 0 0 OH
HO N OO N OH H 0 H
Enterobactin (Escherichia coli & Salmonella typhimurium)
Figure 1-2. Examples of Microbial Siderophores.




11
There are also many siderophores that do not belong to either family, for example, rhizoferrin [N1, N4-bis(1-oxo-3-hydroxy-3,4-dicarboxybutyl)diaminobutane] isolated from Rhizopus microsporus var. rhizopodiformis (45) is based on citric acid. Rhizoferrin has a putrescine backbone that is symmetrically diacylated by citric acid at its 1-carboxylate. The configurations of two asymmetric carbons of citrates in rhizoferrin are suggested to be (R, R) by circular dichroism (CD) spectroscopy. Rhizoferrin is neither a hydroxamate, nor a catecholamide (Figure 1-3).
H 0 HO COOH
HOOC N NA N %cOOH
HOOC OH H
Rhizoferrin
(Rhizopus microsporus var. rhizopodiformis)
Figure 1-3. Structure of Rhizoferrin.




12
In the transport apparatus the primary transport sites are the outer
membrane receptor proteins (46). These receptors are induced in a low iron concentration, and are able to recognize the siderophore-iron complex on the surface of the cells. Typically, these proteins are in the molecular mass range of 70 to 90 kDa. For example, in Escherichia coi, an 80-kDa outer membrane ferric enterobactin receptor has been purified and characterized with a Kd of approximately 0.3 pM (47,48). Strong genetic evidence supports the proposed role in iron transport of the ferric enterobactin receptor encoded by the fepA gene, which has been mapped and sequenced (49).
Most iron transport mechanisms in microorganisms are associated with the interactions between siderophores and outer membrane receptors. Studies of iron transport are a useful means to examine the microbial iron utilization facilitated by siderophores and the matching membrane receptors, including the stereospecificity of iron transport mechanisms. A number of methods have been developed for iron transport assays (50,51). For example, the fate of the siderophore-iron complexes can be examined by using double radio-labeled materials. The natural siderophores can be modified by chemical synthesis so that microbial responses to different ligands can be characterized by the chemistry of those ligands. The redox potentials of Fe(ll) and Fe(lll) complexes can be measured, thus a possible reduction mechanism involved in microbial iron transport can be studied. In addition, chelators such as siderophores usually form colorful complexes with iron in solution.




13
Consequently these complexes can be followed by visible spectroscopy, such as UV and circular dichroism.
The study of iron transport mechanisms in microorganisms has become an active field of research in the past two decades. First, iron transport in prokaryotic species such as Escherichia coil provides an ideal opportunity for the application of the powerful techniques of modern molecular biology. Interest in this research has extended beyond iron transport to more general transmembrane transport processes and their regulation. Second, the systems used by microbes to gather iron have stimulated the search for substances that deferrate human patients who suffer from iron overload (52,53) or actinide poisoning (54). Finally, and more importantly, iron has been identified as a virulence factor within selected pathogenic species. The capacity of invading pathogens to acquire enough of this metal to satisfy demands of growth may constitute one aspect of virulence and pathogenicity
(55).
Iron and Bacterial Infection
The role of iron in infection has been an interesting topic for many years. In an early report, a possible association between iron and infection was indicated when Schade and Caroline demonstrated that iron could overcome the bacteriostatic effect of egg white (56). Subsequent studies demonstrated that siderophore production is a common trait among pathogenic microorganisms and suggested that siderophores may be important virulence factors in animal and plant disease (57,58).




14
Virulence factors have been described as substances that are either
directly toxic to the host or antagonize the antibacterial mechanisms of the host. A pathogen does not have to contend with the insolubility of ferric iron, but can obtain its iron from its host. Although all the iron in human plasma is essentially sequestered by such iron-binding proteins as transferrin and ferritin, a pathogen can still successfully compete with those iron-binding proteins by way of siderophore production. It has been assumed that the production of siderophores of either the hydroxamate or catechol type might facilitate microbial growth by removing non-transferrin-bound iron from plasma. For example, it has been showed that the inability to synthesize enterobactin diminishes both the virulence of Salmonella typhimurium in laboratory animals and the capacity of the organism to grow in human serum (59). The presence of enterobactin-specific immunoglobulin was reported in normal human serum
(60). In other examples, when the rabbit pathogen Pasteurella multocida was injected into live organisms, a dramatic fall in plasma iron and elevated body temperature (61) was obtained. The critical temperature sensitivity of siderophore production was also observed in a feverish man (62). It suggested that fever might be a host defense mechanism designed to deprive the pathogen of iron.
Virulence is dependent upon a multitude of interactions between a host and a pathogen. In specific circumstances, iron assimilation may emerge as the critical element in the host-pathogen interplay. While the unambiguous correlation of iron assimilation with virulence remains to be established, the




15
virulence of many pathogenic organisms has been closely associated with their ability to synthesize siderophores. Siderophores apparently antagonize the iron-restricting mechanism of the host, an antibacterial defense mechanism (61,62).
It is not surprising that agents preventing organisms from utilizing iron
also prevent their growth (63). This implies that iron chelators could be used in antimicrobial chemotherapy. Iron deprivation of a microorganism as a result of iron chelation therapy is based on the premise that microorganisms have receptors specific to their own siderophores. Thus modified natural siderophores or other synthetic ligands have emerged as potential candidates.
The Implications and Applications of Siderophores
Siderophores are iron-specific ligands produced by microorganisms in a low iron concentration environment. They have relatively low molecular weights and form very tight complex with iron, for example, the stability constant of ferric enterobactin is as high as 1052. The main function of a siderophore is to form a soluble complex with ferric ion; thus, a microorganism can use a siderophore to obtain this essential metal from the environment. Interestingly, this ability of a siderophore has been exploited in clinical use for the treatment of iron overload secondary to anemia (64-66).
Although substantial efforts have been made to treat iron overload
anemia, subcutaneous infusion of desferrioxamine B (DFO), a hexacoordinate hydroxamate iron chelator produced by Streptomyces pilosus (40) is still regarded as the method of choice for handling transfusional iron overload.




16
However, this drug suffers from a number of shortcomings, such as its high cost of production, poor to moderate efficiency, and marginal oral activity. Furthermore, DFO has a very short half life in the body, and therefore must be administered by continuous subcutaneous infusion over long periods of time. In addition, it can be very immunogenic (67), and patient compliance has become a real problem (68).
The use of an orally effective iron chelator has been a therapeutic strategy for many years. Although a number of synthetic ligands have been studied in recent years as potential orally active therapeutics, for example, pyridoxal isonicotinic hydrazone (69), hydroxypyridones (70) and bis(o-hydroxybenzyl) ethylenediaminediacetic acid analogs (71), none has yet proven to be satisfactory. Interestingly, siderophores, microbial iron chelators, also provide model compounds as orally active iron chelators. One example is 2-(3'hydroxypyrid-2'-yl)-4-methylthiazoline-4(S)-carboxylic acid (desferrithiocin, DFT) isolated from Streptomyces antibioticus (72). Studies in a bile duct-cannulated rat model as well as in a cebus monkey model suggested that it was indeed an orally active iron chelator, but it exhibited nephrotoxicity (73).
Moreover, polyamine catecholamides such as parabactin (74) and
vibriobactin (75) (Figure 1-4) have demonstrated bacteriostatic activities in some bacteria (76). It was concluded that these compounds inhibited bacterial growth because they sequestered iron and the bacteria were unable to use the siderophore-iron complexes (76). Based on this and other studies, iron deprivation has emerged as a strategy in antimicrobial therapy. For example,




17
? O H O~N
H 0 OH
HO Nr ON
OHO0
L-Parabactin (Paracoccus denitrificans)
OH
OH
OH
OH 0 N
N H N.< OH
O 0 OH Vibriobactin (Vibrio cholerae)
Figure 1-4. Structures of L-Parabactin and Vibriobactin.




18
an enantiomeric enterobactin has been synthesized and its bacteriostatic activity demonstrated against Eschericha coli (77). These initial findings prompted the study of the importance of the polyamine backbone and the stereochemistry of the oxazoline ring of a polyamine catecholamide siderophores in microbial iron metabolism.
Of the more than 100 siderophores that have been isolated, their
usefulness as iron-clearing agents has not at all paralleled the rate of their isolation and structural elucidation. For example, among polyamine catecholamides, parabactin, isolated from Paracoccus denitrificans (74), was studied and shown to remove iron from human transferrin in vitro (78). Furthermore, in a bile duct-cannuated rat model, parabactin was shown to be far more efficient than DFO at removing Fe from the animal (64).
N-[2-(2,3-Dihydroxyphenyl)-trans-5-methyl-2-oxazoline-4-yl]carbonyl-N' ,N
bis(2,3-dihydroxybenzoyl)norspermidine (Fluviabactin) is a new siderophores isolated from Vibrio fluvias (79). It is structurally similar to agrobactin, but contains a norspermidine backbone instead of spermidine in agrobactin. (Figure 1-5). Fluviabactin is utilized in iron acquisition by Vibrio fluvias, a pathogen that is widely associated with pediatric enterocolitis (80). V. fluvias is a halophilic, polarly flagellated, gram-negative rod, and grows well on the media supplemented with sodium chloride. Since 1981, 14 cases of enterocolitis associated with V. fluvias have been reported in the United States
(80).




19
OH
OH HO NN l IOH
OHO0 H
L-Agrobactin
(Agrobacterium tumefaciens)
OH
OOH
0" N HO N I K N OH
OH O O OH
L-Fluviabactin (Vibrio fluvialis)
Figure 1-5. Structures of L-Fluviabactin and L-Agrobactin.




20
An interesting phenomenon in siderophore chemistry is that some varieties of siderophores have a citric acid component. For example, the hydroxamates aerobactin, arthrobactin, schizokinen (81), and nannochelin (82) have a symmetrically 1,3-disubstituted citric acid; whereas rhizoferrin (45) and staphyloferrin A (83), isolated from Rhizopus microsporus var. rhizopodiformis and Staphylococcus hyicus, respectively, have two asymmetrically functionalized citric acids (Figure 1-6). This implies that citric acid might play an important role in microbial iron utilization. In addition, the chiral citric acidbased siderophores, for example, rhizoferrin and staphyloferrin A, provide new synthetic challenges for chemical synthesis.
H HOOC H 0 HO COOH
HOOC N COOH
HOOC' OH H
Staphyloferrin A
(Staphylococcus hyicus)
Figure 1-6. Structure of Staphyloferrin A.




21
This present study involves the synthesis of rhizoferrin, staphyloferrin A,
and fluviabactin as well as fluviabactin analogs. The absolute configurations of two chiral citrates in rhizoferrin and staphyloferrin A were confirmed by X-ray crystallography. An efficient synthetic scheme was developed for the synthesis of fluviabactin and its analogues in order to provide enough materials for animal tests in a bile duct-cannuated rat model. The study also includes the iron transport assays of 55Fe radio-labeled compounds in Paracoccus denitrificans, including growth rate, iron accumulation, and kinetic studies, in order to examine the issue of cross utilization of microbial ferric siderophores, including stereospecificity of the iron transport mechanism.




CHAPTER II
SYNTHESIS OF RHIZOFERRIN
Introduction
In response to a low iron concentration microorganisms produce
siderophores, low molecular weight and iron-specific ligands. Siderophores can form water-soluble siderophore-iron complexes. Interestingly, this ability of siderophores has been found to be of clinical use for the treatment of iron overload diseases (64,65). Due to the critical role siderophores play in microbial growth processes and the potential they offer as therapeutic agents, they have received intense attention from chemists (84) and biochemists (85).
A rather large number of siderophores have been isolated. For the most part, they are separated into two basic structural groups: hydroxamates and catecholamides (43). Although these siderophores vary substantially in overall structure, molecules of both classes are usually predicated on their polyamine backbones, specifically 1,4-diaminobutane (putrescine), 1,5-diaminopentane (cadaverine), norspermidine or spermidine, or on their biochemical precursor ornithine or lysine. The presence of such common structural units has led to the efficient total synthesis of the siderophores parabactin (86) and desferrioxamine (DFO) (84) in this and other laboratories.
There are additional groups of siderophores that do not belong to either of these two major families. Examples include a variety of citrate-based
22




23
siderophores. Citrate in those siderophores is either symmetrically disubstituted, as seen in aerobactin, arthrobactin, schizokinen (81), and nannochelin (82), or unsymmetrically monosubstituted, as in rhizoferrin (45) and staphyloferrin A (83). The latter case is somewhat complicated since the prochiral carbon of the citrate became asymmetric because of the acylation, thus giving rise to new synthetic challenges in the synthesis of these natural products.
N1 ,N4-Bis(1-oxo-3-hydroxy-3,4-dicarboxybutyl)diaminobutane (rhizoferrin) has a putrescine center symmetrically diacylated by citric acid at its 1carboxylate. It is a hydroxyl polycarboxylate along with rhizobactin (87) and staphyloferrin A, which are predicated on L-lysine and D-ornithine, respectively. The configurations of two asymmetric carbons of the citrates in rhizoferrin are suggested to be (R,R) by comparing its circular dichroism (CD) spectroscopy with that of natural (R,R)-tartaric acid (88).
Rhizoferrin was first isolated from Rhizopus microsporus var.
rhizopodiformis, an organism associated with mucormycosis seen in dialysis patients, and occurs in several Zygomycete strains of fungi (89). Like the natural chelators parabactin and DFO, rhizoferrin forms a 1:1 complex with ferric ion (88), however, the formation constant of ferric rhizoferrin has not been measured, and the absolute configurations of the two chiral citrate centers in rhizoferrin were not yet confirmed. The first total synthesis of this natural product could also be the first example of the conversion of a chiral citric acid fragment to a chelator.




24
Retrosynthetic analysis revealed there are two components in rhizoferrin: putrescine and chiral citrate (Figure 2-1).
H o HO COOH HOOC N N )L -COOH
HOOC" OH O H
H
Bn'N NBn
H
H3COOC0H 0 HO COOCH3
H3COO~ OH HOJj -,COOCH3
H2N NH2
H3COOC OCH3 H3COOC OCH3
H3COOC OH O H3COOC OH O
Figure 2-1. Retrosynthetic Scheme of Rhizoferrin.




25
In this synthesis, putrescine has to be converted to N,N'-dibenzylputrescine by condensation of putrescine with benzaldehyde, followed by reduction with sodium borohydride. The principal challenge to the synthesis of rhizoferrin, however, was to access a citrate synthon of correct configuration for coupling to both termini of putrescine in order to unequivocally define the absolute configuration of the siderophore. Since citric acid is a prochiral molecule, it is necessary to convert citric acid into an appropriate 1,2-disubstituted citric acid. Although 1,2-dibenzyl citrate was considered and synthesized by reacting citric acid with benzyl alcohol and then kinetic hydrolysis with sodium hydroxide, it does not form crystals with natural bases, for example, brucine. Consequently, 1,2-dimethyl citrate was utilized.
Synthesis (90)
The synthesis of rhizoferrin began with trimethyl citrate (3), which was
converted to 1,2-dimethyl citrate (~) by a sterically controlled saponification (91). The enantiomers of carboxylic acid (2_) were converted to their (-)-brucine diastereomeric salts. After five fractional crystallizations from water, the crystalline salt was shown by single crystal X-ray diffraction to contain 1,2dimethyl citrate in the R-configuration (Figure 2-2). Treatment of the salt with 1 N HCI and extraction with ethyl acetate furnished (R)-1,2-dimethyl citrate (3).
The determination of the enantiomeric purity of a chiral citrate is another important issue in this synthesis. Both the racemic (2_) and chiral (Q) diesters were used to acylate (S)-(-)-sec-phenethyl alcohol (1,3-dicyclohexylcarbodiimide/catalytic DMAP/CH2CI2) to give unsymmetrical triesters (4) and (.5),




26
respectively. An examination of the methyl ester region in the 600 MHz NMR spectrum of (4) and (5) showed that the latter contained an enantiomeric excess (ee) of 99%.
C5
04B
04A
C8 07B C4 03B C1 5C
4 03A C7 -0
C2
C6 01 07A
Figure 2-2. X-Ray of R-Enantiomer of 1,2-Dimethyl Citric Acid.




27
With the correct enantiomeric acid in hand, N1,N4-dibenzyl-1,4diaminobutane (92) was acylated with (3) (2 equivalents) utilizing diphenylphosphoryl azide (Et3N/DMF) (93). The diamide (6) was obtained in 26% yield after flash column chromatography, which removed by-products, including olefins due to elimination of the tertiary alcohol as indicated by 1H NMR. The methyl esters of (6) were hydrolyzed with sodium hydroxide in aqueous methanol; acidification yielded N',N4-dibenzyl rhizoferrin (7).
Finally, since N-benzyl amides are resistant to hydrogenolysis (94), deprotection of tetraacid (7) under dissolving metal reduction conditions (Li/NH3/THF) (95), protonation of the salts on a cation exchange resin column and purification on a C-18 reversed-phase column furnished the final product, rhizoferrin (Figure 2-3). The high-field 1H NMR and high-resolution mass spectrum of this synthetic compound were essentially identical to the published spectra of the natural product (89). The absolute configurations (R,R) of the synthetic sample and the natural material were identical, since both exhibited a negative Cotton effect at the same wavelength (88). It is fortunate that the correct crystalline diastereomeric (-)-brucine salt of 1,2dimethyl citrate crystallized out and that no subsequent synthetic step could compromise the stereochemical integrity of the chiral centers.
Since rhizoferrin can be considered an aspartic acid derivative that usually cyclize in peptide synthesis (96), the polycarboxylic acid side chain of rhizoferrin could catalyze the spontaneous cyclization process. Indeed, rhizoferrin cyclizes upon standing through dehydration to imidorhizoferrin and




28
H3COOC OH
H3COOCX.,,COOCH3 ()
a
(2)R=H
H3COOC OH (2)
H3COOCQ COOR (4) R = Ph b
Sc,d,e
H3C(OC.) R = H
H3COOC COOR (5) R H CPh b
PhPh
r OHO COOR (6) R=Me
ROOC-. I N N U COOR ( g
ROOC' OH h (7) R = H
Figure 2-3. Synthesis of Rhizoferrin. (a) NaOH/MeOH (39%);
(b) (S)-(-)-sec-phenethyl alcohol/DCC/catalytic DMAP/CH2CI2;
(c) (-)-brucine; (d) fractional crystallization; (e) HCI;
(f) N',N4-dibenzyl-1,4-diaminobutane/DPPNA/Et3N/DMF (26%);
(g) NaOH/MeOH (77%); (h) Li/NH3/THF (64%)




29
bis-imidorhizoferrin which possess one or two five-membered rings, respectively. We observed by 1H NMR spectroscopy that the zero order rate constant for this ring formation at pH 5.0 is 6.9 x 10-2 h. At pH 3 our findings on the extent of cyclization were similar to the literature (88); thus the analytical data were obtained before this cyclization occurred.
The synthetic methodology for rhizoferrin will also be used to prepare the hydroxy polycarboxylated siderophore staphyloferrin A, in which D-ornithine is Nc,N5-diacylated with citric acid at its 1-carboxylate. Thus, the configuration of the citrates in this amino acid chelator will be determined by total synthesis as presented in chapter III.
Experimental
Trimethyl citrate (1) was obtained from CTC Organics, Atlanta, GA. Other reagents were purchased from Aldrich Chemical Company, and were used as received. Fisher Optima grade solvents were routinely employed. Silica gel 32-63 (40 jM "flash") from Selecto, Inc. (Kennesaw, GA) or silica gel 60 (70-230 mesh) from EM Science (Darmstadt, Germany) was used for column chromatography. Optical rotations were run in CH30H at 589 nm (Na lamp) at room temperature with c as g of compound per 100 mL.
H NMR spectra were recorded at 300 or 600 MHz and run in the
deuterated organic solvent indicated or in D20 with chemical shifts given in parts per million downfield from tetramethylsilane or 3-(trimethylsilyl)propionic2,2,3,3-d4 acid, sodium salt, respectively.




30
X-Ray Diffraction
Data were collected at 173 K on a Siemens SMART PLATFORM equipped with a CCD area detector and a graphite monochromator utilizing MoKa radiation (X = 0.71073 A). Cell parameters were refined using up to 6233 reflections. A hemisphere of data (1381 frames) was collected using the coscan method (0.30 frame width). The first 50 frames were remeasured at the end of data collection to monitor instrument and crystal stability (maximum correction on I was < 1%). Psi scan absorption corrections were applied based on the entire data set.
Circular Dichroism
CD spectra were obtained with a Jasco Model J500C spectropolarimeter equipped with a Jasco IF-50011 interface and CompuAdd 286 computer; data collection and processing were performed with Jasco DP-500/PC System version 1.28 software. The cell path length was 2.00 cm. Ultraviolet Spectroscopy
UV spectra were obtained with a Shimadzu UV-2501PC equipped with an AST 486/33 computer data station. The cell path length was 1.00 cm.
(RS) 1,2-Dimethyl Citrate (2)
Compound (2_) was prepared by a published method (91). Sodium
hydroxide (0.1 N, 215 mL) was added to a solution of trimethyl citrate (10.0 g, 42.7 mmol) in 50% aqueous CH3OH (200 mL) over 2 hours with vigorous stirring at room temperature. The solution was concentrated to about 150 mL and extracted with EtOAc (3 x 150 mL). The aqueous layer was acidified with 1




31
N HCI (45 mL) and extracted with EtOAc (3 x 150 mL). The organic layer was dried (MgSO4) and concentrated, providing 3.70 g (39%) of (2) as a colorless oil: 1H NMR (d6-DMSO) 8 5.60 (br s, 1 H, OH), 3.64 (s, 3 H, CO2CH3), 3.57 (s, 3 H, CO2CH3), 2.87 (d, 1 H, J = 15 Hz, 1/2 CH2), 2.81 (d, 1 H, J = 15 Hz, 1/2 CH2),
2.73 (d, 1 H, J = 15 Hz, 1/2 OH2), 2.65 (d, 1 H, J = 15 Hz, 1/2 CH2).
(-)-Brucine Salt of (R)-1,2-Dimethyl Citrate
To a solution of (-)-brucine (12.5 g, 31.8 mmol) (CAUTION: toxic) in EtOAc (460 mL) was added (2_) (7 g, 31.8 mmol) with vigorous stirring overnight. After filtration the precipitate (10.5 g) was recrystallized from water (5 x) and dried to afford 2.04 g of white crystals: mp 165-168 OC.
The diastereomeric salt crystallizes in the monoclinic space group C2
and has cell dimensions: a = 13.8947 (3), b = 12.4224 (3), and c = 17.5408 (3) A; a = 900, 13 = 104.5560 (1), and y = 900. The structure was solved by the Direct Methods in SHELXTL20 and was refined using full matrix least squares. The non-H atoms were treated anisotropically. The methyl hydrogen atoms were calculated in ideal positions and were riding on their respective carbon atoms; the rest of the H atoms were refined without constraints. Two water molecules were located in the asymmetric unit. One was refined with full occupancy, and its H atoms were located. The other, located on a 2-fold axis of rotation, was refined to a 30% occupancy. An absolute configuration of (R) was assigned to the citrate portion of the salt based on knowledge of the stereochemistry of brucine. Parameters (521) were refined in the final cycle of refinement using




32
3855 reflections with I > 2a(l) to yield R and wR of 0.0434 and 0.1040,
2
respectively. Refinement was done using F2
(R)-1,2-Dimethyl Citrate (3)
HCI (1 N, 4 mL) was added to a solution of the (-)-brucine salt of (R)-1,2dimethyl citrate (2.04 g, 3.32 mmol) in water (50 mL), and stirring was continued for 5 min. Extraction with EtOAc (3 x 50 mL), drying over Na2SO4, and concentration gave 630 mg (86%) of (3) as a colorless oil: [a] +4.0 (c 1.00); the NMR was identical to (2).
(R. S)1,2-Dimethyl-3-[(S)-sec-Phenethyl] Citrate (4)
1,3-Dicyclohexylcarbodiimide (103 mg, 0.5 mmol) was added to a
solution of (2) (110 mg, 0.5 mmol), (S)-(-)-sec-phenethyl alcohol (61 mg, 0.5 mmol), and 4-dimethylaminopyridine (3 mg) in dry CH2CI2 (10 mL) at 0 oC, and the mixture was stirred overnight. The mixture was filtered, and the filtrate was concentrated and purified by flash chromatography (1:2 EtOAc/hexane)
1
resulting in 60 mg (37%) of (4) as a colorless oil: H NMR (CDCl3) 8 7.35-7.28 (m, Ph), 5.97 (q, J = 7 Hz, CHPh), 5.88 (q, J = 7 Hz, CHPh), 3.77 (s, CH30), 3.73 (s, CH30), 3.69 (s, CH30), 3.68 (s, CH30), 2.98-2.74 (m, CH2), 1.54 (d, J = 7 Hz, C-CH3), 1.52 (d, J = 7 Hz, C-CH3).
(R)-1,2-Dimethyl-3-f(S)-sec-Phenethyll Citrate (5)
Esterification of (3) with (S)-(-)-sec-phenethyl alcohol by the method of (4) gave (5). The ratio of CH30 peaks 8 3.77 and 3.69 to 83.73 and 3.68 in the 600




33
1
MHz H NMR (CDCl3) showed an ee of more than 99% based on the intergrations.
N1, N4-Dibenzyl Rhizoferrin, Tetramethyl Ester (6)
Diphenylphosphoryl azide (760 mg, 2.76 mmol) and NEt3 (1.5 mL, 11 mmol) were added to a solution of (3) (610 mg, 2.77 mmol) and N1,N4dibenzyl-1,4-diaminobutane (370 mg, 1.38 mmol) in DMF (20 mL) at 0 OC under nitrogen. The solution was stirred at 0 oC for 1 h and then at room temperature for 23 h. After solvents were removed under high vaccum, the residue was taken up in EtOAc (25 mL) and was washed with saturated NaHCO3 (25 mL), water (25 mL), 0.5 N HCI (25 mL), and water (25 mL). The organic layer was dried (MgSO4) and concentrated. Flash chromatography, eluting with 4:1 EtOAc/hexane, generated 240 mg (26%) of (5) as a pale yellow oil: [a] +8.25 (c
1.00); 'H NMR (CDCI3) 8 7.42-7.24 (m, 10 H), 4.65-4.48 (m, 4 H), 3.81 (s, 3 H, OCH3), 3.79 (s, 3 H, OCH3), 3.69 (s, 3 H, OCH3), 3.65 (s, 3 H, OCH3), 3.40-3.12 (m, 4 H), 3.10-2.67 (m, 8 H), 1.57-1.41 (m, 4 H).
Anal. Calcd. for C34H44N2012: C, 60.70; H, 6.59; N, 4.16. Found: C, 60.64; H, 6.61; N, 4.15.
N, N'-Dibenzyl Rhizoferrin (7)
A solution of (6) (170 mg, 0.253 mmol) in CH3OH (7 mL) and 1 N NaOH (7 mL) was stirred at room temperature for 5 h. HCI (1 N, 8 mL) was added, and the solution was concentrated to about 15 mL. After extraction with EtOAc (3 x 15 mL), the organic layer was dried (Na2SO4) and concentrated to give 120 mg




34
(77%) of (7) as a colorless glass: [a] +12.27 (c 1.00); 1H NMR (CD3OD) 8 7.427.20 (m, 10 H, 2 Ph), 4.67-4.47 (m, 4 H, CH2Ph), 3.35-3.23 (m, 4 H, 2 NCH2),
3.19-2.69 (m, 8 H, 4 CH2CO), 1.58-1.41 (m, 4 H, 2 CH2).
Anal. Calcd. for C30H36N2012H20: C, 56.78; H, 6.04; N, 4.41. Found: C, 56.88; H, 6.08; N, 4.34.
Rhizoferrin
A solution of (7) (110 mg, 0.178 mmol) in distilled THF (1.5 mL) was added to Li (33 mg, 4.8 mmol) in NH3 (100 mL), and the mixture was maintained at -780C for 3 h. Aqueous CH3OH (50%, 10 mL) was added until the blue color disappeared. Ammonia was evaporated, and the residue was taken up in water (50 mL) and concentrated to dryness. The colorless residue was dissolved in water and filtered through a cation exchange resin column (Bio Rad, AG 50W-X8). The eluant-containing product (pH = 3) was extracted with EtOAc (50 mL), which was concentrated to dryness. The residue was dissolved in distilled EtOH (2 mL), filtered, and concentrated to yield 50 mg (64%) of rhizoferrin as a colorless glass: HRMS (FAB, m-nitrobenzyl alcohol matrix) calcd. for 016H25N2012 437.1407 (M + H), found 437.1407 (base).
Anal. Calcd. for C16H25N2012 H20: C, 42.29; H, 5.77; N, 6.17. Found: C, 42.49; H, 5.80; N, 5.84.
A solution of crude product (10 mg) was purified by reversed-phase HPLC (C-18 preparative column, 21.4 mm x 25 cm, obtained from Rainin). The initial mobile phase concentration of 3% CH3CN in 0.1% TFA was held for 15 min,




35
followed by gradient elution of 3-11% CH3CN in 0.1% TFA over 35 min, then held at 11% CH3CN in 0.1% TFA for 20 min. Flow rate was maintained at 4 mL/min. Retention time was 56 min. Lyophilization gave 4.32 mg (9.90 mmol) of purified rhizoferrin as a colorless glass: [a] -16.7 (26 oC) (c 0.1613); H NMR (D20) 8 3.21-3.15 (m, 4 H), 3.02 (d, 2 H, J = 16.0 Hz), 2.79 (d, 2 H, J = 16.0 Hz),
2.76 (d, 2 H, J= 14.6 Hz), 2.65 (d, 2 H, J= 14.6 Hz), 1.53-1.47 (m, 4 H).
A stock solution was prepared by dissolving the purified product in 50.00 mL distilled water; a 10.00 mL aliquot was diluted to 20.00 mL and adjusted to pH = 3.02 with 1.90 mL of 0.010 N HCI (final rhizoferrin concentration = 9.04 x 10-5 M). CD and UV spectra were taken immediately after pH adjustment. All spectra were baseline corrected with distilled water blank, which was acidified as above.
CD Results
The CD spectrum of rhizoferrin exhibited a negative Cotton effect from 200 to 220 nm, with a single minimum at 205 nm, Ae= -2.7 compared to a recorded single minimum at 204 nm, As = -4.3 (88). UV Results
nm C e (88)
196 12200 (13900) 200 10800 (13150) 210 5230 (5600) 215 2770 (3000) 220 1200 (1400)




36
Discussion
The main problem in this synthesis is the separation of 1,2-dimethyl citrate enantiomers. There are several ways to obtain resolution of enantiomeric carboxylic acids (97). Among these are biological or enzymatic resolution of the racemic acid or its ester (97). The most primitive form of biological resolution is to grow microorganisms in the presence of the racemic acid or ester. Depending on the choice of microorganism, one or the other of the enatiomers will be utilized. This is because the chiral compound that reacts at different rates with the two enantiomers may be present in a living organism. For instance, a certain esterase may cleave an ester of one enatiomer but not the other. Because it is necessary to find the proper organism and since one of the enantiomers is destroyed in the process, this method is limited. When the proper organism is found, however, a good resolution can be achieved since biological processes are highly stereoselective. A more sophisticated approach utilizes highly purified enzymes from such microorganisms. Lipases are especially useful enzymes in organic synthesis because of their stability in organic solvents, the fact that they can accept a wide range of substrates, and their availablity (98). For example, about 20 different lipases from microbial, plant, and animal sources are commercially available. Lipases are characterized by the involvement of a lipid-water interface in their catalytic process. This unique feature of interfacial catalysis provides the lipases with an inherent affinity for hydrophobic environments and distinguishes them from other hydrolytic enzymes (98).




37
Lipase-catalyzed reactions include esterification, transesterification, amidation, peptide synthesis, and macrocyclic lactone formation (98). Of these, enantioselective ester synthesis/interchange is of particular interest because it provides a facile method for the preparation of optically active acids.
Interestingly, lipases have been utilized as catalysts for the hydrolysis of triesters of citric acid. The triethyl or trimethyl ester of citric acid was hydolysed with porcine liver esterase (PLE) or the proteinase subtilisin forming a mixture of the two diesters (99). It was evident that there was a slight preference for hydrolysis of the ester in the mid-position. The hydrolysis of the trimethyl ester of citric acid was not regioselective. However, hydrolysis of triethyl ester of citric acid yields only one product, the symmetric diester (Figure 2-4).
CH2CO2R CH2CO2H CH2CO2R
I I I
C(OH)CO2R PLE or C(OH)CO2R + C(OH)CO2H I Subtilisin I I
CH2CO2R CH2CO2R CH2CO2R
R = CH3 or CH2CH3
Figure 2-4. Hydrolysis of Triesters of Citric Acid with PLE or Subtilisin.




38
In this project, a number of lipases (lipase type VII from Candida rugosa and type Xl from Rhizopus arrhizus) were tested for their abilities to stereoselectively cleave trimethyl, triethyl, triheptyl and tribenzyl esters of citric acid. To each 2 mL vial was added 900 pL phosphorate buffer solution, 100 pM trimethyl citrate ester, 100 units lipase, and 100 pL various organic solvents. The solvents used in the tests included acetone, acetonitrile, butanol, butylacetate, tert-butylmethyl ether, chloroform, cyclohexane, DMF, dioxane, ethanol, ether, ethyl acetate, n-heptane, n-hexane, methanol, methylene chloride, methyl acetate, octanol, n-propyl ether, THF, and toluene. The mixtures were stirred overnight and the reactions were checked by TLC. New compounds were only found in the reactions that contained dioxane and they were not 1,2-dimethyl citric acid as indicated by TLC. Further attempts to obtain 1,2-dimethyl citric acid enatiomers using enzymatic means was abandoned.
Another means to resolve a racemic carboxylic acid is by fractional
crystallization of its diastereomeric salt with a chiral base. The most frequently used chiral bases include brucine, ephedrine, strychnine, and morphine. The base, brucine, was used in this project. If, for the sake of argument, the base is in the 'S' form, the two salts produced will have the configurations SS and RS. The acids are enantiomers, however, the salts are diastereomers and have different properties (Figure 2-5).
The property of the diastereomer salts most often used for separation is differential solubility. The mixture of 1,2-dimethyl citric acid-brucine salts was allowed to crystallize from water. Because the solubilities are different, the




39
initial crystals formed will be richer in one diastereomer. In this instance, the (R, S) 1,2-dimethyl citrate-brucine salt dominated in the crystal form. Filtration at this point resulted a partial resolution. The total separation of the diastereomers was achieved by several such fractional crystallization steps.
CH2CO2CH3 CH2CO2CH3
HO- -CO2CH3 HO- -CO2CH3
CH2CO2H CH2CO2 brucine-H'
S S S
+ S+ S-Brucine +
CH2CO2CH3 CH2CO2CH3 H3CO2C- -OH H3CO2C- -OH
CH2CO2H CH2CO2 brucine-H+
R R S
Figure 2-5. The Diastereomeric Salts Formed Between
1,2-Dimethyl Citrate and Brucine.




