This item is only available as the following downloads:
SYNTHESIS, SOLUTION CONFORMATION, AND MICROBIOLOGICAL
PROPERTIES OF SPERMIDINE CATECHOLAMIDE SIDEROPHORES
STEVEN J. KLINE
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
The patience, understanding, and continual encouragement from my
parents, Mr. and Mrs. Jerrold J. Kline, in large measure account for
this moment in my career. It is to them that I proudly dedicate this
I would like to express my sincere gratitude to my advisor, Dr. Ray-
mond Bergeron, under whose guidance these investigations were accomplished.
I would also like to express appreciation to my supervisory committee,
Dr. Kenneth Sloan, Dr. Margaret James, Dr. William Dolbier, and Dr. Cemal
Kemal. In addition, I would like to express my sincere thanks to Dr.
Richard Streiff for his many helpful discussions.
The actinide absorption studies were done by Dr. James Navratil at
Rocky Flats, Colorado. I wish to extend my appreciation to him for these
determinations, as the data were a significant part of this work.
I would also like to thank my friends and colleagues in the lab
for all of their help and suggestions.
I wish to thank my parents, Jerrold and Ruth, for their continual
love and support. I would also like to express my gratitude to my sis-
ter, Cindy, for her love, understanding, and friendship.
Most of all, I wish to express my deepest appreciation to my fian-
cee, Erica St. Onge, who has stood beside me throughout all of the
crises, both real and imagined, accentuating the positive and always
TABLE OF CONTENTS
ABSTRACT ........................................................ Vi
I INTRODUCTION AND BACKGROUND........................ 1
Iron in Man...................................... 1
Iron in Microorganisms............................ 3
Iron Overload in Man.............................. 7
Potential of Catecholamide Iron Chelators
as Therapeutic Iron Clearing Devices............. 12
II FLEXIBLE SYNTHESIS OF POLYAMINE CATECHOLAMIDES....... 15
Experimental ..................................... 28
III OCTADENTATE CATECHOLATE LIGANDS AS
ACTINIDE CHELATORS ................................ 69
Introduction........... .......................... 69
Synthesis................................... ...... 74
Materials and Methods............................. 76
Results and Discussion............................ 81
IV SYNTHESIS AND SOLUTION CONFORMATION OF PARA-
BACTIN AND ITS GALLIUM(III) CHELATE...............
Experimental ..................................... 90
Synthesis of Parabactin, Enantioparabactin, and
the Home and Nor Homologs of Parabactin.......... 108
V PARABACTIN-MEDIATED IRON TRANSPORT IN PARACOCCUS
Materials and Methods............................. 172
Results ....................................... 177
Discussion. .................................... 196
VI CONCLUSION...................................... 237
BIOGRAPHICAL SKETCH......................................... 242
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, SOLUTION CONFORMATION, AND MICROBIOLOGICAL
PROPERTIES OF SPERMIDINE CATECHOLAMIDE SIDEROPHORES
Steven J. Kline
Chairman: Raymond Bergeron
Major Department: Medicinal Chemistry
The synthesis of a large number of polyamine catecholamides is
described. Among the catecholamides synthesized is the microbial iron
chelator, parabactin, and many of its analogs.
The synthesis of the first of a novel class of actinide ligands is
described as the preliminarywork in the development of a whole class of
octadentate catecholamides. The results indicate the octadentate cate-
cholamide ligands are excellent sequestering agents of the actinides
plutonium and americium.
High field 300 MHz 'H NMR spectroscopy was used to examine the solu-
tion dynamics of parabactin, and the solution conformation and stereo-
chemistry of the gallium(III) chelate of parabactin. Parabactin was
found to exist in solution as a mixture of interconverting conformers.
The gallium(III) chelate of parabactin was shown to exist exclusively as
the A-cis coordination isomer. Furthermore, there exists two diastereo-
meric forms of the A-cis parabactin-Ga(III) chelate that differ only in
the disposition of the spermidine backbone of the siderophore.
The parabactin-mediated iron uptake system of the soil bacterium
Paracoccus denitrificans was examined by using labeled ferric parabactin
and other analogs including its enantiomer, ferric enantioparabactin.
Parabactin functions as an "iron taxi" in the delivery of iron to the
microorganism, with release of siderophore-bound iron presumably taking
place at the outer cell surface and the ligand never penetrating the
organism. The parabactin-mediated iron transport system of Paracoccus
denitrificans was also shown to be stereospecific. The synthetic enan-
tiomer of parabactin, enantioparabactin, was unable to supply iron to the
INTRODUCTION AND BACKGROUND
Iron in Man
Iron is an absolutely essential element for almost all forms of
living organisms (1-4). The reason for the importance of this metal is
related to its versatility as a biological catalyst. In humans, iron
is found at the active centers of the biomolecules responsible for oxy-
gen storage and transport, myoglobin and hemoglobin (5). Iron is also
present in the heme-containing proteins that function as electron car-
riers, the cytochromes (6). The metal is also found in a large number
of nonheme-containing enzymes including various oxidoreductases, dehy-
drogenases, and dehydrases (6).
In man, the majority of the body's iron is bound to the porphyrin
rings of hemoglobin. This heme-containing protein found in erythrocytes
is responsible for picking up oxygen in the lungs and carrying it to
tissues via the circulatory system. Approximately 65% of the 4 g of
iron present in a 75 Kg man is found as hemoglobin (5). Another 6% of
the body's iron exists in the form of another oxygen-binding protein,
myoglobin, which acts as a cellular oxygen storage site (5).
About 25% of the iron present in man is bound to the two iron-
storage proteins, ferritin and hemosiderin (5). The iron-storage pro-
teins provide the body with a means of storing surplus iron in a non-
toxic form that can be released as required. Ferritin is thought to
consist of a protein shell surrounding an internal core of ferric
hydroxyphosphate (7). The exact structure of hemosiderin is unknown,
but is thought to be similar to ferritin. It is believed that up to
5000 iron atoms occupy a single molecule of ferritin or hemosiderin (7).
Although the iron-binding storage proteins are widely distributed through-
out the body in mammals, these proteins are mainly found in the paren-
chymal cells of the liver, the reticuloendothelial cells in the spleen,
and the bone marrow (7).
At physiological pH in an oxidizing environment such as human blood,
the predominant oxidation state of iron is the iron(III) state. Since
ferric hydroxide is quite insoluble (Ksp=10-8) (8), it would seem that
the body is faced with a difficult task of circulating this essential,
but very insoluble metal throughout the system. The problem of low
solubility is circumvented by the use of an iron-binding transport pro-
tein that shuttles the metal from the iron storage sites to the various
tissues. The iron shuttle protein, transferring, has a molecular weight
of about 80,000 daltons, and binds two molecules of ferric ion tightly
but reversibly (3,7,9). Transferrin is found in the serum and is also
present in the various extracellular fluids in the body (7). Because
transferring passes freely out of the blood vessels into the extravascu-
lar fluids, it has the ability to deposit or remove iron from almost
any body tissue.
Iron enters the body by absorption from food in the small intes-
tine (7). The heme-iron containing foods such as red meat are the most
effective sources of dietary iron (7). However, only a small proportion
of the iron in food is absorbed by the intestine; the usual intake of
iron is less than 1 mg per day (7). Despite the exceptionally low
dietary intake of iron in humans, iron balance is maintained due to the
extremely efficient recycling of the iron already present in the body.
Very little of the body's iron store is excreted, less than 1 mg per day;
the remaining 4000 mg of iron are simply recycled (7).
The red blood cells contain the vast majority of the iron present
in humans (5). Normally, red blood cells survive for about 120 days in
the bloodstream. Senescent red blood cells are removed from the circu-
lation by the reticuloendothelial cells in the spleen, and the iron is
removed from hemoglobin to be recycled. This iron can then be stored
in the reticuloendothelial cells, but is more frequently picked up by
transferring and shuttled to the erythropoietic cells of the bone marrow
(5). Transferrin also shuttles iron from the main iron storage site,
the liver, and from the intestine where the metal is absorbed, to the
bone marrow where the iron is incorporated into the hemoglobin of de-
veloping red blood cells.
As can be seen, the body possesses a highly regulated system to
deal with the storage and transport of iron. The danger of toxicity
resulting from storing large amounts of ferric ion is overcome by stor-
ing the metal as an insoluble hydroxyphosphate complex that is buried
within a protein sheath, in the form of ferritin. The free metal is
too insoluble to freely travel through the blood and other body fluids
to the sites where it is needed. This problem is dealt with by having
the iron-transport protein transferring chelate and solubilize ferric
ion and transport it to the tissues.
Iron in Microorganisms
In addition to being an essential nutrient to humans, iron also
plays a major role in the microbiological world. The involvement of
iron in the cytochrome and nonheme emzymes in the respiratory chain
of aerobic and facultative anaerobes underscores the importance of this
element to energy metabolism in microorganisms (2,3). Iron is also in-
volved in the hydroperoxidases, catalyses, and peroxidases in addition
to playing a role in DNA biosynthesis (2).
Iron is the second most abundant metal found on the earth's sur-
face, and is only outranked by aluminum, silicon, and oxygen in terms
of the total amount of any element found in the earth's crust (2, 10).
It has been reported that iron represents about 5% of an average soil
sample (2). Considering that this essential metal is so prevalent in
the environment, it is a paradox that it is exceedingly difficult for
the microbes to acquire iron. The reason for this apparent contradic-
tion is directly related to the extreme insolubility of iron at pH
values near neutrality. In aqueous solutions at or above neutrality,
ferric ion is most commonly found as its hydroxide. Due to the extreme
insolubility of ferric hydroxide, the equilibrium concentration of fer-
ric ion at a pH of 7 is only about 10-18 M (11). The concentration of
iron required to support microbial growth lies in the range of 5 x 10"-
to 1 x 10 6 M, at least ten powers of ten higher than the concentration
of available soluble iron (12).
However, the microbes have dealt with this apparently insurmount-
able problem very effectively over the millennium. Microorganisms pro-
duce and excrete into the external cell environment lowmolecularweight
(500 to 1000 daltons), virtually ferric ion specific ligands that se-
quester exogenous ferric ion and facilitate the transport of the biolo-
gically essential metal into the cell. These microbial iron chelators
are referred to as siderophores (2,10-13), a term derived from the greek
word meaning iron-carrying. Because these compounds are responsible
for iron acquisition, the biosynthesis of the siderophores is strictly
controlled by the iron nutritional status of the cell (2). In condi-
tions of low iron availability, which normally prevail in the environ-
ment, microorganisms produce and excrete sometimes very large amounts
of the siderophores into the surrounding medium. For example, the yeast
Rhodotorula pilimanae excretes up to 10 g/l of the siderophore rhodo-
torulic acid (Figure 1) (11).
Two main classes of siderophores exist, the hydroxamates and the
catecholamides (2,3,10). To date, a far greater number of hydroxamic
acid-containing siderophores have been isolated and characterized than
catecholamides. Fungi generally seem to produce siderophores of the
hydroxamate variety, while bacteria are known to produce both hydroxa-
mate as well as catecholate siderophores (2). The vast majority of
siderophores, both hydroxamate and catecholamide, are hexadentate
ligands. The hexadentate siderophores contain three sets of bidentate
ligands, each capable of chelating the metal in the form of a five-
membered ring (14). These microbial iron chelating agents form very
stable octahedral, high-spin iron(III) complexes with formation con-
stants on the order of 10"0 to 10s5 (2,9-11).