CHAPTER III
SYNTHESIS OF STAPHYLOFERRIN A Introduction
Citric acid is a component in many siderophores produced by
microorganisms. The citrate-based siderophores can be divided into two groups according to the stereochemistry of citrate in these molecules. The first group consists of aerobactin, arthrobactin, schizokinen (81), and nannochelin
(82), in which citric acid acylates same functional groups via its lor 3 carboxylic acids. Thus, citrate is still prochiral in these siderophores. Examples in the second group include rhizoferrin presented in Chaper II, and staphyloferrin A described in this chapter. Both of these siderophores contain two citric acid residues. Each terminal carboxylic acid of these two citric acids forms a bond with an amino group in putrescine or D-ornithine to give rhizoferrin and staphyloferrin A, respectively. Consequently, the citric acids are asymmetric in these two siderophores. The absolute configurations of the two asymmetric citrates in rhizoferrin were confirmed to be (R,R) in Chapter II.
2 5
N 2 N5-Bis(1-oxo-3-hydroxy-3,4-dicarboxybutyl)-D-ornithine (staphyloferrin A) was isolated from cultures of Staphylococcus hyicus DSM 20459. Staphyloferrin A has one ornithine backbone and two citric acid residues linked by two amide bonds (83). The natural product was obtained as a colorless, acidic compound, which is soluble in water and water/methanol mixtures
40




41
containing up to 80% methanol and virtually insoluble in all common organic solvents. It is stable in neutral aqueous solution over the range 4-800C, and stable up to pH 10 at room temperature (100). However, under acidic conditions, staphyloferrin A was degraded.
The stereochemistry of D-ornithine was determined by amino acid
analysis of the staphyloferrin A acid hydrolysate and by gas chromatography of the N-trifluoroacetylated, n-propyl ester amino acid derivatives on a Chirasil-Val chiral HPLC support (100). However, the configurations of the two chiral citric acid residues were undetermined. The only evidence bearing on the stereochemistry of the two chiral citric acids is the CD spectrum of ferric staphyloferrin A with one positive cotton effect at 350 nm (As = 0.24 M"'cm1) and two negative effects at 302 nm (As = -0.46 M'1cm') and 250 (As = -0.28 M'cm-) nm (100).
The goals in this synthesis are to complete the synthesis of staphyloferrin A and to confirm the absolute configurations of the two chiral citric acid residues in this natural product. As previously mentioned, (R)-1,2-dimethyl citrate developed in the synthesis of rhizoferrin can be used in this synthesis. The stereochemistry of the two chiral citrates was assumed to be (R,R) and could then be verified by the CD spectrum of synthetic ferric staphyloferrin A, as compared to that of the natural ferric siderophore.
Synthesis
While staphyloferrin A has an amino acid, D-ornithine, instead of
putrescine as in rhizoferrin, it is almost impossible to use the synthetic strategy




42
of rhizoferrin for the synthesis of staphyloferrin A, because synthesis of dibenzylated D-ornithine is impractical. Furthermore, the reactivities of the two amines in D-ornithine are different. The a-amine can be acylated with a chiral 1,2-dimethyl citrate in the presence of DCC/DMAP, while attempts to acylate the 8-amine with the same reagents failed. Therefore, a different synthetic scheme was developed for this synthesis.
The total synthesis of staphyloferrin A started with the Cbz protected Dornithine. The carboxylic acid of di-Cbz-D-ornithine was protected as its tertbutyl ester by using HCIO4 in tert-butyl acetate in low yield (23%). Next, the CBZ protecting groups in ester (8) were removed by 10% Pd/C in hydrogen, affording compound (9).
(R)-1,2-Dimethyl citrate (3) generated by the method described in Chapter II, was reacted with pentachlorophenol in the presence of DCC/DMAP to form an active ester (10), which was reacted with compound (9) to provide (11) in 66% yield. Finally, the tert-butyl ester of (1.) was removed by TFA in methylene chloride giving (12), followed by hydrolysis of the methyl esters in NaOH/MeOH to furnish staphyloferrin A (13) as its sodium salt (Figure 3-1).
The final product was only partially epimerized as determined by comparing the 1H NMR spectrum of the D-ornithine methine peaks of compound (12) to that of compound (13). There are two types of peaks matching the methine hydrogen in the 'H NMR spectrum of compound (13) instead of one type in that of compound (12). The ratio between these two




43
types of peaks appeared to be sensitive to the basicity of the solvent and the reaction time. Although different bases, for example, NaOH, KOH, and LiOH,
COO-t-Bu
COOH H
%- a RI-N -~
Cbz-HN NH-Cbz aR-N NN-R
--,- NH-Cbz H
(8), R = Cbz b
(9), R = H
CI Cl
H3COOC OH c H3COOC O -CI H3CO H (H H3COOCOH- o C
(Q) (10)
H R2-OOC 0 HO COOR3 d, e, f R3OoC N N COOR3
de R300 OH 0 H
(11), R2 = tBu; R3= CH3
(12), R2= H; R3= CH3
(13), R2 = H; R3 = H
Figure 3-1. Synthesis of Staphyloferrin A. a) t-Butyl acetate/HCIO4; b) Pd/C/HCIIH2; c) DCC/DMAP/Pentachlorophenol;d) Et3N/(2); e) TFA; f) NaOH (ee 80%).
were tested, partial racemization still resulted. Nonetheless, the racemization did not affect the CD spectrum of the ferric complex or adduct with the final product (368 nm, As = 0.31 M'1cm'; 316 nm, As = -0.60 M1cm1; 266nm As = -




44
0.16 M-'cm-1) nm, which is similar to that of the ferric complex or adduct natural product. Therefore, the absolute configurations of the two chiral citrate in staphyloferrin A are suggested to be (R, R) as in rhizoferrin.
Experimental
General Synthesis (see synthesis of rhizoferrin) Circular Dichroism
The CD spectra were obtained with a Jasco Model J710
spectropolarimeter equipped with a Jasco IF-710 interface. Cell path length was 0.01 cm, and the concentration of ferric staphyloferrin A was 21 mM in water (pH = 5.6).
Synthesis of NN -Bis(benzyloxycarbonyl)-D-ornithine-tert-butyl Ester (8)
A mixture of perchloric acid (1 ml of a 6% solution), N,Nbis(benzyloxycarbonyl)-D-ornithine (2.4 g, 6.0 mmol) (Bachem) in 10 ml CH2CI2, and 10 ml of tert-butyl acetate was stirred for two days. The solvent was removed, and the residue was dissolved in ethyl acetate (20 mL) and washed with saturated aqueous NaHCO3 (20 mL) and brine (20 mL). The organic layer was dried over MgSO4, filtered, and evaporated, affording L8) (0.63 g, 23%) as a colorless oil: [a] = -8.1 (c = 0.67 in CHCI3); 1H-NMR (CD3OD) 5 1.43 (s, 9H, t-Buorn), 1.5-1.82 (m, 4 H, 2 x CH2-orn), 3.10-3.14 (m, 2H, CH2-N), 4.01-4.11 (m,
1H, CH-orn), 5.05 (s, 2H, CH2-Ph), 5.08 (s, 2H, CH2-Ph), 7.31-7.42 (m, 10OH, Ph); Anal. Calcd. for C25H32N206: C, 65.77; H, 7.07; N, 6.14; found C 65.90; H
7.03; N 6.23.
Synthesis of D-Ornithine-tert-butyl Ester Dihydrochloride (9)
To a 100 mL flask containing (8) (0.62 g, 1.36 mmol) in 20 ml of ethanol
was added 10% Pd/C (80 mg ) and 1 N HCI (2.7 ml ) cautiously. The flask was degassed with N2 by three times, then filled with H2. The mixture was stirred for 1 h, filtered through celite, and concentrated to give (9) (0.34 g, 95%) as a white




45
solid: mp 163-165C; [a]24 = -9.84 (c = 0.028 in MeOH); 1H NMR, (CD3OD) 6
1.55 (s, 9H, t-butyl), 1.70-2.01 (m, 4H, 2 x CH2-orn), 3.01-3.10 (m, 2H, N-CH2), 3.90-3.96 (m, 1H CH-orn); Anal. Calcd. for C9H22 C12N202: C, 41.39; H, 8.49; N 10.73. found: C, 41.59; H, 8.31; N, 10.53. Synthesis of (S)-1,2-Dimethyl-3-pentachlorophenyl citrate (10)
A mixture of 1,3-dicyclohexylcarbodiimide (0.82 g, 3.99 mmol), (3) (0.88 g,
3.99 mmol), pentachlorophenol (1.06 g, 3.99 mmol), and 4-dimethylaminopyridine (5 mg) in CH2C12 (60 mL) was stirred overnight at 0 OC. The mixture was filtered. Column chromatography of the residue with EtOAc/hexane (7:3) gave (Q) (1.4 g, 75%) as a white solid (mp 97-990C): [a] = -9.2 (c = 0.5 in CHCl3); 1H-NMR (CDCI3) 8 2.90 (d, J= 15.5, CH2), 2.98 (d, J= 15.6, CH2), 3.23 (d, J = 16.4, CH2), 3.3 (d, J = 16.4, CH2), 3.73 (s, CH30), 3.85 (s, CH30); Anal. Calcd. for C14H1C1507: C, 35.89; H, 2.37; found: C, 36.13; H 2.38. Synthesis of N, N-1,2-(R)-Dimethyl Citric Acid-D-ornithine-t-butyl Ester (11)
A mixture of (9) (0.33 g, 1.27 mmol) in DMF (20 ml), Et3N (0.25 g, 2.5
mmol), and (10) (1.4 g, 3.0 mmol) was stirred overnight at room temperature. Solvents were removed in vacuo. Flash chromatography of the residue with EtOAc gave (11) (0.49 g, 65%) as a colorless oil (ee, 99%): [a] = 3.2 (c = 0.025, MeOH); 'H-NMR (CDC3) 8 1.56 (s, 9H, t-Bu), 1.52-1.83 (m, 4H, 2 x CH2-orn),
2.60-3.00 (m, 8H, 4 x CH2-citri), 3.15-3.18 (m, 2H, N-CH2-orn), 3.65 (s, 3H, OCH3), 3.66 (s, 3H, OCH3), 3.76 (s, 6H, 2 x OCH3), 4.21-4.23 (m, 1H, CH-orn); Anal. Calcd. for 025H40N2014: C, 50.67; H, 6.80; N 4.73; found: C, 50.60; H, 6.75; N, 4.70.




46
Synthesis of N, N-1,2-(R)-Dimethyl Citric Acid-D-ornithine (12)
A mixture of trifluoroacetic acid (8 ml) and (j1) (0.47 g, 0.79 mmol) was stirred at room temperature for 30 min. Evaporation of excess TFA in vacuo afforded (12_) (0.42 g, 99%) as an oil: [ax] = -9.1 (c = 0.14, MeOH); 'H-NMR (CD3OD) 5 1.52-1.90 (m, 4H, 2 x CH2-orn), 2.61-3.03 (m, 8H, 4 x CH2-citric), 3.16-3.19 (m, 2H, N-CH2-orn), 3.65 (s, 6H, 2 x OCH3), 3.76 (s, 6H, 2 x OCH3), 4.35-4.37 (m, 1H, CH-orn); Anal. Calcd. For 012H32N2014: C, 47.02; H, 6.01; N,
5.22; found: C, 46.76; H, 5.89; N, 5.13. N2. N5-(RR)-Bis(1 -oxo-3-hydroxy-3,4-dicarboxybutyl)-D-ornithine (Staphyloferrin A) (13)
To a solution of (12) (0.42 g, 0.78 mmol) in 10 ml of MeOH/H20 (1:1) was added sodium hydroxide (10 ml, 1N). The mixture was stirred at room temperature for 3 h, and then passed through a cation exchange resin column (18/3, Bio Rad, AG 50W-X8, molecular biology grade, 63-150 gm wet bead size, 200-400 dry mesh size, ammonium form). The resulting solution was lyophilized, providing (13) (0.43 g, 98%, ee. 80%) as its penta ammonium salt: []= -11.47 (c= 7.5, H20); 1H NMR (D20) 8 1.48-1.86 (m, 4H, 2 x CH2-orn),
2.46-2.80 (m, 8H, 4 x CH2-citric), 3.18-3.20 (m, 2H, N-CH2-orn), 4.11-4.14 (m,
1H, CH-orn); Chromatography of its ammonium salt (0.20 g, 0.35 mmol) (18/3, Bio Rad, AG 50W-X8, molecular biology grade, 63-150 gm wet bead size, 200-400 dry mesh size, acid form) gave the free acid form of staphyloferrin A (0.11 g, 99%). Anal. Calcd. for 017H24N2014. 2 H20: C, 39.54; H, 5.47; N, 5.42; found: C, 39.71; H, 5.40; N, 5.32.
Staphyloferrin A Ferric Complex
To a solution of staphyloferrin A ammonium salt (14 mg, 0.025 mmol) in 10 ml distilled deionized water was added a solution of ferric acetylacetonate




47
(12 mg, 0.03 mmol) in 10 ml ethyl acetate. The mixture was stirred for 1 h. The colorless water layer turned pale yellow during that time. The two layers were separated, and the water layer was washed with EtOAc (5 x 10 ml) until the organic layer was colorless. Lyophilization of the aqueous layer gave the staphyloferrin A ferric complex (12 mg, 90%) as a yellow solid.
Discussion
While it is common that many microorganisms produce siderophores, (iron specific chelating ligands), when they are grown under iron starved conditions, the production of identical enantiomeric siderophores in fungi and bacteria is a special case. For example, R, R-rhizoferrin has been found to be produced by members of several families of the Zygomycetes, and ferric R, Rrhizoferrin can be utilized not only by the zygomycetous fungi, but also by a nonproducing bacterium Morganella morganii (101).
Recently, a new citric acid-based siderophore, S,S-rhizoferrin has been isolated from Pseudomonas pickettii DSM 6297, a human pathogen responsible for occasional nosocomial infections (102). S, S-Rhizoferrin (enantio-rhizoferrin) has same formula as R, R-rhizoferrin, but has an inversion of the stereochemistry of the two chiral citrates (Figure 3-2). These two siderophores and staphyloferrin A are relatively simple compounds formed by combining three structural subunits. The citric acids became asymmetric because of the combinations in these three siderophores.
The different stereochemistry of the bacterial and fungal rhizoferrins has raised the question whether enantiomeric recognition of these ferric siderophores by microbial transport systems exists. Transport experiments




48
with radiolabelled iron using S, S- and R, R-rhizoferrin showed that the transport of R, R-rhizoferrin had twice as high iron uptake rates in the fungal Rhizopus strain than S, S-rhizoferrin (102). It suggested that transport of ferric R, R-rhizoferrin is stereoselective in this fungus. However iron transport of ferric S, S- and R, R-rhizoferrin were very similar to one another in Pseudomonas picketti (102).
The chiral separation developed in the synthesis of rhizoferrin and
staphyloferrin A allows for the synthesis of all possible enantiomeric citric acidbased chelators. While (R)-1,2-dimethylcitric acid was obtained from the crystals, the enantiomer (S)-1,2-demethylcitric acid can be obtained from the liquid. The possible (R, S)-rhizoferrin, (R, S)-staphyloferrin A, or (S, R)staphyloferrin A can be synthesized by using both enantiomers. This provides a unique opportunity to study the role of chiral citric acid in bacterial iron utilization and transport of ferric chiral citric acid-based chelators.




49
H O HO COOH
HOOCN N COOH
HOOC OH ( H
R, R-Rhizoferrin
(Rhizopus microsporus var. rhizopodiformis)
H 0HQOCOOH
HOOC N jCOOH
HOOC OH 0 H
S, S-Rhizoferrin
(Pseudomonas pickettii)
H HOOC OHO COOH
HOOCO N N OOH
HOOC OH 0 H
R, R-Staphyloferrin A
(Staphylococcus hyicus)
Figure 3-2. Examples of Chiral Citric Acid-based Siderophores.




CHAPTER IV
EFFICIENT SYNTHESIS OF
POLYAMINE CATECHOLAMIDES
introduction
Hydroxamates and catecholamides are two major classes of
siderophores isolated from bacteria (43). Both can form six coordinate octahedral complexes with iron. While hydroxamates donate three hydroxamates as the chelating funtional groups, catecholamides display two different cases: the chelating groups in enterobactin are six phenolic hydroxyls, whereas in polyamine catecholamides, e.g., parabactin, the chelating functionalities are five phenols as well as the nitrogen on the oxazoline ring (103). The major functional difference between hydroxamate and catecholamide siderophores is related to level of iron concentration (104). Microorganisms produce hydroxamates when the iron concentration is relatively high, while the catecholamides are generated in a low iron concentration. Thus, the iron binding constant of a catecholamide is much higher than that of a hydroxamate.
Except for enterobactin, many catecholamides have structural similarities, that is, they all contain a polyamine backbone as well as chiral oxazoline ring(s). They can have a symmetrical (e.g., norspermidine in vibriobactin and fluviabactin) or an asymmetrical (e.g., spermidine in parabactin and agrobactin) polyamine backbone. They can also be separated into two classes on the 50




51
basis of the substituents on the two terminal primary amines of the polyamine backbone. For example, in parabactin isolated from Paracoccus denitrificans
(74) and agrobactin from Agrobacterium tumefaciens (105), the primary amines of spermidine in both compounds are acylated with 2,3dihydroxybenzoic acid. On the other hand, in vibriobactin isolated from Vibrio cholerae (75), one primary amine is connected with a chiral L-oxazoline ring, while the other primary amine is acylated with 2,3-dihydroxybenzoic acid.
Recently, a novel siderophore, fluviabactin, was isolated from Vibrio fluvialis (79). The chemical structure of L-fluviabactin is similar to that of Lagrobactin, but contains a norspermidine backbone. In this project, the stereochemistry of the oxazoline ring in fluviabactin was changed from Lthreonine to D-threonine to form D-fluviabactin, and the polyamine backbone was changed from norspermidine to homospermidine and spermidine to generate L-homofluviabactin and L-agrobactin, respectively (Figure 4-1).
Polyamine catecholamides, for example, parabactin have emerged as potential iron chelators as an alternative to desferrioxamine B (DFO) in a bile duct-cannuated rat model as well as in a Cebus monkey model. When both ligands were administered as a sc bolus, parabactin was 5.8 times as efficient as DFO in the rodent model (P < 0.001) and was 1.8 times as effient as DFO in the primate model (P < 0.001). Both DFO and parabactin were effective at putting the primates in negative iron balance (64). Furthermore, the catecholamide chelators have been shown to be very potent cell synchronization agents, holding the cells at GU/S border (106). Both parabactin




52
and vibriobactin have been shown to strongly inhibit the growth of L1210 cells. Parabactin also inhibits HSV-1 virus (107,108).
OH
OH
O N
HOJ:)YN N N OH
OH O O OH D-Fluviabactin
OH
OH
O N
OH 0 0 0 OH
HH
L-Homofluviabactin
Figure 4-1. Structures of Two L-Fluviabactin Analogues:
D-Fluviabactin and L-Homofluviabactin.




53
In each of the catecholamide chelators to be synthesized, the catechol
functions of L-fluviabactin, D-fluviabactin, L-agrobactin, and L-homofluviabactin are all in the form of 2,3-dihydroxybenzoyl groups. In fact, the 2,3dihydroxybenzoyl forms a very tight, three to one, high spin, hexacoordinate, octahedral complex with Fe(lll). When the 2,3-dihydroxybenzoyl group is fixed to polyamine backbones, as in the cases of fluviabactin, agrobactin, or vibriobactin, or to a triserine macrocycle, as in enterobactin, the formation constants of the Fe(lll) complexes become even higher. For example, the enterobactin-Fe(lll) complex has a formation constant of 1052 M-1 and parabactin forms a tight Fe(lll) complex with a formation constant of 1048 M1. Synthetically, then, in the cases of fluviabactin and agrobactin, the objective becomes to fix the 2,3-dihydroxybenzoyl functionality to the appropriate anchors, e.g., polyamine or threonine.
The similarities of polyamine catecholamides led to the development of several synthetic schemes in this and other laboratories (109,110,111). The main synthetic obstacle was the selective acylation of a triamine. Because the synthesis of parabactin and agrobactin involve the acylations the primary amines of a spermidine with a 2,3-dihydroxybenzoyl group, they present the same kind of selectivity problem: the acylation of the primary vs. secondary amines of a spermidine. In the synthesis of parabactin and agrobactin, the two terminal primary amines of a spermidine could be acylated with 2,3dihydroxybenzoyl groups followed by the introduction of an oxazoline ring. Alternatively, an oxazoline ring first can be introduced on to the secondary




54
amine of a spermidine followed by the acylations of the two primary amines by 2,3-dihydroxybenzoyl groups.
Vibriobactin synthesis consists of the fixing of a 2,3-dihydroxybenzoyl functionality to only one primary amine of a norspermidine. The synthetic problem is associated with the selective acylation of a primary amine in the presence of another primary amine and a secondary amine and thus, requires different synthetic schemes. The synthetic alternatives could include the monoacylation of one primary amine of a norspermidine with a 2,3-dihydroxybenzoyl group followed by the bis-acylation of the other two amines of the norspermidine with two oxazoline fragments, or the opposite order of attachment. Therefore, appropriately protected triamines become the common denominators in the synthesis of all the polyamine catecholamide iron chelators.
In our laboratory, a protected triamine, an interrally N-benzylated triamine, was developed (112). The reagent synthesis begins with a suitable N-benzyl diamine, which can be obtained either from cyanoethylation of benzylamine followed by reduction of nitrile to diamine with a Raney nickel catalyst, or from the condensation of putrescine with benzaldehyde in formic acid followed by reduction of the imine. The primary amine of the diamine can be regioselectively protected with one equivalent of 2-(tertbutoxycarbonyloxyimino)-2-phenylacetonitrile (BOC-ON). The remaining amine is alkylated with either acrylonitrile or 4-chlorobutyronitrile to furnish the homologous nitriles that can be used directly in the next reaction. The nitriles




55
were reduced to the corresponding amine by Raney nickel. The BOC protecting group of the amines can be readily removed by brief exposure to trifluoroacetic acid to provide the corresponding secondary N4-benzylated triamines (Figure 4-2).
a
N NH2 > N N-BOC
n= 1, or2
b C.H c
SNC N ., N-BOC m"
m= 1, or2
H2N NN-BOC d H2N N(4, NH2
mm
Figure 4-2. Synthesis of Benzyl-protected Triamine: a) BOC-ON; b) acrylonitrile or 4-chlorobutyronitrile; c) Raney Ni; d) TFA.
The protected triamines are versatile reagents for the synthesis of all
polyamine catecholamides, and have been successfully used in the synthesis of parabactin (113) and vibriobactin (110). The synthesis of parabactin was




56
initiated with the polyamine reagent N4-benzylspermidine, which was first reacted with 2,3-dimethoxybenzoyl chloride, providing a bisamide. Next, the N4benzyl group is removed by hydrogenolysis over palladium at atmospheric pressure. The free secondary amine of the diamide can then be acylated with L-N-(tert-butoxycarbonyl)threonine, activated as the N-hydroxysuccinimide ester, affording the corresponding triamide. The BOC group was removed with TFA, and the methyl protecting groups were removed by BBr3 in CH2CI2 (81). The resulting amino alcohol was condensed stereospecifically with ethyl 2hydrobenzimidate to form the acid-sensitive oxazoline ring of parabactin (Figure 4-3).
Because vibriobactin has no symmetry with respect to the terminal acyl groups, it thus presents a new synthetic challenge. The synthesis of vibrobactin began with a primary, secondary amino-diprotected norspermidine, 4-benzyl-N' -(tert-butoxycarbonyl)norspermidine. The free primary amine was acylated with 2,3-dimethoxybenzoyl chloride in the presence of triethylamine, generating the trisubstituted noespermidine. Both the tert-butoxycarbonyl protecting group and the N4-benzyl group can be removed using TFA and Pd/C/HCI, respectively. The order of deprotection is not important. The free primary amine and secondary amine were then bisacylated with the activated ester of L-N-(tert-butoxycarbonyl)threonine. Next, the BOC protecting group was removed with TFA, and the catechol methyoxyl protecting groups were removed with BBr3 in methylene chloride. Finally, the condensation between the threonyl




57
groups and excess of 2,3-dihydroxybenzimidate in refluxing methanol provided vibriobactin (Figure 4-4).
c aOCH3
H2N N NH2 CI O
H H 0 OCH3
H3CO N N ~ O b N & OCH3 c, d
OCHP
HO NH2 .TFA
H OCH3 H3CO N N N CH
jP H
OCH3
OH
OH OH
Figure 4-3. Synthesis of L-Parabactin Using N4-Benzylspermidine: a) Et3N; b) Pd/C/H2; c) L-N-(tert-butoxycarbonyl)threonine/DCC/ N-hydroxysuccinimide; d) TFA; e) BBr3; f) Ethyl 2-hydrobenzimidate.




58
OCH3
H + OCH3
H2N .N N-BOC
H HO H b, c, d H3COJ N .N N-BOC
H3CO 0
BOC BOC I I
HO NH HN OH
HO O-41 O He, f H3CO-1 N N NH
H3CO O
OH OH
OH OH
0 N N O
HO N NNH
OH O
Figure 4-4. Synthesis of Vibriobactin Using Diprotected Triamine: a) Et3N; b) Pd/C/H2; c) TFA; d) L-N-(tert-butoxycarbonyl)threonine/ DCCI N-hydroxysuccinimide; e) BBr3; f) ethyl 2-hydrobenzimidate.




59
As can be seen from the above synthesis, the protected triamines are useful reagents for the synthesis of polyamine catecholamide siderophores. However, in the synthesis of polyamine catecholamides and their analogues, some commercially available coupling reagents, for example, 1,1carbonyldiimidazole (CDI) can be taken advantage of. The selective acylation of the primary amines of an unprotected triamine can be achieved by the reaction a carboxylic acid which has been activated with CDI. The phenol hydroxyl protecting groups can be changed from methyl to benzyl (the latter is much more easily removed), and the intermediate can directly utilized in the following reactions without further purification. Therefore, a more concise scheme for the synthesis of polyamine catecholamides is presented.
Synthesis
The synthesis of polyamine catecholamide iron chelators in our lab
depends heavily on the availability of the appropriately protected polyamines norspermidine, spermidine, and homospermidine. In the synthesis of vibriobactin, the selective acylation of one primary and one secondary amine of a norspermidine is almost impossible without a protected polyamine reagent. However, in the synthesis of parabactin, agrobactin, and fluviabactin, the selective acylation of primary amines vs. secondary amine of a triamine can be achieved by utilizing N-hydroxysuccinimide or imidazole activated acids. Although the yields in the formation of diamides are moderate (60-70%), it can save many steps in the synthesis of a protected polyamine reagent. Thus, it is possible to increase the yield of a total synthesis.




60
In the present synthesis, 1,1-carbonyldiimidazole (CDI) (114,115) was
utilized for the selective acylations of two terminal primary amines of triamines. These reactions involved the attachment of two 2,3-dihydroxybenzoyl groups to the primary amines of norspermidine (L- or D-fluviabactin), spermidine (Lagrobactin), or homospermidine (L-homofluviabactin). The CDI coupling reagent was first reacted with 2,3-dibenzoxylbenzoic acid for about one hour, after which the corresponding free triamines were added. CDI was an especially convenient, selective amide-forming reagent in that the imidazole byproduct was washed out during workup. In this project, the reactions of spermidine, norspermidine, or homospermidine with 2,3-bis(benzoxy)benzoic acid in the presence of CDI produced the corresponding bisamides (14, 15, 16) as illustrated in Figure 4-5 (64-73%). 2,3-Bis(benzyloxyl)benzoic acid was utilized instead of 2,3-dimethoxylbenzoic acid because the benzyl protecting groups are more easily removed than the methyl groups. The synthesis of 2,3bis(benzyloxyl)benzoic acid (111) began with 2,3-dihydroxybenzaldehyde, in which the catechol hydroxyls were protected by benzyl groups using benzyl chloride and potassium carbonate, followed by oxidation with sodium chlorite and sulfamic acid, to provide 2,3-bis(benzyloxy)benzoic acid (90%).
However, unlike commercially available spermidine and norspermidine, homospermidine (81) has to be synthesized by alkylation of mesitylenesulfonylamide (116) with N-(4-bromobutyl)phthalimide and NaH to form fully protected homospermidine (67%). Free homospermidine was obtained by removal of the phthalimide and mesitylenesulfonyl protective




61
OBn
I CDI H2N-I N'H NH2 + OBn CDI
HOH
O OH
norspermidine m = n = 1 spermidine m = 1, n = 2 homospermidine m = n = 2
OBn O O OBn
BnO OBn
(14) m=n=1
(15) m=l,n=2
(16) m = n = 2
Figure 4-5. CDI Coupling Reactions.




62
groups with NH2NH2 (82%) (117) and with 30% HBr/HOAc and phenol (98%)
(81), respectively.
Terminally-diacylated triamines (14. 15. 16) were then acylated with Ncarbobenzoxy-L-threonyl-O-hydroxysuccimide (N-Cbz-L-Thr-OSu) for Lfluviabactin, L-agrobactin, L-homofluviabactin, or its D enantiomer N-Cbz-D-ThrOSu for D-fluviabactin (see compounds 17, 18, 19, 20.) (50-76%). The hydroxysuccimide active ester of a Cbz-protected threonyl can be readily prepared using the protected threonines and N- hydroxysuccimide in the presence of DCC in methylene chloride, and do not need to be separated from the solvent; even the insoluble DCU can stay in the reactions. The carbobenzoxyl and benzyl groups of the resulting trisubstituted amide were removed by hydrogenolysis over palladium in methanolic HCI at atmospheric pressure, providing amine salts (21, 22, 23, 24) (96-98%).
The stereospecific formation of the acid-sensitive trans-oxazoline ring (118) can be achieved by the condensation of a threonine residue and ethyl 2,3-dihydroxybenzimidate. However, ethyl 2,3-dihydroxybenzimidate cannot be obtained simply by Pinner's reaction, as in the preparation of ethyl 2hydroxybenzimidate (119). Another approach to the imidate ester started with 2,3-dibenzyloxybenzoic acid, which was also used in the CDI coupling reactions. The acid was converted to 2,3-di(benzyloxy)benzoyl chloride, followed by aminolysis of 2,3-bis(benzyloxy)benzoyl chloride, furnished 2,3di(benzyloxy)benzamide (98%), which was selectively O-alkylated with triethyloxonium hexafluorophosphate in CH2CI2, followed by basification (120)




63
to provide ethyl 2,3-bis(benzyloxy)benzimidate in 80% yield. Finally, the cleavage of benzyl-protecting groups by hydrogenolysis under mild conditions (10% Pd/C, 1 atm) led to ethyl 2,3-dihydroxybenzimidate (73%) (111) (Figure 4-6).
OH OBn
a b
OH OBn
H OH0
OBn OBn
| d
Bn 'OBn
HO CI O
OBn OH
e, f 70H OBn OH
H2N HN OEt
Figure 4-6. Synthesis of Ethyl 2,3-Dihydroxybenzimidate: a) benzyl chloride/K2CO3; b) NaCIO2/sulfamic acid; c) SOCI2; d) NH4OH; e) triethyloxonium hexafluorophosphate; f) Pd/C/H2.




64
Therefore, the threnonyl residues (21, 22, 23, 24) were condensed with ethyl 2,3-dihydroxybenzimidate in methanol to furnish the final products L- and D-fluviabactin, L-agrobactin, and L-homofluviabactin (60-70%) (Figure 4-7). The mechanism for the oxazoline-forming reaction was suggested to be the following. The hydroxyl of a threonine residue first attacks the imine of ethyl 2,3dihydrobenzylimidate, in which the ethoxy serves as a leaving group. After the protons transfers intramolecularly from the protonated amine of the threonyl residue to the imine, the amine of the threonyl residue acts as another nucleophile to react with the imine, forming a five-membered ring. Finally, the elimination of ammonia between the nitrogen of the threonine and the carbon of the imidate provides an oxazoline ring.
In summary, the scheme developed here is shortened to four steps. The starting material, 2,3-bis(benzyloxy)benoic acid, can be prepared on a large scale by recrystallation. Other starting compounds, such as CDI, norspermidine, and spermidine are commercially available. The intermediates of hydogenolysis can be directly used in subsequent reactions. Thus, it is possible to make polyamine catecholamides on a large enough scale to provide materials for animal tests. The iron clearing properties of L- and Dfluviabactin were evaluated in a bile duct-cannuated rat model. Both compounds were shown to be very effective at clearing iron from the rodent. When given sc at a dose of 150 pmol/kg, both ligands had iron clearing efficiences of > 13%.




65
OBn 0 0 OBn
BnO OBn (14) m = n = 1
N N "N (15) m=l,n=2 H (16) m=n:2
a
HO NH-Cbz
OBn O 0 O OBn (17, 18) m= n = 1 BnO OBn (Lm9)m 1, n2=2 H H IV(20) m=n=2
b
HO NH.HCI
HO O OH (21, 22) m = n = 1
N N- "N(23) m=1,n=2
H K (24) m =n
c
OH
OH
O N
OH O O OH
HO N N N H H OH
L- or D- Fluviabactin m = n = 1
L-Agrobactin m = 1, n = 2
L-Homofluviabactin m = n = 2
Figure 4-7. Synthesis of L- and D- Fluviabactin, L-Agrobactin, L-Homofluviabactin. a) N-Cbz-L-Thr-OSu or N-Cbz-D-Thr-OH / N-hydroxysuccimide / DCC; b) Pd / C / EtOH / HCI; c) ethyl 2,3-dihydroxybenzimidate / MeOH.




66
Experimental
General
N-Carbobenzyloxy-L-threonine and N-Carbobenzyloxy-D-threonine were purchased from Sigama Chemical Co. (St. Louis, MO), N-Carbobenzyloxy-Lthreonine-1-(N-succinimidyl) ester was purchased from Bachem Bioscience Ino. (King of Prussia, PA) and all other reagents were purchased from Aldrich Chemical Co. (Milwaukee, WI). Fisher Optima grade solvents (Fisher Scientific, Pittsburgh, PA) were used. DMF was distilled under N2 and stored over molecular seives. Distilled solvents were employed for reactions involving chelators. Organic extracts were dried with sodium sulfate unless otherwise indicated. Glassware for chelator reactions and purification steps were soaked in 3 N HCI for 15 min, rinsed with distilled water then distilled ethanol and dried prior to use. Silica gel 32-63 (40 gM "flash") from Selecto, Inc. (Kennesaw, GA) or silica gel 60 (70-230 mesh) obtained from EM Science (Darmstadt, Germany) or Lipophilic Sephadex LH-20 from Sigma Chemical Co. (St. Louis, MO) was used for column chromatography. Optical rotations were determined at 589 nm (Na lamp) with a Perkin Elmer 341 polarimeter and 1 decimeter cell path length in the indicated solvent; c is expressed as g of compound per 100 mL. Proton NMR spectra were obtained on a Varian Unity 300 at 300 MHz in CD3OD at ambient temperature unless otherwise indicated; chemical shifts are given in parts per million downfield from an internal tetramethylsilane standard. High resolution mass spectra were obtained utilizing FAB ionization from a




67
glycerol matrix on a fennigan 4516. Elemental analyses were performed by Atlantic Microlabs (Norcross, GA).
N1', N-Bisf2,3-bis(benzyloxy)benzoyl]norspermidine (14)
2,3-Bis(benzyloxy)benzoic acid (43) (3.34 g, 10 mmol) and 1,1carbonyldiimidazole (1.62 g, 10 mmol) were dissolved in dry CH2Cl2 (100 mL) and stirred for 1 h at room temperature under a nitrogen atmosphere. A solution of norspermidine (0.67 g, 5.1 mmol) in dry CH2CI2 (10 mL) was added and the mixture was stirred overnight. The resulting solution was washed with 2% NaOH (100 mL), water (100 mL) and brine (100 mL), then dried over MgSO4 and filtered. Solvent removal in vacuo followed by flash chromatography of the residue on silica gel with 5% EtOH in EtOAc gave (14) (2.6 g, 67%) as a colorless oil: 1H NMR (CDCI3) 5 1.52-1.63 (m, 4 H), 2.45-2.50 (m, 4 H), 3.283.36 (m, 4 H), 5.07 (s, 4 H), 5.15 (s, 4 H), 7.10-7.49 (m, 26 H), 7.64-7.68 (m, 2 H), 8.04-8.08 (br, 1 H).
N1, N8-Bis[2,3-di(benzyloxy)benzoyl]spermidine (15)
2,3-Bis(benzyloxyl)benzoic acid (2.60 g, 7.78 mmol), CDI (1.26 g, 7.78
mmol), and spermidine (0.56 g, 3.86 mmol) in dry CH2CI2 (10 mL) were reacted and worked up by the method of 14). Flash chromatography of the resulting residue on silica with 10% MeOH in CHCI3 gave (15) (2.2 g, 74%) as a pale yellow oil: 1H NMR (CD3OD) 8 1.37-1.44 (m, 4 H), 1.56-1.66 (m, 2 H), 2.38-2.44 (m, 2 H), 2.44-2.55 (m, 2 H), 3.20-3.26 (m, 4 H), 5.08 (s, 4 H), 5.16-5.17 (m, 4 H), 7.10-7.50 (m, 26 H).
N1'.No-Bis[2.3-di(benzyloxy)benzoyl]homospermidine (16)




68
A mixture of 2,3-di(benzyloxyl)benzoic acid (2.10 g, 6.28 mmol), 1,1carbonyldiimidazole (1.02 g, 6.29 mmol), and homospermidine (0.5 g, 3.14 mmol) in dry CH2CI2 (10 mL) were reacted and worked up by the method of
(14). Flash chromatography of the resulting crude material on silica with 10% MeOH in CHCI3 gave 1.6 g (64%) (16) as a pale yellow oil: 1H NMR (CD3OD) 5 1.41-1.48 (m, 8 H), 2.51-2.58 (m, 4 H), 3.22-3.28 (m, 4 H), 4.85 (s, 4 H), 5.089
(s, 2 H), 5.18 (s, 2 H). 7.01-7.50 (m, 26 H). N4-(N-Carbobenzyloxy-L-threonyl)-N1'-N 7-bis[2,3bis(benzyloxy)benzoyl]norspermidine (j.7)
A solution of DCC (0.34 g, 1.65 mmol) in dry CH2CI2 (20 mL) was added to a solution of N-carbobenzyloxy-L-threonine (0.33 g, 1.30 mmol) and Nhydroxysuccimide (0.19 g, 1.65 mmol) in dry CH2CI2 (20 mL) and stirred for 18 h at room temperature. The mixture was filtered and (14) (1.0 g, 1.3 mmol) and triethylamine (200 mg, 2.0 mmol) added. The resulting mixture was stirred for 24 h, concentrated in vacuo, then dissolved in EtOAc (100 mL). The solution was washed with H20 (2 x 50 mL), 10% citric acid (50 mL), and H20 (100 mL), dried and filtered. Solvent removal in vacuo followed by flash chromatography on silica with EtOAc/CHCI3 (1:1) gave (17) (0.68 g, 53%): 'H NMR (CD3OD) 8 1.08-1.11 (m, 3 H), 1.51-1.72 (m, 4 H), 2.92-3.31 (m, 8 H), 3.83-3.92 (m, 1 H),
4.38-4.42 (m, 1 H), 5.02-5.21 (m, 10 H), 7.11-7.51 (m, 31 H). N4-[N-Carbobenzyloxy-D-threonyll-N'-N 7-bis[2,3bis(benyloxy)benzoyl]norspermidine (18)
A solution of DCC (0.63 g, 3.04 mmol) in dry CH2CI2 (40 mL), Ncarbobenzyloxy-D-threonine (0.76 g, 3.0 mmol), N-hydroxysuccimide (0.35 g,




69
3.05 mmol) in dry CH2CI2 (40 mL), and compound (14) (1.5 g, 1.96 mmol) were combined and worked up by the method of (17), and afforded (. 8) (0.62 g, 50%): 1H NMR (CD3OD) 8 1.08-1.11 (d, 3 H, J= 7), 1.51-1.72 (m, 4 H), 2.923.31 (m, 8 H), 3.83-3.92 (m, 1 H), 4.38-4.42 (m, 1 H), 5.02-5.21 (m, 10 H), 7.117.51 (m, 31 H).
Anal. Calcd. for C60H62N4010: C, 72.13; H, 6.25; N, 5.61; Found: C, 71.87; H,
6.54; N, 5.53.
N5-(N-Carbobenzyloxy-L-threonyl)-N'-N'-bis[2,3di(benzyloxy)benzoyl]spermidine (19)
DCC (0.63 g, 3.0 mmol) in 40 mL dry CH2CI2, N-carbobenzoxy-L-threonine (0.76 g, 3.0 mmol), N-hydroxysuccimide (0.35 g, 3.0 mmol) in dry CH2CI2 (40 mL), and (15) (1.5 g, 1.93 mmol) were combined and worked up by the method of (1). Flash chromatography of the resulting residue on silica with EtOAc/CHCI3 (1:1) afforded (9) (1.08 g, 55%) as a viscous oil: 1H NMR (CD3OD) 5 1.13-1.16 (m, 3 H), 1.56-1.76 (m, 6 H), 2.94-3.35 (m, 8 H), 3.72-3.86
(m, 1 H), 4.32-4.41 (m, 1 H), 5.06-5.25 (m, 10 H), 7.10-7.50 (m, 31 H).
Anal. Calcd. for C61H64N4010: C, 72.31; H, 6.37; N, 5.53; Found: C, 72.23; H,
6.54; N, 5.44.
N6-(N-Carbobenzoxy-L-threonyl)-N'-No-bis[2,3di(benyloxy)benzoyl]homospermidine (20) A mixture of (16) (1.5 g, 1.9 mmol), N-Carbobenzoxy-L-threonyl-Nhydroxysuccimidate (Bachem, 1.0 g, 2.85 mmol), and triethylamine (0.20 g,
1.98 mmol) in CH2CI2 (100 mL) were stirred overnight. The resulting solution was washed with H20 (100 mL), 10% citric acid (100 mL), and brine (100 mL),




70
then dried and filtered. Solvent removal in vacuo followed by flash chromatography on silica with EtOAc/CHCI3 (1:1) gave (2_) (1.49 g, 76%): 1H NMR (CD3OD) 8 1.09-1.11 (d, 3 H, J= 7), 1.31-1.54 (m, 8 H), 3.01-3.28 (m, 8 H), 3.84-3.93 (m, 1 H), 4.42-4.45 (m, 1 H), 5.07-5.16 (m, 10 H), 7.09-7.48 (m, 31 H).
Anal. Calcd. for C62H66N4010: C, 72.49; H, 6.48; N, 5.45. Found: C, 72.26; H,
6.56; N, 5.46.
N4-L-Threonyl-N1-N'-bis(2,3-dihydroxybenzoyl)norspermidine Hydrochloride
A mixture of (17) (0.75 g, 0.75 mmol) and 10% Pd/C (0.1 mg) in HCI
saturated MeOH (50 mL) was stirred at room temperature for 1 h under a H2 atmosphere at ambient pressure. The mixture was filtered through acidwashed Celite and was concentrated in vacuo to give (21) (0.40 g, 99%) as a white solid, which was used without further purification. An pure sample was obtained by column chromatography on LH-20 with 10-15% EtOH/toluene: 1H NMR (CD3OD) 8 1.25-1.28 (d, 3 H, J = 7), 1.84-2.02 (m, 4 H), 3.38-3.78 (m, 8 H),
4.02-4.10 (m, 1 H), 4.18-4.21 (m, 1 H), 6.68-6.75 (m, 2 H), 6.91-6.95 (m, 2 H),
7.18-7.21 (m, 2 H).
N4-D-Threonyl-N'-N'-bis(2,3-dihydroxybenzoyl)norspermidine Hydrochloride
(22)
A mixture of (18) (0.30 g, 0.30 mmol) and 10% Pd/C (0.15 g) in HCI
saturated methanol (50 mL) were combined and worked up by the method of
(21) to yield (22) (0.16 g, 99%) as a white solid, and used without further purification. An analytical sample was obtained by column chromatography on LH-20 with 15% EtOH in toluene. 1H NMR (CD3OD) 5 1.25-1.28 (d, 3 H, J = 7),




71
1.84-2.02 (m, 4 H), 3.38-3.78 (m, 8 H), 4.02-4.10 (m, 1 H), 4.18-4.21 (m, 1 H),
6.68-6.75 (m, 2 H), 6.91-6.95 (m, 2 H), 7.18-7.21 (m, 2 H).
Anal. Calcd. for C24H33CIN4Os: C, 53.28; H, 6.15; N, 10.36. Found: C, 53.12; H, 6.36; N, 10.12.
N4-(L-Threonyl)-N'-N'-bis(2,3-dihydroxybenzoyl)spermidine Hydrochloride (23)
A mixture of (19) (0.31g, 0.31 mmol) and 10% Pd/C (100 mg) in HCI
saturated methanol (30 mL) were reacted and worked up by the method of (211 to give (23) (0.16 g, 93%) as a white solid. The compound was used without further purification. An analytical sample was obtained by column chromatography on LH-20 with 15% EtOH in toluene: 1H NMR (CD3OD) 8 1.071.10 (m, 3 H), 1.52-1.73 (m, 6 H), 2.91-3.32 (m, 8 H), 3.82-3.91 (m, 1 H), 4.364.44 (m, 1 H), 6.67-6.73 (m, 2 H), 6.90-6.93 (m, 2 H), 7.19-7.23 (m, 2 H).
Anal. Calcd. for C25H35CIN408s: C, 54.20; H, 6.19; N, 10.11. Found: C, 54.11; H, 6.23; N, 9.87.
N 5-(L-Threonyl)-N'-N'-Bis(2, 3-dihydroxybenzoyl)homospermidine Hydrochloride (24)
A mixture of (20) (0.6 g, 0.58 mmol) and 10% Pd/C (200 mg) in HCI
saturated methanol (50 mL) were reacted and worked up by the method of (21). The resulting white solid (2A4) (0.3 g, 91%) was used without further purification. An analytical sample was obtained by column chromatography on LH-20 with 15% EtOH in toluene. 'H NMR (CD3OD) 8 1.19-1.21 (d, 3 H, J = 7), 1.41-1.54 (m, 8 H), 3.11-3.36 (m, 8 H), 4.44-4.53 (m, 1 H), 4.92-5.12 (m, 1 H), 6.84-7.28 (m, 6
H).