The prototype of the catecholamide siderophores is enterobactin,
the cyclic trimer of 2,3-dihydroxybenzoyl serine, found in all enteric
bacteria studied to date (Figure 1) (15,16). Enterobactin is the
strongest ferric ion chelating agent known; at physiological pH, the
iron(III)-enterobactin formation constant has been calculated to be
10s2 (17). An interesting way to look at what this unimaginable num-
ber means has been recently presented. It has been calculated that at
pH 7, the concentration of unchelated hexaaquoiron(III) in an aqueous
CH3C-N H 11 3
0 HO 0
Figure 1. Structures of the Siderophores a) Enterobactin, and
b) Rhodotorulic acid.
solution which contains 10 pM enterobactin and 1 pM iron is 10-30 pM
(18). This means there would exist only one unchelated ferric ion in
every 1012 liters of water (12).
Much less is known about the catecholamide siderophore parabactin,
first isolated by Tait in 1975 from iron-depressed cultures of the soil
bacterium Paracoccus denitrificans (Figure 2) (19). Parabactin was the
first spermidine catecholamide siderophore isolated. Shortly after the
discovery of parabactin, another catecholamide siderophore containing a
spermidine backbone, agrobactin, was isolated from the plant pathogen
Agrobacterium tumefaciens (20). Tait had originally isolated three
iron-binding catechol containing compounds from cultures of Paracoccus
denitrificans, which he referred to as "Compounds I, II, and III" (19).
Tait correctly determined the structures of compounds I and II to be
2,3-dihydroxybenzoic acid and N',N'-bis(2,3-dihydroxybenzoyl)spermidine,
respectively. Compound III was reported by Tait to have the structure
depicted in Figure 3. However, later it was shown that Tait's original
structure elucidation of compound III was incorrect; the siderophore's
true structure is shown in Figure 2 (21). Tait was also able to show
that the microorganism synthesized compound II from compound I, and com-
pound III from compound II, L-threonine, and salicylic acid (19).
Iron Overload in Man
As mentioned earlier, iron is conserved in the body (7). Less
than 1 mg from about 4 g of iron present in humans is lost from the
body each day (5,7). This small amount of iron that is lost is due to
the sloughing off of the intestinal epithelium rather than from a true
excretion mechanism (7). In fact, no major excretory pathway for iron
exists in man (22). Since normal individuals absorb only minute quan-
tities of iron from food and water each day, this economical policy of
HO CH3 CO
CONH N NHC
Figure 2. Structures of the Spermidine Catecholamide Siderophores
Parabactin (R=H), and Agrobactin (R=OH).
HO HO CO OH
CONHCH2C CH2CH 2CH2CH2CHHCHNHCO
Figure 3. Structures of the Catechol Compounds Reported by Tait
as a) "Compound I", b) "Compound II", and
c) "Compound III".
iron regulation is very beneficial for maintaining adequate amounts of
this vital element. However, in certain disease states this conserva-
tion of iron proves to be a liability. Abnormal increases in either
oral or parenteral iron cannot be balanced by corresponding losses since
no major excretory pathway for iron exists in humans. Thus, the inevit-
able result is an increase in the total body iron content.
In certain syndromes, iron accumulates in the body in such large
quantities that tissue damage results (23). These disorders are referred
to as the primary and secondary hemochromatoses; they are somewhat rare
disorders, and are very often fatal (23). The disorder classified as
primary hemochromatosis is due to an inherited defect in iron metabolism
resulting in an inappropriately increased mucosal absorption of iron
(23). Primary hemochromatosis is easily treated by periodic venesection;
usually, about 500 ml of blood are removed weekly (23).
The secondary hemochromatoses pose a substantially more complicated
problem (23). The term secondary hemochromatosis is most often used in
reference to patients with iron overload secondary to anemia. Patients
suffering from severe forms of anemia require repeated blood transfusions
for survival (23). As the normal lifespan of a red blood cell is only
about 120 days, the iron from the old transfused red cells is eventually
removed by the reticuloendothelial cells of the spleen and stored in
the parenchymal cells of the liver or in the reticuloendothelium (5,24).
Unfortunately, the efficient iron-recycling system of the body ensures
that little of the iron introduced into the body via the transfused red
cells is ever excreted (24). The result is an increase in the total
amount of iron stored in the body, which eventually reaches toxic levels.
This excess iron is initially accumulated in the liver and spleen, and
eventually is deposited in the myocardium and endocrine organs; death
usually results from cardiac hemosiderosis (22,23). The most common
form of this type of anemia is B-thalassemia (22). This is an inherited
disorder in which patients are unable to properly manufacture their own
hemoglobin, and therefore require repeated blood transfusions (23).
These iron-loading anemias, as they are referred to, are quite fatal
(23). Children with B-thalassemia can usually be kept in good health
through the first decade of life with regular blood transfusions (22).
However, the vast majority of them die from the toxic effects of iron
overload in the second or third decade (23).
Unlike primary hemochromatoses, the iron-loading anemias are much
more difficult to treat. Obviously, this excess iron cannot be removed
by phlebotomy in the case of secondary hemochromatosis; the patients
are anemic to begin with. Therefore, other means of removing the large
amounts of the metal are needed. To date, the most effective means of
promoting the excretion of iron from patients suffering from the iron-
loading anemias is chelation therapy. The iron chelating agent desfer-
rioxamine (DFO) is currently the most widely employed therapeutic agent
for the treatment of secondary hemochromatoses (5,22). This iron che-
lating agent is a siderophore isolated from Streptomyces pilosus (23).
It has been shown that DFO is able to promote significant increases in
urinary iron excretion in iron overloaded patients (23).
Although DFO can promote the excretion of a small amount of the
excess iron present in thalassemic patients, there are several drawbacks
to the use of DFO in iron chelation therapy. Firstly, DFO is not orally
effective. In addition, since the chelator has such a short half-life
of clearance from the blood (5-10 minutes), it must be administered as
a subcutaneous or intravenous infusion typically over a 12 hour period,
six days a week (23,25). This poses a major inconvenience at the very
least to patients who must receive this drug chronically. In addition,
extensive DFO chelation therapy has not yet been shown to prolong the
life of, or even prevent cardiac disease in, thalassemic patients (22).
Intensive DFO chelation therapy is also quite expensive. Unfortunately,
countries that have the highest incidence of thalassemia are least likely
to have patients that can afford this type of treatment. There is clear-
ly the need for a less expensive, more convenient therapeutic device
that could be used to treat the iron-loading anemias.
Potential of Catecholamide Iron Chelators as
Therapeutic Iron Clearing Devices
Shortly after the isolation of the spermidine catecholamides from
Paracoccus denitrificans, workers began to examine the possibility of
using these compounds in iron chelation therapy (26-28). Most of the
preliminary studies with the catecholamides had been quite promising.
One of the earliest studies was conducted by Jacobs and coworkers,
in which the ability of Tait's Compound II and Compound III (parabactin)
to remove iron from human transferring in vitro was examined (26). Al-
though the total amount of the body's supply of iron that exists bound
to transferring at any one time is less than 1%, this shuttle protein
represents a natural target for iron chelation therapy. Transferrin is
most prevalent in the serum; however, the protein has an almost unlimited
access to all of the various tissues in the body. It is continuously
supplied with iron from the reticuloendothelium and the parenchymal
cells of the liver, the two main iron storage sites in the body. There-
fore, even if an iron chelator were unable to access the storage iron
directly, it would nevertheless come into contact with transferring as
long as the chelator could be absorbed into the serum (29). If an iron
chelating agent could remove iron from transferring, the protein would
be resupplied with iron from the stores, and eventually the storage iron
would be depleted (29). In this way, an iron chelating device could
remove iron from storage sites indirectly, without coming in contact
with the iron storage tissues directly.
Jacobs and coworkers were able to show that both compound II and
parabactin were better than DFO in removing "Fe from labeled human
transferring (26). With the chelators present at a concentration of 1 nMi
it was found that after 6 hours the amount of iron removed from serum
containing transferring by compound II, parabactin, and DFO were 36%,
18%, and 5% respectively (26). These initial results were very encour-
aging in showing the potential of the catecholamides as iron clearing
Chang cells, cultured liver cells, were used as an in vitro model
for studying the relative abilities of compound II, parabactin, and DFO
to act as iron clearing devices (27). Chang cells take up iron and
incorporate it into ferritin, just as the liver parenchymal cells do
in vivo (27). Since the liver parenchyma constitutes the largest store
of iron in the body and is the site of considerable tissue damage in
hemochromatosis, a chelator that could remove iron from the liver would
be of considerable importance. The abilities of compound II, parabactin,
and DFO to inhibit "Fe incorporation into ferritin, to inhibit "5Fe
uptake, and to inhibit ferritin synthesis were studied (27). All three
of these parameters were decreased in the presence of the chelators,
but compound II and parabactin reduced cellular iron uptake considerably
better than DFO (27). Compound II also proved to be better than DFO at
inhibiting ferritin synthesis in Chang cells (27). Chang cells that
were prelabeled with "5Fe were used to determine if the chelators could
diffuse into and out of the cells (27). Both parabactin and compound II
were significantly better than DFO at removing the "5Fe from the cells
(27). This result reinforced the idea that these catecholamides may
be useful iron-clearing devices.
The ability to remove iron from animals is a crucial test of pro-
posed iron-clearing compounds. One of the methods for determining this
ability is the iron-overloaded rat model of hemochromatosis (28). In
this assay, rats are administered large quantities of iron labeled with
"9Fe, and the "5Fe excretion in the urine and feces is monitored (28).
Several of the catecholamide siderophores have been tested with this
model, both orally and parenterally (28). It was found that compound
II was as effective as DFO in increasing the total iron excreted (28).
In addition, while oral administration of DFO had no effect on iron
excretion, compound II given orally caused a significant increase in the
urinary iron output (28).
In these studies catecholamide iron chelators proved to be more ef-
fective than DFO in removing iron from transferin, from Chang cells, and
from rats. These initial findings prompted us to pursue a program of
synthesizing catecholamide iron chelators for possible use as thera-
peutic iron chelating devices.
FLEXIBLE SYNTHESIS OF
In recent years there has been a great deal of effort focused on
the synthesis of polyamine catecholamide iron chelators (30-41). There
are many reasons for the stimulation of interest in this particular
area. First, the polyamine catecholamides are closely related struc-
turally to microbial iron chelators, the siderophores (3,13). The syn-
thetic siderophore analogs have proven to be useful tools in the study
of the mechanisms of siderophore-mediated iron uptake in microorgan-
isms (12). By preparing a series of siderophore analogs whose struc-
tures differ slightly from one another, the sensitivity of the microbial
iron uptake system to changes in the siderophore's structure can be
There also exist more practical applications of the synthetic poly-
amine catecholamides. The lack of an effective iron clearing device to
treat the iron overload syndromes is probably responsible for the vast
majority of recent interest in polyamine catecholamide iron chelators
(42). Several catecholamide iron chelating agents have been prepared
during the past few years that show some potential to be useful thera-
peutic devices for iron chelation therapy (26,28,31).
Another application of synthetic catecholamide compounds concerns
the need for the development of ligands that have the ability to prefer-
entially chelate actinides for example, plutonium (43-48). The
current devices used to treat individuals that have been exposed to the
transuranium elements are not very effective at removing the contaminat-
ing metal from the body without also promoting the excretion of other
essential metals, such as zinc (46).
Two main classes of catecholamide siderophores are known. The first
class is really only comprised of a single compound, enterobactin (Figure
1) (15,16). The first total synthesis of enterobactin was reported by
Corey in 1977 (49). Since that time a large number of enterobactin ana-
logs have been synthesized, all of which essentially replace the labile
cyclic triester platform of the siderophore with some other less labile
The other main class of catecholamide siderophores are the spermi-
dine catecholamide siderophores parabactin and agrobactin (19,20) (Fig-
ure 2). Soon after the isolation of parabactin in 1975, several groups
began programs aimed at completing the total synthesis of this micro-
bial iron chelator, as well as synthesizing a number of its analogs
(33,40,41). So far, Raymond has synthesized a number of catecholamides
that are similar to parabactin and agrobactin in that they have the
linear triamine spermidine as their backbone (40,47,50). These linear
catecholamides are all essentially variations of one simple synthetic
scheme to acylate all three of the amine groups of spermidine with
the same 2,3-dihydroxybenzoic acid derivative. Unlike these synthetic
catecholamides which contain three identical acyl groups at each of the
three spermidine nitrogens, parabactin contains a different substituent
at the N4 position of its spermidine backbone other than the two 2,3-
dihydroxybenzoyl groups at the terminal nitrogens.
Until only recently, the total synthesis of parabactin had eluded
researchers working in the field (32). The reason for the difficulty
in preparing this siderophore is not immediately apparent when examin-
ing the structure of the compound. Figure 4 shows the most obvious syn-
thetic disconnections that would be made in designing the synthesis of
parabactin. Figure 5 shows the resulting products of these synthetic
disconnections. All of the compoundsin Figure 5 are commercially avail-
able, and the process of simply linking these units together might seem
to be a trivial task at first thought. However, a major stumbling block
exists in the synthesis of parabactin that has hindered the successful syn-
thesis of this compound for years the selective acylation of spermidine.
Clearly, an attempt at the selective acylation of the terminal amino
groups of spermidine with an appropriately protected 2,3-dihydroxyben-
zoic acid derivative would result in a mixture of mono, di- and tri-
acylated adducts. Nevertheless, such fatuous attempts at achieving
N',N'-bis-acylation of spermidine are surprisingly frequent in the
literature. For example, attempted NI,Ne-bis-acylation of spermidine
with 2,3-dihydroxybenzoic acid derivatives result in yields typically
around 14%, with the isolation of desired product from complex mixtures
requiring time-consuming procedures (52).
Other workers have attempted to improve on N',Ne-bis-acylation of
spermidine by employing bulky, sterically hindered acylating agents.
However, such approaches have not resulted in a great deal of success.
For example, it was recently attempted to prepare the N,N'-bis-acyl
spermidine adduct of 2,3-bis(benzoyloxy)benzoic acid and spermidine
(41). The above condensation provided a mixture of products that re-
quired chromatography and offered the desired product in only "about
50% yield" (41).
Figure 4. Obvious Synthetic Disconnections of Parabactin.
Figure 5. Compounds Resulting from Disconnections in Figure 4.
The above examples clearly underscored the necessity of developing
a synthetic scheme that could allow for the selective N',N8-bis-acyla-
tion of spermidine. Surprisingly enough, there are only three reagents
available to allow for the selective acylation of spermidine. The first
of these synthetic schemes offers N ,N'-di-t-butoxycarbonyl spermidine
as a reagent for introducing an acyl group at the N' position of sper-
midine (53). The reagent is available in three steps in a 49% yield,
but is only useful in instances where it is not necessary to differen-
tiate between the nitrogens at N4 and N'.
Another reagent has recently been prepared that has two different
protecting groups attached to the N4 and No nitrogens of spermidine,
N'-tosyl-NO-phthaloyl spermidine (54). One could theoretically employ
this reagent to selectively effect the acylation of the terminal nitro-
gens of spermidine in route to the synthesis of parabactin and its
analogs. However, the eight steps required for the synthesis of this
protected spermidine, and the conditions required for the removal of
the protecting groups, do not render this reagent particularly suitable
for the synthesis of parabactin.
The most recent of the three methods takes advantage of a transient-
ly protected spermidine. The condensation of spermidine with formalde-
hyde produces l-(4-aminobutyl)hexahydropyrimidine which can then be
bis-acylated (55). Deprotection involves cleavage of the hexahydropyri-
midine ring, which affords the N',Ne-bis-acyl spermidine. Although this
technique would be the most favorable of the three in terms of being
applicable to the synthesis of spermidine siderophores such as parabac-
tin, it suffers the drawbacks of not being able to attain bis-acyl
norspermidine derivatives as well as requiring somewhat harsh conditions
to effect the ring opening.
It was clear that there existed a need to develop a reagent that
would allow for the selective bis-acylation of the primary amino groups
of spermidine while at the same time keeping the secondary nitrogen
protected. The development of such a reagent that allowed for the se-
lective NI,N'-bis-acylation of spermidine, and its homo and nor homo-
logs, is discussed in this chapter. Using this protected spermidine,
a large number of polyamine catecholamides were synthesized.
The boundary conditions set for developing a reagent that would
enable the selective N',N'-bis-acylation of spermidine required that
the scheme be composed of a short number of steps which proceed in high
yield from relatively inexpensive starting materials. Additionally, it
was required that the eventual deprotection of the secondary amine
could be effected under mild conditions so that there would be minimal
restrictions on what acyl groups could be attached to the N' and NB
positions of the spermidine derivative. Another condition of develop-
ing an N'-blocked spermidine was that the scheme must be applicable
to the homo and nor homologs of spermidine.
Perhaps the most obvious approach to developing a protected sper-
midine would be to begin with spermidine itself and try to selectively
introduce the protecting group to the N4 position. However, using this
approach there exists the original problem of selectively introducing
a substituent to one of the amino groups of spermidine, in this case
a protecting group at N'. The approach that was taken was to develop
a reagent that would allow for the differentiation of the primary and
secondary amine groups of spermidine by having the secondary nitrogen
protecting group "built in" to the molecule. The reagent decided on
was the N -benzyl derivative of spermidine (33).
This simple scheme begins with the inexpensive reagents benzylamine
and acrylonitrile in a cyanoethylation reaction to afford 2-cyanoethyl
benzylamine in high yield (Figure 6). This amine is then alkylated with
4-chlorobutyronitrile in refluxing butanol using potassium carbonate
as the base in high yield. The cyano groups of the resulting di-nitrile
are then reduced using lithium aluminum hydride in the presence of
aluminum trichloride in diethylether, again in good yield. Recently,
it was found that by using W-2 Raney nickel as the catalyst it was pos-
sible to reduce the cyano groups of the di-nitrile to amino groups via
catalytic hydrogenation without debenzylating the secondary nitrogen.
This modification of the above scheme for obtaining N4-benzylspermidine
has greatly enhanced the synthetic route since the reduction now pro-
ceeds quantitatively and eliminates the elaborate workup conditions re-
quired when using lithium aluminum hydride.
One of the major advantages of the above scheme is its ability to
be extended to the homo and nor homologs of N'-benzylspermidine (Figure
6). Benzylamine is used as the starting material for the preparation of
all three benzylspermidine derivatives. By simply altering the reaction
conditions of the cyanoethylation of benzylamine, bis-cyanoethylation
can be effected. Thus, heating benzylamine and excess acrylonitrile
in a sealed tube for several days in the presence of hydroquinone yields
the bis-cyanoethylation adduct, bis-N-(3-cyanopropyl)benzylamine, in a
95% yield. In the absence of hydroquinone typical yields were only
15-30% with large amounts of polymeric side products. The scheme is
also easily extended to yield the homo derivative of N4-benzylspermi-
dine. Di-alkylation of benzylamine can be effected in a 70% yield with
4-chlorobutyronitrile. Reduction of the homo or nor bis-nitriles is
then conducted in the usual manner providing the homo and nor homologs
of benzylspermidine. Using the above scheme the selective acylation
of the primary amino groups of spermidine as well as its homo and
nor homologs can be accomplished.
Each of the secondary N-benzylated amines of Figure 6 may now be
bis-acylated with an acylating agent. An excess of 2,3-dimethoxyben-
zoyl chloride was used to acylate benzylspermidine, and its homologs
in methylene chloride using triethylamine as a base to produce the
bis-catecholamides in excess of 95% yield. The excess acyl chloride
is removed by adding the acyl halide scavenger 3-(dimethylamino)propyl-
amine to the reaction mixture prior to an acid wash. The bis-acyl
spermidine remains in the organic phase when the reaction mixture is
then washed with dilute hydrochloric acid, while the adduct formed be-
tween the acyl chloride scavenger and 2,3-dimethoxybenzoyl chloride,
as well as triethylamine, extracts into the aqueous phase. In this
manner, the bis-acylation proceeds at near quantitative yields with no
need for chromatography of the bis-adduct. The terminal bis-acyl ad-
ducts are then quantitatively debenzylated in acetic acid over a pal-
ladium chloride catalyst, the product of hydrogenolysis not requiring
a chromatography step.
Thus, the bis-acyl spermidine, N',N'-bis(2,3-dimethoxybenzoyl)-
spermidine, and its homo and nor homologs can now be used to attach
any of a large number of acyl groups to the compound's secondary nitro-
gen. In fact, NI,N'-bis(2,3-dimethoxybenzoyl)spermidine has been a
critical reagent for the synthesis of a large number of catecholamides
including parabactin analogs, the siderophore parabactin itself, and
was also used to prepare the first of a new class of octadentate cate-
In the synthesis of the parabactin analogs, the secondary nitrogen
N',N'-bis(2,3-dimethoxybenzoyl)spermidine was acylated with either 2-
hydroxyhippuric acid, N-(2,3-dimethoxybenzoyl)glycine, N-(2,3-dimethoxy-
benzoyl)-4-aminobutyric acid or N-(2,3-dimethoxybenzoyl)-a-alanine. While
2-hydroxyhippuric acid is commercially available, the other N-acyl amino
acids had to be synthesized. All three were prepared by reacting the N-
hydroxysuccinimidyl active ester of 2,3-dimethoxybenzoic acid with the
appropriate amino acid (Figure 37). The active ester was generated in
dioxane at 150C by coupling 2,3-dimethoxybenzoic acid with N-hydroxy-
succinimide using dicyclohexylcarbodiimide (DCC). The dicyclohexylurea
was removed by filtration and an aqueous bicarbonate solution of the
amino acid was added to the filtrate. In this manner the three N-(2,3
dimethoxybenzoyl)amino acids were obtained in 80-90% yields. These
three acids were used to acylate the secondary nitrogen of N',N'-bis-
(2,3-dimethoxybenzoy1)spermidine and its homologs, using DCC as the con-
densing agent (Figure 38).
The only N-acyl amino acid that was not attached to the bis-acyl
spermidines via DCC was 2-hydroxyhippuric acid. This compound was
activated with trifluoroacetic anhydride. It has been shown that
when 2-hydroxyhippuric acid is reacted with trifluoroacetic anhydride,
the active acylating agent is not the mixed anhydride but rather, 2-
(2-trifluoroacetoxyphenyl)-5-oxazolone (30,34). It is likely that
this oxazolone is generated through the mixed anhydride.
The last step of the synthesis involved deprotection of the cate-
chol protecting groups using boron tribromide (BBr3) in methylene chloride.