72
Anal. Calcd. for C26H37CIN408: C, 54.88; H, 6.55; N, 9.85. Found: C, 54.76; H, 6.78; N, 9.78.
N4-[2-(2.,3-Dihydroxyphenyl)-(4S, 5R)-trans-5-methyl-2-oxazoline-4carbonyl]-N1'-N7-bis(2,3-dihydroxybenzoyl)norspermidine (L-Fluviabactin) To a solution of (21) (0.16 g, 0.30 mmol) in dry, degassed methanol (20 mL) was added ethyl 2,3-dihydroxybenzimidate (0.22 g, 1.21 mmol). The mixture was heated at reflux under nitrogen for 30 h then concentrated in vacuo. Column chromatography on LH-20 with 10% EtOH/toluene gave L-fluviabactin (0.12 g, 64%) as a white glass: 1H NMR (CD3OD, 50 oC) 5 1.40 (d, 3 H, J = 7), 1.84-1.94 (m, 2 H), 2.03-2.14 (m, 2 H), 3.40-3.92 (m, 8 H), 4.67-4.88 (m, 1 H),
5.22-5.30 (m, 1 H), 6.01-6.76 (m, 3 H), 6.86-6.96 (m, 3 H), 7.12-7.21 (m, 3 H).
HRMS Calcd. (M+H) 623.2353, found 623.2339; [a] = + 89.18 (c = 1, MeOH).
Anal. Calcd. for C31H34N4010 0.5H20: C, 58.95; H, 5.59; N, 8.87; Found: C, 59.24; H, 5.71; N, 8.81.
N4-[2-(2,3-Dihydroxyphenyl)-(4R, 5S)-trans-5-methyl-2-oxazoline-4carbonyll-N'-N 7-bis(2,3-dihydroxybenzoyl)norspermidine (D-Fluviabactin) A mixture of (22) (150 mg, 0.28 mmol) and ethyl 2,3-dihydroxybenzimidate (120 mg, 0.66 mmol) were reacted and worked up by the method of LFluviabactin. Column chromatography of the resulting residue on LH-20 with 15% EtOH/toluene gave D-fluviabactin (0.11 g, 63%) as a white glass: 'H NMR (CD3OD, 50 oC) 8 1.40 (d, 3 H, J= 7), 1.84-1.94 (m, 2 H), 2.03-2.14 (m, 2 H),
3.40-3.92 (m, 8 H), 4.67-4.88 (m, 1 H), 5.22-5.30 (m, 1 H), 6.01-6.76 (m, 3 H),
6.86-6.96 (m, 3 H), 7.12-7.21 (m, 3 H). HRMS Calcd. (M+H) 623.2353, found 623.2334; [a] = -85.86 (c = 1, MeOH).




73
Anal. Calcd. for C31H34N4010EtOH0.5H20: C, 58.49; H, 6.10; N, 8.27. Found: C, 58.85; H, 6.13; N, 8.41.
N'-[2-(2,3-Dihydroxyphenyl)-(4S, 5R)-trans-5-methyl-2-oxazoline-4carbonyl-N1'-N8-bis(2,3-dihydroxybenzoyl)spermidine (L-Agrobactin) To a solution of (23) (0.14 g, 0.25 mmol) in dry, degassed methanol (30 mL) was added ethyl 2,3-dihydroxybenzimidate (0.12 g, 0.66 mmol). The mixture was heated at reflux under nitrogen for 30 h then concentrated in vacuo. Column chromatography of the resulting residue on LH-20 with 15% EtOH/toluene gave L-agrobactin (86) (0.10 g, 63%) as a white glass: 1H NMR (CD30D, 50 oC) 8 1.37-1.46 (m, 3 H), 1.58-2.12 (m, 6 H), 3.34-3.88 (m, 8 H), 4.70-4.83 (m, 1H), 5.20-5.29 (m, 1 H), 6.60-6.76 (m, 3 H), 6.85-6.95 (m, 3 H),
7.11-7.22 (m, 3 H).
Anal. Calcd. for C32H36N4010 0.5H20: C, 59.53; H, 5.78; N, 8.68. Found: C, 59.17; H, 5.76; N, 8.54.
NS-[2-(2,3-Dihydroxyphenyl)-(4S,5R)-trans-5-methyl-2-oxazoline-4carboxamido]-N'-N10o-bis(2. 3-dihydroxybenzoyl)homospermidine (LHomofluviabactin).
A mixture of (24) (0.30 g, 0.53 mmol) in dry, degassed methanol (50 mL) and ethyl 2,3-dihydroxybenzimidate (200 mg, 1.1 mmol) were reacted and worked up by the method of L-Fluviabactin. Column chromatography of the resulting residue on LH-20 with 15% EtOH/toluene gave L-homofluviabactin as a white glass (0.22 g, 64%): 'H NMR (CD30D, 50 0C) 8 1.42-1.45 (d, 3 H, J = 7),
1.55-1.88 (m, 8 H), 3.35-3.78 (m, 8 H), 4.75-4.94 (m, 1 H), 5.18-5.28 (m, 1H),
6.63-6.76 (m, 3 H), 6.85-6.96 (m, 3 H), 7.12-7.23 (m, 3 H); Anal. Calcd. for




74
C33H38N4010 0.5H20: C, 60.08; H, 5.96; N, 8.49. Found: C, 60.01; H 5.92; N,
8.75.
Discussion
The NMR analysis of polyamine catecholamides is widely studied (43,76, 86, 120). Previously, a 10:1 mixture of CDCI3:d6-DMSO was used in this and other labs for the 1H NMR studies. The chloroform signal was set at 7.24 ppm as a reference in this solvent. The most useful information in the NMR is the 7ymethyl protons appeared between 1.2 and 1.4 ppm, as well as the o- and P3methine protons between 4.4 and 5.4 ppm. However, the polyamine backbone proton peaks (1.42 ppm 2.01 ppm) are very close to those of the y-methyl region (1.26 1.39 ppm), water (1.56 ppm), and DMSO (2.22 ppm).
In this paper, another solvent, CD30D, was utilized. Using this solvent,
there are some dramatic changes in the NMR. The water peak is shifted from
1.56 ppm in CDCI3 to 4.87 ppm in CD3OD, and the solvent residual peak appears at 3.31 ppm. Thus, the polyamine backbone proton peaks and the ymethyl proton peaks are isolated and easily compared. Therefore, using CD3OD as a solvent is a good method to identify and differentiate these compounds. Because all the amide and catechol protons are exchanged with CD3OD, these protons did not interfere with the peaks of aoc-, P3- and y-methyl occurred, and the 'H NMR spectra became simpler.
In setting either the CD3OD signal at 3.31 ppm or the water signal at 4.87 ppm, the 1H NMR spectra of L- and D-fluviabactin, L-agrobactin and Lhomofluviabactin can be separated into four basic sections. The aromatic




75
protons appear between 6.6 ppm and 7.2 ppm; the a- and P-methine protons fall between 4.7 ppm and 5.3 ppm; the internal polyamine backbone protons extend between 1.55 ppm and 2.1 ppm; the y-methyl protons appear between
1.3 ppm and 1.5 ppm (Figure 4-8 to Figure 4-11).
The a-methine proton peaks overlapped with the water peak in the NMR spectra of final productsat room temperature: L- and D-fluviabactin, Lagrobactin, and L-homofluviabactin, although this did not happen in the 1H NMR spectra of all precursors. By setting the temperature at 500C, the a-methine proton peaks were then separated from the water peak. The 300 MHz NMR spectra of L-and D-fluviabactin, L-agrobactin, and L-homofluviabactin were illustrated in Figure 4-8 to 4-11.
It is worth noting that the 'H NMR spectrum of agrobactin is different from those of D- and L- fluviabactin, as well as from L-homofluviabactin. The most outstanding feature of agrobactin 1H NMR spectrum is reflected in the duplicity of the NMR signals, whereas the 1H NMR spectra of the other three compounds lack this duplicity. This is most evident when comparing the y-methyl protons. The 'H NMR spectra of D- and L- fluviabactin and L-homofluviabactin display a single doublet for y-methyl protons between 1.36 ppm to 1.42 ppm. However, the 'H NMR spectrum of agrobactin has double doublets for y-methyl protons from 1.3 ppm to 1.5 ppm. The ratio between these two doublets is not equal, and is apparently sensitive to temperature and the solvent used.
The duplicity of signals originating in the threonyl moiety of agrobactin is of particular interest. The threonine-coupling pattern in all four compounds




HOD D30D
7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 ppm
Figure 4-8. 300 MHz 'H NMR Spectrum of L-Fluviabactin in CD3OD at 50oC.




IOD D3OD
/
7 6 5 4 3 2
Figure 4-9. 300 MHz 1H NMR Spectrum of D-Fluviabactin in CD3OD at 50oC.




IOD C3OD
-41
OD
7 6 5 4 3 2
Figure 4-10. 300 MHz H NMR Spectrum of L-Agrobactin in CD3OD at 500C.




IOD D3OD
-4
6 5 4 3 2
Figure 4-11. 300 MHz 1H NMR Spectrum of L-Homofluviabactin in CD3OD at 50oC.




80
should display a simple spin system if there was not any conformational effects. The a-methine should be split by the P-methine, producing a doublet. The P-methine is coupled to the three y-methyl protons and should show a quartet, which is split again by the a-methine to produce eight lines, which would result in a complex multiplet. Finally, the y-methyl should be split mainly by the single P-methine proton, displaying a doublet with an estimated coupling constant between 5.5 to 6.5 Hz, since the long-range a-y coupling interaction is expected to be small.
The threonyl y-methyl protons of agrobactin display two unequal doublets, unlike those of L- and D-fluviabactin, and L-homofluviabactin. This suggests that there are at least two distinct magnetic environments for the threonine substituent in agrobactin. It is worthwhile to note that the spermidine backbone displays a nearly linear conformation in the X-ray crystallographic study of agrobactin (105). It is assumed that fluviabactin would have a similar conformation as agrobactin. Then, relatively simple models could be established for these compounds, although they do not represent all of the possible orientations of the spermidine backbone and aromatic rings.
Because the y-methyl group is closer to the terminal 2,3-dihydroxybenzoyl groups than P-methine or c-methine, there would be a hindered rotation about the central amide, more changes in the magnetic environment and the protons will be easily influenced by anisotropic effects. Thus, in agrobactin, a proton located in a conformation lying to one side of the polyamine backbone could demonstrate a different signal from its counterpart on the other side of the




81
polyamine backbone. In symmetrical analogues, such as fluviabactin and homofluviabactin, however, the intramolecular distances on either side will be equalized, eliminating the duplicity of signals observed for the y-methyl of agrobactin. As expected, fluviabactin and homofluviabactin showed a simplified spectra. The y-methyl of both compounds exhibit only a single doublet, located at 1.40 ppm in fluviabactin and 1.44 ppm in homofluviabactin (Figure 4-12 to 4-14)
Interestingly, the ratio of the two doublets of y-methyl in agrobactin is not equal, implying that either the cis- or trans-isomer is preferred. The chemical shift (AS) of y-methyl between these two doublets is 0.04 Hz, which is very close to the difference of chemical shifts between the y-methyl of fluviabactin and homofluviabactin (0.035 Hz). When the 1H NMR spectrum of agrobactin was run at room temperature, the integration of downfield doublet of y-methyl was always larger than that of the upfield one. On looking at the 1H NMR spectra of these compounds, the downfield doublet of y-methyl of agrobactin matches that of homofluviabactin, whereas the upfield one matches that of fluviabactin. This suggests that agrobactin prefers to adapt a conformation close to that of homofluviabactin.
There are two kinds of isomers in agrobactin: cis and trans isomers.
When the polyamine side containing four methylenes is on the same side of the y-methyl, it forms a cis isomer; whereas when the polyamine backbone side is on the opposite side of the y-methyl, it forms a trans isomer. The steric hindrance in cis isomer might be less than that of the trans isomer due to the




2.2 2.0 1.8 1.6 1.4
Figure 4-12. 300 MHz 1H NMR of y-Methyl Group and Polyamrnine Backbone C-Methylene Group of Fluviabactin in CD3OD.




00
I 03
2.2 2.0 1.8 1.6 1.4
Figure 4-13. 300 1H NMR Spectrum of y-Methyl Group and Polyamine backbone C-Methylene Groups of Agrobactin.




U0,
..0
1.8 1.6 1.4
Figure 4-14. 300 MHz 'H NMR Spectrum of the y-Methyl group and Polyamine Backbone C-Methylene Group of Homofluviabactin in CD3OD.




85
longer methylene chain. Also, the central amide carbonyl can be hydrogen bonded to the propyl amide hydrogen. It is predicted that agrobactin would prefer to adapt a cis isomer rather than a trans isomer (Figure 4-15).
H3C H3C R4
NN
N R OL-Fluviabactin L-Homofluviabactin
N~N
R / 4
/N
trans (E) L-Agrobactin cis (Z)
HO
H OH
RR NH
R4: OH R3:O
0OH 'Il
Figure 4-15. Conformations of L-Agrobactin, L-Fluviabactin and L-Homofluviabactin in CD3OD.




CHAPTER V
THE EFFECTS OF RHIZOFERRIN, D- AND L-FLUVIABACTIN, L-AGROBACTIN AND L-HOMOFLUVIABACTIN
ON IRON TRANSPORT IN PARACOCCUS DENITRIFICANS Introduction
Iron is required by virtually all microorganisms with the exception of some lactobacilli. However, iron tends to form insoluble polymers under physiological conditions, so the solubility product of ferric hydroxides is only 10-38 at pH 7.4
(6). The efficient utilization of iron by microorganisms requires the production of iron-specific chelating ligands known as siderophores to solubilize iron in a biologically available form (121,122). A wide variety of bacteria and fungi produce siderophores as well as membrane transport proteins to facilitate uptake of hydrophilic siderophore complexes across nonpolar cell membranes (123,124). Siderophore-mediated iron transport has been investigated primarily by following the fate of radiolabeled siderophore iron complexes in transport assays, although recent investigations have applied electron paramagnetic resonance and Mossbauer spectroscopy to the experiments (125,126,127). Iron transport studies can also be conducted by substituting a transition metal ion that has a similar ionic radius, the same charge, and the same coordination geometry as Fe(III). For example, Cr(IIl) forms complexes with siderophores that are structurally similar to ferric siderophores but are kinetically inert to ligand substitution (128), while Ga(lll) forms complexes with
86




Full Text

PAGE 1

SYNTHESIS OF RHIZOFERRIN STAPHYLOFERRIN A, AND POLYAMINE CATECHOLAMIDES AND THE STUDY OF THEIR EFFECTS ON IRON TRANSPORT IN PARACOCCUS DENITRIFICANS BY MEI GUOXIN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2000

PAGE 2

This dissertation is dedicated to my wife, Hua Yao, and my daughter, Anni Xin.

PAGE 3

ACKNOWLEDGMENTS Six years ago, when I first entered the graduate program in the College of Pharmacy, University of Florida, I only had the mind for the research, not the background or means to accomplish it. I had not synthesized an organic compound before, let alone tested something in a biological system. Now that I am a Ph.D. candidate in medicinal chemistry, I shall always be grateful to the people who helped me and made all that possible. I was fortunate to have this opportunity to do research in Dr Raymond J Bergeron's laboratory. I owe a great debt to him for his endless patience, daily guidance, supervision, and friendship For six years, he has taught me as much about life as about science. His farsightedness and dedication to science have always inspired me, even under the most difficult conditions. His quick thinking, vast knowledge, humor and enthusiasm for life and work have affected me as strongly as his annual Fourth of July parties. Dr. Bergeron has a special way to guide a graduate student. With a project seeming quite impossible at first he would always encourage me to keep on with it and get to the core of the problem. At the key time of a project, he would point out what was wrong and what was right. He was so evenhandedly critical of every single mistake I made in my work. This not only impressed me deeply, but also gave me plenty of chances to understand the research and science In his lab, I have developed the ability to iii

PAGE 4

be an independent scientist learned through my failures and through the chances that were always there if I wanted to try again. My sincere appreciation is also extended to my supervisory committee, Dr. John Perrin, Dr. Kenneth Sloan, Dr. Richard Streiff, and Dr. Steven Baker, for their fine instruction and effective advice. I would like further to express my appreciation to Dr. William Weimar. It is with his kind assistance that the iron transport assays were done. In addition, I would like to express my thanks to Dr James McManis, who spent much time assisting with synthesis. I would also like to thank my many friends and colleagues in and out of the laboratory, Mr. Richard Smith, Mr. Sam Algee, Dr. K. A. Abboud, Ms Jan Wiegand, and Dr Christian Ludin for all their assistance. I will always be indebted to my parents, Jianbin Xin and Yingmei Li, and my parents-in law, Guowei Yao and Huimin Song, who have always stood behind me and cared for me no matter how far apart we were I am also very pleased to express my deepest gratitude to my wife, Yao Hua, whose unconditional love and full-hearted support was always there when I needed it most and deserved it least. Without her efforts during these years, this dissertation would not have been possible. iv

PAGE 5

TABLE OF CONTENTS PAGE ACKNOWLEDGEMENTS .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. iii LIST OF FIGURES .. .. .. .. .. ... .. .. .. ... .. .. ... .. ... .. ... .. ... .. vii LIST OF ABBREVIATIONS .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. x ABSTRACT . . . . . . . . . . . . . . . . . . xiii CHAPTER I. INTRODUCTION AND BACKGROUND ...... .............. .... 1 The Properties of Iron in Metabolism.......... .... ....... 1 Iron Metabolism in Humans................................. 4 Iron Overload in Humans........ .... .... .... .... .... ........ 6 Iron Utilization and Transport in Microorganisms...... 8 Iron and Bacterial Infection .......... .............. ......... 13 The Implications and Applications of Siderophores............... ... ........... ...... ........ ... 15 II. SYNTHESIS OF RHIZOFERRIN............ .... .. .... .. .. .... ... 22 Introduction ........ ..... ..... ,... ............ ... ............... 22 Synthesis... ............................................ ......... 25 Experimental. ..... . ... .................... ........ ..... 29 Discussion... ... .... ...... ...... ....... ...... ...... ... 35 Ill. SYNTHESIS OF STAPHYLOFERRIN A............ ...... ..... 40 Introduction........................ ..... ............... ....... 40 Synthesis...... ... ..... ....... ........... ........ .............. 41 Experimental... . . . . . . . . . . . . 44 Discussion...... ..... .......... ....................... .... ...... 47 V

PAGE 6

IV. EFFICIENT SYNTHESIS OF POLYAMINE CATECHOLAMIDES... ... ... .... ... ... ... ... ... ... ... ... . .. 50 Introduction. .. ... ......................... ........ ...... . 50 Synthesis.. .. .............. ..................................... 58 Experimental... . . . . . . . . . . . . 66 Discussion....... .................. .................... ...... 74 IV. THE EFFECTS OF RHIZOFERRIN, oAND LFLUVIABACTIN, L-AGROBACTIN AND LHOMOFLUVIABACTIN ON IRON TRANSPORT IN PARACOCCUS DENITRIFICANS..... ... .. . . .. ... ... . 86 Introduction............ ....... . ... ........................... 86 Experimental... . . . . . . . . . . . . 89 Results .................. ... ... .... ... .... ... ... ......... 93 Discussion. .. .... ....... .... ..... .. ... ....... .... ... 110 VI. CONCLUSIONS .................. ..... .. ... ... .. ............. 133 REFERENCES . . . . . . . . . . . . . . . . 135 BIOGRAPHICAL SKETCH 151 vi

PAGE 7

LIST OF FIGURES Figure Page 1-1 Fenton Reactions and Production of Hydroxyl Radicals.......... 3 1-2 Examples of Microbial Siderophores......................... ....... .. 10 1-3 Structure of Rhizoferrin.................. ....... .... .................. .. 11 1-4 Structures of L-Parabactin and L-Vibriobactin......... ... .......... 17 1-5 Structures of L-Fluviabactin and L-Agrobactin............... ...... 19 1-6 Structure of Staphyloferrin A...... ....................................... 20 2-1 Retrosynthetic Scheme of Rhizoferrin........ ......................... 24 2-2 X-ray of R-Enantiomer of 1 ,2-Dimethyl Citrate.................. 26 2-3 Synthesis of Rhizoferrin.. .................... ........................... .. 28 2-4 Hydrolysis of Triesters of Citric Acid with PLE or Subtilisin ... .. 37 2-5 The Diastereomeric Salts Formed Between 1 ,2-Dimethyl Citrate and Brucine........... ...................... ... .................... 39 3-1 Synthesis of Staphyloferrin A..................... ...... .......... ...... 43 3-2 Examples of Chiral Citric Acid-based Siderophores . 49 4-1 Structures of Two L-Fluviabactin Analogues ...... ......... .... .... 52 4-2 Synthesis of Benzyl-protected Triamine. ... .. ....................... 55 4-3 Synthesis of L-Parabactin Using N4-Benzylspermidine... ... ... .. 57 4-4 Synthesis of L-Vibriobactin Using Diprotected Triamine. ...... 58 4-5 COi Coupling Reactions......................... .................. ........ 61 vii

PAGE 8

4-6 Synthesis of Reagents Ethyl 2,3-Dihydroxybenzimidate.. ... .. 63 4-7 Synthesis of L-and oFluviabactin, L-Agrobactin, LHomofluviabactin............... ... ...... ..... ........... .... ......... ... ... 65 4-8 300 MHz 1H NMR Spectrum of L-Fluviabactin in CD30D at 50C................. .... ............. ...................... ...... .... ..... ... 76 4-9 300 MHz 1H NMR NMR Spectrum of o-Fluviabactin in CD 30D at 50C... ... . ..... ... ...... ... ... ...... ... ...... ... ... ... ...... ... ... ... ... 77 4-10 300 MHz 1H NMR NMR Spectrum of L-Agrobactin i n CD30D at 50C... ... .... ... . ... ... ... ...... ... ...... ... ... ... ... ... ... 78 4-11 300 MHz 1H NMR NMR Spectrum of L-Homofluviabactin in CD30D at 50C........... ............................ ............ ........ ... 79 4-12 300 MHz 1H NMR NMR Spectrum of y-Methyl Group in LFluviabactin ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 82 4-13 300 MHz 1H NMR NMR Spectrum of y-Methyl Group in LAgrobactin..................... ...... ... .. .. ... ... ... ... .. ... .. ... .. .. .. 83 4-14 300 MHz 1H NMR NMR Spectrum of y-Methyl Group in LHomofluv i abactin... ... ... .. .. .. ... ... ... .. .. .. .. .. .. ... ... .. .... 84 4-15 Conformations of L-Agrobactin L-Fluviabactin, and L-Homofluviabactin... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...... 85 5-1 Growth Rate of P. denitrificans in 0.5 M Minimal Salts Solution and Induced Siderophore Production...... ................ 94 5-2 Growth Rate of P. denitrificans in the presence of 2 0 M Catecholamide Ligands. ..... ........ ... ................. ...... 96 5-3 Iron Uptake of [55Fe] from [55Fe]Ferric L-Fluviabactin, o Fluviabactin, L-Agrobactin, L-Homofluviabactin, L-Parabactin.. 97 5-4 K i netic Characteristics of Iron Transport of [55Fe]Ferric LFluviabactin and o-Fluviabactin ... .. . .. ... ... .. . . ... . . 99 5-5 Kinetic of Iron Transport of [55Fe]Ferric L-Fluviabactin .......... 101 5-6 Kinetic of Iron Transport of [55Fe ]Ferric L-Agrobactin.... .. .... .... 103 viii

PAGE 9

5-7 Kinetic of Iron Transport of[55Fe]Ferric L-Homofluviabactin... .. 104 5-8 CD Spectra of Ferric L-and o-Fluviabactin .......... .-. ... .. .. .. .. 107 5-9 Configuration of Ferric L-Fluviabactin... ... .. . .. ... ... . . 108 5-10 Configuration of Ferric o-Fluviabactin... .. . . .. ... ... .. . 109 5-11 Growth Rate of P. denitrificans in the presence of 1.0 M Rhizoferrin and Parabactin.... .. . . ... ... ... ... ... . . .. .. 111 5-12 Structures of the Catechol Compounds Isolated from P. denitrificans... ... ... .. .. .. .. ... .. ... . .. .. ... ... ... ... .. .. 114 ix

PAGE 10

AcOEt [a]25D Anal. aq Ar ATCC atm, atmos ATP BOC BOC-ON bp anti Bu Bz BzOH C oc ca Cale (d) Cbz CD CDCb COi CH2 CH3 CH2Cl2 CHCb Chem CH30H CH2Ph Ci cis cm C02CH3 o (delta) LIST OF ABBREVIATIONS ethyl acetate specific optical rotation at 25C for D (sodium) line in a 1 decimeter cell, neat. analytic aqueous aryl American Type Culture Collection atmosphere(s), atmospheric adenosine triphosphate tert-butoxycarbonyl 2-(tert-butoxycarbonyloxyimino)-2-phenylacetonitrile boiling point stereodesciptor butyl benzoyl C5HsCObenzoic acid concentration (g/100 ml) centigrade degree; Celsius degree (circa) about calculate, calculated carbobenzoxy circular dichroism deutero-chloroform 1, 1 '-carbonyldiimidazole methylene methyl methylene chloride chloroform chemical methanol benzyl curie stereodescriptor centimeter( s) methyl ester chemical shift X

PAGE 11

~E (delta epsilon) DCC DCU DFO DMAP DMF DMSO DNA (E)E (epsilon) Ed(s) EDDA e.g. Et EhN EhO EtOH FMN (FMNH2) g g/1 GI Hb HBr HCI H20 HOAc HPLC h i.e. I "A (lambda) L Me MeOH mg MgSQ4 min ml mm mp MS N NaH NaOH NH2NH2 NH3 molar ellipticity 1,3-dicyclohexylcarbodiimide 1,3-dicyclohexylurea desferrioxamine B dimethylaminopyridine dimethylformide dimethyl sulfoxide deoxyribonucleic acid entgegen stereodescriptor molar extinction coefficient (M"1cm1 ) editor(s) ethylenediamine di-o-hydroxyphenylacetic acid (exempli gratia) for example ethyl triethylamine diethyl ether ethyl alcohol flavin mononucleotide (reduced form) grams grams per liter gastrointestinal hemoglobin hydrogen bromide hydrogen chloride water acetic acid high performance liquid chromatography hour (id est) that is liter wavelength ligands methyl methanol milligram magnesium sulfate minute(s) milliliter ( cubic centimeter) millimeter melting point mass spectrometry normal ( equivalent per liter) sodium hydride sodium hydroxide hydrazine ammonia xi

PAGE 12

NMR OD 0Su p. PAGE Pd/C Ph pH ppm Pr (R) RNA mRNA (S)satd Sec soln tert TFA THF Thr TLC Tris UV wt nuclear magnetic resonance optical density N-hydroxysuccinimide page(s) polyacrylamide gel electrophoresis palladium carbon phenyl acid-base scale: log of reciprocal of hydrogen ion concentrtion parts per million propyl (normal) rectus (right) stereodescriptor ribonucleic acid messenger RNA sinister (left) stereodescriptor saturated secondary solution tertiary trifluoroacetic acid tetrahydrofuran threonine thin-layer chromatography tris(hydroxymethyl)-aminomethane, 2-amino-2hydroxymethylpropane-1,3-diol ultraviolet weight xii

PAGE 13

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SYNTHESIS OF RHIZOFERRIN AND POL YAMINE CATECHOLAMIDES AND THE STUDY OF THEIR EFFECTS ON IRON TRANSPORT IN PARACOCCUS DENITRIFICANS By Mei Guo Xin December 2000 Chairman: Raymond Bergeron Major Department: Medicinal Chemistry The synthesis of rhizoferrin, a novel siderophore isolated from Rhizopus microsporus var. rhizopodiformis, is presented The absolute configurations of two chiral citric acid moieties in this natural product were confirmed to be (R,R) by X-ray crystallography. Also, synthesis of staphyloferrin A, a new siderophore isolated from Staphylococcus hyicus DSM 20459, was completed. The absolute configurations of two chiral citrates in staphyloferrin A were also verified as (R R). Furthermore, fluviabactin, a polyamine catecholamide isolated from Vibrio f/uvials, and its analogs and homologues (L-agrobactin, L-homofluviabactin and o-fluviabactin) were synthesized. The effects of rhizoferrin and the polyamine catecholamides on iron transport and growth of Paracoccus denitrificans were studied Rhizoferrin was unable to promote the bacterial growth when the iron concentration in the medium was low. On the other hand, L-fluviabactin, L-agrobactin, and Lxiii

PAGE 14

homofluviabactin derived from L-threonine were able to stimulate the growth rate of Paracoccus denitrificans, whereas o-fluviabactin did not show any promoting effects on bacterial growth In iron accumulation studies with 55Fe radiolabeled materials, Paracoccus denitrificans acquired 55Fe from 55Fe ferric L-fluviabactin, ferric L-agrobactin, and ferric L-homofluviabactin much more efficiently than from 55Fe ferric o-fluviabactin Furthermore, in kinetic studies, all ferric polyamine catecholamide chelators derived from L-threonine displayed both high-affinity and low-affinity components on the Lineweaver-Burk double reciprocal plots, while ferric o-fluviabactin lacked a high-affinity iron transport The CD spectra of the two chelates, ferric L-fluviabactin and ferric D-fluviabactin, were essentially mirror images. This suggests that it is the change of stereochemistry of oxazoline ring and not the polyamine backbone of the polyamine catecholamides that most affects iron transport in Paracoccus denitrificans, and that the molecular asymmetries of these chelates may contribute to their distinctive kinetic features. xiv

PAGE 15

CHAPTER I INTRODUCTION AND BACKGROUND The Properties of Iron in Metabolism Virtually all living cells require iron as a nutrient. It plays critical roles in many biological redox processes Although iron has a variety of oxidation states, from -2 to +6, it is most simply characterized by the Fe(ll)/Fe(III) redox couple When appropriately coordinated, Fe(II) can easily and reversibly be converted to Fe(III), and act as an electron donor or acceptor. The main advantage of iron, however is that in its interaction with coordinating ligands the properties of iron can be modified. Consequently, the ease of electron acceptance can vary over a wide range. There are six molecular orbitals available for cr bonding, and these can form bonds along the three axes of symmetry, thus giving the octahedral geometry commonly found in iron complexes (1). The equilibrium between Fe(II) and Fe(III), however, is very sensitive to both pH and the nature of the interacting ligands (2). Because the metal serves in a variety of biological redox systems such as cytochromes, oxidases, peroxidases, ribonucleotide reductase (3,4), and various other enzymes as essential cofactors, this sensitivity has been exploited at a cellular level. Despite its prevalence in biological redox systems, there are numerous problems associated with iron in metabolism. One is related to its solubility 1

PAGE 16

2 (5). Iron in early biomolecules was essentially in the Fe(II) state due to a reducing atmosphere. With the evolution of photosynthesizing blue-green algae a surplus of oxygen was produced which resulted in the conversion of Fe(II) to Fe(III). This shift in favor of a higher oxidation state had profound effects on the solubility of iron While both Fe(II) and Fe(III) exhibit exceedingly high affinities for hydroxyl ions in aqueous solution the two oxidation states 2+ 3+ exist as Fe(H20)6 and Fe(H20)6 which readily hydrolyze and polymerize at physiological pH values, forming insoluble complexes. In the case of Fe(III), 38 the solubility product may be less than 10 (6) This translates to a maximum -18 3+ 2+ solubility of 1 O M for Fe(H20)6 In contrast the solub i lity of Fe(H20)6 -1 permits solutions of 10 M Thus, living systems have had to develop rather sophisticated methods for acquiring iron from ferric hydroxides, the predominant oxidation state prevailing in an aerobic environment. While mammalian cells use proteins for example, transferrin (7) or ferritin (8), to solve this problem microorganisms produce a group of relatively low molecular weight virtually iron-specific ligands, termed siderophores, for the purpose of accessing and utilizing this transition metal (9 10 11) The main problem with iron metabolism, however, may be related to its toxicity (12). The reaction between iron and oxygen produces hydroxyl radicals and this highly reactive species reacts with biological membranes and combines with a number of natural products to produce carcinogens (13, 14). The origins of hydroxyl radical generation lie in the existence of redox cycling iron and the reaction driven by the superoxide anion. The superoxide anion can

PAGE 17

3 either be transformed to hydrogen peroxide and oxygen by superoxide dismutase or react with Fe(III), reducing Fe(III) to Fe(II). The problematic species in this chemistry is hydrogen peroxide. Hydrogen peroxide is generally converted to water by glutathione peroxidase or decomposed to water and oxygen by catalase. However, when the level of hydrogen peroxide is above the Km of these enzymes and some Fe(II) is available, hydrogen peroxide will be converted to the hydroxyl radical. Furthermore, this event can be cyclical, as shown in the following equations: 02 + LFe(III) 02 + LFe(II), 2 02 + 2 H+ H202 + 02 H+ + LFe(II) + H202 HQ+ LFe(III) + H20. Figure 1-1. Fenton Reactions and Production of Hydroxyl Radicals. Hydroxyl radical-related cell damage is seen in inflammatory bowel disease (15) as well as in certain neurodegenerative disorders, including Alzheimer and -Parkinson's (16, 17). While iron is the second most common metal in the earth's crust (5% of total elements), it is not readily extracted from its insoluble oxides by living cells, thus its bioavailability is generally limited and higher species often exhibit

PAGE 18

4 deficiency states (18). Paradoxically, iron overload conditions also occur (19). Iron overload disorders have been widely associated with many diseases, including Cooley's anemia (20,21) and often need special chelation therapy (22,23) Considering its relative insolubility, the question has become why does iron overload occur in living cells? Iron Metabolism in Humans Eukaryotes have developed large proteinaceous molecules to solve the access, storage, and utilization problems associated with iron. The iron proteins in humans may be broadly grouped as heme proteins, iron flavoproteins, and a heterogeneous group of proteins (24). Iron plays important roles in many biological processes in humans. For example, iron is found at the active centers of iron-binding proteins responsible for oxygen storage and transport (24). Iron i s important for electron transport in the heme-containing protein, cytochrome (25). This metal is also found i n a large number of nonheme-containing enzymes including various oxidases, reductases, and dehydrogenases (25). Iron metabolism in humans is generally divided into different functional compartments. The largest iron compartment i n humans is hemoglobin iron, normally containing approximately 67% of the 4 g of iron present in a 70-kg man (25). This heme-containing protein found in erythrocytes is responsible for reversibly binding oxygen in the lung, as well as carrying oxygen to tissues via the circulatory system (25).

PAGE 19

5 The iron storage compartment in humans has two distinct forms of proteins: ferritin and hemosiderin. Ferritin is a water-soluble complex of ferric hydroxide and a protein, apoferritin. Apoferritin forms a shell within which ferric ions, hydroxyl ions, and oxygen are dispersed in a lattice-like relationship (26 ). Hemosiderin, a water-insoluble iron-storage protein, is found predominantly in cells of the monocyte-macrophage system, which contains approximately 25 to 30 percent iron by weight (8). Under pathologic conditions, it may accumulate in large quantities in almost every tissue of the body (8). Myoglobin is structurally similar to hemoglobin, but it is monomeric: each myoglobin molecule consists of a heme group surrounded by loops of a long polypeptide chain. Myoglobin is present in small amounts in all skeletal and cardiac muscle cells, and may serve as an oxygen reservoir to protect against cellular injury during periods of oxygen deprivation (25). Based on its total iron content, normally 3 0 mg (25), the iron transport compartment in humans is the smallest of all human iron compartments, but it is kinetically the most active, since its iron is replaced at least 10 times every 24 hours (25). This iron is bound to the specific protein transferrin, a somewhat elongated glycoprotein. Transferrin has a molecular weight of 80 kDa, and binds two molecules of ferric ion tightly but reversibly. This iron shuttle protein is found in the serum and is also present in various extracellular fluids in the body (7,27). In addition to the iron storage protein, ferritin, humans have evolved a highly efficient iron transport and delivery system using transferrin. The

PAGE 20

6 regulation of iron mobilization and storage proteins occurs at the translational level. Transferrin carries iron into the cell via the transferrin receptor. The receptor message RNAs of both ferritin and transferrin contain an iron regulatory element in untranslated regions of the messager RNA (28,29). When the iron concentration in cells is low, an iron regulatory protein (30) binds to the ferritin iron regulatory element, and the translation is blocked. However, when an iron regulatory protein binds to the iron regulatory element of transferrin receptor message RNAs the half-life of the messenger is increased, thus there is a decrease in ferritin synthesis and an increase in the transferrin receptors on the cell surface (28,31). Iron Overload in Humans Iron metabolism in primates is characterized by a high l y efficient recycling process with no specific mechanism for eliminating and absorbing this transition metal. A healthy adult male has a total of 4 g of iron in his body Iron absorption is most efficient in the duodenum and becomes progressively less so further along the alimentary canal (32,33). On the other hand, iron excretion is relatively fixed, less than one one-thousandth of the iron absorbed is lost each day (34). Thus, iron balance is actually maintained by the extremely efficient recycling of the iron already present in the body. Iron requirements are largely determined by the amount lost from the body. Normally, the small obligatory loss due to desquamation from epithelial surfaces, gastrointestinal blood loss, and biliary excretion amounts to about 1 mg daily or less, and menstrual loss averages about 15-20 mg monthly (19). It

PAGE 21

7 is possible to compensate for all physiological losses with a dietary iron intake of 11 mg daily Iron balance is maintained by physiological adjustments of the absorption mechanism so that the amount of iron crossing the epithelial barrier of small intestines is related to, and regulated by, internal iron status (35). Iron overload may be caused by increased iron absorption. For example, there may be an increase in the amount of available dietary iron or increased absorptive activity by the gut epithelium. An increase in iron content of the diet alone may not result in iron loading. Iron overload is often associated with dietary siderosis (18) or a genetic metabolic disorder, for example, idiopathic haemochromatosis (36). Iron overload also may be caused by transfusional iron loading, as seen in patients with severe anemia requiring repeated blood transfusions (37). The iron overload disorders are often classified as primary and secondary hemochromatosis. Primary hemochromatosis refers to an inappropriately increased mucosal absorption and is easily treated by periodic venesection (38). Secondary hemochromatosis, however, presents a much more complicated problem It refers to patients with iron overload secondary to anemia, who require repeated blood transfusions (39). Since the normal lifespan of a red blood cell is only about 120 days, the iron from the transfused red cells is eventually removed by the reticuloendothelial cells of the spleen and stored in the reticuloendothelium or in the parenchyma cells of the liver (19). However, the efficient iron-recycling system in humans ensures that little

PAGE 22

8 of the iron introduced by transfused red cells is ever excreted. The result is an increase in the total amount of iron stored in the body, and eventually iron can reach toxic levels. The most common anemia of the secondary hemochromatoses is ~ thalassemia (39). Patients with this disease require repeated blood transfusions until the iron in their body is overloaded. Most of them die from the toxic effects of iron overload in their second or third decade. Obviously, this excess iron cannot be removed by phlebotomy, as in the case of primary hemochromatosis, because the origin of the excess iron is the transfused red blood cells. Consequently, the only alternative is chelation therapy (39). Desferrioxamine B (40), a siderophore from Streptomyces pilosus, has been used for the treatment of ~-thalassemia for the last 30 years. This drug has a number of shortcomings, such as its marginal oral activity and its poor to moderate efficiency, as well as its high cost. Thus, considerable effort has been invested in the search for alternative iron chelators from siderophores and other synthetic ligands. Iron Utilization and Transport in Microorganisms Except for certain members of the lactobacilli, iron is a nutrient required by all microorganisms. Iron plays important roles in many microbial biological processes, such as electron transport, tricarboxylic acid cycle, and DNA biosynthesis. For example, iron is involved in the electron transport metabolism of microorganisms because of the presence of cytochromes and non-heme iron in the respiratory chains of aerobic and anaerobic species. Iron

PAGE 23

9 sulfur protein has been considered as fully equivalent to that of the cytochromes in microbial metabolism (41). Iron also plays a critical role in ribonucleotide reductase, the enzyme responsible for synthesis of deoxyribonucleotides required for DNA formation (41,42). Although iron in its hydroxide polymer is profoundly insoluble, most microbes have developed clever methods to deal with this problem. It is generally accepted that microorganisms utilize two kinds of iron uptake systems, low-affinity systems and high-affinity systems (41), to acquire iron from the environment. Low-affinity systems are only seen when the levels of iron are high. No specific solublizing and transporting compounds or membrane receptors seem to be required. More importantly, microbes have developed high-affinity systems for the assimilation of iron for use when the concentrations of iron are low High-affinity systems consist of two important parts: siderophores and the transport apparatus. Siderophores are low molecular weight, iron-specific ligands produced by microorganisms under iron-starved conditions. They form tight but soluble complexes with Fe(III). More than 100 sideophores have been isolated from bacteria and fungi. Most of these siderophores fall into two basic families: hydroxamates and catecholamides (43). The catecholamides include both cyclic tricatechols as exemplified by enterobactin and many linear polyamine containing catecholamides. The hydroxamates consist of a number of structurally different compounds including desferrioxamines, rhodotorulic acid, and citrate-containing ligands. Examples of these two families include

PAGE 24

10 hydroxamate desferrioxamine B (40) and catecholamide enterobactin (44) (Figure 1-2) ............ ~H H/O ~~N'\O OH n ~N Ho H N 0 ~NH2 0 0 Desferrioxamine B ( Streptomyces pilosus) 0 OH HN~OH oYJ OH O r01 .~ 0 OH HOvN~O~N~OH I H o H I Enterobactin (Escherichia coli & Salmonella typhimurium) Figure 1-2. Examples of Microbial Siderophores.

PAGE 25

1 1 There are also many siderophores that do not belong to either family, for example, rhizoferrin [N1 N4-bis(1-oxo-3-hydroxy-3,4-dicarboxybutyl) diaminobutane] isolated from Rhizopus microsporus var. rhizopodiformis (45) is based on citric acid. Rhizoferrin has a putrescine backbone that is symmetrically diacylated by citric acid at its 1-carboxylate. The configurations of two asymmetric carbons of citrates in rhizoferrin are suggested to be (R, R) by circular dichroism (CD) spectroscopy. Rhizoferrin is neither a hydroxamate, nor a catecholamide (Figure 1-3). H OHO pooH HOOC~N~N~COOH Hooe OH O H Rhizoferrin (Rhizopus microsporus var. rhizopodiformis) Figure 1-3. Structure of Rhizoferrin.

PAGE 26

12 In the transport apparatus the primary transport sites are the outer membrane receptor proteins (46). These receptors are induced in a low iron concentration, and are able to recognize the siderophore-iron complex on the surface of the cells. Typically, these proteins are in the molecular mass range of 70 to 90 kDa. For example, in Escherichia coli, an 80-kDa outer membrane ferric enterobactin receptor has been purified and characterized with a Kd of approximately 0.3 M (47,48). Strong genetic evidence supports the proposed role in iron transport of the ferric enterobactin receptor encoded by the fepA gene, which has been mapped and sequenced (49). Most iron transport mechanisms in microorganisms are associated with the interactions between siderophores and outer membrane receptors. Studies of iron transport are a useful means to examine the microbial iron utilization facilitated by siderophores and the matching membrane receptors, including the stereospecificity of iron transport mechanisms. A number of methods have been developed for iron transport assays (50,51). For example, the fate of the siderophore-iron complexes can be examined by using double radio-labeled materials. The natural siderophores can be modified by chemical synthesis so that microbial responses to different ligands can be characterized by the chemistry of those ligands. The redox potentials of Fe(II) and Fe(III) complexes can be measured, thus a possible reduction mechanism involved in microbial iron transport can be studied In addition, chelators such as siderophores usually form colorful complexes with iron in solution.

PAGE 27

13 Consequently these complexes can be followed by visible spectroscopy, such as UV and circular dichroism. The study of iron transport mechanisms in microorganisms has become an active field of research in the past two decades. First, iron transport in prokaryotic species such as Escherichia coli provides an ideal opportunity for the application of the powerful techniques of modern molecular biology. Interest in this research has extended beyond iron transport to more general transmembrane transport processes and their regulation. Second, the systems used by microbes to gather iron have stimulated the search for substances that deferrate human patients who suffer from iron overload (52,53) or actinide poisoning (54) Finally, and more importantly, iron has been identified as a virulence factor within selected pathogenic species. The capacity of invading pathogens to acquire enough of this metal to satisfy demands of growth may constitute one aspect of virulence and pathogenicity (55). Iron and Bacterial Infection The role of iron in infection has been an interesting topic for many years In an early report, a possible association between iron and infection was indicated when Schade and Caroline demonstrated that iron could overcome the bacteriostatic effect of egg white (56). Subsequent studies demonstrated that siderophore production is a common trait among pathogenic microorganisms and suggested that siderophores may be important virulence factors in animal and plant disease (57 58).

PAGE 28

14 Virulence factors have been described as substances that are either directly toxic to the host or antagonize the antibacterial mechanisms of the host. A pathogen does not have to contend with the insolubility of ferric iron, but can obtain its iron from its host. Although all the iron in human plasma is essentially sequestered by such iron-binding proteins as transferrin and ferritin, a pathogen can still successfully compete with those iron-binding proteins by way of siderophore production. It has been assumed that the production of siderophores of either the hydroxamate or catechol type might facilitate microbial growth by removing non-transferrin-bound iron from plasma. For example, it has been showed that the inability to synthesize enterobactin diminishes both the virulence of Salmonella typhimurium in laboratory animals and the capacity of the organism to grow in human serum (59) The presence of enterobactin-specific immunoglobulin was reported in normal human serum (60). In other examples when the rabbit pathogen Pasteurella multocida was injected into live organisms a dramatic fall in plasma iron and elevated body temperature (61) was obtained. The critical temperature sensitivity of siderophore production was also observed in a feverish man (62). It suggested that fever might be a host defense mechanism designed to deprive the pathogen of iron. Virulence is dependent upon a multitude of interactions between a host and a pathogen In specific circumstances, iron assimilation may emerge as the critical element in the host-pathogen interplay While the unambiguous correlation of iron assimilation with virulence remains to be established, the

PAGE 29

15 virulence of many pathogenic organisms has been closely associated with their ability to synthesize siderophores. Siderophores apparently antagonize the iron-restricting mechanism of the host, an antibacterial defense mechanism (61,62). It is not surprising that agents preventing organisms from utilizing iron also prevent their growth (63). This implies that iron chelators could be used in antimicrobial chemotherapy. Iron deprivation of a microorganism as a result of iron chelation therapy is based on the premise that microorganisms have receptors specific to their own siderophores. Thus modified natural siderophores or other synthetic ligands have emerged as potential candidates. The Implications and Applications of Siderophores Siderophores are iron-specific ligands produced by microorganisms in a low iron concentration environment. They have relatively low molecular weights and form very tight complex with iron, for example, the stability constant of ferric enterobactin is as high as 1052 The main function of a siderophore is to form a soluble complex with ferric ion; thus, a microorganism can use a siderophore to obtain this essential metal from the environment. Interestingly, this ability of a siderophore has been exploited in clinical use for the treatment of iron overload secondary to anemia (64-66). Although substantial efforts have been made to treat iron overload anemia, subcutaneous infusion of desferrioxamine B (DFO), a hexacoordinate hydroxamate iron chelator produced by Streptomyces pilosus (40) is still regarded as the method of choice for handling transfusional iron overload.

PAGE 30

16 However, thi s drug suffers from a number of shortcomings, such as its high cost of production, poor to moderate efficiency, and marginal oral activity. Furthermore DFO has a very short half life in the body, and therefore must be administered by continuous subcutaneous infusion over long periods of time. In addition, it can be very immunogenic (67), and patient compliance has become a real problem (68). The use of an orally effective iron chelator has been a therapeutic strategy for many years Although a number of synthetic ligands have been studied in recent years as potential orally active therapeutics, for example pyridoxal isonicotinic hydrazone (69), hydroxypyridones (70) and bis(o-hydroxybenzyl) ethylenediaminediacetic acid analogs (71) none has yet proven to be satisfactory. Interestingly, siderophores, microbial iron chelators, a l so provide model compounds as orally active iron chelators. One example is 2-(3' hydroxypyrid-2'-yl) 4-methylthiazoline-4(S)-carboxylic acid ( desferrithiocin, OFT) isolated from Streptomyces antibioticus (72). Studies in a bile duct-cannulated rat model as well as in a cebus monkey model suggested that it was indeed an orally active iron chelator, but it exhibited nephrotoxicity (73). Moreover, polyamine catecholamides such as parabactin (7 4) and vibriobactin (75) (Figure 1-4) have demonstrated bacteriostatic activities in some bacteria (76). It was concluded that these compounds inhibited bacterial growth because they sequestered iron and the bacteria were unable to use the siderophore-iron complexes (76) Based on this and other studies, iron deprivation has emerged as a strategy in antimicrobial therapy For example,

PAGE 31

17 n _I .. ,OH 0 ~N LJ 0 / -~.f O OH HO~~.._,,-...._,,N~N~OH OHO H u L-Parabactin (Paracoccus denitrificans) -...::::: OH OH I A -...::::: OH ;-J.)(~~N~~NOH 0 0 OH Vibriobactin (Vibrio cholerae) Figure 1-4 Structures of L-Parabactin and Vibriobactin

PAGE 32

18 an enantiomeric enterobactin has been synthesized and its bacteriostatic activity demonstrated against Eschericha coli (77). These initial findings prompted the study of the importance of the polyamine backbone and the stereochemistry of the oxazoline ring of a polyamine catecholamide siderophores in microbial iron metabolism. Of the more than 100 siderophores that have been isolated, their usefulness as iron-clearing agents has not at all paralleled the rate of their isolation and structural elucidation. For example, among polyamine catecholamides, parabactin, isolated from Paracoccus denitrificans (74), was studied and shown to remove iron from human transferrin in vitro (78). Furthermore, in a bile duct-cannuated rat model, parabactin was shown to be far more efficient than DFO at removing Fe from the animal (64) N4-[2-(2, 3-Dihydroxyphenyl)-trans-5-methyl-2-oxazoline-4-yl]carbonyl-N 1 N 7 -bis(2, 3-dihyd roxybenzoyl) norspermidine (Fluviabactin) is a new siderophores isolated from Vibrio f/uvias (79). It is structurally similar to agrobactin, but contains a norspermidine backbone instead of spermidine in agrobactin. (Figure 1-5). Fluviabactin is utilized in iron acquisition by Vibrio f/uvias, a pathogen that is widely associated with pediatric enterocolitis (80) V. fluvias is a halophilic, polarly flagellated, gram-negative rod, and grows well on the media supplemented with sodium chloride Since 1981, 14 cases of enterocolitis associated with V. f/uvias have been reported in the United States (80).

PAGE 33

(;!OH ._l~,-OH 0 N 19 )-{to O OH HO~~~N~NJVvoH OHO H u L-Agrobactin (Agrobacterium tumefaciens) (XOH _r .. -OH 0 ~N ~o H 'f H HO~N~N~NyY'oH OH O O OH L-Fluviabactin (Vibrio f/uvialis) Figure 1-5. Structures of L-Fluviabactin and L-Agrobactin.

PAGE 34

20 An interesting phenomenon in siderophore chemistry is that some varieties of siderophores have a citric acid component. For example, the hydroxamates aerobactin, arthrobactin, schizokinen (81), and nannochelin (82) have a symmetrically 1,3-disubstituted citric acid; whereas rhizoferrin (45) and staphyloferrin A (83), isolated from Rhizopus microsporus var rhizopodiformis and Staphylococcus hyicus, respectively, have two asymmetrically functionalized citric acids (Figure 1-6) This implies that citric acid might play an important role in microbial iron utilization. In addition, the chiral citric acid based siderophores for example, rhizoferrin and staphyloferrin A, provide new synthetic challenges for chemical synthesis H Hooq_ H ~o HO ,,COOH N -...._./ COOH HOOC~ ~N HOOc/'oH H Staphyloferrin A ( Staphylococcus hyicus) Figure 1-6 Structure of Staphyloferrin A.

PAGE 35

21 This present study involves the synthesis of rhizoferrin, staphyloferrin A, and fluviabactin as well as fluviabactin analogs. The absolute configurations of two chiral citrates in rhizoferrin and staphyloferrin A were confirmed by X-ray crystallography. An efficient synthetic scheme was developed for the synthesis of fluviabactin and its analogues in order to provide enough materials for animal tests in a bile duct-cannuated rat model. The study also includes the iron transport assays of 55Fe radio-labeled compounds in Paracoccus denitrificans, including growth rate, iron accumulation, and kinetic studies, in order to examine the issue of cross utilization of microbial ferric siderophores, including stereospecificity of the iron transport mechanism.

PAGE 36

CHAPTER II SYNTHESIS OF RHIZOFERRIN Introduction In response to a low iron concentration microorganisms produce siderophores, low molecular weight and iron-specific ligands Siderophores can form water-soluble siderophore-iron complexes. Interestingly, this ability of siderophores has been found to be of clinical use for the treatment of iron overload diseases (64,65). Due to the critical role siderophores play in microbial growth processes and the potential they offer as therapeutic agents, they have received intense attention from chemists (84) and biochemists (85). A rather large number of siderophores have been isolated For the most part, they are separated into two basic structural groups: hydroxamates and catecholamides (43) Although these siderophores vary substantially in overall structure, molecules of both classes are usually predicated on their polyamine backbones, specifically 1,4-diaminobutane (putrescine), 1,5-diaminopentane (cadaverine), norspermidine or spermidine, or on their biochemical precursor ornithine or lysine. The presence of such common structural units has led to the efficient total synthesis of the siderophores parabactin (86) and desferrioxamine (DFO) (84) in this and other laboratories. There are additional groups of siderophores that do not belong to either of these two major families. Examples include a variety of citrate-based 22

PAGE 37

23 siderophores Citrate in those siderophores is either symmetrically disubstituted, as seen in aerobactin, arthrobactin, schizokinen (81), and nannochelin (82), or unsymmetrically monosubstituted, as in rhizoferrin (45) and staphyloferrin A (83). The latter case is somewhat complicated since the prochiral carbon of the citrate became asymmetric because of the acylation thus giving rise to new synthetic challenges in the synthesis of these natural products. N1 N4-s is( 1-oxo-3-hydroxy-3 ,4-d icarboxybutyl)d iam inobutane ( rhizoferrin) has a putrescine center symmetrically diacylated by citric acid at its 1carboxylate. It is a hydroxyl polycarboxylate along with rhizobactin (87) and staphyloferrin A, which are predicated on L-lysine and o-ornithine respectively. The configurations of two asymmetric carbons of the citrates in rhizoferrin are suggested to be (R,R) by comparing its circular dichroism (CD) spectroscopy with that of natural (R,R)-tartaric acid (88). Rhizoferrin was first isolated from Rhizopus microsporus var rhizopodiformis an organism associated with mucormycosis seen in dialysis patients, and occurs in several Zygomycete strains of fungi (89) Like the natural chelators parabactin and DFO rhizoferrin forms a 1 : 1 complex with ferric ion (88), however, the formation constant of ferric rhizoferrin has not been measured, and the absolute configurations of the two chiral citrate centers in rhizoferrin were not yet confirmed. The first total synthesis of this natural product could also be the first example of the conversion of a chiral citric acid fragment to a chelator

PAGE 38

24 Retrosynthetic analysis revealed there are two components in rhizoferrin: putrescine and chiral citrate (Figure 2-1). H OHO ,COOH HOOC~N~ Jl.XcooH ,/""'\I"' ,_.. II N HOOC' OH O H HCOOC~OH 3 ,,....,.... ,_. n H3COOC' OH o HO~_FOOCH3 HO~COOCH3 Figure 2-1. Retrosynthetic Scheme of Rhizoferrin.

PAGE 39

25 In this synthesis, putrescine has to be converted to N,N -dibenzylputrescine by condensation of putrescine with benzaldehyde, followed by reduction with sodium borohydride The principal challenge to the synthesis of rhizoferrin, however, was to access a citrate synthon of correct configuration for coupling to both termini of putrescine in order to unequivocally define the absolute configuration of the siderophore. Since citric acid is a prochiral molecule, it is necessary to convert citric acid into an appropriate 1,2-disubstituted citric acid. Although 1,2-dibenzyl citrate was considered and synthesized by reacting citric acid with benzyl alcohol and then kinetic hydrolysis with sodium hydroxide it does not form crystals with natural bases, for example brucine Consequently, 1,2-dimethyl citrate was utilized Synthesis (90) The synthesis of rhizoferrin began with trimethyl citrate (1) which was converted to 1,2-dimethyl citrate (2) by a sterically controlled saponification (91 ). The enantiomers of carboxylic acid (2) were converted to their (-)-brucine diastereomeric salts After five fractional crystallizations from water, the crystalline salt was shown by single crystal X-ray diffraction to contain 1,2dimethyl citrate in the R-configuration (Figure 2-2) Treatment of the salt with 1 N HCI and extraction with ethyl acetate furnished (R)-1 2-dimethyl citrate (~). The determination of the enantiomeric purity of a chiral citrate is another important issue in this synthesis. Both the racemic (2) and chiral rn) diesters were used to acylate (S)-(-)-sec-phenethyl alcohol (1, 3-d i cyclohexyl carbodiimide/catalytic DMAP/CH2Cl2 ) to give unsymmetrical triesters (~) and @),

PAGE 40

26 respectively. An examination of the methyl ester region in the 600 MHz NMR spectrum of (4) and (.Q) showed that the latter contained an enantiomeric excess (ee) of 99%. 038 07A Figure 2-2. X-Ray of R-Enantiomer of 1,2-Dimethyl Citric Acid.

PAGE 41

27 With the correct enantiomeric acid in hand, N1 ,N4-dibenzyl-1,4diaminobutane (92) was acylated with (J) (2 equivalents) utilizing diphenylphosphoryl azide (Et3N/DMF) (93). The diamide () was obtained in 26% yield after flash column chromatography, which removed by-products, including olefins due to elimination of the tertiary alcohol as indicated by 1 H NMR. The methyl esters of() were hydrolyzed with sodium hydroxide in aqueous methanol; acidification yielded N1 ,N4-dibenzyl rhizoferrin (I). Finally, since N-benzyl amides are resistant to hydrogenolysis (94), deprotection of tetraacid (?) under dissolving metal reduction conditions (Li/NH/THF) (95), protonation of the salts on a cation exchange resin column and purification on a C-18 reversed-phase column furnished the final product, rhizoferrin (Figure 2-3). The high-field 1H NMR and high-resolution mass spectrum of this synthetic compound were essentially identical to the published spectra of the natural product (89) The absolute configurations (R,R) of the synthetic sample and the natural material were identical, since both exhibited a negative Cotton effect at the same wavelength (88). It is fortunate that the correct crystalline diastereomeric (-)-brucine salt of 1,2dimethyl citrate crystallized out and that no subsequent synthetic step could compromise the stereochemical integrity of the chiral centers Since rhizoferrin can be considered an aspartic acid derivative that usually cyclize in peptide synthesis (96), the polycarboxylic acid side chain of rhizoferrin could catalyze the spontaneous cyclization process Indeed, rhizoferrin cyclizes upon standing through dehydration to imidorhizoferrin and

PAGE 42

28 j C, d, e (2.) R = H H CH3 (1) R=x Ph (1) R=H H CH3 (.5.) R = XPh Ph ( OHO ~OOR N II .. -\.. .... COOR (.6) R=Me] g (7 ) R = H ROOC~~N~ ROOC'. OH o l_ Ph Figure 2-3 Synthesis of Rhizoferrin. (a) NaOH/MeOH (39%); (b) (S)-(-)-sec-phenethyl alcohol/DCC/catalytic DMAP/CH2Cl2 ; (c) (-)-brucine; (d) fractional crystallization; (e) HCI; (f) N1 ,N4-dibenzyl-1,4-diaminobutane/DPPA/Et3N/DMF (26%); (g) NaOH/MeOH (77%); (h) Li/NH3!THF (64%)

PAGE 43

29 bis-imidorhizoferrin which possess one or two five-membered rings, respectively We observed by 1H NMR spectroscopy that the zero order rate constant for this ring formation at pH 5 0 is 6.9 x 10-2 h-1 At pH 3 our findings on the extent of cyclization were similar to the literature (88); thus the analytical data were obtained before this cyclization occurred. The synthetic methodology for rhizoferrin will also be used to prepare the hydroxy polycarboxylated siderophore staphyloferrin A, in which 0-ornithine is NCX,Nb-diacylated with citric acid at its 1-carboxylate. Thus, the configuration of the citrates in this amino acid chelator will be determined by total synthesis as presented in chapter Ill. Experimental Trimethyl citrate (1) was obtained from CTC Organics, Atlanta, GA. Other reagents were purchased from Aldrich Chemical Company, and were used as received Fisher Optima grade solvents were routinely employed. Silica gel 32-63 (40 M "flash") from Selecto, Inc. (Kennesaw, GA) or silica gel 60 (70-230 mesh) from EM Science (Darmstadt, Germany) was used for column chromatography. Optical rotations were run in CH30H at 589 nm (Na lamp) at room temperature with c as g of compound per 100 ml. 1 H NMR spectra were recorded at 300 or 600 MHz and run in the deuterated organic solvent indicated or in D20 with chemical shifts g i ven in parts per million downfield from tetramethylsilane or 3-(trimethylsilyl)propionic-2, 2, 3, 3-d4 acid sodium salt, respectively.

PAGE 44

30 X-Ray Diffraction Data were collected at 173 K on a Siemens SMART PLATFORM equipped with a CCD area detector and a graphite monochromator utilizing MoKa radiation (A= 0.71073 A). Cell parameters were refined using up to 6233 reflections. A hemisphere of data (1381 frames) was collected using the ffi scan method (0.3 frame width). The first 50 frames were remeasured at the end of data collection to monitor instrument and crystal stability (maximum correction on I was < 1 %). Psi scan absorption corrections were applied based on the entire data set. Circular Dichroism CD spectra were obtained with a Jasco Model J500C spectropolarimeter equipped with a Jasco IF-50011 interface and CompuAdd 286 computer; data collection and processing were performed with Jasco DP-500/PC System version 1.28 software. The cell path length was 2.00 cm. Ultraviolet Spectroscopy UV spectra were obtained with a Shimadzu UV-2501 PC equipped with an AST 486/33 computer data station. The cell path length was 1.00 cm. (R,S)1.2-Dimethyl Citrate (2) Compound (2) was prepared by a published method (91). Sodium hydroxide (0.1 N, 215 ml) was added to a solution of trimethyl citrate (10.0 g, 42. 7 mmol) in 50% aqueous CH30H (200 ml) over 2 hours with vigorous stirring at room temperature. The solution was concentrated to about 150 ml and extracted with EtOAc (3 x 150 ml). The aqueous layer was acidified with 1

PAGE 45

31 N HCI (45 ml) and extracted with EtOAc (3 x 150 ml). The organic layer was dried (MgS04 ) and concentrated, providing 3.70 g (39%) of (Z) as a colorless oil: 1H NMR (d6-DMS0) 3 5.60 (br s, 1 H, OH), 3.64 (s, 3 H, C02CH3), 3.57 (s, 3 H, C02CH3), 2.87 (d, 1 H, J = 15 Hz, 1/2 CH), 2.81 (d, 1 H, J = 15 Hz, 1/2 CH), 2.73 (d, 1 H, J = 15 Hz, 1/2 CH2), 2.65 (d, 1 H, J = 15 Hz, 1/2 CHJ (-)-Brucine Salt of (R)-1,2-Dimethyl Citrate To a solution of (-)-brucine (12 5 g, 31.8 mmol) (CAUTION: toxic) in EtOAc (460 ml) was added (2) (7 g, 31.8 mmol) with vigorous stirring overnight. After filtration the precipitate (10.5 g) was recrystallized from water (5 x) and dried to afford 2 04 g of white crystals: mp 165-168 C. The diastereomeric salt crystallizes in the monoclinic space group C2 and has cell dimensions : a= 13.8947 (3), b = 12.4224 (3), and c = 17.5408 (3) A; a= 90, = 104 556 (1), and y = 90 The structure was solved by the Direct Methods in SHElXTl20 and was refined using full matrix least squares The non-H atoms were treated anisotropically. The methyl hydrogen atoms were calculated in ideal positions and were riding on their respective carbon atoms; the rest of the H atoms were refined without constraints. Two water molecules were located in the asymmetric unit. One was refined with full occupancy, and its H atoms were located. The other, located on a 2-fold axis of rotation, was refined to a 30% occupancy An absolute configuration of (R) was assigned to the citrate portion of the salt based on knowledge of the stereochemistry of brucine. Parameters (521) were refined in the final cycle of refinement using

PAGE 46

32 3855 reflections with I > 2cr(I) to yield Rand wR of 0 0434 and 0.1040, respectively. Refinement was done using F2 (R)-1,2-Dimethyl Citrate (.3.) HCI (1 N 4 ml) was added to a solution of the (-)-brucine salt of (R)-1 2dimethyl citrate (2 04 g, 3.32 mmol) in water (50 ml), and stirring was continued for 5 min. Extraction with EtOAc (3 x 50 ml), drying over Na2S04 and concentration gave 630 mg (86%) of(~ as a colorless oil: [a] +4.0 (c 1 00); the NMR was i dentical to (2). (R. S) 1,2-Dimethyl-3-[(S)-sec-Phenethyl] Citrate (~) 1,3-Dicyclohexylcarbodiimide (103 mg, 0 5 mmol) was added to a solution of (2) (110 mg, 0.5 mmol), (S)-(-)-sec-phenethyl alcohol (61 mg, 0.5 mmol) and 4-dimethylaminopyridine (3 mg) in dry CH2Cl2 (10 ml) at O C, and the mixture was stirred overnight. The mixture was filtered, and the filtrate was concentrated and purified by flash chromatography (1 :2 EtOAc/hexane) 1 resulting in 60 mg (37%) of (4) as a colorless oil: H NMR (CDCl3 ) 8 7 35-7 28 (m, Ph) 5 97 (q, J = 7 Hz, CHPh), 5 88 (q, J = 7 Hz, CHPh), 3.77 (s CH30), 3.73 (s, CH30), 3.69 (s, CH30), 3.68 (s, CH30), 2 98-2.74 (m, CH2), 1.54 (d, J = 7 Hz, C-CH3), 1.52 (d, J = 7 Hz, C CH3). (R)-1 2-Dimethyl-3-[( S)-sec-Phenethyl] Citrate (_g) Esterification of (.3.) with ( S)-(-)-sec-phenethyl alcohol by the method of (~) gave@). The ratio of CH30 peaks 8 3 77 and 3.69 to 83.73 and 3 68 in the 600

PAGE 47

33 1 MHz H NMR (CDCl3 ) showed an ee of more than 99% based on the intergrations. N1, N4-Dibenzyl Rhizoferrin, Tetramethyl Ester (.Q) Diphenylphosphoryl azide (760 mg, 2.76 mmol) and NEt3 (1.5 ml, 11 mmol) were added to a solution of(~) (610 mg, 2.77 mmol) and N1,N4 -dibenzyl-1,4-diaminobutane (370 mg, 1 .38 mmol) in DMF (20 ml) at O 0c under nitrogen The solution was stirred at O C for 1 h and then at room temperature for 23 h. After solvents were removed under high vaccum, the residue was taken up in EtOAc (25 ml) and was washed with saturated NaHC03 (25 ml), water (25 ml), 0.5 N HCI (25 ml), and water (25 ml). The organic layer was dried (MgS04 ) and concentrated. Flash chromatography, eluting with 4:1 EtOAc/hexane, generated 240 mg (26%) of (Q) as a pale yellow oil : [a] +8.25 (c 1.00); 1H NMR (CDCl3 ) 8 7.42-7 .24 (m, 10 H), 4 65-4.48 (m, 4 H), 3.81 (s, 3 H, OCH3), 3.79 (s, 3 H, OCH3 ) 3.69 (s, 3 H OCH3), 3.65 (s, 3 H, OCH3), 3.40-3.12 (m, 4 H), 3.10-2.67 (m, 8 H), 1.57-1.41 (m, 4 H). Anal. Calcd for C34H44N2012 : C, 60.70; H, 6 .59; N, 4.16. Found: C 60 64; H 6 .61; N, 4.15. N. N'-Dibenzyl Rhizoferrin (1) A solut ion of() (170 mg, 0.253 mmol) in CH30H (7 ml) and 1 N NaOH (7 ml) was stirred at room temperature for 5 h. HCI (1 N, 8 ml) was added, and the solut i on was concentrated to about 15 ml. After extraction with EtOAc (3 x 15 ml), the organic layer was dried (Na2SO 4 ) and concentrated to give 120 mg

PAGE 48

34 (77%) of (1) as a colorless glass : [a] +12 27 (c 1.00); 1H NMR (CD30D) 8 7.427 20 (m, 10 H, 2 Ph), 4 67-4.47 (m, 4 H, CH2Ph), 3.35-3.23 (m, 4 H, 2 NCH2), 3.19-2.69 (m, 8 H 4 CH2CO), 1 58-1.41 (m, 4 H, 2 CH2). Anal. Calcd. for C30H36N2012H20 : C, 56.78; H, 6 04; N, 4.41 Found: C, 56 88; H 6.08; N, 4.34. Rhizoferrin A solution of (Z) (110 mg, 0.178 mmol) in distilled THF (1.5 ml) was added to Li (33 mg, 4 8 mmol) in NH3 (100 ml), and the mixture was maintained at -78 C for 3 h. Aqueous CH30H (50% 10 ml) was added until the blue color disappeared. Ammonia was evaporated and the residue was taken up in water (50 ml) and concentrated to dryness. The colorless residue was dissolved in water and filtered through a cation exchange resin column (Bio Rad, AG 50W-X8). The eluant-containing product (pH = 3) was extracted with EtOAc (50 ml), which was concentrated to dryness. The residue was dissolved in distilled EtOH (2 ml), filtered, and concentrated to yield 50 mg (64%) of rhizoferrin as a colorless glass: HRMS (FAB, m-nitrobenzyl alcohol matrix) calcd for C16H25N2012 437.1407 (M + H), found 437.1407 (base). Anal. Calcd. for C16H25N2012 H20 : C 42.29; H, 5.77; N, 6 17. Found : C, 42.49 ; H, 5.80 ; N, 5.84 A solution of crude product (10 mg) was purified by reversed-phase HPlC (C-18 preparative column, 21.4 mm x 25 cm, obtained from Rainin). The initial mobile phase concentration of 3% CH3CN in 0.1 % TFA was held for 15 min,

PAGE 49

35 followed by gradient elution of 3-11% CH3CN in 0.1% TFA over 35 min, then held at 11 % CH3CN in 0.1 % TFA for 20 min. Flow rate was maintained at 4 mUmin. Retention time was 56 min. lyophilization gave 4.32 mg (9.90 mmol) of purified rhizoferrin as a colorless glass: [a]-16. 7 (26 C) (c 0.1613); 1H NMR (D20) 8 3 21-3 .15 (m, 4 H), 3.02 (d, 2 H, J = 16.0 Hz), 2.79 (d, 2 H, J = 16 0 Hz), 2 76 (d, 2 H, J = 14.6 Hz), 2.65 (d, 2 H, J = 14.6 Hz), 1.53-1.47 (m, 4 H). A stock solution was prepared by dissolving the purified product in 50.00 ml distilled water; a 10.00 ml aliquot was diluted to 20.00 ml and adjusted to pH = 3.02 with 1.90 ml of 0.010 N HCI (final rhizoferrin concentration = 9.04 x 1 o-5 M). CD and UV spectra were taken immediately after pH adjustment. All spectra were baseline corrected with distilled water blank, which was acidified as above. CD Results The CD spectrum of rhizoferrin exhibited a negative Cotton effect from 200 to 220 nm, with a single minimum at 205 nm, .0.E= -2.7 compared to a recorded single minimum at 204 nm, .0.E = -4.3 (88). UV Results nm E E (88) 196 12200 (13900) 200 10800 (13150) 210 5230 (5600) 215 2770 (3000) 220 1200 (1400)

PAGE 50

36 Discussion The main problem in this synthesis is the separation of 1,2-dimethyl citrate enantiomers. There are several ways to obtain resolution of enantiomeric carboxylic acids (97). Among these are biological or enzymatic resolution of the racemic acid or its ester (97). The most primitive form of biological resolution is to grow microorganisms in the presence of the racemic acid or ester. Depending on the choice of microorganism, one or the other of the enatiomers will be utilized This is because the chiral compound that reacts at different rates with the two enantiomers may be present in a living organism For instance, a certain esterase may cleave an ester of one enatiomer but not the other Because it is necessary to find the proper organism and since one of the enantiomers is destroyed in the process, this method is limited. When the proper organism is found however, a good resolution can be achieved since biological processes are highly stereoselective. A more sophisticated approach utilizes highly purified enzymes from such microorganisms. Lipases are especially useful enzymes in organic synthesis because of their stability in organic solvents, the fact that they can accept a wide range of substrates, and their availablity (98) For example, about 20 different lipases from microbial, plant, and animal sources are commercially available. Lipases are characterized by the involvement of a lipid-water interface in their catalytic process. This unique feature of interfacial catalysis provides the lipases with an inherent affinity for hydrophobic environments and distinguishes them from other hydrolytic enzymes (98).

PAGE 51

37 Lipase-catalyzed reactions include esterification, transesterification, amidation, peptide synthesis, and macrocyclic lactone formation (98). Of these, enantioselective ester synthesis/interchange is of particular interest because it provides a facile method for the preparation of optically active acids. Interestingly, lipases have been utilized as catalysts for the hydrolysis of triesters of citric acid. The triethyl or trimethyl ester of citric acid was hydolysed with porcine liver esterase (PLE) or the proteinase subtilisin forming a mixture of the two diesters (99). It was evident that there was a slight preference for hydrolysis of the ester in the mid-position. The hydrolysis of the trimethyl ester of citric acid was not regioselective. However, hydrolysis of triethyl ester of citric acid yields only one product, the symmetric diester (Figure 2-4). CH2C02R CH2C02H CH2C02R I PLE or I I C(OH)C02R C(OH)C02R + C(OH)C02H I Subtilisin I I CH2C02R CH2C02R CH2C02R Figure 2-4 Hydrolysis of Triesters of Citric Acid with PLE or Subtilisin.