It was found that the dimethoxy protecting group was an ideal mask-
ing group for the catechols in this synthesis. Not only was
OCH3 2 DCC |n C 1, 2. 3
OCH 3 004
3 H2N(CH2) -CO2H/NaHCO3
Figure 7. Synthesis of N-(2,3-Dimethoxybenzoyl)amino Acids.
NH,-- (CH2)-N --(CH2)b-- NH
0r,0CH 2 CH303
CONH-- (CH2 )'---(CH)b--NH O
O.-OCH: CH O^
ONH-(cH2)- NN -(CH b
o (OH NN H 0 .
0cCH: NH c
NH- --(CH )--k-(CH ) b-NHT
Figure 8. Overview of Flexible Synthesis of Polyamine
the starting material commercially available, therefore eliminating the
need of protecting the catechol group of 2,3-dihydroxybenzoic acid, but
the 0-methyl groups are taken off quantitatively using BBr3. It was
found that the acidic conditions involved in the hydrolytic workup of
the BBr3 reaction did not adversely affect the catecholamides.
All reagents were purchased from Aldrich Chemical Co. and were
used without further purification. Unless otherwise specified, sodium
sulfate was used as a drying agent. Melting points were taken on a
Fisher-Johns apparatus and are uncorrected. Preparative thin-layer
chromatography was done on 20 x 20 cm silica gel plates obtained from
Analtech Co. Sephadex LH-20 was purchased from Pharmacia Fine Chemicals
Co. Elemental analyses were performed by Galbraith Laboratories, Knox-
ville, Tenn. Samples for 'H NMR were prepared in d-chloroform; chemi-
cal shifts are given in parts per million relative to an internal te-
tramethylsilane standard unless stated otherwise. The spectra were re-
corded on a Varian T-60 spectrometer. Samples for IR spectra were pre-
pared in potassium bromide unless stated otherwise and the spectra were
recorded on a Beckman IR 4210 spectrophotometer.
A solution of 4-chlorobutyronitrile (19.30 g, 0.186 mol) in 100
ml dry 1-butanol was added dropwise over 2 hours to a stirred mixture
of benzylamine (10.30 g, 96 mmol), anhydrous sodium carbonate (30.48 g,
288 mmol) and potassium iodide (5.66 g, 34 mmol) at 1150C. After re-
fluxing an additional 20 hours under a nitrogen atmosphere, the mixture
was allowed to cool to room temperature and was filtered; then, the
salts were washed well with diethyl ether. The combined filtrate and
washings were extracted 3 x 100 ml with 3N hydrochloric acid and 2 x
100 ml water. The acid and water extracts were combined and washed 2
x 100 ml with ether, made basic with sodium carbonate and extracted
3 x 100 ml with ether. The resulting ethereal solution was dried, fil-
tered, evaporated and distilled to give (1): 15.71 g (70%); b.p. 166*
(0.05 mm). An analytical sample was purified on silica gel, using 25%
petroleum ether in diethyl ether as the eluent. 'H NMR: 6 1.45-2.08
(m, 4H), 2.10-2.77 (m, 8H), 3.50 (s, 2H), 7.22 (s, 5H).
Anal. calcd. for C15H,9N3: C, 74.65; H, 7.94; N, 17.41. Found:
C, 74.83; H, 8.06; N, 17.45.
A solution of aluminum chloride (5.05 g, 38 mmol) in 100 ml anhy-
drous diethyl ether was added to lithium aluminum hydride (1.44 g,
38 mmol) in 300 ml anhydrous diethyl ether. The mixture was stirred
under a nitrogen atmosphere for 15 minutes followed by the dropwise
addition of (1) (4.16 g, 17 mnol) in 50 ml anhydrous diethyl ether.
After stirring the reaction mixture an additional 15 hours, it was cooled
to 00C and quenched with 200 ml aqueous 30% potassium hydroxide (w/v).
The contents of the reaction vessel were loaded into a continuous ex-
traction apparatus and extracted with diethyl ether for 48 hours. The
resulting solution was dried, filtered, evaporated, and distilled:
yield: 2.79 g (65%); b.p. 185C (0.075 mm). 'H NMR: 6 1.12 (s, 4H),
1.23-1.63 (m, 8H), 1.93-2.74 (m, 8H), 3.40 (s, 2H), 7.10 (s, 5H).
Anal. calcd. for Ci5H27N3: C, 72.24; H, 10.91; N, 16.85. Found:
C, 72.15; H, 11.10; N, 16.76.
N -Benzyl-N',N8-bis(2,3-methylenedioxybenzoyl)spermidine Hydrochloride(3)
A solution of N -benzylspermidine (33) (0.396 g, 1.68 mmol) and
1,8-bis(dimethylamino)naphthalene (0.750 g, 3.50 mnol) in 150 ml
methylene chloride (CHC12) was cooled to 0OC under nitrogen. Dropwise
addition of 2,3-methylenedioxybenzoyl chloride (56) (0.630 g, 3.41 mnol)
in 100 ml methylene chloride was completed over a one hour period, and
the reaction mixture was allowed to warm slowly to room temperature.
After 16 hours, the reaction mixture was cooled to 0C, washed with
3 x 10 ml ice-cold 3% (w/v) aqueous hydrogen chloride (HC1), and 3 x
10 ml ice water. The organic layer was then dried, filtered, and the
filtrate evaporated. The product was dissolved in a minimum amount
of methylene chloride and precipitated with several volumes of diethyl
ether yielding 0.902 g (94% crude yield) of the desired product as a
white, crystalline, hygroscopic solid.
An analytical sample was dissolved in methanol and sodium methoxide
added to a pH of 11. After stirring the mixture for 30 minutes, the
methanol was evaporated in vacuo. The residue was taken up in methylene
chloride and washed with water. The organic layer was then dried, fil-
tered, and the filtrate evaporated. Chromatography of the resulting
amine on silica gel, using 8% methanol/chloroform as the eluent afforded
a brown oil: 1H NMR: 6 1.30-2.00 (m, 6H), 2.20-2.73 (m, 4H), 3.10-3.68
(m, 4H), 3.82 (s, 2H), 5.78 (s, 2H), 5.85 (s, 2H), 6.57-7.63 (m, 13H).
Anal. calcd. for C3oH33N306: C, 67.78; H, 6.28; N, 7.90. Found:
C, 67.70; H, 6.25; N, 7.78.
A solution of N'-benzyl-N',N8-bis(2,3-methylenedioxybenzoyl)-
spermidine hydrochloride (3) (0.867 g, 1.53 mmol) in 45 ml glacial
acetic acid was prepared and palladium chloride (0.103 g, 0.581 mmol)
added. The reaction was allowed to proceed at room temperature until
hydrogen was no longer taken up. The mixture was then filtered, and
the filtrate evaporated. The residue was then dissolved in 35 ml
absolute methanol and the pH adjusted to pH 11 with sodium methoxide
and evaporated. The resulting solid was taken up in 25 ml methylene
chloride. This solution was washed with 2 x 25 ml cold water, dried,
and was filtered. The filtrate was then evaporated to give 0.660 g
(98% crude yield) of product, a brown oil.
An analytical sample was purified on silica gel, using 3% ammonium
hydroxide/chloroform as the eluent: 1H NMR: 6 1.40-2.02 (m, 7H), 2.42-
2.90 (m, 4H), 3.20-3.78 (m, 4H), 5.92 (s, 4H), 6.60-7.68 (m, 1H).
Anal. calcd. for C23H27N306-H0z: C, 60.25; H, 6.16; N, 9.17.
Found: C, 59.94; H, 5.96; N, 9.15.
Trifluoroacetic anhydride (4.46 g, 21.24 mmol) was added to a sus-
pension of N-(2-hydroxybenzoyl)glycine (0.402 g, 2.06 mmol) in 10 ml
methylene chloride, and the mixture was refluxed at 450C for 2 hours.
The solution was evaporated in vacuo and the N-(2-trifluoroacetoxy-
benzoyl)glycyl trifluoroacetic anhydride was redissolved in 15 ml
methylene chloride. Upon cooling this mixture to -78C, 1,8-bis(di-
methyldiamino)naphthalene (0.634 g, 2.96 mmol) in 10 ml methylene
chloride was added, followed by the dropwise addition of N1,N'-bis-
(2,3-methylenedioxybenzoyl)spermidine (0.653 g, 1.48 mmol) in 30 ml
methylene chloride. The reaction mixture was allowed to warm to room
temperature under nitrogen. After 45 hours, the solution was washed
with cold 3% (w/v) aqueous HCI, dried, filtered, and the filtrate
evaporated. The residue was then dissolved in methanol and the pH
adjusted to 9 by the addition of sodium methoxide. After stirring
the solution under nitrogen for 30 minutes, methanolic HC1 was added
at OOC to give a pH of approximately 3. The resulting solution was
filtered and the filtrate evaporated. The residue was chromatographed
on silica gel eluting with 5% methanol/ethyl acetate. This purification
procedure resulted in 816 mg (89% yield) of the product a white crys-
talline solid. 'H NMR: 6 1.30-2.30 (m, 6H), 3.42 (m, 8H), 4.17 (s, 2H),
6.00 (s, 4H), 6.53-7.77 (m, 13H), 11.99 (s, 1H).
Anal. calcd. for C32H34H409: C, 62.13; H, 5.54; N, 8.96. Found:
C, 61.95; H, 5.59; N, 8.90.
N4-[N- 2-Hydroxybenzoyl)glycyll-N ,N'-bis(2,3-dihydroxybenzoyl)spermi-
To a solution of N -[N-(2-hydroxybenzoyl)glycyl]-N',N'-bis(2,3-
methylenedioxybenzoyl)spermidine (5) (0.230 g, 0.370 mmol) in 20 ml
methylene chloride was added boron tribromide (0.50 ml, 5.29 mmol)
dropwise under nitrogen, and the reaction vessel was cooled to 0C.
The mixture was allowed to warm slowly to room temperature. After 22
hours, 20 ml cold water were added dropwise with vigorous stirring.
After stirring the mixture an additional 2 hours, the crude product
was collected by filtration, washed thoroughly with water, and was dis-
solved in methanol. The solvent was then evaporated to yield 0.210 g
(95% crude yield) of the desired product a white solid.
An analytical sample was preadsorbed on Sephadex LH-20 and eluted
with an ethanol/benzene gradient (5+50% v/v). 'H NMR (CDOCIl): 6 1.32-
3.28 (m, 6H), 3.04-3.72 (m, 8H), 4.22 (s, 2H), 6.48-8.12 (m, 15H),
12.10 (s, 1H), 12.74 (s, 1H), 13.09 (s, IH).
Anal. calcd. for CsOHs4N40s92H0O: C, 57.14; H, 6.07; N, 8.88.
Found: C, 57.20; H, 5.67; N, 8.77.
N'-Benzyl-N' ,N-bis(2,3-dimethoxybenzoyl)spermidine Hydrochloride ()
A solution of N'-benzylspermidine (8.60 g, 36.5 mmol) and tri-
ethylamine (8.87 g, 87.7 mmol) in 500 ml methylene chloride was stirred
at 0C under nitrogen. Dropwise addition of 2,3-dimethoxybenzoyl
chloride (57) (15.39 g, 76.7 mmol) in 100 ml methylene chloride was
completed over one hour, and the reaction mixture allowed to warm slowly
to room temperature. After 18 hours, the reaction mixture was cooled
to O0C and 3-dimethylaminopropyl amine (58) (10 ml, 79.5 mmol) in 100
ml methylene chloride was added slowly. After stirring the reaction
mixture for 2 hours, the reaction vessel was again cooled to 0OC, 150
ml ice-cold 3N HC1 added, and the mixture was stirred an additional
15 minutes. The organic phase was washed 3 x 100 ml with Ice-cold 3N
HC1, dried, filtered, and evaporated to 21.45 g (98% yield) of product-
a white, hygroscopic solid.
An analytical sample was dissolved in methanol and sodium methoxide
was added to obtain pH 11. After stirring this mixture for 30 minutes,
the mixture was evaporated in vacuo. The resulting solid was redis-
solved in methylene chloride, washed with water, dried, filtered, and
evaporated. Silica gel chromatography (10% methanol/chloroform) of
the resulting amine gave a tan oil. 'H NMR: 6 1.32-1.98 (m, 6H), 2.22-
2.68 (m, 4H), 3.13-3.82 (m, 6H), 3.83-3.90 (s, 12H), 6.78-8.05 (m, 13H).
Anal. calcd. for C32HiN306-H20: C, 66.07; H, 7.45; N, 7.22. Found:
C, 66.30; H, 7.18; N, 7.19.
Preparation and purification of (8) was in the same manner as (7)
(97% yield). 'H NMR: 6 1.55-2.03 (m, 4H), 2.27-2.72 (m, 4H), 3.22-
3.68 (m, 6H), 3.67 (s, 6H), 3.72 (s, 6H), 6.77-7.95 (m, 13H).
Anal. calcd. for C31H39N306: C, 67.74; H, 7.15; N, 7.64. Found:
C, 67.64; H, 7.20; N, 7.47.
Preparation and purification of (9) was in the same manner as (7)
(97% yield). 'H NMR: 6 1.48-1.85 (m, 8H), 2.20-2.65 (m, 4H), 3.08-3.63
(m, 6H), 3.77 (s, 12H), 6.78-8.02 (m, 13H).
Anal. calcd. for C3aH3sN3Os: C, 68.61; H, 7.50; N, 7.27. Found:
C, 68.43; H, 7.39; N, 7.12.
To a solution of (7) (7.31 g, 12.2 mmol) in 50 ml glacial acetic
acid was added palladium chloride (0.5 g, 2.8 mmol). The reaction was
stirred at room temperature until hydrogen was no longer taken up. The
reaction mixture was then filtered and evaporated, and the residue was
dissolved in 50 ml methanol. The solution was adjusted to pH 11 with
sodium methoxide. After stirring this mixture for 30 minutes, it was
evaporated in vacuo. The resulting solid was redissolved in 100 ml
methylene chloride and this solution was washed with 2 x 50 ml cold
water, dried, and filtered, and the filtrate evaporated to give 5.65 g
(98% yield) of desired product a light tan oil.
An analytical sample was purified by silica gel chromatography
(10% methanol/chloroform). 'H NMR: 6 1.38-2.02 (m, 7H), 2.45-2.87
(m, 4H), 3.18-3.70 (m, 4H), 3.83 (s, 12H), 6.78-8.37 (m, 8H).
Anal. calcd. for C25H35N306: C, 63.41; H, 7.45; N, 8.17. Found:
C, 63.55; H, 7.33; N, 8.89.
Preparation and purification was in the same manner as (10),
(98% yield). 'H NMR: 6 1.63-2.12 (m, 4H), 2.53-2.95 (m, 4H), 3.27-
3.83 (m, 5H), 3.85 (s, 12H), 6.90-8.25 (m, 8H).
Anal. calcd. for C24H33N306sH20: C, 60.36; H, 7.39; N, 8.80. Found:
C, 60.10; H, 7.29; N, 8.56.
Preparation and purification was in the same manner as (10), (99%
yield). 'H NMR: 6 1.30-1.87 (m, 8H), 2.30 (s, 1H), 2.48-2.82 (m, 4H),
3.10-3.60 (m, 4H), 3.80 (s, 12H), 6.73-8.33 (m, 8H).
Anal. calcd. for C26H37N306: C, 64.31; H, 7.27; N, 8.65. Found:
C, 64.46; H, 7.20; N, 8.86.
N'-[N-(2-Hydroxybenzoyl)glycyl ]-Nl,N"-bis(2,3-dimethoxybenzoy1 )sermi-
dine ( _)
Trifluoroacetic anhydride (5.17 g, 24.6 mnol) was added to a sus-
pension of 2-hydroxyhippuric acid (1.53 g, 7.9 mmol) in 35 ml methylene
chloride, and the resulting mixture stirred under nitrogen for 2 hours.
The solution was evaporated in vacuo and the N-(2-trifluoroacetoxyben-
zoyl)glycyl trifluoroacetic anhydride redissolved in 35 ml methylene
chloride. After cooling this mixture to -780C, triethylamine (2.0 g,
19.8 mmol) in 20 ml methylene chloride was added, followed by the drop-
wise addition of (10) (3.10 g, 6.5 mmol) in 25 ml methylene chloride.
The mixture was allowed to warm slowly to room temperature. After 40
hours, the reaction vessel was cooled to 0OC, and its contents were
washed with 3 x 30 ml ice cold 3% (w/v) aqueous HC1, dried, filtered,
and the filtrate evaporated. The residue was dissolved in 100 ml
methanol and the pH was adjusted to 9 with sodium methoxide under ni-
trogen. After stirring this mixture for 30 minutes, methanolic HC1
was added at 00C to obtain a pH of 3. The solution was evaporated
in vacuo. The residue was redissolved in 100 ml methylene chloride
and this solution was washed with 2 x 75 ml cold water, dried, filtered,
and the filtrate was evaporated. The residue was chromatographed on
silica gel (5% methanol /ethyl acetate) yielding 4.05 g (95% yield) of
the product a white solid. 'H NMR: 6 1.33-2.13 (m, 6H), 3.08-3.70
(m, 8H), 3.71-4.03 (m, 12H), 4.03-4.32 (d, 2H), 6.48-8.42 (m, 13H),
12.12 (s, 1H).
Anal. calcd. for C,3HiN409: C, 62.76; H, 6.51; N, 8.61. Found:
C, 62.62; H, 6.62; N, 8.55.
(3-aminopropyl amine (1A)
Preparation and purification was in the same manner as (13), (93%
yield). 'H NMR: 6 1.53-2.27 (m, 4H), 3.10-3.77 (m, 8H), 3.87 (d, 12H),
4.17 (s, 2H), 6.53-8.40 (m, 13H), 12.12 (s, lH).
Anal. calcd. for C33H4oN40s: C, 62.25; H, 6.33; N, 8.80. Found:
C, 62.06; H, 6.41; N, 8.70.
Preparation and purification was in the same manner as (13), (95%
yield). 'H NMR: 6 1.40-1.93 (m, 8H), 3.03-3.67 (m, 8H), 3.80 (s, 12H),
4.13 (s, 2H), 6.60-8.10 (m, 13H), 12.15 (s, 1H).
Anal. calcd. for C3sH4.N09: C, 63.24; H, 6.67; N, 8.43. Found:
C, 63.16; H, 6.72; N, 8.39.
Preparation and purification was in the same manner as (20),
(95% yield). 'H NMR (d6-acetone): 6 1.40-2.22 (m, 6H), 3.27-4.24
(m, 6H), 4.40-4.86 (s, 2H), 6.42-8.19 (m, 13H).
Anal. calcd. for Cz9H32N409: C, 59.99; H, 5.56; N, 9.65. Found:
C, 60.20; H, 5.67; N, 9.92.
Preparation and purification was in the same manner as (20), (94%
yield). 'H NMR (d6-acetone): 6 1.41-2.20 (m, 8H), 3.25-4.25 (m, 8H),
4.45-4.87 (s, 2H), 6.41-8.25 (m, 13H).
Anal. calcd. for CiHa36N40s: C, 61.18, H, 5.96; N, 9.21. Found:
C, 61.38; H, 6.02; N, 9.11.
2,3-Dimethoxybenzoyl Glycine (18)
Following the procedure of Van Brussel and Van Sumere (59), a mix-
ture of 2,3-dimethoxybenzoic acid (2.21 g, 12.1 mmol) and N-hydroxysuc-
cinimide (1.68 g, 14.6 mmol) in 35 ml dioxane was stirred at 150C under
nitrogen. Dropwise addition of dicyclohexylcarbodiimide (3.02 g, 14.6
mmol) in 30 ml dioxane was completed over 30 minutes, and the reaction
was allowed to warm slowly to room temperature. After 13 hours, the
reaction mixture was filtered and the precipitate was washed with 30 ml
of cold dioxane. A mixture of glycine (1.21 g, 16.1 mmol) and sodium
bicarbonate (1.35 g, 16.1 mmol) in 50 ml water was added to the filtrate
and the resulting mixture was stirred at room temperature. After 27
hours, the solvent was reduced to one third its original volume in vacuo.
Concentrated HC1 was added to this mixture at 0OC to a pH of 2 and the
resulting solid collected by filtration. Recrystallization of the solid
from water yielded 2.64 g (91% yield) of the desired product: m.p.
134.0-134.50C. 'H NMR (CD3OD): 6 3.83 (s, 3H), 3.88 (s, 3H), 4.10
(s, 2H), 6.95-7.63 (m, 3H).
Anal. calcd. for C1zHiaNOs: C, 55.23; H, 5.48; N, 5.89. Found:
C, 55.40; H, 5.61; N, 5.81.
To a solution of (18) (335 mg, 1.4 mmol) and (10) (616 mg, 1.3
mmol) in 15 ml dry methylene chloride under nitrogen was added DCC
(270 mg, 1.3 mmol) in 10 ml dry methylene chloride. After 48 hours
the reaction was filtered, the precipitate was washed with 5 ml methy-
lene chloride, and the filtrate was evaporated. Chromatography on
silica gel (chloroform/benzene/methanol) (10:10:1) afforded 723 mg
(80%) of (19) as a white solid. 3H NMR: 6 1.43-2.20 (m, 6H), 3.10-
3.68 (m, 8H), 3.83 (m, 18H), 4.23 (d, 2H), 6.77-9.17 (m, 12H).
Anal. calcd for C36sHsNOo4: C, 62.23; H, 6.67; N, 8.06. Found:
C, 62.09; H, 6.67; N, 7.96.
To a 1 M solution of boron tribromide in methylene chloride (19
ml, 19 mmol) and dry methylene chloride (30 ml) at 00C was added (19)
in methylene chloride (30 ml) dropwise, under nitrogen. The reaction
mixture was allowed to warm slowly to room temperature. After 14 hours,
the reaction vessel was cooled to 0%C, and ice cold water (40 ml) was
added slowly with vigorous stirring. After continued stirring of this
mixture for 2 hours, the crude product was collected by filtration and
washed with cold water. The resulting solid was purified on Sephadex
LH-20 (20% ethanol/benzene) yielding 0.79 g (90%) of product as a white
solid. 'H NMR (d6-acetone): 6 1.47-2.20 (m, 6H), 3.19-4.30 (m, 8H),
4.25-4.81 (s, 2H), 6.42-8.31 (m, 12H).
Anal. calcd. for CsoHa3N010o: C, 59.01; H, 5.61; N, 9.18. Found:
C, 58.90; H, 5.83; N, 9.01.