PAGE 52

38 In this project a number of lipases (lipase type VII from Candida rugosa and type XI from Rhizopus arrhizus) were tested for their abilities to stereoselectively cleave trimethyl, triethyl, triheptyl and tribenzyl esters of citric acid. To each 2 ml vial was added 900 L phosphorate buffer solution, 100 M trimethyl citrate ester, 100 units lipase, and 100 L various organic solvents. The solvents used in the tests included acetone, acetonitrile, butanol, butylacetate, tert-butylmethyl ether, chloroform, cyclohexane DMF, dioxane, ethanol, ether, ethyl acetate, n-heptane, n-hexane, methanol, methylene chloride, methyl acetate, octanol, n-propyl ether, THF, and toluene. The mixtures were stirred overnight and the reactions were checked by TLC. New compounds were only found in the reactions that contained dioxane and they were not 1,2-dimethyl citric acid as indicated by TLC. Further attempts to obtain 1,2-dimethyl citric acid enatiomers using enzymatic means was abandoned. Another means to resolve a racemic carboxylic acid is by fractional crystallization of its diastereomeric salt with a chiral base. The most frequently used chiral bases include brucine, ephedrine, strychnine, and morphine. The base, brucine, was used in this project. If, for the sake of argument, the base is in the 'S' form the two salts produced will have the configurations SS and RS. The acids are enantiomers, however, the salts are diastereomers and have different properties (Figure 2-5). The property of the diastereomer salts most often used for separation is differential solubility. The mixture of 1,2-dimethyl citric acid-brucine salts was allowed to crystallize from water. Because the solubilities are different, the

PAGE 53

39 initial crystals formed will be richer in one diastereomer In this instance, the (R S) 1,2-dimethyl citrate-brucine salt dominated in the crystal form Filtration at this point resulted a partial resolution. The total separation of the diastereomers was achieved by several such fractional crystallization steps s + CH2C02CH3 H3C02C+OH CH2C02H R s + S-Brucine + R Figure 2-5 The Diastereomeric Salts Formed Between 1,2-Dimethyl Citrate and Brucine. s s

PAGE 54

CHAPTER Ill SYNTHESIS OF STAPHYLOFERRIN A Introduction Citric acid is a component in many siderophores produced by microorganisms. The citrate-based siderophores can be divided into two groups according to the stereochemistry of citrate in these molecules The first group consists of aerobactin, arthrobactin, schizokinen (81), and nannochelin (82), in which citric acid acylates same functional groups via its 1 or 3 carboxylic acids. Thus, citrate is still prochiral in these siderophores. Examples in the second group include rhizoferrin presented in Chaper II, and staphyloferrin A described in this chapter. Both of these siderophores contain two citric acid residues. Each terminal carboxylic acid of these two citric acids forms a bond with an amino group in putrescine or 0-ornithine to give rhizoferrin and staphyloferrin A, respectively. Consequently, the citric acids are asymmetric in these two siderophores. The absolute configurations of the two asymmetric citrates in rhizoferrin were confirmed to be (R,R) in Chapter II. N2 N5 -Bis( 1-oxo-3-hydroxy-3,4-dicarboxybutyl)-D-ornithine ( staphyloferrin A) was isolated from cultures of Staphylococcus hyicus DSM 20459. Staphyloferrin A has one ornithine backbone and two citric acid residues linked by two amide bonds (83). The natural product was obtained as a colorless, acidic compound, which is soluble in water and water/methanol mixtures 40

PAGE 55

41 containing up to 80% methanol and virtually insoluble in all common organic solvents. It is stable in neutral aqueous solution over the range 4-80C, and stable up to pH 10 at room temperature (100). However, under acidic conditions, staphyloferrin A was degraded. The stereochemistry of D-ornithine was determined by amino acid analysis of the staphyloferrin A acid hydrolysate and by gas chromatography of the N-trifluoroacetylated, n-propyl ester amino acid derivatives on a Chirasil-Val chiral HPLC support (100). However, the configurations of the two chiral citric acid residues were undetermined. The only evidence bearing on the stereochemistry of the two chiral citric acids is the CD spectrum of ferric staphyloferrin A with one positive cotton effect at 350 nm (L1E = 0.24 M-1cm-1 ) and two negative effects at 302 nm (L1E = -0.46 M-1cm-1 ) and 250 (L1E = -0.28 M-1cm-1 ) nm (100). The goals in this synthesis are to complete the synthesis of staphyloferrin A and to confirm the absolute configurations of the two chiral citric acid residues in this natural product. As previously mentioned, (R)-1,2-dimethyl citrate developed in the synthesis of rhizoferrin can be used in this synthesis. The stereochemistry of the two chiral citrates was assumed to be (R, R) and could then be verified by the CD spectrum of synthetic ferric staphyloferrin A, as compared to that of the natural ferric siderophore. Synthesis While staphyloferrin A has an amino acid, D-ornithine, instead of putrescine as in rhizoferrin, it is almost impossible to use the synthetic strategy

PAGE 56

42 of rhizoferrin for the synthesis of staphyloferrin A, because synthesis of dibenzylated 0-ornithine is impractical. Furthermore, the reactivities of the two amines in 0-ornithine are different. The a-amine can be acylated with a chiral 1,2-dimethyl citrate in the presence of DCC/DMAP, while attempts to acylate the 8-amine with the same reagents failed. Therefore, a different synthetic scheme was developed for this synthesis. The total synthesis of staphyloferrin A started with the Cbz protected Dornithine. The carboxylic acid of di-Cbz-0-ornithine was protected as its tert butyl ester by using HCI04 in tert-butyl acetate in low yield (23%) Next, the CBZ protecting groups in ester (ID were removed by 10% Pd/C in hydrogen, affording compound rn). (R)-1,2-D i methyl citrate Q) generated by the method described in Chapter II, was reacted with pentachlorophenol in the presence of DCC/DMAP to form an active ester (1.Q), which was reacted with compound cm to provide (11) in 66% yield. Finally, the tert-butyl ester of (11) was removed by TFA in methylene chloride giving (12), followed by hydrolysis of the methyl esters in NaOH/MeOH to furnish staphyloferrin A m) as its sodium salt (Figure 3-1). The final product was only partially epimerized as determined by comparing the 1 H NMR spectrum of the 0-ornithine methine peaks of compound (12) to that of compound m). There are two types of peaks matching the methine hydrogen in the 1 H NMR spectrum of compound m) instead of one type in that of compound (.12.). The ratio between these two

PAGE 57

43 types of peaks appeared to be sensitive to the basicity of the solvent and the reaction time. Although different bases, for example, NaOH, KOH, and LiOH, COOH Cbz-HN ~NH-Cbz HCOOC~OH 3 ,._,../' ..... II H3COOC' OH o (}) a C Cl Cl H3COOC~o---U--c1 H3COOC' OH o (1.Q) d, e, f --'--'--:~ H R2-00(2 ?i Ho,i:;OOR3 R300C::Y'l(N~N~COOR3 R30oc OH O H (11), R2 = tBu; R3 = CH3 (.12), R2 = H; R3 = CH3 (ll), R2 = H; R3 = H Figure 3-1. Synthesis of Staphyloferrin A. a) t-Butyl acetate/HCI04 ; b) Pd/C/HCI/H 2 ; c) DCC/DMAP/Pentachlorophenol;d) Et 3 N/(~); e) TFA; f) NaOH (ee 80%). were tested, partial racemization still resulted. Nonetheless, the racemization did not affect the CD spectrum of the ferric complex or adduct with the final product (368 nm, .0.E = 0 .31 M-1cm1 ; 316 nm, .0.E = -0.60 M-1cm 1; 266nm .0.E = -

PAGE 58

44 0.16 M-1cm-1 ) nm, which is similar to that of the ferric complex or adduct natural product. Therefore, the absolute configurations of the two chiral citrate in staphyloferr i n A are suggested to be (R, R) as in rhizoferrin. Experimental General Synthesis (see synthesis of rhizoferrin) Circular Dichroism The CD spectra were obtained with a Jasco Model J710 spectropolarimeter equipped with a Jasco IF-710 interface Cell path length was 0.01 cm, and the concentration of ferric staphyloferrin A was 21 mM in water (pH = 5.6) Synthesis of N,N -Bis(benzyloxycarbonyl)-o-ornithine-tert-butyl Ester (8) A mixture of perchloric acid (1 ml of a 6% solution), N Nbis(benzyloxycarbonyl)-o-ornithine (2.4 g, 6.0 mmol) (Sachem) in 10 ml CHzClz, and 10 ml of tert-butyl acetate was stirred for two days The solvent was removed and the residue was dissolved in ethyl acetate (20 ml) and washed with saturated aqueous NaHC03 (20 ml) and brine (20 ml). The organic layer was dried over MgS04 filtered, and evaporated, affording CfD (0.63 g, 23%) as a colorless oil: [a]= -8.1 (c = 0.67 in CHCl3); 1H-NMR (CD300) o 1.43 (s, 9H, t-Bu orn), 1.5-1.82 (m, 4 H, 2 x CHz-orn), 3.10-3 14 (m, 2H, CHz-N), 4 01-4 .11 (m, 1 H, CH-orn), 5.05 (s, 2H, GHz-Ph), 5.08 (s, 2H, GHz-Ph), 7 31-7.42 (m, 1 OH, Ph); Anal. Calcd. for Cz5H3zNz06 : C, 65.77; H, 7 .07; N, 6.14 ; found C 65 .90; H 7.03; N 6.23. Synthesis of o-Ornithine-tert-butyl Ester Dihydrochlor i de (9) To a 100 ml flask containing (.a) (0. 62 g, 1.36 mmol) in 20 ml of ethanol was added 10% Pd/C (80 mg) and 1 N HCI (2.7 ml) cautiously. The flask was degassed with Nz by three times then filled with Hz. The mixture was stirred for 1 h, filtered through celite, and concentrated to give ffi) (0.34 g, 95%) as a white

PAGE 59

45 solid: mp 163-165C; [ah4 = -9. 84 (c = 0.028 in MeOH); 1 H NMR, (CD30D) o 1 55 (s, 9H, t-butyl), 1.70-2.01 (m, 4H, 2 x CH2-orn), 3 01-3.10 (m, 2H, N-CH2), 3.90-3.96 (m, 1 H CH-orn); Anal. Calcd. for C9H22 Cl2N202 : C, 41.39; H 8.49; N 10.73 found: C 41.59; H, 8.31; N, 10.53. Synthesis of ( S)-1 ,2-Dimethyl-3-pentachlorophenyl citrate ( 10) A mixture of 1,3-dicyclohexylcarbodiimide (0.82 g, 3.99 mmol), (~ (0.88 g, 3.99 mmol), pentachlorophenol (1. 06 g, 3.99 mmol), and 4-dimethylamino pyridine (5 mg) in CH2Cl2 (60 ml) was stirred overnight at O C. The mixture was filtered. Column chromatography of the residue with EtOAc/hexane (7 :3) gave (1.Q) (1.4 g, 75%) as a white solid (mp 97-99C): [a] = -9. 2 (c = 0.5 in CHCl3 ); 1 H-NMR (CDCl3 ) o 2.90 (d, J = 15.5, CH2), 2.98 (d, J = 15 .6, CH2), 3.23 (d, J = 16.4, CHz), 3.3 (d, J = 16.4, CHz), 3.73 (s, CH30), 3.85 (s, CH30); Anal. Calcd for C14H11Cl50i C 35.89; H, 2.37; found: C, 36.13; H 2 .38. Synthesis of N N-1,2-(R)-Dimethyl Citric Acid-0-ornithine-t-butyl Ester (11) A mixture of rn) (0. 33 g, 1.27 mmol) in DMF (20 ml), Et3N (0 25 g, 2 5 mmol), and (1.Q) (1.4 g, 3.0 mmol) was stirred overnight at room temperature. Solvents were removed in vacuo Flash chromatography of the residue with EtOAc gave (1j) (0.49 g, 65%) as a colorless oil (ee, 99%) : [a]= 3.2 (c = 0.025 MeOH ); 1 H-NMR (CDCl3 ) o 1 .56 (s, 9H, t-Bu), 1 52-1.83 (m, 4H, 2 x CH2-orn), 2.60-3.00 (m, 8H, 4 x CHrcitri), 3 15 3 18 (m, 2H, N-CH2-orn), 3.65 (s, 3H, OCH3), 3 66 (s, 3H, OCH3), 3 76 (s, 6H, 2 x OCH3), 4.21-4.23 (m, 1 H, CH-orn) ; Anal. Calcd. for C25H40N2014: C 50. 67; H, 6.80; N 4 .73; found: C, 50 60; H, 6.75; N, 4.70.

PAGE 60

46 Synthesis of N, N-1,2-(R)-Dimethyl Citric Acid-D-ornithine (12) A mixture of trifluoroacetic acid (8 ml) and W) (0.47 g, 0 79 mmol) was stirred at room temperature for 30 min. Evaporation of excess TFA in vacuo afforded (12) (0.42 g, 99%) as an oil: [a]= -9.1 (c = 0.14, MeOH); 1H-NMR (CD30D) 8 1.52-1.90 (m, 4H, 2 x CHz-orn), 2.61-3.03 (m, 8H, 4 x GHz-citric), 3 16-3.19 (m, 2H, N-CHz-orn), 3.65 (s, 6H, 2 x OCH3), 3.76 (s, 6H, 2 x OCH3), 4 35-4 .37 (m, 1 H CH-orn) ; Anal. Calcd. For C1zH3zNz014 : C, 47 02; H, 6 .01; N, 5.22; found: C, 46 76; H 5 .89; N, 5.13. N2, N5-(R, R)-Bis( 1-oxo-3-hyd roxy-3,4-d icarboxybutyl)-o-orn ith i ne (Staphyloferrin A) (13) To a solution of (.12) (0.42 g, 0 78 mmol) in 10 ml of MeOH/H20 (1:1) was added sodium hydroxide (10 ml, 1 N). The mixture was stirred at room temperature for 3 h, and then passed through a cation exchange resin column (18/3 Bio Rad, AG 50W-X8, molecular biology grade, 63-150 m wet bead size 200-400 dry mesh size, ammonium form) The resulting solution was lyophilized, providing (1]) (0.43 g, 98%, ee 80%) as its penta ammonium salt : [a]= -11.47 (c = 7 .5, HzO); 1H NMR (DzO) o 1.48-1.86 (m, 4H, 2 x CH2-orn), 2.46-2.80 (m, 8H, 4 x GHz-citric), 3.18-3.20 (m, 2H, N-CHz-orn), 4.11-4.14 (m, 1 H CH-orn); Chromatography of its ammonium salt (0. 20 g, 0.35 mmol) (18/3, Bio Rad, AG 50W-X8, molecular biology grade, 63-150 m wet bead size, 200-400 dry mesh size acid form) gave the free acid form of staphyloferrin A (0.11 g, 99%). Anal. Calcd. for C17Hz4Nz014 2 HzO: C 39.54; H, 5.47; N, 5.42 ; found: C, 39.71; H, 5.40; N 5.32. Staphyloferrin A Ferric Complex To a solution of staphyloferrin A ammonium salt (14 mg, 0.025 mmol) in 10 ml distilled deionized water was added a solution of ferric acetylacetonate

PAGE 61

47 (12 mg, 0.03 mmol) in 10 ml ethyl acetate. The mixture was stirred for 1 h. The colorless water layer turned pale yellow during that time. The two layers were separated, and the water layer was washed with EtOAc (5 x 10 ml) until the organic layer was colorless. Lyophilization of the aqueous layer gave the staphyloferrin A ferric complex (12 mg, 90%) as a yellow solid. Discussion While it is common that many microorganisms produce siderophores, (iron specific chelating ligands), when they are grown under iron starved conditions, the production of identical enantiomeric siderophores in fungi and bacteria is a special case. For example R, R-rhizoferrin has been found to be produced by members of several families of the Zygomycetes, and ferric R, Rrhizoferrin can be utilized not only by the zygomycetous fungi, but also by a non producing bacterium Morganel/a morganii (101) Recently, a new citric acid-based siderophore, S,S-rhizoferrin has been isolated from Pseudomonas pickettii DSM 6297, a human pathogen responsible for occasional nosocomial infections (102). S, S-Rhizoferrin (enantio-rhizoferrin) has same formula as R, R-rhizoferrin, but has an inversion of the stereochemistry of the two chiral citrates (Figure 3-2). These two siderophores and staphyloferrin A are relatively simple compounds formed by combining three structural subunits. The citric acids became asymmetric because of the combinations in these three siderophores. The different stereochemistry of the bacterial and fungal rhizoferrins has raised the question whether enantiomeric recognition of these ferric siderophores by microbial transport systems exists. Transport experiments

PAGE 62

48 with radiolabelled iron using S, Sand R, R-rhizoferrin showed that the transport of R, R-rhizoferrin had twice as high iron uptake rates in the fungal Rhizopus strain than S, S-rhizoferrin (102). It suggested that transport of ferric R, R-rhizoferrin is stereoselective in this fungus. However iron transport of ferric S, S-and R, R-rhizoferrin were very similar to one another in Pseudomonas pickettii (102). The chiral separation developed in the synthesis of rhizoferrin and staphyloferrin A allows for the synthesis of all possible enantiomeric citric acid based chelators While (R)-1,2-dimethylcitric acid was obtained from the crystals, the enantiomer (S)-1,2-demethylcitric acid can be obtained from the liquid. The possible (R, S)-rhizoferrin, (R, S)-staphyloferrin A, or (S, R) staphyloferrin A can be synthesized by using both enantiomers. This provides a unique opportunity to study the role of chiral citric acid in bacterial iron utilization and transport of ferric chiral citric acid-based chelators.

PAGE 63

49 H OHO ~OOH HOOe~N~N~eOOH Hooe OH o H R, R-Rhizoferrin (Rhizopus microsporus var. rhizopodiformis) H O HO, eOOH HOOe~N~ Jcx.;eooH ___ r-~ .. II N Hooe OH o H S S-Rhizoferrin (Pseudomonas picketti1) H HOO~ ?i Ho,pooH HOOe~N~ ~eOOH ,....,._."\__ 11 N Hooe OHO H R, R-Staphyloferrin A ( Staphylococcus hyicus) Figure 3-2. Examples of Chiral Citric Acid-based Siderophores.

PAGE 64

CHAPTER IV EFFICIENT SYNTHESIS OF POLYAMINE CATECHOLAMIDES 1ntroduction Hydroxamates and catecholamides are two major classes of siderophores isolated from bacteria (43). Both can form six coordinate octahedral complexes with iron. While hydroxamates donate three hydroxamates as the chelating funtional groups, catecholamides display two different cases: the chelating groups in enterobactin are six phenolic hydroxyls, whereas in polyamine catecholamides, e.g., parabactin, the chelating functionalities are five phenols as well as the nitrogen on the oxazoline ring (103). The major functional difference between hydroxamate and catecholamide siderophores is related to level of iron concentration (104) Microorganisms produce hydroxamates when the iron concentration is relatively high, while the catecholamides are generated in a low iron concentration. Thus, the iron binding constant of a catecholamide is much higher than that of a hydroxamate. Except for enterobactin, many catecholamides have structural similarities, that is, they all contain a polyamine backbone as well as chiral oxazoline ring(s). They can have a symmetrical (e.g., norspermidine in vibriobactin and fluviabactin) or an asymmetrical (e.g spermidine in parabactin and agrobactin) polyamine backbone. They can also be separated into two classes on the 50

PAGE 65

51 basis of the substituents on the two terminal primary amines of the polyamine backbone. For example, in parabactin isolated from Paracoccus denitrificans (7 4) and agrobactin from Agrobacterium tumefaciens (105), the primary amines of spermidine in both compounds are acylated with 2,3dihydroxybenzoic acid. On the other hand, in vibriobactin isolated from Vibrio cholerae (75), one primary amine is connected with a chiral L-oxazoline ring, while the other primary amine is acylated with 2,3-dihydroxybenzoic acid Recently, a novel siderophore, fluviabactin, was isolated from Vibrio f/uvialis (79) The chemical structure of L-fluviabactin is similar to that of Lagrobactin, but contains a norspermidine backbone. In this project the stereochemistry of the oxazoline ring in fluviabactin was changed from Lthreonine to o-threonine to form o-fluviabactin, and the polyamine backbone was changed from norspermidine to homospermidine and spermidine to generate L-homofluviabactin and L-agrobactin, respectively (Figure 4-1). Polyamine catecholamides, for example, parabactin have emerged as potential iron chelators as an alternative to desferrioxamine B (DFO) in a bile duct-cannuated rat model as well as in a Cebus monkey model. When both ligands were administered as a sc bolus, parabactin was 5.8 times as efficient as DFO in the rodent model (P < 0.001) and was 1.8 times as effient as DFO in the primate model (P < 0.001). Both DFO and parabactin were effective at putting the primates in negative iron balance (64). Furthermore, the catecholamide chelators have been shown to be very potent cell synchronization agents, holding the cells at G1/S border (106). Both parabactin

PAGE 66

52 and vibriobactin have been shown to strongly inhibit the growth of L 1210 cells. Parabactin also inhibits HSV-1 virus (107,108). ()'OH __ l~.-oH r) H ;~O H HO~N~N~NyY'oH OHO O OH D-Fluviabactin ()'OH __ l~.-oH 0 N OH O )--1. ... f O O OH HOijN~N~N~OH I H H I b L-Homofluviabactin Figure 4-1. Structures of Two L-Fluviabactin Analogues: o-Fluviabactin and L-Homofluviabactin.

PAGE 67

53 In each of the catecholamide chelators to be synthesized, the catechol functions of L-fluviabactin, D-fluviabactin, L-agrobactin, and L-homofluviabactin are all in the form of 2,3-dihydroxybenzoyl groups In fact, the 2,3dihydroxybenzoyl forms a very tight, three to one high spin, hexacoordinate, octahedral complex with Fe(III). When the 2,3-dihydroxybenzoyl group is fixed to polyamine backbones, as in the cases of fluviabactin, agrobactin, or vibriobactin, or to a triserine macrocycle, as in enterobactin, the formation constants of the Fe(III) complexes become even higher. For example, the enterobactin-Fe(III) complex has a formation constant of 1052 M-1 and parabactin forms a tight Fe(III) complex with a formation constant of 1048 M-1 Synthetically, then, in the cases of fluviabactin and agrobactin the objective becomes to fix the 2 3-dihydroxybenzoyl functionality to the appropriate anchors e.g., polyamine or threonine. The similarities of polyamine catecholamides led to the development of several synthetic schemes in this and other laboratories (109,110 111). The main synthetic obstacle was the selective acylation of a triamine. Because the synthesis of parabactin and agrobactin involve the acylations the primary amines of a spermidine with a 2,3-dihydroxybenzoyl group they present the same kind of selectivity problem : the acylation of the primary vs. secondary amines of a spermidine. In the synthesis of parabactin and agrobactin, the two terminal primary amines of a spermidine could be acylated with 2,3dihydroxybenzoyl groups followed by the introduction of an oxazoline ring. Alternatively, an oxazol i ne ring first can be introduced on to the secondary

PAGE 68

54 amine of a spermidine followed by the acylations of the two primary amines by 2,3-dihydroxybenzoyl groups. Vibriobactin synthesis consists of the fixing of a 2,3-dihydroxybenzoyl functionality to only one primary amine of a norspermidine. The synthetic problem is associated with the selective acylation of a primary amine in the presence of another primary amine and a secondary amine and thus, requires different synthetic schemes The synthetic alternatives could include the mono acylation of one primary amine of a norspermidine with a 2,3-dihydroxybenzoyl group followed by the bis-acylation of the other two amines of the norspermidine with two oxazoline fragments, or the opposite order of attachment. Therefore, appropriately protected triamines become the common denominators in the synthesis of all the polyamine catecholamide iron chelators In our laboratory, a protected triamine, an interrally N-benzylated triamine, was developed (112). The reagent synthesis begins with a suitable N-benzyl diamine, which can be obtained either from cyanoethylation of benzylamine followed by reduction of nitrile to diamine with a Raney nickel catalyst, or from the condensation of putrescine with benzaldehyde in formic acid followed by reduction of the imine. The primary amine of the diamine can be regioselectively protected with one equivalent of 2-(tert butoxycarbonyloxyimino)-2-phenylacetonitrile (BOC-ON) The remaining amine is alkylated with either acrylonitrile or 4-chlorobutyronitrile to furnish the homologous nitriles that can be used directly in the next reaction. The nitriles

PAGE 69

55 were reduced to the corresponding amine by Raney nickel. The BOC protecting group of the amines can be readily removed by brief exposure to trifluoroacetic acid to provide the corresponding secondary N4-benzylated triamines (Figure 4-2). v~~NH2 a n = 1 or 2 b C m = 1, or 2 d Figure 4-2. Synthesis of Benzyl-protected Triamine: a) BOC-ON; b) acrylonitrile or 4-chlorobutyronitrile; c) Raney Ni; d) TFA. The protected triamines are versatile reagents for the synthesis of all polyamine catecholamides, and have been successfully used in the synthesis of parabactin (113) and vibriobactin (110). The synthesis of parabactin was

PAGE 70

56 initiated with the polyamine reagent N4-benzylspermidine, which was first reacted with 2,3-dimethoxybenzoyl chloride, providing a bisamide. Next, the N4 -benzyl group is removed by hydrogenolysis over palladium at atmospheric pressure The free secondary amine of the diamide can then be acylated with L-N-(tert-butoxycarbonyl)threonine, activated as the N-hydroxysuccinimide ester, affording the corresponding triamide. The BOC group was removed with TFA, and the methyl protecting groups were removed by BBr3 in CH2Cl2 (81). The resulting amino alcohol was condensed stereospecifically with ethyl 2hydrobenzimidate to form the acid-sensitive oxazoline ring of parabactin (Figure 4-3) Because vibriobactin has no symmetry with respect to the terminal acyl groups, it thus presents a new synthetic challenge The synthesis of vibrobactin began with a primary, secondary amino-diprotected norspermidine, N4-benzyl-N1-(tert-butoxycarbonyl)norspermidine. The free primary amine was acylated with 2 3-dimethoxybenzoyl chloride in the presence of triethylamine generating the trisubstituted noespermidine Both the tert-butoxycarbonyl protecting group and the N4-benzyl group can be removed using TFA and Pd/C/HCI, respectively The order of deprotection is not important. The free primary amine and secondary amine were then bisacylated with the activated ester of L-N-(tert-butoxycarbonyl)threonine. Next, the BOC protecting group was removed with TFA, and the catechol methyoxyl protecting groups were removed with BBr3 in methylene chloride. Finally, the condensation between the threonyl

PAGE 71

57 groups and excess of 2,3-dihydroxybenzimidate in refluxing methanol provided vibriobactin (Figure 4-4). p + a, b ,... H2N~N~NH2 r'n H H O OCH3 H3CO~N~N~~~OCH3 c d ,.._ OCH:P u HO NH2.TFA r'n H ryo O OCH3 H3CO~N~N~NJVvocH3 e f OCH:P H u n ~I ... ~OH 0 N r'n )-J...fo O OH HoyY~~N~N~OH OHO H u Figure 4-3. Synthesis of L-Parabactin Using N4-Benzylspermidine: a) Et3N; b) Pd/C/H2 ; c) L-N-(tert-butoxycarbonyl)threonine/DCC/ N-hydroxysuccinimide ; d) TFA; e) 8Br3 ; f) Ethyl 2-hydrobenzimidate.

PAGE 72

58 + OH x:OH I OH A OH 0 N N 0 H )-{fo oy-~ HO~N~N~NH OH 0 a b,c,d e,f Figure 4-4. Synthesis of Vibriobactin Using Diprotected Triamine: a) Et3N; b) Pd/C/H2; c) TFA; d) L-N-(tert-butoxycarbonyl)threonine/ DCC/ N-hydroxysuccinimide; e) 8Br3 ; f) ethyl 2-hydrobenzimidate.

PAGE 73

59 As can be seen from the above synthesis, the protected triamines are useful reagents for the synthesis of polyamine catecholamide siderophores However in the synthesis of polyamine catecholamides and their analogues, some commercially available coupling reagents, for example, 1, 1carbonyldiimidazole (CDI) can be taken advantage of. The selective acylation of the primary amines of an unprotected triamine can be achieved by the reaction a carboxylic acid which has been activated with COi. The phenol hydroxyl protecting groups can be changed from methyl to benzyl (the latter is much more easily removed) and the intermediate can directly utilized in the following reactions without further purification Therefore a more concise scheme for the synthesis of polyamine catecholamides is presented. Synthesis The synthesis of polyamine catecholamide iron chelators in our lab depends heavily on the availability of the appropriately protected polyamines norspermidine, spermidine, and homospermidine In the synthesis of vibriobactin, the selective acylation of one primary and one secondary amine of a norspermidine is almost impossible without a protected polyamine reagent. However, in the synthesis of parabactin, agrobactin, and fluviabactin, the selective acylation of primary amines vs secondary amine of a triamine can be achieved by utilizing N-hydroxysuccinimide or imidazole activated acids. Although the yields in the formation of diamides are moderate (60-70%), it can save many steps i n the synthesis of a protected polyamine reagent. Thus, it is possible to increase the yield of a total synthesis.

PAGE 74

60 In the present synthesis, 1, 1-carbonyldiimidazole (COi) (114,115) was utilized for the selective acylations of two terminal primary amines of triamines. These reactions involved the attachment of two 2,3-dihydroxybenzoyl groups to the primary amines of norspermidine (L-or D-fluviabactin), spermidine (Lagrobactin), or homospermidine (L-homofluviabactin) The COi coupling reagent was first reacted with 2,3-dibenzoxylbenzoic acid for about one hour, after which the corresponding free triamines were added. COi was an especially convenient, selective amide-forming reagent in that the imidazole by product was washed out during workup In this project, the reactions of spermidine, norspermidine, or homospermidine with 2,3-bis(benzoxy)benzoic acid in the presence of COi produced the corresponding bisamides (14, .1.,Q, .1.Q) as illustrated in Figure 4-5 (64-73%). 2,3-Bis(benzyloxyl)benzoic acid was utilized instead of 2,3-dimethoxylbenzoic acid because the benzyl protecting groups are more easily removed than the methyl groups. The synthesis of 2,3bis(benzyloxyl)benzoic acid (111) began with 2,3-dihydroxybenzaldehyde, in which the catechol hydroxyls were protected by benzyl groups using benzyl chloride and potassium carbonate, followed by oxidation with sodium chlorite and sulfamic acid, to provide 2,3-bis(benzyloxy)benzoic acid (90%). However, unlike commercially available spermidine and norspermidine, homospermidine (81) has to be synthesized by alkylation of mesitylenesulfonylamide (116) with N-(4-bromobutyl)phthalimide and NaH to form fully protected homospermidine (67%). Free homospermidine was obtained by removal of the phthalimide and mesitylenesulfonyl protective

PAGE 75

61 COi --->. norspermidine m = n = 1 spermidine m = 1, n = 2 homospermidine m = n = 2 0Bn O O 0Bn BnO~ ~0Bn () (J (14) m = n = 1 (.15.) m=1,n=2 (.1.) m = n = 2 Figure 4-5. COi Coupling Reactions.

PAGE 76

62 groups with NH2NH2 (82%) (117) and with 30% HBr/HOAc and phenol (98%) (81), respectively. Terminally-diacylated triamines (H., 15, 16) were then acylated with N carbobenzoxy-L-threonyl-0-hydroxysuccimide (N-Cbz-L-Thr-OSu) for Lfluviabactin, L-agrobactin, L-homofluviabactin, or its D enantiomer N-Cbz-D-Thr OSu for o-fluviabactin (see compounds 17, ~. ~ 20.) (50-76%). The hydroxysuccimide active ester of a Cbz-protected threonyl can be readily prepared using the protected threonines and N-hydroxysuccimide in the presence of DCC in methylene chloride, and do not need to be separated from the solvent; even the insoluble DCU can stay in the reactions. The carbobenzoxyl and benzyl groups of the resulting trisubstituted amide were removed by hydrogenolysis over palladium in methanolic HCI at atmospheric pressure, providing amine salts (21, 22, 23, 24) (96-98%). The stereospecific formation of the acid-sensitive trans-oxazoline ring (118) can be achieved by the condensation of a threonine residue and ethyl 2,3-dihydroxybenzimidate However, ethyl 2,3-dihydroxybenzimidate cannot be obtained simply by Pinner's reaction, as in the preparation of ethyl 2 hydroxybenzimidate (119) Another approach to the imidate ester started with 2,3-dibenzyloxybenzoic acid, which was also used in the CDI coupl ing reactions. The acid was converted to 2,3-di(benzyloxy)benzoyl chloride, followed by aminolysis of 2,3-bis(benzyloxy)benzoyl chloride, furnished 2 3di(benzyloxy)benzamide (98%), which was selectively 0-alkylated with triethyloxonium hexafluorophosphate in CH2Cl2 followed by basification (120)

PAGE 77

63 to provide ethyl 2,3-bis(benzyloxy)benzimidate in 80% yield. Finally, the cleavage of benzyl-protecting groups by hydrogenolysis under mild conditions (10% Pd/C, 1 atm) led to ethyl 2,3-dihydroxybenzimidate (73%) (111) (Figure 4-6). ,.._ ~::: b a ,.._ C ___ ,.._ d ,.._ e, f ,.._ Figure 4-6. Synthesis of Ethyl 2,3-Dihydroxybenzimidate: a) benzyl chloride/K2C03; b) NaCI02/sulfamic acid; c) SOCl 2 ; d) NH40H; e) triethyloxonium hexafluorophosphate; f) Pd/C/H 2

PAGE 78

64 Therefore, the threnonyl residues (21, 22, 23, 24) were condensed with ethyl 2,3-dihydroxybenzimidate in methanol to furnish the final products L-and o-fluviabactin, L-agrobactin, and L-homofluviabactin (60-70%) (Figure 4-7). The mechanism for the oxazoline-forming reaction was suggested to be the following. The hydroxyl of a threonine residue first attacks the imine of ethyl 2,3dihydrobenzylimidate, in which the ethoxy serves as a leaving group. After the protons transfers intramolecularly from the protonated amine of the threonyl residue to the imine, the amine of the threonyl residue acts as another nucleophile to react with the imine, forming a five-membered ring. Finally, the elimination of ammonia between the nitrogen of the threonine and the carbon of the imidate provides an oxazoline ring. In summary, the scheme developed here is shortened to four steps. The starting material, 2,3-bis(benzyloxy)benoic acid, can be prepared on a large scale by recrystallation. Other starting compounds, such as COi, norspermidine, and spermidine are commercially available. The intermediates of hydogenolysis can be directly used in subsequent reactions Thus, it is possible to make polyamine catecholamides on a large enough scale to provide materials for animal tests. The iron clearing properties of L-and o fluviabactin were evaluated in a bile duct-cannuated rat model. Both compounds were shown to be very effective at clearing iron from the rodent. When given sc at a dose of 150 mol/kg, both ligands had iron clearing efficiences of > 13%.

PAGE 79

65 OBn O O OBn BnO~ JVyosn (JJ (JJ (.14) m = n = 1 (.1fil m=1,n=2 (.1.fil m = n = 2 HO NH-Cbz OBn O -'>--1... 0 O OBn (.11, 1.a) m = n = 1 BnO~ ~,-. "'F:. ~OBn -7 N~W''t,-Yn'N -7 (~) m = 1, n = 2 I H H I (20) m = n = 2 -I b y HO NH.HCI OH O )-J... ,... 0 OH HOV ,-. r::, ~OH '71 I IC y (;r'OH _X .. ~OH 0 N OH O )-J... o O OH HO~ ~,-. "F:_ JVyoH (JJ () L-or D-Fluviabactin m = n = 1 L-Agrobactin m = 1, n = 2 L-Homofluviabactin m = n = 2 (21, 22) m = n = 1 (23) m = 1, n = 2 (24) m = n = 2 Figure 4-7. Synthesis of L-and D-Fluviabactin, L-Agrobactin, L-Homofluviabactin. a) N-Cbz-L-Thr-0Su or N-Cbz-o-Thr-OH / N-hydroxysuccimide / DCC; b) Pd/ C / EtOH / HCI; c) ethyl 2,3-dihydroxybenzimidate I MeOH

PAGE 80

66 Experimental General N-Carbobenzyloxy-L-threonine and N-Carbobenzyloxy-D-threonine were purchased from Sigama Chemical Co. (St. Louis, MO), N-Carbobenzyloxy-L threonine-1-(N-succinimidyl) ester was purchased from Sachem Bioscience Ina. (King of Prussia, PA) and all other reagents were purchased from Aldrich Chemical Co (Milwaukee, WI). Fisher Optima grade solvents (Fisher Scientific, Pittsburgh, PA) were used. DMF was distilled under N2 and stored over molecular seives Distilled solvents were employed for reactions involving chelators. Organic extracts were dried with sodium sulfate unless otherwise indicated. Glassware for chelator reactions and purification steps were soaked in 3 N HCI for 15 min, rinsed with distilled water then distilled ethanol and dried prior to use Silica gel 32-63 (40 M "flash") from Selecto, Inc. (Kennesaw, GA) or silica gel 60 (70-230 mesh) obtained from EM Science (Darmstadt, Germany) or Lipophilic Sephadex LH-20 from Sigma Chemical Co. (St. Louis, MO) was used for column chromatography. Optical rotations were determined at 589 nm (Na lamp) with a Perkin Elmer 341 polarimeter and 1 decimeter cell path length in the indicated solvent; c is expressed as g of compound per 100 ml. Proton NMR spectra were obtained on a Varian Unity 300 at 300 MHz in CD30D at ambient temperature unless otherwise indicated; chemical shifts are given in parts per million downfield from an internal tetramethylsilane standard. High resolution mass spectra were obtained utilizing FAB ionization from a

PAGE 81

67 glycerol matrix on a fennigan 4516. Elemental analyses were performed by Atlantic Microlabs (Norcross, GA) N1,N7-Bis[2,3-bis(benzyloxy)benzoyl]norspermidine (14) 2,3-Bis(benzyloxy)benzoic acid (43) (3.34 g, 10 mmol) and 1, 1carbonyldiimidazole (1.62 g, 1 O mmol) were dissolved in dry CH2Cl2 (100 ml) and stirred for 1 h at room temperature under a nitrogen atmosphere. A solution of norspermidine (0 67 g, 5.1 mmol) in dry CH2Cl2 (10 ml) was added and the mixture was stirred overnight. The resulting solution was washed with 2% Na OH (100 ml), water (100 ml) and brine (100 ml), then dried over MgS04 and filtered. Solvent removal in vacuo followed by flash chromatography of the residue on silica gel with 5% EtOH in EtOAc gave (14) (2.6 g, 67%) as a colorless oil: 1H NMR (CDCl3 ) 8 1.52-1.63 (m, 4 H), 2.45-2.50 (m, 4 H), 3.283.36 (m, 4 H), 5.07 (s, 4 H), 5 15 (s, 4 H), 7.10-7.49 (m, 26 H), 7 64-7.68 (m 2 H), 8.04-8.08 (br, 1 H). N1 N8-Bis[2,3-di(benzyloxy)benzoyl]spermidine (.1fil 2,3-Bis(benzyloxyl)benzoic acid (2.60 g, 7.78 mmol), COi (1.26 g, 7.78 mmol), and spermidine (0.56 g, 3.86 mmol) in dry CH2Cl2 (10 ml) were reacted and worked up by the method of (.H) Flash chromatography of the resulting residue on silica with 10% MeOH in CHCl3 gave(~) (2.2 g, 74%) as a pale yellow oil: 1H NMR (CD30D) 81. 37-1.44 (m, 4 H), 1.56-1.66 (m, 2 H), 2.38-2.44 (m, 2 H), 2.44-2.55 (m, 2 H), 3.20-3.26 (m, 4 H), 5.08 (s, 4 H), 5.16-5.17 (m, 4 H), 7 10-7.50 (m, 26 H). N 1 N 10 -Bis[2,3-di(benzyloxy)benzoyl]homospermid ine (1.Q)

PAGE 82

68 A mixture of 2,3-di(benzyloxyl)benzoic acid (2.10 g, 6.28 mmol), 1, 1carbonyldiimidazole (1.02 g, 6.29 mmol), and homospermidine (0. 5 g, 3.14 mmol) in dry CH2Cl2 (10 ml) were reacted and worked up by the method of ( 14). Flash chromatography of the resulting crude material on silica with 10% MeOH in CHCl3 gave 1 6 g (64%) (1) as a pale yellow oil: 1H NMR (CD30D) 8 1.41-1.48 (m, 8 H), 2.51-2.58 (m, 4 H), 3.22-3.28 (m, 4 H), 4.85 (s, 4 H), 5.089 (s, 2 H), 5.18 (s, 2 H). 7.01-7 .50 (m, 26 H). N4-(N-Carbobenzyloxy-L-threonyl)-N 1-N 7 -bis[2,3bis(benzyloxy)benzoyl]norspermidine (11) A solution of DCC (0.34 g, 1 .65 mmol) in dry CH2Cl2 (20 ml) was added to a solution of N-carbobenzyloxy-l-threonine (0.33 g, 1.30 mmol) and N hydroxysuccimide (0.19 g, 1 65 mmol) in dry CH2Cl2 (20 ml) and stirred for 18 h at room temperature. The mixture was filtered and (14) (1.0 g, 1.3 mmol) and triethylamine (200 mg, 2.0 mmol) added The resulting mixture was stirred for 24 h, concentrated in vacuo, then dissolved in EtOAc (100 ml). The solution was washed with H20 (2 x 50 ml), 10% citric acid (50 ml), and H20 (100 ml), dried and filtered. Solvent removal in vacuo followed by flash chromatography on silica with Et0Ac/CHCl3 (1 :1) gave (11) (0.68 g, 53%): 1H NMR (CD30D) 8 1 08-1.11 (m, 3 H), 1.51-1.72 (m, 4 H), 2.92-3 .31 (m, 8 H), 3.83-3.92 (m, 1 H), 4 38-4.42 (m, 1 H), 5 02-5 .21 (m, 10 H), 7.11-7 .51 (m, 31 H). N4-[N-Carbobenzyloxy-o-threonyl] N 1-N 7 -bis[2, 3-bis(benyloxy)benzoyl]norspermidine (1) A solution of DCC (0.63 g, 3 .04 mmol) in dry CH2Cl2 (40 ml), N carbobenzyloxy-D-threonine (0.76 g, 3.0 mmol), N-hydroxysuccimide (0.35 g,

PAGE 83

69 3.05 mmol) in dry CH2Cl2 (40 ml), and compound (14) (1. 5 g, 1.96 mmol) were combined and worked up by the method of (11), and afforded (1.a) (0. 62 g, 50%): 1 H NMR (CD300) 8 1 08-1.11 (d, 3 H, J = 7), 1.51-1 72 (m 4 H), 2.923.31 (m, 8 H), 3.83-3 .92 (m, 1 H), 4.38-4.42 (m, 1 H), 5 02-5.21 (m, 10 H), 7.117.51 (m, 31 H). Anal. Calcd. for C60H62N4010: C, 72.13; H, 6 25 ; N 5.61; Found : C, 71.87; H, 6 54; N, 5 53 N5-(N-Carbobenzyloxy-L-threonyl)-N 1-N8-bis[2.3 di(benzyloxy)benzoyl]spermidine (1m DCC (0.63 g, 3.0 mmol) in 40 ml dry CH2Cl2 N-carbobenzoxy-L-threonine (0.76 g, 3 0 mmol), N-hydroxysuccimide (0.35 g 3 0 mmol) in dry CH2Cl2 (40 ml), and (.15.) (1.5 g, 1 .93 mmol) were combined and worked up by the method of (11). Flash chromatography of the resulting residue on silica with Et0Ac/CHC'3 (1:1) afforded (1]) (1.08 g 55%) as a viscous oil : 1 H NMR (CD300) 8 1.13-1.16 (m, 3 H), 1.56-1 76 (m, 6 H), 2.94-3.35 (m, 8 H), 3.72-3.86 (m, 1 H), 4 32-4.41 (m, 1 H), 5.06-5 25 (m, 10 H), 7.10-7 50 (m, 31 H). Anal. Calcd for C61H64N4010: C, 72.31; H, 6.37; N 5.53; Found: C, 72.23; H, 6.54; N, 5.44. N6-(N-Carbobenzoxy-L-threonyl)-N 1 -N10-bis[2,3di(benyloxy)benzoyl]homospermidine @) A mixture of (.:lg) (1.5 g, 1.9 mmol), N-Carbobenzoxy-L-threonyl-N hydroxysuccimidate (Sachem 1 0 g 2.85 mmol), and triethylamine (0 20 g, 1 98 mmol) in CH2Cl2 (100 ml) were stirred overnight. The resulting solution was washed with H20 (100 ml), 10% citric acid (100 ml), and brine (100 ml),

PAGE 84

70 then dried and filtered. Solvent removal in vacuo followed by flash chromatography on silica with Et0Ac/CHCl3 (1:1) gave (2Q) (1.49 g, 76%): 1H NMR (CD300) 8 1.09-1.11 (d, 3 H, J = 7), 1 31-1.54 (m, 8 H), 3.01-3.28 (m, 8 H), 3 84-3.93 (m, 1 H), 4.42-4.45 (m, 1 H), 5.07-5.16 (m, 10 H), 7.09-7.48 (m, 31 H). Anal. Calcd. for C62H66N4010 : C, 72.49 ; H, 6.48; N, 5.45. Found: C, 72.26 ; H, 6 .56; N, 5.46. N4-L-Th reonyl-N 1-N 7 -bis(2, 3-dihydroxybenzoyl) norspermidine Hydrochloride (21) A mixture of (11) (0. 75 g, 0.75 mmol) and 10% Pd/C (0. 1 mg) in HCI saturated MeOH (50 ml) was stirred at room temperature for 1 h under a H2 atmosphere at ambient pressure. The mixture was filtered through acid washed Celite and was concentrated in vacuo to give (2..1) (0.40 g, 99%) as a white solid, which was used without further purification An pure sample was obtained by column chromatography on lH-20 with 10-15% EtOH/toluene: 1 H NMR (C0300) 8 1 25-1.28 (d, 3 H J = 7), 1.84-2 .02 (m, 4 H), 3.38-3.78 (m, 8 H), 4.02-4.10 (m, 1 H), 4.18-4.21 (m, 1 H), 6 68-6.75 (m, 2 H), 6 91-6.95 (m, 2 H), 7 18-7 .21 (m, 2 H). N4-D-Threonyl-N1N 7-bis(2,3-dihydroxybenzoyl)norspermidine Hydrochloride @) A mixture of (113) (0.30 g, 0 .30 mmol) and 10% Pd/C (0.15 g) in HCI saturated methanol (50 ml) were combined and worked up by the method of (2.1) to yield@) (0 .16 g, 99%) as a white solid, and used without further purification An analytical sample was obtained by column chromatography on lH-20 with 15% EtOH in toluene 1H NMR (CD300) 8 1.25-1.28 (d, 3 H, J = 7),

PAGE 85

71 1.84-2.02 (m, 4 H), 3.38-3.78 (m, 8 H), 4.02-4.10 (m, 1 H), 4.18-4 .21 (m, 1 H), 6.68-6 75 (m, 2 H), 6.91-6.95 (m, 2 H), 7.18-7.21 (m, 2 H). Anal. Calcd for C24H33CIN408 : C, 53.28; H, 6.15 ; N, 10.36. Found: C, 53.12; H, 6.36; N, 10.12. N4-(L-Threonyl)-N 1-N8-bis(2,3-dihydroxybenzoyl)spermidine Hydrochloride (23) A mixture of (.11}) (0.31g, 0.31 mmol) and 10% Pd/C (100 mg) in HCI saturated methanol (30 ml) were reacted and worked up by the method of (211 to give~) (0.16 g, 93%) as a white solid. The compound was used without further purification. An analytical sample was obtained by column chromatography on lH-20 with 15% EtOH in toluene: 1H NMR (CD300) 8 1 071 10 (m, 3 H), 1.52-1 73 (m, 6 H), 2.91-3.32 (m, 8 H), 3.82-3.91 (m, 1 H), 4.364.44 (m, 1 H), 6.67-6 73 (m 2 H), 6 90-6 93 (m, 2 H), 7.19-7 23 (m, 2 H). Anal. Calcd. for C25H35CIN408 : C, 54 20; H, 6 19; N, 10.11. Found: C, 54 11; H, 6.23; N, 9.87. N5-(L-Threonyl)-N1-N9-Bis(2,3-dihydroxybenzoyl)homospermidine Hydrochloride @) A mixture of (29) (0.6 g, 0 58 mmol) and 10% Pd/C (200 mg) in HCI saturated methanol (50 ml) were reacted and worked up by the method of (2.1). The resulting white solid (21) (0.3 g, 91 %) was used without further purification. An analytical sample was obtained by column chromatography on lH-20 with 15% EtOH in toluene 1H NMR (C0300) 81.19-1.21 (d, 3 H, J = 7), 1.41-1 54 (m, 8 H), 3 11-3.36 (m, 8 H), 4.44-4.53 (m, 1 H), 4.92-5 12 (m, 1 H), 6.84-7 28 (m, 6 H).

PAGE 86

72 Anal. Calcd. for C26H37CIN4 08 : C, 54 .88; H, 6.55; N, 9 85. Found: C, 54.76; H, 6.78; N, 9.78. N 4 -[2-(2, 3-Dihyd roxyphenyl)-( 4S, 5 R)-trans-5-methyl-2-oxazoline-4-ca rbonyl]-N 1 N 7 -bis(2. 3-d ihyd roxybenzoyl) norspermid ine (L-Fluviabactin) To a solution of (21) (0.16 g, 0.30 mmol) in dry, degassed methanol (20 ml) was added ethyl 2,3-dihydroxybenzimidate (0.22 g, 1.21 mmol) The mixture was heated at reflux under nitrogen for 30 h then concentrated in vacuo. Column chromatography on LH-20 with 10% EtOH/toluene gave L-fluviabactin (0. 12 g, 64%) as a white glass: 1 H NMR (CD 30D, 50 C) 8 1.40 (d, 3 H, J = 7), 1.84-1.94 (m, 2 H), 2.03-2.14 (m, 2 H), 3.40-3 .92 (m, 8 H), 4 67-4.88 (m, 1 H), 5.22-5.30 (m, 1 H), 6.01-6.76 (m, 3 H), 6.86-6.96 (m, 3 H), 7 12-7.21 (m, 3 H). HRMS Calcd. (M+H) 623.2353, found 623.2339; [a]=+ 89.18 (c = 1, MeOH). Anal. Calcd. for C31H34N 4 010 0.5H20: C, 58.95; H, 5.59; N, 8.87; Found: C, 59 24; H, 5.71; N, 8.81. N 4 -[2-(2, 3-D ihyd roxyphenyl)-( 4R, 5S)-trans-5-methyl-2-oxazol ine-4carbonyl]-N 1-N 7 -bis(2,3-dihydroxybenzoyl)norspermidine (D-Fluviabactin) A mixture of (n) (150 mg, 0.28 mmol) and ethyl 2,3-dihydroxybenzimidate (120 mg, 0.66 mmol) were reacted and worked up by the method of L Fluviabactin Column chromatography of the resulting residue on LH-20 with 15% EtOH/toluene gave D-fluviabactin (0 .11 g, 63%) as a white glass: 1 H NMR (CD 30D, 50 C) 8 1.40 (d, 3 H, J = 7), 1.84-1.94 (m, 2 H), 2 03-2.14 (m, 2 H), 3.40-3.92 (m, 8 H), 4 67-4.88 (m, 1 H), 5.22-5 .30 (m, 1 H), 6.01-6 76 (m, 3 H), 6 86-6.96 (m, 3 H), 7 12-7.21 (m, 3 H). HRMS Calcd. (M+H) 623.2353, found 623.2334; [a]= -85.86 (c = 1, MeOH).

PAGE 87

73 Found : C, 58.85; H, 6.13; N, 8.41. N5 -[2-(2, 3-Dihyd roxyphenyl)-( 4S, 5R)-trans-5-methyl-2-oxazoline-4carbonyl]-N 1-N8 -bis(2, 3-d ihyd roxybenzoyl)sperm i d i ne (L-Ag robactin) To a solution of~) (0.14 g, 0 25 mmol) in dry, degassed methanol (30 ml) was added ethyl 2,3-dihydroxybenzimidate (0.12 g, 0 66 mmol) The mixture was heated at reflux under nitrogen for 30 h then concentrated in vacuo Column chromatography of the resulting residue on lH-20 with 15% EtOH/toluene gave L-agrobactin (86) (0.10 g, 63%) as a white glass: 1 H NMR (CD30D, 50 C) 8 1 37-1.46 (m, 3 H), 1.58-2.12 (m, 6 H), 3.34-3.88 (m, 8 H), 4 70-4 .83 (m, 1 H), 5.20-5.29 (m, 1 H), 6.60-6.76 (m, 3 H), 6.85-6 95 (m, 3 H), 7 11-7 22 (m, 3 H). Anal. Calcd. for C32H36N4010 0.5H20: C, 59.53; H 5 78 ; N, 8.68. Found: C, 59.17; H, 5 .76; N, 8.54. N5-[2-{2,3-Dihydroxyphenyl)-(4S,5R) trans-5-methyl-2-oxazoline-4carboxamido]-N1-N10-bis(2,3-dihydroxybenzoyl)homospermidine (L Homofluviabactin) A mixture of@) (0.30 g, 0 .53 mmol) in dry, degassed methanol (50 ml) and ethyl 2,3-dihydroxybenzimidate (200 mg, 1.1 mmol) were reacted and worked up by the method of L-Fluviabactin. Column chromatography of the resulting residue on lH-20 with 15% EtOH/toluene gave l-homofluviabactin as a white glass (0.22 g, 64%): 1H NMR (CD30D, 50 C) 8 1.42-1.45 (d, 3 H J = 7), 1.55-1.88 (m, 8 H), 3 35-3 .78 (m, 8 H), 4 75-4 .94 (m, 1 H), 5 18-5.28 (m, 1 H), 6.63-6 76 (m, 3 H), 6.85-6.96 (m, 3 H), 7.12-7 23 (m, 3 H); Anal. Calcd for

PAGE 88

74 C33H38N4010 0.5H20: C, 60.08; H, 5.96; N, 8.49 Found: C, 60.01; H 5 .92; N, 8.75 Discussion The NMR analysis of polyamine catecholamides is widely studied (43,76, 86, 120). Previously, a 10:1 mixture of CDCl3:d6-DMS0 was used in this and other labs for the 1H NMR studies. The chloroform signal was set at 7.24 ppm as a reference in this solvent. The most useful information in the NMR is they methyl protons appeared between 1 2 and 1.4 ppm, as well as the aand ~ methine protons between 4.4 and 5.4 ppm. However, the polyamine backbone proton peaks (1.42 ppm -2.01 ppm) are very close to those of they-methyl region (1. 26 -1 .39 ppm) water (1.56 ppm), and DMSO (2.22 ppm). In this paper, another solvent, CD30D, was utilized. Using this solvent, there are some dramatic changes in the NMR. The water peak is shifted from 1.56 ppm in CDCl3 to 4.87 ppm in CD30D, and the solvent residual peak appears at 3.31 ppm. Thus, the polyamine backbone proton peaks and they methyl proton peaks are isolated and easily compared. Therefore, using CD30D as a solvent is a good method to identify and differentiate these compounds. Because all the amide and catechol protons are exchanged with CD300, these protons did not interfere with the peaks of a-, ~and y-methyl occurred, and the 1 H NMR spectra became simpler. In setting either the CD30D signal at 3.31 ppm or the water signal at 4.87 ppm, the 1H NMR spectra of L-and D-fluviabactin, L-agrobactin and Lhomofluviabactin can be separated into four basic sections. The aromatic

PAGE 89

75 protons appear between 6.6 ppm and 7 2 ppm; the aand ~-methine protons fall between 4.7 ppm and 5 3 ppm; the internal polyamine backbone protons extend between 1.55 ppm and 2.1 ppm; they-methyl protons appear between 1 3 ppm and 1 5 ppm (Figure 4-8 to Figure 4-11). The a-methine proton peaks overlapped with the water peak in the NMR spectra of final productsat room temperature: Land o-fluviabactin Lagrobactin, and L-homofluviabactin, although this did not happen in the 1H NMR spectra of all precursors. By setting the temperature at 50C the a-methine proton peaks were then separated from the water peak. The 300 MHz NMR spectra of L-and o-fluviabactin, L-agrobactin, and L-homofluviabactin were illustrated in Figure 4-8 to 4-11. It is worth noting that the 1 H NMR spectrum of agrobactin is different from those of oand Lfluviabact in, as well as from L-homofluviabactin. The most outstanding feature of agrobactin 1 H NMR spectrum is reflected in the duplic i ty of the NMR signals, whereas the 1H NMR spectra of the other three compounds lack this duplicity This is most evident when comparing they-methyl protons The 1 H NMR spectra of oand Lfluviabactin and L-homofluviabactin display a single doublet for y-methyl protons between 1.36 ppm to 1.42 ppm. However the 1 H NMR spectrum of agrobact i n has double doublets for y methyl protons from 1.3 ppm to 1 5 ppm The ratio between these two doublets is not equal, and is apparently sensitive to temperature and the solvent used. The duplicity of signals originating in the threonyl moiety of agrobactin is of particular interest. The threonine-coupling pattern in all four compounds

PAGE 90

-------HOD 0300 ------------/ / ___________ ,---/' _) .-~1--1 I I j I --,-----,-,I ., I 1-, ...--r--r1 .--, I I -rr--,--:-~ r r 1-r I 7 0 6 5 6 I 5 5 5 0 4.5 4.0 3 5 3 0 2.5 __,,,.----J ____/-,-, -,, 2.0 r-, pp10 Figure 4 8 300 MHz 1H NMR Spectrum of L-Fluviabactin in CD300 at 50 C

PAGE 91

100 0300 r -/ I r j 1 -I r-, 1 t 1-r-, J ----,.-r r T--J ,---r-, r 7 6 5 4 3 2 Figure 4-9 300 MHz 1H NMR Spectrum of D Fluviabactin in CD30D at 50 C

PAGE 92

OD ___ )J -.-..,-r 1 I r 1 -r r I r I -,--r-T----r--1 1 I r l r -f I 1 1 7 6 S 4 3 2 Figure 4-10 300 MHz 1H NMR Spectrum of LAgrobactin in CD30D at 50 C ........ co

PAGE 93

-----------------,-, -.-7 -----------I 6 5 --------------------------OD D30D ,----,~--r--r----, r-----r---r ----,---,1 1 -r----r -.-4 3 2 Figure 4-11. 300 MHz 1H NMR Spectrum of L-Homofluviabactin in CD30D at 50 C ........ c.o

PAGE 94

80 should display a simple spin system if there was not any conformational effects The a-methine should be split by the ~-methine producing a doublet. The ~-methine is coupled to the three y methyl protons and should show a quartet which is split again by the a-methine to produce eight lines, which would result in a complex multiplet. Finally, the y-methyl should be split mainly by the single ~-methine proton displaying a doublet with an estimated coupling constant between 5 5 to 6.5 Hz, since the long-range a-y coupling interaction is expected to be small. The threonyl y-methyl protons of agrobactin display two unequal doublets unlike those of L-and D-fluviabactin, and L-homofluviabactin. This suggests that there are at least two distinct magnetic environments for the threonine substituent in agrobactin It is worthwhile to note that the spermidine backbone d i splays a nearly linear conformation in the X-ray crystallographic study of agrobactin (105). It is assumed that fluviabactin would have a similar conformation as agrobactin. Then, relatively simple models could be established for these compounds, although they do not represent all of the possible orientations of the spermidine backbone and aromatic rings Because the y -methyl group is closer to the terminal 2 3-dihydroxybenzoyl groups than ~-methine or a-methine there would be a hindered rotation about the central amide, more changes in the magnetic environment and the protons will be easily influenced by anisotropic effects Thus in agrobactin, a proton located in a conformation lying to one side of the polyamine backbone could demonstrate a different signal from its counterpart on the other side of the

PAGE 95

81 polyamine backbone. In symmetrical analogues, such as fluviabactin and homofluviabactin however, the intramolecular distances on either side will be equalized, elim i nating the duplic i ty of signals observed for they-methyl of agrobactin. As expected, fluviabactin and homofluviabactin showed a simplified spectra. The y-methyl of both compounds exhibit only a single doublet, located at 1.40 ppm in fluviabactin and 1.44 ppm in homofluviabactin (Figure 4-12 to 4-14) Interestingly, the ratio of the two doublets of y-methyl in agrobactin is not equal, implying that either the cis or trans isomer is preferred. The chemical shift (L18) of y methyl between these two doublets is 0.04 Hz, which is very close to the difference of chemical shifts between the y-methyl of fluviabactin and homofluviabactin (0.035 Hz). When the 1H NMR spectrum of agrobactin was run at room temperature, the integration of downfield doublet of y-methyl was always larger than that of the upfield one. On looking at the 1 H NMR spectra of these compounds, the downfield doublet of y-methyl of agrobactin matches that of homofluviabactin whereas the upfield one matches that of fluviabactin This suggests that agrobactin prefers to adapt a conformation close to that of homofluviabactin There are two kinds of isomers in agrobactin : cis and trans isomers. When the polyamine side containing four methylenes is on the same side of they-methyl, it forms a cis isomer ; whereas when the polyamine backbone side is on the opposite side of the y-methyl, it forms a trans isomer. The steric hindrance in cis isomer might be less than that of the trans isomer due to the

PAGE 96

"' ... "' ...... <") .... ... m I 1 1 1 1 I 1 1 1...,-,,-,--,--rm1 1 1 1 1 I I r,,1-, -rr, I 1 1 1 1 I rr 2.2 2.0 1.8 1.6 1.4 Figure 4-12 300 MHz 1H NMR of y-Methyl Group and Polyamine Backbone C-Methylene Group of Fluviabactin in CD300. 00 N

PAGE 97

Z 2 2 0 1. 8 1. 6 l. ii Figure 4-13 300 1H NMR Spectrum of y Methyl Group and Polyamine backbone C-Methylene Groups of Agrobactin

PAGE 98

... "'"' "'" .... ... ... I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 1. 8 1. 6 1. 4 Figure 4-14 300 MHz 1H NMR Spectrum of they-Methyl group and Polyamine Backbone C Methylene Group of Homofluviabactin in CD30D.

PAGE 99

85 longer methylene chain. Also, the central amide carbonyl can be hydrogen bonded to the propyl amide hydrogen. It is predicted that agrobactin would prefer to adapt a cis isomer rather than a trans isomer (Figure 4-15). L-Fluviabactin L-Homofluviabactin trans (E) L-Agrobactin cis (Z) Figure 4 15 Conformations of L-Agrobactin, L-Fluviabactin and L-Homofluviabactin in CD300.

PAGE 100

CHAPTERV THE EFFECTS OF RHIZOFERRIN, D-AND L-FLUVIABACTIN, L-AGROBACTIN AND L-HOMOFLUVIABACTIN ON IRON TRANSPORT IN PARACOCCUS DENITRIFICANS Introduction Iron is required by virtually all microorganisms with the exception of some lactobacilli. However iron tends to form insoluble polymers under physiological conditions, so the solubility product of ferric hydroxides is only 10-38 at pH 7.4 (6). The efficient utilization of iron by microorganisms requires the production of iron-specific chelating ligands known as siderophores to solubilize iron in a biologically available form (121,122) A wide variety of bacteria and fungi produce siderophores as well as membrane transport proteins to facilitate uptake of hydrophilic siderophore complexes across nonpolar cell membranes (123,124) Siderophore-mediated iron transport has been investigated primarily by following the fate of radiolabeled siderophore iron complexes in transport assays, although recent investigations have applied electron paramagnetic resonance and Mossbauer spectroscopy to the experiments (125,126,127). Iron transport studies can also be conducted by substituting a transition metal ion that has a similar ionic radius, the same charge, and the same coordination geometry as Fe(III). For example, Cr(III) forms complexes with siderophores that are structurally similar to ferric siderophores but are kinetically inert to ligand substitution (128), while Ga(III) forms complexes with 86

PAGE 101

87 kinetically inert to ligand substitution (128), while Ga(III) forms complexes with siderophores that are kinetically labile but cannot be reduced (129). A primary goal of these iron transport studies has been to examine the structural requirements for microbial iron uti lization of siderophore-iron complexes, including stereospecificity of transport mechanisms. The studies contained in this paper examine these requirements in Paracoccus denitrificans by structural modifications of the natural siderophore catecholamide L-parabactin. Structural homologues of L-parabactin include L-fluviabactin, L-agrobactin and L-homofluviabactin. D-Fluviabactin is studied to examine the issue of stereospecificity. Finally, the ability of Paracoccus denitrificans to utilize rhizoferrin a non-catecholamide citric acid-based siderophore is studied L-Fluviabactin is a siderophore isolated from Vibrio fluvialis one of the pathogenic vibrios associated with infantile gastroenteritis (79) It has been demonstrated that L-fluviabactin is important to iron acquisition in Vibrio fluvialis (79) and may play an important role in the virulence of this pathogen. Thus, it is important to understand how ferric L-fluviabactin can be utilized by microorganisms This study involves the examination of utilization of ferric L fluviabactin in Paracoccus denitrificans a well-studied non-pathogen Paracoccus denitrificans is a gram-negative soil bacterium that produces L-parabactin in response to low iron concentration (130) Previous studies suggested that Paracoccus denitrificans utilized ferric L-parabactin as well as such other ferric siderophores as ferric desferrioxamine B and ferric Lvibriobactin (131). Iron utilization in Paracoccus denitrificans was sens i tive to

PAGE 102

88 the stereospecificity of the ligand, with o-parabactin unable to promote its growth (132) Furthermore, the Ga(III) L-parabactin complex was accumulated in a manner similar to ferric L-parabactin in the transport experiments, which implied a non-reductive mechanism in Paracoccus denitrificans (131) The iron transport mechanism in Paracoccus denitrificans was also examined by conducting kinetic studies of iron acquisition from ferric siderophores (132). The data demonstrated a stereospecific high affinity binding and suggested a possible nonstereospecific postreceptor processing in iron accumulation by Paracoccus denitrificans L-Parabactin displayed biphasic kinetics with both high-affinity and low-affinity components in the Lineweaver-Burk reciprocal plot, however, o-parabactin, L-parabactin A, and o parabactin A had only the low affinity components It was determined that the Km of the high-affinity transport system for ferric L-parabactin in Paracoccus denitrificans was 0.24 0.06 M (132) A stereospecific high-affinity receptor for ferric L-parabactin has been isolated and partially purified from the Paracoccus denitrificans outer membrane (46). The five phenolic oxygens and the oxazoline-ring nitrogen of L-parabactin have been implicated as the iron chelating functional groups in the ferric Lparabactin complex ( 132,133). This current study focuses on the effects of structural changes of the polyamine backbone and the stereochemistry of the oxazoline ring in catecholamides on iron transport in Paracoccus denitrificans. The natural siderophore L-fluviabactin was modified by either changing the polyamine backbone to generate L-agrobactin and L-homofluviabactin or

PAGE 103

89 inverting the stereochemistry of the oxazoline ring to yield D-fluviabactin. The abilities of the natural siderophore and its derivatives were then evaluated in the microbial iron transport system. As mentioned above, iron utilization in Paracoccus denitrificans is characterized by its ability to utilize exogenous non-native siderophores to acquire iron. It has been demonstrated that some bacteria and fungi are able to accumulate iron from ferric citrate as well as ferric citrate-based siderophores (134,135). This study also involves the investigation of iron utilization of Paracoccus denitrificans from ferric rhizoferrin, a novel chiral citric acid-based siderophore. Experimental Preparation of Siderophore Chelates Ferric nitrilotriacetate [ferric(NTA)z] was prepared by the addition of 2.2 volumes (10% excess) of 30 mM trisodium NTA to 1 volume of 30 mM 55FeCl3 stock solution in 0.1 N HCI followed by immediate adjustment of the pH to 7 0 with 500 mM Tris HCI to give a final stock concentration of 5 mM ferric(NTA)z. The 55Fe ferric chelates of the catecholamides were prepared by ligand exchange with ferric(NTA)z A 10% molar excess of catecholamide in the ethanol was added to 5 mM ferric(NTA)z followed by sufficient 500 mM Tris HCI (pH 7.4) to give a 300 mM ferric catecholamide solution Preparation of Low-iron Defined Minimal Salts Medium To a 1 O liter flask was added 70 8 g of succinic acid, 48.0 g of KH2P04 60.6 g of NaH2P04 and 19.2 g of NH4CI. These materials were dissolved in 8

PAGE 104

90 liters of distilled deionized water. The pH was adjusted to 4 5 and the solution was autoclaved and allowed to sit at 4C for 3-4 days to allow iron salts to coagulate The solution was filtered through a 0.2 M membrane and after the pH had been adjusted to 7.0 with 50% NaOH, was passed through a column of 500 g of Chelex 100 resin. The pH was again adjusted to 7 0 with 1.0 N HCI, then 300 ml of chelex-treated 20% Tween 80 solution was added to prevent clumping of cells, followed by trace element medium described below and the final volume was adjusted to 12 liters. The complete minimal salts medium contained an optimized mixture of trace elements : Mg2+ (1. 7 mM), Ca2+ (182 M), Mn2+ (1 O M), Zn2+ (1 M), Cu2+ (0.1 M), and Co2+ (0.01 M). The iron concentration of this medium was 0.05-0.1 M confirmed by atomic absorption spectroscopy. Bacterial Strain and Culture Conditions Paracoccus denitrificans ATCC 177 41 was incubated in tryptic soy broth medium and maintained in sterilized vials by freeze-drying Individual colonies were inoculated into 20 ml of tryptic soy broth and incubated with rotary shaking for 12 h at 30C. Inoculations were then made from this culture into 50 ml of minimal salts liquid medium containing 1 m Fe(III) and incubated with shaking at 30C for 12 h. This culture was then used to inoculate a 250 ml culture flask of 50 ml of minimal salts medium containing 0.5 M Fe (Ill) to give a starting [00]660 of 0 .05. The flask was incubated with shaking at 30C for 10 h. The medium tested positive for catechol ([00]515 > 0.1) when estimated

PAGE 105

91 using Arnow's reagents (131). This culture was used in the studies described below. Growth Promotion Study Growth studies were carried out in minimal salts media without additional iron in the presence of 1.1 mM ethylenediamine-di(o-hydroxyphenyl)acetic acid (EDDA), a non-utilizable iron chelator. This effectively created an iron-starved environment. The ability of a siderophore to promote growth under these conditions was assessed by the addition of the siderophore to the medium which was inoculated at a starting [00]660 = 0.016-0.018, then monitoring the growth rate at [00]550. Iron Accumulation Studies The cells from 0.5 m Fe minimal salts solution described above were prepared for iron accumulation studies by centrifugation at 15-20C followed by washing with iron-free minimal salts medium at 15-20C. The washed pellet was then suspended in iron-free minimal salts medium used as transport buffer adjusted to [00]660 = 1.00 and incubated with shaking in a water bath at 30C for 15 minutes. The studies were initiated by the addition of 5 M of radiolabeled chelate to each 5.00 ml of cell suspension. Portions (500 l) were withdrawn at 1, 5, 10, 20, 30, 40, 50, and 60 minutes The transport reaction was stopped by adding the 500 l portion to 5 ml of prechilled (0C) transport buffer. The cells were separated from the external medium by rapid filtration using Whatman GF/F glass filters that had been presoaked for 24 h in 500 M unlabeled chelate and rinsed with 5 ml of chilled transport buffer. After

PAGE 106

92 two rinses with 5 ml of chilled transport buffer, the filters were dried at 70C and placed in a scintillation vial with 10 ml of Biofluor. Cell-associated radioactivity was measured by liquid scintillation counting. All data were corrected for cell growth during the assay period. Kinetic Transport Assay The cells from 0.5 m Fe minimal salts solution were prepared for transport assay by centrifugation at 15-20C followed by washing with iron-free minimal salts medium at 15-20C The washed pellet was then suspended in transport buffer adjusted to [0D]660 = 1.00 and incubated with shaking in a water bath at 30C for 15 minutes. The assays were initiated by the addition of 5 M of radiolabeled chelate to 5.00 ml of cell suspension. Portions (500 l) were withdrawn every 20 seconds within the initial 2 minutes. The transport reaction was stopped by adding the portion to 5 ml of prechilled (0C) transport buffer. The cells were separated from the external medium by rapid filtration using Whatman GF/F glass filters that had been presoaked for 24 h in 500 M unlabeled chelate and rinsed with 5 ml of chilled transport buffer. After two rinses with 5 ml of chilled transport buffer, the filters were dried at 70C and placed in a scintillation vial with 10 ml of Biofluor. Cell-associated radioactivity was measured by liquid scintillation counting. All data were corrected for cell growth during the assay period. The six datum points taken during the first 2 minutes for each assay tube were fitted to a regression line with a slope equal to the initial reaction velocity (V0). These initial rates, expressed as picogram-atoms of 55Fe per minute per

PAGE 107

93 milligram of protein, were used to generate Lineweaver-Burk plots to estimate Km, V max, and standard errors. Rates varied linearly with the number of cells present from [OD]660 values of 1 .00 to 2.00. All data were corrected for cell growth during the assay period. Since all the assays were finished within two weeks, the modest variation in transport competence among culture batches could be ignored. In addition, an assay of 55Fe ferric L-parabactin was performed for comparison. Results Growth Rate and Induction of Catecholamide Production The growth rate of Paracoccus denitrificans in iron-free minimal salts medium to which had been added 0.5 M Fe was checked by measuring the [OD] value at 660 nm every four hours for 24 hours. In addition, the Arnow's assay (136) was used to determine the production of catechols by measuring [OD] value at 515 nm. The growth rate and the production of catechols were plotted (Figure 5-1). After 10 hours in 0.5 M Fe medium, at a point when the [OD] value of the bacterium-containing medium reaches about 1.00, catechol production by Paracoccus denitrificans was induced. The cells at this phase of growth were used for iron transport assays. Growth Promotion Studies Of the compounds tested for their abilities to promote growth, Land o fluviabactin contained a norspermidine backbone, but with inversion of the stereochemistry of oxazoline ring, the polyamine backbones in L-agrobactin and L-homofluviabactin are spermidine and homospermidine, respectively.

PAGE 108

94 4 2.0 1 8 1 6 3 1.4 ,,i 1 2 0 / l!)
PAGE 109

95 The native siderophore L-parabactin with a sperm i dine backbone was used as a reference. The effects of L-parabactin L-and o-fluv i abactin, L-agrobactin and L-homofluviabactin on the growth rate of Paracoccus denitrificans under iron starvation conditions as described above in Methods" are illustrated in Figure 5-2. Growth stimulation is shown to be highly stereospecific, with L-fluviabactin promoting more growth than D-fluviabactin For example, starting at the same [00]660 (0 016-0.018) of bacterium-containing media, the [00]660 in the presence of 2 0 M L-fluviabactin reaches about 0 75 after 35 hours. This value in the presence of 2 0 M 0-fluviabactin is only about 0.3, which is not different from control (Figure 5-2). In spite of the different polyamine backbones all the catecholamides derived from L-threonine (Lfluviabactin, L-homofluviabactin, Lagrobactin, and L-parabactin) have pronounced effects on bacterial growth These results imply that the stereochemistry of the oxazoline ring in catecholamides, not the polyamine backbone, plays the more important role in iron utilization by Paracoccus denitrificans. Iron Accumulation Studies Iron accumulation from 55Fe ferric L-and D fluviabactin, L-agrobactin Lhomofluviabactin, and L-parabactin by Paracoccus denitrificans is illustrated in Figure 5-3. The bacterium accumulates 55Fe from 55Fe ferric L fluviabactin much more efficiently than from 55Fe ferric D-fluviabactin For instance, Paracoccus denitrificans accumulates more than 20 ng-atoms of 55Fe per mg protein from 55Fe ferric L-fluviabactin over a period of 30 minutes, while it only accumulates about 2 ng-atoms of 55Fe per mg protein from 55Fe ferric D-

PAGE 110

0 co co ,......., 0 0 96 1 0 ....... -------------------0.8 0.6 L-PB L AB L-FB L HB 0 4 0.2 D-FB control o o~-.....---.....---.....---.....---.....---......--......--......--......... 0 1 0 20 30 40 50 time(h) Figure 5-2. Growth Rate of P denitrificans in the Presence of 2.0 M Ligands: L-Parabactin (L-PB), L-Fluviabactin (L-FB), L-Agrobactin (L-AB) L Homofluviabactin (L-HB), and 0-fluviabactin (D-FB) (mean s.d., n=3).