N',N'-bis(2,3-Dihydroxybenzoyl)bis-(4-aminobutyl)amine Hydrobromide (21)
To a solution of boron tribromide (6.39 g, 25.5 mmol) in 50 ml
methylene chloride at 0C was added (10) (1.13 g, 2.3 mnol) in 45 ml
methylene chloride dropwise under nitrogen. The mixture was allowed to
warm slowly to room temperature. After 15 hours, the reaction mixture
was cooled to OOC and 75 ml of ice-cold water was added slowly with
vigorous stirring. After continued stirring for 2 hours, the crude prod-
uct was collected by filtration and washed well with water and methylene
chloride. The resulting solid was crystallized from methanol to give
1.10 g (93% yield) of the desired product.
Anal. calcd. for C22H3oNs06Br: C, 51.57; H, 5.90; N, 8.20. Found:
C, 51.75; H, 6.0; N, 7.95.
NI,N7-bis(2,3-Dihydroxybenzoyl)bis-(3-aminopropyl)amine Hydrobromide (22)
Preparation and purification was in the same manner as (21), (93%
Anal. calcd. for C20H20N306Br: C, 49.60; H, 5.41; N, 8.68. Found:
C, 49.77; H, 5.40; N, 8.42.
NI,N8-bis(2,3-Dihydroxybenzoyl)spermidine Hydrobromide (23)
Preparation and purification was in the same manner as (21), (94%
yield). Spectral characteristics were identical to those reported in
the literature (31).
in i *-
ID 0 3
NOmmSsN m1 1A3114
NOGrw$r val trawla
, 1fffi R
.Os i.s D-I---
-^^ S,4 ^.g t||^^_ __^ "
. lisa |^la0e BI_^ __^ 8 .
F 7 tT v'-j
.. V7VV7^ ^^ "z "!"""
,"'i-I L-- '
L- .il .. -I ...^
"033 '.51101 1N11)11
",'n ha ,_,-.! Fe F '; I.-I.. "_ ." I I ,
L-iTr' -, T .... .... ...... .. -. ,, t
rr-- F. -T-.--
5 o t1
---. ----, ----- -..-.-. -
h; is;. F _:-IE^ ^'Z^ DI i
aaf~i~a~SM^II- -- I-
in I iw .... .
U p 0 7 7 F1]; y4 12 E ^ r ~ ~ "
II SK -7k
ph q JI.LT a g. I
>! : :
I kL .LLJLALL I I i i
8 ~ S9" 0 0
0SSI N V S I Y I 1N S)fld
i o= '
NOISSMI SN I :Mn
-.1-' 15, ir-ll _ti ,
" I I I S| 1' ';5 i
NOISMSMbI Jl Ini3 a
N0O iwIsW va iN3id
SA lit A5'01 J
.. t. -i.I
0 N-_- ,-- .
^ .- _^ .(__- -
: ll**;;:.fi!1 ~ i _
l~tI.I;4o 7t.- h!t-utihtltt i 'I .T
8 2 g o 0
I j S
S.B- I E LAE I I Lt 1 [ I I II
'- ''''11111-1 1?1 I I IU
OCTADENTATE CATECHOLAMIDE LIGANDS
AS ACTINIDE CHELATORS
The potential biological hazards associated with the nuclear fuel
cycle and nuclear military systems has rapidly increased along with
their development. Unfortunately, the methods fordeallngwith such po-
tential radiation hazards have not paralleled the expanding nuclear
industry. Of the actinides used in nuclear systems plutonium is the
element associated with the greatest potential for lethal radiation
damage. Plutonium is one of the most toxic substances known (43,46).
The high toxicity of plutonium in man is a result of both its radioac-
tivity and its similarity to iron (60,61). Although plutonium is able
to exist in each of the oxidation states from III to VI in aqueous solu-
tion (43,60), it is believed that the metal exists almost exclusively
as plutonium (IV) at physiological pH in vivo. The charge to radius
ratio of plutonium (IV) is quite similar to that of iron(III) (444 to
460 e/pm) (62). And, as expected, much of the chemistry of the two
metals is very similar, including their coordination chemistry. Both
Fe(III) and Pu(IV) are considered to be hard acids (46). The metals
both display a high positive charge, small ionic size, and high acidity.
In fact, the similarities between plutonium(IV) and iron(III) are so
great the body cannot effectively discriminate between the two metals.
The coordination chemistry of the two metals is similar enough that
many of the bioorganic ligands present in the body that are normally
quite specific for iron(III) will also chelate plutonium(IV) (63). Like
Fe(III), Pu(IV) is bound by the human serum transport protein transfer-
rin (64). Transferrin shuttles the plutonium throughout the body, stor-
ing a large amount of the metal in the parenchymal cells of the liver
where it is bound to the iron storage protein ferritin (65,66). Once
plutonium is absorbed by the body, a large amount of the actinide is
transported to the bone marrow and the reticuloendothelial tissue of
the spleen (46). Like Fe(III), once Pu(IV) makes its way into the deep
iron storage sites of the body, it is very difficult to remove (43,46).
The current treatment for exposure to plutonium involves the ad-
- ministration of the chelating agent DTPA (Figure38) (67,68). Although
DTPA has been shown to facilitate the excretion of plutonium if the
chelator is administered very shortly after exposure to the actinide,
it is not a very specific ligand for plutonium (69,70). The agent DTPA
chelates and facilitates the excretion of a large number of the metal
ions found in the body, including cobalt, magnesium, and zinc. In fact,
it is the tendency of DTPA to chelate and clear zinc that is responsible
for the toxic effects when DTPA is administered over a long period of
time (69,70). Clearly, there exists the need for the development of a
more specific sequester of Pu(IV) to treat individuals that have been
exposed to the transuranium elements such as plutonium.
Based on the similarities of plutonium(IV) and iron(III), it would
appear that catecholamide ligands might be able to form stable complexes
with the metal. Catecholamides have indeed been shown to chelate Pu(IV)
and other actinides very effectively (43-48). However, in order to ef-
fect the specific chelation of plutonium(IV), a ligand of the appropriate
denticity and coordination geometry would have to be constructed. Unlike
HOOCCH2 \ /CH2COOH
Figure 38. Structure of Diethylenetriaminepentate Acetic
Fe(III), which forms octahedral coordination compounds with catechol-
amides, the actinide(IV) catecholate complexes have been reported to
exist in the geometry of a trigonal-faced dodecahedron (44).
Several octadentate catecholamide ligands have recently been syn-
thesized by Raymond that have both cyclic and linear tetraamine back-
bones (45,46). It was found that the synthetic catecholamide ligands
which possessed a large degree of freedom in their polyamine platforms
were more effective chelating agents (45). Thus, the linear octaden-
tate catecholamides for example, N.NS,NsN,N1 -tetra(2,3-dihydroxy-
benzoyl)tetraazatetradecane (LICAM) were judged better actinide che-
lators than the catecholamides with cyclic platforms (45).
With the above information in mind, the synthesis of the first of
a novel class of octadentate catecholamide ligands was undertaken.
This new class of catecholamide ligands is referred to as octadentate
"H-shaped" ligands, referring to the shape of the polyamine backbone
of the catecholamides (Figure 39). Each of the four arms of the "H-
shaped" ligands has a 2,3-dihydroxybenzoyl group attached to it.
It is hoped that the new octadentate "H-shaped" ligands will have the
flexibility of the linear octadentate chelators, but also possess the
ability to encapsulate the metal without a large degree of internal
rotation, something that the simple linear octadentate ligands cannot
The development of this new class of octadentate catecholamides
was undertaken in the hope that systems capable of environmental decon-
tamination as well as biological decontamination could be achieved.
O OH HH
--NH-(CH) -- --(CH)b--NH--C
--H--(CH,)I N--(CHI) -NH-CO
Figure 39. Structure of the First of a New Class of Catecholamide
Octadentate Ligands (PLUTO).
The synthesis of the "H-shaped" catecholamide ligands was designed
such that the length of each of the four arms hooking the four cate-
cholate groups together, as well as the central tether linking the two
spermidine backbones of the ligands, could be shortened or lengthened
(Figure 40). In this way, it is hoped that the geometry of the ligands
can be tailored to allow for the preferential chelation of one actinide
over other metals.
The synthesis of the first octadentate "H-shaped" ligand was ac-
complished by linking together two molecules of N',N6-bis(2,3-dimethoxy-
benzoyl)spermidine with glutaryl dichloride in high yield. The methyl
protecting groups were then removed with BBr3 yielding the free octa-
dentate catecholamide again in high yield.
Of course, the synthetic scheme can be applied to the homo and nor
homologs of N',N'-bis(2,3-dimethoxybenzoyl)spermidine, and by using the
acid chloride of any dicarboxylic acid oxalicc, succinic, glutaric,
etc.), a large number of octadentate catecholamides can be synthesized.
From the above description it can be seen that a number of octadentate
ligands, in which both of the spermidine backbones present in the mole-
cule are the same length, could be synthesized. It will also be pos-
sible to synthesize catecholamide ligands in which each of the two poly-
amine chains in the ligand are of a different length. For example, by
first acylating the homo derivative of N',N'-bis(2,3-dimethoxybenzoyl)-
spermidine with glutaric anhydride, followed by linking the resulting
monoglutaramide with the nor derivative of N',N'-bis(2,3-dimethoxy-
benzoyl)spermidine, a ligand with both a nor and homo spermidine back-
bone can be prepared. It will be possible to eventually prepare a large
2 j 01j:I
CO'H -(Cll. ) --;:. -( C17 ), --HI;:0
OCH N CH0.O
00a3 C 0
OH--(CH:),- -1 -(CH,)b -HO
(O.--( )- H .)
Soc.. I 11 0
C; 1--(Cl)Y--li--(C1i )b--;81CO
Figure 40. Synthesis of Catecholamide "H-shaped Ligands.
number of octadentate "H-shaped" catecholamide ligands that possess
slightly different dimensions from one another. It is hoped that by
lengthening or shortening the various appendages of the ligand, it will
be possible to prepare ligands that specifically chelate plutonium over
any other competing metal ions that may be present. Furthermore, the
synthetic scheme described offers an opportunity to covalently attach
an octadentate catecholamide ligand to a polymeric support. For exam-
ple, N-(carbobenzyloxy)glutamic acid could be used to link two mole-
cules of NI,N'-bis(2,3-dimethoxybenzoyl)spermidine together. After re-
moval of the carbobenzyloxy protecting group, the free amino group could
serve as a synthetic handle to covalently link the ligand to a matrix.
Removal of the methyl protecting groups of the ligand that is now at-
tached to the resin would generate a polymeric form of an octadentate
catecholamide ligand. Such a device may be useful in the decontamination
of actinide-tainted water.
Materials and Methods
All reagents were purchased from Aldrich Chemical Co. and, except
where noted, used without further purification. Unless otherwise
specified, sodium sulfate was used as a drying agent. Resins (20-50
mesh) were supplied by Rohm and Haas. Melting points were taken on a
Thomas-Hoover apparatus and are uncorrected. Boiling points are also
uncorrected. Unless otherwise indicated, 'H NMR chemical shifts are
given in parts per million downfield from an internal tetramethylsilane
standard. The spectra were recorded on a Varian T-60 spectrometer.
Elemental analyses were performed by Galbraith Laboratories. Preparative
thin-layer chromatography was done on Analtech 20 x 20 cm silica gel
A mixture of NI,N'-bis(2,3-dimethoxybenzoyl)spermidine (1.00 g,
2.11 mmol) and triethylamine (0.64 g, 6.33 mmol) in CH2C12 (40 ml) was
cooled to 0OC under nitrogen. A solution of glutaryl dichloride (0.16
g, 0.95 mmol) in CHzCI2 (20 ml) was added dropwise over 10 minutes.
After 16 hours, the mixture was washed with 3% (w/v) aqueous HCI (3 x
30 ml) and water (3 x 30 ml); the organic phase was then dried, fil-
tered, and the filtrate evaporated. Purification by silica gel chroma-
tography (10% methanol/ethyl acetate) yielded 0.95 g (96%) of product -
a white solid. 'H NMR (CDC13): 6 1.38-2.11 (14H), 2.18-2.51 (4H),
3.01-3.61 (16H), 3.64-3.98 (24H), 6.74-8.31 (16H).
Anal. calcd. for Cs5H74N6014: C, 63.32; H, 7.15; N, 8.06. Found:
C, 63.09; H, 7.12; N, 7.97.
To a 1 M solution of BBr3 in CH2C12 (25.0 ml, 25 mmol) and dry
CH2C12 (30 ml) at 0OC was added bis-[N',Na-bis(2,3-dimethoxybenzoyl)-
spermidine]-glutaramide (1.05 g, 1.01 mmol) in CH2C12 (30 ml) dropwise
under nitrogen. The reaction mixture was allowed to warm slowly to
room temperature. After 16 hours, the reaction vessel was cooled to
00C, and ice cold water (40 ml) was added slowly to the mixture with
vigorous stirring. After continued stirring of the mixture for 2 hours,
it was filtered and the resulting solid washed with cold water. The
solid was dissolved in methanol and then evaporated, and the process re-
peated several times. Purification on Sephadex LH-20 (20% ethanol/
benzene) yielded 0.8 g (85%) of product a white solid. 'H NMR d6-acetone:
6 1.35-2.18 (14H), 2.25-2.75 (4H), 3.02-3.75 (16H), 6.45-7.42 (12H),
Anal. calcd. for C47HsaeNeO12H20: C, 58.38; H, 6.46; N, 8.69.
Found: C, 58.54; H, 6.55; N, 8.43.
The catecholamides N4-[N-(2,3-dihydroxybenzoyl)butyryl]-N,Ne-bis-
(2,3-dihydroxybenzoyl)spermidine (30) (GABA), PLUTO, and LICAM (LICAM
was supplied by Dr. Kenneth Raymond, Department of Chemistry, University
of California, Berkeley) were tested for their ability to sequester plu-
tonium(IV) and americium(III) from aqueous solutions after the ligands
had been adsorbed on macroreticular resins. These determinations were
performed by Dr. James Navratil's group at Rockwell International,
Golden, Colorado using the following procedures.
The solubilities of the catecholamides in various reagent grade
solvents were tested by mixing 50 mg of each catecholamide with 10 ml
of solvent at ambient temperatures (22-240C). The samples were visually
inspected after one hour. Samples showing no dissolution were left
overnight and reinspected the next day.
The catecholamides were dissolved in methanol and mixed with a
macroreticular resin. The XAD resins are nonionic macroreticular poly-
styrene-divinyl benzene materials. The methanol was allowed to slowly
evaporate. The loaded resin was washed with -20 column volumes of
distilled water and air-dried. The amount of ligand absorbed by the
resin was determined by the weight differences of the resin before and
after contact with the ligand.
For the study of sequestering of americium and plutonium, two
aqueous solutions were prepared: synthetic buffered solutions and actual
Rocky Flats plutonium process waste adjusted to various pH levels. Stan-
dard buffers were prepared according to the National Bureau of Standards
procedure, and plutonium and/or americium were added from purified, con-
centrated actinide stock solutions in dilute HC1 to prepare the syn-
thetic waste samples. Process waste samples were adjusted to appropriate
pH with 12 M nitric acid (HNO3). All the solutions were filtered through
Whatman 42 paper prior to analysis.
Weighed amounts of XAD-4 macroreticular resin containing the octa-
coordinate catecholamide (PLUTO), were added to 10 ml of buffered or
waste solution and equilibrated overnight on a rotary mixer. After the
solution was separated from the resin by filtering through Whatman 42
paper, it was analyzed for americium and plutonium. In the americium
analysis by gamma spectroscopy ona Canberra 80 multichannel analyzer,
the bulk aqueous sample was run before and after contact with the se-
quester. This minimized sampling error. Precision and accuracy were de-
termined by running standards with samples. Three readings of either
sample or standard were made. The average precision is 3%, while the
average accuracy of the standard with independently determined value
is 10%. Plutonium was analyzed on a Nuclear Measurement Corporation
proportional counter with an argon/methane atmosphere. Aliquots of 250
ul from either a standard or the test solution were affixed to stainless
steel planchets and run. For low steel (<10-6 g/l) samples, 1000 ul
aliquots were used. The precision and accuracy from simultaneously
run standards are 6% and 9%, respectively.
Results and Discussion
It was originally desired to examine the ability of the catechol-
amides to sequester the actinides plutonium and americium by partitioning
the ligands between an organic solvent and aqueous solutions containing
the actinides. However, the synthetic catecholamides were found to be
generally quite insoluble in water-imiscible organic solvents such as
carbon tetrachloride, diisopropylbenzene or kerosene. Only LICAM and
GABA were found to be somewhat soluble in octanol. Of these two ligands,
only GABA was judged soluble enough in both water and octanol to allow
a partition experiment to be conducted.
GABA was partitioned between a solution of octanol and an aqueous
process waste solution at pH 6.6 which contained plutonium and americium.
The decontamination factors, which represent the ratio of the initial
actinide concentration to that concentration after contact with the
ligand, for GABA in the partition experiment are presented in Table 1.
The ability of the hexadentate catecholamide ligand, GABA, to chelate
plutonium(IV) is evident by the fairly large plutonium decontamination
factor of 390, which corresponds to a removal of 97% of the plutonium
from the aqueous phase. In addition to being a good sequester of plu-
tonium, it can be seen that GABA shows a large degree of specificity
towards the chelation of plutonium over americium. The americium de-
contamination factor for the GABA partition experiment was only 18,
which corresponds to only 29% of the americium removed from the aque-
Since the octadentate catecholamides were not soluble in water-
imiscible solvents, partition experiments could not be conducted. In-
stead, the ligands were adsorbed onto XAD-4 resin, and the resin washed
with an aqueous solution of the actinides. Table I compares the plu-
tonium and americium decontamination factors for resins coated with
the hexadentate ligand GABA, and the octadentate ligand LICAM. As
Preliminary Adsorption Study Sequestering of Am and
Pu from Filtered Waste at pH 6.6
Waste pH 6.6
5.4 x 10-4 g/k Pu
9.8 x 10-6 g/L Pu
2.9 x 10-5 g/s Am
1.4 x 10-7 g/z Am
Pu g/i in Raffinate
3.3 x 10-7
7.2 x 10-'
1.3 x 10-6
Am g/t in Raffinate
9.9 x 10-8
9.6 x 10-8
1.1 x 10-7
Per G. Liaand
Per G. Liand
Decon. Factor = Actinide Content in Feed /t. ligand
Actinide Content in Raffinate)/w
expected, the octadentate catecholamide chelator LICAM displayed a greater
plutonium decontamination factor than the hexadentate catecholamide
ligand GABA. This is understandable since plutonium(IV) is an octacoor-
dinate metal ion. The americium decontamination factors for the two
ligands were both rather low compared to the plutonium decontamination
factors. It should be pointed out that the reason for the much smaller
decontamination factors obtained when the ligands are coated onto the
resin compared to the octanol-water partition experiment is due to the
large amount of ligand used in the resin experiments. It has been de-
termined that by using less ligand, larger decontamination values are
obtained in the resin experiments.
Table 2 shows the data from a study examining the ability of XAD
resin coated with PLUTO to remove plutonium and americium from two dif-
ferent types of actinide solutions as a function of pH. As can be seen
from the table, the optimum pH region for the removal of both actinides
is in the pH range 6 to 7. While the plutonium removal was greatest at
pH 6 in process waste solutions, the removal of plutonium from standard
buffer solutions was quite effective over a large pH range of 2 to 6.
It should be noted that, although the plutonium and americium decontami-
nation factors appear to vary considerably in some instances, the actual
difference in the amount of metal removed may be quite small. For ex-
ample, when the standard americium buffer at pH 6 is exposed to the
PLUTO-XAD resins system, a plutonium decontamination factor of 677 is
obtained. This decontamination factor corresponds to the removal of
94% of the plutonium from the standard solution. At pH 7 a substantially
smaller americium decontamination factor of 300 is obtained. However,
a decontamination factor here of 300 means that 94% of the metal has been
removed from the buffer, the same amount removed at pH 6.
Sequestering of Plutonium and Americium from Aqueous Solutions
with PLUTO (octadentate catecholamide) on XAD-4 Resin
2 Buffer d
1.6 x 104
1.6 x 10
1.1 x 10 7
2.1 x 10
1.4 x 10 4
1.7 x 10
9.1 x 107
3.1 x 10
1.2 x 10 4
2.9 x 10
8.5 x 10_6
2.4 x 10
1.7 x 10 4
1.7 x 10-
1.2 x 10 5
1.2 x 10
1.8 x 10 4
1.8 x 10
1.1 x 10 6
3.3 x 10
1.7 x 10 4
1.6 x 10
1.5 x 10_7
1.5 x 10
1.5 x 10
1.1 x 10
1.5 x 10 7
1.5 x 10-
1.5 x 10 9
8.6 x 10
1.5 x 10_7
2.7 x 10
1.2 x 10 8
7.6 x 10
1.5 x 10 7
1.6 x 10
1.2 x 10 9
6.6 x 10
1.4 x 10_7
1.5 x 10
1.3 x 10_8
5.1 x 10
1.6 x 10_7
1.5 x 10
a. Error: Pu 6%, Am 3%
b. Decontamination Factor =
Actinide Conc. in Feed )/t PLUTO
Actinide Conc. in Raffinate)/wt PLUTO
D.F. Error: Pu 12%, Am 6%
Waste-II: Waste spiked with Pu and Am after first filtration and refil-
tered after pH adjustment
Factor (per g PLUTO) ,c
As opposed to the standard plutonium buffer solutions, when process
waste is treated with the PLUTO-XAD resin system, there is a signifi-
cantly different amount of plutonium removed at pH 2 than at pH 6.
At pH 6 the plutonium decontamination factor of 300 corresponds to a
removal of 95% of the metal, while at pH 2 the decontamination factor
of 28 indicates that only 44% of the plutonium has been removed from the
process waste solution.
Like plutonium, the greatest amount of americium removed from the
two solutions by the PLUTO-XAD system occurred at pH 6-7. With few ex-
ceptions, the ability of the PLUTO-XAD resin system to remove americium
from these solutions represents a bell-shaped curve where the removal
of metal at very high or very low pH values was substantially poorer
than at pH values around neutrality. For example, 81% of the americium
in the standard buffer solution was removed by the PLUTO-XAD resin
system at pH 7, while only 2% and 44% are removed at pH values of 2
and 12, respectively.
It should be mentioned that the results of Table 2 represent ex-
periments done on two different batches of PLUTO. The even numbered
pH series was done on the first batch of the ligand while the odd num-
bered pH series was conducted on a different batch of the ligand. It
is possible that some of the differences observed among the data are
due to batch differences. However, the differences may be also a re-
sult of metal and pH-dependent hydrolysis of plutonium and americium
at pH values near neutrality.