PAGE 111

30 -C) E 20 Cl) Q) 0 E C Q) LL LO LO Q) .::t:. 10 ct1 a. ::, 0 0 97 [55Fe]Ferric L-PB --0[55Fe]Ferric L-AB [55Fe]Ferric L-FB -0-[55Fe]Ferric L -HB 1 0 20 30 40 50 60 time(min) Figure 5-3. Iron Accumulation of [55Fe]Ferric L-Fluviabactin (L-FB) 0-Fluviabactin (D-FB) L-Homofluviabactin (L-HB) L-Agrobact i n (L-AB) and L-Parabactin (L-PB) by P. denitrificans ( mean s d., n=3). 70

PAGE 112

98 fluviabactin during the same time. Thus, iron accumulation in Paracoccus denitrificans is shown to be highly stereospecific. The accumulation of 55Fe from 55Fe ferric L-fluviabactin, L-homofluviabactin, L-agrobactin, and Lparabactin do not differ to nearly same degree. This is consistent with the idea that the changing of the stereochemistry of the oxazoline not the changing of the polyamine backbone on the polyamine-containing catecholamides, has the more pronounced effect on iron utilization. This was further confirmed by the related kinetic transport studies described below Kinetic Studies The kinetics of 55Fe transport from 55Fe ferric L-fluviabactin, D-fluviabactin, L-homofluviabactin and L-agrobactin by Paracoccus denitrificans are presented in Lineweaver-Burk reciprocal plots. The X-axis of the plot is the reciprocal of substrate concentration expressed as 1/M and the Y-axis is the reciprocal of velocity represented as 1/picogatoms 55Fe per minute per milligram of protein. An estimated Km or V max can be calculated from the intercept of the X-axis or the Y-axis, respectively Figure 5-4 illustrates the dramatic differences in transport of ferric D-fluviabactin as compared with ferric L-fluviabactin. The individual point data in Figure 5-4 represent the means of three experiments at each substrate concentration. The fitted line in Figure 5-4 for the D-fluviabactin data is representative of each of the three experiments with D-fluviabactin, i.e., data for each separate experiment was also fitted well by a single straight line with a similar X-intercept corresponding to a Km of about 3 5 M. There was, however

PAGE 113

99 1.00 D -fl uv i abactin (0 1-5 0 M) 0.75 0 .50 0.25 L-Fluviabactin ( 0 1-5 0 M) 0 .00 +-,-.1...f"'-.,.......-,--,-,-..,......,-,-~-r-......... ...,......,.......-........,-r-T""T-t -1 0 2 3 4 5 6 7 8 9 10 11 1/(S) uM-1 Figure 5-4. Kinetic Characteristics of Iron Transport of [55Fe] Ferric L-Fluviabactin and 0-Fluviabactin in P. denitrificans (mean)

PAGE 114

100 some variability in the slopes of these three o-fluviabactin lines as indicated by the variation in the value of the Y-intercept shown by the standard deviation of Vmax It is important to note that even given this large standard deviation, the range of Vmax values for ferric D-fluviabactin is substantially lower than Vmax values for ferric L-fluviabactin. These differences in velocity of transport are especially pronounced at lower iron concentrations (e. g., < 1 M), as illustrated in Figure 5-4 by the small magnitude of the ferric L-fluviabactin data in this double reciprocal plot. These findings again suggest that the stereochemistry of the chiral oxazoline ring plays a critical role in the iron utilization of ferric catecholamides by Paracoccus denitrificans. In addition to illustrating marked differences in the Vmax of transport, Figure 5-4 suggests other differences of mechanistic interest. Iron accumulation data from 55Fe ferric D-fluviabactin fit a straight line, and thus obey a simple Michaelis-Menten model (Km= 3.48 0.81 M, Vmax = 96.2 62 6 picogram atoms of 55Fe/min/mg protein). For 55Fe ferric L-fluviabactin, it is apparent that the data do not show a linear relationship Figure 5-5 is a detail of the ferric Lfluviabactin data in Figure 5-4 and shows that iron accumulation data from 55Fe ferric L-fluviabactin is nonlinear and apparently biphasic. If analyzed separately, the data from 1 0 M to 5.0 M of 55Fe ferric L-fluviabactin fit one line (r = 0.92, Km= 2.15 1.63 M), implying a low-affinity system, while the data from 0.1 M to 1.0 M of of 55Fe ferric L-fluviabactin fit another line (r = 0.95, Km= 0 233 0.03 M), suggesting a high-affinity system. Paracoccus denitrificans thus appears to have a stereospecific high-affinity system to accumulate iron from

PAGE 115

> T"" 101 0 .014-------------------A High Affinity System ([SJ <1 0.012 Km = 0.233 Vmax = 129 0 .010 0.008 0.006 0.004 0 .002 Km= 2.15 Vmax = 413 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 1 /[ s 1 Figure 5-5 Kinetic of Iron Transport of [55Fe] Ferric L-Fluviabactin in P denitrificans (mean s.d., n=3).

PAGE 116

102 ferric L-fluviabactin under low iron conditions (< 1 M). In addition, the organism has a low-affinity system that may not be stereospecific since the organism is able to effectively obtain iron from ferric o-fluviabactin under high iron conditions A high-affinity iron uptake system is involved in the iron accumulation of ferric L-fluviabactin, probably because there is an efficient and stereospecific binding between ferric L-fluviabactin and the outer membrane receptors. Previous studies demonstrated the induction of a stereospecific high-affinity ferric L-parabactin receptor in Paracoccus denitrificans membranes under low iron concentrations with a Kd estimated to be 0.7 M (46). It thus seems likely that this receptor plays a role in the high affinity system illustrated in Figure 5-5. The other catecholamide chelates containing L-oxazoline system, ferric Lagrobactin, and ferric L-homofluviabactin also appear to exhibit biphasic kinetics with both a high-affinity (0. 1 M < [S] < 1.0 M) and a low-affinity (1. 0 M < [S] < 5 0 M) component. The Kms for high-affinity systems of ferric Lagrobactin (Figure 5-6) and ferric L-homofluviabactin (Figure 5-7) are 0.17 0.04 M and 0.11 0.002 M, respectively All the data for ferric L-and o fluviabactin, L-agrobactin, and L-homofluviabactin are summarized in Table 1. Ferric L-fluviabactin, L-agrobactin, and L-homofluviabactin have the same sterochemistry of oxazoline ring derived from L-threonine, but different polyamine backbones. Nonetheless, they all display biphasic iron transport kinetics with both a high-affinity and a low-affinity component. Ferric o fluviabactin derived from o-threonine lacks the high-affinity component. These

PAGE 117

> T""" 103 0.010------------------0.009 A 0.1-1.0M y = 3.4071 e-3 + 3.9786e-4x RA2 = 0.965 0.008 Km = 0 .11 M, Vmax = 1 0.007 0.006 0.005 0 .004 0.003 0.002 0.001 RA2 = 0.999 0. 000 -t,n,nnr...,.,........,rnm,mnn'l'l'ffi'lm,mnnnffl'fflrnmi"""'fflffl"""""'"""""'"""'rrnl -9-8-7-6-5-4-3-21 0 1 2 3 4 5 6 7 8 9 1 01 11 2 1 /[ S] Figure 5-6. Kinetic of Iron Transport of [55Fe] Ferric L-Agrobactin in P. denitrificans (mean s.d., n=2).

PAGE 118

> ....... T"" 104 0 .014 0 .012 0.1-1 .0M 0 010 y = 3.7046e3 + 6.4081e-4 x 0 008 0 006 0 .004 0 002 R"2 = 0 952 -7-65 -4-3-2-1 0 1 2 3 4 5 6 7 8 9 101112 1 / [ S] Figure 5-7. Kinetic of Iron Transport of [55Fe] Ferric L-Homofluviabactin in P denitrificans (mean s.d., n=3).

PAGE 119

Chelate a Ferric L-Fluviabactin (3) Ferric D-Fluviabactin (3) Ferric L-Homofluviabactin (3) Ferric L-Agrobactin (2) Table 1. Iron accumulation kinetic data High Affinity Component Kn (M) b Vmax c 0 233 0.03 0 17 0 04 0.11 0 002 129.0 1 9 137 9 18 7 154 8 1 7 Low Affinity Component Kn (M) b Vmax c 2.15 1.63 3.48 0 .81 1.03.12 0.54 0 007 413.4 149. 96 2 62.6 229.2 28.6 244.9 0 6 a Number in parentheses refers to the number of assays at each chelator concentration (0.1, 0.2, 0.5, 1.0, 2 .0, 5 .0, 10.0 M) b Km standard error determined as described in Material and Methods c Picogram-atoms of 55Fe per minute per miligram of protein ( standard deviation). 0 01

PAGE 120

106 results further support the concerpt that it is the stereochemistry of the oxazoline ring, not the polyamine backbone, that is the dominant factor in the iron accumulation by Paracoccus denitrificans. It suggests that iron accumulation in Paracoccus denitrificans is probably initiated by stereospecific binding between the chelate and the outer membrane receptor. Ferric o fluviabactin is not recognized by these receptors, thus, is not utilized in this high-affinity iron transport system. To summarize, all the ferric catecholamides containing L-oxazoline rings displayed biphasic kinetics, while the o-oxazoline analogue ferric D-fluviabactin exhibited a different kinetic profile without a high affinity component. Th i s may be explained by the potential contributions of molecular dissymmetry. The similarities in ring size and the nature of the chelator donor centers but differences in the stereochemistry of oxazoline rings of ferric L-fluviabactin and ferric o-fluviabactin allow for the direct comparison of the CD spectra of both complexes (Figure 5-8). The positive CD band maximum (~ = +1.4 at 530 nm) associated with the low-energy transition in the visible spectrum of ferric L-fluviabactin (Amax = 518 nm) is characteristic of the A chelate enantiomer when compared to the CD spectrum of ferric L-parabactin ; the CD spectrum of ferric D-fluviabactin on the other hand indicates a~ chelate enatiomer, the mirror-image of the A chelate enantiomer (Figure 5-8). The configuration of ferric L-fluviabactin (Figure 5-9) differs from that of ferric D-fluviabactin (Figure 5-10) at three chiral centers : two asymmetric carbons in the oxazoline ring derived from L-(2S,3R)-threonine and

PAGE 121

f:3 0 E 107 8 6 1111111111 Ferric L-Fluviabactin 4 2 0 ----------.. ------------2 -4 -6 -8 -10 : ... .. .. . . .. .. . . . . .. .. .. .. I .. .. .. .. : .. .. . .. ... .. .. ;; .. .. .. .. .. .. .. ... .. .,. Ferric D-Fluviabactin -12 --~---~--~--.--..-....-........ --.--..--........ --.--..-....--340 390 440 490 540 590 Wavelength (nm) Figure 5-8. CD Spectra of Ferric L-and D-Fluviabactin (400 M in 100 mM Tris HCI buffer [pH 7.41).

PAGE 122

108 0 0 I\ Figure 5-9. Configuration of Ferric L-Fluviabactin Represented a A Isomer.

PAGE 123

109 0 a 0 Figure 5-10. Configuration of Ferric D-Fluviabactin Represented As a Isomer.

PAGE 124

110 o-(2R, 3S)-threonine as well as a metal center chiral configuration. Ferric Lfluviabactin forms a A chelate, a left-handed propeller, while ferric D-fluviabactin is a chelate, a right-handed propeller, as illustrated in Figure 5-9 and 5-10, respectively The Effect of Rhizoferrin on the Growth Rate of Paracoccus denitrificans The fact that Paracoccus denitrificans can utilize such other siderophores as L-fluviabactin and L-agrobactin rather than its native Lparabactin suggested it would be worthwhile to test other structurally dissimilar exogenous siderophores. When, rhizoferrin, a novel chiral citric acid-based siderophore, was examined in the growth promotion studies of Paracoccus denitrificans the addition of synthetic rhizoferrin did not promote bacterial growth when compared with L-parabactin and control (Figure 5-11). Previous studies demonstrated that citrate was also unable to stimulate the growth of Paracoccus denitrificans under these same conditions (131) These results imply that there are no efficient and specific bindings between citrate or citrate based siderophores and the bacterial outmembrane receptors. Discussion Iron Utilization by Bacteria and Fungi With the possible exception of lactobacilli, iron an essential nutrient, is required by almost all microorganisms (137) However microorganisms have to overcome two problems in order to obtain this metal. First they have to overcome the insolubility of Fe(III). The solubility product of ferric hydroxide is only 10-38 (138) Second, they must regulate iron uptake because of iron's

PAGE 125

0 co co -Cl 0 ....... 111 0.8 .....-----------.;.__ ________ 0.6 L-parabactin 0.4 0 2 rhizoferrin --frcontrol 0.0 --~.....-.--.--.-..--.--.-..--.--.-..--.--.-..--.--.-..--.--.-~-~-~--1 0 5 1 0 1 5 20 25 30 Time (hr) Figure 5-11. Growth Rate of Paracoccus denitrificans in the Presence of 1.0 Mligands: L-Parabactin and Rhizoferrin (mean s.d. n=3).

PAGE 126

112 potential toxicity. Fe(II) and Fe(III) can react with oxygen to generate hydroxyl radicals, which are the most potent oxidizing agents known (139). Therefore, microbes have developed various strategies for acquiring iron while at the same time protecting themselves from iron's toxic effects. The major strategies used by bacteria and fungi to acquire iron include production and utilization of siderophores, and reduction of Fe(III) to Fe(II) with subsequent transport of Fe(II) and utilization of host iron compounds such as heme transferrin and lactoferrin Many microbes rely on chelators, which they may or may not synthesize themselves, to solubilize iron The natural iron chelators produced by microbes have been collectively termed: siderophore (from the Greek meaning iron carrier) and are defined as low molecular weight, virtually iron-specific ligands that facilitate the solubilization and transport of Fe(III). More than 100 siderophores have been isolated, and most can be divided into two families: hydroxamates and catecholamides. Despite the considerable structural variations found among siderophores, they all form six-coordinate octahedral complexes with Fe(III) Siderophores can be isolated and identified either by bioassay or chemically (137,138). Bioassays were used long before the chemical structures of various siderophores were known and remain the most sensitive analytical methods for siderophore detection. The development of the chrome azurol S (CAS) assay, which is independent of sideophore structure, has made it easy to detect siderophore production on an agar plate. Early chemical assays for siderophores were specific for hydroxamates and

PAGE 127

113 catecholamides For example Arnow's reagent was used for the detection of catechol production (136). Siderophore-iron complexes are transported into cells via specific transport systems. The transport may be initiated by specific bindings between the siderophore-iron complexes and the outer membrane receptors. Generation of both siderophores and their receptors is primarily regulated by the iron concentration in the medium. Paracoccus denitrificans use siderophores to acquire iron from the environment. When the bacterium was incubated in 0.5 M iron minimal salts solution, Paracoccus denitrificans in response to this iron concentration started to produce catechols which can be detected by Arnow's assay Tait had originally isolated three iron-binding catechol containing compounds from cultures of Paracoccus denitrificans, which he referred to as "Compounds I, II, Ill" (74) (Figure 5-12) Tait correctly determined the structures of compounds I and II to be 2 3-dihydroxybenzoic acid and N1 ,N8-bis(2,3-dihydroxybenzoyl) spermidine, respectively. Later it was shown that Tait's compound Ill was Lparabactin (130). Tait was also able to show that the microorganism synthesized compound II from compound I, and compound Ill from compound II plus biosynthesized salicyl i c acid. Furthermore, L-parabactin and its analogs L-fluviabactin, and L-agrobactin all can reverse the EDDA-inhibiting growth of Paracoccus denitrificans.

PAGE 128

114 (XOH _J ___ -oH HO 0 Compound I H H O OH HO~N~N~N~OH OHO H u Compound II rOH 0 N )-{fo O OH HO~~~N~N~OH OHO H u Compound Ill Figure 5-15. Structures of the Catechol Compounds Isolated from Paracoccus denitrificans Reduction is another possible way microbes may obtain iron. Ferric polymer can be bound onto the cell surface Subsequent reduction of the metal by the cells via a membrane flavin reductase would solubilize iron from the

PAGE 129

115 polymers and generate the appropriate form for transport by a membrane ferrous translocating system. Although much is known about bacterial transport of Fe(III), much less is known about utilization of Fe(II) Some microbes, including Bifidobacterium bifidum (140), Legionel/a pneumophila (141) and Streptococcus mutans (142) use Fe(II) exclusively as an iron source Both L. pneumophila and S. mutans are believed to reduce Fe(III) before transporting Fe(II) (141,142). An example of bacteria utilizing Fe(II) as an iron source is Streptococcus mutans (142). Because the presence of oxygen in the environment of S. mutants would favor the formation of profoundly insoluble ferric iron the possibility of catechol or hydroxamate siderophore excretion by S. mutans was studied Generally, the 2,3-dihydroxybenzoic acid could be quantitatively detected at 0 9 g/ml and lower concentrations could be qualitative visualized. However, ethyl acetate extraction of 600 ml of the low-iron culture supernatant of S. mutans and concentration (100-fold) of the extract by vacuum evaporation failed to demonstrate Arnow's reagent-reactive substances in the concentrated extract. Additionally, hydroxamates also could not be detected in the supernatant. However, in an iron uptake assay done aerobically, the reductant sodium ascorbate (5 mM) markedly increased radioiron uptake accumulated by viable cells from 13 pmol/mg to 111 pmol/mg (142). This suggested that the reducing capacity of ascorbate was responsible for the accumulation. Ascorbate probably reduced the radioiron to the ferrous state, which was then transported by the cells.

PAGE 130

116 Another strategy that microoragnisms utilize to acquire iron is to use host iron compounds, such as heme, transferrin, and lactoferrin. Several pathogens can utilize heme or hemoglobin as iron sources. Yersiniae pestis, the causative agent of the bubonic plague or Black Death, is the most notorious heme-utilizing bacteria (143). However, the ability to utilize heme and hemoglobin is often not the only strategy an organism has at its disposal for acquiring iron For example, the vibrios genus synthesize siderophores in addition to utilizing heme (144). Some pathogens can utilize transferrin or lactoferrin-binding iron or both. Transferrin and lactoferrin are both single polypeptides with two high-affinity iron-binding sites. Unlike a siderophore mediated system in which siderophores are released from cells that compete with host proteins for iron, utilization of transferrin and/or lactoferrin requires direct contact between the host iron-binding protein and the bacterial cell. For example, Neisseria spp. acquires iron from human transferrin by using a periplasmicbinding protein-mediated ,active transport system (145). Despite the insolubility of Fe(III), microbes are also able to utilize polymeric forms of Fe(III) via a low-affinity system. This system has been designated "low affinity" because relative high levels of iron are required to achieve maximal bacterial growth rates. The evidence for the presence of such a system is that microorganisms that have lost their high-affinity iron uptake systems can still grow in minimal media (146). The low-affinity system appears to be nonspecific and does not require carriers or membrane receptors. Lacking these entities, it is difficult to conceptualize the molecular

PAGE 131

117 mechanics of the system and hence very little is known about its basic features In a low-affinity system, it is possible that some of the surface atoms of ferric hydroxide polymers may be less firmly bound and hence available to the cell. An alternative possibility is that metal-binding sites may be built into the cell envelope (147). It is also suggested from studies with mutants of Neurospora that hydroxy acids could aid in the dissolution of adsorbed ferric polymers (148). The low-affinity system is repressed by the iron chelators nitriloacetic acid and 2,2-dipyridyl. Nitrilotriacetate is often used to block low affinity uptake in E. coli (149). A high-affinity iron uptake system is comprised of two parts, namely the siderophore itself and the cognate transport apparatus. Siderophores are low molecular iron-specific ligands produced by microorganisms under a low iron condition Microorganisms use siderophores to convert insoluble ferric hydroxides to soluble siderophore-iron complexes. During iron transport, the iron-laden form of the siderophore first contacts a surface receptor. In the case of gram-negative bacteria, the outer membrane constitutes a permeability barrier for water-soluble solutes larger than ca 500 Daltons (147). In the evolution of siderophores, it was necessary to synthesize a ligand that could contact all six octahedrally directed bonds of the Fe(III) atom. The receptors are believed to behave as pores or channel forms in the sealed outer membrane. The high-affinity pathway of iron assimilation appears to be unique for Fe(III) and has not been found for other essential metal ions. Possibly this is due to several factors, including the insolubility of Fe(III) at biological pH, the

PAGE 132

118 significantly higher requirements for iron as opposed, for example, to copper, and the fact that the divalent metal ions require nothing more sophisticated than an a-amino carboxyl group to act as a chelation center. Paracoccus denitrificans apparently developed both high-affinity and low affinity systems for the acquisition of iron. Many factors such as the nature of ligands iron concentration or both, differentiate these two systems. For example, Paracoccus denitrificans is able to use both high-affinity and low affinity systems to obtain iron from such ferric polyamine catecholamides based on L-threonine as L-parabactin and L-fluviabactin When the i ron concentration is low Paracoccus denitrificans utilizes a high-affin i ty system to obtain iron from these chelates A rapid growth rate and a high Vmax in the kinetic studies under a low iron concentration has demonstrated the existence of such an efficient iron transport system However when the iron level is high a low-affinity iron uptake system was utilized and verified by different Km and Vmax on a double reciprocal plot (132) The contribution of a high-affinity system to a low-affinity system is unknown but the differences between these two systems are marked and apparently subject to iron concentration On the other hand, Paracoccus denitrificans was unable to use its high-affinity system to obtain iron from ferric o-analogues of polyamine catecholamides derived from o-threonine including o-parabactin and o-fluviabactin no matter what iron concentration was used. Iron utilization by most gram-negative bacteria and fungi is an energy related process. In Paracoccus denitrificans L the high-affinity iron uptake

PAGE 133

119 system is apparently an energy-dependent process, while the low-affinity system is energy independent. This was demonstrated by the absence of a high-affinity iron transport at a very low temperature. The low-affinity iron transport system in Paracoccus denitrificans is not affected by the temperature. For example, during the transport assay of ferric D-fluviabactin in this study similar kinetics were obtained whether the cells were treated at 4 C or 30 C during the transport assay of ferric D-fluviabactin. Or furthermore, an energy independent iron transport was observed in Mycobacterium smegmatis The iron uptake from ferric exochelin was facilitated by diffusion and occured readily at 4 C (150). Most importantly, iron utilization in most gram-negative bacteria and most fungi is characterized by high stereospecificity. For example, Paracoccus denitrificans can only utilize the high-affinity system to obtain iron from ferric Lcatecholamides derived from L-threonine. This probably because the outer membrane receptors could only recognize one type of isomers of the chelates, t1 or A isomer, but not both. Ferric L-Parabactin tends to form a A chelate, while D-parabactin has a t1 isomer (131) The stereospecific ferric parabactin binding activity in Paracoccus denitrificans was checked by electrophoresis gels. When the outer membrane proteins were incubated with [55Fe]ferric L-or o-parabactin and then washed with a nondenaturing, nonionic detergent, Triton X-100, antoradiograms showed that the more intensely labeled band was seen with [55Fe]ferric L-parabactin While [55Fe]ferric D-parabactin did label some

PAGE 134

120 bands, the most intensely labeled band was clearly different from those seen with [55Fe]ferric L-parabactin as the ligand The stereospecificity of iron transport in Paracoccus denitrificans was further investigated in the kinetic studies. Paracoccus denitrificans has a preference for all catechoamides derived from L-threonine, such as natural Lparabactin, L-vibriobactin, and L-fluviabactin although the bacterium does not produce those siderophores. For example, iron accumulation by Paracoccus denitrificans from ferric L-fluviabactin and o-fluviabactin apparently is different. Paracoccus denitrificans only uses a low-affinity system to acquire iron from ferric o-fluviabactin, but it applies both highand low-affinity systems to obtain iron from L-fluviabactin. It has been determined that all L-catecholamides form A isomers with ferric ions, while ferric o-catecholamides form 8. isomers. This was further supported by the CD spectra of L-parabactin compared to o parabactin and L-fluviabactin compared to D-fluviabactin It is possible that the outer membrane protein of Paracoccus denitrificans only recognizes the A isomers formed by ferric L-catecholamides Interestingly, neither the chelate's center chirality nor the stereochemistry of the oxazoline ring alone determines iron utilization by Paracoccus denitrificans. Although the outer membrane receptors of Paracoccus denitrificans only recognize A isomers formed in ferric L-catecholamides, it still can accumulate 55Fe from [55Fe]ferric o-catecholamides. On the other hand, ferric L-parabactin A forms a A isomer, while ferric o-parabactin A forms a 8. isomer, although the marked stereospecific kinetic difference which

PAGE 135

121 distinguishes ferric D-parabactin from ferric L-parabactin transport was not observed in the iron transport of L-parabactin A and D-parabactin A. Iron Transport Mechanisms in Bacteria and Fungi The role of iron in infection and neoplasia has received considerable attention in recent years ( 151-153). Siderophores play important roles in the acquisition of iron by bacteria and fungi. Although certain microbes can interchangeably utilize the siderophores of other prokaryotes, this ability is not universal. Thus, siderophores have potent bacteriostatic and fungistatic effects and appear to act by out-competing some microbes for available iron. The molecular mechanisms of siderophore-mediated iron transport provide unique opportunities to facilitate the development of siderophores and their analogs as antibacterial agents (154). Although great efforts have been taken to study microbial iron transport systems using double radiolabeled ferric siderophores and other methods, the mechanisms still remain the same : iron shuttle and iron taxi mechanisms (134,155) In the iron shuttle mechanism, the intact siderophore-iron complex is incorporated into the cells, followed by intracellular dissociation of the metal ligand complex (50, 156) This process can be initiated by a stereospecific binding between a siderophore-iron chelate and a specific membrane receptor on the outer cell surface. The intact siderophore-iron complex then penetrates into the cell, perhaps facilitated by certain membrane transport apparatus. Once in the cell, iron is released from the complex either by hydrolysis or

PAGE 136

122 reduction. The deferrated ligands then may be used by the biological system for another round of iron transport. Previous examples of the iron shuttle mechanism include ferrichrome in Ustilago sphaerogena (157), schizokinen in 8. Megaterium (125), aerobactin in A. aerogenes ( 125), and enterobactin in E. Coli ( 158). A recent study demonstrated that the uptake of 55Fe mediated by rhizoferrin in Morganella morganii was also accomplished using the iron shuttle mechanism (159). In this study a chase with excess of nonradioactive ferric rhizoferrin was performed. 55Fe was rapidly taken up and remained in the cells during the chase period. Further, ferric [3H] ketorhizoferrin was incubated with M. morganii 13 and chased with nonradioactive ferric rhizoferrin. Radioactive rhizoferrin was taken up by the cells and remained in them during the chase period demonstrating that rhizoferrin also functions as an iron carrier and does not only solubilize ferric iron by delivering it to a cell-bound iron carrier [3H]Ketorhizoferrin not loaded with iron was only slow taken up, and this residual uptake might have been caused by traces of iron in the transport assay that binds to the radioactive ketorhizoferrin The name iron taxi mechanism i mplies that the intact siderophore is not transported into the cells There is only a transient association between the siderophore-metal complex and a specific outer membrane receptor (123). The initial step is still the binding between a siderophore-metal chelate and a specific outer membrane receptor However compared to the iron shuttle mechanism, after donating the metal ion to the outer membrane receptors, the

PAGE 137

123 siderophore will stay outside the cell. The iron is then transferred into the cells by an iron-binding acceptor protein at the membrane surface (123). Other examples of iron taxi mechanism include rhodotorulic acid in Rhodotorula pilimanae (50) and ferric exochelins in Mycobactin smegmatis (150,160) Interestingly, iron uptake from ferric citrate by Mycobacterium smegmatis is also demonstrated to use the iron shuttle mechanism (135). In uptake studies, 55Fe was solubilized by adding citrate at a molar ratio of 200: 1. M. Smegmatis and M Bovis BCG were able to uptake 55Fe from the solublized ferric dicitrate complex. The rate of this uptake was comparable with the rate of uptake of iron from ferric exochelin into cells of M. Smegmatis Further, uptake of 14C from [1,5-14C]citrate was studied in an iron-deficient cell suspension of M Smegmatis. Only a slight absorption of [14C]citrate was observed over 1 hour. Thus it appears that, although 55Fe is taken into the cell from ferric citrate, the citrate is not taken into the cells under similar conditions The iron uptake mediated by L-parabactin in Paracoccus denitrificans_ was shown to be due to an iron taxi mechanism (131). In double radiolabelled experiments, when [55Fe]ferric parabactin was added to the cell suspensions at 30C, a rapid uptake of the [55Fe] was observed. However, when [3H]ferric parabactin was substituted for [55Fe]ferric parabactin, only a small amount (<10%) of tritium label was obtained during the first minute, and no additional uptake of [3H] was observed over the entire course of the transport assay. The initial small absorption of tritium was attributed to some form of binding of the label at the cell surface, and not the actual transport of parabactin into the cell.

PAGE 138

124 The iron shuttle and iron taxi mechanisms can be used to explain most siderophore-mediated iron transport by microorganisms, but are not all inclusive Nonetheless both mechanisms have to solve the same problem, that is, how to release iron from a siderophore-iron complex. Ferric siderophores are usually very tight complexes with iron-binding constants as high as 1030 Microorganisms have to develop a method to compete with the intense binding between a siderophore and iron. Although the mechanisms of iron transport have been studied in detail, the mechanisms responsible for the release of siderophore-bound iron to microorganisms are not yet fully elucidated. The possible ways of removing iron from ferric siderophores by microorganisms include simple ligand exchange reduction of siderophore bound iron from ferric to ferrous, hydrolytic cleavage of the ligand, or a combination of the above methods The ligand exchange mechanism is based on the fact that the iron binding constants between ligands are quite different. Therefore, it is possible that iron could be transferred from the loose complex to a tight chelator. Microorganisms may use their natural siderophores to obtain iron from other ligands when the iron-binding constant of a natural ferric siderophore is much greater than that of the ligand, although use of this mechanism is quite unlikely despite the different iron formation constants. For example Paracoccus denitrificans was able to obtain iron from ferric parabactin as well as from ferric ferrioxamine B (132). Although the iron associate constant strongly favored ferric L-parabactin (1048 ) over ferrioxamine B (1030) on thermodynamic grounds, it was unlikely that

PAGE 139

125 a ligands-metal exchange mechanism was involved. The ligand exchange between parabactin and desferrioxame B was checked by spectrophotometrically recording the catecholamide-iron absorption band (A = 530 nm). After 5 minutes, less than 1 % iron exchange occurred, however, during the 2 min duration of a transport assay, more than 10% [55Fe] was accumulated from the ferric hydroxamate siderophore complex. Because siderophores typically have much weaker affinity for Fe(II) than Fe(III), the reduction of the siderophore-bound ferric ion to the ferrous state has been suggested as a possible mechanism for metal release in microorganisms (42,161). With the siderophore-bound iron in the Fe(II) state, the water soluble Fe(II) can either be used directly by microorganisms or transported by a ferrous ion specific system. For example, the iron uptake in yeast was proposed to reduce Fe(III) to Fe(II) by a plasma membrane ferric reductase, so the Fe(II) crossed the membrane via an Fe(ll)-specific transport system (162). In the case of Saccharomyces cerevisiae it has been demonstrated that the yeast has two transporters for Fe(II), a high-affinity transporter at Fe(II) concentrations below 5 M and a low-affinity transporter operating at concentrations above 5 M (163). In fact, ferric siderophore reductase activities have been found in the cell extracts of many microorganisms. For example, a portion of the ferrisiderophore reductase activities in Bacillus subtilis is catalyzed by a flavin reductase that is responsible for aromatic biosynthesis. The reductase activity is shown to be stimulated by flavin mononucleotide (FMN) plus magnesium

PAGE 140

126 (164) This suggests that ferrisiderophore reductase may be a multifunctional enzyme responsible for iron transport as well as molecular biosynthesis (120,164). Another example of microorganisms demonstrating such reductase activity toward ferric chelates of catecholamide siderophores is Agrobacterium tumefaciens. Reduction of the iron in ferriagrobactin by the cytoplasmic fraction of Agrobacterium tumefaciens strictly required NADH as the reductant (165). Although various ferric siderophore reductase activities in microbial cells are increasingly well documented, the problem of iron release from the natural siderophore is still unresolved. For example, it was found that iron reductase activity in membrane preparations of Ataphylococcus aureus was associated with L-lactate dehydrogenase, glycerol 3-phosphate dehydogenase, and nitrate reductase activities using L-lactate and glycerol 3-phosphate as reductants (166). However, the success of the iron reductase depended on using ferric citrate. Little or no reduction occurred when ferri-ferrichrome and ferri ferrioxamine 8 were used with crude extracts from this organism, as well as from Spiril/um itersonii, Rhodopseudomonas spheroides, and E. coli ( 167). There is no evidence that uptake of iron is via formation of ferric citrate, an easily reducible complex It was also suggested that ferrimycobactin reductase activity in mycobacteria may be generally distributed rather than being the property of a specific enzyme (168). Another possible mechanism of releasing iron from ferric siderophore is that siderophores may be only used to transport a single molecule of ferric ion into a bacterium, for example L-parabactin, and are then hydrolyzed the

PAGE 141

127 oxazoline of L-parabactin by enzymes The degraded molecule has a much lower iron binding constant than a natural siderophore and can not be used again. Evidence for such a mechanism is that the degraded products could be isolated from the supernatant of the bacterial culture. Examples of such an iron transport system include enterobactin in E. coli and triacetylfusign in Penicillium sp (169,170,171). It was reported that hydrolases were responsible for the intracellular hydrolysis of both siderophores (170). The hydrolytic product L-parabactin A of L-parabactin was isolated and characterized from Paracoccus denitrificans. In the kinetic studies, the open form of the oxazoline ring had profound effects on iron transport in the bacterium: Paracoccus denitrificans was unable to use its high-affinity system to acquire iron from ferric L-parabactin A (132) This implied that the chemistry of the oxazoline system may contribute to the special properties of siderophores of Paracoccus denitrificans and confer biological advantages. For example, the oxazoline ring nitrogen provides a nonionic ligand that participates in the chelation of ferric ion; the oxazoline ring may also increase conformational rigidity and may result in a much more compact ferric chelate than is otherwise possible; and lipophilicity is increased because of the oxazoline ring. Furthermore, the properties of both the free ligand and the ferric chelate may be changed by simple hydrolytic cleavage of the oxazoline ring; This is important in deferration of the chelate, allowing the organism access to the iron.

PAGE 142

128 Iron release from siderophore-bound iron could also be a combination of ligand exchange, reduction, or hydrolysis The destruction of the ligands may be an essential step before reduction in order to save metabolic energy. In the case of ferric enterobactin, the hydrolytic product, ferric-tr i s (dihydrobenzoyl)serine has a redox potential well in the range of known physiological reducing agents, while the redox potential of ferric enterobactin at physiological pH is beyond the range of known biological reductants. Therefore, for a long time it was generally accepted that hydrolysis of the intact ferric enterobactin complex must be a prerequisite for iron release, followed by reduction of bound Fe(III) to Fe(II) (2,6). However the ferric chelates of a synthetic enterobactin without the ester linkages was shown to be as effective as ferric enterobactin in promoting the growth of E. coli and Bacillus subtilis (172,173). This suggested that the reductase activities or protonation in the microorganisms may overcome the exceptionally low redox couple of ferric catecholamides. Because an iron transport in Paracoccus denitrificans was operating via an iron-taxi mechanism and ferric siderophore reductase activity has been observed on ferric L-parabactin in Paracoccus denitrificans (74) it appears that there is no need for the destruction of the ligands since the siderophore never enters the cell and therefore could not compete within the cell for intracellular iron. Additionally, if ferric L-parabactin is hydrolyzed to ferric L-parabactin A, that might facilitate iron transport The iron transport could be initiated by stereospecific binding of ferric L-parabactin to outer membrane receptors,

PAGE 143

129 followed by hydolysis of ferric L-parabactin to ferric L-parabactin A. The L parabactin A-bound iron is presumably reduced to ferrous iron and removed by ferrous ion acceptor protein present at the cell surface. Soon after metal removal from L-parabactin A, the free ligand is replaced at the binding site by another ferric L-parabactin complex. Such an iron transport system in Paracoccus denitrificans has several advantages over microorganisms operating an iron shuttle mechanism. First, the lower redox potential of the open-chain threonyl structure of ferric L parabactin A (E0 = -0.400 mV; pH 7.0) over the oxazoline ring of ferric L parabactin (E0 = -0.673 mV; pH 7.0) strongly favors the hydrolytic process (174). Second, since there is no accumulation of free ligand in the cytoplasm, the bacterium might save the energy that is required to modify and expel the deferrated siderophore from the cell. Finally, this reduction and hydolysis iron taxi system may be highly specific for the transport and assimilation of iron. In iron kinetic studies, it was observed that such other metals as Al, Mg, and Co, which have close coordination stereochemistry with Fe(III), did not demonstrated any effect on the high-affinity iron transport in Paracoccus denitrificans. The Role of Citrate in Microbial Iron Transport Many microbes can produce and utilize more than one siderophore. A more powerful chelator might be produced when the first, less powerful, chelator fails to provide enough iron to the bacterium. This stepwise

PAGE 144

130 mechanism of siderophore production has been observed in Azocobacter vinelandii ( 175,176) as well as Paracoccus denitrificans ( 131 132). Citrate is known to be a component in many siderophores. Citrate itself is a relatively poor ligand for Fe(III), with a formation constant that lies many orders of magnitude below those of the usual siderophores, which tend to cluster around 1030 Nonetheless, citrate is able to chelate ferric ions and provide iron to microorganisms. For example, in a survey of meningococca/ strains, the production of conventional siderophores failed to be detected, although citrate was especially active in supplying iron to the cells (177). Thus, citrate was defined as a "functional siderophore" (177). It has been shown that the relative concentrations of iron and citrate are important in determining the structure of the ferric-citrate complex formed (178,179). At an acidic pH, an equimolar concentration of iron and citrate can form an anionic chelate. Fe(III) citrate polymerizes between pH 8 and 9 to give an insoluble spherical complex with a molecular weight of about 2 X 105 This complex probably has an iron hydroxide core with citrate ions bound to the surface. At a physiological acidity (pH 7.4) and in the presence of excess citrate, the soluble anionic chelate ferric dicitrate (Fe(cit)2)5 is formed which competes kinetically with the formation of the monocitrate chelate. At pH 7.7 with Fe(III) at 1 mM and citrate at 31 mM, the concentration of Fe(III) monocitrate and Fe(III) dicitrate becomes equal. Thus the pH and relative concentration of iron and citrate in growth media and uptake systems appear to be crucial to the function of citrate as an iron transport compound.

PAGE 145

131 Many bacteria were assumed to have citrate-mediated iron transport systems; thus, citrate iron uptake systems were widely studied (134,135) In E. coli, a citrate-mediated iron transport system was induced by the presence of citrate in the growth medium that was independent of the enterobactin system (180) The ferric citrate receptor induced had been detected on gel and the subunit size was about 80 kDa (181) Although Salmonella typhimurium LT-2 and E. coli K12 are closely related, they may be distinguished on the basis of their responses to citrate or its iron complex Salmonella typhimurum uses citrate as its carbon source whereas E. coli does not; the organisms exhibit an inverse specificity for iron citrate (146). Although great efforts have been made to establish the role of citrate in iron metabolism in bacteria and fungi, considerable controversies still exist about how and under what circumstance microorganisms use citrate to transport iron Bacteria and fungi may use citrate instead of siderophores to accumulate iron when they lack the energy to biosynthesize the natural products Ferric citrate was also shown to stimulate reductase activity, which is responsible for reduction of Fe(III) to Fe(II), but there is no evidence to support iron transport via the formation of ferric citrate (181). The citrate-mediated system of iron acquisition may be important in the pathogenicity of mycobacteria, since citrate is able to chelate iron from lactoferrin and other naturally occurring iron-binding compounds. The normal range of citrate concentrations in adult human serum is between 100 to 135 M. Although this is high enough to facilitate the growth of mycobacterial pathogens, assuming it

PAGE 146

132 were to complex 1/10 or even a 1/100 of its weight as iron, it is far from clear whether this would happen in situ. In this study, Paracoccus denitrificans did not utilize citrate or citrate-based siderophore rhizoferrin. Probably there are no binding sites specific for citrate or citrate-based siderophores on the cell surface. Ferric rhizoferrin was unable to bind to the outer membrane receptors, thus, the consequent iron transport did not occur. On the other hand, the citrate-mediated system, if it exists in Paracoccus denitrificans, may be only operational under special conditions, such as a very acidic medium, but not under these conditions.

PAGE 147

CHATER VI CONCLUSIONS The synthesis of rhizoferrin was completed in this work. The absolute configurations of two chiral citrates in rhizoferrin were confirmed to be (R,R) by the X-ray crystallography of the 1,2-dimethyl citrate-brucine salt. The chiral citrate separated in the scheme was also used in the synthesis of another chiral citrate-containing siderophores, staphyloferrin A. The stereochemistry of the two chiral citrates in staphyloferrin A was suggested as (R R) by CD spectrum, although the final product was partial! racemized in the o-ornithine segment. An efficient synthetic scheme was applied for the synthesis of polyamine catecholamides L-fluviabactin and its analogs (L-agrobactin, L-homofluviabactin and o-fluviabactin). The effects of rhizoferrin and the polyamine catecholamides on iron transport in Paracoccus denitrificans were evaluated by the studies of growth rate as well as iron accumulation and kinetics All catecholamides except o fluviabactin stimulated the growth rate of Paracoccus denitrificans; however, rhizoferrin was not able to show any effects on the bacterial growth under the same conditions. In accumulation studies, Paracoccus denitrificans accumulated Fess from Fess radio labeled chelates ferric L-fluviabactin, Lagrobactin, and L-homofluviabactin much more efficiently than from ferric o-133

PAGE 148

134 fluviabactin In kinetic studies, the catecholamides derived from L-threonine displayed biphasic kinetic features on Lineweaver-Burk plot with both high affinity and low affinity components The Km for the high-affinity kinetics of Lfluviabactin was determined to be 0.23 0 .06 M. On the other hand the data from Fess ferric D-fluviabactin displayed a single linear relationship on a double reciprocal plot, implying a low-affinity component in the iron uptake kinetics. In addition, the CD spectra of L-fluviabactin and D-fluviabactin were mirror images. Thus the following conclusions were drawn from the above studies: 1) Paracoccus denitrificans can use catecholamides other than native siderophores such as L-fluviabactin, but not the citrate-based siderophore rhizoferrin. 2) The changing of the polyamine backbone of polyamine catecholamides has little effects on the iron transport in Paracoccus denitrificans, while the inverting of stereochemistry of oxazoline ring has profound effects 3) All catecholamides derived from L-thereonine displayed biphasic kinet i c featu res with both high-affinity and low-affinity components; only D-fluviabactin lacks the high-affinity component. 4) The molecular dissymmetries of the chelates contributed to the distinctive kinetics. While the configuration of ferric L-fluviabactin is A chelate, ferric o-fluviabactin has a chelate. 5) Citric acid-based siderophore, rhizoferrin, did not promote the growth of Paracoccus denitrificans under a low iron condition.

PAGE 149

REFERENCES 1 Peter, M. M., David, R. W. 1980. The inorganic chemistry of iron metabolism. In A. Jacobs, and M. Worwood (ed.), Iron in biochemistry and medicine. Academic Press, London and New York. 2 Cooper, S. R., McArdle, J. V., Raymond, K. N. 1978. Siderophore electrochemistry: relation to intracellular iron release mechanism. Proc. Natl. Acad. Sci. U S.A. 75:3551. 3 Ortiz de Montellano, P.R. 1986. Ortiz de Montellano, P.R. (ed.), Cytochrome p450-structure, mechanism and biochemistry. Plenum Press, New York. 4. Sahlin, M., Peterson, L., Graslund, A., Ehrenberg, A., Sjoberg, B. M., Thelander, L. 1987 Magnetic interaction between the tyrosyl free radical and the antiferromagnetically coupled iron center in ribonucleotide reductase. Biochemistry. 26:5541. 5. Winkelmann, G., Van Der Helm, D., Neiland, J. B. 1987 Winkelmann, G Van Der Helm, 0., Neiland J. B. (ed.), Iron transport in microbes, plants and animals. VCH Publishers, Weinheim, West Germany. 6. Raymond, K. N., Carrano, C. J. 1991. Coordination chemistry and microbial iron transport. Biol. Metals. 4: 1-6. 7. Fletcher, J., Huehns, E. R. 1968. Significance of the binding of iron by transferrin. Nature. 218:1211. 8. Fischbach, F. A., Gregory, P. W., Harrison, P. M. 1971. On the structure of hemosiderin and its relationship to ferritin. J. Ultrastruct. Res. 37:495. 9. Bergeron, R. J. 1986. Iron: a controlling nutrient in proferative processes Trends Biochem. Sci. 11 :133. 10. Neilands, J. B. 1981. Microbial iron compounds. Annul. Rev. Biochem. 50:715. 135

PAGE 150

136 11. Hider, R. C. 1984. Siderophore-mediated absorption of iron Struct. Bonding (Berlin) 58:25 12 Crichton, R. R. 1991. In inorganic biochemistry of of iron metabolism Ellis Horwood, New York. 13 Babbs, C. F. 1992. Oxygen radicals in ulcerative colitis. Free Radical. Biol. Med 13:169-181. 14. Liu, D., Yang, R., Yan, X., McAdoo, D. J. 1973. Hydroxyl radicals generated in vivo kill neurons in the rat spinal cord: electrophysiological histological and neurochemical results J Biol. Chem. 248:7792-7797. 15. Keshavarzian, A., Morgan, G., Sedghi, S., Gordon, J. H., Poria, M. 1990. Role of reactive oxygen metabolites in experimental colitis Gut 31 :780-790. 16. Ben-Shachar, D., Eschel, G., Riederer, P., Youdim M. B. 1992. Role of iron and iron chelation in dopaminergic-induced neurodegeneration: implication for Parkinson's disease. Ann. Neural. 32:S105-S110 17. Halliwell, B. 1992 Oxygen radicals as key mediators in neurdogical disease: fact or fiction. Ann. Neural. 32:S1O-S15 18. Hallberg, L. 1981. Bioavailability of dietary iron in man Ann Rev. Nutri. 1 :123-47. 19. Jacobs, A. 1980. The pathology of iron overload. In Jacobs A., and M. Worwood (ed.), Iron in Biochemistry and Medicine II. p 428456. Academic Press London and New York. 20. Brittenham, G. M., Cohen, A. R., Mclaren, C. E., Martin, M. 8., Griffith, P. M., Nienhuis, A. W., Young, N. S., Allen, C. J., Farrell D. E., Marris, J. W. 1993 Hepatic iron stores and plasma ferritin concentration in patients with sickle cell anemia and thalassemia major. Am. J Hematol. 42:81-85 21. Gordeuk, V. R., Thuma P. E., Brittenhan, G. M. 1994 Iron chelation therapy for malaria. Adv Exp. Med Biol. 356:371 382. 22. Halliwell, B. 1989. Protection against tissue damage in vivo by desferrionxamine : what is its mechanism of action? Free Radical. Biol. Med. 7:645-651.

PAGE 151

137 23. Hershko, C. 1994. Control of disease by selective iron depletion: a novel therapeutic strategy utilizing iron chelators. Bailleres Clin Haematol. 7:965-1000. 24. May, P. M., Linder, P. W., William, P.R. 1978. Sigel, H. (ed.), Metal Ions in Biological Systems. Marcel Dekker, New York. 25. Bothwell, T. H., Charlton, R. W., Cook, J. 0., Finch, C. A. 1979. Iron Metabolism in Man, Blackwell Scientific Publications London. 26. Harrison, P. M., Fischbach, F. A., Hoy, T. G., Haggis, G. H. 1967. Ferric oxyhydroxide core of ferritin Nature 216: 1188 27. Morgan, E. H. 1979 Studies of the mechanism of iron release from transferrin. Biochim. Biophys. Acta 580:312. 28. Casey, J. L., Hentze, M. W., Koeller, D. M., Caughman, S. W., Rouault, T. A., Klausner, R. D., Harford, J. B. 1988. Iron responsive elements:regulatory RNA sequences that control mRNA levels and translation. Science 240:924-928 29 Rouault, T. A., Hentze, M. W., Koeller, D. M., Caughman, S. W., Harford, J. 8., Klausner, R. D. 1988. Binding of a cytosolic protein to the iron-responsive element of human ferritin messenger RNA. Science 241: 1207-1210. 30. Leibold, E. A., Munro, H. N. 1988 Cutoplasmic protein binds in vitro to a highly conserved sequence in the 5 -untranslated region of ferritin heavyand lightsubunit mRNAs. Proc. Natl. Acad. Sci U.S .A. 85:2171-2175. 31. Schalinske, K. L., Blemings, K. P., Steffen, D. W., Chen, 0. S., Eisenstein, R. S. 1997. Iron regulatory protein 1 is not required for the modulation of ferritin and transferrin receptor expression by iron in a murine pro-~ lymphocyte cell line Proc. Natl. Acad Sci. U.S.A. 94:10681-10686 32. Moore, C. V., Arrowsmith, W.R., Welch, J., Minnich, V. 1939. Studies in iron transportation and metabolism IV. Observations on the absorption of iron from the gastro-intestinal tract. J. Clin. Invest. 18:553 33. Rhodes, J., Seton, D., Brown, D. A. 1968. Absorption of iron instilled into the stomach, duodenum, and jejunum Gut 9:323.

PAGE 152

138 34. Worwood, M. 1979. in Iron Excess, Aberrations of iron and Porphyrin Metabolism, Seminars in Hematology, Vol. 14. 35. Huebers, H., Huebers, E, Rummel, W. 1970. Mechanism of iron absorption: iron-binding proteins and dependence of iron absorption on an elutable factor. In Kirf, H (ed.), Iron Metabolism and Its Disorders. 13-21. American Elsevier, New York. 36. Seligman, P.A., Klausner, R. D., Huebers, H. A. 1987. Molecular mechanisms of iron metabolism. In Stamatoyanno poulos, G., Nienhuis, A. W., Leader, P. Majeris, P W. (eds.), The Molecular Basis of Blood Disease. 219 Philadelphia. 37. Hoffbrand, A. V. 1980 Transfusion siderosis and chelation therapy In Jacobs, A., Worwood, M. (eds .), Iron in Biochemistry and Medicine II. P. 499. Academic Press, Lodon. 38. Weatherall, D. J. 1981. The iron-loading anemias. In Martell, A. E., Anderson, W. F. and Badman, D. G. (eds.), Development of Iron Chelator for Clinical Use" Elsevier, North-Holland, New York. 39. Aksoy, M., Seyithanoglu, B. Y., Bozbora, A. 1984. Thalassemia and desferrioxamine (Desferal) introduction remarks, and clinical laboratory observations. In Aksoy, A., Birdwood, G F B (eds.), Hypertension and Iron Chelation in Thalassemia P 11. Berne, Switzerland, Huber. 40. Bickel, H., Hall, G. E., Keller-Schierlein, W., Prelog, V., Vischer, E., Wettstein, A. 1960 Metabolic products of actinomycetes ferrioxamine B. Helv. Chim. Acta. 43:2129-2138. 41. Neilands, J. B. 1981 Iron absorption and transport in microorganisms Ann Rev. Nutr. 1 :27. 42. Neilands, J. B. 1957 Some aspects of microbial iron metabolism. Bacterial. Rev. 21:101. 43. Bergeron, R. J. 1984. Synthesis and solution structure of microbial siderophore. Chem. Rev. 84:587-602. 44. Corey, E. J., and Bhattacharyya, 5. 1977 Total synthesis of enterobactin, a macrocyclic iron transporting agent of bacteria. Tetrahedron Lett. 3919. 45. Drechsel, H., Metzger, J., Freund, 5., Jung, G., Boelaert, J. R., Winkelmann, G. 1991 Rhizoferrin -a novel siderophore from the

PAGE 153

139 fungus Rhizopus microaporus var rhizopodiformis. Biol. Met. 4:238-243. 46. Bergeron, R. J., Weimar, W. R., Dionis, J. B. 1988. Demonstration of ferric L-parabactin-binding activity in the outer membrane of Paracoccus denitrificans. J. Bacterial. 170:3711-3717 47. Fiss, E. H., Samuelson, P. 5., Neiland, J.B. 1982 Properties and proteolysis of ferric enterobactin outer membrane receptor in Escherichia coli K12. Biochemistry. 21:4517-4523. 48 Hollifield, W. C ~ Neilands, J. B. 1978. Ferric enterobactin transport system in Escherichia K12 extraction, assay and specificity of the outer membrane receptor Biochemistry 17: 1922-1928. 49 Kadner, R. J., Lunderigan, M. D., Heller, K. 1987 Sequences and interactions of proteins participating in the transport of iron and vitamin B12 in Escherichia coli. In Winkelman, G., D. van der Helm and Neilands J.B. (eds ), Iron Transport in Microbes, plants and animals. p 85-97. VCH Publishers, New York. 50 Leong, 5. A., Neilands, J. B. 1981. Relationship of siderophore-mediated iron assimilation to virulence in crown gall disease J Bacterial. 147:482-489. 51. Schwyn, 8., Neilands, J. 8. 1987. Universal chemical assay for the detection and determination of siderophores. Anal. Biochem. 160:47-56. 52 Bergeron, R. J., Brittenham G. M. 1994. The Development of Iron Chelators for Clinical Use CRC Press, Inc. Boca Raton 53. Bergeron, R. J., Wiegand, J., Dionis, J. 8., Egli-Karmakka, M., Frei, J., Huxley-Tencer, A., Peter, H. H. 1991. Evaluation of desferrinthiocin and its synthetic analogues as orally effective iron chelators. J Med. Chem 34:2072-2078. 54. Bergeron, R. J., Navratil, J. D 1986. Catecholamide chelators for actinide environmental and human decontamination In C. J King and J T. Navratil (eds.), Chemical separations volume II, Appl i cations Litarvan Literature, Denver, CO 55. Kochan, J. 1977. Role of siderophores in nutritional immunity and bacterial parasitism. In E. D. Weinberg (ed ) Microorganisms and minerals P251-289. Marcel Dekker, Inc. New York.

PAGE 154

140 56 Schade, A. L., Caroline, L. 1944. Raw hen egg white and the role of iron in growth inhibition of Shihella dysenteriae Staphylococcus aurens Escherichia coli and Saccharomyces cerevisiae Science 100:14-15. 57 Haas, 8., Kraut, J., Marks, J., Zanker, 5. C., Castignetti, D. 1991. Siderophore presence in sputa of cystic fibrosis patients Infect. lmmun 59:3997-4000 58. Griffiths, E., Humphreys. 1980 Isolation of enterochelin from the peritoneal washing of guinea pigs lethally infected with Escherichia coli. Infect. lmmun 28:286 289. 59. Yancey, R. J., Breeding, 5. A. L., Lankford, C. E. 1979. Enterochelin (enterobactin): virulence factor for Salmonella typhimurium. Infect. lmmun. 24:174-180 60 Moore, D. G., Yancey, R. J., Lankford, C. E., Earhart, C. F.1 979. Bacteriostatic enterochelin-specific immunoglobulin from normal human serum Infect. lmmun. 27:418-423. 61. Kluger, M. J., Rothenburg, B. A. 1979 Fever and reduced iron: their interaction as a host defense response to bacterial infection. Science 203:374-76 62 Garibaldi, J. A. 1972 Influence of temperature on the biosynthesis of iron transport compounds by Salmonella typhimurium. J Bacterial. 110:262-65. 63 Reiter, 8., Brock, J. H., Creel, E. D 1975 Inh i bition of Escheria coli by bovin colost 11.Teffect of lactoferrin on a serum susceptible and serum resistant strain. Immunology 28:83-95. 64. Bergeron, R. J., Streiff, R. R., King, W., Daniel, R. D. Jr., Wiegand, J. 1993 A comparison of the iron-clearing properties of parabactin and desferrioxamine. Blood 82:2552-2557. 65. Bergeron, R. J., Liu, Z. R., McManis, J. 5., Wiegand J. 1992 Structural alterations i n desferrioxamine compatible with iron clearance in animal. J Med Chem 35:47394744. 66 Bergeron, R. J., Wiegand, J. Dionis, J. 8., Egli-Karmakka, M., Frei, J., H-Tencer, A., Peter, H. H. 1991. Evaluation of desferrithiocin and its synthetic analogues as orally effective iron chelators. J Med. Chem. 34:2072-2078.

PAGE 155

141 67 Miller, K. 8., Rosenwasser, L. J., Bessette, J. A. M., Beer, D. J. Rocklin, R. E. 1981. Rapid desensitization for desferrioxamine anaphylactic reactions Lancet, 1: 1059. 68 Kirking, M. H. 1991. Treatment of chronic iron overload Clin. Pharm. 10:775-783 69. Ponka, P., Borova, J., Neuwirt, J., and Fuchs, 0. 1979. Mobilization of iron from reticulocytes:identification of pyridoxal isonicotinoyl hydrazone as a new iron chelating agent. FESS Lett. 97:317-321. 70. Kontoghiorghes, G. J., Aldouri, M.A., Sheppard, L., Hoffbrand, A. V. 1987. 1,2-Dimethyl-3 hydroxypyrid-4-one an orally active chelator for the treatment of iron overload Lancet 1: 1294-1295. 71. Grady, R. W., Hershko, C. 1990. HBED: A potential oral iron chelator. Ann. N. Y. Acad Sci 612:361-368 72 Naegeli, H. U., Zahner, H. 1980. Metabolites of microorganisms part 193, ferrithiocin. Helv Chim. Acta. 63:1400-1406. 73 Bergeron, R. J., Wiegand, J., Ratliff-Thompson, K., Weimar, W. 1997 The origin of the differences in (R)and (S) desmethyldesferrithiocin iron clearing properties. Ann N Y. Acad. Sci. 718:202-216. 7 4 Tait, G. T 1975 The identification and biosynthesis of siderochemes formed by Micrococcus denitrificans Biochem J. 146:191-204. 75. Griffiths, G. L., Sigel, 5. P., Payne, 5. M., Neilands, J.B. 1984 Vibriobactin, a siderophore from Vibrio cholerae J Biol. Chem 259:383. 76 Peterson, T., Falk, K. E., Leong, 5. A., Klein, M. P., Neilands, J.B. 1980. Structure and behavior of spermidine siderophores J Am. Chem. Soc. 102:7715. 77 Rastetter, W. H., Erikson, T. J., Venuti, M. C. 1981. Synthesis of iron chelators enterobactin, enantioenterobactin and a chiral analogue J. Org. Chem. 46:3579-3590. 78 Bergeron, R. J., Cavanaugh, P. F. Jr., Kline, 5. J., Hughes, R. G. Jr., Elliott, G. T., Porter, C. W. 1984. Antineoplastic and antiherpetic activity of spermidine catecholamide iron chelators

PAGE 156

142 Biochem Biophy. Res. Commun. 121 :848-854. 79 Yamamoto, S., Okujo, N., Fujita, Y., Saito, M., Yoshida, T., Shinoda, S. 1993 Structures of two polyamine-containing catechol siderophores from Vibrio f/uvialis J. Biochem. 113:538-544. 80. Weinberg, E. D. 1998 Patho-ecological implications of microbial acquisition of host iron Rev. Med. Microbial. 9:171-178. 81. Bergeron, R. J., McManis, J. S. 1991 Synthesis of catecholamide and hydroxamate siderophores. In Winkelmann G (ed .), Handbook of Microbial Iron Chelators. p. 271-307 CRC Press, Inc. : Boca Raton Florida 82. Bergeron, R. J., Phanstiel, 0 IV. 1992. The total synthesis of nannochelin : a novel cinnamoyl hydroxamate-containing siderophore J Org Chem. 57:7140-7143. 83 Meiwes, J., Fiedler, H.P., Haag, H., Zahner, H., Konetschny-Rapp, S., Jung, G. 1990 Isolation and characterization of staphyloferrin A a compound with siderophore activity from staphylococcus FEMS Microbial. Lett 67:201-206. 84 Bergeron, R. J., McManis, J. S., Phanstiel, 0. IV., Vinson, T. Jr. 1995 A versatile synthesis of desferrioxamine J Org Chem. 60:109-114 85. Bergeron, R. J., Wollenweber, M., Wiegand, J. 1994. An investigation of desferrithiocin metabolism. J Med. Chem. 37:2889-2895 86. Bergeron, R. J., Kline, S. J. 1982. Short synthesis of parabactin. J Am Chem. Soc. 104:4489-4492. 87 Smith, M. J. 1989. Total synthesis and absolute configuration of rhizobactin, a structurelly novel siderophore Tetrahedron Lett. 30:313-316. 88. Drechsel, H., Jung, G., Winkelmann G. 1992 Stereochemical characterization of rhizoferrin and identification of its dehydration products. Biol. Met. 5:141-148 89. Shenker, M., Ghirlando, R., Oliver, I., Heimann, M., Hadar, Y., Chen, Y. 1995. Chemical structure and biological activity of a siderophore produced by Rhizopus arrhizus Soil. Sci. Soc. Am. J. 59:837-843

PAGE 157

143 90. Bergeron, R. J., MeiGuo, X., Smith, R. E., Wollenweber, M., McManis, J. S., Ludin, C., Abboud, K. A. 1997. Total synthesis of rhizoferrin, an iron chelator. Tetrahedron. 53:427-434. 91. Hirota, K., Kitagawa, H., Jhimamura, M., Ohmori, S. 1980 A facile preparation of asym-monomethyl and asym-dimethyl citrate. Chem. Lett. 191-194. 92. Samejima, K., Takeda, Y., Kawase, M., Okada, M., Kyogoku, Y. 1984. Synthesis of 15N-enriched polyamines. Chem. Pharm Bull. 32:3428-3435. 93. Shioiri, T., Ninomiya, K., Yamada, S. 1972. Diphenylphosphoryl azide: a new convenient reagent for a modified curtius reaction and for the peptide synthesis. J. Am. Chem. Soc. 94:6203-6205. 94 Williams, R. M., Kwast, E. 1989. Carbanion-mediated oxidative deprotection of non-enolizable benzylated amides. Tetrahedron Lett. 30:451-454. 95. Kim, M. Y., Starrett, J. E. Jr., Weinreb, S. M. 1981. Synthetic approach to cytochalasins. J. Org. Chem 46:5383-5389. 96. Bodanszky, M., Martinez, J. 1979 Side reactions in peptide synthesis. 8. On the phenacyl group in the protection of the ~ carboxyl function of aspartyl residues. J. Org. Chem. 43:3071. 97 Jerry, M. 1992. Advanced Organic Chemistry. P 120 John Wiley and sons. New York. 98 Ching-Shih, C., Charles, J. S. 1989. General aspects and optimization of enantioselective biocatalysis in organic solvents: the use of lipases. Angew. Chem. Int. Engl. 28:695-707. 99. Lister, K. P. L., Raymond, A.H., Jones, J.B. 1986. Enzymes in organic synthesis. 35. Stereoselective pig liver esterase catalyzed hydrolyses of 3-substituted glutarate diesters. Optimization of enantiomeric excess via reaction conditions control. J Org. Chem. 51 :2047-2050. 100 Silvia, K. R., Gunther, J., Johannes, M., Hans Z. 1990. Staphyloferrin A: a structurally new siderophore from staphylococci. Eur J. Biochem. 191 :65-74 101. McCord, J. M. 1996. Effects of positive iron status at a cellular

PAGE 158

144 level. Nutr Rev. 54:85-88 102 Munzinger, M., Taraz, K., Budzikiewicz, H., Drechsel, H., Heymann, P., Winkelmann, G., Meyer, J-M. 1999. S S-rhizoferr i n (enantio-rhizoferrin) -a siderophore of Ralstonia (Pseudomonas) pickettii DSM 6297 the optical antipode of R, R-rhizoferrin isolated from fungi. Biometal 12:189-193. 103 Neilands, J. 8., Peterson, T., Leong, S. A. 1980. High-affinity iron transport in microorganisms. ACS Symp Ser. 140:263-278. 104. Martell, A. E., Anderson, W. F., Badman D. G. 1980. Martell A. E., Anderson, W. F., and Badman D. G (eds.) Development of Iron Chelators for Clinical Use. Elsevier/North-Holland, New York. 105 Ong, S A., Peterson, T., Neilands, J. 8. 1979 Agrobactin, a siderophore from Agrobacterium tumefaciens J Biol. Chem. 254: 1860-65. 106 Bergeron, R. J., lngeno, M. J. 1987. Microbial iron chelator induced cell cycle synchronization in L 1210 cells : potential in combination chemotherapy. Cancer Research. 47:6010-6016. 107 Cavanaugh, P. F. Jr., Porter, C. W., Tukalo, D., Frankfurt, 0. S., Pavelic, Z. P., Bergeron, R. J. 1985 Characterization of L 1210 cell growth inhibition by the bacterial iron chelators parabactin and compound II. Cancer Research, 45:4754-4759 108 Raymond, T., Honeysett, J. M., Bergeron, R. J. 1989. Combination iron deplet ion therapy. J. Natl. Cancer Institute 81:1229-1235. 109 Bergeron, R. J., Kline, S. J., Stolowich, N. J., McGovern, K. A. Burton, P. S. 1981. Flexible synthesis of polyamine catecholamides. J Org. Chem 46:4524. 110 Bergeron, R. J., Garlich, J. R. and McManis, J. S. 1985. Total synthesis of vibriobactin. Tetrahedron, 41:507-510 111 Bergeron, R. J., McManis, J. S., Dionis, J. 8., Garlich, J. R. 1985. An efficient total synthesis of agrobactin and its gallium (Ill) chelate. J. Org. Chem. 50:2780-2783. 112 Bergeron, R. J., McGovern, K. A. Channing, M.A., Burton, P. S. 1980 Synthesis of 4N-acylated 1 N, 8N-bis(acyl)spermidines: an approach to the synthesis of siderophores. J. Org. Chem. 45:1589-1592.

PAGE 159

145 113. Bergeron, R. J., Garlich, J. R., Stolowich, N. J. 1982. Reagents for the stepwise functionalization of spermidine homospermidine and bis-(3-aminopropyl)amine J. Org. Chem. 49:2997-3001. 114. Joshua, A V., Scott, J. R. 1984. A simple method for the direct bis acylation of the primary amino groups in spermidine and other linear triamines. Tetrahedron Lett. 25:5725-5728. 115. Lee, B. H., Miller, M. J. 1983. Total synthesis of schizokinen, schizokinen A, and arthrobactin. J Org Chem. 48:24-31. 116. Schreinemakers, F. A.H. 1897. Some amides of aromatic sulfonic acids. Reel. Trav. Chim. Pays-Bas Belg. 16:411-424. 117. Ing, H. R., Manske, R.H. F. 1926. A modification of Gabriel synthesis of amines J Chem. Soc. 2348-2351. 118. Elliott, D. F. 1949 The stereochemistry of an oxazoline derivative of threonine: improvement of a recent threonine synthesis. J. Chem. Soc. 589-594. 119. Bergeron, R. J., Dionis, J. 8., lngeno, M. J. 1987. Synthesis of a parabactin photoaffinity label. J. Org. Chem. 52:144-149. 120. Bergeron, R. J. 1987 Synthesis and properties of polyamine catecholamide chelators. In G. Winklemann D. Vander Helm J. B. Neilands (eds.) Iron Transport in Microbes Plants and Animals. p. 285-315. VCH VerlagsgsellschaftNCH Publishers. 121. Neilands, J.B. 1973. Eichhorn G L. (ed .), Microbial Iron Transport Compounds: Siderophores p. 167-200 Elsevier Scientific Publishing Co. New York. 122. Swinburne, T. R. 1986. Iron Siderophore and Plant Diseases. Nata ASI series Series A: Life Sciences 117, Plenum Press New York. 123. Silver, S. 1978. Rosen, B. P. (ed ) Bacterial Transport p 243-324. Marcel Dekker Inc. New York. 124. Muller, G., Raymond, K. N. 1984 Specificity and mechanism of ferrioxamine-mediated iron transport in Streptomyces pilosus. J. Bacterial. 160:304312.

PAGE 160

146 125. Arceneaux, J. E. L., Davis, W. 8., Downer, D. N., Haydon, A. H., Byers, B. R. 1973. Fate of labeled hydroxamate during iron transport from hydroxamate-iron chelates. J. Bacterial. 115:919. 126 Matzanke, 8., Winkelman, G. 1981 Siderophore iron transport followed by Mossbauer spectroscopy. FESS Lett. 130:50-53. 127. Ratledge, C., Patel, P. V., Mundy, J. 1982. Iron transport in Mycobacterium smegmatis: the location of mycobactin by electron microscopy J. Gen. Microbial. 128:1559-1565. 128 Snow, G. A. 1969. Metal complexes of Mycobactin P and of desferrisideramines. Biochem. J 115: 199-205. 129 Emery, T., Hoffer, P. B. 1980. Siderophore-mediated mechanism of gall i um uptake demonstrated in the microorganism Ustilago sphaerogena. J Nucl. Med 21 :935-939 130. Peterson, T., Neilands, J. 8. 1979. Revised structure of a catecholamide spermidine siderophore from Paracoccus denitrificans Tetrahedron Lett 50:4805-4808. 131 Bergeron, R. J., Dionis, J. B., Elliott, G. T., Kline, S. J. 1985. Mechanism and stereospecificity of the parabactin-mediated iron transport system in Paracoccus denitrificans. J. Biol. Chem. 260:7936-7944 132. Bergeron, R. J., Weimar, W. R. 1990. Kinetics of iron acquisition from ferric siderophores by Paracoccus denitrificans. J. Bacterial. 172:2650-2657. 133. Wiebe, C., Winkelmann, G. 1975 Kinetic studies on the specificity of chelate-iron uptake in Aspergillus J. Bacterial. 123:837-842. 134. Ecker, D. J., Emery, T. 1983. Iron uptake from ferrichrome A and iron citrate in Ustilago sphaerogena J. Bacterial. 155:616-622. 135. Messenger A. J. M., Ratkedge C. 1981. Iron transport in Mycobacterium smegmatis: uptake of iron from ferric citrate J. Bacterial. 149: 131-135 136. Corbin, J. L., Bulen, W. A. 1969. The isolation and identification of 2,3-dihydroxybenzoic acid and 2N,6N-di(2,3-dihydroxybenzoyl)-L lysine formed by iron-deficient Azotobacter vinelandii Biochemistry 8:757-762.

PAGE 161

147 137 Neilands, J. B. 1984. Siderophores of bacteria and fungi. Microbial. Sci. 1 :9-14. 138. Avdeef, A., Sofen, S. R., Bregante, T. L., Raymond, K. N. 1978. Coordination chemistry of microbial iron transport compounds. IX. Stability constants for catechol models of enterobactin. J. Am. Chem. Soc. 100:5362-5370. 139. Guerinot M. L., Yi, Y. 1994 Iron: nutritious, noxious and not readily available. Plant Physiol. 104:815-820. 140. Bezkorovainy, A., Soberg, L., Poch, M., Miller-Catchpole, R. 1987 Ferrous iron uptake by Bifidobacterium bifidum var. pennsy/vanicum: the effect of metals and metabolic inhibitors. Int. J Biochem. 19:517-522. 141. Johnson, W., Varney, L., Poch, M.1991. Acquisition of iron by Legionella pneamophila: role of iron reductase. Infect. lmmun. 59:2376-2381 142. Evans S. L., Arcebeaux J. E. L., Byers, B. R., Martin M. E., Aranha N. 1986. Ferrous iron transport in Streptococcus mutans. J. Bacterial. 168: 1096-1099. 143. Perry, R. D., Brubaker R. P. 1979. Accumulation of iron by yersinia. J. Bacterial. 137:1290-1298. 144 Stoebner, J. A., Payne, S. M. 1988 Iron-regulated hemolysin production and utization of heme and hemoglobin by Vibrio cholerae. Infect. lmmun. 56:2891-2895. 145. Chen, C-Y., Berish, S. A., Morse, S. A., Mietzner, T. A. 1993. The ferric iron-binding protein of pathogenic Neisseria spp. Function as a periplasmic transport protein in iron acquisition from human transferrin. Mal. Microbial. 10:311-318. 146. Pollack, J. R., Ames B. N., Neilands J. B. 1970. Iron transport in Salmonella typhimurium: mutants blocked in the biosynthesis of enterobactin. J. Bacterial. 104:635-639. 147. Decad, G. M., Nikaido, H. 1976. Outer membrane of gram negative bacteria. Molecular-sieving function of cell wall. J. Bacterial. 128:325-336 148 Winkelmann, G. 1979. Evidence for stereospecific uptake of iron chelates in fungi. FEBS Lett. 97:43-46.

PAGE 162

148 149. Frost G. E., Rosenberg H. 1973. The inducible citratedependent iron transport system in Escherichia coli. K 12. Biochim. Biophys Acta 330:90-101. 150 Brown, K. A., Ratledge, C.1974 .. Iron transport in Mycobacterium smegmatis: ferrimycobactin reductase [NAD(P)H: ferrimycobactin oxidoreductase], the enzyme releasing iron from its carrier. FESS Letters 53:262-266. 151. Bullen, J. J., Rogers, H.J., Griffiths E. 1974. In J. B. Neilands (ed.), Microbial iron metabolism Academic Press, Inc. New York. 152. Bergeron, R. J. 1986. Iron: a controlling nutrient in proliferative processes. Trends in Biochem. Sci. 11:133-139. 153. Weinberg, E. 0.1984. Iron withholding: a defense against infection and neoplasia. Physiol. Rev. 64:65-102. 154. Bergeron, R. J., Elliott, G. T., Kline, S. J., Ramphal!, R., and Luther, St. J. 1983 Bacteriostatic and Fungostatic action of catecholamide iron chelators. Antimicrob. Agents Chemother., 24:725-730. 155. Leong, J., Neilands, J.B. 1976. Mechanism of siderophore iron transport in enteric bacteria. J Bacterial. 126:823. 156. Greenwood, K. J., Luke, R. K. J. 1978 Enzymatic hydrolysis of enterochelin and its iron complex in Escherichia coli K12. Biochim Biophys. Acta 525:209 157 Emery, T. 1979. Role of ferrichrome as a ferric ionophore in Ustilago sphaerogena Biochemistry 10:1483 158. Cox, G. 8., Gibson, F., Luke, R. K. J., Neuton, N. A., O'Brien, I. G., Rosenberg, H. 1970. Mutations affecting iron transport in escherichia coli J. Bacterial. 104:219-226. 159. Sabine, K., Volkmar, 8., Wolfgang, K.1996 Ferric rhizoferrin uptake into Morganella morganii characterization of genes involved in the uptake of a polyhydroxy carboxylate siderophore. J. Bacterial. 178:496-504. 1 60. Stephenson, M. C., Ratledge, C. 1979. Iron transport in Mycobacterium smegmatis: uptake of iron from ferriexochelin. J Gen. Microbial. 110:193.

PAGE 163

149 161. Emery, T. 1978. In Sigel, H. (ed.) Metal Ions in Biological System. Vol. 7, pp. 77-126, Marcel Dekker, New York. 162 Guerinot, M. L. 1994 Iron uptake and metabolism in the Rhizobial legume symbioses Plant Soil. 130:199-209. 163. Eide, D., Davis-Kaplan S., Jordan, I., Sipe, D., Kaplan J. 1992. Regulation of iron uptake in Saccharomyces cerevisiae. J Biol. Chem 267:20774-81. 164. Gaines, C. G., J. S. Lodge, J.E. L. Arceneaux, 8. R. Byers. 1981. Ferrisiderophore reductase activity associated with an aromatic biosynthetic enzyme complex in Bacillus subtilis. J Bacteriol. 148:527-533. 165. Braun, V. 1978. Structure-function relationships of the gram negative bacterial cell envelope. Symposia of the Society for General Microbiology. 28: 111-138 166. Lascelles, J., Burke, K. A. 1978 Reduction of ferric iron by L Lactate and DL-Glycerol-3-phosphate in membrane preparations from Staphylococcus aureus and interaction with the nitrate reductase system. J. Bacteriol. 134:585-589. 167 Dailey, H. A., Lascelles, J. 1977 Reduction of iron and synthesis of protoheme by Spirillum itersonii and other organisms J. Bacteriol. 129:815-820. 168. Ratledge, C., Marshall, 8. J. 1972. Iron transport in Mycobacterium smegmatis: the role of mycobactin. Biochim. Biophys. Acta. 279:58-74 169 Emery, T. 1976. Fungal ornithine esterases: relationship to iron transport Biochemistry 15:2723. 170. O'Brien, I. G., Cox, G. 8., Gibson, F.1971 Enterochelin hydrolysis and iron metabolism in Escherichia coli. Biochim. Biophys. Acta 237:537. 171. Moore, R. E., Emery, T.1976. NaAcetylfusarinines: isolation, characterization and properties. Biochemistry 15:2719 172. Venuti, M. C., Rastetter, W. H Neilands, J. 8. 1979 1,3,5Tris(N, N' ,N"-2,3 dihydroxybenzoyl) aminomethylbenzene, a synthetic iron chelator related to enterobactin. J. Med. Chem. 22:123.

PAGE 164

150 173. Harris, W. R., Carrano, C. J., Raymond, K. N. 1979. Coordination chemistry of microbial iron transport compounds. XVI. Isolation, characterization, and formation constants of ferric aerobactin. J Am. Chem. Soc. 101 :2722. 174. Rastetter, W. H., Erikson, T. J., Venuti, M. C. 1981. Synthesis of iron chelators, enterobactin, enantioenterobactin and a chiral analogue. J Org. Chem. 46:3579-3590. 175 Page W. J., Huyer, M. 1984. Derepression of the Azotobacter vinelandii siderophore system, using iron-containing minerals to limit iron repletion. J Bacterial. 158:496-502. 176 Sevinc M. S., Page, W. J. 1992. Generation of Azotobacter vinelandii strains defective in siderophore production and characterization of a strain unable to produce known siderophore. J Gen. Microbial. 138:587-96 177. Archibald, F. S., DeVoe, I. W.1980. Iron acquisition by Neisseria meningitidis in vitro. Infect. lmmun. 27:322-324 178. Spiro, T. G., Pape, L., Saltman, P. 1967. The hydrolytic polymerization of ferric citrate I. The chemistry of the polymer J. Am. Chem. Soc. 89:5555-5559. 179 Spiro, T.G., Bates, G., Saltman, P .. 1967. The hydrolytic polymerization of ferric citrate II. The influence of excess citrate. J. Am. Chem Soc 89:5559-5562. 180 Langman, L., Young, I. G., Frost, G. E., Rosenberg, H., Gibson, F. 1972. Enterochelin system of iron transport in Escherichia coli: mutations affecting ferric-enterochelin esterase. J. Bacterial. 112:1142-1149. 181. Hancock, R. E.W., Hantke, K., Braun, V. 1976. Iron transport in Escherichia coli K12: involvement of the colin B receptor and of a citrate-inducible protein. J. Bacterial. 127: 13701375.

PAGE 165

BIOGRAPHICAL SKETCH Mei Guo Xin was born in TaiHu, Anhui province in the people's Republic of China, on August 08, 1968. In July 1989, he received his Bachelor of Science degree in Food Chemistry from Jiangxi University, in Nangchang, China. Meiguo then entered the graduate program at the TianJin Institute of Light Industry, where in 1992 he received his Master of Science in Chemical Engineering. After graduation, He worked at the Xi'an Institute of Modern Chemistry for two years. In 1995 Mei Guo entered a doctoral program in the University of Florida, Department of Medicinal Chemistry under the supervision of Dr. Raymond Bergeron. 151

PAGE 166

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. /"/ I c_ f )) wt-1 aymond Berg ron, Chainman Graduate Research Professor of Medicinal Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the degree of Doctor of Philosophy Jo(n,Rerrin Professor of Medicinal Chemistry Kenneth Sloan Professor of Medicinal Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Professor of Medicinal Ch I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the de~~ Professor of Pharmacology and Therapeutics

PAGE 167

This dissertation was submitted to the Graduate Faculty of the College of Pharmacy and to the Graduate Scholl and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 2000 Dean, Graduate School