Despite the uncertainties regarding the nature of the metal ion
species present in the solutions at different pH values, PLUTO has been
shown to be effective in removing americium and plutonium from aqueous
solutions of the actinides. These results suggest that the octadentate
catecholamide H-shaped ligands may be useful in the removal of actinides
from patients exposed to the metals, as well as being effective devices
to serve in environmental actinide decontamination.
SYNTHESIS AND SOLUTION CONFORMATION OF
PARABACTIN AND ITS GALLIUM(III) CHELATE
In 1975 Tait isolated a catecholamine iron chelator, which he re-
ferred to as "Compound III," from iron-depressed cultures of Paracoccus
denitrificans (19). Compound III was shown to consist of a spermidine
backbone with a 2,3-dihydroxybenzoyl group at each of the two terminal
nitrogen atoms. The central N4 position of the spermidine chain was
proposed to be acylated with an N-salicyl-L-threonine moiety. Later,
Neilands was able to show that Tait's original proposed structure of
compound III was incorrect (21). Based on the ability of compound III
to form a hydrochloride salt with a concomitant red shift in its ultra-
violet absorption spectrum, which is characteristic of oxazoline ring
systems, it was concluded that the proposed salicyl threonine group was
actually a 2-hydroxyphenyl-threonyloxazoline ring (21). The true struc-
ture of the natural catecholamide Tait has isolated is now referred to
as parabactin A (Figure 3).
Since the isolation of parabactin, a great deal of interest has
been focused on the biological and physical properties of this sidero-
phore and its ferric chelate (19,21,41). It has recently been deter-
mined that the formation constant for the iron(III) complex of parabac-
tin at physiological pH is 1048, making parabactin one of the strongest
ferric ion chelators known to date (71). The coordination of iron(III)
by parabactin is believed to take place with the siderophore encapsulating
the metal ion to form an octahedral coordination complex (72). Unlike
some other catecholamide siderophores, parabactin chelates iron(III) with
oxygen as well as nitrogen atoms serving as the lighting groups. Para-
bactin is believed to chelate ferric ion via the five phenolic oxygens,
as well as the oxazoline ring nitrogen of the threonyloxazoline ring
In addition to the physical properties of the metal chelate of para-
bactin, the properties of the free ligand, mainly its solution conforma-
tion, have been the subject of considerable interest (32,41). A high
field 1H NMR study of parabactin and agrobactin, a siderophore isolated
from Agrobacterium tumefaciens (Figure 2), was undertaken which provided
some insight about the solution behavior of these siderophores (41).
It was observed that the peaks in the spectra of the siderophores were
present in duplicate and that, upon heating, the duplicate signals co-
alesced into single peaks. Two theories were offered to explain the
behavior of the NMR spectra of parabactin and agrobactin. The first
involved a rotation of the oxazoline ring moiety about the alpha-carbon
and carbonyl-carbon single bond that results in two interconvertible
conformers of the siderophore (73). The second explanation suggested
that the duplicate signals were a result of a simple cis-trans isomeri-
zation of the N' amide bond (41).
Part of the reason the solution conformation of the catecholamide
siderophores has received so much attention is because of the exception-
ally high formation constants of their metal chelates. It was thought
that one of the reasons for the large ferric ion formation constants of
these ligands may be due to a similar solution conformation of both
ligand and chelate. Another reason for studying the solution behavior
of the siderophores in general is concerned with the role these iron
chelating compounds play in microbial iron transport.
Evidence has been presented recently that shows certain microorgan-
isms have the ability to discriminate between various ferric siderophore
chelates (74-81). While some ferric chelates are transported and uti-
lized by a microorganism, others are not presumably because receptor
proteins on the outer membrane surface of the microorganism can only re-
cognize certain siderophore iron complexes (77,80). The exact nature
of how microorganisms can discriminate between ferric siderophore com-
plexes is unclear. It is believed that recognition may depend on the
configuration of various groups of the ligand or, perhaps, on the con-
figuration of the chelating groups about the metal center itself (12,
It was desired to ultimately evaluate the specificity of the ferric
parabactin iron transport system in Paracoccus denitrificans. However,
before the specificity of the microbial iron uptake system of Paracoccus
denitrificans could be understood, the solution conformation and stereo-
chemistry of the complex formed between parabactin and trivalent metal
ions needed to be examined in more detail. This was accomplished via
300 MHz 'H NMR spectroscopy. To avoid line broadening by the paramag-
netic ferric ion, gallium(III) was used to examine the parabactin metal
All reagents, with the exception of (L)-N-tert-butoxycarbonylthreo-
nine (Sigma Chemical Co.) and gallium(III)nitrate-9 H20 (Alfa Co.) were
purchased from Aldrich Chemical Co. and were used without further puri-
cation. Sodium sulfate was used as a drying agent. Melting points were
taken on a Fisher-Johns apparatus and are uncorrected. Preparative thin-
layer chromatography was done on 20 x 20 cm silica gel plates obtained
from Analtech Co. Sephadex LH-20 was purchased from Pharmacia Fine
Chemicals Co. Optical rotations were measured with a Perkin-Elmer model
141 polarimeter. Elemental analyses were performed by Galbraith Labora-
tories, Knoxville, Tenn., or Atlantic Microlab Inc., Atlanta, Ga. Pro-
ton NMR spectra were obtained on a Nicolet Instrument Corp. NT-300 spec-
trometer and NIC-1180 E data system. Probe temperature was determined
using test samples of ethylene glycol. Computer simulation and curve
analysis/deconvolution programs used were included in the NMCFT software
package provided by Nicolet Technology Corp. Resolution enhancement,
when necessary, was performed by apodization of the FID by a double ex-
ponential multiplication followed by zero filling. Samples for the de-
termination of temperature of coalescence were prepared by dissolving
5-10 mg of the compound in 500 pl d6-DMSO. Coalescence temperatures,
1C, were then determined by observing the coalescence of the beta-
decoupled gamma methyl signals. The coalescence temperatures were meas-
ured both on heating and cooling cycles and activation energies, 0.2
kcal/mol, subsequently calculated by the method of Gutowsky and Cheng
(82). Chemical shifts are reported downfield from an external DSS stan-
dard. By dissolving a sample of 5 mg of parabactin in 500 pl d-chloro-
form (CDCI3) in a 5 mm O.D. NMR tube, d6-DMSO/CDCI3 titrations were
carried out. Additions of d6-DMSO were made directly into the NMR tube
via a microliter syringe, and spectra recorded after mixing the sample.
In this manner, spectra were recorded at d6-DMSO concentrations (volume
percent) of 0%, 2%, 5%, 20%, 30%, 40%, and 50%. pH measurements were
obtained with an Ingold microelectrode. Gallium chelates were prepared
by adding a slight excess of gallium(III) nitrate to an aqueous solution
of the catecholamide at -pH 10 and immediately adjusting the pH to -7.4
under N2. Water was removed in vacuo after allowing the mixture to stir
overnight at room temperature.
To a solution of (L)-N-tert-butoxycarbonylthreonine (2.96 g, 13.5
mmol) and N-hydroxysuccinimide (1.63 g, 14.2 mmol) in dry tetrahydro-
furan (THF) (100 ml) at 0OC was added DCC (2.97 g, 14.4 mmol) in THF
(100 ml). The mixture was allowed to warm slowly to room temperature
with continued stirring under N2. After 14 hours, the mixture was
filtered and the DCU washed with THF (25 ml). The filtrate was evaporated
and the residue crystallized from diethyl ether to yield 3.84 g (90%)
of (26) as white crystals: mp 134-135C, [a] -33.7 0.7 (C=2.7 EtOAc)
(lit. mp 134-135C, [a]D 33.3 (EtOAc)) (83); 'H NMR (CDC13): 6 1.25
(d, 3H), 1.47 (s, 9H), 2.78 (s, 4H), 3.89-4.98 (m, 3H), 6.38 (m, 1H).
(L)-N -[N-tert-Butoxycarbonylthreonyl ]-N1,N'-bis(2,3-dimethoxybenzoyl)-
A solution of NI,N8-bis(2,3-dimethoxybenzoyl)spermidine (10) 2.50
g, 5.28 mmol) in DMF (150 ml) was cooled to 00C. A solution of (26)
(1.75 g, 5.54 mmol) in DMF (50 ml) was added at once and the mixture
allowed to warm slowly to room temperature. After 48 hours the solvent
was evaporated. The residue was dissolved in CH2C12 (100 ml), washed
with cold 3% (w/v) aqueous HC1 (3 x 30 ml) and cold water (10 x 30 ml).
The organic phase was then dried, filtered, and the filtrate evaporated.
Silica gel chromatography (5% methanol/ethyl acetate) yielded (27) as a
hygroscopic, white solid: 2.99 g (84%); 'H NMR (CDCl3): 6 1.13 (d, 3H),
1.40 (s, 9H), 1.50-2.13 (m, 6H), 3.07-3.63 (m, 8H), 3.83 (s, 12H), 4.17-
4.57 (m, 3H), 5.33-5.60 (d, 1H), 6.77-8.27 (m, 8H).
Anal. calcd. for C30HsoN401o: C, 60.52; H, 7.47; N, 8.30. Found:
C, 60.60; H, 7.38; N, 8.45.
A solution of (27) (1.70 g, 0.25 mmol) in trifluoroacetic acid
(TFA) (40 ml) was stirred at room temperature for 35 min. The solvent
was then evaporated and the residue dissolved in cold CHICl2 (100 ml)
and washed with ice cold 30% (w/v) aqueous sodium carbonate (3 x 50 ml).
The organic phase was then dried, filtered, and the filtrate evaporated.
Purification on silica gel (10% methanol/chloroform) provided 1.38 g
(95%) of (28) as a hygroscopic, white solid. 'H NMR (CDCls): 6 1.10-
1.20 (d, 3H), 1.43-2.00 (m, 6H), 2.94-4.08 (25H), 6.84-8.25 (m, 8H).
Anal. calcd. for C29H42N408: C, 60.61; H, 7.37; N, 9.75. Found:
C, 60.50; H, 7.36; N, 9.62.
L)-N -Threonyl-NL,Ne-bis(2,3-dihydroxybenzoyl)spermidine Hydrobromide
To a 1 M stirred solution of BBr3 (20 ml, 20.0 mmol) in dry CH2CIz
(30 ml) at 0OC was added (28) (0.83 g, 1.44 mmol) in CH2Cl~ (30 ml) drop-
wise under nitrogen. The reaction mixture was allowed to warm slowly
to room temperature. After 12 hours the reaction vessel was cooled to
0%C, and ice cold water (15 ml) was added dropwise with vigorous stir-
ring. The resulting suspension was allowed to warm to room temperature
with continued stirring over one hour and the product collected by fil-
tration. The residue was dissolved in methanol and evaporated, this
process being repeated several times. Chromatography on Sephadex LH-
20 (20+40% ethanol/benzene) gave 0.78 g (90%) of (29) as a white solid.
'H NMR (CD3OD): 6 1.10-1.37 (d, 3H), 1.43-2.23 (m, 6H), 3.13-3.80 (m, 8H),
3.83-4.40 (m, 3H), 6.33-7.33 (m, 6H).
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E1AUCKU8C_NVB64I INGEST_TIME 2012-09-24T13:55:12Z PACKAGE AA00011825_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC